Molecular Forces Governing the Biological Function of Per-Arnt-Sim-B (PAS-B) Domains: A Comparative Computational Study

Round 1

Reviewer 1 Report

In this manuscript, a number of molecular dynamics simulations are performed on several PAS domains. Concerning PAS-B, the authors also perform simulations of truncation mutants and show that this leads to differences in the structure and dynamics. I agree with the authors that these simulations can be informative for stimulating additional experiments. However, I have a few concerns, which I have detailed below:

1. The authors should report what’s the extent of the motion of loop1, for example by drawing several conformations of the loop, while the rest is kept fixed.
2. The connection between the authors findings and the importance of loop1 for dimerization confuses me. What exactly the authors want to convey at lines 259-265? Is this just speculation? Even if the authors have shown that the PAS-B truncation mutations bear differences in several calculated quantities, how these differences are connected to hsp90 and dimerization?
3. As evidenced by the sheer number of conditional verbs in the Discussion section, many connections between simulations and biological relevance are just speculations. Do the authors have additional data to prove their statements involving the PAS-B ligand, HSP90 interactions etc…?
4. Lines 297-299: this is one example of unclear writing that makes the results of this work and its implications hard to understand.

Author Response

On behalf of my coworkers and myself, I would like to thank the Reviewer 1 for assessing our manuscript and for providing some very helpful comments. All the corrections suggested were applied through the revised manuscript. Responses and clarifications to further specific comments are given below.

1. The authors should report what is the extent of the motion of loop1, for example by drawing several conformations of the loop, while the rest is kept fixed.

To highlight the extend of the motion, a new panel was added to Figure 5, showing a gradient scale for conformations representing the transition along the largest principal component.

1. The connection between the authors findings and the importance of loop1 for dimerization confuses me. What exactly the authors want to convey at lines 259-265? Is this just speculation? Even if the authors have shown that the PAS-B truncation mutations bear differences in several calculated quantities, how these differences are connected to hsp90 and dimerization?

The effects of dimerisation arise from the relationship between the dynamics of  loop1 (which modulates the ligand binding) and the interface where the termini are located. Our simulations showed that the motions on loop1 covariate with the motions located in the dimerisation interface. This highlights how a change in intrinsic dynamics caused by a ligand binding modulates  the dimerisation event. This was clarified in the manuscript main text in lines 264-265 and 270-273.

1. As evidenced by the sheer number of conditional verbs in the Discussion section, many connections between simulations and biological relevance are just speculations. Do the authors have additional data to prove their statements involving the PAS-B ligand, HSP90 interactions etc.…?

The Hsp90-AhR dimerisation discussion is based on the recently published work by Anatoly Soshilov and coworkers (Int. J. Mol. Sci. 2020, 21, 2474, doi:10.3390/ijms21072474), and previously published data. The structural characteristics for the AhR PAS-B that drives the dimerisation have been discussed in several studies (e.g. Corrada et al., PLoS Comput Biol. 2016 Jun 13;12(6):e1004981; Sakurai et al., J Biol Chem. 2017 Oct 27;292(43):17609-17616, and references therein) . To our knowledge, there is no experimental structural evidence that concludes how the assembly is formed. Computational studies have been carried out, and several concluded that the interaction interface is in the opposite region to the loop1. Hence, we aimed to show that any modulation in the loop1 region should be covariant to the dimerisation interface discussed in these previous studies.

Lines 297-299: this is one example of unclear writing that makes the results of this work and its implications hard to understand.

These lines were edited, and several clarifications were added.

I hope that the response thoroughly addresses all concerns regarding the results and their discussion, data analysis and presenting of the data, and I am looking forward to the feedback.

Yours Sincerely,

Dr. Joao Victor de Souza

Reviewer 2 Report

In this manuscript, the authors study the mechanisms underlying the function of the Per-Arnt-Sim (PAS) domains. The authors hypothesize that the entropy of the disordered terminal region of the domain impacts the function of the domain. To test this hypothesis, they perform molecular dynamics simulations and compare the dynamics of terminal truncation mutants with the standard length PAS domain of AhR (Aryl hydrocarbon Receptor) PAS-B domain. They also compare the dynamics of several structurally similar but sequentially distinct PAS domains. Finally, they explore the role of disulfide bridges in increasing the stability of the binding region of the domain. The results from this study provide important evidence for the role of the terminal region in modulating the function of the domain. There are few suggestions and  minor edits but other than that I do not have any reservations against recommending the manuscript for publication.

Suggestions and minor edits

• Given the low sequence similarity between the different PAS domains, it would be valuable to include some quality statistics for the homology models that they built, in the Materials and Methods section.
• Scales should be included in the heatmap in Figure 1
• Given that this study was performed with only the PAS domain and not with the whole protein, it will be valuable to include what they authors think about this study in the context of the whole protein, in the Discussions section. 
• Missing citation on line 93
• 4C instead of 4B in line 195
• “driving” instead of “driven” in line 199
• Missing citations in lines 243, 246 and 249
• “N” instead of “n” in line 285

Author Response

On behalf of my coworkers and myself, I would like to sincerely thank the Reviewer 2 for assessing our manuscript and for providing some very helpful comments. All the corrections suggested were applied through the revised manuscript. Responses and clarifications to further specific comments are given below.

• Given the low sequence similarity between the different PAS domains, it would be valuable to include some quality statistics for the homology models that they built, in the Materials and Methods section.

The Phyre2 statistics (confidence and sequence identity) were added in the methodology  (Lines 110 and 113).

• Scales should be included in the heatmap in Figure 1

The colour scale was added in Figure 1, as suggested.

• Given that this study was performed with only the PAS domain and not with the whole protein, it will be valuable to include what they authors think about this study in the context of the whole protein, in the Discussions section. 

A brief discussion on how it should affect the PAS-A domain was added to the manuscript (lines 283 to 286).

• Missing citation on line 93
• 4C instead of 4B in line 195
• “driving” instead of “driven” in line 199
• Missing citations in lines 243, 246 and 249
• “N” instead of “n” in line 285

All the above were addressed.

I hope that the response thoroughly addresses all concerns regarding the results and their discussion, data analysis and presenting of the data, and I am looking forward to the feedback.

Yours Sincerely,

Dr. Joao Victor de Souza

Author Response File: Author Response.docx