DDX3 regulates the cap-independent translation of the Japanese encephalitis virus via its interactions with PABP1 and the untranslated regions of the viral genome

Reviewer #1 (Public review):

Summary:

In cells undergoing Flavivirus infection, cellular translation is impaired but the viruses themselves escape this inhibition and are efficiently translated. In this study, the authors use very elegant and direct approaches to identify the regions in the 5' and 3' UTRs that are important for this phenomenon and then use them to retrieve two cellular proteins that associate with them and mediate translational shutoff evasion (DDX3 and PABP1). A number of experimental approaches are used with a series of well-controlled experiments that fully support the authors' conclusions.

Strengths:

The work identifies the regions in the 5' and 3' UTRs of the viral genome that mediate the escape of JEV from cellular transcriptional shutoff, they evaluate the infectivity of the mutant viruses bearing or not these structures and even explore their pathogenicity in mice. They then identify the cellular proteins that bind to these regions (DDX3 and PABP1) and determine their role in translation blockade escape, in addition to examining and assessing the conservation of the stem-loop identified in JEV in other Flaviviridae.

In almost all of their systematic analyses, translational effects are put in parallel with the replication kinetics of the different mutant viruses. The experimental thread followed in this study is rigorous and direct, and all experiments are truly well-controlled, fully supporting the authors' conclusions

Reviewer #2 (Public review):

Summary:

The authors use a combination of techniques including viral genetics, in vitro reporters, and purified proteins and RNA to interrogate how the Japanese encephalitis virus maintains translation of its RNA to produce viral proteins after the host cell has shut down general translation as a means to block viral replication. They report a role for the RNA helicase DDX3 in promoting virus translation in a cap-independent manner through binding a dumbbell RNA structure in the 3' untranslated region previously reported to drive Japanese encephalitis virus cap-independent translation and a stem-loop at the viral RNA 5' end.

Strengths:

The authors clearly show that the Japanese encephalitis virus does not possess an IRES activity to initiate translation using a range of mono- and bi-cistronic mRNAs. Surprisingly, using a replicon system, the translation of a capped or uncapped viral RNA is reported to have the same translation efficiency when transfected into cells. The authors have applied a broad range of techniques to support their hypotheses.

Weaknesses:

(1) The authors' original experiments in Figure 1 where the virus is recovered following transfection of in vitro transcribed viral RNA with alternative 5' ends such as capped or uncapped ignore that after a single replication cycle of that transfected RNA, the subsequent viral RNA will be capped by the viral capping proteins making the RNA in all conditions the same.

(2) The authors report that deletion of the dumbbell and the large 3' stem-loop RNA reduce replication of a Japanese encephalitis virus replicon. These structures have been reported for other flaviviruses to be important respectively for the accumulation of short flaviviral RNAs that can regulate replication and stability of the viral RNA that lacks a polyA tail. The authors don't show any assessment of RNA stability or degradation state.

(3) The authors propose a model for DDX3 to drive 5'-3' end interaction of the Japanese encephalitis virus viral genome but no direct evidence for this is presented.

(4) The authors' final model in Figure 10 proposes a switch from a cap-dependent translation system in early infection to cap-independent DDX3-driven translation system late in infection. The replicon data that measures translation directly however shows identical traces for capped and uncapped RNAs in all untreated conditions so that which mechanism is used at different stages of the infection is not clear.

Reviewer #3 (Public review):

Summary:

This work is a valuable study that aims to decipher the molecular mechanisms underlying the translation process in Japanese encephalitis virus (JEV), a relevant member of the genus Flavivirus. The authors provide evidence that cap-independent translation, which has already been demonstrated for other flaviviruses, could also account in JEV. This process depends on the genomic 3' UTR, as previously demonstrated in other flaviviruses. Further, the authors find that cellular proteins such as DDX3 or PABP1 could contribute to JEV translation in a cap-independent way. Both DDX3 and PABP1 had previously been described to have a role in cellular protein synthesis and also in the translation step of other flaviviruses distinct from JEV; therefore, this work would expand the cap-independent translation in flaviviruses as a general mechanism to bypass the translation repression exerted by the host cell during viral infection. Further, the findings can be relevant for the development of specific drugs that could interfere with flaviviral translation in the future. Nevertheless, the conclusions are not fully supported by the provided results.

Strengths:

The results provide a good starting point to investigate the molecular mechanism underlying the translation in flaviviruses, which even today is an area of knowledge with many limitations.

Weaknesses:

The main limit of the work is related to the fact that the role of the 3' UTR structural elements and DDX3 is not only circumscribed to translation, but also to replication and encapsidation. In fact, some of the provided results suggest this idea. Particularly, it is intriguing why the virus titer can be completely abrogated while the viral protein levels are only partially affected by the knockdown of DDX3. This points to the fact that many of the drawn conclusions could be overestimated or, at least, all the observed effect cannot be attributed only to the DDX3 effect on translation. Finally, it is noteworthy that the use of uncapped transcripts could be misleading, since this is not the natural molecular context of the viral genome.

Author response:

Reviewer #1 (Public review):

Summary:

In cells undergoing Flavivirus infection, cellular translation is impaired but the viruses themselves escape this inhibition and are efficiently translated. In this study, the authors use very elegant and direct approaches to identify the regions in the 5' and 3' UTRs that are important for this phenomenon and then use them to retrieve two cellular proteins that associate with them and mediate translational shutoff evasion (DDX3 and PABP1). A number of experimental approaches are used with a series of well-controlled experiments that fully support the authors' conclusions.

Strengths:

The work identifies the regions in the 5' and 3' UTRs of the viral genome that mediate the escape of JEV from cellular transcriptional shutoff, they evaluate the infectivity of the mutant viruses bearing or not these structures and even explore their pathogenicity in mice. They then identify the cellular proteins that bind to these regions (DDX3 and PABP1) and determine their role in translation blockade escape, in addition to examining and assessing the conservation of the stem-loop identified in JEV in other Flaviviridae.

In almost all of their systematic analyses, translational effects are put in parallel with the replication kinetics of the different mutant viruses. The experimental thread followed in this study is rigorous and direct, and all experiments are truly well-controlled, fully supporting the authors' conclusions.

We greatly appreciate the reviewer's recognition of this study. We elucidated the role of UTR in translation blockade escape of JEV from the perspective of the RNA structure of the UTR and its interaction with host proteins (DDX3 and PABP1), and we hope that this study could gain wider recognition.

Reviewer #2 (Public review):

Summary:

The authors use a combination of techniques including viral genetics, in vitro reporters, and purified proteins and RNA to interrogate how the Japanese encephalitis virus maintains translation of its RNA to produce viral proteins after the host cell has shut down general translation as a means to block viral replication. They report a role for the RNA helicase DDX3 in promoting virus translation in a cap-independent manner through binding a dumbbell RNA structure in the 3' untranslated region previously reported to drive Japanese encephalitis virus cap-independent translation and a stem-loop at the viral RNA 5' end.

Strengths:

The authors clearly show that the Japanese encephalitis virus does not possess an IRES activity to initiate translation using a range of mono- and bi-cistronic mRNAs. Surprisingly, using a replicon system, the translation of a capped or uncapped viral RNA is reported to have the same translation efficiency when transfected into cells. The authors have applied a broad range of techniques to support their hypotheses.

We are grateful for the reviewer’s recognition of the thoroughness and multi-faceted nature of our study.

Weaknesses:

(1) The authors' original experiments in Figure 1 where the virus is recovered following transfection of in vitro transcribed viral RNA with alternative 5' ends such as capped or uncapped ignore that after a single replication cycle of that transfected RNA, the subsequent viral RNA will be capped by the viral capping proteins making the RNA in all conditions the same.

Thank you for your suggestion. We share the same viewpoint as the reviewer. After the first round of translation of the uncapped viral RNA, the subsequent viral RNA will inevitably be capped by the viral capping proteins. However, there is no doubt that the transfected cells do not contain viral capping proteins in the initial transfection stage, which directly proved that JEV possesses a cap-independent translation initiation mechanism.

(2) The authors report that deletion of the dumbbell and the large 3' stem-loop RNA reduce replication of a Japanese encephalitis virus replicon. These structures have been reported for other flaviviruses to be important respectively for the accumulation of short flaviviral RNAs that can regulate replication and stability of the viral RNA that lacks a polyA tail. The authors don't show any assessment of RNA stability or degradation state.

Thank you for your suggestion. We agree that a rigorous supplementary experiment for the assessment of RNA stability or degradation state is desirable. To address this, the relative amounts of viral RNA with the deletion of DB2 or sHP-SL will be determined by real-time RT-PCR analysis in transfected cells at multiple time points, which will allow us to test whether the deletion of the dumbbell and the large 3' stem-loop RNA reduce the RNA stability of JEV.

(3) The authors propose a model for DDX3 to drive 5'-3' end interaction of the Japanese encephalitis virus viral genome but no direct evidence for this is presented.

Thank you for your suggestion. In this study, we did not have direct evidence to suggest that DDX3 can drive the 5'-3' end interaction of the Japanese encephalitis virus viral genome, which is indeed a limitation of our research. In the revision, we will more explicitly discuss the interrelationship between DDX3 and 5'-3' UTR, as well as incorporate a discussion of these points into the main text, acknowledging the limitations of our current models.

(4) The authors' final model in Figure 10 proposes a switch from a cap-dependent translation system in early infection to cap-independent DDX3-driven translation system late in infection. The replicon data that measures translation directly however shows identical traces for capped and uncapped RNAs in all untreated conditions so that which mechanism is used at different stages of the infection is not clear.

Thank you for your suggestion. The replicon transfection system was used to evaluate the key viral element for cap-independent translation. We only monitored reporter gene expression from 2 hpt to 12 hpt, which can’t fully recapitulate the different stages of JEV infection. In the experimental results Figure 1 and Figure 1-figure supplement 1, we demonstrated that JEV significantly induced the host translational shutoff at 36 hpi, while the expression level of viral protein gradually increased as infection went on, suggesting that JEV translation could evade the shutoff of cap-dependent translation initiation at the late stage of infection. As shown in the growth curves in Figure 5Q, JEV replicated to similar virus titers in WT and DDX3-KO cells from 12 hpi to 36 hpi, but higher level virus yields were observed in WT cells from 48 hpi, suggesting that DDX3 is important for JEV infection at the late stage. DDX3 was demonstrated to be critical for JEV cap-independent translation. Based on these data, we proposed that the DDX3-dependent cap-independent translation is employed by JEV to maintain efficient infection at the late stage when the cap-dependent translation imitation was suppressed.

Reviewer #3 (Public review):

Summary:

This work is a valuable study that aims to decipher the molecular mechanisms underlying the translation process in Japanese encephalitis virus (JEV), a relevant member of the genus Flavivirus. The authors provide evidence that cap-independent translation, which has already been demonstrated for other flaviviruses, could also account in JEV. This process depends on the genomic 3' UTR, as previously demonstrated in other flaviviruses. Further, the authors find that cellular proteins such as DDX3 or PABP1 could contribute to JEV translation in a cap-independent way. Both DDX3 and PABP1 had previously been described to have a role in cellular protein synthesis and also in the translation step of other flaviviruses distinct from JEV; therefore, this work would expand the cap-independent translation in flaviviruses as a general mechanism to bypass the translation repression exerted by the host cell during viral infection. Further, the findings can be relevant for the development of specific drugs that could interfere with flaviviral translation in the future. Nevertheless, the conclusions are not fully supported by the provided results.

Strengths:

The results provide a good starting point to investigate the molecular mechanism underlying the translation in flaviviruses, which even today is an area of knowledge with many limitations.

Thank you to the reviewer for providing positive feedback. The research on the molecular mechanism underlying cap-independent translation is still a limited field in the flaviviruses, and its mechanism has not been well elucidated at present. We only hope that this study could reveal a novel mechanism of translation initiation for flaviviruses.

Weaknesses:

The main limit of the work is related to the fact that the role of the 3' UTR structural elements and DDX3 is not only circumscribed to translation, but also to replication and encapsidation. In fact, some of the provided results suggest this idea. Particularly, it is intriguing why the virus titer can be completely abrogated while the viral protein levels are only partially affected by the knockdown of DDX3. This points to the fact that many of the drawn conclusions could be overestimated or, at least, all the observed effect cannot be attributed only to the DDX3 effect on translation. Finally, it is noteworthy that the use of uncapped transcripts could be misleading, since this is not the natural molecular context of the viral genome.

Thank you for your suggestion. We agree with the reviewer's comments that the role of the 3' UTR structural elements and DDX3 may not only be circumscribed to translation. However, not as described by the reviewer, DDX3 knockdown did not completely abrogate JEV infection. As indicated in Figure 5E-5F, the recombinant virus was successfully rescued at 36 hpt and 48 hpt using the uncapped viral genomic RNA, although the viral titer rescued with the uncapped genomic RNA at 24 hpt was below the limit of detection. We have confirmed that the DB2 and sHP-SL elements in 3' UTR play a decisive role in the replication of viral RNA in our research (Figure 2G and Figure 2-figure supplement 4C), and we will further analyze the role of DDX3 in viral RNA replication and encapsidation, thereby clarifying the multiple functions of DDX3 in JEV life cycle. Meanwhile, we will incorporate a discussion of these points into the main text, acknowledging the limitations of our current research.

To eliminate the misleading effects of using uncapped transcripts, we will use a natural molecular background of the viral genome with cap methylation deficiency. The methyltransferase (MTase) of the flavivirus NS5 protein catalyzes N-7 and 2’-O methylations in the formation of the 5’-end cap of the genome, and the E218 amino acid of the NS5 protein MTase domain is one of the active sites of flavivirus methyltransferase (PLoS Pathogens. 2012. PMID:22496660; Journal of Virology. 2007. PMID: 1866096). We will construct a mutant virus of the E218A mutation to abolish 2'-O methylation activity and significantly reduce N-7 methylation activity and then analyze the roles of UTR structure and DDX3 in recombinant viruses with the type-I cap structure functional deficiency.