A Comparison in the Use of the Crystallographic Structure of the Human A1 or the A2A Adenosine Receptors as a Template for the Construction of a Homology Model of the A3 Subtype
Round 1
Reviewer 1 Report
In this manuscript homology models of human A3 adenosine receptor subtype using the template A1 or A2A structures is presented and validated by docking of antagonists and agonists.
Overall the manuscript looks ok. However some parts are not well described:
1/
In the introduction is missing a short overview / description of all available pdb structures of hA1 and hA2A active/ inactive, with agonists/antagonists bound ...
2/
In material and methods, the Homology modeling section gives a very poor description about how the homology model building is performed: which template structure for which homology model? what is the method/software used to build the model? Was the model then 'validated' by some standard methods like the molprobity server? I see that you plan MD simulations later, but this should maybe be standard in the validation proces.
Author Response
1/
In the introduction is missing a short overview / description of all available pdb structures of hA1 and hA2A active/ inactive, with agonists/antagonists bound..
We thank the Reviewer for this suggestion. A short review of available crystal structures is now included in the main text (lines 63-65, 73-77):
“[…] Nowadays, just 46 crystal structures are already available, bound to a wide variety of agonists and antagonists, either natural or synthetic, representing the majority of structural determinants on adenosine receptors ligands’ activity.”
“[…] These structures were co-crystallized with the endogenous agonist adenosine (PDB code: 6D9H), as with the antagonist 4-{[3-(8-cyclohexyl-2,6-dioxo-1-propyl-1,2,6,7-tetrahydro-3H-purin-3-yl)propyl]carbamoyl}benzene-1-sulfonyl fluoride (PDB code: 5UEN), which, actually, constitute the most representative examples of interaction scheme for the system of interest.
2/
In material and methods, the Homology modeling section gives a very poor description about how the homology model building is performed: which template structure for which homology model? what is the method/software used to build the model? Was the model then 'validated' by some standard methods like the molprobity server? I see that you plan MD simulations later, but this should maybe be standard in the validation process.
We also thank the Reviewer for this comment. The PDB structures used as templates are now reported (lines 159-162):
“For inactive state model of the hA3 AR, the following template pdb structures were adopted and then subjected to comparison (see Section 3): 3PWH, 5NM2 (A2a) and 5UEN (A1). For the active state model 2YDO (A2a) and 6D9H (A1) pdb templates were used.”
1. As reported in lines 142-143, we used the MOE Homology Model Tool to build our models. Homology tool restitutes energy terms and scores related to the wellness of the generated output, as now specified in line 154. Furthermore, we used also the Swiss Model / Structure Assessment web server for further general validation, although the main purpose of our study was to use a docking-driven approach for assessment.
2. We decided to not include MD simulations at this stage, referring to the most common virtual screening procedures, which mainly use docking simulations, as reported in reference 50. Indeed, docking results can be more easily related to the sequence/structural comparison proposed and the templates that have been used as rigid reference structures. Surely, MD simulations will give us a deeper understanding of time-dependent processes, as we are already observing in our research team, but we did not consider them necessary at this validation stage.
Reviewer 2 Report
The authors compared the propensity of both models to accommodate a series of known potent and selective human A3 agonists and antagonists. They concluded that it is possible to affirm that the human A3 receptor model based on the crystallographic structure of the A1 subtype can represent a valid alternative to the one conventionally used today, based on the available A2A structures.
This is a very interesting and important work.
Comments:
1:Fig.1 and 2; what does "%" (in some Ki value) mean?
2:The Table1 is not easy to undarstand. Which is the AAs sequence of A3, upper one or lower one? Their residue numbers also should be indicated. I suggest the authors make a reader-friendly figure of multiple sequence alignment by using some software (ex. Clustal https://en.wikipedia.org/wiki/Clustal )
3:Line 284-285 : No orange color is shown in ADO and MRS5127 cells of Fig.6. Is the sentence correct?
Author Response
1:Fig.1 and 2; what does "%" (in some Ki value) mean?
We thank the Reviewer for this note. These values generally represent the ligand effect on the adenylate cyclase activity. Unfortunately, by literature, some instability constants are not reported, as a consequence, whenever lacking we have annotated only the reported enzyme activity assay percentages. For clarity, I have now specified them in both the figures.
2:The Table1 is not easy to undarstand. Which is the AAs sequence of A3, upper one or lower one? Their residue numbers also should be indicated. I suggest the authors make a reader-friendly figure of multiple sequence alignment by using some software (ex. Clustal https://en.wikipedia.org/wiki/Clustal )
We also thank the Reviewer for this comment.
1. As reported in the description (A3/A1), the sequence of A3 (second row) is aligned against the reference domain sequences of A1 and A2a (first row). Now, I have just reported the numbers as indicated and specified which row A3 sequence belongs to.
2. Unfortunately, multiple sequence alignment is not very suitable for this case: we risk obtaining less accurate alignments than single ones because multiple sequence alignment strongly mathematically depends on the maximum sequence length, i.e. gaps are inserted in each sequence in order to conform to such maximum (matrix normalization). Indeed, in our case we wanted to check similarity and identity for specific crystallographic structural domains observed for each subtype template in detail, considering each domain independent on other ones.
3:Line 284-285 : No orange color is shown in ADO and MRS5127 cells of Fig.6. Is the sentence correct?
We thank the Reviewer for this note. In fact, the sentence was considered a misprint and it is properly removed.
Reviewer 3 Report
While A3 adenosine receptor (AR) is important for therapeutic applications, its 3D structure has not been available. In such case, homologous proteins whose 3D structures are known are used as the template in homology modelling. For A3 AR, crystal structures of A1 AR and A2A AR are available. While A2A AR is widely used as the modeling template, the amino acid sequence as well as the protein function of A1 AR are more similar to those of A3 AR. Authors constructed the 3D structural model of A3 AR using the structures of A1 AR and A2A AR. Authors further performed docking simulations of agonist and antagonist, and compared the results quantitatively. Based on the results, it was concluded that the model generated by using A1 AR structure is valid. The results are not novel and quite as expected, but the quantitative evaluation is valuable in this field. The manuscript would be improved if the following points are taken into consideration.
1. The figure showing the difference between two template structures (A1A AR and A2A AR) would be informative.
2. In table 1, the residue numbers are not provided. It makes it difficult to understand the conservation of interaction between the amino acid residue and ligand among AR subtypes (for example Figure 5).
3. Figure 9 is not referred in the text.
Author Response
1. The figure showing the difference between two template structures (A1A AR and A2A AR) would be informative.
We thank the Reviewer for this suggestion. We implemented the following figure reporting superposition of A1 and A2a active states (6D9H, 2YDO). It is very easy to observe that helices and loops do not superpose well on each other.
Figure 1 Structural superposition of hA1 (dark blue) and hA2a (gold) ARs crystal structures in their active state (PDB code: 6D9H and 2YDO, respectively). Panel A and B representations differ only for camera orientation. Additional A2A disulfide bridges are highlighted in red.
2. In table 1, the residue numbers are not provided. It makes it difficult to understand the conservation of interaction between the amino acid residue and ligand among AR subtypes (for example Figure 5)
We thank the Reviewer for this note. Numbers are now properly reported.
3. Figure 9 is not referred in the text.
We thank the Reviewer also for this note. Figure 9 (10 after inclusion of the new figure) reference is now reported in line 408.
