Viruses are obligate intracellular parasites and therefore their replication completely depends on host cell factors. In case of the hepatitis C virus (HCV), a positive-strand RNA virus that in the majority of infections establishes persistence, cyclophilins are considered to play an important role in RNA replication. Subsequent to the observation that cyclosporines, known to sequester cyclophilins by direct binding, profoundly block HCV replication in cultured human hepatoma cells, conflicting results were obtained as to the particular cyclophilin (Cyp) required for viral RNA replication and the underlying possible mode of action. By using a set of cell lines with stable knock-down of CypA or CypB, we demonstrate in the present work that replication of subgenomic HCV replicons of different genotypes is reduced by CypA depletion up to 1,000-fold whereas knock-down of CypB had no effect. Inhibition of replication was rescued by over-expression of wild type CypA, but not by a mutant lacking isomerase activity. Replication of JFH1-derived full length genomes was even more sensitive to CypA depletion as compared to subgenomic replicons and virus production was completely blocked. These results argue that CypA may target an additional viral factor outside of the minimal replicase contributing to RNA amplification and assembly, presumably nonstructural protein 2. By selecting for resistance against the cyclosporine analogue DEBIO-025 that targets CypA in a dose-dependent manner, we identified two mutations (V2440A and V2440L) close to the cleavage site between nonstructural protein 5A and the RNA-dependent RNA polymerase in nonstructural protein 5B that slow down cleavage kinetics at this site and reduce CypA dependence of viral replication. Further amino acid substitutions at the same cleavage site accelerating processing increase CypA dependence. Our results thus identify an unexpected correlation between HCV polyprotein processing and CypA dependence of HCV replication.

Influenza A virus (IAV) uses the low pH in late endocytic vacuoles as a cue for penetration by membrane fusion. Here, we analyzed the prefusion reactions that prepare the core for uncoating after it has been delivered to the cytosol. We found that this priming process occurs in two steps that are mediated by the envelope-embedded M2 ion channel. The first weakens the interactions between the matrix protein, M1, and the viral ribonucleoprotein bundle. It involves a conformational change in a linker sequence and the C-terminal domain of M1 after exposure to a pH below 6.5. The second step is triggered by a pH of <6.0 and by the influx of K ؉ ions. It causes additional changes in M1 as well as a loss of stability in the viral ribonucleoprotein bundle. Our results indicate that both the switch from Na ؉ to K ؉ in maturing endosomes and the decreasing pH are needed to prime IAV cores for efficient uncoating and infection of the host cell. IMPORTANCEThe entry of IAV involves several steps, including endocytosis and fusion at late endosomes. Entry also includes disassembly of the viral core, which is composed of the viral ribonucleoproteins and the RNA genome. We have found that the uncoating process of IAV is initiated long before the core is delivered into the cytosol. M2, an ion channel in the viral membrane, is activated when the virus passes through early endosomes. Here, we show that protons entering the virus through M2 cause a conformational change in the matrix protein, M1. This weakens interactions between M1 and the viral ribonucleoproteins. A second change was found to occur when the virus enters late endosomes. The preacidified core is then exposed to a high concentration of K ؉ , which affects the interactions between the ribonucleoproteins. Thus, when cores are finally delivered to the cytosol, they are already partially destabilized and, therefore, uncoating competent and infectious.

Most systemic viral gene therapies have been limited by sequestration and degradation of virions, innate and adaptive immunity, and silencing of therapeutic genes within the target cells. Here we engineer a high-affinity protein coat, shielding the most commonly used vector in clinical gene therapy, human adenovirus type 5. Using electron microscopy and crystallography we demonstrate a massive coverage of the virion surface through the hexon-shielding scFv fragment, trimerized to exploit the hexon symmetry and gain avidity. The shield reduces virion clearance in the liver. When the shielded particles are equipped with adaptor proteins, the virions deliver their payload genes into human cancer cells expressing HER2 or EGFR. The combination of shield and adapter also increases viral gene delivery to xenografted tumors in vivo, reduces liver off-targeting and immune neutralization. Our study highlights the power of protein engineering for viral vectors overcoming the challenges of local and systemic viral gene therapies.

Cullin-3 (Cul3) functions as a scaffolding protein in the Bric-a-brac, Tramtrack, Broad-complex (BTB)–Cul3–Rbx1 ubiquitin E3 ligase complex. Here, we report a previously undescribed role for Cul3 complexes in late endosome (LE) maturation. RNAi-mediated depletion of Cul3 results in a trafficking defect of two cargoes of the endolysosomal pathway, influenza A virus (IAV) and epidermal growth factor receptor (EGFR). IAV is able to reach an acidic endosomal compartment, coinciding with LE/lysosome (LY) markers. However, it remains trapped or the capsid is unable to uncoat after penetration into the cytosol. Similarly, activation and subsequent ubiquitination of EGFR appear normal, whereas downstream EGFR degradation is delayed and its ligand EGF accumulates in LE/LYs. Indeed, Cul3-depleted cells display severe morphological defects in LEs that could account for these trafficking defects; they accumulate acidic LE/LYs, and some cells become highly vacuolated, with enlarged Rab7-positive endosomes. Together, these results suggest a crucial role of Cul3 in regulating late steps in the endolysosomal trafficking pathway.

Actin and actin-binding proteins are incorporated into HIV-1 particles, and F-actin has been suggested to bind the NC domain in HIV-1 Gag. Furthermore, F-actin has been frequently observed in the vicinity of HIV-1 budding sites by cryo-electron tomography (cET). Filamentous structures emanating from viral buds and suggested to correspond to actin filaments have been observed by atomic force microscopy. To determine whether the NC domain of Gag is required for actin association with viral buds and for actin incorporation into HIV-1, we performed comparative analyses of virus-like particles (VLPs) obtained by expression of wild-type HIV-1 Gag or a Gag variant where the entire NC domain had been replaced by a dimerizing leucine zipper [Gag(LZ)]. The latter protein yielded efficient production of VLPs with near-wild-type assembly kinetics and size and exhibited a regular immature Gag lattice. Typical HIV-1 budding sites were detected by using cET in cells expressing either Gag or Gag(LZ), and no difference was observed regarding the association of buds with the F-actin network. Furthermore, actin was equally incorporated into wild-type HIV-1 and Gag-or Gag(LZ)-derived VLPs, with less actin per particle observed than had been reported previously. Incorporation appeared to correlate with the relative intracellular actin concentration, suggesting an uptake of cytosol rather than a specific recruitment of actin. Thus, the NC domain in HIV-1 Gag does not appear to have a role in actin recruitment or actin incorporation into HIV-1 particles. IMPORTANCEHIV-1 particles bud from the plasma membrane, which is lined by a network of actin filaments. Actin was found to interact with the nucleocapsid domain of the viral structural protein Gag and is incorporated in significant amounts into HIV-1 particles, suggesting that it may play an active role in virus release. Using electron microscopy techniques, we previously observed bundles of actin filaments near HIV-1 buds, often seemingly in contact with the Gag layer. Here, we show that this spatial association is observed independently of the proposed actin-binding domain of HIV-1. The absence of this domain also did not affect actin incorporation and had a minor effect on the viral assembly rate. Furthermore, actin was not enriched in the virus compared to the average levels in the respective producing cell. Our data argue against a specific recruitment of actin to HIV-1 budding sites by the nucleocapsid domain of Gag.