Direct-Acting Antivirals Targeting Flavivirus Entry and Fusion

Abstract

There is a need for effective direct-acting antivirals (DAAs) against flaviviruses such as dengue virus (DV). Targeting conserved sites on the DV envelope (E) glycoprotein involved in viral entry and fusion that are under structural or functional constraint can limit DAA resistance. DV E engages host cell receptors and attachment factors required for entry. Following endocytosis, DV E undergoes conserved, pH-dependent, large-scale conformational rearrangements required for membrane fusion. The fusion process transiently exposes cryptic, conserved regions on E that are potential therapeutic targets. While DAAs can readily target surface-accessible receptor/attachment factor sites on E, DV fusion proceeds within a protected endosomal compartment, presenting a significant barrier for DAAs targeting this critical step, as any DAA must be present in this same compartment as the fusing virion. This thesis describes the identification, development, and characterization of DAAs targeting conserved sites on E required for entry and fusion. In Chapter 2, we show that lectins, which bind conserved glycans on the DV E, can be engineered to have increased affinity and impart specificity for the DV E protein; these engineered lectins are potential DAAs that inhibit viral attachment. Briefly, we made random mutagenesis libraries of select lectins, followed by directed evolution using yeast and phage display platforms, to enrich lectin variants with increased affinity to DV E. We then used rational design to incorporate enriched mutations into homologous regions of the lectin sequences and identified a mutation that had ~10-fold improved affinity to DV E compared to wild type. Additionally, we designed lectin variants with ablated binding domains to control glycan engagement and create lectins with monovalent interactions. In Chapter 3, we describe the identification of DAAs that engage transient E intermediates exposed during the fusion-promoting cascade during DV entry. Briefly, we designed two hyperglycosylated E domain III (hgDIII) probes. These hgDIII probes have introduced glycans that mask i) a buried interface or ii) the surrounding area of DIII, allowing for their use as negative or positive selection probes, respectively. We recombinantly expressed these probes and confirmed their conformational integrity, presence of introduced glycans, and feasibility to identify DAAs that engage cryptic regions of DIII. We performed preliminary yeast display selections using a naïve fibronectin library to enrich for binders to DIII. In Chapter 4, we describe a platform to deliver DAAs to the protected endosomal compartment. Briefly, we designed a protein-based nanoparticle (NP) that binds DV virions prior to endocytosis and concomitantly delivers DAAs to the same compartment to inhibit viral fusion. These molecular “Trojan horses” display on their surface a targeting moiety (TM) that binds to DV E, as well as a payload that inhibits viral fusion. We show that these chimeric NPs self-assemble into nanocages and bind the prefusion form of E. We further show that the stoichiometry of displayed proteins can be tuned to maximize the payload display. We identified non-neutralizing antibodies to function as TMs and used a previously described payload as a proof-of-concept to assess antiviral activity of this platform. Collectively, the work described in this thesis identifies optimal targets for therapeutic intervention in DV entry while providing the conceptual framework for developing future DAAs that target conserved sites. More broadly, this work can be applied to other viruses that engage host cells through glycan-lectin interactions and/or fuse from internal compartments.