Elevated mitochondrial DNA copy number found in ubiquinone-deficient clk-1 mutants is not rescued by ubiquinone precursor 2-4-dihydroxybenzoate

Abstract

Inside mitochondria reside semi-autonomous genomes, called mtDNA. mtDNA is multi-copy per cell and mtDNA copy number can vary from hundreds to thousands of copies per cell. The variability of mtDNA copy number between tissues, combined with the lack of variability of copy number within a tissue, suggest a homeostatic copy number regulation mechanism. Mutations in the gene encoding the Caenorhabditis elegans hydroxylase, CLK-1, result in elevated mtDNA. CLK-1’s canonical role in ubiquinone biosynthesis results in clk-1 mutants lacking ubiquinone. Importantly, clk-1 mutants also exhibit slowed biological timing phenotypes (pharyngeal pumping, defecation, development) and an activated stress response (UPR^mt). These biological timing and stress phenotypes have been attributed to ubiquinone deficiency; however, it is unknown whether the mtDNA phenotype is also due to ubiquinone deficiency. To test this, in animals carrying the uncharacterized clk-1(ok1247) mutant allele, we supplemented with an exogenous ubiquinone precursor 2–4-dihydroxybenzoate (DHB), which has previously been shown to restore ubiquinone biosynthesis. We measured phenotypes as a function of DHB across a log-scale range. Unlike the biological timing and stress phenotypes, the elevated mtDNA phenotype was not rescued. Since CLK-1’s canonical role is in ubiquinone biosynthesis and DHB does not rescue mtDNA copy number, we infer CLK-1 has an additional function in homeostatic mtDNA copy number regulation.

1. INTRODUCTION

Mitochondrial DNA (mtDNA) is closely linked to energy requirements as the mitochondrial genome encodes essential components of the electron transport chain, which generates ATP. Unlike nuclear DNA replication, mtDNA replication occurs independently of the cell cycle mtDNA copy number ranges from hundreds to thousands of copies per cell and predictably varies within one cell over time during development and aging [1, 2]. Further, copy number varies between cell types and correlates with energy requirement [3]. A minimum copy number is crucial for survival and development; insufficient copy number can result in mtDNA depletion syndromes characterized by myopathies and infertility [4–6]. Excess mtDNA is positively correlated with risk of some cancers [7]. The non-static nature of mtDNA copy number suggests that copy number is actively maintained. Thus, the dynamic nature of mtDNA copy number, coupled with the consequences of incorrect copy number, suggest a mechanism of tightly controlled homeostatic copy number regulation.

While the mechanism of copy number regulation is not yet fully understood, many mechanisms have been posited [8]. 1) A limiting factor hypothesis has been proposed, in which the depletion of some factor involved in mtDNA maintenance limits mtDNA replication [8]. Factors involved in replication (ex: mtDNA transcription factor A [TFAM] and mtDNA polymerase [POLG-1]) have been implicated, because modulation of these factors results in alterations in mtDNA copy number [9–12]. However, as these factors are involved in replication, it is difficult to differentiate between active copy number regulation and the inability to replicate. 2) It has been proposed that overall mtDNA mass is being regulated [8]. Supporting this theory are data demonstrating that mtDNA copy number is inversely proportional to mtDNA genome size [13]. One explanation for these data is that mtDNA copy number is regulated by pools of nucleotides that are incorporated during mtDNA replication. The exhaustion of nucleotides would halt mtDNA replication, such that a tightly regulated pool of nucleotides would result in a consistent mass of mtDNA. However, this model has not been explicitly tested. 3) It has been suggested that mtDNA levels are maintained based on a feature (ex: sequence or product) of the mtDNA molecules. mtDNA can exist in a state of heteroplasmy, in which two different mtDNA haplotypes reside within the same cell. In one study using human muscle harboring both wild type and mutant mtDNA molecules, researchers measured both wild type and mutant mtDNA molecules and found that wild type mtDNA levels are maintained, even as mutant mtDNA levels vary [14]. In a previous study, our lab recapitulated this work using the nematode Caenorhabditis elegans. We found that in one heteroplasmic strain (uaDf5), worms containing both wild-type and uaDf5-deletion mtDNA haplotypes exhibit an elevation in total mtDNA copy number. Wild-type mtDNA copy number in uaDf5-bearing worms was similar to that of worms containing only wild-type mtDNA molecules, even as mutant mtDNA levels varied [15]. These data suggest the presence of a homeostatic copy number regulation mechanism working to maintain wild type mtDNA levels in C. elegans.

C. elegans provides a robust model for studying copy number regulation for a few reasons. First, in animals, mtDNA content is highest in oocytes: oocytes contain orders of magnitude more mtDNA than somatic cells. Due to the high copy number within the germline, many previous studies working to understand the mechanism of active copy number regulation have been conducted in oocytes. To quantify mtDNA, many organisms that contain germlines require time-consuming oocyte dissection. Oocyte dissection is not required in C. elegans because the C. elegans germline is a large syncytium containing a shared pool of organelles and ooplasm which accounts for 95% of the total mtDNA content of the worm. Because of the germline syncytium, we can lyse whole individual worms to measure mtDNA in the germline. Second, C. elegans has highly stereotypical development and behavior, enabling phenotypic characterization. Third, C. elegans has a robust set of genetic variants and pharmacologic manipulations available. These characteristics, along with the data suggesting active copy number regulation in C. elegans, position worms as a useful model for identifying the mechanism of homeostatic copy number regulation.

To understand the mechanism of homeostatic copy number regulation, we sought a condition in which copy number is both reproducibly and predictably perturbed. Previous work in C. elegans demonstrated that adult clk-1 mutants exhibit a 30% increase in mtDNA copy number compared to wild-type [16]. The clk-1 gene encodes the mitochondria-localized hydroxylase CLK-1, which catalyzes the penultimate step in ubiquinone biosynthesis [17–23]. CLK-1’s role in ubiquinone biosynthesis is to catalyze the conversion of 5-demethoxyubiquinone (DMQ) to demethylubiquinone (DMeQ) which the enzyme COQ-3 then converts to ubiquinone (Supplementary Fig S1). Consequently, CLK-1-deficient cells accumulate DMQ (except in - coq7/cat5 null yeast where the starting substrate 3-hexaprenyl-4-hydroxybenzoic acid [HHB] is accumulated instead), and exhibit a significant lack of ubiquinone [18, 19, 24–27]. Given ubiquinone’s role in introducing electrons into the electron transport chain, it is not surprising that clk-1 mutants would exhibit phenotypic consequences.

clk-1 mutants exhibit additional phenotypes including slowed biological timing and mitochondrial unfolded protein response (UPR^mt) activation [17, 28–30]. CLK-1 is conserved throughout eukaryotes, and clk-1 mutants exhibit similar phenotypes in all organisms that have been studied. Studies of CLK-1 and its orthologs COQ7/CAT5 and MCLK-1 have been carried out in yeast and mouse models. In budding yeast, deletion of coq7/cat5 prevents growth on non-fermentable carbon sources [19, 31]. MCLK-1-deficient mouse embryos arrest in early development and cultured cells derived from those embryos exhibit slow cell multiplication and reduced differentiation and mice heterozygous for MCLK-1 are long lived [32–34]. Only three cases of humans harboring coq-7 deletions appear in the literature. In one case, a coq-7 mutation resulted in early infantile death [27], while the other two patients exhibited varying levels of hearing and motor impairment [25, 26]. Together, these phenotypes highlight the widespread consequences of clk-1 mutations.

While CLK-1’s ubiquinone biosynthesis role is well known, it has long been suggested that CLK-1 serves an additional cellular function. The widespread impacts of clk-1 mutations led researchers to posit that CLK-1 maintains a role as a transcription factor, acting in service of mitochondrial-nuclear communication [28]. One group identified a distinct nuclear form of CLK-1 and proposed an additional role in cellular stress responses [35]. In the most recent study, another group demonstrated a lack of nuclear CLK-1 and a rescue of all tested CLK-1 phenotypes solely with mitochondrial CLK-1 [36]. In this study, clk-1 mutant worms were fed the unnatural ubiquinone precursor 2–4-dihydroxybenzoate (DHB), which is converted to ubiquinone in a CLK-1 independent manner (Supplementary Fig S1). DHB supplementation rescued the biological timing and mitochondrial stress phenotypes [36]. This experimental approach demonstrates that these phenotypes are due to lack of ubiquinone in clk-1 mutants and thus are related to CLK-1’s canonical ubiquinone biosynthesis function [36]. This study did not measure mtDNA copy number. The elevated mtDNA copy number phenotype found in clk-1 mutants has previously been attributed to a lack of ubiquinone. It has been proposed that this lack of ubiquinone results in decreased respiration, causing mitochondrial compensation and over replication of mtDNA [16]. However, one group provides in vitro evidence that CLK-1 serves a role in regulating mtDNA [37]. In vitro, mouse and C. elegans CLK-1 protein binds directly to short oligonucleotide probes containing sequences corresponding to mtDNA from the respective species, suggesting a possible additional role for CLK-1 [37]. These conflicting studies demonstrate that there is lack of clarity in CLK-1’s cellular role(s). Taken together, we sought to test whether the elevated copy number phenotype in clk-1 mutants is due to an additional role for CLK-1.

It has not previously been reported whether the clk-1 mutant mtDNA copy number phenotype can be rescued with DHB supplementation, like the other phenotypes. Since CLK-1 serves a known role in ubiquinone biosynthesis, it is challenging to decipher any additional roles for CLK-1. To answer this question, we measured five phenotypes as a function of DHB concentration across a log-scale in individual worms lacking CLK-1. This design is novel as it addresses the relationship between DHB concentration and the magnitude of phenotype rescue. By determining the relationship between DHB supplementation and each mutant phenotype, we infer causality of each phenotype. We reasoned that any phenotype that is not rescued proportionally with DHB concentration occurs independently of ubiquinone levels. A clk-1 mutant phenotype that occurs independently of ubiquinone levels would support an additional role for CLK-1. Taken together, we employed a methodical log-scale DHB supplementation approach to determine whether CLK-1’s ubiquinone biosynthesis role is responsible for the elevated clk-1 mutant copy number phenotype.

Here we characterize C. elegans clk-1(ok1247), a previously uncharacterized clk-1 mutant allele. We employ a systematic approach to test the relationship between DHB supplementation and mtDNA copy number. We demonstrate that the clk-1(ok1247) mutant copy number phenotype is not rescued by DHB supplementation, even at concentrations that rescue the biological timing and mitochondrial stress phenotypes. Given that the clk-1 mutant copy number phenotype is not rescued by DHB supplementation, we infer that the elevated copy number phenotype is likely due to another role of CLK-1. We speculate that CLK-1 may directly regulate mtDNA copy number, independent of CLK-1’s canonical ubiquinone biosynthesis function.

2. RESULTS

2.1. clk-1(ok1247) mutation is a null allele with no protein produced.

The two previously characterized alleles (qm30 and e2519) may both be hypomorphs [29, 32, 38] (Fig 1A). Thus, we used the clk-1(ok1247) allele to prevent confoundment of results that might occur in those hypomorphs. We began by establishing the specific nature of the mutation. The clk-1(ok1247) deletion breakpoints had not been mapped, thus, we determined the precise deletion location and found a near-complete deletion of the clk-1 gene (Fig 1B and C). As expected, we observed no CLK-1 protein in clk-1(ok1247) animals using the CLK-1 antibody we generated (Fig 1D and E). We confirmed the utility of the CLK-1 antibody in the two previously characterized alleles. We found that clk-1(qm30) animals did not contain CLK-1 protein, while clk-1(e2519) animals did contain CLK-1 protein, which is consistent with findings in the literature (Supplementary Fig S2A-D) [32].

Figure 1. The previously uncharacterized clk-1(ok1247) allele is a large deletion and does not produce protein.

A) Schematic of region containing C. elegans clk-1 gene and relevant alleles. Exons are numbered. Black arrow indicates location of point mutation; nucleotide substitution is indicated above the arrow. Dashed region indicates deletion. Allele colors in this figure correspond to allele colors in other figures within this manuscript: wild-type (grayscale); clk-1(e2519) red; clk-1(qm30) green; clk-1(ok1247) purple. Nucleotide positions of start and end of each exon are shown with numbering beginning at start of exon 1. B) PCR genotyping of clk-1(ok1247) mutation using previously published from the Caenorhabditis Genetics Center (CGC). Expected wild-type band is 2,648 base pairs, mutant band is 1,334 base pairs. First lane wild-type (N2), second lane mutant clk-1(ok1247), third lane no template control (ntc), final lane 1kb Plus DNA ladder. C) clk-1(ok1247) deletion breakpoint sequence as determined by Sanger sequencing. D) Due to the lack of a C. elegans-specific CLK-1 antibody, we generated a novel polyclonal antibody in rabbits. This antibody shows good specificity. Using this in-house generated antibody, we demonstrated the absence of CLK-1 protein in clk-1(ok1247). N2 and clk-1(ok1247) “undiluted” samples were diluted in “diluted” lanes. Mutant lanes at same dilution contain more total worm lysate than wild-type lanes. However, they appear to have less protein, which may be due to clk-1 mutant worms being smaller. E) Loading control gel visualization of total protein using TGX stain-free gel activation of the gel used for the blot in D.

2.2. clk-1(ok1247) mutants exhibit elevated mtDNA copy number phenotype.

Since the clk-1(ok1247) allele is a null deletion, we proceeded with recapitulating the previously reported elevated copy number phenotype. To assess total copy number, we measured mtDNA using multiplex qPCR. We normalized mtDNA copy number to an internal nuclear gene control (actin). Given that our experiments are conducted at a single developmental time point and are not in a mixed population of worms, we found the nuclear gene control to be effective. We compared total mtDNA copy number in whole larval stage 4 (L4) individual wild-type (N2) and clk-1(ok1247) mutant worms. Here we report for the first time an elevation in mtDNA copy number in the clk-1(ok1247) mutant allele (Fig 2). To demonstrate the rigor of our quantitation method, we used two primer pairs (primer schematic shown in Supplementary Fig S3); both primer pairs detected a similar 30% elevation in copy number. Further, to demonstrate reproducibility of this phenotype, we repeated the experiment in a) another cohort of clk-1(ok1247) L4 worms, b) clk-1(ok1247) L4 worms on an alternative bacterial food source (HT115) and c) clk-1(ok1247) adult worms at day 4, day 5, and day 6 adulthood (Supplementary Fig S4A-C). This elevation in the clk-1(ok1247) allele is congruent with the previously published elevated copy number phenotype in the clk-1(qm30) allele (Supplementary Fig S5) [16]. While previous work identified this copy number phenotype in adult worms, the bulk of our study was conducted in L4 worms, suggesting continued maintenance of this phenotype throughout development.

Figure 2. clk-1(ok1247) mutant copy number phenotype is reproducible across two primer pairs.

Relative mtDNA copy number measured by qPCR in individual worms and normalized to sample’s nuclear DNA copy number (actin). Wild-type (N2) gray circles (n=16), clk-1(ok1247) mutants purple triangles (n=11). Striped bars depict primer pair A, dotted bars depict primer pair B. Bars represent mean and SD. One-way ANOVA with two pairwise comparisons, primer pair A p= 0.0139, primer pair B p= 0.0080.

2.3. C. elegans can use bacterial ubiquinone to survive, but not to thrive.

After confirming the presence of the copy number phenotype in worms carrying the uncharacterized allele, we sought to test the link between clk-1 mutant phenotypes and ubiquinone biosynthesis. Ubiquinone biosynthesis is crucial for survival because ubiquinone (UQ) introduces electrons into the electron transport chain. UQ is made up of a benzoquinone head group capable of electron exchange, and a lipid-based side chain containing isoprenoid units [32, 39]. The number of isoprenoid units in the side chain is species-specific and denoted by the subscript after UQ: E. coli produce UQ₈, while C. elegans and M. musculus produce UQ₉ and H. sapiens produce UQ₁₀ [39]. C. elegans deficient in CLK-1 grown on a UQ-deficient E. coli food source arrest and eventually develop into sterile adults [40]. clk-1 mutant worms grown on UQ-replete E. coli develop and produce progeny but exhibit the biological timing and stress phenotypes described in the introduction [40]. Together, these data suggest that worms can utilize bacterial-derived UQ₈ to survive, but cannot take up enough UQ₈ to thrive, resulting in the biological timing and stress phenotypes [41].

2.4. Supplementation with a ubiquinone precursor can be achieved by feeding.

Until recently, the lack of UQ was only inferred to be the cause of the biological timing and stress phenotypes. Recent work has demonstrated the causal relationship between ubiquinone and some phenotypes by feeding worms the unnatural UQ precursor 2–4-dihydroxybenzoate (DHB) [36]. Unlike the endogenous UQ precursor, DHB is converted to UQ via a CLK-1-independent mechanism (Supplementary Fig S1). We reasoned that any mutant phenotypes caused by loss of UQ in clk-1 mutants will be rescued by DHB supplementation. Conversely, any phenotype not rescued by DHB supplementation likely represents a process independent of CLK-1’s UQ biosynthesis role (Fig 3). To elucidate the relationship between UQ and clk-1 mutant phenotypes, we measured phenotype rescue across a log-scale range of DHB dose. We treated bacteria-seeded plates with DHB, plated five larval stage 4 (L4) worms, maintained them for one generation, and conducted all phenotype measurements on F1 L4s. Because direct UQ supplementation only restores some clk-1 mutant phenotypes but not all phenotypes, we used the more consistent DHB supplementation method [40].

Figure 3. Schematic representation of experimental design to determine which CLK-1 role is responsible for elevated mtDNA copy number.

Wild-type animals with CLK-1 protein and ubiquinone maintain typical mtDNA copy number (left). clk-1 null mutants do not produce ubiquinone and exhibit elevated mtDNA copy number (right). To determine if the clk-1 mutant elevated mtDNA copy number phenotype is caused by lack of ubiquinone or due to lack of CLK-1 protein, worms were treated with an exogenous ubiquinone precursor, 5-demethoxyubiquinone (DHB) (center). We reasoned that if the elevated mtDNA copy number in clk-1 worms is caused by lack of ubiquinone, copy number should be rescued upon DHB supplementation. If copy number is not rescued, we infer an additional role for CLK-1 in regulating mtDNA copy number. Solid lines indicate presence, dashed lines indicate absence.

2.5. Development time from embryo to L4 is strongly rescued by DHB treatment.

To confirm that our system of DHB supplementation is effective, we began by assessing the canonical clk-1 mutant phenotype for which the “clock” mutants are named: delayed time to development. C. elegans exhibit a stereotypical development, with distinct stages: embryo, larval stages 1–4, young adult, and gravid adult. Previous work has demonstrated both developmental delay and lengthened lifespan phenotypes in clk-1 mutants [41, 42]. The lengthened lifespan phenotype is rescued by DHB supplementation [36]. The developmental delay phenotype has not been studied with respect to DHB supplementation. clk-1 mutants exhibit both slowed embryonic and post-embryonic development, compared to wild-type (N2) worms [16, 17, 41, 42]. In our study, we measured how many hours lapsed from embryo lay to larval stage 4 (L4). The experiment was completed once all animals reached L4 stage. We found that clk-1(ok1247) mutant development time was slow on 0mM DHB treatment. Development time visibly hastened in a DHB dose-dependent manner and at 100mM DHB treatment, mutant development time was visually similar to wild type (Fig 4A). Using a Cox proportional hazards model, we determined that mutant genotype had a significantly lower hazard ratio compared to wild-type (0.22), while DHB treatment and genotype-by-treatment interaction had hazard ratios of 1.00 and 1.01, respectively (Supplementary Fig S6).

Figure 4. clk-1 mutant biological timing and stress phenotypes are rescued as a function of DHB treatment concentration.

A) Time to development as measured by counting and removing worms that have reached L4 or older at each given time point. For each genotype and treatment combination, three replicates were plated, and vertical lines represent the mean and standard error of the mean of those three plates. Wild-type (N2) shown as gray solid line, clk-1(ok1247) shown as purple dashed line. clk-1(ok1247) mutants exhibit slow growth on 0mM DHB but visibly hastened growth on 100mM DHB. Statistics were determined based on a multiple linear regression and Cox proportional hazards measurements. Hazard ratio: DHB dose 1.00, genotype 0.2233, DHB dose by genotype interaction 1.0070. B) Pharyngeal pump rate measured as pumps per minute. Wild-type (N2) shown as gray circles (n=20), clk-1(ok1247) shown as purple triangles (n=20). Bars represent mean. clk-1(ok1247) mutants on 0mM DHB exhibit slowed pump rate phenotype (one-way ANOVA comparison, p< 0.0001) but a smaller magnitude difference on 100mM DHB (one-way ANOVA comparison p= 0.0062). Linear regression: wild-type (slope= −0.0942) and clk-1 mutants (slope= 0.8600). Wild-type did not deviate from zero (p= 0.1988), while clk-1 mutants did deviate from zero (p< 0.0001). Slopes are significantly different from each other (p< 0.0001). C) Defecation cycle time measured as seconds between pBoc. Wild-type (N2) shown as gray circles (n=15), clk-1(ok1247) shown as purple triangles (n=15). Bars represent mean. clk-1(ok1247) mutants exhibit increased time between pBocs on 0mM DHB (one-way ANOVA comparison p< 0.0001), but normal time between defecations on 100mM DHB (one-way ANOVA comparison p= 0.8831). Linear regression: wild-type (slope = −0.03315) and clk-1 mutant (slope = −0.2990). Wild-type slightly deviated from zero (p= 0.0145), while clk-1 mutants strongly deviated from zero (p< 0.0001). Slopes are significantly different from each other (p< 0.0001). D) UPR^mt activation as measured by GFP fluorescence driven by the hsp-6 promoter. Wild-type (N2) shown as gray circles (n=6–9), clk-1(ok1247) shown as purple triangles (n=7–9). Bars represent mean. clk-1(ok1247) mutants exhibit UPR^mt activation on 0mM DHB (one-way ANOVA comparison p<0.0001) but no UPR^mt activation on 100mM DHB (one-way ANOVA comparison p=0.6548). Linear regression: wild-type (slope= −0.4936) and clk-1 mutants (slope= −8.423). Wild-type did not deviate from zero (p= 0.6035), while clk-1 mutants strongly deviated from zero (p= 0.0024). Slopes are significantly different from each other (p= 0.0114).

2.6. Pharyngeal pumping rate is strongly rescued by DHB treatment.

Following the robust rescue of overall developmental timing, we next tested individual biological timing phenotypes, starting with pharyngeal pump rate. C. elegans feed by pulling bacteria in liquid into the pharynx. Liquid is then expelled, and bacteria remain in the pharynx. The bacteria are then passed to the terminal bulb and ground by the grinder. This behavior is rhythmic, and controlled by motor neurons [43, 44]. clk-1 mutants exhibit slowed pharyngeal pumping, and previous work demonstrates rescue by DHB supplementation [16, 17, 36, 42]. To measure the pharyngeal pump rate per minute, we counted the number of movements of the grinder in thirty seconds and multiplied the result by two (pharyngeal pumping shown in Supplementary Video S1). We found that the clk-1(ok1247) mutant pharyngeal pump rate was slow on 0mM DHB. Pharyngeal pump rate increased in a DHB dose-dependent manner and at 100mM DHB, the pharyngeal pump rate in mutants was rescued nearly to that of wild-type (Fig 4B).

2.7. Defecation cycle length is completely rescued by DHB treatment.

Given the strong rescue of pharyngeal pump rate, we tested another biological timing phenotype: defecation cycle length. C. elegans have a thoroughly-studied ultradian defecation program, marked by five distinct phases: intercycle, posterior body wall muscle contraction (pBoc), relaxation, anterior body wall muscle contraction (aBoc), and expulsion [45]. clk-1 mutants exhibit increased defecation cycle time length, and previous work demonstrates rescue by DHB supplementation [17, 36]. To measure defecation cycle time, we asked how many seconds from one pBoc to the next pBoc, cycling through each of the other phases (pBoc shown in Supplementary Video S2). We found that clk-1(ok1247) mutant defecation cycle time was slow on 0mM DHB treatment. Defecation cycle time decreased in a DHB dose-dependent manner and at 100mM DHB, defecation cycle time in wild type and mutant did not differ (Fig 4C).

2.8. Mitochondrial Unfolded Protein Response (UPR^mt) activation is completely rescued by DHB treatment.

Following the rescue of biological timing phenotypes, we investigated whether DHB could rescue clk-1(ok1247) mitochondrial stress. The mitochondrial unfolded protein response (UPR^mt) is activated by mitochondrial stressors that impact protein homeostasis, such as defective chaperones and impaired mitochondrial protein import or synthesis [46]. It has previously been demonstrated by fluorescence imaging and mRNA transcript quantification that clk-1(qm30) mutants exhibit strong activation of the UPR^mt [36]. clk-1 mutant UPR^mt activation appears to be due to the stress induced by loss of CLK-1’s enzymatic role in generating ubiquinone, since it is rescued by DHB supplementation [36]. We measured the activation of UPR^mt by quantifying GFP fluorescence under the control of the promoter of heat-shock protein hsp-6 (Phsp6::gfp), a marker of mitochondrial stress (representative images shown in Supplementary Fig S7). We found high fluorescence in clk-1(ok1247) mutants with 0mM DHB treatment, demonstrating significant UPR^mt activation. We found that fluorescence decreased in a DHB dose-dependent manner and at 100mM DHB treatment fluorescence in mutants was rescued to that of wild type (Fig 4D).

2.9. Elevated mitochondrial DNA (mtDNA) copy number phenotype is not rescued by DHB treatment.

Given the robust rescue of the biological timing and mitochondrial stress phenotypes, we evaluated whether DHB could rescue the clk-1(ok1247) mtDNA copy number phenotype. We measured mtDNA copy number in wild-type and clk-1(ok1247) mutants. We again confirmed the elevation in copy number on 0mM DHB. However, we found no change in mtDNA copy number, regardless of DHB dose. At 100mM DHB treatment, mutant copy number was still significantly different from wild type (Fig 5A). We repeated the experiment with an additional cohort of clk-1(ok1247) mutants using 0mM and 100mM DHB treatments to validate the experimental reproducibility, and again found no rescue of mtDNA levels (Supplementary Fig S8).

Figure 5. clk-1(ok1247) mutant copy number is not rescued by DHB supplementation and appears to be caused by lack of CLK-1 protein, not lack of ubiquinone.

A) Relative mtDNA copy number measured by qPCR in individual worms and normalized to sample’s nuclear DNA copy number (actin). Wild-type (N2) depicted as gray circles (n=15–16), clk-1(ok1247) depicted as purple triangles (n=15–16). Bars represent mean. clk-1(ok1247) mutants exhibit elevated mtDNA copy number on 0mM DHB (one-way ANOVA comparison p= 0.0003), and on 100mM DHB (one-way ANOVA comparison p< 0.0001). Linear regression: wild-type animals (slope= 0.0003) and clk-1 mutant animals (slope= 0.0010). The slopes did not significantly differ from each other (p= 0.5469). Elevation of lines differ (p< 0.0001), indicating a strain difference in copy number. Primer pair A used for mtDNA quantitation. B) Mutants harboring the clk-1(ok1247) null allele, produce no CLK-1, and subsequently no ubiquinone. Worms treated with the exogenous ubiquinone precursor DHB exhibit rescue of other clk-1 mutant phenotypes, but do not exhibit rescue of the mutant elevated copy number phenotype (left). Thus, we infer that CLK-1 regulates mtDNA copy number through a ubiquinone-independent mechanism (right).

3. DISCUSSION

3.1. clk-1 mutants exhibit biological timing, stress, and mtDNA copy number phenotypes.

We searched for a condition or mutation that leads to elevated copy number in a manner separable from mitochondrial dysfunction. Previous work demonstrated a 30% increase in copy number in a clk-1 mutant allele [16]. Since CLK-1 catalyzes the penultimate step in ubiquinone biosynthesis, clk-1 mutants exhibit a lack of ubiquinone, which leads to mitochondrial dysfunction. clk-1 mutant biological timing and mitochondrial stress phenotypes have been attributed to this loss of ubiquinone [36]. It has not been previously reported whether the clk-1 mutant copy number phenotype is linked to ubiquinone biosynthesis. In this study, using a pharmacologic intervention, we separate the copy number phenotype from ubiquinone biosynthesis.

We studied a previously uncharacterized large deletion allele clk-1(ok1247). Since we wondered if the elevated mutant mtDNA copy number phenotype was caused by lack of ubiquinone, we endeavored to unlink the two. To do this, we supplemented clk-1(ok1247) mutants with a CLK-1 independent ubiquinone precursor, DHB. Since conversion of DHB to ubiquinone does not require CLK-1, we expected that a ubiquinone-dependent phenotype will be rescued proportionally with the dose of DHB. We reasoned that any phenotype not rescuable by DHB is not caused by clk-1 mutant ubiquinone deficiency.

3.2. clk-1 mutant biological timing and stress phenotypes are rescued by ubiquinone supplementation.

In this study, we demonstrate that the previously studied biological timing and mitochondrial stress phenotypes (time to development, pharyngeal pump rate, defecation cycle length, and UPR^mt) were rescued in a DHB dose-dependent manner, such that DHB dose correlated with intensity of mutant phenotype rescue. This dose-dependent rescue suggests a direct coupling between these phenotype and ubiquinone biosynthesis. Given that the highest dose (100mM) rescues phenotypes to that of wild-type or near wild-type, we conclude that the range of DHB dose is appropriate for answering the question of dose-response. Further, previous work using DHB supplementation in clk-1 mutants shows that while DHB supplementation only partially rescues ubiquinone levels, partial rescue of ubiquinone fully rescues biological timing and stress phenotypes [36]. These data support the previous findings that the biological timing and mitochondrial stress phenotypes are caused by a lack of ubiquinone and confirm that our system for testing the cause of clk-1 mutant copy number phenotype is efficacious.

3.3. clk-1 mutant copy number phenotype is not rescued by ubiquinone supplementation.

Finally, we asked whether the clk-1 mutant copy number phenotype could be rescued by DHB supplementation in the same manner as the other phenotypes. Given the dose-dependent rescue of the other phenotypes, we expected that the copy number phenotype should also be rescued in a dose-dependent manner if it is caused by lack of ubiquinone. Further, even if the phenotype is not entirely rescued, if the phenotype is caused by lack of ubiquinone, the phenotype should inversely correlate with DHB concentration. We found that clk-1 mutant copy number was not rescued to wild-type levels at any concentration of DHB supplementation. Notably, the same scale of DHB concentration was sufficient to rescue the biological timing and stress phenotypes but was not sufficient to rescue the mtDNA copy number phenotype, even partially. Since DHB treatment does not restore copy number to wild-type levels, we infer that clk-1(ok1247) mutant copy number phenotype is not due to lack of ubiquinone. This exciting finding suggests an additional role for CLK-1.

3.4. Does CLK-1 serve another function?

Our study suggests CLK-1 could serve another cellular role. Previous studies suggested an additional role for CLK-1. A 2015 paper suggests that a nuclear-localized form of CLK-1 serves a homeostatic signaling role in regulating lifespan that is responsive to reactive oxygen species [35]. However, more recent work suggests that CLK-1 does not localize to the nucleus [36]. Additional work has suggested that CLK-1 could, like other proteins, exhibit dual functionality by serving two roles [37]. It has not been previously reported whether the clk-1 mutant copy number phenotype is linked to ubiquinone biosynthesis. However, it has been suggested that the clk-1 elevated mtDNA copy number phenotype is due to compensation for respiration rate defects caused by ubiquinone deficiency [16]. It will be interesting to test the biochemical consequences of elevated copy number due to loss of CLK-1. In this study, using a pharmacologic intervention we separate the copy number phenotype from ubiquinone biosynthesis and infer that CLK-1 serves an additional role (Fig 5B).

3.5. CLK-1 and mtDNA reportedly bind in an ADP-dependent manner.

One group assessed CLK-1’s ability to bind to mtDNA [37]. In vitro they combined purified C. elegans CLK-1 and various oligonucleotides with sequences corresponding to various regions of C. elegans mtDNA. They found that CLK-1 bound to a specific sequence in the predicted Origin of Light strand replication (O_L) [37]. The O_L is known to be required for mtDNA replication in other species and is hypothesized to play a role in mtDNA replication in C. elegans. Further, the binding of CLK-1 to the O_L oligonucleotide was strongly inhibited by the addition of ADP and mildly increased by the addition of ATP [37]. Importantly, they found the same results using mouse CLK-1 (MCLK-1) and corresponding mouse O_L sequences. These findings suggest that CLK-1’s interaction with mtDNA is broadly conserved.

3.6. Could CLK-1 serve a regulatory role?

Given that mtDNA encodes essential components of the electron transport chain (ETC), we predict that mtDNA copy number could be regulated based upon energy levels. Taken together, the following data position CLK-1 as a regulator of mtDNA copy number. First, only a few proteins directly interact with mtDNA [47]. Second, at least in vitro, the binding of CLK-1 to mtDNA is directly responsive to a cofactor involved in ETC function: ADP [37]. Since ADP is phosphorylated to ATP via complex V of the ETC, ADP levels are predicted to negatively correlate with ETC function. Third, if C. elegans O_L function is conserved and is involved in mtDNA replication, the binding of CLK-1 to O_L could serve to inhibit mtDNA replication. Given the inhibition of CLK-1 binding to mtDNA by ADP, this positions CLK-1 as a physical regulator of mtDNA replication in response to energy levels [37]. In this model, CLK-1 would act as a homeostatic sensor to negatively regulate mtDNA copy number such that lack of CLK-1 would result in elevated copy number. Our clk-1 mutant copy number data support this homeostatic sensor model. In fact, earlier modeling work predicted the existence of a homeostatic copy number regulation mechanism and was predicated upon a homeostatic sensor that detects a functional output of mtDNA [48]. Since CLK-1’s in vitro binding capacity is ADP-dependent, and ADP is phosphorylated to become a functional output of the electron transport chain, we speculate that CLK-1 could serve the role of the proposed functional output sensor. This model links mtDNA copy number, energy output, and CLK-1 as a homeostatic copy number regulator.

3.7. Limitations.

Previous work measuring clk-1 mutants treated with DHB demonstrated that even when biological timing and stress phenotypes are rescued, ubiquinone is not rescued to wild-type levels [36]. These data suggest that wild-type levels of ubiquinone are not required for restoration of mitochondrial function. We note that it is possible that, unlike the other phenotypes, the mtDNA copy number phenotype could have a non-linear response to DHB supplementation. There could be a threshold effect for mtDNA copy number such that until a sufficient level of ubiquinone is reached, copy number will not decrease to wild-type levels. If a threshold effect exists, the mtDNA copy number phenotype must require more than 100-fold greater DHB supplementation than the biological timing and stress phenotypes, which show partial rescue even at 1mM DHB. Higher DHB concentrations (500mM and 1M) could not be tested since both concentrations cause lethality of both the bacterial food source and the worms.

3.8. Conclusions and Future Directions.

Taken together, our finding that the clk-1 mutant copy number phenotype is not caused by lack of ubiquinone, combined with the previous work describing CLK-1’s mtDNA binding capacity, suggest it is possible that CLK-1 possesses an additional functional role in the mitochondria related to mtDNA. If CLK-1 serves a role in mtDNA copy number regulation, then instances of copy number dysregulation may occur via CLK-1 dysfunction. Since improper mtDNA copy number is implicated in a variety of cancers, infertility, and embryonic arrest, the 30% increase in mtDNA copy number we report in clk-1 mutants may have consequences [7]. For example, excess mtDNA copy number may cause mitochondrial stress or activate signaling pathways. Additional experiments should be conducted to explore the impact of elevated copy number on lifespan and healthspan measures. Further, since mtDNA copy number changes across developmental time and aging, the status of CLK-1 should be investigated in contexts of copy number dysregulation and development. It will also be interesting to determine whether biochemical phenotypes, including reactive oxygen species (ROS) and oxidative phosphorylation, are associated with the elevation in copy number. We encourage further work to elucidate CLK-1’s role including mechanistic experiments exploring CLK-1-mtDNA interaction in vivo, as well as functional experiments investigating CLK-1 status in fertility, development, cancer, and aging.

4. METHODS

4.1. General methods and strains

Animals were grown at 20°C and raised on standard nematode growth media (NGM) plates. Worms were reared on E. coli strain OP50 (https://cgc.umn.edu/strain/OP50) in all experiments except the copy number experiment in which they were raised on E. coli strain HT115 (https://cgc.umn.edu/strain/HT115(DE3)), as noted. No antibiotics were used. The Bristol strain N2 was used as wild type. The mutations used in these experiments were: clk-1(ok1247) III, clk-1(qm30) III, and clk-1(e2519) III. The transgenic strain used in these experiments was the UPR^mt reporter strain SJ4100 zcIs13 V[hsp-6::GFP], which was obtained from the Caenorhabditis Genetics Center (CGC).

4.2. DHB supplementation

The effects of 2,4-dihydroxybenzoic acid (DHB) (Sigma W379808) on clk-1 mutant worms were measured as previously described [36]. For 1M stock, DHB was dissolved in 100% ethanol (Fisher Scientific CE500). For 1mM, 10mM, and 100mM working stocks, 10uL, 100uL and 1mL, respectively, were added to 100% ethanol to reach a final volume of 10mL. Five drops (approximately 250uL) of respective working stock were added to each pre-seeded 6cm plate with 8.7mL NGM using a 10mL serological pipette. Plates were stored at 4°C, protected from light, for up to one month. Unused stock solutions were discarded after one week. P0 L4 animals were plated on DHB-supplementation plates and F1 L4 offspring were used for subsequent phenotype measurement experiments. Treatment with DHB at higher doses (500mM and 1M working stock solution) was detrimental to the worms and was not pursued beyond pilot testing.

4.3. mtDNA copy number quantification via qPCR

Copy number was measured in larval stage four (L4) worms except the copy number experiment in which adult worms were used, as noted. Five larval stage four (L4) worms of each genotype were plated in duplicate. This process was repeated over three additional days to ensure synchronous availability of F1 L4s across genotypes. To measure mtDNA copy number, single F1 L4 animals were crudely lysed (see Crude worm lysis) to generate DNA template. To obtain synchronous adult worms for lysis, the same process was followed as above, and when animals reached L4 stage, they were moved to new plates. The day that animals reach L4 is considered Day 0 adulthood; one day after L4 is reach is considered Day 1 adulthood, and so on. At Day 2 adulthood, the adults were moved again, to prevent young worms from consuming all the bacteria. At Day 4 adulthood some worms were lysed, and others were moved to new plates to prevent young worms from consuming all the bacteria. At Days 5 and 6 worms were lysed. Multiplex qPCR reactions (see Multiplex qPCR details) were run in triplicate using primers and probes for one nuclear DNA target and two mitochondrial DNA targets. After obtaining C_T values, technical outliers were eliminated using a 20% threshold for variation between replicates of a sample. Finally, to account for differences in animal size, C_T values of both mitochondrial DNA targets were normalized to the C_T value of the nuclear DNA target (actin). Normalized values were then again normalized to the average of the normalized wild type values.

4.4. Development time (time to L4)

To measure development time, five P0 L4s were picked to a seeded plate. After two days, the adults were moved to a new seeded plate, allowed to lay embryos for five hours, and removed. At frequent intervals, F1 animals were removed and counted if they had reached L4 stage or left on the plate to continue developing. Worms were considered L4 when a stereotypical developing vulva was visible, including the dark concentric rings. To measure development time, 5 P0 L4s were picked to a seeded plate. After 2 days, the adults were moved to a new seeded plate, allowed to lay embryos for 5 hours, and removed. At frequent intervals, F1 animals were removed and counted if they had reached L4 stage or left on the plate to continue developing. The experiment was completed once all animals reached L4 stage. This was repeated twice for a total of three replicates per genotype-treatment combination.

4.5. Pharyngeal pumping rate

To measure the pharyngeal pump rate, the number of movements of the grinder in thirty seconds was recorded and this value was multiplied by two to obtain pumps per minute. A total of three measurements were made to calculate mean pumps per minute. Sample video was obtained on a Zeiss Axio Zoom V16 stereo zoom microscope at 112x magnification using an AxioCam MRm camera at 14.23 frames per second and playback is at 5 frames per second.

4.6. Defecation cycle length (pBoc)

To measure defecation length, the number of seconds elapsed between pBoc contractions was counted. A total of six measurements were made to calculate the mean time between pBocs. Sample video was obtained on a Zeiss Axio Zoom V16 stereo zoom microscope at 112x magnification using an AxioCam MRm camera at 14.23 frames per second and playback is at 5 frames per second.

4.7. UPR^mt quantification via fluorescence intensity

Activation of UPR^mt was measured by quantifying GFP fluorescence under the control of the promoter of heat-shock protein hsp-6 (Phsp6::gfp) [36, 49]. L4 worms were synchronized from a time-limited embryo laying and were used for measuring whole-worm fluorescence. Worms were mounted on 2% agarose pads. 10uL of 10mM levamisole were added to the agarose pad to immobilize the worms. Fluorescence images were obtained on a Zeiss Axio Zoom V16 stereo zoom microscope at 112x magnification using an AxioCam MRm camera. Fluorescence intensity was quantified using FIJI [50].

4.8. Crude worm lysis

After thawing lysis buffer (see Lysis buffer) aliquot(s), 1uL of 20mg/mL Proteinase K (Invitrogen 2106565) was added per 100uL lysis buffer to make LBMix. One worm was picked into 40uL of LBMix. Worms in LBMix were frozen overnight for a minimum of 12 hours. Lysis was done in a thermal cycler (see Thermal cycler conditions) and lysate was stored at −20°C until used. Samples were pelleted in a microfuge to remove debris before dilution. Samples were first diluted 1:10 in UltraPure Water (ThermoFisher 10977015), and this dilution was subsequently diluted 1:5 in UltraPure Water (ThermoFisher 10977015) for use in qPCR.

4.8.1. Lysis buffer:

50 mM KCL; 10 mM Tris pH 8.3; 2.5 mM MgCl2; 0.45% Tween 20; 0.45% NP-40 (IGEPAL); 0.01% Gelatin in UltraPure Water (ThermoFisher 10977015). Kept at −20°C in 1mL aliquots.

4.8.2. Thermal cycler conditions:

1. 60°C for 90m to lyse worms

2. 95°C for 15m to inactivate Proteinase K

3. 4°C briefly until collected from machine to maintain sample integrity

4.9. Multiplex qPCR details

Pipette tips should not be low-retention (Mettler-Toledo 30389225).

Plates are optical 384-well reaction plates with barcode (Thermo Fisher 4343814).

Plate covers are optically clear adhesive film (Thermo Fisher 4311971).

4.9.1. qPCR MasterMix:

TaqPath™ ProAmp™ Multiplex Master Mix (Thermo Scientific A30869).

4.9.2. Primers:

Concentrations of primers are 50uM each.

Target 1: Total mtDNA (forward) CK078 GCT CAT GTA GAG GCT CCT ACA AC

Target 1: Total mtDNA (reverse) CK079 GCTAAAGCCTTTGAATCTCTTTGAAAC

Target 2: Total mtDNA (forward) CK043 GGTAGAACATCTAGGTTATATTGCCACG

Target 2: Total mtDNA (reverse) CK044 GCATCACCTAAATTAAACGGGTAAATCA

Target 3: Actin nuclear DNA (forward) CK046 GGTGAAAGAGTAACCACGCTC

Target 3: Actin nuclear DNA (reverse) CK047 CCACACGCCATCCTCC

Target 4: Spike-in control (forward) CGT AGT GCG AAA CGT GGA G

Target 4: Spike-in control (reverse) GTGTGGACTGGGCATCG

4.9.3. Probes:

Concentrations of probes are 10uM each.

Target 1: Total probe CKP08 GGCACAGCGGGATTTTTACGTATTTTAGG (6FAM)

Target 2: Total probe CKP02 GCATCTTTACCTAAGTACTCAGGTCTAAAACAAAC (VIC)

Target 3: Actin probe CKP04 GGC TGG ACG TGA TCT TAC TGA TTA CCT C (JUN)

Target 4: Spike-in control probe GCAACCATGTAAGGGTACTAGCTGATAAC (ABY) All probes use QSY quencher.

This assay was optimized with an exogenous spike-in control and corresponding primers and probe (Target 4). However, over time, we found that the spike-in control was not providing additional sensitivity, so we discontinued use of the spike-in. Because the assay was optimized with primers and probe against Target 4, for continuity, we continued use of those oligonucleotides, even though there was no amplification of those targets. If used, these primers and probes should not result in any amplification.

4.9.4. qPCR reaction details:

Each reaction contains: 10uL TaqPath™ ProAmp™ Multiplex Master Mix (Thermo Scientific A30869), 1.52uL UltraPure Water (ThermoFisher 10977015), 0.36uL each primer, and 0.4uL each probe to total 16uL. After dispensing the reactions into the plate, 4uL of template is added to each well. qPCR Reactions were set up using an electronic multichannel pipette (Rainin 17013798) at the slowest speed to minimize between-well variability.

4.9.5. qPCR thermal cycler details:

qPCR machine used was an AppliedBiosystems QuantStudio 6 Flex machine.

4.9.6. qPCR thermal cycler settings:

384-well block. Comparative C_T (ΔΔC_T). TaqMan Reagents. Standard properties. MustangPurple passive reference dye.

4.9.7. qPCR thermal cycler conditions:

1. 60°C for 30s (1.6°C/s ramp rate) Data collection off.

2. 95°C for 5m (1.6°C/s ramp rate) Data collection off.

3. 95°C for 15s (1.6°C/s ramp rate) Data collection off.

4. 60°C for 1m (1.6°C/s ramp rate) Data collection on.

5. Steps 3 and 4 were repeated for 40 cycles.

4.10. PCR and agarose gel electrophoresis

4.10.1. PCR reactions:

Per each reaction: 36.5uL nuclease-free water, 5uL DreamTaq Buffer, 1uL dNTPs, 1uL forward primer, 1uL reverse primer, 0.5uL DreamTaq Polymerase, and 5uL template to total 50uL.

4.10.2. Primer sequences:

clk-1 forward primer: GCTGGCCCAGTACATTTGTT (from CGC: https://cgc.umn.edu/strain/RB1234)

clk-1 reverse primer: CAGTGTTCCGGATTTCAGGT (from CGC: https://cgc.umn.edu/strain/RB1234)

4.10.3. Thermal cycler conditions:

1. 95°C for 2m

2. 55°C for 30s

3. 60°C for 2m

4. Repeat steps 2 and 3 for 35 cycles

5. 60°C for 5m

4.10.4. Agarose gel electrophoresis:

No loading dye was needed as DreamTaq Buffer contains loading dye. GeneRuler 1kb Plus DNA Ladder was used to estimate size of PCR product (Thermo SM1333). 2% UltraPure agarose gel (Invitrogen 16–500-500).

4.10.5. Sequencing reactions:

Per each reaction: 8uL nuclease-free water 2.5uL forward OR reverse primer (see above for clk-1 primer sequences), and 4.5uL PCR product purified using MinElute kit (Qiagen 28004). Sanger sequencing was conducted by GeneWiz.

4.11. Western blot

4.11.1. Antibody generation:

clk-1 cDNA with an added 6-HIS tag was cloned into the pET-17b plasmid and further transfected into Rosetta ™ (DE3) competent cells for protein expression. CLK-1–6-HIS was expressed in pET-17b cells on a large scale and subsequently purified using Ni-NTA Agarose beads (Qiagen 30210). Purified CLK-1–6-HIS was then sent to CoCalico Biologicals for injection into rabbits. Pre-bleeds and test-bleeds were tested during antibody generation. Final production bleeds were stored, and aliquots were purified in batches to produce polyclonal anti-CLK-1 antibody. This in-house generated anti-CLK-1 antibody robustly recognizes CLK-1 and appears to have high specificity (Fig 1D + E, Fig S1A-D).

4.11.2. Western blot protocol:

Western analysis of unquantified boiled total worm lysates from N2, clk-1(ok1247), clk-1(qm30), and clk-1(e2419) samples were performed by electrophoresis on 4–20% Mini-PROTEAN gels (Bio-Rad 4561093), followed by transfer to 0.45 um low fluorescence PVDF membrane (Bio-Rad 1704274). Precision Plus Protein Dual Color Standard was used (Bio-Rad 1610374). The primary antibody to C. elegans CLK-1 was used at a 1:1000 dilution. Horseradish peroxidase-linked secondary antibodies (Jackson ImmunoResearch C839M39) to rabbit IgG were used in a 1:2000 dilution. Visualization was done using Clarity ECL substrate (Bio-Rad 1705060) and imaging was conducted on an Amersham™ Imager 600.

4.12. Statistical analyses

4.12.1. Regression analyses measuring DHB dose-response rescue of phenotypes:

A linear regression analysis was conducted to determine the relationship between phenotype and DHB supplementation concentration. For time to development analysis, a multiple linear regression was used. Phenotype was coded as the dependent variable, and DHB dose as a continuous independent variable. For UPR^mt, pharyngeal pump rate, defecation cycle time, and mtDNA copy number analyses, simple linear regression analyses were conducted, with phenotype coded as the dependent variable and DHB dose as the independent variable. The slope of the linear regression line was calculated. Slopes were compared to determine if slopes differed significantly from zero and whether wild-type and mutant slopes differed significantly from each other, for each phenotype.

4.12.2. One-way ANOVA comparisons:

A one-way ANOVA analysis was conducted for each phenotype at the lowest and highest DHB treatment concentrations (0mM and 100mM DHB working-stock solutions, respectively). Sidak’s multiple comparisons test was used to account for the repeated comparisons. Wild-type and mutant phenotypes were compared via one-way ANOVA at only the highest and lowest DHB concentrations since the linear regression analysis identifies the relationship between each phenotype and DHB dose. The one-way ANOVA is useful not for determining the relationship between these two variables, but for simply comparing the two DHB concentrations.

Highlights.

• An antibody for CLK-1 protein in C. elegans was developed.

• A novel clk-1 allele was established as a null.

• clk-1 mutants, which fail to synthesize ubiquinone suffer from biological timing defects, activate mitochondrial stress pathway, and exhibit an increase in mitochondrial DNA (mtDNA) copy number.

• DHB supplementation, which restores ubiquinone in clk-1 mutants, rescues biological timing and mitochondrial stress phenotypes in a dose-dependent manner, but does not restore mtDNA copy number to baseline levels.

• In addition to its role in ubiquinone synthesis, CLK-1 may serve an additional role in regulating mtDNA copy number in C. elegans.