Filovirus-host interactions are important for efficient egress of virus particles; however, mechanistic details of the formation, dynamics, and trafficking of these virus-host complexes in the natural environment of the host cell have been elusive. In this report, we used a BiMC approach to visualize eVP40-Tsg101 complexes as they formed in the cell. The specificity of this interaction was confirmed by using an L-domain deletion mutant of eVP40 and by using Tsg101 specific siRNAs. Importantly, the NYFP-Tsg101 fusion protein was stably expressed in mammalian cells, and CYFP-VP40 fusion proteins retained their ability to bud independently from cells as VLPs in an L-domain-dependent manner. The BiMC approach is ideal for detecting weak and/or transient protein-protein interactions in living cells [27
]. We demonstrated that eVP40-Tsg101 complexes formed between 3–4 hours after transfection and colocalized at early times (6
hrs. p.t.) with pericentrin-B, an MTOC marker. We postulate that the initial eVP40-Tsg101 complexes may then migrate from the MTOC to the site of budding at the plasma membrane by 12–24 hours p.t. This working model correlates with previous reports which suggested that filovirus VP40 proteins may interact with and utilize the host cytoskeletal network during assembly and egress [43
Elucidation of the molecular complexities and dynamics of virus-host interactions in the natural cell environment will enhance our ability to effectively screen and validate new antivirals [29
]. Recent studies have identified two compounds, FGI-104 and FGI-106, that showed activity against filoviruses in cell culture and in animals [45
]; however, the targets of FGI-104 and FGI-106 remain to be determined. In addition, small molecule inhibitors of filovirus entry have recently been identified and characterized [47
]. Viral L-domain/host interactions remain an attractive target for the development of novel, broad-spectrum budding inhibitors [33
]. The successful development of the BiMC assay to assess filovirus-host interactions (this manuscript) and the use of our well-established VLP budding assay represent powerful tools that will allow us to screen and validate small molecule inhibitors of filovirus budding. By using the known 3D atomic structure of Tsg101 binding to the PTAP motif, we employed an in silico
strategy to identify and rank commercially available compounds with predicted drug-like properties that could potentially block this interaction. From this screen, we identified compound 5539-0062 and used both BiMC and VLP budding assays to demonstrate that this small molecule specifically inhibited PTAP-dependent budding of eVP40 VLPs, but not PTAP-independent budding of mVP40 VLPs. Moreover, compound 5539-0062 also exhibited specific antiviral activity in cells infected with a recombinant VSV (VSV-M40) expressing the PTAP L-domain of eVP40. The reason for the low level of inhibition of PPxY-dependent budding of VSV-WT observed in the presence of compound 5539-0062 remains to be determined (), as does the potential effect of compound 5539-0062 on other stages of VSV replication. The PTAP-specific inhibitory activity exhibited by this single, first-generation compound is encouraging and serves as proof-of-principle for using this strategy and the BiMC assay to identify and validate budding inhibitors. Compound 5539-0062 could potentially be optimized to improve binding affinity by either searching for chemically similar molecules in the commercially available databases, or by using the structural interaction fingerprint method which involves selecting compounds which are predicted to have the same protein-ligand interactions as 5539-0062 [51
]. One advantage of L-domain inhibitors is their potential broad-spectrum activity against a wide-array of RNA viruses. Indeed, we were able to show that in addition to inhibiting budding of eVP40 VLPs, compound 5539-0062 was also able to block the PTAP-mediated interaction between Tsg101 and HIV-1 Gag, leading to inhibition of HIV-1 Gag budding.
Although the focus here was on inhibitors of PTAP-type L-domain activity, similar studies are underway to identify inhibitors of PPxY-type L-domains as well (Liu, Lee, Olson, and Harty, unpublished data). Viral PPxY-type L-domains are known to interact with host proteins such as Nedd4 E3 ubiquitin ligase [4
]. One could envision that the use of a cocktail of budding inhibitors containing both PTAP- and PPxY-specific compounds, for example, might be more effective at blocking virus egress than single L-domain inhibitors. Indeed, both EBOV and MARV appear to utilize both PTAP and PPxY L-domains for efficient egress [6
In addition to these virus-host interaction studies, we are investigating the contributions of viral VP40-VP40, VP40-NP, and VP40-VP35 interactions to filovirus assembly and egress [5
]. The broad applicability of the BiMC approach as well as a multicolor fluorescence strategy will enhance our understanding of the biological relevance of these complex multiprotein interactions to RNA virus egress. Extension of these studies to include live virus and/or animal models is essential and will help lead to new insights into filovirus pathogenesis and novel treatment options.