Ubiquitin ligase and signalling hub MYCBP2 is required for efficient EPHB2 tyrosine kinase receptor function

Reviewer #1 (Public Review):

The Eph receptor tyrosine kinase family plays a critical function in multiple physiological and pathophysiological processes. Hence, understating the regulation of these receptors is highly important question. Through extensive experiments in cell lines and cultured neurons Chang et.al show that the signaling hub protein, MYCBP2 positively regulates the overall stability of a specific member of the family, EPHB2, and by that the cellular response to ephrinBs.
Overall, this work sheds light on the divergent in the regulatory mechanisms of the Eph receptors family. Although the physiological importance of this new regularly mechanism in mammals awaits to be discovered, the authors provide genetic evidence using C.elegans that it is evolutionarily conserved.

Reviewer #2 (Public Review):

Members of the EphB family of tyrosine kinase receptors are involved in a multitude of diverse cellular functions, ranging from the control of axon growth to angiogenesis and synaptic plasticity. In order to provide these diverse functions, it is expected that these receptors interact in a cell-type specific manner with a diverse variety of downstream signalling molecules.

The authors have used proteomics approaches to characterise some of these molecules in further detail. This molecule, myc-binding protein 2 (MYCBP2) is also known as highwire, has been identified in the context of establishment of neural connectivity. Another molecule coming up on this screen was identified as FBXO45.

The authors use classical methods of co-IP to show a kinase-independent binding of MYCBP2 to EphB2. They further showed that FBXO45 within a ternary complex increased the stability of the EphB2/MYCBP2 complex.

To define the interacting domains, they used clearly designed swapping experiments to show that the extracellular and transmembrane domains are necessary and sufficient for the formation of the ternary complex.

Using a cellular contraction assay, the authors showed the necessity of MYCBP2 in mediating the cytoskeletal response of EphB2 forward signalling. Furthermore, they used the technically challenging stripe assay of alternating lanes of ephrinB-Fc and Fc to show that also in this migration-based essay MYCBP2 is required for EphB mediated differential migration pattern.

MYCBP2 in addition is necessary to stabilize EphB2, that is in the absence of MYCBP2, EphB2 is degraded in the lysosomal pathway.

Interestingly, the third protein in this complex, Fbxo45, was further characterized by overexpression of the domain of MYCBP2, known to interact with Fbxo45. Here the authors showed that this approach led to the disruption of the EphB2 / MYCBP2 complex, and also abolished the ephrinB mediated activation of EphB2 receptors and their differential outgrowth on ephrinB2-Fc / Fc stripes.

Finally, the authors demonstrated an in vivo function of this complex using another model system, C elegans where they were able to show a genetic interaction.

Data show in a nice set of experiments a novel level of EphB2 forward signalling where a ternary complex of this receptor with multifunctional MYCBP2 and Fbxo45 controls the activity of EphB2, allowing a further complex regulation of this important receptors. Additionally, the authors challenge pre-existing concepts of the function of MYCBP2 which might open up novel ways to think about this protein.
Of interest is this work also in terms of development of the retinotectal projection in zebrafish where MYCBP2/highwire plays a crucial role, and thus might lead to a better understanding of patterning along the DV axis, for which it is known that EphB family members are crucial.

Overall, the experiments are classical experiments of co-immunoprecipitations, swapping experiments, collapse assays, and stripe assays which all are well carried out and are convincing.

Reviewer #3 (Public Review):

In this improved version of the manuscript, Chang et al set out to find direct interactions with the Eph-B2 receptor, as our knowledge of its function/regulation is still incomplete. Using proteomic analysis of Hela cells expressing EPHB2, they identified MYCBP2 a potential binder, which they then confirm using extensive biochemical analyses, an interaction that seems to be negatively affected by binding of ephrin-B2 (but not B1). Furthermore, they find that FBXO45, a known MYCBP2 interaction, strongly facilitates its binding to EPHB2. Intriguingly, these interactions depend on the extracellular domains of EPHB2, suggesting the involvement of additional proteins as MYCBP2 is thought to be a cytoplasmic protein. Finally, they find that, in contrast to what could be expected given the known function of MYCBP2 as a ubiquitin E3 ligase, it actually positively regulates EPHB2 protein stability, and function.

The strength of this manuscript is the extensive biochemical analysis of the EPHB2/MYCBP2/FBXO43 interactions. The vast majority of the conclusions supported by the data.

The attempt to extend the study to an in vivo animal using the worm is important, however the additive insight is, unfortunately, minimal.

Author Response

The following is the authors’ response to the original reviews.

We thank the reviewers for their thoughtful assessment of our work and their valuable critiques which we will address in the “Recommendations for the authors” section below. In particular, we appreciate Reviewer #3 noting the value of the C. elegans model system and our efforts to bridge models with our study. We agree with the reviewer that there is a need to clarify the rationale, presentation and interpretation of our results. We have substantially revised the text in our manuscript and Figure legend to address this issue, and provided extensive new commentary and citations to lay out the logic behind our experiments. Indeed, it was our oversight not being more thorough about this initially. We have further adjusted our conclusions to be less unequivocal. Finally, we added an RPM-1 signaling diagram (Fig. 8A) to more clearly annotate the players in the RPM-1/MYCBP2 signaling network that were evaluated genetically in Fig. 8. Importantly, we provide clearer commentary on how genetic enhancer effects with known RPM-1 binding proteins and the absence of genetic suppression in vab-1/Eph receptor double mutants with components of the RPM-1/FSN-1 ubiquitin ligase complex are consistent with the biochemical finding that MYCBP2 stabilizes but does not degrade EphB2. Text edits reflecting these points are in the abstract, the C. elegans results section starting on line 411, and the discussion on lines 499, 502-504 and 541.

Following extensive discussions between the three reviewers, all three agree that the C. elegans data, as presented, does not add to, and in fact might harm, your bottom line. Our combined suggestion is to take this data out unless you plan to improve it substantially. All reviewers are perplexed by Figure 2F and the presumed interactions of cytosolic proteins with the extracellular domain of EPHB2. At the very least, please provide some suggestions/model/interpretation.

We have adjusted our manuscript substantially to address this. Please see detailed comments in the individual Reviewer sections below.

We would like to thank the reviewers for their thorough examination of our manuscript, constructive criticisms, and helpful suggestions.

Reviewer #1 (Recommendations For The Authors):

The work is extensive in my view, and mostly of high quality. See minor comments on some of the figures below.

Thank you very much.

Two more major comments :

• I don't think the C. elegans work adds to - in fact I think it hurts - the statement that this regulatory mechanism is specific to EphB2. I would advise the authors to take it out.

We agree that C. elegans has a sole Eph receptor called VAB-1 and is therefore not a specific model for EPH2B. However, testing MYCBP2 specificity for EPHB2 was not the goal or our perceived value for the C. elegans experiments. We now clarify this in the text of the Results section.

Rather, we are providing evidence that the C. elegans ephrin receptor interacts genetically with known MYCBP2/RPM-1 binding proteins. Moreover, we now provide an extensive array of citations to note that genetic enhancer interactions between different RPM-1/MYCBP2 binding proteins is well established. The reviewer has nicely highlighted for us that we handled the C. elegans genetics in too cursory a fashion in our original manuscript. We appreciate this being noted and have now aimed to make this substantially clearer. We hope the reviewer agrees that our revised C. elegans section accomplishes this goal.

Furthermore, we extensively revised the text of the Results to emphasize a key point: our observation that axon termination defects are not suppressed in vab-1; fsn-1 and vab-1; rpm-1 double mutants excludes the possibility that the VAB-1 Eph receptor is a substrate that is inhibited or degraded by the RPM-1/FSN-1 ubiquitin ligase complex. If the VAB-1 Eph receptor were ubiquitinated and degraded by the RPM-1/FSN-1 complex, we would have observed a suppression of phenotype in vab-1; rpm-1 double mutants. The precedent for this genetic relationship between the RPM-1 ubiquitin ligase and its substrates that are degraded has been established by several prior studies (PMID: 15707898; PMID: 31676756; PMID: 35421092). We now more clearly note that the absence of genetic suppression in vab-1; rpm-1 double mutants and vab-1; fsn-1 double mutants is consistent with the non-canonical stabilizing role of MYCBP2 on EPHB2 that was observed in our biochemical experiments with mammalian cells.

We also adjusted the text of the manuscript to stress that we are testing genetic interactions between the VAB-1 Eph receptor and known RPM-1 binding proteins. This is a key point, as genetic enhancer interactions are consistent with the Eph receptor functioning in the RPM-1 signaling network. This concept has been well established for RPM-1 binding proteins as now noted in our revised text with an extensive number of additional citations to published work.

Based on the above arguments, we respectfully disagree with the reviewer that our C. elegans data should be removed from the paper. To re-iterate, we are not trying to evaluate specificity for MYCBP2 and EPHB2 in C. elegans. Rather, our goals are twofold: 1) To ask whether there is an evolutionarily conserved functional genetic link between Eph receptors and known RPM-1 binding proteins. 2) To provide further in vivo genetic evidence invalidating the hypothesis that Ephrin receptors could be ubiquitination substrates that are inhibited/degraded by MYCBP2.

Text edits reflecting these points are in the abstract, the C. elegans results section starting on line 411, and the discussion on lines 499, 502-504 and 541.

• The cellular responses are not robust and the effects of MYCBP2 KO - although significant - are minor in most cases. But I don't think more experiments will help here.

We interpret the comment about the robustness to mean that the extent to which a given cellular response is affected by the loss of MYCBP2 is minor. First, the cellular responses themselves are typical of previous studies and depend on the cellular biology underlying them. For example, a growth collapse of ~50-60% over a background of 10% (Fig. 7) is typical for these sorts of assays (PMID: 37369692; PMID: 33972524; PMID: 17785182). A decrease of cell area by ~25% (Fig. 3) is quite substantial if one considers how much of a cell’s volume is taken up by the nucleus and organelles. Second, the phenotypes elicited by the loss of MYCBP2 are likely brought on by a decrease in EphB2 protein levels, but not its complete absence, as suggested by our biochemical experiment. Given that EphB2 complete loss only affects the cellular responses to a limited extent, the minor effects are not a surprise (e.g. for GC collapse: PMID: 23143520). Nevertheless, the subtle changes in cellular phenotypes, elicited by EPHB2 signaling are often sufficient to achieve proper cell positioning and cell response to guidance cues. For instance, regulation of the growth cone collapse of the outgrowing axons requires delicate changes that are dynamic and temporal.

Minor:

Fig 1C - EPHA3 and EPHB2 seem to run in different sizes, is this the case? In 2A they run at the same size.

We believe this size discrepancy is due to different percentages of SDS-PAGE gels used to resolve proteins. In Fig. 1C, we used a 6% gel for a Western blot analysis of both EPHA3/-B2-FLAG (~130 kDa) and MYCBP2 (~510 kDa). In Fig. 2A however, we performed Western blot analysis using 10% resolving gel to separate and detect EPHA3/-B2-FLAG along with MYC-FBXO45 (~30 kDa). We have reviewed the results obtained from additional biological replicates of this experiment, and observed a similar pattern in gel migration of EPHA3/-B2-FLAG across all replicates.

Fig1F - I can't trust the MYCBP2 blot.

Indeed, the MYCBP2-EPHB2 co-IP with endogenous proteins was not convincing. We now repeated this experiment using rat cortical neurons, and the results replace the previous Fig. 1F panel as mentioned on line 158.

In Fig2b the authors claim that there is enhancement in the binding of MYCBP2 and EPHB2 upon FBXO45 expression. For this type of statement quantification is required.

The quantification is now included in Fig. 2C and its significance is mentioned on line 180. Our conclusion about the enhancement stands.

Fig2G - it remained unclear to me where the binding site to MYCBP2 is, how long is the cytoplasmic tail in the DeltaICD protein?

Based on our experimental observations from Fig. 2E-H, we concluded that the fragment encompassing the extracellular domain(s) and/or transmembrane (TM) domain of EPHB2 is necessary for the protein complex formation with MYCBP2. We would like to accentuate that the EPHB2-MYCBP2 interaction might not be direct, and might involve other transmembrane protein(s) acting as a scaffold for EPHB2 and MYCBP2 binding. We did not pursue experiments to determine the exact region of the extracellular-TM portion of EPHB2 that is required for the interaction with MYCBP2.

The cytoplasmic tail in ΔICD protein consists of 25 aa of the N-terminal fragment of EPHB2 juxtamembrane (JM) region, which is adjacent to the TM helix, and followed by the 8 aa FLAG tag (EPHB2 ΔICD domain composition: extracellular domains – TM domain – 25 aa fragment of JM region – FLAG). We have determined the TM and JM sequences based on Hedger et al. (PMID: 25779975) and included the N-terminal portion of the JM region to facilitate proper ΔICD protein localization within the plasma membrane (PMID: 35793621). We modified the schematic in Fig. 2G to better visualise the EPHB2 truncations and now provide information on their size in the figure legend.

Always good to have a model of how all these proteins work together.

While we acknowledge that this would be helpful, we do not have a clear answer on how the EPHB2-MYCBP2 complex formation occurs. This requires further elucidation of the putative proteins involved in this ternary complex or testing the possibility that a MYCBP2 fragment is extruded extracellularly. Without these experiments there are too many possibilities to summarise into a clear model figure. We thus did not make any edits regarding these possibilities in the section starting on line 195.

Reviewer #2 (Recommendations For The Authors):

Overall, the experiments are classical experiments of co-immunoprecipitations, swapping experiments, collapse assays, and stripe assays which all are well carried out and are convincing.

Thank you for your encouraging comments.

Controls for the stripe assay may include Fc / Fc stripe assays.

We have performed these control experiments and now include their quantifications in the results sectioning concerning Fig. 3, starting on line 249, and those concerning Fig. 6 on line 381.

It is not clear to me why SD and not SEM has been used here for presentations.

Standard deviation (SD) measures the dispersion of a dataset relative to its mean. The standard error of the mean (SEM) measures how much discrepancy is likely in a sample’s mean compared with the population mean. Thus, SEM includes a statistical inference about the sampling distribution while SD is a less “processed” measurement that by definition is larger than SEM. SEM might make the data look less dispersed and many journals encourage the use of SD in bar graphs (PMID: 16223828).

Fig 7A: it is rather difficult to see 'branches' in Fig. 7A, better pictures and close-ups should be provided. How are branches defined? This piece of work needs more attention.

To remedy this shortcoming, we now provide inverted images with GFP signal in dark pixels overlaid on Fc (white) / eB2 (pink) stripes next to the original images.

Reviewer #3 (Recommendations For The Authors):

1. My most important suggestion to the authors would be to more carefully describe the results and their interpretation of the results. Sometimes, the distinction is not clear.

We modified the text throughout the manuscript to address this.

1. There are several cases, when the authors report on trends that are not statistically significant (1D, for example), or report no change, when it is clear that the addition of one more sample could have dramatically made a difference (4M - see point 12).

We agree that some of the nonsignificant differences could become significant if we added more Ns. But we prefer not to move our experimental design towards N-chasing and p-hacking (PMID: 25768323). The number of biological replicates is normally pre-determined before the onset of the experiment. Of course, some replicates can be discarded if there is a valid reason, such as a technical issue with the experiment or a positive control not working but this is not relevant for the dataset we have provided.

1. Data in 1F is very difficult to interpret.

As in response to Reviewer #1: Indeed, the MYCBP2-EPHB2 co-IP with endogenous proteins was not convincing. We now repeated this experiment using rat cortical neurons, and the improved results are in revised Fig. 1F.

1. Figure 2 puts Figure 1 in a strange perspective. If I understand correctly, fig 2 claims that EPHB2 interaction with MYCBP2 depends on FBXO45 - if that is the case then how does the binding in Figure 1 occur?

Indeed, we propose that the EPHB2-MYCBP2 interaction depends on FBXO45. In Fig. 2, we reveal that FBXO45 enhances the formation of the EPHB2-MYCBP2 complex. Thus, we suspect that the endogenous FBXO45 present in HeLa cells and neurons would mediate the interaction between EPHB2 and MYCBP2 in Fig. 1 experiments. We were unable to show this by Western blotting due to lack of reliable commercial antibodies against FBXO45, the complex containing endogenous FBXO45 and EPHB2 is also implied by our AP-MS data (Fig. 1B) and published databases.

1. I am still trying to wrap my mind around the results in 2G-H. So do MYCBP2 and FBXO45 bind the extracellular domain of EPHBP2? What does that mean?

(see also our response to Reviewer #1, end of their section) Based on our experimental observations from Fig. 2G-H, we conclude that the fragment encompassing the extracellular domain(s) and/or transmembrane domain of EPHB2 is necessary for the protein complex formation with MYCBP2 and FBXO45. Although there is a possibility that MYCBP2 directly binds the extracellular portion of EPHB2, we have not formally tested this hypothesis. MYCBP2 has been previously shown to interact with the extracellular portion of transmembrane N-cadherin (CDH2) via BioID proximity labeling and AP-MS proteomics approaches (PMID: 32341084).

Considering the results in Fig. 2A-B, we suspect that EPHB2-MYCBP2 interaction is indirect, as FBXO45 enhances this association. Secretion of FBXO45 and direct binding of FBXO45 to the extracellular cadherin (EC1-2) domains of N-cadherin has been documented (PMID: 25143387; PMID: 32341084). Although, not tested, this is also a possibility for EPHB2-FBXO45 mode of interaction. Nevertheless, we also cannot rule out the possibility that an unknown transmembrane protein binds EPHB2 extracellularly and the same unknown protein binds MYCBP2/FBXO45 intracellularly. Resolving this model is beyond the scope of this study and will require us to pursue extensive new lines of investigation.

1. I don't understand the stable Hela cell line CRISPR - is this a stable MYCBP2 deletion? In which case why is there only a reduction, not complete elimination of the protein? Or, is this a stable integration of a plasmid generating gRNA against MYCBP2? In which case, I would expect a homozygous null to emerge at some point. In any case, this is not well explained.

These lines are not derived from single cells infected with the CRISPR sgRNA-carrying viruses, therefore they are not clonal and probably contain some cells that express normal levels of MYCBP2, hence its detection on a Western. This is now clarified starting on line 221 and on line 608.

1. In 3C - is this the right statistical analysis?? I would say you want to claim the different effect of the control +/- eB2 compared to the effect in the mutant +/- eB2. Still should be significant but I think a more correct analysis.

We now include this comparison in Fig. 3C as well in the results section starting on line 234.

1. The robustness of the assay in Figure 3D is underwhelming – how was the area measured?

This is a live imaging experiment. Fig. 3D plots cell area at 60 minutes after ephrin-B2 addition as a fraction of the same cell’s area at 0 minutes (ephrin-B2 addition). For control cells that is a decrease of ~25%. If one considers that a cell’s nucleus and organelles like the Golgi Apparatus take up most of its volume, the magnitude is not that surprising.

1. Figure 3F – did you try to plot the relative area of overlap divided by the total cellular area? You might get a more striking phenotype. Also – claiming that this confirms that MYCBP2 is REQUIRED for EPHB2 function is a bit overstated, especially given that we don’t know (do you?) the EPHB2 mutant phenotype in this assay.

We preferred to stay with the original method of image quantification which we use for other assays. With respect to the requirement of MYCBP2 for EPHB2 function in the stripe assay, our logic is rooted in the observation that native HeLa cells do not respond to ephrin-B2 stripes (45.46 ± 7.62% of cells on eB2 stripes v. Fc; data not shown). When they are transfected with EPHB2 expression plasmids they do, therefore we assume that EPHB2 expression endows them with a sensitivity to eB2 stripes. A loss of MYCBP2 attenuates this sensitivity. We clarified this starting on line 246 and on line 251.

1. I didn't quite get the difference between 4A and 4B.

We apologize for the confusion. In Fig 4A, we used a stable HeLa cell line that has tetracycline-inducible expression of EPHB2-FLAG. Using these cells, we subsequently generated CTRLCRISPR or MYCBP2CRISPR cells. In these cells we then induced EPHB2 expression with tetracycline and observed that deletion of MYCBP2 resulted in the reduction of EPHB2 protein levels. To confirm this observation and to rule out the possibility that EPHB2 protein reduction is an effect of the CRISPR lines generation, we tested whereas MYCBP2 deletion reduces EPHB2, which has been transiently overexpressed (Fig. 4B). We hence conclude that loss of MYCBP2 decreases EPHB2 that was either expressed from a stable locus (Fig. 4A) or from transient transfection (Fig. 4B). We modified the Results section starting on line 262 to make this point clear.

1. The entire link to lysosomal degradation should be strengthened. Perhaps I am confused, but if the reduced EPHB2 levels in MYCBP2 mutant cells result from impaired lysosomal degradation then inhibiting the lys-deg should bring the protein levels back to normal (i.e. CRISPR control) - no? As currently presented, I do not understand nor do I think the claim is strongly supported by the data.

Before treatment with inhibitors, EPHB2 levels in MYCBP2CRISPR cells are already 40% lower than they are in CTRLCRISPR cells and in all our attempts, inhibitors can only rescue/restore EPHB2 in MYCBP2CRISPR cells to a level that is lower than in CTRLCRISPR cells. But this restoration is greater in MYCBP2CRISPR than in MYCBP2CTRL cells (BafA1: 19% increase in CTRL cells and 40% in MYCBP2CRISPR cells; CoQ: 10% comparing to 35%). This indicates that EPHB2 degradation through the lysosomal pathway in MYCBP2CRISPR cells is stronger, explaining why EPHB2 degradation is promoted in MYCBP2CRISPR cells, compatible with reduced EPHB2 levels and enhanced EPHB2 ubiquitination.

1. 4M, O - reporting ns based on these data seems a bit strange to me... Add one point and it will be strongly significant.

See our response to point (2), above. We prefer not to invoke potential p-hacking.

1. 7d - so what are you claiming? That the cellular response to eB1 but not eB2 is affected by the addition of FBD1? this is almost the opposite of what you wrote in the text...

We treated the cells with two different ephrin-B ligands to make a stronger conclusion. When using ephrin-B1, growth cone collapse in FBD1 WT is not significant comparing to Fc treatment. When using ephrin-B2, growth cone collapse in FBD1 WT is not as significant as it is in FBD1 mut group (*** versus **). We interpret this as meaning that the EPHB2-mediated growth cone collapse to both ligands is dampened, when we disrupt the EPHB2-MYCBP2 association. The difference between these two ligands might be due to their different affinities for the receptor or signalling kinetics.

1. By far the weakest link in this paper is the worm part. I think it's a pity because strengthening this would affect the significance of the finding. First, the authors mention new genes without introducing their relationship to the signaling pathway tested. Second, the textual logics should be strengthened. Finally and most importantly, when the difference between the phenotypic severity is so strong (vab-1 and rpm-1) then I think it's impossible to say anything from the double mutant.

We appreciate the reviewer noting that they appreciate the value and importance of the C. elegans model. The goals of our C. elegans experiments were twofold:

1. To evaluate genetic interactions between the VAB-1 Eph receptor and known RPM-1 binding proteins. This was not clearly explained in the original manuscript nor was the published precedent for these types of genetic enhancer experiments provided. We have now rectified this by substantially revising the text of the Results C. elegans section starting on line 431 and by adding several citations.

2. Our C. elegans genetics confirmed that the VAB-1 Eph receptor is not inhibited/degraded by the RPM-1/MYCBP2 ubiquitin ligase complex. We have now revised the text to draw this point out more clearly.

To further address the reviewer’s concerns, we have added a new schematic (Fig. 8A) to show the relationship between the RPM-1 and the RPM-1 binding proteins (FSN-1/FBXO45 and GLO-4/SERGEF) we are testing. We chose FSN-1 because it is part of the RPM-1 ubiquitin ligase complex and we chose GLO-4 because it functions outside the context of RPM-1 ubiquitin ligase signaling via the GLO-1 Rab GTPase to influence late endosomal/lysosomal biogenesis.

Regarding the reviewer’s concern that different penetrance/frequency of defects between rpm-1 mutants and vab-1 mutants means outcomes with vab-1; rpm-1 double mutants cannot be interpreted. We respectfully disagree. An extensive number of published studies have demonstrated that RPM-1 binding proteins have milder phenotypes than rpm-1 mutants and display genetic enhancer effects as double mutants with one another (PMID:17698012, PMID: 22357847, PMID: 25010424, PMID: 24810406). We now make this point much more clearly. While the frequency of axon termination defects in rpm-1 mutants is high it is not completely saturated as the defect is not 100%. Moreover, a major point of the vab-1; rpm-1 double mutants is that they do not have a significant reduction in phenotypic penetrance/frequency. Thus, our system is fully capable of resolving genetic suppression, which did not occur. We now make this point much more carefully and clearly.

To further address the reviewer’s concern, we have softened language about the VAB-1/Eph receptor functioning in the same pathway as RPM-1 throughout the manuscript. While we think this is still the case, because the frequency of axon termination defects is not fully saturated in rpm-1 mutants and defects could potentially become more severe (i.e. the hook might occur closer to the head of the animal rather than in the midbody). Nonetheless, this is not a critical point and we think it is more important to be clear about the two major goals and objectives of our C. elegans experiments. We hope the reviewer agrees that our rationale, logic and conclusions are more clearly and accurately drawn in the revised paper.