tRNA Fusion to Streamline RNA Structure Determination: Case Studies in Probing Aminoacyl-tRNA Sensing Mechanisms by the T-Box Riboswitch

Round 1

Reviewer 1 Report

Manuscript Number: crystals_1691964

Title: tRNA Fusion to Streamline RNA Structure Determination: Case Studies in Probing Aminoacyl-tRNA Sensing Mechanisms by the T-box Riboswitch

The manuscript describes how to use tRNA scaffolds for crystallographic studies of arbitrary RNA motifs that may be difficult to crystallize on their own. The authors present in detail a case study of the T-box antiterminator domain fused to the anticodon arm of tRNA(Gly). The manuscript is well written, and begins with a comprehensive review of the topic that was a pleasure to read. I do not have major comments to the manuscript except two issues that I believe should be at least briefly discussed.  

Authors present structures solved for 3 out of 25 different constructs, majority of which either did not crystallize or did not yield good quality crystals. Judging from the lowest  resolution structure presented (6Å), the rejected crystals must have been of relatively poor quality indeed. On the other hand, the authors wrote that tRNA(Gly) structure remained consistent across designs with different insertions. I wonder, if the template structure was not affected, why did the crystals not grow? Was the consistency of secondary structure validated e.g. using RNA 2d structure prediction software?

I’m also a bit confused with the data collection statistics presented in Table 1. In both refined models #4 and #19 CC1/2 value and I/sigma in the outer resolution shell are relatively high and it seems that the data could be processed to higher resolution. Why were the resolution limits selected? Are there any objective criteria the authors used? I think that this issue should be (very briefly) discussed in the manuscript. I also could not find PDB validation reports for the two deposited models.

Author Response

Reviewer #1

The manuscript describes how to use tRNA scaffolds for crystallographic studies of arbitrary RNA motifs that may be difficult to crystallize on their own. The authors present in detail a case study of the T-box antiterminator domain fused to the anticodon arm of tRNA(Gly). The manuscript is well written, and begins with a comprehensive review of the topic that was a pleasure to read. I do not have major comments to the manuscript except two issues that I believe should be at least briefly discussed. 

We appreciate the positive response from reviewer #1

Authors present structures solved for 3 out of 25 different constructs, majority of which either did not crystallize or did not yield good quality crystals. Judging from the lowest  resolution structure presented (6Å), the rejected crystals must have been of relatively poor quality indeed. On the other hand, the authors wrote that tRNA(Gly) structure remained consistent across designs with different insertions. I wonder, if the template structure was not affected, why did the crystals not grow? Was the consistency of secondary structure validated e.g. using RNA 2d structure prediction software?

It is true that only 2/25 led to high-resolution diffracting crystals. However, this was carried out in batches and crystals were not extensively optimized in growth or in post-crystallization methods. To get 2 high-resolution diffracting crystal forms of a construct that had never crystallized on its own was deemed a great success. This level of iterative construct design is typical for a new RNA target. To that point, the following has been added to Pg 7, Lines 270-273:

“This level of iterative engineering is common during any typical RNA structure determination effort. The fact that we obtained two different, well diffracting crystal forms for a challenging target with little condition optimization or post crystallization manipulation was encouraging. Upon recognizing…”

The template structures were all validated for predicted folding using the structure prediction software, MFold. This is described in the Methods and an additional statement was added to clarify on Pg 2, Line 84-85: “…MFold webserver, such that intended secondary structures were predicted to be maintained [33].”

I’m also a bit confused with the data collection statistics presented in Table 1. In both refined models #4 and #19 CC1/2 value and I/sigma in the outer resolution shell are relatively high and it seems that the data could be processed to higher resolution. Why were the resolution limits selected? Are there any objective criteria the authors used? I think that this issue should be (very briefly) discussed in the manuscript. I also could not find PDB validation reports for the two deposited models.

The is a good observation. We were conservative when deciding on our resolution cut-offs. A statement to explain that CC1/2 and I/sigma were both used to decide on final cut-offs was added to the methods at Pg 3, Lines 117-120: “The relatively conservative resolution cut-offs were determined by evaluating CC1/2 and I/I values. The resolution cut-offs for both high-resolution structures were chosen just be-low the point that CC1/2 values rapidly declined while I/I was greater than 1.5.”

The validation reports were not included. We did not see this as a requirement for initial submission, but they can be made available. To that point, submission IDs have been updated in the text.

Reviewer 2 Report

This nice paper provides very detailed information on how tRNA fusion can simplify the crystallization and structure determination of target RNA molecules. All information is very useful to researchers in related fields. In addition, the paper is well-structured and the language is readable. Research methods are well described and results are clearly displayed.

They begin by describing common methods and challenges for RNA crystal growth. Then, T-box Riboswitch was taken as an example to demonstrate the fusion mode of target RNA and tRNA with a clear graph. Finally, the crystal structures of several different RNA constructs were obtained and low-resolution interactions between tRNA receptor and antiterminator were observed.

This paper successfully demonstrates that tRNA fusion is a powerful RNA structure determination method.

Author Response

Reviewer #2

This nice paper provides very detailed information on how tRNA fusion can simplify the crystallization and structure determination of target RNA molecules. All information is very useful to researchers in related fields. In addition, the paper is well-structured and the language is readable. Research methods are well described and results are clearly displayed.

They begin by describing common methods and challenges for RNA crystal growth. Then, T-box Riboswitch was taken as an example to demonstrate the fusion mode of target RNA and tRNA with a clear graph. Finally, the crystal structures of several different RNA constructs were obtained and low-resolution interactions between tRNA receptor and antiterminator were observed.

This paper successfully demonstrates that tRNA fusion is a powerful RNA structure determination method.

We appreciate the positive response from reviewer #2!