Replication of Bacteriophage Ribonucleic Acid: Analysis of the Ultrastructure of the Replicative Form and the Replicative Intermediate of Bacteriophage R17

(1)

Replication

of

Bacteriophage

Ribonucleic

Acid:

Analysis

of

the Ultrastructure

of

the

Replicative

Form

and

the Replicative Intermediate

of

Bacteriophage R17

NICOLE GRANBOULAN ANDRICHARD M. FRANKLIN

Laboratoire deMicroscopie Electronique, CentredelaRecherchesScientifiquessurleCancer, 94 Villejuif, France,and The Public HealthResearch Instituteofthe City ofNew York, Inc.,

NewYork,NewYork 10016

Receivedforpublication2November1967

A detailed qualitative and quantitative comparison was made of the

ultrastruc-tureofsingle-stranded ribonucleic acid (RNA)frombacteriophageR17and double-strandedreplicative form (RF) and replicative intermediate (RI) fromcellsinfected with this bacteriophage. The nucleic acids were prepared for electron microscopy by the protein monolayer spreading technique of Kleinschmidt. Single-stranded RNAaggregated duringspreadingintheabsenceofurea, whereasRF and RIdid

not. On the other hand, RF and RI appeared to be susceptible to shear during spreading, whereas R17 RNA was not. From the maximal length ofRF, a base

translation of 3.14 A was calculated. This value favors a 10-fold helix model of double-stranded RNA. The same base translation was found for R17 RNA, in-dicating a stacked basestructure for single-stranded RNA spreadin the presence ofurea. RI isa branchedstructure and thebranchesareremovedby ribonuclease

treatment.The branchesarebelievedtobenascentsingle-stranded viral RNA. The

contourlength ofthebranchwasequaltothecontourlength of the main chainup to the branch point, as predicted from theoretical analysis of the replication of

viral RNA. The structure of RF and the main chain of RI was also analyzed by plotting the log (end-to-end distance squared)versuslog (contour length). This dem-onstratedstructuresintermediate instiffness betweenarandom coil andarigid rod. The key role of double-stranded ribonucleic

acid (RNA) in the replication of the genetic materialof RNAbacteriophageshasbeen demon-strated by a wide variety of in vivo and invitro experiments (9, 26, 27). Two species of

double-stranded RNA have been isolated from cells infected with RNA bacteriophage and their properties have been described (1, 10-13). Repli-cative form (RF) isdouble-stranded RNA com-posedofaviralRNAmolecule (+ strand) hydro-gen-bonded to acomplementary RNAmolecule

(- strand). Replicative intermediate (RI) is double-strandedRNAwithnascentsinglestrands.

It is presumably derived from a more complex

structureinwhichthe RNA isprobablyassociated

with RNA polymerase and may sometimes be associated withpolysomes (16).

Studies of the in vitro reaction have demon-strated a conversion of parental RNA to RF, subsequent conversion ofRFtoRI, andfinally a

transfer of label from RI to newly synthesized

si-igle-strandedviral RNA(26).Semiconservative and conservative models for this process have

been published (9). Some of the experimental datafavor thesemiconservative model of replica-tion (7, 8, 24).

Inthispaper, theultrastructural characteristics of RF and RI are analyzed. Since RI contains both single- and double-stranded components, it wasalsonecessarytoinvestigate the ultrastructure of single-stranded RNA. Unfortunately, single-stranded and double-stranded RNA molecules had the same width in shadowed preparations, precluding any unequivocal identification ofthe twocomponentsof RI. However, twoproperties of single-stranded RNA could be exploited in comparative ultrastructural studies. These were

itssusceptibilitytoribonuclease digestionand its

intra- and intermolecular aggregation during

preparation for electron microscopy.

129

Vol.2,No. 2 Printed inU.S.A.

on November 11, 2019 by guest

http://jvi.asm.org/

(2)

GRANBOULAN AND FRANKLIN

Some preliminary data have been publishedon certain of the parameters to be discussed in this

paper (14).

MATERIALS AND METHODS

Preparation ofRNA. Single-stranded RNA from bacteriophage R17, double-stranded RF, and RI

were prepared as previously described (11, 14). RI

wasdenaturedby heatingto95 Cfor 3minin10-3M

potassium phosphate buffer (pH 6.6) with 10-s M

ethylenediaminetetraacetic acid (EDTA) or by heat-ing to 37 C for 10 min in phosphate-EDTA con-taining 85.7%7o dimethylsulfoxide (DMSO) as described earlier (13). DMSO lowers the melting

temperature (T,0) of double-stranded RNA so that

it maybedenatured underverymildconditions (18). In some experiments, RI was treated with

ribo-nucleaseto digest the single-stranded branches (12). Ribonuclease A (Worthington Biochemical Corp., Freehold, N.J.) was used at aconcentration of 0.08

to 0.1 ,ug/ml, and the RI in 0.1 M NaCl, 0.001 M

EDTA, 0.05 M

tris(hydroxymethyl)aminomethane-chloride buffer atpH6.98 (25 C) was treatedfor 10 min at 37 C (11). After digestion, the ribonuclease

wasremoved byasingle phenolextraction (11). Controlsweremadeby mixingRF andR17RNA. Some preparations were simply mixed (2.3 optical density units,at 260m,, for each RNA) and others

were mixed and extracted once with phenol after

mixing.

Preparation for electroni microscopy: dilutioni of nucleic acids before spreadinig. The single-stranded

RNA extracted from bacteriophageR17 wasdiluted

inurea atafinalconcentrationofeither 6or 8M, to

avoid aggregation of the RNA molecules due to

intra- or intermolecular hydrogen-bonding (15; Granboulan, Scherrer, and Franklin, in press;

Scherrer and Granboulan, inpreparation). The

con-centration ofthe single-stranded RNA in urea was

about 50 Ag/ml. RF and RI were diluted to a final

concentration of 10 to 30 ag/ml, either in (i) 6 or 8

M ureaunder thesameconditionsas the viral single-stranded RNA or in (ii) 0.02 M EDTA at neutral

pH. Single-stranded RNA preparations spread from 6to8Murea were always made at the same time in ordertocheck the action ofurea ontheultrastructure

ofRI.

RI which had been denatured with DMSOatlow temperature or by heatinginthe absence ofDMSO

was diluted in the same manner as single-stranded RNA (in 6or 8 M urea and at final concentrations of 2.5or5 jug/ml).

In control reconstitution experiments, samples of RF and single-stranded RNA were mixed in equal concentrations and were then diluted in either 8 M

ureaor0.02 MEDTAaccordingtotheprocedureused

for RF and RI. The same procedure was also

fol-lowedoncomparable "reconstituted" sampleswhich

had been treated with phenol aftermixing. The final

concentrationof these sampleswas 20 MAg/ml. Spreading and shadow-casting. All of the various

preparations of nucleic acids described above were

spread bythe protein monolayer technique of

Klein-schmidt and co-workers (19-22). Several proteins weresuccessfully used to form the monolayer: cyto-chrome c (Fluka, Buks, Switzerland), chymotrypsino-gen(Nutritional Biochemicals Corp., Cleveland, Ohio), diisopropylphosphoryltrypsin (Worthington Bio-chemical Corp.), and lysozyme (kindly provided by G. Cohen, Gif sur Yvette). Lysozyme was adopted for routine studies because it seemed to give the best differentiation between RNA molecules and back-ground. Itwasusedat afinal concentration of 0.01% in ammonium acetate buffer (1.0M, pH 8.0) contain-ing0.05%- isopropanol. The solutions of protein and ammonium acetate buffer were always filtered on membrane filters (average pore diameter, 0.22 us;

Millipore Corp., Bedford, Mass.) before use. De-pending on the concentration of RNA, 0.1 or 0.2 mlof the RNA solution was added to 1 ml of lysozyme in ammonium acetate buffer; 0.2 ml of this final mix-ture wasgently dropped onto ameticulouslycleaned glass slide from which it flowed onto the surface of ammonium acetate buffer (0.015 M, pH 8.0) in a glass dish. The glass slides were cleaned by keeping them ina 1:1 solution of nitric acid in distilled water and then rinsing them with distilled water followed by ammonium acetate buffer (0.015 M, pH 8.0) just before use. They were always handled with forceps. After spreading, the film was picked up on 300-mesh grids covered with a carbon-coated Formvar film. The specimens were dried in alcohol and iso-pentane and were then shadowed with uranium oxideat anangle of 7°at a vacuumof5 X 10-5 mm

of Hg. The specimens were rotated at 50 rev/min during shadowing. Some samples of RI were shadowed with platinum-carbonatthe same vacuum, but with-outrotation.

Measurements. Thespecimens were examined with a Siemens Elmiskop I or a Philips EM 200 electron microscope at nominal magnifications between 15,000 and 30,000. The actual magnifications were determined with the aid ofacarbon grating replica. The lengths of the molecules were measured on prints withamapmeasurer atfinal magnifications of 45,000 to 130,000. Theend-to-end distance was also

determined.

RESULTS

Single-stranded RNAfrom bacteriophage R17.

Although the single-stranded RNA was spread from solutions whichwereapproximately two to five timesas concentrated as RFor RI, the fre-quency ofmoleculesperunitareaofgridwasfar

less thanfordouble-stranded RNA. Thismay be due to the loss of single-stranded molecules in theammoniumacetatesolutionduringthe spread-ing process (14). It isnot duetothe presenceof

nucleases in the protein,since ribosomalRNA is

notlostduringspreading (Granboulan, Scherrer,

andFranklin, in press; Scherrerand

Granboulan,

inpreparation).

All ofthemolecules spread from solutions of6

to 8 M urea werelinear, i.e., noncyclic and non-branched (Fig. 1). An analysis of 117 molecules

130 J. VIROL.

on November 11, 2019 by guest

http://jvi.asm.org/

(3)

REPLICATION OF PHAGE RNA

FIG. 1. Single-strandedRNAfrombacteriophage RI7,spread in the presence of urea. Examples of whole mole-cules (a,b) and ofafragment (b). X80,000.

revealed a modal length between 1.05 and 1.10

,l

(Fig. 2)

and a mean

length

of 1.06 i 0.12

,u.

RF.There was nodifferencein theshape ofRF spread in the presence orabsence

of

urea.Inboth

casesthe

molecules

werelinear

(Fig. 3);

therewas

no tendencyto aggregate inthe absence

of

urea

(Fig. 3a)andnobranchedmoleculeswere

fotnd.

Besides the prominent population of molecules distributed aroundamean

length

of 1.05 i 0.03 ,u anda modal

length

between 1.05 and 1.10

A.,

there were smaller molecules with

lengths

cor-responding

to

one-half

and

one-fourth

that

of

the

former

population

of molecules

(Fig. 4).

An analysis of the shape of the spread RF molecules was made by means of a double-logarithmic plot

of

the mean square end-to-end

distance versus the contour length (19, 20, 22).

Accordingto this analysis, the

molecules

spread

in theabsence ofurea wereintermediate in struc-turebetween a

rigid

rod anda random coil (Fig.

5).

RI.Whether spreadin the presence orabsence ofurea, all

preparations

of RI contained linear

molecules, some of which were branched and othersofwhich were unbranched (Fig.

6-8).

No

circularmoleculeswerefound inany RI

prepara-tions. Branches were attached at all

possible

positions

along

the

presumed

main chain. The

leng.hof the branch varied accordingto thepoint

of

attachment to the main chain

(Fig. 6).

The variation in branch lengthwasverystriking in the

rare cases

of

RIwithtwobranches

(Fig.

7). When

RI was spread in the absence

of

urea, branches located closeto oneendofachainwerefrequently coiled

(Fig.

8c). This configuration was not

observedwhen RI was

spread

in thepresence of

urea.

DataonthebranchesofRImolecules are

pre-sented in Table 1. Similar values were

found

for molecules spread in the presence or absence of

urea. Only 50%

of

the molecules had branches, and almost all

of

thosewith brancheshad only a

singlebranch. Veryrarely,molecules with two or three branches were seen, but none with more than three branches. Inafew instances,RI

mole-culesappearedto have twobranchesat onepoint. Such molecules were examined closely at high

magnification; inevery case,these were two

over-lapping molecules. An example of two

over-lappingmolecules is shown in Fig. 9b, a

prepara-131 VOL. 2, 1968

on November 11, 2019 by guest

http://jvi.asm.org/

[image:3.490.47.440.77.365.2]

(4)

.40

r-R 17 RNA

I-I

I I I I I I I Il I __ X

0.2 0.3 0.4 0.5 Q6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4

[image:4.490.61.459.60.423.2]

l (length inai)

FIG. 2. Histogram ofthelengthdistributionofR17RNAspreadin thepresenceofurea;atotal of117molecules

wereanalyzed. For 1.45to1.50uandfor1.65 to1.70I,afrequency of0.009wasfound. Thesearenotincludedin

thehistogram forlackofspace.

tion of ribonuclease-treated RI spread in the

presenceofurea.

RI which had been treated with ribonuclease had fewor nobranches(Fig. 9).Ofatotal of 295

molecules spread without urea, only 13 (4.4%)

werebranched; ofatotalof361moleculesspread

with urea, only 7 (1.9%) were branched. These

preparations behaved like RF, since there was

notastrikingloss of moleculesduringthe spread-ing process as there was in the case of

single-stranded R17 RNA. Furthermore, the molecules

spreadintheabsence ofureadidnotform

aggre-gates.

The maximal length of the presumed

double-strandedcomponentofRIwasthesameasthatof RF(Fig. 10,11).Thehistogram for881 molecules

spreadintheabsence ofureahadapeakbetween 1.00and 1.10A (Fig. 10).Themeanlength

calcu-lated for the 333 molecules oflength equalto or

greaterthan 1.00,uwas 1.05 i0.06 ,u. The histo-gramfor 809molecules spread inthepresence of ureaalsohadapeak between 1.00 and 1.10 ,u (Fig. 11). Themeanlength calculated forthe 434

mole-cules of length equal to or greater than 1.00 ,u

was 1.05 + 0.04j,. Therewereshorter molecules

present in both types of preparations and there were peaks corresponding in length to approxi-mately one-half and one-fourththe length ofthe

longest molecules.

The shapeorstiffnessof thepresumed

double-stranded component of RI was analyzed by a

double logarithmic plot of end-to-end distance squared versus contour length. When spread in

the absence of urea, these molecules were more rigidthanarandom coil butnotasrigidas astiff rod (Fig. 12). When spread in the presence of urea, the stiffness of the double-stranded

com-ponentwasincreased,since theplotcontinuedas .35

_.30

C

0

O.25

0 0

u

.20

S

0

u C

T0

U-.10

.05

132 _J. _VIROL.

---I

on November 11, 2019 by guest

http://jvi.asm.org/

(5)

FIG.3. Replicativeform(a)spreadintheabseniceofureaanid(h) spreadinthepreseniceofurea. Wlholemolecules (lenigth : IIA),aswellasole-halfanidonie-quiartermolecules,areillustrated. X60,000.

133

VOL. 2, 1968

on November 11, 2019 by guest

http://jvi.asm.org/

[image:5.490.52.440.74.630.2]

(6)

GRANBOULAN AND FRANKLIN J. VIROL.

Replicative Form (spread without urea)

496 molecules

0.1 0.2 3 0.4 Q5 0.6 0.7 0,8 0.9 1.0 1.1

L (length in,u)

FIG. 4. Histogram of replicative formspreadinthe absence ofurea; atotal of 496 molecules were analyzed.

.5/ .4

3

D-.2/

/

0

N08

0

04 - ,

0

.03

/ /

.02 - ,

.1 .2 .3 A 5 6 8 10 20

Contour length (microns)

FIG. 5. Doublelogarithmic plot ofthemeant

end-to-elid distanice squared versus the contour length (h2'

versuscontourlength) of replicative form spreadinthe

absence of urea. In this plot, contour lengths were

pooled into groups of O.OS-,u steps; variable sample

numbers were available for each group (G = more

thani50samplespergroup; 0 = less thaniSOsamples

pergroup). The dotted line represenits therigidrod.

a straight line up to the maximal length of RI

(Fig. 13). Thiscurve wasalmost identicaltothat of RF (Fig. 5).

According to the semiconservative model of replication, the branches of RI are the partially

synthesized single strands of viral RNA which are displaced from the double strandas newviral

RNAissynthesized (7,9,12). Sincesynthesisand

displacement are presumed to occur simultane-ously,thelengthofthebranchshould be thesame asthe length of the double strandto the branch point, providedthat the mass per unit length of the single-stranded RNA isjust one-half that of

the doublestrand (9). Thisconditionseemstobe

true when one compares the maximal lengthsof R17 RNA and RF (see Discussion). Therefore,

themodelwasinvestigated by measuring the

con-tour length of the presumed double strand up

tothe branchpoint.

Asalready mentioned, it hasnot been possible

to distinguish the single-stranded and

double-strandedcomponentsofRIby theirwidth.

Never-theless,from evidencepresentedhere(see

Discus-sion) and elsewhere (12), it is clear that one of the branches of the branched molecules is single-stranded. Theoretically, the single-stranded

branch could haveall possible lengths up to the

length of the viral RNA (12). The double-stranded chain shouldbeofconstantlength equal

to the length of RF. Since there seemed to be breakage of the presumed double-strand during

134

.24

J

-C

a

e .20 0

0

0 .16

E

0

c . 12

0

.08

.04

I

on November 11, 2019 by guest

http://jvi.asm.org/

[image:6.490.61.447.48.305.2] [image:6.490.56.247.337.570.2]

(7)

FIG. 6. Replicative intermediate spreadin thepresence ofurea. The arrows point out whole molecules with variation inthe branch point andin thelengthofthebranch. Rosette structuressuchasthat seenin 6c werevery rarelyfoundinbothRIand RF. Thenatureof thesestructures is notknown. X 60,000.

135

on November 11, 2019 by guest

http://jvi.asm.org/

[image:7.490.41.444.44.623.2]

(8)

spreading, the analysis of the relation between length of branch and its position on the main chain wasdoneonly onmain chainsof maximal length.

Despite the problem of identification of the branch, contour lengths of the presumed branch and

of

thepresumedmain chainuptothe branch

point

were measured. Often therewas a definite

FIG. 7. Replicative intermediate spreadinthe presence ofurea. Whole molecules withtwobranches (arrows);

variationinthebranch point andinthelength of the branch. X 60,000.

136 J. VIROL.

on November 11, 2019 by guest

http://jvi.asm.org/

[image:8.490.62.445.123.627.2]

(9)

[image:9.490.48.444.78.607.2]

FIG. 8. Replicative intermediate spreadintheabsenceof urea;w,idirectionalshadowingwith PtC; (a, b) whole molecules withsinglebranches; (c) long coiledbranch attachedto oneendofawhole molecule. (a) X 105,000; (b, c) X 102,500.

137

VOL. 2, 1968

on November 11, 2019 by guest

http://jvi.asm.org/

(10)

[image:10.490.62.442.51.600.2]

FIG. 9. Replicativeintermediate treated with ribonuclease A (0.08

pg/ml,

10 min, 37C)followed by removal of the enzyme by phenolextraction: (a) spreadinthe absenceof urea;(b) spreadinthepresenceofurea.Nobranches wereobserved in these preparations. There is a pair of crossed molecules in theupperleft-handcornerof 9b. X 60,000.

138

on November 11, 2019 by guest

http://jvi.asm.org/

(11)

break inthedirection ofthemoleculeat the branch

point. Theoretically, the branch contour length (X) shouldequalthelength ofthedouble-stranded

molecule up to the branch point. Extensive measurementsofthese two parameters were made on whole RI molecules spread in the absence or presence of urea.

A scatter diagram ofthe datafor RIspread in the presence of urea is shown iii Fig. 14. There was agoodcorrelation betweenXand y, the

con-touLrlengthof the main chain tothe branch point.

TABLE 1. Ultrastructure of replicative i,ztermediate

Determination

Total number of whole moleculesb... Total number of

branched whole molecules ... Whole molecules with

No branches... Onebranch... Two branches... Threebranches...

Sp)read Spread withouturea with ureaa

No. %c No. %

50.1 49.8 44.0 5.1 1.2 333 169 164 147 18 1 434 218 216 191 22 5 50.1 49.9 44.1 5.4 0.3

a Spread in the presence of 6 to 8 M urea.

IMolecules with contour length of 1.0 to 1.1 ,.

Branches of all possible lengths were found, but

most ofthebranches were 0.2 to 0.4piin length.

Themaximal length of the branches was 1.05 ,u, approximately equal to the length of the single-strandedR17 RNA.

Theapparent correlationi between X and y was confirmed by least-squares analysis from the equation y = bX + c, where b and c are the con-stants to be determined (Table 2). This

analysis

was performed on a LOCI 2a computer with a programwhich handles a sample of 99 at any one time.Therefore, the samples of 266 molecules (RI

spread in the absence of urea) and of 210 mole-cules (RI spread in the presence of urea) were

split into three parts. Since the measurements were made at random, thevalues ofband c were very similar for each subsample. Whether spread in the absence or presence of urea, the resulting average values of b and c were the same and very

closetotheexpectedvalues ofb = 1 and c = 0. ThedistributionsoflengthsofXfromRIspread in theabsence and thepresence of urea is shown in Fig. 15 and 16, respectively. Neither

distribu-tion was uniform, but the distribution was more uniform with RI spread inthe absence of urea. When spreadin the presence of urea, there was askeweddistribution with a maximum at 0.20 to 0.25 ,u.Thiscanalsobeseeninthe scatter diagram (Fig. 14). In both preparations of RI, there were

Replicative Intermediate(spread without urea) 881 molecules

.08 H

.04 H

F-r

1 .2 .3 A .5 .6 .7 .8 .9 1.0 1.1 1.2

.28H .24 -.20 e-.16 W-.12 H c C 0 -0 0 0) E 0 c 0) cr 0 LL L(length inp)

FIG. 10. Histogram of thepresumed

maini

chtaint

ofRIspread

inz

the

absentce

ofurea; a totalof881molecules were

anialyzed.

VOL. 2, 1968 139

-I

on November 11, 2019 by guest

http://jvi.asm.org/

[image:11.490.43.236.211.366.2] [image:11.490.45.436.355.627.2]

(12)

140 GRANBOULAN AND FRANKLIN J. VIROL.

.28 Replicative Intermediate (spread with urea)

809 molecules

. .24

0 .20

u

E

0

u .12

U_ .08

.04

.1 .2 .3 .4 .5 .6 7 .8 .9 1.0 11 1.2

L

(length

in

Es)

FIG. 11. Histogram of the presumed mainchlain of RIspreadinthepresenceofurea;atotalof809molecules

wereanalyzed.

.5

.5-c /

/

~~~~~~~~~~~~.4/

.2

.1 / .0 00 '

(>08~~~~~~~~~~~

.0 /

~~~~~~.05~~~~~~~~~~~~~~~~~.8/

bj .05

.04 0

-/04

.03 - /.03K

.02 5 8. .

02 / i10

.1 .2 .3 .4 .5 .6 s 1.0 2.0 _01

l/_I

l_ l_ l_I

*l2 .3 A .S .6 .8 l.0 2.0

Cnurlength of double strand (microns)

[image:12.490.59.448.54.334.2]

Contour length of doublestrand(microns)

FIG. 12.Double-logarithmicplot

oJf/T2

versus contour

length for thepresumed main chain of RI, spread in FIG. 13.Double-logarithmic plot

oJ'h2

versuscontour the absence ofurea. At allpoints the sample was 20 lengthfor thepresumed main chain ofRI, spread in

orgreater. 0 =individuallengths;for lengths within the the presenceofurea. Thenumberof samplesperpoint

ranges indicated,data werepooled to give a sample of variedbetween 26 and 69. The dotted line represents

20orgreater. Thedottedline represents the rigid rod. the rigid rod.

on November 11, 2019 by guest

http://jvi.asm.org/

[image:12.490.56.248.374.615.2] [image:12.490.260.452.376.618.2]

(13)

Number of occurrences

00 00 0 0 0 0

0 00

. A n

oo 0 0 w

00 0

0 0

00 0 00 0 0

-~~~~G

oo e 0

00 0 o

_0 0 0 00 0

0,&&Q0%00 oo _ 0 0 1I 0.2 0.4 2 3 4 5 6 . 7 8

0.6 0.8 1.0 1.2

CONTOUR LENGTH OF BRANCH (microns)

FIG. 14. Scatter diagram of theconitourlentgth of the presumedmaint chainito thebranich pointt (y) anldthe con-tourlenigthofthebranich (X). Thisisthesample of210whlolemolectulesspreadiln thepresenice ofuireaanldusedfor

the liniearleast-squaresanalysis ofTable 2.

TABLE 2. Correlationt betweeni branich poinlt anld

branzcha

Urea Group Samplesize b c

Absent 1 98 0.978 0.028

2 98 1.009 0.004

3 70 0.959 0.052

Avg 0.98 0.03

Presentb 1 89 1.016 0.018

2 59 0.958 0.039

3 62 0.958 0.036

Avg 0.98 0.03

aFrom the formula Y =

bX

+ c where y =

distance to branch point measured on double-stranded molecule and X = length of the branch.

bConcentration,6to 8 M.

branches ofall possible lengths uptothe length ofthe

single-stranded

R17molecules.

Denatured RI. Samples of RI denatured with DMSO or by heating could be examined only

after spreading in the presence of urea. One attempt was made to examine the denatured RI

spread in the absenceofurea.There wasa great dealofinter- andintramolecularaggregation,and

thereforethe remainder of thepreparations were

made in thepresenceofurea.Linear molecules of highlyvariablelengthwerefoundinsuch prepara-tions (Fig. 17).No branchedmoleculeswere seen. The length distribution of 860 molecules of RI denatured with DMSO showed a peak

cor-responding to molecules of R17 RNA (1.05 ,u)

and a heterogeneous population with lengths

between 0.02 and 1.05 ,u (Fig. 18). The length

distribution of 389 RI molecules denatured by heatingshowedasimilarheterogeneity (Fig. 19).

Mixedpoptulations of RFandR17 RNA.

Con-trolexperiments were done with 1:1 mixtures of double-stranded RF and single-stranded R17 molecules. In the presence of urea, branched molecules amounted toabout 3 of the

popula-tion (Fig. 20). Therewas noincrease in the

per-centage of branchedmolecules in the absence of

urea, but aggregates were frequently found.

Exactly the same results were obtained with

mixtures which had been extracted witb phenol before spreading and mixtures which had not beenextracted with phenol.

DIscUSSION

Since measurements of base translation are available for double-stranded RNA but not for single-stranded RNA, the length of the

double-1.2 1.0 0.8 0.6 0.4 0.2 U) c 0 E F-z 5

0D

m I 0 z m 0 II 'r H z w -J 0 z 0 0 141

VOL. 2, 1968

0 0

I

on November 11, 2019 by guest

http://jvi.asm.org/

[image:13.490.79.402.74.349.2] [image:13.490.41.232.390.543.2]

(14)

GRANBOULAN AND FRANKLIN .1 2 I .1 0 z w U. 0

(,,.0

8 w 11 Q z

,.06

U-0 0 z w04 w U-.0 2

0.1 0.2 0.3 0.4 0.5

RI SPREAD WITHOUT UREA

XILu

[image:14.490.92.424.67.339.2]

0.6 0.7 0.8 0.9 1.0 1.1 X(CONTOUR LENGTH OF BRANCH in,u)

FIG. 15. Histogram ofXforRIspread ihtheabselice ofairea;atotalof266wholemolecules ofRIwereanialyzed.

stranded RF will be considered before thatof the single-stranded viral RNA. The histogram ofRF

has three peaks, the peaks atthe shorter lengths corresponding to approximately one-half and one-fourth the longest length. The mean of the

cluster around the longest length was 1.05 ,.

RFshould have 3,342 base pairs, by comparison with R17 (28); therefore, the calculated base translation distance for spread RF was 3.14 A.

According to the original analysis of X-ray diffraction patterns of double-stranded RNA

fromreovirus, the RNAwasthought to bea

10-fold helixwithapitch length of30.0 Aperresidue

(23, 29). More recent data and theoretical analyses of thestructures have revealedan

ambi-guity, since the diffraction patterns could be

accounted for by either a 10-fold or an 11-fold helix (2-4). Since the pitch length is thesamefor

either helix, the base translation may be either 3.00or 2.73 A perresidue. The ambiguity arises

from two possible ways of packing molecules in thecrystallographic unitcell. Sincecontourlength

measurements are made on isolated

molecules,

this ambiguity does notariseinthepresent deter-mination. With abase pair translation of3.14 A

anda pitchlength of30.0A, the number of base

pairsperpitch length would be 9.6, thusfavoring the original 10-foldhelixmodel ofLangridgeand

Gomatos (23).

Thesingle-strandedR17 RNAdistribution has

a single peak, and themeanlength ofR17 RNA is1.06,g.Thenumber of nucleotidespermolecule is 3,342 (28), and therefore the translation per base

residue is 3.17 A for this RNA spread on

mono-layers from solutions of 6to8 M urea.This is the

sameinternucleotide distance determined for RF,

implying that the spreadsingle-stranded RNA has

astacked baseconfiguration. Evidence for stacked

base structures insingle-stranded nucleic acidsin

solutionmay be found in recent literature. Only

onesuchexamplewillbecited.Themassper unit

length of single-stranded DNA, measured by small-angle X-ray scattering, is just half that of double-stranded DNA, implying stacking and

orderingof the base planes inthe single-stranded

DNA (25).

The length distribution of the presumedmain chainofRI is similartothatof RF, withapeak

between 1.00and 1.10 ,u and smaller peaks cor-respondingtoone-half and one-fourth this length. NeitherRFnorRIhad suchpopulations of mole-culespriortospreading, accordingtoanalyses of

the molecular weight distributions by

ultra-centrifugation (12). The presence of short

seg-ments ofRFfromRNA bacteriophage M12 has

also been noted inelectron micrographs, but no

distributionshave beenpublished (1). Populations

of "one-half" and "one-fourth" molecules of

142 _J. _VIROL.

on November 11, 2019 by guest

http://jvi.asm.org/

(15)

REPLICATION OF PHAGE RNA 143

.2 0

.18

.16

.14

.12

RI SPREAD WITH UREA

z

.0a

z

co.06

UA.

0

W.04

w

2

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 .

X(CONTOUR LENGTH OF BRANCH in/)

FIG.16. HistogramofXforRIspreadihthepreseniceofurea;atotalof2JO whiole molecuiles ofRIwerealnalyzed.

bacteriophageT2 DNA have alsobeen observed inspreadpreparations, butit isnotclearwhether these preparations already contained such mole-culespriortospreading (19).The double-stranded RNA has a greater rigidity than the single-strandedRNA(12), andismaybethat theformer

RNA issubject toshearing forces during spread-ing on monolayers. Breakage of the double strands byshearing would explainthedistribution into discrete whole, half, and quarter molecules (6).RIspreadinthe absence ofureahadahigher

proportion ofwhole molecules thanRIspreadin the presence of urea (Fig. 10, 11). Since the RI spread in the presence of urea would have a

slightly greaier rigidity than that spread in the absence ofurea, it would havea slightly greater

susceptibilitytoshear. Syntheticpolymers (poly-isobutylenes) as small as 500,000 in molecular weight are susceptibleto shearing (5), but ithas

been reported that double-stranded RNA from cells infected with fr bacteriophage is not de-graded inablenderat 10,000rev/min (17).

The problem of unequivocally identifying the single-strandedcomponent of RI is verydifficult

to solve. Single-stranded R17 RNA has four characteristics which distinguish it from double-stranded RF: (i) aggregation in the absence of urea, (ii) loss during spreading, (iii) sensitivityto

ribonuclease, and (iv) lengtlh distribution around asingle mode ratherthan the tri-modal distribu-tionof RF.Noneof thesecharacteristicsisuseful

fortheidentification of the single-stranded

com-ponent of individualRI molecules. Inthepresent

study, only RI approximately 1 ,u in length was

used for structure analysis. The I-, component

was considered to be the double-stranded main chain,andthecomponentsofshorter length were considered to be the single-stranded branches.

VOL. 2, 1968

on November 11, 2019 by guest

http://jvi.asm.org/

[image:15.490.67.413.56.441.2]

(16)

FIG. 17. Replicative

inttermediate deniatured

with (a)

dimethyl

sulfoxideor (b) by

heatin1g

at 37 C for 3

miii;

both samples were spread inithepreseniceofurea. X 46,800.

144

on November 11, 2019 by guest

http://jvi.asm.org/

[image:16.490.70.461.36.632.2]

(17)

This approach is in keeping with theoretical and experimental analysis of RI (11-13). It enabled us to determine the length distribution of branches andcorrelate branch length with position on the main chain. There is admittedly an unavoidable

ambiguity in identifying a "starting end" of the

.12

.10

0

r .08

e 0

E .06

0 u 04

0 0 .02

F-F

F-H

.1 .2 .3 1 .5

main chain. However, this can be eliminated by taking a starting end such that the branch point would beofapproximately the samelength as the main chain. If one measured from the other end

of the main chain, then y would equal L - X where L isthe maximum length of the mainchain.

.6 7 .9 1.0 1.1 1.2

l (length in,a)

FIG.18. HistogramofRI denatured withDMSO; a total of860 molecules wereanalyzed.

.12

.1_

0

- .10

0 0

E

- 06 0~

c

O .04

U-.02

1.0 1.1 1.2

L(length

ink,;)

FIG.19. HistogramofRIdenaturedbyheating(3miii,97 C); a total of389 moleculeswere analyzed.

145

VOL. 2, 1968

on November 11, 2019 by guest

http://jvi.asm.org/

[image:17.490.59.440.162.393.2] [image:17.490.53.446.422.640.2]

(18)

But rearranging the analysis so that y X for all determinations does not detract from the

established geometrical relationship between the branch and the branchpoint.

The RI used inthisanalysishadapproximately

one single strand per double strand (11, 12).

Assuming aPoissondistribution ofsingle strands, approximately 37%ofthe population should be

unbranched whereas only 50% of the population had visible branches. This may be due to two factors: loss of branches by shearing during spreading and the presenceof very short branches which cannot be distinguished from the back-ground grain. The presence of a smaller propor-tion oflonger branches in RI spread in the pres-enceof urea may be indicative of losses by

shear-FIG.20. Example ofa mixtureofRFandR17RNAspreadinthepresenceofurea;inthisexamplethetwotypes ofmoleculesweresimplymixedtogetherandthenspread. X 46,800.

146 J. VIROL.

on November 11, 2019 by guest

http://jvi.asm.org/

[image:18.490.59.438.170.634.2]

(19)

ing, since the single-stranded branches should be

stifferin thepresenceofurea; verylongbranches might be denatured from the double strand in urea, since onlyafew hydrogen bondsmaybind

itto the main chain.

Denatured RI behaves like single-stranded RNA in that it aggregates if spread in the absence of urea. Similardistributions wereobtained after denaturation atlow temperature in the presence

ofDMSO or athigh temperature inthe absence of DMSO. The highest peak corresponded to the complete RNA molecule. No molecules longerthan R17 RNA werefound,but molecules

of allpossible shorterlengthswerefound. Neither

distribution couldbefitted to the expected

distri-bution ofequal numbersof molecules of all

pos-sible chain lengths from N = 0 to 3,341, plus twice this number of N = 3,342 (12, 13). Con-sidering thedifficultiesofaccurate measurements, however, the distribution of heat-denatured RI (Fig. 19) approaches the theoretical distribution. Except for the peaks at 0.30 to 0.35 ,uand 0.40 to 0.45 u, the

frequency

of0 < N < 3,342 was

approximately 0.5

a.-d

the frequency for N =

3,342 wasapproximately 1.0. Thus, this

distribu-tion wouldalsofavoran RIstructure containing onlyonebranchperdoublestrand.

No branched molecules were found when RF and R17 RNA were mixed, even ifthe mixture

was treated with phenol before examination by

electron

microscopy.

This makes it

unlikely

that

RI arises as an artifact by association of single-stranded viral RNA and double-stranded RNA

during phenol extraction of infected cells.

Forfurther investigation ofstructures such as

RI, it is imperative that some ultrastructural

technique

be

developed

to

distinguish

single-stranded anddouble-strandedRNA. Theanalysis ofthebranchesofRIcould then bemadewithout

any ambiquity.

ACKNOWLEDGMENTS

We acknowledgethe excellent technical assistance of Elizabeth Hinckley and Marianne Salditt, who prepared thenucleic acids used in this study, and of Alain Niveleau, who assisted in the electron

micro-scopic preparation.

Thisinvestigation was supported by Public Health

Service grant AI-07645 from the National Institute ofAllergyand Infectious Diseases andby grant GB 5365 from the National Science Foundation.

LITERATURE CITED

1. AMMAN, J., H. DELIUS,ANDP.H. HOFSCHNEIDER. 1964. Isolation and properties of an intact

phage-specific replicative form of RNA phage

M12. J. Mol. Biol. 10:557-561.

2. ARNoTr, S., F.HUTCHINSON, M. SPENCER,M. H. F. WILKINS, W. FULLER, AND R. LANGRIDGE.

1966. X-ray diffraction studies of double helical ribonucleic acid. Nature 211:227-232. 3. ARNorT,S., M. H. F. WILKINS, W. FULLER, AND

R. LANGRIDGE. 1967. Molecular and crystal

structuresof double-helical RNA. II.

Determi-nationand comparison of diffracted intensities for the a and d crystalline forms of reovirus RNA andtheirinterpretation intermsofgroups

of threeRNA molecules. J. Mol. Biol. 27:525-533.

4. ARNOTT,S., M. H. F. WILKINS,W. FULLER, AND R. LANGRIDGE. 1967. Molecular and crystal structures of double-helical RNA. III. An 11-fold molecular model and comparison of the agreement between the observed and calcu-lated three-dimensional diffraction data for 10- and 11-fold models. J. Mol. Biol. 27:535-548.

5. BESTUL, A. B., AND H. V. BELCHER. 1953. Deg-radation of polyisobutylenes on shearing in solution. J. Appl. Phys. 24:1011-1014. 6. DAVISON, P.F. 1959. Theeffect ofhydrodynamic

shear on the deoxyribonucleic acid from T2 and T4 bacteriophages. Proc. Natl. Acad. Sci. U.S. 45:1560-1568.

7. ERIKSON, R. L., AND E. ERIKSON. 1967. Structure and function ofbacteriophage R17replicative

intermediate ribonucleic acid. II. Properties of the parental labeled molecule. J. Virol. 1:523-528.

8. ERIKSON, R. L., M. L. FENWICK, AND R. M. FRANKLIN. 1964. Replication of bacteriophage RNA:studies onthe fate of parental RNA. J. Mol. Biol. 10:519-529.

9. ERIKSON, R. L., AND R. M. FRANKLIN. 1966. Symposium on replication of viral nucleic acids. I. Formation and properties of a repli-cative intermediate in the biosynthesis ofviral ribonucleic acid. Bacteriol. Rev. 30:267-278. 10. FRANCKE, B., AND P. H. HOFSCHNEIDER. 1966. Infectious nucleic acids of E. coli bacterio-phages. IX. Sedimentationconstantsand strand integrity of infectious M12 phage replicative-form RNA. Proc. Natl. Acad. Sci. U.S. 56: 1883-1890.

11. FRANKLIN, R. M. 1966. Purification and proper-ties of thereplicativeintermediate of the RNA bacteriophage R17. Proc. Natl. Acad. Sci. U.S. 55:1504-1511.

12. FRANKLIN, R. M. 1967. Replication of bacteri-ophage ribonucleic acid: somephysical proper-ties of single-stranded, double-stranded, and branched viral ribonucleic acid. J. Virol. 1: 64-75.

13. FRANKLIN, R. M. 1967. Replication of bacteri-ophage ribonucleic acid: some properties of native and denatured replicative intermediate. J.Virol 1:514-522.

14. GRANBOULAN, N., AND R. M. FRANKLIN. 1966. Electron microscopy of viral RNA, replicative

form, and replicative intermediate of bacteri-ophage R17. J. Mol. Biol. 22:173-177. 15. GRANBOULAN, N., AND K. SCHERRER. 1966.

147

VOL. 2, 1968

on November 11, 2019 by guest

http://jvi.asm.org/

(20)

Examen au microscope electronique de RNA monocatenaires. J. Microscop. 5:54a. 16. HOTHAM-IGLEWSKI, B., AND R. M. FRANKLIN.

1967. Replication of bacteriophage ribonucleic acid: alterationsin polyribosome patterns and association of double-stranded RNA with

polyribosomesin Escherichia coliinfected with bacteriophage R17. Proc. Natl. Acad. Sci. U.S. 58:743-749.

17. KAERNER, H. C., AND H. HOFFMANN-BERLING. 1964. DieBildungvonRNS-Doppelstrang zur Vermehrung eines RNS enthaltenden Bak-teriophagen. Z. Naturforsch. 19:593-604. 18. KATZ, L., AND S. PENMAN. 1966. The solvent

denaturation of double-stranded RNA from

poliovirus infected HeLa cells. Biochem. Biophys. Res. Commun. 23:557-560.

19. KLEINSCHMIDT, A. K. 1967. Nucleinsauren. Ar-chitektur und Wandlungsfahigkeit im mole-kularen Bild. Naturwissenschaften 54:417-428. 20. KLEINSCHMIDT, A. K., T. H. DUNNEBACKE, R. S. SPENDLOVE, F. L. SCHAFFER, AND R. F. WHIT-COMB. 1964.Electron microscopy ofRNAfrom reovirus and wound tumorvirus. J. Mol. Biol. 10:282-288.

21. KLEINSCHMIDT, A. K., D. LANG, D. JACHERTS, AND R. K. ZAHN. 1962. Darstellung und Langemessungen des gesamten desoxyribo-nucleinsaure-Inhaltes von T2-Bakteriophagen. Biochim. Biophys. Acta 61:857-864. 22. LANG, D., A. K. KLEINSCHMIDT, AND R. K.

ZAHN. 1964. Electronen mikroskopische Dars-tellung, Langenverteilung und Konfiguration einstrangiger Desoxyribonucleinsaure im Cyto-chrom-c-Film. Biophysik 2:73-78.

23. LANGRIDGE, R., AND P. J. GOMATOS. 1963. The structureof RNA. Science 141:694-698. 24. LODISH, H. F., AND N. D. ZINDER. 1966.

Semi-conservative replication of bacteriophage f2 RNA. J. Mol. Biol. 21:207-209.

25. LUZZATI, V., J. WITZ, AND A. MATHIS. 1967. The structure of nucleic acids in solution, as determined by x-ray scattering techniques: DNA, RNA, poly A, p. 41-55. In D. Shugar

[ed.], Genetic elements, properties and func-tions. Academic Press, Inc.,New York. 26. MILLS, D. R., N. R. PACE, AND S. SPIEGELMAN.

1966. The in vitrosynthesis ofanoninfectious complex containing biologically active viral RNA. Proc. Natl. Acad. Sci. U.S. 56:1778-1785.

27. PACE, N. R., D. H. L. BISHOP,ANDS. SPIEGELMAN.

1967. The kinetics of product appearance and template involvement in the iin vitro replication of viral RNA. Proc. Natl. Acad. Sci. U.S. 58:711-718.

28. SINHA,N.K., R. K. FUJIMURA,ANDP.KAESBERG. 1965.Ribonucleasedigestion of R17 viral RNA. J. Mol. Biol. 11:84-89.

29. TOMITA, K. I., AND A. RICH. 1964. X-ray dif-fraction investigations ofcomplementary RNA. Nature201:1160-1163.

148 J. VIROL.