Cytoplasmic Structures Associated with an Arbovirus Infection: Loci of Viral Ribonucleic Acid Synthesis PHILIP M

Home , Semliki Forest virus

JOURNAL OF VIROLOGY, Nov. 1968, p. 1326-1338 Vol. 2, No. 11 Copyright @ 1968 American Society for Microbiology Printed in U.S.A. Cytoplasmic Structures Associated with an Arbovirus Infection: Loci of Viral Ribonucleic Acid Synthesis PHILIP M. GRIMLEY, IRENE K. BEREZESKY, AND ROBERT M. FRIEDMAN Pathologic Anatomy Branch and Laboratory of Pathology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20014 Received for publication 31 July 1968 Unique cytoplasmic structures, herein designated as type I cytopathic vacuoles (CPV-I), are found in chick embryo cells early in the logarithmic phase of Semliki Forest virus replication. High resolution autoradiography demonstrated that the CPV-I are loci of 3H-uridine incorporation. This evidence correlates well with previous biochemical data and electron microscopy of the subcellular fractions active in Semliki Forest virus ribonucleic acid synthesis. Origin of the CPV-I within host cell cytoplasm is confirmed by the distribution of electron-dense tracer particles and sequential ultrastructural observations.

The formation of unique intracytoplasmic sodium deoxycholate (12, 14). We demonstrated vacuoles within cells infected by Semliki Forest the participation of CPV-I in viral RNA synthesis virus (SFV) and related members of the arbovirus by means of high resolution autoradiography. group has been demonstrated by electron micros- Ultrastructural studies of the infected cells at copy in both cell cultures and the cerebral tissues sequential intervals and the distribution of an of experimental animals (21, 24, 32). The cyto- electron-dense tracer have provided additional pathic vacuoles (CPV) may be subdivided into information concerning the origin and fate of two morphologically distinct groups. (i) A type the arbovirus CPV. of CPV recently recognized in chick embryo cells infected by SFV (1, 14) or Sindbis virus (un- MATERIALS AND METHODS published data). These occur relatively early in Virus and cell cultures. The Kumba strain of the infectious cycle and are, therefore, designated Semliki Forest virus was grown at 37 C in primary CPV-I. They are characterized by a regular series cultures of chick embryo cells (15). Monolayers of of membrane buds or spherules which are at- chick embryo fibroblasts were infected at virus-cell tached to the interior surface of the vacuoles. The multiplicities of 20:1 to 500:1. Virus was titrated by buds or spherules are formed of membranes with plaque assay (16). In some experiments, virus was unit structure (1), and may contain an electron- absorbed to the cells overnight at 4 C, either in dense central spot. (ii) The type of CPV that was medium containing 1 jug of actinomycin D per ml or originally described in cells infected by Western in medium without actinomycin D. In other experi- equine encephalitis virus (23). These vacuoles are ments, virus was added directly to the monolayers, with or without actinomycin D, and maintained at encircled on the cytoplasmic aspect by closely 37 C. Cultures were studied at selected intervals (0.5 spaced particles of uniform size (25 to 30 nm). to 16 hr) after the initiation of infection. In cold (4 C) They are most numerous in the late (linear) phase virus absorption, the period of infection was con- of the virus replication cycle and are designated sidered to begin from the moment of rewarming to CPV-II. CPV-II have been observed in cell 37 C (13, 15). Eight separate experiments were con- systems infected by SFV and several other group ducted to study the ultrastructural sequence of virus A arboviruses (3, 10, 23-26). development at relatively low multiplicity of infection Preliminary studies in this laboratory showed (20:1 to 40:1). In addition, two experiments employ- that CPV-I arise the early stages of SFV ing a high virus multiplicity (500:1) were conducted during to trace virus entry and to detect early evidence of infection and are present in the cytoplasmic frac- replication (up to 3 hr). The preparation of high titer tions which incorporate 3H-uridine (14). In SFV SFV pools for experiments requiring a high virus to infection, the new viral ribonucleic acid (RNA) is cell multiplicity has been reported previously (13). closely associated with cytomembranes and may Electron microscopic preparations. Cell monolayers be released from the membrane sediment by were scraped from the petri dish with a rubber police- 1326 VOL. 2. 1 968 ARBOVIRUS INFECTION 1327 man and transferred to centrifuge tubes along with the RESULTS original culture fluid. Some of the supernatant fluid was withdrawn before centrifuging at 785 X g for Virus grouwth chlaracteristics. In the chick 5 min. The resulting pellet was fixed for I to 2 hr embryo cells treated with actinomycin D and with 3'( glutaraldehyde (4) in 0.1 M phosphate buffer incubated at 37 C with 20 plaque-forming units (pH 7.3). In experiments employing Thorotrast (PFU) of SFV per cell, the titers of virus infec- (thorium dioxide) or sH-uridine, cell monolayers were tious units increased exponentially during a period fixed directly in the petri dishes before centrifugation. of I to 5 hr (Fig. 1). After 8 hr, production of Glutaraldehyde-fixed pellets were washed in 0.1 M infectious virus reached a plateau. The rising phosphate buffer containing sucrose, prior to 1 hr of titer of cell-associated virus closely paralleled the postfixation in 1'c osmium tetroxide which was also virus in the culture both buffered with 0.1 M phosphate. After rapid dehydra- titration curve of fluid, tion in graded 70 to 100', ethyl alcohol solutions, the in slope and in maximal height. pellets were embedded in Luft's Epon formula (14). Uninfected cells. Chick embryo cells in primary Sections were cut by an automatic ultratome with culture displayed a considerable pleomorphism of diamond and glass knives. They were mounted on cytological appearance, but a majority resembled uncoated copper mesh grids, stained with 10'( uranyl fibroblasts or histiocytes. Most cells contained an acetate in methanol, and examined at original magnifi- abundance of dispersed ribosomes in the cyto- cations up to 18,000 in an RCA 3F electron micro- plasmic matrix. Aggregation of ribosomes into scope operated at 50 kv. crystalloid arrays was occasionally noted within Piulse labelinig. At indicated times after virus ilnfec- the cytoplasmic matrix or adjoining the limiting tion, the culture medium was decanted, and I ml of reticulum. Eagle's medium containing 100 Ac of :H-uridine (20 membranes of endoplasmic Cyto- c/mm) was added for 3 min. The monolayers were plasmic lipid droplets, phagocytic vacuoles, and then quickly washed five times with chilled 0.85', fenestrations were common. The latter ofter NaCl (containing uridine) and fixed immediately with measured more than 2 ,m in diameter. In cells glutaraldehyde. In control tests, monolayers were treated with actinomycin D, the nucleolar fixed with Carnoy's ethyl alcohol-ether fluid, and changes were characteristic (29), and cytoplasmic acid-precipitable radioactivity was counted with a vacuoles containing autophagic whorls of mem- Packard Tri-Carb scintillation spectrometer model brane remnants frequently appeared (9). Neither 3003 (Parkard Instrument Co., Downers Grove, 111.). control Autoradiograpliy. Thick epoxy sections (0.5 to 1.0 in treated nor untreated cells, however, ,um) for light microscopy were picked up oni 'subbed' were the arbovirus CPV ever identified. slides (2) and air-dried. They were coated with Kodak AR-10 stripping film (27) or Ilford L-4 Nuclear Research Emulsion Illford. Ltd., Ilford, Essex, Re/eosed Virus England (2)]. The procedures for electron-microscopy were based on the method of Caro et al. (2). Control -J grids without sections were covered with emulsion. exposed to room light, and examined in the electron Associoted Virus microscope with or witlhout development to check grain distribution. Other control grids with unlabeled sections or without any sections were coated, stored zZ - / - for 2 to 10 weeks, and developed to determine background grains. Thick (0.5 to 1 ;Am) sections were stained through the emulsion with 0.1'( Toluidine ID Blue in aqueous solution. Thin sections were ex- J - amined in the electron microscope without staining. 0 Tlhoroirat.st tracer. In four groups of experiments, a colloidal suspension of 24 to 26%'' Thorotrast was either added to cell pellets during fixation or intro- duced inito the culture medium (I to 2 cc/10 cc of medium) for 5 to 60 min precedinig fixation (no tracer added). For the tracer-fixation procedure, some cell pellets were prefixed in 3' glutaraldehyde (4) for 5 min before thie addition of 3' glutaraldehyde containing I to 2:10 parts ol the volume of the 24 to 26', Thorotrast. Other pellets were fixed directly in the I_l '0 2 3 4 5 6 7 8 9 10 Thorotrast-glutaraldehyde mixture. A similar volu- HOURS AFTER INFECTION WITH SFV metric proportion of the 24 to 26' Thorotrast was FIG. 1. Titers of SFV in click embrYo cells at maintained in the phosphate buffer-sucrose rinse and various times after inf'ction, wit/i a miultiplicity of 20 in the 1 (- osmium tetroxide postfixation. plaqule-forming unlits per cell. 1328 GRIMLEY, BEREZESKY, AND FRIEDMAN J. VIROL.

In.fected cells. In six different series of experi- et. al. (10) occurred in the latest stages examined ments, the chick cells were examined after infec- (16 hr). tion with 20 to 50 PFU of SFV per cell. The The cycle of virus development was asyn- earliest ultrastructural changes attributable to chronous at all multiplicities of infection studied; SFV were the occurrence of surface budding virus nevertheless, certain trends in the development particles and CPV-I at 2.5 to 3 hr. By 3 to 4 hr of cytopathic vacuoles were evident (Fig. 11). after the initiation of infection, cross sections The CPV-I were most prominent during the through 10 to 20% of cells observed in random early phase of infection. After 4 hr, the relative planes disclosed CPV-I (Fig. 2-6). Virus nucleoids proportion of CPV-II increased rapidly, and the budding through the plasmalemma to form proportion of CPV-I observed in any plane of mature virions were often present in the same section of a particular cell appeared inversely cells as the CPV-I (Fig. 3) or in planes of section related to the numbers of CPV-II. Some images lacking the vacuoles. Empty membrane projec- suggested a degeneration of the membrane tions were often observed adjacent to or in spherules lining CPV-I at about the same time as regions of virus budding (Fig. 2, 3). Cells con- CPV-II appeared (Fig. 7). The CPV-III were not taining CPV-II (Fig. 7) were noted after 4 hr of observed before 8 hr of infection and were never SFV infection. In the late stages of infection (8 to numerous. 16 hr), a few "mixed" cytopathic vacuoles (CPV- Intracytoplasmic aggregates of ribosomes were III) combined morphological features both of occasionally noted during the course of SFV in- CPV-I and of CPV-II (Fig. 8). fection (Fig. 8). In some instances, these images Typical CPV-I were distinguished by their probably represented tangential views along the relatively large diameter (0.6 to 2 pm) and the surface of endoplasmic reticulum. A peculiar regularly spaced membranous spherules that phenomenon in the terminal stages of infection projected from their interior surfaces (Fig. 2-6). was a proliferation of smooth endoplasmic re- The CPV-I appeared to originate from compact ticulum in the host cell. This effect has also been clusters of 50-nm spherules which were observed noted in the late stages of some DNA virus in- in the cytoplasmic matrix (Fig. 4-6). Rarefaction fections (P. M. Grimley, unpublished data). of cytoplasm within the clusters apparently Autoradiography. In trial experiments, control produced a central vacuolar space. Mature CPV-I cells incubated with 3H-thymidine or pulse- were clearly defined by a limiting membrane labeled with 3H-uridine displayed an appropriate (Fig. 4), and the characteristic inner spherules and discrete localization of developed silver grains often displayed a dense central spot (Fig. 5). The over the nucleus (Fig. 12) within periods of 2 to inner membranous projections of CPV-I were 10 weeks of emulsion exposure. Control cells not always uniform and circular in profile. In treated with actinomycin D and labeled with some sections, they appeared digital or droplike 3H-uridine in an identical fashion displayed no and often resembled the empty membrane pro- evidence of localized grain development after 10 jections of similar size which appeared in aggre- weeks of exposure. Background grains were gates on the plasmalemma of infected cells (Fig. negligible (<1 per oil immersion field). Direct 2). Membrane degeneration within CPV-I was a recording of acid-precipitable counts in the common feature, particularly as infection pro- control, actinomycin D-treated, and actinomycin gressed (Fig. 7). An intimate relationship of D-treated and virus-infected (20 PFU, 4 hr) CPV-I to the Golgi zone or rough endoplasmic cultures disclosed: (i) 639 counts per min per g reticulum was often noted (Fig. 5, 6). of protein, (ii) 27 counts per min per ,g of protein, The CPV-II appeared to arise from dilated and (iii) 142 counts per min per ,g of protein. regions of the endoplasmic reticulum (Fig. 9). Specific activity in the virus-infected cells at 4 hr As in other group A arbovirus infections, the was thus approximately five times that in cells CPV-II were surrounded by dense particles be- treated with actinomycin D only. lieved to represent viral nucleoids (23, 25). In In cells infected with 20 PFU of SFV for 6 hr planes tangential to the surface of CPV-II, the and pulse-labeled for 3 min with 3H-uridine, 30-nm nucleoids were occasionally arranged in localized cytoplasmic grains were detected by compact crystalloid arrays without intrinsic light microscopy within 2 to 5 weeks of exposure. symmetry. CPV-II were generally smaller in Approximately 30% of the cell sections were very diameter than the CPV-I, averaging 0.35 ,im. As heavily labeled after 5 weeks, and grains were the infection progressed, many CPV-II contained generally concentrated near the cell periphery. mature, membrane-enveloped virions within their Grains were rarely detected over nuclei or outside lumens (Fig. 10), and complex intravacuolar of cell sheets. Few grains appeared over the herniations of the types described by Erlandson identically processed, actinomycin D-treated ~~ ~ ~ ~ Pa .4 W ,4~~~~~~~~~~~~~~~~~~~~~~~~~~~~44 '4N$,@*

* * s * ;;4 F SW * t}XEaXif

,eS~ .r4r % r_.p

k~~~~~~~~~~~~~~~~~~~~ 3j43~ c :**f .ti7 ; . * C.

-IV~~~44~~~~

4 -&~~~~~~~~~~~~~~~r~~~~2 4urt.,Y; V3h,rj....

*~~~~~~~~~~~~~~*

'*1k *

3 v r v.i.; *. v. 3! - z . t §*.

FIG. 2. CPVt'>-'~~~~~~l i c a - '''sK'2or 2"S''' je -i" W

"I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~L

unfrmycicla.PrjctcsFI.2reflsoN0tmmmrnP- nacl netdwtnIhae oeF lngto . rndrsehe reglr rtrson rjcinf h ajcetplsaihnI r lemma(p).)<24,600.~ ~ ~ FIG.3.CVIincoepoiiytpasaem.Mtrviupatce(Varhudnincoepoii' oftohlloplsmaemhaneprotusins1iXp). 24600 FIG.4.roupof PV-I incell infcted ]or5 hr. Custers o cytoplsmict4 sphrulesV, (arows) apear to epresen anealysaglemma0(1p) . X$2400itevauoebmaio.Flccletmaeralwthn atr CV- m'rereet arfidrenat ofherignalcytplsmi marix ><37,00

1329 1330 GRIMLEY, BEREZESKY, AND FRIEDMAN J. VIROL.

G& I S&,

'Qt,¶1tt ;zh+t.431[ Ei? ¢) . * 9.

N~~~~~~~~~~~~~~~~~~~~~~~~~"

~~AL 9,.' ..~ ~~ I' X tSpj T ,k,*~ ~ ~ ~ , ."V

*44 OC.- '4 4

.4~~~~~~~~~~~~~~~~-..4~ FICSCP-IillVi-ilit' oX olg apalltls G) Mcll o th r?ew?bcile pil'/ll/S'Clp4C}-O lbl. tC 'le .-S~~~~~~~2 e~~~~~~~~~~~~l!EolicecllsmTll;Zcilet}(1.X2S4

FIG7.Cl .vinfccelfrSli Ieeeant.aul i)pesmbyceKdfoF PVI(i.I) PVI (icosdrabysnie,ndaetr(l.> 640 Fio.8.ixedvauole(C' V-Ill. Innermembran projectolis reenible PV-Ib (ig. II) but suroundoi desepatilersebl ncloilsofCI VIl Ntecrstdlidagreatonofriosme o a/facntenopasn reiulo(E) X2,40 conro scton eamne 5to14weksafer man istncs,a igifcat umerofth cto expsue. n ecton exmied y lecro plsmc gais wreconenratd verCP-l mirscp, icrt gan eelpen oudbe ote n etin i hihfe r ogris e detectedoverinfectedcells within 4 weeks of veloped over other portions of the cell (Fig. 13). exoue.BcgrudgrisattibcueiAteustietm er iceteCP- er nelgbe Insm elcosscin,tect- eulincae etoseaie yeeto plsmclaein ppaedrndmmcrsop,uniJatrhwve,in eelpen f h

Hxo.ur5. BcPV-ghioiundy ofin atolgisaprtuse (e). Manyc/le mhem/PVIan spieulscapereee tothae destacned fromgileI weastoseIn inel tires regularly aranedcmnd- intact 26,200.yeecro-cote setin braneiwolawithingI aperrow iandictemerlfhowvrm of CPVi.rsoy> 26,400.lpmntofth VOL. 2, 1968 ARBOVIRUS INFECTION 1331

*- S,^ -n . .* t

]I 'e* 44 4 's-a~~ ~ ~ ~ ~~ z Sri~~~~~~i

4~~~~~ :~~e P.Lr.LLJ''4Wsq..

tir' 4 1'~ 1 4 -~~~~~~~~.1Natk ~~~

'' * Hw .4~~~~~~~~~~~~~~~~~~~~~~~~~~~~Ajv2):,ira'If4 A '' A-~~~~~~ - e~~~~~~~~~~~Af ,s.-.@* P

z~~~~~~~~~~~t-~~~~~~~~~~-

packing ofI/.-reid(arow i47veAAlCPV-II 540 *

RN4~~~~~

FIG. 90. Typicl C-lctincellimfatuedo4iuharicNoes closeielationshipustersome vacuocles (a,Jib)ltsfcisteinaer) mlay hlave been released from a similar vacuole inl process Of "exocytosis." X 52,800. 1 332 GRIMLEY, BEREZESKY, AND FRIEDMAN j.i V IROL

FIG. i 1. Scliematic liaigi-CllRl based upon electron micrographs aind sluggestinig possible tranisitionts almolng C Pv. CPV-la with typical membrane splherulies appear to atrise from spherular clusters in cytoplasmic matrix. CPV-lb vith elonigaite membrane projectiolis may represenit a seconidary alteractioni. A relationslhip of this striuecture to suirfiace protrusions1s fom tile plasmalemma is suiggested by morphology cand Tlhorotrast tracet experiments. Degenieraition to CPV-Ic could be initiaited by hydrolytic enzymes fiom the Golgi regiont. Nucleoids su-rrount1dinig CPV-IIa may mature by buiddinig thr-olugh plasmalenmma or limiting membrantie of vacuole (CPV-JIb). Exit of matuire virionis in transport vacutoles (CPV-III) is sluggestedl by Thorotrast tracer experiments anid by the relaitively high proportion of cell-associated vi,iuis (Fig. /). negative plates, subjective bias in field selection icles containing mature virions also disclosed a on the fluorescent screen was largely prevented. few tracer particles. In a majority of the cell sections labeled by When the Thorotrast was added to the tissue tritium, the largest proportion of grains were culture media before cell fixation, results paral- concentrated near the cell surface (Fig. 13). This leled those described above during the first 15 to phenomenon was not observed in identically 45 min of incubation. By 1 hr, however, rare washed, actinomycin D-treated or untreated con- CPV-Ib (Fig. 11) contained electron-dense trols incubated in the same concentrations of particles. Tracer particles were never detected in 3H-uridine. CPV-Il. Significantly, vacuoles located just be- Tlhorotrcast experiments. The electron-dense neath the cell surface and containing mature tracer particles of Thorotrast were present in virions (CPV-IIc) were usually free of tracer sufficient concentration to outline the surface of particles even in cells heavily outlined by the more than 25(" of cell cross sections examined. Thorotrast. In pellets of control chick embryo cells, fixed High multiplicity infrctions. With purified and prior to introduction of Thorotrast or fixed in concentrated virus preparations (15), mature a Thorotrast mixture, electron-dense tracer virions with cell detritus from the heavy innocu- particles were frequently identified within large lum were identified at cell surfaces for up to 1 hr. cytoplasmic fenestrae or surface-connected re- Under these conditions, many particles entered cesses. Similar findings were the rule after 6 hr of the host cells by phagocytosis, sometimes in the SFV infection. Tracer particles were never identi- vicinity of surface "pits" (Fig. 15). Phagocytic fied in CPV, even though closely proximate vacuoles containing complete virions or empty fenestrae or digestive vacuoles contained them virus envelopes persisted within many cells for as (Fig. 14). Occasionally, subplasmalemmal ves- late as 2 hr after inoculation (Fig. 16). Actual VOL. 2, 1968 ARBOVIRUS INFECTION 1333 penetration of individual virus particles, from after high titer inoculation, CPV were not yet phagocytic vacuoles into the cytoplasmic com- present; however, in some cells, condensation of partment, could not be detected. In representa- ribosomes along the rough endoplasmic reticulum tive culture samples examined at 1 and 2 hr was unusually prominent (Fig. 17). As in the

*. t ; ~ -

>#T g~~~~~ p~~~4

*~~~~~~~ r t. *44

-..

- , C t *l W'TK St ~ it'l a.F fn~~~~~~~~~~ , r *4jT,RP~~~' * siM I t

12 "*. - 4,13A,

-~~~~~~~~~~~~~~~~~~~~~~~e

1'jI~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~aIll

3.. \r*t444.,

16~~ ~

FIG. 12. Autoradiogrcaph of cell from chick embryo cuiltuire pulse lIbeled with 3H-uridine. Graini developmenlt localized over ntucleuis (N). X 11,000. with FIG. 13. Autoradiographs of actiniomycini D-treated cells inJPcted jr 6 hIr with SFV anid pulse labeled 3H-uridine. A. Graiins conicenitrated niear plasmalemma. X 13,500. B. Grcinls conlcentracted in regioth of CPV-I. X 25,870. 1334 GRIMLEY, BEREZESKY, AND FRIEDMAN J. VIROL..

4/~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~4 U

11 1~~~~~~~~~~.~ ~

41, t.w ... Mf QM~/TS, e 11$6 6" Vri. v-spc-; i 8

FIG. 14. Section2 of cell from iufrcted culture, incubated with colloidal thorium dioxide for 1 hr. Electroni-dentse tracer hcas el7tered a small subplasmalemmal vacuole (arrow), but it is ntot observed withini CPV-I. X 32,000. FIG. 15. Clusters ofmaturle SFV virionls at surface ofcell 30 min after high multiplicity inifectiont. Note proximity to coated "pits" (arrows) andfilopodial extenisioni of cytoplasm (F). X 39,400. FIG. 16. Clutsters ofmature SFV virionis withini ctyoplasmic vacuoles at 30 miii after highl multiplicity inifectiont. X 39,600. VOL. 2, 1968 ARBOVIRUS INFECTION 1 335

A .1

is *- r

j1 -b

*%- i-1

FIG. 17. Crystalloid condensation of ribosomes int cell itifected for 2 hr (high multiplicity). Note proximity of endoplasmic reticulum (ER) to plasma membrane. X 42,200. FIG. 18. Cell infected with high multiplicity ofSFV. CPV-II are well developed at 3 hr after infection, anid some containi mature virions (lIb, Ifc). These contrast with degenerating membranous spherules (Ic) in CPV-I. X 42,600. 1 336 GRIMLEY, BEREZESKY, AND FRIEDMAN J. VIROL. low multiplicity infections, well-developed CPV- We initially considered the possibility that I appeared at about 3 hr (Fig. 18). These occurred CPV-I might represent cytoplasmic depots related only in a small fraction of cell cross sections. to virus entry. A remarkably similar vacuole oc- CPV-I and CPV-II were often present simul- curs within 15 min of high multiplicity polyoma taneously (Fig. 18). Some CPV-1I contained infection (11), and structures reminiscent of ar- mature virions, and surface budding virus was bovirus CPV-II may arise by "pinocytotic activ- also evident at this time. Digital membrane ity" (22) during the multiplication of mammary protrusions with hollow cores were occasionally tumor agent. In SFV infection, CPV-I were never observed in rows along the surface of infected observed less than 2 hr after high- or low-titer in- cells. fections; phagocytic vacuoles containing clumps of degenerating viruses could easily be distin- DISCUSSION guished during this period. The failure of the electron-dense tracer to penetrate CPV-Ia in Membrane association of viral RNA. Recent several experiments helps to exclude an origin studies of chick embryo cells infected with SFV from invaginating surface membranes. The have suggested that replication of both viral presence of Thorotrast in occasional CPV-Ib or RNA and protein occurs in association with CPV-Ic after 45 min of incubation indicates that cytomembranes (12-14). The replicative inter- the latter structures might communicate with mediate (12) and the RNA polymerase (20) of surface membrane, perhaps in the zones display- SFV have been recovered with membrane frac- ing the unusual configuration of multiple vesicu- tions of the host cell cytoplasm. Up to 75% of the lar projections (18). It is also conceivable that the viral RNA produced early in infection may be Thorotrast observed in CPV-I after 45 min reflects released from the cytomembrane fraction by secondary fusion of CPV-I with lysosomal bodies simple treatment with deoxycholate (1, 2). CPV-I (17). This could explain the CPV-I degeneration are present in cytoplasmic fractions containing and their proportionate decline in numbers with the early forms of viral RNA (14), and auto respect to CPV-II. A close association of CPV-I radiographic and electron microscopic observa- with the Golgi region has occasionally been tions now suggest that they may serve as im- noted (Fig. 5). portant sites for SFV RNA synthesis. The dis- Formation of CPV-II and release of SFV. tribution of silver grains suggests that the CPV-I Several investigators have suggested previously limiting membrane, rather than the vacuole that the 25- to 30-nm "nucleoids" surrounding interior, is the actual site of specific 3H-uridine arbovirus CPV-II are virus precursors (10, 23-25). uptake (6). Heavy grain localization at the cell These nucleoids evidently correspond to the 140S surfaces demonstrates that the plasmalemma may cell fraction isolated by Friedman (12, 14). also provide a site for viral RNA replication. The Precursor particles investing the CPV-II relative contributions of cell surface and intra- evidently acquire an external membrane coat in cytoplasmic loci of SFV-RNA synthesis will be transit to the vacuole lumen (10, 23). The rela- difficult to assess until more refined differential tively higher titer of cell-associated virus obtained separation of cytomembrane sediments is feasible. in our experiment and a close parallel to the rising Concentration of new viral RNA in discrete curve of released virus lend strong support to the cytoplasmic loci or "factories" finds precedent in morphological evidence of mature virus forma- both RNA and deoxyribonucleic acid virus tion in CPV-II. Since the findings of Cheng (5) infections (7, 19, 28). indicate an exceptionally high ratio of infectious Significance of CPV in arbovirus development. to visible virions in SFV infection, we may reason- Our observations of CPV-I as early as 2 to 3 hr of the virions after high- or low-titer virus inoculations and ably expect that most intravacuolar CPV-II within 3 to 4 hr refute an earlier conten- observed during the log growth phase are indeed tion (1) that CPV are unrelated to the phase of biologically active. The absence of electron- effective virus multiplication. We also observed dense tracer within many CPV-II (Fig. 11) after CPV-I in the early logarithmic phase of Sindbis 45 min of incubation in Thorotrast medium virus replication (unpublished data). Careful indicates that mature virions observed in these reexamination of other cell systems during the vacuoles may be leaving rather than entering the log phase of arbovirus growth may disclose close infected cells. morphological as well as biochemical parallels Correlation of morphological and biochemical (30) to the events described in SFV infection. The observations. Several forms of viral RNA are "globoid bodies" described in nerve cells infected produced during SFV infection (12, 14, 15). A by Japanese B encephalitis bear some resemblance partially ribonuclease-resistant, polydisperse form to the CPV-I of SFV (34). containing a 16S "core" apparently is the replica- VOL. 2, 1968 ARBOVIRUS INFECTION 1337 tive intermediate (12). It is membrane-associated LITERATURE CITED and more rapidly labeled than the other RNA 1. Acheson, N. H., and I. Tamm. 1967. Replication forms. Infectious RNA sediments at 42S and is of Semliki Forest Virus: an electron micro- present in a 140S particle that also contains viral scopic study. Virology 32:128/43. protein (13). Autoradiographic labeling with 2. Caro, L. G., R. P. van Tubergen, and J. A. Kolb. 3H-uridine now localizes some of the early viral 1962. High resolution autoradiography. l. RNA to the plasma membrane region and CPV-I. Methods. J. Cell Biol. 15:173-188. The bulk of protein synthesis in SFV-infected 3. Chain, M. T., F. W. Doane, and D. M. McLean. 1966. Morphological development of Chikun- chick cells is also membrane-associated (13). A gunya virus. Can. J. Microbiol. 12:895-900. spatial proximity of virus assembly sites to 4. Chambers, R. W., M. C. Bowling, and P. M. membrane-bound host cell ribosomes thus might Grimley. 1968. Glutaraldehyde fixation in be anticipated. Our electron micrographs are routine histopathology. Arch. Pathol. 85:18-30. consistent with some earlier impressions that 5. Cheng, P. Y. 1961. Purification, size and mor- CPV-II derive from the endoplasmic reticulum of phology of a mosquito-borne animal virus, infected cells (3, 25). The frequent aggregation of Semliki Forest virus. Virology 14:124-131. ribosomes observed in infected cells, particularly 6. Comings, D. E., and T. Kakefuda. 1968. Initia- infections tion of deoxyribonucleic acid replication at the high multiplicity (Fig. 8, 17), may nuclear membrane in human cells. J. Mol. Biol. reflect virus-directed protein synthesis, since 95 c 33:225. of host-cell protein synthesis is inhibited by the 7. Dales, S. 1963. The uptake and development of actinomycin D pretreatment (13). vaccinia virus in strain L cells followed with Migration of early viral RNA from CPV-I to labeled viral deoxyribonucleic acid. J. Cell CPV-II could plausibly occur prior to incorpora- Biol. 18:51-72. tion of the 42S infectious RNA form into the 8. Dales, S., H. J. Eggers, I. Tamm, and G. E. 140S "nucleoids" which provide the first mor- Palade. 1965. Electron microscopic study of the phological evidence of virus core formation. formation of poliovirus. Virology 26:379-389. 9. Eisenstein, R., H. Battifora, and E. Axelrod. This postulated migration of early viral RNA 1966. Vitamin D-actinomycin D interactions in would be compatible with the temporal sequence bone: an electron microscopic study of hyper- of ultrastructural events; CPV-I precede CPV-II trophy and hyperinvolution of osteoblasts. in appearance and decline in relative numbers as Lab. Invest. 15:1589-1600. infection progresses. Some micrographs (Fig. 6) 10. Erlandson, R. A., V. I. Babcock, C. M. Southam, suggest a close proximity of CPV-I to the rough R. B. Bailey, and F. H. Shipkey. 1967. Semliki endoplasmic reticulum; considering the mobility Forest virus in Hep-2 cell cultures. J. Virol. of the latter, frequent contacts must occur. Con- 1:996-1009. tact of the reticulum with the 11. Fraser, K. B., and E. M. Crawford. 1965. Im endoplasmic munofluorescent and electron microscopic plasmalemma may occur with significant fre- studies of polyoma virus in transformation quency (Fig. 17), and has also been observed in reactions with BHK 21 cells. Expt. Mol. Pathol. herpes virus infections (V. Bedoya et. al., J. 4:51-65. Natl. Cancer Inst., in press). Such a phenomenon 12. Friedman, R. M. 1968. Replicative intermediate of may explain the frequent maturation of SFV an arbovirus. J. Virol. 2:547-552. virions near the cell surface. A spatial dissociation 13. Friedman, R. M. 1968. Protein synthesis directed of RNA replicative and viral assembly sites also by an arbovirus. J. Virol. 2:26-32. appears to occur in poliovirus infection (8). 14. Friedman, R. M., and I. K. Berezesky. 1967. the observations to a tentative Cytoplasmic fractions associated with Semliki Integrating date, Forest virus ribonucleic acid replication. J. sequence of events in the SFV reproductive cycle Virol. 1:374-383. may be postulated. (i) The RNA replicative 15. Friedman, R. M., H. B. Levy, and W. B. Carter. intermediate is formed on the plasma membrane 1966. Replication of Semliki Forest virus: three or at intracellular loci (CPV-I). (ii) Assembly of forms of viral RNA produced during infection. infectious RNA within an RNA-protein core or Proc. Natl. Acad. Sci. U.S. 56:440-446. nucleoid proceeds on the exterior of CPV-I1, 16. Friedman, R. M., and J. A. Sonnabend. 1965. derived from endoplasmic reticulum or at Inhibition of interferon action by puromycin. the plasma membrane. (iii) Mature virions form J. Immunol. 95:696-703. as traverse either 17. Gordon, G. B., L. R. Miller, and K. G. Bensch. from the 140S nucleoids they 1965. Studies on the intracellular digestive the plasma membrane or CPV-II limiting mem- process in mammalian tissue culture cells. J. brane and obtain an external envelope. Cell Biol. 25 (no. 2, part 2) :45-56. 18. Gosselin, R. E. 1967. Kinetics of pinocytosis. ACKNOWLEDGMENTS Federation Proc. 26:987-993. We are grateful to R. Bogoroch for valuable advice 19. Granboulan, N., and R. M. Franklin. 1966. High and to S. Mims for able technical assistance. resolution autoradiography of Escherichia coli 1338 GRIMLEY, BEREZESKY, AND FRIEDMAN J. VIROL.

cells infected with bacteriophage R17. J. growth of Venezuelan equine encephalomyelitis Bacteriol. 91:849-857. virus in KB cells. Virology 16:52-62. 20. Martin, E. M., and J. A. Sonnabend. 1967 26. Ota, Z. 1965. Electron microscope study of the Ribonucleic acid polymerase catalyzing syn- development of Japanese B encephalitis virus thesis of double-stranded arbovirus ribonucleic in porcine kidney stable (PS) cells. Virology acid. J. Virol. 1:97-109. 25:372-378. 21. McGee-Russell, S. M., and G. Gosztonyi. 1967. 27. Pelc, S. R. 1956. Design of stripping film. Intern. Assembly of Semliki Forest virus in brain. J. Appl. Radiation Isotopes 1:172-177. Nature 214:1204-1206. 28. Silverstein. S. C., and S. Dales. 1968. The penetra- 22. Moore, D. H. 1962. The milk agent, p. 113-150. tion of reovirus RNA and initiation of its In A. J. Dalton and F. Haguenau (ed.), Tumors genetic function in L-strain fibroblasts. J. Cell induced by viruses: ultrastructural studies. Biol. 36:197-230. Academic Press, Inc., New York. 29. Simard, R., and W. Bernhard. 1966. Le pheno- 23. Morgan, C., C. Howe, and H. M. Rose. 1961. mene de la segregation nucleolaire: specificite Structure and development of viruses as ob- d'action de certains antimetabolites. Intern. J. served in the electron microscope: V. Western Cancer 1:463-479. equine encephalomyelitis virus. J. Exptl. Med. 30. Wecker, E., and A. Richter. 1962. Conditions for 113:219-234. replication of infectious viral RNA. Cold 24. Mussgay, M. 1964. Growth cycle of arboviruses in Spring Harbor Symp. Quant. Biol. 27:51-55. vertebrate and arthropod cells. Progr. Med. 31. Yasuzumi, G., and I. Tsubo. 1965. Analysis of the Virol. 6:193-267. development of Japanese B. encephalitis 25. Mussgay, M., and J. Weibel. 1962. Electron (JBE) virus It. EM studies of neurons infected microscopic and biological studies on the with JBE virus. J. Ultrastruct. Res. 12:304-316.