Lack of Peroxisomal Catalase Causes a Progeric Phenotype in Caenorhabditis elegans * □ S

Studies using the nematode Caenorhabditis elegans as a model system to investigate the aging process have implicated the insulin/insulin-like growth factor-I signaling pathway in the regulation of organismal longevity through its action on a subset of target genes. These targets can be classified into genes that shorten or extend life-span upon their induction. Genes that shorten life-span include a variety of stress response genes, among them genes encoding catalases; however, no evidence directly implicates catalases in the aging process of nematodes or other organisms. Using genetic mutants, we show that lack of peroxisomal catalase CTL-2 causes a progeric phenotype in C. elegans . Lack of peroxisomal catalase also affects the developmental program of C. elegans , since (cid:1) ctl-2 mutants exhibit decreased egg laying capacity. In contrast, lack of cytosolic catalase CTL-1 has no effect on either nematode aging or egg laying capacity. The (cid:1) ctl-2 mutation also shortens the maximum life-span of the long lived (cid:1) clk-1 mutant and accelerates the onset of its egg laying period. The more rapid aging of (cid:1) ctl-2 worms is apparently not due to increased carbonylation of the major C. elegans proteins, although altered peroxisome morphology in the (cid:1) ctl-2 mutant suggests that changes in peroxisomal function, including increased production of

A subset of genes of the nematode Caenorhabditis elegans are direct targets (1,2) of the FOXO family transcription factor DAF-16 (3), a key regulator of the insulin/insulin-like growth factor I signaling pathway implicated in the aging process (1,4). These target genes usually contain specific nucleotide sequences in their upstream regulatory regions capable of binding DAF-16 and can be directly repressed or activated upon binding of DAF-16 (5). Specific transcriptional regulation of target genes by DAF-16 is purported to lead to an extended or shortened life-span for the nematode. Among the target genes regulated by DAF-16 are those encoding heat shock proteins and cytochrome p450s involved in the cell stress response and genes encoding proteins responsible for antioxidant defense, such as the mitochondrial superoxide dismutase SOD-3, the metallothionein homolog MTL-1, and the catalases CTL-1 and CTL-2 (5).
Superoxide dismutases (SODs) 1 and catalases are scavengers of reactive oxygen species (ROS) and H 2 O 2 , respectively, and have been suggested to play important roles in the aging process in C. elegans (6). Catalase levels are increased in the long-lived C. elegans mutants age-1 (6 -8), eat-2 (9), and daf-2 (10,11), whereas catalase gene expression is decreased in a short lived strain mutated for DAF-16 (12). The daf-16 mutation also largely suppresses the increases in catalase activity observed in age mutants (10).
In addition to the insulin/insulin-like growth factor-I signaling pathway, dietary restriction and oxidative stress are also thought to be major determinants of the aging process in C. elegans and other model organisms (10,13). Under conditions of dietary restriction that extend monoxenic, and especially axenic, worm life-span, catalase and SOD levels have been found to be increased dramatically in a DAF-16-independent manner (10). A correlation between increases in SOD and catalase activity and hyperresistance to oxidative stress has also been observed in some long lived mutants (8). During the dauer larval stage when the nematode is developmentally arrested, nonfeeding, and able to survive several times the normal life-span, catalase and/or SOD activities are substantially up-regulated (8,14). Moreover, treatment of worms with SOD/ catalase mimetics extends their life-spans (15,16). These and other findings from different model organisms (13,(17)(18)(19)(20) suggest that catalases play an important role in the aging process; however, a definitive statement on the role of individual catalases in the aging process in C. elegans requires investigation of this process in mutant strains deleted for the individual catalase genes.
Two forms of catalase, the cytosolic CTL-1 and the peroxisomal CTL-2, have been reported for C. elegans (21,22). A previous attempt to engage CTL-1 directly in the control of aging in C. elegans was unsuccessful (21). Molecular features of the catalase gene locus in C. elegans, namely the presence of different catalase genes exhibiting a high level of sequence identity, make it difficult, if not impossible, to use the results of methods such as RNA interference (4,5,23) to permit definitive statements on the roles of individual catalases in the aging process in C. elegans. Here, using C. elegans strains mutated for individual catalase genes, we present evidence directly im- plicating peroxisomal catalase CTL-2, but not cytosolic CTL-1, in the aging process and developmental program of C. elegans and show that a lack of CTL-2 causes a progeric phenotype in the nematode.
Determination of Catalase Activity-Catalase activity of lysates of mixed stage animals prepared by sonication in 20 mM Tris-HCl, pH 7.5, 50 mM potassium acetate, 2 mM EDTA, 100 mM sorbitol was measured either colorimetrically (27) or by staining in situ with ferricyanide following native gel electrophoresis (28).
Determination of Life-span-To measure nematode life-span, plates containing 10 -14 worms were incubated at 20°C and scored daily for surviving animals. Worms were transferred to fresh plates every 2-3 days. The starting point for life-span determination was at hatching. Lost, bagged, and exploded animals were excluded from analysis. To measure chronological life-span in wild-type and mutant strains of the yeast Saccharomyces cerevisiae, cells were grown to stationary phase in YNBD medium (0.67% yeast nitrogen base without amino acids, 2% glucose, 0.08% Complete Supplement Mixture (Bio 101, Carlsbad, CA)), pelleted, and washed three times in distilled, deionized water and then transferred to distilled, deionized water at a density of 2-3 cells/l. Cells were incubated at 30°C with rotational shaking in Erlenmeyer flasks at a flask volume/medium volume ratio of 5:1. To score for surviving cells, cell suspensions were seeded onto YEPD (1% yeast extract, 2% peptone, 2% glucose) agar plates, and the individual yeast colonies were counted.
Measurement of Protein Carbonylation-Protein carbonylation was determined using the OxyBlot Protein Oxidation Detection Kit (Intergen, Purchase, NY) according to the manufacturer's protocol and quantitated by analysis of autoradiograms with a GS-800 calibrated densitometer using Quantity One version 4.2.3 software (Bio-Rad).
Confocal and Electron Microscopy-Confocal microscopy was performed with a LSM510 META laser-scanning microscope (Carl Zeiss MicroImaging, Thornwood, NY). Peroxisomes were visualized with a fluorescent protein chimera of green fluorescent protein tagged at its carboxyl terminus with the peroxisome targeting signal 1 tripeptide Ser-Lys-Leu (23). Samples were prepared for electron microscopy as described (29) and examined with a Philips 410 electron microscope. Electron microscopic images were captured with a MegaView III CCD camera (Soft Imaging System, Lakewood, CO). Sizes of peroxisomes and lipid droplets were determined using UTHSCSA Image Tool 2.00.

Characterization of the Catalase Genes of C. elegans and
Analysis of Catalase Enzymatic Activity in the Nematode-The C. elegans genome contains three catalase genes in tandem: ctl-3, ctl-1, and ctl-2 (Figs. 1 and S1). 2 DAF-16 binding motifs (1) are present upstream and downstream of all three genes ( Fig. 1). The catalase genes exhibit extensive sequence identity. ctl-1 is identical to ctl-2 between nucleotides 1 and 304 and to ctl-3 from nucleotide 304 until its end (Figs. 1 and S1). The coding region of ctl-2 exhibits 76.7% sequence identity to the coding region of ctl-3 (Figs. 1, S1, and S2). The intervening region between the ctl-1 and ctl-3 genes is 100% identical in sequence to the intervening region between the ctl-1 and ctl-2 genes (Figs. 1 and S1). Extensive sequence identity among the catalase genes makes it difficult to analyze their individual expression by techniques such as Northern blotting and suggests that the results of RNA interference and gene microarray analysis should be interpreted with caution.
The catalases themselves can be distinguished from one another because of differences in their biochemical properties. Because the pI values of the three catalases differ significantly (Table I), they can be readily separated by native gel electrophoresis and detected by staining for enzymatic activity ( Fig.  2A). We isolated a worm strain, LB90 (ua90)II, harboring a deletion of the ctl-2 gene (Figs. 1 and S1, ⌬ctl-2). The ⌬ctl-2 mutant strain exhibits no detectable CTL-2 enzymatic activity ( Fig. 2A). DNA sequencing of the strain ctl-1(u800)II, which had been claimed to exhibit decreased catalase activity due to a premature termination codon in the ctl-1 gene (21), showed that the entire ctl-1 gene and extensive sequence upstream were missing in this mutant (Figs. 1 and S1, ⌬ctl-1). Native gel electrophoresis confirmed the absence of CTL-1 enzymatic activity in the ⌬ctl-1 strain ( Fig. 2A). Attempts to isolate mutants of the ctl-3 gene were unsuccessful. In dauer larvae, a developmental stage during which animals do not feed and fat metabolism is shifted to fat storage (30), the level of CTL-1 activity is increased, whereas the level of peroxisomal CTL-2 activity appears to be similar to that found in wild-type worms ( Fig.  2A). ⌬ctl-2 and ⌬ctl-1 mutants do not fail to form the dauer stage (data not shown).
Quantitative colorimetric analysis of catalase activity showed that deletion of the ctl-2 gene (Fig. 2B, ⌬ctl-2, and   Table II) reduced catalase activity to ϳ20% of the total catalase activity observed in wild-type worms, whereas deletion of the ctl-1 gene (Fig. 2B, ⌬ctl-1, and Table II) led to a much smaller reduction in catalase activity to ϳ75% of the total catalase activity observed in wild-type worms. Therefore, peroxisomal catalase CTL-2 contributes most of the total catalase activity in C. elegans.
Lack of Peroxisomal, but Not Cytosolic, Catalase Causes Accelerated Aging of C. elegans-We compared the mean and maximum life-span and egg laying capacity of catalase mutant worms versus wild-type worms to determine whether the dif-ferent catalases in C. elegans could have a role in the development and aging of the nematode. ⌬ctl-1 mutants showed no difference in life-span or egg laying capacity compared with wild-type worms, whereas ⌬ctl-2 mutants had a significantly shortened (16%) life-span and decreased egg laying capacity ( Fig. 3 and Table II). Introduction of the ⌬ctl-2 mutation into long lived ⌬clk-1 mutant worms did not result in a shortening of their mean life-span ( Fig. 3 and Table II), consistent with a previous finding that the extended life-span of ⌬clk-1 mutant worms is not related directly to the antioxidant action of catalase (34). Nevertheless, we observed a significantly shortened

FIG. 3. Effects of mutation of the ctl-1 and ctl-2 genes on life-span and egg laying capacity of wild-type and ⌬clk-1 mutant worms.
A, mutation of the ctl-2 gene (⌬ctl-2) shortens the mean life-span of wild-type (WT) worms (strain N2) by 16% and the maximum life-span of the ⌬clk-1 long living mutant by 14%. In contrast, the ⌬ctl-1 mutation does not influence significantly the mean life-span of wild-type worms. B, ⌬ctl-2 mutants lay 25% less eggs than do wild-type or ⌬ctl-1 mutant worms (see also Table II). Lack of CTL-2 also advances the egg laying period in the ⌬clk-1 mutant by 12 h.

FIG. 2. Activity of catalases in C. elegans.
A, native gel of total nematode lysates stained for catalase activity. In mutant worms, enzymes other than the three catalases that manifest catalase activity are up-regulated (arrowheads). Equal amounts of protein were loaded in each lane except for the lane labeled Dauer, which contained 7.5 times less protein than each of the other lanes. B, total catalase activity measured in vitro in mutant mixed stage worms as a percentage of total catalase activity in wildtype worms.
(14%) maximum life-span and an acceleration of ϳ12 h in the onset of the egg laying period in ⌬clk-1;⌬ctl-2 double mutant worms as compared with ⌬clk-1 mutant worms ( Fig. 3 and Table II).
Increased protein carbonylation has been observed during cell aging (31). Carbonylation of major worm proteins such as vitellogenin-6 (32) increases from 5 days to 10 days in both wild-type worms and ⌬ctl-1 and ⌬ctl-2 mutant worms, but the overall increase in carbonylation is less for both mutants than for wild-type worms (Fig. 4). This smaller increase in carbonylation observed for the ⌬ctl-1 and ⌬ctl-2 mutants may result from a compensatory up-regulation of other antioxidant enzymes in these mutants. Evidence for such compensatory changes in the synthesis of unknown enzymes with catalase activity could be seen for the ⌬ctl-1 and ⌬ctl-2 mutants ( Fig. 2A,  arrowheads). The carbonylation of major protein species therefore appears not to be the cause of the progeric phenotype of the ⌬ctl-2 mutant.
Cells of the ⌬ctl-2 Mutant Exhibit Abnormal Peroxisome Morphology-A deficiency in the peroxisomal ␤-oxidation enzyme acyl-CoA oxidase leads to increased levels of intraperoxisomal H 2 O 2 and alters the morphology of peroxisomes (33). A lack of peroxisomal catalase CTL-2 might be expected to exert a similar effect on peroxisome morphology due to decreased breakdown of intraperoxisomal H 2 O 2 . We therefore examined the morphology of peroxisomes in wild-type and ⌬ctl-2 mutant worms by fluorescent confocal microscopy and electron microscopy. The mean size of peroxisomes is increased in ⌬ctl-2 mutant worms versus wild-type worms (Fig. 5A), and peroxisomes tend to cluster in the ⌬ctl-2 mutant (Fig. 6A). Electron microscopy revealed that these clustered peroxisomes are often associated with lipid vesicles (Fig. 6B, left panel) or with specific multivesicular structures of unknown origin and function (Fig.  6B, right panel). Fat metabolism is apparently normal in ⌬ctl-2 mutant worms, since they accumulate lipid droplets in their cells of the same size accumulated by cells of wild-type worms (Fig. 5B). DISCUSSION The C. elegans genome contains three genes encoding different catalases. To our knowledge, no other metazoan organism has been reported to have multiple catalase genes. The most active form of catalase in C. elegans is peroxisomal catalase CTL-2, which contributes ϳ80% of total catalase activity in the nematode. The remaining catalase activity is contributed by cytosolic catalase CTL-1, a previously unreported catalase CTL-3 encoded by the open reading frame Y54G11A.13, and unknown enzymes exhibiting catalase activity (cf. Fig. 2A). Attempts to obtain a strain mutant for the ctl-3 gene were unsuccessful; however, expression of a green fluorescent protein reporter under the control of a promoter fragment extending 657 bp upstream from the ctl-3 initiating codon was localized to pharyngeal muscle cells and neurons (Supplemental Data, Fig. S3).
Previous results from studies employing RNA interference had suggested the importance of CTL-2 activity for longevity in C. elegans (5). However, because of the high degree of nucleotide identity among the catalase genes, the results of RNA interference studies must be interpreted with caution. Using worms deleted for the ctl-2 gene, we have demonstrated directly for the first time that a lack of peroxisomal catalase accelerates aging in C. elegans. A lack of CTL-2 resulted in a 16% reduction in the mean life-span of worms. The reduced viability of ⌬ctl-2 mutant worms was also reflected in the significantly smaller number of eggs they laid and their slightly delayed and extended egg laying period. Importantly, mutation of the ctl-1 gene did not affect either nematode lifespan or egg laying capacity, demonstrating that the effects of ctl-2 mutation on these characteristics were not simply the result of reduced overall catalase activity. Introduction of the ⌬ctl-2 mutation into the long lived ⌬clk-1 mutant increased the FIG. 4. Carbonylation levels of the major proteins of wild-type and catalase mutant worms. At 10 days, the level of protein carbonylation in wild-type worms (WT) is more than twice that observed at 5 days (taken as 100%), whereas only a modest increase is observed for the ⌬ctl-1 and ⌬ctl-2 mutant worms over the same period.
FIG. 5. Cells of ⌬ctl-2 mutant worms contain enlarged peroxisomes but lipid droplets of normal size. A, the mean diameter of peroxisomes in hypodermal cells of ⌬ctl-2 mutant worms is significantly larger than that of peroxisomes in hypodermal cells of wild-type (WT) worms. B, no significant difference in the mean diameter of lipid droplets of gut cells is observed between wild-type worms and ⌬ctl-2 mutant worms.
survival of ⌬clk-1 mutant worms during the first 19 days of life but resulted in the rapid death of the ⌬clk-1;⌬ctl-2 double mutant worms after 19 days (cf. Fig. 3A), resulting in a shortened maximum life-span, but no overall change in mean lifespan, for the double mutant worms. These findings are consistent with previous data showing that the extended mean lifespan of the ⌬clk-1 mutant is not related directly to the antioxidant action of catalase (34). Interestingly, introduction of the ⌬ctl-2 mutation into the ⌬clk-1 mutation background accelerated the egg laying period of the ⌬clk-1;⌬ctl-2 double mutant by 12 h compared with the ⌬clk-1 mutant. Development and behavior are considerably retarded, and reproduction is strongly reduced in ⌬clk-1 mutant worms as compared with wild-type worms (35). Reduced reproduction in ⌬clk-1 worms is the result of slower germ line development, which in turn is dependent on ROS-mediated oxidation of low density lipoprotein-like lipoproteins (36). Suppression of the sod-1 gene encoding cytosolic SOD has been shown to stimulate germ line development, apparently through the increased oxidation of lipoproteins (36). The synthesis of complex lipids and the assembly of lipoproteins depend on peroxisomes (41,42). An expected increase in the levels of ROS resulting from increased levels of H 2 O 2 inside peroxisomes lacking catalase would be expected to lead to increased oxidation of lipids, including those that are assembled into low density lipoproteins in the nematode. Oxidation of lipids in cells of ⌬ctl-2 mutant worms could also happen outside peroxisomes due to leakage of H 2 O 2 from abnormal peroxisomes. Both confocal and electron microscopy showed the presence of peroxisomes with abnormal morphology in cells of ⌬ctl-2 mutant worms (cf. Fig. 6). Therefore, the progeric phenotype of ⌬ctl-2 mutant worms and the observed advancement in the egg laying period of the ⌬clk-1;⌬ctl-2 double mutant may be due to increased levels of local ROSs produced in the absence of CTL-2 that act in the oxidation of lipids both within and outside the peroxisome. This is particularly significant given that the carbonylation of major protein species is not the cause of the progeric phenotype of the ⌬ctl-2 mutant.
The effects of a lack of peroxisomal catalase on organismal longevity are not restricted to C. elegans. In the yeast S. cerevisiae, loss of cytosolic catalase T has little effect on chronological life-span, although its levels are highly induced during the stationary phase of growth (18). We measured the chrono-logical life-spans of wild-type S. cerevisiae and of the mutants ⌬cta1 and ⌬ctt1, respectively. We found that a lack of peroxisomal catalase, but not cytosolic catalase, decreased the viability of yeast by ϳ15-fold (Supplemental Data, Fig. S4). Therefore, a shorter life-span may be a general consequence of a lack of peroxisomal catalase.
In closing, a lack of peroxisomal catalase CTL-2, but not cytosolic catalase CTL-1, causes a progeric phenotype in the nematode C. elegans. The ⌬ctl-2 mutant of C. elegans represents a convenient model not only for the study of aging but possibly also for the study of the human diseases acatalasemia and hypocatalasemia (23,43,44). Our results support and extend recent findings that peroxisomes not only have important roles in cell metabolism but also are involved in the developmental processes of eukaryotic organisms (23,43,25).