W R Boorstein and E A Craig  HSP70 gene of Saccharomyces cerevisiae.Structure and regulation of the SSA4:

1990, 265:18912-18921.J. Biol. Chem. 

  http://www.jbc.org/content/265/31/18912Access the most updated version of this article at

  .JBC Affinity SitesFind articles, minireviews, Reflections and Classics on similar topics on the

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/265/31/18912.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

at IND

IAN

INST

OF SC

IEN

CE

on May 6, 2014

http://ww

w.jbc.org/

Dow

nloaded from

at IND

IAN

INST

OF SC

IEN

CE

on May 6, 2014

http://ww

w.jbc.org/

Dow

nloaded from

THE JOURNAL OF BIOLOGXAL CHEMKWRY 0 1990 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 26.5, No. 31, Issue of November 5, pp. 18912-18921, 1990 Printed in Cl S. A.

Structure and Regulation of the SSA4 HSP70 Gene of Saccharomyces cerevisiae*

(Received for publication, June 4, 1990)

William R. BoorsteinS and Elizabeth A. CraigSV From the $Molecular and Cellular Biology Program and the ?lDepartment of Physiological Chemistry, University of Wisconsin, Madison, Wisconsin 53706

SSA4 is the only one of five heat-inducible HSP70 genes in Saccharom yces cereuisiae whose expression is restricted to conditions of stress. Comparison of the nucleotide sequences of the SSA4 gene with other HSP70 genes indicates that it diverged from its most closely related yeast homologues hundreds of millions of years ago. However, a high degree of identity has been maintained between Ssa4p and other yeast 70- kDa heat-shock proteins at the amino acid level sug- gesting, in light of its distinct pattern of regulation, that it performs an important function. A 44-base pair region of the SSA4 promoter containing an extended match to the conserved eukaryotic heat-shock element (HSE) is necessary and sufficient to mediate heat-in- ducible regulation. HSE~S~~ is capable of promoting only a low level of transcription under nonstress con- ditions. We present evidence in support of a revised definition of the functional HSE in S. cereuisiae, sim- ilar to the recently proposed modular Drosophila HSE. Elevated expression of several heat-shock proteins in an ssalssa2 double-mutant strain has previously been reported. The SSA4 promoter is activated in this strain. The increase in expression of SSA4 caused by deletion of these closely related genes is mediated via the same upstream activating sequences that activate transcription in response to heat shock. Activation of HSE-mediated transcription by disruption of consti- tutively expressed HSP70 genes supports an autoreg- ulatory model of control of the heat-shock response.

Cells exposed to a variety of environmental stresses, in- cluding elevated temperature, respond by rapidly synthesizing a small set of evolutionarily conserved proteins, known as the heat-shock proteins (HSPs).’ This heat-shock or stress re- sponse is a universal biological phenomenon (reviewed in

* This work was funded by United States Public Health Service grants from the National Institutes of Health (to E. A. C.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduer- tisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) 505637.

§ Supported by a United States Public Health Service training grant in molecular and cellular biology. Present address: Howard Hughes Medical Institute, Division of Biology 156-29, California Institute of Technology, Pasadena, CA 91125.

’ The abbreviations used are: HSP, heat-shock protein; hsp70, 70- kDa heat-shock protein; HSE, heat-shock element; HSF, heat-shock transcription factor; UAS, upstream activation sequence; ORF, open reading frame; bp, base pair; kb, kilobase pair; ATF, activating transcription factor.

Craig, 1985). Stress induction of HSPs is regulated primarily at the level of transcription. The mechanism by which heat- shock genes are activated has also been highly conserved (reviewed in Bienz and Pelham, 1987). Identification of re- lated heat-shock elements (HSEs) in many eukaryotic species, including yeast, plants, and mammals, has led to the deriva- tion of the HSE consensus sequence, CNNGAANNTTC- NNG, where N is any nucleotide (Pelham, 1985). Recently, the definition of HSEs in Drosophila has been revised, based on the effects of site-directed mutations (Amin et al., 1988; Xiao and Lis, 1988). Functionally defined Drosophila HSEs can be viewed as modular, consisting of at least three repeats of the sequence GAA in alternating orientation and separated from each other by two nucleotides. HSEs are specific binding sites for the heat-shock transcription factor, HSF (Top01 et al., 1985; Perisic et al., 1989; Wiederrecht et al., 1987; Wu et al., 1987). Transcriptional activation of heat-shock genes oc- curs by a posttranslational mechanism believed to involve phosphorylation of HSF (Sorger et al., 1987; Sorger and Pelham, 1988; Zimarino and Wu, 1987). Studies of Drosophila suggest that HSF/HSE-mediated control acts at an early step in transcriptional elongation (Rougvie and Lis, 1988).

The 70-kDa heat-shock proteins (hsp70s) are among the most highly conserved of all proteins. Species as distantly related as archeobacter and human contain related genes (Craig, 1985). In S. cereuisiae, a large multigene family en- codes eight hsp7Os that are both expressed and localized differentially within the cell. These HSP70 genes have been divided into four subfamilies based on conservation at the amino acid level (reviewed in Lindquist and Craig, 1988).* Genetic analyses suggest that members of the subfamilies are functionally related as well (reviewed in Craig, 1989). SSA4, the subject of this study, is one of four members of the most complex group, the SSA subfamily. While none of the indi- vidual SSA genes are required for viability, this subfamily is essential (Werner-Washburne et al., 1987). SSAl and SSAQ are the only two members of the SSA subfamily that are expressed at detectable levels in cells under nonstressed growth conditions. ssalssa2 double mutants are temperature- sensitive for growth but viable at 30 “C. ssalssa2ssa4 cells are inviable, indicating that the SSA4 gene can partially compen- sate for the ssalssa2 defects (Werner-Washburne et al., 1987).

SSA4 is the only classic HSP70 gene in yeast; it is expressed at extremely low levels under normal conditions and is rapidly and dramatically induced following heat-shock treatment. The previously analyzed SSAl and SSA3 genes are also heat- inducible; however, they differ from SSA4 in that SSAl is expressed at high basal levels and SSA3 is induced under conditions of lowered intracellular CAMP (such as starvation) (Werner-Washburne et al., 1989). Heat-shock activation of

2 W. R. Boorstein and E. A. Craig, unpublished observations.

18912

at IND

IAN

INST

OF SC

IEN

CE

on May 6, 2014

http://ww

w.jbc.org/

Dow

nloaded from

Structure and Regulation of a Yeast Heat-shock Gene 18913

the SSAl and SSA3 promoters is dependent upon upstream activating sequences (UAS) similar to the HSEs of higher eukaryotes (Boorstein and Craig, 1990b; Slater and Craig, 1987). HSEs are also involved in mediating basal expression of SSAl and enhancing the level of expression of SSA3 under conditions of lowered CAMP. The role of HSEs in promoting transcription under nonstress conditions is consistent with evidence demonstrating that yeast HSF is bound to HSEs under normal, as well as stressed conditions (Jakobsen and Pelham, 1988; Sorger and Pelham, 1987).

Here we present the initial characterization of the SSA4 gene and its transcriptional regulation. We define the pro- moter element that mediates heat-inducible regulation of SSA4 and compare its activity with that of HSEs from SSAl and SSA3, which contribute both to heat induction and to high expression levels under nonstressed conditions. We have also analyzed SSA4 promoter activity in mutant strains lack- ing constitutively expressed, closely related homologues, SSAl and SSAB. We evaluate the modular HSE concept, defined by experiments in Drosophila, in light of the data presented here and recent analyses of heat-inducible S. cere- uisiae promoters.

MATERIALS AND METHODS

S. cereuisiae Strains-DSlO (MAT a, hisd-11, hk3-15, h&3,leu2- 112, lysl, lys2, Atrpl, ura3-52) was utilized in all experiments pre- sented here unless specifically noted. DS16 is isogenic to DSlO and contains disrupted alleles of SSAl and SSAP (ssal::HZS3 and ssa2xLEU2) (Craig and Jacobsen, 1984; Werner-Washburne et al., 1987).

Plasmid Constructions-An SSA4/lacZ translational fusion (Fig. 1A) was constructed in a derivative of the promoter cloning vector pHT102 (Tu and Casadaban, 1990). The EcoRI to Sac1 fragment of pHT102 was replaced with the EcoRI to SmaI fragment of the Escherichiu coli uric operon from pAP55 (Brusilow et al., 1983) ligated to the SmaI to Sac1 fragment of pSKS105 (Casadaban et al., 1983) to generate the translational lacZ fusion vector pWB201. pWB201 in- cludes an extended polylinker (5’-AGCATTCCCGGGGATCCGT- CGACCTGCAGCCAAGCTTGGCACTGGCC-3’) immediately up- stream of lacZ, as well as uric operon sequences, to facilitate construc- tion of promoter deletions by providing an upstream buffer of non- essential vector sequences and an XhoI site to provide constant vector sequences upstream of the SSA4 deletion end points. An SphI linker (5’-pGGCATGCC-3’) was cloned into the SmaI site of the polylinker of pWB201. The 856-bp pSSA4H NlaIII fragment (Fig. 2B) was

A

nmpR

ori

CENll

FIG. 1. Plasmids utilized to identify sequences controlling SSA4 expression. A, pWB213, the plasmid containing the SSAI/ 1acZ hybrid gene (used to construct the 5’ deletion series plasmids). B, the centromeric UAS cloning vector pWB220, a derivative of pWB213. pWB220 and pWB213 differ only in the region between 1acZ and ARSl sequences. Genes and other functional plasmid se- quences are labeled in bold type; the location, size, and orientation of genes are indicated by curved arrows. The two large unshaded arcs in each plasmid represent sequences of pBR322 origin. Lightly stippled arcs represent sequences from the E. coli uric operon. SSA4 sequences are represented by solid black arcs. Restriction enzyme recognition sites used in plasmid constructions are labeled. The entire nucleotide sequence of these vectors is known with the exception of approxi- mately 400 nucleotides at each end of the CENll sequence. RI, EcoRI; Sa, SacI; Sp, SphI; X, XhoI.

subsequently inserted into this SphI site. The central four base pairs of the SalI-recognition site in the polylinker were removed with mung bean exonuclease to generate the correct reading frame (pWB209). The SSAlllacZ fusion junction was verified by DNA sequence analy- sis. Subsequently, an SphI linker (5’-pGGCATGCC-3’) was inserted into the filled in BglII site at position -594 (indicated in Fig. 2B) to provide a unique restriction site for the construction of deletion derivatives. The EcoRI site upstream of the SSA4 sequences was eliminated with mung bean exonuclease, creating pWB213 (Fig. 1A).

Deletion derivatives of the SSA4/lacZ fusion plasmid, pWB213, were created by cleavage with SphI followed by Bal31 exonuclease treatment for various time intervals, creation of blunt ends, ligation of XhoI linkers (5’-pCCTCGAGG-3’), XhoI cleavage, and unimo- lecular reclosure, joining the XhoI linker at the SSA4 deletion end point to a common XhoI site present upstream of the deleted region. pWB213A-594 was created by inserting an XhoI linker into the SphI site and recircularizing as described above, without exonuclease treat- ment. The precise position of pWB213A-594 A-234, and A-174 deletion end points were determined by DNA sequence analysis. The remaining deletion end points were mapped to within f3 nucleotides by polyacrylamide gel electrophoresis of small restriction fragments.

A centromeric UAS cloning vector, pWB220 (Fig. lB), was con- structed by replacing the SSA4 sequences of pWB213 with CYCl promoter sequences. Specifically, the small XhoI to Sac1 region of pWB213 (containing the entire SSA4 region and the amino-terminal portion of the &Z-coding region) was replaced with the correspond- ing restriction fragment of pLG669-Z (Guarente and Ptashne, 1981). The CYCI sequences in pWB220 extend from 178 bp upstream of the primary initiation site, through the amino-terminal codon. pWB220 also contains different linker and 1acZ fusion sites from those in pWB213. pWB227 and pWB227-R were constructed by inserting a synthetic double-stranded oligonucleotide matching the 44 bp indi- cated in Fig. 8A, flanked by sequences necessary to generate an XhoI site and an XhoI complementary overhang, into the XhoI site of pWB220. The pWB227 and pWB227-R inserts were verified to be identical with the native SSA4 sequence. All plasmid constructions utilized standard methods (Ausubel et al., 1987).

DNA Sequence Determination and Analysis-Appropriate regions of the SSA4H plasmid were subcloned into M13mp vectors and sequenced by the dideoxy chain termination method (as described in Ausubel et al., 1987). Sequence was obtained across all restriction sites used for cloning and on both DNA strands. DNA and protein sequence analyses were performed with University of Wisconsin Genetics Computing Group programs (Devereux et al., 1984).

mRNA Analysis-RNA was isolated by vigorous vortexing of S. cerevisiae cells suspended in 0.1 M Tris (pH 7.5), 0.1 M LiCl, 0.01 M dithiothreitol, in the presence of glass beads, 0.5% sodium dodecyl sulfate, phenol, and chloroform. Lysis was followed by multiple phenol/chloroform extractions and ethanol precipitations as de- scribed (Ellwood and Craig, 1984). Yields were determined by AzGO “~ and normalized by comparison of ethidium bromide-stained rRNA following gel electrophoresis.

A Sau3A to RsaI (position +147 to +87) fragment was used as a DNA primer for extension analysis of transcripts from the native gene. A Sau3A to AuaI (+147 to -402) fragment was used as a probe for Sl analysis (Fig. 2B). The DNA fragments were 5’ end-labeled at the Sau3A termini. The standard 17-bp lncZ sequencing primer 5’- GTAAAACGACGGCCAGT-3’ was utilized for priming extension of SSAl/lacZ fusion transcripts. DNA restriction fragments were hy- bridized to RNA in 80% formamide, 0.4 M NaCl following brief (45- s) denaturation at 80 “C; hybridization reactions were slowly cooled from 48 to 37 “C over 7 h. Oligonucleotide primers were prepared and hybridized under aqueous conditions in the presence of poly(A) RNA as carrier as described (Boorstein and Craig, 1989). 40-80 pg of total RNA were utilized per reaction. Nuclease protection samples were treated with 100 units of nuclease Sl at 28 “C for 60 min in the presence of denatured salmon sperm DNA as described (Berk and Sharp, 1977). Primer extension reactions were performed in the presence of actinomycin D as described (Boorstein and Craig, 1989). The samples were run on denaturing 8 M urea, 6% polyacrylamide gels and subsequently visualized by autoradiography.

Yeast Growth and Enzyme Assays-Transformants freshly selected on minimal medium lacking tryptophan were inoculated into YPD medium (Sherman et al., 1982) and grown (-20 h) to AeWn,,, = 0.4 f 0.05. The culture was then divided and one portion was transferred to a prewarmed flask at 39 “C while the remainder was retained at 23 “C. Samples were collected 1 h (or indicated times) following temperature upshift. Stationary phase cells were collected after 4.5-

at IND

IAN

INST

OF SC

IEN

CE

on May 6, 2014

http://ww

w.jbc.org/

Dow

nloaded from

18914 Structure and Regulation of a Yeast Heat-shock Gene

5 days of growth at 23 “C. Cells were collected in siliconized 1.5-ml microcentrifuge tubes; cell pellets were rapidly frozen in a -70 “C bath and were stored at this temperature. Optimization of growth conditions by using rich medium resulted in lower basal and higher heat-shocked levels of expression than observed in defined, plasmid- selective, medium (data not shown). Plasmid stability was routinely assessed as follows: an aliquot of cells was plated on rich medium and the fraction of resultant colonies that contained the TRPI marker carried on the plasmid was determined following replica plating to the medium lacking tryptophan. More than 80% of the cells routinely retained plasmid at the time of sample collection.

fl-Galactosidase activity was determined by microfluorometric as- say utilizing the substrate 4-methylumbelliferyl-fl-n-galactoside (to be described in detail elsewhere). Briefly, cells were suspended and permeabilized in 1 ml of Z buffer (Miller, 1972) with 0.05% sodium dodecyl sulfate, 50 mM P-mercaptoethanol, 1 mM phenylmethylsul- fonyl fluoride, vortexed vigorously, diluted to yield fluorogenic prod- uct in the linear detection range of the fluorimeter, and incubated with 1 mM 4-methylumbelliferyl-/3-n-galactoside for 15-120 min at room temperature. The reactions were performed in a final volume of 200 ~1 in 96-well microtiter plates. Fluorescence of samples with extremely low fi-galactosidase levels was enhanced by the addition of 50 ~1 of 1 M Na2C03 after 2 h of incubation. Fluorescence was determined using a Titertek Fluoroskan I fluorimeter. fi-Galactosid- ase activities are presented as (f2-l/(tz-1 x AGNnm x u x s)) x C x N, where fi-, is the increase in fluorescence between times 1 and 2, tzmL is the time in minutes, between the two fluorescence readings, Asw nm is a measure of cell density at the time of cell collection, u is the volume of the culture assayed, in milliliters, and s is correction factor for plasmid stability that is equal to the fraction of cells retaining the plasmid at the time of cell collection. The arbitrary fluorescence units were converted to Miller units (Miller, 1972) by the empirically derived constant, C = 5.7 x 10’. N is a normalization factor deter- mined by assay of a known standard quantity of fl-galactosidase to control for differences in temperature and pH which affect @-galac- tosidase activity and fluorescence of methylumbelliferone, respec- tively. Selected samples were permeabilized as described above and were assayed in parallel with the substrate o-nitrophenyl-fl-n-galac- topyranoside, as described (Slater and Craig, 1987); results of the two assays were similar, with the exception of the low activity samples that were below the accurate detection limits of the 0-nitrophenyl-@- D-galactopyranoside assay. Data presented are averages from two to six determinations of at least two independent transformants. Indi- vidual determinations were almost always within 20% of the averages shown here. Occasionally, high or low values were obtained. In these cases, samples from all treatments of a single culture were similarly affected (i.e. basal, heat-shock, and stationary phase); thus, patterns of induction were highly reproducible.

RESULTS

Characterization of the SSA4 Gene and Encoded Tran- scripts-The SSA4 gene was isolated on a 9-kb genomic clone.3 We constructed a restriction map of the pSSA4H genomic insert and localized the SSA4 coding region by DNA hybridization analysis utilizing fragments of the related SSAI gene as probes (Fig. 2A and data not shown). The nucleotide sequence of the coding region, as well as 1.2 kb of flanking DNA upstream of the gene, was determined (Fig. 28). A single large open reading frame with similarity to SSAl was identi- fied. This SSA4 ORF has the potential of encoding a protein of 642 amino acids with a predicted molecular mass of 69,650 Da and an isoelectric point of 4.86. These values are consistent with previous estimates based on polyacrylamide gel electro- phoresis analysis of the SSA4 protein, Ssa4p (Werner-Wash- burne et al., 1987).

While Ssa4p is closely related to hsp70s and noninducible cognates of distantly related species (68-75% identity to Dro- sophila and human homologues), it is most closely related to the yeast hsp70 Ssalp and Ssa2p, which are 97% identical. Ssalp and Ssalp are each 642 amino acids in length, although optimal alignment involves the introduction of at least three

“J. C. A. Bardwell and E. A. Craig, unpublished data.

FIG. 2. Restriction map and nucleotide sequence of SSA4 and flanking genomic DNA. A, restriction map of the 9.0-kb SSA4 Hind111 fragment, SSA4H. The SSA4 protein-coding region and the upstream open reading frame are represented by arrows. The filled region of the map indicates the part of the clone whose DNA sequence

at IND

IAN

INST

OF SC

IEN

CE

on May 6, 2014

http://ww

w.jbc.org/

Dow

nloaded from

Structure and Regulation of a Yeast Heat-shock Gene 18915

FIG. 3. Comparison of predicted complete SSA4 and SSAI hsp70 protein sequences. Solid bars, colons, or dots between cor- responding amino acids of the two proteins indicate identity, or high or low levels of similarity, respectively. This alignment was obtained by optimizing matches while introducing the smallest number of gaps (represented by dots in the sequence lines) according to the algorithm of Needleman and Wunsch (1970). Matches between nonidentical amino acids were weighted as described by Gribscov and Burgess (1986). Alternative alignments can be obtained with a larger number of gaps, particularly in the C-terminal region, but these do not significantly increase the similarity. Ssalp is from Slater and Craig (1989).

gaps, indicating that insertion/deletion mutations occurred during the divergence of these genes (Fig. 3). Ssalp and Ssa4p are 82% identical; they are 90% similar, accounting for con- servative amino acid substitutions as described by Gribscov and Burgess (1986). There is an abrupt decrease in similarity near the carboxyl termini of the proteins. The amino-terminal 525 amino acids of Ssa4p are 89% identical with Ssalp, while the carboxyl-terminal 117 amino acids share only 50% iden-

is shown in B. B, Rⅈ Bs. BsrEII; H. HindIII: HD, HnaI: K, KnnI: M, MluI; N, NdeI; p, I%tI; Pv, Pⅈ RI, EcoRI;~$ salI;’ Sk, SacIt SC, ScaI; S,o, SphI; St, StuI; X, XhoI. The following enzymes do not cleave the SSA4H fragment: ApaI, BarnHI, BclI, NruI, PuuI, SacII, SmaI. B, nucleotide sequence of SSA4 and flanking DNA. The 5’ ends of the heat-shock SSA4 mRNA map to the double-overlined nucleotides; the center of this region has been designated +l. The arrowhead at position 32 marks the putative initiation site for SSA4 transcription under control conditions (see the legend to Fig. 5). Long bars above the sequence indicate occurrences of matches to the canonical heat-shock consensus sequence, CNNGAANNTTCNNG, where N is any nucleotide, in at least six of the eight conserved positions. The only match to the extended modular HSE definition is in the region of three overlapping canonical HSEs centered at -188. The putative TATA element and the initiation and termination codons are enclosed in boxes. A match to the ATF consensus binding site is underscored with a broken line (positions -103 through -97). Positions of restriction enzyme recognition sites used in cloning and in construction of primers and probes for mRNA mapping are indi- cated. The amino acid translation of the incomplete ORF upstream of SSA4 is included. The translation of the SSA4 ORF is shown in Fig. 3.

tity. However, two short highly conserved domains are present within the C-terminal region; the eight extreme terminal amino acids of Ssa4p are identical to those of Drosophila hsp70s, as well as to those of Ssalp. 85% of the amino acids between these two conserved domains (610 through 636) are glycine, alanine, and proline, including six glycine-alanine repeats; therefore, this region is likely to be highly flexible.

The nucleotide sequence similarity is strictly limited to the coding region (Fig. 4), in which the genes share 67% identity. The high degree of divergence between SSA4 and SSAl based on silent (synonymous) substitutions (182%, calculated and corrected for multiple substitution events by the method of Perler et al. (1980)) indicates that these genes are likely to have arisen from a duplication event a minimum of 100 million years ago. The relatively low divergence, 13%, based on replacement (nonsynonymous) substitutions indicates that a high level of selective pressure has maintained the similarity in the coding potential of these two genes.

An additional open reading frame of at least 434 nucleotides was identified on the same strand as SSAC (Fig. 2, and data not shown). This ORF begins 5’ to the region included in Fig. 2B and terminates 889 bp upstream of the SSA4 coding domain. The upstream ORF is predicted to encode a protein by the TESTCODE algorithm, an empirically derived method based primarily on the periodic compositional bias of coding sequences (Fickett, 1982). The ORF sequence is not closely

TAA I’

00 05 10 1.5 20

SSA4 distance from transcription slaR (kb)

SSAI 2000 .ccucGGticMcroo ~=T~~TAA;rT1\CA~~T~~~~T~~~

IIIII IIIIIIII IIIIIIII III III CCUCCCTTGAAGAAG -=WGCCMTTGGTGCGGC~TTGAT:

ssA.1 2012

FIG. 4. Comparison of SSA4 to SSAI DNA sequences. A, graphic matrix comparison of the entire SSA4 and SSAl protein- coding regions plus flanking DNA (Maize1 and Lenk, 1981). Wide bars on the ares indicate the positions of the coding regions; initiation and termination codons are labeled. Each dot represents a match of 17 of 24 adjacent nucleotides between the two sequences. Diagonal lines indicate contiguous colinear similarity between the two genes. B, alignment of SSA4 and SSAl nucleotide sequences flanking the initiation (top) and termination (bottom) codons. Vertical lines indi- cate identity between nucleotides at corresponding positions in the two genes. Coding portions of the genes are in bold type. Numbers indicate positions, as in Fig. 2B, for SSA4 and, as in Slater and Craig (1987), for SSAI.

at IND

IAN

INST

OF SC

IEN

CE

on May 6, 2014

http://ww

w.jbc.org/

Dow

nloaded from

18916 Structure and Regulation of a Yeast Heat-shock Gene

related to entries in GenBank, EMBL, and NBRF data bases. Mapping the 5’ Termini of SSA4 mRNA and Identification

of Putative SSA4 Transcriptional Regulatory Sequences-The 5’ ends of SSA4 transcripts were mapped to the gene by primer extension and sl nuclease techniques (Fig. 5). Heat- induced transcripts from the native gene mapped to a five- nucleotide region. The HSP70-related open reading frame begins at the first AUG downstream of this putative initiation site. The region flanking this AUG matches the S. cereuisiae translational initiation consensus sequence in 8 of 10 posi- tions (Cigan and Donahue, 1987). The length, 53 nucleotides, and the adenine-rich content of the predicted 5’ untranslated SSA4 mRNA leader are typical of yeast genes transcribed by RNA polymerase II.

The sequence TATAAA, which matches the functionally defined TATA element (Chen and Struhl, 1988), is present 75 bp upstream of the putative transcriptional initiation site(s). Interestingly, several occurrences of sequences that have been shown to act as preferred initiation sites in CYCl, PH05, and other yeast promoters, TC(G/A)A and RRYRR (where R and Y are purine and pyrimidine nucleotides, re- spectively) (Hahn et al., 1985; Rudolph and Hinnen, 1987),

A.

Extension Sl Nuclease -- Ss’123456 ----- ---

B. C A-234 A-594 Standards

HS:- - - + - + 113456J-8910

E

I

-*

l -c

- r 104-5 * ma

e 0 l w c 145-9

e.

FIG. 5. Analysis of SSA4 RNA. A, Sl and primer extension analvses of the native SSA4 RNA isolated from heat-shocked DSlO. S and S’ are DNA size standards (pBR322 cleaved with MspI and HaeIII, respectively). Lanes 1 and 6, controls containing the same quantities of primer and probe, respectively, used in the mapping reactions; Lanes 2 and 5, reaction controls containing tRNA, but no yeast mRNA; Lanes 3 and 4, SSA4 primer extension and Sl nuclease mapping reactions with 13 fig of total RNA per lane. S1 probes and extension primers were labeled at a common 5’ terminus, therefore, bona fide 5’ ends mapped by each method should yield comigrating products. Unextended primer and undigested probe bands are indi- cated by arrowheads to the right of A. Higher resolution electropho- resis of aliquots of the samples shown here suggest the presence of a single predominant 5’ terminus; the less abundant products from the two reactions do not comigrate (data not shown). B, primer extensions of SSA4/lacZ fusion RNAs from pWB213n-234 and J-594 trans- formants in DSlO under control (-) and heat-shock (+) conditions (Lanes 3-6). Lanes 1 and 2 contain controls: primer only (no reaction) and primer plus poly(A) RNA (no yeast RNA), respectively. A dideoxy DNA-sequencing ladder is included as a size standard in Lanes 7-10. The 5’ termini of native and fusion transcripts were indistinguishable. The range of sizes of Sl nuclease and primer extension products are indicated in nucleotides. Comigrating single faint bands were detected on long exposures of extension products from nonheat-shocked pWB2131-234 and pWB213J-594 transformants (data not shown). These may represent a very low level of transcriptional initiation from position -+32 (as marked in Fig. 2H).

occur near the actual initiation sites; however, these do not appear to be active sites of transcriptional initiation.

Six matches to the canonical dyad heat-shock element CNNGAANNTTCNNG, where N is any nucleotide (Pelham, 1985), occur upstream of the transcribed region of SSA4. Three of these, including the two closest matches (seven of eight) to the consensus, are present in an overlapping arrange- ment centered at position -188. The -188 region contains four perfect plus three imperfect GAA blocks with relative spacing and orientation consistent with the extended modular HSE definition from Drosophila (Amin et al., 1988). In addi- tion, an exact match to the ATF consensus binding site, (G/ T)(A/T)CGTCA, occurs at position -100, 17 bp upstream of the TATA sequence (reviewed in Jones et al., 1988). Very similar sequences have been shown to bind yeast ATF and exhibit UAS function (Kornuc et al., 1988; Jones and Jones, 1989; Lin and Green, 1989). It is intriguing that an HSP70 gene from both yeast and human genomes contain putative ATF-binding sites, although the function of these sequences in heat-shock regulation is not known (Greene et al., 1987)

Expression of an SSA4/lacZ Fusion Gene-To facilitate identification of SSA4 transcriptional regulatory sequences, SSA4 DNA from -801 to the initiation codon was fused to the 1acZ gene of E. coli. The fusion gene included three HSE- like regions, the putative TATA element, the transcriptional initiation region, and the entire untranslated leader from SSA4. @-Galactosidase activity from cells transformed with this construct was very low under optimal growth conditions at 23 “C. Activity increased 70-fold following 39 “C heat-shock treatment (Fig. 6, top line). Basal activity was only slightly (0.8 Miller units) higher in stationary phase than in exponen- tially growing cells at 23 “C.

Localization of Regulatory Sequences-A progressive series of deletions of SSA4 upstream sequence was constructed to delineate regions involved in transcriptional control. The deleted fusion constructs retaining sequences from at least -234 to +59 were regulated in a heat-inducible manner (Fig. 6). The largest and smallest deletion derivatives of the SSA4/ 1acZ hybrid gene that gave rise to heat-inducible P-galactosid- ase activity were analyzed in greater detail. Both constructs exhibited a high degree of heat-shock regulation at the RNA level (Fig. 5B). Furthermore, the 5’ termini of the fusion mRNAs mapped to the same nucleotides in the SSA4 se- quence as did the native SSA4 transcript termini (Fig. 5B). P-Galactosidase activity from both constructs increased dra- matically within 30 min of temperature upshift, reached a maximum after 60 min, and then gradually declined (Fig. 7). Therefore, expression of the fusion genes corresponds to that of the native SSA4 gene.

Removal of an additional 60 bp downstream of the -234 deletion endpoint essentially abolished heat-inducible regu- lation, indicating that the -188 HSE-like region is critical to SSA4 heat induction. This region contains the only match to the newly extended HSE definition within 1.2 kb upstream of the SSA4 gene, as well as the closest matches to the canonical HSE 14-bp palindromic sequence. In addition, /3- galactosidase activity under basal (23 “C) conditions is 3-fold lower from constructs lacking the -188 HSE region than from slightly larger constructs that retain these HSE-like se- quences, suggesting that the HSE sequences may be capable of mediating a low level of expression under optimal condi- tions, as well as increased expression under conditions of stress. The deletion analysis also demonstrates that the HSE match at position -549 is not essential for heat-shock regu- lation and that both the ATF- and HSE-like sequences at -100 and -14, respectively, are not sufficient to drive heat-

at IND

IAN

INST

OF SC

IEN

CE

on May 6, 2014

http://ww

w.jbc.org/

Dow

nloaded from

Structure and Regulation of a Yeast Heat-shock Gene

f3-galactosidase activity

-801 -188 tl ATG 23X T39"C f3943' - -- 0.2 14 70

0.2 12 60

0.1 10 100

0.1 10 100

0.2 11 55

0.7 17 24

0.6 15 25

0.2 0.4 2.0

0.2 0.6 3.0

0.3 0.3 1.0

Vector SSA4 IacZ FIG. 6. ,%Galactosidase activities from the SSAsCllacZ fusion gene and the promoter deletion deriv-

atives. The solid black lines represent SSA4 sequences. The positions of the initiation codon, the transcription initiation site, and the deletion end points are indicated, pWB21311-234 contains 1 bp beyond that indicated identical with that in the native gene because of the addition of linker sequences. Diamonds represent partial matches to the HSE consensus, as overlined in Fig. 2B. The putative TATA element is denoted by T. The plasmid on the top line, pWB213, is the parental plasmid from which the deletions were derived as described under “Materials and Methods.” Deletion derivative plasmids, from top to bottom, are named pWB213A-594 to pWB213A-61. pWB213 contains an 8-bp linker inserted at position -594 of the SSA4 DNA as described under “Materials and Methods.” pWB213 also contains 665 bp of vector sequence immediately upstream of SSA4 that are not present in the deletion derivatives. All other plasmids represented here differ from each other only in the extent of 5’ noncoding SSA4 sequence included (and consequently the relative spacing of DNA sequences flanking the deletion end points and the actual junction sequences). P-Galactosidase activities from transformants growing exponentially at 23 “C and following a l-h 39 “C heat-shock treatment, are presented in Miller units.

18917

inducible expression, at least from their native locations. An increase in both basal and induced levels of expression upon removal of sequences between -366 and -240 suggests ele- ments that negatively modulate SSA4 expression may be present in this region. However, no match to the upstream repression sequence that negatively modulates SSA 1 tran- scription was identified in the SSA4 promoter (Park and Craig, 1989).

Definition of UASHs-To determine whether the sequences shown to be required for heat-shock activity of the SSA4 promoter by 5’ deletion analysis were sufficient to mediate heat-inducible transcription in the absence of flanking SSA4 sequences, a short region containing the HSEs centered at -188 was tested for upstream activation sequence activity. The vector employed for these experiments, pWB220, is a derivative of the SSA4/lucZ deletion plasmids in which the SSA4 sequences were replaced with DNA from the down- stream portion of the CYCl promoter that includes TATA elements and initiation sites, but lacks upstream activation sequences. Thus, expression of this la& fusion is dependent upon UAS activity from DNA inserted into a cloning site adjacent to the upstream boundary of CYCl sequences. An oligonucleotide identical with 44 bp encompassing the HSE- like sequences at -188 was synthesized and cloned into the UAS test vector (Fig. 8). The insertion of this HSEs,, oligomer upstream of the CYCl/lucZ fusion resulted in a 4- fold increase in basal levels of activity of the CYCl/lucZ hybrid protein. Following heat shock, the heterologous pro- moter mediated a further 12-fold increase in activity. The kinetics of the change in /3-galactosidase activity following temperature upshift mediated by the hybrid SSA4-oligonucle- otide/CYCl promoter was similar to that from the “complete” SSA4 promoter (Fig. 7). Since the 44-bp SSA4 sequence was able to mediate heat-inducible transcription in either orien- tation (Fig. 8B), we designate it UASHs.

UAS, Mediates High Levels of Transcription in ssalssa2 Mutants-SSAl and SSAP are the only SSA genes actively transcribed in cells under normal growth conditions. Viability of ssalssa2 double-mutant strains at 23 “C! depends on the

presence of an intact SSA4 gene, even though SSA4 is not expressed in wild-type cells growing at this temperature. The SSA4 protein has been shown to be abundant in ssalssa2 cells (Werner-Washburne et al., 1987). To determine if the SSA4 promoter exhibits altered regulation in ssalssa2 cells at 23 “C, expression of the A-594 construct, which is regulated like the native SSA4 gene in wild-type cells, was assayed in a double- mutant background. p-Galactosidase activity from the SSA4/ la&Z fusion in an ssalssa2 double-mutant strain was 600-fold greater than levels in wild-type isogenic strains during growth at 23 “C (Table I). To determine whether UA&s was alone sufficient to drive this abnormally high basal activity, expres- sion driven by the UAS&CYCl hybrid promoter was assayed in cells with disrupted SSAl and SSA2 genes. The UAS cloning vector, pWB220, yielded an extremely low basal level of P-galactosidase activity, 0.3 units, in the double-mutant strain. Insertion of the 44-bp HSESsA4 oligonucleotide into the UAS cloning site of this vector resulted in 185 units of p- galactosidase activity under the same growth conditions (Table I). This confirms that the constitutive activation in the mutant strains is at the level of transcription and dem- onstrates that a common upstream activating sequence, UASHS, mediates heat-inducible expression in wild-type strains and high basal levels of transcription in cells lacking SSAl and SSAB.

Neither the SSA4jlacZ fusion nor the UAS&CYCl hybrid promoter were activated to higher levels by a 39 “C! heat shock in the ssalssa2 background, although both constructs exhib- ited induction in wild-type cells following heat shock. Heat- shock treatment may have failed to activate the SSA4 pro- moter over the high basal expression levels in ssalssa2 mu- tants because the promoter was already maximally induced. Alternatively, the prolonged activation of the stress response may have induced a steady state that was unable to readily respond to increased stress.

DISCUSSION

We have localized the heat-inducible UAS of SSA4 to a 44- bp region containing extensive similarity to sequences that

at IND

IAN

INST

OF SC

IEN

CE

on May 6, 2014

http://ww

w.jbc.org/

Dow

nloaded from

18918 Structure and Regulation of a Yeast Heat-shock Gene

A.

15-- B

A-594

UAS,

0 60 li0 180 240 300 360 Time (minutes)

FIG. 7. &Galactosidase activity from transformants of SSAl/lacZ fusion plasmids under heat-shock and control con- ditions. Each pair of curues represents data from a single exponen- tially growing culture at 23 “C; at time 0, an experimental aliquot was transferred to 39 “C (solid black symbols) while a control was main- tained at 23 “C (open symbols). Samples were collected at the indi- cated times and fl-galactosidase activities were determined. &rues shown are representative for each plasmid. /3-Galactosidase activity is presented in Miller units. A, pWB213A-594. B, pWB213A-234. C, pWB227 (circles), the HSEssar 44-bp oligonucleotide in pWB220, as well as a heat-shock curve only from the UAS vector, pWB220, with no insert (triangles). The plasmids are illustrated in Figs. 6 and 8B.

regulate stress-induced transcription of heat-shock genes from distantly related eukaryotic species. These HSEaa,+, sequences alone mediate a very similar pattern of transcrip- tion to those regulated by the isolated HSEs of SSA3 and SSAl (Boorstein and Craig, 1990b; Slater and Craig, 1987). Therefore, the role of HSEs in contributing to the high expression of the SSAl and SSA3 promoters under nonstress conditions (in which the HSE-containing SSA4 promoter is not active) appears to result from differences in regions of the promoters distinct from the HSEs (Slater and Craig, 1987; Boorstein and Craig, 1990a). For example, expression of SSA3 in stationary phase is dependent upon a distinct promoter element, UASPDS, but is further enhanced by HSEssa3 (Boor- stein and Craig, 1990a).

The difference in regulation between the SSA subfamily genes, resulting from different promoter composition, may provide an explanation for the maintenance of similar genes over a long period of evolutionary time; the gene products may be functionally equivalent, multiple genes being required for an appropriate pattern of expression and/or rapidity of

B.

Insert

none

44-mer

P-galactosidase activity

Iz!z!zs 3%

0.3 0.2 0.7

1.1 13 12

44-mer/Reverse *3ocr FIG. 8. The SSA4 HSE region centereyat -lk. A,‘;NA

sequence of the 44-bp HSEsa..,, oligonucleotide. Diamonds over the sequence indicate overlapping matches to the canonical HSE se- quence. Triangles below the sequence mark appropriately positioned matches to the repeating core units of the revised HSE definition; filled triangles represent exact matches and open triangles mark sequences with a single mismatch to GAA (in either orientation). B, /3-galactosidase activity from the CYCl/lacZ UAS cloning vector alone (pWB220) or with the 44-bp oligonucleotide (shown in A) inserted at the XhoI site of CYCI, 178 bp upstream of the transcrip- tional initiation site. pWB227 and pWB227-R contain the SSA4 oligonucleotide in native and reverse orientations, respectively, rela- tive to the orientation in the SSA4 promoter. The plasmids are diagrammatically represented as in Fig. 6. These plasmids are cen- tromeric and differ from those in Fig. 6 only in the region depicted here. &Galactosidase activities from transformants growing exponen- tially at 23 ‘C and following a l-h 39 ‘C heat-shock treatment are presented in Miller units.

TABLE I

Activity of the SSA4 promoter and UASHs in an ssalssa:! versus an SSAlSSA2 strain

Hybrid gene

&Galactosidase activity”

ssolssa2 SSAlSSAP

23 “C t39 “C 23 ‘C

SSAl/lacZ A-594’

UAS/CYCl/lacZ No insert’ HSEd

122 122 0.2

0.3 0.2 0.2 185 162 1.1

o /3-Galactosidase activity was determined in DS16 and DSlO under indicated conditions as described in Fig. 6.

* pWB213A-594. The promoter region of this plasmid is depicted in Fig. 6.

’ pWB220. The promoter region of this plasmid is depicted in Fig. 8B.

d pWB227. The pWB220 derivative containing the 44-bp HSE,sA, oligonucleotide in the native orientation, as depicted in Fig. 8B.

induction under certain circumstances. However, it is possible that the amino acid differences between the proteins, perhaps in the divergent C termini, may confer distinct functions, or at least specialized abilities to perform a common function under different environmental conditions.

HSEs in Yeast-The functional Drosophila HSE is com- prised of at least three inverted GAA modules separated from each other by 2 bp (Amin et al., 1988; Xiao and Lis, 1988). The three modules need not be contiguous, a gap of precisely one module width (5 bp) is tolerated, provided the modules flanking the gap are direct repeats (i.e. oriented as if the gap contained a GAA sequence). All seven heat-shock elements that have been functionally defined from heat-inducible yeast

at IND

IAN

INST

OF SC

IEN

CE

on May 6, 2014

http://ww

w.jbc.org/

Dow

nloaded from

Structure and Regulation of a Yeast Heat-shock Gene 18919

genes match the extended modular HSE definition derived from analysis of a Drosophila HSP70 promoter. The yeast HSEs are comprised of at least three and as many as five appropriately positioned GAA modules spanning up to four turns of the DNA helix (Table II). The functionally defined HSE of SSA4 meets the criteria of the modular HSE defini- tion; sequences upstream of the transcribed portion of the SSA4 gene that match the canonical HSE but not the modular HSE appear to be nonfunctional. Strong matches to the modular HSE are also found upstream of heat-inducible yeast genes whose promoters have not yet been analyzed (Table II: KARP and SZ’II). Yeast HSE regions also contain as many as four appropriately positioned sequences with a single mis- match to the core consensus GAA (TTC) in addition to the perfect matches; one such imperfect core sequence, TTA, is

present with particularly high frequency (for example SSA3; see Table II). These occurrences of imperfect GAA blocks suggest that they function as components of HSF-binding sites. In fact, in vitro binding experiments indicated that HSF contacts at least some imperfect GAA blocks (Shuey and Parker, 1986).

Direct evidence supporting an extended HSE definition in yeast comes from reevaluation of previous analysis of HSE2ssa,. An oligonucleotide containing 22 bp of the HSE2ssa, region contains three GAA blocks and is sufficient to mediate heat-inducible transcription (Table II, line 3). However, a shorter (15bp) region that includes the identical 14-bp match with the canonical HSE, but only two of the three GAA modules, was unable to function as a UAS (Table II, line 11). Interestingly, an even shorter HSEBssAl oligonu-

TABLE II Heat-shock elements from Saccharomyces cerevisiae genes

Sources: line 1, this work; line 2, Boorstein and Craig, 1990b; lines 3 and 4, Slater and Craig, 1987, and Wiederrecht et al., 1987; line 5, Farrelly and Finkelstein, 1984, and McDaniel et al., 1989; line 6, Piper et al., 1988; line 7, Tanaka et al., 1989, and Ozkaynak et al., 1987; line 9, Rose et al., 1989; line 10, Nicolet and Craig, 1989; line 11, Slater, 1986; line 12, Slater and Craig, 1987.

Gene Function" sequence* 1. SSA4

2. SSAJ

3. SSAZ-HSEP (22-mer’)

4. SSAl-HSE3

5. HSP82

6. PGK

7. iJBI4

8. Consensus’ Frequent:

Core:

Rare:

+ CAATGAAGTACAT-&TAGAAGT~TAGAACCTTATGGAAGCAC - = --- + TAATTAGGGATCGCTGTGGAAAGTTATAGAATATTACAGAAGCAG -- ----- -

+, B TTTT&ZAGA&ZGECATCGGC B TGTAAACTZCAGAACATTCTAGAAAGA

B, M AGTTTCGCAGAACTTTTT=TT;CTTTTTTTCTAGAACGCCGTGGAAGAA --- -- - - -- = M ATCGAAGGTTCTGGAATGGCGGGAAAGGGTTTAGTACCA M TGTGTAEZGTTCTAGAATAATCCTGGATAA ~~~~ --

TA cg TA TA cg Examples : ..E..TTC..GAA, ..TTC..GAA..TTC == ==-

RY wy RY ac RY wy

9. KAR2 ND ATAGAACCTTCTGGAAATTTCACC - 10. STIl ND ATC~GTGAAGATTCGT~GT~TAGAACA~CAGAAAAA -c- -cc 11. SSAI-HSE2 (15-mei) - ttaggctcgaCCAGj&CGCGTTCCATCtcgagcaga

12. SSAI-HSE2+ (9-mer’) + aggctcgAG&CGEtcgagcag

’ + or -, sequence is or is not sufficient to function as a heat-inducible UAS, respectively; M, sequence has been shown to function in heat-inducible regulation by mutational analysis; B, sequence is known to act as a protein- binding site, and in particular as a site of HSF binding in the case of lines 3 and 4; ND, function of HSE-like region of these heat-inducible promoters has not been determined.

* HSE region sequences are shown, including all GAA modules, in alternating orientation and separated by two nucleotides (allowing for a gap of one module, provided GAA elements flanking the gap are positioned as if the gap contained a GAA sequence) according to the HSE definition of Amin et al. (1988). All sequences are from within 400 bp upstream of the initiation codons. Plus strands are shown. Exact matches and single mismatches to the consensus core module sequence (GAA and TTC) are double- and single-underlined, respectively. Lines l-3, 11, 12: large uppercase characters represent the full extent of native sequences tested for UASns activity; line 2: smaller uppercase letters represent native SSA3 sequence beyond the region shown to be sufficient for UASns function; lines 11 and 12: lowercase characters represent vector sequences flanking the SSAI-HSE2 sequences in the UAS test constructs.

’ SSAI-HSE2 22-mer, 15mer, and 9-mer are from pZJHSE2-26, pZJHSE2-19, pZJHSE2-12, respectively. ‘Two example consensus sequences, each consisting of three adjacent exact GAA modules, illustrate the

sequence bias at positions between GAA blocks. Nucleotides that occur frequently are indicated on the upper line, whereas nucleotides that occur rarely, if at all, are shown below the core modules. Consensus sequences were derived from the plus strand of functionally defined HSEs (lines l-7) only. Consensus nucleotides upstream of exact GAA blocks that are also downstream of exact TTC blocks (TTC--GAA) were derived only from comparably positioned nucleotides in the native sequences. Consensus nucleotides upstream of exact GAA blocks, but not between two adjacent blocks (NNN--GAA) were derived from nucleotides found upstream of all exact GAA modules. Sequence bias at positions upstream of TTC sequences was determined as described for positions upstream of GAA sequences. No strong bias was observed in positions downstream of exact GAA or TTC modules that are not also upstream of exact core modules. For line 8, R represents G or A; Y represents C or T; W represents A or T. Uppercase letters in the consensus sequence indicate a greater degree of sequence bias than lowercase letters. The requirements for a functional HSE are discussed in the text. As indicated by the sequences in lines l-7, arrangements of GAA and GAA-like core modules other than those shown in the consensuses allow HSE function.

at IND

IAN

INST

OF SC

IEN

CE

on May 6, 2014

http://ww

w.jbc.org/

Dow

nloaded from

18920 Structure and Regulation of a Yeast Heat-shock Gene

cleotide was able to function efficiently as a heat-inducible UAS; the vector nucleotides flanking the short SSAl insert created appropriately oriented and positioned matches (2 of 3 bp) to the GAA module on either side of the SSAl oligo- nucleotide (Table II, line 12). This result further supports the idea that partial matches to the core consensus, GAA, can contribute to HSE function in yeast.

The base pairs flanking highly conserved core GAA modules of an HSE are also conserved, but to a lesser extent. The bias of nucleotides between GAA blocks is indicated in line 8 of Table II for two model HSEs that each consist of three adjacent GAA repeats. The bias is strongest upstream of 5’- GAA-3’ sequences on the plus strand, particularly between two exact consensus modules (TTC--GAA). The sequence bias we observed between GAA blocks differs somewhat from that in HSE-like regions from heat-shock genes of higher eukaryotic species (Amin et al., 1988). Most notably, G was not present in the position two nucleotides upstream of GAA in any of the 16 perfect or imperfect core GAA modules in the region of functional HSEs (Table II); Amin et al. (1988) observed that G occurs frequently in this position of HSEs from Drosophila and other species and demonstrated its abil- ity to contribute to HSE function in Drosophila cells.

via sequences distinct from the HSEs (Stone and Craig, 1990). Therefore, it is likely that there are other components, pos- sibly additional HSPs, in the control loop modulating the HSF activity state. In E. coli, mutations in any of three HSP genes, grpE, dnaJ, or the HSP70 homologue, dnaK, result in cr3’-dependent increased transcription of heat-shock genes4 (Tilly et al., 1983). Since DnaK can interact directly with DnaJ and GrpE (Sell, 1987; Johnson et al., 1989), analogues of the latter two proteins may interact with hsp70 in yeast to control the activation state of HSF. Perhans the mechanism

It is clear from examination of the sequences in Table II and the analysis of HSE-containing constructs that HSEs of different yeast genes share many common features. However, the flexibility of the functional HSE sequence is also striking (Wei et al., 1986). The required degree of specificity for HSF binding appears to be achieved by a long but degenerate recognition site, rather than a short strictly conserved UAS as, for example, is the case for Gcn4-binding sites (Hill et al., 1986). Perhaps differences between sequence composition of UASHs elements allows a single transcription factor, HSF, to stimulate transcription of many genes at different relative levels in response to a common signal.

Autoregulation of the Heat-shock Response-Transcription of heat-shock genes has been proposed to be negatively regu- lated by heat-shock proteins. This autoregulatory hypothesis is based on studies in which blocked translation caused ab- normally high levels of heat-shock gene transcription follow- ing stress exposure (McAlister and Finkelstein, 1980; Di- Domenico et al., 1982; Plesset et al., 1982). We have shown that disruption of two constitutively expressed HSP70 genes causes increased transcription of the heat-inducible SSA4 gene. Because the same short UASns element is necessary and sufficient to mediate high basal expression in an ssalssa2 strain as mediates stress-activated transcription in wild-type strains, we surmise that the general stress response is consti- tutively activated in the double-mutant strain. Additional mutant phenotypes of ssalssa2 strains, including constitutive thermotolerance and active synthesis of Hsp82, support this conclusion (Craig and Jacobsen, 1984).

These results suggest that SSA proteins function in a negative feedback loop, regulating the transcription of heat- shock genes. Thus, hsp70 is critical to maintenance of the heat-shock response in the OFF (uninduced) state, and a decrease in hsp70 availability, either by disruption of the constitutively expressed HSP70 genes or the increased de- mand for hsp7Os under conditions of stress, is sufficient to activate the response. HSF may be modulated by hsp70 di- rectly or indirectly via proposed activities in protein folding and mediation of interactions between polypeptides (reviewed by Lindquist and Craig, 1988). Overexpression of the SSAl HSP70 gene (from a heterologous promoter) does not appear to decrease either basal or heat-induced HSF/HSE-mediated transcription, although it does repress SSAl promoter activity

by which the primary intracellular stress stimulus is trans- duced, resulting in induction of the stress response, is con- served between prokaryotes and eukaryotes, although the mechanism of activating RNA polymerase on heat-shock promoters has diverged. Elucidation of the mechanisms by which HSP70 expression controls transcription of heat-shock genes will require further definition of the function of the HSP70 gene products and the means by which the stress pathway is activated.

Acknowledgments-We wish to thank Jeffrey Shilling and the University of Wisconsin Biotechnology Center for sequencing por- tions of the SSA4 gene and Diane Gambill, Lynn Manseau, and Carolyn Norris for comments on the manuscript. We also thank Helen Tu for pHT102 and Dave Stone for the yeast strains used in this study.

REFERENCES

Amin, J., Ananthan, J., and Voellmy, R. (1988) Mol. Cell. Biol. 8, 3761-3769

Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Smith, J. A., and Struhl, K. (1987) Current Protocols in Molecular Biology, John Wiley and Sons, Inc., New York

Berk, A. J., and Sharp, P. A. (1977) Cell 12, 721-732 Bienz, M., and Pelham, H. R. B. (1987) Adv. &net. 24,31-72 Boorstein, W. R., and Craig, E. A. (1989) Methods Enzymol. 180,

347-369 Boorstein, W. R., and Craig, E. A. (1990a) EMBO J. 9, 2543-2553 Boorstein, W. R., and Craig, E. A. (199Ob) Mol. Cell. Biol. 10, 3262-

3267 Brusilow, W. S. A., Porter, A. C. G., and Simoni, R. D. (1983) J.

Bacterial. 155, 1265-1270 Casadaban, M. J., Martinez-Arias, A., Shapira, S. K., and Chou, J.

(1983) Methods Enzymol. 100,293-308 Chen, W., and Struhl, K. (1988) Proc. N&l. Acad. Sci. U. S. A. 85,

2691-2695 Cigan, M. A., and Donahue, T. F. (1987) Gene (Am&) 59, l-18 Craig, E. A. (1985) Crit. Rev. Biochem. 18, 239-280 Craig, E. A. (1989) Bioessays 11, 48-52 Craig; E. A., and Jacobsen: K. (1984) Cell 38,841-849 Craig. E. A.. Kramer. J.. Shilline. J.. Werner-Washburne. M., Holmes.

S.:Kosic:Smithers, J., and tiicoiet, C. M. (1989) Mol.‘Celi. Biol. 9; 3000-3008

Devereux. J.. Haeberli. P.. and Smithies. 0. (1984) Nucleic Acids Res. 12,387-395

DiDomenico, B. J., Bugaisky, G. E., and Lindquist, S. (1982) Cell 31, 593-603

Ellwood, M. S., and Craig, E. A. (1984) Mol. Cell. Biol. 4, 1454-1459 Farelly, F. W., and Finkelstein, D. B. (1984) J. Biol. Chem. 259,

5745-5751 Fickett, J. W. (1982) Nucleic Acids Res. 10, 5303-5318 Greene, J. M., Larin, Z., Taylor, I. C. A., Prentice, H., Gwinn, K. A.,

and Kingston, R. E. (1987) Mol. Cell. Biol. 7,3646-3655 Gribscov, M., and Burgess, R. R. (1986) Nucleic Acids Res. 14,6745-

6763 Guarente. L., and Ptashne, M. (1981) Proc. Natl. Acad. Sci. U. S. A.

78,2199-i203 Hahn. S.. Hoar. E. T.. and Guarente. L. (1985) Proc. N&l. Acad. Sci.

L? k/i. 82,8562-8566 ,

Hill, D. E., Hope, I. A., Macke, J. P., and Struhl, K. (1986) Science 234,451-457

4 D. Straus, W. Walter, and C. Gross, personal communication.

at IND

IAN

INST

OF SC

IEN

CE

on May 6, 2014

http://ww

w.jbc.org/

Dow

nloaded from

Structure and Regulation of a Yeast Heat-shock Gene 18921

Jakobsen, B. K., and Pelbam, H. R. B. (1988) Mol. Cell. Biol. 8,5040- 5042

Johnson, C., Chandrasekhar, G. N., and Georgopoulos, C. (1989) J. Bacterial. 171, 1590-1596

Jones, R. H., and Jones, N. C. (1989) Proc. N&l. Acod. Sci. U. S. A. 86,2176-2180

Jones, N. C., Rigby, P. W., and Ziff, E. B. (1988) Genes Deu. 2, 267- 281

Kornuc, M., Altman, R. Harricb, D., Garcia, J., Chao, J., Kayne, P., and Gaynor, R. (1988) Mol. Cell. Biol. 8, 3717-3725

Lin, Y. S., and Green, M. R. (1989) Proc. N&l. Acad. Sci. U. S. A. 86, 109-113

Lindquist, S., and Craig, E. A. (1988) Annu. Rev. Genet. 22,631-677 Maizel, J. V. J., and Lenk, R. P. (1981) Proc. Natl. Acad. Sci. U. S.

A. 78,7665-7669 McAlister, L., and Finkelstein, D. B. (1980) J. Bacterial. 143, 603-

612 McDaniel, D., Caplan, A. J., Lee, M. S., Adams, C. C., Fishel, B. R.,

Gross. D. S.. and Garrard. W. T. (1989) Mol. Cell. Biol. 9. 4789- 4798

Miller, J. H. (1972) Experiments in Molecular Genetics, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY

Needleman. S. B.. and Wunsch. C. D. (1970) J. Mol. Biol. 48. 443- 453

Nicolet, C. M., and Craig, E. A. (1989) Mol. Cell. Biol. 9, 3638-3646 Gzkaynak, E., Finley, D., Solomon, M. J., and Varshavsky, A. (1987)

EMBO J. 6,1429-1439 Park, H.-O., and Craig, E. A. (1989) Mol. Cell. Biol. 9, 2025-2033 Pelham, H. (1985) Trends Genet. 1,31-35 Perisic, O., Xiao, H., and Lis, J. T. (1989) Cell 59, 797-806 Perler, F., Efstratiadis, A., Lomedico, P., Gilbert, W., Kolodner, R.,

and Dodgson, J. (1980) Cell 20,555-566 Piper, P. W., Curran, B., Davies, M. W., Hirst, K., Lockheart., A.,

Oaden. J. E.. Stanwav. C. A.. Kinasman. A. J.. and Kinasman. S. G (1988) Nucleic AC& Rex '16, 1333-1348 -

Plesset, J., Foy, J. J., Chia, L. L., and McLaughlin, C. S. (1982) in Interaction of Translational and Transcriptional Controls in the Regulation of Gene Expression (Grunberg-Manago, M., and Safer,

B., eds), pp. 495-514, Elsevier Science Publishing Co., New York Rose, M. D., Misra, L. M., and Vogel, J. P. (1989) Cell 57,1211-1221 Rougvie, A. E., and Lis, J. T. (1988) Cell 54, 795-804 Rudolph. H.. and Hinnen, A. (1987) Proc. Nutl. Acad. Sci. U. S. A.

84,-134o-i344 Sell, S. M. (1987) Studies of Bacterial Genes Whose Products Are

Involved in Heat-shock Regulation and Bacteriophage h DNA Rep- lication Ph.D. dissertation, University of Utah, Salt Lake City

Sherman. F.. Fink. G. R.. and Hicks. J. B. (1982) Methods in Yeast Genetics, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY

Shuey, D. J., and Parker, C. S. (1986) J. Biol. Chem. 261, 7934-7940 Slater. M. R. (1986) Transcrintional Regulation of a Yeast Heat-shock

Gene. Ph.D. dissertation, eniversityof Wisconsin, Madison Slater. M. R.. and Craig. E. A. (1987) Mol. Cell. Biol. 7. 1906-1916 Slateri M. R.; and Craig; E. A. (1989) Nucleic Acids Res. 17, 805-806 Sorger, P. K., and Pelham, H. R. B. (1987) EMBO J. 6, 3035-3041 Sorger, P. K., and Pelham, H. R. B. (1988) Cell 54, 855-864 Sorger, P. K., Lewis, M. J., and Pelham, H. R. B. (1987) Nature 329,

81-84 Stone, D. E., and Craig, E. A. (1990) Mol. Cell. Biol. 10, 1622-1632 Tanaka, K., Yatomi, T., Matsumoto, K., and Toh-e, A. (1989) UCLA

Svmn Mol. Cell. Biol. New Series 96. 63-72 Tiliy, K., McKittrick, N., Zylicz, M., and Georgopoulos, C. (1983) Cell

34,641-646 Topol, J., Ruden, D. M., and Parker, C. S. (1985) Cell 42, 527-537 Tu, H., and Casadaban, M. J. (1990) Nucleic Acids Res. 18, 3923-

3931 Wei, R., Wilkinson, H., Pfeifer, K., Schneider, C., Young, R., and

Guarente, L. (1986) Nucleic Acids Res. 14, 8183-8188 Werner-Washburne. M., Stone. D. E.. and Craig, E. A. (1987) Mol.

Cell. Biol. 7, 256812577 I Werner-Washburne. M.. Becker. J.. Kosic-Smithers. J.. and Craie. E.

A. (1989) J. Bacterial.’ 17 1, 2686-2688 I Yl

Wiederrecht, G., Shuey, D. J., Kibbe, W. A., and Parker, C. S. (1987) Cell 48,507-515

Wu, C., Wilson, S., Walker, B., Dawid, I., Paisley, T., Zimarino, V., and Ueda, H. (1987) Science 238, 1247-1253

Xiao, H., and Lis, J. T. (1988) Science 239, 1139-1142 Zimarino, V., and Wu, C. (1987) Nature 327, 727-730

at IND

IAN

INST

OF SC

IEN

CE

on May 6, 2014

http://ww

w.jbc.org/

Dow

nloaded from