Fig 1.
Rotavirus structure, entry, and conformational changes of its membrane-penetration protein VP4 (VP5*/VP8*).
(A) Schematic illustration of an infectious rotavirus particle. Subunits are shown in surface representation and the particle is partially cut to allow inside view. The “double-layer” particle (DLP) is composed of VP2 (VP2A and VP2B, colored in blue and cyan, respectively) and VP6 (green). The outer layer of the “triple-layer” particle (TLP) also contains VP7 (yellow) and VP4 (VP5*/VP8* after proteolytic cleavage). VP5* is in red, orange and salmon; VP8*, in magenta; VP1, the capsid-bound RNA-dependent RNA polymerases (RdRp), in gray. The RNA cap-generating VP3, which has not been localized at a defined position in the virion, is not shown. (B) Schematic drawing of steps in rotavirus entry, summarizing observations from live-cell imaging [24,26,61] and biochemical and structural studies [16,18]. Entry requires trypsin-catalyzed cleavage of the VP4 spike protein into VP5* and VP8*. VP8* binding to glycolipid headgroups allows the TLP to attach to the surface of a host cell. This interaction, potentially in concert with membrane engagement of hydrophobic loops on VP5* (see panels C and D), allows the particle to enclose itself in an inward-budding vesicle. Subsequent steps include loss of Ca2+ from the vesicle interior, dissociation of VP7 and VP8*/VP5* from the DLP, free diffusion of the DLP in the cytosol, and initiation of RNA transcription. (C) Structure of the VP5*/VP8* spike in upright conformation (PDB-ID 6WXE) [18], colored as in (A). The linear domain organization of the upright conformation is shown at the bottom. (D) Structure of the VP5* spike in reversed conformation (PDB-ID 6WXG) [18], colored as in (A). The linear domain organization of the reversed conformation is shown at the bottom. Residues of the foot VP5* domains that were extruded from the VP7 layer cavity upon structural transition are shown schematically in gray.
Fig 2.
Interaction with rotaviruses makes liposomes Ca2+ permeable.
(A) Schematic overview of the TIRF experimental setup. (B) Representative fluorescence images, showing a field of view for the control (left, liposomes only), ionophore (center, liposomes incubated with ionomycin) and TLP (right, liposomes incubated with TLPs) samples. The 640 nm channel (Cy5-DOPE, liposome signal) is in cyan and shifted 7 pixels to the right in all images to visualize colocalization; the 488 nm channel (Fluo-4, Ca2+ sensor signal) is in green; the 561 nm channel (Atto 565 NHS ester dye, virus signal) is in magenta and shifted 14 pixels to the right in the TLP sample. The scale bar corresponds to 5 μm. Representative examples of liposomes are boxed (1–6) and displayed beneath the images at higher magnification (scale bar = 0.5 μm) with the three channels shown next to each other. (C) Normalized frequency of the liposome intensity signal in arbitrary units (AU) from the TLP sample (experiment 1). The grey line shows the distribution for internal control liposomes in the TLP sample (liposomes that did not colocalize with a virus particle in the TLP sample) (n = 28,985). The hatched grey bars show the distribution for the liposomes that colocalized with a virus (n = 415). (D) Ca2+-positive liposomes (%) for the control sample (white bars), the ionophore sample (grey bars), the internal control in the TLP sample (liposomes in the TLP incubated sample that do not colocalized with TLP, hatched white bars), and TLP sample (liposomes that colocalized with TLP, hatch grey bars) from experiments 1, 2 and 3, respectively. "Ca2+-positive liposomes" means percent of the liposomes that had Ca2+ signal intensity greater than 1.5 standard deviations from the mean in the corresponding internal control sample (see red dashed lines in S2A and S2C Fig for the cutoffs). n = total number of liposomes in each condition. (E) Average percent (%) Ca2+-positive liposomes for the control sample (white bar), the ionophore sample (grey bar), the internal control in the TLP sample (hatched white bar), and TLP sample (hatched grey bar) from the three independent experiments (S1 and S2 Figs). Error bars are the standard error calculated from the mean of the independent experiments. Statistical analysis used a t-test between samples. * = p<0.05; ** = p<0.01; *** = p<0.001.
Fig 3.
Exposure of rotaviruses to pH 11.0 before liposome incubation prevents Ca2+ permeabilization.
(A) Average percent (%) Ca2+-positive liposomes, from at least three independent experiments (S3 and S7A Figs), for control liposomes (white bars), ionophore-exposed liposomes (grey bars), internal control in the wt rcTLP sample (liposomes in the wt rcTLP sample that did not colocalize with wt rcTLP: dotted white bars) and wt rcTLP sample (liposomes that colocalized with wt rcTLP: dotted grey bars). (B) Average % Ca2+-positive liposomes, from three independent experiments (S5 and S7B Figs), for the internal control and wt rcTLP sample pre-incubated at pH 8.0 (dotted white and dotted grey bars, respectively), and for the internal control and the wt rcTLP sample and wt rcTLP pre-incubated at pH 11.0 (dotted white and dotted grey bars on grey background, respectively). (C) Average % Ca2+-positive liposomes, from three independent experiments (S6 and S7C Figs), for the internal control and the mutant rcTLP sample pre-incubated at pH 8.0 (crossed hatched white and crossed hatched grey bars, respectively), and for the internal control and the mutant rcTLP sample pre-incubated at pH 11.0 samples (crossed hatched white and crossed hatched grey bars on grey background, respectively). "Ca2+-positive liposomes" designates liposomes that had a Ca2+ signal intensity greater than 1.5 standard deviations from the mean in the corresponding control sample (see red dashed lines in S3A, S3C, S5A, S5C, S6A and S6C Figs for the cutoffs). Error bars are the standard error calculated from the mean of the independent experiments. Statistical analysis used a t-test between samples. n.s. = p>0.05; * = p<0.05; ** = p<0.01; *** = p<0.001.
Fig 4.
Cryo-EM analysis of rotavirus in the presence of liposomes.
(A) Sample preparation scheme and representative micrograph showing rcTLP particles bound to liposomes. Scale bar: 50 nm. (B) Illustration of how template matching was used to locate virions in the micrographs and estimate their z position based on per-particle defocus fitting. (C) Histogram of estimated ice thickness of the cryo-EM samples. For each micrograph that contained at least two particles, we calculated the largest z difference and added 100 nm (approximately two times the particle radius). (D) Mapping of spike positions and their class assignment onto the micrographs of liposome-bound viruses, color-coded as shown in panel (E). The asterisks indicate a cluster of VP5* trimers around the five-fold axis of the virus. Scale bar: 50 nm. (E) Representative reconstructions of the different spike conformations after subparticle classification.
Fig 5.
Cryo-EM reconstruction of membrane-bound VP5* in the reversed conformation.
(A) Reconstruction of class 6 (S13B Fig) in context of the full virus. The map was low-pass filtered at 8.0 Å resolution. The outer-shell VP7 is colored yellow; VP5* is colored red, orange, and salmon; membrane density is gray. The asterisk denotes the VP5* trimer on which the subparticle classification was focused. The direction of one of the icosahedral five-fold axis is indicated. (B) View around the five-fold axis showing multiple VP5* trimers interacting with liposome. The asterisk denotes the VP5* trimer on which the subparticle classification was focused. (C) Close-up view of the membrane-bound VP5* trimer. The map is partially cut. (D) Slab views of the reconstruction showing diffuse lipid bilayer and membrane-embedded VP5* foot domain density.
Fig 6.
Interactions between the reversed VP5* trimer and the membrane.
(A) Placement of the rotavirus-bound VP5* trimer structure in reversed conformation (PDB-ID 6WXG) [18] into the density map of the class 6 reconstruction (Fig 5). Density is shown as gray mesh, docked viral protein subunits are shown in ribbon representation and colored yellow (VP7), red, orange and salmon (VP5*). Close-up views are shown for the hydrophobic loops at the tip of one VP5* β-barrel that interact with the proximal membrane leaflet (1), for the hydrophobic loops at the tip of a second VP5* β-barrel that interact with the proximal membrane leaflet (2), and for the trimeric VP5* coiled-coil extending into the membrane (3).