Curcumin Analogue C1 Promotes Hex and Gal Recruitment to the Plasma Membrane via mTORC1-Independent TFEB Activation

The monocarbonyl analogue of curcumin (1E,4E)-1,5-Bis(2-methoxyphenyl)penta-1,4-dien-3-one (C1) has been used as a specific activator of the master gene transcription factor EB (TFEB) to correlate the activation of this nuclear factor with the increased activity of lysosomal glycohydrolases and their recruitment to the cell surface. The presence of active lysosomal glycohydrolases associated with the lipid microdomains has been extensively demonstrated, and their role in glycosphingolipid (GSL) remodeling in both physiological and pathological conditions, such as neurodegenerative disorders, has been suggested. Here, we demonstrate that Jurkat cell stimulation elicits TFEB nuclear translocation and an increase of both the expression of hexosaminidase subunit beta (HEXB), hexosaminidase subunit alpha (HEXA), and galactosidase beta 1 (GLB1) genes, and the recruitment of β-hexosaminidase (Hex, EC 3.2.1.52) and β-galactosidase (Gal, EC 3.2.1.23) on lipid microdomains. Treatment of Jurkat cells with the curcumin analogue C1 also resulted in an increase of both lysosomal glycohydrolase activity and their targeting to the cell surface. Similar effects of C1 on lysosomal glycohydrolase expression and their recruitment to lipid microdomains was observed by treating the SH-SY5Y neuroblastoma cell line; the effects of C1 treatment were abolished by TFEB silencing. Together, these results clearly demonstrate the existence of a direct link between TFEB nuclear translocation and the transport of Hex and Gal from lysosomes to the plasma membrane.

TFEB recognizes and binds to a regulatory sequence, the coordinated lysosomal expression and regulation (CLEAR) motif, which is present in the promoter region of several lysosomal genes [8]. TFEB modulates and coordinates the main lysosome-dependent degradative pathways to promote intracellular clearance. TFEB activity depends on its phosphorylation status, which is mainly regulated exocytosis [31,36]. In a recent report, Medina and collaborators [37] demonstrated that lysosomal exocytosis is regulated by TFEB. Interestingly, it has been shown that TFEB nuclear translocation induces Hex and Gal recruitment to the plasma membrane in HEK-293 cells [17]. It is noteworthy that HEXB, HEXA, and GLB1 genes contain the CLEAR element in their promoter, making these genes putative targets of TFEB [1]. In order to investigate if T-cell activation promotes TFEB nuclear translocation, resting and phytohaemagglutinin (PHA)-stimulated Jurkat cells were treated as indicated in 'Materials and Methods', and cytosolic and nuclear fractions were blotted with TFEB antibody. β-Actin and H3 were used as cytosolic and nuclear markers, respectively. As reported in Figure 1A, the densitometric analysis of immunoblotting shows an increase in TFEB nuclear expression levels in PHA-stimulated with respect to resting Jurkat T-cells (p < 0.001), indicating the translocation of TFEB to the nucleus upon cell activation. To verify if TFEB activation, induced by T-cell stimulation, was able to promote lysosomal exocytosis, the activity of secreted horseradish peroxidase (HRP) on the culture medium after cell stimulation was evaluated. Jurkat cells were treated with HRP and then stimulated using PHA. The results reported in Figure 1B show an increase in secreted HRP activity of approximately 1.6-fold in PHA-stimulated compared to resting cells (p < 0.001).

External Leaflet Microdomain-Associated Hex and Gal Increase after Jurkat Cell Stimulation
A great deal of evidence indicates that gangliosides associated with lipid microdomains are involved in T-cell activation and they segregate in distinct T-cell subsets following cell stimulation, resulting in asymmetric specific redistribution [25]. As previously reported [31], Jurkat Tlymphocyte stimulation up-regulates the expression and activity of both Hex and Gal and increases their targeting to lipid microdomains where they may participate in the local reorganization of GSL. To verify if TFEB activation, induced by T-cell stimulation, was able to promote lysosomal exocytosis, the activity of secreted horseradish peroxidase (HRP) on the culture medium after cell stimulation was evaluated. Jurkat cells were treated with HRP and then stimulated using PHA. The results reported in Figure 1B show an increase in secreted HRP activity of approximately 1.6-fold in PHA-stimulated compared to resting cells (p < 0.001).

External Leaflet Microdomain-Associated Hex and Gal Increase after Jurkat Cell Stimulation
A great deal of evidence indicates that gangliosides associated with lipid microdomains are involved in T-cell activation and they segregate in distinct T-cell subsets following cell stimulation, resulting in asymmetric specific redistribution [25]. As previously reported [31], Jurkat T-lymphocyte stimulation up-regulates the expression and activity of both Hex and Gal and increases their targeting to lipid microdomains where they may participate in the local reorganization of GSL.
Quantitative PCR showed that there was an increase of HEXB, HEXA, and GLB1 mRNA levels in stimulated Jurkat cells compared to resting cells (Figure 2A). Quantitative PCR showed that there was an increase of HEXB, HEXA, and GLB1 mRNA levels in stimulated Jurkat cells compared to resting cells (Figure 2A).
.  Moreover, total Hex, Hex A, and Gal activity in crude extract from stimulated cells was 1.5, 1.4, and 1.6-fold higher compared to resting cells, respectively, according to our previous publication [31].
To determine if the increase in Hex and Gal activity also concerns the plasma membrane-associated forms, lipid microdomains from stimulated and resting cells were isolated using a discontinuous sucrose-density gradient. Fractions collected from the top to the bottom of the tube were tested by immunoblotting analysis for the presence of the microdomain markers flotillin-2 (flot-2) and the lymphocyte-specific protein tyrosine kinase (lck). As shown in Figure 2B, flot-2 and lck were highly enriched in the light-density fractions 2-4.
The collected fractions were also assayed for the activity of Hex, both Total Hex and the Hex A isoform, using the 4-methylumbelliferyl-N-acetyl-β-D-glucosaminide (MUG) and the 4-methylumbelliferyl-N-acetyl-β-D-glucosaminide-6-sulphate (MUGS) substrates, respectively, and Gal using the 4-methylumbelliferyl-b-D-galactopyranoside (MUGal) substrate. As reported in Figure 2C, Total Hex, Hex A and Gal showed a peak of enzymatic activity corresponding to fraction 3, which co-distributed with the lipid microdomain markers. Furthermore, the increase of Total Hex, Hex A, and Gal activity in light-density fraction 3 of stimulated Jurkat cells was 2.6, 3.0, and 2,6-fold higher compared to resting cells, respectively.
As gangliosides are inserted into the external leaflet of membranes, we investigated the Hex and Gal localization in the outer leaflet of plasma membrane-lipid microdomains. For this purpose, cell surface biotinylation of Jurkat cells followed by lipid microdomains isolation was carried out. Successively, lipid microdomain proteins were recovered from flot-2-positive fraction 3, and biotinylated proteins were recovered by avidin affinity chromatography in the eluate (fraction E), as shown in Figure 3A. Moreover, total Hex, Hex A, and Gal activity in crude extract from stimulated cells was 1.5, 1.4, and 1.6-fold higher compared to resting cells, respectively, according to our previous publication [31]. To determine if the increase in Hex and Gal activity also concerns the plasma membraneassociated forms, lipid microdomains from stimulated and resting cells were isolated using a discontinuous sucrose-density gradient. Fractions collected from the top to the bottom of the tube were tested by immunoblotting analysis for the presence of the microdomain markers flotillin-2 (flot-2) and the lymphocyte-specific protein tyrosine kinase (lck). As shown in Figure 2B, flot-2 and lck were highly enriched in the light-density fractions 2-4.
The collected fractions were also assayed for the activity of Hex, both Total Hex and the Hex A isoform, using the 4-methylumbelliferyl-N-acetyl-β-D-glucosaminide (MUG) and the 4methylumbelliferyl-N-acetyl-β-D-glucosaminide-6-sulphate (MUGS) substrates, respectively, and Gal using the 4-methylumbelliferyl-b-D-galactopyranoside (MUGal) substrate. As reported in Figure 2C, Total Hex, Hex A and Gal showed a peak of enzymatic activity corresponding to fraction 3, which co-distributed with the lipid microdomain markers. Furthermore, the increase of Total Hex, Hex A, and Gal activity in light-density fraction 3 of stimulated Jurkat cells was 2.6, 3.0, and 2,6-fold higher compared to resting cells, respectively.
As gangliosides are inserted into the external leaflet of membranes, we investigated the Hex and Gal localization in the outer leaflet of plasma membrane-lipid microdomains. For this purpose, cell surface biotinylation of Jurkat cells followed by lipid microdomains isolation was carried out. Successively, lipid microdomain proteins were recovered from flot-2-positive fraction 3, and biotinylated proteins were recovered by avidin affinity chromatography in the eluate (fraction E), as shown in Figure 3A.  The enzymatic assay of fraction E highlighted the presence of either Hex and Gal in both resting and stimulated cells, revealing their presence in the outer leaflet of plasma membrane lipid microdomains. The increase of Hex and Gal activities in stimulated Jurkat cells was also demonstrated as shown in Figure 3B. Furthermore, the presence of both Hex and Gal activity in the flow-through fraction (F) indicated that a portion of the lipid microdomain-associated enzymes was not confined on the cell surface but could be associated with the transit vesicles inside the cell.

TFEB Nuclear Translocation Induced by the Curcumin Analogue C1 Increases both the Expression of HEXB, HEXA, and GLB1 genes and the Recruitment of Glycohydrolases to Lipid Microdomains
In order to investigate if the observed increase of lysosomal glycohydrolases activity was correlated to the TFEB nuclear translocation rather than to an indirect effect of the Jurkat cell stimulation, we treated Jurkat cells with the curcumin analogue C1, a potent TFEB activator.
To this end, Jurkat cells were treated for 6 h with C1 at a concentration of 1 µM, according to [35]. The cytotoxicity of C1 in Jurkat cells was assessed by (3-(4,5-dimethylthiazol-2-yl)-2,5diphenyltetrazolium bromide) tetrazolium reduction (MTT) assay, and it was found that the compound was not toxic at the concentration used. As a positive control for TFEB nuclear translocation, cells were treated for 2 h with torin 1 at a concentration of 0.1 µM. As shown in Figure 4A, torin 1, PHA, and the curcumin analogue C1 induced TFEB nuclear translocation to a similar extent.
TFEB activation was confirmed by the increased gene expression of HEXB, HEXA, and GLB1 on C1-treated cells. Moreover, the expression of TFEB gene also increased after C1 treatment ( Figure 4B). As expected from the increase in HEXB, HEXA, and GLB1 mRNAs, the C1 treatment resulted in both an increase in Hex and Gal activity ( Figure 4C) and their recruitment to lipid microdomains as attested by the increase of their activities in flot-2 enriched fractions 2-4 ( Figure 4D). Therefore, it appears that Jurkat cell stimulation induces TFEB nuclear translocation, which in turn promotes an increase in Hex and Gal activity and their recruitment to the cell surface.

Curcumin Analogue C1 Promotes TFEB Activation without Inhibiting mTORC1 Activity
Since TFEB activation and the associated recruitment of glycohydrolases to the plasma membrane may be relevant for the treatment of pathological conditions such as neurodegenerative disorders [26,33], the effect of C1 on TFEB activity was also investigated on a neuronal cell model. SH-SY5Y cells were treated for 24 h with curcumin or its analogue C1 at concentrations of 5 and 1 µM, respectively. Curcumin and C1 cytotoxicity in SH-SY5Y cells was assessed by MTT assay (data not shown). TFEB nuclear translocation was determined by both immunoblotting and immunofluorescence analysis. As positive controls, cells were starved for 16 h in Hank's Balanced Salt Solution (HBSS) medium for immunoblotting analysis or treated for 2 h with torin 1 for immunofluorescence experiments. As reported in Figure 5A,B, TFEB nuclear translocation was activated by both C1 (p < 0.001) and, to a lesser degree, by curcumin (p < 0.01).

Curcumin Analogue C1 Promotes TFEB Activation without Inhibiting mTORC1 Activity
Since TFEB activation and the associated recruitment of glycohydrolases to the plasma membrane may be relevant for the treatment of pathological conditions such as neurodegenerative disorders [26,33], the effect of C1 on TFEB activity was also investigated on a neuronal cell model. SH-SY5Y cells were treated for 24 h with curcumin or its analogue C1 at concentrations of 5 and 1 μM, respectively. Curcumin and C1 cytotoxicity in SH-SY5Y cells was assessed by MTT assay (data not shown). TFEB nuclear translocation was determined by both immunoblotting and immunofluorescence analysis. As positive controls, cells were starved for 16 h in Hank's Balanced Salt Solution (HBSS) medium for immunoblotting analysis or treated for 2 h with torin 1 for immunofluorescence experiments. As reported in Figure 5A,B, TFEB nuclear translocation was activated by both C1 (p < 0.001) and, to a lesser degree, by curcumin (p < 0.01).  To demonstrate that C1 treatment of SH-SY5Y cells promotes TFEB nuclear translocation without effects on mTORC1 activity, we performed an immunoblotting analysis on phospho-S6 ribosomal protein (S235/236), which is known to be a p70S6K target. As reported in Figure 5C, 16 h of cell starvation strongly inhibited mTORC1 activity, as demonstrated by the reduced levels of phospho-S6 ribosomal protein. Curcumin treatment (24 h, 5 µM) showed a mild effect on the reduction of mTORC1 activity. However, even though it was possible to see a slight decrease in phospho-S6 ribosomal protein, there were no statistical differences between C1-treated (24 h, 1 µM) cells with respect to the control cells. Moreover, C1 treatment activated the autophagy flux in a manner similar to starvation as demonstrated by immunoblotting of microtubule-associated proteins 1A/1B light chain 3B (LC3B), which showed a significant increase in the autophagosome membrane-bound form LC3B-II ( Figure 5C).

Curcumin Analogue C1 Promotes an Increase in both Expression and Activity of Hex and Gal and Their Recruitment on the Cell Surface of SH-SY5Y Cells
To demonstrate that TFEB activation is correlated with the increased expression of Hex and Gal glycohydrolases, we performed quantitative analysis on mRNA and evaluated the enzymatic activity of Hex and Gal in SH-SY5Y cells treated with 1 µM of the curcumin analogue C1 for 24 h. As reported in Figure 6A, the increased levels of TFEB, HEXB, HEXA. and GLB1 mRNAs were correlated with the observed TFEB nuclear translocation promoted by C1 treatment. Moreover, as expected, the enzymatic activity of Hex, both Total Hex and Hex A, and Gal was increased after both curcumin and C1 treatments ( Figure 6B). To demonstrate that C1 treatment of SH-SY5Y cells promotes TFEB nuclear translocation without effects on mTORC1 activity, we performed an immunoblotting analysis on phospho-S6 ribosomal protein (S235/236), which is known to be a p70S6K target. As reported in Figure 5C, 16 h of cell starvation strongly inhibited mTORC1 activity, as demonstrated by the reduced levels of phospho-S6 ribosomal protein. Curcumin treatment (24 h, 5 μM) showed a mild effect on the reduction of mTORC1 activity. However, even though it was possible to see a slight decrease in phospho-S6 ribosomal protein, there were no statistical differences between C1-treated (24 h, 1 μM) cells with respect to the control cells. Moreover, C1 treatment activated the autophagy flux in a manner similar to starvation as demonstrated by immunoblotting of microtubule-associated proteins 1A/1B light chain 3B (LC3B), which showed a significant increase in the autophagosome membrane-bound form LC3B-II ( Figure 5C).

Curcumin Analogue C1 Promotes an Increase in both Expression and Activity of Hex and Gal and Their Recruitment on the Cell Surface of SH-SY5Y Cells
To demonstrate that TFEB activation is correlated with the increased expression of Hex and Gal glycohydrolases, we performed quantitative analysis on mRNA and evaluated the enzymatic activity of Hex and Gal in SH-SY5Y cells treated with 1 μM of the curcumin analogue C1 for 24 h. As reported in Figure 6A, the increased levels of TFEB, HEXB, HEXA. and GLB1 mRNAs were correlated with the observed TFEB nuclear translocation promoted by C1 treatment. Moreover, as expected, the enzymatic activity of Hex, both Total Hex and Hex A, and Gal was increased after both curcumin and C1 treatments ( Figure 6B).  To analyze the effect of the curcumin and C1 treatment on plasma membrane-associated glycohydrolase levels, we purified and recovered lipid microdomains. Fractions from the gradient were analyzed for the presence of the specific microdomain markers GM1, by dot blotting, and flot-2, by immunoblotting analysis. As shown in Figure 7A,B, GM1 and flot-2 were enriched in the light-density fractions 2-4 which were then collected and assayed for Hex, both Total Hex and Hex A, and Gal activities. As reported in Figure 7C, the activity of Hex and Gal was strongly increased in flot-2-positive fractions 2-4 of starved and C1-treated cells and, to a lesser extent, in curcumin-treated cells. To analyze the effect of the curcumin and C1 treatment on plasma membrane-associated glycohydrolase levels, we purified and recovered lipid microdomains. Fractions from the gradient were analyzed for the presence of the specific microdomain markers GM1, by dot blotting, and flot-2, by immunoblotting analysis. As shown in Figure 7A,B, GM1 and flot-2 were enriched in the lightdensity fractions 2-4 which were then collected and assayed for Hex, both Total Hex and Hex A, and Gal activities. As reported in Figure 7C, the activity of Hex and Gal was strongly increased in flot-2-positive fractions 2-4 of starved and C1-treated cells and, to a lesser extent, in curcumintreated cells. Finally, in order to unambiguously demonstrate that the effect of the curcumin analogue C1 on the recruitment of lysosomal glycohydrolases to the plasma membrane was associated with TFEB activation, SH-SY5Y cells were transfected with shRNA for TFEB. As shown in Figure 8B, C1 treatment of TFEB knock-down cells failed to promote the increase of Hex and Gal enzymatic activity. Moreover, after C1 cell treatment, the recruitment of the glycohydrolases to the cell surface, even if slightly increased with respect to the scramble cells, was significantly lower than in normal cells ( Figure 8C). These results strongly support the hypothesis of a correlation between TFEB activation and the recruitment of lysosomal glycohydrolases to the plasma membrane. Finally, in order to unambiguously demonstrate that the effect of the curcumin analogue C1 on the recruitment of lysosomal glycohydrolases to the plasma membrane was associated with TFEB activation, SH-SY5Y cells were transfected with shRNA for TFEB. As shown in Figure 8B, C1 treatment of TFEB knock-down cells failed to promote the increase of Hex and Gal enzymatic activity. Moreover, after C1 cell treatment, the recruitment of the glycohydrolases to the cell surface, even if slightly increased with respect to the scramble cells, was significantly lower than in normal cells ( Figure 8C).
These results strongly support the hypothesis of a correlation between TFEB activation and the recruitment of lysosomal glycohydrolases to the plasma membrane.

Discussion
Curcumin is a natural polyphenol, derived from the turmeric Curcuma longa, showing several pharmacological activities. Recently, it has been demonstrated that curcumin can induce autophagy through inhibition of the Akt-mTOR pathway [38] and by directly binding to TFEB, promoting its nuclear translocation [34].

Discussion
Curcumin is a natural polyphenol, derived from the turmeric Curcuma longa, showing several pharmacological activities. Recently, it has been demonstrated that curcumin can induce autophagy through inhibition of the Akt-mTOR pathway [38] and by directly binding to TFEB, promoting its nuclear translocation [34].
In this study, by using the human neuroblastoma cell line SH-SY5Y, which is a widely used cell model in the study of autophagic pathways and neurodegeneration [39][40][41], we confirmed that curcumin has an inhibitory effect on mTORC1 and promotes both lysosomal functions and autophagy by inducing TFEB nuclear translocation. Moreover, the curcumin-dependent TFEB activation was accompanied by an increase in both lysosomal enzyme Hex and Gal activity and their recruitment to the plasma membrane, where these enzymes may be involved in the in situ remodeling of GSL [42]. Notably, we previously demonstrated the alteration of lysosome-to-plasma membrane transport in the TgCRND8 mouse model of Alzheimer's disease, revealing an abnormal localization of glycohydrolases in post-synaptic density microdomains starting from the pre-symptomatic stage of the disease [26]. Interestingly, a role played by ganglioside metabolism on the pathogenesis of the disease has been reported [25]. Moreover, the involvement of TFEB in most neurodegenerative diseases has been extensively documented [11,12,15,16].
Moreover, in this study we demonstrated for the first time that PHA stimulation of Jurkat T-lymphocytes induced TFEB nuclear translocation and resulted in both an increase of the enzymatic activity of Hex and Gal and their targeting to the plasma membrane as promoted by lysosomal exocytosis. The treatment of Jurkat cells with the curcumin analogue C1, a potent mTORC1-independent TFEB activator [35], also induced TFEB nuclear translocation, increased Hex and Gal expression and activity, and enhanced plasma membrane-associated glycohydrolases in an extent very similar to the PHA-stimulated cells. These results clearly indicate that the TFEB nuclear translocation induced by Jurkat cell stimulation is closely associated with the delivery of lysosomal glycohydrolases to the cell surface and support the hypothesis of a possible involvement of TFEB in T-lymphocyte stimulation. Of note is that the implication of TFEB in the immune response has recently been reviewed [43], and the involvement of specific gangliosides in T-cell activation has also been reported [32]. Furthermore, our results suggest that lysosomal exocytosis may be relevant not only for cellular clearance and membrane repair but also for not yet clearly elucidated processes which require in situ remodeling of GSL.
The treatment of the neuroblastoma cell line SH-SY5Y with the curcumin analogue C1 also promoted TFEB nuclear translocation and the increase of lysosomal glycohydrolase expression and activity and their recruitment to the cell surface. Additionally, by a TFEB silencing experiment, we clearly demonstrated the ability of C1 to induce the transport of Hex and Gal to the plasma membrane by direct TFEB activation. In brief, the overall results (i) provide strong evidence of the correlation between TFEB nuclear translocation and the recruitment of Hex and Gal glycohydrolases to the plasma membrane; and (ii) demonstrate the ability of C1 to promote the recruitment of lysosomal glycohydrolases to the plasma membrane microdomains via an mTORC1-indipendent TFEB activation mechanism.
In conclusion, because C1 directly targets TFEB, we herein unambiguously demonstrated the link between TFEB activation, which promotes lysosome-to-plasma membrane fusion, and the transport of active forms of lysosomal glycohydrolases to the cell surface. The role of plasma membrane glycohydrolases has not yet been fully elucidated, but their implication in remodeling the glycosphingolipid pattern has been clearly confirmed [18,19]. Based on the role played by TFEB in recruiting lysosomal glycohydrolases to the plasma membrane, this transcription factor may represent an effective target to modulate this pathway.
Moreover, by using the SH-SY5Y neuroblastoma cell line, we confirmed the ability of the curcumin analogue C1 to activate TFEB without inhibiting mTORC1 activity, thus suggesting its potential therapeutic efficacy for the treatment of neurodegenerative diseases by promoting autophagy.

Drugs and Cell Treatments
The cells were treated for 24 h with the following drugs: PHA (1 mg/mL) from Sigma-Aldrich; curcumin (5 µM) from Sigma-Aldrich; and curcumin analogue C1 (1 µM). Cells were treated for 2 h with torin 1 (0.1 µM) from Cell Signaling.

TFEB RNA Interference
TFEB RNAi was performed by using shRNA expression constructs which were purchased from Origene (Rockville, MD, USA). A scrambled shRNA was used as the control. Briefly, SH-SY5Y cells were transfected using Lipofectamine LTX (Invitrogen). Stable transfected cells were obtained using 0.1 mg/mL puromycin (Sigma).

Cytosolic, Nuclear, and Enriched Plasma Membrane Extracts
Cytosolic and nuclear fractions were isolated from resting and PHA-stimulated Jurkat cells [17] and from SH-SY5Y cells treated for 24 h with curcumin and C1, or after starvation, as previously described. Plasma membrane proteins were isolated from soluble proteins using the Mem-PER Eukaryotic Membrane Protein Extraction Kit (Pierce, Thermo Scientific, Waltham, MA, USA) in accordance with the manufacturer's procedure. Protein concentration was determined by Bradford's assay.

Isolation of Lipid Microdomains
Lipid microdomains from Jurkat and SH-SY5Y cells were isolated by discontinuous sucrose-density gradient centrifugation as previously reported [31]. After centrifugation, eleven fractions of 450 µL were collected from the top to the bottom of the tubes.

Quantitative PCR
Total RNA was extracted from 2 × 10 6 Jurkat or SH-SY5Y cells using a standard Trizol protocol (Sigma-Aldrich). Isolated RNA was treated with a TURBO DNA-free™ Kit (ThermoFisher) according to the manufacturer's procedure. cDNA was obtained by reverse transcription of 2 µg of RNA using a Maxima H Minus First Strand cDNA Synthesis Kit (Thermo Fisher) according to the manufacturer's procedure. cDNA was used for the evaluation of HEXB, HEXA, GLB1, and TFEB gene expression by quantitative PCR (Q-PCR) in a Stratagene Mx3000P Q-PCR machine (Agilent Technologies) as previously reported [31]. Data were analyzed by the ∆∆Ct method. The sequences of specific primers used in this work are listed in Table 1.

Horseradish Peroxidase Assay
Jurkat cells (1 × 10 6 /mL) were treated for 2 h with 2 mg/mL of HRP (Sigma-Aldrich). Successively, cells were recovered, washed with Dulbecco's PBS (DPBS), and resuspended in complete media. Cells were incubated for 16 h and then stimulated or not (resting) with PHA for 24 h. The culture medium was recovered, and HRP activity was determined as previously reported [45]. One unit (U) corresponds to the amount of enzyme that oxidizes 1 µmol of substrate/min at pH 5.0 at 25 • C.

Isolation of Cell Surface Lipid Microdomain Proteins
Cell surface proteins of resting and PHA-stimulated Jurkat cells (1 × 10 8 ) were biotinylated using 1 mg/mL of EZ-Link Sulfo-NHS-LC-Biotin (Thermo Scientific), and the reaction was performed according to the manufacturer's procedure.
Cell surface lipid microdomain protein isolation was performed as previously reported [28]. Briefly, after discontinuous sucrose-density gradient centrifugation, flot-2-positive fraction 3 was recovered from the ultracentrifugation tube and diluted 1:4 with 10 mM Tris, 150 mM NaCl, 5 mM EDTA (TNE), pH 7.4, containing 1% (v/v) Triton X-100 (TX-100). Lipid microdomain vesicles were recovered by ultracentrifugation at 60,000 rpm at 4 • C for 2 h using a TLA-100.3 rotor and an Optima Max ultracentrifuge. The pellet was resuspended in 100 µl of PBS containing 1% (v/v) TX-100 and disaggregated by incubating for 10 min at 37 • C (LM3). Lipid microdomain proteins were loaded at the top of a 0.5 mL column containing the UltraLink™ Monomeric Avidin resin (Thermo Scientific). After washing, biotinylated proteins were eluted by 5 mM D-biotin in PBS.

Immunoblotting and Dot Blot Analysis
The protein extracts from cytosolic, nuclear, and sucrose-density gradient fractions were subjected to SDS-PAGE [46]. Separated proteins were transferred to the PVDF membrane (Biorad, Hercules, CA, USA), blocked in 50 mM Tris-HCl, 150 mM NaCl (TBS), pH 7.6, containing 5% (w/v) BSA/0.1% (v/v) Tween 20 and reacted over-night at 4 • C with one of the primary antibodies reported in Table 2. After being washed, blots were incubated with the appropriate HRP-conjugated secondary antibody and developed by an ECL detection system (GE Healthcare, Chicago, IL, USA). Western blot images were acquired using an ImageScanner calibrated densitometer (Amersham Pharmacia), and densitometry analysis was performed using ImageJ. GM1 ganglioside was revealed by Dot Blot analysis as reported in [31].

Determination of Enzyme Activities
Total Hex, Hex A, and Gal activities were determined by using the artificial substrates MUG, MUGS, and MUGal (Sigma-Aldrich) as previously reported [31].
One enzymatic unit (U) corresponds to the amount of enzyme that hydrolyses 1 mmol of substrate/min at 37 • C.

Immunofluorescence Analysis
SH-SY5Y cells were plated on glass coverslips, previously coated with poly-L-lysine (Sigma-Aldrich), for 30 min at RT, and incubated for 24 h in DMEM media before the drug treatments. The cells were fixed with 4% paraformaldehyde/DPBS for 20 min at RT, washed three times with DPBS, and blocked in DPBS containing 5% (v/v) FBS and 0.3% (v/v) TX-100 for 1 h at RT. After further washings, the cells were incubated for 1 h in the antibody solution (D-PBS, 1% (w/v) BSA, 0.3% (v/v) Tx-100) with the rabbit anti-TFEB primary antibody (1:600, Bethyl Laboratories, Cat. n. A303-673A). After being washed, the cells were incubated with the donkey anti-rabbit IgG Alexa Fluor ® 488 secondary antibody (Thermo-Fisher, Cat. n. A-21206) for 1 h in an antibody solution. Successively, the coverslips were mounted on glass slides using Vectashield with DAPI (Vector Laboratories Inc. Burlingame, CA, USA), and fluorescence microscopy analysis was performed using a Nikon TE2000 microscope (Nikon Instruments S.p.A, Florence, Italy). Image processing was performed by using Adobe Photoshop CS software (Adobe Systems Incorporated).

Statistical Analysis
All data were expressed as mean ± SEM. Statistical differences were evaluated using unpaired Student's t-test. The threshold for statistical significance was set at p < 0.05.

Conflicts of Interest:
The authors declare no conflict of interest.