Imaging individual retroviral fusion events: from hemifusion to pore formation and growth

Abstract

Viral fusion proteins catalyze merger of viral and cell membranes through a series of steps that have not yet been well defined. To elucidate the mechanism of virus entry, we have imaged fusion between single virions bearing avian sarcoma and leukosis virus (ASLV) envelope glycoprotein (Env) and the cell membrane. Viral particles were labeled with a lipophilic dye and with palmitylated enhanced YFP that was incorporated into the inner leaflet of the viral membrane. When individual virions were bound to target cells expressing cognate receptors, they transferred their lipids and contents only when exposed to low, but not neutral, pH. These data are consistent with the proposed two-step mechanism of ASLV entry that involves receptor-priming followed by low pH activation. Most importantly, lipid mixing commonly occurred before formation of a small fusion pore that was quickly and sensitively detected by pH-dependent changes in palmitylated enhanced YFP fluorescence. Nascent fusion pores were metastable and irreversibly closed, remained small, or fully enlarged, permitting nucleocapsid delivery into the cytosol. These findings strongly imply that hemifusion and a small pore are the key intermediates of ASLV fusion. When added before low pH treatment, a peptide designed to prevent Env from folding into a final helical-bundle conformation abolished virus-cell fusion and infection. Therefore, we conclude that, after receptor-activation, Env undergoes low pH-dependent refolding into a six-helix bundle and, in doing so, sequentially catalyzes hemifusion, fusion pore opening, and enlargement.

Fig. 1.

Lipid and content mixing between virions and target cells. Fluorescence and DIC images of cells expressing Tva receptor bound to pYFP/DiD- or to gagYFP/DiD-labeled virions (yellow) are shown in A and C, respectively. Particles that contain only gagYFP or pYFP are pseudocolored green, and particles labeled only with DiD are shown in red. The first images (0 sec) were taken at neutral pH. Fusion was triggered by applying pH 5.4 Mes buffer at 37°C; extracellular solution was reneutralized at the end of experiment (last images). The onset of perfusion with acidic and neutral buffers in B and D is indicated by inverted and upright triangles, respectively. The virions that transferred pYFP and/or DiD to the cell membrane are indicated by arrows in A and C.(A) For the virions that contained (^**) or did not contain (^*) DiD, lowering the pH resulted in an immediate loss of pYFP fluorescence (without changes in DiD signal), whereas raising the pH led to its recovery (last image). (B) Changes in pYFP (light green) and DiD (red) signals of the virion are indicated by an arrow in A. For comparison, pYFP (dark green) and background DiD (purple) fluorescence is shown for the virion marked by an asterisk. (C) Lipid mixing between gagYFP/DiD-labeled virions and target cells (arrows). (D) Plots of fluorescence as a function of time for the viruses indicated by arrows in C. Light green (gagYFP) and red (DiD) traces show fluorescence of the upper particle; dark green and purple traces are the respective fluorescence intensities of the lower particle in C.

Fig. 2.

Kinetics, pH-dependence, and receptor requirement of virus-cell fusion. (A) Waiting times from lowering the pH to the onset of lipid mixing were measured, ranked, and plotted as a fraction of fused virions vs. time. Viruses labeled with gagYFP (+ and ○) and pYFP (▵) were induced to fuse to 293Tva800 cells by applying pH 5.4 Mes buffer at 37°C. Alternatively, acetate buffer of the same acidity was used to trigger fusion of pYFP-labeled virions (▴). In control experiments, gagYFP-labeled viruses were spinoculated at 4°C onto adherent cells and exposed to pH 5.4 Mes buffer (+). (B) Analysis of DiD (open bars) and pYFP (filled bars) mixing events between virions and 293Tva800 cells within 5 min at pH 5.4 (first column, n = 1,341 particles) and within 10 min at neutral pH (third column, n = 272). Low pH-induced lipid and content transfer from viruses to the noncognate Tvb receptor-expressing (293Tvb) cells (fourth column, n = 197) and to 293Tva800 cells pretreated with 15 μg/ml R99 peptide (second column, n = 258) are also shown.

Fig. 3.

Evolution of small fusion pores. Target cells were prebound to EnvA-expressing pseudoviruses, and fusion was induced by pH 5.4 Mes buffer. (A-C) DiD (dotted line) and pYFP (solid line) fluorescence of individual virions as a function of time. The black bars mark the time during which the pH was lowered. pYFP dequenching events are marked by arrows. The inferred changes in permeability of a small nonenlarging pore, transiently opened pore, and enlarging pore are shown in A-C, respectively (○). (D) Schematic representation of changes in intraviral pH of a “leaky” virion caused by fusion pore formation.

Fig. 4.

Open time (t_O) distribution of transient fusion pores (A) and delay times (t_D) between lipid mixing and pore formation (B). (A) (Inset) The open time was determined as the interval between dequenching of pYFP fluorescence and its subsequent disappearance (solid line). The black bar marks the time interval during which the low pH was applied. Recovery of pYFP fluorescence upon returning to neutral pH shows that the pore (open circles) had closed at low pH without allowing pYFP transfer. (B) Analysis of the delay times from DiD mixing to formation of small fusion pores determined by pYFP dequenching (•) or to formation of somewhat enlarged pores that permit transfer of pYFP (○). DiD and pYFP transfer events that started within one interframe interval (7 sec) were defined as simultaneous (t_D = 0). (Inset) The interval between DiD and pYFP transfer was measured as shown.

Fig. 5.

Viruses enter receptor-expressing cells after low pH-induced fusion at the cell surface. ASLV-A particles were spinoculated onto 293Tva800 cells, after a 1-h pretreatment of the cells with 200 nM BafA1. Virus-cell fusion was then triggered by incubation at either pH 7.0 or 5.0 for 30 min (first and second bars, respectively). Alternatively, pH 7.0 was applied for 10 min, followed by pH 5.0 for 20 min (third bar). Cells were then incubated for 6 h in the presence of BafA1, and the early viral DNA products were quantified by using a quantitative PCR amplification assay. For control purposes, ASLV-A was spinoculated onto the cells and infection was allowed to proceed in the absence of BafA1 (fourth bar). The data are representative of three independent experiments that were each performed in triplicate. The error bars indicate standard deviations.