The Phosphoinositide Kinase PIKfyve/Fab1p Regulates Terminal Lysosome Maturation in Caenorhabditis elegans

Caenorhabditis elegans Strains and General Methods

C. elegans strains were derived from the wild-type Bristol strain N2. Worm cultures, genetic crosses, and other C. elegans methods were performed according to standard protocols (Brenner, 1974). The following mutations and rearrangements were used: ppk-3(mc46), vac-2(n2668), vac-2(n2835) (this work), cup-5(ar465) (Fares and Greenwald, 2001), and szT1 (provided by the Caenorhabditis Genetics Center, University of Minnesota, St. Paul, MN), which balances the right half of LG X and the left half of LGI. Transgenic strains used for this study are cdIs36[punc-122::C31E10.7::GFP] for endoplasmic reticulum, cdEx88[pcc1::gyMANS] for Golgi, pwIs50[lmp-1::GFP] for lysosomes, cdEx49[punc-122::GFP::rab-5] and pwIs72[pvha-6::GFP::rab-5] (provided by B. Grant, Rutgers University, Piscataway, NJ; Chen et al., 2006) for early endosomes, bIs33[rme-8::GFP] for late endosomes (provided by B. Grant; Zhang et al., 2001), cdIs67[punc-122::GFP::rab-7] and pwIs429[pvha-6::RFP::rab-7] (provided by B. Grant; Chen et al., 2006) for late endosomes and lysosomes, and Ex[lgg-1::GFP] for autophagosomes [provided by B. Levine (Columbia University College of Physicians and Surgeons, New York); Melendez et al., 2003]. cdEx46[punc- 122::2XFYVE::GFP] expresses a probe which specifically binds PtdIns(3)P. bIs1[vit-2::GFP] (provided by B. Grant), expressing a fusion between the yolk protein YP170 and green fluorescent protein (GFP), was used to monitor yolk degradation in embryos. Coelomocyte endocytosis was assayed using arIs37[pmyo-3::ssGFP] and cdIs5[pmyo-3::ssDsRed2].

ppk-3 genomic DNA (8 kb), including 1.5 kb of the upstream intergenic region and the whole ppk-3 gene, were PCR-amplified using the primers F1 (5′-CATGGATCCTACTGAAATTTGCCATATGGGG-3′) and R1 (5′-ATTGGATCCTTATAAGACCTAATTGATCCGAG-3′) containing BamHI restriction sites and cloned in the C. elegans pmyo3::YFP vector pPD136-64 [provided by A. Fire (Stanford University School of Medicine, Stanford, CA)]. From this construct, we PCR amplified a fragment containing 1.5 kb of the intergenic region, the ppk-3 genomic DNA fused with yellow fluorescent protein (YFP) in its C terminal part and the 3′-untranslated region of unc-54 using the primers F1 and R2 (5′-AAGGGCCCGTACGGCCGACTAGTAGG-3′). The resulting PCR-amplified ppk-3::YFP fusion (40 ng/μl) was coinjected with pRF4 (100 ng/μl) as a transformation marker into N2 and ppk-3(n2668) young animals to determine ppk-3 expression and to test the rescue of the phenotype.

Lethality assays were done by allowing adult hermaphrodites to lay eggs for 4 h and scoring the number of eggs. For embryonic lethality, larvae were counted 15 h later, and postembryonic lethality was assayed 48 h after egg laying. Lethality percentages were calculated by dividing the number of deduced dead animals (for embryonic lethality, number of eggs laid − number of larvae at 15 h; for postembryonic lethality, number of larvae at 15 h − number of worms alive at 48 h) by the number of eggs laid for each indicated phenotype. The growth delay was assayed by determining the stages of live animals present on the Petri plates 3 and 5 d after egg laying.

RNA interference (RNAi) against rab-7 was performed by the feeding method (Timmons and Fire, 1998) and the corresponding clone W03C9.3 is part of the RNAi library of J. Ahringer (University of Cambridge, Cambridge, United Kingdom).

Isolation and Identification of ppk-3 Mutants

The allele ppk-3(mc46) was isolated by PCR screening, after 25 mM ethane methyl sulfonate (EMS)-induced mutagenesis on wild-type L4 hermaphrodites (Wei et al., 2002), by using two pairs of nested primers located around the active site of ppk-3: F1 (5′-TCACCACCGACACAATCTACC-3′), R1 (5′-CAGTATCAAAGCATCAATGATTTC-3′), F2 (5′-ACTAATCAGAATCAATG CTCGC-3′), and R2 (5′-TCACAACAAGTGAATTGGTAGAC-3′). The mutations vac-2(n2668) and vac-2(n2835) were isolated in an EMS screen for larval lethal mutants with defects in morphogenesis and tissue integrity (Chisholm, unpublished results). ok1150 was provided by the Caenorhabditis Genetics Center. The mutations affecting ppk-3(mc46), vac-2(n2668), and vac-2(n2835) alleles were identified by sequencing each ppk-3 exon after PCR amplification. Animals segregating mc46 or n2668 or n2835 were outcrossed eight times to wild-type animals, and mc46 was balanced with szT1. Complementation tests between mc46 and n2668 were performed by crossing n2668/0 males with mc46/mc46 hermaphrodites. The mc46 mutation was followed through genetic crosses by PCR using three primers: two around the deletion F (5′-ACAGATGGATGATGGGGAAA-3′) and R (5′-CGGAAAGTACGAGTCGATGG-3′), and one hybridizing into the deletion region F(5′-TGGATAAGGTCAGTGTTTGAAC-3′). n2668 is a missense mutation, which destroys a BanII restriction site. We then followed the mutation by cutting with BanII the fragment PCR-amplified using F (5′-ATGTCACGGTTTGAGATTCAG-3′) and R (5′-AGGAAGTGGCTGTCATTGC-3′).

Immunofluorescence

For anti-GFP immunostaining of embryos, mixed staged wild-type or mutant embryos were mounted on polylysine-coated slides, freeze cracked in liquid nitrogen, and fixed in successive baths in −20°C methanol, −20°C acetone, and graded acetone:phosphate-buffered saline mixes. Embryos were then incubated overnight with a monoclonal anti-GFP antibody (Institut de Génétique et de Biologie Moléculaire et Cellulaire, Euromedex, Souffelweyersheim, France) diluted 1/1000 in phosphate-buffered saline (PBS)/0.5% Tween/0.1% bovine serum albumin (BSA), washed with PBS/0.5% Tween/0.1% BSA, incubated overnight with a Cy3-coupled secondary antibody, washed with PBS/0.5% Tween containing 1 μg/ml final 4,6-diamidino-2-phenylindole (DAPI) to locate all nuclei, and viewed with a confocal microscope.

Anti-GFP immunostaining of adult worms was done according to the protocol of McIntire with minor modifications (McIntire et al., 1992). Briefly, wild-type and mutant young adults were fixed 1 d in 4% paraformaldehyde and then incubated overnight in 5% β-mercapthethanol, 0.125 M Tris, pH 6.9, and 1% Triton at 37°C. The cuticle was disrupted by treatment with 1000 U/ml collagenase type IV (Sigma-Aldrich, St. Louis, MO) in 100 mM Tris, pH 7.5, 1 mM CaCl₂, for 30 min at 37°C, and membranes were impermeabilized by an incubation of 1 h in 0.125 M Tris, pH 6.9, 1% Triton. Worms were then incubated 24 h with a monoclonal anti-GFP antibody (Institut de Génétique et Biologie Moléculaire et Cellulaire, Euromedex) diluted 1/1000, washed, incubated 4 h with a Cy3-coupled secondary antibody, washed, and viewed with a confocal microscope. Antibodies were diluted in PBS/0.5% Triton/0.5% BSA, and washes were done with PBS/0.5% Triton/0.3% BSA.

For anti-PtdIns(3,5)P₂ immunostaining, mixed staged embryos of wild-type and mutants were mounted and freeze cracked, incubated in −20°C methanol, fixed overnight in 1% paraformaldehyde, and then incubated overnight with a monoclonal anti-PtdIns(3,5)P₂ antibody (Echelon Biosciences, Salt Lake City, UT) diluted 1/40 (12.5 μg/ml final) in PBS/0.1% BSA without any detergent. Embryos were then washed, incubated overnight with a fluorescein isothiocyanate-coupled goat anti-mouse secondary antibody (Beckman Coulter France, Roissy CDG cedex, France), washed with PBS/DAPI, and viewed with a confocal microscope. Experiments were done twice.

Confocal Microscopy

Animals were mounted on 2% agarose pads with 0.2% tricaine/0.02% tetramisole in M9. For differential interference contrast, we used a Zeiss Axioplan microscope coupled to a CoolSNAP camera (Photometrics, Tucson, AZ) under a PlanApo 10× objective. For confocal imaging, we used a Leica SP2-AOBS confocal microscope with an argon 488 excitation to visualize the green dyes and the autofluorescence of worm intestine and a HeNe 543 excitation for the red dyes. Pictures of the same experiment were taken at the same exposure. We then processed images with the Tcstk software (Jean-Luc Vonesch, Institut de Génétique et de Biologie Moléculaire et Cellulaire) and edited pictures using the Dvrtk software (Jean-Luc Vonesch, Institut de Génétique et de Biologie Moléculaire et Cellulaire) and Adobe Photoshop 7.0. Fluorescence intensities were quantified in whole embryos using the Leica Confocal software (Leica, Wetzelar, Germany). Dot-like structures of the anti-PtdIns(3,5)P₂ staining in embryos were scored using the MetaMorph software (Molecular Devices, Sunnyvale, CA) in squares of 280 μm². The average value obtained with 25 different embryos per strain was calculated and difference between wild-type and ppk-3(n2668) mutant was tested with a Student’s t test.

Electron Microscopy

Worms were processed for electron microscopy as described previously (Michaux et al., 2000). Briefly, young adults were sectioned and fixed for 24 h in 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer, pH 7.2, at 4°C. Worms were postfixed for 5 h with 2% osmium tetroxide in the same sodium cacodylate buffer at 4°C. Animals were dehydrated in graded alcohol:water mixes and then embedded in Epon. Ultrathin sections of 70 nm were contrasted with uranyl acetate and lead citrate. Sections were observed with a Philips CM12 electron microscope. At least eight animals per strain were examined.

Endocytosis Assays and Time Course

As described previously (Treusch et al., 2004), Texas-Red BSA (TR-BSA; Sigma) was injected at 1 mg/ml in water into the body cavity of wild-type or n2668 young adult hermaphrodites expressing LMP-1::GFP. At defined time points, animals were mounted on slides, put on ice to stop endocytosis, and fluid-phase internalization of the dye into the coelomocytes was viewed with a confocal microscope. For each time point, similar results were obtained with the coelomocytes of at least six different worms. LysoTracker Red DND-99 (Invitrogen, Carlsbad, CA) is a red fluorescent dye that selectively accumulates in cellular compartments with low internal pH. LysoTracker Red DND-99 was injected into the pseudocoelomic space of wild-type and mutant worms expressing LMP-1::GFP at 0.4 mM in M9 buffer, and its localization inside the coelomocytes was viewed 4 h after injection with a confocal microscope at a 514-nm excitation wavelength and a 580- to 660-nm range emission. The dynamic of the abnormal vacuole biogenesis in ppk-3(mc46) and ppk-3(n2668) mutant coelomocytes was followed through 20 min. Focal plane was changed during experiment to ensure proper visualization of the biggest diameter of the fusing vacuole observed, and pictures were taken at key time points.

Lipid Extraction and Analysis

Worms were collected and washed in phosphate-free RPMI-1640 medium at 20°C and labeled with 1 mCi/ml [³²P]orthophosphate (GE Healthcare, Little Chalfont, Buckinghamshire, United Kingdom) during 12 h in phosphate-free RPMI-1640 medium at 20°C. ³²P-labeled worms were then washed once in phosphate-free RPMI-1640 medium at 20°C, and lipids were immediately extracted following the modified procedure of Bligh and Dyer (1959). Lipids were then separated by TLC (Merck, Nogent-sur-Marne, France) using CHCl₃/CH₃COCH₃/CH₃COOH/H₂O [80/30/26/24/14 (vol/vol)] as a solvent. The spots corresponding to [³²P]PtdInsP₂ were visualized by a PhosphorImager 445 SI (GE Healthcare) using appropriate standards, scraped off, deacylated, and analyzed by high-performance liquid chromatography (HPLC) on a Whatman Partisphere 5 SAX column (Whatman International, Maidstone, United Kingdom) as described previously (Payrastre, 2004).

Developmental and Vacuolar Defects in ppk-3 Mutants

To identify the Caenorhabditis elegans orthologue of PIKfyve/Fab1p, we compared the human and Saccharomyces cerevisiae protein sequences to the C. elegans predicted transcript and genomic sequences using tblastn. We identified VF11C1L.1 (also named ppk-3) on the X chromosome as the C. elegans orthologue. We updated the predicted sequence found in the database by sequencing cDNA clones [OSTR10224F1, provided by M. Vidal (Dana–Farber Cancer Institute, Boston, MA)], by comparison of expressed sequence tags on WormBase [yk246g4.5, yk1185g10.5 from Y. Kohara (National Institute of Genetics, Shizuoka, Japan)] to confirm that the predicted intron 6 is indeed coding, and by comparison of the PIKfyve protein family sequences through evolution to extend the N-terminal sequence. We predicted that the ppk-3 gene encompasses 14 exons and encodes a 1497-amino acid-long protein with a start ATG codon at position 12 975 563 on the X chromosome (Figure 1 and Supplemental Figure 1). Like its orthologues in yeast, plants, and other eukaryotes, PPK-3 has an N-terminal Fab1p, YOTB, Vac1p, EEA1 (FYVE) domain, which might bind to PtdIns(3)P, a domain related to the Cpn60/TCP-1 chaperonin family, and a lipid kinase domain related to yeast Mss4 and class I and II phosphatidylinositol kinases. The protein is, however, significantly smaller than the human protein and lacks the Dishevelled, Egl-10, and Pleckstrin domain found between the FYVE and the chaperonin-like domains in Drosophila, mouse and human orthologues.

To investigate the expression pattern of ppk-3, we fused genomic sequence, from 1.5 kb upstream of the initiation codon to the last codon, with YFP and observed, at all stages, a mosaic expression in the pharynx and in the nervous system: some head and tail neurons, some neurons of the dorsal and ventral nerve cords, and some commissures. At early larval stages, an intestinal expression was detected (Supplemental Figure 2). This is compatible with an expression pattern determined using a ppk-3::GFP transcriptional fusion sEx13072 (McKay et al., 2003). In neurons, our ppk-3::YFP fusion protein was localized as cytoplasmic dots in cell bodies and axons. In the intestine, PPK-3 displayed a diffuse localization with some punctate staining (Supplemental Figure 2).

We screened by PCR for ppk-3 mutants generated by EMS mutagenesis and isolated a deletion mutant (mc46) lacking most of the kinase domain (Δ5213-6077 on the genomic sequence) (Figure 1). This mutant displayed large vacuoles and by comparison with this swiss-cheese phenotype, we identified two other mutants, n2668 and n2835, mapping to the same chromosomal region as the ppk-3 gene, which were obtained by EMS mutagenesis and screened for larval lethal mutants with defects in morphogenesis and tissue integrity. n2668 displayed enlarged vacuoles similar to mc46 as monitored by Nomarski optics (Figure 2B). ok1150 (Δ4027–5640 on the genomic sequence) was created by the C. elegans Gene Knockout Consortium. There was no allelic complementation between mc46 and n2668 (Figure 2A) nor between n2668 and n2835. n2668 causes a serine-to-phenylalanine (S1356F) change within the conserved catalytic loop DLKGS of the kinase domain, close to the aspartate mutated in the S. cerevisiae kinase dead Fab1p mutant (D2134R; Gary et al., 1998). n2835 causes a nonsense mutation (Q1487X), which removes the last 11 amino acids conserved in mammals (Figure 1).

These different mutants can be arranged in an allelic series composed of strong mutants (mc46 and ok1150), a hypomorphic mutant (n2668), and a weak hypomorphic mutant (n2835) (Figure 2A). Homozygous mc46 and ok1150 animals laid by heterozygous mothers were viable but displayed numerous vacuoles, which filled a large part of the body. These homozygous progeny laid eggs that also contained vacuoles, did not hatch, and died at different embryonic stages. This maternal effect was not strict because it was rescued in heterozygous progeny obtained by crossing the viable homozygous mutants mc46/mc46 with wild-type +/0 males: the 46% worms that lived were all mc46/+ hermaphrodites, and the 54% dead eggs corresponded probably to the mc46/0 males (Figure 2A). n2668 also displayed a developmental defect: 27% embryonic lethality, 8% postembryonic lethality. Viable n2668 animals had growth retardation as they reached adult stage up to 2 d later than the wild-type (wild-type animals are adults 2 d after hatching) (Figure 2A). Compared with n2668, n2835 presented a reduced lethality and no growth retardation (Figure 2A). Both mutants displayed abnormal vacuoles of different sizes, which occurred at the one-cell stage in n2668. In n2668 larvae and adult mutants, vacuoles were present in intestine, epidermis, coelomocytes (one big vacuole), and to a lesser extent in the pharynx (Figure 2B). To confirm that the observed phenotypes are due to mutations in ppk-3, we injected the ppk-3::GFP fusion construct into n2668 mutants and obtained a rescue of the developmental defects (Figure 2A) and vacuolar phenotypes (our unpublished data). In conclusion, mutations of ppk-3 in a multicellular eukaryote are responsible for developmental defects and for the appearance of abnormal vacuoles.

Abnormal Vacuoles in ppk-3 Mutants Are Enlarged Lysosomes

To characterize these abnormal vacuoles and their origin, we used electron and fluorescent microscopy. The vacuoles were of different sizes and could reach up to 30% of the worm width (Figure 3A). This striking vacuolar phenotype did not prevent 65% of the worms from reaching adulthood in the n2668 mutant. The vacuoles were surrounded by a membrane and were either empty, filled with membrane whirls in huge vacuoles, or electron-dense deposits in smaller vacuoles. We noted numerous pictures of aggregation and fusion between the vacuoles (Figure 3B). Normal multivesicular bodies were visible in the epidermis of the ppk-3 mutant (Figure 3C). Most of the coelomocytes were filled with a unique huge vacuole that leads to an increase of the cell size of approximately two- to fourfold (Figure 3E). These results suggested a defect of membrane trafficking in ppk-3 mutants.

To better define the origin of the enlarged vacuoles, we focused our analysis on coelomocytes, which are scavenger cells harboring strong endocytic and degradation processes (Fares and Grant, 2002). We chose the n2668 mutant, which is viable and has a missense alteration in the catalytic loop of the kinase domain, to focus on the impact of PI dysregulation. We crossed n2668 with transgenic worms expressing specific markers of different endocytosis and membrane compartments. There was no difference in Golgi (mannosidase II) and endoplasmic reticulum (C31E10.7 cytochrome b₅) patterns between wild-type and mutant (our unpublished data). The endocytic markers, RAB-5 for early endosomes, and RME-8 for late endosomes (Zhang et al., 2001), did not label the enlarged vacuole (Figure 4). In contrast, LMP-1, a lysosomal marker orthologous to human lysosomal-associated membrane protein 1 (Kostich et al., 2000), always stained the membrane of the enlarged vacuoles, strongly suggesting a lysosomal identity. We obtained similar results for LMP-1 in the mc46 mutant. RAB-7 weakly labeled the enlarged vacuoles of the mutant (Supplemental Figure 3A). In C. elegans, RAB-7 labels not only late endosomes but also lysosomes because all the vacuoles accumulating the fluid-phase marker Texas-Red–coupled BSA after 24 h were positive for RAB-7 and LMP-1, whereas they were negative for RME-8 (Supplemental Figure 3B).

In the intestine of n2668, LMP-1 stained the bigger vacuoles, whereas the small abnormal vacuoles were negative for RAB-5 and positive for RAB-7 (Figure 5). In contrast to coelomocytes, there was also a slight enlargement of late endosomes in the intestine. LGG-1, an autophagosome marker orthologous to LC3 (Melendez et al., 2003), showed aggregations distinct from the abnormal vacuoles in the intestine of the ppk-3 mutant (Figure 4). Thus, ppk-3 mutants have enlarged lysosomes and a possible accumulation of autophagosomes.

To decipher the mechanism of enlargement of the lysosomes, we followed LMP-1–positive vacuoles in ppk-3 mutants (mc46 and n2668) over time in coelomocytes. We observed fusion between LMP-1–positive vacuoles. Smaller vacuoles fused with the huge vacuole by two apparent mechanisms: either they fused rapidly or progressively (Figure 6A). For the first case, this led to an intermediate where the membrane of the huge vacuole was distorted after the fusion of the two membranes (fusion event within seconds). For the second case, the fusing vacuole decreased in size progressively (fusion event over a 20-min interval). Together, our results suggest that lack of ppk-3 activity leads to enlargement of lysosomes, and addition of membranes is mediated through fusion with lysosomes.

Involvement of ppk-3 in Terminal Maturation of Lysosomes

To place more precisely the defect leading to lysosome enlargement, we first performed fluid uptake assays by injecting Texas-Red–coupled BSA into the body cavity of wild-type and ppk-3 mutants. This fluorescent marker enters coelomocytes by fluid-phase–mediated endocytosis and is transported through the different endocytic compartments to the lysosomes. In wild-type animals, Texas-Red BSA was present in early endocytic compartments after 15 min, and we noted that it was concentrated in subdomains that are presumably budding domains (Treusch et al., 2004). By 30 min, Texas-Red BSA was mostly in LMP-1–negative vesicles, which are late endosomes as verified with a time course using RME-8::GFP. From 30 min, it started to accumulate in lysosomes labeled with LMP-1::GFP (Figure 6B; our unpublished data). By 1 h and after 24 h of internalization, Texas-Red BSA had accumulated into most of the lysosomes (Figure 6B) and was absent from the RME-8::GFP late endosomes (Supplemental Figure 3B). In ppk-3 mutants, the fluid phase endocytosis followed the same kinetics. Texas-Red BSA first localized to LMP-1–negative endosomes with concentrations in subdomains. By 30 min and 1 h, the Texas-Red BSA accumulated in normal LMP-1–positive compartments and in the LMP-1 large vacuoles. This is consistent with our finding that the large vacuoles are competent to fuse with other compartments. The fluorescent marker localization to the enlarged vacuole increased during the time course, indicating the progressive accumulation of the tracer (Figure 6B). Thus, lack of ppk-3 activity does not prevent the fluid phase endocytosis to lysosomes, suggesting that the biogenesis of lysosomes is normal.

We further tested the maturation status of the enlarged lysosomes. We used worm strains expressing GFP or DsRed2 in the body wall muscle under the myo-3 muscle-specific promoter. These fluorescent proteins are secreted into the body cavity and nonspecifically endocytosed primarily by coelomocytes. It is known that GFP is degraded, or bleached by the low pH, in mature lysosomes, whereas DsRed2 is stable (Fares and Grant, 2002; Patton et al., 2005). The enlarged lysosomes of the ppk-3(n2668) mutant were strongly positive for DsRed2, whereas there was no accumulation of the GFP signal (Figure 7A). As a control, we used coelomocyte uptake defective (cup)-5 mutants, which have a defect in lysosome formation (Treusch et al., 2004). cup-5 encodes a calcium-permeable channel orthologous to h-mucolipin-1, mutated in human mucolipidosis type IV (Bargal et al., 2000). In contrast to ppk-3 mutants, cup-5 mutants accumulated the GFP in aberrant late endosome/lysosome hybrid organelles. To confirm that ppk-3 lies downstream of cup-5, we generated a double mutant cup-5(ar465);ppk-3(n2668). GFP accumulated similarly in this double mutant compared with the single cup-5 mutant (Figure 7A). CUP-5:: GFP accumulated around the enlarged vacuoles of the ppk-3 mutants, and lethality increased in the double cup-5(ar465); ppk-3(n2668) (supplementary Figure 4), suggesting that these two proteins are implicated in the same process, PPK-3 being downstream of CUP-5. In conclusion, enlarged vacuoles of the ppk-3 mutants are maturing lysosomes. It suggests that PPK-3 is not essential for lysosome formation but plays a role in lysosome maturation.

We further checked whether the enlarged lysosomes of the ppk-3 mutants were functional by investigating protein degradation in the ppk-3 mutant. Because the mutant worms expressing pmyo-3::GFP from Figure 7A did not accumulate the GFP in the enlarged lysosomes, we checked that the GFP was indeed degraded and not bleached. Anti-GFP immunostaining of the ppk-3 mutant confirmed that GFP was degraded showing that, although there is no protein degradation in the enlarged hybrid organelles of cup-5 mutants, protein degradation occurred in the enlarged lysosomes of the ppk-3 mutants (Figure 7B).

As a second test for protein degradation, we monitored the presence and level of the yolk protein vitellin-2 (YP170) in embryos (Britton and Murray, 2004). Yolk proteins are synthesized by intestinal cells and transported to the oocytes to be used as amino acid source by the developing eggs. It was previously reported that cathepsin-L1 mutants and worms fed with RNAi against rab-7, two genes implicated in lysosome formation and function, are defective for yolk protein processing and degradation, leading to an accumulation of YP170::GFP in abnormal vacuoles (Grant and Hirsh, 1999; Britton and Murray, 2004), as shown for rab-7 RNAi in Figure 8A. We compared the fluorescence intensity of YP170::GFP in wild-type and ppk-3 mutant transgenic embryos. No detectable difference was noted (p = 0.07) (Figure 8, B and C). To rule out possible bleaching of the GFP fluorescence due to an acidic environment, we performed an anti-GFP immunostaining and obtained similar results (p = 0.29) (Figure 8, B and C). This was sustained by the presence of autofluorescent granules in the intestine of ppk-3 mutants, believed to correspond to lysosomes containing lipofuscin, a degradation product (Clokey and Jacobson, 1986).

Acidification was monitored by injection of LysoTracker Red DND-99 into the coelom of LMP-1::GFP-expressing worms. LysoTracker Red DND-99 fluoresces in acidic compartments and is retained in their membranes. In wild-type worms, all lysosomes were positive for LysoTracker Red DND-99, whereas in ppk-3(n2668) mutant, the enlarged vacuoles did not stain, suggesting that they were not acid (Figure 8D). Together, our results suggest that degradation occurs in ppk-3 mutants, probably before the enlargement of the lysosomes and neutralization of pH.

Defect of PtdIns(3,5)P₂ Production in ppk-3 Mutants

We monitored the localization of PtdIns(3)P and PtdIns(3,5)P₂, substrate and product of the PIKfyve/Fab1p kinase activity, respectively, in the ppk-3(n2668) mutant, which has a S1356F missense in the catalytic loop. PtdIns(3)P was specifically detected with a probe consisting of a tandem repeat of human HRS FYVE domain (Gillooly et al., 2000; Dang et al., 2004) under the control of a coelomocyte promoter. The limiting membrane of the enlarged lysosome was not labeled for the majority of the coelomocytes, although other vesicles were positive for this probe; one-third of the coelomocytes contain an abnormal vacuole faintly positive for the 2XFYVE probe (Figure 9A). PtdIns(3,5)P₂ was detected with a specific mouse monoclonal antibody in embryos and occurred as cytoplasmic dots. The number of dot-like structures in ppk-3(n2668) embryos was highly significantly decreased compared with wild-type (p = 2 ×10⁻⁷) (Figure 9, B and C). The remaining PtdIns(3,5)P₂ staining in the mutant might be due either to nonspecific labeling, or to residual production of PtdIns(3,5)P₂ by the mutated PPK-3 or synthesis by an uncharacterized pathway.

To confirm a decrease in PtdIns(3,5)P₂ production, we set up a protocol to measure its level in C. elegans by ³²P_i labeling, separation of [³²P]phospholipids by thin layer chromatography (TLC) and quantification of the amount of the different isomers of PtdInsP₂ by HPLC. We report for the first time the detection of PtdIns(3,5)P₂ in C. elegans where it is present at low level, as it is in yeast and mammalian cells (Dove et al., 1997; Sbrissa et al., 2001). Although ppk-3 mutants consistently incorporated more ³²P in phospholipids than wild-type worms, we reproducibly observed that the mutants had a much lower level of PtdIns(3,5)P₂ (Figure 9D). We normalized the level of PtdIns(3,5)P₂ on PtdIns(4,5)P₂ and calculated in the representative experiment Figure 9D that PtdIns(3,5)P₂ on PtdIns(4,5)P₂ ratio was 1/40 when comparing ppk-3 mutant and wild type. Thus, the enlargement of lysosomes seems to be linked to a defect in PtdIns(3,5)P₂ production.