- Record: found
- Abstract: found
- Article: found
PRMT1-mediated methylation of MICU1 determines the UCP2/3 dependency of mitochondrial Ca 2+ uptake in immortalized cells
research-article
Author(s):
Corina T. Madreiter-Sokolowski
1 ,
Christiane Klec
1 ,
Warisara Parichatikanond
1 ,
Sarah Stryeck
1 ,
Benjamin Gottschalk
1 ,
Sergio Pulido
2 ,
Rene Rost
1 ,
Emrah Eroglu
1 ,
Nicole A. Hofmann
1 ,
Alexander I. Bondarenko
1 ,
Tobias Madl
1
,
3
,
4 ,
Markus Waldeck-Weiermair
1 ,
Roland Malli
1 ,
Wolfgang F. Graier
a
,
1
Publication date (Electronic):
19 September 2016
Journal:
Nature Communications
Publisher:
Nature Publishing Group
Read this article at
There is no author summary for this article yet. Authors can add summaries to their articles on ScienceOpen to make them more accessible to a non-specialist audience.
Abstract
Recent studies revealed that mitochondrial Ca 2+ channels, which control energy flow, cell signalling and death, are macromolecular complexes that basically consist of the pore-forming mitochondrial Ca 2+ uniporter (MCU) protein, the essential MCU regulator (EMRE), and the mitochondrial Ca 2+ uptake 1 (MICU1). MICU1 is a regulatory subunit that shields mitochondria from Ca 2+ overload. Before the identification of these core elements, the novel uncoupling proteins 2 and 3 (UCP2/3) have been shown to be fundamental for mitochondrial Ca 2+ uptake. Here we clarify the molecular mechanism that determines the UCP2/3 dependency of mitochondrial Ca 2+ uptake. Our data demonstrate that mitochondrial Ca 2+ uptake is controlled by protein arginine methyl transferase 1 (PRMT1) that asymmetrically methylates MICU1, resulting in decreased Ca 2+ sensitivity. UCP2/3 normalize Ca 2+ sensitivity of methylated MICU1 and, thus, re-establish mitochondrial Ca 2+ uptake activity. These data provide novel insights in the complex regulation of the mitochondrial Ca 2+ uniporter by PRMT1 and UCP2/3.
Abstract
MICU1 is a regulatory subunit of mitochondrial Ca
2+ channels that shields mitochondria from Ca
2+ overload. Here the authors show that MICU1 methylation by PRMT1 reduces Ca
2+ sensitivity, which is normalized by UCP2/3, re-establishing mitochondrial Ca
2+ uptake activity.
Related collections
Most cited references53
- Record: found
- Abstract: found
- Article: not found
MICU1 encodes a mitochondrial EF hand protein required for Ca2+ uptake
Fabiana Perocchi, Vishal M. Gohil, Hany S. Girgis … (2010)
Mitochondrial calcium uptake plays a central role in cell physiology by stimulating ATP production, shaping cytosolic calcium transients, and regulating cell death. The biophysical properties of mitochondrial calcium uptake have been studied in detail, but the underlying proteins remain elusive. Here, we utilize an integrative strategy to predict human genes involved in mitochondrial calcium entry based on clues from comparative physiology, evolutionary genomics, and organelle proteomics. RNA interference against 13 top candidates highlighted one gene that we now call mitochondrial calcium uptake 1 (MICU1). Silencing MICU1 does not disrupt mitochondrial respiration or membrane potential but abolishes mitochondrial calcium entry in intact and permeabilized cells, and attenuates the metabolic coupling between cytosolic calcium transients and activation of matrix dehydrogenases. MICU1 is associated with the organelle’s inner membrane and has two canonical EF hands that are essential for its activity, suggesting a role in calcium sensing. MICU1 represents the founding member of a set of proteins required for high capacity mitochondrial calcium entry. Its discovery may lead to the complete molecular characterization of mitochondrial calcium uptake pathways, and offers genetic strategies for understanding their contribution to normal physiology and disease.
- Record: found
- Abstract: found
- Article: not found
EMRE is an essential component of the mitochondrial calcium uniporter complex.
Yasemin Sancak, Andrew Markhard, Toshimori Kitami … (2013)
The mitochondrial uniporter is a highly selective calcium channel in the organelle's inner membrane. Its molecular components include the EF-hand-containing calcium-binding proteins mitochondrial calcium uptake 1 (MICU1) and MICU2 and the pore-forming subunit mitochondrial calcium uniporter (MCU). We sought to achieve a full molecular characterization of the uniporter holocomplex (uniplex). Quantitative mass spectrometry of affinity-purified uniplex recovered MICU1 and MICU2, MCU and its paralog MCUb, and essential MCU regulator (EMRE), a previously uncharacterized protein. EMRE is a 10-kilodalton, metazoan-specific protein with a single transmembrane domain. In its absence, uniporter channel activity was lost despite intact MCU expression and oligomerization. EMRE was required for the interaction of MCU with MICU1 and MICU2. Hence, EMRE is essential for in vivo uniporter current and additionally bridges the calcium-sensing role of MICU1 and MICU2 with the calcium-conducting role of MCU.
- Record: found
- Abstract: found
- Article: not found
MICU1 and MICU2 Finely Tune the Mitochondrial Ca2+ Uniporter by Exerting Opposite Effects on MCU Activity
Maria Patron, Vanessa Checchetto, Anna Raffaello … (2014)
Introduction Mitochondrial calcium homeostasis bears unique properties in the orchestrated flow of the cation across plasma membrane and organelle channels. Indeed, the key biological function of mitochondria, oxidative phosphorylation, requires the steady presence of a very large electrical gradient across the ion-impermeable inner membrane (∼180 mV), which represents a continuous, huge driving force for the accumulation of a divalent cation into the matrix (Rizzuto et al., 2012). At the same time, whereas pulsatile, transient increases in matrix [Ca2+] ([Ca2+]mt), which occur in stimulated cells, are essential for stimulating Ca2+-sensitive dehydrogenases and upregulating ATP production (Jouaville et al., 1999; McCormack et al., 1990), continuous Ca2+ accumulation in resting conditions would on the one hand imply vicious cycling of the cation, with conspicuous energy drain, and on the other cause morphological alterations underlying the release of apoptotic cofactors and/or the bioenergetic dysfunction of necrosis (Rizzuto et al., 2012). Accordingly, the mitochondrial calcium uniporter (MCU) (Baughman et al., 2011; De Stefani et al., 2011), the uptake route of Ca2+ into mitochondria, has long been known to have a sigmoidal response to [Ca2+], with very low uptake rates at the resting [Ca2+] of the cytoplasm, which rapidly rise when the [Ca2+] in the proximity of mitochondria reaches a threshold value (>10 μM) (Csordás et al., 1999, 2010; Giacomello et al., 2010; Rizzuto et al., 1998). Thus, after its recent molecular identification, attention was directed toward the gatekeeping mechanisms of the mitochondrial Ca2+ channel, given the fundamental relevance of this process for the suppression of mitochondria-dependent cell-death pathways. Two recent elegant papers (Csordás et al., 2013; Mallilankaraman et al., 2012) showed that molecular ablation of MICU1 leads to mitochondrial Ca2+ accumulation at low, near to resting cytosolic [Ca2+], thus indicating that this protein, originally identified as a positive effector of MCU (Perocchi et al., 2010), is necessary also for inhibiting its activity at low [Ca2+]. This observation, which implied a complex functional interaction between MCU and MICU1, was the starting point of this study, in which we demonstrate that MICU2 is the genuine gatekeeper of MCU. Indeed, MICU2 inhibits the channel activity of purified MCU in planar lipid bilayers and, accordingly, reduces channel opening at resting [Ca2+] in intact HeLa cells. In contrast, MICU1 has a stimulatory role in MCU activity, both in electrophysiological analyses of purified MCU and in agonist-challenged intact cells. In cells, MICU2 forms an obligate heterodimer with MICU1 located in the mitochondrial intermembrane space. In MICU1-silenced cells, MICU2 is also eliminated and mitochondrial Ca2+ gatekeeping is abolished, in agreement with previous reports (Csordás et al., 2013; Mallilankaraman et al., 2012). Results MICU1 and MICU2 Form a 95 kDa Dimer through a Disulfide Bond To unravel the extraordinary complexity of the role of MICU1 in the control of the mitochondrial calcium uniporter, we investigated the possibility that MICU1 is part of a higher-order complex including molecules with different effects on MCU function. MICU1 ablation could thus destabilize such a complex, accounting for a dual effect on MCU activity (Csordás et al., 2013). In search of the putative partner, we drew our attention to MICU2, a recently identified homolog of MICU1 (Plovanich et al., 2013) that differs in molecular weight (45 kDa versus 50 kDa of MICU1) and has a sequence similarity of 42% at the protein level. Both MICU1 and MICU2 showed exclusive mitochondrial distribution, as revealed by colocalization with the mitochondrial marker Tom20 (Figure 1A). Moreover, in order to confirm the previously reported MICU1-MICU2 interaction (Plovanich et al., 2013), we performed a Förster resonance energy transfer (FRET)-based protein-protein interaction assay in living cells. We thus generated and imaged different combinations of GFP- and mCherry-tagged MICU1 and MICU2 proteins. Significant FRET (evaluated by acceptor photobleaching as previously described; Raffaello et al., 2013) was observed in all tested conditions, thus confirming that MICU1-MICU2 interaction occurs in situ in living cells (Figure 1B). Finally, we checked for the intraorganellar distribution of the MICU1-MICU2 complex: proteinase K protection experiments in isolated mitoplasts indicated localization in the mitochondrial intermembrane space (Figure 1C), in agreement with previous reports (Csordás et al., 2013). We then investigated whether MICU1 and MICU2 are present as higher molecular weight complexes in conditions in which disulfide bonding is preserved. We thus carried out immunoblot analyses in reducing and nonreducing conditions. In the latter case, MICU1 immunoblotting revealed essentially only a 95 kDa band (instead of the predicted 50 kDa band of the monomer), whereas in the presence of 100 mM dithiothreitol (DTT) the immunoblot revealed the unique presence of the 50 kDa monomeric form (Figure 2A). Similarly, MICU2 immunoblotting revealed the same 95 kDa protein in nonreducing conditions, suggestive of MICU1-MICU2 heterodimerization, and a 45 kDa band, corresponding to the predicted molecular weight of the MICU2 monomer, in reducing conditions (Figure 2A). We then independently silenced MICU1 and MICU2. In the former case, as expected, the MICU1 immunoblot showed a drastic reduction of the 95 kDa and 50 kDa bands of the nonreducing and reducing gels, respectively. Strikingly, when MICU1 was silenced, the MICU2 immunoblot revealed a dramatic reduction also of MICU2. This was not due to an off-target effect of the siRNA used, because the MICU2 mRNA levels were unaffected (Figure S1A available online), indicating that MICU1 is required for MICU2 stability, as also previously reported (Plovanich et al., 2013). Conversely, when MICU2 was silenced, the reduction of MICU2 levels was not paralleled by a MICU1 decrease. Rather, a higher molecular weight band (∼100 kDa) appeared, suggestive of MICU1 homodimerization (Figure 2A). To test the hypothesis that the MICU1-MICU2 interaction depends on the formation of a disulfide bond, we generated epitope-tagged mutants of MICU1 and MICU2 in which a highly conserved cysteine was mutated (MICU1C465A and MICU2C410A). Then, epitope-tagged wild-type or mutant MICU1 or MICU2 was overexpressed alone or in combination and the effect on the formation of MICU1-MICU2 dimers was investigated. When MICU1-HA was overexpressed alone, it appeared as a prominent 100 kDa dimer, with a minor fraction appearing as a monomer (Figure 2B). If, however, MICU2 was coexpressed, then the prominent band was at 95 kDa (corresponding to the MICU1-MICU2 heterodimer), again with a minor fraction as a monomer. Conversely, when MICU2 was overexpressed alone, it appeared mostly as a 45 kDa monomer, with a very faint 95 kDa band corresponding in size to the MICU1-MICU2 heterodimer (Figure 2B). These data not only confirm the notion that MICU1 and MICU2 dimerize but also that MICU2, distinct from MICU1, only forms MICU1-MICU2 heterodimers. Then we expressed the MICU1C465A and MICU2C410A mutants. When expressed alone, they gave rise only to the 50 kDa and 45 kDa bands of MICU1 and MICU2, respectively. The same results, however, were obtained when they were coexpressed with the wild-type MICU2 and MICU1 counterparts, indicating that both conserved cysteine residues are necessary for the formation of MICU1-MICU2 dimers (Figure 2B). Moreover, mutation of the other cysteine residues in MICU1 did not impair dimer formation (Figure S1B), confirming the specificity of the Cys465-dependent dimeric structure. The notion was corroborated by coimmunoprecipitation data. HeLa cells were cotransfected with epitope-tagged MICU1 and MICU2 moieties (MICU1-HA, MICU1C465A-HA, MICU2-Flag, and MICU2C410A-Flag) and lysed, and then the antibody to Flag-tagged MICU2 immunoprecipitated MICU1 and the antibody to HA-tagged MICU1 immunoprecipitated MICU2. Conversely, when the indicated cysteine residues were mutated no coimmunoprecipitation occurred, indicating that the MICU1-MICU2 interaction was lost (Figure S1C). The last piece of information we derived from these biochemical analyses was the structural requirements for the interaction of the MICU complex with MCU. Based on recent, detailed analysis of MCU topology (Martell et al., 2012), which proposed that only a short protein loop of MCU is exposed in the intermembrane space (EYSWDIMEP), the interaction domain could in principle be mapped to a very defined region. For this purpose, we took advantage of an MCU mutant (MCUD260N,E263Q) in which ion permeation through MCU was altered by modifying critical residues of this short sequence (Raffaello et al., 2013). Strikingly, whereas wild-type MCU could efficiently immunoprecipitate both MICU1 and MICU2, the MCUD260N,E263Q mutant was almost unable to bind MICU1-MICU2 (Figure S2A). For the interaction with MCU, the MICU1-MICU2 dimer appeared critical. Indeed, when dimer formation was abolished (e.g., by overexpressing MICU2C410A-Flag), MCU was not coimmunoprecipitated by the anti-Flag antibody (Figure 2C). Notably, MICU1 could still interact with MCU in the absence of MICU2, as demonstrated by coimmunoprecipitation carried out in MICU2-silenced cells (Figure S2B). Moreover, we verified that the overexpression of both the wild-type and mutant MICU1 and MICU2 did not alter the assembly of the MCU-containing complex through blue native PAGE (Figure S2C). Overall, these data suggest that MICU2 necessarily dimerizes with MICU1. We thus analyzed the functional role of this complex by dissecting the genuine contribution of MICU1 and MICU2 to the regulation of MCU in the next set of experiments. MICU1 Increases MCU Open Probability in Planar Lipid Bilayers We initially characterized the effect of MICU1 on the isolated MCU protein. For this purpose, recombinant MCU and MICU1 were produced (Figure S3A), and then MCU was inserted into planar lipid bilayers and electrophysiological recordings were carried out, as previously reported (De Stefani et al., 2011; Raffaello et al., 2013). To rule out the effects of Ca2+ itself on MICU1 activity, all experiments were carried out with Na+ as the permeating ion. When MCU currents were recorded in a sodium-gluconate low-divalent solution (10 pM calculated [Ca2+]), channel activity with a conductance of 55 pS (versus 6.2 pS as previously reported; De Stefani et al., 2011) could be observed at negative voltage ranges applied to the cis compartment (Figure 3A, left), in accordance with the known characteristic of calcium channels to allow the flux of Na+ upon removal of Ca2+ (Talavera and Nilius, 2006). Indeed, in agreement with results on MCU studied in mitoplasts (Kirichok et al., 2004), the conductance of the channel was significantly higher in the low-divalent solution (Figure 3A) than in 100 mM Ca2+ (De Stefani et al., 2011), and even a small increase of [Ca2+] resulted in the decrease of the unitary Na+ current (Figure S3B). Under low-divalent conditions (10 pM Ca2+), addition of MICU1 to MCU did not alter channel activity in terms of conductance and open probability (1.2 ± 0.3-fold change; Figures 3A and 3B). In contrast, when MICU1 was added to MCU recorded in sodium-gluconate containing ∼1 μM free calcium a clear effect was observed within tens of seconds (Figures 3C and 3D), namely the increase in open probability (3.1 ± 0.9-fold increase), whereas the conductance and sensitivity to ruthenium red inhibitor were unchanged (Figure S3C). Overall, a clear stimulatory effect of MICU1 on MCU activity was observed in the presence of Ca2+ but no inhibitory effect at low or absent Ca2+, arguing against a direct gatekeeping mechanism. Both MICU1 Silencing and Overexpressing Increase Mitochondrial Ca2+ Loading We thus proceeded to studies in intact HeLa cells, in which MICU1 was either overexpressed or silenced. In the first case, the MICU1-HA construct was recombinantly expressed in HeLa cells. Parallel [Ca2+]mt measurements showed a marked enhancement of the transient rise evoked by stimulation with the inositol 1,4,5-trisphosphate-generating agonist histamine (100 μM; Figure 4A), in agreement with the stimulatory effect on MCU revealed by electrophysiological analyses. Then, MICU1-silencing experiments were carried out by using specific siRNAs. Under those conditions, the histamine-dependent [Ca2+]mt peaks were markedly greater (174% increase in MICU1 overexpression, 156% increase in MICU1 silencing), as revealed by a coexpressed aequorin-based mitochondrial Ca2+ probe (Figure 4B). In addition, mitochondrial Ca2+ loading of nonstimulated cells, measured by coexpressing a high-affinity GFP-based fluorescent Ca2+ indicator (2mtGCaMP6m), was markedly increased (430% of control; Figure 4C). Finally, the intrinsic mitochondrial Ca2+ uptake efficiency of MICU1-silenced HeLa cells was verified in permeabilized cells. After permeabilization in EGTA-containing Ca2+-free intracellular buffer (IB/EGTA), Ca2+ accumulation was initiated by switching the perfusion buffer to IB containing an EGTA-buffered [Ca2+] of 400 nM. MICU1-silenced cells showed an 8-fold greater rate of Ca2+ accumulation (Figure 4D). As an additional control, we verified that MICU1 silencing does not alter mitochondrial morphology (Figure S4) or the expression levels of other components of mitochondrial calcium uptake machinery such as MCU and the recently identified EMRE (Figure S1A). Overall, in agreement with previous studies (Csordás et al., 2013; Mallilankaraman et al., 2012), MICU1 silencing appears associated with the release of a constitutive MCU inhibition. However, considering that the electrophysiological data suggested that MICU1 itself is not the inhibitory subunit and that MICU1 silencing causes the ablation of the whole MICU1-MICU2 dimer, we focused our attention on the detailed role of MICU2 in the control of mitochondrial calcium uptake. MICU2 Is the Genuine Negative Regulator of MCU We thus investigated the effect of MICU2 silencing on agonist-dependent [Ca2+]mt transients of intact HeLa cells. We first evaluated the effect of MICU2 silencing, namely the condition that not only removes the MICU1 partner, possibly endowed with an inhibitory effect on MCU, but also causes the formation of homodimers of the MCU stimulator, MICU1. Indeed, in MICU2-silenced cells, the [Ca2+]mt peak evoked by stimulation with 100 μM histamine was markedly greater (Figure 5A). Again, this effect is not secondary to changes in mitochondrial morphology (Figure S4) or in the expression of MCU and EMRE (Figure S1A). We then proceeded to the analysis of MICU2-overexpressing HeLa cells. When these cells were challenged with histamine (100 μM), the [Ca2+]mt peak was significantly reduced (Figure 5B). These data, indicating an inverse correlation between MICU2 levels and MCU activity, strongly argue for a role of MICU2 as the authentic MCU gatekeeper. We thus proceeded to the direct investigation of the channel activity of purified MCU in the presence or absence of MICU2 purified in Escherichia coli (Figure S5). In order to fully appreciate the gatekeeping effect (i.e., channel activity at low [Ca2+]), measurements were carried out in a sodium-gluconate low-divalent solution (10 pM calculated [Ca2+]) as described above. In these conditions, addition of MICU2 inhibits the MCU-dependent current (Figures 5C and 5D). Overall, the data support the notion that MICU2 is a bona fide MCU inhibitor. Functional Crosstalk between MICU1 and MICU2 Finally, we investigated the functional interaction between MICU1 and MICU2, and the putative regulatory mechanism of the heterodimer. In untransfected cells, the MICU1-MICU2 heterodimer was the only detected protein form (see Figure S6 for quantification of protein levels in these series of experiments). When MICU1 was overexpressed, the MICU1 homodimer was predominant (higher 100 kDa band) and [Ca2+]mt transients were increased (Figure 6A). When MICU1 was overexpressed in MICU2-silenced cells, the MICU1 homodimer was the only available form, and histamine-induced [Ca2+]mt peaks were even higher (150% greater than in control; Figure 6A). When MICU1 and MICU2 were both overexpressed (thus leading to an increase of the heterodimer only), the MICU1-dependent potentiation of [Ca2+]mt responses was almost abolished (Figure 6B). We thus tested the MICU2 gatekeeping properties by overexpressing its target, MCU. MCU-overexpressing cells not only showed greater agonist-induced [Ca2+]mt peaks (De Stefani et al., 2011) but also higher resting [Ca2+]mt values (Figure 6C), most likely due to the imbalanced overexpression of the channel without its gatekeeper. In this configuration, coexpression of MCU and MICU1 failed to recover normal resting [Ca2+]mt, showing just a modest reduction, probably due to partial MICU2 stabilization (Plovanich et al., 2013). On the contrary, the concomitant expression of MCU, MICU1, and MICU2, which substantially increased the MICU1-MICU2 heterodimers, efficiently rescued normal values of basal [Ca2+]mt, thus demonstrating the MICU2 requirement of guaranteeing gatekeeping capabilities (Figure 6C). Calcium-Dependent Functional Regulation of the MICU Complex As for the regulatory mechanisms, both the gatekeeper (MICU2) and the stimulator (MICU1) have conserved EF-hand Ca2+ binding sites in their sequences that sense the [Ca2+] increase of an agonist-stimulated cell. Mutation of these sites (see Experimental Procedures for details) demonstrated their critical role for the function of both proteins, in agreement with the electrophysiological data. For MICU1, the mutant acted as a dominant negative, as overexpression of MICU1EFmut reduced the [Ca2+]mt peaks (Figure 6D), thus genuinely uncovering the loss of cooperativity upon MCU opening (Csordás et al., 2013). The [Ca2+]mt peaks were also markedly reduced with the MICU2 EF-hand mutant (MICU2EFmut), indicating that Ca2+ binding relieves the inhibitory effect of MICU2 on MCU (Figure 5D). Interestingly, the effect of the two mutations was additive, further confirming the nonredundancy of MCU regulation by MICU1 and MICU2 (Figure 6D). Discussion The recent identification of MCU has opened the way to the molecular elucidation and experimental manipulation of a key event in calcium signaling, but has also raised new intriguing biological questions. The well-established notion that MCU is a highly selective channel with an enormous Ca2+ carrying capacity (Kirichok et al., 2004) coupled to the constant presence of a huge driving force for cation accumulation into the matrix would suggest that mitochondria are being overloaded with Ca2+ in most conditions. However, experimental evidence shows that this is not the case; not only does resting [Ca2+]mt almost match that of the cytosol (∼100 nM) but mitochondria also promptly respond to any Ca2+-mobilizing agonist by transiently increasing [Ca2+]mt over a wide range, an event that is instrumental to specific organelle functions. It is thus clear that MCU needs on the one hand a gatekeeper that prevents channel opening at resting [Ca2+], avoiding energy dissipation and cell-death triggering, and on the other hand an activation mechanism that guarantees mitochondrial calcium uptake even when increases in [Ca2+]cyt are spatially and temporally limited. We here demonstrated that this sigmoidal Ca2+ response of the organelle Ca2+ uptake pathway is not intrinsic to the channel per se but is rather provided by the interaction of the MICU complex with MCU. Whereas three different MICU isoforms exist in vertebrates, only two are conserved in protozoa and plants. Moreover, whereas MICU1 and MICU2 have broad tissue expression, MICU3 appears to be expressed at significant levels only in the nervous system. This suggests a tissue-specific role for MICU3, with MICU1 and MICU2 representing core regulators of MCU. The intimate relationship between MICU1 and MICU2 is supported by the demonstration that they form a dimer through a covalent disulfide bond (Figure 2A) at a specific, conserved, and homologous cysteine residue (Figures 2B, S1B, and S1C). Moreover, the dimer appears to be the elementary unit interacting with MCU, as demonstrated by several pieces of experimental evidence: (1) the dimer is apparently the only endogenous available form (Figure 2B); (2) when used as bait, MCU can immunoprecipitate the dimeric form only (Figure S2A); and (3) dimerization-incompetent mutants of MICU2 cannot immunoprecipitate MCU (Figure 2C). However, higher-order MICU complexes cannot be excluded, in particular considering that the functional MCU channel is a multimer (Raffaello et al., 2013) and both MICU1 and MICU2 possess EF-hand domains (Plovanich et al., 2013), suggestive of Ca2+-induced conformational changes. Very recent data suggest that the interaction of the MICU complex with MCU is mediated by a component of the channel named EMRE (Sancak et al., 2013). This is compatible with our model, because we do not exclude that the interaction between MICU1-MICU2 dimers could be settled by other elements. As for the function of MICU1, apparently conflicting reports exist in the literature. MICU1 had been initially identified as a protein necessary for mitochondrial calcium uptake (Perocchi et al., 2010), thus suggesting a positive regulation of the channel itself. Later, Madesh and colleagues showed that MICU1-depleted cells have higher resting [Ca2+]mt with no apparent differences in stimulus-induced mitochondrial calcium uptake (Mallilankaraman et al., 2012), thus clearly ascribing to MICU1 the role of MCU gatekeeper. Finally, Hajnóczky and coworkers added another piece to the puzzle, demonstrating a dual role of MICU1, confirming its inhibitory effect at low [Ca2+] but also uncovering MICU1-dependent cooperativity in MCU opening (Csordás et al., 2013). Thus, a highly complex picture emerged, with a single protein acting as both a channel closer and a channel activator, depending on [Ca2+]. However, all these studies preceded the identification of MICU2 (Plovanich et al., 2013). Strikingly, the silencing of MICU1 leads to a concomitant disappearance of MICU2 at protein (but not mRNA) level, indicating that MICU2 requires MICU1 for its stability, but not vice versa (Figures 2B and S1A). Most importantly, this opens the possibility that the apparent complex effect of MICU1 could be unexpectedly due to the loss of a second regulator. We thus tested this hypothesis in a clean in vitro system, the planar lipid bilayer, and in situ in living cells. In fact, the apparent dual role of MICU1 is the result of a binary complex composed by a bona fide inhibitor, MICU2, and an MCU activator, MICU1. Indeed, specific silencing of MICU2 leads to an increase in mitochondrial calcium transients, indicating the release of an inhibitory mechanism (Figure 5A). Consistently, its overexpression leads to a decrease of mitochondrial calcium uptake (Figure 5B), which is rather small in absolute terms, most likely because in this context MICU1 represents a bottleneck for MICU2 overexpression. Indeed, in the planar lipid bilayer, a condition where protein stability is not an issue, the inhibitory effect of MICU2 is unleashed and becomes clear (Figures 5C and 5D). Interestingly, this inhibition can be appreciated when sodium is used as the permeating ion (thus mimicking low basal [Ca2+]), further supporting the gatekeeping role of MICU2. Finally, we tested this hypothesis in living cells by measuring mitochondrial calcium levels in unstimulated cells. In these conditions, MCU overexpression substantially increased resting [Ca2+]mt, most likely due to the overexpression imbalance of MCU and MICU1/MICU2 levels. In this background, the reintroduction of only MICU1 cannot restore normal Ca2+ load, but the concomitant expression of MICU1 and MICU2 completely recovers the resting [Ca2+]mt value (Figure 6C), indicating that MICU2 is required for gatekeeping. As for MICU1, assessing its genuine role in living cells is complicated by the impossibility of specifically downregulating its expression without compromising MICU2 stability. However, its overexpression clearly shows a stimulation of mitochondrial calcium uptake (Figure 4A) paralleled by the appearance of a specific band that represents the MICU1-MICU1 homodimer (Figures 2B and S6). Along the same lines, the specific silencing of MICU2 led to the obvious decrease of the MICU1-MICU2 heterodimer together with the appearance of the MICU1-MICU1 homodimer (Figures 2A and S6). Interestingly, this condition leads to an increase of mitochondrial calcium responses that is quantitatively similar to the MICU1 overexpression alone (Figure 4A). In addition, overexpressing MICU1 in a MICU2-silenced background makes the MICU1-MICU1 homodimer the only available form (Figure S6); importantly, this leads to an increase of mitochondrial calcium uptake that is even larger than that due to the overexpression of the channel itself (De Stefani et al., 2011), thus indicating that the MICU1-MICU1 homodimer has a genuine stimulatory effect on mitochondrial calcium uptake (Figure 6A). Finally, MICU1 causes a net increase in channel open probability when added to a planar lipid bilayer (Figure 1), arguing for a bona fide stimulatory effect on MCU. Finally, in order to investigate the dynamic mechanism of this complex, we pointed our attention to the most obvious regulatory domain found in both MICU1 and MICU2, namely the EF hands. Mutation of these domains in either MICU1 or MICU2 leads to a decrease in mitochondrial calcium responses, indicating that Ca2+ itself is necessary for the regulation of the MICU1-MICU2 complex. On one hand, it appears necessary to activate the stimulatory effect of MICU1, because (1) MICU1EFmut not only fails to increase mitochondrial calcium uptake in cells (Figure 6D) but also prevents MICU1-mediated enhancement of MCU opening, as previously reported (Csordás et al., 2013), and (2) MICU1 does not increase MCU open probability in planar lipid bilayers when measurements are carried out in a sodium-only-containing media (Figures 3A and 3B). On the other hand, Ca2+ is not necessary for the inhibitory effect of MICU2, as demonstrated by the strong decrease of MCU current in the planar lipid bilayer when sodium is the only available cation (Figures 5C and 5D). This notion is indeed confirmed in situ, where MICU2EFmut strongly decreases mitochondrial calcium uptake, thus indicating that MICU2 Ca2+-binding domains are necessary to release MCU inhibition (Figure 6D). Interestingly, the effect of the two mutations was additive, confirming specialized roles in MCU regulation by MICU1 and MICU2. Our data indeed suggest that increases in [Ca2+] triggered by activation of Ca2+ signaling have a dual role in the MICU complex: channel closure due to MICU2 is relieved and, in parallel, stimulation of MCU opening by MICU1 is triggered. In conclusion, the data identify MICU2 as the genuine gatekeeper of MCU and unravel the mechanism responsible for the sigmoidal Ca2+ response of the organelle Ca2+ uptake pathway. MICU1 and MICU2, direct modulators of the pore-forming subunit (MCU) with opposite effects on channel activity, form a regulatory dimer. At low [Ca2+], the prevailing inhibitory effect of MICU2 ensures minimal Ca2+ accumulation in the presence of a very large driving force for cation accumulation, thus preventing the deleterious effects of Ca2+ cycling and matrix overload. As soon as extramitochondrial [Ca2+] increases, Ca2+-dependent MICU2 inhibition and MICU1 activation guarantee the prompt initiation of rapid mitochondrial Ca2+ accumulation (Figure 7), thus stimulating aerobic metabolism and increasing ATP production in stimulated cells (Denton, 2009). The remarkable clarification, in a few years, of the molecular machinery of a crucial checkpoint of Ca2+ signaling provides insight for understanding, and potentially affecting, disease conditions (Logan et al., 2014) in which ATP deficiency and/or apoptotic or necrotic cell death play a pathogenetic role. Experimental Procedures All experiments were performed in HeLa cells. MICU1 and MICU2 were silenced using specific siRNAs. Mock vectors (i.e., pcDNA3.1 or pEGFP-N1) were used as controls in all overexpression experiments. In parallel, a nontargeting siRNA (i.e., siRNA-scrambled) was used as a control in all silencing experiments. We verified that any siRNA causes an ∼20% drop in mitochondrial calcium uptake when compared to cells transfected with the aequorin probe only or the aequorin probe transfected together with a mock vector (referred to as the “control” condition). Because this modest decrease can be reliably appreciated with a variety (>5) of control siRNAs (with both scrambled and universal nontargeting sequences) we tested, we concluded this is due to a nonspecific effect whenever RNAi machinery is stimulated; as a consequence, a control siRNA (siRNA-scrambled) is always present in all silencing experiments because it represents the appropriate control. Aequorin Measurements For measurements of [Ca2+]cyt and [Ca2+]mt, HeLa cells grown on 13 mm round glass coverslips at 50% confluence were transfected with cytosolic (cytAeq), mitochondrial (mtAeq), or low-affinity mitochondrial (mtAeqMut) probes (as previously described; Pinton et al., 2007) together with the indicated siRNA or plasmid. pcDNA3.1 was used as a control for transfection unless otherwise indicated. The coverslip with the cells was incubated with 5 μM coelenterazine for 1–2 hr in KRB (Krebs-Ringer modified buffer: 125 mM NaCl, 5 mM KCl, 1 mM Na3PO4, 1 mM MgSO4, 5.5 mM glucose, 20 mM HEPES [pH 7.4]) at 37°C supplemented with 1 mM CaCl2, and then transferred to the perfusion chamber. All aequorin measurements were carried out in KRB. Agonists and other drugs were added to the same medium as specified in the text. The experiments were terminated by lysing the cells with 100 μM digitonin in a hypotonic Ca2+-rich solution (10 mM CaCl2 in H2O), thus discharging the remaining aequorin pool. The light signal was collected and calibrated into [Ca2+] values by an algorithm based on the Ca2+ response curve of aequorin at physiological conditions of pH, [Mg2+], and ionic strength, as previously described (Pinton et al., 2007). Representative traces are shown in the figures, whereas the full data set is included in Table S1. Alternatively, aequorin measurements were carried out on a PerkinElmer EnVision plate reader equipped with a two-injector unit. Cells were transfected as described above in 24-well plates and then replated into 96-well plates (1:5 dilution) the day before the experiment. After reconstitution with 5 μM coelenterazine, cells were placed in 70 μl of KRB and luminescence from each well was measured for 1 min. During the experiment, histamine was first injected at the desired concentration to activate calcium transients, and then a hypotonic, Ca2+-rich, digitonin-containing solution was added to discharge the remaining aequorin pool. Output data were analyzed and calibrated with a custom-made macro-enabled Excel workbook. In the experiments with permeabilized cells, a buffer mimicking the cytosolic ionic composition (IB) was employed: 130 mM KCl, 10 mM NaCl, 2 mM K2HPO4, 5 mM succinic acid, 5 mM malic acid, 1 mM MgCl, 20 mM HEPES, and 1 mM pyruvate (pH 7) at 37°C. IB was supplemented with either 100 μM EGTA (IB/EGTA) or a 2 mM EGTA-buffered [Ca2+] of the indicated concentration (IB/Ca2+). Calculated [Ca2+]free was predicted with CHELATOR software (Schoenmakers et al., 1992) and confirmed fluorimetrically with the Fura2 free acid form. HeLa cells were permeabilized by a 1 min perfusion with 100 μM digitonin (added to IB/EGTA) during luminescence measurements. Mitochondrial Ca2+ uptake speed was calculated as the first derivative by using the SLOPE Excel function and smoothed for three time points. The higher value reached during Ca2+ addition represents the maximal Ca2+ uptake speed. All of the results are expressed as means ± SD, and Student’s t test was used for the statistics. All the materials were from Sigma-Aldrich unless otherwise specified. Electrophysiology Planar lipid bilayer experiments were performed as described previously (De Stefani et al., 2011; Raffaello et al., 2013). Briefly, bilayers of ∼150–200 pF capacity were prepared using purified soybean azolectin. The standard experimental medium was 100 mM Ca2+-gluconate and 10 mM HEPES (pH 7.2). Gluconate was used as the anion to exclude anion-channel activity. As a low-divalent solution, 100 mM Na-gluconate, 10 mM HEPES (pH 7.2), and 5 mM EDTA was used. The calcium-containing solution was the same but EDTA was omitted, resulting in 1 μM free calcium as measured by atomic absorbance spectroscopy and confirmed fluorimetrically with Calcium Green-5N (Life Technologies). For calculation of the free-calcium concentration, WebMaxC version 2.2 was used (http://www.stanford.edu/∼cpatton/webmaxcS.htm). Control experiments with empty membrane or with detergents used for the purification showed no activity in any of the above solutions. All voltages reported are those of the cis chamber, zero being assigned by convention to the trans (grounded) side. Currents are considered positive when carried by cations flowing from the cis to the trans compartment. Data were acquired at 100 μs/point, filtered at 300 Hz, and analyzed offline using the pClamp program set (Axon Instruments). Histograms were fitted using the Origin 7.5 program set (OriginLab). Mitochondrial Targeted GCaMP6m Measurements HeLa cells were grown on 24 mm coverslips and transfected with 2mtGCaMP6m-encoding plasmids together with the indicated constructs. After 24 or 48 hr, coverslips were placed in 1 ml of KRB and imaging was performed on a Zeiss Axiovert 200 microscope equipped with a 63×/1.4 N.A. Plan Apochromat objective. Excitation was performed with a DeltaRAM V high-speed monochromator (Photon Technology International) equipped with a 75 W xenon arc lamp. Images were captured with a high-sensitivity Evolve 512 Delta EMCCD (Photometrics). The system is controlled by MetaMorph 7.5 (Molecular Devices) and was assembled by Crisel Instruments. In order to test resting [Ca2+]mt with high sensitivity, we developed a calcium probe based on the last-generation GCaMP probe (Chen et al., 2013) targeted to the mitochondrial matrix. We chose the GCaMP6m version, due to its highest calcium affinity, with its Kd (167 nM) in the range of normal resting [Ca2+]mt. In order to be truly quantitative, we took advantage of the isosbestic point in the GCaMP6m excitation spectrum: we experimentally determined in living cells that exciting GCaMP6m at 410 nm leads to fluorescence emission, which is not Ca2+ dependent. As a consequence, the ratio between 474 and 410 nm excitation wavelengths is proportional to [Ca2+] while independent of probe expression. HeLa cells were thus alternatively illuminated at 474 and 410 nm and fluorescence was collected through a 515/30 nm band-pass filter (Semrock). Exposure time was set to 200 ms at 474 nm and to 400 ms at 410 nm, in order to account for the low quantum yield at the latter wavelength. At least ten fields were collected per coverslip, and each field was acquired for 10 s (1 frame/s). Analysis was performed with the Fiji distribution of ImageJ (Schindelin et al., 2012). Both images were background corrected frame by frame by subtracting mean pixel values of a cell-free region of interest. Data are presented as the mean of the averaged ratio of all time points. Immunofluorescence and FRET were performed on a Leica TCS-II SP5 STED CW system equipped with a 100×/1.4 N.A. Plan Apochromat objective. Statistical data are presented as mean ± SD unless otherwise specified; significance was calculated by Student’s t test, ∗p < 0.05. Full procedures and associated references are available in the Supplemental Experimental Procedures.
Author and article information
Journal
Journal ID (nlm-ta): Nat Commun
Journal ID (iso-abbrev): Nat Commun
Title:
Nature Communications
Publisher:
Nature Publishing Group
ISSN
(Electronic):
2041-1723
Publication date
(Electronic):
19
September
2016
Publication date Collection: 2016
Volume: 7
Electronic Location Identifier: 12897
Affiliations
[1
]Center for Molecular Medicine, Institute of Molecular Biology and Biochemistry, Medical
University of Graz , Harrachgasse 21/III, Graz 8010, Austria
Author notes
Author information
Emrah Eroglu
http://orcid.org/0000-0002-9373-0808
Tobias Madl
http://orcid.org/0000-0002-9725-5231
Wolfgang F. Graier
http://orcid.org/0000-0003-1871-3298
Article
Publisher Item ID: ncomms12897DOI: 10.1038/ncomms12897
PMC ID: 5031806
PubMed ID: 27642082
SO-VID: 84bbe8fa-f8fe-4f7d-b618-ae104470feb9
Copyright © Copyright © 2016, The Author(s)
License:
This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/
History
Date
received
: 18
February
2016
Date
accepted
: 12
August
2016
Categories
Subject:
Article
ScienceOpen disciplines: Uncategorized
Data availability:
ScienceOpen disciplines: Uncategorized