Copyright 6 1997 by the Genetics Society of America
The Isolation and Characterization of Saccharomyces cer&ae Mutants That Constitutively Express Purine Biosynthetic Genes
Maria L. Guetsova," Karine Lecoqt and Bertrand Daignan-Fornier**+ *Institut de GCnCtique et Microbiologie, CNRS URA1354, Universitk Paris Sud, 91405 Orsay Cedex, France and +Institut de Biochimie et GCnCtique Cellulaires, CNRS UPR 9026, 33077 Bordeaux Cedex, France Manuscript received March 13, 1997 Accepted for publication June 16, 1997
ABSTRACT In response toan external sourceof adenine, yeast cells repress the expressionof purine biosynthesis pathway genes. To identify necessary components of this signalling mechanism,we have isolated mutants that are constitutively active for expression. These mutants were namedbra (for hypass of fepression by -adenine). BRA 7 is allelic toFCYZ, the gene encoding the purine cytosine permease BRA9 and is ADEl2, the gene encoding adenylosuccinate synthetase.BRA6 and BRA1 are new genes encoding, respectively, hypoxanthine guanine phosphoribosyl transferase and adenylosuccinate lyase. These results indicate that uptake and salvage of adenine are important steps in regulating expression of purine biosynthetic genes. Wehave also shown that two other salvage enzymes, adenine phosphoribosyl transferase and adenine deaminase, areinvolved in activating the pathway. Finally, using mutant strains affectedin AMP kinase or ribonucleotide reductase activities, we have shown that AMP needs to be phosphorylated to ADP to exert its regulatory role while reduction of ADP into dADP by ribonucleotide reductaseis not required for adenine repression. Together these data suggest that ADP or a derivative of ADP is the effector molecule in the signal transduction pathway.
ICROORGANISMS alter their metabolism in re- bind to the promoters of purine biosynthetic (AD@ M sponse to the presence of metabolic precursors genes. Although Bas2p is involvedin multiple metabolic in the environment. This adaptation requires theability pathways (BRAUSet al. 1989; VOGELet al. 1989; BRAZAS to sense nutrient levels and then transduce a signal to and STILLMAN1993), Baslp appearsspecific for purine redirect the synthesis of metabolic enzymes. We are and histidine biosynthesis genes (ARNDT et al. 1987, interested in the signalling cascade that leads to repres- DAIGNAN-FORNIERand FINK1992, SPRINGERet al. 1996). sion by adenine in Saccharomyces cermisiae. Because all the genes known to be activated by Baslp Coordinate repression of de novo purine synthesis are also repressed by adenine, it is an appealinghypoth- genes has been reported in bacteria and bakers yeast esis that Baslphas a directrole in regulating the purine (MOMOSE et al. 1966; NEUHARDand NYGAARD 1987; biosynthetic pathway. There are several possibilities for DAIGNAN-FORNIERand FINK 1992). Interestingly, this re- how the availabilityof externaladenine might be pression is achieved by very different processes in Esche- sensed. The purine bases themselves might be the sig- richia coli and Bacillus subtilis. In both bacteria the regu- nal. Alternatively, purine availability could affect tran- lation is mediated by a specific repressor named purR. scription indirectly througha signalling cascade. Fi- Although in E. coli binding of the repressor to its 16 nally, Baslp and/or Bas2p might be directly regulated bp target site depends on thepresence of hypoxanthine by this signal. or guanine (ROLFESand ZALKIN 1990), inB. subtilis the As a first step toward answering these questions we repressor binding site is 110 bp long and its interaction have isolated mutants that constitutively express purine with the regulatory protein is inhibited by 5-phosphori- biosynthetic genes and aretherefore candidate compo- bosyl 1-pyrophosphate (PRPP) (WENGet al. 1995). nents of the signalling cascade that respondsto environ- In yeast adenine repression is less well understood. mental adenine. Herewe report theisolation and char- We have previously shown (DAIGNAN-FORNIERand FINK acterization of these mutations and theircognate genes. 1992) that expression of several genes of the purine biosynthetic pathway is repressed in the presence of MATERIALS AND METHODS adenine in the growth medium. Derepression requires the transcription factors Baslp and Bas2p. Both factors Yeast strains and media: Yeast strains are listed in Table1. Yeastmedia were preparedaccording to SHERMANet al. Corresponding autho?.: Bertrand Daignan-Fornier, Institut de Biochi- (1986). Adenine, guanine and hypoxanthine were used at a mie et Gnetique Cellulaires, 1, rue Camille Saint-SaEns, 33077 Bor- final concentrationof 0.15 mM. The XGal synthetic medium deaux Cedex, France. (DANGet al. 1994) and the 5-fluoro-orotic acid (5-FOA) me- E-mail: b.daignan-f0rnieraibgc.u-bordeaux2.fr dium (BOEKEet al. 1984) were prepared using the methods
Genetics 147: 383-397 (October, 1997) 384 M. L. Guetsova, K. Lecoq and B. Daignan-Fornier
TABLE 1 Yeast strains used in this study
StrainSource name Genotype
PLY1 2 1 MATa his3-A200 leu2-3,112 lysZ-A201 ura3-52 P. LUNJDALL PLY122 MATa leu2-3,112 lys2-A201 ura3-52 P. LUNJDALL L3861 MATa ah2 leu2-3,112 lys2-A201 ura3-52 G. FINK L3862 MATa ade2 his?-A200 lysZ-A201 ura3-52 G. FINK NC247-1B MATa ura3A fq2A M. R. CHEVALLIER w109-9C MATa ade2 trpl ura? his3 hptl-27 WOODS R. AH215 MA Ta leu 2 his3 M. KONRAD AH215 adkl MATa leu2 his? adkl::HZS3 M. KONRAD Y203 MATa leu2-3,112 lys2 ura3-A100 ade2-1 his3 trpl rnr3::RNm-URA3-TRPI S. ELLEDGE -21 MATa ku2-3,112 lys2 ura3-A100 ade2-1 his3 trpl ctr6-68 rnr3::RNW-URA3-TRPl S. ELLEDGE + pZZl3 (HZS3) Y399 MATa ade2 leu2-3,112 lysZ-A201 his3-A200 ura3-52 bra91 This work Y531 MATa leu2-3,112 lys2-A201 ura3-52 bra62 This work Y508 MATa leu2-3,112 lys2-A201 ura3-52 hptl::URA3 This work Y511 MATa leu2-3,112 lys2-A201 ura3-52 aptl::URA3 This work Y520 MATa leu2-3,112 lysZ-A201 ura3-52 aahl::URA? This work Y548 MATa leu2-3,112 lys2-A201 ura3-52 aptl::URA3 This work Y549 h4ATa leu2-3,112 lysZ-A201 ura3-52 aptl::URA? This work Y550 MATa leu2-3,112 lys2-A201 ura3-52 apt1::URM aahl::URA3 This work Y55 1 MA79 leu2-3112 lys2-A201 ura3-52 aptl::URA3aahl::URA3 This work Y552 MATa leu2-3,112 lys2-A201 ura3-52 aptl::URA3aahl::URA3 This work Y608 MA Ta ade2 leu2-3,112 lysZ-A201 ura3-52 ADEl2::ADEl2-LEUZ This work Y610 MATa leu2-3,112 lys2-A 201 his3-A200 ura3-52 ADEl?::DE13-LEU2 This work previously described. 5-fluorocytosine (5FC) was added at a were performed, each assay was done on three independent final concentration of 0.1 mM to sc medium containing 0.03 transformants. Variation between assays in each experiment mM uracil. Base analogues 8-azaadenine (8AA) and 8-azagua- was <20%. Variations between experiments are due to the nine (8AG) were added to themedia at a final concentration use of different spectrophotometers. of 0.2 mg/ml. Mutagenesis: EMS mutagenesis was doneon strains Plasmids: P78, the plasmid carrying the ADE5,7-URA3 chi- PLY121 and PLY122 transformed with the P78 plasmid. A 1-ml mera, was constructed by fusing the ADE5,7 promoter and liquid culture of each straingrown overnight was centrifuged, the first 28 codons of ADE5,7 to the coding sequenceof URA3 washed twice with water and resuspended in 1 ml of EMS (AIANI and KLECKNER 1987).For this purpose, theP4 plasmid buffer (0.1 M KPO,, pH 8). To each tube, 30 p1 of EMS was carrying an ADE5,7-lacZ fusion (DAIGNAN-FORNIERand FINK added and vigorously mixed. After 1 h incubation at room 1992) in the vector YEp367R (MYERSet al. 1986) was digested temperature, the cells were pelleted and washed three times with BamHI and BglII and ligated to the BamHI-BamHI frag- with 5% sodium thiosulfate. Finally cells were diluted in water ment carrying the 'URA3 gene from pNKY48 (ALAN1 and and plated on WD medium to estimate the loss of viability KLECKNER 1987). due to the mutagenesis. Comparison between treated and LucZ fusions and pGal assays: The LacZfusions used in this untreated cells allowed us to estimate at 60% the rate of sur- study were constructed as follows.P2 and P115 have been vival after EMS treatment. previously described (DAIGNAN-FORNIERand FINK1992). P2 Integration of LEU2 at theADEI2 locus: An EcoRI-Nszl frag- is a plasmid carrying an ADE2-lac2 fusion in a2p LEU2 vector ment carrying the ADE12 gene from plasmid P103 was cloned YEp368R (MYERS et al. 1986). P115 is a plasmid carrying an into an integrative LEU2vector namedYIpLacl28 (GIETZand ADEl-lacZ fusion in a 2p URA3 vector YEp356R (MYERSet al. SUGINO 1988) linearized with EcoRI and PstI. The resulting 1986). Another DEl-lacZ fusion was constructedin the plasmid, named P659, was linearized at the PstI site in the course of this work using a two steps procedure. First a Nszl- ADEl2coding region and used to transform the L3861 strain. SpeI DNA fragment carrying the ADEl gene was cloned at the Tandem integration of the plasmid at the ADEl2 locus was PstI-XbaI sites of the pRS315 vector (SIKORSKIand HIETER verified by Southern blot on genomic DNA extracted from 1989) generating plasmid P68. Second, a 1600-bp KpnI-XbuI tranformants and cut with Nstl. One of these transformants, DNA fragment from P68 starting 900 bp upstream from the named Y608, was used for linkage analysis. ATG initiation codon of the ADEl gene was cloned in YEp367 Integration of LEU2 at the ADE13 locus: An EcoRI-BglII (MYERSet al. 1986). fragment carrying the ADE13gene was cloned into YIpLac128 PGal assays were performed as described by RUBYet al. (integrative LEU2, GIETZand SUCINO1988) linearized with (1984), with the exception of Table 8, assays that were per- EcoRI and BamHI. The resulting plasmid, named P661, was formed by the method of KIPPERT (1995). In all cases, PGal linearized at the BamHI site in the DE13 coding region and units are defined as follows: used to transform the PLY121 strain. Tandem integration of the plasmid at the ADE13 locus was verified by Southern blot ODq2,, X 1000/ODti,,oX t (min) X vol (ml). on genomic DNA extracted from tranformants and cut with BglII. One of these transformants, named Y610, was used for In each experiment, at least two independent /?Gal assays linkage analysis. M utants constitutiveMutants for ADE genes 385
Disruption of the HPTl gene: Disruption of the HPTl gene Isolation of bra mutants We have used a ADE5,7- was performed asfollows. An EcoRV-EcoRV DNA fragment URA3 translational fusion to select for mutations that carrying the YDR399w open reading frame (OW) from chro- affect repression of ADE genes in the presence of ade- mosome IV was inserted in pUCl8 (YANISCH-PERRONet al. 1985) linearized at the unique SmaI site. The resulting plas- nine. A fusion between the ADE5,7 promoter and the mid, namedP385, was deleted for its internal BgllI-BglII DNA URA3 coding sequence (ALAN1 and KLECKNER 1987) fragment thus removing the promoter region and the 5' half was constructed on a 2p yeast expression plasmid. Ex- of the coding region (this deletion does not affect the up- pression of URA3 in this construct is under the control stream ORF named YDR398w). The deleted fragment was of ADE5, 7regulatory elements,i.e., activated in thepres- replaced by a BamHI-BamHI fragment carryingthe URA3 gene from plasmid YDpURA3 (BERBENet al. 1991). The resulting ence of Baslp and Bas2p and repressed by adenine. A plasmid, named P399, was digested with Nszl and usedto ura3 strain containing this plasmid, grows very slowly transformthe PLY122 strain.Uraf transformants resulting on medium lacking uracil and containing adenine (re- from one-step gene disruption at the HPTl locus were ob- pression conditions). In the absence of adenine the tained and verified by Southern blot (data not shown). growth is faster but still slower than in the presence of Disruption of the AAHI gene: For disruption of the AAHl gene, a 1300-bp EcoRI-SphI fragment carrying the AAHl gene uracil, indicating that the OMP decarboxylase activity from pCG22 (DEELEY1992) was cloned in YIpLac128 (GIETZ provided by the ADE5,7-URA3fusion isnot optimal even and SUCINO1988). This plasmid, named P389, was deleted under derepression conditions. The growth rate differ- for its PstI-XbaI 116bp fragment that is located 65 bp down- ence between repression and derepression conditions stream of the ATG initiation codon of the AAHl OW, and this fragment was replaced by a Nszl-SpeI fragment carrying in the absence of uracil makes it possible to identify the URA3 gene from YEp24. The resulting plasmid, named constitutive mutants as rapidly growing colonies above P456, was digested with both Ssp1 and SphI and used to trans- the background of slowly growing colonies. form the PLY122 strain. Disruption was verified by Southern The ADE5,7-m3 fusion carried on a plasmid was (data not shown). introduced ina and CY isogenic wild-type strains, PLY121 Disruption of the APT1 gene: A 1450-bp EcoRV-KpnI frag- ment carryingthe APTl gene frompCG42 (DEELEY1992) was and PLYl22. These transformed strains were mutagen- cloned in pBluescript KS digested with EcoRV and KpnI. The ized with ethyl methanesulfonate to 60% survival and resulting plasmid, named P393, was deleted for its internal plated at low density (-2000 cfu/plate) on SC -leu HindIII fragment containing mostof the APTl ORF. The -ura + ade. After 4-5 days, 84 mutants (42 in each deleted fragmentwas replaced by the 1.1-kb HindIII fragment mating type) that grew faster than the wild-type strain carryingthe URA3 gene. The plasmidcarrying the APTI::URA3 construct (P397) was digested with EcoRV and under repression conditions were selected for further KpnI and used to transform the yeast strain PLYl22. Disrup analysis. Mutants named 101-142 were isolated from tion was verified by Southern (data not shown). PLY121 and mutants 201-242 from PLYl22. Adenylosuccinate lyase enzymatic assay: Adenylosuccinate We anticipated two genetic events thatwould lead to lyase activity wasmeasured accordingto the method of WOOD- a fast growing phenotype and thatwould not be due to WARD (1978). Briefly, yeast strains were grown in 50ml of SD medium to an ODGooof 0.6 ? 0.1. Cells were harvested, washed changes in the regulationpathway for purinebiosynthe- with breaking buffer (0.05 M TrisCl pH 8, 20% glycerol and sis: gene conversion at the ura3locus or &acting muta- 1 mM DTT) and resuspended in 0.250 ml breaking buffer. tions in the ADE5,7-URA3 reporter. To eliminate these Phenylmethyl sulfonyl fluoride (PMSF) was added to a final two classes of mutations, we independently tested ex- concentration of 2 mM and the cells were then broken with pression from an ADEl-lacZ fusion. URA3 gene conver- glass beads by vortexing four times for 30 sec in the cold. After addition of 0.250 ml of breaking buffer,glass beads and sions orpromoter mutations in the ADE5,7 fusion unbroken cells were pelleted in a microfuge for 5 min and would be expected to show normal regulation of the either 50 or 100p1 of the supernatant wereused for the ADEl-lacZ reporter. Bona jide transactingmutations enzymatic assay. The assay was done in 50 mM TrisCl pH 8, should show constitutive expression of this reporter. 48 pM adenylosuccinate monophosphate (AMPS) in a final This second reporter also allowed us to measure the volume of 1 ml. Conversion of AMPS to AMP by adenylosucci- nate lyase was followed as the decrease of absorbance (10.7/ degree of the derepression in the mutant strains. nmol/ml) at 280 nm. The specific activities are expressed as All the 84 candidateconstitutive mutants were grown nanomoles of AMPS consumed per min per mg of protein. on medium containing 5-fluoroorotate(5FOA) (BOEKE Protein concentration was determined using the Bio-Rad Pro- et al. 1984) to cure the ADE5,7-URA3 plasmid. These tein MicroAssay System, with crystalline bovine serum album- strains were then retransformedwith a plasmid carrying ine serving as the reference standard. an ADEl-lac2 fusion (named P115). 0-galactosidase (0- Gal) activity was then determined underrepression and RESULTS derepression conditions. Thirty-nine mutants were not In this section we will first describe isolation of mu- studied further because they either did not grow on tantsthat constitutively express adenine biosynthetic 5FOA or displayed normal regulation of DEI-lac2 ex- genes, then we will present our analysis of the mutant pression. The remaining mutants were called bra for phenotypes and the results of the complementation bypass- of repression by adenine. analysis. Finally, characterization of several genes in- The bri mutants fallinto three phenotypic classes volved in the signalling cascade will be shown. (see Figure 1 for examples of each class). The class 1 386 M. L. Guetsova, K Lecoq and B. Daignan-Fornier 1 TABLE 2 Expression of ADEl-LacZ, ADE2LacZ and a mutated version of ADE2LacZ deprived of its two Baslp binding sites in the wild-type strain PLY122 and five bra mutants
PGal units ADEl-Lad ADE2-LacZ mutADE2-LacZ Strain +ade -ade +ade -ade +ade -ade PLY122 17 182 52 15713 15 128 106 150 230 20916 17 200 131 136 158 175 195 15 18 216 142 139 177 176 19 17 217 661 757 598 5 560 6 100 220 245 208 174 18213 14
0 wt(PLY122) l(221) l(215) 2 (139) binding site in its promoter (DAIGNAN-FORNIERand Phenotypic class FINK1992). In these five bra mutants, expression of the FICL~RE1.-Examples of the three classes of bra mutants. mutated ADE2"lacZ fusion was as low as inthe wild-type Constitutive expression was tested using an ADEI-lacZ fusion. control strain,showing that Baslpis absolutely required Verticalbars represent PGal units under derepression and for the expression of the constitutive phenotype of repression conditions for the wild-typestrain (wt) and the these mutants (Table 2). three different phenotypic classes numbered 1-3 below the Additional phenotypes conferred by the bra muta- line. For each example, the mutant name is indicated between parentheses. tions: All the mutants were tested for growth defects, eight of them exhibit a slower growth at 15". It has not been shown yetwhether this phenotype is linked to the mutants express the fusion at the wild-type level under derepression of adenine biosynthetic genes, but thefact derepression conditions. Under repression conditions that several of these mutants belong to common com- (in the presence of adenine), expression of the fusion plementation groups (see below) is consistent with the is increased in themutant relative to the wild-type idea that the cold-sensitivity and derepression are the strain. The repression factor, defined as the ratio of result of the same mutation. As stated above, two mu- fusion expression in the absence and in the presence tants (115 and 241) are complete adenine auxotrophs of adenine, varied from 0.9 to 5.9 in this classof mutants and another mutant (217) is a partial auxotroph. compared to 7.6 and 10.7 in the two wild-type strains. A Mutations in genes encoding enzymes in purine sal- total of 19 mutants fall into class 1. The class 2 mutants vage pathwaysmight be expected to produce a bra dere- express more fusion protein than wild type under both pression phenotype. Such mutations should be unable repression and derepression conditions. For the two to convert base analogues into the the toxic derivatives mutants in this class the repression factor is the same that act in vivo and are therefore likely to confer resis- as inthe wild-type control. Class 3 mutants express more tance to these drugs. Possible resistance of the bra mu- PGal activity than wild-type under both repression and tants to three such analogues, 8-azaadenine (SAA), €4- derepression conditionsbut the repression factor is azaguanine (SAG) and 5-fluorocytosine (5FC), was much lower than in wild-type cells (1.1 to 5.8 compared therefore tested by a petri plate growth assay. Several to 7.6 and 10.7 in the two wild-type strains). The size of the bra mutants do confer resistance to one or more of this class is 20. Finally, four mutants could not be of these drugs (Table 3). placed into classes either because they revert with a very The bra mutants fall into more than seven comple- high frequency (mutants 134 and 224) or are unable mentation groups: Each of the bra mutant was crossed to grow in the absence of adenine (mutants 115 and to an isogenic wild-type strain. The phenotype of the 241) and could therefore not be tested for derepres- heterozygous diploids was then scored using the growth sion. assay for expression of the ADE5,7-URA3fusion. Thirty A similar analysis was performed on five bra mutants mutants were found to be fully recessive and 15 were with an ADE2-lac2 fusion (Table 2). These results show dominant or semidominant. Pairwise crosses of the re- thatthe phenotypic classes arenot specific forthe cessive mutants from opposite mating types were made ADEl-lad fusion. We also tested whether the bra mu- and a complementation analysis was performed. The tants require Baslp to express their derepression phe- partially dominant mutants were also crossed to oppo- notype. The same five mutants were tested for their site mating type mutants and growth of these diploids ability to express an ADE2-1acZ fusion with no Baslp in the absence of uracil and in the presence of adenine Mutants constitutive for ADE genes 387
TABLE 3 Phenotypical analysis of the 20 mutants belonging to complementation groups bral-bru7
BGal units (ADEI-~ucZ) Repression Additional Repression M utant" Class Mutant" +ade -adephenotypesb factor
~~~ 17 182 10.7PLY122 (wt) 182 17 bral-1 (115) NA" Ade- h~l-2(213) 3 2.7 109 299 h~l-3(231) 3 3.8 109 410 h~2-1(1 18) 3 56 5.2 291 cs, SD, 8AAR h~2-2( 124) 3 172 408 2.4 cs, 8AAR h~2-38AAR(211) 3 2.1 197 413 bra3-1 (129) 8AAR,1 2.861 171 SAGR bra3-2 ( 130) 8AGR 18AAR, 1141.6 186 bru3-3 (220) 18AAR, 2450.8 208 SAGR h3-4 (239) 8AGR 18AAR, 1021.6 162 bra4-1 (132) 2 688.2 560 cs, 8AAR bra4-2 (242) 3 1952.8 537 cs, 8AAR hu5-l (109) 1 5.9 26 153 h5-2 (206) 3 278 479 1.7 cs, SD, 8AAR bra5-3 (208) 3 253 471 1.9 cs, SD, 8AAR bra61 (134) d ha62 (216) 1391 142 1.o 8AGR bra63 (224) d h7-l (119) 1 1.6 89 143 5FCR, 8AAR, 8AGR bra7-2 (232) 1 142 204 1.4 5FCR, 8AAR, 8AGR wt, stands for wild-type control, numbers inside brackets indicate originalmutant name. cs, SD, 8AAR, 5FCRand 8AGRstand, respectively, for cold-sensitive,semidominant, resistantto 8-azaadenine, resistant to 5-fluorocytosine and resistant to 8-azaguanine. Assignmentof semi-dominant mutants to comple- mentation groups is uncertain (see text for details). NA, not applicable. dThese two mutant strains belonging to the bra6 complementation group revert with high frequency and were not tested for additional phenotypes. was estimated by comparison to the wild-type isogenic Two mutants (1 19 and 232), belonging to the same cross. Since these partially dominant mutants may not complementation group (bra7), are resistant to 5FC. be due to loss-of-function mutations, assignments of Both mutants are also resistant to other purine base these mutants to a given complementation group are analogues, probably because thetransport of these uncertain. As shown in Table 3, 20 bra mutants define drugs is diminished in these mutants. When bra7-1 was at least seven complementation groups. The remaining crossed to a wild-type strain, the absence of repression mutants complement all mutants of opposite mating by adenine (followed by the growth assay for the expres- type. In most cases, the mutants in a given complemen- sion of the ADE5,7-URA3 fusion) segregated as a muta- tation group show similar additional phenotypes and tion in a single nuclear gene (2:2 in 16 tetrads). fall in the same phenotypic class (see previous section Complementation analysis suggested that bra7-2 is al- and Table 3). lelic to FCY2. A centromeric plasmid carrying the FCY2 BRA7 is FCY2: One possible mechanism by which gene (pRFF2 kindlyprovided by M. R. CHEVALLIER)re- some of our mutants fail to repress gene expression in stores normal regulation of an ADEI-lucZ fusion when response to adenine is that these mutants cannot take introduced into bra7-2 (data not shown).Also we found up adeninefrom the media. We have used a drug resis- that a strain containing a deletionof FCY2 fails to com- tance test previously described (CHEVALLIERet al. 1975) plement a bra7-1 strain. The adenine repression factor to identify such mutants. This test is based on the fact in a fq2/BRA7 diploid is 7.0 and falls to 2.9 in a fq2/ that mutations in the gene FCY2, encoding the purine- bra7 isogenic diploid. Furthermore, the latter diploid cytosine permease, lead to resistance to a toxic cytosine is resistant to 5FC but the former is sensitive. Finally, analogue, 5-fluorocytosine (5FC). Thebra mutants were both diploids were sporulated and after tetrad dissec- therefore tested for growth on medium supplemented tion the resistance to 5FC segregated 2:2 in the fq2/ with 5FC. BRA7 diploid (eight tetrads) while only 5FCR spores M. L. Guetsova, K. Lecoq and B. Daignan-Fornier
PRPP bral-1 and bra9-l are as follows: (1) poor growth at low concentrations of adenine even on rich medium, (2) ATP GTP t spores carrying the mutation germinatepoorly, (3) ger- dADP I de novo mination defect that can be rescued by a mutation up- biosynthesis stream in the pathway. RNR \IADP i GDPt Because of the germination defect of the adenine- specific mutants, it was not possible to study the segrega- tion of bru9-1.To bypass thisproblem, thebra9-1 mutant was crossed to an isogenic wild-type strain (L3862) car- rying an ade2 mutation, since the germination defect of adel2 strains was rescued by mutations upstream in AMP IMP “----) GMP ADE13- ADEl2 - the pathway. A spore carrying both ade2 and bra9-1 mu- tations was selected on the basis of itsred color (charac- teristic of ade2 mutants) and growth on adenine and APT1 AAH 1 pRpp+Tf not hypoxanthine due to the bra9-1 mutation. This 1“--I- strain was then crossed to an isogenic wild-type strain carrying an ade2 mutation (3’608) ; the resulting diploid Adenine -Hypoxanthine Guanine is homozygous for ade2 and heterozygous for bra9-1. L Fourteen tetrads from this cross were analyzed. We ob- FCY2 my2 FCY2 1 served 2:2 segregation for lack of growth on hypoxan- thine supplemented medium and for derepressed ex- (Extracellular purines I pression of the ADEl-LacZ fusion. Both phenotypes co- segregate. FIGURE2.-Schematic representation of purine intercon- To isolate the BRA9 gene, the bra91original mutant version in yeast. The following abbreviations are used: PRPP, was transformed with a yeast genomic library carried 5-phosphoribosyl-1-pyrophosphate; IMP, inosine 5”mOnO- on a centromeric vector. Two plasmids able to comple- phosphate. Genes names are indicated in italic and encode the following enzymatic activities: MI,adenine deaminase; ment the adenine requirement of bra9-1 were isolated. ME15 adenylosuccinatesynthetase; ADE13, adenylosucci- These plasmids wereshown to carry overlapping se- nate lyase;ADKl, AMP kinase; AMDl, AMP deaminase; APTl. quences that hybridize to chromosome XIV (data not adeninephosphoribosyl transferase; FCY2, purinecytosine shown). The plasmid P103, able to complementthe permease; HPTl, hypoxanthineguanine phosphoribosyl bra9-1 auxotrophy, was recently also shown to carry the transferase; RNR, ribonucleotide reductase. The? symbol rep DE12 gene (GALLERTet al. This conclusion is resents possible GMP reductase activity discussed in the text. 1996). For simplification purpose nucleosides are not represented. based on several lines of evidence: (1) it can specifically suppress the growth defect of a purA mutant in E. coli were recovered in the nine tetrads dissected from the (purA is the gene encodingadenylosuccinate synthetase fcy2/bru7 diploid. Cosegregation of 5FC resistance and in E. coli), (2) the P103 plasmid contains a sequence constitutive expression of ADEl-lucZ expression was encoding a polypeptide highly similar to adenylosucci- confirmed, establishing linkage between fcy2 and bru7. nate synthetases from other organisms, (3) thisse- From these data we infer that the purine cytosine per- quence maps physically very close to the adel2 locus. mease is required for adenine repression, most likely The conclusion that bra9-1 and adel2 are mutations because the transduction pathway monitors the intra- in the same gene is further supported by the fact that cellular concentration of adenine or a derivative. the bra9-1 mutation cannot complement the auxotro- BRA9 is ADE12: Two mutants, brul-1 (115) and bru9- phic phenotype of an udel2 mutant. Both mutations 1 (241), are adenine auxotrophs. Both were tested for also lead to an absence of adenylosuccinate synthetase complementation by strains containing mutations that activity in vitro (GALLERTet al. 1996). Finally, linkage affect steps in the de novo pathway for adenine biosyn- between bra9and the clonedADE12 gene was also dem- thesis (adeI-ude9); both mutants complement all the onstrated. The LEU2 marker was integratedat the tested mutations. bral-1 and bru9-1 do not grow on a DE12locus by transformation inan ade2 mutant strain medium supplemented with hypoxanthine, a base that (see MATERIALS AND METHODS). The resulting strain, is converted into IMP and can therefore rescue all the named Y608, was crossed to the ade2 Y399 strain car- mutants affecting the de novo pathway leading to synthe- rying the bru9-1 mutation. In 14 tetrads only parental sis of IMP (see Figure 2). brul-1 and bra91 therefore ditypes were observed (all the Leuf spores were Ade’ appear to be “adenine-specific.” Two mutants with sim- and all the Leu- spores were Ade-). Together, these ilar adenine-specific phenotypes, adel2 and ude13, have results show that bra9-1 isa mutationin the ADE12 gene been described (DORFMAN1969). Additional pheno- encoding adenylosuccinate synthetase. types found in udel2 mutants (DORFMAN1969) and in bra1 mutants affect adenylosuccinatelyase activ- Mutants constitutive for ADE genes 389 ity: The second mutant showing an adenine-specific re- A 2.9-kb EcoRI fragment carrying only the YLlU59w quirement is bral-1. Interestingly, the bral-2 and bral- ORF (see Figure 3) was subcloned in a LEU2 CEN vector ? mutants are not adenine auxotrophs, demonstrating and shown to be able to restore adenine prototrophy that the constitutive expression phenotype and the ade- to the bral-1 and bra8-1 mutants. Furthermore, thesame nine requirement can be separated. The Bra- pheno- plasmid can complement the derepression phenotype type (monitored with an ADEl-lacZ fusion) of the bral- of bral-2 and bral-? mutant strains (data not shown). 3 mutant was shown to segregate 2:2 in a cross with an Linkage between bral and the ADE13 gene was estab- isogenic wild-type strain (19 tetrads). The BRAl gene lished asfollows: the LEU2 marker was integrated at was cloned by complementation of the adeninerequire- the ADEl? locus by transformation (see MATERIALS AND ment of a bral-1 strain. Two plasmids carrying overlap- METHODS for details), the resulting strain (Y610)was ping genomic inserts were isolated. As expected, the then crossed to the bral-? mutant and expression of BRAl gene also complemented bral-2 derepression an ADEl-lacZ fusion in 19 tetrads from this cross was phenotype. determined. Constitutive expression of the fusion in Surprisingly a third mutation, bra8-1 (mutant 217), the presence of adenine was found inall the Leu- was also complemented by BRAl both for derepression spores and in none of the Leu+spores. bral and ADEl? and for a slight adenine requirement that cannot be are therefore tightly linked. It is very likely that bral rescued by hypoxanthine. When bra8-1 was crossed to and bra8 are allelic to the previously described adel3 an isogenic wild-type strain (PLYl21), the subtle ade- locus (DORFMAN,1969) but linkage analysis could not nine requirement was shown to segregate 2:2 in 10 tet- be performed because the original adel? mutants have rads. Furthermore, using an ADE2-lacZ reporter, the been lost. adenine requirement of bra8-1 was found to be linked To further test if the bral and bra8 mutants are muta- to the derepression phenotype. Since bra8-1 and bral- tions in the structural gene for adenylosuccinate lyase, 1are fully recessiveand fully complement for derepres- enzymatic activitywas measured in these mutant strains. sion, these two mutations were placed into different We found that enzymatic activity is low in all the bral complementation groups, but the complementation of mutants and in the bra8 mutant (Table 4). It is not both mutants by the same plasmid strongly suggest that affected in the bra9mutant,which is not complemented they are complementing alleles of the same gene. Link- by the plasmid carrying the ASL encoding gene. Both age between bra1 and bra8 was studied by crossing bral- bral and bra8 mutants therefore have decreased adenyl- 2 and bra8-I strains, the resulting diploid was sporulated osuccinate lyaseactivity, consistent with our genetic and the spores were analyzed for adenine repression analysis. Of note, the lowest ASL activity was observed using an ADEl-laczfusion. All the tetrads (20) contain in the bral-1 mutant; bral-1 is the allele that confers an four spores showing a Bra- phenotype demonstrating adenine auxotrophy. The prototrophic alleles lead to that bral and bra8 are tightly linked. This result is not decreased enzymatic activity but apparently a low level surprising since intragenic complementation has been of activity is sufficient to sustain growth in the absence described previously at homologous loci in other organ- of adenine. This demonstrates that mutations at this isms (WOODWARD et al. 1958; FOLEYet al. 1965). locus causing derepression of the purine pathway can The structure of the BRAl genomic locus was further cause either adenine auxotrophy or not. characterized. It contains fourinternal BamHI frag- BRA6 is HFZ'l, the gene encoding hypoxanthine-gua- ments (between 1.5 and 2.5 kb, see Figure 3).Deletion ninephosphoribosyl transferase: The bra6 comple- of all four fragments abolished complementation. None mentation group contains three alleles. Two of the bra6 of the four fragments alone complemented bral-1, sug- alleles (134 and 224) are difficult to study because they gesting that one of the BamHI site was in BRAl (see revert at a high frequency. Further studies on this com- Figure 3).The ends of the BamHI fragments were there- plementation group were therefore done with bra6-2. fore sequenced and two of them were found to be in This mutant belongs to the first phenotypic class (Table an open reading frame(YLR359w) encoding apolypep- 3). The bra6-2 mutant was crossed to an isogenic wild- tide highly similar to adenylosuccinate lyase (ASL) in type strain (PLYl21) and the resulting diploid was spor- other organisms (see Figure 4). Of note, a histidine ulated. The derepression phenotype associated to the residue that is part of the B. subtilis enzyme active site bra mutation was followed using an ADEl-lacZ fusion (residue 141, LEEet al. 1997) is conserved in the yeast and shown to segregate 2:2 in 11 tetrads (datanot sequence (at position 134). ASL catalyzes two steps in shown). thepurine biosynthesispathway, one in the de novo BRA6was cloned by complementation of bra6-2. A biosynthesispathway and one in the interconversion genomic library carried on a LEU2 centromeric vector of IMP into AMP. Adenylosuccinate lyase activity was was introduced into bra6-2 and complementing clones previously shown to be abolished by adel3 mutations were identified by the ability to restore wild-type regula- (DOWMAN1969). The adel? locus has not been tion of the ADEl-ZacZfusion in the presence of adenine. mapped. Two complementing plasmids (P132 and P133) were 390 M. L. Guetsova, K. Lecoq and B. Daignan-Fornier
YLR357w YLR356w + ADE13 YLR360w 'YLR361c +IL v5 YLR35&* Complementation 1 1000 2000 3000 4000 60005000 7000 8000 9000- 10000 of Mal-1 1 I I I I I I I I I L I I I I + B B B E BI E E
+ FIGURE3.-Schematic physical map of the yeast DNA insert carried by P127 plasmid that complements the brul-1 auxotrophy for adenine. Numbers over the line refer to the size of the DNA fragment in base pairs. E and B stand for EcoRI and BumHI restriction sites, respectively. ORFs longer than 100 codons deduced from the complete nucleotide sequence of chromosome XI1 are represented as arrows on the top of the drawing. Restriction fragments used for subcloning experiments are drawn at the bottom of the drawing with complementation result presented on the right part of the figure. isolated. One plasmid (P133) was used for physical map- when adenine is provided as a purine source because ping. Hybridization to separated yeast chromosomes adenine phosphoribosyl transferase (APRT) can con- demonstrated that theP133 insert is derived from chro- vert adenine to AMP, which is then converted to IMP mosome N.This result was confirmed by hybridization by AMP deaminase (see Figure 2). Therefore this mu- to an ordered lambda library (kindly provided by L. tant will grow on adenine but not on hypoxanthine as RILES and M. OLSON)revealing two spots correspond- a purine source. The P386 plasmid carrying only the ing to two overlapping clones mapping between ade8 YDR399zu ORF (see Figure 5) and a control plasmid and snfl. The chromosome IV sequences from the yeast (pRS316)were introducedinto the W109-9C strain. genome sequence allowed us to precisely position the Growth of the transformants was then monitored on P132 and P133 insert sequences. As shown in Figure 5, minimal medium with either adenine or hypoxanthine the two plasmids contain overlapping sequences, the as a purine source. By contrast with the controlplasmid, common part of which contains four ORFs. Further P386 supported growth on media containing hypoxan- subcloning (see Figure 5) established that YDR39% thine as a purine source. Because the YDR39% ORF (plasmid P386) is the locus that complements the Bra carried on a centromeric vector complements the hptl phenotype (Table 5A). This ORF encodes a polypeptide mutation, this ORF was renamed HPTl. presenting some similarities with hypoxanthine phos- The HPTl gene was disrupted in the PLY122 strain phoribosyl transferases (HPRT) from several organisms (MATERIALS AND METHODS). Isogenic wild-type and dis- (see Figure 6). Oneof the most conserved regions (resi- rupted strains were then transformed with plasmid P473 dues 103 to 119 in the yeast sequence) is a potential carrying a ADEl-lacZ fusion. In the hptl::URA3 strain PRPP binding site as defined by HERSHEYand TAYLOR (Y508) the fusions are derepressed in the presence of (1986). Mutations named hptl hadbeen previously adenine (Table 5B) and are resistant to 8AG. Linkage characterized by WOODSand coworkers that lead to the analysis also confirmed that bra6 and hptl are the same loss of HPRT activityand resistance to a base analogue locus. A ha6-2 strain 01'531)was crossed tothe named 8-azaguanine (SAG) (WOODS et al. 1983). In- hptl::URA3 strain and the diploids were sporulated. In deed, the bra62 mutant is specifically resistant to SAG, 10 tetrads the resistance to 8AG segregated 4:O suggesting that ha6 and hptl could be the same locus (8AGR:8AGS).Furthermore, with the ADEl-lacZ fusion (Table 3). introduced into the spores from five tetrads, we found The following experiments were performed to con- that all the meiotic progeny from this cross showedthe firm that bra62 is allelic to hptl. First, we took advantage derepression phenotype. The factor of repression by of a strain (W109-9C) mutated at the hptl and ah2 adenine in the mutant spores varied from 0.9 to 2.5, loci. Such a double mutant cannot perform thede nouo while it was 4.1 in the wild-type control. As a final test synthesis of purines and cannot use hypoxanthine as a of the idea that bra6 and hptl are allelic, the bra62 purine source. However, this double mutant can grow mutant strain Y531 and the isogenic wild-type strain Mutants constitutive for ADE genes 391
-do mmPcvm OOCuWd mmmdd aoaoomao ro\oaoPao dddom aoaoaowm rnrnPWo\ hlcvddv) aoao0Q)d ddPdP 00Cuom -0 mmmdu) OIOIddrl rlrlddd CuCucvcUcv cvcvmcvm mmmmm ddddd zki
-x2 *r.Ll* * * .4 W4>4Z . . . .J o>w...... 4 0~0..
H WWZ * *
ncr . .
Z>EHH UV>4G hC444U nnwws JJ>EW XaHnW . . . .o-J
* exax ~.sMB
* .WHO * wcf-lwx - * -a .a zz . .pl .m . . . .pl .> . .
* -2*I3 . .a .J * .>I .tl) a .n .W * .a .J
a; la Cziumu S3vmu Cgvmu Sgumu Sz3umu Sziumu ~gumu~gumu - " 3 umwmw uzmmw umwmw umcnmw uzwmw uzwmw umwmw umwmw IIIII dld'dW dLl1dW d~d~d~dldld~dld~d~dl d~d~dLLI1 dld~d~d~d~ dldldld'dl d d rlr( d 536 mmmmm mmmmm mmmmm mmmmm mmmmm mmmmm mmmmm mmmmm 202 44444 44444 44444 44444 4444444444 44444 44444 y2 392 M. L. Guetsova, K. Lecoq and B. Daignan-Fornier
TABLE 4 -10 Chromosome IV Adenylosuccinate lyase activity in the bral, brag, bra9 E B EB mutants and isogenic and wild-type strains PLY121 PLY122 .-bad-4A YDR3BBw- Hprl DITZ Dm RBP~MRP~ Complementation StrainActivity (nmolAMPS/min/mg prot) YDR399w of bras-2 P132 + PLY1 2 1 10.0 t- 2.0 P133 + bral-l 0.2 0.2 t- P484 - PLY122 13.3 2.6 t- P388 - bra 1-2 1.1 t- 0.4 + bral -3 1.2 t- 0.5 FIGURE5.-Physical map of the chromosome IV region bra8-I 2.1 5 0.3 carrying the HPTl gene. E and B stand for EcoRI and BamHI bra9-1 18.0 t- 0.8 restriction sites, respectively.ORFs deduced from the nucleo- tide sequence are represented as arrows below the line. Sub- Activity is the average of three assays. See MATERIALS AND cloningstrategy and results of complementationare pre- METHODS for details. sented at the bottom of the figure.
PLY121 were crossed to the previously characterized tive (data not shown),suggesting that they act through hptl mutant from WOODSand coworkers (W109-9C). the same signalling pathway. Second, the aptl mutation Although the heterozygous HPTl/hptl diploid was sen- does not affect repression by adenine, suggesting that sitive to SAG, the ha6 X hptl diploid is resistant (data the adenine repression signal could be carried by a not shown). Both diploids were tranformed with the metabolite in the HPRT route. If this is correct, aahl P115 plasmid carrying an DEI-lac2 fusion and the re- mutations should have the same effect on derepression pression by adenine was estimated by measuring PGal as hptl mutations. Surprisingly although an hptl null activity in the presence or absence of adenine. Results allele leads to derepression, mutation of the aahl locus presented in Table 5C clearly show that hptl cannot has no effect on adenine repression. We interpret this complement bra6-2 for derepression. The ha6 X hptl observation as follows. In wild-type strains most of the diploid was sporulated and segregation of resistance to available adenine is metabolized to hypoxanthine by 8AG was monitored in 14tetrads. All of the spores from adenine deaminase; however, in the hptl mutant strain, this cross wereresistant to SAG, demonstrating thatbra6 hypoxanthinecannot be further metabolized and and hptl loci are tightly linked. In sum, we conclude therefore does not activate a repression signal. In the that bra6 and hptl are the same locus and that they aahl mutant,the adenine normally deaminated by encode yeast HPRT. Aahlp is available for utilization by Aptlp, thus allowing Role of the APTl and AAHl genes in the process of synthesis of the factor that activates repression. repression by adenine: Once inside the cell, adenine By this hypothesis, adenine that is not used in one can take two different metabolic routes (see Figure 2). route is used in the other. A prediction of this model It can be metabolized into AMP by APRT (encoded by is that a double aahl apt1 mutant should be fully dere- the APTl gene, ALFONZO et al. 1995). Alternatively, it pressed. The desired double mutant was isolated from can be deaminated to hypoxanthine by the adenine a cross between aptl::CTRA3 (Y548 or Y549) and deaminase (encoded by the AAHl gene, WOODS et al. aahl::URA3 (Y520), haploid strains. Three such double 1984, DEELEY1992), and then transformed intoIMP by mutant spores (Y550, Y551 andY552) were isolated and HPRT (encoded by the HPTl gene, WOODSet al. 1983 the presence of the double disruption was confirmed and this work). Our finding that ha6 is allelic to hptl by Southern blot analysis (data not shown). TheADEl- demonstratedthat the “HPRT route” plays an im- 1acZreporter was introduced into these strains, and the portant role in the process of adenine repression. To effect of adenine, hypoxanthine and guanine on ex- further evaluate the contributions of these two path- pression of the fusion in the transformed strains was ways,we constructed isogenic strains with disruptions determined (Table 6B). As predicted, in thedouble of aptl, aahl and hptl (see MATERIALS AND METHODS). aahl aptl mutant regulation by adenine is abolished Adenine regulation was analyzed in these strains while regulation by hypoxanthine is unaffected. An- (termed Y511,Y520 and Y508, respectively). We also other prediction is that overexpression of APTl should tested two other purine bases, hypoxanthine and gua- increase the flux of adenine used for synthesis of AMP nine, for effects on transcriptional repression of the through APRT and should therefore at least partially ADEl-lacZ gene fusion. Several conclusions can be bypass the deregulation in the hptl mutant. The ADEl- drawn from the results presented in Table 6A. First, lacZ reporter and a multicopy plasmid carrying either adenine and hypoxanthine cause similar levels of re- the APTl gene (pCG3, DEELEY1992) were introduced pression but guanine causes only partial repression. It into bra6-2 mutant strain. Results presented in Table 7 is noteworthy that theeffects ofadenine, hypoxanthine clearly show that overexpression of APTl abolishes the and guanine on transcriptional repression are notaddi- derepression phenotypeof the hptl mutation, therefore M utants constitutive for constitutive Mutants ADE genes 393 TABLE 5 Expression of ADEhZ fusions in different HPTl genetic backgrounds
Plasmids PGal activity PGal Plasmids Relevant Strain genotype Strain +ade OtherLacZ fusion -ade RF" 8AGb A. PLY122 (BRA@ P2 (ADE2-Lac9 pRS316 (control) 27 80 3.0 s' PLY122 (BRA@ P2 (ADE2-Lad) P386 (HPTl CEN) 28 83 3.0 S 216 (br~6-2) P2 (ADE2-La4 pRS316 (control)80 66 1.2 R 216 (br~6-2) P2 (ADEZ-Lac4 P386 (HPTl CEN) 28 77 2.8 S B. PLY122 (HPT1) P473 (ADEl-LacZ) - 6.0 28 169 S Y508 (hptl::URA3) P473 (ADEl-Lad) - 134 56 2.4 R C. W109-9C X Y349 (hptl/HPTl) P115 (ADEl-Lad) - 2 25 12.5 S W109-9C X Y531 (hptl/b~~6-2) P115 (ADEI-LucZ) - 21 32 1.5 R
a RF, repression factor. 8AG, 8-azaguanine. R and S, resistant and sensitive, respectively. establishing that the total flux between the two routes blocked using hydroxy urea (HU). A strain carrying an is critical for the repression mechanism. ADEl-LacZ fusion integratedat the ADEl locus was To exert its regulatory effect, adenine has to be con- grown in SD medium with or without adenine that con- verted into ADP but not into dADP Results presented tained increasing concentrations of HU (5-80 mM). in the previous sections strongly suggest that adenine After 13 h expression of the fusion was monitored and once inside the cell needs to be metabolized into AMP no effect of HU on regulation of the fusion could be to exert its regulatory role. We have tested whether detected even under conditions where growth is se- transformation of AMP into ADP was required for ade- verely affected by the drug (data not shown). nine repression. For this purpose, we have used a strain disrupted at the ADKl locus (KONRAD1992). In this DISCUSSION strain only 10% of wild-type AMP kinase activity can be detected in a crude extract (KONRAD1992). Results To investigate the signalling pathway controlling ade- presented in Table 8 clearly showthat expression of an nine responsive genes, we have isolated constitutive mu- ADEl-LacZ fusion in this mutant strain is totally unaf- tants that relieve the transcriptional repression of ADE fected by addition of adenine in the medium. There- genes by adenine. Afull understanding of the pathway fore we conclude that AMP has to be converted into will require the identification of the following: (1) the ADP for correct transduction of the repression signal. signal (the effector molecule) (2) the transcription fac- Finally, we have tested whether reduction of ADP into tors responding to the signal and (3) the proteinfactors dADP was required for repression by adenine. This was that are required for the perception of the signal and done by two different approaches. First, we used a tem- for its transduction to the transcription factors. Our perature-sensitive allele (named mt6-68) of the RNR2 phenotypical and molecular analysis of the bra mutants gene, which encodes a subunitof ribonucleotide reduc- sheds light on the two first points. tase (ZOU and ELLEDGE 1992). This strain was cured ADP or a derivative of ADP is the effector mole- for the pZZ13 plasmid and then cotransformed with an cule: Because S. cermisim doesnot take up external ADE2-CEN vector (pASZ11, STOTZand LINDER1990) nucleotides, the natureof the effector cannot be simply to make it Ade' and with the P473 plasmid carrying tested by adding nucleotides to the medium. Our ge- the ADEl-LacZ fusion. Expression of ADEl-LacZ in this netic analysis provides strong clues about the identity mutant strain and in the wild-type isogenic strain was of the effector molecule. First, the fact that BRA7 is then measured after growth at 30". This temperature allelic to FCY2, the gene encoding purine permease, was chosen because at 30" the E21mutant strain grows indicates that purines need tobe taken up into cells to much slower than the isogenic wild-type strain (named trigger repression of biosynthetic genes. It is unlikely Y203),indicating that under these conditions synthesis that purine bases themselvesare the effector molecules of dNTPs is most probably limiting for growth. Results because mutations that block their metabolism abolish presented in Table 8 show that the mt6-68 mutation has their regulatory effect (for example, hptl mutation in no effect on repression by adenine. The same result the case of hypoxanthine or a double aahl apt1 muta- was obtained when ribonucleotide reductase activity was tion in the case ofadenine). Second, ourresults suggest 394 M. L. Guetsova, K. Lecoq and B. Daignan-Fornier
00 t-rlddul Wo\WOO WOOPJo\ MCrLnFW that the major route for adenine utilization under the MMeWd, Waddo\ ddddrl ocJoFd e, WWddPJ & tested conditions is its deamination into hypoxanthine ddPJcJN -3E .[I)&Xc) JJc)[3 * a:>H>a: -us followed by transformation into IMP. This result is in &v)ZXc) " 2s good agreementwith direct measurement of enzymatic .w .c) XSdU. sa . B!!!::::: c . .J .c) HJgU. ~E~Lcu. .a:. . up: activities (DEELEY1992). Third, the fact that mutations fs ;I WWUb'5r: * .w - * 5,- that decrease AMP synthesis (mutations at the ALE12 : :; :; k$ . .c) . . n . .#a .H ZLI ..a: .. Sp and ADE13 loci) were obtained in this screen strongly aoz *H WL3ZHCx ..a: .. tt: suggests that AMP plays an important role in the ade- BWV *J U!XUOW ..Ms. a, nine repression process (see Figure 2). This conclusion ~za~WBUc)> . .a * . zg 3S;V)DL V)PWB[3 E-b>L?> * -a ' * sg is strengthened by the experiment inwhich overex- mzs 3WMv)H >UJ4h * .w '01 5- pression of APT1 in the hptl mutant restores wild-type :4a:I;wa l3ZBcY5 * *J .M e, 2 m@Fa repression. Finally, mutation of the gene encodingAMP UIWCIMZ V)0[3[I)a * . * .c7 * .w .x gg- l3 . . . .h * .x4w kinase, ADKI, abolishes transcriptional regulation by e.. .z .og:o1 d . $g adenine. This result suggests that ADP or a derivative W3lXCIZ v) . . . .@ -.z>w 28. * * * *M of ADP is the effector. 3 Wwl3'5r:h 2%; S **-*EWQW4H XcE The basis for the different levels of constitutive ex- 3 * * * .w >;I>Mh .ec. wi3ila> Qb!X:dO * * * .a >J>c)H c. M pression of the ADE genes in the bra mutants is not yet sr;. . . .v) a:Hv)l3'CI ?d% clear. Mutations in the same complementation group >HXEEV .. . .'5r: &>&wZ i?*yfL5s usually fall into the same phenotypic class (Table 3). Wpmq.1 --w-'% * * . .w p4wqwc) 2 2 %
-clvl sis. It was therefore thought that lacks GMP 2.dQ C V 2.dQ C V 3.4 Q C V 3.dQ C V S. cereviside VrdQav) VrdQaW VrdQ!Uv) mdQav) '.ha VOhE I UOhE I VOhE I UOhE I aos reductase activity (see discussion in BURRIDGEet ul. avLlt3u avLl7u avLlt30 avLlt3u mwbxh mwbxh mwwxh mwBxh 28W 1977). Our results indicate that guanineexerts a partial AduIA2 Au'AA2 ;Jul;g duIdJg 2 5.s repression effect that requires HGPRT activity (see Ta- hLlLlLl ULlLlLl LlLiLlLl LlhLlLl aaaa aaaa 2222 2422 -5 0" ble 6A). Since this repression by guanine is abolished xxxx xzxx v11 M utants constitutiveMutants for ADE genes 395
TABLE 6 TABLE 8 Expression of the DEI-LacZ fusion in the presence of Effect of mutations in the ADKl and RNR2 genes on different purine bases in strains carrying different expression of an ADEl-LacZ fusion combinations of hpt, dland apt1 mutations PGal activity PGal activity Relevant Relevant Strain genotype -ade +ade RF Strain genotypeStrain 0 ade gua hyp AH215 Wild type 412 25 16.7 A. AH215adkl adkl::HZS3 334 296 1.1 PLYl22 98 16 40 16 Wild type 10.8 Y508 hpt I 149 121 126 133 Y203 Wild type 29.8 2.8 rnr2 19.0 1.9 10.0 Y511 apt I 145 24 54 22 Y221 Y520 aahl 152 23 81 29 RF, repression factor. B. PLY122 Wild type 68 18 32 55 Y550 aptl aahl 90 92 24 53 pathway wouldconverge on thetranscription factor and Y55 1 aptl aahl 100 104 26 54 it would therefore be expected thatonly a few dominant Y552 aptl aahl 83 81 19 43 mutations at theBASl locus could lead to the derepres- 0, growth on SD medium containing no purine base; ade, sion phenotype. We have isolated several dominant mu- paand hyp, growthon SD medium supplemented withade- tations in our screen. Itwill be interesting to determine nine, guanine or hypoxanthine, respectively. if some of these mutations are in BASl. Is adenylosuccinatesynthetase a bifunctional pro- in a adel3 mutant (data not shown), this suggests the tein? From previous work (DORFMANet al. 1970, LOW existence of a GMP reductase activity providing a suffi- and WOODS1970) it has been proposed that theADEl2 cient amount of IMP and AMP to cause repression (see gene could encode both catalytic and regulatory func- Figure 2). This weak activity might not be sufficient to tions. This conclusion was based on the isolation of allow ade mutants to grow in the presence of guanine prototrophic regulatory mutants of adenylosuccinate as a purine source. Theexistence of such an enzymatic synthetase. This conclusion is at variance with our re- activity is supported by studies on intracellular purine sults. Although we have isolated a mutant (bra9-1) that content of cellsfed with radioactive guanine (BURRIDGE has lost both adenylosuccinate synthetase activity and et al. 1977). regulatory properties, we have also found that muta- What are the protein factors involved in the signal tions at the ADE13 locus are similarly deregulated. De- transduction pathway? We have shown that an ADE2- regulation is therefore not specific to ADE12, but is lac2 fusion mutated for its Baslp binding sites is not observed with any block in the pathway from IMP to regulated by adenine and is not derepressed by the bra AMP. Furthermore, we have shown, for some alleles mutations. This suggests an important role for Baslp of ADE13, that decreased enzymatic activity leads to a in regulation as well as activation. Since Baslp carries derepressed phenotype but no growth requirement for a potential nucleotide binding site in its protein se- adenine. It would be interesting to know whether the quence (TICE-BALDWINet al. 1989), it is tempting to adel2 prototrophic regulatory mutants previously de- propose that the effector could bind to Baslp, directly scribed have wild-type levels adenylosuccinateof synthe- affecting its capacity to bind DNA, interact with other tase activity. factors, or activate transcription. The central role for Yeast as a model to study purine metabolism regula- Baslp in this process is confirmed by the fact that all tionin higher eucaryotes: The genesthat we have the genes regulated by adenine isolated so far are also shown play central roles in yeast adenine regulation activated by Baslp. If there is a direct interaction be- correspondto important human disease genes. The tween Baslp and the effector, the signal transduction best understoodexample at the molecular level is Lesch-Nyhan syndrome, a syndrome whose symptoms TABLE 7 include hyperuricemia, severe mental retardation and Effect of overexpression of theAPT1 gene on expression automutilation (LESCHand NYHAN1964). Lesch-Nyhan of the ADEl-LacZ fusion in the h62 mutant strain syndrome results from the absence of HPRTactivity due to mutations in the HPRT gene (SEEGMILLERet al. PGal activity 1967). Patients with a partial defect in HPRT activity Plasmid +ade -ade RF have also been described, and they develop hyperuri- cemia but not the other features of Lesch-Nyhan syn- YEpl3 (control LEU2, 2y)1.9 73 39 drome (KELLEY et al. 1967). These HPRT-deficient pa- pCG3 (AFT1 in YEpl3 L.EU2, 2y) 11 66 6.0 tients show an increased synthesis ofpurine nucleotides RF, repression factor. and it was proposed that this could be dueto increased 396 M. L. Guetsova, K. Lecoq and B. Daignan-Fornier
PRF’P levels due to the lack of salvage of hypoxanthine CHEVALLIER,M. R., R. JUND, and F. LACROUTE, 1975 Characteriza- tion of cytosine permeation in Saccharomyces cereuisiae.J. Bacteriol. and guanine by HPRT (ROSENBLOOMet al. 1968). The 122: 629-641. excess of PRPP wouldbe shunted into thede novo path- DAIGNAN-FORNIER,B., and G. R. FINK,1992 Coregulation of purine way leading to increased purine biosynthesis. Lack of and histidine biosynthesis by the transcriptional activators BAS1 adenylosuccinate synthetase and adenylosuccinate lyase and BAS2. Proc. Natl. Acad. Sci. USA 89: 6746-6750. DANCV-D., M. VALENS, M.BOLOTIN-FUKUHARA, and B. DAIGNAN-FOR- have alsobeen shown to be associated, respectively, with NIER, 1994 A genetic screen to isolate genes regulated by the purine oversecretion (ULLMANet al. 1982) and mental yeast CCAAT-box binding protein Hap2p. Yeast 10: 1273-1283. retardation (STONEet al. 1992). DEELEY,M. C., 1992 Adenine deaminase and adenine utilization in Saccharomyces cereuisiae. J. Bacteriol. 174 3102-3110. Our findings in yeast suggest an appealing hypothesis DORFMAN,B-Z., 1969 The isolation of adenylosuccinate synthetase to explain the purine overproduction in Lesch-Nyhan mutants in yeast by selection for constitutive behaviour in pig- syndrome and related disorders. We found that muta- mented strains. Genetics 61: 377-389. DORFMAN,B., B. A. GOLDFINGER,M. BERGER,and S. GOL.DSTEIN,1970 tions in the yeast genes encoding HPRT, adenylosucci- Partial reversion in yeast: genetic evidence fora newtype of nate synthetase or adenylosuccinate lyase lead to dere- bifunctional protein. Science 168: 1482-1484. pressed synthesis of the de novo pathway enzymes. Fur- FOLLY,J. M., N. H. GILES,and C. F. ROBERTS,1965 Complementa- tion at the adenylosuccinase locus in Aspergallus nidulans. Genet- thermore,purine secretion has been described for ics 52: 1247-1263. certain alleles of the adel2 locus (LOMAXand WOODS GALLERT,K. C., T. OHANJAN,B. DAIGNAN-FORNIER,F. LOTTSPEICH, 1970) and can be associated with increased de novo syn- and G. KRAUSS, 1996 Enzymatic properties and inhibition by single-stranded ARS sequences of adenylosuccinate synthetase thesis of purines (BURRIDGEet al. 1978). Thereforemu- from Saccharomyces cereuisiae. Eur. J. Biochem. 239 487-493. tations at these loci could lead to purine overproduc- GIETZ,R. D., and A. SUGINO,1988 New yeast-Escherichia coli shuttle tion by deregulating thesynthesis of the de novo pathway vectors constructed with in vitro mutagenized yeast genes lacking rather than by increasing substrate (PRPP) availability. six-base pair restriction sites. Gene 74: 527-534. HERSHEY,H. V., and M. W. TAYLOR,1986 Nucleotide sequence and It would be interesting to test whether deregulationand deduced amino acid sequence of Escherichia coli adenine phos- overexpression of the de novo pathway enzymes is also phoribosyl-transferase and comparison with other analogous en- observed in human cell lines deficient in either HPRT, zymes. Gene 43: 287-293. KEI.I.EY,W. N., F. M. ROSENBLOOM,J. F. HENDERSON,and J. E. SEEG adenylosuccinate lyase or adenylosuccinate synthetase. MILLER,1967 A specific enzyme defect in gout associated with overproduction of uric acid. Proc. Natl. Acad. Sci. USA 57: 1735- We are grateful to Drs. R. A. WOODS,M. KONRAD,S. ELLEDGE,M. 1739. DEELEYand M. R. CHEVALLJERfor sending strains and plasmids. The KIPPERT,F., 1995 A rapid permeabilization procedure for accurate authors also thank Dr. D. PEI.I.MANfor critical reading of the manu- quantitative determination of pgalactosidase activity in yeast script and C. BOHNfor technical help in the ASL assays. This work cells. FEMS Microbiol. Lett. 128: 201-206. was supported by grants from Fondation pour la Recherche Medicale, KONRAU,M., 1992 Cloning and expression of the essential gene for Conseil Regional d’Aquitaine and CNRS (URA1354 and UPR9026). guanylate kinase from yeast. J. Biol. Chem. 267: 25652-25655. M.L.G. was supported by a NATO fellowship. LEE,T. T., C. WORBY,J. E. DIXON, andR. C. COLMAN,1997 ldentifica- tion of His141 in the active site of Bacillus subtilis adenylosucci- nate lyase by affinity labeling with ~(4bromo-2,3dioxobutyl)thi- LITERATURE CITED oadenosine 5’-monophosphate. J. Biol. Chem. 272: 458-465. LESCH,M., and W. L. NYHAN,1964 A familial disorder of uric acid AIANI, E., and N. KLECKNER, 1987 A newtype of fusion analysis metabolism and central nervous system function. Am. J. Med. applicable to many organisms: protein fusions of the URA3 gene 36: 561-570. of yeast. Genetics 117: 5-12. LOMAX,C. A,, and R. A. WOODS,1970 Prototrophic regulatory mu- ALFONZO, J. D., A. SAHOTA,M. C. DEELEY,P. RANJEKAR and M. W. tants of adenylosuccinate synthetase in yeast. Nature 229: 116. TAYLOR,1995 Cloning and characterization of theadenine MOMOSE,H., H.NISHIKAWA, and I. SHIIO,1966 Regulation of purine phosphoribosyl transferase-encoding gene (APTI) from Saccharo- nucleotide synthesis in Bacillus subtilis. 1. Enzyme repression by myces cereuisiae. Gene 161: 81-85. purine derivatives. J. Biochem. 187: 373-379. ARNDT, K. T., SIXES, C. and G. R. FINK,1987 Multiple global regula- MYERS,A. M., A. TZAGOLOFF,D. M. JSINNEY,and C. J. LUSTY,1986 tors control HIS4 transcription in yeast. Science 237: 874-880. Yeast shuttle and integrative vectors with multiple cloning sites BERBEN,G., J. DUMONT,V. GILLIQUET,P. A. BOLLEand F. HILGER, suitable for construction of lacZ fusions. Gene 45: 299-310. 1991 The YDp plasmids: a uniform set of vectors bearing versa- NEUHARD,J., and P. NYGAARD,1987 Biosynthesis and conversion of tile gene disruption cassettes for Saccharomyces cereuisiae. Yeast 7: nucleotides, purines and pyrimidines, pp. 445-473 in Escherichia 475-477. coli and Salmonella typhimurium Cellular and Molecular Biology, ed- BOEKE,J. D., F. LACROUTE, and G. R. FINK,1984 A positive selection ited by F. C. NEIHARDT,J. L. INGRAHAM,K. B. LOW, B. MAGASANIK, for mutants lacking orotidine-5’-phosphate decarboxylase activ- M. SCHAECHTERand H. E. UMBARGER.American Society for Mi- ity in yeast: 5fluoro-orotic acid resistance. Mol. Gen. Genet. 197: 345-346. crobiology, Washington, DC. BRAIJS,G., H-U. MOSCH,K. VOGEI.,A. HINNEN, andR. HOTTER, 1989 ROLFES,R. J., and H. ZALKIN,1990 Purification of the Escherichia coli Interpathway regulation of the TRP4 gene of yeast. EMBO J. 8: purine regulon repressor and identification of corepressors. J. 939-945. Bacteriol. 172: 5637-5642. BRAZAS,R. M., and D. J. STILLMAN,1993 Identification and purifica- ROSENBLOOM,F. M., J. F. HENDERSON,1. C. CAI.DWELL,W. N. KELLEY, tion of a protein that binds DNA cooperatively with the yeast and J. E. SEEGMILLER,1968 Biochemical bases of accelerated SWI5 protein. Mol. Cell. Biol. 13: 5524-5537. purine biosynthesis de novo in human fibroblasts lacking hypo- BURRIDGE,B. W., R. A. WOODS,and J. F. HENDERSON,1977 Purine xanthine-guanine phosphoribosyltransferase. J. Biol. Chem. 243 metabolism in Saccharomyces cereuisiae. Can. J. Biochem. 55: 935- 1166-1173. 941. RUBY,S. W., J. W. SZOSTAK,and A. W. MURRAY,1983 Cloning regu- BURRIDGE,B. W., R. A. WOODS,and J. F. HENDERSON,1978 Altered lated yeast genes from a pool of lacZfusions. Methods Enzymol. pnrine metabolism in regulatory mutants of Saccharomyces cereuis- 101: 235-269. me. J. Gen. Microbiol. 107: 403-406. SEEGMILLER,J. E., F. M. ROSENBLOOM,and W. N. KELLEY, 1967 An Mutants constitutive for ADE genes 397
enzyme defect associated with a sex linked human neurological VOGEL,K., W. HOE, and A. HINNEN,1989 The two positively acting disorder and excessive purine synthesis. Science 155 1682-1684. regulatory proteins PH02 and PH04 physically interact with SHERMAN,F., G. R. FINK,and J. B. HICKS, 1986 Methodr in Yeast Genet- PH05upstream activation regions. Mol. Cell. Biol.9 2050-2057. ics, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. WENG,M., P. L. NAGY, and H. ZALKIN, 1995 Identification of the SIKORSKI,R. S., and P. HIETER,1989 A system of shuttle vectors and Bacillus subtilis pur operon repressor. Proc. Natl. Acad. Sci. USA yeast host strains designed for efficient manipulation of DNA in 92: 7455-7459. Saccharomyces cereuisiae. Genetics 122: 19-27. WOODS,R. A,, D. G. ROBERTS,T. FRIEDMAN,D.JOLLY, and D. RLPULA, SPRINGER,C., M. KI~NZLER,T., BALMELLI,and G. H. BRAUS,1996 1983 Hyp0xanthine:guanine phosphoribosyltransferase mu- Amino acid and adenine cross-pathway regulation act throught tants in Saccharomyces cerevisiae. Mol. Gen. Genet. 191: 407-412. the same 5’-TGACTC3’ motif in the yeast HIS7 promoter. J. WOODS,R. A., D. G. ROBERTS,D. S. STEIN,and D. FILPUIA,1984 Ade- Biol. Chem. 271: 29637-29643. nine phosphoribosyltransferase mutants in Saccharomyces cerevis- STONE,R. L., J. AMI, B. A. BARSHOP,J. JAEKEN,G. VAN DEN BERGHE iae. J. Gen. Microbiol. 130 2629-2637. et al., 1992 A mutation in adenylosuccinate lyase associated with WOODWARD, D. O., 1978 Adenylosuccinate AMP-lyase (Neurospma mental retardation and autistic features. Nature Genetics 1: 59- crassa). Methods Enzymol. 51: 202-207. 63. WOODWARD,D. O., C. W. H. PARTRIDGE,andN. H. GILES,1958 Com- STOTZ,A,, and P. LINDER,1990 The ADE2 gene from Saccharomyces plementation at the AD4 locus in Neurospora crmsa. Genetics 44 cerevisiae: sequence and new vectors. Gene 95: 91-98. 1237-1244. TICE-BALDWIN,K., G. R. FINK,and K. T. ARNDT,1989 BAS1 has a YANISCH-PERRON,C., J. VIEIRA,and J. MESSING,1985 Improved MI3 Myb motif and activates HIS4 transcription only in combination phage cloning vectors and host strains: nucleotide sequences of with BAS2. Science 246: 931-935. the M13mp18 and pUC19 vectors. Gene 33 103-119. ULLMAN,B., M. A. WORMSTED,M. B. COHEN,and D. W. MARTIN,JR., ZOU, Z., and S.J. ELLEDGE,1992 Isolation of &mutants constitutive 1982 Purine oversecretion in cultured murine lymphoma cells for transcription of the DNA damage inducible gene RNR3 in deficient in adenylosuccinate synthetase: genetic model for in- Saccharomyces cerevisiae. Genetics 131: 851 -866. herited hyperuricemia and gout. Proc. Natl. Acad. Sci. USA 79 5127-5131. Communicating editor: F. WINSTON