A high throughput SEC23 functional interaction screen reveals a role for focal adhesion and extracellular matrix signalling in the regulation of COPII subunit SEC23A
This article has been Reviewed by the following groups
Listed in
- Evaluated articles (Review Commons)
Abstract
Proteins that enter the secretory pathway are transported from their place of synthesis in the Endoplasmic Reticulum to the Golgi complex by COPII coated carriers. The components of the COPII transport machinery have been well characterized, but the network of proteins that regulate these components in response to extracellular cues have remained largely elusive. The discovery of such regulatory proteins is crucial for understanding how cells integrate extracellular and intracellular information to fine-tune membrane traffic in the secretory pathway. A key group of proteins that plays a central role in the regulation of membrane traffic is associated with the cytoskeleton. Using high throughput microscopy of the well-established VSVG transport from the ER to the plasma membrane, we comprehensively screened 378 cytoskeleton-associated proteins for their functional interaction with the COPII components SEC23A and SEC23B using a double knockdown approach. Ranking of the transport effectors identified 20 cytoskeleton-associated proteins as the strongest functional interactors of SEC23, most of them not previously associated with the secretory pathway. Knockdown of a subgroup of these interactors (FERMT2, MACF1, MAPK8IP2, NGEF, PIK3CA, ROCK1), associated with cell adhesion, not only induced changes in focal adhesions and the expression of adhesion-related genes but led to the specific downregulation of SEC23A, a SEC23 paralogue that has been implicated in the secretion of extracellular matrix (ECM) components. Furthermore, SEC23A downregulation could also be recapitulated by plating cells on ECM and was dependent on focal adhesions function. Altogether, our results identify a network of cytoskeleton-associated proteins connecting focal adhesions and ECM-related signalling with the gene expression of the COPII secretory machinery.
Article activity feed
-
Note: This rebuttal was posted by the corresponding author to Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Reply to the reviewers
1. General Statements [optional]
We thank the reviewers for their critical comments and suggestions. We are glad that the reviewers appreciated the quality of the data and the novel findings connecting the secretory trafficking machinery with extracellular matrix-related signaling.
2. Description of the planned revisions
Reviewer #1 (Evidence, reproducibility and clarity (Required)):
The manuscript by Jung et al reports on an interesting finding that focal adhesion signaling regulates the expression of Sec23A and thereby regulates COPII-dependent trafficking. The data presented a mostly solid and the finding itself is highly novel, as it tackles an area …
Note: This rebuttal was posted by the corresponding author to Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Reply to the reviewers
1. General Statements [optional]
We thank the reviewers for their critical comments and suggestions. We are glad that the reviewers appreciated the quality of the data and the novel findings connecting the secretory trafficking machinery with extracellular matrix-related signaling.
2. Description of the planned revisions
Reviewer #1 (Evidence, reproducibility and clarity (Required)):
The manuscript by Jung et al reports on an interesting finding that focal adhesion signaling regulates the expression of Sec23A and thereby regulates COPII-dependent trafficking. The data presented a mostly solid and the finding itself is highly novel, as it tackles an area of secretory trafficking that remains poorly understood, namely the connection between the ECM and secretion.
I will list below all comments that I have mixing both technical and conceptual topics:
**Technical issues:**
1-The authors should provide a better description of how the designed this siRNA library. What were the inclusion criteria for these 378 genes? I might have missed it, but I could not find this information easily.
Reply: The library has been designed in-house based on gene annotations and literature to include cytoskeleton structural proteins, motor proteins, and other associated and regulatory proteins. We will add this information in the Materials and Methods section.
2-Figure 2: I know this is challenging for EM images, but is there a way the authors could quantify these data? How many images were looked at? What was the average width of ER cisterne?
__Reply: __We will provide image quantifications and statistics
3-Figure 4: I think that the characterization of the FA phenotype is a bit underdeveloped. There is no quantification of these data. Is the size of FA changing? Is the number of FA per cell changing? Is the length of FAs changing? I think that more work is needed to increase the confidence in these data.
I could also not easily see what type of cells these are. __A better description of this experiment is also required. Also, how many cells were analyzed. __I think it is important that this experiment is done with a sufficient number of cells to increase the confidence in the data.
__Reply: __We agree with the reviewer that our observations regarding the focal adhesion (FA) phenotype will benefit from image quantification and we intend to include this in the revised manuscript. All FA experiments were performed on HeLa cells. We will update the materials and methods sections to better describe this experiment.
**Conceptual issues:**
1-The finding that focal adhesion signaling negatively affects ER-export is surprising, because cancer cells that grow on stiff substrates have more focal adhesions and are more invasive and migratory. Both migration and invasion are expected to depend on ER-export. Although the authors did not formally test Sec23A expression under different stiffnesses, I would expect that stiff substrates would lower Sec23A expression and thereby negatively affect ER-export. It would certainly increase the breadth of this work to include data like this and to also discuss this highly surprising finding. However, it is of course the decision of the authors and the editors to decide whether such an experiment would benefit the entire story.
__Reply: __In this work, we have shown that cells plated on ECM or matrigel have decreased SEC23A expression compared to control cells. We have also shown that inhibition of FA kinase leads to an increase in SEC23A expression (Figure 5). Whether this translates into a change in ER transport, is a fair point that we will address in the revision. Regarding stiffness, we have done a preliminary experiment that shows that cells plated on a soft synthetic substrate have less SEC23A than cells plated on plastic.This goes in line with our ECM experiments because Matrigel and fibroblast-derived ECM are softer than plastic.
2-The authors postulate that this novel mechanism could be part of a feedback loop. If this were the case one would expect the acute effect of FA to increase ER-export (or secretion) and the negative feedback will then reduce secretion. However, the acute effect of FA is not addressed in this manuscript. In order to postulate a feedback loop, the authors would need to test the individual nodes of this loop.
__Reply: __The question appears to be whether an acute effect on FA would affect the expression of SEC23A and therefore ER transport. If by the acute effect the reviewer means a pharmacological manipulation, we have shown that upon treatment with the FAK inhibitor the expression of SEC23A increases (Fig 5A). Whether this increase in SEC23A expression translates into a corresponding increase in ER transport remains to be seen. This will be tested in our revised manuscript as mentioned above in reply to point # 1.
Our data encouraged us to propose a hypothetical feedback loop that would connect the deposition of ECM through the expression of SEC23A. We will have more data to support (or reject) this idea once we do the transport experiments as mentioned above. However, we think that a full characterization of this hypothetical loop by testing individual nodes is beyond the scope of this manuscript
Reviewer #1 (Significance (Required)):
I think that the basic finding of this manuscript is highly novel, by showing the impact of the ECM and focal adhesions on COPII-dependent trafficking. I think that this will not only appeal to people from the trafficking community, but also to people working on cell migration and on mechanobiology. The work in its current form does not require much extra efforts (max. 3 month). However, if the authors would decide to increase the breadth of data, they would require 3-6 months.
Reply: We thank reviewer #1 for the comments. We also believe that this story will appeal to a broader audience and would help to bridge the gap between membrane trafficking and mechanobiology communities.
**Referees cross-commenting**
I went through the comments of the two other reviewers and agree with their verdict. Some extra work on the characterization of the early secretory pathway would be good. Both reviewers provided a nice catalogue of possible experiments to choose from.
Reply: We have characterized the early secretory pathway in terms of ER exit sites, Beta-COP, and Golgi morphology (FIG. 2B-H and S1A-B). Together, these data strongly characterize the nature of ER-block. Moreover, the finding that our interactors affect the expression of SEC23A allows us to explain mechanistically why an ER transport block occurs. This is further strengthened by the rescue experiments (FIG. 3F). We believe that further characterization of the secretory pathway will not contribute substantially to the main message of this manuscript.
Reviewer #2 (Evidence, reproducibility and clarity (Required)):
The manuscript by Jung et al which based on a targeted siRNA screen, demonstrates regulation of SEC23A (component of the SEC23 complex of the COP coat) levels at transcriptional level downstream of focal adhesion signaling. By regulating siRNA mediated downregulation, the authors were able to identify proteins which either increased or decreased traffic of VSVG through the secretory pathway when combined with downregulation in the levels of with either SEC23A or SEC23B. Authors have focused on a group of SEC23B functional interactors, downregulation of which shows them increased size of focal adhesions which also downregulate SEC23A levels, thus providing an explanation for reduced secretory traffic. Authors further show that plating cells on fibronectin or Matrigel, which activate Focal adhesion kinase signaling also results in downregulation of SEC23A transcript levels. The screen is conducted in a well-controlled manner for most parts with a clear explanation of the analysis routines and the data presentation if of very good quality. Most important results have been validated by more than one experimental strategy which lends substantial confidence to the findings. The results also open further avenues for understanding the transcriptional regulation in different physiological and disease contexts.
There are certain issues, which the authors should address with regards to controls and some conflicting observations with published results with respect the phenotypes associated with downregulating proteins on focal adhesions size. Additionally, authors don't tie the ends by monitoring secretory traffic in cells grown on different matrices but include it in the model. Addressing/explaining these issues could improve this manuscript and the model may have to be tweaked a bit.
**Major comments:**
1)I wonder why the authors only used siRNA control in their screen when the effects are scored in context of double knockdown fashion in combination with mild knockdown of SEC23A and SEC23B to get functional interactors. Control siRNA in combination with SEC23A and SEC23B should have been two ideal negative controls in the screen. Nevertheless, in data presented Figure 1E and whole of Figure 2, using control siRNA in combination with SEC23B siRNA would have been ideal control to show that the combination does not induce any trafficking defects which could impact the findings of the study. Hence, a few of the data presented from some of these figures should have sicontrol+SEC23B siRNA combination as a control.
Reply: There seems to be a misunderstanding. In the screen, the negative controls are only used as a reference as the scoring is based on a 5X5 matrix centered on the siRNA of interest. This is done to overcome possible plate effects and to normalize data across different biological replicas. As seen in figure 1B, the negative controls (Control siRNA or Control siRNA + SEC23A siRNA or Control siRNA + SEC23B siRNA are very close to 0 (but not exactly 0) as they were not used in the normalization process. It is important to mention that all single knockdowns also contain our control siRNA to keep the same final siRNA concentration in single and double knockdowns. In Fig 1E we will include the images from Control + SEC23A siRNAs and Control + SEC23B siRNA as a reference. For Figure 2 all except 2A and 2H have the single knockdowns as controls.
2)What is the identity of post-ER structures which authors refer to in Figure 2A? Could the images represent VSVG concentrated at ER exit sites? Authors should stain with markers for ERES to see if the VSVG puncta colocalize with it.
__Reply: __We have done the experiment, and indeed these structures colocalize with an ER exit site marker (SEC31A). We intend to include this data into the revised manuscript. Our observations are in agreement with what is known in the literature about VSVG transport.
3)Based on RNA sequencing results, authors chose to follow up on SEC23A levels in background of siRNA knockdown of components (like MACF1, ROCK1, FERMT2 etc.) which regulate Focal adhesions in cells and show that there is a reduction in both transcript and protein levels of SEC23A. In images shown in Figure 2B and Figure 2C, levels or SEC31A and β-Cop1 are reduced. Authors should test using qPCR and western blots whether there is a downregulation SEC31A, β-Cop1 and SEC23B in siRNA knockdowns of MACF1, ROCK1, FERMT2 etc. It would provide new insights if there were a co-regulation of secretory machinery to modulate the secretory traffic in response to Focal Adhesion based signaling.
__Reply: __Our transcriptomics data (FIG 3C and Table 5) shows that SEC31A and COPB1 mRNAs are not altered upon any of the knockdowns. For SEC23B, we observed only a slight decrease in ROCK1 knockdown. This data suggests that a co-regulation of the secretory machinery might not be present. Instead, the curation of secretory pathway genes in our transcriptome data shows that SEC23A is the only commonly differentially expressed gene.
4)Most major concern in this manuscript surrounds around results presented in Figure 4C. Authors show that in response to all the knockdowns, they see more focal adhesions as monitored by Vinculin staining and this along with the experiments with cells plated on Matrigel and Fibronectin arrive at the conclusion that increased Focal adhesion signaling downregulates SEC23A levels which presumably modulates secretory traffic. I am not an expert on Focal adhesions but based on my understanding of the literature on that topic, downregulation of ROCK1, FEMRT2 disrupts focal adhesions. (See: Theodosiou et. al., Elife, 2016 or Lock et. al., Plos One, 2012 for example). How do authors explain their results in siRNA knockdown of ROCK1 and FEMRT2 which leads to an increased size of focal adhesions which seems contradictory to the published results? To clarify these results authors should test phosphorylation of FAK in their siRNA backgrounds which is another read out of focal adhesion signaling.
The experiments from cells grown on Fibronectin and Matrigel favor the argument which authors put forth, but authors may have to tweak the model a bit based on FAK phosphorylation and FAK signaling in context of above-mentioned knockdowns.
__Reply: __Based on the images for vinculin staining, in our current manuscript we propose that changes in FAs occur upon knocking down our interactors. In our revised manuscript we will provide a more robust quantitative assessment of those changes (change in number, size, or intensity) as mentioned in our reply to Reviewer #1.
As for the discrepancies in the relation of FA phenotype upon depletion of ROCK1 and FERMT2, we want to point out that this effect depends on the cell type used. For instance, the papers listed by the reviewer here use fibroblasts and keratinocytes respectively while we have used Hela Kyoto cells which are epithelial in nature. Another example is that while in fibroblasts depletion of FERMT2 leads to a rounded morphology and almost an absence of FAs (Theodosiou et. al., Elife, 2016), in podocytes (Qu et al JCS, 2011), it leads to fewer FAs but an increase in their size. Nonetheless, this is a very keen observation from the reviewer and we will address this point in our revised manuscript discussion.
5)What happens to VSVG traffic or RUSH-Cadherin traffic when cells are plated on Matrigel and Fibronectin? Reduction in secretory traffic of these is an important experiment which is missing to close the loop and validate the model presented. Authors must test these experiments either with cells grown on matrix alone or in combination with siRNA to SEC23B. Authors should also monitor ERES and transport carriers in this background.
__Reply: __We agree with the reviewer and intend to perform these experiments.
6)This is not such a major issue, but it would be good to see a comparison in SEC23A levels in siRNA knockdown condition in comparison to those when cells are grown on different substrates and in ROCK1, FEMRT2 knockdowns (blots of which authors already have in this manuscript).
__Reply: __We will assess the level of SEC23A at the protein level for cells plated on matrigel or Fibroblast-derived ECM.
**Minor comments:**
1)Scale bars are missing in EM images in Figure 2H.
__Reply: __We will add the scales in our EM images
2)Show molecular weight markers in Western blots in main figure 3E and supplementary figure S1E.
__Reply: __We will add molecular weight markers in our Western-Blots
Reviewer #2 (Significance (Required)):
I have looked at the manuscript from through the lens of a cell biologist as that is predominantly my area of expertise. In that respect I find the screen conducted by authors particularly interesting as they aim to connect how extracellular cues regulate the secretory pathway. A screen seems justified as there is no comprehensive understanding linking the two above-mentioned processes. Authors have done a functional interaction screen and analyzed a lot of images to identify candidates which either increase or decrease secretory traffic in combination with SEC23A and SEC23B. Such a functional screen has helped authors identify candidates which were otherwise missed in single siRNA knockdowns in their previous work from 2012. This definitely opens up interesting avenues to test the candidates identified in the screen in different physiological contexts and in disease as also the transcriptional program connecting Focal adhesion signaling with the regulation of components governing secretion. Such functional interaction screens could also be employed to identify crosstalk of different cellular processes with the regulation secretory pathway at ER as well as at the Golgi apparatus.
Reply: We thank reviewer #2 for the comments. As we mentioned in our reply to reviewer #1, we strongly believe that these results will encourage further research at the crossroads of membrane trafficking and mechanobiology.
**Referees cross-commenting**
I agree with the comments from both the referees that the manuscript is very interesting, most experiments are well controlled, but the quantification of focal adhesion phenotype in knockdowns need to be done in an extensive manner and secretion phenotypes need to measured upon plating cells on different matrix to validate the model presented.
__Reply: __These two experiments will be included in our revision
Reviewer #3 (Evidence, reproducibility and clarity (Required)):
**Summary**
The authors use a synchronized cargo release assay following codepletion of either Sec23 paralog with cytoskeletal and associated proteins to identify potential functional interactions between COPII trafficking and the cytoskeleton. This screen yields a number of Sec23b functionally interacting molecules that stall cargo trafficking to various degrees within the secretory pathway upon codepletion, and in the case of MACF1 reduce ERES number despite not physically interacting. Depletion of the majority of the identified Sec23b functional interactors alone surprisingly caused the downregulation of Sec23a at the mRNA and protein levels, and cargo trafficking could be partially or fully rescued by Sec23a overexpression depending on the codepleted cytoskeletal factor. RNA-seq enrichment analysis and imaging of a focal adhesion marker suggest that genes involved in cell adhesion were differentially regulated following depletion of the cytoskeletal functional interactors. Finally, the authors show that Sec23a expression levels are reduced when cells are cultured on dishes with high amounts of ECM to induce focal adhesions, and that inhibition of focal adhesion kinase can rescue Sec23a expression levels.
**Major comments**
#1 The authors successfully implicate a group of cytoskeletal proteins and their actions at focal adhesions in negatively regulating Sec23a expression levels and COPII trafficking. This description of a shared, novel mode of COPII transcriptional regulation by cytoskeletal factors is convincingly shown to be at least a contributor to the delayed trafficking in the presence of focal adhesions. In general, the data are reproducible and use appropriate statistical analysis. However, a more robust description of the architecture of early secretory pathway would be beneficial, especially in the case of MACF1 codepletion which cannot be fully rescued by Sec23a-YFP overexpression. In contrast, trafficking during codepletion of FERMT2 is fully rescued by Sec23a-YFP despite both MACF1 and FERMT2 showing similar loss of Sec23a mRNA levels upon codepletion. This data suggests that while the trafficking delay in FERMT2 codepletion might be exclusively due to reduced Sec23a expression levels, there are likely additional causes for the trafficking delay observed in MACF1 codepletion.
__Reply: __We thank the reviewer for the appreciation of our results and the importance they might bear for the field. The reviewer has very neatly highlighted that each of our interactor hits might have roles in the secretory pathway beyond the ER or independent of the expression levels of SEC23A. This phenomenon could also explain the differential rescue of the arrival of VSVG at the plasma membrane upon SEC23A overexpression in FERMT2 and MACF1 knockdowns (FIG 3F). For instance, MACF1 has been involved in Golgi to Plasma Membrane transport as well (Kakinuma et al. Exp. Cell Res. 2004, Burgo et al. Dev. Cell 2012). So a possibility is that SEC23A overexpression rescues only ER to Golgi transport but the lack of rescue in the compartment between Golgi and plasma membrane independent of SEC23A expression levels would result in reduced rescue In the case of MACF1 compared to FERMT2. To support this, in our revised manuscript, we will provide example images from the experiment.
Nonetheless, we agree that these are very important observations from Reviewer #3 and warrant a detailed discussion in the light of other interactors as well, which we intend to highlight in our revised manuscript.
#2 While there is indeed a reduction in the number of ERESs following MACF1 codepletion, the authors report an even more dramatic reduction in 'transport intermediates / cell' as marked by COPI. However, as recent cyro-EM analysis of ERESs has definitively show, COPI exists stably at ERGIC membranes (1). Thus, an alternative possibility for the more dramatic reduction of COPI sites compared to Sec31a sites in Figures 2B-E is that ERGIC membranes are destabilized following MACF1 codepletion in a manner independent of Sec23a expression, and this destabilization compounds with reduced ERES number to ultimately delay trafficking. To more directly determine whether ERGIC membranes stability is regulated by MACF1, the authors should compare COPI and ERGIC-53 staining among MACF1 codepleted and FERMT2 codepleted cells with and without Sec23a-YFP overexpressed to levels that rescue cargo trafficking. If Sec23a-YFP restores the number of ERGIC puntae marked by these stains in FERMT2 but not MACF1 codepleted cells, it would suggest a role for MACF1 in forming or stabilizing ERGIC membranes which are known to associate with microtubules and WHAMM, an actin nucleator. Additionally, it would be useful to costain COPII with COPI or ERGIC-53 in control, MACF1 depleted, MACF1 codepleted, and MACF1 codepleted and Sec23a-YFP rescued cells to determine their colocalization. COPII and ERGIC membranes should be almost entirely coupled and juxtaposed in control cells and may be decoupled upon loss of MACF if plays a role in ERGIC membrane localization and stability. These proposed experiments are relevant because ERGIC membranes are sites of COPII cargo delivery and changes in ERGIC stability or localization would suggest an additional mechanism for cytoskeletal regulation of COPII trafficking. These immunofluorescence studies should be straightforward and completed in a few weeks.
__Reply: __Although a possible additional role of MACF1 in the organisation of early secretory pathway, stability of ERES, etc., independent of the expression of SEC23A is interesting on its own, we believe that an extensive characterization of these possible roles/ pathways as proposed by the reviewer is beyond the scope this manuscript.
#3 The choice to use VSVG and E-Cadherin for the synchronized release assays unfortunately convolutes interpreting the 'transport ratios' used by the authors to compare the effects of the various codepletions. Each protein progresses beyond the Golgi during secretion, and the authors choose to calculate the ratio of cargo intensity at the plasma membrane normalized to the total cellular cargo. This means that the synchronized release assays and calculated 'transport ratios' assay not only ER to Golgi trafficking, but also trafficking from the Golgi to the plasma membrane. In instances where Sec23a-YFP overexpression does not fully rescue the codepletion, it is possible that additional trafficking delays occur during Golgi to plasma membrane trafficking that cause the 'transport score' to decrease. Thus, the 'transport score' as the authors calculate it is needlessly nonspecific to COPII trafficking and should not be used to compare the codepletions for COPII functional interactors.
__Reply: __We agree that the “transport score” used here and in our previous genome-wide screen (Simpson et. al Nat. Cell Biol. 2012) does not allow us to distinguish between the individual transport substeps in the transport of VSVG from the ER to the plasma membrane. However, as we see in Fig 1E, the proteins that we have decided to follow in more detail in this study do have a clear ER transport block phenotype (except for CRKL). So for 6 out of 7 of these proteins, the images clearly show that the decrease in the “transport score” is due to a decreased ER to Golgi transport.
*#4 To mitigate unwanted contributions of post-COPII trafficking events from altering 'transport scores,' the authors should use a cargo for synchronized release assays that does not progress past the Golgi such as α-Mannosidase II and quantify a ratio of the perinuclear cargo signal to whole cell signal. Ideally, the screen would be repeated with a more appropriate cargo generating new 'transport scores' for the full list of cytoskeletal proteins. However, this may not be feasible, and as such 'transport scores' based on a Golgi resident protein should at least be produced for the 7 Sec23b functional interactors featured in this manuscript. These Golgi 'transport scores' would add much needed quantification of ER to Golgi transport delays that currently can only be inferred from the representative images in Figure 1E, which unfortunately show significant heterogeneity among cells from the same image. The authors should also explicitly state that any 'transport score' from a synchronous release assay using a cargo destined for the plasma membrane will take into account trafficking rate changes due not only to COPII, but also COPI from the ERGIC to the Golgi, and transport carriers departing from the TGN. These synchronized release assays would likely take between a few weeks to a few months depending on their ability to automate image analysis. *
Reply: We consider that having a “Golgi transport score” won't add any new information as the proteins that we have chosen to follow are the ones that show a strong ER-block phenotype. However, we agree that such a “Golgi score” would indeed be useful if one would like to study other interactors, for instance, the ones that induce transport acceleration.
Also, we don't expect all cells to behave similarly as the level of knockdown might be slightly different or because of the cell to cell variability. Even in control conditions (no knockdown), this heterogeneity is evident. As suggested by the reviewer, in our revised manuscript we will explicitly state that a change in the transport scores could mean a change in any sub-step of the transport from the ER to the PM in our assay.
**Minor comments**
It would be useful for the authors to quantify the number of focal adhesions present from Vinculin stains from Figure 4C and 5C instead of just showing representative images. It would be interesting to determine if there is a meaningful relationship between focal adhesion number induced by the codepletions or tissue culture coating and Sec23a expression levels like in Figure 3D. Generally, the figures, text, and references were appropriate.
__Reply: __As also pointed out by the other reviewers we will quantify the FA changes
Reviewer #3 (Significance (Required)):
In recent years, significant effort has been devoted to elucidating mechanisms by which COPII trafficking is modulated in response to cellular cues. These studies have revealed that changes in nutrient availability, growth factors, ER stress, autophagy, and T-cell activation all cause changes in COPII trafficking via unique gene expression, splicing, or post-translational control (2-7). This work elucidates a novel mechanism of transcriptional control driven by focal adhesions. Additionally, it provides a number of potentially useful Sec23a and Sec23b functional interactors among cytoskeletal factors for further study. These unexplored factors may have unique mechanism of COPII regulation that could contribute to our understanding ER export modulation. Altogether, this and similar works are building an increasingly complex set of regulatory pathways that when integrated ultimately dictate COPII trafficking kinetics.
The reported findings are not only relevant to those who study COPII trafficking, but also other fields where secretion is studied in the context of the ECM. This work would suggest that secretion of factors involved in crosstalk between cells, including in tumors, is likely to be controlled by the interactions of cells with ECM.
__Reply: __We thank reviewer #3 for the comments and insightful discussion about the limitations of our assay that we will highlight in the revised manuscript and in general for the insight into the early secretory pathway regulation. Furthermore their explicit summary of how our study could bridge COPII trafficking, ECM signaling and the relevance to various pathophysiologies is highly appreciated.
Expertise keywords: cell biology, light microscopy, membrane trafficking
References
*1.Weigel A V., Chang CL, Shtengel G, Xu CS, Hoffman DP, Freeman M, et al. **ER-to-Golgi protein delivery through an interwoven, tubular network extending from ER. *Cell. 2021 Apr;184(9):2412-2429.e16.
*2.Farhan, H., Wendeler, M. W., Mitrovic, S., Fava, E., Silberberg, Y., Sharan, R., Zerial, M., & Hauri, H. P. (2010). *MAPK signaling to the early secretory pathway revealed by kinase/phosphatase functional screening. Journal of Cell Biology, 189(6), 997-1011.
*3.Zacharogianni, M., Kondylis, V., Tang, Y., Farhan, H., Xanthakis, D., Fuchs, F., Boutros, M., & Rabouille, C. (2011). ERK7 is a negative regulator of protein secretion in response to amino-acid starvation by modulating Sec16 membrane association. *EMBO Journal, 30(18), 3684-3700.
*4.Lillmann, K.D., V. Reiterer, F. Baschieri, J. Hoffmann, V. Millarte, M.A. Hauser, A. Mazza, N. Atias, D.F. Legler, R. Sharan, et al 2015. *Regulation of Sec16 levels and dynamics links proliferation and secretion. J. Cell Sci. 128:670-682.
5.Liu, L., Cai, J., Wang, H., Liang, X., Zhou, Q., Ding, C., Zhu, Y., Fu, T., Guo, Q., Xu, Z., Xiao, L., Liu, J., Yin, Y., Fang, L., Xue, B., Wang, Y., Meng, Z. X., He, A., Li, J. L., ... Gan, Z. (2019). Coupling of COPII vesicle trafficking to nutrient availability by the IRE1α-XBP1s axis. Proceedings of the National Academy of Sciences of the United States of America, 116(24), 11776-11785.
6.Jeong, Y.-T., Simoneschi, D., Keegan, S., Melville, D., Adler, N. S., Saraf, A., Florens, L., Washburn, M. P., Cavasotto, C. N., Fenyö, D., Cuervo, A. M., Rossi, M., & Pagano, M. (2018). The ULK1-FBXW5-SEC23B nexus controls autophagy. ELife, 1-25.
7.Wilhelmi, I., Kanski, R., Neumann, A., Herdt, O., Hoff, F., Jacob, R., Preußner, M., & Heyd, F. (2016). Sec16 alternative splicing dynamically controls COPII transport efficiency. Nature Communications, 7, 12347. https://doi.org/10.1038/ncomms12347
3. Description of the revisions that have already been incorporated in the transferred manuscript
4. Description of analyses that authors prefer not to carry out
Reviewer #3 suggested to robustly characterise the early secretory pathway, in response to the depletion of our interactors, for instance, the role of MACF1 in the organization and the stability of ERES. This view is also supported by reviewer #1. However, in our revised manuscript we would like to focus more on the novel aspect of our study (as highlighted by all the reviewers), namely how ECM signaling and changes in FAs affect SEC23A and possibly ER transport. For this, we would like to present a more quantitative outlook of the FA phenotype and concentrate on the transport experiments. The reason for not dwelling into a more extensive characterization of the early secretory pathway is that these experiments are very interesting on their own, and merit a separate study that would deconvolve in detail the individual trafficking steps, and their relation to SEC23A expression, ERES stability, and ECM signaling.
Reviewer #2 suggested that to better characterize the FA phenotype and solve the apparent discrepancies between our data and the literature, we could test FAK phosphorylation. As we mentioned in our reply to this point, we think that most of the discrepancies arise from the different cell types used. Nevertheless, we agree that a quantitative approach is needed for a better characterisation of FA phenotype, therefore we intend to perform quantification of the vinculin stainings.
-
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #3
Evidence, reproducibility and clarity
Summary
The authors use a synchronized cargo release assay following codepletion of either Sec23 paralog with cytoskeletal and associated proteins to identify potential functional interactions between COPII trafficking and the cytoskeleton. This screen yields a number of Sec23b functionally interacting molecules that stall cargo trafficking to various degrees within the secretory pathway upon codepletion, and in the case of MACF1 reduce ERES number despite not physically interacting. Depletion of the majority of the identified Sec23b functional interactors alone surprisingly caused the downregulation of Sec23a at the mRNA and …
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #3
Evidence, reproducibility and clarity
Summary
The authors use a synchronized cargo release assay following codepletion of either Sec23 paralog with cytoskeletal and associated proteins to identify potential functional interactions between COPII trafficking and the cytoskeleton. This screen yields a number of Sec23b functionally interacting molecules that stall cargo trafficking to various degrees within the secretory pathway upon codepletion, and in the case of MACF1 reduce ERES number despite not physically interacting. Depletion of the majority of the identified Sec23b functional interactors alone surprisingly caused the downregulation of Sec23a at the mRNA and protein levels, and cargo trafficking could be partially or fully rescued by Sec23a overexpression depending on the codepleted cytoskeletal factor. RNA-seq enrichment analysis and imaging of a focal adhesion marker suggest that genes involved in cell adhesion were differentially regulated following depletion of the cytoskeletal functional interactors. Finally, the authors show that Sec23a expression levels are reduced when cells are cultured on dishes with high amounts of ECM to induce focal adhesions, and that inhibition of focal adhesion kinase can rescue Sec23a expression levels.
Major comments
The authors successfully implicate a group of cytoskeletal proteins and their actions at focal adhesions in negatively regulating Sec23a expression levels and COPII trafficking. This description of a shared, novel mode of COPII transcriptional regulation by cytoskeletal factors is convincingly shown to be at least a contributor to the delayed trafficking in the presence of focal adhesions. In general, the data are reproducible and use appropriate statistical analysis. However, a more robust description of the architecture of early secretory pathway would be beneficial, especially in the case of MACF1 codepletion which cannot be fully rescued by Sec23a-YFP overexpression. In contrast, trafficking during codepletion of FERMT2 is fully rescued by Sec23a-YFP despite both MACF1 and FERMT2 showing similar loss of Sec23a mRNA levels upon codepletion. This data suggests that while the trafficking delay in FERMT2 codepletion might be exclusively due to reduced Sec23a expression levels, there are likely additional causes for the trafficking delay observed in MACF1 codepletion.
While there is indeed a reduction in the number of ERESs following MACF1 codepletion, the authors report an even more dramatic reduction in 'transport intermediates / cell' as marked by COPI. However, as recent cyro-EM analysis of ERESs has definitively show, COPI exists stably at ERGIC membranes (1). Thus, an alternative possibility for the more dramatic reduction of COPI sites compared to Sec31a sites in Figures 2B-E is that ERGIC membranes are destabilized following MACF1 codepletion in a manner independent of Sec23a expression, and this destabilization compounds with reduced ERES number to ultimately delay trafficking. To more directly determine whether ERGIC membranes stability is regulated by MACF1, the authors should compare COPI and ERGIC-53 staining among MACF1 codepleted and FERMT2 codepleted cells with and without Sec23a-YFP overexpressed to levels that rescue cargo trafficking. If Sec23a-YFP restores the number of ERGIC puntae marked by these stains in FERMT2 but not MACF1 codepleted cells, it would suggest a role for MACF1 in forming or stabilizing ERGIC membranes which are known to associate with microtubules and WHAMM, an actin nucleator. Additionally, it would be useful to costain COPII with COPI or ERGIC-53 in control, MACF1 depleted, MACF1 codepleted, and MACF1 codepleted and Sec23a-YFP rescued cells to determine their colocalization. COPII and ERGIC membranes should be almost entirely coupled and juxtaposed in control cells and may be decoupled upon loss of MACF if plays a role in ERGIC membrane localization and stability. These proposed experiments are relevant because ERGIC membranes are sites of COPII cargo delivery and changes in ERGIC stability or localization would suggest an additional mechanism for cytoskeletal regulation of COPII trafficking. These immunofluorescence studies should be straightforward and completed in a few weeks.
The choice to use VSVG and E-Cadherin for the synchronized release assays unfortunately convolutes interpreting the 'transport ratios' used by the authors to compare the effects of the various codepletions. Each protein progresses beyond the Golgi during secretion, and the authors choose to calculate the ratio of cargo intensity at the plasma membrane normalized to the total cellular cargo. This means that the synchronized release assays and calculated 'transport ratios' assay not only ER to Golgi trafficking, but also trafficking from the Golgi to the plasma membrane. In instances where Sec23a-YFP overexpression does not fully rescue the codepletion, it is possible that additional trafficking delays occur during Golgi to plasma membrane trafficking that cause the 'transport score' to decrease. Thus, the 'transport score' as the authors calculate it is needlessly nonspecific to COPII trafficking and should not be used to compare the codepletions for COPII functional interactors.
To mitigate unwanted contributions of post-COPII trafficking events from altering 'transport scores,' the authors should use a cargo for synchronized release assays that does not progress past the Golgi such as α-Mannosidase II and quantify a ratio of the perinuclear cargo signal to whole cell signal. Ideally, the screen would be repeated with a more appropriate cargo generating new 'transport scores' for the full list of cytoskeletal proteins. However, this may not be feasible, and as such 'transport scores' based on a Golgi resident protein should at least be produced for the 7 Sec23b functional interactors featured in this manuscript. These Golgi 'transport scores' would add much needed quantification of ER to Golgi transport delays that currently can only be inferred from the representative images in Figure 1E, which unfortunately show significant heterogeneity among cells from the same image. The authors should also explicitly state that any 'transport score' from a synchronous release assay using a cargo destined for the plasma membrane will take into account trafficking rate changes due not only to COPII, but also COPI from the ERGIC to the Golgi, and transport carriers departing from the TGN. These synchronized release assays would likely take between a few weeks to a few months depending on their ability to automate image analysis.
Minor comments
It would be useful for the authors to quantify the number of focal adhesions present from Vinculin stains from Figure 4C and 5C instead of just showing representative images. It would be interesting to determine if there is a meaningful relationship between focal adhesion number induced by the codepletions or tissue culture coating and Sec23a expression levels like in Figure 3D. Generally, the figures, text, and references were appropriate.
Significance
In recent years, significant effort has been devoted to elucidating mechanisms by which COPII trafficking is modulated in response to cellular cues. These studies have revealed that changes in nutrient availability, growth factors, ER stress, autophagy, and T-cell activation all cause changes in COPII trafficking via unique gene expression, splicing, or post-translational control (2-7). This work elucidates a novel mechanism of transcriptional control driven by focal adhesions. Additionally, it provides a number of potentially useful Sec23a and Sec23b functional interactors among cytoskeletal factors for further study. These unexplored factors may have unique mechanism of COPII regulation that could contribute to our understanding ER export modulation. Altogether, this and similar works are building an increasingly complex set of regulatory pathways that when integrated ultimately dictate COPII trafficking kinetics.
The reported findings are not only relevant to those who study COPII trafficking, but also other fields where secretion is studied in the context of the ECM. This work would suggest that secretion of factors involved in crosstalk between cells, including in tumors, is likely to be controlled by the interactions of cells with ECM.
Expertise keywords: cell biology, light microscopy, membrane trafficking
References
1.Weigel A V., Chang CL, Shtengel G, Xu CS, Hoffman DP, Freeman M, et al. ER-to-Golgi protein delivery through an interwoven, tubular network extending from ER. Cell. 2021 Apr;184(9):2412-2429.e16.
2.Farhan, H., Wendeler, M. W., Mitrovic, S., Fava, E., Silberberg, Y., Sharan, R., Zerial, M., & Hauri, H. P. (2010). MAPK signaling to the early secretory pathway revealed by kinase/phosphatase functional screening. Journal of Cell Biology, 189(6), 997-1011.
3.Zacharogianni, M., Kondylis, V., Tang, Y., Farhan, H., Xanthakis, D., Fuchs, F., Boutros, M., & Rabouille, C. (2011). ERK7 is a negative regulator of protein secretion in response to amino-acid starvation by modulating Sec16 membrane association. EMBO Journal, 30(18), 3684-3700.
4.Lillmann, K.D., V. Reiterer, F. Baschieri, J. Hoffmann, V. Millarte, M.A. Hauser, A. Mazza, N. Atias, D.F. Legler, R. Sharan, et al 2015. Regulation of Sec16 levels and dynamics links proliferation and secretion. J. Cell Sci. 128:670-682.
5.Liu, L., Cai, J., Wang, H., Liang, X., Zhou, Q., Ding, C., Zhu, Y., Fu, T., Guo, Q., Xu, Z., Xiao, L., Liu, J., Yin, Y., Fang, L., Xue, B., Wang, Y., Meng, Z. X., He, A., Li, J. L., ... Gan, Z. (2019). Coupling of COPII vesicle trafficking to nutrient availability by the IRE1α-XBP1s axis. Proceedings of the National Academy of Sciences of the United States of America, 116(24), 11776-11785.
6.Jeong, Y.-T., Simoneschi, D., Keegan, S., Melville, D., Adler, N. S., Saraf, A., Florens, L., Washburn, M. P., Cavasotto, C. N., Fenyö, D., Cuervo, A. M., Rossi, M., & Pagano, M. (2018). The ULK1-FBXW5-SEC23B nexus controls autophagy. ELife, 1-25.
7.Wilhelmi, I., Kanski, R., Neumann, A., Herdt, O., Hoff, F., Jacob, R., Preußner, M., & Heyd, F. (2016). Sec16 alternative splicing dynamically controls COPII transport efficiency. Nature Communications, 7, 12347. https://doi.org/10.1038/ncomms12347
-
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #2
Evidence, reproducibility and clarity
The manuscript by Jung et al which based on a targeted siRNA screen, demonstrates regulation of SEC23A (component of the SEC23 complex of the COP coat) levels at transcriptional level downstream of focal adhesion signaling. By regulating siRNA mediated downregulation, the authors were able to identify proteins which either increased or decreased traffic of VSVG through the secretory pathway when combined with downregulation in the levels of with either SEC23A or SEC23B. Authors have focused on a group of SEC23B functional interactors, downregulation of which shows them increased size of focal adhesions which also downregulate SEC23A …
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #2
Evidence, reproducibility and clarity
The manuscript by Jung et al which based on a targeted siRNA screen, demonstrates regulation of SEC23A (component of the SEC23 complex of the COP coat) levels at transcriptional level downstream of focal adhesion signaling. By regulating siRNA mediated downregulation, the authors were able to identify proteins which either increased or decreased traffic of VSVG through the secretory pathway when combined with downregulation in the levels of with either SEC23A or SEC23B. Authors have focused on a group of SEC23B functional interactors, downregulation of which shows them increased size of focal adhesions which also downregulate SEC23A levels, thus providing an explanation for reduced secretory traffic. Authors further show that plating cells on fibronectin or Matrigel, which activate Focal adhesion kinase signaling also results in downregulation of SEC23A transcript levels. The screen is conducted in a well-controlled manner for most parts with a clear explanation of the analysis routines and the data presentation if of very good quality. Most important results have been validated by more than one experimental strategy which lends substantial confidence to the findings. The results also open further avenues for understanding the transcriptional regulation in different physiological and disease contexts.
There are certain issues, which the authors should address with regards to controls and some conflicting observations with published results with respect the phenotypes associated with downregulating proteins on focal adhesions size. Additionally, authors don't tie the ends by monitoring secretory traffic in cells grown on different matrices but include it in the model. Addressing/explaining these issues could improve this manuscript and the model may have to be tweaked a bit.
Major comments:
1)I wonder why the authors only used siRNA control in their screen when the effects are scored in context of double knockdown fashion in combination with mild knockdown of SEC23A and SEC23B to get functional interactors. Control siRNA in combination with SEC23A and SEC23B should have been two ideal negative controls in the screen. Nevertheless, in data presented Figure 1E and whole of Figure 2, using control siRNA in combination with SEC23B siRNA would have been ideal control to show that the combination does not induce any trafficking defects which could impact the findings of the study. Hence, a few of the data presented from some of these figures should have sicontrol+SEC23B siRNA combination as a control.
2)What is the identity of post-ER structures which authors refer to in Figure 2A? Could the images represent VSVG concentrated at ER exit sites? Authors should stain with markers for ERES to see if the VSVG puncta colocalize with it.
3)Based on RNA sequencing results, authors chose to follow up on SEC23A levels in background of siRNA knockdown of components (like MACF1, ROCK1, FERMT2 etc.) which regulate Focal adhesions in cells and show that there is a reduction in both transcript and protein levels of SEC23A. In images shown in Figure 2B and Figure 2C, levels or SEC31A and β-Cop1 are reduced. Authors should test using qPCR and western blots whether there is a downregulation SEC31A, β-Cop1 and SEC23B in siRNA knockdowns of MACF1, ROCK1, FERMT2 etc. It would provide new insights if there were a co-regulation of secretory machinery to modulate the secretory traffic in response to Focal Adhesion based signaling.
4)Most major concern in this manuscript surrounds around results presented in Figure 4C. Authors show that in response to all the knockdowns, they see more focal adhesions as monitored by Vinculin staining and this along with the experiments with cells plated on Matrigel and Fibronectin arrive at the conclusion that increased Focal adhesion signaling downregulates SEC23A levels which presumably modulates secretory traffic. I am not an expert on Focal adhesions but based on my understanding of the literature on that topic, downregulation of ROCK1, FEMRT2 disrupts focal adhesions. (See: Theodosiou et. al., Elife, 2016 or Lock et. al., Plos One, 2012 for example). How do authors explain their results in siRNA knockdown of ROCK1 and FEMRT2 which leads to an increased size of focal adhesions which seems contradictory to the published results? To clarify these results authors should test phosphorylation of FAK in their siRNA backgrounds which is another read out of focal adhesion signaling. The experiments from cells grown on Fibronectin and Matrigel favor the argument which authors put forth, but authors may have to tweak the model a bit based on FAK phosphorylation and FAK signaling in context of above-mentioned knockdowns.
5)What happens to VSVG traffic or RUSH-Cadherin traffic when cells are plated on Matrigel and Fibronectin? Reduction in secretory traffic of these is an important experiment which is missing to close the loop and validate the model presented. Authors must test these experiments either with cells grown on matrix alone or in combination with siRNA to SEC23B. Authors should also monitor ERES and transport carriers in this background.
6)This is not such a major issue, but it would be good to see a comparison in SEC23A levels in siRNA knockdown condition in comparison to those when cells are grown on different substrates and in ROCK1, FEMRT2 knockdowns (blots of which authors already have in this manuscript).
Minor comments:
1)Scale bars are missing in EM images in Figure 2H.
2)Show molecular weight markers in Western blots in main figure 3E and supplementary figure S1E.
Significance
I have looked at the manuscript from through the lens of a cell biologist as that is predominantly my area of expertise. In that respect I find the screen conducted by authors particularly interesting as they aim to connect how extracellular cues regulate the secretory pathway. A screen seems justified as there is no comprehensive understanding linking the two above-mentioned processes. Authors have done a functional interaction screen and analyzed a lot of images to identify candidates which either increase or decrease secretory traffic in combination with SEC23A and SEC23B. Such a functional screen has helped authors identify candidates which were otherwise missed in single siRNA knockdowns in their previous work from 2012. This definitely opens up interesting avenues to test the candidates identified in the screen in different physiological contexts and in disease as also the transcriptional program connecting Focal adhesion signaling with the regulation of components governing secretion. Such functional interaction screens could also be employed to identify crosstalk of different cellular processes with the regulation secretory pathway at ER as well as at the Golgi apparatus.
Referees cross-commenting
I agree with the comments from both the referees that the manuscript is very interesting, most experiments are well controlled, but the quantification of focal adhesion phenotype in knockdowns need to be done in an extensive manner and secretion phenotypes need to measured upon plating cells on different matrix to validate the model presented.
-
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #1
Evidence, reproducibility and clarity
The manuscript by Jung et al reports on an interesting finding that focal adhesion signaling regulates the expression of Sec23A and thereby regulates COPII-dependent trafficking. The data presented a mostly solid and the finding itself is highly novel, as it tackles an area of secretory trafficking that remains poorly understood, namely the connection between the ECM and secretion.
I will list below all comments that I have mixing both technical and conceptual topics:
Technical issues:
1-The authors should provide a better description of how the designed this siRNA library. What were the inclusion criteria for these 378 genes? I …
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #1
Evidence, reproducibility and clarity
The manuscript by Jung et al reports on an interesting finding that focal adhesion signaling regulates the expression of Sec23A and thereby regulates COPII-dependent trafficking. The data presented a mostly solid and the finding itself is highly novel, as it tackles an area of secretory trafficking that remains poorly understood, namely the connection between the ECM and secretion.
I will list below all comments that I have mixing both technical and conceptual topics:
Technical issues:
1-The authors should provide a better description of how the designed this siRNA library. What were the inclusion criteria for these 378 genes? I might have missed it, but I could not find this information easily.
2-Figure 2: I know this is challenging for EM images, but is there a way the authors could quantify these data? How many images were looked at? What was the average width of ER cisterne?
3-Figure 4: I think that the characterization of the FA phenotype is a bit underdeveloped. There is no quantification of these data. Is the size of FA changing? Is the number of FA per cell changing? Is the length of FAs changing? I think that more work is needed to increase the confidence in these data. I could also not easily see what type of cells these are. A better description of this experiment is also required. Also, how many cells were analyzed. I think it is important that this experiment is done with a sufficient number of cells to increase the confidence in the data.
Conceptual issues:
1-The finding that focal adhesion signaling negatively affects ER-export is surprising, because cancer cells that grow on stiff substrates have more focal adhesions and are more invasive and migratory. Both migration and invasion are expected to depend on ER-export. Although the authors did not formally test Sec23A expression under different stiffnesses, I would expect that stiff substrates would lower Sec23A expression and thereby negatively affect ER-export. It would certainly increase the breadth of this work to include data like this and to also discuss this highly surprising finding. However, it is of course the decision of the authors and the editors to decide whether such an experiment would benefit the entire story.
2-The authors postulate that this novel mechanism could be part of a feedback loop. If this were the case one would expect the acute effect of FA to increase ER-export (or secretion) and the negative feedback will then reduce secretion. However, the acute effect of FA is not addressed in this manuscript. In order to postulate a feedback loop, the authors would need to test the individual nodes of this loop.
Significance
I think that the basic finding of this manuscript is highly novel, by showing the impact of the ECM and focal adhesions on COPII-dependent trafficking. I think that this will not only appeal to people from the trafficking community, but also to people working on cell migration and on mechanobiology. The work in its current form does not require much extra efforts (max. 3 month). However, if the authors would decide to increase the breadth of data, they would require 3-6 months.
Referees cross-commenting
I went through the comments of the two other reviewers and agree with their verdict. Some extra work on the characterization of the early secretory pathway would be good. Both reviewers provided a nice catalogue of possible experiments to choose from.
-
-
-