This is really interesting! I think it could also be interesting to see if any of the features (these or others) correlate or if any features could be predictive of others?

This is super cool! I can't wait to try it out!

I'm curious if you can use this with structures predicted by AlphaFold or ESMFold. Related to that, I'm curious if you need to do any sort of pre-processing of the structures (mostly for AlphaFold and ESMFold structures because they're known to not always have optimal side chain placement).

I think this figure might also be mixed up.

I think I only see one panel in this figure

Thanks for providing the code! It helped me better understand some of the examples in the paper.

Most of your examples seem to be embedding or vector-based, which is very cool. But I think it could be useful to see some examples that use sequence or even structures since that is also presumably doable with your approach.

I really love this example!

I'm having a bit of trouble separating what your approach enables vs what just Protein-Vec alone does in this example. I know that your approach tells us about confidence in the annotations, but it might be interesting to discuss what comes out of Protein-Vec alone vs what comes out with your approach?

Might have missed this, but which proteins are which color?

I think that the idea of using conformal prediction to generate some sort of confidence about which proteins to experiment with could be extremely useful! I really enjoyed reading this paper, and one of my favorite things about this paper is that the authors include so many different examples of how this could be applied. I think it could be very cool to take some of these predictions into the lab in the future!

Because you use Protein-Vec quite a bit throughout this paper, it might be useful to give a bit more context up front.

I am curious if there were differences in expression or yield between the different proteins. I could see that being important if the thought is to produce a bunch of this protein and use it to degrade PET.

This is interesting! Do the protein design methods you used specifically try to preserve the catalytically active parts of the protein or was this something that you assessed when picking proteins? If not, might that be useful to consider when selecting proteins to test? For example, maybe the proteins with negligible activity had a lot of differences in the triad. That seems like something you could catch before purifying and testing activity.

I'd love to see some discussion around your working hypothesis as it seems that you maybe disproved it? It might even be cool to have a bivariate plot with thermal stability on one axis and enzymatic activity on the other to see if there is a correlation. Additionally, because you tried 3 different methods to generate proteins and then evaluated them the same way, it might be useful to talk about which worked best, why that might be, pros/cons of the three methods, etc.

Again, it would be interesting to see this data, even if you just put it in the supplement!

This sounds like such a cool result, it would be great to be able to see the data.

It might still be useful to determine the Tm for these proteins to help with testing the working hypothesis that you discussed in the intro that proteins with higher stability have higher enzymatic activity.

What a cool paper combining computational modeling of protein structure with experimental analysis to support it! I'm excited for the next steps listed here (crystallography or NMR) to see how those results line up with what was found here, but either way, I'm a big fan of the combined computational and experimental work!

Did you do technical replicates? I'm curious about how consistent the data is between trials.

I'm curious if you happened to do replicates of this and/or stats to determine if your quantification is significant?

Love that you included the accession number and a direct link to the sequence! It can be a pain when papers don't reference the exact protein that they're working with. It might be worth explicitly stating how you identified the Myb domain.

I really appreciate the direct comparison of the 4 methods on a single sample. This is totally beyond the scope of the paper, but I wonder if these observations would hold true with other proteins/protein domains.

Is this not just because you're BLASTing a single domain against a database of mostly full proteins?

It's a bit confusing to refer to the Myb domain as a "protein" when you aren't talking about the full-length protein.

Just curious if there's a structure available of the full protein? It might be useful context to see how your analysis of the Myb domain fits with the full structure (if it's available).

Do you know the domains that these different interacting proteins bind? The first sentence suggests that you do, and it might be useful for better understanding the particular domain that you focus on here (Myb) to know what interacts with it.

This is very minor, but when I first read this (before looking at Figure 1), I thought this phrase meant that this specific region was 323-445 amino acids long. It might be less confusing to use language like "a specific region located between amino acid 323 and 445"? Otherwise, this section introducing the functional domains is very thorough and got me right up to speed on TTF1!

This is an interesting deep dive into the ADF family, particularly in Arabidopsis! There's a ton of data and I appreciate the open code!

Are these datasets provided somewhere? I might have missed them, but couldn't find them!

Are motifs 3 and 6 related since they're both green?

Can you draw any sort of hypotheses from this analysis? It's a lot of data and I'm just wondering exactly what it all means in the context of the rest!

I think this is maybe supposed to be "candidates"?

Is it known if any of these ADF proteins have overlapping functions? Can they compensate for another if needed?

I don't see a Fig 2b.

The clusters aren't super obvious to me, could you maybe highlight the 3 clusters in the figure?

I'm curious how similar these genes are within organisms and between organisms. Do you have sequence alignments or sequence identities you could share to give an idea?

Apart from choosing one to represent each phylum of eukaryotes, how did you pick these specific species? You might have this somewhere up ahead, but it could also be nice to have a table of these species and the other set of 8 species.

I really like the inclusion of the stats to highlight the need for characterization of these proteins! This is no big deal but the last sentence of this chunk (The remaining 30% remain without any...) has slightly odd wording. Maybe something like "the remaining 30% lack functional annotation or characterization"?

This is a really cool study of the role of mechanosensing and actin in regeneration using a very cool model! I also really enjoyed all the microscopy images. Looking forward to learning more about Hydra and actin in future papers!

I like the use of these for comparison! It's cool that you have built-in examples of when things didn't work to use for a comparison like this.

Woah cool images!

It might be helpful to annotate an image with what these different things look like or add arrows to the existing figure for folks who aren't as familiar with looking at these images.

This is somewhat related to a previous comment, but is the orientation of head-regenerating tissues affected by agarose compression? It's not super clear from the data in extended figure 1a, but it seems like it could be somewhat related since at 0.5% AC there's only lateral orientation.

It seems like there's quite a bit of phenotypic variation even between these 2 examples of biaxial, could this variation be meaningful?

It might be worth showing what this looks like (especially compared to the biaxial ones) in the main figure since you do have a portion of organisms with this ectopic tentacle phenotype.

Is this expected? Is it because they're less squished so they have room to orient? Just wondering about the biological significance

I really enjoyed reading this paper about septin evolution! It opens the door to help us understand more about septins by looking outside of Opisthokonta. The paper is also thorough, and in addition to learning about their newly identified septins, I was also able to learn a lot about septin biology in general. I'm excited to see how this data is used in the future!

This is a really cool section! I love the use of structural analysis + phylogeny to uncover something really cool about the function of these proteins.

Do you have any ideas about the functional consequences of this variability? Do you think it could affect the G-interface dimerization?

I'm curious how/why you selected these specific species and how you decided which septin to use within each species?

It sounds like your searches were mostly sequence-based. I know septins can be large and can contain regions of disorder, which may make them difficult to work with structurally. However, I'm curious if you think you might identify more septin sequences using a structure-based search (like Foldseek)?

Was your search limited to these taxa or did you search more broadly as well?

One of the main reasons I might use a structural search is if I have a novel protein that maybe isn't in UniProt. I don't know if it's possible with the way your tool works, but it could be a cool thing to think about for the future - is there a way to support user provided or determined PDBs that aren't in UniProt?

The examples are really great! It is a bit hard to really see what's happening in the overlays of all the structures. It might be helpful to see overlays for each hit protein with the the input as separate panels or something.

I like that you brought up some of these other tools and described how your tool is different. Are there other benefits that the user might care about? For example, I noticed that the web tool is really fast! This is probably beyond the scope here, but a comparison of these different structure search tools would be useful.

I appreciate the limitations section! I'm curious if there are plans to eventually incorporate the newer version of the AlphaFold database? Also wondering about things like the PDB database itself and the ESM metagenomic atlas?

I really appreciate this super easy to use and fast web application tool for finding proteins similar to an input!

Thanks so much for sharing this tool! I really enjoyed reading about it and can't wait to try it myself.

I really love this walkthrough!!

You mentioned that several of the top predicted PPIs have been previously experimentally validated. It would be interesting to know which of these annotations predicted by your pipeline are supported by experimental validation. Could you somehow denote this in this figure?

Could you include a key with the color scale for the heatmap?

If you run the analysis multiple times, do you see variation between runs or is it only when you switch the chains?

I'm curious if you're pipeline supports more than 2 proteins. For example, could I put in a handful of proteins to see if they form a complex?

I think based on your analysis that the answer to this question is yes, but can you do multiple bait proteins as well as multiple candidate proteins? Also curious about how many proteins this can handle?

This is a really cool analysis combining sequence-based and structure-based functional domain prediction! It provides tons of new info about trypanosomatid biology (even giving us some possible new drug targets!), but could also tell us a lot about sequence/structure conservation and how we might leverage this for annotation and drug target ID purposes.

It would be interesting to think about how your sequence-based predictions and structure-based predictions compare. For example, do you have proteins that were similar based on sequence but not structure and vice versa? And even between different sequence-based and structure-based methods since you applied a bunch of tools? This could be an interesting opportunity to learn more about sequence and structure conservation in an organism that's more distant from humans!

It might be useful to put this information sooner. I was confused why you were talking about IFT70 previously and this context was really helpful!

Because it seems like you're sort of starting a new section here, it might be useful to add a title so that it doesn't run together with the previous section.

I love this bit of information about what you found in your analyses. It's very helpful for thinking about these proteins.

Would also mention what this abbreviation stands for in case you have readers that aren't super familiar with this protein domain.

Again would be interesting to include what you found in your analysis about AAA 18 domains.

Might be good to include what this abbreviation means when you first reference it. Would also be useful to include some information about what you found in your analysis that led you to include this section.

I'm sure you considered it, but you could also try employing Foldseek to search for matches in a couple subsets of the AlphaFold database as well as CATH50, MGnify, and others. It's also super fast! https://search.foldseek.com/search

It would be cool to see a figure showing how similar the sequences and structures are of these proteins compared to known versions. Since it seems like we know quite a bit about them (ex that muts like Arg23Gln can affect binding in UFC1), it would be interesting to see if particular important residues are conserved.

Why these criteria? Also by requiring that you get matches in both the sequence based and structure based tools, are you missing proteins that maybe have very similar structures (and possibly function) but very different sequences?

Should this be "out"?

Very cool! I love studies taking advantage of this differential regulation to assay how different parts of the actin cytoskeleton contribute to cellular functions!

I wonder if the intensities of Cdc42 are different in these various treatment conditions? for example, it looks like there's lots more signal in the for3 mutants, but this could just be the images shown here. Additionally, the changes in the level of nuclear localization here are interesting!

Maybe it's just the individual cells in the image, but the oscillations, while still present, do look somewhat different in the for3 mutants. For example, according to the red arrows, the frequency of oscillations between the DMSO treated and for3 mutants do not look the same in Figure 1A. I agree with Cameron that showing something like supplemental figure 1A would be super helpful, because with that you can really see the difference between DMSO treated/for3 mutants and CK-666 or LatA treated cells.

Really fun paper that digs into protein function with a lot of cool assays! Overall, I really enjoyed reading it!

Could it also be related to the lack on NTD in the wtMR construct?

I understand the effort to eliminate confounding variables. However, based on the western blots in Fig S3, it seems like in most cases in your strains, you're ending up with CsoS2B (which I think is the long isoform?), but here you're forcing all mutants to the short version? Or maybe I'm misunderstanding! Either way, this part is a bit confusing and I'd love a bit of clarification here.

Do you think this would be the case if you had mutated the cysteine to an alanine?

This is super helpful and it might be worth moving some version to the main text or adding a table of your mutants or something since you have so many and it can be a little tricky to keep track of.

Would be great if you could quantify these results? I also think it could be helpful to check all mutants with both antibodies to compare 2B to 2A for all instead of just the VTG mutants.

I totally see the reasoning behind mutating to C to S instead of A. However, I think it would also be interesting to see if mutating this to an A would cause further functional changes, especially because you changed the rest to A.

This is super interesting! Would love some more information - could you provide citations and/or data to demonstrate this?

I really loved reading this paper! Although actin and the Arp2/3 complex are well-studied proteins in mammalian cells and common model organisms, this paper really highlights how much we still have to learn about both actin biology and especially the Arp2/3 complex!

This is also something that we encountered in Chlamydomonas! I wonder if a structural search for some of those NPFs might give you better results than a sequence search.

It might be helpful to discuss the roles of the different subunits in actual complex function. For example, ARPC2 and ARPC4 generally form the primary connection with the mother filament and Arp2 and Arp3 help nucleate the new daughter filament. You identified those essential subunits! It might also be helpful to talk a bit about other non-canonical Arp2/3 complexes throughout the tree of life (for example Chlamydomonas), and to talk about previous studies where people have mutated or removed a subunit and the complex still retained some function.

I'm curious if there are any other known actin-dependent phenotypes that you could probe with your ARPC1 null to support this more in future work.

Some docking studies using CK-666 could help support this since you already have the structures available!

I'm wondering how healthy cells were with this dosage of CK-666? And if you tried any concentrations between 50uM and 250uM?

This section is so cool! I like the model of your putative Arp2 and Arp3 with a couple actin monomers. Did you try modeling a full complex with your identified proteins?

Just pointing out that this sentence appears to be incomplete!

I'm curious if there are differences in axoneme formation or structure. I'm not sure about plasmodium cells, but the Arp2/3 complex has been found to be involved in flagellar assembly in other organisms.

Very interesting!

I know that you'll get to structure in your work, but was there anything known about structure conservation previously? I know 20% is super low for sequence identity, but I imagine that the structure could potentially be better conserved.

Since you mention phylogenetic studies, I'm curious if this apparent loss of most of the Arp2/3 complex is common in Plasmodium's closest relatives or if this is unique to Plasmodium species.

I really loved this section!

A very cool paper, with some very cool images showing tunneling nanotubes in zebrafish embryos!

Is there a third intermediate concentration you could try to see if you get an intermediate effect?

Panels C and E and the significance labels are a bit confusing. I'm wondering if there's a different a way to display this data?

I really love this approach, and I think this data is very cool! I don't know much about these specific labels but I'm curious if you could see transfer of colors between cells at all?

Could you quantify the lifetime of different types of connections? Just wondering about different ways to classify TNTs vs other types of connections.

I love this figure and video of the two filopodia reaching out and connecting!

I'm interested in the thickness of the connections - the bridges seem quite thick.

This is a clever way to differentiate between cytokinetic bridges and the TNT-like connections, but I don't know that it fully rules out that the TNT-like connections could be related to division. Is there a way to block division and see if you still get TNT-like connections?

Did you only do this staining and analysis with TNT-like protrusions? It's clear that there's some diversity of the cytoskeletal composition of these structures, but I wonder if compared to other types of protrusions, some pattern might emerge. For example, comparatively, filopodia are likely only actin the majority of the time, making them quite different than these structures.

Might be nice to add a citation here since you have specific number values.

Have these been seen in living embryos before or is your study the first report of this?

I'm sort of interested in the pairs that are close to the interaction cutoff but on the expected side. For example, would a pair with an interaction score of like 0.56 or so have the structure and interaction that you would expect? I guess, how accurate is the 0.5 cutoff?

This is really exciting! I can think of tons of really interesting hypotheses that could be tested by doing this kind of analysis at a larger scale, and I think your analysis here does a great job opening that up.

Do you see other cases of disorder influencing the predictions of binding? Maybe not to this degree, but are there other cases where disordered regions cause a lower or higher ipTM than expected?

Might consider adding a citation here

Is this similar to what you see with non-interacting pairs? Are the protein structures themselves quite different or just the contact between the two models?

I really appreciate the use of this previous dataset for this purpose! I went back to the previous paper, and this dataset seems like the perfect fit for this kind of analysis. I think this is a great test of the limits of this particular feature of AlphaFold2 and provides some great insight into what it can be useful for in the future.

This is really interesting, and I'm curious how often these aggregates show up for the different mutants.

This paper is so cool and such a fun read! I found the data that showed PFN1 in the mitochondria (figure 6) especially interesting and the mutant data throughout particularly convincing. I'm looking forward to following this story and learning more about PFN1 and how it affects mitochondria!

I think that the mutant data throughout is quite convincing since they have varying actin polymerization activity. It could be really useful for digging into this more to know a little more about these mutants. Do they block binding to other proteins, affect overall profilin folding or expression, or affect other functions that we know of?

This experiment and this data is so cool!!! I am curious if you think actin is also present inside the mitochondria?

Are the stars here signifying significance between each mutant and the GFP control or the PFN1 rescue?

In Figures 1 and 2, I don't think you mention what the stars represent. I'm guessing that they are the usual significance cutoffs, but you might add them in the legends!

M114T looks a bit more promising than the others. I'd be interesting to know more about how this mutant is different!

Does decreasing the amount of polymerized actin further eventually result in mitochondrial defects and activation of mitophagy? In the discussion you mentioned that this process doesn't actually require that much actin so I wonder if the 50-60% that's left is actually sufficient. Additionally, because profilin has such a complex role in actin function (polymerization promotion, monomer sequestering, ATP hydrolysis promoting), I imagine that it is probably quite difficult to mimic what loss of profilin would look like in this way. Is there some other cellular function that is affected by loss of profilin function that you could use as a readout to show that this is affecting cells in a similar way to loss of profilin?

I'm interested in knowing a bit more about this expression dataset in the context of this work. Do actin and related proteins express at normal levels in PFN1 KO cells?

unpaired?

Could you measure the intensity of the Arp3 fluorescence in the cytosol and outside of the aggregates in the HTTQ15 compared to the HTTQ138 cells to see the extent of sequestration? Same with the actin! It might help support some of your conclusions if you can actually point to some numbers and data.

Really striking differences, super cool! I'm curious how many cells were measured to make these graphs and if there were any stats done?

A really interesting paper using some very cool techniques to look at actin (and the Arp2/3 complex!) and clathrin-mediated endocytosis in disease states!

Are there ways to make the cells stiffer without the HTTQ138 defect that could support this statement? Has stiffness been shown to impair CCS movement in previous papers that you could cite here?

You might consider some reorganization of figures 2-4 as it would be helpful to see the controls, the knockdown, and the rescue all together in one figure. So for example, putting the Hip1 knockdown and coexpression stuff in the same figure would be helpful so readers don't have to scroll back and forth between figures.

Because you have a lot of proteins that you've found do influence this process, it might be helpful to have a diagram/model maybe at the end showing how each protein is important.

This is a very minor wording thing, but the movement in 2E actually does look quite directional (it's not distributed around the circle), just not the direction that you would expect.

This is a bit confusing - are you using WT and HTTQ15 interchangeably or are you showing wild-type in Figure 3A-C instead of of the HTTQ15 expressing cells? I think the HTTQ15 is probably the better control so I'd like to see it in the figure.

Really interesting study with some beautiful microscopy and analysis! Studies looking at interactions between the different cytoskeletal elements are always great, and I'm left wondering how the third biopolymer, actin, factors in.

Super minor comment but there are a few places throughout the paper where words are randomly hyphenated in the middle of the word.

Very cool!

It would be nice to see images with just microtubules and just vimentin for each condition.

This might be beyond the scope of this paper or already done somewhere, but the networks look quite different between the two cell shapes even in just wild-type cells. I'm curious if there are any more in depth characterizations between these?

Do the cytoskeletal and other components of the cell look mostly normal in these compared to cells not grown on these micropatterns?

Do you know if they also typically contain actin based on immunofluorescence or staining?

Cofilactin in microtubule lumens?! What an amazing finding and paper! So much cool data here that I think will also lead to tons of really interesting hypotheses

It might be nice to mention and discuss a little bit the link between Cytochalasin D and the actin cytoskeleton. Do you think that this could at all be affecting your results even with the low concentration? I think this could probably just be addressed in the text, but are there other methods to induce these protrusions that you could use as a control to show that this isn't related to Cytochalasin D or could you stain some cells with phalloidin and treat with this low concentration of CytD to show that it isn't affecting filaments?

This is a really interesting question! This is probably outside the scope of this paper, but related to this question, I'm curious about the timescale of formation of the protrusions? How dynamic are they?

Does CtyD also decreases this phosphorylation a little? I'm wondering if the few luminal filaments in non-TG treated cells are also likely cofilactin as opposed to bare actin.

Does the number of protofilaments change throughout the length of the microtubule? Like one end having more or less protofilaments or are they uniform throughout?

To really illustrate this it would be helpful to show a bivariate analysis between like number of lamellipodia per cell and number of TNTs per cell for 2B. Then you could actually graphically show this anti-correlation and provide a correlation coefficient to strengthen this conclusion. It would also be interesting to show this with CK-666 treatment as I wondered while reading this if the increase in TNTs/filopodia with CK-666 treatment was accompanied by a decrease in lamellipodia.

What a cool paper! Beautiful images and I love papers that deal with this concept of a limited pool of actin in the cell that is directed to its different functions by its interactions with actin binding/interacting proteins.

I really like this data with the formin agonist!

I noticed in figure 4A the tubulin loading control isn't super consistent. Did you normalize to your loading control? Might also be worth doing total protein as your loading control instead of tubulin, because it is possilbe that tubulin is being affected since you're probing other cytoskeleton related proteins.

This experiment is very cool, but I wonder if you could show a similar thing with naturally forming TNTs? Do you see more actin in natural TNTs when the Arp2/3 complex is inhibited? I worry that stretching the cell like this could be causing other things to happen in the cell and isn't fully representative of a TNT forming on its own.

It would be really great to see representative images of the data quantified in Figure 2C.

Interesting that ARPC5 disappeared, but not ARPC5L if I'm reading that correctly?

Maybe i'm missing this, but I don't see Eps8 in this map. Might just be worth discussing in the text why it's not there?

There are a couple of these graphs where we're looking at the distribution of data and there's mention that there's more of a skew in a sample compared to another, but sometimes this is really difficult to see in the graphs. It might be useful to split out arp2 mutants vs arpc2 mutants so that we're comparing 2 instead of 4. It might also be useful if you could include some measure of skew, especially for the ones where you think there might be a difference in skew.

What a cool paper! I think papers dealing with different proteins interacting with a single pool of actin are so interesting. Also super awesome that you validate CK-666 in these cells and start dealing with potential off-target effects of SMIFH2 in these cells - so much added value to the field!

I think you do a great job here and in the next figure validating CK-666 (definitely showing that there isn't added effect in the Arp2/3 complex mutants is great!), but you might consider adding CK-689 as an inactive control just to cover all your bases.

It would be interesting to see specific data points here instead of bars for these graphs. Based on the representative images, it looks like there could be a couple population with different lengths, especially in the arpc2 hypocotyls. If that's the case, it would also be interesting to hear why that might be.

Do we know that SMIFH2 broadly inhibit all of the formins in Arabidopsis?

Interesting, is the hypothesis that some of these formins that nucleate filaments off the side of existing filaments are responsible for these specific branched filaments?

I really enjoyed reading this paper - what a great story! It really highlights how little we still know about actin regulation (even in human cells!), but it also does a great job filling in some of that knowledge with this pathway involving this atypical MAPK.

Super interesting! Do these mutants have motility defects too?

Very cool! Did you check other subunits or did they not precipitate?

Woah! Curious if the kinase activity (that you test in the next section) is required for this change in actin polymerization?

It would be interesting to see this in ERK3 knockdowns as well to see if the localization of Cdc42 changes with loss of ERK3. I'm also curious about the localization of Rac1.

By eye, I agree that it looks like cell shape is altered in Figure 1A, but it would be cool to include a quantification (maybe like how round the cells are or something). I also think that Figures 1H-J are relevant here to help connect some dots for this conclusion and show that the loss of ERK3 does cause changes in specific actin dynamics that result in shape and motility defects, but you don't talk about Fig 1H-J until later. Finally, I think the conclusion about polarized migration is more relevant in the next section than at the end of this section.

Are there stats for Figure 2F? It seems like there's more variance with the Rac1 blot quantifications than with the Cdc42 blot quantifications.

It might be useful to quantify protrusions for all of 6F because it seems like there are definitely some big differences!

I find this interesting also in terms of the effects of ERK3 knockdown on the morphology of the cells and the specific actin structures. You mentioned previously that Rac1 leads to the formation of lamellipodia and Cdc42 leads to the formation of filopodia, and looking at the images of your cells, it looks like filopodia are more affected than lamellipodia. Is that something you noticed looking through your images?

You might specify that this is true for human isoforms. There are lots of weird actins out there in other organisms that are well under that 95%

I think it would be useful to talk again about the effect of the different constructs that you used in the context of what that means biologically in the cell.

I think this is obviously quite visible in the images, but it would be interesting to see a quantification of filament length in each case and also the occurrence of these bright bundle-like regions. I know you do this more precisely in the next section but I think this would also help set that section up a little bit.

It would be super cool to see some images along with these graphs looking at single filament elongation.

I am interested what this would look like in cells! Maybe a next step?

These are amazing!

Actin + Arp2/3 + VCA + delta-mod (cyan) is interesting - it looks like nucleation is delayed, but then is able to rapidly at least approach the same level as Arp2/3 + VCA. I'd love to hear some more thoughts about this.

This is a very interesting paper looking out how different actin binding proteins work together to regulate actin in specific functions!

I think it would go a long way to support your concluding hypotheses if you did a quantification of branching for Figure 6C because it is a little hard to make out with the density of the filaments and because it is so important for your conclusions.

It would be interesting to look at direct binding between these proteins, especially because they are mutants that might not behave quite as expected.

How often did you see these different morphologies? I'm curious if some are more common than others.

Interesting.. I'd expect these structures to be really dynamic. Is the turnover of actin at the membrane decreased in these deflated GUVs?

Did you do any experiments where you polymerized actin filaments with the Arp2/3 complex, but with non-membrane anchored VCAs?

This is a really interesting paper with a really cool technique, some beautiful images, and a ton of data that tells us a lot about the mechanics of cell protrusions. Additionally, their GUVs with minimal actin cortices could be a super useful tool for better understanding what different proteins and combinations of proteins are doing at the cortex.

Really clever way to drive formation of an actin cortex at the membrane of the GUVs. But obviously VCAs aren't attached to the membrane like this in biological systems, so are there biological consequences that might affect some of this data? I think it would be interesting to hear a little bit about this in the discussion maybe?

I was still kind of questioning if the GUV misshaping was really due to the actin vs just having extra membrane around (even though you had the bare GUVs being more spherical), but this experiment and the CytoD experiment are really convincing. Although I wonder if you could include a quantification of roundness or a similar parameter pre and post ablation?

In figure 1D, it looks like the width of the bleached region didn't change, but it looks like the overall intensity of the signal around the whole GUV deceases - do you think this is from cycling through depolymerization, then monomeric actin, then repolymerizing at the bleached spot?

I think this FRAP experiment is really cool, and definitely important to show that the actin at the cortex turns over. Also, really cool that the turnover is close to that of living cells even without debranching, severing, disassembling proteins! Can you distinguish between filamentous actin and monomeric actin with your labelling technique? (It might be worth mentioning how you're labelling the actin here.) I'm curious what the ratio of G-actin to F-actin looks like and how this might affect turnover. You have a really great paragraph about this in discussion, but might be useful to mention briefly here.

The integrin localization looks weird compared to the control too (at least in this image), do you think hat loss of Arf6 is affecting trafficking or polarity or something?

Could we get a little more information about this experiment? Is it in vitro? Is it just actin from lysate? What is the buffer composition? Do you have some sort of control to show that you started with the same amount of actin in each case?

Cool! But this isn't true for all proteins correct? For example, it looks like VEGFR2 doesn't require the Arp2/3 complex, which is interesting, especially because Arf6 does seem to be required. Any hypotheses about this?

Couldn't there still be active actin polymerization by formins and the like in other parts of the cell? I think you could add that the active actin polymerization is Arp2/3 complex-mediated.

Is this the actin-related protein Arp2 that is a part of the Arp2/3 complex? I think this is an important distinction that you might take advantage of and discuss a little bit more here.

Does loss of Arf6 cause differences in Arp2/3 complex recruitment to clathrin sites?

Could be related to off-target effects of PitStop2

This is really interesting! I'd love to know more about the mechanism behind this phenotype.

I think DMSO is generally a fine control, but you might consider CK-689 (the inactive version of CK-666) just because some of these are so close.

Beautiful images throughout the paper!

Is there some sort of control you could include here to show that you loaded the same amount of total protein? Maybe a coomassie dye or something?