INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The yeast Saccharomyces cerevisiae can use tryptophan, phenylalanine, or tyrosine as the only source of cellular nitrogen (11). The main products of their catabolism are tryptophol, phenylethanol, and tyrosol, respectively, constituents of the mixture of alcohols collectively known as fusel oil in fermentations (41, 42, 66, 69). Fusel oil formation from amino acids is believed to proceed through the so-called Ehrlich pathway involving three enzymatic steps. A first transamination produces the -keto-acid analog of the amino acid, a decarboxylation step yields an aldehyde, and a reduction step converts the aldehyde to a primary alcohol (79). Tryptophan is thus converted to tryptophol via the metabolic intermediates indole-3-pyruvate and indole-3-aldehyde (41, 69); tyrosine and phenylalanine are converted to tyrosol and phenylethanol in a similar way (42, 66). Yet as recently stressed for leucine catabolism by Dickinson et al. (15), this general scheme of amino acid degradation stems mainly from studies of metabolic intermediates and end product formation. Too little is known about the specific permeases and enzymes involved in aromatic amino acid utilization, their genetic determinants, and their regulation. Aromatic aminotransferase II, the product of the recently cloned ARO9 gene (34), is the first characterized enzyme proposed to be involved in the catabolism of aromatic amino acids. Specifically, (i) ARO9 gene transcription is induced by the presence of tryptophan, phenylalanine, or tyrosine in the growth medium and remains at a very low level in their absence (34, 41, 77); (ii) aromatic aminotransferase II catalyzes the first step of the catabolism of these amino acids (41, 77); (iii) the enzyme is dispensable for growth on phenylalanine or tyrosine as the only source of nitrogen, but aro9 mutants grow poorly on tryptophan or kynurenine (77); and (iv) preliminary expression studies (34) show that, in the presence of inducer, ARO9 expression levels are more than 10 times higher on urea medium than on ammonia medium, indicating that ARO9 is subject to an ammonia effect (80). Thus, as originally proposed by Kradolfer et al. (41), the physiological function of aromatic aminotransferase II is most likely to participate in the catabolism of aromatic amino acids, mainly tryptophan.

Like many other nitrogen-catabolic genes, ARO9 appears to be regulated in at least two complementary ways: by an induction mechanism, with aromatic amino acids acting as inducers, and by an ammonia effect which modulates expression levels according to nitrogen source quality.

In S. cerevisiae, the synthesis of permeases and enzymes of many nitrogen-catabolic pathways is inducible. Examples include the pathways of arginine, proline, urea, and allantoin (for reviews, see references 47, 52, and 80) and 4-aminobutyrate (reference 76 and references therein) and serine/threonine (31) utilization. Induction is mediated in cis by a class of upstream activating sequences (UASs), specific to genes of a particular catabolic pathway. It has been demonstrated in several cases that these UASs are binding sites for specific trans-acting factors that can stimulate transcription of adjacent genes in response to inducer availability (31, 70). Many of these regulatory proteins, such as ArgRIIp (53), Uga3p (1), Put3p (48), and Cha4p (31), belong to the C₆ zinc cluster family of fungal transcription factors, whose prototypes are Gal4p (50, 60) and Prp1p (51). As deduced from the analysis of the complete genome sequence, the family comprises no less than 56 members in S. cerevisiae, most of them of unknown function (64).

Inducer availability is by no means sufficient to ensure high-level transcription of the genes of a nitrogen catabolic pathway. S. cerevisiae can use some 30 different compounds as the sole source of cellular nitrogen (11). Some nitrogen sources such as glutamine and ammonia (good or preferred nitrogen sources) support optimal growth, while others, such as proline and urea (poor or secondary nitrogen sources), support slower growth. When a good nitrogen source is present, utilization of the poorer nitrogen sources is prevented. This general phenomenon, termed the ammonia effect (80) or nitrogen regulation (47), is the result of at least two distinct regulatory mechanisms: nitrogen catabolite repression (NCR), which affects the synthesis of enzymes and permeases responsible for the utilization of secondary nitrogen sources, and nitrogen catabolite inactivation (NCI), which down-regulates the activity of several permeases that import these substrates (24, 47, 80). NCR of susceptible genes is achieved through upstream 5'-GATA-3' sequences variously characterized as UAS_NTR (58, 59), UAS_N (55), or UAS_GATA (4). Four distinct GATA factors can bind to these UASs and affect transcription: activators Gln3p and Nil1p/Gat1p and repressors Uga43p/Dal80p and Gzf3p/Nil2p/Deh1p (references 10, 62, and 71 and references therein). In addition, nitrogen repression of most NCR-sensitive genes is relieved by mutations in the URE2/GDHCR/USU locus (16, 22, 25). Ure2p is not a GATA factor, does not bind to DNA, and seems to act mainly by inactivating Gln3p, the main activator of nitrogen regulation (47). Hence, nitrogen source repression involves negative factors counteracting the function of positive factors when preferred nitrogen sources are available.

NCI, on the other hand, affects the activity of most permeases subject to NCR, including Gap1p and the proline permease Put4p (24). Addition of ammonium ions to cells growing on poor substrates such as proline or urea triggers rapid and complete inactivation (23) and degradation (30) of Gap1p. This process requires the Npi1p/Rsp5p ubiquitin ligase and the integrity of the C-terminal region of the permease (29, 30, 72). Ubiquitination is required for endocytic internalization of the permease and its subsequent degradation in the vacuole (72). In npi1 mutants, Gap1p ubiquitination is impaired and the permease remains located in the plasma membrane and active (23, 72).

Here we have sought to identify regulatory elements involved in the induction and nitrogen regulation of ARO9 expression. We have identified two elements required for specific induction of the ARO9 gene by aromatic amino acids: a UAS (UAS_aro) and the transcriptional activator Aro80p. We have further shown that the ARO10 (YDR380w) gene, encoding a putative indole-3-pyruvate decarboxylase, possesses an upstream sequence similar to UAS_aro and that this gene is also subject to Aro80p-dependent induction by tryptophan. Nitrogen control analyses indicate that the general factors of nitrogen transcriptional regulation do not directly affect ARO9 expression. Transcription of this gene appears, rather, to be regulated essentially through inducer exclusion. A preliminary report of this work has already appeared (32).

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Strains, media, and methods. The S. cerevisiae strains used are listed in Table 1. All are isogenic with wild-type strain 1278b (MAT) (6). The laboratory's standard minimal medium has been described previously (36); the nitrogen source was either 10 mM ammonium sulfate or another nitrogenous compound at 1 g/liter; the carbon source was 3% glucose. Inducing media were supplemented with 0.5 g of tryptophan, phenylalanine, or tyrosine per liter. When needed, 50 µg of uracil or leucine per ml was added to the medium. Complex medium 868 has been described previously (17). Standard yeast genetic techniques were applied (68). Yeast cells were transformed by the lithium acetate method (21, 35). The Escherichia coli strain used in this work was JM109. All E. coli and DNA manipulations were performed according to standard procedures (5, 63).

Formation of 5' and internal deletions in the ARO9 upstream region. Construction of the YCpARO9-lacZ plasmid has been described previously (34). In the plasmid, the upstream ARO9 DNA fragment is flanked at the 5' end by a portion of the pFL38 polylinker. To isolate the 5' nested deletions of the ARO9-lacZ reporter gene, the plasmid was cleaved at the polylinker BamHI and SphI restriction sites, treated with exonuclease III and S1 nuclease, and religated, as instructed for the double-stranded nested deletion kit (Pharmacia LKB).

View larger version (49K): [in this window] [in a new window]   FIG. 1.   (A) Sequence of the 5'-flanking region of the ARO9 gene. The sequence includes the SPL2 (YHR136c) gene, recently shown to encode a novel inhibitor of the Pho80p/Pho85p cyclin-dependent protein kinase (20). Putative TATAAA sequences are underlined, GAT(A/T)(A/G) (75) sequences are shaded, and the UASaro sequence is boxed. (B) Sequence of the UASaro indispensable for ARO9 gene induction. The arrows indicate the CCG triplets. (C) Comparison of UASaro and a similar sequence located upstream from the ARO10 gene. The arrows indicate the direct repeats.

Plasmids carrying heterologous lacZ fusions. Plasmid PAL1-15 carries a heterologous UAS_aro-CYC1-lacZ fusion gene in which the TATA box and transcriptional start site are provided by the CYC1 gene and the UAS is the UAS_aro element of the ARO9 gene. Plasmid PAL1-15 was constructed by introducing UAS_aro at the XhoI restriction site position of the CYC1 promoter in plasmid pLG670-Z (28). This was done in several steps with the Altered Sites in vitro mutagenesis system (Promega). First, a BamHI-SalI fragment of the CYC1 promoter was isolated and inserted into the pALTER-1 phagemid restricted with BamHI and SalI. UAS_aro was then introduced into the CYC1 promoter by directed mutagenesis with an oligonucleotide comprising the 36-bp element flanked by targeting sequences of CYC1. The correct insertion of UAS_aro into pALTER-1 was checked by sequencing, and finally the BamHI-SalI fragment containing the CYC1-UAS_aro-CYC1 construct was isolated from pALTER-1 and inserted into plasmid pLG670-Z (28), yielding plasmid PAL1-15. Using appropriate oligonucleotides, we followed the same procedure to obtain the PAL series of heterologous fusions carrying variously altered forms of the 36-bp sequence (PAL60, PAL50, PAL2-13, and PAL2-10). PAL1-31 and PAL1-46 were fortuitously obtained in the course of PAL1-15 construction. All the constructs were sequenced on both strands.

Isolation of mutants affected in the ARO9 gene induction. Cells of the ura3 strain 30332d bearing episomal plasmid PAL1-15 were mutagenized with ethane methylsulfonate and plated on proline minimal medium supplemented with 500 µg of tryptophan per ml and 40 µg of X-Gal (5-bromo-4-chloro-3-indolyl--D-galactopyranoside) per ml, the medium being buffered at pH 7.0 (28). About 32,500 colonies were screened, white or pale blue colonies were restreaked, and 17 isolates were chosen for enzymatic tests.

Isolation, subcloning, and deletion of the WAP1 (YCL025c) gene. S. cerevisiae 32102c, derived from IHI611, carries mutations gap1-92 and ura3, plus a mutation affecting an unknown locus (see Results). The strain is unable to grow on minimal medium containing either tryptophan or phenylalanine at a 500-µg/ml concentration as the only source of nitrogen. The mutant was transformed with a genomic library from wild-type strain 1278b (49), and transformants growing on tryptophan or phenylalanine were selected. A second cycle of transformation was carried out with purified plasmids, and the extremities of the inserts were sequenced. From transformants selected on tryptophan, only genomic fragments carrying the GAP1 gene (38) were obtained (four different complementing clones). From transformants selected on phenylalanine, 11 different genomic fragments were analyzed. Of these, one carried the GAP1 gene. The 10 other genomic inserts were partially overlapping fragments from a region of chromosome III (56). YCL025c was the only open reading frame common to all of them. YCL025c was isolated as a 3-kb BsmI-BsmI subfragment and cloned into vector pFL38 (8) at the SmaI restriction site. The resulting pII406 plasmid, when introduced into cells of strain 32102c, restored their ability to grow on phenylalanine medium.

Isolation, subcloning, and deletion of the ARO80 (YDR421w) gene. S. cerevisiae 32054d (aro80-2 aro8-1 ura3), auxotrophic for phenylalanine and tyrosine, was transformed with a wild-type 1278b genomic library (49), and prototrophic transformants were selected. Among them, two different growth phenotypes were observed on aromatic amino acid substrates because both aro80 ARO8 and ARO80 aro8 single mutant strains are prototrophs. Only transformants displaying the Aro8 phenotype (77) were further analyzed, yielding complementing plasmids pII242 and pII251. Sequencing revealed that their inserts overlapped over almost their entire length and came from a region of chromosome IV extending from bp 1305529 to bp 1316570 (37). This chromosomal region comprises two internal open reading frames: YDR420w, corresponding to the HKR1 gene, and YDR421w (see Results). The function of the latter is unknown, but the deduced polypeptide contains a Zn₂Cys₆ binuclear cluster domain and other features characteristic of Gal4p family transcriptional regulators (64). YDR421w was cloned as a 5.6-kb BamHI-BamHI fragment extracted from plasmid pII242 and inserted into the BamHI restriction site of centromeric plasmid pFL38 (8). The resulting pII270 plasmid complemented the auxotrophy of strain 32054d (aro80-2 aro8-1 ura3) and the growth defects of the aro80 mutants on medium containing an aromatic amino acid as the sole source of nitrogen.

Tryptophan uptake. The time course of L-[side chain-3-¹⁴C]tryptophan (specific radioactivity, 56 mCi/mmol; from DuPont NEN) uptake was determined as previously described for 4-aminobutyric acid accumulation (27), with radiolabeled amino acids at a 20 µM final concentration.

Enzyme assays. Tryptophan-phenylpyruvate transamination activity was assayed as described elsewhere (77). -Galactosidase was assayed according to the method of Miller (54); the unit used was nanomoles of o-nitrophenol formed per minute per milligram of protein. Protein concentrations were determined according to the method of Lowry et al. (43) with bovine serum albumin as the standard.

RNA preparation and gene expression analysis by RT-PCR. Cells were grown to a density of approximately 10⁷ cells/ml on minimal medium with the nitrogen source specified in the legends to the figures. Total RNA was isolated as instructed for the RNeasy Mini kit by the supplier (Qiagen); the final RNA concentration was approximately 1 µg/µl. The purified RNA was treated with RNase-free DNase I (Boehringer Mannheim) to eliminate any trace of DNA. Reverse transcription-PCR (RT-PCR) was performed with approximately 10 ng of RNA per reaction with the Titan One RT-PCR kit (Boehringer Mannheim) and a Progene thermal cycler, from Techne (Cambridge) Ltd. The RT stage was performed at 55°C for 30 min. After 1 min at 94°C to denature the DNA, amplification was carried out as follows: (i) 10 cycles of 30 s at 94°C, 30 s at 55°C, and 1 min at 72°C; (ii) 15 cycles of 30 s at 94°C, 30 s at 55°C, and 1 min at 72°C, the elongation time of each cycle being increased by 5 s; and (iii) a final elongation step of 7 min at 72°C. The primers used to amplify a 606-bp subfragment of the ARO9 gene situated 84 bp downstream from the start codon were 5'-CCGTTCAGTAGTCAGGTCGCT-3' and 5'-GGATCCCCCAGTGAGCTACCCATT-3'. For the ARO10 (YDR380w) gene, the 745-bp amplicon is located 60 bp downstream from the start codon, and the primers used were 5'-CCATACCTATCAGTCAGTACGTCTCC-3' and 5'-CGATCAGCAACAATTCCGTTTGGTG-3'. The ACT1 amplicon was produced with the primers 5'-GACTCCTACGTTGGTGATGA-3' and 5'-CTGGAGGAGCAATGATCTTG-3'.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Detection of a UAS element required for induction of the ARO9 gene. To roughly locate the cis-acting sequences involved in ARO9 induction, we examined the effects of a series of 5' deletions produced in the ARO9 promoter region (Fig. 2). The starting plasmid was YCpARO9-lacZ, containing 1,231 bp of upstream region and the first 11 codons of the ARO9 gene fused to a promoterless lacZ gene (34). Our previous work had shown that this fusion product responds to the same regulatory signals as the chromosomal ARO9 gene. In particular, it is very weakly expressed in the absence of inducer, induced by aromatic amino acids, and subject to the ammonia effect (34). The effect on induction of removing different promoter sequences was monitored by measuring the -galactosidase activity in plasmid-bearing yeast strain 23346c (ura3) grown on minimal urea medium in the presence or absence of inducer. Deletions extending from 1231 through 255 had no major effect on gene induction. Further removal of sequences up to 169 resulted in a two- to threefold decrease in the level of induction, and no further change was observed with deletions extending to 153. This segment of the promoter thus seems to contribute modestly to activation. It contains several sequences of potential interest: the sequence 234-TTGCCGCCGA-225 almost perfectly matches (9 of 10 nucleotides) the URS1 element (44), the binding site of Car80p/Ume6p, a key regulator of nitrogen metabolism and early meiosis gene expression (74). The imperfect palindrome 231-CCGCCGACGGCCG-219 overlaps with this element. Further deletion of sequences up to 145 and beyond caused almost complete loss of tryptophan-induced expression, indicating that an element essential to activation is present in the promoter near position 145. This region contains, from 168 to 133, a remarkable sequence consisting of four CCG repeats regularly spaced by 7 bp (Fig. 1A and B). An internal deletion removing this 36-bp sequence was produced in plasmid pII472. The deletion abolished induction of the gene (Fig. 2), showing that this 36-bp promoter fragment is necessary for ARO9 induction.

View larger version (21K): [in this window] [in a new window]   FIG. 2.   Effects of 5'-nested deletions on induced ARO9-lacZ expression. Strain 23346c (ura3) transformed with a low-copy-number plasmid carrying the different promoter deletion constructs was grown on minimal urea medium with (+ Trp) or without ( Trp) tryptophan at 500 µg/ml. -Galactosidase activities, expressed in nanomoles per minute per milligram of protein, are averages of two independent experiments. -Galactosidase activity variations were 15%. The black box indicates the position of UASaro. The grey box indicates the position of the sequence similar to URS1. The hatched box represents the position of the SPL2 (YHR136c) gene (20). Vertical lines represent the positions of two TATA consensus sequences (TATAAA).

View larger version (25K): [in this window] [in a new window]   FIG. 3.   UASaro renders the UASaro-CYC1-lacZ gene inducible by aromatic amino acids. Strain 23346c (ura3) transformed with a high-copy-number plasmid carrying the wild-type form (PAL1-15) or a mutated form of UASaro was grown on minimal urea medium with (+ Trp) or without tryptophan at 500 µg/ml. -Galactosidase activities, expressed in nanomoles per minute per milligram of protein, are averages of two independent experiments. -Galactosidase activity variations were 15%. The underlined lowercase "t" in the PAL1-46 plasmid indicates substitution of a thymine for a cytosine.

Isolation of mutants impaired in ARO9 induction. Cells of strain 30332d (ura3) transformed by multicopy plasmid PAL1-15 bearing the UAS_aro-CYC1-lacZ fusion gene form blue colonies during growth on urea plus tryptophan medium supplemented with X-Gal. To isolate mutants affected in ARO9 induction through the UAS_aro element, we mutagenized transformed 30332d cells and analyzed clones forming white or pale blue colonies. Among 17 analyzed mutants, strains IHI602, IHI606, and IHI614 displayed no induction of either -galactosidase or tryptophan-phenylpyruvate aminotransferase activity. Strain IHI611 was only partially deficient (Table 2). The same induction deficiencies were observed in mutant cells cured of the plasmid and back-transformed with PAL1-15 (data not shown). The mutants were defective in induction of both the chromosomal ARO9 gene and the multiple plasmid-borne copies of the UAS_aro-CYC1-lacZ gene. The mutations most likely affect trans-acting elements involved in UAS_aro-mediated induction.

ARO9 induction is impaired in mutants lacking the Gap1p and Wap1p permeases. Mutant strain IHI611 failed to grow on minimal medium containing tryptophan, phenylalanine, tyrosine, or citrulline as the sole nitrogen source and grew at a very reduced rate on isoleucine. Genetic analyses indicated that two mutations were responsible for this growth phenotype (Table 3). One of them affected the GAP1 locus and was called gap1-92. The second mutation did not by itself produce a particular phenotype, but in double mutants it markedly worsened the growth deficiencies caused by the gap1-92 mutation (Table 3). To identify the altered genetic locus, we searched a wild-type genomic library for fragments complementing the growth deficiencies of the double mutant strain 32102c (see Materials and Methods). Subsequent subcloning experiments indicated that the second mutation affected the YCL025c gene on chromosome III (56). On the basis of sequence similarities, this gene had been postulated to encode an amino acid permease (2, 56). We temporarily named the second mutation ycl025-1 and constructed a ycl025-1 deletion mutant strain. Growth phenotype analysis, complementation tests, and analysis of 21 tetrads derived from the cross between strain IHI726 (gap1-92 ycl025-1 ura3) and strain 32102c (gap1-92 ycl025-1 ura3) showed that the recessive mutations ycl025-1 and ycl025-1 are allelic.

gap1 ycl025-1

gap1 ycl025-1

The ARO80 gene is essential to induced expression of ARO9. Mutant strains IHI602, IHI606, and IHI614 failed to grow on minimal tryptophan medium; their growth rate was reduced on minimal phenylalanine or tyrosine medium and unaffected on other sources of nitrogen. Genetic analyses showed that all three mutations segregate 2:2; are recessive; complement the aro8, aro9, gap1, and wap1 mutations; and belong to the same complementation class, which we have named ARO80. When an aro80 mutation was associated with an aro8 mutation affecting aromatic aminotransferase I (77), the aro80 aro8 double mutant cells grew very slowly on ammonia medium unsupplemented with phenylalanine or tyrosine. This confirmed the absence of inducible aromatic aminotransferase II activity in the double mutants, our previous studies having shown that the simultaneous absence of aromatic aminotransferases I and II results in phenylalanine and tyrosine auxotrophy (77).

aro81-1 aro80-2 ura3

ydr421w-1

ydr421w-1

ydr421w-1

aro80-1

View larger version (16K): [in this window] [in a new window]   FIG. 4.   (A) Map of chromosome IV region around the ARO80 gene. (B) Schematic drawing of the DNA fragments that complement the auxotrophy of the aro8 aro80-2 double mutation. (C) Schematic diagram of predicted Aro80p domains: putative bipartite nuclear targeting sequence (), C6 zinc finger domain (), coiled-coil region (), middle homology region (), and acidic domain (). (D) Alignment of the C6 zinc finger sequence of the ARO80 gene and the consensus sequence of the domain (64).

View larger version (83K): [in this window] [in a new window]   FIG. 5.   RT-PCR analysis of ARO9 and ARO10 transcripts. Strains 23346c (ura3), 30701a (aro80 ura3), and 30312d (aro9 ura3) were grown on minimal medium containing the following nitrogen sources: urea (U), urea plus tryptophan (U+T), ammonia (A), or ammonia plus tryptophan (A+T). Total RNA was used to perform the analysis (see Materials and Methods). wt, wild type.

Aro80p is a member of the Zn₂Cys₆ transcription factor family. The ARO80 (YDR421w) open reading frame encodes a 950-residue polypeptide with a calculated molecular mass of 118 kDa. Its codon bias index is 0.022 (7), and its codon adaptation index is 0.12 (67), both values being indicative of a low expression level, typical of yeast transcriptional activators. Two poor matches to the TATA consensus sequence lie at positions 49 and 72 from the putative initiation codon. This upstream region also contains a cluster of three 5'-GATA-3' sequences from positions 137 to 114, raising the possibility that GATA factors control ARO80 expression. The deduced amino acid sequence of Aro80p contains several motifs (Fig. 4C and D) that characterize the Gal4p class of transcription factors, as already analyzed by Schjerling and Holmberg (64) in their study of the 56 Gal4p family members present in the entire yeast genome: the most salient feature is an N-terminal segment highly similar to the consensus sequence of the Zn₂Cys₆-type zinc finger motif (Fig. 4D), an established DNA-binding domain (40). The Aro8p motif has all the features found in the other cloverleaf-like Zn₂Cys₆ binuclear clusters (57), but the region of variable length linking the two cysteine-rich substructures comprises 12 amino acids instead of 6 to 9 as commonly found in other proteins of the family; it shares this property with the putative yeast transcription factor Yml076p of unknown function; several clusters of acidic residues are present in the predicted Aro80p. The cluster with the highest negative charge (9) is located at the C-terminal end spanning positions 884 to 909. Similar C-terminal acidic regions have been shown to mediate protein-protein interactions and to contribute to the ability of the protein to activate transcription (46). In Aro80p, this C-terminal acidic segment is followed by an asparagine-rich region; several coiled-coil structures with a potential to form heptad repeats of an amphipathic -helix are predicted by the Pepcoil algorithm (45). It is common for this family of factors to act as homodimers, and crystal structures have revealed that coiled-coil structures mediate formation of Gal4 and Ppr1 homodimers (50, 51). Other noteworthy structural features of Aro80p have not been reported previously: (i) a recognizable part of the middle homology region (64) spans positions 370 to 412; (ii) a putative bipartite nuclear targeting sequence (61) spans amino acids 5 to 20; (iii) an ATP-GTP binding site spans residues 647 to 654; and (iv) a putative PEST region extends from residue 836 to residue 854.

The ARO80 gene is also essential to induced expression of the ARO10 (YDR380w) gene encoding a putative indole-3-pyruvate decarboxylase. On media containing tryptophan, phenylalanine, or tyrosine as the sole nitrogen source, growth of the aro80 mutant was clearly affected more strongly than that of the aro9 mutant (Fig. 6). This indicates that Aro80p activates other genes besides ARO9 in the aromatic amino acid utilization pathway. Genes involved in aromatic amino acid transport were obvious candidates for this regulation. Yet the tryptophan, phenylalanine, and tyrosine uptake rates were unaffected by the aro80 mutation, and the ARO80 gene proved not to be involved in induction of the new Wap1p permease (data not shown). We searched the genome for sequences similar to UAS_aro. The pattern 5'-CCG(7X)CCG(7X)CCG(7X)CCG-3' was found, spanning bp 342 to 310 upstream from the open reading frame YDR380w. The deduced Ydr380p sequence is highly similar to those of yeast pyruvate decarboxylases Pdc6p, Pdc5p, and Pdc1p and to the recently described -ketoisocaproate decarboxylase Kid1p (15). It is also highly similar to the three bacterial indolepyruvate decarboxylase sequences present in databases. In fact, together with pyruvate decarboxylases and Kid1p, Ydr380p is the yeast protein most similar to them, as established by sequence comparisons (data not shown). Thus, Ydr380p may be the yeast indole-3-pyruvate decarboxylase catalyzing conversion of indole-3-pyruvic acid to indole-3-acetaldehyde, the second reaction involved in tryptophan degradation to tryptophol. In line with this hypothesis, we found the YDR380w and ARO9 genes to be regulated in a very similar manner. RT-PCR transcript analysis (Fig. 5) showed weak expression of YDR380w on both urea and ammonia medium. On both media, YDR380w transcription was induced upon cell exposure to tryptophan, the activation level being markedly higher on urea medium. In both cases, induction of transcription was abolished in aro80 mutant cells, showing that ARO80 is also essential to YDR380w induction by tryptophan. These facts highlight a second target of Aro80p and integrate the YDR380w gene into tryptophan metabolism. We have thus given YDR380w the name ARO10. Further characterization of this gene is under way.

View larger version (101K): [in this window] [in a new window]   FIG. 6.   Growth phenotypes of strains 23346c (ura3), 30312d (aro9 ura3), and 30701a (aro80 ura3) on minimal medium containing ammonia, phenylalanine, tyrosine, or tryptophan as the nitrogen source. All media were supplemented by 50 µg of uracil per ml. wt, wild type.

Nitrogen regulation of ARO9 expression. Like many other nitrogen catabolic genes, ARO9 is subject to nitrogen source regulation (see the introduction). A systematic study was undertaken to discover the mechanisms underlying this control. Wild-type and appropriate mutant strains transformed with the ARO9-lacZ plasmid were grown under conditions favoring repression (ammonia medium) or derepression (urea medium) and in the presence or absence of the inducer tryptophan. The transformed strains each carried one or several mutations affecting factors involved in either NCR or NCI. -Galactosidase activity was taken as a measure of ARO9 expression.

(i) Role of the NCR general negative factors. In the absence of inducer, wild-type cells grown on either ammonia or urea medium displayed very low, similar levels of ARO9 expression; hence, no ARO9 derepression was observed on urea medium (Table 6, row 1). In the absence of inducer, likewise, we observed no increased expression of ARO9 in cells harboring, separately or in combination, the mutations ure2, uga43, and gzf3, disabling factors that mediate nitrogen repression (Table 6, rows 2 to 8). This contrasts with the behavior of the inducible genes CAR1, GDH2, DUR1, DUR2, and DAL1 (47, 80); PUT1, PUT2, and UGA1 (81); and UGA4 (4), whose basal expression in inducer-free media is subject to nitrogen repression. Yet despite the apparent insensitivity of ARO9 to nitrogen repression, once wild-type cells were exposed to the inducer, much higher levels of ARO9 transcription were observed on urea medium than on ammonia medium (Table 6, row 1). This ammonia effect can hardly be attributed to nitrogen repression exerted by Ure2p, by GATA factor Gzf3p or Uga43p, or by a combination of these, since expression increased very little in NH₄⁺-grown cells lacking these factors (Table 6, rows 2 to 8). Furthermore, the threefold increase observed in cells carrying the ure2 mutation, alone or associated with the gzf3 and uga43 mutations (Table 6, rows 2, 5, and 7 and 8), was lost in ure2 aro80 cells (Table 6, row 9), showing that it is dependent on induction.

(ii) Role of the positive factors Gln3p and Nil1p. Nitrogen-regulated genes are typically under the positive control of the GATA family factors Gln3p and Nil1p (47, 52). On tryptophan-supplemented urea medium, ARO9 transcription was unaffected in cells lacking one of these factors (Table 6, rows 10 and 11). Surprisingly, the activity in the gln3 mutant is higher than that in the wild-type strain after growth in minimal ammonia plus tryptophan (Table 6, rows 10 and 1; Table 7). This unexpected result has also been observed in the case of the arginine induction of the CAR1 gene encoding arginase (18). ARO9 expression was greatly reduced upon elimination of both activators (Table 6, row 12), indicating that they both can contribute to the amplification of induction observed under conditions of nitrogen derepression. Yet under conditions of nitrogen repression, elimination of both activators had no depressing effect on ARO9 induction (Table 6, row 12), in contrast to what happens with other nitrogen-regulated genes (9, 71, 73). Furthermore, Gln3p and Nil1p activate expression of nitrogen-sensitive genes via 5'-GATA-3' UASs. The ARO9 promoter contains several dispersed GATA motifs (Fig. 1A) but no GATA element similar to either UAS_NTR (58, 59), UAS_N (55), or UAS_GATA (4). Moreover, we failed to obtain evidence that Gln3p and Nil1p exert their stimulatory effect at a site different from UAS_aro: none of our various ARO9 promoter constructs enabled us to distinguish Gln3p and Nil1p target elements from the induction element UAS_aro. First, we introduced the constructs obtained by exonuclease III treatment (Fig. 2) into wild-type and gln3 nil1 strains and compared their expression in both genetic backgrounds. In the set of deletion mutants, loss of induction closely paralleled loss of the amplifying action of the two GATA activators (data not shown). Removal of the sequence between 255 and 169 resulted in a two- to threefold decrease in the level of induction in the double gln3 nil1 mutant from in the wild-type strain. Next, we confirmed that the amplifying effect of Gln3p and Nil1p requires the integrity of the UAS_aro element, as judged from the effect of deleting the element in the ARO9-lacZ fusion construct (Table 7). Finally, we observed that UAS_aro alone suffices to drive a much higher level of tryptophan-induced transcription on urea medium than on ammonia and that this amplification is no longer observed in gln3 nil1 double mutant cells (Table 7). Hence, Gln3p and Nil1p appear to amplify induction of ARO9 mainly through UAS_aro. The low constitutive activity measured in the gln3 nil1 double mutant (Table 6, row 12) is possibly due to a particular phenomenon connected to its very low growth in minimum medium.

(iii) Role of inducer exclusion. The above results suggest that NH₄⁺ and the GATA factors Gln3p and Nil1p may affect ARO9 induction indirectly, by inducer exclusion. Inducer uptake as a limiting step in ARO9 induction is indeed consistent with the observation that ARO9 expression is severely reduced in gap1 wap1 double mutants which take up tryptophan at a very reduced rate (Table 4). To test this hypothesis, we monitored induction of ARO9 in gap1^pgr and npi1 mutants grown on ammonia medium. These strains are defective in NCI and incorporate tryptophan faster than do wild-type cells on ammonia medium (see the introduction). The gap1^pgr mutation protects only the Gap1p permease from inactivation, whereas the npi1 mutation prevents inactivation of other NCI-sensitive permeases as well (23). On tryptophan-supplemented ammonia medium, gap1^pgr mutant cells displayed four-times-higher ARO9 expression levels than wild-type cells (Table 6, row 14). In npi1 cells grown on the same medium, ARO9 expression was increased more than 10-fold, reaching 80% of the level found in wild-type cells grown on urea plus tryptophan (Table 6, row 15). ARO9 can thus be expressed to near-maximum levels under conditions of nitrogen repression in mutants released from NCI. Simultaneous release from NCR and NCI in a ure2 npi1 double mutant caused but a slight further increase in expression compared to the effect of the npi1 mutation alone (Table 6, row 16). These results strongly indicate that the ammonium effect on ARO9 expression is produced mainly by inducer exclusion through permease inactivation. The Gap1p permease is neither the only nor the principal tryptophan permease involved in exclusion, since (i) the ARO9 induction level was only slightly affected in gap1 cells compared to that in wild-type cells (Table 6, row 13) and (ii) the npi1 gap1 strain (Table 6, row 17) attains about 60% of the induction level reached by the npi1 single mutant. The experiments for Table 6 were also carried out with phenylalanine or tyrosine as an alternative inducer, yielding very similar results (data not shown).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The ARO9 gene encodes aromatic aminotransferase II, the enzyme catalyzing the first reaction of aromatic amino acid catabolism (34, 41, 77). Transcription of the ARO9 gene is induced by aromatic amino acids and subject to nitrogen regulation. In this work, we have characterized cis- and trans-acting elements mediating this induction. We have also shown that nitrogen source regulation of ARO9 expression is due mainly to inducer exclusion, in contrast to the situation observed with many genes involved in secondary nitrogen source utilization. We further show, in this report, that the two main uptake systems for utilization of aromatic amino acids are Gap1p, the general amino acid permease, and Wap1p, a newly discovered inducible amino acid permease with wide substrate specificity. Finally, we present evidence that transcription of the ARO10 gene, which very likely encodes the second enzyme of the aromatic catabolic pathway, indolepyruvate decarboxylase, is subject to the same new induction mechanism.

Induction-specific regulatory elements. Induction of ARO9 by aromatic amino acids requires UAS_aro, a 36-bp element, and the trans-acting factor Aro80p. UAS_aro and Aro80p behave functionally like the pathway-specific elements ArgRIIp (53), Put3p (48), Uga3p (1), and Cha4p (31) and their cognate UASs, which control induction of the arginine, proline, GABA, serine, and threonine catabolic pathways in yeast, respectively. Like ArgRIIp, Put3p, Uga3p, and Cha4p, Aro80p is a member of the Zn₂Cys₆ family of transcriptional activators. Its deduced sequence displays all the features that characterize this protein family. Aro80p is required for specific induction of ARO9 expression, and its function depends on sequences present in the UAS_aro element. This 36-bp element, on the other hand, is necessary and sufficient to mediate transcriptional activation of a reporter gene in response to aromatic amino acids. UAS_aro could thus be the DNA element to which Aro80p binds. Its structural features support this hypothesis: binding sites for Zn₂Cys₆ proteins often comprise two symmetrically placed CGG elements, separated by a characteristic spacing (64). UAS_aro possesses one CGG triplet which could form an imperfect palindrome with any of several CCG triplets in the element (Fig. 1B). The CCGCGG core of one such palindrome, for example, is identical to the core of the Pdr3p binding site (39). Yet not all zinc cluster proteins bind to sites with inverted CGG repeats. Hap1p binding sites, for instance, are composed of two direct CGG repeats, and their spacer displays no symmetry (82). The most obvious feature of the UAS_aro sequence is the pattern of four regularly spaced direct CCG repeats. A similar pattern was found in the ARO10 (YDR380w) promoter. When the putative UAS of ARO10 is aligned with UAS_aro, the consensus sequence of the motif is extended to T(T/A)(G/A)CCG followed by four variable nucleotides (Fig. 1C). ARO10 is also inducible by tryptophan, and its transcriptional activation is Aro80p dependent. The ARO9 and ARO10 genes are thus coregulated by aromatic amino acids.

Nitrogen regulation of ARO9 is exerted mainly through inducer exclusion. ARO9 is less actively transcribed in ammonia-grown than in urea-grown cells, indicating that the gene is subject to nitrogen source regulation. Nitrogen catabolic genes are commonly regulated by nitrogen repression, i.e., by the competitive action at upstream 5'-GATA-3' elements of positive and negative GATA family factors and by the inhibition exerted by Ure2p on Gln3p (47, 52). Our results strongly suggest that ARO9 is an exception. Nitrogen regulation of ARO9 transcription would appear to be achieved mainly or even exclusively by inducer exclusion, via regulation of permease synthesis and activity. This conclusion is based on the fact that mutations relieving nitrogen repression only slightly affect ARO9 expression, compared with mutations relieving NCI of permeases. After growth on ammonia medium containing tryptophan (or phenylalanine or tyrosine), npi1 cells display a high level of ARO9 transcription, reaching 80% of the level displayed by wild-type cells grown on tryptophan-supplemented urea medium. Under growth conditions promoting nitrogen repression, ARO9 expression can thus reach near-maximum levels provided that permease inactivation is prevented. The CAR2 gene encoding ornithine aminotransferase in the arginine catabolic pathway (14) is likewise subject to an ammonia effect but apparently insensitive to nitrogen repression. In this case, it was possible to correlate closely the amount of enzyme synthesized with the cellular pool of arginine, the inducer (14). The hypothesis that the nitrogen source regulates ARO9 expression principally by modulating inducer availability is consistent with other observations presented here, notably the fact that ARO9 expression is induction dependent under all examined conditions. The very low level of ARO9 expression in cells grown on inducer-free medium is in line with the very small cellular pools of aromatic amino acids (13, 19, 41). The negative effect of the gln3 nil1 double mutation can also be attributed to reduced inducer uptake. We have shown that Gap1p and Wap1p are the main transporters of aromatic amino acids and that gap1 wap1 cells display severely reduced levels of ARO9 expression on tryptophan-supplemented urea medium. It is known that the GAP1 transcription level is very low in gln3 nil1 cells (71, 73). We show elsewhere (33) that WAP1 responds in a very similar manner, a result also published recently by another group (65). Thus, the marked negative effect of the double gln3 nil1 mutation is most likely a consequence of poor permease synthesis. The inducer exclusion hypothesis also seems to best explain why the gln3 nil1 double mutation does not affect induced ARO9 expression under conditions of nitrogen repression. Analysis of the npi1 mutant has indeed shown that permease inactivation, and not permease synthesis, limits inducer availability under these growth conditions.