Replication process of the parvovirus H-1. VI. Characterization of a replication terminus of H-1 replicative-form DNA.

Copyright © 1977 American Society for Microbiology Printed in U.S.A.

Replication Process of the Parvovirus

H-1.

VI.

Characterization of a Replication Terminus of H-1

Replicative-Form DNA

SOLON L. RHODE III

PutnamMemorial Hospital Institute for Medical Research, Bennington, Vermont 05201

Receivedfor publication 11June1976

Thelinearduplexreplicative form (RF) DNA of the parvovirus H-1 has been

characterized withrespect tocleavageby thebacterial restriction endonuclease

of Escherichia coli, EcoRI. RF DNA has a single cleavage site 0.22 genome

length fromthe left end ofthemolecule. Themolecular weight ofH-1RF DNA determined by gel electrophoresis is 3.26 x 106. H-1RF DNA has beenfoundto dimerize by hydrogen-bonded linkage at the molecular left end, and in some molecules the viral strand is covalently linked to the complementary strand.

Some 10%ofmonomeric RFDNA also hasacovalent linkage betweenthe viral

and complementary strandsatthe left end. TheEcoRI-B fragment, containing the left end of the RF molecule, appears to be a replication terminus by its labelingcharacteristics forboth RFand progeny DNA synthesis. These findings

suggest that the left end of H-1 RF DNA has some type of "turn-around"

structureandthat this endisnotan origin forDNAsynthesis.

The parvovirus H-1 contains a genome of

Nevertheless,

itis clear that the

genetic

capac-single-stranded

DNAwithamolecular

weight

ity

of H-1 is very

limited,

andthus its

replica-of

approximately

1.7 x 106(15). We have

previ-

tion must be executed

largely by

cellular

en-ously characterized

a double-stranded DNA zymes. Anumber of conditionallethal mutants

replicative form (RF)

synthesized

during

the ofH-1have been isolated andnoneofthese has replication ofH-i-infected cells (17, 18). H-1RF shownaviralprotein

requirement

for RFDNA DNAhas been foundtobe linearby

isopycnic

replication (19, and unpublished data). It is centrifugation in ethidium bromide-CsCl gra- possible, but not established, that RF

replica-dients, velocity

sedimentation(17, 18),

gel

elec- tionis

totally

under the control of cellular pro-trophoresis (in this paper), and

by

electronmi- teins. Inany case,analysis ofRF

replication

is

croscopy (23). a powerful probe of cellular DNA replication

In this study, H-1 RF DNA and its dimer and could be usefulinthestudy of the mecha-have been characterized further. An

improved

nismsofreplication of linearmolecules.

estimateofH-1 RF DNAmolecular

weight,

3.26 Ihave examined thephysicalmapof H-1RF

x 106, has beenobtained by gelelectrophoresis. DNA with respect to the bacterial restriction

Basedon a DNAsize of about 4,920 basepairs endonuclease of Escherichia coli, endo

(bp),

the H-1 genome has a total

capacity

to R-EcoRI (EcoRI). This enzyme has alloweda

code forproteins with acumulative molecular moredetailed characterization of one end of the

weightof about 172,000. Todate, only thetwo RF DNA molecule, and evidence for several polypeptides of the H-1

capsid,

VP1 and

VP2,

structural alternatives for this end are

pre-have been demonstrated as viral-induced pro- sented. Experiments described here indicate

teins ininfectedcells, andtheir total molecular that thisend is a terminus for the

semiconser-weightof 164,000 is close tothecoding capacity vative replication of RF DNA and for progeny

of the genome (14). It hasnotbeendetermined DNAsynthesis. This terminus is in some cases

whether the H-1 genome is transcribed to the self-cohesive, as manifestedby the formation of

extentof the otherparvoviruses (50to 100%of dimer molecules linked in aninvertedmanner

the rat virus viral strand is transcribed

[20],

ratherthan in tandem. The viral (V) and

com-and 75% of the adenovirus-associated virus plementary (C) strands in some monomer or

[AAV]

genome istranscribed

[3-5]),

orwhether dimer RF molecules are covalently linked at

VP1 andVP2 havea commonstructural gene. this end. These findings suggest a structure for

694

on November 10, 2019 by guest

http://jvi.asm.org/

the RF DNA replication terminus similar to (Haemophilus aegyptius, kindly provided by David one recently proposed for the AAV DNA ter- T. Denhardt)werecarriedoutfor 16 hat37°Cin50

mini (8, 24). ml of 10 mM Tris (pH 7.5)-10 mM NaCI-10 mM

MgCl2-lmMdithiothreitol and with5,lI of enzyme.

MATERIALS AND METHODS Reactions were stopped by addition of EDTA to a Virus and cell strains. Parasynchronous cultures finalconcentration of 50 mM.

ofsecondary hamster embryoor NBcellswerepre- Analytical gel electrophoresis. Electrophoresis pared and infectedaspreviously described (14, 16). wasconducted in cylindrical gels (0.6 by 15 cm) with TheH-1 strains used here were wild type (wt) and various concentrations of agaroseor 3% acrylamide-ts 1 (19). AnE.coli (Wi100)lysogen ofXcI857 susS 0.5% agarose in E buffer (40 mM Tris-20 mM Na wasobtained from Herbert Boyer, San Francisco. acetate-1 mM EDTA) (9) at 23°C under constant Infection, labeling, and viral DNA extraction. voltage as specified in the text. DNA samples were Parasynchronous hamster embryo cultures were in- mixedwith 0.5 volume of a solution of E buffer, 50% fected at12h poststimulation, and NB cell cultures glycerol, and 0.01% bromophenol blue for layering wereinfected at the time of reversal of the metho- on thegel. Gels were sliced into 1.0-mm or 1.17-mm trexate block with medium containing thymidine slices,and two slices pervial were incubated2to16

(TdR; 10-5M), adenosine (13 ,ug/ml), and minimal h in 0.5 ml of NCS solubilizer (Amersham/Searle) essential medium nonessential amino acids as a and assayed for radioactivity by liquid scintillation

source of glycine (19). Virus adsorptions were per- spectrometry in 10 ml of a toluene-based scintilla-formed for 30 min at33°C. After virus adsorption, tionfluid. Ninety percent or more of the activity is the cultures were incubated at the temperatures eluted from the gel slices by this method.

indicated. Viral DNA was labeled by changing the Preparative gel electrophoresis. Viral DNAwas mediumafter two washes with Hanks balanced salts fractionated in cylindrical gels (1.2 by 6 cm) of 1% (5 ml per petri dish) to aprewarmed medium con- agaroseinEbuffer using a constant voltage of 150 taining the appropriatelabeled precursor. In most V. DNA infractions of 2 ml was collected every 10 cases, labeled TdR was usedinthe presence of 5'- minby continuous electroelution into a chamber of fluorodeoxyuridine (FUdR) at 0.5

Ag/ml

after en- ourdesign.Aliquotsweretaken fordetermination of dogenous pools of TdR had been depleted by a prein- radioactivity, and appropriate fractions were col-cubation in medium with FUdR alone. Labeling lectedand precipitated with0.1 volumeof3 M Na

with32po4wasdoneinmedium buffered by 25mM acetate(pH 5.0)and2.5volumes ofethanol for16or

HEPES (N-2-hydroxyethylpiperazine-N'-2-ethane- more h at -20°C.

sulfonicacid)and with the appropriate Cl salts sub- Slab-gel electrophoresis. Slab-gelelectrophoresis stituted for PO4inthe medium. wascarriedoutingels (0.3 by12by16cm) of 1.8% Viral DNA was extracted by amodification of the agarose in an EC470 apparatus (E-C Apparatus

Hirt procedure (12). The monolayer cultures were Corp., Philadelphia, Pa.). The gel and electrode washed once with 5 ml ofTris-buffered saline per buffers were buffer E, and electrophoresiswasfor18 100-mmdish and lysed with 0.5 ml of0.6%sodium hat 16°C witha constant voltage of50 V. Thegel dodecyl sulfatein 50 mMTris (pH 7.5)-10 mM EDTA wasdriedinvacuoand exposed toKodakno-screen

perdish. After15to 20minat roomtemperature,0.5 X-rayfilmNS2T for4daysbeforedevelopment.

mlof asolution containing 5 mg of Pronase per ml Velocity centrifugation. Neutral sucrose

gra-was added to each dish, and the incubation was dientswere 5 to20% sucrose in 50 mMTris(pH 8.0)-continuedfor30minat37°C. A solution of5MNaCl 1MNaCl-1 mM EDTA-0.2% Sarkosyl centrifugedin

was then added dropwise to each dish to a final anSW25rotorfor18 to 22hat24,000 rpm.Alkaline concentration of 1 M and gently mixed with the sucrose gradients were 5 to 20% sucrose in 0.2 N lysate. The lysates were collected into centrifuge NaOH-0.8 M NaCl-1 mM EDTA centrifuged inan tubes and incubatedovernight at0°C.After centrif- SW50 rotorfor 16hat 30,000 rpm. Fractions were

ugation at 25,000 xgfor 50 min, the supernatant, collectedthrough the bottoms of the tubes. Aliquots containing viral DNA, was precipitated with 2.5 weretaken fordetermination ofradioactivity either volumes of ethanol for 6 to 16 h at -20°C. The by solubilization directlyin 10 volumes of NCSor precipitateswereredissolvedin 50 mMTris (pH 7.5)- after precipitation on cellulose acetate filters with

1 mM EDTA anddigested with pancreatic RNase 10%trichloroaceticacid and100,ug of bovine serum (heatedto80°C for20min) at 50 ug/ml. Theywere albumin, followed by liquid scintillation spectrome-theneither extracted with phenolordirectly frac- try. Alkalinegradientswerefractionatedon 25-mm

tionatedby centrifugation in 5to 20% neutral su- Whatman no. 3 filter paper disks, which were

crose gradients as detailed previously (17). Gener- washed twice in cold 5% trichloroacetic acid and ally,abouttwo-thirdsofthe totalisotopic incorpora- once in acetone, dried, andcounted with 0.2 ml of tion was recovered asviralDNA in the Hirtsuper- NCS solubilizerandtoluene scintillationfluid. natant. [3H]TdR-labeled X DNA was prepared as Isopycnic centrifugation. DNA was banded to described (11). equilibrium inCs2SO4 as previouslydescribed (17, Digestion with restriction endonuclease. Viral 19),andaliquotsweremeasured forradioactivityas DNAdissolved in 50to 100 ,ul of100 mM Tris (pH for alkalinesucrosegradients.

7.5)-50mMNaCl-10 mMMgCl2wasincubated with Benzoylated DEAE-cellulose (BDC) chromatog-10 to chromatog-100Uof EcoRI for2or 3h at37°C. Endonucle- raphy. BDC was purchased from Schwarz/Mann

ase R*EcoRI was purchased from Miles Laborato- (Orangeburg, N.Y.).Columns(1by5cm)were con-ries, Inc. Digestions with endonuclease R-HaeIII structedand washedwith 100to200volumesof0.3

on November 10, 2019 by guest

http://jvi.asm.org/

M NaCI-50 mM Tris (pH 7.5)-l mM EDTA (0.3 M above. Final recoveries were 80 to 100% of theDNA

TBS). DNA in 2 to 5 mlof0.3 MTBSwasappliedto applied to the columns. the columns with aflowrateof1 ml/min, and the

column waswashed with15-, 30-,and 30-ml volumes RESULTS of0.3 MTBS. Double-stranded DNA waselutedby

two 10-ml applications of 1 MTBS, and DNA with Endo R-EcoRI digestion of H-1 RF DNA.

single-stranded regions was similarly eluted with H-1 RF DNA labeled with [3H]TdR was

pre-two 10-ml applications of 2% caffeine in 1 M TBS pared from

tsl H-1-infected

parasynchronous

warmed to40°C.Insome cases,the caffeine elution NB cells incubated at the restrictive

tempera-was carriedout with 20 ml of a linear gradient of 0 to ture of39.5°C

(17).

A portion of the major peak

2%caffeinein 1MTBS. Thewashesand aliquots of

thel M TBS and 2% caffeine fractions were assayed at

i6S

in the preparative neutral sucrose

gra-for radioactivity bytrichloroacetic acidprecipitation dient was digested with EcoRI for 2 h at 37°C

as above. DNA in the eluates was precipitated for and a control incubated in the absence of the

further studies by 2.5 volumes of ethanol overnight enzyme, and thedigestsweresubjected to

elec-at -200C. trophoresis in replicate agarose gels; the results

Alkaline denaturation-BDCchromatography was are illustrated in Fig. 1.EcoRI-cleavedH-1RF

done by incubating the DNA in a final concentration DNA into a large

fragment,

RI-A, and one or

of 0.2 N NaOHina volumeof 200

Al

for 20 min at more smaller fragments. This digestion did not

ambient temperature; the solution was cooled to

0C, neutralized, and diluted 25-fold with cold0.2 M go to completion, but identical results were

ob-TBS containing 0.2 ml of 0.2 N HCl. To minimize tamned with complete digests using RF DNA

intermolecular annealing, the DNA solution was prepared with either tsi or wt H-i and with

immediately applied to the column of BDC within1 electrophoresis in3% acrylamide-0.5% agarose

to 2 min.The elutions were carried out as described gels. The EcoRI-Afragment containedan

aver-* I

I

I

I1

I

|

50-

RF

40-EcoRIA

30 -0

E

20 -i

Ig~~~~~~~~~~cR

It~~~~~~~~~~~~~

B2

Fraction

FIG. 1. Analyticalgel electrophoresis ofH-1RF DNA andEcoRI-digestedH-1 RF DNA. H-1 RF DNA

labeledwithP3H]TdRwasisolatedfromhamsterembryocellsinfectedwithwtH-1at37°C.AportionofRF DNAthatsedimentedat16.5S(cf. Fig.3,fractionA)wasdigestedin 50piofdigestion buffer for20 minat 37°Cwith 5 ,ulofEcoRI,andacontrol wasincubatedsimilarlywithoutenzyme. The reaction wasstopped

with 20 ,ulof2.5% sodiumdodecyl sulfate-50%glycerol-0.01% bromophenol blue-10 mMEDTA,and the

samplesweresubjectedtoelectrophoresis incylindrical gels of1.4% agarose.Inthiscasetheelectrophoresis

bufferwas90mMTris-borate(pH8.3)-2.5mMdisodiumEDTA,andelectrophoresiswasfor20hat35Vat

roomtemperature.Intheillustration, theresultsoftheseparategelsarecombinedtosimplify comparison.

on November 10, 2019 by guest

http://jvi.asm.org/

[image:3.501.106.421.298.583.2]

age of 78.4% of the [3H]TdRor32P when

32PO4

wasused as alabel, and the smaller fragments A.Haem

contained the remaining 21.6%. The dispersed 4 electrophoretic pattern of the smaller

frag-ments suggested several possible interpreta- 3

tions: (i) threeEcoRI cleavage sites, producing

one large fragment migrating as RI-A, one as A B C

RI-B2, andtwo asRI-B1 aslabeled in Fig. 1; (ii) 2 D E

twoEcoRI

cleavage

sites,

onewithin

EcoRI-B%,

F

yielding

asmall

fragment

lostfrom the

gel

and 1 dme

LAAALGL1

converting

RI-B,

to RI-B2; and (iii) one EcoRI A

cleavage

site,

with structural

heterogeneity

of\ x

the EcoRI-B fragment giving heterogeneous E

migration inthe gelelectrophoresis. u B.

EcO

RI

The firstalternativewas ruled out by deter- .

A'

mining the size of the EcoRI-B fragments. EcoRIA

Treatment of H-1RF DNA withthe restriction

endonuclease endo RNHaeIII of H. aegyptius .* ' A.^

produces

12unique fragments,designatedA to 10 103 -Eco RIB

L, ranging in size from about 1,550 to 30 bp as

B.

.e

determinedbytheir contents of 32p as shown in . S s H

Fig. 2A. In areplicate gel (Fig. 2B), theEcoRI-H.* F

Bfragmentswere notresolved and migrated as

onebroadpeak at a sizeposition ofabout 1,050 Ha H

bp,

or 21% of the genome, in

agreement

with 5

_0

a 1

the proportionofcountsrecoveredin all of the B Relative

Mobility

fragments (Fig. 2G). Thus, the B fragments B

represent about 22% of the genome by size and composition and must be aheterogeneous

col-lectionof the samebasicfragment.

Theobserved8%differenceinrelative mobil- o 3 ity of

EcoHI-B,

and EcoRI-B2 with respect to _ Fraction

EcoRI-A in 3%acrylamidegelssuggests that a FIG. 2. Analytical gel electrophoresis of

32P-la-difference in size of about 50 to 70 bp would be beledH-1RF DNA digested by HaeIII or EcoRI.H-1

expected. Althoughtheproportionsof

EcoRI-B1

RF DNA labeled with 32po4 was prepared in the toEcoHl-B2were constant at 2:1independent of usualmannerthrough thesucrosegradient step. RF

whether the digestion was complete or notasin DNA infinal volumes of 50p1 wasdigested for 16 h

Fig. 1, afurtherexperiment to detect a possible at37°C with 5

pi

ofHaeIII (A)orfor 2 h at 37°Cwith

second cleavage site near the first was con- 5 p of EcoRI (B).TheDNA wasanalyzed by

electro-ducted. Arelativelylargeamount of[3H]TdR- phoresis in

cylindrical gels

of

3%

acrylamide-0.5%

ducled.

reA. la

rigesamoun

to

[omH]tdR

agaroseataconstantvoltage of30Vfor16 h atroom

labeled RF DNIA was digested to completion temperature. The counts recovered in each fraction

with EcoRI.Thisdigest contained40,000 cpm of were converted to percent total counts, and the

EcoRI-B and was analyzed by electrophoresis number of base pairs for each fraction was calculated

for 3 h at50Vina3%acrylamide gel (datanot using a total of4,920 (cf. Fig. 4). The electrophoretic

shown). Using the mobility of the EcoRI-B frag- mobility ofeach fragment expressed as a relative

ment and a linear extrapolation- of mobility mobility in comparison to EcoRI-A fragment was

versus logarithm of molecular size as in Fig. plottedagainstthefragment size on alog scale in(C).

2C,

the3-h

gel

wouldhavecontainedfragments The mobility ofEcoRI-B indicates a fragment size of

2Ci

tharelative

3-hmgelwouldt

of9withrespecto about1,050

bp

or21.3%

of

thegenome.In

Fig.

2B,

with a relative

mobility

of 9 with

respect

EcoRI-B contained 20% of the total 32P

recovered.

EcoHI-B and an estimated sizeof 2 to 4 bp. A The

HaeIII

fragments I-L

migrated

off this gel.

putative fragment of 50 to 70 bp would have

containedapproximately1%of the totalcounts extracts of H-1-infected cultures that the RF

per minute or about 2,000 cpm. No evidence for DNA peak has a fast-sedimenting shoulder

suchafragmentwasobtained. Evidenceinsup- (e.g.,Fig. 3). Similarly, gel electrophoresishas

port of the thirdalternative,namely, that there revealed a species of DNA migrating more

is oneEcoRI cleavagesiteandthat the EcoRI-B slowlythanRFDNA,e.g., fraction6in Fig. 1.

fragment is heterogenous, ispresentedbelow. To determine the molecular weight of this

H-1 dimer RF DNA. It has been observed DNA, the

[14C]TdR-labeled

viral DNA

illus-routinely in neutral sucrose gradients ofHirt trated in Fig. 3 was pooledasillustrated, and

on November 10, 2019 by guest

http://jvi.asm.org/

[image:4.501.252.444.51.412.2]

B. .A . beleddimerwas

prepared

asabove and

purified

30

I-+--A-I

by preparative gel electrophoresis. A portion of

this preparation was digested by EcoRI and

analyzed by electrophoresis in a 1.4% agarose

25 - gel(Fig. 5).Threefragmentswereobtained;the

largest had the electrophoretic mobility of

EcoRI-A, the second migrated with a mobility

20 - equivalent to 2,100 bp or twice the size of

o EcoRI-B, andasmaller portionwasat the

posi-x tionof

EcoRI-B.

This is the

expected

result fora

E lS dimer RF DNA linked through the EcoRI-B

E 15 fragments, rather than in tandem. A similar

u l result was obtained after a digestion of a

mix-10 ._lo} ture of equal parts of monomer and dimer RF

. DNA. The small amount of 3H recovered at the

position of EcoRI-B probably arose from

con-RI

I

1 5 10 15 20 25 o \

Fraction 9g

FIG. 3. Preparative sucrose gradient centrifuga- L - H 1 dimer RF tion of H-1 RF DNA. H-1 RF DNA was labeled with

['4C]TdR by incubating parasynchronous hamster .C 5 - RI 2

cultures infected with tsl H-1 at 39.5°C 12 to 16 h p.i. *j R

in medium containing FUdR (0.5 pg/ml) and RI 3

[f4C]TdR (1 pCilml, 57.1 mCi/mmol). ViralDNA , H1 RF

wasextractedbythe Hirt method and sedimentedin - RI5

agradient of 5 to 20% sucrose in 50 mM Tris (pH \

8.0)-i MNaCl-1 mMEDTA-0.2% Sarkosyl in the 0- RI

SW25rotorfor 18 h at 24,000 rpm. Fractions of1 ml 6

werecollected through the bottom of the tube,10-pl

aliquots were assayed for radioactivity, and the frac-tions were pooled as in the figure. The DNA was precipitated with 2.5 volumes of ethanol at -20°C for

16or more h.The precipitates were collected by cen- 106 ' I ' I ' I I

trifugation and washed once with 70% ethanol-50 0 20 40 60 80 100

mMTris (pH 7.5)-1.5 MNaCl-1 mMEDTA, and the Distance Migrated

DNA was stored in 10 mM Tris (pH 7.5)-10 mM FIG. 4. Determination of the molecularweights of NaCl-0.1 mM EDTA at -20°C.

H-1

monomer and dimer RF DNA. The molecular weights of H-1 monomer and dimer RF DNA were pool B, previously determined to contain some measured by coelectrophoresis of[f4C]TdR-labeled

of the RF DNA of pool A, was subjected to H-1DNApreparedas in Fig.3 with the

[3H]TdR-coelectrophoresis with an EcoRI digest of labeled EcoRIfragments of X phage DNA.

[3H]TdR-[3H]TdR-labeled

bacteriophage X DNA in0.7% labeled XDNA was prepared as described(11) and

agarose gels. The results are presentedasthe digestedtocompletion with100 U of EcoRIin a

vol-umeof100pi.Thereactionwas stopped by the

addi-log

molecular

weight

versus

distance

of

migra-

tion ofEDTA to 50 mM, and 20pdoftheXdigest and

tion in Fig. 4, and the molecular weights de- a

5-pi

aliquot of the

["4C]TdR H-1

DNA were

ana-rivedfor thetwopeaks ofH-1DNAwere 3.26x lyzed by electrophoresis in duplicate in 1%agarose

106forH-1RF DNAand6.5 x 106for the faster- cylindrical gels as in Materials and Methods. The

sedimenting species. These values represent figure illustratesthe distance migratedand thelog

the average of two separate determinations, molecular

weight of

EcoRIX

fragments

(a)andthe

which differed by less than 1%. It should be electrophoretic mobility ofH-1 RF DNA monomer

noted thatthe curve isnonlinearinthe regions and dimer

(0).

The molecular weight of the smaller

where thefaster-sedimenting H-1DNA is rep-

H-1

DNA was calculated using a linearregression

whresent.

Thes

D * constructed by the methodof least squares for log

respeted.ofa

This

erDN

Has

the

moleculAr

weig molecular

weight

as a

function

of

distance

migrated

expected of a dimerofH-i RF DNA. for the XEcoRI fragments

2-6

(11).

The molecular

Toclarifyfurther the identity and structure weight of the larger

H-1

DNA was estimated

graphi-of this putative dimer RF DNA, [3H]TdR-la- cally, since it lay on the nonlinear part of thecurve.

on November 10, 2019 by guest

http://jvi.asm.org/

[image:5.501.60.286.46.293.2] [image:5.501.265.456.96.453.2]

EcoRIA' ' ' '

'IF

cohesive termini of themonomer RF DNA

dur-15 - E _ ing extraction and subsequent handling. To

de-termine whether all RF DNAcould anneal, a

portion of the 11-to 12-h RF DNApreparation

abovewas incubated in 50%formamidefor4 h at 400C before electrophoresis. No increase in

10_ dimerRF DNAoccurred.

or

.A more directtest for the possibility of

ran-x - dom annealing of cohesive RF monomers was

E made

using

the

density

label

bromodeoxyuri-u dine (BUdR). In this experiment, wt RF DNA

^5 - _ was

prelabeled

with

[14C]TdR

and incubated30

dimer

minwith

FUdR,

followed

by

30 minof

labeling

EcoRIB with [3H]BUdR in thepresence of FUdR. This

EcoRI B

0 10 20 30 40 50 60 20 A. + _10

Fraction +

FIG. 5. Analytical gel electrophoresis of an EcoRI 15- t

digest ofH-1 dimer RF DNA. Electrophoresis was in o L \

a1.4% agarose gel as in Fig. 1 except buffer E was x x

used for 16 h at 25 V at room temperature. DNA Elo_, 5 E

peaks were identified from the position of monomer u O- u

fragments on a replicate gel and their mobilities rela- I \ i tivetothe marker EcoRI-A.o

taminationof the dimer

preparation

withmon- 0 O

omerRF DNA.

To determine whether the dimer DNA con- B. tained covalently linked dimer length strands,

[3H]TdR-labeled dimer or monomerand 32P-la-

10

beledmonomerRF DNA werecomparedinal- °

kaline sucrose gradients (Fig. 6). Virion DNA

tE

was sedimented in a replicate gradient (not X ls shown),thepositionof which is indicated by the I

arrow

(V),

and thearrow

(D)

indicates the

posi-

_

tion of dimer length DNA expected using the /

equationsofStudier(25).Itcanbeseenthat the ° / t

majorityof bothmonomerand dimer DNAsare

genome length in the alkaline gradient and d

thus not covalently linked. Both preparations

show

fast-sedimenting

shoulders not seenwith

01

5 10 15 20 25

30

virion

DNA,

which

we will consider in detail

Fraction

i-9

below.

_Thelw

relative proportions of dimer andmon- FIG. 6.

Centrifugationofdimerand

monomerH-1 RF DNAinalkalinesucrosegradients.Labeled H-1

omer RF DNA synthesized at different times RF DNA was prepared as in

Materials

and Methods.

postinfection (p.i.) were examined in wt H-1- The B fraction of the 3H-labeled

H-1

DNA shownin

infected hamster embryo cells labeled with Fig. 3 was fractionated into monomer and dimer [3H]TdR11to 12 hp.i.,11to 12 hp.i. plus a 5-h fractions by preparative gel electrophoresis in 1%

incubation in the presence of unlabeled TdR, agarose. 32P-labeled monomer

H-i

RF DNA was

and 16 and 17 h p.i. The viral DNA was ana- addedto the 3H monomer (A) ordimer(B)fractions,

lyzed by electrophoresis in 0.7% agarose gels. and after the mixtures were adjusted to 0.2NNaOH,

Theproportionof viral DNAthatwasdimer RF they were sedimented in a 5 to 20% alkalinesucrose

DNAvaried little: 17%o for the 11- to 12-h

pJ

i gradient

(cf.

A in

Fig.

3)

for

6hat42,000rpmat

4°C

in the SW50 rotor. A replicate tube contained 3H-labeling period, 26% for 11 to 12 h plus 5-h labeled virion H-I DNA, and its sedimentation

posi-chase, and 22% for the 16- to 17-h label. Since tion is indicated by the arrow V. The arrow Ddenotes

the majority of the dimers are not covalently the expected sedimentation position of dimer DNA

linked, they may be formed by annealing of relative to monomerby theequation of Studier (25).

on November 10, 2019 by guest

http://jvi.asm.org/

[image:6.501.246.442.226.520.2]

period ofdensity labelingwaschosentoproduce AIim

*.l

alabeling of onlyasmall portion of the total RF H Light

pool. Therefore, if dimers formed by random Hybrid Light

association ofmonomers, the majority of the W

V

dimerslabeledwith3Hwould containonlyone LO)

[3H]BUdR-substituted monomer RF DNA and ° x (

beless than hybrid density. On the otherhand, x 7

ifdimers replicate semiconservatively during E 5 \ E

thedensity labelingoriftheyaregenerated by u

afailure of

daughter

RFmonomersto

segregate

° 1

after replication, 3H-labeled dimers would have - 135

hybrid density.

Dimer RF DNAwas

prepared

, l

bysucrosegradient centrifugation and bypre- *

parativegel electrophoresis, and the dimer and

monomer (sucrose

gradient only)

ugtin-n-

RF DNAs

were analyzed

by

isopycnic

centrifugation

in - .50

Cs2SO4

(Fig. 7A and B). About 60% of the3H- \

Bronomer

m

labeledmonomer isof hybriddensity, whereas O

the density of the 3H dimer is intermediate 15 51*

between light and hybrid, and virtuallyno 3H > 4 \ 4

dimer isofhybrid density. This is compatible 0 - o t

witharandom annealing ofmonomerRF DNA x x14 E

asthemechanism of dimer formation and rules E 5 10

outafailure ofdaughter RFmonomersegrega- u °

tionas a meansof dimergeneration. However, -X

this result would also be obtained if dimers 5

replicated more slowly than monomers, such o

thata 30-mindensity labelisless thanorequal

to the replication time for dimer RF under I

these conditions.

_1

To examine this possibility, the experiment C 10 20 30 40

was repeated in asimilar manner, except the Fraction

period of density label was extended to60min FIG. 7. Isopycniccentrifugationofdensity-labeled

(not

shown). After the

1 h

of

density labeling,

H-i

RF monomer and dimer DNA.

H-1

DNA was mostof the 3H-labeleddimer DNA was of hy- labeled in parasynchronous NB cultures infected

brid density; however, appreciable amounts of withtsl at 39.50C first with['4C]TdRfrom 12 to 15.5 thetotalmonomer RF DNA werecalculatedto hp.i. (FUdR,0.5pg/ml; [14C]TdR, 1 XCi/ml,57.1

have become heavy (13%) and hybrid (38%) mCi/mmol), followed by a 0.5-h treatment with

densities. Thus, the distribution of density la- FUdR alone, and then labeled with

[3H]BUdR

from

bel inthedimerRF DNA wascompatiblewith 16 to 16.5 h p.i. (FUdR, 0.5 pglml;[3H]BUdR,2

a slower dimer replication process, random

XCi/ml,

10-6

M). Viral DNA was extracted by the

annsalow

dimrF

DNA* Hirt method and

fractionated

intomonomerand

di-annea

of RF

DnA

merby

centrifugation

in

a neutral sucrose

gradient

complex situation. and bypreparative gel electrophoresis in 1 % agarose.

Covalentlinkage of theV toC strandinthe Monomer and dimer DNA were banded to equilib-EcoRI-B fragment ofmonomer RF DNA. It rium inCs2SO4:(A) dimer RFDNA, (B) monomer

waspreviously observed that the RF monomer RF DNA.Centrifugationwasfor48 hat35,000 rpm

and dimer DNApeaks had afaster-sediment- at

4°C

in thetype 40rotor.

ing shoulder in alkaline gradients, suggestive

of the presence of single-stranded DNA of guanine plus cytosine content that might fail to

greater than genome length. Tattersall et al. denature with heat under the conditions used

have presented evidence for the existence of forMVM RF DNA (23). Nevertheless, if such a

some RFDNAoftheparvovirus MVM (minute covalent cross-linkage is present in a portion of

virus of mice) that renatures unimolecularly theH-1RF monomer and dimer DNAs, then it

after heat denaturation, andsuggested that a could be isolated by using its rapid renatur-covalently closed turnaround may exist in RF ationrate. This was tested by alkaline

denatur-DNA (27). We have obtained evidence for a ationof wt monomerH-1 RFDNA followed by

region at one end of H-1 RF DNA with a high rapidneutralization and immediate application

melting temperature implicating a very high to a BDC column. Double-stranded DNA was

on November 10, 2019 by guest

http://jvi.asm.org/

[image:7.501.269.458.49.399.2]

H-1

TABLE 1. Akaline denaturation-BDCchromatography of H-1 RF DNAa

Cycle1 Cycle2

Procedure

cpm % Total cpm %Total

Washes 1,295 3,108 1.7

1 MTBS-total 261,560 18.5 128,560 73.1

Caffeine-total 1,151,340 81.4 44,300 25.2

a [3H]TdR-labeled wt H-1 RF DNA was denatured by 0.2 N NaOH, neutralized at

0°C,

and

chromato-graphed on a column of BDC as in Materials and Methods. The DNA eluted with 1 M TBS was collected by ethanol precipitation, and theprocesswasrepeatedincycle2. Values shown are total counts per minute recovered and theirrespective percentages.

selectively elutedwith 1 M TBS and DNA con- oldandnewmolecules. This DNA was fraction-taining single-stranded regions recovered by ated by alkaline denaturation-BDC

chromatog-elution with 2% caffeine. The results are sum- raphy, followed byanalkalinesucrose gradient

marized in Table 1. When the double-stranded as above. The results in Fig. 8B show that the

DNA was putthrough asecondcycleofalkaline 3H/1'4C ratios were lower for the dimer length

denaturation-BDC chromatography, the per- DNA than for the unfractionated monomer RF

centagerecoveredasdoublestrandedincreased DNA, 0.29 versus 1.1. Thus, newlyreplicated

from 18.5 to 73%. A portion of this DNA was monomerRF molecules (within 30min)contain

then analyzed by alkalinesucrosegradientcen- less of the covalent linkage than the general trifugation (Fig. 8A). A majority of the DNA pool,asrepresented by the[14C]TdR-labeledRF

(68%) sedimented at the expected position of DNA.

dimer length singlestrands, in constrast tothe Covalentlinkage of Vto C strand in dimer [14C]TdR-labeled tsl RF DNA monomer RFDNA. Experimentssimilar to those used to marker. The final recovery indicated that at show linkage ofV toCstrandsinmonomerRF least 9%of the RFmonomercontaineda cova- DNAwerecarriedoutwith[3H]TdR-labeledwt

lentlinkagebetween the Vand C strands. dimerRF DNApurifiedby gelelectrophoresis.

The localization of the V-C linkage was de- Afteronecycle ofalkalinedenaturation, 24%of termined by digesting RF DNA with EcoRI the 3H was recovered as double stranded in before alkalinedenaturation andBDCchroma- BDC chromatography, and 87% of this DNA tography. The double-stranded DNA so ob- sedimentedasdimerlengthin an alkaline su-tained was analyzed by electrophoresis in a crose gradient (not shown). This represented

1.4% agarose gel and compared with the total about 21% of the original dimer preparation.

digest (without selection for the rapidly rena- The gel electrophoresis of this dimer prepara-turing DNA) inareplicate gel. The resultsare tion after EcoRI digestion is shownin Fig. 5.

illustratedtogether forconvenience inFig. 9A. This DNA contained about 5%

EcoRI-B

mon-TheEcoRI-Bfragmentwas clearlyselected for omer would be dimer length

single-stranded

by the procedure isolating rapidly annealing 20% contamination with monomer RF DNA. DNA, indicating that the covalentlinkageis in Sincewehaveseenthat about10%of the mon-the EcoRI-B fragment. A control experiment omer would be dimer-length

single-stranded

showedthat BDCchromatography ofanEcoRI DNA, this would constitute only2%of the total digest of RF DNA without alkalinedenatura- dimer preparation, so that it can be deduced tion resulted in the retention ofabout 30% of that some of the dimers contained covalently theDNAafter 1MTBSelution,suggestingthe linkedVand C strands.

presenceofsomesingle-strandedregions,butit To test whether the covalent linkage inthe

was notselective for eitherfragment. This sub- dimers is at the internal left end of the

mon-ject will beconsideredin moredetailin asubse- omer, and probably at the site oflinkage

be-quent sectionof this

study.

tweenthemonomers, dimerDNA was

digested

Ifthe covalent linkage of V to C strand in with EcoRI and then subjected to the alkaline

monomer RF DNA was involved in the initia- denaturation-BDC chromatography procedure.

tionofreplication of RF DNA or progeny DNA The rapidly renaturing DNA was

analyzed

by

synthesis, then

rapidly

labeled molecules gel electrophoresis (Fig. 9B). Itwasfoundthat would be enriched in this property compared theBfragmentwasisolatedina40-foldmolar

withthetotal collection of RF DNA molecules. excesscomparedwith the Afragment and thus

The [14C]TdR-prelabeled RF monomer DNA containedthe V-Cstrandlinkage.It isunlikely

used for the

[3H]BUdR

densitylabel in Fig. 7B that a covalent linkage occurred during the

was a ready source of DNA for comparison of Hirt extraction and subsequent

handling,

so

on November 10, 2019 by guest

http://jvi.asm.org/

A. crudelocalization of the

origin

of

replication.

In

these experiments, tsl H-1 was used at the

restrictive temperature to inhibit progeny DNA

20 - > \ ,0, _

synthesis.

This method has been

successfully

x 8 | , z used by Danna and Nathans for simian virus 40

E (7) and

by

Godson for

OX174

(10). The first

problem to be overcome with the linearH-1 RF

Ut

I I

0

I

'10~ ~ q10

~~~~10I

~~~~~~~~~~~A.

X 0

* -Eco RIA Eco RI B

B. dimer monomer Eo

II I

60 30 Eo

I

~~~~~~~~~~~~~~~~~~~~~~Eco

RI B

9 ~~

~~~~~~~~~~~~~RF

~ darner

140 20 A

9, ~~~~~~~~~~0

x x

E E B.

FIG. Alaln -ursrdetcnrfgto

a-0~~~ ~ ~~~~~~~.

020 01

4

0~~~~~~~~~

C1

~~~10

20 30 E

Fraction 9-ga

FIG. 8. Alkaline sucrose gradient centrifugation 2 -dimer

dirner

of rapidly renaturing RF DNA. (A) H-1 RF DNA 2 RF Eco RIA Eco RIB EcoRIB monomer was selected for its ability to reanneal

rap-idly by two cycles of alkaline denaturation-BDC chromatography (Table 1). The DNA in the 1 M TBS

eluate

of

thesecond

chromatography

was

adjusted

to a

JLAb2Ahl

A

0.2 NNaOH andsedimented in agradient of5 to

20% sucrose in 0.2 N NaOH-0.8 M NaCl-1 mM 1 10 20 30 40 50 60

EDTA for 16 h at 32,000 rpm at 4°C. A marker of - Fraction +

[14C]TdRmonomer RF DNA obtained frompoolA of FIG. 9.

Analytical

gelelectrophoresisofEcoRI

di-Fig. 3 was included. (B) H-1 RF DNA monomer gsso RFDAbyeoeandaferoakaie

dEntRa-prelabeled with ['4C]TdR (12 to 15.5 h p.i.) and gestsofRFDNAbeforeand

afiter

alkaline

denatura-pulse-labeledwith[H]BUdR (16to16.5 hp.i.) as tion-BDC chromatography. (A) RF DNA was

di-puselbeldwi

h.

t

7,HBUd

(16 to16. h

yci)

as gestedto completion withEcoRI and then

fraction-describe in Fig.7, sctdmtog

onehycleo

ated byalkaline denaturation-BDCchromatography.

raplk

annalineDN

denaturation-BDC chr1MaTo

Tewa AsampleoftherapidlyannealingDNA inthe 1 M

rapidlyannealing DNA inthe1 M TBS eluate was TBS

eluate

and an aliquot removed before alkaline

centrifugedinanalkalinesucrosegradientasin(A). denaturation-BDC chromatography were analyzed

separately in replicate gels of 1.4% agarose.

Electro-thisresult strongly suggests that the dimer RF phoretic conditions were 17 h at 25 V at room

temper-DNAexists in vivo. ature. Results ofthe replicate gels wereplotted

to-Evidence thatthe Bfragment is a terminus gether for convenience: before alkaline

denaturation-in RF DNA

replication.

By

analyzing

the dis- BDC(), andafter alkaline

denaturation-BDC

(0).

trinution

repltion.d

Be

zIAang

te- (B) DimerRF DNA was digested with EcoRIand

tribution of [3H]TR in the EcoRI-A and B fractionatedbyalkalinedenaturation-BDC

chroma-fragments of H-i RF DNA that have

just

com- tography. The1 M TBS eluate was analyzed in a gel

pleted replication during a short incubation in of1.4% agarose as in (A),except electrophoresis was

thepresence of[3H]TdR,it is possible to make a for 20 h at 25V.

on November 10, 2019 by guest

http://jvi.asm.org/

[image:9.501.65.268.56.410.2] [image:9.501.268.456.158.501.2]

DNA is to devise a method of isolating the .

moleculesthat have just completed replication A.

during the short pulse of [3H]TdR. In these IMTBS 0-27 Caffeine molecules, the specific activity of TdR at each

position in the molecule will increase as the

distance of that position from the origin in- o

creases. In comparison, RImolecules will have x

adeclining specific activity in TdRas the TdR E 5

-positionfrom theorigin increases.For this pur- l

pose, I have usedchromatography on columns I

ofBDC, which separates double-stranded DNA

fromthe more tightly bound DNA that contains °

single-strandedregions (2, 27). Thismethodor °

a similar one using benzoylated napthoylated i DEAE-cellulose has been used for the

separa-tion of viral RI molecules from mature DNA for 0

lambda (13),adenovirus(26),

OX174

(21), MVM B.

(27) and, most recently, herpes simplex virus I-I - 15

(22).

Rapidly labeled H-1 RF DNA was found to bind preferentially to BDC and to require caf- °

feine for elution, as has been shown for other x 5 - 10 x

E E

examples of replicative intermediates cited Q lI

above. Rapidly labeled viralDNA wasprepared I

bylabelingtsl-infectedNBcultures,afterincu- "V

bationat39.5°C, for5minat22°C, orfor5min

50

or 15minat39.5°C, with [3H]TdR before

extrac-tion of viral DNA. A

uniformly

labeled RF DNAwas

similarly prepared by

incubating

the 0

cultures from 12 to 16 h p.i. in FUdR and 1 2 3 1 10 20 30 40 [14C]TdR (cf. fractionA in Fig. 3).Afterapre- Fraction

liminary purification in neutral sucrose gra- FIG. 10. BDC

chromatography of rapidly

labeled dients, the [3H]DNA that sedimented as mon-

H-1

RF DNA.ParasynchronousculturesofNBcells

omer RF DNA (but which may also include were infected with tsl H-i at 39.5C and treatedwith

early RI molecules), wascombined with apor- FUdR for 30 min, 14.5 to 15 hp.i.At 15 h p.i. the tionof['4C]TdR-labeledRF DNA andwasthen cultures were labeled for5or 15 min with [3H]TdR chromatographed on BDC columns (Fig. 10). (20 p.Ci/ml, 50Ci/mmol) in thepresence ofFUdR.

Thepercentage of the labeled DNA that eluted Viral DNA

sedimenting

as monomerRF DNA was

withcaffeinewas 95%forthe22°C,5-min label prepared in the usual manner and was chromato-(not

shown),

80% for the 5-min 3H label at graphed with a

['4C]TdR

marker (cf. Fig. 3) on

39.50C,

61%for the15-min 3Hlabel,and38%for columns ofBDC-cellulose using elutions

of2O

ml of1

the_the4-h_-h[14C]TdR-labeled_{[4C]TR-laeledDNA.Thus}DNA. Thus,

_{therap-caffeine}

theMrap-

M TBS and_{in 1 M}a 20-mi linear_{TBS. One-milliliter}

gradient

_fractions

of

0 to 2%

were

idly labeled RF DNA was preferentially re- collected and

50-pi

aliquots were assayedfor

radio-tained after the 1 M TBSelution, even though activity; (A)5-minlabel, (B)15-minlabel. The1 M

the more rapidly sedimenting RI molecules TBS and 2% caffeine eluates were pooled as illus-were already partially excluded by the zonal trated and the DNA wasprecipitated with 2.5

vol-centrifugation

inneutralsucrose. umesof ethanol at

-20°C

overnight.

The rather high fraction of the

[14C]TdR-la-beledDNAthat remained boundafter salt elu- wouldbe retained

preferentially

after the 1 M tion deserves further study. Electron micro- TBS elution. This wastested

by

digesting

3H-scopevisualizationof thisfraction showed that labeled RF DNA

(4-h labeling period)

with

onlyabout 25% of the molecules were RI mole- EcoRI. The

digest

was extracted with

phenol,

cules (23), so that one explanation would be

precipitated

with

ethanol,

redissolved in0.3M

that these RF DNA molecules have large

TBS,

and

chromatographed

on BDC.

Thirty-enough

single-stranded

DNA regions to bind seven percent of the total 3H eluted with

caf-BDC in 1 M TBS, as recently reported for feine. Portions of the 1MTBSand 2% caffeine

herpes simplex virus (22). Ifthis were true, eluate were analyzed by gel electrophoresis,

then theseregions may beuniquelylocated, so and there wasnopreferential binding

ofEcoRI-that afterEcoRI cleavage,oneof thefragments A

compared

with -B

by

the BDC

(Table

2).

on November 10, 2019 by guest

http://jvi.asm.org/

[image:10.501.252.446.50.379.2]

TABLE 2. Compositionof BDCchromatographic visualization by electron microscopy.

There-fractions ofEcoRI-digestedH-i RF DNAa fore, the saltfractioncontains the

nonreplicat-%Composition ing molecules, even though some are retained

Elution with: by thecolumn for anunknownreason.

RF EcoRI-A EcoRI-B Returningnow to the ts 1 RF DNAlabeled for

1 MTBS 11 67 22 15minat39.5°C and fractionated by BDC

chro-2%Caffeine 23 60 17 matography (cf. Fig. 10), the1 MTBS fraction

aH-1 RF DNAlabeled with 3H wasdigested to and 2%caffeine fraction wereseparately

precip-near completion with EcoRI. The digest was ex-

itated

with ethanol and

digested

With

EcoRI,

tracted with phenol andprecipitated with ethanol. and the digests were analyzed by gel

electro-Theprecipitate wasredissolved in 0.3 M TBS and phoresis (Fig. 12). Additional

[14C]TdR-labeled

fractionatedon acolumn of BDC. The 1 M TBS and marker RFDNA,not fractionatedbyBDC,was 2%caffeineeluates wereprecipitatedwithethanol, added before the digest to improve the 3H/14C

and the DNA was analyzed by electrophoresis in ratios (theabsolutevaluesforthe 3H/14C ratios

1.4% agarose gels. The values shown represent the of the twocolumn fractionswere notadjustedto

percent composition of RF, EcoRI-A, and EcoRI- be equal). The 3H/14C ratios for the 1 M TBS

Bin each eluate. eluate EcoRI-A and -B fragments prepared

from this mixture were 3.32 and 4.42, respec-The 2% caffeine eluate was enriched for the

undigested RF DNA.

If the postulated single-stranded regions A' I were distributed within both EcoRI-A and

EcoRI-B, then RF DNA in the caffeine eluate shouldshowadecreaseinsedimentation

veloc-ity inalkalinesucrosegradientsincomparison

10

to the 1 M TBS eluate. In this experiment, uniformly labeled monomerRFDNAwas frac-tionatedonaBDCcolumn, and theDNA inthe 1MTBSeluate andthe 2%caffeinewasprecipi- 5

tatedby ethanoland thensubjectedto

sedimen-'l

tation inalkaline sucrose gradients. Amarker

of

[14C]TdR-labeled

RF DNA (Fig. 3) was in- x

cluded, and its modal sedimentation is indi- E °

catedbythearrow (Fig. 11). Althoughthe caf- X 20 -B.

feine eluatewasenrichedslightly for DNA sed- I

imenting slower thanthegenomelength DNA, 1

both preparations were largely intact dimer

15

length ormonomerlengthsingle strands. This indicatesthat

internal

single-strandedgaps are

notcommon in the RF DNA of the2% caffeine

eluate.

10

_

Another possibility is that the DNA is re-tained nonspecifically andnotbecause internal single-strandedregions exist. Ifthisisthecase,

rechromatography

of the 1 M TBS eluate 5

shouldagain produce a sizable fractionretained

after 1 M TBS elution. This experiment was

doneusing aportionofthe first 1 M TBS eluate

of

Fig.

11, which

represented

one-tenth of the

01

10

20

30

total DNAused for the first BDCcolumn. The

result obtained was that 24% ofthe 1 M TBS Fraction g

fraction remained tightly bound and eluted in FIG. 11. Alkalinesucrose

gradient

centrifugation

the 2% caffeine eluate. This is a smaller per- of BDCchromatographic fractions of H-1 RF DNA.

centage than duringthe first chromatography

wt

RF DNA labeled4 h with

[3H]TdR

was

fraction-(37%),

but it does indicate a

relatively

high

ated by BDC chromatography using batch elutions of (37%, bu

_evel.of

_nonspeclinc

it

einic

does

_retentIon

ten

a

_after

rltel

_1rM

gT

i

_rTB

1_{collected by ethanol precipitation and sedimented}MTBSand 2%caffeine. The DNAfractions were_as

elution. In conclusion, RI molecules are re- alkalinesucrosegradientsas inFig. 8. The marker

tained on BDC columns after salt elution on DNA of[14C]TdR-labeledRFsedimentedatthe

ar-the basis ofthe kinetics oflabelingand direct row; (A) 1 M TBS eluate, (B) 2%caffeineeluate.

on November 10, 2019 by guest

http://jvi.asm.org/

[image:11.501.263.455.266.582.2]

IA.

E

RIA

labeled DNA, migrating faster than EcoRI-A,

5-A.EcoRIA arose as aresultof the BDC chromatographyor

4-20 R h 5 EcoRI

digestion,

thebehavior of

rapidly

labeled

4-20 RF DNA and RI DNA in

gel

electrophoresis

3 -15 cJ > \was examined.

3H-labeled

tsl RF DNA was

x;

3-31~

prepared by

a5-min incubation with

[3H]TdR

I _E at

39.5°C

and fractionatedinaneutralsucrose

z2 a Fg ubsfo h

".ffi , _0 gradient

(Fig.

13). Tubes from the gradient

. EcoRIB were pooled to make threefractions, A, B, and

1

C,

as illustrated. After precipitation with

---

ethanol,

each fraction was

subjected

toagarose

o_ gel electrophoresis in the presence of a

10 30

[14C]TdR-labeled

RF preparation containing

monomer and dimer RF DNA (Fig. 14). Frac-tionAshows[3H]DNAexhibitingahigh 3H/14C

-15 20 ratio migrating in advance of monomer RF

IO

s I DNA

(Fig.

14A).

Therefore,

the

rapidly

labeled

x -o < v 9 10 DNA

migrated

ahead ofRF DNA before BDC

a- chromatography and EcoRI digestion, with a

I o 1 _

°0

mobility equivalent to a duplex DNA 10 to 15%

-5a 1' smaller than RF DNA.

After

EcoRI

digestion

-o

Q ! zthis DNA migrated ahead of EcoRI-A, and its

mobility

increased about the amount

expected

10 20 30 40 50 if EcoRI has cleaved theEcoRI-B region from

Fraction +

FIG. 12. Analytical gel electrophoresis of EcoRI

digests of RF and RI DNA labeled15 min.H-1 tsl 15 I

RFDNAlabeled15minwith[3H]TdRwas fraction-ated by BDC chromatography as in Fig. 10. The DNA in the1 MTBS and2%caffeine eluates were

separatelydigested with EcoRI after the addition of C B A

analiquot ofuniformlylabeled[14C]TdR-labeledtsl I

RF DNA and analyzed by electrophoresis in 1.4%

agarosegelsasinFig.9.Electrophoresiswasfor17 h 10_

at25Vat roomtemperature;(A)1MTBS eluate, (B) ||

2%caffeine eluate. 0

x

tively. Thus, if the 3H/ 4C ratio ofEcoRI-A is E

normalized to 1, the ratioforEcoRI-B was 1.33.

This increase is

expected

if EcoRI-B does not 5

contain the origin ofreplication. The caffeine

eluate shows 3H/14C ratiosof7.94 and8.27 for

EcoRI-A and -B, respectively. The normalized

value for the

EcoRI-B/EcoRI-A

was 1.04. In the

caffeine eluate the EcoRI-A

fragment

did not show the

expected

higher

3H/14C

ratio

compared

0

with EcoRI-B, but it can be seen in Fig. 12B 1 10 20 thatasmall

portion

ofthe

rapidly

labeled[3H]- Fraction g DNA with high 3H/14C ratios (as high as 40) FIG. 13. Preparative sucrose gradient centrifuga-migrated aheadof EcoRI-A. This DNA was not tion of rapidly labeledH-1 RF DNA.

Parasynchro-included in the average ratio of 7.94 cited for nous NB cultures infected with tsl H-1 at 39.5°Cwere

EcoRI-A above. treated 14 to15 hp.i. with FUdR (0.5 pglml) and

It would beexpected that RI moleculeswould labeled at39.5°Cwith [3H]TdR (50 pXi/ml,20

Cil

migrate moreslowly than linear RF DNA due mmol, + FUdR 0.5 pg/ml) for 5 min. Viral DNA

to their larger size, nonlinear configuration, was extracted as in Materials and

Methods

and

sedi-andby analogywith the resultsreportedfor RI mented ina

gradient

of5to

20%

sucroseasinFig.3.

moleculeofsimianvirus4028. Ontheot Fractions of1 ml were collected, and 20 pi was

moleculesof simian virus40 (28). On the other assayed for

radioactivity

directly,accounting for the hand, single-strandedregions andgreater flexi- acid-soluble[3H]TdR seen in fractions 23 to 28. Se-bility might accelerate their electrophoretic lectedfractions were pooled as illustrated, and the

mobility. To determine whether this rapidly DNA was collectedby ethanolprecipitation.

on November 10, 2019 by guest

http://jvi.asm.org/

[image:12.501.51.242.55.318.2] [image:12.501.255.450.326.542.2]

A.

d;mRF RF

0

x~~~~~~~

O

6

5

15B.

~~~~~~~~~dimer

RF RF

'o

C

4 I

M I,

9t

2t5

C.

1 10 S S Z 9 K

- ~~~~~Fraction +

FIG. 14. Analyticalgel electrophoresis of rapidly labeledH-iRF DNA.H-iRFDNA waslabeled for 5mi

at39.5°Casin Fig. 13. The DNA fractions pooled as A, B, and C from the preparative sucrose gradient ofFig.

13 were analyzed in 0.7% agarose gels.A marker DNA of['4C]TdR-labeled mnonomerand dimer RF DNA (pool B of Fig. 3) was added before electrophoresis. Pools A, B, and C are illustrated in (A), (B), and (C), respectively. It should be noted that thealiquot used for A was 10% of the total, and B and C were 25% of the total. Gel electrophoresisof the total DNA of apreparativesucrose gradient equal to a composite offractions A + B + C of a tsiRF DNA labeledsimilarly for 5minat22°Cwith[3H]TdR is shown in(D).

these molecules. This suggests that this DNA the following papers of this series.

contains a high specific activity of 3H in the Gel electrophoresis of fractions B and C of

EcoRI-A portion of the molecule, which pre- Fig. 13 is shown in Fig. 14B and C. The

pat-sumably contains the origin ofreplication. Ex- terns revealed predominantly dimer RF DNA

periments to demonstrate this more conclu- and some [3H]DNA exhibiting high

3H/"4C

ra-sively using other restriction endonucleases are tios migrating more slowly than dimer RF.

inprogress, anda similarconclusion has been A DNA

preparation

from tsl H-i-infected

reached by electron microscopy, asdescribed in

min

cells that had been labeled

for

5 at

22°C

on November 10, 2019 by guest

http://jvi.asm.org/

[image:13.501.120.397.64.501.2]

REPLICATION TERMINUS OF 707

wasfractionatedon aBDC columnasdescribed vatedslightly by the 30 to 40%contaminationof above, and the caffeine-eluted fraction was the2%caffeine eluate with the 1MTBS DNA banded by isopycnic centrifugation in a CsCl aspreviously discussed. The expectedhigh 3H/ gradient, precipitated with ethanol, and then 321pratio for EcoRI-A may be absentbecausethe subjected toelectrophoresis in a 0.7% agarose RImolecules are selectively lost bythis

proce-gel (Fig. 14D). Thiselectrophoreticpattern rep- dure. The sucrose gradient fractionused here

resents an approximate compositeof Fig. 14A, excluded DNA migrating faster than monomer

B, and C. Thus, even with relativelyshort la- RF DNA. Also, examinationof the 2%caffeine

beling times the majority of the label has the eluate byelectron microscopyrevealed that RI mobility of eitherthe RFdimeror monomer. molecules were not the majorformpresent (23). Origin forprogeny DNAsynthesis is inthe Anotherfactorcomplicating theanalysis of the EcoRI-A fragment. In the preceding section, 2% caffeine eluate isthat [3H]DNA represent-evidencewaspresented which implied that the ing one-third of the total of the fraction and origin for semiconservativereplication ofts 1H- with thehighest3H/32Pratio(38.25,normalized

1RF DNA is in theEcoRI-A fragment. Since it to EcoRI-A) did not migrate as EcoRI-A or

was found earlier (17) that about 80% of the

EcoRI-B,

butwasinthefirst fraction of the

gel.

labeled BUdR incorporated bywtRF DNAlate

Why

this DNA didnot

migrate

into the

gel

is ininfectionat370Cwas intheVstrand, it can unknown. In summary, the distribution of la-be estimated that approximately 60% of the bel in

newly completed

RF

DNA,

as repre-totalincorporationis asymmetricV-strandsyn- sentedby the3Hlabelinthe 1MTBS

fraction,

thesis. Thisprobably represents progenyDNA is notuniform and hasahigher

specific activity

synthesis by displacement of the preexistingV in the EcoRI-B fragment thanin theEcoRI-A strand from RF DNAby the nascentVstrand. fragment.

Thus,

this

fragment

appears to

in-Using theexperimentaldesign of the preceding cludethe terminusforbothRFreplication and

section, the distribution of

[3H]TdR

in newly progeny DNA synthesis. The

complementary

completed wt RF DNA wasused todetermine analysis ofRI molecules asrepresented by the the polarity of synthesis and thusacrudeesti- 2% caffeine eluate is complicated by several

mate ofthe location of the origin for progeny factors andis notconclusive for the localization

DNAsynthesis. ofthe origin.

Parasynchronous NB cultures infected with

EcoRI-B2

fragment contains the

turna-wtH-1 at

370C

werelabeled for 5 min at37°C round. The left end of H-1 RF DNAhas been

with [3H]TdRin thepresence ofFUdR after a shown to be heterogenous in its structure, as

15-min preincubation with FUdR. The

mon-omerRFregion of the neutralsucrosegradient

of the Hirt-extracted viral DNA wasfraction- TABLE 3. Distribution of nascent DNAin wt RF ated by BDC

chromatography

and 10% of the DNAa

applied DNA was eluted with 1 M TBS. This

DNAand the2%caffeine eluateswerecollected 1 M TBS

2%

Caffeine asbefore and

digested

with EcoRIinthepres- Prepn

3H/32P

Normal-

3H/32P

Normal-enceof

uniformly

labeled 32Ptsl RF DNA, and ized ized

the digestwas

analyzed

by gel electrophoresis Gel top 7.65 38.25

as in Fig. 11 (not

shown).

The 3H/32P ratios RF 0.457 3.07

wereaslistedinTable3.In the1MTBSeluate, EcoRI-A 0.149 1.0 0.200 1.0 the EcoRI-B

3H/32P

ratio was 7.4-fold higher EcoRI-B 1.107 7.43 0.512 2.56 than the ratio for EcoRI-A. There were only wtRF DNAwaslabeled with[3H]TdR for5

mm

small differences (no more than 12%) in at16 h p.i. at

370C

(50 ,uCi/ml; specificactivity, 20

[3H]TdR/32

P ratios for EcoRI or HaellI frag- Ci/mmol). The DNA

sedimenting

in a neutral

su-ments when the [3H]TdR-labeled RF was la-

crose

gradientasmonomer RF DNA (e.g., fraction

beled for a4-h

period.

Therefore,

thedistribu- A, Fig. 3) was fractionated by BDC chromatog-tionof TdRinRFDNA isnot

sufficiently

asym- raphy.The 1 M TBSand 2% caffeine fractions were

metric to affect these results. This suggests

precipitated

with70% ethanolat

-220C,

collected

by

that EcoRI-B is a terminusforwtDNAsynthe- centrifugation, and dried in vacuo. The DNA was

sis. redissolved in 50

,ul

ofEcoRIdigestion buffer

com-Thedigestof the2%caffeine eluate was com- bined with a portion of 32P-labeled RF DNA and

plete,

and the3H/32Pratiowasagain

higher

for

digested

for2 h at37°C with 100 U ofEcoRI. The

digest

wasfractionated by gel electrophoresisasin

EcoRI-B (0.512) than for EcoRI-A (0.200), but

Fig. 12.

The 3H and

32P

for each DNAspecies were

the difference was not as greatas for the salt- summed, and the 3H/32P ratios were calculated.

Val-elutable RF DNA, 2.5 versus 7.4

(normalized

uesshownare 3H/32P ratios and theratios

normal-values). The ratio for EcoRI-B would be ele- ized to theratio for Eco RI-A.

on November 10, 2019 by guest

http://jvi.asm.org/

manifested by twoforms of the EcoRI-B frag- tweenthetwoelectrophoretic species for

EcoRI-ment, i.e.,

B1

and B2. We have also presented B wereduetodifferencesin sizeandnot

strand-other evidence for structural diversity inthis edness,sinceneitherB1norB2 wasremoved by

region: a covalent linkage of V strand to C BDC chromatography. TheEcoRI-B fragment

strandin 10%of the RF forms and self-cohesive- with the covalent Vto Clinkage had the

elec-nessof the leftendasmanifested by dimeriza- trophoretic mobility of the smaller

EcoRI-B2

tion. When the distribution of [3H]TdR in fragment. The twotypesofEcoRI-B

fragments

newlyreplicated RF DNA wasanalyzed,itwas were produced bya single cleavagebyEcoRI,

observed inbothexperiments that the ratio of andasimilar result has beenobtained with the

3Hto theuniformlabel of14Cor32p washigher H. aegyptius restriction endonuclease endo

forthe slower-migrating portion of the EcoRI-B R HaeII. This enzyme also cleaves the mole-fragment peak corresponding to B1- than for cule once near the EcoRI site (0.20 genome

EcoRi-B2 (cf. Fig. 12).The slightly faster migra- length from theleft end), producingtwo

HaeII-tionof the"snap-back" EcoRI-Bversusthetotal B fragments (unpublished data). Since both

EcoRI-B in Fig. 9 suggeststhatit migrates as

EcoRI-B1

and -B2 were found with

tsi

RF as

EcoRI-B2. A more definitive analysis of these well as wtRF, both fragmentswere

preserved

findingswascarriedoutbyadirectcomparison

during

thecloning ofts1. It isprobable that the ofanEcoRI digest of 32P-labeled H-1 RF DNA EcoRIfragments differ by about 50 to 70 bpin before and after alkaline denaturation-BDC length, asthedifferenceintheir electrophoretic chromatography by slab-gel electrophoresis mobilities suggests, or they may have some

(Fig. 15). It isclear thattherapidly renaturing more complicated structural heterogeneity,

EcoRI-Bfragment migrates asEcoRI-B2.Inad- such as proposed for AAV DNA (8). Probably

dition, digestion of dimer RF DNA

produced

the simplestexplanationisthat there isa turn-the expected dimer EcoRI-B fragment, which aroundorhairpin atthe left endthatisabout shows a single

electrophoretic

mobility.

Frac- 50 bp in length in the B2 fragment, as

illus-tionation of theEcoRI-B digest on aBDC col- trated in Fig. 16 and, in the

B1

fragment, it is

umnwasalso donetodetermine whether either extended by a full-length complementary fragment hada

significant single-stranded

re- strand, but this remains to be confirmed.

gion,andelectrophoresis of the1MTBS eluate Analysis ofH-1 intracellular viral DNA by

shows that both

EcoRI-B,

and -B2arepresent. agarose-gelelectrophoresishas shown that H-1

Therefore, it canbeconcluded thatnewly repli- RF DNA exists in two forms, a monomerand a

cated RFDNAmolecules contain the

configura-

dimer. The identity of the larger species as a tion characterized by an

EcoRI-B,

fragment dimer RF wasmade by establishing its molecu-and that these molecules are probably con- larweight astwicethat ofmonomer RF DNA

verted to the foldback configuration with or and by the specificity of cleavage of the

mole-without the covalent linkage of the V to C cule by EcoRIendonuclease. H-1 monomer and strand afterward. Thedifferencesin

electropho-

dimer RF DNA molecular weights were

mea-reticmobilites of

EcoRI-B,

and-B2suggestthat sured by agarose-gel electrophoresis using the

the foldback region is about 50 to 70 bp in EcoRIfragments2through6ofbacteriophage X length, assuming thattheyareboth linear du- as markers of known molecular weights (11).

plexes. Monomer RFDNAhadan estimatedmolecular

DISCUSSION

weight

of3.26 x 106,and the dimerwasfoundto

DISCUSSION

be a "tail-to-tail" linkage of monomer RF DNA Thelinear duplexRF DNA oftheparvovirus attheir leftends rather than atandem union. H-1replicatessemiconservatively andprovides Thiswasrevealedby digestion of the dimer RF

the template for the asymmetricprogeny DNA DNA with

EcoRI,

and the EcoRI fragments

synthesis (17, 18). Thelatterprocess was found obtained included twoEcoRI-Amoleculesanda

torequire at least one of thecapsid proteins by new fragment equal to twice the molecular

analysis of temperature-sensitive capsid pro- weight of EcoRI-B. Sedimentation of dimer RF

tein mutants of H-1 (19). In this study, the DNA in alkaline sucrosegradients showed that

cleavage of H-1 RF DNA by the bacterial re- most of the DNA sedimented as monomer

strictionendonucleaseEcoRI is shown to occur length single strands. Thus, the dimers were

at a unique cleavage site 0.22 genome length inferred to be linkedby hydrogenbonds in the

from theempirically designated left end of the self-cohesiveterminiatthe left end of monomer

molecule. The smaller EcoRI fragment B was RF DNA. Incubation of RF DNA in50%

form-heterogeneousin electrophoreticmobility, and amide for 4 h at 37°C resulted in a slight

de-anumber ofalternative structures were shown crease in thepercentage of the preparation that

to exist in this fragment. The differences be- wasdimerlength, revealing that the majority

on November 10, 2019 by guest

http://jvi.asm.org/

1

2

3

4

_w

*-4---DI

RF

_RF

a,-

R

I

A~~I

L.mp

---Di

RI B

,--

.--RI

B1

,

_

-

RIB2

FIG. 15. Slab-gelelectrophoresis ofEcoRI

digests

of

RF DNA.RFDNA labeled with 32Pwas

digested

for

3

hat37°Cwith 100 UofEcoRI. Thedigestwasextracted with

phenol,

precipitated

with ethanolat

-20'C,

and

redissolvedin 20M Tris(pH

7.5)-i

mM EDTA.

Aliquots

were

separately

fractionated

by

alkaline

denatura-tion-BDCorneutral BDCchromatography.Aportion

of

32P-labeleddimer RF DNA

(cf.

pool

B,

Fig.

3)was

alsodigestedwith EcoRI.Electrophoresiswasfor18hat

16°C

ataconstant

voltage of

50V.

Samples

were:(1)

EcoRIdigestofRFDNA; (2)1MTBS eluateofalkalinedenaturation-BDC

chromatography

of

(1);(3)1 M TBS eluate ofneutral BDCchromatography of (1); (4) EcoRI

digest of

dimer RFDNA. Viral DNA and

fragments are labeledasfollows: dimer RFDNA,

DIRF;

RFDNA, RF; EcoRI-A, RIA;dimer

EcoRI-B,

DIRIB; and

EcoRI-B,

and-B2,RIB,andRIB2,respectively.

ofthe monomer RF DNA molecules do notdi- 15to25% of thetotalat severaltimes

p.i.

and

merize invitro. alsowith

pulse-chase labeling. Experiments

us-Dimer RF DNA was found to be

relatively

ingthe

density

label BUdRweredoneto

deter-constantin

proportion

tomonomerRF DNAat minewhether dimer moleculesmayform

by:

(i)

on November 10, 2019 by guest

http://jvi.asm.org/

[image:16.501.42.441.44.508.2]

b' a'

Eco RI B2 b 4- iC

b a a b b' a'

Eco RIB, _______________

a' b' b a

dimer Eco RI B a b b a'C a

b'

b a

sI .. . I . . . ..

100 0 100 990

Base Pairs Eco RI A

cleavage

site

FIG. 16. Modelfor the structure proposed for the left end ofH-i RF DNA. Letters representterminal sequences of 50 to 70 nucleotides, and sequence complementarity is indicated by a primed letter. The structuresshownareillustrated withagap(arrow) betweenadjacent terminiofthe V and Cstrands,whichin

some cases iscovalently closedasindicatedby a dot.

random annealing ofmonomers with cohesive AAV DNA(23). It wassuggestedthat covalent

left termini, perhaps in vitro; (ii) replication of linkage between plus and minus strands of

previously formed dimer DNA; or (iii) failure of AAV may be the result of the 3' end of one

replicature intermediates of monomer RF DNA strandacting as the primer for the initiation of

to segregate. The results showedthat dimer RF the C strand synthesis, as had also been

pro-DNA incorporated the density label more posed previously for MVM DNA (26). If this

slowly than monomer RF DNA (i.e., became were the case for H-1, then newly labeled RF

hybrid density)andruledoutnonsegregation of DNA would be expected to contain relatively

monomerRI moleculesastheimmediatemech- moreof the covalentlinkage thanthe total RF

anism of formation ofdimer DNA. The com- DNA pool. In the present study the contrary

plexity of densitylabelingwithlabeling periods resultwasobtained; RF DNAlabeledfor 3.5 h

greater than 30 min precluded a definitive dis- followedby a 1-h chase had a higher proportion

tinction between (i) possible formation of la- ofthe covalently linked V to C strands than

beled dimers by random association of mon- DNAlabeled for the 0.5 hprecedingthe extrac-omerRFDNA,or(ii) by replication of parental tion.

dimer DNA. The location of the covalent linkage was

Tattersall et al.reported whatappearedtobe shown to be in the EcoRI-B fragment, and it

intramolecular reannealing ofMVMRF DNA seems mostlikely that it is locatedat the left afterheatdenaturationat100°Cinasolution of endin an invertedself-complementary "turna-moderate ionic strength (27). In a subsequent round" suggested by the dimerizationobserved

report inthisseries, we will describe evidence occurring atthissite. Dimer RF DNA was

ana-foraregionattheright end of RF DNA witha lyzed in asimilar manner, and covalent link-veryhigh meltingtemperature, which, if pres- ageofVtoC strandswasfoundfor at least 19% ent in MVM DNA, could have accounted for ofthepopulation. This linkagewas again dem-incomplete denaturation at

1000C.

Infact, the onstrated to occur in the EcoRI-B fragmentand alkaline sucrose

gradient analysis

of MVM RF probablyatthemolecular left end of RFDNA. DNA did not show the expected proportion of The presence of dimer length V-C single single-stranded DNA greater than monomer strands in some of the dimer RF DNA

mole-length (26). Nevertheless, a morerigorous test cules strongly implies that atleast this type of

forcovalentlinkageof VtoCstrandsinH-1 RF dimerexistedinvivo, whatever themechanism

DNAwas carried out usingalkaline denatura- of their formation.

tionfollowed byrapidneutralization andBDC Finally, thedistributionof isotope in the two

chromatography to separate double-stranded EcoRI fragments, after short labeling times,

DNA that had reannealed rapidly from DNA wasexaminedtogaininsight as to the location

with single-stranded regions. After two cycles ofthe origin for replication of RF DNA and ofthis procedure, a fraction was isolated that progeny DNA synthesis. When isotope is first

represented19% of thetotal RF DNA, and this introduced into the precursorpools, it is

incor-fraction sedimented as a dimer length single- porated into RI molecules. After a labeling pe-stranded DNA in alkaline sucrose gradients. riod that does not greatly exceed the time

re-During the preparation of this manuscript a quired for a round of replication, RI molecules

similar structure wasreportedfor intracellular that have completed replicationduring the

on November 10, 2019 by guest

http://jvi.asm.org/

[image:17.501.137.372.66.188.2]