PREreview of Influenza virus antagonizes self sensing by RIG-I to enhance viral replication

In this manuscript, Ledwith, Nipper, Davis and colleagues investigate the identity of the RNAs that activate RIG-I, a cytoplasmic sensor that triggers immune response to infection, identifying that RIG-I binds to viral RNAs, but more crucially to host RNAs that activate this antiviral state during infection. This manuscript also explores a novel mechanism whereby influenza virus nucleoprotein (NP) antagonizes innate immune sensing by RIG-I through binding host-derived non-coding RNAs, particularly those transcribed by RNA polymerase III. Using insights derived through sophisticated quantitative CLIP-Seq and rigorous functional assays (e.g., ISRE activation assays, RNA-immunoprecipitations), the authors provide compelling evidence of a previously unknown role for NP in innate immune evasion. These findings significantly advance our understanding of the modulation of innate immunity during infection, and could reveal potential new targets for antiviral therapeutics in Influenza Virus A (FLUAV) infection.

This manuscript is very thorough, with extensive innovative and elegant experiments to explore NP’s RIG-I antagonism mechanisms, and identify the RNAs targeted by NPs. However, there are a few areas where additional clarification or revision would strengthen the work and make the conclusions more robust. With the proposed edits to improve the paper’s scientific narrative and statistical rigor, this paper has the potential to greatly contribute to scientific efforts to understand innate immunity and how viruses can antagonize such mechanisms. I have outlined my suggestions below.

• Confusion regarding the specific RNAs that NPs bind to: on page 6, the authors state, “NP binds RNA non-specifically, forming sequence-independent interactions with the phosphate backbone allowing it to coat the entirety of the viral genome.” Later on in the manuscript, the authors outline specific regions that NPs bind to (Figure 2G-H), and show that ~25% of the RNAs that NP bind to are host non-coding RNAs. It can be argued that NP binds to specific sequence motifs (repetitive elements being an example) and that NP has preferential binding to certain RNA types, sequences, or structural motifs. This lack of clear hierarchy or binding preference may leave readers uncertain about functional significance (whether NP preferentially binds RNAs critical for immune response, versus binding non-specifically to RNAs abundant during infection). To address this, I recommend the authors explicitly state whether NP binding shows any preference or specificity toward certain RNA classes or structures, and support this using statistical analysis comparing binding enrichment across different RNA biotypes or sequence motifs.

• Provide clearer analysis or discussion on the temporal dynamics of NP-RNA interactions: the manuscript shows significant differences in RNA-binding patterns of NP between constitutively expressed conditions and those induced by infection (Figure 3D). It notes that "more peaks were identified in the absence of infection," indicating widespread changes in RNA binding dependent on infection, yet it does not clearly explain the underlying mechanism driving these differences​. Such mechanisms could be hypothesized with more consistent sampling across the experiment’s time course, as opposed to only at the beginning and end (at 24 hours) where multiple rounds of infection possibly happened. The manuscript further suggests that the subcellular localization of NP, regulated dynamically during infection, partly determines the RNAs bound by NP (Figure 3G-H). For example, MALAT1 RNA, which is exclusively nuclear, is preferentially bound when NP is induced constitutively. This implies that subcellular localization significantly influences binding differences, but a detailed explanation of how localization changes precisely modulate binding preferences across conditions is missing, and this could be addressed with experiments with increased temporal resolution​. Therefore, I recommend the authors provide a clearer explanation connecting the differential RNA binding patterns to the subcellular localization changes and other infection-specific factors. Specifically, I recommend the authors explicitly discuss whether these differences result from infection-induced factors that alter RNA availability, NP's structural or functional modifications, or cellular compartmentalization during different expression states, and suggest further experiments to support such hypotheses.

• Increasing details regarding statistical significance analyses: on page 10 (and in Figure 4B), it is mentioned that “RNA binding sites for RIG-I and NP showed significant overlap” but there is no number for sites for which there is overlap, or a t-test to verify this. On page 26, the authors broadly mention the use of Student's T-tests for pairwise comparisons and one-way ANOVA for multiple comparisons but provide no clear justification or rationale for selecting these specific statistical tests for each dataset or data type presented. It remains ambiguous why certain datasets were deemed appropriate for a T-test versus ANOVA, and no criteria are clearly stated for the choice of specific post-hoc corrections used (e.g., Šídák’s or Tukey’s multiple comparisons tests, as mentioned in the legends for Figures 4 and 5). Therefore, I recommend:

1. the authors clearly describe the rationale behind the selection of each statistical test, why it is suitable for that specific type of experimental data, and report what levels of statistical significance the statistical notations (*,**, etc) imply.

2. the authors conduct multiple hypothesis testing corrections, especially when identifying differentially-bound RNAs (given that CLiP-seq analysis is done on a genome-wide scale).

In addition, there are minor areas of improvement that could improve the manuscript’s readability and flow.

• Expanding the introductory literature review to better outline NP’s RIG-I antagonism mechanism: I recommend that the current literature review be expanded to clearly differentiate NP mechanism novelty from known RIG-I antagonists, particularly clarifying RIG-I discrimination mechanisms of immunogenic RNAs, as the current abstract description may confuse readers. For example, on page 2, it is mentioned that “discrimination occurs after RIG-I binds dsRNA where it has a faster off rate and lower affinity for dsRNA lacking a 5´ PPP that promotes translocation along the length of the RNA and rapid recycling” but it is not mentioned how these differences are  mechanistically relevant to innate immune activation. This point should be expanded upon for clarity. 

• Justifying the choice of cell line: I believe it is crucial for the authors to elaborate upon why they chose human A529 cell lines for this set of experiments. I also recommend that the authors recommend what possible other cell lines or animal models might they carry out further experiments in, which can address their mentioned need for further experiments with increased physiological relevance (page 17).

• Elaboration on non-human models of NP-mediated RIG-I antagonism: on page 6, where the authors discuss how NPs bind host non-coding RNAs, the authors mention, “These findings are reinforced by similar interactions with RIG-I derived from FLUAV hosts like duck and pig.” even if it is not referenced later on in the results section, or elaborated upon with novel experiments. Therefore, this can cause confusion to the readers, and I recommend this sentence be either eliminated or moved to the discussion section, since it is still scientifically relevant to highlight previous work outlining how NPs interact with other mammals’ immune systems. 

• Elaboration on the relationship between NS1 and NP’s functional relationship: similar to the previous point, the authors acknowledge a crucial question regarding the role of NP in antagonizing RIG-I’s response, without explaining how such inquiry fits within the narrative of the paper. For example, The authors mention a positive functional relationship between NS1 and NP in antagonizing RIG-I signaling, as in page 16, without addressing this using experiments. Therefore, I do believe that the authors should revise the manuscript to possibly eliminate the discussion of NS1’s role in antagonizing RIG-I in the introduction, and instead mention it as a possible future direction.

• Improving the resolution and organization of figures for visual accessibility: the figures in this paper are very thorough and detailed, and impart significant support for the authors’ results and conclusions. However, the figures are very dense which make them difficult to read. Specifically: in Figures 2 and 4, some panels cannot be distinguished from each other clearly, and the panels in Figures 1D and 2E are blurry. 

• Incorporating gene names alongside overlap diagrams: the venn diagrams in Figures 1F and 3D are very helpful in highlighting which genes overlap across different infectious conditions. I do recommend, though, that the authors include some specific names of genes that are shared (such as specific U6 and Y RNAs) in the body or in a table, instead of in the supplement, to create more context for the discussion of how these RNAs are involved in immune response.

• Increasing details regarding methods: I recommend the authors elaborate on the ASO purification methodology and how it preserves the RNA’s molecular features (page 13), as it is not clear how those features are preserved or why they are important in identifying the role of the RNAs bound by NP in host immune response. In addition, there is missing information regarding technical controls. For example, in the experiments addressing the RNA dependence of the NP-RIG-I interaction (page 20-21), there needs to be increased information on RNAse concentration and treatment duration to verify that the RNAs were completely eliminated, as partially-digested RNAs could be immunogenic. 

• Ensuring an adequate number of replicates: especially in the high-throughput sequencing analysis, the authors mentioned that they generated “pseudoreplicates” by randomly splitting samples’ reads. This does not genuinely capture biological variability across replicates, and could further highlight batch effects. Therefore, I recommend the authors conduct further experiments to increase the number of biological and technical replicates to increase statistical rigor.

• Spelling errors: multiple minor spelling errors, such as “Principle Component Analysis” (pages 4, 5, 8) and “human A549 lungs cells” (page 6) were identified that should be corrected.

• Sentence syntax: on page 8, the sentence, “Unlike many previously profiled RNA-binding proteins, NP enriched for RNAs spanning several diverse classes and localizing to multiple cellular compartments” needs adjustment for clarity.

Competing interests

The authors declare that they have no competing interests.