Lysosomal Regulation of Inter-mitochondrial Contact Fate and Motility in Charcot-Marie-Tooth Type 2

SUMMARY

Properly regulated mitochondrial networks are essential for cellular function and implicated in multiple diseases. Mitochondria undergo fission and fusion events, but the dynamics and regulation of a third event of inter-mitochondrial contact formation remain unclear. Using super-resolution imaging, we demonstrate that inter-mitochondrial contacts frequently form and play a fundamental role in mitochondrial networks by restricting mitochondrial motility. Inter-mitochondrial contact untethering events are marked and regulated by mitochondria-lysosome contacts which are modulated by RAB7 GTP hydrolysis. Moreover, inter-mitochondrial contact formation and untethering are further regulated by Mfn1/2 and Drp1 GTP hydrolysis respectively. Surprisingly, endoplasmic reticulum tubules are also present at inter-mitochondrial contact untethering events, in addition to mitochondrial fission and fusion events. Importantly, we find that multiple Charcot-Marie-Tooth Type 2 disease-linked mutations in Mfn2 (CMT2A), RAB7 (CMT2B) and TRPV4 (CMT2C) converge on prolonged inter-mitochondrial contacts and defective mitochondrial motility, highlighting a role for inter-mitochondrial contacts in mitochondrial network regulation and disease.

Graphical Abstract

ETOC blurb:

Wong et al. demonstrate that inter-mitochondrial contacts frequently form and restrict mitochondrial motility. Inter-mitochondrial contact dynamics are regulated by Rab7 GTP hydrolysis at mitochondria-lysosome contacts, and by Mfn1/2 and Drp1 GTP hydrolysis. Moreover, inter-mitochondrial contacts are prolonged in multiple Charcot-Marie-Tooth Type 2 mutants (Mfn2/RAB7/TRPV4), highlighting this pathway’s role in disease.

INTRODUCTION

Mitochondrial networks are critical for functional cellular metabolism and must be properly regulated (Wai and Langer, 2016), as reflected by multiple human diseases linked to defective mitochondrial dynamics (Alexander et al., 2000; Burte et al., 2015; Delettre et al., 2000; Zuchner et al., 2004). Mitochondria undergo both fission and fusion events to regulate the mitochondrial network, but the dynamic role and regulation of a third event - the formation of inter-mitochondrial contact sites - in shaping mitochondrial network dynamics is still not well understood. Indeed, the majority of previous studies have viewed mitochondrial tethering between two mitochondria merely as a transition step occurring prior to membrane fusion for generating a single mitochondria (Mishra and Chan, 2016), rather than as a distinct event which can dynamically modulate the mitochondrial network.

Mitochondria can additionally interact with other organelles such as the endoplasmic reticulum (ER) or lysosomes/late endosomes at membrane contact sites to mediate multiple forms of inter-organelle communication and maintain cellular homeostasis (Eisenberg-Bord et al., 2016; Gottschling and Nystrom, 2017; Phillips and Voeltz, 2016; Wong et al., 2018; Wu et al., 2017). However, while mitochondria fission events are marked by ER tubules and lysosomes (Friedman et al., 2011; Wong et al., 2018), how inter-organelle contacts interact with and modulate inter-mitochondrial contact tethering dynamics is still not known.

In addition, whether dysfunctional inter-mitochondrial contact site dynamics might contribute to human disease pathogenesis is also unclear. Charcot-Marie-Tooth (CMT) disease is the most common hereditary peripheral neuropathy, with Type 2 forms representing autosomal dominant forms linked to axonal degeneration (Harel and Lupski, 2014). Mitofusin 2 (Mfn2) is an outer mitochondrial membrane protein which regulates mitochondrial fusion and has been implicated in additional mitochondrial functions including mitophagy, mitochondrial transport, ER-mitochondria contacts and mtDNA stability (Chen et al., 2010; Chen and Dorn, 2013; de Brito and Scorrano, 2008; Misko et al., 2010; Santel and Fuller, 2001; Schrepfer and Scorrano, 2016), with Mfn2 mutations leading to CMT Type 2A (Zuchner et al., 2004). In contrast, Rab7 is a lysosomal/late endosomal regulator (Hutagalung and Novick, 2011), with Rab7 mutations resulting in CMT Type 2B (Houlden et al., 2004; Manganelli et al., 2012; Meggouh et al., 2006; Verhoeven et al., 2003a; Wang et al., 2014). However, whether and how CMT2A (Mfn2) and CMT2B (RAB7) disease-linked mutations converge in disease pathogenesis is unclear.

Using super-resolution confocal imaging of mitochondrial membranes in living cells, we identify inter-mitochondrial contact tethering as fundamental events which modulate mitochondrial network dynamics. Inter-mitochondrial contacts form 10x more frequently than mitochondrial fission/fusion events and functionally help to restrict mitochondrial motility. Interestingly, while the ER marks but does not promote inter-mitochondrial contact untethering, late endosomes/lysosomes actively promote inter-mitochondrial contact untethering which are regulated by lysosomal RAB7 GTP hydrolysis at mitochondria-lysosome contact sites. Inter-mitochondrial contact formation and untethering dynamics are further mediated by Mfn1/2 GTP hydrolysis and Drp1 GTP hydrolysis respectively, and modulated by mitochondrial respiration and nutrient availability. Importantly, multiple Charcot-Marie-Tooth Type 2 disease-linked mutants including Mfn2 (CMT2A), RAB7 (CMT2B) and TRPV4 (CMT2C) converge on defective inter-mitochondrial contact untethering and disrupted mitochondrial motility. Our study thus identifies a crucial role for inter-mitochondrial contacts in modulating mitochondrial homeostasis and demonstrate its dysfunction in the disease pathogenesis of multiple forms of CMT2.

RESULTS

Super-resolution Imaging of Inter-mitochondrial Contacts

To investigate the formation and dynamics of inter-mitochondrial contacts, we performed electron microscopy and super-resolution live cell imaging of mitochondria in human cells. Consistent with previous studies demonstrating mitochondria in close apposition spanning <20 nm between the outer mitochondrial membrane in fixed cells without membrane fusion (Pease, 1962; Picard et al., 2015; Vernay et al., 2017), we observed inter-mitochondrial contact sites using transmission electron microscopy (TEM) in healthy untreated HeLa cells (Figures 1A, S1A and S1B). Next, we examined their dynamics using super-resolution imaging of the outer mitochondrial membrane (mApple-TOM20) by time-lapse structured illumination microscopy (SIM) in living cells. Inter-mitochondrial contacts were stably tethered together over time (Figures 1B–1E, S1C–S1E; Video S1), with contacts forming between mitochondria of different lengths including tubular mitochondria as observed by TEM (Figures S1A and S1B), super-resolution (Figures 1B) and confocal imaging (Figure S1F). Importantly, contacts were indeed the result of two distinct mitochondria (Figures S1G and S1H) and were positive for both outer mitochondrial membrane (TOM20) and matrix (mito-BFP; COX4) proteins (Figures 1F and S1I). At any given time, ~30% of mitochondria were in stable inter-mitochondrial contacts for >10 sec (Figure 1G) and remained tethered for 39.8 ± 2.6 sec (Figure 1H).

Figure 1. Super-Resolution Imaging of Inter-mitochondrial Contacts as Important Contributors to Mitochondrial Networks.

(A) Transmission electron microscopy of inter-mitochondrial contacts (M, arrows) in untreated HeLa cells, corresponding to Figure S1B. Scale bar, 200 nm.

(B–E) Super-resolution time-lapse structured illumination microscopy (N-SIM) of inter-mitochondrial contact tethering (white arrows) and subsequent untethering (yellow arrows) in live HeLa cells (outer mitochondrial membrane (OMM) label mApple-TOM20). Scale bars, 0.5µm. See Video S1, corresponding to Figure 1D.

(F) Linescan of inter-mitochondrial contact positive for OMM label mApple-TOM20 and matrix label Mito-BFP, corresponding to line from Figure S1I (t = 0 s).

(G–H) Percentage of mitochondria forming inter-mitochondrial contacts and frequency of contact duration from confocal time-lapse images of live HeLa cells (OMM label mApple-TOM20, matrix label Mito-BFP) (n = 176 events from 10 cells).

(I–J) Percentage of tethering outcomes and time to outcome for inter-mitochondrial contact untethering and mitochondrial fusion events from confocal time-lapse images of live HeLa cells (OMM label mApple-TOM20, matrix label Mito-BFP) (n = 176 untethering and n = 19 fusion from 10 cells).

(K–L) Confocal time-lapse images of inter-mitochondrial contacts (white arrows) which do not undergo mitochondrial matrix mito-PAGFP transfer prior to untethering (yellow arrows) in live HeLa cells (OMM label SNAP-Omp25, matrix label DsRed2-mito, individual mitochondrial matrix label mito-PAGFP). The matrix of individual mitochondria were selectively locally photoactivated and tracked over time. Scale bar, 0.5µm. See Video S4.

(M) Quantification of mitochondrial event rates from confocal time-lapse images of live HeLa cells (OMM label mApple-TOM20, matrix label Mito-BFP) (n = 231 tether, n = 275 untether, n = 16 fusion, n = 21 fission from 10 cells).

(N) Percentage of events marked by ER tubules for tethered mitochondria (confirmed as distinct mitochondria by selective local photoactivation of individual mitochondria by mito-PAGFP (OMM label SNAP-Omp25, individual mitochondrial matrix label mito-PAGFP, ER label mCherry-ER)) and mitochondrial untethering, fusion, and fission events (OMM label mEmerald-TOM20, matrix label Mito-BFP, ER label mCherry-ER) in live HeLa cells (n = 88 tethered, n = 36 untether, n = 18 fusion, n = 17 fission).

Mean ± SEM; *p <0.05; **p <0.01; ***p <0.001; N.S., not significant (unpaired two-tailed t test for (I) and (J), ANOVA with Tukey’s post-hoc test for (M) and (N)). See also Figures S1–S4 and Videos S1 and S4.

Using super-resolution imaging of the outer mitochondrial membrane (mApple-TOM20), we further confirmed that inter-mitochondrial contacts were indeed distinct from previously described mitochondrial fusion or fission events. Mitochondrial fusion events involved two separate mitochondria that tethered together for >20s on average (white arrows) but subsequently fused to form a single mitochondria (Figures S2A and S2B; Video S2). Conversely, mitochondrial fission events involved a single mitochondria which divided into two transiently tethered distinct daughter mitochondria (white arrows) prior to separating (Figures S2C and S2D; Video S3). In contrast, the formation and untethering of tethered mitochondria began and ended with two distinct mitochondria (Figures 1D and 1E).

We further found that the majority of tethered mitochondria in the cell represented inter-mitochondrial contacts which subsequently untethered from each other (Figures 1C–1E; Video S1), rather than mitochondrial fusion events (89.3% untether versus 10.7% fusion, p<0.0001, n = 195 events) (Figure 1I) with their outcomes independent of tethering duration (Figure 1J). Thus, inter-mitochondrial contacts represent a previously underappreciated dynamic event which frequently occurs and are distinct from mitochondrial fission or fusion events.

Inter-mitochondrial Contacts Form Without Matrix Transfer and Modulate the Mitochondrial Network

As mitochondrial kiss-and-run transient fusion events between two mitochondria have previously been observed without requiring complete fusion of the outer mitochondrial membrane (Huang et al., 2013; Liu et al., 2009), we examined whether tethered mitochondria represented sites of transient fusion by selectively locally photoactivating the matrix of individual mitochondria labeled with mito-PAGFP (photoactivatable GFP) and tracking their fate. We found that the majority of mitochondria tethered to another mitochondria (labeled for both outer mitochondrial membrane (SNAP-Omp25) and matrix proteins (Dsred2-mito)) without mito-PAGFP transfer (Figures 1K, 1L and S2E). Moreover, mitochondria remained stably tethered in inter-mitochondrial contacts over time (Figure S2F), and even untethered from one another without mito-PAGFP transfer (Figure 1K; Video S4). While we did observe several examples of transient fusion events leading to mito-PAGFP transfer without outer mitochondrial membrane fusion (mApple-TOM20) (Figure S2G), the majority of mitochondrial tethering events resulted in untethering without bulk matrix transfer rather than in transient fusion (Figure S2H), and was not dependent on the amount of time mitochondria were tethered (Figure S2I).

To further differentiate between physical tethering and non-specific close apposition of mitochondrial membranes, we traced the movements of tethered inter-mitochondrial contact partners by tracking the point of inter-mitochondrial contact over time (10s) from time-lapse confocal live cell images. As expected, we found that tethered mitochondria could move together over time consistent with physical tethering between mitochondria (Figure S2J), resulting in inter-mitochondrial contact sites which persisted but could move within the cell.

Next, we investigated the rate of different mitochondrial events to determine which occurred the most frequently in the cell. Surprisingly, while the rates of both mitochondrial fission and fusion were <2 events/min in live HeLa cells, inter-mitochondrial contact tethering and untethering events occurred ten times more frequently (23.1 ± 3.8 tethering events/min, 27.4 ± 4.7 untethering events/min) (Figure 1M). We further confirmed that inter-mitochondrial contact tethering and untethering occurred significantly more frequently in other cell types including HEK293 and H4 neuroglioma cells (Figures S2K and S2L). Taken together, these results demonstrate that dynamic inter-mitochondrial contacts occur frequently without matrix transfer and significantly more often than mitochondrial fission or fusion events, suggesting they play a critical role in modulating the mitochondrial network.

ER Tubules Mark Inter-mitochondrial Contact Untethering Events

Inter-organelle contacts are essential for regulating organelle function and homeostasis (Eisenberg-Bord et al., 2016; Gottschling and Nystrom, 2017; Phillips and Voeltz, 2016; Wu et al., 2017), but their role in inter-mitochondrial contact dynamics has never been investigated. As we found that both the ER (Figure S1A) and lysosomes/late endosomes (Figure S1B) could simultaneously contact inter-mitochondrial contacts, we investigated whether the fate of inter-mitochondrial contacts might be coupled to contacts with other organelles.

We first examined the interaction between ER tubules and inter-mitochondrial contacts using confocal time-lapse imaging of the ER (mCherry-ER), outer mitochondrial membrane (mEmerald-TOM20) and mitochondrial matrix (Mito-BFP). Interestingly, while ER tubules did not always mark stably tethered inter-mitochondrial contacts during the initial tethering (<40%, n = 88), they marked the majority of inter-mitochondrial contact untethering events (>75%, n = 73), which was significantly greater than expected by random chance (46.1%; *p<0.05, Fisher’s exact test) (Figures 1N, S3A–S3D and Video S5).

We further confirmed that ER tubules marked inter-mitochondrial contacts which were indeed two distinct mitochondria using selective local matrix photoactivation of individual mitochondria (Figure S3E). However, ER tubule localization at inter-mitochondrial contact untethering events occurred regardless of ER-mitochondria contacts during the initial tethering (Figure S3F). Importantly, ER-mitochondria contacts did not regulate the timing of inter-mitochondrial contact untethering (Figure S3G), suggesting that they are able to mark but do not preferentially promote inter-mitochondrial contact untethering.

Moreover, in addition to marking mitochondrial fission sites (Friedman et al., 2011) (Figures 1N and S4A), ER tubules were also present at the majority of mitochondrial fusion (Guo et al., 2018) (Figures 1N, S4B, S4C and Video S6) and transient fusion events (Figures S4D and S4E). Interestingly, ER tubules marked inter-mitochondrial contact untethering or fusion events immediately prior to the event, but marked fission sites for the majority of the event (Figures S4F and S4G). Thus, ER tubules ubiquitously form ER-mitochondria contacts at multiple mitochondrial events including mitochondrial fission, fusion and inter-mitochondrial contact untethering events.

Lysosomes Both Mark and Regulate Inter-mitochondrial Contact Untethering Events

As lysosomes/late endosomes could also simultaneously contact inter-mitochondrial contacts (Figure S1B), we next examined whether the fate of inter-mitochondrial contact might be directly coupled to mitochondria-lysosome contacts. Surprisingly, using time-lapse confocal imaging of mitochondria (mApple-TOM20) and the lysosome/late endosome membrane marker LAMP1-mGFP, we found that the majority of inter-mitochondrial contact untethering events had at least one of the mitochondria simultaneously in contact with a LAMP1-positive vesicle within 10 s of untethering (94%, n = 50 events), which was significantly greater than expected by random chance (10.7%; ***p < 0.001, Fisher’s exact test) (Figures 2A–2B and S5A; Video S7). In contrast, very few tethered inter-mitochondrial contacts (Figure 2C) or mitochondrial fusion events (Figures 2C and S5B) were marked by lysosomes, suggesting that in contrast to ER tubules (Figure 1N), mitochondria-lysosome contacts preferentially mark inter-mitochondrial contact untethering events over fusion events.

Figure 2. Lysosomes Regulate Inter-mitochondrial Contact Untethering Events.

(A) Confocal time-lapse images of inter-mitochondrial contact (M-M) formation (white arrow) and subsequent untethering (yellow arrow) temporally coupled to mitochondria-lysosome (M-L) contact formation and untethering in live HeLa cells (OMM label mApple-TOM20, lysosome/late endosome label LAMP1-mGFP). Scale bar, 0.5µm.

(B) Observed localization of lysosomes marking inter-mitochondrial contact untethering events compared to lysosomal localization by random chance (Expected) in live HeLa cells (OMM label mApple-TOM20, lysosome/late endosome label LAMP1-mGFP) (n = 50 untethering events from 26 cells).

(C) Percentage of inter-mitochondrial contact untethering events and mitochondrial events (tethered state, untethering event, fission event, fusion event) marked by LAMP1-mGFP vesicles in live HeLa cells (OMM label mApple-TOM20, lysosome/late endosome label LAMP1-mGFP) (n = 25 tethered, n = 50 untether, n = 9 fission, n = 17 fusion events from 26 cells).

(D) Histogram of time between mitochondria-lysosome (M-L) contact formation and inter-mitochondrial (M-M) contact untethering (n = 47 total events from 21 cells).

(E) Histogram of time between mitochondria-lysosome (M-L) contact untethering and inter-mitochondrial (M-M) contact untethering (n = 46 total events from 18 cells).

(F) Histogram of time between mitochondria-lysosome (M-L) contact formation and inter-mitochondrial (M-M) contact formation (n = 40 total events from 18 cells).

(G) Histogram of time between mitochondria-lysosome (M-L) contact untethering and inter-mitochondrial (M-M) contact formation (n = 46 total events from 18 cells).

(H) Graph of time between mitochondria-lysosome (M-L) contact formation and inter-mitochondrial (M-M) contact untethering (n = 47 total events from 21 cells).

(I) Graph of time between mitochondria-lysosome (M-L) contact formation and inter-mitochondrial (M-M) contact formation and untethering (n = 40 total M-M formation events from 18 cells, n = 47 total M-M untether events from 21 cells).

(J) Graph of time between mitochondria-lysosome contact untethering and inter-mitochondrial (M-M) contact untethering (n = 46 total events from 18 cells).

(K) Graph of time between mitochondria-lysosome (M-L) contact untethering and inter-mitochondrial (M-M) contact formation and untethering (n = 46 total M-M formation events from 18 cells, n = 46 total M-M untether events from 18 cells).

(L) Fate of inter-mitochondrial contacts either remaining tethered or untethering in ≤10s depending on whether mitochondria simultaneously contact lysosomes (n = 45 events from 21 cells (no lysosomes); n = 47 events from 21 cells (with lysosomes)).

Mean ± SEM; *p < 0.05; **p < 0.01; ***p <0.001 (Fisher’s exact test for (B), unpaired two-tailed t test for (H-L), ANOVA with Tukey’s post-hoc test for (C)).

See also Figure S5 and Video S7.

We further investigated the timing of inter-mitochondrial contact untethering and found that it was closely coupled to both mitochondria-lysosome (M-L) contact formation (Figure 2D) and subsequent M-L contact untethering (Figure 2E). In contrast, the formation of inter-mitochondrial contacts was not temporally coupled to either M-L contact formation (Figure 2F) or M-L contact untethering (Figure 2G). Indeed, M-L contacts which formed on a mitochondria that was in contact with another mitochondria resulted in rapid inter-mitochondrial contact untethering (≤10s) (Figures 2H and 2I), as well as rapid subsequent M-L contact untethering (≤10s) (Figures 2J and 2K). Moreover, we saw multiple examples of inter-mitochondrial contacts and M-L contacts which were tethered together (Figures 2A and S5C; linescans in Figures S5D and S5E), followed by simultaneous untethering of both contact sites (Figures 2A and S5C; linescans in Figures S5F and S5G).

Finally, to examine whether lysosomes directly promoted inter-mitochondrial contact untethering, we analyzed the fate of inter-mitochondrial contacts. In the absence of M-L contacts, the majority of inter-mitochondrial contacts remained tethered (87.8%, n = 45 contacts from 21 cells) with very few subsequently untethering within 10s (12.2%, ***p<0.001) (Fig. 2L; left). In contrast, upon formation of M-L contacts, the majority of inter-mitochondrial contacts subsequently untethered within 10s (69.6%, n = 47 contacts from 21 cells) with few remaining tethered (30.3%,***p<0.001) (Fig. 2L; right), suggesting that mitochondria-lysosomal contacts directly promote the untethering of inter-mitochondrial contacts.

Inter-mitochondrial Contacts are Regulated by RAB7 GTP Hydrolysis at Mitochondria-lysosome Contact Sites

Mitochondria-lysosome contact dynamics are regulated by RAB7 (Wong et al., 2018), a GTPase master regulator of lysosomal/late endosomal dynamics (Hutagalung and Novick, 2011). GTP-bound RAB7 promotes mitochondria-lysosome contact formation, while contact untethering is driven by RAB7 GTP hydrolysis from an active GTP-bound state into an inactive cytosolic GDP-bound state mediated by TBC1D15 (Wong et al., 2018), a RAB7-GAP recruited to the mitochondria by the outer mitochondrial membrane protein FIS1 (Onoue et al., 2013; Peralta et al., 2010; Yamano et al., 2014; Zhang et al., 2005). To examine whether mitochondria-lysosome contact dynamics directly regulated the fate of inter-mitochondrial contacts, we expressed GTP-bound mutant RAB7 (Q67L) which is unable to undergo GTP hydrolysis and prevents efficient mitochondria-lysosome contact untethering (Wong et al., 2018).

While the majority of tethered mitochondria in control cells (LAMP1-mGFP) soon untethered by 90s (Figure S6A), the RAB7 (Q67L) GTP hydrolysis mutant dramatically increased the duration of inter-mitochondrial contacts which were unable to untether even after 90s compared to wild-type RAB7 (Figures 3A–3C and S6B–S6C). Interestingly, inter-mitochondrial contacts could simultaneously contact multiple lysosomes (yellow arrows) but still could not untether in RAB7 (Q67L) expressing cells (Figures 3A and S6B), suggesting that both efficient mitochondria-lysosome contact formation and subsequent untethering are critical for promoting inter-mitochondrial contact untethering.

Figure 3. Inter-mitochondrial Contacts are Regulated by RAB7 GTP Hydrolysis at Mitochondria-lysosome Contact Sites.

(A–C) Confocal time lapse images over 90s and quantification of inter-mitochondrial contact (white arrows) with increased tethered duration despite the presence of adjacent mitochondria-lysosome contacts (yellow arrows) in cells expressing mutant RAB7 (Q67L) compared to wild-type RAB7 (WT) (OMM label mApple-TOM20, lysosome/late endosome label LAMP1-mGFP (Control), RAB7(WT)-GFP or RAB7(Q67L)-GFP). Scale bar, 0.5µm. (n = 20 events from 15 cells (Control), n = 41 events from 8 cells (RAB7 (WT)), n = 55 events from 11 cells (RAB7 (Q67L))).

(D–I) Confocal time lapse images over 90 seconds and quantification of inter-mitochondrial contact (white arrows) with increased tethered duration in cells expressing mutant HA-tagged TBC1D15 (D397A) or FLAG-tagged FIS1 (LA) compared to wild-type TBC1D15 (WT) or FIS1 (WT) (OMM label mApple-TOM20, lysosome/late endosome label LAMP1-mGFP). Scale bars, 0.5µm. (E: n = 21 events from 17 cells (Control), n = 40 events from 8 cells (TBC1D15 (WT)), n = 45 events from 9 cells (TBC1D15 (D397A)); H: n = 21 events from 15 cells (Control), n = 40 events from 8 cells (FIS1 (WT)), n = 45 events from 9 cells (FIS1 (LA))).

Mean ± SEM; *p < 0.05; **p < 0.01; ***p <0.001 (ANOVA with Tukey’s post-hoc test for (B,E,H)).

See also Figure S6.

To further confirm the role of RAB7 GTP hydrolysis at mitochondria-lysosome contact sites in regulating inter-mitochondrial contact dynamics, we expressed the RAB7 GAP mutant TBC1D15 (D397A) which is unable to drive RAB7 GTP hydrolysis (Onoue et al., 2013) and the mutant FIS1 (LA) which is unable to recruit TBC1D15 to mitochondria (Onoue et al., 2013), both of which inhibit efficient mitochondria-lysosome contact untethering (Wong et al., 2018). Consistent with our hypothesis, both mutant TBC1D15 (D397A) (Figures 3D–3F and S6D) and FIS1 (LA) (Figures 3G–3I and S6E) resulted in prolonged inter-mitochondrial contacts that were unable to untether, even when in contact with lysosomes (yellow arrows) (Figures 3D and 3G). Thus, RAB7 GTP hydrolysis at mitochondria-lysosome contact sites are able to simultaneously regulate both mitochondria-lysosome and inter-mitochondrial contact untethering events.

Mfn1/2 and Drp1 GTP Hydrolysis Respectively Regulate Inter-mitochondrial Contact Formation and Untethering

We next investigated whether other GTPases including mitochondrial Mfn1/2 and Drp1 might further regulate inter-mitochondrial contact dynamics. Both Mfn1 and Mfn2 localized to sites of inter-mitochondrial contact tethering (Figures 4A and 4B) which were positive for both outer mitochondrial membrane (TOM20) and matrix (mito-BFP; COX4) proteins. Moreover, while the majority of tethered inter-mitochondrial contacts (>70%) were marked by Mfn1 (Figure 4C) and Mfn2 (Figure 4D), significantly fewer untethering events (<35%) were marked by Mfn1 or Mfn2 (Figures 4C and 4D), suggesting that mitofusins might promote contact tethering rather than untethering events.

Figure 4. Mfn1/2 GTP Hydrolysis Regulates Inter-mitochondrial Contact Formation.

(A-B) Confocal images of mCherry-Mfn1 (A) and mCherry-Mfn2 (B) localized to sites of inter-mitochondrial contact tethering (white arrows) in live HeLa cells (OMM label mEmerald-TOM20, matrix label Mito-BFP). Scale bar, 0.5µm.

(C-D) Percentage of inter-mitochondrial contact (in tethered state) and untethering events marked by mCherry-Mfn1 (C) and mCherry-Mfn2 (D) oligomers from confocal time-lapse images in live HeLa cells (OMM label mEmerald-TOM20, matrix label Mito-BFP) (Mfn1: n = 105 events from 15 cells; Mfn2: n = 91 events from 13 cells).

(E-G) Percentage of mitochondria in inter-mitochondrial contact (E) and minimum time to untethering (F, histogram in G) in live HeLa cells expressing myc-tagged Mfn1 wild-type or GTP hydrolysis mutant K88T (n = 146 events from 21 cells (Control), n = 119 events from 17 cells (Mfn1 (WT)), n = 140 events from 20 cells (Mfn1 (K88T))).

(H-J) Percentage of mitochondria in inter-mitochondrial contact (H) and minimum time to untethering (I, histogram in J) in live HeLa cells expressing myc-tagged Mfn2 wild-type or GTP hydrolysis mutant K109A (n = 146 events from 21 cells (Control), n = 140 events from 20 cells (Mfn2 (WT)), n = 98 events from 14 cells (Mfn2 (K109A))).

Mean ± SEM; *p < 0.05; **p <0.01; ***p <0.001 (unpaired two-tailed t test for (C, D), ANOVA with Tukey’s post-hoc test for (E, F, H, I).

To directly examine the role of mitofusin GTP hydrolysis on inter-mitochondrial contact dynamics, we compared the effect of wild-type Mfn1 and Mfn2 and GTP hydrolysis deficient mutants on inter-mitochondrial contact tethering. Both wild-type Mfn1 and Mfn2 expression significantly increased the percentage of tethered inter-mitochondrial contacts by ~1.5 fold (Figures 4E and 4H). In contrast, GTP hydrolysis deficient mutants Mfn1 (K88T) or Mfn2 (K109A) did not affect the percentage of tethered mitochondria (Figures 4E and 4H), suggesting that inter-mitochondrial contact tethering is directly regulated by mitofusin GTP hydrolysis. Moreover, wild-type Mfn1 or Mfn2 expression inhibited inter-mitochondrial contact untethering events resulting in significantly prolonged contact durations (Figure 4F–4G, 4I–4J, ***p<0.001). In contrast, GTP hydrolysis deficient mutants Mfn1 (K88T) or Mfn2 (K109A) did not disrupt inter-mitochondrial contact untethering events, leading to decreased contact durations compared to wild-type Mfn1 and Mfn2 (Figure 4F–4G, 4I–4J, ***p<0.001), suggesting that both Mfn1 and Mfn2 GTP hydrolysis promote inter-mitochondrial contact tethering.

We also found that Drp1 oligomers were gradually recruited to inter-mitochondrial contact sites and marked the majority of untethering events (>75%, n = 79), which was significantly greater than expected by random chance (18.7%; ***p<0.001, Fisher’s exact test) (Figures 5A, 5B and S7A–S7C). In addition, Drp1 oligomers preferentially marked inter-mitochondrial contact untethering events over stably tethered contacts (<40%; n = 94 tethered mitochondria from 19 cells) (Figure 5C), suggesting that they might regulate inter-mitochondrial contact untethering dynamics.

Figure 5. Drp1 GTP Hydrolysis Regulates Inter-mitochondrial Contact Untethering.

(A) Confocal time-lapse images of inter-mitochondrial contact tethering (white arrows) and untethering (yellow arrows) showing Drp1 oligomerization in live HeLa cells (OMM label mEmerald-TOM20, matrix label Mito-BFP, Drp1 label mCherry-Drp1). Scale bar, 0.5µm.

(B, C) Observed localization of Drp1 oligomers marking inter-mitochondrial contact untethering events compared to Drp1 localization by random chance (Expected) or on contacts from confocal time-lapse images in live HeLa cells, confirmed as distinct mitochondria by selective local photoactivation of individual mitochondria by mito-PAGFP (OMM label SNAP-Omp25, individual mitochondrial matrix label mito-PAGFP, Drp1 label mCherry-Drp1) (n = 94 tethered mitochondria and n = 79 untethering events from 19 cells).

(D-E) Percentage of events marked by Drp1 oligomerization (D) and percentage of event time marked by Drp1 oligomers (E) for mitochondrial fusion, untethering and fission events in live HeLa cells (OMM label mEmerald-TOM20, matrix label Mito-BFP, Drp1 label mCherry-Drp1). (D: n = 11 fusion, n = 24 untether, n = 12 fission; E: n = 7 fusion, n = 14 untether, n = 12 fission from 6–9 cells).

(F-G) (F) Confocal image of inter-mitochondrial contact marked by Drp1 oligomerization confirmed as distinct mitochondria by selective photoactivation of individual mitochondria by mito-PAGFP, and (G) Drp1 oligomerization dynamics in live HeLa cells (OMM label SNAP-Omp25, individual mitochondrial matrix label mito-PAGFP, Drp1 label mCherry-Drp1). Scale bar, 0.5µm. (G: n = total 79 events from 19 cells).

(H-J) Percentage of mitochondria forming inter-mitochondrial contacts (H) and minimum time to untethering (I, histogram in J) in live HeLa cells expressing mCherry-tagged Drp1 wild-type or GTP hydrolysis mutant K38A (n = 146 events from 21 cells (Control), n = 119 events from 17 cells (Drp1 (WT)), n = 133 events from 19 cells (Drp1 (K38A))).

Mean ± SEM; *p <0.05; **p <0.01; ***p <0.001; N.S., not significant (Fisher’s exact test for (B), unpaired two-tailed t test for (C,G), ANOVA with Tukey’s post-hoc test for (D, E, H, I)).

See also Figure S7.

Surprisingly, Drp1 oligomers were also recruited to the majority of both mitochondrial fission (Figure S7D) and fusion events (Figures 5D and S7E) as well as transient fusion events (Figures S7F and S7G). While Drp1 oligomers were present for the majority of time during mitochondrial fission events, Drp1 oligomers were only present immediately prior to mitochondrial fusion or inter-mitochondrial contact untethering events (Figures 5E and S7H). We further confirmed using selective local mito-PAGFP photoactivation of individual mitochondria that inter-mitochondrial contacts marked by Drp1 oligomers were indeed two distinct mitochondria lacking mito-PAGFP transfer, and thus not the result of fission or fusion events (Figure 5F). Moreover, Drp1 oligomerization at mitochondrial untethering events occurred regardless of whether Drp1 oligomers were present during the initial tethering (Figure 5G).

To directly examine the role of Drp1 in inter-mitochondrial contact dynamics, we compared the effect of wild-type Drp1 and its GTP hydrolysis deficient mutant on inter-mitochondrial contact tethering. Expression of the GTP hydrolysis mutant Drp1 (K38A) significantly increased the percentage of mitochondria forming inter-mitochondrial contacts (Figure 5H, ***p<0.001), as compared to wild-type Drp1 or control conditions. Additionally, Drp1 (K38A) significantly prolonged the duration of inter-mitochondrial contacts by preventing efficient untethering (Figures 5I and 5J, ***p<0.001), suggesting that Drp1 GTP hydrolysis regulates inter-mitochondrial contact site untethering.

Of note, two distinct tethered mitochondria at inter-mitochondrial contacts labeled by matrix markers (DsRed2-mito or mito-BFP) (Figures 1K, 1L, S2E, S2F) appeared very similar to previously described mitochondrial constriction sites. Since inter-mitochondrial contact tethering occurred significantly more frequently than mitochondrial fission events (Figures 1M, S2K, S2L), the majority of previously described mitochondrial constriction sites may have been tethering of two distinct mitochondria at inter-mitochondrial contact sites followed by untethering, rather than prolonged constriction of a single mitochondria followed by fission into two daughter mitochondria, which occurs much faster (Figure S4F). Interestingly, we found that similar to what was previously proposed for mitochondrial constriction sites, Drp1 oligomers and ER tubules also both marked inter-mitochondrial contact untethering and mitochondrial fusion events (Figures 1N and 5D).

Inter-mitochondrial Contacts Functionally Restrict Mitochondrial Motility and are Modulated by Mitochondrial Respiration and Nutrient Availability

We then analyzed how inter-mitochondrial contacts might functionally regulate the mitochondrial network, as they represented significant events occurring more frequently than fission or fusion (Figures 1M, S2K, S2L). While mitochondria tethered in inter-mitochondrial contacts could move (Figure S2J), their motility was significantly decreased compared to that of free mitochondria not in inter-mitochondrial contacts (Figures 6A and 6B), suggesting that inter-mitochondrial contacts may functionally act to restrict individual mitochondrial motility. Indeed, disruption of inter-mitochondrial contact untethering by inhibiting RAB7 GTP hydrolysis via RAB7(Q67L), TBC1D15(D397A) or Fis1 (LA) mutants all resulted in significantly decreased mitochondrial motility (Figures 6C–6E).

Figure 6. Inter-mitochondrial Contacts Functionally Restrict Mitochondrial Motility and are Modulated by Mitochondrial Respiration and Nutrient Availability.

(A-B) Decreased mitochondrial motility (B, histogram in A) in inter-mitochondrial contact (30s before contact untethering) compared to free mitochondria (30s after contact untethering) (n = 30 events per condition from 15 cells).

(C-E) Decreased mitochondrial motility upon inhibition of RAB7 GTP hydrolysis by mutant RAB7(Q67L), TBC1D15 (D397A) or FIS1(LA) (OMM label mApple-TOM20, matrix label Mito-BFP) (n = 75 mitochondria from 15 cells (RAB7(WT), RAB7(Q67L), TBC1D15(D397A), FIS1(WT)), n = 60 mitochondria from 12 cells (TBC1D15(WT)), n = 85 mitochondria from 17 cells (FIS1(LA)).

(F) Quantification of distribution of mitochondrial events from confocal time-lapse images of live HeLa cells treated with mitochondrial respiration Complex I inhibitor Rotenone (3h, 1µM) (OMM label mApple-TOM20, matrix label Mito-BFP) (n = 493 events from 10 cells).

(G) Increased percentage of mitochondria forming inter-mitochondrial contacts in live HeLa cells treated with Rotenone (3h, 1µM) compared to DMSO (3h) (n = 105 events from 15 cells (DMSO), n = 105 events from 15 cells (Rotenone)).

(H) Decreased mitochondrial motility in live HeLa cells treated with Rotenone (3h, 1µM) compared to DMSO (3h) (n = 90 mitochondria from 18 cells per condition).

(I) Quantification of distribution of mitochondrial events from confocal time-lapse images of live HeLa cells cultured in low nutrient media (2h HBSS) (OMM label mApple-TOM20, matrix label Mito-BFP) (n = 297 events from 10 cells).

(J) Increased percentage of mitochondria forming inter-mitochondrial contacts in live HeLa cells cultured in low nutrient media (2h HBSS) compared to high nutrient media (Control) (n = 112 events from 16 cells (Control), n = 112 events from 16 cells (2h HBSS)).

(K) Decreased mitochondrial motility in live HeLa cells cultured in low nutrient media (2h HBSS) compared to high nutrient media (Control) (n = 90 mitochondria from 18 cells per condition).

Mean ± SEM; *p < 0.05; **p <0.01; ***p <0.001 (unpaired two-tailed t test for (B-E,G,H,J,K), ANOVA with Tukey’s post-hoc test for (F,I).

See also Figure S7.

We further examined whether inter-mitochondrial contact dynamics might respond to mitochondrial or cellular stress such as defective mitochondrial respiration or decreased nutrient availability as a way to modulate the mitochondrial network and its motility. Disruption of mitochondrial respiration with the mitochondrial respiratory chain complex I inhibitor Rotenone (1µM) or decreased nutrient availability (2h HBSS) resulted in inter-mitochondrial contact tethering/untethering rates which were still significantly higher than mitochondrial fusion/fission rates (Figures 6F and 6I). This was consistent with our observations in control cells (Figure 1M), suggesting that mitochondrial tethering is an important dynamic process that persists upon inhibition of mitochondrial respiration or decreased nutrient availability. Moreover, these conditions disrupted mitochondria-lysosome contact site dynamics (Figure S7I and S7K), significantly increased the percentage of mitochondria forming inter-mitochondrial contacts (Figure 6G and 6J), and led to a functional Mfn-dependent decrease in mitochondrial motility (Figures 6H, 6K, S7J, S7L). Thus, upregulation of inter-mitochondrial contacts in response to mitochondrial or cellular stress may contribute to restricted mitochondrial motility as a potential mechanism for preventing unnecessary energy expenditure.

Multiple Charcot-Marie-Tooth Disease Type 2 Mutants Converge on Defective Inter-mitochondrial Contact Dynamics and Mitochondrial Motility

Mutations in the mitochondrial GTPase Mitofusin2 (Mfn2) result in Charcot-Marie-Tooth Disease Type 2A (Zuchner et al., 2004), while mutations in the GTPase RAB7 lead to Charcot-Marie-Tooth Disease Type 2B (Houlden et al., 2004). However, whether these different mutants functionally converge in CMT2 disease pathogenesis is not well understood. We first compared the effect of wild-type Mfn2 and the CMT2A disease-linked Mfn2 mutation (T105M) located in its GTPase domain (Chung et al., 2006; Lawson et al., 2005) on inter-mitochondrial contact dynamics. Mutant Mfn2 (T105M) significantly prolonged the duration of inter-mitochondria contacts leading to inefficient untethering compared to wild-type Mfn2 (Figures 7A and 7B) and significantly decreased mitochondrial network motility (Figure 7C), consistent with previous studies showing mitochondrial defects (Bannerman et al., 2016; Detmer et al., 2008; El Fissi et al., 2018; Franco et al., 2016; Rocha et al., 2018).

Figure 7. Multiple Charcot-Marie-Tooth Disease Type 2 Mutants Converge on Defective Inter-mitochondrial Contact Dynamics and Mitochondrial Motility.

(A-B) CMT2A disease-linked Mfn2 mutant (T105M) prevents efficient untethering of inter-mitochondrial contacts (A, histogram in B) (n = 77 events from 11 cells per condition).

(C) CMT2A disease-linked Mfn2 mutant (T105M) disrupts mitochondrial motility (n = 90 events from 18 cells per condition).

(D) Model of RAB7 showing GTP hydrolysis mutant (Q67L; red circle) and Charcot-Marie-Tooth Type 2B (CMT2B) disease-linked mutations (green circles).

(E-G) RAB7 CMT2B disease-linked V162M mutation increases the percentage of lysosomes in mitochondria-lysosome (M-L) contacts (E) and the duration of M-L contacts (F, histogram in G). (n = 133 events from 19 cells (RAB7 (WT)), n = 119 events from 17 cells (RAB7 (V162M))).

(H-I) RAB7 CMT2B disease-linked V162M mutation prevents efficient untethering of inter-mitochondrial contacts (H, histogram in I) (n = 133 events from 19 cells (RAB7 (WT)), n = 119 events from 17 cells (RAB7 (V162M))).

(J) CMT2B disease-linked mutant RAB7 (V162M) disrupts mitochondrial motility (n = 90 events from 18 cells per condition).

(K–L) CMT2C disease-linked mutant TRPV4 (R269H) prevents efficient untethering of inter-mitochondrial contacts (K, histogram in L) (n = 50 events from 10 cells (TRPV4 (WT)), n =50 events from 10 cells (TRPV4 (R269H))).

(M) CMT2C disease-linked mutant TRPV4 (R269H) disrupts mitochondrial motility (n = 90 events from 18 cells per condition).

Mean ± SEM; *p < 0.05; ***p < 0.001; (unpaired two-tailed t test for (A, C, E, F, H, J, K, M)).

We next examined whether CMT2B mutations in RAB7 (Figure 7D) which lead to defective GTP hydrolysis and increased GTP-binding (McCray et al., 2010; Spinosa et al., 2008; Zhang et al., 2013) might similarly disrupt inter-mitochondrial contact dynamics. Using time-lapse confocal imaging, we investigated whether the most common CMT2B disease-linked RAB7 mutation (V162M) (Manganelli et al., 2012; Verhoeven et al., 2003b) might disrupt mitochondria-lysosome contacts and further misregulate mitochondrial dynamics. Indeed, we found that similar to the RAB7 (Q67L) GTP hydrolysis mutant which locks RAB7 in a GTP-bound state, RAB7 (V162M) also significantly increased the formation of stable mitochondria-lysosome contacts (lasting >10s) (Figure 7E) and increased the duration of mitochondria-lysosome contacts resulting in inefficient untethering (Figures 7F and 7G). Moreover, consistent with our data suggesting that mitochondria-lysosome contact dynamics regulate inter-mitochondrial contact untethering events, we found that RAB7 (V162M) led to prolonged inter-mitochondrial contacts unable to efficiently untether (Figures 7H and 7I) and also decreased mitochondrial motility (Figure 7J).

Finally, to test whether defective mitochondrial dynamics might be a common pathway across multiple forms of CMT2, we examined the CMT2C disease-linked TRPV4 mutation (R269H) (Auer-Grumbach et al., 2010; Deng et al., 2010; Landoure et al., 2010). Surprisingly, we found that mutant TRPV4 (R269H) also resulted in significantly prolonged inter-mitochondrial contacts which could not efficiently untether (Figure 7K and 7L), and also decreased mitochondrial motility (Figure 7M). Together, these results highlight a role for inter-mitochondrial contacts in regulating mitochondrial network motility and suggest that this pathway may be converging mechanism relevant to multiple forms of Charcot-Marie-Tooth Disease Type 2.

DISCUSSION

Our study highlights an important role for inter-mitochondrial contact formation in regulating mitochondrial network dynamics. Indeed, the majority of tethered mitochondria in the cell did not result in fusion or bulk mitochondrial matrix exchange, consistent with EM studies showing the absence of membrane fusion at sites of close mitochondrial apposition in fixed cells (Pease, 1962; Picard et al., 2015; Vernay et al., 2017). Rather, they represented dynamic inter-mitochondrial contact sites which formed under normal conditions in multiple cell types and whose dynamics were acutely regulated by multiple GTPases. Inter-mitochondrial contact formation was further modulated by mitochondrial respiratory activity and nutrient availability, suggesting that contact dynamics may be regulated to counter mitochondrial and cellular dyshomeostasis (Ichas et al., 1997; Picard et al., 2015; Rambold et al., 2011; Santo-Domingo et al., 2013; Vernay et al., 2017). Moreover, mitochondria in inter-mitochondrial contacts had significantly restricted motility, potentially allowing for contact formation to further modulate the dynamics of the global mitochondrial network.

Inter-mitochondrial contact dynamics were further coupled to inter-organelle contact sites. Interestingly, ER tubules marked inter-mitochondrial contact untethering events, in addition to their previously reported localization at both mitochondrial fission and fusion events (Friedman et al., 2011; Guo et al., 2018) suggesting their ubiquitous localization at multiple mitochondrial events. In addition, while studies labeling mitochondria with matrix markers previously described mitochondrial constriction events prior to fission as marked by both ER tubules and Drp1 oligomers (Chakrabarti et al., 2018; Cho et al., 2017; Friedman et al., 2011), some of these events may have been two distinct mitochondria tethered at inter-mitochondrial contact sites prior to an untethering event which appear similar when visualized with mitochondrial matrix markers, and are also marked by ER tubules and Drp1 oligomers. In contrast, lysosome contacts preferentially marked inter-mitochondrial untethering events over mitochondrial fusion events, and actively promoted untethering via lysosomal RAB7 GTP hydrolysis at mitochondria-lysosome contact sites (Wong et al., 2018) mediated by the mitochondrial Rab7-GAP (TBC1D15) recruited to mitochondria by Fis1 (Onoue et al., 2013; Peralta et al., 2010; Yamano et al., 2014; Zhang et al., 2005). Inter-mitochondrial contact formation and subsequent untethering were further regulated by Mfn1/2 and Drp1 GTP hydrolysis, consistent with early studies showing their effects on mitochondrial clustering and aggregation (Santel and Fuller, 2001; Smirnova et al., 2001; Smirnova et al., 1998). Thus, multiple lysosomal and mitochondrial GTPases together converge to regulate inter-mitochondrial contact dynamics.

As multiple human diseases are linked to defective mitochondrial dynamics (Alexander et al., 2000; Burte et al., 2015; Delettre et al., 2000; Zuchner et al., 2004), further elucidating the dynamic inter-organelle interactions shaping mitochondrial networks is critical for understanding disease pathogenesis. Surprisingly, we found that multiple forms of disease-linked Charcot-Marie-Tooth Type 2 including Mfn2 (CMT2A), RAB7 (CMT2B) and TRPV4 (CMT2C) converged on inefficient inter-mitochondrial contact untethering dynamics and defective mitochondrial motility. Of note, mitochondrial motility was also previously shown to be impaired by CMT2-disease linked mutations in Mfn2 (Baloh et al., 2007; Misko et al., 2010; Misko et al., 2012; Rocha et al., 2018) and recently by mutant Rab7 (Cioni et al., 2019). Our study thus points to a potentially important role for this pathway in CMT Type 2, consistent with a growing list of CMT genes implicated in regulating mitochondrial dynamics (Korobova et al., 2013; Lee et al., 2016; Wong et al., 2018; Zuchner et al., 2004). Ultimately, understanding the pathways involved in regulating mitochondrial networks will provide insight into both cellular metabolism and pathogenic mitochondrial dysfunction in disease.

STAR METHODS

CONTACT FOR REAGENT AND RESOURCE SHARING

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Dimitri Krainc (dkrainc@nm.org).

EXPERIMENTAL MODEL DETAILS

Cell lines

HeLa cells (gift from Michael Schwake (ATCC)) and HEK293 cells (human embryonic kidney cell line 293FT (Life Technologies)) were cultured in DMEM (Gibco; 11995–065) supplemented with 10% (vol/vol) FBS, 100 units/mL penicillin, and 100 µg/mL streptomycin. H4 neuroglioma cells (Mazzulli et al., 2011) were cultured in Optimem + 5% FBS, 200ug/ml geneticin and hygromycin, 1% penicillin / streptomycin (Life Technologies) and treated with 1ug/ml doxycycline (Sigma) for 3 days. Cells were derived from female (HeLa, HEK293) or male (H4) subjects. All cells were maintained at 37 °C in a 5% CO2 incubator and previously verified by cytochrome c oxidase subunit I (COI) and short tandem repeat (STR) testing and tested for Mycoplasma contamination. Cells were transfected using Lipofectamine 2000 (Invitrogen). For live imaging, cells were grown on glass bottom culture dishes (MatTek; P35G-1.5-14-C).

METHOD DETAILS

Plasmids

The following plasmids were obtained from Addgene: LAMP1-mGFP was a gift from Esteban Dell’Angelica (Addgene #34831) (Falcon-Perez et al., 2005), mito-BFP and mCh-Drp1 were gifts from Gia Voeltz (Addgene #49151, #49152) (Friedman et al., 2011), Snap-Omp25 was a gift from David Sabatini (Addgene # 69599) (Katajisto et al., 2015), mito-PAGFP was a gift from Richard Youle (Addgene # 23348) (Karbowski et al., 2004), EGFP-RAB7A WT was a gift from Qing Zhong (Addgene #28047) (Sun et al., 2010), mApple-TOMM20-N-10, mEmerald-TOMM20-C-10, DsRed2-Mito-7, and mCherry-ER-3 were gifts from Michael Davidson (Addgene #54955, #54281, #55838, #55041), and Mfn1-Myc, Mfn1(K88T)-10xmyc, Mfn2-myc and Mfn2 K109A-myc were gifts from David Chan (Addgene #23212, #26050, #23213, #26051) (Chen et al., 2003). GFP-RAB7-Q67L was a gift from Aimee Edinger (Romero Rosales et al., 2009). N-terminal HA-tagged TBC1D15 (wild-type and D397A mutant) and Flag-FIS1 (wild-type and LA mutant) were generous gifts from Naotada Ishihara (Jofuku et al., 2005; Onoue et al., 2013), mCherry-Mfn1 and mCherry-Mfn2 were generous gifts from Elena Kolobova (Mason et al., 2014), and pIRES2-ZsGreen1- TRPV4 (WT) and pIRES2-ZsGreen1- TRPV4 (R269H) were generous gifts from Han-Xiang Deng (Deng et al., 2010). mCherry-Drp1(K38A), bicistronic Mfn2(WT)+mApple, bicistronic Mfn2(T105M)+mApple, mNeonGreen-RAB7(WT) and mNeonGreen-RAB7(V162M) were generated using VectorBuilder.

Cell treatment

To modulate mitochondrial and cellular homeostasis, HeLa cells were incubated with control DMSO (3h, Sigma), mitochondrial respiration Complex I inhibitor Rotenone (3h, 1µM, Sigma) or cultured in low nutrient media Hank’s Balanced Salt Solutions (2h HBSS, Corning) prior to confocal live-cell imaging experiments.

Confocal microscopy

All confocal images were acquired on a Nikon A1R laser scanning confocal microscope with GaAsp detectors using a Plan Apo λ 100× 1.45 NA oil immersion objective (Nikon) using NIS-Elements 4.20 (Nikon). Live cells were imaged in a temperature-controlled chamber (37 °C) at 5% CO2 at 1 frame every 2–3 sec. Cells transfected with Snap-OMP25 were visualized by incubation for 30 min with 0.6 µM SNAP-Cell® 647-SiR (New England Biolabs; S9102S) and subsequently washed 4x with warm media before imaging. The matrix of individual mitochondria were selectively labeled by localized photoactivation of cells transfected with photoactivatable mitochondrial matrix marker mito-PAGFP using a 405 nm laser (100% for 4 sec) and mitochondrial matrix transfer at fusion or transient fusion events was subsequently tracked.

Electron microscopy

For transmission electron microscopy (TEM), cells were grown on coverslips and fixed in a mixture of 2.5% glutaraldehyde and 2% paraformaldehyde in 0.1M cacodylate buffer for 2–24 h at 4 °C. After post-fixation in 1% osmium tetroxide and 3% uranyl acetate, cells were dehydrated in series of ethanol, embedded in Epon resin and polymerized for 48 h at 60 °C. Ultrathin sections were made using UCT ultramicrotome (Leica Microsystems) and contrasted with 4% uranyl acetate and Reynolds’s lead citrate. Samples were imaged using a FEI Tecnai Spirit G2 transmission electron microscope (FEI Company, Hillsboro, OR) operated at 80 kV. Images were captured with an Eagle 4k HR 200kV CCD camera.

Structured illumination microscopy

Structured illumination microscopy (SIM) super-resolution images were taken on a Nikon N-SIM system with an oil immersion objective lens 100×, 1.49 NA, Nikon. Images were captured using NIS-Elements (Nikon) at 1 frame every 6 sec and reconstructed using slice reconstruction in NIS-Elements (Nikon). Images for live cell imaging (live N-SIM) were taken at a single z-plane. Cells used for live cell imaging were maintained in a temperature-controlled chamber (37 °C) at 5% CO2 in a TokaiHit stagetop incubator.

QUANTIFICATION AND STATISTICAL ANALYSIS

Image analysis

Inter-mitochondrial contacts were defined as those which clearly showed two distinct mitochondria which came together to form a tether and subsequently untethered, and remained as two distinct mitochondria throughout the process without fusion. Mitochondrial fusion events were defined as those which clearly began as two distinct mitochondria and subsequently tethered together prior to outer mitochondrial membrane fusion to form a single mitochondria. Mitochondrial fission events were defined as those which clearly showed division of a single mitochondria into two distinct daughter mitochondria which transiently tethered prior to subsequently separating from one another. The percentage of mitochondria in inter-mitochondrial contacts at a given point in time were defined as those which were tethered (<0.1 µm) to another mitochondria for >10s. The frequency of duration of inter-mitochondrial contact tethering was calculated from contacts which showed both clear formation of tether and subsequent untethering. Mitochondrial events were defined as tether (formation of inter-mitochondrial contacts not leading to fusion), untether (untethering of inter-mitochondrial contacts), fusion and fission as analyzed from confocal time-lapse of outer mitochondrial membrane markers. Mitochondrial transient fusion events were defined as events involving transfer of photoactivated mitochondrial matrix marker mito-PAGFP from individual mitochondria selectively labeled by localized photoactivation, without the obvious fusion of outer mitochondrial membrane marker (mApple-TOM20). Tethered mitochondrial partners were tracked over time (10s) for their inter-mitochondrial contact site using ImageJ 1.51j8 (NIH) from time-lapse confocal images of outer mitochondrial membrane marker (mApple-TOM20), with the starting location for all contact sites set to the center of the graph (Figure S2J). For motility analysis, mitochondria which were tethered in an inter-mitochondrial contact (distance traveled over 30 sec prior to untethering) were compared to free mitochondria (distance traveled over 30 sec after untethering from an inter-mitochondrial contact).

Mitochondrial untethering, fusion and fission events marked by Drp1 or ER were analyzed from confocal time-lapse images of outer mitochondrial membrane marker (TOM20-mEmerald), matrix marker (mito-BFP) and Drp1 marker (mCherry-Drp1) or ER marker (mCherry-ER) (Figure 1N, 5D, 5E, S4G and S7H). Mitochondrial untethering events and tethered mitochondria (the state of being tethered) were further analyzed for the presence of ER or Drp1 and confirmed to be distinct from fission/fusion/transient fusion events by confocal time-lapse images of individually labeled locally photoactivated mitochondria (mito-PAGFP) with outer mitochondrial membrane marker (Snap-OMP25) (Figure 1N (tethered), 5B, 5C, 5G, S3B, S3F, and S3G). The expected probability that Drp1 oligomers would be at the site of a mitochondrial division event by random chance was calculated as the density of mCherry-Drp1 on the outer mitochondrial membrane (mApple-TOM20) from n = 19 living cells, using ImageJ 1.51j8 (NIH). The expected probability that ER would be at the site of a mitochondrial division event by random chance was calculated as the density of mCherry-ER in the cytosol from n = 11 living cells, using ImageJ 1.51j8 (NIH). Analysis for Figure 5H–5J was conducted from confocal time-lapse images of outer mitochondrial membrane marker (mEmerald-TOM20), matrix marker (mito-BFP) and mCherry-Drp1 (WT or K38A) and subsequently normalized to Control. The duration for different mitochondrial events (Figure S4F) were defined for fusion (time from formation of mitochondrial tether to fusion into a single mitochondria), untether (time from formation of mitochondrial tether to untethering as two distinct mitochondria throughout the process), and fission (time from the formation of transient tethering of two daughter mitochondria to their separation). The minimum time to untethering (Figure S3G) or transient fusion (Figure S4E and S7G) were quantified from mitochondrial tethers already formed at the beginning of the video.

Mitochondrial untethering, fusion and fission events and tethered state marked by lysosomes/late endosomes contacts (M-L) were analyzed from confocal time-lapse images of outer mitochondrial membrane marker (mApple-TOM20) and lysosome/late endosome marker (LAMP1-mGFP) (Figure 2). The expected probability that a LAMP1 vesicle would be at the site of a mitochondrial division event by random chance was calculated as the density of LAMP1-mGFP vesicles in the cytosol from n = 16 living cells, using ImageJ 1.51j8 (NIH). The fate of inter-mitochondrial contacts (either remaining tethered or untethering in ≤10s) was analyzed for tethered mitochondria which did not simultaneously contact lysosomes compared to those that did. Inter-mitochondrial (M-M) contact and M-L contact formation were defined as the time when two distinct organelles clearly came together to form a tether, while M-M and M-L contact untethering were defined as the time when two distinct organelles clearly separated. M-L contacts were only analyzed for LAMP1-mGFP-positive vesicles which were clearly in contact (<0.1 µm) with a mitochondria that was simultaneously in a M-M contact. M-L contact formation was quantified for any events occurring either prior or subsequent to M-M formation but prior to M-M untethering. M-L contact untethering was quantified for any events occurring after M-M contact formation but either prior or subsequent to M-M untethering. The minimum time to M-M untether (duration of M-M contacts) (Figure 3) was analyzed from mitochondria already tethered at the beginning of the video, and was defined as the time prior to mitochondrial untethering over a 150s video. Any contacts which lasted throughout the entire 150s video and which were still in contact by the end of the video were categorized as 150s in bar graphs, and as > 120s in histograms.

Inter-mitochondrial contacts marked by mitofusins were analyzed from confocal time-lapse images of outer mitochondrial membrane marker (mEmerald-TOM20), matrix marker (mito-BFP) and mitofusin markers (mCherry-Mfn1, mCherry-Mfn2). The percentage of inter-mitochondrial (M-M) contacts was quantified as the percentage of mitochondria tethered to another mitochondria for >10s divided by the total number of mitochondria in the region of interest, and the minimum duration of M-M contacts was analyzed from mitochondria already tethered at the beginning of the video, and defined as the time prior to mitochondrial untethering over a 180s video, with any contacts which lasted throughout the entire 180s video and which were still in contact by the end of the video categorized as 180s and as > 120s in histograms. Analysis for Figure 4E–4J was conducted from confocal time-lapse images of outer mitochondrial membrane marker (TOM20-mApple) and matrix marker (mito-BFP) with or without myc-tagged mitofusins. All analysis was subsequently normalized to Control (Figure 4E–4F, 4H–4I).

Analysis for Figure 7A–7C was conducted from confocal time-lapse images of outer mitochondrial membrane marker (mEmerald-TOM20), matrix marker (mito-BFP) and bicistronic Mfn2 WT or V162M in cells simultaneously expressing mApple, and was subsequently normalized to Mfn2 (WT). For comparison of mNeonGreen-RAB7 WT with V162M using outer mitochondrial membrane marker (mApple-TOM20) (Figure 7E–7J), the percentage of lysosomes in M-L contact was quantified as the percentage of RAB7 vesicles that formed contacts lasting >10s with mitochondria divided by the total number of RAB7 vesicles in the region of interest, the rate of mitochondrial fission was analyzed from videos >180s, and the minimum duration of both M-L and M-M contacts were analyzed from mitochondria already tethered at the beginning of the video, and defined as the time prior to mitochondrial untethering over a 180s video, with any contacts which lasted throughout the entire 180s video and which were still in contact by the end of the video categorized as 180s and as > 120s in histograms, and all analysis was subsequently normalized to RAB7 WT. Analysis for Figure 7K–7L was conducted from confocal time-lapse images of outer mitochondrial membrane marker (mApple-TOM20) and the minimum duration of M-M contacts were analyzed from mitochondria already tethered at the beginning of the video, and defined as the time prior to mitochondrial untethering over a 180s video, with any contacts which lasted throughout the entire 180s video and which were still in contact by the end of the video categorized as 180s and as > 120s in histograms, and was subsequently normalized to TRPV4 (WT). For motility analysis in Fig. 6, 7 and S7, the motility of 5 randomly selected mitochondria per cell was tracked over 3 min (only for mitochondria which remained in the field of view for the entire 3 min) and normalized to WT/control/DMSO conditions.

Statistical analysis, graphing and figure assembly

Data were analyzed using unpaired two-tailed Student t test (for two datasets), Fisher’s exact test or one-way ANOVA with Tukey’s post hoc test (for multiple data sets) (see figure legends for details). Data presented are means ± SEM (except in histograms). Statistics and graphing were performed using Prism 7 (GraphPad) software. All experiments were analyzed from n ≥3 independent experiments (biological replicates) per condition. All line scans were generated using ImageJ 1.51j8 (NIH) and normalized per protein. Examples of transient fusion and M-L contacts are shown after linear rescaling in NIS-Elements 4.20 (Nikon) and all videos and images were assembled using ImageJ 1.51j8 (NIH). All final figures were assembled in Illustrator CC (Adobe).

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Chemicals, Peptides, and Recombinant Proteins
Lipofectamine 2000	Invitrogen	Cat#11668019
SNAP-Cell® 647-SiR	New England Biolabs	Cat#S9102S
Rotenone	Sigma	Cat#R8875–1G
HBSS	Corning	Cat#21020CV
Experimental Models: Cell Lines
HeLa cells	ATCC	CCL-2
HEK293FT cells	Life Technologies	R70007
H4 cells	Mazzulli et al., 2011	N/A
Recombinant DNA
LAMP1-mGFP	Falcon-Perez et al., 2005	Addgene Plasmid #34831
mito-BFP	Friedman et al., 2011	Addgene Plasmid #49151
mCh-Drp1	Friedman et al., 2011	Addgene Plasmid #49152
Snap-Omp25	Katajisto et al., 2015	Addgene Plasmid #69599
mito-PAGFP	Karbowski et al., 2004	Addgene plasmid #23348
EGFP-Rab7A WT	Sun et al., 2010	Addgene Plasmid #28047
GFP-Rab7-Q67L	Romero Rosales et al., 2009	N/A
mApple-TOMM20-N-10	Michael Davidson	Addgene Plasmids #54955
mEmerald-TOMM20-C-10	Michael Davidson	Addgene Plasmids #54281
DsRed2-Mito-7	Michael Davidson	Addgene Plasmids #55838
mCherry-ER-3	Michael Davidson	Addgene Plasmids #55041
Mfn1-Myc	Chen et al., 2003	Addgene Plasmids #23212
Mfn1(K88T)-10xmyc	David Chan	Addgene Plasmids #26050
Mfn2-myc	Chen et al., 2003	Addgene Plasmids #23213
Mfn2 K109A-16xmyc	David Chan	Addgene Plasmids #26051
HA-TBC1D15 WT	Onoue et al., 2013	N/A
HA-TBC1D15 D397A	Onoue et al., 2013	N/A
Flag-Fis1 WT	Onoue et al., 2013; Jofuku et al., 2005	N/A
Flag-Fis1 LA	Onoue et al., 2013; Jofuku et al., 2005	N/A
mCherry-Mfn1	Mason et al., 2014	N/A
mCherry-Mfn2	Mason et al., 2014	N/A
pIRES2-ZsGreen1- TRPV4 (WT)	Deng et al., 2010	N/A
pIRES2-ZsGreen1- TRPV4 (R269H)	Deng et al., 2010	N/A
mCherry-Drp1(K38A)	VectorBuilder	N/A
Mfn2(WT)+mApple	VectorBuilder	N/A
Mfn2(T105M)+mApple	VectorBuilder	N/A
mNeonGreen-Rab7(WT)	VectorBuilder	N/A
mNeonGreen-Rab7(V162M)	VectorBuilder	N/A
Software and Algorithms
NIS-Elements	Nikon	https://www.nikoninstruments.com/Products/Software
ImageJ 1.51j8	NIH, USA	https://imagej.nih.gov/ij/
GraphPad Prism 7	GraphPad Software	https://www.graphpad.com/
Adobe Illustrator CC	Adobe	https://www.adobe.com/creativecloud/plans.html?single_app=illustrator
Other
Glass bottom culture dishes	MatTek	P35G-1.5–14-C
Laser scanning confocal microscope	Nikon	A1R
Objective Lens, 100x (for laser scanning confocal microscope)	Nikon	Plan Apo λ 100× 1.45 NA oil immersion objective
Structured Illumination Super-resolution Microscope (SIM)	Nikon	N-SIM
Objective Lens, 100x (for SIM)	Nikon	100× 1.49 NA oil immersion objective
Transmission electron microscope (TEM)	FEI Company	FEI Tecnai Spirit G2 transmission electron microscope
Camera (for TEM)	FEI Company	Eagle 4k HR 200kV CCD camera

Highlights:

• Inter-mitochondrial contacts dynamically form and restrict mitochondrial motility

• Lysosomes regulate inter-mitochondrial contact untethering via Rab7 GTP hydrolysis

• The ER marks fission, fusion and inter-mitochondrial contact untethering events

• Multiple CMT2 mutants (Mfn2/RAB7/TRPV4) converge on impaired mitochondrial dynamics