Distinct role of TGN-resident clathrin adaptors for Rab5 activation in the TGN-endosome trafficking pathway

  1. Makoto Nagano
  2. Kaito Aoshima
  3. Hiroki Shimamura
  4. Daria Elisabeth Siekhaus
  5. Junko Y. Toshima
  6. Jiro Toshima

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

Clathrin-mediated vesicle trafficking plays central roles in the post-Golgi transport pathways from the trans -Golgi network (TGN) to endosomes. In yeast, two clathrin adaptors – AP-1 complex and GGA proteins (GGAs) – are predicted to generate distinct transport vesicles at the TGN, and epsin-related Ent3p/Ent5p act as accessories for these adaptors. Recently, we showed that vesicle transport from the TGN, rather than from the plasma membrane, is crucial for Rab5-mediated endosome formation, and that Ent3p/5p are crucial for this process, whereas AP-1 and GGAs are dispensable. However, these observations were incompatible with previous studies showing that these adaptors are required for Ent3p/5p recruitment to the TGN, and thus the overall mechanism responsible for regulation of Rab5 activity remains ambiguous. Here we investigated the functional relationships between clathrin adaptors in post-Golgi-mediated Rab5 activation. We were able to show that AP-1 disruption in ent3 Δ/ 5 Δ mutant impairs Rab5-GEF Vps9p transport to the Rab5 compartment, and severely reduces Rab5 activity. Additionally, GGAs, Golgi-resident PI4 kinase Pik1p and Rab11 GTPases Ypt31p/32p were found to have partially overlapping functions for recruitment of AP-1 and Ent3p/5p to the TGN. These findings suggest a distinct role of clathrin adaptors for Rab5 activation in the TGN-endosome trafficking pathway.

Article activity feed

  1. Review Commons

    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

    Manuscript number: RC-2023-01939

    Corresponding authors: Jiro Toshima, Junko Y. Toshima

    1. __ General Statements __ We are grateful for the reviewer’s evaluation of our study. In the new manuscript, we have answered all of the points raised by the two reviewers (the altered or added text is indicated in red in the new manuscript). Reviewer #1 pointed out that definition of "Vps21 activity" is unclear throughout the manuscript. In this study we have developed a novel biochemical method capable of detecting Vps21p activity with high sensitivity (Fig. 2) and utilized this method to measure Vps21p activity, which is clearly stated in the new manuscript. The reviewer #1 also pointed out the issue that we have not clearly explained about difference of two Vps21p-residing structures, small endosome-like puncta and aberrant large structure. To clearly distinguish them, in the new manuscript we have added data showing the size distribution of Vps21p-residing structures (Fig. S2). Regarding comment #2, we think that the reviewer may have misunderstood the data (please see the response to this comment described below). Reviewer #2 did not request any additional experiments but gave us many helpful comments to improve the manuscript. In the new manuscript, we have revised all the places that the reviewer pointed out.

    __ Point-by-point description of the revisions__

    __Reviewer #1 (Evidence, reproducibility and clarity (Required)): __

    (Reviewers’ comments are in italics)

    *Summary: *

    In the present study Nagano et al. identify an overlapping function of clathrin adaptors in the activation of the yeast Vps21 Rab GTPase. This activation is regulated in a concerted manner by two TGN cargo adaptors, AP-1 and GGA1/2. The basis of this study is derived from the previous work Nagano et al., 2019 where authors reported that Ent3p and Ent5p are important for the formation of the Vps21p-positive endosome. By utilizing a synthetic genetic approach, the authors observed that disruption/loss of the AP-1 complex (apl4 mutant), Ent3p, Ent5p or Pik1 decreased fluorescence intensity for GFP-Vps21p and increased number of Vps21p puncta. They found that these effects for AP-1 disruption are additive, that is, each makes a distinct contribution, at least in ent3∆/ent5∆ mutant cells. They next examined the role of factors required for TGN localization of Ent3p/5p and AP-1 in Vps21p activation. The authors reported that GGA1/2, Pik1p and the Ypt31/32 Rab GTPases make modest contributions to targeting of AP-1 and Ent3/5 to the TGN. The observation that accumulation of GFP-Vps21 next to vacuolar compartments in pik1-1 ent3D* mutants similar to that of ent3Dent5Dapl4D, lead authors to conclude that both PI(4)P as well as PI(4)P independent Ent3p recruitment to TGN plays a crucial role in Vps21p activation. Further they found that compared to the pik1-1 ypt31ts mutant (41%), activity of Vps21p (14%) was severely reduced in the pik1-1 ypt31ts gga1D gga2D mutant pointing towards redundancy among these factors in Vps21p activation. Finally using a class E Vps mutant authors found a fall in endosomal population of GFP-Vps9p ~29% in the ent3D ent5D mutant, which was further reduced to 0% in the ent3D ent5D apl4D mutant. Collectively this study suggests a differential role of TGN adaptors, AP-1 and GGA in early endosome formation. Ent3p/5p and AP-1 are proposed to activate Vps21p by localizing Vps9p on endosomes and thus facilitating its transport whereas GGAs act redundantly along with Pik1p and Ypt31/32 in regulating TGN localization of Ent3p/5p and AP-1. *

    *Major comments: *

    There is a considerable amount of data that address the roles of AP-1, Ent3, Ent5, Gga1/2, and Pik1 in targeting of Vps21 and related trafficking pathway components to the TGN/endosome. The experiments are essentially genetic epistasis tests that compare the fluorescence patterns of GFP-Vps21 in a sophisticated set of strains. The genetic data are interpreted in terms of spatiotemporal dynamics of Vps21: proportion Vps21GTP on a compartment and number of GFP-Vps21 positive compartments. *Being genetic in nature, the data are open to wide interpretations in terms of molecular mechanisms that target candidate proteins Vps21p and Vps9 to the TGN/endosome. The authors presentation (Fig. 7) is based on well controlled experiments and is logical, but key questions regarding Vps9 trafficking as it relates to Vps21 endosome formation are not resolved. *

    Response:

    In this study, in addition to comparison of the fluorescence patterns of GFP-tagged yeast Rab5 (Vps21p), we have developed a novel biochemical method capable of detecting the amount of active Vps21p with high sensitivity. The amount of active Vps21p obtained by this method correlated well with the results obtained by imaging analysis, and we think this approach significantly increased the reliability of our results.

    Using this new biochemical method and fluorescence imaging analysis, we have clarified the overall regulatory mechanisms of Vps21p by vesicle transport from the TGN. In particular, we believe that this is an important study that links the activation of Vps21p that mediates endosome formation with numerous previous studies involving vesicle transport from the TGN to the endosome.

    Comment #1(a)

      • Throughout their study the authors conflate measurements of GFP-Vps21 puncta intensity and number of Vps21p puncta as readouts of Vps21 "activity". Figure 7 exemplifies this especially: "Vps21p Activity: 100%; Vps21p Activity: 45%; Vps21p Activity: 10%". *
    2. *a) Would the authors please explicitly define how they use "activity" in the manuscript? * Response:

    We appreciate the reviewer’s pointing out our error. As the reviewer pointed out, since we have used the word “activity” when we explained the result obtained by the fluorescence intensity and the number of Vps21p puncta in lines 312-315 (in the new manuscript), we have revised this sentence “~ a decreased PI(4)P level reduces Vps21p activity and thus inhibits fusion of Vps21p compartments.” to “~a decreased PI(4)P level seems to inhibit fusion of Vps21p compartments.” (lines 314-315).

    In other parts of the manuscript, we have used the word “activity” only when we explained the result obtained by measuring the amount of active Vps21p by the biochemical method (Fig. 2). “Vps21p Activity” depicted in Fig. 7A-C are also based on the results obtained by the biochemical assay, and thus we have added explanatory sentences in the Discussion section (lines 432-433, 447) and figure legend (lines 996-998) in the new manuscript.

    Comment #1(b)

    1. *b) The amounts of Vps21-GTP were measured for the ent3D ent5 and ent3D ent5 apl4D mutants (Fig. 2). Other mutant backgrounds should be analyzed in order to address the specific requirements of gga1/2, pik1 and ypt31/32 genes and to challenge the assumption that aspects of GFP-Vps21 localization correlate with the proportion of Vps21GTP. * Response:

    We agree with the reviewer’s comment that it is crucial to confirm that aspects of GFP-Vps21 localization correlate with the proportion of Vps21GTP. In the previous manuscript, we have already measured the amount of active Vps21p (GTP-bound form of Vps21p) in the pik1-1, and pik1-1 ent3D mutants (Fig. 4E) and shown that it decreases to ~62% in the pik1-1 mutant, or to ~22% in the pik1-1 ent3D mutant relative to wild-type cells (Fig. 4E). The relative amount of GTP-bound form of Vps21p in these mutants correlated well with the results obtained by imaging analyses of GFP-Vps21p (Fig. 4B and C). To make it clearer, we have added sentences “and the amounts of active Vps21p in these mutants correlate well with the results obtained by imaging analyses of GFP-Vps21p (Fig. 4B, C, and H).” in lines 326-327. We have also demonstrated that the amount of active Vps21p correlated with the fluorescence intensity of GFP-Vps21p at puncta in the pik1-1 ypt31ts or the pik1-1 ypt31ts gga1D2D mutant (Figs 4F-J, S4E), and explained about this in lines 334-341.

    Comment #1(c)

    1. *c) Regarding the measurements of fluorescence intensity of GFP-Vps21 puncta, how were distinct puncta identified, particularly in the large clusters of puncta shown in Figs. 1D, 3A, 4F, 5A, 5C. * Response:

    As the reviewer pointed out, in the previous manuscript we have not clearly explained about how we had distinguished two Vps21p-residing structures, small endosome-like puncta and aberrant large structure. To clearly distinguish them, in the new manuscript we examined the size and number of these structures and showed the data in Fig. S2. This result revealed that the ent3D5D apl4D mutant contains single large Vps21p-residing structure with a size of >100 pixels and many small Vps21p-residing puncta with a size of ~50 pixels. To explain about this, we have added sentences in lines 235-239. Regarding Fig. 5A and 5C, since these figures do not show the localization of Vps21p, we have not added explanation about them.

    Comment #2

    • In the representative micrographs shown in Fig. 1A (Vph1-mCH), 1B (Hse1-tdTom), 1D (Sec7-mCH) and 5A, why do only (roughly) half of the cells in each micrograph express the tagged organelle marker protein? Shouldn't all of the cells? What is especially concerning is that the appearance of GFP-Vps9 in cells that express Sec7-mCH is different than in cells that do not. Specifically, there are fewer GFP-Vps9 puncta in expressing cells and GFP-Vps9 appears to be largely cytosolic in these cells. Have the authors noted the same? *

    Response:

    In Fig. 1, we expressed mCherry/tdTomato-tagged protein only in wild-type cells (Fig. 1A and B) or in ent3D5D mutants (Fig. 1D) to distinguish the mutant cells from the wild-type cells, as described in the Result section (lines 156-159) and figure legends. As explained in the text (lines 156-159), by labeling only wild-type or mutant cells, we precisely evaluated the differences in the localization of GFP-Vps21p by comparing mutant cells directly alongside wild-type cells.

    In Fig. 5A, we expressed Sec7-mCH only in the ent3D5D mutants to distinguish the mutants from wild-type cells (the upper panels) or the ent3D5D apl4D mutants (the lower panels), as described in figure legend. Therefore, the reviewer’s comment that “the appearance of GFP-Vps9 in cells that express Sec7-mCH is different than in cells that do not. Specifically, there are fewer GFP-Vps9 puncta in expressing cells and GFP-Vps9 appears to be largely cytosolic in these cells.” is exactly what we wanted to show in this figure. To show this more clearly, we labeled cells with “WT” or “mutant” in these micrographs (Fig. 1A, 1B, 1D, and 5A).

    Comment #3

    • Figure 4A: How were the proportional contributions of each factor to the TGN localization of Ent3/5, AP-1 determined? What do the percentiles indicate? *

    Response:

    As described in the Result section (lines 293-297), we have shown that deletion of the GGA1 and GGA2 genes significantly decreased the localization of Ent3-GFP at the TGN to ~33% of wild-type cell, without changing the localization of Ent5-GFP and Apl2-GFP (Fig. S3A, B). Based on these results, the contribution of Gga1/2p to the localization of Ent3p, Ent5p, or AP-1 was evaluated to be 37%, 0%, or 0%, respectively (Fig. 4A). To make this clearer, we have added sentence “~ and thus, we evaluated the contribution of Gga1p/2p to the localization of Ent3p, Ent5p, or AP-1 to be 37%, 0%, or 0%, respectively (Fig. 4A)” in line 296-297. Similarly, we have determined the contribution of PI(4)P by assessing the localization of Ent3p, Ent5p and Apl2p at the TGN in the pik1-1 (Fig. S3C and D), as described in lines 297-305. Regarding Rab11s (Ypt31p/32p), we have evaluated the contribution based on the data in our previous study, as described in line 305-309.

    Comment #4

    • In the model presented in Figure 7, the authors proposed that AP-1 is required to target Vps9 from the late TGN to the early TGN. The best characterized function of AP-1 is to concentrate integral membrane proteins to form the inner layer of a clathrin coated vesicle. Vps9 is a soluble protein that fractionates with cytosolic proteins (Burd et al., 1996). Despite measuring intensity and localizing Vps9p with different endosomal markers (Fig. 6), the basis of membrane recruitment of Vps9 by TGN clathrin adaptors is unclear. How do the authors envision AP-1 to function in targeting of Vps9, a soluble protein, between compartments? *

    Response:

    Like other many Rab-GEFs (e.g., Sec2p, the GEF for Sec4p or Mon1p/Ccz1p, the GEF for Rab7), we think that Vps9p transiently localizes to the donor organelle to activate Rab proteins and load them on the transport vesicle. We have previously demonstrated that Arf1p, a Golgi-resident GTPase, plays an important role in the recruitment of Vps9p to the Golgi (Nagano et al., Comm. Biol., 2019). In this study we have shown that deletion of AP-1 in the ent3D5D mutant increases the localization of Vps9p at the TGN (Fig. 5A and B). These suggest that AP-1, like Ent3p/5p (Nagano et al., Comm Bio, 2019), is dispensable for the recruitment of Vps9p to the TGN but required for the transport of Vps9p from TGN to endosomes.

    In a recent study Casler et al. proposed a role of AP-1 function that maintain Golgi-resident proteins by mediating intra-Golgi recycling pathway (Casler et al., JCB, 2021). Based on this model, we have speculated that AP-1 also functions to maintain Vps9p in the TGN by recycling from the late TGN to early TGN and discussed about this in the second paragraph of the Discussion section (lines 434-454 in the new manuscript). However, as the reviewer #2 pointed out (please see comment #6 of the reviewer #2), Casler et al proposed AP-1’s role in transport from the TGN back to earlier Golgi compartment but did not discuss compartmentalization within the TGN, we have modified sentence in the Discussion from “~ the role of AP-1 that recycles Vps9p back to the early TGN might become apparent” to “~ the role of AP-1 that recycles Vps9p back to the earlier Golgi compartment might become apparent” (lines 444-445).

    __Minor Comment: __

    • The interchangeable terminology used to refer to Rab GTPases throughout the manuscript made it exceptionally difficult for me to focus on the presentation of the experiments. Vps21 and Rab5 are used interchangeably, but this study investigated Vps21, not Rab5. Vps21 does not even appear in the title or abstract. Similarly, Ypt31/32 is used interchangeably with Rab11, but this study investigated Ypt31/32, not Rab11. The accurate names of the yeast proteins should be used. A discussion regarding significance of the yeast proteins for understanding mammalian Rab5 and Rab11 belongs in the Discussion. *

    Response:

    In accordance with the reviewer’s suggestion, we have replaced Rab5 with yeast Rab5 or Ypt21p. We have also replaced Rab11 with yeast Rab11 or Ypt31p/32p.

    __Reviewer #1 (Significance (Required)): __

    *General assessment: In general, this is a well-executed and controlled study. The major strengths are the large quantity of data from complementary experiments that provide a rationale for the proposed mechanistic model proposed (Fig. 7). The major weaknesses lie with the genetic approach, which does not lend itself to the mechanistic interpretations that the authors propose, and the narrow scope of the work such that the study will be of interest to a small group of colleagues. The audience will likely include researchers who use yeast to investigate proteins sorting in the endo-lysosome network of organelles and colleagues who investigate signaling by Rab GTPases. *

    Response:

    We cannot agree with the reviewer’s comment that “the narrow scope of the work such that the study will be of interest to a small group of colleagues”, because the regulation of endosome formation by Rab5 is one of the major topics in the field of membrane traffic, and many mechanisms still remain to be elucidated. Moreover, the model we have proposed in this study is adaptable not only to yeast but to higher organisms, as discussed in the last paragraph of the Discussion section. The endolysosomal pathway is important for the regulation of a wide variety of crucial cellular processes, including mitosis, antigen presentation, cell migration, cholesterol uptake, and many intracellular signaling cascades. Our work thus also has implications for development, immunity, and oncogenesis. We believe that the studies described in our paper represent an advance in our understanding of the cellular biology of endocytic trafficking and therefore would be interesting to researchers in other fields, as well as membrane traffic filed.

    __ __

    __Reviewer #2 (Evidence, reproducibility and clarity (Required)): __

    (Reviewers’ comments are in italics)

    *Summary: *

    *The manuscript by Nagano et al. describes the results of extensive analysis on the roles of clathrin adaptors for activation of Rab5 during TGN-to-endosome traffic in budding yeast. They examined the localization and activation status of Vps21, a major Rab5 member in yeast, in a variety of mutants and showed that AP-1 had a cooperative role with Epsin-related Ent3/5 in transport of Vps9 (Rab5 GEF) to endosomes. GGAs, PI4 kinase Pik1, and Ypt31/12 (Rab11) had partially overlapping functions in recruitment of AP-1 and Ent3/5 to TGN. *

    *It is an indeed extensive study but the interpretation of the results is complicated and somewhat speculative. It is most probably because the differences between mutants are partial (even though the authors tried to show statistics) and the logics to lead conclusions are not always compelling. To be honest, I had a hard time to follow rationales to justify arguments. The conclusions the authors make, that is, multiple clathrin adaptors cooperate in the TGN-to-endosome traffic, are reasonable, but I have several questions as follows, which I would like the authors to address. *

    Comment #1

      • The description about Vps21 fluorescence is often quite confusing. When the authors say fluorescence intensity, is it the total intensity of a whole cell or the average fluorescence intensity of individual puncta? For example, in Fig. 1D, it doesn't look to me at all that the GFP intensity of ent3/ent5 is lower than WT. How did the authors obtain the data of Fig. 1E? If the authors measured the fluorescence of individual puncta, how did they do it? * Response:

    We agree that in the previous manuscript explanation about how we measured Vps21p fluorescence intensity was insufficient. In this study, we have measured the whole fluorescence intensity of single GFP-Vps21p punctate structure, which was subtracted the cytoplasmic fluorescence background, and shown it as the fluorescence intensity of Vps21p compartment (the aberrant large GFP-Vps21p structure (Fig. 3A) were excluded). The graphs of fluorescence intensity of GFP-Vps21p show the average of three data (each average of 50 puncta) from three independent experiments. To clarify where and how Vps21 fluorescence was measured, in the new manuscript we have revised text (lines 160-161, 163, 166, 177, 179) and added explanatory sentences in “Materials and Methods” (lines 542-546).

    Regarding Fig. 1D and E, since the fluorescence intensity of GFP-Vps21p at the cytosol was increased in the ent3D5D mutant (Fig. 1D), the fluorescence intensity in the mutant may not have appeared lower than that in wild-type cell. To show the decrease of the fluorescence intensities of individual Vps21p puncta in the mutant cells more clearly, we have added the higher magnification view of GFP-Vps21p puncta in Fig. 1D in the new manuscript.

    Comment #2

    • Related to the previous question, how the images were taken is very important. In the legend to Fig.1, there is no description about the image analysis. Are they epifluorescence images or confocal images, and if the latter, are they ones of 2D confocal images or maximum intensity projections of Z stacks as mentioned in the legend to Fig. 3A? It matters very much. *

    Response:

    We appreciate the reviewer’s helpful suggestion. In Fig. 1, we have used epifluorescence images for analyzing the fluorescence intensity or number of GFP-Vps21p puncta, because Vps21p puncta have high mobility (please see also the responses to comment #9). In accordance with the reviewer’s suggestion, we have added the description about imaging method in the legend of Fig. 1 (lines 831-832, 837 and 843).

    Comment #3

    • It is also confusing when the authors say increase or decrease of fluorescence. Is it the intensity or the number of puncta? Please clarify which the authors intend to mention whenever relevant. There are many places that bother readers. *

    Response:

    We appreciate the reviewer’s helpful suggestion. In accordance with the reviewer’s suggestion, we have revised manuscript (lines 274 and 316).

    Comment #4

    • The method the authors developed to estimate the activation states of Vps21 is intriguing. It may provide important information without direct measurements of the GTP-binding activity. However, the results should be carefully interpreted because this kind of tricky experiments may not reflect the exact biochemical statuses in the cell. For example, I am concerned about whether release of GTP or spontaneous GTPase activity during the preparation processes is ignored. *

    Response:

    As the reviewer pointed out, we cannot rule out the possibility that the GTP-bound status might be changed during the preparation processes. However, this problem also occurs in the conventional pull-down assay, which assesses the amount of the GTP-bound form of Rab proteins. To confirm whether the activity of Vps21p assessed by this method reflects in vivo activation level, we have demonstrated that the level of active Vps21p correlated with the in vivo phenotypes, such as fluorescence intensity of GFP-Vps21p at the endosome and number of GFP-Vps21p puncta, that implicate defect of endosomal fusion. Thus, in the new manuscript we have added some sentences to explain about this (lines 221-222).

    Comment #5

    • In Discussion (p. 20, line 410), the authors describe that "Gga2p is localized predominantly at the Tlg2-residing compartment," but this is wrong. In the BioRxiv paper (2022), the authors showed that "Gga2p appears around the Sec7p-subcompartment and disappears at a similar time as Sec7p." I understand that, to explain the roles of GGAs in endosomal transport, it is reasonable to assume their presence in the Tlg2 compartment (and I agree on that), but the above description is wrong and must be corrected. *

    Response:

    We appreciate the reviewer’s helpful suggestion. As the reviewer described, we have recently demonstrated that Gga2p localization well overlapped with the Tlg2p-residing TGN sub-compartment that is structurally distinct from the Sec7p-residing sub-compartment (Toshima et al., BioRxiv, 2022). Thus, in accordance with reviewer's suggestion, we have changed this sentence to “Interestingly, Gga2p appears to reside at the Tlg2p sub-compartment, which is distinct from the Sec7p sub-compartment.” in the new manuscript (lines 427-428).

    Comment #6

    • Hypothesizing the role of AP-1 in the recycling from the late TGN to the early TGN is new. Glick's group proposed its role in transport from the TGN back to earlier compartment (Golgi) but did not discuss compartmentalization within the TGN. The authors' speculation is a fancy idea, but I am afraid there is no direct evidence for that. *

    Response:

    We appreciate the reviewer’s appropriate and helpful suggestion. As the reviewer pointed out, Glick's group has proposed its role in transport from the TGN back to earlier Golgi compartment, but not discussed compartmentalization within the TGN (Casler et al., 2021, JCB), and thus we modified sentence in the Discussion section from “~ the role of AP-1 that recycles Vps9p back to the early TGN might become apparent.” to “~ the role of AP-1 that recycles Vps9p back to the earlier Golgi compartment might become apparent.” (lines 444-445).

    Comment #7

    • The role of Ypt31/32 (Rab11) is also puzzling to me. It could be an indirect effect, which might be due to the complex network of GTPases as proposed by Chris Fromme (2014). Am I correct? *

    Response:

    As the reviewer pointed out, Fromme’s group has shown that Ypt31/32 forms the complex networks with several GTPases and their GEFs (McDonold and Fromme, 2014, Dev Cell; Thomas and Fromme, 2016, JCB, Thomas et al., 2019, Dev Cell), in which Ypt31/32 promotes the activation of Arf1p via its GEF Sec7p. We have previously shown that Arf1p plays an important role in the recruitment of Vps9p to the Golgi (Nagano et al., Comm. Biol., 2019). These findings suggest that disruption of Ypt31p/32p may affect the localization of Vps9p through reduced activity of Arf1p. However, arf1D and ypt31ts mutants exhibit different effects on the Vps9p localization: in arf1D mutant the recruitment of Vps9p to the TGN is impaired and in ypt31ts mutant Vps9p localization at the TGN is increased (Nagano et al., 2019, Comm Biol.). Thus, the role of Ypt31/32 in the Vps9p localization appears to be independent of Arf1p activity. In the new manuscript, we have added a brief discussion about this (lines 466-473).

    Comment #8

    • In the legend to Fig. 3D, the authors state that the read arrowheads indicate 50 nm vesicles and black arrowheads indicate vesicle clusters. However, the electron micrograph clearly shows that their morphologies are different. Red ones, which I estimate to be a little larger than 50 nm, often appear to have dense material inside, while those in black are even larger (probably around 200 nm) and do not look like a cluster of the same type of vesicles (I do not even think that such large structures should be called vesicles). How do the authors explain these differences? *

    Response:

    In the previous manuscript explanation about the electron microscopy analysis was insufficient. In the new manuscript, to clearly distinguish two Vps21p-residing structures, small endosome-like puncta and aberrant large structure, observed in ent3D5D apl4D mutant by fluorescence microscopy (Fig. 3A), we examined the size and number of these structures and showed the data in Fig. S2. This result revealed that the ent3D5D apl4D mutant contains single aberrant large aggregate with a size of >100 pixel adjacent to the vacuole and endosome-like structures with a size of Comment #9

    • In Fig. 4F, the authors show different sets of images, Focal plane and Z projection. What is the purpose to do it? The results with Z projection should be more informative. Why the authors use only Focal plane data for the analysis in panel G? *

    Response:

    We measured the fluorescence intensity or number of individual GFP-Vps21p puncta using a single focal plane images (Figs. 1C, 1E, 3I, and 4B), because Vps21p-residing small puncta have high mobility and identical endosome often appears in multiple different planes in the Z-stack image taken by a conventional epifluorescence microscope. In contrast, we analyzed the aberrant large aggregate using Z projection image (Figs. 3B, S3G) because this structure is relatively stable and low motile, and not observed if it is not in the focal plane. In Fig. 4F, since both of small puncta and large aggregate are analyzed, we have shown both of focal plane image and Z-projection image. In new manuscript, we have added about the description about imaging method in each figure legend or text (lines 230-232, 332-334).

    __Reviewer #2 (Significance (Required)): __

    *It is a complicated story but I find most of the conclusions reasonable. It provides important knowledge to the understanding on the Rab5 GTPase regulation in trafficking from the TGN. *

    Response:

    We are very grateful for this reviewer’s favorable evaluation of our studies.

  2. Review Commons

    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 Nagano et al. describes the results of extensive analysis on the roles of clathrin adaptors for activation of Rab5 during TGN-to-endosome traffic in budding yeast. They examined the localization and activation status of Vps21, a major Rab5 member in yeast, in a variety of mutants and showed that AP-1 had a cooperative role with Epsin-related Ent3/5 in transport of Vps9 (Rab5 GEF) to endosomes. GGAs, PI4 kinase Pik1, and Ypt31/12 (Rab11) had partially overlapping functions in recruitment of AP-1 and Ent3/5 to TGN.

    It is an indeed extensive study but the interpretation of the results is complicated and somewhat speculative. It is most probably because the differences between mutants are partial (even though the authors tried to show statistics) and the logics to lead conclusions are not always compelling. To be honest, I had a hard time to follow rationales to justify arguments. The conclusions the authors make, that is, multiple clathrin adaptors cooperate in the TGN-to-endosome traffic, are reasonable, but I have several questions as follows, which I would like the authors to address.

    1. The description about Vps21 fluorescence is often quite confusing. When the authors say fluorescence intensity, is it the total intensity of a whole cell or the average fluorescence intensity of individual puncta? For example, in Fig. 1D, it doesn't look to me at all that the GFP intensity of ent3/ent5 is lower than WT. How did the authors obtain the data of Fig. 1E? If the authors measured the fluorescence of individual puncta, how did they do it?
    2. Related to the previous question, how the images were taken is very important. In the legend to Fig.1, there is no description about the image analysis. Are they epifluorescence images or confocal images, and if the latter, are they ones of 2D confocal images or maximum intensity projections of Z stacks as mentioned in the legend to Fig. 3A? It matters very much.
    3. It is also confusing when the authors say increase or decrease of fluorescence. Is it the intensity or the number of puncta? Please clarify which the authors intend to mention whenever relevant. There are many places that bother readers.
    4. The method the authors developed to estimate the activation states of Vps21 is intriguing. It may provide important information without direct measurements of the GTP-binding activity. However, the results should be carefully interpreted because this kind of tricky experiments may not reflect the exact biochemical statuses in the cell. For example, I am concerned about whether release of GTP or spontaneous GTPase activity during the preparation processes is ignored.
    5. In Discussion (p. 20, line 410), the authors describe that "Gga2p is localized predominantly at the Tlg2-residing compartment," but this is wrong. In the BioRxiv paper (2022), the authors showed that "Gga2p appears around the Sec7p-subcompartment and disappears at a similar time as Sec7p." I understand that, to explain the roles of GGAs in endosomal transport, it is reasonable to assume their presence in the Tlg2 compartment (and I agree on that), but the above description is wrong and must be corrected.
    6. Hypothesizing the role of AP-1 in the recycling from the late TGN to the early TGN is new. Glick's group proposed its role in transport from the TGN back to earlier compartment (Golgi) but did not discuss compartmentalization within the TGN. The authors' speculation is a fancy idea, but I am afraid there is no direct evidence for that.
    7. The role of Ypt31/32 (Rab11) is also puzzling to me. It could be an indirect effect, which might be due to the complex network of GTPases as proposed by Chris Fromme (2014). Am I correct?
    8. In the legend to Fig. 3D, the authors state that the read arrowheads indicate 50 nm vesicles and black arrowheads indicate vesicle clusters. However, the electron micrograph clearly shows that their morphologies are different. Red ones, which I estimate to be a little larger than 50 nm, often appear to have dense material inside, while those in black are even larger (probably around 200 nm) and do not look like a cluster of the same type of vesicles (I do not even think that such large structures should be called vesicles). How do the authors explain these differences?
    9. In Fig. 4F, the authors show different sets of images, Focal plane and Z projection. What is the purpose to do it? The results with Z projection should be more informative. Why the authors use only Focal plane data for the analysis in panel G?

    Significance

    It is a complicated story but I find most of the conclusions reasonable. It provides important knowledge to the understanding on the Rab5 GTPase regulation in trafficking from the TGN.

  3. Review Commons

    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

    Summary:

    In the present study Nagano et al. identify an overlapping function of clathrin adaptors in the activation of the yeast Vps21 Rab GTPase. This activation is regulated in a concerted manner by two TGN cargo adaptors, AP-1 and GGA1/2. The basis of this study is derived from the previous work Nagano et al., 2019 where authors reported that Ent3p and Ent5p are important for the formation of the Vps21p-positive endosome. By utilizing a synthetic genetic approach, the authors observed that disruption/loss of the AP-1 complex (apl4 mutant), Ent3p, Ent5p or Pik1 decreased fluorescence intensity for GFP-Vps21p and increased number of Vps21p puncta. They found that these effects for AP-1 disruption are additive, that is, each makes a distinct contribution, at least in ent3∆/ent5∆ mutant cells. They next examined the role of factors required for TGN localization of Ent3p/5p and AP-1 in Vps21p activation. The authors reported that GGA1/2, Pik1p and the Ypt31/32 Rab GTPases make modest contributions to targeting of AP-1 and Ent3/5 to the TGN. The observation that accumulation of GFP-Vps21 next to vacuolar compartments in pik1-1 ent3∆ mutants similar to that of ent3∆ent5∆apl4∆, lead authors to conclude that both PI(4)P as well as PI(4)P independent Ent3p recruitment to TGN plays a crucial role in Vps21p activation. Further they found that compared to the pik1-1 ypt31ts mutant (41%), activity of Vps21p (14%) was severely reduced in the pik1-1 ypt31ts gga1∆ gga2∆ mutant pointing towards redundancy among these factors in Vps21p activation. Finally using a class E Vps mutant authors found a fall in endosomal population of GFP-Vps9p ~29% in the ent3∆ ent5∆ mutant, which was further reduced to 0% in the ent3∆ ent5∆ apl4∆ mutant. Collectively this study suggests a differential role of TGN adaptors, AP-1 and GGA in early endosome formation. Ent3p/5p and AP-1 are proposed to activate Vps21p by localizing Vps9p on endosomes and thus facilitating its transport whereas GGAs act redundantly along with Pik1p and Ypt31/32 in regulating TGN localization of Ent3p/5p and AP-1.

    Major comments:

    There is a considerable amount of data that address the roles of AP-1, Ent3, Ent5, Gga1/2, and Pik1 in targeting of Vps21 and related trafficking pathway components to the TGN/endosome. The experiments are essentially genetic epistasis tests that compare the fluorescence patterns of GFP-Vps21 in a sophisticated set of strains. The genetic data are interpreted in terms of spatiotemporal dynamics of Vps21: proportion Vps21GTP on a compartment and number of GFP-Vps21 positive compartments. Being genetic in nature, the data are open to wide interpretations in terms of molecular mechanisms that target candidate proteins Vps21p and Vps9 to the TGN/endosome. The authors presentation (Fig. 7) is based on well controlled experiments and is logical, but key questions regarding Vps9 trafficking as it relates to Vps21 endosome formation are not resolved.

    1. Throughout their study the authors conflate measurements of GFP-Vps21 puncta intensity and number of Vps21p puncta as readouts of Vps21 "activity". Figure 7 exemplifies this especially: "Vps21p Activity: 100%; Vps21p Activity: 45%; Vps21p Activity: 10%".
      • a) Would the authors please explicitly define how they use "activity" in the manuscript?
      • b) The amounts of Vps21-GTP were measured for the ent3D ent5 and ent3D ent5 apl4D mutants (Fig. 2). Other mutant backgrounds should be analyzed in order to address the specific requirements of gga1/2, pik1 and ypt31/32 genes and to challenge the assumption that aspects of GFP-Vps21 localization correlate with the proportion of Vps21GTP.
      • c) Regarding the measurements of fluorescence intensity of GFP-Vps21 puncta, how were distinct puncta identified, particularly in the large clusters of puncta shown in Figs. 1D, 3A, 4F, 5A, 5C.
    2. In the representative micrographs shown in Fig. 1A (Vph1-mCH), 1B (Hse1-tdTom), 1D (Sec7-mCH) and 5A, why do only (roughly) half of the cells in each micrograph express the tagged organelle marker protein? Shouldn't all of the cells? What is especially concerning is that the appearance of GFP-Vps9 in cells that express Sec7-mCH is different than in cells that do not. Specifically, there are fewer GFP-Vps9 puncta in expressing cells and GFP-Vps9 appears to be largely cytosolic in these cells. Have the authors noted the same?
    3. Figure 4A: How were the proportional contributions of each factor to the TGN localization of Ent3/5, AP-1 determined? What do the percentiles indicate?
    4. In the model presented in Figure 7, the authors proposed that AP-1 is required to target Vps9 from the late TGN to the early TGN. The best characterized function of AP-1 is to concentrate integral membrane proteins to form the inner layer of a clathrin coated vesicle. Vps9 is a soluble protein that fractionates with cytosolic proteins (Burd et al., 1996). Despite measuring intensity and localizing Vps9p with different endosomal markers (Fig. 6), the basis of membrane recruitment of Vps9 by TGN clathrin adaptors is unclear. How do the authors envision AP-1 to function in targeting of Vps9, a soluble protein, between compartments?

    Minor Comment:

    1. The interchangeable terminology used to refer to Rab GTPases throughout the manuscript made it exceptionally difficult for me to focus on the presentation of the experiments. Vps21 and Rab5 are used interchangeably, but this study investigated Vps21, not Rab5. Vps21 does not even appear in the title or abstract. Similarly, Ypt31/32 is used interchangeably with Rab11, but this study investigated Ypt31/32, not Rab11. The accurate names of the yeast proteins should be used. A discussion regarding significance of the yeast proteins for understanding mammalian Rab5 and Rab11 belongs in the Discussion.

    Significance

    General assessment: In general, this is a well-executed and controlled study. The major strengths are the large quantity of data from complementary experiments that provide a rationale for the proposed mechanistic model proposed (Fig. 7).

    The major weaknesses lie with the genetic approach, which does not lend itself to the mechanistic interpretations that the authors propose, and the narrow scope of the work such that the study will be of interest to a small group of colleagues. The audience will likely include researchers who use yeast to investigate proteins sorting in the endo-lysosome network of organelles and colleagues who investigate signaling by Rab GTPases.

  4. Version published to 10.1101/2023.03.27.534325v1 on bioRxiv