Discovery of a non-nucleoside inhibitor that binds to a novel site in the palm domain of the respiratory syncytial virus RNA-dependent RNA polymerase

Abstract

Respiratory syncytial virus (RSV) is a major cause of severe respiratory tract infections in infants, young children, and the elderly. We report herein the discovery and characterization of a novel RSV polymerase (RSVpol) non-nucleoside inhibitor (NNI) chemotype that binds to a previously undescribed, highly conserved site in the palm domain of the L protein. Consistent with the observed mode of inhibition, cryogenic electron microscopy (cryo-EM) revealed the site to be adjacent to the nucleotide binding site. Minireplicon assays confirmed on-target activity against RSVpol, and cell-based antiviral assays showed that the lead compound effectively inhibited viral mRNA transcription and replication in clinically relevant A and B strains. Together, our data provides valuable insights into the molecular basis of inhibition for a novel mechanism of action and paves the way for structure-based design to deliver effective therapeutics against RSV.IMPORTANCERespiratory syncytial virus (RSV) is a negative-sense, single-stranded RNA virus belonging to the family Pneumoviridae of the order Mononegavirales. Currently, monoclonal antibody treatments are only approved for infants, and vaccines are reserved for pregnant women and adults aged 60 years and older. Prophylaxis is also limited to the pediatric patient population, and there are currently no direct antiviral therapies for post-exposure treatment. Viral polymerases are considered well-validated drug targets due to their critical role in transcription and genome replication. Herein, we disclose the discovery of a spiro-indolinone series as polymerase inhibitors and describe the preliminary structure-activity relationship (SAR). A cryogenic electron microscopy (cryo-EM) structure obtained with an optimized lead revealed a novel binding site located in the palm domain, which will enable future structure-based drug design efforts. Novel RSV antivirals could be beneficial both as therapeutics following diagnosis and as a prophylactic in patients less likely to respond to vaccines.

Keywords: NNI; RSV polymerase inhibitor; cryogenic electron microscopy; high-throughput screen; palm domain; respiratory syncytial virus.

Fig 1

Optimized high-throughput screening hit inhibits RSV RdRp and exhibits cell-based antiviral activity. (A) Domain architecture of RSV L and P protein subunits. Stars indicate known inhibitor binding sites (capping domain: JNJ-8003, MRK-1, S1; connector domain: JNJ-7184, AZ-27, AVG), colored domains indicate structured regions observed by cryo-EM (RdRp: polymerase, PRNTase: capping, CD: connector, MTase: methyltransferase, CTD: C-terminal domain, NTD: N-terminal domain, OD: oligomerization domain). (B) Structure of hit and optimized cell-active lead. (C) Inhibition of RSV L+P screening construct and coupled enzyme system by 1. RSV L+P_Δ1-124 IC₅₀ = 0.61 µM, PP_i counterscreen IC₅₀ > 100 µM. RSV L+P_Δ1-124 assay: 20 nM RSV L+P_Δ1-124, 10 µM ATP/GTP/CTP, 100 µM ATPαS, 5 µM TrC-25. Counterscreen: 1.5 µM PP_i, 1 U/µL ATP sulfurylase, 2 nM luciferase, 5 µM adenosine phosphosulfate (APS), 300 µM luciferin, 1 mM coenzyme A. (D) Biochemical activity of 1 and 22 against full-length RSV L+P and human mitochondrial RNA polymerase (POLRMT). RSV L+P IC₅₀ = 0.67 µM (1), 0.27 µM (22), POLRMT IC₅₀ > 100 µM (1, 22). POLRMT counterscreen: 8 nM POLRMT, 4.5 µM ATP, 0.85 µM GTP, 0.1 µM CTP, 0.60 µM UTP, 0.1 µM RNA primer/ssDNA-FAM template/ssDNA-BHQ quencher. (E) Inhibition of RSV A2 replication in HeLa cell GFP reporter assay and APC-126 replicon assay by 22. RSV eGFP EC₅₀ = 2.28 µM, ATPLite CC₅₀ > 50 µM, Replicon EC₅₀ = 2.22 µM, Replicon CC₅₀ > 50 µM. HeLa assays: 3,000 cells/well, rgRSV224 viral multiplicity of infection (MOI) = 1, 37°C, 5% CO₂, 72 h. APC-126: 3,500 cells/well, 37°C, 5% CO₂, 72 h. (F) Surface plasmon resonance (SPR) sensorgram for RSV-L+P binding to 22. RSV L+P immobilized at 30 µg/mL, 1,200 s. 22 injected for 60 s, 30 µL min⁻¹ flow rate, 500 s dissociation. k_a = 1.28 × 10⁵ M⁻¹ s⁻¹, k_d = 6.09 × 10⁻² s⁻¹, K_D = 0.47 µM. All data are representative of at least two independent experiments and plotted as mean ± SD.

Fig 2

Inhibition of de novo initiation and primer extension of RSV L+P RNA synthesis by compound 22. (A) Reaction scheme of de novo (starting from +3 site on TrC-14 template) initiation assay. (B) Representative sequencing gel image showing RNA products synthesized by RSV L+P at various concentrations of 22 (0.1 nM to 200 µM). The reactions contained 10 nM RSV L+P, 2 µM RNA template (TrC-14), 500 µM GTP, 10 µM ATP, and 170 nM [α-³³P]ATP. RNA synthesis started from the +3 site and paused with a major 11-mer product. Negative control (Neg.) was the reaction without the enzyme. (C) Quantification and analysis of the inhibition of pppGA and the 11-mer RNA products. The data were fit to the Hill equation, and the inhibition of de novo pppGA formation exhibited an IC₅₀ of 0.29 µM with a maximal inhibition of 85%, whereas inhibition of the 11-mer product exhibited an IC₅₀ of 0.48 µM with a maximal inhibition of 71%. Data are representative of three technical replicates and plotted as mean ± SD. (D) Polyacrylamide sequencing gel showing single-nucleotide incorporation from a set of short primers with or without 50 µM 22 with TrC-14 template. (E) Inhibition of primer extension by 22 to primers with various lengths. Data were analyzed from the gel in D. Bars indicate the mean, and data points from two independent experiments are shown.

Fig 3

Compound 22 is an allosteric inhibitor of RSVpol. (A and B) The mode of inhibition for 22 was (A) mixed noncompetitive/uncompetitive (α = 0.33 ± 0.27) with respect to UTP and (B) noncompetitive (α = 1.1 ± 0.6) with respect to the TrC-25 RNA template. (C and D) SPR K_D determined for 22 in the presence of (C) capping domain binder JNJ-8003 (compound 22 k_a = 1.19 × 10⁵ M⁻¹ s⁻¹, k_d = 6.28 × 10⁻² s⁻¹, K_D = 0.53 µM) and (D) connector domain binder JNJ-7184 (compound 22 k_a = 1.40 × 10⁵ M⁻¹ s⁻¹, k_d = 7.25 × 10⁻² s⁻¹, K_D = 0.52 µM). (E and F) Yonetani-Theorell analysis showing that binding of 22 was not mutually exclusive with (E) compound 23 (γ = 1.1 ± 0.2) and (F) compound 24 (γ = 0.88 ± 0.72). (G) Chemical structures of capping 23 and connector 24 domain ligands.

Fig 4

Structural basis of the interactions between RSVpol and compound 22 (PDB 9N36, EMD 48846). (A) Visualization of the Cryo-EM structure illustrating the RSVpol complex with compound 22. The RdRp domain, capping domain, palm domain, four partial P proteins, and compound 22 are color-coded in gray, cyan, red, light green, pink, light orange, green, and magenta, respectively. (B) Depiction of the cryo-EM density corresponding to compound 22. The ligand density is shown at a contour level of 0.339, as displayed in Chimera. (C) Presentation of the binding environment for compound 22. (D) Binding pocket for compound 22, with the compound and residues within 4 Å that form interactions shown in stick representation. (E) The interaction between compound 22 and the GDN motif. The GDN motif is shown as green sticks. (F) Two-dimensional interaction patterns of compound 22 and RSV L were highlighted using dashed lines. (G) Structural comparison of the ligand binding pocket between compound 22-bound and Apo RSV L, featuring rotamer transitions for Lys885, Lys871, Phe868, Asn812, and Glu336, indicated by arrows.

Fig 5

Detailed insights into the structural features and interactions within the compound 22 binding domain of RSVpol. (A) Compound 22 within the RSVpol domain is surrounded by alpha-helical and beta-sheet bundles. The alpha helix and beta strand, significantly stabilized by compound 22, are highlighted in orange and blue, respectively. (B) Raw density visualization of the newly stabilized alpha helix. The new helix density is shown at a contour level of 0.245, as displayed in Chimera. (C) Comparison of compound 22 bound and the Apo form (PDB: 6PZK). In the compound 22 bound form, the palm domain and new helix are depicted in red and orange, respectively, while the Apo form is displayed in cyan. Compound 22 is shown as magenta. (D) Depiction of key interactions surrounding the helix that contribute to its stabilization. The color code in panels C and D in compound 22 bound form corresponds to that in Fig. 4A.

Fig 6

Chemical synthesis of 1 and its derivatives. (A) The three vectors being explored: southern R¹, left-hand-side R², and right-hand-side R³. (B) Chemical syntheses of 1 and 18–22. Conditions: (a) trimethylboroxine, Pd(dppf)Cl₂, K₂CO₃, 1,4-dioxane, water, 100°C; (b) alkyliodide, NaH, DMF, rt; (c) HCl, 1,4-dioxane, 100°C; (d) ArX, DIPEA, NMP, 80°C; (e) 3-bromopyridine, DMEDA, CuI, Cs₂CO₃, toluene, DMF, 100°C. (C) Chemical syntheses of 23–25. Conditions: (f) DIAD, cyclopropane methanol, P(t-Bu)₃, THF, 0°C to rt. (D) Chemical syntheses of 29–32 and 34. Conditions: (g) MeI, Cs₂CO₃, DMF, 0°C; (h) ACE-Cl, DIPEA, CH₃CN, toluene, 60°C; (i) ArX, NaOt-Bu, RuPhos Pd G4, 1,4-dioxane; (j) 4-bromo-2-fluoro-6-methylbenzonitrile, NaOTMS, RuPhos Pd G4, 1,4-dioxane, 80°C; (k) morpholine, RuPhos Pd G4, NaOt-Bu, 1,4-dioxane, 65°C. X, appropriate halide; DMF, dimethylformamide DIPEA, N,N-diisopropylethylamine; NMP, N-methylpyrrolidone; DMEDA, N,N-dimethylethylenediamine; DIAD, diisopropyl azodicarboxylate; ACE-Cl, 1-chloroethyl chloroformate; RuPhos Pd G4, [dicyclohexyl(2′,6′-diisopropoxy-2-biphenylyl)phosphine-κP](methanesulfonatato-κO)[2′-(methylamino-κN)-2-biphenylyl-κC²]palladium.