Analysis of the termini of the DNA of bovine parvovirus: demonstration of sequence inversion at the left terminus and its implication for the replication model.

Analysis of the Termini of the DNA of Bovine

Parvovirus:

Demonstration of Sequence Inversion

at

the Left Terminus and

Its

Implication for the Replication Model

KATHERINE C. CHEN,* BRUCE C. SHULL, MURIEL LEDERMAN, ERNEST R. STOUT,

AND ROBERT C. BATES

Department of Biology, VirginiaPolytechnicInstituteandState University, Blacksburg, Virginia24061 Received29October1987/Accepted6July 1988

The distribution of terminal-sequence orientations in the viral DNA of bovine parvovirus (BPV), an autonomousparvovirus,wasstudiedby endlabeling and restrictionenzymedigestion and also by cloning. The

left (3') end of theminus strandofBPVwasfound intwoalternativesequenceorientations(designatedasflip

andflop, whichare reversecomplementsof each other), witha10-foldexcessof flip. This is incontrasttothe autonomous rodent parvoviruses whichencapsidate'minus-strand DNAwithonly the flip orientation atthis

end. Theright(5') end of the minus strand ofBPVcontainedbothsequenceorientationswith equalfrequencies, asin the rodent parvoviruses. Sequenceinversionswerealsodetected'atboth endsof the plus strand,'which

makesupabout10% ofthe encapsidated BPV DNA. Each terminus of BPV DNA hadacharacteristic ratioof

flip toflop forms, and thisratiowasrestored in the progeny DNA resulting fromtransfection withgenomic

clones of different defined terminal conformations. Replicative-form DNA showed the same distribution of

terminal-sequence orientationsasthe reannealedplus andminusvirion DNAs,suggestingthatthedistribution offlip and flop forms observedinvirionDNA isnotduetoselectiveencapsidation, but rathertothespecific distribution ofreplicative forms. Thecurrentreplicationmodelforautonomousparvoviruses,whichwasbased onthe available datafor the rodent parvoviruses,cannotaccountfortheobserveddistribution ofBPVDNA.

Analternativemodelis suggested.

Parvoviruses are small icosahedral viruses with

single-stranded DNA genomes containing terminal palindromic sequences.Theseterminal regions have cis signals important in various steps of replication (6, 8, 9, 14). A number of related models have been proposed for the replication of parvoviruses (1, 3, 7, 8, 14, 16, 21, 30, 31). An essential feature included in all these models is a hairpin transfer

mechanism originally proposed by Cavalier-Smith (11) for thereplication of linear molecules containing terminal palin-dromes. According to this mechanism, ifa terminal

palin-dromeisimperfect,twoterminal sequenceswillresult from thereplicationprocessand each will be thereverse comple-ment of theother, arbitrarily assigned asfliporflop (3, 25).

Ifarestrictionenzymerecognitionsequencelies withinsuch anunpaired region in thesingle-stranded DNA, that

restric-tion enzymewill bediagnostic for the flip and flop

orienta-tions, since itwill cutdouble-stranded DNAattwo alterna-tivesitesdepending onthe orientation (21).

The dependent parvovirus, adeno-associated virus (AAV), encapsidates equal amounts of plus- and minus-strandDNA. The left andright terminiareidentical, and the two alternative sequence orientations are found withequal

frequenciesatboth endsof theplus- and minus-strand DNA, consistent with thehairpin transfer mechanism (8, 17). The autonomous parvoviruses encapsidate various amounts of

minus-strand DNA, from 99% for the rodent parvoviruses, which include minute virus of mice (MVM), H-1, and rat virus,to90% for bovine parvovirus (BPV),to50% forLuIlI (14, 28). The left and right termini of the autonomous

parvovirusesarenotidentical. Theterminal-sequence

orien-tations have sofar beenstudied indetailonlyfor therodent

parvoviruses (2-4, 24). The right (5') termini ofthe minus

*Correspondingauthor.

strands of these viralgenomes have been shownto exist in either orientation with equal frequencies, but the left (3') termini areunique. Amodified rolling hairpin model (1, 14)

was developed to account for the absence of sequence

inversion at the left end. However, the generality of this model for the otherautonomousparvoviruses remains tobe determined.

In thisstudy, the distribution of terminal-sequence orien-tations in the viral DNA population of BPV, a nonrodent parvovirus, wasinvestigated by end labeling and restriction enzyme analyses and also by cloning. The results

demon-strated the presence of both flip and flop conformations at

the left end of the minus strand ofBPV, incontrast to the

unique conformationatthis endfor therodentparvoviruses.

The relation of this finding to the models of parvovirus replication is discussed.

MATERIALS AND METHODS

Cloneconstruction. The construction and characterization

of BPV infectious genomic clpnes pVT501, pVT502, and

pGCSma20 have been described previQusly (27). Left and

right terminal fragments ofBPVwerecloned intopUC8,and

terminal SmaI fragments were cloned'into M13mpl8 by

standard procedures (18).

End-labelanalysis.Procedures forisolatingDNAs (single-and double-stranded virion DNAs, as well as

replicative-form[RF] DNA)weredescribedpreviously (12). DNA,1 to

2 pugper50-,ulreaction, was'labeledatthe3'-hydroxylends

with 50 ,uCi of [35S]ddATP (Dupont, NEN Research

Prod-ucts,Boston, Mass.)and 20U'ofterminal

deoxynucleotidyl-transferase(Dupont)inas-olutioncontaining100 mM potas-sium cacodylate, 25 mM Tris hydrochloride, 0.2 mM dithiothreitol, and 1 mM CoCl2(pH-7.6) (23). After incuba-tionat37°C for2h,DNAwasprecipitatedandwashedthree 3807

0022-538X/88/103807-07$02.00/0

Copyright© 1988, American Society for Microbiology

on November 10, 2019 by guest

http://jvi.asm.org/

3808 CHEN ET AL.

times

with

70% ethanol.The labeled DNAwas

digested

with

the restriction enzymes

diagnostic

for the

flip

and

flop

orientationsasindicated in Results.The

digestion

products

were

analyzed

on a

sequencing gel (10) by using M13mp18

sequencing

reactions as size markers. End-labeled restric-tion

fragments

would be one base

longer

than the sizes

predicted

from the DNA sequence as a result ofthe extra

base

(ddA)

added

by

theterminal transferase.

RESULTS

BPV

encapsidates approximately

90% minus-strandDNA

and 10%

plus-strand

DNA,

and virion DNAs of

opposite

polarity readily

reanneal under the conditions used for

isolation

(12).

VirionDNA was recoveredintwofractions:

double-stranded

DNA,

consisting

of reannealed

plus

and

minus

strands,

and

single-stranded

DNA of minus

polarity

only.

Weassumethat theminusstrands in reannealedvirion

DNAare a

representative sample

of the total minus-strand

population. By convention, parvoviral

genomes are pre-sented withthe

coding

sequence

(plus strand) reading

from

leftto

right; thus,

the 5' endofthe

plus

strandorthe 3' end of the minus strand is referred to as the left end. This

convention will be followed heretodiscuss the distribution

(the

proportion

of various components in a

population)

of

terminal-sequence

orientations ofBPV DNA.

Sequence

heterogenity

oftheleftendofBPVdemonstrated

by

cloning

of terminal

fragments.

Toobtain a

population

of cloned terminal

fragments,

reannealed double-stranded

vi-rion DNA of BPV was used to generate clones of the left-terminal

(map

units 0

through 17)

and

right-terminal

(map

units92

through

100)

EcoRI

fragments

ofBPV.For the

right

end,

clones with either

flip

or

flop

orientations were

readily

obtained. For the left

end,

most of the clones

obtained were in the

flip

conformation.

Surprisingly,

clone

p3'V8

containing

a

34-base-pair (bp)

deletionwas found to

be in the

flop

conformationatthis terminus. This conforma-tion hadnot been observed from otherautonomous parvo-virus genomes.

Sinceasequenceinversion withintheleft terminus ofBPV could occur

during

cloning

as anartifact

(27),

itwas

neces-sarytoconfirm thepresence ofthe

flop

conformation in the native DNA

by

other means. The

unique

SmaI

recognition

sequence

(CCC/GGG)

ofBPVlies withinan

unpaired region

of the left

hairpin (Fig.

1A). Hence,

the enzyme will cut

double-stranded

DNA at either 44 or 106 nucleotides

(nt),

depending

on whetherthe DNA is in the

flip

or

flop

form. The

ability

to clone a left terminus of 106 nt from the

double-stranded virion DNA after

digestion

with SmaI

would

provide

direct evidence for the existence of the

flop

conformationatthe left endofboth strands of DNA. When

aseriesofterminal

SinaI

cloneswere

sequenced,

onewitha

96-bp

insert was found. This

clone,

M13Smal99,

had the

expected

sequence for the

flop

conformation but with a

10-bp

deletion at the terminus.This

finding

provided

direct

proof,

by cloning,

of the existence of the

flop

conformation

attheleft endofboththe

plus

andminusstrands. Sequence inversion during

cloning

could not

give

rise to a 96-bp

fragment

froma

44-bp fragment.

Demonstration ofboth

flip

and

flop

conformationsatthe left

(3')

end

of

the minus-strandDNAby end-labelanalysis. The occurrenceofthe

flip

or

flop

formscanalsobevisualizedand

the relative amounts ofeach can be assessed by end-label

analysis

asdescribed in

Materials

and Methods. On thebasis

of the sequence data

(12),

the left end of the minusstrandof

BPVDNAcanexist inamoststable

(maximum

basepairing)

Y-shaped configuration (Fig.

1A).

There is a

unique

HhaI (GCG/C) site inoneof itsarms. Endlabeling(whichaddeda

ddA basetothe3'-hydroxylend and made the

fragment

one

baselonger)of the minusstrand,followedby digestionwith

HhaI, shouldyieldbands of 77 and 66 nt,correspotidingto

theflipandflop conformations, respectively.

Whenthe single-stranded BPV DNA(minuspolarity

only)

was analyzed in this way, bands at 79 and 64 nt, which corresponded well with the expected lengths, were ob-served, along with a number ofsecondary bands (Fig.

2).

The band corresponding to theflip conformation at the 3' terminus was about 10-fold more intense than that

corre-sponding to theflop conformationatthe 3' terminus. There-fore, roughly90% of the minus strands had the flip orienta-tion, and only 10% had the flop orientation at the left end. HhaI tends to make single-strand cuts within the

duplex

region containing the recognition site (3). Weaker bands in the end-label analysis (Fig. 2), at nt 71 and 83, agreedwell withthe expected values of 73 and 84predictedfor

cleavage

atthe distalsides of theHhaI sites(Fig. 1A).Bandsat77,

78,

and80nt(Fig. 2)werethe result of variations in the

position

of the terminal nucleotide at the left end, as reported for other parvoviruses (3, 15). The discrepancies between the expected terminal-fragment lengths and the observed ones

aredue, at least in part, to the effect ofsecondarystructure

on themigration of thesefragmentsinasequencing

gel.

Equalfrequencies of flip and flop forms at the right(5')end ofthe minus strand.The conformation of theright (5')end of the minusstrand can be demonstratedby end-labelanalysis

of invitro-replicated minus strand. In vitro replication would yield a double-stranded DNA whose conformation at the right end is defined by the conformation of the minus strand. NdeI(CA/TATG) digestion would cut the invitro-replicated

BPVDNA at nt 5437 inthe flip orientation and at nt 5422 in the flop orientation (12). Replicated BPV virion DNAwas

thus analyzed, and two bands of 55 and 73 nt, with approx-imately equal intensities, were observed (12).

Theright terminus of BPV was previously assignedto nt

5491 on the basis of the data mentioned above and the sequenceanalysis of cloned Si nuclease-resistantfragments

ofsingle-stranded virion DNA(12). However, in vitro rep-lication of palindromic DNA by Escherichia coli DNA

polymerase I Klenow fragment is often incomplete and arrests at specific sites that are capable of forming hairpin

structures,aspreviously reported by Cotmoreand Tattersall

(13)and asdescribedbelow. Duringcloning both of theright

terminus of BPV and of infectious genomic clones from

double-stranded virion DNA (27), several of the clones obtained had therightterminus 26bp beyond thepreviously assignedterminus ofthe genome. Also, one clone of theSi

nuclease-resistant right terminus containing the same 26 additionalbp wasfound(K. C. Chen, unpublishedresults). In

addition,

the fragment lengths of end-labeled,

NdeI-digesteddouble-stranded virion DNA were found to be 20 to 22bp longerthan thoseobtainedfrom the in vitro-replicated DNA(seebelow). On thebasisof this evidence, the revised terminus of BPV isplacedatnt 5517 (Fig. 1B).

Even thoughthe terminal fragments of NdeI-digested in

vitro-replicated minus-strandDNA were truncated because of

incomplete

replication,thebandintensitiesindicatedthat both theflipandflopforms occur with equal frequencies at the

right

end ofthe minus-strand virion DNA, as demon-stratedfor otherparvoviruses (1, 17, 21, 24). This result is

expectedifthehairpintransfer mechanism were involved in the synthesis of minus strand with the plus strand as the

templateand its right-endpalindrome as the self-primer.

J. VIROL.

on November 10, 2019 by guest

http://jvi.asm.org/

(A) Left (3') terminus (1) Flip conformation

A 80-A G

G C,_ C G HhaI

G C 100 120 140 160

C G

IIII

G CGCGTAACA'CCGGATTAGCCCGTCGGTTATACCACCTACGTCCTTTACTCATTATTTTTATATCTCGGACAC---5'

C

A TCGCATTATT-GGCC----CGGGCAGCCAATATGGTGGATGCAGGAAATGAGTAATAAAAATAT 3'

G C\

C G 60 SmaI 40 20 1

G C G C C G G A C

(2) Flop conformation G

C T G C C G C G

80-G C 100 SmaI 120 140 160

C G T

T AGCGTAATA-ACCGG----GCCCGTCGGTTATACCACCTACGTCCTTTACTCATTATTTTTATATCTCGGACAC- ' G

C GCGCATTGTAAGGCCTAATCGGGCAGCCAATATGGTGGATGCAGGAAATGAGTAATAAAAATAT 3'

C G 60 40 20 1

G C HhaI C G T C T

(B) Right (5') terminus

(1) Flip conformation

5341 5360 5380 5400 5420

3'--GAAATATAAAAATATGTGACTAAGTGAGAAAGGTGGTAGAAAATCAACATAAGAATCAATAGTTATTTCCGCGCTTCGCGGTATTTTTTA A 5'

fTATTTTTATACACTGATTCACTCTTTCCACCATCTTTTAGTTGTATTCTTAGTTATfAATAAAGGCGCGAAGCGCCIATATAT

5517 5500 5480 5460 5440

NdeI

(2) Flop conformation

5341 5360 5380 5400 5420 NdeI

II

I

I

I

A T T

3'--GAAATATAAAAATATGTGACTAAGTGAGAAAGGTGGTAGAAAATCAACATAAGAATCAATAGTTATTTCCGCGCTTCGCGGTATACTCTA T

5'

ATATTTTTATACACGGATTCACTCTTTCCACCATCTTTTAGTTGTATTCTTAGTTATCAATAAAGGCGCGAAGCGCCATAAAAAAT

5517 5500 5480 5460 5440

FIG. 1. Nucleotidesequencesof theflipandflopconformationsatthe left(3')terminus(A)andtheright (5')terminus(B)oftheminus strandof BPV. RestrictionenzymerecognitionsequencesforHhaI,NdeI,andSmaIareindicated. Thepositionof the

previously

assigned

right (5') terminus (*) is shown.

on November 10, 2019 by guest

http://jvi.asm.org/

[image:3.612.90.534.82.655.2]

3810 CHEN ET AL.

GTM1 3

|A C G T|

_~_

f jp -_

f

Ilip_--flop

flop

-83nt (84)

--79nt (77)

71n t

-64nt

(73 )

(66 )

FIG. 2. End-label analysis of the left terminus of the minus

strand of BPV. Single-stranded virion DNA was end labeled, digested with HhaI, and analyzed as described in Materials and

Methods. Lane 1 contains wild-type virion DNA. Bands corre-spondingtotheffipandfloporientations areindicatedontheleft.

The observed and the expected (shown in parentheses) sizes are indicatedontheright.

Demonstration ofsequence inversions at both ends of the

plus strand by end-label analysis of double-stranded virion

DNA. The distribution ofterminal-sequence orientations in

theplus-strand virion DNAcanbededuced from the analy-sis of double-stranded (reannealed plus and minus strands) virionDNA, provided thatthedistributionwithin the minus strandis known, asillustrated below.

Reannealing ofBPV virion DNAplus and minus strands

would result indouble-stranded molecules whichare either

fullybasepaired (in the fliporflop conformations)orpartly

unpaired in terminal regions because of various

terminal-sequence orientations. The former would be cut by the diagnostic restriction enzyme and produce fragments char-acteristic of the sequence orientation, but the latter would

not. Forarestrictionenzymetogenerate afragment witha

given conformation, either flip (i) or flop(o), both plus (c)

andminus (v) strands in the double-stranded (d) DNA have

to be inthesame orientation. Theprobability (P) of cutting

atsuchaconformation is the product of theprobabilities that eachstrand is inthatconformation,asgiven byPd(i)=

PC(i)

x P,(i) andPd/o) = Pc(o) x Pv(o), wherePd(i) denotesthe

probability of double-stranded DNA being in the flip orien-tation. The ratio of flip to flop observed in the double-stranded DNA is then the product of the ratios for the individual strands,

Pd(i)/Pd(o) = [P,(i)/P,(o)] X

[PA(i)IPA(o)]

(1)

Endlabeling and SmaI digestion ofdouble-stranded,

rean-nealed virion DNAfrom wild-type virus gaveamajor band

at 42 ntand a minor band at 103 nt(Fig. 3A, lane 1), which

corresponded well with the expected 45 and 107 nt for the flip and flop conformations at the left end, respectively. Bands at 40, 41, and 43 nt were the result of terminal

variations of the leftend,as observedwith HhaI digestion. Theexistence ofabandcorrespondingtotheflop

conforma-tion ofthe left end from double-stranded DNA provided

additional evidence for its presence in the minus strand. The relative intensities indicated that the flip conformation was predominant at the left end. The flip-to-flop ratio in double-stranded, reannealed virion DNA was about the same as that

observedin theminus strand after digestion with HhaI(Fig. 2). Sincethe two ratios were about thesame, this suggested, from equation 1, that theratio for the plus strand mustbe closeto 1.Thatis, for the virion DNA thatwasencapsidated

as plusstrand, half of itwasin theflip orientationatits left

end, and the other half was in the flop orientation. This

would be the result expected if plus strands were synthe-sized bythehairpin transfer mechanism with minus strandas thetemplate and its 3'hairpin asthe self-primer.

Similaranalysis oftherightendofdouble-stranded virion

DNAwith NdeIasthediagnosticenzymeshowedtwobands of77 and 93 nt (Fig. 3B, lane 1). With the reassigned right end of BPV, theexpected sizes of theflip and

flop

fragments

would be 81 and 96 nt(Fig. 1B),whichwere in reasonable

agreement with the observed values. The flop band was

about twice as intense as the flip band. Since the minus strand had equal proportions of flip toflop atthis end, the ratio observed in the double-strandedDNA mustresult from

the flip-to-flop ratio in the plus strands (see equation 1).

Hence, for the plus strand, roughly one-third of itwas

flip

andtwo-thirds of it was flopatthe right end.

Terminal structures of BPV RF DNA. In vivo RF DNA (mostly of monomer length) (22) was isolated from BPV-infected cells 20 to 24 hpostinfection byamodification of the Hirt procedure (29). This DNA was subjected to end-label

analysis by usingSmaI for the left end and NdeIfortheright end. The restriction fragments from these reactions were

comparedtothoseobtainedbydigesting

reannealed,

double-stranded virion DNA with the same enzymes

(Fig.

3A and

B).The ratio offliptoflopateach terminuswasthesamefor

both RF and double-stranded virion DNA. Selective

encap-sidation could not be responsible for the observed ratios of terminal-sequence orientations in virion DNA, because the same ratios of flip to flop forms were observed in the RF

population. Furthermore, the sizes of terminal NdeI

frag-ments obtained from the two DNAs were the same. There

was noextensionof the right endofBPV RF DNAcompared withthe encapsidated virion DNA (1), incontrast to MVM. Fate offlip and flop forms of transfecting genomic clones.

Full-length genomic clones pVT501 (flop at left end) and

pVT502(flip atleft end) have the same right end (in the flip form) and differ only in the orientation at the left end (27). When the left ends ofthe progeny DNA of amplified

trans-fections initiated with either pVT501 or pVT502 were ana-lyzed byend labeling and SmaIdigestion, bothflip andflop conformations wereobserved and theratio ofthe twoforms

wasidentical to that found in a normal infection (Fig. 4). This suggested that, at the left end, one conformation could generate the otherand that the flip form always accumulated preferentially in the viralDNA, regardless of the conforma-tion present in the clone used for transfection. Similar

analyses were performed on a different infectious genomic

clone, pGCSma2O (flipatboth ends [27]), which is deletedby

3 bases at the left end and 35 bases at the right end. Both ends wererepaired towild-type lengthand againtheratio of

J.VIROL.

on November 10, 2019 by guest

http://jvi.asm.org/

[image:4.612.73.280.71.357.2]

TERMINAL-SEQUENCE ORIENTATIONS OF BPV DNA

A LEFT END

B

RIGHT END

__ _

_R

.v_

_5

_ _

__

_ _

_!! __

_

_ _

[image:5.612.75.552.76.406.2]

-..S.

...

-_ _

-_,.,>..

= _

4 _

_w _

_ _

----103 nt ( 107)

flip

ml S1

1 A C G T

am

--fo

cow -..V_

M .M

,. _

000

-93 nt (96)

77nt (81)

- 42 nt (45)

f p

FIG. 3. End-label analyses of double-strandedvirion DNA and in vivoRFDNA.Reannealedplus- and minus-strandvirion DNA (lane 1)

and RFDNA(lane2)wereendlabeledanddigested either with SmaI for analysis ofsequenceinversionsatthe leftend (A)orwithNdeI for

analysisofsequenceinversionsatthe right end (B). The positions and observed and expected (shown inparentheses)sizes ofthe flip andflop

conformationsareindicated.

fliptoflopateach terminus wasindistinguishable from that for wild-type virion DNA (data not shown). The same

phenomenon has been reported forAAV (26) andthe right

terminus ofMVM (cited in reference 14).

DISCUSSION

Thedistribution ofterminal-sequence orientations of BPV virion DNA was analyzed by end labeling and restriction enzyme digestion. The results demonstrated specific ratios offlipandflop conformationsatboth endsfor DNAofeither

plusorminuspolarity.Incontrast,sequenceinversionatthe

left end of the minus strand was absent in all the rodent

autonomousparvoviruses studiedto date (14).

Thepresenceofapproximately equalamountsofflipand

flopforms attherightendsof the minus strands, aswellas at the left ends of the plus strands, suggests that hairpin transferprocesses occurforsynthesisof BPV virion DNA of eitherpolarity.Ineithercase,thecomplementtotheplusor minus strand would serve as the template and self-primer from its 3'-terminalpalindrome.

Transfection with genomic clones of BPV containing

definedconformations atthe endsresulted inprogenyvirion

DNAs with distributions identicaltothatfound inanormal infection. The same ratio ofplus tominus strands (27), as well as the same ratios offlip to flopforms atthe ends of

reannealed virionDNAs (Fig. 4),were observed. This indi-cated that there must be inherent constraints on DNA

replication fortheestablishment of such aspecific distribu-tion. Evidence exists thatbothcellular andviral factors are

involved intheseprocesses. Both H-1 and LuIll can infect NBEcells;however, H-1 encapsidates predominantly minus strands, whereas LuIlI encapsidates equalamountsof plus

and minus strands (5, 19, 20). The cellular influence is

indicatedbythe observation that H-1 encapsidates various

amounts ofplus strand when propagated in different cell

types (S. Rhode, personal communication).

The specificity of the distribution of terminal-sequence

orientations in virion DNA isdeterminedatthe RFstageand

not atthe encapsidation stage(Fig. 3). Thepresenceofthe sameratios offlip to flopforms in RF DNA and in double-stranded virion DNA for BPVarguesagainst encapsidation

of a subpopulation of progeny DNAs. If this were the mechanism for selection, the ratios in reannealed virion DNAwould beexpectedtodiffersignificantlyfrom thosein RFs. The data indicated insteadthat the distribution ofRFs

with different terminal-sequence orientations isresponsible

for the observed distribution ofterminal-sequence orienta-tions in virion DNA. The sameconclusion was reachedfor MVM (1) from the observation that a unique sequence

orientationwasdetected at the left end of both RF andvirion DNA.

1

~~M

13 l

. A C _

*Nw n

.~~~~~~-t am_

flop

VOL. 62,1988 ₃₈₁₁

on November 10, 2019 by guest

http://jvi.asm.org/

3812 CHEN ET AL.

M tj

j,A-:s<

LLOP-t

-FIG. 4. Terminal-sequence orientations at theleft enyDNAresultingfrom amplified transfections with

pVT502. Double-stranded virion DNA waspurified I

turesinfectedwithprogenyvirions ofcellstransfected (lane 1),pVT502 (lane2),orwild-typeBPV(lane3). 1 endlabeledanddigestedwith SmaI.

However, themechanisms forestablishingspe progenyDNAdistributions currently proposed

and MVMreplication (1, 14)are notadequatetc

datafor BPV. The replication mechanism for) allowshairpin transfer withequal efficiencyatb

would predict equal distribution in both RFs

DNAs.

Themodifiedrollinghairpinmodel (1, 14),po:

general model for autonomous parvovirus repli posedamechanismforselectivegeneration ofce

accountforthe distribution of virion DNA inM model,twoobligatory dimerintermediates and

nicks and ligations generate, in equal propc monomerRFswhoseleftendsareinthesameoi thatof theinput virionDNA. ThetwoRFsinterc

onetotheotherwhilegenerating minus strands displacement synthesis onplus-strandtemplate: theconstraints that restrict these RFsto minus thesis werenotindicated.

We have observed a complex distributior strandedBPVDNA. All eightpossible configura

flopateachendof plus orminus strands) occu amounts.Consequently, all fourmonomerRFs(I either end) must also be present, in unequal,

serve as templates forprogenyDNA synthesis.

butionofRFs mustbeestablishedandmaintain

atetheparticular distribution of virion DNAob

To adapt the modified rolling hairpin model to our

obser-vations with BPV, we might propose that the site-specific

nickase-ligase would act on the progeny minus

strand,

as

well as the parental minus strand, of thedimerintermediate.

In this way, it would be possible to generate all four

monomer RFs in the distribution observed for BPV by

postulating certain affinities of the nickase-ligase for these - 0 3 n sites. Butwhenthesetemplate RFs areused for the

synthe-sis of both plus and minus strands (as required for

BPV),

their characteristic asymmetric distribution will be altered,

because thetemplate RFschange fromoneformto

another,

depending on the type of progeny DNA

synthesized.

For example, progenyplus-strand synthesis wouldcausethe left endof parental RF tochangefromfliptoflop orvice versa.

If plus- andminus-strandsyntheses were equally likely, then eventually a uniform distribution ofRFswould be reached.

More likely, plus and minus strands are synthesized with different efficiencies, inwhichcasetheasymmetric

distribu-tion of RFs depends on these efficiencies as well. Thus it

seems that the characteristic distribution ofDNA forms is determined in a complicated fashion by the rates of all the steps in the replication process.

Rather than adapt the modified rolling hairpin model, we have developed a unified kinetic model forparvovirus rep-lication (K. C. Chen, J. J. Tyson, M.Lederman,E. R.Stout,

4 . andR. C.

Bates,

manuscript

in

preparation;

J. J.

Tyson

and

K. C. Chen,manuscriptin preparation) based on the simpler AAVreplication model (8), in which RF and progeny DNAs are synthesized via hairpin transfer. In our model, we assume that the rates ofsynthesis of these DNAforms may bedifferent dependingonthe flip or flopconformationatthe 3' terminus of the template. Dimer RF molecules were end of prog- included as possible intermediates in the generation of pVT501 and monomer RFs. At some point in the infectious cycle, the from cell cul- various forms of RF and progeny DNAs would reach

steady-with

pVT501

statedistributions,based on their rates of synthesis and rates

rhe

DNAwas

of

conversionto otherforms by hairpintransfer. These rate

constants would be characteristic for each parvovirus and

could be modulated bycellular or viral factors or both.

cificRFand For example, suppose that the rates of RF and progeny

forAAV (8) DNA synthesis were the same regardless of the

conforma-)explainthe tion at the ends of thetemplate. Then a uniform distribution kAV, which of RF and virion DNAwould beestablished,asobserved for )othtermini, AAV. To account for BPV distribution, suppose that the and virion rates of synthesis with the flip form at the left terminus were very slow compared with the rates with the flop form at the stulated as a leftterminusor either form atthe rightterminus. Then most ication, pro- ofthe RF would be in theflip form at the left terminus mainly -rtain RFs to because these RFs with the flipconformationat the left end [VM. In that would beused lessfrequentlyastemplates and remain in the

site-specific flip form. Since the left terminus would be used as the primer )rtions, two andtemplateforplus-strand synthesisand the rightterminus rientation as would be used for minus-strand synthesis, mostly minus -onvertfrom strands with the flip form at the left end would be produced byrepetitive from these RFs as a result of the different rates of synthesis s. However, from the ends. This would be the case for BPV, and if the ,-strand syn- rate were even slower for utilization of the flip form of the left terminus, the distribution seen with MVM could be of single- obtained. Withthis kinetic model, the observed virion DNA ttions(flip or and RF DNA distributions of all parvoviruses studied to r inunequal date, including LuIl (N. Diffoot and B. Shull,unpublished

fliporflop at results), can beaccounted for by assumingdifferent sets of amounts, to rate constants for synthesis of RF and progeny DNA with

This distri- differentprimers-templates.

iedtogener- In the rolling hairpin model, dimer intermediates are )served. obligatory and must be processed by asequenceofnicking

J. VIROL.

vmiu

on November 10, 2019 by guest

http://jvi.asm.org/

[image:6.612.63.298.73.393.2]

andligationstepsin ordertogeneratetwomonomer RFs(of MVM) withauniqueconformationatthe left end. However, dimerintermediates may not be obligatory for BPV replica-tion. The four monomer RFs in BPV theoretically can be

generatedfrom any given form of input virion DNA, withor

without a dimer intermediate, by the hairpin transfer pro-cesses. In vivo RF isolated from BPV-infected cells was found tobe predominantly ofmonomer length (22), unlike the large proportion of dimers observed for MVM (32). In

ourkinetic model, dimer RFs are possibleintermediatesbut arenotnecessaryin order to generate aspecificdistribution ofRFs and virion DNA forms. Ifdimer intermediates are formed, they can be processed to two monomers by

stag-gerednickings and repair of both strands.

The observation oftwoalternative orientations at theleft terminus of the minus-strand DNA of the autonomous par-vovirus BPV shows greater diversity in the replication strategy of these viruses than had previously been sus-pected. To account for this diversity, we suggest that the relative abundance of the various viral DNA species are determined by the rate constants for synthesis of RF and progeny DNA from various templates. A more detailed description of this kinetic model forparvovirusDNA repli-cationis inpreparation.

ACKNOWLEDGMENTS

We thank J. J. Tyson for helpful discussions and for critical review of themanuscript.

This researchwassupported by American CancerSocietygrant

MV-220.

LITERATURE CITED

1. Astell, C.R., M. B. Chow, and D. C.Ward. 1985. Sequence analysis of the termini of virionandreplicative forms of minute virus of miceDNA suggests amodifiedrollinghairpin modelfor

autonomousparvovirusDNAreplication.J. Virol.54:171-177. 2. Asteli, C. R., E. M. Gardiner, and P. Tattersall. 1986. DNA sequenceof thelymphotropicvariant of minute virus ofmice, MVM(i), and comparison with theDNA sequenceof the fibro-tropicprototypestrain. J.Virol. 57:656-669.

3. Asteli, C.R., M. Smith,M. B. Chow,and D. C. Ward. 1979. Structure of the 3' hairpin termini of four rodent parvovirus

genomes: nucleotide sequence homology at origins of DNA

replication. Cell 17:691-703.

4. Astell, C. R.,M.Thompson,M.Merchlinsky,and D.C. Ward.

1983. ThecompleteDNA sequenceofminute virus ofmice,an

autonomousparvovirus. Nucleic AcidsRes. 11:999-1018. 5. Bates,R.C., C.E.Synder,P. T.Banerjee,andS.Mitra. 1984.

Autonomous parvovirus Lulll encapsidates equalamounts of plus and minus DNA strands.J. Virol. 49:319-324.

6. Berns,K.I., and R. A.Bohenzky.1987. Adeno-associatedvirus:

anupdate. Adv. Vir. Res. 32:243-306.

7. Berms, K. I., and W. W. Hauswirth. 1982. Organization and

replicationofparvovirusDNA,p. 3-35.InA.S. Kaplan (ed.), Organization andreplication of viral DNA. CRC Press, Inc.,

BocaRaton, Fla.

8. Bems, K. I., and W. W. Hauswirth. 1984. Adeno-associated virusDNA structure andreplication, p. 1-31. In K. I. Berns

(ed.), Theparvoviruses. PlenumPublishingCorp.,NewYork.

9. Berms,K.I.,and M. Labow.1987.Parvovirusgeneregulation.J.

Gen. Virol. 68:601-614.

10. Biggin, M. D., T. J. Gibson, and G. F. Hong. 1983. Buffer

gradient gels and 35S label as anaid to rapid DNA sequence

determination. Proc. Natl. Acad. Sci. USA 80:3963-3965.

11. Cavalier-Smith,T.1974. Palindromic basesequencesand

repli-cation ofeucaryotic chromosome ends. Nature (London)240: 467-470.

12. Chen, K. C., B. C. Shull, E. A. Moses, M. Lederman, E. R.

Stout,and R. C. Bates.1986.Completenucleotide sequence and genome organization of bovine parvovirus. J. Virol. 60:1085-1097.

13. Cotmore, S. F.,and P. Tattersall. 1984. Characterization and molecularcloning ofahumanparvovirus genome. Science226: 1161-1165.

14. Cotmore, S. F., and P. Tattersall. 1987. The autonomously

replicating parvoviruses of vertebrates. Adv. Vir. Res.

33:91-174.

15. Fife, K. H., K. I. Berns, and K.Murray. 1977. Structure and nucleotide sequence of the terminal regions of adeno-associated virusDNA.Virology 78:475-487.

16. Hauswirth, W. M. 1984. Autonomous parvovirus DNA

struc-ture and replication, p. 129-152. In K. I. Berns (ed.), The parvoviruses. Plenum Publishing Corp.,NewYork.

17. Lusby, E., K. H. Fife, and K. I. Berns. 1980. Nucleotide sequenceof the inverted terminalrepetition in adeno-associated virus DNA. J. Virol. 34:402-409.

18. Maniatis, T.,E. F.Fritsch, andJ.Sambrook. 1982. Molecular

cloning: alaboratory manual. ColdSpringHarborLaboratory,

ColdSpringHarbor, N.Y.

19. Muller, D.-E., and G. Siegl. 1983. Maturation of parvovirus LuIll in asubcellular system. I.Optimalconditions for in vitro

synthesis and encapsidation of viral DNA. J. Gen. Virol. 64: 1043-1054.

20. Rhode, S. L. 1978. H-1 DNA synthesis, p. 279-296. In D. C. WardandP.Tattersall(ed.),Replicationofmammalian parvo-viruses. Cold SpringHarborLaboratory, Cold SpringHarbor,

N.Y.

21. Rhode, S. L., and B.Klaassen. 1982. DNAsequence of the 5'

terminuscontaining the replication origin of parvovirus replica-tive formDNA. J.Virol. 41:990-999.

22. Robertson, A. T., E. R. Stout, and R. C. Bates. 1984.Aphidicolin inhibition of the production of replicative-form DNA during

bovineparvovirus infection.J. Virol.49:652-657.

23. Roychoudhury, R., and R. Wu. 1980. Terminal transferase-catalyzed addition of nucleotides to the 3' termini of DNA. Methods Enzymol. 65:43-62.

24. Sahli, R., G. K.McMaster, and B. Hirt. 1985. DNAsequence comparison between twotissue-specific variants ofthe autono-mousparvovirus, minutevirusof mice. NucleicAcidsRes. 13:

3617-3633.

25. Samulski, R. J., K. I. Berns, M. Tan, and N. Muzyczka. 1982. Cloningofadeno-associatedvirus intopBR322: rescue of intact virusfrom therecombinant plasmid in human cells. Proc.Natl. Acad.Sci. USA79:2077-2081.

26. Samulski, R. J., A. Srivastava, K.I. Berns, and N. Muzyczka.

1983. Rescueofadeno-associated virus from recombinant plas-mids:gene correction within the terminalrepeatsofAAV.Cell 33:135-143.

27. Shull, B. C., K. C. Chen, M.Lederman,E.R.Stout, and R. C. Bates. 1988. Genomic clones of bovineparvovirus: construction

and effect of deletions and terminal sequence inversions on

infectivity. J. Virol. 62:417-426.

28. Siegl, G., R. C. Bates, K. I. Berns, B. J. Carter, D. C.Kelly, E. Kurstak, and P. Tattersall. 1985. Characteristics and taxonomy

ofparvoviridae. Intervirology23:61-73.

29. Siegl, G., and M. Gautschi. 1978.Purification andproperties of replicative-form andreplicative-intermediateDNAmolecules of

parvovirus LuIII, p. 315-325. In D.C. Ward and P. Tattersall (ed.), Replication of mammalian parvoviruses. Cold Spring HarborLaboratory, Cold SpringHarbor, N.Y.

30. Straus, S. E., E. D.Sebring,andJ.A. Rose. 1976.Concatemers of alternating plus and minus strands are intermediates in adenovirus-associated virus DNA synthesis. Proc. Natl. Acad.

Sci. USA73:742-746.

31. Tattersall, P., and D. C. Ward. 1976. Rolling hairpin model for replication ofparvovirusand linearchromosomal DNA. Nature (London) 263:106-109.

32. Ward, D. C., and D. K.Dadachanji.1978.Replicationofminute virusof mice DNA, p. 297-313. In D.C. Ward andP.Tattersall

(ed.), Replication of mammalian parvoviruses. Cold Spring

HarborLaboratory, Cold Spring Harbor, N.Y.

on November 10, 2019 by guest

http://jvi.asm.org/