|Home | About | Journals | Submit | Contact Us | Français|
Loss-of-function mutations in TRPML1 (Transient Receptor Potential Mucolipin 1) cause a lysosomal storage disorder called mucolipidosis type IV (MLIV). Previously, we established a Drosophila model for MLIV by knocking out the single TRPML1 homolog expressed in flies . The mutant animals displayed impaired autophagy, and reduced viability during the pupal period—a phase when wild-type animals rely on autophagic sources for nutrients. However, the specific defect in autophagy has remained unclear. Here, we show that TRPML, which was localized to the membranes of late-endosomes and lysosomes in vivo, was required for fusion between late-endosomes with lysosomes. We report that loss of TRPML led to accumulation of vesicles of significantly larger volume, and loss of TRPML from the late-endosomes/lysosomes led to the accumulation of higher luminal Ca2+ within those vesicles. We also found that trpml1 mutant cells showed decreased TORC1 signaling, and a concomitant upregulation of autophagy induction. Both of these defects were reversed by activating TORC1 in the mutants genetically and by feeding the mutant larvae a high-protein diet. Feeding the larvae a high-protein diet also reduced both the pupal lethality, and the increased volume of acidic vesicles. Conversely, further inhibition of TORC1 activity by rapamycin exacerbated the mutant phenotypes. Finally, TORC1 exerted reciprocal control on TRPML function. A high amino acid diet enhanced cortical localization of TRPML, and this effect was blocked by rapamycin. Our findings delineate the interrelationship between TRPML- and TORC1-mediated nutrient sensing pathways, and also raise the intriguing possibility that amino acid supplementation might reduce the severity of the clinical manifestations associated with MLIV.
Drosophila TRPML localizes to late-endosomes (LEs)/lysosomes in cultured cells . To identify the subcellular localization of TRPML in vivo, we expressed a UAS-trpml::myc transgene in flies using the GAL4/UAS system . TRPML::MYC decorated the periphery of LysoTracker positive vesicles (Figures 1A–C), and colocalized with the LE/lysosomal markers YFP::Rab7 (Figures S1A–C) and LAMP::GFP (Figures 1D–F). These data indicated that TRPML::MYC is a LE/lysosomal membrane protein, as is the case for mammalian TRPML1 .
The trpml1 flies are unable to complete lysosomal degradation of autophagosomes . To identify the step in the lysosomal degradation pathway affected in trpml1, we evaluated the degradation of the Drosophila Wnt homolog, Wingless (Wg) . Following binding to its receptor, Wg is internalized into endosomes, and degraded in lysosomes . We found that there was increased accumulation of Wg in the wing pouch and notum of trpml1 wing-discs (Figures 1G–H and 1K), and this phenotype was rescued by a trpml+ genomic transgene (P[trpml+];trpml1) (Figure 1K).
Wg could be accumulating either in early endosomes or in LEs/multivesicular bodies (MVBs) in trpml1 discs. To discriminate between these possibilities, we relied on the observation that Wg transmits signals at both the plasma membrane and early endosomes. Only after the formation of MVBs is the signal terminated. Therefore, increased Wg signaling in trpml1 would suggest that Wg is accumulating in early endosomes, while unchanged Wg signaling would be consistent with Wg accumulating in MVBs. Therefore, we evaluated activation of the Wg target gene Hindsight (Hnt) in wing-discs as described . Nuclear Hnt expression was indistinguishable between wild-type and trpml1 (Figure 1I–J) indicating that Wg signaling was not increased in trpml1.
We also evaluated accumulation of Notch using an antibody specific for the endocytosed domain of Notch, and found that levels of Notch increased dramatically in trpml1 wing-discs (Figures S1D–E and 1K). Notch levels appeared higher than Wg in trpml1 because while Wg accumulated in the wing pouch and notum, Notch was elevated over the whole disc. In support of the conclusion that the vesicles that accumulated in trpml1 were LEs, the Notch positive vesicles co-stained with LysoTracker (Figures S1F–H).
Autophagy is a pathway required for the degradation of cellular macromolecules that are too big to fit through the proteosomal barrel [6, 7]. During autophagy, double membrane bound vesicles called autophagosomes isolate the cytosolic material destined for degradation. Subsequently, autophagosomes fuse with LEs/MVBs to form amphisomes [8, 9]. Amphisomes then coalesce with lysosomes leading to the formation of autolysosomes. Since lysosomes carry degradatory enzymes, the contents of amphisomes are broken down following autolysosome formation [8, 9].
We previously reported that trpml1 adults display hallmarks of decreased autophagic flux . To provide evidence that there was accumulation of autophagosome and amphisomes, we stained wild-type and trpml1 fat bodies with LysoTracker and GFP::ATG8. Autophagosomes are labeled with GFP::ATG8 only, while amphisomes are stained with both GFP::ATG8 and LysoTracker. Although wild-type showed virtually no GFP::ATG8 staining (Figure 1L), there were many trpml1 vesicles that were labeled with GFP::ATG8 only or both GFP::ATG8 and LysoTracker (Figures 1M–N). These data indicated that loss of trpml led to an elevation of autophagosomes and amphisomes. Defects in receptor degradation have also been reported in human cells lacking TRPML1, and in C. elegans with a mutation disrupting the worm TRPML1 homolog [10–12]. These trpml1 phenotypes closely resemble those of flies lacking fab1 (vesicular Phosphatidylinositol 3-Phosphate 5-Kinase) . These phenotypic similarities are consistent with the finding that the mammalian TRPML1 is activated by the product of Fab1/PIK-FYVE-Kinase, PI(3,5)P(2) .
We also performed EM on larval fat-bodies to confirm the nature of the accumulating vesicles based on previously described morphological criteria . Wild-type fat-body cells contained large electron-dense, lysosome-like structures (Figures 2A and C). Many of these lysosomes were in the process of fusing, and appeared as “figure-8”-like structures without internal separations (Figure 2A and blue lines in Figure 2C). In contrast, trpml1 fat-body cells contained substantially fewer and smaller lysosomes (Figure 2B). This phenotype resembles that observed in deep orange (dor) mutant fat-bodies, which are characterized by diminished fusion of LEs with lysosomes . Furthermore, trpml1 mutant cells accumulated large single-membrane bound vesicles that contained smaller internal vesicles indicating that they were MVBs/ amphisomes (Figures 2B and D–E). Many of these vesicles were touching lysosomes, apparently unable to undergo fusion (Figures 2D–E). We propose that these “fusion clamped” vesicles arise from a defect in fusion of MVBs and/or amphisomes with lysosomes in trpml1. The trpml1 fat-bodies also showed fusion clamped lysosomes (Figure 2F), indicating that both homotypic and heterotypic fusion of vesicles were disrupted.
To quantify the increased vesicle accumulation in trpml1, we performed EM on adult photoreceptor cells (PCs), which are more amenable to quantification than the cells of the fat-bodies because the cell boundaries are more clearly evident in the former. We found that while wild-type PCs showed very few MVBs (single membrane bound vesicles with internal vesicular structures) and lysosomes (electron-dense multilamellar structures) per section, trpml1 PCs displayed a dramatic elevation of both MVBs and lysosomes (Figures S2A–B and S2G). The mutant cells also showed increased accumulation of autophagosomes (double membrane bound vesicles containing cytosolic material) (Figure S2G). Higher levels of autophagosomes in adult PCs was consistent with the finding in larval fat-bodies. The relative numbers of autolysosomes, which formed after fusion of MVBs and lysosomes (single membrane bound vesicles containing internal vesicular structures and multilamellar electron-dense lysosomes), were not significantly different in wild-type and trpml1 cells despite of an increase in amphisomes and lysosomes (Figure S2G). Furthermore, the ratio of MVBs/autolysosomes was 10-fold higher in trpml1 compared to wild-type (11.7 and 1.2, respectively; Figure S2G). The trpml1–PCs also contained many MVBs that were touching lysosomes (2.2 ±0.8 fusion-clamped vesicles/ommatidia in trpml1 mutants) (Figures S2C–G). The number of these “fusion-clamped” vesicles was significantly lower in wild-type cells (0.1 ±0.1 fusion-clamped vesicles/ommatidia, p=0.04, Student’s t-test) (Figure S2G). Taken together, our data point to a defect in the fusion of MVBs with lysosomes in trpml1 cells. Furthermore, since the total number of autolysosomes is unchanged in the mutants, we suggest that they display an additional defect in lysosomal degradation.
Mammalian TRPML1 is a LE/lysosomal channel responsible for the release of luminal Ca2+ [13, 15, 16]. Therefore, loss of TRPML1 might lead to elevated LE Ca2+ levels. To investigate whether elimination of Drosophila trpml led to elevated LE Ca2+ levels, we measured the releasable pool of LE luminal Ca2+ in Fura-2 loaded fat-bodies. For releasing LE Ca2+ we treated the fat-bodies with bafilomycin A1—a cell-permeant inhibitor of the vacuolar-type H+-ATPase. Treatment with bafilomycin A1 results in loss of the LE H+ gradient and eventual depletion of LE Ca2+, which increases the cytosolic free Ca2+ concentration . Consistent with the notion that the trpml1 mutants had elevated LE Ca2+ levels, bafilomycin A1 treatment caused a significant elevation in cytosolic free Ca2+ in trpml1 compared to wild-type (0.03 ±0.006 and 0.15 ±0.05 in wt and trpml1 respectively) (Figure S3A and Figures 3A–B). These data indicated that trpml1 LEs contained higher levels of Ca2+ due to diminished Ca2+ release.
We compared the sizes of LysoTracker positive vesicles in wild-type and trpml1 larval fat-bodies to assess autophagy induction . Consistent with increased induction of autophagy in trpml1, there was a striking elevation in the volume of LysoTracker-positive vesicles in mutant fat-bodies (Figures 3C–D and F). This change became evident, and was most pronounced in fat-bodies from 2nd instar larvae (8.14 ±1.7-fold larger in trpml1) (Figure 3F). The difference between wild-type and trpml1 was less pronounced but still significant in 3rd instar larvae (2.77 ±0.5-fold increase in trpml1) (Figure 3F). The smaller elevation in trpml1 3rd instar larvae likely reflects ecdysone-dependent autophagy activation in wild-type tissues at this developmental stage [14, 19].
We hypothesized that induction of autophagy without its completion should suppress TORC1 activity due to two factors. First, a decline in autophagic flux would decrease net availability of amino acids that are produced via autophagic degradation of proteins . Reduced amino acid levels would diminish activity of the Target of Rapamycin Complex 1 (TORC1) [21–23]. Indeed, diminished autophagic flux by Atg7 knockdown led to reduced TORC1 activity as determined by phosphorylation of the TORC1 substrate, S6-kinase (pS6K) [21, 24]. We found a similar decrease in TORC1 activity after knocking down Atg5 in wild-type fat-bodies using RNAi (Figure S3B–C). Second, increased autophagy will directly suppress TORC1 function because autophagy and TORC1 activity are mutually antagonistic . Decreased TORC1 will induce further induction of autophagy leading to the generation of larger LysoTracker-positive vesicles.
Several lines of evidence support the preceding proposal. First, feeding trpml1 3rd instar larvae protein-rich yeast paste suppressed the increase in the volume of LysoTracker-positive vesicles (Figures 3E–F). Second, phosphorylation of S6-kinase was diminished in trpml1 fat bodies (Figure 3G–H). The decrease in pS6K in trpml1 was reversed by driving wild-type trpml+ in fat-bodies using cg-GAL4  (Figure 3H). Furthermore, yeast-feeding suppressed the reduction in pS6K levels in trpml1 (Figure 3G–H). To investigate whether decreased TORC1 activity occurs in other mutants with diminished fusion of LEs with lysosomes, we evaluated pS6K phosphorylation in dor mutants [14, 26]. Extracts from dor4 mutant larvae showed a reduction in pS6K (Figure 3H). Therefore, a block in the fusion of LEs with lysosomes results in decreased in cellular amino acid levels and decreased TORC1 activity.
Third, genetically up-regulating TORC1 activity in mutant fat-bodies by overexpressing Rheb and constitutively active Rag (RagQ61L) [27, 28] decreased the LysoTracker-positive vesicular volume (Figure S3D). Vesicular volume in trpml1 did not increase any further when we expressed dominant negative Rag (RagT16N)  indicating that Rag activity was already maximally reduced in trpml1. The finding that elevating TORC1 activity is sufficient to suppress lysosomal storage argues that the increase in acidic vesicles in trpml1 reflects a decrease in TORC1 activity. Therefore, elevating TORC1 activity is sufficient to prevent vesicle accumulation despite the persistence of vesicle fusion defects.
Fourth, the half-maximal time to pupation was increased in trpml1 (Figure S3E). Since decreased TORC1 activity causes a developmental delay , these data are also consistent with a decrease in TORC1 activity in trpml1. Feeding the larvae protein-rich yeast paste restored pS6K levels to wild-type (Figures 3G–H) and rescued the defect in developmental timing (Figure S3E).
LE/lysosomal Ca2+ is required for the homotypic and heterotypic fusion of these vesicles . Therefore, the increased LysoTracker staining in trpml1 may have resulted from diminished Ca2+ release from the vesicles leading to impaired fusion of the LEs/amphisomes with lysosomes, ultimately resulting in reduced degradation of their contents. Consistent with the proposal that increased vesicular volume stemmed from diminished Ca2+ release, treatment of the fat-bodies with thapsigargin, which blocks the SERCA pump and causes Ca2+ release from ER stores, resulted in a significant decrease in the volume of LysoTracker positive vesicles in trpml1 (Figures 3I–K). Therefore, despite the absence of a LE Ca2+ release mechanism in trpml1, elevation of cytosolic Ca2+ levels from a different Ca2+ reserve was sufficient to partially suppress the increase in LysoTracker-positive vesicle volume. Although our data are most consistent with a defect in the fusion of vesicles in trpml1, we cannot rule out that there may also be a defect in vesicular trafficking, thereby reducing encounters between fusible vesicles.
During the pupal period Drosophila do not feed, and depend on autophagy for the amino acids necessary for morphogenesis and survival. Loss of trpml causes semi-lethality during the pupal period, as <10% of adults eclose from the pupal cases (Figures 4A and S4C) . To test whether this reduced viability resulted from an insufficient supply of amino acids, we fed the mutant larvae a high-protein diet (food supplemented with 20% w/v yeast). We found that this diet significantly suppressed the lethality (Figure 4A and S4C). However, the effects of another mutation that caused pupal-lethality (P-element inserted in vamp-7, CG1599EY09354; Figure S4A–B) were not suppressed by yeast supplementation (Figure S4C).
The suppression of the trpml semi-lethality by yeast paste could have been due to either protein or carbohydrates in this supplement. Therefore, we tested whether supplementation of either tryptone or sucrose diminished the lethality. We found that while tryptone supplementation reduced the semi-lethality, sucrose supplementation did not (Figure 4B and S4D). The lack of suppression with sucrose indicated that the phenotype was not a result of caloric deprivation, but rather reflected a requirement for increased dietary amino acids by trpml1 larvae.
Next, we considered whether the suppression of the semi-lethality by the high-protein diet was due to increased TORC1 activity. We fed yeast paste to trpml1 larvae in the presence of the TORC1-inhibitor, rapamycin. We found that rapamycin prevented suppression of the pupal semi-lethality by yeast paste (Figure S4E). Moreover, rapamycin enhanced the lethality when trpml1 larvae were reared on normal food (Figure 4C). However, feeding the dor mutants rapamycin did not decrease their viability (data not shown). These data indicate that not all mutants with deficient fusion of LEs to lysosomes show increased sensitivity to rapamycin.
TORC1 simultaneously increases protein translation and decreases autophagy. One of the ways through which TORC1 increases protein translation is phosphorylation and inhibition of the translational suppressor Thor—fly homolog of 4E-BP1 [31, 32]. Therefore, if the effects of TORC1 activation in trpml1 depend upon protein translation, then rapamycin should not reverse the beneficial effect of yeast feeding in thor2;trpml1 double mutant animals. However, yeast paste still suppressed the semi-lethality in thor2;trpml1 double mutants (no thor2;trpml1 adults eclosed in the absence of yeast supplemented diet), and the effect of rapamycin remained unchanged (Figure S4E). These results suggest that activation of TORC1 by high levels of amino acids may have suppressed the pupal semi-lethality of trpml1 by decreasing autophagy rather than by increasing protein translation.
To investigate whether TORC1 activity reciprocally affected TRPML, we examined the spatial distribution of TRPML::MYC under conditions in which we manipulated TORC1 activity. On a normal diet, TRPML::MYC was exclusively intracellular (Figure 4D). However, on a high protein diet, TRPML::MYC colocalized with the cortical F-actin marker phalloidin (Figure 4D), indicating that TRPML was at the plasma membrane (PM). In larvae maintained on a high-protein diet and rapamycin, we detected TRPML::MYC exclusively in intracellular vesicles (Figure 4E). These data indicate that the PM localization of TRPML::MYC depends on the activity of TORC1. Further supporting this conclusion, TRPML::MYC was predominantly localized to the PM in larval salivary glands (Figure S4F), which are characterized by low levels of autophagy (and therefore high TORC1 activity) till the onset of the pupal phase, when autophagy is required for the degradation of the pupal salivary gland during metamorphosis .
The effect of TORC1 activity on the subcellular location of TRPML is unlikely to reflect alterations in bulk endocytosis because TORC1 enhances rather than suppresses bulk endocytosis [34, 35]. Rather, we suggest that by occluding entry of TRPML into endosomes and diminishing the levels of TRPML in the LEs, TORC1 exerts feedback regulation on the completion of autophagy. Therefore, in addition to suppressing the initiation of autophagy , TORC1 also inhibits the completion of autophagy by regulating the subcellular location of TRPML.
TRPML is a Ca2+ channel, which is endocytosed from the PM and eventually enters the LEs (Figure 4F). The LEs fuse with autophagosomes creating amphisomes. TRPML present in amphisomes releases luminal Ca2+ to facilitate Ca2+ dependent fusion of amphisomes with lysosomes. The amino acids generated by degradation of proteins in the autolysosomes promote TORC1 activation. In addition to inhibiting the initiation of autophagy, activated TORC1 also diminishes endocytosis of TRPML.
In the absence of TRPML, fusion of amphisomes and lysosomes is impaired. This leads to a decrease in autophagic flux of amino acids causing a reduction in TORC1 and up regulation of autophagy. Supplementing the trpml1 diet with protein-rich yeast reverses the effects of diminished TORC1 activity. These findings raise the possibility that MLIV patients may also show diminished TORC1 activity. If so, it is intriguing to speculate that amino acid supplementation might reduce the severity of the clinical manifestations associated with MLIV.
We thank the Bloomington Stock Center for fly stocks, and FlyBase for valuable genome information. We also thank Geoffrey Broadhead and Hongxiang Hu for technical assistance, the IBP Cytodynamic Imaging Facility, and Dr. Hugo Bellen for providing reagents. C.M. was supported by grants from the March of Dimes and the National Eye Institute (EY10852).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Supplemental Information includes three figures and Supplemental Experimental Procedures and can be found with this article online.
The authors declare no conflicts of interest.