Respiratory chain complexes in dynamic mitochondria display a patchy distribution in life cells

Abstract

Background: Mitochondria, the main suppliers of cellular energy, are dynamic organelles that fuse and divide frequently. Constraining these processes impairs mitochondrial is closely linked to certain neurodegenerative diseases. It is proposed that functional mitochondrial dynamics allows the exchange of compounds thereby providing a rescue mechanism.

Methodology/principal findings: The question discussed in this paper is whether fusion and fission of mitochondria in different cell lines result in re-localization of respiratory chain (RC) complexes and of the ATP synthase. This was addressed by fusing cells containing mitochondria with respiratory complexes labelled with different fluorescent proteins and resolving their time dependent re-localization in living cells. We found a complete reshuffling of RC complexes throughout the entire chondriome in single HeLa cells within 2-3 h by organelle fusion and fission. Polykaryons of fused cells completely re-mixed their RC complexes in 10-24 h in a progressive way. In contrast to the recently described homogeneous mixing of matrix-targeted proteins or outer membrane proteins, the distribution of RC complexes and ATP synthase in fused hybrid mitochondria, however, was not homogeneous but patterned. Thus, complete equilibration of respiratory chain complexes as integral inner mitochondrial membrane complexes is a slow process compared with matrix proteins probably limited by complete fusion. In co-expressing cells, complex II is more homogenously distributed than complex I and V, resp. Indeed, this result argues for higher mobility and less integration in supercomplexes.

Conclusion/significance: Our results clearly demonstrate that mitochondrial fusion and fission dynamics favours the re-mixing of all RC complexes within the chondriome. This permanent mixing avoids a static situation with a fixed composition of RC complexes per mitochondrion.

Figure 1. Labelling of RC complexes in human cells by fusion to fluorescent proteins.

A. Mitochondrial RC complexes were labelled by fusion of single subunits to monomeric fluorescent proteins mEGFP, mDsRed and mDendra2, resp. at the indicated sites. B.C. Expression of the complex II subunit SDHB fused to GFP (CII-G) and its mitochondrial localization in stably transfected HeLa cells as visualized by fluorescence microscopy. D. Expression of the complex III subunit 10 (6.4 kDa UQCR) fused to GFP (CIII-G) and its mitochondrial localization in stably transfected HeLa cells as visualized by fluorescence microscopy. E. Expression of complex IV subunit 8a fused to hAGT and labelled with BG-TMR in stably transfected HeLa cells (CIV-hAGT-TMR). F. Expression of complex IV subunit 8a fused to GFP in stably transfected HeLa cells (CIV-G). G. Mitochondrial localization of complex V labelled at its γ-subunit by fusion to DsRed in stably transfected HeLa cells (CV-R). Scale bars: 20 µm (D, F, G); 5 µm (B, C), and 10 µm (E).

Figure 2. Assembly of fusion proteins into mitochondrial complexes in HeLa cells was tested by in-gel fluorescence after 2-D BN/SDS-PAGE.

GFP and Dendra2 fluorescence was preserved after SDS-PAGE and signals are shown in grayscale (lower panels). A. The same gels were silver-stained. The gel of CIV-G expressed HeLa cells was shown as an example. The positions of the fusion proteins were marked (left axis). B. Complex I – 30 kDa-GFP (CI-G) was detected in complex I, in supercomplexes (S–G) and in subcomplexes of complex I. C. Complex II-subunit B-GFP (CII-G) was predominantly found in complex II. D. Complex III-6.4 kDa subunit -GFP (CIII-G) was predominantly found in the dimer of complex III (CIII₂-G), additionally in small and large supercomplexes (S–G and III₂IV₁-G). E. Complex IV-Cox8a-GFP (CIV-G) was detected predominantly in complex IV, and smaller amounts were assembled into supercomplexes (S–G and III₂IV₁-G). F. Complex V-γ-Dendra2 (CV-D) was detected in holo complex V (V–D) and, in F₁ subcomplexes (F₁-D). At the gel front, individual fusion proteins (*) together with degradation products (**) were detected (B–D, F). S, large supercomplexes composed of respiratory chain complexes I, III₂, and IV; III₂IV₁, small supercomplexes composed of respiratory chain complexes III₂, and IV; V, complex V or ATP synthase; IV, complex IV or cytochrome c oxidase; III₂, dimer of complex III or cytochrome c reductase; II, complex II; F₁, subcomplex of complex V; G, GFP; D, Dendra2, BNE, Blue native electrophoresis.

Figure 3. Tagging of subunits of RC complexes does not interfere with mitochondrial localization, membrane-potential generation, respiration, or ultrastructure in stably transfected HeLa cells.

A–C. Expresssion of complex III subunit K fused to GFP (CIII-G) and labelling with the mitochondrial-specific dye MitoTracker Deep Red FM. C. The merged image shows clear co-localization indicating correct targeting of complex III subunit 10-GFP to mitochondria. D. The mitochondrial membrane potential in cells stable expressing the modified γ-subunit of complex V (CV-R-cells) is sustained. The merge (F) of the Complex V-γ-DsRed-signal (D) with the DASPMI-signal (E) shows the co-localization of the signal. G. Respiration of CV-R cells and cells stable expressing CIV-hAGT is not altered when endogenous substrate is oxidized, nor when substrates for the different RC complexes are added (complex I is supplied with NADH by the corresponding dehydrogenase feed with malate/glutamate, M/G; complex II is fueled with succinate, succ; and complex IV receives electrons from cytochrome c, which is reduced by TMPD in presence of ascorbate, A/T). Oxygen consumption rates derive from at n≥3 (CV-R: 4; HeLa: 3; CIV-hAGT: 5) independent cell preparations and measurements. H. Electron micrographs of mitochondria in HeLa cells. I. Electron micrographs of mitochondria in CV-R cells stably expressing the modified γ-subunit. Scale bars: 10 µm (A–C), 20 µm (D–F), 300 nm (H, I).

Figure 4. Spreading of RC in single cells due to mitochondrial fusion and fission dynamics and motility.

The re-mixing of RC-complexes in intact living cells was shown by using photoactivatable PAGFP fused to complex I. A. Scheme of events leading to a re-mixing of RC. A single event already spreads the signal due to mitochondrial fission dynamics (a), later mitochondrial fusion and fission dynamics overlay with mitochondrial motility distributing the RC further (b). B. Quantification of CI-PAGFP re-location in single cells (n = 7) due to mitochondrial fusion and fission dynamics and motility. C. Time-course of CI-PAGFP spreading in a single cell after area-restricted visualization of CI-PAGFP by photoactivation. After recording a z-t-series, a maximum intensity image from each z-stack was generated and converted into a binary image after background substraction. For better contrast, the images are depicted in the inverted mode.

Figure 5. Hybrid mitochondria following mitochondrial fusion display a patterned distribution of RC complexes.

Ongoing fusion and fission of mitochondria subsequent to cell fusion caused the generation of hybrid mitochondria with a mixed population of RC complexes. A. Appearance of mitochondria with CIII-R and CII-g 5d after coexpression. Lower part: Line plot of CIII-R and CII-G fluorescence intensities along the longitudinal axis of a single mitochondrion, indicated with a white arrow head in A. The corresponding Pearson's cross correlation coefficient is 0.89 calculated from the fluorescence distribution in a single mitochondrion as depicted in the lower part. B. CV-R and CV-G distribution in hybrid mitochondria 5 h after cell fusion. The corresponding Pearson coefficient is 0.49 calculated from the line plot in the lower part. C. CV-R and CIII-G fusion and analysis 4 h later. Lower part: Line plot of CIII-G and CV-R fluorescence intensities along the longitudinal axis of a single mitochondrion, indicated with a white arrow head in C. The corresponding Pearson's cross correlation coefficient is 0.38. D. CI-G and CV-R in mitochondria 4.5 h after cell fusion. Lower part: Line plot of CI-G and CV-R fluorescence intensities along the longitudinal axis of a single mitochondrion, indicated with a white arrow. The corresponding Pearson's cross correlation coefficient is 0.49. Some fragmented mitochondria display only CV-R fluorescence, because they did not fuse with CI-G-mitochondria (arrow). E. CI-G and CIII-R fluorescence distribution in mitochondria 5 h after fusion. The calculated Pearson for the correlation of the fluorescence distribution is 0.62. F. CI-G and mtDsRed fluorescence distribution in mitochondria in cells co-expressing both plasmids. The corresponding Pearson's cross correlation coefficient is 0.66. Scales bars 1 µm  = 30−40 pixel (A, C, D E), 1 µm  = 108 pixel (B), 1 µm  = 134 pixel (F). After recording z-stacks, the images were deconvolved by ®Autoquant software before quantitative analysis. G. Comparison of fluorescence distribution in CI-G and CV-R fused and co-expressing cells under different conditions reveal an increase in homogeneity by time due to ongoing fusion dynamics. 24 h after cell fusion the mixing of RC is complete as proved by the comparison with co-expressing cells (n≥3 independent cell assays for 4 h, 24 h CHX and coexpression; n = 2 cell assay for 23 h no CHX). H. Complex II mixes better with other complexes in co-expressing cells than complex I and V, resp. (students t-test, significance level p = 0.05) I.J. Cells with red and green fluorescent RC complexes were fused and fluorescence distribution of RC complexes in hybrid mitochondria 4–5 h after fusion as well as in mitochondria from co-expressing cells was determined. The cross-correlation of the two fluorescence channels, calculated according to Pearson, was taken as a measure for the homogeneity of RC complex distribution. For each combination at least 15 heterokaryons from 3 different independent cell fusion/co-expression assays were investigated. The data are shown as mean ± s.e.m (n≥3). Significance of differences from the relevant controls was calculated by Students' t test.

Figure 6. Model of putative cristae preservation during mitochondrial fusion.

A. Hybrid mitochondria with patterned distribution of CI-G and CV-R 4.5 h after cell fusion. B. Possible arrangement of cristae in recently fused mitochondria explaining the pattern of hybrid mitochondria. In HeLa mitochondria with regularly arranged cristae, the mean distance between cristae is 84±16 nm.