SLC17A9 Protein Functions as a Lysosomal ATP Transporter and Regulates Cell Viability*

Background: Lysosomes contain abundant ATP. Results: SLC17A9 transports ATP into lysosomes. SLC17A9-deficient cells exhibit lysosome dysfunction and undergo subsequent cell death. Conclusion: SLC17A9 functions as a lysosomal ATP transporter and maintains cell viability. Significance: This work identifies a molecular mechanism underlying lysosomal ATP accumulation and suggests that ATP functions not only as energy currency but plays a key role in lysosome function. Lysosomes contain abundant ATP, which is released through lysosomal exocytosis following exposure to various stimuli. However, the molecular mechanisms underlying lysosomal ATP accumulation remain unknown. The vesicular nucleotide transporter, also known as solute carrier family 17 member 9 (SLC17A9), has been shown to function in ATP transport across secretory vesicles/granules membrane in adrenal chromaffin cells, T cells, and pancreatic cells. Here, using mammalian cell lines, we report that SLC17A9 is highly enriched in lysosomes and functions as an ATP transporter in those organelles. SLC17A9 deficiency reduced lysosome ATP accumulation and compromised lysosome function, resulting in cell death. Our data suggest that SLC17A9 activity mediates lysosomal ATP accumulation and plays an important role in lysosomal physiology and cell viability.

Lysosomes contain abundant ATP, which is released through lysosomal exocytosis following exposure to various stimuli. However, the molecular mechanisms underlying lysosomal ATP accumulation remain unknown. The vesicular nucleotide transporter, also known as solute carrier family 17 member 9 (SLC17A9), has been shown to function in ATP transport across secretory vesicles/granules membrane in adrenal chromaffin cells, T cells, and pancreatic cells. Here, using mammalian cell lines, we report that SLC17A9 is highly enriched in lysosomes and functions as an ATP transporter in those organelles. SLC17A9 deficiency reduced lysosome ATP accumulation and compromised lysosome function, resulting in cell death. Our data suggest that SLC17A9 activity mediates lysosomal ATP accumulation and plays an important role in lysosomal physiology and cell viability.
The solute carrier (SLC) 2 group of membrane transport proteins includes over 300 members subdivided into 52 families (1,2). These proteins are located in the plasma membrane and the membranes of intracellular organelles where they transport a variety of solutes across cell membranes. Among these, the SLC17 family mediates activities requiring transmembrane transport of organic anions, including urate secretion (SLC17A1 and SLC17A3), lysosomal export of sialic acid (SLC17A5), and vesicular accumulation of aspartate (SLC17A5) and glutamate (SLC17A5-8). The most recently described family member, SLC17A9, represents a vesicular nucleotide transporter (VNUT) and loads ATP into secretory vesicles in adrenal chromaffin cells (3), T cells (4), and zymogen granules of pancreatic cells (5).
Recent studies indicate that lysosomes store large amounts of ATP (6,7). However, the molecular mechanisms underlying lysosomal ATP accumulation have not been identified. Because cells contain high (millimolar) levels of ATP in the cytosol, one mechanism could be that lysosomes take up ATP from the cytosol through an ATP transporter located in the lysosomal membrane. Several molecules have been suggested to mediate that activity, including connexins, pannexins, the ATP-binding cassette transporter, maxi-anion channels, and SLC proteins (3,8,9). The ATP receptor P2X7 is also implicated in ATP release across the plasma membrane (10). Because SLC17A9 has been shown to transport ATP in intracellular secretory vesicles, we hypothesized that it could also function in lysosomal ATP transport. Here we report that SLC17A9 is highly enriched in lysosomes and actively transports ATP into lysosomes. Compromised SLC17A9 function impaired lysosomal function, subsequently causing lysosomal storage and cell death. Our data suggest a molecular mechanism for lysosomal ATP accumulation and a role for lysosomal ATP in cell viability.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfection-COS1, HEK293T, and C2C12 cells were obtained from the ATCC and maintained in DMEM:nutrient mixture F-12 (F12) supplemented with 10% fetal bovine serum (Invitrogen). Cells were cultured at 37°C in a 5% CO 2 atmosphere. For some experiments, cells were seeded on 0.01% poly-lysine-coated coverslips and cultured 24 h before further experiments. COS1 and HEK293T cells were transfected using Lipofectamine 2000. C2C12 cells were transfected by electroporation with the Neon transfection system (Invitrogen) following the optimized protocol of the manufacturer (1650 V, 30 ms, 3 pulses). An ϳ80% transfection efficiency was achieved regularly.
Immunocytochemistry-Cells grown on coverslips were washed with PBS twice and fixed in 4% paraformaldehyde in PBS for 15 min at room temperature. Fixed cells were permeabilized with 0.1% Triton X-100 in PBS for 5 min and then blocked with 3% BSA in PBS for 60 min at room temperature. After three PBS washes, cells were incubated with primary antibodies at 4°C overnight. After three more PBS washes, cells were incubated with fluorescence-conjugated secondary antibodies for 45 min at room temperature in the dark. Images were acquired using a confocal microscope (LSM510, Zeiss) with a ϫ63 oil immersion objective lens and captured using ZEN2009 software (Zeiss).
Confocal Microscopy-Confocal fluorescent images were taken using an inverted Zeiss LSM510 confocal microscope with ϫ10 plain or ϫ63 oil immersion objectives. Sequential excitation wavelengths at 488 nm and 543 nm were provided by argon and helium-neon gas lasers, respectively. BP500 -550 and LP560 emission filters were used to collect green and red images in channels 1 and 2, respectively, and then images of the same cell were saved and analyzed with ZEN software. Here the term colocalization refers to the coincident detection of an above-background green and red fluorescent signal. Image size was set at 1024 ϫ 1024 pixels.
Lysosome Isolation by Subcellular Fractionation-Lysosomes were isolated as described previously (30,31). Briefly, cell lysates were obtained by Dounce homogenization in a homogenizing buffer (0.25 M sucrose, 1 mM EDTA, 10 mM HEPES (pH 7.0)) and then centrifuged at 1500 ϫ g (4200 rpm) (Fisher, ST-16R, F15 rotor) at 4°C for 10 min to remove nuclei and intact cells. Postnuclear supernatants were then fractionated on a Percoll density gradient using a Beckman Optima L-90K ultracentrifuge. For that analysis, the ultracentrifuge tube was layered with 2.5 M sucrose (bottom), 18% Percoll in homogenizing buffer (middle), and sample (top). Samples were centrifuged at 90,000 ϫ g (31,300 rpm) at 4°C for 1 h using a Beckman Coulter 70.1 Ti rotor to obtain light, medium, and heavy membrane fractions. Heavy membrane fractions containing cellular organelles were further layered over a discontinuous iodixanol (v/v; 27, 22.5, 19, 16, 12, and 8%) gradient generated by mixing iodixanol in homogenizing buffer with 2.5 M glucose. The osmolarity of all solutions was 300 mosmol. After centrifugation at 4°C for 2.5 h at 180,000 ϫ g (44,200 rpm), the sample was divided into 12 fractions (0.5 ml each) for further analyses. The biological properties and ionic composition of lysosomes were largely maintained during this procedure because of slowed transport across lysosomal membranes occurring at 4°C.
Western Blotting-Proteins were analyzed using standard methods. Proteins derived from whole cells or lysosome lysates were resolved on 10% SDS-PAGE. Proteins were then transferred onto PVDF membranes using a semidry transfer apparatus (Bio-Rad). Nonspecific binding was blocked using 5% skim milk in TBS-T (0.1% Tween 20 in 1ϫ Tris-buffered saline (pH 7.4)) for 2 h at room temperature. Membranes were then incubated with specific primary antibody solution at 4°C overnight with gentle constant shaking. After three thorough TBS-T washes, the membranes were incubated with corresponding HRP-conjugated second antibody at room temperature for 1 h. Immunoreactive bands were visualized using Clear ECL (Bio-Rad) and autoradiography.
Staining for ATP-Cells were pretreated with various transporter inhibitors (CBX, 20 M; NFA, 20 M; DIDS or Evans Blue, 5 M or 20 M, respectively) for 3 h and then incubated with quinacrine (5 M) together with LysoTracker Red DND-99 (50 nM) for 30 min at 37°C and chased for 1 h. Images were acquired using a confocal microscope (Zeiss) with a ϫ63 oil immersion objective and then collected and analyzed using ZEN2009 (Zeiss). A low laser power (Ͻ 0.5%) was used to minimize fluorescence bleaching.
Bioluminescence Measurement of ATP and ATP Uptake Assay-The ATP content of the culture medium or buffer containing freshly isolated lysosomes was measured in triplicate with a microplate luminometer (Fluoroskan Ascent FL microplate fluorometer and luminometer, Thermo Scientific) using the adenosine 5Ј-triphosphate bioluminescent assay kit (FLAA, Sigma) according to the instructions of the manufacturer. A standard curve was established for each measurement. Samples were adjusted to pH 7.8 to establish conditions optimal to detect ATP. All samples were kept on ice before measurement. Briefly, 100 l of the ATP assay mix solution was added to the assay tube and allowed to stand at room temperature for 3 min to hydrolyze endogenous ATP. Then, 100 l of the sample was added rapidly, mixed gently, and then the plate was read immediately using a luminometer. Blank medium or buffer samples were read in parallel as background, which was subtracted from sample readings. For some experiments, freshly isolated lysosome (1-5) and mitochondrial (8 -12) fractions were treated with 200 M GPN or 1% TritonX100 for 5 min at 4°C (to minimize ATP degradation) to lyse lysosomes or mitochondria. Samples containing same amount of protein were loaded for the ATP assay (75 g protein/well).
For ATP uptake assays, 0.2 l of 2 mM ATP was added to 100 l of samples (containing 75 g of total protein) for an initial ATP concentration of 2 ϫ 10 Ϫ6 M. Samples luminescence was read over time, and the luminescence decay indicated active ATP uptake. Sample buffer served as a blank, and the decay of intrinsic luminescence was measured as background. In some experiments, freshly isolated lysosomes from cells pretreated with the inhibitors described above for 30 min at 37°C were assayed immediately assayed for ATP uptake in the presence of those inhibitors. At the end of the assay, 0.1% TritonX100 was added to each sample to measure the total ATP released from vesicles.
Cell Viability Assay-Briefly, cells were transfected with the indicated DNA using Lipofectamine 2000 or electroporation. Cells were pretreated overnight with the inhibitors described above. Cells were stained with propidium iodide (Sigma, P-4170) (10 g/ml) and incubated at room temperature in PBS for 15 min. Cells were then examined immediately by fluorescence microscopy (emission at 615 nm and excitation at 488 nm). For quantitative analysis, adherent cells were collected by trypsinization, and PI ϩ and GFP ϩ cells and GFP ϩ /PI ϩ cells were counted using a hemocytometer under an upright fluorescent microscope.
For some assays, lactate dehydrogenase activity was determined by measuring NADH oxidation with pyruvate as substrate. For each assay, a 50-l sample was incubated with 50 l of reaction buffer (CytoTox-ONE TM homogeneous membrane integrity assay, Promega) for 10 min at 22°C, and the fluorescence at 340 nm was measured using a spectrophotometer. Results were scaled to the complete cell death induced by exposure to 1% Triton X-100 at room temperature for 20 min. The percent of cell death was defined as 100 ϫ (sample optical den-sityϪ medium optical density)/(Triton X-100 optical densityϪ medium optical density).
Data Analysis-Data are presented as mean Ϯ S.E. Statistical comparisons were made using analysis of variance and Student's t test. p Ͻ 0.05 was considered to be statistically significant.

Inhibition of ATP Transport across the Lysosome Membrane by DIDS and Evans
Blue-To directly measure ATP transport across the lysosomal membrane, we isolated lysosomes from HEK cells using density gradient centrifugation and assayed ATP uptake. We monitored ATP transport by evaluating ATP concentrations in the buffer using a bioluminescent ATP assay (9). A decrease of ATP concentration in the buffer suggested ATP uptake by lysosomes. We compared the luminescence decay over time following the repeated addition of ATP (2 M) to lysosomes. As shown in Fig. 1A, control samples not treated with inhibitors showed ATP decrease rates of 12.26 Ϯ 0.07%/ min after the first ATP addition (11-19 min). Interestingly, that rate slowed to 5.00 Ϯ 0.46%/min when lysosomes were pretreated with DIDS (20 M) or Evans blue (20 M), two known inhibitors of mammalian VNUT (3), for 30 min, and ATP was added in the presence of the inhibitors. In contrast, treatment of samples with CBX (20 M) or NFA (20 M), two inhibitors targeting both connexin and pannexin channels (11), had no effect.
For comparison, buffer ATP concentration 19 min after the first ATP application and 29 min after the second were replotted in Fig. 1B. Treatment with DIDS or Evans blue but not CBX or NFA significantly reduced ATP uptake. Buffers of DIDS and Evans blue-treated groups contained 1.36 Ϯ 0.08 M and 1.61 Ϯ 0.12 M ATP, respectively, at the 19-min time point after the first ATP addition, whereas the buffer of the non-treated group contained only 0.04 Ϯ 0.01 M ATP in the buffer (Fig. 1B). Consistent with the data from HEK cells, we also detected inhibition of ATP transport by DIDS, but not CBX, in lysosomes prepared from C2C12 cells (Fig. 4, E-G). Taken together, these FIGURE 1. ATP transport across lysosomal membrane is blocked by DIDS or Evans blue. A, ATP uptake in isolated lysosomes is blocked by DIDS or Evans blue but not by CBX or NFA. Lysosomes isolated from HEK293T cells were treated with CBX, NFA, DIDS, or Evans blue (20 M) for 30 min and then subjected to ATP uptake assays. Lysosomes were suspended in sodium-free homogenizing buffer (0.25 M sucrose, 1 mM EDTA, 10 mM HEPES (pH 7.0)) which was used under the "buffer alone" condition to detect intrinsic decay of ATP luminescence after ATP addition. At 11 and 21 min, ATP was added to 100 l of sample to a final concentration (Conc) of 2 M. Decay of ATP luminescence was detected over time as an indicator of active ATP uptake. At the end of the assay, 0.1% Triton X-100 was added to measure total ATP released from vesicles. B, ATP levels at 19 and 29 min. DIDS or Evans blue blocked ATP transport into lysosomes and resulted in higher extralysosome ATP levels relative to controls, whereas CBX or NFA treatment had no significant effect. Data were obtained from three independent experiments. NT, no treatment. **, p Ͻ 0.01. C, luminescence decay was observed in intact lysosomes but not in GPN (200 M)-treated lysosomes, which release lysosomal ATP, or in buffer alone.
data suggest the presence of a VNUT-like transporter in the lysosomes.
Loss of luminescence could be caused by ATP breakdown. To exclude this possibility, isolated lysosomes were treated with GPN (200 M), a substrate of the lysosomal exopeptidase cathepsin C. GPN induces lysosome osmodialysis (6) and promotes the release of lysosomal ATP. After the addition of 2 M ATP, we observed no significant luminescence decay in the GPN-treated group or the buffer control group. In contrast, non-treated lysosomes showed a dramatic loss of luminescence (Fig. 1C), suggesting that decreased luminescence was due to ATP uptake by lysosomes rather than ATP hydrolysis.
Inhibition of Lysosomal ATP Accumulation by DIDS and Evans Blue-If ATP accumulation is mediated by an ATP transporter, blocking that transporter should antagonize ATP deposition in lysosomes. To test this hypothesis, COS1 cells were treated with DIDS, Evans blue, CBX, or NFA, and ATP deposition was evaluated by the accumulation of quinacrine, a fluorescent dye used to locate intracellular ATP stores (6,7,12). As shown in Fig. 2A, quinacrine-labeled COS1 cells displayed a punctate granular pattern consistent with a vesicular origin, and puncta colocalized with LysoTracker and Lamp1staining (Fig. 5A), suggesting ATP accumulation in lysosomes. DIDS (20 M) or Evans blue (20 M) treatment significantly decreased quinacrine fluorescence (Fig. 2, A and B) by 84.71 Ϯ 3.01 and 84.45 Ϯ 7.25%, respectively. By contrast, treatment with CBX (20 M) or NFA (20 M) had no effect.
Lysosomal ATP levels were further measured by a bioluminescence assay using purified lysosomes from COS1 cells. High ATP concentrations in lysosome fractions, which were positive for the lysosome marker Lamp1, and in mitochondrial fractions, which were positive for the mitochondrial marker complex II, were detectable only after 5 min of treatment with 1% Triton X-100, indicating that ATP was contained within membrane-enclosed compartments (Fig. 2, C and D). Treatment of GPN (200 M for 30 min) selectively increased the ATP levels in the lysosomal fractions but not mitochondrial fractions, suggesting a high lysosomal ATP concentration. Lysosomal ATP content (defined by the concentration of ATP released following GPN treatment) decreased remarkably from 15.15 Ϯ 3.9 M to 1.64 Ϯ 0.89 M or 2.62 Ϯ 1.23 M following treatment with DIDS or Evans blue, respectively (Fig. 2C), whereas CBX or NFA treatment (Fig. 2D) had no effect on comparable samples. Taken together, these data suggest that a VNUT-like transporter may mediate lysosomal ATP accumulation. They also suggest that DIDS and Evans blue slowly penetrate the cell membrane, although the mechanism remains unclear.
SLC17A9 Localization in Lysosomes-SLC17A9 reportedly localizes in secretory vesicles/granules (3)(4)(5). Our pharmacology data obtained using isolated lysosomes suggest that lysosomal membranes contain substantial amounts of SLC17A9 protein. To confirm this observation, lysosomal fractions of COS1 cells or COS1 cells expressing mSLC17A9-GFP (mouse SLC17A9-GFP) were enriched using density gradient centrifu- gation, lysed, and then incubated with anti-SLC17A9 antibody. As shown in Fig. 3, A and B, both endogenous SLC17A9 and mSLC17A9-GFP were distributed to the Lamp1-positive fraction.
To further assess SLC17A9 subcellular localization, COS1 cells were transfected with Lamp1-GFP and then assessed with an SLC17A9 antibody. A large proportion of SLC17A9-positive puncta were also Lamp1-GFP-positive (Fig. 3C), suggesting that endogenous SLC17A9 is enriched in lysosomes. To evaluate the localization of exogenous SLC17A9, COS1 cells were transfected with mSLC17A9-GFP and labeled with Lyso-Tracker or cotransfected with mSLC17A9-mCherry and Lamp1-GFP. Confocal imaging indicated a large proportion of exogenous mSLC17A9-GFP fluorescence in LysoTracker-or Lamp1-GFP-positive organelles (Fig. 3, D and E). Overall, our data suggest that both endogenous and heterologously expressed SLC17A9 is enriched in lysosomes.
SLC17A9 Mediates Lysosome ATP Transport-To test whether SLC17A9 is required for lysosomal ATP transport, C2C12 cells (a mouse cell line) were transfected with mSLC17A9-shRNA and then assayed for ATP transport. SLC17A9 expression levels in either C2C12 whole cell lysate (Fig. 4A) or lysosome lysate (Fig. 4B) from C2C12 culture were suppressed dramatically by mSLC17A9-shRNA expression. SLC17A9 silencing significantly decreased ATP transport in lysosomes isolated from C2C12 cells (Fig. 4, C and D). Although the initial ATP concentrations in buffer were comparable (ϳ2 M), by 21 min, ATP concentrations decreased to 0.14 Ϯ 0.03 M or 1.49 Ϯ 0.23 M in cells expressing scramble control shRNA or mSLC17A9-shRNA, respectively (Fig. 4D). Conversely, mSLC17A9-GFP overexpression dramatically increased lysosomal ATP uptake (Fig. 4, E and F). ATP concentrations in the buffer decreased to 0.023 Ϯ 0.001 M and 0.010 Ϯ 0.001 M in WT cells and cells expressing mSLC17A9-GFP, respectively, by 19 min after the first ATP addition, although initial ATP levels were comparable (ϳ2.0 M) (Fig. 4, E and F). As in WT cells, DIDS treatment inhibited ATP transport in cells expressing mSLC17A9-GFP. However, in mSLC17A9-GFP-expressing cells, a higher concentration (20 M) of DIDS was required to suppress ATP transport to the same level as 5 M DIDS did in WT cells. In contrast, 20 M CBX (Fig. 4, G and H) did not inhibit ATP transport. Our data suggest that SLC17A9 is both sufficient and required for ATP transport across lysosomal membranes.
Although quinacrine is commonly used as an ATP indicator (6,7,12), it has also been suggested to function as a pH indicator (13). Identification of SLC17A9 as a lysosome ATP transporter allowed us to clarify this. As shown in Fig. 5C, SLC17A9-deficient lysosomes showed strong LysoTracker staining, suggestive of an acidic pH, despite decreased lysosomal ATP content (Fig. 5E) and a decrease in quinacrine staining (Fig. 5A). These data suggest that quinacrine is likely an indicator of lysosome ATP but may not be a reliable indicator of pH. This is also in agreement with data showing decreased quinacrine staining but strong LysoTracker staining in DIDS-treated cells (Fig. 2A).
Lysosomal ATP accumulation was further evaluated by measuring ATP content in lysosomes purified from C2C12 cells. As shown in Fig. 5, D and E, mSLC17A9-GFP overexpression did not significantly increase GPN-induced ATP release from purified lysosomes. This outcome might be due to contamination of lysosomes isolated from non-transfected cells. However, mSLC17A9-shRNA dramatically decreased lysosomal ATP by 80.6% (from 12.04 Ϯ 1.68 M to 2.38 Ϯ 0.49 M). In contrast, ATP content in mitochondria was unchanged. Taken together, our data suggest that suppression of SLC17A9 significantly decreases lysosomal ATP accumulation, whereas its up-regulation has an opposing effect.
SLC17A9-mediated Lysosome ATP Accumulation Is Blocked by DIDS and Evans Blue-To further assess the role for SLC17A9 in lysosomal ATP accumulation, COS1 cells expressing mSLC17A9-mCherry were treated with inhibitors and then stained with quinacrine. As shown in Fig. 6, A and B, although DIDS (5 M) or Evans blue (5 M) treatment significantly decreased quinacrine intensity in WT cells by 84.64 Ϯ 6.39% or 84.32 Ϯ 10.39%, respectively, those treatments had little effect on quinacrine intensity in mSLC17A9-mCherry-expressing cells. However, a higher concentration of DIDS (20 M) or Evans blue (20 M) was sufficient to eliminate quinacrine staining in mSLC17A9-mCherry-expressing cells (Fig. 6, A and B). In contrast, CBX treatment did not decrease quinacrine intensity in wild-type or mSLC17A9-mCherry-expressing cells. Accordingly, DIDS (5 M) or Evans blue (5 M) treatment significantly decreased ATP content (from 20.67 Ϯ 2.96 M to 1.60 Ϯ 0.70 M or 0.88 Ϯ 0.14 M, respectively) in lysosomes purified from WT cells. However, higher (20 M) concentrations of DIDS or Evans blue were required to significantly decrease ATP contents in lysosomes isolated from mSLC17A9-mCherry-expressing cells (Fig. 6C) (20 M) or Evans blue (20 M) was applied. These data suggest that SLC17A9 overexpression increases ATP transport efficiency, and, hence, higher levels of DIDS and Evans blue are required to inhibit ATP accumulation. These data are also in accordance with findings from the ATP uptake assay (Fig. 4, E and F). Overall, lysosomal SLC17A9 expression is sufficient to accumulate ATP, an activity blocked by DIDS or Evans blue.
SLC17A9 Deficiency Promotes Cell Death-C2C12 cells expressing mSLC17A9-shRNA appeared rounder, suggesting that SLC17A9 loss compromises cell survival. To assess this possibility, we monitored cell death in C2C12 cells using propidium iodide (PI) staining. In cells expressing control scramble shRNA, PI staining was observed in 4.11 Ϯ 0.99% of attached cells. In contrast, 38.44 Ϯ 5.63% of attached cells expressing mSLC17A9-shRNA were PI positive (Fig. 7, A and B). In agreement with this, DIDS (5 M or 20 M) treatment, but not CBX or NFA treatment, dramatically increased PI-positive cells in COS1 cells (Fig. 7, C and D). PI-positive attached cells were increased from 3.31 Ϯ 1.28% in control to 33.00 Ϯ 5.86% (5 M . ATP is transported into lysosomes through SLC17A9. A, detection of mSLC17A9 in WT or mSLC17A9 knockdown C2C12 cells. Whole cell lysates were used, and GAPDH levels served as loading controls. B, detection of endogenous mSLC17A9 in isolated lysosomes of WT and mSLC17A9 knockdown C2C12 cells showing that endogenous protein levels are robustly down-regulated in the latter. C, SLC17A9 knockdown impairs lysosomal ATP uptake. Lysosomes of WT and mSLC17A9 knockdown C2C12 cells were isolated and assayed for ATP uptake. Conc, concentration. A and D, ATP levels at 21 and 43 min. The mSLC17A9-shRNA group showed significantly higher levels of ATP at 21 and 43 min, suggestive of impaired ATP transport. **, p Ͻ 0.01; NS, no significance. E and F, mSLC17A9-GFP overexpression facilitated ATP transport across lysosome membranes. 19 min after ATP addition, lysosomes from cells overexpressing mSLC17A9-GFP showed significant lower ATP levels than the WT group. DIDS (5 M) significantly blocked ATP uptake in the WT group but not in the mSLC17A9-GFP-overexpressing group. DIDS (20 M) blocked ATP uptake in both WT and mSLC17A9-GFPoverexpressing groups. *, p Ͻ 0.05; **, p Ͻ 0.01. G and H, CBX (20 M) did not suppress ATP uptake mediated by mSLC17A9-GFP. The experiment was repeated three times in triplicate. *, p Ͻ 0.05. DIDS) and 76.00 Ϯ 4.58% (20 M DIDS) in DIDS-treated groups. Given that 5 M DIDS inhibited lysosomal ATP accumulation in WT cells but not mSLC17A9-GFP-expressing cells, overexpression may rescue the cell death phenotypes observed in cells treated with 5 M DIDS. As shown in Fig. 7, C and D, the number of PI-positive attached COS1 cells decreased to 8.67 Ϯ 1.76% when mSLC17A9-GFP, but not Lamp1-GFP (data not shown), was overexpressed. Such rescue was not observed in mSLC17A9-GFP-expressing cells treated with 20 M DIDS, a finding in agreement with the observation that 20 M DIDS inhibited lysosomal ATP accumulation in COS1 cells expressing mSLC17A9-GFP (Fig. 6). We also assessed cell death associated with SLC17A9 loss by assaying the levels of lactate dehydrogenase, a cytosolic enzyme that is released into cell culture medium when cells lose membrane integrity. Both DIDS (5 M or 20 M) treatment and SLC17A9 knockdown dramatically increased cell death on the basis of the lactate dehydrogenase assay. Specifically, DIDS treatment induced 55.58 Ϯ 3.07% (5 M) and 68.13 Ϯ 3.34% (20 M) cell death, whereas SLC17A9 silencing induced 39.95 Ϯ 8.00% cell death in C2C12 cells (Fig. 7E).
SLC17A9 Suppression Promotes Lysosomal Dysfunction-To confirm that SLC17A9 knockdown promotes lysosomal dysfunction potentially associated with cell death, we as-sessed lysosomal function using DQ-BSA, a dye that indicates active proteolysis in lysosomes. Strong DQ-BSA signals were seen in control C2C12 cells, as anticipated, but not in cells expressing mSLC17A9-shRNA (Fig. 8A), suggesting that mSLC17A9 loss compromises lysosomal function. In agreement, treatment of C2C12 cells with DIDS (at 5 and 20 M) but not CBX (20 M) reduced DQ-BSA staining (Fig.  8B). mSLC17A9-GFP overexpression rescued attenuated DQ-BSA staining induced by 5 M but not 20 M DIDS, suggesting that lysosome dysfunction induced by 5 M DIDS results from SLC17A9 loss.
Dysfunction of lysosomes promotes accumulation of proteins, lipids, and carbohydrates, which may be further converted to a non-degradable oxidized polymeric substance called lipofuscin, also referred to as "aging pigment." Autofluorescence (green) detectable within the excitation wavelength of 480 nm is an indicator of lipofuscin accumulation (14). Although we observed no significant autofluorescence in control cells, mSLC17A9 knockdown cells showed strong autofluorescence (Fig. 8C), which was also observed after cells were treated with DIDS but not CBX (Fig. 8D). Taken together, our data suggest that lysosomal SLC17A9 is critical for lysosomal function and that lysosomal SLC17A9 might underlie cell death.

DISCUSSION
Previous studies have indicated that ATP is enriched in lysosomes of astrocytes and microglia cells and is released through lysosomal exocytosis in response to various stimuli (6,7). We have also reported that lysosomes in COS1, HEK293, and C2C12 cells contain high (ϳmillimolar) ATP levels (12). However, the mechanisms underlying ATP accumulation in those organelles remained unclear. Emerging evidence suggests that VNUT (also called SLC17A9) could be a lysosomal ATP transporter (15), but controversy remains (16). In particular, evidence supporting a lysosomal localization of endogenous SLC17A9 or lysosomal ATP transport activity via SLC17A9 was lacking. Here, using quinacrine staining, a bioluminescent ATP assay, and ATP uptake assays, we provide strong evidence that SLC17A9 is responsible for lysosomal ATP accumulation.
Previously, vesicular neurotransmitter transporters (vesicular glutamate transporters and VNUT/SLC17A9) have been shown to take up glutamate or ATP into synaptic vesicles or secretory vesicles using membrane potential driven by V-ATPase (3,(17)(18)(19). Like synaptic vesicles and secretory ves-icles, V-ATPase in the lysosome membranes generates substantial voltage potential and an H ϩ gradient (20,21). Therefore, is likely that SLC17A9 transports ATP into lysosomes using the membrane potential established by V-ATPase, although confirmation of this awaits further investigation.
Although DIDS is often used as a SLC17A9 blocker (3,17), it also reportedly blocks other transporters (22). Here, however, we propose that defects in ATP uptake and accumulation induced by DIDS were primarily due to SLC17A9 inhibition. We conclude this for the following reasons. We showed that DIDS directly inhibited SLC17A9-mediated ATP uptake in isolated lysosomes, we observed a correlation between SLC17A9 levels and DIDS concentrations required to inhibit ATP transport as well as ATP accumulation (Figs. 4 and 6), and effects on DIDS treatment were rescued by overexpression of SLC17A9-GFP (Figs. 7C and 8B).
Quinacrine has been used as an indicator of lysosomal pH (13). However, quinacrine also binds ATP with a high affinity (23,24) and is commonly used to visualize ATP-containing vesicles, including lysosomes (6,7,25,26). We provide new evi-   dence that quinacrine intensity correlates with lysosomal ATP levels rather than lysosomal acidic pH. Several lines of evidence support this. Although quinacrine staining was altered by both suppression (Figs. 2A and 5A) or overexpression (Fig. 5A) of SLC17A9, lysosomal pH, as indicated by LysoTracker staining, remained unchanged in either case (Figs. 2A, 3D, and 5C). SLC17A9 knockdown eliminated quinacrine staining in lysosomes (Fig. 5). Quinacrine intensity was correlated with lysosomal ATP content, as measured using the ATP bioluminescent assay in all experiments. Therefore, our data suggest that quinacrine is an indicator for ATP content in acidic compartments.
SLC17A9 reportedly functions in insulin secretion through vesicular ATP release in ␤ cells (5), in exocytosis of ATP-enriched vesicles in biliary epithelial cells (25), and in exocytosis of ATP-enriched vesicles during T cell activation (4). However, the identities of the ATP-containing vesicles in these studies are unclear. We showed that SLC17A9 is highly enriched in lysosomes in several cell types, including C2C12, COS1, and HEK293T cells ( Fig. 1 and data not shown), although a portion of SLC17A9 is distributed in non-lysosomal vesicles. Therefore, in these cells, lysosomes may be the major secretory vesicles that release ATP. Previously, others have reported that ATP accumulated in lysosomes of astrocytes (6,26) and microglia (7) and lysosomal ATP release via lysosomal exocytosis is implicated in ischemic insult (6) and cell migration (7). However, the molecular mechanisms underlying ATP accumulation in these cells remain elusive. On the basis of the data presented here, SLC17A9 might be an important lysosomal ATP transporter in astrocytes and microglia cells as well and potentially function in microglia migration and ischemic damage.
The cell types used in this study exhibit little lysosomal exocytosis. This raises the question of why these cells need to store ATP in lysosomes. Lysosomes contain hydrolases that degrade organelles and macromolecules delivered to the lysosomal compartment by autophagy, endocytosis, or phagocytosis. Lysosomal dysfunction causes lysosomal storage and lipofuscin accumulation and, subsequently, leads to cell death (27). We showed that cells with compromised SLC17A9 lacked DQ-BSA staining and accumulated lipofuscin in lysosomes, suggesting defective lysosome function. Likewise, inhibition of SLC17A9mediated ATP transport into lysosomes was correlated with significant cell death, although the mechanism underlying that outcome remains unknown. ATP reportedly activates proteases such as cathepsin D and increase proteolysis within lysosomes (28), and cathepsin D deficiency induced lysosomal storage of lipofuscin (29). Our data suggest that SLC17A9 regulates lysosomal proteolysis by maintaining ATP at levels sufficient to activate proteases, such as cathepsin D, within the compartment. Therefore, compromised SLC17A9 function could attenuate cathepsin D activity and promote subsequent lipofuscin accumulation and cell death.
In summary, we present the first direct evidence that that SLC17A9 actively transports ATP across the lysosomal membrane in several cell types. Compromised SLC17A9 leads to reduced ATP accumulation in lysosomes and dysfunctional lysosomes and, subsequently, to cell death. Therefore, ATP functions not only as cellular energy currency but is also critical for lysosome function.