Cloning and characterization of oriL2, a large palindromic DNA replication origin of herpes simplex virus type 2.

Vol. 58, No.2 May

0022-538X/86/050513-09$02.00/0

Copyright© 1986, American Society forMicrobiology

Cloning

and

Characterization

of

oriL2, a

Large

Palindromic

DNA

Replication Origin

of

Herpes

Simplex Virus Type

2

DANIEL LOCKSHONAND DENISE A. GALLOWAY* FredHutchinson Cancer Research Center, Seattle, Washington98104

Received 2 December1985/Accepted7February 1986

Anorigin ofreplication within the long unique sequence of herpes simplex virustype2designated

OriL2

has been identified in aposition homologous to its type 1 counterpart,oriLl, between map coordinates 0.398 and 0.413. The difficulties encountered inpreviousattemptstoclone both

OriL2

and

OriLl

in anundeletedform were

surmountedby (i)minimizingthe growth of the host

Escherichia

col,(ii)using a recBCsbcBE.colihost, and (iii)purifyingthefull-length plasmid from deleted forms byusing a novel method which exploits theabilityof apalindrome-containing plasmidtoadoptacruciformconformation,therebydecreasingitssupercoiling.Ina previously developed assay for functional origin

activity,

OriL2 was localized to a 241-base-pair ApaI-SstII

fragment. DNAsequenceanalysisrevealed a 136-basepair,almostperfect palindrome. Comparisonwith

OriLl

showed a very high degree of conservation: the two origins differ in only 16 of the 144-base-pair

OriLl

palindromic region. Mostsignificantly, thedifferencesbetween

OriLl

and OriL2 mainlyoccur inpairsso as to

generallypreserve thepotentialfor intrastrand basepairing.ThecentralregionofOriL2ishomologouswiththe shorter palindromic structures detected in origins located within the repetitive sequences of the short component ofherpessimplexvirus type 1or2.

Initiation of replication of herpes simplex virus (HSV)

DNA can occur atmultiple, widely separated loci. Although this wassuggested originallyby electron microscopic

visu-alization of the replicatinggenome (9, 14), subsequent

stud-ies which characterized the DNA sequences present in defective HSV type 1 (HSV-1) were more definitive

(re-viewed by Frenkel [8]). Thesedefective HSVs fall intotwo

classes. The class I defective genome consists of multiple tandem reiterations ofa fraction of the full-length genome

andalways containsasegmentof the "c"repeatunitof the Ssection of HSV-1DNA. That the classIdefectivegenome

bearsareplication originwasdirectly shown by

Vlazny

and Frenkel (38), who demonstrated that cotransfection of monomer-length class I defective DNA, generated by

re-striction ofthe defective genome, with full-length HSV-1

DNA as helper resulted in the replication ofthe defective

DNA and the regeneration ofits tandemly repeated form.

Origin

activity within the "c" sequence was subsequently

confirmed by Mocarski and Roizman (22). Stow(35) delim-ited this so-called

oris

to a 995-base-pair (bp) restriction fragment,andStow and McMonagle(37)subsequently iden-tified amere 90-bp segmentwhich sufficed for origin func-tion.Inthoseexperimentsarecombinantplasmid containing

the

oris

segment was shown tobeamplified after its

trans-fection into cells which were subsequently infected with

HSV-1. Since

oris

resides in a repeat unit, each HSV standard genome contains two copies (or, in the case of HSV-2, four copies [43]) of this replication origin.

Class II genomes contain instead a minimum of 3.4

kilobases(kb) from aboutthemiddle oftheULregion of the HSV-1genome (33).Attempts to clone this region have been

frustrated, however,by its markedtendency to suffer

dele-tions when grown in Escherichia coli (33, 41, 42). When a

plasmid containing the deleted form of this region of the genome is transfected into cells along with HSV-1 DNA,

defectivegenomes indeed arisefromthe plasmid, but their

*Correspondingauthor.

deleted region has been restored, presumably by recombi-nation with the helper HSV-1 DNA (33). This suggests that the deleted region contains the origin ofreplication, oriL, of the class II defective genome. GrayandKaerner (10), after mapping the region of an HSV-1 (strain ANG) class II defective DNA which deletes in E. coli, sequenced this region of the noncloned defective DNA. Similarly,Weller et al. (42) mapped the deleting region of the nondefective HSV-1 (strain KOS) genome to a 425-bp BstEII-BamHI

fragment. In addition, Weller et al. were successful in cloning the 2.3-kb BamHI-V fragment, which contains the 425-bp fragment, in a yeast vector in an apparently undeleted

form. They showed this construct, p1499, to exhibitorigin

function since after its transfection into animal cells, infec-tion withHSV-1 caused a dramatic amplification ofplasmid.

Bycontrast,deletedclones would notamplifyinthis assay. Both groups(10,42) found oriL of HSV-1 (hereafter referred to asoriLl) to contain a144-bp (i.e., two arms of 72 bpeach)

perfect palindromic sequence. As wasexpected from previ-ous demonstrations ofthe instability of palindromes in E. coli(2, 6, 12, 17, 21, 29), the sequences deleted from p1499 were always at least partially within its palindromic region (42). Thus, the palindrome contains elements required for oriLl function.

orisn

(23, 37, 40) andsubsequently ori2 aswell (43) were

also shown to contain palindromes. These are only about one-third as long the oriLl palindrome, but the middle

regions of allthree arehighly homologous. Knowledge ofthe sequence of oriL2 would completethispicture. Attempts by

Spaete and Frenkel tocloneoriL2inE. coliin anundeleted

form have been unsuccessful (34). The present report dem-onstratesthe construction of E. coliplasmids containingthe undeleted oriL2 sequence. We have also demonstrated clonedoriL2 to be functional. Finally, we have determined theoriL2 sequence. Asexpected,itishighly homologous to

oriLl,

yet severalimportant differencesareapparent as well.

We argue that these differences between the two oriL se-quences further substantiate a physiological role for a cruciformconformation of the palindrome.

513

on November 10, 2019 by guest

http://jvi.asm.org/

MATERIALS ANDMETHODS

Cloning procedures. AllDNA-modifyingenzymes (restric-tion enzymes, T4 DNA ligase, E. coli DNA polymerase I, Klenowfragment ofDNApolymerase,DNaseI) werefrom

New England Biolabs, Inc. or Bethesda Research

Labora-tories, Inc. and were used according to the supplier's instructions or established procedures (19). The cloning vectors used were Kosl (19), pKC7 (27), and pUC19 (45).

Electrophoresis was typically done through a 1% agarose

(Sigma, medium EEO) gel in submerged "mini"orhorizontal

(10by 17 cm) slabs. Preparative gelelectrophoresis(used to

purify theSall fragment of pDL017, to separate the deleted andfull-lengthforms of pDL701, andoccasionally forDNA

sequencing) employed 1% low-melting-point agarose (Be-thesdaResearchLaboratories,Inc.). Thedesired bandswere cut from the ethidium bromide-stained preparative gel,

melted at68°C, extracted with 1volume ofbuffer-saturated

phenol andthenwith 1volume of chloroform, and run over an Elu-Tip column (Schleicher& Schuell, Inc.). The DNA was theneluted from the column as suggested by the supplier and ethanol precipitated. HSV-2 DNA was prepared as

previously described (39). The cosmid pBglGJO was

constructed by ligation ofthe product of the partial BglII digestion of HSV-2DNAinto the Kosl vector.

The host for all plasmids except those otherwise noted was E. coli HB101 (4). E. coli JC9387 (recB21 recC22

sbcB15;J. R. Gillen,Ph.D. thesis, University of California, Berkeley,1974) wasprovided by G.Smith,andstrain JC811

(recB21 recC22 recF143 sbcB15; 15) was kindly sentby F. Stahl. E. coli cells were transformed by established

proce-dures (19). All E. coli growth, both on plates and in

liquid

culture, used L broth (19). After

transformation,

E. coli growthwasoften keptto aminimumby

permitting

growth of

thetransformed bacteriumonlyuntil it gaveacolonyof less than 0.5 mm in diameter. Most ofthe cells in this colony

were then removed into 1 to 11 ml ofampicillin-containing brothandgrownfor12 to 16h to astationary-phase culture, from which plasmidwasthenpurified.Plasmid wasprepared from either 1, 10, or 1,000 ml ofculture by the alkaline

detergent lysis procedure (19). Thorough

plasmid

purifica-tion was achieved using CsCl-ethidium bromide gradients

(19).

Replicon assays. Theprocedure of Stow andMcMonagle (37)wasclosely followed forrepliconassays except that the amount of cells used per transfection and the amount of

DNA per transfected cell were both scaled down.

Briefly,

35-mm-wide wells(sixwells perplate)ofsubconfluentBHK

cells,grown onDulbecco modifiedEagle medium

containing

5% fetal calf serum, were transfected with

equimolar

amounts ofplasmid (20 to 50 ng of plasmid per well) plus

shearedcalfthymusDNA(2.4,ug perwell) ascarrierin 0.2 ml ofcalcium phosphate precipitate. Cells were refed 1 h aftertransfection and then shocked with dimethylsulfoxide 4hafter the DNA was added. At 6 hposttransfection, cells were either mock infected or infected with HSV-2 (strain 333) in Dulbecco modified Eagle medium containing 2% serum at amultiplicity of infection between 5 and 20. After incubation at37°Cfor 15h,all thecellsoftheinfected wells

exhibited cytopathic effects. The medium was then

aspi-rated, 0.5 ml of 0.6% sodium

dodecyl

sulfate-10 mM EDTA-10 mM Tris hydrochloride (pH 7.5)-0.5 mg of pronase per ml was added to the wells, and incubation was continued at 37°C for 4 h. The lysates were transferred to tubes and the wellswere rinsed with 0.5 ml of 10 mMTris

hydrochloride(pH

8.0)-i

mMEDTA.NaClwasaddedto0.2

M, and phenol and chloroform extractions were then

per-formed. The nucleic acid was ethanol precipitated and

adjusted to a concentration of 140 ,ug/ml, as determined

spectrophotometrically.

The "minifold" apparatus (Schleicher & Schuell, Inc.)

wasusedas atemplate for spottingDNAontonitrocellulose (BA85; Schleicher &

Schuell,

Inc.). A 50-,ul volume of

normalizedlysate or control samplewascombined with 50

[L of 0.4 N NaOH ina96-well serocultureplate,and 10

RI

of

thiswas removed into90 [L of0.2 N NaOH for succesive 10-fold dilutions. Beforethe denatured samplewas

applied

tothe

nitrocellulose,

itwas neutralized with0.35 mlof 0.75

MTris

hydrochloride

(pH

7.5)-i

MNaCl. Thefilterwasthen

gently

rinsed, baked,

and probed with nick-translated pUC19, using established procedures (19). When

pUC19-hybridizable

material was analyzed by gel

electrophoresis,

between 1 and 6

pLg

of

lysate

DNAwas

digested

(or not) and electrophoresed through 1%agarose,afterwhichthe

gel

was

blottedbythe Southern method(32) ontonitrocellulose and

thefilterwasprobed asabove.

DNA sequence determination. The chemical

degradation

method of Maxam and Gilbert (20) and the method of

dideoxynucleotide

extension of

synthetic

oligonucleotide

primers by the large fragment ofDNA polymerase using a

closed circular DNA

template (5, 30)

were both used to

determine DNA sequence. The sequenceofboth strandsof

the

326-bp

segment

spanning

theStuI-SstII sites

(Fig. 1)

was

determined,

and any

discrepancies

were

convincingly

re-solved

by

additional

sequencing

runs. The

following

two

fragments were analyzed by the chemical method:

(i)

pDL601, linearized at the KpnI site in the

polylinker

se-quence of its vector, was 3'-end labeled

using

terminal tranferase

(New

England

Nuclear

Corp.)

and

cordycepin

[a-32P]triphosphate

(New

England

Nuclear

Corp.),

and the

labeled insert was then

sequenced

after

being

released

by

HindlIl

digestion

and

purified

on a

preparative gel;

(ii)

after

3'

end-labeling

of the SstII

fragments

of

pDL601

using

terminal transferase and

cordycepin triphosphate,

the

Stul-SstII

fragment

was made

by

BamHI

digestion

and then

purified

on a

preparative

agarose

gel.

The

following

three

determinations

using

the

primer

extension method were done:

(i)

pDL601was

hybridized

with

d(GTAAAACGACG-GCCAGT) (New England

Biolabs, Inc.);

(ii) 701A,

the

full-length

subpopulation

of

pDL701 (see

Fig. 3),

was

hy-bridizedwith the "reverse

sequencing primer"

d(AACAG

CTATGACCATG) (New England

Biolabs);

and

(iii) 701B,

the deleted

subpopulation

of

pDL701 (see

Fig. 3),

was

hybridized

with the same reverse

primer.

RESULTS

Cloning of

OrlL2

in an undeleted form. In

cloning

oriL2,

difficulties

comparable

tothoseseen whenoriLl was

propa-gatedinE.coli(33, 41, 42)were

anticipated. However,

in the presentstudy, problems onlyaroseafter extensive

subclon-ing.Figure1 illustrates theseriesof

subcloning

stepswhich

purified

oriL2 from the

surrounding

HSV-2 sequence.

Be-cause the HSV-1 and HSV-2 genomes are known to be

colinear,

oriL2was

anticipated

tolie within the 0

fragment

of

BglII-digested

HSV-2.

Digestion

of cosmid

pBglGJO

with

BglII

and

ligation

of the

product

intothe

BglII

siteof

pKC7

gave

pBglO.

Comigration

of the authentic

4,650-bp BglII

0

fragment

ofHSV-2

(Fig. 2,

lane

2)

with this insert

(Fig.

2,

lane3,lower

band)

indicates thatnodetectabledeletionhad occurred. In

fact, however,

after

purification

of the whole insert from

pBglO

and

digestion

withSalI

(Fig. 2,

lane

5),

a

smallamountof deleted

fragment

was seeninadditiontothe

on November 10, 2019 by guest

http://jvi.asm.org/

D RP G J 0 C N I H M L K

119 II

"I

I

Construct

nRnl C1

Insert size(kb)

Tendency vector to delete 34 KosI N.D.

PvuI[ _

tUl

HindM SalI StuI Sst1 Pst I SalBgIIE

I

S5I

SalIKpnI

SolI 'ApaI

pBgl 0 4.65 pKC7 verylow

pDL017 2.28 pKC7 verylow

pDL506 1.68 pUCl9 low

pDL601 1.40 pUC19 low

pDL701 0.33 pUC19 moderote pDL411 0.25 pUC19 N.D.

FIG. 1. Restrictionenzymecleavagemapofthe regionsurrounding oriL2. ThetoplineportraystheBglII fragments of HSV-2 (7),and the

repeatedsections of the L and S regionsareindicatedby boxes. The lines belowportraythesuccessive subcloneswhich containoriL2.The

tendency for the subclonestodelete thepalindrome, aqualitativeassessmentdescribedin the text, issummarizedintherightward column. ClaI,PvuII, SstI, KpnI, StuI,Sall, HindIII,andSphIcleave the entireBglII0fragment onlywhereindicated.PstIcleavesthepDLO17 insert

only where indicated, but has additional sites in therestofBglII 0.SstIl cleavesthepDLO17insertatleast fourtimes,butonlytheonesite whichwasmappedis shown. TheClaI site is cleavable in pBglOgrownindam, butnotdam',E.coli. The ApaI sitewasdeduced from DNA

sequence.The156-bppalindromic regionisrepresented byadouble-headedarrow.

full-length, 1,680-bp internal Sall fragment. The unique Hindlll sitewithin theBglII 0fragmentwas usedtobisect the insertto give pDLO17 (containing the rightwardhalfof the BglII 0fragment) and pDLO18 (containing its leftward

half).Inanassaydescribedbelow, bothpBglOandpDLO17 contained a functional HSV DNA replication origin,

whereaspDLO18 didnot.TheSalI fragmentwithin the insert ofpDL017wasthenpurifiedandligatedintotheSalIsiteof pUC19, ahigh-copy-number vector,to givepDL506.

Unlike pBglO and pDL017, it was obvious that pDL506 suffered deletions. Some E. coli transformants from this construction yielded homogeneous populations ofplasmid from which between 60 and 150 bp of insert was deleted,

whileother transformants containedamixtureof deleted and

apparently full-length molecules. As lanes 7 and 8 ofFig. 2 show,thefraction of deleted material could be kept accept-ably small (<10%) by limiting the extent of E. coli growth before plasmid preparation. By comparing restriction pat-ternsofplasmid from fully deletedtransformants with those ofpDL506, the deletionswerealwaysfoundtomapbetween the unique StuI site and the SstII site portrayed in Fig. 1. Thus, the next subcloningstepinvolvedtheexcision of the smallerSalI-StuIsegmentfrom theinsert by digestionwith StuI andXbaI, the latterof which cleaves within thevector polylinker. Repairof theends by DNApolymerase (Klenow fragment) treatment and recircularization by DNA ligase produced pDL601.

pDL601 suffered deletiontoabout the sameextentasdid

pDL506, and again, the fraction of deleted material in its preparations could be lowered sufficiently (toless than 10% of thepopulation) by minimization of E. coli growth (Fig. 2, lane9). Thefragmentof inserton therightward side of the

StuI-SstIIsegmentwasthenexcisedby digestionwith SstII

and PstI, the latter of which cuts in the vector polylinker rightward of the Sall site, to produce pDL701. This

con-struct, however,showed an evenhigher tendency todelete. Even when its E. coli H13101 host was grown as little as

possible, approximately half of the population of plasmid

molecules was deleted (Fig. 2, lane 10). Since DNA

se-quence cannot be determined from such a heterogeneous

population, the deletion problem now had to be solved in otherways.

Theeliminationof deleted moleculeswasachievedbytwo different methods. Initially, the closedcircular deleted and full-length populations were purified from eachother on a

preparative agarosegel. The unfractionated plasmid prepa-ration ("701"; Fig. 3, top) contained, in additionto nicked circles, both a highly supercoiled fraction (comprised ofa

series oftopological isomers which, because of their high supercoiling, were notresolvableon thisgel)and afraction

containing the less highly supercoiled topological isomers. Thesetwofractionswereseparated by cuttingapreparative

gel throughwhichalargeamountof 701waselectrophoresed

togive701A and 701B (Fig.3, top). Releaseof the insertby BamHIandHindIII digestion (Fig. 3, bottom)showed701to containtwospecies ofinsert,whereas701A, the lesshighly supercoiled fraction,containedonlythelarger insert (shown laterto be undeleted) and 701B containedonlythe smaller, deleted form.The nicked forms whicharose during

fraction-ation also illustrated that 701A is largerthan701B (Fig. 3, top).

By changing the E. colihost strain, thedeletionproblem

was subsequently more easily solved. A recBC sbcC host

genotype has been shown to enhance palindrome stability (16, 25, 44). Another studyhas demonstratedthat, in addi-tion,aninactiverecFgenewillallowmaintainence ofalarge

palindrome (3). Whenpurified701A(Fig. 3, top) was

intro-duced into E. coliJC9387(recBC sbcB),mosttransformants, aftergrowthto at least the same extentas the E. colifrom whichthe701 preparationinFig. 3waspurified,containeda

high fraction (>90%) offull-length plasmid (the remaining transformants bore exclusively deleted material). In the

same experiment, transformants of E. coli JC8111 (recBC recF sbcB) had either too low a proportion offull-length

material (never more than 60%) or only deleted material (data not shown). Thus the palindromic sequence in oriL2

G J 0

58,

L =

IJLYI%vv%

- -j

Bgl]jr---

---.StIPvull

ClaI HindM SphI _eBglll

on November 10, 2019 by guest

http://jvi.asm.org/

[image:3.612.129.489.72.270.2]

C---650 4370- 2320--2030'

HE

-

-t-.i 5-:

I-O. C) f

I-i

0 00

0n (i!

_i -j

-L... C-2

2-CIL C.) :L

1350

i 1080

-870 --600 -275

2 3 4

6

7

6

9

lull

FIG. 2. Restriction enzyme analysis of oriL2 subclones. Lane 2, BglII-digested HSV-2 DNA. Lane 3, Intact insert ofpBglO,released byBglII digestion. Lane 4, Intact pBglO, after SalI digestion, gave the 1,680-bp Sallfragment of the insert in addition to two vector-insert junction fragments and a vector-derived fragment. After purification of the insert from pBglO, digestion with Sall (lane 5) gave the internal 1,680-bp SalI fragment and, in addition, a small amount of deleted SalI fragment migrating slightly faster than the full-lengthform.SalIdigestion ofpDL017(lane 6)again revealed the SalIfragment, but the small amount of deleted Sall fragmentwas obscured, asin lane 4, by the vector-derived band. Lanes 7 and 8, Two different preparations of pDL506 were digested with Sallto -release the full insert. The plasmid in lane 7 was prepared from extensively grown E. coli, whereas in lane 8, host growthwaskept toaminimum. Lane 9, Release of the complete insert of pDL601 (from minimally grown E. coli) by restriction with BamHI and HindIII, two enzymes whose sites reside in thevectorsurrounding thisinsert, providing the 1.4-kb insert andasmallamountof deleted material as well. Lane 10, Digestion of pDL701 (prepared from minimally grown HB101 host) withBamHI andHindIII gave the full-length 335-bp insert andacomparable amount of190-bp deleted insert. Fragment size markers areHindIII-digested X phage DNA (lane 1) andHaeIII-digested4X174RFform(lane 11).

appears to be more stable in strain JC9387 than in strain JC8111.

The DNA sequence analysis during the course of the

subcloning revealed a unique

ApaI

site within the 326-bp insert ofpDL701 and to theleftofthepalindromicsequence.

The finalsubcloningstep wasthus the excision of this85-bp leftwardfragment by digestion withApal

andKpnl

(another polylinker site) to produce pDL411. In this construction, strain JC9387 served as the immediate host of the

ligation

product. Again, >90% of the plasmid preparation was full

length (data notshown).

Analysis of origin function in vivo. As was shownfirstwith simian virus 40 (24) and later with HSV-1 (35, 37), cloned

viral origins can drive the replication of their bacterial vectors when thetrans-acting viral factors are present in the

transfected cell. Using this method, analysis of a series of

BAL 31 exonuclease-generated deletants of cloned HSV-1 enabled Stow and McMonagle (37) to delimit sequences necessary forthefunction of

orisl.

We have employedthis

repliconassay to show that cloned oriL2 is functional.Inthe

simplest application ofthisprocedure(Fig. 4),HSV-2 infec-tionwasshown toamplifysometransfectedplasmidsbut not

others. Total DNA from BHK cells, transfected with a

plasmidandsubsequentlyinfected withHSV-2,wasspotted

onto a nitrocellulose filter and probed for

plasmnid

vector

sequence. Therewas anenormousincreasein dotintensity,

uponinfection with HSV-2, of DNA fromcells transfected

with eitherpBglO, pDLO17, pDL601, pDL701, or pDL411

(Fig. 4).Incontrast, amuchsmallerincreaseindot

intensity

was observed when cells were transfected with pDLO18 (containingthe leftward half of BgiII 0) or pUC19. Thus,

oriL2isprobably contained within the former plasmids, but notthelatter.

Thevector-hybridizable materialinthese lysateswasalso

tested formethylation of theN6 position of adenine in the sequenceGATC. Thispropertypermitsdistinctionbetween

DNA replicated in E. coli, which modifies this sequence using dam methylase, from that synthesized inaeucaryotic cell, which is apparently incapable of methylating (or demethylating) the N6positionof adenine(28).Digestion of

lysates with either MbcI, which cleaves only at

unmethyl-atedGATC,orwith DpnI,which cleaves thesame sequence only when methylated, serves to testthis property. Figure

SA examines lysates from pDL411-transfected cultures.

Lane 1 demonstrates the vector-hybridizable material from thepDL411-transfected andHSV-infected culturesto be of high molecular weight. This was expected, since earlier reports of replication of cloned

oris1

(35) and ofdefective

HSV-1 (38) have shown the producttobeatandemrepeatof high molecular weight. Digestion with MboI (Fig. 5, lane 2) completely converted this slow-migrating material into frag-ments which displayed the same mobility as the digestion prQducts of authentic pDL411 (Fig. 5, lane 12). However,

digestion withDpnI had noapparenteffect (Fig. 5, lane 3). DpnIwasindeedactive, since DpnI digestion after addition of E. coli-derived pUC19 to the lysate (Fig. 5, lane 5)

produced bands characteristic of DpnI-digested pUC19(lane 10). Lane 6 of Fig. 5 contains undigested lysate from a culture which was not infected with HSV-2 after pDL411

transfection. As demonstrated by the dots of Fig. 4, far less vector-hybridizable material was apparent than in the HSV-2-infected counterpart (Fig. 5, lane 1). The mobility of the uninfected material corresponded to that of authentic pDL411 (lane 11), and none ofthe high-molecular-weight

material of lane 1 was seen. As expected, MboI digestion

(Fig. 5, lane 7) could not alter its mobility, whereasDpnI

Uncut

Born HI

HindITT

N: ie:c %irc'es

*SJpercoIle,d

1-3C;7l

S1ltP

FIG. 3. Separation of the full-length and deleted forms of pDL701. The upper panel contains undigested samples, and the lower shows samples digestedwithbothBamHIandHindlll, two enzymes whichcut oneitherside oftheinsert, therebyreleasingit. Lane701, Plasmid pDL701 preparedfrom minimallygrownstrain HB101host. The resultof excision ofalltopoisomers(i.e., exclud-ingthenickedcircles) except thehighly supercoiled forms froma

preparativegel of701is shown in lane 701A. Materialinlane 701B waspurifiedfrom ahighlysupercoiledbandof thesamepreparative gel.

.: .." , F, ., ,.;: .1

on November 10, 2019 by guest

http://jvi.asm.org/

[image:4.612.98.266.69.257.2] [image:4.612.375.498.496.632.2]

digestion (lane 8) converted it to fragments which corre-sponded in mobility to authentic DpnI-digested pDL411 (lane 12). Thus, Fig. SA demonstrates that the large amount of vector-hybridizable material which appears upon infec-tion of a pDL411-transfected culture with HSV-2 is indeed synthesized aftertransfection since (i) it has the character-istic high molecular weight and (ii) it is unmethylated on the N6position of adenine in the sequence GATC.

Can HSV-2 infection induce any replication of transfected DNA whichlacks HSV origin sequence? Thiswas apossible interpretation of the10-fold increase in dot intensity of either pDL018- orpUC19-transfected cultures. The methyl group assay was used to test for small amounts of replication of these transfected plasmids (Fig. 5B). Unlike the slow-migrating material seen with pDL411 transfection (Fig. SA),

transfection with pDLO18 (Fig. SB, lane 14) or with pUC19 (lane 23), followed by HSV-2 infection, gave

vector-hybridizable material which comigrated with input DNA. Moreover, this material was DpnI sensitive(Fig.5B,lane 15,

pDL018; lane 24, pUC19) and MboI insensitive (lane 17,

pDL018,lane25,pUC19).Thus,neitherpDL018 nor pUC19 wascapable ofreplication. Origin function istherefore quite specific: only when the transfected DNA bears a certain sequence is it capable of replicating after the trans-acting

factors are provided by HSV-2. The increase in

vector-hybridizable material upon HSV-2 infection of cells transfected withplasmids whichlackfunctional HSV origins

couldperhaps be attributed to an increasedpermeabilityof

cellsto residualcalcium phosphate-precipitated DNA.

SequenceofOrHL2. WhenpDL601 and 701A, the material

purified from pDL701 growth in HB101 (Fig. 3), were

subjectedtoboththedideoxy triphosphatechain termination

(5, 30) and the chemical modification (20) methods, the sequence ofthe 326-bp StuI-SstII segment was determined

(Fig. 6). The replicon experiment described aboveconfines

thesequenceimportant fororiL2functionto bases 86 to 326.

Thetwo armsofthepalindrome, which extend frombase 122

pBqlO p3iO, pl_018 pDL601 pX720 4'

+

10-fold

dilutiorns

1.0-foG!d

didutiors

...12 3 4 5

[image:5.612.319.557.74.275.2]

pUCi CT L

FIG. 4. Hybridization, to vector sequence, of lysates ofcells transfected with oriL2plasmidsfollowedbyHSV-2ormock

infec-tion. Two wells of BHKcells were transfected withthe indicated

plasmids combined with carrier DNAor with carrier DNA alone

(CTRL). Afterdimethyl sulfoxide shock, HSV-2 infection (+) or

mockinfection (-), andovernightincubationat37°C,cells from 16 wellswerelysed, and DNAwasprepared. Tenfold dilutionsof the

lysateswere then spotted onto nitrocellulose. Inaddition, 10-fold dilutions ofvarious controlswerespotted.Column1, 2ngofpUC19 plus undiluted CTRL, + lysate; column 2, 2 ng ofpUC19 plus undilutedCTRL,-lysate; column3, 2ngof pUC19 only;column 4,

0.2,ugof HSV-2DNA;column5,2ngofpKC7.Thefilterwasthen

probedwithnick-translated pUC19.

A

1 2 3 45 6 7 8 90 1 I2 13

qm a

I

..

xIt

4 .~~~~~~~4

_ _ do @

4

I

B

14 15 16 17181,920 21 22 23 2425

*~~~~~~~~~.E

7')

FIG. 5. Digestion of lysates with MboI and DpnI. pDL411-transfected lysates (A) and pDL018 and pUC19 lysates (B) were analyzed. (A) Lanes 1 through 5 contain lysate from pDL411-transfected andHSV-2-infectedcells.Lane 1, Undigested; lanes 2 and 4,MboI-digestedlysatealone (lane 2) orlysate plus 0.2 ng of pUC19 (lane 4); lane 3 and 5, DpnI-digested lysate alone (lane 3)or lysate plus 0.2 ng ofpUC19 (lane 5). Lanes 6 through 8 contain lysatefrompDL411-transfected and mock-infected cells. A ninefold longer autoradiographic exposure of these three lanes is shown. Lane 6,Undigested; lane 7,MboIdigested; lane 8,DpnI digested. Lanes 9through 13 contain various markers. Lane 9, pUC19 (0.2 ng), undigested; lane 10, pUC19 (0.2 ng), DpnI digested; lane 11, pDL411 (0.3 ng), undigested; lane 12, pDL411 (0.3 ng) DpnI digested; lane 13, X phage DNA, HindIII digested and end-labeled with 32p. (B) Lanes 14 through 17 contain lysate from pDL018-transfected and HSV-2-infected cells.Lane14,Undigested;lane15, DpnIdigested; lane 16, lysate combined with 0.2 ng ofpUC19 and thendigestedwithDpnI; lane 17, MboIdigested. Lanes 18 and 19 contain pDL018-transfected, mock-infected lysate and are undi-gested (lane 18)orDpnIdigested (lane 19).Lanes 20through 22 bear various markers. Lane 20, pDLO18, undigested; lane 21, DpnI-digestedpDLO18;lane22,DpnI-digestedpUC19.Lanes 23through 25containpUC19-transfectedandHSV-2-infectedlysate.Lane23, Undigested; lane 24,DpnIdigested; lane 25,MboI digested.

to base 257, are denoted in Fig. 6 by brackets. Note the

absence of any intervening nonpalindromic sequence. The

ATrichness ofthe oriLl palindrome has been noted previ-ously (42)and isreadilyapparent fororiL2 as well. The37%

A+T content ofthe entire 326-bp segment is only slightly higher than the 31% A+T content of the entire HSV-2 genome. The 136 palindromic base pairs are 51% A+T,

while the 190 bp of flanking nonpalindromic sequence is

merely 28%A+T.The middle 96bp ofthepalindromehas an A+T content of 64%. Thus, as far as base composition is concerned, oriL2, likeitscounterpart inHSV-1,iscomposed of GC-rich domains which flank the AT-rich palindrome. ThishighAT contentis concentrated in the centralregion of thepalindrome.

The deleted subpopulation ofpDL701 grown in HB101, 701B(Fig. 3),wasalsosequenced usingthedideoxy triphos-phatechain termination method. It was found to beidentical

in sequence to 701A except for the deletion of bases 94

through 241 of Fig. 6. This directly demonstrates, as had been shown previously in cloning oriLl (42), that deletion occursatleast partiallywithin thepalindrome.

4- -r- -

-0 0

-AW&WOMMAL. A..

on November 10, 2019 by guest

http://jvi.asm.org/

[image:5.612.84.277.460.604.2]

CCTGCGGATG GCGGCCCAGA GCTGGGTATG CTCGGCCGGG GCGGCCGGTA TATGTACGGC 60 GTGCTGGGAG GGGCGGCGTC GGGCCCCGCC CACGGTCCGC CACGCCCCGC GCGTCATCGG 120 C[AGGGGGCGT GGTCGCCCTT CTAAAAAAAG TGAGAACGCG AAGCGTTCGC ACTTTGTCCT 180 AATAGTATH ATATTATTAG GACAAAGTGC GAACGCTTCG CGTTCTCACT TTTTTTAGAA 240 GGGCGGCCAC GCCCCCTJTTG ACGTCACGCT CACCCGGGCG GCCGGCCGCC CATAAGCGCG 300

GCCTGCCGGG CCGATAAAAA GAAACC 326

FIG. 6. DNA sequenceoftheinsertofpDL701.Onestrand,reading 5' to 3',of theHSV-2sequence contained inpDL701isshown. Each 68-bparmof the palindrome is enclosed inbrackets.ApaIcleaves between bases 85 and 86.

DISCUSSION

Cloning of oriL2. A 241-bp fragment of HSV-2 which

containsoriL2has beencloned.Weconcludethat nodeletion has occurred during the course of this construction (aside

from that inasmallfraction of the plasmid population)since

(i) the cloned oriL2 is functional in vivo and (ii) when the

clonedoriL2 sequenceiscompared to the sequenceof

oriLl

derived from viral (uncloned)DNA(10,42), thesesequences are highly homologous and colinear, with no indication of

sequence present inoriLl thatismissingin oriL2 (Fig. 7).

Deletion did pose certain problems, butthese were sur-mounted by minimizingthegrowthof the host, bypurifying deletedfrom undeleted plasmid, and by using abetterhost

strain. The problem is most simply solved by limiting the

growth ofthe E. colitransformant. If this is unsuccessful,

the use ofadifferenthost strain is necessary. Since oriL2 is

more stable in strain JC9387 (recBC sbcB) than in strain JC8111 (recBC recF sbcB), whereas the large palindrome from minute virus of mice is more stable in the latter (3),

both strains should be tested whencloning palindromes.

Thepurification of undeletedpalindrome-containing

plas-mid reported here could be used when the simpler

ap-proaches prove unsuccessful. This procedure (Fig. 3)

ex-ploits the ability of palindromes to adopt the cruciform

structure(21, 26). The numberofsupercoils is determinedin part by the number of primary helical turns in a closed

circular molecule of DNA. The extrusion of palindromic

sequenceintoacruciformconformation depletesamolecule

of

primary

helical turns. The

negative supercoiling

ofthe

plasmid thereby decreases byanamountwhichdepends on

the size ofthe palindrome. In the case oforiL2, about 13

supercoils are removed by cruciform formation (136 bp of palindrome + 10.4bpperhelicalturn). Theaverage number

of supercoils in noncruciform pDL701 is about 18 (0.096 helicalturnsperbp x 3,100 bp x 0.06supercoilsperhelical

turn; 1). Crucifortn formation, which

presumably

occurs

only during the course of plasmid purification, therefore causesasubstantialreductioninsupercoiling. Since

electro-phoretic mobilitydepends on the degree of supercoiling, the

plasmid which is capable of a decrease in supercoiling by virtue ofcruciform formation is separable from that which

lacks the palindrome.

Palindromic

sequences are known tobepoorlypropagated

inE.coli(2, 6, 12, 17, 21, 29). oriLof HSVwasfirst reported

to deletein E. coli by Spaete andFrenkel (33), who

exam-ined HSV-1. Weller. et al. (42) were subsequently able to

cloneoriL, butonly ina yeast vectorsincedeletionororiLl inE.coli provedto betoo severe.SinceoriLl

cloned

in yeast was functional in a replicon assay similar to the one

de-scribed here, it appears that at least part of the plasmid preparation used was undeleted (42). Previous attempts to

maintain

oriL

of HSV-2 in E. coli havebeen unsuccessful. Spaeteand Frenkel(34) indicatethat theSall

fragment

(Fig.

1,this paper) suffers "small"deletions. Haywardand Reyes (13) had earlier commented on such deletion in HSV-2, although they claimed thatthe ClaI site, which is 1,200 bp

away from the palindrome (Fig. 1, this paper), is always

within the 120 bp of deleted sequence. Since we find no evidenceof deletionaround theClaI site, andsincetheClaI siteis cleavablein plasmid fromdam butnot

dam'

E. coli

(data not shown), it islikelythat theearlierstudy(14) didnot detect aClaIsite in theclonedDNAdue todammethylation

rather than to deletion. The present study, in successfully cloning full-length oriL2 in E. coli, allowsthepreparation of relatively largequantities ofanoriLclonecompared with the

small amountsobtainable fromayeast vector. Theability of

the present study, but not previous ones, to obtain full-length oriL in E. coli can be most likely attributable to

technical differences andperhapsto aninherentdifference in thebehavior oforiLl andoriL2 in E. coli, although the

high

degree of similarity between the oriLl and oriL2 sequences argues againstthe latterinterpretation.

What might account for the instability of

palindromic

DNA inE.coli?Onefact whichmayprove to be relevant to this problem is the marked decrease in stability of the smaller subclones of oriL2. ThiswasalsoseenwithoriLl(42).

S2 Si

GCCC TC A A C T C TA A A

5 -CCCACGGTCCGCCACGCCCCGCGCGTCATCGGCAGGGGGCGTGGTCGCCCTTCTAAAAAAAGTGAGAACGCGAAGCGTTCGCACTTTGTCCTAATAGTATA

3 -CCGCCGGCCGGCGGGCCCACTCGCACTGCAGTTTCCCCCGCACCGGCGGGAAGATTTTTTTCACTCTTGCGCTTCGCAAGCGTGAAACAGGATTATTATAT

T G C CTGTC TCGG AA AT T

FIG. 7. Palindromic regionof oriL2 and itscomparisonwithoriLl. Onestrandof the oriL2 sequence(bases89through290 ofFig. 6) is portrayed inasnakelikefashiontoillustratethepotential for intrastrandbasepairing. The A and Tontherightwardendsofthetwolinesare

adjacentinitssequence. Eachasteriskdenotes apotential intrastrandbasepair of oriL2.The basesoforiLl(42) whicharedifferentfromoriL2

areindentifiedbyletters above orbelowthe lines oforiL2 sequence. Forinstance, oriLl containsanAinsteadof G in the fifthpositionfrom therightendofthe topline.The caratsindicatethat the A and T are inserted inoriLl relativetooriL2. Theextentof thepalindromes of

oris1,

oriS2, oriL2, and oriLl are indicatedbyarrows labeledSi,S2, L2,and Li,respectively.

on November 10, 2019 by guest

http://jvi.asm.org/

[image:6.612.86.544.608.670.2]

HSV-2

20 40 60 80 O1

Orn SI CATTGG-CCGGCGGGCGGGN-NNCCCCA---AGG-GGGCGGGGCCGCCGGG-TAAAAGAAGTGAGAACGCGAAGCGT

ori S2

CGTTGGTCC-GCGGGCGT

CTCCGCCCCAA---AGG-GGGCGGGGCCGCAGGG-TAAAAGAAGTGAGAACGCGA

I I I11 IIII II II IIH il II

ori

L2

CCACGGTCC-GC .

CACGCCCCGCGCGTCATCGGC..ugu--orn LI CCAGCCCCT-CC.---CACACCCCGCGCGTCATC

0

midpoint 120 140 160 180 200

on, SI CACTGGCGCCGTGCCCGAC-TCCGCGCCGGCCCCGGGGGCGGGCCCGGGCGGCGGGGGGCGGGTCTC

11111111111111111111111 11111111111 11111 1 111 11i11 111 11 11111

ori S2 GCGCCCTGCCCGGGGCCCGCGTCATCCCGCGGTGCCCGCCCCAAAGGGGGCGGGGCCGCAGG

or0 L2 TTGACGTCACGCTCACCCGGGC

ori LI

1W

ACGTCACGCTCCTGTCGGC

FIG. 8. FourHSV replicationorigins. Sequencesarealignedsothat thecentersof thepalindromes (notedas"midpoint") coincide. The palindromic regions areindicated with white lettering. To allow the optimal alignment ofthe sequences, gaps (indicated atdashes) are

inserted. Thus,for example, G (base 60)and T(base 61)areadjacent intheoris1 sequence.Verticallines indicate that correspondingbases inadjacent sequencesarethesame. The

oris1

sequence is takenfrom reference23(bases 2345through2526; N indicates bases whichwere

undetermined); theoriS2sequence is taken from reference43(bases711through896), theoriL2 sequenceshown is bases 90through 279 of Fig. 6,and the oriLl sequence is taken from reference42 (bases147through 338).

In the present study, pBglO and pDL017 were initially thoughttoberefractorytodeletion. In retrospect, however, it appears that a small fraction of pBglO prepared from extensively grown E. coli is deleted (Fig. 2,lane5).pDL506 andpDL601 haveahighertendencytodelete, and extensive E. coligrowth had tobe avoided togive a sufficiently high fraction of full-length material. Further subcloning to give pDL701 caused an even higher deletion rate. Presumably

pDL411 tends to delete to a similar, or perhaps greater extent than does pDL701, although this was not tested.

Unfortunately,nosimplewaytoquantitatethisphenomenon is apparent. Theincrease indeletion tendency withsmaller

insert size is not due simply to a change in plasmid size. When the Sall fragment of pDLO17 was ligated into pKC7 (5.8 kb) instead ofpUC19 (2.7

kb),

theproducthad asgreat atendencytodeleteasdidpDL506 (datanotshown). Thus,

it appears that some feature of the surrounding HSV se-quence has a stabilizing effect on palindromes in E. coli. Examination of these flanking sequences would be most

interesting if other natural eucaryotic palindromes could be shown to actsimilarly inthisregard.

Features of theOrlL2 sequence. The most salient featureof oriL2-indeed, the feature which prevented it from being easily cloned-is its 68-bppalindrome. Thisis illustrated in

Fig. 7 by noting with asterisks the 66 potentialintrastrand base pairs. Note that aside from two mismatches, the

palindromeisperfect, i.e.,therearenobases whichserveto

interruptthe register ofthetwo arms.

Figure7also illustratesthedifferences between oriL2 and

oriLl,

notedaslettersabove and belowtheportrayal oforiL2

sequence. The 16differences whichoccurwithin the 144-bp

palindromic region (rightward ofthearrowlabeled

Li

inFig. 7)areoftwotypes: those in oriL2which eliminate potential

intrastrand basepairs, and those whichoccurin pairs so as topreservethe potentialof intrastrand basepairing. Ten of the16differences fall into this latter group of paired changes, as far higher fraction than should occur if there were no selection pressure to maintain the relationshipbetween the twoarmsofthepalindrome. Fortwochangestoconstitutea

pairwisechange, theymust notonly occuratcorresponding

positions onthetwo arms, butthey must also be alteredto the correct base as well. Four of the pairs of changes are

simply thesubstitution of Gand CforAand T or viceversa. The fifth pair of changes, the insertion ofTand Ain oriLl

relativetooriL2,is

particularly striking.

Thus,despitethe 16

differences in the oriLl and oriL2palindrome, the twodiffer by only sixin thenumber ofpotential intrastrand basepairs.

At least two functions ofpalindromes can be envisioned

(21). It is possible that the twofold rotational symmetry inherent in a DNA palindrome is important only for its interaction with a dimeric protein complex which also

ex-hibits twofoldsymmetry. Eachhalf ofthedimerwould thus

recognize one arm of the palindrome. To explain the high degree of

perfection

ofthepalindrome, theprotein complex

would have to be atleast as long as thepalindrome, binding

to it veryintimately soas torecognize all the bases over its

entire length. Ifthis were notthe case, there would be no

basistoexplaintheconservation of

palindromic perfection.

Considering

the extreme structural demands which this

model would placeon suchahypothetical protein complex,

we consideritunlikely.

Inthe other model, the palindrome exists because of its ability toadoptacruciform conformation suchthatthetwo arms of the palindrome base pair with each other in an

intrastrand fashion. This is suggested merely by the

exist-enceofthese large perfect palindromes. Thecomparison of

thetwooriL sequencesevenfurthersuggeststhis model:the

differences, occurring mainlyin pairs, serveto maintainthe

potential for intrastrand base pairing. We thus consider it

likelythat the oriLpalindromebecomesacruciformduringat least partofthe virallifecycle. Cruciform formation iswell knowntobe influencedbyDNAtorsional strain(11, 21,26),

and thisiscertainlya

plausible

mechanism forahypothetical

oriLstructural transition. It should benoted, however, that there is no evidence for such strain in HSV-1 in vivo as determined by the rate oftrimethylpsoralen photobinding (31).This doesnoteliminate thepossibleexistenceofasmall class ofreplicating viral genomes which have unrestrained supercoils, ashas beenseen in simian virus 40 (18).

Itisunlikelythat the basepairingin thecruciformcould be maintained all the way to thecenter of thepalindrome, yet thereisno nonpalindromic sequencebetween itstwo arms.

Interestingly, this central region is not simply palindromic

but is, moreover, an alternatingAT sequence. Perhaps the VOL.S58,1986

I

I

I

on November 10, 2019 by guest

http://jvi.asm.org/

[image:7.612.57.561.64.239.2]

unique sequence in this region serves a function different from that of the rest of the palindrome. Its AT composition makesiteasily meltable, and the alternating sequencecould

be necessary to allowthis region to melt as easily aspossible

(11) to allow the initiation of cruciform formation or of

replication itself. Stow (36), in fact, has shown the center of

the

oris1

palindrome to be essential fororigin function.

Figure 8 compares the four HSV ori sequences. The

relationship between oriLl and oriL2 has been discussed above. Whittonand Clements (43) havepreviously discussed

the relationships between

oris1

and

oriS2.

As reiterated in Fig. 8, these two sequences arevirtually identical. Welleret

al. (42) have discussed the relationship between oriLl and

orisl,

and these are notdirectly comparedinFig.8. Because

ofthe close similarity between oriLl andoriL2and between

oris1

and

oriS2,

general comparisons between oriL and

oris

can be made. As noted previously (42), the sequences

leftward ofthe centersofthe oriLand

oris

palindromesare

highly homologous, whereas rightward of this center, the

homology extends only for about 25 bases. Thus, the

oris

palindrome is quite similartothat oforiL, but isonly about

25 bases long.

Acomparison of

oris

and

oriL

suggeststhat notall of the

moreextensive palindrome oforiLisrequired for replication. After all, since

oris1

is functional with merely a 23-base palindrome(37), and 47 or soadditionalpalindromicbases in oriL should be nonessential. Furthermore, since Stow and

McMonaglehave shown that aplasmid (Sl1 in reference 37)

which contains 298 bp of HSV-1 sequencerightward of base

51 (Fig. 8) is afunctional

oris1

replicon, whereas aplasmid (S13 in reference 37) whoseHSV-1 insert extendsrightward from base 84(Fig. 8) isnonfunctional, a sequenceessential

for

oris5

function resideswithin the sequencehomologousto

the 25ththrough the 57th base oftheleftwardarmoftheoriL palindrome (bases51 through 83 of

Fig.

8). In other

words,

asequencewhich isessential for

oris

function lies outside of its palindrome, while itshomolog in oriLiscontained within

the oriL palindrome. If

oris

and oriLfunction

similarly,

an

assumptionatthis

point

based

solely

ontheir

homology,

this

again suggests that the outer

region

ofthe

palindrome

is

required forsomefunction otherthan

origin activity.

A role for the oriL palindrome in the control of

replication

or

transcription thus becomes a

tempting possibility.

If this viewwere correct, the outerregion ofthe oriL2

palindrome

should be superfluous with respectto

origin

functionin the

repliconassay. We arecurrently

examining

the

significance

ofvariouselementswithin theoriL2regionfororiginactivity.

ACKNOWLEDGMENTS

We thank Jim Wallace and Margaret Swain for help with the computerassisted analysis of the DNA sequence, Toni Higgs for typing, and Paul Su for artwork.

This study was supported by Public Health Service grants CA35568 andCA26001from the NationalCancer Institute.

LITERATURE CITED

1. Bauer, W. R. 1978. Structure andreactions ofclosedduplex DNA. Annu. Rev. Biophys. Bioeng. 7:287-313.

2. Behnke, D., H. Malke, M. Hartmann, and F. Walter. 1979. Post-transformational rearrangement of an in vitro recon-structed group-A streptococcal erythromycin resistance plas-mid. Plasmid2:605-616.

3. Boissy, R., and C. R. Astell. 1985. An Escherichia coli recBCsbcBrecFhostpermitsthedeletion-resistantpropagation of plasmid clones containing the 5' terminal palindrome of minute virus of mice. Gene35:179-185.

4. Bolivar, F., and K. Backman. 1979. Plasmids of E. coli as

cloningvectors.MethodsEnzymol.68:245-280.

5. Chen,E.Y.,and P. H.Seeburg. 1985.Supercoilsequencing:a

fast and simple method for sequencing plasmid DNA. DNA 4:165-170.

6. Collins, J., G. Volckaert, and P. Nevers. 1982. Precise and nearly-preciseexcision of the symmetricalinverted repeats of Tn5;commonfeatures ofrecA-independentdeletioneventsin Escherichiacoli.Gene 19:139-146.

7. Cortini, R., and N. M. Wilkie. 1978. Physical maps for HSV type 2 DNA with restriction endonucleases. J. Gen. Virol. 39:259-280.

8. Frenkel,N.1980.Defectiveinterfering herpesviruses,p.91-120. InA. J.Nahmias,W.R.Dowdle,andR. F.Schinazi(ed.),The humanherpesviruses. Elsevier, NewYork.

9. Friedmann, A., J. Shlomai, and Y. Becker. 1977. Electron microscopy of herpes simplex virus DNA molecules isolated from infected cellsby centrifugationinCsCldensitygradients. J.Gen. Virol. 34:507-522.

10. Gray,C. P. and H.C.Kaerner. 1984. Sequenceof theputative origin ofreplication in the ULregionofherpes simplex virus type 1 ANGDNA.J. Gen. Virol. 65:2109-2119.

11. Greaves,D.R.,R. K.Patient,and D. M.J. Lilley. 1985. Facile cruciform formation byan (A-T)34 sequence from aXenopus globingene.J. Mol. Biol. 185:461-478.

12. Hagan,C.E.,and G.J.Warren. 1983.Viabilityofpalindromic

DNA is restored by deletions occurring at low but variable frequency inplasmidsof Escherichiacoli. Gene 24:317-326. 13. Hayward,G.S.,andG. R.Reyes. 1983.Biochemical aspects of

transformation by herpes simplex viruses. Adv. Viral. Oncol. 3:271-306.

14. Hirsch, I., G.Cabral,M.Patterson,and N. Biswal.1977.Studies

on the intracellular replicating DNA of herpes simplexvirus type 1.Virology81:48-61.

15. Horii, Z.-I.,and A.J.Clark. 1973.Geneticanalysisof theRecF pathwaytogeneticrecombination inE. coliK12: isolation and characterization ofmutants.J. Mol. Biol. 80:327-344. 16. Leach,D. R. F., and F. W. Stahl. 1983. Viability of Aphages

carryingaperfectpalindromein the absence ofrecombination nucleases. Nature(London)305:448-451.

17. Lilley,D. M.J.1981.Invivo consequences ofplasmid topology.

Nature(London)292:380-382.

18. Luchnik, A. N., V. V. Bakayev, I. B. Zbarsky, and G.P. Georgiev.1982.Elastic torsional strain in DNA withinafraction ofSV40 mini-chromosomes: relationtotranscriptionallyactive chromatin. EMBOJ. 1:1353-1358.

19. Maniatis, T., E. F. Fritsch,andJ. Sambrook. 1982. Molecular cloning:alaboratorymanual. ColdSpringHarborLaboratory,

ColdSpring Harbor, N.Y.

20. Maxam,A.M.,and W. Gilbert.1980. Sequencingend-labelled DNAwithbasespecificchemicalcleavage. MethodsEnzymol.

65:499-560.

21. Mizuuchi, K., M. Mizuuchi, and M. Gellert. 1982. Cruciform

structuresinpalindromicDNAarefavoredbyDNA

supercoil-ing.J. Mol. Biol. 156:229-243.

22. Mocarski,E.S.,andB. Roizman.1982.Herpesvirus-dependent amplification and inversion of cell-associated viral

thymidine

kinase gene flankedbyviralAsequencesand linkedtoan

origin

of viral DNA replication. Proc. Natl. Acad. Sci. USA 79:5626-5630.

23. Murchie, M.-J., and D. J. McGeoch. 1982. DNA sequence analysisofanimmediate-earlygeneregionoftheherpes

simplex

virus type1 genome.J.Gen. Virol.62:1-15.

24. Myers,R. M.,andR.Tjian.1980.Constructionand

analysis

of simianvirus 40 originsdefectiveintumorantigenbindingand DNAreplication.Proc. Natl. Acad. Sci. USA77:6491-6495. 25. Nader,W.F.,T. D.Edlind,A.Huettermann,andH. W.Sauer.

1985.CloningofPhysarumactin sequences inan exonuclease-deficient bacterial host. Proc. Natl. Acad. Sci. USA 82: 2698-2702.

26. Panayotatos,N., and R.D. Wells. 1981.Cruciformstructuresin supercoiledDNA.Nature(London)289:466-470.

27. Rao, R. N.,and S. G. Rogers. 1979. PlasmidpKC7: a vector

on November 10, 2019 by guest

http://jvi.asm.org/

containing ten restriction endonuclease sites suitable for cloning DNA segments. Gene. 7:79-82.

28. Razin, A., H. Cedar, andA.D. Riggs (ed.).1984. DNA methyla-tion:biochemistry and biological significance, p. 1-10. Springer-Verlag, Inc., New York.

29. Sadler, J. R., J. L. Betz, and M.Tecklenburg. 1978. Cloning of chemically synthesized lactose operators. II. EcoRI-linked op-erators.Gene 3:211-232.

30. Sanger, F., S. Nicklen,and A.R.Coulson. 1977.DNA sequenc-ing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74:5463-5467.

31. Sinden, R. R., D. E. Pettijohn, and B. Francke. 1982. Organiza-tionof herpes simplex virus type 1 deoxyribonucleic acid during replication probed in living cells with 4,5',8-trimethylpsoralen. Biochemistry21:4484-d490.

32. Southern, E. 1979.Gelelectrophoresis of restriction fragments. Methods Enzymol.68:152-176.

33. Spaete, R. R., andN.Frenkel. 1982. The herpessimplex virus amplicon: anew eucrayoticdefective-virus cloning-amplifying vector.Cell 30:295-304.

34. Spaete, R.R.,andN. Frenkel. 1985. The herpessimplex virus amplicon: analyses of cis-acting replication functions. Proc. Natl.Acad. Sci. USA82:694-698.

35. Stow,N. D. 1982. Localization ofanorigin ofDNAreplication within theTRs/IRsrepeated region of the herpes simplex virus type1genome. EMBO J. 1:863-867.

36. Stow, N.D.1985.Mutagenesis ofaherpessimplex virus origin of DNAreplication and its effect on viral interference. J. Gen. Virol.66:31-42.

37. Stow, N. D., andE. C. McMonagle. 1983. Characterization of

theTRs/IRsoriginof DNA replication of herpes simplex virus type 1.Virology 130:427-438.

38. Vlazny, D. A., and N. Frenkel. 1981. Replication of herpes simplex virus DNA: localization of replication recognition sig-nals within defective virus genomes. Proc. Natl. Acad. Sci. USA 78:742-746.

39. Walboomers,J. M. M.,andJ.T.Schegget. 1976. A new method for the isolation of HSV-2DNA. Virology 74:256-258. 40. Watson,R.J.,andG.F.VandeWoude. 1982. DNAsequenceof

animmediate-early gene (IE mRNA-5) of herpes simplex virus type I.Nucleic Acids Res. 10:979-991.

41. Weller, S. K., K. J. Lee, D. J. Sabourin, and P. A. Schaffer. 1983. Geneticanalysis oftemperature-sensitive mutantswhich define the gene for the major herpes simplex virus type 1 DNA-binding protein. J. Virol.45:354-366.

42. Weller, S. K.,A.Spadaro, J.E.Schaffer,A. W.Murray,A. M. Maxam, and P.A. Schaffer. 1985. Cloning, sequencing, and functionalanalysis of oriL, a herpes simplex virus type1origin ofDNAsynthesis. Mol. Cell. Biol. 5:930-942.

43. Whitton,J. L.,andJ.B.Clements. 1984.Replication origins and asequence involved in coordinate induction of the immediate early genefamilyareconserved inanintergenic region of herpes simplex virus. Nucleic AcidsRes. 12:2061-2079.

44. Wyman,A.R.,L. B.Wolfe,and D.Botstein.1985. Propagation ofsome human DNA sequences in bacteriophage vectors re-quires mutant Escherichia coli hosts. Proc. Natl. Acad. Sci. USA 82:2880-2884.

45. Yanisch-Perron, C., J. Vieira, andJ. Messing. 1985. Improved M13 phage cloning vectors and host strains: nucleotide se-quences oftheM13mpl8andpUC19vectors.Gene 33:103-119. 58,

on November 10, 2019 by guest

http://jvi.asm.org/