The UPR mt is required for the enhanced longevity of daf-2 mutants and PHBdepleted daf-2 mutants
PHB depletion has a differential effect on lifespan as well as in the induction of the mitochondrial unfolded protein response (UPRmt) depending on the genetic background: lack of PHB shortens longevity and triggers a very strong induction of the UPRmt in otherwise wild-type worms while in the longlived insulin/insulinlike growth factor 1 (IGF-1) receptor daf-2(e1370) mutants, the induction of the UPRmt is reduced and lifespan is extended (Artal-Sanz et al. 2009, Gatsi et al. 2014). Since the link between the UPRmt and longevity remains unclear, we analysed whether the UPRmt was necessary for the enhanced lifespan of daf-2 mutants upon PHB depletion. UPRmt activation is controlled by mitochondrial import of ATFS-1 (Nargund et al. 2012), thus we investigated whether atfs-1 knockdown affects lifespan. As previously shown (Bennett et al. 2014, de la Cruz-Ruiz et al. 2021)], depletion of atfs-1 by RNAi did not affect the lifespan of wild-type worms(Fig. 1a, Table S1). However, lack of ATFS-1 suppressed the extension conferred by phb-1 depletion to daf-2 mutants. Moreover, ATFS-1 depletion drastically reduced the lifespan of daf-2 mutants (Fig. 1b, Table S1). This result shows that the UPRmt plays an important role in the longevity of insulin/ IGF-1 signalling (IIS) mutants, as mitophagy does (Palikaras et al. 2015).
Identification of PHB interactors affecting the UPRmt
Our objective was then to identify new genetic interactors of PHB, to shed light into the molecular mechanisms of the mitochondrial stress response, both in wild-type and in IIS mutants, to assess the role of the IIS pathway in mitochondrial protection. For this, we performed an RNAi screening in PHB depleted worms and IIS;PHB depleted mutants expressing GFP driven by the hsp-6 promoter (a widely used UPRmt reporter) using a double RNAi strategy (Fig. 1c). The screening was performed by mixture of bacterial cultures (Lehner et al. 2006, Tischler et al. 2006, Jadiya et al. 2014) using the OrthoList RNAi sublibrary (Kamath et al. 2003, Shaye et al. 2011, Hernando-Rodriguez et al. 2018) and phb-1(RNAi) bacteria. Each plate of the screening contained a positive control (phb-1;atfs-1 RNAi) and a negative control (control;phb-1 RNAi) (Fig. 1c). We first confirmed that control;phb-1(RNAi) reduced expression of PHB in both, wild-type and daf-2(e1370) mutants (Figure S1a). Dilution of phb-1 RNAi with control bacteria also recapitulated the longevity (Figure S1b) (Artal-Sanz et al. 2009) and the UPRmt induction phenotypes of PHB depletion in wild-type and daf-2 mutants (Figure S1cd) (Gatsi et al. 2014). In both cases, induction of the UPRmt was completely abolished upon atfs-1(RNAi) (Figure S1cd). These results confirmed that diluting phb-1(RNAi) with control(RNAi) is efficient in silencing gene expression and leads to the already described phenotypes upon PHB depletion.
We assessed whether positive controls (phb-1;atfs-1 RNAi) and negative controls (control;phb-1 RNAi) were easily separable by calculating the Strictly Standardized Mean Difference (SSMD) (Zhang 2007). All screened plates (72 96well plates x 2 backgrounds x 2 biological replicates) showed a SSMD > 3 (Figure S1e), which is an indicator of good quality. We also calculated the Pearson correlation coefficient to assess the reproducibility of the replicates (Figure S1f). We observed a good reproducibility in phb-1(RNAi) and daf-2;phb-1(RNAi) animals, although in daf-2;phb-1(RNAi) the variability was slightly increased. All together, these results suggest that the double RNAi strategy by mixing bacterial culture in equal proportions is suitable for large scale screenings.
We set up a size criterion and removed worms with a length smaller than 660 µm (Figure S2a) as hsp-6 expression increases during development and the activation of the UPRmt is strongest at L4 and young adult stages (Yoneda et al. 2004). We found 88 genes in phb-1(RNAi) and 194 genes in daf-2;phb-1(RNAi) whose depletion caused developmental defects (File S1). IIS mutants have several developmental defects (Gems et al. 1998) and are more sensitive to RNAi, having a more intense response (Wang et al. 2004), explaining the bigger number of RNAi clones affecting development in daf-2 mutants. By performing functional annotation cluster, we found in both cases genes encoding for ribonucleoproteins, proteasome and genes involved in protein transport (Figure S2bc). In addition, in daf-2;phb-1(RNAi) treated mutants, we found a cluster of genes involved in RNA splicing (Figure S2c). As expected, the bigger cluster contained genes related to larval development in both groups (Figure S2bc).
Based on the p value and the log fold change (LFC) of the remaining genes (p value < 0.05 and LFC < 0.58 or LFC > 0.58), accessible at the searchable online tool https://phbscreen.shinyapps.io/phbscreen/, we found a big number of genes whose depletion reduced the UPRmt (File S2): 355 (5.7%) in phb-1(RNAi) treated worms and 419 (6.8%) in daf-2;phb-1(RNAi) treated mutants (Fig. 1d). Additionally, we identified 102 genes (1.7%) in phb-1(RNAi) worms and 287 genes (4.6%) in daf-2,phb-1(RNAi) mutants whose depletion further increased the UPRmt (Fig. 1d). By looking at the overlap of the identified clones between the two backgrounds, we observed common candidates in both backgrounds and others specifically affecting one of the two backgrounds (Fig. 1e and File S2). We also identified candidates that had opposing effects: 8 RNAi clones reduced the signal in phb-1(RNAi), while increased it in daf-2;phb-1(RNAi), and 3 genes whose depletion increased the signal in phb-1(RNAi) while reduced it in daf-2;phb-1(RNAi) (Fig. 1e and File S2).
Among all the significant genes, we performed a Gene Ontology (GO) term enrichment analysis using the topGO package in R (Alexa et al. 2006) and ReviGO (Supek et al. 2011) to group them by biological processes. We examined separately the genes whose depletion upregulated and downregulated the mitochondrial stress response. (Figure S3 and File S3). Amidst the ones increasing the UPRmt, we found an enrichment of genes encoding for general pathways such as embryo and larval development, determination of adult lifespan, and behaviour (Figure S3a). Interestingly, we also found in both genetic backgrounds enrichment in specific pathways such as mitochondrial organization, mitochondrial transmembrane transport and mitochondrial ATP synthesis coupled to electron transport, which is in line with previous publications (Lee et al. 2003, Durieux et al. 2011, Baker et al. 2012, Bennett et al. 2014, Munkacsy et al. 2016); as well as pathways related with cellular response to different stresses and protein catabolism. This data suggest that the mitochondrial stress response is associated with a complex network of biological pathways, which may play a crucial role in maintaining mitochondrial homeostasis under stress conditions.
Among the RNAi clones that decreased the UPRmt we found genes belonging to general pathways such as gonad development, reproduction, molting, body morphogenesis, growth and locomotion, in both genetic backgrounds (Figure S3b). The most enriched processes in both backgrounds are related with protein synthesis such as translation, ribosomal assembly and mRNA processing. Such pathways have already been described to reduce the UPRmt (Haynes et al. 2010, Nargund et al. 2012, Shore et al. 2012, Houtkooper et al. 2013, Runkel et al. 2013, Hernando-Rodriguez et al. 2018). Moreover, we also found enrichment of genes related with intracellular pH reduction and ATP hydrolysis coupled to proton transport in both backgrounds as previously reported (Runkel et al. 2013). Interestingly, we found mRNA splicing genes only in phb-1(RNAi) treated worms, as in daf-2;phb-1(RNAi) treated mutants depletion of those genes caused developmental defects and were not included in the analysis.
By performing GO enrichment analysis, we found almost the same processes enriched in both backgrounds. Therefore, in order to select potential candidates that could explain the opposing lifespan phenotypes occurring upon PHB depletion we selected genes related to stress responses or lifespan that were strong candidates in the screening (bigger fold change and significance), not reducing size and having a good distribution of the intensity data in both biological repeats (Kim et al. 2018). We selected 38 clones for retest, from which 7 were confirmed as UPRmt regulators (Figure S4a and Table S2). The reduced number of confirmed clones could be explained by differential growth conditions of the parents subjected to bleaching, that were grown in liquid media for the screening and in solid plates for the retest (see materials and methods) (Lewis et al. 1995, Van Voorhies et al. 1999). From the confirmed clones, we validated them on solid plates and focused on two transcription factors whose depletion reduced the mitochondrial stress response in phb-1(RNAi) worms and in daf-2;phb-1(RNAi) mutants: ZNF-622 and TLF-1; and two epigenetic factors, HIS-65 and USP-48, whose depletion further induced the mitochondrial stress response more strongly in daf-2;phb-1(RNAi) mutants. The 4 clones showed good intensity and length values in the screening (Figure S4be). The other 3 clones, a translation elongation factor (eef-1B.1), the antioxidant transcription factor skr-1, and the epidermal development regulator blmp-1 will be the focus of future studies as they seem to specifically regulate the UPRmt in wild-type animals.
ZNF-622 and TLF-1, new transcription factors regulating the UPRmt and longevity
ZNF-622, the orthologue of a human conserved C2H2 zinc finger gene ZNF622/ZPR9 (Ma et al. 2021), and TLF-1, a TATA binding protein (TBP)like factor (Dantonel et al. 2000, Kaltenbach et al. 2000), are two transcription factors whose depletion by RNAi on plates reduced the mitochondrial stress signal consistently in both backgrounds, phb-1(RNAi) treated animals and daf-2;phb-1(RNAi) treated mutants, without affecting hsp-6 expression in the absence of mitochondrial stress (Fig. 2ab). To assess whether these genes were also regulators of other stress responses, such as the UPRER or the heat shock response (HSR), we depleted znf-622 and tlf-1 in worms expressing the UPRER reporter, Phsp-4::GFP, and the HSR reporter, Phsp-16.2::GFP (Figs. 2c and d, respectively). Treating worms with tunicamycin, an inhibitor of protein Nglycosylation, a critical step for the synthesis of glycoproteins, induced the expression of Phsp-4::GFP, which was abolished upon depletion of ire-1, a key regulator of the UPRER. We observed a significant inhibition of hsp-4 expression upon depletion of znf-622, and no effect upon tlf-1(RNAi) (Fig. 2c). When subjecting worms to heat shock, there was a high induction of the Phsp-16.2::GFP reporter, that was reduced upon depletion of the heat shock transcription factor, hsf-1(RNAi). We observed that depletion of the two transcription factors did not affect the response (Fig. 2d). This suggests that these two transcription factors are specifically modulating the mitochondrial stress response, instead of being general stress regulators, although ZNF-622 may play a role also in the UPRER regulation, as the mammalian ortholog ZNF622 has been related with ER quality control and ERassociated degradation (Shaban et al. 2021).
Next, we asked whether these two genes were specific regulators of the PHBelicited mitochondrial stress response or were general regulators of the UPRmt. For this we analysed the effect of depleting the two transcription factors in worms subjected to other mitochondrial stress inducers, such as lack of SPG-7, a component of the mAAA protease (Yoneda et al. 2004) (Fig. 2e), or knocking down cco-1, the nuclearencoded cytochrome c oxidase-1 subunit Vb/COX4 (Dillin et al. 2002) (Fig. 2f). SPG-7 regulates proteolytic degradation of mitochondrial proteins and is required for mitochondrial ribosome biogenesis. Depletion of spg-7 causes slow growth and reduced brood size, increases pathogen resistance (Pellegrino et al. 2014), perturbs mitochondrial morphology and induces the UPRmt (Benedetti et al. 2006, Haynes et al. 2010, Nargund et al. 2012, Nargund et al. 2015). Moreover, the yeast mAAA protease orthologue interacts with prohibitins (Steglich et al. 1999). In our case, depletion of SPG-7 strongly induced the UPRmt, which was abolished by atfs-1(RNAi). We observed that both, znf-622 and tlf-1 depletion significantly reduced the UPRmt upon spg-7(RNAi) (Fig. 2e). Depletion of cco-1 causes also developmental delay, induces the UPRmt (Bennett et al. 2014, Tian et al. 2016) and increases lifespan (Tullet et al. 2008, Durieux et al. 2011). Accordingly, we observed that cco-1(RNAi) induced a milder UPRmt than spg-7(RNAi), which was again abolished by atfs-1(RNAi). Interestingly, depletion of either znf-622 or tlf-1, did not significantly affect the induction of the UPRmt caused by cco-1(RNAi) (Fig. 2f). All together, these results suggest that ZNF-622 and TLF-1 are two new regulators of the UPRmt caused by altered membraneprotein homeostasis upon phb-1 and spg-7 depletion.
In order to assess the role of the two transcription factors on ageing we analysed the lifespan of wild-type animals and daf-2 mutants under normal and mitochondrial stress conditions (Fig. 2gj, Table S1). Depletion of znf-622 reduced lifespan of wild-type but not prohibitin depleted worms (Fig. 2g), however, it shortened lifespan in both daf-2 and daf-2;phb-1(RNAi) treated mutants (Fig. 2h). Interestingly, the percentage of lifespan shortening was much higher in daf-2 mutants than in wild-type (38% in control, 44.9% in phb-1 RNAi for daf-2 and only 6.9% in wild-type), suggesting that ZNF-622 is an important factor in regulating ageing, specially under reduced insulin signalling conditions. On the other hand, depletion of tfl-1 reduced lifespan of all tested conditions (Fig. 2ij). Again, the lifespan shortening was bigger in daf-2 mutants than in wild-type (48.8% in control and 58% in phb-1 RNAi for daf-2 and only 15.8 in control and 3.8% in phb-1 RNAi wild-type worms), suggesting that their role in longevity is IIS dependent. In particular tlf-1 depletion partially suppressed the lifespan extending phenotype caused by PHB depletion in a daf-2 mutant background. Moreover, it is interesting to point that depletion of both transcription factors, znf-622 and tlf-1, reversed the effect of PHB depletion in lifespan as phb-1; znf-622(RNAi) and phb-1; tlf-1(RNAi) treated worms live longer than znf-622(RNAi) and tlf-1(RNAi) treated worms respectively (Figs. 2g,i). Altogether, we uncovered two conserved transcription factors required for mounting a proper UPRmt that are differentially required for lifespan in wild-type worms and IIS mutants upon PHB depletion.
The histone deubiquitinase USP-48 regulates the UPRmt and lifespan
Among the validated candidates (Figure S4A, Table S2) we selected two epigenetic factors, HIS-65 and USP-48, that specifically induced the UPRmt in daf-2 PHB-1 depleted animals. When retested on plate, depletion of his-65, a H2A histone, induced the mitochondrial stress response under all genetic conditions except in wild-type animals (Figure S5a). Indeed, lack of HIS-65 further increased the UPRmt caused by PHB depletion. Moreover, it also induced the mitochondrial stress reporter in daf-2 mutants in the absence of mitochondrial stress, suggesting that HIS-65 might have a role in regulating mitochondrial function upon insulin signalling deficiency. Nevertheless, the UPRmt increase caused by HIS-65 depletion was stronger in daf-2;phb-1(RNAi) treated mutants, fully suppressing the reduced UPRmt triggered by daf-2 mutation (Figure S5a). To test whether HIS-65 was specifically involved in the UPRmt induction, we next investigated the role of HIS-65 in the UPRER and HSR (Fig. S5bc). We found that depletion of HIS-65 does not further increase the UPRER signal caused by tunicamycin treatment, while it decreased the hsp-16 reporter signal under heat shock stress. This data suggests that HIS-65 may be required for a proper HSR, but it does not play a role in the UPRER. Depletion of his-65 did not affect lifespan in any of the conditions (Figure S5de), therefore, we did not further investigate HIS-65.
We then analysed the role of USP-48 in longevity. We observed that usp-48(RNAi) reduced lifespan in all the tested conditions (Fig. 3ab). Interestingly, lifespan shortening was stronger in daf-2 mutants (65.5%) and in daf-2;phb-1(RNAi) mutants (75.5%) compared to wild-type (27.5%) and phb-1(RNAi) (26.5%) backgrounds respectively, showing that daf-2 mutants rely more on usp-48 function, just as the ubiquitinated proteome during reduced insulin signalling or dietary restrictions (Koyuncu et al. 2021). We therefore further investigated the role of USP-48 in stress responses and in longevity.
In order to confirm the role of USP-48 in the UPRmt we introduced the UPRmt reporter Phsp-6::gfp in a usp-48(ot872) mutant (Rahe et al. 2019). Interestingly, in the absence of mitochondrial stress, usp-48 mutants showed an induction of the UPRmt, which was suppressed in a daf-2 mutant background (Fig. 3c). As previously shown, depletion of phb-1 strongly induced the UPRmt, which was reduced in daf-2 mutants (Gatsi et al. 2014). While the strongest induction of the UPRmt was observed upon depletion of phb-1 in an usp-48 mutant background, usp-48 deficiency prevented the reduction of the UPRmt conferred by daf-2 mutation (Fig. 3c). These data suggest that USP-48 plays a role in mitochondrial functionality and is required for the attenuation of the UPRmt caused by reduced IIS. This also suggest that the relationship between USP-48 and the UPRmt may be complex and contextdependent, and that additional factors may play a role in modulating its effect.
We next assessed whether USP-48 was also involved in other types of stress, such as the UPRER and the HSR by introducing the respective stress reporters in the usp-48(ot872) allele. In the absence of ER stress, and contrary to what happened with the mitochondrial stress reporter, usp-48 mutants did not induce the UPRER. However, when ER stress was induced with tunicamycin, usp-48 mutants further increased the UPRER reporter (Fig. 3d). Interestingly, USP-48 induction of mitochondrial (Fig. 3c) or ER (Fig. 3d) reporters is strong when the stress is already triggered by other sources, such as phb-1 depletion or tunicamycin treatment, suggesting that USP-48 may play a role in the adaptation and survival of cells under stress conditions just as observed in the lifespan experiments. Since mitochondrial function and the PHB complex are also involved in the UPRER (Lourenco et al. 2021), additional research is needed to fully elucidate the precise mechanisms underlying this observation. The HSR reporter, Phsp-16.2::gfp, however, was mildly induced in usp-48 mutant animals specifically in the head under basal conditions, (Figure S6). On the contrary, heat shock did not promote a higher response in usp-48 mutant worms as compared to wild-type (Fig. 3e). This suggest that USP-48 may not be necessary for the full activation of the HSR but may play a modulatory role in the response. It has been shown that lowdose mitochondrial stress suppresses the programmed repression of the HSR inducing a specific subset of HSF-1 target genes, including hsp-16.2 (Labbadia et al. 2017, Williams et al. 2020). Whether USP-48 regulation of the HSR pathway under basal conditions is linked to mitochondrial dysfunction deserves further investigation. Altogether this data suggests that USP-48 plays a more prominent role in regulating mitochondrial function and the response to mitochondrial stress.
In order to assess the role of USP-48 in other long-lived mutant backgrounds where PHB depletion enhances lifespan we performed lifespans in daf-7(e1372) and isp-1(qm150) mutants (Artal-Sanz et al. 2009). We observed that depletion of usp-48 shortened the lifespan of daf-7 mutants and fully suppressed the lifespan extension conferred by PHB depletion to daf-7 mutants (Fig. 3f). Likewise depletion of usp-48 reduced lifespan of isp-1 mutants as well as isp-1;phb-1(RNAi) treated mutants, but in this case PHB depletion was still beneficial for isp-1 mutants (Fig. 3g). Thus, USP-48 is generally required for lifespan, however, its requirement is exacerbated under stress conditions. Overall, these findings provide important insights into the complex relationship between USP-48 and stress response pathways, and highlight its crucial role in promoting longevity, particularly in stress conditions.
Nuclear levels of USP-48 increase upon PHB-1 depletion and IIS deficiency
usp-48 encodes the C. elegans ortholog of the human ubiquitin hydrolase USP48 (Komander et al. 2009) involved in restricting cellular plasticity (Rahe et al. 2019). Mammalian USP48 function is not well understood, but some studies have suggested multiple roles, among them, as a deubiquitylating enzyme (DUB) for H2A, antagonizing BRCA1 E3 ligase function and regulating DNA damage response (Uckelmann et al. 2018). Also, Chromatin ImmunoprecipitationMass Spectrometry (ChIPMS) on multiple histone marks associated with active transcription or promoter/enhancer activity coprecipitated USP48 (Engelen et al. 2015, Ji et al. 2015).
To further investigate the possible role USP-48 in stress responses, an endogenous translational Cterminal wrmScarlet fusion was built using the Nested CRISPR approach (Vicencio et al. 2019, Vicencio et al. 2022). USP-48::wrmScarlet showed nuclear localization and ubiquitous expression as previously reported (Rahe et al. 2019) (Fig. 4). To study whether phb-1 depletion could affect USP-48 regulation and to discard a USP-48unrelated UPRmt induction, we analysed USP-48 protein levels in the intestine, hypodermis and gonads upon phb-1 depletion in wild-type and daf-2 worms by quantifying fluorescence intensity of nuclei in the three tissues (Fig. 4). We observed a general increase of USP-48 protein levels upon phb-1 depletion and in daf-2 animals (Fig. 4). These data suggest that USP-48 levels are regulated upon mitochondrial stress and insulin deficiency to regulate gene expression. Interestingly, phb-1 RNAi did not further increase the USP-48 signal in the intestine in daf-2 worms as compared to wild-type animals in the same conditions (Fig. 4a), but it did so in the hypodermis and gonads (Fig. 4b-c). This data suggests that USP-48 expression may be tissue and contextdependent.
UPR mt activation in usp-48 mutants is largely independent of the canonical UPR mt transcription factors
Compromising USP-48 nuclear function is sufficient to induce the UPRmt (Fig. 3c). UPRmt activation is dependent upon nuclear accumulation of the transcription factor ATFS-1 (Nargund et al. 2012, Nargund et al. 2015), while the transcription factor DVE-1 is required for full UPRmt activation (Haynes et al. 2010, Tian et al. 2016). Therefore, we tested the requirement of both transcription factors for activation of the UPRmt in usp-48 mutants. RNAi depletion of atfs-1 only partially suppressed the UPRmt activation of usp-48 mutants, while dve-1 depletion, surprisingly further increased expression of the UPRmt reporter (Fig. 5a). This suggests that a mechanism independent of the canonical UPRmt must largely be responsible for UPRmt activation by USP-48 deficiency. The strong UPRmt induction occurring upon PHB depletion is fully mediated by ATFS-1 and partially DVE-1 (Hernando-Rodriguez et al. 2018) and enhanced by usp-48 loss of function (Fig. 3c). Therefore, we investigated whether the involvement of ATFS-1 and DVE-1 on the PHBmediated UPRmt was suppressed in a usp-48 mutant background. We observed that usp-48 mutants retained higher levels of the UPRmt after depletion of both transcription factors when compared to wild-type worms upon PHB depletion (Fig. 5b). This again suggests that the mitochondrial stress response of usp-48 mutants is partially independent of the canonical UPRmt transcription factors.
Puzzled by the opposing effect of dve-1 depletion on the UPRmt mediated by usp-48 loss of function, we examined the expression and distribution of DVE-1::GFP in wild-type and usp-48(ot872) mutants. As previously reported, we observed DVE-1 expression in nuclei of several tissues of the head and tail, and nuclei of the hypodermis and intestine along the body, with a stronger intensity in the intestinal cells closer to the tail (Haynes et al. 2007). We found a strong increase of DVE-1 in hypodermis of usp-48(ot872) animals as compared with wild-type (Fig. 5c). Similarly, we observed in usp-48(ot872) worms an increase of DVE-1 in the intestinal cells of the tail (Fig. 5d), but not in the first two intestinal cells closer to the head of the worms (Figure S6c). This data further supports a tissuespecific role of USP-48 in UPRmt induction and suggests that it may be regulated by DVE-1.