"Endless" viral DNA in cells infected with channel catfish virus.

(1)

May 0022-538X/83/050405-08$02.00/0

CopyrightC1983,AmericanSocietyforMicrobiology

"Endless" Viral

DNA

in Cells Infected with Channel Catfish

Virus

JOStCEBRIAN,DANIELLE BUCCHINI,tAND PETERSHELDRICK*

Institutde RecherchesScientifiquessurle Cancer,94802Villejuif Cedex, France

Received 7September 1982/Accepted 21 January 1983

Thestate of intracellular viral DNA in cells infected with channel catfish virus

has been studied by the Hirt selective extraction procedure and by restriction

endonucleasedigestion. The sedimentation properties and restriction patternsof

viral DNAin the Hirtsupernatantfraction indicate that themajority, ifnotall,of

the DNAis in the form of linear unit-length (Mr 85 x 106) molecules. However,

restriction digests of viral DNA in the pellet fraction lacked two fragments

corresponding to the molecular ends of unit-length DNA. In addition, there

appeared in HpaI digests of pellet DNAa newrestrictionfragment interpretable

astheproduct offusionbetween theendsofunit-lengthmolecules. The size of the

new fragment requires that fusion occur in such a way that one copy of the

terminallyrepeatedsequences(Mr 12.3 x 106) of the unit-length DNA is lost in

the process. In pulse-chase experiments, radioactivity flowed from the pellet

fraction tothe supernatantfraction, suggestinga precursor-product relationship

for these DNA species. The results are easily understood if unit-length virion

DNAisgenerated by excision from concatemericstructures.

Asextracted from virions, the genome of the

herpesvirus channel catfish virus (CCV) is a

linear molecule of double-strandedDNA with a

molecular weight of approximately 85 x 106(5,

7). CCV DNA is terminally repetitive; that is,

thesamenucleotidesequence (Mr 12.3 x 106)

ispresent atboth ends of the molecule (5). The

repeated sequences are mutually oriented with

direct polarity, a configuration which permits

theformation of circlesor concatemers by

intra-molecular orintermolecular recombination,

re-spectively(4, 17-19). As circlesorconcatemers

are formed in this way from linear DNA, the

molecular endsareeither lost intheformercase,

orin thelattercase,theirfrequency relativeto a

giveninternalsequenceis reduced inproportion

tothe number of unit-length molecules making

upthe concatemer.

When linear DNA isdigested with restriction

endonucleases, the molecular ends generate a

subset offragments in which one end of the

fragment is definedby aninternalrestriction site

and the other end by the molecular end itself.

For CCV virion DNA, in which the terminal

sequences are notaltered by permutation (17) or

inversion (11), this subset contains two

frag-ments. The size and relative position of these

have been determined for the restriction

endo-nucleasesEcoRI,HindIII, HpaI, andXbaI (5).

t Present address: Institut de RecherchesenBiologie Mol6-culaire,75221Paris Cedex05, France.

In this communication, we describe

experi-ments which use the Hirt selective extraction

technique (12)tofractionate intracellular DNA

late in the replicative cycle of CCV. HpaI and

XbaI digests of CCV DNA in the supernatant

fraction contained normalamounts ofend

frag-ments, but these were notdetectable indigests

of viral DNAfrom the pellet fraction. The data

indicate the existence oftwoforms of

intracellu-lar CCV DNA, one of which corresponds to

unit-length genomes and the second to

"end-less" structures such as circles, long

conca-temers, or possibly some combination of the

two. Moreover, pulse-chase experiments

sug-gest that the endless form of CCV DNA is a

precursorof theunit-length genomes.

Results similar tothose presented here have

been previously published by Ben-Porat and

Rixon (2)forpseudorabies virus and by Jacobet

al. (13)for herpes simplex virus.

(A preliminary report ofthis work was pre-sented at the Herpes Virus Workshop,

Cam-bridge,

England, in August1978.)

MATERIALSANDMETHODS

Cells and virus.Thecontinuous catfish cell line BB

(20)wascultivatedat30°CinEagle minimum essential medium (MEM) supplemented with 5% fetal calf se-rum.Confluentmonolayers in plastic petri dishes

(60-or100-mmdiametercontaining, respectively,3 x 106 to 4 x 106cells and 10x 106to12 x 106cells)were

infected with CCVat amultiplicity of10to20PFUper cell. Virusadsorption was carried out in0.2-ml

(60-405

on November 10, 2019 by guest

http://jvi.asm.org/

(2)

406 CEBRIAN, BUCCHINI, AND SHELDRICK

mm dish) or 0.5-ml (100-mm dish) volumes of MEM containing10% (vol/vol) dimethyl sulfoxide for 20 min at22°C. Two milliliters (60-mm dish) or 5 ml (100-mm dish) of MEM without serum was then added, and the infected cells were incubated for the desired time at 22°C under an atmosphere of95%air-5%CO2.

Radiolabeling ofintracellular DNA. Cultures were labeled with [methyl-3H]thymidine (-50 Ci/mmol;

Commissariat al'Energie Atomique, Saclay, France, and NewEngland NuclearCorp., Boston, Mass.) by adding the isotope to the culture fluid toobtainafinal concentration of5 to50 ,uCi/ml, depending upon the experiment. Chaseswereperformedby removal of the radioactive medium, followed by awash with MEM

containing 2 mMthymidine and0.1mMdeoxycytidine (1),with additional incubation in the same medium.

Nonselective extraction of intracellular DNA. The extractionprocedure was derived from the method of Gross-Bellard et al. (10). Infected cell monolayers werelysed with 0.5% sodium dodecyl sulfate (SDS) in 10 mMTris-hydrochloride (pH7.5)-10 mM EDTA (1 ml per106 cells), and the lysate was incubated in the presence of 100 ,ug of proteinase K(E. MerckAG, Darmstadt,Germany) per ml for 12 h at 37°C. After a second addition ofproteinase (100,ug/ml,6hat37°C), thelysate was adjusted to 1% SDS and extracted twice with an equal volume of phenol previously equilibrat-ed with 0.5 M Tris-hydrochloride (pH 8.3)-0 mM EDTA. The aqueous phasewasdialyzed againsttwo

changes of 10 mMTris-hydrochloride(pH7.5)-l mM EDTAandstoredat4°C.

Selective extraction of intracellular DNA. Selective DNAextractionbythe methodof Hirtwasperformed aspreviouslydescribed(12), except thatproteinaseK (100 ,ug/ml) was included in the lysis buffer. The presence ofproteinasewasfound tofacilitate subse-quent resuspensionofethanol-precipitated DNA

(be-low); otherwise, no essential differences were ob-served in the velocity sedimentation properties or

buoyant densities of supernatant ofpelletDNAsif the proteinase was omitted, in agreement with Hirt's

originalobservations (12). Thelysate(1 ml of

extrac-tionbuffer per1 x 106to2 x 106cells)wasincubated

at37°Cfor1h before theaddition, withgentle mixing,

of 0.25 volumesof 5 M NaCI. After being held forat

least 2 h at 4°C, the precipitate was pelleted by centrifugationat27,000 x gfor 30 min. ForCsClor sucrose gradient analyses, the supernatant wasused

directly. For restriction endonuclease analysis, the supematant was first dialyzed at 4°C against two

changes of100volumes of10mMTris-hydrochloride

(pH7.5)- mMEDTA,adjustedto0.3 Mammonium acetate, andprecipitatedwith 2 volumesof95% etha-nol. Thecentrifugedprecipitatewasthensuspended in

asmall volume of theappropriate restriction endonu-cleasebuffer(seebelow).Thehigh-salt pelletfraction wasgentlyresuspended(usually24 h at room

tempera-turewithoutagitation) in theoriginal volume of

extrac-tion buffer(10mMTris-hydrochloride [pH 7.5]-10mM

EDTA)lacking SDS andproteinase Kand

reprecipi-tated withNaClasdescribed above. Pellet fractionsto be subjected to CsCl density gradient centrifugation

were first extracted withphenolasindicated for total

DNA.

CsCI gradient centrifugation. Solutions ofthe

mate-rialtobeanalyzedwereadjustedto adensityof1.7g/

cm3by the addition ofcrystalline CsCl.Centrifugation

wascarried out at 20°C in polyallomer tubes for 72 h at 35,000 rpm in a 50Ti (Beckman Instruments, Inc., Fullerton,Calif.) rotor, or for 20 h at 50,000 rpm in a 65 VTi (Beckman) vertical rotor. For radioactivity mea-surements,samples ofgradientfractions were deposit-ed on GF/C filters (Whatman, Inc., Clifton, N.J.) whichwere sequentially soaked in 10%trichloroacetic acid (10 min) and70%ethanol (10 min) and then dried.

Sucrose gradient centrifugation. Samples (usually 1 ml) werelayered onto 10-ml, 5 to 20% (wt/wt) linear sucrosegradients (in1 MNaCI-1 mM EDTA-10 mM sodium phosphate [pH 7.5]) and centrifuged in an SW41 (Beckman) rotor at 36,000 rpm for 150 min at 20°C. Measurements of radioactivity were performed

asdescribed above.

Restrictionendonuclease analyses. All enzymes were purchased from New England Biolabs (Beverly, Mass.).Incubationbuffers forHpaIand XbaI were as previously described (5).

DNA was digested with an excess (2 to 5 times basedontheactivity specified by the manufacturer) of enzymefor 2 hat37°C. Incubation mixtures included 100 ,ug of bovine serum albumin per ml. Reactions were terminated by adding 1/10 volume of 0.1 M EDTA-2% SDS-70o glycerol-0.2% bromophenol blue, and thedigestion products were electrophoresed

on0.5% agarose horizontal slab gels (27 by 10 by 0.4 cm)at1.5V/cmfor 15 to 20 h at ambient temperature (-22°C). The electrophoresis buffer was 40 mM Tris-hydrochloride-30 mM acetic acid-20 mM sodium ace-tate(pH7.8)-2mM EDTA. Radioactivity was detect-edbyfluorographyasdescribedpreviously by Bonner andLaskey (3),usingpresensitized Kodak RPX-omat radiographic film (14).

RESULTS

Kinetics of thymidine incorporation during CCV infection.Wolf andDarlington (20)

estimat-edthelength of the CCV infectious cycleat22°C

to beapproximately 12h. Inthepresentstudy,

wewishedto place ourexperiments at a time in

the infectious cyclenear the maximum level of

viralDNA synthesis, so wefirstdetermined the

kinetics of thymidine incorporation into CCV

DNA. Confluent BB monolayers at 22°C were

infected with 10to20 PFUof CCVpercell,and,

aftera20-minadsorptionperiod,weresubjected

hourly to 30-minpulsesof

[3H]thymidine.

Imme-diately afterthelabeling period,total DNAwas

extracted andanalyzedbyCsClgradientdensity

centrifugationas described above. The relative

proportion of label incorporated into cellular

DNA (p = 1.700

g/cm3)

and viral DNA (p =

1.717

g/cm3

[9])atvarioustimes afterinfection is shown in Fig. 1. Label associated with CCV DNA wasfirstdetectable inthepulsebeginning

at 2h postinfection (p.i.) and became maximal

withintheinterval defined bythe 3 h-p.i. and 5

h-p.i. pulses. In the experiments described

be-low, we thereforeused 30-min labeling periods

beginning at 3.5 h p.i. Dixon and Farber (7)

obtained kinetics for CCV DNA synthesis at

30°C whicharesimilar tothosereported here.

J. VIROL.

http://jvi.asm.org/

(3)

407

!

0

0 1 2 3 4 5 6 7

HOURS POSI-INFECTION

FIG. 1. Kinetics of [3H]thymidine incorporation into CCV-infected cellular DNA. Infected cellular

DNA was extracted and analyzed by CsCI density gradient centrifugation as described inthe text.

Re-sults areexpressed as the viral (A) and cellular(0)

componentsof the total(A) incorporation. The

maxi-mumtotal incorporationwasassignedavalue of100o. The proportion ofcountsinviral and cellular DNAs

wasestimated fromdensity gradient profilessimilarto

those of Fig. 3.

Fractionation of 3H-labeled DNA by Hirt

ex-traction. Hirt(12) developedaselective

precipi-tation procedure permitting the separation of

low-molecular-weight polyoma virus DNA (Mr

3 x 106) from high-molecular-weight cellular

DNA(Mr > 100 x 106). Since itsintroduction,

the technique has been extended to studies of

viruses with genomes of considerably higher

molecular weight than that of polyoma virus,

notably pseudorabies virus(Mr 90 x 106[1])

andherpes simplexvirus(Mr 100x 106[15]).

Inapplyingthe HirtproceduretoCCV-infected

cells, we observed that successive extractions

continued to liberate labeled material into the

supernatantfractionindecreasingamounts(Fig.

2). Of the total radioactivity recovered in the

supernatants of four serial extractions, -50%

wasreleased in thefirst extraction, -25%in the

second, -10%o in the third, and -5% in the

fourth; furtherextractions failed torelease

sig-nificantamountsofradioactivity fromthepellet.

In a representative experiment, -30%o of the

total label incorporated in a 30-min pulse was

found in the combined supematants,and -70%o

remainedpelletassociated. When pulse-labeled

uninfected BB cells were extracted in thesame

way(datanotshown), >95% oftheincorporated

radioactivity remained associated with the pel-let.

Toinvestigate the partitioning of radioactivity

between viral and cellular DNAs, successive

supematants andthe exhaustively extracted

pel-let were centrifuged to equilibrium inCsCl

gradi-ents. Roughly90% of the radioactivity in each

supernatant banded at the density of viral DNA (Fig. 3A), whereas about 60% of the pellet radioactivity had the density of viral DNA (Fig.

3B). Thesupernatants and pellet were also

char-acterized by velocity sedimentation in 5 to 20%

sucrose gradients (Fig. 4). Sedimentation

pro-files ofradioactivity in successive supernatant

fractions were typically heterodisperse (Fig. 4A

and B) and covered a range of sedimentation

coefficients from -50S to -20S. Unit-length

CCVDNAextractedfrom virions sediments at

-53S(P.Sheldrick, N. Berthelot, and S.

Chous-terman,unpublished data). Similar experiments

with the final pellet fraction (not shown)

re-vealed that 80o of the radioactivity sedimentd

to the bottom of the gradient.

We next examined the behavior of

pulse-labeledsupernatantand pellet DNA during a 4-h

chase with excess nonradioactive thymidine.

The results (Fig. 5) show that as the chase

progressed, radioactivity was lost from the

pel-let andaccumulated with similarkinetics in the

supematant. In controlexperiments not

present-ed,the presenceof excess thymidine from 3.5 h

p.i. on did not measurably affect the yield of

infectious virus from afull growth cycle

(mea-suredat 24hp.i.). Velocitysedimentation

analy-At5

0

1 2 3 4

EXTRACTION

FIG. 2. Successive Hirt extractions of CCV-infect-ed cells. Infected cellswere labeled with

[3H]thymi-dine for 30 minat3.5hp.i.andextractedasdescribed in thetext.Resultsareexpressedasthe percentageof totalradioactivity extracted.

VOL.46, 1983

http://jvi.asm.org/

[image:3.490.44.230.58.290.2] [image:3.490.269.414.435.626.2]

(4)

I?

I'½

0t 20 40 60 SO 0 40 60 S0

FRACTION FRACTION

FIG. 3. Buoyantdensity gradientsofsupernatant andpelletDNAs.(A) Combined supernatant fractions from theexperiment shown inFig.2.(B) Pellet fraction after four successive extractions (Fig. 2). As a buoyant density reference, [14C]thymidine-labeledCCV virion DNA (arrows)wasadded to the gradients. The gradient bottoms are totheleft.

04

V

M.

I

FRACTION FRACTION

FIG. 4. Velocity sedimentation ofsupernatantDNA.Sucrosegradients (5to20%)of the first (A) and second (B) successivesupernatantfractions shown inFig.2. Arrows indicate thepositionof'4C-labeledCCV virion DNAaddedtothegradients.Sedimentation wasfromrighttoleft.

J. VIROL.

http://jvi.asm.org/

[image:4.490.60.444.54.333.2] [image:4.490.60.445.371.640.2]

(5)

INTRACELLULAR

t:

0

50~~~~~~~

0 ,

0 t 2 3 4

CHASE PERIOD (HOURS)

FIG. 5. Partitioning ofradioactive CCVDNA be-tweensupernatantandpelletfractionsduringa pulse-chase. Infected cells were pulse-labeled as in the

legendtoFig. 2 andincubated fortheindicated times in the presence of excess thymidine (see text). Radio-activity in

CCV

DNA was determined by density

gradientcentrifugation(as inFig. 1). Values for each

timepoint were obtainedfrom four (combined)

super-natantfractions(0)andfrom the final pelletfraction

(0).

sisof the supernatantfraction preparedat

vari-oustimes duringthe chase(Fig. 6) showed that

the major part, if not all, ofthe accumulating radioactivity sedimented at -53S, the position

ofunit-length CCV DNA.

Restriction endonuclease analysis of

superna-tant, peUlet, and nonselectively extracted DNAs.

In restriction endonuclease digests of linear

DNA, the ensemble of fragments contains a

subset produced on the one hand by internal

restriction sites and on the other hand by the

molecular ends of the DNA. For unit-length

CCV DNA, two fragments appeared in this

subset. The present experiments were carried

outwithHpaI, whose recognition sites lie

out-side the terminally repeated sequences, and with XbaI, one of whose recognition sites lie within

these sequences (see Fig. 9 and reference5).

CsCl gradient-purified CCV DNA from the

supernatant and pellet fractions was digested

with each of these enzymes, and the resulting

fragmentswereresolved by agarose gel

electro-phoresis (see above). The fragment proffies of

restricted supernatant DNA are

indistinguish-able from those of CCV virion DNA digested

with the respective enzymes (Fig. 7A through

D), supporting our conclusion that this fraction

contains principally unit-length viral genomes.

Withpellet-fractionDNA,however, quite

differ-ent results were obtained. The profile of XbaI

digests lacked band Faltogether, and, as

esti-mated from densitometer tracings of

photo-graphic negatives, theintensityof the 2x molar

band (D, E) relative to bands, C, G, and H

diminishedtotheequivalentof -1.1 molar(Fig.

7 G and H). Similarly, in gelsofHpaI digests,

bands B andC wereno longer detectable (Fig.

7Eand F).

Visual inspection of the gels permitted us to

FRACTION

FIG. 6. Velocity sedimentation ofsupernatantDNAduringapulse-chase. Thefirstsupernatantfractions of a

30-minpulseat3.5 hp.i. (A) and after chases of 1, 3, and 4 h (B, C, and D, respectively) were sedimented in S to

20osucrosegradients. Arrows and direction of sedimentation are as in Fig. 4.

46, 409

http://jvi.asm.org/

[image:5.490.60.213.53.253.2] [image:5.490.93.390.439.642.2]

(6)

A B C D

Bw -:

B"

D"....Cwe&..

E _w

E

FF

F*%M

E :F

A_ L

E E

F'

H

G G

J

10-fold lower than in virion

(or supernatant)

DNA. A similar estimate can be obtained by

comparing

HpaI

fragments

BandC

(Mr

= 17 x

B 106 and 13.2 x 106, respectively) with HpaI

fragment

G(Mr = 1.5 x

106).

Inadditiontothe

above changes, a new band (B') appeared in

profiles of HpaI digests for pellet DNA (Fig.

7F). The molecular weight

(Mr

18 x 106) of

this new

fragment

is

precisely

that

expected

E from a fusion (via overlap of the terminal re-F- peats) of theHpaIfragments B and C of virion

DNA (seeFig. 9). Thiscanbecalculated by the

followingequation: fragment B (17 x 106)plus

fragment C (13.2 x 106) minus the terminal

G repeatsequence

(12.3

x

106) equals

17.9 x

106.

For another project,we have recentlycloned

fragmentB' in the HpaI site of cosmid pHC 79.

Restriction endonuclease site maps(M. Lacasa

H and J. Cdbrian, unpublished data) agree

com-pletely with the presentassignment.

Thefusion fragment HpaI B' is more

conve-niently observed in digests of intracellularDNA

A B c D E F

[image:6.490.56.240.62.377.2]

J

FIG. 7. Restriction endonuclease digests of viral DNAfrom a Hirt extraction of CCV-infected cells.

Infectedcellswerepulse-labeled (30 min)at3.5 hp.i.,

and viral DNAwasselectivelyextracted andpurified byCsCldensitygradientcentrifugation. Supematant

(lanes B andD) andpellet (lanesF andH)DNAswere digestedwithHpaI(lanesBandF)andXbaI(lanesD andH). Digestsof14C-labeledCCV virion DNA with therespectiveenzymes(HpaI, lanes AandE;XbaI,

lanes C and G) were included for reference. Condi-tions forelectrophoresisandautoradiographyare giv-enin thetext.

roughly estimate an upperlimit for the relative

concentration of the missing fragments. For

example, inthe XbaIprofile,fragment F, witha

molecularweight of5.6 x

106,

was not

detect-able under theconditions of filmexposureused, whereas fragmentJ, at 0.65 x 106, was clearly

visible (Fig. 7H). Since the molecular weights

(and supposedly the amount of[3H]thymidine)

ofthese fragments differby a factor of10, the

concentration ofXbaIfragment F must be less

than1/10thatofXbaIfragment J; otherwise,the

bandwould bevisible.Moreover,invirionDNA one copy of XbaI fragment J is present per genome, which meansthattheconcentrationof

ends (strictly speaking, one end for XbaI) per

genome equivalent in pellet DNA is morethan

mow,...

C _ .. ..

[image:6.490.279.423.315.595.2]

D t IIiw;

FIG. 8. Restriction endonuclease digests of total DNA from nonselective extraction ofCCV-infected cells. HpaIdigests of DNA from a nonselective

ex-traction (see text) of infected cells pulse-labeled (30

min) at 3.5 h p.i. (lane B) and chased with excess

thymidine for 1 (lane C),2(lane D),3(lane E), and4h

(lane F). HpaI-digested, "4C-labeledvirion DNAwas

includedas amarker(lane A).

J. VIROL.

http://jvi.asm.org/

(7)

INTRACELLULAR CHANNEL CATFISH VIRUS DNA

VIRION DNA

TR TR

C B

.~~~, . ^

D E A F

a a apA I.3

Hpa I

XbaI

INTRACELLULAR DNA

B B

A E A E

A

p . a a Ip ..a _

HpaI

Xba I

FIG. 9. Restriction endonuclease site map of virion and intracellular CCV DNAs. Thick lines represent the M =12.3 x 10 terminal repeat sequences referred to in the text. Only the restriction fragments relevant tothe

presentstudy arelettered.

from a nonselective extraction (see above),

since it andfragmentHpaI B form adoublet in

gel profiles (Fig.8B). Achase with

nonradioac-tive thymidine was also incorporated into the

experiment shown in Fig. 8, and, asthe figure

shows, thegelprofilesevolvedduringthecourse

of the chase. Ina30-minpulse,only band B' (as

in pellet DNA) was apparent, but asthe chase

proceeded its intensity decreased whereas the

intensities of bands B and C increased. These

results areinagreement with theprevious

sug-gestion that pellet DNA isaprecursor to

super-natantDNA.

Certain bands other than thosejustdiscussed

werealso found to have modifiedintensities in

gel profilesofpelletandtotal DNA.Thesewere

XbaI band A(Mr= 29 x

106)

andHpaIband A

(Mr = 31 x 106), both of which are clearly

underrepresentedrelativetovirion and

superna-tantDNAs in the gels of Fig. 7and 8. We have

investigatedthismoreclosely withBglII, which

gives smaller fragments (Mr < 15 x 106;

S. Chousterman andM. Lacasa, personal

com-munication) in thegenomic regions covered by

XbaIband A andHpaIbandA. Inthiscase(not

shown), no reduction in the intensities of the

corresponding bands was observed. We

there-fore think it likely that this effect is due to

nonspecific breakage of intracellularDNA

dur-ing the isolation procedures used (see above).

Largerfragments would be affectedtoa greater

extentby such breakage.

DISCUSSION

By applying the selective extraction

proce-dureof Hirt (12)toCCV-infectedcells, we have

obtained evidence fortheexistence oftwo

class-es of intracellular viral DNA which differ in

several important respects. DNA released into

thesupernatantfraction sedimentedatthesame

rate (-53S) as did DNA extracted from CCV

virions (Fig. 4). This result and the fact that

restriction endonuclease digest profiles (Fig. 7)

were indistinguishable from profiles of virion

DNAindicate that supernatant DNA consists of

unit-lengthCCVgenomes.

Pellet DNA, on the other hand, besides the

fact that it doesnot enterthesupernatant under

theextractionconditions,differed from

supema-tantDNAin that itsedimentedtothebottom of

gradients in which virion DNA sedimented

ap-proximately halfway. Inaddition, theXbaI and HpaIfragmentsthatarediagnostic of the

molec-ular ends of theunit-length CCV genome were

not detectable (<1/10 the expected

concentra-tion) indigests of pelletDNA (Fig. 7). By this

lattercriterion,therefore, pelletDNAis endless.

Thepresenceofanew"fusion"fragment(B')

which replaced the missing end fragments in

HpaIdigestsofpelletDNA(Fig. 7) isimportant for two reasons. First, it effectively eliminates the possibility that end fragments are missing

from pellet DNA because they areattached to,

forinstance,aproteinwhich couldprevent them

from entering the gel. This phenomenon has

beenfoundtoapplyunder certainconditionsto

the end fragments of the adenovirus genome

(16). Second,thesize ofthefusionfragment(Mr

= 18 x 106) requires that the fusion leading to

loss ofthe molecular ends takeplacethrough a mechanism in which one copy per genome

equivalent ofthe terminal repetition of12.3 x

106 is also lost, rather than by a "blunt end"

fusion in which case a new fragment with a

VOL.46,1983 41

http://jvi.asm.org/

[image:7.490.59.421.52.252.2]

(8)

molecular weight of 30 x 106(HpaI fragment B

plus HpaI fragment C) would be formed. In

XbaI digests of pellet DNA (Fig. 7), nofusion

fragment is expected, because arestriction site

lieswithin the terminalrepeat(Fig. 9).

Fusion with loss of a copy of the terminal

repetition couldoccurbygeneticrecombination

betweenrepeats (4,17, 18)orby exonucleolytic

removal of complementary strands from two

repeats with subsequent annealing of the

un-paired strandsto regenerate asinglerepeat(e.g.,

reference 19). Whatever theprecise mechanism,

fusion of intragenomic repeats would lead to

unit-length (less oneterminalrepeat) circle

for-mation whereas concatemeric structures (in-cluding multimeric circles) would result from

intergenomic fusion. Subsequent encapsidation

of unitlengthgenomescould thenproceedviaa

replicative mechanismof the typeproposedfor

bacteriophage T7(19) whereby the terminal re-peats are "restored" tovirion DNA.

Either unit-length circles or concatemers

could accountforourresults. However, unless

heldin placebyaproteinase K-resistantlinkage,

unit-length circles would not be expected to

remain in thepelletfraction. Norhave webeen

ableto detectcircular molecules by direct

elec-tronmicroscopicexamination of CsCl

gradient-purified pellet DNA (N. Berthelot, personal

communication). Admittedly, theseare not

ab-solutely irrefutable argumentsbecause of

possi-bletrappingeffects inhigh-molecular-weight

cel-lularDNA, buttheydo leadustofavor the idea

thatpellet DNA consists principallyof

concate-meric forms (or"rollingcircle" forms[8],which

wouldnotbedistinguishable fromconcatemers

in thepresentexperiments)of theCCVgenome.

Similarconclusions derived from restriction

en-zymeanalysishavebeenreached for the

replica-tiveforms ofDNAinpseudorabiesvirus(2) and

herpes simplexvirus (13) infections and for the

intracellular DNA ofbacteriophage SPOt (6).

ACKNOWLEDGMENTS

We thank Madeleine Laithier forprovidingcell cultures and virus stocks.

This workwassupportedby generalgrants fromtheCentre Nationale de la RechercheScientifique and a specialgrant ATPCNRS 655-2108.J.C.was aFellowof theLigue Nation-aleFrancaisecontrele Cancer.

LITERATURE CITED

1. Ben-Porat, T., A. S.Kaplan, B. Stehn, and A. S. Ruben-stein. 1976.Concatemeric forms of intracellular herpesvi-rusDNA.Virology69:547-560.

2. Ben-Porat, T., and F. J. Rlxon. 1979. Replication of herpesvirus DNA. IV. Analysis ofconcatemers. Virology 94:61-70.

3. Bonner,W.M., and R.A.Laskey. 1974. Afilmdetection method for tritium-labelled proteins and nucleic acids in polyacrylamide gels. Eur. J. Biochem. 46:83-88. 4. Campbell,A.1962. Theepisomes. Adv. Genet.

11:101-145.

5. Chousterman, S., M. Lacasa, and P. Sheldrick. 1979. Physical map of the channel catfish virus genome: loca-tion of sites forrestriction endonucleases EcoRI, HindIII, HpaI, and XbaI. J. Virol. 31:73-85.

6. Cregg, J. M., and C. R.Stewart. 1978.Terminal redun-dancy of"highfrequency of recombination" markers of Bacillus subtilis phage SPOl. Virology 86:530-541. 7. Dixon, R. A. F., and F. E. Farber. 1980. Channelcatfish

virus:physicochemical properties of the viral genome and identification of viralpolypeptides. Virology 103:267-278. 8.Gilbert, W., and D. Dressler. 1968. DNAreplication: the rolling circle model. Cold Spring Harbor Symp. Quant. Biol.33:473-484.

9.Goodheart, C. R., and G. Plummer. 1975. Thedensities of herpesviral DNAs.Prog. Med. Virol. 19:324-352. 10. Gross-Bellard, M., P. Oudet, and P. Chambon. 1973.

Isolationofhigh molecular weight DNA from mammalian cells. Eur. J. Biochem. 36:32-38.

11. Hayward, G. S., R. J. Jacob, S. C. Wadsworth, and B. Rolzman. 1975. Anatomy of herpes simplex virus DNA: evidence for four populations of molecules that differ in therelative orientationsof their long and short segments.Proc.Natl. Acad. Sci. U.S.A. 72:4242-4247. 12. Hirt, B.1967.Selective extraction of polyoma DNA from

infected mouse cell cultures. J. Mol. Biol. 26:365-369. 13. Jacob, R. J., L. S. Morse,andB.Rolzman. 1979. Anatomy

ofherpessimplex virus DNA. XII. Accumulation of head-to-tail concatemers in nuclei of infected cells and their role in thegeneration of the fourisomeric arrangements of viral DNA. J. Virol. 29:448-457.

14. Laskey, R. A., and A. Mills. 1975. Quantitative film detectionof 3H and14Cinpolyacrylamidegels by fluorog-raphy.Eur. J. Biochem. 56:335-341.

15. Pater, M. M., R. W. Hyman, and F. Rapp. 1976. Isolation ofherpes simplex virus DNA from the "Hirt superna-tant."Virology75:481-483.

16. Sharp, P. A., C. Moore, andJ. L.Haverty. 1976. The infectivityof adenovirus 5DNA-protein complex. Virolo-gy 75:442-456.

17. Streisinger,G., J. Emrich, and M. M.Stahl.1967. Chro-mosomestructureinphageT4. III. Terminalredundancy andlength determination. Proc. Natl. Acad. Sci. U.S.A. 57:292-295.

18. Thomas, C.A. 1966.Recombination ofDNA molecules. Prog. Nucleic Acid Res. Mol. Biol.5:315-349. 19. Watson, J. D. 1972. Origin of concatemeric 17 DNA.

Nature(London)NewBiol. 239:197-201.

20. Wolf, K., andR. W. Darlington. 1971. Channel catfish virus: a newherpesvirus of ictalurid fish. J. Virol. 8:525-533.

J.VIROL.

http://jvi.asm.org/