DNA sequence of the 5' terminus containing the replication origin of parvovirus replicative form DNA.

(1)

0022-538X/82/030990-10$02.00/0

DNA

Sequence

of the

5' Terminus Containing

the

Replication

Origin

of

Parvovirus Replicative Form DNA

SOLON L. RHODE III* AND BELIAKLAASSENt

Institutefor Medical Research ofBennington, Bennington, Vermont05201 Received 6 August 1981/Accepted 13 October 1981

The nucleotide sequence of the 5' terminus of the parvovirus H-1 was determined. There are twoorientations of the 242-base-pair terminalpalindrome in nativereplicative form DNA, one inverted with respect to the other. Adjacent tothe terminalpalindrome is an AT-rich regionthat isnoncodingandcontainsa

55-base-pair tandem repeat. The addition mutant of H-1, DI-1, was also

se-quenced in this region and shown to have three copies of the tandem repeat sequence. Similarly, the related parvovirus H-3 contains only one copy of this repeat sequence. This region contains the replication origin for parvovirus replicative form DNA replication. Some of the implications of these results are

discussed.

Thenondefective parvoviruses are small ico-sahedral viruses that contain a linear single-stranded DNA of about 1.6 x 106 daltons (13,

34, 36). The virion DNA is predominantly, but

notentirely, minus strand and is the template for allvirus-inducedmRNA's that have been identi-fied (9, 30). During the replication process, the virion single-stranded DNA is converted to a linear duplex replicative form (RF) DNA (16, 35). This RF DNA undergoes amplification by a semiconservative modeofDNAreplicationthat requiresaviral RF rep geneproduct which has not been characterized (16, 22). The RF DNA serves as atemplate for viral RNAsynthesisand the synthesis of single-stranded minus strand DNA that becomes encapsidated into progeny virions (17, 37).

The virion DNA containsself-complementary sequences atboth its 3' and 5' termini,and RF DNAhas been shown tocontain these terminal palindromes in both extended and foldback

con-figurations (3, 5, 19). The 5'termini of both the plus and minus strandsofRFDNAare covalent-ly bound to a protein(s) whose function is un-known (15). The pools of RF DNA extracted from infected cells are also characterized by a

significantfraction(=25%)ofdimer-length mol-ecules that arelinked at the 3' terminus(18). The 3' termini of several parvoviruses have been sequenced, and evidence was discussed which indicated thatthe orientation ofthe3'-terminal

palindrome is unique, that is, not inverted, in both virions andRF DNA(1, 2).

Inthisreportwepresenttheprimarystructure

of the 5' terminus of the parvovirus H-1 and

t Laboratorium voor Fysiologische Scheikunde, Leiden,

The Netherlands.

portions of the terminal regions of DI-1, an

addition mutant of H-1, and H-3, aparvovirus related to H-1. This region of themolecule has previously been shown by us to contain the

replication origin for RF DNA replication (26,

27). Our evidence indicates that the5'-terminal

palindromeis present in two orientations inRF

DNA, one inverted with respect to the other. This ffip-flop relationship suggests that the mechanism for synthesis of 5' termini of DNA proposed by Cavalier-Smith applies to this end

ofparvovirusRF DNA(4).Finally,wepresenta

model forRF DNAreplicationthatis

compati-ble with the currentlyavailable factspertaining

tothis process.

MATERIAUSANDMETHODS

Materials.Restriction endonucleases were obtained

from New England Biolabs or Bethesda Research Laboratories andwereusedasdirected by the

suppli-er.Thelargefragment of Escherichia coli DNA poly-meraseI(pol I) was obtained from Boehringer-Mann-heim. T4 DNA ligase was from New England Biolabs. The deoxynucleotide triphosphates and dideoxynu-cleotide triphosphates were from P-L Biochemicals, Inc. The HindIII linker and theM13-specific primer werepurchased fromCollaborativeResearch, Inc. [a-32P]dATP was from New England Nuclear.

Cells and viruses. The parvoviruses used in this

study were wild-type H-1, H-1 DI-1, and H-3 (20). Virus and RF DNA were obtainedfrom infected NB

cellsaspreviously described (19). M13 mp2, mp5, and

mp7, and JM103, werekindlyprovidedby J.Messing.

Moleculardoning ofparvovirusDNA. Approximate-ly 20 ,ug ofparvovirusRF DNAwasdigestedwith the appropriate restriction endonucleaseaspreviously de-scribed (19). The reactions were terminated by the addition of 50 mMTris, pH 8.0, 0.1 MNaCl, 1 mM EDTA, and 0.2%Sarkosyl,to afinalvolume of 200

tl,

andwereextracted oncewith phenol andonce with 990

on November 10, 2019 by guest

http://jvi.asm.org/

(2)

phenol-chloroform-isoamyl alcohol (25:24:1). Phenol was removed by ether extraction; the solution was

adjustedto0.3 MNa+with 3 M sodium acetate, pH 5.5, and the DNA was precipitated with ethanol. HindII or HincIl digests were ligated to a l1-mer HindIll linker as described (8). In some cases the linkerswerephosphorylated withunlabeled ATP, and themobility shift of the fragmentduring its isolation by gel electrophoresis providedamonitorof the effective-nessof the linker ligation step. Reactionswerecarried

out in 66 mMTris, pH 7.4, 6.6 mMMgCl2, 10mM

dithiothreitol, 1 mMspermidine,0.5 mMATP,3,uM

linker,50,ugof bovineserumalbumin perml,and 10 to20 Uof T4 DNA ligase per mlat14°Cfor 20 h. The reactionswerestopped byheatingat65°Cfor 10 min. The reactionmixture was adjusted to100 ,ul and 60

mM NaCl-0.03% sodium azide and was incubated

with 10 U ofHindIIl for 16 h at 37°C. The digestion was terminated by the additionof EDTAto20mM, and the DNA was precipitated with ethanol. The restrictionfragments were isolated byelectroelution after fractionation by electrophoresison agelof 5%

acrylamide. Slightlybroadened bands wereobtained

atthegel positions expected after the addition of the linker sequences. After ethanol precipitation, the

DNA fragments were ligated to HindIII-cleaved

pBR322 or, in thecaseofHincIl 83/91, directly into M13mp5by standard methods (8). E.coli HB101 was

transformedbythe methodof Dagert and Ehrlich(6). Recombinant plasmids were identified by screening the Apr Tcscolonies for the proper size insertionsby

gelelectrophoresis.Plasmid DNAwasprepared bythe

extraction method ofGuerryetal. (10),concentrated

by precipitation with 10%o polyethylene glycol,

ex-tractedwithphenol, andpurified by centrifugation in

CsCVlethidium bromide gradients. Parvovirus

frag-ments weresubcloned in M13mp5 bydigestionof 0.5

to1 p.gof plasmid with HindIII and ligation of the

fragmentsto50to100 ngofHindIII-cleavedM13mp5

RF DNA. Wedidnot recoveranypBR322-M13mp5 recombinants.

TheHincIl 95.7/99.9 fragment of H-1was isolated

from pBR322 by gel electrophoresis and restricted

with TaqI; the single-stranded endswerefilled inby

treatment with the large fragment ofE. coli pol l.

HindIll linkerswere addedas described above, and

thefragmentswereisolatedby gelelectrophoresisand

finallywerecloned in M13mp5.

The87/97.5fragmentof H-1wasobtainedbycloning

theTaqI 38/97.5 fragmentin M13mp7atthe AccI site.

RF DNA of this recombinant was restricted with Sau3A, whichcuts H-1 atmapposition 87 and M13 mp7atits Bamsite. The properfragmentwasisolated

by gel electrophoresis and clonedatthe Bam site of

M13mp7.

The strands represented in the M13 clones were

determined by incubating approximately 1 p.g of recombinant single-stranded DNA with 1 pgof H-1 minus strandDNAin SxSSC(SSC:0.15MNaClplus 0.015 Msodiumcitrate)at65°C for1to2h. Plus strand recombinants hybridize to H-1 DNA and show a

marked reduction of mobility in 1.6% agarose gels. Some clones contained tandem repeats of the

frag-ment.Inverted repeatswere notrecovered.Organisms

containing parvovirusrecombinantDNA were

propa-gated under Pl-EK1 conditions in accordance with the Guidelines for Recombinant DNA Research.

DNAsequencing. DNA sequencingwas done by the chain-terminating method of Sanger et al. (23, 24). We used either the 96-base-pair (bp) primer asdescribed (11) or a 12-bp synthetic primer. Electrophoresiswas in 0.3-mm gels of 8% acrylamide at about 50°C (29). Most gelswere 27 by 40 cm, buta few were 15 by60 cm.Samples were applied in two loadings 3 h apart.

RESULTS

Orientation of the 5'-terminal palindrome. The physical map ofH-1 RF DNA was determined for therestriction endonucleases HhaI andTaqI by partial digestions and reciprocal digestions with enzymes having known sites, as previously described (19) (Fig. 1). These studies revealed two smallfragments generated by both of these enzymes that were present in less than equimo-lar amounts and which mapped to the 5' termi-nus of the molecule. This can be illustrated by the double digestion of RF DNA withHhaI and HpaII(Fig. 2). HpaII is known to have a number of cleavage sites, all located in a small area of the 5'-terminal palindrome (19). The twoHhaI

fragments of about 140 and 100 bp were both cleaved by HpaII, localizing them to the right end. Isolation of the HpaI 91/100 fragment and then digestion with HhaI produced four frag-ments of 100, 140, 340, and 380 bp (data not shown). The sums of the first plus fourth and the second plus third of these equal the original

480-bp 91/100fragment. The simplest interpretation of these results is that the 5' terminus has HhaI sites at either 140 or 100 nucleotides from the end in approximately equal proportions.

Fur-thermore, we would predict that although these

sites are within the terminal foldback they are not in a palindromic portion of it which has its axis of symmetry at nucleotides 115 through 120. This interpretation was confirmed by

demon-strating that HhaI did not cleave this region when in the foldback configuration (data not

shown) and more directly by DNA sequencing

as described below. A similar result was

ob-tained with TaqI except that its cleavage sites were located about 135 or 105 nucleotides from the 5' terminus. Thus, these data indicate that the 5'-terminal foldback region ofH-1 exists in

two orientations inverted with respect to each

Hh

I8

0.98

.190

205

TaqI 19.6

53.0 56.5

38 \0 / 72.5

I r/ 38.5

97.4or98.1 97 5

97.5or98.0

FIG. 1. Physical map ofwild-type H-1 RF DNA for the restriction endonucleases HhaI and TaqI. It is assumed thatthe terminalpalindromesareextended, with map position 0 at the 3' end of the minus strand and 100atthe 5' end. TheHhaI sitesat0.8 and 0.98

were obtained from the DNA sequence (1). The

re-maining sites were derived by partial digestion and

double digestions with enzymes with knowncleavage sites.

http://jvi.asm.org/

[image:2.491.254.443.564.598.2]

(3)

ci

Hhoc-; tlp H1jeii

190-t40

-100- _

30- 1

23

-FIG. 2. Electrophoresis patternof32P-labeledH-1

RF DNA fragments produced by: HhaI, HhaI + HpaII, and HpaII. Fragment sizeswereestimated by

usingaHaeIII digest of+X174RFDNAinanadjacent

lane.Electrophoresiswasina15%acrylamide gel with

a 1-cm top layer of 5% acrylamide containing the

sample slots for 6 hat150 V. The gelwasstained with

ethidiumbromide, photographed, and then driedfor

autoradiography.

other such that there are two alternative

posi-tionsforaHhaIorTaqI site thatareequidistant

from the axis ofsymmetry.

DNA sequence of the 5' terminus of H-1.' To

determine

the

structureof the 5'-terminal

palin-drome and adjacent regions ofH-1, we cloned theHindll 95.7/99.9and91/95.7fragmentswith

Hindllllinkers into theplasmid pBR322. Subse-quent subcloning ofthe 95.7/99.9fragment into the single-stranded phage vector M13 mp5

re-sulted in a deletion of about 500 bp from the

recombinantsathighfrequency.Thisinstability

was circumvented by subcloning 95.7/99.9 in two fragments using the TaqI site at 97.5 (for

purposes of discussion only one orientation is

assumedhere),thusdividingthepalindromeinto parts. These fragments were sequenced by the

chain-terminating method ofSanger etal. (24), andarepresentative gelisshowninFig. 3.The

polymerizationreaction encountered avery

in-hibitory G-rich sequence in the palindrome

when the reaction was in the direction 97.5 to

95.7 or97.5 to99.9, and itwas notpossibleto

read the sequence past this point on these

strands. The sequencing reaction was often

slightly ambiguous atthis site on the opposite

strands as aresultofanunderrepresentation of

someof the Gbands,asshownin Fig.3a,buton some gels these regions were clearly deter-mined, e.g., Fig. 3b. Another complication is that the 97.5/99.9 fragment contains a small region with a9-bp palindrome with two HpaII

sites thatyieldedadifferentsequencedepending

onwhich strand of the subcloned fragmentwas

thetemplate (see Fig. 5). Threedifferentclones

were sequencedin the 5' to 3' direction 99.9 to

97.5 and two clones were sequenced from 97.5

toward99.9withthe same result. The sequence

derived in the 99.9 to 97.5 direction predicts a

HpaIIfragment of23 bpthat is cleaved by HhaI.

Thealternativesequence,whichis shown in the

legend of Fig. 5, would have HhaI cleaving a

HpaIIfragment of36bp. The results of such a

digestion (Fig. 2) indicate that the former

se-quence is the correct one. The other sequence

represents aninversionof 28 nucleotides (except

theone underlined inthelegend of Fig. 5). The

sequence change in the fragment must have

occurred during propagation of the original

HindII 95.7/99.9 fragment cloned in pBR322or

during the subsequent subcloning in M13 mp5. This inversion is not tobe confused with that occurring duringtheparvovirus RF DNA

repli-cationdiscussed above.

The HindII 91/95.7 fragments of H-1, the addition mutant, DI-1, and H-3 were similarly cloned inpBR322 and subcloned in M13 mp5. In addition,aRsaI siterevealedby sequencingwas

used to subdivide this region further for the sequencing of H-1 and DI-1. The sequencing strategy used is summarized inFig. 4, andthe sequence for wild-type H-1 minus strand is presentedinFig. 5.

The first seven nucleotides atthe 5' terminus werenotdirectly determinedbut are inferred for

reasonsofsymmetry sinceit ishighlylikely that

they belong tothe palindromic portions ofthe terminus which includes the HindII sites at

positions7and236. Because the 5' terminus was

FIG. 3. (a)Sequencing gelof theHindII-TaqIfragment95.7/98.0(nucleotides135to235)in both directions.

TheG-richregionat188to201producesastrongstoptothepolymerizationreaction,preventingconfirmation of

the sequencebeyond thatpoint(2). ThecomplementaryC-richregionwhich isadjacenttothepile-upbanddoes

notslowpolymerization(3).Reading from bottomtotop,panel1readsnucleotides170to135(plusstrand), panel

2reads 135to178(minusstrand),andpanel3reads 235to180(plus strand;numbersrefertothoseofFig. 5).On

panel2the H-1 sequencebeginsafter the TTGG of theHindIII linker.Inpanel3theoriginal HindIIIlinkerwas

filled.inbypolIandasecondlinkerligatedsothat theH-1 sequencebeginsafter the TTGGAGCTTG.Itwas confirmed by sequencingacrosstheHindll siteat235 andusingtheRsaIfragment (nucleotides157to313)that

oneGwaslost from the linkerduringthefirstcloningofHindII95.7/99.9.(b)Sequencefrom nucleotides 34to 105 across the palindrome at 41 to 57. Symbols: aC, arabinofuranosyl cytosine; A, deoxyadenosine; C,

deocycytidine; G, deoxyguanosine; T,thymidine.

VIROL.

http://jvi.asm.org/

[image:3.491.100.205.65.267.2]

(4)

TERMINUS

2

3

a

CACOGT

Am~~~~~~~~~~

AG

|

K

s

a*G

TTTT^G

9,~~~

-a

CACOGT CA C GT C A C G T

a

1

41,

I

http://jvi.asm.org/

(5)

Toq I HindII Rso I

136 SS 235 312

HindEl Sou3A matedfor the 3'terminus of parvovirus mRNA

43' species(9, 30), buttwoother AATAAAsignals are present in this region. This region is also

FIG. 4. Sequencingstrategy used for determining

the 5'-terminal sequences of parvovirus H-1. The Hindll fragments: nucleotides7to235 and235to488

werecloned inpBR322 withHindIlI linkersand 488to

=880wascloned inM13mp5withEcoRIlinkers.The HindlI-TaqI fragments7-136 and 136-235and HindII-RsaIfragments235-312 and 312-488weresubcloned in M13mp5.The Sau3Afragment136-640wassubcloned

inM13 mp7fromaTaqIfragment136--3,270cloned atthe AccI site inM13 mp7. The symbolssindicates

theG-rich palindrome thatactedasastrong stoptothe

sequencing reactions. Theclone marked withan aster-isk had a region thatwas inverted within it, as

dis-cussedinthetext.

covalently bound to protein and remained blocked after deproteinization, we have not

identified the terminal nucleotide precisely. Its approximate location wasderived from the

ob-servationthat restriction of the terminal HpaII fragment with Hindllcauses itto shift down in mobility by approximately7bp(19).Therefore,

we are placing the 5' terminus tentatively at

sevennucleotides from the first HindII site.

The most likely foldback structure that we

predict for the 5'-terminal palindrome in virion

DNA, usingthe methods of Tinocoetal.(32),is

shown inFig. 6. Thestemof thehairpin,

nucleo-tides 1 to 100 and 143 to 242, is a perfectly

matchedpalindrome, and only the portion

con-taining the asymmetries is shown in Fig. 6. A T-shapedstructureis also possible if thesequences

from 109to 119are pairedto those at89to94 and149to154, but this is lesslikely since it hasa

lower freeenergy.Also, therestriction ofnative

RF DNA with KpnI, which cleaves at 76 and within thepalindrome, produced onlyoneband onagarosegelsattheposition expectedfor the

76/97.1 fragment (data not shown). If the

fold-back were inthe T-shaped structure, the KpnI

sitewouldbeatthe branchpoint ofthe T andnot

cleaved.Thiswouldresult intwobands,namely

the 76/97.1 and 76/100*, as found with the

HaeIII 83.5/100 bands (19). Thesequence

con-firms thepredictedasymmetries oftheHhaI site

at100 andthe TaqI site at136.

The 88/95 region immediately adjacenttothe terminal foldback is a noncoding region ofthe

genomebecause therearechain-terminating

co-dons inallthreereadingframes between

nucleo-tides 600 and 640. A canonical polyadenylate

signalof AATAAA(plus strand) followedby CA

21 nucleotides downstream is present at 277.

This is compatible with the map position

esti-HindII 10 20 30 40 50

51-TTCAGTTGAC CAACTGAACC TATAGTATCA CTATGTTTTT AGGGTGGGGG 60 70 80 K_nI9O HnaII 100

GGTGGGAGAT ACATACGTTC GCTATGGACC AAGTGGTACC GGTTGGTTGC

HhaI 110 HpaII 120 130 TaqI 140 150

GCTCAACCAA CCGGACCGGC TTAGCCGTCTTGTTCGAGCT TAGCAACCAA

HpaII

KpnI 160 170 180 190 200

CCGGTACCAC TTGGTCCATA GCGAACGTAT GTATCTCCCA CCCCCCCACC 210 220 230HindrI 240 250

CTAAAAACAT AGTGATACTA TAGGTTCAGT TGGTCAACTGAATAGACAGT

260 270 280 290 300

AGTCCTTGTT CATTTTAATT ATTTTATTAT TTTTTTGGTCCCTAACATTC

310 RsaI 320 330 340 3S0

AGTCTAAGGG AGTACCAACC AACCACCCAA CCACCCTTTTTATTTCCC>A

360 370 380 390 400

ATACAAAATC TTTTTATATT CTATCTAACA TTAACCAACT AACTATATTT

410 420 430 440 450

ATTTTAAAATACAAAATCTTTTTATATTCT ATCTAACATT AACCAACTAA

460 470 480 HpaI 490 500

CTATATTTAT TCCAAAAGTTCTATTATTTTTTACAGTTAA CCTACCATAT

-j

510 520 530 540 550

TATGTTATTT TTAATAATCTATTATTAAGT GTATATTTTT ATGTTGTAGT

560 570 580 590 600

CTAGTTAGTT TAAGTATTAT GTGAAGCAAACAGAGAAACATAGTTGGTTG

610 620 630 Sau3A

[image:5.491.56.250.56.120.2]

GTTAGTATGT CATGTGAGGC ACAGGTCTAC AAATCAATGGATC-3'

FIG. 5. The nucleotidesequenceof the 5' terminus of theparvovirus H-1. Thefirst fournucleotides are

inferredfrom the expected symmetryof the inverted

repeatencompassing nucleotides 0to100and 143to 242,andwerenotdirectlydetermined. The5'-terminal nucleotide was bound to a protein; its removal by protease digestion leaves a 5' terminus blocked to labeling by polynucleotide kinase. Its location was

estimatedbythemobilityshifttotheHpaIIterminal

fragmentafterdigestionwithHindll. Thesequencesin

thebracketsarethe55-bp tandemrepeatsofwild-type H-1. The viable defective mutantofH-1, DI-1,was

sequenced for themapregioncorrespondingtoHindII

236to 488. DI-1 containedthree copiesof the 55-bp

repeatsequencewithtwointerveningTTA.Similarly,

the parvovirus H-3 was shownto contain only one copyofthis sequence. H-3 also differed fromH-1 at three otherpositionsinthisregion: nucleotides291 =

T, 362 = G, and 479= G. The alternative sequence

obtained from twoclones marked in Fig. 4 withan

asteriskwas

100 HpaII HpaII

GCGCTCA-ACAAGACGGCTAGCCGGTCCGGTT-TaqI GGT-TCGA

This sequence (between the dashes) is the inverted complementof thatinFig.5exceptfor the I,which is

the same. The chain terminators in three reading

frames between 600 and 630 are underlined. The

terminator at 602 ends the only large open reading frame,soitverylikely marks the carboxy terminus of

the H-1coatproteins (Rhode, unpublished data).

Hind 0

7 SS

5'1

J. VIROL.

4 lb64 lo

ke ---w .4

http://jvi.asm.org/

[image:5.491.264.451.119.382.2]

(6)

PARVOVIRUS 5' TERMINUS 995

5'-O---85-GGTACCGGTTGGT

242----] 58-CCATGGCCAACCAI

5' -o---85-GGTACCGGTTGGT1 242----]58-CCATGGCCAACCAJ

FIG. 6. Estimated structure

5'-terminal palindromes in th4

tions(32). The nucleotides 0tc

completely base paired. This s

by a reviewer of this manusc

lowerfreeenergy than the one

us.

distinguished by its richne4

two regions of entirely A-I

and471 to 481. H-1 contain

55 bp (348 to 402, 406 to 460

TTA. The addition mutant

tandem repeats of this seqi intervening TTA, and H-3

copyand is missing the inte

thanthe repeat sequence,H

gous to H-1 since it differe

tionsbetween 235 and488.

DISCUSSII

The fortuitous asymmetr TaqI sites in the

5'-termin;

DNA have demonstrated t]

contains approximately eqi

types ofterminal palindron

complement of the other. was derived from a freshl

finding indicates that thefi ofthe 5'-terminal palindron ess, but it does not prove

t}

each round of replication. been established for the c

adeno-associated virus (12

strongly support the

foldbE

forgenerating5' termini de Smith (4). In the case of the

viruses,this mechanism

app

one end of the DNA, the minus strand. Because ofti

the mere existence of coval toplus strands at the 5' tern ingevidence in favor of a I nism ofinitiation of RFD? 31).

Ahypothetical replicatior

HhaI CCC ratesthe knownfeatures of

parvovirus

RF

repli-rGCGCTC_GA

GGC

cation is outlined in Fig. 7. In particular, it CGCCGAGTTTCTGCCG depends on the

finding

of

Y-shaped replicative

AT

T C intermediates described for

H-1

(26, 27). These

TaqI data

conffict

with the failure of Tseng et al. to

detect Okazaki fragments in nascent

H-1

DNA (33). Resolution of thiscontradiction

will

require further study. The synthesis of parental RF

TaqI DNA is hypothesized to initiate on the native 3'

TNA

G terminus of virion DNA, step 2 (3). Direct

rk6CT&ACAAGACGGc!"\

evidence of this is lacking, but it has been shown

',CGCGAGTTTCTGCCG,

that virion DNA chases into

RF

DNA with a

HhaI

d

9G

covalently

closed

3'-terminal

foldback

(35).

The transcription of parental RF DNA has been

esr

foldthe

totnfigeuraf

inferredbyindirectmethods through its

sensitiv-D85 and 158 to 242 are

ity

to

inhibition by

substitution of

bromode-.tructurewas suggested

oxyuridine

for

thymidine during parental

RF

:ript

and has a slightly

synthesis

(17, 37).

originally

proposed by One principal feature ofthis model is that it utilizes a Cavalier-Smith type foldback transfer process to generate the 5' termini of minus

strands,

which is shown at step 5 or 9. A second ss in AT (73%) and feature is that an asymmetrical processing of the pairs at 261 to 285 minus strand

3'-terminal

palindrome is executed

Is

a tandem repeat of on a dimer substrate for generating the 5' termini ) with an intervening of plus strands. In this way the unique orienta-DI-1 contains three tion of the

3'-terminal

palindrome of minus uence, including the strands in virion DNA and RF DNA is con-possesses only one served as previously reported (1), steps 5 and 6.

~rvening

TTA. Other At step 3 replication is presented as an

1-3is highly homolo-

internally

primed event on the plus strand d at only three posi- template ofRF DNA with the 5' foldback region in its extended duplex configuration. There is no direct evidence to support or refute this

hypo-'ON thesis at this time. This step may be analogous

ies of the

HhaI

and to the D loop of mitochondria. In fact, we are al region of

H-1

RF proposing that the function of the

G-rich

se-hat native RF DNA quence in the plus strand shown as f in Fig. 7 ual amounts of two prevents minus strand synthesis on the 5' side of ne, one the inverted this site as suggested for mitochondrial DNA Since the

RF

DNA replication (25). It is interesting that this site in y cloned virus, this

H-1

is a G-rich palindrome, AGGGTGGGG

lip-flop configuration GGGTGGGA, that is remarkably

similar

to the ne is an active proc- sequence AGGGGGTGGGGGGGT found adja-iat it happens during cent to the origin of H-strand synthesis in rat This result has also mitochondrial DNA (25). By inhibiting DNA

lefective parvovirus synthesis, this signal sequence would ensure ), and both results that the plus strand palindrome remains free to ack transfer process renature as a foldback structure when initiation scribed by Cavalier- and rightward chain extension on the minus nondefective parvo- strand template drives the separation of the ears to apply to only parental foldback region, as shown in steps 3

5' terminus of the and 4.

Progression

of the replication fork toward he foldback transfer, the covalently closed 3' foldback converts the

lentlinkage of minus monomer

RF

to a dimer, step 5. Synthesis of the ninus is not convinc- plus strand would be by a discontinuous method DNA-primed mecha- that presumably uses Okazaki fragments as NA replication (e.g., intermediates.

As an

alternative,

a DNA-primed initiation

amodel that incorpo- using the plus strand 3' terminus as primer

VOL.41, 1982

http://jvi.asm.org/

[image:6.491.46.240.73.217.2]

(7)

1.

b

2.

cQa:

5.

5'-f f' f' f

6. 5'

f f'

f' f a' b' c' c d' a 7. 5'

5f f -a b c c' d a'

IRFrep

f', f

o f f'

L

b a

I

f f _

d" 5'afi

doa

f fU

8.

[image:7.491.110.402.70.437.2]

9.

FIG. 7. Hypothetical model for thereplicationofparvovirusDNA. Theviral DNA is notdrawn to scale. Complementarysequencesarerepresentedbysmalllettersandtheirprimes.Abbreviations: RF rep,transacting

viralproductrequired for RFreplication; VP,viralproteins;N1orN2site-specific single-strandendonucleaseor

nickase; Pol,DNA polymerization; L, ligation; f,thepalindrome sequence GGGTGGGGGGGTGGG. All or

nearlyallofthe5' terminiareboundtoaprotein(15).Step1depictsthestructureof virion DNA. Atstep2the synthesisofparentalRF DNA hasoccurred,andtheparentalRF istemplatefor viral mRNAsynthesis. When RFrep orvirionproteins,orboth,aresynthesized,RFreplicationproceedsatstep3. Initiation isproposedto begin with minus strand DNAjust inboard of the terminal inverted repeat sequences. An early replicative intermediateisshownatstep3.Progressionof thereplicationforktothe 3'hairpinconvertstheparentalRFtoa

dimer RF. A foldback transferprocess proceedsatone ofthedaughter molecule's 5' termini toproduce a5'

terminus for thenascentminus strand insteps 5 and6. Thisisrepresentedbynickase N1cleavingtheplusstrand

toproducea3'-OHprimersite for stranddisplacement synthesiswhich extends the foldback. Inasimilarmanner

nickaseN2cleavesattheboundaryof the 3'hairpinsequences(step 5),but stranddisplacementsynthesisoccurs

atonlyoneof thesesites(step 6).Instep7 thedimer has converted to twomonomerRF DNAmolecules,and the

orientation of the 3'hairpinremains thesame asfor theparentalvirion DNA. RF DNAwiththe extended3' end replicatesasinsteps3 and 4except that twomonomerRFmoleculesaregenerated(steps 7-9). MonomerRF

DNA with the3' end in thehairpin configurationmustrecyclethroughthe dimer intermediatewhen itreplicates.

Thismodelwasformulatedtofit thefollowingexperimentaldata:(i)thereplicativeintermediates ofRF DNAare

double-stranded branched molecules(26); (ii)theorientation of the 3'hairpinisunique (1); (iii)therearetwo

orientations of the 5'hairpin (this work; 1); (iv)dimer moleculesareprominentmembers(25%)of thepoolof replicatingRF DNA(18, 26, 27); (v)asignificantfractionof the 3'foldback structuresarecovalentlyclosed(18); (vi) onlythe extended andfoldbackconfigurationsof the 5' terminus have been identified(19).Arabbit-eared5'

terminushas notbeen detectedby gel electrophoresis of 5'-terminal restrictionfragments. _f'

5 f

_ f

If

J.VIROL.

f 5

f'

http://jvi.asm.org/

(8)

997

would require a mechanism for converting the extendedstructureof the palindrometo a "rab-bit-eared"structure. Wehaveproduceda termi-nal fragment tentativelyidentified as a

"rabbit-eared" structure after denaturation and renaturation of the native terminal fragment. There is no evidence for such a terminal frag-ment in native RF DNA (S. L. Rhode, unpub-lished data).

The foldback transfer processes which are symmetrical at the 5' terminus, that is, indepen-dentof the ffip or flop orientation, and

asymmet-rical for the 3' terminus are shown as a site-specific nicking reaction by nickase N1 or N2

followed by strand

displacement

by the DNA replication apparatus,!°.Therefore, DNA repli-cation itself would drive the strand separations necessary for these reactions. Examination of the reported sequence for the various 3'-termi-nal structures of H-1 and comparison with the appropriate region of the 5' terminus suggests how the template could direct these reactions

(Fig. 8). Sequence 1 is the plus strand sequence attheborder of the 5'foldback with the foldback

turnaroundtothe right. Sequences 2 and 3 are the twoportions of minus strand sequences from thedimer junction, that is, the abc and ad'c of

Fig. 7, and sequence 4 is the 3' terminus in foldback configuration. The minus strand 3'-terminal sequences at the monomer junctions in the dimer are shown one above the other for easeofcomparison with the axis of symmetry at therightindicated by an asterisk. The tentative

nickingandprimer sites for strand displacement

are again indicated by N !$ A most striking feature of all four sequences is thehomologous positions of the possible recognition signals

TGAAC, TGACCAAC, and

TGAA(k)C

with

respect to thepostulated nicking sites. Equally

important, the 3' foldback region of the dimer, which we have suggested is nicked but is not a substrate for strand displacement synthesis, is missing the TGAAAC sequence present in the other two templates. Similarly, the same signal

site in thefoldbackconfiguration(4) is located at an interior loop rather than a duplex site, as

found in the dimer. This model proposes that

this configuration isprocessed byligationrather thannicking. Therefore,themodelpredictsthat a duplex

TGAA(k)C

signal is required for the strand displacement DNA synthesis. Thus, these regions contain the required symmetries

andasymmetries of sequence necessaryto regu-late the foldback transfer mechanism as de-scribed in the model. It is possible that the nicking enzymes N1 and N2 are the same even though the nick sites do not seem to have identicalrelationships to thepostulated

recogni-tion signals. It is alsopossible that the terminal boundproteins are the nickases (15).

This modelpredictsarecyclingofsomeofthe monomer RF DNAs through adimer molecule and allows someof the monomer RF DNAs to

replicatewithout a dimerintermediate, steps 7-9. Dimers arepostulated to beformed by

repli-cation and notrecombination.This is supported

by our finding that the dimersproducedincells doubly infected with standard virus anda

defec-+ strand

Npol

1 5'-TTAAAATGAACAAGGACTACTGTCTATTCAGTTGACCAACTGAACCTATAGTATCACTATGTTTTTAGGGTGGGGGGGTGGGAGATA

Npo1

- strand

2 5'-ACTCTCTGAACCGCTTATCATTTTTAGAACTGACCAACCATGTGAAACGCTAAGTGACGTGTGACGTCCGAAGGCGCG*

Nno pol

- strand

3 5'-ACTCTCTGAACCGCTTATCATTTTTAGAACTGACCAACCATGTTC ACGCAAGTGACGTGATGACGCGCGCTGCGCG*

No N

+no pol

4 5'-ACTCTCTGAACCGCTTATCATTTTTAGAACTGACCAACCATGT ACGC AAGTGACGTGTGA

TGAGAGACTTGGCGAATAGTAAAAATCTTGACTGGTTGGTACAA

TGCG

TTCACAQPCT

FIG. 8. Comparison of the 5'-terminal (1) and3'-terminal(2-4) sequences in theregion of thehypothetical sitesfor the nickasesN1and N2ortheligation (L) of thereplication model of Fig. 7. Sequence1is the plus strand sequenceattheboundary of the5'-terminal palindrome. Sequences 2 and 3 are the minus strand sequencesatthe dimer RFjunction brought into vertical alignment with the axis of symmetry indicated by the asterisk. Sequence

4is the 3' terminus in itsfoldback configuration (1). The # signs below the line mark points of asymmetry where theoneside of the dimer

sequence

(3) or thefoldback sequence (4) does not haveaduplexTGAAAC which we postulatetoberequired for the strand displacement, P°, to occur. The exact site of nicking for sequence 1 awaits theprecise placement of the 5'-terminal nucleotide.

VOL.41,1982

http://jvi.asm.org/

[image:8.491.50.421.438.600.2]

(9)

RHODE AND KLAASSEN

tive deletion mutant are of two types: standard length x2 or deleted length x2 (Rhode, unpub-lished data). This conclusion should be tem-pered by the fact that we cannotprove that the replicating pools of standard and deleted mole-cules are not compartmentalized within the nu-cleus, thus precluding recombination between them(28).

Because of the symmetry within 100 nucleo-tides of the nick site at the 5' terminus, one might expect that the nicking-strand displace-mentmechanism might not distinguish between 5' termini in foldback and extended configura-tions. If this is the case, this process acting on extended termini would strand displace free copies of the 5' hairpin as a wasteful side reac-tionof the replication process, and we intend to search for these. Otherpredictions of this model which can be tested are that the 5' termini of

dimermolecules willbe in the same orientation (step 6, Fig. 7) and that inhibition of DNA

synthesis will block the interconversion of the various terminal structures.

Finally, the model does not show progeny

minus strand synthesis. Virion DNA synthesis couldoccur by aninteractionof anappropriately mature viral capsid with an early replication intermediate (justbefore step 3 or 8) in such a way as to block theantipolar plus strand synthe-sis. Evidence has beenpresented which suggests thatencapsidation proceeds in closeconjunction with the displacement of minusstrand from the RFtemplate (7).

The sequence derived for the 5' terminus of H-1 includes the replication origin which maps to this site by electron microscope analysis of replicative intermediates(26, 27). The sequence data haveconfirmed ourearlierhypothesisthat the addition in DI-1 is a tandem repeat of a sequence that is repeated already in wild-type

H-1 andis present inonlyonecopy in H-3(20).

A similar relationship has been suggested by physicalmapping data for minute virus of mice (MVM) and MVM(i) (14). Itis thusverylikely

that the further additions reported in H-1 DI genomes between map positions 91and95.7are

tandemrepeatsof this same sequence (21). What role this sequence has in DNA replication is unknown, but it is clear fromtheexample of H-3 that only one copy is required for efficient RF DNAreplication. The tentativeorigin region has many featuresreported for otherorigins ofDNA

replication. The noncoding region inboard of the terminalpalindrome 244 to 640 isA-Trich(73%)

and contains a 25-nucleotideATblockat 262 to

286 andan11-nucleotideATblockat472to 482. Italsocontains a23-nucleotide AC stretchat314

to 336. There are true palindromes of eight

nucleotides orlargerat3 to10,25to35,41 to57,

111to119,186 to202,208 to218,233 to240,267

to 278, and 313 to 324. We are attempting to develop a functionalassay for origin activity so thatstructure-function studies can be carried out todetermine which features of this sequence are important in the initiation of RF DNA replica-tion.

ACKNOWLEDGMENTS

We gratefully acknowledge Jeanne Helft for expert techni-calassistance,Virginia Haas forsecretarial duties, and Elea-norSpicer for helpful discussions.

This work was supported by Public Health Service grants CA-25866andCA26801 fromtheNational CancerInstitute.

LITERATURE CITED

1. Asteil, C., M. Smith, M. Chow, and D. Ward. 1979. Structureof the 3'hairpintermini of four rodent parvovi-rusgenomes:nucleotide sequence homology at origins of DNAreplication. Cell 17:691-703.

2. Asteil, C., M. Smith, M. Chow, and D. Ward. 1979. Sequenceof the3'-terminusof the genome fromKilham rat virus, anon-defective parvovirus. Virology 94:669-674.

3.Bourgulgnon,G., P.Tattersall, and D. Ward. 1976.DNA of minutevirusofmice:self-priming, nonpermuted, sin-gle-strandedgenomewith a5-terminal hairpinduplex. J. Virol.20:290-306.

4. Cavalier-Smith, T. 1974.Palindromic basesequencesand replicationof eukaryote chromosome ends. Nature (Lon-don)250:467-470.

5. Chow, M. B., and D. C. Ward. 1978.Comparison ofthe terminal nucleotide structures in the DNA ofnondefective parvoviruses, p. 205-217. In D. C. Ward and P. Tattersall (ed.), Replication of mammalian parvoviruses. Cold SpringHarborLaboratory, Cold Spring Harbor,N.Y. 6. Dagert, M., and S. D.Ehrlich.1979.Prolongedincubation

in calcium chlorideimproves the competence of Esche-richiacoli cells.Gene 6:23-28.

7. Gautschi, M., and P. Reinhard. 1978. Synthesis of a

mammalian parvovirus onBrij 58-lysedcells. J. Virol. 27:453-456.

8;Goodman, H. M., and R. J. MacDonald. 1979.Cloningof hormone genes from a mixture of cDNA molecules. MethodsEnzymol.68:75-89.

9. Green, M., R. Lebovitz, and R. Roeder. 1979.Expression

ofthe autonomousparvovirusH-1 genome: evidencefora

single transcriptional unit and multiple spliced polyad-enylatedtranscripts.Cell17:967-977.

10. Guerry, P., D. J. LeBlanc, and S. Falkow.1973.General methodfortheisolation ofplasmiddeoxyribonucleic acid. J.Bacteriol. 116:1064-1066.

11. Heidecker, G., J. Mesdng,andB.Gronenborn. 1980. A versatile primer for DNA sequencing in the M13mp2 cloningsystem.Gene 10:69-73.

12. Lusby, E., K. H.Fife,andK. I.Berns.1980.Nucleotide sequence of theinverted terminalrepetitionin adenoasso-ciated virus DNA. J. Virol. 34:402-409.

13. McGeoch,D. J.,L.V. Crawford, andE. A.C. Foilett. 1970. TheDNAs of threeparvoviruses. J. Gen. Virol. 6:33-40.

14. McMaster, G. K.,P.Beard,H. D.Engera,andB.Hirt. 1981.Characterization ofanimmunosuppressive

parvovi-rusrelated to minute virus ofmice. J. Virol. 38:317-326.

15. Revie, D., B. Tseng, R. Grafstrom,and M.Goullan.1979. Covalent association ofproteinwith thereplicativeform DNAofparvovirus H-1. Proc. Natl.Acad. Sci. U.S.A. 11:5539-5543.

16. Rhode, S.L.1974.Replicationprocess of theparvovirus H-1.II.Isolation andcharacterizationof H-1 replicative form DNA. J. Virol. 13:400-410.

17. Rhode, S.L.1974.Replicationprocess of theparvovirus H-1. III. Factors affecting H-1 RF DNA synthesis. J. Virol. 14:791-801.

J. VIROL.

http://jvi.asm.org/

(10)

18. Rhode, S. L. 1977.Replicationprocessoftheparvovirus H-1. VI.Characterizationof areplicationterminus of H-1 replicative-form DNA synthesis. J. Virol. 21:694-712. 19. Rhode, S. L. 1977.Replication process of the parvovirus

H-1. IX. Physical mappingstudies of the H-1 genome. J. Virol.22:446 458.

20. Rhode, S. L. 1978. Replication process of the parvovirus H-1. X.Isolation of a mutant defective in replicative-form DNAreplication. J. Virol. 25:215-223.

21. Rhode, S. L. 1978. Defectiveinterfering particles of parvo-virusH-1. J. Virol. 27:347-356.

22. Rhode, S. L. 1977. H-1 DNA synthesis, p.279-2%.In D. Wardand P. Tattersall(ed.),Replication of mammalian parvoviruses. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.

23. Sanger, F., A. R.Coulson,B. G.Barrell, A. J. H. Smith, andB. A. Roe. 1980.Cloning in single-stranded bacterio-phage as anaidtorapidDNAsequencing. J.Mol. Biol. 143:161-178.

24. Sanger, F.,S. Ndcklen, and A. R.Coulson. 1977. DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad.Sci.U.S.A. 74:5463-5467.

25. Sekiya, T., M. Kobayashi, T.SekI, and K. Kolke. 1980. Nucleotide sequence of a clonedfragment ofrat mito-chondrial DNA containingthe replication origin. Gene 11:53-62.

26. Shiger,I.I., and S. L. Rhode. 1977.Replication process of theparvovirus H-1. VII. Electron microscopy of replica-tive form DNA synthesis. J. Virol. 21:713-723. 27. Singer,1.I., and S. L. Rhode.1977.Replicationprocessof

the parvovirusH-1. VIII. Partialdenaturation mapping andlocalization of the replicationoriginof H-1replicative formDNA with electron microscopy. J. Virol. 21:724-731.

28. Shiger, I. I., and S. L. Rhode. 1978.Ultrastructural studies

of H-1 parvovirus replication. VI. Simultaneous

autora-diographicand immunochemical intranuclear localization of viral DNA synthesis and protein accumulation. J. Virol. 25:349-360.

29. Smith, A.J.H.1980. DNA sequence analysisbyprimed synthesis. Methods Enzymol. 65:560-579.

30. Tal, J., P. Ron, P. Tattersall, S.Bratosin, and Y. Aloni. 1979. About 30%o of minute virus of mice RNA isspliced out followingpolyadenylation. Nature (London) 279:649-651.

31. Tattersall, P., and D. C. Ward. 1976. Rolling hairpin model for replication of parvovirus and linear chromo-somalDNA. Nature (London)263:106-109.

32. Tlinoco, I., Jr., P. N. Borer, B. Dengler, M. D. Levine, 0. C.Uhlenbeck,P. M. Crothers,andJ.Gralla. 1973. Im-proved estimation of secondary structure in ribonucleic acids.Nature(London) 246:4041.

33. Tseng,R.Y., R. H.Grafstrom,D.Revie, W. Oertel, and M.Goulian. 1979. Studies on early intermediates in the synthesisof DNA in animal cells. Cold SpringHarbor Symp. Quant. Biol. 43:263-270.

34. Usategul-Gomez, M., F. Al-Lami, S. M. Hopkins, N. Ledinko, and H. Toolan. 1969. Single-stranded DNA from theparvovirus H-1.Virology 39:617-621.

35. Ward, D. C., K. Dlnshaw,and D. K.Dadachanj. 1978. Replication of minute virus of mice DNA, p. 297-313. In D.Ward and P. Tattersall (ed.), Replication of mammalian parvoviruses. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.

36. Ward, D. C., and P. Tattersall (ed.). 1978.Replicationof mammalian parvoviruses. Cold Spring Harbor Labora-tory,Cold Spring Harbor, N.Y.

37. Wolter, S., R. Rkhards, and R. W. Armentrout. 1980.Cell cycle-dependentreplication of the DNA ofminutevirusof mice,aparvovirus.Biochim. Biophys.Acta 607:420-431.

http://jvi.asm.org/