Effect of genomic location on expression of beta-galactosidase mRNA controlled by the herpes simplex virus type 1 UL38 promoter.

0022-538X/92/052973-09$02.00/0

Copyright ©1992, American Society for

Microbiology

Effect

of

Genomic Location

on

Expression

of

1-Galactosidase

mRNA

Controlled by the

Herpes

Simplex Virus

Type

1 UL38

Promoter

SHERYL A. GOODART, JOHNF. GUZOWSKI, MARCIA K. RICE,ANDEDWARD K. WAGNER*

DepartmentofMolecular BiologyandBiochemistry, Universityof California Irvine, Irvine, California 92717 Received 14 November 1991/Accepted5February 1992

To examine the effect of genomic location on the details ofexpression of selected herpes simplex virus

promoters,wehaveconstructedrecombinationvectorsfor placing suchpromoterscontrollingthe

1-galactosi-dasereportergeneintotworegions oftheviral genomelackinganynearbypromoterorregulatory elements.

The first vector generates the promoter-B-galactosidase reporter gene inverted within the locus of the gC

(UL44) translational reading frame; the second replaces the LAT promoter and the first 600 bases of the

primary transcript in both copies of the RL region. These locations were chosen to obviate any possible

influenceofupstreambutnoncontiguous heterologousorhomologous DNAsequenceelementsuponpromoter activity.When thereportergenecontrolledby the strict late (-y) UL38promoterwasplaced inthegClocation, itwas significantlyless active than in its normal location; in contrast, promoteractivitywas comparable to

wild-typevalues when the promoterwasrecombined into the RL region. The low levelofactivity in the gC locationcould bepartially alleviated by theincorporation of additionalDNAsequencesupstreamof theUL38 promoter.Despitetheeffect of genomiclocationuponthelevelof expression, the kineticsof expressionineither

locationmirrors the wild-typeUL38 strict late kineticsof expression. Finally, weuseddeletional analysis to demonstrate thatno morethan29 basesofDNAsequence5' of the mRNAcapsitearerequiredforpromoter activity in either location; this result is consistent with earlier results of transient-expression assays and indicates that the UL38 promoter shares general features with other strict late (-y) herpes simplex virus promoters.

Transcripts encoded by herpes simplex virus type 1

(HSV-1)areexpressed duringthelytic replication cycle ina

sequential cascade pattern (14). Regulation of this expres-sionpatternoccursprimarilyatthe level oftranscription (21, 26, 29, 31). Transcripts encodingthe fiveimmediate-early(a)

genes are the first to be expressed; these gene products,

especiallyICP4 (ao4), are required for expression of subse-quentclasses of HSV-1genesduringlytic expression. Early

ordelayed-early

(pi)

genesarethesecondgroupofgenesto be expressed. Generallythe polypeptide products of these

genesareinvolved inreplication of viralDNA; their

expres-siontendstoreachmaximal levels around thepeakof DNA

replication and then decrease (31). Late transcripts are maximally expressed only after the onset of viral DNA

replication. They can begrouped into leaky-late (,3yor

yl)

genes and true or strict late (y or _Y2) genes; the difference between these two subclasses is in the stringency of the

requirement for DNA replication. Transcripts encoding

P-y

genes are expressed at low but readily detectable levels

before DNA synthesis, whereas -y transcripts evidence a strict requirement for DNA replication for any significant

expression (10).

The promoterscontrollingthe -yor strict late genes have been the subject of some investigation, especially those

controllingthegH (UL22), UL38, gC(UL44),and

Us11

genes

(11, 13, 16, 17, 22, 24). Although these promoters do not evidencearequirementforviral DNAreplicationfor

expres-sion in transient-expression assays, the extent of the

pro-motersdelineatedgenerallyagreeswiththosedeterminedby

generation of recombinant virusescontainingmodified

pro-* Correspondingauthor.

motersequences. Currentdataindicate that ypromotersare

quitelimited inlength, extendingfromarecognizableTATA

homology 28 to 30 bases 5' of the mRNA cap site and

includingsequences nearthetranscriptinitiationsite

extend-ing to 10to 20 bases 3' of thispoint.

The limited number ofDNAsequence elements

responsi-ble for the control ofytranscription is of interest becauseof

its potential utilityin saturation mutagenesis studies and as

probes for theidentification of transcription factors impor-tant in late gene expression. As noted, to date the most clearly recognizable promoter element found in these

pro-moters is the TATAhomology. Since this element is such a

notablefeatureofall-ypromotersstudied, it isreasonableto speculate that a protein factor(s) which recognizes this element might play a primary role in the expression of y

genes. Further, since-ygenes requireviral DNAreplication

forexpression,it is plausibleto speculatethatsuchafactor

might be altered in infectiontoenable it tobetterrecognize promoters.

Despitetheubiquityof-y promoterTATAboxes,there is some controversy concerning the actual sequence

require-ments for function. Johnson and Everett reported that in

transient-expression assaystheycould change theearlygD

(U,6)

promoterintoan operationallystrict latepromoteron aplasmid containing the HSV

ori,

by truncating it at -33, leaving the TATA element as the only recognizable

pro-moterelement(16).Morerecently, SteffyandWeirreported

thatthree TATA elements derived from theao4, Lthymidine

kinase (TK; UL23), and y gC (UL44) promoters could substitute for the y gH(UL22) TATA elementina

recombi-nantvirus, although the levels ofexpression differed signif-icantly dependingupon the exactelement used (24). These results argue against a strict sequence requirement for a 2973

on November 9, 2019 by guest

http://jvi.asm.org/

functional -y TATA box. However, Homa et al. concluded that the TK (UL23) promoter-leader (-37 through +57)

possessing

the intact TATA box

(ATATTAA)

could not

substitute for the gC promoter-leader

(-146 through

+124) inrecombinant virus

(11,

12). NogCmRNA wasexpressed from theTK

promoter-leader-gC

fusion until the TK TATA element was mutated to

GTATAAA,

which more closely resembles the gC sequence, TATAAATT. This mutated TATA box generated 10% of

wild-type (wt)

levels of gC mRNA.

Although

results of studies with recombinant viruses

bearing

mutated -y promoter

regions generally

agree with results of

transient-expression

assays, these

experiments

have beencarriedout atthe normallocation of the promoter in

question (11,

12,

24).

This allows the

possibility

that

noncontiguous

DNA sequence elements further upstreamof the TATA box will influence mRNA

expression.

Further, such

experiments

require

that the promoter in

question

control expressionofa genewhich is nonessential for viral

replication

in cell culture. We have

developed

two recombi-nation vectors which allow the insertion of any HSV pro-moter

controlling

the

3-galactosidase

indicator gene intotwo

regions

ofthe HSV-1 genome,onewithin the location of the nonessential gC

(UL44)

gene and the other replacing a

portion

ofthe

latency-associated transcription

unitand pro-moterinthe

long

repeat

(RL)

region.Thesetwolocationsof recombination will obviate the influence of any specific promoter elements

occurring

upstream of the "core" pro-moter and allow a rough estimate of the effect ofgenomic

position

on

expression.

In this system, thewt

transcription

unit and control sequences of the gene whose promoter is

being

studied are not altered and the

expression

of the wt mRNA

provides

an absolute internal standard for measure-ment.

We have chosen the promoter

controlling

the

UL38

gene as a useful model for the further

analysis

of -y promoters. This promoter iswithin a

280-bp region

of DNA which also contains the promoter for the

1B

UL37

mRNAtranscribed in the

opposite

direction

(2, 31).

These two promoters are

completely separable

in

transient-expression

assays. Se-quences

extending

from -29 to +10 relativeto the cap site

provide essentially

full

expression

uponviral

superinfection,

and DNA

extending

from -48to +99 basesissufficientfor the

expression

of a marker gene

(13-galactosidase)

in a recombinant virus. Three T+A-rich

regions

occur in this latter

region:

AAATat -38, TTTAAAat-29,andTATAat

-17.

Further, antibody-mediated

DNA supershift assays indicate that thecx4

protein

interacts with DNAnearthe cap

site; and,

finally,

the target for this interaction has sequence

homology

with promoter elements of the

yy42

(UL49)

gene

(6).

Inthecurrentreport,wedescribe the details of

expression

of UL38-controlled

1-galactosidase

in these recombinant virus systems. With the UL38 promoter, we found that

genomic position strongly

influences the level butnot kinet-ics of

expression.

Deletionalanalysisshows that in thevirus, the sequence

T'`TTAAA

between -29 and -23 of theUL38 mRNAcapsite is absolutely required fortranscription, but levelsofexpressioncanbereadilyincreasedbytheaddition

offurthersequence elements farther upstream.Thus,we can

clearly

differentiate sequence elements required for the kinetics of

expression

from those

important

in

influencing

levels of

expression.

The unusual nature of theUL38TATA

homology

supports the major body of experimental data

indicating

that a

specific

sequence ofthe -y TATAbox is of minor

importance

in

mediating

transcription.

MATERIALS AND METHODS

Cells and virus. HSV-1

(17syn')

was the parental strain usedto generate the recombinant viruses described below. Parental and recombinant virus infection of rabbit skin fibroblastsat amultiplicity of infection of 5 PFU per cellwas

used to generate the RNA used for Northern (RNA) blots and RNase protection assays. Vero cells were used for plaque assays, and both cell types were used during

recom-binant virus screening and isolation. The cells were main-tained at 37°C under 5% carbon dioxide in Eagle minimum essential medium containing 5% cadet calf serum, 100 U of penicillin perml, and 100 ,ug of streptomycin per ml (3, 5,7, 30).

Enzymes. Restriction enzymes were purchased from Boehringer Mannheim Biochemical Corp.; Bethesda Re-searchLaboratories,Inc.;orNewEngland Biolabs. Restric-tionbuffers recommended by the suppliers were used for all digestions. T7 RNA polymerase and RNasin used in RNase protection were purchased from Promega Biotec, Madison, Wis. T4polynucleotide kinase (Bethesda Research Labora-tories) was used to 5' end label appropriate markers for RNase protection. Finally, Klenow enzyme used for both labeling and blunting the ends of DNA fragments was purchased from Boehringer Mannheim.

Recombinant DNA. pBR322 and pGEM plasmid deriva-tives were grown in Escherichia coli DH5a. Growth and purification of plasmids were essentially as previously de-scribed (1, 7, 18). Oligonucleotide linkers were purchased from Collaborative Research, Inc., Waltham, Mass. DNA fragments for probe generation were isolated by digestion with the appropriate restriction enzymes followed by elec-trophoresis and either electroelution from agarose or over-nightelution from acrylamideat37°C inasolution of 500 mM ammonium acetate, 10 mM magnesium acetate, 1 mM EDTA, 0.1% (wt/vol) sodiumdodecyl sulfate, and 10 pgof E. coli tRNA(Sigma) per ml. Probeswereuniformly labeled by random priming andsynthesis by using Klenow enzyme with [ct-32P]dCTP (3 Ci/,umol; Amersham).

RNAisolationand fractionation. Methodsfor isolation of RNAbyusing guanidinium isothiocyanate-hot phenol have been previously described (27, 28). RNA generated from infected cells was isolated at specific times postinfection (p.i.) and either useddirectlyastotal RNAorfirst fraction-ated into poly(A)+ and poly(A)- fractions by use of oli-go(dT)-cellulose (Collaborative Research) chromatography (27).

Phosphonoacetic acid (PAA) or thymine

1-0-D-arabi-nofuranoside (ara-T)wasusedtoinhibit viral DNA replica-tion. Cellswereincubated with PAA(300

,ug/ml)

for30 min before infection or with ara-T (50 ,ug/ml) for 1 h before infection.Drugswerepresentduringviraladsorptionand for thetimes noted in thetext.

Actinomycin D(10 ,ug/ml)wasusedto study thestability of the mRNAs of interest by its addition 6 h p.i. The infectionwasallowedtoproceedforanother 1or2h before RNAisolation.

Recombinant virus. All sequence numbers are based on thecompletesequenceof the17syn+ strain of HSV-1(19).A gCrecombinantvectorextendingfrom theSall site atbase 94853in thegC regionto theHindIIIsiteatbase98823was

constructed by using appropriate clones from the 17syn+ strain of HSV-1. This construct contains an intactgC

pro-moter. The 3.9-kbpXbaI-BamHI cassette of pCAL5Div17 (6)contains 48basesof theUL38 promoter 5' of the cap site and 99 bases ofUL38 leader fused tothe bacterial

on November 9, 2019 by guest

http://jvi.asm.org/

tosidase gene with a 4-base linker. It is terminated with the bidirectional simianvirus 40(SV40) termination-polyadeny-lation sequences. The BamHI site was blunted with Klenow enzyme, andthecassettewasligated into the

XbaI-EcoRV-digested gC

recombinant vector, replacing 22 bp of gC sequences. This construct places the chimeric gene in the

opposite

orientation to and downstream of the gC cap and promoter. As constructed, both a truncated gC and a chi-meric

3-galactosidase

transcriptcanterminate at the bifunc-tional SV40 termination-polyadenylation signal. This con-structcontainsa

360-bpXbaI-Kp4nI

cassettecontaining UL38

promoter-leader

sequencesand the 5' portion of the

P-galac-tosidase gene. Promoterconstructswith less DNA 5' of the cap site were constructed by replacement of this cassette with

appropriate

XbaI-Kpnl

cassettes from the UL38 pro-moter constructs described elsewhere (6). A larger UL38

promoter-,3-galactosidase

recombinant vector including all the

UL37

and UL38 promoterregionsdriving

0-galactosidase

was

generated

by first inserting an XbaI oligonucleotide linker at the SalI site near the UL37-CAT junction in

pCALSDiv

and thencloningthis

XbaI-KpnI

cassette into the recombinantvector.

The

generation

of the recombinant virus FLA5, which contains 48 bases of the UL38 promoter and the 99-base leader fusedtothe,B-galactosidasegene recombined into the

RL

regions

of the genome substituting for the latency-associated

transcript

(LAT) promoter and 5' sequence, has been describedpreviously (6). Itwasconstructed by using a recombination vector plasmid in which 1,800 bp of HSV DNA of the RL

region

containing the LAT promoter and 5' sequencesof the

primary

latency-associatedtranscript were

replaced

with the same XbaI-BamHI fragment containing the indicator gene as used in the gC recombination vector

plasmid

described above. OtherRLrecombinant virus con-structs were

generated

by replacing either the

XbaI-KpnI

cassette or the XbaI-BamHI cassette of the recombinant vectorwith

appropriate

modifications or in reversed orien-tation in the

RL

region.

Recombinant virus was generated by first releasing the HSV-1 sequences in the plasmidvectors bydigestion with

appropriate

restriction enzymes and then cotransfecting cloned DNA sequences with infectious

17syn'

DNA into rabbit skinfibroblasts. Single-virus plaquesderived by

lim-iting

dilutionwere screened in 96-well plates by hybridiza-tion of transfechybridiza-tion

aliquots

with

32P-labeled

0-galactosidase

probe.

Wells containing recombinant viral DNAwere then

plaque purified

atleastthreemoretimesby repetition of the

following procedure.

Isolatedplaqueswerepickedfrom agar

overlays

of

limiting

dilutions of virus infecting Vero cells into 96-well plates containing rabbit skin fibroblasts and screened

by hybridization.

Theresulting isolates were then

analyzed by

Southern blot. Purified isolates containing the

complete promoter-p-galactosidase

construct inserted into the

gC region

for

gC

recombinant virusorinto both the IRL andTRL

regions

for the RL recombinant virus were used in this report.

Plaque-purified

isolates from different transfec-tionswere takenfor each construct, and thelevel of expres-sion of the chimeric

3-galactosidase

genewas the same as

assayed by

colorimetric methods. Theequivalency of levels of RNA

expression

from different recombinants bearing

differing lengths

ofupstreampromoter sequenceorin

differ-ent orientations confirmed that the expression seen was a result of the

specific

recombinations characterized andwas not due to some nonspecific mutation in another region of

the viral genome.

RNase protection assays. RNase protection assays have

been described elsewhere (4, 9). pGEM templates were used with T7 polymerase (Promega Biotec) by following instruc-tions supplied to generate the RNA probes. The probe was

labeled with 100

,uCi

of

[32PJUTP

(800 Ci/mmol; Amersham) per

,ug

of template DNA. The UL38 protection probe was a

HhaI fragment generated from the

UL38-,B-galactosidase

chimeric recombinant virus construct. The HhaI fragment spansthe region of DNA from 5 bp 5' of the UL38 cap to 28 bpwithin the

,-galactosidase

leader. This probe protects 99 bases of wild-type UL38 and 127 bases of

UL38-,B-galactosi-dase recombinant mRNA. A control dUTPase protection probe was a

BglII-Kjpnl

fragment extending 62 bp 5' of the capto 95 bp3' of the cap. This dUTPase probe protects 95 bases of 17syn'wild-type dUTPase mRNA.

More than 95% of total RNase protection probes synthe-sized with T7 polymerase were full length on the basis of results with analytical gels. A total of 1.5 x

106 cpm

(Cerenkov) of probe was hybridized with 10 to 20 pLg of appropriate RNA samples in 50,ulof80% formamide-40mM

piperazine-N,N'-bis(2-ethanesulfonic acid) (PIPES; pH

6.4)-400mM

NaCl-lmM EDTA at

60°C

for 11 h. The

mix-turewasthendigested by addition of 400,ulof RNase T1 (2 ,ug/ml; Sigma) in 10 mM Tris (pH 7.5)-300 mM

NaCl-SmM

EDTAfor 1 h at

30°C.

The digestion mix was next treated with 60 pLg of proteinase K (Sigma) at

37°C

for 15 min and phenol-chloroform extracted. RNase-resistant material was fractionated on an 8% acrylamide-8 M urea sequencing gel with DNA size standards.

Northern blot analysis. RNA transfer blots were hybrid-ized and washed essentially as described previously (6, 31). An 8-ml volume containing 4 x

107

cpm (Cerenkov) of radiolabeled random-primed DNA in the presence of 50% formamide, 0.4 M Na+, 0.1 M HEPES (N-2-hydroxyeth-ylpiperazine-N'-2-ethanesulfonic acid; pH 8.0), 0.005 M EDTA, and Denhardt's solution containing 100

,ug

of dena-tured calf thymus DNA per ml was used at

50°C

for 40h.

RESULTS

Placement ofUL38promoter-controlled expression marker in different locations in the HSV-1 genome. To eliminate the influence of transcription signals upstream of the UL38 core promoter, we generated recombinant viruses in which the

promoter-indicator gene cassette was placed in the inverted

orientation in the gC locus. The general structure of such

recombinants is shown in Fig. 1A; the 5' terminus of the UL38 promoter is bounded by anXbaI site at base 97669. This is 250 bases 5' of the cap site of the 730-base -y mRNA encoding the UL45 gene, which is coterminal with the 3' end of gC (8). The genomic structure and purity ofrecombinants were confirmed by Southern blots of viral DNA. As shown in Fig. 1A, the 3,600-bp Sall fragment was replaced by the 7,440-bp fragment containing the 3-galactosidase gene, var-ious portions of the UL38 promoter, and the SV40 termina-tion-polyadenylation sequences described in Materials and Methods.

The bidirectional activity of the SV40 polyadenylation signal was confirmed by analysis of RNA expressed by recombinant viruses containing the chimeric UL38-I-galac-tosidase reporter gene in the gC region. Figure 2A showsan

example in which Northern blots of poly(A)+ RNA from

cells infected with the UL38 promoter construct from -48 to +99 expressed a truncated 1.5-kb portion of gC mRNA (track ii) compared with the 2.7-kb wt gC mRNAexpressed in cellsinfected with the 17syn' strain (wt, track iii).

on November 9, 2019 by guest

http://jvi.asm.org/

SoIl-3600bp

Xba IRL IRSUS TRS

I J

UL38,SoSoUTPose

SV40(A+)

So 1500bpj, 3400bp Xba Sa

(948531) (9B1 (97669)(98422)

Sall* -7440bp

aQ)J3? bz4P" aR

VD C

lX lv eDC

-23

-9.4

-6.6 j-44

-23 !-2.0

23

-9.4

-66

-4.4

B. _TRL

IIR IRS -Rs

Xbo Xba Xbo Xbc

-_1 I;

poUUSalSal-- UL38 "-l So! Us

C-8767bp f E-6385bp

a20 <O~ 017TRL

X

'a-Sal 3400bp -Ps-, Xba Sa' C)

(5469) Xba (10636) 114232) / (7707)

SoIC -10,851 bp / IR, l9.

Sal -Pst- 3400bp So!

(114517) Xba (120,902)

(118659)

SolE*-8468bp

so

,-ga:probe

-4C

2.3-2.0

-gC probe

LATprobe 7_

934 -6; 6

-44

-2 3

-20.

rXC,/;51 ::! CXD/ O'

FIG. 1. HSV-1 recombinant viruses. The location and precise sequence numbers where the reporter gene-promoter constructs were

inserted into the HSVgenome areindicatedin themaps.(A) Schematic representationof theexpectedstructureof recombinantscontaining the UL38 promoter-controlled 1-galactosidase gene recombined into the gC region of the HSV-1 genome. Insertion of the promoter-n-galactosidase constructsgenerates aSalI fragment (7,440 bp) which is largerthan thewtfragment (3,600 bp). Below themapareshown diagnostic Southern blotsdemonstrating the insertion ofthe ,-galactosidase fragment. DigestionwithSall generates the expected-sized fragmentfor therecombinant viruseswhich hybridizestoboth 1-galactosidase-specific andHSV-1-specific probes.The latterprobewas a

945-bp PstI-SalI fragment (correspondingtobases 97477through98422of thecomplete17syn'sequence[191)cloned from thegC regionof

wtHSV-1 DNA. Bycontrast, wtviral DNA digested withSallgeneratestheexpected fragment which hybridized onlytothegC-specific probe. (B) Schematic representation of theexpected structureoftheRLrecombinants with theUL38 promoter-controlled 3-galactosidase

generecombinedinto boththeTRLandtheIRL regions ofthe HSV-1genome.Insertionintobothregionsgeneratestwonewfragmentsupon digestionwithSall:SalC*(10,851bp)andSalE* (8,468 bp).ThesefragmentsarelargerthanthosegeneratedbySalldigestionofwtHSV-1 DNA:fragments SaiC(8,768 bp)andSalE(6,385 bp), respectively.Therecombinantfragments hybridizetoboth 3-galactosidase-specificand

wtHSV-1-specific probes.The latterprobewas a605-bpHpaI-SalI fragment (correspondingtobases 120297through 120902) from theRL regionofwtDNA.Digestionwith XbaI andSalI generateda3.9-kbfragmentfrom theRLrecombinants.Thisfragment hybridizeswith both

3-galactosidaseand HSV-1probes.ThesamedigestofwtDNAgenerated6.3- and 5.2-kbfragments,whichweredetectableonlywith the HSV-1-specific probe.

Hybridization of blots with probes specific for UL38

mRNA revealed both the normal 1.9-kbUL38mRNA and a 3.4-kb chimeric mRNA containing the 99 bases of UL38

leader and ,B-galactosidase sequences, whereas RNA from

wt-infected cells expressed only UL38 mRNA (Fig. 2B,

tracks i andii). Hybridization of blots with ,B-galactosidase-specific probes reveals only the 3.4-kb mRNA (data not shown).

TheUL38

promoter-p-galactosidase

reporter cassettewas also recombined in both orientations into the RL regionof the HSV-1 genome by the substitution of the proximal promoterand5'regionsof thelatency-associated

transcrip-tion unit.Examples of Southernblots of such recombinants

showingrecombination into both repeatsareshown inFig. 1B; recombination results in thegeneration of the 10.8-kbp SalI fragment C* and the 8.5-kbp Sall fragment E* deriva-tivescontainingthe

P-galactosidase

reportergeneconstruct replacingthe8.8-and 6.4-kbp parental fragments.

Recombination of the reporter construct into the RL

region also yielded recombinant virus which exhibited

sig-nificant expression of the chimeric

UL38-p-galactosidase

mRNApromoteras wellas normal UL38mRNA (Fig. 2B, track iii). These recombinants expressed normal-sized gC

mRNA (Fig. 2A, tracki). Inotherexperiments (results not

shown)wedetermined thatrecombination ineither regionof theHSVgenome had noeffectuponthesizeor theamount ofaOmRNA,whichterminatesjust3' of the site ofinsertion

ofthereporter constructinto the RL region.

Amounts ofreportergenemRNAexpressedareinfluenced

by the genomic position of the promoter controlling them.

AlthoughUL38 promoter-controlled ,B-galactosidase mRNA wasefficiently expressed fromrecombinantviruses bearing

thecassetteeither in thegClocationorin theRL regions,the levels of chimeric

UL38-I-galactosidase

mRNA seen in Northern blotswasalwaysgreaterfor the latterconstructs. We confirmed and quantitated this difference in mRNA

levels by use of RNase protection assays with an RNase

probe extending from a HhaI site in the p-galactosidase

sequencesof the reportercassette28 bases 3'ofthejunction

of theUL38leaderto aHhaIsite 5' of the mRNAcap site (see Materials and Methods). Chimeric

UL38-,B-galactosi-dase mRNAprotected 127 bases of thisprobe,whereaswt UL38 mRNA protected 99 bases. Since the probe is

uni-formly labeled and since protected fragments have

essen-tianly

thesamesize for bothmRNAs,adirectcomparisonof the intensity of RNase-resistant hybrids results in a direct

measureof the relativeamountsof both mRNAs in infected

cells in which the amount ofwt UL38 mRNA provides an internal standard. Further, since the structure of the chi-meric mRNAwasthesamefor allrecombinants, any

differ-encesin levels of mRNAmustreflect differences in levels of

expression and not differential mRNA stability or other factors.

Our datawereidentical forUL38promoterscontaining48 and 29 bases of DNA sequence upstream of the cap site. Results ofatypical experimentareshown inFig.3;the ratio

A.

TRL UL

on November 9, 2019 by guest

http://jvi.asm.org/

[image:4.612.97.535.76.285.2]

o J

U L

3

03R

5Kb-.-(I) (ii) 1

030

034-A.

c

Xct CMj

P MIM2

03

I II

P M2

aw~~~~~~1

:,WI" 0

2Kb

UL38/'.2g-r3l-~

UL'38

-U[L38probe (i) ( ii (lii)

gCprobe

FIG. 2. Northern blot analysis of chimeric and gC mRNAs expressed during infection with recombinant viruses. Samples of poly(A)+ RNAwereisolated from rabbit skin fibroblasts 4 h after infection (hpi) with recombinant orwtvirus,fractionated, blotted, and hybridized. (A) Representative Northern blot probed with an

896-bp EcoRI-EcoRV DNA fragment specific for gC (UL44) mRNA (correspondingto bases96524 through 95628). The 1.5-kb mRNA expressed by the recombinants (trackii) indicates that recombina-tionintothe gC region results in truncation of the gCmRNA. Both the RL recombinant (track i) and parental wt (track iii) viruses

express full-length gC mRNA (2.7 kb), as expected. (B) Repre-sentative Northernblot of gC and RL recombinantsalong with their parental wt RNAprobedwith theXbaI-KpnI fragment from the -48/+99-,-galactosidaseconstruct.Thisfragment hybridizesto wt

UL38 mRNAand to ,-galactosidase-containing mRNA. All three virustypesexpresswtUL38 mRNA(1.9 kb), whereas only thetwo

recombinantsexpresstheUL38-3-galactosidasemRNA (3.4 kb).

of chimericto wtmRNAwas determineddensitometrically in two to four separate gels. Ratios varied ±10% between gels.Chimeric

U_38-3-galactosidase

mRNAexpressed from the UL38 promoter in the gC region recombinants was

reproduciblypresent at20%of thelevel ofwtUL38mRNA

(tracks i and iii). Incontrast, recombinant viruses with the samepromoter constructrecombined intoeitherorientation in the RL region expressed essentially twice as much chi-meric mRNAaswtUL38mRNA(tracksvandvi).Since the recombinantpromoter-reporter geneconstructispresent in

twocopies in the RLconstructs, ourdata indicate that the UL38 promoter is as active in the RL region as it is in its normal location.

Thelower level ofexpressionof theUL38promoterinthe

gClocation could bepartiallyalleviatedbytheincorporation

of DNAsequencesextending beyond48 bases of the mRNA cap site. Thus, a construct containing 232 bases of DNA

sequence inserted upstream ofthe 48-base UL38promoter resulted ina two- tothreefold increase in chimeric mRNA levels (Fig. 3, tracks ii and iv). This additional sequence

contained the DNAsequencesextendingfrom thecapsite of theUL37 mRNA, throughits promoter, andincludingUL38

sequences5' of -48. Onceinserted, itreconstructed thewt

UL37-UL38 juxtaposed promoters with the UL38promoter

linked to ,3-galactosidase.

Weconfirmed thevalidityofourmeasurementsof mRNA levels as an index of promoter activity by comparing the

stabilityof chimeric

UL38-3-galactosidase

mRNAwith that ofthewttranscriptby using actinomycinchaseexperiments.

Theratios ofwt tochimeric mRNA donotchange following

19*.-UL4 UL38/A -gal

i)38

(1)(11) OIO1 Ov) (v)(VI)

gC RL

Recombinant

B-48/+~48/99RL

B.~ ~~~~~~~~~~

ml _

:i

tX I- UL38/i3-gal

6 6+1h hpi hpi Actino

FIG. 3. Quantitation of the levelofchimeric

UL38-I-galactosi-dase RNAexpressionin recombinantvirus. Total RNAwasisolated

from rabbit skin fibroblasts and hybridized to theHhaI

UL38-1-galactosidase probedescribed in Materials and Methods. Thisprobe

protects 99bases ofwt UL38 mRNA and 127 basesof chimeric UL38-P-galactosidasemRNA. Afterhybridizationand RNase diges-tion, the RNase-resistant material was fractionated on an 8%

denaturing acrylamide gel with EcoRI-Hinfl- (Ml) and HaeIII (M2)-digested pBR322 DNA fragments as size markers. (A) Ten micrograms (tracks i, ii,v,andvi)or20p.g(tracksiii andiv)of RNA isolated 6 h p.i.was used for eachvirus. Shown areRNAprobe fragments protected by RNAisolated fromcells infectedwithtwo

gCrecombinant viruses(-48/+99,tracks i andiii;-280/+99,tracks

ii and iv) and two RL recombinant viruses [-48/+99, track vi; -48/+99(R)in which the whole promoter-reportergeneconstructis reversed, trackvJ. A lane containing 600,000 cpm (Cerenkov) of undigested probe(P)wasincludedtocheck thelabeling efficiencyof theprobe.The size of theundigested probeis 197bp owingtothe

presence of the polylinkersequence. (B) RNAwasisolated from

cellsinfected with the-48/+99 RLrecombinant virusat6 hp.i.and

again after a 1-h treatmentwith 10 p.g ofactinomycin D per ml starting6 hp.i.The markers Ml and M2werethesame asdescribed above.

60to120minofchaseafter 6hof infection with recombinant virus; an example of a 60-min chase is shown in Fig. 3B. Therefore, the incorporation of bacterial 3-galactosidase

sequences in the reporter mRNA does not lead to any

significantincreaseordecrease in thestabilityof chimericas

comparedwithwtUL38mRNAwhich would interfere with

Ij(9)

QO cy)

A 4- 4-YCl}

A.

ODC

52Kb-..*

-4,

I

4

on November 9, 2019 by guest

http://jvi.asm.org/

[image:5.612.76.282.80.250.2] [image:5.612.318.554.83.437.2]

t-) U a

A -48/t99 -48/+99 B + + +i

U~8f-AI.

B.~

*

R of chmeoCcDac m

M M M 3'm 13

expressionToa.RAa m in

UL38/a-9a1-

E

t |t~~~~~~~~~~-

I-

JPose

denangacla gl w eIds P D

UL38-__Ii

2 6 12 2 6 12

[image:6.612.327.552.74.334.2]

hp.

FIG. 4. Timecourse of chimeric UL38-B-galactosidase mRNA expression.Total RNAwasisolatedatvarious timesafterinfection, hybridized toanRNaseprotection probe,and subjectedtoRNase treatment, and the protected specieswerefractionated on an 8% denaturing acrylamide gelwithHaelll-digestedPBR322 DNA frag-ments as size standards (lanesM). (A) Total RNAs isolated from cells infected with either the-48/+99gCortheRLvirusat2, 6,and 12 hp.i. werehybridized tothe HhaIUL38-4-galactosidase probe described inMaterialsandMethods andin thelegendtoFig.3.(B) Total RNAs isolated fromcells infected withwtvirusorwith the -48/+99 gCorRL virusesat4and 6hp.i.werehybridized with the v dUTPase (IJL5O)-specific RNase protection probe described in Materials and Methods, which protects 95 bases ofwt dUTPase mRNA.

the use of the ratio ofchimeric to wt as a measure of the actual amounts of these mRNAs expressed in the infected cell.

Genomic location has no effect upon the kinetics ofUL38 promoter activity. We next examined the influence of ge-nomic position on the kinetics of UL38 promoter activity. TheexpressionofwtUL38 mRNA conformstostrict late

(y)

kinetics; thusexpressionof this mRNA absolutely requires viral DNAreplication (2, 31). Despite the variability in the levels of chimeric mRNA expressed from this promoter in the different recombinants, we found that its expression kinetics wereessentially thesamewhenit wasrecombined into either thegCorRLlocation in the genome. Results ofa

typical experiment are shown in Fig. 4. RNA isolated at

various times after infection withwtorrecombinant viruses

wasanalyzed for the presence of UL38 and chimeric UL38-,-galactosidase mRNA. NoUL38or chimeric mRNA from recombinants in gCorin theRLregion could be detectedat

2 hp.i.Incontrast,these mRNAswerereadily detectable by 6 to 12 h p.i. when DNA replication is occurring at high rates. As acontrol, RNaseprotection assays with aprobe specific for the

P

class dUTPase (UL50) mRNA yielded relatively high levels at 2to 4 hp.i., with significantly less RNArecoverable from cells by 6to8hp.i.Sample data for

4and8 hp.i. RNA are shown in Fig. 4B.

Astrictrequirement for DNA replication was also shown by using DNA synthesis inhibitors. In the presence of either 50 ,ug ofara-T perml or300 ,ugof PAA per ml, no wt or chimeric mRNA was expressed (Fig. SA). By contrast, similar samples of RNA hybridized with the

Pi

dUTPase

mRNAprobeyielded levels as highas orhigher than those in

the untreated controls(Fig. 5B).

BasalUL38 promoteractivity requires sequences extending

nofarther than 29 bases5' of the transcript cap site. We next

A. -48 2+99 -280/+99

cI--<0

U 38/

U,38

gC Recomt

B.

dT:Pose

--48/+99

r-1-~~

M l

| :

-s:._E r

I

< 0-u CLq<

40 q JUL38/0901

L_ Recomb

-48 /+99 -280/+99

1

I..~~~~~~~~~~.

2-.r 2.0:- CM

hA r n) r C, 1 Cr

gCRecomb

FIG. 5. Effect of DNAreplicationontheexpressionof chimeric

UL38-p-galactosidase

mRNA.Total RNAwasisolated 6 hp.i. from rabbit skinfibroblasts infected with recombinant virusin the

pres-ence orabsence(control[lanesC]) ofeither50 ,ugofara-T permlor

300 ,ug of PAA per ml. The RNAwas hybridized to an RNase protection probe, treated with RNase, and fractionatedon an 8% denaturing acrylamide gelwithHaeIII-digestedpBR322DNA frag-ments assize standards(lanes M). (A) RNAwashybridizedtothe HhaIUL38-13-galactosidaseprobe describedin Materials and Meth-ods and in the legend to Fig. 3. The panel on the left shows an experimentwith the -48/+99 and the -280/+99 gC recombinant viruses. The panel on the right shows an experiment with the -48/+99 RL recombinant virus. (B) ThesameRNAsamples from the gC recombinant virus-infected cells used in panel A were hybridized with the dUTPase probe described in Materials and Methods and thelegend toFig.4.

investigated the minimal sequence elements required for

measurable UL38 promoter activity with recombinants in both gC and the RL regions. Reporter constructs with promoters extending to -29 bases expressed the same

levels of chimeric

UL38-p-galactosidase

mRNA as did those with -48 bases of promoter in both locations (Fig. 6A). A promoter construct extending to only -23 bases, however, expressed nomeasurable level of chimeric mRNA. This is shown for a gC recombinant in Fig. 6B. These results are somewhat at variance with previously reportedresults oftransient-expression assaysinwhich the -23 promoter still had about30% of theactivityof themore

extensive ones (6). Extremely long exposures of similar RNaseprotection gelsandexaminationofhybridsformedby using more than twice the amount of infected-cell RNA revealedno RNaseprotection productcorrespondingtothe chimeric

UL38-p-galactosidase

mRNA(datanotshown).On the basis of controlexperiments with mixtures ofchimeric andwt mRNAs, this places levels ofexpression from the -23 promoter constructs as being lower than 5% of wt

levels.

on November 9, 2019 by guest

http://jvi.asm.org/

[image:6.612.68.301.77.230.2]

AC

A.

wt M P

gC Recomb.

8.

UL38(G-gal

I1

+ +

C) a)

N. N.3M

c'J M

UL38/

-k-go'

-:

q8

-

J838/)S-ga

- 138

() o)

c) r()

N~ C'.m

RL Recomb.

[image:7.612.64.291.74.386.2]

gCRecomb.

FIG. 6. UL38promoteractivity requiressequencesextendingno

furtherthan 29 bases 5' of thetranscript capsite. Total RNAwas

isolated 8 h p.i. from rabbit skin fibroblasts and hybridizedto the

HhaIUL38-,-galactosidaseprobedescribed inMaterials and

Meth-ods and the legend to Fig. 3. After treatment with RNase, the protected fragmentswerefractionated on an8% denaturing

acryl-amide gel with HaeIII-digested pBR322 DNA fragments as size

standards(M). (A)The leftpanelshows results ofRNaseprotection

with 10 p,g of RNA from cells infected with wt, -48/+99, and

-29/+99 gCrecombinant viruses.Thispanelalso includesalane in

which600,000cpm(Cerenkov)ofundigested probewasloaded(lane

P). The full-length probe was 197 bp because it included pGEM polylinkersequences. The panelonthe rightshowsthe results of using10p.gof total RNAfrom cells infectedwith wt,-48/+99,and -29/+99RLrecombinant viruses.(B)RNaseprotectionwith 10pg

of total RNA isolated from twogCrecombinantviruses, -29/+99

and-23/+99.

DISCUSSION

Oneof thereasonsforinitiatingtheexperimentsdescribed in this communication was to develop an internally con-trolled system in which the expression ofHSV promoters bearingdefined modifications could be measured and

com-paredwith that of the unmodifiedparentalpromoter from the sameviralgenome. Itisimportanttoemphasize that

differ-ences in expression do not reflect some sort ofnonspecific

interference due to transcription through the inserted

re-portergene-promoter construct since all RNase protection and Northern blot analyses confirm the lack of detectable levels ofsuch readthrough RNA. Such a system is clearly feasibleon the basis of the datapresented here,which also demonstrate that theactivity ofamodel promoter

control-lingthesametranscriptcanvarybymorethan afactor of 5 depending on where it is expressed on the viral genome.

Sucharesult haspractical significance in the design of HSV vectors for moving and expressing specific genes into spe-cific infected tissuesorcells. It is also important in interpret-ing the results of deletional and substitutional mutagenesis studieson HSV promoter function and in the identification of promoter sequences controlling the kinetics of expression ofparticularHSV genes.

Ourpicture of the essential sequence elements in -y

pro-moters is generally based on the modifications of the gC promoter in situby Homa et al. (11-13). These researchers havedemonstrated thataverylimited amount of sequence 5' of the capsite, aswellas some elements 3', of it are all that

are required for full activity. Further, these workers have suggestedthat the sequenceof the TATA box element itself (-28to -23 of thecap) isadetermining factor in its activity (11-13).Similar in situ modification analyses on other model -y promoters such as that controlling the expression of mRNAs encoding the gH (UL22),

-Y42

(UL49),

and

U.11

proteins generally confirm the limited extent of the 5' se-quencerequiredtocontrol strict latetranscription, although it is clear that no one specific TATA box sequence is conserved (16, 17, 22, 24). The present data on the UL38 promoter, which removes the influence of any possible sequence elements extending farther upstream than the regionsmodified in such in situ promoter modification stud-ies, fully support this general picture of HSV y promoters and demonstrate that very limited amounts of sequence elements upstream of thetranscription cap, along with other potentialelements 3' of thissite, provideall theinformation necessary for the expression oftranscripts following viral DNAreplication (6).

The dataontheextentof theUL38 promoter upstream of the cap sitepresentedhereareingeneralagreementwithour transient-expression data presented earlier (6), with the notableexceptionthat in the viral genome theregion extend-ing between -29 and -23 corresponding to the sequence TaITAAAisabsolutelyessential for promoterfunction; thus, the canonical sequence TATAoccurring at -17of the cap site cannot alone mediate any appreciable transcriptional activity.This result is incontrastwith the situation found in transient-expression assays, in which constructs extending to -27 and -23 of the cap site (the latter eliminating the whole upstream T+A-rich element) had activities on the order of30% of full promoterexpression.Such dataprovide agood demonstration ofthe limitations of the use of tran-sient-expressionassaysas asole measureof promoter activ-ityinmutational modificationanalyses.

Although all available evidence demonstrates that the 5' extentof HSVy promotersisquite limited and is transport-ableas cassettes todifferent locations of the viral genome, it isclear that the environment of thiscorepromotercanhave a significant role in the levels of its activity. In the UL38 promoter, the addition ofarelativelysmallamount ofextra

sequencecorrespondingtotheUL37promoter and upstream sequencesin their normal

position

relativeto

UL38

make the

y promoter

significantly

more active in thegC region used forgeneratingrecombinants.

Two observations suggest that the low level ofactivityof the

minimally

activeUL38promoter in

gC

isnot duetothe general inabilityofapromotertobe

efficiently expressed

in this region. First, inclusion ofextra sequence

significantly

increases

activity.

Second,

experiments currently

in

prog-ress in which 219 bases ofthe dUTPase promoter is

being

usedtodrive the

expression

of similarchimeric

P-galactosi-dase-containing

constructs demonstrate

essentially

equiva-lent levels of mRNA

expression

compared

with the wt

on November 9, 2019 by guest

http://jvi.asm.org/

mRNA. It isnot

clear, however,

whether the lowactivityof

the minimal

UL38

promoter is due to its kinetic class or whether minimal promoter elements of other classes of

transcripts

would also be expressed at lower than wt levels

in this region.

AlthoughthecoreUL38promoteris sufficienttodefine the

kinetic class of this gene, its

significantly

lower activity in

the

gC

location thanin theRLimpliesthat this minimal core

promoter does not contain all the information requirements for normal levels of mRNA

expression.

Although DNA

sequence elements inserted upstream of the core UL38 promoter in the

gC

region are important in influencing its

level of

expression, they

donotappeartobehighlyspecific. This conclusion is based on the fact that the same core

promoter

placed

inthe RL

region

inanorientation inwhich

the promoter

replaces

220 bases of the LAT promoter is

active at levels

approaching

those seen for the promoter in

its normal location. Further, the same level of activity is

seenwhenthewholecassetteisinverted in the RLregionso

that the sequences

immediately

5' of the UL38corepromoter are

completely

different. Thus, three different sets of up-stream sequences, onlyone of whichencodes another

pro-moter-regulatory region, permit

essentially full activity of thecore promoter.

Other data supportourconclusion that the local environ-ment of the promoter, as well as the specific sequences within

it,

is

important

in

influencing

levels ofexpression of HSV

transcripts.

Roemeret al.

(20)

found that the location ofan inserted SV40promoter-enhancer-controlled reporter gene in the viral genome affected the levels ofexpression.

Further,

removal of

recognizable

ao4-bindingsites within the promoter of the L TK

(UL23)

transcript has a profound

activity

upon its

expression

in transient-expression assays, but substitutionofsuchmodified promoters for thewtonein the viral genomedoesnotleadtoloss ofTKexpression (15). Asimilar observation has beenreportedforgD(23, 25).This result has been

interpreted

as indicating that the viral ge-nomehasa

high

degreeofredundancyin sequence elements which bind either cellular or viral transcription factors so

that loss ofone or a few in anygiven region canbereadily

compensated

for. Our data would suggest that the DNA sequence upstream of the core UL38 promoter which in-cludes the

UL37

promoter

acting

in the opposite direction and the DNA sequence in the RL regions in which this promoter was recombined contain enough such sites that UL38 promoter

activity

is

high.

However, our data also show that such sitesare not uniformlydistributed

through-out the viral genome since the region within the gC locus used forrecombination isa

significantly

poorerenvironment for

transcription

than is theRLregion.

ACKNOWLEDGMENTS

This work was supported byU.S. Public HealthService grants CA11861 and A106246. Further support was provided by the UC Irvine ResearchUnit in Animal Virology. J.F.G. isa predoctoral trainee and therecipientoftraininggrantT32GM07311 ("Structure andsynthesis ofbiologicalmacromolecules").

REFERENCES

1. Anderson, K.P., R.J. Frink, G. B. Devi-Rao, B. H. Gaylord, R. H.Costa,and E.K.Wagner.1981.Detailed characterization of the mRNAmappingintheHindIlI fragmentKregion of the herpessimplexvirus type 1 genome. J.Virol. 37:1011-1027. 2. Anderson, K. P., L. E. Holland, and E. K. Wagner. 1980.

Characterizationofherpessimplexvirus type 1 RNA present in theabsence of de novoprotein synthesis.J.Virol. 34:9-27. 3. Blair, E. D., C. C. Blair, and E. K. Wagner. 1987. Herpes

simplexvirusvirion stimulatoryproteinmRNAleader contains sequenceelements which increase bothvirus-induced transcrip-tion and mRNA stability. J.Virol. 61:2499-2508.

4. Dobson, A. T., F. Sedarati, G. B. Devi-Rao, W. M. Flanagan, M. J. Farrell, J.G. Stevens, E. K. Wagner, and L. T. Feldman. 1989. Identification of the latency-associated transcript pro-moter by expression of rabbit beta-globin mRNA in mouse sensory nerve ganglia latently infected with a recombinant herpes simplex virus. J. Virol.63:3844-3851.

5. Draper, K. G.,G. B. Devi-Rao, R. H.Costa,E. D.Blair,R. L. Thompson, and E. K. Wagner. 1986. Characterization of the genesencodingherpes simplexvirus type 1 and type 2alkaline exonucleases andoverlappingproteins.J. Virol. 57:1023-1036. 6. Flanagan,W.M.,A. G.Papavassiliou,M.Rice,L. B.Hecht, S. Silverstein, and E. K. Wagner. 1991. Analysis of the herpes simplex virus type 1 promoter controlling the expression of UL38, a true late geneinvolved in capsid assembly. J. Virol. 65:769-786.

7. Flanagan, W. M., and E. K. Wagner. 1987. A bi-functional reporter plasmid forthe simultaneous transient expression

as-say of two herpessimplex viruspromoters. Virus Genes 1:61-71.

8. Frink,R. J., R.Eisenberg,G.Cohen,and E.K. Wagner. 1983. Detailed analysisof the portionofthe herpessimplex virus type 1 genomeencodingglycoprotein C.J. Virol.45:634-647. 9. Gilman,M. 1989. Preparation and analysisofRNA, p.

4.7.1-4.7.8. In F. Ausubel, R. Brent, R. Kingston, J. Moore, J. Seidman, J. Smith, and K. Struchl (ed.), Current protocols in molecular biology.Wiley-Interscience, NewYork.

10. Holland, L. E., K. P. Anderson, C. Shipman, and E. K.Wagner. 1980. Viral DNAsynthesis isrequired for the efficient expres-sion of specific herpes simplex virus type 1 mRNAspecies. Virology 101:10-24.

11. Homa, F. L., J. C. Glorioso, and M. Levine. 1988. Aspecific 15-bp TATA boxpromoterelement is requiredfor expressionof aherpes simplexvirustype 1 lategene. Genes Dev. 2:40-53. 12. Homa, F. L., A. Krikos,J. C.Glorioso, and M. Levine. 1991.

Functional analysisof regulatoryregionscontrolling strict late HSV gene expression, p. 207-222. In E. K. Wagner (ed.), Herpesvirus transcription and its regulation. CRC Press, Inc., Boca Raton,Fla.

13. Homa, F. L., T. M. Otal, J. C. Glorioso, and M.Levine. 1986. Transcriptional control signals of a herpes simplex virus type1 late (gamma2) gene lie within bases -34 to + 124 relative to the 5' terminus of the mRNA. Mol.Cell. Biol. 6:3652-3666. 14. Honess, R. W., and B. Roizman. 1974. Regulation of herpesvirus

macromolecular synthesis. I. Cascade regulation of the synthe-sis of three groups of viral proteins. J. Virol. 14:8-19. 15. Imbalzano, A. N., A. A. Shepard, and N. A. DeLuca. 1990.

Functional relevance of specific interactions between herpes simplex virus type 1ICP4 and sequencesfrom the promoter-regulatory domain of the viral thymidine kinase gene. J. Virol. 64:2620-2631.

16. Johnson, P. A., and R. D. Everett. 1986. The control ofherpes simplex virus type-1 late gene transcription: a 'TATA-box'/cap site region is sufficient for fully efficient regulated activity. Nucleic Acids Res. 14:8247-8264.

17. Johnson, P. A., C.MacLean,H. S. Marsden, R. G. Dalziel,and R.D. Everett.1986. The product of gene

U.11

of herpes simplex virus type 1 is expressed as a true late gene. J. Gen. Virol. 67:871-883.

18. Maniatis, T., E. F. Fritsch, and J. Sambroolk 1982. Molecular cloning: alaboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.

19. McGeoch, D. J. 1989. The genomes of the humanherpesviruses: contents, relationships, and evolution. Annu. Rev. Microbiol. 43:235-265.

20. Roemer, K., P. A. Johnson, and T.Friedmann.1991. Activityof the simian virus 40 earlypromoter-enhancer in herpessimplex virus type 1vectors is dependent on its position, the infected cell type, and the presence ofVmwl75. J. Virol. 65:6900-6912. 21. Roizman, B., and A. E. Sears. 1990. Herpes simplex viruses and theirreplication, p. 1795-1842. In B. N.Fields, D. M. Knipe,

on November 9, 2019 by guest

http://jvi.asm.org/

R. M. Chanock, M. S. Hirsch, J. L. Melnick, T. P. Monath, and B. Roizman(ed.), Virology, 2nd ed. Raven Press, New York. 22. Shapira, M., F. L. Homa, J. C. Glorioso, and M. Levine. 1987.

Regulation of the herpes simplex virus type 1 late (gamma 2) glycoprotein Cgene:sequencesbetween base pairs -34to+29 control transientexpression and responsiveness to transactiva-tion by the products of the immediate early (alpha) 4 and 0

genes.NucleicAcids Res. 15:3097-3111.

23. Smiley, J. R.,D.C. Johnson, L.I.Pizer, and R. D. Everett. 1992.

The ICP4 binding sites in the herpes simplex virus type 1 glycoprotein D(gD)promoterarenotessential forefficient gD transcriptionduring virus infection. J. Virol. 66:623-631. 24. Steffy, K. R., and J. P. Weir. 1991. Upstreampromoterelements

oftheherpes simplex virustype1glycoprotein Hgene.J. Virol. 65:972-975.

25. Tedder,D. G., R. D. Everett, K. W. Wilcox, P. Beard, and L.I.

Pizer. 1989. ICP4-binding sites in the promoter and coding regions of the herpes simplex virus gD gene contribute to

activationofin vitrotranscription by ICP4. J. Virol. 63:2510-2520.

26. Wagner, E. K. 1985. Individual HSV transcripts:

characteriza-tion of specific genes, p. 45-104. In B. Roizman (ed.), The herpesviruses,vol. III.Plenum Press,NewYork.

27. Wagner, E. K., G. B. Devi-Rao, L. T. Feldman, A. T.Dobson, Y. F.Zhang, W. M. Flanagan, and J. G. Stevens. 1988.Physical characterization of the herpes simplex viruslatency-associated transcriptin neurons.J. Virol.62:1194-1202.

28. Wagner, E. K., W. M. Flanagan, G. B. Devi-Rao, Y. F.Zhang, J. M. Hill, K. P. Anderson, and J. G. Stevens. 1988. Theherpes simplexviruslatency-associatedtranscript is spliced during the latentphase of infection. J. Virol. 62:4577-4585.

29. Weinheimer, S. P., and S. L. McKnight. 1987. Transcriptional andpost-transcriptional controls establish the cascade ofherpes simplexvirusprotein synthesis.J. Mol.Biol. 195:819-833. 30. Zhang, Y. F., G. B. Devi-Rao, M. Rice, R. M. Sandri-Goldin,

andE. K.Wagner. 1987. The effect of elevated levels ofherpes simplex virus alpha-geneproductsonthe expression of model early and lategenesin vivo. Virology157:99-106.

31. Zhang, Y. F., and E. K. Wagner. 1987. The kinetics of

expres-sion ofindividual herpes simplex virustype1transcripts. Virus Genes1:49-60.

on November 9, 2019 by guest

http://jvi.asm.org/