(C) PLOS One
This story was originally published by PLOS One and is unaltered.
. . . . . . . . . .



Optogenetic cleavage of the Miro GTPase reveals the direct consequences of real-time loss of function in Drosophila [1]

['Francesca Mattedi', 'Department Of Basic', 'Clinical Neurosciences', 'Maurice Wohl Clinical Neuroscience Institute', 'Institute Of Psychiatry', 'Psychology', 'Neuroscience', 'King S College London', 'London', 'United Kingdom']

Date: 2023-08

Miro GTPases control mitochondrial morphology, calcium homeostasis, and regulate mitochondrial distribution by mediating their attachment to the kinesin and dynein motor complex. It is not clear, however, how Miro proteins spatially and temporally integrate their function as acute disruption of protein function has not been performed. To address this issue, we have developed an optogenetic loss of function “Split-Miro” allele for precise control of Miro-dependent mitochondrial functions in Drosophila. Rapid optogenetic cleavage of Split-Miro leads to a striking rearrangement of the mitochondrial network, which is mediated by mitochondrial interaction with the microtubules. Unexpectedly, this treatment did not impact the ability of mitochondria to buffer calcium or their association with the endoplasmic reticulum. While Split-Miro overexpression is sufficient to augment mitochondrial motility, sustained photocleavage shows that Split-Miro is surprisingly dispensable to maintain elevated mitochondrial processivity. In adult fly neurons in vivo, Split-Miro photocleavage affects both mitochondrial trafficking and neuronal activity. Furthermore, functional replacement of endogenous Miro with Split-Miro identifies its essential role in the regulation of locomotor activity in adult flies, demonstrating the feasibility of tuning animal behaviour by real-time loss of protein function.

Funding: This work was supported by a NC3Rs David Sainsbury fellowship (NC/N001753/2) and NC3Rs SKT grant (NC/T001224/1), an Academy of Medical Sciences Springboard Award (SBF004/1088), an ARUK King’s College London Network Centre Grant (ARUK-NC2020-KCL), a van Geest Fellowship in Dementia and Neurodegeneration, and a van Geest Studentship to A.V. A MRC-DTP Studentship supports E.L.M. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Here, we present the generation of “Split-Miro,” a photocleavable variant of Drosophila Miro, to achieve rapid and controlled loss of protein function. Mitochondrial network remodelling is a rapid response to Split-Miro photocleavage in Drosophila S2R+ cells. This effect, mediated by loss of anchorage to the microtubule network, impacts on mitochondrial distribution into cell processes where the use of Split-Miro shows that this protein is sufficient to increase mitochondrial motility but dispensable for the maintenance of elevated mitochondrial velocities. This effect is mirrored in adult fly neurons in vivo, where Split-Miro affects the motility, but not the processivity, of the mitochondria. Unexpectedly, we show that Split-Miro photocleavage neither directly impacts mitochondrial calcium homeostasis nor the association between the mitochondria and the endoplasmic reticulum in S2R+ cells. Finally, we demonstrate that Split-Miro modulates neuronal activity in adult flies, and, by rescuing the lethality associated with classical Miro LoF mutations via pan-neuronal expression, we provide proof of concept that Split-Miro affords control of fly locomotor activity through exposure to blue light.

This issue is especially problematic when studying complex cellular processes such as mitochondrial dynamics, and it is particularly evident for Miro proteins, mitochondrial Rho GTPases that influence the motility, morphology, and physiology of mitochondria [ 1 – 3 ]. In Drosophila, homozygous mutants of miro are developmentally lethal [ 4 ], while knockout of Miro1 in mice leads to perinatal lethality [ 5 , 6 ]. Chronic loss of Miro is detrimental for mitochondrial transport in Drosophila and mammalian neurons [ 7 – 9 ], where it leads to alteration of synaptic strength [ 4 , 10 ], whereas disruption of Miro in Drosophila additionally impairs mitochondrial calcium homeostasis [ 11 , 12 ]. Reducing Miro abundance has also profound effects on mitochondrial morphology with fragmented mitochondria observed in yeast [ 13 ], Drosophila larval motor neurons [ 7 ], and mouse embryonic fibroblasts [ 6 ]. It is not yet clear, however, to what extent these phenotypes are directly consequences of Miro disruption and whether they might arise independently. Any attempt to dissect the causality of cellular and organismal phenotypes after Miro manipulation is challenging as the current tools do not allow acute disruption of function.

Methods to observe loss of function (LoF) phenotypes are used to study many biological processes. Although important tools for elucidating gene function, disruption of genes by genomic mutations or RNA interference (RNAi) often does not have the spatiotemporal resolution to capture the direct cellular and organismal consequences of LoF and to report specifically on a protein’s primary function. The observed “end-point” phenotypes might thus be the result of compensatory mechanisms to loss of protein, and gene pleiotropy means that, even in cell culture models, it is often difficult to dissect the causality of the observed phenotypes.

Strikingly, 8-day-old miro Sd32/B682 , UAS-Split-Miro flies displayed a rapid and sustained increase in motor activity when exposed to blue light compared to control, while no difference was observed in their baseline activity prior to exposure to blue light ( Fig 7F–1H ). While the fly speed returned to control level as the blue light was turned off, the action initiation revealed a prolonged effect on locomotor activity, similar to what observed in younger flies ( Fig 7F–1H ). Remarkably, Split-Miro–dependent effects on fly activity and action initiation could be reversed by coexpressing human SNPH (UAS-SNPH) with Appl-Gal4 ( Fig 7I–1K ). Thus, SNPH can also suppress the effect of Miro inactivation in the context of animal behaviour. Together, these results show that acute loss of Miro function leads to fly hyperactivity, and this phenotype is exacerbated by ageing and suppressed by expression of SNPH. Collectively, these experiments reveal that Split-Miro is effective in an adult animal and uncover a previously unknown role of Miro in the regulation of Drosophila locomotor behaviour.

(A) Adult miro Sd32/B682 flies eclosed after Appl-Gal4-driven expression of UAS-wt-Miro, UAS-Split-Miro or UAS-Split-Miro + UAS-EGFP-SNPH (UAS-SNPH) were counted from 6, 5, and 3 independent crosses, respectively. Data are reported as percentage of flies expected from mendelian ratios. Data are mean ± SEM, one-way ANOVA with Tukey’s post hoc test showed no statistical difference. (B) Schematic of the “Opto-DART” behavioural setup consisting of a custom-made optogenetic enclosure equipped with LEDs for blue light stimulation and a camera to record fly activity. Flies are transferred to a 1D platform for automated recording of activity using the DART system [ 45 , 46 ]. DAC: Digital-to-Analog Converter for multiplatform integration. (C-H) Overall activity, action initiation, and average speed of 2-day-old flies (C-E) and 8-day-old flies (F-H) expressing UAS-wt-Miro or UAS-Split-Miro, before, during, and after blue light exposure (shaded blue rectangles). In (C) and (F), the average activity of flies expressing UAS-wt-Miro vs. UAS-Split-Miro is: before light exposure, 3.78% ± 0.94 vs. 2.34% ± 0.72 (p = 0.11) (C) and 5.62% ± 1.02 vs. 8.7% ±1.81 (p = 0.14) (F); under blue light, 9.52% ± 0.96 vs. 7.32% ± 0.81 (p = 0.13) (C) and 8.45% ± 0.82 vs. 21.87% ± 2.07**** (F); after blue light exposure, 3.36% ± 1.18 vs. 7.85% ± 1.96 (p = 0.13) (C) and 3.23% ± 1.1 vs. 13.19% ± 2.39** (F). (I, K) Overall activity, action initiation, and average speed of 8-day-old flies expressing UAS-Split-Miro or UAS-Split-Miro + UAS-SNPH, before, during, and after blue light exposure (shaded blue rectangles). In (I), the overall activity of flies expressing UAS-Split-Miro vs. UAS-Split-Miro + UAS-SNPH is: 6.09% ± 0.89 vs. 0.95% ± 0.35**** before light exposure; 18.62% ± 1.4 vs. 3.55% ± 0.75**** under blue light; 9.9% ± 1.4 vs. 3.91% ± 1.25* after blue light exposure. In (C-K), n = number of flies. Values are means ± SEM, Mann–Whitney test (C, F, I) and multiple unpaired t test (D, E, G, H, J, K) with Holm–Sidak correction for multiple comparisons. * p < 0.05, ** p < 0.01, **** p < 0.0001. The data underlying the graphs shown in the figures can be found in S1 Data .

Having demonstrated that Split-Miro affects mitochondrial functionality in S2R+ cells and neuronal physiology in vivo, we turned to behavioural genetics to assess Split-Miro versatility for the study of organismal phenotypes. Pan-neuronal expression of UAS-mCherry-MiroN-LOV2 and UAS-Zdk1-MiroC (“UAS-Split-Miro”) with the Appl-Gal4 driver rescued the lethality associated with classic null mutations of the Miro gene (miro Sd32/B682 ) (Figs 7A and S5A ). Adult flies were assayed for their motor behaviour at day 2 and day 8 before, during, and after blue light exposure using the “Opto-DART” system ( Fig 7B ). miro Sd32/B682 , UAS-Split-Miro flies at 2 days of age showed a modest increase in locomotor activity after an hour of exposure to blue light, which was sustained also after the exposure to light ceased ( Fig 7C–7E ). These observations suggest that the effects of Split-Mito photocleavage were not rapidly reversed after removal of blue light.

Genetic mutations and RNAi have shown that Miro is critical for mitochondrial functionality in the nervous system [ 5 , 7 , 8 ]. Homozygous miro gene loss of function alleles are lethal, thus precluding a comprehensive analysis of Miro function in adult animals. Conditional loss of Miro1 in mouse neurons causes severe movement defects within 30 days postnatal [ 5 ], and both increased and reduced Miro abundance in the Drosophila nervous system can rescue fly climbing activity in models of neurodegeneration [ 43 , 44 ]. These findings suggest that real-time Miro disruption in the adult nervous system could be exploited to manipulate animal behaviour.

(A, B) Examples of spontaneous bouts of neuronal activity, indicated by transient fluorescence increase of the GCaMP5 Ca 2+ indicator in the cell body (A) and axonal bundle (B) of adult wing neurons. White circle, neuronal soma (A). In (B), the arrows indicate a single axon within the axonal bundle. Scale bars: 5 μm. (C) Pan-neuronal expression of UAS-Split-Miro significantly reduces the amplitude of Ca 2+ transients, compared to UAS-wt-Miro. Traces indicate the average GCaMP5 fluorescence intensity at individual time points in neuronal axons. N = number of bouts of activity from 16 wings (UAS-wt-Miro = 8; UAS-Split-Miro = 8) and 2 independent experiments. (D, E) Peak of GCaMP fluorescence expressed as fluorescence fold increase from baseline (D) and time to reach the peak (E). Circles, combined bouts of activity from neuronal soma and axons, relative to (A-C). Data are mean ± SEM, Mann Whitney test. (F) Quantification of the number of active neuronal regions after 5 minutes of time-lapse imaging with 488-nm blue light in UAS-wt-Miro and UAS-Split-Miro flies. The number of neuronal regions (cell bodies and axons) that respond were counted in 16 (UAS-wt-Miro) and 15 (UAS-Split-Miro) fields of view, Fisher’s exact test. * p < 0.05, *** p < 0.001, **** p < 0.0001. Refer to S3 Table for full genotypes. The data underlying the graphs shown in the figures can be found in S1 Data .

Having observed a strong reduction in mitochondrial trafficking in 2-day-old Split-Miro flies, we next asked whether rapid Split-Miro photocleavage could affect wider neuronal physiology. Reduced mitochondrial motility has been associated with alterations in neuronal activity [ 4 , 10 , 39 , 40 ]. Thus, we decided to measure the activity of adult wing neurons by recording spontaneous Ca 2+ transient with the GCaMP5 Ca 2+ indicator ( Fig 6A and 6B ). As done for the trafficking experiment, we expressed Split-Miro and the wt-Miro control in a miro +/− background and recorded basal neuronal responses under exposure to blue laser light. We found that UAS-Split-Miro + neurons displayed a milder and slower response compared to UAS-wt-Miro + control neurons ( Fig 6C–6E ). Interestingly, the dampened neuronal firing was offset by an enlarged area of active neurons ( Fig 6F ), suggesting that Split-Miro photocleavage induced distributed network performance, which characterises neural ensembles and manifolds [ 41 , 42 ].

(A, B) Stills from movies of GFP-labelled mitochondria in wing neuronal axons expressing UAS-wt-Miro (A) or UAS-Split-Miro (B) in miro Sd32/+ background during the first minute (top panels) and fifth minute (bottom panels) of blue light exposure. Traces of transported mitochondria in corresponding movies are overlayed onto the images. (C, D) Number of motile mitochondria captured in a 50-μm axonal tract in wing neurons expressing UAS-wt-Miro (C) or UAS-Split-Miro (D). Bar charts show the average mitochondrial content at each time point. Filled circles represent the number of mitochondria within each axonal bundle at minute 1 and 5. Data were analysed by paired Student t test. Number of wings analysed: UAS-wt-Miro = 8, UAS-Split-Miro = 8, from 2 independent experiments. (E, F) Anterograde velocity (E) and run length (F) of axonal mitochondria in wing neurons expressing UAS-wt-Miro or UAS-Split-Miro, during the first and fifth minute of blue light exposure, relative to (C, D). Due to the overall lower number of bidirectional and retrograde-moving mitochondria, a meaningful statistical analysis of their velocity and run length is not possible. Circles represent tracked mitochondria. Data mean ± SEM, Mann–Whitney test. *** p < 0.001. The data underlying the graphs shown in the figures can be found in S1 Data .

It has been shown that a subset of Miro is found at the mitochondria-ER interface to regulate the contacts between these 2 organelles [ 12 , 32 , 33 ], and the mitochondria-ER contacts sites (MERCS) are known to mediate Ca 2+ exchange between the 2 organelles [ 34 ]. However, we did not detect any significant difference in the number of MERCS visualised using a split-GFP assay [ 35 , 36 ] after Split-Miro photocleavage ( Fig 4F–4H ). Furthermore, the morphological changes displayed by mitochondria after Split-Miro photocleavage were not associated with detectable changes in their membrane potential ( S4 Fig ), implying that mitochondria do not become dysfunctional during this rapid morphological transition. Together, these findings suggest that acutely modulating mitochondrial network integrity via Split-Miro/SNPH or blocking mitochondrial motility via SNPH ( S3F and S3G Fig ) are not sufficient to perturb [Ca 2+ ] m homeostasis in this context.

(A) Representative image of an S2R+ cell showing mitochondrial targeting of mito-GCaMP6f (cyan) in the perinuclear region (white arrows) and in single mitochondria (arrowheads). Yellow is mCherry-tagged wt-Miro (wt-Miro); magenta is MitoTracker DeepRed (MTDR). Scale bar: 5 μm. (B) Cells transfected with mito-GCaMP6f and either mCherry-wt-Miro (wt-Miro) or mCherry-Split-Miro (Split-Miro) were stained with MTDR; and the ratio of mito-GCaMP6f/MTDR signal intensity was analysed at the beginning (first minute) and at the end (seventh minute) of the time-lapse imaging under blue light. Number of cells: wt-Miro = 12, Split-Miro = 16; from 3 independent experiments. Data are shown as mean ± SEM. Repeated measures one-way ANOVA followed by Tukey’s post hoc test did not show any significant difference between groups, indicating that the basal [Ca 2+ ] m does not significantly change after Split-Miro photocleavage. (C) Expression of mCherry-Split-Miro (Split-Miro) or EBFP-SNPH (SNPH) does not significantly alter [Ca 2+ ] m uptake in cells challenged with ionomycin, compared to control conditions. Traces indicate the average mito-GCaMP6f fluorescence intensity values (circles) at individual time point before and after cell exposure to ionomycin (arrow). N = number of cells, from 5 independent experiments. (D, E) Normalised response peak and time to reach the peak, respectively, relative to the data shown in (C). Circles, number of cells. Kruskal–Wallis test followed by Dunn’s multiple comparisons showed no difference between conditions. (F) wt-Miro or Split-Miro (magenta) were coexpressed with the ER-mito::SPLICS probe (cyan) in S2R+ cells. The SPLICS probe displays a typical punctuated stain in these cells, as previously observed in mammalian and Drosophila cells [ 34 , 36 ]. (G, H) Quantification of the SPLICS puncta and statistical analysis by Wilcoxon test showed no significant difference between the beginning (first minute) and the end (seventh minute) of the time-lapse imaging under blue light in wild-type and Split-Miro–transfected cells. Data are presented as % change of SPLICS puncta at seventh minute compared to first minute. Number of contacts analysed are in brackets from 5 (wt-Miro) and 9 (Split-Miro) cells from 2 independent experiments. The data underlying the graphs shown in the figures can be found in S1 Data .

Mitochondria buffer calcium to help maintain cellular homeostasis and loss of Miro reduces calcium levels in the mitochondria of the Drosophila brain [ 11 , 12 , 30 ]. However, the mechanisms underlying decreased calcium uptake when Miro is disrupted are not understood. In S2R+ cells, Split-Miro photocleavage did not induce any changes in the fluorescent intensity of the mitochondrial calcium [Ca 2+ ] m indicator mito-GCaMP6f when compared to control cells ( Fig 4A and 4B ), indicating that the steady-state level of [Ca 2+ ] m is not affected by this manipulation. Optogenetic inactivation of Split-Miro also did not affect [Ca 2+ ] m uptake when S2R+ cells were challenged with ionomycin, an ionophore that causes a sharp increase in cytosolic calcium [ 27 , 31 ] ( Fig 4C–4E ). This result indicates that the mitochondrial morphological changes induced by Split-Miro are not sufficient to alter [Ca 2+ ] m homeostasis. Likewise, overexpression of SNPH, which rescues the Split-Miro–induced mitochondrial collapse, did not have any effect on [Ca 2+ ] m uptake when Split-Miro was cleaved ( Fig 4C–4E ).

The Miro-milton-motor complex provides a link for the attachment of mitochondria onto the microtubules [ 27 , 28 ]. To test the hypothesis that loss of microtubule tethering is responsible for the collapse of the mitochondrial network, we set out to induce mitochondrial tethering to the microtubules in a Miro-independent manner. In mammals, syntaphilin (SNPH) anchors mitochondria onto the microtubules [ 29 ], although a Drosophila homologue has not yet been found. Thus, we coexpressed EGFP-tagged human SNPH with Split-Miro in S2R+ cells stained with MitoTracker. SNPH signal in Drosophila cells overlaps with mitochondria and, as observed in mammalian neurons, SNPH puncta often localise at mitochondrial ends and are associated with strong reduction in mitochondrial dynamics (Figs 3E–3E’ , S3H and S3I ). While in cells devoid of SNPH the mitochondrial network retracts after Split-Miro photocleavage, the presence of SNPH prevents this phenotype ( Fig 3E–3G ). These observations support the notion that loss of mitochondrial anchoring on microtubules is responsible for the rapid mitochondrial network collapse when Split-Miro is cleaved.

(A) Representative images of mitochondria in Drosophila S2R+ cells at the beginning (0 minutes) and after 3 and 7 minutes of exposure to blue light, which leads to Split-Miro, but not wt-Miro, photocleavage. Control cells (top panels) were cotransfected with mCherry-tagged wt-Miro (magenta) and EGFP targeted to the mitochondria via the Zdk1-MiroC anchor (EGFP-mito, grey). Split-Miro (bottom panels) was tagged with both EGFP and mCherry to independently follow the C-terminal and N-terminal half, respectively. The mCherry is shown at the beginning and end of the imaging period to confirm retention on and release from the mitochondria in wt-Miro and Split-Miro transfected cells, respectively, under blue light. Scale bar: 10 μm. (B) Quantification of mitochondrial aspect ratio (AR) and (C) of the number of mitochondrial branches within the network, relative to A. (D) Quantifications of the mitochondrial collapse phenotype after 7 minutes of time-lapse imaging with 488-nm blue light in S2R+ cells overexpressing wt-Miro and Split-Miro with or without a Miro dsRNA construct (Miro RNAi ). Number of cells: wt-Miro = 15, Split-Miro = 15, Split-Miro + Miro RNAi = 10, Fisher’s exact test. (E) Representative images of cells expressing mCherry-Split-Miro (Split-Miro, yellow) with either an empty vector (top panels) or with EGFP-SNPH (SNPH, bottom panels, grey) and stained with MitoTracker DeepRed (MTDR, magenta). White arrows show examples of mitochondria that have retracted after Split-Miro photocleavage. Diffuse cytoplasmic yellow signal indicates release of Split-Miro N-terminus from the mitochondria. Scale bar: 10 μm. (E’) Magnified inset shows examples of stable SNPH-positive mitochondria (white arrowheads) and dynamic mitochondrial membranes devoid of SNPH (magenta arrowheads). Scale bar: 2 μm. Not shown, Split-Miro. (F) Quantification of mitochondrial AR and (G) number of mitochondrial branches at the time points indicated, relative to E. Circles represent the average AR calculated from single mitochondria within the same cell (B, F) and the average number of branches per cell normalised to the average group value (Split-Miro, Split-Miro + SNPH) at time point 0 (C, G). Comparison across time points was performed by repeated measures one-way ANOVA followed by Tukey’s post hoc test (B, C, F) and Friedman test followed by Dunn’s post hoc test (G), from 3 independent experiments. Data are reported as mean ± SEM. * p < 0.05, ** p < 0.01, **** p < 0.0001. The data underlying the graphs shown in the figures can be found in S1 Data .

We next examined the effect of Split-Miro on the integrity of the mitochondrial network in the cell soma. Strikingly, exposing Split-Miro-transfected S2R+ cells to blue light triggered a rapid (<3 minutes) and dramatic remodelling of the entire mitochondrial network, which progressively collapsed towards the centre of the cell ( Fig 3A and S3 Movie ). Mitochondria shortened along their long axis taking up a rounder shape ( Fig 3B ), which was associated with a strong reduction in the number of branches ( Fig 3C ) and reduction of the total area covered by mitochondria ( S3A Fig ). Again, the presence of endogenous Miro was dispensable for Split-Miro functionality as reducing Miro levels by RNAi did not affect the mitochondrial phenotype in Split-Miro–transfected cells ( Fig 3D ). This phenotype is not dependent on adverse effects caused by increased cytoplasmic concentration of Split-MiroN, as overexpression of the Miro N-terminal moiety alone, which diffuses throughout the cytoplasm, does not cause any gross mitochondrial morphological aberration ( S3B Fig and [ 24 , 25 ]). Of note, neither the motility and distribution of peroxisomes in the cell processes ( S3C and S3D Fig ) nor the overall organisation of the microtubule network ( S3E Fig ) were affected by this rapid cellular-scale change. Nevertheless, we reasoned that mitochondrial network collapse might impede organelle delivery towards the periphery. The number of mitochondria in the cell processes showed a progressive decline in mitochondrial content in cells expressing Split-Miro under blue light, while no effect was observed in the presence of wt-Miro ( S3F and S3G Fig ). We conclude that the rapid mitochondrial shape transition with associated loss of network integrity, clearly detectable within 3 minutes under blue light, strongly contributes to the progressive depletion of mitochondria from the cell processes.

To test whether Split-Miro photocleavage reverses the observed Miro gain-of-function effects on mitochondrial transport, we analysed mitochondrial motility under blue light in wt-Miro and Split-Miro transfected cells. Time-lapse imaging with the 488-nm laser line ensured that Split-Miro was not reconstituted while recording mitochondrial motility via the EGFP-tag ( Fig 2E , green). wt-Miro, which cannot be photocleaved, was used as a control. We found that while there was no detectable difference in the motility after imaging for 1 minute ( S2I Fig ), sustained Split-Miro photocleavage (7 minutes) reduced the time mitochondria spent on long runs and the proportion of organelles in the processes that were motile to the levels observed prior to overexpression, while there was no such effect in wt-Miro controls ( Fig 2E–2G , 488-nm laser). Cleaving Split-Miro from mitochondria did not, however, have any major effects on the velocity and run length of the moving organelles, which remained elevated and did not return to control levels ( Fig 2H , 488-nm laser, and S2J Fig ). This was also the case when endogenous Miro was depleted in the Split-Miro condition by RNAi ( S2K Fig ). Thus, using a strategy combining up-regulation (via overexpression) and down-regulation (via Split-Miro photocleavage), these results indicate that the proportion and processivity of motile mitochondria are controlled by separate Miro-dependent and Miro-independent mechanisms ( Fig 2I ).

Miro links kinesin and dynein motors to mitochondria via milton (TRAK1/2 in mammals) [ 8 , 9 , 22 – 25 ] ( S2D and S2E Fig ), and so overexpression of Miro is predicted to favour the recruitment of motor proteins on mitochondria for processive transport. Consistent with this idea, mitochondria spent more time on long runs, paused less, and engaged less frequently in short runs after either Miro or Split-Miro overexpression compared to controls ( Fig 2F , 561-nm laser, and S2F and S2G Fig ). Thus, both Miro isoforms turn the motility of mitochondria from predominantly bidirectional with frequent reversals to markedly more processive. Further supporting this notion, both Miro isoforms caused a higher proportion of mitochondria in cells processes to be motile ( Fig 2G , 561-nm laser), and their duty cycle increased in both the anterograde and retrograde directions ( S2H Fig ), suggesting that Miro participates in the activation of transport complexes for bidirectional mitochondrial transport. Interestingly, while there was no difference in the retrograde run velocities when Miro or Split-Miro were overexpressed, mitochondria traveling in the anterograde direction moved at nearly double the speed when compared to control ( Fig 2H , 561-nm laser). This observation might reflect a preference for Miro to recruit milton-kinesin complexes, consistent with what was observed in larval segmental neurons after Miro overexpression [ 26 ]. Collectively, these experiments shed light on Miro’s role in regulating mitochondrial motility in S2R+ cells and demonstrate that, in absence of blue light exposure, Split-Miro and Miro are functionally equivalent.

(A) Example of S2R+ cell treated with cytochalasin D to induce the formation of microtubule-rich processes. The microtubules and the mitochondria are stained with Tubulin Tracker and with MitoTracker Green (MitoTracker), respectively. White arrows indicate examples of cellular processes containing mitochondria. Scale bar: 10 μm. (B) Representative kymographs of mitochondrial transport in the cellular processes of S2R+ cells treated with control (upper panel) or Miro dsRNA (bottom panel). Yellow highlights indicate examples of short (<2 μm) and long runs (≥2 μm). Scale bars: 1 μm (distance) and 5 seconds (time). (C) Duty cycle analysis describes the average time mitochondria spend on long runs, short runs, or pausing. For each parameter, all mitochondrial values from each cell were averaged and compared between control and Miro dsRNA condition. (D) Percentage of motile mitochondria in cellular processes. Number of mitochondria analysed are in brackets, from 29 (Ctrl dsRNA) and 36 (Miro dsRNA) cells, from 2 independent experiments. Data are shown as mean ± SEM, multiple Student t tests (C) and Mann–Whitney test (D). This analysis suggests that Miro is necessary to drive mitochondrial motility in Drosophila S2R+ cells, and this is mainly due to modulation of long-range transport, while short-range motility appears to be largely independent of Miro. (E) Representative kymographs showing mitochondrial transport in S2R+ cellular processes before and during exposure to blue light. Control cells were cotransfected with mCherry-tagged wt-Miro, which cannot be photocleaved, and EGFP targeted to the mitochondria via the Zdk1-MiroC anchor (EGFP-mito). Split-Miro was tagged with both EGFP and mCherry to independently follow the C-terminal and N-terminal half, respectively. Mitochondrial transport was first imaged with a 561-nm laser, to capture the mCherry signal, and then with a 488-nm laser, to capture the EGFP signal while photocleaving Split-Miro ( S4 and S5 Movies). Cartoon depicts schematic of the transfected constructs. Scale bars: 5 μm (distance) and 30 seconds (time). (F) Duty cycles analysis and (G) percentage of motile mitochondria in cellular processes. In (F, G), mitochondrial transport was first analysed following the mCherry tag in cells expressing either mCherry-tagged wt-Miro or Split-Miro (561-nm laser). In a separate experiment (488-nm laser), mitochondrial transport was quantified following the EGFP tag at the beginning (first minute) and at the end (seventh minute) of the time-lapse imaging with 488-nm blue light, as shown in panel E. In (F), for each parameter, all mitochondrial values from each cell were averaged and compared to the control condition (561-nm laser) or to the first minute of blue light exposure (488-nm laser). Data are shown as mean ± SEM, from 3 independent experiments. Statistical significance was evaluated by one-way ANOVA followed by Tukey’s post hoc test (561-nm laser) or multiple paired t tests (488-nm laser) in (F) and by Kruskal–Wallis test followed by Dunn’s post hoc test (561-nm laser) or by Mann–Whitney tests (488-nm laser) in (G). Number of mitochondria analysed are in brackets, from 16 (Control), 15 (wt-Miro), and 15 (Split-Miro) cells under 561-nm laser, and 11 (wt-Miro) and 17 (Split-Miro) cells under 488-nm laser. (H) Distribution of anterograde and retrograde long run velocity after wt-Miro and Split-Miro overexpression (561-nm laser) and after irradiation with blue light to photocleave Split-Miro (488-nm laser). Solid, dashed, and dotted lines are fitted curves. Statistical significance was calculated with a Kruskal–Wallis test followed by Dunn’s post hoc test (561-nm laser) and a Mann–Whitney test (488-nm laser). N = number of mitochondrial runs. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001. (I) Model for Split-Miro–mediated regulation of mitochondrial motility. The Split-Miro trafficking data are consistent with a model in which 2 different pools of mitochondria coexist in S2R+ cells. After overexpression, Split-Miro recruits the motor complexes on the mitochondria and activates transport. On one mitochondrial pool (mitochondrial pool 1), Split-Miro links directly or indirectly (for example, via Milton/Trak, not depicted) to the motor proteins. On another subset of mitochondria (mitochondrial pool 2), the motor complexes are instead stabilised on the organelle by an unknown factor (question mark) following Miro-dependent recruitment. After Split-Miro photocleavage, mitochondrial pool 1 can reverse to control levels of processivity or become stationary (Fig 2F and 2G). However, the processivity (e.g., velocity and run length) of the mitochondrial pool 2 is not affected as the recruited motors are not directly linked to the mitochondria via Miro (Figs 2H, S2J and S2K ). This model is consistent with previous reports describing Miro-independent mitochondrial motility [ 6 , 24 ]. Split-Miro–dependent recruitment of motor complexes on an interconnected mitochondrial network, such as the one found in the perinuclear area of the cell soma, could cause significant tension on the network. Releasing the Split-Miro anchor by photocleavage, even if only on a subset of the mitochondria, would be sufficient to release the tension and cause mitochondrial network retraction, as shown in Fig 3 . Not depicted, dynein motor complex. The data underlying the graphs shown in the figures can be found in S1 Data .

Drosophila S2R+ cells are often used for intracellular trafficking experiments [ 19 , 20 ] and can be induced to extend processes ( Fig 2A ) with a stereotypical plus-end out microtubule array [ 20 , 21 ] that display Miro-dependent long-range, bidirectional mitochondrial motility (Figs 2B–2D and S2A–S2C ). With the aim of studying whether mitochondrial motility can be manipulated via Split-Miro, we first transfected S2R+ cells with either Split-Miro or wild-type Miro (wt-Miro) N-terminally tagged with mCherry. We initially exploited the mCherry tag to follow mitochondria with a 561-nm laser line, which does not photocleave Split-Miro ( Fig 2E–2G ; 561-nm laser).

(A) Schematic of Split-Miro functional domains. (B) Schematic representation of reconstituted and photocleaved mitochondrial-bound Split-Miro in its dark and lit state, respectively. Split-MiroN and Split-MiroC moieties can be followed independently by tagging them with different combinations of fluorescent proteins (e.g., EGFP/mCherry). (C) Localisation of Split-Miro in S2R+ cells cotransfected with mCherry-Split-MiroN (magenta) harbouring the T406A, T407A mutations in the N-terminus of LOV2 [ 15 , 18 ], and EGFP-tagged Split-MiroC, before and after a 570-ms pulse of blue light. Before irradiation, the mCherry and EGFP signals colocalise, while mCherry-Split-MiroN is fully released into the cytoplasm (indicated by the homogenous magenta colour) immediately after irradiation (16 seconds), and it fully reconstitutes after 3 minutes. Cartoon depicts photocleavage and reconstitution of mCherry-Split-Miro-EGFP at different time points. s, seconds. (D) Quantification of mCherry-Split-MiroN fluorescence at the mitochondria and in the cytoplasm before irradiation, at the point of maximum release and recovery indicates that the N-terminus moiety is completely released into the cytoplasm (mean ratio = 1 at release) and fully reconstituted afterwards. (E) Quantification of EGFP-Split-MiroC fluorescence indicates that the localisation of the C-terminus moiety does not change during the experiment. (F, G) Quantification of mCherry-Split-MiroN half-time release (F) and recovery (G) after photocleavage, relative to C. In (G), left panel shows levels of cytosolic Split-Miro N-terminus quantified after the maximum release is reached; right panel: recovery half-time. Data are shown as mean ± SEM. Solid line in (G) is exponential curve fit. Circles, number of cells, from 3 independent experiments. Statistical significance in (D, E) was calculated with a repeated measures one-way ANOVA followed by Tukey’s post hoc test. Scale bars: 10 μm. **** p < 0.0001. The data underlying the graphs shown in the figures can be found in S1 Data .

Discussion

Using optogenetics to implement a real-time LoF paradigm by targeting Miro, we show that collapse of the mitochondrial network is an immediate response to Miro photocleavage in S2R+ cells, which temporally precedes the defects observed in mitochondrial trafficking. We found that Miro overexpression increases the proportion and the processivity of mitochondria transported in the processes of S2R+ cells. Surprisingly, although sustained Split-Miro photocleavage reverted the proportion of transported mitochondria to control levels, the velocities and run lengths of the motile organelles were largely unaffected. Interestingly, Split-Miro photocleavage decreases the number, but not the velocities and run length, of motile mitochondria in adult fly neurons in vivo, suggesting that a potentially similar mechanism might account for the regulation of mitochondrial motility in the processes of S2R+ cells and in adult neurons. This scenario is consistent with an essential role for Miro in the recruitment of transport complexes for activation of bidirectional transport, likely by recruiting [47,48] or directly activating [49] molecular motor complexes. However, we hypothesise that once motors have been recruited onto mitochondria, they may link to the organelles via additional factors in a Miro-independent way, at least on a proportion of mitochondria (Fig 2I). Future studies should focus on discovering these factors, for example, by testing if the functional homologs of metaxins [50] in Drosophila could fulfil this role.

We previously showed that mitochondrial motility declines with age in the axons of Drosophila neurons [20,38], which can be partly rescued by boosting the cAMP/PKA signalling pathways and the abundance of the kinesin-1 motor protein [38]. However, the mechanisms underlying transport decline are still not clear. In the current study, we observed that Split-Miro photocleavage decreases mitochondrial motility in 2-day-old flies while there was no effect in 7-day-old flies. This result is intriguing as it suggests that, while important to maintain transport in young flies, Miro is largely dispensable to sustain the less abundant transport typically observed at later stages [20], implying that Miro-independent mechanisms may become predominant.

Rapid retraction of the mitochondrial network in the cell soma after Split-Miro photocleavage is conceivably a consequence of releasing membrane tension that accumulates under stretch, reminiscent of the recoil of daughter mitochondria after fission [51]. Increased mitochondrial tension following Miro overexpression is consistent with the idea that more motors are recruited and pull onto the mitochondrial membrane via cytoskeletal interaction [52–55], which would likely contribute to the buildup of mechanical energy onto an interconnected network. According to this view, releasing the link between mitochondria and the microtubules then triggers the rapid collapse of the network. It would be interesting in future investigations to establish whether the phenotype that we observe could be regarded as a “mitoquake,” i.e., rapid mitochondrial network disruption with associated release of mechanical energy, similar to the sudden cytoskeletal rearrangements (“cytoquakes”) that were proposed to underpin mechanical adaptivity during cellular dynamic processes [56].

We found that using SNPH to tether mitochondria onto the microtubule network prevented Split-Miro–induced mitochondrial reorganisation, indicating that Miro stabilises the mitochondrial network by providing an anchor to the cytoskeleton. It is possible, however, that Miro may stabilise the mitochondrial network by simultaneously bridging mitochondria to different cellular structures. In this regard, we did not find that the actin cytoskeleton plays a significant role in S2R+ cells (S6 Fig), although it is conceivable that the actin network contributes to mitochondrial stability via Miro-Myosin interactions in other cell types [6,57–59]. We showed that Split-Miro photocleavage does not affect mitochondria-ER contacts using a split-GFP reporter, suggesting that the associations mitochondria establish with the ER do not significantly contribute to the stability of the mitochondrial network. Testing whether Miro-dependent mitochondrial interaction with other cellular structures is necessary to maintain mitochondrial network stability is a goal for future studies.

Decreasing the abundance of Miro by RNAi reduced [Ca2+] m levels in the neurons of Drosophila brain [11,12], and mutating the Ca2+-binding EF domains of Miro reduced [Ca2+] m uptake in mouse hippocampal neurons [60], although Miro1-KO and Miro-EF mutant MEFs did not show any disturbances in [Ca2+] m homoeostasis [5,27]. In S2R+ cells, shedding Split-Miro functional domains (including the Ca2+-binding EF-hands motifs) did not have any effect on [Ca2+] m uptake. We hypothesise that impaired [Ca2+] m homeostasis shown with classical Miro LoF approaches (i.e., RNAi, knockout) may be a secondary effect of Miro LoF, potentially a consequence of sustained morphological and transport defects of mitochondria. However, because Miro was shown to interact with MCU and the Sam/MICOS complexes [31,32], presumably via direct interaction with the Miro transmembrane domain, we cannot exclude that the short mitochondrial targeting sequence might still mediate [Ca2+] m homeostasis. Our finding that MERCS were not affected by Split-Miro photocleavage supports the idea that [Ca2+] m uptake, known to be regulated by MERCS, is not a primary role of Miro in this context.

Overexpressing SNPH to rescue Split-Miro–dependent mitochondrial network retraction also did not affect [Ca2+] m uptake. Because SNPH also locks mitochondria into a stationary state with little network dynamic, these results raise the intriguing possibility that mitochondrial movements are not critical to maintain [Ca2+] m homeostasis, as long as mitochondria maintain their functionality.

Elegant methods for light-induced repositioning of trafficked vesicles and mitochondria have been developed, which are based on the recruitment of truncated forms of motor proteins to overpower the endogenous transport machinery and so to achieve controlled redistribution of cellular cargoes [61–64]. Engineering the LOV2-Zdk1 domain into subunits of motor and adaptor proteins could offer a complementary strategy for studying intracellular trafficking when a real-time LoF approach is preferred. Because protein photocleavage is reversible, the LOV2-Zdk1 methodology also offers significant advantages over existing methods based on the rapid, nonreversible, degradation of a target protein by the proteasome, such as the degron [65] or the TRIM-away systems [66].

Opsin-based optogenetic approaches to activate/repress specific neurons and study associated animal behaviour have been extensively used in live animals [67]. A LOV2-controlled CaMKII inhibitor was used to impair memory formation in live mice after blue light stimulation for 1 hour [68]. By creating Split-Miro flies, we combine optogenetics with Drosophila behaviour and neuronal specificity to perform LoF experiments in adult animals in real time. Although the locomotor behaviour of Split-Miro flies is indistinguishable from the wild-type counterpart before exposure to blue light, their activity is enhanced under blue light, and this hyperactivity becomes more pronounced with age.

The mechanisms underlying the exacerbated hyperactivity phenotype in older flies are unexplained. It is conceivable that the alteration of neuronal activity observed in young Split-Miro flies is linked to an acute imbalance of synaptic transmission and leads to augmented locomotor activity. We speculate that, in older flies, a potentially compromised cellular state might contribute to amplify this response. In line with this notion, ageing has been associated with increased activity of excitatory neurons in C. elegans, flies, and mice [69–72]. We are mindful that we performed these experiments in the wing neurons of the flies, and, although these cells can relay signals to affect motor phenotypes [73,74], we are cautious not to generalise our findings to all neurons in the fly.

The observed reduction in mitochondrial axonal transport induced by Split-Miro may also be a contributing factor towards enhanced fly activity, at least in young flies, by causing an imbalance in neuronal activity. It is known that reducing mitochondrial number positively correlates with activity-dependent vesicular release at the presynapses of hippocampal and cortical neurons [10,39] and with miniature excitatory junction potentials at the Drosophila NMJ [4]. In this view, the remarkable rescue of Split-Miro–induced hyperactivity in older animals by SNPH overexpression could conceivably occur via further reduction of the remaining mitochondrial transport in older neurons or via a direct effect on synaptic mitochondria [40]. Overall, our data point to an important role of neuronal mitochondrial mobility for animal behaviour and suggest that Miro could play a crucial role in preventing hyperexcitation in the ageing nervous system with potential ramification in the context of neurodegeneration.

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3002273

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.