Mitochondria-lysosome contacts regulate mitochondrial Ca²⁺ dynamics via lysosomal TRPML1

Significance

Mitochondria and lysosomes are critical for cellular homeostasis and defects in both organelles are observed in several diseases. Recently, contact sites between mitochondria and lysosomes were identified and found to modulate mitochondrial dynamics. However, whether mitochondria–lysosome contacts have additional functions is unknown. Here, we identify a function of mitochondria–lysosome contacts in facilitating the direct transfer of calcium from lysosomes to mitochondria. Transfer of calcium at mitochondria–lysosome contacts is mediated by the lysosomal channel TRPML1 and is disrupted in mucolipidosis type IV, a lysosomal storage disorder caused by loss-of-function mutations in TRPML1. Calcium transfer from lysosomes to mitochondria at mitochondria–lysosome contacts thus presents an additional mechanism of intracellular calcium regulation that may further contribute to various disorders.

Abstract

Mitochondria and lysosomes are critical for cellular homeostasis, and dysfunction of both organelles has been implicated in numerous diseases. Recently, interorganelle contacts between mitochondria and lysosomes were identified and found to regulate mitochondrial dynamics. However, whether mitochondria–lysosome contacts serve additional functions by facilitating the direct transfer of metabolites or ions between the two organelles has not been elucidated. Here, using high spatial and temporal resolution live-cell microscopy, we identified a role for mitochondria–lysosome contacts in regulating mitochondrial calcium dynamics through the lysosomal calcium efflux channel, transient receptor potential mucolipin 1 (TRPML1). Lysosomal calcium release by TRPML1 promotes calcium transfer to mitochondria, which was mediated by tethering of mitochondria–lysosome contact sites. Moreover, mitochondrial calcium uptake at mitochondria–lysosome contact sites was modulated by the outer and inner mitochondrial membrane channels, voltage-dependent anion channel 1 and the mitochondrial calcium uniporter, respectively. Since loss of TRPML1 function results in the lysosomal storage disorder mucolipidosis type IV (MLIV), we examined MLIV patient fibroblasts and found both altered mitochondria–lysosome contact dynamics and defective contact-dependent mitochondrial calcium uptake. Thus, our work highlights mitochondria–lysosome contacts as key contributors to interorganelle calcium dynamics and their potential role in the pathophysiology of disorders characterized by dysfunctional mitochondria or lysosomes.

Interorganelle contact sites have become increasingly appreciated as essential regulators of cellular homeostasis. Contact sites, which form dynamically between two distinct organelles in close proximity, have been shown to have a variety of functions, including the ability to act as platforms for the direct transfer of ions, such as calcium (1–6). Recently, interorganelle contact sites between mitochondria and lysosomes were characterized, revealing a novel mechanism of cross-talk between the two organelles (7–18). Interestingly, both mitochondria and lysosomes are also important players in cellular homeostasis, including intracellular calcium dynamics (19–22), and a number of diseases presenting with mitochondrial and lysosomal dysfunction also exhibit dysregulation of cellular calcium (23–30). Although the calcium dynamics of mitochondria and lysosomes have previously been studied individually or in relation to other organelles (1–5, 31, 32), whether mitochondria and lysosomes can interact directly to modulate their calcium states has not been elucidated. Mitochondria–lysosome contacts may thus enable the direct transfer of calcium between lysosomes and mitochondria and function as an additional pathway in regulating intracellular calcium homeostasis.

Transient receptor potential mucolipin 1 (TRPML1) is a lysosomal/late-endosomal cation channel that mediates lysosomal calcium efflux (33–38) and function (39–44), and dysfunction in TRPML1 has been associated with several mitochondrial defects (45, 46). In addition, loss-of-function mutations in TRPML1 cause mucolipidosis type IV (MLIV), an autosomal recessive lysosomal storage disorder characterized by psychomotor retardation, retinal degeneration, and developmental delay (33, 47–49), and which has been associated with various lysosomal and mitochondrial aberrations (45, 46, 50–54). However, whether TRPML1-mediated lysosomal calcium release modulates mitochondrial calcium dynamics via mitochondria–lysosome contact sites, and the role of mitochondria–lysosome contact site dysfunction in the pathophysiology of lysosomal storage disorders such as MLIV, has not previously been studied.

Using live-cell high spatial and temporal resolution microscopy, we show that TRPML1 lysosomal calcium release mediates the direct transfer of calcium into mitochondria. Calcium transfer from lysosomes to mitochondria is modulated by mitochondria–lysosome contact site tethering and is modulated by the outer and inner mitochondrial membrane proteins, voltage-dependent anion channel 1 (VDAC1) and mitochondrial calcium uniporter (MCU), respectively. Importantly, MLIV patient fibroblasts with loss of TRPML1 function exhibit disrupted mitochondria–lysosome contact site dynamics and contact-dependent calcium transfer, suggesting a potential contribution of dysregulated mitochondria–lysosome contact site dynamics in lysosomal storage disorders. Our results thus elucidate an additional mechanism for regulating intracellular calcium dynamics via mitochondria–lysosome contact sites, which are further implicated in disease pathophysiology.

Results

TRPML1-Mediated Lysosomal Calcium Efflux Leads to Mitochondrial Calcium Influx.

To evaluate whether lysosomal TRPML1 calcium efflux modulated mitochondrial calcium (Fig. 1A), we used live-cell confocal microscopy at high spatial and temporal resolution to image mitochondrial calcium dynamics using the mitochondria-targeted genetically encoded calcium sensor Mito-R-GECO1 (55) (Fig. 1B). We first verified correct localization of Mito-R-GECO1, which was found to localize to the mitochondrial matrix as demonstrated by colocalization with the mitochondrial matrix-targeted BFP-mito (SI Appendix, Fig. S1A). Next, mitochondrial calcium responses were measured in wild-type HeLa cells upon activation of TRPML1 lysosomal calcium release with the TRPML1 agonist ML-SA1 (54). Following treatment with ML-SA1, total mitochondrial calcium was significantly increased (Fig. 1 C and D and Movie S1). Compared to control cells, cells treated with ML-SA1 showed a sustained elevation in mitochondrial calcium (Fig. 1D) and a significant increase in maximum mitochondrial calcium, mean mitochondrial calcium, and mitochondrial calcium at multiple time points (Fig. 1 E–G). These results were further validated by activation of TRPML1 with an additional small-molecule agonist, MK6-83 (56) (SI Appendix, Fig. S1 B–F), as well as with its physiological activator, PI(3,5)P₂ (57) (SI Appendix, Fig. S1 G–J), both of which resulted in a sustained increase in mitochondrial calcium. Thus, lysosomal TRPML1 calcium efflux robustly modulates mitochondrial calcium dynamics by increasing calcium influx into the mitochondrial matrix.

Fig. 1.

TRPML1-mediated lysosomal calcium efflux leads to mitochondrial calcium influx. (A) Model of activation of lysosomal calcium release by TRPML1 agonist, ML-SA1, resulting in mitochondrial calcium influx at mitochondria–lysosome contacts. (B) Experimental design for the assessment of mitochondrial calcium responses (ΔF/F) to TRPML1 activation in live cells. (C and D) Mitochondrial calcium response in live HeLa cells expressing mitochondrial-matrix targeted calcium sensor, Mito-R-GECO1, in response to TRPML1 activation with ML-SA1 (31.25 µM) (yellow arrow) or control treatment (white arrow) at t = 0 s with representative time-lapse confocal images (C, n = 23 cells for ML-SA1, n = 20 cells for control) and mitochondrial calcium traces (ΔF/F) (D, n = 23 cells for ML-SA1, n = 20 cells for control). (Scale bars, 10 µm; 1 µm in zoom images.) (E–G) Quantification of maximum mitochondrial calcium response (E), mean mitochondrial calcium response (F), and mitochondrial calcium response at 30, 60, 90, and 120 s (G) after TRPML1 activation with ML-SA1 (31.25 µM) or control treatment from confocal time-lapse images in C (n = 23 cells for ML-SA1, n = 20 cells for control). Data are means ± SEM (***P < 0.001, ****P < 0.0001, unpaired two-tailed t test).

TRPML1 Activation Preferentially Increases Mitochondrial Calcium at Mitochondria–Lysosome Contacts.

Because activation of lysosomal calcium release via TRPML1 led to increased mitochondrial calcium, we next evaluated whether this increase in mitochondrial calcium preferentially occurred at mitochondria–lysosome contact sites. We found that stable mitochondria–lysosome contacts dynamically formed in wild-type HeLa cells, defined as lysosomes remaining tethered to mitochondria for over 10 s (Fig. 2A), as recently described (8, 9). To assess whether TRPML1 mediated the direct transfer of calcium at mitochondria–lysosome contacts, we analyzed the calcium dynamics of mitochondria that were either in contact or not in contact with lysosomes upon TRPML1 activation (Fig. 2B, SI Appendix, Fig. S2A, and Movie S2). Mitochondria stably in contact with lysosomes (>10 s) had a significantly higher increase in calcium after TRPML1 activation, compared to mitochondria not in contact with lysosomes (Fig. 2 B and C). This preferential increase in mitochondrial calcium at mitochondria–lysosome contacts was observed in multiple cell types, including fibroblasts and HCT116 cells, which similarly showed a mitochondria–lysosome contact-dependent increase in mitochondrial calcium upon activation of TRPML1 lysosomal calcium release (SI Appendix, Fig. S2 B–D and Movies S3 and S4).

Fig. 2.

TRPML1 activation preferentially increases mitochondrial calcium at mitochondria–lysosome contacts. (A) Representative time-lapse confocal images showing stable mitochondria–lysosome contact site tethering (white arrows) in live HeLa cells expressing lysosomal marker, lamp1-mGFP, and mitochondrial-matrix targeted calcium sensor, Mito-R-GECO1. (Scale bar, 1 µm.) (B) Representative time-lapse confocal images of increase in mitochondrial calcium at t = 30 s following TRPML1 activation with ML-SA1 (31.25 µM) in mitochondria in contact with lysosomes (Lower, yellow arrow) versus those not in contact with lysosomes (Upper, white arrow) in live HeLa cells expressing mitochondrial matrix calcium sensor, Mito-R-GECO1, and lysosomal marker, lamp1-mGFP (n = 100 events from 20 cells for each condition). (Scale bars, 1 µm.) (C) Quantification of mitochondrial calcium responses from mitochondria in contact and not in contact with lysosomes following TRPML1 activation with ML-SA1 (31.25 µM) from confocal time-lapse images (from B) (n = 100 events from 20 cells for each condition). (D and E) Mitochondrial calcium response (ΔF/F) in live wild-type or TBC1D15 knockout (KO) HCT116 cells expressing mitochondrial-matrix targeted calcium sensor, Mito-R-GECO1, following TRPML1 activation with ML-SA1 (31.25 µM) at t = 0 s with representative time-lapse confocal images (D, n = 18 cells for each condition) and mitochondrial calcium traces (ΔF/F) (E, n = 18 cells for each condition). (Scale bars, 5 µm.) (F–H) Quantification of maximum mitochondrial calcium response (F), mean mitochondrial calcium response (G), and mitochondrial calcium response at 30 and 60 s (H) after treatment with ML-SA1 (31.25 µM) in live wild-type or TBC1D15 KO HCT116 cells from confocal time-lapse images (from D) (n = 18 cells for each condition). Data are means ± SEM (*P < 0.05, **P < 0.01, ****P < 0.0001, unpaired two-tailed t test).

We then investigated whether directly modulating mitochondria–lysosome contact sites could increase the transfer of lysosomal calcium into mitochondria. We previously showed that mitochondria–lysosome contact site untethering is regulated by the activity of the mitochondrial-localized Rab7 GTPase-activating protein (TBC1D15) driving Rab7 GTP hydrolysis on lysosomes/late endosomes, and consequently that TBC1D15 knockout significantly prolonged mitochondria–lysosome contact site tethering (8). To further investigate whether mitochondrial calcium dynamics could be regulated at mitochondria–lysosome contacts, we compared HCT116 wild-type with TALEN-generated HCT116 TBC1D15 knockout cells (58), as TBC1D15 knockout cells have significantly increased mitochondria–lysosome contact tethering duration (8). Upon TRPML1 activation with ML-SA1, TBC1D15 knockout cells showed a significantly greater increase in total mitochondrial calcium compared to wild-type cells (Fig. 2 D and E), as well as significantly increased maximum mitochondrial calcium, mean mitochondrial calcium, and mitochondrial calcium at multiple time points (Fig. 2 F–H). Thus, directly modulating mitochondria–lysosome contact site tethering is sufficient to increase lysosomal calcium transfer into mitochondria.

We further confirmed that mitochondrial uptake of lysosomal calcium was not dependent on the endoplasmic reticulum (ER) (31, 32), as blocking ER calcium release using an inositol 1,4,5-triphosphate receptor (IP3R) antagonist (Xestospongin-C, pretreatment for 20 min) (SI Appendix, Fig. S3 A–D) or ryanodine receptor antagonist (DHBP, pretreatment for 10 min) (SI Appendix, Fig. S3 E–G) did not prevent an increase in mitochondrial calcium (Mito-R-GECO1) upon TRPML1 activation. Similarly, neither blocking store-operated calcium entry with a stromal interaction molecule 1 (STIM1) inhibitor (SKF-96365, pretreatment for 10 min) (SI Appendix, Fig. S3 H–J) nor depleting ER calcium using a SERCA (sarco/endoplasmic reticulum Ca²⁺-ATPase) pump inhibitor (thapsigargin, pretreatment for 10 min) (SI Appendix, Fig. S3 K–M) altered mitochondrial calcium increase (Mito-R-GECO1) upon TRPML1 activation. TRPML1-mediated mitochondrial calcium influx was also unaltered upon chelation of cytosolic calcium (BAPTA-AM, pretreatment for 20 min), further suggesting that calcium transfer primarily occurs at mitochondria–lysosome contacts (SI Appendix, Fig. S4). Altogether, these findings indicate that TRPML1-mediated calcium influx into mitochondria occurs preferentially at mitochondria–lysosome contacts and, furthermore, can be directly regulated by modulating mitochondria–lysosome contact tethering machinery.

Lysosomal TRPML1 Specifically Modulates Mitochondrial Calcium and Mitochondria–Lysosome Contact Dynamics.

To further demonstrate that lysosomal TRPML1 activity modulates mitochondrial calcium, we expressed either wild-type TRPML1 (TRPML1 WT-Halo) or the dominant-negative, nonconducting TRPML1 pore mutant (TRPML1 D471K-Halo) in HeLa cells and examined mitochondrial calcium dynamics (Mito-R-GECO1) upon ML-SA1 treatment. We first confirmed that both wild-type TRPML1 and dominant-negative TRPML1 were localized to the lysosomal/late-endosomal compartment as evidenced by colocalization with the lysosomal membrane marker, BFP-lysosomes (SI Appendix, Fig. S5 A and B), and that both wild-type and dominant-negative TRPML1 were expressed at similar levels (SI Appendix, Fig. S5C). Upon ML-SA1 treatment, cells expressing dominant-negative TRPML1 showed a significant reduction in mitochondrial calcium influx compared to wild-type TRPML1-expressing cells (Fig. 3 A and B). In addition, expression of the dominant-negative TRPML1 mutant significantly reduced the maximum mitochondrial calcium, mean mitochondrial calcium, and mitochondrial calcium at multiple time points (Fig. 3 C–E). These results thus suggest that TRPML1 activity is important for modulating mitochondrial calcium dynamics.

Fig. 3.

Lysosomal TRPML1 specifically modulates mitochondrial calcium and mitochondria–lysosome contact dynamics. (A and B) Mitochondrial calcium response in live HeLa cells expressing mitochondrial-matrix targeted calcium sensor, Mito-R-GECO1, and either wild-type or dominant-negative (D471K) TRPML1 mutant, in response to TRPML1 activation with ML-SA1 (31.25 µM) at t = 0 s with representative time-lapse confocal images (A, n = 20 cells for each condition) and mitochondrial calcium traces (ΔF/F) (B, n = 20 cells for each condition). (Scale bars, 10 µm.) (C–E) Quantification of maximum mitochondrial calcium response (C), mean mitochondrial calcium response (D), and mitochondrial calcium response at 30, 60, 90, and 120 s (E) after TRPML1 activation with ML-SA1 (31.25 µM) in live HeLa cells expressing TRPML1 wild-type or TRPML1 D471K mutant from confocal time-lapse images (from A) (n = 20 cells for each condition). (F) Quantification of mitochondrial calcium responses of mitochondria in contact and not in contact with lysosomes following TRPML1 activation with ML-SA1 (31.25 µM) in live HeLa cells expressing mitochondrial matrix calcium sensor, Mito-R-GECO1, lysosomal marker, lamp1-mGFP, and either wild-type (TRPML1 WT-pHcRed) or dominant-negative (TRPML1 D471-472K-pHcRed) TRPML1 mutant (n = 100 events from 20 cells for each condition). (G–I) Quantification of percentage of lysosomes contacting mitochondria (for >10 s; G, n = 10 cells for each condition) and minimum duration of mitochondria–lysosome contacts with corresponding histogram (H and I, n = 70 events from 10 cells for each condition) in live HeLa cells expressing mitochondrial outer membrane marker, Tom20-mEmerald, and lysosomal marker, BFP-lysosomes along with either wild-type (TRPML1 WT-pHcRed) or dominant-negative (TRPML1 D471-472K-pHcRed) TRPML1 mutant. Data are means ± SEM (*P < 0.05, **P < 0.01, ****P < 0.0001, ns, not significant, unpaired two-tailed t test).

In order to investigate the role of TRPML1 at contact sites, we analyzed whether TRPML1 specifically modulated mitochondrial calcium at mitochondria–lysosome contact sites. HeLa cells expressing wild-type TRPML1 displayed a significantly higher increase in mitochondrial calcium after TRPML1 activation for mitochondria that were in contact with lysosomes, compared to mitochondria not in contact with lysosomes (Fig. 3F). In contrast, this difference in contact-dependent calcium transfer was entirely abolished in cells expressing the dominant-negative TRPML1 mutant (Fig. 3F). To probe downstream effects of calcium transfer at mitochondria–lysosome contacts, we investigated whether TRPML1-mediated mitochondrial calcium influx promoted mitochondrial permeability transition pore (mPTP) opening, which can be assessed by the quenching of mitochondrial-localized calcein in the presence of CoCl₂ (SI Appendix, Fig. S6A). In contrast to ionomycin treatment, which induced mPTP opening and rapid quenching of calcein, neither the TRPML1 agonist ML-SA1 nor vehicle control reduced mitochondrial calcein fluorescence (SI Appendix, Fig. S6 B and C), suggesting that TRPML1-mediated mitochondrial calcium influx does not induce sustained mPTP opening. We also investigated whether TRPML1-mediated mitochondrial calcium influx activated apoptotic pathways via endogenous cytochrome C staining (SI Appendix, Fig. S6D). Relative to staurosporine, which induced release of cytochrome C from the mitochondria into the cytosol, we did not observe significant changes in mitochondrial distribution of cytochrome C in cells treated with TRPML1 agonist ML-SA1 or vehicle control at multiple time points (SI Appendix, Fig. S6 E–G).

In addition to modulating contact-dependent calcium transfer, we found that TRPML1 modulated mitochondria–lysosome contact site dynamics. In cells expressing the dominant-negative TRPML1 mutant, there was both a higher percentage of lysosomes in stable contact with mitochondria (Fig. 3G) and a significant increase in minimum duration of mitochondria–lysosome contacts (Fig. 3 H and I). We further verified whether TRPML1 localized preferentially to mitochondria–lysosome contacts in cells expressing HA-tagged TRPML1 and an HA-mCherry nanobody. TRPML1-HA puncta marked a subset of contact sites (SI Appendix, Fig. S7A) and TRPML1-HA localization at the contact sites was significantly greater than expected by random chance (SI Appendix, Fig. S7B). These findings highlight a role for TRPML1 in regulating mitochondria–lysosome contact site dynamics and contact-dependent calcium transfer into mitochondria.

VDAC1 and MCU Mediate Mitochondrial Uptake of Lysosomal Calcium at Mitochondria–Lysosome Contact Sites.

Having shown that TRPML1-mediated lysosomal calcium release led to increased mitochondrial calcium at mitochondria–lysosome contacts, we next sought to identify the mitochondrial components promoting uptake of lysosomal calcium. VDACs on the outer mitochondrial membrane have been implicated in mitochondrial calcium uptake (59–64) and specifically, VDAC1 was previously identified as a potential interactor of lysosomal TRPML1 (65). To first confirm the interaction of TRPML1 with VDAC1, we conducted coimmunoprecipitation experiments in cells expressing TRPML1-GFP and found that TRPML1 interacted with endogenous VDAC1 (SI Appendix, Fig. S7C), but not endogenous VDAC2 or VDAC3 (SI Appendix, Fig. S7C). We next investigated whether VDAC1 is important for the mitochondrial uptake of lysosomal calcium by assessing mitochondrial calcium dynamics upon TRPML1 activation in cells expressing either wild-type human VDAC1 or a VDAC1 mutant (E73Q) with a single amino acid substitution in a putative calcium-binding site (66, 67). In response to TRPML1 activation with ML-SA1, cells expressing the VDAC1 mutant showed significantly lower increase in mitochondrial calcium (Mito-R-GECO1) (Fig. 4A and SI Appendix, Fig. S7D), as well as significantly decreased maximum mitochondrial calcium, mean mitochondrial calcium, and mitochondrial calcium at multiple time points compared to wild-type cells (Fig. 4 B–D). After verifying that wild-type and mutant VDAC1 were expressed at similar levels (SI Appendix, Fig. S7F), we next assessed whether VDAC1 regulated calcium transfer preferentially at mitochondria–lysosome contact sites. Mitochondria in contact with lysosomes had significantly elevated calcium influx after TRPML1 activation, compared to mitochondria not in contact with lysosomes in cells expressing wild-type VDAC1. In contrast, there was no difference in mitochondrial calcium response between mitochondria in and not in contact with lysosomes in VDAC1 mutant-expressing cells (Fig. 4E). These findings thus suggest that VDAC1 on the outer mitochondrial membrane serves as a mediator of mitochondrial uptake of lysosomal calcium at mitochondria–lysosome contacts.

Fig. 4.

VDAC1 and the MCU modulate mitochondrial uptake of lysosomal calcium at mitochondria–lysosome contact sites. (A) Mitochondrial calcium response (ΔF/F) in live HeLa cells expressing mitochondrial-matrix targeted calcium sensor, Mito-R-GECO1, and either wild-type or mutant (E73Q) VDAC1 in response to TRPML1 activation with ML-SA1 (31.25 µM) at t = 0 s (n = 22 cells for each condition). (B–D) Quantification of maximum mitochondrial calcium response (B), mean mitochondrial calcium response (C), and mitochondrial calcium response at 30, 60, 90, and 120 s (D) after TRPML1 activation with ML-SA1 (31.25 µM) in live HeLa cells expressing VDAC1 WT or VDAC1 E73Q mutant from confocal time-lapse images (from A) (n = 22 cells for each condition). (E) Quantification of mitochondrial calcium responses of mitochondria in contact and not in contact with lysosomes following TRPML1 activation with ML-SA1 (31.25 µM) in live HeLa cells expressing mitochondrial matrix calcium sensor, Mito-R-GECO1, lysosomal marker, lamp1-mGFP, and either wild-type or mutant (E73Q) VDAC1 (n = 100 events from 20 cells for each condition). (F) Mitochondrial calcium response (ΔF/F) in live HeLa cells expressing mitochondrial-matrix targeted calcium sensor, Mito-R-GECO1, and either wild-type or mutant (E264A) MCU in response to TRPML1 activation with ML-SA1 (31.25 µM) at t = 0 s (n = 20 cells for each condition). (G–I) Quantification of maximum mitochondrial calcium response (G), mean mitochondrial calcium response (H), and mitochondrial calcium response at 30, 60, 90, and 120 s (I) after TRPML1 activation with ML-SA1 (31.25 µM) in live HeLa cells expressing MCU WT or MCU E264A mutant from confocal time-lapse images (from F) (n = 20 cells for each condition). (J) Quantification of mitochondrial calcium responses of mitochondria in contact and not in contact with lysosomes following TRPML1 activation with ML-SA1 (31.25 µM) in live HeLa cells expressing mitochondrial matrix calcium sensor, Mito-R-GECO1, lysosomal marker, lamp1-mGFP, and either wild-type or mutant (E264A) MCU (n = 100 events from 20 cells for each condition). Data are means ± SEM (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, ns, not significant, unpaired two-tailed t test).

The MCU is the major transporter of calcium across the inner mitochondrial membrane into the mitochondrial matrix (68, 69). To evaluate the role of the MCU in uptake of lysosomal calcium across the inner mitochondrial membrane, we expressed either wild-type MCU or MCU mutant (E264A), which disrupts calcium uptake (69, 70), and first confirmed that their expression levels were similar (SI Appendix, Fig. S7G). We then examined mitochondrial calcium dynamics (Mito-R-GECO1) in these cells in response to TRPML1 activation. Compared to cells expressing wild-type MCU, cells expressing mutant MCU had reduced total mitochondrial calcium increase upon TPRML1 activation (Fig. 4F and SI Appendix, Fig. S7E), as well as significantly lower maximum mitochondrial calcium, mean mitochondrial calcium, and mitochondrial calcium at multiple time points after ML-SA1 treatment (Fig. 4 G–I). Importantly, the MCU was also important for mitochondria–lysosome contact-dependent calcium transfer. Wild-type MCU-expressing cells showed significant differences in calcium influx for mitochondria in contact with lysosomes compared to mitochondria not in contact, while expression of the MCU mutant (E264A) completely abolished this difference (Fig. 4J). These results indicate that MCU on the inner mitochondrial membrane modulates mitochondrial calcium dynamics at mitochondria–lysosome contact sites.

Loss of TRPML1 Function in MLIV Patient Fibroblasts Disrupts Mitochondria–Lysosome Contact and Calcium Dynamics.

Loss-of-function mutations in TRPML1 cause the autosomal recessive lysosomal storage disorder, MLIV (47, 48), which has been associated with both lysosomal (50–54) and mitochondrial aberrations (45, 46). Given that we found TRPML1 to be important for the regulation of mitochondrial calcium dynamics via direct transfer of calcium at mitochondria–lysosome contact sites, we evaluated whether MLIV patient fibroblasts had defective mitochondrial calcium dynamics due to loss of TRPML1 function. We treated fibroblasts from MLIV patients and age-matched healthy controls with ML-SA1 to activate TRPML1 and examined mitochondrial calcium dynamics (Mito-R-GECO1). While control fibroblasts showed a significant increase in total mitochondrial calcium upon TRPML1 activation, this was reduced in multiple MLIV patient fibroblast lines (Fig. 5A). Consistent with these findings, maximum mitochondrial calcium, mean mitochondrial calcium, and mitochondrial calcium at 30 s was also significantly decreased in MLIV patient lines (Fig. 5 B–D). Moreover, MLIV patient fibroblasts showed defects in contact-dependent calcium transfer. In control fibroblasts, mitochondria in contact with lysosomes showed a significantly higher increase in calcium influx following TRPML1 activation compared to those not in contact with lysosomes. In contrast, there was no difference in calcium influx between mitochondria in and not in contact with lysosomes in MLIV fibroblasts (Fig. 5E).

Fig. 5.

Loss of TRPML1 function in MLIV patient fibroblasts disrupts mitochondria–lysosome contact and calcium dynamics. (A) Mitochondrial calcium response (ΔF/F) in fibroblasts from MLIV patients (MLIV #1, MLIV #2, MLIV #3) or age-matched healthy controls (Con #1, Con #2, Con #3) expressing mitochondrial-matrix–targeted calcium sensor, Mito-R-GECO1 in response to TRPML1 activation with ML-SA1 (31.25 µM) at t = 0 s (n = 20 cells for each condition). (B–D) Quantification of maximum mitochondrial calcium response (B), mean mitochondrial calcium response (C), and mitochondrial calcium response at 30 s (D) after TRPML1 activation with ML-SA1 (31.25 µM) from fibroblasts from MLIV patients and controls (from A) (n = 20 cells for each condition). (E) Quantification of mitochondrial calcium responses of mitochondria in contact and not in contact with lysosomes following TRPML1 activation with ML-SA1 (31.25 µM) in fibroblasts from MLIV patients and controls expressing mitochondrial matrix calcium sensor, Mito-R-GECO1, and lysosomal marker, lamp1-mGFP (n = 100 events from 20 cells for each condition). (F and G) Quantification of percentage of lysosomes contacting mitochondria (for >10 s; F, n = 10 cells for each condition) and duration of mitochondria–lysosome contacts (G, n = 70 events from 10 cells for each condition) in fibroblasts from MLIV patients and controls expressing mitochondrial outer membrane marker, Tom20-mApple, and lysosomal marker, lamp1-mGFP. (H) Model of the regulation of calcium dynamics at mitochondria–lysosome contacts showing calcium transfer from lysosomes to mitochondria via TRPML1 (lysosome), VDAC1 (outer mitochondrial membrane), and MCU (inner mitochondrial membrane) leading to increased mitochondrial calcium in healthy cells (Left). In the lysosomal storage disorder MLIV, loss-of-function TRPML1 mutations lead to dysregulation of mitochondrial calcium dynamics (Right). Data are means ± SEM [*P < 0.05, ***P < 0.001, ****P < 0.0001, ns, not significant, one-way ANOVA with Tukey’s post hoc test (B, C, D, F, and G), unpaired two-tailed t test (E)].

In addition to changes in contact-dependent mitochondrial calcium responses, MLIV fibroblasts also displayed abnormal mitochondria–lysosome contact dynamics. Interestingly, MLIV fibroblasts had a significantly increased percentage of lysosomes in stable contact (>10 s) with mitochondria compared to control fibroblasts (Fig. 5F) and the duration of mitochondria–lysosome contact tethering was also significantly prolonged (Fig. 5G). Together, these data suggest that loss of TRPML1 function in MLIV may contribute to disease pathogenesis by dysregulating mitochondrial calcium dynamics at contact sites and additionally disrupting mitochondria–lysosome contact tethering dynamics. We thus propose a model in which TRPML1-mediated lysosomal calcium efflux results in mitochondrial calcium influx preferentially at mitochondria–lysosome contacts through the mitochondrial channels VDAC1 and the MCU, and that calcium transfer at mitochondria–lysosome contact sites is consequently disrupted in the lysosomal storage disorder MLIV due to loss-of-function TRPML1 mutations (Fig. 5H).

Discussion

We identified a role of mitochondria–lysosome contact sites in modulating intracellular calcium dynamics, whereby TRPML1-mediated lysosomal calcium efflux leads to mitochondrial calcium influx preferentially at mitochondria–lysosome contact sites. TRPML1-mediated increase in mitochondrial calcium is further modulated by VDAC1 and the MCU on the outer and inner mitochondrial membranes, respectively. Importantly, we show that mitochondrial calcium dynamics are disrupted in the lysosomal storage disorder MLIV, which results from loss of TRPML1 function. We additionally find that altered mitochondrial calcium dynamics in MLIV are dependent on mitochondria–lysosome contacts, providing further evidence for the convergence of lysosomal and mitochondrial dysfunction in disease.

Our work further establishes the growing importance of interorganelle contact sites in the regulation of cellular homeostasis. Defects in interorganelle contact sites have been implicated in multiple human diseases, including lysosomal storage disorders (6), peroxisomal diseases (71), and neurodegenerative disorders (27, 72–75). Recently, mitochondria–lysosome contact sites have been shown to be an important regulator of mitochondrial and lysosomal cross-talk independent of lysosomal degradation of mitochondria (8, 9, 11–18), and to be involved in regulating mitochondrial fission and intermitochondrial contact untethering (8, 16, 27), as well as the transfer of metabolites, such as cholesterol (10). Here, we find that mitochondria–lysosome contact sites further play a key role in regulating calcium transfer between these two organelles, which is disrupted in a lysosomal storage disorder. Uncovering the diverse functions of mitochondria–lysosome contact sites shall inform our understanding of how contact sites contribute to both physiological and pathophysiological states.

Our findings also expand upon the previously described physiological roles of TRPML1. TRPML1 is a nonselective cation channel that mediates lysosomal calcium efflux (33–38) and regulates various lysosomal functions, including lysosomal exocytosis, membrane trafficking, and lysosomal biogenesis (39, 41, 43, 56, 76–78). Our data suggest that in addition to regulating lysosomal dynamics and function, TRPML1 directly impacts mitochondrial homeostasis by modulating mitochondrial calcium dynamics via mitochondria–lysosome contact sites. Mitochondria–lysosome contacts may act as platforms to provide localized pockets of high calcium concentration required for influx into mitochondria and, moreover, contact-dependent transfer of calcium from lysosomes to mitochondria may serve as a mechanism to spatially regulate calcium transfer to a subset of mitochondria to facilitate downstream, calcium-dependent mitochondrial functions, including oxidative phosphorylation, motility, and reactive oxygen species (ROS) signaling (79–83). Indeed, mitochondrial function has also been shown to reciprocally regulate TRPML1 as increased mitochondrial ROS potentiates TRPML1 activity (77). Further studies investigating how TRPML1-mediated mitochondrial calcium influx modulates mitochondrial structure, dynamics, and function will provide additional insights into the direct communication between lysosomes and mitochondria.

In our study, we also identified VDAC1 on the outer mitochondrial membrane and the MCU on the inner mitochondrial membrane as mediators of mitochondrial calcium influx at mitochondria–lysosome contact sites. While our finding that TRPML1 preferentially interacts with VDAC1 and not other VDAC isoforms is consistent with previous studies (65), it is possible that the other VDAC isoforms, which have also been implicated in mitochondrial calcium uptake (62–64), may play a role in contact-dependent calcium transfer in other cell types. In addition, while we observed that mutant MCU impaired mitochondrial calcium uptake following TRPML1-mediated lysosomal calcium release, it did not completely abolish mitochondrial uptake at contact sites. Thus, our results suggest that there may be alternative, MCU-independent mechanisms of calcium influx (83) into the mitochondrial matrix at mitochondria–lysosome contacts. Indeed, prior studies have proposed additional transporters mediating mitochondrial matrix calcium influx (84–87), which may also play a role in the uptake of lysosomal calcium.

Of clinical relevance, loss-of-function mutations in TRPML1 cause the autosomal-recessive lysosomal storage disorder MLIV, which is characterized by psychomotor retardation, retinal degeneration, and neurodevelopmental delay (33, 47–49). Although the pathophysiology of MLIV remains unclear, various cellular phenotypes, including defective lysosomal biogenesis, altered lysosomal pH, impaired autophagy, and mitochondrial fragmentation have been described (43, 45, 46, 50, 76, 77, 88, 89). Notably, our results demonstrate that MLIV is also associated with defective mitochondria–lysosome contact dynamics and contact-dependent calcium transfer. As increasing mitochondria-lysosome contact duration in healthy cells increases contact-dependent calcium transfer, it is possible that prolonged mitochondria-lysosome contacts in MLIV compensate for reduced calcium transfer from lysosomes to mitochondria due to loss of TRPML1 function. Indeed, prolonged mitochondria-lysosome contacts have also been observed in other disorders affecting lysosomal/late endosomal genes including Charcot-Marie Tooth type 2B (27). It would be important in future studies to elucidate whether altered organelle contacts in MLIV contribute to previously observed mitochondrial phenotypes and whether MLIV mutations in TRPML1 alter additional mitochondrial functions, such as ATP production and mitochondrial fission/fusion dynamics in a contact-dependent manner. Given that many critical mitochondrial functions are regulated by calcium (83), it is possible that dysregulation of mitochondrial calcium, in conjunction with decreased lysosomal function, potentiates defects in mitochondrial metabolism and dynamics, which may consequently contribute to downstream phenotypes, such as mitochondrial fragmentation in MLIV disease pathogenesis.

In addition to its role in MLIV, TRPML1 has also been implicated in various neurological and lysosomal storage diseases (19, 90–93). Several disease models have shown down-regulation of TRPML1, which impairs lysosomal function and promotes accumulation of toxic proteins (94, 95). Other studies have reported misregulation of TRPML1 activity due to alterations in lipids or lysosomal pH (83, 89). TRPML1 activity is highly regulated by specific lipids, including its endogenous activator PI(3,5)P₂ (57), which has been suggested to be misregulated in Charcot-Marie Tooth disease and amyotrophic lateral sclerosis (96–98), and sphingomyelins and cholesterol, which when accumulated in Niemann-Pick type C, impair TRPML1-mediated lysosomal calcium release (54). Moreover, TRPML1 activity is likely regulated upstream by lysosomal pH as it has been shown that TRPML1-mediated calcium dyshomeostasis and autophagic defects are rescued by restoration of lysosomal pH, but not calcium, in an Alzheimer’s model (19). While these previous studies predominantly describe the role of TRPML1 in regulating lysosomal function in disease (43, 54, 76, 92), our findings suggest that TRPML1 may also contribute to disease pathogenesis by modulating mitochondrial calcium dynamics. Altered TRPML1-mediated calcium transfer at mitochondria–lysosome contact sites and subsequent dysregulation of mitochondrial calcium dynamics may be an additional contributory mechanism to pathophysiology. Indeed, many of these diseases share cellular hallmarks including mitochondrial and lysosomal dysfunction and calcium dyshomeostasis (23–26). Importantly, TRPML1 has recently emerged as a potential therapeutic target for the treatment of neurodegenerative and lysosomal storage disorders. Studies suggest that TRPML1 agonists may act multimodally by activating various lysosomal pathways, including autophagy and lysosomal exocytosis (43, 54, 92). Thus, our results further highlight a potential role for therapeutically targeting TRPML1 in modulating mitochondrial calcium dynamics in disease.

In summary, our work shows that lysosomes can directly transfer calcium to mitochondria at mitochondria–lysosome contacts, thereby further supporting the emerging role for both organelles as critical players in modulating calcium dynamics (20–22) and elucidating an additional pathway by which intracellular calcium can be regulated. A broader understanding of the mechanisms underlying intracellular calcium regulation will ultimately inform our understanding of the role of mitochondrial and lysosomal cross-talk in both cellular homeostasis and disease.

Materials and Methods

Detailed materials and methods are included in SI Appendix, Materials and Methods.

Cell Culture and Transfection.

HeLa, HEK293, and HCT116 cells and fibroblasts were cultured according to standard procedures at 37 °C in a 5% CO₂ incubator. Cells were transfected with either X-tremeGENE HP DNA transfection reagent (Roche; XTGHP-RO) or Lipofectamine LTX with PLUS reagent (Invitrogen; 15338100).

Live-Cell Confocal Microscopy.

All confocal images were acquired on a Nikon A1R laser scanning confocal microscopy with GaAsp detectors using a Plan Apo λ 100× 1.45 NA oil-immersion objective (Nikon) using NIS-Elements (Nikon). Live cells were imaged in a temperature-controlled chamber (37 °C) at 5% CO₂. Dual color videos were acquired as consecutive green–red or blue–green images. For calcium imaging, growth media was replaced with Krebs-Ringer solution (without calcium) (Boston BioProducts; BSS-255) immediately prior to imaging and live cells were imaged at one frame every 1 to 5 s.

Image Analysis.

Calcium responses for total mitochondria and contact-dependent mitochondrial calcium responses were assessed using ΔF/F₀ analysis. Analysis of mitochondria–lysosome contact dynamics were analyzed as previously described (8). All image analysis was conducted using ImageJ and NIS-Elements (Nikon).

Data Availability.

All data that support the findings of this study are included in the article or are available at the Open Science Framework, https://osf.io/wf83s/.