GALI-GALIO Divergent Promoter Contains Negative Control Elements in Addition to Functionally Separate and Possibly Overlapping U

Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

GALI-GALIO divergent promoter region of Saccharomyces cerevisiae contains negative control elements in addition to functionally separate and possibly overlapping upstream activating sequences

Robert W. West Jr., Shiming Chen, Henry Putz, Geraldine Butler, 1 and Mary Banerjee Department of Biochemistry and Molecular Biology, SUNY Health Science Center, Syracuse, New York 13210 USA

The upstream activating sequence (UASG) of the adjacent and divergently transcribed GALl and GALIO promoters of Saccharomyces cerevisiae regulates the induction of the corresponding genes in response to the presence of galactose. We constructed chimeric yeast promoters in which a different UAS, UASc from the iso-l-cytochrome c (CYC1) gene of S. cerevisiae, was fused at different locations upstream of GALl (UASc-- GALl promoters) or GALIO (UASc- GALIO promoters) and used to monitor the activity of UASG in cells grown in the presence or absence of galactose. Though the CYC1 promoter is fully induced in yeast grown in glycerol medium, UASc-GAL chimeric promoters containing UASG were repressed as much as 400-fold (UASc-GAL1) or 1350-fold (UASc- GALIO) in this growth medium. Several distinct portions of the GAL1- GALIO divergent promoter region blocked the UASc-induced expression of the GALl and GALIO promoters, whereas others did not, suggesting that several distinct negative control elements are present that may repress transcription of GALl and GALIO in the absence of galactose. The approximate locations of these negative control elements were delimited to sites adjacent to or possibly overlapping the sites at which the positive control protein GAL4 binds in UASG. Deletion derivatives of GAL4 that fail to induce transcription from the wild-type GAL promoters but retain the DNA binding domain significantly derepressed the expression of the UASc-GAL chimeric promoters. These results, combined with those of earlier studies, suggest the possibility that GAL4 normally induces transcription of GALl and GALIO by blocking the activity of these negative control elements, in addition to stimulating transcription by a mechanism of positive control. [Key Words: Upstream activating sequence; negative control; GALl- GALIO divergent promoter; Saccharomyces cerevisiae; CYC1 gene] Received June 9, 1987; revised version accepted October 6, 1987.

Transcriptional control of eukaryotic protein-coding distal regulatory elements have been termed upstream genes requires specific regulatory sequences located ad- activating sequences, or UASs (Guarente 1984). jacent to each gene. Proximal regulatory sequences in- One well-characterized yeast UAS is UASG, which clude the TATA box, which in yeast is known to be in- controls the adjacent and divergently transcribed GALl volved in determining the precise start sites for tran- and GALI O genes (Guarente et al. 1982; Johnston and scription initiation (Chen and Struhl 1985; Hahn et al. Davis 1984; West et al. 1984; Yocum et al. 1984). UASG 1985; Nagawa and Fink 1985; McNeil and Smith 1986). is about 120 bp in size (nomenclature of Giniger et al. Distal regulatory sequences (upstream promoter ele- 1985), resides about midway between the translation ments) respond to specific physiological stimuli to con- start sites of GALl and GALIO, and is required for galac- trol the amount of transcription initiating downstream tose-mediated induction of both genes. In galactose (Gal) (Guarente et al. 1984; Giniger et al. 1985; Hope and medium, the positive control protein GAL4 binds to Struhl 1985; Arndt and Fink 1986; McKnight and Tjian four related, dyad-symmetric sequences in UASG and in- 1986; Pfeifer et al. 1987). In Saccaromyces cerevisiae the duces transcription of GALl and GALI O (Brain and Kornberg 1985; Giniger et al. 1985). In glycerol (Gly)medium, though GAL4 is produced constitutively (John- ~Permanent address: Department of Genetics, Trinity College, Dublin, Ireland. ston and Hopper 1982; Laughon and Gesteland 1982), its

1118 GENES& DEVELOPMENT 1:1118-1131 © 1987 by Cold Spring Harbor Laboratory ISSN 0890-9369/87 $1.00 Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

Negative elements in a yeast GAL promoter activity is inhibited by the negative regulatory protein elements in the GAL1-GALIO divergent promoter re- GAL80 and transcription of GALl and GALI O is pre- gion. To pursue this, we fused a 150-bp fragment con- vented (Lue et al. 1987). GAL80 probably inhibits GAL4 taining UASc from the iso-l-cytochrome c (CYC1) gene by binding to a distinct region at its carboxyl terminus of S. cerevisiae (Guarente et al. 1984; Pfeifer et al. 1987), and blocking a specific domain that is rich in acidic at the end points of 5' deletions of GALl or GALl 0, as amino acids (see Struhl 1987) and is involved in GAL4's shown in Figure 1. Our rationale was that UASc would positive control (transcription-activating) function provide the GAL promoters a basal level of expression (Johnston et al. 1987; Ma and Ptashne 1987b). independent of galactose and GAL4, allowing us to The DNA-binding domain of GAL4 is located within monitor the capacity of portions of UASG and flanking the amino-terminal 75 amino acids of the 881-amino- sequences to inhibit UASc-induced expression of GALl acid protein (Brent and Ptashne 1985; Keegan et al. 1986; or GALI O. Activation of CYC1 transcription by UASo Johnston 1987; Johnston and Dover 1987). Data derived though inhibited about fivefold by glucose, is essentially from in vivo DMS protection (Giniger et al. 1985), constitutive under the growth conditions we normally DNase I footprinting (Lohr and Hopper 1985), and pho- use to induce or repress the GAL promoters (galactose tofootprinting studies (Selleck and Majors 1987a, b) indi- vs. glycerol and lactate medium, respectively). cate that GAL4 may be bound at UASG in Gly medium To measure the amount of expression from the hybrid as well as Gal medium. This result, and the fact that promoters, the GALl and GALI O genes were fused to GAL80 represses transcription of GALl and GALI O by the Escherichia coli lacZ gene, and levels of transcrip- blocking GAL4's transcription-activating domain, sug- tion were determined by assaying for fl-galactosidase gests a model where galactose induces transcription by produced in yeast. The UASc-GAL-lacZ gene fusions causing GAL80 to dissociate from GAL4 molecules were maintained on multicopy plasmids derived from bound at UASG, thus exposing GAL4's transcription-ac- YEp24, as described previously (West et al. 1984; Yocum tivating domain to the cellular transcription apparatus et al. 1984). Since UASc and UASG are differentially reg- (Johnston et al. 1987; Ma and Ptashne 1987b; Selleck ulated, expression of the hybrid UASc-GAL promoters and Majors 1987b). should reflect the physiological and genetic conditions Four other S. cerevisiae genes, GAL7, GAL2, GAL80, normally controlling CYC1 or GAL gene transcription. and MEL1, are also induced by GAL4 and have GAL4 For this purpose, the plasmids were transformed into the binding sites in their 5' control regions (Bram et al. yeast strains YM256 (GAL4 +) or YM335 (Agal4)and the 1986). The MEL1 and GAL80 genes are transcribed at a transformants grown in synthetic medium containing detectable level in uninduced cells (Post-Beittenmiller et either glycerol and lactate (Gly)or galactose plus glyc- al. 1984; Shimada and Fukasawa 1985), whereas GALl, erol and lactate (Gal), prior to assaying for f3-galactosi- GALIO, GAL7, and GAL2 are not (St. John and Davis dase production. 1981; West et al. 1984; Yocum et al. 1984; Tajima et al. 1986; Tschopp et al. 1986). Previous deletion-mapping Expression in GAL medium analysis of the 600-bp GALl-GALl 0 divergent promoter region revealed that a certain portion of UASG (located The activities of the hybrid promoters shown in Figure proximal to the GALl promoter), when deleted, in- 1, when transformed into YM256 (GAL4 ÷) and grown in creased the uninduced level of GALl transcription from Gal medium, are presented in Table 1. Only hybrid pro- an undetectable level to about 5% of the fully induced moters containing a single copy of UASo inserted in the level (West et al. 1984). This raised the possibility that a normal orientation with respect to a TATA box, are in- negative control element(s)is also present in UASG that cluded in Table 1 (see Materials and methods). The re- normally represses transcription of GALl and GALl 0 in sults indicated that if UASG was present in a particular uninduced cells. UASc-GAL1 or UASc-GAL10 hybrid promoter, expres- Here we show evidence suggesting that multiple nega- sion of the UASc-GAL1- or UASc-GALIO-lacZ fu- tive control elements are present in the GAL1-GALIO sions was normally induced, and f3-galactosidase levels divergent promoter region, which may account, in part, were often as much as twofold higher than in cells con- for the lack of detectable expression of the respective taining a respective GALl- or GALI O--lacZ fusion genes in Gly medium. The negative control elements lie lacking UASc. If UASG was absent in a given chimeric adjacent or possibly overlap the GAL4 binding sites in promoter, as with plasmids UASc-GAL1-8 and UASc- UASG, and neither GAL4 nor GAL80 is required for their GALl-9 in Table 1 for example, expression derived function. Deletion derivatives of GAL4 that apparently solely from the activity of UASc. These results show bind to UASG, but fail to activate transcription of the that UASc does not affect the activity of UASG other wild-type GAL genes, significantly block the activity of than to increase the total amount of expression from the the negative control elements, suggesting that normally hybrid promoters and that the GALl and GALI O pro- GAL4 regulates the activity of this repression mecha- moters can be induced by UASc alone when UASG is nism. absent.

Results Repression m Gly medium Experimental design In contrast to the results above, Table 1 also shows that Our goal was to test for the presence of negative control when the same plasmid-containing cells were grown in

GENES & DEVELOPMENT 1119 Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

West et al.

(1) (2)

i' I 9~ 7v

t12v '~ A lw

1o ,' -''- fG | -- I 1 ¢* , _1 B- d, d .,-3 i 2) ,d

a6 [ ill ..9 t -- i - ~ 100 300 500 700 900 L I I I 1 | I ! I | Figure 1. Experimental design. The GAL1-GALIO divergent promoter region, shown at center, is drawn approximately to scale. Stars indicate the positions of the major 5' mRNA cap sites and open arrows the transcriptional orientations of GALl 0 (left) and GALl (right). Closed boxes denote the positions of TATA boxes, whereas the open box designated G indictes the location of UASG. Thin lines with arrows located above and below the map represent DNA sequences remaining in 5' deletions of GALl (A) and GALIO (B), with closed triangles denoting the deletion end points. The open box designated C represents the 150-bp XhoI fragment containing UASc. Numbers located next to the thin lines with arrows designate specific UASc-GAL1-{A) for UASc-GALIO-(B)hybrid promoters characterized in the text and listed in subsequent tables and figures (not all of the promoters are shown here). Closed bars (top) show the sequences included in the upstream (1) and downstream (2)probes that were used for S1 mapping analysis. The scale at the bottom correlates the position of each feature with the DNA sequence of this region, described in Yocum et al. (1984).

Gly medium, [~-galactosidase levels decreased substan- or about UASG. Alternatively, GAL4 may act as a re- tially as large portions of UASG were present, separating pressor in the absence of galactose, either alone or as UASc from the respective promoter. Figure 2 shows that part of a GAL4-GAL80 complex (see introductory sec- for both the GALl promoter (Fig. 2A) and the GALI O tion). Thus, we tested to see if GAL80 or GAL4 (or a promoter (Fig. 2B), [3-galactosidase levels dropped drasti- GAL4-GAL80 complex) was required for this repres- cally (about 400-fold altogether for GALl and over 1350- sion. We transformed each of the UASc-GALI-lacZ fu- fold for GALl O) as a function of the linear distance sepa- sions shown in Table 1 into yeast strains YM335 (Agal4), rating UASc from the promoter. The results of other ex- YJ1 (Agal4), and YM709 (Agal4 Agal80) and analyzed periments suggested that this repression was not due to their expression. The relative level of [~-galactosidase altered spacing between regulatory elements of the chi- synthesized by each fusion in YM335 {Table 2), YJ1, and meric promoters (Guarente and Hoar 1984; R.W. West, YM709 (data not shown)was approximately the same as unpubl.), implying instead that inhibition of UASc-in o that in wild-type strain YM256 (Table 1 ), suggesting that duced expression of GALl and GALI O was due to the neither GAL80 nor GAL4 is required for this repression. presence of negative control elements in the GAL1- S1 mapping studies confirmed that [3-galactosidase GALI O divergent promoter region. Figure 2 also shows levels in YM335 cells accurately reflected the amount of that the same portions of UASG that repress the UASc- specific mRNA made from each of the hybrid promoters GALl or UASc-GAL10 promoters in cells grown in Gly (Fig. 3) and that no transcripts were initiated at positions medium induce their expression in cells grown in Gal upstream of the normal start sites (data not shown). This medium. This suggests that the putative negative con- indicates that inhibition of UASc activity by portions of trol elements reside in the GALl-GALl 0 divergent pro- UASG was not a consequence of transcription (and trans- moter region in the same proximity as the GAL4 binding lation) starting at aberrant upstream locations. sites of UASG. Table 2 also shows that YM335 (Agal4)transformants grown in Gal (galactose plus glycerol and lactate) medium expressed 2- to 20-fold more [~-galactosidase than GAL80 and GAL4 are not required for repression transformants grown in Gly (glycerol plus lactate) me- Since GAL80 negatively regulates transcription of the dium, indicating that galactose partially derepresses the GAL structural genes, it was possible that repression in expression of UASc-GAL1- and UASc-GALIO-lacZ Gly medium was due to GAL80 binding at sequences in fusions in the absence of GAL4. The basis of this partial

1120 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

Negative elements in a yeast GAL promoter derepression by galactose is unclear. Nevertheless, 55-bp fragment (UASG-55; formerly designated UASG', factors other than galactose are plainly required to over- West et al. 1984) containing the GAL4 binding sites 2 come the repression mechanism. and 3 (nomenclature of Giniger et al. 1984) did not repress the CYC1 promoter, suggesting that a GAL4 binding site per se is insufficient for repression. A 75-bp Multiple negative control elements reside in or about fragment (UASG-75) containing GAL4 binding sites 1, UASG 2, and 3 repressed CYC1 promoter activity twofold, The data of Tables 1 and 2 suggested the presence and whereas a l l0-bp fragment (UASG-110)containing possible locations of several distinct negative control el- GAL4 binding site 4 and a 120-bp fragment (UASG-120) ements residing in the GAL1-GALIO divergent pro- containing GAL4 bindings sites 2, 3, and 4 repressed the moter region. To analyze this possibility further, restric- CYC1 promoter fivefold and sevenfold, respectively. A tion fragments containing various parts of the GAL1- 145-bp fragment (UASG-145) containing all four GAL4 GALIO divergent promoter region were inserted binding sites (Bram and Kornberg 1985; Giniger et al. between UASc and the CYC1 TATA box in the wild- 1985) reduced expression of the CYC1 promoter 100- type CYCI-lacZ fusion plasmid pLG669&-312 (Guar- fold. The orientation in which these fragments were in- ente et al. 1984), as depicted in the schematic diagram of serted did not significantly affect their ability to repress Figure 4. The ability of each fragment to repress the ex- the CYC1 promoter (data not shown). pression of the CYC1 promoter in cells (YM335; agal4) Combined, the data of Tables 1-3 suggest the pres- grown in Gly medium was then examined. A 365-bp ence of approximately three negative control elements fragment containing UASG and flanking sequences in the GALl-GALl 0 divergent promoter region, residing (UASG-365) reduced expression of the CYC1 promoter adjacent to and possibly overlapping (but separate from) 1200-fold (Fig. 4). However, smaller portions of the the GAL4 binding sites in UASG. Their apparent loca- GAL1-GALIO divergent promoter region repressed the tions are defined by small portions of the GALl-GALl 0 CYC1 promoter much less, if at all. Table 3 shows that a divergent promoter region having the most notable ef-

Table 1. Activities of UASc--GAL1- and UASc-GALIO-IacZ fusions in a GAL4 + strain 5' Deletion UASc-UASG Presence ~3-Galactosidase activity in Promoter end point a Distance b of UASGc Gly Gal a CYC1 wild type - 877 784 GALl wild type + <0.1 2400 UASc-GAL1- 1 274 220 + 5 4050 (2210) 2 301 193 + 6 4195 (2601) 3 330 164 + 14 4965 (2136) 4 365 129 + 98 4013 (1464) 5 376 118" --- 79 3046 (1593) 6 390 104" -+ 173 3365 (734) 7 423 - 547 529 (0) 8 578 - 2014 790 (0) 9 632 - 1387 862 (0) GALIO wild type + <0.1 450 UASc-GALIO- 1 592 240 + <0.1 882 (576) 2 552 200 + <0.1 914 (363) 3 473 121 * --- 1.0 850 (293) 4 428 76" - 0.5 330 (90) 5 412 60" +- 1.0 486 (5) 6 394 42" --- 1.0 194 (0) 7 390 38* -+ m 131 (0) 8 326 - 135 158 (0) 9 261 - 92 69 (0) 13-Galactosidase activities were from YM256 cells {GAL4 ÷) containing the indicated plasmids, grown in Gly or Gal medium. a Number refers to the 5' deletion end point position in the GALl-GALl 0 divergent promoter region for GALl or GALl O, according to the nomenclature of Yocum et al. (1984). See also Fig. 1. b Distances, denoted in base pairs, are taken from the center of UAS c (Guarente et al. 1984) to the center of UASG (Giniger et al. 1985). Asterisks indicate that one or more of the four GAL4 binding sites of UASG have been removed by the deletion. c (+) Contains all four GAL4 binding sites; (_+) contains one to three GAL4 binding sites; (-) lacks all four GAL4 binding sites. a Activities for GALI-lacZ fusions that lack UASc are denoted by parentheses and are provided for comparison.

GENES & DEVELOPMENT 1121 Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

West et al.

I I I I I I I I j I I I I I I I I I divergent promoter region between positions 330 and 365 (nomenclature of Yocum et al. 1984), proximal to s~k A GALI O and adjacent to or possibly overlapping GAL4 binding site 1. The position of GAL O] is defined by dif- 1000- ferences in the expression (in Gly medium) of the two iS" sets of promoters UASc-GAL1-3 and UASc-GAL1-4 T' (sixfold) and UASc-GALIO-7 and UASc-GALIO-8 (twofold). Additional support for this assignment was ob o tained from UASc-GAL1-3 and UASc-GAL1-4 pro- 100- 5 moters that were integrated into the yeast genome, where a 15-fold difference in their expression was ob- served (see Table 4, described below). A second negative ~,-~ Gly control element appears to be located between positions A--A Gal 10- 365 and 394 and is arbitrarily designated GAL 02. GAL

1 02 is defined by differences in repression of the CYC1 promoter by DNA fragments UASc-55 and UASG-75 (twofold), as well as differences in the expression of the

I/t 1- two sets of promoters UASc-GAL1-4 and UASc- dim B omm GALl-7 (15-fold) and UASc-GALIO-6 and UASc- l: GALl 0-7 (90-fold; see Discussion). GAL 02 probably overlaps GAL4 binding site 1 and may overlap GAL4 B binding site 2 as well. A third negative control element, 1000- arbitrarily designated GAL 03, appears to be located in 5 ,,~" UASG proximal to GALl, between positions 473 and 510. GAL 03 lies adjacent to or overlaps GAL4 binding & site 4 and is defined by differences in repression of the 100- 9 8"JL7 CYC1 promoter by DNA fragments UASG-55 and 9 UASG-120 (fivefold), as well as by differences in expression of the two sets of promoters UASc-GAL1-7 and UASc-GAL1-8 (twofold)and UASc-GALIOo2 and UASc- 10- GALl0-3 (tenfold). A search for similarities in the DNA sequences at the sites corresponding to GAL O], 02, and 03 revealed no striking homologies; further refinement ~Gly of the sizes and locations of the negative control ele- A--A Gal ments may require alternative experimental approaches such as DNase I footprinting. Though GALl and GALI O transcription is regulated by glucose repression as well as by GAL4 and GAL80 (West et al. 1984; Yocum et al. 1984), partly mediated by I .1 I ' I , ', ,.., ,.~ : " ' cis-acting glucose repression elements present in the 1 O0 300 500 700 900 GAL promoter region (West et al. 1984; S. Chen and R. West, unpubl.), GAL O~, 02, and 03 acted independently Position {bp} of the glucose repression pathway. Figure 2. Repression of UASc-GAL hybrid promoters in Gly medium. The GAL1-GALIO divergent promoter region is shown at center (drawn approximately to scale)with the rele- GAL4 deletion derivatives significantly derepress vant features of Fig. 1 included. (A) Semilogarithmic plot of UASc-GAL1 promoters units of ~-galactosidase (from 1 to 5000 units) vs. position of The fact that the GAL operators lie in close proximity to UAS c for UASc-GALI-lacZ fusions. Numbers (only alternates the GAL4 binding sites of UASG (Fig. 5)and that galac- are labeled) correspond to the promoters of Fig. 1 and Table 1./~ tose alone is insufficient to derepress significantly the and A show the results for GAL4 + (YM256) cells grown in Gly (glycerol plus lactate) or Gal (galactose plus glycerol and lactate) UASc-GAL1 and UASc-GALIO promoters in a Agal4 medium, respectively. (B) Semilogarithmic plot of units of [3- strain (Table 2) suggested the possibility that GAL4 it- galactosidase (from 0.1 to 1000 units)vs, position of UAS c for self regulates the activity of the GAL operators when UASc-GAL10--lacZ fusions. bound at UASG. To examine this possibility, we devised a specific genetic selection procedure to obtain GAL4 mutants that, although unable to activate transcription fects on the expression of both UASc-GAL1 and UASc- of the wild-type GAL promoters, might block the ac- GALl 0 promoters (Tables 1 and 2) and on the CYC1 pro- tivity of the GAL operators and allow expression of moter in the plasmid pLG669A-312 (Table 3). Figure 5 UASc-GAL promoters (for details, see Materials and shows that one negative control element, arbitrarily des- methods). Three mutant gal4 genes whose products ac- ignated GAL operator 1 (GAL O~), appears to lie in the tivated an integrated chimeric promoter, UASc-GAL1 o

1122 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

Negative elements in a yeast GAL promoter

GLY GAL MPr234 5 6 78 9W'C /M\ P'l 2 34 5 6 7 8 9W'C M 309

242 o

190

160 q

122

Figure 3. S1 mapping of RNA made in vivo from UASc-GALI-lacZ fusions. Total cellular RNA was isolated from YM335 (Agal4) cells containing plasmids, grown in Gly or Gal medium. The downstream probe (number 2 of Fig. 1) was used for each lane shown. Lane numbers correspond to the UASc-GAL1 hybrid promoters of Table 2. Lane W contained RNA from cells harboring the wild-type GALI-lacZ fusion plasmid pRY131. Other lanes: (M) molecular size markers (HpaII-digested pBR327, West et al. 1984); (P) undigested probe; (C) control reaction (probe plus E. coli tRNA). The arrow at right indicates the position of migration of the major RNA species protected by the probe. Numbers at left indicate the sizes of the markers in base pairs.

1, but not the endogenous wild-type GAL promoters, order of magnitude less than that produced by the re- were obtained by this procedure. Each contained a single spective promoter located on a multicopy plasmid, and base transition that created a translation termination the activity of each was inversely proportional to the codon between the DNA-binding domain at the amino number of GAL operators it contained. Multicopy terminus and the transcription-activating domain at the plasmids containing the gal4-174 or ga14-404 genes were carboxyl terminus of GAL4. Two of these mutations then transformed into each of the strains of Table 4, and placed a stop codon at amino acid position 174 (GAL4- the amount of f~-galactosidase in cells grown in Gly or 174), whereas the third placed a stop codon at amino Gal medium was measured. Table 4 shows that acid position 404 (GAL4-404)of the 881-amino-acid se- GAL4 o174 significantly increased the expression of each quence, yielding derivatives of GAL4 with carboxy-ter- UASc-GAL1 promoter in both media. For example, the minal truncations. Figure 6 shows that both of these chimeric promoter UASc-GAL1-2 was expressed at a GAL4 deletion mutants contain the DNA-binding do- level 63-fold higher in Gly medium and 133-fold higher main (Fig. 6, box 1), but lack a region required for tran- in Gal medium in the presence of GAL4-174. GAL4- scription-activation (Fig. 6, box 2). Both the gal4-174 and 404 significantly derepressed the UASc-GAL1 pro- the ga14-404 genes, when located on a multicopy moters in Gal medium but not Gly medium (Table 4). plasmid and overexpressed by a yeast constitutive pro- Figure 7 shows that in Gal medium, in the presence of moter, gave rise to protein products that failed to cause GAL4-404, integrated chimeric promoters containing significant expression of a wild-type GALI-lacZ fusion one or more GAL operators produced only slightly less (lacking UASc) that had been integrated at the URA3 {about two- to threefold) B-galactosidase than UASc- locus of chromosome 5 (Fig. 6 and Table 4). GALl-8 which lacks the GAL operators. Furthermore, To determine how efficiently GAL4-174 and each chimeric promoter was expressed at roughly the GAL4-404 activated the expression of the UASc-GAL same level, regardless of the total number of GAL oper- promoters, seven of the nine different UASc-GAL1- ators present. These results suggest that GAL4 is a major lacZ fusion plasmids of Table 1 were integrated into the factor responsible for blocking the activity of the GAL URA3 gene of strain YJ1 {Agal4)(see Materials and operators and that its derepressing function normally methods). Seven independent strains resulted, desig- may be physiologically regulated in response to the pres- nated 274.3, 301.1,330.3, 365.1,390.1, and 578.1 (Table ence or absence of galactose. 4). Table 4 shows that the amount of B-galactosidase Even in the presence of GAL4-404 or GAL4-174, produced by each integrated promoter was roughly an UASc-GAL1 promoters containing one or more GAL

GENES & DEVELOPMENT 1123 Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

West et al.

Table 2. Activities of UASc-GAL1 and UASc--GALIO-lacZ tion of the GAL structural genes is also inhibited by fusions in a Agal4 strain GAL80, we concur that at least two distinct pathways are involved in repressing GALl and GALl 0 in Gly me- ~-Galactosidase Fold dium. The pathway we characterized is independent of activity in derepression GAL4 and GAL80, but may impose an equal amount of Promoter Gly Gal by galactose control on the expression of the GAL structural genes. CYC1 Combined, the two control mechanisms moderate tran- wild type 842 777 scriptional levels of GALl and GALI O by over four GALl orders of magnitude (West et al. 1984; Yocum et al. wild type <0.1 <0.1 1984). (<0.1) (2400) To a first approximation, the three putative negative UASc-GAL1- control elements we defined, tentatively designated 1 3 11 4 GAL O~, Oz, and O3, map to positions either adjacent to 2 2 17 9 3 7 40 6 or overlapping GAL4's binding sites in UASG. This fact 4 44 118 3 may have precluded their identification during the 5 68 210 3 course of biochemical procedures used to characterize 6 115 257 2 the sites of GAL4 binding in UASG. To confirm the 7 582 1194 2 number of negative control elements and their precise 8 1380 1372 1 sizes and locations will require considerable more study. 9 1535 1509 1 Nevertheless, our results suggest that in the case of GALIO UASG, the term UAS may be a misnomer and that the wild type <0.1 <0.1 structure and function of UASG are more complicated {<0.1) (450) UASc-GALIO- than previously imagined. 1 <0.1 <0.1 The mechanism by which the GAL operators inhibit 2 <0.1 <0.1 transcriptional activation is unclear. In conjunction 3 1.0 22 22 with previous data showing that a 365-bp fragment con- 4 1.0 9 9 taining UASG (see Fig. 4) did not repress expression of 5 0.4 8 20 the CYC1 promoter when positioned upstream of UASc 6 1.0 18 18 in the plasmid pLG669A-312 (Guarente and Hoar 1984), 7 85 213 3 our data suggest that the GAL operators repress tran- 8 180 319 2 scription if located downstream of a UAS but not up- 9 166 184 1 stream of it. We recently confirmed this notion by in- The Agal4 strain used was YM335. Activities in brackets were serting a single copy of UASc at two independent posi- from an isogenic GAL4 + strain {YM256). For additional infor- tions between UASG and the GALl TATA box; mation see Table 1. expression of these "reverse" chimeric promoters was roughly equivalent to that of UASc-GAL1-8, which was nonrepressed (Table 1; R.W. West, unpubl.). Thus, the operators were normally repressed two- to ninefold in GAL negative control elements may be functionally Gly or Gal medium (Table 4; Fig. 7). Apparently, either the loss of a portion of the GAL4 protein diminishes the capacity of GAL4-174 or GAL4-404 to inhibit the ac- Table 3. Repression of the CYC1 promoter by parts of the tivity of the GAL operators or the GAL4-404 and GAL promoter region GAL4-174 proteins themselves partially inhibit the function of these promoters (see, e.g., Keegan et al. J3-Galactosidase 1986). activity in Activation of the UASc-GAL1 chimeric promoters by Promoter Gal Gly Fold repression GAL4-174 or GAL4-404 was dependent on the presence CYC1 (wild type) 1050 1240 of UASc as well as the presence of a wild-type allele of UASG-365 2713 1 1240 the HAP1 gene, which encodes a positive control protein UASG- 145 2617 12 100 that binds to UASc (Guarente et al. 1984; Pfeifer et al. UASG-55 1980 1188 1 1987), showing that the GAL4 deletion derivatives acti- UASG- 75 2516 650 2 vated expression by removing a block on UASc-induced UASG-110 1421 171 7 transcription of the GALl promoter (data not shown). UASG- 120 2031 178 5 The plasmid pLG669A-312 {Guarente et al. 1984), containing Discussion the wild-type CYC1 promoter fused to lacZ, was used as a control. The strain used was YM335 (Agal4). Fold repression values We have shown evidence suggesting that transcriptional indicate the ratio of the f]-galactosidase activity produced from regulation of the S. cerevisiae GALl and GALI O genes the wild-type CYC1 promoter relative to that of the given hy- involves negative control elements as well as inducing brid promoter, in Gly medium. For additional information, see sequences in the GAL promoter region. Since transcrip- Figs. 4 and 5 and the text.

1124 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

Negative elements in a yeast GAL promoter Activity UAS .. o ..,,~ Sma I C Xho I Gly Gal ,

A@ CYC 1 1240 1050 | • • t i II \ UAS c i I., 'rnm B@ I " CYC 1 1 2713 365 bp t UAS G O~

C@ GALIO GALl 0 3200

~4--,,- i Ddel Sau3a Figure 4. Repression of the CYC1 promoter by UASG. A 365-bp SalI fragment containing UASG (UASG-365), which spans the GAL1- GALIO divergent promoter region from a unique DdeI site at position 300 to a unique Sau3a site at position 660, was inserted into the unique XhoI site of the wild-type CYCI-lacZ fusion plasmid pLG669A-312. At right are shown 13-galactosidase activities for the wild-type CYCI-lacZ fusion plasmid pLG669A-312 (A), the chimeric UASc- UASc-CYCI-lacZ fusion plasmid (B), and the wild-type GALI-lacZ fusion plasmid pRY131 (C) in YM335 (&gal4) cells grown in Gly or Gal medium. analogous to bacterial operators since their activity is that the products of the ga14-174 and ga14-404 genes position dependent. In this respect they are dissimilar to bind to UASG in vivo or in vitro, other investigators have other yeast negative control elements recently charac- shown that a variety of GAL4 mutants that retain the terized, which inhibit the activity of a UAS from either DNA-binding domain but contain other structural alter- an upstream or downstream location (Brand et al. 1985; ations can bind to UASG in an apparently normal fashion Brent 1985; Johnson and Herskowitz 1985; Miller et al. (Brent and Ptashne 1985; Keegan et al. 1986; Johnston 1985; Wright and Zitomer 1985; Siliciano and Tatchell and Dover 1987; Johnston et al. 1987; Ma and Ptashne 1986). 1987a, b). In view of the fact that the GAL operators and A striking result was that small portions of UASG the GAL4 binding sites of UASG lie in close proximity, containing individual GAL operators inhibited UASc-in- we favor a mechanism where the GAL4 deletion deriva- duced expression of the GALl and GALI O promoters or tives directly inhibit the activity of the GAL operators the CYC1 promoter only two- to sevenfold, whereas when bound to one or more of the four GAL4 binding larger portions containing two or three GAL operators sites in UASG. The fact that GAL4-404 derepresses the inhibited expression more significantly (see Table 3). UASc-GAL1 promoters in Gal medium but not Gly me- One interpretation of this result is that two or more dium may indicate that GAL4-404 (and GAL4 + )harbors GAL operators act synergistically to increase the a specific domain lacking in GAL4- 174 that responds to amount of repression. Thus, for example, GAL 02 alone the presence or absence of galactose to regulate the blocked transcription of the CYC1 promoter in the function of the GAL operators. One possibility is that plasmid pLG669A-312 only twofold (compare UASG-55 GAL80 binds to GAL4 at a position corresponding to the and UASG-75; Fig. 5), whereas GAL 02 together with region located between the translation stop codons of GAL O~ blocked GALIO transcription more than 90-fold GAL4-174 and GAL4-404 (in addition to binding to in the promoter UASc-GALIO-6 (compare UASc- GAL4's carboxyl terminus; Johnston et al. 1984; Ma and GALl 0-6 and UASc-GALIO-7, Fig. 5). A synergistic in- Ptashne 1987b) to control GAL4's derepressing function. teraction that may occur when the GAL operators are If a repressor protein(s) modulates the activity of the present in various combinations is reminiscent of a GAL operators, then the fact that its binding sites reside specfic property of bacterial operators, where repressor adjacent to or overlap those of GAL4 implies that the dimers bind cooperatively to two operator sites to render repressor and GAL4 may compete for their respective a greater degree of repression than a repressor dimer target sites in UASc. This view is consistent with fea- bound to a single operator (Ptashne 1986). tures of other yeast promoters, where transcription is Several different mechanisms may be envisioned to regulated by the mutually exclusive binding of positive account for the ability of GAL4-174 and GAL4-404 to and negative control proteins at overlapping or adjacent activate the expression of the UASc-GAL1 chimeric sites (Pfeifer et al. 1987; Nasmyth et al. 1987). Alterna- promoters. Although we did not provide direct evidence tively, GAL4 and a hypothetical repressor protein may

GENES & DEVELOPMENT 1125 Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

West et al.

i i|. I I w I I ! 2S0 35O 450 55O 650 UAS G

O1 02 03 GAL14 097) Hpa Alu Bst A

GALIO 55 :(,)

• 11o (,)

129

•. 145 __, 0003

| 365 (1240) , •

Figure 5. GAL negative control elements map to three locations. At center is shown the GAL1-GALIO divergent promoter region, drawn approximately to scale, from the TATA box (T) of GALIO (left) to GALl (right). Positions are denoted by the scale at the top of the figure. The locations of UASG and GAL4 binding sites 1-4 (Bram and Kornberg 1985; Giniger et al. 1985) are noted. The approximate positions of three negative control elements, tentatively designated GAL O~, 02, and 03, are indicated by double lines above the sequence. (A) Arrows descending from above the sequence indicate deletion end point positions of 5' deletions of GALl used in constructing a respective UASc-GALI-lacZ fusion. The number adjacent to the arrow refers to the respective promoter of Fig. 1 and Table 1. The number in parentheses above the arrow indictes the fold-repression of UASc at that position. (B) Arrows ascending from below the sequence denote deletion end point positions of 5' deletions of GALIO used to construct a respective UASc-GALIO-lacZ fusion, and numbers adjacent to the arrows denote the corresponding hybrid promoters of Fig. 1 and Table 1. Fold-repression values for A and B were calculated from the average activity obtained for a given plasmid from Tables 1 and 2 (in Gly medium in YM256 and in Gly or Gal medium in YM335), relative to the activity of UASc-GAL1-8 for UASc-GALI-lacZ fusions or of UASc-GALIO-8 for UASc-GALIO-lacZ fusions. (C)Bars designate the position and extent (bp)of sequences of UASo that were cloned into the wild- type CYC1 promoter, whereas the number in parentheses adjacent to each bar denotes the fold-repression occurring in Gly medium (see Table 3).

bind simultaneously to the GAL promoter region in Gly Ma and Ptashne 1987b). Subsequently, and possibly asa medium, consistent with the notion than an inactive consequence of the dissociation of GAL80, a structural GAL4-GAL80 complex occupies the normal GAL4 alteration occurs that somehow allows GAL4 to block binding sites of UASG in uninduced cells (see introduc- the activity of the repressor protein. Thereafter, GAL4 tory section; Johnston et al. 1987; Lue et al. 1987; Ma stimulates transcription by a mechanism of positive and Ptashne 1987b; Selleck and Majors 1987b). control (Ptashne 1986; Struhl 1987}. Based on the latter hypothesis, the galactose-inducible derepressing function of GAL4-404, as well as the close proximity of GAL4's binding sites with respect to the Materials and methods GAL operators, suggest a possible regulatory model like Strains and plasmids that shown in Figure 8. In Gly medium, a hypothetical S. cerevisiae YM335 (aAga14-537 ura3-52 ade2-101 lys2- repressor protein(s) binds to each of the GAL operators 801 his3-200 met-)and the isogenic GAL4 + strain YM256, as and reduces basal level transcription of GALl and well as YM709 (a Aga14-542 Aga180-538 ura3-52 his3-200 ade2- GALl 0, whereas inactive GAL4-GAL80 complexes re- 101 lys2-801 trpl-901 tyrl-501 met - CAN r) were kindly pro- side at the GAL4 binding sites of UASG. In Gal medium, vided by M. Johnston. S. cerevisiae YJ1 (a Agal4 leu2-3 leu2- galactose (or a metabolic derivative) binds to GAL80, l12ura3-52his-MEL1) was a gift of S. Johnston. JGll5 causing it to dissociate from GAL4 (Johnston et al. 1987; (cxAgal7 ade8 HIS+) was provided by Jim Yarger. The strain

1126 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

Negative elements in a yeast GAL promoter

Amino Acids I I I I I l 100 300 $00 7OO 881

m Activity 174 404 Gly ~1 GAL4 + "".,2 t t .. coo. | 96 234

TGA GAL4-174 .~ -I 0.1 0.8

GAL4"404 :-- TtAG o.1 1 Figure 6. Structure and activity of GAL4 and its deletion derivatives. Linear representations of wild-type GAL4 protein (881 codons) and the GAL4 deletion derivatives, GAL4-174 (173 codons) and GAL4-404 (403 codons), are drawn approximately to scale. Positions of the in-frame translation stop codons of the respective GAL4 deletion derivatives are indicated. The major functional domains of GAL4 are denoted by boxes above the GAL4 + map, designating the position of the DNA-binding domain at the amino terminus (box 1) and the transcription-activating domain at the carboxyl terminus (box 2). Activities of GAL4 and its deletion derivatives are shown at right. The GAL4 + gene and the gal4-174 and ga14-404 genes were present on the multicopy plasmid AAH5 and transcribed from the constitutive ADC1 promoter. Values indicate the amount of f~-galactosidase synthesized in strain 131.1 (Agal4 GALl-lacZ::URA3) following growth in Gly or Gal medium. Overproduction of wild-type GAL4 protein causes the GALI-lacZ fusion to be expressed in noninduced cells (Gly medium; see Johnston and Hopper 1982). For further information, see Tables 4 and 5 and the text.

YJ1-7 (Aga14Aga171eu2-3,112ura3-52) was constructed by dIII fragment containing the wild-type GAL4 gene, was de- mating strains YJ1 and JGll5 and screening for progeny that scribed previously (Silver et al. 1984; Ma and Ptashne 1987a). were Ura- and Leu-, slow growing on YEP(I% yeast extract, The wild-type GALl- and GALI O-lacZ fusion plasmids, 2% peptone)plates containing 2% galactose (YEP-Gal), and that pRY131 (or pRY121)and pRY133 (or pRY123), respectively, and were unable to grow on YEP-Gal plates after transformation their 5' deletion derivatives have been described previously with a plasmid containing the wild-type GAL4 gene. S. cerevi- (West et al. 1984; Yocum et al. 1984). siae BWG1-7a (a leu2-2,112 his4-519 adel-lO0 ura3-52) and the isogenic hapl-1 derivative strain were kindly provided by L. Guarente. E. coli strains RR1 and DHSa were used for routine Media and chemicals cloning work. S. cerevisiae cells were grown in synthetic defined (SD)me- Plasmid pLG669A-312, containing a wild-type CYCI-lacZ dium (0.67% yeast nitrogen base without amino acids)con- fusion, was described previously (Guarente et al. 1984}. Plasmid taining either 3% glycerol and 2% lactate (Gly medium)or 2% pLPK-C15 (a gift from L. Keegan), which harbors a 2.9-kb Hin- galactose plus 3% glycerol and 2% lactate (Gal medium). Ethyl

Table 4. GAL4 deletion derivatives significantly derepress integrated UASc.--GAL1 promoters

[3-Galactosidase activity ¢ Fold derepression a Integrated Control gal4-174 ga14-404 gal4-174 ga14-404 Strain a promoter b Gly Gal Gly Gal Gly Gal Gly Gal Gly Gal 131.1 GALl {wild type) 0.1 0.1 0.1 0.8 0.1 1 .... 274.3 UASc-GALI-1 0.6 0.6 17 25 0.5 30 28 42 1 50 301.1 UASc-GAL1-2 0.3 0.3 19 40 0.4 30 63 133 1 100 330.3 UASc-GAL1-3 0.2 0.2 8 20 0.2 24 40 100 1 120 365.1 UASc-GAL1-4 3 3 21 11 2 40 7 4 1 13 376.1 UASc-GAL1-5 4 4 17 19 3 42 4 5 1 11 390.1 UASc-GAL1-6 13 7 30 49 19 56 2 4 1.5 8 578.1 UASc-GA L 1-8 57 60 72 40 76 68 1 1 1 1 Strains were transformed with either the expression vector AAH5 (control plasmid, lacking GAL4 coding sequences) or AAH5 containing gal4-174 or ga14-404. a Strains were derived from YJ1 (Agal4) by integrating a plasmid containing the respective GALl or UASc-GAL1 promoter, fused to lacZ, into the URA3 gene (see Materials and methods). Strain designations for UASc-GAL1 promoters also indicate the position in the GAL promoter region at which UASc was inserted. b See Fig. 1 and Table 1. c Values indicate units of ~-galactosidase activity for strains (containing the respective plasmids) grown in either Gly or Gal medium. a Values indicate ratios of the amount of expression in strains containing AAH5 plus gal4-174 or ga14-404 relative to strains containing AAH5 alone, for a given medium.

GENES & DEVELOPMENT 1127 Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

West et al.

UAS G __

., il II1111 1 100- 123456 8 .°i~8

4.;~ .... -2s .2

A ul ,--, Oal 4-- 10" em C ~--~ Oly

Figure 7. Galactose-induced derepression of inte- V grated UASc-GAL1 promoters by GAL4-404. Shown at 4 top is a map of the GALl promoter, drawn approxi- ~s mately to scale, with the positions of UASG, GAL4 el> binding sites 1-4, GAL operators Oz, 02, and 03, GALl el TATA box (filled box), GALl gene, and 5' mRNA cap 6= 1 t,0 site (star)and transcriptional orientation (arrow)de- < 1 noted. Vertical arrows beneath the map indicate the positions at which UAS c was inserted in the GAL promoter region, corresponding to the respective chimeric promoters UASc-GALI-1 (arrow 1)to UASc-GAL1-8 A 3 (arrow 8). Below the map is a semilogarithmic plot of units of f~-galactosidase (from 0.1 to 100 units) pro- 0.1 ' ' I I duced by each corresponding integrated UASc-GAL1 -2 o ' promoter, in cells grown in Gal (A) or Gly (A) medium. For additional information, see Table 4. Position (bp)

methanesulfonate (EMS)was obtained from Sigma Co. 5- described by West et al. (1984), and the precise end points of Bromo-4-chloro-3-indolyl-f3-D-galactoside (Xgal)was purchased each deletion in the GAL1-GALIO divergent promoter region from Boehringer-Mannheim Co. are provided in Table 1. A 134-bp SmaI-XhoI fragment harboring UASc was obtained from the plasmid pLG669-Z (Guar- Yeast transformation and fl-galactosidase assays ente and Ptashne 1981 ). An 8-bp linker was placed at the SmaI end, giving a nominal 150-bp XhoI fragment. The XhoI frag- Yeast were transformed using either the spheroplast (Sherman ment containing UASc was then inserted into the unique XhoI et al. 1986) or the lithium acetate {Ito et al. 1983) techniques. sites located at the end points of the GALl and GALIO 5' dele- Transformants were selected on SD medium (lacking the ap- tion mutants (Fig. 1). The orientation and number of copies of propriate amino acid) containing 2% glucose. ~3-Galactosidase UASc inserted at each 5' deletion end point of GALl and assays were performed as described previously (West et al. GALl 0 were determined by performing restriction mapping ex- 1984). Samples were analyzed in triplicate cultures and the re- periments using the enzymes MluI and EcoRI. The former en- sults averaged. zyme cuts at an asymmetric location within the 150-bp XhoI fragment, whereas the latter cleaves the plasmid vectors at a S1 mapping site corresponding to the fusion junction of GALl or GALI O with the lacZ gene (see, e.g., Fig. 1 of West et al. 1984). The Total S. cerevisiae RNA was isolated as described by Sherman presence of two or more copies of UASc in a particular hybrid et al. (1986), and the 5' ends of GALl-encoded transcripts were promoter was detected by the presence of an additional 150-bp mapped by S1 nuclease analysis (Weaver and Weissman 1979). MluI fragment (derived from tandem 150-bp XhoI fragments Reaction mixtures contained 10 ~g of RNA, 500 U/ml of S1 containing UASc)on polyacrylamide gels. nuclease (Sigma Co.), and an excess of single-stranded a2P-la- The original mutant GALl promoters and the corresponding beled DNA probe. The downstream probe extended from posi- hybrid UASc-GAL1 promoters are as follows: 121-274, tion 688 to 930 {Figs. 1 and 3) and was isolated and 3zP-labeled UASc-GALI-1., 121-301, UASc-GAL1-2; 121-330, UASc- as described previously (West et al. 1984). The upstream probe, GALl-3; 121-365, UASc-GAL1-4; 121-376, UASc-GAL1-5; which extended from position 300 to 660, was isolated as a 121-390, UASc-GAL1-6; 121-423, UASc-GAL1-7; 121-578, 365-bp BglII fragment from the plasmid pRY24 (a gift of R. UASc-GAL1-8; 121-632, UASc-GAL1-9. The original mutant Yocum) and 32P-labeled and strand separated as described pre- GALIO promoters and the corresponding hybrid UASc-GALIO viously (West et al. 1984). promoters are: 123-592, UASc-GALIO-1; 123-552, UASc- GALIO-2; 123-473, UASc-GALIO-3; 123-428, UASc- Construction of UASc-GAL1 and UASc-GALIO hybrid GALIO-4; 123-412, UASc-GALIO-5; 123-394, UASc- promoters GALIO-6; 123-390, UASc-GALIO-7; 123-326, UASc- The 5' deletion mutants of GALl and GALl 0 were previously GALIO-8; 123-261, UASc-GALIO-9. Wild-type CYC1, GALl,

1128 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

Negative elements in a yeast GAL promoter

OFF • TATA TATA ~ OFF 10 l mm n , I 1

.,o,.c,o.. .,o,.c,o..,.,

O O K; 4Y° aP

Figure 8. Hypothetical scheme depicting GALl and GALIO regulation in the presence ( + ) or absence ( - ) of galactose. In the absence of galactose (Gly medium; top), an inactive GAL4-GAL80 complex binds to UAS G (for clarity, only a monomer of the GAL4-GAL80 complex is shown), and a hypothetical repressor protein (R) binds to GAL O1, 02, and 03. In the presence of galactose (Gal medium; bottom), galactose itself or a metabolic derivative (A) binds to GAL80, causing it to dissociate from GAL4. Subsequently, GAL4 undergoes a structural transition (O), allowing it to block the activity of the hypothetical repressor (possibly causing the repressor to dissociate from the GAL operators; "a" arrows)and to stimulate RNA polymerase II to transcribe GALl and GALl 0 ("b" arrows). The effects of glucose repression (Yocum et al. 1984; Selleck and Majors 1987b) are not considered in this model.

and GALI O promoters, as well as the UASc-GAL1 and UASc- Integrated UASc-GALI-lacZ fusion strains GALI O hybrid promoters, were fused to the lacZ gene (de- Plasmids containing UASc-GALI-lacZ fusions were inte- scribed previously: Yocum et al. 1984; West et al. 1984)and grated by gene transplacement (Rothstein 1983)into the URA3 promoter activities measured as a function of B-galactosidase locus of S. cerevisiae strain YJ1, using the procedure described synthesis. by Brent and Ptashne (1984). YIp derivatives (lacking 2p replicon sequences) of muhicopy plasmids containing UASc- Subcloning parts of UASc in pLG669A-312 GALI-lacZ fusions were constructed by partial digestion with EcoRI and reclosure with T4 DNA ligase. The plasmids were A 365-bp SalI fragment (UASG-365) containing the sequence then linearized in the URA3 coding sequence by cutting with from position 300 to position 660 of the GAL1-GALIO diver- either ApaI or StuI, and the linearized plasmids were trans- gent promoter region (Guarente et al. 1982; Yocum et al. 1984) formed into strain YJ1. Ura + transformants were selected, and was obtained from the plasmid pRY26, a gift of R. Yocum. The stable integrants were identified by plasmid segregation anal- 145-bp fragment containing UAS G (UASG-145) was obtained by ysis and Southern blotting procedures. The strains derived from digesting the GALI-lacZ fusion plasmid 121-365 (West et al. YJ1 in this manner were designated as follows: 131.1 (wild-type 1984) with XhoI and AluI, which cleave at positions 365 and GALI-lacZ::URA3); 274.3 (UASc-GALI-I::URA3); 301.1 510 of the divergent promoter region (respectively), and placing (UASc-GAL1-2::URA3); 330.1 (UASc-GAL1-3::URA3); 365.1 an 8-bp XhoI linker at the AluI end. The 75-bp fragment con- (UASc-GAL1-4::URA3); 376.1 (UASc-GAL1- 5::URA3); 390.1 taining UAS G (UASG-75) was obtained by digesting 121-365 (UASc-GAL1-6::URA3); 578.1 (UASc- GAL1-8::URA3). with XhoI and HpaII, the latter enzyme cutting within UAS G at position 440. The HpaII end was filled in using Klenow fragment and dNTPs prior to ligation to a XhoI linker. The 120-bp Genetic selection of gal4 mutants fragment (UASG- 120)was obtained by cutting the GALI-lacZ fusion plasmid 121-390 with XhoI and AluI. The former en- We first constructed a Agal7 derivative of the strain YJ1 (&gal4) zyme cuts at position 390 in UASG. The 55-bp (UASG-55) and designated YJ1-7 (Agal4 AgalT) and then integrated a YIp deriva- l l0-bp (UASG-110) fragments were obtained by digesting tive of a plasmid that contained the chimeric promoter UAS c- 121-390 with XhoI and HpaII, as well as BstNI (position 550), GALl-1 fused to lacZ (Fig. 1; Table 1)into the URA3 gene of gel purifying the 55-bp and 110-bp fragments, filling in the ends YJ1-7 to yield the strain PC1-3 (Agal4 Agal7 UASc-GAL1- using Klenow fragment and dNTPs, and ligating to XhoI I::URA3). We then transformed PC1-3 with a muhicopy linkers. The 55-bp fragment, formerly designated UASG', was plasmid (pLPK-C15)containing the wild-type GAL4 gene, tran- described previously (West et al. 1984). The SalI or XhoI frag- scribed by the constitutive ADC1 promoter (Ammerer 1983). ments derived from these procedures were then cloned into the When the resulting strain is grown in the presence of galactose, unique XhoI site of the plasmid pLG669A-312. the Agal7 mutation causes the toxic substrate galactose-1-

GENES & DEVELOPMENT 1129 Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

West et al.

phosphate to accumulate, providing a positive selection for Bram, R.J., N.F. Lue, and R.D. Kornberg. 1986. A GAL family of mutations (including gal4- mutants)in the GAL pathway upstream activating sequences in yeast: Roles in both in- (Matsumoto et al. 1980). We then mutagenized this strain with duction and repression of transcription. EMBO J. 5: 603- EMS, plated the cells onto selective medium containing galac- 608. tose plus glycerol and lactate, as well as the indicator dye Xgal, Brand, A.H., L. Breeden, J. Abraham, R. Sternglanz, and K. Nas- and screened for survivors of gal7 killing (gall- or gal4-) that myth. 1985. Characterization of a "silencer" in yeast: A were blue. GAL4 mutants that were incapable of binding to DNA sequence with properties opposite to those of a tran- UASG produced white colonies, whereas GAL4 mutants that scriptional enhancer. Cell 41: 41-48. bound to UAS G and blocked the activity of the GAL operators Brent, R. 1985. Repression of transcription in yeast. Cell (but failed to activate transcription of the endogenous GALl 42: 3-4. gene on chromosome 2)produced blue colonies. Plasmid DNA Brent, R. and M. Ptashne. 1984. A bacterial repressor protein or from 16 blue PC1-3 colonies was then isolated, using the a yeast transcriptional terminator can block upstream acti- method of Sherman et al. (1986). By recombining complemen- vation of a yeast gene. Nature 312: 612-615. tary pairs of restriction fragments containing either the GAL4 1985. A eukaryotic transcriptional activator bearing the gene or vector sequences obtained from wild-type or mutagen- DNA specificity of a prokaryotic repressor. Cell 43: 729- ized plasmid DNA, we determined that 15 of the 16 plasmids 736. contained mutant gal4 genes whose products failed to activate Chen, W. and K. Struhl. 1985. Yeast mRNA initiation sites are transcription of the wild-type GALI-lacZ fusion in strain determined primarily by specific sequences, and not by dis- 131.1. For four of the 15, the approximate positions of muta- tance from the TATA element. EMBO J. 4" 3273-3280. tions in the GAL4 coding sequence (Laughon and Gesteland Giniger, E., S.M. Varnum, and M. Ptashne. 1985. Specific DNA 1984) were mapped by recombining complementary pairs of re- binding of GAL4, a positive regulator protein of yeast. Cell striction fragments obtained from mutagenized and wild-type 40: 767- 774. GAL4 DNA. After localizing a given mutation to a specific re- Guarente, L. 1984. Yeast promoters: Positive and negative ele- striction fragment, the precise base alteration was found by ments. Cell 36: 799-800. DNA sequencing (both strands)using either the Sanger or Guarente, L. and E. Hoar. 1984. Upstream activation sites of the Maxam and Gilbert techniques. CYC1 gene of Saccharomyces cerevisiae are active when in- verted but not when placed downstream of a TATA box. Proc. Natl. Acad. Sci. 81" 7860-7864. Expression of wild-type and mutant gal4 genes in yeast Guarente, L. and M. Ptashne. 1981. Fusion of Escherichia coli The yeast expression vector AAH5 (Ammerer 1983)was used to lacZ to the cytochrome c gene of Saccharomyces cerevisiae. express the wild-type and mutant gal4 genes in S. cerevisiae. Proc. Natl. Acad. Sci. 78: 2199-2203. AAH5 contains the ADC1 promoter, a 2 pm replicon sequence Guarente, L., R.R. Yocum, and P. Gifford. 1982. A GALIO- for maintenance in multiple copies, a LEU2 ÷ gene for selection CYC1 hybrid yeast promoter identifies the GAL4 regulatory in yeast, and sequences required for selection and maintenance region as an upstream site. Proc. Natl. Acad. Sci. 79" 7410- in E. coli. HindIII fragments 2.9 kb in size, containing the wild- 7414. type or mutant gal4 genes, were cloned into a unique HindIII Guarente, L., B. Lalonde, P. Gifford, and E. Alani. 1984. Dis- site located immediately downstream of the ADC1 promoter. tinctly regulated tandem upstream activation sites mediate catabolite repression of the CYC1 gene of S. cerevisiae. Cell 36:503-511. Acknowledgments Hahn, S., E.T. Hoar, and L. Guarente. 1985. Each of three "TATA elements" specifies a subset of the transcription ini- We especially thank Lenny Guarente and Rog Yocum for many tiation sites at the CYC-1 promoter of Saccharomyces cere- helpful discussions during the early phases of this work and for visiae. Proc. Natl. Acad. Sci. 82: 8562- 8566. providing many of the necessary strains and plasmids. R.W.W. Hope, I.A. and K. Struhl. 1985. GCN4 protein, synthesized in wishes to thank Mark Ptashne, in whose laboratory parts of this vitro, binds HIS3 regulatory sequences: Implications for project were initiated. We thank Mark Johnston, Stephan John- general control of amino acid biosynthetic genes in yeast. ston, and Jim Yarger for yeast strains and Ray Judware for Cell 43:177-188. helping to construct a YIp derivative of the plasmid containing Ito, H., F. Yasuki, K. Murata, and A. Kimura. 1983. Transfor- UAS c- GALl-8. We also thank Russ Finley, Peter Hahn, Joseph mation of intact yeast cells treated with alkali cations. J. L. Messina, Dave Mitchell, and Michael Schechtman for a crit- Bacteriol. 153: 163-168. ical reading of the manuscript. Johnson, A.D. and I. Herskowitz. 1985. A repressor (MATa2 This work was supported by grants to R.W.W. from the product) and its operator control expression of a set of cell American Cancer Society (MV-269), the Alexandrine and Alex- type specific genes in yeast. Ce11 42: 237-247. ander Sinsheimer Fund (PN72675), and the New York State Johnston, M. 1987. Genetic evidence that zinc is an essential Health Research Council (15-068). co-factor in the DNA binding domain of GAL4 protein. Na- ture 328: 353-355. Johnston, M. and R.W. Davis. 1984. Sequences that regulate the Reterences divergent GALI--GALIO promoter in Saccharomyces cere- Ammerer, G. 1983. Expression of genes in yeast using the visiae. Mol. Cell. Biol. 4: 1440- 1448. ADC1 promoter. Methods Enzymol. 101: 192-201. Johnston, M. and J. Dover. 1987. Mutations that inactivate a Arndt, K. and G.R. Fink. 1986. GCN4 protein, a positive tran- yeast transcriptional regulatory protein cluster in an evolu- scription factor in yeast, binds general control promoters at tionarily conserved DNA binding domain. Proc. Natl. Acad. all 5' TGACTC 3' sequences. Proc. Natl. Acad. Sci. Sci. 84: 2401-2405. 83:8516-8520. Johnston, S.A. and J.E. Hopper. 1982. Isolation of the yeast regu- Brain, R.J. and R.D. Kornberg. 1985. Specific protein binding to latory gene GAL4 and analysis of its dosage effects on the far upstream activating sequences in polymerase II pro- galactose/melibiose regulon. Proc. Natl. Acad. Sci. moters. Proc. Natl. Acad. Sci. 82" 43-47. 79: 6971-6975.

1130 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

Negative elements in a yeast GAL promoter

Johnston, S.A., J.M. Salmeron, and S.S. Dincher. 1987. Interac- Siliciano, P.G. and K. Tatchell. 1986. Identification of the DNA tions of positive and negative regulatory proteins in the ga- sequences controlling the expression of the MATs locus of lactose regulon of yeast. Cell 50:143-146. yeast. Proc. Natl. Acad. Sci. 83: 2320-2324. Keegan, L., G. Gill, and M. Ptashne. 1986. Separation of DNA Silver, P.A., L.P. Keegan, and M. Ptashne. 1984. Amino ter- binding from the transcription-activating function of a eu- minus of the yeast GAL4 gene product is sufficient for nu- karyotic regulatory protein. Science 231: 699-704. clear localization. Proc. Natl. Acad. Sci. 81: 5951-5955. Laughon, A. and R.F. Gesteland. 1982. Isolation and prelimi- St. John, T.P. and R.W. Davis. 1981. The organization and tran- nary characterization of the GAL4 gene, a positive regulator scription of the galactose gene cluster of Saccharomyces. J. of transcription in yeast. Proc. Natl. Acad. Sci. 79: 6827- Mol. Biol. 152: 285-315. 6831. Tajima, M., Y. Nogi, and T. Fukasawa. 1986. Duplicate up- 1984. Primary structure of the Saccharomyces cerevi- stream activating sequences in the promoter region of the siae GAL4 gene. Mol. Cell. Biol. 4: 260-267. Saccharomyces cerevisiae GAL7 gene. Mol. Cell. Biol. Lohr, D. and J. Hopper. 1985. The relationship of regulatory 6: 246-256. proteins and DNase I hypersensitive sites in the yeast Struhl, K. 1987. Promoters, activator proteins, and the mecha- GALI-IO genes. Nucleic Acids Res. 13: 8409-8423. nism of transcriptional imitation in yeast. Cell 49: Lue, N.F., D.I. Chasman, A.R. Buchman, and R.D. Kornberg. 295-297. 1987. Interaction of GAL4 and GAL80 gene regulatory pro- Tschopp, J.F., S.D. Emr, C. Field, and R. Schekman. 1986. GAL2 teins in vitro. Mol. Cell. Biol. 7:3446-3451. codes for a membrane-bound subunit of the galactose per- Ma, J. and M. Ptashne. 1987a. Deletion analysis of GAL4 de- mease in Saccharomyces cerevisiae. I. Bacteriol. 166: 313- fines two transcriptional activating segments. Cell 48: 847- 318. 853. Weaver, R.F. and C. Weissman. 1979. Mapping of RNA by a 1987b. The carboxy-terminal 30 amino acids of GAL4 modification of the Berk-Sharp procedure: The 5' termini are recognized by GAL80. Cell 50: 137-142. of 15S ~-globin mRNA precursor and mature 10S ~-globin Matsumoto, K., Y. Adachi, A. Toh-e, and Y. Oshima. 1980. mRNA have identical map coordinates. Nucleic Acids Res. Function of positive regulatory gene gal4 in the synthesis of 7: 1175-1193. galactose pathway enzymes in Saccharomyces cerevisiae: West, Jr., R.W., R.R. Yocum, and M. Ptashne. 1984. Saccharo- Evidence that the GAL81 region codes for part of the gal4 myces cerevisiae GAL1-GALIO divergent promoter region: protein. J. Bacteriol. 141: 508-527. Location and function of the upstream activating sequence McKnight, S. and R. Tjian. 1986. Transcriptional selectivity of UASG. Mol. Cell. Biol. 4: 2467-2478. viral genes in mammalian cells. Cell 46: 795-805. Wright, C.F. and R.S. Zitomer. 1985. Point mutations implicate McNeil, J.B. and M. Smith. 1986. Transcription initiation of the repeated sequences as essential elements of the CYC7 nega- Saccharomyces cerevisiae iso-l-cytochrome c gene. ]. Mol. tive upstream site in Saccharomyces cerevisiae. Mol. Cell. Biol. 187: 363- 378. Biol. 5:2951-2958. Miller, A.M., V.L. Mackay, and K.A. Nasmyth. 1985. Identifica- Yocum, R.R., S. Hanley-Way, R.W. West Jr., and M. Ptashne. tion and comparison of two sequence elements that confer 1984. Use of lacZ fusions to delimit regulatory elements of cell-type specific transcription in yeast. Nature 314: 598- the inducible divergent GAL1-GALIO promoter in Saccha- 603. romyces cerevisiae. Mol. Cell. Biol. 4: 1985-1998. Nagawa, F. and G.R. Fink. 1985. The relationship between the "TATA" sequence and transcription initiation sites at the HIS4 gene of Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. 82" 8557-8561. Nasmyth, K., D. Stillman, and D. Kipling. 1987. Both positive and negative regulators of HO transcription are required for mother-cell-specific mating-type switching in yeast. Cell 48: 579-587. Pfeifer, K., B. Arcangioli, and L. Guarente. 1987. Yeast HAP1 activator competes with the factor RC2 for binding to the upstream activation site UAS1 of the CYC1 gene. Cell 49: 9-18. Post-Beittenmiller, M.A., R.W. Hamilton, and J.E. Hopper. 1984. Regulation of basal and induced levels of the MEL1 transcript in Saccharomyces cerevisiae. Mol. Cell. Biol. 4: 1238-1245. Ptashne, M. 1986. Gene regulation by protein acting nearby and at a distance. Nature 322: 697-701. Rothstein, R.J. 1983. One-step gene disruption in yeast. Methods Enzymol. 101: 202-211. Selleck, S. and J. Majors. 1987a. Photofootprinting in vivo de- tects transcription-dependent changes in yeast TATA boxes. Nature 325:173-177. 1987b. In vivo DNA-binding properties of a yeast transcription activator protein. Mol. Cell. Biol. 7: 3260-3267. Sherman, F., G.R. Fink, and J.B. Hicks. 1986. Methods in yeast genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. Shimada, H. and T. Fukasawa. 1985. Controlled transcription of the yeast regulatory gene GAL80. Gene 39: 1-9.

GENES & DEVELOPMENT 1131 Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press

GAL1-GAL10 divergent promoter region of Saccharomyces cerevisiae contains negative control elements in addition to functionally separate and possibly overlapping upstream activating sequences.

R W West, S M Chen, H Putz, et al.

Genes Dev. 1987, 1: Access the most recent version at doi:10.1101/gad.1.10.1118

References This article cites 54 articles, 25 of which can be accessed free at: http://genesdev.cshlp.org/content/1/10/1118.full.html#ref-list-1

License

Email Alerting Receive free email alerts when new articles cite this article - sign up in the box at the top Service right corner of the article or click here.