Protection of C. elegans from Anoxia by HYL-2 Ceramide Synthase

See allHide authors and affiliations

Science  17 Apr 2009:
Vol. 324, Issue 5925, pp. 381-384
DOI: 10.1126/science.1168532


Oxygen deprivation is rapidly deleterious for most organisms. However, Caenorhabditis elegans has developed the ability to survive anoxia for at least 48 hours. Mutations in the DAF-2/DAF-16 insulin-like signaling pathway promote such survival. We describe a pathway involving the HYL-2 ceramide synthase that acts independently of DAF-2. Loss of the ceramide synthase gene hyl-2 results in increased sensitivity of C. elegans to anoxia. C. elegans has two ceramide synthases, hyl-1 and hyl-2, that participate in ceramide biogenesis and affect its ability to survive anoxic conditions. In contrast to hyl-2(lf) mutants, hyl-1(lf) mutants are more resistant to anoxia than normal animals. HYL-1 and HYL-2 have complementary specificities for fatty acyl chains. These data indicate that specific ceramides produced by HYL-2 confer resistance to anoxia.

The molecular pathways underlying resistance of Caenorhabditis elegans to oxygen deprivation (18) appear to be conserved, at least in part, between vertebrates and invertebrates. These pathways differ in C. elegans according to the developmental stage and oxygen tension. For example, hypoxia-inducing factor 1 (HIF-1) is not required for resistance of either nematode embryos (2) or young adults (table S1) to anoxia. However, it stimulates survival of embryos to hypoxia (9). On the other hand, mutation of the daf-2 gene, an insulin and insulin-like growth factor receptor–like gene, promotes survival of C. elegans to hypoxia or anoxia (1, 5), an effect that is mediated by the glyceraldehyde-3-phosphate dehydrogenases GPD-2 and GPD-3 (5). In C. elegans, mutations have not been detected that cause a reduction in viability under anoxia in adult hermaphrodites (5). We sought to identify essential components of the pathways regulating the adaptation of C. elegans to anoxia.

We tested the ability of young adult wild-type N2 Bristol C. elegans (72 hours after the L1 stage grown at 20°C) to survive anoxia at 20°C (0.001% O2). Survival rates were 99.5% ± 1.5% (n > 5), 81% ± 14.5% (n > 5), and 8.5% ± 9.5% (n > 5) after 24, 48, and 72 hours of anoxia, respectively (Fig. 1A). Animals that failed to survive anoxia became rodlike and showed the presence of propidium iodide–stained necrotic cells throughout their entire body (Fig. 1B).

Fig. 1.

The hyl-2(gnv1) mutation confers hypersensitivity to anoxia in C. elegans. (A) Survival of wild-type Bristol N2 animals exposed to anoxia for 24, 48, and 72 hours. Results are mean ± SD, n > 5. (B) Young adults N2 worms were stained with propidium iodide postanoxia and observed with a fluorescent microscope. After 16 hours of anoxia, dead worms displayed stained nuclei (white arrowhead) throughout the cell body (left and middle images). At 48 hours of anoxia, dead animals (top right image, black arrow) were intensively stained as a result of necrosis in all tissues, whereas living animals were unstained (top right image). (C) Survival of hyl-2(gnv1) and hyl-2(tm2031) mutants after 48 hours of anoxia (mean ± SD, n > 5).

In order to identify genes required for resistance to anoxia, we screened a library of mutants for animals with increased sensitivity to anoxia (table S1). After eliminating a role for other mutations in the initial mutant line, we characterized a mutation in the homolog of yeast longevity assurance gene 2 (hyl-2 gene), hyl-2(gnv1) (10) (fig. S1), that conferred increased sensitivity to anoxia (Fig. 1C and fig. S2). The mutation consists of two consecutive base substitutions at positions 1297 and 1298 (CATCAT → CAATAT) that result in the conversion of His168His169 residues into Gln168Tyr169 in the Lag motif of HYL-2. Sensitivity to anoxia was also observed in hyl-2(tm2031) mutants carrying a deletion of the hyl-2 gene (Fig. 1C). Both hyl-2(gnv1) and hyl-2(tm2031) mutants were indistinguishable from N2 animals and had a normal fecundity and life span (fig. S3). hyl-2(gnv1) mutants, as hyl-2(tm2031), were also more sensitive to heat shock at 36°C (fig. S4A) but had normal responses to thermal stress at 30°C for 7 days and to hypotonic shock (fig. S4, B and C). Introduction in hyl-2(gnv1) animals of a green fluorescent protein (GFP)::WT hyl-2 transgene under the control of the endogenous hyl-2 promoter (fig. S5C) showed gene expression from the larval to the adult stage with strong expression in the gut, the posterior bulb of the pharynx, the hypoderm, and unidentified cells of the head and the tail (Fig. 2A). Expression of the transgene partially restored resistance to anoxia and heat shock in hyl-2(gnv1) animals (Fig. 2B and fig. S4D). Collectively these experiments indicate that hyl-2 is required for adaptation of the nematode to anoxia.

Fig. 2.

Sensitivity to anoxia results from loss of function of HYL-2 ceramide synthase. (A) GFP staining of hyl-2(gnv1) adult animals expressing the transgene GFP::hyl-2 under the control of the hyl-2 promoter. (B) Survival after 48 hours of anoxia of hyl-2(gnv1) animals complemented with the transgene GFP::hyl-2 under the control of the hyl-2 promoter (mean ± SD, n > 5). (C) Functional complementation of mutant yeast cells by C. elegans hyl-1 and hyl-2 in lag1Δlac1Δ S. cerevisiae transformed with pRS424–hyl-1, pRS424–hyl-2, and pRS424–hyl-2(gnv1) (TRP1 plasmid). The compound 5-FOA (5-fluorouracil-6-carboxylic acid monohydrate) is a counterselection for cells that carry the URA3-based plasmid with yeast LAG1. Therefore, only cells that have lost this plasmid can grow, making growth dependent on function of the heterologous ceramide synthase.

hyl-2 belongs to a eukaryotic gene family known as longevity assurance genes (Lass genes, fig. S6). Several members encode dihydroceramide synthases for the de novo ceramide pathway (11) and are therefore called ceramide synthase (CerS) genes (12). These genes have sequence similarity in a domain called the Lag motif that is essential for enzyme activity (fig. S6) (13). In Saccharomyces cerevisiae, LAG1 and LAC1 are required for de novo ceramide synthase activity, and yeast lacking both genes are almost inviable. C. elegans has three ceramide synthase genes, hyl-1, hyl-2, and lagr-1. Simultaneous deletion of both hyl-1 and hyl-2 is lethal (10). Sphingolipids from C. elegans are somewhat different from their counterparts in other eukaryotes because they contain exclusively isosphingoid bases (14), which are presumably used by the dihydroceramide synthases from worms. Nevertheless, hyl-1 can complement the loss of function of yeast LAG1 and LAC1 (15). To test whether HYL-2 is a ceramide synthase, we transformed lag1Δlac1Δ yeast with an expression vector carrying the cDNA of either LAG1, hyl-1, or hyl-2. All expression vectors rescued the growth phenotype of lag1Δlac1Δ strain (Fig. 2C), indicating that hyl-2, like hyl-1, is an ortholog of LAG1. Expression of hyl-2(gnv1) cDNA in lag1Δlac1Δ yeast failed to rescue the lethal phenotype, demonstrating that His168, His169, or both are required for HYL-2 function (Fig. 2C). This result supports previous reports on the essential function of the evolutionarily conserved His residues of the Lag motif (13, 16). Thus, the inability of hyl-2 mutants to adapt to oxygen deprivation appears to result from a loss of function of the HYL-2 ceramide synthase.

In contrast to hyl-2 deficient animals, hyl-1(gk203) and hyl-1(ok976) deletion mutants (fig. S5A) resisted anoxia better than N2 worms (Fig. 3, A and B) (17). We therefore tested whether HYL-1 could substitute for HYL-2 in hyl-2–deficient worms. We fused hyl-1 to a sequence encoding GFP and expressed the construct under the control of hyl-2 promoter. This ensured appropriate temporal and spatial expression of the GFP::hyl-1 transgene (fig. S5D). Expression of the GFP::hyl-1 transgene conferred a small protection against anoxia but was significantly less efficient than the GFP::hyl-2 transgene (Fig. 3C). This finding was confirmed in three different transgenic lines, making it unlikely that suboptimal expression of the GFP::hyl-1 transgene was responsible for the weak beneficial effect. Thus, although HYL-1 can complement growth defects in yeast as efficiently as does HYL-2, it cannot completely substitute for HYL-2 to confer strong resistance to anoxia. Thus, anoxia tolerance may require one or more ceramide species that are either specifically or preferentially synthesized by HYL-2.

Fig. 3.

Incomplete rescue of hyl-2 loss of function by hyl-1. (A and B) Survival of hyl-1(gk203) and hyl-1(ok976) after 48 or 72 hours of anoxia. Results are mean ± SD [(A) n = 7 for 48 hours, n = 4 for 72 hours, *P < 0.05, two-tailed t test. (B) n > 5 for 48 hours and n > 5 for 72 hours, **P < 0.01, two-tailed t test.]. (C) Survival of hyl-2(gnv1) mutants complemented with the transgene GFP::hyl-1 under the control of hyl-2 promoter after 48 hours of anoxia. Results are mean ± SD, n > 5, **P < 0.01, ***P < 0.001, analysis of variance (ANOVA) test.

Dihydroceramide synthases combine a sphingoid base with a fatty acyl–coenzyme A (CoA) to form dihydroceramide, which is then desaturated by a dihydroceramide desaturase into ceramide in animals. In many organisms, multiple ceramide synthases are expressed, each displaying fatty acyl-CoA specificity to produce a diversity of ceramides differing in their fatty acyl chains (1820). Ceramides are the precursors for sphingolipids, including sphingomyelins, which are present in C. elegans (14). We quantified the major ceramide (Cer) and sphingomyelin (SM) species of C. elegans N2, hyl-1(ok976), and hyl-2(gnv1) animals by electrospray ionization mass spectrometry (ESI-MS). hyl-1(ok976) and hyl-2(gnv1) mutants expressed different types of Cers and SMs than N2 worms did (Fig. 4, fig. S7, and table S2). hyl-2–deficient worms [hyl-2(gnv1)]had fewer Cers and SMs with C20 to C22 fatty acyl chains and more with C24 to C26 fatty acyl chains than did N2 animals. In contrast, hyl-1–deficient worms [hyl-1(ok976)] expressed more C20 to C22 Cers and SMs than did N2 worms, but they contained the same or lesser amounts of C24 to C26 Cers and SMs. These data indicated that efficient synthesis of C20 to C22 Cers requires HYL-2 whereas that of C24 to C26 Cers is mainly dependent on HYL-1. To verify this, we measured dihydroceramide synthase activity in microsomes isolated from N2 and mutant worms with 3H-sphinganine and acyl-CoA substrates of various lengths (fig. S8, A and C). Membranes from hyl-2(tm2031) mutants catalyzed synthesis of more C26 Cers, whereas membranes from hyl-1(ok976) mutants catalyzed synthesis of more C22 Cers. All membranes were equally active with use of C24-acyl CoA (fig. S8B). These results are consistent with the ESI-MS results and confirm that the two Cer synthases have different specificities for fatty acyl chains. They indicate that HYL-2 may promote survival of animals in anoxic conditions by producing C20 to C22 Cers and SMs. In support of this hypothesis, the hyl-1(ok976) mutant worms produced larger total amounts of Cers and SMs than did N2 and hyl-2(gnv1) mutants (fig. S9A). Moreover, expression of the GFP::hyl-1 transgene under the control of the hyl-2 promoter in hyl-2(gnv1) worms failed to restore normal amounts of C20 to C22 ceramides (fig. S9B), which could explain why the transgene did not restore normal resistance to anoxia in hyl-2(gnv1) animals.

Fig. 4.

Ceramide, sphingomyelin, and phosphatidylcholine (PC) quantification in N2, hyl-2(gnv1), and hyl-1(ok976) worms. Cers, SMs, and PC species were quantified by ESI-MS (see fig. S7 for quantification of PC species). Cers and SMs were quantified after base hydrolysis of glycerophospholipids (23). Graphs of the amounts are shown on a logarithmic scale ± SD (n = 4). The amount of each lipid species found in mutant worms was divided by the amount found in wild-type Bristol N2 worms and shown in a graph below the corresponding graph of the total amounts. Acyl chain lengths corresponding to charge/mass (m/z) values are shown in boxes (see also table S2). The entire profiles of Cer and SM species of N2, hyl-2(gnv1), and hyl-1(ok976) animals are different with P = 0.002 for Cers and P = 0.007 for SMs (multivariate ANOVA test).

Ceramides function in radiation-induced apoptosis of cells in the germ line (21). However, apoptosis appears not to account for the sensitivity of HYL-2–deficient animals to anoxia because lack of CED-3 caspase activity in ced-3(n717);hyl-2(gnv1) double mutants did not extend survival of these animals during anoxia (fig. S10). Sphingosine 1-phosphate, an anti-apoptotic derivative of ceramide, also appears not to function in resistance to anoxia because resistance of sphk-1(ok1097) animals, in which conversion of isosphingosine to isosphingosine 1-phosphate is prevented by a null allele of sphingosine kinase, was normal and those of sphk-1(ok1097);hyl-1(ok976) and sphk-1(ok1097);hyl-2(tm2031) double mutants were not modified (fig. S11, A and B). A block in ceramide synthase is expected to lead to an increase in isosphingoid bases, which are ceramide precursors. Therefore, we examined the amounts of isosphingoid bases in the sphk-1(ok1097) and hyl mutants by MS-ESI. Isosphingoid bases accumulated in hyl-2 mutants and were less abundant in hyl-1 mutants (fig. S11C). However, sphk-1(ok1097) worms accumulated even more isosphingoid bases than either hyl mutant but adapted to anoxia normally (fig. S11, A to C), suggesting that accumulation of isosphingoid bases did not influence survival under anoxic conditions.

The daf-2/daf-16 insulin-like signaling pathway is involved in oxygen deprivation survival in C. elegans (1, 5). We also found that daf-2(e1370) mutants survived 72 hours of anoxia better than did N2 worms (fig. S12A). To determine whether hyl-2 and daf-2 interact genetically, we generated daf-2(e1370);hyl-2(gnv1) double mutants and tested their sensitivity to 48 hours of anoxia. Resistance of the double mutants was significantly improved compared with that of hyl-2(gnv1) mutants, whereas it was decreased compared with that of daf-2(e1370) mutants (fig. S12B). Thus, it appears that, with respect to anoxia resistance, HYL-2 and DAF-2 are acting in parallel pathways that mutually influence each other.

We have shown that a dihydroceramide synthase, with a distinct substrate specificity, provides an important function in the anoxia response in C. elegans. Rather than their quantity, it is the chemical structure of ceramide species that seems to be important for resistance to anoxia. Ceramides have been reported previously to be effectors of kinases or phosphatases in various biological processes (22). It is most likely that the activity of key ceramide species during anoxia relies on interaction of ceramides with other molecules integrated in a cell survival pathway.

Supporting Online Material

Materials and Methods

Figs. S1 to S12

Tables S1 and S2


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article