Figure 1.
The rotavirus particle and conformational transformations during entry.
A. Surface representation of the mature rotavirus TLP [9]. VP4 is in magenta (VP8*) and red (VP5*); VP7, in yellow; VP6 (from underlying DLP), in green; VP2 (beneath VP6 and visible just through the opening directly facing viewer), in blue. The projecting part of the VP4 spike is twofold clustered, but the base (buried beneath VP7 in this surface view) is a threefold symmetric trimer. Compare the ribbon diagram on the left in panel. C. Scale bar = 100 Å. B. Domains of VP4. Trypsin cleavage generates VP8* (magenta) and VP5* (red), by excising residues 231–248 from two of the subunits in a trimer and residues 27–248 from the third. Refer to the left-hand image in C for a three dimensional ribbon representation of the mature trimer. “Head” is the receptor-binding, lectin domain of VP8*; “body”, the β-barrel domain of VP5*, with hydrophobic loops at its tip; “foot”, the largely α-helical domain clamped beneath VP7 in the TLP. Numbers above the bar refer to amino-acid residues; lines immediately beneath the bar, substructures of the protein (α: N-terminal helix, anchors VP8* into the foot; lectin: receptor-binding domain of VP8*; β-barrel: projecting domain of VP5*; c-c: coiled-coil structure that forms when VP5* folds back, as in the right-hand image in C). C. Proposed correlation of conformational changes in the mature VP4 spike with remodeling and breaking of a bilayer membrane. VP7 (yellow), VP6 (green), and VP2 (blue shell) are shown schematically. Left: mature spike, anchored into the DLP by VP7, a Ca2+-stabilized trimer. VP4 is activated by cleavage to VP8* (magenta ribbons) and VP5* (red, orange and gold ribbons). Red stars: hydrophobic loops of VP5* β-barrel; blue diamonds: glycolipid head groups to which the lectin domains of VP8* attach. Center: Schematic diagram of the product of an initial conformational change, in which VP8* lectin domains dissociate from their positions covering the hydrophobic loops of VP5*, exposing them for interaction with the membrane. VP8* lectin domain is simplified here to a magenta oval; VP5*, to red, orange, and gold ovals (β-barrels) connected by flexible links to the foot (light red cylinder). Right: loss of Ca2+ causes VP7 (yellow) to dissociate from DLP, releasing VP5*, which is now free to fold back to its most stable, postcleavage conformation. Interaction between membrane and VP5* hydrophobic loops (red stars) couples membrane distortion and rupture (dashed lines) to folding back of VP5*. See [2], [9] and text of this paper for further structural details.
Figure 2.
Fluorescent labeling of recoated TLP.
A. Images of particles, adsorbed to coverslip, at each of three wavelengths and overlay. B. Analysis of recoated, purified TLPs by 12% SDS-PAGE. C. Fluorescent focus assay measuring effect of labeling on reacoated particle infectivity, with labeling as shown. D. Images of doubly labeled (VP7, green; DLP, red) recoated particles, 10 min and 45 min following addition to BSC-1 cells. Uncoated particles (red) are evident at the later time point. E. Time lapse sequence of triply labeled particles uncoating in BSC-1 cells. Time frames at 5 sec intervals from an arbitrary time point following addition of particles to cell. Pseudo-colors: VP7, blue; VP4, green; DLP, red. F. Graphic representation of time lapse sequence in panel E; images recorded at 1 sec intervals. Intensities (colors as in E) evaluated in a box moving with the DLP; orange curve: velocity of DLP.
Figure 3.
Statistics of entry and uncoating.
A. Percent of doubly labeled TLPs (labels on DLP and VP7) that have uncoated, as a function of time after addition of pulse of virus to BSC-1 cells. (Bolus of virus added, excess removed and replaced with fresh medium). Images recorded at 6-sec intervals. Average of three experiments; bars show standard errors. B. Histogram of interval between binding and uncoating; images recorded at 6-sec intervals. C. Detailed traces for two individual particles. Four stages color coded: lateral motion (red), capture (purple), uncoating (green), and DLP release (blue). Left-hand panels show projected particle positions (scale bar = 1 µm); right-hand panels, particle displacement for each 6-sec interval between successive frames. The dashed line shows the expected displacement in 6 sec for a particle of radius 350 Å undergoing Brownian motion in a medium of effective viscosity 10 times that of water at 37°C, as estimated from the particle-radius dependence of cytoplasmic viscosity for cells in culture [48]. D. Distribution of projected 6-sec displacements of rotavirus particles during each phase; one hundred particles analyzed for each panel. E–H. Time (in minutes) spent in each phase, for one hundred particles. Images recorded at 6-sec intervals.
Figure 4.
A. Image of three particles that uncoat after addition of fluorescently labeled m159 antibody to the medium. (Virus pre-bound to cells, excess medium withdrawn, and antibody-containing medium then added.) Particles that uncoat do not bind antibody. White line represents edge of cell; particles on the coverslip all bind antibody. B. Time to initiation of uncoating (decrease in VP7 and VP4 label intensities) of particles that do not bind antibody m159. The antibody was added 5–7 minutes after addition of virus to cells, as indicated by the arrow; images were collected every 6 seconds. C. EDTA pulse. Top panel: doubly labeled virus bound to cell, before and after EDTA pulse. Middle panel: DIC images of the same cell; Bottom panel: particles bound to coverslip. D. Percent of virus resistant to EDTA uncoating, as a function of time between addition of virus and pulse of EDTA (black open circles) and cumulative representation of data in 2B (gray solid circles).
Figure 5.
VP4 fusion-loop mutant and VP7 disulfide-linked trimer.
A. Effects of mutations on uncoating. Particles recoated with VP4 fusion-loop mutant (top panel) and VP7 disulfide-linker trimer (bottom panel) added to BSC-1 cells and imaged at 2 min and 45 min. All particles in the field at 45 sec retain VP7. B. Percent uncoating for particles recoated with wt proteins, with VP4 mutant and VP7 wt, and with VP4 wt and VP7 mutant. Average of three experiments; bars show standard error. C. Percent of particles protected from EDTA dissociation (for wt and VP4-391D) and from access by m159 (for VP7 C-C) as a function of time. Average of three experiments; bars show standard error.
Figure 6.
Association of productive (A) and non-productive (B) particles with AP-2 clathrin adaptor, dynamin, and Rab5.
“Productive particles” are those that bind and uncoat as in Figure 3; “non-productive particles”, those that fail to uncoat within 30 mins of binding. Of 62 productive particles followed in cells transfected with σ2-eGFP (A, left), 50 did not colocalize with σ2 at any time, while 12 colocalized with σ2 at an early time point. Likewise, of 62 productive particles followed in Rab5-eGFP transfected cells (A, right), only 11 appeared to colocalize with Rab5, while of 166 non-productive particles (B, right), a substantial majority ended up in Rab5 endosomes. Cells stably transfected with σ2-eGFP or Rab5-eGFP or transiently transfected with dynamin-eGFP.
Figure 7.
Thin-section electron microscopy: stages of rotavirus entry.
Experiments done with authentic virions, not recoated DLPs. A. Surface-bound particles. The center of the particle (RNA) takes up stain particularly avidly. B. Engulfment phase. In the right-hand panel, the particle may be entering by a clathrin-coated pit. C. Compartmentalized virions.
Figure 8.
Whole-cell cryo-tomography of surface-bound rotavirus particles.
A–C. Tomographic slices through individual virions, free (A) and bound (B,C) to cell membrane. Note differences in spike extension in B and C (red arrowheads). D–I. Tomographic slices (D–F) and isosurface renderings (G–I) of subtomogram averages of rotavirus particles. Subtomogram averages: (D,G & F,I) after applying icosahedral symmetry (D,G: 1080 repeats from 18 virions; F,I: 420 repeats from 7 virions); for (E,H), we classified the repeats and averaged only those 78 repeats from 5 virions for which spikes were in contact with membrane (blue density).