Vol. 65, No. 11

Analysis of the Specificity and Mechanism of TranscriptionalActivation of the Human hsp7O Gene during

Infection by DNA VirusesBENETTE PHILLIPS, KLARA ABRAVAYA, AND RICHARD I. MORIMOTO*

Department of Biochemistry, Molecular Biology, and Cell Biology,Northwestern University, Evanston, Illinois 60208-3500

Received 29 April 1991/Accepted 12 July 1991

We have examined the transcriptional regulation of the 70-kDa (70K) heat shock gene family followinginfection of human and monkey cells with four different DNA viruses: adenovirus type 5 (Ad5), herpes simplexvirus type 1 (HSV-1), simian virus 40, and vaccinia virus. Our results indicate that induction of these genes isnot a general response to the stress of viral infection but is instead a highly specific response, both with regardto the inducing virus and with regard to the target gene. Of three 70K heat shock genes examined, only hsp70was induced during viral infection, and induction occurred only after infection by Ad5 and HSV-1. As revealedby genomic footprinting analysis, the mechanism of transcriptional activation of hsp70 during Ad5 or HSV-1infection does not involve changes in the avidity of binding of basal transcription factors to the hsp70 promoter.In HSV-1-infected HeLa cells, transcriptional activation of hsp70 was quite transient, following whichtranscription was rapidly repressed; this was accompanied by the release of bound factors from the hsp70promoter. In addition to the selectivity which characterizes the viral activation of hsp70 transcription, ourresults indicate that the consequences of this activation, as measured by changes in hsp70 mRNA levels andprotein synthesis, are also virus specific.

Induction of stress genes following viral infection of cellsin culture has been widely observed. Both DNA and RNAviruses have been reported to induce such genes in a widevariety of cell types (9, 26, 27, 28, 31, 39, 44, 47). There hasbeen considerable interest in the significance of the elevatedexpression of stress proteins in infectious disease, e.g., as

markers of disease or as determinants in the etiology,progression, or immunological status of the disease. Thisinterest has been heightened by the recognition that productsof stress genes not only protect cells from environmentalstress but also play major roles in a variety of fundamentalbiological processes such as protein folding (4) and translo-cation (7, 12, 25, 42) and the assembly/disassembly ofmacromolecular complexes (18). This recognition expandsthe potential implications of the virus-induced stimulation ofstress gene expression. In addition to considering suchinduction as a defensive response of the host to the physio-logical disruption of virus infection or as a sensitivity ofcertain host genes to viral signalling, it is also quite possiblethat these genes are induced because sufficient levels of theirproducts are vital to the successful outcome of a viralinfection. It has, on the other hand, also been suggested thatincreased expression of stress proteins in virus-infected cellsmay alert the immune system to the presence of a viralinfection (61). These issues, as well as the utility of thissystem for studying viral effects on host gene expression,provide the impetus for continued examination of the phe-nomenon of stress protein induction during virus infection.

Several viruses have been reported to activate members ofthe 70-kDa (70K) family of stress genes, which includehsp70, a highly heat inducible member, grp78 (Bip), a gene

whose product is localized to the endoplasmic reticulum,and p72 (hsc70, clathrin-uncoating ATPase). Induction of

* Corresponding author.

hsp70 in adenovirus (Ad)-infected HeLa cells has been welldocumented and characterized (26, 59). Increased synthesisof hsp70 has also been reported in simian virus 40 (SV40)-infected CV1 cells (28), as have elevated levels of hsp70RNA in SV40-infected HeLa cells (52). There have also beennumerous reports of accumulation of 70K stress proteins inseveral different cell lines during herpes simplex virus (HSV)infections. The initial study used an HSV-1 temperature-sensitive mutant in the immediate-early protein ICP4. Infec-tion with this mutant virus at the nonpermissive tempera-ture, which resulted in the overproduction of all of the viralimmediate-early proteins, including mutant ICP4, also led toincreased synthesis of host heat shock proteins (39). A laterreport demonstrated that the overproduction of an abnormalform of ICP4, rather than the presence of wild-type HSV-1proteins, was in fact responsible for induction of stressprotein synthesis in this system (46). Another series ofstudies used as probes monoclonal antibodies made againsthost DNA-binding proteins which were purified from HSV-2-infected BHK cells. These investigators reported in-creased levels of 57 to 61K heat shock proteins, whoseidentity is unclear, in both HSV-1- and HSV-2-infected cells(27, 31). Another member of the herpesvirus family, humancytomegalovirus, has been reported to induce expression ofhsp70 in human foreskin fibroblasts (47). The paramyxovi-ruses SV5 and Sendai virus were reported to induce expres-sion of grp78 in infected cells (44), and the transcriptionalinduction of grp78 following SV5 infection of CV1 cells hasrecently been further characterized (56a).A consideration of the published literature, which consists

primarily of isolated studies focusing on a single virus andhost cell, prompted us to undertake a comprehensive studyof the activation of the 70K family of human heat shockgenes during infection of primate cells by DNA viruses.Specifically, we have concentrated on the transcriptionalinduction of these genes, in host cells which are fully

5680

JOURNAL OF VIROLOGY, Nov. 1991, p. 5680-56920022-538X/91/115680-13$02.00/0Copyright © 1991, American Society for Microbiology

TRANSCRIPTIONAL ACTIVATION OF THE hsp7O GENE 5681

permissive for infection, by DNA viruses which are mem-bers of four different viral classes: AdS, HSV-1, SV40, andvaccinia virus. One of our primary goals was to determinethe specificity of viral induction, i.e., to establish within asingle study which members of the 70K heat shock genefamily were induced by each virus. Our data indicate thatthere is a high degree of specificity to induction of heat shockgenes by DNA viruses: hsp70 is the only member whosetranscription increases upon infection, and in our hands suchenhancement is observed only during infection with AdS andHSV-1. These results were then used as a basis for genomicfootprinting studies to examine possible mechanisms ofhsp70 activation during viral infection.

MATERIALS AND METHODS

Cells and virus. HeLa cells, monkey CV1 cells, and 293cells, a human embryonic kidney cell line transformed byand expressing the El region of Ad5 (20), were maintained asmonolayers in Dulbecco modified Eagle medium (DME)supplemented with 5% calf serum. IC4 cells (6), a HeLaderivative obtained by transformation with a mouse mam-mary tumor virus-ElA chimeric gene such that ElA expres-sion is dexamethasone inducible, were a generous gift of A.Berk and were maintained as monolayers in DME containing10% fetal bovine serum and 200 ,ug of G418 per ml.AdS strain d1704 (phenotypically wild type) was propa-

gated and titered on HeLa cells and was used as a crudelysate. HSV-1 (KOS) was kindly supplied by P. Spear andwas used either as a crude lysate or after dextran purification(with equivalent results). Vaccinia virus (WR), originallyobtained from B. Moss, was purified on sucrose gradientsand generously provided by R. Paterson. SV40 wild-typestrain 830 was grown and titered on CV1 cells and was usedas a crude lysate. For infections, HeLa or CV1 cells wereplated at a density of 1.8 x 106 cells per 100-mm dish or 7.0x 105 cells per 60-mm dish 36 h prior to infection and weresubconfluent at the time of infection. Medium was removed,saved at 37°C, and replaced with 1.0 ml (100-mm dishes) or0.4 ml (60-mm dishes) of virus, which was diluted in serum-free DME to yield a multiplicity of infection of 10. Virusadsorption was allowed to proceed for 1 h at 37°C, followingwhich spent medium was returned to the dishes. Mockinfections were carried out by using either serum-free DMEor lysate prepared from mock-infected cells.IC4 cells were plated at a density of 3 x 106 cells per

100-mm dish 18 h prior to treatment with dexamethasone,which was added to the culture medium as an ethanolsolution (final concentrations, 0.1% ethanol and 10-6 Mdexamethasone).Measurement of transcription rates. Isolation of nuclei

and transcription run-on analysis were done as previous-ly described (36). Nuclei from 1.5 x 107 to 2.0 x 107 cellswere used in each analysis. The following plasmids orplasmid fragments were immobilized on filters and hybrid-ized to labeled transcripts: pH2.3 (human hsp70 [58]); pUC801, kindly provided by L. Weber (human hsp90a [22]);pHG23.1.2 (human grp78 [55a]); pHA7.6 (human p72 [23a]);pBR322 (vector); pHFPA-1 (human ,B-actin [21]); tkLS-119/-109 (HSV-1 tk [34]); KpnI-BamHI fragment of pJYM(SV40 early region [24]); and XbaI-XhoI fragment of pA5130(AdS E3 region [5]).

Analysis of RNA levels. Isolation of cytoplasmic RNA,RNA dot blot analyses, and S1 nuclease protection assayswere done as previously described (56, 58). 32P-labeledprobes for dot blot analysis were prepared by nick transla-

tion of plasmids described above. For Si nuclease analysisof hsp70, a probe 5' end labeled at the BamHI site (+ 153) ofhsp70 was used; for analysis of ElA, the probe was 5' endlabeled at an XbaI site (nucleotide 1340) in exon 2 of the AdSElA gene (6).

Analysis of protein synthesis rates. For labeling of HeLacells during adenovirus or HSV-1 infection, cells in 60-mmdishes were washed and then incubated in methionine-freeDME containing [35S]methionine (15 p.Ci/ml) for 1 h. Forlabeling of CV1 cells during SV40 infection, cells werewashed and incubated in DME containing 1/17 the normalconcentration of methionine, [35S]methionine (60 ,uCi/ml),and 2.0% calf serum for 6 h. Labeled proteins were analyzedby isoelectric focusing in the first dimension (ampholytes,pH 5 to 7) and sodium dodecyl sulfate-polyacrylamide gelelectrophoresis in the second dimension (40).

Gel mobility shift assay. Conditions for the gel shift assay,a description of the 32P-labeled heat shock element (HSE)oligonucleotide, and preparation of whole-cell extracts wereas previously published (36).Genomic footprinting analyses. Genomic footprinting c.

samples taken during virus infection of HeLa cells wiperformed as previously described (1). Briefly, four 100-mn-dishes of infected cells were trypsinized, pooled, resuspended in a small volume of DME plus 5% calf serum, andexposed to dimethyl sulfate (DMS). Genomic DNA wasisolated, and genomic footprinting was performed by using aligation-mediated polymerase chain reaction-based method(37). Duplicate samples were analyzed at each time point.Primers which allowed visualization of the DMS reactivitypatterns upstream of nucleotide +48 on the coding strandand downstream of nucleotide -131 on the noncoding strandwere described previously (1).

RESULTS

Specificity of transcriptional activation. HeLa cells wereinfected with AdS, HSV-1, or vaccinia virus, and monkeyCV1 cells were infected with SV40, in order to determinewhether, under conditions of a fully lytic infection, there isspecificity with respect to which of these DNA viruses caninduce 70K heat shock genes and which of these genes isvirus inducible. At selected times during infection, nucleiwere harvested for use in transcription run-on assays. Sam-pling times, which varied with the virus, were chosen so asto obtain samples prior to and during the early phase of viralinfection as well as during the late phase of infection, whenviral replication and expression of viral late proteins wasoccurring. Samples were also harvested at each time pointfrom mock-infected cells (Fig. la); frequently, although notconsistently, mock infection resulted in a mild (usuallyapproximately twofold) and transient transcriptional induc-tion of hsp70.As has been previously reported by our laboratory and

others, the transcription rate of the hsp70 gene in HeLa cellsincreased during the early phase of Ad infection (26, 59). Ourresults indicate that this induction is specific for hsp70;transcription of p72 and grp78 was not enhanced (Fig. lb).Although minor fluctuations in the transcription rates of p72and grp78 often occur during viral or mock infection, re-peated analyses of adenovirus-infected HeLa cells failed toreveal any consistent induction of these genes.A similar selective induction of hsp70 transcription was

observed when HeLa cells were infected with HSV-1 (Fig.lc). The transcription rate of hsp70 was elevated at 3 hpostinfection (p.i.) and was induced almost 10-fold by 5 h

VOL. 65, 1991

5682 PHILLIPS ET AL.

a 3 7 11 15 24

vec tor

hsp70 * 40.

grp78

p72

a c t n I-

b 3 7 11 15 24

vector

hsp70 e Igrp78

p72 .ac t In

cu 1 MocklHeLa

1z

cz ..

c)I *'

-u 4> actin

Hours p t i ngr p 7 8+

H c urs post-in fe rIion

1D 8

o 6

.acnc 4

V 2

U)cc 0

Ad5/HeLa

0 10 20

Hours post-infection

3 5 7 9 11

1 3 5 7 10 13

............. .. ...I* a I* * W.w.*.. _ I

HSV-1 HeLa

ar.< / ~~~~~~hs p 7 0

p72o

Hours post-infeclion

w 8-(-

O 6-._

.'

C 4-

. 2

4)cc

0

V V/He La

actin

hrp78*~~ ~~~* I II 1

2 4 6 8 10 12 14

Hours post-infection

- s~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.

*~ * 0 0

* *- *e.1.0

C

vector

hsp70

grp78

p72

actin

h v - tk

dvector

hsp70

grp78

p72

actin

J. VIROL.

TRANSCRIPTIONAL ACTIVATION OF THE hsp7O GENE

1 0 1 6 22 35 cos7

Is__ _I

2 5 8 1 1 14

nw1

20

m

6-

0

'UC

0

0.

IX 1 0

IC

0

0.

SV40/CV1

10 20 30 40Hours post-infection

Ad

1 3 5 7 9l.- HSV

FIG. 1. Run-on analysis of transcription of the indicated genes in mock-infected HeLa cells (a), AdS-infected HeLa cells (b),HSV-1-infected HeLa cells (c), vaccinia virus (VV)-infected HeLa cells (d), and SV40-infected CV1 cells (e) and run-on analysis of hsp7Otranscription in Ad- and HSV-1-infected CV1 cells (f). Column numbers indicate the time (hours postinfection) at which nuclei were harvestedfor run-on analysis. At the initial time point shown for each experiment in panels b to f, the transcription rates of the hsp7O, grp78, p72, and,-actin genes were indistinguishable from those in mock-infected cells from the same experiment. There was, however, variation betweenexperiments in the relative transcription rates of these four cellular genes at the initial time point, the basis for which is not clear. Labeledtranscripts were hybridized to plasmid DNA immobilized on nitrocellulose. Values for transcription rates were from laser densitometry ofautoradiographs or from scanning and quantification of the filters with a Molecular Dynamics 400A Phosphoimager. Within each experiment,transcription rates were normalized to the value at the initial sampling time.

p.i., following which transcription of hsp70 was repressed,coincident with a generalized shutoff of host transcriptionwhich virtually abolished transcription of the p72, grp78,and ,-actin genes by 11 h p.i. Repeated infections withHSV-1 revealed no consistent activation of either p72 orgrp78. These results constitute the first unequivocal demon-stration of activation of hsp7o during infection by wild-typeHSV-1.

Induction of hsp70 transcription is not a general responseof HeLa cells to viral infection, however. Infection of HeLacells with vaccinia virus did not activate transcription of anyof the 70K heat shock genes examined (Fig. ld), although theefficacy of the infection was apparent both in a rapidrounding up of the infected cells and by a change in thepattern of protein synthesis from host to viral proteins (datanot shown).

It has been previously reported that the synthesis ofmonkey hsp70 increases following infection of CV1 cellswith SV40 (28); this was later suggested to be due tolarge-T-antigen-mediated transactivation of the hsp70 pro-moter (29). We have consistently failed to detect transcrip-tional induction of any of the 70K genes during the first 36 hof infection, despite abundant transcription of the SV40early region encoding large-T and small-t antigens (Fig. le).(Transcriptional induction of the ,-actin gene during SV40infection was consistently observed, however.) Further-more, the relative rates of transcription of hsp70, p72, andgrp78 in COS7 cells, which constitutively express large-Tantigen, were similar to those in CV1 cells, suggesting that

the expression of large-T antigen in COS7 cells does notselectively enhance expression of any single member of the70K family (Fig. le). The monkey hsp70 gene is virusinducible, however, since activation was observed followinginfection of CV1 cells with Ad5 or HSV-1 (Fig. lf).

It thus appears that of the four DNA viruses examined,only AdS and HSV-1 activate transcription of any of the 70Kheat shock genes during infection of permissive host cells,and this activation is selective for hsp7o. Previous reportsfrom our laboratory and from other investigators have sug-gested that the adenovirus transactivator ElA mediatesinduction of hsp70 transcription, since Ad mutants which failto synthesize ElA, and specifically the product of the ElA13S mRNA, also fail to activate hsp70 transcription duringinfection (26, 59). To directly establish that ElA, in theabsence of other viral proteins, is sufficient to induce tran-scription of the endogenous hsp70 gene, we used a cell line,IC4, in which expression of ElA is dexamethasone inducible(6). As shown in Fig. 2, induced expression of the integratedElA gene was accompanied by a fivefold increase in thelevels of hsp7o cytoplasmic RNA. For purposes of compar-ison, hsp70 RNA levels in 293 cells, which constitutivelyexpress the Ad5 ElA gene, are shown in the same figure.Induction of hsp70 expression in IC4 cells was not a re-sponse to dexamethasone, since expression of the hsp70gene was not altered when HeLa cells were treated with thesame concentration of dexamethasone (data not shown).This result provides convincing evidence that ElA alone iscapable of inducing expression of the endogenous hsp70

evector

hsp7O

grp78

p72

actin

S V40 -T

f

* * *bl- L

VOL. 65, 1991 5683

5684 PHILLIPS ET AL.

a

ElA _

b

hsp70 _

293 0 3a

6 9 18 Mnt154

HSF

CHBA

NS

b

vec.tor154

FIG. 2. Levels of hsp70 mRNA during induction of ElA expres-sion. Cytoplasmic RNA was isolated at the indicated times aftertreatment of IC4 cells with 106 M dexamethasone or from un-treated 293 cells, and 10 ,ug was used for S1 analysis. (a) S1 analysisusing an ElA-specific probe. The protected fragment (111 nucleo-tides [nt]) is indicated on the left. (b) Si analysis using an hsp70-specific probe. The protected fragment (153 nucleotides) is indicatedon the left.

gene and indicates that the IC4 cell line may provide a usefulsystem in which to demonstrate the ElA responsiveness ofcellular genes.Although transcriptional activation of hsp70 during HSV-1

infection is less clearly defined, it exhibits two propertieswhich implicate the viral immediate-early gene products asinducers. First, activation of hsp70 did not occur when HeLacells were infected in the presence of cycloheximide, indi-cating a requirement for new, presumably viral, proteinsynthesis. Second, the kinetics of hsp70 activation paralleledthat of the HSV-1 thymidine kinase gene (Fig. lc), whichrequires immediate-early proteins for its expression (23, 32).

Activation of HSF. During heat shock or upon exposure toother conditions which result in protein damage, transcrip-tional induction of the human hsp7O gene is mediated by aspecific heat shock transcription factor (HSF), whose acti-vation can be monitored by gel retardation assay (19, 36). Toestablish whether HSF activation could be detected duringinfection by Ad or HSV-1, gel shift assays were used tomonitor levels of activated HSF in extracts of infected cells.Although activated HSF was never detected during infectionof HeLa cells with HSV-1, low levels of the activated factor,which were first detectable at 3 h p.i., were consistentlydetected in Ad-infected HeLa cells (Fig. 3a).The kinetics of appearance of activated HSF in Ad-

infected HeLa cells, coupled with a previous observation inour laboratory (44a) that in 293 cells, which constitutivelyexpress ElA and EBB, HSF is constitutively activated in theabsence of heat shock, suggested the possible involvementof ElA in the activation of HSF. This possibility wasespecially intriguing in view of recent speculations that HSFmay be held in an inactive form by interaction with othercellular proteins (8) and by a recent report that ElA is ableto activate the cellular transcription factor E2F by disruptinga complex which maintains E2F in an inactive state (3).However, examination of IC4 cells, induced to express ElA

hsp7O

p72

grp78

actin

hsp9O

Ad-E3

l 3 6 9 1 2 15

FIG. 3. Activation of HSF during Ad infection of HeLa cells. (a)Gel retardation assay using an HSE-containing oligomer and whole-cell extracts prepared from HeLa cells at the indicated times (hours)of infection with Ad or from heat-shocked (42°C, 45 min) HeLa cells(HS). Complexes due to nonspecific (NS) DNA-binding proteins, aconstitutive HSE-binding activity (CHBA) which is present innon-heat-shocked HeLa cells (36), and HSF are indicated. (b)Run-on analysis of transcription of the indicated genes, using theinfected cells analyzed in panel a.

at levels equivalent to those in 293 cells, failed to revealdetectable levels of activated HSF. The mechanism bywhich HSF is activated during Ad infection is thus not yetunderstood.The presence of low levels of activated HSF in Ad-

infected HeLa cells also raises the issue of its possiblecontribution to the transcriptional induction of hsp70. In thisregard, our failure to observe activation of hsp9o during Adinfection should be noted (Fig. 3b). The lack of induction ofanother HSF-responsive gene would argue against a signifi-cant contribution to hsp70 induction by the low levels ofHSF present during Ad infection. This argument is strength-ened by the observation that HSF levels are still rising at 12h p.i., at which time hsp7O transcription has peaked and isdeclining (Fig. 3).Genomic footprinting of the hsp7O promoter during Ad and

HSV-1 infection. To explore possible mechanisms of virus-mediated transcriptional activation of hsp70, we performedin vivo genomic footprinting to examine interactions offactors with the hsp7O promoter at selected time pointsduring infection of HeLa cells with Ad or HSV-1. Such ananalysis has been previously conducted using the Ad E2Apromoter; the results indicated that ElA-mediated transcrip-tional induction of the viral gene did not perturb the bindingof factors to E2A promoter elements (13). There have beenno reports of similar studies on HSV-1 promoters to assesswhether changes in promoter occupancy correlate with thetranscriptional state of the viral gene, although it has beenproposed that the DNA-binding ability and transactivatingcapabilities of the major viral immediate-early protein ICP4

. M

1itbim i

"MWj.,&AJ-3 6 9 12

** * .0

4w

0 0 0 0

1 1

J. VIROL.

TRANSCRIPTIONAL ACTIVATION OF THE hsp7O GENE 5685

may be linked (43, 53). Such studies on viral promoters arecomplicated by the possibility that only a fraction of the totalnumber of viral genomes in the infected cell are beingactively transcribed, thus hindering detection of subtle dif-ferences in factor binding. This drawback is avoided byconducting a similar analysis of the endogenous hsp70 pro-moter. The sequence of the promoter between nucleotides-20 and -120 is shown in Fig. 4. The proximal 75 nucleo-tides are sufficient to confer basal expression and ElAinducibility upon a reporter gene when such a construct isintroduced into HeLa cells by transfection (57, 60). Thisregion contains consensus binding sites for several knowntranscription factors, among them CTF and Spl. There isalso a partial (five of eight bases) match to the consensusbinding site for ATF/CREB at -35. Genomic footprintingstudies performed previously in our laboratory have re-vealed that in untreated HeLa cells, the promoter regionswhich include the CCAAT and Spl sites are protected frommethylation, as is a G residue on the coding strand justdownstream of the TATA site. No factor binding is apparentat the ATF-like site. During heat shock, a characteristicpattern of protections and hypersensitivities is observed onboth strands of the promoter in a 30-bp region termed theHSE, which is centered at -105 (1). These alterations in themethylation pattern are summarized in Fig. 4 (patterns C andHS).Genomic footprinting of the proximal 120 nucleotides of

the hsp70 promoter in samples taken during the first 12 h ofmock or Ad infection revealed identical patterns for thecoding strand (Fig. 5a). When the noncoding strand wasexamined, a hypersensitivity at G-95 was seen in Ad-in-fected, but not mock-infected, samples at 9 and 12 h p.i. Thishypersensitivity, which is the most sensitive diagnosticfeature of HSF binding (1), was much less intense than thehypersensitivity observed for a sample of HeLa cells heatshocked at 42°C for 40 min (Fig. Sb). This finding presumablyreflects the lower levels of HSF in Ad-infected cells than inheat-shocked HeLa cells. There was no perturbation inbinding of factors to the proximal 75 bp of the hsp70promoter during the period of Ad-induced transcriptionalactivation (summarized in Fig. 4). These results indicate thatElA transactivation is not mediated by changes in avidity ofbinding of basal transcription factors to this region of thehsp70 promoter.

Footprinting of the coding strand of the hsp70 promoterduring HSV-1 infection (Fig. 6a) revealed that no changes infactor binding could be detected at 3 and 5 h p.i., whentranscriptional activation was peaking (Fig. 6b). However, at7 and 10 h p.i., when host transcription, including that of thehsp70 gene, was repressed, a release of bound factors fromthe promoter was quite apparent. This was most dramatic inthe case of the Spl site: in the 10-h mock-infected sample,there was a characteristic set of three protections and asingle hypersensitivity, indicative of a bound factor. Incontrast, the pattern in this region in the 10-h HSV-infectedsample closely resembled that of naked DNA, indicatinglittle if any bound protein at this site (Fig. 6a). Decreases infactor binding to the CCAAT and TATA sites were concur-rently observed. Footprinting of the noncoding strandshowed, as expected from gel shift analysis, no occupancy ofthe HSE in any of the HSV-1-infected samples (data notshown). Although this analysis (summarized in Fig. 4) sug-gests that no changes in interactions of either cellular or viralfactors with the hsp70 promoter occur during the period ofHSV-1-mediated transcriptional activation, it does provide a

window on events which accompany transcriptional repres-sion during HSV-1 infection.

Distinct features of posttranscriptional regulation. Al-though the primary focus of this study was the transcrip-tional regulation of the 70K heat shock genes during viralinfection, regulation of these genes at the posttranscriptionallevel was also examined. Measurement of cytoplasmic RNAlevels by both dot blot analysis and SI analysis (data notshown) and of protein synthesis rates by two-dimensional gelanalysis revealed that while transcriptional activation ofhsp7o during Ad infection was reflected in increased levels ofboth mRNA and protein synthesis (Fig. 7a and 8a), such wasnot the case during HSV-1 infection ofHeLa cells. Levels ofhsp70 mRNA remained relatively constant during the first 11h of HSV-1 infection; however, during this same period, thelevels of grp78, p72, and actin mRNA declined rapidly (Fig.7b). Protein synthesis rates were also analyzed at 2-h inter-vals during the first 9 h of HSV-1 infection. We observed thatin a background of declining protein synthesis rates for mostcellular genes, including the other members of the 70K heatshock family, hsp70 protein synthesis rates remained rela-tively constant. This contrast is most readily apparent whenthe patterns of protein synthesis at 1 and 9 h p.i. arecompared (Fig. 8b). HSV-1-induced degradation of hostmessages, with consequent effects on protein synthesis, is awell-established phenomenon, and hsp70 mRNA has beenshown to be susceptible to such degradation (41, 51). Pre-sumably, the steady-state level of mRNA represents a bal-ance between HSV-1-mediated transcriptional activationand degradative effects on host messages.To rule out the possibility that there is posttranscriptional

regulation of hsp70 during SV40 infection, which wouldaccount for the discrepancy between our failure to detecttranscriptional activation of hsp70 and previous reportsshowing an increase in the synthesis rate of hsp70, wecarried out a two-dimensional gel analysis of protein synthe-sis rates during SV40 infection of CV1 cells (Fig. 9). We alsoincluded a sample of uninfected COS7 cells. As is apparentfrom this analysis, synthesis of hsp70 in both mock-infectedCV1 cells and COS7 cells was virtually undetectable, al-though heat shock strongly induced hsp70 synthesis (Fig. 9).No induction of hsp70 synthesis was detected during the first42 h of SV40 infection; these same gels showed viral lateprotein synthesis commencing at 22 h p.i., indicating thatviral infection was proceeding normally.

DISCUSSION

Selectivity of transcriptional activation. The results of thisstudy, in which transcriptional regulation of several mem-

bers of the 70K heat shock gene family was examinedfollowing infection of human and monkey cells with fourdifferent DNA viruses, indicate that induction of heat shockgenes is not a general response to the stress of viral infectionbut is, rather, a highly specific response. This specificity ismanifested with regard to both the inducing virus and thetarget gene. Of three 70K family members examined, onlyhsp70 was induced during viral infection, and induction wasobserved only in response to infection by Ad and HSV-1.These two viruses also failed to activate transcription ofhsp90, which, like hsp70, is highly heat inducible, thusindicating that characteristics which confer a susceptibilityto rapid transcriptional activation do not necessarily alsoconfer susceptibility to viral transactivation. The absence ofgrp78 induction during HSV-1 infection is noteworthy inview of recent observations that transcriptional activation of

VOL. 65, 1991

5686 PHILLIPS ET AL. J. VIROL.

-120 -110 -100 -90 -80 -70 -60 -50 -40 -30 -2:

NGAAN CCAAT GGCGGG TGACGTCA TATA

HSF CTF Spi ATF TFIID

* * #4t_ GGAGGCGAAACCCCTGGAATATTCCCGACCTGGCAGCCTCATCGAGCTCGGTGATTGGCTCAGAAGGGAAAAGGCGGGTCTCCGTGACGACTTATAAAA-\-CCTCCGCTTTGGGGACCTTATAAGGGCTGGACCGTCGGAGTAGCTCGAGCCACTAACCGAGTCTTCCCTTTTCCGCCCAGAGGCACTGCTGAATATTTT-

4 +4 4

* tttkI- z GGAGGCGAAACCCCTGGAATATTCCCGACCTGGCAGCCTCATCGAGCTCGGTGATTGGCTCAGAAGGGAAAAGGCGGGTCTCCGTGACGACTTATAAAAGu-u) CCTCCGCTTTGGGGACCTTATAAGGGCTGGACCGTCGGAGTAGCTCGAGCCACTAACCGAGTCTTCCCTTTTCCGCCCAGAGGCACTGCTGAATATTTTC

GGAGGCGAAACCCCTGGAATATTCCC6,ACCTGGCAGCCTCATCGAGCTCGGTGATTGGCTCAGAAGGGAAAAGGCGGGTCTCCGTGACGACTTATAAAAG

-o

9-12

I_

GGAGGCGAAACCCCTGGAATATTCCCGACCTGGCAGCCTCATCGAGCTCGGTGATTGGCTCAGAAGGGAAAAGGCGGGTCTCCGTGACGACTTATAAAAG15 CCTCCGCTTTGGGGACCTTATAAGGGCTGGACCGTCGGAGTAGCTCGAGCCACTAACCGAGTCTTCCCTTTTCCGCCCAGAGGCACTGCTGAATATTTTC

71 GGAGWGCGAACCCCTGGAATATTCCCGACCTGGCAGCCTCATCGAGCTCGGTGATTGGCTCAGAAGGGAAAAGGCGGGTCTCCGTGACGACTTATAAAAG71 CCTCCGCTTTGGGGACCTTATAAGGGCTGGACCGTCGGAGTAGCTCGAGCCACTAACCGAGTCTTCCCTTTTCCGCCCAGAGGCACTGCTGAATATTTTC7- 0 t

FIG. 4. Summary of DMS reactivity patterns of the hsp7O promoter at the indicated times of infection of HeLa cells with Ad or HSV-1.The promoter region sequence analyzed by genomic footprinting is shown at the top. Sites which are perfect or imperfect matches toconsensus sites for known transcription factors are underlined. The consensus sequence and transcription factors which bind to these sitesin vitro are indicated above each underlined site. Summaries of footprinting analyses of DNA from control (C) and heat-shocked (HS) HeLacells, shown for comparison, were published previously (1). Arrows denote guanine residues protected from methylation as compared withnaked DNA, and stars denote guanines hypersensitive to methylation; the sizes of these symbols indicate the relative degree of protectionor hypersensitivity.

TRANSCRIPTIONAL ACTIVATION OF THE hsp70 GENE 5687

a

-120

HSE

-20TATA IATF I

Sp1CCAAT I

Spl

ATF

TATA

b

CCAAT

HSE

-120

1 3 6 9 12 1 3 6 9 12 HSM A M A M A M A M A M A M A M A M A M A

FIG. 5. DMS reactivity patterns of the human hsp70 promoter during Ad infection of HeLa cells. Samples were from the infected cellsanalyzed in Fig. 3. (a) Coding-strand methylation patterns in genomic DNA isolated from mock-infected (M) and Ad-infected (A) HeLa cellsat the indicated times (hours postinfection). (b) Noncoding-strand methylation patterns. Genomic DNA from HeLa cells which were heatshocked at 42°C for 5 or for 40 min (HS) is included in this set. The star denotes hypersensitivity to methylation. Although the overall intensityof bands in the lanes representing DNA from Ad-infected cells at 9 and 12 h p.i. is higher than in adjacent lanes, the indicated guanine residuein these samples was judged to be hypersensitive on the basis of its intensity relative to the intensity of the adjacent guanine residues.

this gene (but not of hsp70) occurs during infection of CV1cells with the paramyxovirus SV5 and is associated with ahigh flux of the viral HN glycoprotein through the endoplas-mic reticulum (56a). During HSV-1 infection, there is abun-dant production of several viral glycoproteins without acti-vation of grp78 transcription.Although it has been reported that infection of monkey

CV1 cells with SV40 resulted in an increase in the synthesisrate of hsp70 (28), nuclear run-on analysis performed in ourlaboratory has consistently failed to reveal transcriptionalinduction of hsp70 (or of grp78 or p72) during the first 48 h of

SV40 infection. Although it is possible that a difference inviral strain or a cell line variation is responsible for thedisparity between our data and previously reported results,nuclear run-on analysis and two-dimensional gel analysisusing COS7 cells also suggest that large-T antigen does notenhance expression of the endogenous hsp70 gene.Our results thus suggest that transcriptional induction of

hsp7O by DNA viruses may be limited to those virusesencoding immediate-early proteins with promiscuous trans-activating capabilities. Although SV40 large-T antigen issometimes classified as a promiscuous transactivator, in the

VOL. 65, 1991

7

-zO

5688 PHILLIPS ET AL.

LJ L1 IlJ N1 3 5 7 10

M a ME N H M E M E

-70

I CAAT

Sp1

ATF

I TATA

-20

a4 8 12 24

. *.

F~~~~~~~~~~~~~~~~~~_..

b

hsp70

p72

grp78

actLi n

1 3 5 7 9 11

I~~~~~~~~~~

.. |* . 4

iii

FIG. 7. Analysis of mRNA levels in HeLa cells infected with Ad(a) or HSV-1 (b). Cytoplasmic RNA was isolated from HeLa cells atthe indicated times (hours) of infection, quantitated, immobilized onnitrocellulose, and hybridized to gene-specific 32P-labeled probes.Top and bottom rows correspond to 1 and 2 p.g of RNA, respec-tively.

b

vector

hsp70

act in

HSV-TK

1 3 5 7 10

FIG. 6. DMS reactivity patterns of the human hsp7O promoterduring infection of HeLa cells with HSV-1. (a) Coding-strandmethylation patterns in genomic DNA isolated from mock-infected(M) and HSV-1-infected (H) HeLa cells at the indicated times (hourspostinfection). N, protein-free (naked) DNA, methylated in vitro.Arrows denote guanine residues protected from methylation; starsdenote guanines hypersensitive to methylation. (b) Run-on analysisof transcription of the indicated genes. Samples were from theinfected cells analyzed in panel a.

context of a natural infection it need only be capable oftransactivating the single SV40 late promoter. This is incontrast to the situation for Ad and HSV, in which theimmediate-early protein(s) must transactivate multiple viralearly and late promoters with disparate compositions andorganizations. Pertinent to this point was a study by Everettand Dunlop (15) which contrasted the ability of HSV-1 andAd to transactivate a variety of plasmid-borne promoterswith the inability of SV40 to induce expression from thesesame promoters. Vaccinia virus, which also failed to inducehsp7O transcription, does not utilize the host transcriptional

machinery and carries out a productive infection withoutentering the nucleus (35), so that transactivation of hostpromoters by viral products is unlikely. Vaccinia virus does,however, have immediate cytopathic effects on HeLa cellswithout evoking a stress response.

Possibility that Ad and HSV-1 utilize similar mechanismsfor transactivation of hsp7O. It is quite possible that similar oridentical mechanisms underlie activation of hsp7O by Ad andHSV-1; previous studies provide precedent for this sugges-tion. Both ICP4 and the pseudorabies virus (PRV) immedi-ate-early protein (IE), a homolog of the HSV-1 ICP4 protein,can functionally substitute for ElA in activating ElA-re-sponsive adenovirus promoters (16, 55). Moreover, thestudy of Everett and Dunlop (15) demonstrated that both theHSV-1 glycoprotein D and the rabbit P-globin promoters,transfected into HeLa cells, could be transactivated bysubsequent infection with either HSV-1, PRV, or Ad2.

Sensitivity of hsp7O to viral transactivation. Although theability of viruses such as Ad to transactivate cellular genes isoften alluded to, this claim is usually based on cotransfectionstudies; hsp7O is one of the few well-documented examplesof a cellular gene that is responsive to viral transactivation.The paradox of promiscuous activation of promoters intro-duced by transfection or infection but very selective induc-tion of promoters resident in the genome remains unre-solved. This study reiterates the notion that hsp7O is ratherdistinctive in its sensitivity to viral activation. Although thefeatures of the endogenous gene which confer such respon-siveness have not been identified, there are several possibil-ities worthy of consideration. One possibility is that anindividual promoter element or the particular arrangement ofpromoter elements in hsp7O confers sensitivity to viraltransactivators. It is also possible that the chromosomal

a

T

_. ~~_. -_ _

a. S

_w:tZ'

~~~~~~~~4-

__.0

J. VIROL.

40 0

0 0

0 0 * *

TRANSCRIPTIONAL ACTIVATION OF THE hsp70 GENE 5689

a 4 12

b 1

a .

0X-*N'4

I* _w

I .... ...

qo -NW

q

40

9

v

....Nq - 4

FIG. 8. Synthesis of 70K heat shock proteins during virus infection. HeLa cells infected with Ad (a) or HSV-1 (b) were labeled with[35S]methionine for 1 h at the indicated times (hours postinfection), and labeled proteins were analyzed by two-dimensional gelelectrophoresis. The position of hsp70 is indicated by an arrow, the position of p72 is denoted by a closed arrowhead, and the position of grp78is indicated by an open arrow.

location of the gene, its chromatin structure, or the positionsof nuclear matrix-associated regions within or adjacent tothe gene constitute features which are recognized by viraltransactivators. In the case of Ad, attachment of the terminiof the viral genome to the nuclear matrix is required fortranscriptional activation by ElA (48). Alternatively, theremay be a special feature of the assembled transcriptioninitiation complex or the initiation process itself whichallows hsp7O to respond to viral inducers.Mechanism of transactivation of hsp7O by ElA. Results of

cotransfection studies (57) and of previous experimentsemploying the E1A- mutant d1312 (26) and 12S and 13Smutant viruses (59), and the high constitutive expression ofhsp7O in 293 cells, had all implicated the product of the ElA13S message as the transactivator of hsp7O. The IC4 cell line,in which ElA is dexamethasone inducible, has allowed us todemonstrate that activation of the endogenous gene can beeffected by ElA in the absence of an accompanying viralinfection. This is especially valuable in the case of hsp70,whose expression is extremely sensitive to environmentalperturbations.

In an attempt to identify promoter elements which conferElA responsiveness, our laboratory has previously carriedout an extensive 5' deletion and linker scanner mutationanalysis of the proximal region of the hsp7O promoter, whichwas linked to a reporter gene and cotransfected into HeLacells with an ElA-expressing plasmid (57). Results obtainedwith the series of 5' deletions indicated that sequences

upstream of the CCAAT element at -74 did not contributeto ElA inducibility. Our conclusion from studies of thelinker scanner mutants was similar to conclusions drawnfrom similar studies on ElA responsive-adenovirus promot-ers: no single element is responsible for ElA transactivation,and ElA may be acting through a basal transcription com-plex whose composition can vary. However, other models,which invoke ElA-mediated modifications or increases inconcentrations of cellular factors or which suggest thatspecific factors, especially ATF, are targets for ElA action,also have experimental support (11, 17, 33, 38, 45). With theexception of HSF binding to the HSE, whose relevance toElA-mediated activation is unclear, we can detect no

changes in interactions of factors with elements of the basalpromoter accompanying increased transcription of hsp7Oduring Ad infection. Although there is an ATF-like site at-35 in the hsp7O promoter, factors do not appear to bind tothis region prior to or during activation, suggesting thatactivation of the hsp7O promoter is not mediated throughATF or through changes in amounts or properties of otherbasal transcription factors which affect their avidity ofbinding to the promoter.

Activation of HSF during Ad infection. The basis for andsignificance of the low levels of activated HSF which appearat approximately 3 h p.i. and then persist through at least 12h p.i. have not been elucidated. Unlike HSV-1, Ad is notthought to exert deleterious effects on host macromoleculesduring the early phase of infection. Our laboratory has

V

* .4- 40 lo

40 a

VOL. 65, 1991

5690 PHILLIPS ET AL.

a b

FIG. 9. Synthesis of 70K heat shock proteins in SV40-infected

CV1 cells. Mock-infected (a) orSV40-infected (b) cells were labeled

with [35S]methionine at 22 to 28 h p.i. Untreated COS7 cells

also labeled for 6 h. CV1 cells (d) were labeled at 37°C for

immediately following a 2-h, 42.5°C heat shock. Labeled proteins

were analyzed by two-dimensional gel electrophoresis. The position

of hsp7O is indicated by an arrow, the position of p72 is denoted

a closed arrowhead, and the position of grp78 is indicated

open arrow.

previously reported activation of HSF upon treatment

cells with agents not commonly thought to induce protein

damage, however (54). Studies usingIC4 cells rule out the

possibility that activation of HSF is ElA mediated and that

such activation is necessary for induction of the endogenous

gene. This result confirms earlier transfection studies from

our laboratory which demonstrated that the HSE region

the promoter did not contribute to ElA inducibility. The

confounding observation of activated HSF in 293 cells may

reflect coexpression ofE1B in these cells (2). Two observa-

tions suggest that the levels of activated HSF present

Ad-infected HeLa cells do not contribute significantly

transcriptional induction of hsp70. First, hsp9o, a

which is strongly responsive to HSF-mediated induction,

not stimulated during adenovirus infection; second,

scription of hsp70 is declining during the period that HSF

levels are still rising. It should also be noted that HSF,

has DNA-binding activity as measured by a gel shift

may not necessarily be transcriptionally active (24a).

Mechanism of hsp7O induction during HSV-1 infection. The

kinetics of hsp70 induction during HSV-1 infection, as well

as its dependence on protein synthesis, imply but do

establish a role for one or more of the HSV-1 immediate-

early proteins in hsp70 transactivation. In contrast to Ad,

HSV-1 encodes five immediate-early proteins, three ofwhich have been implicated in transactivation. A major rolein transactivation has been assigned to ICP4; however,ICP27 is also necessary for appropriate regulation of earlyand late viral gene expression (14). ICPO, through synergisticaction with ICP4, may also contribute to transactivation ofviral promoters (14). Use of viruses which are null mutantsin each of these immediate-early genes, as well as cotrans-fection of the cloned immediate-early genes with hsp70promoter-reporter gene constructs, is required to rigorouslydemonstrate that one or more of the HSV-1 immediate-earlyproteins mediates the transcriptional induction of hsp7O.Both PRV IE and the homologous HSV-1 transactivator

ICP4 have been shown to possess DNA-binding activity.DNase footprinting of the hsp7o promoter, using extracts ofPRV-infected HeLa cells, revealed protection of two regionsof the hsp70 promoter, one at -66 to -87 and a secondregion just downstream of the TATA site (10). Furthermore,when extracts of cells infected with a virus encoding a

temperature-sensitive IE were heat treated so as to abolishthe ability of IE to stimulate transcription in vitro, binding tothe hsp70 promoter was similarly abolished. From theseresults, it was suggested that transcriptional activation byPRV IE may involve binding to promoter elements (10).Similar hypotheses have been proposed for the homologousHSV-1 transactivator ICP4 (43, 53), although recent reportsof ICP4 mutants with impaired DNA-binding capacitieswhich retain the ability to stimulate transcription do notsupport this concept (49, 50). Footprinting analysis of theproximal 120 bp of the hsp70 promoter during the first 5 h ofHSV-1 infection, when transcription of hsp70 is rapidlyinduced, revealed no detectable changes. Assuming thathsp70 transcriptional induction is indeed mediated by one ormore HSV-1 immediate-early proteins, our footprinting re-sults suggest that this transcriptional activation does notinvolve binding of viral proteins to proximal promoter ele-ments and that, as in the case of Ad-mediated transcriptionalstimulation, it is also not mediated by increased binding ofhost transcription factors.HSV-1-mediated release of factors from the hsp7O pro-

moter. The release of host transcription factors from thehsp70 promoter during transcriptional repression is a novelobservation. Although degradation of host mRNAs, includ-ing that of hsp70, during HSV-1 infection has been wellstudied, repression of host transcription has not been closelyexamined. Studies to establish the basis for and significanceof this phenomenon are currently under way in our labora-tory; moreover, it is evident from this observation thatgenomic footprinting can be used to detect virus-mediatedalterations in promoter-protein interactions.Consequences and significance of transcriptional activation

of hsp7O. In addition to the selectivity which characterizesthe viral activation of hsp70 transcription, our results indi-cate that the consequences of this activation, as measured bychanges in hsp70 mRNA levels and protein synthesis, arealso virus specific, presumably reflecting differences in viraleffects on host macromolecules. However, both Ad5- andHSV-1-infected cells show greater synthesis of hsp70 than ofother host proteins, which may reflect a requirement for thisprotein in viral replication or morphogenesis.

ACKNOWLEDGMENTS

We thank S. Fox for help and patience in preparation of thefigures and L. Sistonen, B. Wu, and S. Murphy for valuablecomments on the manuscript.

v -2* ~~~~~~~~~~~~~44w 0 :%O

r

- 00

v 0

^~~~~~ "

x I

d

**

..4.. . \

J. VIROL.

qo 'T _:-0 :. 0

w

40- 0.-W qw

- 00

4r,

40 0

TRANSCRIPTIONAL ACTIVATION OF THE hsp7O GENE 5691

This work was supported by grants from the National Institutes ofHealth, American Cancer Society, and March of Dimes to R.I.M.,by NIH and American Cancer Society postdoctoral fellowships toK.A., and by an NIH postdoctoral fellowship to B.P.

REFERENCES1. Abravaya, K., B. Phillips, and R. I. Morimoto. 1991. Heat

shock-induced interactions of heat shock transcription factorand the human hsp70 promoter examined by in vivo footprint-ing. Mol. Cell. Biol. 11:586-592.

2. Babiss, L. E., C. S. H. Young, P. B. Fisher, and H. S. Ginsberg.1983. Expression of adenovirus Ela and Elb gene products andthe Escherichia coli XGPRT gene in KB cells. J. Virol. 46:454-465.

3. Bagchi, S., P. Raychaudhuri, and J. R. Nevins. 1990. AdenovirusElA proteins can dissociate heteromeric complexes involvingthe E2F t-ranscription factor: a novel mechanism of ElA trans-activation. Cell 62:659-669.

4. Beckmann, R. P., L. A. Mizzen, and W. A. Welch. 1990.Interaction of hsp70 with newly synthesized proteins: implica-tions for protein folding and assembly. Science 248:850-854.

5. Bhat, G., L. SivaRaman, S. Murthy, P. Domer, and B. Thim-mappaya. 1987. In vivo identification of multiple promoterdomains of adenovirus EIIA-late promoter. EMBO J. 6:2045-2052.

6. Brunet, L. J., and A. J. Berk. 1988. Concentration dependenceof transcriptional transactivation in inducible ElA-containinghuman cells. Mol. Cell. Biol. 8:4799-4807.

7. Chirico, W. J., M. G. Waters, and G. Blobel. 1988. 70K heatshock related proteins stimulate protein translocation into mi-crosomes. Nature (London) 332:805-810.

8. Clos, J., J. T. Westwood, P. B. Becker, S. Wilson, K. Lambert,and C. Wu. 1990. Molecular cloning and expression of ahexameric Drosophila heat shock factor subject to negativeregulation. Cell 63:1085-1097.

9. Coilins, P. L., and L. E. Hightower. 1982. Newcastle diseasevirus stimulates the cellular accumulation of stress (heat shock)mRNAs and proteins. J. Virol. 44:703-707.

10. Cromlish, W. A., S. M. Abmayr, J. L. Workman, M. Horikoshi,and R. G. Roeder. 1989. Transcriptionally active immediate-early protein of pseudorabies virus binds to specific sites onclass II gene promoters. J. Virol. 63:1869-1876.

11. deGroot, R., N. Foulkes, M. Mulder, W. Kruijer, and P.Sassone-Corsi. 1991. Positive regulation of jun/AP-1 by ElA.Mol. Cell. Biol. 11:192-201.

12. Deshaies, R. J., B. D. Koch, M. Werner-Washburne, E. A.Craig, and R. Schekman. 1988. A subfamily of stress proteinsfacilitates translocation of secretory and mitochondrial precur-sor polypeptides. Nature (London) 332:800-805.

13. Devaux, B., G. Albrecht, and C. Kedinger. 1987. Identicalgenomic footprints of the adenovirus Ella promoter are de-tected before and after Ela induction. Mol. Cell. Biol. 7:4560-4563.

14. Everett, R. D. 1987. The regulation of transcription of viral andcellular genes by herpesvirus immediate-early gene products.Anticancer Res. 7:589-604.

15. Everett, R. D., and M. Dunlop. 1984. Transactivation of plas-mid-borne promoters by adenovirus and several herpes groupviruses. Nucleic Acids Res. 12:5969-5978.

16. Feldman, L. T., M. J. Imperiale, and J. R. Nevins. 1982.Activation of early adenovirus transcription by the herpes virus.Proc. Natl. Acad. Sci. USA 79:4952-4956.

17. Flint, J., and T. Shenk. 1989. Adenovirus ElA protein paradigmviral transactivator. Annu. Rev. Genet. 23:141-161.

18. Georgopoulos, C., D. Ang, K. Liberek, and M. Zylicz. 1990.Properties of the Escherichia coli heat shock proteins and theirrole in bacteriophage lambda growth, p. 191-221. In R. I.Morimoto, A. Tissieres, and C. Georgopoulos (ed.), Stressproteins in biology and medicine. Cold Spring Harbor Labora-tory Press, Cold Spring Harbor, N.Y.

19. Goldenberg, C. J., Y. Luo, M. Fenna, R. Baler, R. Weinmann,and R. Voellmy. 1988. Purified human factor activates heat-shock promoter in a HeLa cell-free transcription system. J.

Biol. Chem. 263:19734-19739.20. Graham, F. L., and J. Smiley. 1977. Characteristics of a human

cell line transformed by DNA from human adenovirus type 5. J.Gen. Virol. 36:59-72.

21. Gunning, P., P. Ponte, H. Okayama, J. Engel, H. Blau, and L.Kedes. 1983. Isolation and characterization of full-length c-DNAclones for human a-, P- and -y-actin mRNAs: skeletal but notcytoplasmic actins have an amino-terminal cysteine that issubsequently removed. Mol. Cell. Biol. 3:787-795.

22. Hickey, E., S. E. Brandon, G. Smale, D. Lloyd, and L. A. Weber.1989. Sequence and regulation of a gene encoding a human89-kilodalton heat shock protein. Mol. Cell. Biol. 9:2615-2626.

23. Honess, R. W., and B. Roizman. 1974. Regulation of herpesvirus macromolecular synthesis. I. Cascade regulation of threegroups of viral proteins. J. Virol. 14:8-19.

23a.Hunt, C. Unpublished data.24. Jog, P., B. Joshi, V. Dhamankar, M. J. linperiale, J. Rutila, and

K. Rundell. 1990. Mutational analysis of simian virus small-tantigen. J. Virol. 64:2895-2900.

24a.Jurivich, D. Unpublished data.25. Kang, P. J., J. Ostermann, J. Shilling, W. Neupert, E. A. Craig,

and N. Pfanner. 1990. Requirement for hsp70 in the mitochon-drial matrix for translocation and folding of precursor proteins.Nature (London) 348:137-143.

26. Kao, H.-T., and J. R. Nevins. 1983. Transcriptional activationand subsequent control of the human heat shock gene duringadenovirus infection. Mol. Cell. Biol. 3:2058-2065.

27. Kennedy, P. G. E., N. B. LaThangue, W. L. Chan, and G. B.Clements. 1985. Cultured human neural cells accumulate aheat-shock protein during acute herpes simplex virus infection.Neurosci. Lett. 61:321-326.

28. Khandjian, E. W., and H. Turler. 1983. Simian virus 40 andpolyoma virus induce synthesis of heat shock proteins inpermissive cells. Mol. Cell. Biol. 3:1-8.

29. Kingston, R. E., A. Cowie, R. I. Morimoto, and K. A. Gwinn.1986. Binding of polyomavirus large T antigen to the humanhsp70 promoter is not required for trans activation. Mol. Cell.Biol. 6:3180-3190.

30. Kingston, R. E., T. J. Schuetz, and Z. Larin. 1987. Heat-inducible human factor that binds to a human hsp70 promoter.Mol. Cell. Biol. 7:1530-1534.

31. LaThangue, N. B., K. Shriver, C. Dawson, and W. L. Chan.1984. Herpes simplex virus infection causes the accumulation ofa heat-shock protein. EMBO J. 3:267-277.

32. Leung, W.-C. 1978. Evidence for a herpes simplex virus-specificfactor controlling the transcription of deoxypyrimidine kinase.J. Virol. 27:269-274.

33. Liu, F., and M. R. Green. 1990. A specific member of the ATFtranscription factor family can mediate transcription activationby the adenovirus Ela protein. Cell 61:1217-1224.

34. McKnight, S. L., and R. Kingsbury. 1982. Transcriptionalcontrol signals of a eukaryotic protein-coding gene. Science217:316-324.

35. Moss, B. 1991. Poxviridae and their replication. In B. N. Fieldsand D. M. Knipe (ed.), Fundamental virology, 2nd ed. RavenPress, Ltd., New York.

36. Mosser, D. D., N. G. Theodorakis, and R. I. Morimoto. 1988.Coordinate changes in heat shock element-binding activity andhsp70 gene transcription rates in human cells. Mol. Cell. Biol.8:4736-4744.

37. Mueller, P. R., and B. Wold. 1989. In vivo footprinting of amuscle specific enhancer by ligation mediated PCR. Science246:780-785.

38. MuDer, U., M. P. Roberts, D. A. Engel, W. Doerfler, and T.Shenk. 1989. Induction of transcription factor AP-1 by adeno-virus ElA protein and cAMP. Genes Dev. 3:1991-2002.

39. Notarianni, E. L., and C. M. Preston. 1982. Activation ofcellular stress protein genes by herpes simplex virus tempera-ture-sensitive mutants which overproduce immediate earlypolypeptides. Virology 123:113-122.

40. O'Farrell, P. H. 1975. High resolution two-dimensional electro-phoresis of proteins. J. Biol. Chem. 250:4007-4021.

41. Oroskar, A. A., and G. S. Read. 1989. Control of mRNA

VOL. 65, 1991

5692 PHILLIPS ET AL.

stability by the virion host shutoff function of herpes simplexvirus. J. Virol. 63:1897-1906.

42. Ostermann, J., A. L. Horwich, W. Neupert, and F.-U. Harti.1989. Protein folding in mitochondria requires complex forma-tion with hsp60 and ATP hydrolysis. Nature (London) 341:125-130.

43. Paterson, T., and R. D. Everett. 1988. The regions of the herpessimplex virus type 1 immediate-early protein Vmwl75 requiredfor site-specific DNA binding closely correspond to those in-volved in transcriptional regulation. Nucleic Acids Res. 16:11005-11025.

44. Peluso, R. W., R. A. Lamb, and P. W. Choppin. 1978. Infectionwith paramyxoviruses stimulates synthesis of cellular polypep-tides that are also stimulated in cells transformed by Roussarcoma virus or deprived of glucose. Proc. Natl. Acad. Sci.USA 75:6120-6124.

44a.Phillips, B. Unpublished data.45. Richter, J. 1989. In vivo photocrosslinking reveals that tran-

scription factor binding to the mammalian ATF recognitionsequence is required for ElA-induced transactivation in injectedXenopus laevis oocytes. Nucleic Acids Res. 12:4503-4516.

46. Russell, J., E. C. Stow, N. D. Stow, and C. R. Preston. 1987.Abnormal forms of the herpes simplex virus immediate earlypolypeptide Vmwl75 induce the cellular stress response. J.Gen. Virol. 68:2397-2406.

47. Santomenna, L. D., and A. M. Colberg-Poley. 1990. Induction ofcellular hsp70 expression by human cytomegalovirus. J. Virol.64:2033-2040.

48. Schaack, J., W. Y.-W. Ho, P. Freimuth, and T. Shenk. 1990.Adenovirus terminal protein mediates both nuclear matrix as-sociation and efficient transcription of adenovirus DNA. GenesDev. 4:1197-1208.

49. Shepard, A. A., and N. A. DeLuca. 1991. A second-site revertantof a defective herpes simplex virus ICP4 protein with restoredregulatory activities and impaired DNA-binding properties. J.Virol. 65:787-795.

50. Shepard, A. A., A. N. Imbalzano, and N. A. DeLuca. 1989.Separation of primary structural components conferring auto-regulation, transactivation, and DNA-binding properties to theherpes simplex virus transcriptional regulatory protein ICP4. J.Virol. 63:3714-3728.

51. Strom, T., and N. Frenkel. 1987. Effects of herpes simplex viruson mRNA stability. J. Virol. 61:2198-2207.

52. Taylor, I. C. A., W. Solomon, B. M. Weiner, E. Paucha, M.Bradley, and R. E. Kingston. 1989. Stimulation of the humanheat shock protein 70 promoter in vitro by simian virus 40 largeT antigen. J. Biol. Chem. 264:16160-16164.

53. Tedder, D. G., R. D. Everett, K. W. Wilcox, P. Beard, and L. I.Pizer. 1989. ICP4 binding sites in the promoter and codingregions of the herpes simplex virus gD gene contribute toactivation of in vitro transcription by ICP4. J. Virol. 63:2510-2520.

54. Theodorakis, N. G., D. J. Zand, P. T. Kotzbauer, G. T.Williams, and R. I. Morimoto. 1989. Hemin-induced transcrip-tional activation of the hsp70 gene during erythroid maturationin K562 cells is due to a heat shock factor-mediated response.Mol. Cell. Biol. 9:3166-3173.

55. Tremblay, M. L., S.-P. Yee, R. H. Persson, S. Bacchetti, J. R.Smiley, and P. E. Branton. 1985. Activation and inhibition ofexpression of the 72,000-Da early protein of adenovirus type 5 inmouse cells constitutively expressing an immediate early pro-tein of herpes simplex virus type 1. Virology 144:35-45.

55a.Watowich, S. S. 1990. Ph.D. thesis. Northwestern University,Evanston, Ill.

56. Watowich, S. S., and R. I. Morimoto. 1988. Complex regulationof heat shock- and glucose-responsive genes in human cells.Mol. Cell. Biol. 8:393-405.

56a.Watowich, S. S., R. I. Morimoto, and R. A. Lamb. 1991. Flux ofthe paramyxovirus hemagglutinin-neuraminidase glycoproteinthrough the endoplasmic reticulum activates transcription of theGRP78-BiP gene. J. Virol. 65:3590-3597.

57. Williams, G. T., T. K. McClanahan, and R. I. Morimoto. 1989.Ela transactivation of the human HSP70 promoter is mediatedthrough the basal transcriptional complex. Mol. Cell. Biol.9:2574-2587.

58. Wu, B., C. Hunt, and R. I. Morimoto. 1985. Structure andexpression of the gene encoding the major heat shock proteinHSP70. Mol. Cell. Biol. 5:330-341.

59. Wu, B. J., H. C. Hurst, N. C. Jones, and R. I. Morimoto. 1986.The ElA 13S product of adenovirus 5 activates transcription ofthe cellular human HSP70 gene. Mol. Cell. Biol. 6:2994-2999.

60. Wu, B. J., R. Kingston, and R. I. Morimoto. 1986. Human hsp7Opromoter contains at least two distinct regulatory domains.Proc. Natl. Acad. Sci. USA 83:629-633.

61. Young, R. A., and T. J. Elliott. 1989. Stress proteins, infection,and immune surveillance. Cell 59:5-8.

J. VIROL.