Molecular mechanism of respiratory syncytial virus fusion inhibitors

Abstract

Respiratory syncytial virus (RSV) is a leading cause of pneumonia and bronchiolitis in young children and the elderly. Therapeutic small molecules have been developed that bind the RSV F glycoprotein and inhibit membrane fusion, yet their binding sites and molecular mechanisms of action remain largely unknown. Here we show that these inhibitors bind to a three-fold-symmetric pocket within the central cavity of the metastable prefusion conformation of RSV F. Inhibitor binding stabilizes this conformation by tethering two regions that must undergo a structural rearrangement to facilitate membrane fusion. Inhibitor-escape mutations occur in residues that directly contact the inhibitors or are involved in the conformational rearrangements required to accommodate inhibitor binding. Resistant viruses do not propagate as well as wild-type RSV in vitro, indicating a fitness cost for inhibitor escape. Collectively, these findings provide new insight into class I viral fusion proteins and should facilitate development of optimal RSV fusion inhibitors.

Figure 1. Inhibitors bind to a three-fold-symmetric cavity in prefusion RSV F

(a) ITC data for the binding of JNJ-2408068 (left) and JNJ-49153390 (right) to prefusion RSV F (DS-Cav1). Red lines represent the best fit of titration data to a single-binding-site model. (b) Crystal structure of JNJ-2408068 bound to RSV F as viewed from the side. RSV F is shown as a molecular surface, except for the hydrophobic fusion peptides, which are shown as ribbons. An oval-shaped window into the central cavity has been created by removing a portion of the tan-colored protomer. Surfaces inside the cavity lack specular reflections. JNJ-2408068 is shown as a ball-and-stick representation with corresponding electron density shown as a blue mesh. (c) Top view of the structure shown in b, with the upper hemisphere of RSV F removed for clarity.

Figure 2. Inhibitors tether hydrophobic residues in two structurally labile regions

(a,b) Top (left) and side views (middle) for JNJ-2408068 (a) and JNJ-49153390 (b) bound to RSV F. Each RSV F protomer is a different color (tan, pink and green), and hydrophobic side chains are shown with transparent molecular surfaces. Inhibitors are shown as ball-and-stick representations with carbon atoms colored in cyan, nitrogen atoms in blue, oxygen atoms in red, bromine atoms in dark red and sulfur atoms in yellow. At bottom are 2Dligand-interaction diagrams generated in Molecular Operating Environment; A, Band Crefer to the green, tan and pink protomers, respectively. Bonds with RSV F main chain and side chain atoms are shown as blue and green dashed lines, respectively, and an ionic interaction is shown as a purple dashed line. When present, arrowheads point toward the acceptor.

Figure 3. RSV F rearrangements required for inhibitor binding are prevented by the D489Yresistance mutation

(a) Top view of RSV F apo (PDB ID 4MMS, green) superposed with the JNJ-2408068-bound (light purple) and D489Y (tan) RSV F crystal structures. The electron density of JNJ-2408068 in the bound structure is shown as a black mesh. The three RSV F protomers (labeled A, Band C) are separated by dashed gray lines emanating from the center of the three-fold axis. Salt bridges and interprotomeric hydrogen bonds between Lys394, Thr400 and Asp489 are shown as dotted lines in the lower left for the bound structure and to the right for the apo structure, and are absent in the D489Y structure at the top left. (b) Side view of the D489Y and JNJ-2408068-bound RSV F structures, colored as in a. Hydrogen bonds and salt bridges in the bound structure are shown as dotted lines. Hydrophobic side chains in the fusion peptide (Leu138, Phe140 and Leu141) and in the HRB (Tyr489) are shown with transparent molecular surfaces.

Figure 4. Inhibitors stabilize prefusion RSV F

(a) The relative percentage of surface-expressed RSV F remaining in the prefusion conformation after a 55 °C heat shock performed in the presence of increasing concentrations of fusion inhibitors. (b) Relative amount of prefusion RSV F on the surface of cells, normalized to wild type (Fwt), as assessed by the binding of the prefusion-specific antibody CR9501. MFI, mean fluorescence intensity of the Alexa Fluor 647–conjugated antibody. (c) Relative amount of total RSV F on the surface of cells, normalized to wild type, as assessed by the binding of the conformation-independent antibody CR9503. (d) Fraction of RSV F in the prefusion conformation on the surface of cells, as determined by the ratio of CR9501-binding to CR9503-binding. Many of the inhibitor-escape mutations (gray) reduce the fraction of RSV F in the prefusion conformation on the surface of HEK293 cells, although some mutations (blue) increase the fraction or maintain a similar level as wild type (black). For a, data represent the mean (n = 1–2). For b–d, data represent the mean ± s.d., where n = 7 for Fwt, n = 5 for D486N and D489Y, n = 4 for E487D, n = 3 for T400A and n = 2 for all other mutants.

Figure 5. Effects of inhibitor-escape mutations on cell-cell fusion activity and viral fitness

(a) Relative fusion activity, normalized to wild-type F (Fwt), for RSV F variants containing inhibitor-escape mutations. Data represent the mean of two independent experiments, each replicated six times. (b) Relative RSV F surface expression, normalized to Fwt, as assessed by the binding of the conformation-independent antibody motavizumab to cells transfected in parallel with those used for the fusion assay. Data represent the mean of two independent experiments, each replicated three times. (c) The percentage of A549 cells infected with either wild-type (WT) rgRSV224 or inhibitor-escape variants (D486N or L141W) was measured by analyzing cellular GFP expression in individual cells every 60 min for 48 h starting 5 h after infection. Data represent the mean (n = 2). (d) The growth of WTRSV and inhibitor-escape variants was determined by a plaque titration assay in Vero cells. Data represent the mean of three replicates (n = 1 for each replicate).