The cell-non-autonomous nature of electron transport chain-mediated longevity

Abstract

The life span of C. elegans can be increased via reduced function of the mitochondria; however, the extent to which mitochondrial alteration in a single, distinct tissue may influence aging in the whole organism remains unknown. We addressed this question by asking whether manipulations to ETC function can modulate aging in a cell-non-autonomous fashion. We report that the alteration of mitochondrial function in key tissues is essential for establishing and maintaining a prolongevity cue. We find that regulators of mitochondrial stress responses are essential and specific genetic requirements for the electron transport chain (ETC) longevity pathway. Strikingly, we find that mitochondrial perturbation in one tissue is perceived and acted upon by the mitochondrial stress response pathway in a distal tissue. These results suggest that mitochondria may establish and perpetuate the rate of aging for the whole organism independent of cell-autonomous functions.

Figure 1

Life-span analysis of cco-1 hairpin transgenic animals. A. Wild-type worms allow import of dsRNA from surrounding tissues, but sid-1(qt9) mutant worm can not import dsRNA and RNAi knockdown is no longer systemic but is maintained locally within the tissue in which the dsRNA is produced (Winston et al., 2002). B. Intestine-specific knockdown of cco-1 results in life-span extension. sid-1(qt9)/rol-6 control (black line, mean 18.8 +/- 0.7 days), ges-1p∷cco-1hairpin (green line, 23.9 +/- 0.8 days, p<.0001). C. Body wall muscle knockdown of cco-1 does not significantly affect life span. sid-1(qt9)/rol-6 control (black line, mean 18.6 +/-0.5 days), myo-3p∷cco-1 hairpin (blue line, mean 16.6+/- 0.5 days, p=0.0574). D. Neuronal knockdown of cco-1 extends life span. sid-1(qt9)/rol-6 control (black line, 18.2 +/-0.2 days), rab-3p∷cco-1 hairpin (purple line, 21.7 +/- 0.5 days, p<.0001). E. Neuronal knockdown of cco-1 driven by the unc-119 promoter also extends life span. sid-1(qt9)/rol-6 control (black line, mean 19.8 +/-0.7days), unc-119p∷cco-1 hairpin (red line, mean 23.8 +/-0.8 days, p=.0001). Please see Table S1 for all statistical analyses and also Figure S1 and Movie S1 for additional experiments.

Figure 2

Life-span analysis of tissue-specific complementation of rde-1 with cco-1 feeding RNAi. A. Tissues exposed to dsRNA from feeding RNAi initiate knockdown if rde-1 has been rescued in the corresponding tissue. Neighboring tissues are unable to initiate RNAi if rde-1 is absent. B. rde-1(ne219) mutants do not respond to cco-1 feeding RNAi. Animals fed bacteria harboring an empty vector (black line, mean 18.0 +/- 0.3 days), cco-1 RNAi (red line, mean 18.16 +/-0.4 days, p<0.4043). C. rde-1 rescued in the intestine (VP303) extends life span when fed cco-1 dsRNA producing bacteria. Animals fed vector only bacteria (black line, mean 14.7 +/- 0.6 days), cco-1 RNAi (green line, mean 22.0 +/- 0.2 days, p<.0001). D. rde-1 rescued in the body wall muscle (NR350.5) decreases life span when fed cco-1 dsRNA bacteria. Animals fed bacteria harboring empty vector (black line, mean 13.5 +/- 0.3 days), cco-1 RNAi (blue line, mean 11.8 +/- 0.3 days, p<0.0002. E. rde-1 rescued in the hypodermis (NR222) has no effect on life span when fed cco-1 dsRNA producing bacteria. Vector only (black line, mean 13.5+/- 0.3days), cco-1 RNAi (purple line, mean 14.3 +/- 0.4 days, p=0.148). F. Life-span extension by cco-1 feeding RNAi in the intestine of ges-1:;rde-1 rescued animals is independent of daf-16. Intestinal rde-1 rescued worms were fed empty vector (black line), mean 16.4+/- 0.6 days, daf-16 RNAi diluted 50% with empty vector (gold line, mean 16.3+/- 0.4days) or daf-16 diluted 50% with cco-1 RNAi (green19.9+/- 0.5 days, p<.0001). Please see Table S1 for all statistical analysis. G. Double transgenic animals carrying rab-3:;cco-1HP and ges-1:cco-1HP (blue line, mean life span 23.4 +/-0.6 days) did not live longer than either the rab-3∷cco-1HP (red line, mean life span 23.2 +/-0.6 days, p =.64) or the the ges1∷cco-1HP (green line, mean life span 22.9+/-0.6 days, p=.75) animals. sid-1 control (black line, mean life span 19.2+/-0.5 days).

Figure 3

Induction of the UPR^mt is specific to the ETC longevity pathway. A. hsp-6p∷GFP reporter worms fed empty vector (EV) containing bacteria have low levels of background GFP (i) overlay; (ii) GFP. hsp-6p∷GFP reporter worms fed cco-1 RNAi upregulate the UPR^mt. Relative fluorescence was quantified using a fluorescence plate reader (iii). B. daf-2 RNAi does not induce hsp-6p∷GFP (i) overlay; (ii) GFP. hsp-6p∷GFP reporter worms were hatched on empty vector, cco-1, or daf-2 dsRNA expressing bacteria and allowed to grow to day 1 of adult hood. Relative fluorescence was quantified (iii). C. Dietary restricted eat-2(ad1116) mutant worms do not upregulate hsp-6p∷GFP reporter (i) overlay; (ii) GFP. Relative fluorescence was quantified (iii). D The UPR^ER is not induced by cco-1 RNAi, (i) overlay; (ii) GFP. hsp-4p∷GFP transgenic reporter worms were fed empty vector containing bacteria or cco-1 dsRNA bacteria. No fluorescence upregulation was detected (iii). Both EV and to a lesser extent cco-1 RNAi fed worms were able to upregulate the UPR^ER upon treatment with tunicamycin, (i and ii) which is known induce UPR^ER. Relative fluorescence of was quantified (iii). E. cco-1 RNAi does not induce a marker of cytosolic protein misfolding stress, (i) overlay; (ii) GFP. hsp16.2p∷GFP reporter worms were fed EV or cco-1 dsRNA bacteria. No fluorescence upregulation was detected (iii). As positive controls, heat shock for 6 hours at 31°C could induce the heat-shock response (HSR) and cco-1 RNAi did not block this response (i and ii). In all panels, error bars indicate standard deviations (SD).

Figure 4

ubl-5 is necessary and specific for ETC mediated longevity. A. The long life span of isp-1(qm150) mutant animals is dependent upon ubl-5. isp-1(qm150) (empty vector, black line, mean 25.8 +/- 1.0 days), isp-1(qm150) fed ubl-5 dsRNA bacteria (orange line, mean 15.5 +/-0.7 days, p<.0001), N2 wild-type (grey line, mean 19 +/- 0.5 days). B. daf-2(e1370) mutant life span is unaffected by ubl-5 knockdown. daf-2(e1370) mutant animals grown on empty vector bacteria (black line, mean 40.1+/- 1.2 days), daf-2(e1370) fed ubl-5 dsRNA bacteria (orange line, mean 39.9 +/-1.2 days, p=.327). C. Dietary restricted eat-2(ad1116) mutant life span is not dependent upon ubl-5. N2 on empty vector (grey line, mean life span 18.2+/-0.4 days); eat-2(ad1116) on empty vector (black line, mean 26.4+/-0.6 day)s; eat-2(ad1116) fed ubl-5 dsRNA bacteria (orange line, mean 23.3+/-0.7 days, p<0.0004). D. N2 wild-type life span is unaffected by ubl-5 knockdown. N2 grown on empty vector bacteria (black line, mean 18.2 +/-0.4 days), N2 fed ubl-5 dsRNA bacteria (orange line, mean 20.3 +/- 0.4 days, p=0.0834). All statistical data can be found in Table S1. See also Figure S2 for additional experiments.

Figure 5

The temporal activation of ETC generated longevity signal is coincident with induction of the UPR^mt. A. hsp-6p∷GFP reporter worms were transferred to cco-1 RNAi at each larval developmental stage and early adulthood. GFP fluorescent measurements were taken 16 hours after reaching young adulthood in all cases. B. hsp-6p∷GFP is upregulated if transfer occurs before the L4 stage of development. C. Quantification of hsp-6p∷GFP in (A); error bars represent standard deviation (SD). D. cco-1 knockdown during larval development is sufficient to induce the hsp-6p∷GFP reporter in adulthood. hsp-6p∷GFP reporter worms we grown on cco-1 dsRNA bacteria during development and then moved to dcr-1 dsRNA producing bacteria at the L4 larval stage, to disrupt the RNAi machinery allowing CCO-1 levels to return to normal (Dillin et al., 2002). UPR^mt remains induced. E. hsp-6p∷GFP fluorescence 48 hours after transfer to dcr-1 RNAi as described by schematic D. See also Figure S3.

Figure 6

Cell non-autonomous upregulation of the UPR^mt. A. Representation of cell-autonomous and non-autonomous upregulation of UPR^mt. “X's” depict tissue where cco-1 is knocked down (intestine or neurons). Green indicates location of upregulation of hsp-6p∷GFP reporter (intestine upon knockdown in intestine or neurons). B. hsp-6p∷GFP reporter worms were crossed to tissue-specific cco-1 hairpin lines. Control hsp-6p∷GFP shows only background GFP (i). Neuron-specific cco-1 hairpin results in upregulation of hsp-6p∷GFP in the intestine (rab-3 (ii) and unc-119 (iii) lines shown). Intestine-specific ges-1p∷cco-1 hairpin (iv) also results in upregulation of the hsp-6p∷GFP reporter in the intestine. C. Fluorescent quantification of B; error bars represent standard deviation (SD). D. Intestinal knockdown of ubl-5 does not block UPR^mt induction caused by neuronal cco-1 reduction. Worm strains were created with rab-3∷cco-1HP; hsp-6p∷gfp with gly-19p∷ubl-5HP. E. Intestinal reduction of ubl-5 does not block the life-span extensions of rab-3∷cco-1HP animals. N2 fed empty vector (black line, mean=19.1+/-0.4); sid-1(qt9) (blue line, mean=19.4+/-0.6 days); sid-1(qt9); gly-19p∷ubl-5-KD (red line, mean=20.1+/-0.7 +/-0.7days); rab-3p∷cco-1-HP; gly-19p∷ubl-5-KD (yellow line, mean= 24.1+/-0.7days); rab-3p∷cco-1HP (green line, mean= 24.5+/-0.7days); N2 on cco-1 RNAi (mean= 27.3+/-0.6days. p>0.66 (green vs. yellow)) F. Intestinal reduction of ubl-5 (gly-19p∷ubl-5HP) blocks UPR^mt induction of animals fed bacteria expressing cco-1 dsRNA. See also Figure S4.

Figure 7

Model for the cell non-autonomous nature of the UPR^mt. Cells experiencing mitochondrial stress, in this scenario neuronal cells (circles) marked within the yellow box, produce a signal that is transmitted from the mitochondria to the nucleus to regulate the expression of genes regulated by UBL-5 and possibly DVE-1. These cells serve as sending cells and produce an extracellular signal (mitokine) that can be transmitted to distal, receiving cells, in this case intestinal cells marked in the green box. Receiving cells perceive the mitokine and induce the mitochondrial stress response. See also Figure S5.