Classification
The Coronaviridae were recognized as a new virus family in 1968 because their virion
morphology and intracellular budding site distinguished them from other RNA viruses.
Characteristic features of their genome, replication strategy, structural proteins
and polymerase later supported this classification. The toroviruses and coronaviruses
were recognized as separate genera within the Coronaviridae family in 1993.
The Coronaviridae and Arteriviridae are now classified as members of the Nidovirales
order, viruses with monopartite plus-strand RNA genomes that are transcribed to yield
a nested set of overlapping subgenomic mRNAs that have a common 3′ end.
Virion Structure and Proteins
A model of coronavirus virions is shown in Fig. 1
. The enveloped virions are approximately 100 nm in diameter, and are characterized
by large, petal-shaped spikes. The spikes are oligomers of the 180–200 kDa S glycoprotein
that binds to receptor glycoproteins and induces fusion of the viral envelope with
cell membranes and, sometimes, cell–cell fusion. S proteins of several coronaviruses
also bind to 9-O-acetylated sialic acid. A small envelope glycoprotein, M, traverses
the lipid bilayer three times and interacts with the nucleocapsid in the virion. Intracellular
transport of the M glycoprotein is arrested in the Golgi, which may determine the
intracellular budding site of coronaviruses. The small E glycoprotein, originally
believed to be a nonstructural protein, is present in small amounts in virions, and
is essential for virus budding. Envelopes of some coronaviruses also contain a hemagglutinin-esterase
glycoprotein, HE, which forms short spikes that bind N-acetyl-9-O-acetylneuraminic
acid or N-glycolylneuraminic acid and have esterase activity. The nucleocapsid protein,
N, encapsidates the monopartite, linear, single-stranded genomic RNA. The internal
structure of coronaviruses was originally believed to be a helical nucleocapsid as
shown in Fig. 1. Recently, however, cores that appear to have cubic symmetry were
observed in detergent-treated virus preparations. Thus, the internal structure of
coronaviruses is not yet understood.
Figure 1
Model of coronavirus virion. The helical nucleocapsid containing the 30 kb, plus-sense
RNA genome is coiled within an envelope that contains the S, M and E glycoproteins.
Some coronaviruses related to MHV also express the HE glycoprotein. (Modified and
published with permission from Fields BN, Knipe DM, Howley PM et al (eds) (1996) Fields
Virology, 3rd edn. Philadelphia: Lippincott-Raven).
Genome Structure
The 27–32 kb, plus-strand RNA genomes of coronaviruses are capped and polyadenylated.
Transfection of genomic RNA into cells leads to production of infectious virions.
The genomes of several coronaviruses have been sequenced, although no infectious coronavirus
cDNA has yet been obtained.
A map of the genome of murine coronavirus MHV (Fig. 2
) illustrates the characteristic features of coronavirus genomic RNA. At the 5′ end
of the genome is a cap with a leader RNA of approximately 70 bp. The order of the
genes encoding the polymerase and structural proteins is the same in genomes of all
coronaviruses (Pol, S, E, M, N), but several additional open reading frames (ORFs)
are interspersed among these genes. These ORFs encode nonstructural (NS) proteins
of unknown functions, and they differ in number, position and sequence among different
coronaviruses.
Figure 2
Map of the genome of murine coronavirus MHV and intracellular coronavirus RNAs. The
polyadenylated plus-strand RNA genome is approximately 32 kb long. ORFs that encode
structural proteins are shown in open boxes with the protein encoded by the gene inside.
This gene order is the same in all coronaviruses. The RNA polymerase gene at the 5′
end of all coronavirus genomes consists of two ORFs, 1a and 1b, with a pseudoknot
and a slippery sequence in the overlapping region. ORFs encoding putative nonstructural
proteins (shown in filled boxes) vary in number and position among different coronaviruses.
The conserved intergenic sequence on all plus-strand RNAs is shown by IS. The cap
and approximately 70 bp leader RNA at the 5′ ends of each plus-strand RNA are indicated
by a shaded box. Minus-strand templates are shown by dashed lines with a sequence
complementary to the leader at the 3′ end indicated by a barred box, and a poly (U)
tract at the 5′ end. Each subgenomic mRNA is translated to yield only the protein
encoded by the 5′ end of the plus-strand RNA.
For both Coronaviridae and Arteriviridae, the RNA-dependent RNA polymerase gene at
the 5′ end of the genome consists of two ORFs in different reading frames, joined
by a pseudoknot at a ribosomal frame shift site. The 5′ and 3′ ends of the genome
contain predicted complex stem loop structures, and additional stem loops are predicted
within some ORFs.
Preceding every ORF in the genome is an intergenic sequence (IS) consisting of one
or more UCUAAAC sequences for MHV or related sequences for other coronaviruses. The
IS is included in a promoter that regulates the transcription of different viral genes.
The location on the genome of packaging signals for the initiation of encapsidation
of genomic RNA by the N protein appears to differ among coronaviruses.
Transcription, Replication, Translation and Genetics
Upon infection, the two large ORFs in the 20 kb polymerase gene of the genomic RNA
are translated, via ribosomal frameshift, as a single, very large polyprotein that
is cleaved during translation by 3-C-like and papain-like proteases within the nascent
protein to yield a variety of peptides. The processing of the large polyprotein and
the functions of the resulting peptides in RNA-dependent RNA polymerase activity are
being analyzed in many laboratories.
Using the genomic RNA as template, the polymerase generates a full-length, minus-strand
RNA with a 5′ poly (U) sequence. Infected cells also contain a 3′ coterminal nested
set of polyadenylated subgenomic plus-strand mRNAs that have a cap and the approximately
70 bp leader sequence at their 5′ ends (Fig. 2). A negative-strand RNA template corresponding
to each subgenomic mRNA is also found in infected cells (Fig. 2). A nonprocessive,
leader-primed polymerase activity was postulated to account for the presence of the
leader on each mRNA. A complex of polymerase and leader transcribed from the 3′ end
of the full-length negative-strand template would bind to an IS on the template and
act as a primer for synthesis of the subgenomic mRNA. This hypothesis does not explain
the presence of the negative-strand templates for each subgenomic RNA. An alternative
hypothesis suggests that polymerase jumping occurs during negative-strand synthesis,
generating the subgenomic negative-strand templates for subsequent mRNA synthesis.
Mutations in the IS preceding an ORF can prevent transcription of the mRNA with that
ORF at its 5′ end. Although virus strains differ in the relative amounts of various
mRNAs in infected cells, there is little temporal regulation of transcription.
Coronavirus replication is associated with a high frequency of mutations that arise
by two mechanisms. First, during the replication of each approximately 30 kb RNA genome,
several point mutations would be expected to occur, based upon the known error frequency
of RNA polymerases. Second, coronavirus genomes undergo a very high frequency of RNA
recombination.
Small or large deletions at the sites of recombination can produce mutant viruses,
or defective interfering (DI) RNAs and subgenomic replicons. Large deletions in certain
sites of the gene encoding the S glycoprotein and mutations in the N gene yield viruses
with altered virulence, tissue tropism, or thermal stability. Thus, even after sequential
single plaque isolations, coronavirus stocks always contain a mixture of quasispecies.
Different virus variants may be selected during passage in different cell lines or
under a variety of culture conditions. Therefore it is important to document the passage
history of coronavirus strains or mutants.
Rarely, recombination may occur between coronavirus genomic RNA and mRNAs of cells
or unrelated viruses. Such an illegitimate RNA recombination event may have resulted
in the incorporation of the mRNA that encodes the HE glycoprotein of influenza C into
the genome of a coronavirus ancestral to the MHV group. Similarly, some of the NS
ORFs may have been acquired through recombination with foreign mRNAs.
Genetic analysis of coronaviruses is somewhat limited because no full-length, infectious
cDNA clones of coronavirus genomes are yet available. DI RNAs transfected into cells
infected with wild-type virus can be replicated and have been used to study encapsidation
and transcription initiation signals. Site-specific mutations have been introduced
into several coronavirus genomes by targeted RNA recombination in infected cells transfected
with a mutagenized subgenomic cDNA.
Coronavirus replication occurs in the cytoplasm and does not require the nucleus.
Only the ORF at the 5′ end of each of the coronavirus mRNAs is translated within the
infected cell or in vitro. The phosphorylated N protein, which is translated from
one of the smallest mRNAs, is the most abundant viral protein in infected cells. It
assembles with genomic plus-strand RNA in the cytoplasm to form helical nucleocapsids.
Viral glycoproteins are translated on the rough endoplasmic reticulum, where S oligomerizes,
then they are transported to the Golgi, where S proteins of some coronaviruses are
cleaved by a trypsin-like host cell protease to yield S1 and S2 peptides. The S and
HE glycoproteins are expressed on the plasma membrane. The viral S, HE, M and E glycoproteins
and the nucleocapsid assemble by budding at a special pre-Golgi compartment, and the
virions are apparently transported to the cell surface in large vesicles that fuse
with the plasma membrane to release virions from the intact cell by exocytosis. Infected
cells are characteristically coated with a thick layer of adsorbed virions.
Host Range, Tissue Tropism and Virus Propagation
Most coronaviruses cause epidemic disease in only one species, although limited replication,
usually without disease, may result from experimental inoculation of other species.
Coronaviruses typically cause respiratory or enteric diseases, although several can
also cause hepatitis, infectious peritonitis, nephritis, myocarditis, sialodacryadenitis,
or neurological, reproductive or immunological disorders. The viruses were named for
their natural host and sometimes for the associated disease: for example, avian infectious
bronchitis virus (IBV); mouse hepatitis virus (MHV); sialodacryadenitis virus of rats
(SDAV); bovine coronavirus (BCoV); porcine hemagglutinating encephalomyelitis virus
(HEV); turkey bluecomb coronavirus (TCoV);
human respiratory coronaviruses (HCoV); transmissible gastroenteritis virus of swine
(TGEV); porcine respiratory coronavirus (PRCV); canine coronavirus (CCoV); feline
infectious peritonitis virus (FIPV) and feline enteric coronavirus (FeCoV); and rabbit
coronavirus (RbCoV).
In vivo, coronaviruses bind to receptors expressed on the apical membranes of enteric
and respiratory epithelial cells, and are released either from the apical or basolateral
borders or both, depending upon the virus.
Although strains of many coronaviruses have been adapted to growth in continuous cell
lines, isolation of human coronaviruses from infected patients may require human fetal
tracheal organ cultures. Some coronaviruses, such as rabbit coronavirus or enterotropic
strains of MHV, cannot be propagated in cell culture, but require animal passage.
Coronavirus-like particles seen in electron micrographs of human feces (human enteric
coronaviruses, HECV) have not been adapted to serial passage in cell culture.
IBV and TCoV can be propagated in embryonated eggs, and some strains grow in avian
cell lines. Although cells of the natural host species are generally required for
infection by virions, purified coronavirus genomic RNA can infect cells across species
barriers. Receptors have been identified for several coronaviruses: MHV uses murine
biliary glycoproteins in the immunoglobulin superfamily; HCoV-229E and TGEV use human
and porcine aminopeptidase N (APN), respectively; FIPV and FeCoV use feline APN, which
can also be utilized by HCoV-229E and TGEV; and BCoV and HCoV-OC43 use N-acetyl-9-O-acetyl
neuraminic acid moieties. Expression of the cloned receptor glycoproteins in cells
of a foreign species can render them susceptible to infection with coronavirus virions.
Thus, coronavirus–receptor interactions are an important determinant of the species
specificity of coronavirus infection.
Coronavirus infection of cells may be inapparent or cause cell fusion, vacuolization,
rounding and/or cell death. Cytopathic effects are minimized, and virus yield and
stability are increased at acid pH. Some strains of MHV infect cells by receptor-dependent
fusion of the viral envelope with the plasma membrane, but other strains of MHV and
other coronaviruses appear to enter via fusion with endosomal membranes. For some
coronaviruses, cleavage near the middle of the S glycoprotein at a sequence of basic
amino acids yields the noncovalently linked S1 and S2 peptides and enhances viral
infectivity and/or cell fusion. S glycoproteins of many other coronaviruses, such
as FIPV, lack the protease target sequence and do not require protease activation
for infectivity or cell fusion.
Serologic and Evolutionary Relationships
There are three coronavirus serogroups. One group includes HCoV-229E, TGEV, PRCoV,
CCoV, FCoV, and others. A second group includes MHV, BCoV, SDAV, HEV, HCoV-OC43 and
others. Avian IBV strains make up the third serogroup. Phylogenetic analysis of coronavirus
N and S genes correlate well with the division of coronaviruses into these three groups.
Nucleic acid sequence analysis also shows that certain pairs of coronaviruses that
cause different syndromes in the same host should probably be considered strains of
a single virus. Thus, PRCoV and TGEV of swine, which cause epizootic respiratory and
enteric disease, respectively, are highly homologous, except for a deletion of more
than 675 nucleotides in the S gene of PRCoV. The feline coronavirus FeCoV causes epizootic
enteric disease in cats, and mutant FIPV viruses that arise within FeCoV-infected
animals cause fatal systemic disease.
The HE glycoprotein is encoded by a gene found in the MHV group of coronaviruses,
but not in the HCoV-229E or IBV groups. Phylogenetic analysis of coronavirus HE gene
suggests that BCoV and HCoV-OC43 are more closely related to each other than to MHV.
The organization of the polymerase gene with two large slightly overlapping ORFs in
different reading frames joined by a pseudoknot is conserved among all coronaviruses
and shared by toroviruses such as Berne virus (BEV) and Arteriviridae. In addition,
BEV encodes a protein with significant homology to HE of coronaviruses and influenza
C virus.
Recombination between genomes of related coronaviruses can occur in experimentally
inoculated cell cultures or animals, and also during natural outbreaks of disease
in the natural host. Recombinants between IBV strains have been isolated from infected
birds. Nucleotide sequencing shows that one biotype of feline coronavirus that causes
epizootic disease is a recombinant between canine coronavirus and the other biotype
of feline coronavirus. Mutants of MHV derived from persistently infected murine cell
lines have acquired the ability to infect nonmurine cells. These observations indicate
that there may be naturally occurring interactions between the genomes of coronaviruses
from different species.
Epidemiology
Most coronaviruses cause epidemic or epizootic disease in only one species. Because
of the great antigenic variability among coronavirus strains and because many coronaviruses
replicate only in epithelia where protective immunity is relatively short-lived, reinfection
is common. Infection is often inapparent, and virus particles or antigens and subsequent
seroconversion are observed in healthy individuals. The viruses are enzootic or endemic
in their host species, causing sporadic disease and seasonal outbreaks when enough
susceptibles are available. Adults with acute, self-limited, inapparent infection
transmit virus to neonates that develop clinical disease. Coronavirus diseases are
more severe in neonatal animals than adults, either because pre-existing immunity
to related virus strains moderates infections in adults, or because the immature immune
system permits higher levels of virus replication.
In immunocompromised hosts, infection may be inapparent but virus shedding may be
prolonged. Such individuals serve as reservoirs of virus. For example, enterotropic
MHV strains are endemic in colonies of laboratory mice, sustained both by persistent
infection of immunocompromised nude mice and by the continuous availability of new
susceptibles due to birth or importation.
Outbreaks of coronavirus diseases are often seasonal. In humans, coronaviruses cause
15–30% of colds, predominantly during the winter months, and outbreaks of different
human coronaviruses alternate at 2–3 year intervals. Outbreaks of BCoV-induced dysentery
in cattle also occur in the winter. Severe enteritis due to TGEV or BCoV infection
of suckling pigs or calves, respectively, occurs seasonally in correlation with breeding
cycles. Outbreaks of IBV-induced respiratory disease in chickens can occur at any
time, but are most common during the winter. Stress can exacerbate coronavirus-induced
diseases. Mice with inapparent MHV infection can develop hepatitis if they are subjected
to immunosuppression, thymectomy, transplanted tumors or infection with unrelated
organisms. Cattle with inapparent BCoV infection may develop respiratory disease during
shipping.
Coronaviruses have broad geographic distribution. Human infections with viruses related
to HCoV-229E or to HCoV-OC43 occur worldwide, as do IBV, TGEV, BCoV, FCoV and CCoV
infections. Occasionally, strains of these viruses that cause unusual manifestations
of disease arise and spread locally, and some of these viruses become very widely
distributed. For example, IBV strains that cause severe nephropathy arise and spread
locally. PRCoV, which causes a mild epizootic respiratory infection in swine, has
arisen several times in Europe and in the USA from TGEV by means of a very large deletion
in the S glycoprotein. The epizootic spread of PRCoV in European pigs has apparently
acted as a natural vaccine that has decreased the incidence of serious TGEV-induced
enteric disease in piglets.
Pathology
Most coronaviruses cause only respiratory or enteric disease in one host species.
These viruses generally replicate only in respiratory or enteric epithelial cells,
where the apical membranes express specific glycoprotein receptors for the viruses.
Coronaviruses are shed in respiratory secretions and/or feces. Some coronaviruses,
such as some MHV strains, FIPV and rabbit coronavirus, cause disseminated disease
and can replicate in macrophages, hepatocytes, neurons, glial cells, endothelial cells,
kidney epithelium, lymphocytes, urogenital tract and/or myocardium. In general, coronavirus
titers in the respiratory or enteric tract rise during the first 3–5 days postinoculation,
and recovery of infectious virus from an immunocompetent host is usually not possible
after 10–14 days, although viral antigens and RNA may continue to be detectable for
several weeks. Infectious coronavirus can be shed for months by immunocompromised
hosts.
Although most coronaviruses do not persist in immunocompetent hosts, coronaviruses
such as the neurotropic MHV-JHM strain can sometimes be detected in the brain months
or years after inoculation. Reverse transcriptase–polymerase chain reaction (RT-PCR)
can detect HCoV-229E and/or HCoV-OC43 RNA in up to 40% of brains from patients with
neurological diseases and from healthy individuals, but the significance of this observation
for human disease is unknown.
Coronavirus-induced lesions vary markedly depending upon the virus strain, dose and
tissue tropism and the genetic background of the host. Intestinal infections with
BCoV, MHV, TGEV, FeCoV, TCoV and CCoV cause loss of apical epithelial cells of the
intestinal villi and shortening and broadening of the villi.
Some enterotropic coronaviruses cause necrotizing enterocolitis, particularly in young
animals, while others cause only watery diarrhea or inapparent enteric infection.
Diarrhea is probably due to altered transport of fluids and electrolytes by the immature
epithelial cells that cover the blunted villi. Intestinal absorption of certain sugars
in TGEV-infected pigs remains altered for several days after the diarrhea has ceased.
Mononuclear inflammatory cells infiltrate the lamina propria. Reinfection with a different
strain of the virus generally causes more moderate disease than primary infection.
However, kittens are more likely to develop FIP following their second infection with
feline coronavirus, than after their first infection. Thus, an immunological response
to the primary infection may somehow facilitate the later development of disseminated
disease.
Human respiratory coronaviruses related to HCoV-229E or HCoV-OC43 infect the epithelial
cells in the upper respiratory tract and cause colds. Infection of young asthmatic
children can exacerbate wheezing, and lower respiratory tract coronavirus infection
has occasionally been observed in adults. Reinfection is frequent, even in volunteers
inoculated with the same strain of human coronavirus. Infection is usually demonstrated
by RT-PCR or by rising serum antibody titers because primary isolation of these viruses
from respiratory washings is difficult. It is not yet clear how many coronavirus strains
cause human respiratory disease or whether coronaviruses may play a role in other
human diseases.
IBV-induced respiratory diseases in chickens is of great economic importance. In addition,
some IBV strains are nephrotropic and cause kidney disease, while others infect the
oviduct and reduce egg laying. Recombinants between several IBV strains have been
isolated from infected flocks.
Neurological disease and hepatitis can result from coronavirus infections. MHV strains
cause local and/or systemic infections via respiratory or fecal/oral routes. While
some MHV strains are strictly enterotropic, most strains infect both the respiratory
and enteric tracts. Some MHV strains causes focal hepatitis, acute encephalitis, and
subacute or chronic focal demyelinating disease in the murine brain and spinal cord.
Cell fusion, necrosis and infiltration with mononuclear cells are observed in acutely
infected tissues, depending upon the strain of virus and strain of mice. Susceptibility
to MHV is affected by at least three murine genes that determine the virus receptor
isoforms expressed, the yield of infectious virus and the ability to generate monocyte
procoagulant activity, a prothrombinase in the coagulation pathway, in response to
MHV infection. Mutations in the MHV and TGEV S glycoproteins can alter tissue tropism,
virulence and persistence.
HEV causes respiratory infection and outbreaks of encephalomyelitis in suckling pigs,
and virus infection of neurons that innervate the stomach causes vomiting and wasting
syndrome.
Unusual syndromes associated with coronaviruses include feline infectious peritonitis,
rabbit myocarditis and rat sialodacryadenitis. FECV causes enteritis in kittens and
inapparent infection in adult cats, while the closely related FIPV can cause infectious
peritonitis, with ascites, wasting and death in a small percentage of infected cats.
A cardiotropic rabbit coronavirus causes dilated myocardiopathy and death within 7–12
days after intravenous inoculation of rabbits with serum from an infected rabbit.
Rat coronaviruses infect the respiratory tract, salivary and lacrimal glands, causing
sialodacryadenitis, and can also infect the urogenital tract, interfering with breeding.
Immune Responses
Infection of respiratory or enteric tracts of adults results in antiviral antibodies
in the serum, secretory antibody in the respiratory and enteric tracts, colostrum
and milk, and development of virus-specific T cells. These immune responses are useful
for diagnostic purposes, terminate infection, and probably ameliorate subsequent infections,
but they do not necessarily prevent reinfection. In newborns, passive oral immunization
with neutralizing antibody can sometimes prevent fatal coronavirus enteritis. Because
coronavirus diseases are generally most severe in newborns that do not respond well
to active immunization, an attractive strategy to protect these newborn animals that
is being explored is to vaccinate pregnant dams in order to maximize antiviral antibodies
in the colostrum and milk.
Coronaviruses can sometimes infect cells of the immune system and may modulate cytokines
and immune responses to unrelated immunogens. Infected macrophages can spread infection
to distant tissues. Mononuclear cell infiltrates in infected tissues consist of macrophages,
plasma cells, and CD4+ and CD8+ lymphocytes. Infection of glial cells with MHV can
upregulate expression of major histocompatibility complex (MHC) class I or II antigens,
making these cells potential targets for cytotoxic T cells. MHV infection causes thymic
atrophy, and the virus can also infect B lymphocytes. Polyclonal B cell activation
and hypergammaglobulinemia are associated with FIPV infection. Cheetahs, which have
little polymorphism in their MHC genes, have a much higher fatality rate after feline
coronavirus infection than cats.
Prevention and Control of Coronavirus Diseases
Because of the economic importance of coronavirus diseases of domestic animals, modified
live vaccines against IBV, TGEV, BCV, CCV and FIPV have been developed. However, they
do not provide solid protection from infection with wild-type coronaviruses. Other
approaches to protection or control of coronavirus diseases include use of recombinant
S proteins, synthetic peptides that mimic neutralization epitopes, passive immunization
with antibody against S glycoproteins, and treatment with interferon α or monoclonal
antireceptor antibody. Improvement of vaccines will require understanding of coronavirus
virulence factors, strain variation and mechanisms of immunopathology.
Future Perspectives
Full-length cDNA copies of coronavirus genomes and a way to express them to obtain
infectious virions are needed to investigate coronavirus replication and pathogenesis.
Analysis of the complex synthesis, processing and functions of the coronavirus polymerase
peptides will provide unique insight into mechanisms of RNA recombination, tools for
coronavirus genetics, and possibly new targets for drugs to inhibit coronavirus replication.
Improvements in diagnostic tests to identify coronavirus infections in humans and
animals will elucidate the epidemiology of these viruses and may implicate coronaviruses
in the etiology of additional diseases.
The characterization of coronavirus–host interactions during persistent infection
in vitro and in vivo will provide additional insight into coronavirus epidemiology
and pathogenesis. Further understanding of virus variants and host responses to coronavirus
infections of the respiratory and enteric tracts may lead to improved coronavirus
vaccines.
See also:
ARTERIVIRUSES (ARTERIVIRIDAE); TOROVIRUSES (CORONAVIRIDAE); IMMUNE RESPONSE | Cell
Mediated Immune Response; IMMUNE RESPONSE | General Features; RESPIRATORY VIRUSES;
RECOMBINATION OF VIRUSES; RHINOVIRUSES (PICORNAVIRIDAE).