Vol. 21, No. 1 Printed in U.S.A.
JOURNAL OF VIROLOGY, Jan. 1977, p. 292-300 Copyright © 1977 American Society for Microbiology
Structural Studies on the Polyhedral Inclusion Bodies, Virions, and DNA of the Nuclear Polyhedrosis Virus of the Cotton Bollworm Heliothis zea DAVID W. SCHARNHORST, KAY L. SAVING, SUSAN B. VUTURO, PETER H. COOKE,' AND ROBERT F. WEAVER* Departments of Biochemistry* and Physiology and Cell Biology, McCollum Laboratories, The University of Kansas, Lawrence, Kansas 66044
Received for publication 1 July 1976
The polyhedral inclusion body of the cotton bollworm nuclear polyhedrosis virus contains virions occluded in an orthogonal crystalline matrix. The virions appear as rods or, more frequently, as oval structures that form upon bending of the nucleocapsid within the viral membrane. The nucleocapsid consists at least of DNA surrounded by a capsid composed of subunits, possibly helically arranged. The viral DNA is circular and supercoiled. It is heterogeneous in size, with contour lengths ranging from 15 to 45 ,um.
The nuclear polyhedrosis virus (NPV) of the cotton bollwonn Heliothis zea is one of a group of occluded viruses that includes the cytoplasmic polyhedrosis viruses and the granulosis viruses. The NPVs and granulosis viruses are DNA viruses; the cytoplasmic polyhedrosis viruses contain RNA. A striking feature of the NPV is that the viruses are occluded within polyhedral bodies that are crystals of protein (4). The protein forms ordered high-molecular-weight aggregates at neutral pH, but it is dissociated into monomers under alkaline conditions. Thus, treatment of polyhedra at high pH can be used to free the virions. Released virions are reported to be very unstable (5). Initially the envelope of the virion was degraded, and this step was followed by loss of the capsid that surrounds the core containing the DNA. Some controversy surrounds the analysis of the genome size of NPVs. Onodera et al. (8) report sedimentation experiments that indicate that the infectious DNA of the NPV in the silkworm (Bombyx mori) has a molecular weight of only 2 x 106. Other investigators report much higher molecular weights (30 x 106 to 100 x 106) based on sedimentation and electron microscopic studies of a variety of NPVs and granulosis viruses (7, 11, 12). In this paper we present morphological analyses of the polyhedra, virions, and DNA of the NPV infecting the cotton bollworm H. zea. MATERIALS AND METHODS Purification of PIBs. A preparation of NPV from I Present address: Department of Physiology, University of Connecticut Health Center, Farmington, CT 06032.
H. zea was obtained from Sandoz-Wander, Homestead, Fla. This was a dry powder containing >1011 polyhedral inclusion bodies (PIBs)/g. These PIBs were purified by differential centrifugation following the procedures of Van der Geest (13) and Scott et al. (10). First, 2 g of powder was taken up in 20 ml of water and layered onto solutions of 61.7% (wt/wt) sucrose in water in JA-20 tubes and centrifuged for 30 min at 17,000 x g in the JA-20 rotor on a Beckman J-21 centrifuge. This and all subsequent steps were carried out at 0 to 4°C. The PIBs formed a layer on top of the sucrose solution. This layer was removed with a spatula, taken up in 20 ml of water, and layered onto solutions of 43.9% (wt/wt) sucrose in water in JA-20 tubes and centrifuged for 30 min at 17,000 x g. The PIBs formed a pellet, which was taken up in 5 to 10 ml of water and layered onto gradients of 10 to 80% (wt/ vol) sucrose in water. These gradients were centrifuged for 30 min at 50,000 x g in the SW41 rotor on a Beckman L2-65B ultracentrifuge. The PIBs formed a band near the bottom, at 1.28 g/ml. Lysis of PIBs and purification of virions. The PIBs were collected from the sucrose gradients with a Pasteur pipette and diluted with at least 1 volume of water. The PIBs were then pelleted by centrifugation for 15 min at 17,000 x g in the Beckman 35 rotor. This pellet was taken up in 20 ml of 0.1 M Na2CO3, pH 11.2, and allowed to stand at room temperature for 30 min to lyse the PIBs. Then the suspension was diluted with 5 to 10 volumes of sodium borate buffer, pH 9.0, and pelleted by centrifugation for 60 min at 18,000 x g in the JA-20 rotor. The pelleted virions were taken up in 5 ml of borate buffer and layered onto 10 to 80% (wt/vol) sucrose gradients in borate buffer in SW41 tubes. These gradients were centrifuged for 60 min at 50,000 x g. The virions formed a band approximately at the middle point in the gradient, at a density of 1.16 g/ ml. The virions were collected with a Pasteur pipette, 292
VOL. 21, 1977
STRUCTURE AND DNA OF COTTON BOLLWORM NPV
293
by 36 nm in diameter; see arrow no. 1 in Fig. 1B). This is surrounded by a 12-nm-thick coat (no. 2). This coat is enclosed in a 4-nm-wide space (no. 3) that is peripherally limited by an 8-nm layer (no. 4), which separates the virion from the matrix. The crystalline lattice of the matrix appears as two sets of orthogonal electron-dense lines (Fig. 1C). Regularly spaced projections (P) are observed, especially at the bottom surface of the PIB in Fig. 1A. The spacing between these projections is approximately 14 nm. This bears no obvious relationship to the 10-nm spacing between lattice planes in the PIB matrix. Structure of the virions. PIBs were disrupted in a solution of sodium carbonate, and the virions were isolated on sucrose gradients. The density of the virion was 1.16 g/ml. Figure 2A shows a field of negatively stained virions. The virions are morphologically heterogeneous. There are cylinders or rods (R) and ellipsoids or oval structures (0). Some virions have lost their outer membrane, leaving a dense core, or nucleocapsid (N). Several envelopes (E) are in the process of lysing, frequently revealing bent nucleocapsids (B) within. The free nucleocapsids are also heterogeneous. In some cases, they appear crimped, as if they had been bent (C). Evidence for such bending is presented below. One half-empty nucleocapsid (H) appears to be losing its DNA. Some nucleocapsids have lost all their DNA and appear as empty coats, or ghosts (G). In some cases, the envelope of the virion was not obviously lysed but was penetrated by stain. The enclosed core of these virions was bent into a V-shaped rod (Fig. 2B). Thin sections of purified virions also reveal the V-shaped cores (Fig. 2C). The nucleocapsid in the lower virion is cut in cross section, so the two arms of the bent nucleocapsid appear as dense circles. Within the preparation of isolated virions there are structures that correspond to the components seen in thin sections of PIBs (i.e., intact virions, cores with coat, and isolated coats). The isolated coats (Fig. 2D and E) consist of subunits arranged in lines almost perpendicular to the long axis of the coat. The spacing between these parallel lines is 4.8 nm. The angle of their deviation from the perpendicular is approximately 2°. Viral DNA. NPV DNA was purified by isopycnic banding on CsCl gradients containing RESULTS ethidium bromide. Breakage of DNA was miniMorphology of PIBs. Crude PIBs were puri- mized by lysing the purified virions directly on fied by isopycnic banding in sucrose density top of the gradients with sodium lauryl sarcosigradients. Their density is 1.28 g/ml. Figure 1A nate. Two bands of DNA are resolved, with shows a thin section of a PIB containing virions densities of 1.61 and 1.57 g/ml (Fig. 3A). This transected in various orientations. Each virion result is consistent with the hypothesis that the contains a dense cylindrical core (280 nm long dense band contains supercoiled DNA and the
diluted, pelleted, and resuspended as above and then subjected to rate-zonal centrifugation in 5 to 20% (wt/vol) sucrose gradients for 10 min at 50,000 x g in the SW41 rotor. The band of virions was collected, diluted, and pelleted as above and resuspended in borate buffer. Electron microscopy. Standard techniques were used for negative staining and sectioning and staining of PIBs and virions. Samples of viral DNA were prepared for electron microscopy by the protein monolayer technique of Kleinschmidt (6), as modified in the aqueous technique of Davis et al. (2). The purified viral suspension was diluted 1:20 with a solution containing 0.1 mg of cytochrome c per ml, 0.5 M ammonium acetate, and 1 mM EDTA (pH 7.5). DNA was then extracted from the virions by one of two methods: repeated freezing and thawing of the diluted viral suspension or heating at 60°C for 30 min. Approximately 50 ,ul of this spreading solution was allowed to run slowly down an acid-cleaned microscope slide onto a hypophase of 0.25 M ammonium acetate (pH 7.5). Samples of the DNA-protein monolayer 2 to 4 mm from the slide were picked up on carbon film-bearing copper grids by touching the grid to the surface. Excess liquid adhering to the grid was drawn off with filter paper. Prepared grids were stained in an acidic uranyl acetate solution. An aqueous stock solution containing 0.5 M HCI and 0.05 M uranyl acetate was diluted with 9 parts of absolute ethanol. Prepared grids were dipped in this ethanolic solution for 30 s, rinsed in absolute ethanol, and air dried. To increase contrast, the stained grids were shadow-cast with Pt-Pd at an angle of 50. All viewing was done with a Philips EM300 electron microscope operating at 60 kV. Purification of viral DNA by isopycnic gradient centrifugation. A modification of the method of Summers and Anderson (12) was used. Purified virions were mixed with 5 volumes of a solution containing 4% sodium lauryl sarcosinate in 0.05 M NaCl, 0.05 M Tris-hydrochloride, pH 7.9. This solution was heated at 60°C for 60 min to aid lysis of virions. Ethidium bromide was added to a concentration of 200 ug/ml, and the solution was layered gently with a wide-tipped pipette onto a solution of CsCl in an SW65 tube. This solution contained 200 ug of ethidium bromide per ml, 0.05 M NaCl, 0.05 M Trishydrochloride, pH 7.9, and had a density of 1.55. A density gradient was established by centrifuging for 24 h at 117,000 x g. During this time the viral DNA formed bands, which were detected by illumination with a UV lamp and removed with a wide-tipped pipette.
A
M.
B1
C
Now 'War
...or:
FIG. 1. (A) Thin section of a polyhedral inclusion body containing several enveloped virions. The crystalline lattice of the matrix is apparent. Virions are cut in transverse and longtitudinal planes; x108,000. (B) Higher magnification of a single virion cut transversely; x324,000. (C) Higher magnification of the matrix showing both lattice planes; x270,000. 294
I .$
;I
G"-.K
'H I,, I
-.t
I
I
-
-
.sO
/
.A *4iE,
k
'
B~~~ FIG. 2. (A) Negatively stained preparation of virions showing various stages of degradation: intact rodshaped virion (R); intact oval virion (0); bent nucleocapsid (B) shedding envelope (E); rod-shaped nucleocapsid (N); crimped nucleocapsid (C); nucleocapsid ghost (G); half-empty nucleocapsid (H); x60,OOO. (B) Negatively stained oval virion penetrated by stain; x217,000. (C) Thin section of oval virions; x205,000. (D and E) Negatively stained nucleocapsid ghosts showing surface structure; x104,000. 295
296
SCHARNHORST ET AL.
lo
J. VIROL.
300
20
0.6I
e:
0
~~~~~z
7~~~~~w
0.4
~~~~~~~~1.5 ~~fAC IO
NO. 14
0.2
pietin itwt
ietpedLn-eymcoi
10
20
FRACTION
30
NO.
FIG. 3. Isopycnic banding of viral DNA. CsCl carried out described gradient centrifugation was
in Materials and Methods. The DNA
pipetting
was
it
(A)
as
Unsheared DNA.
(B)
sheared before
with
banding by repeatedly a fine-tipped Lang-Levy micropi-
pette.
less dense band contains nicked circles and lin-
pieces of DNA. Further support for this comes from the finding that shearing the DNA by repeated pipetting with a Lang-Levy micropipette abolishes the dense band (Fig. ear
notion
3B). DNA from the dense band was collected and observed by electron microscopy. Some superhelical DNA was indeed observed (data not shown), but much of the DNA was broken, presumably during fractionation of the gradient and preparation of DNA for electron microscopy. To ameliorate this problem, virions were lysed gently and subjected immediately to spreading, staining, and shadowing for electron microscopy.
Because sedimentation through ethidium bromide-CsCl gradients indicated a circular or superhelical molecule, DNA molecules were selected for measurement by their clarity and absence of free ends. Micrographs were enlarged, and the contour lengths of the molecules were determined. A heterogenous popula-
tion of DNA molecules was observed, with contour lengths ranging from 15 to 45 jum. Some of these molecules are shown in Fig. 4A through D. Their lengths are 21, 23, 44, and 39 um, respectively. Figure 5 is a histogram showing the abundances of different size classes in the population of DNA circles we have measured. The most prominent size class is 20 to 25 ,m; many circles larger than this are observed, but their distribution is comparatively broad, with no distinct peak. At least part (36%) of the molecules viewed existed in the form of supercoils with a very high degree of twistedness. Some of the these supercoils are shown in Fig. 6A through C. The same heterogeneity of contour lengths seen in open circles is seen in the supercoils. The observation that the viral DNA exists as a supercoil suggests that the molecule may be packed into the virion in this form, but some additional folding must be required. This possibility is reinforced by our study. In Fig. 7, a nucleocapsid is seen releasing its DNA. Four distinct strands may be seen emerging from this particle, while part of the still densely packed DNA is visible within the nucleocapsid as a very thick, rodlike structure. No free ends are seen, suggesting that the four strands form two loops of DNA.
DISCUSSION Gregory et al. (4) have presented electron micrography of PIBs and virions of the NPV of H. zea that show the general morphology of these structures. Our studies confirm and extend these findings and permit the following conclusions. The virions are contained within a crystalline matrix in the PIB. The purified virions readily lose their envelopes, or viral membranes, revealing the nucleocapsids within. These nucleocapsids contain a sheath or capsid, with a regular subunit structure, surrounding the viral DNA. This capsid structure consists of subunits, presumably protein, arranged in parallel lines almost normal to the long nucleocapsid axis. The angle (880) and spacing (4.8 nm) of these lines is consistent with their arrangement in a helix with a pitch of 20. On the other hand, Beaton and Filshie (1) have perfonned optical diffraction experiments on the capsids of a number of NPVs and granulosis viruses. Their results favor a stacked-disk subunit structure, rather than a helical one. It is possible that our observed deviation from the perpendicular is an artifact due to distortion of the capsid. Harrap (5) has observed a similar fine structure in ghosts of the NPV of the gypsy moth
.: \t.||
r
B or
*;
wr s
,?
*5>t
..
xxl-e.
*
4
B
v\
i.
.-
l
*-:
.:
.
a
r0
-4
-A 9-
4
p
4l-~
b1
N. 4
.. i
A
B
C.
D
FIG. 4. Circular viral DNA molecules. Virions were prepared and lysed, and the resulting DNA was spread, stained, and shadow-cast as described in Materials and Methods. (A) x18,400; (B) x21,600; (C) x7,800; (D) x10,300. 297
298
SCHARNHORST ET AL.
J. VIROL.
4 E 0
c
2
0
50
20
30
40
50
length,p FIG. 5. Size distribution of circular DNA molecules. Micrographs were enlarged, and the DNA contour lengths were determined with a map measurer.
Porthetria dispar. A slight deviation of the lines of subunits from a line perpendicular to the longitudinal axis was also observed. We find no counterpart to the ringlike arrangement of subunits reported by Harrap. We favor his suggestion that they may be artifacts generated by superimposing front and back images, although we believe that the "back images" are the circular structures in the background of his micrograph. The nucleocapsids frequently bend, distorting their envelopes into V-shaped or, more usually, ellipsoidal structures. Alternatively, the envelopes may collapse into ellipsoidal shape, bending the nucleocapsids. These distorted virions also appeared in the preparations of Gregory et al. (4) but were not as abundant. It seems certain that this bending is an artifact of preparation, since it is not seen in intact polyhedra or in virions from freshly dissolved polyhedra (4). Two lines of evidence support the idea that the viral DNA is supercoiled. First, two bands of DNA are observed in isopycnic CsCl gradients in the presence of ethidium bromide. The separation between these two bands (0.04 g/ml) is what is observed in other systems with both superhelical and open circular DNA (9). The dense band disappears upon shearing the DNA, consistent with its identification as supercoiled DNA. Second, highly supercoiled DNA is observed in electron micrographs. Of course, the finding of supercoiled DNA implies that it is circular, and we also observe open circles by electron microscopy. Formation of a superhelix may be the first
order of compacting necessary to fit the large viral DNA into the small volume of the nucleocapsid sheath. A second order of packing is suggested by our electron micrograph showing two loops of a relaxed circle extending from one end of a nucleocapsid. It appears that the supercoiled DNA is folded over on itself at least once. This idea was previously proposed by Shvedchikova and Tarasevich (11) to explain their observations of abrupt thinning of DNA emerging from NPVs of the silkworm B. mori. We observe similar "despiralization" in heavily shadowed DNA preparations (data not shown). The size of the DNA of the bollworm NPV is still difficult to judge. We have observed apparently closed circles, including supercoils, ranging in size from 15 to 45 ,m. Some variation is to be expected, due to inconsistencies in stretching and shrinking during the preparation of grids for electron microscopy (10). A threefold variation, however, is outside the range of this experimental error. We can only conclude that we have sampled DNA molecules varying considerably in size. The distribution of size classes of DNA circles shows a peak at 20 to 25 ,u m, with a more or less continuous distribution of larger molecules ranging up to 40 to 45 ,um in length. We believe that the relatively small sample size and the lack of a marker DNA preclude firm conclusions about a possible bimodal distribution of DNA circles. It would seem, however, that the real genome size is approximately 20 to 25 ,um, and the larger molecules contain partial duplications of this fundamental unit. Assuming that one base pair contributes 0.34 nm to the length of the DNA, we estimate that 20 to 25 ,um corresponds to 40 x 106 to 50 x 106 daltons. This is in accord with our unpublished observations on the reassociation kinetics of NPV DNA, which suggest a genetic complexity of approximately 50 x 106 daltons. Finally, the question arises about the ability of the nucleocapsids to contain DNA of greater than one genome length. We calculate that the volume of the average cylindrical NPV nucleocapsid is greater than that of the icosohedral herpes simplex I nucleocapsid. Since the latter holds a DNA with a molecular weight of 100 x 106 (3), the NPV nucleocapsid would appear to be spacious enough to hold even the largest DNA molecules we observe. ACKNOWLEDGMENTS We thank Lorraine Hammer for expert technical assistance, H. Stone and R. Bussell for helpful suggestions on the manuscript, and the Sandoz-Wander Co., Homestead, Fla., for gifts of virus. This work was suported by Public Health Service grant GM 22127 from the National Institute of General Medical
0:X~~~~~~~¾ ; * v(
*
- *-i @ l s;N;
I.: *-6 ~w@eX-e;B*Sfi||; K| A
pt : Of0t-B:
A FIG. 6. Highly twisted viral DNA molecules. Preparation x10,700; (B) x10,500; (C) x20,400. 299
was described in Materials and Methods. (A)
300
SCHARNHORST ET AL.
J. VIROL.
4?
.6.
e
FIG. 7. DNA escaping from a virion. Preparation was described in Materials and Methods. x28,800. Sciences and by a grant from the General Research Fund of the University of Kansas. LITERATURE CITED 1. Beaton, C. D., and B. K. Filshie. 1976. Comparative ultrastructural studies of insect granulosis and nuclear polyhedrosis viruses. J. Gen. Virol. 31:151-161. 2. Davis, R. W., M. Simon, and N. Davidson. 1971. Electron microscope heteroduplex methods for mapping regions of base homology in nucleic acids. Methods Enzymol. 21D:413-428. 3. Frenkel, N., and B. Roizman. 1972. Separation of the herpes-virus deoxyribonucleic acid duplex into unique fragments and intact strand on sedimentation in alkaline gradients. J. Virol. 10:565-572. 4. Gregory, B. G., C. M. Ignoffo, and M. Shapiro. 1969.
Nucleopolyhedrosis of Heliothis: morphological de-
scription of inclusion bodies and virions. J. Invertebr. Pathol. 14:186-193. 5. Harrap, K. A. 1972. The structure of nuclear polyhedrosis viruses. Virology 50:114-123. 6. Kleinschmidt, A. K. 1968. Monolayer techniques in electron microscopy of nucleic acid molecules. Methods Enzymol. 12B:361-377.
7. Kok, I. P., A. V. Chirtynhova-Ryndich, and A. P. Gudz-Gorban. 1972. Macromolecular structure of DNA of the silkworm nuclear polyhedrosis virus. Mol. Biol. (USSR) 6:323-331. 8. Onodera, K., T. Komano, M. Himeno, and F. Sakai. 1965. The nucleic acid of nuclear-polyhedrosis virus of the silkworm. J. Mol. Biol. 13:532-539. 9. Radloff, R., W. Bauer, and J. Vinograd. 1967. A dyebuoyant density method for the detection and isolation of closed circular duplex DNA: the closed circular DNA in HeLa cells. Biochemistry 57:1514-1521. 10. Scott, H. A., S. Y. Young III, and J. A. McMasters. 1971. Isolation and some properties of components of nuclear polyhedra from the cabbage looper, Trichoplusia ni. J. Invertebr. Pathol. 18:117-182. 11. Shvedchikova, N. G., and L. M. Tarasevich. 1971. Electron Microscope investigation of granulosis viruses of Dendrolimus sibiricus and Agrotis segetum. J. Invertebr. Pathol. 18:25-32. 12. Summers, M. D., and D. L. Anderson. 1973. Characterization of nuclear polyhedrosis virus DNAs. Virology 12:1336-1346. 13. Van der Geest, L. P. S. 1968. A method for the purification of polyhedra. J. Invertebr. Pathol. 11:502.
JOURNAL OF VIROLOGY, Jan. 1977, p. 292-300 Copyright © 1977 American Society for Microbiology
Structural Studies on the Polyhedral Inclusion Bodies, Virions, and DNA of the Nuclear Polyhedrosis Virus of the Cotton Bollworm Heliothis zea DAVID W. SCHARNHORST, KAY L. SAVING, SUSAN B. VUTURO, PETER H. COOKE,' AND ROBERT F. WEAVER* Departments of Biochemistry* and Physiology and Cell Biology, McCollum Laboratories, The University of Kansas, Lawrence, Kansas 66044
Received for publication 1 July 1976
The polyhedral inclusion body of the cotton bollworm nuclear polyhedrosis virus contains virions occluded in an orthogonal crystalline matrix. The virions appear as rods or, more frequently, as oval structures that form upon bending of the nucleocapsid within the viral membrane. The nucleocapsid consists at least of DNA surrounded by a capsid composed of subunits, possibly helically arranged. The viral DNA is circular and supercoiled. It is heterogeneous in size, with contour lengths ranging from 15 to 45 ,um.
The nuclear polyhedrosis virus (NPV) of the cotton bollwonn Heliothis zea is one of a group of occluded viruses that includes the cytoplasmic polyhedrosis viruses and the granulosis viruses. The NPVs and granulosis viruses are DNA viruses; the cytoplasmic polyhedrosis viruses contain RNA. A striking feature of the NPV is that the viruses are occluded within polyhedral bodies that are crystals of protein (4). The protein forms ordered high-molecular-weight aggregates at neutral pH, but it is dissociated into monomers under alkaline conditions. Thus, treatment of polyhedra at high pH can be used to free the virions. Released virions are reported to be very unstable (5). Initially the envelope of the virion was degraded, and this step was followed by loss of the capsid that surrounds the core containing the DNA. Some controversy surrounds the analysis of the genome size of NPVs. Onodera et al. (8) report sedimentation experiments that indicate that the infectious DNA of the NPV in the silkworm (Bombyx mori) has a molecular weight of only 2 x 106. Other investigators report much higher molecular weights (30 x 106 to 100 x 106) based on sedimentation and electron microscopic studies of a variety of NPVs and granulosis viruses (7, 11, 12). In this paper we present morphological analyses of the polyhedra, virions, and DNA of the NPV infecting the cotton bollworm H. zea. MATERIALS AND METHODS Purification of PIBs. A preparation of NPV from I Present address: Department of Physiology, University of Connecticut Health Center, Farmington, CT 06032.
H. zea was obtained from Sandoz-Wander, Homestead, Fla. This was a dry powder containing >1011 polyhedral inclusion bodies (PIBs)/g. These PIBs were purified by differential centrifugation following the procedures of Van der Geest (13) and Scott et al. (10). First, 2 g of powder was taken up in 20 ml of water and layered onto solutions of 61.7% (wt/wt) sucrose in water in JA-20 tubes and centrifuged for 30 min at 17,000 x g in the JA-20 rotor on a Beckman J-21 centrifuge. This and all subsequent steps were carried out at 0 to 4°C. The PIBs formed a layer on top of the sucrose solution. This layer was removed with a spatula, taken up in 20 ml of water, and layered onto solutions of 43.9% (wt/wt) sucrose in water in JA-20 tubes and centrifuged for 30 min at 17,000 x g. The PIBs formed a pellet, which was taken up in 5 to 10 ml of water and layered onto gradients of 10 to 80% (wt/ vol) sucrose in water. These gradients were centrifuged for 30 min at 50,000 x g in the SW41 rotor on a Beckman L2-65B ultracentrifuge. The PIBs formed a band near the bottom, at 1.28 g/ml. Lysis of PIBs and purification of virions. The PIBs were collected from the sucrose gradients with a Pasteur pipette and diluted with at least 1 volume of water. The PIBs were then pelleted by centrifugation for 15 min at 17,000 x g in the Beckman 35 rotor. This pellet was taken up in 20 ml of 0.1 M Na2CO3, pH 11.2, and allowed to stand at room temperature for 30 min to lyse the PIBs. Then the suspension was diluted with 5 to 10 volumes of sodium borate buffer, pH 9.0, and pelleted by centrifugation for 60 min at 18,000 x g in the JA-20 rotor. The pelleted virions were taken up in 5 ml of borate buffer and layered onto 10 to 80% (wt/vol) sucrose gradients in borate buffer in SW41 tubes. These gradients were centrifuged for 60 min at 50,000 x g. The virions formed a band approximately at the middle point in the gradient, at a density of 1.16 g/ ml. The virions were collected with a Pasteur pipette, 292
VOL. 21, 1977
STRUCTURE AND DNA OF COTTON BOLLWORM NPV
293
by 36 nm in diameter; see arrow no. 1 in Fig. 1B). This is surrounded by a 12-nm-thick coat (no. 2). This coat is enclosed in a 4-nm-wide space (no. 3) that is peripherally limited by an 8-nm layer (no. 4), which separates the virion from the matrix. The crystalline lattice of the matrix appears as two sets of orthogonal electron-dense lines (Fig. 1C). Regularly spaced projections (P) are observed, especially at the bottom surface of the PIB in Fig. 1A. The spacing between these projections is approximately 14 nm. This bears no obvious relationship to the 10-nm spacing between lattice planes in the PIB matrix. Structure of the virions. PIBs were disrupted in a solution of sodium carbonate, and the virions were isolated on sucrose gradients. The density of the virion was 1.16 g/ml. Figure 2A shows a field of negatively stained virions. The virions are morphologically heterogeneous. There are cylinders or rods (R) and ellipsoids or oval structures (0). Some virions have lost their outer membrane, leaving a dense core, or nucleocapsid (N). Several envelopes (E) are in the process of lysing, frequently revealing bent nucleocapsids (B) within. The free nucleocapsids are also heterogeneous. In some cases, they appear crimped, as if they had been bent (C). Evidence for such bending is presented below. One half-empty nucleocapsid (H) appears to be losing its DNA. Some nucleocapsids have lost all their DNA and appear as empty coats, or ghosts (G). In some cases, the envelope of the virion was not obviously lysed but was penetrated by stain. The enclosed core of these virions was bent into a V-shaped rod (Fig. 2B). Thin sections of purified virions also reveal the V-shaped cores (Fig. 2C). The nucleocapsid in the lower virion is cut in cross section, so the two arms of the bent nucleocapsid appear as dense circles. Within the preparation of isolated virions there are structures that correspond to the components seen in thin sections of PIBs (i.e., intact virions, cores with coat, and isolated coats). The isolated coats (Fig. 2D and E) consist of subunits arranged in lines almost perpendicular to the long axis of the coat. The spacing between these parallel lines is 4.8 nm. The angle of their deviation from the perpendicular is approximately 2°. Viral DNA. NPV DNA was purified by isopycnic banding on CsCl gradients containing RESULTS ethidium bromide. Breakage of DNA was miniMorphology of PIBs. Crude PIBs were puri- mized by lysing the purified virions directly on fied by isopycnic banding in sucrose density top of the gradients with sodium lauryl sarcosigradients. Their density is 1.28 g/ml. Figure 1A nate. Two bands of DNA are resolved, with shows a thin section of a PIB containing virions densities of 1.61 and 1.57 g/ml (Fig. 3A). This transected in various orientations. Each virion result is consistent with the hypothesis that the contains a dense cylindrical core (280 nm long dense band contains supercoiled DNA and the
diluted, pelleted, and resuspended as above and then subjected to rate-zonal centrifugation in 5 to 20% (wt/vol) sucrose gradients for 10 min at 50,000 x g in the SW41 rotor. The band of virions was collected, diluted, and pelleted as above and resuspended in borate buffer. Electron microscopy. Standard techniques were used for negative staining and sectioning and staining of PIBs and virions. Samples of viral DNA were prepared for electron microscopy by the protein monolayer technique of Kleinschmidt (6), as modified in the aqueous technique of Davis et al. (2). The purified viral suspension was diluted 1:20 with a solution containing 0.1 mg of cytochrome c per ml, 0.5 M ammonium acetate, and 1 mM EDTA (pH 7.5). DNA was then extracted from the virions by one of two methods: repeated freezing and thawing of the diluted viral suspension or heating at 60°C for 30 min. Approximately 50 ,ul of this spreading solution was allowed to run slowly down an acid-cleaned microscope slide onto a hypophase of 0.25 M ammonium acetate (pH 7.5). Samples of the DNA-protein monolayer 2 to 4 mm from the slide were picked up on carbon film-bearing copper grids by touching the grid to the surface. Excess liquid adhering to the grid was drawn off with filter paper. Prepared grids were stained in an acidic uranyl acetate solution. An aqueous stock solution containing 0.5 M HCI and 0.05 M uranyl acetate was diluted with 9 parts of absolute ethanol. Prepared grids were dipped in this ethanolic solution for 30 s, rinsed in absolute ethanol, and air dried. To increase contrast, the stained grids were shadow-cast with Pt-Pd at an angle of 50. All viewing was done with a Philips EM300 electron microscope operating at 60 kV. Purification of viral DNA by isopycnic gradient centrifugation. A modification of the method of Summers and Anderson (12) was used. Purified virions were mixed with 5 volumes of a solution containing 4% sodium lauryl sarcosinate in 0.05 M NaCl, 0.05 M Tris-hydrochloride, pH 7.9. This solution was heated at 60°C for 60 min to aid lysis of virions. Ethidium bromide was added to a concentration of 200 ug/ml, and the solution was layered gently with a wide-tipped pipette onto a solution of CsCl in an SW65 tube. This solution contained 200 ug of ethidium bromide per ml, 0.05 M NaCl, 0.05 M Trishydrochloride, pH 7.9, and had a density of 1.55. A density gradient was established by centrifuging for 24 h at 117,000 x g. During this time the viral DNA formed bands, which were detected by illumination with a UV lamp and removed with a wide-tipped pipette.
A
M.
B1
C
Now 'War
...or:
FIG. 1. (A) Thin section of a polyhedral inclusion body containing several enveloped virions. The crystalline lattice of the matrix is apparent. Virions are cut in transverse and longtitudinal planes; x108,000. (B) Higher magnification of a single virion cut transversely; x324,000. (C) Higher magnification of the matrix showing both lattice planes; x270,000. 294
I .$
;I
G"-.K
'H I,, I
-.t
I
I
-
-
.sO
/
.A *4iE,
k
'
B~~~ FIG. 2. (A) Negatively stained preparation of virions showing various stages of degradation: intact rodshaped virion (R); intact oval virion (0); bent nucleocapsid (B) shedding envelope (E); rod-shaped nucleocapsid (N); crimped nucleocapsid (C); nucleocapsid ghost (G); half-empty nucleocapsid (H); x60,OOO. (B) Negatively stained oval virion penetrated by stain; x217,000. (C) Thin section of oval virions; x205,000. (D and E) Negatively stained nucleocapsid ghosts showing surface structure; x104,000. 295
296
SCHARNHORST ET AL.
lo
J. VIROL.
300
20
0.6I
e:
0
~~~~~z
7~~~~~w
0.4
~~~~~~~~1.5 ~~fAC IO
NO. 14
0.2
pietin itwt
ietpedLn-eymcoi
10
20
FRACTION
30
NO.
FIG. 3. Isopycnic banding of viral DNA. CsCl carried out described gradient centrifugation was
in Materials and Methods. The DNA
pipetting
was
it
(A)
as
Unsheared DNA.
(B)
sheared before
with
banding by repeatedly a fine-tipped Lang-Levy micropi-
pette.
less dense band contains nicked circles and lin-
pieces of DNA. Further support for this comes from the finding that shearing the DNA by repeated pipetting with a Lang-Levy micropipette abolishes the dense band (Fig. ear
notion
3B). DNA from the dense band was collected and observed by electron microscopy. Some superhelical DNA was indeed observed (data not shown), but much of the DNA was broken, presumably during fractionation of the gradient and preparation of DNA for electron microscopy. To ameliorate this problem, virions were lysed gently and subjected immediately to spreading, staining, and shadowing for electron microscopy.
Because sedimentation through ethidium bromide-CsCl gradients indicated a circular or superhelical molecule, DNA molecules were selected for measurement by their clarity and absence of free ends. Micrographs were enlarged, and the contour lengths of the molecules were determined. A heterogenous popula-
tion of DNA molecules was observed, with contour lengths ranging from 15 to 45 jum. Some of these molecules are shown in Fig. 4A through D. Their lengths are 21, 23, 44, and 39 um, respectively. Figure 5 is a histogram showing the abundances of different size classes in the population of DNA circles we have measured. The most prominent size class is 20 to 25 ,m; many circles larger than this are observed, but their distribution is comparatively broad, with no distinct peak. At least part (36%) of the molecules viewed existed in the form of supercoils with a very high degree of twistedness. Some of the these supercoils are shown in Fig. 6A through C. The same heterogeneity of contour lengths seen in open circles is seen in the supercoils. The observation that the viral DNA exists as a supercoil suggests that the molecule may be packed into the virion in this form, but some additional folding must be required. This possibility is reinforced by our study. In Fig. 7, a nucleocapsid is seen releasing its DNA. Four distinct strands may be seen emerging from this particle, while part of the still densely packed DNA is visible within the nucleocapsid as a very thick, rodlike structure. No free ends are seen, suggesting that the four strands form two loops of DNA.
DISCUSSION Gregory et al. (4) have presented electron micrography of PIBs and virions of the NPV of H. zea that show the general morphology of these structures. Our studies confirm and extend these findings and permit the following conclusions. The virions are contained within a crystalline matrix in the PIB. The purified virions readily lose their envelopes, or viral membranes, revealing the nucleocapsids within. These nucleocapsids contain a sheath or capsid, with a regular subunit structure, surrounding the viral DNA. This capsid structure consists of subunits, presumably protein, arranged in parallel lines almost normal to the long nucleocapsid axis. The angle (880) and spacing (4.8 nm) of these lines is consistent with their arrangement in a helix with a pitch of 20. On the other hand, Beaton and Filshie (1) have perfonned optical diffraction experiments on the capsids of a number of NPVs and granulosis viruses. Their results favor a stacked-disk subunit structure, rather than a helical one. It is possible that our observed deviation from the perpendicular is an artifact due to distortion of the capsid. Harrap (5) has observed a similar fine structure in ghosts of the NPV of the gypsy moth
.: \t.||
r
B or
*;
wr s
,?
*5>t
..
xxl-e.
*
4
B
v\
i.
.-
l
*-:
.:
.
a
r0
-4
-A 9-
4
p
4l-~
b1
N. 4
.. i
A
B
C.
D
FIG. 4. Circular viral DNA molecules. Virions were prepared and lysed, and the resulting DNA was spread, stained, and shadow-cast as described in Materials and Methods. (A) x18,400; (B) x21,600; (C) x7,800; (D) x10,300. 297
298
SCHARNHORST ET AL.
J. VIROL.
4 E 0
c
2
0
50
20
30
40
50
length,p FIG. 5. Size distribution of circular DNA molecules. Micrographs were enlarged, and the DNA contour lengths were determined with a map measurer.
Porthetria dispar. A slight deviation of the lines of subunits from a line perpendicular to the longitudinal axis was also observed. We find no counterpart to the ringlike arrangement of subunits reported by Harrap. We favor his suggestion that they may be artifacts generated by superimposing front and back images, although we believe that the "back images" are the circular structures in the background of his micrograph. The nucleocapsids frequently bend, distorting their envelopes into V-shaped or, more usually, ellipsoidal structures. Alternatively, the envelopes may collapse into ellipsoidal shape, bending the nucleocapsids. These distorted virions also appeared in the preparations of Gregory et al. (4) but were not as abundant. It seems certain that this bending is an artifact of preparation, since it is not seen in intact polyhedra or in virions from freshly dissolved polyhedra (4). Two lines of evidence support the idea that the viral DNA is supercoiled. First, two bands of DNA are observed in isopycnic CsCl gradients in the presence of ethidium bromide. The separation between these two bands (0.04 g/ml) is what is observed in other systems with both superhelical and open circular DNA (9). The dense band disappears upon shearing the DNA, consistent with its identification as supercoiled DNA. Second, highly supercoiled DNA is observed in electron micrographs. Of course, the finding of supercoiled DNA implies that it is circular, and we also observe open circles by electron microscopy. Formation of a superhelix may be the first
order of compacting necessary to fit the large viral DNA into the small volume of the nucleocapsid sheath. A second order of packing is suggested by our electron micrograph showing two loops of a relaxed circle extending from one end of a nucleocapsid. It appears that the supercoiled DNA is folded over on itself at least once. This idea was previously proposed by Shvedchikova and Tarasevich (11) to explain their observations of abrupt thinning of DNA emerging from NPVs of the silkworm B. mori. We observe similar "despiralization" in heavily shadowed DNA preparations (data not shown). The size of the DNA of the bollworm NPV is still difficult to judge. We have observed apparently closed circles, including supercoils, ranging in size from 15 to 45 ,m. Some variation is to be expected, due to inconsistencies in stretching and shrinking during the preparation of grids for electron microscopy (10). A threefold variation, however, is outside the range of this experimental error. We can only conclude that we have sampled DNA molecules varying considerably in size. The distribution of size classes of DNA circles shows a peak at 20 to 25 ,u m, with a more or less continuous distribution of larger molecules ranging up to 40 to 45 ,um in length. We believe that the relatively small sample size and the lack of a marker DNA preclude firm conclusions about a possible bimodal distribution of DNA circles. It would seem, however, that the real genome size is approximately 20 to 25 ,um, and the larger molecules contain partial duplications of this fundamental unit. Assuming that one base pair contributes 0.34 nm to the length of the DNA, we estimate that 20 to 25 ,um corresponds to 40 x 106 to 50 x 106 daltons. This is in accord with our unpublished observations on the reassociation kinetics of NPV DNA, which suggest a genetic complexity of approximately 50 x 106 daltons. Finally, the question arises about the ability of the nucleocapsids to contain DNA of greater than one genome length. We calculate that the volume of the average cylindrical NPV nucleocapsid is greater than that of the icosohedral herpes simplex I nucleocapsid. Since the latter holds a DNA with a molecular weight of 100 x 106 (3), the NPV nucleocapsid would appear to be spacious enough to hold even the largest DNA molecules we observe. ACKNOWLEDGMENTS We thank Lorraine Hammer for expert technical assistance, H. Stone and R. Bussell for helpful suggestions on the manuscript, and the Sandoz-Wander Co., Homestead, Fla., for gifts of virus. This work was suported by Public Health Service grant GM 22127 from the National Institute of General Medical
0:X~~~~~~~¾ ; * v(
*
- *-i @ l s;N;
I.: *-6 ~w@eX-e;B*Sfi||; K| A
pt : Of0t-B:
A FIG. 6. Highly twisted viral DNA molecules. Preparation x10,700; (B) x10,500; (C) x20,400. 299
was described in Materials and Methods. (A)
300
SCHARNHORST ET AL.
J. VIROL.
4?
.6.
e
FIG. 7. DNA escaping from a virion. Preparation was described in Materials and Methods. x28,800. Sciences and by a grant from the General Research Fund of the University of Kansas. LITERATURE CITED 1. Beaton, C. D., and B. K. Filshie. 1976. Comparative ultrastructural studies of insect granulosis and nuclear polyhedrosis viruses. J. Gen. Virol. 31:151-161. 2. Davis, R. W., M. Simon, and N. Davidson. 1971. Electron microscope heteroduplex methods for mapping regions of base homology in nucleic acids. Methods Enzymol. 21D:413-428. 3. Frenkel, N., and B. Roizman. 1972. Separation of the herpes-virus deoxyribonucleic acid duplex into unique fragments and intact strand on sedimentation in alkaline gradients. J. Virol. 10:565-572. 4. Gregory, B. G., C. M. Ignoffo, and M. Shapiro. 1969.
Nucleopolyhedrosis of Heliothis: morphological de-
scription of inclusion bodies and virions. J. Invertebr. Pathol. 14:186-193. 5. Harrap, K. A. 1972. The structure of nuclear polyhedrosis viruses. Virology 50:114-123. 6. Kleinschmidt, A. K. 1968. Monolayer techniques in electron microscopy of nucleic acid molecules. Methods Enzymol. 12B:361-377.
7. Kok, I. P., A. V. Chirtynhova-Ryndich, and A. P. Gudz-Gorban. 1972. Macromolecular structure of DNA of the silkworm nuclear polyhedrosis virus. Mol. Biol. (USSR) 6:323-331. 8. Onodera, K., T. Komano, M. Himeno, and F. Sakai. 1965. The nucleic acid of nuclear-polyhedrosis virus of the silkworm. J. Mol. Biol. 13:532-539. 9. Radloff, R., W. Bauer, and J. Vinograd. 1967. A dyebuoyant density method for the detection and isolation of closed circular duplex DNA: the closed circular DNA in HeLa cells. Biochemistry 57:1514-1521. 10. Scott, H. A., S. Y. Young III, and J. A. McMasters. 1971. Isolation and some properties of components of nuclear polyhedra from the cabbage looper, Trichoplusia ni. J. Invertebr. Pathol. 18:117-182. 11. Shvedchikova, N. G., and L. M. Tarasevich. 1971. Electron Microscope investigation of granulosis viruses of Dendrolimus sibiricus and Agrotis segetum. J. Invertebr. Pathol. 18:25-32. 12. Summers, M. D., and D. L. Anderson. 1973. Characterization of nuclear polyhedrosis virus DNAs. Virology 12:1336-1346. 13. Van der Geest, L. P. S. 1968. A method for the purification of polyhedra. J. Invertebr. Pathol. 11:502.
