The mechanism of translation initiation on Aichi Virus RNA mediated by a novel type of picornavirus IRES

Transcription

1 Manuscript EMBO The mechanism of translation initiation on Aichi Virus RNA mediated by a novel type of picornavirus IRES Yingpu Yu, Trevor R. Sweeney, Panagiota Kafasla, Richard J. Jackson, Tatyana V. Pestova, Christopher U.T. Hellen Corresponding author: Christopher U.T. Hellen, State University of New York Review timeline: Submission date: 01 June 2011 Editorial Decision: 24 June 2011 Revision received: 18 July 2011 Accepted: 01 August 2011 Transaction Report: (Note: With the exception of the correction of typographical or spelling errors that could be a source of ambiguity, letters and reports are not edited. The original formatting of letters and referee reports may not be reflected in this compilation.) 1st Editorial Decision 24 June 2011 Thank you for submitting your manuscript for consideration by The EMBO Journal. Three referees have now evaluated it, and you will be pleased to see that they are all very positive and would support publication here after minor revision (see below). I would thus like to invite you to prepare a revised manuscript in which you need to address the referees' points in an adequate manner. Furthermore, there are a number of editorial issues that need further attention. First, please add an author contribution section to the main body of the manuscript text. Second, prior to acceptance of every paper we perform a final check for figures containing lanes of gels that are assembled from cropped lanes. While cropping and pasting may be considered acceptable practices in some cases (please see Rossner and Yamada, JCB 166, 11-15, 2004) there needs to be a proper indication and explanation in all cases where such processing has been performed according to our editorial policies. Please note that it is our standard procedure when images appear like they have been pasted together without proper indication (like a white space or a black line between) or explanation in the figure legend to ask for the original scans. In the case of the present submission there are a number of panels that appear to not fully meet these requirements: figure 1B, 3H, 3I, 6E. I therefore like to kindly ask you to include suitably amended versions of these figures in the revised manuscript and to explain in the figure legend that all lanes come from the same gel. Please be reminded that according to our editorial policies we also need to see the original scans for the panels in question. I should add that it is EMBO Journal policy to allow only a single round of revision, and acceptance of your manuscript will therefore depend on the completeness of your responses in this revised version. European Molecular Biology Organization 1

2 When preparing your letter of response to the referees' comments, please bear in mind that this will form part of the Peer Review Process File, and will therefore be available online to the community. For more details on our Transparent Editorial Process, please visit our website: We generally allow three months as standard revision time. As a matter of policy, competing manuscripts published during this period will not negatively impact on our assessment of the conceptual advance presented by your study. However, we request that you contact the editor as soon as possible upon publication of any related work, to discuss how to proceed. Thank you for the opportunity to consider your work for publication. I look forward to your revision. Yours sincerely, Editor The EMBO Journal REFEREE REPORTS: Referee #1 (Remarks to the Author): Picornavirus mrnas use internal ribosome entry sites (IRESs) to initiate viral protein synthesis during infection, when host protein synthesis is down-regulated. This manuscript describes a new kind of IRES contained within the mrna of Aichivirus (AV), a member of the Kobuvirus genus of Picornaviridae, that differs structurally from previously discovered viral IRESs. Its function similarly involves interaction with host initiation factor eif4g, but its eif4g-interacting domain is structurally distinct. A unique feature of this IRES is that its initiation codon is sequestered within a stable hairpin structure. Therefore, translation initiation is strongly dependent on the RNA helicase protein DHX29. Overall, the study is interesting and the results support the conclusions of the work. Publication is recommended pending revisions to address the following specific comments: 1. The structural environment of the AV initiation codon is reminiscent of the silent (but functionally important) initiation codon in Type 1 picornavirus IRESs, as they are both in the Yn- Xm-AUG motif following the eif4g interacting domain and are both located within a stable hairpin. It would be helpful if the authors could compare these two IRES types in the Discussion. 2. On page 11, the requirement for DHX29 is described as "conditional". I cannot agree with this statement, as in this study DHX29 was found to be non-essential only for those AV IRES mutants that were designed to destabilize domain L. Because a stable domain L is conserved in all AV isolates, wild type viruses would appear to be always dependent on DHX For Figs. 4C and 4D, it would be useful if the authors could show the toe-printing and in vitro translation results from the same set of mutants. Does the strength of the stall at the stage of 48S initiation complex assembly correlate with translation efficiency? This information would add important mechanistic insight to the proposed initiation pathway. 4. Minor point: in Fig. 3I, XL' (stands for cyclin, as in Fig. 2B legend) in the left lanes and cyclin in the right lanes are of different sizes. Is this because different truncations were used? Referee #2 (Remarks to the Author): This manuscript describes a carefully conducted study on the function of an Aichivirus IRES, a type of picornavirus with novel structure and activity that conforms to neither of the establishedpicornavirus types. This is an extensive, carefully conducted study with very high levels of technical expertise. The findings are novel and of broad interest, as they define a new type of picornavirus IRES. I have a few comments in details which the authors may wish to adress: The authors conclude that eif4a enhances the interaction between eif4g and the IRES, based on European Molecular Biology Organization 2

3 the observation that the radical-induced cleavage of the IRES RNA by Cys-substituted eif4g is enhanced in the presence of eif4a. It could be the case that the affinity of eif4g is unaltered but that eif4a induces a conformation which cleaves the RNA more efficiently. The authors should either tone down their conclusion, or conduct assays that directly monitor affinity (eg pull-downs or similar). In the biochemical fractionation experiments, PTB is identified as a trans-acting factor and this is then confirmed using recombinant PTB. It is not clear whether the same was done for DHX29, or whether DHX29 was only used purified from RRL? The materials and methods mention expression of DHX29 in E. coli, but then it is not clear to me from the results section where this is used. In the introduction (3rd sentence), the authors state that the 43S initiation complex comprises the 40S subunit, eif2 TC, eif3, eif1 and eif1a. Should this not also include eif5? It would be nice if the authors could introduce Aichiviruses with one or two sentences - who/what does it infect, what diseases does it cause? Referee #3 (Remarks to the Author): This paper provides a very detailed characterization of the structure and factor requirements of the picornovirus Aichi virus IRES. Its factor requirements overlap that of type 1 and type 2 picornavirus IRESs in requiring eif2, eif3, the middle domain of eif4g and eif4a, but not eif4e, and in being stimulated by PTB. A distinctive requirement is the need for the helicase DHX29, which was shown to be required to destabilize the helical region of domain L of the IRES, which harbors the AUG codon in base-paired form. Consistently, the presence of the AUG in a duplex is a unique feature of the AV IRES. Similar to the EMCV IRES, eif4gm binds to the apical region of domain K, dependent on the conserved motif AGGU, which is also crucial for AV IRES function, and in a manner stimulated by eif4a. Although the AV IRES contains a cruciform domain J, as found in other picornavirus IRESs, the tetraloop at its apex is dispensable for AV IRES function, indicating the presence of other, unknown sequences in domain J important for IRES function. Association of eif4gm with the IRES seems to induce conformational changes near the AUG codon, which are proposed to be involved in the ability of eif4gm to induce 43S complex binding to the correct initiation codon. These changes are enhanced by PTB, suggesting that PTB might stimulate AV IRES function by promoting proper interactions of eif4gm with the IRES, similar to previous suggestions about the function of PTB in modulating eif4g-ires interactions for poliovirus. The data presented here are of the highest quality and can be regarded as providing definitive information regarding the requirements for eif4gm, eif4a, PTB, and DHX29 in AV IRES function in rabbit reticulocyte lysates, and of the binding sites for eif4gm, eif4a and PTB within the AV IRES structure. Although the data do not provide any especially novel insights into how eif4g and PTB stimulate IRES function, the results support previous suggestions for other picornavirus IRESs that they remodel the region surrounding the AUG codon to facilitate scanning-independent 43S attachment at the start codon. This extends previous conclusions about a common mechanism for different picornavirus IRESs despite an overall lack of primary sequence similarity among them. The most novel finding is that the AV IRES is dependent on DHX29 for 43S binding at the IRES AUG, owing to sequestration of the AUG in a base-paired segment of domain L. It would greatly enhance the significance of the study if a requirement for DHX29 for AV IRES function and/or viral replication could be demonstrated in cultured cells, but I don't have the expertise to say whether or not such experiments are feasible. Minor comments: 1. some explanation is required for the fact that nearly half of the RNAse T1 cleavage sites involve G's involved in base-pairs. 2. As the assays in Fig. 2E-F involve monocistronic constructs, it seems important to show that the mutant RNAs that appear to be defective for IRES activity are of comparable stability to the wildtype RNA or, alternatively, verify the results with dicistronic constructs, at least in the case of the AGGUtoUCCA and the del mutants. 3. It should be stipulated which toe-printing experiments in Fig. 3, and also later in the study, employ recombinant DHX29. European Molecular Biology Organization 3

4 1st Revision - authors' response 18 July 2011 Editor: To comply with your recommendations concerning editorial issues, I have added an author contribution section and also revised Figures 1B, 1C, 1D, 3G, 3I, 7A and 7C to indicate those lanes that were not contiguous in the original gels. The relevant Figure legends have been revised accordingly to indicate that all lanes in one panel were derived from the same original gel. The scans of the original gels for each revised panel are included in the Attachment, and have been annotated to indicate which lanes were used in the final version. The original gels were obtained from experiments that included additional conditions (i.e. additional combinations of factors, or, as in Fig. 3G, titration of factor concentrations) that we excluded from the final version because they did not add further useful information. Your letter suggested that Figures 3H and 6E were also composites, but this is not the case, and they have therefore not been revised. Reviewer #1 1. The structural environment of the AV initiation codon is reminiscent of the silent (but functionally important) initiation codon in Type 1 picornavirus IRESs, as they are both in the Yn- Xm-AUG motif following the eif4g interacting domain and are both located within a stable hairpin. It would be helpful if the authors could compare these two IRES types in the Discussion. This is an excellent point: the AUG triplet of the Yn-Xm-AUG motif in Type 1 IRESs is also sequestered within a hairpin (domain VI), but it is either silent or used very inefficiently, at least in part due to its sub-optimal nucleotide context, and substitutions that optimize the context enhance utilization of this codon significantly (Pestova et al., 1994; Kaminski et al., However, initiation has not yet been reconstituted in vitro on this or the authentic initiation codon of any Type 1 IRES, and we therefore do not yet know whether DHX29 is required for ribosomal recognition of this AUG or for initiation on the authentic start codon. Since there is thus no mechanistic data to discuss yet, and because the manuscript is already at the upper limit of length, we reluctantly decided against mentioning the possible involvement of DHX29 in initiation on Type 1 IRESs in the Discussion. Kaminski A, Pöyry TA, Skene PJ, Jackson RJ. (2010). Mechanism of initiation site selection promoted by the human rhinovirus 2 internal ribosome entry site. J Virol. 84: Pestova TV, Hellen CU, Wimmer E. (1994). A conserved AUG triplet in the 5' nontranslated region of poliovirus can function as an initiation codon in vitro and in vivo. Virology 204: On page 11, the requirement for DHX29 is described as "conditional". I cannot agree with this statement, as in this study DHX29 was found to be non-essential only for those AV IRES mutants that were designed to destabilize domain L. Because a stable domain L is conserved in all AV isolates, wild type viruses would appear to be always dependent on DHX29. In response to this comment, we have modified the title (which now reads "The requirement for DHX29 results from sequestration of the initiation codon in a stable hairpin") and the last sentence of this paragraph. 3. For Figs. 4C and 4D, it would be useful if the authors could show the toe-printing and in vitro translation results from the same set of mutants. Does the strength of the stall at the stage of 48S initiation complex assembly correlate with translation efficiency? This information would add important mechanistic insight to the proposed initiation pathway. We have included additional data to Fig. 4D to show toe-printing on the Aichvirus (mut ) mutant: consistent with the results for other mutants in which IRES domain L had been destabilized, 48S complex formation was no longer dependent on DHX29, and was more efficient than on the wild-type RNA in which this domain is intact. Unfortunately, we were unable to obtain meaningful translation data for the (Δ ) mutant mrna (for which toe-printing data are shown in Fig. 4D, lanes 4, 5) because the deletion of 31nt from the 5 -terminal region of the ORF almost immediately put it out-of-frame with the initiation codon, so that the resulting translation product is European Molecular Biology Organization 4

5 severely truncated and has a much lower methionine content than the translation product from the wild-type and all other mutant mrnas. Our experience indicates that quantitatively, results obtained from in vitro translation experiments in RRL and from experiments involving formation of 48S complexes in the in vitro reconstituted system are not directly comparable, because RRL contains RNAs that can compete with mrna for initiation factors, RNA-binding proteins that can compete with initiation factors for binding to mrna and different (undetermined) concentrations of initiation factors. However, the trend of the effects of mutations in the two systems is invariably the same, as in the present case. 4. Minor point: in Fig. 3I, XL' (stands for cyclin, as in Fig. 2B legend) in the left lanes and cyclin in the right lanes are of different sizes. Is this because different truncations were used? The cyclin B2 ORF in the dicistronic Aichi virus IRES-containing mrna (now designated as ΔXL) has a truncation that removes seventeen amino acids from the C-terminus. This is now clarified on page 7 (lines 16-19), on page 23 (lines 17-22), and in the Fig. 2A legend. Reviewer #2 1. The authors conclude that eif4a enhances the interaction between eif4g and the IRES, based on the observation that the radical-induced cleavage of the IRES RNA by Cys-substituted eif4g is enhanced in the presence of eif4a. It could be the case that the affinity of eif4g is unaltered but that eif4a induces a conformation, which cleaves the RNA more efficiently. The authors should either tone down their conclusion, or conduct assays that directly monitor affinity (eg pull-downs or similar). We agree that our data do not directly address the influence of eif4a on the affinity of eif4g for the IRES, and have accordingly toned down this conclusion on p12, lines and on p20, lines In the biochemical fractionation experiments, PTB is identified as a trans-acting factor and this is then confirmed using recombinant PTB. It is not clear whether the same was done for DHX29, or whether DHX29 was only used purified from RRL? The materials and methods mention expression of DHX29 in E. coli, but then it is not clear to me from the results section where this is used. We have modified the text in the Methods and the Results sections (p10, lines 3-5; p26, line 5) to indicate clearly that recombinant DHX29 was used in all reconstitution experiments once the requirement for this factor had been determined in the initial fractionation experiments (Fig. 3C). 3. In the introduction (3rd sentence), the authors state that the 43S initiation complex comprises the 40S subunit, eif2 TC, eif3, eif1 and eif1a. Should this not also include eif5? The text has been amended to indicate that eif5 is a component of the 43S complex. 4. It would be nice if the authors could introduce Aichiviruses with one or two sentences - who/what does it infect, what diseases does it cause? We have now introduced Aichi virus at the end of the Introduction section (p6, lines 11-12). Reviewer #3 1. It would greatly enhance the significance of the study if a requirement for DHX29 for AV IRES function and/or viral replication could be demonstrated in cultured cells, but I don't have the expertise to say whether or not such experiments are feasible. Our laboratory is equipped for the biochemical study of proteins and nucleic acids and their complexes, but we unfortunately do not have the resources or expertise for the suggested experiments. European Molecular Biology Organization 5

6 2. Some explanation is required for the fact that nearly half of the RNAse T1 cleavage sites involve G's involved in base-pairs. There are 12 sites of observed RNAse T1 cleavage in nt of the AV IRES (Fig. 1), and seven of them occur in unpaired regions of RNA. The other five are not in unpaired regions, but instead occur in the terminal base-pair of short helical elements next to unpaired regions (G548, G550, G650 and G740) or between large internal loops, adjacent to a G-U pair (G651). It is thus reasonable to assume that the partial accessibility of these nucleotides to RNAse T1 may reflect conformational flexibility in corresponding regions of the IRES: this possibility is supported for G650 by the partial reactivity of C705 to CMCT modification. We note that probing analyses that led to well-accepted models of the Type 1 IRES (Bailey and Tapprich (2007) J. Virol. 81: ) and of the Type 2 IRES (Pilipenko et al. (1989) Nucl. Acids Res. 17: ) also identified partial sensitivity of ostensibly base-paired regions to chemical modification and to S1 nuclease cleavage, respectively. 3. As the assays in Fig. 2E-F involve monocistronic constructs, it seems important to show that the mutant RNAs that appear to be defective for IRES activity are of comparable stability to the wild-type RNA or, alternatively, verify the results with dicistronic constructs, at least in the case of the AGGUtoUCCA and the del mutants. We did not observe any difference in the stability of [32P]-labeled in vitro transcribed wt and mutant mrnas over the course of translation reactions (1 hour). We would also like to note that in our experience, unlike other in vitro translation systems, RRL does not have significant levels of nuclease activity. 4. It should be stipulated which toe-printing experiments in Fig. 3, and also later in the study, employ recombinant DHX29. As noted in response to Specific Comment #2 raised by reviewer 2 above, we have revised the text in the Methods and the Results sections (p10, lines 3-5; p26, line 5) to indicate those toeprinting experiments that employed recombinant DHX29. Acceptance letter 01 August 2011 Thank you for sending us your revised manuscript. I have now had a chance to go through the manuscript and the changes you made in detail, and you will be pleased to learn that you have addressed all criticisms in a satisfactory manner. The paper will now be publishable in The EMBO Journal and you will receive a formal acceptance letter shortly. Thank you very much again for considering our journal for publication of your work. Yours sincerely, Editor The EMBO Journal European Molecular Biology Organization 6