Autophagy Genes Are Essential for Dauer Development and Life-Span Extension in C. elegans

See allHide authors and affiliations

Science  05 Sep 2003:
Vol. 301, Issue 5638, pp. 1387-1391
DOI: 10.1126/science.1087782


Both dauer formation (a stage of developmental arrest) and adult life-span in Caenorhabditis elegans are negatively regulated by insulin-like signaling, but little is known about cellular pathways that mediate these processes. Autophagy, through the sequestration and delivery of cargo to the lysosomes, is the major route for degrading long-lived proteins and cytoplasmic organelles in eukaryotic cells. Using nematodes with a loss-of-function mutation in the insulin-like signaling pathway, we show that bec-1, the C. elegans ortholog of the yeast and mammalian autophagy gene APG6/VPS30/beclin1, is essential for normal dauer morphogenesis and life-span extension. Dauer formation is associated with increased autophagy and also requires C. elegans orthologs of the yeast autophagy genes APG1, APG7, APG8, and AUT10. Thus, autophagy is a cellular pathway essential for dauer development and life-span extension in C. elegans.

Under conditions of high population density, limited food, or increased temperature, Caenorhabditis elegans nematodes reversibly arrest in an alternate third larval stage, the dauer diapause, which is specialized to survive in an unfavorable environment (1). Entry into dauer is regulated by different signaling pathways, including transforming growth factor (TGF-β), cyclic guanosine monophosphate, and insulin-like signaling. Single-gene mutations in the insulin-like signaling pathway that promote constitutive dauer phenotypes also prolong adult life-span (2). Despite extensive characterization of regulatory signals, the cellular pathways involved in dauer morphogenesis and longevity are not well understood.

The regulation of dauer entry in C. elegans shares similarities with the induction of autophagy in other eukaryotes. The environmental cues that promote dauer morphogenesis in C. elegans are also potent stimulators of autophagy in yeast and mammalian cells (3, 4). Through autophagy, cells generate sufficient pools of amino acids to synthesize proteins that are essential for survival when the environmental food supply is limited. Components of the insulin-like signaling pathway that positively (e.g., daf-18) and negatively (e.g., age-1 and akt1/akt2) regulate dauer entry in C. elegans may also regulate autophagy in mammalian cells (5, 6). Furthermore, autophagy may be involved in life-span extension induced by dietary restriction in mammals (7), in delaying leaf senescence (8, 9), and in developmental processes that require extensive cellular and tissue remodeling (10). We therefore postulated that autophagy is essential in both dauer development and life-span extension in C. elegans.

To test this, we used RNA interference (RNAi) (11) to inhibit autophagy genes in C. elegans that carry a loss-of-function mutation in daf-2, the C. elegans insulin-like tyrosine kinase receptor (12). daf-2(e1370) animals are temperature-sensitive daf-c class mutants (13), with inappropriate constitutive dauer entry at 25°C and extended adult life-span at lower temperatures (such as 15° to 22.5°C) (12, 14) that allow reproductive growth. C. elegans contains candidate homologs of several yeast autophagy (APG) genes. We focused on bec-1, the C. elegans ortholog (T19E7.3) of yeast APG6/VPS30, because its mammalian counter-part, beclin 1, is important in autophagy, embryonic development, and tumor suppression (15, 16). In both yeast and mammalian cells, Apg6/Vps30/Beclin 1 is part of a Class III phosphatidylinositol 3-kinase complex that is thought to be important in mediating localization of other Apg proteins to preautophagosomal structures (17, 18).

C. elegans bec-1 encodes a predicted 375–amino acid, coiled-coil protein that shares 28% homology with yeast Apg6/Vps30p and 31% homology with human Beclin 1. To determine whether bec-1 is a functional homolog of yeast APG6/VPS30, we tested its ability to restore autophagy and vacuolar protein sorting in APG6/VPS30-disrupted (Δapg6/vps30) yeast. Using differential interference contrast microscopy, we found that Δapg6/vps30 yeast transformed with C. elegans bec-1 or human beclin 1, but not with empty vector, showed an increase in nitrogen starvation–induced autophagy similar to that of Δapg6/vps30 yeast transformed with yeast APG6/VPS30 (Fig. 1A). In contrast, Δapg6/vps30 yeast transformed with either C. elegans bec-1 or human beclin 1 were unable to properly sort and mature the vacuolar protein carboxypeptidase Y (CPY) (Fig. 1B, top). This defect was not due to a lack of stable C. elegans BEC-1 or human Beclin 1 protein expression in yeast (Fig. 1B, bottom) and could not be overcome by overexpression of the Apg6/Vps30p binding partner, Vps38p, that is part of the yeast vacuolar protein sorting complex (18) (Fig. 1B, middle). Thus, similar to human beclin 1 (15), C. elegans bec-1 complements the autophagy but not the vacuolar protein sorting function of APG6/VPS30 in yeast. In addition, we examined the C. elegans bec-1 expression pattern using a reporter gene in which the bec-1 promoter was fused to green fluorescent protein (GFP) coding sequences (19). We found that the bec-1 promoter is transcriptionally active in several different tissues (including the pharynx, intestine, hypodermis, nervous system, and reproductive organs) that undergo morphologic changes during dauer formation (1) (fig. S1).

Fig. 1.

C. elegans bec-1 complements the autophagy but not the vacuolar protein sortingfunction of yeast APG6/VPS30. (A) Quantitation of autophagy in nitrogen-starved wild-type (wt) yeast and in Δapg6/vps30 yeast transformed with empty vector or with plasmids that encode either yeast Apg6/Vps30p, C. elegans (Ce) FLAG epitope–tagged BEC-1, or human FLAG epitope–tagged Beclin 1 (19). Results represent the mean (±SEM) percentage of cells with autophagic bodies within the vacuole for triplicate samples. Similar results were observed in three independent experiments. (B) Sorting of vacuolar protein CPY in (top) Δapg6/vps30 yeast or (middle) Δapg6/vps30 yeast with VPS38 overexpression, transformed with empty vector or with plasmids that encode either yeast Apg6/Vps30p, C. elegans FLAG epitope–tagged BEC-1, or human FLAG epitope–tagged Beclin 1 (19). p2, precursor form; m, sorted mature form. Bottom: A Western blot (WB) analysis of yeast cell lysates with a polyclonal antibody to FLAG (19).

To investigate the effects of inhibiting bec-1 expression on the constitutive dauer phenotype of daf-2(e1370) mutants at 25°C, we used bec-1 RNAi treatment that decreased bec-1 mRNA levels but did not interfere with wild-type C. elegans development (19) (Fig. 2A). At 25°C, both untreated daf-2(e1370) (Fig, 2A) and GFP RNAi control–treated daf-2(e1370) formed normal dauers. These normal dauers are characterized by hyperpigmented granules in the intestine; radial constriction and elongation of the body and pharynx; the presence of a lateral, ridged, cuticular thread known as the dauer alae; increased fat storage as detected by Sudan black staining (Fig. 2B); resistance to sodium dodecyl sulfate (SDS) treatment; long-term survival at 25°C; and the ability to resume reproductive growth upon transfer to 15°C (19).

Fig. 2.

bec-1 RNAi blocks normal dauer formation of daf-2(e1370) mutants. (A) Developmental stages of untreated and bec-1 RNAi-treated wild-type N2 and daf-2(e1370) animals grown at 25°C. Abnl, abnormal. (B) Representative Nomarski photomicrographs of the normal dauer phenotype of untreated daf-2(e1370) animals and the abnormal dauer phenotype of daf-2(e1370); bec-1 (RNAi) animals grown at 25°C. The left column shows control, wild-type, nondauer, L3 larvae grown at 25°C. Top center insert: A dauer lateral alae. Middle row: The pharynx. Bottom row: Animals stained with Sudan black.

In contrast, at 25°C, the majority of daf-2(e1370); bec-1 (RNAi) animals formed abnormal dauers (Fig. 2A). Like the daf-2(e1370) dauers, these animals underwent a developmental arrest in the L3 gonadal stage and had some hyperpigmented granules in the intestine, some increased fat storage, and some radial constriction and elongation of the body and pharynx. However, the hyperpigmented granules were unevenly distributed and the magnitudes of fat storage, radial constriction, and elongation of the body and pharynx were less than that observed in untreated daf-2(e1370) animals (Fig. 2B). Furthermore, the daf-2(e1370); bec-1 (RNAi) animals failed to form dauer alae, were not resistant to SDS, died within a few days at 25°C, and usually did not resume reproductive growth on transfer to 15°C. These observations demonstrate that bec-1 is not required for dauer initiation, but is required for normal dauer morphogenesis in daf-2(e1370) mutant animals (20). The death of these animals also suggests that autophagy genes may be critical for dauer survival, a finding that is consistent with the essential role of autophagy genes in survival during nutrient deprivation states in other eukaryotic organisms (21, 22).

Next, we evaluated the effects of bec-1 RNAi treatment on the life-span extension of daf-2(e1370) worms during reproductive growth at 15°C (Fig. 3). Consistent with previous studies (12, 14), daf-2(e1370) animals had a significant life-span extension as compared to wild-type animals (median survival was 48 versus 28 days; P < 0.001, log-rank test). This life-span extension of daf-2(e1370) animals was reduced by bec-1 RNAi treatment (median survival was 28 days; P < 0.001, log-rank test). Although bec-1 RNAi had a slight inhibitory effect on the survival of wild-type animals (24 days for bec-1 RNAi-treated animals versus 28 days for untreated animals), the effect of bec-1 RNAi on shortening survival of the daf-2(e1370) mutants was significantly greater than its effects on shortening survival of wild-type animals. [Cox proportional-model hazard ratios were 5.4 for bec-1 RNAi-treated daf-2(e1370) animals and 1.7 for bec-1 RNAi-treated wild-type animals; P < 0.001]. These data indicate that bec-1 is essential for the adult life-span extension of daf-2(e1370) nematodes.

Fig. 3.

bec-1 is required for life-span extension in daf-2(e1370) mutants. A Kaplan-Meier survival curve shows the life-span of adult offspring of wild-type N2 (blue) and mutant daf-2(e1370) (red) untreated (solid line) or bec-1 RNAi-injected (dotted line) worms maintained at 15°C (n = 81 to 273 animals per experimental group) (19).

The requirement for bec-1 in daf-2(e1370) dauer formation and life-span extension suggested that autophagy may be involved in these phenotypes. To visualize autophagy in intact nematodes, we generated transgenic wild-type N2 and daf-2(e1370) animals that expressed a GFP-tagged version of the C. elegans gene product LGG-1 (19), which is an ortholog of yeast Apg8/Aut7p and mammalian MAP-LC3. During autophagy, Apg8/Aut7p and MAP-LC3 localize to preautophagosomal and autophagosomal membranes (2325), and their punctate staining pattern (versus a diffuse pattern in the absence of autophagy) provides a useful marker of autophagy (2527). Wild-type N2 and daf-2(e1370) animals carrying an extrachromosomal array that expressed GFP::LGG-1 had diffuse cytoplasmic GFP expression in multiple tissues in different stages of development, including the nervous system, pharynx, intestine, hypodermis, somatic gonad, and vulva (28). The tissue distribution of GFP::LGG-1 expression was similar to that observed for GFP regulated by the endogenous bec-1 promoter.

During dauer formation, we observed a marked change in the subcellular localization pattern of GFP::LGG-1 in hypodermal seam cells, a cell type known to be important for certain dauer-associated morphological changes, including the formation of the dauer alae and the radial constriction of the body (2931) (Fig. 4A). Therefore, to quantitate differences in autophagic activity, we counted the number of GFP-positive punctate areas per hypodermal seam cell in N2 and daf-2(e1370) L3 larvae grown at 15°C (during reproductive growth) and at 25°C [during dauer diapause for daf-2(e1370) but not N2 animals] (Fig. 4B). In N2 animals, similar numbers of GFP-positive punctate areas were observed during growth at 15°C and at 25°C. In contrast, there was a marked increase in the number of seam cell, GFP-positive, punctate areas in daf-2(e1370) animals during dauer formation at 25°C as compared to during reproductive growth at 15°C (P < 0.001, t test). This increase demonstrates that dauer morphogenesis is associated with autophagy induction in hypodermal seam cells. In addition, daf-2(e1370) mutant animals grown at 15°C had significantly more seam cell GFP-positive punctate areas than wild-type N2 animals grown at 15°C (P < 0.001, t test). This latter observation demonstrates that the wild-type daf-2 gene product also functions to suppress autophagy during reproductive growth and that the insulin-like signaling pathway negatively regulates autophagy in C. elegans.

Fig. 4.

Dauer formation in daf-2(e1370) mutant animals is associated with an increase in autophagy in lateral hypodermal seam cells. (A) GFP::LGG-1 expression in the seam cells of wild-type N2 and daf-2(e1370) L3 larvae grown at 15°C and 25°C. The arrow shows representative GFP-positive punctate areas that label preautophagosomal and autophagosomal structures. (B) Quantitation of GFP::LGG-1–positive punctate areas in the seam cells of wild-type N2 and daf-2(e1370) animals grown at 15°C and 25°C (mean + SEM for at least 50 seam cells from 10 different animals). (C to I) Representative electron micrographs of lateral hypodermal seam cells in a wild-type N2 control animal in (C) the L2 larval stage and (D) the L3 larval stage; a daf-2(e1370) animal grown for (E) 1 day at 25°C and [(F) to (H)] 2 days at 25°C; and (I) a daf-2(e1370); bec-1 (RNAi) animal grown for 2 days at 25°C. In (F), arrowheads denote representative vacuolar structures in the late stages of the autophagolysosomal pathway; the white arrowhead denotes structure that is shown at higher magnification in (H); the arrow denotes the dauer cuticle. (G) A representative early autophagosome and (H) a multilamellar body/autolysosome are shown in a seam cell of a daf-2(e1370) dauer animal. Scale bars, 1 μM [(C) to (F) and (I)]; 0.1 μM [(G) and (H)].

We performed electron microscopy to further study the process of seam cell autophagy during dauer development (Fig. 4, C to I). In wild-type L2 and L3 animals, we observed rare early autophagosomes, but no evidence of later structures in the autophagy pathway (Fig. 4, C and D). In daf-2(e1370) “pre-dauer” animals grown at 25°C for 1 day, we saw an increase in the number of multivesicular bodies (a structure of endosomal origin), but very few autophagosomes or autolysosomes (Fig. 4E). In contrast, the seam cells from daf-2(e1370) animals that entered dauer diapause after 2 days of growth at 25°C displayed marked ultrastructural differences (Fig. 4F) from those in the pre-dauer daf-2(e1370) or wild-type L2 or L3 animals. Although some early autophagosomes were detected (Fig. 4G), the most notable finding was the accumulation of vacuolar structures that appear similar to those described in the late stages of autophagy in mammalian cells (Fig. 4, F and H). Furthermore, seam cells from the daf-2(e1370); bec-1 (RNAi) abnormal dauer animals completely lacked any early or late autophagic structures (Fig. 4I). Together with the immunofluorescent studies of GFP::LGG-1 localization in daf-2(e1370) animals, these data indicate that there is active autophagy in hypodermal seam cells during dauer development that is blocked by inactivation of the bec-1 autophagy gene.

To further evaluate whether autophagy is required for daf-2(e1370) dauer formation, we tested the effects of RNAi of other C. elegans orthologs of yeast autophagy genes on daf-2(e1370) dauer formation (Fig. 5). In addition, because the yeast ortholog, APG6/VPS30, of bec-1 is involved in both vacuolar protein sorting and autophagy, we examined the RNAi phenotypes of two VPS genes, VPS5 and VPS35, that have no documented role in autophagy. RNAi-mediated interference of the C. elegans orthologs of yeast APG4, APG17, and the C. elegans lgg-2 ortholog of yeast APG8, as well as of the C. elegans orthologs of yeast VPS5 and VPS35, had minimal effects on dauer morphogenesis in daf-2(e1370) mutant animals (Fig. 5A) and did not alter the autophagosomal punctate staining pattern in the seam cells of GFP::LGG-1 dauer animals (Fig. 5B). In contrast, RNAi inactivation of the C. elegans orthologs of yeast APG1 (unc-51), yeast APG7, yeast AUT10, and the C. elegans lgg-1 ortholog of yeast APG8/AUT7, as well as a loss-of-function mutation in unc-51 [unc-51(e369)] (32), resulted in the formation of abnormal daf-2(e1370) dauers (Fig. 5A). These abnormal dauers displayed characteristics similar to those described for daf-2(e1370); bec-1 (RNAi) animals and also had aberrant localization of GFP::LGG-1 (Fig. 5B) (33). Thus, several C. elegans orthologs of yeast and mammalian APG genes, including APG1, APG6/VPS30/beclin1, APG7, APG8, and AUT10, are required for normal dauer morphogenesis in C. elegans.

Fig. 5.

The effect of apg and vps gene inactivation on dauer formation and autophagy in daf-2(e1370) mutant animals at 25°C. (A) The percentage of abnormal dauers observed with different C. elegans gene inactivations. Black bars correspond to gene inactivations that resulted in ≥20% abnormal dauers, and white bars correspond to gene inactivations that resulted in <20% abnormal dauers (n = 296 to 1840 animals per experimental group). Asterisks denote that undiluted RNAi is lethal in the N2 background and that data are shown for a 1:1 dilution of RNAi that is not lethal in the N2 background [see (19)]. (B) Representative photomicrographs of GFP::LGG-1 stainingin seam cells from daf-2(e1370) animals treated with RNAi as indicated.

Our findings demonstrate that autophagy genes are required for normal dauer morphogenesis and life-span extension in C. elegans. In the unicellular organism Saccharomyces cerevisiae, autophagy is essential for survival during starvation and for sporulation and differentiation (21). We propose that autophagy is also essential for the cellular and tissue remodeling that permits multicellular organisms to successfully adapt to environmental stress. Furthermore, our results suggest that autophagy functions in multicellular organisms not only in development and in stress responses but also in cellular processes that regulate life-span during nonstressed conditions. Thus, the degradation of cellular contents by autophagy may underlie, in part, the diverse metabolic, developmental, and life-span processes controlled by the insulin-like signaling pathway in nematodes and, possibly, in higher eukaryotic organisms.

Supporting Online Material

Materials and Methods

Fig. S1

Table S1

References and Notes

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article