Accumulating evidence implicates acidic organelles of the endolysosomal system as mobilisable stores of Ca2+ but their relationship to the better-characterised endoplasmic reticulum (ER) Ca2+ store remains unclear. Here we show that rapid osmotic permeabilisation of lysosomes evokes prolonged, spatiotemporally complex Ca2+ signals in primary cultured human fibroblasts. These Ca2+ signals comprised an initial response that correlated with lysosomal disruption and secondary long-lasting spatially heterogeneous Ca2+ oscillations that required ER-localised inositol trisphosphate receptors. Electron microscopy identified extensive membrane contact sites between lysosomes and the ER. Mobilisation of lysosomal Ca2+ stores is thus sufficient to evoke ER-dependent Ca2+ release probably through lysosome–ER membrane contact sites, and akin to the proposed mechanism of action of the Ca2+ mobilising messenger nicotinic acid adenine dinucleotide phosphate (NAADP). Our data identify functional and physical association of discrete Ca2+ stores important for the genesis of Ca2+ signal complexity.
Changes in cytosolic Ca2+ form the basis of a ubiquitous and highly versatile signalling pathway indispensable for cell function (Berridge et al., 2000). Cytosolic Ca2+ signals display marked temporal and spatial heterogeneity in the form of oscillations and waves. They can be confined to a particular subcellular locale or spread through the cell and into neighbouring ones in order to regulate Ca2+-dependent targets (Berridge et al., 2000). Many of these signals rely upon the endoplasmic reticulum (ER) which fills with and releases Ca2+ through well-defined pumps and channels (Berridge, 2002). Ca2+ uptake by mitochondria serves to both shape ER-derived Ca2+ signals and to regulate mitochondrial function (Rizzuto et al., 2004). In addition, Ca2+ is dynamically regulated by a variety of morphologically eclectic acidic organelles that include the endolysosomal system, lysosome-related organelles, secretory vesicles and the Golgi complex (Patel and Docampo, 2010). Much less is known concerning Ca2+ handling by these so called acidic Ca2+ stores. This is despite the importance of Ca2+ release from endosomes and lysosomes during their fusion (Luzio et al., 2007; Pryor et al., 2000). Indeed, growing evidence links dysfunction of acidic Ca2+ stores in lysosomal storage disorders (Lloyd-Evans et al., 2008; Shen et al., 2012), acute pancreatitis (Gerasimenko et al., 2009) and Alzheimer’s disease (Coen et al., 2012).
The best characterised route for mobilisation of acidic Ca2+ stores is via the potent Ca2+ mobilizing messenger nicotinic acid adenine dinucleotide phosphate (NAADP) (Galione et al., 2010). Evidence stems from the use of the lysosomotropic agent GPN (Gly-Phe β-naphthylamide) which causes osmotic permeabilisation of lysosomes through its hydrolysis by the acid hydrolase, cathepsin C (Jadot et al., 1984) resulting in Ca2+ depletion. In many cells, GPN selectively blocks NAADP-induced Ca2+ release and cytosolic Ca2+ signals in response to NAADP-forming agonists (Brailoiu et al., 2006; Churchill et al., 2002; Galione et al., 2010; Yamasaki et al., 2004). NAADP-evoked Ca2+ signals are thought to initiate through the activation of the recently described endo-lysosomal two-pore channels (TPCs) (Brailoiu et al., 2009; Calcraft et al., 2009; Patel et al., 2010; Zong et al., 2009). The resulting release of Ca2+ via NAADP has been proposed to ‘trigger’ further Ca2+ release from ER Ca2+ stores through inositol trisphosphate (InsP3) and/or ryanodine receptors (Cancela et al., 1999; Guse and Lee, 2008). ‘Chatter’ between channels occurs via Ca2+-induced Ca2+ release (CICR) and/or through the priming of ER Ca2+ channels as a result of store overloading (Cancela et al., 1999; Churchill and Galione, 2001; Patel et al., 2001). However, several studies have concluded that NAADP may have more direct effects on the ER through activation of ryanodine receptors (Dammermann et al., 2009; Gerasimenko et al., 2003; Hohenegger et al., 2002). The potential action of NAADP at multiple Ca2+ stores, possibly though promiscuous low molecular weight NAADP-binding proteins, has hampered efforts to study the functional relationship between acidic organelles and the ER (Guse, 2012; Lin-Moshier et al., 2012; Marchant et al., 2012). Thus, whether mobilisation of lysosomal Ca2+ stores is sufficient to trigger ER Ca2+ release is not known.
Recently, we showed that mutation of a dileucine targeting motif in TPC2 (an NAADP-sensitive Ca2+-permeable channel) redirects it from its normal lysosomal location to the plasma membrane (Brailoiu et al., 2010). Plasma-membrane-targeted TPC2 was functional as evidenced by both single channel recordings and Ca2+ imaging (Brailoiu et al., 2010; Yamaguchi et al., 2011). Importantly, NAADP-evoked Ca2+ signals were largely insensitive to blockade of ER Ca2+ stores (Brailoiu et al., 2010). Thus, TPCs appeared to be uncoupled from ER Ca2+ channels despite being able to mediate global, albeit sluggish, Ca2+ increases. These data are indicative of a more intimate relationship between acidic organelles and the ER under normal conditions perhaps akin to membrane contact sites between the ER and the plasma membrane/mitochondria (Patel and Brailoiu, 2012). The latter drive processes such as excitation-contraction coupling in muscle cells and Ca2+ uptake by mitochondria through privileged communication between organelles be it via local Ca2+ changes or protein–protein interactions (Protasi, 2002; Rizzuto et al., 2004; Toulmay and Prinz, 2011). Indeed, physical coupling between TPCs and ryanodine receptors at putative lysosome–ER junctions might well harmonize conflicting views regarding NAADP action (Patel et al., 2010; Patel and Brailoiu, 2012). ‘Trigger zones’ as sites of NAADP action have been proposed but ultra-structural evidence is scant (Kinnear et al., 2004; Patel and Brailoiu, 2012).
In the present study, we recapitulate complex Ca2+ signals in human fibroblasts upon direct lysosomal permeabilisation in the absence of overt cellular stimulation. We also provide electron microscopic evidence for the existence of extensive membrane contacts between lysosomes and the ER. Our data provide new insight into cross-talk between Ca2+ stores associated with acidic organelles and the ER.
Results and Discussion
To probe the relationship between lysosomal and ER Ca2+ stores, we used primary cultured human fibroblasts. These cells are large (∼120 µm) and flat thus facilitate imaging. As shown in Fig. 1A, immunocytochemical analysis using a primary antibody raised to LAMP-1, revealed well resolved structures that were distributed throughout the cell. A similar distribution was obtained when live cells were incubated with Lysotracker Red, a fluorescent weak base that accumulates in acidic organelles (Fig. 1B). Addition of the lysosomotropic agent GPN resulted in a prompt loss of Lysotracker Red fluorescence consistent with the labelling of lysosomes (Fig. 1B). To quantify lysosomal Ca2+ content, we monitored cytosolic Ca2+ concentration in response to acute addition of GPN using cells loaded with the Ca2+ sensitive indicator fura-2. Surprisingly, despite rapid permeabilisation of lysosomes, GPN evoked long-lasting Ca2+ signals that persisted up to 15 min (Fig. 1C). Parallel measurements of Lysotracker Red fluorescence confirmed that GPN abolished Lysotracker Red fluorescence within ∼3 min of addition under these conditions (Fig. 1D). Removal of extracellular Ca2+ had little effect on Ca2+ signals evoked by GPN indicating that Ca2+ influx does not contribute significantly to GPN responses (Fig. 1E). For comparison, the effect of the SERCA pump inhibitor, thapsigargin, was used to estimate ER Ca2+ content (Fig. 1F). GPN thus evokes relatively large cytosolic Ca2+ signals that outlast lysosome integrity.
The kinetic mismatch between lysosome permeabilisation and cytosolic Ca2+ responses was further characterised by simultaneous measurements of fura-2 and Lysotracker Red fluorescence at the single cell level (Fig. 2A–C). In all 36 co-loaded cells examined (n = 3), we noted a primary linear increase in Ca2+ upon GPN stimulation which correlated with the decrease in Lysotracker Red fluorescence. We term this increase in cytosolic Ca2+ concentration the ‘pacemaker’ response. This was followed by one of three secondary responses. In most cells, the rise in cytosolic Ca2+ accelerated, resulting in a prominent Ca2+ spike. These spikes either repeated resulting in a series of Ca2+ oscillations (Fig. 2A) or were abortive (Fig. 2B). In the remainder of cells, cytosolic Ca2+ levels continued to increase monotonically primarily peaking after complete loss of Lysotracker Red fluorescence (Fig. 2C). Thus, there was marked variation in the time to peak response upon GPN stimulation. Similar results were obtained in the absence of Lysotracker Red. From all 15 experiments analysed (with and without Lysotracker Red), the proportion of oscillatory (Fig. 2A) and abortive cells (Fig. 2B) were 48±5% and 29±5%, respectively (total number of cells = 177). At the spatial level, secondary responses to GPN were non-uniform. In cells displaying repeated Ca2+ spikes, the responses derived from a particular sub-cellular location and spread throughout the cell in wave-like manner. These responses either initiated at one-end of the cell (67±5%, n = 15; Fig. 2D) or in the cell periphery (33±5%, n = 15; Fig. 2E). In the latter case, the wave propagated in a centripetal fashion resembling the closing of an iris. In all 104 oscillatory cells examined, the waves repeatedly initiated from the same sub-cellular locale. Taken together, these data indicate that GPN-evoked Ca2+ signals are biphasic comprising an initial pacemaker response, which is probably lysosome dependent and a secondary response displaying spatial heterogeneity that appears lysosome independent.
GPN-evoked Ca2+ responses are highly reminiscent of those evoked by cell-surface agonists. We therefore examined the effect of the physiological agonist bradykinin, which mobilises intracellular Ca2+ stores, on GPN responses. Bradykinin generated robust cytosolic Ca2+ signals in Ca2+-free media. Subsequent treatment with GPN induced largely monotonic responses (Fig. 3B). Thus, the proportion of cells displaying oscillations was 57±10% (n = 5, 74 cells) and 10±4% (n = 4, 60 cells) in the absence and presence of bradykinin, respectively (Fig. 3A,B). To probe the contribution of ER Ca2+ stores to GPN-evoked Ca2+ signals, cells were pre-treated with thapsigargin. As shown in Fig. 3C, GPN stimulation evoked only a transient response after thapsigargin treatment similar to bradykinin. Secondary responses were not observed in any of the 279 cells examined (n = 17). To define the role of inositol trisphosphate (InsP3) receptors, cells were pre-treated with the InsP3 receptor antagonist 2-aminoethoxydiphenyl borate (2-APB). 2-APB did not induce a Ca2+ signal (Fig. 3D). However, in its presence, the secondary responses to GPN were again largely eliminated (11±8% of oscillatory cells, n = 3, 45 cells). In contrast, GPN evoked typical complex Ca2+ signals upon blockade of ryanodine receptors by ryanodine pre-treatment (Fig. 3E; 63±13% oscillatory cells, n = 3, 45 cells). Prolonged GPN-evoked Ca2+ signals thus require InsP3 receptors likely on the ER but the possible contribution of Golgi-localised InsP3 receptors (Pizzo et al., 2011) cannot be excluded.
The ER is known to form several inter-organellar contact sites important for both Ca2+ and lipid dialogue (Toulmay and Prinz, 2011). But much less is known concerning contacts between the ER and acidic organelles. We and others recently described contact sites between endosomes/multi-vesicular bodies (MVBs) and the ER (Eden et al., 2010; Rocha et al., 2009). As shown in Fig. 4, membrane contact sites (regions of close apposition of <20 nm) were clearly visible between lysosomes and the ER. Quantification in random sections revealed that 82±2% of 140 lysosomes analysed form contact sites with the ER (n = 3). This is probably an underestimate, as lysosomes have a diameter of 200–500 nm and thus will extend over a number of experimental sections (70 nm thick) where additional contacts in different planes may have been missed. That such a large proportion of lysosomes display contacts in random sections indicates extensive physical coupling between lysosomes and the ER. Thus, a considerable fraction of the perimeter membrane of each lysosome is coupled. Consistent with our previous observations on ER–MVB contact sites, regions where fibres appear to connect opposing ER and lysosome membranes (Fig. 4A, white arrows) were visible. The fibres resemble the tethers described within ER–mitochondria membrane contact sites (Csordás et al., 2006). Additionally, we identify areas where the apposing membranes appear to make direct contact with no discernible space between them (Fig. 4B, arrowhead).
To conclude, we functionally dissect GPN-evoked Ca2+ signals into lysosomal and ER-derived components. Mobilisation of lysosomal Ca2+ is thus sufficient to evoke complex Ca2+ signals. These data provide direct support for the trigger hypothesis for NAADP action whereby physiological activation of endolysosomal Ca2+ channels is proposed to trigger secondary Ca2+ release via ER Ca2+ channels (Cancela et al., 1999). In our experiments, potential confounding effects of NAADP on ER channels are circumvented. We also demonstrate that the majority of lysosomes make contact with the ER via extensive membrane contact sites which likely represent an important site for the triggering of complex Ca2+ signals through juxtaposition of trigger and amplification Ca2+ channels (Fig. 4C). The tethered contacts we identify would facilitate CICR between TPCs and ER Ca2+ release channels (Fig. 4C, left). This is similar to the physical and functional arrangement of voltage-gated Ca2+ channels and ryanodine receptors at diadic junctions of cardiac muscle (Protasi, 2002). More direct contacts provide a basis for potential protein-protein interactions between Ca2+ release channels (Fig. 4C, right). This echoes the architecture of the excitation-contraction machinery in skeletal muscle (Protasi, 2002). Further ultra-structural and molecular definition of these novel membrane contact sites is warranted.
Materials and Methods
Human skin fibroblasts established from healthy females were maintained in DMEM supplemented with 10% v/v fetal bovine serum, 100 units/ml penicillin and 100 µg/ml streptomycin (all from Invitrogen) at 37°C in a humidified atmosphere with 5% CO2. Cells were plated on glass coverslips before experimentation.
Fibroblasts were fixed for 10 min with 4% w/v paraformaldehyde, washed three times in phosphate-buffered saline (PBS) and then permeabilised for 5 min with 40 µM β-escin. Cells were washed again (three times in PBS), and blocked for 1 h with PBS supplemented with 1% w/v BSA and 10% v/v fetal bovine serum. Fibroblasts were sequentially incubated for 1 h at 37°C with a primary anti-LAMP-1 (lysosomal-associated membrane protein 1) mouse antibody (1∶100 dilution; H4A3 clone from Developmental Hybridoma Bank) and a secondary mouse antibody conjugated to Alexa Fluor 647 (1∶100 dilution; Invitrogen) in blocking solution. Nuclei were labelled for 5 min with 1 µg/ml DAPI (4′,6-diamidino-2-phenylindole). Cells were washed three times in PBS containing 0.1% v/v Tween in between incubations and mounted onto microscope slides with DABCO (1,4-diazabicyclo[2,2,2]octane).
Experiments were performed in HEPES-buffered saline (HBS) comprising (in mM) 1.25 KH2PO4, 2 CaCl2, 2 MgSO4, 3 KCl, 156 NaCl, 10 glucose and 10 HEPES (pH 7.4; all from Sigma). For measurement of cytosolic Ca2+ concentration, fibroblasts were incubated with Fura-2AM (2.5 µM) and 0.005% v/v pluronic acid (from Invitrogen) for 1 h. For measurement of lysosome distribution/integrity, fibroblasts were incubated with 100 nM Lysotracker Red (Invitrogen) for 30 min. Cells were stimulated with 200 µM GPN (Santa Cruz Biotech), 10 nM bradykinin (Sigma), 1 µM thapsigargin, 100 µM 2-APB (2-aminoethoxydiphenyl borate) and 100 µM ryanodine (all sourced from Merck). Where indicated, Ca2+ in HBS was replaced with 1 mM EGTA (Sigma).
Epifluorescence images were captured every 3–8 s with a cooled coupled device camera (TILL Photonics) attached to an Olympus IX71 inverted fluorescence microscope fitted with a 20× objective, and a monochromator light source. Fura-2 and Lysotracker Red were excited at 340/380 nm and 560 nm and emitted fluorescence was captured using a 440 nm long pass filter and 590 nm filter, respectively.
Confocal images were captured using an LSM510 confocal scanner (Zeiss) attached to a Zeiss Axiovert 200 M inverted microscope fitted with a 63× Plan Apochromat water-immersion objective. DAPI, Lysotracker Red and Alexa Fluor 647 were excited at 364 nm, 560 nm and 633 nm and emitted fluorescence was captured using long pass 385 nm, 560–615 nm band pass or 655–719 nm band pass filters, respectively.
BSA-gold was prepared by coupling 5 nm colloidal gold (BBI) to BSA as described previously (Slot and Geuze, 1985). Fibroblasts cultured on thermanox coverslips (Agar Scientific) were pulsed for 15 min with either BSA-gold or horseradish peroxidase (Sigma) and chased for 3 h in complete medium. This procedure labels exclusively lysosomes, as shown by their degradative capacity and LAMP-1 content (Futter et al., 1996). Cells were fixed and processed for electron microscopy essentially as described (Tomas et al., 2004) except that uranyl acetate (1%, 40 min incubation) was used in place of tannic acid, and samples were embedded in TAAB 812 hard formulation resin. Samples were viewed on a JEOL 1010 TEM and images acquired by a Gatan Orius SC100B charge-coupled device camera. Contact sites between lysosome and ER membranes were defined as regions of close apposition (<20 nm).
We thank Tim Levine and Chi Li for useful discussions.
This was supported by the Biotechnology and Biological Sciences Research Council [grant number BB/K000942/1 to S.P.], Parkinson’s UK [grant number K-1107 to S.P. and A.H.S.] and an IMPACT studentship from University College London [to B.S.K.].