Cytoplasmic and mitochondrial aminoacyl-tRNA synthetases differentially regulate lifespan in Caenorhabditis elegans

Summary Reducing the rate of translation promotes longevity in multiple organisms, representing a conserved mechanism for lifespan extension. Aminoacyl-tRNA synthetases (ARSs) catalyze the loading of amino acids to their cognate tRNAs, thereby playing an essential role in translation. Mutations in ARS genes are associated with various human diseases. However, little is known about the role of ARSs in aging, particularly whether and how these genes regulate lifespan. Here, using Caenorhabditis elegans as a model, we systematically characterized the role of all three types of ARS genes in lifespan regulation, including mitochondrial, cytoplasmic, and cyto-mito bifunctional ARS genes. We found that, as expected, RNAi knockdown of mitochondrial ARS genes extended lifespan. Surprisingly, knocking down cytoplasmic or cyto-mito bifunctional ARS genes shortened lifespan, though such treatment reduced the rate of translation. These results reveal opposing roles of mitochondrial and cytoplasmic ARSs in lifespan regulation, demonstrating that inhibiting translation may not always extend lifespan.

In this study, we systematically examined the role of ARS genes in lifespan regulation in Caenorhabditis elegans, a genetic model organism widely used for aging research (Kenyon, 2010). We first characterized a mutant allele of ears-2 gene that encodes the mitochondrial glutamyl-tRNA synthetase, and found that ears-2 mutant worms or worms with ears-2 knocked down by RNAi are long-lived. This long-lived phenotype is mediated by UPR mt . Much to our surprise, knockdown of ears-1 gene, which encodes the cytoplasmic homolog of ears-2, substantially shortened lifespan, though the rate of translation is reduced in these worms. We then characterized all the remaining ARS genes in the C. elegans genome, and found that RNAi of mitochondria-specific ARS genes all extended lifespan, whereas RNAi of cytoplasm-specific or cyto-mito bifunctional ARS genes all shortened lifespan. These results identify opposing roles of mitochondrial and cytoplasmic ARSs in lifespan regulation, revealing the rather complex nature of translation in aging. We suggest that global suppression of cytoplasmic protein synthesis is detrimental to animals, leading to shortened lifespan. Thus, reducing the rate of translation may not always promote longevity.

ears-2 mutant worms are long-lived
In search of long-lived mutants, we found that xu120, a mutant in our collection, showed increased lifespan ( Figure 1A and Table S1). By whole-genome sequencing, we mapped the mutation to the ears-2 gene that encodes the sole C. elegans mitochondrial glutamyl-tRNA synthetase. A single C to T transition in xu120 leads to a predicted change of amino acid residue 152 from methionine to isoleucine ( Figure 1B). To verify the phenotype, we inactivated ears-2 by RNAi and found that RNAi of ears-2 extended lifespan, similar to the phenotype of ears-2(xu120) mutant worms ( Figure 1C and Table S1). Of interest, mutations in another mitochondrial ARS gene lars-2/lrs-2, also extend lifespan (Lee et al., 2003). These data together support the notion that inactivation of mitochondrial ARS genes prolongs longevity.
Inactivation of ears-2 alters mitochondrial function and activates UPR mt ears-2 encodes a mitochondrial ARS. Mitochondrial ARSs catalyze the loading of tRNA with cognate amino acids, which represents a key step in the translation of proteins encoded by the mitochondrial genome. In C. elegans, the mitochondrial genome encodes 12 proteins, all of which participate in the oxidative phosphorylation process essential for ATP production (Okimoto et al., 1992). We therefore examined whether inactivation of ears-2 affects normal mitochondrial functions. We measured the ATP content and reactive oxygen species (ROS) level in ears-2(RNAi) worms, and found that RNAi of ears-2 markedly reduced the ATP level while increasing the production of ROS ( Figures 1D and 1E), indicating compromised mitochondrial functions.
In addition to compromising mitochondrial functions, interfering with mitochondrial mRNA translation is also known to trigger UPR mt , resulting in lifespan extension (Houtkooper et al., 2013). We thus wondered if inactivation of ears-2 induces UPR mt . Using hsp-6::GFP as a UPR mt reporter (Benedetti et al., 2006), we observed that ears-2 RNAi induced UPR mt , similar to the effect of cco-1 RNAi ( Figures 1F and 1G), which served as a positive control (Durieux et al., 2011). The observed UPR mt in ears-2(RNAi) worms is specific, since RNAi of ears-2 did not activate UPR in the ER (UPR ER ); nor did it stimulate heat shock response or DAF-16 signaling pathway ( Figure S1). Taken together, these results demonstrate that inactivation of ears-2 compromises mitochondrial functions and triggers UPR mt .

ears-2-dependent lifespan extension requires UPR mt
We then asked if ears-2-dependent lifespan extension requires UPR mt . To address this question, we examined atfs-1, a key player in UPR mt (Nargund et al., 2012), and found that mutations in atfs-1 suppressed the lifespan-extending phenotype of ears-2 RNAi worms ( Figure 2A and Table S1). As a control, we also assayed mutants of several lifespan regulating transcription factors, such as daf-16, nhr-49, hif-1, skn-1 and iScience Article hlh-30, and found that they are not required for ears-2-dependent lifespan extension (Figures 2B-2F and  Table S1). Notably, ears-2(RNAi) appears to suppress the short-lived phenotype of hlh-30 mutant worms, suggesting a genetic interaction between the two genes ( Figure 2B and Table S1). These data suggest that ears-2-dependent lifespan extension requires UPR mt , indicating an essential role of UPR mt in mediating ears-2 longevity.

Inactivation of ears-1 shortens lifespan
While EARS-2 uploads glutamic acid to its cognate tRNA in the mitochondria, its cytoplasmic homolog EARS-1 does so in the cytoplasm. To evaluate whether inactivation of ears-1 can also extend lifespan, we knocked down ears-1 by RNAi. Surprisingly, RNAi of ears-1 greatly shortened lifespan ( Figure 3A and Table S1). This is rather surprising because it is well known that reducing translation extends lifespan. In the case of EARS-1, a key enzyme that produces charged tRNA glu needed for the synthesis of all the proteins in the cytosol, its inactivation should presumably reduce the rate of translation and thereby extend lifespan. Indeed, using an FRAP-based assay to quantify the rate of translation of a fluorescent reporter (C) ears-2 RNAi extends lifespan. All lifespan assays were performed at 20 C and were repeated at least twice. See Table S1 for lifespan statistics. Logrank (Kaplan-Meier) was used to calculate p values. (D) ears-2 RNAi decreases the ATP level. The same number ($120) of day1 adult worms treated by feeding RNAi were used for ATP level test. n = 3. ***p< 0.001 (t test). (E) ears-2 RNAi enhances the ROS level. 10 mMol/L DCFH-DA was used for ROS measurement. n = 10. ***p< 0.001 (t test). (F-G) ears-2 RNAi induces UPR mt , similar to the effect of cco-1 RNAi. Representative images (F) and quantification graph (G) are shown. Scale bars, 300 mm. n = 11-15. Error bars represent SE of mean. ***p< 0.001, ****p< 0.0001 (ANOVA with Dunnett's test). Also see Figure S1.

OPEN ACCESS
iScience 25, 105266, November 18, 2022 3 iScience Article protein expressed in the cytoplasm (Papandreou et al., 2020), or using the O-propargyl-puromycin (OPP) incorporation assay to quantify the translation rate of total proteins (Somers et al., 2022), we found that RNAi of ears-1 indeed substantially inhibited the translation rate of the reporter protein ( Figure 3B) or total proteins ( Figure 3C). As a control, RNAi of ears-2 only had a minimal effect ( Figures 3B and 3C). We thus conclude that unlike ears-2, inactivation of its cytoplasmic homolog ears-1 shortened lifespan, though such treatment indeed reduced the rate of translation.

ARS genes encoded in the C. elegans genome
The observation that the mitochondria-specific ARS gene ears-2 and the cytoplasm-specific ARS gene ears-1 possess opposing roles in lifespan regulation prompted us to ask the question whether this is a phenomenon unique for glutamic acid-specific ARS genes like ears-1 and ears-2 or a general phenomenon for all ARS genes. We thus decided to systematically examine the role of all C. elegans ARS genes in longevity. As a first step to address this question, we surveyed the C. elegans genomic database (Wormbase) for ARS genes. There are 34 ARS genes encoded in the C. elegans genome, with 15 predicted to act exclusively in the cytoplasm, 13 in the mitochondria, and 6 in both the cytoplasm and mitochondria that are thus considered bifunctional (Table 1). These genes are named in a way similar to ears-1 and ears-2, with cytoplasmic/ bifunctional ARS members named class ''1'' genes and mitochondrial ARS members class ''2'' genes (Antonellis and Green, 2008). However, four genes do not follow this nomenclature, i.e. aars-1, aars-2, vars-1, and vars-2. Specifically, based on the sequence homology, aars-1 and vars-1 are predicted to be mitochondria-specific ARS genes, whereas aars-2 and vars-2 are expected to encode cytoplasm-specific ARSs (Table 1); yet their gene names are switched. To clarify this, we set out to determine the subcellular protein localization of these four ARSs by expressing them as an mCherry fusion protein in transgenic animals. We focused on body wall muscles, a cell type commonly used for visualizing mitochondrial morphology. We found that as expected, AARS-1::mCherry and VARS-1::mCherry co-localized with a , hif-1(ia4) (E) or skn-1(zu135) (F) mutation fails to fully suppress the long-lived phenotype of ears-2 RNAi worms. B-F shared the same WT; vector RNAi curve and WT; ears-2 RNAi curve as they were done at the same time. All lifespan assays were performed at 20 C and repeated at least twice. See Table S1  iScience Article mitochondrial GFP marker (mitoGFP), similar to the case with EARS-2::mCherry ( Figure 4). By contrast, AARS-2::mCherry and VARS-2::mCherry did not, but rather exhibited a more diffuse pattern similar to EARS-1::mCherry ( Figure 4). We thus conclude that as predicted by their gene sequences, aars-1 and vars-1 likely encode mitochondrial ARSs, while aars-2 and vars-2 may instead encode cytoplasmic ARSs.

Opposing roles of cytoplasmic and mitochondrial ARSs in lifespan regulation
We then inactivated all these 34 ARS genes by RNAi and assayed their lifespan. As was the case with ears-2, RNAi of all the 13 mitochondrial ARS genes extended lifespan ( Figure 5). Similarly, inactivation of these mitochondrial ARS genes also induced UPR mt (Figures 6 and S3), but not UPR ER ( Figure S4), heat shock response ( Figure S5) or DAF-16 signaling pathway ( Figure S6). By contrast, inactivation of all the 15 cytoplasmic ARS genes shortened lifespan, similar to the case with ears-1 ( Figure 5 and Table S1). RNAi of nars-1 was previously reported to extend lifespan (Chen et al., 2007). We thus repeated this experiment using multiple conditions and observed a similar short-lived phenotype ( Figures S7B-S7D). Thus, inactivation of cytoplasm-and mitochondria-specific ARS genes shortens and extends lifespan, respectively.
We performed RNAi on worms beginning at the L4 stage to avoid affecting growth and development. Indeed, we found that if initiated at the egg stage, RNAi knockdown of cytoplasmic-specific ARS genes caused lethality or larva arrest. As cytoplasmic ARSs mediate the translation of the vast majority of proteins in the cell, global shutdown of translation is expected to be detrimental to worms, thereby affecting their growth and development. By contrast, this phenomenon was not observed when we knocked down those 13 mitochondria-specific ARS genes. In fact, if we inactivated mitochondria-specific ARS genes beginning at the egg rather than L4 stage, the lifespan-extending effect from some of these genes, such as pars-2, was even more robust ( Figure 5G vs. Figure S7A and Table S1). These results provide additional data supporting that inactivation of cytoplasmand mitochondria-specific ARSs shortens and extends lifespan, respectively.
Notably, inactivation of those 6 cyto-mito bifunctional ARS genes all shortened lifespan (Figure 7 and Table S1), similar to the case with ears-1. As RNAi of tars-1 was previously reported to extend lifespan (Chen et al., 2007), we repeated this experiment under multiple conditions and observed a similar short-lived phenotype ( Figures S7E-S7G). We thus conclude that inactivation of cyto-mito bifunctional ARS genes shortens lifespan. Because inhibiting the expression of these bifunctional ARS genes would reduce translation in both the cytoplasm and mitochondria, this suggests that the lifespan-shortening iScience Article effect resulting from the inhibition of cytoplasmic translation predominates. This is expected, as nearly all the proteins in the cell are translated in the cytoplasm, whereas the mitochondria are merely in charge of the synthesis of 12 proteins. Together, these data demonstrate that reducing translation in the cytoplasm and mitochondria shortens and extends lifespan, respectively, revealing opposing roles of cytoplasmic and mitochondrial translation in lifespan regulation.  The notion that reducing translation can increase adult lifespan has been widely accepted, although the underlying mechanisms are largely unknown (Gonskikh and Polacek, 2017;Kennedy and Kaeberlein, 2009;Steffen and Dillin, 2016;Tavernarakis, 2008). Thus far, this view has been supported by experiments testing the major components in the translation apparatus, such as eIFs, eEFs, ribosomal proteins and ribosomal RNAs (Chen et al., 2007;Ching et al., 2010;Chiocchetti et al., 2007;Curran and Ruvkun, 2007;Essers et al., 2015;Hansen et al., 2007;Heissenberger et al., 2020;Houtkooper et al., 2013;Pan et al., 2007;Rogers et al., 2011;Schosserer et al., 2015;Syntichaki et al., 2007;Tiku et al., 2017;Xie et al., 2019). Nevertheless, whether this is the case for ARSs, a group of enzymes that play an essential role in translation, has not been extensively explored. In this study, we systematically assessed the role of all the ARS genes in longevity in C. elegans. We found that while inactivation of mitochondrial ARS genes extends lifespan and does so via UPR mt , inhibiting those ARSs acting in the cytoplasm, surprisingly, suppressed longevity, although such treatment indeed reduced the rate of translation. Thus, simply reducing the rate of translation may not extend lifespan.
While our observation that inactivation of mitochondrial ARS genes extends lifespan is consistent with the notion that reducing translation extends lifespan, our findings regarding cytoplasmic ARS genes are not. The question arises as to how to reconcile these findings with those demonstrating that reducing cytoplasmic translation extends lifespan. Notably, a previous study reported the intriguing observation that RNAi knockdown of the eukaryotic translation initiation factor 4G (eIF4G) results in a relative increase in the expression of stress response genes with long mRNA, although the overall rate of translation is reduced (Rogers et al., 2011). Knockdown of some of these stress response genes can suppress the long-lived phenotype induced by eIF4G inhibition, suggesting that a relatively increased expression of a select group of genes may underlie the longevity phenotype associated with eIF4G inhibition (Rogers et al., 2011). It is possible that inhibiting the expression or function of other eIFs, eEFs, and ribosomal proteins and RNAs may extend lifespan through a similar mechanism. One common feature of many such factors, such as eIFs and ribosomal proteins, is that they play a regulatory, rather than an essential role, in translation (Jackson et al., 2010;Wilson and Doudna Cate, 2012). For some other factors that are essential for translation such as eEF2 and ribosomal RNAs, the interventions that led to lifespan extension did not directly target such factors, but instead targeted their regulators that are likely not essential for translation (Heissenberger et al., 2020;Schosserer et al., 2015;Tiku et al., 2017;Xie et al., 2019). However, ARSs, particularly cytoplasmic ARSs, are enzymes essential for the synthesis of all the proteins in the cytoplasm, as each such ARS catalyzes the production of tRNA charged with a specific amino acid and thereby participates in the translation of nearly all the proteins in the cell (Kuo and Antonellis, 2020). Thus, direct inactivation of ARS genes would uniformly reduce the expression of all the proteins in the cytoplasm or mitochondria, making it less likely to induce a relatively increased expression of a select group of genes as shown for eIF4G. This may contribute to the short-lived phenotype associated with knockdown of cytoplasmic ARS genes.
By contrast, as the mitochondrial genome merely encodes 12 or 13 proteins, knockdown of mitochondriaspecific ARSs would only affect a very small number of proteins. Thus, it is less likely for such knockdown to induce a severe detrimental effect on the cell; yet it can trigger UPR mt that is known to promote longevity. We therefore suggest that the differential effects resulting from knockdown of cytoplasmic and mitochondrial ARS genes may underlie their opposing roles in lifespan regulation.
As an essential group of enzymes required for protein synthesis, ARSs are associated with a variety of human inherited diseases, ranging from peripheral neuropathies to infantile liver diseases and sensorineural hearing loss (  . Subcellular localization of EARS-1, EARS-2, AARS-1, AARS-2, VARS-1 and VARS-2 (A-F) EARS-2::mCherry (B), AARS-1::mCherry (C) and VARS-1::mCherry (E) co-localize with mitoGFP, a mitochondriatargeted GFP marker, in body-wall muscle cells. EARS-1::mCherry (A), AARS-2::mCherry (D) and VARS-2::mCherry (F) do not co-localize with mitoGFP, but rather exhibit a more diffuse pattern in body-wall muscle cells. Transgenes were expressed in body wall muscle cells under the myo-3 promoter, except ears-2::mCherry that was expressed under its own promoter. A commercial structured illumination microscope (HiS-SIM, High Sensitivity Structured Illumination Microscope) was used to acquire and reconstruct the images. Scale bars, 10 mm. iScience Article different human ARS genes are linked to distinct clinical symptoms; yet they all are ubiquitously expressed in the same tissues/cells and all play a similar role in protein synthesis (Antonellis and Green, 2008;Kuo and Antonellis, 2020). The underlying mechanisms are not clear. Clearly, much remains to be learned about these enzymes. Our research in C. elegans points to a novel, complex role of ARSs in aging. Work in C. elegans has also revealed other roles of these enzymes in mediating stress responses such as resistance to hypoxia and starvation (Anderson et al., 2009;Webster et al., 2017). As ARSs genes are highly conserved from worms to mammals, research in C. elegans will continue to provide valuable insights into the function of ARS genes in health and disease.

Limitations of the study
We show that RNAi knockdown of the cytoplasmic glutamyl-tRNA synthetase gene ears-1 greatly shortens lifespan and substantially inhibits the translation rate of the reporter protein or total proteins. Nevertheless, we do not rule out the possibility that a slight reduction in the expression level of all cytosolic proteins might exhibit a beneficial effect, whereas too much of such reduction would be deleterious. To test this possibility, it would require more precise manipulation of the translation rate than that offered by the RNAi approach used in this study.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following:   -1 (B), alanyl tRNA synthetases gene aars-2 (C), leucinyl tRNA synthetases gene lars-1 (D), isoleucinyl tRNA synthetases gene iars-1 (E), tryptophanyl tRNA synthetases gene wars-1 (F), prolyl tRNA synthetase gene pars-1 (G), valinyl tRNA synthetase gene vars-2 (H), aspartyl tRNA synthetase dars-1 (I), asparaginyl tRNA synthetase gene nars-1 (J), serinyl tRNA synthetase gene sars-1 (K), tyrosyl tRNA synthetase gene yars-1 (L) and glutaminyl tRNA synthetase gene qars-1 (M) shorten lifespan. RNAi of cytoplasmic phenylalanyl tRNA synthetases gene fars-3 (B) does not affect lifespan. A, C-E share the same vector RNAi curve as they were done at the same time. B, G and H share the same vector RNAi curve as they were done at the same time. I-L share the same vector RNAi curve as they were done at the same time. All lifespan assays were performed at 20 C and were repeated at least twice. See Table S1 Table S1 for lifespan statistics. Logrank (Kaplan-Meier) was used to calculate p values. Also see Figure S7. iScience Article reconstruct the worm images. To further improve the resolution and contrast in reconstructed images, sparse deconvolution was used as previously described (Zhao et al., 2021). Images were further processed with software ImageJ (NIH).

QUANTIFICATION AND STATISTICAL ANALYSIS
For the lifespan assays, survival graphs were generated using Prism 8 (GraphPad Software, Inc.) and IBM SPSS Statistics 21 (IBM, Inc.) software. Log-rank (Kaplan-Meier) was used to calculate p values. ImageJ Fiji (NIH) was used to analyze fluorescent images.