InsP3R signaling and actomyosin-dependent mitochondrial dynamics play essential roles in mitochondrial stress-induced longevity

Abstract

Certain forms of mitochondrial impairment confer longevity, while disease-associated mitochondrial dysfunction triggers pathogenesis. The adaptive pathways that distinguish benefit from pathology remain unclear. Here we reveal that longevity induced by mitochondrial Complex I/nuo-6 mutation in C. elegans is dependent on the endoplasmic reticulum (ER) Ca²⁺ channel, InsP3R. To explain this connection, we test multiple candidate links between Ca²⁺ and mitochondrial homeostasis previously established in vitro, including mitochondrial calcium uniporter (MCU)-dependent stimulation of respiration and cytosolic pathways regulating mitochondrial dynamics. We find that MCU is dispensable for both respiration and longevity in Complex I mutants. Conversely, transcriptomic profiling and imaging reveal InsP3R impairment results in maladaptive expansion of compromised mitochondrial networks. We provide evidence that this aberrant mitochondrial expansion results from disruption of a conserved, InsP3R-dependent actin remodeling network centered on Arp2/3. Disruption of actin remodeling or autophagy mimics the mitochondrial expansion and longevity suppression of InsP3R mutants. Conversely, driving mitochondrial fragmentation ameliorates mitochondrial expansion and rescues longevity in InsP3R mutants, supporting a model in which InsP3R-dependent actin remodeling is required for segregation and clearance of mitochondria. These findings identify an inter-organelle signaling axis linking ER calcium release and cytoskeletal remodeling to adaptive mitochondrial responses associated with longevity.

Mitochondrial perturbations confer longevity or disease depending on if cells successfully adapt. Here, the authors show that ER calcium signalling promotes longevity during mitochondrial stress by regulating actin-dependent mitochondrial remodeling.

Introduction

Various genetic and pharmacological approaches that perturb mitochondrial electron transport chain (ETC) function can extend lifespan in worms, flies, rodents and potentially humans^1–7. On the other hand, severe mitochondrial impairment results in toxicity and mitochondrial disease. The cellular adaptations that determine whether mitochondrial perturbation results in longevity versus pathogenesis thus possess enormous therapeutic potential but remain surprisingly unclear. Previously, the search for mechanisms acting downstream of these mitochondrial longevity paradigms has yielded insights on roles for signaling from reactive oxygen species (ROS), metabolic rewiring, and retrograde signals to the nucleus that remodel chromatin and transcriptional outputs, such as the mitochondrial unfolded protein response (UPRmt)^8–10. However, outputs such as the UPRmt and ROS levels do not consistently correlate with lifespan, nor are they sufficient for longevity^9,11–13. This latter point indicates that (i) important adaptive mechanisms remain to be discovered, and/or (ii) longevity from mitochondrial perturbation arises through the collective action of multiple integrated processes. Adding further complexity, perturbing different components of the ETC can promote longevity through distinct mechanisms^14,15. Even within the same complex, different types of perturbation—such as mutation versus knockdown—may engage genetically separable pathways that promote longevity, the basis for which remains poorly understood¹⁶. Thus, identifying pathways specifically required for longevity in defined mitochondrial stress contexts is critical.

The current challenges in explaining how mitochondrial stress leads to life-extension has led to suggestions that cellular processes that operate outside of mitochondria must be playing important, poorly understood roles in supporting longevity⁹. Along these lines, one underexplored avenue in understanding mitochondrial longevity paradigms relates to central mitochondrial roles in buffering and responding to fluctuations in intracellular Ca²⁺ levels. Mitochondrial impairment results in altered calcium uptake and storage in the matrix^17–20, thus leading to potential broadscale rewiring of calcium signaling processes in other cellular compartments²¹. Indeed, dysregulation of cellular calcium handling and mitochondrial dysfunction commonly co-occur in age-related diseases and are thought to antagonize one another^22,23, pointing to calcium-dependent processes as key determinants of healthy vs. pathology-associated aging trajectories. How aberrant intracellular calcium signaling exacerbates mitochondrial dysfunction in these contexts remains unclear, however. This lack of clarity stems in part from the inherent challenges in dissecting the multifaceted and highly interconnected interplay between intracellular Ca²⁺ signaling and mitochondrial homeostasis, which involves links to bioenergetic control in the matrix and multiple pathways regulating mitochondrial dynamics and organization.

Inositol triphosphate receptor (InsP3R) channels on the endoplasmic reticulum (ER) membrane act as a unique control point for intracellular calcium signaling, as they are ubiquitously expressed in metazoans and linked directly to both mitochondrial and cytosolic Ca²⁺ pools. InsP3R channels are enriched at ER-mitochondrial contact sites and capable of controlling bioenergetic tone through direct Ca²⁺ flux into the mitochondrial matrix^20,24–26. The mitochondrial outer membrane is relatively permeant to Ca²⁺ due to high levels of voltage-dependent anion channels (VDAC), while the mitochondrial calcium uniporter (MCU) mediates low-affinity, high-capacity uptake of Ca²⁺ into the matrix. Notably, MCU Ca²⁺ uptake is sensitive to mitochondrial membrane potential, and mitochondrial stress can result in cytosolic Ca²⁺ spillover that can reciprocally modulate activity of the InsP3R^17–19,21, thus speculatively providing an avenue for InsP3R channels to sense and respond to mitochondrial dysfunction. In vertebrates, InsP3R-dependent increases in matrix Ca²⁺ stimulate the activity of multiple dehydrogenases associated with the tricarboxylic acid (TCA) cycle, including pyruvate dehydrogenase (PDH), isocitrate dehydrogenase, oxoglutarate dehydrogenase, and potentially even the ETC itself²⁷.

Beyond bioenergetic stimulation, regulation of mitochondrial dynamics has emerged as another major factor in determining the outcome of mitochondrial stress responses and aging outcomes²⁸. Recent studies have revealed that calcium plays important roles in activating mitochondrial fission machineries^29,30 and actin remodeling processes in response to mitochondrial damage^31–33. Notably, these mitochondrial actin networks cage and segregate depolarized mitochondria from the larger network, alter regulation of mitochondrial motility and fission/fusion dynamics, and help to promote adaptive metabolic shifts, such as glycolytic upregulation^31–34. Generally, these studies have focused primarily on acute damage paradigms in vitro, however, and little is known about whether calcium-dependent remodeling of mitochondrial networks plays causal roles in organismal aging or mitochondrial longevity paradigms.

Here, we employed C. elegans to test whether inter-organelle Ca²⁺ signaling plays a role in the adaptive benefits of mitochondrial stress. We focus on the InsP3R, a conserved ER Ca²⁺ channel linked to mitochondrial homeostasis, but whose physiological role in aging and stress adaptation is unclear. We demonstrate that InsP3R activity is required for the longevity conferred by mutation in Complex I. Consistent with observations in vertebrate cells, InsP3R mutants present with bioenergetic defects, but surprisingly, both InsP3R-dependent effects on respiration and lifespan are independent of matrix Ca²⁺ uptake through MCU. In conjunction with InsP3R effects on respiration, a combination of transcriptomic, imaging, and genetic analyses reveal that InsP3R signaling is also linked to calcium-sensitive actomyosin remodeling during Complex I dysfunction, which constrains maladaptive mitochondrial expansion and generates a permissive landscape for successful adaptation to mitochondrial dysfunction. Together, these findings define a previously unrecognized cell biological role for cytosolic Ca²⁺ signaling and actin dynamics in mitochondrial longevity, adding a structural dimension to existing models involving metabolic and transcriptional rewiring.

Results

InsP3R signaling is essential for adaptation and lifespan extension following mitochondrial perturbation

The roles of cellular calcium signaling in mitochondrial longevity paradigms remain poorly defined. Here we set out to test the hypothesis that ER-mitochondrial calcium crosstalk may act as an important driver of longevity in these contexts. To modulate this crosstalk in the aging model, C. elegans, we focused on the InsP3R, which is closely linked to both mitochondrial bioenergetics in vertebrates and to longevity in C. elegans in the contexts of epidermal growth factor (EGF) signaling and ER stress^35,36. While mammals possess three InsP3R genes, C. elegans possess only one, itr-1, presenting a simplified system for genetic modulation. While complete loss of InsP3R/itr-1 function compromises viability³⁷, a temperature-sensitive point mutant, itr-1(sa73), exhibits mild to moderate reductions in the frequency, velocity and magnitude of physiologic Ca²⁺ oscillations under normal culture conditions (20 °C)^38,39. Consistent with previous links to bioenergetic signaling in worms and direct bioenergetic regulation in vertebrates^25,36, we found that InsP3R/itr-1(sa73) mutants exhibit reductions in basal and FCCP-induced oxygen consumption rates (OCR) that are similar in magnitude to animals harboring a mutation in Complex I subunit, gas-1 (Fig. 1a–c). This effect on OCR is independent of gross ER stress or dysfunction, as impairing ER homeostasis more directly through the ire-1/xbp-1 pathway has no effect on respiration (Supplementary Fig. 1a–c). These results indicate a key role for InsP3R/itr-1 signaling in maintaining mitochondrial respiration at the organismal scale.

Fig. 1. InsP3R/itr-1 is required for longevity in Complex I/nuo-6 mutants.

a Representative traces of OCR in InsP3R/itr-1 and gas-1 mutants; mean ± SEM. Additions of FCCP and sodium azide are indicated by arrows. b, c Quantification of basal (b) and FCCP-induced (c) OCR; n = 3 independent trials; mean ± SEM; one-way ANOVA with Tukey’s multiple comparisons test. d Percent of wild-type and InsP3R/itr-1 mutant worms reaching adulthood 10 days post egg lay after dsRNA knockdown against long-lived ETC complexes. n = 201, 110, 211, 137, 194, 123, 207, 136, 199, 129 worms from 4 independent trials (left to right); mean ± SD; two-way ANOVA with Fisher’s LSD test. e Representative images showing the development of itr-1(sa73) worms 10 days post egg lay on RNAi against ETC complexes. All images taken at 4×. f Lifespan of InsP3R/itr-1; Complex I/nuo-6 double mutants. n = 100 worms per condition. Statistical significance of lifespan curves was determined by a log-rank Mantel–Cox test. See Supplementary Data 1 for lifespan statistics. Source data are provided as a Source Data file.

To investigate if InsP3R/itr-1 signaling is broadly required for organismal adaptation to mitochondrial dysfunction, we first tested how InsP3R/itr-1 mutants tolerate RNA interference (RNAi)-mediated impairment of ETC complexes I (nuo-6), III (cyc-1), IV (cco-1), and V (atp-3). Intriguingly, we found that knockdown of Complexes III, IV, and V caused profound developmental arrest selectively in the InsP3R/itr-1 background (Fig. 1d, e). In contrast, Complex I/nuo-6 knockdown was compatible with development in these animals. These differences were not attributable to RNAi efficacy (Supplementary Fig. 1d, e) and instead suggest that InsP3R function is essential for developmental adaptation to ETC inhibition at Complexes III–V. Because Complex I/nuo-6 impairment was uniquely tolerated in InsP3R/itr-1 mutants, we next asked whether InsP3R function was also required for the lifespan extension typically observed in Complex I mutants. We crossed itr-1(sa73) into a long-lived nuo-6(qm200) background¹⁶ and found that reduced InsP3R function completely suppressed the extended lifespan of nuo-6 mutants (Fig. 1f), which typically ranges from 75 to 120% on standard media (see Supplementary Data 1 for complete statistics). This genetic suppression supports a model in which InsP3R signaling is required for the adaptive responses that mediate longevity in the context of Complex I impairment.

To determine whether this requirement was specific to the nuo-6 mutant or extended to other Complex I–related longevity paradigms, we also tested the effect of InsP3R loss on lifespan extension induced by RNAi knockdown of nuo-6. Prior work demonstrated that nuo-6(RNAi) and nuo-6 mutation extend lifespan via genetically independent mechanisms, though those precise mechanisms remain undefined¹⁶. In line with this, we found nuo-6(RNAi) longevity was unaffected by InsP3R/itr-1 (Supplementary Fig. 1f), indicating that InsP3R signaling is not globally required for ETC-induced longevity. To further assess the specificity of InsP3R’s requirement for longevity, we tested two other paradigms distinct from ETC perturbation. First, lifespan extension from age-1 insulin-signaling mutants was only modestly suppressed by InsP3R/itr-1 loss (Supplementary Fig. 1g), and dietary restriction robustly extended lifespan in both wild-type and InsP3R/itr-1(sa73) mutants (Supplementary Fig. 1h). These findings demonstrate that InsP3R/itr-1 mutants do not block longevity broadly by causing pleiotropic physiological dysfunction, but rather InsP3R signaling is selectively required for the pro-longevity program triggered by Complex I mutation.

Finally, we also used an independent approach for reducing InsP3R function without potential disruption of channel structure in InsP3R point mutants. We generated animals ubiquitously expressing an InsP3 sponge, a validated construct that sequesters InsP3R agonist to reduce InsP3R-driven calcium efflux³⁷. In this model, we observed that lifespan extension was substantially diminished ~40% when the sponge is expressed (Supplementary Fig. 1i). While the suppression in this case is not complete, likely stemming from reduced penetrance relative to sa73 mutation or compensatory upregulation of inositol phosphate species, these results nevertheless support a causal role for InsP3R signaling in mitochondrial stress-induced longevity.

InsP3R-dependent control of respiration and lifespan is independent of MCU

We next sought to understand how InsP3R function promotes longevity in the context of Complex I inhibition. In cancer cells, constitutive InsP3R-dependent Ca²⁺ transfer to the mitochondrial matrix supports oxidative metabolism, and impairing ER-mitochondrial Ca²⁺ coupling via either InsP3R or MCU inhibition can trigger bioenergetic collapse^25,40,41. Because RNAi-based ETC perturbation in C. elegans lifespan studies can exhibit similar threshold-dependent effects⁴², we considered a model where InsP3R-mediated Ca²⁺ transfer into Complex I/nuo-6 mutant mitochondria could sustain function above a critical threshold, thereby enabling adaptive responses and longevity. This model predicts that respiration will be robustly impaired when InsP3R or MCU loss is combined with Complex I/nuo-6 mutation, and that matrix Ca²⁺ levels in these mutants would similarly correspond with longevity phenotypes. Upon validating that nuo-6 mutants indeed disrupt Complex I activity in alignment with prior studies^16,43 (Supplementary Fig. 2a), we tested how the interaction between InsP3R/itr-1 and Complex I/nuo-6 mutants impacts respiration. While both Complex I/nuo-6 and InsP3R/itr-1 single mutants exhibit reduced basal and FCCP-induced OCR, combining mutations did not produce additive declines in basal OCR relative to Complex I/nuo-6 mutants alone (p = 0.17) (Fig. 2a–c). In contrast, FCCP-induced OCR was ~48% lower in double mutants than Complex I/nuo-6 mutants alone, indicating additional reductions in reserve respiratory capacity in the double mutants. These data suggest that InsP3R loss does not impose a steady-state respiratory limitation on Complex I mutants but instead constrains expansion of respiratory flux under uncoupled conditions or high demand. While these data do not strongly support a bioenergetic collapse, the additive impairment of FCCP-induced OCR in Complex I; InsP3R double-mutants does suggest that InsP3R signaling either enhances or acts in parallel of Complex I function to promote mitochondrial bioenergetics.

Fig. 2. InsP3R roles in Complex I/nuo-6 longevity are uncoupled from MCU and matrix calcium.

a Representative OCR traces in InsP3R/itr-1; Complex I/nuo-6 mutants; mean ± SEM. Additions of FCCP and sodium azide are indicated by arrows. b, c Quantification of basal (b) and FCCP-induced (c) OCR in InsP3R/itr-1; Complex I/nuo-6 mutants. n = 3 independent trials; mean ± SEM; one-way ANOVA with Tukey’s multiple comparisons test. d, e Representative images of Mito-GCaMP7 and Mito-mKate2 imaging in intestine of day 1 (d) and day 7 (e) adults of InsP3R/itr-1; Complex I/nuo-6 mutants. Scale bar: 10 µm. f Quantification of the ratio of GCaMP7 to mKate2 fluorescence as an indicator of basal mitochondrial calcium level. n = 115, 115, 123, 122, 138, 103, 120, 59 worms from 3 independent trials (left to right); mean ± SD; two-way ANOVA with Tukey’s multiple comparisons test. g Representative images of TMRE staining in hypodermal mitochondria of day 1 adults of InsP3R/itr-1; Complex I/nuo-6 mutants. Scale bar: 10 µm. h Quantification of mean TMRE intensity. n = 50, 50, 50, 59 worms from 3 independent trials (left to right); mean ± SD; one-way ANOVA with Tukey’s multiple comparisons test. i Representative OCR traces in mcu-1; Complex I/nuo-6 double mutants; mean ± SEM. Additions of FCCP and sodium azide are indicated by arrows. j, k Quantification of basal (j) and FCCP-induced (k) OCR in mcu-1; Complex I/nuo-6 double mutants. n = 3 independent trials; mean ± SEM; one-way ANOVA with Tukey’s multiple comparisons test. l Lifespan of mcu-1; Complex I/nuo-6 double mutants. n = 100 worms per condition. Statistical significance of lifespan curves was determined by a log-rank Mantel–Cox test. See Supplementary Data 1 for lifespan statistics. Source data are provided as a Source Data file.

Next, we asked whether matrix Ca²⁺ correlates with the effects of Complex I/nuo-6 and InsP3R/itr-1 on lifespan. To quantify relative matrix Ca²⁺ levels, we co-targeted Ca²⁺-sensitive GCaMP7 and Ca²⁺-insensitive mKate2 as a normalization factor to the mitochondrial matrix (Fig. 2d–f). After confirming the reporter responded as expected to conditions that reduce (mcu-1 loss) or enhance (SERCA/sca-1 inhibition) matrix Ca²⁺ levels (Supplementary Fig. 2b, c), we analyzed Complex I mutants. We observed robust decline of resting matrix Ca²⁺ in Complex I/nuo-6 mutants; however, we observed no additive effects when InsP3R and Complex I mutants were combined (Fig. 2d–f), indicating that Complex I status is the dominant determinant of matrix Ca²⁺ levels. Notably, at advanced ages, matrix Ca²⁺ in wild-type and itr-1 mutants also decline to comparable levels as Complex I mutants (Fig. 2d–f), mirroring previously reported age-dependent declines in mitochondrial membrane potential^44,45. As mitochondrial Ca²⁺ uptake is membrane potential-dependent, we surveyed mitochondrial membrane potential across genotypes via tetramethylrhodamine ethyl ester (TMRE) staining. Consistent with the observed changes in matrix Ca²⁺, we observed a substantial decrease in TMRE staining of mitochondria in Complex I/nuo-6 mutants and no apparent contributions from InsP3R/itr-1 (Fig. 2g, h). These results reveal that long-lived Complex I mutants indeed exhibit clear alterations in calcium homeostasis. However, Complex I- and InsP3R-dependent effects on matrix Ca²⁺ levels and membrane potential do not correlate with their respective effects on lifespan, arguing against a model where longevity is determined through direct ER-to-mitochondrial Ca²⁺ transfer.

While resting matrix Ca²⁺ does not correlate with OCR or the lifespan outcomes, we also reasoned that difficult-to-detect Ca²⁺ transients from the InsP3R could still stimulate bioenergetics and other mitochondrial behaviors^30,46,47. We first examined whether calcium-sensitive matrix dehydrogenases are responsive to InsP3R signaling in C. elegans. In vertebrates, rises in matrix Ca²⁺ levels stimulate PDH phosphatase activity, subsequently leading to dephosphorylation of PDH subunit E1α at S293²⁷, a site fully conserved in C. elegans pdha-1. While Western blots of phospho-PDH revealed higher levels of the inactive, phosphorylated PDH enzyme in InsP3R/itr-1 mutants (Supplementary Fig. 2d, e), potentially consistent with reduced phosphatase activity, we also observed a corresponding increase in total PDH levels (Supplementary Fig. 2f, g). This result suggests that an expansion of mitochondrial enzymes or networks in InsP3R/itr-1 mutants may drive the effect on phospho-PDH levels rather than a robust, Ca²⁺-dependent shift towards inactivation of PDH. Finally, to rule out potential roles for matrix transients and the MCU complex more generally, we incorporated deletion mutants of mcu-1, the sole channel-forming MCU subunit in C. elegans. While mcu-1 mutants exhibited reduced resting [Ca²⁺]_m as expected (Supplementary Fig. 2b), we found that animals lacking MCU have no observable reduction in OCR (Fig. 2i–k). Furthermore, lifespan extension of Complex I mutants is unaffected by MCU function (Fig. 2l).

Taken together, these results indicate that while the InsP3R stimulates mitochondrial respiration at the organismal scale in C. elegans, the mechanisms by which InsP3R signaling supports mitochondrial function and Complex I longevity are largely independent of direct transfer of Ca²⁺ from ER stores to the matrix via MCU. Thus, while these data do not rule out a role for InsP3R-dependent bioenergetic contributions in Complex I-dependent longevity, neither basal respiration nor MCU-dependent matrix Ca²⁺ appear sufficient to explain the requirement of InsP3R signaling. We therefore investigated alternative Ca²⁺-linked mechanisms acting downstream of the InsP3R.

Transcriptomics reveal dysregulation of mitochondrial biogenesis in InsP3R mutants under mitochondrial stress

We next took an unbiased approach to identify candidate mechanisms by which the InsP3R supports longevity in Complex I mutants. We performed bulk RNA-seq analysis of a panel of mutants that included wild-type controls, InsP3R/itr-1, Complex I/nuo-6, and InsP3R/itr-1; Complex I/nuo-6 double mutants. We defined differentially expressed (DE) transcripts as those which exhibited a fold change (FC) of at least 1.5× (P_adj < 0.01) relative to wild-type (see Supplementary Data 2). This approach revealed widespread transcriptional changes, yielding >2000 DE transcripts in each mutant (Fig. 3a). We categorized DE transcripts to highlight the genes and cellular processes predicted to drive the lifespan phenotypes downstream of Complex I and InsP3R signaling (Fig. 3b). We reasoned the most promising candidates would fall into Category I: transcripts altered in Complex I mutants that are InsP3R/itr-1 dependent. Additionally, Category II transcripts are those altered in Complex I mutants independently of InsP3R function, thus likely uncoupled from lifespan, while Category III transcripts exhibit a synthetic expression profile, altered only when both Complex I and InsP3R function are impaired (Fig. 3a, b and Supplementary Data 3).

Fig. 3. Transcriptomic analysis reveals InsP3R-dependent regulation of mitochondrial biogenesis and adaptive responses to Complex I impairment.

a Venn diagram representing differentially expressed genes (Log₂ FC of 1.5×, P_adj < 0.01) in Complex I/nuo-6 (green), InsP3R/itr-1 (orange), and InsP3R/itr-1; Complex I/nuo-6 (blue). Boxes represent categories (WormCat–Category 2) of genes enriched in the indicated groups. WormCat p-values were calculated using one-sided Fisher’s exact test with Bonferroni correction; categories with P_adj < 0.01 are shown. See Supplementary Data 2 for DESeq2 results and Supplementary Data 3 for a summary of differentially expressed genes. b Model depicting how DE gene categories correlate with genetic regulation of lifespan by nuo-6 and itr-1. c–h Heatmaps comparing the expression of mRNA transcripts in Complex I/nuo-6, InsP3R/itr-1, and InsP3R/itr-1; Complex I/nuo-6 mutants for c ribosome biogenesis, d autophagy, e glycolysis, f lipid beta-oxidation, g electron transport chain, and h mitochondrial ribosome genes. All genes normalized to wild-type. Source data are provided as a Source Data file.

Given the large-scale remodeling of gene expression we observed in these contexts, we next performed gene enrichment analysis within these categories⁴⁸. Notably, our analysis of Complex I/nuo-6 mutants closely matched established signatures of the UPRmt observed upon a variety of forms of mitochondrial stress^49–51. Some of the strongest transcriptional changes we observed in Category II (Complex I dependent, InsP3R-independent) involved transcriptional modules associated with the UPRmt (Fig. 3a), including reduced translation and cytosolic ribosome biogenesis (Fig. 3c), upregulated protein turnover and autophagy (Fig. 3d), and upregulation of glycolysis (Fig. 3e). Interestingly, altered lipid metabolism was a signature present in both Category II and III, indicating that some aspects of lipid metabolism are remodeled similarly in Complex I mutants regardless of InsP3R status, while others are uniquely altered in InsP3R/itr-1; Complex I double-mutants (Fig. 3a). These results suggest that Complex I mutants and InsP3R signaling may converge on distinct sets of lipid metabolism genes. Consistent with this notion, we observed similar overall trends in lipid beta-oxidation transcripts between Complex I/nuo-6 and InsP3R/itr-1; Complex I/nuo-6 conditions, while a subset were strongly InsP3R-dependent (Fig. 3f).

Finally, the processes within Category I include those that are altered during Complex I dysfunction and dependent on intact InsP3R function (Fig. 3a, b), and which we therefore expected to be enriched for roles in mediating InsP3R-dependent lifespan extension. Strikingly, only two categories were enriched in this subset of transcripts: collagens and mitochondrial functions, strongly suggesting that InsP3R signaling was likely to be mediating adaptation and longevity in Complex I mutants at least in part through a mechanism involving mitochondrial homeostasis. Upon closer examination of the genes and processes within this Mitochondrial category, we observed robust induction of mitochondrial ETC components and mitochondrial ribosomes in Complex I/nuo-6 mutants, indicative of a compensatory mitochondrial biogenesis response during chronic mitochondrial dysfunction. However, this induction was substantially blunted in InsP3R/itr-1 mutants (Fig. 3g, h), revealing a role for InsP3R signaling in regulating adaptive expansion of the mitochondrial compartment.

Altogether, our analysis confirms overlap with prior mitochondrial stress signatures and suggests that certain UPRmt transcriptional modules—particularly those related to mitochondrial biogenesis—are selectively dependent on InsP3R function. Interestingly, InsP3R signaling emerged as a hit from two prior genetic screens for UPRmt modulators, though neither study pursued mechanistic validation^52,53. Our data build on these findings by demonstrating that InsP3R function shapes specific transcriptional modules of the UPRmt during Complex I impairment.

InsP3R signaling controls mitochondrial network scaling in response to Complex I dysfunction

To explain the effects of InsP3R signaling on mitochondrial biogenesis-associated transcripts, we hypothesized that the InsP3R may play a role in mitochondrial network expansion and quality control during stress. We analyzed the impacts of InsP3R and Complex I/nuo-6 mutants on mitochondrial networks directly via confocal microscopy of animals harboring natively labeled Complex IV subunit, COX-4::eGFP, focusing on the primary metabolic tissues of C. elegans, the hypodermis (Fig. 4) and intestine (Supplementary Fig. 3). Mirroring the transcriptional signature of upregulated biogenesis, the mitochondrial networks of young Complex I/nuo-6 mutants undergo robust, compensatory expansion and adopt a predominantly tubular network morphology, while InsP3R/itr-1 mutants alone exhibit mild mitochondrial phenotypes (Fig. 4a–c). In a surprising contrast with the prediction from both our transcriptomic results and respiration measurements, we observed an exaggerated expansion of the mitochondrial network in InsP3R/itr-1; Complex I/nuo-6 double mutants (Fig. 4a–c). This effect was generally consistent in the intestine (Supplementary Fig. 3), where double mutants display an expanded and highly fused mitochondrial footprint similar to Complex I mutants, but with further elevated levels of COX-4::eGFP. As mitochondrial dynamics and function are highly dynamic during aging²⁸, we followed this phenotype as the animals reached ages where subsets of mitochondria in control animals begin exhibiting characteristic fragmentation and swelling (Fig. 4a, d). Consistent with this role for the InsP3R in restraining mitochondrial expansion, Complex I/nuo-6 driven mitochondrial expansion is further enhanced over age when InsP3R function is impaired (Fig. 4d, e). Indeed, the mitochondrial network phenotype in aged InsP3R/itr-1; Complex I/nuo-6 double mutants becomes relatively extreme, adopting such a dense and hyperfused configuration that COX-4::eGFP mitochondria appear to fill the cytosol, punctuated only by nuclei and non-fluorescent vesicular structures (Fig. 4a). To better resolve these structures, we performed transmission electron microscopy (TEM) imaging of transverse cross-sections (Fig. 4f–i). Consistent with the expanded tubular network observed in confocal images, we found mitochondrial cross-sections in the Complex I/nuo-6 mutants that appear relatively densely packed together (Fig. 4f, h). In Complex I; InsP3R double mutants, the density of individual mitochondrial tubules appears consistent, but the mitochondria are substantially larger in both young and aged animals, suggesting InsP3R mutants enhance or are additive to processes resulting in mitochondrial expansion (Fig. 4f–i). Additionally, TEM images frequently indicated lipid droplets punctuating the hyper-expanded mitochondrial network in InsP3R mutants (Fig. 4f, h), suggesting that these may be the vesicular structures that appear to accumulate in the fluorescence imaging as well.

Fig. 4. InsP3R signaling constrains mitochondrial network expansion during Complex I dysfunction.

a Representative images of mitochondrial networks (COX-4::eGFP) in hypodermis of young (day 1) and aged (day 7) worms. Scale bar: 10 µm. “N” = nucleus, arrows indicate lipid droplets. b Representative examples of fragmented, intermediate, fused, and hyperfused mitochondrial morphologies. Scale bar: 5 µm. c Categorization of mitochondrial morphologies in young (day 1) worms. n = 46, 41, 43, 35 worms from 2 independent trials (left to right); Kruskal–Wallis test with Dunn’s multiple comparisons test. d Categorization of mitochondrial morphologies in aged (day 7) worms. n = 45, 44, 42, 36 worms from 2 independent trials (left to right); Kruskal–Wallis test with Dunn’s multiple comparisons test. e Quantification of total COX-4 (mean COX-4::eGFP intensity × mitochondrial footprint) in young (day 1) and aged (day 7) worms. n = 46, 41, 43, 35, 45, 44, 42, 36 worms from 2 independent trials (left to right); mean ± SD; two-way ANOVA with Tukey’s multiple comparisons test. f TEM imaging of mitochondria in hypodermal cells of young (day 1) worms. Scale bar: 500 nm. g Quantification of mitochondrion size in young (day 1) worms. n = 23, 202, 49, 105 individual mitochondria from 1 to 2 images each from 2 to 4 worms (left to right); mean with 95% CI; one-way ANOVA with Sidak’s multiple comparisons test on log-transformed data. h TEM imaging of mitochondria in hypodermal cells of aged (day 7) worms. Scale bar: 500 nm. i Quantification of mitochondrion size in aged (day 7) worms. n = 25, 109, 104, 136 individual mitochondria from 1 to 2 images each from 2 to 4 worms (left to right); mean with 95% CI; one-way ANOVA with Sidak’s multiple comparisons test on log-transformed data. f, h Arrows indicate lipid droplets; arrowheads indicate mitochondria. Source data are provided as a Source Data file.

Overall, we find that while the transcriptional and physical expansion of mitochondrial networks are tightly coupled in Complex I mutants, this coupling is lost when InsP3R function is impaired. Instead, InsP3R dysfunction exacerbates mitochondrial network expansion while reducing biogenesis at the transcriptional level. This inverse relationship is consistent with a negative feedback mechanism blunting the transcriptional response in these animals, though a dedicated dissection of the transcriptional activation of these genes is required to confirm this hypothesis. Here, the fact that double mutants exhibit reduced transcript levels yet enhanced, progressive expansion of mitochondrial networks suggests that post-transcriptional mechanisms, such as impaired structural remodeling and turnover, are the potential drivers of mitochondrial phenotypes downstream of InsP3R. As this mitochondrial gene expression signature was also the strongest correlate with lifespan extension from our RNA-seq results (Fig. 3), we hypothesized that control of mitochondrial dynamics and turnover are also mechanistically linked to the effects of InsP3R signaling on lifespan extension in Complex I mutants. As the marked expansion of mitochondrial networks in Complex I/nuo-6; InsP3R/itr-1 mutants occurs despite significantly reduced respiratory output (Fig. 2), we additionally reasoned these observations may reflect a maladaptive response in which dysfunctional organelles accumulate in the absence of proper regulatory control.

Mitochondrial damage-induced actin is required for Complex I mutant longevity

Based on the aberrant mitochondrial scaling and morphologies in InsP3R/itr-1; Complex I/nuo-6 double mutants (Figs. 3 and 4), we next hypothesized that cytosolic calcium signaling might intersect with regulation of mitochondrial dynamics to play important roles in potentiating Complex I longevity. To begin dissecting potential roles for cytosolic mediators of calcium signaling, we first tested a role for calmodulin/cmd-1, which localizes to the cytosol and nucleus and directly binds Ca²⁺ released by the InsP3R to initiate diverse downstream signaling pathways. While calmodulin/cmd-1 knockdown from hatch leads to developmental arrest, strikingly, we found that even post-developmental knockdown of calmodulin/cmd-1 suppresses Complex I longevity (Fig. 5a). This requirement for calmodulin coupled with the lack of an effect from mcu-1 ablation strongly supports a model where cytosolic calcium signals potentiate mitochondrial-mediated longevity.

Fig. 5. InsP3R signaling promotes actin remodeling in Complex I/nuo-6 mutants to regulate mitochondrial dynamics.

a Lifespan of Complex I/nuo-6 mutants fed cmd-1(RNAi) from L4. n = 100 worms per condition. Statistical significance of lifespan curves was determined by a log-rank Mantel–Cox test. See Supplementary Data 1 for lifespan statistics. b Putative working model of mitochondrial stress-associated actin remodeling derived from mammalian studies. Previously described acute paradigms³¹ involve parallel Cdc42 and WAVE pathways regulating formin and Arp2/3-dependent remodeling, while prolonged paradigms³² rely on Cdc42 and WASP-dependent Arp2/3 activation. A combination of myosins mediate actomyosin contractility and mitochondrial trafficking to autophagosomes for turnover. c Heatmap comparing the expression of actin cytoskeletal mRNA transcripts in Complex I/nuo-6, InsP3R/itr-1, and InsP3R/itr-1; Complex I/nuo-6 mutants. All genes normalized to wild-type. d Representative images of F-actin networks (LifeAct::GFP) in hypodermis of L4 worms. Scale bar: 10 µm. e Representative examples of minimal, intermediate, and dense F-actin networks. Scale bar: 5 µm. f Categorization of F-actin network density. n = 89, 102, 52, 77 worms from 4 independent trials (left to right); Kruskal–Wallis test with Dunn’s multiple comparisons test. g Quantification of actin footprint. n = 90, 98, 48, 76 worms from 4 independent trials (left to right); median with quartiles; one-way ANOVA with Tukey’s multiple comparisons test. h Representative images of F-actin networks (LifeAct::GFP) in hypodermis of L4 worms fed arx-2(RNAi). Scale bar: 10 µm. i Categorization of F-actin network density in worms fed arx-2(RNAi). n = 47, 50, 54, 21 worms from 3 independent trials (left to right); Kruskal–Wallis test with Dunn’s multiple comparisons test. j Quantification of actin footprint in worms fed arx-2(RNAi). n = 47, 50, 54, 21 worms from 3 independent trials (left to right); median with quartiles; one-way ANOVA with Tukey’s multiple comparisons test. k Representative images of mitochondrial networks (COX-4::eFGP) in hypodermis of young (day 1) worms fed arx-2(RNAi). Scale bar: 10 µm. l Quantification of total COX-4 (mean COX-4::eGFP intensity × mitochondrial footprint) in worms fed arx-2(RNAi). n = 98, 87, 96, 84 worms from 4 independent trials (left to right); mean ± SD; one-way ANOVA with Sidak’s multiple comparisons test. Source data are provided as a Source Data file.

We next sought to test calcium signaling mediators with more direct links to regulation of mitochondrial dynamics. We focused first on calcium/calmodulin-dependent protein kinase II (CaMKII)/unc-43 and calcineurin/tax-6, which are both linked to lifespan as well as regulation of mitochondrial fission through dynamin-related protein (DRP)-1^29,54–56. While calcineurin/tax-6 knockdown alone was sufficient to extend lifespan as previously shown⁵⁵, Complex I/nuo-6 dependent lifespan extension was independent of both tax-6 and unc-43 (Supplementary Fig. 4a, b). Though we observed no clear role for these potential upstream regulators of fission, we also attempted to determine more directly whether DRP-1 fission complexes are altered in InsP3R and Complex I mutants. We employed a GFP::DRP-1 line to visualize the formation of DRP-1 oligomers in vivo, but discovered fluorescence labeling of DRP-1 significantly impacts its function in vivo (Supplementary Fig. 4c), consistent with prior in vitro findings⁵⁷. To mitigate this issue, we maintained gfp::drp-1 as a heterozygote alongside a wild-type drp-1 copy across all backgrounds and used nuo-6(RNAi) in place of nuo-6(qm200) mutants to avoid synthetic phenotypes (“Methods”). With these caveats, we found that both the number and mean intensity of GFP::DRP-1 complexes were actually increased in the Complex I/nuo-6(RNAi) background, and to an even greater extent in InsP3R/itr-1; Complex I/nuo-6(RNAi) animals (Supplementary Fig. 4d–i). This inverse correlation between fused mitochondrial morphologies and increased DRP-1 puncta suggests the presence of fission complexes alone is insufficient to drive effective fragmentation in these conditions. Notably, given that nuo-6(RNAi)-mediated longevity is independent of InsP3R/itr-1, we found that the mitochondrial hyperexpansion observed in InsP3R/itr-1; nuo-6(qm200) animals did not occur in InsP3R/itr-1; nuo-6(RNAi) conditions (Supplementary Fig. 4j, k). This result is consistent with previous findings that nuo-6(RNAi) and nuo-6(qm200) mutants extend lifespan through independent mechanisms¹⁶, and highlight these differential effects on mitochondrial network dynamics as an important avenue in understanding how these distinct modes of mitochondrial perturbation each extend lifespan. Overall, genetic analysis of upstream mediators, CaMKII and calcineurin, does not support a model in which altered Ca²⁺-dependent signaling directly regulate DRP-1 fission machinery to drive lifespan phenotypes. However, technical challenges limit our ability to directly probe fission complexes in vivo. Improved tools will be needed to more precisely dissect the molecular regulation of DRP-1 downstream of InsP3R signaling in these contexts.

Importantly, alternative Ca²⁺-linked pathways for regulating mitochondrial dynamics have also emerged from recent in vitro studies. These studies reveal that Ca²⁺ signals play a role in triggering actin remodeling around dysfunctional mitochondria, and that these actin “cages” can prevent fusion of damaged mitochondria within the network, enhance metabolic rewiring, and promote dispersal and trafficking of damaged mitochondria for turnover^31–34 (Fig. 5b). Intriguingly, if these damage-induced actin networks are impaired during prolonged mitochondrial dysfunction, mitochondrial networks expand and individual mitochondria become enlarged or swollen³², similar to what we observe in InsP3R; Complex I mutants (Fig. 4). Intriguingly, we also noticed a trend of downregulation (14–43%) of 4 out of 5 actin genes arising in the double mutants in our transcriptomic dataset (Fig. 5c), consistent with a potential link between InsP3R signaling and actin dysregulation in these contexts.

While roles for this type of actin remodeling are so far unexplored in vivo or in aging contexts, we reasoned that impaired InsP3R signaling during Complex I dysfunction could result in defective actin remodeling, thus explaining the larger and more hyperfused network we observe in double mutants. To observe actin remodeling, we employed a LifeAct::GFP reporter for F-actin and tested whether mitochondrial dysfunction causes similar actin remodeling in C. elegans. Indeed, we observed a roughly threefold increase in F-actin networks in Complex I/nuo-6 mutants relative to wild-type controls (Fig. 5d–g), demonstrating pronounced actin remodeling in response to mitochondrial dysfunction in C. elegans. Furthermore, the increase in F-actin formation was substantially impaired in InsP3R/itr-1 mutants, consistent with its role in the early responses to mitochondrial depolarization in vitro³¹ (Fig. 5d–g). By contrast, we found that nuo-6(RNAi) does not induce significant actin network formation (Supplementary Fig. 4l). This observation indicates that nuo-6 mutation engages an InsP3R-dependent actin remodeling program distinct from RNAi-based perturbation, reinforcing again the notion that distinct mitochondrial stress paradigms evoke mechanistically distinct responses¹⁶. Overall, these results reveal that mitochondrial damage-associated actin remodeling is conserved in C. elegans in certain contexts and promoted by InsP3R signaling.

Temporally, actin remodeling has been described across distinct acute versus prolonged phases of mitochondrial stress, with each phase involving molecular mechanisms that are partly overlapping and partly unique^31,32,58. Chronic mitochondrial dysfunction, for example, in models of Leigh syndrome, also induces actin remodeling, though the regulation and consequences in these chronic contexts are less well defined^34,58. Importantly, the factor shown to be important across all phases of mitochondrial damage-associated actin remodeling is the nucleating and branching Arp2/3 complex^31,32,58, which led us to focus on this complex as a potential mediator of InsP3R/itr-1 effects in Complex I mutants. First, we tested whether Arp2/3 is also required for damage-associated F-actin polymerization in C. elegans during Complex I/nuo-6 impairment. When we targeted Arp2/3 function via RNAi of obligatory subunit, arx-2, we found that arx-2 had little effect on basal F-actin networks, but fully suppressed the actin networks induced by Complex I/nuo-6 dysfunction, confirming a key, conserved role for Arp2/3 in forming and/or stabilizing these stress-dependent actin networks (Fig. 5h–j). We also tested whether Arp2/3 plays a role in regulating mitochondrial networks in Complex I/nuo-6 mutants by imaging mitochondrial COX-4::eGFP markers. Again, arx-2 knockdown phenocopied InsP3R mutants by demonstrating an enhanced expansion of mitochondrial networks upon arx-2 inhibition (Fig. 5k, l). Together, these results support functional conservation of the pathway mapped in mammalian cells³¹, where InsP3R signaling contributes to Arp2/3-dependent actin remodeling to regulate mitochondrial dynamics during mitochondrial dysfunction.

Next, if InsP3R/itr-1 effects on lifespan were mediated via actin remodeling, we predicted that impairing damage-associated actin remodeling downstream of InsP3R signaling would be sufficient to suppress Complex I/nuo-6 associated longevity. We fed animals arx-2 dsRNA, and while Arp2/3 impairment reduced lifespan compared to controls, this treatment completely suppressed Complex I-dependent lifespan extension (Fig. 6a). We also tested an alternative Arp2/3 subunit, arx-4, which yielded identical results (Fig. 6b). Overall, these findings demonstrate a role for the Arp2/3 complex and damage-associated actin remodeling more broadly in promoting longevity during mitochondrial stress contexts.

Fig. 6. An actomyosin regulatory network is required for lifespan extension during Complex I impairment.

a Lifespan of Complex I/nuo-6 mutants fed arx-2(RNAi). n = 97–100 worms per condition. b Lifespan of Complex I/nuo-6 mutants fed arx-4(RNAi). n = 97–100 worms per condition. c Lifespan of Complex I/nuo-6 mutants fed fhod-1(RNAi). n = 97–100 worms per condition. d Lifespan of Complex I/nuo-6 mutants fed cdc-42(RNAi). n = 67–100 worms per condition. e Lifespan of Complex I/nuo-6 mutants fed wsp-1(RNAi). n = 80–105 worms per condition. f Lifespan of Complex I/nuo-6 mutants fed mlck-1(RNAi). n = 100 worms per condition. Statistical significance of lifespan curves was determined by a log-rank Mantel–Cox test. See Supplementary Data 1 for lifespan statistics. Source data are provided as a Source Data file.

While Arp2/3 activity is centrally important to damage-associated actin responses, current models of acute damage-associated actin (ADA) and prolonged damage-associated actin (PDA) formation each involve larger regulatory networks. For example, some forms of damage-associated actin also involve formin-mediated linear filament assembly in addition to Arp2/3^31,58. Indeed, we found that RNAi targeting of the formin FHOD-1 completely suppressed Complex I/nuo-6 mutant lifespan extension (Fig. 6c). This finding suggests an important role for linear actin nucleation in our mitochondrial longevity model, potentially by promoting stable actin structures required for prolonged mitochondrial segregation or trafficking. We also sought to test factors proposed to act upstream of actin nucleators. Specifically, Cdc42 and Wiskott-Aldrich syndrome protein (WASP)-family proteins were found to activate the nucleation by Arp2/3 and formins in mitochondrial damage paradigms⁵⁸, so we tested roles for these genes in mediating mitochondrial longevity. We found that knockdown of both cdc-42 and wsp-1 substantially suppressed lifespan extension of Complex I mutants from 45 to 37.5% and 45 to 20% (Fig. 6d, e). We reason that these partial effects may represent greater redundancy among upstream regulators than observed for direct actin nucleators, Arp2/3 or FHOD-1. Finally, myosins cooperate with actin networks to perform diverse roles in mitochondrial dynamics, including actin-based scaffolding and constriction of mitochondria for fission, stabilization of damage-associated actin networks, and disassembly of “aggregated” mitochondria for delivery to autophagosomes following damage^33,58. While genetic redundancy among non-muscle myosins makes genetic dissection of specific myosin genes challenging in this system, these functions are often regulated by upstream phosphorylation of myosin regulatory light chains via myosin light chain kinase (MLCK), notably a well-established target of calmodulin/cmd-1, which we previously showed is required for Complex I/nuo-6 longevity. Additionally, some unconventional myosins associated with damage-associated actin subpopulations also use calmodulin/CMD-1 itself as a regulatory light chain⁵⁹. Indeed, we found that RNAi against MLCK-1 substantially suppressed Complex I/nuo-6 lifespan extension (Fig. 6f). Further, MLCK-1 appears to play no role in lifespan extension downstream of insulin signaling pathways (Supplementary Fig. 5), indicating specificity to mitochondrial longevity paradigms. Overall, our findings integrate emerging in vitro models with in vivo genetic evidence to support a model where InsP3R-derived Ca²⁺ signals are required for actomyosin-based scaffolding and remodeling of mitochondrial networks, which enables lifespan extension during chronic mitochondrial stress. Importantly, while our genetic data highlight the essential, collective roles of this calcium- and actin-regulatory network in vivo, explicit ordering of these mediators into a linear pathway in vivo will require additional molecular and biochemical dissection.

Restoring mitochondrial fragmentation rescues lifespan extension in InsP3R mutants

Because mitochondrial damage-induced actomyosin remodeling collectively shifts the network balance towards fragmentation by promoting DRP-1 fission⁶⁰ and by inhibiting re-fusion³², we hypothesized that restoring the ability of mitochondrial networks to fragment would rescue lifespan extension in InsP3R/itr-1 mutants. To test this, we inhibited the outer mitochondrial membrane fusion factor Mfn/fzo-1, revealing robust mitochondrial fragmentation across all genotypes (Fig. 7a, b). Furthermore, Mfn/fzo-1 impairment uniquely reverses the mitochondrial hyperexpansion and elevated COX-4 accumulation in double mutants, restoring both footprint and COX-4 levels to those observed in Complex I/nuo-6 single mutants (Fig. 7c and Supplementary Fig. 6a, b). Strikingly, Mfn/fzo-1(RNAi) also restored lifespan extension in Complex I/nuo-6; InsP3R/itr-1 double mutants (Fig. 7d, e), indicating that reversing outer membrane fusion is sufficient to bypass defects in stress adaptation in InsP3R mutants. Neither inhibition of mitochondrial fission (drp-1(RNAi)) nor disruption of inner membrane fusion (Opa1/eat-3(RNAi)) produces this rescue effect (Supplementary Fig. 6c, d).

Fig. 7. Driving mitochondrial fission rescues longevity in InsP3R mutants and autophagy impairment mimics InsP3R/itr-1 loss.

a Representative images of mitochondrial networks (COX-4::eGFP) in hypodermis of young (day 1) worms fed control or fzo-1(RNAi). Scale bar: 10 µm. b Categorization of mitochondrial morphologies in worms fed control or fzo-1(RNAi). n = 44, 44, 43, 44, 41, 40, 35, 38 worms from 2 independent trials (top to bottom); Kruskal–Wallis test with Dunn’s multiple comparisons test. c Quantification of total COX-4 (mean COX-4::eGFP intensity × mitochondrial footprint) in worms fed control or fzo-1(RNAi). n = 43, 44, 42, 43, 41, 40, 35, 38 worms from 2 independent trials (left to right); mean ± SD; two-way ANOVA with Sidak’s multiple comparisons test. d, e Lifespan of InsP3R/itr-1; Complex I/nuo-6 double mutants fed control (d) and fzo-1(RNAi) (e). n = 100 worms per condition. f Lifespan analysis of Complex I/nuo-6 mutants fed lgg-1(RNAi). n = 100 worms per condition. g Representative images of mitochondrial networks (COX-4::eGFP) in hypodermis of young (day 1) and aged (day 7) worms fed control or lgg-1(RNAi). Scale bar: 10 µm. h Quantification of total COX-4 (mean COX-4::eGFP intensity × mitochondrial footprint) in worms fed control or lgg-1(RNAi). n = 63, 50, 54, 53, 74, 66, 62, 61 worms from 4 independent trials (left to right); mean ± SD; two-way ANOVA with Tukey’s multiple comparisons test. i, j Overview of how the InsP3R regulates mitochondrial dynamics (i) and the effect of impaired Ca²⁺ signaling during Complex I impairment (j). Statistical significance of lifespan curves was determined by a log-rank Mantel–Cox test. See Supplementary Data 1 for lifespan statistics. Source data are provided as a Source Data file.

Given the diverse functional consequences of altered mitochondrial morphology, multiple mechanisms likely contribute to the lifespan rescue following Mfn/fzo-1 impairment, including potential alterations in damage segregation, calcium and lipid handling, metabolic and redox signaling, and/or mitochondrial transport²⁸. However, prior work shows that a key outcome of PDA-like actin remodeling is mitophagic turnover³², while elongated mitochondrial networks are resistant to autophagic clearance^61,62. These observations, combined with our findings that InsP3R signaling both promotes actin remodeling and constrains mitochondrial expansion, led us to hypothesize that InsP3R signaling prevents excessive accumulation of dysfunctional mitochondria by facilitating autophagy-dependent clearance. Supporting this hypothesis, we impaired autophagy via LC3/lgg-1(RNAi) and found that autophagy is indeed required for Complex I/nuo-6–mediated lifespan extension (Fig. 7f). Furthermore, directly impairing macroautophagy in Complex I/nuo-6 mutants phenocopies the shift towards mitochondrial retention (Fig. 7g, h) seen when both InsP3R/itr-1 and Arp2/3 are impaired (Figs. 4a–e and 5k, l). Notably, lgg-1(RNAi) also resulted in a unique swollen and potentially fragmented morphology with age (Fig. 7g), which is likely related to the intact F-actin remodeling processes acting upstream as well as the potential for indirect impacts from the chronic loss of bulk macroautophagy. Lastly, to determine whether InsP3R signaling acts genetically upstream of actin remodeling and autophagy during Complex I/nuo-6 longevity, we attempted to combine knockdown of core downstream mediators in this pathway (cmd-1, arx-2, and lgg-1) with InsP3R/itr-1 loss to examine gene epistasis relationships. Perhaps unsurprisingly, the combined impairment of three essential processes (mitochondrial function/nuo-6, InsP3R signaling/itr-1, and cmd-1/arx-2/lgg-1) in an animal model resulted in non-viability. However, when gene knockdowns were initiated at the L4 stage, we found that in each case, there was no additive effect on lifespan when itr-1 loss was combined with downstream mediators (Supplementary Fig. 6e). Thus while we are unable to rule out contributions during early development, the combination of our gene epistasis analysis and pathway mapping performed previously in vitro⁵⁸ together support a working model where these genes act in a common regulatory network rather than independent pathways.

Together, these results establish essential roles for InsP3R-dependent calcium signaling, cytoskeletal regulation, and mitochondrial dynamics in a Complex I-based longevity paradigm. Integrated with emerging molecular models from in vitro studies, our findings support a coherent, evolutionarily conserved regulatory network in which InsP3R signaling coordinates cytoskeletal remodeling with mitochondrial clearance to create a permissive cellular landscape for lifespan extension.

Discussion

Here, we uncovered a critical role for the InsP3R, and calcium signaling more broadly, in mediating lifespan extension during Complex I impairment. Specifically, our findings support a model in which cytosolic Ca²⁺ signals from the InsP3R coordinate mitochondrial dynamics through actomyosin remodeling, thereby supporting mitochondrial network maintenance and permissive conditions for longevity (Fig. 7i, j). Notably, we also observed significant effects of InsP3R signaling on both organismal bioenergetics (Fig. 1) and the transcriptional response to mitochondrial stress (Fig. 3), processes which also have roles in mitochondrial longevity paradigms⁹. While the molecular mechanisms underlying InsP3R-dependent effects on metabolic and transcriptional rewiring remain to be elucidated, our data suggest that these processes operate alongside cytoskeletal and organelle dynamics as part of an integrated physiological response to mitochondrial stress. These results collectively place the InsP3R as a major node in promoting mitochondrial stress adaptation and longevity. Additionally, while our study focuses on manipulating InsP3R signaling in contexts of life-extending ETC mutations, advanced age and age-associated disease states are also linked to the co-occurrence of cellular Ca²⁺ handling defects and mitochondrial dysfunction^23,63. Our findings thus suggest plausible cell biological explanations for how age-associated disruptions in both mitochondrial function and ER calcium dynamics can trigger a pathological vicious cycle. In this way, our work illuminates inter-organelle calcium dynamics and actin remodeling as processes that contribute to defining whether mitochondrial stress outcomes are beneficial or pathological.

One of the key advances of this study is the demonstration that mitochondrial stress-dependent actomyosin remodeling, which had previously been defined in vitro, plays an essential role in promoting mitochondrial homeostasis and longevity in an animal model of chronic ETC impairment. Specifically, we show that Arp2/3, formin FHOD-1, and upstream regulators CDC-42 and WSP-1 all contribute to Complex I-induced longevity, revealing a conserved calcium and cytoskeletal remodeling network with key roles in aging and lifespan. Based on pathways carefully delineated in tissue culture studies^31,32, our in vivo data suggest the existence of a coherent signaling and actin remodeling axis emanating from the InsP3R, but additional genetic tools and molecular characterization are required to precisely order these actin remodeling factors in vivo. Similarly, identifying specific myosin mediators and understanding how these actin subpopulations intersect with mitochondrial fission/fusion, mitophagy, and transport machineries in these chronic contexts remain important questions for future studies. Finally, our findings establish the regulation of mitochondrial expansion as a critical factor distinguishing adaptive stress responses from maladaptive outcomes. An important next step will be to determine how unchecked mitochondrial expansion promotes cellular dysfunction. Potential explanations include enhanced propagation of ROS, disruption of inter-organelle interactions and communication, or simply draining metabolic resources towards futile maintenance of irreparable organelles.

Our finding that the canonical InsP3R-MCU Ca²⁺ flux pathway is dispensable for Complex I-mediated lifespan extension reinforces the importance of Ca²⁺ signaling pathways operating outside the matrix in regulating mitochondrial homeostasis^25,40,64,65. Our data from mcu-1 mutant worms are consistent with the relatively minor phenotypes associated with MCU knockout mice^66,67, and support a model where Ca²⁺ signaling pathways and redox shuttles in the cytosol and intermembrane space may play larger roles in mitochondrial metabolism and homeostasis than previously anticipated^{41,58,68–71}. In our model, reduced membrane potential in Complex I mutants likely limits matrix Ca²⁺ uptake. This impaired Ca²⁺ uptake during Complex I dysfunction provides a potentially simple mechanism by which InsP3R activity may respond to mitochondrial stress, as previous work showed that mitochondrial Ca²⁺ uptake suppresses InsP3R activation by removing positive feedback by cytosolic Ca²⁺ ions²¹. In the future, it will also be useful to explore whether Ca²⁺-regulated redox pathways, such as the glycerol phosphate or malate-aspartate shuttles, are important for InsP3R-dependent control of bioenergetics and stress adaptation in long-lived mitochondrial mutants. These shuttles are activated by Ca²⁺ ions in the intermembrane space, thus not requiring MCU, and function to carry reducing equivalents from cytosolic NADH to the respiratory chain^72,73. These shuttles thus represent plausible candidates in mediating the reductions in basal respiration in InsP3R/itr-1 mutants and may also play roles in ameliorating NADH accumulation and reductive stress in Complex I mutants.

While early models suggested a unified mechanism promoting longevity during mitochondrial stress^2,3,42, mounting evidence points to mechanistic heterogeneity across different ETC perturbations. For example, a recent suppressor screen showed that virtually all compensatory responses to a mutation in Complex I acted locally within the affected complex itself⁴³, underscoring the specificity of these adaptations. Our results show that while InsP3R/itr-1 mutants tolerate Complex I knockdown, they exhibit profound developmental arrest in response to ETC inhibition at Complexes III, IV, or V, indicating differential requirements for InsP3R function when distinct Complexes are perturbed. One potential explanation for this difference is that Complex I mutation allows alternative electron entry points into the ETC, limiting the need for redox compensation, while impairments in Complexes III, IV, and V present more severe energetic bottlenecks that increase reliance on processes that regulate redox balance. Likewise, distinct modes of mitochondrial perturbation can engage parallel pathways: interventions that extend lifespan through stable ETC mutations are additive when combined with RNAi-induced ETC disruption¹⁶. Consistent with this, we show that InsP3R function is required for longevity in nuo-6 mutants but dispensable for nuo-6 knockdown, revealing a striking example of genetic divergence based on how Complex I is perturbed. Why these different modes of ETC perturbation elicit distinct responses requires more dedicated investigation, but likely involves the proteostasis-related impacts of altered stoichiometric balances of ETC subunits when expression is reduced by RNAi^16,74. An additional caveat to interpretation of RNAi results in C. elegans is also the issue of tissue-specificity, as RNAi is sometimes less penetrant in C. elegans neurons relative to other tissues. Future work to map the tissues in which nuo-6 and itr-1 perturbations are required for their effects on lifespan will be informative in this regard. Collectively, these findings reveal InsP3R signaling plays important, context-specific roles in adapting to ETC dysfunction. More broadly, these results suggest differences in how ETC perturbations impact calcium signaling and cytoskeletal remodeling could underlie the mechanistic diversity in mitochondrial longevity paradigms.

Overall, our work provides a framework for understanding how organelle-scale calcium dynamics intersect with cytoskeletal remodeling to mediate stress adaptation and lifespan regulation. Our findings reveal previously unknown roles for InsP3R-dependent calcium signaling pathways in promoting lifespan extension under chronic mitochondrial stress. These results broaden the scope by which ER calcium signaling is promoting mitochondrial homeostasis, expanding from the canonical ER-to-matrix Ca²⁺ flux model and highlighting critical roles for cytosolic signaling, cytoskeletal remodeling and autophagy-mediated turnover. Given the widespread association between impaired ER-mitochondrial communication, cytoskeletal dysfunction, and aging-related diseases, the pathways defined here offer promising therapeutic targets for promoting mitochondrial health and longevity.

Methods

C. elegans strains and husbandry

See Supplementary Data 4 for a complete list of strains used in this study and Supplementary Data 5 for a complete list of reagents. Worms were grown and maintained at 20 °C on standard nematode growth media (NGM) seeded with E. coli (OP50-1). E. coli cultures were grown from single colonies overnight in LB at 37 °C. 100 µL of liquid culture was then seeded onto NGM plates and bacterial lawns were allowed to grow for 2 days at room temperature before use.

RNAi experiments

RNAi experiments were conducted by feeding worms sequence-confirmed, dsRNA-producing clones of HT115 E. coli from the Ahringer RNAi library (Source Bioscience)—see Supplementary Data 6 for RNAi sequences. The empty vector (pL4440) was used as control. dsRNA-producing bacterial lawns were produced by growing single colonies overnight in LB cultures and seeding 100 µL onto NGM plates, with cultures and plates both containing 100 µg/mL carbenicillin to maintain selection. dsRNA was induced using 100 µL 0.1 M IPTG (supplemented with 100 µg/mL carbenicillin and 12.5 µg/mL tetracycline) at least 1 h before worms were added to plates. atp-3(RNAi) was diluted 1 in 10 with empty vector RNAi-producing cells, which was previously shown to bypass toxicity and confer lifespan extension⁴². All RNAi experiments were treated from hatch unless otherwise indicated. See Supplementary Fig. 7 for validation of RNAi knockdown.

Developmental assay

Worms were synchronized by timed egg lay at 20 °C on NGM plates supplemented with carbenicillin and then ~50 eggs per condition were transferred to activated RNAi plates (time = 0). Worms were inspected and scored for 10 days following egg lay, and vulval development was used to distinguish between worms arrested at late larval stages and sterile adults. Unhatched eggs and desiccated worms were censored. Four independent replicates were performed.

InsP3 sponge construction

To generate an InsP3 sponge system, we adapted the design of ref. ³⁷ with minor modifications. To create plasmid pGF1 (eft-3p::supersponge::AID::mKate2::unc54 3'UTR), we digested pDB6 using XmaI/NheI to create a backbone with 5′ eft-3p and 3′ mKate2 sequences. The InsP3 sponge was generated by cloning the sequence representing amino acids M254 to K670 from itr-1d cDNA using forward primer: GACCCGGGATGTTTTTGCTTTTTGATG and reverse primer: CTGGATCTTTAGGCTTTGGATTATTGTGC. cDNA was derived from itr-1(sy290) allele, which contains the mutated InsP3 binding site that doubles affinity for InsP3³⁷. Auxin-inducible degron (AID) sequence was amplified from pLZ29 (eft-3p::AID::EmGFP) using forward primer: ACAATAATCCAAAGCCTAAAGATCCAGCC and reverse primer: TTAGCTAGCACTCTTCACGAACGCC. The InsP3 binding domain was fused with AID via PCR using the IP3 binding domain forward primer (containing XmaI site) and AID reverse primer (containing NheI site). The resultant InsP3 supersponge::AID sequence was then ligated into the backbone. This plasmid was microinjected into N2 animals to produce bugEx7[eft-3p::IP3::supersponge::AID::mKate2] and subsequently gamma irradiated and outcrossed 3× to produce integrated strain BUZ117.

Complex I activity assay

50,000 worms per condition were synchronized for collection at day 1 of adulthood, and mitochondria were isolated in line with a previously described protocol⁴³. Briefly, worm pellets were washed in M9 buffer then resuspended in mitochondrial isolation buffer (mannitol, 210 mM; sucrose, 70 mM; 0.5 M EDTA pH 8.0, 1 mM; 1 M Tris HCl pH 7.4, 5 mM; 1 tablet protease inhibitor cocktail per 100 mL; prepared in ddH2O) and kept on ice. To isolate mitochondria, worm pellets were homogenized on ice with ~200 gentle strokes using a tissue grinder (Kimble 885302-0007). Subsequent centrifugation steps, 200 × g for 5 min at 4 °C and 800 × g for 10 min at 4 °C, were performed to remove worm debris and nuclei, respectively. A final spin at 12,000 × g for 30 min at 4 °C was performed to pellet the isolated mitochondrial fraction. After the isolated mitochondrial fraction was collected, protein concentration was measured using a Pierce BCA protein assay kit. To measure Complex I activity, the isolated mitochondrial fraction was prepared following kit instructions (Novus Biologicals NBP3-25843) and the NADH oxidation rate was measured at 340 nm using a BioTek Synergy HT multi-mode microplate reader. Complex I activity was normalized to protein level and calculated according to kit instructions.

Confocal fluorescence microscopy

Synchronized worms at indicated ages were picked onto a 10% agarose pad with ~3 µL 0.1 µm Polybead microsphere suspension (Polyscience) for immobilization and then a cover slip (VWR, #1.5) was added on top to mount worms⁷⁵. Confocal microscopy was performed on a Nikon Eclipse Ti2 with CSU-W1 spinning disk and Plan-Apochromat 100×/1.49 objective. Mitochondrial targeted GCaMP7b, COX-4::eGFP, LifeAct::GFP, and GFP::DRP-1 were imaged by 488 nm laser excitation and ET525/36 m emission filter. Mitochondrial targeted mKate2 and TMRE were imaged with 561 nm laser excitation and ET605/52 m emission filter. DIC images were also taken alongside the fluorescent images. Nikon NIS-Elements Advanced Research (NIS-AR) was used for image processing and analysis.

Calcium imaging

Images of mitochondrial targeted GCaMP7b and mKate2 were taken in the anterior-most intestinal cells. Mitochondrial targeted GCaMP7b was imaged with 2× integration to increase signal strength. Mitochondrial networks were identified and masked via setting threshold of mKate2 signal after rolling ball background subtraction. Samples were blinded and thresholding parameters were manually adjusted for every image. Regions of interest (ROIs) were manually drawn to encompass a single anterior intestine cell. Subsequently, the mean intensity of GCaMP7b and mKate2 were measured in the entire mitochondrial network of single intestinal cells, and then the ratio of GCaMP7b:mKate2 signal was calculated.

Mitochondrial imaging

Endogenously tagged COX-4::eGFP was imaged in the anterior intestine and hypodermis of adult worms at indicated ages (typically day 1 and day 7). Deconvolution and denoising were applied to get sharper images. ROIs were manually drawn to include a single intestine cell in the anterior-most intestine or the hypodermis. In aged intestine images, bright spot detection (Nikon Elements) was used to exclude auto-fluorescent gut granules. Binary masks representing mitochondrial networks were generated by setting fluorescence intensity thresholds. Because overall fluorescence intensity varies across age and genotype, thresholding parameters were manually adjusted for every image while blinded. Using these binary masks, the mean intensity and area of mitochondria was measured. Mitochondrial footprint (mitochondrial area divided by ROI area) and total COX-4 (footprint multiplied by mean intensity) were calculated. To quantify mitochondrial networks based on their overall morphology, blinded image files were used to manually classify networks into one of four categories: fragmented, intermediate, fused, or hyperfused. Neurons, nuclei, and seam cells were omitted from ROIs.

Actin network imaging

LifeAct::GFP was used to visualize filamentous actin in the hypodermis of mid-L4 worms rather than day 1 adults, as the promoter activation declines after larval development. Worms were synchronized to mid-L4 by identifying the “Christmas tree” stage of vulva development. All images were taken using 2× integration in the hypodermis just anterior to the vulva in a plane with visible nuclei. To enhance visualization of the actin networks, all images were denoised and sharpened (Nikon Elements). All files were blinded prior to analysis. ROIs were manually drawn to include the hypodermis while omitting seam cells and nuclei. Binary masks were generated for actin filament networks by manually setting intensity-based threshold parameters. Actin footprint (total actin area divided by ROI area) was calculated. Additionally, blinded image files were also used to manually classify F-actin networks into one of three categories: dense, intermediate, or minimal.

DRP-1 imaging, processing, and analysis

Multiple technical challenges arose in establishing lines for the study of DRP-1 interactions with mitochondria. First, animals homozygous for GFP-labeled DRP-1 exhibit grossly aberrant mitochondrial morphology, as suggested also in tissue culture studies⁵⁷. However, animals heterozygous for DRP-1::GFP with one wild-type copy exhibit mitochondrial morphology similar to controls (Supplementary Fig. 4c), leading us to adopt this heterozygous approach across all backgrounds. Secondly, we were unable to generate the 4-locus strain (itr-1; nuo-6; GFP::drp-1; mito-mKate2) ideally needed to image DRP-1-mitochondrial interactions. We therefore used RNAi to knockdown nuo-6 in itr-1; drp-1::GFP(+/−); mito-mKate2 animals. Egg lays were performed on activated RNAi plates (empty vector vs. nuo-6). At L4 stage, worms were sorted under a fluorescence stereoscope to include only those heterozygous for the GFP::drp-1 transgene. These worms were maintained on the corresponding RNAi lawns. At days 1 and 7, 20–30 worms for each condition were removed and mounted for imaging as described above. All images were taken around the middle portion of the anterior hypodermis. Because of fluorophore crosstalk, spectral unmixing was performed (Nikon Elements AR) with spectra for “eGFP” and “mTangerine” providing the cleanest separation of signal. Unmixed images were then subjected to automated 2-D deconvolution (green ex/em 488/535; red ex/em 561/620) and denoising (Denoise.ai module). Rolling ball background removal was performed on each channel with identical settings for all images in the experimental replicate. ROIs were drawn around hypodermis cell hyp7, excluding seam cells, neurons, and unfocused regions. Mitochondria and DRP-1 puncta were segmented via thresholding and bright spot detection, respectively. Within each ROI, the total number of DRP-1 puncta was counted and normalized to the total area of segmented mitochondria. Mean DRP-1 puncta intensity was measured collectively across all puncta within the ROI.

Respiration analysis

OCR measurements were conducted on synchronized day 1 adults using Agilent Seahorse XFe96 flux analyzer and Wave desktop software according to the manufacturer’s instructions and adapted for C. elegans⁷⁶. The heater of XFe96 flux analyzer was turned off before use and experiments were performed at 20–25 °C. The XFe96 sensor cartridge was hydrated with sterile water overnight at 37 °C without CO₂ and then hydrated with Seahorse XFe96 calibrant for a few hours before use. Worms were washed off plates and washed three times with M9 buffer. ~20 worms were loaded to each well of cell culture microplates (Agilent Seahorse XF96) except the wells set for background correction. After calibration and equilibration, basal OCR was measured for 5 loops and each loop contains mixing for 2 min, waiting for 0.5 min and measuring for 2 min. 10 µM FCCP (final concentration) was injected to induce maximal oxygen consumption rate for 9 loops. 40 mM sodium azide (final concentration) was injected to inhibit mitochondrial respiration and non-mitochondrial OCR was measured for 4 loops. The precise worm numbers for each well were counted after experiments and raw OCR readings were divided by worm numbers to calculate the OCR per worm (pmol/minute per worm). The well was excluded if its raw OCR was over 600 due to high worm number. To diminish the impact of transient noise arising from worm movements and allow time for FCCP to fully equilibrate in the animals, basal OCR was calculated by averaging the OCR across measurement loops 2–5 and the maximal OCR across loops 11–14. Each condition had at least 6 wells for technical replicates in one experiment and at least three independent experiments were performed.

Lifespan analysis

All lifespan experiments were performed at 20 °C. Timed egg lays or hypochlorite bleaching were used to obtain synchronized populations. One day before adulthood (day 0), at least 100 mid-L4 worms were transferred to 5 or 10 fresh plates at a density of 10 or 20 worms per plate. To separate from progeny and avoid starvation, worms were transferred to fresh plates every day or every other day until the first deaths (~10–12 days). Survival was scored every 1–2 days and worms were marked dead when they failed to respond to three taps on the head, tail, and middle body part. Worms were censored when they crawled off the plate and desiccated, lost vulval integrity during reproduction, or when embryos hatched internally. Log-rank Mantel–Cox tests were performed to determine statistical significance between two conditions.

Lifespan analysis on solid dietary restriction (DR) plates

Solid dietary restriction assays were performed as previously described in ref. ⁷⁷ with the following modifications. Plates were prepared with a bacterial concentration of 10¹¹ cfu/mL for ad libitum (AL) and 10⁸ cfu/mL for DR. After diluting bacteria to the correct concentrations in LB, the diluted bacterial stocks were put in a shaking incubator at 37 °C for 1 h. After shaking for 1 h, 50 mg/mL kanamycin and 100 mg/mL carbenicillin were diluted and mixed into the diluted bacterial stocks at 1:11 ratio prior to seeding plates. All plates were prepared in advance and stored at 4 °C before use. Worms were synchronized with timed egg lays on standard NGM plates with OP50-1 grown using standard techniques. For each condition, 150 day 1 adults were transferred to 15 AL or DR plates at 10 worms per plate.

RNA-seq analysis

Experiments were performed with three biological replicates. About 1000 worms were synchronized by hypochlorite bleaching and grown to L4 on NGM plates seeded with HT115 E. coli. Worms were collected and washed three times with M9 buffer to remove bacteria. After the final wash, pelleted worms were resuspended in Qiazol, snap frozen in liquid nitrogen, and stored at −80 °C until RNA extraction. For RNA extraction, worm samples were disrupted in three freeze thaw cycles with Qiazol between 37 °C water bath and snap freezing in liquid nitrogen. Then samples were incubated and vortexed with chloroform at room temperature to extract RNA. RNA was purified through Qiagen RNeasy mini kit per manufacturer instructions and treated with Qiagen RNase-Free DNase on the column to get rid of DNA contamination.

RNA quality analysis, cDNA library, and Next-Generation sequencing were performed by VUMC Vanderbilt Technologies for Advanced Genomics (VANTAGE). RNA quality was analyzed (BioAnalyzer) with all samples having an RNA integrity number of 10. cDNA library was constructed with stranded mRNA (polyA-selected) library preparation kit (NEB). Paired-end 150 bp sequencing was performed on the Illumina NovaSeq6000, targeting 50 M reads per sample.

Analysis was performed through the Illumina Dragen pipeline. Quality control analysis via MultiQC indicated high-quality reads with mapping rates >96% over >75 million total reads per replicate. Alignment to the genome (UCSC ce10) was performed via STAR⁷⁸; gene counts and normalization were generated through Salmon (https://combine-lab.github.io/salmon/); and the DESeq2 algorithm was used to identify DE genes using a Wald test with FDR correction (https://bioconductor.org/packages/release/bioc/html/DESeq2.html). The list of DE genes from each genotype relative to wild-type were then trimmed to remove genes with 0 reads or status = LOW (< 10 transcripts per kilobase million).

To process for enrichment analysis, cutoffs for DE genes were set as >1.5-fold change and P_adj < 0.01, and these subsets of genes were contrasted across genotypes. WormCat 2.0⁴⁸ (http://www.wormcat.com/) was used to analyze gene set enrichment using annotations from the whole genome (v2). Terms from WormCat Category 2 annotations were used to identify enriched cellular processes in the Venn diagram and Category 3 annotations defined subcategories used in focused heatmaps.

qRT-PCR

For wild-type and itr-1(sa73) comparisons, synchronized worm populations were grown on RNAi from hatch and collected at L2 prior to larval arrest. To validate RNAi knockdown in wild-type worms alone, synchronized worm populations were grown on RNAi from hatch and collected as day 1 adults. For RNA extraction, worms were resuspended in Qiazol then treated to three freeze-thaw cycles alternating between 37 °C and liquid nitrogen. Samples were then treated with chloroform and the RNA-containing phase was collected and purified using a Qiagen RNeasy mini kit. cDNA was generated using the SuperScript VILO cDNA synthesis kit per manufacturer’s instructions. TaqMan qRT-PCR was run on a Bio-Rad CFX96 Thermal Cycler. Fold change was calculated following the 2^−ΔΔCT method. Fold change values were normalized to the mean of the control group, and statistical analyses were performed on normalized values. The following TaqMan gene expression assays were used: iscu-1 (Ce02467253_g1) as the internal control, cco-1/cox-5b (Ce02412619_g1), atp-3 (Ce02417729_g1), arx-2 (Ce02480573_m1), cmd-1 (Ce02471037_g1), fzo-1 (Ce02433118_g1), lgg-1 (Ce02433594_g1), and mlck-1 (Ce02499590_m1).

TMRE staining and imaging

500 µL of a 0.1 mM TMRE solution (0.1 M TMRE stock in DMSO diluted 1:1000 in M9) was added to an NGM plate pre-seeded with OP50-1. Plates were allowed to dry for ~8 h, then L4 worms were picked onto TMRE plates and fed overnight. After 16 h, worms were picked to standard NGM plates with OP50-1 for 1 h to remove excessive dye before imaging. TMRE solution, plates, and treated worms were protected from light before imaging. TMRE images were taken in the hypodermis of day 1 adults. Deconvolution and rolling ball background subtraction were applied to get sharper images; bright spot detection was used to exclude TMRE aggregates. Prior to analysis, all image files were blinded. Mitochondria were segmented by manually setting the threshold. ROIs were manually drawn to include the hypodermis and within the ROI the mean intensity of the segmented mitochondria was measured.

Western blotting

Synchronized young adults were collected, washed three times with M9 buffer and snap frozen in liquid nitrogen. Three independent biological replicates were collected. Identical volumes of RIPA buffer (50 mM Tris, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate, 0.1% SDS, pH 7.5) supplemented with protease inhibitors and phosphatase inhibitors were added to worm samples and then samples were lysed by sonication. Lysates were centrifuged at 16,000 × g for 15 min at 4 °C and supernatant protein concentration was measured using a Pierce BCA protein assay kit. An equal amount of protein was heated at 95 °C for 15 min with 4× Laemmli Sample Buffer added. Samples containing 40–60 µg protein were loaded to a gel and once separated, proteins were wetly transferred to a PVDF transfer membrane. The PVDF membrane was blocked with 5% milk in TBST or 5% BSA-TBS. Following incubation with the primary antibody (for p-PDH: anti-p-PDH (1:1000), anti-beta-actin (1:2000); for total PDH: anti-PDH (1:1000), anti-tubulin (1:5000); for CMD-1: anti-calmodulin (1:5000), anti-tubulin (1:5000)) the PVDF membrane was washed three times for 10 min each with TBST and then incubated with the secondary antibody (for p-PDH: HRP-conjugated anti-rabbit (1:2000), HRP-conjugated anti-mouse (1:10,000); for total PDH: IRDye 800 anti-rabbit (1:10,000), IRDye 680 anti-mouse (1:10,000); for CMD-1: IRDye 800 anti-rabbit (1:10,000), IRDye 680 anti-mouse (1:10,000)). To detect HRP-conjugated antibodies, signals were developed using Pierce ECL Western Blotting Substrate, imaged with an Amersham Imager 600, and bands were quantified using ImageJ. Fluorescently labeled antibodies were imaged with an Odyssey Imaging System and bands were quantified using Image Studio.

Transmission electron microscopy (TEM)

Worms were transferred to a conical with 5 mL of M9 buffer and spun down at 0.6 × g three times to wash bacteria off. Excess supernatant was removed. A pellet of live C. elegans was resuspended in 0.15 M sucrose prepared in M9 buffer and loaded into 200 µm deep well of a A-type carrier (Leica). The assembly was covered with flat side of B-type carrier (Leica) and vitrified using high-pressure freezing machine (Leica EM ICE). The frozen specimens were stored in liquid nitrogen until further processing. The freeze-substitution (FS) process was adapted from ref. ⁷⁹. FS was performed using a cocktail of 0.5% glutaraldehyde and 0.1% tannic acid in acetone. Samples were transferred into automatic freeze-substitution machine (AFS2, Leica) pre-cooled to −140 °C and warmed to −90 °C over a period of 3 h, where they were then kept for 114 h. At the end of a hold period, samples were washed 4 times for 30 min each with acetone. The final acetone wash was replaced with 2% osmium tetroxide in acetone. Following this, samples were warmed to −20 °C over a period of 12 h after which temperature was maintained at −20 °C for an additional 10 h before further warming to 0 °C over a period of 4 h. Samples were then transferred into an ice filled bath and washed 4 times for 30 min each with acetone and infiltrated with Spurr’s resin (Electron Microscopy Sciences). During final steps of resin infiltration, individual worms were transferred from carriers into resin molds with a fine needle and samples were cured in an oven at 60 °C for 72 h. Thin sections (70 nm) were cut using a Leica UC7 and a Diatome diamond knife. Thin sections were positioned on grids and stained with uranyl acetate and Reynold’s lead citrate. These sections were then imaged on a Tecani T12 TEM operating at 100 keV using an AMT CMOS camera. Mitochondrial area and footprint were quantified using Amira software. A mask of the mitochondria and hypodermis were manually drawn using the freehand masking tool in Amira, then area was measured.

Quantification and statistical analysis

GraphPad Prism (versions 9 & 10) was used for statistical analysis and plotting graphs. Comparisons between two groups were determined by two-tailed unpaired student’s t-test. For multiple comparisons, one- or two-way ANOVAs were performed with post-hoc tests. Two-way ANOVAs were used to compare multiple genotypes between young and old ages. Data for mitochondrion size was distributed on a log normal scale. Confocal and TEM image analyses were performed on blinded files. Statistical tests, sample size, and error bars are indicated in figure legends.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

K.B. and G.F. conceived the study; G.F., E.M.R., A.G.M., E.K.F.D., A.L.H., B.J.C., L.P., and K.B. designed and executed experiments and data analysis; K.B. pursued funding; all authors contributed to manuscript drafting and revision.

Peer review

Peer review information

Nature Communications thanks Nektarios Tavernarakis and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Funding

This work was supported by the Glenn Foundation for Medical Research/American Federation for Aging Research and NIH/NIA R00AG052666 (KB), R01AG073354 (KB).

Data availability

Complete raw RNA sequencing data are available through GEO: GSE297429. Raw image files were not publicly deposited due to size restrictions and file complexity, but are available upon request from the corresponding author. Worm strains generated in this study will be deposited to the Caenorhabditis Genetics Center for anticipated highly used strains or made available upon request. Further information and requests for data, resources, and reagents should be directed to and will be fulfilled by the lead contact, Kristopher Burkewitz (kristopher.burkewitz@vanderbilt.edu). Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-76514-3.