Mitochondrial movement during its association with chloroplasts in Arabidopsis thaliana

Plant mitochondria move dynamically inside cells and this movement is classified into two types: directional movement, in which mitochondria travel long distances, and wiggling, in which mitochondria travel short distances. However, the underlying mechanisms and roles of both types of mitochondrial movement, especially wiggling, remain to be determined. Here, we used confocal laser-scanning microscopy to quantitatively characterize mitochondrial movement (rate and trajectory) in Arabidopsis thaliana mesophyll cells. Directional movement leading to long-distance migration occurred at high speed with a low angle-change rate, whereas wiggling leading to short-distance migration occurred at low speed with a high angle-change rate. The mean square displacement (MSD) analysis could separate these two movements. Directional movement was dependent on filamentous actin (F-actin), whereas mitochondrial wiggling was not, but slightly influenced by F-actin. In mesophyll cells, mitochondria could migrate by wiggling, and most of these mitochondria associated with chloroplasts. Thus, mitochondria migrate via F-actin-independent wiggling under the influence of F-actin during their association with chloroplasts in Arabidopsis.

The authors interpret that wiggling movement is mediated in a f-actin independent manner. However, the movements are also found on the continuous line of directional movement around/in the chlorophyll autofluorescence, and vice versa (Supplementary movie 1). It does not rule out the possibility that wiggling movement could be mediated by chloroplasts (as physical barriers) on the moving paths of mitochondria along f-actin. Fig 2c and Fig 3. The authors represent the data in two groups: low speed with high angle change (gray region) and high speed with low angle change (pale-blue region). It is hard to follow the interpretation. First, what does angle change mean? Please describe detailed method to measure angles. The low speed population exhibit a broad range of angles indicating that the mitochondria movements are not homogeneously regulated as Brownian movement. Actually, a few populations of mitochondria (NA in Fig. 7) exhibit high speed and low angle change and most mitochondria exhibit low speed and random movement.
Chloroplasts are associated with the plasma membrane (Oikawa et al., 2003 Plant Cell;Oikawa et al., 2008 Plant Physiology). Mitochondria with directional movement could be hindered by chloroplasts. Therefore, to insist the F-actin-independent interaction of mitochondria with chloroplasts should be confirmed by other experiments such as biochemical analysis and chloroplast movement.   This paper describes the different types of motility that are exhibited by mitochondria, changes in this behavior when in the vicinity of a chloroplast and the effect of actin and microtubule depolymerisation on motility of mitochondria.
When I read the manuscript I was surprised that many relevant papers that (in part) come to similar conclusions as in this paper were not cited. For example, the work of the Mathur group is completely missing (see e.g. Barton et al., JCS 2018), no literature is cited that covers the specific organisation and function of the peri-plastidal actin cytoskeleton (for review see e.g. Wada, Plant Sci. 2013) and Akkerman et al., Plant Cell Physiol 2011. There are significant overlaps between the work and conclusions in the current manuscript and the literature that is not cited.
Besides this problem, there are multiple issues, some of which are mentioned below: Why were protoplasts taken as a modelsystem? The authors repeat the experiments in intact leaf mesophyll cells, but the main results are obtained in protoplasts. It is uncertain how the absence of a cell wall or the removal from a tissue context affect the intracellular organisation or motility.
I presume that single z-plane time series were collected. How do the authors consider mitochondria that move in or out of the focal plane during their analyses?
The conclusion that microtubules do not affect the movement of mitochondria is premature. Since microtubules are likely to have a cortical localisation, only the motility in this plane should be considered. To link motility to microtubules, co-localisation of mitochondria and microtubules would be essential.
How do the authors define association with a chloroplast? An attachment is not shown, but it depends on the definition how this is interpreted. Since it is likely that these protoplasts have large central vacuoles, there may be a thin layer of cytoplasm that surrounds plastids. Mitochondria trapped in this cytoplasm may not necessarily associated with a chloroplast. Besides this problem, the mitochondria in the vicinity of a chloroplast show a similar wiggling behavior as the mitochondria that are not associated (see e.g. movie 4). In the absence of actin, something needs to generate a force for directional movement. That might be a flux in the cytoplasm, perhaps originating from membrane transport from plastid to cytoplasm and vice versa.
I would love to see decent statistics on the performed analyses.

Response to Reviewer #1
We sincerely express our appreciation to the Reviewer #1 for giving us valuable comments and suggestions, which have improved our manuscript. We respond to the reviewer's comments point by point as follows:

Response 1
We thank the reviewer for providing the fundamental and significant comments on our study. We improved the introduction and the discussion to clear the importance of classification of directional movement and wiggling, based on the reviewer's comments. The interesting point of the classification of two mitochondrial movements is that directional movement depends on cytoskeleton, whereas wiggling depends on interaction with chloroplast. These different types of mitochondrial movements are related to different cellular mechanisms and functions such as metabolic pathway and energy supply. These are significant points, however, it has not well been characterized in plant biology field. As the reviewer pointed out, F-actin-dependent-mitochondrial movement has been well characterized in plant cells (Logan and Leaver, 2000 8 ;Van Gestel et al, 2002 11 ;Sheahan et al, 2005 14 ;Doniwa et al, 2007 15 ;Zheng et al, 2009Zheng et al, 12, 2010 ) and the wiggling in F-actin-disrupted plant cells has also been reported (Sheahan et al, 2005 14 ;Zheng et al, 2009 12 ). However, the meaning and mechanism of the wiggling, which is induced by associating with chloroplast in normal leaf mesophyll cell, has not been clarified in detail to date. To explain these backgrounds, we inserted the additional sentences in Introduction and Discussion sections as follows: (Page 4, Lines 61-63) "This report suggested that mitochondrial movements would be influenced by chloroplast 16 . However, to date, how the mitochondrial movement is influenced by its association with chloroplast has not been clarified." (Page 31, "Taken together, the two types of mitochondrial movement, wiggling and directional movement, would be related to energy supply and metabolic pathway among cellular compartments under the influence of cytoskeleton, cytoplasmic streaming, membrane transport, and undefined tethering factors in plant cells. The wiggling induced by association with chloroplast would have an significant meaning to affect chloroplast and cellular function." Comment 2. The classification of "wiggling" as opposed to "directional" is very ad hoc. There is a huge amount of work on this by now. While the authors are not physicists, there is no reason why they cannot analyze the mean squared displacement of the mitochondria and use that for the classification and the analysis. Please check out this paper for some ideas: Particle tracking in living cells: a review of the mean square displacement method and beyond Naama Gal · Diana Lechtman-Goldstein · Daphne Weihs Rheol Acta (2013) 52:425-443.

Response 2
We thank the reviewer for suggesting that the mean squared displacement (MSD) analysis should be applied to this study. We examined MSD analysis of the mitochondrial movements for the classification following to the reviewer's comments (Pages 38-39, Lines 549-569; Methods for MSD analysis). We calculated the two-dimensional MSD <r 2 > for each trajectory of the mitochondrial movements (Gal et al. 2013 21 ;Saxton & Jacobson, 1997 23 ). At first, we fitted the MSD plot to an equation (Eq 1): where D is a diffusion coefficient, t is the time between frames, and α is the MSD scaling exponent (0≤α<2). However α value of the MSD in the oryzalin-treated cell and PA (partially associating with chloroplast) exhibited over ballistic limitation (2 < α) (Gal et al. 2013 21 ). The results indicated that the Eq 1 wasn't fitted to the MSD plots. Therefore, we fitted the MSD plots to an Eq 2: where v is a mean velocity, ∆t is the time between frames, and D is a diffusion coefficient (Supplementary Figs. 10,11,and 12;Supplementary  Comment 3. Previous work has shown that even seemingly random "wiggling" motion in cells has an active component. Analysis of the MSD may help in figuring out whether this active component is present in the mitochondria that associate with chloroplasts.

Response 3
We could separate the wiggling, directional movement, and Brownian motion using MSD analysis. The MSD of mitochondrial movements with both short-distance migration (< 5 µm) and the continuous associating with chloroplast group (CA) revealed directed-and diffusive motion with high-coefficient value (D), whereas mitochondrial movements with long-distance migration (5 µm <), in oryzalin-treated cell, the no associating (NA) and partially associating (PA) with chloroplast group showed directed motion with high-value of the mean v with or without low value of the D. The 5 µm was determined as the threshold from which the second peak appeared in distribution of the migrate distances in Fig. 2b  Taken together with the analysis of the trajectory, the MSD, biochemical co-precipitation, and physiological interaction, we redefined the wiggling and directional movement as follows: (Page 30, Lines 408-415) "Overall, the wiggling is defined as a mitochondrial movement possessing a short-distance migration with lower speed and high angle changes associated with high diffusion and low mean velocity derived from the MSD analysis, which is induced by interacting with chloroplast independently of Factin. The short-distance below 5 µm is related to chloroplast size, which mitochondria associate with.
The directional movement is defined as a mitochondrial movement possessing a long-distance migration with high speed and low angle changes associated with low diffusion and high mean velocity derived from the MSD analysis, which depends on F-actin." Comment 4. The results of the paper suggest that mitochondria that are in association with chloroplasts are not connected to active molecular motors and may be tethered to the chloroplasts.
Comparison of the MSD between cytochalasin treated cells and WT cells will help in figuring out whether the actin cytoskeleton or microtubules serves to confine the "wiggling" population or not.

Response 4
Following to the reviewer's suggestion, we performed the MSD analysis of mitochondrial movements in cytochalasin-and oryzalin-treated protoplasts (i.e., F-actin-and microtubules-disrupted cells, respectively), and compared the MSD patterns with the MSD pattern of mitochondrial movements in CA. By the MSD analysis, we could reveal that CA was independent of actin cytoskeleton or microtubule.
In cytochalasin-treated cells (F-actin-disrupted cells), the MSD pattern of the mitochondrial movements revealed both directed-and diffusive motion ( Supplementary Fig. 10e), similar to that in CA (Supplementary Fig. 11c; Supplementary Table 4). However, both D and v values were lower in cytochalasin-treated cell than those in CA. Therefore, we concluded that the wiggling was induced independently of F-actin, while migration distance of the mitochondrial movement on chloroplast would be expanded by F-actin. We added the explanations for the contribution of F-actin to mitochondrial movements in Result as follows: (Pages 28-29, Lines 379-388) "However, the MSD analysis of mitochondrial movement in cytochalasin-treated cell revealed lowdiffusion coefficient and low velocity, meaning that mitochondria exhibited diffusion in short range in F-actin free condition ( Supplementary Fig. 10e). It indicated that wiggling is independent of Factin, but F-actin would contribute to extend migrate distance of the mitochondrial movement on chloroplast, because speed of mitochondrial movement in cytochalasin-treated cell was dramatically reduced, as compared to a mitochondrial movement with a short-distance migration (

Response 1
We thank the reviewer for providing the insight about the wiggling movement is mediated by chloroplasts as simple physical barriers. Supplementary Movie 1 shows the details of the mitochondrial movements, which the reviewer pointed out. The mitochondria associating with chloroplasts revealed both the wiggling or liner movement on chloroplast before left it. However, we thought that the liner mitochondrial movements on chloroplast were not the wiggling defined in the present study. We would like to explain the reason to rule out the possibility that the reviewer suggested.
We defined the wiggling as a mitochondrial movement that consists of low speed with high angle changes, such as a short-distance migration (less than 5 µm in 30 s) on chloroplast. The 5 µm was determined as the threshold of the migration length ( Fig. 2b; Supplementary Fig. 1), related to chloroplast diameter and association with chloroplast as described in Discussion (Page 25, Lines 328-331). We inserted the sentence about the definition of the wiggling in Results as follows: (Pages, 9-10, Lines 133-138) "Taken together, mitochondria migrating less than 5 µm in 30 s moved at low speed with high angle change rates defined as wiggling, and mitochondria migrating more than 5 µm in 30 s moved at high speed with low angle change rates defined as directional movement. These results indicate that our method for evaluating and quantifying mitochondrial movements was sufficient to further explore the differences between wiggling and directional movement." We showed that the wiggling still occurred and increased in F-actin disrupted-cells (Figs. 4, 5, and 6, Supplementary Movies 4 and 5). In the revised manuscript, by the MSD analysis, we confirmed to clearly separate the wiggling from the F-actin-dependent directional movement and concluded that the wiggling is independent of F-actin. Based on the results, we rule out the possibility that wiggling

Response 2
We thank the reviewer for pointing out our obscurity in Fig To clarify "two groups: low speed with high angle change (gray region) and high speed with low angle change (pale-blue region)" as the reviewer pointed out, we classified mitochondrial movements based on criteria as follows. At first, we separated mitochondria by the migrate distance obtained from trajectory analysis of mitochondrial movements at every 1 s for 30 s ( Fig. 2b; Supplementary Fig. 1). Two different colors of plots in Fig. 2c mean different categories based on mitochondrial migration distance of less or more than 5 µm in 30 s. The 5 µm was determined as the threshold from which the second peak appeared in the distribution of the migrate distances ( Fig. 2b; Supplementary Fig. 1). As described above, we thought that the 5 µm is related to chloroplast diameter. Secondly, we plotted all sets of the speed and the angle changes, which each mitochondrial movement possessed. We separated two different types of mitochondrial movements according to distribution patterns of scatter plot (low speed with high-angle changes; gray region, and high speed with low angle changes; pale-blue region) (Fig. 2c). The result corresponded to the migrate distance. We showed each representative image (Fig. 3)

Response 3
In accordance with the reviewer's comment, mitochondrial movements of low speed (< 0.4 µm/s) have a broad range of angle change (Fig. 7c). The MSD analysis revealed that mitochondrial movement were independent of Brownian movement, as described above. The scatter plot for the speed and angle changes of mitochondrial movements in mitochondria continuously associating with chloroplast (CA) exhibited mostly with low speed -high angle changes (Fig.7c, CA). These results depended on how long mitochondria associated with chloroplasts. As described above, we examined the MSD analysis of mitochondria movements in NA, partially associating with chloroplast group (PA), and CA, and clearly separated that in CA from the others (Supplementary Fig.11in Response 1). These movements were different from Brownian movement ( Supplementary Fig.10f). The MSD of mitochondrial movements in NA and PA revealed directed motion with high value of the v, differing from that in CA, which has both directed-and diffusive motion with high value of the D. We inserted the sentence about the MSD analysis of mitochondria movement in NA, PA, and CA in Result as follows: (Pages 24-25, Lines 309-320)

Response 4
We appreciate the reviewer's comment for additional experiments to confirm the F-actin-independent interaction between mitochondria and chloroplasts. To elucidate the interaction, we examined both biochemical and physiological approaches. At first, we examined biochemically co-isolation assay, revealing that mitochondria were detected in chloroplast fraction isolated by centrifugation using Percoll gradient (Co-isolation of mitochondria and chloroplasts in Methods; Pages 37-38, Lines 520-538). We inserted the sentences about co-isolation assay and Supplementary Figure. Next, we examined the direct interaction between mitochondrion and chloroplast focusing on chloroplast movement in F-actin-disrupted cell. Without F-actin, usual light-and actin-dependent chloroplast movement stopped, however chloroplasts moved vigorously with mitochondria in accordance with protoplast. We measured the distance between a centroid of chloroplast and mitochondrion during time-lapse analysis (Methods; Page 36, Lines 514-518), revealing that interaction between two organelles was stable. We showed physiologically that interaction between mitochondrion and chloroplast was Factin-independent. We inserted the sentences about the physiological experiment in Results as follows: (Pages 29-30, Lines 401-405) "We further examined the interaction of mitochondrion with chloroplast by measuring a distance between them in mobile chloroplasts during time-lapse analysis. The result showed that the distances were kept stably under half width of chloroplast diameter even though chloroplast moved vigorously, meaning that the interaction tightly occurred (Supplementary Fig. 15, Supplementary Movie 9 Comment 5. Fig 4c. The chloroplast is abnormal in the size (>20 µm) in comparison with those presented in other Figs (~10 µm).

Response 5
We thank the reviewer for pointing out the abnormal size of chloroplasts in Fig 4b (Fig 4c in the previous   manuscript). We corrected the size of scale bar to 2 µm in the Fig. 4b. Comment 6. It is hard to recognize the MTrackJ line. Please use the thicker line.

Response 6
We modified the lines to be clearer in Fig. 4b (Fig. 4c in the previous manuscript) put in the Response 5.

Response 7
We thank the reviewer for providing us insufficient and unclear points. We confirmed the interaction between mitochondrion and F-actin on chloroplast, if the interaction was kept for more than 12 s during the time-lapse analysis. We inserted the sentence about how we determined the interaction in Methods as follows: (

Response 8
Fig . 5 shows the effect of cytochalasin B on F-actin structure. F-actin structures were mostly disrupted in 50 µM or 500 µM cytochalasin B-treated cells. We selected representative images of the wiggling in Fig. 5b, which shows similar pattern of the trajectory and scatter plot (speed -angle changes in Supplementary Fig. 5) to that in F-actin-disrupted cell (Fig. 4c). In addition, the MSD analysis defined mitochondrial movements quantitatively in F-actin-disrupted cells as the wiggling as described in Response 1. We inserted sentences about how we defined the wiggling in F-actin-disrupted cell and the quantitative data as follows:  "In 500 µM cytochalasin-treated protoplasts, the mitochondrial movement was confirmed to be the wiggling by analyzing the tracking and speed -angle changes of mitochondrial movement ( Supplementary Fig. 5 Comment 9. Line 194-196, It is hard to understand the meaning. Please rewrite it.

Response 9
We considered the reviewer's opinion and decided to remove the sentence.
Comment 10. Line 200-202, The data is not meaningful because the range of data is less than the resolution of confocal microscopy.

Minor points:
Comment 11. Please check the positions of >5 μm and <5 μm in Fig 3b

Response 11
We thank the reviewer for pointing out our mistakes. We corrected the positions (Distance <5 µm and 5 µm < Distance) in Fig 3b. Comment 12. In Supplementary Fig. 6, Constant association (CA) 0 Continuous association.

Response 12
We corrected the "constant" to "continuous" in Supplementary Fig.6 (Supplementary Fig.9 in revised manuscript). We changed "Constant association" to "Continuous association (CA)" in the other Figures.

Response to Reviewer #3
We sincerely appreciate Reviewer #3 for giving us insightful comments and suggestions, which have significantly improved our manuscript. We have carefully addressed to the reviewer comments as follows: Reviewer #3 (

Response 1
Thank you very much for pointing out insufficiency of the references, which are related to our study. We Comment 2. There are significant overlaps between the work and conclusions in the current manuscript and the literature that is not cited.

Response 2
We thought that the significant overlaps, which the reviewer pointed out, were relationship between Factin and mitochondrial movements and interaction between chloroplasts and mitochondria. We sincerely referred to the related papers in Introduction and Discussion (Pages 3-4, Lines 41-63; References No. 4 -16 please see in References). We also referred to wiggling (Page 4, Lines 54-55) as follows: "During treatment with F-actin-disrupting drugs, F-actin-independent wiggling of mitochondria was observed 12,13 ." In accordance with the reviewer's comments, our results partially overlapped with the current reports from other groups. However, our main finding that the wiggling is induced by the association with chloroplasts is different from the previous reports and an intriguing mechanism guided by trajectory analysis of mitochondrial movement in leaf mesophyll cells (migrate distance, speed, angle changes). We also performed MSD analysis of mitochondrial movements and concluded that directed-and diffusive motion with high value of the diffusion coefficient is a future of the wiggling. We inserted the additional sentences, Figures Table 4). Fig. 10e). It indicated that wiggling is independent of F-actin, but F-actin would contribute to extend migrate distance of the mitochondrial movement on chloroplast, because speed of mitochondrial movement in cytochalasin-treated cell was dramatically reduced, as compared to a mitochondrial movement with a short-distance migration (less than 5 µm) and CA (Figs. 2, 4, and 7, Supplementary Figs. 2 and 3). While F-actin apparently has role in a long-distance migration of mitochondria in cytosol as component of actomyosin system."

Both movements (wiggling and directional movement) are apparently different from Brownian motion. However, the MSD analysis of mitochondrial movement in cytochalasin-treated cell revealed low-diffusion coefficient and low velocity, meaning that mitochondria exhibited diffusion in short range in F-actin free condition (Supplementary
Conclusively, we defined the wiggling and directional movement as follows: (   We agreed with the reviewer that intracellular organization or motility in the protoplast is different from that in the intact leaf cell. In fact, mitochondrial activity was slightly reduced in the protoplast, as compared to that in the intact leaf cell (Supplementary Fig. 13). However, we thought that the mesophyll protoplast has many advantages in CLSM analysis, as compared to the intact leaf mesophyll cell. Firstly, we focused on a single leaf mesophyll cell for preventing contamination of images of chloroplasts (small and poorly developed) and mitochondria from leaf epidermis cells, which are in the immediate upper layer of the mesophyll cells. Distance between bottom of a leaf epidermis cell and top of first layer of a leaf palisade mesophyll cell is too narrow to separate these cell border. Therefore, we thought that the mesophyll protoplast is suitable for CLSM analysis to obtain clear images by easily adjusting focus on mitochondria and chloroplasts. As shown in this study, the mesophyll protoplast is easy to examine the pharmacological (e.g. cytochalasin, oryzalin) effects of mitochondrial movements. In addition, protoplast can be easily transformed by the appropriate plasmid vectors using PEG method 43 (Figs. 4 and 5).
However, we also thought that the mesophyll protoplast has also disadvantages, as the reviewer pointed out. Therefore, we performed additional experiments using the intact leaf mesophyll cells to confirm the wiggling induced by association with chloroplasts (Supplementary Figs. 1,7,8,and 13) and could obtain the similar results to that from the mesophyll protoplast. We discussed about these topics as follows: (Pages 25-26, Lines 334-345) "In addition, we applied our methods for evaluating mitochondrial movement to intact leaf mesophyll cells and obtained similar results (Supplementary Figs. 1, 7, 8, and 13, Supplementary Fig. 13a) than that in protoplasts ( Supplementary Fig. 3a, DMSO). Both the mean and maximum speeds of mitochondrial movement were higher in leaf mesophyll cells than that in protoplasts ( Supplementary Fig. 13b, c), likely due to differences in cell shape or culture conditions between protoplasts and intact leaf cells. Since

mitochondria in both cell types are active, we mainly used leaf mesophyll protoplasts to obtain clear image with avoiding contamination of mitochondrial images from leaf epidermis cells. Leaf mesophyll protoplasts is useful material for fluorescence imaging in only mesophyll cell."
Comment 4. I presume that single z-plane time series were collected. How do the authors consider mitochondria that move in or out of the focal plane during their analyses?

Response 4
We configured each parameter in the CLSM for tracking most mitochondria within 30 s at high-speed rate (250 ms/frame) and slightly expanded a pinhole size from 1 airy unit to cover mitochondria in deep area.
As a result, we could mostly track the mitochondria, however a few mitochondria still disappeared from focus plane to go behind chloroplasts as the reviewer's comment. In that case, we could not trace the mitochondria for 30 s successively and removed it from data set. We added the detail information in Comment 5. The conclusion that microtubules do not affect the movement of mitochondria is premature. Since microtubules are likely to have a cortical localisation, only the motility in this plane should be considered. To link motility to microtubules, co-localisation of mitochondria and microtubules would be essential.

Response 5
In the previous-version manuscript, we concluded that microtubule was not involved in both directional movement and wiggling, because both the movements were kept in oryzalin-treated cell similar to that in control cells. However, we performed some additional experiments and reconsidered the conclusions in this revised manuscript, according to the reviewer's comment. Based on the additional results, we concluded that microtubule would slightly give effect on mitochondria directional movement. We would like to explain how microtubule contributed to mitochondrial movements as follows. Firstly, the scatter plot in oryzalin-treated cell revealed reduction of the spots of high speed -low angle change (Fig. 4c). Frequency of speed distribution of mitochondrial movements revealed reduction in high-speed area at more than 0.5 µms -1 in oryzalin-treated cell as compared to that in DMSO ( Supplementary Fig.3). "However, plots of speed (Fig. 4c) and speed frequency (Supplementary Fig. 3) at more than 0.5 µms -1 were slightly decreased in oryzalin-treated cells. Therefore, both directional movement and wiggling occurred independently of microtubules, but microtubule may have effect on mitochondria directional movement." (Pages 26-27, Lines 352-360) "However, treatment with oryzalin, a microtubule-disrupting drug, did not inhibit mitochondrial wiggling, whereas directional movement seemed to be slightly reduced in A. thaliana mesophyll protoplasts (Fig. 4a, c, Supplementary Fig. 3, Supplementary Movie 2), suggesting that this wiggling occurs independently, but the directional movement would be slightly affected by microtubules and related motors, such as kinesins. The result would be related to microtubule function affecting actin filament organization leading to affecting mitochondria velocity and trajectory in directional movement 11,12 , or to the event that mitochondria trapped at F-actin -microtubule junction 27 ." On the other hand, we concluded that microtubule doesn't contribute to the wiggling, because the MSD of the mitochondrial movement in oryzalin-treated cell divided by the migrate distance (less or more than 5 µm) was similar to that in control cell. We inserted the sentences about MSD analysis about microtubule and Fig. 12 as follows: (Page 27, Lines 360-364) "Our MSD analysis showed that mitochondrial movement with a short-distance migration (< 5 µm) in oryzalin-treated cell maintained both directed-and diffusive motion with slow slope (D: 0.038, v: Fig. 12, Supplementary Table 4), which is characteristic of the mitochondrial wiggling. Thus, it appears that microtubules are not a driving force for the mitochondrial wiggling." Supplementary Fig. 12 MSD analysis of mitochondrial movement in oryzalin-treated cell. Each plot is fitted to curve models using the least-squares method with the Eq 2. The mean squared displacement (MSD) of mitochondrial movement in oryzalin-treated cells, which are separated to shorter (a) or longer (b) than 5.0 µm-migrate distance.

Comment 6. How do the authors define association with a chloroplast? An attachment is not
shown, but it depends on the definition how this is interpreted.

Response 6
We thank the reviewer for providing such insightful comments. We defined interaction between mitochondrion and chloroplast based on the experimental results obtained from the following experiments in the present study. Firstly, we separated the mitochondrial movements to the wiggling and directional movement based on trajectory analysis of mitochondrial movements taken by CLSM at high-speed rate.
Then, we defined association of mitochondria with chloroplast by measuring time when a mitochondrion associates with a chloroplast from the trajectory analysis described in Methods as follows: (Page 36, Lines 511-513): "To classify the dependence of mitochondrial movement on the association with chloroplasts, we analyzed the mitochondrial trajectories in the serial images. We classified all mitochondria into no,

partial (more than 3 s), and continuous (within 30 s) association with chloroplasts."
We showed that a mitochondrial movement in continuous association with chloroplast (CA) was the wiggling. In addition, we performed the MSD analysis of mitochondrial movements (Pages 24-25, Lines 303-320; Page 27, Lines 360-364; Page 28, Lines 375-388; and Page 30, Lines 408-415, see Response 1) for separating the wiggling associating with chloroplast from other type of mitochondrial movements. In the present study, we gained the evidence of the direct interaction between mitochondria and chloroplast by performing biochemical co-isolation assay (Supplementary Fig. 14) and physiological analysis by measuring stable distance between mitochondria and chloroplast during time-lapse analysis (Supplementary Fig. 15; Supplementary Movie. 9). We inserted the corresponding sentences and put

Response 7
We observed not only wiggling but also directional movement on and near chloroplasts (e.g., see S1 Video).
Given that mitochondria move freely on and near chloroplasts, it is considered that cytoplasmic thin layer around chloroplast does not mediate the association between mitochondria and chloroplasts. Therefore, we believed that there would be undefined factors directly connecting chloroplast with mitochondrion .
Comment 8. Besides this problem, the mitochondria in the vicinity of a chloroplast show a similar wiggling behavior as the mitochondria that are not associated (see e.g. movie 4).

Response 8
We checked the Supplementary Movie 4 again to doublecheck the wiggling behavior. Most of the mitochondria revealed the association with chloroplasts as marked with white asterisks (*) in the attached image below. As the reviewer pointed out, however, a few mitochondria in cytosol showed Brownian motion similar to that in fixed cells. To clear this technical issue, we removed those mitochondria from wiggling mitochondria in statistical analysis.

Supplementary Movie 4. Wiggling of mitochondria associated with a chloroplast in a CB-treated cell.
Wiggling of mitochondria (arrow) associated with a chloroplast around the central area on the chloroplast at 0 s.
White asterisks mean mitochondrion associated with chloroplast at the periphery.
We also performed statistical analysis about the number of mitochondria associated with chloroplasts in F-actin-disrupted cell, resulting that most of the mitochondria interacted with chloroplasts ( Supplementary Fig. 6). Thus, to clear these points, we inserted the sentences about the interaction as follows: (Page 29, Lines 393-395) "Moreover, disrupting F-actin with cytochalasin increased the number of wiggling mitochondria beside chloroplast (Figs. 4 and 5, Supplementary Figs. 2, 3, and 6, Supplementary Comment 9. In the absence of actin, something needs to generate a force for directional movement. That might be a flux in the cytoplasm, perhaps originating from membrane transport from plastid to cytoplasm and vice versa.

Response 9
The reviewer's idea that membrane transport from plastid to cytoplasm tethering mitochondria is attractive and we added that discussion into Discussion as follows: (

Response 10
We followed to the reviewer's suggestion and performed additional statistical analyses. We counted the number of the wiggling mitochondria in cytochalasin-treated protoplasts expressing Lifeact-Citrine ( Supplementary Fig. 6), performed MSD analysis of mitochondrial movements (Supplementary Figs. 10,11,and 12), measured time-dependent distance between a mitochondrion and a chloroplast (Supplementary  Table. 5) The quality of the analysis of the mitochondrial motion has significantly improved. However I feel that the data is richer than the analysis and the authors should, in future, consider collaborations that could enhance the mathematical and statistical analysis. I still have a few concerns however.
1. Some parts of the analysis are still confusing. On lines 128-133 the authors claim that mitochondria that inhabit the grey region move less than 5 microns while those that inhabit the blue region move more than 5 microns. However this is contradicted by Fig 2c where the pink dots representing motion greater than 5 microns are found in both the grey and the blue regions. There may be a few more pink dots in the blue region than the grey but (i) this is not claimed and (ii) no numbers or statistical tests have been performed to suggest that the two populations are statistically different. This is even more pronounced in the leaf mesophyll data in Supp Fig 1. While the distinction between the two forms of movement are apparent in Fig 2c, it appears that mitochondria move long distances by "wiggling" too. However this is not noted or discussed.
2. Please correct the typos in the MSD analysis section. You have written liner for linear, sloop for slope. Supp Fig 10. What is the difference between panel (a) and panel (c)? Does the data in panel (c) include all mitochondria, irrespective of the distance travelled? If that is so shouldn't we expect that it would be the same as the MSD analysis of panel (a) and (b) together? Or is DMSO making a difference to mitochondrial mobility? Again in Supp. Table 4 the DMSO treated mitochondria show only directed motion. There is similarly a difference between Fig 4c and Fig 2c. 4. There is no data on the goodness of fit (i.e confidence intervals or p-values) in the MSD analysis fits in Supp Table 4. So we cannot judge whether, for example, the velocity of 0.046 microns/s is significant or should be treated as effectively zero.

I am confused about the results in
5. The conclusion that "the wiggling has diffusive motion with low velocity, and that directional movement has high velocity." in lines 319 and 320 could be enhanced. In this section you currently only report the numbers in Supp table 4. However the important point here are the comparisons between the numbers. Thus you could point out that CA mitochondria show negligible directed motion compared with NA mitochondria. Thus all the previous conclusions can be recapitulated through this analysis. There are some new points that you could also make. For example, the diffusion constant is greatly reduced by cytochalasin, even when comparing with the <5micron population, suggesting that mitochondrial wiggling may be related with actin. Thus wiggling does not represent thermal diffusion but random motion due to cytoskeleton related activity (possibly powered by the mitochondria?). I suggest that instead of merely reporting the numbers you report on the comparisons between treatments, concentrating on the fit parameters that are statistically significant

Response to Reviewer #1
We sincerely express our appreciation to the Reviewer #1 for carefully reading our revised manuscript and providing us critical comments and advices to further improve our manuscript. We have addressed the reviewer's comment point by point as follows: Reviewer #1 (Remarks to the Author): The quality of the analysis of the mitochondrial motion has significantly improved.
However I feel that the data is richer than the analysis and the authors should, in future, consider collaborations that could enhance the mathematical and statistical analysis.
I still have a few concerns however.
Comment 1. Some parts of the analysis are still confusing. On lines 128-133 the authors claim that mitochondria that inhabit the grey region move less than 5 microns while those that inhabit the blue region move more than 5 microns. However this is contradicted by Fig 2c where the pink dots representing motion greater than 5 microns are found in both the grey and the blue regions. There may be a few more pink dots in the blue region than the grey but (i) this is not claimed and (ii) no numbers or statistical tests have been performed to suggest that the two populations are statistically different. This is even more pronounced in the leaf mesophyll data in Supp   Fig 1. While the distinction between the two forms of movement are apparent in Fig 2c, it appears that mitochondria move long distances by "wiggling" too. However this is not noted or discussed.

Response 1
Thank you for pointing out our insufficient interpretation of distribution of different type (short and long distance migration, black and magenta) of mitochondria in two area (grey and blue). To address the reviewer's comments, we have additionally performed statistical tests for the two different type of the mitochondria belong to short-and long-distance migration.
As the reviewer mentioned, the mitochondrion migrates more than 5 µm has also plots belong to low speed and high-angle changes and vice versa. Representative three mitochondria plots in Fig.3

Response 3
We are sorry for this confusing description. As the reviewer's comment, the Fig. 10 (c) is the MSD about all mitochondrial movement in DMSO treated-protoplasts examined as control for oryzalin-or cytochalasin-treated protoplast. The Fig. 10 (a) and (b) are the MSD about mitochondrial movement, which were separated depend on the migrate distance in non-treated protoplast. We used low concentration of DMSO, however we could not ignore the effect of DMSO on the mitochondrial movement. In addition, number of mitochondria traveling at long distance are less than that at short distance, therefore the MSD of the Fig. 10 (c) would be theoretically less than the MSD about simple sum of (a) and (b). The MSD curve of the DMSO (c) shows only directed motion, however the Fig. 4 (c) and Fig. 2 (c) contain the wiggling. We think that diffusive parameter would be hidden behind direct motion of mitochondrial movement, which mostly increased speed in time-dependent manner. We examined the MSD analysis about mitochondrial movement separated to the short-and long-migrate distance.
Comment 4. There is no data on the goodness of fit (i.e confidence intervals or p-values) in the MSD analysis fits in Supp Table 4. So we cannot judge whether, for example, the velocity of 0.046 microns/s is significant or should be treated as effectively zero.

Response 4
Based on the reviewer's comment, we additionally calculated the chi-squared value ( 2 )  Comment 5. The conclusion that "the wiggling has diffusive motion with low velocity, and that directional movement has high velocity." in lines 319 and 320 could be enhanced. In this section you currently only report the numbers in Supp There are some new points that you could also make. For example, the diffusion constant is greatly reduced by cytochalasin, even when comparing with the <5micron population, suggesting that mitochondrial wiggling may be related with actin. Thus wiggling does not represent thermal diffusion but random motion due to cytoskeleton related activity (possibly powered by the mitochondria?).
I suggest that instead of merely reporting the numbers you report on the comparisons between treatments, concentrating on the fit parameters that are statistically significant.

Response 5
Thank you for providing significant suggestion, we have considered the MSD results and inserted the sentence as follows： (Page 24, [This means that CA shows negligible directed motion compared with NA, and is well suited for wiggling.] We have also inserted the sentence about MSD results based on the reviewer's suggestion as follows; (Pages 24-25, Lines 324-326) [These MSD results conclude that the wiggling has diffusive motion with low velocity, and that directional movement has high velocity, supporting the previous conclusion in this study.] About reduction of the diffusion constant in the MSD analysis for cytochalasin, we discussed at Page 28, lines 384-393 in previous revised manuscript. Here, we have again modified the sentence following to the reviewer's suggestion in Abstract, and Discussion as follows; (Page 2, Lines 28-33 in Abstract) [The MSD analysis could separate these two movements. Directional movement was dependent on filamentous actin (F-actin), whereas mitochondrial wiggling was not, but slightly influenced by F-actin. In mesophyll cells, mitochondria could migrate by wiggling, and most of these mitochondria associated with chloroplasts. Thus,

mitochondria migrate via F-actin-independent wiggling under the influence of F-actin during their association with chloroplasts in Arabidopsis.]
(Pages 28-29, Lines 386-393 in Discussion) [However, the MSD analysis of mitochondrial movement in cytochalasin-treated cell revealed low-diffusion coefficient and low velocity, even when comparing with a mitochondrial movement with a short-distance migration (less than 5 µm) and CA (Figs.   2, 4, and 7, Supplementary Figs. 3, 4 and 10), suggesting that the wiggling may be influenced by F-actin. It means that F-actin would contribute to extend migrate distance of the mitochondrial movement on chloroplast. Therefore, the wiggling does not represent thermal diffusion, but random motion including cytoskeleton related activity.]