Sequential and separable fusions of OMM and IMM observed during transient and complete fusion events. (A) Scheme showing the possible paths of transient/complete fusion observed by PEG fusion of cells expressing IMS-targeted cyto c GFP with mtRFP-expressing cells. Fusion of the OMM results in diffusion of the cyto c GFP into the ‘red' mitochondrion usually followed by IMM fusion and diffusion of the mtRFP into the ‘green' mitochondrion. In rare cases, the OMM fusion is reversed without IMM fusion having occurred (dashed arrow). (B) Close temporal coupling of OMM and IMM mergers observed with cyto c GFP and mtRFP as described above. The separation of events is not visible in the images, although the graph reveals an ∼1-s lag between exchange of IMS and matrix contents. (C) Clearly separated OMM and IMM fusions during transient fusion. Mitochondrion #2 becomes yellow from intake of cyto c GFP ∼6 s before mtRFP enters #1. (D) OMM fusion can be completely uncoupled from IMM fusion: an example of the rare OMM-only transient fusion. Mitochondrion #1 donates cyto c GFP to #2 and re-separates (60 s) without having merged matrices. Minutes later #1 undergoes a matrix fusion with another mitochondrion (382 s). (E) Slow and partial mixing of integral IMM proteins during transient fusion after photoactivation in H9c2 cell expressing both IMM-PAGFP and mtRFP. (F) Comparison of the transfer of soluble matrix and integral IMM proteins during transient and complete fusions in cells expressing mtPAGFP/mtRFP or IMM-PAGFP/mtRFP (mean±s.e.m.; ^*P⩽0.001).

Opa1 amount controls both fusion quantity and quality. (A) Effects of Opa1 siRNA, overexpression (OE) or AV on expression levels of Opa1. (B) Effects of Opa1 siRNA, OE, or AV on mitochondrial morphology. (C) Effects of Opa1 siRNA on mitochondrial fusion quantity and quality. Means±s.e.m.; P⩽0.03. (D) Effects of Opa1 OE or AV mitochondrial fusion quantity and quality. Means±s.e.m.; ^*P⩽0.003. (E) The time course of typical side-by-side complete fusion after photoactivation in Opa1-overexpressing cells. The donor (yellow) mitochondria moves into oblique opposition with the accepter. At 0 s the transfer of photoactived mtPAGFP has begun. Two mitochondria completely fuse to one at 56 s. (F) Summary of Opa1 amount effects on fusion quantity and quality.

Dependence of transient and complete fusion on mitochondrial movements and anchorage to the cytoskeleton (A) Longitudinal (end-to-end) interaction between two mitochondria along a common microtubular track resulted in complete fusion. Tubulin-GFP and mtRFP fluorescence is shown in green and red, respectively. mtPAGFP fluorescence gained after photoactivation is also shown in green. The white dash marks the area of photoactivation, the white arrows indicate the vector for the force produced by the motor proteins, and the numbers in white indicate the mtPAGFP donor (#1) and acceptor (#2) mitochondria, respectively. (B) Oblique (side-by-side) interaction between two mitochondria travelling along separate microtubular tracks. Here, tubulin-GFP fluorescence is shown in grayscale and mtRFP fluorescence is shown in red, whereas mtPAGFP is shown in green only in the linescan images. The mtPAGFP donor is marked by #1 and the acceptor is labeled #2. The direction of movement is marked by white arrows. A line scan-type image was created for both mtRFP (red) and mtPAGFP (green, after subtraction of the pre-photoactivation image), respectively, by selecting a line parallel to the direction of movement of the acceptor, obtaining the corresponding fluorescent signal from every 2D images in the time series and stacking the successive lines horizontally. mtRFP is visible in the acceptor mitochondrion at every time point, but mtPAGFP appears only after the interaction with the donor took place at 0 s. (C) The number of microtubular tracks supporting the movement of the pair of mitochondria undergoing complete or transient fusions. (means±s.e.m.). (D) Merging of mitochondria is suppressed in NCD-pretreated and VP-stimulated cells. The KFP fusion assay was conducted in naive, NCD-pretreated (10 μM for 30 min) cells and in cells stimulated with 100 nM VP after photoactivation. Occurrence of fusion events/measurement is plotted (naive, n=43; NCD, n=14; VP, n=15). (E) Rapid re-separation of merged mitochondria is inhibited by VP. The fraction of merging mitochondria that showed re-separation is shown for both naive (n=47) and VP-stimulated (n=9) cells. (F) Comparison of the kinetics of the inhibition of mitochondrial motility and suppression of the merging of the mitochondria induced by VP (100 nM). Confocal images show both mtYFP fluorescence (grayscale) and at each time point, the sites of mitochondrial movement calculated by subtraction of sequential images (red for positive changes and green for negative changes). To quantitate mitochondrial motility, the pixels that change more than a threshold value were calculated and normalized as the percentage loss from the average prior to stimulation (control) (Yi et al, 2004, means±s.e.m., upper graph). The time of the merging of mitochondria was also recorded and the frequency of events in the VP-treated cells (n=23) was normalized to the control (n=158, lower graph).

Visualization of transient mitochondrial fusion by photoactivation or cell fusion. (A) Labelling of three subsets of mitochondria by photoactivation of KFP in an H9c2 cell expressing both mtKFP and mtGFP. (B) Time course of a typical mitochondrial transient fusion after photoactivation. The acceptor (green) mitochondrion moves into oblique apposition with the photoactivated donor (red). At 0 s the transfer of photoactivated KFP has begun. By 73 s the pair has re-separated at the apparent site of fusion and moved apart. (C) Duration of transient fusion. (D) A PEG-induced fusion between mtRFP- and mtYFP-expressing cells just before the intermixing of contents begins and 12 min later. (E) An individual transient fusion event from an mtRFP–mtYFP cell hybrid demonstrates the bidirectional exchange of contents. At 4 s both mitochondria have become yellow and re-separation occurs at 12 s. Bidirectionality is confirmed by the graphs showing a decrease in green and increase in red in the green donor (solid lines) and the inverse in the red donor (dashed lines). (F) A representative fusion from the PEG fusion assay shows the canonical end-to-end interaction of tubular mitochondria and also the bidirectional redistribution of matrix contents.

IMM ΔΨ_m and transient and complete mitochondrial fusion. To avoid light-induced damage, the far-red fluorescent ΔΨ_m-sensor dye DiIC₁(5) (50 nM) was added to cells immediately after photoactivation and was monitored as it accumulated in the mitochondria. (A) Mitochondrial fusion after dissipation of the ΔΨ_m. After 7-min monitoring in the presence of DiIC₁(5), uncoupler (5 μM FCCP, 5 μg/ml oligomycin) was added resulting in a loss of ΔΨ_m and consequent release of the probe to the cytosol. The graph shows that redistribution of the dye was complete ∼30 s after addition of uncoupler (dotted line); still, ∼60 s later, an exchange of matrix contents between two mitochondria occurred (solid line). These mitochondria are marked by white arrows in the upper left image. Immunoblots show time-dependent loss of the long form of Opa1 in cells treated with uncoupler. (B) ΔΨ_m is not lost during transient or complete fusion. Two representative events are shown in a cell after photoactivation and addition of DiIC₁(5). In neither case is there a disturbance in the ΔΨ_m during the exchange of contents as indicated by the steady increase in DiIC₁(5) fluorescence shown in blue in the lower row of images and the dashed lines in the graphs.

Drp1 is required for separation during transient fusion. (A) Effect of Drp1 domain negative or Drp1YFP overexpression on mitochondrial morphology. H9c2 cells were transfected with mtRFP, mtRFP/Drp1K38ACFP, and mtRFP/Drp1YFP. (B) Effect of Drp1 domain negative or Drp1YFP overexpression (OE) on mitochondrial fusion quantity and quality. Statistically significant differences from the control are marked by an asterisk (P⩽0.04). (C) Effect of Drp1 domain negative or Drp1YFP overexpression on durations of transient fusion. (D) The time course of Drp1YFP fluorescence involved in mitochondrial transient fusion. At 0 s the photobleaching of mtRFP fluorescence is finished. (E) Transient fusion and Drp1YFP fluorescence time chart for individual mitochondria.

Mitochondrial movements and metabolism depend on transient fusion. (A) Differential control of mitochondrial motility during transient and complete fusion. Left: Fraction of merging mitochondria that showed >1.5 μm/5-min displacement after complete and transient fusion is shown (n=10 and 11, respectively). Right: Maximal displacement is plotted against the elapsed time for complete fusion (0–279 s, n=10), for the merging (0–42 s, n=11) and post re-separation (47–67 s, n=6) phases of transient fusion. The grey arrows indicate the temporal sequence of the events. (B) Mitochondrial morphology in a control cell, a control cell with fragmented mitochondria, and an Opa1-overexpressing (high level) cell. (C) Comparing the fusion quantity and quality between control cell, control cell with fragmented mitochondria, and Opa1-overexpressing (high level) cells. (D) TMRE loading of control cells with fragmented mitochondria, and Opa1-overexpressing (high level) cells. The TMRE loading of mitochondria was normalized to the response in the non-transfected control cells. Means±s.e.m.; ^*P⩽0.01. (E) Oxygen depletion rate of Opa1OE AV and control Ins1 cells in the absence and presence of an uncoupler. (F) Mitochondrial movement and fusion dynamics. Mitochondrial composition and position relative to the microtubules (gray) is depicted during the time course of the transient (upper) and complete fusion (lower). The IMS/matrix space contents are marked in green and red and the membrane contents are shown in black and sky for the interacting mitochondria, respectively. Direction of the movement is marked by arrows.