JOURNAL OF BACTERIOLOGY, Feb., 1966Copyright © 1966 American Society for Microbiology

Vol. 91, No. 2Printed in U.S.A.

Biophysical Studies of Vesicular Stomatitis VirusROBERT M. McCOMBS, MATILDA BENYESH-MELNICK, AND JEAN P. BRUNSCHWIG

Department of Virology and Epidemiology, Baylor University College of Medicine, Houston, Texas

Received for publication 3 September 1965

ABSTRACT

MCCOMBS, ROBERT M. (Baylor University College of Medicine, Houston, Tex.),MATILDA BENYESH-MELNICK, AND JEAN P. BRUNSCHWIG. Biophysical studies ofvesicular stomatitis virus: J. Bacteriol. 91:803-812. 1966.-The infectivity andmorphology of vesicular stomatitis virus (VSV) were studied after density gradientcentrifugation in cesium chloride (CsCI), potassium tartrate (KT), and sucrose.Centrifugation in CsCl revealed two equally infectious bands corresponding todensities of 1.19 and 1.22 g/ml, and a third (density, 1.26 g/ml) band of low infec-tivity. Two bands (densities of 1.16 and 1.18 g/ml) were observed in the KT gradient,in which the lighter band contained most of the infectivity. Centrifugation in sucroseresulted in a single broad infectious band (density, 1.16 g/ml). The typical rod-shaped VSV particles were found mainly in the lighter bands obtained in CsCl (1.19g/ml) and KT (1.16 g/ml) and in the single sucrose gradient band (1.16 g/ml). Bentparticles equally as infectious as the rod-shaped particles were a constant findingin the CsCl preparations, and were observed mainly in the second band (density,1.19 g). Numerous strands 15 m,u wide were found in the third CsCl (density, 1.26g/ml) and the second KT (1.18 g/ml) bands. Similar strands could be liberatedfrom VSV particles after treatment with deoxycholate. Internal transverse striationswere found to be a regular feature of VSV particles examined with the pseudorepli-cation negative-staining technique. For crude virus stocks, the physical particle-to-infectivity ratio ranged from 73 to 194. Several morphological similarities betweenVSV and myxoviruses were observed, including 10 m, surface projections, pleomor-phic morphological forms, and 15 mMi seemingly nucleoprotein strands.

The morphology of vesicular stomatitis virus(VSV) has been studied by several investigators(4, 5, 9) using either the negative staining proce-dure of Brenner and Home (2) or the modificationby Howatson and Whitmore (5). The virus par-ticles were shown to be rod-shaped, rounded atone end and flat at the other, and measuringabout 65 by 175 m,u. Other structural details thatwere observed included surface projections, acentral axial hollow extending through a portionof the particle from the flat end, and occasionallyinternal transverse striations. A buoyant densityof 1.20 g/ml in cesium chloride (CsCl) (40-hrcentrifugation) has been reported for VSV (8).

In the present study, VSV was found to band ina different fashion after density gradient centrifu-gation in CsCl for short (3.5 hr) and long (24 hr)periods of time. The morphology and quantifica-tion of the particles after density gradient centrif-ugation and after chemical treatment was studiedby use of the pseudoreplication-negative stainingprocedure as described by Smith and Melnick(12). Some of the particles recovered from the

density gradient bands, as well as particles ob-tained after chemical treatment, exhibited a strik-ing similarity to typical myxoviruses.

MATERIALS AND METHODSTissue culture. Serially propagated human em-

bryonic lung (HEL) fibroblasts were used. These cellswere grown in Eagle's medium with 10% fetal bovineserum and 0.075% sodium bicarbonate (for cells instoppered vessels) or 0.225% sodium bicarbonate(for cells in petri dishes in a 5% CO2 atmosphere).Primary monkey kidney (MK) cells grown in l-ozprescription bottles with the use of the proceduresand the M-H medium described by Melnick (6) werealso employed.

Virus. The Indiana serotype of VSV was kindlysupplied by Werner Henle. Virus stocks were preparedin HEL fibroblasts. Infected monolayers showing 80to 90% cytopathic effects were frozen and thawedtwice, and the crude suspension was pooled and clari-fied by low-speed centrifugation. The resulting super-natant fluid was stored at -90 C. The virus for den-sity gradient centrifugation and chemical treatmentwas concentrated 10- to 20-fold and was partiallypurified by two cycles of differential ultracentrifuga-

803

on May 3, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

McCOMBS, BENYESH-MELNICK, AND BRUNSCHWIG

tion (19,620 X g for 15 min; 54,500 X g for 60 min)and resuspended in 0.05 M tris(hydroxymethyl)amino-methane (Tris) buffer (pH 7.4).

Infectivity assay. VSV was titrated by plaque as-say by use of either MK monolayers in 1-oz prescrip-tion bottles or HEL monolayers in 60-mm petridishes. Decimal virus dilutions in 0.05 M Tris bufferwere used. For assays in MK cells, 0.1 ml of virus wasadded to drained monolayers and, after absorption at37 C for 1 hr, the cultures were overlaid with 5 ml ofEarle's salt solution containing 0.4% sodium bicar-bonate, 0.0017% neutral red, 0.1% skim milk, and1.5% agar (Difco) (14). The HEL monolayers werefirst washed with Tris buffer and then infected with 0.5ml of virus. After 1 hr of incubation at 37 C, they wereoverlaid with 5 ml of Eagle's medium containing 5%fetal bovine serum, 0.9% agar, and 0.225% sodiumbicarbonate. A second overlay (3 ml) of the same com-position, but with neutral red (final concentration1:25,000), was added after 48 hr, and the plaques werecounted after overnight incubation. Since assays onMK and HEL monolayers yielded identical results,they were used interchangeably throughout this study.

Density gradient centrifugation. VSV was bandedby density gradient centrifugation in cesium chloride(CsCl), potassium tartrate (KT), potassium citrate(KC), and sucrose. In general, 0.5 ml of the partiallypurified and concentrated virus was layered over 4.5ml of preformed gradient and centrifuged at 99,972X g for 3.5 hr at 4 C in the SW 39 rotor of a Spincomodel L2 ultracentrifuge. The bands obtained wereremoved either from the side of the tube by use of asyringe and needle (25 g), or by puncturing the bottomof the tube with a Buchler piercing unit (BuchlerInstruments, Inc., Fort Lee, N.J.) and collecting thedrops. In the latter procedure, 80 to 90 fractions of5 drops each were collected. The density of the bandswas computed by measurement of the refractive indexin an Abbe 3L refractometer, and the density wasextrapolated from standard curves obtained by use ofdata from the International Critical Tables.

Chemical treatment. Partially purified VSV prep-arations were treated with trypsin, sodium deoxy-cholate, or sodium lauryl sulfate under various condi-tions (see Results). Immediately after incubation withthe reagent, specimens were prepared for electronmicroscopy.

Electron microscopy. Two methods of specimenpreparation were employed. The first was that de-scribed by Brenner and Home (2), in which an equalvolume of 2% phosphotungstic acid (PTA) at pH7 and virus suspension were mixed. A drop of thismixture was placed on a collodion-coated steel grid(200 mesh), and the excess was removed after severalminutes. The second method was the pseudoreplica-tion method described by Smith and Melnick (12)and later modified to include osmic acid fixation (11).A drop of the virus suspension was first allowed todialyze to dryness on 2% agar and was then immedi-ately fixed in osmic acid vapor for 7 min. Parlodion(0.75% in amyl acetate) was spread over the surfacecontaining the virus and, after drying, the film wasstripped from the agar by floating on a 0.2% aqueoussolution of PTA (pH 7) containing 0.05% sucrose.

The film was picked up on a steel grid and air-driedprior to electron microscopic examination. All photo-micrographs were taken with a Hitachi 11A electronmicroscope at instrument magnifications of 5,000 and15,000.

Particle counts were initially made by the sedi-mentation technique (10) with the use of the modifi-cation for negative staining (12). Later, it was foundthat the droplet pseudoreplication method, withoutsedimentation, yielded reproducible particle counts(see Results), and this technique was also used forthe enumeration of particles in the density gradientbands.

RESULTS

Morphology of VSV. The two methods of nega-tive staining with PTA gave significantly differentresults. The particles seen in Fig. la were stainedby the procedure of Brenner and Home (2), andhad the morphology previously described byothers (4, 5, 9). The surface projections were evi-dent, and the PTA had penetrated only a centralhollow area extending part way from the flat endof the particle. When the pseudoreplicationmethod of Smith and Melnick (12) was used, theparticles had a different appearance (Fig. lb). Thesize of the particles prepared by this method was65 by 175 mju, in good agreement with that pre-viously reported. Considerably more internaldetail, in particular the internal transverse stria-tions, was routinely seen, probably because ofpenetration of the PTA. Such internal structuraldetail was observed by the above-mentionedauthors mainly in particles that were examinedafter prolonged incubation at 4 C and that wereconsidered to be partially disrupted.The technique of pseudoreplication was em-

ployed for all of the following electron micro-scopic examinations.

Virus particle quantification. A comparison ofparticle counts determined by the sedimentationmethod and the droplet method was carried out.Two different unconcentrated stocks of VSV wereused. The infectivity titers were 1.6 x 107 plaque-forming units (PFU) per ml for stock A and 1.5 X107 PFU per ml for stock B. For the sedimentationmethod, the formula described by Smith andMelnick (13) was applied for the calculation ofthe particle concentration per milliliter of virussuspension: particles/ml = (average number ofparticles/field) X (1/dilution) X (1/height ofcolumn) X (1 /unit area in cm2) = (average num-ber particles/field) X (1/dilution) x (1.25) x(1.2 X 106).The calculations for the droplet method were

as follows. It was found in repeat measurementsthat 1 ml of virus suspension contained 50 dropsof the virus suspension used for the examina-

804 J. BACrRIOL.

on May 3, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

VESICULAR STOMATITIS VIRUS

FIG. la. VSV particles negatively stained by the method ofBrenner and Horne (2). X 150,000.FIG. lb. VSV particles negatively stained after pseudoreplication by the procedure of Smith and Melnick (12).

X 150,000. The bar in this and subsequent micrographs represents 100 m,u.

tions. Thus, the following formula was applied:particles/ml = (average number of particles/field) X (1/dilution) x (50) X (1/unit area incm2) = (average number of particles/field) x(1/dilution) X (50) X (1.2 x 106).A good correlation was found between the

counts obtained and the values determined bythe two methods in two experiments with virusstock A and one experiment with virus stock B.Stock A yielded 3.1 x 109 particles per milliliterby the sedimentation method and 2.1 X 109 par-ticles per milliliter by the droplet method. Thevalues determined for stock B were 1.1 X 109 and1.6 X 109 particles per milliliter, respectively.Because of its simplicity, the droplet method wasemployed in all further experiments.The ratio of physical particles to infectivity

(PFU) calculated from these data varied from 73to 194, with an average of 126.

Density gradient centrifugation of VSV. In pre-liminary experiments, virus was incubated at 4 Cfor 3.5 hr with the appropriate compound usedfor making the gradient, at a concentration cor-responding to the approximate buoyant densityof the virus. About 50% infectious virus was lostafter incubation with CsCl, KT, and KC. Abetter recovery (79 to 100%) was obtained afterincubation with sucrose. Gradients prepared inTris buffer yielded sharper bands as compared

with those prepared in water, and thus wereutilized throughout the centrifugation experi-ments.The bands obtained after equilibrium density

gradient centrifugation (3.5 hr at 99,972 X g) inCsCl, KT, and sucrose are illustrated in Fig. 2.Three bands were obtained in CsCl, two bands inKT, and a single band in sucrose. The correspond-ing buoyant density of each band is indicated inthe figure. The bottom bands in both the CsCland KT had a granular appearance.The distribution of infectivity in fractions col-

lected from the bottom of the tube with the threegradients is illustrated in Fig. 3, 4, and 5. Thepeaks of infectivity shown in these figures corre-sponded to the visible bands seen in Fig. 2. Resultsof density gradients with KC were similar to thosefound in KT, and will not be presented.The separate bands in all three gradients were

collected from the side of the tube and examinedin an electron microscope. The morphologicalforms observed in each of the three CsCl bandsare presented in Fig. 6. The first band (density,1.19 g/ml) contained the typical rod-shaped virusparticles, as well as some "bent" particles thatappeared to be broken almost exactly in half andfolded together (Fig. 6a). The second band (den-sity, 1.22 g/ml) consisted almost entirely of thebent particles (Fig. 6b). The third band (density,

VOL. 91, 1966 805

on May 3, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

MCCOMBS, BENYESH-MELNICK, AND BRUNSCHWIG

1.26 g/ml) contained strands, about 15 m, indiameter, as well as some rod-shaped and bentparticles trapped within the strands (Fig. 6c),which probably caused the granular nature of theband as seen in the centrifuge tube.

In KT, band 1 (density, 1.16 g/ml) containedintact, rod-shaped particles and some bent forms,whereas band 2 (density, 1.18 g/ml) containedmasses of strands which had entrapped manyparticles, thus bearing a great similarity to the ap-pearance of the third CsCl band. In the sucrosegradient, the single band obtained (density, 1.16

- 1.191.22

- 1.26

1.16I 1.18

- 1.16

CsCI KT Suc

1.20

1.18 I-

z

a)1.14

30 40 50 60 70 s0FRACTION NUMBER

FIG. 4. Distribution of VSV infectivity in fractionsafter density gradient centrifugation in potassiumtartrate (3.5-hr centrifugation at 99,972 X g). Symbols:0 = density; 0 = infectivity.

FIG. 2. CsCl, potassium tartrate and sucrose densitygradients of VSV (3.5-hr centrifugation at 99,972 Xg) and the buoyant density of the bands obtained.

CESIUM CHLORIDE

7r-

-6

>

1. I

1.24 FI.22~V'1.20 w

SO.1a

20 30 40 50 60 70

FRACTION NUMBER

FIG. 3. Distribution of VSV infectivity in fractionsafter density gradient centrifugation in CsCl (3.5-hrcentriguation at 99,972 X g). Symbols: 0 = density;0 = infectivity.

SUCROSE

_4r-

I I I

1.20

.18 >.I.-

z1.16 o

1.14

10 20 30 40 50 60

FRACTION NUMBER

FIG. 5. Distribution of VSV infectivity in fractionsafter density gradient centrifugation in sucrose (3.5-hrcentrifugation at 99,972 X g). Symbols: 0 = density;0 = infectivity.

g/ml) consisted of intact, rod-shaped particles, aswell as some bent particles and isolated strands.A quantitative correlation between infectivity

and particle counts of the density gradient bands,

806 J. BAUEERIOL.

7

-J

U.)

z

2CD0 5-J

4

on May 3, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

VESICULAR STOMATITIS VIRUS

C

FIG. 6. Electron micrographs ofportions of the band occurring at (a) density 1.19 glml in CsCI, (b) density1.22 glml in CsCl, and (c) density 1.26 glml in CsCl. X 35,000 (inserts, X 150,000).

807VOL. 91, 1966

on May 3, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

MCCOMBS, BENYESH-MELNICK, AND BRUNSCHWIG

collected from the side of the tube, was carriedout. An average of the results of two experimentsis given in Table 1. CsCl band 1 contained 48%oi the total infectivity recovered in the threebands; band 2 contained 44% and band 3 con-

tained 8%. Upon further centrifugation in CsCl,tor 24 hr, much of the infectivity previously con-

tained in band 1 was found in the second band.In KT, band 1 contained 36% of the infectivityrecovered and band 2 contained 64%. Of partic-ular interest were the particle counts made frompseudoreplicas of droplet preparations taken fromeach band. There was a high degree of correlationbetween the per cent infectivity (PFU) and theper cent total particles recovered from the indi-vidual bands. From the data presented in the twolast columns in Table 1, it appeared that both therod-shaped and bent particles were equally infec-tious. This is illustrated clearly in the CsCl experi-ments. After 3.5 hr of centrifugation, the infectiv-ity was highest in the first band. This band alsocontained 54% of the total particles counted, withnearly equal numbers of rod-shaped and bentforms. However, after 24 hr of centrifugation,63% of the infectivity and 51 % of the total par-ticles were found in the second band. The increasein particles consisted of an increase in the bentforms rather than the rod-shaped forms.

Morphological degradation of VSV. The effectof several chemical agents on the morphology ofVSV was studied. Partially purified virus was in-cubated in the presence of 0.05% trypsin (30 minat 37 C) with approximately 50% loss of infectiv-

ity. The particles appeared devoid of the 10-m,isurface projections or spikes (Fig. 7a), as com-pared with the control, nontreated preparation(Fig. 7b). However, there seemed to be a trypsin-resistant area at the rounded end of the particles,as seen in the insert of Fig. 7a. No other morpho-logical alteration of the particles was observed bythis treatment.Complete disruption of VSV particles, leaving

only the 15 m,u strands, was achieved using 0.05%sodium deoxycholate (30 min at 37 C; Fig. 7c).Evidence that these strands form an integral com-ponent of the virion is illustrated in Fig. 8, inwhich strands can be seen emanating from virusparticles obtained from the third band in CsCl orfrom the second band in KT. The strands wereassociated with the partially disrupted particles,with most strands originating from the inner por-tion of the particle. When these strands, obtainedwith the deoxycholate treatment, were furthertreated with trypsin, they were no longer detect-able in the electron microscope, indicating a com-plete digestion by the enzyme. Preliminary resultsindicated that neither ribonuclease nor deoxyribo-nuclease altered the morphology of these strands.Treatment of VSV particles with 0.5% sodium

lauryl sulfate (10 min at room temperature) re-sulted in complete disruption of the particles;electron microscopy revealed that the preparationwas devoid of either particles or strands.

Occurrence of myxovirus-like forms in VSVpreparations. The electron microscopic examina-tion of VSV preparations, in particular those

TABLE 1. Quantitative correlation between infectivity (PFU) and particle counts of density gradient bandswith VSV

Centrifugation* Per cent per bandt

ParticlesGradient Time Band no. Density PFU

Total Rod forms Bent forms

hr g/liCsC1 3.5 1 1.19 48 54 31 23

2 1.22 44 36 7 293 1.26 8 10 7 3

24 1 1.19 31 34 17 172 1.22 63 51 3 483 1.26 6 15 4 11

Potassium tartrate 3.5 1 1.16 36 48 27 212 1.18 64 52 21 31

Sucrose 3.5 1 1.16 100 100 70 30

* At 99,972 X g.no. per band

t Per Cent = total no. recovered X 100

808 J. BA=rRIOL.

on May 3, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

VESICULAR STOMATITIS VIRUS

a

r6m

FIG. 7. Electron micrographs of VSV particles (a) after treatment with trypsin, (b) in the control preparation,and (c) after treatment with sodium deoxycholate. X 35,000 (inserts, X 150,000).

treated with CsCl, revealed particles closely re-sembling myxoviruses. Several of these particlesare illustrated in Fig. 9. Particle "a" had the usualrod-shaped appearance, while particles "b" and"c" seemed to be assuming the pleomorphic shapecharacteristic of myxoviruses. The morphologyof particle "d" appeared indistinguishable fromthat of typical myxoviruses, including pleo-morphic shape, surface projections, and even anill-defined internal helix.

DiscussIoN

Reczko (9), Howatson and Whitmore (5), andHackett (4) were the first to study in detail the

morphology of VSV by use of the negative stain-ing technique of Brenner and Horne (2). In theirpreparations, typical virus particles were pene-trated by PTA only in the area of the central axialhollow, and the internal structure was seldomvisible, although the surface projections wereclearly outlined. The failure of the PTA to pene-trate intact VSV particles seemed to be due tothe presence of a lipid membrane surrounding theparticles (8). With the pseudoreplication methodemployed in this study, the parlodion used to stripthe virus from the agar surface was applied as asolution in amyl acetate. This solvent probablysolubilized enough of the outer lipid to allow thePTA to penetrate.

809VOL. 91, 1966

::::::-:.1:;.

I-M

.: -TIM, ::.,

;:M

W

.4 31"

t.,i.

-A&VA jf,:

.4.. 1.

.Wr

on May 3, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

McCOMBS, BENYESH-MELNICK, AND BRUNSCHWIG

FIG. 8. VSV particles showing the relation of the 15-m, strands to the virus particle. Top, from CsCI thirdband; bottom, from potassium tartrate second band. X 150,000.

Particle counts obtained by the sedimentationand the droplet methods were in agreement, thusindicating that the droplet method can be usedas a rapid and simple means of quantification.The average ratio of physical particles to PFUobtained for the two unconcentrated virus stocksuspensions was 126, varying from 73 to 194.Hackett (4) taking into consideration only un-penetrated intact particles, reported a ratio closeto 10 for potent VSV preparations. With thepseudoreplication technique used in our study, allparticles were penetrated by the PTA and theratio obtained was based on total counts of in-tact particles.VSV was previously banded in CsCl by Prevec

and Whitmore (8), who reported a single band ofinfectivity having a density of 1.20 g/ml, but theydid not study the morphology of the particles in

this band. After 3.5 hr of centrifugation of VSV,we found that the peaks of infectivity in bands 1and 2 were approximately equal, but, upon furthercentrifugation for 24 hr, the second band (con-taining bent forms) increased in content over thefirst band (Table 1). If centrifugation was con-tinued for 40 hr, as done by Prevec and Whitmore(8), the second band would probably increase inquantity so that the first and third bands wouldbe obscured, thus giving the appearance of asingle, broad band.The recovery of three bands of infectivity from

CsCl density gradients, compared with a singleband from sucrose gradients, may be explainedon the basis of the morphology of the particlesrecovered. As a result of centrifugation in CsCl,the VSV particles were partially disrupted intobent forms (seen in band 2) that may retain their

810 J. BACTrERIOL.

on May 3, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

VESICULAR STOMATITIS VIRUS

FIG. 9. VSV particles (from CsCI-treated preparationis) showilig varyinlg degrees of myxovirus-like structure.X 150,000.

infectivity, and were completely disrupted into15 m,u strains (seen in band 3). KT also causedthe disruption of particles into bent forms andthe 15 m,u strands. This gradient salt, however,was not able to resolve the rod-shaped and thebent forms into different bands. The high percent-age of infectivity (64%c) found in the second bandwas due to large numbers of virus particlestrapped in the masses of strands.Even though VSV had been widely used, it re-

mains an unclassified animal virus. Mussgay andSuarez (7) suggested that VSV was related to thearboviruses because of its multiplication in mos-quitoes, maturation at the cell membrane, and thepresence of lipid as an integral viral component.However, Howatson and Whitmore (5) suggestedthat the surface projections seen on the virus par-ticles closely resembled those of the myxoviruses.In our study, three distinct similarities betweenVSV and the myxoviruses were noted. The firstand most common was the spikes (10 my in size)seen on the VSV particles. In addition, the spikescould be removed with trypsin, as has been re-ported for some myxoviruses (1, 3). The secondsimilarity was the occurrence of myxovirus-likeforms in the preparations, and the third similaritywas the strands present in the bottom bands andafter treatment with deoxycholate. The viralorigin of these strands is well illustrated in Fig. 8.The strands measured about 15 m,u in diameter,which is similar to that of the nucleoproteinstrands reported for the paramyxoviruses (15).

ACKNOWLEDGMENTSThis investigation was supported by Public Health

Service Research Grant Al 05382 and Training Grant5 TI Al 74 from the National Institute of Allergy andInfectious Diseases.The assistance of David Lennon is greatly appre-

ciated.

LITERATURE CITED1. BORECKY, L., V. LACKOVIC, AND E. MRENA.

1964. The effect of trypsin on influenza virus.Acta Virol. 8:555.

2. BRENNER, S., AND R. W. HORNE. 1959. A nega-tive staining method for high resolution elec-tron microscopy of viruses. Biochim. Biophys.Acta 34:103-1 10.

3. GRESSER, I., AND J. F. ENDERS. 1961. The effectof trypsin on representative myxoviruses. Virol-ogy 13:420-426.

4. HACKETT, A. J. 1964. A possible morphologicalbasis for the autointerference phenomenon invesicular stomatitis virus. Virology 24:51-59.

5. HOWATSON, A. F., AND G. F. WHITMORE. 1962.The development and structure of vesicularstomatitis virus. Virology 16:466-478.

6. MELNICK, J. L. 1956. Tissue culture methods forthe cultivation of poliomyelitis and otherviruses, p. 97. In Diagnostic Procedures forViral and Rickettsial Diseases, 2nd ed. Ameri-can Public Health Association, New York.

7. MUSSGAY, M., AND 0. SUAREZ. 1962. Multipli-cation of vesicular stomatitis virus in Aedesaegypti (L.) mosquitoes. Virology 17:202-204.

8. PREVEC, L., AND G. F. WHITMORE. 1963. Purifi-cation of vesicular stomatitis virus and the

811VOL. 91,~1966

on May 3, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

McCOMBS, BENYESH-MELNICK, AND BRUNSCHWIG

analysis of P3I-labeled viral components.Virology 20:464-471.

9. RECZKO, E. 1960. Electron microscopic studiesof vesicular stomatitis virus. Arch. Ges. Virus-forsch. 10:588-05.

10. SMITH, K. O., AND M. BENYESH-MELNICK. 1961.Particle counting of polyoma virus. Proc. Soc.Exptl. Biol. Med. 107:409-413.

11. SMITH, K. 0., M. BENYESH-MELNICK, AND D. J.FERNBACH. 1964. Studies on human leukemia.It. Structure and quantitation of myxovirus-like particles associated with human leukemia.J. Natl. Cancer Inst. 33:557-570.

12. SMITH, K. O., AND J. L. MELNICK. 1962. A methodfor staining virus particles and identifying their

nucleic acid type in the electron microscope.Virology 17:480-490.

13. SMITH, K. O., AND J. L. MELNICK. 1962. Electronmicroscopic counting of virus particles bysedimentation on aluminized grids. J. Immunol.89:279-284.

14. WALLIS, C., J. L. MELNICK, AND M. BIANCHI.1962. Factors influencing enterovirus andreovirus growth and plaque formation. TexasRept. Biol. Med. 20:693-702.

15. WATERSON, A. P., AND J. G. CRUICKSHANK.1964. Form of the nucleoprotein componentas a taxonomic criterion for the myxoviruses.Nature 201:640-641.

812 J. BACTERIOL.

on May 3, 2018 by guest

http://jb.asm.org/

Dow

nloaded from