mTORC1 activity is supported by spatial association with focal adhesions

Rabanal-Ruiz and Byron et al. present a novel mechanism of nutrient signaling that identifies FAs as key cellular hubs that coordinate growth factor signaling and amino acid input into the cell and are required for efficient downstream activation of mTORC1.

You will see that It's clear that your model connecting mTORC1 activity to focal adhesions (FAs) spatially is interesting to adhesion and mTOR experts. Based on their interest, which we share, we'd like to consider the work further. However, the referees raise similar, significant issues with the work: the evidence suggesting that focal adhesions promote mTORC1 signaling and are sites of mTORC1 activation is not fully convincing because you don't rule out other factors that could contribute to the phenotypes seen (e.g., upon Arl8B depletion, Rev#1 #5 and Rev#3; upon impairment of FA proteins, per Rev#3) and the work lacks the necessary rescues and controls to show specific effects of adhesions on mTORC1 signaling (see Rev#1 main comment and points #2 -lack of specificity in Ab used; #3/4 lack of rapamycin control; #6; Rev#2 #3-4-5). The referees additionally comment on the multiple systems used (Rev#1, Rev#2 #2). In addition, their remarks suggest that the results are not sufficiently well integrated with the literature (e.g., Rev#2 points #6 and #1) -including prior work linking mTOR to fibrillary adhesions as stressed by Rev#2. Both Revs #1 and #3 have questions about the hits from the proteomics studies (Rev#1 #1 and Rev#3 minor point #1).
We have editorially discussed these issues and find them valid and important. Because there are a number of factors that could contribute to the effects of FAs on mTORC, as the reviewers highlighted with great detail, definitive evidence is needed to implicate FAs or particular types of adhesions in promoting peripheral mTORC1 activation. More work is needed to provide much stronger and definitive evidence that should convince experts in the fields of metabolism and adhesion for publication.
Please let us know if you are able to address the major issues outlined above and wish to submit a revised manuscript to JCB. Note that a substantial amount of additional experimental data likely would be needed to satisfactorily address the concerns of the reviewers in full. It may be necessary to extend your manuscript to a full Research Article. As you may know, the typical timeframe for revisions is three to four months. However, we at JCB realize that the implementation of social distancing and shelter in place measures that limit spread of COVID-19 also pose challenges to scientific researchers. Lab closures especially are preventing scientists from conducting experiments to further their research. Therefore, JCB has waived the revision time limit. We recommend that you reach out to the editors once your lab has reopened to decide on an appropriate time frame for resubmission. Please note that papers are generally considered through only one revision cycle, so any revised manuscript will likely be either accepted or rejected.
If you choose to revise and resubmit your manuscript, please also attend to the following editorial points. Please direct any editorial questions to the journal office.
GENERAL GUIDELINES: Text limits: Character count for a Report is < 20,000; a full Research Article is < 40,000, not including spaces. Count includes title page, abstract, introduction, results, discussion, acknowledgments, and figure legends. Count does not include materials and methods, references, tables, or supplemental legends. If you choose to resubmit, please include a cover letter addressing the reviewers' comments point by point. Please also highlight all changes in the text of the manuscript.
Regardless of how you choose to proceed, we hope that the comments below will prove constructive as your work progresses. We would be happy to discuss them further once you've had a chance to consider the points raised. You can contact the journal office with any questions, cellbio@rockefeller.edu or call (212) 327-8588.
Thank you for thinking of JCB as an appropriate place to publish your work.
Reviewer #1 (Comments to the Authors (Required)): The manuscript by Rabanal-Ruiz et al proposes an intriguing hypothesis that localization of mTORC1 near the focal adhesions (FA) at the plasma membrane brings mTORC1 activity "into close association with nutrient inputs". Specifically, the authors claim that (1) "mTORC1 is activated in specific regions colocalizing with FAs" and that (2) "FAs are required for maximal mTORC1 activation in response to nutrients". To support their claims, the authors present the following evidence: • Nearly half of proteins identified in a BioID experiment with an mTORC1 component Raptor overlap with proteins identified in FAs • mTOR, LAMP1 and paxillin show partial FCS/aa-induced colocalization • in about half of cells, IF of phosphorylated mTORC1 substrates 4EBP1 and S6 shows increase both in the cell periphery and interior, partially overlapping with LAMP1 and paxillin staining • a ratiometric FRET reporter for the mTORC1-specific substrate 4EBP1 indicates lower FRET in sites overlapping with GFP-paxillin when treated with FCS • KD of ARL8B leading to perinuclear lysosome accumulation decreases phosphorylation of mTORC1 substrate S6 • Talin1/2 KO cells show reduced S6 phosphorylation (both in the periphery and in cell interior) • ROCK-specific inhibitor and integrin antagonist reduced S6 phosphorylation • Cells with constitutive mTORC1 signaling show increased number of FAs Overall, the presented evidence points in favor of the authors' hypothesis, albeit indirectly. I personally find the hypothesis is indeed interesting, generally in line with previous reports and worth publishing; yet, the presented evidence is mostly correlative and insufficient to examine the causeand-effect relationship. While some of the authors' models and tools (e.g., Talin1/2 KO cells, Arl8B KD, ratiometric FRET reporter) appear to be well-suited for examining their working hypothesis in detail, the key experiments, such as Talin-or Arl8B rescue, inhibition of mTORC1/mTOR with rapamycin/Torin1, Akt inhibition, using FRET reporter in Talin1/2 KO cells, etc, are missing. I'd be supporting publication of the report, but in my opinion it would require major revision. I'd recommend the authors to clarify the working hypothesis (what does "SUPPORTING mTORC1 activity at the periphery" exactly mean?) and concentrate on 1-2 model systems.
Major comments: 1. As mTORC1 could be both membrane-bound and cytosolic, I'd expect a lot of proteins to be biotinylated by a cytosolic protein over 16 hrs. The labeling radius of 10 nm the authors refer to was estimated for a stable transmembrane complex (NPC) and therefore may not necessarily apply here. Further, from the presented data it is not clear what fraction of BioID2-fused Raptor is integrated into a complex. As the enzyme could potentially interfere with mTORC1 formation, this estimation is very relevant for the BioID/MS results. I'd expect some sort of negative control to be quite useful to demonstrate labeling specificity. A statement on the reproducibility of the proteomic data would also be desirable: the authors indicate that bead-bound proteins from four independent experiments we used; were they pooled or analyzed separately? 2. The authors use mTOR-specific antibody as a proxy for mTORC1. Given that mTOR is a part of both mTORC1 and mTORC2, and at least a significant portion of mTORC2 localizes to the PM, Raptor-specific Ab appears to be a better choice for IF. 3. The authors indicate that >40% of cells show phospho-S6 staining at the periphery. What happens to the other cells? Do they not have peripheral phospho-S6 or do not respond to refeeding? Further, in such experiments, rapamycin treatment is a natural control needed to establish mTORC1 specificity; why didn't the authors use it? 4. I am somewhat worried about the results obtained using TORCAR FRET reporter, which appears to be one of the strongest evidence for the authors' conclusions. First, rapamycin should be used after FCS treatment to demonstrate both the reversibility and mTORC1 specificity of the FRET signal. Loss of FRET is often associated with unequal photobleaching, especially with a relatively low dynamic range of the FRET signal. Next, GFP and YPet excitation spectra overlap significantly, which could result in weird ratio even after unmixing; the lack of the signal in the "control" (i.e., GFPfree) areas could equally well be interpreted as an unmixing artifact. I'd strongly recommend using far-red FPs for paxillin. Finally, I could not find an example of the FRET ratio image when cells are treated with cell-permeable leucin analog; please add. 5. Arl8b KO was shown to significantly delay mTORC1 as well as mTORC2/Akt activity upon refeeding, with ~30-40% recovered only after 1 hour (Jia and Bonifacino, Mol Cell 2019). The authors, however, compare phospho-S6 staining only after 10 min. Does the peripheral mTORC1 pool still contribute to S6 phosphorylation 1 hr after re-feeding? Could PM-localized Akt (and not FAs) play a role in mTORC1 activation at the PM (to remove Tsc2)? 6. Experiments in Talin1/2 KO cells, adherent/suspension cells and with ROCKi and integrin antagonist are the strongest evidence supporting the authors' hypothesis, but they require proper specificity controls with mTORC1, mTORC2, PI3K and Akt substrates and the corresponding rescue/inhibitor washout experiments wherever possible. 7. Finally, I think it is important to better define the authors' working hypothesis. They propose that "peripheral dispersal of mTORC1-positive lysosomes supports mTORC1 activation by bringing it into close proximity to incoming mitogenic signaling cascades and amino acids". Of course this could be the case, but discrimination between the contribution of the peripheral and interior pools of mTORC1 seems to be a non-trivial task, as such spatially distinct pools could have different catalytic activities, flux of substrates or distinct inactivation kinetics. Unraveling the contribution of these parameters would require careful experimentation.
The reviewer cannot comment on the quality of the proteomic data (out of expertise). The reviewer claims no competing interests.
Reviewer #2 (Comments to the Authors (Required)): In the present study the authors expand on their earlier exciting data showing that peripheral lysosomes are a feature of fed cells and that starvation triggers perinuclear lysosome localisation. Here the authors have exploited the powerful proximity-biotinylation method and find that adhesion components are significantly overrepresented in the RPTOR BioID dataset. They go on to show that mTOR and lysosomes are tethered to focal adhesion, not directly through FA components but via FA associated growth factor receptors. This is a predominantly carefully conducted study. However, I have some suggestions that the authors may want to consider addressing to further increase the impact of the findings and the relevance to the adhesion biology field.
1) The authors provide a nicely illustrated and informative representation of their RPTOR BioID2 dataset that convincingly demonstrates the proximity of mTORC1 with focal adhesions. Interestingly, two of the nine proteins were found significantly enriched in RPTOR/Kindlin/paxillin proximity proteomes are linked predominantly with fibrillar adhesions (tensin-1 and tensin-3) and two are components of the FA proximal structures that link focal adhesion to microtubules (Kank and liprin). These data would imply that mTORC1 proximity is not linked to focal adhesion specifically but possibly engages different types of adhesions in cells. This would also be in line with a previous study showing recruitment of lysosomes and active mTOR to fibrillary adhesions (DOI:https://doi.org/10.1016/j.celrep.2014.12.037). Could the authors expand on their data in Figure 2 to also other adhesion types than paxillin-positive adhesions?
2) The proximity biotinylation was carried out in osteosarcoma U2OS cells. Why is the validation data of mTOR and LAMP1 recruitment to paxillin positive adhesion with HeLa cells? The authors should consider showing this (and possibly localization to other adhesion types as well) in other cell types to demonstrate the generality/specificity of these data (there is some data from MEFs and MRC5 cells later, but would be nice to have a broader validation here).
3) Figure 3a. Could the authors somehow demonstrate that the "significant increase in colocalisation between p-S6 and paxillin" is not just a reflection of having low p-S6 in the cell in starved conditions to having it abundantly throughout the cells in fed conditions? 4) Figure 4a. How have the authors validated the specificity/ruled out off-target effects of the ARL8B siRNA? Does ARL8B siRNA influence adhesion number or distribution in the cells. Starvation in other cell models has been shown to impact on the balance between peripheral focal and adhesions and elongated adhesions localising under the cell body. Figure 4G would be easier to interpret if starved and fed condition were shown like in the other experiments.

5)
6) In the discussion the authors speculate that "Activation of mTORC1 at FAs by nutrients may therefore suppress autophagy of FA components". They should mention earlier work describing already that mTOR activity suppressed integrin and ECM protein uptake into spatially localised adhesion proximal lysosomes.
Minor points: Figure 2c. What is SDHA? Why was it chosen as control? It might be better to score the number of PLA spots/cell or the number of PLA spots overlapping with Paxillin rather than the % of PLA positive cells. Figure 3c would benefit from adding labelling in the images (to indicate what the green signal is). Page 7 top, this sentence may need to be edited "although at this point is cannot necessarily be completely ruled out." Reviewer #3 (Comments to the Authors (Required)): The present work provides evidence that focal adhesions are required for maximal mTORC1 activation in response to nutrients. The authors based investigations on their earlier work demonstrating that the localization of lysosomes near the plasma membrane increases mTORC1 activity by bringing it into close association with nutrient inputs into the cell.
Here it is shown that mTORC1 interacts with focal adhesion proteins by several proteomics mass spectrometry-based approaches that were confirmed and analyzed in detail by complementary immunofluorescence approaches. Although data analyses were performed at a state-of-the-art level, and the results are clearly represented the manuscript falls short in demonstrating under which physiological conditions the postulated "maximal mTORC1 activation" would be required. In addition, although in the Arl8b KD experiment mTORC1 signaling is compromised, it remains to my opinion unclear if the results obtained can be really assigned to a lack of targeting focal adhesions. It could be equally well possible that Arl8b KD might interfere with mTORC1 signaling by increasing the amount of ragulator associating with BORC and thereby sequestering it away from Rags. If one took this into account, the conclusions would be completely different. Therefore, the authors could for instance artificially target mTORC1 to FA (eg. FAT domain on Rheb) and look if this would rescue the signaling defect detected upon Arl8b KD. Importantly, depletion of focal adhesions, either by Talin KD or ROCK inhibitors, has a significant impact on the cytoskeleton. Chemotactic changes and/or alterations in Rac1 activity have previously been shown to regulate both mTORC1 and mTORC2. Therefore, presented experiments should be interpreted with caution since some of the effects observed might be independent of the here described focal adhesion dependent activation of mTORC1.
Minor points that should be revisited by the authors: Figure 1: The results from the MS analysis of the BioID2-RPTOR pulldowns is in part surprising. Why is the log2 enrichment for the autophagy related proteins considerably higher than that observed for core mTORC1 subunits? Figure 3: The authors state that "FAs play an important role in the early events of mTORC1 activation in response to nutrients". Unfortunately, there are a few points unclear to this reviewer concerning the Leucine methyl ester stimulation experiment. In the legend from figure 3 it is mentioned that in panel E, "each data point represents a coverslip including several cells" whereas in panel F, "each data point represents a single cell". Also, in the legend from supplementary figure 3 panel A, "each data point represents a coverslip including several cells". Could the authors please explain why the analysis from the Leucine methyl ester stimulation was apparently performed differently from the remaining TORCAR biosensor experiments? Further, it would make the authors claim stronger if they could include an independent positive control, e.g. immunoblotting, to detect mTORC1 activation upon Leucine methyl ester stimulation.
Supplementary Figure 2: The image quality on subcellular localization of p-4EBP1 after AA refeeding is not convincing. The entire Panel would also profit from colocalization analysis with either LAMP1 or Paxillin, similarly to what is provided in the remaining panels of the same figure. Figure 4b shows quantified integrated densities of pS6 signals in IF experiments. Somewhat surprisingly the decrease of pS6 signal upon ARL8B knock-down is almost the same for peripheral and intracellular pS6. Such data presentation might give an impression to the readers that peripheral (10% of the cell area) vs intracellular regions were not properly defined. Considering an approximately 2.5-fold decrease in total cell number with peripheral p-S6 upon ARL8b depletion (panel c), one would expect a more specific effect for peripheral staining. The same criticism applies to corresponding Supplementary Figures 4b-c.
Corresponding to panels a-c in Figure 4, panel d shows immunoblotting analyses of total pS6. It is partially misleading because the conditions (aa and aa plus FCS) are different from panels a-c (FCS starvation and FSC stimulation), and more related to Supplementary Figures 4a-d. The exact conditions are not described, but one could assume 18 h FCS starvation (panels a-c) and 1h aa/FCS starvation in panel d. We appreciate the constructive criticism and suggestions made by all three reviewers on our manuscript entitled "mTORC1 activity is supported by spatial association with focal adhesions". We have now performed extensive revisions and are pleased to offer the modified manuscript for your consideration. Please also find below a point-by-point response to the reviewers' comments, where our replies are in red.

Reviewer #1 (Comments to the Authors (Required)):
Major comments: 1. As mTORC1 could be both membrane-bound and cytosolic, I'd expect a lot of proteins to be biotinylated by a cytosolic protein over 16 hrs. The labeling radius of 10 nm the authors refer to was estimated for a stable transmembrane complex (NPC) and therefore may not necessarily apply here. Further, from the presented data it is not clear what fraction of BioID2-fused Raptor is integrated into a complex. As the enzyme could potentially interfere with mTORC1 formation, this estimation is very relevant for the BioID/MS results.
We agree that soluble cytosolic proteins could potentially interact with significantly more molecules than transmembrane components. However, we could not find any evidence from the literature that this would be the case, and the labelling radius for BioID2 would nevertheless still be expected to remain the same. Therefore, we added an acknowledgement in the text that the proximity is only an estimate from the studies with NPC components. At the same time, as suggested, we performed additional controls for our BioID2-fused RPTOR construct. Specifically, we show that RPTOR-BioID2 biotinylates mTOR and is efficiently incorporated into mTORC1 complex (Fig. S1 B).
I'd expect some sort of negative control to be quite useful to demonstrate labeling specificity. A statement on the reproducibility of the proteomic data would also be desirable: the authors indicate that bead-bound proteins from four independent experiments we used; were they pooled or analyzed separately?
A negative control was included in all four independent experimental replicate analyses (see  Table S1). The four independent experiments were not pooled, so we were able to assess reproducibility (Fig. S1 C and Table S1 of the manuscript). All four independent experiments were highly correlated (RPTOR BioID Spearman rank correlation coefficients > 0.91; negative control Spearman rank correlation coefficients > 0.81; Fig. S1 C). We have now included a statement of reproducibility and clarified the specific experimental conditions that were analysed.
2. The authors use mTOR-specific antibody as a proxy for mTORC1. Given that mTOR is a part of both mTORC1 and mTORC2, and at least a significant portion of mTORC2 localizes to the PM, Raptor-specific Ab appears to be a better choice for IF.
The main finding of the paper is that mTORC1 is activated in the vicinity of FAs, thus we use phospho-S6, which is specific to mTORC1, not mTORC2. Additionally, we use the mTORC1-specific FRET reporter TORCAR. As there are no reagents that would allow us to characterise intracellular localisation of RPTOR, we cannot carry out the proposed experiment; however, we made it clear in the text that both phospho-S6 and TORCAR are the measure of mTORC1 but not mTORC2 activity and included rapamycin controls for all our assays to demonstrate their specificity for mTORC1.
3. The authors indicate that >40% of cells show phospho-S6 staining at the periphery. What happens to the other cells? Do they not have peripheral phospho-S6 or do not respond to refeeding? Further, in such experiments, rapamycin treatment is a natural control needed to establish mTORC1 specificity; why didn't the authors use it?
As shown in our earlier publication (Korolchuk et al. (2011) Nat Cell Biol), translocation of lysosomes to the cell periphery in response to nutrients is a stochastic event which is observed in approximately half of the cell population at any given timepoint. However, what is important for our current manuscript is that, in all cells where lysosomes do translocate to the periphery, we also observe peripheral phospho-S6 staining, thus indicating direct and extremely high correlation between the two events (Fig. S2 E). To go beyond correlation, in addition to ARL8B KD data, which was in our original manuscript, we have now also performed experiments where we forced lysosomes to the periphery. Specifically, we increased the % cells with peripheral lysosomes by over-expressing ARL8B, which we previously showed to promote mTORC1 activity (Korolchuk et al. (2011) Nat Cell Biol). We found that, consistent with our model, ARL8B increases peripheral phospho-S6 levels and the proportion of cells with peripheral phospho-S6 (Fig. S2 F). Rapamycin controls have been done in the initial phases of the project, however were not included as our starvation/recovery experiments clearly show the response of phospho-S6 to mTORC1 inhibition. However, we agree with the Reviewer that rapamycin controls for phospho-S6 and other readouts for mTORC1 are useful and these are now shown in Fig. 3 and Fig. S3.
4. I am somewhat worried about the results obtained using TORCAR FRET reporter, which appears to be one of the strongest evidence for the authors' conclusions. First, rapamycin should be used after FCS treatment to demonstrate both the reversibility and mTORC1 specificity of the FRET signal.
According to the Reviewer's suggestion, we have now performed TORCAR FRET measurements in the presence of rapamycin. Results of these experiments revealed that rapamycin treatment blocks FCS-mediated increase in mTORC1 activity in focal adhesions (Fig. S3 D).
Loss of FRET is often associated with unequal photobleaching, especially with a relatively low dynamic range of the FRET signal.
We fully agree with the Reviewer that the quantitative FRET analysis is a difficult issue with multiple pitfalls. Therefore, during the control experiments (i.e. TORCAR FRET measurements without FCS), we performed recordings of the fluorescence intensity separately for the donor and acceptor. In Fig. R1 below, you can see a representative graph illustrating that, under our experimental conditions, we did not observe significant and/or unequal bleaching of donor (CFP) and acceptor (YPet) fluorophores. Moreover, experiments in which cells were treated with DMEM without FCS show no difference in FRET ratio over time between focal adhesions and control sites (Fig. S3 D), further confirming physiological origin of FCS-mediated mTORC1 activation in focal adhesions instead of bleaching artefacts. Next, GFP and YPet excitation spectra overlap significantly, which could result in weird ratio even after unmixing; the lack of the signal in the "control" (i.e., GFP-free) areas could equally well be interpreted as an unmixing artifact. I'd strongly recommend using far-red FPs for paxillin.
We appreciate the opportunity to clarify this point. To overcome the problem with overlapping GFP and YPet spectra, we combine reference-based spectral imaging using the highly sensitive spectral detector of Zeiss LSM 780 ("lambda mode") with an advanced offline algorithm developed in our lab (Wlodarczyk et al.  Finally, I could not find an example of the FRET ratio image when cells are treated with cellpermeable leucin analog; please add. In line with the Reviewer's suggestion, we have added representative examples of TORCAR FRET images after treatment of cells with leucine methylester to the revised manuscript (Fig.  S3 C). 5. Arl8b KO was shown to significantly delay mTORC1 as well as mTORC2/Akt activity upon re-feeding, with ~30-40% recovered only after 1 hour (Jia and Bonifacino, Mol Cell 2019). The authors, however, compare phospho-S6 staining only after 10 min. Does the peripheral mTORC1 pool still contribute to S6 phosphorylation 1 hr after re-feeding? Could PMlocalized Akt (and not FAs) play a role in mTORC1 activation at the PM (to remove Tsc2)?
We would like to point out that our manuscript has a focus on the early events in mTORC1 activation in response to growth-promoting stimuli, which is the major area of research in the mTOR field. However, we now performed re-feeding of cells both for 10 and 60 minutes and, in agreement with Jia and Bonifacino (2019) Mol Cell, found that ARL8B KD suppresses mTORC1 re-activation at both time points (Fig. S6 E). Furthermore, we found that constitutive targeting of mTORC1 to FAs (see below) rescues the mTORC1 activation defect in ARL8B KD cells at both time points (Fig. S6 E).
Re. the role of PM-localized Akt; although monitoring activation of Akt at the subcellular level is challenging with reagents currently available in the field, we have demonstrated that GFRs and mTORC1 are specifically activated in paxillin-positive FA, so it is highly likely that Akt is also phosphorylated in these regions. The mechanism by which Akt removes TSC from peripheral lysosomes is plausible, however we were unable to detect TSC2 on peripheral lysosomes and therefore could not test this further, but as the mechanisms controlling TSC localisation to the lysosome are still relatively unknown, future studies may shed light on this question.

Experiments in Talin1/2 KO cells, adherent/suspension cells and with ROCKi and integrin
antagonist are the strongest evidence supporting the authors' hypothesis, but they require proper specificity controls with mTORC1, mTORC2, PI3K and Akt substrates and the corresponding rescue/inhibitor washout experiments wherever possible.
In addition to the previously shown mTORC1 readouts, we now show phospho-Akt S308 and phospho-Akt T473 blots, readouts for PI3K/Akt and mTORC2 activities, respectively (Fig. 5  B). Both readouts are partially suppressed by Talin knockout, consistent with suppression of the entire GFR/PI3K/Akt/mTOR axis upon FA perturbation.
In our experiments with ROCKi and integrin antagonist cilengitide, cells were pre-treated with the drugs for the last 2 hours of the starvation regime. We found that, whether the 10 min re-feeding was performed in the presence or absence (effectively drug washout) of drugs, mTORC1 was equally suppressed (not shown). This indicates to us that either the drugs are irreversible, or 10 min of washout is not sufficient to re-form the FAs. In either case, we do not believe we can provide the informative drug washout data within the feeding protocols used in our study.
7. Finally, I think it is important to better define the authors' working hypothesis. They propose that "peripheral dispersal of mTORC1-positive lysosomes supports mTORC1 activation by bringing it into close proximity to incoming mitogenic signaling cascades and amino acids". Of course this could be the case, but discrimination between the contribution of the peripheral and interior pools of mTORC1 seems to be a non-trivial task, as such spatially distinct pools could have different catalytic activities, flux of substrates or distinct inactivation kinetics. Unraveling the contribution of these parameters would require careful experimentation.
We have considered this point carefully and believe that our new experiments in the revised manuscript provide support for our original working hypothesis. We do agree, however, that we are only starting to understand the complexity of mTORC1 signalling at the sub-cellular level and there are limitations to our study; we highlighted this in the Discussion section of the revised manuscript.

Reviewer #2 (Comments to the Authors (Required)):
1) The authors provide a nicely illustrated and informative representation of their RPTOR BioID2 dataset that convincingly demonstrates the proximity of mTORC1 with focal adhesions. Interestingly, two of the nine proteins were found significantly enriched in RPTOR/Kindlin/paxillin proximity proteomes are linked predominantly with fibrillar adhesions (tensin-1 and tensin-3) and two are components of the FA proximal structures that link focal adhesion to microtubules (Kank and liprin). These data would imply that mTORC1 proximity is not linked to focal adhesion specifically but possibly engages different types of adhesions in cells. This would also be in line with a previous study showing recruitment of lysosomes and active mTOR to fibrillary adhesions (doi: 10.1016/j.celrep.2014.12.037). Could the authors expand on their data in Figure 2 to also other adhesion types than paxillin-positive adhesions?
We have consulted with Prof. Martin Humphries at the University of Manchester, an expert in cell adhesion, and we are of the understanding that there is no definitive marker for fibrillar adhesions but that localisation within the cell is commonly used to identify such structures and they can, in fact, be paxillin-positive. We also note that tensin-1 and tensin-3 have been described to localise to peripheral focal adhesions (Barber-Perez et al. (2020) J Cell Sci). We therefore analysed the distribution of paxillin-positive structures and how it correlates with phospho-S6 distribution, i.e. at the cell periphery versus under the centre of the cell. As can be seen in Fig. 3, C and L, and Fig. S3 A, paxillin accumulates in cell periphery regions (corresponding to FAs) and also at lower density under the cell body (which we interpret as being at least in part localised to fibrillar adhesions). Importantly, following refeeding, phospho-S6 (and a readout for newly synthesised proteins, HPG) are enriched in FA regions and also gradually increase towards the middle of the cell. Therefore, we cannot rule out the role of fibrillar adhesions under the cell body in mTORC1 activation, which we stated in the Discussion and cited the study by Rainero et al. (2015) Cell Rep. However, we believe that our experiments with FA disruption and constitutive targeting of RPTOR to FAs strongly implicate FAs as the key hub for mTORC1 activation.
2) The proximity biotinylation was carried out in osteosarcoma U2OS cells. Why is the validation data of mTOR and LAMP1 recruitment to paxillin positive adhesion with HeLa cells? The authors should consider showing this (and possibly localization to other adhesion types as well) in other cell types to demonstrate the generality/specificity of these data (there is some data from MEFs and MRC5 cells later, but would be nice to have a broader validation here).
The reason for using U2OS cells in our proteomics studies was that these were engineered to carry RPTOR-BioID. To validate localisation of mTORC1 activity, we now performed phospho-S6 stainings in U2OS, and also in Cos7 and HEK293 cell lines. In all of these models, we see a clear phospho-S6 signal in the proximity of FAs upon feeding (Fig. S3 B).
3) Figure 3a. Could the authors somehow demonstrate that the "significant increase in colocalisation between p-S6 and paxillin" is not just a reflection of having low p-S6 in the cell in starved conditions to having it abundantly throughout the cells in fed conditions?
To address this question in a formalised way, we used two approaches. First, we employed analyses similar to those used in our TORCAR experiments, i.e. by quantifying intensity of phospho-S6 in the regions of interest (ROI) defined by peripheral focal adhesions vs neighbouring ROI which are devoid of focal adhesions. New data presented in Fig. 3, D, E and M, clearly indicate that phospho-S6 (and a newly introduced readout for mTORC1, HPG) is significantly higher in the peripheral areas where focal adhesions accumulate compared to adjacent (control) regions. Second, we analysed fluorescence intensity profiles for paxillin and mTORC1 readouts from cell periphery to the nucleus. The histograms shown in Fig. 3, C and L, and Fig. S3 A indicate striking enrichment of phospho-S6 and HPG in the paxillin-positive areas. Based on both approaches we conclude that peripheral distribution of mTORC1 activity is not random but tightly associates with FA areas. 4) Figure 4a. How have the authors validated the specificity/ruled out off-target effects of the ARL8B siRNA?
The effects of our ARL8B siRNA on mTORC1 activity and specificity were validated in our previous paper (Korolchuk et al. (2011) Nat Cell Biol), using multiple siRNAs (including the ones used in this study) and rescue with ARL8B OE. We made a reference to the original study in the text.

5) Does ARL8B siRNA influence adhesion number or distribution in the cells.
We analysed distribution of paxillin staining in cells with/without ARL8B depletion and did not detect differences (Fig. S6 D), which is consistent with unperturbed growth factor signalling in ARL8B KD cells (Fig. 6 D). We also quantified the number and size of focal adhesions, which are not affected by ARL8B depletion (Fig. R3). 6) Starvation in other cell models has been shown to impact on the balance between peripheral focal and adhesions and elongated adhesions localising under the cell body.
The majority of our protocols involve comparison between starved cells and cells re-fed for 10 minutes. Even within this short stimulation period, we see an increase in the intensity of paxillin staining at the cell periphery (FAs), which is concomitant with the reduction of its intensity inside the cell (potentially associated with other types of complexes, such as fibrillar adhesions) (Fig. S3 A). This is consistent with previous reports (Rainero et al. Rep) and also supports the role for FAs as the hubs of nutrient sensing. 7) Figure 4G would be easier to interpret if starved and fed condition were shown like in the other experiments.
There is essentially no phospho-S6 signal in starved cells, which is shown repeatedly in the manuscript; therefore, to avoid duplication and save space, we only show FA inhibitor data for cells stimulated with nutrients (Fig. 5 D). However, starvation conditions have been shown in western blot data (Fig. S5 E) for consistency with other blots.
8) In the discussion the authors speculate that "Activation of mTORC1 at FAs by nutrients may therefore suppress autophagy of FA components". They should mention earlier work describing already that mTOR activity suppressed integrin and ECM protein uptake into spatially localised adhesion proximal lysosomes.
We have referenced the study indicating that inhibition of mTORC1 increases integrin uptake into adjacent lysosomes (Rainero et al. (2015) Cell Rep) in the Discussion. We also believe that the experiments suggested by the reviewer earlier allowed us to more clearly put our findings in the context of published literature.
Minor points: 9. Figure 2c. What is SDHA? Why was it chosen as control? It might be better to score the number of PLA spots/cell or the number of PLA spots overlapping with Paxillin rather than the % of PLA positive cells.
We have clarified that SDHA is a mitochondrial protein and used here as a negative/unrelated control as mTORC1 has not been previously shown to associate with mitochondria. Quantification of PLA as number of PLA dots/cell shows the same pattern as % of PLA-positive cells (Fig. 2 C).
10. Figure 3c would benefit from adding labelling in the images (to indicate what the green signal is). Page 7 top, this sentence may need to be edited "although at this point is cannot necessarily be completely ruled out." The text/figures have been modified.

Reviewer #3 (Comments to the Authors (Required)):
1. Although data analyses were performed at a state-of-the-art level, and the results are clearly represented the manuscript falls short in demonstrating under which physiological conditions the postulated "maximal mTORC1 activation" would be required.
The conditions where cells were fed with a full complement of amino acids and growth factors (i.e. FCS) were considered to induce maximal activation of mTOR. As the term could be misleading, we replaced it with "full activation of mTORC1" throughout the text.
2. In addition, although in the Arl8b KD experiment mTORC1 signaling is compromised, it remains to my opinion unclear if the results obtained can be really assigned to a lack of targeting focal adhesions. It could be equally well possible that Arl8b KD might interfere with mTORC1 signaling by increasing the amount of ragulator associating with BORC and thereby sequestering it away from Rags. If one took this into account, the conclusions would be completely different. Therefore, the authors could for instance artificially target mTORC1 to FA (eg. FAT domain on Rheb) and look if this would rescue the signaling defect detected upon Arl8b KD.
As suggested by the Reviewer, we generated constructs expressing RPTOR fused to the focal adhesion targeting (FAT) domain sequence of vinculin. The new data clearly demonstrate that, in agreement with our model and the Reviewer's prediction, forced targeting of mTORC1 to FAs rescues the signalling defect in ARL8B-depleted cells ( Fig. 6; and Fig. S6).
3. Importantly, depletion of focal adhesions, either by Talin KD or ROCK inhibitors, has a significant impact on the cytoskeleton. Chemotactic changes and/or alterations in Rac1 activity have previously been shown to regulate both mTORC1 and mTORC2. Therefore, presented experiments should be interpreted with caution since some of the effects observed might be independent of the here described focal adhesion dependent activation of mTORC1.
We acknowledged in the text the potential pleiotropic effect of interference with FA assembly as a limitation of this study. We then went on to rationalise that forced targeting of mTORC1 to FAs would provide another, and potentially more specific, way to investigate the role of FAs in nutrient sensing.
Minor points that should be revisited by the authors: 4. Figure 1: The results from the MS analysis of the BioID2-RPTOR pulldowns is in part surprising. Why is the log2 enrichment for the autophagy related proteins considerably higher than that observed for core mTORC1 subunits?
Autophagy proteins represent an important set of mTORC1 substrates, and since RPTOR plays a significant role in the recruitment of cargo to the mTORC1 complex (e.g. Ahmed et al. (2019) Sci Rep), it is not surprising that there is a high relative stoichiometry of autophagy proteins encountered by RPTOR.
5. Figure 3: The authors state that "FAs play an important role in the early events of mTORC1 activation in response to nutrients". Unfortunately, there are a few points unclear to this reviewer concerning the Leucine methyl ester stimulation experiment. In the legend from figure 3 it is mentioned that in panel E, "each data point represents a coverslip including several cells" whereas in panel F, "each data point represents a single cell". Also, in the legend from supplementary figure 3 panel A, "each data point represents a coverslip including several cells". Could the authors please explain why the analysis from the Leucine methyl ester stimulation was apparently performed differently from the remaining TORCAR biosensor experiments?
We thank the Reviewer for this comment and apologise for the technical mistake. In the revised manuscript, we unified data presentation in Fig. 3 and Fig. S3 in the way that, in all FRET-related graphs, each data point represents a coverslip.
Further, it would make the authors claim stronger if they could include an independent positive control, e.g. immunoblotting, to detect mTORC1 activation upon Leucine methyl ester stimulation. The quality of the phospho-4EBP1 antibody is significantly less than phospho-S6, and is particularly weak in response to amino acids alone (in the absence of serum). Therefore, instead of pursuing the use of phospho-4EBP1 antibody further, we removed phospho-4EBP1 data and focused on the TORCAR reporter, which is based on 4EBP1 phosphorylation by mTORC1, for more detailed kinetics and localisation analyses ( Fig. 3; and Fig. S3).
7. Figure 4b shows quantified integrated densities of pS6 signals in IF experiments.
Somewhat surprisingly the decrease of pS6 signal upon ARL8B knock-down is almost the same for peripheral and intracellular pS6. Such data presentation might give an impression to the readers that peripheral (10% of the cell area) vs intracellular regions were not properly defined. Considering an approximately 2.5-fold decrease in total cell number with peripheral p-S6 upon ARL8b depletion (panel c), one would expect a more specific effect for peripheral staining. The same criticism applies to corresponding Supplementary Figures 4b-c.
As indicated in the Methods section, the regions were identified by automated segmentation of the cell images, and therefore quantification is completely unbiased and precisely defined. The data show changes in the percentage of phospho-S6 signal, and the two regions (peripheral and intracellular) have been analysed separately. As such, we do not agree that there is a discrepancy in our data. Instead, it strongly suggests (and this is explicitly written in the manuscript) that preventing peripheral distribution of lysosomes not only prevents activation of mTORC1 at the periphery but also intracellularly. As such, we believe our data indicate that peripheral activation is also important for the overall activation of mTORC1 (which is also confirmed by immunoblotting; The old Fig. 4d showed western blot data for conditions that correspond to both, FCS stimulation (shown as IF data in old Fig. 4a-c) and amino acid stimulation (old Supplementary Fig. 4a-c). As such, one blot was used to illustrate changes in mTORC1 activation in response to both stimulation protocols.
In the revised manuscript we have also investigated the effect of ARL8B depletion in cells expressing RPTOR and RPTOR FAT . Since studies in two cell lines increased the amount of data significantly, we decided to focus on feeding these cell lines with complete media and not investigate amino acids alone. 1st Revision -Editorial Decision Thank you for submitting your revised manuscript entitled "mTORC1 activity is supported by spatial association with focal adhesions". Thank you for your patience with the re-review process. You will see that the reviewers are largely supportive of publication. We feel that you have done an adequate revision and also addressed the main points from Ref. #3, who was not available to rereview, so we would be happy to publish the manuscript in JCB pending final minor revisions that address the remaining points from Ref. #1 as well as changes necessary to meet our formatting guidelines (see details below). Please let us know if you would like to discuss any of the changes needed for publication.
To avoid unnecessary delays in the acceptance and publication of your paper, please read the following information carefully. Please be sure to reorganize the supplemental material to meet the limit of the format desired. A figure can use up to one entire page as long as all panels fit on the page. We feel that the paper is a good fit for the Article format and would suggest resubmitting in that format; we would be happy to further discuss as needed.
2) eTOC summary: A 40-word summary that describes the context and significance of the findings for a general readership should be included on the title page. The statement should be written in the present tense and refer to the work in the third person. to S1A (main image and magnifications) Molecular weight or nucleic acid size markers must be included on all gel electrophoresis. Please add molecular weight with unit labels on the following panels: 6D, S1B, S6BE 4) Statistical analysis: Error bars on graphic representations of numerical data must be clearly described in the figure legend. The number of independent data points (n) represented in a graph must be indicated in the legend. Statistical methods should be explained in full in the materials and methods. For figures presenting pooled data the statistical measure should be defined in the figure legends.
5) Materials and methods: Should be comprehensive and not simply reference a previous publication for details on how an experiment was performed. Please provide full descriptions in the text for readers who may not have access to referenced manuscripts.
-For all cell lines, vectors, constructs/cDNAs, etc. -all genetic material: please include database / vendor ID (e.g., Addgene, ATCC, etc.) or if unavailable, please briefly describe their basic genetic features *even if described in other published work or gifted to you by other investigators* -Please include species and source for all antibodies, including secondary, as well as catalog numbers/vendor identifiers if available.
-More information is needed to meet our policy: even if described in other published work, basic procedures must be described. Please double-check the M&M. For instance, more info is needed for FA isolation.
-Microscope image acquisition: The following information must be provided about the acquisition and processing of images: a. Make and model of microscope b. Type, magnification, and numerical aperture of the objective lenses c. Temperature d. imaging medium e. Fluorochromes f. Camera make and model g. Acquisition software h. Any software used for image processing subsequent to data acquisition. Please include details and types of operations involved (e.g., type of deconvolution, 3D reconstitutions, surface or volume rendering, gamma adjustments, etc.). 6) References: There is no limit to the number of references cited in a manuscript. References should be cited parenthetically in the text by author and year of publication. Abbreviate the names of journals according to PubMed.
-Please edit the reference formatting throughout to meet this style.

7)
A summary paragraph of all supplemental material should appear at the end of the Materials and methods section.
-Please include one brief descriptive sentence per item.
A. MANUSCRIPT ORGANIZATION AND FORMATTING: Full guidelines are available on our Instructions for Authors page, https://jcb.rupress.org/submission-guidelines#revised. **Submission of a paper that does not conform to JCB guidelines will delay the acceptance of your manuscript.** B. FINAL FILES: Please upload the following materials to our online submission system. These items are required prior to acceptance. If you have any questions, contact JCB's Managing Editor, Lindsey Hollander (lhollander@rockefeller.edu).
--An editable version of the final text (.DOC or .DOCX) is needed for copyediting (no PDFs).
--Cover images: If you have any striking images related to this story, we would be happy to consider them for inclusion on the journal cover. Submitted images may also be chosen for highlighting on the journal table of contents or JCB homepage carousel. Images should be uploaded as TIFF or EPS files and must be at least 300 dpi resolution.
**It is JCB policy that if requested, original data images must be made available to the editors. Failure to provide original images upon request will result in unavoidable delays in publication. Please ensure that you have access to all original data images prior to final submission.** **The license to publish form must be signed before your manuscript can be sent to production. A link to the electronic license to publish form will be sent to the corresponding author only. Please take a moment to check your funder requirements before choosing the appropriate license.** Thank you for your attention to these final processing requirements. Please revise and format the manuscript and upload materials within 7 days. If complications arising from measures taken to prevent the spread of COVID-19 will prevent you from meeting this deadline (e.g. if you cannot retrieve necessary files from your laboratory, etc.), please let us know and we can work with you to determine a suitable revision period.
Thank you for this interesting contribution, we look forward to publishing your paper in Journal of Cell Biology.