JOURNAL OF STRUCTURAL BIOLOGY 124, 70–76 (1998)
ARTICLE NO. SB984055
Structure of the Human Cytomegalovirus B Capsid by Electron
Cryomicroscopy and Image Reconstruction
S. J. Butcher,*,1 J. Aitken,† J. Mitchell,* B. Gowen,‡ and D. J. Dargan*
*MRC Virology Unit and †Institute of Biomedical and Life Sciences, Division of Virology, University of Glasgow, Church Street,
Glasgow G11 5JR, United Kingdom; and ‡Department of Biochemistry, Imperial College of Science, Technology, and Medicine,
London SW7 2AZ, United Kingdom
Received June 22, 1998, and in revised form October 23, 1998
The three-dimensional structure of B capsids of
the b-herpesvirus human cytomegalovirus (HCMV)
was investigated at a resolution of 3.5 nm from
electron cryomicrographs by image processing and
compared with the structure obtained for the a-her-
pesvirus herpes simplex virus type 1 (HSV-1). The
main architectural features of the HSV-1 and HCMV
capsids are similar: the T 5 16 icosahedral lattice
consists of 162 capsomers, composed of two distinct
morphological units, 12 pentamers and 150 hexa-
mers, with triplex structures linking adjacent cap-
somers at positions of local threefold symmetry. The
main differences in the HSV-1 and HCMV capsids
are found in the diameter of the capsids (125 and 130
nm, respectively); the hexamer spacing and relative
tilt (center-to-center hexon spacing at outer edge,
17.9 and 15.8 nm, respectively); the morphology of
the tips of the hexons (similar in length but 33%
thinner in HCMV); and the average diameter of the
scaffold (44 and 76 nm, respectively). By analogy
with HSV-1, the mass on the HCMV hexon tip is
attributed to the smallest capsid protein (HCMV
gene UL48/49). The differences in capsid structure
are discussed in relation to the ability of the HCMV
structure to package a genome some 60% larger than
that of HSV-1. r 1998 Academic Press
Key Words: human cytomegalovirus; three-dimen-
sional structure; electron cryomicroscopy.
INTRODUCTION
Herpesviridae is a family of large and complex
viruses ubiquitous in nature and infecting diverse
animal species. The family is classified into a-, b-,
and g-herpesvirus subfamilies on the basis of shared
biological properties (i.e., host range, replication
1 Current address: Institute of Biotechnology and Department
of Biosciences, Division of Genetics, Biocenter 2, FIN-00014
University of Helsinki, Finland.
1047-8477/98 $25.00
Copyright r 1998 by Academic Press
All rights of reproduction in any form reserved.
kinetics, and ability to spread in culture) and genetic
relatedness. Herpesviruses genomes vary consider-
ably in size (125–245 kb), G 1 C content (32–75%),
and sequence organization, but these characteristics
are independent of subfamily classification (Roiz-
man et al., 1992). DNA sequence and genetic analy-
sis of viruses from each subfamily have revealed a
set of common ‘‘core’’ genes that are conserved in the
viral DNA sequence. Core genes provide common
functions and encode proteins required for viral DNA
metabolism, DNA replication, and assembly of the
viral capsid structure and some virus envelope glyco-
proteins with a role in virus entry (Davison, 1993).
Other genes appear to be specifically conserved only
within subfamilies and must have a role in confer-
ring subfamily biological characteristics.
Herpesviruses share a common virion architec-
ture. The double-stranded DNA genome is contained
within an icosahedrally symmetric capsid, though it
is not yet clear how this large DNA molecule folds
itself into the liquid crystalline packaged form (Booy
et al., 1991). However, there is no convincing evi-
dence that the DNA is spooled around a protein core
in the shape of a torus as earlier reported (Furlong et
al., 1972). The nucleocapsid is surrounded by a
complex protein layer known as the tegument and
enveloped by a lipid-rich membrane decorated with
virus glycoprotein spikes.
The capsid shell architecture of the a-herpesvi-
ruses herpes simplex virus type 1 (HSV-1) (Booy et
al., 1988; Schrag et al., 1989; Zhou et al., 1994),
equine herpesvirus type 1 (EHV-1) (Baker et al.,
1989), and the more distantly related channel catfish
virus (CCV) (Booy et al., 1996) has been studied by
cryomicroscopy and image reconstruction. The cap-
sid is organized on a T 5 16 lattice with 150
hexamers (hexons) and 12 pentamers (pentons).
Intercapsomeric linkage is facilitated by triplex com-
plexes. The HSV-1 capsid shell proteins have been
localized by image reconstruction combined with
70
 STRUCTURE OF HUMAN CYTOMEGALOVIRUS
TABLE I
Location of the Four Major Components of the HSV Capsid Shell and Comparison to the HCMV Homologues
(Gibson, 1996)
HSV-1
Copy
Location in capsid number Gene Gene product
Hexons and pentons 960 UL19 VP5 (149 kDa)
Triplexes 640 UL18 VP23 (34 kDa)
Triplexes 320 UL38 VP19c (50 kDa)
Hexon tips 900 UL35 VP26 (12 kDa)
difference imaging (Trus et al., 1995; Zhou et al.,
1988b, 1995) and antibody labeling (Trus et al., 1992)
and are shown in Table I. Herpesvirus capsid assem-
bly and viral DNA packaging takes place in the
nucleus of infected cells and is dependent on the
virus scaffolding proteins [e.g., the product of HSV-1
genes UL26 and UL26.5: and HCMV genes UL80
and UL80.5 (Desai et al., 1994; Preston et al., 1992;
Tatman et al., 1994; Welch et al., 1991a,b)]. Three
types of capsid structure are generated: A, which are
devoid of any core structure; B, which contain a
scaffold structure but lack viral DNA; and C, which
are mature DNA-containing structures (Gibson and
Roizman, 1972). The B-capsid structures are the
most abundant in the infected cell nucleus and most
easily isolated.
This paper extends the investigation of capsid
shell architecture to the b-herpesvirus subfamily.
Using electron cryomicroscopy and image reconstruc-
tion, we have examined and compared the B-capsid
shell structures of two important human pathogens:
human cytomegalovirus (HCMV), the prototype vi-
rus of the b-herpesvirus subfamily, and HSV-1, the
prototype virus of the a-herpesvirus subfamily. The
HCMV capsid shell proteins are compared with
those of HSV-1 in Table I. The HSV-1 and HCMV
major capsid proteins (MCPs) are of similar molecu-
lar size and their respective genes share 25% nucleo-
tide and 26% amino acid homology (Chee et al.,
1989). The HSV-1 triplex structures are composed of
two proteins, VP23 (gene UL18) and VP19c (gene
UL38), in the ratio of 2:1. Similarly, the HCMV
triplex structures are composed of the minor capsid
protein-binding protein (mC-BP, gene UL46) and the
minor capsid protein (mCP, gene UL85), also in the
ratio of 2:1. However, while the HCMV mCP shares
22.5% amino acid homology with HSV-1 UL18, the
HCMV mC-BP protein is only two-thirds of the size
of its HSV-1 functional homologue and shares no
appreciable amino acid homology (Gibson et al.,
1996a). The small HSV-1 protein, VP26, located at
the tip of the hexons is not required for capsid shell
assembly (Tatman et al., 1994). HSV-1 VP26 and its
HCMV counterpart, the ‘‘smallest capsid protein’’
(SCP, reported as gene UL48.5, or UL48/49), have
71
HCMV
Gene Gene product
UL86 Major capsid protein (154 kDa)
UL85 Minor capsid protein (35 kDa)
UL46 Minor capsid protein-binding protein (33 kDa)
UL48/49 Smallest capsid protein (8.5 kDa)
only limited amino acid homology, but both are small
proteins with a basic pI (Baldick and Shenk, 1996;
Gibson et al., 1996b).
We report differences between the HCMV and
HSV-1 capsid shell structures that are relevant to
the observation that the HCMV capsid structure
accommodates a DNA molecule up to 60% larger (at
245 kb) (Cha et al., 1996; Chee et al., 1990; Dargan et
al., 1997) than that of the HSV-1 genome (152 kb)
(McGeoch et al., 1988).
MATERIALS AND METHODS
Cells
Human fetal foreskin fibroblast cells (HFFF-2; European Collec-
tion of Animal Cell Cultures, No. 86031405) were propagated in
Dulbecco’s modified Eagle’s medium containing 10% fetal bovine
serum (DMEM 10% FB).
Virus Propagation
The AD169 (Glasgow) stock of HCMV (Dargan et al., 1997) was
used throughout. Virus-infected cell preparations were produced
by infecting HFFF-2 cell monolayers (4 3 106 cells) with HCMV at
a multiplicity of infection of 0.01 PFU/cell and, when extensive
cytopathic effect was evident, seeding virus-infected cells (,105
cells) onto subconfluent monolayers of HFFF-2 cells in roller
bottles (,2 3 107 cells/bottle) and incubating at 37°C. At 24 h
postinfection (PI) the culture medium was replaced with fresh
DMEM 10% FB and the infected cultures were incubated again at
37°C. Cytopathic effect was usually apparent by the fifth day PI
and was extensive by the ninth day. The infected cells were
detached from the surface of the roller bottle by gentle shaking
and then pelleted by centrifugation at 1000g.
Capsid Preparation
Nuclei were isolated from the infected cell pellets by washing
the cells with NTE buffer (0.5 M NaCl, 20 mM Tris, pH 7.4, 1 mM
EDTA) containing 1% Nonidet P-40. (NP40). The resulting nuclei
were pooled and stored at 270°C until required. The nuclei were
resuspended in NTE containing 1% NP40 and subjected to three
cycles of freeze–thawing. This suspension was then probe soni-
cated to break residual nuclei open and the cellular debris
pelleted at 1000g. The virus particles present in the supernatant
were pelleted through a 40% (w/v in NTE) sucrose cushion
(Sorvall TST41, 25 000 rpm, 1 h, 4°C), resuspended in 1% NP40 in
NTE, and banded on a 10–40% sucrose gradient in NTE (Sorvall
TST41, 40 000 rpm, 20 min, 4°C). The B-capsid band was har-
vested, pelleted (Sorvall TST41, 25 000 rpm, 1 h, 4°C), and
resuspended in 0.15 M NaCl, 20 mM Tris, pH 7.4, 1 mM EDTA.
HSV capsids were purified from baby hamster kidney cells
72 BUTCHER ET AL.

infected for 16 h at 37°C with 5 PFU HSV-1 (strain 17) per cell, as
described previously (Zhou et al., 1994).

Electron Cryomicroscopy
Electron cryomicroscopy allows us to examine samples under
native, hydrated conditions free from some of the artifacts that
are commonly associated with negatively stained specimens, as
well as reveal internal details. Freshly prepared capsids were
vitrified on holey carbon coated grids by plunge freezing into
liquid ethane (Adrian et al., 1984; Fukami and Adachi, 1965).
Samples were held at less than 2170°C in an Oxford CT3500
cryostage and observed using low-dose techniques in a Jeol
JEM-1200 EX II electron microscope (120 kV, tungsten filament)
fitted with an Oxford TAC100 twin-blade anticontaminator. Focal
pairs were recorded on Kodak SO163 film using a dose of 10–15
e/Å2 at a nominal magnification of 25 000 or 30 000 at defocus
values between 3 and 4 µm. FIG. 1. Electron cryomicrograph of B capsids from HCMV.

Image Processing than those for HSV1 (see Results). The data from the Jeol
Selected micrographs were scanned on a Zeiss-SCAI scanner micrographs were then scaled accordingly. In addition the scaling
using a step size of 14 µm per pixel, corresponding to 0.58 nm per of the HSV-1 reconstruction was checked against that of a 1.9-nm
pixel (14/24 100) and 0.48 nm per pixel (14/29 000) on the speci- HSV-1 reconstruction kindly provided by Dr. Hong Zhou (Zhou et
men. Image quality was assessed and the defocus estimated by al., 1995).
evaluating the position of the first zero in the sum of power
spectra from the particles of each micrograph. In total six RESULTS
micrographs were scanned from which 144 particles were se-
lected, boxed, floated, and normalized using SPIDER (Frank et al., Electron Cryomicroscopy and Image Reconstruction
1981). Orientations and centers were determined and refined
We have studied the structure of the HCMV capsid
using the common lines procedure (Crowther, 1971; Fuller et al.,
1996). The polar Fourier transform method was used to check the by electron cryomicroscopy and image processing.
accuracy of the model, further refinement, and orientation searches Figure 1 shows part of an electron micrograph of
for new data (Baker and Cheng, 1996). Three-dimensional recon- vitrified HCMV B capsids. The capsids have a diam-
structions were calculated using cylindrical expansion methods eter of approximately 130 nm. This can also be seen
(Crowther, 1971). Only particles with icosahedral phase residuals
below 75° at 3.5-nm resolution were included in the reconstructions.
from the radial density profile (Fig. 2). The capso-
The map of HCMV was calculated to 3.5-nm resolution from a meres covering the surface of each capsid are clearly
subset of 59 particles from four micrographs (taken at 3 µm under evident as projections at the edge of the particle and
focus). All the eigenvalues were greater than 10 and 97% were as regularly spaced dots across the surface (,15 nm
greater than 100, showing that reciprocal space was adequately in diameter). In addition the scaffold can be seen as a
sampled. Independent reconstructions calculated from subsets of
the data were consistent with all the features described below. For
dense area, roughly in the center of each particle.
comparison, a micrograph of HSV-1 B capsids (taken at 3 µm The HCMV capsids were difficult to fully dissoci-
under focus) was also processed in a similar fashion. Here, 38 of 43 ate from the nuclear matrix without breakage, and
particles were included in the reconstruction that was also could not be stored at 270, 220, or 4°C without
calculated to 3.5-nm resolution. For this reconstruction, all of the further breakage occurring. Consequently, only rela-
eigenvalues were greater than 1 and 87% were greater than 100,
reflecting the lower number of particles included in the reconstruc- tively few particles from each field could be used
tion. Visualization and rendering of the reconstructions were done (typically 30 particles were processed per micro-
in SPIDER. All of the programs were run on a Digital Alpha 255/233
AXP running Digital Unix OSF1 v3.2, an SGI Indigo Elan 4000
running IRIX 5.2, and a Digital Alpha 2100 running Open VMS 7.0.
To address the scaling of the micrographs, additional data were
collected on a CM200 FEG microscope at 200 kV, using prepara-
tions mixed with tomato bushy stunt virus [TBSV, 38.5 nm in
diameter in the presence of 1 mM EDTA, pH 7.4 (Robinson and
Harrison, 1982)] to act as an internal standard. These data (taken
at defocus values of 5605 and 6120 nm for HSV-1 and 5200, 5605,
and 5610 nm for HCMV, at a magnification of 319 400, and
scanned with a 7-µm step size) could be corrected for the micro-
scope contrast transfer function due to the presence of several
minima in the power spectra, unlike the original data collected on
the Jeol 1200EX with a tungsten filament. For each micrograph,
at least 90 TBSV particles were aligned and a radial profile made
of the average to check the magnification. The HSV1 data (134
particles) and the HCMV data (70 particles) were averaged FIG. 2. Graph showing the radial density distribution of B
separately and gave rise to radial profiles similar to those shown capsids from HCMV (solid line) and HSV-1 (broken line) calcu-
in Fig. 2, where the shell peaks for HCMV are at a larger radius lated from the reconstructions shown in Fig. 3.
 STRUCTURE OF HUMAN CYTOMEGALOVIRUS 73

FIG. 3. Comparison of HCMV and HSV-1 reconstructions. Surface representations of HCMV (A) and HSV-1 (B) viewed down an
icosahedral twofold axis of symmetry contoured at one standard deviation above the mean density value. The inset shows the positions of
hexamers (white dots) and pentamers (green dots) within 1 facet of a T 5 16 lattice, in an orientation similar to that of the reconstructions.
Central 0.58-nm-thick sections through the HCMV (C) and the HSV-1 (D) reconstructions. Bar 5 50 nm.

graph), so data from several micrographs were com-
bined to yield the final reconstruction. There was
also a high rejection rate during the common lines
refinement as up to 60% of the particles per micro-
graph gave icosahedral phase residuals greater than
75°, even at 3.5-nm resolution, and were therefore
excluded from the reconstruction.
Three-Dimensional Structure
The major structural components of the HCMV B
capsid are 150 hexons, 12 pentons, and 320 triplexes
organized on a T 5 16 icosahedral lattice (Fig. 3A).
The relative positions of the capsomers (hexamers
and pentamers) within a facet are shown in the top
inset of Fig. 3.
Capsomers
The HCMV hexons and pentons form cylindrical
protrusions (Figs. 3A, 3C, 4) penetrated by an axial
channel constricted 7.4 nm from the base of the
capsomer. The hexons appear slightly skewed, so
that the axial channel is oval rather than round. The
74
FIG. 4. (A) Enlarged view of a HCMV1 hexamer from an
icosahedral 2f axis. (B) Enlarged view of a HCMV1 pentamer from
an icosahedral 5f axis. Adjacent triplexes can also be seen at the
base of each capsomer.
tip of each hexon subunit is topped by a rod-shaped
protrusion (Figs. 3A, 3C, 4A), so the hexon looks like
an open flower with six petals (Figs. 3A, 4A). In
contrast, the penton subunit tips are more pointed,
giving a triangular cross section (Figs. 3A, 3C, 4B).
Triplexes
The triplexes are the structures found at the local
threefold axes of symmetry, interacting with the
hexon and penton walls (Figs. 3A, 3C, 4). The
triplexes do not appear to be threefold symmetric
despite their position. Instead they vary depending
on the location within the asymmetric unit (Figs. 3A,
4). The most striking triplexes are those interacting
with the pentons, where the triplex forms a buttress
against the penton wall (Figs. 3A, 3C, 4B).
Floor of the Capsid
The floor of the capsid is seen as a peak of density
in the radial profiles (Fig. 2) at a radius of 50 nm in
HCMV. This floor is penetrated by holes at the base
of each capsomer and underneath the triplexes (Fig.
3C). The thinnest part of the shell occurs between
the triplexes.
Internal Details
The scaffold core is a very strong feature in the
HCMV micrographs (Fig. 1). However, it is not
evident in the reconstructions (Fig. 3C). Inside the
capsid, there is only a low signal from icosahedrally
ordered mass above the background noise level,
which varies between different HCMV reconstruc-
tions, suggesting that it is not significant (data not
shown). Although the density from the scaffold pro-
teins does not reveal an icosahedrally ordered struc-
ture (Fig. 3C), we can still estimate the extent of the
majority of the scaffold from the radial density plot
(Fig. 2). This gives an average scaffold core radius of
38 nm The dips seen at low radii (less than 5 nm) are
probably artifacts of the Fourier Bessel procedures,
as analysis of the aligned raw data did not show
similar fluctuations.
BUTCHER ET AL.
Comparison with HSV-1 B Capsids
The basic elements of the HSV-1 reconstruction
described here are similar to those reported by
others (Booy et al., 1994; Newcomb et al., 1993; Trus
et al., 1992, 1995; Zhou et al., 1998a, 1995) although
the structure was calculated independently.
The HCMV B capsid has a diameter larger than
that of HSV-1 B capsids. This is illustrated from the
radial density plots shown in Fig. 2. HCMV has an
average outer radius of 65 nm compared with 62.5
nm for HSV-1. The internal radius of the capsid was
estimated using the peak from the radial density
plot (Fig. 2) corresponding to the floor of the shell in
Figs. 3C and 3D (HCMV, 50 nm; HSV-1, 47.5 nm).
The average height of the capsomers is 15 nm in both
cases. In addition the maximum hexon diameter is
similar in both structures (13.7 nm, Figs. 3C, 3D).
However, a distinct difference is the relative tilt
between adjacent hexons. For instance, the angle
between the hexon on the 2f axis (9 o’clock, Figs. 3C, 3D)
and the adjacent hexon is 9.5° in HCMV but only 5° in
HSV-1. Consequently, the separation between the hex-
ons at the base, measured from the center of one hexon
to the center of the next, is 14.7 nm in both structures.
However, at the tip of the hexons, it has increased to
17.9 nm in HCMV and only 15.8 nm in HSV-1, directly
accounting for a change in capsid diameter.
The tips of the HSV-1 hexons that have been
identified as a single protein species, VP26 (Booy et
al., 1994; Trus et al., 1995; Zhou et al., 1995), are not
as prominent as those of HCMV at this resolution
(Figs. 3, 4A). Those on the HCMV hexons are similar
in length, but are 33% thinner (Figs. 3C, 3D); the
choice of contour level did not appreciably affect the
appearance or dimensions of these features. The
HSV-1 pentons lack VP26, giving the pentons a
distinctive top compared with the hexons. In HCMV,
the penton tips are also distinct from those of the
hexons (Fig. 4). This may reflect a difference in
composition between the hexons and the pentons
due to the SCP, or a conformational difference be-
tween the hexameric form and the pentameric form
of the MCP.
The final difference between the two capsids is the
distribution of the scaffold core, which can be seen
from the radially averaged profiles (Fig. 2). The
profile of the HSV-1 B capsid agrees well with that of
Zhou et al. (1998b). However, the average density falls
off at radii of 22 nm in HSV-1 and 38 nm in HCMV.
DISCUSSION
The overall architectural design of the herpesvirus
capsid has been conserved among different subfami-
lies. However, our studies have demonstrated that
limited changes can be made to the capsid structure
by virus-specific differences in the component proteins.
 STRUCTURE OF HUMAN CYTOMEGALOVIRUS
The HCMV capsid is made up of hexamers and
pentamers arranged on a T 5 16 lattice. By analogy
with HSV-1, we expect that the major capsid protein
(MCP) forms the major part of these structures and
that the smallest capsid protein (SCP) forms the tips
of the hexons. The minor capsid protein (mCP) and
the mCP-binding protein (mC-BP), which associate
together in a ratio of 2:1, are analogous to the triplex
proteins of HSV-1 and probably form the equivalent
structures seen in the HCMV reconstruction at the
positions of local threefold symmetry (Fig. 4).
One of the motivating questions for performing
this study was: How does the capsid of HCMV
package a genome that is 60% larger than that of
HSV-1? There is, as predicted from a comparative
study between HSV-1 and channel catfish virus
(Booy et al., 1996), no change in the T number of the
capsid to accommodate the larger genome; however,
there is an increase in the capsid volume. The
internal radius of the capsid was estimated using the
peak from the radial density plot which corresponds
to the floor of the shell (HCMV, 50 nm; HSV-1, 47.5
nm) because the edge of the shell density is harder to
estimate. Assuming a sphere, this gives a volume
ratio of 1.17 for HCMV compared with HSV-1. This is
much smaller than the ratio of the genome lengths
(1.51) of the two virus strains used in this study
[HCMV AD169, 230 kb (Dargan et al., 1997); HSV-1
strain 17, 152 kb (McGeoch et al., 1988)]. Interest-
ingly, HCMV assembles this larger shell with pro-
teins that have molecular weights similar to those of
HSV-1 (Table I). In fact one of the HCMV triplex
proteins is appreciably smaller (mC-BP). It has been
shown for HSV-1 that the triplex proteins are pivotal
in in vitro capsid assembly, linking the capsomers
together (Booy et al., 1996; Newcomb et al., 1996).
Thus it is possible that the HCMV triplex proteins
are important in determining the greater center-to-
center spacing of the HCMV capsomers and their
relative tilt. The size of the scaffold core may also be
important for the assembly of a larger capsid.
The observed difference in the HCMV and HSV-1
B capsid volumes is insufficient by itself to account
for packaging of the larger HCMV genome. There are
at least three additional factors that have been
shown to be important for DNA packaging in other
viruses: DNA density, variable capacity of the capsid,
and expansion associated with DNA packaging. The
average interduplex spacing for most viruses, includ-
ing HSV-1 (Booy et al., 1991), has been reported to be
,2.6 nm, with L-A virus (at 4–4.5 nm) as an excep-
tion (Caston et al., 1997). HCMV could have more
densely packed DNA than HSV-1. In support of this,
studies of bacteriophage T7 have shown that interdu-
plex spacing increased as genome length fell (Cerri-
telli et al., 1997), but on the contrary, HSV-1 and the
75
smaller channel catfish virus were reported to have
similar genome packing densities (Booy et al., 1996).
Alternatively HSV-1 may underutilize its packaging
capacity like bacteriophage l, which can package
DNA 9% larger than its genome without affecting
viability (Weil et al., 1972). A third possibility is that
an expansion of the HCMV capsid is associated with
DNA packaging. B capsids from simian CMV have
been estimated to have a diameter 7% smaller than
that of A capsids when examined in thin sections of
infected cell nuclei (Lee et al., 1988). Studies of the
DNA-containing capsid are required to address these
issues in more detail. We have endeavored to isolate
such capsids from both virions and infected cells, but
invariably ended up with damaged, empty capsids.
This implies that HCMV is more fragile than HSV-1.
In conclusion, this preliminary study of HCMV
emphasizes the similarities between evolutionary
distinct herpesviruses. It also supports the idea that
although the three-dimensional structures of the
capsid proteins are similar at the gross level, we can
expect changes in the interactions between mono-
mers at the atomic level, giving rise to differences in
spacing between capsomers.
Dr. Hong Zhou is thanked for providing a reconstruction of
HSV-1 B capsids. Dr. Frazer Rixon is thanked for helpful discus-
sions. Dr. Stephen Fuller kindly provided his programs, as did
Felix deHaas and Ralph Heinkel. Ilaria Ferlenghi and Erica
Mancini are thanked for help with scanning of the micrographs.
This work was funded by the U.K. Medical Research Council.
REFERENCES
Adrian, M., Dubochet, J., Lepault, J., and McDowall, A. W. (1984)
Cryo-electron microscopy of viruses, Nature (London) 308, 32–36.
Baker, T. S., and Cheng, R. H. (1996) A model-based approach for
determining orientations of biological macromolecules imaged
by cryo-electron microscopy, J. Struct. Biol. 116, 120–130.
Baker, T. S., Newcomb, W. W., Booy, F. P., Brown, J. C., and Steven,
A. C. (1989) Three-dimensional reconstructions of ‘light’ and
‘intermediate’ capsids of equine herpes virus, Electron Microsc.
Soc. Am. Proc. 47, 822–823.
Baldick, C. J., and Shenk, T. (1996) Proteins associated with
purified human cytomegalovirus particles, J. Virol. 70, 6097–6105.
Booy, F. B., Newcomb, W. W., Brown, J. C., and Steven, A. C.(1988)
Herpesvirus nucleocapsids imaged in the frozen-hydrated state,
Electron Microsc. Soc. Am. Proc. 46, 164–165.
Booy, F. P., Newcomb, W. W., Trus, B. L., Brown, J. C., Baker, T. S.,
and Steven, A. C. (1991) Liquid-crystalline, phage-like packing
of encapsidated DNA in herpes simplex virus, Cell 64, 1007–1015.
Booy, F. P., Trus, B. L., Davison, A. J., and Steven, A. C. (1996) The
capsid architecture of channel catfish virus, an evolutionary
distant herpesvirus, is largely conserved in the absence of
discernible sequence homology with herpes simplex virus,
Virology 215, 134–141.
Booy, F. P., Trus, B. L., Newcomb, W. W., Brown, J. C., Conway,
J. F., and Steven, A. C. (1994) Finding a needle in a haystack:
Detection of a small protein (the 12-kDa VP26) in a large
complex (the 200-MDa capsid of herpes simplex virus), Proc.
Natl. Acad. Sci. USA 91, 5652–5656.
Caston, J. R., Trus, B. L., Booy, F. P., Wickner, R. B., Wall, J. S.,
76 BUTCHER ET AL.
and Steven, A. C. (1997) Structure of L-A virus: A specialized
compartment for the transcription and replication of double-
stranded RNA, J. Cell Biol. 138, 975–985.
Cerritelli, M. E., Cheng, N., Rosenberg, A. H., McPherson, C. E.,
Booy, F. P., and Steven, A. C. (1997) Encapsidated conformation
of bacteriophage T7 DNA, Cell 91, 271–280.
Cha, T. A., Tom, E., Kemble, G. W., Duke, G. M., Mocarski, E. S.,
and Spaete, R. R. (1996) Human cytomegalovirus clinical
isolates carry at least 19 genes not found in laboratory strains,
J. Virol. 70, 78–83.
Chee, M., Rudolph, S.-A., Plachter, B., Barrell, B., and Jahn, G.
(1989) Identification of the major capsid protein gene of human
cytomegalovirus, J. Virol. 63, 1345–1353.
Chee, M. S., Bankier, A. T., Beck, S., Bohni, R., Brown, C. M.,
Cerny, R., Horsnell, T., Hutchison, C. A., Kouzarides, T., Marti-
gnetti, J. A., Preddie, E., Satchwell, S. C., Tomlinson, P.,
Weston, K. M., and Barrell, B. G. (1990) Analysis of the
protein-coding content of the sequence of human cytomegalovi-
rus strain AD169, Curr. Top. Microbiol. Immunol. 154, 125–169.
Crowther, R. A. (1971) Procedures for three-dimensional recon-
struction of spherical viruses by Fourier synthesis from electron
micrographs, Phil. Trans. R. Soc. London B 261, 221–230.
Dargan, D. J., Jamieson, F. E., Maclean, J., Dolan, A., Addison, C.,
and McGeoch, D. J. (1997) The published DNA sequence of
human cytomegalovirus strain AD169 lacks 929 base pairs
affecting genes UL42 and UL43, J. Virol. 71, 9833–9836.
Davison, A. J. (1993) Herpesvirus genes, Rev. Med. Virol. 3,
237–244.
Desai, P., Watkins, S. C., and Person, S. (1994) The size and
symmetry of B capsids of herpes simplex virus type 1 are
determined by the gene products of the UL26 open reading
frame, J. Virol. 68, 5365–5374.
Frank, J., Shimkin, B., and Dowse, H. (1981) SPIDER: A modular
software system for electron image processing, Ultramicroscopy
6, 343–358.
Fukami, A., and Adachi, K. (1965) A new method of preparation of
a self-perforated microplastic grid and its application, J. Elec-
tron Microsc. (Japan) 14, 112–118.
Fuller, S. D., Butcher, S. J., Cheng, R. H., and Baker, T. S. (1996)
Three-dimensional reconstruction of icosahedral particles: The
uncommon line, J. Struct. Biol. 116, 48–55.
Furlong, D., Swift, H., and Roizman, B. (1972) Arrangement of
herpesvirus deoxyribonucleic acid in the core, J. Virol. 10,
1071–1074.
Gibson, W. (1996) Structure and assembly of the virion, Intervirol-
ogy 39, 389–400.
Gibson, W., Baxter, M. K., and Clopper, K. S. (1996a) Cytomegalo-
virus ‘‘missing’’ capsid protein identified as heat aggregable product
of human cytomegalovirus UL46, J. Virol. 70, 7454–7461.
Gibson, W., Clopper, K. S., Britt, W. J., and Baxter, M. J. (1996b)
Human cytomegalovirus (HCMV) smallest capsid protein iden-
tified as product of short open reading frame located between
HCMV UL48 and UL49, J. Virol. 70, 5680–5683.
Gibson, W., and Roizman, B. (1972) Proteins specified by herpes
simplex virus: VIII. Characterization and composition of mul-
tiple capsid forms of subtypes 1 and 2, J. Virol. 10, 1044–1052.
Lee, J. Y., Irmiere, A., and Gibson, W. (1988) Primate cytomegalo-
virus assembly: Evidence that DNA packaging occurs subse-
quent to B capsid assembly, Virology 167, 87–96.
McGeoch, D. J., Dalrymple, M. A., Davison, A. J., Dolan, A.,
Frame, M. C., McNab, D., Perry, L. J., Scott, J. E., and Taylor, P.
(1988) The complete DNA sequence of the long unique region in
the genome of herpes simplex virus type 1, J. Gen Virol. 69,
1531–1574.
Newcomb, W. W., Homa, F. L., Thomsen, D. R., Booy, F. P., Trus,
B. L., Steven, A. C., Spencer, J. V., and Brown, J. C. (1996)
Assembly of the herpes simplex virus capsid: Characterization
of intermediates observed during cell-free capsid formation, J.
Mol. Biol. 263, 432–446.
Newcomb, W. W., Trus, B. L., Booy, F. P., Steven, A. C., Wall, J. S.,
and Brown, J. C. (1993) Structure of the herpes simplex virus
capsid: Molecular composition of the pentons and the triplexes,
J. Mol. Biol. 232, 499–511.
Preston, V. G., Rixon, F. J., McDougal, I. M., McGregor, M., and
M. F., A. K. (1992) Processing of the herpes simplex virus
assembly protein ICP35 near its carboxy terminal end requires
the product of the whole UL26 reading frame, Virology 186,
87–98.
Robinson, I. K., and Harrison, S. C. (1982) Structure of the
expanded state of tomato bushy stunt virus, Nature (London)
297, 563–568.
Roizman, B., Desrosiers, R. C., Fleckenstein, B., Lopez, C.,
Minson, A. C., and Studdert, M. J. (1992) The family herpesviri-
dae: An update. The herpesvirus study group of the interna-
tional committee on taxonomy of viruses, Arch. Virol. 123,
425–449.
Schrag, J. D., Prasad, B. V. V., Rixon, F. J., and Chiu, W. (1989)
Three-dimensional structure of th eHSV1 nucleocapsid, Cell 56,
651–660.
Tatman, J. D., Preston, V. G., Nicholson, P., Elliott, R. M., and
Rixon, F. J. (1994) Assembly of herpes simplex virus capsids
using a panel of recombinant baculoviruses, J. Gen. Virol. 75,
1101–1113.
Trus, B. L., Homa, F. L., Booy, F. P., Newcomb, W. W., Thomsen,
D. R., Cheng, N., Brown, J. C., and Steven, A. C. (1995) Herpes
simplex virus capsids assembled in insect cells infected with
recombinant baculoviruses: Structural authenticity and localiza-
tion of VP26, J. Virol. 69, 7362–7366.
Trus, B. L., Newcomb, W. W., Booy, F. P., Brown, J. C., and Steven,
A. C. (1992) Distinct monoclonal antibodies separately label the
hexons or the pentons of herpes simplex virus capsid, Proc.
Natl. Acad. Sci. USA 89, 11508–11512.
Weil, J., Cunningham, R., Martin, R., Mitchell, E., and Bolling, B.
(1972) Characteristics of lambda p4, a lambda derivative con-
taining 9 per cent excess DNA, Virology 50, 373–380.
Welch, A. R., McNally, L. M., and Gibson, W. (1991a) Cytomegalo-
virus assembly protein nested gene family: Four 38-coterminal
transcripts encode four in frame-overlapping proteins, J. Virol.
65, 4091–4100.
Welch, A. R., Woods, A. S., McNally, L. M., Cotter, R. J., and
Gibson, W. (1991b) A herpesvirus maturational protease assem-
blin: Identification of its gene, putative active site domain, and
cleavage site, Proc. Natl. Acad. Sci. USA 88, 10792–10796.
Zhou, Z. H., Chiu, W., Haskell, K., Spears, H., Jakana, J., Rixon, F. J.,
and Scott, L. R. (1998a) Refinement of herpesvirus B capsid
structure on parallel supercomputers, Biophys. J. 74, 576–588.
Zhou, Z. H., He, J., Jakana, J., Tatman, J. D., Rixon, F. J., and
Chiu, W. (1995) Assembly of VP26 in herpes simplex virus-1
inferred from structures of wild-type and recombinant capsids,
Nature Struct Biol. 2, 1026–1030.
Zhou, Z. H., Macnab, S. J., Jakana, J., Scott, L. R., Chiu, W., and
Rixon, F. J. (1998b) Identification of the sites of interaction
between the scaffold and outer shell in herpes simplex virus-1
capsids by difference electron imaging, Proc. Natl. Acad. Sci.
USA 95, 2778–2783.
Zhou, Z. H., Prasad, B. V., Jakana, J., Rixon, F. J., and Chiu, W.
(1994) Protein subunit structures in the herpes simplex virus
A-capsid determined from 400 kV spot-scan electron cryomicros-
copy, J. Mol. Biol. 242, 456–469.