Coronaviruses (CoVs) comprise a large family of viruses within the order Nidovirales containing single-stranded positive-sense RNA genomes of 27-32 kilobases.

Divided into four genera (alpha, beta, gamma, delta) and multiple newly defined subgenera, coronaviruses include a number of important human and livestock pathogens responsible for a range of diseases.

Historically, human coronaviruses OC43 and 229E have been associated with up to 30% of common colds, while the 2002 emergence of severe acute respiratory syndrome associated coronavirus (SARS-CoV) first raised the specter of these viruses as possible pandemic agents.

Although the SARS-CoV pandemic was quickly contained and the virus has not returned, the 2012 discovery of Middle East respiratory syndrome-associated coronavirus (MERS-CoV) once again elevated coronaviruses to a list of global public health threats.

The genetic diversity of these viruses has resulted in their utilization of both conserved and unique mechanisms of interaction with infected host cells.

Like all viruses, coronaviruses encode multiple mechanisms for evading, suppressing, or otherwise circumventing host antiviral responses.

Coronaviruses (CoVs) are a family of RNA viruses that cause significant diseases in humans such as severe acute respiratory syndrome (SARS) and other respiratory infections, as well as a variety of respiratory, gastrointestinal and other infections in an increasingly large variety of mammals and birds.

The capacity CoVs possess for trans-species movement and adaptation, long recognized in the laboratory, was confirmed in nature by the recent emergence of several animal coronavirus pathogens of domesticated animals and SARS-CoV.

Data on the latter emergence event indicated that SARS likely resulted from human infection and adaptation by a Bat SARS-like CoV.

Finally, the “post-SARS” identification and analysis of a vast diversity of newly identified coronaviruses across bat species, along with the successful synthetic resurrection of a Bat SARS-like CoV, suggests that many, if not all, mammalian CoVs may originate from bats.

However, the mechanisms of CoV host species movement and adaptation for replication and pathogenesis are poorly understood.

In order to survive and propagate, RNA viruses must achieve a balance between the capacity for adaptation to new environmental conditions or host cells with the need to maintain an intact and replication competent genome.

Several virus families in the order Nidovirales, such as the coronaviruses (CoVs) must achieve these objectives with the largest and most complex replicating RNA genomes known, up to 32 kb of positive-sense RNA.

The CoVs encode sixteen nonstructural proteins (nsp 1–16) with known or predicted RNA synthesis and modification activities, and it has been proposed that they are also responsible for the evolution of large genomes.

The CoVs, including murine hepatitis virus (MHV) and SARS-CoV, encode a 3′-to-5′ exoribonuclease activity (ExoN) in nsp14.

Genetic inactivation of ExoN activity in engineered SARS-CoV and MHV genomes by alanine substitution at conserved DE-D-D active site residues results in viable mutants that demonstrate 15- to 20-fold increases in mutation rates, up to 18 times greater than those tolerated for fidelity mutants of other RNA viruses.

Thus nsp14-ExoN is essential for replication fidelity, and likely serves either as a direct mediator or regulator of a more complex RNA proofreading machine, a process previously unprecedented in RNA virus biology.

Elucidation of the mechanisms of nsp14-mediated proofreading will have major implications for our understanding of the evolution of RNA viruses, and also will provide a robust model to investigate the balance between fidelity, diversity and pathogenesis.

The discovery of a protein distinct from a viral RdRp that regulates replication fidelity also raises the possibility that RNA genome replication fidelity may be adaptable to differing replication environments and selective pressures, rather than being a fixed determinant.

Coronavirus genome structure, replication, and transcription

Coronaviruses fall within the Cornidovirineae suborder, family Coronaviridae, and subfamily Orthocoronavirinae of the order Nidovirales.

These viruses are grouped together for their unique genome organization, ribosomal frameshift in the first open reading frame, and expression of 3’ structural and accessory genes via transcription of nested, subgenomic mRNAs .

Also within the order Nidovirales is the suborder Tornidovirineae which includes the genus Torovirus, that includes the toroviruses, which are distinguished from coronaviruses primarily by differences in their accessory genes.

The most notable feature of coronaviruses is their extraordinary genome size of up to 32 kilobases, among the largest RNA genomes, which is capped at the 5’ end and 3’ polyadenylated.

Recently, two nidoviruses have been discovered with genomes of 36 and 41 kilobases in length, highlighting the increase in coding capacity afforded by the expression of a viral proofreading exonuclease and a non- icosahedral nucleocapsid .

The 5’ ~20 kb of the coronavirus genome comprises the replicase gene (ORF1ab), which is translated from the genome as a single polyprotein and proteolytically processed into up to 16 constituent non-structural proteins (nsp).

In the majority of translation events, protein synthesis terminates at the end of ORF1a while a low-frequency ribosomal frameshift allows translation of the full length ORF1ab.

Downstream of the conserved replicase gene, in the 3’ 10 kb of the genome, are the coronavirus structural genes.

The four major structural genes, in their 5’->3’ order are spike (S), memberane (M), envelope (E), and nucleocapsid (N).

Some viruses within the Betacoronavirus genus contain a fifth structural gene, hemagglutinin-esterase (HE) .

Interspersed in the 3’ region are genes encoding non-structural (NS) accessory proteins, which are unique to distinct genera and subgenera of coronaviruses and often mediate critical interactions between the virus and host innate immune pathways.

As such, these genes and the proteins they encode have been of particular interest to the Weiss laboratory.

As with other positive sense ssRNA viruses, coronavirus replication and transcription occurs at ER-derived and localized replication-transcription complexes (RTCs) that are formed by several of the non-structural proteins and contain the RNA-dependent RNA polymerase (nsp12), RNA primase (nsp8), RNA helicase (nsp13), and RNA proofreading exonuclease (nsp15) among other viral and host proteins .

While ORF1ab translation occurs directly from the genome, synthesis of all other proteins requires transcription of nested subgenomic mRNAs which is regulated by a transcription regulatory sequence (TRS) upstream of each structural and accessory gene .

The first step in this unusual process is transcription of a negative-sense subgenomic (sg) RNA corresponding to each gene.

The negative-sense sgRNA is joined to a leader sequence from the 5’ end of the genome and used as a template for viral mRNA synthesis, with in most cases only the 5’ gene of each subgenomic mRNA being translated.

Coronaviridae phylogeny

The taxonomic organization of the order Nidovirales was substantially revised in July 2018 by the International Committee on the Taxonomy of Viruses. Previously, the family Coronaviridae comprised two sub-families, Coronavirinae and Torovirinae which contained the coronaviruses and toroviruses, respectively. Under the 2018 revision, Nidovirales is now divided into seven sub-orders with the coronaviruses reclassified into the sub-order Cornidovirineae, family Coronaviridae, sub-family Orthocoronavirinae (Fig 1.1).

Subordinate to this sub- family, they are further divided into four genera: Alphacoronavirus, Betacoronavirus, Gammacoronavirus, and Deltacoronavirus that are further separated into numerous subgenera (Fig 1.1).

The majority of human and agricultural pathogens among the coronaviruses fall within Alphacoronavirus and Betacoronavirus, divided into twelve and five subgenera, respectively.

The human upper respiratory tract pathogen human coronavirus 229E (HCoV-229E) and HCoV-NL63 are among the alphacoronaviruses, along with the highly virulent agricultural pathogen porcine epidemic diarrhea virus (PEDV) .

Betacoronavirus had previously been divided into four lineages (A-D), but now consists of the subgenera Embecovirus (HCoV-OC43, HCoV-HKU1, mouse hepatitis virus/MHV), Hibecovirus (bat Hp-betacoronavirus Zhejiang 2013), Merbecovirus (MERS-CoV, MERS-like bat coronaviruses), Nobecovirus (Rousettus bat coronavirus HKU9), and Sarbecovirus (SARS-CoV, SARS-like bat coronaviruses).

The distinction between the Betacoronavirus subgenera lies in the degree of sequence similarity between their replicase genes and the suite of accessory genes they contain, while the basic genomic organization is highly conserved.

Aside from providing taxonomic classification, the accessory genes likely underlie virus-host interactions that are unique between subgenera.

Further, they are likely acquired from host organisms or via horizontal gene transfer with other coronaviruses or unrelated viruses, as appears to have occurred with a recently discovered deltacoronavirus that contains a small reovirus-derived open reading frame .

The toroviruses, previously grouped into Coronaviridae with the coronaviruses, are now in their own suborder Tornidovirineae, family Tobaniviridae, and subfamily Torovirinae, which contains the genus Torovirus.

These viruses are primarily characterized as agricultural pathogens, the best studied of which is equine torovirus, with the most intensively studied variant known as Berne virus (BEV) .

SARS-CoV – First pandemic of the 21^st century

For over forty years after the discoveries of HCoV-OC43 and HCoV-229E, coronaviruses were considered only mildly important to human health.

The discovery of SARS-CoV in 2003 and its rapid global spread dramatically changed this paradigm.

The first indications that a new infectious threat had emerged came from reports of a large outbreak of atypical pneumonia in Guangdong Province, southern China, in November 2002 .

This disease, termed severe acute respiratory syndrome (SARS) before its etiological agent had been identified, spread from Guangdong to Hong Kong, where a remarkable “super spreader” event took place on February 21, 2003 that led to large outbreaks in Canada, Vietnam, and Singapore (39-42).

Ultimately, 8,096 cases and 774 deaths were recorded in 27 countries by the time the outbreak was declared to be over on July 31, 2003.

Upon the recognition of a novel respiratory disease efforts began to identify the etiological agent and its source. By May 2003 three independent groups, in Hong Kong, Germany, and the United States had identified a novel coronavirus as the likely cause of SARS.

This determination was made based on morphology of the virus, partial genome sequencing, and virus isolation in cell culture.

The rapid identification of SARS-CoV and its receptor, angiotensin converting enzyme-2 (ACE-2) would ultimately serve as a model for later studies that rapidly identified MERS-CoV and its cellular receptor.

While the outbreak was still ongoing, numerous SARS-CoV isolates were fully sequenced and characterized , representing one of the first examples of modern molecular biology being brought to bear against a novel infectious disease threat.

Modern molecular biology techniques were combined with classic epidemiology to identify the source of the outbreak, a critical step in bringing it to an end.

Experts quickly suspected that SARS-CoV was a zoonotic agent, transmitted from animals to humans. As early SARS-CoV cases were particularly common among restaurant workers who handled exotic game animals, suspicions about the source turned to live animal markets in southern China.

SARS-CoV was isolated from palm civets and raccoon dogs in a single market, and there was a high prevalence of SARS-CoV antibodies found in animal handlers working in the same market .

A second, and the last known SARS- CoV outbreak at the end of 2003 provided further evidence for civet-to-human transmission, as employees and civets from the same restaurant were found to be infected with virtually identical SARS-CoV isolates .

Although a strong link was drawn between civet and human SARS-CoV infections, extensive surveys failed to find SARS-CoV in civets outside animal markets, suggesting they were not the true wild reservoir , as did the observation that SARS-CoV infection of civets results in overt disease.

Nevertheless, the live animal markets clearly served to amplify transmission between civets and humans, and shutting down these markets was critical to ending the epidemic.

Because various species of bat had previously been identified as reservoirs of zoonotic viruses such as Ebola , Marburg , Nipah , and Hendra viruses, they were also considered possible reservoirs of SARS- CoV.

A 2004 survey of more than 400 bats from four different regions of China found high seroprevalence and wide geographical distribution of anti-SARS-CoV antibodies in bats from the genus Rhinolophus, or horseshoe bats , as did a contemporary study conducted in Hong Kong .

Both reported the first full genome sequences of SARS-like coronaviruses (SL-CoVs), showing a much greater degree of genetic diversity among these viruses than among civet and human SARS-CoV isolates.

Notably, the bat SL-CoV genomes have an intact ORF8, similar to civet and early human isolates, while later human isolates feature a 29 nucleotide deletion in this region .

Continued studies into the prevalence of SL-CoVs in Chinese bats has provided significant insight into the potential for a re-emergence of SARS-CoV or spillover of SL-CoVs into the human population.

SARS-CoV is unusual among zoonotic viruses in that has not caused any additional human outbreaks, aside from a small cluster of cases in late 2003.

However, SL-CoVs with the potential to infect humans continue to circulate among bat populations in China. In 2013 the first bat SL-CoV that can infect human cells via ACE-2 was isolated from Rhinolophus sinicus bats and genome sequences of other viruses predicted to do were found in the same population.

Serological evidence from humans in southern China suggests that SL-CoVs do infect humans without erupting into large outbreaks.

Whether these infections result in any clinical disease is unclear, but experimental evidence suggests that ACE-2 utilizing SL-CoVs, while capable of infecting human cells, may be less virulent than SARS-CoV . The potential for SARS-CoV itself to re-emerge remains unclear. Unlike Marburg or Nipah viruses, SARS-CoV itself has not been found in a wild reservoir.

The simplest explanation for why it has not re-emerged may be that it is extinct in the wild, and recent evidence suggests it arose due to a series of recombination events in a bat population rich with SL-CoVsthat may only rarely recombine to produce a virus with the potential to cause human outbreaks.

Whatever the reason for the apparent disappearance of SARS-CoV, the diversity of SL-CoVs in the wild makes clear that the threat of such viruses remains, and the emergence of MERS-CoV in 2012 demonstrates that this threat is far from geographically isolated.

Coronavirus interactions with the OAS-RNase L antiviral pathway

OAS-RNase L activation by viral dsRNA

Like all viruses, coronaviruses employ mechanisms to evade, suppress or otherwise counteract host innate immune responses.

The most important aspect of this response early in infection is the interferon (IFN)-dependent antiviral response, which is triggered by host sensing of viral dsRNA. Without evasion or antagonism of this pathway, a virus is unlikely to establish a productive infection and replicate to a sufficient level that transmission to new hosts can occur.

IFNs are a large class of cytokines, with type I (IFNa and IFNb) and type III (IFNl) IFNs being most critical for initiating innate antiviral immune responses .

Though both classes of IFNs induce expression of a similar suite of IFN stimulated genes (ISGs), IFNl plays a dominant role at mucosal barriers such as the airway and intestinal tract (, where it is preferentially produced by and acts upon epithelial cells . In the context of human coronavirus infections which primarily occur in the airway, IFN may therefore play a predominant role in the early antiviral response.

The expression of these IFN genes occurs downstream of viral dsRNA sensing by host RIG-I-like receptors (Fig 1.2), specifically MDA5 in the context of coronavirus infection .

Cytoplasmic dsRNA is a hallmark activator of the antiviral response because it is rare in uninfected cells or modified by the cellular adenosine deaminase ADAR1 such that it loses the capacity to activate antiviral signaling .

IFN signaling induces the expression of up to several hundred ISGs, including other dsRNA binding proteins such as protein kinase R (PKR), and oligoadenylate synthetases (OAS), though high basal expression of PKR and OAS genes in some cell types allows their activation in the absence of additional IFN .

OAS proteins, upon binding dsRNA, catalyze the synthesis of a small second-messenger molecule 2’,5’- oligoadenylate (2-5A) which binds to ribonuclease L (RNase L) monomers, inducing their dimerization and activation. Activated RNase L targets singlestranded (ss) RNA of both cellular and viral origin and as such can have direct antiviral effects , cause translational arrest , and lead to cell death .

Through the combination of these events downstream of its activation, RNase L can potently restrict the replication of diverse RNA and DNA viruses . These include flaviviruses such as West Nile virus , hepatitis C virus , and Sindbis virus (SINV) . Consequently, numerous viruses encode well characterized antagonists of this pathway.

These include dsRNA binding proteins such as E3L encoded by Vaccinia virus (VACV) and the NS1 protein of influenza A viruses .

Other known RNase L antagonists include the L* protein encoded by Theiler’s murine encephalomyelitis virus and the VACV D9 protein, an mRNA decapping enzyme .

The Weiss lab has extensively studied another class of OAS-RNase L antagonists, the host-derived phosphodiesterases (PDEs) encoded by select rotaviruses and, most significantly, by mouse hepatitis virus (MHV).

Coronavirus Transcription, Replication and Recombination
CoVs are enveloped viruses that have positive-sense, non-segmented RNA genomes 27–32 kb in length. The basic gene organization and replication is similar for all CoVs and is illustrated
for SARS-CoV (Fig. 1).

Gene 1 encodes all predicted replicase/transcriptase proteins, which are translated from
input genome RNA (RNA1).

Genes 2–9 encode structural and accessory proteins, which are translated from separate subgenomic (sg) mRNAs. CoVs, as members of the Nidovirales order, generate not only new genome RNA, but also a 3′-nested set (Nido = nest) of subgenomic mRNAs (sgRNAs). Along with a portion of the 3′ genome sequence, each CoV sgRNA also contains the first approximately 70 nt of the 5′ leader sequence (Fig. 2).

Coronavirus RNA synthesis can be conceptualized as involving two stages: genome replication and subgenomic RNA
transcription.

In genome replication, the plus-strand genome RNA is transcribed into a full-length minus-strand template RNA, and then significantly more plus-strand genome RNAs are synthesized from that minus-strand template.

In subgenomic RNA transcription, 3′-nested subgenomic RNAs are transcribed to serve as templates for translation of the viral structural and accessory proteins.

This stage of viral RNA synthesis involves a discontinuous RNA transcription model termed transcription attenuation during negative strand RNA synthesis.

During negative strand synthesis, the viral RdRp recognizes virus-specific conserved sequences termed transcriptional regulatory sequences (TRSs), located just upstream of
each subgenomic ORF.

At these points, the polymerase either reads through to the next TRS or dissociates from the template strand, then re-associates with the leader TRS, located in the 5′ UTR, and completes synthesis of a set of subgenomic length negative strand RNA containing an antileader RNA
sequence and equivalent in size to each viral mRNA.

These subgenomic negative strand RNAs then function as the principal templates for the production of subgenomic mRNAs

that are 3′-coterminal, and that all possess an ~70 nt 5′ leader sequence.

Because of this transcription mechanism, alterations in TRS sequences can influence viral replication efficiency.

Importantly, the primary TRS sequence seems less important, as recombinant viruses engineered to encode completely new TRS networks are viable, suggesting that regulatory sequences flanking the TRS elements are critical regulators of subgenomic transcription.

Recently, Wu and Brian have shown that artificial, marked positive-sense subgenomic mRNAs can function as templates for minus strand synthesis and likely contribute to amplifying the amounts of plus-strand sg mRNAs synthesized during infection.

Moreover, mRNAs also can serve as templates for the synthesis of smaller sg mRNAs by recognition of
internal TRS elements as well.

The specific mechanism that confers the capacity of the polymerase to dissociate and reassociate is not well understood, but is thought to be mediated by complementary binding of the anti-TRS on the nascent minus strand with the leader template TRS on the plus strand.

As rare misaligned leader-body junctions are occasionally seen during transcription, it is possible that one or more unique RNA modifying activities, like nsp14 ExoN which encodes a
3′-5′ exonuclease activity (see below), may process the ends of incomplete negative strand RNAs to promote base-pairing and the priming of antileader RNA synthesis.

Coronaviruses and Recombination

Recombination has shaped the population genetic structure of coronaviruses, promoting cross-species transmission and patho- genesis while complicating vaccine  design.

Coronaviruses are quite capable of mediating homologous RNA recombination, with rates approaching 20% during mixed infection of closely related strains in the same group.

This high recombination frequency is likely due to the large size of the genome, paired with replication machinery that is already equipped to dissociate and reassociate from the template RNA (site-assisted copy choice recombination), as well as the availability of full-length and subgenomic-length strands for template switching.

This view Is supported by studies showing targeted recombination between specially designed mutant subgenomic mRNAs and genome length templates.

Viable mutant CoVs can be recovered with artificial TRS sequences, as long as the leader and intergenic TRS sequences match. Further, these artificial TRS sequence viruses prevent recombination with virus containing the native TRS sequences.

These factors combine to allow for the predicted rapid evolution of structural genes, especially within the Spike gene, which undergoes high positive selective pressure during emergence and host-shift events.

Limits of Replication Fidelity and Mutational Diversity in RNA Viruses

The potential for genetic variability of RNA viruses has long been considered to be fundamental to their evolution, adaptation and escape from host responses. However, the effects of changes in replication fidelity, susceptibility to accumulation of deleterious mutations and lethal mutagenesis are not well studied for many viruses.

Genetic determinants including size of genome and pres- ence of repair mechanisms such as proofreading, replicase fidelity and recombination, as well as other as yet undetermined factors may have evolved quite differently in distinct  virus  families.

The high mutation rates of RNA viruses also render them par- ticularly susceptible to repeated genetic bottleneck events during replication, transmission between hosts or spread within a host, resulting in progressive deviation from the consensus sequence associated with decreased viral fitness and sometimes extinction.

The process by which populations of asexual organisms tend to accumulate deleterious mutations in the absence of recombination is referred to as Muller’s ratchet.

Muller’s ratchet has been shown to be applicable to multiple RNA viruses during plaque-to-plaque passage  and to result in accumulation  of mutations and lethal mutagenesis and extinction of plaque- passaged viruses.

For example, some FMDV clones are suscep- tible to genetic bottleneck-mediated extinction, while others are resistant.

Mutagenesis has also been proposed to work as an antiviral strategy.

A major mechanism of action of ribavirin and other RNA mutagens is lethal mutagenesis, as demonstrated with poliovirus, and other RNA viruses, including HIV, FMDV, LCMV and Hantaan virus.

In contrast, SARS-CoV  is highly resistant to ribavirin in vitro and in vivo; in fact, in some instances, the drug exacerbates in vivo disease.

Genome Evolution

The availability of mutator phenotype mutants of both MHV and SARS that can tolerate up to 20-fold increase in mutation rate, accumulate massive mutational diversity at the population level and survive extended population and plaque passage, rep- resents a powerful tool to directly test long-standing questions concerning the role of diversity and fidelity in virus replication, pathogenesis and evolution.

How do CoVs maintain a large and complex genome over time while allowing sufficient mutation rates for adaptation and trans-species movement?

Is the fidelity of wt CoV replication fixed by highly selected interactions between the RdRp, ExoN and possibly other viral and cellular proteins or is it flexible in response to altered environmental conditions as has been shown for bacteria such as E. coli?

What are the limitations to CoV genome diversity in vitro and in vivo? Does the ExoN mutator phenotype result in more rapid adaptation or attenua- tion associated with lethal mutational load and rapid extinction during infection in vivo?

Does the mutator phenotype increase susceptibility to mutagens for lethal mutagenesis?

Clearly, CoVs provide a rich empirical platform to address these interesting and unique questions in virus evolution and adaptation.

Replication and Recombination

CoVs use a unique discontinuous mechanism to transcribe a series of progressively larger subgenomic mRNAs, and each con- tains a leader RNA sequence that is derived from the 5′ end of the genome.

RNA recombination and coronavirus subgenomic transcription occur by template switching mechanisms, which can occur either by sequence-assisted base pairing and hydrogen bonding networks or sequence independent processes in mismatched regions.

Given the potential need to appropriately process and match the 3′ ends of nascent RNAs and their templates, the ExoN mutator phenotype provides a novel approach to investigate the role of fidelity in regulating both full length and subgenomic length RNA synthesis, transcription attenuation during negative strand RNA synthesis and as potential regulators of RNA recombination distributions and frequencies across the genome.

The possibility that ExoN interacts with other RNA modifying enzymes such as nsp15 EndoU and nsp16 2′-O-MTase or interact with putative RNA processivity components encoded in nsp7 and 8 to modify RNA repair rates or recombination fre- quencies may serve as a rich environment for future research.

Experimental Evolution

Of the 335 new emerging infectious diseases that were identi- fied between 1940–2004, 60.3 percent were zoonoses, 71.8 per- cent of which originated in wildlife.

As pathogen emergence  has also been increasing overtime, coupled with greater rates of global dissemination, the threat to global health and economies is profound.

Strategies that can identify viral threats that emerge as a consequence of advantageous mutations in response to select evolutionary pressure(s) would provide profound advances in the ability of the Global Health Response Network to control emerg- ing diseases.

The existence of a defined ExoN-mediated muta- tor phenotype may allow for mechanistic insights and modeling of the mutation repertoires that govern:

(a) the rapid selection  of host range expanded mutant viruses which represent precursors to future epidemic emergences;

(b) pathways of escape from therapeutic human monoclonal antibodies and drugs;

(c) limits of genome variation and stability; and

(d) replicase mutations and interacting networks which restore fidelity in passaged ExoN mutant viruses.

For example, CoV phylogeny is punctuated by numerous shifts between host species and cross-species trans- mission is readily achieved in co-cultures and during persistent infections in vitro.

In nature, human coronavirus (HCoV) OC43 emerged around 1,900 from closely related bovine CoVs, whereas HCoV 229E likely emerged from closely related group 1 bat coronaviruses around 1,800 in Africa, leading some to propose that nearly all human and animal CoVs originated from a vast reservoir of strains circulating in bat species.

SARS-CoV is also a zoonotic virus that crossed the species barrier, most likely originating from bats, following amplification in other species (civet cats, raccoon dogs), prior to transmission to humans.

Our group has used synthetic biology to reconstruct full-length genomes of SARS-like bat CoV precursors to the 2002–2003 epidemic strains.

Although these strains replicate but do not spread because they lack the appropriate receptor-binding domain, recombinant bat coronaviruses harboring the SARS- CoV receptor binding domain (RBD) replicate efficiently  and use angiotensin 1 converting enzyme 2 (ACE2) as a receptor for docking and entry.

These data suggest that the trimeric spike glycoprotein of CoVs may be plastic and modular in design, readily allowing for protein domains to be exchanged between divergent S glycoproteins from different strains.

We propose that blend- ing the ExoN mutator phenotype into synthetically reconstructed zoonotic viruses provides a strategy to rapidly identify pathway components and mutation sets that govern trans-species movement in cell or organ cultures and in vivo.

We hypothesize that the CoV ExoN mutator phenotype constitutes a robust investiga- tive platform to predict mutations and possibly recombinants in advance of their occurrence by identifying advantageous muta- tions governing host range and virus cross-species transmission.

Pathogenesis

Genetic diversity within a quasispecies has been proposed to contribute to pathogenesis by cooperative interactions among engineered variant viruses within a population.

However, the impact of reduced replication fidelity on pathogenesis remains largely untested for RNA viruses. The CoV ExoN mutator phenotype represents a unique property with unclear consequences for adaptation and pathogenesis in animals.

Although increased fidelity attenuates virulence of poliovirus and restricts tissue tropism, the ExoN mutator activity might increase virus virulence and tissue tropism because of the increased population diversity and spread into novel tissues.

Alternatively, ExoN decreased fidelity might attenuate virulence because the mutation frequency may approach error catastrophe and drive self-annihilation of S-ExoN in vivo.

The growth defects observed in culture seem to support the prediction that S-ExoN will be attenuated in animals, but these impairments could be trumped by potential enhanced adaptability of S-ExoN.

To assess these possibilities, pathogenesis studies in animal models are currently underway. Of note, a low-fidelity exonuclease mutant of cytomegalovirus showed accelerated evolution of drug resistance in cell culture.

Increased mutation rates in the GII.4 noroviruses have been proposed to account for their epochal evolution, increased diversity and the striking increase in the frequency of human epidemics in winter.

Thus, fidelity regula- tion is a broadly relevant topic with farranging appeal in RNA virus evolution and pathogenesis.

Importantly, ExoN represents an important and unique high impact target for understanding CoV replication fidelity, quasispecies diversity and pathogenesis and is strongly coupled with the potential of developing universal strategies to build safe live attenuated CoV vaccines.

Vaccines and Therapeutics

Live attenuated vaccines elicit strong protective immune responses with low risk of disease, leading to robust tools that protect the public health against pathogens like measles, poliovirus, mumps, smallpox, herpesviruses and rubella.

Safety concerns clearly exist as evidenced by vaccine reversion to virulence and the development of serious and lethal disease in a low percentage of vaccinees.

The present regulatory environment in the US is now limiting live attenuated virus vaccine use because of safety concerns, attesting to the need for rational approaches that prevent reversion to virulence.

The high conservation of nsp14 ExoN sequences among CoVs and lack of close orthologs in cells suggests that nsp14 ExoN may be a promising target for live attenuated virus design or antiviral therapeutics.

Clearly, studying the ExoN mutator phenotype in pathogenesis and as a rational approach to develop reversion-resistant live attenuated vaccines provides a potential rapid response strategy to control future emerging CoV diseases in human and domesticated animals.

Current treatment regimens for SARS-CoV include ribavirin, a nucleoside analog that induces lethal mutagenesis of other RNA viruses such as poliovirus, foot and mouth disease virus, hepatitis C virus among others.

However, its precise mechanism of action against CoVs has not been determined and the high replication fidelity of WT MHV and SARS-CoV in cell culture suggests that drug-induced viral extinction thera- pies employed against other RNA viruses might not be as effective against CoVs.

A possible recalcitrance of CoVs to RNA mutagens is also suggested by our demonstration that at least in cell culture SARS-CoV tolerates a-fold increase in substitution frequency, whereas a two- to six-fold increase in mutation frequency was suf- ficient to cause lethal mutagenesis of poliovirus in cell culture.

Moreover, ribavirin is clearly ineffective against mouse-adapted SARS-CoV and appears to exacerbate disease, suggesting that the ExoN activity in wildtype viruses may reduce the efficacy of this important antiviral.

The high conservation of nsp14 ExoN sequences among CoVs and lack of close orthologs in cells suggests that nsp14 ExoN might represent a promising target for design and development of antiviral drugs and raises the possibility that a single drug targeting ExoN might be effective against multiple coronaviruses, including potential zoonotic SARS-CoV-like viruses from bats that emerge in the future.

However, given numerous examples of viruses evolving drug resistance, an ExoN-targeted companion drug in a combination therapy would not only attenuate pathogenesis by altering error rates, but also prevent reversion from other compounds in the cocktail.

Investigating the potential of ExoN targeted mutations as a universal strategy to construct live attenuated, reversion proof CoV vaccines and antivirals seems broadly relevant.

Studies investigating the pathogenesis of ExoN mutants in animal models, along with their tenability as vaccine candidates, are currently in progress.

Many of the symptoms caused by 2019 -nCoV, such as acute respiratory syndrome, are similar to those caused by SARS coronavirus (SARS -CoV).

Severe acute respiratory syndrome (SARS) coronavirus (SARS-CoV) is a novel virus that caused the first major pandemic of the new millennium (89, 180, 259). The rapid economic growth in southern China has led to an increasing demand for animal proteins including those from exotic game food animals such as civets.

Large numbers and varieties of these wild game mammals in overcrowded cages and the lack of biosecurity measures in wet markets allowed the jumping of this novel virus from animals to human (353, 376). Its capacity for human-to-human transmission, the lack of awareness in hospital infection control, and international air travel facilitated the rapid global dissemination of this agent.

Over 8,000 people were affected, with a crude fatality rate of 10%. The acute and dramatic impact on health care systems, economies, and societies of affected countries within just a few months of early 2003 was unparalleled since the last plague.

The small reemergence of SARS in late 2003 after the resumption of the wildlife market in southern China and the recent discovery of a very similar virus in horseshoe bats, bat SARS-CoV, suggested that SARS can return if conditions are fit for the introduction, mutation, amplification, and transmission of this dangerous virus (45, 190, 215, 347).

Here, we review the biology of the virus in relation to the epidemiology, clinical presentation, pathogenesis, laboratory diagnosis, animal models or hosts, and options for treatment, immunization, and infection control.

TAXONOMY AND VIROLOGY OF SARS-CoV

SARS-CoV is one of 36 coronaviruses in the family Coronaviridae within the order Nidovirales. Members of the Coronaviridae are known to cause respiratory or intestinal infections in humans and other animals (Fig. 1). Despite a marked degree of phylogenetic divergence from other known coronaviruses, SARS-CoV together with bat SARS-CoV are now considered group 2b coronaviruses (190, 282).

Primary isolation of SARS-CoV was achieved by inoculation of patients’ specimens into embryonal monkey kidney cell lines such as FRhK-4 or Vero E6 cell lines, which produced cytopathic changes at foci, where cells become round and refractile within 5 to 14 days (259). These initial cytopathic changes spread throughout the cell monolayers, leading to cell detachment within 24 to 48 h.

Subcultures can be made on Vero (monkey kidney), Huh-7 (liver cancer) (301), CACO-2 (colonic carcinoma) (79) or other colorectal cancer, MvLu (mink lung epithelial) (104), and POEK and PS (pig) cell lines (122).

Transmission electron microscopy of infected cell lines showed characteristic coronavirus particles within dilated cisternae of rough endoplasmic reticulum and double-membrane vesicles. Clusters of extracellular viral particles adhering to the surface of the plasma membrane were also seen.

Negatively stained electron microscopy showed viral particles of 80 to 140 nm with characteristic surface projections of surface proteins from the lipid envelope (89, 180, 259). SARS-CoV has a higher degree of stability in the environment than other known human coronaviruses (91, 276).

It can survive for at least 2 to 3 days on dry surfaces at room temperature and 2 to 4 days in stool (276). The electron microscopic appearance and genome order of 5′-replicase (Orf1ab)-structural proteins (spike [S]-envelope [E]-membrane [M]-nucleocapsid [N])-poly(T)-3′ are similar to those of other members of the Coronaviridae (236).

Similar to other coronaviruses, it is an enveloped positive-sense single-stranded RNA virus with a genome size of almost 30 kb (Fig. 2). The genome is predicted to have 14 functional open reading frames (ORFs) (290). Their functions and putative roles are outlined in Table 1. Two large 5′-terminal ORFs, ORFs 1a and 1b, encode 16 nonstructural proteins, 7 of which are likely to be involved in the transcription and replication of the largest genome among all RNA viruses (92, 95, 158, 166, 242, 284, 309, 316, 343, 414). The two proteases are involved in posttranslational proteolytic processing of the viral polyprotein (5, 15, 121, 224, 394). The surface S protein is involved in the attachment and entry of the host cell and is therefore the main target for neutralizing antibody and antiviral peptides (159, 206, 227, 301, 334). N together with M, E, and Orf7a are involved in the assembly of the virion (97, 147, 150, 245, 359).

Orf3a is an ion channel protein that is likely to be involved in viral budding and release (234). Analysis of genome sequences of many isolates of SARS-CoV from humans with civet SARS-CoV and bat SARS-CoV showed that the most variable genes with nucleotide homologies of less than 90% are the S gene, Orf3, Orf8, nsp2, nsp3, and nsp4 (190, 215, 282).

Deletions of 82 and 415 nucleotides in Orf8 were found in some human isolates, whereas a unique 29-nucleotide signature insertion in Orf8 can be found in animal isolates (64, 117). Therefore, the more conserved Orf1b is generally chosen to be the molecular target for the design of clinical diagnostic tests rather than these less conserved regions.

TABLE 1.

Nomenclature and functional characteristics of SARS-CoV gene products and their interactions with host cells in disease pathogenesis

Gene nomenclature (no. of amino acid residues in product)	Gene product and/or characteristic(s) (reference[s])	Effect on cellular response of host (reference[s])
Orf1a/b
nsp1 (180)	Expression promoted degradation of host endogenous mRNAs, which may inhibit host protein synthesis and prevented endogenous IFN-β mRNA accumulation (167)	Induce CCL5, CXCL10 (IP10), and CCL3 expression in human lung epithelial cells via activation of NF-κB; increases cellular RNA degradation, which might facilitate SARS-CoV replication or block immune responses (81, 192)
nsp2 (638)	Deletion attenuates viral growth and RNA synthesis (106)
nsp3 (1,922)	Papain-like protease 2; proteolytic processing of the viral polyprotein at 3 sites and participation in synthesis of subgenomic RNA segment (15, 121, 224)	Putative catalytic triad (Cys1651-His1812-Asp1826) and zinc-binding site have deubiquitinating activity; this unexpected activity in addition to its papain-like protease suggests a novel viral strategy to modulate the host cell ubiquitination machinery to its advantage (15, 224, 279)
	ADP-ribose 1-phosphatase; dephosphorylates Appr-1‴-p, a side product of cellular tRNA splicing, to ADP-ribose (271)
nsp4 (500)	Not known
nsp5 (306)	3C-like protease; proteolytic processing of the replicative polyprotein at 11 specific sites and forming key functional enzymes such as replicase and helicase (5, 394)	Growth arrest and apoptosis via caspase-3 and caspase-9 activities demonstrated in SARS-CoV 3CLpro-expressing human promonocyte cells with increased activation of the nuclear factor-κB-dependent reporter (222)
nsp6 (290)	Not known
nsp7 (83)	Three-dimensional structure by nuclear magnetic resonance study found potential sites for protein-protein interactions (261)
nsp8 (198)	Putative RNA-dependent RNA polymerase; crystal structure of the hexadecameric nsp7-nsp8 possesses a central channel with dimensions and positive electrostatic properties favorable for nucleic acid binding; it is probably another unique RNA-dependent RNA polymerase for its large genome (158, 414)
nsp9 (113)	Three-dimensional crystal structure of a dimer which binds viral RNA and interacts with nsp8 (92, 316)
nsp10 (139)	Crystal structure suggests a nucleic acid binding function within a larger RNA binding protein complex for viral gene transcription and replication (166, 309)	Interacts specifically with the NADH 4L subunit and cytochrome oxidase II with depolarization of inner mitochondrial membrane of transfected human embryo lung fibroblast and extensive cytopathic effect (210)
nsp11 (13)	Not known
nsp12 (932)	RNA-dependent RNA polymerase; replication and transcription to produce genome- and subgenome-sized RNAs of both polarities (158)
nsp13 (601)	Helicase (dNTPase and RNA 5′-triphosphatase activities) (95)
nsp14 (527)	3′→5′-exoribonuclease; this unusual 3′→5′-exoribonuclease activity supplements the endoribonuclease activity in the replication of the giant RNA genome (242)
nsp15 (346)	Uridylate-specific endoribonuclease; RNA endonuclease that is critically involved in the coronavirus replication cycle (284)
nsp16 (298)	Putative 2′-O-ribose methyltransferase (343)
Orf2 (1,255)	Spike protein; binds to the host cell receptor ACE2 and other coreceptors, mediates viral entry into host cells as a type 1 viral fusion protein; required acidification of endosomes for efficient S-mediated viral entry; proteolytic cleavage by abundantly expressed infected cell membrane-associated factor Xa into S1 and S2; protease activation required for cell-cell fusion (159, 162, 206, 214, 227, 301, 334)	293 T cells transfected with ACE2 can form multinucleated syncytia with cells expressing the spike; intraperitoneal injections of spike protein into mice reduced ACE2 expression in lungs and worsened acute lung failure in vivo that can be attenuated by blocking the renin-angiotensin pathway (181); recombinant baculovirus expressing different deletion and insertion fragments identified the functional region of S protein from amino acids 324-688, which can induce the release of IL-8 in lung cells (43); induces unfolded protein response in cultured cells as SARS-CoV with a substantial amt of S protein accumulation in the endoplasmic reticulum, which may modulate viral replication (30)
Orf3a (274)	Forms potassium-sensitive ion channel, may promote virus budding and release (234)	Overexpression in cell line may trigger apoptosis; its expression in A549 lung epithelial cells up-regulates mRNA and intracellular and secreted levels of all three subunits, alpha, beta, and gamma, of fibrinogen, which is also observed in SARS-CoV-infected Vero E6 cells; it is highly immunogenic and induces neutralizing antibodies (193, 321); 3a/X1 and 7a/X4 were capable of activating NF-κB and c-Jun N-terminal kinase and significantly enhanced IL-8 promoter activity in A549 cells; enhanced production of inflammatory chemokines that were known to be up-regulated in SARS-CoV infection (169)
Orf3b (154)	Predominately localized to the nucleolus in different transfected cells (409)	Vero E6 but not 293T cells transfected with a construct for expressing Orf3b underwent necrosis as early as 6 h after transfection but underwent simultaneous necrosis and apoptosis at later time points; Orf3b inhibits expression of IFN-β at synthesis and signaling (175, 178)
Orf4 (76)	Envelope protein; synthetic peptides form ion channels in planar lipid bilayers, which are more permeable to monovalent cations than to monovalent anions; putatively involved in viral budding and release (359)	Induced apoptosis in transfected Jurkat T cells especially in the absence of growth factors; a novel BH3-like region was located in the C-terminal cytosolic domain of SARS-CoV E protein can bind to Bcl-xL, whose overexpression can antagonize apoptosis; this may explain the consistent lymphopenia found in SARS patients (397)
Orf5 (221)	Membrane protein; surface protein responsible for viral assembly and budding	M protein induced apoptosis in HEK293T cells, which could be suppressed by caspase inhibitors (29)
Orf6 (63)	Novel membrane protein that accelerates replication and virulence of a recombinant mouse coronavirus expressing Orf6; an important virulence factor in vivo demonstrated in a mouse model (327)	Inhibits both IFN synthesis and signaling; inhibited nuclear translocation but not phosphorylation of STAT1 (178); Orf6 is localized to the endoplasmic reticulum/Golgi membrane of infected cells; it binds and disrupts nuclear import complex formation by tethering karyopherin alpha 2 and karyopherin beta 1 to the membrane; this retention of the complex at the endoplasmic reticulum/Golgi membrane leads to a loss of STAT1 transport into the nucleus despite viral RNA-induced IFN signaling; thus, it blocks the expression of STAT1-activated genes, which are essential for establishing an antiviral state (100)
Orf7a (122)	Unique type I transmembrane protein; involved in viral assembly by interacting with M and E, which are essential for virus-like particle formation when coexpressed with S and N (97, 150, 245)	Expression of Orf7a induces apoptosis via a caspase-3-dependent pathway and in cell lines derived from different organs including lung, kidney, and liver (179, 320, 408)
Orf7b (44)	Not known
Orf8a (39)	Not known	Orf8a was localized in mitochondria, and overexpression resulted in increases in mitochondrial transmembrane potential, reactive oxygen species production, caspase-3 activity, and cellular apoptosis; Orf8a enhances viral replication and induces apoptosis through a mitochondrion-dependent pathway (49)
Orf8b (84)	May modulate viral replication; expression of E was down-regulated by Orf8b but not Orf8a or Orf8ab (172)
Orf9 (422)	Nucleocapsid protein; binding and packaging of viral RNA in assembly of the virion (147)	N antagonized IFN by inhibiting synthesis of IFN-β (130); NF-κB activation in Vero E6 cells expressing the N protein is dose dependent (220); N may cause inflammation of the lungs by activating COX-2 gene expression by binding directly to the promoter, resulting in inflammation through multiple COX-2 signaling cascades (393); induced apoptosis of COS-1 monkey kidney but not 293T cells in the absence of growth factors; induced actin reorganization in cells devoid of growth factors (315)
Orf9b (98)	Crystal structure of Orf9b, an alternative ORF within the N gene, may be involved in membrane attachment and associates with intracellular vesicles, consistent with a role in assembly of the virion (241)

VIRAL LIFE CYCLE

Trimers of the S protein form the peplomers that radiate from the lipid envelope and give the virus a characteristic corona solis-like appearance under an electron microscope. S is a class I fusion protein that consists of the amino-terminal S1 and carboxyl-terminal S2 subunits connected by a fusion peptide.

The two subunits are indispensable for receptor binding and membrane fusion, respectively. The receptor binding domain of S1 has been mapped to residues 318 to 510 (9, 365). The binding of S1 to the cellular receptor will trigger conformational changes, which collocate the fusion peptide upstream of the two heptad repeats of S2 to the transmembrane domain, and, finally, fusion of the viral and cellular lipid envelopes.

Moreover, this process could be facilitated by the infected cell membrane-associated protease, such as factor Xa, which can cleave S into S1 and S2. This proteolytic cleavage is specifically inhibited by a protease inhibitor, Ben-HCl (90).

The key receptor of the host cell attached by S is angiotensin-converting enzyme 2 (ACE2), which is a metalloprotease expressed in the cells of the lung, intestine, liver, heart, vascular endothelium, testis, and kidney (119).

Since ACE2 was shown to protect against acute lung injury in a mouse model and since the binding of the S protein to host cells results in the downregulation of ACE2, this mechanism may contribute to the severity of lung damage in SARS (181). Cells expressing some lectins, including DC-SIGN, L-SIGN, and LSECtin, have been shown to augment the cellular entry of pseudotype virus expressing S but only in the concomitant presence of ACE2 (40, 107, 162, 398).

Nonsusceptible cells expressing these lectins in the absence of ACE2, such as dendritic cells, were able to promote the cell-mediated transfer of SARS-CoV to susceptible cells (40). Although lysosomotropic agents can block viral entry, which indicates that endosomal acidification is required for entry, the activation of the S protein by protease can bypass this inhibition and result in cell-to-cell fusion.

Despite the role of the pH-sensitive endosomal protease cathepsin L in the entry pathway (151, 300), viral culture does not require pretreatment with trypsin. However, this pH-sensitive cathepsin L may be a target for agents such as chloroquine, which elevates endosomal pH (174, 341).

The process of viral disassembly in the cytoplasm for the release of viral RNA for translation and replication remains elusive. Translation starts with two large polyproteins from Orf1a and Orf1ab, which are posttranslationally cleaved by the two viral proteases into nsp1 to nsp16.

These cleavage products form the replication-transcription complex, which replicates the viral genome and transcribes a 3′-coterminal nested set of eight subgenomic RNAs. It is therefore conceivable that infected cells contain a higher number of transcripts containing genes towards the 3′ terminus of the viral genome. On this basis, reverse transcriptase PCR (RT-PCR) using the N gene may have a better sensitivity than those using the other genes.

As in other coronaviruses, SARS-CoV may attach by the hydrophobic domains of their replication machinery to the limiting membrane of autophagosomes and form double-membrane vesicles.

Once sufficient viral genomic RNA and structural proteins are accumulated, viral assembly by budding of the helical nucleocapsid at the endoplasmic reticulum to the Golgi intermediate compartment occurs. Here, the triple-membrane-spanning M protein interacts with the N protein and viral RNA to generate the basic structure.

It also interacts with the E and S proteins to induce viral budding and release. Unlike other coronaviruses, the M protein of SARS-CoV also incorporates another triple-membrane-spanning protein of Orf3a into the virion (161).

The N protein is the most abundantly expressed viral protein in infected cells in which the mRNA levels were amplified 3 to 10 times higher at 12 h postinfection than other structural genes (138) and is therefore an important target for immunohistochemistry and antigen detection in clinical specimens. Various diagnostic tests, antiviral agents, and vaccines are designed on the basis of our understanding of the structure and function of the various viral proteins involved in the life cycle of this virus.

SEQUENCE OF THE SARS EPIDEMIC AND MOLECULAR EVOLUTION OF THE VIRUS

Sequence of EventsSARS was the first known major pandemic caused by a coronavirus. During the epidemic in 2003, 8,096 cases with 774 deaths had occurred in over 30 countries among five continents (89, 117, 144, 180, 182, 197, 236, 250, 259, 260, 270, 290, 292, 303, 336, 377). The disease emerged in late 2002, when an outbreak of acute community-acquired atypical pneumonia syndrome was first noticed in the Guangdong Province (Table 2).

Retrospective surveillance revealed severe cases of the disease in five cities around Guangzhou over a period of 2 months (431). The index case was reported in Foshan, a city 24 km away from Guangzhou. The second case involved a chef from Heyuan who worked in a restaurant in Shenzhen. The patient had regular contact with wild game food animals.

His wife, two sisters, and seven hospital staff members who had contact with him were also affected. From 16 November 2002 to 9 February 2003, a total of 305 cases were reported in mainland China, with 105 of those cases involving health care workers. The devastating pandemic started in Hong Kong, Special Administrative Region (HKSAR), when a professor of nephrology from a teaching hospital in Guangzhou who had acquired the disease from his patients came to HKSAR on 21 February 2003. Within a day, he transmitted the infection to 16 other people in the hotel where he resided. His brother-in-law, one of the secondary cases, underwent an open lung biopsy from which the etiological agent was discovered and first isolated (259). It was a novel coronavirus, named SARS-CoV.

TABLE 2.

Sequence of events and molecular evolution of SARS-CoV throughout the epidemic^a

The secondary cases unknowingly carried the disease to hospitals in the HKSAR and to other countries and continents including Vietnam, Canada, Singapore, the Philippines, the United Kingdom, the United States, and back again to China.

Carlo Urbani, a physician working at the World Health Organization (WHO) office in Hanoi, Vietnam, was the first to notify the WHO of cases outside Guangdong after witnessing an explosive nosocomial outbreak of SARS in a hospital in Hanoi, which resulted from a person who had returned from the hotel in HKSAR. Carlo Urbani’s description of the disease, to which he later succumbed, alerted health authorities throughout the world and accelerated collaborative research to identify the virus and combat the disease (281).

Molecular EvolutionSoon after the isolation of SARS-CoV, SARS-CoV-like viruses were found in palm civets and a raccoon dog from wild-animal markets in the Guangdong Province of China (117), suggesting that these animals could be the source of human infections. As a result, massive numbers of palm civets were culled to remove sources for the reemergence of SARS in Guangdong in January 2004.

The virus was found in many civets and raccoon dogs from the wildlife market prior to culling but not in over 1,000 civets later sampled at 25 farms in 12 provinces (168). The evolutionary starting point was a prototype group consisting of three viral genome sequences of animal origin. This prototype group representing low-pathogenicity virus has seven single-nucleotide variation (SNV) sites that caused six amino acid changes, at positions 147, 228, 240, 479, 821, and 1080 of the S protein, which were involved in generating the early phase of the 2002 and 2003 epidemic.

One of these was found in the first SARS patient in the subsequent epidemic of 2003 to 2004. A further 14 SNVs caused 11 amino acid residue changes, at positions 360, 462, 472, 480, 487, 609, 613, 665, 743, 765, and 1163. This resulting high-pathogenicity virus group caused the middle phase of the epidemic of 2003.

Finally, the remaining six SNVs caused four amino acid changes, at positions 227, 244, 344, and 778, which resulted in the group of viruses responsible for the late phase and the global epidemic (168).

The neutral mutation rate of this virus during the epidemic in 2003 is almost constant, at around 8 × 10⁻⁶ nt⁻¹ day⁻¹, which is similar to those of most known RNA viruses (64, 304). The most recent common ancestor was estimated to be present around mid-November, which is epidemiologically compatible with the first case of SARS found in Foshan.

After the epidemic was over, a second interspecies-jumping event occurred in late 2003 to early 2004, resulting in the reemergence of four human cases in China (45, 347).

These four cases were believed to be due to an independent interspecies transmission event, instead of residual cases of the major epidemic, because of the much lower affinity for human ACE2 (hACE2) of the S proteins of SARS-CoV isolated from these patients and palm civets than that of the major 2003 epidemic isolates from SARS patients, which utilized both human and palm civet ACE2 efficiently (216).

Since S contains the receptor binding domain for the host receptor and is immunogenic, it is under selection in the host and becomes the most rapidly evolving protein, with most mutations located in the S1 domain and especially the receptor binding domain. Bioinformatic analysis has identified three key amino acid residues at positions 360, 479, and 487 that are responsible for host-specific binding (17).

Most human isolates in the 2003 epidemic have N479 and T487 in their S, whereas most civet isolates have K/R479 and S487. The low affinity of the S proteins bearing K479 and S487 combinations for hACE2 was confirmed by pseudotype binding assays. However, the human and civet isolates of the outbreak of 2003 to 2004 had N479 and S487, which suggested that this is an intermediate stage of mutation of the S protein.

Further change to the N479 and T487 combination will allow efficient human-to-human transmission (275). Apart from the subsequent minor outbreak, three laboratory-associated outbreaks were reported in Singapore, Taiwan, and Beijing from September 2003 to May 2004 (221, 251, 252, 256). In Beijing, the outbreak also involved secondary and tertiary cases.

Phylogenetic analysis of the S protein of 139 SARS-CoV isolates in the Hong Kong outbreak showed that several introductions of viruses had occurred but that only one of them was associated with the major outbreak in HKSAR and the rest of the world (116).

Some of the strains found in the early stages of the outbreak were phylogenetically distinct from the major cluster and were closer to some of the Guangdong and Beijing strains. This concurred with the fact that the index patient of the HKSAR outbreak was a Guangzhou medical doctor who had traveled to HKSAR.

Another molecular epidemiological study of the Guangdong outbreak suggested that the disease spread from Guangdong to HKSAR and the rest of the world, and the index case was a chef who handled game animals (431). Subsequent animal surveillance in China recovered coronavirus isolates that had 99.8% nucleotide identity with SARS-CoV (117).

A characteristic 29-bp insertion between Orf8a and Orf8b (also initially known as Orf10 and Orf11) was found in these animal isolates (117, 302). This 29-nucleotide segment was deleted either before or soon after crossing the species barrier to humans. The biological effect of this deletion remains elusive.

A number of SARS-CoV isolates in the later stages of the epidemic showed larger deletions around this site (64). Two independent molecular epidemiological studies comparing the complete genomes of 12 and 63 virus isolates also found evidence of strong positive selection at the beginning of the epidemic, which was followed by a purifying selection, as indicated by the amino acid substitution rate at S, Orf3a, and nsp3 (64, 304, 402).

Both studies suggested that molecular adaptation of the virus had occurred after interspecies transmission from animals to humans. In the small outbreak in Guangzhou in 2004, all four human isolates belonged to a separate sublineage of the concurrent animal isolates that were distinct from the human pandemic or animal viruses in 2003.

Although SARS-CoV is distinct from the three existing groups of coronaviruses, it may be closer to group II because 19 out of 20 cysteines found in the S1 domain of the S protein are spatially conserved compared with the group II consensus sequence, whereas only five cysteine residues are conserved compared with those of groups I and III (93, 302).

Since coronaviruses are believed to have coevolved with their animal hosts, it is possible that rats, mice, and cattle harboring group II coronaviruses are more likely to be the animal host for SARS-CoV than cats, which harbor group I coronavirus.

However, when a comparison of the phylogenetic trees for 11 known host species and nucleocapsid sequences of 36 coronaviruses was done using an inference approach with sliding-window analysis, there was statistical incongruence, which indicates multiple host species shifts between the coronaviruses of many animals that are phylogenetically distant (283).

Thus, it would not be too unexpected if other mammals are the true animal reservoir rather than mice and rats.

Nevertheless, civets and other related mammals had at least served as a major amplification host in the markets of southern China irrespective of the original animal reservoir. The control of these animals and the markets played a pivotal role in the epidemiological control of SARS (304).

In view of the low rate of detection of SARS-CoV in wild and farm civets (338), in contrast to a very high rate in caged civets in wildlife markets, efforts were made to find the natural reservoir of SARS-CoV in birds, pig, cattle, sheep, mice, and rats, which all turned out to be negative. However, SARS-CoV-like viruses with around 90% genomic identity with SARS-CoV were independently discovered in horseshoe bats (Rhinolophus spp.) in HKSAR and mainland China (190).

The high seroprevalence and viral load of infected Chinese horseshoe bats, Rhinolophus sinicus, strongly suggested that bats are the natural reservoir of SARS-CoV-like viruses, similar to the situation of fruit bats carrying Hendra virus or Nipah virus (363).

EPIDEMIOLOGICAL CHARACTERISTICS

The epidemiological linkage of the initial human cases of the 2003 pandemic to wild game animals suggested that SARS-CoV is zoonotic in origin (431). The isolation of SARS-CoV-like viruses from palm civets and subsequently horseshoe bats further supported this contention (117, 190).

It was reported that a seroprevalence rate of about 80% was found in civets in animal markets in Guangzhou (338).

However, person-to-person transmission has been the primary mode of spread of the epidemic, which has occurred in health care facilities, workplaces, homes, and public transportation.

The most important route of person-to-person spread appears to be direct or indirect contact of the mucosae with infectious respiratory droplets or fomites (296). SARS-CoV has been detected in respiratory secretions, feces, urine, and tears of infected individuals (42, 229).

Nosocomial transmission of SARS was facilitated by the use of nebulizers, suction, intubation, bronchoscopy, or cardiopulmonary resuscitation on SARS patients, when large numbers of infectious droplets were generated (70, 197, 340).

In fact, almost half of the SARS cases in HKSAR were nosocomial infections that were acquired within health care facilities and institutions (202).

The attack rate among health care workers was higher where the number of SARS patients was greater (187). Although airborne transmission is considered uncommon, a unique form of airborne transmission was considered a likely explanation for a large community outbreak in a private housing estate called Amoy Garden in HKSAR.

Contaminated aerosols generated in toilets by exhaust fans coupled with dried U traps of sewage drains, which ascended the light well connecting different floors, caused an explosive outbreak affecting hundreds of people (71, 405).

The presence of viruses in stool, often with high viral loads (156, 258), also suggested the possibility of feco-oral transmission, although this has not been proven conclusively. It was suggested that SARS was transmitted in commercial aircraft during the epidemic. Out of a total of 40 flights investigated, 5 were associated with probable in-flight SARS transmission, affecting 37 passengers (254).

Most of the affected passengers sat within five rows of the index case. The overall risk of transmission appears to be low, at around 1 in 156 (358). In the largest incident, during a 3-h flight carrying 120 passengers traveling from HKSAR to Beijing, a superspreading event (SSE) infected 22 passengers (254).

The pattern of involvement was atypical, considering the short duration of exposure of 3 h and the widespread involvement of patients sitting within seven rows in front of and five rows behind the index case.

Although airborne transmission was considered to be a possible explanation, other potential modes of transmission, such as contact of passengers with the index case before or after the flight, cannot be excluded, especially since 17 out of the 22 people infected were from two tourist groups (254).

In another study, a SARS patient traveled between HKSAR and European countries during the presymptomatic and early symptomatic period, and no transmission among passengers seated in close proximity to the index patient was found, suggesting that in-flight transmission of SARS is not common (23).

Symptomatic SARS patients appeared to transmit infections on board much more readily than presymptomatic ones (23, 254, 358). Initiation of screening procedures to detect people with fever prior to boarding has been used in an attempt to reduce the risk of in-flight transmission of SARS, but the efficacy is still uncertain (342).

In 17 studies that reported on seroepidemiology, the seroprevalence varied from 0 to 1.81% for the general population, 0 to 2.92% for asymptomatic health care workers, 0 to 0.19% for asymptomatic household contacts, and 12.99 to 40% for asymptomatic animal handlers (28, 37, 45, 69, 117, 141, 198, 201, 203, 207, 209, 228, 352, 369, 387, 406, 429).

The last finding is quite expected, since frequent zoonotic challenges by low-level-pathogenic strains of SARS-CoV before 2003 in animal handlers of southern China would probably have caused such a high seroprevalence in this at-risk group.

Genuine asymptomatic infection with antigenemia detected by enzyme immunoassay (EIA) and seroconversion confirmed by neutralization antibody assay was documented in a restaurant worker who worked in the same restaurant as the index case of the outbreak of 2003 to 2004 (45).

However, in 2003, sustained exposure of the animal handlers to these infected civets and other wild animals would result in the introduction of a moderately transmissible and more virulent SARS-CoV strain, which would have mutated from the animal strain and adapted to infect humans more efficiently.

The result was a massive global outbreak, but the overall asymptomatic infection rate was still relatively low with this more virulent human-adapted virus in the general population, health care workers, and household contacts.

A meta-analysis gave overall seroprevalence rates of 0.1% for the general population and 0.23% for health care workers (203). It is also important to remember that these seroprevalence studies are not directly comparable since different serological methods of various sensitivities or specificities were used with or without confirmation by another test. Thus, the true incidence of asymptomatic infection remains elusive.

The incubation period of SARS is 2 to 14 days, although occasional cases with longer incubation periods have been reported (41). The average number of secondary cases resulting from a single case was two to four (225, 285).

Unlike influenza virus, where the patients were most infectious in the first 2 days of illness, transmission from symptomatic SARS patients usually occurred on or after the fifth day of onset of disease, which is in line with the rising viral load in nasopharyngeal secretions that peaked at around day 10 (258).

There have been speculations about the incidence of SARS and ambient temperature (319), but a definite seasonality could not be concluded. SSEs have been noted to play an important role in the propagation of the SARS outbreak, which gives rise to a disproportionate number of secondary cases, as in the Amoy Garden of HKSAR. A study comparing the clinical and environmental features of SSE and non-SSE cases showed that SSEs were likely to be related to a combination of factors including delayed isolation, admission to a nonisolation ward, and severe disease at the time of isolation (53).

CLINICAL FEATURES

The typical clinical presentation of SARS is that of viral pneumonia with rapid respiratory deterioration (Table 3). Fever, chills, myalgia, malaise, and nonproductive cough are the major presenting symptoms, whereas rhinorrhea and sore throat are less frequently seen (7, 21, 37, 149, 197, 258, 259, 270, 278, 336, 411, 425).

Clinical deterioration, often accompanied by watery diarrhea, commonly occurs 1 week after the onset of illness (58, 258). Similar to other causes of atypical pneumonia, physical signs upon chest examination are minimal compared with the radiographical findings. Chest radiographs typically show ground-glass opacities and focal consolidations, especially in the periphery and subpleural regions of the lower zones.

Progressive involvement of both lungs is not uncommon (113, 148, 184, 362). Shifting of radiographic shadows and spontaneous pneumomediastinum may occur (74, 258). A retrospective analysis of serial chest radiographs in all SARS patients from HKSAR showed that the initial extent and progression of radiographic opacities may be useful for prognostic prediction (6).

TABLE 3.

Correlation between clinical, virological, immunological, and histopathological findings

Clinical and laboratory features (% positive isolates [no. of isolates studied/total no.]) (reference)a	Viral load for indicated day(s) after onset of symptoms (reference)	Blood immune profile or histopathological feature (reference)
Systemic involvement	Mean 1.1 log copies/ml between days 10 and 15 in serum (156)	Increased mean serum concentrations of IL-16, TNF-α, and transforming growth factor β1 but decreased IL-18 between days 3 and 27 (16); increased IFN-γ and inflammatory cytokines IL-1, IL-6, and IL-12 for at least 2 wk; chemokine profile demonstrated increased neutrophil chemokine IL-8, MCP-1, and Th1 chemokine IP-10 (360); increased serum concn of IP10, MIG, and IL-8 during the first wk was associated with adverse outcome or death (325)
Fever (99.9 [751/752])
Chill or rigors (51.5 [377/732])
Malaise (58.8 [317/539])
Respiratory involvement	Mean 2.4 log copies/ml between days 10 and 15 for NPA (156), 9.58 × 102-5.93 × 106 copies/ml for throat swab and 7.08 × 102-6.38 × 108 copies/ml for saliva between days 2 and 9 (349), and 2 × 104-1 × 1010 copies/ml between days 5 and 51 for lung tissue (96)	IP10 highly expressed in both lung and lymphoid tissues, with monocyte-macrophage infiltration and depletion of lymphocytes (163); increased alveolar macrophages and CD8 cells, decreased CD4-to-CD8 ratio, and increased TNF-α, IL-6, IL-8, RANTES, and MCP-1 levels in bronchoalveolar lavage samples (124, 344); IP10 was increased in lung tissue from patients who died of SARS (325); increased differential expression of cytokines within these pulmonary tissues, including Stat1, IFN-regulatory factor 1, IL-6, IL-8, and IL-18, often characteristic of patients with acute respiratory distress syndrome (8)
Rhinorrhea (13.8 [50/362])
Sore throat (16.5 [91/552])
Cough (65.5 [460/702])
Dyspnea (45.9 [282/614])
Cardiovascular involvement	1 × 104-2.8 × 107 copies/ml between days 5 and 23 for cardiac tissue (96)	Subclinical diastolic impairment without systolic involvement but no interstitial lymphocytic infiltrate or myocyte necrosis in histology (211); gross pulmonary thromboemboli and marantic cardiac valvular vegetations in some autopsies (67)
Tachycardia (46.1 [71/154])
Bradycardia (14.9 [18/121]) (403)
Hypotension (50.4 [61/121]) (403)
Gastrointestinal involvement	Mean 6.1 log copies/ml between days 10 and 15 for stool (156), with higher mean viral load in NPA obtained on day 10 significantly associated with diarrhea (58); 2.7 × 103-2.7 × 109 copies/ml between days 10 and 29 for small intestinal tissue and 5.3 × 103-3.7 × 108 copies/ml between days 10 and 43 for large intestinal tissue (96)	Minimal architectural disruption despite active viral replication in enterocytes of both terminal ileum and colonic biopsy specimens; no villous atrophy or inflammation (205); atrophy of mucosal lymphoid tissue (298)
Diarrhea (20.1 [130/647])
Other symptoms
Myalgia (48.5 [365/752])		Focal myofiber necrosis with scanty macrophage infiltration may be related to steroid treatment (204)
Headache (38.8 [292/752])	RT-PCR positive for some cerebrospinal fluid (188)	Necrosis of neuron cells and broad hyperplasia of gliocytes (389)
Dizziness (27.3 [163/597])
Hematological involvement		Prolonged lymphopenia with nadir during days 7-9 returning to normal after 5 wk; death and severity are associated with profound CD4+ and CD8+ lymphopenia; little change in CD4/CD8 ratio (136)
Anemia (12.6 [17/135])
Leukopenia (24.2 [114/472])
Lymphopenia (66.4 [296/446])
Thrombocytopenia (29.7 [140/472])
Biochemical involvement
Increased serum alanine aminotransferase levels (44.1 [208/472])	Positive RT-PCR for liver tissue (44), 6 × 103-5 × 104 copies/ml between days 2 and 9 for liver tissue (96)	Ballooning of hepatocytes and mild to moderate lobular lymphocytic infiltration (44)
Impaired serum creatinine (6.7 [36/536]) (76)	Mean 1.3 log copies/ml between days 10 and 15 for urine (156) and 4.3 × 103-7.4 × 105 copies/ml between days 11 and 27 for kidney tissue (96)	Acute tubular necrosis (76)
Decreased serum tri-iodothyronine and thyroxine		Extensive cell apoptosis and exfoliation of the follicular epithelium into distorted, dilated, or collapsed follicles (354)
Other
Histological orchitis (388)		Widespread germ cell destruction, few or no spermatozoa in the seminiferous tubule, thickened basement membrane, and leukocyte infiltration with T lymphocytes and macrophages in the interstitial tissue (388)

• ↵a See references 7, 21, 37, 149, 197, 258, 259, 270, 278, 336, and 425 for clinical and laboratory features unless otherwise specified in the table.

Diarrhea is the most common extrapulmonary manifestation, followed by hepatic dysfunction; dizziness, which may be related to diastolic cardiac impairment and pulmonary arterial thrombosis; abnormal urinalysis; petechiae; myositis; neuromuscular abnormalities; and epileptic fits (44, 58, 188, 211, 248, 335, 346, 383).

The elderly may present atypically without fever or respiratory symptoms (68, 361). While infections in children appear to be milder than those in adults (20, 144, 183), SARS in pregnant women carries a significant risk of mortality (364, 410).

Higher nasopharyngeal and serum viral loads were associated with oxygen desaturation, mechanical ventilation, and mortality; higher stool viral loads were associated with diarrhea; and higher urine viral loads were associated with abnormal urinalysis (58, 75, 156). The significant correlation of the viral loads in these specimens to the severity of clinical or laboratory findings suggested that extrapulmonary viral replication was contributing to clinical manifestations (156).

As for hematological parameters, peripheral blood lymphopenia and elevated hepatic parenchymal enzymes are common with or without thrombocytopenia or increases in D dimers and activated partial thromboplastin time (197).

About 20% to 30% of patients developed respiratory failure requiring mechanical ventilation, and the overall mortality rate was around 15%. Age, presence of comorbidities, increased lactate dehydrogenase level, hypouricemia, acute renal failure, more extensive pulmonary radiological involvement at presentation, and a high neutrophil count at the time of admission are poor prognostic indicators (153, 197, 385).

Restrictive lung function abnormalities due to residual lung fibrosis and muscle weakness are common in the convalescent phase (34, 247, 255). Among survivors of SARS in HKSAR 1 year after illness, significant impairment in diffusion capacity was noted in 23.7% of studied subjects.

The exercise capacity and health status of SARS survivors were also remarkably lower than those of the healthy population (154). A study on the pathological changes of testes from six patients who died of SARS indicated that orchitis was also a complication and suggested that reproductive functions in male patients who recovered from SARS should be monitored (388).

Depression and posttraumatic stress disorder are especially common among health care workers and patients with affected family members (57, 66, 238, 310). Complications due to the use of corticosteroids including psychosis, adrenal insufficiency, and avascular osteonecrosis were also reported (36, 112, 145, 195, 200).

HISTOPATHOLOGICAL CHANGES OF SARS

Histological Changes

Acute diffuse alveolar damage with air space edema was the most prominent feature in patients who died before the 10th day after onset of illness (99, 250).

Hyaline membranes, interstitial edema, interstitial infiltrates of inflammatory cells, bronchiolar injury with loss of cilia, bronchiolar epithelial denudation, and focal deposition of fibrin on the exposed basement membranes were other observed features (157).

Patients who died after the 10th day of illness exhibited a mixture of acute changes and those of the organizing phase of diffuse alveolar damage. There was interstitial and airspace fibroblast proliferation, type II pneumocyte hyperplasia, and squamous metaplasia of bronchial epithelium.

The alveolar spaces contained a combination of macrophages, desquamated pneumocytes, and multinucleated giant cells. Hemophagocytosis in the alveolar exudates and thrombosis of venules were noted in some cases. Other pulmonary complications might include secondary bacterial bronchopneumonia and invasive aspergillosis (345).

Systemic vasculitis involving the walls of small veins with edema, fibrinoid necrosis, and infiltration by monocytes, lymphocytes, and plasma cells were noted in one report (87).

No tissue destruction or severe inflammatory process associated with viral infection was noted in other organs or tissues, but viral particles could be detected in pneumocytes and enterocytes by in situ hybridization (331).

Inflammation, cellular apoptosis, or microvillus atrophy of a significant degree was not found in the intestinal mucosa to account for the watery diarrhea. Immunohistochemical staining showed the presence of viral nucleoproteins in type II pneumocytes and occasionally pulmonary macrophages. Necrosis or atrophy in the lymphoid tissue of lymph nodes and white pulp of the spleen are commonly observed extrapulmonary pathologies.

Immunological Profiles

Flow cytometric examination of the peripheral blood at the time of admission before the use of steroid showed decreases in levels of dendritic cell subsets, natural killer cells, CD4⁺ and CD8⁺ T lymphocytes, and B lymphocytes (82, 213, 420).

A study of three SARS patients suggested that a self-limiting or abortive infection of peripheral blood mononuclear cells can occur, as evident by the presence of minus-strand RNA, the replicative intermediate of the virus during the initial week of illness (208). Studies of the cytokine profile of SARS patients showed conflicting results, which may be due to the use of many immunomodulators including steroids.

However, those studies generally showed consistent and significant elevations of the plasma chemokines gamma interferon (IFN-γ)-inducible protein 10 (IP10 [CXCL10]), monocyte chemotactic protein 1 (MCP-1 [CCL2]), and interleukin-8 (IL-8). In some studies, levels of the Th1-related cytokines IFN-γ and IL-12 and the inflammatory cytokines IL-1β and IL-6, which can induce an intense inflammatory response, were also increased (63, 152, 163, 165, 325, 360).

In one study, patients with severe disease tended to have increased plasma levels of IFN-α, IFN-γ, and CXCL10 and decreased levels of IL-12p70, IL-2, and tumor necrosis factor alpha (TNF-α) during the acute phase.

In the late phase, patients with severe disease had significantly increased plasma chemokine levels of IL-8, CXCL10, and CCL2 but decreased cytokine levels of IL-12p70, IL-2, TNF-α, and IFN-γ compared with mild cases of SARS (26). These host responses may account for the recruitment and accumulation of alveolar macrophages and polymorphs and the activation of Th1 cell-mediated immunity by the stimulation of natural killer and cytotoxic T lymphocytes, respectively. Since SARS-CoV appears to evade the triggering of IFN-α and IFN-β in human macrophages in vitro (61, 280), the lack of an antiviral innate immune response may permit uncontrolled viral replication with progressive increases in viral load and the accompanying proinflammatory systemic response.

This situation continues into the second week of illness until the appearance of the adaptive immune response, which brings viral replication under control. Moreover, comparative transcriptomal microarray analysis showed that SARS-CoV rather than CoV-229E markedly upregulated genes associated with apoptosis, inflammation, the stress response, and procoagulation during the early phase of infection of a human liver cancer cell line (Huh7) (322).

Both observations help to explain the clinical severity of SARS in relation to the high viral load at up to 2 weeks of illness and the intense inflammatory response as evident from serum cytokine profiles and histopathology.

The majority of SARS patients resolved the proinflammatory cytokine and chemokine responses at the acute phase and expressed adaptive immune genes.

In contrast, patients who later succumbed showed deviated IFN-stimulated gene and immunoglobulin gene expression levels, persistent chemokine levels, and deficient anti-SARS spike antibody production.

It was speculated that unregulated IFN responses during the acute phase may lead to a malfunction of the switch from innate immunity to adaptive immunity. Indeed, recovered patients were found to have higher and sustainable levels of N-specific antibody and S-specific neutralizing antibody responses, whereas patients who later succumbed had an initial rise and then a fall in antibody levels just before death, suggesting that antibody response is likely to play an important role in determining the ultimate disease outcome (417).

PATHOGENESIS, IMMUNE RESPONSE, AND HOST SUSCEPTIBILITY

Interaction between Viral and Cellular Factors

The exact mechanism of how the virus produces damage at cells, tissue, and organs to clinical levels remains elusive. Similar to other viruses such as influenza A virus, Nipah virus, or Ebola virus, SARS-CoV must possess the ability to evade the innate antiviral response of the cells in order to replicate efficiently in the host.

Transfection experiments with Orf3b, Orf6, and N in 293T cells showed that these viral proteins are IFN antagonists that can interfere with the synthesis of IFN and its downstream signaling pathways (178).

However, this cannot explain the apparent discrepancy of IFN-β/α production in infected human intestinal Caco-2 cell line (253) and the lack of such production in SARS patients’ peripheral blood mononuclear cells or in human primary macrophages abortively infected with SARS-CoV despite the activation of several IFN-stimulated genes in the latter case (61).

On the other hand, this may explain the increased serum level of IFN of some SARS patients, which may have an intestinal source. Due to the lack of a type 2 pneumocyte cell line that is susceptible to SARS-CoV, the relevance of these findings cannot be ascertained for lung epithelial cells.

Once the virus can overcome the innate immune response at the cellular level, it can take over the host metabolic apparatus through the degradation of host mRNA by nsp1 and the modulation of the ubiquitination pathway of the host by nsp3 (15, 81, 192, 224, 279).

Efficient viral replication ensues, and cell damage occurs by virus-induced cytolysis or immunopathology.

Infected cell lines and postmortem lung tissues have shown cytopathic changes due to apoptosis, necrosis, or occasionally syncytium formation. Expression of nsp5, nsp10, Orf3a, Orf3b, Orf7a, Orf8a, E, M, and N in different cell lines by transfection can cause cellular apoptosis (Table 1). Expression of S in transfected cells can lead to syncytium formation with cells expressing ACE2 (181).

Paradoxically, little cytopathic effect or inflammation was found in intestinal biopsy specimens of SARS patients despite marked viral replication seen with electron microscopy (205). The transcriptomal profile of infected Caco-2 cells showed a marked upregulation of the potent immunosuppressive cytokine transforming growth factor β and the antiapoptotic host cellular response, which may explain the noninflammatory secretory diarrhea and huge amount of viral shedding in stool (79).

Therefore, the clinical or histopathological manifestations at various organs or tissues do not depend solely on the presence of the relevant receptor and coreceptors or the viral productivity as reflected by the viral load.

The inflammatory and apoptotic responses of the cell triggered by the virus and the compensatory regenerative power or functional reserve of that organ may be equally important in determining the manifestations and the outcome of infection. nsp1 expression in human lung epithelial A549 cells can increase the expression of the chemokines IP10, CCL3, and CCL5 through the NF-κB pathway (192).

This correlated well with the plasma chemokine profile of SARS patients and the immunohistochemical staining of infected lungs. IP10 expressed on pneumocytes is a potent chemoattractant for activated cytotoxic T lymphocytes, natural killer cells, and monocytes, which may therefore infiltrate the interstitium and alveoli of lungs of SARS patients.

Administration of a recombinant S fragment between positions 324 and 688 and Orf3a expression in lung cells can excite the production of IL-8 (43, 169). The expression of N in transfected cells can also activate the Cox2 inflammatory cascade (393). If SARS-CoV can indeed suppress the early innate immune response of IFN-β/α in type 2 pneumocytes without activating the IFN-stimulated genes and therefore also allowing an uncontrolled viral replication in the adjacent cells, the concomitant activation of proinflammatory chemokines and cytokines would explain the dominant and highly fatal manifestation of SARS in the lungs.

Adaptive Immune Response

In general, specific serum antibody against whole SARS-CoV by indirect immunofluorescence or neutralization tests starts to appear at around day 7, plateaus at around the second month, and is maintained for over 12 months. Immunoglobulin M (IgM) and IgG appeared at around the same time, but the former was not detected after 2 to 3 months (371).

Serum testing by recombinant nucleocapsid EIA can detect such an antibody as early as the fifth day after the onset of symptoms (46). The virus-specific T-cell-mediated immune response is not clearly defined. In one study, S-specific cell-mediated immunity mediated by CD4 and CD8 cells was found to last for more than 1 year (395).

Host SusceptibilitySome studies suggested a possible association of HLA-B*4601 with susceptibility to and severity of SARS among the Chinese population in Taiwan (223), but the finding was not confirmed in HKSAR SARS cases.

Among the Chinese population in HKSAR, similar associations with HLA-B*0703 and the genetic variant ICAM3 Gly143 have been found (35, 249). Low-mannose-binding lectin producing the YB haplotype has an increased risk of acquiring SARS (160, 416). On the other hand, individuals with HLA-DRB1*0301 or that are homozygous for CLEC4M tandem repeats were found to be less susceptible to SARS-CoV infection (40, 249). However, the latter finding was strongly disputed in two subsequent studies (324, 430).

CLINICAL MANAGEMENT AND ANTIVIRALS

Since there is no proven effective antiviral agent by randomized placebo control trial (Table 7), clinical management of SARS has relied largely upon supportive care. Broad-spectrum antimicrobial coverage for community-acquired pneumonia should be given while virological confirmation is pending.

Such antibiotics should be stopped once the diagnosis of SARS is confirmed, but nosocomial infections as a result of prolonged intubation and the use of corticosteroids should be appropriately managed.

TABLE 7.

Antiviral agents and immunomodulators against SARS-CoV in vivo

Antiviral agent and/or immunomodulator (no. of subjects) (study design)	Main findingsa	Reference
Ribavirin (144 patients) (retrospective case series)	126 patients (88%) treated; side effects of hemolysis (76%) and lowered hemoglobin of 2 g/dl (49%)	21
Ribavirin (229 patients) (retrospective uncontrolled cohort analysis)	97 patients (42.2%) treated; crude death rate of 10.3% (treatment) vs 12.9% (control) (P = 0.679)	199
Ribavirin and corticosteroids (75 patients) (prospective case series)	9 patients (12%) had spontaneous pneumomediastinum; 20% developed ARDS in wk 3	258
Ribavirin and MP (31 patientsb) (retrospective case series)	No patient required intubation or mechanical ventilation; no mortality noted in this series	303
Ribavirin and corticosteroidsc (71 patientsd) (prospective cohort study)	Crude mortality rate of 3.4% (only in patients aged >65 yr); none of the discharged survivors required continuation of oxygen therapy	186
Ribavirin and corticosteroidse (138 patients) (prospective uncontrolled study)	None responded to antibacterials; 25 patients (18.1%) responded to ribavirin and low-dose corticosteroid; 107 patients required high-dose MP, 88.8% of whom responded; 21 patients (15.2%) required mechanical ventilation; mortality rate, 10.9%	314
Ribavirin and MP (72 patientsf) (retrospective uncontrolled study)	Patients treated with initial pulse MP therapy had no better rate on mechanical ventilation (5.9% vs 9.1%) (NS) and mortality (5.9% vs 5.5%) (NS)	139
Lopinavir-ritonavir and ribavirin (41 patients) (retrospective study with historical control)g	ARDS and death were lower in treatment group than in historical control (2.4% vs 28.8%) (P < 0.001) at day 21 after symptom onset	72
Lopinavir-ritonavir as initial therapy (44 patients) (retrospective matched cohort study)h	Intubation rate of 0% vs 11% (P < 0.05); mortality rate of 2.3% vs 15.6% (P < 0.05)	33
Lopinavir-ritonavir as rescue therapy (31 patients) (retrospective matched cohort study)i	Intubation rate of 9.7% vs 18.1% (NS); mortality rate of 12.9% vs 14% (NS)	33
IFN-alfacon-1 and corticosteroids (22 patients) (open-label study)	9 patients (40.9%) were treated; 1 (11.1%) patient required mechanical ventilation, and no patient died; of 13 patients (59.1%) treated with corticosteroid alone, 3 (23.1%) required mechanical ventilation and 1 (7.7%) died	231
Pentaglobin, an IgM-enriched immunoglobulin (12 patientsj) (retrospective analysis)	Improvement in radiographic scores compared with day 1 (median, 9.5) on days 6 (median, 6) (P = 0.01) and 7 (median, 6) (P = 0.01) and in oxygen requirement compared with day 1 (median, 2.5 liters/min) on days 6 (median, 1 liter/min) (P = 0.04) and 7 (median, 0.5 liters/min) (P = 0.04) after commencement of pentaglobin treatment	140
Convalescent plasma (1 patientk) (case report)	Convalescent plasma (200 ml) was given at day 15 after onset of illness without adverse reaction; patient recovered uneventfully	366
Convalescent plasma (3 patients)	Viral load decreased from 4.9 × 105-6.5 × 105 copies/ml to undetectable 1 day after transfusion	401
Convalescent plasma (80 patients)	A higher day 22 discharge rate was observed in patients treated before day 14 of illness (58.3% vs 15.6%) (P < 0.001), and in patients with positive PCR, SARS-CoV antibodies were negative at the time of plasma infusion (66.7% vs 20%) (P = 0.001)	60
Two herbal formulas (Sang Ju Yin and Yu Ping Feng San) (37 healthy volunteers)	Given oral traditional Chinese medicine regimen daily for 14 days with transient increase in CD4/CD8 ratio	269

• ↵a ARDS, acute respiratory distress syndrome; MP, methylprednisolone; NS, P value was not significant.
• ↵b One patient recovered on antibacterial treatment alone.
• ↵c A 3-week step-down course of corticosteroids and pulsed methylprednisolone rescue for deterioration.
• ↵d Three patients recovered on antibacterial treatment alone.
• ↵e Low-dose corticosteroid and selective use of high-dose methylprednisolone.
• ↵f Initially treated with high-dose pulse (n = 17) versus nonpulse (n = 55) methylprednisolone.
• ↵g One hundred eleven patients treated with ribavirin as a historical control.
• ↵h Six hundred thirty-four patients selected as matched cohort.
• ↵i Three hundred forty-three patients selected as matched cohort.
• ↵j Patients who continued to deteriorate despite ribavirin and corticosteroid therapy.
• ↵k Patient who continued to deteriorate despite ribavirin and corticosteroid therapy.

The correlation between viral loads and clinical outcome suggests that suppression of viral replication by effective antiviral drugs should be the key to preventing morbidity and mortality. However, in vitro susceptibility test results were often conflicting, as in the case of IFN-β1a (78, 137, 318) and IFN-α2b (308, 318).

Nevertheless, it appears that IFN-β, IFN-αn1, IFN-αn3, and leukocytic IFN-α have some potential activity and warrant evaluation by clinical trials (50, 305, 426). Although a very high 50% cytotoxic concentration exceeding 1,000 mg/liter has been demonstrated for ribavirin (77), and although its low level of in vitro activity against SARS-CoV was initially attributed to cellular toxicity (318), ribavirin has good activity when tested in other human Caco-2 and pig kidney cell lines despite its lack of activity in Vero cells (243). The use of different cell lines, testing conditions, and virus strains may have contributed to these discrepancies.

Numerous other potential antiviral agents have been identified using different approaches (Table 8). Replication of SARS-CoV requires proteolytic processing of the replicase polyprotein by two viral cysteine proteases, a chymotrypsin-like protease (3CL^pro) and a papain-like protease (PL^pro).

These proteases are important targets for the development of antiviral drugs. Protease inhibitors (especially nelfinavir) (386, 392), glycyrrhizin (77), baicalin (50), reserpine (381), aescin (381), valinomycin (381), niclosamide (380), aurintricarboxylic acid (129), mizoribine (293), indomethacin (4), chloroquine (174), and many herbal formulations, have also been found to possess some antiviral activity against SARS-CoV in vitro.

In addition, an organic nitric oxide donor, S-nitro-N-acetylpenicillamine, appeared to have inhibitory activity against SARS-CoV (2), which has formed the basis for the use of nitric oxide inhalation as an experimental form of rescue therapy for SARS (52). Several agents with good in vitro antiviral activities, including ACE2 analogues, helicase inhibitors, and nucleoside analogues, were also reported to have some activity in vitro (14, 332).

Antiviral peptides designed against the S protein and especially those derived from heptad repeat region 2 of S2 were shown to inhibit membrane fusion and cell entry (22, 177, 227). Small interfering RNA (siRNA) also demonstrated activities in reducing cytopathic effects, viral replication, and viral protein expression in cell lines (125, 232, 351, 418, 419, 428).

Screening of chemical libraries has identified several inhibitors of protease, helicase, and spike-mediated cell entry (170). Most of the above-mentioned chemicals or approaches have not been evaluated in human or animal models. In mouse models, nelfinavir, β-d-N⁴-hydroxycytidine, calpain inhibitor VI, 3-deazaneplanocin A, human leukocyte IFN-αn3, and anti-inflammatory agents including chloroquine, amodiaquin, and pentoxifylline did not significantly reduce lung virus titers in mice.

When not given in combination with other antivirals, the IMP dehydrogenase inhibitors, including ribavirin, suppress the proinflammatory response while augmenting viral replication in this mouse model (13).

TABLE 8.

Antiviral agents and immunomodulators tested against SARS-CoV in animals and in vitro

Antiviral agent(s) and/or immunomodulator(s)	Study setting and methods (virus strain)	Main findingsa	Reference
IFN-αB/D (hybrid IFN)	BALB/c mice (Urbani)	i.p. IFN-αB/D once daily for 3 days beginning 4 h after virus exposure reduced SARS-CoV replication in lungs by 1 log10 at 10,000 and 32,000 IU; at the highest dose of 100,000 IU, virus lung titers were not detectable	13
Ampligen [poly(I:C124)] (mismatched double-stranded RNA IFN inducer)	BALB/c mice (Urbani)	i.p. Ampligen at 10 mg/kg 4 h after virus exposure reduced virus lung titers to undetectable levels	13
Pegylated IFN-α as prophylactic treatment	Cynomolgus macaques (Macaca fascicularis) (patient 5668)	Significantly reduced viral replication and excretion, viral antigen expression by type 1 pneumocytes, and pulmonary damage; postexposure treatment with pegylated IFN-α yielded intermediate results	118
IFN-α2b (Intron A)	Vero (FFM-1, HK isolate)	Mean (SD) EC50 = 4,950 (890) IU/ml (SI of >2) for FFM-1 isolate; mean (SD) EC50 = 6,500 (980) IU/ml (SI of >105) for HK isolate	78
	Caco2 (FFM-1, HK isolate)	Mean (SD) EC50 = 1,530 (220) IU/ml (SI of >6.5) for FFM-1 isolate; mean (SD) EC50 = 880 (130) IU/ml (SI of >11.4) for HK isolate	78
IFN-β1b (Betaferon)	Vero (FFM-1, HK isolate)	Mean (SD) EC50 = 95 (17) IU/ml (SI of >105) for FFM-1 isolate; mean (SD) EC50 = 105 (21) IU/ml (SI of >95) for HK isolate	78
	Caco2 (FFM-1, HK isolate)	Mean (SD) EC50 = 21 (3.9) IU/ml (SI of >476) for FFM-1 isolate; mean (SD) EC50 = 9.2 (2.1) IU/ml (SI of >1,087) for HK isolate	78
IFN-γ1b (Imukin)	Vero (FFM-1, HK isolate)	Mean (SD) EC50 = 2,500 (340) IU/ml (SI of >4) for FFM-1 isolate; mean (SD) EC50 = 1,700 (290) IU/ml (SI of >5.9) for HK isolate	78
	Caco2 (FFM-1, HK isolate)	Mean (SD) EC50 = >10,000 IU/ml (SI NA) for FFM-1 isolate; mean (SD) EC50 = >10,000 IU/ml (SI NA) for HK isolate	78
IFN-β1a	Vero E6 (Tor2, Tor7, and Urbani)	IFN with p.i. IC50 = 50 IU/ml; IFN added postinfection IC50 = 500 IU/ml	137
IFN-β, IFN-α, IFN-γ	Vero, MxA-expressing Vero (FFM-1)	SARS-CoV strongly inhibited by IFN-β (with p.i.) and less so with IFN-α and IFN-γ; MxA does not interfere with viral replication	305
IFN-α, IFN-β	FRhK-4 (NMf)	↓ intracellular viral RNA copies; IFN-α IC50 = 25 U/ml; IFN-β IC50 = 14 U/ml	426
IFN-α2b	Vero E6 (Tor2, Tor3, Tor7, and Tor684)	IC50 = ∼500 IU/ml	308
Leu-IFN-α	FRhK-4 (HKU39849)	EC50 at 48 h = 5,000 μg/ml	50
	Vero E6 (HKU39849)	EC50 at 48 h = 19.5 μg/ml	50
IFN-α (p.i. for 16 h before viral inoculation)	FRhK-4 (HKU39849)	EC50 at 48 h = 39 μg/ml	50
	Vero E6 (HKU39849)	EC50 at 48 h = 19.5 μg/ml	50
IFN-β	FRhK-4 (HKU39849)	EC50 at 48 h = 200 μg/ml	50
	Vero E6 (HKU39849)	EC50 at 48 h = 106 μg/ml	50
IFN-β (p.i. for 16 h before viral inoculation)	FRhK-4 (HKU39849)	EC50 at 48 h = 625 μg/ml	50
	Vero E6 (HKU39849)	EC50 at 48 h = 19.5 μg/ml	50
IFN-β1b (Betaferon)	Vero E6 (2003VA2774)	IC50 = 0.2 IU/ml; IC95 = 8 IU ml	318
IFN-αn3 (Alferon)	Vero E6 (2003VA2774)	IC50 = 0.8 IU/ml; IC95 = 200 IU/ml	318
Human leukocyte IFN-α (Multiferon)	Vero E6 (2003VA2774)	IC50 = 2 IU/ml; IC95 = 44 IU/ml	318
IFN-β	Vero E6 (FFM-1)	IC50 = 110 IU/ml at 10 TCID50; IC50 = 625 IU/ml at 100 TCID50	83
Multiferon	Vero E6 (FFM-1)	IC50 = 540 IU/ml at 10 TCID50; IC50 = 2,400 IU/ml at 100 TCID50	83
IFN-α2b	Vero E6 (FFM-1)	IC50 = >3,125 IU/ml at 10 TCID50; IC50 = >3,125 IU/ml at 100 TCID50	83
IFN-α2a	Vero E6 (FFM-1)	IC50 = >3,125 IU/ml at 10 TCID50; IC50 = >3,125 IU/ml at 100 TCID50	83
IFN-alfacon1 (Infergen)	Vero (Urbani)	IC50 = 0.001 μg/ml	257
IL-4 and IFN-γ	Vero E6 (HKU39849)	IL-4 and IFN-γ downregulated cell surface expression of ACE2; ACE2 mRNA levels were also decreased after treatment	84
IFN-β and ribavirin	Caco2 (FFM-1)	Mean (SD) CI = 0.45 (0.07)	243
HR2-8 (HR2-derived peptide)	Vero 118 (NM)	EC50 = 17 μM	22
CP-1 (HR2-derived peptide)	Vero E6 (WHU)	IC50 ≈ 19 μmol/liter	227
HR1-1 (HR1-derived peptide)	Vero E6 (BJ01 and pseudovirus)	EC50 = 3.68 μM for wild-type virus assay; EC50 = 0.14 μM for pseudotyped virus assay	407
HR2-18 (HR2-derived peptide)	Vero E6 (BJ01 and pseudovirus)	EC50 = 5.22 μM for wild-type virus assay; EC50 = 1.19 μM for pseudotyped virus assay	407
HR2	Vero E6 (WHU)	CPE inhibition IC50 = 0.5-5 nM (synthetic HR2 peptide) and 66.2-500 nM (fusion HR2 peptide)	432
Peptides representing various regions of ACE2	TELCeB6, HeLa, and VeroE6 (pseudovirus)	IC50 = 50 μM (peptide aa 22-44); IC50 = 6 μM (peptide aa 22-57); IC50 = 0.1 μM (peptide aa 22-44 and 351-357) artificially linked by glycine	120
Peptides analogous to viral spike protein	Vero E6, L2 (Urbani)	Inhibit viral plaque formation by 40-70% at 15-30 μM; peptides analogous to regions of the N terminus or the pretransmembrane domain of the S2 subunit; inhibit viral plaque formation by >80% at 15-30 μM (peptides analogous to the SARS-CoV loop region)	295
siRNA, RL004, RL005	Vero E6 (Y3)	siRNA (600 pmol/liter) targeting conserved regions of SARS-CoV, ↓ virally induced CPE at 67 h	418
siRNA	FRhK-4 (HKU66078)	siRNA duplexes targeting regions in entire viral genome, ↓ virally induced CPE and viral production at 72 h	428
siRNA targeting viral RP	Vero (NM)	↓ virally induced CPE, ↓ viral production, ↓ viral protein synthesis at 1.5 or 3 μg of siRNA	351
RNA interference targeting viral RP	Vero E6, 293, HeLa (SARS-CoV-p9)	↓ expression of RP (293 and HeLa cells); ↓ plaque formation at 1 μg of siRNA	232
siRNA targeting S gene	Vero E6, 293T (BJ01)	↓ S gene expression in SARS-CoV-infected cells at 2, 3, and 4 μg of siRNA	419
siRNAs targeting S gene and 3′ untranslated region	Vero E6 (HK strain)	↓ viral antigen synthesis of 64% (by siSARS-S2), 51% (siSARS-S3), 40% (siSARS 3′ untranslated region) at 100 pmol of siRNA	379
Glycyrrhizin	Vero (FFM-1, FFM-2)	CPE assay mean (SD) CC50 = >20,000 mg/liter; EC50 = 300 (51) mg/liter (SI of >67)	77
	FRhK-4 (HKU39849)	EC50 at 48 h = >400 μg/ml	50
	Vero E6 (HKU39849)	EC50 at 48 h = 100 μg/ml	50
Mizoribine	Vero E6 (FFM-1)	IC50 = 3.5 μg/ml; CC50 = >200 μg/ml	293
	Vero E6 (HKU39489)	IC50 = 16 μg/ml	293
Ribavirin	Vero E6 (FFM-1)	IC50 = 20 μg/ml; CC50 = >200 μg/ml	293
	Vero E6 (HKU39489)	IC50 = 80 μg/ml	293
	FRhK-4 (HKU39849)	EC50 at 48 h = 50-100 μg/ml	50
	Vero E6 (HKU39849)	EC50 at 48 h = >200 μg/ml	50
Rimantidine	FRhK-4 (HKU39849)	EC50 at 48 h = 16 μg/ml	50
	Vero E6 (HKU39849)	EC50 at 48 h = 8-16 μg/ml	50
Lopinavir	FRhK-4 (HKU39849)	EC50 at 48 h = 16 μg/ml	50
	Vero E6 (HKU39849)	EC50 at 48 h = 8-16 μg/ml	50
Baicalin	FRhK-4 (HKU39849)	EC50 at 48 h = 12.5 μg/ml	50
	Vero E6 (HKU39849)	EC50 at 48 h = 100 μg/ml	50
Aurintricarboxylic acid	Vero (NM)	EC50 = 0.2 mg/ml; CC50 = 37.5 mg/ml; SI = 187	129
Reserpine	Vero E6 (HK strain)	EC50 = 3.4 μM; CC50 = 25 μM; SI = 7.3	381
Aescin	Vero E6 (HK strain)	EC50 = 6 μM; CC50 = 15 μM; SI = 2.5	381
Valinomycin	Vero E6 (HK strain)	EC50 = 0.85 μM; CC50 = 68 μM; SI = 80	381
Niclosamide	Vero E6 (Taiwan strain)	EC50 = 1-3 μM; CC50 = 250 μM	380
Nelfinavir	Vero E6 (FFM-1)	Mean (SD) EC50 = 0.048 (0.024) μM; CC50 = 14.75 (2.75) μM; SI = 302.1	392
Chloroquine	Vero E6 (FFM-1)	Mean (SD) IC50 = 8.8 (1.2) μM; CC50 = 261.3 (14.5) μM; SI = 30	174
	Vero E6 (Urbani)	Mean (SD) EC50 = 4.4 (1.0) μM; refractory to infection if pretreated with chloroquine (10 μM) for 20 h	341
Indomethacin	Vero E6 (Tor2)	IC50 = 50 μM	4
3C-like proteinase inhibitors
Cinanserin (SQ 10,643)	Vero (NM)	IC50 = 5 μM	51
TG-0205221	Vero E6 (NM)	↓ viral load by 4.7 logs at 5 μM	396
Octapeptide AVLQSGFR	Vero (BJ01)	EC50 = 0.027 mg/liter; CC50 = >100 mg/liter; SI = >3,704	101
Peptidomimetic inhibitor	NM	IC50 = 45-70 μM	103
Calpain inhibitors, Val-Leu-CHO	Vero E6 (Urbani)	EC90 = 3 μM	14
Calpain inhibitors, Z-Val-Phe-Ala-CHO	Vero E6 (Urbani)	EC90 = 15 μM	14
Cyclopentenyl carbocyclic nucleosides	NM	EC50 = 47 μM for 1,2,3-triazole analogue (17c); EC50 = 21 μM for 1,2,4-triazole analogue (17a)	65
Nucleoside analogue inhibitor, β-d-N4-hydroxycytidine	Vero E6 (Urbani)	EC90 = 6 μM	14
Nitric oxide, S-nitroso-N-acetylpenicillamine	Vero E6 (FFM-1)	Mean (SD) IC50 = 222 (83.7) μM; SI = 3	173
Pyridine N-oxide derivatives	Crandel feline kidney (CRFK) and Vero (FFM-1)	Among 192 compounds tested, the oxide part on pyridine moiety was indispensable for antiviral activity with CC50 of 50-100 mg/liter	11
Stilbene derivatives	Vero E6 (NM)	Inhibited by compounds 17 and 19 at 0.5 mg/ml, and no significant cytotoxic effects were observed in vitro	217
Peptide-conjugated antisense morpholino oligomers (P-PMO)	Vero E6 (Tor2)	Several virus-targeted P-PMO (AUG1, AUG2, AUG3, 1AFBS, and 3UTR) consistently reduced CPE at a concn of 20 μM	246
20-mer synthetic peptides (S protein fragments)	FRhK-4 (GZ50)	IC50 = 24.9-113 μg/ml; IC90 = 0.9-15.9 μg/ml	427
Diverse small molecules,b MP576, HE602, and VE607	Vero (HKU39849)	EC50 = <10 μM	170
Adamantane-derived compoundsc	FRhK-4 (NM)	IC50 = 0.5-3 μM; CC50 = >300 μM	328
Semisynthetic derivatives of glycopeptide
Vancomycin	Vero E6 (FFM-1)	EC50 = 22->100 μM; CC50 = >80 μM	10
Eremomycin	Vero E6 (FFM-1)	EC50 = 14->100 μM; CC50 = 45->100 μM	10
Teicoplanin, ristocetin A, and DA-40926	Vero E6 (FFM-1)	EC50 = >80 μM; CC50 = >80 μM	10
Lycoris radiata (Chinese medicinal herb)	Vero E6 (BJ001, BJ006)	Mean (SD) EC50 = 2.4 (0.2) μg/ml; CC50 = 886.6 (35.0) μg/ml	212
Artemisia annua (Chinese medicinal herb)	Vero E6 (BJ001, BJ006)	Mean (SD) EC50 = 34.5 (2.6) μg/ml; CC50 = 1,035 (92.8) μg/ml	212
Pyrrosia lingua (Chinese medicinal herb)	Vero E6 (BJ001, BJ006)	Mean (SD) EC50 = 43.2 (14.1) μg/ml; CC50 = 2,378 (87.3) μg/ml	212
Lindera sp. (Chinese medicinal herb)	Vero E6 (BJ001, BJ006)	Mean (SD) EC50 = 88.2 (7.7) μg/ml; CC50 = 1,374 (39.0) μg/ml	212

• ↵a i.p., intraperitoneal; EC₅₀, 50% effective concentration; SI, selectivity index; NA, not available; p.i., preincubation; NM, not mentioned; IC₅₀, 50% inhibitory concentration; TCID₅₀, 50% tissue culture infective dose; CI, combination index (combination index of <1 indicates synergism); CPE, cell culture cytopathic effect; aa, amino acids; RP, RNA polymerase; CC₅₀, 50% cytotoxic concentration; ↓, decreased.
• ↵b MP576, HE602, and VE607 were validated to be inhibitors of SARS-CoV M^pro, Hel, and viral entry, respectively.
• ↵c Bananin, iodobananin, vanillinbananin, and eubananin were effective inhibitors of the ATPase activity of the SCV helicase.

Before the demonstration of viral load as an important factor in determining clinical outcome, immunomodulators were empirically used for the treatment of SARS during the initial epidemic (59). These immunomodulators include corticosteroids, intravenous immunoglobulins, pentaglobulin, thymosin, thalidomide, and anti-TNF (140, 421).

Corticosteroids were previously found to reduce mortality in patients with pneumonia due to varicella-zoster virus and influenza virus (1, 109).

High-dose hydrocortisone was shown to reduce the expression of the proinflammatory chemokines CXCL8 and CXCL10 in infected Caco-2 cells (80).

However, without an effective antiviral agent, the early use of high doses of corticosteroids for prolonged periods could be detrimental.

It may increase the plasma viral load and the risk of nosocomial infections and avascular osteonecrosis (196). Pegylated IFN-α2a was shown to be useful for prophylaxis and reducing respiratory viral shedding and lung pathology when used as an early treatment in a monkey model (118).

Among clinical treatments studied, combinations of steroid with either alfacon-1, a recombinant consensus IFN-α (231), or protease inhibitors and ribavirin were found to improve outcomes in two different treatment trials using historical controls (33, 72).

Due to the very short time course of this epidemic and the initial lack of suitable animal models, randomized control treatment trials are difficult to be organized and executed despite the finding of some commercially available candidate agents that appeared to be active in vitro.