Cell,
Vol. 12, 243-249,
September
1977, Copyright
0 1977 by MIT
Viral DNA Sequences from Incomplete Human Adenovirus Type 7 Clark Tibbetts Department of Microbiology School of Medicine University of Connecticut Health Farmington, Connecticut 06032
Center
Summary Large pools of empty viral capsids accumulate in cells infected by subgroup B human adenoviruses. Such infected cells also yield DNA-containing incomplete particles in larger quantities than cells infected with serotypes representing other adenovirus subgroups. DNA isolated from carefully purified classes of Ad7 incomplete particles was analyzed by restriction endonuclease cleavage, gel electrophoresis and electron microscopy. At least 90% of the DNA molecules in each sample consisted of sequences that extended from the left end of the viral genome map by variable lengths toward the right end. The average length of DNA is linearly related to the average buoyant density of the incomplete particles from which the DNA is isolated. The results indicate that each capsid contains one DNA molecule. There is also a specific association of the left end of the viral genome with assembled or assembling capsids. The characteristic distributions of Ad7 incomplete particles may result from intracellular pools of assembly intermediates in which the incompletely packaged DNA has been fragmented in vivo or by shear during preparative procedures. Introduction Buoyant CsCl density gradient centrifugation has been used in procedures for the purification and characterization of many different viruses. Early applications in the preparation of adenoviruses led to observations of viral particles with buoyant densities significantly lower than those of infectious virions. The lightest of these particles were identified by electron microscopy and polypeptide analysis as empty adenovirus capsids (Kohler, 1962; Smith, 1965; Shimojo et al., 1967; Maizel, White and Scharff, 1968). These empty capsids have been implicated as intermediates in virion assembly, formed prior to the interaction with newly synthesized viral DNA in the nuclei of infected cells (Sundquist et al., 1973). Adenovirus particles have also been observed at positions in CsCl density gradients between the virions and empty capsids. These incomplete particles also have an adenovirus morphology and contain, in general, viral DNA sequences shorter than full-length virion DNA (Prage, Hoglund and Phiiipson, 1972; Wadell,
Particles
of
Hammerskjold and Varsanyi, 1973; lshibashi and Maizel, 1974; Burlingham, and Brown Doerfler, 1974; Rosenwirth et al., 1974; Daniell, 1976). Serotypes representing different adenovirus subgroups appear to give rise to characteristic distributions of virions, incomplete particles and empty capsids in preparative CsCl density gradients. Preparations from cells infected by subgroup 6 human adenoviruses (Ad3, Ad7, Ad16) generally have the largest quantities of incomplete particles and empty capsids (Prage et al., 1972; Wadell et al., 1973; Daniell, 1976; Tibbetts, unpublished observations). A variety of reports have dealt with the biological properties of incomplete adenovirus particles. The infectivity associated with such particles is orders of magnitude less than that of virions (Prage et al., 1972; Wadel.1 et al., 1973; Burlingham et al., 1974; Niiyama et al., 1975). The infectivity is most probably due to contamination of preparations with trace quantities of infectious virions. Incomplete particles of human Ad3 and Ad7 do not appear to interfere with infection by virions nor to alter the eventual yield or distribution of virions, incomplete particles and empty capsids (Daniell, 1976; C. Tibbetts, unpublished observations). incomplete particles of human Ad12 and bovine Ad3, which contain nearly full-length viral DNA, seem to be prevalent only after passaging the virus by high multiplicity infection (Mak, 1971; Niiyama et al., 1975; lgarishi et al., 1975). There have been reports of adenovirus incomplete particles having potential for malignant transformation of cells in culture (Schaller and Yohn, 1974; lgarishi et al., 1975). This seems to be quite feasible, since only a limited region of the adenovirus genome, in particular the left-end and 7-l 4%, appears to be necessary for the transformation of cells (Graham, van der Eb and Heijneker, 1974; Gallimore, Sharp and Sambrook, 1974). Some incomplete particle DNA may represent the transforming DNA sequences, yet lack other viral genes that have cytotoxic effects compromising transformation assays. Studies of viral polypeptides specifically associated with virions or incomplete and empty particles have led investigators to suggest relationships of these particles to the in vivo sequence of steps in virion assembly (Sundquist et al., 1972; lshibashi and Maizel, 1974; Edvardsson et al., 1976). Parallel studies of the DNA associated with incomplete adenovirus particles may lead to further understanding of their origin as well as their various biological properties. This study describes incomplete particles of human Ad7 found in CsCl density gradients throughout the region between virions and empty capsids. Viral DNA was extracted from nine purified buoyant density classes of incomplete particles and characterized by gel electrophoresis, restric-
Cell 244
tion endonuclease cleavage and electron microscopy. The results are consistent with a model based on specific association of the left end of one Ad7 genome with assembled capsids during the synthesis of mature virions. The incomplete particles observed in density gradients may be the result of fragmentation of incompletely packaged DNA in vivo or by shear during preparative procedures.
Results Incomplete
Particles of Ad7
HeLa cells were infected with Ad7 virus and labeled with “C-formate (Experimental Procedures) to provide radioactivity in viral DNA and protein. Sonicates of infected cells were cleared of debris by low speed centrifugation (1000 g) and banded to equilibrium in self-generating CsCl density gradients (Figure 1). Opalescent bands corresponding to virions, empty capsids and the predominant species of incomplete particles [as described for Ad3 by Prage et al. (1972)J were readily observed. In addition to these three major bands, there was a more diffuse
c1 I
10
Virions
+ ...... .. ... .. .. .. . ..
._
..
\
.
E
op3hl-
FRACTIONS Figure 1. Equilibrium beled Ad7 Particles
CsCl
from Density
CsCl Gradients
GRADIENTS of %-Formate-La-
Approximately lo9 HeLa cells (2 liters) were infected with Ad7 virus and labeled with 3 mCi of “C-formate. The cells were harvested (64 hr post-infection), and virus was prepared as described in Experimental Procedures. The distribution of radioactivity associated with unfractionated particles and residual soluble protein is shown in the upper panel. Material in the indicated interval “upper pool” was combined and centrifuged to equilibrium with the resultant distribution shown in the lower panel. Eleven fractions (IP-a to IP-k) were pooled as indicated, recentrifuged to equilibrium and subjected to the viral DNA extraction procedure.
distribution of similarly opalescent material in the region between the virions and major incomplete particles. In the intermediate region, there were also bands of presumed adenovirus particles which appeared in characteristic positions and relative amounts. The overall appearance was quite similar to the density gradient distributions seen with other subgroup B human adenoviruses [Ad3, (Daniell, 1976); Ad1 6 (Wade11 et al., 1973)]. Material appearing above the virions was pooled as shown in the upper panel of Figure 1 and rebanded to equilibrium in CsCl density gradients. This gradient was fractionated, as indicated in the lower panel of Figure 1, to provide eleven pools corresponding to density classes throughout the range from virions to empty capsids. These pools were each rebanded to equilibrium in CsCl density gradients (not shown). Each of the gradients showed a predominant band at a position corresponding to the density of the region from which the particles had been originally pooled. Pools from IP-b and IP-g also showed minute amounts of material in a second band at the position corresponding to buoyant virions. The major bands from each gradient were recovered by fractionation and subjected to the viral DNA extraction procedure (Experimental Procedures). Each sample, except pool IP-k, yielded labeled DNA (about 2 x lo4 cpm pg) at positions near 1.7 g/ml in the buoyant density gradients. The analysis of these DNA samples presented below specifically relates to the samples pooled as in Figure 1. Several similar preparations of Ad7 incomplete particles have been examined with equivalent results.
Incomplete
Particle DNA
Samples (1 pg) of DNA from incomplete particle pools IP-b to IP-j were mixed with 0.1 pg Ad7 viral DNA (2.3 x 10’ daltons) and subjected to electrophoresis in 0.4% agarose gels (Figure 2). The DNA from each pool represents linear duplex DNA (see electron microscopy results in this section) and migrates as a rather unimodal distribution of sizes. The average lengths of DNA from the pools varied from 80% (IP-b) to 12% (IP-j) of full-length AD7 DNA, and in each case, the average length was linearly related to the density of the pooled particles from which the DNA had been prepared. This is similar to earlier observations of the DNA associated with more discrete bands of incomplete particles of Ad3 (Prage et al., 1972; Daniell, 1976) and of Ad2 and Ad12 (Burlingham et al., 1974). The result suggests that each incomplete particle of adenovirus includes only one less than full viral length DNA molecule. Restriction endonucleases have recently been used to provide physical maps of Ad7 virion DNA (Tibbetts, 1977). Viral DNA and DNA from pools
DNA from 245
Incomplete
Adenovirus
Particles
IP-a through IP-J were analyzed by digestions with endoR.Sma I (Figure 3), endoR.Hind III (Figure 4), endoR.Bam HI (not shown), and a mixture of endoR.Sma I and endoR.Hpa I (not shown). Cleavage of DNA from the Ad7 incomplete particle COpools led to restriction fragments which migrated with the fragments of Ad7 DNA generated with the same endonuclease. In general, >90% of the restricted DNA from the incomplete particles was found to co-migrate with one or more Ad7specific restriction fragments. Quantitative analysis of these results by scanning densitometry of photographic negatives of the ethidium-stained gels (as for the Hind III experiment in Figure 4) was based upon the assumption of sequence identity of co-migrating fragments from virion and incomplete particle DNA. This analysis revealed two groups of fragments in the restricted DNA from Ad7 incomplete particles. Some fragments were present in equimolar quantities (when related to equivalent quantities of viral DNA fragments and corrected for the smaller size of incomplete particle DNA molecules), while the remaining fragments apoeared in an order of decreasing, less than equimolar, amounts. These patterns of distributions of restriction fragments, for three of the enzymes used in this study, are sum-
RI
b
c
d
e
f
g
h
i
i
marized in Figure 5 by relating the equimolar set of fragments and the next most abundant fragment of each digest to the known maps of restriction sites in Ad7 DNA. It is readily apparent that the bulk of DNA associated with the incomplete particle pools of Ad7 represents the linear progression of viral DNA sequences from the left end of the genome map to a variable right terminus determined by the average size of DNA in the particles representing each particular buoyant density class. Electron microscopy of DNA preparations from pools IP-a to IP-j revealed exclusively linear duplex DNA. Contour length analysis of the DNA from each pool led to results such as those shown for pool IP-f in Figure 6. Each distribution was unimodal; the mean length of each distribution was sufficient to account for only single-copy representation of each of the observed equimolar restriction fragments of DNA from the respective pool. The mean lengths were linearly related to the density of the pool of particles from which the DNA had been isolated. Most of the viral DNA sequences associated with the Ad7 incomplete particles seemed to be free of inverted repetitions of the Ad3 type described by Daniel1 (1976). Samples of DNA representing Ad7 incomplete particle pools were denatured with alkali, neutralized and held at low DNA concentration (0.2-0.5 pg/ml) under reassociation conditions
a
:
bCdefehii7
B
Figure
2. Electrophoresis
of Ad7 Incomplete
Particle
DNA
Samples of DNA (about 1 fig) from incomplete particle pools Ip-b through IP-j were mixed with about 0.2 pg full-length Ad7 DNA and subjected to electrophoresis in 0.4% agarose gels (Experimental Procedures). The left gel shows two of the three Eco RI restriction fragements of Ad7 DNA (A = 35.4%; 6 = 13.0%) to indicate the approximate range of DNA lengths associated with the different incomplete particle pools. The DNA from the denser pools may appear to be less heterogeneous in size than that from the iighter pools, but this must be attributed to the lower resolution of the electrophoresis system with DNA larger than about 1 x 10’ daltons. The distributions of DNA are rather unimodal, but there is some indication of discrete lengths of DNA appearing as bands within the continuous distribution of most of the incomplete particle DNA.
Figure 3. Restriction endoR.Sma I
of Ad7 Viral and Incomplete
Particle
DNA by
Each DNA sample (about 1 pg) was mixed with about 0.05 pg of Ad7 viral DNA prior to incubation with the restriction enzyme to control for complete cleavage. Prior to electrophoresis. uncut Ad7 viral DNA (0.2 pg) was added to each sample as a position marker. At least 90% of the incomplete particle pool DNA samples could be recovered as fragments co migrating with known Sma I cleavage fragments of Ad7 DNA.
Cell 246
i
E
+
l +
I-:::. .O
.
.2 FRACTfON
Figure 5. Physical Particles
Maps
.
.
.6 .4 of Ad 7 GENOME
of DNA Associated
.6
.
’
1.0
LENGTH
with
Ad7 Incomplete
DNA from each pool of incomplete particles was analyzed as shown in Figure 4 with restriction endonucleases Sma I, Barn I and Hind ill (upper, middle and lower bar of each set). The solid lines represent those fragments which were recovered in singlecopy yield from each DNA sample. The gap between the right end of each bar and the + symbol represent the next most abundant cleavage fragment of the digest. The + symbols together indicate the descending region of the DNA length distributions where only longer molecules will yield the restriction fragments indicated by the gaps. These lengths are in agreement with contour length analysis by electron microscopy (Figure 6) and are linearly related to the buoyant density of the fractions pooled for the preparation of each DNA sample (Figure 1).
Figure 4. Relative Amounts of Hind III Cleavage Restricted Ad7 Viral and Incomplete Particle DNA
Fragments
in
Film densitometry of negatives permitted quantitative estimation of the amounts of each Hind Ill cleavage fragment associated with the DNA from incomplete particle pools. The residual DNA at positions corresponding to viral DNA cleavage fragments is about 5% of the original amount shown in the IP-a pool scan and represents the viral DNA added to each sample to control for complete cleavage. These were not present in other samples run with the incomplete particle DNA alone. Negatives had exposure times limiting the regions of the most intense absorbance to the range of linear film response to ethidium fluorescence.
[50% formamide, 0.10 M Tris (pH 84, 0.01 M EDTA, 37”C] until subsequent preparation for electron microscopy. The observed distributions of contour lengths of the linear single strands (>95% of the molecules in samples prepared within 5 min) were somewhat broader than seen with the duplex DNA before denaturation, but the mean lengths were the same (k 10% of Ad7 genome length) for the duplex and denatured DNA of each pool. Fewer than 5% of the molecules, held up to 2 hr before spreading with cytochrome c, were observed to form single-stranded circles. Some of these singlestranded structures had rather long duplex-appearing panhandles, reminiscent of the stem-loop structures described by Daniel1 (1976) for Ad3 incomplete particle DNA. Renatured molecules had the generally simple appearance of linear duplex
DNA from 247
Incomplete
Adenovirus
Particles
t
.O
4
.2 FRACTION
Figure 6. Contour cle Pool IP-f
Length
.4 .6 of Ad 7 GENOME Analysis
.8
1.0
LENGTH
of DNA from
Incomplete
Parti-
DNA from pool IP-f was mixed with form II plasmid P,,, DNA (length marker) for contour length analysis by electron microscopy (Experimental Procedures). The histogram shows the results of 175 length measurements plotted as 2% intervals of Ad7 genome length. The IP-f pool represents DNA from particles having a buoyant position nearly midway betwen virions and empty capsids in CsCl density gradients (Figure 1). Notice that the bulk of the unimodal distribution represents molecules slightly longer than the left-end sequences corresponding to restriction fragments observed in single-copy amounts, but shorter than the restriction sites corresponding to fragments obtained in less than equimolar yield (Figure 5). Analysis of the other incomplete particle pools gave results similar to these obtainedwith IP-f.
DNA with a short single-stranded DNA tail of variable length at only one end of each molecule. Similar structures have been observed in DNA from incomplete particles of a murine adenovirus (D. Nathans, personal communication. Discussion The results presented in this report suggest that Ad7 capsids can interact with only one viral DNA molecule, and that this interaction is in some manner specifically associated with the molecular left end of the conventional viral genome map. Most of the DNA associated with Ad7 incomplete particles are simple linear duplex sequences from the left end of the Ad7 genome to a variable right terminus. This left end specificity is intriguing, since the molecular ends of adenovirus DNA are identical with regard to their nucleotide sequences (regions of inverted terminal repetition) and the presence of the covalently bound 5’ end proteins on each complementary strand of viral DNA (Garon, Berry and Rose, 1972; Wolfson and Dressler, 1972; Roberts, Arrand and Keller, 1974; R. Roberts, personal communication). Studies of adenovirus polypeptides in virions, incomplete particles and empty capsids have led
other investigators to suggest a precursor-intermediate-product relationship of these particles through virion assembly in vivo (Sundquist et al., 1973; lshibashi and Maizel, 1974; Edvardsson et al., 1976). The framework of the hypothesis is supported by radioactive labeling results of pulsechase experiments and the clear evidence for cleavage processing of certain precursor polypeptides, present in empty capsids and incomplete particles, to their respective mature virion-associated products. This background leads to an appealing interpretation of the properties of viral DNA sequences associated with incomplete particles of Ad7 (Figure 7). Empty capsids are proposed as preassembled intermediates, destined to interact with one viral DNA molecule and its associated proteins (perhaps a precursor to the virion core). The molecular left end of the viral DNA is specifically inserted, leaving about 85% of the viral DNA extending from the capsid. The following steps in the packaging of the viral genome would require substantial compression of the volume occupied by the DNA and possible displacement or cleavage processing of polypeptides associated with the previously empty capsids. Final steps in virion assembly may involve restoration or completion of external capsomeres in the region where DNA enters the capsid. The incomplete particles which are observed in preparations of adenovirus from infected cells may be derived from intermediates of the assembly process proposed above. Incompletely packaged DNA may be fragmented in vivo during abortive assembly. It would certainly be fragmented in vitro by shear forces associated with preparative procedures. The proposed model suggests that the initial interaction of empty capsids with the DNA (or precore), and the subsequent packaging steps are rate-limiting in virion assembly, permitting accumulation of intermediate structures predisposed to yield abundant quantities of empty capsids or incomplete particles. The types of particles which are most abundant in extracts of ceils infected by subgroup B adenoviruses are empty particles, an incomplete species with about IO-15% of the viral genome (pool IP-j, Figure 1) and mature virions. An alternative model might suggest that the incomplete particles represent breakdown products of labile assembly intermediates. This could imply a preferential extrusion and fragmentation of the molecular right portion of the viral genome from more nearly complete assembly intermediates. It was suggested by Edvardsson et al. (1976) that cetain assembly intermediates are not stable in CsCl density gradients. When such particles were isolated from their preparative Ficoll gradients, fixed with gluteraldehyde and banded in GsCl density gra-
Cell 240
IN PACKAGING Ad7 VIRAL DNA
MATURE VIRIONS
Figure
7. Packaging
of Adenovirus
II
1.340 V
DNA
The results of this study suggest that viral DNA interacts with empty capsids by specific association (entry?) of one end of the viral genome (the left end of conventional physical maps). The subsequent packaging of the remaining DNA probably reflects a concerted process of altering DNA conformation within the particle and displacement or cleavage of internal polypeptides of the capsids. Subgroup B human adenoviruses yield large quantities of incomplete particles in density gradients which may reflect DNA fragmentation, in vivo or during preparative procedures, in intermediates which accumulate in cells during the DNA packaging process. Subtle differences in virus-coded polypeptides specified by serotypes representing other subgroups could account for different rate-limiting steps in the course of virus assembly. This would then lead to accumulation of intracellular intermediates and thus different types and relative amounts of incomplete particles in preparative density gradients.
dients, however, the predominate breakdown products appeared to be empty capsids and DNAcontaining cores. There are several types of DNA structures which have been described in preparations of incomplete particles from different adenovirus serotypes and subgroups. Preparations of human Ad12 (Mak, 1971; Garon et al., 1972; C. Tibbetts, unpublished observations) and bovine Ad3 (Niiyama et al., 1975) suggest that dense species of incomplete particles in these systems represent defective but replicating DNA of less than full length, with heterologous sequences located specifically near one end of the viral genome maps. DNA structures observed in preparations of human Ad3 (Daniell, 1976) and human Ad1 8 (Garon, Berry and Rose, 1975) appear to contain long inverted terminal repetitions of sequences from one end of the viral DNA. In these cases, electron microscopy of denatured DNA molecules reveals panhandle structures much longer than those resulting from association of the inverted terminal repetition sequences of denatured virion DNA (Wolfson and Dressler, 1972; Garon et al., 1972). None of these cases precludes, however,
the type of model which has been proposed for adenovirus assembly. The results presented for Ad3 incomplete particle DNA by Daniel1 (1976) are in some respects qualitatively similar to those presented for Ad7 in this report. This should be expected, since Ad3 and Ad7 are very closely related subgroup B human adenoviruses. The DNA structures emphasized by Daniel1 (linear duplex DNA with long inverted terminal repetitions) have also been observed in my own preparations of Ad3 and Ad7 incomplete particle DNA. I have not, however, found such structures to represent more than about 5% of the total incomplete particle DNA preparations. The abortive DNA replication scenario proposed by Sambrook (Daniell, 1976) may indeed relate to these novel DNA structures. It is the quantitative consideration of the viral DNA sequences associated with incomplete particles of Ad7 which has led me to emphasize their possible relationship to events in adenovirus assembly. Experimental
Procedures
Cells and Viruses Adenovirus type 7 (Gomen) and type 3 (G.B.) werepropagated in suspension cultures of HeLa S3 cells as described previously (Tibbetts, Johansson and Philipson, 1973). DNA Adenovirus DNA was extracted from virions or incomplete particles by incubation with sarkosyl and pronase 6, and purified by extraction with phenol, chloroform, isoamyl alcohol followed by centrifugation in isopycnic CsCl density gradients (Tibbetts et al., 1973). Labeling of Ad7-infected cell cultures with “C-formate was performed as previously described (Tibbetts et al., 1973). Restriction Conditions sites in Ad7 manuscript
Endonucleases for enzyme digests and DNA have been presented submitted).
physical maps of restriction elsewhere (Tibbetts, 1977).
Density Gradients Sonicates of virus-infected cells or partially purified preparations were adjusted to1.34 g/ml with CsCl and centrifuged in angle rotors (Ti50 or Ti75) at 30,000 rpm, 4°C for 36-48 hr. Gradients were fractionated by tube puncture at the bottom; 50 or 70 PI fractions (7 or 10 drops) were collected. Radioactivity in the gradient was monitored by drying 5 ~1 aliquots of each fraction on paper filters and counting in a liquid scintillation spectrometer with a toluene-based fluor (Tibbetts et al., 1973). Gel Electrophoresis DNA samples were subjected to electrophoresis in agarose gels as originally suggested by Sharp, Sugden and Sambrook (1972). Specific details of my applications were recently described (Tibbetts, 1977). Electron Microscopy DNA samples were prepared by the formamide modification of the Kleinschmidt technique as reported by Davis, Simon and Davidson (1971). Contour length measurements were made and analyzed as described recently (Tibbetts, 1977).
DNA from 249
Incomplete
Adenovirus
Particles
Acknowledgments
of the noninfectious (defective) component in pools of adenoviruses type 12 and simian adenovirus 7. J. Virol. 14, 392-401.
This paper is dedicated This investigation was Cancer institute. Received
to the memory of Jerome Vinograd. supported by a grant from the National
May 20. 1977;
revised
June
Sharp, P. A., Sugden, B. and Sambrook, J., (1973). Detection of two restriction endonuclease activities in H. parainfluenzae using analytical agarose-ethidium bromide electrophoresis. Biochemistty 12, 3055-3063.
16, 1977
Shimojo, H., Yamashita, T., Yamamoto, C., (1967). Empty particles of adenovirus Biol. 19, l-8.
References Burlingham, B. T., Brown, D. T. and Doerfler, W. (1974). Incomplete particles of adenovirus: I. Characteristics of the DNA associated with incomplete adenovirions of types 2 and 12. Virology 60, 41 g-430. Daniell, E., (1976). Genome structure adenovirus. J. Virol. 19, 685-708.
of incomplete
particles
of
Davis, R. W., Simon, M. and Davidson, N. (1971). Electron microscope heteroduplex methods for mapping regions of base sequence homology in nucleic acids. In Methods in Enzymology, 21D, L. Grossman and K. Moldave, eds. (New York: Academic Press), pp. 413-426. Edvardsson, B., Everitt, L. (1976) Intermediates 547.
E., Jbrnvall, in adenovirus
H., Prage, assembly.
L. and Philipson, J. Virol. 19, 533-
Gallimore, P. H., Sharp, P. A. and Sambrook, J. (1974). Viral DNA in transformed cells: II. A study of the sequences of adenovirus 2 DNA in nine lines of transformed rat cells using specific fragments of the viral genome. J. Mol. Biol. 89, 49-71. Garon, C. F., Berry, K. W. and Rose, terminal redundancy in adenovirus Acad. Sci. USA 69, 2391-2395.
J. A. (1972). A unique form of DNA molecules. Proc. Nat.
Garon, C. F., Berry, K. W. and Rose, J. A. (1975). Arrangement sequences in the inverted terminal repetition of adenovirus DNA. Proc. Nat. Acad. Sci. USA 72, 3039-3043.
of 18
Graham, F. L., van der Eb, A. J. and Heijneker, H. L. (1974). and location of the transforming region of human adenovirus 5 DNA. Nature 251, 687-691.
Size type
Igarishi, K., Niiyama, Y.,Tsukamoto, K., Kurokawa, T. and Sugino, Y. (1975). Biochemical studies on bovine adenovirus type 3: II. Incomplete virus. J. Virol. 76, 634-641. Ishibashi, M. and Maizel, J. V., Jr. (1974). The polypeptides of adenovirus: V. Young virions, structural intermediates between top components and aged virions. Virology 57, 409-424. KBhler, K., (1962). Gewinnung und Reinigung 2 mit Hilfe der CsCl Gradientzentrifugation. 544-547.
von Adenovirus-Typ Z. Naturforsch. 776,
Maizel, J. V., Jr., White, D. 0. and Scharff, M. D., (1968). The polypeptides of adenovirus: II. Soluble proteins, cores, top components and the structure of the virion. Virology 36, 126-136. Mak, S. (1971). Defective Virol. 7, 426-433.
virions
in human
adenovirus
type
12. J.
Niiyama, Y., Igarishi, K., Tsukamoto, K., Kurokawa, T. and Sugino, Y., (1975). Biochemical studies on bovine adenovirus type 3: I. Purification and properties. J. Virol. 76, 621-633. Prage; L., HBglund, S. and Philipson, L., (1972). Structural teins of adenoviruses: VIII. Characterization of incomplete cles of adenovirus type 3. Virology 49, 745-757. Roberts, R. J., Arrand, the terminal repetition USA 71, 3829-3833. Rosenwirth, Incomplete polypeptide 437. Schaller,
proparti-
J. R. and Keller, W. (1974). The length of in adenovirus-2 DNA. Proc. Nat. Acad. Sci.
B., Tjia, S., Westphal, M. and Doerfler, W., (1974). particles of adenovirus: II. Kinetics of formation and composition of adenovirus Type 2. Virology 60, 431J. P. and Yohn,
D. S., (1974).
Transformation
potentials
H., Suzuki, 12. Jap.
Smith, K. 0. (1965). Studies on adenovirus-12: correlations between some physical, antigenic properties. J. lmmunonol. 94, 976-989. Sundquist, Assembly
B., Everitt, E. Philipson, of adenoviruses. J. Virol.
Tibbetts, adenovirus
C. (1977). Physical organization genomes. J. Virol., in press.
E. and Abe, J. Med. Sci. I. Quantitative and infectious
L. and Hdglund. 17, 449-459. of subgroup
S. (1973). B human
Tibbetts, C., Johansson, K., and Philipson, L. (1973). Hydroxyapatite chromatography and formamide denaturation of adenovirus DNA. J. Virol. 72, 218-225. Wadell, G., Hammerskjdld, M. and Varsanyi, T., (1973). Incomplete virus particles of adenovirus type 16. J. Gen. Virol. 20; 287302. Wolfson, inverted 3057.
J. and Dressler, D. (1972). terminal repetition. Proc.
Adenovirus Nat. Acad.
2 DNA contains an Sci. USA 69, 3054-
Vol. 12, 243-249,
September
1977, Copyright
0 1977 by MIT
Viral DNA Sequences from Incomplete Human Adenovirus Type 7 Clark Tibbetts Department of Microbiology School of Medicine University of Connecticut Health Farmington, Connecticut 06032
Center
Summary Large pools of empty viral capsids accumulate in cells infected by subgroup B human adenoviruses. Such infected cells also yield DNA-containing incomplete particles in larger quantities than cells infected with serotypes representing other adenovirus subgroups. DNA isolated from carefully purified classes of Ad7 incomplete particles was analyzed by restriction endonuclease cleavage, gel electrophoresis and electron microscopy. At least 90% of the DNA molecules in each sample consisted of sequences that extended from the left end of the viral genome map by variable lengths toward the right end. The average length of DNA is linearly related to the average buoyant density of the incomplete particles from which the DNA is isolated. The results indicate that each capsid contains one DNA molecule. There is also a specific association of the left end of the viral genome with assembled or assembling capsids. The characteristic distributions of Ad7 incomplete particles may result from intracellular pools of assembly intermediates in which the incompletely packaged DNA has been fragmented in vivo or by shear during preparative procedures. Introduction Buoyant CsCl density gradient centrifugation has been used in procedures for the purification and characterization of many different viruses. Early applications in the preparation of adenoviruses led to observations of viral particles with buoyant densities significantly lower than those of infectious virions. The lightest of these particles were identified by electron microscopy and polypeptide analysis as empty adenovirus capsids (Kohler, 1962; Smith, 1965; Shimojo et al., 1967; Maizel, White and Scharff, 1968). These empty capsids have been implicated as intermediates in virion assembly, formed prior to the interaction with newly synthesized viral DNA in the nuclei of infected cells (Sundquist et al., 1973). Adenovirus particles have also been observed at positions in CsCl density gradients between the virions and empty capsids. These incomplete particles also have an adenovirus morphology and contain, in general, viral DNA sequences shorter than full-length virion DNA (Prage, Hoglund and Phiiipson, 1972; Wadell,
Particles
of
Hammerskjold and Varsanyi, 1973; lshibashi and Maizel, 1974; Burlingham, and Brown Doerfler, 1974; Rosenwirth et al., 1974; Daniell, 1976). Serotypes representing different adenovirus subgroups appear to give rise to characteristic distributions of virions, incomplete particles and empty capsids in preparative CsCl density gradients. Preparations from cells infected by subgroup 6 human adenoviruses (Ad3, Ad7, Ad16) generally have the largest quantities of incomplete particles and empty capsids (Prage et al., 1972; Wadell et al., 1973; Daniell, 1976; Tibbetts, unpublished observations). A variety of reports have dealt with the biological properties of incomplete adenovirus particles. The infectivity associated with such particles is orders of magnitude less than that of virions (Prage et al., 1972; Wadel.1 et al., 1973; Burlingham et al., 1974; Niiyama et al., 1975). The infectivity is most probably due to contamination of preparations with trace quantities of infectious virions. Incomplete particles of human Ad3 and Ad7 do not appear to interfere with infection by virions nor to alter the eventual yield or distribution of virions, incomplete particles and empty capsids (Daniell, 1976; C. Tibbetts, unpublished observations). incomplete particles of human Ad12 and bovine Ad3, which contain nearly full-length viral DNA, seem to be prevalent only after passaging the virus by high multiplicity infection (Mak, 1971; Niiyama et al., 1975; lgarishi et al., 1975). There have been reports of adenovirus incomplete particles having potential for malignant transformation of cells in culture (Schaller and Yohn, 1974; lgarishi et al., 1975). This seems to be quite feasible, since only a limited region of the adenovirus genome, in particular the left-end and 7-l 4%, appears to be necessary for the transformation of cells (Graham, van der Eb and Heijneker, 1974; Gallimore, Sharp and Sambrook, 1974). Some incomplete particle DNA may represent the transforming DNA sequences, yet lack other viral genes that have cytotoxic effects compromising transformation assays. Studies of viral polypeptides specifically associated with virions or incomplete and empty particles have led investigators to suggest relationships of these particles to the in vivo sequence of steps in virion assembly (Sundquist et al., 1972; lshibashi and Maizel, 1974; Edvardsson et al., 1976). Parallel studies of the DNA associated with incomplete adenovirus particles may lead to further understanding of their origin as well as their various biological properties. This study describes incomplete particles of human Ad7 found in CsCl density gradients throughout the region between virions and empty capsids. Viral DNA was extracted from nine purified buoyant density classes of incomplete particles and characterized by gel electrophoresis, restric-
Cell 244
tion endonuclease cleavage and electron microscopy. The results are consistent with a model based on specific association of the left end of one Ad7 genome with assembled capsids during the synthesis of mature virions. The incomplete particles observed in density gradients may be the result of fragmentation of incompletely packaged DNA in vivo or by shear during preparative procedures.
Results Incomplete
Particles of Ad7
HeLa cells were infected with Ad7 virus and labeled with “C-formate (Experimental Procedures) to provide radioactivity in viral DNA and protein. Sonicates of infected cells were cleared of debris by low speed centrifugation (1000 g) and banded to equilibrium in self-generating CsCl density gradients (Figure 1). Opalescent bands corresponding to virions, empty capsids and the predominant species of incomplete particles [as described for Ad3 by Prage et al. (1972)J were readily observed. In addition to these three major bands, there was a more diffuse
c1 I
10
Virions
+ ...... .. ... .. .. .. . ..
._
..
\
.
E
op3hl-
FRACTIONS Figure 1. Equilibrium beled Ad7 Particles
CsCl
from Density
CsCl Gradients
GRADIENTS of %-Formate-La-
Approximately lo9 HeLa cells (2 liters) were infected with Ad7 virus and labeled with 3 mCi of “C-formate. The cells were harvested (64 hr post-infection), and virus was prepared as described in Experimental Procedures. The distribution of radioactivity associated with unfractionated particles and residual soluble protein is shown in the upper panel. Material in the indicated interval “upper pool” was combined and centrifuged to equilibrium with the resultant distribution shown in the lower panel. Eleven fractions (IP-a to IP-k) were pooled as indicated, recentrifuged to equilibrium and subjected to the viral DNA extraction procedure.
distribution of similarly opalescent material in the region between the virions and major incomplete particles. In the intermediate region, there were also bands of presumed adenovirus particles which appeared in characteristic positions and relative amounts. The overall appearance was quite similar to the density gradient distributions seen with other subgroup B human adenoviruses [Ad3, (Daniell, 1976); Ad1 6 (Wade11 et al., 1973)]. Material appearing above the virions was pooled as shown in the upper panel of Figure 1 and rebanded to equilibrium in CsCl density gradients. This gradient was fractionated, as indicated in the lower panel of Figure 1, to provide eleven pools corresponding to density classes throughout the range from virions to empty capsids. These pools were each rebanded to equilibrium in CsCl density gradients (not shown). Each of the gradients showed a predominant band at a position corresponding to the density of the region from which the particles had been originally pooled. Pools from IP-b and IP-g also showed minute amounts of material in a second band at the position corresponding to buoyant virions. The major bands from each gradient were recovered by fractionation and subjected to the viral DNA extraction procedure (Experimental Procedures). Each sample, except pool IP-k, yielded labeled DNA (about 2 x lo4 cpm pg) at positions near 1.7 g/ml in the buoyant density gradients. The analysis of these DNA samples presented below specifically relates to the samples pooled as in Figure 1. Several similar preparations of Ad7 incomplete particles have been examined with equivalent results.
Incomplete
Particle DNA
Samples (1 pg) of DNA from incomplete particle pools IP-b to IP-j were mixed with 0.1 pg Ad7 viral DNA (2.3 x 10’ daltons) and subjected to electrophoresis in 0.4% agarose gels (Figure 2). The DNA from each pool represents linear duplex DNA (see electron microscopy results in this section) and migrates as a rather unimodal distribution of sizes. The average lengths of DNA from the pools varied from 80% (IP-b) to 12% (IP-j) of full-length AD7 DNA, and in each case, the average length was linearly related to the density of the pooled particles from which the DNA had been prepared. This is similar to earlier observations of the DNA associated with more discrete bands of incomplete particles of Ad3 (Prage et al., 1972; Daniell, 1976) and of Ad2 and Ad12 (Burlingham et al., 1974). The result suggests that each incomplete particle of adenovirus includes only one less than full viral length DNA molecule. Restriction endonucleases have recently been used to provide physical maps of Ad7 virion DNA (Tibbetts, 1977). Viral DNA and DNA from pools
DNA from 245
Incomplete
Adenovirus
Particles
IP-a through IP-J were analyzed by digestions with endoR.Sma I (Figure 3), endoR.Hind III (Figure 4), endoR.Bam HI (not shown), and a mixture of endoR.Sma I and endoR.Hpa I (not shown). Cleavage of DNA from the Ad7 incomplete particle COpools led to restriction fragments which migrated with the fragments of Ad7 DNA generated with the same endonuclease. In general, >90% of the restricted DNA from the incomplete particles was found to co-migrate with one or more Ad7specific restriction fragments. Quantitative analysis of these results by scanning densitometry of photographic negatives of the ethidium-stained gels (as for the Hind III experiment in Figure 4) was based upon the assumption of sequence identity of co-migrating fragments from virion and incomplete particle DNA. This analysis revealed two groups of fragments in the restricted DNA from Ad7 incomplete particles. Some fragments were present in equimolar quantities (when related to equivalent quantities of viral DNA fragments and corrected for the smaller size of incomplete particle DNA molecules), while the remaining fragments apoeared in an order of decreasing, less than equimolar, amounts. These patterns of distributions of restriction fragments, for three of the enzymes used in this study, are sum-
RI
b
c
d
e
f
g
h
i
i
marized in Figure 5 by relating the equimolar set of fragments and the next most abundant fragment of each digest to the known maps of restriction sites in Ad7 DNA. It is readily apparent that the bulk of DNA associated with the incomplete particle pools of Ad7 represents the linear progression of viral DNA sequences from the left end of the genome map to a variable right terminus determined by the average size of DNA in the particles representing each particular buoyant density class. Electron microscopy of DNA preparations from pools IP-a to IP-j revealed exclusively linear duplex DNA. Contour length analysis of the DNA from each pool led to results such as those shown for pool IP-f in Figure 6. Each distribution was unimodal; the mean length of each distribution was sufficient to account for only single-copy representation of each of the observed equimolar restriction fragments of DNA from the respective pool. The mean lengths were linearly related to the density of the pool of particles from which the DNA had been isolated. Most of the viral DNA sequences associated with the Ad7 incomplete particles seemed to be free of inverted repetitions of the Ad3 type described by Daniel1 (1976). Samples of DNA representing Ad7 incomplete particle pools were denatured with alkali, neutralized and held at low DNA concentration (0.2-0.5 pg/ml) under reassociation conditions
a
:
bCdefehii7
B
Figure
2. Electrophoresis
of Ad7 Incomplete
Particle
DNA
Samples of DNA (about 1 fig) from incomplete particle pools Ip-b through IP-j were mixed with about 0.2 pg full-length Ad7 DNA and subjected to electrophoresis in 0.4% agarose gels (Experimental Procedures). The left gel shows two of the three Eco RI restriction fragements of Ad7 DNA (A = 35.4%; 6 = 13.0%) to indicate the approximate range of DNA lengths associated with the different incomplete particle pools. The DNA from the denser pools may appear to be less heterogeneous in size than that from the iighter pools, but this must be attributed to the lower resolution of the electrophoresis system with DNA larger than about 1 x 10’ daltons. The distributions of DNA are rather unimodal, but there is some indication of discrete lengths of DNA appearing as bands within the continuous distribution of most of the incomplete particle DNA.
Figure 3. Restriction endoR.Sma I
of Ad7 Viral and Incomplete
Particle
DNA by
Each DNA sample (about 1 pg) was mixed with about 0.05 pg of Ad7 viral DNA prior to incubation with the restriction enzyme to control for complete cleavage. Prior to electrophoresis. uncut Ad7 viral DNA (0.2 pg) was added to each sample as a position marker. At least 90% of the incomplete particle pool DNA samples could be recovered as fragments co migrating with known Sma I cleavage fragments of Ad7 DNA.
Cell 246
i
E
+
l +
I-:::. .O
.
.2 FRACTfON
Figure 5. Physical Particles
Maps
.
.
.6 .4 of Ad 7 GENOME
of DNA Associated
.6
.
’
1.0
LENGTH
with
Ad7 Incomplete
DNA from each pool of incomplete particles was analyzed as shown in Figure 4 with restriction endonucleases Sma I, Barn I and Hind ill (upper, middle and lower bar of each set). The solid lines represent those fragments which were recovered in singlecopy yield from each DNA sample. The gap between the right end of each bar and the + symbol represent the next most abundant cleavage fragment of the digest. The + symbols together indicate the descending region of the DNA length distributions where only longer molecules will yield the restriction fragments indicated by the gaps. These lengths are in agreement with contour length analysis by electron microscopy (Figure 6) and are linearly related to the buoyant density of the fractions pooled for the preparation of each DNA sample (Figure 1).
Figure 4. Relative Amounts of Hind III Cleavage Restricted Ad7 Viral and Incomplete Particle DNA
Fragments
in
Film densitometry of negatives permitted quantitative estimation of the amounts of each Hind Ill cleavage fragment associated with the DNA from incomplete particle pools. The residual DNA at positions corresponding to viral DNA cleavage fragments is about 5% of the original amount shown in the IP-a pool scan and represents the viral DNA added to each sample to control for complete cleavage. These were not present in other samples run with the incomplete particle DNA alone. Negatives had exposure times limiting the regions of the most intense absorbance to the range of linear film response to ethidium fluorescence.
[50% formamide, 0.10 M Tris (pH 84, 0.01 M EDTA, 37”C] until subsequent preparation for electron microscopy. The observed distributions of contour lengths of the linear single strands (>95% of the molecules in samples prepared within 5 min) were somewhat broader than seen with the duplex DNA before denaturation, but the mean lengths were the same (k 10% of Ad7 genome length) for the duplex and denatured DNA of each pool. Fewer than 5% of the molecules, held up to 2 hr before spreading with cytochrome c, were observed to form single-stranded circles. Some of these singlestranded structures had rather long duplex-appearing panhandles, reminiscent of the stem-loop structures described by Daniel1 (1976) for Ad3 incomplete particle DNA. Renatured molecules had the generally simple appearance of linear duplex
DNA from 247
Incomplete
Adenovirus
Particles
t
.O
4
.2 FRACTION
Figure 6. Contour cle Pool IP-f
Length
.4 .6 of Ad 7 GENOME Analysis
.8
1.0
LENGTH
of DNA from
Incomplete
Parti-
DNA from pool IP-f was mixed with form II plasmid P,,, DNA (length marker) for contour length analysis by electron microscopy (Experimental Procedures). The histogram shows the results of 175 length measurements plotted as 2% intervals of Ad7 genome length. The IP-f pool represents DNA from particles having a buoyant position nearly midway betwen virions and empty capsids in CsCl density gradients (Figure 1). Notice that the bulk of the unimodal distribution represents molecules slightly longer than the left-end sequences corresponding to restriction fragments observed in single-copy amounts, but shorter than the restriction sites corresponding to fragments obtained in less than equimolar yield (Figure 5). Analysis of the other incomplete particle pools gave results similar to these obtainedwith IP-f.
DNA with a short single-stranded DNA tail of variable length at only one end of each molecule. Similar structures have been observed in DNA from incomplete particles of a murine adenovirus (D. Nathans, personal communication. Discussion The results presented in this report suggest that Ad7 capsids can interact with only one viral DNA molecule, and that this interaction is in some manner specifically associated with the molecular left end of the conventional viral genome map. Most of the DNA associated with Ad7 incomplete particles are simple linear duplex sequences from the left end of the Ad7 genome to a variable right terminus. This left end specificity is intriguing, since the molecular ends of adenovirus DNA are identical with regard to their nucleotide sequences (regions of inverted terminal repetition) and the presence of the covalently bound 5’ end proteins on each complementary strand of viral DNA (Garon, Berry and Rose, 1972; Wolfson and Dressler, 1972; Roberts, Arrand and Keller, 1974; R. Roberts, personal communication). Studies of adenovirus polypeptides in virions, incomplete particles and empty capsids have led
other investigators to suggest a precursor-intermediate-product relationship of these particles through virion assembly in vivo (Sundquist et al., 1973; lshibashi and Maizel, 1974; Edvardsson et al., 1976). The framework of the hypothesis is supported by radioactive labeling results of pulsechase experiments and the clear evidence for cleavage processing of certain precursor polypeptides, present in empty capsids and incomplete particles, to their respective mature virion-associated products. This background leads to an appealing interpretation of the properties of viral DNA sequences associated with incomplete particles of Ad7 (Figure 7). Empty capsids are proposed as preassembled intermediates, destined to interact with one viral DNA molecule and its associated proteins (perhaps a precursor to the virion core). The molecular left end of the viral DNA is specifically inserted, leaving about 85% of the viral DNA extending from the capsid. The following steps in the packaging of the viral genome would require substantial compression of the volume occupied by the DNA and possible displacement or cleavage processing of polypeptides associated with the previously empty capsids. Final steps in virion assembly may involve restoration or completion of external capsomeres in the region where DNA enters the capsid. The incomplete particles which are observed in preparations of adenovirus from infected cells may be derived from intermediates of the assembly process proposed above. Incompletely packaged DNA may be fragmented in vivo during abortive assembly. It would certainly be fragmented in vitro by shear forces associated with preparative procedures. The proposed model suggests that the initial interaction of empty capsids with the DNA (or precore), and the subsequent packaging steps are rate-limiting in virion assembly, permitting accumulation of intermediate structures predisposed to yield abundant quantities of empty capsids or incomplete particles. The types of particles which are most abundant in extracts of ceils infected by subgroup B adenoviruses are empty particles, an incomplete species with about IO-15% of the viral genome (pool IP-j, Figure 1) and mature virions. An alternative model might suggest that the incomplete particles represent breakdown products of labile assembly intermediates. This could imply a preferential extrusion and fragmentation of the molecular right portion of the viral genome from more nearly complete assembly intermediates. It was suggested by Edvardsson et al. (1976) that cetain assembly intermediates are not stable in CsCl density gradients. When such particles were isolated from their preparative Ficoll gradients, fixed with gluteraldehyde and banded in GsCl density gra-
Cell 240
IN PACKAGING Ad7 VIRAL DNA
MATURE VIRIONS
Figure
7. Packaging
of Adenovirus
II
1.340 V
DNA
The results of this study suggest that viral DNA interacts with empty capsids by specific association (entry?) of one end of the viral genome (the left end of conventional physical maps). The subsequent packaging of the remaining DNA probably reflects a concerted process of altering DNA conformation within the particle and displacement or cleavage of internal polypeptides of the capsids. Subgroup B human adenoviruses yield large quantities of incomplete particles in density gradients which may reflect DNA fragmentation, in vivo or during preparative procedures, in intermediates which accumulate in cells during the DNA packaging process. Subtle differences in virus-coded polypeptides specified by serotypes representing other subgroups could account for different rate-limiting steps in the course of virus assembly. This would then lead to accumulation of intracellular intermediates and thus different types and relative amounts of incomplete particles in preparative density gradients.
dients, however, the predominate breakdown products appeared to be empty capsids and DNAcontaining cores. There are several types of DNA structures which have been described in preparations of incomplete particles from different adenovirus serotypes and subgroups. Preparations of human Ad12 (Mak, 1971; Garon et al., 1972; C. Tibbetts, unpublished observations) and bovine Ad3 (Niiyama et al., 1975) suggest that dense species of incomplete particles in these systems represent defective but replicating DNA of less than full length, with heterologous sequences located specifically near one end of the viral genome maps. DNA structures observed in preparations of human Ad3 (Daniell, 1976) and human Ad1 8 (Garon, Berry and Rose, 1975) appear to contain long inverted terminal repetitions of sequences from one end of the viral DNA. In these cases, electron microscopy of denatured DNA molecules reveals panhandle structures much longer than those resulting from association of the inverted terminal repetition sequences of denatured virion DNA (Wolfson and Dressler, 1972; Garon et al., 1972). None of these cases precludes, however,
the type of model which has been proposed for adenovirus assembly. The results presented for Ad3 incomplete particle DNA by Daniel1 (1976) are in some respects qualitatively similar to those presented for Ad7 in this report. This should be expected, since Ad3 and Ad7 are very closely related subgroup B human adenoviruses. The DNA structures emphasized by Daniel1 (linear duplex DNA with long inverted terminal repetitions) have also been observed in my own preparations of Ad3 and Ad7 incomplete particle DNA. I have not, however, found such structures to represent more than about 5% of the total incomplete particle DNA preparations. The abortive DNA replication scenario proposed by Sambrook (Daniell, 1976) may indeed relate to these novel DNA structures. It is the quantitative consideration of the viral DNA sequences associated with incomplete particles of Ad7 which has led me to emphasize their possible relationship to events in adenovirus assembly. Experimental
Procedures
Cells and Viruses Adenovirus type 7 (Gomen) and type 3 (G.B.) werepropagated in suspension cultures of HeLa S3 cells as described previously (Tibbetts, Johansson and Philipson, 1973). DNA Adenovirus DNA was extracted from virions or incomplete particles by incubation with sarkosyl and pronase 6, and purified by extraction with phenol, chloroform, isoamyl alcohol followed by centrifugation in isopycnic CsCl density gradients (Tibbetts et al., 1973). Labeling of Ad7-infected cell cultures with “C-formate was performed as previously described (Tibbetts et al., 1973). Restriction Conditions sites in Ad7 manuscript
Endonucleases for enzyme digests and DNA have been presented submitted).
physical maps of restriction elsewhere (Tibbetts, 1977).
Density Gradients Sonicates of virus-infected cells or partially purified preparations were adjusted to1.34 g/ml with CsCl and centrifuged in angle rotors (Ti50 or Ti75) at 30,000 rpm, 4°C for 36-48 hr. Gradients were fractionated by tube puncture at the bottom; 50 or 70 PI fractions (7 or 10 drops) were collected. Radioactivity in the gradient was monitored by drying 5 ~1 aliquots of each fraction on paper filters and counting in a liquid scintillation spectrometer with a toluene-based fluor (Tibbetts et al., 1973). Gel Electrophoresis DNA samples were subjected to electrophoresis in agarose gels as originally suggested by Sharp, Sugden and Sambrook (1972). Specific details of my applications were recently described (Tibbetts, 1977). Electron Microscopy DNA samples were prepared by the formamide modification of the Kleinschmidt technique as reported by Davis, Simon and Davidson (1971). Contour length measurements were made and analyzed as described recently (Tibbetts, 1977).
DNA from 249
Incomplete
Adenovirus
Particles
Acknowledgments
of the noninfectious (defective) component in pools of adenoviruses type 12 and simian adenovirus 7. J. Virol. 14, 392-401.
This paper is dedicated This investigation was Cancer institute. Received
to the memory of Jerome Vinograd. supported by a grant from the National
May 20. 1977;
revised
June
Sharp, P. A., Sugden, B. and Sambrook, J., (1973). Detection of two restriction endonuclease activities in H. parainfluenzae using analytical agarose-ethidium bromide electrophoresis. Biochemistty 12, 3055-3063.
16, 1977
Shimojo, H., Yamashita, T., Yamamoto, C., (1967). Empty particles of adenovirus Biol. 19, l-8.
References Burlingham, B. T., Brown, D. T. and Doerfler, W. (1974). Incomplete particles of adenovirus: I. Characteristics of the DNA associated with incomplete adenovirions of types 2 and 12. Virology 60, 41 g-430. Daniell, E., (1976). Genome structure adenovirus. J. Virol. 19, 685-708.
of incomplete
particles
of
Davis, R. W., Simon, M. and Davidson, N. (1971). Electron microscope heteroduplex methods for mapping regions of base sequence homology in nucleic acids. In Methods in Enzymology, 21D, L. Grossman and K. Moldave, eds. (New York: Academic Press), pp. 413-426. Edvardsson, B., Everitt, L. (1976) Intermediates 547.
E., Jbrnvall, in adenovirus
H., Prage, assembly.
L. and Philipson, J. Virol. 19, 533-
Gallimore, P. H., Sharp, P. A. and Sambrook, J. (1974). Viral DNA in transformed cells: II. A study of the sequences of adenovirus 2 DNA in nine lines of transformed rat cells using specific fragments of the viral genome. J. Mol. Biol. 89, 49-71. Garon, C. F., Berry, K. W. and Rose, terminal redundancy in adenovirus Acad. Sci. USA 69, 2391-2395.
J. A. (1972). A unique form of DNA molecules. Proc. Nat.
Garon, C. F., Berry, K. W. and Rose, J. A. (1975). Arrangement sequences in the inverted terminal repetition of adenovirus DNA. Proc. Nat. Acad. Sci. USA 72, 3039-3043.
of 18
Graham, F. L., van der Eb, A. J. and Heijneker, H. L. (1974). and location of the transforming region of human adenovirus 5 DNA. Nature 251, 687-691.
Size type
Igarishi, K., Niiyama, Y.,Tsukamoto, K., Kurokawa, T. and Sugino, Y. (1975). Biochemical studies on bovine adenovirus type 3: II. Incomplete virus. J. Virol. 76, 634-641. Ishibashi, M. and Maizel, J. V., Jr. (1974). The polypeptides of adenovirus: V. Young virions, structural intermediates between top components and aged virions. Virology 57, 409-424. KBhler, K., (1962). Gewinnung und Reinigung 2 mit Hilfe der CsCl Gradientzentrifugation. 544-547.
von Adenovirus-Typ Z. Naturforsch. 776,
Maizel, J. V., Jr., White, D. 0. and Scharff, M. D., (1968). The polypeptides of adenovirus: II. Soluble proteins, cores, top components and the structure of the virion. Virology 36, 126-136. Mak, S. (1971). Defective Virol. 7, 426-433.
virions
in human
adenovirus
type
12. J.
Niiyama, Y., Igarishi, K., Tsukamoto, K., Kurokawa, T. and Sugino, Y., (1975). Biochemical studies on bovine adenovirus type 3: I. Purification and properties. J. Virol. 76, 621-633. Prage; L., HBglund, S. and Philipson, L., (1972). Structural teins of adenoviruses: VIII. Characterization of incomplete cles of adenovirus type 3. Virology 49, 745-757. Roberts, R. J., Arrand, the terminal repetition USA 71, 3829-3833. Rosenwirth, Incomplete polypeptide 437. Schaller,
proparti-
J. R. and Keller, W. (1974). The length of in adenovirus-2 DNA. Proc. Nat. Acad. Sci.
B., Tjia, S., Westphal, M. and Doerfler, W., (1974). particles of adenovirus: II. Kinetics of formation and composition of adenovirus Type 2. Virology 60, 431J. P. and Yohn,
D. S., (1974).
Transformation
potentials
H., Suzuki, 12. Jap.
Smith, K. 0. (1965). Studies on adenovirus-12: correlations between some physical, antigenic properties. J. lmmunonol. 94, 976-989. Sundquist, Assembly
B., Everitt, E. Philipson, of adenoviruses. J. Virol.
Tibbetts, adenovirus
C. (1977). Physical organization genomes. J. Virol., in press.
E. and Abe, J. Med. Sci. I. Quantitative and infectious
L. and Hdglund. 17, 449-459. of subgroup
S. (1973). B human
Tibbetts, C., Johansson, K., and Philipson, L. (1973). Hydroxyapatite chromatography and formamide denaturation of adenovirus DNA. J. Virol. 72, 218-225. Wadell, G., Hammerskjdld, M. and Varsanyi, T., (1973). Incomplete virus particles of adenovirus type 16. J. Gen. Virol. 20; 287302. Wolfson, inverted 3057.
J. and Dressler, D. (1972). terminal repetition. Proc.
Adenovirus Nat. Acad.
2 DNA contains an Sci. USA 69, 3054-
