Effects of the RNA-Polymerase Inhibitors Remdesivir and Favipiravir on the Structure of Lipid Bilayers—An MD Study

Round 1

Reviewer 1 Report

Manuscript ID: membranes-1915857

Title: Effects of the RNA-polymerase inhibitors Remdesivir and Favipiravir on the structure of lipid bilayers - A MD study

Authors: Mauro Bringas, Meike Luck, Peter Müller, Holger A. Scheidt, Santiago Di Lella

# Comments

The article attempts to understand how RNA-polymerase inhibitors Remdesivir and Favipiravir are distributed in the bilayer structure of cell membranes. It is motivated by previous experimental studies by the same group indicating that both compounds can be incorporated in the bilayer, however, Remdesivir disturbs the bilayer structure while Favipiravir does not.

This is a well-designed and thorough study, analyzing the mechanism by which the molecules accumulate in the membrane considering two different membrane compositions (POPC and POPC+Cholesterol), two different molecule concentrations, and also considering the molecules to start from the aqueous or lipidic phase. The authors conclude that both molecules spontaneously penetrate the membrane, and show a tendency to accumulate in the region close to the phospholipid heads, which is in accordance with the experiments. The authors also examine aspects of the interactions between the drugs and membrane, and the effect of drug insertion on the membrane structure.

This is an interesting study, and I recommend it to be published after some considerations that I list below.

1) Those molecules are actually prodrugs that are metabolized to the active form, as acknowledged by the authors on Page 2, line 66. However, all the study was done with the original molecules. It would be important to provide some justification for the use of the prodrug instead of the active molecule. What is the biological significance of the results for the drugs?

2) Page 1, line 41: This last sentence, which ends on page 2 (line 45) is way too long and confusing. I recommend separating it into multiple smaller sentences with a well-defined message in each.

3) In Page 2, line 62, there is a mention of "Figure 1A". However, Figure1 does not have "A" or "B" areas defined. Please either remove the reference or add the labels to the Figure.

4) Figure 1: Add the "A" and "B" labels to the figure if needed. I recommend writing the structure names under the molecules.

 

## Computational Methods

5) On page 3, line 86, the authors indicate using 128 lipid molecules per membrane. However, it is unclear how these were distributed into lower/upper leaflets. Please describe in more detail the final membrane compositions.

6) Also, identify what was the actual number of each molecule in each simulation.

7) On Page 3, line 96, the authors mention an "equilibration" phase. Although the term is indeed commonly used, it is inaccurate and does not really describe the process. It would be better to change the phase to a "relaxation" phase.

8) Page 3, line 94: How long was the heating stage of the simulation?

9) Page 3, lines 101-102: The authors mention using bond, dihedral, and angle constants from the Amber force field, but do not specify which version was used. Was it from GAFF, GAFF2, or some other force field? Please specify.

10) Page 3, line 100: "631-G*" should be "6-31G*".

 

## Results

11) Page 5, Figure 2. 

  (a) What is the meaning of the different colored lines in the plots?

  (b) The plots seem to show 1,200 ns dynamics, but the description of the methods mentions 3 independent 400ns production runs for each system. If this is correct, the plots should only show 400ns simulations, not 1200ns. The three independent simulations do NOT sum up to a large simulation. It would be better to show the plot for 1 "typical" simulation, and present the other two in the SI. 

  c) There seems to be a certain degree of convergence in those plots. It would be interesting to have the converged values in a table somewhere.

12) Page 6, Figure 3. The distribution for Remdesivir (Fig 3B) looks bimodal. Is there a known reason for that or should it just be considered a "convergence" problem?

13) Page 7, Table 1, shows an interesting result of the distribution of species around the drug molecules. However, to understand the significance, it would be important to see how those numbers change during the simulation, and how they converge to the final values reported. For example, 

  (a) what is the distribution for the initial structure, and at selected points during the dynamics?

  (b) If cholesterol molecules are indeed approaching the ligands (Favipiravir in particular), how does the distribution of drug-cholesterol or cholesterol-cholesterol distances change? (It should probably be random in the beginning and evolve to a bimodal distribution toward the end of the simulation).

 

14) Page 8, line 225. The authors mention the approximation of cholesterol molecules to the drug molecules as a "possible" explanation of the effect on the membrane structure. Can this be verified, again, by monitoring the cholesterol-cholesterol distances?

 

15) page 10, lines 273-281. The authors comment on the relative orientation of the ligands relative to the membrane. There could be a more detailed comparison of the computational and experimental results here.

---

Once the authors address those issues, I believe the article should be published.

 

Author Response

Please see file attached

Author Response File: Author Response.pdf

Reviewer 2 Report

The study “Effects of the RNA-polymerase inhibitors Remdesivir and Favipiravir on the structure of lipid bilayers - A MD study”, by Bringas and Colleagues, is a very interesting application of MD simulations approach to evaluate the interaction between membrane and remdesivir/favipiravir. Authors show that these drugs can spontaneously bind to model lipid membranes, in the presence or absence of cholesterol, according to experimental results.

Despite the study has been correctly designed, being clear in its development and conclusions, I have some minors:

i)               Figure 2: The figure is not clear. Authors should modify all the panels in order to make figure 2 more readable.

ii)              Figure 4: the representation of drugs in membrane is low quality. Authors should improve the graphical quality.

iii)             I think that Authors should show the frame 0 for all the MD simulations (in supplementary) and replace the box 1 with a workflow (or a pipeline). 

Author Response

Please see file attached

Author Response File: Author Response.pdf