Structural and kinetic analyses of herpes simplex virus type 1 latency associated transcripts in human trigeminal ganglia and in cell culture

Structural and kinetic analyses of herpes

simplex virus type 1 latency-associated

transcripts in human trigeminal ganglia and in

cell culture.

P R Krause, … , J M Ostrove, S E Straus

J Clin Invest.

1990;

86(1)

:235-241.

https://doi.org/10.1172/JCI114689

.

Only one herpes simplex virus type 1 (HSV-1) gene is expressed in sensory neurons of

latently infected animals and humans, yielding two RNAs, called latency-associated

transcripts (LATs). The LATs appear to modulate virus reactivation. In mice and rabbits the

5' origins, kinetics of synthesis, and splicing pattern of the LATs are well established.

Because these details of LAT structure and expression have not been defined in humans,

we sought to do so. Using primer extension and Northern hybridization analyses, we

demonstrate that in human trigeminal ganglia, the smaller (1.35 kb) HSV-1 transcript differs

from the larger (1.85 kb) LAT by excision of an intron near its 5' end; they are otherwise

colinear, and 5' coterminal. In infected cells only the 1.85 kb LAT is detected. Its expression

is inhibited by cycloheximide or acyclovir, indicating this LAT is synthesized late in the viral

replicative cycle. All of these features of the LATs in humans are consistent with those

reported in rabbits and mice and further validate the animal models of human HSV-1

infection.

Research Article

Find the latest version:

http://jci.me/114689/pdf

Structural

and

Kinetic Analyses

of

Herpes Simplex

Virus

Type

I

Latency-associated

Transcripts in Human Trigeminal

Ganglia

and

in

Cell

Culture

Philip R. Krause, Kenneth D. Croen, Jeffrey M. Ostrove, and Stephen E. Straus

Medical Virology Section, Laboratory of Clinical Investigation, NationalInstitute ofAllergy and InfectiousDiseases,

National Institutes ofHealth, Bethesda, Maryland 20892

Abstract

Only one herpes simplex virus type 1 (HSV-1) gene is

ex-pressed in sensory neurons oflatently infected animals and humans, yielding two RNAs, called latency-associated tran-scripts(LATs).TheLATs appeartomodulate virus reactiva-tion. In mice and rabbitsthe 5'origins, kineticsofsynthesis,

andsplicingpattern of the LATsarewell established. Because thesedetails ofLATstructure and expressionhave notbeen

definedin humans,wesoughttodoso.Using primer extension

andNorthern hybridization analyses,we demonstrate that in humantrigeminal

ganglia,

thesmaller (135 kb)HSV-1

tran-script differs fromthelarger(1.85 kb)LATbyexcision of an intron near its5' end;theyareotherwise colinear,and5'

coter-minal. Ininfected cellsonly the1.85kb LATis detected.Its

expression isinhibitedby cycloheximide oracyclovir, indicat-ingthisLATissynthesized lateinthe viralreplicativecycle. Allof these featuresofthe LATs in humansareconsistent with

those reported in rabbits and mice and further validate the animal models ofhuman HSV-1 infection. (J. Clin. Invest. 1990. 86:235-241.) Key words: herpes infection * viral latency-neurons

Introduction

A fundamental hallmark ofherpessimplex virus (HSV)' in-fectionof humans is its abilitytorecur(1). Abundant evidence indicatesthat recurrentinfections ariseas a resultof

reactiva-tion oflatentvirus insensory nerveganglia. The virus gains

access tothe

ganglia

early in infection by local spread from

epithelial

cells to sensory nerve endings and ascent up the

axon. The latent state is

characterized

by

persistence

of the

viralgenomeinanepisomal form in neuronal nuclei (2). Only one ofthe80ormoreviralgenesis active untilasyet unde-finedsignals trigger reactivation via expression ofthe remain-inggenesand the resultantreplicationand assembly of prog-enyvirions. Virusthenspreads centrifugallytocausea

recur-rentmucocutaneousinfection.

Transcription ofthesinglegenethatis active inlatency has

primarily

beenexamined in rabbitsand

mice,

inwhich studies

Addressreprint requests to Dr. Stephen E. Straus, Bldg. 10,Room 1

INI

13, 9000 RockvillePike, Bethesda,MD20892.

Receivedfor publication4December 1989 and inrevisedform27

February1990.

1. Abbreviations usedinthis paper:HSV, herpes simplex virus;HSV-1,

HSV type 1; LAT,latency-associated transcripts.

The Journal of ClinicalInvestigation,Inc.

Volume86,July 1990,235-241

of HSVtype 1 (HSV- 1) infectedsensoryganglia revealedthe

expression

oftwoabundant viralRNAs,termed

latency-asso-ciated transcripts (LATs) (3-9). Thesmaller LAT, orpossibly a pair of transcripts estimated to be 1.3-1.5 kb in size, are

spliced forms ofalarger2.0-2.3-kb transcript.The LATs ap-pear to be localized to the ganglion neuronal nuclei. They possess common 5' startsites within the long inverted repeat sequencesofthe genome and there are thus twocopiesof the LAT gene pervirus. They overlap the geneencodingthe viral

immediate-early protein ICP0 (see reference 10 for a

discus-sion of

ICPO),

but are transcribed from the opposite DNA

strand relativeto the ICPO message. Studies of

ganglia

from mice latently infected with HSV-1 indicate thatthe LATs are notrequired fortheestablishmentormaintenance oflatency. Rather, they appear to play aroleinpromoting viral reactiva-tion(1 1-13).

Studies of human tissues also demonstrated abundant

HSV-I LATsin latently infected

trigeminal

nerve

ganglion

cell

nuclei. The LATs were estimated to be 1.85-2.0 kb and

1.35-1.5kb inlength

(14-19).

Noother viral

transcription

has beendetected inhumantissues.

Inthepresentworkweundertooka detailed examination ofthe nature, structure, and

expression

of HSV-1 LATs in

human tissues. Should these aspects ofHSV-1 latency in

humantissues resemble the

findings

inrabbits and

mice, they

would further validate the use of the animal models in the

study ofHSVdisease

pathogenesis.

Methods

Cellsandinfections.Vero cells weregrown at

370C

in 5%CO2

incuba-torsina1:1mixture of minimal essential medium(MEM)/199 (Qual-ity Biological, Inc., Rockville, MD), supplemented with 10% FBS

(Quality Biological, Inc.),glutamine (29.2 mg/ml), aureamycin(2.6 mg/ml), streptomycin (25 mg/ml), and penicillin (100 U/ml). For some experiments, cellswere treated withacyclovir (30

gM;

Bur-roughs-Wellcome Co., Research Triangle Park, NC), beginning 24h

before infection, while in others, cycloheximide (50 gg/ml,

Boehringer-MannheimBiochemicals, Indianapolis, IN) was added 3 h before infection.

Cells wereinfected with a plaque titered stockof

HSV-l

strain KOS,orwithalow passage

HSV-l

clinical isolate recovered from a

recurrent lip lesion ofan immune competent patient. Human

trigeminal gangliaremovedatautopsyperformed within 24hof death were frozen and stored at -70'C until the RNA was extracted as described(15).

RNApurification. RNAwaspurified from infected oruninfected

Vero cells by lysis in situ and from human trigeminal ganglia by

poly-tronhomogenizationin4Mguanidinium thiocyanate (Fluka, Buchs,

Switzerland)with 0.5%sarkosyl.RNA waspurified by the guanidin-ium thiocyanate/cesium chloride method as described (20). Yields wereestimated by spectrophotometry, and averaged 150

Mg/75

cm2 tissue culture flask and 100

_jtg

per

trigeminal

ganglion.

DNAhomologoustothelong repeats of HSV-1 strain KOSwas

single-stranded RNA transcription from inserts in the plasmid templates. RNA markers for hybridization controls were prepared by in vitro transcription, from the appropriate clones using SP6 or T7 polymer-ases (21),transcribed in the same direction as the LATs. These reac-tionsyielded full length RNAs homologous to the HSV-I 1,660-bp Eco RV-Hinc II fragment (bases 1,479-3,139 relative to thejunction ofthe unique long segment[UL]and the internal repeat[IRLI,seeFig. 1 for a map), the 2,866-bp Sph I-Sph I fragment (bases2,913-5,779),and the 786-bp Sph I-Sph I fragment (bases 2,127-2,913).

Northern hybridizations. RNA samples (- 5 tg per lane) were subjected to electrophoresis in 20 mM morpholinopropanesulfonic

acid (MOPS; Sigma Chemical Co., St. Louis, MO) buffer (pH 7.0) through 1.5% agarose-6%formaldehydegels,aspreviously described (22). RNA was transferred to nitrocellulose paper or Nytran mem-branes (Schleicher & Schuell, Inc., Keene, NH), hybridized, and washed according to the instructions of the manufacturer. Membranes were autoradiographed with Kodak XAR-5 film in cassettes with in-tensifyingscreens.

Hybridization probes. Oligonucleotide probes designedto detect

RNAtranscription in the region ofthe LATs (see TableI)werechosen based on the knownDNAsequence of HSV-1 strain 17,kindly pro-vided by D. McGeoch (10, 23). The probes (20-mers) were manufac-tured ona DNAsynthesizer(model380A;AppliedBiosystems, Inc., FosterCity, CA), purified free of partial lengtholigonucleotidesby use of oligonucleotidepurificationcolumnspurchased fromApplied Bio-systems,Inc., and 5' end-labeled in the presence of32P-_y-ATP (Amer-sham Corp., Arlington Heights,IL) usingT4 polynucleotidekinase (Life Technologies/Bethesda Research Laboratories, Gaithersburg,

MD)or3' end-labeled in the presence of32P-a-dATPwith terminal

deoxynucleotidyl transferase (Boehringer-Mannheim).

Unincorpor-atedcounts were removedbyethanol precipitation. Approximately

6-10 X 106 cpmwereused per Northern hybridization.Gel-purified

double-stranded DNA probes from the HSV-l KOS Sal I-Hinc II

subfragmentregion(seeFig. 1)werepreparedaspreviouslydescribed (15) and radiolabeled in vitroto aspecific activityof - 108_109

cpm/ggby nick translation in the presence of 32P-a-dCTP (Amersham Corp.).

Primer extension. Primerextension wasdone byaslight

modifica-tion of the procedure ofSpalholz (24). RNA(10 ug)purified from tissue culture cellsorhumantrigeminal gangliawasincubatedwith 10

mM methyl-mercury hydroxide (Morton Thiokol, Denver, CO) at

42°Cfor 10 min. 2-mercaptoethanol (Sigma Chemical Co.)wasadded

to afinal concentration of 200mMandtheRNA washeldatroom

temperature for 5 min.5'end-labeled probes (- 2 pmol perreaction)

forprimerextensionwereannealedtotheRNAin the presence of 100

mM sodium chloride, 20mMTris (pH 7.6), and 0.1 mMEDTA by

heating at90°Cfor 3 min,incubatingat56°Cfor 20 min, and then

coolingto42°Cover20 min. Primer extension with clonedMoloney

Murine Leukemia Virus(M-MLV)reversetranscriptase(Bethesda

ResearchLaboratories)wasconductedat37°Cfor 1haccordingtothe instructions of themanufacturer, using I mMdNTPs (Bethesda Re-search Laboratories), 50 ug/mlactinomycin D

(Boehringer-Mann-heim),and1000g/mlnuclease-free BSA (BethesdaResearch

Labora-tories).Primer extended productswerethen extracted withphenoland

chloroform, precipitated,and washed in ethanol, andone-quarterof theproductwassubjectedto electrophoresis through8 M urea, 6%

polyacrylamidedenaturinggelsat60Wfor4h,dried,and autoradio-graphed asdescribed above forNorthern blot analyses. Primer ex-tendedproductsweresized relativeto a DNAsequenceladderfroman unrelatedgene.

Results

Splicing of HSV-J

LATs. In earlier

studies

of sequences that

are

transcribed during

acute and latent HSV-1

infections,

cloned genomic

fragments spanning

the region of the viral long repeats (Fig. 1) weregel purified and used as probes in

HSV-1

TRL

I

UL

I a - N- XR

--- -l! F4

-;

IR, IRS Us TRS

i=

}

\ wN

\dt

.1

I

Eco RV SphI SphIHincII SalI HincI1 SphI BanHI

LATENCY- | *

--ASSOCIATED

TRANSCRIPTS < -

-A _-CPO

Figure1. Mapof HSV-Igenome and summaryof data onphysical

characterization of thelatency-associated transcripts.Basesare la-beled relativetothestartof thelongrepeats.

Northern hybridizations.Inthoseexperimentsweshowedthat

a

_{probe consisting}

ofthe

1,256-bp

Sal I-Hinc II

fragment

(bases 3,743-4,999

relativetothe junctionof the repeatwith

the

unique long

_MULl

segment) hybridizes toboth the

larger

and smallerLATsin

latently

infected human trigeminal

gan-glia,

but

only

to the

larger

LAT and the overlapping

ICPO

transcript

in acutely infectedVero cells(15). To more

accu-rately

map the LATs andtodiscern regionsofidentity between

the

larger

and smallertranscripts, Northern

hybridizations

wereperformed using synthetic oligonucleotideprobes

repre-senting

the sequences spanning this area of the genome

(Fig. 1).

Each oligonucleotide probewashybridizedtofour differ-entRNA preparations.Theseincludedaganglion RNA dem-onstratedby previous Northern hybridizationtocontain both thelargerand smallerLATs,a ganglion RNAwhich showed

noevidence ofLAT transcription upon Northern

hybridiza-tion,

RNA from Vero cellsinfectedfor 24 h withHSV-1 ata multiplicityoftwoplaque formingunits(pfu)percell, andin

vitro transcribedRNAmarkersrepresenting the 786-bp Sph I

fragment

(spanning bases 2,127-2,913 relativetothe

junction

ofthe repeat and unique sequences), the 1,660-bp Eco

RV-Hinc II fragment (spanning bases 1,479-3,139), and the

2,866-bp SphIfragment (spanningbases2,914-5,779). These markerswere usedto verifythe quality of thehybridization conditions and, moreover, these controls allowed us toeasily discriminatebetween the 1.85-kb HSV-1 LAT and 18S ribo-somalRNA,whichmigratetosimilar positions in the formal-dehyde-agarose gels.

Typical Northern hybridizationsusing theoligonucleotide probesare shownin Fig. 2. The oligonucleotide probe

span-ning bases 2,282-2,301 detected only the 786- and 1,660-b

control marker RNAs. The oligonucleotide probe representing

bases 2,400-2,419 detectedthese same markerRNAs,both the 1.85-kb and the 1.35-kb LAT RNAs from a known positive ganglion, but only the 1.85-kb LAT in RNA from HSV-1-in-fected Vero cells. Neither LAT was detected in thenegative

ganglion. The probe spanning bases2,711-2,730detected only the 1.85-kbLAT in RNA from HSV- 1-infected Vero cells and

fromthepositive ganglion,aswell as the appropriate marker RNAs. The 2,866-b marker could now be detected duetothe location of this oligonucleotide's sequence. Failure of the probe to detect the smallertranscriptinseveral similar hybrid-izations indicates that the 1.35-kb LAT lackssignificant

4,089-M I +

.... .e

1:i~::-

2866-1850 -1660 1350

-786

-2282-2301 2400-2419 2711-2730 4089-4108 4499-4518

Figure 2.Typical Northern hybridizations using oligonucleotide probes to detect theHSV-ILATs. The larger LAT is seen at 1.85 kb, the smaller LATis at 1.35 kb. Eachblotcontains RNA fromHSV-1-infected Vero cells (MOI 2, harvested at 20 h, lanes marked I), RNA from a humantrigeminal ganglion preparation previously shown to include (lanes marked +) or not to include (lanes marked -) both LATs by

North-ernhybridization, or in vitro synthesized control marker RNAs at 2,866 b (from bases 2,913 to 5,779), 1660 b (from bases 1,479 to 3,139), and 786 b(frombases 2,127 to 2,913) (all in lanes marked M).

4,108, however, detected both the 1.85-kb and the 1.35-kb LAT RNAs in ganglia extracts, andagain only the 1.85 kb LATfrom infected cells. The probe representing bases 4,499-4,518detectedonly the control RNA, and neitherLATfrom ganglion orinfectedVerocell RNAs. Smallamountsof 28 S ribosomalRNA

(migrating

at - 4.4kb)werealsoseenin these

hybridizations. Resultsofhybridizations using these and other

oligonucleotide

probes

spanning

theLAT

region

are

summa-rizedinTableI.

Theobvious interpretation of the results ofthis series of

experimentsis thattranscriptionof the HSV-1 LATs is initi-ated somewhere in the region of bases2,301-2,307 andthat the smaller RNA is a spliced form of the larger one with an

intronexcised from theregionbetween bases2,419-2,711 and 3,134-3,201. We found no evidence for additional splices using these probes.

LA Tinitiation site. Topreciselylocalize the 5' end of these LATs, we performed a primer extension analysis on RNA

Table I. ResultsofNorthernHybridizations ofRNAExtractedfromHSV-1KOS-infectedVero Cells andLatently InfectedHuman

TrigeminalGanglia Using

Synthetic

Oligonucleotide

Probes

Transcriptdeted Ganglia Verocells

Position Sequence 1.85 kb 1.85 kb 1.35 kb

2282-2301 TGGAAACGCGGCGTCTTTGT - -

-2307-2326 GGGAGAAGCAGGTGTCTAAC + + +

2400-2419 CTGGTGTGCTGTAACACGAG + + +

2711-2730 CGTATCCTCGCTTTAGGAAC + +

-3115-3134 CTATAGGGCAACAAAGGATG + +

-3202-3221 GTGTGGGTGTTAAGTTTCCG + + +

3487-3506 CTGTCTGTGTTGGATGTATC + + +

4089-4108 GGTCTCTAGCGTGGTCGCCC + + +

4176-4195 GAGGAAGTGTGCCCGGAAGA + + +

4499-4518 GAAAACCCCTCCCCCCAGTC

Shownarethepositionsand sequences of theoligonucleotide probesused and the results of NorthernhybridizationwithRNAextractedfrom

HSV-1 infectedVerocells andlatentlyinfected humantrigeminalganglia. Oligonucleotide probesarenumberedrelativetothe

junction

of the

uniquelong (UL)andlonginternal repeat(IRL)segments.

hA I - _{+ M I} - _{+ m I +}

1.

:.

M I - +

fromhuman trigeminalganglia. Theprimerused in these

reac-tionsspannedbases2,400-2,419, andhybridized toboth the

1.85-kband the 1.35-kbLATs, as shown in Fig. 2. As controls, RNAswere also chosenfromextractsofuninfectedVero cells, HSV- 1 KOS-infected cellsandfromcellsinfected witha low passage HSV-1 clinical isolate. RNAsfromthree human tri-geminalgangliawerestudied, twoof whichwerepositive and

oneof whichwasnegative forthe LATsby Northern

hybrid-ization. The sizes oftheextension products in each reaction

weredetermined relativeto a DNA sequence ladder.

Autoradiogramsofgelscontaining primer extendedRNA

from Verocells infected with HSV-l strain KOS showedan

intensediscreteband at 1 17basesfromthe 5' endofthe oligo-nucleotideprimer,nocorresponding bandwasseenusing

un-infected Vero cells (Fig. 3). Primer extension ofRNAfrom

cellsinfected withaclinicalisolategaveriseto a band onebase smaller.Extension reactions with ganglionRNApreparations previouslyshown tocontainboth LATs byNorthern hybrid-ization eachyielded asingle intense discrete band(ganglia 1

and 2), and the primer extension product ofRNA from a

negative ganglion preparation (ganglion 3) yielded no band.

The extendedproducts of the positive ganglia showed slight variation in the size of this band, compatible either with a

single base differenceinthe 5'startsiteor a

single

basedeletion

in the sequenceofonelatent HSV- 1 genomerelative tothe

otherinthis region. Afainter band isseen - _{35bases above}

Z (n

A

G C

T

U

°SI

..

I'

fl_'- '..bm

_I-7 ":.# 6b.

_u~ _. -4

_ .. *Pip

UL

2

2

:D

this band in ganglion number 1. As this is not seen in either the otherpositiveganglion or the negative ganglion, it is not clear whetherit represents the 5' end of a minor latency-associated RNAspeciesinitiated further upstream from the more abun-dant LATs.Becauseofasinglebasedifferencein thesequences

near the 5' endof LAT between strain KOS (8) and strain 17 (22), these data indicate that the 5' end of the major HSV-l LATsis2,304 bases from the junctions ofthe long repeats and the UL segment in virus strain 17.

LAT kinetics. In acute infection, HSV gene expression

occursin a coordinated, cascade fashion over a period of 12-18 h. By 1-2 h after infection, 5 viral "immediate-early" (alpha) genes are transcribed. These regulate and promote the synthesis of over a dozen viral "early" (beta) gene products necessary for DNA replication. By 6-10 h after infection, nu-merous viral "late" (gamma) gene products are synthesized, which participate invirusassembly,maturation, and release.

Inthe followingexperimentsourgoal was todeterminethe temporal order of HSV-l LAT gene expression by using drug blocktechniques. To do this, RNAs were extracted at various time points after HSV- 1 infection from cells treated or un-treated with cycloheximide or acyclovir. Cycloheximide is known to block transcription of all HSV-l RNAs that are

dependentupon earliersynthesisof viralproteins, namelythe viral early and late genes. Acyclovir inhibits thetranscription

of HSV- 1 late genes, whoseexpressionisdependentupon

ear-GANGLIA

1 2

3

Figure 3. Primer extensiontodetermine the5'endof the LATs. TheRNAusedwasextractedfromVero

cells acutely infected withalowpassageHSV-I clini-calisolate (lane marked CLIN)orstrain KOS (lane markedKOS), from uninfected Vero cells (lane

markedUNINF),andfrom humantrigeminal

gan-glia showntocontain both species ofLATby North-ernhybridization (lanesmarked1 and 2),andfroma

humantrigeminalganglion shownnot tocontain

de-tectable levelsof eitherLATbyNorthern hybridiza-tion(lanemarked3).Anunrelated humanDNA

se-quenceis shownas asizemarker.The 5' end of the LATswasdetected1 7basesupstreamof the end of theprimer(whichspanned bases2,400-2,419).

117-

*1

am ob..

a_,M

tm t-W,

4%-*,

am 89

t:=

-%, *AM gmftm "

ilk,.

him-lier viral DNA synthesis. The 1,256-bp Sal I-Hinc II subfrag-ment was gelpurifiedand used as a DNA probe in Northern

hybridization analyses of the RNAs. This probe detects the 2.6-kb immediate-earlyICPO transcript and both the 1.85-kb

and 1.35-kbLATs(15).

Inthefirstdrug block experiment Vero cells were treated withcycloheximide(50

_,g/ml)

starting at 3 h before infection

with5 pfu/cell of a low passageHSV-1 clinical isolate. Total RNA was extracted 18 h afterinfection and analyzed by

Northern hybridization (Fig. 4). Cycloheximide treatment substantially reducedLATlevels as compared to that seen in RNAfrom untreated infectedcells. RNA species migrating at

- 4.2 kb proved upon subsequent hybridization with a

single-strandedoligonucleotide probeto behomologousto theICPO

X

0

-)

z

+

ZUL

z

ML

L

v

Z Z

geneandtobe transcribed fromthesame strand as theICPO

message (data not shown), thus suggesting it to be an ICPO splicing precursor. In cycloheximide-treated cells, transcrip-tion of boththe2.6-kbICPOmessage andofthe 4.2-kb

ICPO-related message were increased.

Inthesecond series ofstudies thetimecourse of, and the

effect of acycloviron LAT gene expression was investigated (Fig. 5). Hybridizations ofRNA extractedfromcells infected

with a low multiplicity

(0.1

pfu/cell) of

HSV-l

strain KOS (lanesmarked LO) showedno LATexpressionat 6 h, although

the 2.6-kb ICPORNA wasalready evident. By24 hafter infec-tion, the 1.85-kbLATwasdetectable. UsingRNAfromcells

infected withahigher virus multiplicity(2pfu/cell,HIlanesin Fig. 5),

ICPO

RNAcould beseen asearlyas 2 hafter infection, whileno LAT was detectedeither2 h or 6 hafter infection.By 24h,the 1.85-kb LAT wasreadilyseen. In cellsinfected with2

pfu/cellandtreated with30

_,M

acyclovir

toinhibit viralDNA

synthesis, ICPORNA was seen at 6h, butwasnolonger appar-ent at 24h.LAT RNAswere neverevident inextractsfromthe

acyclovir-treated

HSV-l

infected cells.

L -~

.z

qqjr

Z

_co

_cN

Z

_O

CONj

_O

-s z

_{_-_}

t~~~

(0

CN

N (D CN

>

>

u C

I

I:

I

<

<

-

2.6 Kb

-1.85 Kb

1.4-Figure4.LATexpressioninthepresenceofcycloheximide.RNA

ex-tractedfromuninfectedVerocells(lane UNINF)orVerocells in-fected withanHSV-I clinical isolateinthepresence(laneINF+ Cy-clohex)orabsence(lane INF)ofcycloheximidewashybridizedwith

agelpure I,256-bpSalI-HincIIDNAprobe.

Figure5. Kineticanalysisof LATexpressionand effect ofacyclovir onLATexpression.RNA from uninfectedVero cells(laneUNINF)

orfromVero cells infectedatalowmultiplicity (0.1 pfu/cell)

ex-tractedat6(laneLO6h)or24(laneLO24h)hafterinfection;Vero

cells infectedatahigher multiplicity (2 pfu/cell)extracted at2_(lane

HI2h),6(laneHI6h),or24(laneHI24h)hafterinfection;or

acy-clovir-treatedVerocells infectedatamultiplicityof 2pfu/cell

ex-tractedat6(laneACV6h)or24(laneACV24h)hwas_hybridized

withagelpure1,256-bpSalI-HincIIDNAprobe.

9.5-7.5

4.4

2.4

-

4.2

-

2.6

The data in Figs.4and5reveal that HSV-l 1.85-kb LAT is not expressed early in infection, and that its synthesis, or at leastits accumulation to detectable levels, requires both pro-tein and viralDNAsynthesis. Thus, HSV-l LAT is expressed in a manner similar to that of the many HSV-l "late" (gamma) genes.

Discussion

Inthisreport we summarize our findings with respect to the structure, localization, splicing and kinetics of the HSV-1 LATs in human ganglia and in cell culture. By Northern hy-bridization with oligonucleotide probes, two LATs, 1.85 kb and 1.35 kb in length, are recognized (Table I, Fig. 2). Primer

extension analysisfixed the 5' ends of these LATs from the

diploidlatency gene at 2,304 bases from the junctions of the repeat andunique long segments (or at genome bases 6,911 and 119,461 relative to the strain 17 sequence; Fig. 3), i.e., within two bases of the location assigned in rabbit and murine

studies (8, 9). Both LATs possess identical 5' ends but the smaller transcript differsfromthelargeroneby excision of a single intron that is 400-800 bases in length.

Comparison of our data on the structure of the 1.85-kb LATinhumantrigeminal gangliashows no differences from that foundinacutelyinfected Vero cells. Northern hybridiza-tions of each yielded equivalent resultswith oligonucleotide probes, andprimer extension analysisshowed that these RNAs also have the same 5' end. Therefore, we believe that tissue culture is a valid system in which to study the nature and

expressionof these genesduringacuteinfection.Duringacute infection,LATexpression requiresbothprotein synthesis and viralDNAsynthesisto accumulate to detectable levels (Figs. 4 and 5). This isin fullconcordance with Spivack andFraser,

who also found LATtorequire protein and DNA synthesis

(25), but differs fromthe report ofWagner et al. (7), which concludedthat LATis transcribed as an immediate-early gene.

DuringHSVlatency,there isnoevidence for viral transcrip-tion(otherthanthatleadingtotheproductionofLAT) or for

viralDNAsynthesis. Thissuggestseitherthat LAT transcrip-tion occurs only during theacuteganglionic infection when

viral protein and DNA synthesis occur and that the LATs

persist

duetohigh stability,orthat LATexpressionin

ganglia

isnotdependentupontheexpression of other viralgenes or uponviralDNA

synthesis,

asitappearstobein cultured

non-neuronal cells. Given the maintenance of HSV-1 latency in humansformany decades the former possibility is unlikely. Thus, weconclude thatthe LATs must be synthesized in

la-tently infectedneurons.Spivackand Frasersuggestedthatthe

HSV-l

latencygene beconsidereda "lambdagene"to

distin-guish

itfrom the classical

alpha,

beta,and gamma viral genes

defined by studies of

acutely

infectedcells(25). LAT

expres-sion inneurons would haveto beregulated by, as-yet

unde-fined,

host,orviral factors whichevolve inthecontextofthe

unique milieu of these terminally differentiated, nondividing cells.

The

HSV-l

LATshavebeen shown toplayarolein

pro-moting

reactivation by cocultivation intissue culture of

gan-glia from latently infected mice. Viruses with mutations up-stream of theLAT gene

produce

noLAT butarecapable of

establishing

latency (1

1-13).

Aviralmutant

lacking

sequences

in the longrepeatsand

producing

no LATwas able to

reacti-vate upon explant cocultivation of latently infected ganglia

fromlatent infection, but at a lower rate than virus without thesemutations. Nine oftenganglia from mice infected with

nonmutatedvirushadreactivatedby 7 d,whilethe first reacti-vationof mutantvirusoccurredafter 10 d, and by 31 d virus had reactivatedfromonly 14 of 18 ganglia. This mutant virus was also less lethal to mice (21 vs. 62% death rate) (13). In separatestudies,viruscontainingadifferent deletion mutation preventingLATexpression was also able toestablishlatency.

Only 21 of43 cocultivated ganglia yielded this virus, while virusrecovery rates from ganglia infected with the wild-type andmarker rescued viruswere 17of18 and 17 of20, respec-tively. Thismutantvirus exhibitednormalgrowth characteris-tics in culture, andexhibitedno alteration in virulence (12). Although these data failed to elucidate the mechanism of LAT

action, theydid establish a role forthe LATs in modulating viralreactivation and possibly virulence.

Thestructure andexpressionof the HSV-I LATsin

exper-imentally infected miceandrabbits,and in naturallyinfected

humans seem virtually identical. These data further validate thefindings in miceandrabbitsasbearingonthe nature and

pathogenesis ofhuman HSVdisease.

Acknowledgments

Sunil Nagpal, Alison Freifeld, and William Reinholdprovidedhelpful suggestions. We thank Barbara Spalholz forherprimerextension

pro-tocol, Duncan McGeoch for unpublished

HSV-l

sequence

informa-tion,HarryMalech for critical review ofthis paper, and Pat Stallsmith

formanuscript preparation.Also,specialthankstoJohn Smialek and Mario Golle of the Baltimore Medical Examiners office for assistance inobtaining thetissues. We areparticularly indebtedto Dr. David

Katzof the NationalInstituteof Neurological Disorders andStrokefor sharing his expertise intheremovaland handling ofhumantrigeminal ganglia.

References

1.Corey, L., andP.G.Spear. 1986. Infections with herpessimplex

viruses.N.Engl. J. Med. 314:686-691.

2.Mellerick,D.M.,andN. W.Fraser. 1987.Physicalstateof the latentherpes simplex virusgenomeinamousemodelsystem:evidence suggestinganepisomalstate. Virology. 15:265-275.

3. Stevens,J.G., E.K. Wagner, G. B. Devi-Rao,M. Cook,and

L. T. Feldman. 1987. RNA complementary to a herpesvirus alpha

gene mRNAisprominentinlatentlyinfectedneurons.Science(Wash.

DC). 235:1056-1059.

4.Deatly,A.M., G.Spivack,E.Lavi,andW.Fraser. 1987. RNA

from an immediate early regionof the type 1 herpessimplex virus

genomeispresentin thetrigeminal gangliaoflatentlyinfected mice.

Proc.Natil.Acad. Sci. USA. 84:3204-3208.

5. Spivack, J.G., and N. W. Fraser. 1987. Detection ofherpes simplex virus type 1 transcripts duringlatent infection in mice. J.

Virol. 61:3841-3847.

6. Spivack, J. G.,and N. W. Fraser. 1988. Expressionofherpes simplexvirustype 1 latency-associated transcriptsin thetrigeminal gangliaof miceduringacuteinfection andreactivation oflatent infec-tion.J. Virol. 62:1479-1485.

7.Wagner,E.K.,W. M.Flanagan, G. Devi-Rao,Y.F.Zhang,J. M.

Hill, K. P. Anderson,and J. G. Stevens. 1988. The herpessimplex

viruslatency-associated transcript is spliced during thelatentphase of infection. J. Virol. 62:4577-4585.

8.Wagner, E. K., G.Devi-Rao,L. T.Feldman,A.T.Dobson,Y.

character-ization of the herpes simplex virus latency-associated transcript in

neurons. J. Virol. 62:1194-1202.

9. Wechsler,S. L.,A.B. Nesburn,R. Watson,S. M.Slanina,and H.

Ghiasi. 1988. Fine mapping ofthe latency-related geneofherpes

sim-plex virustype 1: alternate splicingproducesdistinctlatency-related RNAscontainingopenreading frames.J. Virol. 62:4051-4058.

10.Perry, L.F., F. J.Rixon,R. D. Everett, M.G.Frame, and D. J. McGeoch. 1987.Characterization ofthe IEI10geneofherpessimplex virustype 1.J. Gen. Virol. 67:2365-2380.

11.Javier,R.T., J.G.Stevens,V.B.Dissette,and E. K. Wagner.

1988. The herpes simplex virus transcript abundant in latently infected

neuronsis dispensable for establishment of the latentstate. Virology. 166:254-257.

12. Leib,D. A., C. L.Bogard, M.Kosz-Vnenchak, K. A.Hicks,

D. M.Coen,D.M.Knipe, andP. A.Schaffer. 1989.Adeletionmutant

ofthelatency-associated transcript ofherpessimplex virustype 1

reac-tivates from the latent state with reduced frequency. J. Virol. 63:2893-2900.

13. Steiner, I.,J. G. Spivack, R. P. Lirette, S. M.Brown, A. R.

Maclean, J. Subak-Sharpe, and N. W. Fraser. 1989.Herpessimplex viruslatency-associated transcriptsareevidentlynotessential forlatent

infection. EMBO (Eur. Mol. Biol. Organ.) J. 8:505-511.

14.Croen,K.D., J. M.Ostrove,L. J.Dragovic,J. E.Smialek,and

S.E.Straus. 1987. Latent herpes simplex virus inhumantrigeminal ganglia.N.Engl. J. Med. 314:1427-1432.

15. Krause, P. R.,K. D.Croen, S. E.Straus, andJ. M.Ostrove.

1988. Detectionand preliminary characterization ofherpessimplex virustype 1 transcripts inlatentlyinfectedhumantrigeminal ganglia.

J. Virol. 62:4819-4823.

16. Stevens, J. G.,L.Haarr,D. D.Porter, M. L.Cook, andE. K. Wagner. 1988. Prominence ofthe herpessimplexvirus

latency-asso-ciatedtranscript in trigeminal ganglia from seropositive humans. J. Infect.Dis. 158:117-123.

17. Steiner, I., J.G. Spivack, D. R. O'Boyle, E. Lavi, and N. W. Fraser. 1988. Latentherpessimplex virus type 1 transcription in

humantrigeminal ganglia. J. Virol. 62:3493-3496.

18. Gordon, Y.J., B. Johnson, E. Romanowski,and T.

Araullo-Cruz. 1988.RNAcomplementarytoherpessimplextype 1ICP0gene

demonstrated in neurons of human trigeminal ganglia. J. Virol. 62:1832-1835.

19. Wechsler, S.L., A. B.Nesburn, R.Watson, S. Slanina,and H.

Ghiasi. 1988. Fine mapping of the major latency-related RNA of

herpessimplex virustype 1in humans.J.Gen. Virol. 69:3101-3106.

20. Chirgwin,J. M., A. E. Przbyla, R. J.MacDonald, andW. J. Rutter. 1979. Isolation of biologically active ribonucleic acid from

sourcesenriched inribonuclease.Biochemistry. 18:5294-5299. 21.Melton, D. A., P. A.Kreig,M. R.Rebagliati,T. Maniatis,K.

Zinn, andM. R.Green. 1984. Efficient in vitro synthesis of biologi-cally activeRNAandRNAhybridization probes from plasmids

con-tainingabacteriophage SP6promoter.Nucleic AcidsRes.

12:7035-7056.

22.Ostrove,J. M., W.Reinhold, C.Fan,S. Zorn,J. Hay,and S.E.

Straus. 1985. Transcription mapping of the varicella-zoster virus ge-nome.J. Virol. 56:600-606.

23. McGeoch, D. J., M. A. Dalrymple, A. J.Davison,A. Dolan, M.C. Frame,D.McNab,L.J.Perry,J. E.Scott,and P.Taylor. 1988. ThecompleteDNA sequenceof the long unique region in thegenome

of herpessimplex virustype 1.J. Gen. Virol. 69:1531-1574.

24.Spalholz,B. A., P. F.Lambert, C.L.Yee, andP. M.Howley. 1987.Bovine papilloma virus transcriptional coagulation: localization of the E2-responsive elements of the long control region. J. Virol. 61:2128-2137.

25.Spivack, J. G., andN. W. Fraser. 1988. Expression of herpes simplex virustype 1 (HSV- 1)latency-associated transcriptsand