The mitochondrial unfolded protein response (UPRmt) is a surveillance pathway that defends proteostasis in the “powerhouse” of the cell. Activation of the UPRmt protects against stresses imposed by reactive oxygen species, respiratory chain deficits, and pathologic bacteria. Consistent with the UPRmt’s role in adaption, we found that either its pharmacological or genetic activation by ethidium bromide (EtBr) or RNAi of the mitochondrial AAA-protease spg-7 was sufficient to reduce death in an anoxia-based Caenorhabditis elegans model of ischemia-reperfusion injury. The UPRmt-specific transcription factor atfs-1 was necessary for protection and atfs-1 gain-of-function (gf) mutants were endogenously protected from both death and dysfunction. Neurons exhibited less axonal degeneration following non-lethal anoxia-reperfusion (A-R) when the UPRmt was pre-activated, and consistent with the concept of mitochondrial stress leading to cell non-autonomous (ie. “remote”) effects, we found that restricted activation of the UPRmt in neurons decreased A-R death. However, expression of the atfs-1(gf) mutant in neurons, which resulted in a robust activation of a neuronal UPRmt, did not upregulate the UPRmt in distal tissues, nor did it protect the worms from A-R toxicity. These findings suggest that remote signaling requires additional component(s) acting downstream of de facto mitochondrial stress.
Citation: Peña S, Sherman T, Brookes PS, Nehrke K (2016) The Mitochondrial Unfolded Protein Response Protects against Anoxia in Caenorhabditis elegans. PLoS ONE 11(7): e0159989. https://doi.org/10.1371/journal.pone.0159989
Editor: Alexander A. Mongin, Albany Medical College, UNITED STATES
Received: May 11, 2016; Accepted: June 27, 2016; Published: July 26, 2016
Copyright: © 2016 Peña et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All data is available in the manuscript and supplemental table.
Funding: This work was supported in part by grants from the National Science Foundation (www.nsf.gov/): IOS 1352836 (KN), the National Institutes of Health (www.nih.gov/): R01HL127891 (KN and PSB), T32GM07356 (SP) and F31GM109496 (SP). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Abbreviations: UPRmt, mitochondrial unfolded protein response; ETC, electron transport chain; I-R, ischemia reperfusion; MosSCI, Mos-mediated Single Copy gene Insertion; RNAi, RNA interference; NGM, normal growth media; gf, gain-of-function; lf, loss-of-function; EtBr, ethidium bromide
The mitochondrial unfolded protein response (UPRmt) is an adaptive signaling pathway that was first identified in mammals , but has been best characterized genetically in the nematode Caenorhabditis elegans [2–5]. Misfolded proteins in the mitochondria trigger the expression of a repertoire of genes that relieve respiratory burden and restore proteostasis (for review, see reference ). Activation of the UPRmt occurs in response to disruptions in the stoichiometric equilibrium between nuclear and mitochondria-encoded electron transport chain (ETC) subunits , reactive oxygen species (ROS) [8, 9], and by exposure to pathogenic bacteria , suggesting that it is part of a surveillance mechanism that responds to disruption of core physiologic processes [8, 10–13]. Interestingly, mitochondrial stress in C. elegans neurons has been shown to activate the UPRmt in distal tissues and to regulate lifespan, leading to the idea that mitochondrial proteostasis is a sounding board that can trigger adaptation remotely through cell-cell communication .
In C. elegans, mito-nuclear retrograde signaling via the bZip transcription factor, ATFS-1, is a critical component of the UPRmt axis [15, 16]. ATFS-1 normally traffics to mitochondria, however during mitochondrial dysfunction a fraction of ATFS-1 fails to reach the mitochondria and instead accumulates in the nucleus. Additionally, the UPRmt is regulated by dimerization and nuclear localization of the dve-1 encoded transcription factor along with a small ubiquitin like protein ubl-5, as well as the activity of mTOR regulator, rheb-1. [2, 3]. Importantly, ATFS-1 coordinates mito-nuclear genomic output to properly balance the stoichiometry of electron transport chain complex assembly in order to maintain oxidative phosphorylation during mitochondrial stress . Several mitochondrial electron transport chain (ETC) mutants show UPRmt activation and require atfs-1 for survival  suggesting that this signaling pathway helps the organism to cope with mitochondrial dysfunction.
In the cellular stress imposed by oxygen deprivation, mitochondria are central to both death and survival (for review, see reference ). Significant events leading to death are irreversible inhibition of oxidative phosphorylation, proton leak across the mitochondrial inner membrane, calcium overload, reactive oxygen species (ROS) generation and permeability transition pore opening [18–23]. However, interventions that preserve mitochondria function, for example by scavenging ROS or inhibiting PT pore opening, are protective [24–26].
Recently, anoxia-reperfusion (A-R) has been shown to cause the accumulation of misfolded proteins in the mitochondria of C. elegans . A-R has been used extensively as a surrogate for ischemia-reperfusion in worms, due to their diminished sensitivity to low oxygen levels compared to mammals . Protecting the mitochondrial protein folding environment by simply over expressing mitochondrial molecular chaperones (Hsp-60 and mtHsp-70) can protect mammalian cells from ischemia-reperfusion (I-R) injury [29, 30] and inhibiting mitochondrial translation through knockdown of aminoacyl-tRNAs has been shown to protect against A-R injury in worms . Collectively, it seems that interventions that support mitochondrial proteostasis can protect against the detrimental effects of low oxygen. In fact, pre-activation of the UPRmt was recently shown to protect against anoxia in worms .
Herein, we build on the foundation that atfs-1 is required for protection from A-R injury, to show that atfs-1 gain-of-function (gf) mutants that exhibit constitutive activation of the UPRmt are endogenously protected. Based on this result, we hypothesized that protection might occur through intrinsic, cell autonomous processes or through remote effects that are coordinated throughout the organism (ie, “mitokine” signaling). To address this idea, we developed of a novel genetic model for single-copy cell restricted expression of an atfs-1(gf) mutant. Our results indicate that cell autonomous protection from A-R injury can be elicited through an atfs-1(gf), but that remote UPRmt activation and protection from death may require other pathway(s) that responds to de facto mitochondrial stress.
Materials and Methods
Worm strains were routinely propagated using standard culture techniques at 20°C on nematode growth media (NGM) agar plates containing 5 μg/ml cholesterol and seeded with OP50 bacteria, as described . The strains used in this work are:
SJ4058, zcIs9 [Phsp-60::GFP] V; CMH5, atfs-1(tm4525) V; QC115, atfs-1(et15) V; EG4322, ttTi5605 II; unc-119(ed9) III;; KWN176, rnyls13 [Pmec-4::mCherry; unc-119(+)] X; KWN 456, rnySi37 [Patfs-1::wCherry::atfs-1::GFP, cb-unc-119(+)] II; KWN 495, pha-1 (e2123ts) III, him-5 (e1490) V, rnyEx289 [Patfs-1::flp, Pmyo-2::mCherry, pha-1(+)]; KWN484, atfs-1(tm4525) V, rnySi37, zcIs9, rnyEx289; KWN525 rnySi49 [Patfs-1::wCherry::atfs-1(et15)::GFP, cb-unc-119(+)] II; unc-119(ed3) III; KWN529, pha-1 (e2123ts) III, him-5 (e1490) V, rnyEx297 [Prab-3::flp, Pmyo-2::mCherry, pha-1(+)]; KWN566, rnyls13, zcIs9; KWN567, atfs-1(tm4525), rnyls13, zcIs9; KWN568, atfs-1(et15), rnyls13, zcIs9; KWN562, atfs-1(tm4525), rnySi49, zcIs9, rnyEx297; KWN569, rnySi49, zcIs9, rnyEx297; KWN591, atfs-1(tm4525), rnySi49, zcIs9, rnyls013, rnyEx289; KWN590, rnySi49, zcIs9, rnyls013, rnyEx289; HC196, sid-1(qt9); AGD724, sid-1(qt9), uthIs243 [Prab-3::cco-1HP, Pmyo-2::tdTomato]
Standard genetic protocols were used for crosses. The zcIs9 insertion on chromosome V is closely linked to atfs-1. Hence, GFP signal was used to follow atfs-1 following mating. To confirm that the genomic atfs-1 allele co-segregated with zcIs9, the final strains were PCR genotyped using primers 5’-GCCTCCTTTCGCCTTTTGTCATC-3’, 5’- GCACAGCTTCTCCGATTCAGTG-3’, and 5’-gggggatttttagtcggcaatg-3’. These primers do not amplify the chromosome II MosSCI insertions. PCR products were sequenced to identify the atfs-1(et15) allele, which is a missense mutation. MosSCI insertions were confirmed using primers OG967, 5’-AGGCAGAATGTGAACAAGACTCG-3’ and OG970, 5’-ATCGGGAGGCGAACCTAACTG-3’, as described .
The atfs-1 promoter (~2.5 kb) was PCR amplified from genomic DNA using primers 5’-ACACGTCGACCAAACAATTTGATGGTACTGTTTCAGAT-3' and 5'-ACACCCTAGGCGAAGTTACACCTGCAAATGTACAAG-3' and cloned into pFH6.II-linker  as an AvrII-SalI fragment to create pSEP1. The 3’ untranslated region from atfs-1 was PCR amplified from genomic DNA using primers 5’-ACACGCGGCCGCCAATAATCAGAATTCGAAACAATTGTTC-3’ and 5’-ACACACTAGTCCTAGGTTGGCTAAACAGGTAACG-3’ and inserted into pSEP1 as a NotI-SpeI fragment to create pSEP2. A SalI-Acc65I FRT-wCherry-FRT fragment was cloned from pWD176-myo-2-flop-wCherry  into pSEP2 to create pSEP3. The atfs-1 open reading frame was PCR amplified from wild-type genomic DNA with primers 5’-ACACGGTACCAAAAATGTTTTCCCGTGTGGGACGTCTC-3’ and 5’-ACACGGTACCTGAATAATGGCGCCCATTTTACGAAG-3’ and inserted into pSEP3 as an Acc65I fragment to create pSEP4. The atfs-1(et15) open reading frame was PCR amplified using identical primers and inserted into pSEP3 to create pSEP6. pSEP4 and pSEP6 were digested with AvrII-SpeI and cloned into the SpeI site in the MosSCI vector pCFJ151-p5605  to create pSEP5 and pSEP7, respectively. The AgeI-AvrII fragment from pWD79-hsp-2u-flp-RV (obtained courtesy of Dr. Wayne Davies) was cloned into the AgeI-SpeI sites of pFH6.II in place of GFP to create pFLF1. The rab-3 promoter was PCR amplified from genomic DNA using primers 5’-ATATGTCGACCAGCCGCAATCTGAAAATAGGGCTACTGTAG-3’ and 5’-ATATCCTAGGGACGACGACGACCTCGACGG-3’ and cloned into pFLF1 as a SalI-AvrlI fragment to create pSEP10. The atfs-1 promoter was also cloned into pFLF1 to create pSEP11. All PCR-amplified fragments were sequenced to verify fidelity.
Transgenesis and MosSCI Single Copy Gene Insertion
Injection of DNA into the gonads of young adult animals was performed as described . The protocol for single copy insertion has been similarly well described [33, 36]. Briefly, a strain containing the Mos transposon ttTi6505 in chromosome II was injected with pSEP5 or pSEP7, along with vectors coding for positive and negative selectable markers and an inducible Mos transposase. Following recombination between pSEP5 or pSEP7 and the ttTi5605 site, the chromosome II integration locus was PCR amplified and sequenced to confirm the presence of a full-length insert. All resulting strains were backcrossed to wildtype N2 prior to use.
Developmentally-synchronized populations of hermaphrodites were established by using alkaline-bleach treatment (20% commercial bleach, 500 mM NaOH) to generate embryos from gravid adults. These embryos were then cultured under standard conditions at 20°C until adulthood, which was defined as 1 day after vulva development was complete, then subject to anoxia. For experiments examining axonal degeneration, staged L1/L2 larvae (24 hours post-hatch) were used in place of adults. Experimental anoxia was established as described . Briefly, the worms were washed 3x in M9 solution (22 mM KH2PO4, 42 mM Na2HPO4, 86 mM NaCl, 1mM MgSO4, pH 7) and transferred to an anaerobic chamber (Coy Lab Products, USA) equipped with a palladium catalyst to scavenge oxygen (O2<0.01 ppm) for 20 h under 95% N2 and 5% H2 at ~26°C. Worms were scored for viability 20–24 hours after anoxic exposure as previously described . In general, 2–6 replicates (or individual plates) were assayed per day per condition. The percentage lethality from each plate was then averaged to create a single data point, representing from 49–946 worms (average of 223 worms per data point; see S1 Table for raw data). Experimental manipulations were always assessed in parallel with controls, in order to normalize to index lethality, which fluctuated on a daily basis and was the greatest source of experimental variability. For presentation purposes, the values from individual paired trials were plotted side-by-side, together with their averages.
Synchronized larval stage 2/3 worms were moved to plates supplemented with transformed HT115 E.coli bacteria where an spg-7 insert flanked by T7 promoter sites or empty vector control (pPD129.36) was induced using IPTG. Worms were incubated on these plates for at least 24 hours prior to experimental use.
Worms were imaged on a 2% agarose pad under anesthetic (0.1% tetramisole). Neuronal puncta were visualized from transgenic worms expressing a red mCherry fusion in mechanosensory neurons, in the L2–L3 stage as previously described , using a Nikon Eclipse TE2000-U microscope (Nikon USA, Melville, NY, USA) with a Polychrome V monochromator (TILL Photonics, Gräfelfing, Germany) running TILLvisION software. Neuronal cell bodies were identified under a Texas Red filter set using a 10x objective, then viewed at higher magnification, typically with a 100x oil objective, to assess axonal breaks and abnormalities. Most of these consisted of puncta where the neuronal process was visibly enlarged and at least 2-fold more fluorescent than the surrounding process. The number of abnormalities in either the PVML or PVMR, one of two mechanosensory neurons in the posterior of the worm, was scored in a region extending from the cell body to the vulva. These processes under normal conditions are extremely homogenous with respect to the distribution of fluorescent reporter. Since they run close to the surface or “skin” of the worm, they are easily scored. In general, ten worms were analyzed per day, the data was averaged, and the experiments were repeated 3–4 times (see S1 Table for raw data). Other strains expressing various gene fusions, reporters, or fluorescent markers were examined using an Olympus FV1000 confocal microscope (available as part of the University of Rochester Confocal Core) under appropriate illumination.
Mitochondria are pivotal for determining survival following exposure to low oxygen levels, and ROS generated under these conditions can lead to both damage and adaptation. Worms are resistant to hypoxia  and hence A-R has been utilized to simulate I-R in this genetic model organism [38, 40]. In addition, worms are normally cultivated from 15–25°C and the A-R model utilizes a modestly elevated temperature (26°C) to sensitize the worms to anoxic toxicity. Hence, it is possible that interventions that modulate A-R toxicity may have their basis in thermal tolerance. However, this is true in mammalian systems, as well, where I-R injury is also temperature sensitive, and there is a large body of evidence suggesting that A-R translates well as a surrogate for I-R in mammalian cells (for review, see reference ). Here, we utilized C. elegans to determine the influence of UPRmt activation on surviving A-R.
The UPRmt reporter Phsp-60::GFP is strongly activated by RNAi mediated knockdown of spg-7 (spg-7(RNAi)), a mitochondrial quality control m-AAA protease (Fig 1A). A priori treatment with spg-7(RNAi) reduced death following A-R (vector control 71 ± 7%; spg-7(RNAi) 49 ± 8%, error is SEM, p = 0.02; Fig 1D), as did 30 mcg/mL EtBr, an established chemical activator of the UPRmt (S1 Fig). Protection elicited from both regiments was lost in the atfs-1(tm4525) loss-of-function (lf) mutant (atfs-1(tm4525) 58 ± 14%; spg-7(RNAi) 73 ± 11%; error is SEM, p = 0.14; Fig 1E).
Representative fluorescent photomicrographs of Phsp60::GFP activation in A.) vector control and B.) UPRmt inducer spg-7(RNAi) treated worms and in C.) an atfs-1(et15) gain-of-function mutant. D.) Box-and-whisker plot showing that pre-induction of the UPRmt with spg-7(RNAi) reduces death following A-R (n = 4, p*[Student’s t test] <0.05). E.) Box-and-whisker plot showing that spg-7(RNAi) does not protect atfs-1(tm4525) loss-of-function worms from A-R toxicity (n = 5, p*[Student’s t test] <0.05). F.) Box-and-whisker plot showing that atfs-1(et15) mutants are constitutively less susceptible to A-R (n = 6, p*[Student’s t test] <0.05). Experimental replicates are shown as black dots; lines between replicates indicate that they were run on the same day. Grey diamonds are means with the error shown as standard deviations.
Next we examined whether activation of atfs-1 is sufficient for protection. We took advantage of forward genetic screen for mutations that protected against the toxic non-cholesterol effects of statins . The atfs-1(et15) gf allele contains a missense mutation in the first six amino acids that is predicted to reduce ATFS-1 mitochondrial import, and the resulting activation of the UPRmt was shown to protect against statin toxicity . As expected, atfs-1(et15) mutants exhibited constitutive induction of the UPRmt reporter gene hsp-60 (Fig 1C). The atfs-1(et15) mutants also exhibited decreased death following A-R (wildtype 80 ± 7%; atfs-1(et15) 36 ± 10%, error is SEM, p = 0.006; Fig 1F).
We then determined whether induction of the UPRmt could protect tissues from sub-lethal A-R injury. The fluorescent reporter mCherry was expressed in mechanosensory neurons under control of the mec-4 promoter in order to measure axonal damage  (Fig 2A). Process abnormalities including breaks and “bead-on-a-string” type puncta (Fig 2B) have been reported to occur following A-R in these neurons . Following A-R in young larva, which are less susceptible to A-R than adults and hence were used here as a sub-lethal model, atfs-1(gf) mutants formed fewer axonal puncta than control animals (wildtype 2.4 ± 0.6; atfs-1(et15) 1.4 ± 0.3, error is SD, p = 0.02; Fig 2C). While the difference in puncta between genotypes was small, which is expected given the nature of the model, the data is robust and the difference statistically significant. Previous reports have suggested that upwards of 10 puncta may form following A-R ; however, in our hand the only worms that exhibited this many abnormalities were dead or close to death and hence unsuitable for analysis. It is also possible that the differences we observe are due to labeling with mCherry rather than GFP, as was used previously .
A.) Schematic representation of mechanosensory neurons (red dots indicate soma) in the worm, and the PLM axonal processes that were used to measure degeneration. B.) Representative fluorescent photomicrograph of an integrated Pmec-4::mCherry reporter in wildtype worms (top image) and following A-R (bottom image). Scored puncta are denoted by stars. C.) atfs-1(et15) worms develop fewer puncta following A-R (n = 4 independent trials with ~10 worms per genotype per trial), p*[Student’s t test] <0.05). The error is the standard deviation.
Neuronal mitochondrial stress can be sensed remotely, suggesting that cells are able to coordinate adaptive responses based upon communicating the status of their mitochondria (Fig 3A) . We found that a small hairpin loop RNA that targets complex IV (cco-1HP) expressed under the control of the pan-neuronal rab-3 promoter decreased A-R death (wildtype 59 ± 12%; cco-1HP 30 ± 9%, error is SEM, p = 0.03; Fig 3B). This suggests that neurons are able to coordinate remote adaptation among the whole organism in response to intrinsic mitochondrial stress. However, the cco-1HP model presumably combines de facto mitochondrial proteotoxicity through ROS generation with compensatory activation of the UPRmt. As such, we were motivated to ask whether neuronal expression of the atfs-1(gf) mutant, which exhibits compensation in the absence of stress, would also increase survival.
A.) Schematic of how cco-1HP interference with Complex IV leads to mitochondrial stress in neurons, which is transduced remotely through an unidentified “mitokine” to elicit adaptive signaling measures in distal cells. B.) Box-and-whisker plot showing that neuronal cco-1HP reduces death following A-R (n = 5, p*[Student’s t test] <0.05). Experimental replicates are shown as black dots; lines between replicates indicate that they were run on the same day. Grey diamonds are means with the error shown as standard deviations.
To achieve this, we generated a novel genetic system that combines two techniques for restricting gene expression. Transgenesis in worms is generally achieved through multi-copy extra-chromosomal arrays that drive recombinant expression. However, the UPRmt reflects a balance between mitochondrial and nuclear ATFS-1 targeting, which is likely to be prone to overexpression artifacts. Moreover, cell-specific promoters exhibit a range of temporal and quantitative differences in expression levels. To circumvent these issues, we first generated a genetic model where FLP recombinase is used to activate gene expression . A recombinant atfs-1 transgene was cloned where the promoter is interrupted by a wCherry insert that is itself flanked by FLP recombinase sites (Fig 4A). This transgene was then integrated via Mos-mediated single copy gene insertion (MosSCI) into the ttTi5605 site on chromosome II . Hence, transgenic ATFS-1 expression should occur at physiological levels from its own promoter but only in a subset of tissues defined by the overlap with a promoter chosen to drive recombinant FLP expression.
A.) Schematic showing relevant genetic constructs to restrict atfs-1 expression using transgenic FLP recombinase and an atfs-1 promoter interruption. The activity of the flp and atfs-1 gene products on their target promoters is indicated by arrows. B.) Representative transmission and fluorescent photomicrographs of Patfs-1::wCherry expressed from a MosSCI single-copy, integrated atfs-1 promoter interruption in the absence of FLP recombinase. C-E.) Photomicrographs of DIC (grayscale), a fluorescent marker of transgenic FLP (red) and UPRmt reporter Phsp-60::GFP (green) in flp(-) and F-H.) flp(+) siblings following spg-7(RNAi). The genomic background is atfs-1(tm4525), a loss-of-function allele.
Under basal conditions in the absence of FLP recombinase, wCherry was widely expressed at low levels (long exposure, Fig 4B), suggesting that the atfs-1 promoter was active. When the MosSCI atfs-1(+) transgene was placed into a genomic atfs-1(lf) background, the worms were unable to activate the UPRmt Phsp-60::GFP reporter in response to spg-7(RNAi) (Fig 4C), suggesting effective suppression of ATFS-1 expression by the wCherry promoter interruption. In contrast, expression of FLP recombinase from the atfs-1 promoter itself restored the ability to activate the UPRmt Phsp-60::GFP reporter (Fig 4C), suggesting effective de-suppression through wCherry excision.
In an attempt to discern regulated trafficking of the ATFS-1 transcription factor, a translational GFP fusion was expressed in the MosSCI system, but there was no detectable fluorescent signal even under conditions where we expected to observe increased nuclear trafficking (S3 Fig). We hypothesize that this is likely due to single copy gene expression and the presumed basal degradation of ATFS-1, as its physiologic effects in the nucleus are clearly observed (Fig 4C).
Next, the atfs-1(et15) gf missense mutation was incorporated into the MosSCI transgene, to generate a cell-specific compensatory response in the absence of stress. When FLP recombinase was expressed using the pan-neural promoter rab-3 (Fig 5A), we observed robust activation of the UPRmt hsp-60::GFP reporter gene in neurons (Fig 5C–5H). We predicted that remote activation of the UPRmt would require atfs-1 acting in distal tissues, and so we expressed the MosSCI/FLP transgene in a genomic atfs-1(lf) mutant background as a presumed negative control. However, we found that regardless of whether the MosSCI atfs-1(gf) transgene was in a genomic atfs-1(lf) or atfs-1(+) background, the UPRmt was not activated in distal tissues (Fig 5E and 5H). The lack of remote UPRmt activation in our model was not due to an inability of distal tissues to sense mitochondrial proteostasis, as they responded normally to EtBr (S2 and S3 Figs). Hence, restricted expression of ATFS-1(gf) can elicit a cell-autonomous UPRmt, but cannot trigger its remote activation in other cells.
A.) Schematic showing the relevant genetic constructs and the activity of the flp and atfs-1 gene products on their target promoters. B.) Schematic representing UPRmt activation in neurons (green) in the presence of Prab-3::FLP (red). The intestine is denoted in pink, as labeled, and activation of the UPRmt is notably absent in this tissue. C-E.) Transmission and fluorescent photomicrographs of cell-autonomous UPRmt activation in a genomic atfs-1(+) background and in a F-H.) atfs-1(4525) background. Note that distal tissues do not express the Phsp-60::GFP transgene, regardless of genetic background.
Finally, we asked whether the cell-autonomous UPRmt activation observed in neurons reduced local A-R injury. The Pmec4::mCherry reporter that was used to label mechanosensory neurons and their processes fluoresced much more brightly than the low level wCherry expressed from the single-copy promoter interruption (Fig 6B). We found that mechanosensory neurons from worms that exhibited FLP-mediated UPRmt reporter activation (Fig 6A) were protected against axonal damage compared to their non-transgenic brood mates (flp null 2.6 ± 1.1; Prab-3:flp 1.7 ± 0.9, error is the SD, p = 0.04; Fig 6D). However, lethality following A-R was not influenced by the presence of the FLP transgene and did not correspond to UPRmt reporter activation in neurons (atfs-1(+) background: flp null 73 ± 11%, Prab-3:flp 69 ± 13%, error is SEM, p = 0.69; Fig 6E). These results coincide with the lack of remote activation of the UPRmt reporter (Fig 5E and 5H; S2 Fig), and hence, we hypothesize that neuronal mitochondrial stress itself is required to synchronize adaptive responses in other tissues.
Representative fluorescent photomicrographs of a FLP out atfs-1(et15) transgenic worm with neuronal FLP (Prab-3::FLP) expressing A.) the UPRmt reporter Phsp60::GFP (green) and B.) the neuronal reporter Pmec-4::mCherry in mechanosensory neurons. Labeling of the PLM cell processes can be clearly distinguished over background Patfs-1::wCherry from the un-excised promoter interruption in other cells. C.) Merge indicating co-localization of UPRmt activation and Pmec-4::mCherry. D.) Expression of the FLP out atfs-1(et15) MosSCI transgene attenuates axonal damage following A-R (n = 3 independent trials of ~10 worms per genotype per trial), p*[Student’s t test] <0.05). The error is the standard deviation. E.) Box-and-whisker plot indicating that neuronal FLP out atfs-1(et15) MosSCI transgene does not reduce death following A-R (n = 5, p*[Student’s t test] <0.05). All data was obtained in an atfs-1(+) genomic background. Experimental replicates are shown as black dots; lines between replicates indicate that they were run on the same day. Grey diamonds are means with the error shown as standard deviations.
It is clear from our results that pharmacologic or genetic pre-activation of the UPRmt protects against 1.) A-R toxicity and 2.) axonal degeneration in neurons subject to sub-lethal A-R. We also find that mitochondrial dysfunction elicited through RNAi targeting of the ETC complex IV subunit cco-1 in neurons can coordinate stress responses remotely to decrease A-R toxicity. Somewhat surprisingly however, neuronal expression of an atfs-1(gf) mutant did not reduce death, although it did function intrinsically to protect neurons themselves from axonal damage. Our results suggest that neuronal survival is not the major determinant of A-R survival and that de facto mitochondrial stress is required to coordinate remote stress responses through an unknown signal(s). However, our results do not discount a role for atfs-1 acting in parallel to this unidentified signal, or being required in the target tissues independent of the signaling tissue. It’s interesting that the homeodomain-containing transcription factor DVE-1 and the small ubiquitin-like protein UBL-5 are both encoded by genes required for signaling the UPRmt  and have been shown to work in parallel to atfs-1. In addition, mitochondrial function has been shown to regulate mTORC1 via ROS signaling (for review, see reference ), which together with its well established sensitivity to ATP may be central for metabolic adaptation and anoxic resistance and could potentially intertwine with UPRmt signaling.
The UPRmt has been suggested to be a surveillance response, whereby interruptions in core physiologic processes such as ETC function coordinate adaptation throughout the organism . It seems intuitive that surveillance would best be served by a sentinel that’s selectively vulnerable to the type of damage that’s being detected (ie. a “first responder”). We speculate here that UPRmt activation in the germline (or stem cells?) may act as a sentinel for mitochondrial damage. Ethidium bromide, which causes mitochondrial stress and induces the UPRmt, has been shown to preferentially affect germline maintenance in C. elegans. There are multiple precedents for germline-soma signaling. For example, translation (like mitochondrial function) is monitored as a surveillance mechanism , and lipid signals generated from the germline in response to translation inhibition can induce stress signaling in distal tissues . The germline is well recognized to generate signals that lead to metabolic remodeling of somatic tissues [46–48]. Within this context, it’s quite interesting that stem cell maintenance also requires a UPRmt mediated metabolic checkpoint . Finally, we note that a previous report has shown that glp-1 mutants which do not generate oocytes have altered fat stores and are greatly protected from A-R .
Genetic manipulations that impair mitochondrial function in C. elegans also alter lifespan . Similarly, interventions that support protein folding are generally protective against low oxygen conditions and also extend lifespan . One common denominator in these processes is that they rely on mitochondrial ROS in one capacity or another. However, an important difference is highlighted by recent work indicating that atfs-1 is dispensable for cco-1 RNAi mediated longevity . In contrast, our work and that of others  indicate that protection from A-R injury via the UPRmt requires atfs-1 and, at least cell autonomously, that an ATFS-1(gf) that favors nuclear targeting is sufficient to reduce A-R death. Undoubtedly there are multiple signaling pathways that respond to mitochondrial stress, and the genetic requirements for protection from stress may overlap, but are likely to have insult specific components.
The UPRmt like other compartment specific unfolded protein responses is aimed at restoring proteostasis in the face of stress. A general concept that emerges from multiple studies of stress is that adaptive signaling is generally beneficial in the short term, but chronic activation of these pathways is maladaptive. However, it’s still unclear whether this holds true for the UPRmt. For example, it’s been established that pre-activation of the UPRmt protects against statin toxicity  and Pseudomonas infection . In addition, several C. elegans mutants in electron transport chain (ETC) subunits require atfs-1 for survival , suggesting that at least within the context of mitochondrial dysfunction any potential drawbacks of chronic activation are outweighed by the benefits. So, can we conclude that the UPRmt is strictly beneficial?
Unfortunately, there are several observations suggesting that this is not as straightforward as would appear, and probably depends upon the degree of activation. Treatment with spg-7(RNAi) can cause larval maturation to stall, as can high concentrations of EtBr, both of which are means of activating the UPRmt. One might argue that these phenotypes result from excess mitochondrial dysfunction that cannot be compensated for by activation of the UPRmt. However, it is worth noting that the atfs-1(et15) gf mutant is generally smaller and matures more slowly than wildtype worms and displays reduced fecundity . In addition, worms expressing an a nuclear targeted ATFS-1 with a deletion of the mitochondrial targeting sequence(Δ1-32.myc ATFS-1) similarly develop slower than worms expressing a wildtype ATFS-1 . This may result from chronic nuclear activation, or alternatively from reduced ATFS-1 function in the mitochondria, whose genome has also been shown to be regulated by this transcription factor . Hence, while there’s reason to believe that the UPRmt may be primarily beneficial, further work is required to delineate the downside of chronic activation.
Collectively, our results demonstrate that atfs-1 and the UPRmt can influence A-R damage and survival. We further hypothesize that mitochondrial stress activates an as-of-yet unidentified signaling pathway(s) to coordinate adaptive responses throughout the organism. While the mammalian ATFS-1 ortholog has yet to be identified, the similarities between the UPRmt in mammals and worms (for review, see reference  and the strong protection from A-R elicited by activation of the UPRmt warrant further work on this topic.
S1 Fig. UPRmt activation by ethidium bromide reduces death following anoxia-reperfusion and requires atfs-1.
A.) Box-and-whisker plot of A-R toxicity following treatment with 30 μg/ml EtBr to activate the UPRmt (n = 8, p*[Student’s t test] <0.05) B.) atfs-1(tm4525) mutants fail to exhibit EtBr mediated protection (n = 5, p*[Student’s t test] <0.05). Grey diamonds are means with the error shown as standard deviations.
S2 Fig. Distal tissues are capable of mounting an UPRmt in the neuron specific atfs-1(gf) FLP-out transgenic model.
Representative fluorescent photomicrographs of MosSCI FLP-out atfs-1(gf) transgenic worms grown on (A, C.) control plates and (B, D.) plates containing 30μg/mL EtBr. The transgenic marker for FLP (Pmyo-2::mCherry) is shown in panels A and B, while activation of the UPRmt reporter Phsp-60::GFP is shown in panels C and D. The genomic background is atfs-1(+). These data indicate that failure to activate distal tissue stress responses in this strain does not result from an intrinsic dysfunction.
S3 Fig. Distal tissues remain responsive to UPRmt inducers in the neuron specific FLP-out transgenic model.
Representative photomicrographs of MosSCI FLP-out atfs-1(gf) transgenic worms grown on (A, C, F, H.) control plates and (B, D, E, G, I, J) plates containing 30μg/mL EtBr. Transmitted light images are shown in panels A-E, while fluorescent images are shown in panels F-J. Relevant information is labeled at the top of the figure. Note that the strain in panels E and J lacks the UPRmt reporter Phsp-60::GFP and that the MosSCI FLP-out atfs-1(gf) is fused to GFP. These panels have been overexposed. The failure to detect either mitochondrial or nuclear GFP (the green is autofluorescence) exemplifies the low expression level of a single copy transgene and its likely degradation under basal conditions.
Several strains were provided by the C. elegans Gene Knockout Consortium, the Japanese National Bioresource Project and the C. elegans Genetics Center. Dr. Cole Haynes and Dr. Andrew Dillin also provided genetic reagents, for which we are grateful. Dr. Yun-Ki Im provided assistance and advice and Dr. Andrew Wojtovich offered helpful discussions.
Conceived and designed the experiments: SP TS PSB KN. Performed the experiments: SP TS. Analyzed the data: SP TS KN. Wrote the paper: SP KN.
- 1. Zhao Q, Wang J, Levichkin IV, Stasinopoulos S, Ryan MT, Hoogenraad NJ. A mitochondrial specific stress response in mammalian cells. The EMBO journal. 2002;21(17):4411–9. Epub 2002/08/29. pmid:12198143; PubMed Central PMCID: PMC126185.
- 2. Benedetti C, Haynes CM, Yang Y, Harding HP, Ron D. Ubiquitin-like protein 5 positively regulates chaperone gene expression in the mitochondrial unfolded protein response. Genetics. 2006;174(1):229–39. pmid:16816413; PubMed Central PMCID: PMC1569816.
- 3. Haynes CM, Petrova K, Benedetti C, Yang Y, Ron D. ClpP mediates activation of a mitochondrial unfolded protein response in C. elegans. Developmental cell. 2007;13(4):467–80. Epub 2007/10/11. pmid:17925224.
- 4. Haynes CM, Yang Y, Blais SP, Neubert TA, Ron D. The matrix peptide exporter HAF-1 signals a mitochondrial UPR by activating the transcription factor ZC376.7 in C. elegans. Molecular cell. 2010;37(4):529–40. pmid:20188671; PubMed Central PMCID: PMC2846537.
- 5. Yoneda T, Benedetti C, Urano F, Clark SG, Harding HP, Ron D. Compartment-specific perturbation of protein handling activates genes encoding mitochondrial chaperones. Journal of cell science. 2004;117(Pt 18):4055–66. pmid:15280428.
- 6. Lin YF, Haynes CM. Metabolism and the UPR(mt). Molecular cell. 2016;61(5):677–82. pmid:26942672; PubMed Central PMCID: PMC4779188.
- 7. Houtkooper RH, Mouchiroud L, Ryu D, Moullan N, Katsyuba E, Knott G, et al. Mitonuclear protein imbalance as a conserved longevity mechanism. Nature. 2013;497(7450):451–7. Epub 2013/05/24. pmid:23698443; PubMed Central PMCID: PMC3663447.
- 8. Runkel ED, Liu S, Baumeister R, Schulze E. Surveillance-activated defenses block the ROS-induced mitochondrial unfolded protein response. PLoS genetics. 2013;9(3):e1003346. pmid:23516373; PubMed Central PMCID: PMC3597513.
- 9. Rauthan M, Ranji P, Abukar R, Pilon M. A Mutation in Caenorhabditis elegans NDUF-7 Activates the Mitochondrial Stress Response and Prolongs Lifespan via ROS and CED-4. G3. 2015;5(8):1639–48. pmid:26038366; PubMed Central PMCID: PMC4528320.
- 10. Pellegrino MW, Nargund AM, Kirienko NV, Gillis R, Fiorese CJ, Haynes CM. Mitochondrial UPR-regulated innate immunity provides resistance to pathogen infection. Nature. 2014;516(7531):414–7. pmid:25274306; PubMed Central PMCID: PMC4270954.
- 11. Dunbar TL, Yan Z, Balla KM, Smelkinson MG, Troemel ER. C. elegans detects pathogen-induced translational inhibition to activate immune signaling. Cell host & microbe. 2012;11(4):375–86. pmid:22520465; PubMed Central PMCID: PMC3334869.
- 12. Liu Y, Samuel BS, Breen PC, Ruvkun G. Caenorhabditis elegans pathways that surveil and defend mitochondria. Nature. 2014;508(7496):406–10. pmid:24695221; PubMed Central PMCID: PMC4102179.
- 13. Govindan JA, Jayamani E, Zhang X, Breen P, Larkins-Ford J, Mylonakis E, et al. Lipid signalling couples translational surveillance to systemic detoxification in Caenorhabditis elegans. Nature cell biology. 2015;17(10):1294–303. pmid:26322678; PubMed Central PMCID: PMC4589496.
- 14. Durieux J, Wolff S, Dillin A. The cell-non-autonomous nature of electron transport chain-mediated longevity. Cell. 2011;144(1):79–91. pmid:21215371; PubMed Central PMCID: PMC3062502.
- 15. Nargund AM, Fiorese CJ, Pellegrino MW, Deng P, Haynes CM. Mitochondrial and nuclear accumulation of the transcription factor ATFS-1 promotes OXPHOS recovery during the UPR(mt). Molecular cell. 2015;58(1):123–33. pmid:25773600; PubMed Central PMCID: PMC4385436.
- 16. Nargund AM, Pellegrino MW, Fiorese CJ, Baker BM, Haynes CM. Mitochondrial import efficiency of ATFS-1 regulates mitochondrial UPR activation. Science. 2012;337(6094):587–90. pmid:22700657; PubMed Central PMCID: PMC3518298.
- 17. Murphy E, Steenbergen C. Mechanisms underlying acute protection from cardiac ischemia-reperfusion injury. Physiological reviews. 2008;88(2):581–609. pmid:18391174; PubMed Central PMCID: PMC3199571.
- 18. Hardy DL, Clark JB, Darley-Usmar VM, Smith DR. Reoxygenation of the hypoxic myocardium causes a mitochondrial complex I defect. Biochemical Society transactions. 1990;18(4):549. pmid:1703502.
- 19. Asimakis GK, Conti VR. Myocardial ischemia: correlation of mitochondrial adenine nucleotide and respiratory function. Journal of molecular and cellular cardiology. 1984;16(5):439–47. pmid:6737484.
- 20. Brookes PS, Yoon Y, Robotham JL, Anders MW, Sheu SS. Calcium, ATP, and ROS: a mitochondrial love-hate triangle. American journal of physiology Cell physiology. 2004;287(4):C817–33. pmid:15355853.
- 21. Borutaite V, Mildaziene V, Brown GC, Brand MD. Control and kinetic analysis of ischemia-damaged heart mitochondria: which parts of the oxidative phosphorylation system are affected by ischemia? Biochimica et biophysica acta. 1995;1272(3):154–8. pmid:8541346.
- 22. Garcia-Rivas Gde J, Carvajal K, Correa F, Zazueta C. Ru360, a specific mitochondrial calcium uptake inhibitor, improves cardiac post-ischaemic functional recovery in rats in vivo. British journal of pharmacology. 2006;149(7):829–37. pmid:17031386; PubMed Central PMCID: PMC2014682.
- 23. Paradies G, Ruggiero FM, Petrosillo G, Quagliariello E. Peroxidative damage to cardiac mitochondria: cytochrome oxidase and cardiolipin alterations. FEBS letters. 1998;424(3):155–8. pmid:9539141.
- 24. Horwitz LD, Fennessey PV, Shikes RH, Kong Y. Marked reduction in myocardial infarct size due to prolonged infusion of an antioxidant during reperfusion. Circulation. 1994;89(4):1792–801. pmid:8149545.
- 25. Piot C, Croisille P, Staat P, Thibault H, Rioufol G, Mewton N, et al. Effect of cyclosporine on reperfusion injury in acute myocardial infarction. The New England journal of medicine. 2008;359(5):473–81. pmid:18669426.
- 26. Argaud L, Gateau-Roesch O, Muntean D, Chalabreysse L, Loufouat J, Robert D, et al. Specific inhibition of the mitochondrial permeability transition prevents lethal reperfusion injury. Journal of molecular and cellular cardiology. 2005;38(2):367–74. pmid:15698843.
- 27. Kaufman DM, Crowder CM. Mitochondrial Proteostatic Collapse Leads to Hypoxic Injury. Current biology: CB. 2015;25(16):2171–6. pmid:26234215.
- 28. Van Voorhies WA, Ward S. Broad oxygen tolerance in the nematode Caenorhabditis elegans. The Journal of experimental biology. 2000;203(Pt 16):2467–78. pmid:10903161.
- 29. Belke DD, Gloss B, Hollander JM, Swanson EA, Duplain H, Dillmann WH. In vivo gene delivery of HSP70i by adenovirus and adeno-associated virus preserves contractile function in mouse heart following ischemia-reperfusion. American journal of physiology Heart and circulatory physiology. 2006;291(6):H2905–10. pmid:16920815.
- 30. Lin KM, Lin B, Lian IY, Mestril R, Scheffler IE, Dillmann WH. Combined and individual mitochondrial HSP60 and HSP10 expression in cardiac myocytes protects mitochondrial function and prevents apoptotic cell deaths induced by simulated ischemia-reoxygenation. Circulation. 2001;103(13):1787–92. Epub 2001/04/03. pmid:11282911.
- 31. Anderson LL, Mao X, Scott BA, Crowder CM. Survival from hypoxia in C. elegans by inactivation of aminoacyl-tRNA synthetases. Science. 2009;323(5914):630–3. Epub 2009/01/31. pmid:19179530.
- 32. Stiernagle T. Maintenance of C. elegans. WormBook: the online review of C elegans biology. 2006:1–11. pmid:18050451; PubMed Central PMCID: PMC4781397.
- 33. Frokjaer-Jensen C, Davis MW, Hopkins CE, Newman BJ, Thummel JM, Olesen SP, et al. Single-copy insertion of transgenes in Caenorhabditis elegans. Nature genetics. 2008;40(11):1375–83. pmid:18953339; PubMed Central PMCID: PMC2749959.
- 34. Hagen FK, Nehrke K. cDNA cloning and expression of a family of UDP-N-acetyl-D-galactosamine:polypeptide N-acetylgalactosaminyltransferase sequence homologs from Caenorhabditis elegans. The Journal of biological chemistry. 1998;273(14):8268–77. pmid:9525933.
- 35. Rieckher M, Kourtis N, Pasparaki A, Tavernarakis N. Transgenesis in Caenorhabditis elegans. Methods in molecular biology. 2009;561:21–39. pmid:19504062.
- 36. Frokjaer-Jensen C, Davis MW, Ailion M, Jorgensen EM. Improved Mos1-mediated transgenesis in C. elegans. Nature methods. 2012;9(2):117–8. pmid:22290181; PubMed Central PMCID: PMC3725292.
- 37. Wojtovich AP, DiStefano P, Sherman T, Brookes PS, Nehrke K. Mitochondrial ATP-sensitive potassium channel activity and hypoxic preconditioning are independent of an inwardly rectifying potassium channel subunit in Caenorhabditis elegans. FEBS letters. 2012;586(4):428–34. pmid:22281198; PubMed Central PMCID: PMC3288502.
- 38. Scott BA, Avidan MS, Crowder CM. Regulation of hypoxic death in C. elegans by the insulin/IGF receptor homolog DAF-2. Science. 2002;296(5577):2388–91. Epub 2002/06/18. pmid:12065745.
- 39. Queliconi BB, Marazzi TB, Vaz SM, Brookes PS, Nehrke K, Augusto O, et al. Bicarbonate modulates oxidative and functional damage in ischemia-reperfusion. Free radical biology & medicine. 2013;55:46–53. pmid:23195687; PubMed Central PMCID: PMC3995138.
- 40. Padilla PA, Nystul TG, Zager RA, Johnson AC, Roth MB. Dephosphorylation of cell cycle-regulated proteins correlates with anoxia-induced suspended animation in Caenorhabditis elegans. Molecular biology of the cell. 2002;13(5):1473–83. pmid:12006646; PubMed Central PMCID: PMC111120.
- 41. Rodriguez M, Snoek LB, De Bono M, Kammenga JE. Worms under stress: C. elegans stress response and its relevance to complex human disease and aging. Trends in genetics: TIG. 2013;29(6):367–74. pmid:23428113.
- 42. Rauthan M, Ranji P, Aguilera Pradenas N, Pitot C, Pilon M. The mitochondrial unfolded protein response activator ATFS-1 protects cells from inhibition of the mevalonate pathway. Proceedings of the National Academy of Sciences of the United States of America. 2013;110(15):5981–6. pmid:23530189; PubMed Central PMCID: PMC3625262.
- 43. Pinan-Lucarre B, Gabel CV, Reina CP, Hulme SE, Shevkoplyas SS, Slone RD, et al. The core apoptotic executioner proteins CED-3 and CED-4 promote initiation of neuronal regeneration in Caenorhabditis elegans. PLoS biology. 2012;10(5):e1001331. pmid:22629231; PubMed Central PMCID: PMC3358320.
- 44. Davis MW, Morton JJ, Carroll D, Jorgensen EM. Gene activation using FLP recombinase in C. elegans. PLoS genetics. 2008;4(3):e1000028. pmid:18369447; PubMed Central PMCID: PMC2265415.
- 45. Groenewoud MJ, Zwartkruis FJ. Rheb and mammalian target of rapamycin in mitochondrial homoeostasis. Open biology. 2013;3(12):130185. pmid:24352740; PubMed Central PMCID: PMC3877845.
- 46. Tsang WY, Lemire BD. Mitochondrial genome content is regulated during nematode development. Biochemical and biophysical research communications. 2002;291(1):8–16. pmid:11829454.
- 47. Ratnappan R, Amrit FR, Chen SW, Gill H, Holden K, Ward J, et al. Germline signals deploy NHR-49 to modulate fatty-acid beta-oxidation and desaturation in somatic tissues of C. elegans. PLoS genetics. 2014;10(12):e1004829. pmid:25474470; PubMed Central PMCID: PMC4256272.
- 48. Wang MC, O'Rourke EJ, Ruvkun G. Fat metabolism links germline stem cells and longevity in C. elegans. Science. 2008;322(5903):957–60. pmid:18988854; PubMed Central PMCID: PMC2760269.
- 49. Mohrin M, Shin J, Liu Y, Brown K, Luo H, Xi Y, et al. Stem cell aging. A mitochondrial UPR-mediated metabolic checkpoint regulates hematopoietic stem cell aging. Science. 2015;347(6228):1374–7. pmid:25792330; PubMed Central PMCID: PMC4447312.
- 50. Mendenhall AR, LeBlanc MG, Mohan DP, Padilla PA. Reduction in ovulation or male sex phenotype increases long-term anoxia survival in a daf-16-independent manner in Caenorhabditis elegans. Physiological genomics. 2009;36(3):167–78. pmid:19050081; PubMed Central PMCID: PMC2646460.
- 51. Dillin A, Hsu AL, Arantes-Oliveira N, Lehrer-Graiwer J, Hsin H, Fraser AG, et al. Rates of behavior and aging specified by mitochondrial function during development. Science. 2002;298(5602):2398–401. pmid:12471266.
- 52. Mabon ME, Mao X, Jiao Y, Scott BA, Crowder CM. Systematic identification of gene activities promoting hypoxic death. Genetics. 2009;181(2):483–96. pmid:19047414; PubMed Central PMCID: PMC2644942.
- 53. Bennett CF, Vander Wende H, Simko M, Klum S, Barfield S, Choi H, et al. Activation of the mitochondrial unfolded protein response does not predict longevity in Caenorhabditis elegans. Nature communications. 2014;5:3483. pmid:24662282; PubMed Central PMCID: PMC3984390.
- 54. Arnould T, Michel S, Renard P. Mitochondria Retrograde Signaling and the UPR mt: Where Are We in Mammals? International journal of molecular sciences. 2015;16(8):18224–51. pmid:26258774; PubMed Central PMCID: PMC4581242.