Dominant and Recessive Suppressors That Restore Glucose Transportin a Yeast Snf3 Mutant

Dominant and Recessive Suppressors That Restore Glucose Transportin a Yeast snf3 Mutant

Linda Marshall-Cadson*, Lenore Neigeborn*, David Coons?, Linda Bisson? and Marian Carlson* *Department of Genetics and Development and Institute of Cancer Research, Columbia University College of Physicians and Surgeons, New York, New York 10032, and tDepartment of Viticulture and Enology, University of Calijornia, Davis, Calvornia 95616 Manuscript received October 29, 1990 Accepted for publication March 23, 1991

ABSTRACT The SNF3 gene of Saccharomycescereuisiae encodes a high-affinity glucose transporter that is homologous to mammalian glucose transporters. To identify genes that are functionally related to SNF3, we selected for suppressorsthat remedy the growth defect ofsnf3 mutants on low concentrations of glucose or fructose. We recovered 38 recessive mutations that fall into a single complementation group, designated rgtl (restores glucose transport).The rgtl mutations suppressa snf3 null mutation and are not linked to snf3. A naturally occurring rgtl allele was identified in a laboratory strain. We alsoselected five dominant suppressors. At least two are tightly linked to one another and are designated RGT2. The RGT2 locus was mapped 38 cM from SNF3 on chromosome N.Kinetic analysis of glucose uptake showedthat the rgtl and RGT2 suppressors restore glucose-repressible high-affinity glucosetransport in a snf3 mutant. These mutations identify genes that may regulate or encode additional glucose transport proteins.

HE transport of glucose into eukaryotic cells is gene, was first identified by isolating mutants defec- T mediated by specific carrier proteins. The genes tive in growth on sucrose or raffinose (NEIGEBORN encoding a variety of glucosetransporters from mam- and CARLSON1984). These sugars are hydrolyzed malian cells have been sequenced, and many of the extracellularly, and the resulting glucose and/or fruc- proteins are closely related, containing 12 putative tose,released at low concentration, must be trans- membrane-spanning regions and conserved sequence ported into the cell. The mutants are also defectivein motifs (MUECKLERet al. 1985; BIRNBAUM,HASPEL growth on medium containing glucose at low concen- and ROSEN1986; THORENSet al. 1988; FUKUMOTOet tration (NEIGEBORNand CARLSON1984; NEIGEBORN al. 1988; BIRNBAUM1989; CHARRONet al. 1989; et al. 1986). Kineticanalysis showed thatthe snf3 JAMES,STRUBE and MUECKLER1989). Different mam- mutants lack high-affinity glucose uptake, but exhibit malian transport systems are subject to different reg- normal low-affinity uptake (BISON et al. 1987). The ulation; for example, some transporters are regulated defect in high-affinity transport accounts for the in response to insulin (CHARRONet al. 1989; JAMES, growth phenotypes of snf3 mutants. STRUBEand MUECKLER1989). The SNF3 gene was cloned (NEIGEBORNet al. 1986) We have studied glucose transport in Saccharomyces and encodes a 97-kilodalton protein, containing 12 cerevisiae with the view that genetic analysis should putative membrane-spanning regions, that is homol- prove useful in studying a complex, highly regulated ogous to mammalian glucose transporters (CELENZA, process that is essential to all eukaryotic cells. Like MARSHALL-CARLSONand CARLSON1988). The SNF3 higher organisms, S. cerevisiae also appears to express protein differs from the mammalian transporters in multiple, differently regulated glucose transport sys- having additional sequences at the N and C termini. tems. Kinetic analysis of glucose uptake in yeast has The large C-terminal extension (303 amino acids) revealed at least two components, a high affinity com- contributes to, but is not essential for, SNF3 function ponent (K, - 1-2 mM) that is dependent on the pres- (MARSHALL-CARLSONet al. 1990). SNF3 is alsoho- ence of a cognate hexose kinase and a low affinity mologous to other yeast and bacterial sugar trans- component (K, -20-50 mM) (BISON and FRAENKEL porters (MAIDENet al. 1987; SZKUTNICKAet al. 1989; 1983; LANG and CIRILLO1987). Bothsystems also CHENGand MICHELS1989; NEHLIN,CARLBERG and transport fructose. The two components are differ- RONNE1989). The SNF3 product is associated with ently regulated: the high-affinity system is repressed membranes and is localized at the cell surface (CE- by glucose, and the low-affinity system is expressed LENZA, MARSHALL-CARLSONand CARLSON 1988). constitutively (BISSONand FRAENKEL1984). Taken together, these data indicate that SNF3 en- A glucose transporter gene of S. cerevisiae, the SNF3 codes a high-affinity glucosetransporter.

Genetics 128: 505-512 uuly, 1991) 506 Marshall-Carlson L. et al.

Previousstudies have identified additional genes TABLE 1 that appear functionally related to SNF3. A selection List of S. cerarisiae strains for multicopy plasmids that complement the growth defect of a snf3 mutant yielded at least five different Strain" Genotype genes (BISSONet al. 1987). One of these, named HXT2, MCY657 MATa snf3-72 ura3-52 lys2-801SUC2 encodes a protein that resembles other glucose trans- (SUC7?) porters, and mutations in HXT2 affect high-affinity MCY659 MATa snf3-72 ura3-52 lys2-801 ade2-101 hexose transport, although not as severely as muta- suc2 (SUC7?) tions in SNF3 (KRUCKEBERGand BISON 1990). Low MCY7 14 MATa snf3-217 ura3-52SUC2 MCY 1093 MATa ura3-52 lys2-801 his4-539SUC2 stringencyblot hybridization analysis of genomic MCY 1094 MATa ade2-I01 ura3-52SUC2 DNA suggested thatthe yeast genome contains a MCY 1408 MATa snf3-A4::HIS3 his3-A200 ura3-52 family of sequences homologous to HXT2, probably lys2-801 ade2-101 SUC2 including additional glucose transporter genes.At MCY 1409 MATa snf3-A4::HIS3 his3-A200 ura3-52 least one additional transporter gene must exist, as 1~~2-801SUC2 MCY1410 MATa snf3-A4::HIS3 his3-A200 ade2-101 neither SNF3 nor HXT2 is responsible for low-affinity 1~~2-801SUC2 transport. MCY1471 MATa rgtl-1 ade2-I01SUC2 In this study, we have used a different approach to MCY1516 MATa rgtl-1 snf3-A4::HIS3 ura3-52 identify genes that are functionally related to SNF3. ade2-I01 (his3-A200?)SUC2 We sought to identifygenes that could mutate to MCY 1520 MATa rgtl-1 snf3-A4::HIS3 ade2-I01 (his3-A200?)SUC2 suppress the transport defect caused by a snf3 muta- MCY1710 MATa RGT2-I snf3-A4::HIS3 his3-A200 tion. We therefore selected for suppressors that re- ura3-52 lys2-801 SUC2 store growth of mutants on raffinose, which requires MCY1711 MAT@ RGT2-1snf?-A4::HIS3 his3-A200 high-affinity fructose uptake. We anticipated that this lys2-801 ade2-101 ura3-52 SUC2 selection could yield mutations that alter other trans- MCY1713 MATa RGT2-2snf3-A4::HIS3 his3-A200 ura3-52 lys2-801 ade2-101 SUC2 porters so that they can bind and transport fructose MCY1714 MATa Rgt#3 snf3-A4::HIS3 his3-A200 with high affinity. Alternatively, the selection could ura3-52 lys2-801 ade2-I01SUC2 yield mutations that increaseexpression of other MCY1717 MATa Rgt#4 snf3-A4::HlS3 his3-A200 transporters or allow expression of normally cryptic ura3-52 lys2-801 ade2-101 SUC2 transporters. We describe here the isolation of two MCY1719 MATa Rgt#5 snf3-A4::HIS3 his3-A200 ura3-52 lys2-801 ade2-I01SUC2 classes of suppressors that restore high-affinity uptake MCY 1807 MATa ccsl snf3-A4::HIS3 (his3-A200?) in snf3 mutants: recessive rgtl mutations and domi- ura3-52 SUC2 nant RGT2 mutations. MCY2035 MATa rgtl-2snf3-72 lys2-801 his4-539 ura3-52::pLSI 1 SUC2 MCY2 157 MATa RGT2-1 his3-A200 lys2-801 SUC2 MATERIALS AND METHODS MCY2 160 MATa cdc9 snf3-A4::HIS3 (his3-A200?) suc2 Strains and general genetic methods: Strains of S. cere- MCY2 162 MATa leu2-3 SUF25-1 ura3-52 his4- uisiae used in this study are listed in Table 1. pLS 1 1carries 519R SUC2 the URA3 gene and a SUCZ-LEU2-lacZ fusion (SAROKINand MCY2166 MATa cdc9 snf3-A4::HIS3 (his3-A200?) CARLSON1985) that is irrelevant to this study. Genetic lys2-801 ura3-52 SUC2 analysis was carried out by standard methods (SHERMAN, MCRY 168 MATa snf3-72 lys2-801 his4-539 ura3- FINKand LAWRENCE1978). Growth phenotypes were de- 52::PLSl I suc2 termined by spotting cell suspensionsonto plates usinga 32- LBY415 MATa hxt2::LEU2 snf3-A4::HIS3 hid- point inoculator and incubating the plates at 30" under A200 ura3-52 lys2-801 ade2-101 trpl- anaerobic conditions in a GasPak disposable anaerobic sys- A43 leu2-AI SUC2 tem (BBL). Growth ofsingle colonies was examined as 1629' MATa leu2-3 described in the legend to Figure 1. Unless otherwise noted, 1695b MATa leu2-3 his4-519R1 ura3-52 plates contained rich medium (YEP) and 2% of the indicated SUF25-1 carbon source. Glucose uptake assays: Cells were grown in yeast nitro- a MCY strains are from the CARLSONlaboratory, and the LBY gen base (0.67%) containing casamino acids (0.2%), auxo- strain is from the BISON laboratory. * Obtained from MICHAELCULBERTSON. trophic requirements, and the indicated carbon source. Cul- tures were harvested in early or mid log phase, and glucose uptake assays were performed by measuring uptake of D- antimycin A (1 rg/ml). Cells were then exposed to 100 J/ [U-'4C]glucose(New England Nuclear) over the concentra- m' of UV radiation. In control experiments, 30% of the tion range of 0.2 to 200 mM, as described previously cells remained viable. The plates were incubated at 30" for (KRUCKEBERGand BISON 1990). Each strain was assayed at 5 days. Revertants arose at frequencies of 1 to 5 X least twice after growth under the specified conditions. Revertants derived from three single coloniesof each strain Isolation of revertants of haploid sn.mutants: Strains (10 from MCY657,8 from MCY659,7 from MCY714, and MCY657,MCY659, MCY714 and MCRY168 were sub- 13 from MCRY 168) were colony purified and retested. jected to UV mutagenesis. Single colonies were suspended Complementationanalysis: Mutations were tested for in water and spread on a YEP-2% raffinose plate containing dominance by crossing each revertant to asnf3 null mutant. in YeastGlucose Transport in 507

To test for complementation, we constructed snf3/snf3 dip- port the low amounts of fructose released by extra- loids that were heterozygous for the suppressor mutations cellularhydrolysis of the trisaccharide. To test in pairwise combinations. Diploids were usually isolated by whether the mutation also restores efficient uti- prototrophic selection, and when no selection was possible, rgtl single colonies were isolated and tested for mating or spor- lizationof glucose at low concentration in a snf3 ulation. Diploids were scored for anaerobic growth on raf- mutant, rgtl-1 snf3-A4::HIS3and control strains were finose after 24 and 48 h. streaked for single colonies on rich medium contain- Identification of rgtl SNF3 strains: Segregants of gen- ing either 0.1% or 2% glucose. All strains produced otype rgtl SNF3 were identified in nonparental ditype te- coloniesof the samesize on 2% glucose. On low trads from crosses of rgtl snj3 strains to wild type. The presence of the rgtl mutation was verified by crossing the glucose, however, the snf3-A4::HZS3 mutant formed putative rgtl SNF3 strain to a snf3 mutant and demonstrat- very small coloniescompared to thewild type, whereas ing segregation of the suppressor in tetrad analysis. the rgtl-1 snf3-A4::HZS3strain grew as well asthe wild Isolation of revertants carrying dominant suppressors: type(Fig. 1). Thus, the growth defect of the snf3 Six single colonies derived from the cross of MCY 1408 X MCY1409 were used to inoculate YEP-glucose liquid me- mutant on low glucose was clearly remedied by rgtZ. dium. After growth overnight, 0.3 ml of each culture (3 X Diploidsof genotype snf3lsnf3 rgtllRGT1 formed 10’ cells) was spread onto YEP-raffinose medium, and plates small colonieson low glucose, confirming that the rgtl were incubated anaerobically for 96 hr. Approximately 10- suppressor is recessive with respect to this phenotype 30 colonies grew on each plate. Six independent revertants, (data not shown). In a wild-type background, one from each plate, were colony purified twice and (SNF3) retested. rgtl caused no obvious phenotype. rgtl is unlinked to snf3: To determine whether RESULTS rgtl is linked to snf3, two of the revertants (snf? rgtl) werecrossed to wild type. Tetrad analysisof the Isolation of revertants of snj3 mutants: The raff- diploids yielded frequent raffinose nonfermenting se- nose-nonfermenting phenotype of mutants is snf3 gregants, presumably of snf3 RGTl genotype. Two caused by the defect in high-affinity glucose/fructose additional crosses heterozygous for snf3 and rgtl also uptake: the mutants are unable to transport the fruc- yielded segregations of 4+:0-, 3+: 1- and 2+:2- for tose that is released at low concentration by the extra- raffinose utilization in ratios approximating 1:4: 1. cellular hydrolysis of raffinose. We selected for sup- The ratio for the combined data from these crosses pressors that restore growth of snf3 mutants on raffi- was 6:19:4. Thus, rgtl is not tightly linked to snf3. nose. Four haploid snf3 mutant strains were subjected rgtl suppresses a snj3 deletion muation: To test to UV mutagenesis, and 38 revertants able to utilize whether an rgtl allele suppressesa snf3 null mutation, raffinosewere selected, as described in MATERIALS the revertant of MCRY168 carrying rgtl-1 (snf3-72 AND METHODS. The strains carried either the snf3-72 pLSl1 carries was crossed or snf3-217 allele. The snf3-72 allelehas been se- rgtl-1 ura3::pLSll; URA3) quenced, and the mutation changes Gly-153 to Arg to MCY 1408 (snf3-A4::HZS3 ura3).Tetrad analysis of the resulting diploid showed 2+:2- segregations for (MARSHALL-CARLSONet al. 1990). Dominancetests: To test for dominance of the raffinose utilization in seven tetrads. Because rgtl is mutation responsible for the revertant phenotype, unlinked to snf3, these data indicate that rgtl-1 sup- each revertant was crossed to a strain carrying the presses a snf3 null mutation. These data also confirm snf3-A4::HZS3 null allele(NEIGEBORN et al. 1986). The that rgtl-1 behaves asa lesion ina single nucleargene. resulting diploids were in each case unable to grow The segregation pattern for rgtl and the centromere- on raffinose anaerobically, indicating that all of the linked marker ura3 (5 tetratype and 2 nonparental suppressor mutations are recessive. ditype asci) did not indicate tight linkage to a centrom- Complementationanalysis: Revertants derived ere for rgtl. from MCY657,MCY714 and MCRY 168 were Further evidence that rgtl-2 suppresses snf3- crossed to MCYl520 (snf3-A4::HZS3 rgtl-1),which Al::HIS3 came from analysis of the cross MCY 1408 was derived from MCRY 168, and revertants derived (snf3-A4::HZS3)by MCY 1471 (rgtl-1SNF3). Segre- from MCY659were crossed to MCY2035 (snf3-72 gations of 4+:0-, 3+:1- and 2+:2- for raffinose rgtl-2)which was derived from MCRY 168. All of the utilization were observed in the ratio 1:4: 1, and the resulting diploids were able to ferment raffinose, in- presence of the snf3-A4::HZS3 or SNF3 allele in Rap dicating that the mutations fall into a single comple- segregants was determined by complementation. mentation group. Additional tests of other pairwise Strains MCY 15 16 and MCY 1520 (snf3-A4::HZS3rgtl- combinations also revealed no complementation. The 1) were recovered from this cross. complementation group was designated rgtl for re- A suppressor of snj3 present in some laboratory stores glucose transport. strains is an rgtl allele: The strains routinely used in rgtl restores growth of sn. mutants on low glu- this laboratory are derived from the S288C genetic cose: The revertants wereselected for growth on background. During analysis of the cross of the raffinose, which normally requires the ability to trans- S288C-derived strain MCY2 160 (snf3-A4::HZS?) by 508 L. Marshall-Carlson et al. MCY2 162 (SNF3), which was derived from strains 1629 and 1695 (obtained from M. CULBERTSON,Uni- versity of Wisconsin), we observed a suppressor of snf3 segregating. Five segregants carrying both snf3- A4::HIS3 andthe suppressor were identified. The suppressor was shown to be recessive by crossing each segregant to asnf3-A4::HIS3 strain; the diploids were raffinose nonfermenters. To test the suppressor for complementation of rgtl, the five segregants were then crossed to snf3-A4::HIS3 rgtl-Z strains (MCY 15 16 or MCY 1520). All five diploids grew on raffinose, indicating that the suppressor fails to complement rgtl. Tetrad analysis of one of the diploids yielded no raffinose-nonfermenting segregants in seven four-spored tetrads and six triads, confirming that the suppressor is linked to rgtl. Thus, the suppressor is a naturally occurring rgtl allele. rgtl is unlinked to hxt2: The HXT2 gene was FIGURE1.-Growth phenotypes of snf? rgtl and snf? RGT2 identified asa multicopy suppressor of the snf3 mutant strains. Strains were streaked on YEP containing 2% glucose (left panels) or 0.1%glucose (right panels). (A) Plates were incubated defect in high-affinity glucosetransport (BISON et al. aerobically at 30" for 48 hr and then photographed. Relevant 1987) and encodes a protein homologous to glucose genotypes: (a) wild type; (b) snf3-A4::HIS3; (c) rgtl-I; (d) rgfl-1 snf?- transporters (KRUCKEBERGand BISON 1990). An hxt2 A4::HIS?. The strain shown in panel (c) is MCY I47 1, and the others null mutation reduces high-affinity glucose transport are segregants of cross MCY 147 1 X MCY 1408. (B) Plates were under derepressing conditions, but not as severely as incubated anaerobically at 30" for 72 hr and then photographed. Strains and relevant genotypes: (a) MCY1093 (wild type); (b) a snf3 mutation, and does not cause a strong growth MCY 1409 (snJ?-A4::HIS?); (c) MCY 17 13 (RGT2-2 snf?-A4::HIS?); defect on medium containing either high or low glu- (d) MCY 17 1 1 (RGT2-I snf?-Al::HIS?). cose (KRUCKEBERCand BISON 1990). The hxt2 snf3 double null mutants resemble snf3 mutants in pheno- raffinose fermenting phenotype segregated 2+:2- in type. all seven tetrads tested, and Rap segregants grew as To determine whether the rgtl suppressors are well as the wild type on YEP-raffinose or YEP-0.1% alleles of HXT2, we carried out tetrad analysis of the glucose under anaerobic conditions (Figure 1). The diploid MCY1516 (snf3-A4::HIS3 rgtl-I) X LBY415 sixth revertant showed a weak phenotype and was not (snf3-A4::HIS3 hxt2::LEU2 leu2). Ten tetrads were characterized further. recovered that showed 2+:2- segregations for leucine The dominance of these mutations was confirmed dependence, corresponding to thenonparental ditype by crossing a Raf snf3-A4::HIS3 segregant from each configuration for the leu2 and hxt2::LEUZ markers. of the five revertants to asnf3-A4::HZS3 mutant. The Tetrads of this class were easily recovered because resulting diploids each showed a Raft phenotype. In leu2 and hxt2 are linked to different centromeres (D. addition, oneof the diploids (MCY 14 10 X MCY 1 7 10) COONSand L. BISON, unpublished results). The hxt2 was sporulated, andtetrad analysisagain showed mutation did not affect the ability of rgtl to suppress 2+:2- segregations for growth on raffinose in seven snf3 because 2+:2- segregations for growth on low tetrads. glucose were observed. The ten tetrads in which the Linkage of the dominant suppressors to one an- segregation of hxt2::LEU2 could be inferred included other: To determine whether the five dominant mu- one parental ditype, one nonparental ditype and eight tations are linked to one another,segregant a obtained tetratypes with respect to rgtl and hxt2. These data from one of the revertants (MCY 1710) was crossed to indicate that rgtl and hxt2 are not tightly linked. segregants from eachof the other four revertants Selection for dominant suppressors of snt: Be- (MCY1713, MCY1714, MCY1717, MCY1719). Te- cause all 38 suppressors of snf3 selected in haploid trad analysis of the resulting diploids showed 4+:0- strains were recessive alleles ofa single locus, we next segregations for'ability to utilize raffinose (46 tetrads carried out a selection for suppressors in a diploid in for MCY 17 10 X MCY 17 13 and seven tetrads for each an attempt to recover dominant mutations. Wese- of the other threediploids). Thus, the suppressors in lected spontaneous raffinose-fermenting revertants of MCYl7lO and MCY1713 are tightly linked to one the diploid MCY 1408 X MCY 1409, which is homo- another and are probably alleles of the same locus. zygous for snf3-A4::HIS3 (see MATERIALS AND METH- These mutations were designated RGT2-1 and RGT2- ODS). Six independent revertants were sporulated and 2, respectively. The other suppressors may alsobe subjected to tetrad analysis. For five revertants, the alleles of this locus. Glucose Transport in Yeast 509

TABLE 2 Linkage data

No. of tetrads" Map distance Cross Parents Cross Gene pairNPD PD T (cWb

LC36 MCY1094 X MCY1710 RGTZ-I-snf3 11 1 14 LC76 MCY1408 X MCY2157 RGTZ-I-snf3 -2 -0 -5 13 1 19 38

LC116 MCY1711 LC116 X MCY2166 RGTZ-l-cdc9 11 180 6

LC66 MCYl7lO X MCY1807 RGTZ-I-CCS~ 18 0 3 LC70 MCY 1408 X LC66.10A RGTZ-1-ccsl 10 0 4 LC7 1 MCY 1409 X LC66.12A RGTZ-1-ccsl 5 0 1 LC73 MCY1408 X LC66.19A RGTZ-I-CCSI -6 -0 - 0 39 0 8 9 PD, parental ditype; NPD, nonparental ditype; T, tetratype. Data were obtained from tetrads with four viable spores and showing 2:2 segregation for the markers. Genetic map distances in centimorpans" were calculated from the tetrad data by the equation of PERKINS(1949): distance = 100(T + 6NPD)/2(PD +'NPD + T).

Genetic mapping of RGT2 near SNF3 on chro- yielded tetratype tetrads containing both a large Raf mosome N: Standard meiotic linkage analysis ofthe and asmall Raf recombinant, thereby confirming the cross MCY 17 10 X MCY 1094 revealed a genetic dis- genotypes of LC66.10A and LC66.12A (Table 2). tance of 38 cM between RGT2-1 and snf3 (Table 2). Thus, RGT2-1and ccsl are clearly not allelic, and the Previous studies established a gene order of cdc9-snf3- calculated genetic distance is 9 cM. From the genetic SUF25 with distances of 32 cM for cdc9-snf3 and 11 distances, the likely gene order is RGT2-ccsl-snf3. cM for snf3SUF25 (MARSHALL-CARLSONet al. 1990). Assaysof invertase activityin LC66.10A and To determine the location of RGT2 relative to cdc9, LC66.12A showed that RGT2-1does not suppress the the cross MCY 17 1 1X MCY2166 was analyzed (Table invertase constitutivity resulting from the combina- 2). The calculated map distance is 18 cM, indicating tion of snf3 and ccsl (data not shown). Also,assays of that RGT2 and cdc9 lie on the same side of snf3. The RGT2 snf3 segregants showed that RGT2-1 does not likely gene order is RGT2-cdc9-snf3-SUF25;however, affect regulation of invertase expression (not shown). this could not be easily confirmed by a three-point These linkage data suggest that ccsl maps close to cross due to problems inherent in scoring these cdc9,and both mutations cause growth defects at 37 O. markers. We therefore tested the two mutations for comple- Analysisof the cross MCY 1710 (snf3 RGT2) X mentation. The heterozygous diploid grew normally MCY 1516 (snf3 rgtl) yielded parental ditype, tetra- at 37", suggesting that the two mutations comple- type and nonparental ditype tetrads in the ratio ment. 3: 12:3, confirming that RGT2 and rgtl are not linked. rgtl and RGT2 mutations restore glucose-repres- RGT2 is not allelic to ccsl: Previous workidentified sible high-affinity glucose transport in a snj3 mu- the recessive ccsl mutation, which is tightly linked to tant: The Raf phenotype of snf3 mutants is caused snf3 (genetic distance 19 cM) (MARSHALL-CARLSONet by the defect in high-affinity glucose/fructoseuptake, al. 1990). This mutation causes poor growth on glu- and suppression of this phenotype by the rgtl and cose and, in conjunction with snf3-39 or snf3-A4::HIS3 RGT2 mutations could most easily be accounted for alleles,causes constitutive invertase expression. To by the restoration of hexose transport. We therefore test whether the RGT2 mutations might be dominant examined the kinetics of glucose transport in snf3- alleles of the same gene, the strain MCY 1710 (RGT2- A4::HIS3 rgtl-1 and snf3-A4::HIS3 RGT2-1 strains. 1 snf3) was crossed to MCY 1807 (ccsl snf3) in cross Both suppressors restored high-affinity uptake in de- LC66 (Table 2). Tetrad analysis yielded 3 tetratype repressed cells. RGT2-l restored wild-type levels of and 18 parental ditype tetrads, as judged by the seg- glucose transport activity, and rgtl-1 caused wild-type regation of the raffinose fermentation phenotype and or slightlyelevated levels of transport (Figure 2). the small spore clone size that is characteristic of ccsl Similar results were obtained for snf3 mutant strains mutants. To confirm that the threesmall, Raf recom- carrying the RGT2-2and rgtl-2 alleles (MCY 17 13 and binants (LC66.1OA, LC66.12A and LC66.19A) in the MCY2035; data not shown). tetratype asci have the genotype RGT2 ccsl snf3, each High-affinity glucose transport in wild-type (SNF3) was crossed to a snf3 mutant. Two of the threediploids strains is regulated by glucose repression (BISSONand 510 L. Marshall-Carlson et al.

50 A B 40

30 V 20

60 C

40 V 30

FIGURE2.-&die-Hofstee plots of glucose uptake in snf3 rgtl and snf3 RGT2 strains. Velocity is expressed as nanomoles of glucose per minute per milligram (dry weight); V/[S]is expressed as velocity per millimolar concentration. Cells were grown to early log phase in medium containing glycerol (2%) and lactate (2%), and assays were carried out as described in MATERIALS AND METHODS. Strains: (A) MCY1093 (wild type), filled circles, and MCY 1409 (snf3-A4::HIS3),open circles; (B) MCY 1710 (snf3-A4::HIS3 RGT2-I);(C) MCY 1516 (snf?-A4::HIS3 rgtl-I).

FRAENKEL1984), as is expression of the SNF3 gene pressing conditions, and high-affinity transport was (NEIGEBORNet al. 1986;CELENZA, MARSHALL-CARL- detected(Figure 3). Assaysof MCY 171 3 and SON and CARLSON1988). To determinewhether the MCY2035 also showed glucose-repressible high-affin- high-affinity uptake that is restored in the rgtl-1 snf3- ity transport (data not shown). A4::HIS3 and RGT2-I snf3-A4::HIS3 strains is glucose- repressible,transport was inassayed glucose-grown DISCUSSION cultures. The high-affinity transportexpressed in these strains was glucose-repressible (Figure 3). As a By selecting forraffinose-fermenting revertants of control,the same cultureswere also shifted to dere- snf3 mutants, we isolated two classesof suppressors Glucose Transport in Yeast 511

50 - restore that high-affinity glucose transport in a snf3 null mutant. The recovery of such revertants supports A previousevidence thatthe yeast genome includes multiplegenes thatare functionally related to or homologous to the SNF3 glucose transportergene (BIsson et a,!. 1987; KRUCKEBERGand BISON 1990). Presumably, the suppressor mutations allow expres- V of productsion the or alter of one of these genes. The 38 mutations isolated in haploid strains are all recessive and define a single complementation group, rgtl. The recessiveness of the alleles suggests that the mutations cause loss of function. We can imagine at least two mechanisms by which an rgtl mutation could a.".l-.--l restore the expression of high-affinity transport. First, 0 5 10 " the wild-type RGTl geneproduct could repress vm1 expressionencodinggene of a a high-affinity transporter that is functionally homologous to the SNF3 transporter. This gene might normally be expressed only at low levels or not at all. An rgtl mutation would then release expression of this gene to levels sufficient B to confer anearly wild-type capability for high-affinity transport. A second possibility is that the RGTl product acts to inhibit the function of a high-affinity transporter thatis expressed but does not contributeto the high-affinity component of glucose/fructose uptakein V wild-type cells. An apparently naturally occurring rgtl suppressor allele was identified in a laboratory strainof S. cerevisiae with a genetic background different from that of S288C. Interestingly, a cross of a snf3 mutant with the wild-type strain DFY 1 (originally D585-11C from F. SHERMAN)also did not yield Mendelian segregation 0 5 10 '' forfailure to grow on low glucose, althoughthe vm segregation pattern was not consistent with the segregation of a single suppressor of snf3 (BISSON1988). Perhaps rgtl suppressors are fairly common in S. cerevisiae strains, and the RGTl allele present in the C S288C genetic background is unusual. A selection for revertants of a homozygous snf3 mutant diploid yielded five dominant suppressors at the RGT2 locus. The dominance of these mutations suggests that suppression results froma change of function. The RGT2 suppressors may elevate the V expression or activate the function of a high-affinity transporter that, in wild-type cells, contributes little to the high-affinity component of uptake. It is also possible that these dominant mutations convert low- a affinity transporter into a high-affinity transporter by increasing the affinity for hexoses or by affecting 0 5 10 . ,; interactions with hexose (BISSONkinases and FRAEN-

~ vas1 either 4% or 0.05% glucose, and incubated with aeration for 4 hr prior to assay for glucose uptake. Symbols: 0,cells resuspended in FIGUREJ.-Glucose repression of high-affinity transportin snf3 4% glucose (repressed); 0, cells resuspended in 0.05% glucose rgtl and snJ3 RGT2 strains. Eadie-Hofstee plots of glucose uptake (derepressed). Strains: (A) MCY1093 (wild type); (B) MCY1710 are shown, as for Figure 2. Glucose-repressed cultures were grown (snfjr-A4::HIS3 RGTP-I); (C) MCY 1516 (snf3-A4::HIS3 rgtI-I). to mid-lag phase in medium containing-glucose (4%). Cells were These data differ from those shown in Figure 2 because growth collected by filtration, resuspended in fresh medium containing conditions were different. 512 L. Marshall-Carlsonet al.

KEL 1983; LANGand CIRILLO1987); however, muta- homologous to the mammalian protein. Proc. Natl. Acad. Sci. tions of this type might be expected to occur rarely. USA 85 2130-2134. CHARRON,M. J., F. C. BROSIUS111, S. L. ALPERand H. F. LODISH, Molecular analysisof the RGT2 gene will be required 1989 A glucose transport protein expressed predominantly to determine whether it encodes a glucose transport in insulin-responsive tissues. Proc. Natl. Acad.Sci. USA 86: protein. RGT2 may prove to be the same as one of 2535-2539. the genes cloned as a multicopy suppressor of snf3 CHENG,Q., and C. A. MICHELS,1989 The maltose permease encoded by the MAL61 gene of Saccharomyces cerevisiae exhibits (BISON et al. 1987), but an identitywith HXT2 is both sequence and structural homology to other sugar trans- unlikely. RGT2 is linked to SNF3, whereas HXT2 is porters. Genetics 123: 477-484. centromere-linked (D. COONSand L. BISON, unpub- FUKUMOTO,H., S. SEINO,H. IMURA,Y. SEINO,R. L. EDDY,Y. lished results), and no linkage was detected between FUKUSHIMA,M. G. BYERS,T. B. SHOWSand G. I. BELL, HXT2 and SNF3 (KRUCKEBERGand BISSON1990). 1988 Sequence, tissue distribution, and chromosomal local- ization of mRNA encoding a human glucose transporter-like Both biochemicaland genetic evidence indicate that protein. Proc. Natl. Acad. Sci. USA 85: 5434-5438. in yeast, as in mammalian cells, glucose transport is a JAMES, D. E.,M. STRUBEand M. MUECKLER, 1989 Molecular complex process involving the products of multiple cloning and characterization of an insulin-regulatable glucose genes. Some of these genes encode transporters, and transporter. Nature 338 83-87. others may regulate or otherwise affect transport. KRUCKEBERG,A. L., and L. F. BISSON, 1990 The HXT2 gene of Saccharomyces cerevisiae is required for high-affinity glucose Perhaps we should expect to find functional redun- transport. Mol. Cell. Biol. 105903-5913. dancy of the genes responsible for a processso critical LANG,J. M., and V. P. CIRILLO,1987 Glucose transport in a to survival of the cell. These studies identified two kinaseless Saccharomyces cerevisiae mutant. J.Bacteriol. 169: genes that affect glucose transport, and molecular 2932-2937. MAIDEN,M. C. J., E. 0. DAVIS,S. A. BALDWIN,D. C. M. MOORE analysis should help to elucidate their roles and the and P. J. F. HENDERSON,1987 Mammalian and bacterial mechanism by which the rgtl and RGT2 suppressor sugar transport proteins are homologous. Nature 325: 641- mutations restore transport in snf3 mutants. 643. MARSHALL-CARLSON,L., J. L. CELENZA,B. C. LAURENTand M. We acknowledge one of the reviewers for the suggestion that CARLSON,1990 Mutational analysisof the SNF3 glucose the dominant suppressors may affect kinase interactions. This work transporter of Saccharomyces cerevisiue. Mol. Cell.Biol. 10: was supported by grants from the National Science Foundation 1105-1 115. (DCB-8709915) and the American Diabetes Association and MUECKLER,M., C. CARUSO,S. A. BALDWIN,M. PANICO, I. BLENCH, an American Cancer Society Faculty Research Award to M.C. H. R. MORRIS,W. J. ALLARD,G. E. LIENHARDand H. F. L.M.C. was supported by National Research Service Award post- LODISH, 1985 Sequence and structure of a human glucose doctoral fellowship F32 GM12232. transporter. Science 229 941-945. NEHLIN,J. O., M. CARLBERGand H. RONNE, 1989 Yeast galactose permease is related to yeast and mammalian glucose transport- LITERATURE CITED ers. Gene 85: 3 13-3 19. NEIGEBORN,L., and M. CARLSON,1984 Genes affecting the reg- BIRNBAUM,M. J., 1989 Identification of a novel gene encoding ulation of SUC2 gene expression by glucose repression in Sac- an insulin-responsiveglucose transporter protein.Cell 57: 305- charomyces cerevisiue. Genetics 108: 845-858. 315. NEIGEBORN,L., P. SCHWARTZBERG,R. REID and M. CARLSON, BIRNBAUM,M. J., H. C. HASPEL and 0.M. ROSEN,1986 Cloning 1986 Null mutations in the SNF3 gene of Saccharomyces cere- and characterization of a cDNA encoding the rat brain glucose visiue cause a different phenotype than do previously isolated transporter protein. Proc. Natl. Acad.Sci. USA 83: 5784- missense mutations. Mol. Cell. Biol. 6 3569-3574. 5788. PERKINS,D. D., 1949 Biochemical mutants in the smut fungus BISON,L. F., 1988Highaffinity glucose transport in Saccharo- Ustilago maydis. Genetics 34 607-626. myces cerevisiae is under general glucose repression control. J. SAROKIN,L., and M. CARLSON,1985 Upstream region of the Bacteriol. 170 4838-4845. SUCP gene confers regulated expression to a heterologous gene BISSON, L. F., and D. G. FRAENKEL,1983 Involvement of kinases in Saccharomyces cerevisiae. Mol. Cell. Biol. 5: 2521-2526. in glucoseand fructose uptake by Saccharomyces cerevisiae. Proc. SHERMAN,F., G. R. FINKand C. W. LAWRENCE, 1978Methods in Natl. Acad. Sci. USA 80 1730-1734. Yeast Genetics. Cold Spring Harbor Laboratory, Cold Spring BISON,L. F., and D. G. FRAENKEL,1984 Expression of kinase- Harbor, N.Y. dependent glucose uptake in Saccharomyces cerevisiae. J. Bacte- SZKUTNICKA,K., J. F. TSCHOPP,L. ANDREWS and V. P. CIRILLO, 1989 Sequence andstructure of the yeast galactose trans- rial. 159 1013-1017. BISSON,L. F., L. NEIGEBORN,M. CARLSONand D. G. FRAENKEL, porter. J. Bacteriol. 171: 4486-4493. THORENS,B., H. K. SARKAR,H. R. KABACKand H. F.LODISH, 1987 The SNF3 gene is required for high-affinity glucose 1988 Cloning and functional expression in bacteria of anovel transport in Saccharomyces cerevisiae. J. Bacteriol. 169 1656- glucose transporter present in liver, intestine, kidney, and P- 1662. pancreatic islet cells. Cell 55 281-290. CELENZA,J. L.,L. MARSHALL-CARLSONand M. CARLSON, 1988 The yeast SNF3 gene encodes a glucose transporter Communicating editor: E. W. JONES