Can Nodamura virus B2 protein boost self-amplifying RNA activity?

dsRNA replication intermediates limit saRNA output

Self-amplifying RNA (saRNA) is an alphavirus-derived vector in which the structural genes are replaced by a gene of interest, while the non-structural proteins nsP1–4 still form a replication complex that copies the RNA inside the cell. That replication is what gives saRNA prolonged antigen expression and self-adjuvanticity, and it is also what creates the problem: copying the genome generates double-stranded RNA (dsRNA) intermediates, and cytosolic dsRNA is exactly what cell-intrinsic antiviral systems are built to detect.

The paper frames the trade-off directly. The same dsRNA that makes saRNA immunostimulatory also activates effectors that impair its stability, replication, transcription or translation. So the design question is not how to silence innate immunity wholesale, but whether a specific intracellular restriction can be relieved while leaving the adjuvant signal intact. The authors' answer is to encode a viral protein, Nodamura virus (NoV) B2, in cis from the saRNA itself, so that the suppressor is produced in the same cells where the vector replicates.

B2 relieves different blocks in stem and somatic cells

The study separates two cell contexts because the dominant restriction differs between them. In E14 mouse embryonic stem cells transfected with a CHIKV-based GFP saRNA, the vector triggered accumulation of 21–23-nucleotide small RNAs mapping to both strands of the genome, the signature of Dicer cleaving dsRNA replication intermediates. Expressing wild-type NoV B2 reduced the accumulation of these small RNAs and raised GFP expression; the R59Q mutant, which cannot bind dsRNA, did not. The authors read this as suppression of antiviral RNA interference, and they are explicit that they did not confirm it by depleting Dicer or Ago2, so other B2-sensitive pathways cannot be excluded in stem cells.

In somatic cells the mechanism shifts to protein kinase R (PKR), an interferon-stimulated kinase that binds dsRNA, autophosphorylates, and phosphorylates eIF2α to shut down translation. In HeLa and THP-1 cells, wild-type B2 raised GFP expression while the R59Q mutant stayed restricted. The decisive comparison is genetic: in HeLa PKR-knockout cells, losing PKR alone relieved the restriction and the advantage conferred by B2 largely disappeared. In HEK293T cells lacking both Dicer and PKR, GFP levels were similar across saRNAs regardless of functional B2, whereas Dicer-only knockouts still showed B2-dependent enhancement. That pattern places PKR, not Dicer, as the relevant brake in these somatic lines.

B2 acts downstream of PKR autophosphorylation

A useful detail is where in the PKR pathway B2 intervenes. PKR autophosphorylation (P-PKR) levels were similar in THP-1 cells receiving B2-expressing or B2-R59Q saRNA, but phosphorylated eIF2α was detectable only with the mutant. Flow cytometry and confocal imaging agreed: P-eIF2α was present in cells transfected with the R59Q vector and absent with wild-type B2, and P-eIF2α levels ran inversely with GFP expression. So B2 does not stop PKR from being activated by dsRNA; it prevents the downstream phosphorylation that would otherwise halt translation.

Two independent checks support the PKR assignment. Reintroducing wild-type PKR into Dicer/PKR double-knockout HEK293T cells restored restriction of the B2-R59Q vector, while the K296R mutant (defective in ATP binding, dimerisation and phosphorylation) and the D414A kinase-dead mutant did not, indicating kinase activity is required. And the small-molecule PKR inhibitor C16 raised GFP expression from the B2-R59Q vector in a dose-dependent way, with the two vectors converging at 1 µM C16 in THP-1 cells. C16 also increased the percentage of GFP-positive cells up to eightfold for both vectors, which the authors interpret as evidence that baseline PKR activity limits how permissive cells are to saRNA at all.

B2 moves dsRNA to the cell periphery

The spatial data explain how a dsRNA-binding protein can blunt PKR signalling without abolishing interferon induction. Tagged NoV B2 expressed from the saRNA localised predominantly at the cell periphery, adjacent to but not overlapping the membrane-anchored replication complex component nsP1, with partial colocalisation with nsP2. The R59Q mutant, by contrast, was diffusely spread through the cytoplasm, and a time course showed wild-type B2 progressively excluded from the perinuclear region while the mutant stayed everywhere.

dsRNA distribution followed the protein. In HeLa PKR-knockout cells, dsRNA was broadly dispersed through the cytoplasm when functional B2 was absent, but became enriched at the cell periphery when wild-type B2 was expressed, with orthogonal views placing B2 just beneath dsRNA foci rather than directly overlapping them. The authors describe this as sequestration of dsRNA at the periphery, shielding it from antiviral effectors. This is a mechanistic interpretation built on colocalisation imaging rather than a demonstrated causal step, and the paper treats it as such.

Type I interferon induction survives B2 expression

The central design claim is that enhancement does not come at the cost of adjuvanticity. In HeLa cells, the B2 advantage persisted in the presence of the JAK inhibitor ruxolitinib and with added universal type I interferon, so the effect was not simply a byproduct of blocking interferon signalling. In THP-1 Dual reporter cells, B2 expression did not suppress interferon induction or NF-κB signalling, and the authors report no increased interferon-stimulated gene induction accompanying B2-mediated enhancement under their conditions.

That combination is what distinguishes this from simply co-delivering a broad interferon antagonist. The paper compares B2 with other viral suppressors of RNAi: Flock house virus B2 enhanced saRNA activity similarly to NoV B2, enterovirus 71 3A did not counteract the restriction, and vaccinia virus E3L improved GFP only modestly and left saRNAs less active in wild-type than in PKR-knockout cells, consistent with only partial PKR inhibition. Among the tested proteins, NoV B2 and FHV B2 most effectively relieved PKR-mediated restriction.

The effect extends to vaccine-relevant antigens

Because a GFP reporter is a proxy, the relevant question is whether the gain holds for antigens and backbones that matter for vaccines. The paper reports enhancement across CHIKV-, Semliki Forest virus- and VEEV-based saRNA vectors, three alphavirus backbones used for vaccination, and in both stem and somatic cells. It also reports increased expression of a codon-optimised prefusion SARS-CoV-2 Spike protein in HEK293T cells from CHIKV-Spike and VEEV-Spike constructs carrying wild-type versus R59Q B2, with western blots in HeLa cells comparing B2 placed upstream or downstream of Spike.

These are still cell-culture measurements of protein abundance, not immunogenicity or protection data. The paper is explicit that in vivo evaluation remains future work, and it notes that prolonged cell viability and sustained transgene expression could in principle shift immune activation kinetics in a context- and dose-dependent manner. The authors also flag that combining cis-expressed B2 with other reported optimisations, such as modified nucleotide substitution or co-expression of other viral proteins, is untested here.

What the evidence supports and what it does not

The strongest, most specific result is the genetic one: in the somatic lines tested, PKR loss removes the benefit of B2, and catalytically active PKR is required to restore restriction in complemented knockout cells. That makes PKR-driven eIF2α phosphorylation the best-supported target of B2 in these cells. The stem-cell arm is weaker in the paper's own accounting, because the RNAi conclusion rests on small-RNA sequencing and B2's dsRNA-binding mutant rather than on Dicer or Ago2 depletion, which the authors say is needed to establish the functional contribution formally.

Scope also matters. The work uses immortalised and stem-cell lines (HeLa, THP-1, HEK293, HEK293T, BHK-21, E14 mESCs) with transfection of defined saRNA doses, and the authors suggest the stem-cell findings may extend to human induced pluripotent stem cells and adult stem cells expressing the antiviral Dicer isoform aviD, but that is an extrapolation, not a measurement here. No animal or clinical data are presented.

Why B2 beats a broad interferon blocker here

The paper's practical priority is a specific one: if the goal is more antigen per saRNA dose, the comparison to make is between suppressing a downstream effector and suppressing the sensing pathway. B2 sits on the downstream side. It leaves PKR autophosphorylation and type I interferon induction intact while preventing eIF2α phosphorylation and redistributing dsRNA to the periphery, so the adjuvant signal that saRNA depends on is not traded away for expression. The head-to-head suppressor comparison points the same way: E3L, which inhibits both Dicer and PKR, performed worse than NoV B2 in wild-type relative to PKR-knockout cells.

The decision threshold the data suggest is cell context. Where PKR is the dominant brake, as in the HeLa and THP-1 lines used here, B2 adds little once PKR is removed, so the gain should be largest in cells with active PKR signalling. Where antiviral RNAi dominates, as in the stem-cell setting, the evidence is suggestive but not yet genetically nailed down. Reading the paper as a validated vaccine design would outrun the data; reading it as a mapped, testable route to relieve a specific intracellular restriction while keeping adjuvanticity is what the experiments support.

Sources used in this answer