The Paradoxes of Viral mRNA Translation during Mammalian Orthoreovirus Infection

Round 1

Reviewer 1 Report

This is an excellent and thorough review on the topic of reovirus translation. It is exhaustive and fully current in its sourcing, and is insightful in areas that require further experimentation to support or define. There are some minor punctuation and grammatical concerns, but the overall structure is appropriate and informative. A pleasure to read!

Author Response

Thank you very much for the very favorable review. We have attended to the punctuation and grammatical concerns

Reviewer 2 Report

The authors provide an exhaustive compilation of research on reovirus mRNA translation to date. Despite numerous studies in this field, the complete process underlying reovirus mRNA translation remains unclear.  This review condenses all these decades of research into a single manuscript in a well-structured narrative.  It also discusses the different hypotheses about this field followed by exciting questions waiting to be unveiled.  The figures are clear and figure 1 summarizes all the information quite well with nice visuals to help the reader.

I have just a few comments, questions, and suggestions provided bellow:

Table 1 For M1 genome segment, the corresponding mRNA should be m2.

Line 67-70 The four early gene segments (L1, M3, S3 and S4) produce four mRNAs that are among the viral mRNA with the longest 5’-UTR (24, 18, 27 and 32 nt long, respectively). According to Nonoyama et al., the next gene segment expressed (after these four) is M2, which produces an mRNA with a 5’-UTR 29 nt in length. It appears that mRNAs with longer 5’-UTRs mRNAs tend to be translated earlier than those with shorter 5'-UTRs. Maybe, these mRNAs with long 5’-UTR could be translated more easily by the cellular translation machinery than those with short 5’-UTR (due to mechanistic issues, as you explain in line 110-118). So, could the length of the 5’-UTR be another factor influencing translation priority?

Line 81-82 ‘VFs are non-membrane bound compartments’. As you explain at lines 313-325, ER membranes are intimately related to reovirus VF. Then, this sentence could lead to misunderstandings. I suggest starting this sentence with: VFs were thought to be non-membrane…’ or just changing it to ‘VFs are membrane-bound compartments’.

Line 108 s4 5’-UTR is 32nt long. Therefore, it conforms to the canonical ribosomal ‘footprint’, and allows the s4 mRNA standard translation in cells. Once σ3 protein is produced, could it act as a translation factor and assist in the translation of the other viral mRNAs?

Line 485: Is there any reason why reovirus early mRNAs are capped and late mRNAs are uncapped (at least in some cell lines)? Is λ2 not capping late mRNAs correctly? Authors may answer this question or add it to the final questions.

Line 538 ‘Are viral mRNAs translated within RNP granules? ‘: I would suggest: ‘Are early viral mRNAs translated within RNP granules?’  (since late ones are translated in the VFs).

Figure 1. σ3 binding to dsRNA is crucial to produce shutdown (line 396-398 and 402-404). It might be interesting to show this in figure 1. Perhaps it could be depicted by adding the viral mRNA icon next to the σ3 protein icon. I also propose to label the viral mRNA icon (see attached image).

Figure 1: σNS recruits translational machinery of SG disassembly and bring it to the VF margin (line 522-525). In this figure, this is represented by a text followed by a small arrow pointing to σ3. It might be a bit confusing. I would suggest reversing the order of the σNS hexagon with the σ3 oval in this area, or simply replacing the small arrow with a plus sign (which does not point to any particular protein). 

Comments for author File: Comments.pdf

Author Response

We thank the reviewer for the complementary comments regarding our manuscript. We have responded to each of the comments and questions as detailed below:

Table 1 For M1 genome segment, the corresponding mRNA should be m2.

Yes, thank you for noticing our error! Corrected.

Line 67-70 The four early gene segments (L1, M3, S3 and S4) produce four mRNAs that are among the viral mRNA with the longest 5’-UTR (24, 18, 27 and 32 nt long, respectively). According to Nonoyama et al., the next gene segment expressed (after these four) is M2, which produces an mRNA with a 5’-UTR 29 nt in length. It appears that mRNAs with longer 5’-UTRs mRNAs tend to be translated earlier than those with shorter 5'-UTRs. Maybe, these mRNAs with long 5’-UTR could be translated more easily by the cellular translation machinery than those with short 5’-UTR (due to mechanistic issues, as you explain in line 110-118). So, could the length of the 5’-UTR be another factor influencing translation priority?

Yes, this is definitely a possibility that might explain differential translational efficiency (or earlier translational detection) from different transcripts. However, it does not explain the lack (or below detectable concentrations) of six viral transcripts early in infection. The study by Nonoyama did not detect the other viral transcripts until later in infection - suggesting either a trancriptional preference of the RNA-dependent RNA polymerase for the early transcripts or as described, a difference in transcript stability. The length of the 5' UTR can have an effect on translational efficiency, but would not explain the difference in transcript levels.

Line 81-82 ‘VFs are non-membrane bound compartments’. As you explain at lines 313-325, ER membranes are intimately related to reovirus VF. Then, this sentence could lead to misunderstandings. I suggest starting this sentence with: VFs were thought to be non-membrane…’ or just changing it to ‘VFs are membrane-bound compartments’.

Thank you. We have changed this sentence to "VFs are compartments within infected cells...". Because VFs are Liquid-liquid phase separated structures (unpublished findings), our best interpretation for what is happening is that membranes are embedded within VFs and can actively pass through VFs (at least early in infection). However, the VFs are not themselves membrane bound, as membranes appear to move through the VFs. Vesicles can be seen to move through VFs (see supplemental movies in Desmet et al mBio, 2014).

Line 108 s4 5’-UTR is 32nt long. Therefore, it conforms to the canonical ribosomal ‘footprint’, and allows the s4 mRNA standard translation in cells. Once σ3 protein is produced, could it act as a translation factor and assist in the translation of the other viral mRNAs?

Yes, σ3 overexpression has been shown to enhance translation of other mRNAs by Giantini & Shatkin, 1989. The mechanism of enhancement is hypothesized to be due to the capacity of σ3 to prevent PKR activation and thus increase the free concentration of ternary complex. We have added two sentences (lines ) to indicate this.  

Line 485: Is there any reason why reovirus early mRNAs are capped and late mRNAs are uncapped (at least in some cell lines)? Is λ2 not capping late mRNAs correctly? Authors may answer this question or add it to the final questions.

The hypothesized reason is that the secondary core particles that assemble de novo in infected cells are responsible for synthesizing the uncapped mRNAs (see Skup et al., 1980). The idea being that the capping enzymes within the λ2 protein in the de novo assembled core particles are 'masked' during extrusion of the viral transcripts leading to the synthesis of uncapped viral mRNAs. These studies were done on secondary core particles that were purified from infected cells and chymotrypsin treatment of such particles allowed them to synthesize capped mRNAs. In light, of later work, these findings suggest the particles isolated were core particles that were partially coated with outer-capsid proteins. Later studies, however, failed to find evidence of uncapped mRNAs in a different cell line. The question whether uncapped mRNAs are synthesized during infection or if their synthesis is cell-type specific remains open. We have added two sentences (lines ) to answer the reviewers question.

Line 538 ‘Are viral mRNAs translated within RNP granules? ‘: I would suggest: ‘Are early viral mRNAs translated within RNP granules?’  (since late ones are translated in the VFs).

Thank you. We have changed the text as you suggest.

Figure 1. σ3 binding to dsRNA is crucial to produce shutdown (line 396-398 and 402-404). It might be interesting to show this in figure 1. Perhaps it could be depicted by adding the viral mRNA icon next to the σ3 protein icon. I also propose to label the viral mRNA icon (see attached image).

Thank you for the suggestions! We have added a label for the viral mRNA. We have decided to not follow your advice regarding σ3. This is because, we have unpublished data that shows that the capacity of σ3 to inhibit PKR can be dissociated from its capacity to bind dsRNA. We have prepared viral mutants that express a mutant form of σ3 that is unable to bind dsRNA, but is still able to suppress activation of PKR. The role of dsRNA-binding by σ3 appears to be to suppress a different type of RNP granule called RNase L-dependent bodies (See Burke, James M., Evan T. Lester, Devin Tauber, and Roy Parker. “RNase L Promotes the Formation of Unique Ribonucleoprotein Granules Distinct from Stress Granules.” Journal of Biological Chemistry 295, no. 6 (February 7, 2020): 1426–38.). Thus, σ3 suppresses PKR independently of dsRNA-binding. We would prefer not to discuss these new findings in this review at the moment, as our other comprehensive paper discussing these findings will be submitted for publication shortly. 

Figure 1: σNS recruits translational machinery of SG disassembly and bring it to the VF margin (line 522-525). In this figure, this is represented by a text followed by a small arrow pointing to σ3. It might be a bit confusing. I would suggest reversing the order of the σNS hexagon with the σ3 oval in this area, or simply replacing the small arrow with a plus sign (which does not point to any particular protein). 

Thank you for pointing this out. We have removed the arrow and replaced it with a plus sign, as suggested.