Identification of small molecule inhibitors for influenza a virus using in silico and in vitro approaches

Abstract

Influenza viruses have acquired resistance to approved neuraminidase-targeting drugs, increasing the need for new drug targets for the development of novel anti-influenza drugs. Nucleoprotein (NP) is an attractive target since it has an indispensable role in virus replication and its amino acid sequence is well conserved. In this study, we aimed to identify new inhibitors of the NP using a structure-based drug discovery algorithm, named Nagasaki University Docking Engine (NUDE), which has been established especially for the Destination for GPU Intensive Machine (DEGIMA) supercomputer. The hit compounds that showed high binding scores during in silico screening were subsequently evaluated for anti-influenza virus effects using a cell-based assay. A 4-hydroxyquinolinone compound, designated as NUD-1, was found to inhibit the replication of influenza virus in cultured cells. Analysis of binding between NUD-1 and NP using surface plasmon resonance assay and fragment molecular orbital calculations confirmed that NUD-1 binds to NP and could interfere with NP-NP interactions essential for virus replication. Time-of-addition experiments showed that the compound inhibited the mid-stage of infection, corresponding to assembly of the NP and other viral proteins. Moreover, NUD-1 was also effective against various types of influenza A viruses including a clinical isolate of A(H1N1)pdm09 influenza with a 50% inhibitory concentration range of 1.8-2.1 μM. Our data demonstrate that the combined use of NUDE system followed by the cell-based assay is useful to obtain lead compounds for the development of novel anti-influenza drugs.

Fig 1. Tail-binding pocket of NP, in silico screening and antiviral activity of 4-hydroxyquinolinone hit compounds.

(A) Crystal structure of the NP monomer (PDB code: 2IQH) represented as a surface. The tail-binding pocket of NP is highlighted in green. (B) Cartoon representation of a closer view of the tail-binding pocket of NP. Target amino acid residues S165, E339, D340 and A387 are represented as sticks. The images for A and B were generated using PyMOL software. (C) In silico screening was performed to identify compounds that bind to the target amino acid residues. The binding scores of 9,430 compounds were plotted as a function of the in silico ranking. Open diamonds indicate 4-hydroxyquinolinone compounds, NUDs 1–5, that were found to suppress virus replication in the cell-based assay (see part D). (D) Structure of NUDs 1–5 and the IC₅₀ (mean ± SD) are presented. IC₅₀ was obtained from three independent experiments.

Fig 2. Chemical structure and anti-influenza activity of NUDs 6–24.

Commercially available analogs of NUDs 1–5 were acquired and evaluated for anti-influenza activity. The results are the average ± SD from two independent experiments.

Fig 3. SPR and FMO analyses of NUD-1 and NP binding.

Molecular interactions of NUD-1 and NP were analyzed by SPR (A–D) and FMO calculations (E). (A) Sensorgram of naproxen, (B) affinity curve of naproxen, (C) sensorgram of NUD-1, (D) affinity curve of NUD-1, (E) interaction energies of NUD-1 with NP tail-binding pocket amino acids calculated by the FMO method. The energies obtained by the HF method (blue) mainly includes electrostatic and charge-transfer interactions, and the energies obtained by the MP2 method (red) additionally includes dispersion interactions.

Fig 4. Dose-dependent inhibitory activity of NUD-1 against influenza virus.

MDCK cells were infected with A/WSN/33 virus in the presence of varying concentrations of NUD-1 or oseltamivir. At 48 h post-infection, the culture supernatant was harvested for virus yield titration using hemagglutination assay. Cells viability was then evaluated by WST-1 assay and crystal violet staining. HA titer/mL (closed triangles) and relative cell viability determined by WST-1 assay (closed circles) and crystal violet assay (open circles) were plotted against the concentrations of (A) NUD-1, (B) oseltamivir.

Fig 5. Mode of action of NUD-1 against influenza virus replication.

MDCK cells were infected with A/WSN/33 virus (MOI = 0.001) and the culture supernatant was collected at 12 h post-infection for TCID₅₀ assay. (A) Different cell and/or virus treatment protocols. (B) NUD-1 (2.5 μM) was added to cells before, during or after virus infection; or virus was incubated with NUD-1 before infection. C) Infected cells were exposed to NUD-1 (2.5 μM) or zanamivir (10 μM) at different time points after virus infection (0–3 h, 3–6 h, 6–9 h, 9–12 h). Virus yield for each treatment condition is represented as a percentage of the untreated control. The results are the mean ± SD obtained from at least two independent experiments. Student’s t-test was performed using Graphpad software. * indicates a p-value of less than 0.05.

Fig 6. Multiple alignment of IAV and IBV NP sequences.

Sequences of A/WSN/33 (AAA43452.1), A/Puerto Rico/8/34 (NP_040982.1), A/Virginia/04/2009 (ACR08603.1), A/Aichi/2/68 (AFM71861.1) and B/Lee/40 (NP_056661.1) were obtained from the NCBI protein database and aligned using Clustal Omega software. The NP sequence for strain A/Virginia/ATCC2/2009 was not available, so strain A/Virginia/04/2009 was chosen as a representative sequence for A(H1N1)pdm09 influenza. Asterisks indicate conserved amino acid residues between IAV and IBV, while conserved residues among IAV strains are highlighted in grey. Amino acid residues important for NP-NP interactions in IAV and IBV are indicated in red and purple font, respectively. Tail loop regions are underlined.