Intragenic complementation of herpes simplex virus ICP8 DNA-binding protein mutants.

(1)

0022-538X/93/020876-10$02.00/0

Intragenic Complementation

of

Herpes Simplex

Virus

ICP8

DNA-Binding

Protein

Mutants

MIN GAOtANDDAVID M. KNIPE*

Department ofMicrobiologyandMolecularGenetics, Harvard MedicalSchool, 200

Longwood

Avenue, Boston,

Massachusetts 02115

Received 13July1992/Accepted30 October1992

ThemajorDNA-binding protein, orinfected-cell protein 8 (ICP8),ofherpes simplexvirus isrequiredfor viralDNAsynthesisand normalregulationofviralgeneexpression. Previous geneticanalysishasindicatedthat the carboxyl-terminal 28 residues aretheonlyportion of ICP8capableofactingindependentlyas a nuclear

localization signal.In this study,weconstructeda mutantvirus (nllSV) inwhich thecarboxyl-terminal28 residuesofICP8werereplaced bythesimian virus 40large-T-antigennuclearlocalization signal.ThenllSV ICP8 localized intothe nucleus and boundtosingle-strandedDNAinvitroastightlyaswild-typeICP8 didbut

wasdefectiveforviral DNAsynthesisand viralgrowthinVero cells. TwomutantICP8proteins(TL4andTLS) containing amino-terminal alterations could complement the nlSV mutant but not ICP8 gene deletion mutants.Celllinesexpressing TILAandTL5ICP8wereisolated,and inthesecells, complementationofnllSV wasobservedatthelevels of both viral DNAreplicationand viralgrowth.Therefore, complementationbetween nllSVICP8 andTL4orTLM ICP8 reconstitutedwild-typeICP8 functions. Our results demonstrate that(i)the carboxyl-terminal28 residues of ICP8arerequiredforafunction(s)involved in viral DNAreplication, (ii)this

function canbe supplied intrans byanother mutant ICP8, and (iii) ICP8 has multiple domains possessing differentfunctions, and atleastsomeofthesefunctions cancomplementintrans.

The major DNA-binding protein of herpes simplexvirus (HSV), infected-cellpolypeptide 8,orICP8,isoneofseven

virus-encoded proteins required forreplication ofthe HSV

genome(3, 4, 20, 24, 39, 40). ICP8, a 130-kDapolypeptide,

is expressed as a

0i

ordelayed-early-gene product (20, 24).

From studies of thephenotypes ofvirus strains containing temperature-sensitive (4, 39), nonsense, deletion, or site-specific mutations (5, 8-10, 12, 23) in the ICP8gene, ICP8 has beenshowntohavetheabilityto(i)bindtoDNA invitro and invivo(1, 16,18, 25, 31), (ii)localizetothe cell nucleus (10), (iii) down-regulate the expression ofviral genes from parental genomes (12), (iv) stimulate late-gene expression fromprogenytemplates (9), and(v)promoteorganizationof nuclearstructures involved in viral and cellular DNA repli-cationproteins (5).

ICP8 bindspreferentiallytosingle-strandedDNA(ssDNA) invitro (1, 16, 18, 25, 32, 33), and the interaction ofICP8 with ssDNA is cooperative (18, 22, 31). On the basis of studies ofICP8 moleculesexpressedbymutantviruses,of in vitrotranscription-translation products ofICP8, and of

par-tialproteasedigestionofpurified ICP8, the ICP8sequences

required for ssDNA-binding activity have been mapped between residues 564 and 849(8, 19, 38). Genetic evidence indicates that ICP8 specifies other nuclear functions in additiontoDNAbinding (8, 9).

Theintranuclearlocalization of ICP8tospecificstructures

appears to be required for ICP8 to exert its nuclear func-tions. In theabsence ofviral DNAreplication, ICP8 local-izes to nuclearframework-associated structures called

pre-replicative sites (5, 26, 27). These structures have been hypothesized to be the sites of viral and cellular protein

*Correspondingauthor.

tPresent address:DepartmentofVirology, Bristol-MyersSquibb Co.,Princeton, NJ08543.

complexes poised to initiate viral DNA synthesis. As viral DNA replication occurs, progeny viral DNA-ICP8

com-plexes and additional viral and cellular proteins migrate to large globular replication compartments(5, 26). HSV

poly-merase colocalizes with ICP8 in both replicative sites and replication compartments. However, localization of HSV polymerase to these structures is dependent on functional

ICP8 molecules. Therefore, ICP8 is required for the

assem-bly of prereplicativesites (5).

Genetic analysis of ICP8genemutantsandICP8-pyruvate kinase fusion proteins demonstrated that the carboxyl-ter-minal 28 residues of ICP8 constitutetheonly portion ofICP8 thatcanfunction aloneas anuclearlocalization signal(NLS) (10). The carboxyl-terminal 28 residues of ICP8 are not needed forssDNA-binding activity(10).

Tostudy the functional domains of ICP8 further,wehave

constructed amutant virus(nllSV)inwhich the carboxyl-terminal 28 residues of ICP8 were replaced by the nuclear

localization signal of simian virus 40 (SV40) T antigen. Despite the fact that nllSV was unable to promote viral DNAsynthesis, nllSV ICP8 still retainssomephysical and

functional properties of wild-type (wt) ICP8. This paper

reports that the carboxyl-terminal 28 residues of ICP8 are

required forafunction(s) involved in viral DNA replication

and that thefunction(s) missinginnllSVcanbesupplied in

trans by anothermutant ICP8 molecule. MATERUILSANDMETHODS

Cellsand viruses. Vero cellswere grownandmaintainedas

described previously (14). The growth medium forthe

neo-mycin-resistant S-2 cell line included 200 ,ug of the anti-bioticG418perml duringthefirstpassageof the cells after

thawing or 500 ,ug of G418 per ml of medium every five passages.

876

on November 9, 2019 by guest

http://jvi.asm.org/

(2)

COMPLEMENTATION OF HSV ICP8 MUTANTS 877

L S

-l II -l 4II

Bamn s

0.41 0A0 039 038

wt ICP8

n1l

nllSV

n9

d301

p8 B-S

TL4

TLS

TLI 6

1169

832

264 932

204

207

732

pSVdlOl1

17 564

FIG. 1. Locations ofICP8genemutations used in this study. The location of theICP8 coding regiononthe HSVgenomeis shownat thetop.Restriction sites shown areBamHI andSalI. The hatched

barinmutantvirusnllSVrepresentstheSV40 T-antigen NLS (see Materials andMethods). L, longcomponent;S, shortcomponentof the viralgenome.

TL4 and TL5 cell lines were isolated by procedures

described previously (6, 8). Vero cells were transformed

with the plasmid TL4 or TL5 and pSVneo (Fig. 1) in the

presenceof theantibiotic G418 (36).Drug-resistant colonies

wereisolated,grownintocultures,andscreened byindirect

immunofluorescence (27) with anti-ICP8 monoclonal anti-body lOE-3 (30) for the ability to express mutant forms of ICP8 upon infection by nlO mutant virus. The lOE-3 anti-body recognizes TL4ICP8 and TL5 ICP8 butnotnlOICP8 (10).

HSVtype1wtstrainKOS1.1waspropagatedandassayed

asdescribedpreviously (16, 17).Mutant virusesweregrown

inICP8-expressingS-2 cells (8).

Plasmids. Plasmids p8B-S, pICP8, pSV8, pSV8.2, and pSVdlOlaswellasthenucleotide-numberingsystemfor the ICP8 gene were described previously (7-9). The plasmid p8B-S contains the ICP8gene, includingits ownpromoter.

Theplasmid pICP8contains the ICP8genecodingsequences

without its promoter. The plasmid pn9 was generated by

linearization of theplasmid pICP8 (whichwas achievedby partial digestion with SmaI) and subsequent insertion of a

14-nucleotide XbaI linker (New England BioLabs, Inc., Beverly, Mass.) containing stop codons in all threereading frames.Thus, pn9encodesthe first 833 amino acid residues ofICP8. Theplasmid pSV8 wasconstructed by inserting a

5.5-kbp SmaI-SacI fragment (map units 0.374 to 0.409) downstream of the SV40 early promoter. The plasmid pSV8.2 was derived from pSV8 by deleting the polylinker region and the SacI-BglII fragment of thepyruvate kinase

gene. The plasmid pSVdlO1 was generated as described previously (8); it lacks codons for residues 17 to 563 but has an insertion of one Arg codon encoded by the BglII linker sequence. The plasmids p8S/3583 and p8E/3583 were de-rived from pICP8 by deleting the NaeI (nucleotide 3583)-SacI(nucleotide 6075) and the SmaI (nucleotide 436)-NaeI (nucleotide 3583) fragments, respectively. The plasmids

TL4, TL5, and TL16 were constructed as described previ-ously (37).Theycontain eight, six, or two codon insertions afterresidues204, 207, or 732, respectively, of ICP8.

The plasmid p8N/12B-1 was constructed by inserting a 12-nucleotide BglIIlinker at an NaeI site (nucleotide 4111). The plasmid pSVnll was generated by linearization of the plasmid p8N/12B-1 and subsequent insertion of an XbaI

linker containing stop codons in all three reading frames. Thus, pSVnll encodes the first 1,169 amino acid residues of ICP8 aswell as 4 additional amino acids (Thr-Ser-Leu-Asp)

encoded by the XbaI linker sequence. The plasmid pSVnllSV was generated by linearization of the plasmid pSVnll withXbaI and subsequent insertion of a 27-nucle-otide linker (5'-CTAGCACCAAAAAAGAAGAGAAAGG

TA-3')encoding the SV40 large-T-antigen NLS. The correct orientation of the insertion was confirmed by DNA sequenc-ing. Thus, pSVnllSV encodes the first 1,169 amino acid residues of ICP8 as well as the additional amino acids

Thr-Ser-Leu-Ala-Pw-L-L

-Ly-Arg-L

s-Yal-Leu-Asp.

The underlined amino acid sequences represent the SV40 NLS (13).

Isolation of nll and nllSV mutantviruses.The strategy for

isolationof nll and nllSVmutantviruses wasasdescribed

previously(8).The HD-2mutantviruscontainsanin-frame

insertion of thelacZgenein the ICP8 coding sequence and forms blue plaques in the presence of

5-bromo-4-chloro-3-indolyl-p-D-galactopyranoside

(X-Gal). This virus served as a recipient in marker transferexperiments tointroduce the

n9, nll, and nllSV mutations in theICP8geneinto theviral

genome.

Indirect immunofluorescence. Indirect

immunofluores-cence was performed as described previously (27) with anti-ICP8 monoclonal antibody lOE-3or793 (1:50dilution;

30) and rhodamine-conjugated goat anti-mouse antibody (1:100 dilution).

Marker rescue and complementation experiments. Marker rescueexperimentswereperformed by cotransfection of S-2 cells with 1.0 ,ug each ofinfectiousmutant virus DNAand linearizedwtICP8 geneplasmid DNA(Fig. 1)aspreviously

described (15). Transfected cultures wereincubated at 37°C

for 2 to3 days, and progeny viruswas assayed inS-2 and Verocells.

Complementation of thegrowth ofn9,

d301,

andnllSV

mutant viruses by prior transfection of the

plasmid

TLA,

TL5,

TL16, or

pSVdlOl

wascarriedoutasdescribed

previ-ously(28).

Analysis

ofviral proteins and viral DNAreplication. Vero

cell monolayer cultures were infected with

KOS1.1,

nll,

nllR, nllSV, ornllSVR and labeled with

[35S]methionine

for 30 min before harvest at9 h

postinfection (hpi).

Sodium

dodecyl

sulfate-polyacrylamide gel

electrophoresis

(SDS-PAGE) of infected-cell lysates was

performed

asdescribed

previously (17). After

electrophoresis,

the

gels

were

fixed,

dried, andexposed toKodak SB5 film.

Analysis ofviral DNA

amplification during

thecourse of infectionwasperformedasdescribed

previously (9,

29).

The

probeusedwasthe

plasmid

pBR3441

(21).

VOL. 67,1993

Li

on November 9, 2019 by guest

http://jvi.asm.org/

[image:2.612.64.301.70.344.2]

(3)

¢ zr,:<X~~EOW.... _

FIG. 2. Subcellular localization of ICP8 encodedbymutantvirusesnll and nllSV.Verocellswereinfected withKOSl.iorICP8mutant

viruses.At 5hpi,the cellswerefixed, permeabilized, and incubatedwith anti-ICP8monoclonalantibody793 andrhodamine-conjugatedgoat anti-mouse immunoglobulin. Immunofluorescencemicrographs areofKOSl.i-infectedcells (A), KOSi.l-infected cells in thepresenceof PAA(B), nll-infected cells(C),andnllSV-infected cells(D).

ssDNA cellulose chromatography. ssDNA cellulose

chro-matography ofinfected-cell extracts was performed as de-scribed elsewhere (8, 14).

RESULTS

Phenotypeof mutant viruses nll andnIlSV.Ourprevious genetic analysis demonstrated that the mutant nlO ICP8, which lacks thecarboxyl-terminal36 residues and localizes to the cytoplasm, does not support viral growth in Vero

cells even though it binds ssDNA in vitro as tightly as wt ICP8 does (8). Further study indicated that the

carboxyl-terminal 28 residues are the onlyportion of ICP8 capable

of acting independently as an NLS (10). Thus, the failure of mutant nlO ICP8 to support viral DNA synthesis may have been solely a consequence of its inability to localize to the cell nucleus. If this were the case, replacement of ICP8 NLS with another known NLS should restore the full functionality of ICP8. Therefore, we constructed two ICP8 mutant viruses. One, nll, lacks the carboxyl-ter-minal 28 residues of ICP8, and the other (nllSV) has the SV40T-antigen NLS (NH2-Pro-Lys-Lys-Lys-Arg-Lys-Val-COOH) replacingthe carboxyl-terminal 28 residues of ICP8

(Fig.

1).

(i) Subcellular locations of nll ICP8 and

nl1SV

ICP8. We first used immunofluorescence microscopy to examine the intracellular localization of the nll and nllSV ICP8

poly-peptideswhentheywereexpressed from plasmids in

trans-fectedVero cells.The

nil

ICP8 polypeptide remained within thecytoplasm, while the nllSV ICP8 polypeptide localized in the nucleus (11). We next determined the intracellular

distribution of mutant ICP8 molecules when they were

expressed from recombinant viruses (Fig. 2). At 5 hpi,

infected Vero cellswerefixed andprocessedfor immunoflu-orescence. Cells infected with wt virus showed that ICP8 localized in the nucleus in characteristicreplication compart-ments in the absence ofphosphonoaceticacid (PAA) (Fig. 2A) andatprereplicativesites in the presenceof PAA(Fig.

2B; 26, 27). As expected, nll ICP8 localized in the

cyto-plasm(Fig. 2C)because of lack of the carboxyl-terminal28 residues. Inmostcells,thenllSV ICP8 localizedthroughout

the cell nucleus (Fig. 2D) and did not appear to localize

specificallyeither inreplication compartmentsor at prerep-licative sites. Therefore, replacement of the ICP8 NLS with the SV40 T-antigen NLS restored the capacity of mutant ICP8tolocalize into the nucleus butnot tothesamesitesas wtICP8.

(ii) Growth properties of nll and nllSV. Plaque assays were performed to determine whether the mutantviruses

wereabletogrowin Vero cells. Both mutant viruses failed togrow in Vero cells but grewtotitersnearthoseofwt on the ICP8-expressing S-2 cellline (Table 1). Thus, although

replacementof the ICP8 NLS with the SV40T-antigen NLS restored the ability ofICP8 tolocalize to the nucleus, the substitution didnot restorecomplete functionality of ICP8. (iii) Marker rescueofmutantviruses nll and nllSV. To verify that the phenotypes of mutant viruses nll and nllSV were due to mutations in the ICP8 gene, we performed marker rescue experiments (Fig. 3). Homologous recombi-nation between nll DNA or nllSV DNAand a fragment containingawt ICP8 DNA sequence(p8S/3583) resulted in viral progeny possessing the ability togrow in Vero cells,

thereby rescuingthegrowth deficiencies of nll and nllSV.

on November 9, 2019 by guest

http://jvi.asm.org/

[image:3.612.117.504.75.351.2]

(4)

[image:4.612.317.560.85.139.2]

COMPLEMENTATION OF HSV ICP8 MUTANTS 879 TABLE 1. Growth ofnllandnllSVa

Titer(PFU/ml)

Virus

S-2cells Vero cells

KOSl.1 1.7 x 109 1.8 x 109

nll 2.3x109 2.0x103

nllSV 2.8x 108 2.0 x 103

a Cultures of ICP8-expressingcells(S-2) and Verocellswereinfected with each virus and incubated at 37°C. Plaqueswerecountedat 3days.

TABLE 2. Growthproperties of nllRandnllSVRa

Titer(10' PFU/ml)

virus

S-2cells Verocells

nllR 3.8 2.8

nllSVR 1.3 2.0

a Cultures ofICP8-expressingcells(S-2)and Vero cellswereinfected with each virus andincubatedat37°C. Plaqueswerecountedat3days.

The high percentage of rescue (1.5 to 8.3%)

rescue was likely due to a recombinational el

the transfectedplasmidsandnotadditionaleve ofrescuedviruses nllR and nllSVRwere

ap

orders of magnitude greater than those of viruses nll and nllSV on Vero cells and roi their titers in S-2 cells (compare Tables ] conclude that the phenotype of mutant Vir nllSVwas aresult of mutations within the IC likely the engineered mutations.

(iv) DNA-binding

properties

of mutant vii

nllSV. To attempt to define the nature of functions of the mutant ICP8 molecules, we

ssDNA-binding properties of mutant nll IC]

ICP8. Vero cells were infected with nll c extract from infected cells was passed ov

cellulosecolumn, andICP8was elutedstepw

containing increasing salt concentrations. P thevarious fractionswere analyzed bySDS-F The saltelution patterns fornll ICP8 (Fig.4

BamHI Sall

0.41 0.40 0.39 0.38

wt ICP8

607 4195

n1l

607 4111

nil SV

607

p8S/3583

p8E/3

583

p8S/3583

pSV8.2

p8E/3583

p8S/3583

pSV8.2

Viru.

=

% wt recc

<O.C 1 .! 3.,

<0.1

1.

FIG. 3. Markerrescueof mutant virusesnllar

bars indicate coding regions of ICP8 genes. Rel

positions of the recombinant plasmids are as fol nucleotides 432 to 3583; p8S/3583, nucleotides 3 mately 5900; andpSV8.2, nucleotides432 to appi

Percentageofwtrecombinants=(virustiterin Ver

inS-2 cells) x 100%.

indicated that

(Fig. 4C)

ICP8 werevery similar to that for wt

ICP8

(Fig.

ventinvolving

4A),

and the

majority

of wt ormutant ICP8waseluted with nts The

tivtergs

0.5 M NaCl

(Fig.

4, lane

12).

Thus, nll andnllSV ICP8 iproximately. bound to ssDNA in vitro as

tightly

aswt ICP8 did.

There-)proximately

5 fore, the inability of the

nllSV

mutant to replicate its DNA

f the

parental

in Vero cellswas

likely

not aresult ofan

inability

to bindto

ughly equal

to ssDNA

L and 2). We Intragenic complementation of ICP8 mutants. Despite its

ups8es

nil

aonsd

failureto promote viral DNA

synthesis,

the mutant

nllSV

ICP8 retained several of the

physical

andfunctional proper-ties ofwtICP8: itlocalizedtothe

nucleus,

boundtossDNA

ruses

nil

and in

vitro,

and was recognized by the 39S anti-ICP8

mono-examined the

clonal

antibody

(not shown),

recognition

by

which is

very P8andnllSV

sensitive

to

conformational

changes

in ICP8. These results

8 nllSVdthn

suggest

that the

mutant

virus

nllSV

may

be altered in

only

er

an1SV

the asingle functional domain of ICP8 which is involved in viral

er

an ssDNA DNA replication. We therefore screened a series of ICP8

lse

with

buffer

linker insertion

mutant

plasmids

to

determine whether they

olypep(Fidegs

4n

could

complement the nllSV defect by providing certain

RAGE (Fig.

_[B)

_and

_niiSV

4)S

functions_(i) _{Intragenic complementation}of ICP8 intrans.

_between

_mutant _virus

nllSV and mutant

plasmid

TL4 or TL5. We

performed

complementation

experiments

todeterminewhether

expres-Sadc

sion of mutant ICP8

polypeptides

from linker insertion mutant

plasmids

could

complement

the

growth

of nllSV mutantvirusin Vero cells. All of the linker insertionmutant

formsofICP8used inthis

study

wereunableto

complement

the ICP8 gene deletion mutantvirus d301 and

appeared

to

ses localize into the nucleus in transfected cells (37). The

plas-mid

pSVdlOl

containing

a

large

deletion in the ICP8 gene

wasused as a

negative control,

and

TL16

wasusedalinker

insertionmutantcontrol

(Table 3).

Neitherof thesemutants

complemented

the

growth

of either d301 or nllSV. In

DNA fragments contrast, cotransfection of the wt ICP8 plasmid (p8B-S)

increased the

yields

ofmutantvirusd301

4,400-

to

10,000-fold andof

nllSV

1,500-

to

3,000-fold.

Two linkerinsertion

mutants, TL4and

TL5,

which could

complement

growth

of themutantvirus nllSV in Vero

cells,

wereidentified

(Table

3).

Neitherof thesemutant

plasmids

wasableto

complement

imbinants

d301,

butboth

plasmids

complemented

nllSV

significantly.

)12

For

plasmid

TIA,

complementation

ranged

from 7to23%of 5 the level ofwt ICP8

plasmid

(Table

3).

The

plasmids

TL4

.4

and TL5 contain 8- and 6-amino-acid insertions after resi-dues 204 and 207of

ICP8,

respectively (Fig.

1).

Because the

0029

insertions in

TL4

ICP8 and TL5 ICP8 are only three residues 9 apart, the same functional

domain(s)

of ICP8 may be

af-.3 fected.

By

looking

for the

ability

ofmutants to form

plaques

on

id

nllSV.

Black

Vero cells (Table

4),

we also examined infected-celllysates

lative

nucleotide'

[lows: p8E/3583, from the

complementation

experiments

shownin Table 3 for

3583

to approxi- wt recombinants. The plasmid

pSVdlOl

was used as a roximately

5900.

negative

control because deletions in the ICP8 genes of ocells/virustiter mutants

pSVdlOl

and d301are

overlapping;

thus,no

recom-binationbetween these twomutants should occur.

Recom-p8E/3583

pSV8.2

Cotransfected Viral DNA fragments

nll nil nil

nl1SV nil SV nl1SV VOL. 67, 1993

on November 9, 2019 by guest

http://jvi.asm.org/

[image:4.612.60.302.86.148.2] [image:4.612.63.298.412.659.2]

(5)

880 GAO AND KNIPE

*owN -ICP8

7 8 9 10 11 12 13

B

# 10,** # '60 -ICP8

_f.,

:-; ..;-i

1 2 3 4 5 6 7 8 9 10 11 12 13

C

,,

_i

s. .

.w:

r ": 'W

R

-1 2 3 4 5 6 7 8 9 10 1112 13

FIG. 4. ICP8 binding to ssDNA-cellulose. Vero cells were in-fected with KOS1.1(A), nll (B),ornllSV(C) in thepresenceof PAA and labeled with [35S]methionine from 4 to 6 hpi. Various protein fractions resolvedon ssDNAcellulose columnswere sub-jectedtoSDS-PAGE.Lanes: 1,pellet from high-saltDNase

extrac-tion; 2, pellet after dialysis; 3, extractput on ssDNA column; 4 through 10,flowthrough and wash; 11, 0.3MNaCleluate; 12,0.5 M NaCleluate; 13,1.0 MNaCl eluate. PositionsofICP8areindicated

attheright.

bination between the transfected wt ICP8 plasmid and in-fectedmutantviruses didoccur atlowfrequency.However, wtvirusfrequencies were almost thesameford301 (where

[image:5.612.86.273.72.383.2]

no complementation occurred) and for TL4 or TL5 (where i I00 -ICP8

TABLE 4. Recombinationbetween ICP8 mutantsa

Plasmid Virus Titer (PFU/ml)

transfected superinfected Expt1 Expt2

pSVdlOl d301 ND ND

pSVdlOl nllSV 10 70

TL16 d301 ND ND

TL16 nllSV ND 50

p8B-S (wt) d301 590 300

p8B-S (wt) nllSV 1,300 800

TIA d301 40 ND

TM4 nllSV 10 60

TL5 d301 10 ND

TL5 nllSV 40 80

[image:5.612.316.555.86.211.2]

a Infected-cell lysates from theexperimentswhoseresults areshown in

Table 3 wereassayedfor wtrecombinantsbyassayingfortheabilitytoform plaqueson Verocells. ND =titerof<10PFU/ml.

complementation occurred). Thus, the enhancement of growth ofmutantvirusnllSVbyTMAorTL5was notdueto recombination. We conclude thatthe

function(s) required

for themutantvirusnllSVDNAreplicationcould be supplied

in transbypartiallyfunctional TL4orTL5ICP8.

(ii) Phenotype of TL4 and TL5 mutants. To study the mechanism ofintrageniccomplementation byICP8 mutants, we examined the properties of the TML and TL5 ICP8 molecules. Our results indicated that both TLM and TL5

mutantgeneproductsexhibitedatrans-dominantphenotype

which prevented the introduction of these mutations into virus. We therefore isolatedtwocelllines,namedTL4.6 and TL5.2, which expressedTL4 ICP8 and TL5 ICP8, respec-tively, upon HSV infection (11).

Mutant ICP4 moleculescontainingalterations in the trans-activation domain can form a heterodimer with another mutant ICP4 moleculecontainingan alteration in the

DNA-bindingdomain andcancomplementeach other intrans(34, 35). To test whether this interpretation could apply to the mutant ICP8also,weexaminedthe DNA-binding abilityof TL4 ICP8 expressed from the cell line. We concluded that TL4ICP8,like nllSVICP8,binds ssDNA in vitroastightly

as wt ICP8 does (11). These results suggested that the mechanism of intragenic complementation between nllSV and

TMA

ICP8 mutants is different from that of the ICP4 mutants, because both mutant ICP8 molecules can bind DNA.

TABLE 3. Intragenic complementation between ICP8mutantsa

Plasmid Virus Virus yield

(PFU/ml)p

Complementationindexc

transfected superinfected

Expt

1 Expt2 Expt1 Expt2

pSVdlOl d301 9.0 x 101 3.4 x 102 1.0 1.0

pSVdlOl nllSV 3.0x 102 1.1 x 103 1.0 1.0

TL16 d301 1.3 x 102 6.0 x 102 1.4 2.0

TL16 nllSV 2.3 x 102 1.1 X 103 0.8 1.0

p8B-S (wt) d301 9.2 x 105 1.5 x 106 10,000 4,412

p8B-S (wt) nllSV 9.0 x 105 1.6 x 106 3,000 1,455

TMA d301 1.1 x 102 2.8 x 103 1.2 8

TL4 nllSV 6.4 x 104 3.6 x 105 210 327

TL5 d301 6.2 x 102 3.3 x 103 6.9 10

TL5 nllSV 1.2 x 104 1.8 x 105 40 164

aVerocellsweretransfected withthe plasmids indicated. At 20 h posttransfection, cells were superinfected with 3 PFU of either d301 ornllSVper cell and

incubated forafurther24h at37'Cbefore being harvested.

bDeterminedby plaqueassay onS-2cells.

c Expressedasvirus yield relativetoyieldafter transfection with

pSVdlOl

DNA.

A

*-,

t0 .1^

...

:.3 is

*''v -."I .. j :.

1 2 3 4 5 6

J. VIROL.

on November 9, 2019 by guest

http://jvi.asm.org/

[image:5.612.58.562.564.687.2]

(6)

Vero

n9

n11SV

KOS

TL4.6

h

FIG. 5. Formation ofreplication compartments in nllSV-infectedTILA.6cells. Vero andTL4.6cells,respectively, were infected with n9 (AandB),nllSV(C andD),orKOSl.1(E and F). At 5 hpi, the cells were fixed, permeabilized, and incubated with monoclonal antibody 10E-3andrhodamine-conjugated goatanti-mouseimmunoglobulin.

(iii) Formation of replication compartments in nllSV-in-fectedTLAandTL5 cells. Because theplasmid TI4resulted inmore-significant complementationof thegrowthof nllSV in Verocells thanTL5did(Table 3),weused TIA.6 cells for most of our further studies. To examine the localization

properties ofTL4 ICP8, we infected Vero and TILA.6 cells

with n9, nllSV,or wtKOS1.1for 5h and thenprocessedthe cells forimmunofluorescencebyusingmonoclonalantibody 10E-3astheprimary antibody (Fig. 5).Thisantibody recog-nizes the carboxyl-terminal 28 residues of ICP8 (10) and therefore recognizes only the ICP8 expressed from TIA.6

cells(Fig. 5B andD)andnotthatexpressed byn9

(Fig.

5A)

ornllSV(Fig. SC).Theantibody,of course, alsorecognizes

wtICP8. KOS1.1-infectedVero cellswereused for controls of theantibody andformation ofreplication compartments.

Replicationcompartmentswereobserved innllSV-infected

(Fig.

SD)

butnotn9-infected(Fig.

SB)

TIA.6 cells.Thus,we conclude that TLA ICP8 could localize into replication

compartmentsin the presence ofnllSV ICP8.

(iv) DNA replication in nllSV-infected TL4.6 cells. To determine whether the mutant virus nllSV replicates its DNAin TL4.6cells,we examined viral DNAaccumulation

in these cells. Total viral DNAwas isolated from

mock-,

wt-virus-,

d301-,

or nllSV-infected Vero cells,

wt-ICP8-expressing S-2 cell line, and TIA.6 cells at 1 or 16 hpi.

wt-virus-infected Vero cells in the presence or absence of PAA were used as positive and negative controls of viral DNAsynthesis, respectively (Fig. 6). As

expected,

mutant

d301-infected TL4.6 cells, like d301-infected Vero

cells,

showedno

amplification

of viral DNAduringthecourse of

infection, indicating that the

TL4

ICP8was unable to pro-mote synthesis of mutant viral DNA. In contrast,

n11SV-VOL. 67,1993

on November 9, 2019 by guest

http://jvi.asm.org/

[image:6.612.92.523.71.508.2]

(7)

S-2

oa

o)

0

Xi

Ci

TL4.6

en

or

80

o

csc

v c, c,

Vero

en o

i- 2e at

'r-

DNA(ng)

1000

200 40

8

I6hr

4m 411 -.4

.4 : *m

1000

200

40

[image:7.612.80.537.72.278.2]

E:

8

FIG. 6. Mutant nllSV DNA replicationin TL4.6 cells.Vero,wt-ICP8-expressing S-2,andTLA.6cellswereinfected withKOS1.1,d301,

ornllSV. Total cellular DNAwasprepared immediatelyafter viralabsorption (1 h) or nearthe endoftheinfection cycle (16 h). Equal

amountsofeach DNAweresubjectedtofivefold serialdilutions,and the DNAswereboundtoanitrocellulosefilter,whichwasprobed with 32P-labeled DNA specific for the VP16gene. Anautoradiographof the blot isshown.

infected TL4cells showed substantial amplificationofviral DNA during the infection. To quantitate these results, the amount of radioactive probe hybridized to each slot was

measured,and the relative amountsofreplicated HSVtype 1 DNA in each sample were determined. In thisparticular

experiment,theamountofnllSVviral DNAsynthesizedin TL4cellswasabout42%of the level of DNAsynthesizedin wt virus-infected Vero cells. We conclude that the func-tion(s) whichisrequired fornllSVtoreplicateits DNAcan

besuppliedintransbythemutant TIAICP8.

(v) ViralgeneexpressioninnllSV-infected TILAcells. ICP8 is not only required for viral DNA replication but is also involved in stimulation of late-gene expression (9). We therefore examined viral late-gene expression in n11SV-infected TL4.6 cells. Vero cells, wt-ICP8-expressing S-2 cells, and TL4.6 cells were infected with KOS, d301, or

nllSVandlabeledwith [35S]methionine from 9.5to 10hpi, andlabeled proteinswereanalyzed by SDS-PAGE (Fig. 7).

Vero cellsinfectedwithwtvirus inthepresenceorabsence

of PAAwereusedaspositive andnegative controls for the

expression of-y2genes. Whenexamined in Vero cells, the

nllSV mutation resulted inaphenotype similartothat of the negative control with respect to late-gene expression. De-tectable levelsof y2 viralpolypeptides (such asICP1-2 and

ICP15)werenotexpressed. Similar resultswereobservedin

d301-infected Vero cells. These results demonstrated that bothnllSV ICP8 and TM4 ICP8areunableto exert wtICP8 functionsforexpression of lategenes.However,thedefects formutantnllSV virallate-proteinsynthesiswerecorrected

by expressionoftheTL4 ICP8 inTL4.6cells. Thepatternof viralpolypeptides synthesizedinnllSV-infectedTL4.6cells

wasindistinguishable from thatobserved innllSV-infected wt-ICP8-expressing S-2 cells or in wt-infected Vero cells

(Fig. 6).

Tofurther define the level atwhich late-geneexpression

wasrestored,weperformed Northern (RNA)blot analysisto measure thesteady-state level ofy2 glycoproteinC mRNA. These resultsindicated that the intragenic complementation

between nllSV ICP8 and TML ICP8 completely restored viral latemRNA accumulation(resultsnotshown),

although

thecomplementationrestored viralDNAreplicationto

only

42% of the normal level(Fig.

6).

KOS d301

+ 11

M :2?2 uz x t

nilSV

<.

".-..: -1-2

-- 5

Mw 6

-8

_ww -pgB

-15

-25

FIG. 7. Protein synthesis innllSV-infected TLA.6 cells. Vero, wt-ICP8-expressingS-2,orTL4.6 cellswereinfected withKOS1.1, d301,ornllSV. At 8.5 hpi, cellswerelabeled for30min with[35S] methionineand then harvested.Equal fractions of each celllysate

weresubjectedtoSDS-PAGE. An autoradiogram of the resultinggel isshown. At the rightof eachpanelarethemigration positions of several HSVtype1 proteins. Lane M, mock-infected cell lysate.

I

hr

-J

0

E

c0

Cl

on November 9, 2019 by guest

http://jvi.asm.org/

[image:7.612.348.529.413.655.2]

(8)

[image:8.612.57.299.89.261.2]

TABLE 5. Growth of wt andmutant viruses indifferent celllinesa

Virus and Yield (PFU/cell) at MOI of:

cell type 1.0 10.0

KOS1.1

Vero 137 393

S-2 110 306

TL4.6 39 245

TL5.2 4 NT

d301

Vero 0.0006 0.0079

S-2 275 361

TL4.6 0.02 0.33

TL5.2 0.008 NT

nllSV

Vero 0.0003 0.0071

S-2 225 119

TL4.6 44 50

TL5.2 0.9 NT

aCellswereinfected with different viruses, incubated at 37°C for 24 h, and harvested.Titers of progeny viruses were determined in S-2 cells. NT, not tested.

(vi) Growth of nllSVinTL4.6 and TL5.2 cells. Because the plaques of nllSV in TL4.6 or TL5.2 cells were very

small, we performed single-cycle growth experiments to examine thegrowthof nllSV in TL4.6 and TL5.2 cells. The

yields ofwtvirus in TL4.6 andTL5.2cells were 3- to 30-fold lower than those in Vero cells or in the wt-ICP8-expressing

S-2 cellsat amultiplicityofinfection(MOI)of1.0(Table 5), suggestingthattheTL4and TL5mutantgeneproductshave sometrans-dominantnegative phenotype. Theinhibition of

growth ofwtvirusinTL4.6 cellswasovercome at anMOI of 10. Thus, the trans-dominant effect appeared to be due to

competitionbetweenwtICP8 and

TM4

ICP8. Theyields of nllSV in TL4.6 cellswere 7to32%of those ofwtvirus in Verocellsat anMOI of 1 and10,respectively.Becauseonly mutantvirus nllSVbutnotd301could grow inTL4.6 and TL5.2cells,weconcludethat thecomplementationbetween nllSV ICP8 and TL4 or TL5 ICP8 could reconstitute wt ICP8 activities.

DISCUSSION

Previous genetic andbiochemical evidence suggests that the functions of the HSV DNA replication protein require

different portions of the molecule: the carboxyl-terminal portion of the molecule is

required

for

ssDNA-binding,

nuclearlocalization,andstimulation oflate-geneexpression,

and the amino-terminal portion of the molecule has an additional nuclear function(s)

(8).

In this study, we have demonstrated that these functionsarediscrete,becausetwo

partially functional ICP8 molecules can complement each other intrans.

Replacement of the ICP8 NLS with the SV40

T-antigen

NLS restoredthe

ability

of ICP8tolocalize intothe

nucleus,

but themutantvirusnllSV exhibitedadefect in viral DNA

replication and late-gene

expression,

suggesting

that the

primary defect of the nllSV ICP8 is a failure to promote viral DNA synthesis or earlier events such as assembly of viral DNAreplication complexes. The nllSV ICP8 defect can be overcome by

supplying

TLM or TL5 ICP8 intrans. The functions defective innllSVICP8 may be carried out

by the carboxyl terminus of ICP8 or other parts of the

molecule; however, extensive conformational

changes

are

not apparent in nllSV ICP8, because it still reacts with anti-ICP8 monoclonal antibody 39S, which is very sensitive to conformational changes in ICP8 (7).

Because monoclonal antibodylOE-3 detectsTM4andTL5

ICP8but notnllSV mutant ICP8, we do not know whether nllSV ICP8 localized alone or formed a complex with

TML

or TL5 ICP8 in thenucleus. However, both

TML

ICP8 and TL5ICP8 localized to replication compartments after super-infection with the mutant virus nllSV but not after super-infection with n9, suggesting that these ICP8molecules can localize to these sites only in the presence of nllSV ICP8. Moreover, because the monoclonal antibody recognized both wt and

TM4

ICP8, we know that both ICP8 molecules localized into replication compartments after TL4.6 cells were infected with wt virus. These results suggest thatnllSV ICP8 may alsolocalize toreplication compartments. Because the alter-ation in

TML

or TL5 ICP8, as innllSV ICP8, is outside of the region required for ssDNA-binding, both mutant ICP8 molecules, as expected, bind to DNA. It is conceivable that

the complementing polypeptides of TIA or TL5 ICP8 and nllSV ICP8 do notinteract with each other directly but that both bind DNA and that each molecule contributes some activities from different domains to achieve wt ICP8 func-tions.

Analternativemechanism for intragenic complementation is theformation of a dimer or multimers of nllSV ICP8 and TL4 or TL5 ICP8. Ithas been reported that ICP8 sediments in aglycerol gradient as a monomer (22), but the cooperative nature of the binding of ICP8 tossDNA suggests that ICP8 mayform polymers along ssDNA (22). In multimeric forms, the monomers may be weakly associated and assemble only uponinteractionwith DNA.Therefore, if multimerization is involved in theintragenic complementation of nllSV ICP8 and TL4 or TL5 ICP8, several models can be considered. First, the fact that nllSV ICP8 did not localize to prerepli-cative sites or replication compartments mayindicate that the SV40 T-antigen NLS did not allow proper targeting of ICP8 to specific sites within the nucleus, where viral DNA replication is initiated. This suggests that the

carboxy!-terminal 28 residues of ICP8 may also be involved in determining intranuclear localization and may interactwitha component of the cell nucleus toinitiate viral DNA replica-tion. Therefore, it ispossiblethat thenllSV ICP8 needs

TML

or TL5ICP8 to form the multimers fortargetingto

replica-tionmachinery.

Second, we (8) and others (19) noted that the amino-terminal portion of ICP8 also contributes to the

DNA-binding activity of the intact protein, although the

DNA-binding region has been identified at the carboxyl-terminal

halfof ICP8 molecule(8, 19,38). This contributioncould be due to the intra- or intermolecular interactions of ICP8. Intermolecularinteractions of this type couldexplain

intra-geniccomplementation. Asecondmodel is thatTL4orTL5 ICP8 and nllSV ICP8 bind DNA but that they can bind DNAcooperatively onlywhentheyinteract with each other. Ifthis is true, the domain which isrequiredforformation of

polymers must be disturbed in at least one mutant ICP8 molecule and the TIAorTL5 ICP8canform aheterodimer with nllSV ICP8 but cannot form polymers. This may

explain whytheefficiencyof

intragenic

complementation

of TL4and nllSV ICP8 reachedonly7to20% of that ofwt:the two mutantICP8 moleculescanform

only

adimer instead of

polymersalongwith DNA. Further

experiments

toexamine

cooperative binding ofnll ICP8 is needed to address this

possibility.

Athird model involves theinteraction of ICP8 with other VOL.67,1993

on November 9, 2019 by guest

http://jvi.asm.org/

(9)

viral and cellular proteins during viral DNA replication. Recently,Capson etal. demonstrated stepwise

ATP-depen-dentassembly of bacteriophage T4 DNAreplicationproteins by usingprotein-DNAcross-linking(2). Inthissystem, gene 32 ssDNA-binding protein but not DNA polymerase is required for assembly of DNA polymerase accessory pro-teins to form a complex on the primer-template molecule. After ATP hydrolysis, DNApolymerase binds to the com-plex to initiate DNA synthesis. It is possible that the alterations in both nllSV ICP8 and TIA or TL5 ICP8 inactivate domains required for interactions with different DNA replication proteins. However, the complementing ICP8 molecules can interactwith these replication proteins after they form multimers along DNA, resulting in the reconstitution of wt ICP8 activity.

In summary,ourresultsdemonstratethat ICP8consists of separate domainswhich possessdifferentfunctions and that somefunctionscancomplementeach other intrans. Further analysisofintrageniccomplementation of ICP8mutantsmay help definetheinteractions betweenICP8and otherviral and cellular proteins which are involved in viral DNA replica-tion.

ACKNOWLEDGMENTS

We thank L.Pereira for anti-ICP8 monoclonal antibody 793 and E. Villarreal for providing the TL mutant plasmids. We thank SumanShirodkarforassistance with isolation of the nllSV mutant virus andSteve Rice forcomments on the manuscript.

This work wassupported by Public HealthService grant CA26345 from theNational Cancer Institute.

REFERENCES

1. Bayliss, G. J., H. S.Marsden, and J. Hay. 1975. Herpes simplex virusprotein: DNA-binding proteins in infected cells and in the virusstructure. Virology68:124-134.

2. Capson, T. D., S. J.Benkovic, and N. G. Nossal. 1991. Protein-DNAcross-linking demonstrates stepwise ATP-dependent as-sembly of T4 DNApolymerase and its accessory proteins on the primer-template. Cell 65:249-258.

3. Challberg,M. D. 1986.A method for identifying the viral genes required for herpesvirus DNA replication. Proc. Natl. Acad. Sci. USA83:9094-9098.

4. Conley, A. J., D. M. Knipe,P. C. Jones, and B. Roizman. 1981. Moleculargenetics ofherpes simplex virus. VII. Characteriza-tion of a temperature-sensitive mutant produced by in vitro mutagenesisanddefective in DNA synthesis and accumulation ofTpolypeptides. J.Virol. 37:191-206.

5. de Bruyn Kops, A., andD. M. Knipe. 1988. Formation of DNA replication structures in herpes virus-infected cells required a viral DNAbindingprotein. Cell 55:857-868.

6. Deluca, N. A., A.McCarthy, and P. A. Schaffer. 1985. Isolation andcharacterization ofdeletion mutants of HSV-1 in the gene encodingtheimmediate-early regulatory protein ICP4. J. Virol. 56:558-570.

7. Gao, M., J.Bouchey,K. Curtin, and D. M. Knipe. 1988. Genetic identification of a portion of the herpes simplex virus ICP8 proteinrequiredforDNA-binding. Virology 163:319-329. 8. Gao, M., and D. M.Knipe. 1989. Genetic evidence for multiple

nuclearfunctionoftheherpes simplex virus ICP8 DNA-binding protein. J. Virol.63:5258-5267.

9. Gao, M., and D. M. Knipe. 1991. Potential role for herpes simplex virus DNA replication protein in stimulation of late geneexpression. J.Virol. 65:2666-2675.

10. Gao, M., and D.M.Knipe. 1992. Distal protein sequences can affect the function of a nuclear localization signal. Mol. Cell. Biol. 12:1330-1339.

11. Gao, M., and D. M. Knipe.Unpublished results.

12. Godowski,P. J.,and D. M.Knipe. 1986. Transcriptional control of herpesvirus gene expression: gene functions required for positiveand negative regulation. Proc. NatI. Acad. Sci. USA

83:256-260.

13. Kalderon,D., B. L. Roberts, W. D. Richardson, andA. E. Smith. 1984. A short amino acid sequence able to specify nuclear location. Cell 39:499-509.

14. Knipe, D. M., M. P. Quinlan, and A. E. Spang. 1982. Charac-terization of two conformational forms of the major DNA-binding protein encoded by herpes simplex virus 1. J. Virol. 44:736-741.

15. Knipe, D. M., W. T. Ruyechan, and B. Roizman. 1979. Molec-ular genetics of herpes simplex virus. III. Fine mapping of a genetic locus determining resistance to phosphonoacetate by two methods of marker transfer. J. Virol. 29:698-704.

16. Knipe, D. M., and A. E. Spang. 1982. Definition of a series of stages in the association of two herpesvirus proteins with the cell nucleus. J. Virol. 43:314-324.

17.Lee, C. K., and D. M. Knipe. 1983. Thermolabile in vivo DNA-binding activity associated with a protein encoded by mutants of herpes simplex virus type 1. J. Virol. 46:909-919. 18. Lee, C. K., and D. M. Knipe. 1985. An immunoassay for the

study of DNA-binding activitiesof herpessimplex virusprotein ICP8. J. Virol. 54:731-738.

19.Leinbach, S. S., and L. S. Heath. 1988. A carboxyl-terminal peptide of the DNA-binding protein ICP8 of herpes simplex virus contains a single-stranded DNA-binding site. Virology 166:10-16.

20. Littler, E., D. Purifoy, A. Minson, and K. L. Powell. 1983. Herpes simplex virus nonstructural proteins. III. Function of the major DNA-binding protein. J. Gen. Virol.

64:983-995.

21. McKnight, J. L. C., T. M. Kristie, S. Silver, P. E.

Pellett,

P. Mavromara-Nazos, G. Campadelli-Fiume, M. Arsenakis, and B. Roizman. 1986. Regulation of herpes simplex virus 1 gene expression: the effect of genomic environments and its implica-tions for model systems. Cancer Cells 4:163-173.

22. O'Donnell, M. E., P. Elias, B. E. Funnell, and I. R. Lehman. 1987. Interaction between the DNA polymerase and single-stranded DNA-binding protein (infected cell protein 8) of herpes simplex virus 1. J. Biol. Chem.262:4260-4266.

23. Orberg, P. K., and P. A. Schaffer. 1987. Expression of herpes simplex virus type 1 major DNA-binding protein, ICP8, in transformed cell lines: complementation of deletion mutants and inhibition of wild-type virus. J. Virol. 61:1136-1146.

24. Powell, K. L., E.Littler,and D. J. M. Purifoy. 1981. Nonstruc-tural proteins of herpes simplex virus. II. Major virus-specific DNA-binding protein. J. Virol. 39:894-902.

25. Powell, K. L., and D. J. M. Purifoy. 1976. DNA-binding proteins of cells infected by herpes simplex virus type 1 and type 2. Intervirology 7:225-239.

26. Quinlan, M. P., L. B. Chen, and D. M. Knipe. 1984. The intranuclear location of a herpes simplex virus DNA-binding protein is determined by the status of viral DNA replication. Cell 36:657-868.

27. Quinlan, M. P., and D. M. Knipe. 1983. Nuclear localization of herpesvirus protein: potential role for the cellular framework. Mol. Cell. Biol. 3:315-324.

28. Quinlan, M. P., and D. M. Knipe. 1985. A genetic test for expression of a functional herpes simplex virus DNA-binding protein from a transfected plasmid. J. Virol. 54:619-622. 29. Rice, S. A., and D. M. Knipe. 1990. Genetic evidence for two

distinct transactivation functions of the herpes simplex virus a proteinICP27. J. Virol.64:1704-1715.

30. Rose, D. S. C., K. Shriver, D. S. Latchman, and N. B. Lathangue. 1986. A filamentous distribution for the herpes simplex virus type 2-encoded major DNA binding protein. J. Gen. Virol. 67:1315-1325.

31. Ruyechan, W. T. 1983. The major herpes simplex virus DNA-binding protein holds single-stranded DNA in an extended conformation. J. Virol. 46:661-666.

32. Ruyechan, W. T., A. Chytil, and C. M. Fisher. 1986. In vitro characterization of thermolabile herpes simplex virus DNA-binding protein. J. Virol. 59:31-36.

33. Ruyechan, W. T., and A. C. Weir. 1984. Interaction with nuclear acids and stimulation of the viral DNA polymerase by the herpes simplex virus type 1 major DNA-binding protein. J.

on November 9, 2019 by guest

http://jvi.asm.org/

(10)

COMPLEMENTATION OF HSV ICP8 MUTANTS 885 Virol. 52:727-733.

34. Shepard, A. A., and N. A. Deluca. 1991. Activities of het-erodimerscomposed ofDNA-binding- and transactivation-defi-cient subunits of the herpes simplex virus regulatory protein ICP4. J. Virol. 65:299-307.

35. Shepard, A. A., and N. A. Deluca. 1991. A second-siterevertant

ofadefectiveherpes simplex virus ICP4 protein with restored regulatory activities and impaired DNA-binding properties. J. Virol. 65:787-795.

36. Southern, P. J., and P. Berg. 1982. Transformation of mamma-lian cells to antibiotic resistance with a bacterial gene under control of theSV40 early regionpromoter.J. Mol.Appl. Genet. 1:327-341.

37. Villarreal, E., and D. M. Knipe. Unpublished results.

38. Wang, Y., and J. Hall. 1990. Characterization of a major DNA-bindingdomain in the herpes simplex virustype1 DNA-binding protein (ICP8). J. Virol. 64:2082-2089.

39. Weller, S. K., J.Lee, D. J. Sabourin, and P. A. Schaffer. 1983. Geneticanalysis of temperature-sensitivemutantswhich define thegenefor the major herpes simplex virus type-1 DNA-binding protein. J. Virol. 45:354-366.

40. Wu, C. A., N. J.Nelson, D. J. McGeoch, and M. D. Challberg. 1988. Identification of herpes simplex virus type 1 genes re-quired for origin-dependent DNA synthesis. J. Virol. 62:435-443.

VOL. 67,1993