Figure 1.
Models of virus cell-to-cell transmission.
Two models have been proposed to explain the observed recruitment of retroviruses to sites of cell–cell contact. In Model I, virus assembly initiates randomly at the surface of virus-producing cell. Once assembly is completed, virus particles are subsequently recruited to sites of cell–cell contact, followed by the spreading of the infection to the target cell. In Model II, virus assembly is initiated preferentially at cell–cell contact sites, and nascent virus spreads to the target cells after the assembly is completed.
Figure 2.
De novo assembly of murine leukemia virus in living cells.
(A) Selective frames from Video S1 illustrate de novo appearance and release of Gag-YFP–labeled MLV particles (red) in HEK293 cells. Four representative particles were labeled A–D. The last panel represents the single-particle tracking analysis for these particles. The size bar corresponds to 10 µm. (B) Quantitative analysis of fluorescence intensity for particles A–D shown in (A) over time. (C) XZ presentation of particle A from (A) over time. (D and E) Single-particle tracking for viral particles that are either released shortly after reaching maximum intensity (D) or continued to stay associated with the plasma membrane prior to release (E). XY tracks for all particles are given below the graphs.
Figure 3.
MLV transmission is contact-dependent.
Producer cells (HEK293, Cos-1 cells) expressing an intron-regulated MLV luciferase reporter inLuc, MLV GagPol, and MLV Env were cocultured with target cells (XC, NIH 3T3, HEK293 cells stably expressing mCAT1) in the presence or absence of 1% methyl cellulose as indicated, and the resulting luciferase activity originating from expression in infected target cells was measured. The ability of 1% methyl cellulose to block infection of target cells by cell-free virus was tested to the right.
Figure 4.
Assembly followed by transmission of MLV from virus-producing cells to target cells.
(A) Selective frames from Video S2 monitoring the assembly of MLV particles (Gag-YFP, red) in HEK293 cells expressing dynamin2-CFP (green) followed by the transmission to target XC cells expressing receptor mCAT1-CFP (green). Six representative particles were labeled A–F. The last panel represents single-particle tracking analysis for these particles. The size bar corresponds to 10 µm. (B) Display of the YZ movement of particle E shown in (A). (C) Single-particle tracking analysis of mCAT1/dynamin2-CFP (green), Gag-YFP (red), and particle motility (blue) from producer to target cell of particle E over time. (D) Analysis as in (C) for particles A–F.
Figure 5.
Fluorescent punctae correlate to single viral particles.
COS-1 cells expressing MLV provirus and Gag-CFP (green) in contact with XC-mCAT1-YFP (red) were imaged by time-lapse microscopy. Following the detection of virus cell-to-cell transmission, the cells were fixed and processed for scanning electron microscopy. An edged in grid was used to re-identify region X in the FEI ESEM scanning electron microscope. Two black bars were introduced for orientations in the correlative images. Correlating fluorescence and SEM identifiable viral particles are indicated by white arrows (left) and black arrows (right). The size bar in the lower right corresponds to 1 µm.
Figure 6.
Quantification of MLV assembly events in and outside of the contact zone.
(A) Six representative frames of a 157-frame video when an initiation of de novo virus assembly (blue cross) was detected. See Video S3 and Table S1 for the full analysis. The accumulation of dynamin2 and receptor-CFP was then used to define the contact zone (red line) in each frame. Homogeneous dynamin2 expression and Gag-YFP in the producer cell was used to mark the remaining plasma membrane surface outside the contact zone (white line). (B) Left panel. The analysis described in (A) identified 44 assembly events in the contact zone and eight outside the contact zone. To calculate the overall assembly frequency per surface unit (in square micrometers), the number of assembly events observed in either zone was divided by their average surface area. The resulting fold enhancement of MLV assembly at sites of cell–cell contact over the remaining plasma membrane is presented at the top right (54.5×). The underlying image used to illustrate this time-resolved analysis represents time point 00:19:43 of Video S3. Right panel. An alternative and simplified approach to define the fold enhancement was based on the accumulative merged image of all 157 frames of Video S3. Due to the dynamics of the contact zones, this approach results the definition of a broader contact zone thereby reducing the fold enhancement observed at sites of cell-cell contact (14.7×). For detailed analysis see Table S1. Size bars correspond to 15 µm.
Figure 7.
MLV assembly is directed towards sites of cell–cell contact.
(A–I) Analysis of nine representative time-lapse videos monitoring MLV assembly in HEK293 cells in contact with XC target cells as described in Figure 6B. The individual panels represent the merged images of all frames of each video. Single-particle tracking was applied to identify all de novo assembly events and to calculate assembly frequency within and outside the cell–cell contact zone. The resulting fold enhancement of MLV assembly at sites of cell–cell contact over the remaining plasma membrane is presented at the top right of each panel. For detailed analysis, see Table S2. For comparison, (A) depicts the analysis from Figure 6B. The corresponding videos for (A, B, and C) are Videos S3, S4, and S5, respectively. (J) An analysis as in (A) was performed for cocultures between a MLV-producing HEK293 cell that expresses dynamin2-CFP to label the cell–cell contact and a XC target cell expressing actin-binding Lifeact-CFP. The corresponding video for panel (J) is Video S6. (K) An analysis as in (A) was performed for cocultures between a MLV-producing HEK293 cells and a HEK293 cell expressing mCAT1-CFP. The corresponding video for panel (K) is Video S7. (L–N) An analysis as in (A) was performed for cocultures between MLV-producing Cos-1 cells and XC target cells. The corresponding video for (N) is Video S8. Size bars correspond to 15 µm.
Figure 8.
Virus assembly is not accelerated by cell–cell contact.
(A) The assembly time (minutes), defined as the time that it takes a particle to reach maximum intensity from background levels, is displayed for 129 assembly events in zones of cell–cell contact (with contact) and for 92 assembly events in the absence of cell–cell contact (without contact). The average assembly times (t) are given above the graphs. (B) Histograph of assembly time for particles assembled in the presence (with contact) and absence (without contact) of cell–cell contact.
Figure 9.
Monomeric Gag is recruited to cell–cell contact.
(A,B) HEK293 cells transfected with full-length MLV provirus lacking capsid domain (MLVΔCA-GFP, green) were cocultured with XC target cells expressing mCAT1-mCherry (red). MLVΔCA-GFP is recruited to sites of cell–cell contact with accumulation of receptor mCAT1-mCherry. Size bars correspond to 17 µm.
Figure 10.
Env localizes to sites where target cell membranes are anchored in infected cells.
(A-F) HEK293 cells expressing MLV GagPol, genome, MLV Env-YFP (red), and dynamin2-CFP (green) were cocultured with XC target cells expressing mCAT1-CFP (green). MLV Env-YFP efficiently localized to sites of cell–cell contact in addition to being incorporated into punctate virions. Size bars correspond to 15 µm.
Figure 11.
Polarized assembly depends on the presence of the cytoplasmic tail of Env.
(A–D) An experiment as in Figure 7 was performed with MLV Env lacking histidine 8 (Env ΔH8). The fold enhancement of MLV assembly at sites of contact is given at the top right of each panel. The corresponding video for (D) is Video S9. (E–H) An experiment as in (A–D) was performed for mutant Env lacking the cytoplasmic tail (Env ΔH8ΔCT). The corresponding video for (F) is Video S10. Size bars correspond to 15 µm.
Figure 12.
Polarized assembly in the context of four phases of cell-to-cell transmission.
(A) COS-1 cells expressing dynamin2-CFP (green) and producing fluorescently labeled MLV (Gag-YFP, red) were cocultured with XC cells expressing the MLV receptor mCAT1-CFP (green). Time-lapse microscopy over the period of 8.5 h revealed four “waves” of cell-to-cell transmission (Videos S11 and S12). Each wave contained four phases, the establishment of cell–cell contact, characterized by the accumulation of dynamin2-CFP and receptor mCAT1-CFP at sites of cell–cell contact (green arrow), virus assembly out of cell–cell contact sites (red arrows), cell-to-cell transmission of viral particles (red arrow), and down-regulation of the contact. The four phases presented here correspond to wave 3 in (B). (B) Graphical presentation and quantitative analysis of the four observed transmission events. The upper panels present the four time points (h∶min) of each “wave” when cell–cell contact was established prior to the induction of virus assembly. The lower panels represent merged frames of the time period of Video S12 that describes the four virus assembly and transmission events (frames 29–58, 82–118, 166–193, and 208–219 of Video S12). Finally, the quantitative analysis of these four transmission events for the adhesion markers receptor/dynamin2 (green) and viral particles (red) is given below the images. (C) The average composite of the four analyses shown in (B) illustrate the four phases of cell-to-cell transmission. Size bars correspond to 15 µm.