A single biochemical activity underlies the pleiotropy of the aging-related protein CLK-1

Abstract

The Caenorhabditis elegans clk-1 gene and the orthologous mouse gene Mclk1 encode a mitochondrial hydroxylase that is necessary for the biosynthesis of ubiquinone (UQ). Mutations in these genes produce broadly pleiotropic phenotypes in both species, including a lengthening of animal lifespan. A number of features of the C. elegans clk-1 mutants, including a maternal effect, particularly extensive pleiotropy, as well as unexplained differences between alleles have suggested that CLK-1/MCLK1 might have additional functions besides that in UQ biosynthesis. In addition, a recent study suggested that a cryptic nuclear localization signal could lead to nuclear localization in cultured mammalian cell lines. Here, by using immunohistochemical techniques in worms and purification techniques in mammalian cells, we failed to detect any nuclear enrichment of the MCLK1 or CLK-1 proteins and any biological activity of a C. elegans CLK-1 protein devoid of a mitochondrial localization sequence. In addition, and most importantly, by pharmacologically restoring UQ biosynthesis in clk-1 null mutants we show that loss of UQ biosynthesis is responsible for all phenotypes resulting from loss of CLK-1, including behavioral phenotypes, altered expression of mitochondrial quality control genes, and lifespan.

Figure 1

2,4-dihydroxybenzoate (DHB) is an unnatural bypass precursor of UQ biosynthesis. UQ is composed of a benzoquinone ring attached to a polyisoprenoid side chain. The length of the side chain varies between species. It is 9 isoprenoid subunits long in C. elegans (UQ₉). All cells depend on endogenous biosynthesis for their supply of UQ. The enzyme encoded by clk-1 and its orthologues (Coq7 in yeast, Mclk1 in mice, and COQ7 in humans) catalyzes the penultimate step in UQ biosynthesis, the hydroxylation of DMQ at position 6. In eukaryotes, the natural biosynthetic precursor for the benzoquinone ring is 4-hydroxybenzoate (4-HB), but 2,4-dihydroxybenzoate (DHB), a hydroxylated analogue of 4-HB, is able to serve as an alternative, albeit unnatural, precursor of UQ synthesis that allows for the biosynthesis of UQ in the absence of CLK-1/COQ7/MCLK1 step^, .

Figure 2

2,4-DHB supplementation restores UQ biosynthesis in clk-1 mutants. HPLC-UV chromatograms of quinone extracts from worms with the indicated genotypes. Elution times for DMQ₉ and UQ₉ are indicated. At the first larval stage, the worms were transferred onto plates supplemented, or not, with DHB. They were fed the UQ-deficient bacterial strain GD1 and grown to the adult stage before being harvested for quinone extraction. Whole worm homogenates with 250 µg of protein were used, expect for the clk-1(qm30) mutants exposed to DHB at 0.5 mM in the plate where a homogenate with 440 µg of protein was used for HPLC analysis. Treatment with DHB restored UQ biosynthesis in clk-1(qm30) mutants in a dose-dependent manner, but showed no significant effect on the wild-type (N2).

Figure 3

The behavioural defects of clk-1(qm30) mutants are fully rescued by 2,4-DHB. (a) clk-1 mutants pump at a significantly slower rate than the wild type. 1 mM 2,4-DHB fully rescues the slow pumping rates of clk-1. At this concentration, 2,4-DHB had no effect on the wild type. The bars represent the mean number of pumps per minute. (b) clk-1 mutants have a significantly lengthened defecation cycle length. 1 mM 2,4-DHB fully rescues the slow defecation phenotype of clk-1. 2,4-DHB had no effect on the wild type. The bars represent the mean defecation cycle of animals that have been scored for three consecutive defecation cycles each in the case of non-treated clk-1(qm30) mutants and for five consecutive defecation cycles for all the other conditions. The error bars represent S.E.M. (n ≥ 20 animals for each condition). The asterisks indicate that the data are significantly different from that of the wild-type. All differences were significant at p < 0.0001 by a t-test.

Figure 4

The longevity of clk-1 mutants is rescued by 2,4-DHB. (a) Supplementation with 0.15 mM and 1 mM 2,4-DHB fully rescues the lengthened lifespan of clk-1 mutants. (b) Supplementation with either 0.15 mM or 1 mM 2,4-DHB has no effect on the wild type lifespan. (c) Treatment with 1 mM 3,4-DHB does not affect the lifespan of either clk-1 mutants or the wild-type. (d) Treatment with 1 mM 2,4-DHB rescues the lengthened lifespan of clk-1(e2519) mutants and has no effect on wild type lifespan. Sample sizes, numerical values and statistical analyses for all lifespan experiments are presented in Table S2.

Figure 5

CLK-1(ΔMTS)::GFP cannot rescue the phenotypes of clk-1 mutants (a) clk-1(qm30) mutants grow slowly when fed OP50 and arrest when fed UQ₈-deficient bacteria GD1. Both phenotypes are rescued by a single copy insertion of full length CLK-1::GFP. Neither phenotype is rescued by a single copy insertion of CLK-1(ΔMTS)::GFP, which lacks the mitochondrial targeting sequence of clk-1. (n ≥ 30 animals for each genotype). (b) The slow pumping rate and slow defecation of clk-1(qm30) mutants can be fully rescued by the expression of full length CLK-1::GFP but not by the expression of CLK-1(ΔMTS)::GFP. Bars represent the mean number of pumps per minute or the mean defecation cycle of animals that have been scored for three consecutive defecation cycles each in the case of clk-1(qm30) mutants and for five consecutive defecation cycles for all the other genotypes. The error bars represent S.E.M. (n ≥ 20 animals for each genotype). The asterisks indicate that the data are significantly different from that of the wild-type. All differences were significant at p < 0.0001 by a t-test. (c), (d) and (f) The expression of CLK-1(ΔMTS)::GFP had no effect on the lifespans of clk-1(qm30), clk-1(qm30); daf-2(e1370) or clk-1(e2519) mutants. (e) Treatment with 1 mM 2,4-DHB rescues the lengthened lifespan of clk-1(qm30); CLK-1(ΔMTS)::GFP and has no effect on the lifespan of the wild type. Numerical values and statistical analyses for all lifespan experiments are presented in Table S2.

Figure 6

Changes in gene expression in clk-1 mutants can be rescued by the expression of CLK-1::GFP or by 2,4-DHB treatment. (a,b) The mRNA level of UPR^mt genes was examined in day 1 adult worms by quantitative real-time PCR. Results were normalized to mRNA levels in wild type. clk-1(qm30) worms showed altered expressions of hsp-60 and hsp-6. The expression of CLK-1::GFP, but not CLK-1(ΔMTS)::GFP, rescues the transcript levels of hsp-60 and hsp-6 in clk-1 mutants. The same rescuing effect can be achieved by treating clk-1 mutants with 2,4-DHB. (c) The spg-7 mRNA levels were the same for the all genotypes that we examined. (d,e) Quantification of GFP intensity from Phsp-6::GFP. Values are expressed as fold change in GFP fluorescence intensities relative to the GFP fluorescence intensity of the Phsp-6::GFP reporter in the wild type background. The treatment of 2,4-DHB decreased the reporter expression in clk-1 mutants, while it had only a small effect on the control strains. Error bars indicate S.E.M. (n ≥ 30 animals). *p < 0.05, **p < 0.01, ***p < 0.001. Scale bar: 100 µm.

Figure 7

Detection of MCLK1 in protein fractions. Western blot analysis of nuclear protein extracts prepared from mouse embryonic fibroblasts (MEFs) (a) and Western blot analysis of nuclear protein fractions from mouse tissues (b). Mclk1 KO cells were generated as previously published and the KO tissue samples were obtained from agoMclk1 KO mice described in ref. . They were used to further confirm band identity. The wild-type controls are Mclk1 floxed MEFs or tissues without Cre expression. In both panels, 40 µg proteins per lane were loaded except for the membrane fraction (MF), where 5 µg were used. HDAC1 was used as a loading control of nuclear soluble fractions (NF). For insoluble (chromatin-bound) nuclear fractions (CBF), lamin A/C and Histone H3 served as markers. The mitochondrial matrix protein SOD2 was used as an additional control to validate the purity of the nuclear protein fractions. (a) In MEFs, MCLK1 could only be detected in membrane fractions (MF). (b) In mouse tissues, MCKL1 could be detected in both nuclear soluble fractions (NF) and membrane fractions (MF). However, the mitochondrial matrix protein SOD2 could also be detected in both fractions suggesting that the nuclear soluble fraction was contaminated by mitochondria. MCLK1 could not be detected in the chromatin-bound nuclear fraction (CBF). No evidence therefore points to the presence of MCLK1 in the nucleus or an association with chromatin. Uncropped western blot scans with size indications are shown in Supplementary Data.