Conserved allosteric inhibitory site on the respiratory syncytial virus and human metapneumovirus RNA-dependent RNA polymerases

Abstract

Respiratory syncytial virus (RSV) and human metapneumovirus (HMPV) are related RNA viruses responsible for severe respiratory infections and resulting disease in infants, elderly, and immunocompromised adults^1-3. Therapeutic small molecule inhibitors that bind to the RSV polymerase and inhibit viral replication are being developed, but their binding sites and molecular mechanisms of action remain largely unknown⁴. Here we report a conserved allosteric inhibitory site identified on the L polymerase proteins of RSV and HMPV that can be targeted by a dual-specificity, non-nucleoside inhibitor, termed MRK-1. Cryo-EM structures of the inhibitor in complexes with truncated RSV and full-length HMPV polymerase proteins provide a structural understanding of how MRK-1 is active against both viruses. Functional analyses indicate that MRK-1 inhibits conformational changes necessary for the polymerase to engage in RNA synthesis initiation and to transition into an elongation mode. Competition studies reveal that the MRK-1 binding pocket is distinct from that of a capping inhibitor with an overlapping resistance profile, suggesting that the polymerase conformation bound by MRK-1 may be distinct from that involved in mRNA capping. These findings should facilitate optimization of dual RSV and HMPV replication inhibitors and provide insights into the molecular mechanisms underlying their polymerase activities.

Fig. 1. Characterization of purified L proteins.

a Chemical structure of MRK-1. b Schematic diagram illustrating the domain organizations of full-length RSV and HMPV L proteins and truncated RSV L protein (L_trunc). Flexible domains not captured in pneumovirus L structures are colored in gray. c SDS-PAGE of purified L-P and L_trunc-P proteins migrated alongside a BenchMark protein ladder (Invitrogen). d Titration of unlabeled MRK-1 against radiolabeled MRK-1 to determine the affinity of binding to wild-type (WT) RSV L-P. The data show the mean and data points for three technical replicates, representative of one of three independent experiments (see also Supplementary Fig. 1a, c). Dissociation constant (K_D) is denoted under the curve. e Analysis of competition between unlabeled NNIs and radiolabeled MRK-1 for binding to WT RSV L-P. Data show the mean and data points for three technical replicates, representative of one of three biological replicate experiments (see also Supplementary Fig. 1b, d). f Schematic diagram illustrating the design of the RNA synthesis assay. g RNA synthesis activities of RSV L-P and L_trunc-P from position 3C of the RSV trailer (tr) promoter. A phosphorimage representing one of three independent experiments is shown. Lane 1 shows a ladder representing products generated from position 3C of a trailer 1-25 promoter. Note that the ladder RNAs have a 5′ monophosphate and products ≤6 nt in length migrate differently than RNA with a 5′ triphosphate. Lane 2 shows a marker for the pppGpA dinucleotide.

Fig. 2. Structure of RSV L in complex with MRK-1.

a The structure of RSV L_trunc-P in complex with the MRK-1 ligand (PDB ID: 8FPI). The RdRp and PRNTase domains are in blue and green, respectively, and the P tetramer is in pink. A magnified view of MRK-1 and its binding pocket are shown. b Magnified view of the MRK-1 binding pocket with L side chains shown as green sticks. Ile1381 is shown in pink and the HR motif in orange. c Comparison of the binding site between MRK-1 bound and apo RSV polymerase structures (PDB IDs: 8FPI and 6PZK, respectively). The MRK-1 liganded structure is superimposed on the apo L structure. The apo RSV L structure is shown in gray, with the MRK-1 bound structure shown in green. Rearrangements are indicated with black arrows.

Fig. 3. Comparison of MRK-1 binding in RSV and HMPV polymerases.

a Aligned electron density maps of MRK-1 bound in the RSV and HMPV polymerases (PDB IDs: 8FPI and 8FPJ, respectively). b Comparison of the binding site between MRK-1 bound RSV (protein side chains in green, compound in gray) and HMPV (protein side chains in brown, compound in light blue) L structures (PDB IDs: 8FPI and 8FPJ, respectively). Labeled residues have undergone similar structural rearrangement in both RSV and HMPV L proteins upon MRK-1 binding.

Fig. 4. MRK-1 inhibits RSV synthesis initiation and early elongation.

a Representative Northern blot analysis of minigenome-specific RNAs generated in the minigenome system. The blots show input minigenome, antigenome, and mRNAs 1 and 2 generated in the presence of varying concentrations of MRK-1. Lane 1 is a negative control in which the plasmid expressing the L protein was omitted from the transfection. b Quantification of mRNA 1 and antigenome RNA signal from two biological replicate Northern blot experiments, normalized to input minigenome. c Representative primer extension analysis of RNA generated from the minigenome in the presence of varying concentrations of MRK-1. The RNA was analyzed with a primer that corresponds to leader nucleotides 15–39 and detects RNA initiated at leader promoter positions 1U and 3C. Lanes 1 and 2 are 39 and 37 nt markers representing sizes of cDNA products generated from RNAs initiated at 1U or 3C, respectively. d Quantification of 1U and 3C products from two or three biological replicate experiments, normalized to RNA with no inhibitor. e Representative primer extension analysis of RNA generated from the minigenome analyzed with a primer that corresponds to nucleotides 12–31 of gene 1 and detects RNA initiated at the first gene start signal. The gel shows capped and uncapped RNAs generated in the presence of varying concentrations of MRK-1 or in the presence of 800 nM BI-D. Lane 1 shows a 31 nt marker representing the cDNA generated from uncapped RNA. Lane 2 shows RNA generated in transfections with an L protein containing H1338A and R1339A substitutions. f Quantification of capped and uncapped mRNAs from two or three biological replicate experiments, normalized to RNA generated by wt L protein with no inhibitor. g Schematic diagram illustrating the design of the RNA synthesis assay. h RNA synthesis products generated by polymerase in an in vitro RNA synthesis assay in the presence of varying concentrations of MRK-1, analyzed by denaturing polyacrylamide gel electrophoresis. A phosphorimage representing one of three independent experiments is shown. Vertical lines indicate where lanes were excised due to loss of the RNA pellet during purification. i Quantification of the different RNA products generated in vitro. Each value was normalized to minus inhibitor control, which was set to 1. The data show the mean individual data points of two or three independent experiments. j–l Comparisons of the VSV and RSV L structures, showing the apo structure of VSV L (PDB ID: 6U1X) (j), RSV L (PDB ID: 6PZK) (k), and an overlay of the RSV L_trunc-MRK-1 and VSV L structures (PDB IDs 8FPI and 6U1X, respectively; l). Likely priming residues in VSV and RSV L proteins are indicated with spheres. In each case, the GDN motif in the RdRp domain is shown as yellow sticks. The indicated priming loops included the following amino acid residues: VSV L: 1146–1172 with Trp1167 indicated as the priming residue; RSV L: 1256–1282 with Pro1261 and Trp1262 indicated as the priming residues.