Journal of Virology

Recombination can generate genetic diversity among RNA viruses; yet, mechanisms governing recombination for segmented, double-stranded RNA viruses remain poorly understood. Reovirus genomes contain ten segments of double-stranded RNA. Previous analyses suggested low rates of reovirus RNA recombination and led to a proposed mechanism of recombination during minus-strand synthesis. These analyses focused on recombination events detected in genomic double-stranded RNA extracted from purified viral particles. In the current study, we looked deeper into the origins of the observed recombination events by assaying total cellular material, which comprises viral genomic material, as well as replication intermediates, mRNAs, and products of recombination that may not be eligible for packaging into virus particles. Intracellular viral RNA, which is almost exclusively plus-strand RNA, exhibited substantially higher recombination frequency than particle-associated double-stranded RNA. The diversity and frequency of identified junctions were more similar among samples of a given type (intracellular or virion-encapsidated) from distinct infections than between distinct sample types from the same infection. Although potential recombination sites are widely distributed across the genome, a more limited set was used in junctions in packaged viral RNA than in intracellular viral RNA. Together, these observations suggest recombination may occur more frequently than previously recognized during reovirus infection, and many such events likely occur during plus-strand synthesis. However, recombination products generated during viral transcription may be selectively incorporated into viral particles.

The live, attenuated vaccine for Varicella-Zoster Virus (VZV) protects against chickenpox in children and adults. However, the vaccine strain Oka (vOka) establishes latency and may reactivate. The molecular basis of vOka attenuation remains unknown. We investigated whether three high-frequency single-nucleotide polymorphisms (SNPs) that differentiate vOka from parental strain Oka (pOka) are responsible: S628G and R958G in the immediate-early regulatory protein IE62, and *130R in the membrane protein ORF0. We evaluated how these fixed SNPs mediate the attenuated growth phenotype of VZV and diminish skin pathology by creating recombinant pOka VZV carrying these SNPs in the wild-type background. These were assessed in cells and human skin organ cultures (SOCs). In epithelial and fibroblast cell lines, the individual IE62 mutations slightly delayed viral growth at 8 - 24 hours post-infection but not in human epidermal keratinocytes (htert-HEK). The fixed ORF62 SNPs also affected ORF62 transcription, IE62 protein abundance at early times, and delayed IE62 accumulation in the cytoplasm later during infection. The *130R mutation in ORF0 conferred a growth advantage in culture, suggesting it is an adaptation to cell culture. While there was donor-to-donor variability and differences in overall virus spread in SOC, the R958G and *130R conferred an attenuated growth phenotype similar to that of vOka in SOC. Histopathologic analysis of VZV-infected skin sections revealed that, similar to vOka, S628G and R958G in IE62 limited the spread of VZV skin lesions. Our findings underscore the critical roles of conserved vOka SNPs in understanding the attenuation of the live vaccine strain vOka.

Herpes simplex virus-1 (HSV-1) has infected more than 67% of the world population causing diseases that range in severity from benign cold sores to encephalitis. To infect and persist in the host, HSV-1 has evolved sophisticated strategies to counteract antiviral responses. The viral immediate early protein ICP0 (Infected Cells Protein 0) plays a fundamental role in this process. ICP0 is an E3 ubiquitin ligase that is required for successful onset of lytic infection and productive reactivation of viral genomes from latency. Most ICP0 studies have been focused on its nuclear functions, but late gene expression drives accumulation of ICP0 to the cytoplasm where its functions are unknown. We recently demonstrated that cytoplasmic ICP0 localizes on the surface of endosomes through its interaction with CIN85 (Cbl-interacting protein of 85 kDa), a scaffolding protein for factors involved in endocytosis. This interaction is required for extrusion of host antiviral factors and immunoevasion. Here, we report that the E3 ubiquitin ligase activity of ICP0 is also required for these effects, as disruption of ICP0 RING finger domain, which disrupts its catalytic activity, resulted in an increase in the number of CIN85 endosomes that cannot be exocytosed. These endosomes colocalize with autophagosome components, but they do not enter the lysosomal pathway for degradation. Mutant viruses that cannot exocytose CIN85 endosomes produce EVs that are more potent at activating antiviral responses.

Kaposi's sarcoma-associated herpesvirus (KSHV) is a human tumor virus. Common marmosets, a potential non-human primate model for vaccine development and infection studies, have previously been shown to support KSHV infection. We established infection of marmosets with BAC-derived, recombinant KSHV to analyze the role of individual glycoproteins (GPs) in tissue tropism. Three groups of four animals were infected intravenously with BAC-derived KSHV wild type (wt), EphA2-receptor-binding-deficient KSHV gHASAELAAN, or KSHV{triangleup}K8.1, lacking glycoprotein K8.1. Viral DNA was detected in multiple tissues of all animals, with the highest loads in spleen, liver, heart, and bone marrow. In most tissues, viral DNA loads did not significantly differ between groups. However, KSHV{triangleup}K8.1-infected animals exhibited significantly lower KSHV DNA levels in liver and bone marrow. KSHV wt infection induced activation of CD4+ and CD8+ T-cells. In animals infected with KSHV wt, KSHV gHASAELAAN or KSHV{triangleup}K8.1, different B cell subpopulations expanded after infection, which suggests that the glycoproteins themselves or the cell tropism imparted by their presence or absence might shape the host response to KSHV. In summary, our findings demonstrate that neither the Eph receptor interaction nor GP K8.1 is essential for intravenous infection, but K8.1 critically contributes to KSHV tissue tropism in vivo.

Respiratory syncytial virus (RSV) is an enveloped virus that assembles and buds at the plasma membrane, producing filamentous infectious particles. Unlike many other enveloped viruses, a substantial fraction of RSV virions remains associated with the surface of infected cells, limiting our understanding of the membrane-scission mechanisms responsible for particle release. Here, we identified the actin-binding protein drebrin as a cellular interactor of the RSV matrix (M) protein, both in vitro and in infected cells. We demonstrated that RSV infection induces the relocalization of drebrin to actin-rich structures at the cell periphery associated with viral filaments. Using a Transwell-based assay to quantify infectious particles release without mechanically perturbing cells, we confirmed that RSV release is independent of the endosomal sorting complex required for transport (ESCRT) pathway and showed that drebrin depletion reduced RSV release, whereas drebrin overexpression enhanced it. Since drebrin depletion did not significantly affect overall infectious virus production, our results suggest that drebrin may act during the final step leading to membrane scission and RSV particle release. Finally, proteomics analysis identified Transmembrane 9 superfamily member 2 (TM9SF2) as a drebrin-associated protein specifically during infection. TM9SF2 depletion reduced RSV particle release and impaired the RSV-induced relocalization of drebrin to the plasma membrane. Together, these findings identify drebrin and TM9SF2 as host factors regulating RSV release and support a model in which TM9SF2 promotes the recruitment of drebrin to viral budding sites.

Interferon gamma (IFN-{gamma}) is a potent cytokine secreted following primary varicella-zoster virus (VZV) infection. Our previous studies demonstrated that pretreatment with IFN-{gamma} completely inhibited VZV replication in lung fibroblast MRC-5 and retinal epithelial ARPE-19 but not in melanoma MeWo cells, suggesting that IFN-{gamma}-stimulated protein(s) inhibit viral replication. Our microarray analysis revealed that 34 interferon-stimulated genes (ISGs) were upregulated by greater than 2.5-fold at 8 h post-treatment in both ARPE-19 and MRC-5 cells compared to those of MeWo cells. The depletion of CTSS, GBP4, MX1, IFIT1, IFITM1 and IRF1 by siRNA in IFN-{gamma}-treated cells significantly increased VZV yields. Ectopic overexpression of interferon-induced myxovirus resistance 1 (MX1) or transmembrane protein 1 (IFITM1) significantly reduced VZV yield both in ARPE-19 and MeWo cells. IFITM1 abrogated viral major immediate-early IE62 protein-mediated trans-activation. IFITM1 reduced the level of VZV immediate-early IE62 protein, but not the IE62 mRNA, suggesting that IFITM1 expression reduces the expression level of IE62 by post-transcriptional regulation. Cell-free VZV failed to infect human primary epidermal keratinocytes, in which the basal level of endogenous MX1 gene expression is very high. Pharmacologic inhibition of JAK signaling reduced MX1 expression and enhanced VZV replication in calcium chloride-induced differentiated keratinocytes. Collectively, these findings identify MX1 and IFITM1 as key IFN-{gamma}-induced restriction factors that inhibit VZV replication and suggest that suppression of JAK-STAT signaling can promote VZV replication in differentiated keratinocytes.

Newborns and infants are particularly vulnerable to influenza, largely due to the immaturity of their immune systems. As such, maternal immunization is a valuable strategy to enhance influenza-neutralizing antibodies in both blood and breast milk, ensuring optimal passive immunity transfer to the infant. In our pregnant and lactating ferret maternal-neonatal model, we evaluated antibody responses to an adenovirus type 5-vectored H1N1 hemagglutinin vaccine delivered enterically, intranasally, or intramuscularly. While all three routes elicited a robust antibody response compared to saline-treated controls, intranasal vaccination significantly increased influenza-specific antibodies in dam serum, upper respiratory tract secretions, and milk, and subsequent transfer to suckling kits. Our results suggest that intranasal immunization was the most effective route for inducing functional influenza-specific antibodies in both dam serum and milk, with efficient passive transfer to suckling kits. Leveraging mucosal immune induction may enhance maternal antibody responses, transfer to offspring, and downstream neonatal protection.

Infectious bronchitis virus (IBV) is an economically important avian gammacoronavirus. IBV uses sialic acids and heparan sulfate for cellular attachment, but the host proteins governing its infection remain poorly defined. Here, we investigated the roles of angiotensin-converting enzyme 2 (ACE2) and transmembrane serine protease 2 (TMPRSS2) during IBV Beaudette (IBV BD) infection. ACE2 knockout impaired IBV replication in DF-1 cells, whereas ACE2 reconstitution restored infection. Furthermore, ACE2 overexpression enhanced infection of DF-1 and primary chicken embryo kidney cells, supporting the role of chicken ACE2 as a proviral host factor. In contrast, TMPRSS2 exerted a restrictive function on IBV BD infection as demonstrated in TMPRSS2 KO DF-1 cells and upon pharmacological inhibition, which significantly increased IBV BD infection in DF-1 and Vero cells. RNA sequencing revealed significant changes in lipid metabolism and membrane-related processes following TMPRSS2 knockout and pharmacological inhibition. These findings highlight the opposing roles of ACE2 and TMPRSS2 in IBV BD infection, emphasizing the combined impact of proviral host factors, protease activity, and the cellular membrane environment during gammacoronavirus infections.

HSV-1 is a neurotropic human pathogen that establishes lifelong latency in sensory neurons. Periodic reactivation causes mucocutaneous lesions and, in some cases, severe corneal and neurological disease. While HSV-1 utilizes numerous viral proteins to extensively reprogram host signaling pathways and evade antiviral responses, how it alters host cytokine receptor networks remains poorly understood. Here, we demonstrate that HSV-1 infection selectively downmodulates multiple members of the gp130 (glycoprotein 130) cytokine receptor family, including gp130, Oncostatin M receptor {beta} (OSMR-{beta}), and leukemia inhibitory factor receptor (LIFR), while sparing IL-6Ra and CNTFRa (Ciliary Neurotrophic Factor Receptor). Because gp130 is the central signaling component of the IL-6 cytokine family and mediates JAK/STAT3 activation, its elimination likely represents a viral strategy to suppress inflammatory and antiviral signaling. Consistent with this, HSV-1 does not induce IL-6 mRNA expression in productively infected cells, suggesting coordinated inhibition of both cytokine production and receptor availability. Mechanistically, we found that the viral RNase VHS is at least partially responsible for gp130 family receptor downmodulation, likely in combination with virus-induced endocytosis/degradation pathways. Functionally, activation of gp130 signaling significantly impairs HSV-1 infection, demonstrating that suppression of antiviral cytokine responses is critical for virus replication. Finally, we determined that CNTF causes reactivation of latent HSV-1 using the LUHMES in vitro latency model, illustrating CNTFRa signaling contribution to latency and reactivation. A similar effect has been reported for IL-6. Together, these findings reveal a previously unrecognized immune evasion mechanism by which HSV-1 remodels gp130 cytokine signaling to promote infection.

Protein kinase R (PKR) is an antiviral sensor that induces translational shutdown and stress granule (SG) formation through the phosphorylation of eIF2. Although previous studies have proposed that hepatitis C virus (HCV) non-structural protein 5A (NS5A) inhibits PKR dimerization, other studies consistently observe PKR activation and SG formation during HCV infection. To address this discrepancy, we interrogated the ability of NS5A from different genotypes and strains to regulate canonical PKR-eIF2 activation in a variety of cell culture models. Regardless of the genotype or strain tested, HCV NS5A failed to inhibit PKR dimerization or dsRNA-induced PKR activation, nor did it rescue downstream steps in the PKR-eIF2 pathway, such as translation halt and SG formation, in contrast to known viral PKR antagonists. Similarly, NS5A did not antagonise PKR activation when expressed in the context of other HCV non-structural proteins. We also tested the role of the host factor cyclophilin A (CypA), which interacts with both NS5A and PKR. However, the absence of CypA did not affect downstream PKR signaling or the activity of viral PKR antagonists or NS5A. Collectively, these findings indicate that NS5A does not effectively antagonise the PKR-eIF2 pathway and further support previous findings that HCV replication occurs in the presence of an active PKR response.

The Herpes Simplex Virus 1 (HSV-1) genome undergoes extensive chromatinization following entry into the nucleus, yet how viral regulatory elements influence this process during lytic infection remains poorly understood. The HSV-1 latency-associated transcript enhancer (LTE) is well established as a key element in the regulation of latency and reactivation, but its role during the productive infection has not been defined. Here, we identify a role for the LTE in regulating chromatin composition on the lytic HSV-1 genome. Deletion of a 310-bp region within the LTE selectively reduced enrichment of the canonical histone H3.1 at specific viral lytic genes, including ICP4 and ICP8, without altering H3.3 enrichment. This phenotype was not reproduced using a second LTE deletion that disrupted enhancer activity, indicating that regulation of H3.1 enrichment is separable from the function of the enhancer. LTE-dependent H3.1 enrichment persisted in the absence of viral DNA replication and was accompanied by reduced occupancy of the H3.1 histone chaperone CAF-1. However, depletion of CAF-1 did not substantially reduce H3.1 enrichment, indicating that additional mechanisms contribute to H3.1 association with the viral genome. Together, these findings reveal that the HSV-1 LTE regulates locus-specific chromatin composition during early lytic infection and identify a previously unrecognized connection between a viral enhancer and histone H3.1 enrichment on the incoming viral genome.

Laguna Negra virus and Black Creek Canal virus are hantaviruses associated with hantavirus cardiopulmonary syndrome (HCPS), but their cellular entry receptors have not been defined. We used replication-competent recombinant vesicular stomatitis viruses (rVSVs) bearing hantavirus Gn/Gc glycoprotein complexes to compare entry by these viruses, the Old World rodent-borne Dabieshan virus and Tigray virus, and the shrew-borne Cao Bang virus and Amga virus. All six Gn/Gc complexes supported infection of primary human pulmonary microvascular endothelial cells, with virus-specific efficiencies and kinetics. Genetic ablation of protocadherin-1 (PCDH1) reduced infection mediated by Laguna Negra virus and Black Creek Canal virus Gn/Gc in human osteosarcoma and endothelial cells, whereas PCDH1 re-expression restored susceptibility. Both viruses were captured by a soluble PCDH1 protein comprising the first two extracellular cadherin domains. Infection was also inhibited by a soluble decoy containing the first cadherin domain, and by an antibody specific to that domain, indicating that the first PCDH1 cadherin domain mediates viral engagement. Black Creek Canal Gn/Gc-bearing rVSVs incorporated less Gn/Gc and showed lower apparent capture by PCDH1 than Laguna Negra virus and Andes virus Gn/Gc-bearing particles, even after normalization for glycoprotein abundance. In contrast, PCDH1 depletion did not detectably alter infection mediated by the other four hantavirus Gn/Gc complexes. Thus, Laguna Negra virus and Black Creek Canal virus use PCDH1 for efficient entry but differ in apparent receptor engagement and responses to PCDH1-directed inhibitors, whereas all six viruses are potently neutralized by the broadly reactive Gn/Gc-directed antibody ADI-42898.

While many cellular and viral mechanisms are associated with proviral latency, the positive transcription factor b (P-TEFb), a cellular co-factor of the nuclear factor kappa B and viral transactiator proteins, plays a critical role in low viral transcription levels in resting infected cells. P-TEFb, comprised of CDK9 and Cyclin T1 (CycT1), is absent in quiescent CD4+T cells that represent the bulk of the viral latent reservoir. In these cells, the CycT1 subunit is dephosphorylated and dissociated from CDK9, leading to its rapid proteasomal degradation. Based on the knowledge of CycT1 phosphorylation and proteolysis, we designed a mutant CycT1 protein [constitutively active (CA)-CycT1] that is resistant to protein degradation and is stably expressed in resting CD4+ T cells. Interestingly, the expression of CA-CycT1 significantly delayed the establishment of viral latency in CD4+ T cells. This is the first example of expressing a high level of CycT1 proteins in resting CD4+ T cells, which modulates viral latency and T cell functions. This finding presents a significant therapeutic potential for the development of effective therapies for latent viral infections.

Inducible expression of type I IFNs is a well-established host innate defense mechanism to limit virus infection. Yet, many viruses have evolved strategies to suppress the induction of these cytokines to enhance replication, spread, and transmission between hosts. Whether additional antiviral mechanisms protect against infection when inducible responses are compromised is not well understood. Here, we demonstrate that human keratinocytes preemptively protect the skin against virus infection through constitutive production of IFN-kappa (IFN-{kappa}), a poorly studied type I IFN family member. We find that constitutive IFNK expression protects keratinocytes against skin tropic RNA and DNA viruses, including vesicular stomatitis virus (VSV) and herpes simplex virus-1 (HSV-1). Using a human skin organoid model, we further demonstrate that keratinocyte-derived IFN-{kappa} establishes an antiviral state in dermal fibroblasts. Genetic and chemical analysis of the type I IFN receptor (IFNAR) signaling pathway in monocultured keratinocytes and skin organoids revealed that distinct pathways control VSV and HSV-1 replication. While canonical JAK-STAT signaling provided protection against VSV, HSV-1 infection was controlled through a JAK-STAT-independent mechanism. Using transcriptomic analysis, we further identified an IFN-{kappa}-dependent gene signature in keratinocytes that is not similarly driven by other type I IFNs. Together, this work establishes constitutive IFN-{kappa} production by keratinocytes as a broadly antiviral tissue autonomous defense mechanism.

Mosquito-transmitted alphaviruses, including chikungunya (CHIKV), Mayaro, Eastern equine encephalitis, and Venezuelan equine encephalitis viruses, cause severe arthritic or neurologic disease in humans. There are currently no approved antiviral treatments against alphavirus infection. Here, we report that the purine ribonucleoside EIDD-3525 is a potent inhibitor of alphavirus replication in vitro with broad spectrum antiviral activity against the Alphavirus genus. Using a trans-replicase assay, we find that EIDD-3525 blocks alphavirus RNA replication. Mutational profiling suggests that EIDD-3525 targets the alphavirus nsP4 RNA-dependent RNA polymerase that is conserved across medically important alphaviruses. In a pre-clinical mouse model of CHIKV infection and disease, oral treatment with EIDD-3525 reduced viral tissue burdens and the severity of virus-induced disease. Collectively, our findings support the continued development of the purine ribonucleoside EIDD-3525 as an orally available antiviral therapeutic for use against a broad group of alphaviruses.

Defective interfering particles (DIPs) profoundly influence viral replication and host responses yet it remains difficult to quantify their activity routinely in the laboratory. Here, using lymphocytic choriomeningitis virus (LCMV) as a model mammarenavirus, we establish complementary approaches for measuring interfering activity of DIPs across a range of experimental scales. First, we optimized plaque- and focus-based interference assays that directly quantify DIP-mediated suppression of standard infectious virus replication through a modified neutralization assay. Second, in cell monolayers saturated with both DIPs and infectious particles, we identified a complete loss of expression of the viral glycoprotein and matrix protein, which are each encoded in the positive-sense orientation of the mammarenavirus genome. Notably, this loss of protein expression mirrors the inhibition of plaque formation traditionally seen in this high DIP setting. The observed DIP-mediated loss of glycoprotein expression is a surrogate marker of interference and was used to develop an alternative assay of DIP activity. This glycoprotein expression assay recapitulated interference detected in traditional high-titer plaque assays (i.e. loss of plaques when all cells are saturated with DIPs and infectious particles) while substantially increasing scalability. These approaches enable sensitive and scalable quantification of DIP activity for LCMV as well as Junin virus and expand our experimental toolkit for studying arenavirus population dynamics, pathogenesis and immunity, and may help advance DIP-based antiviral strategies. This study also provides mechanistic details on how mammarenavirus DIPs interfere with infectious virus propagation.

Severe fever with thrombocytopenia syndrome (SFTS) is an emerging, tick-borne zoonosis caused by severe fever with thrombocytopenia syndrome virus (SFTSV) and is associated with high case-fatality rates in humans and cats. No vaccines or antiviral drugs are currently available for the prevention or treatment of feline SFTS. Because cats with SFTS often develop blood viral RNA loads exceeding those in severe human cases, we hypothesized that some naturally recovered cats develop strong neutralizing antibody responses and may serve as a potential source of antibody-rich serum for passive immunotherapy. We compared 80% plaque-reduction neutralization test (PRNT80) titers in convalescent sera from six human SFTS survivors and 14 cats that had recovered from naturally acquired SFTS, then evaluated the therapeutic efficacy of convalescent cat serum in SFTSV-infected, interferon receptor-deficient AG129 mice. Some SFTS-recovered cats showed high PRNT80 titers. Passive administration of high-titer convalescent cat serum early after infection protected SFTSV-infected mice. A single dose given at 2 or 3 days post-infection completely protected mice, whereas treatment at 4 days post-infection resulted in 40% survival. Administration on two consecutive days was associated with lower late-phase viremia in surviving mice. Sera from three independent SFTS-recovered cats improved survival, with greater effects observed for sera with higher neutralizing antibody titers. These findings indicate that some cats recovering from SFTS can develop high neutralizing antibody titers and their sera can exert therapeutic effects in an SFTSV-infected mouse model. These findings support further evaluation of passive immunotherapy as a treatment strategy for feline SFTS.

Long non-coding RNAs (lncRNAs) are a class of regulatory RNAs with lengths exceeding 200 nucleotides. The lncRNA PURPL (p53 upregulated regulator of p53 levels) is a p53-regulated gene. PURPL is an oncogenic lncRNA that represses p53 function; it is upregulated in many tumors and promotes cancer cell proliferation. Kaposi sarcoma-associated herpesvirus (KSHV) is the causative agent of all forms of Kaposi sarcoma (KS) and is tightly associated with primary effusion lymphoma (PEL). We found upregulation of PURPL in both de novo and chronically KSHV-infected cells. We found that p73 (a p53 homolog) associates with the PURPL promoter and upregulates PURPL expression. Previously, it has been shown that KSHV-encoded latency-associated nuclear antigen (LANA) stabilizes p73. We found that ectopic expression of LANA increased both p73 protein levels and PURPL expression. p73 knockdown abrogates PURPL upregulation following KSHV infection. Antisense oligonucleotides (ASO) targeting PURPL reduced PEL cell proliferation and increased acetylated p53 levels. In contrast to p53, p73 is not repressed by PURPL RNA. In summary, our study reveals how an oncogenic virus can utilize p73 to upregulate the oncogenic lncRNA PURPL as an additional way to repress p53. Moreover, we propose another mechanism by which p73 may promote tumor growth.

Herpes simplex virus 1 (HSV-1) establishes life-long latency in peripheral neurons following neuroinvasion, yet the early determinants governing latency versus productive infection remain poorly understood. Available in vitro models bypass the physiological route of infection by directly infecting neuronal cell bodies under antiviral suppression. Here, we established a compartmented axonal infection model that recapitulates HSV-1 neuroinvasion and latency establishment without pharmacological inhibitors. Using primary superior cervical ganglion (SCG) neurons, we demonstrated that low-dose axonal infection consistently establishes latency, whereas higher axonal doses or direct somatic infection led to productive replication. Latently infected neurons exhibit accumulation of the latency-associated transcript (LAT) and can be efficiently reactivated by UV-inactivated virus or VP16 expression. Remarkably, productive infection can be induced during low-dose axonal infection by simultaneous exposure of neuronal cell bodies, but not axons, to replication-deficient virions, implicating tegument delivery in soma overrides latency establishment. Conversely, excess replication-incompetent particles in axons suppress productive outcomes and subsequent reactivations, suggesting competition for retrograde transport. Together, these findings identify axonal entry and transport as critical regulatory checkpoints in HSV-1 latency establishment and suggest that interference with retrograde trafficking may represent a strategy to limit neuroinvasion and life-long reactivations.

The encapsulation of the viral RNA condensate inside the capsid is a crucial step in the late stages of the HIV-1 replication cycle. While it is agreed that RNA encapsulation depends on RNA-bound integrase (IN), that capsid-bound IN-RNA filaments are present within the capsid, and that the ribonucleoprotein (RNP) complex takes up only a fraction of the capsid volume, the timing of filament formation and ribonucleoprotein (RNP) encapsulation are unclear. Here, we ran virion-scale, coarse grained molecular dynamics simulations to identify how the organization of the RNP and IN in particular affects mature capsid assembly and genome encapsulation. These virion-scale simulations allow for direct comparison between experimental capsid morphologies and those self-assembled in the simulations. It was found, when relatively few tetramers are on the RNP surface, that IN tetramers could facilitate RNA encapsulation without filling a significant fraction of the capsid volume, but do not form a capsid-templated filament. When octameric IN-RNA filaments are modeled, it was found that WT-like capsid morphologies form only when the IN is organized as a single long filament, in agreement with recent experimental suggestions. These capsids are more conical and elongated than the capsids assembled around the RNP with tetrameric IN. Thus, we suggest the IN-RNA filament assembles ahead of the capsid and directs it to its final conical morphology, consistent with the experimental observation of less elongated capsids when IN oligomerization is disrupted.