Roles of CUP-5, the Caenorhabditis elegans orthologue of human TRPML1, in lysosome and gut granule biogenesis

Background CUP-5 is a Transient Receptor Potential protein in C. elegans that is the orthologue of mammalian TRPML1. Loss of TRPML1 results in the lysosomal storage disorder Mucolipidosis type IV. Loss of CUP-5 results in embryonic lethality and the accumulation of enlarged yolk granules in developing intestinal cells. The embryonic lethality of cup-5 mutants is rescued by mutations in mrp-4, which is required for gut granule differentiation. Gut granules are intestine-specific lysosome-related organelles that accumulate birefringent material. This link between CUP-5 and gut granules led us to determine the roles of CUP-5 in lysosome and gut granule biogenesis in developing intestinal cells. Results We show that CUP-5 protein localizes to lysosomes, but not to gut granules, in developing intestinal cells. Loss of CUP-5 results in defects in endo-lysosomal transport in developing intestinal cells of C. elegans embryos. This ultimately leads to the appearance of enlarged terminal vacuoles that show defective lysosomal degradation and that have lysosomal and endosomal markers. In contrast, gut granule biogenesis is normal in the absence of CUP-5. Furthermore, loss of CUP-5 does not result in inappropriate fusion or mixing of content between lysosomes and gut granules. Conclusions Using an in vivo model of MLIV, we show that there is a defect in lysosomal transport/biogenesis that is earlier than the presumed function of TRPML1 in terminal lysosomes. Our results indicate that CUP-5 is required for the biogenesis of lysosomes but not of gut granules. Thus, cellular phenotypes in Mucolipidosis type IV are likely not due to defects in lysosome-related organelle biogenesis, but due to progressive defects in lysosomal transport that lead to severe lysosomal dysfunction.


Background
Lysosomes are the major degradative organelles of endocytosed, phagocytosed, and autophagocytosed material [1,2]. Lysosomes also have specialized functions, for example fusing with the plasma membrane to initiate wound repair and mediating some cell death pathways [3][4][5]. Lysosome biogenesis is a dynamic process, in which late endosomes fuse with lysosomes resulting in a hybrid late endosome/lysosomal organelle [6]. Late endosomes and lysosomes are reformed from these hybrid organelles, a process that requires the release of intra-organellar Ca 2+ [6,7]. Some tissues have additional organelles called lysosome-related organelles (LROs) that are acidic, contain some lysosomal proteins, and have cell type-specific functions [8]. Although LROs derive from the endosomal system, they are different from bona fide lysosomes in composition, morphology, and function. Examples of LROs include organelles with storage or secretion functions, such as melanosomes in melanocytes, plateletdense granules in platelets, and acrosomes in sperm cells [8][9][10]. In C. elegans embryos, gut granules are LROs found in intestinal cells at all stages of development [11][12][13][14]. These gut granules contain lipids, birefringent material that is autofluorescent under several wavelengths of light, and gut granule-specific proteins. The function and biogenesis of gut granules is not completely understood.
In C. elegans, CUP-5 is required for the biogenesis of lysosomes in scavenger cells called coelomocytes [15,16]. CUP-5 is the sole orthologue of mammalian TRPML1 that is encoded by MCOLN1, mutations in which cause Mucolipidosis type IV (MLIV) in humans [15,17]. Many MLIV-associated defects that are linked to lysosomal dysfunction have been described. Some examples include an enlargement of lysosomes that accumulate both lipid and water soluble material, a delay in the transport of endocytosed lactosylceramide (LacCer) from late endosomes/lysosomes to the Golgi Apparatus, a delay in the degradation in and/or transport of endocytosed lipids and proteins to lysosomes, and a delay in the degradation of autophagosome material [18]. Similar to the MLIV phenotypes, worms with a cup-5 mutation have enlarged endo-lysosomal compartments in several cell types, including developing intestinal cells and coelomocytes [15,19]. Pulse-chase studies in coelomocytes have shown that CUP-5 is required for the biogenesis of lysosomes, the earliest MLIV-associated defect in the endocytic pathway that has yet been described [16].
Loss of CUP-5 results in embryonic lethality [20]. In these cup-5 mutant embryos, there is a significant enlargement of yolk granules and a defect in the degradation of endocytosed yolk proteins in developing intestinal cells [20]. Loss of MRP-4 was shown to rescue both the yolk granule/yolk degradation defects and the embryonic lethality of cup-5 mutant [19]. However, MRP-4 is an ATP-Binding Cassette (ABC) transporter that is required for gut granule differentiation in developing intestinal cells [21]. Because of the role of MRP-4 in LRO/gut granule differentiation and the mrp-4 mutant rescue of cup-5 mutants, we hypothesized that CUP-5 may play a role in LRO biogenesis or differentiation, in addition to its role in lysosome biogenesis. We therefore initiated this study to decipher possible roles of CUP-5 in lysosome and gut granule biogenesis in developing intestinal cells.

Results and discussion
Discrete late endocytic organelles and gut granules in the developing intestine Two late-stage endo-lysosomal organelles (yolk granules and lysosomes) and one LRO (gut granules) have been identified in developing intestinal cells of embryos [11,22]. However, because there are no systematic studies that visualized all three organelles in developing intestinal cells, it is not clear whether they indeed represent discrete and separate compartments. We therefore first carried out colocalization studies in wild type embryos to define these organelles using established markers. We confined our analysis to specific stages of embryonic development, "comma" to "1.5-fold" stages, when the yolk granule defect is clearly observed in cup-5 mutants [20].
Yolk granules, lysosomes, and gut granules represent distinct compartments of developing intestinal cells. YP170::GFP did not colocalize with any of the gut granule markers at steady state (Fig. 1A, B; Additional file 1 Fig. S1B). This indicates that yolk granules are discrete and separate from gut granules. Similarly, YP170::GFP did not colocalize with LMP-1::TagRFP(S158T) at steady state (Fig. 1C). This indicates that yolk granules and lysosomes are distinct compartments; yolk is likely transported to lysosomes but is degraded in these organelles. Finally, LMP-1::GFP did not colocalize with PGP-2 or autofluorescence at steady state ( Fig. 1D; Additional file 1 Fig. S1C). This indicates that lysosomes and gut granules are also discrete and separate organelles. Having defined the identities of these organelles by light microscopy studies, we then did localization and mutational analyses to decipher roles of CUP-5 in organelle biogenesis and transport in developing intestinal cells.

cup-5(zu223) lethality is primarily due to defects in the developing intestine
We had previously made a functional GFP::CUP-5 fusion protein that rescued cup-5 mutant lysosomal defects in C. elegans coelomocytes [16]. To assay whether this GFP::CUP-5 also rescues intestinal defects in embryos, we expressed this GFP::CUP-5 under the control of an elt-2 promoter that drives expression in developing intestinal cells [27]. Although most tissues of cup-5(zu223) embryos show lysosomal defects, this Pelt-2::gfp::cup-5 construct rescues the lethality in cup-5 (zu223) embryos: 5.4 ± 5.9% lethality compared with Figure 1 Late endocytic and lysosome-related organelles in developing intestinal cells of wild type embryos. A) Epifluorescence images of a wild type embryo that expresses YP170::GFP and stained with Nile Red. B) Confocal images of a wild type embryo that expresses YP170:: GFP and immunostained to detect PGP-2. C) Epifluorescence images of a wild type embryo that expresses YP170::GFP and LMP-1::TagRFP(S158T). D) Confocal images of a wild type embryo that expresses YP170::GFP and immunostained to detect PGP-2. Bottom panels are magnified images of the regions indicated in the top panels. Scale bars in whole embryo images represent 10 μm; scale bars in magnified images represent 2 μm. Error bars represent the 95% confidence intervals for significance. B) Epifluorescence images of a wild type embryo that expresses LMP-1::GFP and mCherry::CUP-5. C) Confocal images of a wild type embryo that expresses GFP::CUP-5 and LMP-1::TagRFP(S158T). D) Epifluorescence images of a wild type embryo that expresses YP170::GFP and mCherry::CUP-5. E) Confocal images of a wild type embryo that expresses GFP::CUP-5 and immunostained to detect PGP-2. Bottom panels are magnified images of the regions indicated in the top panels. Scale bars in whole embryo images represent 10 μm; scale bars in magnified images represent 2 μm.
100% without the construct ( Fig. 2A). This indicates that the embryonic lethality of cup-5 mutant embryos is mostly due to intestinal defects, which is consistent with the strong suppression of cup-5 mutant lethality by loss of MRP-4 since mrp-4 is only expressed in developing intestinal cells [19,21]. Furthermore, elt-2 promoter becomes active after the yolk granule defects manifest themselves in cup-5 mutant embryos [20]. This suggests that expression of CUP-5 is sufficient to reverse the endocytic defects of the cup-5 mutant, an in vivo model of MLIV.

CUP-5 protein localizes to lysosomes
We had previously localized CUP-5 to lysosomes of coelomocytes [16]. We wanted to determine whether CUP-5 localizes only to lysosomes of developing intestinal cells or whether it is also found on yolk granules and gut granules. We used our transgenic worms expressing either GFP::CUP-5 or mCherry::CUP-5 in intestinal cells for these localization studies.
Therefore, CUP-5 protein localizes to lysosomes at steady state in developing intestinal cells. We next assayed whether loss of CUP-5 causes content mixing between yolk granules, lysosomes, and gut granules.
Yolk Granule and lysosome content mixing in the cup-5 (zu223) mutant To determine whether CUP-5 affects yolk granule, lysosome, and gut granule biogenesis, we did localization studies between markers for yolk granules, lysosomes, and gut granules in cup-5(zu223) embryos. Similar to wild type embryos, YP170::GFP did not colocalize with any of the gut granule markers at steady state in cup-5 (zu223) embryos (Fig. 3A, B; Additional file 3 Fig. S3A). This indicates that yolk granules remain discrete and separate from gut granules in the absence of CUP-5. In contrast to wild type embryos, LMP-1::TagRFP(S158T) localized to the limiting membranes of compartments that contained YP170::GFP in mutant embryos (Fig. 3C). Specifically, we found that all compartments that were labeled with LMP-1::TagRFP(S158T) also contained YP170::GFP (arrowheads in Fig. 3C), but some YP170:: GFP-containing compartments did not show detectable LMP-1::TagRFP(S158T) (arrows in Fig. 3C, Additional file 4 Fig. S4). This indicates that there is a transport defect between yolk granules and lysosomes in the absence of CUP-5. Consistent with these results, LMP-1::GFP did not colocalize with the gut granule marker PGP-2 or gut granule lipofuscin ( Fig. 3D; Additional file 3 Fig. S3B). Thus in the absence of CUP-5, there is an apparent mixing of content of lysosomes and of yolk granules, while the gut granules remain discrete and separate compartments. This indicates that gut granule biogenesis is normal in the absence of CUP-5. We therefore asked whether there is a defect in the morphologies or content of the various organelles.
Alteration in size and content of yolk granules/lysosomes in the cup-5(zu223) mutant We had previously shown an enlargement of the total population of yolk granules and an accumulation of YP170::GFP in C. elegans cup-5(zu223) mutants [20]. We visualized markers for the different organelles in wild type and cup-5(zu223) embryos using identical microscopy parameters that allow us to compare sizes of organelles and intensities of staining of each marker.
To determine the sizes of the compartments, we relied on endogenous markers, YP170::GFP for yolk granules, LMP-1::GFP for lysosomes, and autofluorescence for gut granules; this eliminates potential variability due to exogenous staining with dyes. Indeed, Nile Red staining of gut granules showed significant variability in staining quality between different experiments. We also used PGP-2 antibody immunostaining to confirm sizes of gut granules.
In contrast to the yolk granule/lysosome defect, there was no apparent effect of the loss of CUP-5 on gut granule size. Gut granule sizes were similar in cup-5 (zu223) and in wild type embryos based on UV fluorescence (1.17 ± 0.07 μm 2 for mutant, 1.11 ± 0.06 μm 2 for wild type, p-value = 0.25) and PGP-2 immunofluorescence (0.70 ± 0.04 μm 2 for mutant, 0.76 ± 0.04 μm 2 for wild type, p-value = 0.06) (Fig. 4C, D). However, the intensity per unit area of autofluorescent material was statistically lower per unit area in cup-5(u223) mutant embryos compared with wild type embryos (p-value = 2.3 × 10 -16 ) (Fig. 4D), while the intensity per unit area of PGP-2 immunofluorescence was similar in both wild type and mutant embryos (p-value = 0.57) (Fig. 4C). Thus while the biogenesis of gut granules is normal, the , and UV (D) signals. Images for the same marker were taken using the same microscopy parameters. Scale bars represent 10 μm. Error bars represent the 95% confidence intervals for significance, and asterisks represent statistical significance at p < 0.05, compared to WT. In cup-5(zu223) embryos, the difference in areas of YP170::GFP organelles (A) with or without LMP-1:: TagRFP(S158T) is also statistically significant.
loss of CUP-5 may have indirect effects on the composition of these granules; the severe yolk granule/lysosome defect in the cup-5 mutant may affect the trafficking of some gut granule proteins.

Conclusion
The localization of CUP-5 primarily to lysosomes and the cup-5 mutant defects are both consistent with a model where CUP-5 functions primarily in lysosome biogenesis and/or transport but does not function in the biogenesis of gut granules. We propose that YP170 is transported from early endosomes to late endosomes (yolk granules) in both wild type and cup-5 mutant embryos (Fig. 5). In wild type, YP170, LMP-1, and CUP-5 are efficiently transported to lysosomes where YP170 is degraded. In the absence of CUP-5, the rate of YP170 and LMP-1 transport to lysosomes is reduced leading to the progressive enlargement of yolk granules (Fig. 5). Eventually, this results in a terminal phenotype, the appearance of large vacuoles that accumulate both YP170 and LMP-1. Gut granules/LROs are presumed to originate from early endosomes. The observation that loss of CUP-5 has a minimal effect on the formation of gut granules indicate that CUP-5 functions downstream of this early endosomal sorting step but upstream of late endosome-tolysosome transport. However, as this yolk granule/lysosomal defect gets more severe in the absence of CUP-5, it indirectly impacts the trafficking of some gut granule proteins. Our results suggest that MLIV patients will show normal LROs that would not contribute to the symptoms of the disease. Furthermore, our in vivo studies suggest that TRPML1 has functions in late endosome-to-lysosome transport that are earlier than those inferred from some of the terminal lysosomal defects described in cell culture models of MLIV; these terminal lysosomal defects include defective lysosome/autophagosome fusion and reformation and lysosome exocytosis [28][29][30][31][32].

Microscopy
Nile Red is a lipophilic membrane-permeable dye that diffuses across membranes and fluoresces when it complexes with lipids [24,26,36]. For Nile Red staining, adult worms were placed on OP50 bacteria suspended in 60 ug/ml Nile Red, on NGM plates with 60 ng/ml Nile Red (Invitrogen, Carlsbad, CA). Worms were allowed to lay embryos overnight, and embryos were immediately imaged. For imaging of live worm embryos (Nile Red, UV imaging, and fluorescently tagged constructs), embryos were resuspended in M9 buffer and dropped onto 3% agar pads; coverslips were added and sealed with petroleum jelly.
For PGP-2 immunostaining, embryos were dropped onto slides coated with poly-l-lysine (Sigma, St. Louis, MO) and covered with a coverslip. Slides were quickly frozen on a metal plate cooled over dry ice, and coverslips were flicked off the slides. Slides then underwent one 15-minute 100% methanol wash at -20°. Slides were washed twice with PBST (PBS plus 0.1% Tween), and blocked with PBS, 1% nonfat milk, 1% BSA, 0.1% Tween for 2 hours at 4°. Anti-PGP-2 antibody was added at 1:500 dilution in the above blocking buffer and left overnight at 4°, followed by three 10-minute washes in PBST [26]. Secondary anti-rabbit Cy3 antibody (Jackson ImmunoResearch Laboratories, West Grove, PA) was added at 1:200 dilution in blocking buffer for two hours at room temperature, followed by three 10-minute washes in PBST. Excessive moisture was wicked away from embryos, and mounting media (Invitrogen) was added prior to imaging.
Confocal images were taken on a Zeiss LSM510 (Carl Zeiss, Oberkochen, Germany) laser scanning confocal, using LSM imaging software. Epifluorescence images were taken on a Leica DMRXA compound microscope (Leica Microsystems, Wetzlar, Germany), using Meta-Morph software (Molecular Devices, Sunnyvale, CA).
For colocalization analysis and quantification of organelle size and intensity, MetaMorph (Molecular Devices) was used, and Microsoft Excel was used for statistical analysis. For each embryo measured, two 7 × 7 μm boxes were drawn, and all compartments within each box were measured for area and intensity per unit area. For YP170::GFP and LMP-1::GFP experiments, one z-section for each of five embryos of each genotype was measured. For PGP-2 and UV experiments one z-section of five embryos of each genotype was measured. We used two-tailed t-test p-values to determine significance in all experiments. Figs. 1, 2, 3 and 4 and Additional files were assembled using Adobe Photoshop (Adobe Systems Incorporated, San Jose, CA), and Fig. 5 was assembled using Canvas (ACD Systems, Victoria, BC Canada).