The presence of foreign nucleic acids within a cell is a universal signal of pathogen invasion and the target for a wide range of immunity systems, ranging from restriction enzymes and CRISPR systems in bacteria to a coordinated array of cell-intrinsic and -extrinsic responses in mammals. Following detection of nucleic acids, typically derived from an infecting virus, by a range of dedicated pattern recognition receptors (PRRs), mammals initiate a cascade of signaling through the type I interferons to alert neighboring cells and induce a widespread anti-viral state. At the same time, infected cells induce cell-intrinsic defenses to limit viral replication and spread. One canonical response is the activation of protein kinase R (PKR) and RNase L to shut down cellular translation, preventing viral gene translation and subsequent viral particle production. However, it has been a mystery how the global shutdown of cellular translation and the requirement to continue to produce anti-viral proteins and type I interferons. This paper presents an elegant explanation solving this long-standing mystery in anti-viral responses.
- RNase L regulates stress granule assembly during an anti-viral response
The authors first demonstrated that stress granules, which are RNA-protein complexes formed downstream of PKR, are altered by RNase L activity. By using knockout cells and rescue experiments, the authors demonstrated that RNase L catalytic activity blocks classical stress granule formation at least partially through degradation of specific mRNAs important for stress granule assembly as well as nuclear translocation of a stress granule assembly factor PABPC1. Instead, cells with an active RNase L enzyme form smaller punctate foci the authors term RNase L-dependent bodies (RLBs).
- Anti-viral mRNAs escape RNase L-mediated degradation and continue to translate during an anti-viral response
The authors show widespread decrease in total mRNA levels during an anti-viral response, including a ~70% decrease in total poly(A)+ mRNA transcripts, including multiple transcripts that are not localized to stress granules. However, multiple anti-viral mRNAs, such as the cytokines IFN-b and IL-6, are not decreased. Whole transcriptome analysis revealed that the vast majority of mRNAs are degraded in an RNase L-dependent manner during an anti-viral response, but those that are not are highly enriched for anti-viral and IRF3-responsive genes. Importantly, IFN-b protein continued to be produced and secreted even after the anti-viral response began.
- There is widespread heterogeneity in individual cellular anti-viral responses
In the course of their microscopy experiments, the authors note that there is substantial heterogeneity between cells in their anti-viral response. For example, not all cells that induce RNase L activation produce IFN-b, and vice versa. The same is true for stress granule formation in RNase L knockout cells, as there is no strict connection between stress granule formation and IFN-b production.
This preprint provides strong evidence for a model whereby RNase L activation, during an anti-viral response, leads to remodeling of the cellular transcriptome and translational control through escape from mRNA degradation of anti-viral mRNAs (see Figure 7, reproduced from the preprint below). This allows cells to accomplish the conflicting goals of restricting viral translation and initiating apoptosis while simultaneously producing anti-viral factors for cell-intrinsic defense as well as cell-cell communication. A long-standing mystery in anti-viral immunity is being resolved!
Figure 7. Model of RNase L reprogramming of translation via differential mRNA turnover.
- The system used throughout the paper is poly(I:C) transfection, which activates the dsRNA response. It will be very interesting to examine the cellular heterogeneity and RNase L-dependent responses in different viral infections, particularly in the setting of viruses with known or putative pathway antagonists. How have viruses evolved to subvert this defense system?
- The absence of a predictable sequence motif of RNase L-resistant mRNAs is very intriguing. The authors mention that many of the RNase L-resistant mRNAs contain AU-rich elements, which were very recently shown to be important for RNA localization to a membraneless organelle, the TIGER domain [Ma 2018]. It would be very interesting to see whether RNase L-resistant mRNAs are localized to the TIGER domain or a similar subcellular localization where RNase L may be excluded, providing a general mechanism for escape from degradation.
- There is so much exciting work that can emerge from the observations of heterogeneity in the anti-viral response between single cells. Is this heterogeneity a form of bet-hedging, to evade any single viral escape mechanism (e.g. by decoupling IFN-b production and stress granule assembly, cells can continue to produce interferon even in the presence of a viral PKR antagonist)? Is the total complement of anti-viral proteins produced similar between RNase L-active and inactive cells (or stress granule positive/negative cells), or is there some “specialization” leading to division of labor? How does communication between neighboring cells affect these decisions – the images seem to show neighboring cells taking on different “fates”, is this just from those images or is there paracrine feedback? Further microscopy, single-cell RNA-seq, and systems circuit modeling will all come to play – there’s so much to explore!
- Ma, W. et al. A Membraneless Organelle Associated with the Endoplasmic Reticulum Enables 3’UTR-Mediated Protein-Protein Interactions. Cell 175(6) 1492-1506.e19
Posted on: 21st December 2018