Surveillance-activated defenses block the ROS-induced mitochondrial unfolded protein response

Abstract

Disturbance of cellular functions results in the activation of stress-signaling pathways that aim at restoring homeostasis. We performed a genome-wide screen to identify components of the signal transduction of the mitochondrial unfolded protein response (UPR(mt)) to a nuclear chaperone promoter. We used the ROS generating complex I inhibitor paraquat to induce the UPR(mt), and we employed RNAi exposure post-embryonically to allow testing genes whose knockdown results in embryonic lethality. We identified 54 novel regulators of the ROS-induced UPR(mt). Activation of the UPR(mt), but not of other stress-signaling pathways, failed when homeostasis of basic cellular mechanisms such as translation and protein transport were impaired. These mechanisms are monitored by a recently discovered surveillance system that interprets interruption of these processes as pathogen attack and depends on signaling through the JNK-like MAP-kinase KGB-1. Mutation of kgb-1 abrogated the inhibition of ROS-induced UPR(mt), suggesting that surveillance-activated defenses specifically inhibit the UPR(mt) but do not compromise activation of the heat shock response, the UPR of the endoplasmic reticulum, or the SKN-1/Nrf2 mediated response to cytosolic stress. In addition, we identified PIFK-1, the orthologue of the Drosophila PI 4-kinase four wheel drive (FWD), and found that it is the only known factor so far that is essential for the unfolded protein responses of both mitochondria and endoplasmic reticulum. This suggests that both UPRs may share a common membrane associated mechanism.

Figure 1. Paraquat induces hsp-6 and its reporter in a ROS–dependent manner.

A. Quantitative analysis (by qRT-PCR) of endogenous hsp-6 mRNA in wild type (N2) worms exposed to 0.5 mM paraquat from early L3 stage on for 4 h, 24 h and 30 h, presented as fold induction. Dots indicate single experiments; mean plus SD. B. Quantification of GFP fluorescence intensity in the hsp-6 reporter strain (Phsp-6::gfp) after two days of exposure to 0.5 mM paraquat from early L3 stage on. Paraquat significantly increased (p<0.0001) hsp-6 reporter expression. Columns represent pooled normalized values of three independent experiments plus standard error of the mean (SEM). Numbers in or on columns indicate the number of analyzed animals (n_total = 152). ***: p<0.0001; Mann Whitney test. C. The paraquat-triggered induction of the hsp-6 reporter (Phsp-6::gfp) was decreased by the addition of the ROS scavenger N-acetyl-L-cysteine (NAC). Columns represent normalized values plus standard error of the mean (SEM). Numbers in columns indicate the number of analyzed animals (n_total = 40). ***: p<0.0001; Unpaired t test with Welch's correction.

Figure 2. The induction of hsp-6::gfp by paraquat does not require HAF-1, but ATFS-1.

A. GFP staining after induction of the hsp-6 reporter gene (Phsp-6::gfp) with (PQ) or without (no PQ) 0.5 mM paraquat. L4440: vector control. atfs-1(i): RNAi against atfs-1; haf-1(+): wild type and haf-1(−): ok705 mutant allele. Equal optical settings per row. Scale bar 200 µm. B. Quantification of GFP fluorescence intensity confirms that the ok705 mutation did not block, but increased the induction of hsp-6::gfp with paraquat, whereas knockdown of afts-1 (afts-1(i)) reduced reporter fluorescence to background. Columns represent pooled normalized values of three independent experiments plus standard error of the mean (SEM). Numbers in or on columns indicate the number of analyzed animals (ok705: n_total = 318; afts-1 (i): n_total = 339). ***: p<0.0001; Mann Whitney test (after subtraction of respective background fluorescence).

Figure 3. Mitochondrial ROS generators induce hsp-6::gfp.

A. The hsp-6 reporter strain (Phsp-6::gfp) was induced when exposed during larval development to paraquat (0.5 mM), antimycin A (0.25 µM), or rotenone (0.25 µM). A ROS generating point mutation (qm150) in the ETC complex III gene isp-1 also induced the hsp-6 reporter. Scale bar: 200 µm. B. Acrylamide, whose oxidative stress generating activity has not been linked to mitochondrial metabolism, induced the hsp-6 reporter weakly, but paraquat induction was considerably stronger. Columns represent pooled normalized values of three independent experiments plus standard error of the mean (SEM). Numbers in or on columns indicate the number of analyzed animals (n_total = 234). ***: p<0.0001; Kruskal-Wallis test plus Dunn's Multiple Comparison Test.

Figure 4. Paraquat affects mitochondrial morphology, but does not provoke the unfolded protein responses of the endoplasmatic reticulum (ER) or the cytosol.

A. Representative confocal micrographs of cells in wild type worms after Mitotracker staining. Worms were exposed to 0.5 mM paraquat starting from early L3 stage. Hypodermal mitochondrial staining was analyzed after two days. Arrows indicates the location of vulva. Scale bar: 10 µm. B–C. Quantification of GFP fluorescence intensities in a cytosolic UPR reporter strain (Phsp-16.2::gfp) (B) and an UPR^ER reporter strain (Phsp-4::gfp) after two days of exposure to 0.5 mM paraquat or to their respective inductor (heat shock: 4 h at 34°C at L4, analyzed after one day; tunicamycin: 7.2 µM at L1, analyzed after three days). Both UPR reporters were not induced by paraquat, suggesting that the compound does not affect protein folding environment in cytosol or ER, respectively. Columns represent normalized pooled values of three independent experiments plus standard error of the mean (SEM). Numbers in or on columns indicate the number of analyzed animals (cytosolic UPR n_total = 159; UPR^ER n_total = 188). ***: p<0.001, n.s: p>0.05; Kruskal-Wallis test plus Dunn's Multiple Comparison Test.

Figure 5. The hsp-6 response to paraquat does not require the stress-inducible transcription factors SKN-1, DAF-16, or HIF-1.

A–C. Representative micrographs and the corresponding quantifications of the induction of hsp-6 reporter (Phsp-6::gfp) with paraquat in daf-16(mu68) (A), skn-1(zu67) (B), and hif-1(ia4) (C) strains. Early L3 larvae of the respective mutants carrying the hsp-6 reporter (Phsp-6::gfp) were exposed to 0.5 mM paraquat on OP50 plates and analyzed for GFP fluorescence after two days. Equal optical settings per row. Columns represent normalized pooled values of three independent experiments each plus standard error of the mean (SEM). Numbers in or on columns indicate the number of analyzed animals (daf-16: n_total = 293, skn-1: n_total = 272, hif-1: n_total = 278). ***: p<0.0001, *: p = 0.0214, n.s.: p<0.05; Mann Whitney test (A, B), unpaired t test with Welch's correction (C) after subtraction of respective background expression. Scale bar 250 µm.

Figure 6. Activities of rpl-36, atfs-1 and pifk-1 are required for the hsp-6 response to paraquat.

Representative micrographs (A) and quantification of GFP fluorescence intensity (B) of three screening positives (rpl-36, atfs-1 and pifk-1) show a block of the paraquat triggered induction of the hsp-6 reporter (Phsp-6::gfp) after their RNAi. Worms were raised on respective RNAi plates from L1 larval stage and exposed to 0.5 mM paraquat at early L3 stage. GFP fluorescence was analyzed after two days. Columns represent pooled normalized values of three independent experiments plus standard error of the mean (SEM). Numbers in or on columns indicate the number of analyzed animals (n_total = 648). ***: p<0.001; Kruskal-Wallis test plus Dunn's Multiple Comparison Test; Mann Whitney test. Equal optical settings, scale bar 200 µm. (i): RNAi; L4440: empty vector control.

Figure 7. Knockdown of rpl-36, atfs-1, and pifk-1 suppresses the isp-1(qm150)–mediated induction of the hsp-6 reporter.

The isp-1(qm150) mutant of mitochondrial superoxide constitutively activated the Phsp-6 reporter (Phsp-6::gfp). RNAi of all three tested genes suppressed (p<0.001) the constitutive hsp-6 reporter gene induction. Representative micrographs (A) and quantification of GFP fluorescence intensity (B). hsp-6 reporter worms carrying the qm150 allele were analyzed for GFP expression after one week on the respective RNAi plates. Columns represent pooled values of three independent experiments plus standard error of the mean (SEM). Numbers in columns indicate the number of analyzed animals (n_total = 317). ***: p<0.001; Kruskal-Wallis test plus Dunn's Multiple Comparison Test. Equal optical settings per row, scale bar 100 µm. (i): RNAi; L4440: empty vector control.

Figure 8. The downregulation of rpl-36, atfs-1, and pifk-1 increases paraquat sensitivity.

L1 staged N2 worms were placed on the respective RNAi plates containing 0.4 mM or no paraquat, development was analyzed five days later. Downregulation of all three genes enhanced sensitivity towards paraquat, indicated by delayed development. Columns represent pooled values of three independent experiments in percent. Numbers on columns indicate the number of animals analysed (n_total = 2701). (i): RNAi; L4440: empty vector control.

Figure 9. The knockdown of rpl-36, atfs-1, or pifk-1 does not prevent non-mitochondrial stress responses.

Worms were grown from L1 larval stage on the respective RNAi plates before being exposed to the respective stress and analyzed four days after L1. A. A reporter strain for the SKN-1 dependent phase II response (Pgst-4::gfp) was exposed to 2.1 mM acrylamide starting at early L3 stage. RNAi of rpl-36 and atfs-1 did not prevent, but pifk-1 (RNAi) significantly (p<0.001) reduced reporter gene induction as compared to vector control. Columns represent pooled normalized values of four independent experiments plus standard error of the mean (SEM). Numbers in columns indicate the number of analyzed animals (n_total = 819). ***: p<0.001; Kruskal-Wallis test plus Dunn's Multiple Comparison Test; Mann Whitney test (comparison of vector with and without acrylamide). Equal optical settings, scale bar 200 µm. B. Cytosolic UPR (heat shock) reporter worms (Phsp-16.2::gfp) were exposed to 34°C for 4 h at L4. While the downregulation of none of the three screening positives prevented the heat shock response completely, the knockdown of rpl-36 and atfs-1 significantly decreased heat stress induced reporter expression (p<0.001). Columns represent pooled normalized values of two independent experiments plus standard error of the mean (SEM). Numbers in or on columns indicate the number of analyzed animals (n_total = 354). ***: p<0.001; Kruskal-Wallis test plus Dunn's Multiple Comparison Test; Mann Whitney test (comparison of vector with and without heat shock). Equal optical settings, scale bar 100 µm. C. The UPR^ER reporter strain (Phsp-4::gfp) was raised from L1 stage RNAi plates (with 7.2 µg/ml tunicamycin). UPR^ER induction was not blocked by any RNAi tested here, but pifk-1 (RNAi) and rpl-36 (RNAi) strongly impaired its induction (p<0.001). Columns represent pooled normalized values of four independent experiments plus standard error of the mean (SEM). Numbers in or on columns indicate the number of analyzed animals (n_total = 495). ***: p<0.001; Kruskal-Wallis test plus Dunn's Multiple Comparison Test; Mann Whitney test (comparison of vector with and without tunicamycin). Equal optical settings, scale bar 100 µm. (i): RNAi; L4440: empty vector control.

Figure 10. cSADDs inhibits paraquat-mediated signaling to hsp-6 through KGB-1.

A–B. In kgb-1(um3) mutant animals, which are cSADDs deficient, paraquat induced hsp-6 induction is not blocked by elt-2 RNAi. Thus, ROS induced UPR^mt is activated in the absence of functional cSADDs. In contrast, kgb-1(um3) does not prevent inhibition of hsp-6 induction by afts-1 and pifk-1 knockdown, suggesting that they function downstream of kgb-1 and the cSADDs. Columns represent normalized values plus standard error of the mean (SEM). Numbers in or on columns indicate the number of analyzed animals (n_total = 248). ***: p<0.001, *: p<0.05; Kruskal-Wallis test plus Dunn's Multiple Comparison Test (A). Equal optical settings, scale bar 400 µm. (i): RNAi; L4440: empty vector control; +: wild type allele (B). C. Model: Genes activating the cSADDs (cellular surveillance system) inhibit the paraquat-triggered induction of the UPR^mt.