Mitochondrial dysfunction confers resistance to multiple drugs in Caenorhabditis elegans

Abstract

In a previous genetic screen for Caenorhabditis elegans mutants that survive in the presence of an antimitotic drug, hemiasterlin, we identified eight strong mutants. Two of these were found to be resistant to multiple toxins, and in one of these we identified a missense mutation in phb-2, which encodes the mitochondrial protein prohibitin 2. Here we identify two additional mutations that confer drug resistance, spg-7 and har-1, also in genes encoding mitochondrial proteins. Other mitochondrial mutants, isp-1, eat-3, and clk-1, were also found to be drug-resistant. Respiratory complex inhibitors, FCCP and oligomycin, and a producer of reactive oxygen species (ROS), paraquat, all rescued wild-type worms from hemiasterlin toxicity. Worms lacking mitochondrial superoxide dismutase (MnSOD) were modestly drug-resistant, and elimination of MnSOD in the phb-2, har-1, and spg-7 mutants enhanced resistance. The antioxidant N-acetyl-l-cysteine prevented mitochondrial inhibitors from rescuing wild-type worms from hemiasterlin and sensitized mutants to the toxin, suggesting that a mechanism sensitive to ROS is necessary to trigger drug resistance in C. elegans. Using genetics, we show that this drug resistance requires pkc-1, the C. elegans ortholog of human PKCepsilon.

Figure 1.

Amino acid sequence alignment of the region mutated in (A) HAR-1 and (B) SPG-7. The amino acid sequence near the mutation site is shown for Saccharomyces cerevisiae, C. elegans, Drosophila melanogaster, Mus musculus, and Homo sapiens. The mutated amino acid in each protein is indicated with an asterisk (*). A complete alignment assembled by the ClustalX program is presented in Supplemental Figures 1 and 2.

Figure 2.

PHB-2 and CHCHD2 (Har-1) mutants localize to mitochondria. Hela cells were transfected with plasmid expression vectors expressing myc-tagged human PHB-2, human PHB-2(E130K), CHCHD2 (the human ortholog of Har-1), and CHCHD2(G65E). After 18 h cells were stained with MitoTracker Red, fixed, and stained with anti-myc antibody (green). PHB2 is known to localize to and CHCHD2 suspected of localizing to mitochondria. Each of the mutants also colocalized with the MitoTracker Red. In some cells such as the one shown for PHB-2 wild type, the myc-staining pattern resembled MitoTracker staining, but no MitoTracker was visible in the cells.

Figure 3.

The hemiasterlin analog causes mitochondrial networks to fragment in wild-type worms. Images of worms stained with MitoTracker Red were taken with the focal plane in marginal cells in the region of the metacorpus and isthmus in the head of wild-type or phb-2;har-1 double mutant worms in the presence or absence of the hemiasterlin analog. (A) Wild-type worm. (B) Fragmented mitochondria in a wild-type worm treated with 0.6 μM hemiasterlin analog. (C) One of the more normal appearing phb-2;har-1 worms without drug treatment. (D) phb-2;har-1 treated with 0.6 μM drug. (E) An untreated phb-2;har-1 worm with fragmented mitochondria. All images were taken at the same magnification and with the same camera settings.

Figure 4.

AMPK is activated in phb-2, har-1, and spg-7, suggesting impaired ATP production. Wild-type and mutant adult worms were lysed and analyzed by immunoblotting with an antibody that recognizes phosphorylated T172 on AMPK-α. The blot was reprobed with antibody to actin.

Figure 5.

Not all mutations in mitochondrial proteins confer increased resistance to the hemiasterlin analog. Synchronized L1 larvae of each strain were incubated at the indicated hemiasterlin concentrations for 4–5 d, and healthy mobile adults with eggs were counted. For each concentration, two wells were scored in an experiment. Average values are graphed ± SEM. Note that data for N2 and mev-1 superimpose. Similar results were reproduced in at least three independent experiments.

Figure 6.

The mutants isolated in our screen are not long-lived. phb-2, phb-2;har-1, and spg-7 have slightly shortened life spans, and har-1 is similar to wild type. isp-1worms were used as a long-lived control. Data represent average values of two independent experiments ± SEM.

Figure 7.

Mitochondrial inhibitors rescue wild-type N2 worms from the hemiasterlin analog. About 400 synchronized N2 L1 larvae were grown in liquid culture in the presence of increasing concentrations of the mitochondrial inhibitor alone or mitochondrial inhibitor plus 1.3 μM hemiasterlin analog. After 3 d of exposure healthy, morphologically normal and mobile L4 and adult worms were counted. Worms exposed to 1.3 μM hemiasterlin analog alone die or develop as paralyzed and dumpy animals. In contrast, when exposed to 1.3 μM hemiasterlin analog in the presence of an inhibitor of mitochondrial respiration, N2 worms grow as healthy, mobile, and morphologically normal animals. TTFA, rotenone, FCCP, oligomycin, and antimycin eventually killed worms at higher concentrations, but were protective at lower dosages. Rotenone, oligomycin, and antimycin also slightly delayed worm growth comparable to untreated animals. Stigmatellin, myxothiazol, and FCCP protected worms from hemiasterlin analog and did not affect animal growth. Experiments with each drug were independently reproduced three times, and average values are graphed ± SEM. For most concentrations tested, the error bars are smaller than the symbol that indicates the data point.

Figure 8.

Inducing ROS with paraquat, or loss of MnSODs, increases resistance to hemiasterlin. (A) The sod-2;sod-3 double mutants are more sensitive to paraquat than N2 worms. (B) Like mitochondrial inhibitors, paraquat protects worms from hemiasterlin toxicity, and sod-2;sod-3 double mutants were protected at a lower paraquat concentration than wild-type worms. Worms were grown in 2 μM hemiasterlin analog, a concentration that kills most larvae, with the few survivors remaining as paralyzed dumpy worms. However, in the presence of paraquat, animals survived and were able to move. (C) Loss of MnSODs increases tolerance to hemiasterlin analog. Worms with the double sod-2;sod-3 mutations are modestly resistant to the toxin, and phb-2;sod-2;sod-3, har-1;sod-2;sod-3, and spg-7;sod-2;sod-3 animals are more resistant than phb-2, har-1, or spg-7, indicating that triple mutants acquired their resistance capacity from both parental strains. Experiments shown in A–C were repeated three times and averages are plotted ± SEM. (D) Immunoblot analysis with anti-MnSOD antibody (top panel) detects a 25-kDa MnSOD band present only in N2 animals (1), but not in sod-2;sod-3 (2), phb-2;sod-2;sod-3 (3), har-1;sod-2;sod-3 (4), or spg-7;sod-2;sod-3 (5). The same blot was reprobed with actin antibody, as a loading control (bottom panel).

Figure 9.

phb-2 and har-1 respond differentially to some drugs. har-1 is hypersensitive to antimycin but more tolerant than phb-2 to the inhibitor of fatty acid oxidation, etomoxir sodium. Approximately 400 wild-type (N2) or mutant L1 larvae were treated with indicated drug concentrations for 4 d, and healthy adult worms with eggs were counted. Three wells were scored for each concentration, and average values were graphed ± SEM. Curves were fit to the data using the polynomial option in GraphPad Prism 5.0 (San Diego, CA). Each experiment was reproduced independently three times.

Figure 10.

NAC decreased production of H₂O₂ by N2 worms treated with myxothiazol or FCCP. Worms were grown for 2 d in the presence of DMSO and 20 mM NAC and in each mitochondrial inhibitor at indicated concentrations with or without 20 mM NAC. H₂O₂ production was measured as described in Materials and Methods. Results are graphed as the fold difference in H₂O₂ produced by each sample compared with controls treated with DMSO. Average values of three independent experiments performed on different days were graphed ± SEM.

Figure 11.

NAC counteracts the effects of mitochondrial inhibitors or mutations on drug resistance. NAC restored sensitivity of mutants to the hemiasterlin analog and reduced the protection provided by mitochondrial inhibitors to N2 worms. (A and B) In the presence of 20 mM NAC, much higher concentrations of myxothiazol (A) or FCCP (B) were required to protect N2 worms from the hemiasterlin analog. (C) NAC sensitized phb-2, har-1, and spg-7 mutants to the hemiasterlin analog, but did not enhance the sensitivity of N2 worms. Concentrations which killed most worms and left rare dumpy, paralyzed survivors were graphed ± SEM; n = 4.

Figure 12.

PKC-1, but not HIF-1 or AKT-1 plays a role in drug resistance caused by mitochondrial dysfunction. (A) Wild-type N2, har-1, and double mutants combining har-1 with hif-1, akt-1, or daf-18 were compared for resistance to hemiasterlin. None of the double mutants had increased resistance compared with the har-1 single mutant. (B) About 250 synchronized N2 or pkc-1 L1 larvae were grown in liquid culture in the concentrations of FCCP shown plus 1.3 μM hemiasterlin analog. After 3 d, healthy, morphologically normal and mobile L4 and adult worms were counted. FCCP protects both the wild-type worms and pkc-1 mutants from hemiasterlin, but pkc-1 mutants were less protected and protection required higher concentrations of FCCP. For each FCCP concentration, three wells were scored in an experiment, and average values ± SEM were graphed. This experiment was reproduced four times. (C) pkc-1(nj3) was crossed into either the phb-2, har-1, or spg-7 mutants, and the sensitivity of single or double mutants to hemiasterlin was compared. For each concentration of hemiasterlin, triplicate wells containing 300 of L1 larvae of each strain were grown for 3 d, and healthy L4 and adult worms were counted. The average concentrations of three independent experiments at which <10% of worms are healthy adults are graphed ± SEM. By t test, p = 0.005 comparing phb-2 and pkc-1;phb-2, p = 0.02 for har-1 and pkc-1;har-1; and p = 0.02 for spg-7 and pkc-1;spg-7.