DAF‐2/insulin IGF‐1 receptor regulates motility during aging by integrating opposite signaling from muscle and neuronal tissues

Abstract During aging, preservation of locomotion is generally considered an indicator of sustained good health, in elderlies and in animal models. In Caenorhabditis elegans, mutants of the insulin‐IGF‐1 receptor DAF2/IIRc represent a paradigm of healthy aging, as their increased lifespan is accompanied by a delay in age‐related loss of motility. Here, we investigated the DAF‐2/IIRc‐dependent relationship between longevity and motility using an auxin‐inducible degron to trigger tissue‐specific degradation of endogenous DAF‐2/IIRc. As previously reported, inactivation of DAF‐2/IIRc in neurons or intestine was sufficient to extend the lifespan of worms, whereas depletion in epidermis, germline, or muscle was not. However, neither intestinal nor neuronal depletion of DAF‐2/IIRc prevented the age‐related loss of motility. In 1‐day‐old adults, DAF‐2/IIRc depletion in neurons reduced motility in a DAF‐16/FOXO dependent manner, while muscle depletion had no effect. By contrast, DAF‐2 depletion in the muscle of middle‐age animals improved their motility independently of DAF‐16/FOXO but required UNC‐120/SRF. Yet, neuronal or muscle DAF‐2/IIRc depletion both preserved the mitochondria network in aging muscle. Overall, these results show that the motility pattern of daf‐2 mutants is determined by the sequential and opposing impact of neurons and muscle tissues and can be dissociated from the regulation of the lifespan. This work also provides the characterization of a versatile tool to analyze the tissue‐specific contribution of insulin‐like signaling in integrated phenotypes at the whole organism level.

Decline in physical performance is a universal feature of aging.
The motility phenotype of daf-2 mutants has been controversial in the past, as they showed reduced motility in the presence of food and/or in the absence of stimulation due to a food-seeking defect (Churgin et al., 2017;Hahm et al., 2015;Hsu et al., 2009;Huang et al., 2004;Podshivalova et al., 2017). However, measurement of motility of daf-2 mutant in the absence of food, either on plate or in liquid, has revealed that it is increased in middle-age compared to wild-type worms (Bansal et al., 2015;Hahm et al., 2015;Mulcahy et al., 2013). Thus, daf-2 mutants are also considered to be healthier in later life, which may be due to preserved motoneurons function (Liu et al., 2013).
To date, very little is known about the physiological expression pattern of DAF-2/IIRc. An immunolabelling approach has shown that DAF-2/IIRc is expressed mainly in the nervous system and in a pair of cells called "XXX cells" in the head and also in the epidermis (Kimura et al., 2011). In addition, two studies analyzed the requirement of DAF-2/IIRc or DAF-16/FOXO in different tissues for lifespan regulation using transgene rescue of daf-2 or daf-2; daf-16 mutants (Libina et al., 2003;Wolkow et al., 2000). The first study concluded that DAF-2 acts primarily in the nervous system, consistent with the expression profile of DAF-2/IIRc, while DAF-16/FOXO was later shown to act principally in the gut. This was counter-intuitive because these two proteins function in the same signaling pathway. The paradox was resolved by proposing that intestinal DAF-16/FOXO may trigger a secondary signal from the gut to induce inhibition of DAF-2/ IIRc in distant neuronal tissues (Libina et al., 2003). However, these studies were based on standard transgenic strategies, the only tools available at the time, which lack some of the regulatory elements present in the endogenous loci and also lead to the overexpression of proteins encoded by the transgenes. New technologies have been developed to manipulate the endogenous expression of specific proteins, providing a way to re-examine the tissue-specific activities of DAF2/IIRc, as we still do not know in which tissue wild-type DAF-2/IIRc functions for the regulation of the different phenotypes observed in daf-2 mutants.
In this work, we investigated when and where DAF-2/IIRc is required to maintain worm motility during adulthood and how this phenotype relates to the dauer, lifespan and oxidative stress resistance phenotypes. In order to deplete DAF-2/IIRc protein in a spatially and temporally controlled manner, we generated alleles for conditional degradation by inserting an auxin-inducible degron (AID) (Zhang et al., 2015) and a fluorescent tag into the daf-2 locus and constructed several independent strains to induce DAF-2/IIRc degradation in all cells or in neurons, muscle, intestine, germline or hypodermis. Previous studies used similar tools (Venz et al., 2021;Zhang et al., 2021) or tissue-specific CRISPR and RNAi approaches (Uno et al., 2021) to address the role of DAF-2/IIRc activity in the control of lifespan, dauer and oxidative stress resistance. Our work further explored the involvement of combinations of tissues in these phenotypes and more specifically investigated the regulation of motility, whose maintenance is commonly regarded as a characteristic of healthy aging.
Our results showed that DAF-2/IIRc is ubiquitously expressed in worms and can be efficiently degraded by the AID system.
Degradation of DAF-2/IIRc in all tissues, from adulthood onwards, reproduced the lifespan and motility phenotypes of the reference daf-2(e1370) allele and the constitutive dauer phenotype in the progeny. Depletion of DAF-2/IIRc in neurons or in the gut was sufficient to extend the lifespan of the worms and only intestinal inactivation of DAF-2/IIRc reproducibly increased resistance to oxidative stress.
Still, neither neuronal nor intestinal inactivation improved worm motility in adulthood. Neuronal depletion of DAF-2/IIRc unexpectedly downregulated motility from early adulthood, in a DAF-16/FOXO dependent manner. In contrast, muscle depletion was sufficient to improve motility in middle-aged worms, without affecting lifespan or resistance to oxidative stress. Finally, muscle inactivation of DAF-2/ IIRc induced nuclear accumulation of DAF-16/FOXO in cells but did not require its activity for the regulation of motility which relies on the transcription factor UNC-120/SRF.

| RE SULTS
2.1 | The AID::mNeonGreen-tagged DAF-2 protein is functional and efficiently downregulated in the presence of auxin The mNeonGreen (mNG) and degron sequences were added to the 3′ end (before the STOP codon) of the endogenous daf-2 locus (daf-2::AID::mNG or kr462) so that DAF-2 degradation could be monitored via the loss of mNG fluorescence. DAF-2::AID::mNG was detected in head neurons, XXX cells and epidermis as reported earlier (Kimura et al., 2011) but also in the majority of the worm tissues, from the two-cell stage embryo (Figure 1a). daf-2(kr462) worms were crossed with two independent lines that expressed in all tissues the plant ubiquitin ligase substrate recognition subunit, TIR1, an essential component of the AID system (see Table S1 for strain description). We then verified that neither the tag nor the presence of the TIR1 transgene interfered with DAF-2 function. In the absence of auxin, daf-2(kr462) and daf-2(kr462); Pubiquitous::TIR1 worms exhibited the same lifespan (Figure 1b,c; Table S2) and motility (Figure 1d,e) as wild-type worms. Furthermore, they did not enter the dauer stage at any temperature in the presence of food (Table S3), in contrast to the heat-sensitive reference daf-2(e1370) mutants that showed a fully penetrant dauer constitutive phenotype. Thus, the addition of the degron and mNG sequences to the daf-2 locus or TIR1 ubiquitous expression did not seem to impair DAF-2 function.
In the presence of auxin, the fluorescence signal was strongly downregulated, confirming the efficiency of auxin-induced DAF-2 degradation ( Figure 2 and Figure S1a). We then assessed the dauer, lifespan and motility phenotypes of daf-2(kr462) worms expressing TIR1 in all tissues compared to daf-2(e1370) mutants. All transgenic worms placed on auxin from hatching entered the dauer stage at 15, 20 or 25°C, (Table S3 and Venz et al., 2021) mimicking the fully penetrant dauer phenotype of daf-2(e1370) mutants raised at the restrictive temperature of 25°C. Furthermore, when daf-2(kr462) worms were placed on auxin plates at a later developmental stage (L4) to bypass the dauer arrest, their lifespan was doubled, as daf-2(e1370) worms at 20°C (Figure 1c and Table S2). Remarkably, the ubiquitous degradation of DAF-2 also recapitulated the agedependent motility phenotype of daf-2(e1370) mutants. Indeed, 1-day-old and 13-day-old daf-2(kr462) worms expressing TIR1 in all tissues showed a lower and higher frequency of body bends, F I G U R E 1 Expression pattern of DAF-2::AID::mNG and functional validation of its degradation by the auxin-inducible system. (a) Image of DAF-2::AID::mNG in 1-day-old daf-2(kr462) adult. Scale bar: 100 μm. (b) Survival curves of control (N2) and daf-2(kr462) animals (N = 2, n = 141 and 138 for N2 and daf-2(kr462), respectively). (c) Survival curves of N2, daf-2(e1370) mutants, and worms with ubiquitous depletion of DAF-2. Data have been pooled from two independent experiments (n = 155-160 for each genotype) in which two different Peft-3::TIR1 containing strains were tested. See Table S1 for strain description and Table S2 for detailed lifespan data, replicates, and statistics. (d, e) Body bends frequency at day 1 (d) and day 13 (e) of adulthood of N2, daf-2(e1370), or daf-2(kr462) worms, with or without ubiquitous expression of TIR1. The number of animals scored is indicated in each bar and corresponds to the pool of two experiments (see Figure 4 for more replicates). Bars indicate median values, means are represented by black horizontal lines, and brackets show standard deviations, ns: non-significant, ***: p < 0.001, Kruskal-Wallis and Dunn's post hoc test with FDR method for adjusting p-value. All experiments were performed at 20°C respectively, compared to control animals of the same age ( Figure 1d,e). Overall, these data demonstrate that the downregulation of DAF-2::AID::mNG correlates with a significant reduction of DAF-2 function, validating our experimental approach.
This suggests that some DAF-2 protein persists, consistent with the low level of fluorescence still detected in the presence of auxin ( Figure 2). Nevertheless, our results show that inactivation of DAF-2 during development is not a prerequisite for lifespan extension, in agreement with previous results (Dillin et al., 2002;Venz et al., 2021) and that down-regulation of DAF-2 in adulthood is also sufficient to recapitulate the effect of constitutive downregulation of DAF-2 on worm motility during aging.

| Intestinal and neuronal DAF-2 activities cooperate to regulate lifespan but are differently required for the resistance of worms to oxidative stress
We first examined the tissue-specific contribution of DAF-2 inactivation to the regulation of dauer and lifespan. Transgenic lines were generated to express TIR1 in muscle (Pmyo-3), hypodermis (Pdpy-17), neurons (Prab-3), gut (Pges-1), or germline (Psun-1). In addition to the previously described TIR1 transgenes, we generated new transgenes in order to test two independent lines for each tissue and thus limit potential confounding effect of the genetic background (see Experimental procedures and Table S1). In the presence of auxin, these TIR1 transgenes allowed efficient degradation of the DAF-2::AID::mNG protein in individual tissues, as indicated F I G U R E 2 DAF-2::AID::mNG is effectively downregulated in the presence of ubiquitously expressed TIR1 after auxin treatment. (a-g) Images of DAF-2::AID::mNG in 1-day-old daf-2(kr462) adults expressing ubiquitous TIR1 and grown in the absence of auxin (upper panels) or after 24 h of auxin treatment (lower panels). Images focus on specific body regions: the head (a), showing strong expression in the nerve ring (NR) and the XXX cells; the neuronal cell bodies of the ventral nerve cord (VNC) (b); the proliferating germ cells (c); the embryos (d); the epidermal syncytium (e); the intestine (f) and the body wall muscles (g). For the intestine (f), images were taken in apb-3(ok429) mutant background in order to reduce unspecific intestinal autofluorescence (arrows indicate the specific DAF-2::AID::mNG associated signal). In all images, the remaining staining of the gut after auxin treatment corresponds to nonspecific autofluorescence that varies between animals. Similar results were obtained in 7-day-old animals (data not shown). Scale bars: 20 μm by the loss of fluorescent signals in young and middle-aged adults after 24 h (Figure S1b-h) and by DAF-16 nuclear accumulation (see below, result Section 4). Dauer formation could not be achieved after degradation of DAF-2 in one given tissue (  (Libina et al., 2003). However, recent work on the tissue-specific activities of DAF-16 in a wild-type context argues that DAF-16 is required in several tissues to control the dauer phenotype (Aghayeva et al., 2021), in agreement with our results with DAF-2.
Inactivation of DAF-2 in muscle, hypodermis, or germline did not reproducibly affect lifespan suggesting that depletion of DAF-2 in these tissues may not be sufficient to impact this phenotype (Figure 3a-f; Figure S2d-g and Table S2). However, we cannot completely rule out that a small amount of undegraded DAF-2 remained in these tissues.
Degradation of DAF-2 in neurons or intestine was sufficient to

| DAF-2 neuronal signaling is required for worm motility in early adulthood, while DAF-2 muscle signaling impairs motility from midadulthood
In order to further characterize the fitness of long-lived worms, we measured their body-bend frequency in liquid medium (BBF), as a proxy of physical performance (Duhon & Johnson, 1995;Laranjeiro et al., 2019). The motility of worms with DAF-2 inactivation in the intestine was similar to that of control worms on days 1 and 13 of adulthood ( Figure 4a-c). Thus, the signaling from the intestine upon DAF-2 degradation is sufficient to prolong lifespan, but does not markedly affect the function of the neuromuscular system with age.
Neuronal inactivation of DAF-2 significantly reduced the worm's BBF of 1-day-old animals and mirrored the motility phenotype of animals with whole-body degradation of DAF-2 ( Figure 4a). This reduction was unexpected as Liu et al. reported an increased neurotransmission at the neuromuscular junction (NMJ) of daf-2(e1370) mutants, although for later age (Liu et al., 2013). We thus assessed cholinergic neurotransmission in our strains by treating worms with aldicarb. Aldicarb is an inhibitor of acetylcholinesterase, which induces worm paralysis due to the accumulation of acetylcholine in the synaptic cleft. Consistent with Liu et al. data, we observed an increase in cholinergic neurotransmission, indicated by an accelerated paralysis in response to aldicarb, when DAF-2 was ubiquitously inactivated in middle-aged animals but also in 1-day-old animals ( Figure S3). However, worms with neuronal depletion of DAF-2 behaved on aldicarb like control worms ( Figure S3).
The regulation of worm motility relies on a complex neuronal network that involves different class of interneurons, excitatory cholinergic and inhibitory GABAergic motoneurons (Zhen & Samuel, 2015). Previous report showed that DAF-2 is expressed in both cholinergic and GABAergic neurons (Taylor et al., 2021).
Depletion of DAF-2 in either cholinergic or GABAergic neurons only was sufficient to impede worm's motility in 1-day-old animals, thus suggesting that DAF-2 functions in both types of neurons to control motility in young adults ( Figure 4b).
In contrast to neuronal DAF-2, muscle inactivation of DAF-2 did not affect the BBF of 1-day-old adult animals but was sufficient to recapitulate the higher BBF of 13-day-old adult animals with ubiq- Taken together, these data support a critical role for wild-type DAF-2 activity in muscle in negatively regulating motility in middle age, while neuronal DAF-2 promotes motility in early adulthood.

| DAF-16 nuclear accumulation upon tissuespecific inactivation of DAF-2
Early studies on the tissue-specific activities of DAF-2 and DAF-16 have suggested that down-regulation of the DAF-2 signaling pathway in one tissue induces its inhibition in distant tissues (Libina et al., 2003;Wolkow et al., 2000). However, these results were obtained in a sen-  accumulation of DAF-16::GFP in distant tissues (Uno et al., 2021).
To avoid potential issues associated with reporter transgenes such as overexpression, we endogenously tagged DAF-16/FOXO with wrmSCARLET using CRISPR/Cas9 mediated genome engineering.
DAF-16::wrmSCARLET was detected in all somatic tissues and the germline during adulthood, with highest expression in neurons, and localized mainly in the cytoplasm of all cells (Figure 5a and Table S4).
Ubiquitous depletion of DAF-2 from the L4 stage induced nuclear ac- Overall, while DAF-16 accumulated in the nuclei of neurons or muscles after inactivation of DAF-2 in the same tissue, it is essential for the regulation of motility in neurons but not in muscles in which DA-F2 required UNC-120 for motility regulation.

| Both neuronal and muscular DAF-2 depletion prevents muscle mitochondria fragmentation with age
We  Figure S4). Thus, inactivation of DAF-2 in muscles or neurons is sufficient to prevent muscle mitochondrial fragmentation, suggesting that DAF-2 acts both autonomously and non-autonomously to maintain muscle integrity during aging.

| DISCUSS ION
In this work, we investigated the relationship between lifespan and motility phenotypes caused by DAF-2/IIRc inactivation by analyzing F I G U R E 3 Inactivation of DAF-2 in neurons or in the gut is sufficient to increase lifespan. (a-k) Survival curves of animals with DAF-2 depletion in all cells (a-k), muscle (a, b), germline (c, d), epidermis (e, f), intestine (g, h), neurons (i, j), or neurons and intestine (k). Numbers (e.g. Muscle 1, Muscle 2) refer to distinct alleles driving TIR1 expression (see Table S1). For each condition, n is about 80 individuals. Some experiments were split in separate graphs for clarity; thus, some graphs share the same negative and positive controls: (a), (e), (g), (i); (b), (d), (j); (f) and (h). The control conditions correspond to the N2 or daf-2(kr462) strains, in the presence of EtOH or auxin, whose lifespans did not show significant differences. (l-n) Survival curves of animals with DAF-2 depletion in all cells (l-n), intestine (l), neurons (m), or neurons and intestine (n) in presence of 20 mM paraquat. Controls correspond to daf-2(kr462). All strains have been treated with auxin. For each condition, 75 to 100 individuals have been assayed. For detailed lifespan data, replicates, statistics, and summary of independent assays see Table S2 and Figure S2 the tissue-specific functions of DAF-2/IIRc in the regulation of these phenotypes. To this end, we have created a reporter line that allows the visualization of both DAF-2/IIRc expression and its degradation by the TIR1 auxin-inducible system in live animals.

| daf-2(kr462) transgenic animals as a model to study DAF-2 expression and function
As in other species, multiple isoforms of DAF-2/IIRc have been described, including DAF-2A and DAF-2C, which differ by the size of the α-chain of the α2β2 tetramer, and are analogous to mammalian IR-A and IR-B, respectively (Ohno et al., 2014). The mNeon-Green tag was inserted into the C-terminus shared by these two isoforms which control lifespan, dauer, heat tolerance (A and C) and avoidance behavior (C only). Three shorter isoforms have been reported (Ohno et al., 2014). DAF-2B, which retains the extracellular ligand-binding domain but lacks the intracellular signaling domain, modulates insulin signaling by sequestering insulin peptides, but its expression is restricted to the developmental larval stages and is no longer observed in the adult stage (Martinez et al., 2020).
The DAF-2D and E isoforms lack part of the β-chain and the IRS1 F I G U R E 4 DAF-2 degradation in neurons or muscles differentially alters motility in an age-dependent manner. (a-c) Body bends frequency of 1-day-old (a, b) and 13-day-old (c) adults with depletion of DAF-2 in all cells or in the intestine, muscle, neurons or muscle and neurons as indicated. Control corresponds to daf-2(kr462) in presence of auxin. Numbers (e.g., Muscle 1, Muscle 2) refer to distinct alleles driving TIR1 expression (see Table S1). For numbers and percentage of worms with a strong DAF-16::wrmSCARLET nuclear signal, see Table S4. Scale bars: 20 μm and E isoforms (Kimura et al., 1997). Thus, DAF-2A and/or DAF-2C play a major role in the regulation of motility, lifespan, dauer, and oxidative stress resistance while the DAF-2D and E isoforms are not essential.
We detected DAF-2::AID::mNG receptor expression throughout the body, as in flies and mammals, and more extensively than previously appreciated using immunochemistry (Kimura et al., 2011).
Interestingly, DAF-2::AID::mNG receptors fluorescence appeared as a punctate pattern in the cytoplasm of most cells, rather than enriched at the cell membrane. This pattern may highlight DAF-2/ IIRc biosynthetic pathway and/or endosomal signaling compartment. Recent data obtained in mammalian cell culture also showed that the majority of insulin receptors are localized within intracellular vesicles under regular culture conditions (Boothe et al., 2016).
Internalization of the insulin receptor is necessary to shut down insulin signaling, but it also induces endosome-specific signal transduction (Morcavallo et al., 2014). Increasing evidence suggests that alterations in the insulin receptor trafficking can lead to severe insulin resistance (Chen et al., 2019). daf-2(kr462) transgenic animals will be useful to further study the conservation of insulin/IGF-1 receptor biosynthesis and trafficking and its potential deregulation in the context of different daf-2 mutants.

| DAF-2/IIRc activity in neurons and intestine limits wild-type lifespan via shared mechanisms that do not involve inter-organ inactivation of the DAF-2/IIRc
Our data pointed to the importance of both intestine and neurons for lifespan regulation by the DAF-2/IIRc pathway as reported in two recent studies using similar tools (Venz et al., 2021;Zhang et al., 2021 (Uno et al., 2021;Zhang et al., 2021).
We further showed that combined DAF-2/IIRc degradation in the neurons and in the gut did not further extend the lifespan of the animals compared to worms with DAF-2/IIRc degradation in the gut.
This suggests that DAF-2/IIRc signaling in the gut and neurons share F I G U R E 6 DAF-16 is required for motility regulation when DAF-2 is inactivated in neurons but not in muscles. Body bends frequency of 1-day-old (a) and 13-day-old (b, c) adults with depletion of DAF-2 and DAF-16 in neurons or muscle (a, b) or with down-regulation of unc-120 by RNAi in a rrf-3(pk1426) genetic background (c) (see Experimental procedure and Table S1 for detailed genotype of strains). Data from three independent experiments were pooled. The number of animals scored is indicated under each bar. The bars correspond to the median values, the means are represented by black horizontal lines, and brackets show standard deviations. Comparisons were done with Kruskal-Wallis, Dunn post hoc tests with FDR method to adjust p-value, ns: not significant, *: p adjusted <0.05, ***: p adjusted <0.001 downstream mechanisms for the regulation of lifespan. Surprisingly, although intestinal DAF-2/IIRc inactivation conferred resistance of animals to oxidative stress, the correlation between longevity and oxidative stress resistance phenotypes was not verified with neuronal inactivation of DAF-2/IIRc. The link between oxidative stress resistance and lifespan extension has been a matter of intense debate for years (Dues et al., 2017(Dues et al., , 2019. Our data suggest that while neuronal and intestinal DAF-2/IIRc affect lifespan through common mechanisms, those mechanisms do not seem to involve resistance to oxidative stress. However, we cannot exclude that our results might have been different if we had used another oxidative stressor, as we only studied a severe one (paraquat).
Shared mechanisms may result from the inactivation of DAF-2/ IIRc in distant tissues (i.e., gut or neurons) when DAF-2/IIRc is inactivated in one tissue (i.e., neurons or gut, respectively) as previously proposed (Libina et al., 2003;Uno et al., 2021). However, our observations do not support this model, as neuronal or intestinal inactivation of DAF-2/IIRc triggered nuclear accumulation of DAF-16 in the same tissue, respectively ( Figure 5). This discrepancy is probably linked to the lines used for monitoring DAF-16 subcellular localization. Previous studies used wild-type animals carrying an integrated multicopy array (Pdaf-16::daf-16::gfp) that exhibit several phenotypes that are reminiscent of DAF-16 gain-of-function (Henderson & Johnson, 2001). Therefore, gut and neuronal DAF-2/IIRc function in lifespan does not seem to depend on DAF-2/IIRc inactivation through inter-organ communication. daf-2(e1370) prevents loss of gut integrity (Gelino et al., 2016) and visceral pathologies associated with aging that limits worm survival (Ezcurra et al., 2018). During the last years, several longevity pathways have been shown to involve brain-gut communications in C. elegans (Berendzen et al., 2016;Durieux et al., 2011;Prahlad et al., 2008;Shao et al., 2016;Taylor & Dillin, 2013;Zhang et al., 2018), which may be explored using tissuespecific inactivation of DAF-2/IIRc to better understand the role of DAF-2/IIRc in lifespan regulation.

| Neuronal inactivation of DAF-2/IIRc downregulates motility in a DAF-16 dependent manner during early adulthood
Declined motility of worms with age has been associated with both decreased neuronal stimulation and the loss of muscle cell integrity (Liu et al., 2013;Mergoud Dit Lamarche et al., 2018;Mulcahy et al., 2013). Based on electrophysiological data, Liu et al. (2013) reported that synaptic transmission defects in motor neurons appear as early as day 5 of adulthood, while muscle contraction defects do not occur before day 11 of adulthood. They also showed that daf-2(e1370) mutation delayed the functional decline of neurons at the neuromuscular junction. Consistent with these data, we observed that whole-body inactivation of DAF-2 increased excitatory cholinergic neurotransmission at the neuromuscular junction in middleaged animals, but also as early as day 1 of adulthood. ( Figure S3).
However, when we further tested DAF-2 inactivation in muscle or neurons only, worms did not show this phenotype at neither day 1 or day 14 of adulthood ( Figure S2). Those data strongly suggest that the motility phenotype associated with daf-2 inactivation in either muscle or neurons does not rely on the modulation of cholinergic transmission at the neuromuscular junction, although we cannot rule out subtle defects that would only be apparent at physiological concentrations of acetylcholine. Interestingly, those results also show that the increase in neurosecretion in daf-2 mutants (our results and Liu et al., 2013) can be dissociated from their motility phenotype.
The mechanisms responsible for this discrepancy remained to be elucidated.

| Muscle mitochondria morphology is regulated both autonomously and non-autonomously by DAF-2 activity
Mitochondria are dynamic organelles that undergo cycles of fusion and fission, important for the maintenance of their membrane potential and for mitophagy, respectively (Kleele et al., 2021). Muscle inactivation of DAF-2/IIRc was sufficient to prevent fragmentation of mitochondria, suggesting that the increased motility of the worms relies on improved mitochondrial homeostasis. However, neuronal inactivation of DAF-2/IIRc protein resulted in a similar mitochondrial phenotype, but did not improve motility in middle age.
Thus, although a young mitochondrial network may be a prerequisite for the maintenance of the neuromuscular system with aging, this may not be sufficient for the upregulation of motility by muscle DAF-2/IIRc.

| DAF-2 acts primarily in muscle to control motility in aging worms
Muscle DAF-2/IIRc plays a major role in the loss of worm motility observed in daf-2 mutants, while it did not affect worm lifespan or resistance to oxidative stress. Furthermore, DAF-16/FOXO was dispensable for the regulation of motility by muscular DAF-2/IIRc while UNC-120/SRF was required. These observations agree with our previous data obtained in the context of daf-2(e1370) mutants and support the cell-autonomous impact of DAF-2/UNC-120 on worm motility in middle age. Our results do not exclude the existence of a secondary signal from muscle to neurons that could improve motility.
Indeed, previous work has identified the miRNA mir-1, as a regulator of a retrograde signal from muscle to neurons that modulates neuronal activity (Simon et al., 2008) and mir-1 inactivation improves worm motility under pathological conditions (Schiffer et al., 2021). Interestingly, mammalian SRF has been shown to negatively regulate mir-1 expression (Tritsch et al., 2013;Zhang et al., 2011). Identification of transcriptional targets of UNC-120/SRF in the context of muscle DAF-2 inactivation should help to better define the cellular mechanisms involved in the regulation of motility by muscle DAF-2.
Overall, we have developed and characterized a powerful tool to explore DAF-2 function with aging, uncovering unexpected findings regarding tissue-specific roles of DAF-2 in the regulation of dauer, lifespan, resistance to oxidative stress, and motility, as well as in the cross-talk between tissues.
Although both DAF-2 and DAF-16 are ubiquitously expressed, our approach identifies a tissue-specific, antagonistic, and agedependent role for DAF-2/DAF-16 and DAF-2/UNC-120 signaling in the regulation of motility. Numerous phenotypes have been associated with daf-2 mutants, and the future challenge will be to define the contribution of the different tissue-specific activities of DAF-2 in the regulation of these phenotypes and their downstream effectors.

| Caenorhabditis elegans strains and media
All experiments were performed at 20°C except where specified. All strains were maintained at 20°C, except strain FS428 daf-2(e1370) III (corresponding to the original CB1370 strain outcrossed 6 times) which was maintained at 19°C to prevent larval arrest. Strains were grown on nematode growth medium (NGM) agar plates freshly poured and seeded with Escherichia coli OP50 culture. The wild-type reference strain was C. elegans N2 Bristol. All strains used in this study are described in Table S1.

| Plasmids and generation of single-copy insertion alleles
The plasmids constructed for this study are described in Table S5.
Plasmids used to create single-copy insertion alleles by the mini-Mos method (Frøkjaer-Jensen et al., 2014) are described in previous studies (Zhou et al., 2020(Zhou et al., , 2021, and the newly generated alleles are listed in Table S5. All constructs were verified by Sanger sequencing from GATC Company. For tissue-specific expression, the promoters used were as follows: myo-3-(body-wall muscle), unc-

| Allele generation by CRISPR/Cas9 genome engineering
The generation of kr462 and kr535 alleles is described in Supplemental experimental procedures.

| Aging cohorts and auxin treatment
Auxin plates were prepared by adding auxin indole-3-acetic acid (Sigma-Aldrich) from a 400 mM stock solution in ethanol into NGM at the final concentration of 1 mM (Zhang et al., 2015). For control ethanol plates, the same volume of ethanol was added to NGM. Animals were transferred on auxin or ethanol plates at the L4 stage except for dauer tests for which eggs were laid and grown on auxin or ethanol plates.
For all aging cohorts, 20 μM 5-fluorouracil (5-FU, Sigma-Aldrich) was added to prevent progeny growth. Animals were transferred weekly to fresh plates, without 5-FU after 2 weeks. The day of transition to L4 is counted as day 0 of adulthood for the cohort.

| Dauer, lifespan and oxidative stress assays
Dauer, lifespan, and oxidative stress assays have been performed as described in Supplemental experimental procedures.

| Thrashing assays
For body bends frequency (BBF) measurements, 1-day-old or 13-day-old worms prepared as described in "aging cohorts" were gently transferred into a 12-well cell culture plate (ten worms per

| Aldicarb assay
The tests were performed with synchronized worms. Twenty worms were added to a 100 μL drop of 250 μM aldicarb (Sigma,ref. 33,386) in M9 buffer on polylysine-coated slides and incubated in a humidified chamber. 4, 4.5, 5, and 5.5 h later, the worms were stimulated with a blue light for 50 s and touched with a worm pick. With the blue light still on, their movement was then recorded for 30 s with the same device as for the thrashing tests.
The recordings were made blind and the worms were considered paralyzed if they made less than three body bends during this recording.
Three biological replicates with independent cohorts were performed for each age.

| RNAi
RNAi feeding conditions are described in Supplemental experimental procedures.

| Scoring of DAF-16 nuclear localization
One-day-old adults were observed using a AZ100 Multizoom Microscope (Nikon) equipped with a CMOS flash 4 C11440 (Hamammatsu) camera, for no more than 10 min per slide to avoid postmounting DAF-16::wrmSCARLET nuclear translocation. In absence of auxin, the fluorescence associated with DAF-16::wrmSCARLET appeared diffuse in the cytoplasm and nuclei of most tissues, except in the head neurons where it was completely excluded from the nuclei. A tissue was scored positive when several nuclei were brighter than the cytoplasm. Strains scoring and image analysis were performed blind.

| Scoring of muscle mitochondria morphology
Images of krSi134  worms were acquired on an Axioscop compound microscope (Zeiss) equipped with a Neofluar 63x/NA 1.25 oil-immersion objective and a EMCC CoolSnap HQ (Photomectrics) camera. For each worm, a representative image from the posterior body wall muscle cells was acquired.
Cells with long interconnected mitochondrial networks were classified as interconnected; cells with a combination of interconnected mitochondrial networks along with some smaller fragmented mitochondria were classified as interrupted; cells with sparse small round mitochondria were classified as fragmented. Strains scoring and image analysis were performed blind.

| Statistical analysis
All statistical analyses were performed with R version 4.0.1 (2020-06-06). The R Test Survival Curve Differences (package survival_3.2-3) was used to analyze lifespan assays. This test is based on the Grho family of tests which makes use of the Kaplan-Meier estimate of survival (Fleming et al., 1982). Thrashing assays were analyzed using the non-parametric Kruskal-Wallis rank sum test, followed by Dunn's test of multiple comparisons (package rstatix_0.6.0) with FDR adjusting method as post hoc tests. To compare between different conditions for the mitochondria morphology assay, a Fisher exact test was performed, followed by pairwise tests with FDR adjusting method as post hoc tests (package rcompan-ion_2.4.1). For all tests, compared samples were considered different when statistical test gave an adjusted p-value <0.05 (*p < 0.05; **p < 0.01; ***p < 0.001), ns: non-significant.

ACK N OWLED G EM ENT
We thank the Caenorhabditis Genetic Center (which is funded by NIH Office of Research Infrastructure Programs, P40 OD010440) for strains. We are grateful to Laure Granger for technical assistance. This work was supported by the European Research Council

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The authors confirm that the data supporting the findings of this study are available within the article and its supplementary materials. The strains generated in this study will be shared by the lead contact upon request.