Inhibition of CRM1-mediated nuclear export of influenza A nucleoprotein and nuclear export protein as a novel target for antiviral drug development

Abstract

An anti-influenza compound, DP2392-E10 based on inhibition of the nuclear export function of the viral nucleoprotein-nuclear export signal 3 (NP-NES3) domain was successfully identified by our previous high-throughput screening system. Here, we demonstrated that DP2392-E10 exerts its antiviral effect by inhibiting replication of a broad range of influenza A subtypes. In regard to the molecular mechanism, we revealed that DP2392-E10 inhibits nuclear export of both viral NP and nuclear export protein (NEP). More specifically, in vitro pull-down assays revealed that DP2392-E10 directly binds cellular CRM1, which mediates nuclear export of NP and NEP. In silico docking suggested that DP2392-E10 binds at a region close to the HEAT9 and HEAT10 domains of CRM1. Together, these results indicate that the CRM1-mediated nuclear export function of influenza virus represents a new potential target for antiviral drug development, and also provide a core structure for a novel class of inhibitors that target this function.

Keywords: Chromosome region maintenance 1; Influenza A virus; Nuclear export protein; Nuclear export signal; Nucleoprotein.

Fig. 1

DP2392-E10 inhibits replication of a broad range of influenza A subtypes. (A) Chemical structure of DP2392-E10. (B) MDCK cells were infected with different strain of influenza A viruses and subjected to plaque assays in the presence of DMSO or DP2392-E10 at 10, 30, and 50 µM. After 48 h, the cells were stained with crystal violet. Plaques were counted, and the numbers were used to calculate the relative plaque forming units (PFU)/ml (%DMSO). Data represents means ±SD of virus titer from three independent experiments. (C) The half-maximal inhibitory concentration (IC₅₀), half-maximal cytotoxic concentration (CC₅₀), and selectivity index (SI = CC₅₀/IC₅₀) of DP2392-E10 calculated from the plaque assay in B and the Water Soluble Tetrazolium salt (WST-1) assay. (D) MDCK cells were infected with influenza A/WSN/1933 virus at a multiplicity of infection (MOI) of 0.0001 and treated with DMSO or DP2392-E10 at 10, 30, and 50 µM. Supernatant at 12, 24, and 36 h post infection (hpi) were collected for virus titration by plaque assay. Data represents means ±SD of virus titer from three independent experiments.

Fig. 2

DP2392-E10 inhibits viral replication by preventing nuclear export of viral NP and NEP. MDCK cells were infected with influenza A/WSN/1933 at MOI of 2 and treated with DMSO or 30 µM of DP2392-E10. Intracellular localization of NP and NEP were observed at 2, 4, and 6 hpi by immunofluorescence staining with anti-NP/Alexa Fluor 594 and anti-NEP/Alexa Fluor 488 antibodies. Nuclei were counterstained with Hoechst 33342. Data are representative of two independent experiments.

Fig. 3

DP2392-E10 inhibits CRM1-mediated nuclear exports of NP-NES3 and NEP-NES2. (A) Schematic representation of Aequorea coerulescens green fluorescent protein (AcGFP)-NP-NES3, AcGFP-NEP-NES2, and AcGFP-Rev-NES constructs. (B-D) MDCK cells stably expressing AcGFP-NP-NES3, AcGFP-NEP-NES2, or AcGFP-Rev-NES were treated with DMSO, 10 nM LMB (positive control), or 2.5, 5.0, 7.5, 10.0, 12.5, or 15.0 µM DP2392-E10 for 8 h. The cells were then fixed with 4% paraformaldehyde, nuclei were stained with Hoechst 33342, and the mean fluorescent intensity of nuclear AcGFP was quantified on a CELAVIEW RS100. Data represents mean ±SD of nuclear AcGFP-NP-NES3 (B, bottom panel), AcGFP-NEP-NES2 (C, bottom panel), or AcGFP-Rev-NES (D, bottom panel) and the half-maximal effective concentration (EC₅₀) values from two independent experiments. Each of triplicate samples in one experiment was visualized for 36 fields of observation.

Fig. 4

DP2392-E10 inhibits NP/CRM1 binding by directly targeting CRM1. (A) FLAG– or NP-FLAG–immobilized agarose beads were incubated with purified CRM1-HA protein in the presence of DMSO or 3, 10, 30, or 100 µM of DP2392-E10 overnight, pulled down, washed, and subjected to western blot analysis with anti-HA and anti-FLAG antibodies. Intensity of the pulled-down CRM1 and NP were quantified using the ImageJ software. Data are representative of two independent experiments. (B) Photo-crosslinked Sepharose beads without DP2392-E10 (control) or with DP2392-E10 were co-incubated with purified NP-FLAG or FLAG-CRM1 proteins, pulled down, washed, and subjected to western blot analysis with anti-FLAG antibody. Data are representative of two independent experiments.

Fig. 5

In silico analysis of the interaction between DP2392-E10 and unliganded CRM1. Models of unliganded human and canine CRM1 proteins were constructed by homology modeling and subjected to docking simulation studies in the Molecular Operating Environment (MOE). (A) Structural model of human CRM1. The model consists of twenty-one HEAT repeats (H1–H21), each composed of two α-helices linked by a short loop (Fung and Chook, 2014, Monecke et al., 2013). HEAT repeats for a switch from the inactive to active conformation of CRM1 are highlighted by colors: H9 (magenta), H9 acid loop (orange), H11 (yellow), and H12 (cyan). (B and C) Left panels: 3D distribution of potential binding pockets on human (B) or canine (C) CRM1. The pockets were detected around all the HEAT repeats, except the H21 region. The pockets for the binding of small molecules are shown as grey spheres. Right panels: Distribution of docking scores of U_dock at individual pockets. In each pocket, several U_dock values representing distinct binding conformations were found. Only U_dock scores giving negative values are shown. A red bar at a given site indicates the average of all U_dock scores at the site.