Phenotypic and Suppressor Analysis of Defecation in clk-1 Mutants Reveals That Reaction to Changes in Temperature Is an Active Process in Caenorhabditis elegans

Corresponding author: Siegfried Hekimi, Department of Biology, McGill University, 1205 Dr. Penfield Ave., Montreal, Quebec H3A 1B1, Canada., shekim1{at}po-box.mcgill.ca (E-mail)

Communicating editor: B. J. MEYER

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Mutations in the Caenorhabditis elegans maternal-effect gene clk-1 affect cellular, developmental, and behavioral timing. They result in a slowing of the cell cycle, embryonic and postembryonic development, reproduction, and aging, as well as of the defecation, swimming, and pharyngeal pumping cycles. Here, we analyze the defecation behavior in clk-1 mutants, phenotypically and genetically. When wild-type worms are grown at 20° and shifted to a new temperature, the defecation cycle length is significantly affected by that new temperature. In contrast, we find that when clk-1 mutants are shifted, the defecation cycle length is unaffected by that new temperature. We carried out a screen for mutations that suppress the slow defecation phenotype at 20° and identified two distinct classes of genes, which we call dsc for defecation suppressor of clk-1. Mutations in one class also restore the ability to react normally to changes in temperature, while mutations in the other class do not. Together, these results suggest that clk-1 is necessary for readjusting the defecation cycle length in response to changes in temperature. On the other hand, in the absence of clk-1 activity, we observe temperature compensation, a mechanism that maintains a constant defecation period in the face of changes in temperature.

MUTATIONS in the Caenorhabditis elegans gene clk-1 are highly pleiotropic, affecting the rates of physiological traits that occur over a wide range of timescales (WONG et al. 1995 ). They result in a mean lengthening of the cell cycle of early embryos, embryonic and post-embryonic development, as well as the defecation, swimming, and pharyngeal pumping cycles of adults. clk-1 mutations also affect reproductive features, like the egg-production rate and self-brood size, which are both reduced, and lead to an increased life span.

A number of observations suggest that the phenotypes of clk-1 mutants are the result of an inability to appropriately set the rate of physiological processes (WONG et al. 1995 ; BRANICKY et al. 2000 ). One example is that many of the features affected by clk-1 mutations are more variable, in addition to being slower on average. For instance, although the average length of embryogenesis of clk-1 mutants is slower than that of the wild type, some clk-1 embryos can develop faster than wild-type embryos, while others take more than two times longer, which suggests that timing is deregulated in the mutants (WONG et al. 1995 ). Also, we observed that clk-1 mutant embryos are unable to properly adjust their rate of development in response to changes in temperature. Briefly, when wild-type embryos are cultured to the two-cell stage at a particular temperature and are then transferred to a new temperature, they immediately develop at a rate corresponding to that new temperature. In contrast, when clk-1 mutant embryos are transferred to a new temperature, the rate of development at the new temperature is strongly influenced by the temperature experienced before the shift (WONG et al. 1995 ). This suggests that clk-1 might also be needed to reset physiological rates in response to changes in temperature. Finally, all of the phenotypes affected in clk-1 mutants can be maternally rescued; that is, homozygous mutant progeny issued from a heterozygous hermaphrodite are phenotypically wild type. This maternal rescue extends to adulthood, such that all adult behaviors and even the long life of clk-1 mutants are rescued (HEKIMI et al. 1995 ; WONG et al. 1995 ). We interpreted this to mean that clk-1 affects a regulatory process that is somehow involved in setting physiological rates in the worm (WONG et al. 1995 ; FELKAI et al. 1999 ; BRANICKY et al. 2000 ). In the presence of maternally supplied clk-1 product, timing can be set appropriately early in development so that homozygous clk-1 mutants can subsequently develop and behave like the wild type.

clk-1 encodes a mitochondrial protein that is highly conserved, structurally and functionally, among eukaryotes (PROFT et al. 1995 ; JONASSEN et al. 1996 ; EWBANK et al. 1997 ; VAJO et al. 1999 ) and encodes a putative hydroxylase (STENMARK et al. 2001 ) that is required for the biosynthesis of ubiquinone [UQ; also called coenzyme Q (CoQ)], a prenylated benzoquinone lipid that functions as a transporter of electrons in complexes II and III of the respiratory chain. Mutants of the yeast homolog of clk-1, coq7, do not produce UQ and therefore cannot grow on nonfermentable carbon sources (MARBOIS and CLARKE 1996 ). Mitochondria isolated from clk-1 mutants also do not contain detectable levels of UQ but instead accumulate the UQ biosynthetic intermediate, demethoxyubiquinone (DMQ; MIYADERA et al. 2001 ). In clk-1 mutants this compound functions as an electron carrier, such that the mitochondria can maintain respiration despite the complete absence of UQ (FELKAI et al. 1999 ; MIYADERA et al. 2001 ). However, DMQ cannot substitute entirely for UQ as clk-1 mutants cannot complete development when they are fed Escherichia coli strains that do not produce UQ (JONASSEN et al. 2001 ).

It is, however, not clear how the absence of UQ relates to the other mutant phenotypes as there is no correlation between this biochemical phenotype and the severity of the overall phenotype. Indeed, the quinone phenotype is identical for all three known clk-1 alleles (e2519, qm30, and qm51): UQ is totally absent from mitochondria in all three cases, and all three accumulate the same amount of DMQ. Yet, most of the features affected in clk-1 mutants are slowed down much more severely in the putative null alleles qm30 and qm51 than they are in the partial loss-of-function allele e2519 (WONG et al. 1995 ; FELKAI et al. 1999 ). Also, by various measures of energy metabolism in intact worms, clk-1 mutants have been shown to have metabolic capacities and ATP levels comparable to the wild type (BRAECKMAN et al. 1999 ). Together, these observations suggest that much of the phenotype of clk-1 mutants may not be the direct consequence of an absence of UQ in mitochondria or a decreased level of energy production.

One of the features affected in clk-1 mutants is the defecation cycle. In C. elegans, defecation is effected by a stereotyped defecation motor program (DMP). The DMP consists of three distinct steps: the posterior body muscle contraction (pBoc), the anterior body muscle contraction (aBoc), and the expulsion (Exp), which consists of the enteric muscle contractions (EMC; THOMAS 1990 ). In the presence of adequate food, the defecation cycle period of 56 sec is regular in single animals over time and among animals, with a standard deviation of only a few seconds. In addition to its tight periodicity, the defecation cycle has other properties that suggest that it might be controlled by an endogenous "clock." For example, the phase of the cycle can be reset by lightly touching the animal, and the rhythm is maintained even in the absence of expression of the DMP (LIU and THOMAS 1994 ).

The periodicity of the defecation cycle can be altered by mutations in at least 13 genes (Dec phenotype). These mutations fall into two major classes: short Dec (Dec-s), for mutations that decrease the cycle length, and long Dec (Dec-L), for mutations that increase the cycle length (IWASAKI et al. 1995 ). The molecular identification of the Dec-L gene, dec-4 (lef-1/itr-1), as the inositol triphosphate (IP3) receptor, a protein involved in regulating intracellular calcium levels, suggests that calcium oscillations contribute to the regulation of the rhythm. Indeed, DAL SANTO et al. 1999 showed that calcium levels peak in the intestine just prior to the first muscle contraction of the DMP and that expression of the IP3 receptor in the intestine was sufficient for normal rhythm generation. Although it is not yet clear how the different Dec genes might be interacting to regulate the defecation cycle, the molecular characterization of two other genes, flr-1 and unc-43/dec-8, also supports roles for calcium and the intestine in rhythm regulation. flr-1 mutants, originally identified on the basis of their resistance to fluoride (KATSURA et al. 1994 ), have, among other defecation phenotypes, a very short defecation cycle length (IWASAKI et al. 1995 ). flr-1 encodes an ion channel of the degenerin/epithelial sodium channel superfamily, which is expressed only in the intestine from embryos to adults (TAKE-UCHI et al. 1998 ). Mutations in unc-43 result in multiple behavioral defects, including defecation phenotypes (LIU and THOMAS 1994 ; REINER et al. 1999 ). Loss-of-function mutations result in an increased frequency of defecation, usually the result of a repetition of the DMP 13 sec after the initiation of the primary motor program, whereas the gain-of-function mutation results in a decreased frequency of defecation. unc-43 encodes the C. elegans CaM Kinase II, which is widely expressed in neurons, muscles, and the intestine (REINER et al. 1999 ).

Here, we analyzed the defecation behavior of clk-1 mutants, with particular emphasis on the response of the mutants to changes in temperature. We find that when wild-type worms are grown at 20° and then shifted to a new temperature, the defecation cycle length is significantly affected by that new temperature. In contrast, when clk-1 mutants are shifted, the defecation cycle length is the same at the new temperature as it is at 20°. We reasoned that if clk-1 is actively involved in regulating the timing of the defecation cycle it might work through other gene products. We carried out a screen for suppressor mutations at 20° and identified 10 mutations, which correspond to seven different complementation groups. We find that the suppressor mutations fall into two distinct classes. In addition to suppressing the slow defecation at 20°, one class also suppresses the inability of clk-1 mutants to adjust their defecation cycle length after a temperature shift, while the other class does not. One general conclusion that can be drawn from this study is that adjusting the rate of defecation in response to changes in temperature is an active process in C. elegans, which requires clk-1.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

General methods and strains:
Most strains were derived from the wild-type C. elegans N2 Bristol strain and were cultured as described (BRENNER 1974 ). The wild-type RW7000 strain was used for some of the linkage analyses, using sequence-tagged sites (STS). All animals were grown at 20° unless otherwise indicated. The genes, alleles, and STSs used in this work are as follows:

• LGI: bli-4(e937), stP124;

• LGII: rol-6 (e187), rol-1(e91), unc-52(su250ts), maP1;

• LGIII: daf-2(e1368), dpy-17(e164), clk-1(qm30), unc-32(e189), dec-7 (sa296), vab-7 (e1562);

• LGIV: flr-3(ut9), dpy-9(e12), unc-33(e204), unc-5(e53), unc-31 (e928), dpy-4(e1166sd), sP4;

• LGV: unc-34(e315), dpy-11 (e224), stP192; and

• LGX: lin-15(n765), flr-4(ut7), unc-3(e151), lon-2(e678), unc-2 (e55), stP103.

Isolation of suppressor mutations:
clk-1(qm30) animals were mutagenized with 25 mM ethyl methanesulfonate (EMS), as described (SULSTON and HODGKIN 1988 ). Groups of five mutagenized hermaphrodites (P0) were plated on 60-mm petri dishes and left to self-fertilize. Groups of 25 F₁ animals were transferred to 90-mm plates as young adults and were left to lay eggs for 24 hr. F₂ animals were scored for one defecation cycle each at 20°. On the basis of the assumption that specific suppressor mutations would not cause morphological or other behavioral defects, only wild-type-looking animals were scored. A maximum of 50 F₂ animals were scored from each plate to minimize the probability of scoring multiple worms carrying the same mutation. Animals that had a defecation cycle length of <65 sec were picked to 60-mm plates, singled, and left to self-fertilize. The progeny (F₃) of the singled candidate worms were scored for defecation; only those strains that had a significant proportion of fast defecating worms in the F₃ generation were kept for further genetic and phenotypic analysis. A total of 5421 F₂ animals were screened, a number equivalent to 2134 haploid genomes (ELLIS and HORVITZ 1991 ).

Mapping of suppressor mutations:
Mutations were linked to chromosomes in one of three ways. For mutations that could be mapped on wild-type [clk-1(+)] backgrounds, linkage analysis was performed either by using the DA438 strain (for qm133, qm141; AVERY 1993 ) or by using the wild-type RW7000 strain and testing linkage to STSs by PCR analysis (for qm142, qm183) as described (WILLIAMS 1995 ). For mutations that had less obvious phenotypes on wild-type backgrounds or that could not be linked to an STS, linkage to each chromosome was tested separately, using strains that contained visible marker mutations in the clk-1(qm30) background (for qm178, qm179, qm180, qm182, and qm184). Formal linkage analysis was not needed for qm166, as it was found to be tightly linked to clk-1(qm30). Once linkage to a chromosome was established, mutations were mapped more precisely using two- and three-point mapping strategies.

To avoid marker effects, for most three-point mapping experiments, homozygous recombinant progeny were isolated from F₂ recombinant animals and were crossed with homozygous mutant males. Defecation was scored in the F₁ cross-progeny. Because qm133 has other phenotypes that could be scored in addition to defecation cycle length, including a thin, starved appearance, all mapping of qm133 was done using conventional methods (by singling recombinant F₂ animals and scoring for the presence of ¹/₄ qm133 animals in the F₃ generation). All mapping of qm133, qm141, and qm183 was done with strains in a clk-1(+) background and was scored at 20°; all mapping of qm179, qm180, qm182, and qm184 was done with strains in a clk-1(qm30) background and was scored at 25°. Mapping of qm142 was also done in a clk-1(qm30) background but was scored at 20°. As qm142 has a dominant effect, it was mapped by crossing homozygous recombinant progeny with clk-1(qm30) males and scoring the F₁ animals after 48 hr after they had reached adulthood, when the dominant effect of qm142 is strongest. The mapping data have been submitted to WormBase (http://www.wormbase.org/), and the critical data are summarized in Table 2.

Complementation tests:
Complementation tests were performed when different mutations appeared to map to the same genetic region or when a mutation mapped to a region where a Dec gene had previously been mapped. Generally, males homozygous for one mutation were mated to hermaphrodites homozygous for the other mutation, and defecation was scored in the trans-heterozygous F₁ animals. In this manner it was found that qm166 and qm178 fail to complement each other and the previously identified mutant dec-7(sa296), which are therefore all likely to be allelic. Similarly, qm179, qm184, and qm180 all fail to complement each other and are therefore all likely to be allelic. qm133 maps in the region of flr-4 but complements flr-4(ut7), suggesting that they define distinct genes. Similarly, qm182 maps in the region of flr-3 but complements flr-3(ut9), suggesting that they define distinct genes.

Behavioral analyses:
Defecation was scored in hermaphrodites on their first day of adulthood at 20°, unless otherwise indicated. The defecation cycle length was defined as the duration between the pBoc steps of two consecutive defecations. Each animal was scored for five consecutive cycles (six consecutive pBocs), and the mean and standard deviation were calculated. To prevent the animals from being heated by the microscope lamp during the scoring session, the plates were placed on "heat sinks" (petri dishes filled with water) and animals were scored from them for a maximum of only 15 min.

Temperature shift experiments:
Animals were grown at 20° and transferred to either 15° or 25° as young adults. Animals were then scored at 15° or 25° 2–6 hr after being transferred to that temperature. As it was difficult to maintain the temperature of the plates at 15°, animals were scored for only three consecutive cycles at this temperature, and plates were kept on the microscope for only as long as was required to score one animal.

qm142 time course studies:
To generate heterozygous dsc-2 (qm142)/+ animals, dsc-2(qm142) dpy-11(e224) animals were mated with N2 males. To generate qm30; qm142/+ animals, clk-1(qm30); unc-5(e53) hermaphrodites were mated with clk-1 (qm30); dsc-2(qm142) males. Late L4 stage F₁ generation animals were picked to plates and examined 3 hr later. Animals that had molted to adults during this period were used for the experiment and were considered to be 1.5-hr-old adults at the end of the interval. The sample size of each genotype for each time point is 10. The same sets of animals were scored at the different time points.

Statistical analyses:
We performed two-sample Student's t-tests, taking into account the unequal variances of the samples. We tested whether N2 and clk-1(qm30) mutants were different when grown and scored at 15° and 25° from when they were grown and scored at 20°. Highly significant differences (P < 0.05) were found for both N2 and clk-1(qm30) at 20° vs. at 15°. Significant differences were not detected for N2 at 20° vs. at 25° (P = 0.40) or for clk-1(qm30) at 20° vs. at 25° (P = 0.27). We also tested whether N2 and clk-1(qm30) mutants were different when grown at 20° and scored at 15° and 25° from when they were grown and scored at 20°. Highly significant differences (P < 0.001) were found in all comparisons except for clk-1(qm30) at 20° vs. at 15° (P = 0.29) and at 20° vs. at 25° (P = 0.99). We also tested whether each suppressor mutation had a significant effect on the defecation cycle by comparing, at every temperature, each clk-1(qm30) double mutant strain with clk-1(qm30) and every clk-1(+) mutant strain with clk-1(+). Highly significant differences (generally P < 0.001) were found for all comparisons except the following: clk-1(qm30) dec-7(qm166) vs. clk-1(qm30) at 15° (P = 0.34) and at 25° (P = 0.97); clk-1(qm30) dec-7(qm178) vs. clk-1(qm30) at 15° (P = 0.83) and at 25° (P = 0.15); clk-1(qm30); dsc-3(qm184) vs. clk-1(qm30) at 15° (P = 0.80); and clk-1(qm30); dec(qm183) vs. clk-1(qm30) at 20° (P = 0.12) and at 25° (P = 0.31).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

clk-1 mutants have a lengthened defecation cycle that is not affected by changes in temperature:
Defecation in C. elegans is achieved by the periodic activation of a stereotyped motor program. In wild-type animals, the defecation cycle length is 56 sec, with a standard deviation of only 3.4 sec (at 20°). As previously described (WONG et al. 1995 ; FELKAI et al. 1999 ), in clk-1 mutants the defecation cycle is both increased in length and more irregular: In clk-1(qm30) animals, the cycle length is 88 sec, with a standard deviation of 14 sec, and in the weaker allele clk-1(e2519), it is 77 sec, with a standard deviation of 7 sec (at 20°; Fig 1, Table 1).

View larger version (40K): In this window In a new window Download PPT slide   Figure 1. The effects of temperature on the defecation cycle length of the wild-type (N2) and clk-1 mutants. The bars represent the means of animals that had each been scored for five consecutive defecation cycles at 20° and 25° and for three consecutive cycles at 15°. The error bars represent the standard deviations of the means. The sample sizes are as follows: for 20°, n 50; for 15° and 25°, n 25. (A) Animals were raised at 15°, 20°, and 25° and were scored at the temperature at which they had been raised. The numerical values are as follows: for N2 at 15°, 20°, and 25°, respectively, 78.0 ± 6.7, 55.9 ± 3.4, and 51.2 ± 3.4; for clk-1(qm30) at 15°, 20°, and 25°, respectively, 131.9 ± 24.0, 88.4 ± 13.5, and 94.8 ± 14.3. (B) Animals were raised at 20° and scored at 15°, 20°, and 25°. The numerical values can be found in Table 1.

To examine the effect of temperature on the defecation cycle length of wild-type and clk-1 mutant worms, we raised the worms for two generations at 15° or 25° and scored them at the temperature at which they had been raised (Fig 1A). We found that clk-1 mutants are slower than the wild type at all temperatures and that both genotypes have significantly longer cycles when grown and scored at 15° than when grown and scored at 20°. In contrast, when wild-type or mutant worms are grown and scored at 25° the defecation rates are not significantly different from those at 20°. These results agree with, and extend, those of IWASAKI et al. 1995 .

To examine the effect of temperature shifts on the defecation cycle length of wild-type and clk-1 mutant worms, we raised the worms at 20°, shifted them to 15° and to 25°, and scored them at that new temperature (Fig 1B). We found that when wild-type animals are shifted to either 15° or 25°, the defecation cycle length is significantly affected by the change in temperature. When the worms are transferred from 20° to 25°, the mean cycle length is decreased by 13 sec; when the worms are transferred from 20° to 15°, the mean cycle length is increased by 26 sec. This adjustment happens very rapidly. For example, we found that when wild-type worms are scored 5 min after they are transferred to 25°, they are already as fast as they are when scored 2–6 hr later (data not shown). Note that the defecation cycle of wild-type worms raised and scored at 25° (Fig 1A) is in fact slower than when the worms are raised at 20° and then shifted to 25° for scoring (Fig 1B). This suggests that the worms have become adapted or acclimated to the higher temperature during development and are thus less affected when scored at that temperature. However, whether the worms are raised at 15° or 20° does not appear to affect the length of the defecation cycle when it is scored at 15°.

In contrast to the observations with the wild type, we found that clk-1 mutants are unable to readjust the length of their defecation cycle after they are shifted to a new temperature. When clk-1 mutants are transferred either from 20° to 25° or from 20° to 15° there is no change in the mean cycle length (Fig 1B, Table 1). This suggests that after the mutants have become adapted to 20°, they cannot readjust the defecation cycle length upon a temperature shift, indicating that clk-1(+) activity is required for this adjustment to occur.

A screen for suppressors of clk-1:
To identify genes that interact with clk-1 to regulate the defecation cycle length, we carried out a screen to isolate mutations that could suppress clk-1, that is, mutations that could restore the length of the defecation cycle of clk-1 mutants to that of the wild type. clk-1(qm30) worms were mutagenized with EMS and second-generation (F₂) animals were directly scored for one defecation cycle each at 20°. Animals that had a cycle length of <65 sec were kept for further analysis. In this manner, we screened 5421 F₂ animals (an equivalent of 2134 haploid genomes) and identified eight suppressor mutations, which correspond to five different complementation groups. Seven of these mutations are recessive and one mutation, qm142, has dominant effects (described in detail below). On the basis of the mapping and complementation tests we performed (see MATERIALS AND METHODS and Table 2), it is likely that these mutations define four new complementation groups, which we called dsc for defecation suppressor of clk-1. Two mutations, qm166 and qm178, are alleles of the previously identified Dec-s gene dec-7. From our screen, we also isolated two other recessive mutations, qm141 and qm183, which affect defecation, but which do not suppress clk-1(qm30) (discussed below). None of the mutations can suppress other clk-1 phenotypes, such as slow growth or pharyngeal pumping (data not shown).

Analysis of suppressor mutants:
We analyzed the phenotype of these suppressor mutants in a number of different ways (Table 1). We reisolated all of the mutations on a wild-type [clk-1(+)] background to determine how the mutations affect defecation in the presence of clk-1(+) activity. To see if the suppressors could also suppress the inability of clk-1 mutants to react to changes in temperature, we scored every clk-1(+) and clk-1(qm30) strain at 20° as well as after shifts to 15° and 25°. We also reisolated the mutations on the background of clk-1 (e2519), which has partial clk-1 activity.

There are two classes of suppressors:
By analyzing the effect of the mutations on the clk-1(qm30) background, we find that the suppressors fall into two distinct classes on the basis of their differential abilities to suppress clk-1(qm30) after temperature shifts, particularly to 25° (Fig 2). The class I mutants dsc-3(qm179, qm180, qm184) and dsc-4(qm182) strongly suppress clk-1(qm30) at 20° as well as after shifts to 25° (Fig 2A and Fig 3A). At 25°, dsc-3(qm179), the strongest mutant in this respect, can even shorten the defecation cycle length of clk-1(qm30) mutants to less than the wild-type length. Thus, the profile of defecation in the class I dsc; clk-1(qm30) double mutants at the different temperatures is just like that of the wild type: slowest at 15°, fastest at 25°. Therefore, these mutants can suppress the long defecation cycle of clk-1(qm30) mutants when grown and scored at 20° as well as the temperature insensitivity when grown at 20° and then shifted to another temperature.

View larger version (39K): In this window In a new window Download PPT slide   Figure 2. There are two classes of clk-1 suppressor mutants. Animals were raised at 20° and scored at 15°, 20°, and 25°. The bars represent the means of animals that had each been scored for five consecutive defecation cycles at 20° and 25° and for three consecutive cycles at 15°. The error bars represent the standard deviations of the means. The sample sizes are as follows: at 20°, n 50; at 15° and 25°, n 25. (A) Suppressor mutations on the clk-1(qm30) background. The class I mutations suppress the increased cycle length of clk-1 (qm30) mutants and restore the ability to react to changes in temperature. The class II mutations suppress the increased cycle length at 20° but do not restore the ability to react to changes in temperature. (B) Suppressor mutations on the wild-type background. The class I mutants generally have weaker effects than the class II mutants, particularly at 15° and 20°, but both classes of mutants react to changes in temperature like the wild type. The numerical values can be found in Table 1.

View larger version (25K): In this window In a new window Download PPT slide   Figure 3. Examples of class I and class II suppressors. Each point represents the mean (on the ordinate) and standard deviation (on the abscissa) of one animal, scored for five consecutive defecation cycles. (A) dsc-3(qm179) is an example of a class I mutant: It suppresses clk-1(qm30) at 20° and after a shift to 25°, such that in the double mutant, defecation is faster at 25° than at 20°. (B) dec-7(qm166) is an example of a class II mutant: It suppresses clk-1(qm30) at 20° but cannot suppress after a shift to 25°, such that the defecation rate of the double mutant at 25° is not faster than the defecation rate of clk-1(qm30).

In contrast to class I mutants, the class II mutants suppress only weakly [dsc-1(qm133) and dsc-2(qm142)] or not at all [dec-7(qm166, qm178)] after a temperature shift. This suggests that although these mutants can suppress the long defecation cycle of clk-1(qm30) mutants at 20°, they cannot suppress the inability of clk-1(qm30) mutants to readjust their defecation cycle length after a temperature shift. dec-7, however, might be a special case, as the mutants' slow mean defecation cycle after shifts to 15° or 25° could be due to an inability to carry out the DMP rather than to intrinsically slow cycling (see below).

The phenotypes of class I and class II mutants are much more similar on the wild-type background than on the clk-1(qm30) background, although the class I mutants generally have weaker effects on the wild type than the class II mutants, particularly at 15° and 20° (Fig 2B). All the mutations significantly decrease the length of the defecation cycle at all temperatures and both the class I and class II mutants react to changes in temperature like the wild type (i.e., defecation is fastest at 25°, slowest at 15°). This indicates that the lack of effect of the class II mutants at 25° on the clk-1(qm30) background cannot be due to any of the dsc mutations being intrinsically temperature sensitive. Overall, the observation that on the clk-1(qm30) background there are two distinct classes of interactions suggests that the genes of the different classes interact with clk-1 in different ways.

The effect of the dsc mutations on the clk-1(e2519) background:
We reisolated the dsc mutations on the background of clk-1(e2519), which has a weaker phenotype than clk-1(qm30), to see if they could speed up clk-1 (e2519) more than clk-1(qm30). Although we did consistently find that the defecation cycle lengths of the clk-1 (e2519); dsc double mutants were shorter than those of the clk-1(qm30); dsc double mutants, the difference was generally very small [ Table 1; see below for a description of the special case of dsc-5(qm141)]. The exact significance of this is unclear at present; however, it could suggest that clk-1 also has effects on the defecation cycle that do not depend on the dsc genes. Possibly, part of the residual activity of e2519, which allows the mutants to have a faster defecation cycle than qm30 mutants, acts in such a dsc gene-independent fashion.

Mutations that cannot suppress clk-1(qm30):
While screening for suppressors, we isolated a number of candidate strains from which we could not subsequently isolate mutations that strongly suppress clk-1(qm30). In two cases, however, the candidate strains did segregate mutations that significantly affect the rate of defecation. One of these mutations, dsc-5(qm141), has a strong effect on the defecation cycle of wild-type worms at all temperatures, as well as on clk-1(e2519) mutants, but only a very weak, but significant, effect on clk-1(qm30) mutants at all temperatures. This suggests that dsc-5(qm141) might be acting largely through clk-1, as it affects the null mutant (qm30) only slightly, whereas it fully suppresses the partial loss-of-function mutation (e2519). The other mutation we isolated, qm183, also has significant effects on the defecation cycle of wild-type worms at all temperatures, but it cannot suppress either clk-1(e2519) or clk-1 (qm30) mutants. It is clear, therefore, that the slow defecation of clk-1(qm30) mutants cannot simply be suppressed by every mutation that decreases the defecation cycle length of the wild type. In fact, some mutations may require full (i.e., qm183) or at least partial (i.e., qm141) clk-1(+) activity to affect the defecation cycle.

The dsc-2(qm142) mutation has a semidominant time-dependent effect:
We found that the strength of the qm142 mutant phenotype changes with the age of the animal, in both heterozygotes and homozygotes, albeit at very different rates (Fig 4). We performed a time course study in which we scored defecation in the same animals at different time points after they had molted to adults. In a clk-1(qm30) background, the homozygous dsc-2(qm142) animals are almost as slow as the clk-1(qm30) animals 2 hr after molting to adults, but by 8 hr, the defecation cycle length is restored to the wild-type length, and by 18 hr, the defecation cycle length is significantly shorter than that of the wild type. The defecation cycle length of clk-1(qm30); dsc-2(qm142)/+ animals is very similar to that of clk-1(qm30) animals until about 40 hr. In fact, it takes 48 hr after molting to adults for the defecation cycle length of the dsc-2(qm142)/+ heterozygous animals to become as fast as the age-matched wild-type animals. As the dsc-2(qm142)/+ heterozygotes never become as fast as the homozygotes at any time point and take longer than the homozygotes to speed up significantly, the effect of the qm142 mutation is incompletely dominant over the wild-type allele. This is also confirmed by observations of qm142 heterozygotes and homozygotes on the wild-type background (data not shown), although the effects are much less dramatic. One way in which the dsc-2(qm142) allele could have this semidominant time-dependent effect is that the mutation results in a protein that can interfere with the function of the wild-type DSC-2 protein. An accumulation of the mutant product with time could increase the severity of the mutant phenotype.

View larger version (19K): In this window In a new window Download PPT slide   Figure 4. The severity of phenotype of dsc-2(qm142) mutants is semidominant and time-dependent. A time-course study of mutants heterozygous and homozygous for the dsc-2 (qm142) mutation in the clk-1(qm30) background is shown. Animals were scored at different time points after having molted to adults (at time 0 hr). Each point represents the mean of 10 animals, scored for three consecutive defecation cycles at 20°; the error bars represent the standard errors of the means. clk-1 designates the clk-1(qm30) allele; dsc-2, the dsc-2 (qm142) allele; and +, the dsc-2(+) allele.

Mutations in dec-7 result in multiple discrete defecation cycle lengths:
We characterized all the suppressor mutants by analyzing the mean defecation cycle length of a number of animals that had each been scored for five defecation cycles (Table 1). We also calculated the standard deviations of individual animals. We noted that animals carrying dec-7 mutations in a clk-1(qm30) background have very high standard deviations at 15° and 20° but not at 25° [see Fig 3B for the example of dec-7(qm166) at 20° and 25°]. To analyze this variability further we plotted the frequency of single defecation cycle lengths of clk-1 dec-7 animals at four different temperatures (Fig 5). At all temperatures, there is only one frequency peak for clk-1(qm30) mutants, but there are two peaks for both clk-1 dec-7 double mutant strains at the three temperatures <25° (Fig 5, A–C). One of the peaks occurs at a cycle length that is two times that of where the other peak occurs. At 25°, however, there is only one peak, which coincides with the clk-1 (qm30) peak (Fig 5D).

View larger version (21K): In this window In a new window Download PPT slide   Figure 5. Mutations in dec-7 result in multiple discrete defecation cycle lengths. An analysis of the frequency of single defecation cycle lengths of dec-7 mutants in the clk-1(qm30) background is shown. The sample sizes for all genotypes at 15°, 20°, and 25° are n 25; at 22.5°, n = 10–25. At all temperatures, there is only one frequency peak for clk-1(qm30) mutants. At 15° (A), 20° (B), and 22.5° (C), there are two frequency peaks for both clk-1(qm30) dec-7(qm166) and clk-1 (qm30) dec-7(qm178) double mutants, the second peak occurring at a cycle length double that of the first peak. At 25° (D), there is only one frequency peak for the clk-1(qm30) dec-7 double mutants, which coincides with the clk-1(qm30) peak.

One interpretation of this pattern is that clk-1 and/or dec-7 have a role in coupling the activation of the DMP to the cycle, such that the coupling increasingly fails in clk-1 dec-7 double mutants with increasing temperature. This would result in double cycle lengths and could mean that at 25° every cycle observed is actually a double cycle. Multiple discrete cycle lengths are not observed in dec-7 mutants on a clk-1(+) background or in other class II mutants, and thus this phenomenon appears to be specific to clk-1 dec-7 mutants.

Another interpretation is that, with increasing temperature, there is a decrease in the penetrance of the suppression of clk-1 by dec-7. This is suggested by the observation that at 20°, 22.5°, and 25° the main peak coincides with the unsuppressed clk-1 peak (Fig 5). However, this does not happen at 15°, which is difficult to explain with this interpretation.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Mutations in clk-1 affect numerous features of the worm including the rates of rhythmic behaviors, as well as growth and reproductive features (WONG et al. 1995 ; BRANICKY et al. 2000 ). Results from earlier experiments, which looked at the effect of temperature shifts on the growth rate of embryos, suggested that clk-1 mutants are impaired in their ability to sense or react to changes in temperature (WONG et al. 1995 ). Here we have focused on the defecation cycle to further explore this phenomenon.

Reaction of defecation to changes in temperature:
When we examined the reaction of wild-type animals to changes in temperature we found that when they are raised at 20° and are shifted to either 15° or 25° the defecation cycle length is profoundly altered by the shift. When shifted to 15°, the defecation cycle is significantly lengthened, and when shifted to 25°, the defecation cycle is significantly shortened. The specifics of assay conditions or temperature-shift protocol may affect these results: One previous report found little change in cycle length with temperature (LIU and THOMAS 1994 ) but results similar to those reported here have also been obtained recently by K. IWASAKI and J. THOMAS (personal communication). Our findings do, however, indicate that there is a temperature compensation mechanism. Indeed, we have found that, in contrast to the wild type, clk-1 mutants do not adjust the length of their defecation cycle when they are shifted to a new temperature. Because worms are poikilothermic animals and the rate of the biochemical reactions that underlie physiological processes must be temperature dependent, this implies that there is a temperature compensation mechanism that can maintain the cycle length even at different temperatures. In summary, our findings suggest that the reaction to temperature we observe in the wild type is an active process that requires clk-1 activity. In the absence of clk-1 activity, we can observe a temperature compensation mechanism, which presumably also acts in the wild type, such that the wild-type condition appears to be a combination of temperature compensation and active adjustment.

Suppressors of the defecation phenotypes of clk-1 mutants:
We reasoned that if clk-1 is involved in regulating the timing of the defecation cycle it might work through other gene products. To identify such genes we carried out a screen for mutations that could suppress clk-1(qm30) at 20°. We identified eight suppressor mutations, which fall into five different complementation groups. We found that these mutations also significantly decrease the defecation cycle length of the wild-type and clk-1(e2519) mutants at 20°. If we consider the degree to which the mutations decrease the cycle length in the different backgrounds, (i.e., the percentage by which the mutation decreases the cycle length in each background), all of the dsc mutations [except dsc-5(qm141) and dec(qm183)] have the greatest effect on the clk-1 (qm30) background and the weakest effect on the wild-type background (Table 1). In some cases, this is very striking. For example, the dsc-3 mutants decrease the rate of defecation 25% on the clk-1(qm30) background but only 5% on the wild-type background. This effect is so small that these mutations could probably not have been found in a wild-type background. On the other hand, only the dec-7(qm166) mutation affects the wild type to the same degree as it affects clk-1. It is therefore not surprising that dec-7 was the only gene we identified that was also found in a previous screen aimed at finding mutations that affect the defecation cycle length in a wild-type background (IWASAKI et al. 1995 ).

When we examined how defecation in the dsc or dec; clk-1(qm30) double mutants is affected by changes in temperature, we found that the suppressors fall into two distinct classes. Class I mutants completely suppress the defecation phenotypes of clk-1 mutants, such that they suppress clk-1 at 20° and after a switch to a new temperature. Double mutants carrying class I mutations have a profile of defecation very much like that of the wild type with respect to temperature: shorter after a shift to 25° and longer after a shift to 15°. In contrast, although the class II mutants suppress as well as the class I mutants when grown and scored at 20°, they cannot overcome the inability of clk-1 mutants to react to changes in temperature. After a switch to 15° or 25° these mutations cannot, or only very poorly, suppress the slow defecation cycle. The existence of the class II mutants therefore indicates that the slow defecation and the inability to adjust the cycle length in response to changes in temperature are separable phenotypes of clk-1 mutants.

Although it is not yet clear how the class I mutants restore the reaction to temperature, one interesting possibility is that these genes are actually components of the temperature compensation mechanism. These mutations might therefore be acting by abolishing the mechanism of temperature compensation observed in the absence of clk-1. In this model, the reaction to temperature observed in the clk-1; dsc double mutants might be a more passive (thermodynamic) reaction than the combination of compensation and active adjustment we believe to be the wild-type condition. This would imply that the combined mechanisms of temperature compensation and active adjustment increase the defecation cycle length from what it would be if it was determined passively by temperature. A mechanism that prevents the organism from functioning permanently at the highest possible rate at a given temperature might be adaptive.

Toward understanding the pleiotropic clk-1 phenotype:
Although clk-1 affects ubiquinone biosynthesis (JONASSEN et al. 2001 ; MIYADERA et al. 2001 ; STENMARK et al. 2001 ), it is not clear how this is linked to the other phenotypes of clk-1 mutants. Indeed, clk-1 mutant mitochondria can function properly using DMQ, the biosynthetic intermediate they accumulate, and the severity of the ubiquinone phenotype is not correlated with the severity of the organismal phenotypes. The results presented here also support the notion that the phenotype of clk-1 mutants, specifically the slow defecation phenotype, cannot simply be explained by a lack of ubiquinone. Indeed it is unlikely that the suppressors, which do not affect any other phenotype, act by overcoming the ubiquinone deficiency in only those cells that are responsible for the slow defecation cycle of the clk-1 mutants. One possibility is that clk-1 functions in other processes in addition to the biosynthesis of ubiquinone. Alternatively, it could be that although DMQ can partially replace UQ in the respiratory chain, it is unable to replace UQ for some of its other functions, whose resulting impairment is responsible for some aspects of the clk-1 phenotype, such as the slow and temperature-invariant defecation cycle we described here. UQ is found in almost all biological membranes (DALLNER and SINDELAR 2000 ) and is known to be a cofactor of the uncoupling proteins (UCP) in the mitochondria (ECHTAY et al. 2000 , ECHTAY et al. 2001 ), to regulate the permeability transition pore (FONTAINE et al. 1998 ), and to function in plasma membrane and lysosomal oxido-reductase systems (SANTOS-OCANA et al. 1998 ; GILLE and NOHL 2000 ). Identifying the molecular nature of the dsc genes and their cellular and intracellular localization should thus help us understand the basis of the clk-1 phenotypes and the regulation of the defecation cycle in general.

	FOOTNOTES

¹ These authors contributed equally to this work.

	ACKNOWLEDGMENTS

We gratefully acknowledge Claire Bénard and Jason Burgess for critically reading the manuscript and Ambereen Hussain for helping to isolate mutants. Some nematode strains used in this work were provided by the Caenorhabditis Genetics Centre, which is funded by the National Institutes of Health National Center for Research Resources. S.H. is a Canadian Institutes of Health research scientist and R.B. is supported by a McGill Faculty of Graduate Studies Fellowship.

Manuscript received June 5, 2001; Accepted for publication August 7, 2001.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AVERY, L., 1993  The genetics of feeding in Caenorhabditis elegans. Genetics 133:897-917[Abstract].

BRAECKMAN, B. P., K. HOUTHOOFD, A. DE VREESE, and J. R. VANFLETEREN, 1999  Apparent uncoupling of energy production and consumption in long-lived Clk mutants of Caenorhabditis elegans. Curr. Biol. 9:493-496[Medline].

BRANICKY, R., C. BENARD, and S. HEKIMI, 2000  clk-1, mitochondria, and physiological rates. Bioessays 22:48-56[Medline].

BRENNER, S., 1974  The genetics of Caenorhabditis elegans. Genetics 77:71-94[Abstract/Free Full Text].

DAL SANTO, P., M. A. LOGAN, A. D. CHISHOLM, and E. M. JORGENSEN, 1999  The inositol trisphosphate receptor regulates a 50-second behavioral rhythm in C. elegans. Cell 98:757-767[Medline].

DALLNER, G. and P. J. SINDELAR, 2000  Regulation of ubiquinone metabolism. Free Radic. Biol. Med. 29:285-294[Medline].

ECHTAY, K. S., E. WINKLER, and M. KLINGENBERG, 2000  Coenzyme Q is an obligatory cofactor for uncoupling protein function. Nature 408:609-613[Medline].

ECHTAY, K. S., E. WINKLER, K. FRISCHMUTH, and M. KLINGENBERG, 2001  Uncoupling proteins 2 and 3 are highly active H(+) transporters and highly nucleotide sensitive when activated by coenzyme Q (ubiquinone). Proc. Natl. Acad. Sci. USA 98:1416-1421[Abstract/Free Full Text].

ELLIS, R. E. and H. R. HORVITZ, 1991  Two C. elegans genes control the programmed deaths of specific cells in the pharynx. Development 112:591-603[Abstract].

EWBANK, J. J., T. M. BARNES, B. LAKOWSKI, M. LUSSIER, and H. BUSSEY et al., 1997  Structural and functional conservation of the Caenorhabditis elegans timing gene clk-1. Science 275:980-983[Abstract/Free Full Text].

FELKAI, S., J. J. EWBANK, J. LEMIEUX, J. C. LABBE, and G. G. BROWN et al., 1999  CLK-1 controls respiration, behavior and aging in the nematode Caenorhabditis elegans. EMBO J. 18:1783-1792[Medline].

FONTAINE, E., F. ICHAS, and P. BERNARDI, 1998  A ubiquinone-binding site regulates the mitochondrial permeability transition pore. J. Biol. Chem. 273:25734-25740[Abstract/Free Full Text].

GILLE, L. and H. NOHL, 2000  The existence of a lysosomal redox chain and the role of ubiquinone. Arch. Biochem. Biophys. 375:347-354[Medline].

HEKIMI, S., P. BOUTIS, and B. LAKOWSKI, 1995  Viable maternal-effect mutations that affect the development of the nematode Caenorhabditis elegans. Genetics 141:1351-1364[Abstract].

IWASAKI, K., D. W. LIU, and J. H. THOMAS, 1995  Genes that control a temperature-compensated ultradian clock in Caenorhabditis elegans. Proc. Natl. Acad. Sci. USA 92:10317-10321[Abstract/Free Full Text].

JONASSEN, T., B. N. MARBOIS, L. KIM, A. CHIN, and Y. R. XIA et al., 1996  Isolation and sequencing of the rat Coq7 gene and the mapping of mouse Coq7 to chromosome 7. Arch. Biochem. Biophys. 330:285-289[Medline].

JONASSEN, T., P. L. LARSEN, and C. F. CLARKE, 2001  A dietary source of coenzyme Q is essential for growth of long-lived Caenorhabditis elegans clk-1 mutants. Proc. Natl. Acad. Sci. USA 98:421-426[Abstract/Free Full Text].

KATSURA, I., K. KONDO, T. AMANO, T. ISHIHARA, and M. KAWAKAMI, 1994  Isolation, characterization and epistasis of fluoride-resistant mutants of Caenorhabditis elegans. Genetics 136:145-154[Abstract].

LIU, D. W. and J. H. THOMAS, 1994  Regulation of a periodic motor program in C. elegans. J. Neurosci. 14:1953-1962[Abstract].

MARBOIS, B. N. and C. F. CLARKE, 1996  The COQ7 gene encodes a protein in Saccharomyces cerevisiae necessary for ubiquinone biosynthesis. J. Biol. Chem. 271:2995-3004[Abstract/Free Full Text].

MIYADERA, H., H. AMINO, A. HIRAISHI, H. TAKA, and K. MURAYAMA et al., 2001  Altered quinone biosynthesis in the long-lived clk-1 mutants of Caenorhabditis elegans. J. Biol. Chem. 276:7713-7716[Abstract/Free Full Text].

PROFT, M., P. KOTTER, D. HEDGES, N. BOJUNGA, and K. D. ENTIAN, 1995  CAT5, a new gene necessary for derepression of gluconeogenic enzymes in Saccharomyces cerevisiae. EMBO J. 14:6116-6126[Medline].

REINER, D. J., E. M. NEWTON, H. TIAN, and J. H. THOMAS, 1999  Diverse behavioural defects caused by mutations in Caenorhabditis elegans unc-43 CaM kinase II. Nature 402:199-203[Medline].

SANTOS-OCANA, C., J. M. VILLALBA, F. CORDOBA, S. PADILLA, and F. L. CRANE et al., 1998  Genetic evidence for coenzyme Q requirement in plasma membrane electron transport. J. Bioenerg. Biomembr. 30:465-475[Medline].

STENMARK, P., J. GRUNLER, J. MATTSSON, P. J. SINDELAR, and P. NORDLUND et al., 2001  A new member of the family of di-iron carboxylate proteins: Coq7 (clk- 1), a membrane-bound hydroxylase involved in ubiquinone biosynthesis. J. Biol. Chem. 2:2.

SULSTON, J. E., and J. HODGKIN, 1988 Methods, pp. 587–606 in The Nematode Caenorhabditis elegans, edited by W. B. WOOD. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

TAKE-UCHI, M., M. KAWAKAMI, T. ISHIHARA, T. AMANO, and K. KONDO et al., 1998  An ion channel of the degenerin/epithelial sodium channel superfamily controls the defecation rhythm in Caenorhabditis elegans. Proc. Natl. Acad. Sci. USA 95:11775-11780[Abstract/Free Full Text].

THOMAS, J. H., 1990  Genetic analysis of defecation in Caenorhabditis elegans. Genetics 124:855-872[Abstract].

VAJO, Z., L. M. KING, T. JONASSEN, D. J. WILKIN, and N. HO et al., 1999  Conservation of the Caenorhabditis elegans timing gene clk-1 from yeast to human: a gene required for ubiquinone biosynthesis with potential implications for aging. Mamm. Genome 10:1000-1004[Medline].

WILLIAMS, B. D., 1995 Genetic mapping with polymorphic sequence-tagged sites, pp. 81–95 in Caenorhabditis elegans: Modern Biological Analysis of an Organism, edited by H. F. EPSTEIN and D. C. SHAKES. Academic Press, San Diego.

WONG, A., P. BOUTIS, and S. HEKIMI, 1995  Mutations in the clk-1 gene of Caenorhabditis elegans affect developmental and behavioral timing. Genetics 139:1247-1259[Abstract].

This article has been cited by other articles:

				A. K. Hihi, M.-C. Beauchamp, R. Branicky, A. Desjardins, I. Casanova, M.-P. Guimond, M. Carroll, M. Ethier, I. Kianicka, K. McBride, et al. Evolutionary conservation of drug action on lipoprotein metabolism-related targets J. Lipid Res., January 1, 2008; 49(1): 74 - 83. [Abstract] [Full Text] [PDF]

				K. Nehrke, J. Denton, and W. Mowrey Intestinal Ca2+ wave dynamics in freely moving C. elegans coordinate execution of a rhythmic motor program Am J Physiol Cell Physiol, January 1, 2008; 294(1): C333 - C344. [Abstract] [Full Text] [PDF]

				R. Branicky, P. A. T. Nguyen, and S. Hekimi Uncoupling the Pleiotropic Phenotypes of clk-1 with tRNA Missense Suppressors in Caenorhabditis elegans Mol. Cell. Biol., May 15, 2006; 26(10): 3976 - 3985. [Abstract] [Full Text] [PDF]

				R. Branicky and S. Hekimi Specification of muscle neurotransmitter sensitivity by a Paired-like homeodomain protein in Caenorhabditis elegans Development, November 15, 2005; 132(22): 4999 - 5009. [Abstract] [Full Text] [PDF]

				Y. Shibata, R. Branicky, I. O. Landaverde, and S. Hekimi Redox Regulation of Germline and Vulval Development in Caenorhabditis elegans Science, December 5, 2003; 302(5651): 1779 - 1782. [Abstract] [Full Text] [PDF]

				J. Burgess, A. K. Hihi, C. Y. Benard, R. Branicky, and S. Hekimi Molecular Mechanism of Maternal Rescue in the clk-1 Mutants of Caenorhabditis elegans J. Biol. Chem., December 5, 2003; 278(49): 49555 - 49562. [Abstract] [Full Text] [PDF]

				J. L. Watts, E. Phillips, K. R. Griffing, and J. Browse Deficiencies in C20 Polyunsaturated Fatty Acids Cause Behavioral and Developmental Defects in Caenorhabditis elegans fat-3 Mutants Genetics, February 1, 2003; 163(2): 581 - 589. [Abstract] [Full Text] [PDF]