Replication of Bacteriophage Ribonucleic Acid: Some Properties of Native and Denatured Replicative Intermediate

JOURNALOFVIROLOGY, June 1967,p.514-522 Copyright©1967 AmericanSociety forMicrobiology

Replication of Bacteriophage

Ribonucleic

Acid:

Some

Properties

of

Native

and

Denatured

Replicative Intermediate

RICHARD M. FRANKLIN

ThePublic Health ResearchInstitute oftheCityofNew York,Inic.,

New York,New York10009

Receivedforpublication 18January 1967

Purified replicative form (RF) and replicative intermediate (RI) prepared from Escherichia coli cells infected with the ribonucleic acid (RNA) bacteriophage R17 weredenatured withdimethyl sulfoxideat 37C orinaqueoussolvents of lowionic

strengthat97 C. Denaturationwasdemonstrated for RF andRIby an increase in

specific infectivity and a striking change in the hyperchromicity curves after

treat-ment. RI denaturation wasalso demonstrated bya shift in the buoyant densityin

Cs2SO4 from 1.619 to the buoyant density of single-stranded R17 RNA (1.627). Analysis of the denatured RIhyperchromicitycurvesand the equilibrium

distribu-tions of denatured RI in Cs2SO4 gradients revealed, however, a residual double-strandedcomponent. Velocitysedimentation of denatured RI was performed, and the weight distribution of S values was calculated. From the known relation

be-tween molecularweight and S values, itwas possibleto transformthe weight

dis-tribution into a number distribution of chain lengths. This distribution was com-pared with that predicted from the steady-state hypothesis for RI. Deviations fromthe predicteddistributionmaybedueto the residualdouble-stranded

compo-nent.

The replication of single-stranded viral ribo-nucleic acid (RNA) has been investigated by studying thestructure ofreplicative intermediate (RI), which is double-stranded RNA template with boundnascentviralRNA(7,8).Thus,some

of the details of the kinetic process of RNA

synthesis canbe explored by structural analysis,

since RIisbelieved torepresent asteadystateof moleculesundergoing polymer synthesis (8).

In the first paper to enunciate this principle,

thestructureofRIwasexplored by measuringits

hydrodynamic properties and comparing them with theproperties of single-stranded viral RNA

andreplicative form (RF), which ispure

double-stranded RNA (8). In the present paper, the

emphasiswill beonthepopulation of molecules arising after denaturation of RI.Accordingtothe

theoretical analysis of RI, complete denaturation should yield a population of single-stranded

moleculeswithequalnumbers of molecules from

dinucleotidesupto a chainlength ofX-1, where

X = 3342isthenumber of nucleotides in the viral

RNA, plusalarger number of moleculesof chain length X,which isthe contributionof the

double-strandedtemplate.

There is no possibility for a priori prediction of the average number of single strands per

double-stranded template. When the theoretical analysis of RI was applied to the data on the

relative amounts ofsingle- and double-stranded

component inthepopulation of molecules of RI,

itwasestimatedthattheaveragenumber ofsingle

strandsofanylengthwas oneperRImolecule (8). Itmustbeemphasized that this numbermay vary

from batch tobatch of RI andmayalsodepend

on the method of preparation. Therefore, the

number of molecules of nascent viral RNA

could be more than one in native RI in vivo. After denaturation of preparations having one single strand per template, there should be 2X

moleculesof chainlengthX andXmolecules with

chainlengthsfromX-1to2(8).The 2X molecules

of length X derive from the template with a molecularweightof 2.2 X 106.

As apreliminary toan analysisof the popula-tion ofmolecules derived from RI by

denatura-tion,it isnecessarytoinvestigatemeansbywhich

RI can be denatured. Denaturation of RI was

followed by increase in infectivity, changes in

buoyantdensity,andchanges inhyperchromicity.

514

Vol.1, No. 3 Printed in U.S.A.

on November 11, 2019 by guest

http://jvi.asm.org/

MATERIALS AND METHODS

Preparation of RNA. Growth of Escherichia coli

strain3000 andbacteriophageR17 has been described

(9).Preparationof RNA from purified bacteriophage R17 and preparation and purification of RF and RI

have beendescribed (7, 8, 20).

Denaturationi ofdouble-stranzded RNA (RF orRI).

(i) For heatdenaturation, approximately 20to40/Ag

of RNA per ml was dialyzed overnight against PE

[0.001 M potassium phosphate buffer, pH 6.6; 0.001

M sodium ethylenediaminetetraacetic acid (EDTA);

ionic strength0.0121. This RNAwasthenplaced ina

tightly stoppered tube, the total volume being 1 ml.

Thetubewasplunged intoaboiling-water bath,held

there for 3 min, and thenquick-frozen ina dry

ice-alcohol bath. (ii) For denaturation with dimethyl

sulfoxide (DMSO), double-stranded RNA was

de-natured with DMSO according to the method of

Katz and Penman (15). Approximately 200 to 400

flgof RNAwasdialyzed againstPE. Onepartof this

RNA was mixedwith six parts of DMSO

(Spectro-quality Reagent; Matheson, Coleman and Bell,

Cin-cinnati, Ohio), givingafinalDMSO concentration of

85.7% (v/v). This wasincubatedat37 Cfor 10 min

and thenquick-cooledinanice bath. NaCIwasadded

to afinalconcentration of 0.1 M (calculated onlyon

thebasis ofthe aqueousvolume), and the RNA was

precipitated byadditionof 2volumesofethylalcohol

followedbyincubationovernightat -20 C. The

pre-cipitated RNA was taken up in PE and dialyzed against PEbefore infectivity tests oranalytical

sedi-mentation studies.

Inifectivity ofRNA.Thesetestswerecarried outby

infecting spheroplasts with RNA and plating the

sphereoplasts for infective centers, byuseof standard

assay techniques. The preparation and infection of

spheroplastswerecarried outby modifications of the

procedure of Francke and Hofschneider(6). Aculture

ofE.coli 3000wasgrownovernightin TC Glmedium

(7),diluted 1:2,000 in TCGl,and allowedtogrowto

atiter of2 X 108cells per ml. The cellswereharvested

by centrifugation (5,000 rev/min) at room

tempera-turefor 10 min, andwerethen taken up in one-tenth

the original volume in five partstris(hydroxymethyl)

aminomethane (Tris) buffer(0.1 M,pH 7.7) andthree

parts 40% sucrose. Then EDTA and lysozyme were

added todigest the cell wall and form spheroplasts.

Anexample of thisprocedure, for 100 ml of original

culture, was asfollows. (i) A 6.25-ml amount of 0.1

MTrisbuffer (pH 7.7)wasadded; (ii) 3.75mlof

40%7o

sucrose was added; (iii) 0.26 ml of 0.1 M EDTAwas

added; (iv) 0.26 ml of lysozyme (Worthington

Bio-chemical Corp., Freehold, N.J.) was added at 2

mg/ml. The cells wereincubated for 10 min at room

temperature, resulting in a complete conversion to

spheroplastsasobservedby dark-fieldmicroscopy. At

this stage, further lysozyme action was inhibited by

addition to the above sample mixture of 0.14 ml of

0.5 MMg2SO4and 0.16 mlof30%bovineserum

albu-min (Fraction V from bovine plasma, Armour

Pharmaceutical Co., Kankakee, Ill.).

The spheroplasts were kept at room temperature

for30min. Then0.2-ml samples were exposed to 0.05

ml ofappropriate dilutions of RNA (usually 10, 1,

and 0.1 Mig/mlin 0.1 M Tris buffer, pH 7.7). After an

additional 30 min at room temperature, 0.2 ml ofan

E. coli 3000 suspension (108 cells per ml) wasadded,

followed immediately by the addition of 2 ml of

spheroplast soft agar; this mixture was poured onto

nutrient agar plates. The spheroplast soft agar had

thefollowingcomposition: Difco agar, 0.75%;

suc-rose, 3%,; CaC12, 0.01 M; MgSO4, 0.01 M; NaCl,

0.5% .

After autoclaving, the pH was adjusted between 7

and 8with 1NNaOH. Bottom agarwas the modified

MSagarpreviously used (9).

Hyperchromicity. Hyperchromicity curves were

recorded on aGilford model 2000 spectrophotometric

recorder. Glycerin was circulated through Beckman

thermospacers, and the temperature of the circulating

bath was set athigher than 100 C. The temperature,

measured directly in the cuvettes, increased from 25

to 100C inless than 45 min. The data werecorrected

for the expansion of water and plotted as optical

density (OD) at 260 mM at any given temperature relative to that at 25 C.

Analytical centtrifugationi. Sedimentation velocity

experimentsandanalysis of the results were described

in a previous paper (8).

Distribution functions in the Cs2SO4 equilibrium

gradient were determined by use of a 12-mm

alumi-num centerpiece. The solvent contained 0.001 M

potassium phosphate buffer (pH 6.6) with 0.001 M

sodium EDTA (PE). Solvent saturated at 25 Cwith

Cs2SO4 (optical grade; StanleyH. CohenAssociates,

Yonkers, N.Y.) wasadded toIto 2Mugof RNAin PE,

andthedensity of the solution was thendeterminedby

refractometry(high accuracy Abbe '60'refractometer,

Bellingham and Stanley) by use of the equation p25 =

13.6986 [nD]25 - 173233 (13, 21). Thefinal solvent

density wasadjusted to the appropriate value to band

theparticularRNAapproximately in the center of the

cell. Equilibrium centrifugation was carried out at

42,040 rev/min at25 Cfor 24 to 48 hr. Under these

conditions, no change in the concentration

distribu-tionof nucleic acid occurred after 24 hr.

Densitometer tracings of the photographic record

of theequilibrium distributionweremade by use of a

Spinco model RAnalytrol with a microdensitometer

accessory. Thedistribution ofCs2SO4 in the analytical

cellwascalculated from the usual equation (21):po =

p+ (12$) w2(ro + r'e) (ro- r'e), where po = density

atband center; p = original density of the solution;

c = angularvelocity;ro = distancefrom rotor center

toband center; r'e =

[12

(rb2 - ra2) 1/2; rb =distance

from rotor center to bottom of thecell;ra = distance

fromrotorcentertomeniscus; 1

/O

=constant.

Theconstant 1/dwas determined from the

collec-tion ofdataontheposition of the band center for a

given species of RNA centrifuged inCs2SO4 solutions

ofvariable original density. The average value of 1/d

was 1.252 X 10-9 (cgs), which differs from that

re-ported by Erikson and Szybalski (5). This may be due

to several factors, such as the relative purity of the

Cs2SO4. The purity of the Cs2SO4 would affect the

determination of p and thus alter the value of

1/7.

on November 11, 2019 by guest

http://jvi.asm.org/

FRANKLIN

RESULTS

Demonstration of denaturation of RF and RI:

Infectivity assays. According to Amman, Delius, and Hofschneider (1), double-stranded RNA (RF) fromcells infected with an RNA bacterio-phage is not infectious by itself but canbe con-verted to infectious material by heating. The conversion is due to denaturation of the double-stranded RNA, yielding at least one infectious single-strandedRNA, the"+"strand(1,6, 12a). This conversionhas beendemonstrated for both RFandRI(6,12a). There is alsosomesuggestion that RI is infectious before heatdenaturation and thatthe specific infectivity increases after heating (R. Erikson, E.Erikson, andJ. Gordon,J. Mol. Biol. 22:257-268, 1966).

[image:3.461.44.236.326.447.2]

Data onthe increase ininfectivity afterDMSO denaturation of preparations ofRF or RI are reported in Table 1. The infectivity of a single batch of viral RNA varied by a factor of10 in individualexperiments,twoof which arereported in Table 2. This variation was probably due to variation in competence of the spheroplasts.

TABLE 1. Infectivity ofviral-specific RNA species

PFUperpAg Increase(fold)

Prepn of RNA ~afterDMS0

Prepn of RNA

Ictreatment

R17 RNA(A)a... 1.22X 105

R17 RNA (B)a... 1.35 X 104

Replicative form

(RF)... 492

DMSO-denatured RF.. 6,644 13.5

Replicative

inter-mediate

(RI)

... 41

DMSO-denatured RI.. 474 11.6

aStandard RNA samples wereassayed on two

differentbatchesofprotoplasts. Batch Awasused

in the RF assayand batch B in the RI assay.

TABLE2. Buoyant density ofnative and denatured

viral RNA species

Exptno. Sample PO

5 R17 RNA 1.627

7 R17 RNA 1.627

13 R17RNA 1.628

1 RI 1.619

3 RI 1.619

8 RI 1.619

16 Denatured RIa 1.631

17 Denatured RIa 1.625

10 Denatured RIb 1.633

15 Denatured RIc 1.627

aTreated for 3 min at97 C.

bTreated with

85.7%

DMSOfor10minat37 C.

cTreated with85.7%DMSOfor10minat45C.

Everytest of infectivity of RF or RI was accom-panied by a test of viral RNA. In a typical experi-ment, there was a 10- to 20-fold increase in infectivity after DMSO treatment of either RF orRI.Thus, the experiments of Hofschneider and co-workers onheat denaturation of RF and RI (12a, 15) and the present experiments withDMSO suggest that both methods may be useful in investigating the pattern of single-stranded mole-cules derived from RF and RI bydenaturation. Most of thepresent data areconcernedwith RI, since conversion of RF to 27S single-stranded RNA has been well established (12a). Some comparativedata on the denaturation of RF and RI are also presented.

Hyperchromicity. Figure 1 presents a compari-son ofhyperchromicity of R17 RNA, RF, and RI. As expected, the single-stranded RNA dena-tures over a very broad range of temperadena-tures (2), thedouble-strandedRNA over a very short range (11, 16), and RI has a complex hyperchromicity owingto itspartially single-stranded and partially double-stranded properties (7). Thus, denatura-tion ofRF or RI should result in alterations in the hyperchromicitycurve. This wasbest demon-strated with RF (Fig. 2), which appears to be completely converted to single-stranded RNA after DMSO denaturation (10min at 37 C), at least according to the criterionemployed in this experiment.

RIdenaturedby treatment for 10minat 37 C with 85.7%DMSO still retained a small amount of double-stranded character as determined by the hyperchromicity test (Fig. 3A). This was retained even after more drastic treatment with DMSO (10minat 45 Cwith85.7%DMSO,Fig. 3B), or after heat denatdration (3 minat 97 C, Fig.

3C).

The smallest amount of residual

- 130

t

. 120

110

-Bacteroph.geR17 RNA

-- Replicotwoih. .ed..to

RWpic.ti,efo___.

-/

-X

!

TI

30 40 50 60 70

TEMPERATURE('C)

80 90 100

FIG. 1. Temperaturedependence ofOD at 260m,u

ofR17 RNA, RF, and RI, all in PE (ionic strength

0.012). TheODat260m,uisplottedasrelativeincrease

in OD at260m,uatany given temperaturecompared

withthatat25 C.

516 J. VIROL.

on November 11, 2019 by guest

http://jvi.asm.org/

[image:3.461.246.445.378.608.2] [image:3.461.44.237.463.645.2]

DM5

Replicat.4efo (RF)(5

-^ DMSOtreatedRF(85.7X10min 37'C

double-strandedcomponent was found after heat denaturation.

Equilibrium

distributions in

CS2SO4.

The

buoy-ant density of a single-stranded viral RNA was invariably lower than that of double-stranded RF isolated from cells infected with that virus (cf. 4). Since RI seems to have an intermediate value of buoyantdensity accordingtoequilibrium data in the preparative ultracentrifuge (3), buoyant density provides another criterion for denaturation. With the availability of highly

30 40 50 60 70 80 90 100

TEMPERATURErc)

FIG. 2. Relative increase in OD at 260 m,u as a

function of temperature of native RF (0) and RF

treated with DMSO(@).Bothsamplesweremeasured

inPE.

E 670 665

uJ

06560

655

I 00.50

u 645 5 640 635

DMSOt5e0t RI(857%IO..37-C)

BACTERIOPHAGE R17 RNA

(A) Il 0.1 05 2 510 20 40 60 80 95 99

CUMULATIVE f X 100

(A)

.1 I

30 40 50 60 70

TEMPERATURErc)

130- Rpqicati* int*8*doato(RI)

E1.25~ 4 DM50orta4dRi8(57%10m.i 45C)

110

1.20

1.05-

o-20 30 40 50 60

To_stwe(-C

E 665 -' 660

Z 655 6)50

s0_{eo vo}90 100_loo ~~~~0I _6.45

u 640

635

630

REPLCATIVE INTERMEDIATE

(B)

001 01 05 2 5 10 20 40 60 80 95

CUMULATIVEf X 100

(B)

70 80 90

i30- - Eplicativeinhrm7di3t9(RE)

E1.25- ~ heat dRl(3

97C)-20

S1-2

1.00 5(-CI~~~~

20 30 40 50 60 70 80 90

[image:4.461.28.214.51.209.2]

Twwa,*rc)

FIG. 3A-C. Relative increasein ODat260ml asa

function oftemperatureofnativeRIand RItreatedas

indicated. All samples were measured in PE. InFig.

3C, heatingat97 Cwasfollowed byquick-freezingat

-70 Cpriortodeterminingthehyperchromicitycurve.

E 670

cRE 3c6h inrmediateR17RNA

665 -Z 660

-0655 0 6.50

-645

e 6A0

-(C) 0.01 GI 05 12 5 10 20 40 60 80 95 99 99.9 9999

CUMULATIVEF x100

FIG. 4. Cumulative distribution of the equilibrium

concentration ofR17 RNA and RIpreparations in a

Cs2SO4 gradient. Thecumulativedistributionona

prob-ability scale isplotted versus distance from the rotor

center. (A) R17RNA; (B) native RI; (C) mixtureof

RIandR17 RNA (ratioof RItoR17RNA,1.01).

140

5 130

-vu_

z

O 1.20

-010 _

130 _

<t

>0

!i oc

° R*plicoatiei4*so.*dito.(Ri)

99.9 9999

A .

J

140

_-i.

a,

on November 11, 2019 by guest

http://jvi.asm.org/

[image:4.461.224.418.190.602.2] [image:4.461.29.217.265.604.2]

FRANKLIN

purified RI, the buoyant density of RI was carefully compared with that of R17 RNA in Cs2SO4 in the analytical cell.

Some of the data on the buoyantdensities are shown in Table 2. The average buoyant density of single-stranded R17 RNA was 1.627 and of RI, 1.619, values consistent with the reported values of 1.630 and 1.616, respectively (3). After denaturation, the values scattered somewhat but were all close to the value for single-stranded RNA. Ananalysis oftheequilibriumdistribution, however, indicated that double-stranded RNA was still present in the heated or DMSO-treated RIpreparations. This analysis was carried out by

.12 r

f 10

.08 F

02 k

* DMSO 857%37-C,10 ,,

x xDMSO 85.7%45'C10min

\97-C 3min

/

*W

----I c. T I

6/3 i

6.3 64 6.5 66 67

Distance from rotor center r Incm)

plotting on probability paper the cumulative distribution of the equilibrium concentration of nucleic acid versus distance from the rotor center. The height above base line of the densitometer tracing was taken as a measure of nucleic acid concentration.Thedistributionwasstartedatthe point nearest the rotor center. This plot has several advantages: (i) a Gaussian distribution is transformed intoastraight-line plot; (ii) a- can be read directly as the abscissa height between ordinate values of0.500 and0.841; (iii) themean, which should correspond to the peak for the Gaussian distribution, is read directly from the ordinate value of0.50; (iv) ascomparedwiththe

(B)

E 67 Denatured replcot,einte,medolte

c --- 857° DMSO I10m8 37°C

665 -- 85 7%DMS0 l08n845'C

No DMS0 38in 97°C

Z 660 -0

655

-0 0 650 -J 645

-0 640

635

-x

O01 01 05 2 5 10 20 40 60 80

CUMULATIVEf x 100

[image:5.461.49.440.231.426.2]

95 99 49 9 99.99

FIG. 5. (A)Distributionof theequilibriumcontcentrationofdenatured RI. (B)Sameas5A,plottedas a

cumula-tivedistributiononprobabilitypaper.

(A)

)o 0

6

z 4

25 75 125 175 100

S20,.

200 400 600 1000

Choi.length(N)

FIG.6. (A) WeightdistributionofSvalueforRI denatured with85.7%70DMSOfor10minat 37 C. This RI was

centrifugedinasolvent ofionicstrength0.06 at 52,640rev/min. (B) Numberdistribution ofchain lengths forRI denaturedwith 85.7% DMSOfor10minat 37 C. This isatransformation ofthe datapresentedinFig. 6A. The

arrowindicatesthepositioni ofR17RNA (molecularweight, 106to 1.1 X106;chaiinlength, X =3,342).

518 J.VIROL.

0

x

ci

2

x

0

u

x:

(D 20 175 15 125 10 75

25

(B)

yAv/ \

2000 14000

--,W-A 'X

x

on November 11, 2019 by guest

http://jvi.asm.org/

[image:5.461.49.441.463.614.2]

plot of band

height

versus(band

width)2/8j

(19), the

probability plot

is more accurate, since it avoids thelargeerrorsduetosquaring the largest bandwidths, whichare mostdifficult to measure because theyareclosesttothebaseline.

The Gaussian plot of the equilibrium distribu-tion of

bacteriophage

R17 RNA is shown in Fig. 4A; that ofRI, in Fig. 4B; and that ofan artificial mixture

of

R17 RNAand RI, ina ratio of1.01 RI to 1 R17

RNA,

is shown in

Fig.

4C. TheR17RNAdistribution is

Gaussian,

asisthe RIdistribution. The Gaussian distribution ofRI is

unexpected

and will be commented on in the Discussion. The two-component distribution of the mixture of RI and R17 RNA was

clearly

demonstrated by

the

probability plot (Fig.

4C). The peak of this distribution was located at a

densityof 1.614g/cm3. The

equilibrium

concen-trationdistributionsfor

denatured

RI are shown inFig.5A, and the

probability plot

of the cumula-tive

distributions,

in

Fig.

5B. Under all conditions of

denaturation,

the distributions are skewed

toward

the

buoyant

density of

single-stranded

RNA, and the

departure

from the Gaussian distribution is

clearly

showninFig.5B.The

posi-tion of the

peak

inrelationto

distance

from the rotor center

depended

on p and the

position

of themeniscus inthe

ultracentrifuge

cell.

Thus, theexperiments designedtodemonstrate denaturation ofRI and RF indicated that there was aresidual double-stranded component pres-ent in RI,

although

the

major

component was

denatured

single-stranded

RNA. The

hyper-chromicity

curves indicated that heat-denatured RI hada smaller residual

double-stranded

com-ponentthan

DMSO-denatured

RI.

Velocity sedimentation

analysis of

denatured RI. The above

experiments, although interesting

in

themselves,

were

preliminary

to the

analysis

of denatured RI. The

following

strategy was used. Since the theoretical

analysis

ofRI viewed as a

steady-state population predicted equal numbers offragmentswithchain

lengths

of2to

X-1,

then themost

meaningful

analysis would bea plot of the number distribution of chain

lengths.

This distribution can be derived from the weight distribution ofS values (17)

provided

that the

relationship

between molecular weight and S value is known. Single-stranded R17 RNA in solutions of high ionic strength has, however, a sedimentation constant which is higher than thatexpected fromtheSpirin or Gierer relation-ship between S and M (7). Therefore, the ionic strength ofthe viral RNA was carefully adjusted according to the curve of dependence of the S value ofthis RNA on ionic strength (8) so that the Svalueofthecompletemolecule of molecular weight 106 to 1.1 X 106would correspond to the

value predicted from thewell-knownrelationship betweenSandM.In

particular,

an

ionic

strength

of

0.06M wasusedtoobtain S = 22.5. To

clarify

and recapitulate, RI was

denatured

in solutions of low ionic strength as described in Materials andMethods. Just beforeperforminga standard velocity sedimentationanalysis,the ionic

strength

wasadjustedto0.06Mbyadding the

appropriate

amountofNaCl.Aweightdistribution

of

Svalues [q(S) =

1/Co

x

AC0/.SO,30

x

100]

was cal-culated (17), plotted, and then transformed into a number distribution of chain

lengths.

The weight distributions [q(S)] areshown in

Fig.

6A and Fig. 7A. The number distributions

[N(S)]

areshowninFig. 6Band7B.

The transformationfroma weight distribution to a number distribution [N(S)] was

done

as

follows.

Amolecular

weight

was

assigned

toeach Svalue byuseoftheSpirin

equation,

M =

1,550-20

-z8

6

(A) 28r

24

121

to0 30

S.

B)

o

I--,1

-T

I-

11

K)O 200 400 600 1000

Choi.lgth (N)

[image:6.461.230.419.280.604.2]

2000 14000

FIG. 7. (A) Weight distribution ofS valuefor RI

denaturedbyheating inPEfor3 minat97 C.

Condi-tionsforcentrifugation,same as inFig. 6A. (B)

Trans-formation of the data presentedinFig. 7A tothe

num-berdistributionofchainlengths. Thearrowindicates the

positionofR17 RNA.

(E

on November 11, 2019 by guest

http://jvi.asm.org/

FRANKLIN

521(cf. 7). Theweightdistribution q(S)wasthen divided by the chain length toobtainN(S). The chain length (number of nucleotides) was cal-culated for each S value by first calculating M and then multiplying this by 3,342/1.1 X 106, since the single-stranded RNA of molecular weight 1.1 x 106 has 3,342

nucleotides

(18).

Figure 6A shows the weight distribution ofS

for

RI

denatured

with DMSO, and

Fig.

6Bshows thetransformationtothe number distribution of chain lengths. Figures 7A and B show, respec-tively, the weight distribution of S and the

transformation

to the number distribution of chain

lengths

for

heat-denatured

RI. There was apronounced

bulge

inthe numberdistribution of

DMSO-denatured

RI (Fig.

6B)

and a smaller

bulge

in the number

distribution

of heat-de-natured RI (Fig.

7B).

Since

heat-denatured

RI hadlessresidual

double-stranded

componentthan DMSO-denatured RI, it is not unreasonable to speculate that this prominent

deviation

from the predicted distribution may

be due

tothe

double-stranded

component.

The

height of the

peak

of

complete

viral RNA

(N

=

X)

was about twice that of

the

plateau

region

for DMSO-denatured RI and 2.5 times the plateau region for

heat-denatured

RI.

DIscuSSION

As a

preliminary

to an

analysis

of

denatured

RI,

several

criteria for denaturationwere

studied.

The

infectivity

ofRIaftertreatmentwith DMSO

under conditions

which result in

denaturation

of thepoliovirus

replicative

form

(15)

ledto a10- to

14-fold

increase in

specific

infectivity

for

sphero-plasts,

measured

as

plaque-forming

units

(PFU)

per

microgram

of RNA. This increase is

com-parabletothe ninefold increase in RI

infectivity

after heat

denaturation, reported

by Erikson, Erikson, and Gordon

(J.

Mol. Biol.

22:257-268,

1966).A

similar

increasein

specific

infectivity

was

found

forRF

denatured

withDMSO.

Changes

in

optical

density

during

thermal denaturation of R17

RNA, RF,

andRI

provide

a

basis for a further

analysis

of the

degree

of denaturation.

Single-stranded

RNA denatures over a wide temperature range

(2),

whereas RF and other types of double-stranded RNA de-nature athighertemperaturesand overa narrow

temperature range

(11, 14,

16).

RI

hyper-chromicity curves havetwo distinct components

owing

to the

single-

and double-stranded prop-ertiesof themolecules

(7).

All thesefeatures are

illustratedin

Fig.

1.Whereasthe

hyperchromicity

ofDMSO-denatured RF increases

continuously

with no detectable step, that of RI denatured underanyconditionhasastep inthetemperature range where double-stranded RNA denatures.

This suggests a residual double-stranded com-ponent, and the amount of this component can be estimated from the percentage increase in absorbance of the step region. The residual double-stranded component varied between 17 and25% ofthe total withDMSO-treated RI and was6.4% withheatedRI.

Equilibrium distributions of both R17 RNA and RI in gradients ofCs2SO4 conformed to the Gaussian distribution.Althoughthiswasexpected for the single-stranded RNA, itwas not entirely expected for RI, which is composed of single-strandedRNA of

po

= 1.629 anddouble-stranded RNAof

po

= 1.608to 1.606 (3, 4). The variable lengthsofsingle-stranded RNA might be expected toconferadensityheterogeneity on RI which has anaverage buoyant density of 1.619, intermediate

between

that of single-stranded R17 RNA and RF. However, it may be that the intermediate buoyant density ofRI is not entirely dueto the averaged contributions ofRFandR17 RNA, but also dueto some structural features peculiar to RI, such asthe structure atthe point where the single-stranded branch deviates from the double strand.

From the probability plot of the cumulative distribution, the

a2

forR17RNA was 2.02 X10-3, leadingto amolecularweightestimate of0.77 x 106.TheU2forRI was1.37 X 10-3to1.51 X 10--3 leadingto amolecular weight estimate of1.01 x 106to1.14 x 106.

The

molecular weight estimate for R17 RNA istoolowbya

factor

1.4and that forRI

by

a

factor

2.5, byuse

of

1.1 X 106 asthe best estimatefor R17RNA(10,

18)

and 2.9 x 106 asthe theoretical weightaveragemolecular weight ofRI(7). The

molecular weight

estimates derived from the

band

widthofequilibrium distributions ofT2andT4

deoxyribonucleic

acidareknownto

be low bya factor of2 (19), and the deviation found in the present determinations may be assumed to be due to the same type of error, whatever that may be. This

fact,

plus

the pro-nounced deviation

of

the distributions of de-naturedRIfrom the Gaussian

distribution,

made itpointless toattempt an estimate

of

the molec-ular

weight

ofthe denatured RI from the band width of theequilibrium distribution in

CS2SO4

. However, the distributions

clearly

showed

that,

whereas the bulk of denatured RI was

single-stranded

(po

= 1.627to

1.631),

there was stilla

double-stranded

component.

The number distribution ofchain

lengths

was

predicted to be

composed

of

equal

numbers of molecules with chain

lengths

of 2 to X-1

plus

a

largernumber of molecules with chain

lengths

of X (cf.

7).

The relative number of molecules of chain length X

depends

on two factors.

(i)

The first factor is the average number of

single

520 J.VIROL.

on November 11, 2019 by guest

http://jvi.asm.org/

strandsperdouble-stranded template, whichwas about 1 for the present preparations of RI (7). On this basis, therefore, there should be two molecules of chain length X for every molecule of chain length less than X. (ii) This assumes, however,that completion of chains, i.e., addition of the terminal nucleotide to a nascent chain of length X-1, results in displacement from RI, so that thelongest chaintobefoundassociated with RI maybe oneof length X-1.

Fromthenumber distribution, it isdifficult to obtainmorethan a first-order approximation of the ratio of numbers of molecules of length Xto the number of molecules of length less than X. This is because there is considerable inherent error in the estimate of the number distribution used here, especially at chain lengths less than about200(see nextparagraph).Also,thevariable amountof residual double-stranded RNA inthe population

of

denatured RI will affect the result (see nextparagraph). Anestimate of the relative amountsof material of chain length X andlengths

<X

wasmade by planimetry ofa linear plot of Nversuschainlength.Notethat forconvenience

only

semilogarithmic plots areshownin Fig. 6B and7B.This ratio is0.4rather than the

predicted

value of 2. This would indicatean uppervalueof five single strands (or five growing points) per replicative intermediate. Thus, thepresent experi-mentsandearlierexperiments (7,8) indicate that thereareprobably betweenoneand five growing

points

perRI molecule.

Betweenchain lengths of200and 1,100, there isasatisfactory fit of the numberdistribution to that predicted. Deviations occurring at very low chain lengths are due to two factors. Chains of very short length may be lost during dialysis of denaturedRI.Also, the slopeoftheconcentration distribution during velocity sedimentation isvery small and, therefore,very difficultto measure at values corresponding to the smallest chain lengths. Thus, large errors arise inthisportionof theweightdistribution curveand subsequentlyin the number distribution transformation. The deviation from equal numbers ofmolecules be-tween1,200 and3,000 was seen inallexperiments. The amount of material in this region was some-what higher in DMSO-denatured RI than in heat-denatured RI. Thisleads to thespeculation that thisrepresents the residual double-stranded material. When analysis of hyperchromicity curvesforheat-denaturedRIledto anestimateof 6.4%double strand, the area under the bulge in the number distribution curve was 6.2%. On the other hand, RI treated with

DMSO

for 10 min at 37 C had a 16% residual double-stranded component when analyzed by hyperchromicity, but the area under the bulge in the number

distribution curve was only 7%. In view of the difficulties inherent in this experiment and the problem of the residual double-stranded com-ponent, the fit of the number distribution of denatured RI to the predicted distribution is quite satisfactory. It is not clear asyet whether the double-stranded residue is a special class of double-stranded molecules or some partially denatured RI. No evidence of a special type of double-stranded RNA, suchascircular RNA,was seen inelectron micrographs of RF and RI (12) or denatured RI (Granboulan and Franklin, in preparation).

ACKNOWLEDGMENTS

I am grateful toJohn Cann, Department of

Bio-physics, University of Colorado SchoolofMedicine,

for the useofhis model Eanalytical ultracentrifuge.

Elizabeth Hinckleypreparedthenucleic acidsused in

this study, and additional technical assistance was

providedby MarianneSalditt.

Thisinvestigation wassupported by PublicHealth

Service grant AI 07645fromtheNationalInstituteof

Allergy and Infectious Diseases and by grant GB

5365 from theNationalScience 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. DoTy, P., H. BOEDTKER, J. R. FRESCO, R.

HASELKORN, AND M. LITT. 1959. Secondary

structureinribonucleicacids. Proc. Natl.Acad.

Sci. U.S. 45:482-499.

3. ERIKSON, R. L. 1966. Fractionation of

viral-specific ribonucleic acid by cesium sulfate

equilibrium density-gradient centrifugation. J. Mol. Biol. 18:372-381.

4. ERIKSON, R. L., AND R. M. FRANKLIN. 1966.

Symposiumonreplicationof viralnucleicacids. I. Formation and properties of a replicative intermediate in the biosynthesis of viral

ribo-nucleicacid. Bacteriol. Rev. 30:267-278.

5. ERIKSON, R. L., AND W. SZYBALSKI. 1964. The

Cs2SO4 equilibrium density gradient and its

application for the study of T-even phage DNA: glucosylation and replication. Virology

22:111-124.

6. 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.

7. FRANKLIN, R. M. 1966. Purification and

prop-erties of the replicative intermediate of the

RNA bacteriophage R17. Proc. Natl. Acad.

Sci. U.S.55:1504-1511.

8. FRANKLIN, R. M. 1967. Replication of

bacterio-phage RNA:somephysical propertiesof

on November 11, 2019 by guest

http://jvi.asm.org/

FRANKLIN

stranded, double-stranded, and branched viral

RNA. J.Virol. 1:64-75.

9. FRANKLIN, R. M., AND N. GRANBOULAN. 1966.

Ultrastructure of Escherichia coli cellsinfected

with bacteriophage R17. J. Bacteriol. 91:834-848.

10. GESTELAND,R.F.,ANDH. BOEDTKER. 1964. Some

physical properties of bacteriophage R17 and

itsribonucleic acid. J. Mol. Biol. 8:496-507.

1t. GOMATOS, P. J., AND I. TAMM. 1963. The

sec-ondarystructure of reovirus RNA. Proc. Natl.

Acad.Sci. U.S. 49:707-714.

12. GRANBOULAN, N., AND R. M. FRANKLIN. 1966.

Electron microscopy of viral RNA, replicative

form, and replicative intermediate ofthe

bac-teriophage R17. J. Mol. Biol. 22:173-177.

12a. HOFSCHNEIDER, P.,J.AMMAN, AND B. FRANCKE.

1967. Infectious M12 phage replicative form

RNA: A tool for studyingviral RNA

replica-tion. In J. S. Colter [ed.], Proc. Symp. Mol.

Biol. Virus, Academic Press, Inc., New York,

inpress.

13. JENSEN, R. H.,ANDN. DAVIDSON. 1966.

Spectro-photometric, potentiometric, and density

gradient ultracentrifugation studies of the

binding ofsilver ionby DNA. Biopolymers 4:

17-32.

14. KAERNER, H. C., AND H. HOFFMAN-BERLING.

1964. Synthesis of double-stranded RNA in

RNA-phage infected E. colicells. Nature 202:

1012-1013.

15. KATZ, L., AND S. PENMAN. 1966. The solvent

denaturation of double-stranded RNA from

poliovirus infected HeLa cells. Biochem.

Bio-phys.Res. Commun. 23:557-560.

16. MONTAGNIER, L., AND F. K. SANDERS. 1963.

Replicative form of encephalomyocarditis

virusribonucleic acid. Nature199:664-667.

17. SCHUMAKER, V. N.,ANDH. K. SCHACHMAN. 1957.

Ultracentrifugal analysis of dilute solutions.

Biochim.Biophys. Acta 23:628-639.

18. SINHA, N. K., R. K. FUJIMARA,ANDP. KAESBERG.

1965. Ribonuclease digestion of R17 viral

RNA. J. Mol.Biol. 11:84-89.

19. THOMAS, C. A., JR., AND K. I. BERNS. 1961. The

physical characterization of DNA molecules

releasedfrom T2 and T4 bacteriophage.J.Mol.

Biol. 3:277-288.

20.

VASQUEZ)

C., N. GRANBOULAN, AND R. M.

FRANKLIN. 1966. Structure of the ribonucleic

acid bacteriophage R17. J. Bacteriol.

92:1779-1786.

21. VINOGRAD, J., AND J. E. HEARST. 1962.

Equilib-rium sedimentation of macromolecules and

viruses inadensity gradient. Fortschr. Chem.

Org. Naturstoffe 20:372-422.

522 J.VIROL.

on November 11, 2019 by guest

http://jvi.asm.org/