Regeneration in starved planarians depends on TRiC/CCT subunits modulating the unfolded protein response

Abstract Planarians are able to stand long periods of starvation by maintaining adult stem cell pools and regenerative capacity. The molecular pathways that are needed for the maintenance of regeneration during starvation are not known. Here, we show that down‐regulation of chaperonin TRiC/CCT subunits abrogates the regeneration capacity of planarians during starvation, but TRiC/CCT subunits are dispensable for regeneration in fed planarians. Under starvation, they are required to maintain mitotic fidelity and for blastema formation. We show that TRiC subunits modulate the unfolded protein response (UPR) and are required to maintain ATP levels in starved planarians. Regenerative defects in starved CCT‐depleted planarians can be rescued by either chemical induction of mild endoplasmic reticulum stress, which leads to induction of the UPR, or by the supplementation of fatty acids. Together, these results indicate that CCT‐dependent UPR induction promotes regeneration of planarians under food restriction.


Introduction
Regeneration is widespread in the animal kingdom with almost every phylum having species with the capacity to regrow some parts of their bodies. Planarians are an extreme example of regeneration. They possess a large population of stem cells, approximately 15-25% of the total cell number in the parenchyma, that permits planarians to fully regenerate their bodies in a few days (Baguñ a, 1976b). Planarian stem cells (pluripotent and specialized) are the only proliferative cells and are able to collectively give rise to any planarian cell type (Baguñ a, 1976b;Reddien, 2018). During the process of regeneration, planarians need to cope with a massive demand for new cells to form the regenerative blastema by inducing hyper-proliferation of their stem cells (Reddien, 2018). Indeed, regeneration is a highly energy-demanding process that requires the animals to allocate resources (Maginnis, 2006). Remarkably, planarians are able to stand starvation maintaining the relative number of stem cells and the regenerative power just like fed or growing planarians (Gonz alez-Est evez et al, 2012a; Felix et al, 2019). Stem cell maintenance during starvation serves as a strategy to allow for a rapid growth when a more favourable nutritional environment is encountered or for being primed for a regenerative response to injury (Felix et al, 2019). However, it is currently unknown how starved planarians cope metabolically with regeneration and/or what genetic programmes are necessary for the maintenance of regeneration specifically under conditions of starvation.
Proteome integrity is essential for the functionality of cells and the survival of organisms. The maintenance of protein homeostasis (proteostasis) is regulated by a complex network, which coordinates protein synthesis, folding, trafficking, aggregation, disaggregation and degradation of proteins. This balanced network is constantly affected by many factors including mutations and age (Balch et al, 2008;Hipp et al, 2019). Interestingly, interventions that increase lifespan such as dietary restriction (DR) or reduction of insulin/ insulin-like growth factor 1 (IGF-1) signalling are associated with enhanced mechanisms regulating proteostasis (Cohen et al, 2009).
Chaperones are a component of the proteostasis network and are responsible for assisting the de novo folding and protection of existing proteins from proteotoxic stress (Labbadia & Morimoto, 2015). There are specific chaperones for each of the protein-folding compartments in a cell. The stress response at the cytosol induces the expression of chaperones, co-chaperones and chaperonins under different sorts of stress. Chaperones can act alone or in combination with various co-chaperones to regulate client-substrate interactions, folding, disaggregation, degradation and trafficking within the cell. Compared to other chaperones, the HSP60/chaperonin member TRiC (TCP1-ring complex or chaperonin containing TCP1, also known as CCT) recognizes a smaller repertoire of substrates and is necessary for folding about 5-10% of newly synthesized proteins including actin and tubulin (Saibil, 2013). It also binds to misfolded proteins regulating their aggregation. Indeed, it has been predicted that late folding intermediates or misfolded species are preferred substrate conformers of the TRiC (Horwich et al, 2007). The endoplasmic reticulum (ER) is the major organelle for lipid synthesis and the biosynthesis/folding and maturation of proteins and another component of the proteostasis network. When overloaded with misfolded proteins, the unfolded protein response (UPR) is activated at the ER leading to decreases in global protein translation and specific transcription of stress response genes that promote proteostasis and cell survival. Misfolded, aggregated or damaged proteins are degraded through the proteasome or autophagy. If the UPR fails to restore the ER to normality, ER stress can promote apoptosis (Labbadia & Morimoto, 2015).
Here we perform transcriptional profiling of stem cells at different nutritional states to unravel regulators of regeneration during starvation. We found that down-regulation of TRiC subunits impedes planarian regeneration only during starvation by down-regulating the UPR. The study reveals that CCTs are necessary for mitotic fidelity and the maintenance of genome integrity of planarian stem cells during starvation. We also find that CCTs regulate ATP levels during starvation. Indeed, supplementation of fatty acids or chemical induction of mild ER stress is sufficient to rescue impairments in survival and regeneration of CCT-depleted planarians exposed to food starvation. We validate that CCT depletion abrogates the UPR activation in mouse hematopoietic stem and progenitor cells (HSPCs) under glucose deprivation, a mammalian regenerative system. Our work identifies CCT-mediated UPR induction as a mechanism that contributes to the unique capacity of planarians to be able to fully regenerate even under starved conditions.

Results
Transcriptional profiling reveals TRiC subunits as potential regulators of regeneration depending on nutritional states In order to investigate the transcriptional profiles of stem cells in different nutritional states, we sorted stem cells by FACS (X1 subfraction: S and G2/M cell cycle phase stem cells) (Hayashi et al, 2006) from 1, 7 and 30 days starved planarians (1dS, 7dS and 30dS, respectively) and performed RNA-seq and pairwise comparison between the different nutrient conditions. By performing gene ontology (GO) enrichment, we found that "mitotic S phase" was among the most overrepresented biological processes down-regulated in X1 stem cells at 7dS and 30dS versus 1dS (Fig EV1A,, agreeing with 1dS planarians being under a mitotic response to feeding (Baguñ a, 1976a). We observed that most of the components of the translation machinery were also down-regulated in response to starvation (Fig EV1A and Dataset EV1d) standing in line with the down-regulation of translation in response to nutrient deprivation in other organisms (Hansen et al, 2007).
Interestingly, the term "protein folding" (Fig EV1A) was one of the most up-regulated gene categories during the feeding response (1dS) when compared to 7dS and 30dS with many of the transcripts corresponding to cytosolic chaperones (25/41 up in 1dS versus 7dS and 12/21 up in 1dS versus 30dS) (Dataset EV1e, f). In addition to their functions in folding of de novo synthesized proteins, many chaperones are also induced under conditions of environmental stress and are involved in protein refolding, disaggregation, trafficking and degradation also in humans (Yang et al, 2016). Grouping differentially expressed genes (DEGs) according to expression profile trajectories into clusters (Kumar & Futschik, 2007) identified one U shaped cluster with a relative expression decrease at 7dS versus 1dS but a recovery of gene expression to feeding conditions (1dS) at 30dS (cluster 1 in Fig EV1B and  . Of note, most of the transcripts related to "protein folding" in this cluster (23/27) were enriched in X1 (stem cells) when compared to Xins (differentiated cells) (Dataset EV2c-d) including most of the subunits of the chaperonin TRiC (6/8) (Figs 1A and EV1C and D and Dataset EV2d). A search for other components of TRiC (not included in cluster 1) corroborated that they were also enriched in stem cells and followed a trend of down-regulation at 7dS but a recovery at 30dS compared to 1dS (Fig 1A,Dataset EV2d). In agreement with our observation, it has been suggested that high levels of CCT subunits are hallmarks of hESCs (Noormohammadi et al, 2016). Based on these data, we wondered whether TRiC subunits could be potential regulators of regeneration in certain nutritional states.

TRiC subunits are necessary for blastema formation specifically in starved planarians
The TRiC is formed by a double-ring complex of 8 units codified by different genes (cct1 to cct8) that belong to the chaperone family of HSP60 (Kubota et al, 1995). Homologs of each cct gene and some gonad-specific paralogs have been previously identified in the sexual strain of S. mediterranea (Counts et al, 2017;Rouhana et al, 2017). Since the sexual-specific ccts had almost undetectable expression levels in our RNA-seq data sets of the asexual strain, we focused on the somatic ccts.
Aiming to understand the function of ccts during regeneration of starved planarians, we designed RNAi schedules under different nutritional states. We injected planarians with either dsRNA targeting gfp as control or any of the ccts for three consecutive days and then amputated heads and tails 5 days later to follow the regeneration of the trunks. To analyse regeneration under starvation, planarians were either 14 days (hereafter referred to as "starved conditions") or 37 days starved when performing the amputation (Figs 1B and EV2A and B). The down-regulation of each one of the 8 ccts during the starvation schedule resulted in defective blastema (new regenerating tissue) formation and death of the animals starting at 30 days after amputation (Figs 1B and EV2A and B). To analyse regeneration under feeding, one extra feeding was allocated 1 day before the first injection (hereafter referred to as "feeding conditions"). The downregulation of each of the 8 ccts in feeding conditions did not affect regeneration as most of the amputated planarians formed a normal blastema (Figs 1C and EV2A). Importantly, the level of RNAi downregulation was not altered by the feeding ( Fig EV2C) and feeding did also not rescue the previously described phenotypes of Smed-smg-1 or Smed-tor RNAi (Gonz alez-Est evez et al, 2012b) (Appendix Fig S1A  and B). Together these results showed that the feeding protocol did not interfere with the efficiency of the RNAi.
A previous report on starved planarians which were fed multiple times with dsRNA for ccts showed a rapid demise of the planarians (Counts et al, 2017). Although they performed dsRNA by feeding instead of injection and used sexual instead of asexual planarians, we tried to design a similar schedule adapted to our specific conditions. Starved planarians were fed after RNAi injections, and we observed that there was also a rapid failure (see for cct3A RNAi in Appendix Fig  S1C). However, when planarians that were under the feeding condition were fed after RNAi injections, we observed that they regenerated even better (Appendix Fig S1D). Altogether indicates that the metabolic condition ("starved" or "feeding") before down-regulation of the genes determines the phenotype: no regeneration in case of "starved" and regeneration in case of "feeding".
We conclude that TRiC subunits are required for blastema formation and successful regeneration in starved planarians but dispensable in fed planarians.
Down-regulation of cct3A in starved planarians leads to mitotic failure during regeneration The observation that ccts are necessary for blastema formation suggested that they may regulate stem cell proliferation and/or   Figure 1. The expression of cct3A is enriched in stem cells and necessary for blastema formation during starvation.
A The expression heat map shows that all the somatic cct subunits from TRiC are up-regulated in X1 (stem cells) compared to Xins (differentiated cells) at 30dS. The colour-coded scale indicates Row Z-score of normalized TPMs (transcripts per million) values per replicate. The asterisks indicate that these genes are significant differentially expressed (q-value (FDR) < 0.01). B RNAi injections schedule in starved conditions. Planarians are at 14dS when the amputation is performed (indicated by the grey cross at day 14). Live images show that cct3A(RNAi) planarians form a minimal blastema compared to controls at the time points shown. At the bottom are the number of planarians with the phenotype shown. The remaining planarians are dead by the time point of regeneration shown. dR indicates days of regeneration. C RNAi injections schedule in feeding conditions. One extra feeding in respect to (B) is introduced 1 day prior to injections. Planarians are at 8dS when the amputation is performed. The live images show that most of the cct3A(RNAi) planarians regenerate like controls. At the bottom are the number of planarians with the phenotype shown. At 8-10dR, the remaining planarians are either dead (4/124) or have a tiny blastema (23/124) and either died or regenerated later. At 51-68dR, the remaining planarians are either dead (10/93) or have a tiny blastema and died later (5/93).
Data information: Scale bars, 300 µm. Source data are available online for this figure.
ª 2021 The Authors EMBO reports 22: e52905 | 2021 differentiation. We therefore examined the pattern of mitoses during regeneration in starved planarians by using a Histone H3 phosphorylated at serine 10 antibody (anti-H3P) (Hendzel et al, 1997). We chose one of the ccts with the highest phenotype penetrance  as representative to further investigate TRiC subunits. It has been shown that planarian amputation triggers two mitotic peaks early in regeneration that contribute to blastema formation and growth (Baguñ a, 1976b;Wenemoser & Reddien, 2010). We observed that cct3A(RNAi) animals had an increased number of H3P + stem cells just after both mitotic peaks when compared to control planarians (Fig 2A). Remarkably, these defects did not occur in feeding conditions (Fig 2B). At 15dR cct3A(RNAi), animals also showed a slight increase of mitoses, whereas after 50dR mitotic activity was nearly abolished when compared to controls (Fig 2A). Next, we sought to determine whether cct3A RNAi affects stem cell numbers by conducting fluorescent in situ hybridization (FISH) and qPCR for smedwi-1, a marker that labels proliferating stem cells in planarians (Reddien et al, 2005). At 30dR stem cell loss was evident in some planarians, and by 50dR, most of the planarians showed few if any stem cells (Fig 2C and D). In contrast, cct3A RNAi did not affect stem cell numbers under feeding conditions (Fig 2C and D), fitting with the maintenance of regenerative capacity observed under this condition ( Fig 1C). In agreement with the lack of blastema growth in cct3A(RNAi) planarians in starved conditions, we observed minimal differentiation of eyes (anti-VC1) (Sakai et al, 2000), brain (Smed-gpas) (Iglesias et al, 2011), epidermal cilia (anti-acTUB) (Iglesias et al, 2011) and muscle (anti-TMUS) (Cebri a et al, 1997) in anterior wounds (Figs 1B and EV3).
The increase in mitotic stem cells in combination with the lack of blastema growth in cct3A(RNAi) planarians under starved conditions suggested that cct3A is required for mitotic progression under these conditions. Since mitotic arrest can increase cell death, TUNEL staining was conducted on regenerating cct3A(RNAi) planarians and controls in starved conditions at the time point which showed the highest accumulation of mitosis and the time point just after (72hR and 4dR). Interestingly, cct3A RNAi induced a massive increase in TUNEL + cells in both anterior and posterior wounds ( Fig 3A) including stem cells (Appendix Fig S2A) at 4dR. Since the increase in apoptosis occurred only after the mitotic peak and not during the peak itself, we speculated that cct3A(RNAi) impairs mitotic fidelity. In line with this assumption, double immunostaining with antityrosin-tubulin (de Sousa et al, 2018) and anti-H3P revealed a significant decrease in the percentage of stem cells in metaphase (from 43% in controls to~17%) and an increase in anaphase (from 38% in controls to~62%) ( Fig 3B, Appendix Fig S2B). In addition, more than 70% of all mitotic figures were defective ( Fig 3B,  Appendix Fig S2B). Interestingly, in the feeding condition the distribution of the different phases of cell cycle in cct3A(RNAi) planarians was similar to controls and we only observed a low number (30%) of defective figures (Appendix Fig S2C-E). This indicates that cct3A is required for mitotic fidelity specifically under starvation.
CCT subunits induce the unfolded protein response (UPR), which is essential for mitotic fidelity during regeneration of starved planarians In order to identify cct3A-dependent pathways that control mitosis under starvation, we performed RNA-seq on starved cct3A(RNAi) planarians and controls at 72hR . Strikingly, among the top down-regulated genes (with the lowest q-value) upon cct3A RNAi were the components of the unfolded protein response in the endoplasmic reticulum (UPR ER ) xbp1 (decrease of 15.44%) and atf6 (decrease of 13.55%) ( Fig 4A). Validation in starved cct3A(RNAi) whole planarians by qPCR showed a decrease of 22.15% for xbp1 and 13.88% for atf6 ( Fig 4B). These two transcription factors are crucial for two of the three main UPR ER branches responsible for the recovery of ER homeostasis or the induction of apoptosis (Labbadia & Morimoto, 2015). We also found down-regulated a repertoire of ER chaperones, known to be transcriptionally induced by xbp1 and/or atf6 (Li & Lee, 2006;Shoulders et al, 2013), including hspa5/bip, a master regulator of the UPR sensor activation (Li & Lee, 2006;Malhotra & Kaufman, 2007) and a bona-fide marker of the UPR activation when analysed at the mRNA and/or protein level (Kozutsumi et al, 1988;Yoshida et al, 1998) ( Fig 4A). Of note, we also observed the chaperone dnajb9, a repressor of the UPR (Amin-Wetzel et al, 2017), to be up-regulated ( Fig 4A). Remarkably, the down-regulation of xbp1 and atf6 was stronger in stem cells with a 43.22% reduction of xbp1 and 31.58% of atf6 ( Fig 4C) than in whole cct3A(RNAi) planarians ( Fig 4B) and there was also a strong reduction of bip (40.42% reduction, Fig 4C).
Interestingly, we observed that the effect on the UPR was not specific of cct3A, as down-regulating another randomly chosen subunit (cct4B) also led to a decrease in xbp1 and bip levels ( Fig EV4B).
In order to investigate whether the observed down-regulation of the UPR ER was causally involved in the mitotic failure and defective blastema of cct3A(RNAi) planarians under starved conditions, we performed RNAi experiments for Smed-xbp1, Smed-atf6, Smed-bip-1 and additionally also Smed-bip-2 and Smed-bip-3 (a second and third planarian bip not differentially expressed upon cct3A RNAi (Fig 4D,Appendix Fig S3A and B and Dataset EV2a). Strikingly, RNAi for bip-1 (Appendix Fig S3B) and double RNAi for xbp1/atf6 (Figs 4D and EV4C) phenocopied the regeneration failure in starved planarians but normal regeneration under feeding conditions as seen in cct3A(RNAi) planarians ( Fig 1B and C). Interestingly, regenerative failure of xbp1/atf6(RNAi) planarians under starved conditions mimicked many of the characteristic features of regeneration failure in cct3A(RNAi) animals, such as aberrant mitosis, increased apoptosis during regeneration (including X1 stem cells) and lack of differentiation ( . xbp1/atf6 (RNAi) planarians started to die by 15dR and all the planarians were dead by 35dR (Fig 4D), far faster than cct3A RNAi where 47% of the planarians were still alive at 50-69dR ( Fig 1B). Compared to cct3A that is enriched in S and G2/M phase stem cells (X1 fraction), xbp1 and atf6 are more enriched in G0/G1 (X2 fraction) and differentiated cells (Xins) (Fig EV4D). Therefore, it is possible that downregulation of xbp1/atf6 affects not only mitotic stem cells but also differentiating and differentiated cells. This could explain the faster lethality phenotype which prevents the observation of possible longterm effects in mitotic stem cells such as stem cell depletion in all planarians.
Together, these results indicate that cct3A-dependent xpb1/atf6 expression contributes to maintain the regenerative capacity of planarians under starved conditions by preventing mitotic failure.

Treatment with a mild dose of the UPR inducer DTT further supports that TRiC subunits function through the UPR
To test whether the UPR down-regulation contributes to the regenerative failure of cct3A RNAi during starvation, we employed the chemical ER-stressor dithiothreitol (DTT) (Kaufman, 1999). We observed that the expression levels of the ER stress marker bip-1 increased during early time points of regeneration in starved control planarians (not RNAi injected) when compared to feeding conditions ( Fig EV5A). We reasoned that if the failure in up-regulating the UPR contributed to the regenerative impairment in cct3A RNAi, low dose DTT may have the potential to rescue this deficiency by independently inducing ER stress and the UPR ER . First, we determined a low level of ER stress that could induce an ER stress response (bip-1 expression) without perturbing regeneration in control planarians under feeding and starved conditions (Fig EV5B-D). Next, we conducted a regeneration experiment with or without adding a mild dose of DTT (0.05mM) for 3.5h prior to RNAi injections ( Fig 5). Interestingly, at 58dR we observed that while DTTuntreated cct3A(RNAi) planarians in starved conditions could not regenerate and were all dead (0% survival) as observed before (Fig 1B), 33.33% of DTT-treated cct3A(RNAi) planarians survived with 14.81% completely regenerated (Fig 5).
We also treated planarians with DTT in the feeding condition ( Fig EV5E). We observed that most of the DTT-treated cct3A(RNAi) planarians that were alive by 8dR (9/13) showed blastema formation and differentiation. However, DTT-treated cct3A(RNAi) planarians finally survived significantly less (25%) than DTT-untreated  15dR-starved c8 Figure 2. cct3A RNAi in starved conditions shows increased numbers of mitoses after both mitotic peaks and stem cell depletion at late time points of regeneration.
A The graph shows the mitotic numbers of trunks during different time points of regeneration in the starved condition for cct3A RNAi and controls. Error bars are SD from the mean and asterisks indicate P < 0.001 (three asterisks), P < 0.01 (two asterisks) and n.s. indicates not significant using two-tailed Student's test with equal variance. n ≥ 9 planarians per time point. Also shown are maximum projections of representative trunks labelled with anti-H3P at different time points of regeneration in the starved condition. On the bottom, the number of planarians with the phenotype shown from the total is displayed. Note that at 50dR planarians show few (a4) or no mitoses (a5). The dashed line delimits the body of the planarian. B The graph shows the mitotic numbers during 72 h of regeneration in feeding conditions for cct3A RNAi and controls. Error bars are SD from the mean, and n.s.
indicates not significant using two-tailed Student's test with equal sample variance; n ≥ 8 planarian per time point. C Maximum projections of representative trunks after FISH for smedwi-1 at different time points of regeneration under starved and feeding conditions. On the bottom, the number of planarians with the phenotype shown from the total is displayed. At 15dR, the expression levels are similar. At 30dR, some planarians have almost no expression (c5) and at 50dR almost all planarians show no expression (c10 and c11) in starved conditions. Under the feeding condition the expression levels at 30dS are similar. On the bottom, the number of planarians with the phenotype shown from the total is displayed. The dashed line delimits the body of the planarian. D Relative expression of smedwi-1 at 13 and 50 days of regeneration after either cct3A or gfp RNAi under feeding and starved conditions. Error bars are SD from the mean and asterisks indicate P < 0.001 (three asterisks), P < 0.05 (one asterisk), and n.s. indicates not significant using two-tailed Student's test with equal variance. n = 3 replicates (5 planarians each) per time point.
ª 2021 The Authors EMBO reports 22: e52905 | 2021 cct3A(RNAi) planarians by 26dR (78.6%) (Fig EV5E). This suggests that combining feeding, cct3A(RNAi) and DTT treatment has a negative effect in survival without generally affecting early regeneration. Overall, our results indicate that the UPR down-regulation contributes to the regenerative failure of cct3A(RNAi) planarians during starvation.

Cct3-knockdown attenuates the UPR ER in freshly isolated mouse hematopoietic stem cells and progenitor cells (HSPCs) in response to glucose deprivation
To test whether Cct3-mediated UPR may be important in a mammalian regenerative system, HSPCs (LSK cells: Lineage À , Sca-1 + , c-Kit + cells) (Okada et al, 1992) from mouse bone marrow were lentiviral transduced with a Cct3-shRNA or a scramble control shRNA. Transduced LSK cells were either cultured under normal (6 mM) or low (2 mM) glucose conditions. NanoString analysis of stress response genes showed that shCct3 transduction in glucose-deprived LSK cells resulted in the regulation of genes involved in "UPR and ER genes" 6A). 32% of the "UPR and ER genes" in the dataset were significantly down-regulated, which included BiP and Xbp1 (Fig 6B,  Dataset EV4). In agreement with the expression data, Cct3knockdown also decreased the ratio of Xbp1(spliced)/Xbp1(unspliced), a well-known marker of ER stress signalling (Yoshida et al, 2001), in glucose-deprived LSK cells (Fig 6C). In line with the planarian data, Cct3-knockdown did not affect the expression of UPR-related genes or the ratio of Xbp1(spliced)/Xbp1(unspliced) in LSK cultures exposed to normal glucose conditions.
Together these data support the conclusion that the induction of the UPR and ER stress responses in murine LSK cells in low glucose conditions are dependent on Cct3, the murine homologue of Smed-cct3A.
cct3A(RNAi) planarians have decreased energy levels and their regenerative defects can be rescued by supplementation with fatty acids In order to explore the dependency of ccts RNAi phenotypes on starvation, we determined general ATP levels in whole planarians A The cartoon represents a regenerating trunk and the squares display the region analysed at the anterior "A" and posterior "P" blastemas. The images are TUNEL maximum projections of representative blastemas from 72 h and 4 days regenerating trunks in starved conditions. The dashed line delimits the body of the planarian. The graph shows the number of TUNEL + cells per mm 2 . The differences in cell death are not significant (n.s.) at 72hR (two-tailed Student's test with equal sample variance) and significant (***P < 0.001 using two-tailed Student's test with equal variance) at 4dR. n ≥ 5 planarians per condition. B Percentage of stem cells in different mitotic phases and the percentage of defective mitotic figures in 72hR anterior blastemas of cct3A RNAi and controls in starved conditions after double immunostaining with anti-tyrosine-tubulin and anti-H3P (****P < 0.0001, ***P < 0.001, *P < 0.05 using two-sided Chi-square test; the number of cells analysed is displayed in Appendix Fig S2B); n ≥ 5 planarians. Representative images are shown. Nuclei are stained with DAPI. Arrows indicate abnormal organization or number of spindle poles and the yellow arrowhead indicates chromosome lagging. Asymmetrical karyokinesis is observed in b11 and b12.
Data information: dR, days of regeneration; hR, hours of regeneration. Scale bars, 4 µm (A), 10 µm (B). Source data are available online for this figure.
6 of 18  ( Fig 7A). We observed that controls in starved conditions had ATP levels comparable to controls that were fed. Remarkably, cct3(RNAi) animals showed decreased levels of ATP specifically in starved conditions when compared to controls (Fig 7A), indicating that cct3A is required to maintain ATP levels in starved planarians. We reasoned that if cct3A(RNAi) planarians were deficient in energy, supplementation with a dietary source of fuel such as fatty acids could rescue cct3A(RNAi) phenotypes. We injected palmitic acid (PA), one of the most abundant fatty acids in animals, and its mobilized form palmitoyl-L-carnitine (PC) before RNAi injection in starved conditions as done previously (Deb et al, 2021). Interestingly, PA/PC significantly increased the survival of cct3A(RNAi) planarians (Fig 7B and C). While all non-supplemented cct3A (RNAi) planarians did not regenerate or were dead by 48dR, 44% of the PA/PC supplemented cct3A(RNAi) planarians were able to fully regenerate (Fig 7B and C) and survived at least until ◀ Figure 4. cct3A RNAi down-regulates the UPR components xbp1 and atf6 whose down-regulation leads to a similar phenotype as cct3A.
A Volcano plot displaying DEGs in starved cct3A(RNAi) planarians compared to controls at 72hR (q-value < 0.05). Y-axis indicates the negative log10 of the false discovery rate (FDR) (q-value). X-axis indicates the beta values (b), a biased estimator of the fold change. For better visualization, 13 dots with -log10 (q-value) > 20 are not displayed. This includes cct3A with q = 0. Notice that when down-regulating cct3A in starved planarians, xbp1, atf6, bip-1 and a repertoire of ER chaperones are down-regulated, whereas dnajb9, a repressor of the UPR, and other ccts are up-regulated. B Relative expression of xbp1 and atf6 related to gfp control at 72 h and 13 days of regeneration during either starving or feeding conditions in cct3A(RNAi) animals.
The graph shows that cct3A RNAi down-regulates xbp1 and atf6 specifically during starvation at 72hR. n = 3 replicates (5 planarians each) per time point. C The graph shows that both cct3A RNAi down-regulates xbp1, atf6 and bip-1 specifically at 72hR during starvation in X1 stem cells. 50 planarians per replicate (three replicates) were used to obtain the X1 population (stem cells in S and G2/M). D Live images show that xbp1/atf6(RNAi) planarians in starved conditions form a minimal blastema when compared to controls while at the feeding condition most of them can regenerate as controls. The remaining planarians in the starved condition are dead by the time point of regeneration shown. In the feeding condition, the remaining planarians at 8-10dR are either dead (  65dR when all the non-supplemented cct3A(RNAi) planarians had died. Together, the data show that fatty acid supplementation is able to revert the regenerative failure of cct3A(RNAi) planarians under starved conditions implying that lipid metabolism is a downstream target of cct3A required for the maintenance of regeneration in starved conditions.

Discussion
In this study, we found CCT-mediated UPR modulation as a mechanism that contributes to explain how planarians can regenerate under starved conditions. CCT subunits function as a multiprotein complex called the TRiC, acting as a chaperonin in the process of correct protein folding. It is known that all eight subunits are required for the proper function of TRiC, and depletion of any subunit is able to reduce its activity (Tam et al, 2006). However, free monomeric subunits have been described as having some additional roles outside the CCT oligomer (Vallin & Grantham, 2019). Indeed, using siRNAs to reduce levels of CCT subunits disrupts both CCT oligomer functions and functions associated with specific CCT monomeric forms (Vallin & Grantham, 2019). Also, depleting one CCT subunit results in a reduction of assembled oligomer and that in turn increases monomeric non-targeted CCT subunits. It has been predicted that if targeting different CCT subunits gives the same results, then most likely the oligomer function has been affected (Vallin & Grantham, 2019). Our experiments show the same phenotype after independently down-regulating each of the 8 subunits, and we also see an increase of the non-targeted ccts mRNAs after A Volcano plots displaying shRNA-Cct3-and shRNA-luciferase-infected LSK cells cultured either in normal glucose (upper graph) or low glucose conditions (lower graph).
Dots represent all the NanoString analysed transcripts. Red dots are UPR and ER stress-related genes. The UPR and ER stress genes significantly down-regulated only under low glucose conditions in shCct3 cells compared to controls are indicated near the lower graph. For better visualization, two dots with q ≤ 0.001 have been removed from the volcano plot (see them in Dataset EV4). Y-axis indicates the false discovery rate (FDR) (q-value), and the X-axis indicates the log2 fold changes. The FDR-based method of P-value adjustment was conducted to calculate the q-values by nSolver software using Benjamini-Hochberg methods. Significance is established by q-value < 0.05 and indicated by the dotted horizontal line. n = 3 mice. B "UPR and ER stress" genes are significantly enriched (only down-regulated) in shCct3 cells compared to controls under low glucose conditions (**P < 0.01 using twosided Chi-square test). C The graph shows that the ratio between Xpb1(spliced) / Xpb1(unspliced) is lower (i.e. lower levels of UPR) in LSK which have Cct-3 down-regulated compared to controls only when cultured under low glucose conditions. Error bars are SD from the mean. P < 0.001 (three asterisks), and n.s. indicates not significant using twotailed Student's test with equal sample variance. n = 3 mice.
Source data are available online for this figure.  cct3A down-regulation (Fig 4A), which could be explained by a feedback loop of the cell trying to re-establish sufficient TRiC levels. Therefore, it is likely that the role of CCTs on planarian regeneration during starvation depends on the whole complex rather than on CCT subunits acting as monomers. How TRiC subunits or the TRiC in the cytosol is able to regulate the stress response at the ER and whether the TRiC itself is responding to stress will require further research. The protein HSF1, a transcription factor that responds to stress to protect cells from protein misfolding and that is known to be directly regulated by the TRiC (Neef et al, 2014), could represent a possible link. Alternatively, the TRiC itself, or a TRiC client or TRiC subunits may be able to bind to ATF6 which would explain xbp1 and bip modulation.

UPR ER C g f p ( R N A i) -s t a r v e d c c t 3 ( R N A i) -s t a r v e d g f p ( R N A i) -f e e d in g c c t 3 ( R N A i)
Our experiments suggest that the effects we observe after downregulation of ccts are mainly due to CCTs regulating the UPR in stem cells. In line with previous reports on mammalian stem cells (Noormohammadi et al, 2016), we found that planarian stem cells show increased expression of several chaperones including all TRiC subunits when compared to differentiated cells. We also showed that bip, a well-known ER chaperone which is transcriptionally regulated in response to ER stress, is strongly down-regulated in S and G2/M stem cells (X1 fraction) after cct3A or cct4B RNAi. Furthermore, cct3A RNAi and xbp1/atf6 RNAi shows stem cell-related phenotypes (defects in mitosis, increased apoptosis during regeneration which includes X1 stem cells and stem cell depletion in case of cct3A RNAi) and prevent regeneration. Mild DTT treatment on whole planarians greatly affects stem cells, since it enhances planarian regeneration of cct3A(RNAi) planarians in starving conditions and indeed allows for survival of the lethal cct3A RNAi. As a way to complement our results with an additional regenerative system which does not have limitations regarding UPR assays, we found that mouse shCct3 hematopoietic stem cells and progenitor cells, when grown under low glucose conditions, functionally downregulate the UPR. Altogether the data suggest that the observed phenotypes are predominantly due to CCT regulating the UPR in X1 stem cells. However, because ccts and more importantly xbp1 and atf6 are also expressed in the X2 (G0 and G1 stem cells) and Xins (differentiated cells), we cannot formally rule out the possibility that part of the phenotype is due to defects in differentiating postmitotic and/or differentiated cells.
The human chaperome is formed by 332 chaperones and cochaperones, and each family of chaperones is essential for cell viability indicating that they have non-overlapping functions (Brehme et al, 2014). Indeed, some other planarian chaperones have been linked to regeneration and stem cells (prohibitin-1 and prohibitin-2, mortalin, hsp40 and hsp60) (Conte et al, 2009;Fernandez-Taboada et al, 2011;Rossi et al, 2014;Wang et al, 2019) and our data validate their enrichment in stem cells. It would be interesting to know whether there are other chaperones apart from CCT subunits which are specifically required for regeneration during starvation. Our cluster 1 data set predicts that at least prohibitins (Rossi et al, 2014) and one hsp40 (Fernandez-Taboada et al, 2011) (Dataset EV2a) could also have a nutrient-dependent effect on regeneration. Our work establishes planarians as an excellent in vivo model to investigate functions of proteins in metabolism.
Mechanistically, we found that cct3A, a TRiC subunit, is necessary to prevent mitotic defects during the response of stem cells to amputation. cct3A RNAi leads to gross nuclear alterations, resulting in cell death observed at wound sites at 4dR. These results highlight the importance of proteostasis to maintain genome integrity. Although it is known that altered proteostasis and genome instability are linked to an increased risk of cancer (Adams et al, 2015;Dai & Sampson, 2016), it is not known whether proteostasis is essential for protecting the genome. There are only few evidences indicating a possible role of proteostasis in regulating genome stability. For instance, it has been reported that ER stress and thus activation of the UPR ER induced by tunicamycin or glucose deprivation, suppresses DNA double-strand break repair in cancer cells by stimulating the degradation of Rad51 (Yamamori et al, 2013). Also, Hsp70 has been linked to genome stability in mouse embryonic fibroblasts under heat shock stress (Hunt et al, 2004) and Hsp110 is associated with genome instability in cancer cells (Dorard et al, 2011). Interestingly, recent studies in planarians and Drosophila have shown that the heat shock protein DNAJA1 and HSP90 interact with PIWI proteins suggesting a possible role of these chaperones in suppressing transposition (Gangaraju et al, 2011;Wang et al, 2019). Although the TRiC has a potential role in cancer development by modulating the folding of client proteins related to oncogenesis, cell cycle and cytoskeleton (i.e. actin and tubulin) (Roh et al, 2015), a direct role in genome stability of stem cells is still missing. Since cct3A RNAi alters mitosis specifically in starving conditions, it remains possible that TRiC-dependent protein folding is required for the timed expression of mitosis regulating proteins, especially under starvation. In support of this possibility, we found that the cyclindependent kinase 1 (cdk1) is down-regulated upon cct3A(RNAi) (Dataset EV3c). While our results cannot proof this concept, we think that is an interesting possibility that could contribute to the observed phenotypes. Remarkably, double RNAi for two of the main ◀ Figure 7. cct3A(RNAi) shows decreased ATP levels during starvation and fatty acid administration rescues regeneration defects in a percentage of planarians.
A ATP measurement on 48hR trunk extracts shows that cct3A RNAi under starved conditions contains less ATP than controls. Error bars are SD from the mean and asterisks indicate P < 0.05 (one asterisk) and n.s. indicates not significant using two-tailed Student's test with equal variance. n = 3 replicates (5 planarians per replicate). B The Kaplan-Meier curve demonstrates increased survival of starved cct3A(RNAi) animals when treated with palmitic acid/palmitoylcarnitine (PA/PC). Percentages indicate the number of survivals at the indicated time points. Three asterisks indicate P < 0.001 with log-rank (Mantel-Cox) test. C Images are representative surviving animals for the different conditions in (B). At the bottom are the number of planarians with the phenotype shown. The rest of planarians are dead at this point. Scale bars indicate 500 µm. D In our proposed model the levels of ER stress (UPR) are basal during regeneration in feeding conditions at 72hR and not essential for stem cell proliferation and regeneration. Planarians obtain energy and lipids from administered food, including exogenous fatty acids (exo FA). During starved conditions, ccts increase the levels of the UPR and this allows proper stem cell proliferation and regeneration when exogenous fatty acids are not available.
Source data are available online for this figure.
ª 2021 The Authors EMBO reports 22: e52905 | 2021 transcription factors of the UPR ER xbp1 and atf6 phenocopies some of the features of cct3A RNAi in terms of genome instability. This altogether indicates that one of the major outcomes after ccts or the UPR ER down-regulation during starvation is related to genome integrity in stem cells. Moreover, we show that starvation protects the planarian stem cell genome and thus possibly contributes to the capacity of planarians to circumvent cancer. These results also suggest that the protective effect that fasting confers to stem cells upon chemotherapy (Nencioni et al, 2018) could be related to increased proteostasis. It is known that lipids in the form of lipid droplets (LDs) accumulate under various stress conditions. For instance, mice injected with the ER stress inducer tunicamycin develop hepatic steatosis which is even more severe and showing LD accumulation when any of the UPR branches is knocked down (Rutkowski et al, 2008;Yamamoto et al, 2010). Several of our results suggest a role of TRiC subunits in lipid metabolism. One single feeding during the fasting period is sufficient to prevent any of the ccts RNAi phenotypes and allows regeneration thus indicating that energy production and/or the supply of nutritional building blocks for cellular anabolism prevents the phenotype. Indeed, the administration of fatty acids (PA and PC) increases survival and prevents cct3A(RNAi) phenotype of regeneration impairment in a percentage of planarians. Also, cct3A RNAi lowers ATP levels which overall suggest that starved cct3A(RNAi) planarians cannot generate sufficient energy. One possibility is that cct3A(RNAi) planarians in starved conditions fail to provide sufficient fatty acids and/or cholesterol for membrane expansion in order to pass mitosis, a highly energy-demanding process (Kalucka et al, 2015). Future research is needed to clarify the link of TRiC components with lipid and LD metabolism and to elucidate whether this occurs downstream of the UPR. We propose a model where TRiC components are up-regulated in response to starvation and this stress response in turn activates the UPR ER that enhances the regenerative capacity. The activation of this protective pathway allows mitotic fidelity and induces a metabolic adaptation of starved animals possibly by providing energy and/or lipid membrane components required for cell growth and proliferation ( Fig 7D). During feeding the levels of ER stress are basal and not necessary for the mitotic regenerative response. Energy and lipid components are obtained from exogenous food ( Fig 7D). However, we also observed that ER stress induction by DTT treatment in fed cct3A RNAi planarians but not in controls led to a decrease in survival. These data suggest that cct3A RNAi also has effects under feeding conditions that per se do not disturb regeneration but synergizes with DTT-mediated induction of ER stress/ UPR to impair survival. While this mechanism remains to be elucidated, the data on starved planarians provide experimental evidence that cct3A-dependent induction of ER stress response/UPR is required for regeneration under starvation conditions.

Starvation and feeding experiments
In experiments involving feeding, all animals were observed to make sure they ate. Food was given in excess and removed after 2 h. For RNA-seq experiments, planarians at 1dS, 7dS and 30dS had the same area 4 mm 2 (5-5.5 mm length at 7dS and 1dS and 5.5-6mm length at 30dS). Graph paper placed under the Petri dish was used to preselect animals, and the final selection was done after measuring the areas with the Leica Application Suite (Leica) on photographs of live planarians taken under a stereomicrospe coupled with a Leica camera MC170 HD (Leica). For RNAi experiments, planarians around 5 mm were selected by use of graph paper.

Fluorescence-activated cell sorting (FACS)
Planarian dissociation and cell population analysis were performed as described previously (Hayashi et al, 2006). Briefly, planarians were cut into small pieces on ice and cell dissociated in the presence of papain (Merck; final concentration 1 mg/ml) during 15 min at room temperature (Moritz et al, 2012). Cells were then filtered through a 40 µm filter, counted and resuspended in staining solution containing the cytoplasmic dye Calcein-AM (Biotium; final concentration of 0.5 µg/ml) and the nuclear dye Hoechst 33342 (Thermo Fisher Scientific; final concentration 30 µg/ml) in order to isolate X1, X2 and Xins populations. Staining was performed in the dark at 25°C, with continuous shaking. Propidium iodide (final concentration 1 µg/ml) was added 1 min before flow cytometry analysis to discard dead cells. Around 250,000 events were sorted per sample using BD FACSAria III or BD FACS Fusion. Cells were put into tubes containing TRIzol LS (Ambion), and RNA was extracted following manufacturer's instructions.

RNA-seq
Library preparation was done using the Illumina kit TruSeq. Sequencing was done on an Illumina HiSeq2500 in 51 cycle, single-end, highthroughput mode at the Core facility DNA sequencing at FLI. 4 replicates per FACS population (X1, X2 and Xins) and per time point (1dS, 7dS and 30dS) were sequenced: 36 samples (2,304 million of nonribosomal reads). Three replicates per RNAi condition (cct3A(RNAi) and gfp(RNAi) were sequenced: six samples (256 million of nonribosomal reads). For transcriptomic analysis, the Dresden Schmidtea mediterranea transcriptome (version 4) was used as reference (Plan-Mine) (Rozanski et al, 2019). Kallisto (version 0.43.0) (Bray et al, 2016) was used to perform read pseudo-alignment and quantification with the parameters -l 200 -s 20 -b 100. Previously, only reads mapping to the planarian reference transcriptome were extracted by using Kallisto and the parameter -F 4. The portion of planarian ribosomal RNA contamination was identified by mapping all reads with Bowtie2 (Langmead & Salzberg, 2012) against a pool of Platyhelminthes rRNA index. IDs numbers from Dresden transcriptome of version 4 are equivalent to version 6 (Planmine). For example, dres-den_comp15_c0_seq1 is equivalent to dd_Smed_v6_15_0_1. Differential expression analysis was performed with the R package Sleuth (version 0.28.1) (Pimentel et al, 2017) with the filtered reads mapping to the reference transcriptome. Wald test was used to identify differentially expressed genes (DEGs). Significance was determined by q-value (false discovery rate (FDR) < 0.1 for pairwise comparisons for the different time points of starvation (7dS versus 1dS, 30dS versus 1dS and 30dS versus 7dS) in the X1. For pairwise comparisons, X1 versus Xins at 1dS, 7dS and 30dS, all TPMs were normalized including also X2 values and the significance was determined by q-value < 0.01. Significance was determined by q-value < 0.05 in the cct3(RNAi) RNA-seq analysis. Downstream analysis was done with the R software. The heat map was done with the R package "Pheatmap". Clustering was performed using the software package Mfuzz (Kumar & Futschik, 2007). Genes with a value lower than 0.35 for cluster assignment are not displayed in the plots. Gene ontology enrichment was done creating lists of DEGs in PlanMine by using default parameters (test Benjamini-Hochberg; P < 0.05). Redundant GO terms were removed from the list by using REVIGO with small similarity parameter (Supek et al, 2011).

RNAi experiments
Templates with T7 promoters appended to both strands were generated. Double-stranded RNA (dsRNA) was synthesized by in vitro transcription following MEGAscript RNAi kit (Ambion) instructions. dsRNA was injected into the planarian as previously described (Gonz alez-Est evez et al, 2012b). Control animals were injected with gfp dsRNA, a sequence not present in the planarian genome. The different protocols used are specified in the "Results" section. The following oligos were used to generate the templates for dsRNA production:

Whole-mount in situ hybridization
Whole-mount in situ hybridization was carried out as described previously (Gonz alez-Est evez et al, 2012b) using an InsituPro VSi ª 2021 The Authors EMBO reports 22: e52905 | 2021 (Intavis). Hapten-labelled RNA probes were prepared by using an in vitro RNA labelling kit (Roche).

Measurement of ATP levels
ATP levels were measured by using an ATP bioluminescence assay kit HS II (Roche) according to the manufacturer's instructions.
Regenerating trunks were incubated in cell lysis reagent and boiled for 15 min at 100°C. Then, samples were sonicated for 4 cycles of 30 00 . Luciferase reagent was added to the samples, and luminescence was measured by the Mithras LB940 plate reader (Berthold Technologies) using the injection function. Three biological replicates consisting of five trunks at 48hR were used per condition, and each sample was replicated three times in each ATP experiment. Protein content was measured by the Bradford method (Bio-Rad Protein Assay).

Annexin V staining
Annexin V staining was performed according to previously published (Shiroor et al, 2020) with the following modifications. After staining with Hoechst 33342 and Calcein-AM (as previously explained for the FACS protocol), 500,000 and 2 × 10 6 cells/sample were pelleted and stained with APC Annexin V RUO (BD Pharmingen; 550475) in a ratio of 5 µl Annexin V: 100 µl freshly made 1X Annexin V buffer (10× concentrate RUO, BD Pharmingen): 100,000 cells. After staining for 15 min in the dark at 25°C, cells were washed and suspended in 300 µl of 1× Annexin V buffer with 1 µg/ml Propidium Iodide and analysed on BD FACSAria III. FlowJo TM Software (FlowJo 10.6.2, Ashland, OR) was used to analyse the data.

DTT treatment
DTT at concentrations 0.05, 0.1 and 0.2 mM was added to the planarian water, and planarians were incubated with that solution for the 3.5 h previous to each RNAi injection. In experiments on non-RNAi planarians, the incubations were performed for 3.5 h on the days that RNAi injections would occur.

Administration of palmitic acid (PA) and palmitoylcarnitine (PC)
12mM BSA-conjugated palmitic acid was prepared as previously (Deb et al, 2021). Briefly, 12 mM palmitic acid was prepared by mixing palmitic acid (Sigma-P0500) in EtOH 100%. Vortex was required to completely dissolve it. Then, 0.2680 gr of fatty acid-free BSA was resuspended in 2 ml of PBS (pH 7.2) at 37°C and subsequently treated in an ultrasound bath for 10 min at 37°C. Next, palmitic acid solution was pre-warmed at 37°C for 2 min and 37 µl was added very slowly (in drops of 7 µl) to the fatty acid-free BSA mixture. Vortex was required after every drop. The 12 mM BSAconjugated palmitic acid solution was filtered through a 0.2-µm mesh. 11 mM palmitoyl-L-carnitine chloride (Sigma-P1645) was prepared in water. A solution containing 750 µM BSA-conjugated palmitic acid and 750 µM palmitoyl-L-carnitine chloride was freshly prepared every day of injection. Three pulses of 69 nl were injected into planarians 3 h prior to RNAi injections.

Experiments with mice
Mice used in this study were C57BL/6J wild-type mice (5-to 6month-old) obtained from Janvier or from internal stock (FLI internal license O_KLR_18-20 and §11 Lizenz). Mice were maintained in a specific pathogen-free animal facility in Fritz Lipmann Institute with 12 h of light/dark cycle and fed with a standard mouse chow. Animal experiments were performed according to Institutional and European Union guidelines.
Isolation of mouse bone marrow cells and FACS sorting of hematopoietic stem and progenitor cells (LSK cells: Lineage À , Sca-1 + , c-Kit + cells)

Lentivirus production
shRNA was inserted into the SF-LV-shRNA-EGFP plasmid using mir30 primers (Chen et al, 2019). Lenti-X (Clontech) cells were cultured in DMEM supplemented with 10% FBS, 100 U/ml penicillin and 100 µg/ml streptomycin. Lentivirus was produced in Lenti-X cells using calcium phosphate transfection of 30 µg shRNA plasmid, 18 µg psPAX2 and 9 µg pMD2.G plasmids according to standard procedures (Schambach et al, 2006). Medium was changed 12 h after transfection and virus supernatant was collected 36 h after changing medium. Lentiviruses were concentrated at 106,800 g for 2.5 h at 4°C, and viral pellets were resuspended in sterile PBS.
Luciferase shRNA: 5 0 TGCTGTTGACAGTGAGC GCCCGCCTGAAGTC TCTGATTAATAGTGAAGCCACAGATGTATTAATCAGAGACTTCAG GCGGTTGCCTACTGCCTCGGA-3 0 Cct3 shRNA: 5 0 TGCTGTTGACAGTGAGCGCACGTGGAGTTATGATT AACAATAGTGAAGCCACAGATGTATTGTTAATCATAACTCCACGTA TGCCTACTGCCTCGGA-3 0 NanoString analysis on normal and low glucose cultured LSK cells Measuring of RNA expression by NanoString was done according to previously (Chen et al, 2019) and to manufacturer protocol (Nano-String Technologies). In brief, 5 × 10 4 hematopoietic stem and progenitor cells (LSK cells) were transduced with shRNA targeting murine Cct3 or luciferase control. The cells from both conditions were then cultured in DMEM medium (Gibco; A1443001) with low glucose concentration (2 mM) or normal glucose concentration (6 mM) for 16hrs. Both of the culture conditions contained 4 mM glutamine, 50 ng/ml mTPO (Peprotech; 315-14), 50 ng/ml SCF (Peprotech; 250-03), 100 unit/ml penicillin and 100 µg/ml Streptomycin (Gibco; 15140-122). 1 × 10 4 cells were lysed in 2 µl of lysis/ binding solution (Applied Biosystems; 8500G14). The cell lysate was then used for hybridization reaction as following: 2 µl of cell lysate was mixed with 5 µl of nCounter hybridization buffer (Nano-String), 2 µl of Core Tagset, 2 µl of extension Tagset, 0.5 µl of 0.6 nm Probe A working pool, 0.5 µl of 0.6 nm probe A extension Pool, 0.5 µl of 3 nm Probe B working pool and 0.5 µl of 3 nm Probe B extension pool (IDT technologies). 2 µl of Nuclease-free water was added to each reaction to reach a final volume of 15 µl. The reaction mixture was prepared in Strip tubes (NanoString technologies). Then, it was incubated at 67°C using a thermal cycler for 16 h. The Nanostring chemistry was processed automatically using nCounter prep-station 5s (NanoString Technologies) according to manufacturer protocol. Directly after the run, the nCounter Cartridge was loaded into nCounter digital analyser 5s (NanoString technologies). Data analysis and q-values were obtained after background correction using nSolver advanced analysis software (v.4) and (R software v 3.3.2.). The following housekeeping genes were used for normalization: ActB, B2 M, Gapdh, Gusb, Hprt, PGK1, Polr1b, Polr2a, Ppia, Rpl19, Sdha and Tbp. For analysis of significantly expressed genes by NanoString, the FDR-based method of P-value adjustment was conducted to calculate the q-values by nSolver software (v.4. Based on R-language) using the Benjamini-Hochberg method. Significance was established by q-value < 0.05. Volcano plots were obtained with GraphPad Prism software version 8. The Xpb1(spliced)/Xbp1(unspliced) ratio was calculated by dividing the number of mRNA counts of spliced Xpb1/unspliced Xbp1 after normalization to housekeeping genes.

Statistical analysis in planarian experiments
In all the manuscript, error bars are SD from the mean and asterisks indicate P < 0.05 (one asterisk), P < 0.01 (two asterisks) or P < 0.001 (three asterisks), P < 0.0001 (four asterisks) and n.s. indicates not significant using two-tailed Student's test with equal variance or two-sided chi-square test or log-rank (Mantel-Cox) test for Kaplan-Meier curves. The number of replicates is indicated in figures and figure legends. Wald test was used to identify differentially expressed genes (DEGs). Significance was determined by qvalue (false discovery rate (FDR)) < 0.1 for pairwise comparisons for the different time points of starvation (7dS versus 1dS, 30dS versus 1dS and 30dS versus 7dS) in the X1. Significance was determined by q-value < 0.01 for pairwise comparisons X1 versus Xins at 1dS, 7dS and 30dS. Significance was determined by q-value < 0.05 in the cct3(RNAi) RNA-seq analysis. More details on the RNA-seq statistics can be found in the RNA-seq sections. For analysis of gene expression by NanoString, significance was determined by q-value < 0.05.

acc=GSE134013)
Expanded View for this article is available online.
Aging-Fritz Lipmann Institute (FLI) for their technical support. We are grateful to J. Solana, M. Rodriguez-Orejuela and all past and current members in CGE laboratory for discussions, suggestions and/or help on the project. We would like to thank M. Iglesias for critical reading of the manuscript and technical support. We are grateful to T. Adell for the protocol for staining with anti-tyrosine-tubulin and to Prof. K. Watanabe and H. Orii for providing the VC1 antibody. CGE was funded by the Leibniz Institute on Aging-Fritz Lipmann Institute (FLI). The FLI is a member of the Leibniz Asso- License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.