Accurate virus identification with interpretable Raman signatures by machine learning.

Rapid identification of newly emerging or circulating viruses is an important first step toward managing the public health response to potential outbreaks. A portable virus capture device, coupled with label-free Raman spectroscopy, holds the promise of fast detection by rapidly obtaining the Raman signature of a virus followed by a machine learning (ML) approach applied to recognize the virus based on its Raman spectrum, which is used as a fingerprint. We present such an ML approach for analyzing Raman spectra of human and avian viruses. A convolutional neural network (CNN) classifier specifically designed for spectral data achieves very high accuracy for a variety of virus type or subtype identification tasks. In particular, it achieves 99% accuracy for classifying influenza virus type A versus type B, 96% accuracy for classifying four subtypes of influenza A, 95% accuracy for differentiating enveloped and nonenveloped viruses, and 99% accuracy for differentiating avian coronavirus (infectious bronchitis virus [IBV]) from other avian viruses. Furthermore, interpretation of neural net responses in the trained CNN model using a full-gradient algorithm highlights Raman spectral ranges that are most important to virus identification. By correlating ML-selected salient Raman ranges with the signature ranges of known biomolecules and chemical functional groups­for example, amide, amino acid, and carboxylic acid­we verify that our ML model effectively recognizes the Raman signatures of proteins, lipids, and other vital functional groups present in different viruses and uses a weighted combination of these signatures to identify viruses.

High-resolution asymmetric structure of a Fab-virus complex reveals overlap with the receptor binding site.

Canine parvovirus is an important pathogen causing severe diseases in dogs, including acute hemorrhagic enteritis, myocarditis, and cerebellar disease. Overlap on the surface of parvovirus capsids between the antigenic epitope and the receptor binding site has contributed to cross-species transmission, giving rise to closely related variants. It has been shown that Mab 14 strongly binds and neutralizes canine but not feline parvovirus, suggesting this antigenic site also controls species-specific receptor binding. To visualize the conformational epitope at high resolution, we solved the cryogenic electron microscopy (cryo-EM) structure of the Fab-virus complex. We also created custom software, Icosahedral Subparticle Extraction and Correlated Classification, to solve a Fab-virus complex with only a few Fab bound per capsid and visualize local structures of the Fab-bound and -unbound antigenic sites extracted from the same complex map. Our results identified the antigenic epitope that had significant overlap with the receptor binding site, and the structures revealed that binding of Fab induced conformational changes to the virus. We were also able to assign the order and position of attached Fabs to allow assessment of complementarity between the Fabs bound to different positions. This approach therefore provides a method for using cryo-EM to investigate complementarity of antibody binding.

Honey Bee Deformed Wing Virus Structures Reveal that Conformational Changes Accompany Genome Release.

The picornavirus-like deformed wing virus (DWV) has been directly linked to colony collapse; however, little is known about the mechanisms of host attachment or entry for DWV or its molecular and structural details. Here we report the three-dimensional (3-D) structures of DWV capsids isolated from infected honey bees, including the immature procapsid, the genome-filled virion, the putative entry intermediate (A-particle), and the empty capsid that remains after genome release. The capsids are decorated by large spikes around the 5-fold vertices. The 5-fold spikes had an open flower-like conformation for the procapsid and genome-filled capsids, whereas the putative A-particle and empty capsids that had released the genome had a closed tube-like spike conformation. Between the two conformations, the spikes undergo a significant hinge-like movement that we predicted using a Robetta model of the structure comprising the spike. We conclude that the spike structures likely serve a function during host entry, changing conformation to release the genome, and that the genome may escape from a 5-fold vertex to initiate infection. Finally, the structures illustrate that, similarly to picornaviruses, DWV forms alternate particle conformations implicated in assembly, host attachment, and RNA release. IMPORTANCE: Honey bees are critical for global agriculture, but dramatic losses of entire hives have been reported in numerous countries since 2006. Deformed wing virus (DWV) and infestation with the ectoparasitic mite Varroa destructor have been linked to colony collapse disorder. DWV was purified from infected adult worker bees to pursue biochemical and structural studies that allowed the first glimpse into the conformational changes that may be required during transmission and genome release for DWV.

Single Mutations in the VP2 300 Loop Region of the Three-Fold Spike of the Carnivore Parvovirus Capsid Can Determine Host Range.

UNLABELLED: Sylvatic carnivores, such as raccoons, have recently been recognized as important hosts in the evolution of canine parvovirus (CPV), a pandemic pathogen of domestic dogs. Although viruses from raccoons do not efficiently bind the dog transferrin receptor (TfR) or infect dog cells, a single mutation changing an aspartic acid to a glycine at capsid (VP2) position 300 in the prototype raccoon CPV allows dog cell infection. Because VP2 position 300 exhibits extensive amino acid variation among the carnivore parvoviruses, we further investigated its role in determining host range by analyzing its diversity and evolution in nature and by creating a comprehensive set of VP2 position 300 mutants in infectious clones. Notably, some position 300 residues rendered CPV noninfectious for dog, but not cat or fox, cells. Changes of adjacent residues (residues 299 and 301) were also observed often after cell culture passage in different hosts, and some of the mutations mimicked changes seen in viruses recovered from natural infections of alternative hosts, suggesting that compensatory mutations were selected to accommodate the new residue at position 300. Analysis of the TfRs of carnivore hosts used in the experimental evolution studies demonstrated that their glycosylation patterns varied, including a glycan present only on the domestic dog TfR that dictates susceptibility to parvoviruses. Overall, there were significant differences in the abilities of viruses with alternative position 300 residues to bind TfRs and infect different carnivore hosts, demonstrating that the process of infection is highly host dependent and that VP2 position 300 is a key determinant of host range. IMPORTANCE: Although the emergence and pandemic spread of canine parvovirus (CPV) are well documented, the carnivore hosts and evolutionary pathways involved in its emergence remain enigmatic. We recently demonstrated that a region in the capsid structure of CPV, centered around VP2 position 300, varies after transfer to alternative carnivore hosts and may allow infection of previously nonsusceptible hosts in vitro. Here we show that VP2 position 300 is the most variable residue in the parvovirus capsid in nature, suggesting that it is a critical determinant in the cross-species transfer of viruses between different carnivores due to its interactions with the transferrin receptor to mediate infection. To this end, we demonstrated that there are substantial differences in receptor binding and infectivity of various VP2 position 300 mutants for different carnivore species and that single mutations in this region can influence whether a host is susceptible or refractory to virus infection.

Near-Atomic Resolution Structure of a Highly Neutralizing Fab Bound to Canine Parvovirus.

Canine parvovirus (CPV) is a highly contagious pathogen that causes severe disease in dogs and wildlife. Previously, a panel of neutralizing monoclonal antibodies (MAb) raised against CPV was characterized. An antibody fragment (Fab) of MAb E was found to neutralize the virus at low molar ratios. Using recent advances in cryo-electron microscopy (cryo-EM), we determined the structure of CPV in complex with Fab E to 4.1 Å resolution, which allowed de novo building of the Fab structure. The footprint identified was significantly different from the footprint obtained previously from models fitted into lower-resolution maps. Using single-chain variable fragments, we tested antibody residues that control capsid binding. The near-atomic structure also revealed that Fab binding had caused capsid destabilization in regions containing key residues conferring receptor binding and tropism, which suggests a mechanism for efficient virus neutralization by antibody. Furthermore, a general technical approach to solving the structures of small molecules is demonstrated, as binding the Fab to the capsid allowed us to determine the 50-kDa Fab structure by cryo-EM. IMPORTANCE: Using cryo-electron microscopy and new direct electron detector technology, we have solved the 4 Å resolution structure of a Fab molecule bound to a picornavirus capsid. The Fab induced conformational changes in regions of the virus capsid that control receptor binding. The antibody footprint is markedly different from the previous one identified by using a 12 Å structure. This work emphasizes the need for a high-resolution structure to guide mutational analysis and cautions against relying on older low-resolution structures even though they were interpreted with the best methodology available at the time.

Group Selection and Contribution of Minority Variants during Virus Adaptation Determines Virus Fitness and Phenotype.

Understanding how a pathogen colonizes and adapts to a new host environment is a primary aim in studying emerging infectious diseases. Adaptive mutations arise among the thousands of variants generated during RNA virus infection, and identifying these variants will shed light onto how changes in tropism and species jumps can occur. Here, we adapted Coxsackie virus B3 to a highly permissive and less permissive environment. Using deep sequencing and bioinformatics, we identified a multi-step adaptive process to adaptation involving residues in the receptor footprints that correlated with receptor availability and with increase in virus fitness in an environment-specific manner. We show that adaptation occurs by selection of a dominant mutation followed by group selection of minority variants that together, confer the fitness increase observed in the population, rather than selection of a single dominant genotype.

Global displacement of canine parvovirus by a host-adapted variant: structural comparison between pandemic viruses with distinct host ranges.

Canine parvovirus type 2 (CPV-2) emerged in 1978 and spread worldwide within 2 years. Subsequently, CPV-2 was completely replaced by the variant CPV-2a, which is characterized by four specific capsid (VP2) mutations. The X-ray crystal structure of the CPV-2a capsid shows that each mutation confers small local changes. The loss of a hydrogen bond and introduction of a glycine residue likely introduce flexibility to sites that control interactions with the host receptor, antibodies, and sialic acids.

Specificity of coxsackievirus B3 interaction with human, but not murine, decay-accelerating factor: replacement of a single residue within short consensus repeat 2 prevents virus attachment.

UNLABELLED: Many coxsackievirus B (CVB) isolates bind to human decay-accelerating factor (DAF) as well as to the coxsackievirus and adenovirus receptor (CAR). However, the virus does not interact with murine DAF. To understand why CVB3 binds specifically to human DAF, we constructed a series of chimeric molecules in which specific regions of the human DAF molecule were replaced by the corresponding murine sequences. We found that replacement of human short consensus repeat 2 (SCR2) with murine SCR2 ablated virus binding to human DAF, as did deletion of human SCR2. Although replacement of human SCR4 had a partial inhibitory effect, deletion of SCR4 had no effect. Within human SCR2, replacement of serine 104 (S104) with the proline residue found in murine DAF eliminated virus binding. On the basis of the structure of the CVB3-DAF complex determined by cryo-electron microscopy, DAF S104 is in close contact with a viral capsid residue, a threonine at VP1 position 271. Replacement of this capsid residue with larger amino acids specifically eliminated virus attachment to human DAF but had no effect on attachment to CAR or replication in HeLa cells. Taken together, these results support the current model of virus-DAF interaction and point to a specific role for VP1 T271 and DAF S104 at the virus-DAF interface. IMPORTANCE: The results of the present study point to a specific role for VP1 T271 and DAF S104 at the interface between CVB3 and DAF, and they demonstrate how subtle structural changes can dramatically influence virus-receptor interactions. In addition, the results support a recent pseudoatomic model of the CVB3-DAF interaction obtained by cryo-electron microscopy.

Kinetic and structural analysis of coxsackievirus B3 receptor interactions and formation of the A-particle.

UNLABELLED: The coxsackievirus and adenovirus receptor (CAR) has been identified as the cellular receptor for group B coxsackieviruses, including serotype 3 (CVB3). CAR mediates infection by binding to CVB3 and catalyzing conformational changes in the virus that result in formation of the altered, noninfectious A-particle. Kinetic analyses show that the apparent first-order rate constant for the inactivation of CVB3 by soluble CAR (sCAR) at physiological temperatures varies nonlinearly with sCAR concentration. Cryo-electron microscopy (cryo-EM) reconstruction of the CVB3-CAR complex resulted in a 9.0-Šresolution map that was interpreted with the four available crystal structures of CAR, providing a consensus footprint for the receptor binding site. The analysis of the cryo-EM structure identifies important virus-receptor interactions that are conserved across picornavirus species. These conserved interactions map to variable antigenic sites or structurally conserved regions, suggesting a combination of evolutionary mechanisms for receptor site preservation. The CAR-catalyzed A-particle structure was solved to a 6.6-Šresolution and shows significant rearrangement of internal features and symmetric interactions with the RNA genome. IMPORTANCE: This report presents new information about receptor use by picornaviruses and highlights the importance of attaining at least an ∼9-Šresolution for the interpretation of cryo-EM complex maps. The analysis of receptor binding elucidates two complementary mechanisms for preservation of the low-affinity (initial) interaction of the receptor and defines the kinetics of receptor-catalyzed conformational change to the A-particle.

Assembly and characterization of foot-and-mouth disease virus empty capsid particles expressed within mammalian cells.

The foot-and-mouth disease virus (FMDV) structural protein precursor, P1-2A, is cleaved by the virus-encoded 3C protease (3C(pro)) into the capsid proteins VP0, VP1 and VP3 (and 2A). In some systems, it is difficult to produce large amounts of these processed capsid proteins since 3C(pro) can be toxic for cells. The expression level of 3C(pro) activity has now been reduced relative to the P1-2A, and the effect on the yield of processed capsid proteins and their assembly into empty capsid particles within mammalian cells has been determined. Using a vaccinia-virus-based transient expression system, P1-2A (from serotypes O and A) and 3C(pro) were expressed from monocistronic cDNA cassettes as P1-2A-3C, or from dicistronic cassettes with the 3C(pro) expression dependent on a mutant FMDV internal ribosome entry site (IRES) (designated P1-2A-mIRES-3C). The effects of using a mutant 3C(pro) with reduced catalytic activity or using two different mutant IRES elements (the wt GNRA tetraloop sequence GCGA converted, in the cDNA, to GAGA or GTTA) were analysed. For both serotypes, the P1-2A-mIRES-3C construct containing the inefficient GTTA mutant IRES produced the highest amount of processed capsid proteins. These products self-assembled to form FMDV empty capsid particles, which have a related, but distinct, morphology (as determined by electron microscopy and reconstruction) from that determined previously by X-ray crystallography. The assembled empty capsids bind, in a divalent cation-dependent manner, to the RGD-dependent integrin αvß6, a cellular receptor for FMDV, and are recognized appropriately in serotype-specific antigen ELISAs.

Cryo-EM maps reveal five-fold channel structures and their modification by gatekeeper mutations in the parvovirus minute virus of mice (MVM) capsid.

In minute virus of mice (MVM) capsids, icosahedral five-fold channels serve as portals mediating genome packaging, genome release, and the phased extrusion of viral peptides. Previous studies suggest that residues L172 and V40 are essential for channel function. The structures of MVMi wildtype, and mutant L172T and V40A virus-like particles (VLPs) were solved from cryo-EM data. Two constriction points, termed the mid-gate and inner-gate, were observed in the channels of wildtype particles, involving residues L172 and V40 respectively. While the mid-gate of V40A VLPs appeared normal, in L172T adjacent channel walls were altered, and in both mutants there was major disruption of the inner-gate, demonstrating that direct L172:V40 bonding is essential for its structural integrity. In wildtype particles, residues from the N-termini of VP2 map into claw-like densities positioned below the channel opening, which become disordered in the mutants, implicating both L172 and V40 in the organization of VP2 N-termini.

The novel asymmetric entry intermediate of a picornavirus captured with nanodiscs.

Many nonenveloped viruses engage host receptors that initiate capsid conformational changes necessary for genome release. Structural studies on the mechanisms of picornavirus entry have relied on in vitro approaches of virus incubated at high temperatures or with excess receptor molecules to trigger the entry intermediate or A-particle. We have induced the coxsackievirus B3 entry intermediate by triggering the virus with full-length receptors embedded in lipid bilayer nanodiscs. These asymmetrically formed A-particles were reconstructed using cryo-electron microscopy and a direct electron detector. These first high-resolution structures of a picornavirus entry intermediate captured at a membrane with and without imposing icosahedral symmetry (3.9 and 7.8 Å, respectively) revealed a novel A-particle that is markedly different from the classical A-particles. The asymmetric receptor binding triggers minimal global capsid expansion but marked local conformational changes at the site of receptor interaction. In addition, viral proteins extrude from the capsid only at the site of extensive protein remodeling adjacent to the nanodisc. Thus, the binding of the receptor triggers formation of a unique site in preparation for genome release.

Enterovirus 71 Virus Propagation and Purification.

Since its discovery in 1969, enterovirus 71 (EV71) has emerged as a serious worldwide health threat. This member of the picornavirus family causes hand, foot, and mouth disease, and also has the capacity to invade the central nervous system to cause severe disease and death. This is the propagation and purification procedure to produce infectious virion.

In VITRO ASSEMBLY of the øX174 procapsid from external scaffolding protein oligomers and early pentameric assembly intermediates.

Bacteriophage øX174 morphogenesis requires two scaffolding proteins: an internal species, similar to those employed in other viral systems, and an external species, which is more typically associated with satellite viruses. The current model of øX174 assembly is based on structural and in vivo data. During morphogenesis, 240 copies of the external scaffolding protein mediate the association of 12 pentameric particles into procapsids. The hypothesized pentameric intermediate, the 12S⁎ particle, contains 16 proteins: 5 copies each of the coat, spike and internal scaffolding proteins and 1 copy of the DNA pilot protein. Assembly naïve 12S⁎ particles and external scaffolding oligomers, most likely tetramers, formed procapsid-like particles in vitro, suggesting that the 12S⁎ particle is a bona fide assembly intermediate and validating the current model of procapsid morphogenesis. The in vitro system required a crowding agent, was influenced by the ratio of the reactants and was most likely driven by hydrophobic forces. While the system reported here shared some characteristics with other in vitro internal scaffolding protein-mediated systems, it displayed unique features. These features most likely reflect external scaffolding protein-mediated morphogenesis and the øX174 procapsid structure, in which external scaffolding-scaffolding protein interactions, as opposed to coat-coat protein interactions between pentamers, constitute the primary lattice-forming contacts.