The structure of a prokaryotic viral envelope protein expands the landscape of membrane fusion proteins

Abstract

Lipid membrane fusion is an essential function in many biological processes. Detailed mechanisms of membrane fusion and the protein structures involved have been mainly studied in eukaryotic systems, whereas very little is known about membrane fusion in prokaryotes. Haloarchaeal pleomorphic viruses (HRPVs) have a membrane envelope decorated with spikes that are presumed to be responsible for host attachment and membrane fusion. Here we determine atomic structures of the ectodomains of the 57-kDa spike protein VP5 from two related HRPVs revealing a previously unreported V-shaped fold. By Volta phase plate cryo-electron tomography we show that VP5 is monomeric on the viral surface, and we establish the orientation of the molecules with respect to the viral membrane. We also show that the viral membrane fuses with the host cytoplasmic membrane in a process mediated by VP5. This sheds light on protein structures involved in prokaryotic membrane fusion.

Fig. 1

The overall fold of VP5. a Cartoon representation of HRPV-2 VP5 (left) and HRPV-6 VP5 (right) coloured from the N-terminal (blue) to the C-terminal (red). Residues Q258 and Q262 connecting the N- and C-terminal domains are shown as green spheres. b Cartoon and topology representation of HRPV-2 VP5 coloured by domains. c Structural differences between the N2 domains of superposed HRPV-2 VP5 (green) and HRPV-6 VP5 (pink). d Localization of the potential fusion peptide depicted as magenta sticks within HRPV-2 VP5 shown as grey cartoon

Fig. 2

Cryo-electron tomography studies of HRPV-6. a A slice through a 3D tomogram of HRPV‐6 reconstructed from tilt series data collected using a Volta phase plate. The scale bar corresponds to 40 nm. b Reconstruction of an HRPV-6 virion. 41 VP5 spikes were found from a representative HRPV-6 virion by manual picking. The EM density was projected onto the aligned positions and orientations of these spikes. c VP5 crystallographic structure fitted into its density solved by cryo-electron tomography and subtomogram averaging

Fig. 3

Effect of NaCl concentration and the Volta-phase plate on the tomographic data. a A slice through a 3D tomogram of HRPV-6 in 1.5 M NaCl buffer collected without Volta-phase plate. b A slice through a 3D tomogram of HRPV-6 in 1.5 M NaCl buffer collected with Volta-phase plate (left), and the corresponding cross section of cylindrically average spike and density distribution (right). c A slice through a 3D tomogram of HRPV-6 in 0.25 M NaCl buffer collected with Volta-phase plate (left), and the corresponding cross section of cylindrically average spike and density distribution (right). d Averaged radial density distributions of HRPV-6 in low (0.25 M NaCl) and high salt (1.5 M NaCl) are compared. The radial density distribution for the spike is similar at low- and high-salt concentration. The scale bar corresponds to 100 nm. Source data are provided as a Source Data file

Fig. 4

Microscopy of virus-cell fusion assay. Fluorescence microscopy of the Halorubrum sp. SS7-4 cells infected with fluorescently labelled HRPV-6 (a) or fluorescently labelled spikeless HRPV-6 particles (b). c A close-up of cells infected with HRPV-6 showing spreading of fluorescence evenly along the cell and viral particles fluorescing brightly. Pictures are representatives from at least three separate experiments all of which contained both treatments. Scale bar 5 μm (a, b) and 2 μm (c). DIC, cells imaged using differential interference contrasting; R18, cells imaged using fluorescence for rhodamine label; merge, an overlay of the two channels used. Source data are provided as a Source Data file

Fig. 5

Halorubrum virus membrane fusion. a HPRV-6 fusion kinetics with cells. Fluorescently labelled HRPV-6 was incubated at 37 °C with Halorubrum sp. SS7-4 host cells (grey dots) or the non-host cells Halorubrum PV6 (white dots) at 37 °C under continuous stirring and dequenching measured at 585 nm. As additional controls, HPRV-6 was incubated with Halorubrum sp. SS7-4 host cells at 4 °C (white boxes) or spikeless HRPV-6 was prepared by proteinase K digestion (TCPK) and incubated with the same host cells at 37 °C (black triangles). b HPRV-6 fusion kinetics with liposomes derived from haloarchael SS7-4 cell lipids. Fluorescently labelled HRPV-6 was incubated with liposomes at 37 °C (white boxes) and 55 °C (grey dots). As a control, spikeless HRPV-6 was incubated with haloarchaeal liposomes at 55 °C (white triangles). Results from A and B are representative for n = 3 independent experiments. c Lipid mixing kinetics of fluorescently labelled HRPV-6 particles after incubation with liposomes at 45 °C (black dots), 50 °C (white triangles) and 55 °C (grey dots) normalized to the maximum extend of fusion at 55 °C. Each segmented line represent the single exponential fit to the kinetic data. Source data are provided as a Source Data file

Fig. 6

HRPV VP5 extended conformation and proposed model. a HRPV-6 virions incubated at room temperature do not show extended spikes. b HRPV-6 virions incubated at 55 °C, showing VP5 in more extended conformation. c A cryo-EM image showing a particle with extended spikes (low salt). Scale bars correspond to 50 nm. d Proposed model for the rearrangement of HRPV spike from the pre-fusion to the membrane insertion step conformation. The N- and C-terminal domains of VP5 are coloured in blue and red, respectively, whereas the putative host recognition (N2) domain is coloured in green. VP5 would bind to the host via its N2 domain and opens up an elongated structure of N1