Yeast cells respond to the presence of amino acids in their environment by inducing transcription of several amino acid permease genes including AGP1, BAP2, and BAP3. The signaling pathway responsible for this induction involves Ssy1, a permease-like sensor of external amino acids, and culminates with proteolytic cleavage and translocation to the nucleus of the zinc-finger proteins Stp1 and Stp2, the lack of which abolishes induction of BAP2 and BAP3. Here we show that Stp1—but not Stp2—plays an important role in AGP1 induction, although significant induction of AGP1 by amino acids persists in stp1 and stp1 stp2 mutants. This residual induction depends on the Uga35/Dal81 transcription factor, indicating that the external amino acid signaling pathway activates not only Stp1 and Stp2, but also another Uga35/Dal81-dependent transcriptional circuit. Analysis of the AGP1 gene’s upstream region revealed that Stp1 and Uga35/Dal81 act synergistically through a 21-bp cis-acting sequence similar to the UASAA element previously found in the BAP2 and BAP3 upstream regions. Although cells growing under poor nitrogen-supply conditions display much higher induction of AGP1 expression than cells growing under good nitrogen-supply conditions, the UASAA itself is totally insensitive to nitrogen availability. Nitrogen-source control of AGP1 induction is mediated by the GATA factor Gln3, likely acting through adjacent 5′-GATA-3′ sequences, to amplify the positive effect of UASAA. Our data indicate that Stp1 may act in combination with distinct sets of transcription factors, according to the gene context, to promote induction of transcription in response to external amino acids. The data also suggest that Uga35/Dal81 is yet another transcription factor under the control of the external amino acid sensing pathway. Finally, the data show that the TOR pathway mediating global nitrogen control of transcription does not interfere with the external amino acid signaling pathway.
YEAST cells possess a plasma-membrane-associated sensor for detection of a wide variety of amino acids in their external environment. This sensor comprises Ssy1p, a protein highly similar in sequence to amino acid permeases, but apparently unable to mediate amino acid uptake. Ssy1p also differs from amino acid permeases by its much larger N-terminal cytosolic tail and its relatively low expression level (Didionet al. 1998; Jorgensenet al. 1998; Iraquiet al. 1999; Gaberet al. 2003). Also essential to the response of cells to external amino acids are the Ptr3 and Ssy5 proteins (Barneset al. 1998; Jorgensenet al. 1998; Klassonet al. 1999; Bernard and André 2001a) proposed to be integral parts of the sensor forming a complex with Ssy1 (Forsberg and Ljungdahl 2001a,b). The main function of this sensor complex, called SPS (Forsberg and Ljungdahl 2001b), is to respond to the presence of external amino acids by inducing expression of several genes encoding amino acid and peptide permeases. Among these are the AGP1, BAP2, and BAP3 genes (encoding wide-range-specificity amino acid permeases) and the ditripeptide permease gene PTR2 (Barneset al. 1998; Didionet al. 1998; Iraquiet al. 1999; Regenberget al. 1999). Experiments based on genome-wide microarray analyses have led to the suggestion that the Ssy1 signaling pathway influences the transcription of many other genes (Forsberget al. 2001; Kodamaet al. 2002). Another component essential for cells to respond to external amino acids is the SCFGrr1 ubiquitin-ligase complex (Iraquiet al. 1999; Bernard and André 2001b). The Ssy1 signaling pathway culminates with the proteolytic cleavage of two transcription factors, Stp1 and Stp2 (Andreasson and Ljungdahl 2002). These factors appear to be functionally redundant, contain a zinc-finger domain highly similar to that shared by the Kruppel-family transcription factors in higher eukaryotes, and play an important role in transcription of genes induced by external amino acids (Jorgensen et al. 1997, 1998; de Boeret al. 2000; Nielsenet al. 2001). Endoproteolytic processing of Stp1 and Stp2 is apparently required for these factors to be translocated from the cytosol into the nucleus (Andreasson and Ljungdahl 2002). In the stp1 stp2 double mutant, there is no induction of the BAP2 and BAP3 genes by amino acids, whereas residual induction persists in both single mutants (de Boeret al. 2000; Nielsenet al. 2001). A minimal sequence (25 bp) called UASAA, found in the promoter region of the BAP3 gene, has been shown to drive induction of reporter genes in response to external amino acids. A larger fragment (42 bp) including this sequence also displays UASAA properties, and its ability to activate reporter genes is entirely dependent on Stp1 (de Boer et al. 1998, 2000). A sequence of ∼100 bp isolated from the BAP2 upstream region displays similar UASAA properties (Nielsenet al. 2001). The UASAA elements of the BAP2 and BAP3 genes have in common the presence of 5′-CGGC-3′ doublets separated by three to six nucleotides. Mutagenesis experiments have confirmed that direct or inverted repeats of this tetranucleotide are crucial to UASAA function (de Boer et al. 1998, 2000). Furthermore, there is evidence that Stp1 and Stp2 can bind to these UASAA elements (Nielsenet al. 2001). In another study it was shown that yet another transcription factor, the Uga35/Dal81 protein containing a Gal4-type Zn(II)2-Cys6-cluster DNA-binding domain (Bricmontet al. 1991; Coornaertet al. 1991), is also essential for full induction of the AGP1 gene in response to amino acids (Iraquiet al. 1999; Bernard and André 2001a). Whether this factor also acts through the UASAA element remains undetermined. Uga35/Dal81’s role is not limited to transcription of Ssy1-regulated genes since it is also required for induction of genes involved in γ-aminobutyric acid (GABA), urea, and allantoin utilization (Turoscy and Cooper 1982; Jacobset al. 1985; Visserset al. 1990; Coornaertet al. 1991). Interestingly, experiments show that the Zn(II)2-Cys6-cluster-type DNA-binding domain of Uga35/Dal81 is not required for induction of at least some of its target genes (Bricmontet al. 1991). A similar situation has been described for TamA, an Aspergillus nidulans protein highly similar to Uga35/Dal81 and involved in expression of several genes in response to various nitrogenous compounds, e.g., GABA (Daviset al. 1996).
In this study we show that induction of the AGP1 gene by external amino acids depends mainly on the synergistic action of factors Stp1 and Uga35/Dal81, whereas Stp2 does not significantly contribute. Stp1 or Uga35/Dal81 alone can respond significantly to the Ssy1 pathway, suggesting that in addition to Stp1 and Stp2, Uga35/Dal81 is yet another transcription factor under the control of this pathway. Both Stp1 and Uga35/Dal81 act through a 21-bp sequence similar to the UASAA element. The detailed mutagenesis analysis of this element is presented. When cells grow under poor nitrogen-supply conditions, the positive effect of UASAA on AGP1 transcription is amplified ∼10-fold by yet another factor, the GATA factor Gln3. This amplification is negligible if a favored nitrogen source like ammonium is provided. We propose that Gln3 acts through neighboring 5′-GATA-3′-containing sequences, its action being conditioned by the prior action of Stp1 and Uga35/Dal81 through the adjacent UASAA element.
MATERIALS AND METHODS
Strains, growth conditions, and methods: The Saccharomyces cerevisiae strains used in this study are all isogenic with the wild-type Σ1278b (Béchetet al. 1970) except for the mutations mentioned (Table 1). Cells were grown in a minimal buffered (pH 6.1) medium with 3% glucose as the carbon source (Jacobset al. 1980). To this medium, urea (5 mm), proline (5 mm), (NH4)2SO4 (10 or 50 mm), amino acids (1–5 mm), or combinations of these compounds were added as source(s) of nitrogen. All procedures for manipulating DNA were standard ones (Ausubelet al. 1995; Sambrooket al. 1997; Susan-Resiga and Nowak 2003). The Escherichia coli strain used in our experiments was JM109.
Construction of stp1Δ, stp2Δ, and uga35/dal81Δ deletion strains: The stp1Δ, stp2Δ, and uga35/dal81Δ strains were constructed by the PCR-based gene-deletion method (Wach 1996). The DNA segments used to introduce these mutations were generated by using the kanMX2 gene from plasmid pFA6a-kanMX2 (Longtineet al. 1998) as a template and the following PCR primers: stp1Δ:kanMX2, D5-STP1, and D3-STP1; stp2Δ:kanMX2, D5-STP2, and D3-STP2; and uga35/dal-81Δ:kanMX2, D5-UGA35, and D3-UGA35 (Table 2). The yeast strain 23344c (ura3) was transformed with the PCR fragment by the lithium method (Itoet al. 1983) as described previously (Gietzet al. 1992). Transformants were selected on complete medium containing 200 μg G418·ml-1 (Geneticin; GIBCO BRL, Gaithersburg, MD).
Formation of 5′ and internal deletions in the AGP1 upstream region: The centromere- and URA3-based plasmid YCpAGP1-lacZ has been previously described (Iraquiet al. 1999). Plasmids pFA1 and pFA2 were constructed by inserting into the BamHI- and HindIII-cleaved YCpAJ152 plasmid (Andréet al. 1993) DNA fragments flanked by HindIII and BamH1 restriction sites and spanning the first five codons of AGP1 preceded by 537 bp (pFA1) or 513 bp (pFA2) of upstream sequences. These DNA fragments were obtained by PCR, using plasmid p16.2 bearing the AGP1 gene (Iraquiet al. 1999) as a template and the PCR primers 3-AGP1, 5-AGP1 + GA, and 5-AGP1 + GA + UASa (Table 2). Plasmid pFA3 is a derivative of YCp AGP1-lacZ from which the 21 bp corresponding to the UASAA element (see sequence in Table 4, line 7) have been specifically deleted. The deletion, spanning positions -516 to -538 relative to the ATG initiation codon, was introduced using the QuikChange XL site-directed mutagenesis kit (Stratagene, La Jolla, CA). A similar procedure was used to delete the region spanning positions -513 to -567 (pFA4). The primers (5-Del-UASaa, 3-Del-UASaa, 5-Del-UASc, and 3-Del-UASc) used for these mutagenesis experiments are shown in Table 2. The accuracy of each mutagenized plasmid was checked by sequencing.
UASAA-driven lacZ reporter plasmids: The episomal pFA10 plasmid (alias YEp-UASaa-CYC1-lacZ) was constructed from plasmid pLG670-Z (Guarente 1983). This vector bears a CYC1-lacZ fusion preceded by the CYC1 upstream region deleted of its natural UAS sequences (by excision of an internal XhoI-XhoI restriction fragment). The 21-bp sequence corresponding to UASAA (Table 4; plasmid pFA10) or the same UASAA preceded by 10 additional nucleotides (Table 4; plasmid pFA5) was inserted by site-directed mutagenesis into the reconstituted unique XhoI site of plasmid pLG670-Z. For this, we used the QuikChange XL site-directed mutagenesis kit (Stratagene), oligonucleotides 5- and 3-UASaa (pFA10) or 5- and 3-UASaa2 (pFA5; Table 2), and protocols recommended by the supplier. The correct insertion of sequences into pLG670-Z was checked by sequencing. The AGP1-(A to S) series of plasmids (Table 7) was constructed from pFA10 by similar site-directed mutagenesis. The list of oligonucleotides used for these constructions is shown in Table 2 (5- and 3-ΔAGP1-A to -S). Each mutagenized plasmid construct was checked by sequencing.
Yeast cell extracts and immunoblotting: Crude cell extracts were prepared as previously described (Heinet al. 1995). For Western blot analysis, equal quantities of proteins were loaded on an 8% SDS-polyacrylamide gel in a tricine system. After transfer to a nitrocellulose membrane (Shleicher and Schüll, Keene, NH), the proteins were probed with polyclonal antibodies raised against the N-terminal tail of Agp1 (1:5000; M. El Bakkoury and B. André, unpublished data) or Pma1 (1:1000) used as a loading control. Primary antibodies were detected with horseradish-peroxidase-conjugated anti-rabbit IgG secondary antibody (Amersham Pharmacia Biotech, Piscataway, NJ) followed by enhanced chemiluminescence (Roche Molecular Biochemicals, Indianapolis).
β-Galactosidase assays: All β-galactosidase assays were performed on cells having reached the state of balanced growth. β-Galactosidase activities were measured as described earlier (Andréet al. 1993) and are expressed in nanomoles of o-nitrophenol formed per minute per milligram of protein. Protein concentrations were measured with the Folin reagent and the standard used was bovine serum albumin.
Stp1 and Stp2 are not essential for AGP1 induction: Previous studies have shown that induction by leucine of the BAP2 and BAP3 genes does not occur at all in a stp1Δ stp2Δ mutant, whereas residual induction subsists in stp1Δ and stp2Δ single-mutant strains (de Boeret al. 2000; Nielsenet al. 2001). We thus tested the involvement of Stp1 and Stp2 in induction by phenylalanine of AGP1, another amino acid permease gene under the control of the Ssy1 pathway. Phenylalanine was chosen because this amino acid is one of the most potent inducers of AGP1. All experiments were carried out in a gap1Δ mutant lacking the general amino acid permease, so that the same strains could also be used in growth tests on amino acids used as the sole nitrogen source (see below). Induction of an AGP1-lacZ gene carried on a low-copy plasmid was reduced by >90% in the gap1Δ stp1Δ mutant as compared to the gap1Δ strain, but it was essentially normal in the gap1Δ stp2Δ mutant (Table 3). In the gap1Δ stp1Δ stp2Δ triple mutant, induction of AGP1 was much lower than that in the gap1Δ strain, but residual induction corresponding to ∼15% of the control was unexpectedly observed. Hence, the AGP1 gene is still induced by phenylalanine, although weakly, in the total absence of Stp1 and Stp2. These results were confirmed by means of growth tests indicative of Agp1 function. When an amino acid such as phenylalanine, isoleucine, or methionine is supplied at low concentration (1 mm) as sole nitrogen source, growth of a gap1Δ mutant requires a functional Agp1 permease since a gap1Δ agp1Δ strain does not grow on these media (Iraquiet al. 1999). A similar nongrowing phenotype is displayed by the gap1Δ ssy1Δ mutant, where the AGP1 gene is not induced by amino acids (Iraquiet al. 1999) (Figure 1). The growth test results presented in Figure 1 clearly show that the gap1Δ stp1Δ stp2Δ strain grows as well as the gap1Δ strain on all tested amino acids, indicating that the residual expression of AGP1 recorded in this strain (Table 3) is sufficient to sustain growth under these conditions. Residual induction of AGP1 in the gap1Δ stp1Δ stp2Δ strain was also confirmed in Western-blot experiments using antibodies against the Agp1 permease (Figure 2). In conclusion, Stp1 plays a major role in induction of AGP1 by amino acids. Yet in contrast to the situation described for the BAP2 and BAP3 genes, Stp2 does not significantly contribute to AGP1 expression, and significant induction of AGP1 can still occur in the total absence of both Stp1 and Stp2. This suggests that another transcription factor besides Stp1 and Stp2 can respond to the external amino acid sensing pathway to promote some degree of transcriptional induction of the AGP1 gene.
Stp1 acts synergistically with Uga35/Dal81 to promote AGP1 induction: In previous articles we reported that the Uga35/Dal81 protein [containing a Zn(II)2-Cys6-cluster-type domain] is required for full induction of AGP1 by amino acids (Iraquiet al. 1999; Bernard and André 2001a). We thus sought to test the relative contributions of Stp1, Stp2, and Uga35/Dal81 to AGP1 induction. In the gap1Δ uga35/dal81Δ mutant, induction of AGP1-lacZ by phenylalanine was severely reduced, but weak induction persisted (Table 3). In the gap1Δ uga35/dal81Δ stp1Δ mutant, induction was totally abolished (Table 3). Consistently, no detectable Agp1 signal was observed on immunoblots of the gap1Δ uga35/dal81Δ stp1Δ strain (Figure 2). Furthermore, while the gap1Δ stp1Δ stp2Δ and gap1Δ uga35/dal81Δ strains grow normally on amino acids supplied as the sole nitrogen source, the triple gap1Δ stp1Δ uga35Δ and quadruple gap1Δ stp1Δ stp2Δ uga35Δ mutant strains do not (Figure 1). These results show that Uga35/Dal81 plays an important role in AGP1 induction and is also responsible for residual AGP1 expression in the gap1Δ stp1Δ stp2Δ mutant. It thus appears that normal AGP1 induction requires, in addition to Stp1, activation via the Ssy1 pathway of another transcriptional activation circuit that is Uga35/Dal81 dependent. Finally, the fact that the gap1Δ stp1Δ uga35/dal81Δ and gap1Δ stp1Δ stp2Δ uga35/dal81Δ strains behave similarly in all experiments (Table 3; Figures 1 and 2) confirms that Stp2 does not significantly contribute to AGP1 induction. Hence, Stp1 and Uga35/Dal81 appear as the principal transcription factors mediating induction of AGP1 in response to external amino acids. In addition, they appear to be equally important, since induction of AGP1 is equally impaired after deletion of the STP1 or UGA35/DAL81 gene (Table 1).
Identification of the UASAA element of the AGP1 gene: We next addressed the question of whether Uga35/Dal81 and Stp1 exert their positive effects through the same or separate cis-acting sequences in the AGP1 gene’s upstream region. For this, we carried out experiments aimed at identifying the cis-acting sequences of the AGP1 gene responsible for its induction by amino acids. Previous work has revealed in the BAP3 gene’s upstream region a minimal 25-bp sequence (called UASAA) that is sufficient to drive induction of reporter genes by multiple amino acids (de Boer et al. 1998, 2000). This element contains the sequence 5′-CGGCN6GCCG-3′, and each 5′-CGGC-3′ quadruplet appears to be crucial to the UASAA’s induction-driving properties. Another study centered on the BAP2 gene (Nielsenet al. 2001) led to the isolation of a 65-bp DNA fragment also displaying UASAA properties. This fragment contains the 5′-CGG CN4CGGC-3′ sequence shown in mutagenesis experiments to determine UASAA properties. These studies indicate that direct or inverted repeats of 5′-CGGC-3′ are the core sequences of UASAA elements. Consistently, comparative analysis of the upstream sequences of several Ssy1 target genes (AGP1, BAP2, BAP3, PTR2, TAT1, TAT2, and GNP1) by means of the regulatory sequence analysis tools (http://rsat.ulb.ac.be/rsat/; van Heldenet al. 1998) reveals that the most recurrent oligonucleotide within these promoter regions is the 5′-CGGC-3′ sequence. Moreover, many such tetranucleotides are separated by two to nine nucleotides. Also, in support of the view that these 5′-CGGC-3′ sequences are important for transcriptional control of these genes, most of them tend to be conserved in orthologous gene promoters of closely related yeast species (Cliftenet al. 2003; Kelliset al. 2003). The upstream region of the AGP1 gene (up to position -780 with respect to the ATG initiation codon) contains 13 copies of the 5′-CGGC-3′ sequence. Deletion of the upstream region of AGP1 up to position -538 did not significantly reduce induction of the AGP1-lacZ gene by phenylalanine, but further deletion of 24 bp totally abolished this induction (Table 4, lines 1–3). This deletion eliminates the most upstream 5′-GCCG-3′ of a 5′-CGGCN6GCCG-3′ sequence. We thus specifically deleted this sequence from the full-length AGP1 upstream region. This caused strong reduction of induction, but the resulting AGP1-lacZ construct still responded weakly to the presence of amino acids (Table 4, lines 1 and 4). Just upstream from the deleted 5′-CGGCN6GCCG-3′ is the sequence 5′-CGGCN9CGGC-3′. When the latter sequence was also deleted, residual induction of the AGP1-lacZ was reduced another three-fold (Table 4, lines 1–5). The very weak residual induction displayed by this AGP1-lacZ construct (∼3% of the wild-type level) is likely determined by other 5′-CGGC-3′ core sequences still present in the AGP1 gene upstream region.
A 21-bp sequence from the AGP1 gene’s upstream region (between positions -536 and -515, containing the 5′-CGGCN6GCCG-3′ motif) was then inserted upstream from the CYC1-lacZ reporter gene lacking its own UAS. The resulting high-copy-number plasmid was introduced into the wild type and β-galactosidase activities were measured in cells growing either on urea or on the same medium with added phenylalanine (Table 4, lines 6–8). The results clearly show that the 21-bp sequence drives high-level transcriptional induction in response to phenylalanine (lines 6 and 7). A similar result was obtained when the 21-bp sequence was extended at its 5′ end with 10 additional nucleotides comprising a 5′-GCCG-3′ tetranucleotide (line 8). Further experiments showed that the 21-bp sequence can drive transcriptional activation in response to many other amino acids, the induction level varying markedly according to the amino acid (Table 5). This 21-bp sequence thus has the properties of a UASAA element. For many amino acids like phenylalanine, threonine, leucine, or isoleucine, the responses displayed by the lacZ constructs under the control of the UASAA alone or the entire AGP1 upstream region (AGP1-lacZ gene) are similar; i.e., amino acids causing strong induction of the former also cause strong induction of the latter (Table 5). However, some amino acids like valine and tryptophan are strong inducers of the AGP1-lacZ gene but relatively weak inducers of UASAA-driven transcription. In contrast, others like serine or histidine are rather good inducers of UASAA-dependent transcription although relatively weak inducers of AGP1-lacZ.
Finally, experiments showed that induction mediated by the UASAA element isolated from the AGP1 gene is entirely dependent on the Ssy1, Ptr3, and Ssy5 factors (Table 6). In previous articles, we reported that components of the SCFGrr1 ubiquitin-ligase complex are also required for proper induction of the AGP1 gene (Iraquiet al. 1999; Bernard and André 2001b). The results of Table 6 show that the F-box protein Grr1 is also essential for UASAA function. This observation is consistent with the view that SCFGrr1 is an integral part of the external amino acid signaling pathway.
Mutational analysis of the UASAA element of the AGP1-gene: Site-directed mutagenesis was used to further analyze the 21-bp UASAA element found upstream from the AGP1 gene (Table 7). Single-base replacement of any of the four nucleotides within one or both of the inversely repeated 5′-GCCG-3′ quadruplets drastically reduced induction by amino acids (Table 7, lines 2–8). The same was observed when a single quadruplet was replaced with 5′-ACAT-3′ or 5′-ATGC-3′ (lines 9–11). This clearly shows that at least two copies of the 5′-GCCG-3′ sequence are needed to constitute a UASAA responding strongly to amino acids. In the native UASAA, the quadruplets are separated by six nucleotides. When this number was reduced to two, the UASAA lost its activation potency (line 12); the UASAA resulting from increasing the number of separating nucleotides to eight (lines 13 and 14) was also less active. It thus seems that six nucleotides is the optimal distance between the two inversely repeated 5′-GCCG-3′ quadruplets, although derived UASAA elements with greater spacing can still respond significantly to amino acids. Replacement mutations were also introduced at sites directly flanking the quadruplets, sites among the six nucleotides separating them, or both (lines 15–17). Again, both types of mutation caused a net reduction of expression, but significant induction of the lacZ reporter gene may subsist. When these mutations were combined in the same construct, induction was entirely lost (line 17). It thus seems that the direct environment of the repeated 5′-GCCG-3′ sequences is important for the optimal functioning of the UASAA. Finally, we tested whether direct instead of inverse 5′-GCCG-3′ repeats could also promote induction. Again this caused a significant reduction of expression, but significant induction was maintained (line 18). It thus seems that the orientation, spacing, and immediate context of the 5′-GCCG-3′ repeats of the UASAA element of the AGP1 gene are optimal for high-level induction in response to amino acids, but that other arrangements are also possible, leading to lower but significant induction levels.
Stp1 and Uga35/Dal81 are both essential to UASAA function: We next monitored transcriptional induction driven by the native UASAA element in gap1Δ cells with different combinations of STP1, STP2, and UGA35/DAL81 gene deletions (Table 8). Deletion of STP2 did not affect induction of the reporter gene in response to phenylalanine. In contrast, deletion of STP1 or UGA35/DAL81 completely suppressed this induction. Hence, Stp1 and Uga35/Dal81 act through the same cis-acting sequence (UASAA) and both are absolutely essential to its ability to drive transcriptional induction. This situation somewhat differs from that of the lacZ reporter under the control of the entire AGP1 upstream region, since in the latter case weak but significant induction persisted in cells lacking only one gene, STP1 or UGA35/DAL81 (Table 3). Perhaps Stp1 and Uga35/Dal81 can act alone via other 5′-CGGC-3′ sequences of lesser importance in the AGP1 upstream region, e.g., those responsible for weak residual induction of AGP1 gene expression when the UASAA element characterized above is deleted (Table 4, lines 4 and 5).
AGP1 induction by amino acids is under nitrogen catabolite repression: Regenberg et al. (1999) reported a two-times-higher level of AGP1-lacZ induction by leucine (0.23 mm) in cells growing on proline than in cells growing on ammonium as the source of nitrogen. This suggests that AGP1 is controlled by nitrogen catabolite repression (NCR), even though the negative effect of ammonium was limited (probably because the strain used in these experiments—a derivative of S288C—is largely insensitive to this regulation (Magasanik and Kaiser 2002). We repeated these experiments in strains derived from the Σ1278b wild type, known to respond normally to ammonium regulation. We found induction of AGP1-lacZ gene expression by phenylalanine to be considerably stronger in cells growing on urea (a poor nitrogen source) than in cells growing on 20 mm ammonium (a good nitrogen source; Table 9, line 1). When a fivefold higher ammonium concentration (100 mm) was used, the induction level was further halved (data not shown). Hence, induction of AGP1 by external phenylalanine is highly sensitive to the nitrogen status of the cell. Genes subject to this global regulation (NCR) are known to contain several 5′-GATA-3′ core sequences in their upstream regions, capable of binding several transcription factors of the GATA family (Cooper 2002; Magasanik and Kaiser 2002). The Gln3 factor is the GATA factor playing the main positive role in transcription of genes subject to NCR (Stanbroughet al. 1995). On good nitrogen sources, Gln3 is sequestered in the cytoplasm, but on poor ones, it is translocated into the nucleus and activates transcription of genes via upstream 5′-GATA-3′ sequences (Beck and Hall 1999). Consistently with AGP1 being subject to NCR, this gene’s upstream region contains multiple 5′-GATA-3′ sequences. In the gln3Δ mutant, induction of the AGP1-lacZ reporter gene was severely reduced (Table 9, line 2). The Ure2 protein is required for sequestration of Gln3 in the cytoplasm under good nitrogen-supply conditions (Beck and Hall 1999). In the ure2Δ mutant, AGP1-lacZ was induced to a much higher level on ammonium (Table 9, line 3). The negatively acting GATA factor encoded by the GZF3/DEH1/NIL2 gene is also required for normal NCR of genes (Coffmanet al. 1997; Rowenet al. 1997; Soussi Boudekou et al. 1997). In the gzf3Δ mutant also, AGP1-lacZ was induced to a significantly higher level on ammonium, and in the ure2Δ gzf3Δ double mutant, this induction was still higher (Table 9, lines 4 and 5). Experiments were also performed with mutants lacking the two other GATA factors involved in controlling NCR-sensitive genes (DAL80/UGA43 and NIL1/GAT1), but AGP1 expression was not significantly affected in these strains (data not shown). Taken together, these data clearly show that induction of AGP1 by external amino acids is subject to NCR and that the main trans-factors involved in this control are Gln3, Ure2, and, to a lesser extent, Gzf3/Deh1/Nil2.
The UASAA element is insensitive to nitrogen regulation: The same experiments were repeated with cells transformed with the lacZ reporter gene under the control of the UASAA element isolated from the AGP1 gene (Table 9). The results clearly show that UASAA-driven transcriptional induction is not significantly different in cells growing on urea or ammonium medium. Furthermore, a lack of Gln3, Ure2, or Gzf3 did not affect this expression. These results thus show that UASAA-driven transcription is totally insensitive to the NCR status of the cell. As target of rapamycin (TOR) kinases regulate GATA factor function in response to nutrient availability (Cooper 2002; Crespo and Hall 2002), these data also suggest that the TOR pathway does not interfere with the external amino acid signaling pathway. Induction of AGP1-lacZ is nevertheless much greater on urea medium than on ammonium medium, and a lack of Gln3 on urea medium leads to loss of ∼85% of AGP1 induction (Table 9, lines 1 and 2). This suggests that while Stp1 and Uga35/Dal81 play a crucial role in AGP1 induction, Gln3 is responsible for this induction reaching much higher levels in cells grown under poor nitrogen-supply conditions.
Transcriptional induction of yeast genes in response to external amino acids requires the permease-like protein Ssy1 and the membrane-associated Ptr3 and Ssy5 factors. According to a recent model, these factors are associated in a sensor complex (SPS) able to detect external amino acids and in turn to activate by endoproteolytic processing two transcription factors, Stp1 and Stp2, which are then translocated into the nucleus to activate gene transcription (Andreasson and Ljungdahl 2002). Stp1 and Stp2 appear to be redundant since a residual transcriptional induction of the BAP2 and BAP3 genes in response to leucine subsists in both single stp1Δ and stp2Δ mutants whereas no induction is detected in the double stp1Δ stp2Δ strain (de Boeret al. 2000; Nielsenet al. 2001). The functional redundancy of Stp1 and Stp2 was also illustrated by the observation that the growth phenotypes of the stp1Δ stp2Δ and ssy1Δ mutant strains on several selective media were indiscernible, whereas the stp1Δ and stp2Δ single mutants behaved like the wild type (Andreasson and Ljungdahl 2002). In this study, we show that transcription of the AGP1 gene encoding another amino acid permease induced by external amino acids (Iraquiet al. 1999) is mainly dependent on Stp1, whereas Stp2 does not significantly contribute to this expression. This means that Stp1 and Stp2 are not functionally redundant in the case of AGP1 gene expression and this raises the question of the specific role of each Stp factor in the response of cells to external amino acids. We further show that Stp1 acts synergistically with another transcription factor, namely Uga35/Dal81: lack of Stp1 or Uga35/Dal81 dramatically reduces induction of AGP1 by amino acids whereas lack of both transcription factors results in total loss of induction. The Stp1 factor appears to be specifically involved in the transcription of genes controlled by the Ssy1 pathway. In contrast, Uga35/Dal81 is also required for induction by GABA of the UGA genes (utilization of GABA; Visserset al. 1990; Coornaertet al. 1991) and for induction by allophanate of the DUR and DAL genes (degradation of urea and allantoin; Turoscy and Cooper 1982; Jacobset al. 1985). In both of these induction processes, Uga35/Dal81 acts together with an inducer-specific transcription factor, namely Uga3 (GABA induction; André 1990) or Dal82/DurM (allophanate induction; Jacobset al. 1985; André and Jauniaux 1990; Oliveet al. 1991). Furthermore, each pair of transcription factors acts through regulon-specific UAS elements (Dorrington and Cooper 1993; Talibiet al. 1995). We here report a similar situation, i.e., that Stp1 and Uga35/Dal81 mediate induction of the AGP1 gene by acting via a UASAA element (see below). Uga35/Dal81 thus seems to act in tight conjunction with various pathway- or inducer-specific transcription factors (Stp1, Uga3, and Dal82/DurM) via specific UAS elements to promote transcriptional induction of nitrogen-utilization genes in response to various stimuli.
A significant induction of the AGP1 gene by amino acids thus subsists in the stp1Δ and stp1Δ stp2Δ mutants and this induction is entirely dependent on Uga35/Dal81. Such residual induction, however, is not observed in ssy1Δ, ptr3Δ, ssy5Δ, or grr1Δ mutants. Two models may account for this observation. First, Uga35/Dal81 might act in conjunction with yet another transcription factor, which, like Stp1 and Stp2, would be activated by external amino acids via the SPS sensor complex. The yeast proteome includes two proteins, Stp3 and Stp4, which, like Stp1 and Stp2, harbor a Kruppel-type zinc-finger domain. Yet our experiments failed to show a role of these factors in AGP1 induction: the stp1Δ stp2Δ stp3Δ stp4Δ quadruple mutant still displays Uga35/Dal81-dependent residual induction of the AGP1 gene in response to phenylalanine (our unpublished data). The putative factor acting with Uga35/Dal81 to induce AGP1 transcription in the stp1Δ stp2Δ mutant would thus belong to another family of transcription factors. A second model consistent with our data is that Uga35/Dal81 itself might respond to the external amino acid signaling pathway. In other words, this factor alone might be able to activate AGP1 transcription in response to external amino acids, its effect on gene transcription being much stronger in conjunction with Stp1. This would mean that the external amino acid signaling pathway bifurcates into two branches, one mediating activation by endoproteolytic processing of Stp1 and Stp2 and another leading to activation of Uga35/Dal81 through an unknown mechanism. This hypothesis immediately raises the question: Why would Uga35/Dal81 be involved also in GABA- and allophanate-induced transcription? Sharing of a transcription factor between different induction processes might allow the establishment of a hierarchy between them when the inducers of these pathways are simultaneously present in the medium. For instance, it is possible that upon activation by the Ssy1 pathway, Uga35/Dal81 might be recruited specifically to the promoters of genes induced by external amino acids, rather than to those responding to GABA or allophanate. In support of this view, addition of amino acids to cells growing on urea leads not only to induction of Ssy1-regulated genes but also to downregulation of the allophanate-inducible DUR genes involved in urea utilization (our unpublished data). Further experiments using cells growing on combinations of nitrogen compounds will be needed to test the validity of this model. Finally, three other properties of Uga35/Dal81 are worth a comment. First, whereas Stp1 is translocated into the nucleus in the presence of amino acids (Andreasson and Ljungdahl 2002), Uga35/Dal81 stays in the nucleus whether amino acids are present in the medium or not (our unpublished data). Second, experiments show that the Zn(II)2-Cys6-cluster-type DNA-binding domain of Uga35/Dal81 is not required for its role in allophanate-induced transcription (Bricmontet al. 1991). A similar situation has been described for TamA,an A. nidulans protein highly similar to Uga35/Dal81 and involved in expression of several genes in response to various nitrogenous compounds, e.g., GABA (Daviset al. 1996). It will thus be interesting to test the role of the Zn(II)2-Cys6 cluster domain of Uga35/Dal81 in induction of the AGP1 gene by external amino acids. Third, among the many yeast transcription factors sharing a Gal4-type Zn(II)2-Cys6 cluster domain, Uga35/Dal81 is the factor whose cluster domain most closely resembles that of the Rgt1 transcription factor (D. Coornaert, personal communication) controlled by the external glucose signaling pathway of S. cerevisiae (Ozcan and Johnston 1999). In this pathway, the transporter-like glucose sensors Snf3 and Rgt2 respond to the presence of external glucose to modulate the function of Rgt1, and this signaling pathway also involves the SCFGrr1 ubiquitin-ligase complex (Li and Johnston 1997). Together with our data suggesting that Uga35/Dal81 could be a direct target of the Ssy1 pathway, these observations raise the possibility that the last steps of the external amino acid and glucose signaling pathways might share more structural similarities than thought thus far.
Our results show that Stp1 and Uga35/Dal81 synergistically activate AGP1 transcription by acting through the same cis sequence, a UASAA element similar to those previously found upstream from the BAP3 and BAP2 genes (de Boeret al. 2000; Nielsenet al. 2001). Furthermore, induction by leucine of the BAP2 gene is impaired in the uga35/dal81 mutant (Bernard and André 2001a). It is thus likely that UASAA, Stp1 and/or Stp2, and Uga35/Dal81 are the key regulatory elements involved in transcriptional induction by external amino acids of all Ssy1-regulated genes. The UASAA of AGP1 consists of two inverse copies of the 5′-CGGC-3′ sequence spaced by six nucleotides. Our mutagenesis experiments suggest that the situation created by the inverse orientations, spacing by six nucleotides, and flanking nucleotides of the 5′-CGGC-3′ doublet of the AGP1 gene, is optimal for high-level induction by external amino acids. On the other hand, these experiments also showed that many mutations in this UASAA are permissible, since they strongly reduce but do not completely suppress the activating power of UASAA. Although two copies of the 5′-CGGC-3′ sequence appear crucial to UASAA function, these can be oriented as direct instead of inverse repeats and they can be spaced by more nucleotides. Furthermore, although the nucleotides between the 5′-CGGC-3′ repeats and those flanking them are also important for the function of the UASAA, their replacement with other nucleotides leads to reduction rather than to complete loss of induction.
Finally, we showed that transcription dependent on UASAA alone is totally insensitive to the nitrogen status of the cell and to mutations in GLN3, URE2, or GZF3/DEH1. This suggests that the TOR signaling pathway mediating nitrogen-source control of transcription does not interfere with the external amino acid signaling pathway. Transcription driven by UASAA alone in response to amino acids (Table 5) thus likely reflects the actual external amino acid sensing capability of the cell. Nevertheless, in accord with previous conclusions (Regenberget al. 1999), our results showed that AGP1 transcription is under tight nitrogen control. This control is mediated by the GATA factor Gln3 that enhances ∼10-fold the level of AGP1 transcriptional induction when cells grow under limiting nitrogen supply conditions. Under these conditions, Gln3 is reported to be translocated into the nucleus, where it binds to 5′-GATA-3′ sequences to activate gene transcription (Beck and Hall 1999). Unlike other permease genes under Gln3 control (GAP1, DAL5, or MEP2), however, expression of AGP1 absolutely requires the presence of an inducer, i.e., external amino acids. Hence, Gln3 present in the nucleus is apparently unable to trans-activate AGP1 expression without the prior action of Stp1 and Uga35/Dal81 through the UASAA element. These results suggest that Gln3 is somehow recruited for positive action by Stp1 and/or Uga35/Dal81. Interestingly, studies on TamA of A. nidulans show that this factor can interact with and recruit Area, a GATA-family transcription factor functionally related to Gln3 (Smallet al. 1999). Direct interaction between Uga35/Dal81 and Gln3 might thus account for the probable recruitment of Gln3 for a positive effect on AGP1 transcription. Such a mechanism has also been proposed to account for the much higher induction by GABA of the UGA4 gene under nitrogen derepression than under nitrogen repression (Andréet al. 1995; Talibiet al. 1995). When cells are grown on a good nitrogen source, Gln3 is sequestered in the cytoplasm in a manner dependent on the Ure2 factor (Beck and Hall 1999). This unavailability of Gln3 in the nucleus is likely the main cause of the much lower AGP1 induction in ammonium medium (Table 9), this weak induction likely reflecting the actual activating power of Stp1 and Uga35/Dal81 via the UASAA element.
In conclusion, this study shows that transcriptional induction of the AGP1 gene in response to external amino acids involves SPS-dependent activation of two transcription factors, Stp1 and Uga35/Dal81, acting synergistically via the UASAA element. Furthermore, the amplitude of this induction is modulated according to nitrogen availability by the Gln3 factor, with poor nitrogen-supply conditions leading to much higher induction. It will be interesting to compare the physiological conditions and molecular mechanisms mediating activation of the Stp1 and Uga35/Dal81 pathways and to extend this work to other amino acid permease genes under Ssy1 control.
We gratefully acknowledge the excellent technical assistance of Catherine Jauniaux. We thank S. Vissers and D. Coornaert for a critical reading of the manuscript. We also thank members of the laboratory and also Dominique Thomas for many fruitful discussions. This work was supported by grant FRSM 3.4597.00 F for Medical Scientific Research, Belgium, and by grant 98/03-223 of the Communauté Française de Belgique, Direction de la Recherche Scientifique.
Communicating editor: A. P. Mitchell
- Received October 30, 2003.
- Accepted January 14, 2004.
- Copyright © 2004 by the Genetics Society of America