Crystal Structure of the Human Astrovirus Capsid Protein

Abstract

Importance:
Human astrovirus (HAstV) is a leading cause of viral diarrhea in infants and young children worldwide. As a non-enveloped virus, HAstV exhibits an intriguing feature in that its maturation requires extensive proteolytic processing of the astrovirus capsid protein (CP) both inside and outside of the host cell. Mature HAstV contains three predominant protein species, but the mechanism for the acquired infectivity upon maturation is unclear. We have solved the crystal structure of VP90(71-415) of the human astrovirus serotype 8. VP90(71-415) encompasses the conserved N-terminal domain of the viral CP. Fitting the VP90(71-415) structure into the cryo-EM maps of the HAstV produced an atomic model for the T=3 icosahedral capsid. Our model of the HAstV capsid provides valuable insights into intermolecular interactions required for capsid assembly and trypsin-mediated proteolytic maturation. Such information has potential applications in the development of VLP vaccine as well as small molecule drugs targeting astrovirus assembly/maturation.

Full text

Crystal Structure of the Human Astrovirus Capsid Protein
Yukimatsu Toh,
a
Justin Harper,
a
Kelly A. Dryden,
b
Mark Yeager,
b
Carlos F. Arias,
c
Ernesto Méndez,
c
†Yizhi J. Tao
a
Department of BioSciences, Rice University, Houston, Texas, USA
a
; Department of Molecular Physiology and Biological Physics, University of Virginia School of Medicine,
Charlottesville, Virginia, USA
b
; Departamento de Genética del Desarrollo y Fisiología Molecular, Universidad Nacional Autonoma de México, Cuernavaca, Morelos, Mexico
c
ABSTRACT
Human astrovirus (HAstV) is a leading cause of viral diarrhea in infants and young children worldwide. HAstV is a nonenvel-
oped virus with a Tⴝ3 capsid and a positive-sense RNA genome. The capsid protein (CP) of HAstV is synthesized as a 90-kDa
precursor (VP90) that can be divided into three linear domains: a conserved N-terminal domain, a hypervariable domain, and an
acidic C-terminal domain. Maturation of HAstV requires proteolytic processing of the astrovirus CP both inside and outside the
host cell, resulting in the removal of the C-terminal domain and the breakdown of the rest of the CP into three predominant pro-
tein species with molecular masses of ⬃34, 27/29, and 25/26 kDa, respectively. We have now solved the crystal structure of
VP90
71–415
(amino acids [aa] 71 to 415 of VP90) of human astrovirus serotype 8 at a 2.15-Å resolution. VP90
71–415
encompasses
the conserved N-terminal domain of VP90 but lacks the hypervariable domain, which forms the capsid surface spikes. The struc-
ture of VP90
71–415
is comprised of two domains: an S domain, which adopts the typical jelly-roll ␤-barrel fold, and a P1 domain,
which forms a squashed ␤-barrel consisting of six antiparallel ␤-strands similar to what was observed in the hepatitis E virus
(HEV) capsid structure. Fitting of the VP90
71–415
structure into the cryo-electron microscopy (EM) maps of HAstV produced an
atomic model for a continuous, Tⴝ3 icosahedral capsid shell. Our pseudoatomic model of the human HAstV capsid shell pro-
vides valuable insights into intermolecular interactions required for capsid assembly and trypsin-mediated proteolytic matura-
tion needed for virus infectivity. Such information has potential applications in the development of a virus-like particle (VLP)
vaccine as well as small-molecule drugs targeting astrovirus assembly/maturation.
IMPORTANCE
Human astrovirus (HAstV) is a leading cause of viral diarrhea in infants and young children worldwide. As a nonenveloped vi-
rus, HAstV exhibits an intriguing feature in that its maturation requires extensive proteolytic processing of the astrovirus capsid
protein (CP) both inside and outside the host cell. Mature HAstV contains three predominant protein species, but the mecha-
nism for acquired infectivity upon maturation is unclear. We have solved the crystal structure of VP90
71–415
of human astrovirus
serotype 8. VP90
71–415
encompasses the conserved N-terminal domain of the viral CP. Fitting of the VP90
71–415
structure into the
cryo-EM maps of HAstV produced an atomic model for the Tⴝ3 icosahedral capsid. Our model of the HAstV capsid provides
valuable insights into intermolecular interactions required for capsid assembly and trypsin-mediated proteolytic maturation.
Such information has potential applications in the development of a VLP vaccine as well as small-molecule drugs targeting astro-
virus assembly/maturation.
Members of the Astroviridae family possess a nonsegmented,
positive-sense, single-stranded RNA (ssRNA) genome with
a nonenveloped icosahedral capsid (1). Astroviruses are organized
into two genera, Mamastrovirus and Avastrovirus, that infect
mammals and avian species, respectively. Astrovirus was first de-
tected in 1975 in fecal samples collected from infants, wherein
viral particles were found to display a star-like morphology by
negative-staining transmission electron microscopy (EM) (2,3).
Human astrovirus is considered one of the major causes of child-
hood viral gastroenteritis worldwide (4). Human astrovirus
(HAstV) can be further divided into eight major serotypes, with
HAstV-1 (human astrovirus serotype 1) being the predominant
clinical isolate (5). Transmission of human astrovirus occurs
through the ingestion of contaminated food or water, leading to
infection of gut epithelial cells via receptor-mediated endocytosis,
which ultimately culminates in the lytic release of viral progenies
(6,7).
The ⬃7-kb genomic RNA of human astrovirus is polyadenyl-
ated and contains three open reading frames (ORFs) (8,9). ORF1a
and the downstream overlapping ORF1b encode two nonstruc-
tural polyproteins, nsp1a and nsp1ab, that are proteolytically pro-
cessed into smaller proteins implicated in viral genome replica-
tion (8–13). At the 3=end of the genome, ORF2 encodes the capsid
protein (CP), which is translated from a subgenomic RNA, thus
allowing for the differential regulation of structural and nonstruc-
tural protein synthesis (14). The astrovirus CP, which is initially
synthesized as VP90, can be divided into three distinct domains: a
conserved amino terminus (amino acids [aa] 1 to 415 of VP90
[VP90
1–415
]), a hypervariable central region (VP90
416– 646
), and
an acidic C-terminal domain (VP90
647–782
)(15,16)(Fig. 1a). The
acidic C-terminal domain mediates a transient association be-
Received 13 April 2016 Accepted 20 July 2016
Accepted manuscript posted online 27 July 2016
Citation Toh Y, Harper J, Dryden KA, Yeager M, Arias CF, Méndez E, Tao YJ. 2016.
Crystal structure of the human astrovirus capsid protein. J Virol 90:9008 –9017.
doi:10.1128/JVI.00694-16.
Editor: T. S. Dermody, University of Pittsburgh School of Medicine
Address correspondence to Yizhi J. Tao, ytao@rice.edu.
† Deceased.
Copyright © 2016 Toh et al. This is an open-access article distributed under the
terms of the Creative Commons Attribution 4.0 International license.
crossmark
9008 jvi.asm.org October 2016 Volume 90 Number 20Journal of Virology

tween full-length VP90 and host membranous structures via an
endoplasmic reticulum (ER)-targeting motif, allowing for the co-
localization of capsid assembly and viral genome replication (17,
18). Upon assembly of the VP90 capsid lattice, the acidic C-termi-
nal domain is cleaved intracellularly by host caspases, leading to its
exclusion from the virion and the conversion of VP90 to VP70 (19,
20)(Fig. 1a). When overexpressed in eukaryotic hosts, the astro-
virus CP, in the form of either VP90 or VP70, was able to self-
assemble into virus-like particles (VLPs) (21–23).
After astrovirus particles are released from infected cells, fur-
ther proteolytic processing of the viral capsid by host extracellular
proteases is required for virus infectivity. In cell culture, the inclu-
sion of trypsin is essential for the successful propagation of human
astrovirus (24). Trypsin treatment produces a capsid composed of
three predominant protein species with molecular masses of ⬃34
kDa (VP34), 27/29 kDa (VP27/29), and 25/26 kDa (VP25/26) (14,
19,23,25,26). VP34 is derived from the conserved N-terminal
domain (VP90
1–415
), which comprises the capsid shell (i.e., the
continuous, spherical capsid), whereas both VP27/29 and
VP25/26 are from the hypervariable region with a different N ter-
minus, which forms the capsid surface spike (15,27)(Fig. 1a).
Within the conserved N-terminal domain, the polypeptide span-
ning residues 1 to 70 is an RNA coordination motif that is en-
riched in basic amino acids, likely structurally disordered, and
dispensable for particle assembly (21), a situation similar to that of
many small RNA plant viruses such as tomato bushy stunt virus
(TBSV) and turnip crinkle virus (28,29).
The astrovirus capsid is an external structural barrier that not
only encapsidates nucleic acids but also interacts with the host to
define cell tropism, mediate cell entry, and trigger the host im-
mune response (1). A 25-Å-resolution cryo-EM reconstruction of
an immature astrovirus capsid shows T⫽3 icosahedral symmetry
with a total of 90 spikes distributed along the 2-fold as well as the
pseudo-2-fold symmetry axes (30). In comparison, the recon-
struction of a mature virion shows an overall similar capsid topol-
ogy, but only 30 spikes are observed along icosahedral 2-fold sym-
metry axes (30). The crystal structure of the human astrovirus
capsid spike has also been determined as dimers (15). The overall
structure of each spike/projection domain has a unique three-
layered ␤-sandwich fold, with a core, six-stranded ␤-barrel struc-
ture that is also found in hepatitis E virus (HEV) capsid protru-
sions (31,32).
To further enhance our understanding of astrovirus assembly
and maturation, here we report the crystal structure of VP90
71–415
FIG 1 Astrovirus CP. (a) Proteolytic processing of astrovirus VP90. Astrovirus VP90 consists of a conserved region (white, blue, and yellow), a variable region
(red), and an acidic C-terminal region (gray). Stars represent trypsin digestion sites. (b) Secondary-structure assignment of VP90
71–415
.␣-Helices are indicated
by tubes, ␤-strands are indicated by arrows, loops are indicated by thick lines, and disordered regions are indicated by dotted lines. Stars highlight trypsin
digestion sites confirmed previously and in this study. (c) Crystal structure of VP90
71–415
. The molecule is colored with the S domain in blue and the P1 domain
in yellow. The disulfide bond C82-C254 is highlighted in orange. The equivalent parts of the HEV CP and the TBSV CP are shown on the right.
Structure of the Human Astrovirus Capsid Protein
October 2016 Volume 90 Number 20 jvi.asm.org 9009Journal of Virology

from HAstV-8 at a 2.15-Å resolution. VP90
71–415
, which covers
the conserved N-terminal region of VP70 except for the putative
RNA coordination motif (i.e., aa 1 to 70), was crystallized as
monomers with 1 molecule per crystallographic asymmetric unit
(CAU). The structure of VP90
71–415
shows two domains: an S
domain that adopts the typical jelly-roll ␤-barrel fold and a P1
domain that has the appearance of a squashed ␤-barrel consisting
of six antiparallel ␤-strands. A Dali search indicated that VP90
71–415
is a close structural homolog of the HEV CP in spite of the lack of
any detectable sequence similarity (31,32). By fitting the VP90
71–
415
crystal structure into the available cryo-EM maps (30), an
atomic model of the astrovirus capsid is derived, which highlights
important molecular interactions involved in the formation of
various types of capsomeres found in a T⫽3 icosahedral capsid.
The VP90
71–415
structure also provides insights into the trypsin-
mediated capsid maturation process and the accompanying struc-
tural changes that lead to enhanced viral infectivity.
MATERIALS AND METHODS
Subcloning, protein expression, and purification. The coding sequence
for HAstV-8 CP
71–415
(strain Yuc8) (GenBank accession number AF260508)
was cloned between the NcoI and HindIII sites of the vector pET28a
(Novagen) by using forward primer 5=-GGCGCCCATGGAAAAACAAG
GTGTCACAGGACCAAAACCT-3=and backward primer 5=-GCGCCAA
GCTTTTAGTGATGATGATGATGATGACCACCACCATGACCTAAA
CTAGGCTGATTCATC-3=.
A6⫻His tag and a GGG linker sequence were engineered at the C
terminus of the construct to facilitate protein purification. Recombinant
protein was expressed in Escherichia coli by using the Rosetta 2(DE3)
strain. Cells were first grown to an optical density at 600 nm (OD
600
)of
0.8 to 1.0 at 37°C and then induced with 1 mM IPTG (isopropyl-␤-D-1-
thiogalactopyranoside) for 19 h at 30°C. Cells were harvested and lysed by
using a sonicator (Branson 250 Sonifier) for 15 min at 4°C. The lysis buffer
consisted of 300 mM NaCl, 50 mM Tris (pH 7.5), 5 mM 2-mercaptoeth-
anol, 5 mM imidazole, 1 mM PMSF (phenylmethylsulfonyl fluoride), 0.5
␮g/ml pepstatin, and 0.5 ␮g/ml leupeptin. The lysate was clarified by
centrifugation at 12,000 ⫻gat 4°C for 45 min. The supernatant was
subjected to affinity purification using Ni-nitrilotriacetic acid (NTA)
resin (Thermo Scientific). Bound protein was eluted by using an elution
buffer containing 500 mM NaCl, 50 mM Tris (pH 7.5), 5 mM 2-mercap-
toethanol, and 300 mM imidazole. Recombinant protein was further pu-
rified by ion exchange chromatography using a HiTrap Q HP column (GE
Healthcare) and by size exclusion chromatography using a HiLoad Super-
dex 200 16/60 gel filtration column (GE Healthcare). The final protein was
at least 95% pure as judged by SDS-PAGE. The protein concentration was
determined by the A
280
reading from a NanoDrop 2000/2000c spectro-
photometer (Thermo Scientific). The molar extinction coefficient (ε)of
CP
71–415
was calculated to be 1.58 by the ExPASy ProtParam tool. Ap-
proximately 5 mg of purified protein could be obtained from 6 liters of cell
culture.
To obtain selenomethionine (SeMet)-labeled proteins, VP90
71–415
was overexpressed by using M9 minimal medium containing SeMet and a
mixture of six other amino acids to prevent methionine synthesis (33).
SeMet-labeled proteins were purified by using the same protocol as the
one described above.
Crystallization and structure determination. Purified HAstV-8
VP90
71–415
proteins, both native and SeMet-labeled forms, were concen-
trated to 10 mg/ml and subjected to crystallization screening. The best
crystals were obtained by the hanging-drop vapor diffusion method at
25°C by mixing 2 ␮l of the protein solution with an equal volume of
reservoir solution containing 0.2 M ammonium phosphate and 22%
polyethylene glycol 4000 (PEG 4000). For data collection, crystals were
briefly soaked in a cryoprotectant made from mother liquor supple-
mented with 23% (vol/vol) glycerol and flash-cooled in a nitrogen cryo-
stream. Diffraction data were collected at the Advanced Photon Source
(APS) (Argonne National Laboratory, Chicago, IL, USA) Life Sciences
Collaborative Access Team (LS-CAT) F line. For SeMet-labeled crystals, a
single-wavelength anomalous dispersion (SAD) data set of 720 frames was
collected at the peak wavelength for Se (0.97872 Å) by using a detector-
to-crystal distance of 200 mm, an exposure time of 2 s, and an oscillation
angle of 1°. All of the data were processed by using the HKL2000 program
(34).
The structure of VP90
71–415
was determined to a 2.15-Å resolution by
Se-SAD. The AutoSol program in PHENIX (35) located four out of the six
Se atoms in the asymmetric unit. Model building was carried out by using
the program AutoBuild in PHENIX and Coot (36). Structure refinement
was performed by using the maximum likelihood method with the phe-
nix.refine program from the PHENIX suite (35). The data statistics are
summarized in Table 1.
HAstV-8 VP90
71–415
structure fitted into cryo-EM structures. Pseu-
doatomic models for the astrovirus capsid were generated by fitting the
crystal structure of VP90
71–415
into the HAstV cryo-EM reconstruction
maps calculated to a 25-Å resolution (30). Three copies of VP90
71–41
were
manually fitted into a region of the cryo-EM maps corresponding to an
icosahedral asymmetric unit using the UCSF Chimera program (37). Fit-
ting was further improved by using the “Fit in Map” option in Chimera. A
correlation coefficient of 0.90 was given for the combined S, P1, and P2
domains. The entire T⫽3 capsid model was generated by applying icosa-
hedral symmetry. This model was used for studying CP capsomere inter-
actions.
Cartoon and surface representations were generated with the PyMOL
(http://www.pymol.org/) and UCSF Chimera programs, respectively.
Accession number(s). Atomic coordinates and structure factors have
been deposited in the RCSB Protein Data Bank (PDB) under accession
number 5IBV.
TABLE 1 X-ray data statistics
a
Parameter Value(s) for HAstV VP90
71–415
Data collection
Space group P2
1
Cell dimensions
a,b,c(Å) 52.0, 59.2, 56.2
␣,␤,␥(°) 90.0, 91.5, 90.0
Resolution (Å) 30.0–2.15
CC (1/2) 99.9 (70.6)
R
meas
10.7 (71.6)
I/␴17.9 (2.0)
Completeness (%) 99.8 (97.8)
Redundancy 4.0 (3.2)
Phasing
No. of Se sites 4
Figure of merit 0.47
Refinement
No. of reflections 35,248
R
work
/R
free
0.204/0.258
RMS deviations
Bond length (Å) 0.010
Bond angle (°) 1.372
Ramachandran value (%)
Most favored 308 (96.9)
Additionally allowed 10 (3.1)
Disallowed 0 (0)
a
Statistics in parentheses refer to the outer-resolution shell. CC (1/2), percentage of
correlation between intensities of random half-data sets; RMS, root mean square.
Toh et al.
9010 jvi.asm.org October 2016 Volume 90 Number 20Journal of Virology

RESULTS AND DISCUSSION
Structure of astrovirus VP90
71–415
.Three truncation mutants of
HAstV-8 VP90, including VP90
71–415
(38.1 kDa; pI 9.49), VP90
71–313
(26.8 kDa; pI 8.45), and VP90
71–283
(23.51 kDa; pI 8.45), were
cloned and expressed in E. coli (Fig. 1a). Residues 71 and 415 have
been mapped to roughly the beginning of the S domain and the
end of the P1 domain, respectively (15,22,23,27). Therefore,
VP90
71–415
was expected to contain both the S and P1 domains of
the astrovirus CP, but VP90
71–283
should contain only the S do-
main. VP90
71–313
was designed to mimic VP34, the longest pep-
tide fragment observed in mature astrovirus after trypsin activa-
tion (23). All three constructs were expressed as soluble proteins.
Furthermore, gel filtration chromatography using a HiLoad Su-
perdex 200 16/60 gel filtration column (S200) showed that the
three proteins were all eluted as a single peak with an apparent
molecular mass of ⬃30 kDa, indicating the formation of mono-
mers in solution. Therefore, the S or S-P1 domain alone appeared
to be insufficient to mediate the assembly of high-order oligom-
ers/capsomeres.
Among the three constructs, only VP90
71–415
produced single
crystals. The space group was determined to be P2
1
with aequal to
52.0 Å, bequal to 59.2 Å, cequal to 56.2 Å, and ␤equal to 91.5°.
The structure was determined to a 2.15-Å resolution by Se-SAD
(Fig. 1 and Table 1). There is one VP90
71–415
molecule in each
crystallographic asymmetric unit cell, consistent with VP90
71–415
being a monomer in solution. The final model, which was refined
to an R
work
value of 0.204 and an R
free
value of 0.258, contains 321
out of 345 residues in total. No density was observed for the initial
Met residue and the 6⫻His tag. Additional disordered regions
include residues 71 to 76, 332 to 334, 390 to 398, and 413 to 415
(Fig. 1b and c). An intramolecular disulfide bond is formed be-
tween C82 and C254 (Fig. 1c).
The structure of VP90
71–415
is organized into two domains
called the S domain (residues 71 to 256) and the P1 domain (res-
idues 257 to 415) (Fig. 1b and c). The S domain has the typical
jelly-roll ␤-barrel fold with eight antiparallel ␤-strands that is
broadly conserved among many viral capsid proteins. These eight
␤-strands, often designated by the letters B to I, form two twisted
␤-sheets called BIDG and CHEF (Fig. 1b). The surfaces of the
two ␤-sheets are decorated by a number of loops and also four
helices (i.e., ␣1 from the CD loop, ␣2 from the EF loop, and two
3
10
helices from the GH loop). The P1 domain has an antiparallel
␤-barrel structure composed of seven ␤-strands (i.e., ␤9to␤15)
and three ␣-helices (i.e., ␣3to␣5). The P1 domain is connected to
the S domain through an asparagine-rich linker loop (i.e., residues
257 to 267). A substantial amount of surface area (⬃2,400 Å
2
)is
buried between the S and P1 domains. This interaction, which is
mostly hydrophobic in nature, is mediated by ␤(C)2(C) and the
CD loop, EF loop, and GH loop from the S domain and the do-
main linker loop, ␤9, ␤12, ␤13, ␤15, ␣4, and ␣5 from the P1
domain. The C terminus of the P1 domain is located externally
near the S-P1 domain interface. In the astrovirus capsid, the C
terminus of P1 is expected to connect to the P2 domain, which
forms the dimeric spikes on the outer surface of the viral capsid
(15).
Structural comparison with other viral CPs. A structural ho-
molog search using the Dali server (38) showed that the S domain
of VP90
71–415
was best aligned to the jelly-roll domain of carna-
tion mottle virus (CMV) (Zscore of 17.8, with a value of ⬎2.0
being significant) (39), TBSV (Zscore ⫽16.0) (40), ryegrass mot-
tle virus (RMV) (Zscore ⫽15.8) (41), Orsay virus (OV) (Z
score ⫽15.7) (42), and HEV (Zscore ⫽15.6) (31,32). When the
P1 domain was used as the reference for the Dali search, the hit
with the highest Zscore was HEV (Zscore ⫽7.0). When both the
S and P1 domains of VP90
71–415
were used as the search model,
HEV came up with the highest Zscore, 27.0, which was followed by
CMV (Zscore ⫽16.0), TBSV (Zscore ⫽14.4), RMV (Zscore ⫽
14.0), and OV (Zscore ⫽13.8) (Fig. 1c). The finding that HEV
persistently showed up as a top structural homolog based on ei-
ther individual domains or the entire structure indicates that hu-
man astrovirus and HEV are evolutionarily related, consistent
with previous conclusions based on structural comparison using
the P2 domain alone (15,30).
Tⴝ3 HAstV-8 capsid models. Structures of both immature
human astrovirus (HAstV-8) (EMD-5414) and mature human
astrovirus (HAstV-1) (EMD-5413) have been established by
cryo-EM reconstruction to a 25-Å resolution (30). Therefore, the
structure of VP90
71–415
allowed us to build an atomic model of the
astrovirus capsid. Together with the crystal structure of the capsid
spike (15), we now have atomic coordinates for the entire astro-
virus VP70 protein except for the RNA coordination motif, which
is expected to be mostly structurally disordered.
Comparison of the structures of the immature and mature
astrovirus capsids shows a major difference in the stoichiometry of
the surface spike: while the immature astrovirus capsid shows 90
dimeric spikes along the icosahedral and quasi-2-fold symmetry
axes, the mature astrovirus capsid shows only 30 spikes along
2-fold symmetry axes (Fig. 2a)(30). The crystal structure of
VP90
71–415
fits well into the EM maps of the immature and mature
astrovirus capsids, producing nearly identical models. Due to the
lack of a meaningful difference, only the mature capsid model is
presented (Fig. 2a and b). Although the available EM map for the
mature astrovirus capsid is for HAstV-1, and our crystal structure
is for HAstV-8, the structural difference should not be a concern at
this resolution, as the sequences of HAstV-1 and HAstV-8 are 83%
identical in the conserved N-terminal region of the CP (i.e., aa 1 to
415). The outer diameter of our mature astrovirus model without
the surface spikes is ⬃350 Å. The S domain assembles into a con-
tinuous capsid shell, while the P1 domain forms trimeric clusters
on the capsid surface (Fig. 2c,e, and f). These P1 trimeric clusters
are in close contact across the icosahedral 2-fold symmetry axes,
but the P1 trimeric clusters related by pseudo-2-fold symmetry
axes do not interact, resulting in the breakdown of pseudo-6-fold
symmetry on 3-fold icosahedral symmetry axes (Fig. 2d and e).
Small depressions are observed at both 5-fold and 3-fold symme-
try axes.
Without the surface spike, the VP90
71–415
dimer buries the
smallest amount of surface areas among all types of capsomeres
found in the T⫽3 capsid. There are two types of dimers: one that
is relatively flat, sitting on 2-fold axes (i.e., C-C dimer), and an-
other that has an inwardly bent conformation, located on quasi-
2-fold axes (i.e., A-B dimer). Due to the different bending angles,
we observed dramatic differences in the gap distances between the
two P1 domains in the two types of dimers, with ⬃5 Å for the C-C
dimer and ⬃25 Å for the A-B dimer (Fig. 3a). Consequently, the
two P1 domains from the A-B dimer are completely segregated.
The A-B dimer interface buried a total surface area of only ⬃300
Å
2
, but nearly ⬃1,200 Å
2
of surface area is buried in the C-C
dimer. Previous studies of other T⫽3 viral capsids indicate that
Structure of the Human Astrovirus Capsid Protein
October 2016 Volume 90 Number 20 jvi.asm.org 9011Journal of Virology

the different bending angles observed in A-B and C-C dimers
could be maintained by different viral nucleic acid binding modes
and/or differentially ordered structural elements from the CP N-
terminal sequence at the icosahedral versus quasi-2-fold symme-
try axes (28). The secondary structural elements ␣1 and ␤1 from
the S domain are found at the interface of both dimers (Fig. 3a to
5). The C-C dimer interface has additional contacts made by the
loop connecting the S and P1 domains.
FIG 2 Human astrovirus capsid. (a) Cryo-EM structure of mature HAstV-1. The surface is colored by radial depth cue from blue to red. The colors blue, yellow, and
red roughly match the S, P1, and P2 domains of the astrovirus CP, respectively. (b) Astrovirus capsid model docked into the three-dimensional cryo-EM density map of
mature HAstV-1. (Left) Outside view; (right) central slab. For the capsid model, the S, P1, and P2 domains are colored according to the color key. (c) Atomic model of
the astrovirus capsid by EM docking. The S, P1, and P2 domains are colored according to the color key. (d) Astrovirus capsid with only the P1 and P2 domains. (e)
Astrovirus capsid with the S and P1 domains. (f) Mature astrovirus capsid with only the S domain. Two asymmetric units are highlighted by two triangles panels a and
c to f. Icosahedral symmetry axes (2 for 2-fold, 3 for 3-fold, 5 for 5-fold, and 2=for quasi-2-fold) are also highlighted where space is available.
FIG 3 Capsomeres from the T⫽3 mature astrovirus capsid model. (a) Dimers. C-C and A-B dimers are related by icosahedral and quasi-2-fold symmetry axes,
respectively. (b) Trimer. (c) Pentamer. (d) Hexamer. Two-, 3-, 5-, and 6-fold symmetry axes are represented by black lines and highlighted by an oval, triangle,
pentagon, and hexagon, respectively. The molecules in the top row are viewed from the side, while the molecules in the bottom row are viewed along the
symmetry axes. A reference molecule is colored with the S domain in blue and the P1 domain in yellow. Other symmetry-related molecules are each shown in a
different color, with the S domain and P1 domain from the same molecule shown in lighter and darker shades of the same colors, respectively.
Toh et al.
9012 jvi.asm.org October 2016 Volume 90 Number 20Journal of Virology

Our fitted model of the mature astrovirus capsid shows that the
VP90
71–415
trimer buries a substantial amount of surface (i.e.,
⬃2,000 Å
2
) between adjacent subunits (Fig. 3b and 4). The inter-
action is mediated largely by the following structural elements: (i)
helix ␣2 and the GH loop from the S domain, (ii) helix ␣5 and an
extended loop (residues 305 to 324) connecting ␣3 and ␤12 from
the P1 domain, and (iii) the loop connecting the S and P domains
(Fig. 4). In particular, the 3-fold symmetry axis is surrounded by
the GH loop from the S domain, helix ␣5, and the loop (residues
305 to 324) connecting ␣3 and ␤12 from the P1 domain (Fig. 4).
The astrovirus VP90
71–415
pentamers are maintained exclu-
sively by interactions mediated by the S domain (Fig. 3c and 4).
The 5-fold symmetry axis is surrounded by the ED, FG, and HI
loops. Additionally, helix ␣1 and the BIDG ␤-sheet from one sub-
unit make contact with helix ␣2 and the CHEF ␤-sheet from the
adjacent subunit in the same pentamer. The total buried surface
area between adjacent subunits is ⬃2,000 Å
2
in pentamers. Matsui
et al. reported previously that mutations at T227 resulted in the
disruption of capsid assembly in HAstV-1 (43). Our structural
model shows that T227 is from ␤7(H) located at the pentamer
interface (Fig. 4).
Close inspection of the VP90
71–415
hexamer shows that the S
domain makes interactions similar to those in the pentamer (Fig.
3d and 4). The major difference is observed in the P1 domain.
While the five P1 domains from a pentamer are completely iso-
lated from each other, the six P1 domains in a hexamer interact
with each other in pairs (e.g., molecule 1 interacts with molecule 2,
molecule 3 interacts with molecule 4, and molecule 5 interacts
with molecule 6). The interface between two interacting P1 do-
mains is mediated by helices ␣3 and ␣4. The total buried surface
areas are ⬃2,000 Å
2
for adjacent subunits with noninteracting P1
domains and ⬃2,100 Å
2
for adjacent subunits with interacting P1
domains. It is worthwhile to note that some structural clashes are
observed in both the pentamer and hexamer near the 5- and 6-fold
symmetry axes, suggesting that the three structured loops ED, FG,
and HI from the S domain may adopt a somewhat different con-
formation upon the assembly of a capsid.
Mapping of trypsin cleavage sites required for astrovirus
maturation. Proteolytic cleavage is a common activation mecha-
nism for both enveloped and nonenveloped viruses. A recurring
theme is that proteolytic cleavage either releases cell penetration
factors or triggers conformational changes in the capsid or cell
attachment proteins (44–47). Astrovirus infectivity is also depen-
dent on proteolysis-mediated maturation (24). Upon activation
by trypsin treatment, VP70 from the immature particle is con-
verted to several smaller polypeptide species, including VP34,
VP27, and VP25 (14,19,23,25,26).
With the crystal structure of VP90
71–415
solved, we are able to
map the terminal ends of VP34 and VP27, two of the three major
peptide fragments from mature astrovirus, in the context of the
capsid model. The N-terminal residue of HAstV-8 VP27 was pre-
viously determined to be Q394 by N-terminal sequencing (23),
and the structure of VP90
71–415
shows that R393, which is strictly
conserved in all human astrovirus serotypes (Fig. 6b), is located at
a structurally disordered surface loop (i.e.,
390
ASARQSNPV
398
)
facing the surface depression on 5-fold and quasi-6-fold symme-
try axes (Fig. 4 and 5) and is completely solvent exposed. In
HAstV-2, the 26-kDa protein was presumably generated by a sim-
ilar cleavage event at Arg394 at an equivalent position (26). The
390
ASARQSNPV
398
loop is connected at its C terminus to ␤15,
which is part of a four-stranded ␤-sheet (i.e., made of ␤13, ␤12,
␤15, and ␤9) that closely interacts with the S domain from the
opposing side. Therefore, in principle, VP27 could remain teth-
ered to the capsid through the ␤-sheet interaction mediated by
␤15. In comparison, VP25 starts its N terminus at residue 424,
which is beyond the P1 domain, and therefore would lose its grip
on the capsid after its cleavage (15,30).
VP34 has an intact N terminus, but its C terminus has not yet
been experimentally defined. The apparent molecular mass of 34
kDa suggests that its C terminus is likely to end at around residue
310. Examination of the protein sequence shows three trypsin-
susceptible sites around this area, including R299, R313, and
K347. Because R299 and K347 are located in the ␣3 and ␣4 helices,
respectively, the most likely cleavage site is R313, which is situated
FIG 4 VP90
71–415
molecule highlighting capsomere interactions and trypsin cleavage sites. (a) Side view. The viewing orientation is the same as that described
in the legend of Fig. 1c. (b) Top view at a 90° rotation from the view in panel a. The S domain is in blue, and the P1 domain is in yellow. Residues implicated in
trypsin cleavage are highlighted in red, with side chains shown in a stick-and-ball representation. Likely cleavage sites as discussed in the text are underlined.
Secondary-structure elements involved in capsomere interactions are pinpointed by ovals (2-fold), triangles (3-fold), pentagons (5-fold), and hexagons (6-fold).
Structure of the Human Astrovirus Capsid Protein
October 2016 Volume 90 Number 20 jvi.asm.org 9013Journal of Virology

in a flexible loop located on the surface of the P1 domain facing the
quasi-3-fold symmetry axis (Fig. 4 and 5). Multiple-sequence
alignment shows that R313 is strictly conserved in all astrovirus
serotypes (Fig. 6b). The calculated molecular mass of the polypep-
tide containing residues 1 to 313 is 33.7 kDa, which closely
matches that of VP34 (19,21).
Implications for astrovirus maturation. Previous work by
Méndez et al. indicated that VP34 is generated by the progressive
shortening of VP41 (containing residues 1 to 393, with a theoret-
ical molecular mass 42.9 kDa) through a number of intermediates,
including VP38.5 (23). Why is a cascade of trypsin cleavage events
required to activate virus infectivity? The sequential order of the
observed trypsin cleavages may imply that the final cleavage site
needed to activate virus infectivity is not accessible at the begin-
ning, but the initial cleavages could induce structural changes to
expose the R313 site. Indeed, our capsid model shows that R313 is
partially buried near the quasi-3-fold symmetry axes (Fig. 5d).
While we cannot rule out the possibility that the ␣3-␤12 loop
hosting R313 could undergo some local structural rearrangements
upon capsid formation, it is highly likely that R313 is not fully
revealed in the context of the capsid without the previous cuts by
trypsin.
Another interesting question is what happens to the polypep-
tide region from residue 314 to 393 after trypsin treatment. This
peptide, which covers the region from after the C terminus of
VP34 to the N terminus of VP27, has a calculated molecular mass
of 8.5 kDa. The VP90 polypeptide region from residues 314 to 393
contains a total of six arginines/lysines (i.e., K347, R359, R361,
R365, R366, and K380) that are potentially susceptible to trypsin.
These six residues are all surface exposed in the crystal structure of
VP90
71–415
. In the context of the capsid, the most susceptible sites
are R359 and R361, which are located in a structurally flexible loop
hovering over the P1 domain (Fig. 4 and 5). The accessibility of the
other four residues can be ranked in descending order as R365/
R366 ⬎K347 ⬎K380 due to the following considerations: R365
and R366 are in a ␤-strand underneath the loop containing R359
and R361, K347 is in helix ␣4 adjacent to the continuous capsid
shell, and K380 lies at the quasi-3-fold trimer interface. Indeed,
the theoretical molecular mass of the polypeptide containing res-
idues 1 to 359 is 38.5 kDa, exactly matching that of the cleavage
intermediate VP38.5 (23), suggesting a cleavage event at either
R359 or R361 during astrovirus maturation. In our capsid struc-
ture model, the polypeptide spanning residues 359 to 393, which is
comprised of ␤14 and ␣5, is located near the center of the quasi-
3-fold trimers (Fig. 5d).
The proteolytically induced maturation of the capsid results in
FIG 5 Trypsin-mediated human astrovirus capsid maturation. (a) EM reconstruction of the mature capsid with a surface spike. (b) EM reconstruction of the
mature capsid without a spike. (c) EM reconstruction of the immature capsid without a spike. Molecules A-C and D-F form two separate trimers across the 2-fold
symmetry axis. In panels a to c, only two asymmetry units are shown. (d) Fitted capsid structure model. The P1 domain is in yellow. Trypsin cleavage sites are
highlighted in the top asymmetric unit. In the asymmetric unit at the bottom, residues 360 to 398 are highlighted in dark gray.
Toh et al.
9014 jvi.asm.org October 2016 Volume 90 Number 20Journal of Virology

a⬎100-fold increase in infectivity (23,24). The underlying mech-
anism of acquired infectivity has not yet been determined but
could be related to an altered structural stability of the capsid that
is essential for uncoating, exposure of cell receptor binding sites,
and/or liberation of viral penetration factors residing within the
cleaved C terminus of the VP41 intermediate that would allow
astrovirus internalization (23). The capsid model of VP90
71–415
will provide a valuable structural framework for further biochem-
ical and genetic analyses to pinpoint cleavage sites essential for
astrovirus infectivity and the associated mechanisms.
Insights into astrovirus assembly. Recombinant human as-
trovirus proteins, in the form of either VP70 or VP90, with or
without the first 70 amino acid residues, have been found to self-
assemble into VLPs when overexpressed in mammalian cells (21–
23). It is also evident from these studies that the efficiency and
accuracy of astrovirus VLP formation were rather low compared
to those of HEV as the closest structural homolog. Recombinant
HEV CP self-assembles into T⫽1orT⫽3 VLPs, depending on the
presence or absence of the N-terminal peptide of the CP (48).
However, our results show that astrovirus VP70 expressed in in-
sect cells or E. coli formed predominantly dimers (data not
shown). Comparison of the crystal structure of the HEV VLP and
the HAstV capsid model shows that the HAstV CP-CP interaction
around the 3-fold symmetry axis is substantially weaker, with
⬃2,000 Å
2
of buried surface (S, ⬃1,000 Å
2
; P1, ⬃1,000 Å
2
; P2,
none), compared to ⬃3,000 Å
2
(S, 2,000 Å
2
; P1, 1,000 Å
2
; P2,
none) in HEV. Therefore, the lack of efficient assembly of astro-
virus VLPs may be due partly to a weak interaction around the
3-fold symmetry axis.
Several approaches may help to enhance the efficiency of astro-
virus VLP assembly. For instance, peptide sequences can be either
an insertion or conjugated to the N/C terminus of the astrovirus
CP to promote trimer formation. Disulfide bond engineering at
the trimer interface may also help to stabilize astrovirus CP inter-
actions. Furthermore, molecular modeling based on our VP90
71–415
capsid model should be able to identify additional mutations that
favor trimer formation. The ability to obtain large quantities of
VLPs would greatly facilitate detailed characterization of the as-
trovirus capsid function and the maturation mechanism. Further-
more, development of VLPs that can encapsidate exogenous nu-
cleic acids would allow them to be used as delivery agents of RNA
or DNA for different purposes.
Serotype-related sequence and structural differences. Near
the end of manuscript preparation, the crystal structure of
HAstV-1 VP90
71–415
was reported (49). The structure reported
under PDB accession number 5EWN contains two independent
FIG 6 Structural conservation among the eight HAstV serotypes. (a) Structural comparison between HAstV-1 and HAstV-8 VP90
71–415
. Variable regions
are shown in red, while regions superimposed well are shown in blue. (b) Multiple-sequence alignment for CPs from the eight astrovirus serotypes
HAstV-1 (UniProtKB accession number O12792), HAstV-2 (UniProtKB accession number J7LPD5), HAstV-3 (UniProtKB accession number Q9WFZ0),
HAstV-4 (UniProtKB accession number Q3ZN05), HAstV-5 (UniProtKB accession number Q4TWH7), HAstV-6 (UniProtKB accession number
Q67815), HAstV-7 (UniProtKB accession number Q96818), and HAstV-8 (GenBank accession number AAF85964.1). Only residues 71 to 415 of the CP
are included in the alignment. Superimposed secondary structural elements are taken from the crystal structure of HAstV-8 VP90
71–415
. Conserved
arginines and lysines from the P1 domain are highlighted by red stars, with likely cleavage sites further underlined by red triangles. (c) Sequence variations
mapped to the HAstV-8 VP90
71–415
structure. The structure is colored in rainbow coloring, with conserved regions in blue and poorly conserved regions
in red.
Structure of the Human Astrovirus Capsid Protein
October 2016 Volume 90 Number 20 jvi.asm.org 9015Journal of Virology

HAstV-1 VP90
71–415
molecules. Superimposition of HAstV-8
VP90
71–415
onto each of the two HAstV-1 VP90
71–415
molecules
gives average root mean square deviations (RMSDs) of 0.92 Å and
0.90 Å for 324 common C
␣
atoms, which are slightly higher than
the average RMSD of 0.83 Å from superimposing the two
HAstV-1 VP90
71–415
molecules. Major variations between the two
VP90
71–415
structures from the two different serotypes are
mapped to the following four regions: the S-P1 domain linker, the
loop connecting ␣3 and ␤12, the loop connecting ␤13 and ␤14,
and the loop connecting ␣5 and ␤15 (Fig. 6a). Additionally, ␣4
from HAstV-8 is disordered in both subunits of the HAstV-1
structure. Except for the S-P1 domain linker, the other three
structurally variable regions all have sequences that are poorly
conserved among the eight serotypes (Fig. 6b).
To highlight sequence regions of VP90
71–415
that are highly
variable among the eight HAstV serotypes, HAstV-8 VP90
71–415
is
colored according to the level of sequence conservation (Fig. 6b
and c). The S domain is highly conserved, except for the HI loop,
which forms a small plateau around the 5-fold symmetry axis and
is largely exposed in the viral capsid. The P1 domain contains both
highly conserved and poorly conserved regions. The three most
poorly conserved regions of P1 are the ␣3-␤12 loop, the ␤13-␤14
loop, and the ␣5-␤15 loop, all of which are completely exposed on
the surface of P1. The structural flexibility of these three loop
regions is the highest based on temperature factors of the struc-
ture. The polypeptide regions from ␤9to␣3 are also variable but
to a lesser degree. The structure of the capsid models shows that
this region is partially shielded by other structural elements in the
context of the capsid. Our results thus indicate that the presence of
the P2 domain does not exclude access to the P1 domain by the
host immune system. The top surface of P1, especially the three
surface loops mentioned above, may be immunogenic and con-
tain neutralization epitopes recognized by antibodies generated
during human astrovirus infection.
ACKNOWLEDGMENTS
We thank Corey Hryc for technical support.
This work was supported by grants from the Welch Foundation and
the National Institutes of Health.
FUNDING INFORMATION
This work, including the efforts of Yizhi J. Tao, was funded by HHS |
National Institutes of Health (NIH) (AI103777). This work, including the
efforts of Mark Yeager, was funded by HHS | National Institutes of Health
(NIH) (GM066087). This work, including the efforts of Yizhi J. Tao, was
funded by Welch Foundation (C-1565).
REFERENCES
1. Mendez E, Arias CF. 2007. Astroviruses, p 981–1000. In Knipe DE,
Howley PM, Griffin DE, Lamb RA, Martin MA, Roizman B, Straus SE
(ed), Fields virology, 5th ed, vol I. Lippincott Williams & Wilkins,
Philadelphia, PA.
2. Appleton H, Higgins PG. 1975. Letter: viruses and gastroenteritis in
infants. Lancet i:1297.
3. Madeley CR, Cosgrove BP. 1975. Letter: 28 nm particles in faeces in
infantile gastroenteritis. Lancet ii:451–452.
4. Corcoran MS, van Well GT, van Loo IH. 2014. Diagnosis of viral
gastroenteritis in children: interpretation of real-time PCR results and
relation to clinical symptoms. Eur J Clin Microbiol Infect Dis 33:1663–
1673. http://dx.doi.org/10.1007/s10096-014-2135-6.
5. Matsui SM, Greenberg HM. 2001. Astroviruses, p 875–894. In Knipe
DM, Howley PM, Griffin DE, Lamb RA, Martin MA, Roizman B, Straus
SE (ed), Fields virology, 4th ed, vol 1. Lippincott-Raven, Philadelphia, PA.
6. Donelli G, Superti F, Tinari A, Marziano ML. 1992. Mechanism of
astrovirus entry into Graham 293 cells. J Med Virol 38:271–277. http://dx
.doi.org/10.1002/jmv.1890380408.
7. Mendez E, Munoz-Yanez C, Sanchez-San Martin C, Aguirre-Crespo G,
Banos-Lara MDR, Gutierrez M, Espinosa R, Acevedo Y, Arias CF,
Lopez S. 2014. Characterization of human astrovirus cell entry. J Virol
88:2452–2460. http://dx.doi.org/10.1128/JVI.02908-13.
8. Jiang B, Monroe SS, Koonin EV, Stine SE, Glass RI. 1993. RNA sequence
of astrovirus: distinctive genomic organization and a putative retrovirus-
like ribosomal frameshifting signal that directs the viral replicase synthe-
sis. Proc Natl Acad SciUSA90:10539–10543. http://dx.doi.org/10.1073
/pnas.90.22.10539.
9. Willcocks MM, Brown TD, Madeley CR, Carter MJ. 1994. The complete
sequence of a human astrovirus. J Gen Virol 75(Part 7):1785–1788.
10. Lewis TL, Matsui SM. 1995. An astrovirus frameshift signal induces
ribosomal frameshifting in vitro. Arch Virol 140:1127–1135. http://dx.doi
.org/10.1007/BF01315421.
11. Lewis TL, Matsui SM. 1996. Astrovirus ribosomal frameshifting in an
infection-transfection transient expression system. J Virol 70:2869–2875.
12. Lewis TL, Matsui SM. 1997. Studies of the astrovirus signal that induces
(⫺1) ribosomal frameshifting. Adv Exp Med Biol 412:323–330. http://dx
.doi.org/10.1007/978-1-4899-1828-4_53.
13. Marczinke B, Bloys AJ, Brown TD, Willcocks MM, Carter MJ, Brierley
I. 1994. The human astrovirus RNA-dependent RNA polymerase coding
region is expressed by ribosomal frameshifting. J Virol 68:5588–5595.
14. Bass DM, Qiu S. 2000. Proteolytic processing of the astrovirus capsid. J
Virol 74:1810–1814. http://dx.doi.org/10.1128/JVI.74.4.1810-1814.2000.
15. Dong J, Dong L, Méndez E, Tao Y. 2011. Crystal structure of the human
astrovirus capsid spike. Proc Natl Acad SciUSA108:12681–12686. http:
//dx.doi.org/10.1073/pnas.1104834108.
16. Wang QH, Kakizawa J, Wen LY, Shimizu M, Nishio O, Fang ZY,
Ushijima H. 2001. Genetic analysis of the capsid region of astroviruses. J
Med Virol 64:245–255. http://dx.doi.org/10.1002/jmv.1043.
17. Guix S, Bosch A, Ribes E, Dora Martínez L, Pintó RM. 2004. Apoptosis
in astrovirus-infected CaCo-2 cells. Virology 319:249–261. http://dx.doi
.org/10.1016/j.virol.2003.10.036.
18. Méndez E, Aguirre-Crespo G, Zavala G, Arias CF. 2007. Association of
the astrovirus structural protein VP90 with membranes plays a role in
virus morphogenesis. J Virol 81:10649–10658. http://dx.doi.org/10.1128
/JVI.00785-07.
19. Banos-Lara MDR, Méndez E. 2010. Role of individual caspases induced
by astrovirus on the processing of its structural protein and its release from
the cell through a non-lytic mechanism. Virology 401:322–332. http://dx
.doi.org/10.1016/j.virol.2010.02.028.
20. Méndez E, Salas-Ocampo E, Arias CF. 2004. Caspases mediate process-
ing of the capsid precursor and cell release of human astroviruses. J Virol
78:8601–8608. http://dx.doi.org/10.1128/JVI.78.16.8601-8608.2004.
21. Caballero S, Guix S, Ribes E, Bosch A, Pintó RM. 2004. Structural
requirements of astrovirus virus-like particles assembled in insect cells. J
Virol 78:13285–13292. http://dx.doi.org/10.1128/JVI.78.23.13285-13292
.2004.
22. Dalton RM, Pastrana EP, Sánchez-Fauquier A. 2003. Vaccinia virus
recombinant expressing an 87-kilodalton polyprotein that is sufficient to
form astrovirus-like particles. J Virol 77:9094–9098. http://dx.doi.org/10
.1128/JVI.77.16.9094-9098.2003.
23. Méndez E, Fernández-Luna T, López S, Méndez-Toss M, Arias CF.
2002. Proteolytic processing of a serotype 8 human astrovirus ORF2 poly-
protein. J Virol 76:7996–8002. http://dx.doi.org/10.1128/JVI.76.16.7996
-8002.2002.
24. Lee TW, Kurtz JB. 1981. Serial propagation of astrovirus in tissue culture
with the aid of trypsin. J Gen Virol 57:421– 424. http://dx.doi.org/10.1099
/0022-1317-57-2-421.
25. Monroe SS, Stine SE, Gorelkin L, Herrmann JE, Blacklow NR, Glass RI.
1991. Temporal synthesis of proteins and RNAs during human astrovirus
infection of cultured cells. J Virol 65:641–648.
26. Sanchez-Fauquier A, Carrascosa AL, Carrascosa JL, Otero A, Glass RI,
Lopez JA, San Martin C, Melero JA. 1994. Characterization of a human
astrovirus serotype 2 structural protein (VP26) that contains an epitope
involved in virus neutralization. Virology 201:312–320. http://dx.doi.org
/10.1006/viro.1994.1296.
27. Krishna NK. 2005. Identification of structural domains involved in astro-
virus capsid biology. Viral Immunol 18:17–26. http://dx.doi.org/10.1089
/vim.2005.18.17.
Toh et al.
9016 jvi.asm.org October 2016 Volume 90 Number 20Journal of Virology

28. Harrison SC. 2007. Principles of virus structure, p 60–98. In Knipe DM,
Howley PM, Griffin DE, Lamb RA, Martin MA, Roizman B, Straus SE
(ed), Fields virology, 5th ed, vol I. Lippincott Williams & Wilkins, Phila-
delphia, PA.
29. Hogle JM, Maeda A, Harrison SC. 1986. Structure and assembly of
turnip crinkle virus. I. X-ray crystallographic structure analysis at 3.2 A
resolution. J Mol Biol 191:625–638.
30. Dryden KA, Tihova M, Nowotny N, Matsui SM, Mendez E, Yeager M.
2012. Immature and mature human astrovirus: structure, conformational
changes, and similarities to hepatitis E virus. J Mol Biol 422:650–658.
http://dx.doi.org/10.1016/j.jmb.2012.06.029.
31. Guu TS, Liu Z, Ye Q, Mata DA, Li K, Yin C, Zhang J, Tao YJ. 2009.
Structure of the hepatitis E virus-like particle suggests mechanisms for
virus assembly and receptor binding. Proc Natl Acad SciUSA106:12992–
12997. http://dx.doi.org/10.1073/pnas.0904848106.
32. Yamashita T, Mori Y, Miyazaki N, Cheng RH, Yoshimura M, Unno H,
Shima R, Moriishi K, Tsukihara T, Li TC, Takeda N, Miyamura T,
Matsuura Y. 2009. Biological and immunological characteristics of hep-
atitis E virus-like particles based on the crystal structure. Proc Natl Acad
SciUSA106:12986 –12991. http://dx.doi.org/10.1073/pnas.0903699106.
33. Doublié S. 1997. Preparation of selenomethionyl proteins for phase de-
termination. Methods Enzymol 276:523–530. http://dx.doi.org/10.1016
/S0076-6879(97)76075-0.
34. Otwinowski Z, Minor W. 1997. Processing of X-ray diffraction data
collected in oscillation mode, vol 276, p 307–326. Elsevier, New York, NY.
35. Adams PD, Afonine PV, Bunkóczi G, Chen VB, Davis IW, Echols N,
Headd JJ, Hung L-W, Kapral GJ, Grosse-Kunstleve RW, McCoy AJ,
Moriarty NW, Oeffner R, Read RJ, Richardson DC, Richardson JS, Ter-
williger TC, Zwart PH. 2010. PHENIX: a comprehensive Python-based sys-
tem for macromolecular structure solution. Acta Crystallogr D Biol Crystal-
logr 66:213–221. http://dx.doi.org/10.1107/S0907444909052925.
36. Emsley P, Lohkamp B, Scott WG, Cowtan K. 2010. Features and devel-
opment of Coot. Acta Crystallogr D Biol Crystallogr 66:486–501. http:
//dx.doi.org/10.1107/S0907444910007493.
37. Pettersen EF, Goddard TD, Huang CC, Couch GS, Greenblatt DM,
Meng EC, Ferrin TE. 2004. UCSF Chimera—a visualization system for
exploratory research and analysis. J Comput Chem 25:1605–1612. http:
//dx.doi.org/10.1002/jcc.20084.
38. Holm L, Rosenström P. 2010. Dali server: conservation mapping in 3D.
Nucleic Acids Res 38:W545–W549. http://dx.doi.org/10.1093/nar/gkq366.
39. Morgunova E, Dauter Z, Fry E, Stuart DI, Stel’mashchuk V, Mikhailov
AM, Wilson KS, Vainshtein BK. 1994. The atomic structure of carnation
mottle virus capsid protein. FEBS Lett 338:267–271. http://dx.doi.org/10
.1016/0014-5793(94)80281-5.
40. Olson AJ, Bricogne G, Harrison SC. 1983. Structure of tomato busy stunt
virus. IV. The virus particle at 2.9 A resolution. J Mol Biol 171:61–93.
41. Plevka P, Tars K, Zeltins A, Balke I, Truve E, Liljas L. 2007. The
three-dimensional structure of ryegrass mottle virus at 2.9 A resolution.
Virology 369:364–374. http://dx.doi.org/10.1016/j.virol.2007.07.028.
42. Guo YR, Hryc CF, Jakana J, Jiang H, Wang D, Chiu W, Zhong W, Tao
YJ. 2014. Crystal structure of a nematode-infecting virus. Proc Natl Acad
Sci U S A 111:12781–12786. http://dx.doi.org/10.1073/pnas.1407122111.
43. Matsui SM, Kiang D, Ginzton N, Chew T, Geigenmuller-Gnirke U.
2001. Molecular biology of astroviruses: selected highlights. Novartis
Found Symp 238:219 –233; discussion 233–236. http://dx.doi.org/10.1002
/0470846534.ch13.
44. Hellen CU, Wimmer E. 1992. The role of proteolytic processing in the
morphogenesis of virus particles. Experientia 48:201–215. http://dx.doi
.org/10.1007/BF01923512.
45. Roby JA, Setoh YX, Hall RA, Khromykh AA. 2015. Post-translational
regulation and modifications of flavivirus structural proteins. J Gen Virol
96:1551–1569. http://dx.doi.org/10.1099/vir.0.000097.
46. Mangel WF, San Martin C. 2014. Structure, function and dynamics in
adenovirus maturation. Viruses 6:4536–4570. http://dx.doi.org/10.3390
/v6114536.
47. Ganser-Pornillos B, Yeager M, Sundquist M. 2008. The structural bio-
logy of HIV assembly. Curr Opin Struct Biol 18:203–217. http://dx.doi.org
/10.1016/j.sbi.2008.02.001.
48. Liu Z, Tao YJ, Zhang J. 7 November 2011. Chapter 7. Structure and
function of the hepatitis E virus capsid related to hepatitis E pathogenesis.
In Mukomolov SL (ed), Viral hepatitis—selected issues of pathogenesis
and diagnostics. InTech. http://dx.doi.org/10.5772/27635.
49. York RL, Yousefi PA, Bogdanoff W, Haile S, Tripathi S, DuBois RM.
2016. Structural, mechanistic, and antigenic characterization of the hu-
man astrovirus capsid. J Virol 90:2254–2263. http://dx.doi.org/10.1128
/JVI.02666-15.
Structure of the Human Astrovirus Capsid Protein
October 2016 Volume 90 Number 20 jvi.asm.org 9017Journal of Virology