Design of a photocleavable Miro variant in Drosophila

To gain real-time spatiotemporal control of Miro LoF, we created a Miro variant that contains the LOV2-Zdk1 protein pair (Fig 1A and 1B) that undergoes light-induced dissociation upon exposure to blue light [14–16] and so is predicted to achieve rapid and reversible Miro LoF through protein photocleavage (Fig 1B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Split-Miro is a photocleavable version of Drosophila Miro.

(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 blue light exposure. 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.

https://doi.org/10.1371/journal.pbio.3002273.g001

We fused Zdk1 N-terminally to the mitochondrial targeting transmembrane domain of Miro, with LOV2 fused C-terminally to the rest of the protein (Fig 1A and 1B), and expressed this optogenetic variant of Miro (herein called “Split-Miro”) as 2 components in Drosophila S2R+ cells (Figs 1B, 1C, and S1A). In absence of blue light, LOV2 interacts with Zdk1, thus reconstituting Split-Miro at the mitochondria (Fig 1C). However, upon blue light exposure, the LOV2 Jα helix undergoes a conformational change that prevents Zdk1 from binding [14,15], resulting in the photocleavage of Split-Miro and the complete and rapid release of the N-terminal moiety into the cytoplasm (Fig 1D–1F and S1 Movie). Quantifications of the kinetics of release after blue light exposure showed release half-life of 1.6 ± 0.2 seconds (Fig 1F) and full Split-Miro reconstitution within 3 minutes (Fig 1D, 1E and 1G) after removal of blue light, in line with what was previously reported for the LOV2-Zdk1 association in mammalian cells [14,17]. Tagging Split-Miro only at the N-terminus did not significantly change the kinetic of photocleavage (S1B–S1G Fig and S2 Movie), thus showing the Split-Miro design supports a versatile tagging strategy to facilitate fluorophore multiplexing.

Overexpression of Split-Miro increases the proportion of motile mitochondria and their processivity in cell processes

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).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Miro regulates mitochondrial motility in the processes of S2R+ cells, and Split-Miro photocleavage reverts the effect of Split-Miro overexpression on mitochondrial motility but not velocity.

(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’s 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.

https://doi.org/10.1371/journal.pbio.3002273.g002

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.

Optogenetic cleavage of Split-Miro reduces the proportion, but not the processivity, of motile mitochondria in 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).

Split-Miro photocleavage triggers a rapid collapse of the mitochondrial network, which is rescued by mitochondrial anchoring to the microtubules

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.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Split-Miro–dependent changes of mitochondrial morphology and connectivity are rescued by SNPH expression.

(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.

https://doi.org/10.1371/journal.pbio.3002273.g003

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.

Split-Miro photocleavage does not affect mitochondrial calcium buffering or mitochondria-ER association

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 [Ca²⁺]_m indicator mito-GCaMP6f when compared to control cells (Fig 4A and 4B), indicating that the steady-state level of [Ca²⁺]_m is not affected by this manipulation. Optogenetic inactivation of Split-Miro also did not affect [Ca²⁺]_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 [Ca²⁺]_m homeostasis. Likewise, overexpression of SNPH, which rescues the Split-Miro–induced mitochondrial collapse, did not have any effect on [Ca²⁺]_m uptake when Split-Miro was cleaved (Fig 4C–4E).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Split-Miro photocleavage does not affect [Ca²⁺]_m abundance.

(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 [Ca²⁺]_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 [Ca²⁺]_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.

https://doi.org/10.1371/journal.pbio.3002273.g004

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 Ca²⁺ 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 [Ca²⁺]_m homeostasis in this context.

In vivo regulation of mitochondrial motility by Split-Miro

To gain insight into the in vivo roles of Split-Miro in regulating mitochondrial dynamics, we expressed UAS-mCherry-MiroN-LOV2 and UAS-EGFP-Zdk1-MiroC (herein “UAS-Split-Miro”) with the Appl-Gal4 driver and imaged mitochondrial motility in the axons of the adult wing neurons in situ [37]. Split-Miro is reconstituted in adult fly neurons and can be photocleaved efficiently through exposure to blue light (S5A–S5D Fig). Overexpression of Miro in Drosophila neurons in vivo has been shown to be detrimental for mitochondrial motility and overall neuronal viability [7,20]. However, expressing Split-Miro and the wt-Miro control in a miro^+/− background is permissive towards mitochondrial motility (Figs 5A, 5B, S5E and S5F and S6 Movie). Tracking of mitochondrial movements revealed that, while exposing the UAS-wt-Miro⁺ neurons to pulses of blue laser light did not cause any change in the number of motile mitochondria in 2-day-old flies (Fig 5A and 5C and S6 Movie), UAS-Split-Miro⁺ neurons showed a strong decline after 5 minutes (Fig 5B and 5D). Mitochondrial velocity and run length, however, were not affected by this manipulation (Fig 5E and 5F), similarly to what observed in S2R+ cells after Split-Miro photocleavage. Interestingly, 7-day-old UAS-Split-Miro⁺ neurons did not show any changes in mitochondrial motility under the same experimental conditions (S5E–S5H Fig). This suggests that in older flies, when the rates of mitochondrial transport are known to decline [20,38], alternative mechanisms might become predominant towards controlling mitochondrial trafficking.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Split-Miro–dependent control of mitochondrial motility in vivo.

(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’s 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.

https://doi.org/10.1371/journal.pbio.3002273.g005

Split-Miro controls neuronal activity in vivo

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 Ca²⁺ transient with the GCaMP5 Ca²⁺ 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].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Basal neuronal activity is affected by Split-Miro.

(A, B) Examples of spontaneous bouts of neuronal activity, indicated by transient fluorescence increase of the GCaMP5 Ca²⁺ 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 Ca²⁺ 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.

https://doi.org/10.1371/journal.pbio.3002273.g006

Inducible hyperactivity of Split-Miro flies is exacerbated by age

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.

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.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Optogenetic control of Drosophila locomotor behaviour in Split-Miro flies.

(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.

https://doi.org/10.1371/journal.pbio.3002273.g007

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–7H). 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–7H). 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–7K). 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.

Generation of plasmid DNA

The new constructs produced in this study are reported in S1 Table and were generated either through restriction enzymes mediated cloning or NEBuilder HiFi DNA Assembly (NEB) using the primers listed in S2 Table. The plasmid inserts were amplified by PCR using the Q5 Hot-Start High-Fidelity 2X Master Mix (NEB). Site-directed mutagenesis was performed using the Q5 Site-directed mutagenesis kit (NEB) following the manufacturer’s instructions. The fidelity of all assembled constructs was verified by Sanger sequencing (Eurofins Genomics).

Isolation of Drosophila Miro cDNA

Five Oregon-R flies were anesthetised on dry ice, rapidly grinded to powder with a plastic pestle in an Eppendorf tube and the total RNA extracted with the RNeasy Micro kit (Qiagen) following the manufacturer’s instructions. Reverse transcription was performed with the iScript Select cDNA Synthesis kit (BioRad), using 1 μl of total RNA and Oligo dT primers. Drosophila Miro is encoded by a single gene, located on the third chromosome, from which potentially 4 different transcripts are produced (FlyBase reference: FBgn0039140). The coding sequence of the longest Miro-RE/Miro-RF isoform was amplified from the total Drosophila cDNA with the primer pair #1 and #2 (S2 Table) and was used as a template to engineer wt-Miro and the Split-Miro variants employed in this study.

Cell culture, transfection, and RNAi

Drosophila S2R+ cells were obtained from the Drosophila Genomics Research Centre (DGRC, Indiana University) and cultured in Schneider’s insect medium (Gibco) at 25°C. Cells were transfected using FuGene HD transfection reagent (Promega) using a DNA:FuGene ratio of 1:3 following the manufacturer’s instructions.

For RNAi experiments, 532 bp (Miro) and 599 bp (Control) dsRNA molecules were transcribed with the MEGAScript RNAi kit (Thermo Scientific) from templates generated by PCR of sequences within the Miro 3′-UTR (targeting only endogenous, and not overexpressed, Miro and Split-Miro) and the pT2-DsRed-UAS plasmid backbone, respectively. Primers used are listed in S2 Table. Cells were treated with 15 μg/ml dsRNA for a total of 6 days, replacing the dsRNA every 24 hours in fresh media, prior to imaging.

Drosophila husbandry and generation of transgenic flies

The fly strains used in this study are listed in S3 Table. Flies were maintained on “Iberian” food [70 mg/ml yeast (Brewer’s yeast, MP Biomedicals, 903312), 55 mg/ml glucose (VWR, 10117HV), 7.7 mg/ml agar (SLS, FLY1020), 35 mg/ml organic plain white flour (Doves Farm, UK), 1.2 mg/ml Nipagin (Sigma, H3647), 0.4% propionic acid (Sigma-Aldrich, P5561] at 25°C and 60% humidity with a 12-hour light–12-hour dark cycle.

The transgenic fly lines generated in this study were obtained by phiC31-mediated transgenesis to integrate the relevant constructs into either the attP40 (25C6) or attP2 (68A4) landing sites following embryo injection.

Immunoprecipitation

Cells were transfected in 10-cm dishes, washed with PBS, and incubated with 900 μl IP lysis buffer (50 mM Trizma (pH 7.4), 150 mM NaCl, 5 mM EDTA, 0.5% Triton X-100, 1× PhosSTOP, 1× cOmplete EDTA-free protease inhibitor cocktail) on ice for 20 minutes. The cell lysate was homogenised through a 23G syringe needle and centrifuged at 21,000g for 30 minutes at 4°C. The cleared lysate (1 mg) was incubated at 4°C overnight with 25 μl of GFP-Trap beads (Chromotek) pre-equilibrated by washing 3x times in IP lysis buffer. After incubation, the beads were washed 3x times with IP buffer, and the immunoprecipitated material was eluted in 20 μl of 2X NuPAGE sample buffer supplemented with 40 mM DTT for 10 minutes at 95°C.

Western blotting

Samples were loaded on a NuPAGE 4% to 12% Bis-Tris protein gel and transferred onto an Immobilion PVDF membrane (Merk-Millipore) for 1 hour at 35 V in a wet transfer apparatus buffered with 12.5 mM Trizma-Base (Sigma), 96 mM glycine (Sigma), 10% methanol (Fisher Chemicals). The membrane was incubated in blocking solution [5% milk in PBST (PBS (pH 7.4) with 0.1% Tween-20 (VWR))] for 2 hours and probed overnight with the following primary antibodies: anti-β-tubulin (1:100, Developmental Studies Hybridoma Bank, DSHB #7, AB_2315513), anti-dMiro [75] (1:50,000), anti-Milton [76] (1:1,000, monoclonal antibody 2A108), anti-GFP (1:1,000, Chromotek, PABG1). Membranes were incubated for 1 hour with either an IRDye secondary antibody (for LI-COR Odyssey imaging) or with horseradish peroxidase (HRP)-conjugated secondary antibodies followed by a 3-minute incubation with a chemiluminescent substrate (GE Healthcare) for ChemiDoc imaging. Secondary HRP-conjugated antibodies: anti-rabbit (1:5,000, NIF824), anti-mouse (1:5,000, NIF825), anti-guinea pig (1:5,000, SeraCare 5220–0366). LI-COR secondary antibody: IR Dye 800CW goat anti-mouse IgG (1:10,000).

Live cell imaging of cargo transport and photostimulation

S2R+ cells were seeded in 4-well Nunc Lab-Tek chambered coverglass (Thermo Scientific) coated with 0.5 mg/ml Concanavalin A (Sigma). To induce the formation of cellular processes, cells were treated with 1 μM cytochalasin D (Sigma) for 3 to 4 hours before imaging and throughout the imaging period. Before imaging, cells were stained with the following dyes, depending on the specific experimental setting: 200 nM MitoTracker Green FM, 200 nM MitoTracker DeepRed FM (Thermo Scientific), 100 nM MitoView 405 (Biotium), 1X TubulinTracker DeepRed (Thermo Scientific), 1X ActinTracker (CellMask Actin Tracker Stain, Invitrogen). The MitoView 405 dye was added to the cells 15 minutes prior imaging and kept throughout. All the other dyes were washed off prior to imaging, after 30 minutes of incubation. Live cell imaging was performed using a Nikon A1RHD inverted confocal microscope, equipped with a photostimulation unit, GaAsP and multialkali PMT detectors, 405-nm, 488-nm, 561-nm, and 640-nm laser lines and a Nikon 60X/1.4NA Plan Apochromatic oil immersion objective, unless specified otherwise. Experiments were carried out at a constant 25°C temperature in an Okolab incubation system, and time series were digitally captured with the Nikon NIS-Elements software at 1 frame per seconds (fps), unless specified otherwise.

Sustained Split-Miro photocleavage in S2R+ cells was achieved with a 0.3% (0.017 mW) 488-nm laser, scan zoom of 4, at a framerate of 1 fps for the entire duration of the experiment. For dual colour imaging during sustained Split-Miro photocleavage, a frame every 1 minute was acquired with either the 405-nm, 561-nm, or 640-nm laser line. Transient whole-cell photocleavage of Split-Miro with untagged C-terminus was achieved with a stimulation step of 570 ms using a 4% 488-mn laser at a scan speed of 4 fps and 2 scanning iterations. In these experiments, a Nikon 40X/1.15NA Apochromatic water immersion objective with a scan zoom of 4 was used. The stimulation area was set to 492.7 μm², sufficient to cover the whole cell area. Time series before and after stimulation were acquired at 0.5 fps. To achieve transient photocleavage of double-tagged Split-Miro, cells were imaged with both a 488-nm (0.8%, scan time 0.6 seconds/frame) and a 561-mn (1%, scan time 0.6 seconds/frame) laser for 8 seconds with a frame rate of 0.5 fps, scan zoom of 4. Recovery was monitored by imaging with a 561-nm laser with a frame rate of 0.5 seconds for up to 3 minutes.

To combine optogenetic stimulation with 3-colour imaging and fast acquisition (as in Fig 3), the same settings were maintained, and the 640-nm laser power was set to 0.8%.

Imaging the neurons of the adult Drosophila wing was performed as reported in Vagnoni and colleagues [37,38]. For rapid blue light photostimulation in this tissue, a 5% (0.184 mW) 488-nm laser, scan zoom of 4 was used with a scan speed of 4 fps and 4 scanning iterations. The total duration of the stimulation step was 1.14 seconds, and the stimulation area was set to cover a 46.23 μm² region of the neuronal axons.

Mitochondrial trafficking in adult wing neurons was recorded with a frame rate of 0.5 frames/seconds for 5 minutes and quantified from the bundle of marginal axons along the L1 vein, essentially as we described previously [37,38]. The in vivo trafficking experiments were performed using a Nikon spinning disk system with a CSU-X1 scanning head (Yokogawa) and an inverted microscope stand (Eclipse Ti-E (Nikon)) equipped with a 60 × CFI Apo oil immersion objective (1.4 NA) and an EM-CCD camera (Du 897 iXon Ultra (Andor)). Imaging of mitochondrial motility and concomitant Split-Miro photocleavage were achieved using a 15% 488-nm laser light and 400-ms light exposure.

Analysis of mitochondrial and peroxisomal transport in S2R+ cells

Organelle motility was quantified in 2 to 6 processes for each cell. Only processes that remained in the same focal plane throughout acquisition and had no or minimal crossing with other processes were selected. Each process was straightened using the “Straighten” function in Fiji/ImageJ and kymographs produced and analysed with the “Velocity Measurement Tool” (https://dev.mri.cnrs.fr/projects/imagej-macros/wiki/Velocity_Measurement_Tool) in Fiji/ImageJ.

“Run” was defined as a continuous movement that lasts for at least 3 seconds without any change in direction or velocity. Organelles were defined as “motile” if engaging in at least one run ≥2 μm. A “pause” was defined as a period of at least 3 seconds during which the mitochondrion does not change its position. The duty cycle describes the motile behaviour of the organelles as percentage of time spent on long runs (≥2 μm), short runs (<2 μm) and pausing, or in the anterograde and retrograde direction.

Analysis of mitochondrial morphology and mitochondria-ER contacts

Mitochondrial morphology in the cell soma was analysed from cells showing a clearly interconnected network before photocleavage, using Fiji/ImageJ and a pipeline adapted from Chaudhry and colleagues [77]. Briefly, the “Adaptive Threshold” command was used to create binary images that were individually checked for accuracy in marking the mitochondrial network. The “Analyse particles” command was used to measure mitochondrial size, with a cutoff set at 0.09 μm². A skeleton of the mitochondrial network within each cell was obtained and analysed with the “Skeletonize” and “Analyze skeleton (2D/3D)” functions, respectively. The same analysis pipeline was used to quantifying mitochondria-ER contacts by creating a mask of the SPLICS signal and counting the GFP puncta throughout the cell.

Mitochondrial mass was defined as the total area covered by mitochondria within the cell and normalised to the first time point. The mitochondrial aspect ratio (AR) was measured in single mitochondria smaller than 6 μm² and was defined as the ratio between the longest and the shortest axis of the organelle. For the analysis of the mitochondrial network branching, only branches emanating from the main mitochondrial skeleton were considered.

Analysis of Split-Miro kinetics

The kinetics of the cytoplasmic release of mCherry-tagged Split-Miro N-terminus after photocleavage and the reconstitution at the mitochondria was analysed by measuring the cytosolic fluorescence intensity of the mCherry before and after photocleavage. For cell culture experiments with single-tagged Split-Miro, images were analysed with the General Analysis tool of the Nikon NIS-Elements Software by defining a region of interest (ROI) covering the cellular cytosol (with no mitochondria) and measuring the fluorescence intensity at each time point. For cell culture experiments with double-tagged Split-Miro and for neuronal in vivo experiments, images were analysed with Fiji/ImageJ by measuring the fluorescence intensity of a cytosolic ROI devoid of mitochondria (area is 0.451-μm² and 0.126-μm², respectively). For each cell or fly wing, the fluorescence intensity over time was normalised to a scale from 0 to 1, where 0 corresponds to the minimum value measured before Split-Miro N-terminus release and 1 to the maximum value measured after photocleavage. Mitochondria/cytosolic ratio of both mCherry-tagged Split-MiroN and EGFP-tagged Split-MiroC in S2R+ cells were calculated by measuring the fluorescence intensity at different time points from two 0.451-μm² ROIs located within the mitochondrial network and within a proximal area of the cytosol devoid of mitochondria. An mCherry ratio of 1 indicates the signal is homogeneous throughout the cytoplasm, and, thus, mCherry-Split-MiroN has reached complete release.

Live cell calcium imaging in S2R+ cells

S2R+ cells were imaged with a Nikon Ti-E inverted epifluorescence microscope equipped with a Nikon 60X/1.4NA Plan Apochromatic oil immersion objective, a mercury lamp (Nikon Intensilight C-HGFI) for illumination, and a Dual Andor Neo sCMOS camera for detection. Cells were plated on a Concanavalin A-coated 18-mm coverslip and imaged in a Ludin imaging chamber (Type 1, Life Imaging Services) in S2 media. Cells were imaged at 1 fps for 3 minutes under blue light, which was sufficient to induce Split-Miro–dependent mitochondrial network retraction, prior to ionomycin stimulation. A peristaltic pump system (Ismatec) was used to perfuse the imaging chamber with 2.5 μM ionomycin (Thermo Fisher Scientific) until all the media in the chamber was replaced. Additional 5-minute imaging at 1 fps was performed to record the dynamic mito-GCaMP6 signal. To verify Split-Miro cleavage, a frame with green light (to capture mCherry-Split-MiroN) was acquired at the beginning and the end of each acquisition period.

To quantify mitochondrial calcium, an ROI was drawn to outline single cells and the fluorescence intensity of mito-GCaMP6f signal measured using the Nikon NIS-Elements software. Response curves were aligned at the base of the response peak, and, for each cell, the fluorescence intensity at every time point was normalised to the value of the first frame. Only cells that responded at least 2-fold to ionomycin in the first 100 seconds were included in the quantification. Mito-GCaMP6f “peak” was defined as the maximum fold-change value reached during imaging. The time to reach the peak was calculated as the Δt between the “peak” time point and the time of first response to stimulation.

Live cell calcium imaging in adult neurons

Calcium transients in the wing marginal neurons were acquired using the same spinning disk imaging system described to capture mitochondrial trafficking. Image series were recorded for 5 minutes using a 10% 488-nm laser light with an acquisition time of 2 frames/seconds and exposure time of 200 ms. Image series from 1 to 2 fields of view were acquired from the same areas of each wing (i.e., mid-margin to the L1 to L2 vein intersection). To quantify the GCaMP fluorescence during spontaneous calcium transients, an ROI was drawn around each responding neuronal soma and axons. After background subtraction in Fiji/ImageJ (rolling ball radius = 50 px), the ROI mean intensity over time was calculated using the Plot Z-axis profile command. GCaMP fluorescence peak and time to reach the peak were calculated as done for calcium imaging in S2R+ cells.

Behavioural assay

Flies were anesthetised with CO₂ on the day of eclosion, transferred to new food, and aged for either 2 or 8 days prior to the assay. On the day of the assay, flies were briefly anesthetised on ice and individually loaded into 65-mm glass tubes on custom-made platforms. Fly activity was recorded in an optogenetic enclosure (BFK Lab) equipped with LEDs for 444-nm blue light stimulation and a webcam for video recording. The position of each fly was automatically tracked and the quantification of the fly locomotor activity performed using the Drosophila ARousal Tracking (DART) system [45,46]. Flies were left to acclimatise for 20 minutes before starting the experiment. Fly activity was recorded continuously for 30 minutes before the blue LED light was switched on for 2 hours (power at the platform: 5 mW/cm²). After the blue light was switched off, fly movements were recorded for an additional hour, for a combined experimental time of 3 hours, 30 minutes. Flies were defined as “active” if displaying bouts of activity of at least 2 mm/seconds. The active average speed refers to the speed of fly movement when active. The action initiation is defined as the number of times a bout of activity is started per second. The % activity represents the percentage of time the fly spends moving.

Statistical analysis and image preparation

Data were analysed with Microsoft Excel and GraphPad Prism 9. Statistical tests and number of replicates are reported in the figure legends. The data underlying the graphs shown in the figures can be found in S1 Data. Images were assembled using Fiji/ImageJ. For presentation purposes, “Subtract Background” (rolling ball radius = 20 pixels), and either the “Despeckle” or “Smooth” filters were used to reduce salt and pepper noise. All images in the same experimental series were processed in the same manner.