INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The expression of protein-coding genes in eukaryotes depends on DNA-binding transcription factors that typically have at least two distinct domains: one domain responsible for specific DNA recognition and one responsible for transcriptional activation (5, 27). According to their predominant amino acid composition, activation domains have been classified mainly into acidic, proline-rich, and glutamine-rich domains (for reviews, see references 28 and 40). When tethered to DNA, these classes of activation domains possess different biological properties in their ability to influence gene expression (38). Acidic activation domains rich in acidic and hydrophobic amino acids, such as those found in VP16 or Gal4p proteins, are the most versatile activators, and they stimulate transcription when bound to DNA from proximal and distal enhancer positions in all eukaryotes. Proline-rich activation domains, e.g., those of AP-1 or CTF/NF1, generally activate from proximal and, albeit to a much reduced extent, from distal positions. Glutamine-rich activators like Sp1 or Oct1 on their own fail to activate transcription from a remote enhancer position, but they stimulate gene expression in response to remote enhancers when bound in close proximity to the TATA box in human cells (38). The same glutamine-rich activation domains were reported to be active in the fission yeast Schizosaccharomyces pombe when tethered to a proximal regulatory sequence (35). In contrast, with the budding yeast Saccharomyces cerevisiae, we and others have reported that glutamine-rich activation domains of the mammalian transcription factors Sp1, Oct1, and Oct2 or full-length Sp1, when tethered to DNA, are transcriptionally inactive at either proximal or remote position when tested on their own (24, 32).

In order to stimulate transcription, activation domains have to interact with components of the transcriptional complex (for reviews, see references 2 and 30). For Sp1, it has been shown that the two glutamine-rich activation domains directly interact with TAF_II110, a component of the general transcription machinery (20). TAF_II110 is present in higher eukaryotes (e.g., as dTAF_II110 in Drosophila and as hTAF_II130 in humans), but no homologue could be found in the genome of the budding yeast S. cerevisiae. This finding offered a straightforward explanation for the seeming inability of the Sp1 glutamine-rich domains to activate transcription in the budding yeast.

Another important observation is that the typical arrangement of regulatory sequences controlling gene expression in yeast differs from that of higher eukaryotes. In metazoans, binding sites for transcription factors are often found in close proximity to the transcriptional start site and also at a considerable distance. Proximal sites may contain binding sites for Sp1 and/or Oct1 (22). Distal enhancer elements can influence gene expression when positioned upstream, downstream, or even as part of an intron within the transcription unit (1; for a review, see reference 37). In the budding yeast, probably due to space constraints and the almost complete lack of introns, the majority of genes are controlled by a few binding sites for transcription factors, termed upstream activating sequences, which are located close to the TATA box.

To see whether enhancer or promoter structure could influence the activity of glutamine-rich domains in the budding yeast, we reconstituted a metazoan-like regulatory structure in the yeast chromosomal context by introducing transcription factor binding sites both in close proximity to the TATA box and at a remote upstream enhancer position. Under these conditions, we observed that the glutamine-rich activation domains of mammalian Sp1 or Oct1 and yeast Snf5p readily contributed to gene expression in yeast, in a manner similar to their behavior in mammalian cells tested in parallel. These results indicate that there is no functional difference for glutamine-rich activation domains in stimulating gene expression in yeast and mammalian cells, irrespective of the lack of TAF_II110 in S. cerevisiae.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Plasmids and yeast strains. The yeast-integrating LacZ reporter plasmid pENH200 (see Fig. 2A) was derived via a combination of pJP158 (3), pDE200, and pDEYS2 (13). This resulted in a URA3-marked integrating vector that contains the LacZ gene under the control of two Lex, and three Gal4 binding sites 40 and 240 bp upstream of the GAL1 TATA box, respectively. The reporter plasmids pENH400 and pENH800 are derived from pENH200 and have 400- and 800-bp spacer sequences between the two proximal Lex binding sites and the three Gal4 binding sites, respectively. The integrating yeast plasmid pMG1, with switched Lex and Gal4 binding sites, was derived from pENH200 via replacement of the two proximal Lex (SacII-XbaI) and the three distal Gal4 (AflII-XmaI) sites with a double-stranded oligonucleotide that encodes two synthetic Gal4 and three Lex binding sites and possesses the corresponding SacII-XbaI and AflII-XmaI overhangs, respectively. The TATA-less TRP3 reporter plasmid pENH200TL was derived by subcloning the regulatory cassette that contains three Gal4 and two Lex binding sites (AflII [blunt ended]-XbaI) from pENH200 into the XhoI-filled-in XbaI sites immediately upstream of the TATA-less Trp promoter of pAB135, a URA3-integrating LacZ reporter vector.

MAT ura3-52 his3200 leu21 trp163 lys2385 gal411

Gene expression. Gene expression in yeast was monitored by using liquid -galactosidase assays and were performed as previously described (13). For reliable measurement of the low signals obtained when the glutamine-rich activation domains were tested alone, we routinely extracted whole-cell proteins as described below. All assays were conducted with duplicate samples and were repeated at least once. For HeLa cell experiments, cells were transfected as previously described (14). As a reporter plasmid, we used the OVEC -globin system (42). Subsequent S1 nuclease mapping was done as described (42).

EMSA and immunoblot analysis. Yeast cultures were grown in 10 ml of selective media to an optical density at 600 nm of 1.5 and then were harvested by centrifugation and resuspended in 1 ml of buffer containing 20 mM HEPES (pH 7.5), 10% glycerol, and 0.45 M NaCl, then were reharvested and resuspended in 0.1 ml of the same buffer supplemented to 0.1 M phenylmethylsulfonyl fluoride and 1 M dithiothreitol. From this step on, the samples were kept on ice. After addition of an equal volume of glass beads, samples were vigorously vortexed (5 × 30 s), and 0.1 ml of the same buffer was added before an additional vortexing step (15 s). After a 10-min centrifugation step at 4°C, supernatants were collected and used for electrophoretic mobility shift assay (EMSA) experiments (approximately 2 to 6 µl) or for -galactosidase assays according to standard protocol. As a probe, we used a double-stranded, ³²P-labeled oligonucleotide (5' GCAGTGCTGTATATAAAACGAGTGGTTATATGTACAGTAG 3') that contains two Lex binding sites. EMSA was performed as described (38). Immunoblot analysis of the Lex-Sp1Q2 wild type and the different mutants was carried out by using an anti-LexA antibody (Clontech). The experimental procedure of the immunoblot was carried out as described (13).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Glutamine-rich activation domains have a low activation potential on their own, but strongly synergize with a distal enhancer in both human and yeast cells. To clarify the transcriptional activation by glutamine-rich activation domains, we tested them in parallel in both mammalian and in yeast cells. In human cells, such domains were tethered to promoter-proximal binding sites, either on their own, or in combination with an enhancer. For this, we cloned the minimal glutamine-rich activation domains of Sp1 (Sp1Q1 and Sp1Q2, see Materials and Methods), Oct1 (Oct1Q), and Oct2 (Oct2Q) to the Gal4 DNA-binding domain (DBD) into a mammalian expression vector under the control of the strong cytomegalovirus promoter (Fig. 1A, Transactivators). As reporter, we used the -globin gene, which was under the control of two Gal4 binding sites proximal to the TATA box, with or without a simian virus 40 (SV40) enhancer sequence in a downstream position (Fig. 1A, Reporters). The different transactivators and reporters were cotransfected into HeLa cells. Two days after transfection, we quantified the transcript levels of reporter and reference genes. In the absence of a distal enhancer, proximally bound glutamine-rich activation domains showed a very low intrinsic activation potential (Fig. 1B, lanes 2 to 6). The strongest activation domain (Sp1Q1, lane 3) was able to activate reporter gene transcription five- to sixfold above that of Gal4 DBD alone (lane 2). However, transactivation by Sp1Q1 was less than 5% of that mediated by the strong herpes simplex activator VP16 fused to the Gal4 DBD (data not shown). A remotely positioned SV40 enhancer by itself, without a proximally bound transcription activation domain, activated -globin expression only weakly (lane 7). Binding of the Gal4 DBD (amino acids 1 to 93) in combination with the SV40 enhancer stimulated the reporter gene expression to some extent (lane 8), while a combination of proximal transactivator and distal SV40 enhancer led to strong synergistic gene activation (lanes 9 to 12). Quantification of reference and reporter -globin expression revealed that the activity of a remote enhancer was six- to sevenfold higher in concert with Sp1Q1 (lane 9), compared to the background by the Gal4 DBD alone (lane 8). The other tested glutamine-rich activation domains, Sp1Q2, Oct1, and Oct2, also synergized with the remote SV40 enhancer to yield three-, six-, and fourfold-higher levels of activity (lanes 10 to 12), respectively. These results show that glutamine-rich activation domains in HeLa cells possess a very low intrinsic activation potential, but in combination with additional activators contribute to synergistic gene activation. These data, which are in agreement with previous results (38), were the basis for a direct comparison to the budding yeast, where the same glutamine-rich activation domains were tested. Unlike our previous study where we found no activity of glutamine-rich domains in yeast (24), this time we used a promoter architecture that more closely resembles that of higher eukaryotes by introducing two proximal Lex sites and three distal enhancer binding sites for Gal4p upstream of the TATA box (Fig. 2A). This LacZ reporter construct was integrated into the genomic URA3 locus. We fused the Lex DBD (amino acids 1 to 202) to the same glutamine-rich activation domains of Sp1 (Lex-Sp1Q1 and Lex-Sp1Q2) and Oct1 (Lex-Oct1Q) as used in the HeLa cell experiments. These effector plasmids were tested for their ability to activate LacZ gene expression when bound proximally. The activity of Lex fusion proteins, either alone or in combination with the remotely bound Gal4p activator, was measured by the -galactosidase assay (Fig. 2B). All glutamine-rich activators when bound to the two proximal Lex binding sites were able to activate transcription to a low level only (Fig. 2B, lanes 4 to 6), although significantly above basal expression (lanes 1 to 3). This low expression was less than 5% of the value seen with acidic activators such as Lex-VP16 or Lex-Gal4p (data not shown).

View larger version (41K): [in this window] [in a new window]   FIG. 1.   Glutamine-rich activation domains are dependent on a distal enhancer in human cells. (A) HeLa cells were cotransfected with expression plasmids where the DBD of Gal4 (amino acids 1 to 93) was fused to glutamine-rich activation domains of Sp1, Oct1, or Oct2 (Transactivators), along with a reporter plasmid under the control of two proximal Gal4 binding sites, with or without a distal SV40 enhancer (Reporters). As an internal control, a reference plasmid expressing the 5'-truncated form of the -globin gene was cotransfected. (B) Total RNA was harvested 2 days after transfection of HeLa cells, and S1 nuclease analysis was performed. Lanes 7 to 12, due to the stronger reporter signals, were exposed for a shorter time than lanes 1 to 6. Quantification of the reporter signal (Rep.) relative to the corresponding reference (Ref.) was carried out by using a PhosphorImager. Without a distal enhancer, the glutamine-rich activation domains stimulate reporter gene expression only slightly above the background level (lanes 3 to 6 versus lanes 1 and 2). In combination with the remote SV40 enhancer, the glutamine-rich activation domains synergistically activated gene expression (lanes 9 to 12) when compared to the influence of the SV40 enhancer alone (lanes 7 and 8).

View larger version (28K): [in this window] [in a new window]   FIG. 2.   Glutamine-rich activation domains stimulate transcription from a chromosomally embedded reporter gene in yeast. (A) Schematic drawing of the integrated yeast reporter. The promoter architecture resembles that of a typical higher eukaryote. It contains distant (three Gal4 binding sites) as well as proximal (two Lex binding sites) regulatory elements upstream of the GAL1 TATA box. The reporter was integrated at the chromosomal URA3 locus of S. cerevisiae. (B) The reporter strain was transformed with yeast plasmids that express Lex fusion proteins to the Gal4 dimerization domain (amino acids 58 to 97) as a negative control and the same glutamine-rich activation domains of Sp1 (Q1 and Q2) and Oct1Q as tested in HeLa cells (Fig. 1). Their ability to stimulate transcription when bound to the proximal binding sites either alone or in combination with a distal Gal4p activator was monitored. The factors containing a glutamine-rich activation domain displayed a low intrinsic activation potential when tested on their own (lanes 4 to 6), which was above the background as determined by the Lex (lane 2) or Lex-Gal4 dimerization domain alone (lane 3). The influence on reporter gene activation of the distal Gal4p (lane 7) and the proximal glutamine-rich activation domains (lanes 4 to 6) was more than additive when tested in combination, i.e., they resulted in synergistic gene activation (lanes 10 to 12). As in the mammalian cells shown in Fig. 1, Sp1Q1 (lanes 4 and 10) was the strongest activation domain, followed by Oct1Q (lanes 6 and 12) and Sp1Q2 (lanes 5 and 11). (C) EMSAs were performed with total yeast protein extracts derived from reporter strains that expressed various Lex fusion proteins. A 32P-labeled oligonucleotide duplex with two Lex binding sites was used as a probe. Extracts from reporter strains without a Lex fusion protein did not reveal any protein interaction with the oligonucleotide (lane 1). Lex alone (lane 2) produced a weaker bandshift than the other fusion proteins. Fusion of the Gal4 dimerization domain to the Lex protein (lane 3) increased the stability of the protein to interact with the oligonucleotide and yielded a bandshift comparable to those of Lex fused to Sp1Q1 (lane 4), Sp1Q2 (lane 5), and Oct1Q (lane 6).

View larger version (23K): [in this window] [in a new window]   FIG. 3.   The herpes simplex viral activator VP16 synergizes with a proximal glutamine-rich activation domain of Sp1 over long distances in yeast. (A) The potent activation domain of the herpes simplex transactivator VP16 fused to the Gal4 DBD (Gal4-VP16) was tested for its ability to stimulate gene expression from remote Gal4 binding sites (Gal4 BS). These binding sites were separated from the proximal Lex binding sites (Lex BS) by a 400-bp spacer sequence. A schematic drawing of the chromosomal reporter construct is indicated above the graph. Distal Gal4-VP16 or proximal Sp1Q1 fused to LexA (Lex-Sp1Q1) alone activated to comparable low levels. Combination of both resulted in a more-than-tenfold increase of reporter gene transcription as compared to the effect of each activator tested on its own. (B) Over the very long distance of 800 bp, the Gal4-VP16 transactivator did not stimulate gene expression at all when tested either alone or with the proximal, transcriptionally inert LexA. However, in combination with the proximally bound glutamine-rich domain of Sp1 (Lex-Sp1Q1), Gal4-VP16 strongly contributed to gene expression, i.e., under these conditions, the enhancer-bound activator is strictly dependent on the presence of a promoter-proximal glutamine-rich activation domain.

The serine/threonine-rich domains of Sp1 do not influence the adjacent glutamine-rich activation domains. Several serine/threonine-rich domains have been reported to harbor transcriptional activity (18, 38, 39). The Sp1 transcription factor contains two serine/threonine-rich domains, each immediately N terminal to the two glutamine-rich activation domains (see Fig. 4A for a schematic drawing). We wanted to determine whether these two serine/threonine-rich domains of Sp1 could influence the activation potential of the adjacent glutamine-rich domains. For this, we fused fragments of Sp1 containing the serine/threonine-rich and Q1 domains (amino acids 1 to 161) as well as the serine/threonine-rich and Q2 domains (amino acids 161 to 403) to the Lex DBD. These fusion proteins were compared to their counterparts containing only the glutamine-rich activation domains Q1 and Q2 (Fig. 4B). We observed that the serine/threonine-domain-containing constructs behaved in a manner indistinguishable from that of the pure glutamine-rich activation domains. Therefore, the serine/threonine-rich domains apparently do not influence the transactivation potential of the adjacent Q1 or Q2 activation domain of Sp1 in yeast. These data are in agreement with previous results from Drosophila Schneider cells that indicated that the serine/threonine-rich domains of Sp1 do not contribute to transcriptional activation (9).

View larger version (24K): [in this window] [in a new window]   FIG. 4.   The serine/threonine-rich domains of Sp1 do not influence transactivation by the adjacent glutamine-rich activation domains. (A) Schematic drawing of full-length Sp1 transcription factor (696 amino acids). Sp1 harbors two glutamine-rich activation domains termed Q1 and Q2, two serine/threonine-rich domains, and a DBD at the C terminus. (B) Quantitative -galactosidase assay. The activity of the glutamine-rich activation domains Q1 and Q2 were compared with N-terminal extensions that included the serine/threonine-rich domains. The activity of the different Lex DBD hybrids was assayed in yeast when bound to the proximal position, either alone (no Gal4p) or in combination with a distal Gal4p activator (with Gal4p). The serine/threonine-domain-containing constructs activated reporter gene expression indistinguishably from that of the respective activation domains Q1 or Q2.

Mutations in Sp1Q2 similarly affect its transactivation activity in yeast and in human cells. The Drosophila TATA-binding protein (TBP)-associated factor TAF_II110 has been shown to interact with the human transcription factor Sp1 (20). The level of Sp1 transcriptional activity in Drosophila directly correlated with the strength of interaction between the glutamine-rich activation domains of Sp1 and TAF_II110 (16). Substitutions in the weaker of the two glutamine-rich activation domains (Sp1Q2), changing three leucines to alanines (LA) or one tryptophan to alanine (WA), were reported to be defective in their ability to interact with the Drosophila TAF_II110 in a yeast two-hybrid assay. When fused to the Zn finger DBD of Sp1, they also failed to activate transcription of a reporter gene dependent on six Sp1 binding sites from the SV40 promoter in Drosophila Schneider cells (16). We introduced the same mutations as described (LA and WA) and the combination of them (L, WA) by site-directed mutagenesis (Fig. 5A). The resulting mutant Sp1Q2 domains fused to Lex (amino acids 1 to 202) were tested for their ability to synergize with the distal Gal4p activator in the yeast reporter strain described in Fig. 2A. Unexpectedly, we could not observe a significant difference of the two mutants (LA and WA) relative to transactivation by wild-type Sp1Q2 in yeast (Fig. 5B). Only a combination of the two mutants, with the four amino acids exchanged in the 147-amino-acid-long activation domain reduced activity to near background level.

View larger version (41K): [in this window] [in a new window]   FIG. 5.   Mutations in Sp1Q2 similarly affect its activation potential in both yeast and human cells. (A) Substitutions of three leucines to alanines (LA) and a tryptophan to an alanine (WA), which are known to affect the interaction with dTAFII110 and gene activation in Drosophila cells, were introduced in the Q2 activation domain of Sp1. In addition, both types of mutations were combined (L, WA), resulting in a four-amino-acid exchange in the 147-amino-acid-long activation domain. (B) Yeast strains containing the reporter gene as described in Fig. 2A were transformed with plasmids encoding the Gal4 activator and different Lex-Sp1Q2 mutants or wild type. Quantitative -galactosidase assays showed that the two mutants (LA and WA) still activated gene expression as much as the wild type. Only the combination mutant (L, WA) diminished the transactivation potential of the Q2 domain. (C) Mutant and wild-type Sp1Q2 fusion proteins bind similarly to the Lex binding sites. EMSAs were performed by using total protein extracts from yeast cells and 32P-labeled oligonucleotide duplex containing two Lex binding sites. Equal amounts of protein were used for each EMSA. The Sp1Q2 mutants (lanes 4 to 6) bound to the Lex binding sites as well as the wild type (lane 1), indicating similar protein expression levels. Protein extracts from isogenic yeast cells containing an empty expression plasmid did not yield any detectable bandshifts (lane 3). The immunoblot using anti-LexA antibodies showed similar expression of the wild-type Sp1Q2 and the different mutants. (D) Transfection of HeLa cells with wild-type and mutant Sp1Q2. The mutants tested in yeast were subcloned as Gal4 (amino acids 1 to 93) hybrids into a mammalian expression vector. These transactivator plasmids were cotransfected into HeLa cells along with a reporter plasmid, under the control of two proximal Gal4 binding sites and a downstream SV40 enhancer (see Fig. 1A), and a reference plasmid. S1 nuclease analysis was performed, and the signals were quantified by using a PhosphorImager. As in yeast (Fig. 4B), the mutants LA and WA had the same transactivation potential as wild-type Sp1Q2. Only the combined mutations L, WA showed a reduced ability to stimulate reporter gene expression, which was still above background.

The glutamine-rich domain of Snf5p activates transcription in yeast. The above-mentioned results show that glutamine-rich activation domains from Sp1 and Oct1 on their own stimulate transcription in yeast and human cells only to a minor extent compared to acidic activators, but they readily synergize with a remote enhancer. Next, we wanted to test whether a glutamine-rich domain of an endogenous yeast protein could also activate gene expression. A search of the yeast genome database for open reading frames (ORFs) containing a stretch of at least 10 glutamine residues revealed that out of 64 hits, 48 are known ORFs. Of these 48, 18 (37%) are factors involved in transcriptional activation. They include the following genes: FLO8, DAT1, HAP2, MCM1, MED3, POP2, TAF61, NDD1, IXR1, CRZ1, CCR4, SNF5, DAL81, GAL11, YPR022C, GTS1, SRB9, and HAP1. The rest of the ORFs could be assigned to eight kinase or kinase-associated factors, five RNA-binding proteins, three factors involved in G-protein-coupled complexes, two chaperones, and eight other miscellaneous factors. This distribution indicates that the largest fraction of yeast glutamine-rich proteins are involved in transcriptional regulation and activation.

View larger version (29K): [in this window] [in a new window]   FIG. 6.   The glutamine-rich domain of Snf5p activates transcription in yeast. (A) Schematic drawing of the 905-amino-acid-long Snf5p factor which contains an acidic region (amino acids 490 to 588), three proline-rich domains (amino acids 72 to 132, 287 to 324, and 714 to 882), and two glutamine-rich domains (amino acids 61 to 69 and 185 to 286). The boundary amino acid positions of the corresponding domains are indicated. The larger glutamine-rich domain of Snf5p (amino acids 185 to 286) (Lex-Snf5 185-286) as well as a deletion mutation (deletion of amino acids 211 to 260) of this glutamine-rich domain (Lex-Snf5 211-260) were fused to the Lex DBD. (B) Quantitative -galactosidase assay. The glutamine-rich domain of Snf5p stimulated LacZ expression when bound proximally to the TATA box (lane 4) and synergized with a distal Gal4p activator (lane 8). For comparison, the stronger glutamine-rich activation domain of Sp1 (Lex-Sp1Q1) was used (lanes 2 and 6). The 50-amino-acid deletion abolished the transactivation potential of the Snf5p glutamine-rich domain (lane 3) and hence the synergism with the distal Gal4p activator (lane 7).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Glutamine-rich domains and polyglutamine stretches are integral components of many proteins involved in transcriptional regulation, from yeast to human. We had previously shown that in mammalian cells, glutamine-rich domains poorly activate transcription on their own when tethered to DNA in a promoter-proximal position but strongly synergize with remotely bound transcriptional activators of the acidic type (38). While yeast cells readily respond to mammalian acidic activators, we and others had found them to be nonresponsive to glutamine-rich activation domains of mammalian factors (23, 24, 32), which at that time suggested that yeast lacks an important interaction partner for these latter domains. In Drosophila, which has no homologue of the mammalian Sp1 transcription factor and thus was suitable for testing the activity of ectopically expressed Sp1, its glutamine-rich activation domains were found to bind to TAF_II110, a TBP-associated general transcription factor (8, 16, 20). The completion of the entire S. cerevisiae genome sequence revealed a number of homologues for mammalian TATA binding protein-associated factors, but no counterpart to Drosophila TAF_II110/human TAF_II130, which seemed to offer a straightforward explanation for the lack of activity of the glutamine-rich domains in yeast. More recently, the fission yeast S. pombe was found to be responsive to glutamine-rich domains of mammalian origin (35). In the sequence database of S. pombe we have found a 365-amino-acid ORF with considerable similarity to dTAF_II110/hTAF_II130. This prompted us to introduce this protein into the budding yeast and to study its effect on glutamine-rich activation domains. The S. pombe protein was expressed in S. cerevisiae, but failed to influence the expression of reporter genes driven by glutamine-rich activation domains in any of the enhancer/promoter constellations tested (D. Escher and W. Schaffner, unpublished results). Although we cannot exclude the possibility that an S. pombe-specific cofactor for TAF_II110 is missing in the budding yeast, it has become clear by now that the latter can respond well to glutamine-rich domains in the absence of a TAF_II110 homologue. Recently, it was reported that the glutamine-rich activation domain Sp1Q1 is able to activate transcription of a reporter gene when tethered to DNA in S. cerevisiae (43). This was, however, only observed when the reporter gene was present on a high-copy-number plasmid (2µm), while no activation was detected when it was integrated in a chromosomal locus. The present study clearly shows that glutamine-rich domains, both of mammalian and yeast origin, stimulate reporter gene expression in a yeast chromosomal context if the promoter region is structured in a manner that is prevalent in higher eukaryotes but less common in yeast, namely a promoter with proximal factor binding sites plus enhancer-type regulatory sequences further upstream. Such a configuration was not tested in previous experiments, including the ones from our lab; rather, a simpler promoter version with a few bindings sites in the immediate vicinity of the TATA box had been used. In the present study, the glutamine-rich activation domains of Sp1 and Oct1, when tethered to a proximal position, even synergized to comparable levels with remote acidic activators in yeast and human cells. Furthermore, mutations in the activation domain of Sp1Q2 had similar effects in yeast and HeLa cells, suggesting a common interaction pathway. Thus, it remains to be seen whether the interaction observed between TAF_II110/TAF_II130 and the human Sp1 glutamine-rich activation domains is of general significance. Whatever the role of the TAF_II110-like protein in S. pombe, our results show that glutamine-rich domains behave very similarly in yeast and human and that their activity does not depend on the presence of a TAF_II110 homologue in budding yeast. This again raises the question of the possible target(s) of glutamine-rich domains. The TBP itself was reported to interact with glutamine-rich activation domains. However, the good activity of glutamine-rich domains in yeast observed by us is difficult to reconcile with the report that TBPs from Drosophila and humans, but not from yeast, bind well to glutamine-rich domains (12, 32). Recently, a multiprotein complex termed "cofactor required for Sp1" was found to mediate Sp1 activity in extracts of human cells (36). Some of the characterized components are absent in yeast, whereas others do have a yeast homologue (36) and thus may be involved in mediating the activity of glutamine-rich domains both in human and yeast.