Identifying and prioritizing potential human-infecting viruses from their genome sequences

Determining which animal viruses may be capable of infecting humans is currently intractable at the time of their discovery, precluding prioritization of high-risk viruses for early investigation and outbreak preparedness. Given the increasing use of genomics in virus discovery and the otherwise sparse knowledge of the biology of newly discovered viruses, we developed machine learning models that identify candidate zoonoses solely using signatures of host range encoded in viral genomes. Within a dataset of 861 viral species with known zoonotic status, our approach outperformed models based on the phylogenetic relatedness of viruses to known human-infecting viruses (area under the receiver operating characteristic curve [AUC] = 0.773), distinguishing high-risk viruses within families that contain a minority of human-infecting species and identifying putatively undetected or so far unrealized zoonoses. Analyses of the underpinnings of model predictions suggested the existence of generalizable features of viral genomes that are independent of virus taxonomic relationships and that may preadapt viruses to infect humans. Our model reduced a second set of 645 animal-associated viruses that were excluded from training to 272 high and 41 very high-risk candidate zoonoses and showed significantly elevated predicted zoonotic risk in viruses from nonhuman primates, but not other mammalian or avian host groups. A second application showed that our models could have identified Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) as a relatively high-risk coronavirus strain and that this prediction required no prior knowledge of zoonotic Severe Acute Respiratory Syndrome (SARS)-related coronaviruses. Genome-based zoonotic risk assessment provides a rapid, low-cost approach to enable evidence-driven virus surveillance and increases the feasibility of downstream biological and ecological characterization of viruses.

Surveillance of emerging viruses is one of the first steps to avoid the next pandemic. This study uses machine learning to identify many zoonotic viruses directly from their genomes. This allows rapid assessment of research priorities as soon as new viruses are discovered, focusing research and surveillance efforts on the viruses most likely to infect humans.     

Developing an effective RNA interference strategy against a plus-strand RNA virus: silencing of coxsackievirus B3 and its cognate coxsackievirus-adenovirus receptor

Coxsackievirus B3 (CVB-3) is a plus-strand RNA virus that is believed to be the most common causal agent of viral myocarditis. Since no specific treatment for CVB-3 infections is available to date, we and others have recently started to develop RNA interference (RNAi) approaches to prevent virus propagation. Here we describe our strategy for the development of efficient small interfering RNAs (siRNAs) against viral genomes. Initially, fusion constructs of a reporter (green fluorescent protein) and viral subgenomic fragments were employed to select active siRNAs against the virus. Moreover, in an attempt to achieve sustained virus silencing and reduce the risk of generating escape mutants, only highly efficient siRNAs directed against regions of the viral genome that are unlikely to tolerate mutations were considered for virus inhibition. Two siRNAs directed against the 3D RNA-dependent RNA polymerase were found to inhibit virus propagation by 80-90%. The protective effect of the efficient siRNAs lasted for several days. Furthermore, we have first evidence that inhibition of the cellular coxsackievirus-adenovirus receptor (CAR) by RNAi also reduces the virus titre. Our strategy is likely to be applicable to other (RNA) viruses as well.     

Innate Immunity and the Inter-exposure Interval Determine the Dynamics of Secondary Influenza Virus Infection and Explain Observed Viral Hierarchies

Influenza is an infectious disease that primarily attacks the respiratory system. Innate immunity provides both a very early defense to influenza virus invasion and an effective control of viral growth. Previous modelling studies of virus–innate immune response interactions have focused on infection with a single virus and, while improving our understanding of viral and immune dynamics, have been unable to effectively evaluate the relative feasibility of different hypothesised mechanisms of antiviral immunity. In recent experiments, we have applied consecutive exposures to different virus strains in a ferret model, and demonstrated that viruses differed in their ability to induce a state of temporary immunity or viral interference capable of modifying the infection kinetics of the subsequent exposure. These results imply that virus-induced early immune responses may be responsible for the observed viral hierarchy. Here we introduce and analyse a family of within-host models of re-infection viral kinetics which allow for different viruses to stimulate the innate immune response to different degrees. The proposed models differ in their hypothesised mechanisms of action of the non-specific innate immune response. We compare these alternative models in terms of their abilities to reproduce the re-exposure data. Our results show that 1) a model with viral control mediated solely by a virus-resistant state, as commonly considered in the literature, is not able to reproduce the observed viral hierarchy; 2) the synchronised and desynchronised behaviour of consecutive virus infections is highly dependent upon the interval between primary virus and challenge virus exposures and is consistent with virus-dependent stimulation of the innate immune response. Our study provides the first mechanistic explanation for the recently observed influenza viral hierarchies and demonstrates the importance of understanding the host response to multi-strain viral infections. Re-exposure experiments provide a new paradigm in which to study the immune response to influenza and its role in viral control.     

Tobacco mosaic virus movement protein enhances the spread of RNA silencing

Eukaryotic cells restrain the activity of foreign genetic elements, including viruses, through RNA silencing. Although viruses encode suppressors of silencing to support their propagation, viruses may also exploit silencing to regulate host gene expression or to control the level of their accumulation and thus to reduce damage to the host. RNA silencing in plants propagates from cell to cell and systemically via a sequence-specific signal. Since the signal spreads between cells through plasmodesmata like the viruses themselves, virus-encoded plasmodesmata-manipulating movement proteins (MP) may have a central role in compatible virus:host interactions by suppressing or enhancing the spread of the signal. Here, we have addressed the propagation of GFP silencing in the presence and absence of MP and MP mutants. We show that the protein enhances the spread of silencing. Small RNA analysis indicates that MP does not enhance the silencing pathway but rather enhances the transport of the signal through plasmodesmata. The ability to enhance the spread of silencing is maintained by certain MP mutants that can move between cells but which have defects in subcellular localization and do not support the spread of viral RNA. Using MP expressing and non-expressing virus mutants with a disabled silencing suppressing function, we provide evidence indicating that viral MP contributes to anti-viral silencing during infection. Our results suggest a role of MP in controlling virus propagation in the infected host by supporting the spread of silencing signal. This activity of MP involves only a subset of its properties implicated in the spread of viral RNA.     

Replication of viruses in a growing plaque: a reaction-diffusion model.

An understanding of the viral replication process commonly referred to as "plaque growth" is developed in the context of a reaction-diffusion model. The interactions among three components: the virus, the healthy host, and the infected host are represented using rates of viral adsorption and desorption to the cell surface, replication and release by host lysis, and diffusion. The solution to the full model reveals a maximum in the dependence of the velocity of viral propagation on its equilibrium adsorption constant, suggesting that conditions can be chosen where viruses which adsorb poorly to their hosts will replicate faster in plaques than those which adsorb well. Analytic expressions for the propagation velocity as a function of the kinetic and diffusion parameters are presented for the limiting cases of equilibrated adsorption, slow adsorption, fast adsorption, and large virus yields. Hindered diffusion at high host concentrations must be included for quantitative agreement with experimental data.     

Using Non-Homogeneous Models of Nucleotide Substitution to Identify Host Shift Events: Application to the Origin of the 1918 ‘Spanish’ Influenza Pandemic Virus

Nonhomogeneous Markov models of nucleotide substitution have received scant attention. Here we explore the possibility of using nonhomogeneous models to identify host shift nodes along phylogenetic trees of pathogens evolving in different hosts. It has been noticed that influenza viruses show marked differences in nucleotide composition in human and avian hosts. We take advantage of this fact to identify the host shift event that led to the 1918 ‘Spanish’ influenza. This disease killed over 50 million people worldwide, ranking it as the deadliest pandemic in recorded history. Our model suggests that the eight RNA segments which eventually became the 1918 viral genome were introduced into a mammalian host around 1882–1913. The viruses later diverged into the classical swine and human H1N1 influenza lineages around 1913–1915. The last common ancestor of human strains dates from February 1917 to April 1918. Because pigs are more readily infected with avian influenza viruses than humans, it would seem that they were the original recipient of the virus. This would suggest that the virus was introduced into humans sometime between 1913 and 1918.     

ODE models for oncolytic virus dynamics

Replicating oncolytic viruses are able to infect and lyse cancer cells and spread through the tumor, while leaving normal cells largely unharmed. This makes them potentially useful in cancer therapy, and a variety of viruses have shown promising results in clinical trials. Nevertheless, consistent success remains elusive and the correlates of success have been the subject of investigation, both from an experimental and a mathematical point of view. Mathematical modeling of oncolytic virus therapy is often limited by the fact that the predicted dynamics depend strongly on particular mathematical terms in the model, the nature of which remains uncertain. We aim to address this issue in the context of ODE modeling, by formulating a general computational framework that is independent of particular mathematical expressions. By analyzing this framework, we find some new insights into the conditions for successful virus therapy. We find that depending on our assumptions about the virus spread, there can be two distinct types of dynamics. In models of the first type (the “fast spread” models), we predict that the viruses can eliminate the tumor if the viral replication rate is sufficiently high. The second type of models is characterized by a suboptimal spread (the “slow spread” models). For such models, the simulated treatment may fail, even for very high viral replication rates. Our methodology can be used to study the dynamics of many biological systems, and thus has implications beyond the study of virus therapy of cancers.     

Translational control of coronaviruses

Coronaviruses represent a large family of enveloped RNA viruses that infect a large spectrum of animals. In humans, the severe acute respiratory syndrome coronavirus type 2 (SARS-CoV-2) is responsible for the current COVID-19 pandemic and is genetically related to SARS-CoV and Middle East respiratory syndrome-related coronavirus (MERS-CoV), which caused outbreaks in 2002 and 2012, respectively. All viruses described to date entirely rely on the protein synthesis machinery of the host cells to produce proteins required for their replication and spread. As such, virus often need to control the cellular translational apparatus to avoid the first line of the cellular defense intended to limit the viral propagation. Thus, coronaviruses have developed remarkable strategies to hijack the host translational machinery in order to favor viral protein production. In this review, we will describe some of these strategies and will highlight the role of viral proteins and RNAs in this process.     

Amplification and Spread of Viruses in a Growing Plaque

The two-dimensional propagation of viruses through a "lawn" of receptive hosts, commonly called plaque growth, reflects the dynamics of interactions between viruses and host cells. Here we treat the amplification of viruses during plaque growth as a reaction-diffusion system, where interactions among the virus, uninfected host cells, and virus-producing host-virus complexes are accounted for using rates of viral adsorption to and desorption from the host-cell surface, rates of reproduction and release of progeny viruses by lysis of the host, and by the coupling of these reactions with diffusion of free virus within the agar support. Numerical solution of the system shows the development of a traveling wave of reproducing viruses, where the velocity of the wave is governed by the kinetic and diffusion parameters. The model has been applied to predict the propagation velocity of a bacteriophage plaque. Different mechanisms may account for the dependence of this velocity on the host density during early stages of a growing plaque. The model provides a means to explore how changes in the virus-host interactions may be manifest in a growing plaque.     

Role of the human immunodeficiency virus type 1 envelope gene in viral fitness

A human host offers a variety of microenvironments to the infecting human immunodeficiency virus type 1 (HIV-1), resulting in various selective pressures, most of them directed against the envelope (env) gene. Therefore, it seems evident that the replicative capacity of the virus is largely related to viral entry. In this study we have used growth competition experiments and TaqMan real-time PCR detection to measure the fitness of subtype B HIV-1 primary isolates and autologous env-recombinant viruses in order to analyze the contribution of wild-type env sequences to overall HIV-1 fitness. A significant correlation was observed between fitness values obtained for wild-type HIV-1 isolates and those for the corresponding env-recombinant viruses (r = 0.93; P = 0.002). Our results suggest that the env gene, which is linked to a myriad of viral characteristics (e.g., entry into the host cell, transmission, coreceptor usage, and tropism), plays a major role in fitness of wild-type HIV-1. In addition, this new recombinant assay may be useful for measuring the contribution of HIV-1 env to fitness in viruses resistant to novel antiretroviral entry inhibitors.     

Comparing the plant virus spread patterns of non-persistently transmitted virus and persistently transmitted virus using an individual-based model

To compare the spread patterns between two types of plant viruses, non-persistent virus (NPV) and persistent virus (PV), we developed a spatially-explicit individual-based model. Our probability-based model is driven by the actions of insect vectors that are affected by interactions with host plants and plant viruses, considering both biological and behavioral components of their relationship. As a model system, we used potato virus y and potato leafroll virus, respectively for NPV and PV, potato for host plant, and Myzus persicae for the insect vector; empirical results from previous studies were acquired and adjusted to be used as our parameter values. Our simulation results showed that initial infection of PV in the field resulted in over 1.3 times greater number of insect vectors while causing approximately 7 times greater number of virus-infected plants compared to NPV by the end of simulation. Furthermore, spatial analysis showed that PV-infected plants showed greater aggregation in the field, forming larger patches compared to NPV-infected plants. Our results demonstrated the importance of host plant and insect vector manipulation by plant viruses as well as biological properties such as infectious period in the insect on the difference in overall spread pattern.     

Influenza A virus PB1-F2 protein contributes to viral pathogenesis in mice

The influenza virus PB1-F2 protein is a novel protein previously shown to be involved in induction of cell death. Here we characterize the expression and the function of the protein within the context of influenza viral infection in tissue culture and a mouse model. We show that the C-terminal region of the protein can be expressed from a downstream initiation codon and is capable of interaction with the full-length protein. Using this knowledge, we generated influenza viruses knocked out for the expression of PB1-F2 protein and its downstream truncation products. Knocking out the PB1-F2 protein had no effect on viral replication in tissue culture but diminished virus pathogenicity and mortality in mice. The viruses replicated to similar levels in mouse lungs by day 3 postinfection, suggesting that the knockout did not impair viral replication. However, while the PB1-F2 knockout viruses were cleared after day 5, the wild-type viruses were detectable in mouse lungs until day 7, implying that expression of PB1-F2 resulted in delayed clearance of the viruses by the host immune system. Based on our findings and on the fact that the PB1 genomic segment was always newly introduced into some pandemic influenza viruses of the last century, we speculate that the PB1-F2 protein plays an important role in pathogenesis of influenza virus infection and may be an important contributor to pathogenicity of pandemic influenza viruses.     

Viral biogeography revealed by signatures in Sulfolobus islandicus genomes

Viruses are a driving force of microbial evolution. Despite their importance, the evolutionary dynamics that shape diversity in viral populations are not well understood. One of the primary factors that define viral population structure is coevolution with microbial hosts. Experimental models predict that the trajectory of coevolution will be determined by the relative migration rates of viruses and their hosts; however, there are no natural microbial systems in which both have been examined. The biogeographic distribution of viruses that infect Sulfolobus islandicus is investigated using genome comparisons among four newly identified, integrated, Sulfolobus spindle-shaped viruses and previously sequenced viral strains. Core gene sequences show a biogeographic distribution where viral genomes are specifically associated with each local population. In addition, signatures of host–virus interactions recorded in the sequence-specific CRISPR (clustered regularly interspaced short palindromic repeats) system show that hosts have interacted with viral communities that are more closely related to local viral strains than to foreign ones. Together, both proviral and CRISPR sequences show a clear biogeographic structure for Sulfolobus viral populations. Our findings demonstrate that virus–microbe coevolution must be examined in a spatially explicit framework. The combination of host and virus biogeography suggests a model for viral diversification driven by host immunity and local adaptation.     

Oncolytic adenoviruses in anticancer therapy: Current status and prospects

Lytic virus infection results in production of a virus progeny and lysis of the infected cell. Tumor cells are usually more sensitive to virus infection. Studies indicate that viral oncolysis provides a promising alternative approach to cancer therapy. The ability of viruses to selectively kill cancer cells is long known, but construction of virus variants with an improved therapeutic potential was impossible until recent advances in virus and cell molecular biology and the development of modern methods for directed modification of viruses. Adenoviruses are one of the best studied models of oncolytic viruses. These DNA viruses are convenient for genetic manipulation and show minimal pathogenicity. The review summarizes the data on the directions and approaches to generation of highly efficient variants of oncolytic adenoviruses. The approaches include introduction of directed genetic modifications into the virus genome, accelerated selection of oncolytic virus variants following treatment with mutagens, the use of adenoviruses as vectors to introduce therapeutic gene products, optimization of viral delivery systems, minimization of the negative effects from the host immune system, etc. The dynamic development of studies in the field holds promise that many variants of oncolytic adenoviruses will find clinical application in the nearest future.     

Viral diseases in commercially exploited crabs: a review

Viruses and viral diseases of crabs were observed and investigated earlier than the first observation of viruses in shrimp. In fact, crabs were used as biological models to investigate crustacean virology at the beginning of shrimp aquaculture development. More than 30 viruses have been reported in crabs, including those related to the known virus families Reoviridae, Bunyaviridae, Roniviridae and a group of Bacilliform enveloped nuclear viruses. This review reports data on several important viral diseases of crabs, particularly those associated with pathology of organs and tissues of commercially and ecologically significant host species.     

Essential role of dengue virus envelope protein N glycosylation at asparagine-67 during viral propagation

Dengue virus envelope protein (E) contains two N-linked glycosylation sites, at Asn-67 and Asn-153. The glycosylation site at position 153 is conserved in most flaviviruses, while the site at position 67 is thought to be unique for dengue viruses. N-linked oligosaccharide side chains on flavivirus E proteins have been associated with viral morphogenesis, infectivity, and tropism. Here, we examined the relevance of each N-linked glycan on dengue virus E protein by removing each site in the context of infectious viral particles. Dengue viruses lacking Asn-67 were able to infect mammalian cells and translate and replicate the viral genome, but production of new infectious particles was abolished. In addition, dengue viruses lacking Asn-153 in the E showed reduced infectivity. In contrast, ablation of one or both glycosylation sites yielded viruses that replicate and propagate in mosquito cells. Furthermore, we found a differential requirement of N-linked glycans for E secretion in mammalian and mosquito cells. While secretion of E lacking Asn-67 was efficient in mosquito cells, secretion of the same protein expressed in mammalian cells was dramatically impaired. Finally, we found that viruses lacking the carbohydrate at position 67 showed reduced infection of immature dendritic cells, suggesting interaction between this glycan and the lectin DC-SIGN. Overall, our data defined different roles for the two glycans present at the E protein during dengue virus infection, highlighting the involvement of distinct host functions from mammalian and mosquito cells during dengue virus propagation.     

Entry of Influenza A Virus with a α2,6-Linked Sialic Acid Binding Preference Requires Host Fibronectin

The receptor binding specificity of influenza A virus is one of the major determinants of viral tropism and host specificity. In general, avian viral hemagglutinin prefers to bind to α2,3-linked sialic acid, whereas the human viral hemagglutinin prefers to bind to α2,6-linked sialic acid. Here, we demonstrate that host fibronectin protein plays an important role in the life cycle of some influenza A viruses. Treating cells with anti-fibronectin antibodies or fibronectin-specific small interfering RNA can inhibit the virus replication of human H1N1 influenza A viruses. Strikingly, these inhibitory effects cannot be observed in cells infected with H5N1 viruses. By using reverse genetics techniques, we observed that the receptor binding specificity, but not the origin of the hemagglutinin subtype, is responsible for this differential inhibitory effect. Changing the binding preference of hemagglutinin from α2,6-linked sialic acid to α2,3-linked sialic acid can make the virus resistant to the anti-fibronectin antibody treatment and vice versa. Our further characterizations indicate that anti-fibronectin antibody acts on the early phase of viral replication cycle, but it has no effect on the initial binding of influenza A virus to cell surface. Our subsequent investigations further show that anti-fibronectin antibody can block the postattachment entry of influenza virus. Overall, these results indicate that the sialic acid binding preference of influenza viral hemagglutinin can modulate the preferences of viral entry pathways, suggesting that there are subtle differences between the virus entries of human and avian influenza viruses.     

Viral Host Jumps: Moving toward a Predictive Framework

In order to predict pathogen emergence, we must distinguish between emergence phenomena that occur via different processes. Focusing on the appearance of viral pathogens in new host species, I outline a framework that uses specific molecular characteristics to rank virus families by their expected a priori ability to complete each of three steps in the emergence process (encounter, infection, and propagation). I then discuss the degree to which the patterns expected, based solely on molecular-level structural characteristics, agree with observations regarding the ability of animal viruses to infect humans. This approach yields predictions consistent with empirical observations regarding the ability of specific viral families to infect novel host species but highlights the need for consideration of other factors, such as the ecology of host interactions and the determinants of cellular susceptibility and permissivity to specific virus groups, when trying to predict the frequency with which a virus will encounter a novel host species or the probability of propagation within a novel host species once infection has occurred.     

A Survey of Overlooked Viral Infections in Biological Experiment Systems

It is commonly accepted that there are many unknown viruses on the planet. For the known viruses, do we know their prevalence, even in our experimental systems? Here we report a virus survey using recently published small (s)RNA sequencing datasets. The sRNA reads were assembled and contigs were screened for virus homologues against the NCBI nucleotide (nt) database using the BLASTn program. To our surprise, approximately 30% (28 out of 94) of publications had highly scored viral sequences in their datasets. Among them, only two publications reported virus infections. Though viral vectors were used in some of the publications, virus sequences without any identifiable source appeared in more than 20 publications. By determining the distributions of viral reads and the antiviral RNA interference (RNAi) pathways using the sRNA profiles, we showed evidence that many of the viruses identified were indeed infecting and generated host RNAi responses. As virus infections affect many aspects of host molecular biology and metabolism, the presence and impact of viruses needs to be actively investigated in experimental systems.     

Naturally Occurring Swine Influenza A Virus PB1-F2 Phenotypes That Contribute to Superinfection with Gram-Positive Respiratory Pathogens

A combination of viral, bacterial, and host factors contributes to the severity and overall mortality associated with influenza virus-bacterium superinfections. To date, the virulence associated with the recently identified influenza virus protein PB1-F2 has been largely defined using models of primary influenza virus infection, with only limited assessment in models of Streptococcus pneumoniae superinfection. Specifically, these studies have incorporated isogenic viruses that differ in the PB1-F2 expressed, but there is still knowledge to be gained from evaluation of natural variants derived from a nonhuman host species (swine). Using this rationale, we developed the hypothesis that naturally occurring viruses expressing variants of genes, like the PB1-F2 gene, can be associated with the severity of secondary bacterial infections. To test this hypothesis, we selected viruses expressing variants in PB1-F2 and evaluated outcomes from superinfection with three distinct Gram-positive respiratory pathogens: Streptococcus pneumoniae, Staphylococcus aureus, and Streptococcus pyogenes. Our results demonstrate that the amino acid residues 62L, 66S, 75R, 79R, and 82L, previously proposed as molecular signatures of PB1-F2 virulence for influenza viruses in the setting of bacterial superinfection, are broadly associated with enhanced pathogenicity in swine in a bacterium-specific manner. Furthermore, truncated PB1-F2 proteins can preferentially increase mortality when associated with Streptococcus pyogenes superinfection. These findings support efforts to increase influenza virus surveillance to consider viral genotypes that could be used to predict increased severity of superinfections with specific Gram-positive respiratory pathogens.