Regulation of Dauer Formation by O-GlcNAcylation in Caenorhabditis elegans

Abstract

Modification of proteins at serine or threonine residues with N-acetylglucosamine, termed O-GlcNAcylation, plays an important role in most eukaryotic cells. To understand the molecular mechanism by which O-GlcNAcylation regulates the entry of Caenorhabditis elegans into the non-aging dauer state, we performed proteomic studies using two mutant strains: the O-GlcNAc transferase-deficient ogt-1(ok430) strain and the O-GlcNAcase-defective oga-1(ok1207) strain. In the presence of the dauer pheromone daumone, ogt-1 showed suppression of dauer formation, whereas oga-1 exhibited enhancement of dauer formation. Consistent with these findings, treatment of wild-type N2 worms with low concentrations of daumone and the O-GlcNAcase inhibitor O-(2-acetamido-2-deoxy-d-glucopyranosylidene)amino-N-phenylcarbamate (PUGNAc) enhanced dauer formation, which was dependent on intact O-GlcNAcylation metabolism. We also found that the treatment of daumone enhanced O-GlcNAcylation in vivo. Seven proteins, identified by coupled two-dimensional electrophoresis/liquid chromatography-mass spectroscopy (LC-MS) analysis, were differentially expressed in oga-1(ok1207) worms compared with wild-type N2 worms. The identities of these proteins suggest that O- GlcNAcylation influences stress resistance, protein folding, and mitochondrial function. Using O-GlcNAc labeling with fluorescent dye combined with two-dimensional electrophoresis/LC-MS analysis, we also identified five proteins that were differentially O-GlcNAcylated during dauer formation. Analysis of these candidate O-GlcNAcylated proteins suggests that O-GlcNAcylation may regulate cytoskeleton modifications and protein turnover during dauer formation.

Introduction

Caenorhabditis elegans is a small, free-living soil nematode that has been used as a model system for decades; many known molecular mechanisms have been discovered based on C. elegans research. The standard reproductive life cycle of C. elegans is 3 days under normal conditions and comprises an egg stage, four larval stages, and the adult stage. C. elegans may also enter an alternative diapause stage, becoming what is termed dauer larvae. C. elegans development is influenced by the concentration of the dauer pheromone, available food supply, and temperature (1, 2). These three factors collectively determine whether a developing larva enters dauer stage or continues to grow to an adult. Genetic research of dauer formation has identified a number of genes required for dauer formation and has provided a window into the process by which C. elegans senses external environmental stimuli and adapt its developmental processes accordingly. The regulation of dauer entry has been studied extensively and shown to require a number of genetically distinct mechanisms, including insulin-like signaling, transforming growth factor β-like signaling, and steroid hormone receptor signaling (3–6). These signaling pathways link various protein kinases through phosphorylation-dependent protein-protein interactions (4, 7). Membrane receptors for insulin-like and transforming growth factor β-like signaling contain kinase domains that are critical for their activities. The expression and activity of steroid receptors are also highly dependent on post-translational modifications including phosphorylation (8). Thus, various post-translational modifications, but especially phosphorylation, appear to be important for the proper fine-tuning of signal transduction during dauer formation.

Recent attention has focused on the role of a different post-translational modification, O-GlcNAcylation, in the dauer entry process. O-GlcNAcylation is an intracellular protein modification that links a single N-acetylyglucosamine moiety to serine and threonine residues in various proteins (9). Two enzymes regulate the balance of O-GlcNAcylation: O-GlcNAc transferase (OGT)² and O-GlcNAcase (OGA) (10). OGT contains a hexosyltransferase domain, which mediates the addition of the O-GlcNAc moiety to target proteins, and an N-terminal tetratricopeptide repeat domain. OGA contains a hexominidase domain, which is responsible for removing O-GlcNAc from the target proteins, and a C-terminal acetyltransferase domain. The role of O-GlcNAcylation varies depending on the specific protein modified. O-GlcNAcylation regulates diverse physiological changes mediated by a long list of identified target proteins, including glycolytic proteins, cytoskeleton proteins, nuclear pore proteins, protein-folding machinery proteins, transcription factors, and signal transduction proteins (10). Interestingly, O-GlcNAcylation sites are also typically targets of phosphorylation. Once such a site is modified by O-GlcNAc, O-GlcNAcylation inhibits protein phosphorylation. Thus, in many cases, O-GlcNAcylation acts as an antagonist of serine/threonine phosphorylation. However, there are also cases in which both modifications act synergistically (10, 11).

Traditionally, radioactive hexose has been used to detect O-GlcNAcylation. Although anti-O-GlcNAc antibodies are now widely used for this purpose, the sensitivity of anti-O- GlcNAc antibodies is limited, and O-GlcNAcylated sites tend to be labile. Recently, a new enzymatic O-GlcNAc labeling technique has been developed to detect O-GlcNAcylated proteins (12). This technique exploits the selectivity of the galactosyltransferase (GalT) enzyme toward free N-acetylglucosamine. Using modified GalT(Y289L), the synthetic substrate azidogalactose is linked to free N-acetylglucosamine, and then either a fluorescent dye or biotin tag is added to the azido functional group. This technique facilitates the detection of O- GlcNAcylation and has become a routine assay method.

Because various post-translational modifications have been implicated in the regulation of the dauer process, it is reasonable to suppose that O-GlcNAcylation might be among the post-translational modifications that contribute to dauer formation. In this regard, Hanover et al. (13) reported that dauer formation is abnormal in C. elegans O-GlcNAc-cycling mutants. The C. elegans genome contains a single O-GlcNAc transferase (ogt-1) and one O-GlcNAcase (oga-1) (14). In a daf-2 mutant background, oga-1(ok1207) mutants enter dauer more easily than wild-type N2 worms, whereas ogt-1(ok430) mutants exhibit a dauer-defective phenotype (14). These data strongly implicate O-GlcNAcylation in dauer formation, although the detailed mechanisms involved have remained elusive. In this study, we used the enzymatic O-GlcNAc labeling technique along with oga-1(ok1207) and ogt-1(ok430) mutant and a synthetic dauer pheromone, daumone (15), to investigate the mechanism by which O-GlcNAcylation regulates the dauer formation process. In our proteomic approaches, we addressed the following additional questions. 1) Does O-GlcNAcylation affect dauer formation? 2) If so, at what step during dauer formation does O-GlcNAcylation act? 3) What kinds of proteins are modified by O-GlcNAcylation? 4) What are the physiological roles of O-GlcNAcylation during dauer formation? Here, we report a proteomic view of O-GlcNAcylation-mediated formation of the non-aging dauer state in C. elegans.

EXPERIMENTAL PROCEDURES

Strains and General Handling

Methods used for the maintenance and handling of C. elegans were as described previously (16). The N2 strain was used as the wild-type control. All strains were maintained at 20 °C on nematode growth medium (NGM) agar plates containing Escherichia coli (OP50 strain). The following mutant strains, obtained from the Caenorhabditis Genetics Center (Minneapolis, MN), were used in this study: daf-16(mu86), ogt-1(ok430), daf-2(e1370), F54H12.6(ok2133), unc-60(e723), oga-1(ok1207), and daf-12(rh273).

Chemicals

All chemicals used, regardless of the commercial source, were of the highest grade available. Synthetic daumone was obtained as described previously (15). O-(2-Acetamido-2-deoxy-d-glucopyranosylidene)amino-N-phenylcarbamate (PUGNAc) was purchased from Toronto Research Chemicals. The Click-iT™ O-GlcNAc enzymatic labeling system, Click-iT™ protein analysis detection kit with TAMRA alkyne, and Click-iT™ protein analysis detection kit with biotin alkyne were purchased from Invitrogen. Immobiline immobilized pH gradient (IPG) DryStrips (pH 3–10 nonlinear 18 cm) were purchased from Amersham Biosciences. NANOSEP MF (0.45 μm), and Acrodisc LC syringe filters (0.2 μm) were purchased from Pall Life Sciences. The Quick Start™ Bradford protein assay kit, Coomassie Brilliant Blue G-250, and N-N-N′-N′-tetramethylethylenediamine (TEMED) were from Bio-Rad. Griess reagent, CHAPS, urea, dithiothreitol, Tris base, thiourea, glycine, ammonium persulfate, SDS, and other basic chemicals were purchased from Sigma.

Standard and Suboptimal Daumone Assays

Standard daumone assays for measuring dauer induction were performed as described previously with minor modifications (15). Briefly, daumone-containing plates (5-cm diameter) were prepared with 3 ml of NGM (without peptone) and 320 μg of daumone. Once the plates had dried, 160 μg of dead E. coli that had been heated at 95 °C for 30 min (with vigorous vortexing every 5 min) was added to the center of each plate. Ten adult nematode worms were then placed on the plate and incubated at 20 °C for 4–6 h. After incubation, the adult worms were removed, and the eggs on the plate (generally 100–150 eggs/plate) were incubated further at 25 °C for 72 h and then counted. A 1% SDS solution (1.0 ml) was added to the plate, and the worms that survived were counted as dauer larvae. For suboptimal daumone assays, NGM plates were prepared by adding one-tenth of the standard amount of daumone (i.e. 32 μg) to each plate. For PUGNAc treatment in either standard or suboptimal daumone assays, a stock solution of PUGNAc (50 mm PUGNAc in H₂O:EtOH (1:1)) was added to NGM plates at the same time that daumone was added. These plates, containing a final PUGNAc concentration of 0.5 mm, were used for subsequent daumone assays.

Measuring Dauer Formation in PUGNAc-treated Dauer Formation Constitutive (Daf-c) Mutants

In these assays, NGM plates containing a final PUGNAc concentration of 0.5 mm were used. Ten adult worms with the indicated genotypes were placed on each plate and incubated for 4–6 h. After incubation, adult worms were removed and the eggs on the plate (generally 100–150 eggs/plate) were incubated at 22.5 °C for 72–96 h. The dauer formation rate was determined by counting dauer larvae as described above.

daf-16::GFP Nuclear Translocation

Eggs of the daf-16::GFP strain (TJ356) were placed on a 0.5 mm PUGNAc plate and incubated for 2.5 days. Young adult worms were analyzed for the subcellular localization of intestinal nuclear GFP by fluorescence dissection microscopy. More than 50 worms from PUGNAc-treated and control groups were analyzed. The same experiments were also performed in the presence of daumone. The individual worm was scored as positive if the intestinal nuclear shapes were distinguishable. Unless otherwise specified, three independent experiments were performed and used for data processing.

Western Blot Analysis

Cell lysates were prepared by sonication (five times for 10 s each) with the indicated samples. For anti-O-GlcNAc Western blot, lysates were treated with peptide:N-glycosidase F for 1 h at 37 °C. The samples were dissolved in fresh SDS sample buffer and boiled for 5 min. Samples were separated by SDS-PAGE and analyzed by Western blotting as described elsewhere (17) using the antibody dilutions recommended by the manufacturer. Anti-FTT-2 (sc-1657) and anti-DAF-16 (ce-300) antibodies were purchased from Santa Cruz Biotechnology, and the anti-actin antibody (A4700) was purchased from Sigma. Anti-O-GlcNAc antibody (RL2) and anti-MAPK antibody were purchased from Abcam (Cambridge, MA). DAF-16 band intensities were determined densitometrically and analyzed using Melanie 7.0 software. The band intensity values were normalized by subtracting the intensity of the DAF-16 lane.

Two-dimensional Electrophoresis

For analysis in the first dimension, 1 mg of protein was electrofocused on immobilized pH gradient strips (pH 3–10 nonlinear) as described (18, 19) except that a total of 80,000 volt hour was applied. For separation in the second dimension, isoelectrically focused strips were electrophoresed on 9–16% gradient polyacrylamide gels until the dye front reached the lower end of the gel. The relative abundance of protein spots was quantified by staining the gels with Coomassie Brilliant Blue G-250 followed by scanning them with a GS-710 imaging densitometer (Bio-Rad) and analysis with Melanie 7 image analysis software (GE Healthcare). Labeled images were uniformly processed using Adobe Photoshop (version 7.0) software.

Tagging of O-GlcNAcylated Proteins with TAMRA

Dauer worms were prepared by collecting eggs from gravid adults as described (20) and plating them on standard daumone plates. After incubation for 3 days, worms were harvested and lysed with buffer L (50 mm Tris-HCl, pH 7.5, 3 mm streptozotocin (STZ), and protease inhibitor mixture). The worm lysates were centrifuged to remove cellular debris, and the concentration of proteins was determined using the Quick Start Bradford reagent (Bio-Rad). An equal amount of protein was treated with peptide:N-glycosidase F for 3 h. O-GlcNAcylated proteins were tagged according to the manufacturer's protocol (Invitrogen). Briefly, 1 mg of protein extract was mixed into a solution containing 5 mm MnCl₂, 1.25 mm ADP, 0.5 mm synthetic UDP substrate, and GalT(Y289L) (25 ng/μl) and incubated for 12–14 h at 4 °C. After enzymatic labeling, extracts were dialyzed three times (2 h each) against denaturing buffer (5 m urea, 50 mm NH₄HCO₃, and 100 mm NaCl, pH 7.8). The dialysate was adjusted to 50 mm NaOAc, pH 4.8, by adding the appropriate volume of a 2.7 m NaOAc solution, pH 3.9. Either TAMRA alkyne or biotin alkyne was then added to the reaction mixture, which was then incubated overnight. Protein extracts were adjusted to a final concentration of 50 mm, pH 8, with 3 m NH₄HCO₃, pH 9.6, dialyzed once (for 2 h) against a solution containing 6 m urea, 50 mm NH₄HCO₃, pH 7.8, and 100 mm NaCl, and then dialyzed a second time (for 10 h) against a low salt version (10 mm NaCl) of the same buffer. The TAMRA-labeled samples were separated by two-dimensional electrophoresis as described above, and the resulting images were scanned using the Typhoon system (GE Healthcare). Matched spots were analyzed using Melanie 7 image analysis software (GE Healthcare), and protein spots of interest were removed from the gel and analyzed by LC-MS.

Protein Identification by Mass Spectrometry

Nano-LC-MS/MS analysis was performed on an Agilent 1100 series nano-LC and LTQ mass spectrometer (Thermo Fisher Scientific). The capillary column used for LC-MS/MS analysis (150 × 0.075 mm) was obtained from Proxeon (Odense M), and the Magic C18 stationary phase (5-μm beads/100-Å pore size) was packed in-house (Michrom Bioresources). The mobile phase A for LC separation was 0.1% formic acid in deionized water, and the mobile phase B was 0.1% formic acid in acetonitrile (21). The chromatography gradient was programmed to yield a linear increase from 5% B to 40% B in 50 min, 40% B to 60% B in 20 min, and 60% B to 80% B in 5 min. The flow rate was maintained at 300 nL/min after splitting. Mass spectra were acquired using data-dependent acquisition with full mass scan (400–1800 m/z) followed by MS/MS scans. Each MS/MS scan acquired was an average of 1 microscan on the quadrupole linear ion trap (LTQ) (22). The temperature of the ion transfer tube was controlled at 200 °C, and the spray was 1.5.0–2.0 kV. The normalized collision energy was set at 35% for MS/MS. MASCOT software was used to identify peptide sequences. Peptides that were oxidized at methionine residues and carboxyamidomethylated and carboxymethylated at cysteine residues were considered during the MASCOT search.

Gene Ontology (GO) Categorization

Gene ontology annotations for each gene were obtained from UniProt Knowledgebase (release 14.6) (23). Annotations were processed manually for the categorizations.

RESULTS

Daumone Assays Show That the Level of O-GlcNAcylation May Regulate Dauer Formation

Previous reports have shown that dauer formation at 20 and 25 °C is enhanced in oga-1;daf-2 double mutants and decreased in ogt-1;daf-2 double mutants (14). To analyze dauer formation quantitatively in these two mutants, we performed both standard (15) and suboptimal daumone assays (see “Experimental Procedures”). In a standard daumone assay, ogt-1(ok430) exhibited suppression of dauer formation as predicted, whereas oga-1(ok1207) underwent dauer formation normally (Fig. 1A). A suboptimal daumone assay using one-tenth of the normal amount of daumone (i.e. 38 μm daumone/plate) showed that dauer formation in the oga-1 mutant (67.2%) was enhanced (∼2-fold) compared with wild-type N2 worms (34.3% (Fig. 1A)). These observations confirm that enhanced O-GlcNAcylation is likely to promote dauer formation, whereas lack of it severely impairs the dauer formation process.

FIGURE 1.

O-GlcNAcylation regulates the dauer formation process. A, daumone assays with various O-GlcNAc-cycling mutants. Standard and suboptimal daumone assays were performed as described under “Experimental Procedures.” The percentages of dauers formed in suboptimal (white bars) and standard daumone assays (black bars) are indicated. Three independent plates were counted for each genotype. All asterisks indicate p < 0.001 (Student's t test). Error bars show S.D. B, PUGNAc enhances dauer formation in suboptimal daumone assays. Suboptimal daumone assays were performed with the indicated genotypes in the presence or absence of PUGNAc as described under “Experimental Procedures.” The percentages of dauers formed are indicated. Three independent plates were counted for each genotype. Error bars show S.D. C, Western blot (WB) analysis with O-GlcNAc-specific antibody. To measure the level of O-GlcNAcylation, Western blot analysis with the anti-O-GlcNAc antibody RL2 was performed with various samples. The indicated genotypes were used, and PUGNAc and daumone were applied as indicated. Anti-actin antibody was used as a control. D, PUGNAc enhances dauer formation in Daf-c mutants. Dauer formation at 22.5 °C was observed in the presence or absence of PUGNAc. The indicated genotypes were incubated for 72 h, and the number of dauer larvae formed was counted. Dauer formation percentages are indicated. Three independent plates were counted for each genotype. Error bars show S.D.

Inhibition of O-GlcNAcase Increases the Frequency of Dauer Formation

Because O-GlcNAc-cycling mutants were previously shown to be dauer formation abnormal (14), we sought to determine the relationship between O-GlcNAcylation and dauer formation. Given that OGA-1 may have functions distinct from its O-GlcNAcase activity, we wanted to confirm that the dauer formation abnormal phenotype was due to O-GlcNAcase activity. To this end, we employed PUGNAc (24), an inhibitor of O-GlcNAcase that has been used to induce insulin resistance and is known to block sperm penetration during fertilization in mammals (25). To determine how PUGNAc influences dauer formation in C. elegans, we performed suboptimal daumone assays in the presence or absence of 0.5 mm PUGNAc. The growth rate and morphology of worms treated with PUGNAc alone appeared to be normal (data not shown). However, dauer formation under suboptimal daumone conditions was increased 1.86-fold to 75.6% in the presence of PUGNAc (Fig. 1B). The levels of dauer formation in oga-1 and PUGNAC-treated worms were comparable, suggesting that the biochemical effects of null oga-1 mutant and PUGNAc treatment are likely to be similar in vivo. Also, ogt-1 and oga-1 mutants are resistant against PUGNAc treatment, unlike the wild type. The data imply that the enhancement of dauer formation is dependent on O-GlcNAc and that enhanced O-GlcNAcylation causes more dauer formation.

Western Blot Analysis Shows That Daumone Enhances O-GlcNAcylation

To measure the level of O- GlcNAc in vivo, we performed anti-O-GlcNAc Western blot analysis (Fig. 1C). The level of O-GlcNAc was almost undetectable in ogt-1 but was enhanced in oga-1 as predicted. The treatment of PUGNAc also enhanced the level of O- GlcNAc up to that of the oga-1 null mutant, supporting the idea that both PUGNAc and the oga-1 null mutant may have similar effects in vivo. Interestingly, when we added daumone, the level of O-GlcNAc was slightly enhanced (Fig. 1C, lane 3). This suggests that treatment with daumone elicits the enhancement of O-GlcNAcylation, which may be the major mechanism by which signal transduction can be regulated during dauer formation.

Inhibition of O-GlcNAcase Enhances the Frequency of Dauer Formation of Daf-c Mutants

We next tested whether inhibition of O-GlcNAcase also enhances dauer formation in Daf-c mutants by treating daf-2(e1370) and daf-12(rh273) with PUGNAc and measuring the changes in dauer formation at 22.5 °C. In the absence of PUGNAc treatment, these mutants usually showed a partial Daf-c phenotype (37.7 and 15.9% for daf-2(e1370) and daf-12(rh273), respectively) at this temperature (Fig. 1D). Dauer formation was significantly enhanced by PUGNAc treatment, which increased the frequency of dauer larvae by 1.9-fold (to 71.6%) and 3.5-fold (to 56.5%) in daf-2(e1370) and daf-12(rh273) mutants, respectively (Fig. 1D). Because daf-12 is genetically located far downstream in the dauer formation pathway (26), these data suggest that O-GlcNAcylation may regulate multiple points during dauer formation.

Inhibition of O-GlcNAcase Does Not Induce DAF-16 Nuclear Translocation

In mammalian systems, high dose PUGNAc treatment induces diabetes-like symptoms, suggesting that O-GlcNAcylation regulates the insulin/IGF-1-signaling pathway (27). Given that dauer formation is also regulated by the IGF-1-signaling pathway, it would be reasonable to predict that treatment with PUGNAc might also act through the IGF-1-signaling pathway to regulate dauer formation in C. elegans. To test the capacity of PUGNAc to activate the IGF-1-signaling pathway, we monitored the activity of the forkhead (FOXO) transcription factor DAF-16 (which is regulated by the IGF-1-signaling pathway) using daf-16::GFP transgenic worms that had been used previously for this purpose (28, 29). When placed on 0.5 mm PUGNAc, these transgenic worms showed no change in the nuclear localization of DAF-16::GFP (Fig. 2A). These results are consistent with a previous report showing that other transcription factors do not accumulate in the nucleus in an oga-1 background (14). Surprisingly, even under the slightly activated condition of suboptimal daumone treatment, DAF-16::GFP showed no change in nuclear localization in the presence of PUGNAc (Fig. 2A). Thus, although PUGNAc is able to potentiate the dauer formation process, it does not induce DAF-16 nuclear translocation, suggesting that the ability of PUGNAc treatment to activate dauer formation may be mediated by other targets of O-GlcNAcylation.

FIGURE 2.

PUGNAc treatment is not sufficient for DAF-16 nuclear translocation. A, DAF-16 nuclear translocation with PUGNAc treatment. The indicated strains of transgenic daf-16::GFP worms were observed by fluorescence microscopy after incubation for 72 h in the presence or absence of PUGNAc. Also the same experiments were performed in the presence of daumone. The individual worm is scored as positive if the intestinal nuclear shapes are distinguishable. Three independent experiments were performed. Error bars show S.D. B, Western blot (WB) analysis with O-GlcNAc-cycling mutants. Lysates from the indicated genotypes were analyzed by Western blotting using an anti-DAF-16 antibody. Anti-actin Western blot was used as a control. Normalized band intensities are shown in the lower panel. Three independent experiments were performed. Error bars show S.D.

One possibility is that PUGNAc treatment regulates the expression of DAF-16 itself, thereby stimulating the transcription of downstream target genes. To address this possibility, we analyzed the expression level of DAF-16 in the O-GlcNAc-cycling mutants oga-1(ok1207) and ogt-1(ok430) using Western blotting. In ogt-1 (deficient in O-GlcNAcylation), DAF-16 protein levels were slightly decreased (Fig. 2B). Although the mechanism for this effect is currently unknown, this observation may partially explain the fact that ogt-1 mutants show suppression of dauer formation. In contrast, oga-1 (enhanced O-GlcNAcylation) had little effect on the expression of DAF-16 (Fig. 2B). Taken together, these results suggest that modulation of the O-GlcNAcylation status is not directly linked to DAF-16 nuclear translocation.

Enhanced O-GlcNAcylation Causes Differential Changes in Protein Expression

Using Western blot analysis, both we and Forsythe et al. (14) had demonstrated that oga-1 mutants have enhanced O-GlcNAcylation, an observation that is consistent with the completely null mutation predicted on the basis of the molecular structure. To understand what kinds of proteomic changes are associated with enhanced O-GlcNAcylation, we analyzed protein extracts from wild-type N2 and oga-1(ok1207) mutant worms for differentially expressed proteins using two-dimensional electrophoresis. Representative two-dimensional gels and individual spots are shown in Fig. 3. These seven proteins corresponded to F15E11.12 and F15E11.13, which were increased in oga-1, and to DAF-21 (HSP-90), F₀F₁-type ATP synthase α subunit, NADH-ubiquinone oxidoreductase, ribosomal subunit L7a, and cathepsin L-like cysteine protease, all of which were decreased in oga-1 (see Table 1).

FIGURE 3.

The oga-1 mutant, which has enhanced O-GlcNAcylation, induces proteomic changes. A, two-dimensional electrophoresis with wild-type N2 and oga-1 mutant worms. Lysates from the indicated genotypes were resolved by two-dimensional electrophoresis. The resolved gels were stained with Coomassie Brilliant Blue. A representative image from two independent experiments is shown. Black arrows indicate spots that were differentially expressed by more than 2.5-fold and selected for further analysis. Molecular mass and pI values are also indicated. The identities of spots are shown in Table 1. B, various spots are differentially expressed in oga-1 mutant. Identities for each selected spot (i.e. those differentially expressed by more than 2.5-fold) are indicated. Spot boundaries used for data processing in Melanie 7.0 software are indicated. The -fold changes in differential expression are shown in Table 1.

TABLE 1.

Identified proteins differentially expressed between N2 and oga-1(ok1207)

Spot ID	Protein name	Function	NCBI accession no.	-Fold changea	No. of matched peptides	% Sequence coverage	MASCOT score
1	F15E11.12	Unknown (nematode-specific protein)	gi 25396294	3.13↑	33	32	358
2	F15E11.13	Unknown (nematode-specific protein)	gi 17559734	3.70↑	52	47	209
3	DAF-21	HSP90	gi 17559162	4.90↓	25	29	680
4	RPL-7A	Ribosomal subunit L7a	gi 32565827	2.67↓	7	20	224
5	CPL-1	Cathepsin L-like cysteine protease	gi 17563798	3.05↓	36	28	318
6	H28O16.1a	F0F1-type ATP synthase	gi 71988063	3.04↓	77	34	749
7	NUO-2	NADH-ubiquinone oxidoreductase	gi 32563621	2.62↓	17	32	282

^a Average -fold change from two independent experiments.

Multiple Proteins Are Targets of O-GlcNAcylation in C. elegans

Enhanced O-GlcNAcylation appears to promote dauer formation. However, it is not clear what common mechanism might link multiple O-GlcNAcylated targets to cooperatively regulate this process. To understand the underlying molecular mechanism, we analyzed proteins that were O- GlcNAcylated during dauer formation. O-GlcNAcylated proteins in daumone-treated and nontreated protein samples were labeled enzymatically as described under “Experimental Procedures” and tagged specifically with TAMRA fluorescent dyes (Fig. 4A). After separating the labeled samples by two-dimensional electrophoresis, a number of TAMRA-positive spots were detected, from which we selected the 20 most strongly positive spots in the daumone-treated samples (Fig. 4B). These TAMRA-positive spots, subsequently identified by LC-MS, are listed in Table 2. These O-GlcNAcylated proteins can be classified into various functional categories based on GO annotations (Fig. 4C). A majority of them are cytoplasmic proteins (Fig. 4C, first pie graph), which is consistent with a previous report that O-GlcNAcylation occurs in the cytosolic compartment. As categorized by biological process, a majority of the annotated proteins were also associated with development (Fig. 4C, third pie graph), suggesting that O-GlcNAcylation might be involved in developmental processes other than dauer formation. An analysis of O-GlcNAcylated proteins based on their respective RNAi phenotypes (obtained from WormBase) showed that the embryonic lethal (Emb), larval lethal (Lvl), sick (Sck), and sterile (Ste) phenotypes were more frequent than other phenotypes (Fig. 4D and Table 2). Consistent with the GO categorizations, some of the observed phenotypes suggest that O-GlcNAcylation is also involved in other critical cellular processes. There were a high percentage of cytoskeleton-related proteins (actin, tropomyosin, and UNC-60/cofilin) among the O-GlcNAcylated proteins, suggesting that O-GlcNAcylation might be important in cytoskeleton regulation. Interestingly, O-GlcNAcylated 14-3-3 and a galectin homolog were identified, although the functional significance of these O-GlcNAcylations is not yet clear.

FIGURE 4.

Various proteins are O-GlcNAcylated during dauer formation. A, strategy for specific tagging of O-GlcNAcylated proteins. The schematic diagram shows a summary of tagging methods. O-GlcNAc moieties and azido-containing substrates are linked by modified galactosyltransferase (GalT(Y289L)). Subsequently, the azido functional group is fluorescently tagged with TAMRA. See “Experimental Procedures” for additional details. B, various proteins are O-GlcNAcylated during dauer formation. Lysates from the indicated treatment groups were fluorescently labeled as described under “Experimental Procedures,” and labeled proteins were resolved by two-dimensional electrophoresis. A representative image from two independent experiments is shown. Black arrows indicate spots that were highly stained with TAMRA fluorescence and selected for further analysis. Molecular mass and pI values are also indicated. The identities of spots are listed in Table 2. C, O-GlcNAcylated proteins categorized by GO annotation. Gene ontology annotations, divided into the subcategories of cellular component, molecular function, and biological process, are shown in each pie graph. The identities of 20 selected spots were used for the analysis. D, O-GlcNAcylated proteins categorized by RNAi phenotypes. The pie graph shows RNAi phenotypes obtained from WormBase. The identities of 20 selected spots were used for the analysis.

TABLE 2.

Proteins identified as O-GlcNAcylated during the dauer formation process

Spot ID	NCBI accession no.	Protein name	Gene function	MASCOT score	No. of matched peptides	TAMRA -fold changea	RNAi phenotype	Vol % TAMRA intensityb
1	gi 3875041	C47E8.5	HSP-90 (daf-21)	142	7	1.96	Esp, Ste	0.35
2	gi 3875041	C47E8.5	HSP-90 (daf-21)	135	6	2.13	Esp, Ste	0.10
3	gi 6624	T04C12.6	Actin	1348	339	1.08	Ste, Pvl, Lva, Lvl, Emb, Sck	1.43
4	gi 829164	T04C12.6	Actin	1676	562	1.16	Ste, Pvl, Lva, Lvl, Emb, Sck	2.81
5	gi 829164	T04C12.6	Actin	1442	440	1.06	Ste, Pvl, Lva, Lvl, Emb, Sck	3.82
6	gi 6624	T04C12.6	Actin	1519	400	1.02	Ste, Pvl, Lva, Lvl, Emb, Sck	0.98
7	gi 9857649	LEC-5	Galectin-5	467	13	1.28		0.74
8	gi 9857649	LEC-5	Galectin-5	573	12	1.33		0.90
9	gi 1208411	Y105E8B.1e	LEV-11, tropomyosin	812	22	1.37	Egl, Slu, Ste, Lon, Sck, Lvl, Bmd, Unc, Pvl, Prl, Mlt	0.36
10	gi 1208409	Y105E8B.1d	LEV-11, tropomyosin	175	7	1.08	Egl, Slu, Ste, Lon, Sck, Lvl, Bmd, Unc, Pvl, Prl, Mlt	0.61
11	gi 409298	F54H12.6	Elongation factor-1 β	129	3	1.58	Emb, Gro, Lva	0.50
12	gi 2492485	F52D10.3	14-3-3	987	105	1.02	Ste, Age, Emb, Let, Egl, Sck, Gro, Daf	0.65
13	gi 3892154	F39H11.5	Proteasome β subunit (pbs-7)	432	11	1.35	Lva, Lvl, Emb, Unc, Ste, Ocs, Let	0.75
14	gi 3892154	F39H11.5	Proteasome β subunit (pbs-7)	176	8	2.37	Lva, Lvl, Emb, Unc, Ste, Ocs, Let	0.56
15	gi 3877919	F58G1.4	dct-18	252	7	1.07	Gro, Sck, Clr	0.52
16	gi 963095	M01F1.2	60S ribosomal protein L13A (rpl-16)	438	14	1.20	Emb, Lva	0.43
17	gi 304344	UNC-60	Cofilin/destrin homolog	453	28	1.51	Unc, Lvl, Age, Sck, Ste, Pvl	0.82
18	gi 304344	UNC-60	Cofilin/destrin homolog	255	9	1.08	Unc, Lvl, Age, Sck, Ste, Pvl	0.76
19	gi 304344	UNC-60	Cofilin/destrin homolog	571	55	1.32	Unc, Lvl, Age, Sck, Ste, Pvl	0.78
20	gi 2736418	F15E11.1	Unknown	1288	621	2.34		0.80

^a Average -fold change from two independent experiments.

^b Percentages of TAMRA intensities are measured in daumone-treated sample.

Differences in TAMRA staining between daumone-treated and untreated samples were also measured (Table 2). As expected, a small subset of TAMRA-stained proteins was differentially modified during dauer formation. Five proteins (DAF-21, PBS-7, F54H12.6, UNC-60, and F15E11.1) showed a differential modification greater than 1.5-fold; others showed no significant changes in O-GlcNAcylation. The identities of these proteins suggest that the cytoskeleton and the protein turnover pathway are targets of O-GlcNAcylation during dauer formation and may play a regulatory role in the process.

Validation of O-GlcNAcylated Proteins

To verify O- GlcNAcylation of the identified proteins, we selected candidate proteins, ACT-1 (actin) and FTT-2 (14-3-3), for which antibodies were available. We employed avidin affinity chromatography to isolate biotin-labeled proteins prepared using the enzymatic O-GlcNAc labeling system. The biotin-labeled proteins obtained from this procedure were analyzed by Western blotting using anti-ACT-1 and anti-FTT-2 antibodies. Specific bands for ACT-1 and FTT-2 were detected in the experimental group (GalT(Y289L) treatment), whereas no such bands were present in the control group (no GalT(Y289L) treatment) and non-O-GlcNAcylated MAPK (supplemental Fig. S2). Although the experiments might pull down other proteins that did not have O-GlcNAc on them but interacted with O-GlcNAcylated proteins, these data could be another piece of evidence supporting the possibility that ACT-1 and FTT-2 could be O-GlcNAcylated.

To further validate differential O-GlcNAcylation during dauer formation, we performed suboptimal daumone assays using two of the five proteins that exhibited a greater than 1.5-fold difference in O-GlcNAcylation: unc-60(e723) and F54H12.6(ok2133) (Fig. 5). The analysis was limited to these proteins because no mutants were available for two proteins (F15E11.1 and pbs-7) and the remaining protein (daf-21(p673)) exhibited a Daf-c phenotype. unc-60 showed slightly enhanced dauer formation in the suboptimal daumone plate in relation to N2 but lost PUGNAc-enhanced dauer formation. There was also a reduction in PUGNAc-induced dauer formation in the suboptimal daumone assays of F54H12.6(ok2133). These data suggest that PUGNAc does not enhance dauer formation in unc-60 and F54H12.6 mutants and imply that PUGNAc target genes could be downstream of (or parallel to) unc-60 and/or F54H12.6.

FIGURE 5.

Suboptimal daumone assays with differentially O-GlcNAcylated candidates. Suboptimal daumone assays were performed using the indicated genotypes in the presence (black bars) or absence (white bars) of PUGNAc. The percentages of dauers formed are indicated. Numbers indicate p values (Student's t test). Three independent plates were counted for each genotype. Error bars show S.D.

DISCUSSION

Emerging evidence indicates that O-GlcNAcylation is involved in dauer formation in C. elegans (14). This association is strongly supported by our findings. First, ogt-1 (O- GlcNAcylation-deficient) showed suppression of dauer formation, whereas oga-1 (elevated O-GlcNAcylation) exhibited enhancement of dauer formation in the presence of the dauer pheromone, daumone (Fig. 1A). Second, the O-GlcNAcase inhibitor PUGNAc enhanced dauer formation, which was dependent on the intact O-GlcNAcylation metabolism (Fig. 1B). Third, daumone treatment enhanced O-GlcNAcylation in vivo (Fig. 1C).

Our study also provides insight into the molecular mechanism by which O-GlcNAcylation modulates the dauer formation process, identifying proteins that are differentially expressed in the oga-1 mutant and proteins that are differentially O-GlcNAcylated in association with dauer formation. Notably, our suboptimal daumone assay was capable of detecting subtle changes in dauer formation that might otherwise have been obscured, allowing us to quantitatively analyze dauer formation (Fig. 1).

Previous studies demonstrating the importance of active cycling of O-GlcNAc for proper dauer formation implicate the IGF-1-signaling pathway in the process (14). Our demonstration that treatment with the O-GlcNAcase inhibitor PUGNAc enhanced dauer formation in daf-2 (insulin-like receptor tyrosine kinase-deficient) is consistent with a role for this pathway and supports the idea that the level of O-GlcNAcylation is important for proper dauer formation. Another downstream mediator of this pathway in the C. elegans dauer formation process is the transcription factor DAF-16, activation of which is absolutely required for dauer formation. Although treatment with PUGNAc did not induce DAF-16 nuclear translocation, as determined by our study, it is likely that enhanced O-GlcNAcylation of target proteins may potentiate daf-16 activation and subsequent dauer formation via various mechanisms. Because daf-12 (defective for a nuclear steroid hormone receptor component downstream of DAF-2) is located at a downstream position in the dauer formation pathway, we speculate that multiple O-GlcNAcylations involving genes parallel to or downstream of daf-12 might contribute to the regulation of the dauer formation process.

Candidate alternative O-GlcNAcylated proteins that might contribute to daf-16 activation during dauer formation, some of which have been identified previously as O-GlcNAcylated proteins (30), were identified in our proteomic analysis (Table 2). These studies, conducted in a different model system using O-GlcNAc-specific antibodies to immunopurify O-GlcNAcylated proteins, identified some of the same protein groups and thus strengthen out data. Among the proteins identified in our study is FTT-2, the C. elegans homolog of 14-3-3, which is known to bind various phosphorylated proteins. Previous reports have shown that mutation of ftt-2 induces the dauer formation process by activating DAF-16 (31). Because daf-16 nuclear translocation was not detected upon PUGNAc treatment in our study, the role of O-GlcNAcylated FTT-2 remains uncertain. Several other candidate target proteins, namely actin, tropomyosin, and UNC-60/cofilin, are involved in cytoskeleton structure formation, suggesting that cytoskeleton function could be regulated by O-GlcNAcylation. In particular, in our daumone assay (Fig. 5) unc-60 showed the enhanced dauer formation. unc-60, a homolog of cofilin, is a family of actin-binding proteins. Cofilin causes depolymerization of the actin filament, and the released actin monomers are recycled for their reassembly. Therefore, cofilin plays an important role during the reorganization of actin filaments and helps to maintain the monomeric actin pool for sustained motility (32). In addition to reorganizing the actin filaments, cofilin is also involved in reorganizing microtubules in the neuronal growth cone (33). Thus, the activation and inactivation of cofilin can regulate actin dynamics and alter cell morphology and neuronal outgrowth (34). We speculate that the dynamic nature of the neuronal process is decreased in the absence of cofilin, as in the case of the unc-60 mutant, which might affect the plasticity of amphid neurons that are important for responding to the surrounding environment. Because the daf-19 mutant, which lacks cilia, shows a constitutive dauer formation (35), we speculated that unc-60 may have a partial cilia defect, which causes the slightly enhanced dauer formation, and that O-GlcNAcylation of UNC-60 might regulate the cilia formation process. In addition, UNC-60/cofilin is the differentially O-GlcNAcylated protein identified in this study. Such changes during dauer formation suggest that O-GlcNAcylation could regulate the overall shift in the cytoskeletal structure that occurs during dauer formation, such as that associated with sensory neuron retraction and/or radial shrinkage. Among the five identified differentially O-GlcNAcylated candidates that might play important roles during dauer formation was daf-21 (HSP-90 homolog), which, when mutated, is known to induce constitutive dauer formation. In addition to being differentially O-GlcNAcylated during dauer formation, DAF-21 levels were decreased in the oga-1 mutants (Fig. 3). It is thus possible that the increase in DAF-21 O-GlcNAcylation might induce its own degradation and contribute to the subsequent enhancement of dauer formation. Other differentially O-GlcNAcylated proteins identified include the proteasome subunit pbs-7, which is involved in protein turnover and may regulate selective protein degradation, and F54H12.6, which is an elongation factor-1 β/δ chain homolog and regulates the translation process. Collectively, analysis of these differentially O-GlcNAcylated candidates suggests that O-GlcNAcylation may regulate cytoskeletal modification and protein turnover during dauer formation.

We also identified seven proteins in which expression was differentially regulated in oga-1 mutants. The expression of F15E11.12 and F15E11.13, which are both unknown proteins without identified homologs in other taxa, were increased in oga-1. These two proteins likely serve similar functions based on their high degree of amino acid homology; because their RNAi data are associated with slow growth and cadmium hypersensitivity, they are possibly involved in stress responsiveness (36). The daf-2 mutant also shows high expression of these proteins under stress conditions (37); therefore, oga-1 may partially phenocopy daf-2. The expression of DAF-21 (HSP-90), which is required for larval development and is a negative regulator of dauer formation, is decreased in oga-1. A mutation in daf-21 results in constitutive dauer formation, leading us to speculate that the decrease in DAF-21 expression in oga-1 could facilitate dauer entry. Two of the candidate proteins in which expression is also decreased in oga-1, F₀F₁-type ATP synthase α subunit and NADH-ubiquinone oxidoreductase, are mitochondrial proteins. Previous reports have shown that decreases in these proteins reduce mitochondrial function and induce a life extension pathway (38). Thus, mitochondrial activity is expected to be lower in oga-1, although oga-1 does not show the life extension phenotype predicted by this decrease (supplemental Fig. S1). The remaining differentially expressed proteins, ribosomal subunit L7a and cathepsin L-like cysteine protease, are involved in protein synthesis and degradation. Decreases in the expression of these proteins suggest that protein turnover could be suppressed in oga-1 mutants. Taken together, the properties of the identified proteins, some of which correspond to previously reported products of daf-2-responsive genes (37), help to explain why oga-1 exhibited the enhancement of dauer formation. On the basis of these changes, we propose that oga-1 mutants may mimic many aspects of daf-2 mutants, which exhibit long life and have a high tendency to form dauer larvae.

In mammals, O-GlcNAcylation has been linked to diabetes, cancer, and neurodegeneration (see Ref. 10 for details). One of the most studied pathways in this context is the insulin-signaling pathway. In fact, the O-GlcNAcase inhibitors PUGNAc and STZ have been used to mimic diabetes symptoms in model animals (39, 40). Although both PUGNAc and STZ inhibit OGA-1 activity (39, 40), STZ caused additional side effects in our assays (data not shown), implying that the mechanism of action of these two agents may be quite different. In addition to its involvement of the insulin-signaling pathway, PUGNAc is also known to inhibit neuronal process formation (38), implying that O-GlcNAcylation could regulate this process.

Finally, we considered the physiological roles of O- GlcNAcylation (Fig. 6). Because the treatment of daumone enhances O-GlcNAcylation, enhanced O-GlcNAcylation may be a key modulator during dauer formation. Many components of the IGF-1 pathway have been identified as O-GlcNAcylated, including AKT, FOXO, and IRS, and these modified proteins are important in regulating IGF-1 signal transduction. Because the IGF-1-signaling pathway regulates dauer formation, O- GlcNAcylation might be expected to regulate the dauer formation process by modulating the IGF-1 signal. However, our gel-based analysis did not detect O-GlcNAcylated IGF-1 signaling intermediates. Although PUGNAc potentiated the action of daumone, treatment with PUGNAc did not appear to enhance DAF-16 nuclear translocation, suggesting that O-GlcNAcylation is not sufficient for the regulation of the IGF-1 pathway, at least under our detection conditions. Another possibility is that O-GlcNAcylated proteins important for dauer formation might work in parallel with or downstream of the insulin/IGF-1-signaling pathway. O-GlcNAcylations could conceivably affect the in vivo activity of a number of proteins. A possible example is unc-60/cofilin, which is a regulator of cytoskeletal structure. A previous report has shown that PUGNAc may inhibit neuronal process formation (41), indicative of a cytoskeletal rearrangement defect caused by hyper-O-GlcNAcylation. Neurons play crucial roles during dauer formation; thus, treatment with PUGNAc could disturb the delicate activities of neurons in C. elegans, resulting in a malfunction of the neurons that detect environmental stimuli. All of these various possibilities warrant further investigation.

FIGURE 6.

Proposed working model. Environmental stimuli (e.g. daumone, high temperature, and low food) modulate various protein activities and induce dauer formation. The same stimuli are speculated to enhance overall O-GlcNAcylation during dauer formation. The enhanced O-GlcNAcylations of various components, including DAF-21, UNC-60, F54H12.6, F15E11.1 and PBS-7, might affect the previously known IGF-1-signaling molecules directly or indirectly. However, it is also possible that the O-GlcNAcylated proteins affect dauer formation in parallel with the IGF-1-signaling pathway. During this process, the expressions of various proteins, including F15E11.12, DAF-21, NUO-2, RPL-7, and others, are either enhanced or down-regulated to modulate the dauer formation process.