INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The STA1, STA2, and STA3 genes encode extracellular glucoamylase isozymes which enable Saccharomyces cerevisiae cells to utilize starch as a carbon source (reviewed in references 38, 40, and 50). The three genes have nearly identical sequences, and are located on chromosomes II (STA2), IV (STA1), and XIV (STA3). All three members of the STA gene family are located in subtelomeric positions, similar to the FLO (reviewed in reference 48), SUC (reviewed in reference 15), and MAL (reviewed in reference 32) gene families, which probably evolved through genomic duplications and chromosomal rearrangements. The 5' upstream region of STA1 and STA2 (the nucleotide sequence of STA3 has not previously been determined) is almost identical to that of MUC1, which encodes a large membrane-bound, mucin-like protein that plays an important role in the processes of invasive growth, pseudohyphal development, and flocculation (8, 23, 27, 28). The homology extends over more than 3,500 bp upstream of the ATG start codon and includes the first 60 bp of the open reading frame (ORF) encoding a secretion signal sequence (52). With the exception of a few single nucleotide dissimilarities, the only significant differences between the promoters of STA2 and MUC1 are two inserts of 20 and 64 bp in the MUC1 promoter, which are absent from the STA2 promoter (23). These inserts stretch from nucleotides 1333 to 1313 and 933 to 869, respectively. This very limited sequence divergence between the STA and MUC1 promoter regions suggests a recent origin of the STA genes. The STA genes probably evolved through a recombination and sequence duplication process between the promoter and signal sequence of MUC1 and the ORF of the SGA1 gene that encodes a sporulation-specific intracellular glucoamylase. MUC1 and SGA1 are located on the right and left arms of chromosome IX, respectively (59, 60). Besides the strong sequence conservation between these genes, other arguments in favor of a recent origin of the STA genes and of the proposed molecular mechanism are (i) the subtelomeric position of the STA genes compared to the more central position of both MUC1 and SGA1; (ii) the presence of STA genes in only some S. cerevisiae strains, compared to the general presence of MUC1 (4, 59, 60) and SGA1 (59, 60) in all S. cerevisiae strains investigated so far; and (iii) the existence of homologous repeated sequences on either side of the proposed junctions (59, 60).

Analyses of the upstream areas of STA1 (1, 46), STA2 (22), and MUC1 (44) demonstrated that elements at distances of up to 2,800 bp from the translation start codon (ATG) are involved in the transcriptional control of the respective genes, therefore representing the largest S. cerevisiae promoters identified to date (7, 44). The STA and MUC1 upstream regions are therefore of particular interest from both an evolutionary and a functional point of view.

The extent of the promoter homology would suggest that genes involved in starch metabolism and pseudohyphal differentiation or invasive growth are coregulated to a large extent, and experimental data so far have supported this hypothesis. Lambrechts et al. (23, 24) and Gagiano et al. (8) showed that two transcriptional regulators, Msn1p and Mss11p, strongly induce transcription of both the STA2 and MUC1 genes when present on multiple-copy plasmids. Conversely, msn1 or mss11 strains show strongly reduced transcription of these genes. Furthermore, Lo and Dranginis (28) demonstrated that MUC1 is regulated by Ste12p, a transcription factor responsible for both pheromone-specific (reviewed in reference 21) and, in combination with the TEA (or ATTS) family transcription factor Tec1p, filamentation-specific (9, 30) gene regulation. Gagiano et al. (8) presented evidence that the same factor regulates the STA genes in a similar way.

Other regulatory factors have so far only been associated with regulation of MUC1 or STA1, STA2, and STA3 independently. Recent data suggest that the transcription of MUC1 might be specifically regulated by a network of signal transduction pathways which control invasive growth and pseudohyphal differentiation (8, 28, 44). This network combines inputs from at least three interacting signal transduction modules, including (i) the filamentation-specific mitogen-activated protein (MAP) kinase cascade (25, 31, 42), (ii) the cyclic AMP (cAMP) and cAMP-dependent kinase (29, 43), and (iii) the cyclin-dependent kinase Cdc28p (6). In addition to the result mentioned above that MUC1 was subjected to MAP kinase-dependent regulation by Ste12p and Tec1p, the gene was shown to be regulated by cAMP levels, a regulation that occurs via Flo8p (44), a transcription factor initially identified for its role in flocculation (18). The gene was also shown to be negatively regulated by a suppressor of flocculation, Sfl1p, which interacts specifically with the yeast A kinase, Tpk2p, to repress MUC1 transcription in the absence of a cAMP signal (43).

Numerous data concerning the regulation of the STA genes have been published. Expression of STA1 to -3 is negatively regulated at several levels. Transcription is repressed on most readily metabolized carbon sources, including glucose, sucrose, maltose, and galactose (5, 17, 20, 41, 47). Carbon catabolite repression was reported to involve two separate pathways, of which one requires HXK2 and the other HAP2 (17). It was also reported that repression of STA2 does not require Mig1p, the common repressor of genes under carbon catabolite control. MUC1 was also shown to be repressed in medium containing glucose as a carbon source (8, 27), probably via the same mechanisms as STA1 and STA2. Transcription of STA1 to -3 is repressed in most, but not all, diploid strains of S. cerevisiae (5, 41). The mechanism through which repression occurs is not defined, since the removal of the putative a1-2 repressor binding sites from the STA2 promoter does not relieve the repressive effect observed in diploid strains (22). In rich medium, MUC1 is also repressed in diploid strains, but under nitrogen starvation conditions it seems to be repressed more in haploid than diploid strains (28, 44).

Most laboratory strains of S. cerevisiae contain an undefined repressor, STA10, which reduces transcription of the STA1 to -3 genes at least 20-fold (37, 41). It was reported that the repressive effect of STA10 results from interaction between two unlinked genes, IST1 and IST2 (34), but this was not confirmed. The negative effects of several other genes (i.e., INH1 [58], SGL1 [36], and SNS1 and MSS1 [1]) on the transcription of the STA genes have also been reported, but the relationships between these negatively acting genes and the repressive effect of STA10 remain to be determined.

Transcription of STA1 to -3 is subject to the repressive effect of chromatin on promoters, since SUD1, a component of a global chromatin-associated repressor of promoter activity, was shown to act on the STA1 promoter (57). Furthermore, transcription of STA1 to -3 also requires the presence of components of the SWI-SNF global activation complex (13, 20, 33, 61-63), which associates with the RNA polymerase holoenzyme at specific promoters and relieves the repressive effect of chromatin on transcription (19, 55).

cis-acting promoter elements in several regions within the STA1 (1, 46), STA2 (22), and MUC1 (44) promoters were shown to be required for transcriptional regulation. Two areas hosting upstream activating sequences (UASs) (UAS1 between nucleotides 1390 and 1074 and UAS2 between nucleotides 1940 and 1815), as well as three upstream repression sequences (URSs), were identified in the STA2 promoter (22). URS1 was found to reside in the area between nucleotides 1390 and 1074, which also hosts UAS1. URS2 was identified between nucleotides 1650 and 1390 and URS3 upstream of position 2457. Similar regions were defined for the STA1 promoter (1, 46). A recent, more systematic, analysis of the MUC1 promoter (44) revealed a vast array of regulatory elements which confer the regulation of several nutritional and cell-type signals on MUC1 expression levels. In good agreement with the previous studies of the highly homologous STA1 and STA2 promoters, four areas required for the activation of MUC1 and nine areas required for the repression thereof were identified. The transcriptional activator encoded by FLO8 was found to exert its activating effect through a 200-bp sequence stretching from nucleotides 1200 to 1000 in the upstream region of MUC1 (44).

The 5' upstream areas of MUC1, STA1, and STA2 are predicted to contain a single small ORF, YIR020c, of unknown function, situated from nucleotides 1285 to 882 in the upstream region of MUC1. YIR020c lies in an area identified and experimentally defined as a regulatory region for STA1, STA2, and MUC1, and other regulatory regions were shown to exist upstream of this ORF (1, 22, 44, 46). Its occurrence therefore does not affect conclusions regarding the transcriptional regulation of STA1, STA2, or MUC1, independently of whether this ORF encodes a functional protein or not.

The homologous sequences from nucleotides 1390 to 1074 of the STA2 promoter and from nucleotides 1479 to 1136 of the MUC1 promoter are of particular interest since they (i) have previously been identified as areas hosting a UAS as well as a URS (22), (ii) confer increased levels of activity from a far-upstream position, and (iii) include one of the two significant differences between the upstream areas of MUC1 and STA1 to -3 (a sequence of 20 bp that is deleted in the STA2 promoter). The region might therefore contain an evolutionarily significant molecular change explaining differences in the regulation of STA1 to -3 and MUC1.

In this paper, we compare expression levels conferred by the full MUC1 and STA2 promoters on reporter gene expression. We furthermore present a detailed analysis of the promoter region from nucleotides 1390 to 1074 of STA2 and the corresponding area of MUC1, from nucleotides 1479 to 1136. We show that these regions of MUC1 and STA2 confer both similar and divergent regulation and contain sequences involved in general repression as well as areas for (i) activation by the transcriptional activators encoded by MSN1 and MSS11, (ii) activation by the transcriptional activator encoded by FLO8, (iii) carbon catabolite repression, and (iv) diploid repression. Our data indicate that differences in expression levels observed between MUC1 and STA2 are largely due to the two deletions of 20 and 64 bp that have occurred in the STA promoters. We also show that the repressive effect identified as STA10 in most laboratory S. cerevisiae strains is due to the absence of the FLO8-encoded transcriptional activator. Epistatic analysis furthermore suggests that FLO8 requires or is situated upstream of MSS11, but acts independently of MSN1.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Strains, growth media, and genetic methods. The S. cerevisiae strains used in this study, along with the relevant genotypes, are listed in Table 1. Transformation of S. cerevisiae cells was carried out by the lithium acetate procedure (3). The one-step gene replacement method (3) was used to disrupt the FLO8 loci with the flo8::URA3 cassette, pflo8, in the genomes of strains ISP15 and ISP20 to generate strains ISP15flo8 and ISP20flo8, respectively. Successful disruptions of the FLO8 loci in these strains were verified by Southern blot analysis and confirmed by PCR analysis. The URA3 marker of strains ISP15flo8 and ISP15msn1 was regenerated through transformations with the ura3::Kan^r disruption cassette, pura3::kan, and selected on medium containing 125 mg of kanamycin per ml and 1 mg of 5-fluoroorotic acid per ml. S. cerevisiae FY23 (56) is isogenic to the S288C genetic background, and L5366 (25) and L5366h (8) are isogenic to the 1278b genetic background. Strain JM2508 does not contain any of the STA1 to -3 genes and is from the culture collection of the late Julius Marmur.

Construction of plasmids. FLO8 was isolated as a 3,252-bp SphI-EcoRV fragment from plasmid pF415-1 (18) and ligated to plasmids YEplac112 and YEplac181 (10), digested with SphI and SmaI, to generate plasmids YEplac112-FLO8 and YEplac181-FLO8. YEplac112-FLO8 was subsequently used to construct pflo8, a cassette for disrupting the FLO8 locus. In order to do this, a 760-bp PstI-BglII fragment, comprising the translational start codon (ATG) and a large part of the FLO8 ORF, was removed and replaced with a 1,084-bp NsiI-BamHI fragment containing the URA3 marker, isolated from plasmid pJJ242 (16).

View larger version (19K): [in this window] [in a new window]   FIG. 1.   Construction of a series of plasmids containing sequential deletions of the STA2 and MUC1 UAS1 area, upstream of the lacZ reporter gene in plasmid pHP41. The positions of UAS1 and UAS2 relative to the translation initiation codon (ATG) of the STA2 and MUC1 ORFs are indicated, and the positions of the fragments in the respective promoters are given. The position of the 20-bp insert of the MUC1 promoter is indicated by the black bars.

Sequencing of the STA2 and STA3 promoters. To sequence the 5' upstream region of STA3, a series of nine primers were synthesized, covering the entire promoter area and first part of the STA3 ORF. The primers were designed from the available sequences of STA1 and STA2. Plasmid pSTA3-6-4 (59) was used as a template to determine the nucleotide sequence.

-Galactosidase assays. After transformation, at least three colonies of each transformation were grown overnight in 10 ml of selective SCD medium. From each overnight culture, 10-ml SCD, SCGE, SLAD, and SPD cultures were inoculated to an optical density at 600 nm (OD₆₀₀) of 0.1 and incubated to grow for four to five generations at 30°C to an OD₆₀₀ of ~1.0. To obtain postdiauxic- shift cultures, SCD cultures were incubated for longer periods until they had reached an OD₆₀₀ of >3.0. The effect of osmotic shock on expression levels was determined in 10-ml selective SCD cultures that were grown to an OD₆₀₀ of 1.0. Sterile NaCl was added to a final concentration of 0.7 M, after which the cultures were incubated at 30°C for 1 h. The effect of heat shock was determined in 10-ml selective SCD cultures grown to an OD₆₀₀ of 1.0 and placed at 42°C for 1 h. -Galactosidase assays were done according to the method of Ausubel et al. (3). Margins of error were calculated for each set of assays and were usually less than 7.5% and never higher than 15%.

Invasive growth and pseudohyphal development assays. Three colonies from a transformation were inoculated into SCD medium and grown to an OD₆₀₀ of 1.0. To assess the ability of these yeast cells to grow invasively into the agar, 10 µl of this liquid culture suspension was spotted onto SLAD, SCS, SCGE, and SCD agar plates. Plates were incubated at 30°C and investigated for invasive growth at intervals of 2 days. Yeast colonies were washed off the surface of the agar by rubbing the surface of the plates with a gloved finger under running water. Cells that grow invasively into the agar cannot be washed off and are clearly seen below the surface of the agar.

Plate assays to determine starch utilization. The STA2 gene encodes an extracellular glucoamylase which hydrolyzes starch by liberating glucose molecules from the nonreducing end of the starch molecule (50). The presence of the STA2 gene therefore enables most yeast strains to grow on starch as the sole carbon source. On plates containing starch as a carbon source (SCS), a clear zone is formed around such starch-degrading colonies, and the size of the colony, as well as the diameter of the zone, is indicative of the amount of glucoamylase secreted (39, 59). The levels of expression of STA2 in yeast strains were therefore determined by the size of the colonies and the clear zone around each of the colonies on SCS plates.

Sequence analysis and homology searches. Homology searches in the yeast genome subdivision of GenBank were done with BLAST software (2). Sequence fragment assembly and individual alignments between the STA genes and MUC1 were done with the OMIGA v1.1 package (Oxford Molecular Ltd.).

Nucleotide sequence accession number. A 2,779-bp sequence comprising the STA3 promoter and the first part of the ORF was submitted to the GenBank database and assigned accession no. U95022. A 1,462-bp sequence comprising the far upstream region of the STA2 promoter was submitted to the GenBank database and assigned accession no. AF169185.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Similar and divergent regulation of STA2 and MUC1. To determine the extent of the coregulation between MUC1 and STA2, we determined the -galactosidase activity of the MUC1 and STA2 promoters fused to the lacZ reporter gene with plasmids pPMUC1-lacZ and pPSTA2-lacZ, respectively, under different growth conditions as well as in the presence of multiple copies of the transcriptional activators FLO8, MSN1, and MSS11. The results for these assays are given in Tables 4 and 5.

The STA10 repressive effect in S288C-derived strains is due to a mutation in FLO8. When compared to feral S. cerevisiae strains, most laboratory strains (e.g., S288C) exhibit a 20-fold reduction in STA1 to -3 expression (41). This phenomenon was believed to be due to the presence of a repressor, designated STA10 (37). It was, however, recently reported that most laboratory strains contain a point mutation in FLO8, a transcriptional activator of the flocculation genes, which renders these strains unable to flocculate, grow invasively, or form pseudohyphae (26). Due to the extensive homology between the STA2 and MUC1 promoter regions and since FLO8 was shown to be required for transcription of MUC1 (44), we investigated whether a genetic relationship between STA10 and FLO8 exists.

View larger version (67K): [in this window] [in a new window]   FIG. 2.   (A) The S288C-derived STA10 strain, FY23, transformed with the centromeric plasmids YCplac22 and YCplac33 (Ctrl.), YCplac33 and YCplac22-FLO8 (FLO8), YCplac22 and YCplac33-STA2 (STA2), and YCplac22-FLO8 and YCplac33-STA2 (STA2 + FLO8) on plates containing potato starch as the sole carbon source (SCS). Cells that are unable to express STA2 are unable to grow, whereas cells that do express STA2 sufficiently produce extracellular glucoamylase, which enables them to grow. The clear zone around the Sta+ colony is due to the hydrolysis of the starch in the medium. (B) The sta10 strains ISP15 and ISP20 with wild-type (wt) and disrupted (flo8) FLO8 loci on medium containing starch as the sole carbon source. The wild-type strains express STA2 sufficiently to sustain growth on starch, whereas the flo8 strains show a clear reduction in glucoamylase expression and are therefore unable to grow.

Flo8p acts independently of Msn1p but upstream of Mss11p. FLO8 is one of several transcriptional regulators required for the transcriptional activation of the STA1 to -3 genes and MUC1. The epistatic relationships between these transcriptional regulators revealed a complex network of signal transduction pathways that converge at the promoter of the MUC1 (8, 44) and STA1 to -3 (8) genes. To establish the epistatic relationship between FLO8 and other transcriptional regulators required for MUC1 and STA1 to -3 expression, MSN1 and MSS11, present on 2µm plasmids, were transformed into strains with deleted FLO8 loci, ISP15flo8 and ISP20flo8. A 2µm plasmid carrying FLO8 was also transformed into strains with deleted MSN1 and MSS11 loci. These strains were spotted onto SLAD (limited nitrogen source) and SCS (potato starch as carbon source) plates and scored for their ability to grow invasively into the agar and to utilize starch.

View larger version (95K): [in this window] [in a new window]   FIG. 3.   Determination of the epistatic relationships between MSN1, MSS11, and FLO8 on limited nitrogen (SLAD) medium in strain ISP20. Wild-type ISP20, ISP20flo8, ISP20msn1, and ISP20mss11 were transformed with YEplac112 without insert (Ctrl), YEplac112-FLO8 (2µm FLO8), YEplac112-MSN1 (2µm MSN1), and YEplac112-MSS11 (2µm MSS11). Cells were grown in SCD medium until mid-log phase, whereupon 10 µl of each of the respective cell suspensions was spotted onto limited nitrogen (SLAD) medium. Plates were incubated for 6 days, after which the growth on top of the agar was washed off. Cells that grew invasively into the agar could not be washed off and were photographed. 2µ, 2µm.

Role of UAS1 in determining expression levels of STA2 and MUC1. Deletions of the UAS1 area from the promoters of MUC1 and STA2 (Table 5) indicated that this area is required for glucose repression and transcriptional activation by MSS11, FLO8, and MSN1. The data show that multiple copies of MSN1 or MSS11 were still able to increase expression levels conferred by the MUC1 and STA2 promoters when UAS1 is deleted. This suggests that the corresponding gene products act through regulatory sequences both within and outside of UAS1. Interestingly, the same does not apply for multiple copies of FLO8, which are unable to induce reporter gene expression when UAS1 is deleted. Flo8p is therefore completely dependent on sequences within the UAS1 region to assert its effect on MUC1 and STA2 transcription.

Identification of regulatory regions within the UAS1 area. Both a previous report (44) and the data presented in Table 5 suggest that FLO8 confers regulation via a sequence within the UAS1 area. Our data (Table 5) furthermore show that Msn1p and Mss11p act in part via the same region. To better define this area, sequential deletions of UAS1 were generated through PCR amplification, by using the promoters of MUC1 and STA2 as templates. These fragments were introduced into the UAS-less CYC1 promoter fused to lacZ as a reporter gene on the centromeric vector pHP41 (Fig. 1). These constructs, as well as the vector without any insert as a control, were transformed into different genetic backgrounds, and the levels of -galactosidase conferred by these fragments were determined.

Effect of the small MUC1 promoter inserts on expression levels. Unlike MUC1, the STA1 to -3 genes are not present in the genomes of the S288C-derived laboratory strains that were used in the sequencing of the S. cerevisiae genome. Laboratories working on starch metabolism in S. cerevisiae therefore contributed the sequences of the STA1 and STA2 genes. STA3 is the only member of the STA gene family that had not been sequenced to date. To establish whether the promoter is identical to those of the other members of the family, the 5' upstream region of STA3 was sequenced and compared to the available sequences of STA1 and STA2. The sequence proved to be identical to those of the STA1 and STA2 promoters, with the exception of some single nucleotide substitutions (data not shown).

View larger version (123K): [in this window] [in a new window]   FIG. 4.   (A) SCS plate (containing 2% cornstarch as a carbon source) to investigate the starch utilization of S. cerevisiae JM2508 transformed with plasmids YCplac33 and YEplac112 without any inserts (no. 1), YCplac33-STA2 and YEplac112 (no. 2), YCplac33-STA2 and YEplac112-FLO8 (no. 3), YCplac33-STA2 and YEplac112-MSN1 (no. 4), and YCplac33-STA2 and YEplac112-MSS11 (no. 5). (B) SCS plate (containing 2% cornstarch as a carbon source) to investigate the starch utilization of S. cerevisiae JM2508 transformed with plasmids YCplac33 and YEplac112 without any inserts (no. 1), YCplac33-PMUC1-STA2 and YEplac112 (no. 2), YCplac33-PMUC1-STA2 and YEplac112-FLO8 (no. 3), YCplac33-PMUC1-STA2 and YEplac112-MSN1 (no. 4), and YCplac33-PMUC1-STA2 and YEplac112-MSS11 (no. 5).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The MUC1 and STA1 to -3 promoters are of particular interest, since they (i) consist of evolutionarily closely related sequences, allowing the study of promoter evolution on a molecular level; (ii) represent the largest S. cerevisiae promoters identified to date; and (iii) might integrate the information transmitted by several separated signal transduction pathways to specifically result in an adaptive cellular differentiation process. Our results confirm previously published data suggesting that the expression of the MUC1 and the STA genes is indeed controlled by the complex interaction of a large number of factors which are regulated by several independent signaling pathways.

Our data regarding the transcriptional activity of the entire promoters reveal several general features. First, PMUC1-dependent reporter gene expression is very low under most conditions and is generally well below levels observed for the UAS-less reporter plasmid alone, indicating that the entire promoter has a repressive effect. The STA2 promoter, on the other hand, is in a less repressed state. Indeed, expression levels of the PSTA2-dependent reporter gene are consistently higher than those for the PMUC1-dependent reporter gene.

Second, the data show that overall variations in expression levels conferred by the entire promoters in a wild-type strain are much more important for the STA2 promoter than for the MUC1 promoter, even if the general regulation patterns are very similar. Since the STA genes encode extracellular glucoamylases and can therefore provide otherwise inaccessible nutrients, high expression levels and strong induction can obviously be advantageous to the cell. MUC1 expression, on the other hand, has to be more tightly controlled, since overexpression of the gene could result in profound physiological changes. MUC1 is essential for pseudohyphal differentiation and invasive growth, and both processes can be induced through overexpression of MUC1 from a heterologous promoter or, to a lesser degree, by multiple copies of the gene (8, 23, 28). From an evolutionary perspective, the changes between the two promoters have therefore allowed them to retain a similar regulation pattern, ensuring a coregulation of pseudohyphal differentiation and invasive growth with starch degradation, while allowing for much stronger production of glucoamylases.

The data show that the parameters which affect expression of both genes are (i) the presence or absence of a fermentable carbon source, (ii) the ploidy of the strain, and (iii) the presence or absence of several transcriptional regulators. As stated above, in all of these cases, we found that changes conferred by the STA2 promoter are generally more significant than those conferred by the MUC1 promoter.

The STA2 gene seems to have retained most specific regulatory elements but has evolved a less attenuated or less repressed promoter. This could indicate that the sequences which are found in the MUC1 promoter, but which are deleted in the STA1 to -3 promoters, are required for general repression. Our data suggest that this is indeed the case, since (i) the two inserts reduce STA2 expression strongly when present upstream of this gene and (ii) the 20-bp insert has a repressive effect, as the analysis of the UAS1 region clearly demonstrates. Our data in addition show that this repression is specifically linked to the Flo8p transcriptional activator. Multiple copies of FLO8 indeed result in strong production of glucoamylases when the STA2 gene is controlled by its own promoter, but fail to do so when the two MUC1 promoter inserts are present. Multiple copies of MSN1 and MSS11 do not result in a similar difference between the two promoters but efficiently increase STA2 expression in the presence or absence of the inserts. The repressive effect of these sequences might therefore depend on directly or indirectly inhibiting the Flo8p-dependent regulation of MUC1.

The sequence does not seem to confer a repressive effect on its own, but its regulatory activity seems context specific. When tested in the pHP41 plasmid, both the M20 and the M64 inserts reduce transcription of the reporter gene. However, and surprisingly, both sequences confer activation to a reporter gene when tested in a different reporter plasmid. The strong activation observed in the case of the pLG670-Z plasmid might be due to the creation of a spurious activation sequence, even if this hypothesis is difficult to reconcile with the fact that the two insert sequences do not present any homologies. These results could nevertheless suggest that the two inserts are the target of a DNA-binding protein, whose binding could result in repression in the specific context of the MUC1 gene promoter.

Our study of the entire promoters confirms that MUC1 and STA2 respond similarly to the deletion or the presence of multiple copies of MSN1 and MSS11. However, and as suggested by the effect of the MUC1 promoter inserts on STA2 expression, the responses of the two genes to multiple copies and deletion of FLO8 differ. Multiple copies of FLO8 result in strongly increased expression of the STA2 gene in medium containing either glucose or glycerol-ethanol as a carbon source but induce MUC1 expression only in medium containing glucose as a carbon source. Rupp et al. (44) showed that Flo8p was required for the cAMP-dependent regulation of invasive and pseudohyphal growth. The only physiologically significant variation in intracellular cAMP concentration is observed when glucose is added to cells grown on nonfermentable carbon sources (reviewed in references 14 and 49), and data suggest that the main role of cAMP could be the sensing of fermentable sugars. Our data could therefore indicate that Flo8p is required only for MUC1 induction during growth on substrates containing fermentable sugars, as is the case on nitrogen-limited SLAD medium, which is the main or only medium used for the assessment of filamentation by most authors. Flo8p could interact with other factors to induce filamentation during nitrogen limitation, when glucose levels are still high but might not be required or act differently under other conditions.

The size of the promoter, coupled to the apparent complexity of the regulatory processes, renders detailed molecular analysis of the entire promoter a difficult task. For most promoter studies with yeast published so far, a reasonable correlation between mechanistically (i.e., the binding of a regulatory protein to a specific sequence) and physiologically (i.e., the resulting change in transcription levels) relevant data can easily be achieved. However, in the case of the MUC1 and the STA1 to -3 genes, data suggesting specific molecular interactions and regulatory events in a specific area of the promoters might not result in the expected and corresponding changes in the overall transcriptional activity of the genes. The activating or repressing effect expected after the binding of a transcription factor to a region within the promoter might frequently be masked and covered by other regulatory signals acting through other areas.

Physiologically, the only significant data are those that relate to the activity of the promoter as a whole. However, in order to establish mechanistically relevant data concerning, for example, cis-acting transcription factor binding sites, it is necessary to dissect the promoter by using smaller sequence fragments. For purposes of analysis, these fragments are placed in a new, very different sequence context (i.e., plasmid sequences), and the data obtained have to be interpreted carefully when considering the effects on the native promoters. For this reason, we have focused our investigation on a small section of the STA2 and MUC1 promoters that combines several of the interesting features of the entire, intact promoters within a relatively short stretch of DNA. Our data show that this area (i) confers transcriptional regulation from a far-upstream (>1,000 nucleotides) position in the context of the native promoter and (ii) regulates reporter gene expression very similarly to the entire promoters when analyzed on its own. More specifically, this area of the MUC1 and STA2 promoters indeed (i) confers a general repressive effect on reporter gene transcription under most conditions and (ii) contains sequences responsible for both specific activation and specific repression. Furthermore, the area contains one of the two significant changes between the MUC1 and STA2 promoters.

Our data clearly establish that this additional sequence contributes to the general repression or attenuation of the MUC1 promoter, giving a clear indication of a molecular rearrangement during promoter evolution. In addition, the sequence is a target of glucose repression. The three transcriptional regulators investigated during this study, Flo8p, Msn1p, and Mss11p, all act, at least in part, via UAS1 to activate transcription of MUC1 and STA2. The deletion analysis pinpoints the sequences within UAS1 which confer these regulations, and these short sequences can now be investigated further to establish the binding sites of the factors involved. Flo8p and Mss11p clearly act in the 80-bp region between nucleotides 1160 and 1070 in the STA2 promoter and nucleotides 1210 to 1130 in the MUC1 promoter.

We also identify the STA10 repressive effect as being due to a mutation in the gene encoding the transcriptional regulator Flo8p. Indeed, we clearly demonstrate that a single copy of FLO8 in an S288C genetic background allows production of an amount of glucoamylase similar to that observed in naturally occurring starch-degrading strains. FLO8 was shown to be required for invasive growth in S288C-derived strains (26), since transformation of these strains with a single copy of wild-type FLO8 restored the ability to invade the growth medium. W303, another commonly used laboratory strain, contains, in addition to a mutation in FLO8, mutations in other activators required for invasive growth and pseudohyphal differentiation (26) and is therefore unable to form pseudohyphae or grow invasively. In this strain, a single copy of FLO8 was also unable to restore glucoamylase expression from a plasmid-borne STA2 gene (data not shown), suggesting that the STA10 phenotype in W303 strains might be due to the requirement of FLO8 as well as other transcriptional activator(s). We also show that Flo8p requires Mss11p to induce both starch degradation and pseudohyphal differentiation and invasive growth. Since Mss11p is able to overcome mutations in FLO8, we suggest that Mss11p is situated downstream of Flo8p in a linear signal transduction cascade. However, Flo8p apparently acts independently of Msn1p, which is probably situated in a parallel pathway.

As expected for such a complex promoter and as discussed above, some of the data obtained for UAS1 do not correlate properly with those seen for the entire promoter. Most tendencies are, however, conserved, and the data are mechanistically significant. Only a combination of studies of all UAS and URS sequences of the MUC1 and STA promoters will allow us to reveal a complete picture of how transcription factors combine to result in either repression or activation.