Perturbation of RNA homeostasis impairs mitochondrial respiration during poxvirus infection through excess RNA accumulation

Abstract

Induction of RNA degradation in infected cells is a strategy used by many viruses to promote efficient replication. Vaccinia virus, the prototype poxvirus and the vaccine platform for smallpox and mpox, encodes two decapping enzymes to accelerate mRNA and double-stranded RNA (dsRNA) degradation during infection, through functional coordination with host cell RNA exonuclease. Previous studies have largely focused on RNA degradation as a mechanism for regulating viral gene expression and evading innate immune sensing. Here, we show that impaired RNA degradation in vaccinia virus-infected cells, due to either depletion of viral decapping enzymes or cellular exonuclease, severely compromises mitochondrial respiration and integrity. We further demonstrated that accumulation of excess dsRNA and mRNA, including pseudouridine-modified RNAs, is sufficient to induce profound defects in mitochondrial respiration and integrity. Notably, this impairment occurs independently of interferon induction and dsRNA innate immune sensor Protein Kinase R. Moreover, excess RNA directly suppresses respiration in purified mitochondria and physically associates with mitochondria in cell-free and cellular contexts, supporting an immune-independent mechanism. Excess mRNA and dsRNA reduce mitochondrial membrane potential in both cells and purified mitochondria, indicating disruption of the proton gradient as the mechanism underlying impaired mitochondrial respiration and integrity. Together, these findings identify excess mRNA and dsRNA as perturbants of mitochondrial homeostasis in cells with dysfunctional RNA degradation during vaccinia virus infection, revealing a paradigm-shift concept linking RNA metabolism to mitochondrial function. The discovery carries broad implications for understanding RNA and mitochondrial biology, and RNA-based therapeutics and vaccines.

Introduction

RNA degradation is a fundamental post-transcriptional process that regulates RNA abundance and preserves RNA integrity (1). Many viral infections impose an acute conflict between the virus and the host cell for RNA metabolism. Beyond competing with host cell mRNAs for RNA precursor and the translation machinery required for protein synthesis(2), certain viral transcripts, such as double-stranded RNA (dsRNA) and uncapped single-stranded (ss)RNA, are recognized by host sensors that trigger innate immune and inflammatory responses(3). Many viruses encode proteins that promote RNA degradation, thereby redirecting host resources toward viral replication and dampening antiviral innate immune signaling(4). This strategy is exemplified by poxviruses, a large family of double-stranded DNA viruses that include pathogens responsible for deadly diseases such as smallpox and mpox, which have evolved mechanisms to accelerate RNA degradation in infected cells to enhance viral replication(5, 6). The primary mechanism by which poxviruses promote RNA degradation is through decapping enzymes. Vaccinia virus (VACV), the prototype member of the poxvirus family, encodes two decapping enzymes, D9 and D10(7–9). By removing the 5’-caps from mRNAs, the decapping enzymes render the transcripts to rapid degradation by host exonucleases such as XRN1(9–11). This decapping activity also prevents the accumulation of VACV-synthesized dsRNA (dsRNA)(11), a potent activator of host antiviral innate immune responses, facilitating viral evasion of the immune system.

Mitochondria are the “powerhouses” of cells, producing adenosine triphosphate (ATP) as the primary energy molecule through mitochondrial respiration, also known as oxidative phosphorylation(12). In addition to mitochondrial respiration, cells, especially cancer cells, also use glycolysis to generate ATP(13). Together, mitochondrial respiration and glycolysis constitute cellular respiration and are metabolically interconnected (14). A handful of studies have suggested dysfunctional RNA degradation may affect mitochondrial function. In yeast, activation of XRN1 is necessary for yeast growth in glycerol media, suggesting a requirement for XRN1 in ATP production(15). In C. elegans, local coordination of mRNA storage and degradation in the proximity of mitochondria regulates aging, also suggesting an impact of RNA degradation on mitochondrial function(16). Forebrain-specific XRN1 conditional knockout mice exhibit obesity with decreased energy expenditure and other dysfunctions in metabolic homeostasis(17). However, the nature and mechanisms by which RNA degradation may affect cellular respiration and mitochondrial functions are essentially unknown, including in the context of viral infection, which often profoundly modulates RNA decay pathways. RNAs are strongly negatively charged, whereas mitochondrial respiration relies on a positively charged proton gradient across the inner mitochondrial membrane(18, 19). Whether and how excess RNA may affect mitochondrial respiration remains unexplored.

Here, we demonstrate that dysfunction of RNA degradation in VACV-infected cells causes a significant deficiency in cellular respiration, primarily mitochondrial respiration, as well as impaired mitochondrial integrity. We further identified excess mRNA and dsRNA as the agents causing mitochondrial respiration deficiency in both intact cells and purified cell-free mitochondria. Notably, RNA-induced mitochondrial dysfunction occurs independently of interferon (IFN) signaling and the PKR/RNase L pathway. Excess dsRNA or mRNA impairs mitochondrial membrane potential in both cells and a cell-free system using purified mitochondria, indicating disruption of the proton gradient as the mechanism underlying impaired mitochondrial respiration.

Results

Depletion of XRN1 compromises cellular respiration in VACV-infected cells.

Cells undergo a rapid surge in viral RNA levels shortly after VACV infection(6, 11). Meanwhile, mRNA is rapidly degraded in the infected cells, initiated by the virally encoded decapping enzymes, D9 and D10, and subsequently digested by XRN1(10, 11), a 5’ to 3’ exonuclease essential for executing the degradation during RNA decay(20). We first evaluated the effect of loss of XRN1 using an XRN1 knockout cell line, A549ΔXRN1 (21) (Fig. S1A), on cellular respiration and ATP production with or without VACV infection using a Seahorse assay. The oxygen consumption rate (OCR), an indicator of mitochondrial respiration function(22), and extracellular acidification rate (ECAR), an indicator of glycolysis(22) were similar between the A549ΔXRN1 and A549 cells in OCR, mitochondrial ATP (MitoATP), ECAR, and glycolytic ATP (GlycoATP) production (Fig. 1A–C; Fig. S2A–C). The results suggest that XRN1 depletion alone does not markedly affect mitochondrial respiration, glycolysis, or cellular ATP production, possibly due to compensatory adaptations in XRN1-deficient cells.

Fig. 1. Functional RNA degradation is required to maintain mitochondrial respiration during VACV infection.

Seahorse assays were performed to measure OCR and MitoATP. Oligomycin and Rot/AntiA were sequentially injected to assess mitochondrial respiratory parameters. (A-C) A549 and A549ΔXRN1 cells, uninfected. (A) OCR traces. (B-C) Quantification of OCR and MitoATP. (D-F) A549ΔXRN1 cells, uninfected or VACV-infected (MOI = 5). (D) OCR traces. (E-F) Quantification of OCR and MitoATP at 12hpi. (G-I) A549 cells, uninfected or VACV-infected (MOI = 5). (G) OCR traces. (H-I) Quantification of OCR and MitoATP at 12 hpi. (J-L) A549 cells, uninfected, WT-VACV-, or vΔD9ΔD10-infected (MOI = 5). (J) OCR traces. (K-L) Quantification of OCR and MitoATP at 12 hpi. For infection comparison panels, x-axis labels denote the corresponding cell and infection conditions.

VACV-infected A549ΔXRN1 cells exhibited a substantial reduction in OCR and MitoATP, as well as decreased ECAR and GlycoATP (Fig. 1D–F; Fig. S2D–F). In contrast, WT-A549 cells did not show significant changes in OCR and MitoATP production, while ECAR and GlycoATP decreased upon VACV infection (Fig. 1G–I; Fig. S2G–I). These findings suggest that XRN1 is integral to maintaining cellular respiration and ATP production during VACV infection, particularly mitochondrial respiration. Notably, the decrease in OCR during VACV infection of A549ΔXRN1 cells began around 3–4 hours post-infection (hpi) (Fig. 1D), coinciding with the onset of robust VACV post-DNA replication transcription, when large amounts of viral mRNA and dsRNA begin to accumulate. This temporal association supports the idea that the mitochondrial respiratory defect is linked to loss of XRN1 during the stage of abundant post-replicative viral RNA synthesis. The OCR defect becomes progressively more pronounced as infection advances and is robust by 12 hpi, consistent with progressive accumulation of viral RNA species, particularly dsRNA. Moreover, infection with a VACV mutant lacking the transcription factor A8 (vΔA8), which allows early gene expression and viral DNA replication but prevents viral post-DNA replication RNA synthesis(23), did not dampen OCR and ECAR in A549ΔXRN1 (Fig. S1B–D, E–G), suggesting that VACV RNA synthesis post viral DNA replication is associated with this respiratory phenotype, but not viral DNA and events prior to viral DNA replication. Furthermore, to determine whether viral DNA itself impairs respiration, we transfected DNA isolated from VACV-infected cells into both A549 and A549ΔXRN1 cells. No significant changes in OCR, ECAR, MitoATP, or GlycoATP production were observed in either cell line (Fig. S1H–M, N–S), indicating that viral DNA alone is insufficient to disrupt cellular respiration.

Infection of A549 cells with vΔD9ΔD10, a mutant VACV with both viral decapping enzymes knocked out, led to a significant decline in OCR and MitoATP production (Fig. 1J–L). In contrast, cells infected with WT-VACV did not show a decrease in OCR and MitoATP production (Fig. 1J–L). Likewise, vΔA8 infection of A549 cells did not alter mitochondrial OCR or MitoATP (Fig. S1T–Y), again indicating that the observed respiratory defect is associated with post-replicative viral RNA synthesis rather than viral DNA or events preceding viral DNA replication. Interestingly, we did not observe a decrease in ECAR and GlycoATP production in vΔD9ΔD10-infected cells, as seen with WT-VACV infection (Fig. S2 J–L), suggesting that other viral factors also regulate glycolytic activity. Predictably, infection of A549ΔXRN1 cells with vΔD9ΔD10 resulted in a pronounced reduction in mitochondrial respiration (Fig. S3A–F).

To determine whether the mitochondrial respiratory defect impacts VACV replication, we mimicked this phenotype using FCCP to induce mitochondrial dysfunction. Because the OCR defect emerged around 3–4 hpi (Figs. 1D and J), FCCP was added at 4 hpi, which significantly reduced VACV infectious virus production. A reduction in viral titers was also observed when FCCP was added at 1 hpi (Fig. S4A–B). These findings are also consistent with our previous results using CCCP, which showed that mitochondrial oxidative phosphorylation is important for efficient VACV infection in HeLa cells(24).

Taken together, these findings reveal that loss of XRN1 and viral RNA decapping enzymes during VACV infection significantly impair cellular respiration, particularly mitochondrial respiration, and that mitochondrial dysfunction negatively affects efficient VACV replication, underscoring the importance of functional RNA degradation in safeguarding mitochondrial function during VACV infection.

Dysfunctional RNA degradation perturbs mitochondrial integrity in VACV-infected cells

Because impaired RNA degradation is closely linked to reduced mitochondrial respiration during VACV infection, we investigated whether this dysfunction disrupts other key mitochondrial functions, including mitochondrial membrane potential, mitochondrial morphology, and mitochondrial DNA (mtDNA)(25). mitochondrial membrane potential is a vital indicator of mitochondrial function, reflecting the electrochemical gradient across the inner mitochondrial membrane that drives ATP production. Infection of A549 cells with vΔD9ΔD10 led to a significant reduction in mitochondrial membrane potential, as determined by tetramethylrhodamine methyl ester (TMRM) staining, which selectively accumulates in active mitochondria(26). WT-VACV infection caused a moderate decrease in mitochondrial membrane potential compared to uninfected mock cells (Fig. 2A, C). Interestingly, this moderate loss of mitochondrial membrane potential did not result in a significant reduction of OCR or ATP levels (Fig. 1J–L), suggesting that mitochondrial respiration remains functionally compensated despite partial depolarization. This partial mitochondrial membrane potential decrease likely reflects mild mitochondrial stress during infection, while efficient RNA degradation by WT-VACV prevents the accumulation of inhibitory RNA species that would otherwise suppress mitochondrial respiration. Infection of A549ΔXRN1 cells with vΔD9ΔD10 also caused a significant reduction in mitochondrial membrane potential compared to WT-VACV infection; however, the difference was marginal (Fig. 2B, D), likely due to the preexisting impairment of RNA degradation in A549ΔXRN1 cells. Mitochondrial morphology and network connectivity are critical indicators of mitochondrial health, as fragmentation and loss of interconnected networks signify bioenergetic stress and dysfunction. Infection with vΔD9ΔD10 caused a more pronounced disruption of mitochondrial morphology and network connectivity compared to WT-VACV infection in A549 cells (Fig. 2E, G, H). Quantitative image analysis revealed that vΔD9ΔD10 infection significantly reduced the mitochondrial area, perimeter, form factor, and aspect ratio, as well as the number of branches, branch junctions, and branch lengths (Fig. 2GH), indicating severe fragmentation and loss of mitochondrial networking. Similar to mitochondrial membrane potential, WT-VACV infection also disrupted mitochondrial morphology and network connectivity compared to uninfected mock cells. However, the extent of disruption was consistently milder (Fig. 2E, G, H). In uninfected A549ΔXRN1 cells, parameters of mitochondrial morphology and network connectivity tended to be lower than those in A549 cells, and the differences between uninfected and WT-VACV-infected A549ΔXRN1 cells were less prominent (Fig. 2E–H). Notably, vΔD9ΔD10 infection of A549ΔXRN1 cells still caused significantly greater disruption of mitochondrial morphology and network connectivity than WT-VACV infection (Fig. 2F–H). Together, these results indicate that impaired RNA degradation contributes to the loss of mitochondrial membrane potential and mitochondrial network integrity during VACV infection.

Fig. 2. Dysfunctional RNA degradation perturbs mitochondrial integrity during VACV infection.

(A–B) Representative confocal images of mitochondrial membrane potential in A549 (A) and A549ΔXRN1 (B) cells, either uninfected (Mock) or infected with WT-VACV or vΔD9ΔD10 (MOI = 5, 14 hpi). Cells were stained with TMRM (orange) to assess mitochondrial membrane potential. Nuclei were stained with DAPI (blue). (C–D) Quantification of TMRM fluorescence intensity in A549 (C) and A549ΔXRN1 (D) cells corresponding to panels (A) and (B), respectively. (E-F) Representative confocal images of mitochondria in A549 and A549ΔXRN1 cells stained with MitoTracker (red) following mock, WT-VACV, or vΔD9ΔD10 infection (MOI = 5, 14 hpi). (G) Quantitative analysis of mitochondrial morphology parameters in A549 and A549ΔXRN1 cells based on MitoTracker-stained confocal images, including mean area, perimeter, form factor, and aspect ratio. (H) Quantitative analysis of mitochondrial network connectivity parameters in A549 and A549ΔXRN1 cells, including branch number, branch junctions, branch length, and mean branch length. (I-J) Relative mtDNA levels in A549 (I) and A549ΔXRN1 (J) cells infected with WT-VACV or vΔD9ΔD10, quantified by qPCR using COX2 and normalized to nuclear DNA.

mtDNA is essential for mitochondrial function, and its depletion is a hallmark of mitochondrial stress. In A549 cells, vΔD9ΔD10, but not WT-VACV infection, led to significant mtDNA depletion at 24 hpi, while both WT-VACV and vΔD9ΔD10 infections showed similarly reduced mtDNA levels at 36 hpi (Fig. 2I). In A549ΔXRN1 cells, mtDNA depletion occurred earlier and more severely, with significant reductions detected as early as 8 hpi and progressively worsening through 36 hpi, while WT-VACV infection started to deplete mtDNA at a later time (Fig. 2J). These findings suggest that loss of viral decapping enzymes D9 and D10 exacerbates mitochondrial DNA depletion and that the absence of host exonuclease XRN1 further intensifies this effect, highlighting a synergistic role of viral and host RNA degradation pathways in preserving mitochondrial genome integrity during infection. Future studies using XRN1 rescue systems and catalytically inactive mutants will help define the specific contribution of XRN1 enzymatic activity to this response.

Excess dsRNA accumulation dampens cellular respiration

The finding that dysfunctional RNA degradation impairs mitochondrial respiration during VACV infection prompted us to hypothesize that excess RNA accumulation impairs mitochondrial respiration and ATP production. We transfected A549 cells with total RNA extracted from uninfected cells, WT-VACV-infected cells, or vΔD9ΔD10-infected cells. Transfection with RNA from either WT-VACV or vΔD9ΔD10-infected cells led to significant reductions in OCR and MitoATP, while RNA from uninfected cells caused a moderate but still significant decrease compared to uninfected cells (mock) (Fig. 3AB). ECAR and GlycoATP were mildly but significantly reduced by RNAs from WT-VACV or vΔD9ΔD10-infected cells, but not from uninfected cells (Fig. S5AB). Transfection with increasing concentrations of RNA extracted from vΔD9ΔD10-infected cells resulted in a dose-dependent decrease in OCR and MitoATP (Fig. 3CD), while ECAR and GlycoATP remained mildly decreased (Fig. S5CD), indicating that mitochondrial respiration is the primary metabolic process impaired by excess RNA. Importantly, we did not detect VACV proteins in cells transfected with RNA isolated from infected cells, ruling out the possibility that the observed respiratory defects were due to translation of viral proteins from the transfected RNA (Fig. S6AB). The results corroborate that the impairment of cellular respiration is a consequence of the presence of excess RNA but not its encoded proteins.

Fig. 3. Excess dsRNA accumulation dampens cellular respiration.

Seahorse analysis of OCR and MitoATP in A549 cells at 12 h post-transfection. (A-B) Cells transfected with total RNA (100 ng per well, ~30,000 cells) isolated from uninfected cells or from WT-VACV- or vΔD9ΔD10-infected cells. (C-D) Cells transfected with increasing amounts (25, 50, 100 ng) of RNA isolated from vΔD9ΔD10-infected cells. (E-F) Cells transfected with in vitro-transcribed dsRNA (100 ng, ~30,000 cells). (G-H) Cells transfected with increasing amounts (10, 20, 100 ng) of in vitro-transcribed dsRNA.

Ribosomal RNA (rRNA) and transfer RNA (tRNA) together comprise over 95% of total cellular RNA and are not known to be degraded during VACV infection. Therefore, the ratios of total RNAs from uninfected and VACV-infected cells are similar and unlikely to mediate the response. RNA species that are specific or more abundant in VACV-infected cells with dysfunctional RNA degradation are more likely to be responsible. dsRNA is produced during VACV infection through transcription from both strands of the viral genome(6, 27). Mutation of the Nudix catalytic domains in both decapping enzymes D9 and D10 leads to greater dsRNA accumulation(11). Using immunofluorescence staining with the dsRNA-specific J2 antibody, we confirmed this phenomenon in cells infected with our recombinant virus, vΔD9ΔD10, as we observed markedly higher levels of dsRNA in vΔD9ΔD10-infected cells compared to those infected with WT-VACV. In contrast, little or no dsRNA was detectable in uninfected cells (Fig. S6CD).

To test whether dsRNA is sufficient to impair mitochondrial function, we transfected in vitro-synthesized dsRNA, bearing a 5’-cap and 3’-poly (A)-tail, into uninfected A549 cells. Our results show that dsRNA-transfected cells exhibited significant decreases in OCR, MitoATP (Fig. 3EF), ECAR, and GlycoATP, with OCR and MitoATP being more notably reduced (Fig. S5EF). Moreover, this effect was concentration-dependent: higher amounts of transfected dsRNA resulted in greater suppression of both oxidative and glycolytic ATP production (Fig. 3GH; Fig. S5GH). Additionally, transfection with Poly(I:C), a synthetic dsRNA mimic, produced a similar reduction in mitochondrial and glycolytic activity (Fig. S6E–H), confirming that dsRNA is sufficient to impair mitochondrial respiration.

Excess mRNA, including pseudouridine-modified mRNA, dampens cellular respiration in cells

Transfection of total RNA from uninfected cells caused a mild yet statistically significant reduction in OCR (Fig. 3AB), even though no detectable dsRNA was present (Fig. S6CD). In addition, total RNA from vΔD9ΔD10-infected cells did not impair mitochondrial respiration significantly more than RNA from WT-VACV-infected cells (Fig. 3AB), even though vΔD9ΔD10-infected cells produced substantially more dsRNA (Fig. S6 CD). These findings led to the hypothesis that other RNA species also contribute to mitochondrial dysfunction. Since mRNA is a major target of degradation during VACV infection(7, 9), we examined whether mRNA could affect cellular respiration.

We isolated mRNA by purifying poly(A)-tailed RNA from uninfected, WT-VACV-, and vΔD9ΔD10-infected cells. Transfection of mRNA into A549 cells from all three sources significantly and similarly reduced OCR, MitoATP (Fig. 4AB), ECAR, and GlycoATP (Fig. S7AB). Further, transfection of A549 cells with increasing concentrations of mRNA isolated from VACV-infected cells revealed that higher mRNA levels led to a more pronounced decrease in OCR and MitoATP, but not ECAR and GlycoATP (Fig. S8A–D). To rule out the possibility that this effect was due to viral proteins translated from the transfected RNA, we showed that no viral proteins were detected in cells transfected with mRNA from VACV-infected cells (Fig. S8E). In vitro-transcribed mRNA encoding GFP containing a 5′-m⁷G cap and a 3′ poly(A) tail also led to a significant reduction in OCR, MitoATP (Fig. 4CD), ECAR, and GlycoATP (Fig. S7CD), with OCR and MitoATP being more notably affected (Fig. 4CD). VACV-infection produces many short mRNAs(6, 28). Transfection with short mRNAs of varied lengths into A549 cells all significantly decreased mitochondrial respiration, including OCR, MitoATP, ECAR, and GlycoATP (Fig. S8F–I), indicating that the impairment of cellular respiration by mRNA is not dependent on transcript length.

Fig. 4. Accumulation of excess mRNA, including pseudouridine-modified mRNA, dampens cellular respiration.

Seahorse analysis of OCR and MitoATP in A549 cells following RNA transfection (100 ng per well, ~30,000 cells). Measurements were performed at 12 h post-transfection except for modified mRNA conditions, which were analyzed at 15 h. (A-B) Cells transfected with poly(A)+ RNA isolated from uninfected, WT-VACV-, or vΔD9ΔD10-infected cells. (C-D) Cells transfected with in vitro-transcribed mRNA. (E-F) Cells transfected with pseudouridine-modified mRNA (Ψ mRNA) or unmodified mRNA. (G-H) Cells transfected with N1-methylpseudouridine-modified mRNA (m¹Ψ mRNA) or unmodified mRNA. (I-J) Cells transfected with in vitro-transcribed dsRNA or m¹Ψ-modified dsRNA.

Pseudouridine (Ψ) is the most abundant and ubiquitous post-transcriptional RNA modification, found primarily in structural RNAs such as rRNA and tRNA(29–31). The modification is also widely used in the development of therapeutic and vaccine mRNAs. The incorporation of pseudouridine into mRNA enhances mRNA stability, reduces immunogenicity, and increases translational efficiency(32–34). Transfection of A549 cells with pseudouridine-modified in vitro-transcribed mRNA (ΨmRNA) led to a significant reduction of OCR and MitoATP, though the impairment was less severe than that of unmodified mRNA (Fig. 4 EF). Given the widespread use of N1-methyl-pseudouridine (m¹Ψ) modification in mRNA vaccines and therapeutics(35, 36), we extended our investigation to this modification. Similar to Ψ mRNA, m¹Ψ-modified mRNA induced a weaker but still significant reduction in OCR and MitoATP compared to unmodified mRNA (Fig. 4GH). The ΨmRNA and m¹Ψ modified mRNAs did not or only slightly decrease ECAR and GlycoATP (Fig. S7EF and GH). Interestingly, incorporating m¹Ψ into synthetic dsRNA did not reduce the impairment in cellular respiration relative to unmodified dsRNA (Fig. 4IJ and Fig S7IJ).

Excess dsRNA and mRNA accumulation impairs mitochondrial integrity

Transfection of in vitro-synthesized dsRNA, unmodified mRNA, or m¹Ψ-mRNA all led to a significant reduction in mitochondrial membrane potential, as measured by TMRM staining (Fig. 5AB). mtDNA amounts decreased in all three types of RNA-transfected cells at 16 hours post transfection (Fig. 5C). All three RNA species significantly decreased multiple (though not all) parameters of mitochondrial morphology and network organization compared to transfected mock cells, including area, perimeter, form factor, aspect ratio, branch junctions, branch length per mitochondrion, and mean branch length (Fig. 5D–F). These changes suggest that excess RNA alters mitochondrial morphology and disrupts network connectivity, although m¹Ψ-mRNA generally causes milder changes, with some parameters remaining comparable to those of mock-treated cells. Together, these findings demonstrate that excess dsRNA or mRNA disrupts mitochondrial function, as reflected by reductions in mitochondrial membrane potential, mtDNA content, and increased mitochondrial fragmentation, which are likely interconnected features of mitochondrial dysfunction.

Fig. 5. Excess RNA accumulation impairs mitochondrial integrity.

A549 cells were transfected with in vitro-synthesized dsRNA, mRNA, or m¹Ψ-modified mRNA and analyzed at the indicated time points. (A-B) mitochondrial membrane potential assessed at 14 h post-transfection by TMRM staining (orange) and confocal microscopy. Nuclei were stained with DAPI (blue). (B) Mean TMRM fluorescence intensity was quantified. (C) MtDNA levels quantified by qPCR using COX2 and normalized to nuclear DNA at 8 h and 16 h post-transfection. (D-F) Mitochondrial morphology and network connectivity analyzed at 14 h post-transfection. Cells were stained with MitoTracker (red) and imaged by confocal microscopy (D). Size and shape parameters (area, perimeter, form factor, and aspect ratio) and network connectivity metrics (branch junctions, branch length per mitochondrion, and mean branch length) were quantified (E-F).

Excess RNA accumulation-induced cellular respiration impairment is not dependent on RNA-mediated Interferon responses and PKR/RNase L

Many pathogen-associated RNA species, such as double-stranded RNA (dsRNA), are recognized by cellular pattern recognition receptors (PRRs) that trigger innate immune responses, including interferon (IFN) production(37). Under certain circumstances, PRRs can also recognize mRNA to elicit similar responses(38). We therefore investigated whether the mitochondrial respiration defects caused by excess RNA are due to RNA-induced interferon responses. Infection of A549 and A549ΔXRN1 cells with vΔD9ΔD10 or WT-VACV induced little IFNβ (Fig. 6A; Fig. S9A), consistent with the fact that VACV encodes many genes expressed early during infection that evade host innate immune responses(39, 40). Transfection of synthetic RNAs into A549 cells revealed distinct IFNβ induction profiles. At 16 h post-transfection, dsRNA triggered the highest IFNβ levels, followed by poly (I:C) and mRNA, while m¹Ψ-mRNA induced only modest responses (Fig. 6B). At 8 h, IFNβ remained low for all RNA types, with only dsRNA showing a notable increase. Given that mitochondrial respiration defects occurred as early as ~1 h after RNA transfection (Fig S10), these results indicate that impaired respiration precedes IFNβ production. IFNβ treatment did not significantly alter OCR, MitoATP, ECAR, or GlycoATP compared to mock-treated cells (Fig. 6CD; Fig. S9 BC), indicating that type I IFN does not impair mitochondrial respiration. Collectively, these findings demonstrate that excess RNA impairs mitochondrial respiration regardless of whether IFNβ is produced.

Fig. 6. Excess RNA accumulation impairs mitochondrial respiration independently of RNA-mediated IFN responses and PKR/RNase L.

(A) A549 cells were mock-infected or infected with WT-VACV or vΔD9ΔD10 (MOI = 5). Poly(I:C) transfection was included as a positive control. IFN-β levels in culture supernatants were measured by ELISA at 8 and 16 h post-infection. (B) A549 cells were transfected with synthetic RNAs as indicated, and IFN-β levels were measured by ELISA at 8 and 16 h post-transfection. (C-D) A549 cells were treated with recombinant human IFN-β (0, 5000, or 10000 U/mL) for 12 h, and OCR (C) and MitoATP (D) were measured by Seahorse XF analysis. (E-F) A549DKO cells (PKR/RNase L double knockout) were mock-infected or infected with WT-VACV or vΔD9ΔD10 (MOI = 5) for 12 h, followed by Seahorse XF analysis of OCR (E) and MitoATP (F).

In addition to inducing IFNs through certain PRRs, dsRNA can trigger translational arrest via the protein kinase R (PKR) pathway. PKR activation leads to the phosphorylation of eukaryotic initiation factor 2α (eIF2α), which inhibits protein synthesis and restricts viral replication(11, 41). Simultaneously, RNase L, activated by 2’,5’-oligoadenylates produced in response to dsRNA, degrades both viral and cellular RNA, further limiting viral replication and inducing apoptosis(42). The VACV decapping enzymes D9 and D10 counter these antiviral responses by preventing dsRNA accumulation, thereby avoiding PKR and RNase L activation(11). To explore whether the cellular respiration impairment post vΔD9ΔD10 infection is dependent on the PKR and RNase L pathways, we utilized A549DKO cells, in which both PKR and RNase L genes are knocked out(21). Infection of A549DKO cells with vΔD9ΔD10 resulted in significant decreases in OCR and MitoATP, but not in ECAR and GlycoATP (Fig. 6EF; Fig. S9 DE), mirroring the effects observed in A549 cells (Fig. 1J–L). Moreover, both dsRNA and mRNA caused significant reductions in OCR, MitoATP, ECAR, and GlycoATP in A549DKO cells (Fig. S11A–D; Fig. S11E–H), similar to those observed in A549 cells. Transfection of mRNA isolated from VACV-infected cells into A549DKO also significantly dampens mitochondrial respiration (Fig. S11I–L). Together, these results show that mitochondrial dysfunction caused by defective RNA degradation during VACV infection, or by excess exogenous RNA, occurs independently of PKR and RNase L pathways.

RNA compromises the respiration in purified mitochondria in a cell-free system

We next investigated whether RNA can impair mitochondria in a cell-free system. We isolated highly pure mitochondria from HeLa cells, as confirmed by the absence of cross-contamination in compartment-specific immunoblots, and verified their functional integrity by TMRM staining, which demonstrated preserved mitochondrial membrane potential after isolation (Fig. S12A–C). Incubation of isolated mitochondria with m¹Ψ-mRNA for 40 min caused a marked reduction in OCR (Fig. 7A). Unmodified mRNA, by contrast, had no effect unless mitochondria were pretreated with an RNase inhibitor, which unmasked a similar degree of inhibition to m¹Ψ-mRNA (Fig. 7BC). These results are consistent with the differential susceptibility of RNAs to degradation by ribonucleases and possibly by exonuclease activities co-purified with the organelle: m¹Ψ modification stabilizes mRNA against exonucleolytic attack(43), whereas unmodified mRNA is more readily degraded before exerting an effect.

Fig. 7. RNAs associate with mitochondria and impair respiration in isolated mitochondria.

(A-C) Isolated mitochondria were incubated with in vitro-transcribed unmodified or m¹Ψ-modified mRNA for 45 min, and OCR was measured by Seahorse XF analysis. Where indicated, mitochondria were pretreated with RNase inhibitor (RI; SUPERase·In, 1 U/μL) prior to mRNA incubation. (D-G) Isolated mitochondria were incubated with in vitro-transcribed dsRNA, unmodified mRNA, or m¹Ψ-modified mRNA for 20 or 45 min, and OCR was measured by Seahorse XF analysis. (H-I) Isolated mitochondria were incubated at 37 °C for 40 min with in vitro-transcribed dsRNA, unmodified mRNA, or m¹Ψ-modified mRNA (mRNAs pretreated with SUPERase·In where indicated), and mitochondrial membrane potential was assessed by TMRM staining and quantified as mean fluorescence intensity. FCCP was included as positive control. (J-M) Purified mitochondria were incubated at 37 °C for 20 min with fluorescein (F12U)-labeled in vitro-transcribed dsRNA, unmodified mRNA, or m¹Ψ-modified mRNA, washed, stained with MitoTracker, and imaged by confocal microscopy. RNA-mitochondria contacts were quantified by object-based colocalization analysis using DiAna (Fiji) (M). (N-Q) A549 cells were transfected with fluorescein (F12U)-labeled in vitro-transcribed dsRNA, unmodified mRNA, or m¹Ψ-modified mRNA and analyzed at 10 h post-transfection by MitoTracker staining and confocal microscopy. RNA-mitochondria contacts were quantified using DiAna (N). (R-W) A549 cells transfected with fluorescein-labeled in vitro-transcribed dsRNA, unmodified mRNA, or m¹Ψ-modified mRNA were stained with TMRM at 10 h post-transfection and imaged by confocal microscopy. TMRM fluorescence intensities were quantified for mitochondria in contact with RNA and for mitochondria without contacting RNA.

Interestingly, dsRNA and mRNA exhibited different effects on purified mitochondrial respiration. A short exposure to dsRNA (~20 min) transiently increased OCR, whereas longer incubations (45 min) suppressed OCR (Fig. 7DE). This effect, observed without RNase inhibitor treatment, is consistent with the fact that dsRNA is intrinsically more nuclease-resistant(44). The transient OCR increase, apparent after ~20 min of dsRNA treatment, was absent with both unmodified and m¹Ψ-mRNA (Fig. 7FG), suggesting that dsRNA elicits a rapid, stress-like mitochondrial activation that transitions to suppression with sustained exposure. Moreover, all dsRNA, m¹Ψ-mRNA, and unmodified mRNA (with RNase inhibition) significantly reduced mitochondrial membrane potential (Fig. 7HI).

Total cellular RNA isolated from either uninfected or VACV-infected cells also similarly reduced OCR and mitochondrial membrane potential when applied to isolated mitochondria (Fig. S12D–G). These effects differed from those observed in transfected cells, where total RNA from uninfected cells had only a marginal impact (Fig. 3AB), likely because rRNA and tRNA interact with other cellular components in living cells, limiting their access to mitochondria. To determine whether RNA can directly contact mitochondria, we incubated fluorescently labeled RNAs with purified mitochondria. More than 80% of dsRNA, unmodified mRNA, or m¹Ψ-mRNA molecules were found associated with mitochondria (Fig. 7J–M). dsRNA typically appeared as punctate foci on the organelle surface (Fig. 7J), whereas mRNA, whether unmodified or m¹Ψ-modified, was distributed more uniformly along the mitochondrial outer membrane (Fig. 7K–L). In cells, over 50% of transfected RNA puncta were observed in contact with mitochondria (Fig. 7N–Q). Notably, RNA-associated mitochondria showed lower mitochondrial membrane potential signal intensity than adjacent mitochondria lacking RNA contacts (Fig. 7R–W). Together, these data demonstrate that RNAs can directly associate with mitochondria and impair their function in a cell-autonomous manner.

Discussion

Herein, we demonstrate that dysfunctional RNA degradation impairs mitochondrial respiration and integrity during VACV infection. Excess dsRNA or mRNA accumulation is sufficient to trigger this impairment. Importantly, the process is not dependent on RNA-mediated innate immune responses such as IFN and PKR/RNase L and instead results from disruption of mitochondrial membrane potential. This discovery opens new avenues for a fundamental understanding of how RNA degradation affects cellular energy metabolism and mitochondrial function during viral infections, with broad implications for disorders associated with defective RNA degradation and mitochondrial impairment, as well as for the design and improvement of RNA-based therapeutics and vaccines.

A novel RNA-mitochondria interaction concept in VACV infection with broad impact

Our findings indicate that both dsRNA and mRNA dampen mitochondrial respiration, compromise mitochondrial functions, providing a novel mechanistic explanation for why these RNAs are major substrates of VACV-encoded decapping enzymes and the host exonuclease XRN1 (10, 11). By limiting the accumulation of excess dsRNA and mRNA, effective RNA degradation preserves mitochondrial respiration and integrity. Thus, RNA degradation likely serves not only the roles in gene regulation and immune evasion (45). The mitochondrial dysfunction induced by excess RNA may reflect not only cellular stress caused by RNA burden, but also a host defense mechanism through restrictive metabolic state. By reducing ATP production and limiting the bioenergetic resources required for macromolecular synthesis, trafficking, and virion assembly, this fatigue-like state may indirectly constrain the ability of VACV to hijack host cellular processes for efficient replication. In parallel, mitochondrial stress may further influence antiviral signaling pathways, particularly through mitochondrial genome instability and possible exposure of mitochondrial nucleic acids to cytosolic sensing pathways, adding a subsequent layer to the host responses. Many viruses likewise generate abundant dsRNA and mRNA that are targeted by viral and host RNA decay pathways (46–48), but whether RNA degradation similarly modulates mitochondrial respiration in other infections remains to be determined.

In healthy cells, dsRNA and mRNA levels are tightly controlled (49–51), and our findings provide a new essential rationale for cells to tightly control mRNA and dsRNA levels beyond avoiding innate immune responses and regulating gene expression. A few prior studies have linked defective RNA degradation to altered energy metabolism and mitochondrial dysfunction in yeast, C. elegans, and mice (15–17), with unknown mechanism. Our results now show that excess RNA accumulation can directly impair mitochondrial respiration, providing a conceivable mechanistic basis for these earlier observations. Mitochondria are widely believed to have originated from bacteria(52). Consistent with this idea, in bacteria, simultaneous loss of the exoribonucleases PNPase and RNase PH in E. coli causes marked defects in ribosome assembly and severely impaired growth(53), suggesting that proper RNA turnover is required for normal biosynthetic capacity and proliferation in bacteria, and our findings may raise the interesting possibility that excess RNA could also impact bacterial membranes and energy metabolism. Dysfunctional RNA degradation is implicated in inflammatory and autoimmune disorders, cancer, cardiovascular disease, and aging-related neurodegeneration(54–59). Given that many of these diseases are also associated with mitochondrial dysfunction, excess RNA-induced impairment of mitochondrial respiration may represent a previously unrecognized contributor to disease pathogenesis.

Intriguingly, transfection of mRNA impairs mitochondrial respiration, including the m1Ψ-modified mRNA that is widely used in current mRNA-based vaccine platforms (60), although with less severe effects than unmodified mRNA (Fig. 4). Because the amounts of mRNA administered in vaccines and therapeutic settings are small relative to the total number of cells in an organism, these effects will not cause systemic metabolic disruption. The transient, high local concentrations at administration sites could induce mitochondrial stress to some local cells uptake RNAs. The potential subsequent effects, for example, cell death due to energy deprivation, may induce local inflammation and other innate immune responses to enhance immune activation, as a mechanism to enhance vaccine efficacy. Further studies will be required to determine how RNA-induced mitochondrial perturbations influence cell survival, immune cell recruitment, and therapeutic efficacy. Beyond infectious disease vaccination, both mRNA and dsRNA are being actively developed as cancer therapies (61–63). The additional mechanistic layer uncovered in this study may help refine and optimize mRNA and dsRNA-based cancer treatments.

Mechanism by which excess RNA accumulation dampens mitochondrial respiration

Our data indicate that RNA-induced impairment of cellular respiration is not a downstream consequence of innate immune activation. The direct addition of dsRNA or mRNA impaired the respiration of purified mitochondria in a cell-free system, supporting a direct effect of RNA on mitochondrial function (Fig. 7), rather than an indirect consequence of disrupted protein import, ER-mitochondria interactions, or innate immune signaling. The mitochondrial impairment may, in turn, modulate innate immune responses, given that mitochondria serve as central hubs for antiviral signaling(64). We propose a Direct RNA-Mitochondrial Interference (DRMI) model, in which excess RNA disrupts the proton gradient across the inner mitochondrial membrane (Fig. S13), reducing membrane potential and ATP synthesis (19). Because RNAs are strongly negatively charged (18), excess dsRNA and mRNA that accumulate near mitochondria may disrupt the proton gradient across the inner mitochondrial membrane via an electrostatic mechanism, thereby reducing mitochondrial membrane potential and impairing respiration. The close association of RNA with mitochondria in the cell-free system and in transfected cells (Fig. 7) supports this model.

This model (Fig. S13) also explains why total cellular RNA from uninfected cells (with over 95% of rRNA and tRNA (65, 66) and little dsRNA) significantly dampen mitochondrial respiration in the cell-free system (Fig. 7), but had only marginal effects on mitochondrial respiration in transfected cells (Fig. 3). In intact cells, the transfected rRNA and tRNA are largely complexed with ribosomal proteins and other RNA-binding proteins (67–69), which balance their negative charges, whereas much of the transfected mRNA and dsRNA likely remain more exposed. Pseudouridine-modified mRNA caused a less severe impairment of mitochondrial respiration in transfected cells but not in purified mitochondria, consistent with the known effects that pseudouridine modification promotes ribosome recruitment (70). In cells, this increased ribosome association likely shields the negative charges of modified mRNA, reducing its electrostatic interference with mitochondrial proton gradients. However, in purified mitochondria, where ribosomes and cytosolic RNA-binding proteins are absent, the modified and unmodified mRNAs exert comparable effects. Pseudouridine-modified dsRNA impaired respiration to a similar extent as unmodified dsRNA (Fig. 4), possibly reflecting the fact that dsRNA is not a substrate for translation and thus remains unshielded by ribosomes. Excess dsRNA and mRNA also reduced glycolytic activity, measured by ECAR, an indicator of how rapidly cells release protons (H⁺) into their surrounding environment, in transfected cells. This reduction may also result from the excessive accumulation of negatively charged RNA molecules, which can interfere with proton release during glycolysis. As glycolysis occurs in the less spatially confined cytosol, glycolysis is less impaired by excess RNA than mitochondrial respiration. RNA-mediated inhibition of glycolysis may also arise secondarily from mitochondrial dysfunction.

Altogether, this study establishes a novel conceptual link between RNA degradation and mitochondrial respiration during VACV infection and exogeneous introduction. Further defining the molecular players, regulation and physiological contexts of this RNA-mitochondria crosstalk will open a rich avenue for investigations, providing new insights into how cells coordinate RNA homeostasis with mitochondrial function and bioenergetic balance in physiological, pathological, and therapeutic conditions.

Materials and Methods

Cellular respiration was assessed using Seahorse assays. RNA extraction, purification, synthesis, and labeling, as well as other materials and methods are described in detail in the Supporting Information (SI). Data were analyzed using GraphPad Prism software. For comparisons between two groups, t-tests were used, while one-way ANOVA was applied for analyses involving more than two groups. All experiments were performed with at least three biological replicates, and results are expressed as mean ± standard deviation (SD). A p-value of less than 0.05 (p < 0.05) was considered statistically significant. In all the figures, ns, not significant; *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001; ****, P ≤ 0.0001.

Significance Statement.

Vaccinia virus encodes decapping enzymes that promote rapid turnover of mRNA and double-stranded RNA (dsRNA) through coordinating with host degradation factor. Here, we show that dysfunctional RNA degradation in vaccinia virus-infected cells dampens mitochondrial respiration and integrity during infection. Complementary exogenous RNA and cell-free mitochondrial assays demonstrate that excess mRNA and dsRNA can both impair mitochondrial respiration independently of innate immune signaling pathways. These findings reveal that RNA degradation is needed to preserve mitochondrial function during vaccinia virus infection, extending their roles beyond gene regulation and immune evasion. Together, this work uncovers a previously unrecognized dimension of virus-host interaction and provides a framework for understanding how excess RNA burden may influence mitochondrial homeostasis in broad biological and therapeutic contexts.