Mitochondrial unfolded protein response transcription factor ATFS-1 increases resistance to exogenous stressors through upregulation of multiple stress response pathways

The mitochondrial unfolded protein response (mitoUPR) is an evolutionarily conserved pathway that responds to various insults to the mitochondria through transcriptional changes that restore mitochondrial homeostasis in order to facilitate cell survival. Gene expression changes resulting from the activation of the mitoUPR are mediated by the transcription factor ATFS-1/ATF-5. To further define the mechanisms through which the mitoUPR protects the cell during mitochondrial dysfunction, we characterized the role of ATFS-1 in responding to organismal stress. We found that activation of ATFS-1 is sufficient to cause upregulation of genes involved in multiple stress response pathways, including the DAF-16-mediated stress response pathway, the SKN-1-mediated oxidative stress response pathway, the HIF-mediated hypoxia response pathway, the p38-mediated innate immune response pathway, and antioxidant genes. Moreover, ATFS-1 is required for the upregulation of stress response genes after exposure to exogenous stressors, especially oxidative stress and bacterial pathogens. Constitutive activation of ATFS-1 increases resistance to multiple acute exogenous stressors, while disruption of atfs-1 decreases stress resistance. Although ATFS-1-dependent genes are upregulated in multiple long-lived mutants, constitutive activation of ATFS-1 in wild-type animals results in decreased lifespan. Overall, our work demonstrates that ATFS-1 serves a vital role in organismal survival of acute stresses through its ability to activate multiple stress response pathways, but that chronic ATFS-1 activation is detrimental for longevity.


26
The mitochondrial unfolded protein response (mitoUPR) is a stress response pathway 27 that acts to reestablish mitochondrial homeostasis through inducing transcriptional changes of 28 genes involved in metabolism and restoration of mitochondrial protein folding [1]. Various 29 perturbations to the mitochondria can activate mitoUPR, including excess reactive oxygen 30 species (ROS) and defects in mitochondrial import machinery [2]. The mitoUPR is mediated by 31 the transcription factor ATFS-1 (activating transcription factor associated with stress-1) in C.

32
elegans [3], or ATF5 in mammals [4]. 33 34 ATFS-1/ATF5 regulates mitoUPR through its dual targeting domains, a mitochondrial 35 targeting sequence (MTS) and a nuclear localization signal (NLS). Under normal unstressed 36 conditions, the MTS causes ATFS-1 to enter the mitochondria through the HAF-1 import 37 channel. Inside the mitochondria, ATFS-1 is degraded by the Lon protease CLPP-1/CLP1 [3]. In order to study the role of the mitoUPR in longevity, we previously disrupted atfs-1 in 44 long-lived nuo-6 mutants, which contain a point mutation that affects Complex I of the electron 45 transport chain [6]. nuo-6 mutants have a mild impairment of mitochondrial function that leads 46 to increased lifespan and enhanced resistance to multiple stressors. We found that loss of atfs-47 1 not only decreased the lifespan of nuo-6 worms, but also abolished the increased stress 48 resistance of these worms, thereby suggesting that ATFS-1 contributes to both longevity and 49 stress resistance in these worms [7]. 50 51 While a role for the mitoUPR in longevity has been reported [8][9][10][11], and debated [12, 52 13], little is known about the role of ATFS-1 in response to exogenous stressors. Pellegrino et al. 53 4 found that activation of ATFS-1 can increase organismal resistance to the pathogenic bacteria P.

57
In this study, we use C. elegans to define the relationship between ATFS-1 and 58 organismal stress resistance, and explore the underlying mechanisms. We find that activation of 59 ATFS-1 is sufficient to upregulate genes from multiple stress response pathways and is To activate ATFS-1, we used the nuo-6 mutation. We also examined gene expression in 79 two different gain-of-function (GOF) mutants with constitutively active ATFS-1: atfs-1(et15) and 80 atfs-1(et17). Both of these constitutively active ATFS-1 mutants have mutations in the MTS 81 causing increased nuclear localization of ATFS-1 [16]. We used a loss-of-function (LOF) atfs-1 82 deletion mutation (gk3094) to disrupt ATFS-1 function in wild-type and nuo-6 mutants.

92
We found that compared to wild-type worms, atfs-1(gk3094) deletion mutants did not 93 have decreased expression levels for the target genes of any of the stress response pathways 94 6 ( Fig. 1). This indicates that ATFS-1 is not required for the basal expression levels of these stress 95 response genes.

114
To gain a more comprehensive view of the extent to which mitoUPR activation causes 115 upregulation of genes in other stress response pathways, we compared genes upregulated in 116 the constitutively active atfs-1 mutant, atfs-1(et15), to genes upregulated by activation of 117 different stress response pathways. As a proof-of-principle, we first examined the overlap 118 between upregulated genes in atfs-1(et15) mutants and genes upregulated by activation of the 119 mitoUPR with spg-7 RNAi in an ATFS-1-dependent manner [3].  In comparing genes upregulated in the constitutively active atfs-1 mutant et15 to these 142 previously published gene lists, we found that 51% of genes upregulated by spg-7 RNAi in an 143 ATFS-1-dependent manner are also upregulated by constitutive activation of ATFS-1 ( Fig. 2A).

144
Similarly, we found a highly significant overlap of upregulated genes between atfs-1(et15) 145 mutants and each of the examined stress response pathways. We found that atfs-1(et15) had a 146 25% overlap with genes of ER-UPR pathway ( Having shown that constitutive activation of ATFS-1 can induce upregulation of genes involved 156 in various stress response pathways, we next sought to determine the role of ATFS-1 in the 157 genetic response to different stressors. To do this, we exposed wild-type animals and atfs-1 158 loss-of-function mutants (atfs-1(gk3094)) to six different types of stress and quantified the 159 resulting upregulation of stress response genes using quantitative RT-PCR (qPCR). We found 160 that exposure to either oxidative stress (4 mM paraquat, 48 hours) or the bacterial pathogen 161 Pseudomonas aeruginosa strain PA14 induced a significant upregulation of stress response 162 genes in wild-type worms, but that this upregulation was suppressed by disruption of atfs-1 163 ( Fig. 3A,B). In contrast, exposure to heat stress (35°C, 2 hours; Due to the crucial role of ATFS-1 in upregulating genes in multiple stress response pathways, we 172 next sought to determine the extent to which activating ATFS-1 protects against exogenous 173 stressors. To do this, we quantified resistance to stress in two constitutively active atfs-1 gain-174 of-function mutants (atfs-1(et15), atfs-1(et17)) compared to wild-type worms. For comparison, 175 we also included an atfs-1 loss-of-function deletion mutant (atfs-1(gk3094)), which we have 176 previously shown to have decreased resistance to oxidative, heat, osmotic and anoxic stress [7].

178
We measured resistance to acute oxidative stress by exposing worms to 300 µM 179 juglone. We found that both the gain-of-function mutants, atfs-1(et15) and atfs-1(et17), have 9 increased resistance to acute oxidative stress compared to wild-type worms, while atfs-181 1(gk3094) deletion mutants were less resistant compared to wild-type worms (Fig. 4A). To test 182 resistance to chronic oxidative stress, worms were transferred to plates containing 4 mM 183 paraquat beginning at day 1 of adulthood. Similar to the acute assay, atfs-1(et17) mutants were 184 more resistant to chronic oxidative stress, while atfs-1(gk3094) mutants were less resistant to 185 chronic oxidative stress compared to wild-type worms (Fig. 4B). Oddly, atfs-1(et15) gain-of-186 function mutants exhibited decreased resistance to chronic oxidative stress. wild-type worms, while atfs-1(gk3094) deletion mutants had decreased survival, although the 194 difference was only significant at 500 mM (Fig. 4D, E). Resistance to anoxic stress was measured 195 by placing worms in an oxygen-free environment for 75 hours, followed by a 24-hour recovery.

199
Lastly, to test resistance to bacterial pathogens, worms were exposed to Pseudomonas 200 aeruginosa strain PA14 in either a fast kill assay in which worms die from a toxin produced by 201 the bacteria, or a slow kill assay in which worms die due to the intestinal colonization of the 202 pathogenic bacteria [25]. In the fast kill assay, we found that constitutive activation of ATFS-1 203 increases survival in atfs-1(et15) and atfs-1(et17) mutants (Fig. 4G). We also observed increased 204 survival in atfs-1(gk3094) deletion mutants. For the slow kill assay, we used two established 205 protocols: one in which the assay is initiated at the L4 larval stage and performed at 25°C [14, 206 25, 26] and one in which the assay is initiated at day three of adulthood and performed at 20°C 207 [27]. Surprisingly, at 25°C, we found that atfs-1(et17) mutant had a small decrease in resistance 208 to PA14, while atfs-1(gk3094) mutants exhibited a small increase in resistance to PA14 209 compared to wild-type worms ( Fig. 4H). At 20°C, both atfs-1(gk3094) and atfs-1(et17) mutants 210 had a small increase in resistance to PA14 compared to wild-type worms (Fig. I).

212
All together, these data indicate that activation of ATFS-1 is sufficient to protect against 213 oxidative stress, osmotic stress, anoxia, and bacterial pathogens but not heat stress. They also 214 show that ATFS-1 is required for wild-type worms to survive oxidative stress, heat stress, 215 osmotic stress, and anoxia. After identifying genes that are differentially expressed in each of these long-lived 230 mutants, we compared the differentially expressed genes to genes that are upregulated by 231 ATFS-1 activation. We defined ATFS-1-upregulated genes in two ways: (1) genes that are 232 upregulated by spg-7 RNAi in an ATFS-1-dependent manner [3]; and (2) genes that are 233 upregulated in a constitutively active atfs-1 mutant (et15; [7]).

252
Overall, these results indicate that ATFS-1 target genes are upregulated in multiple long-253 lived mutants, including mutants in which mitochondrial function is not directly disrupted.  While ATFS-1 is not required for longevity in wild-type animals, it plays a significant role in 349 protecting animals against exogenous stressors. Disrupting atfs-1 function decreases 350 organismal resistance to oxidative stress, heat stress, osmotic stress, and anoxia (Fig. 4).

351
Additionally, we previously determined that inhibiting atfs-1 in long-lived nuo-6 worms 352 completely prevented the increased resistance to oxidative stress, osmotic stress and heat 353 stress typically observed in that mutant [7], and that disruption of atfs-1 in Parkinson's disease 354 15 mutants pdr-1 and pink-1 decreased their resistance to oxidative stress, osmotic stress, heat 355 stress, and anoxia [46]. Combined, these results demonstrate that ATFS-1 is required for 356 resistance to multiple types of exogenous stressors.

379
Overexpression of the mitoUPR target gene hsp-60 also increases resistance to P. aeruginosa In exploring the mechanism by which ATFS-1 and the mitoUPR modulate stress resistance, we 385 found that activation of ATFS-1, through either a mutation that mildly impairs mitochondrial 386 function (nuo-6) or through a mutation that constitutively activates ATFS-1 (atfs-1(et15)), 387 causes upregulation of genes involved in multiple stress response pathways including the ER-388 UPR pathway, the Cyto-UPR pathway, the DAF-16-mediated stress response pathway, the SKN-389 1-mediated oxidative stress response pathway, the HIF-mediated hypoxia response pathway, 390 the p38-mediated innate immune response pathway and antioxidant genes (Fig. 2). These The mitoUPR is required for animals to survive exposure to exogenous stressors, and activation 403 of this pathway is sufficient to enhance resistance to stress (Table S4). In addition to 404 upregulating genes involved in restoring mitochondrial homeostasis, the mitoUPR increases 405 stress resistance by upregulating the target genes of multiple stress response pathways.

406
Although increased stress resistance has been associated with long lifespan, and multiple long-407 lived mutants exhibit activation of the mitoUPR, constitutive activation of ATFS-1 shortens 408 lifespan while increasing resistance to stress, indicating that the role of ATFS-1 in stress 409 resistance can be experimentally dissociated from its role in longevity. Overall, this work 410 highlights the importance of the mitoUPR in not only protecting organisms from internal stress, 411 but also improving organismal survival upon exposure to external stressors.

507
For heat stress, young adult worms were incubated in 37°C and survival was measured every 2 508 hours for a total of 10 hours. For osmotic stress, young adult worms were transferred to plates 509 containing 450 mM or 500 mM NaCl and survival was measured after 48 hours.

510
For anoxic stress, plates with young adult worms were put into BD Bio-Bag Type A 511 Environmental Chambers for 75 hours and survival was measured after a 24-hour recovery 512 period.

513
Two different bacterial pathogenesis assays involving P. aeruginosa strain PA14 were 514 performed. In the slow kill assay worms are thought to die from intestinal colonization of the 515 pathogenic bacteria, while in the fast kill assay worms are thought to die from a toxin secreted 516 from the bacteria [25]. The slow kill assay was performed as described previously [14,27]. In 517 the first protocol [14], PA14 cultures were grown overnight and seeded to center of a 35-mm 518 NGM agar plate. Plates were left to dry overnight, and then incubated in 37°C for 24 hours.

519
Plates were left to adjust to room temperature before approximately 40 L4 worms were 520 transferred onto the plates. The assay was conducted 25°C and plates were checked twice a day 521 until death. In the second protocol [27], overnight PA14 culture were seeded to the center of a 522 35-mm NGM agar plate containing 20 mg/L FUdR. Plates were incubated at 37°C overnight, 523 then at room temperature overnight before approximately 40 day three adults were 524 transferred onto these plates. The assay was conducted 20°C and plates were checked daily 525 until death. The fast kill pathogenesis assay was performed as described previously [25]. PA14 526 22 cultures were grown overnight and seeded to Peptone-Glucose-Sorbitol (PGS) agar plates.

527
Seeded plates were left to dry for 20 minutes at room temperature before incubation at 37°C 528 for 24 hours and then at 23°C for another 24 hours. Approximately 30 L4 worms were 529 transferred onto the plates and were scored as dead or alive at 2, 4, 6, 8 and 24 hours. Fast kill 530 plates were kept at 23°C in between scoring timepoints. Acknowledgments. We would like to thank Paige Rudich for carefully reviewing the manuscript 549 prior to submission. Some strains were provided by the CGC, which is funded by NIH Office of 550 Research Infrastructure Programs (P30 OD010440). We would also like to acknowledge the C. To determine the role of ATFS-1 in the activation of genes from different stress response pathways, we activated ATFS-1 by mildly impairing mitochondrial function through a mutation in nuo-6 (blue bars) and then examined the effect of disrupting atfs-1 using an atfs-1 deletion mutant atfs-1(gk3094) (purple bars). We also examined the expression of these genes in two constitutively active atfs-1 mutants, atfs-1(et15) and atfs-1(et17) (green bars). Target genes from the mitochondrial unfolded protein response (mitoUPR) (A, hsp-6), DAF-16-mediated stress response (D, sod-3), SKN-1-mediated oxidative stress response (E, gst-4), HIF-1-mediated hypoxia response (F, nhr-57), p38-mediated innate immune pathway (G, Y9C9A.8), and antioxidant defense (H, trx-2) are all significantly upregulated in nuo-6 mutants in an ATFS-1-dependent manner. Target genes from the mitoUPR, DAF-16-mediated stress response, HIF-1-mediated hypoxia response, p38-mediated innate immune pathway and antioxidant defense are also upregulated in constitutive activation of ATFS-1 (A, D, F, G, H). In contrast, activation of ATFS-1 by nuo-6 mutation or atfs-1 gain-of-function mutations did not significantly affect target gene expression for the endoplasmic reticulum unfolded protein response (B, ER-UPR, hsp-4) or the cytoplasmic unfolded protein response (C, Cyto-UPR, hsp-16.2). atfs-1(gk3094) is a loss-of-function deletion mutant. atfs-1(et15) and atfs-1(et17) are constitutively active gain-of-function mutants. Error bars indicate SEM. **p<0.01, ***p<0.001. A full list of genes that are upregulated by ATFS-1 activation can be found in Table S2.   308   2597  734 13 2892 43 23% overlap; p < 0.05 30% overlap; p = 7.5X10 -28 Figure 2. Constitutive activation of ATFS-1 results in upregulation of genes from multiple stress response pathways. Genes that are upregulated by activation of ATFS-1 were compared to previous published lists of genes involved in different stress response pathways, including the mitochondrial unfolded protein response (A, mitoUPR), the endoplasmic reticulum unfolded protein response (B, ER-UPR), the cytoplasmic unfolded protein response (C, Cyto-UPR), the DAF-16-mediated stress response (D), the SKN-1-mediated oxidative stress response (E), the HIF-1-mediated hypoxia response (F), the p38-mediated innate immune response (G), and antioxidant genes (H). In every case, there was a significant degree of overlap ranging from 22%-51%. Grey circles indicate genes that are upregulated by activation of the stress response pathway indicated. Turquoise circles indicate genes that are upregulated in the atfs-1(et15) constitutively active gain-of-function mutant. The numbers inside the circles show how many genes are upregulated. The percentage overlap is the number of overlapping genes as a percentage of the number of genes upregulated by the stress response pathway. mitoUPR = mitochondrial unfolded protein response. ER-UPR = endoplasmic reticulum unfolded protein response. Cyto-UPR = cytoplasmic unfolded protein response. DAF-16 = DAF-16-mediated stress response pathway. SKN-1 = SKN-1-mediated oxidative stress response pathway. HIF-1 = HIF-1-mediated stress response pathway. Innate immunity = p38-mediated innate immunity pathway. Antioxidant = antioxidant genes. Stress pathway gene lists and sources can be found in Table S3.    Figure 3. ATFS-1 is required for upregulation of stress response genes after exposure to oxidative stress or bacterial pathogen stress. To determine the role of ATFS-1 in responding to different types of stress, we compared the upregulation of stress response genes in wild-type and atfs-1(gk3094) loss-of-function deletion mutants after exposure to different stressors. A. Exposure to oxidative stress (4 mM paraquat, 48 hours) caused a significant upregulation of hsp-6, nhr-57 and trx-2 in wild-type worms that was prevented by the disruption of atfs-1. B. Exposure to bacterial pathogen stress (PA14, 24 hours) resulted in an upregulation of dod-22, gst-4 and sod-5 in wild-type worms that was prevented by the atfs-1 deletion. C. Exposure to heat stress (35°C, 2 hours) increased the expression of hsp-16.2, hsp-4 and ckb-2 in both wildtype and atfs-1 worms. D. Exposure to osmotic stress (300 mM, 24 hours) caused an upregulation of sod-3, gst-4 and Y9C9A.8 in wild-type worms and to a greater magnitude in atfs-1 mutants. E. Anoxia (24 hours) resulted in the upregulation of hsp-16.2, Y9C9A.8 and dod-22 in both wild-type and atsf-1 worms. F. Exposing worms to endoplasmic reticulum stress (tunicamycin, 24 hours) increased the expression of ckb-2 and clec-67 in both wild-type and atfs-1 worms. Error bars indicate SEM. *p<0.05, **p<0.01, ***p<0.001. To determine the role of ATFS-1 in resistance to stress, the stress resistance of an atfs-1 loss-of-function mutants (atfs-1(gk3094)) and two constitutively active atfs-1 gain-of-function mutants (atfs-1(et15), atfs-1(et17)) was compared to wild-type worms. A. Activation of ATFS-1 enhanced resistance to acute oxidative stress (300 µM juglone), while deletion of atfs-1 markedly decreased resistance to acute oxidative stress. B. Disruption of atfs-1 decreased resistance to chronic oxidative stress (4 mM paraquat). atfs-1(et17) mutants showed increased resistance to chronic oxidative stress, while atfs-1(et15) mutants had decreased resistance. C. Resistance to heat stress (37°C) was not enhanced by activation of ATFS-1, while deletion of atfs-1 decreased heat stress resistance. D,E. Activation of ATFS-1 increased resistance to osmotic stress (450 mM, 500 mM NaCl), while disruption of atfs-1 decreased osmotic stress resistance. F. Constitutively active atfs-1 mutants show increased resistance to anoxia (75 hours), while atfs-1 deletion mutants exhibit a trend towards decreased anoxia resistance. G. Activation of ATFS-1 increased resistance to P. aeruginosa toxin in a fast kill assay. A slow kill assay in which worms die from internal accumulation of P. aeruginosa was performed according to two established protocols. H. At 25°C , atfs-1(et17) mutants showed a small decrease in resistance to bacterial pathogens (PA14), while atfs-1(gk3094) mutants showed a small increase in resistance. I. At 20°C, both atfs-1(et17) and atfs-1(gk3094) mutants exhibited a small increase in resistance to bacterial pathogens. Error bars indicate SEM. *p<0.05, **p<0.01, ***p<0.001.   Figure 5. Multiple long-lived mutants from different pathways of lifespan extension show upregulation of ATFS-1-dependent genes. To determine the extent to which long-lived genetic mutants from different pathways of lifespan extension show differential expression of ATFS-1 target genes, we compared genes that are upregulated in nine different long-lived mutants to a published list of spg-7 RNAi-upregulated, ATFS-1-dependent target genes . We found that clk-1, isp-1, nuo-6, sod-2, daf-2, glp-1 and ife-2 worms all show a highly significant degree of overlap with genes upregulated by spg-7 RNAi in an ATFS-1-dependent manner. The grey circles represent the 366 genes that are upregulated by spg-7 RNAi in an ATFS-1 dependent manner. Turquoise circles are genes that are significantly upregulated in the long-lived mutant indicated as determined from our RNA sequencing data. The number of unique and overlapping genes are indicated. Overlap is calculated as the number of genes in common between the two gene sets divided by the total number of genes that are upregulated by spg-7 RNAi in an ATFS-1 dependent manner. Enrichment is calculated as the number of overlapping genes observed divided by the number of overlapping genes predicted if genes were chosen randomly. To determine the effect of ATFS-1 on aging, we quantified the lifespan of an atfs-1 deletion mutant and two constitutively active atfs-1 mutants. A,B. Both constitutively active atfs-1 mutants, et15 and et17, have a significantly decreased lifespan compared to wild-type worms. C. Deletion of atfs-1 does not affect lifespan compared to wild-type worms. atfs-1(gk3094) is a loss of function mutant resulting from a deletion. atfs-1(et15) and atfs-1(et17) are constitutively active gain-of-function mutants.
A B C Figure S1. Multiple long-lived mutants from different pathways of lifespan extension show upregulation of ATFS-1-dependent genes. To determine the extent to which long-lived genetic mutants from different pathways of lifespan extension show differential expression of ATFS-1 target genes, we compared genes that are upregulated in nine different long-lived mutants to genes upregulated in a constitutively active atfs-1 mutant (et15). All of the long-lived mutant worms, except for osm-5, show a highly significant degree of overlap with the constitutively active atfs-1 mutant. The grey circles represent genes that are significantly upregulated in the constitutively active atfs-1(et15) mutant. Turquoise circles are genes that are significantly upregulated in the long-lived mutant indicated. The number of unique and overlapping genes are indicated. Overlap is calculated as the number of genes in common between the two gene sets divided by the smaller gene set. Enrichment is calculated as the number of overlapping genes observed divided by the number of overlapping genes predicted if genes were chosen randomly.