The mitochondrial calcium uniporter is a multisubunit Ca2+ channel that mediates mitochondrial Ca2+ uptake, a cellular process crucial for the regulation of oxidative phosphorylation, intracellular Ca2+ signaling, and apoptosis. In the last few years, genes encoding uniporter proteins have been identified, but a lack of efficient tools for electrophysiological recordings has hindered quantitative analysis required to determine functional mechanisms of this channel complex. Here, we redirected Ca2+-conducting subunits (MCU and EMRE) of the human uniporter to the plasma membrane of Xenopus oocytes. Two-electrode voltage clamp reveals inwardly rectifying Ca2+ currents blocked by a potent inhibitor, Ru360 (half maximal inhibitory concentration, ~4 nM), with a divalent cation conductivity of Ca2+ > Sr2+ > Ba2+, Mn2+, and Mg2+. Patch clamp recordings further reveal macroscopic and single-channel Ca2+ currents sensitive to Ru360. These electrical phenomena were abolished by mutations that perturb MCU-EMRE interactions or disrupt a Ca2+-binding site in the pore. Altogether, this work establishes a robust method that enables deep mechanistic scrutiny of the uniporter using classical strategies in ion channel electrophysiology.
The mitochondrial calcium uniporter (henceforth denoted the uniporter) is a Ca2+ channel that imports cytoplasmic Ca2+ through the inner mitochondrial membrane (IMM) into the mitochondrial matrix (Bernardi, 1999; Gunter et al., 2000; Kirichok et al., 2004). It regulates a wide range of physiological processes, including aerobic metabolism, intracellular Ca2+ signaling, and apoptotic cell death (Rizzuto et al., 2012; Kamer and Mootha, 2015). Dysfunction of this channel has been implicated in a multitude of cardiovascular and neuromuscular pathologies (Logan et al., 2014; Kwong et al., 2015; Luongo et al., 2015; Liu et al., 2016; Penna et al., 2018). Breakthroughs in the last few years have led to identification of uniporter genes (Baughman et al., 2011; De Stefani et al., 2011; Mallilankaraman et al., 2012; Sancak et al., 2013), revealing that this channel is a protein complex containing at least four subunits: MCU, EMRE, MICU1, and MICU2 (Fig. 1). It has been shown that MCU pentamerizes to form a Ca2+ pathway traversing the IMM (Oxenoid et al., 2016), and that EMRE opens the MCU pore by associating with MCU via transmembrane (TM)–helix interactions (Sancak et al., 2013; Tsai et al., 2016). MICU1 and MICU2 are regulatory subunits in the intermembrane space (IMS), functioning to shut the uniporter when free Ca2+ in the IMS (and hence the cytoplasm) falls below 1 µM (Mallilankaraman et al., 2012; Csordás et al., 2013; Tsai et al., 2016; Kamer et al., 2017).
Fundamental properties of ion channels, including ion permeation and gating, are traditionally studied using electrophysiological recordings under voltage-clamp conditions. However, applying this approach to the uniporter has been difficult because of its limited electrical accessibility as an organellar channel. The Clapham laboratory developed a mitoplast patch clamp technique to record uniporter currents in mitochondria (Kirichok et al., 2004), but this method is of the “heroic” variety, not suitable for routine, multiply replicated experiments necessary for in-depth quantitative analysis. Recently, there have been reports of single-channel uniporter recordings in reconstituted planar bilayers (De Stefani et al., 2011; Patron et al., 2014; Wu et al., 2018). However, the quality of proteins used in these studies was unclear, and the observed electrical properties are vastly different from data obtained with the mitoplast patch clamp (Kirichok et al., 2004).
Here, we establish an efficient method using a two-electrode voltage clamp (TEVC) and patch clamp to record electrical activity of the human uniporter targeted to the plasma membrane of Xenopus laevis oocytes. We show that the uniporter mediates inwardly rectifying Ca2+ currents, inhibited by Mn2+ and blocked by a potent and specific inhibitor, Ru360 (Ying et al., 1991; Matlib et al., 1998), with a divalent cation conductivity of Ca2+ > Sr2+ and Ba2+, Mn2+, and Mg2+ currents undetectable. These electrical phenomena are strictly absent in nonfunctional mutants or uninjected controls. With this new tool, it is now possible to perform iterative electrophysiological analysis to determine fundamental molecular mechanisms of the uniporter.
Materials and methods
The human MCU-EMRE fusion construct with a C-terminal 1D4 tag (TETSQVAPA) was cloned into a pOX vector (Jegla and Salkoff, 1997). Site-directed mutagenesis was performed using the QuickChange kit (Agilent) with mutations verified by sequencing. For in vitro transcription, the pOX vector was first digested with NotI, and then 1 µg of the linearized vector was used for complementary RNA (cRNA) synthesis using the mMessage mMachine T3 transcription kit (Thermo Fisher Scientific) following the manufacturer’s instructions. cRNA was purified using the RNeasy MinElute Cleanup kit (Qiagen), dissolved in nuclease-free water to a final concentration of ∼1 µg/µl, and then stored at −80°C. The quality of cRNA was assessed by electrophoresis on denaturing agarose gels containing 4% formaldehyde. The full amino acid sequence of the human MCU-EMRE fusion construct was as follows: MAQNLEGEWQNLGAVYCSTVVPSDDVTVVYQNGLPVISVRLPSRRERCQFTLKPISDSVGVFLRQLQEEDRGIDRVAIYSPDGVRVAASTGIDLLLLDDFKLVINDLTYHVRPPKRDLLSHENAATLNDVKTLVQQLYTTLCIEQHQLNKERELIERLEDLKEQLAPLEKVRIEISRKAEKRTTLVLWGGLAYMATQFGILARLTWWEYSWDIMEPVTYFITYGSAMAMYAYFVMTRQEYVYPEARDRQYLLFFHKGAKKSRFDLEKYNQLKDAIAQAEMDLKRLRDPLQVHLPLRQIGEKDVSAAWSGSGRSLVPSRSVIVTRSGAILPKPVKMSFGLLRVFSIVIPFLYVGTLISKNFAALLEEHDIFVPEDDDDDDTETSQVAPA.
Handling of Xenopus oocytes
Xenopus laevis oocytes were purchased from Nasco and defolliculated by digestion with 1.25 mg/ml type II collagenase (Sigma-Aldrich) at room temperature for 2 h in a solution containing 96 mM NaCl, 2 mM KCl, 2.5 mM MgCl2, 5 mM HEPES, and NaOH, pH 7.4. Stage V–VI oocytes were selected, incubated in 18°C in a ND96 solution (96 mM NaCl, 2 mM KCl, 2 mM CaCl2, 0.5 mM MgCl2, 5 mM HEPES, and NaOH, pH 7.4), and were injected with 50 nl (12–40 ng for TEVC and 40–60 ng for patch clamp) of cRNA using a Nanojet II microinjector (Drummond). Electrophysiological experiments were performed 3–4 d after cRNA injection. During the revision of this paper, we received several batches of oocytes that exhibit low expression of uniporter proteins. The reason is unknown, but it has been reported previously that oocytes could exhibit some seasonal variations in quality. We note that most experiments here were performed in summer and autumn, whereas the revision was in spring.
Oocyte membrane preparations and Western blots
Xenopus oocyte plasma membranes were purified following a published protocol (Wall and Patel, 1989; Kamsteeg and Deen, 2001). In brief, 25 oocytes were homogenized using a Dounce homogenizer in 1.2-ml ice-cold homogenization buffer (HB; 20 mM Tris, 5 mM MgCl2, 5 mM NaH2PO4, 1 mM EDTA, and 80 mM sucrose) supplemented with a protease inhibitor cocktail (Roche), and centrifuged for 30 s at 13.5 g at 4°C. Then, 1 ml of the top of the sample was removed, and 1 ml HB was added. This step was repeated four times, with centrifugation changed from twice at 13.5 g to once at 24 g and then to once at 38 g. After the last centrifugation, the supernatant was removed. Plasma membranes were spun down for 20 min at 14,000 g at 4°C, and then resuspended in 50 µl HB for Western analysis.
Plasma membrane samples were loaded onto 4–20% gels for electrophoresis, and were then transferred to low-fluorescence polyvinylidene difluoride membranes (LI-COR) using a Trans-Blot Turbo transfer system (Bio-Rad). Membranes were blocked in LI-COR Odyssey blocking buffer for 30 min and incubated overnight with an anti-1D4 primary antibody (50 ng/ml, produced in house) in 10 ml TBST (150 mM NaCl, 50 mM Tris, 0.05% Tween-20, and HCl, pH 7.5), and then with a goat anti–mouse IRDye 680RD secondary antibody (1:15,000 in TBST; LI-COR) for 1 h. Western images were acquired using a LI-COR Odyssey CLx imager, and were analyzed with the associated Image Studio software (LI-COR).
TEVC signals were measured using an Oocyte Clamp OC-725B system (Warner), filtered at 1 kHz, and sampled at 2 kHz. Voltage and current electrodes were pulled and filled with 3 M KCl to have a resistance of 0.5–1.5 MΩ. Data acquisition and membrane voltage control were performed with a pCLAMP-10/Digidata-1322A system (Axon). In all experiments, membrane potentials were held at −60 mV, and repeated voltage ramps were applied as described in the text to allow detection of both uniporter currents and Ca2+-activated Cl− currents. Oocytes were continuously superfused with the ND96 solution during recordings. In experiments testing dose dependence of currents or divalent-cation conductivity, MCl2 (M = Ca2+, Sr2+, Mn2+, Ba2+, or Mg2+) and NaCl were varied in pairs as 2 and 96, 5 and 90, 10 and 85, 20 and 70, and 50 and 25 to maintain osmolarities. For experiments requiring intracellular Ca2+ chelators, oocytes were injected with 50 nl 0.1-M EGTA (buffered with 10 mM HEPES and NMG, pH 7.4), and then incubated in ND96 supplemented with 100 nM Ru360 for 10–20 min before TEVC. For recordings requiring zero external Cl−, all anions were substituted with isethionate or gluconate.
Patch clamp recordings were performed using an Axopatch 200B amplifier, interfaced with pCLAMP-10/Digidata-1440A (Axon), with signals filtered at 1 kHz and captured at 5 kHz. Electrodes were filled with 90 mM sodium gluconate, 10 mM NaCl, 5 mM EGTA, 5 mM EDTA, 20 HEPES, and NaOH, pH 7.6, with a pipette resistance of 5–8 MΩ. Oocytes were superfused in ND96 for the formation of gigaohm seals. After obtaining outside-out patches, the bath solution was changed to 100 mM CaCl2, 20 HEPES, and NaOH, pH 7.6 for recording Ca2+ currents.
All recordings were performed at room temperature. Negative (inward) currents are defined as currents flowing from the bath to the oocyte cytoplasm (or the cytoplasmic side of outside-out membranes), with bath defined as zero voltage. Raw patch clamp traces were digitally filtered at 50 Hz using the Clampfit software. Data plotting was performed using Igor Pro 7 (WaveMetrics).
Data are presented as mean ± SEM. All experiments were repeated at least three times.
Online supplemental material
Two supplemental figures are available describing the response of whole-cell currents to niflumic acid (Fig. S1), and documenting uniporter currents recorded in low Cl− environments (Fig. S2).
The mitoplast patch clamp technique is difficult because it requires the challenging maneuver of obtaining gigaohm-seal patches in small mitochondrial membranes. Our strategy to circumvent this problem is to express the uniporter's TM region, containing MCU and EMRE subunits, in plasma membranes for straightforward electrical recordings. To ensure codelivery of both MCU and EMRE to an alternative membrane system, we used a fully functional tandem construct with human EMRE fused to the C terminus of human MCU (Tsai et al., 2016; hME, Fig. 1). This construct was initially expressed in HEK 293 cells, but despite multiple attempts in varying cell culture conditions and protein engineering (i.e., deleting the mitochondrial targeting sequence and/or including an N-terminal glycophorin-A sequence), no plasma-membrane expression was detected using immunostaining of fixed cells.
We thus turned to Xenopus oocyte expression, a classical system for ion channel electrophysiology. Injecting 12 ng cRNA of mitochondrial targeting sequence–truncated hME (hereafter defined as WT hME, sequence available in Materials and methods) into oocytes led to expression of a full-length 45-kD protein in purified plasma membranes after 3–4 d of incubation (Fig. 2). TEVC analysis using repeated voltage ramps from −120 to 80 mV shows that increasing external Ca2+ from 2 to 20 mM induces large Ca2+-activated Cl− currents (ICACC; Hartzell, 1996; Yao and Tsien, 1997), which exhibit a diagnostic outwardly rectifying I–V relation (Fig. 3 A), a reversal potential of −25 mV near the Cl− equilibrium potential (Fig. 3 A), and strong responses to an ICACC inhibitor niflumic acid (NFA; White and Aylwin, 1990; Fig. S1). This ICACC is induced by Ca2+ influx via the uniporter (IMCU), as it is inhibited by Ru360 (Fig. 3 A), and is ∼10-fold smaller in uninjected controls or in oocytes expressing nonfunctional mutants E215A (E264A in MCU; Baughman et al., 2011) and G353W (G81W in EMRE; Tsai et al., 2016; Fig. 2 and 3 B).
The existence of large ICACC poses a problem for our analysis: Ru360-sensitive currents would contain both IMCU and ICACC (Fig. 3 A), thus making it impossible to specifically characterize IMCU. To overcome this problem, we applied NFA along with intracellular Ca2+ chelators (5 nmol EGTA) to inhibit ICACC, as has been done previously to unmask endogenous store-operated Ca2+ currents (Hartzell, 1996; Yao and Tsien, 1997). Under this condition, 20 mM Ca2+ elicits Ru360-sensitive IMCU with an amplitude of 0.5–2 µA at −120 mV, exhibiting an inwardly rectifying I–V relation that plateaus near zero current (Fig. 4, A and B), closely resembling uniporter currents observed in the mitoplast patch clamp. In contrast, adding Ru360 fails to affect currents recorded in control cells with no hME expression (Fig. 4 C). As an alternative means to suppress ICACC, we soaked oocytes in a Cl−-free condition for 3 d to remove 80% of intracellular Cl− (Fig. S2), and then performed TEVC recordings in a Cl−-free extracellular medium supplied with NFA. Currents induced by 20 mM Ca2+ rectify inwardly (Fig. S2), similar to observations made with oocytes containing intracellular EGTA (Fig. 4 B).
Our attempts to enhance uniporter currents by injecting more cRNA (30–40 ng) lead to enormous Ca2+-induced, outwardly rectifying ICACC approaching 100 µA at 80 mV in most oocytes. Surprisingly, in a subset of oocytes (ranging from 5–30% in six independent batches), Ca2+-activated Cl− channels appear to be inactivated: outward currents are small (1–5 µA) and are unaffected by increasing external Ca2+ or adding Ru360 (Fig. 5 A). Moreover, the I–V curve is little affected by adding NFA (Fig. S1). The mechanism underlying this Ca2+-activated Cl− channel down-regulation is unclear, but similar phenomena have been reported before when foreign proteins were expressed in Xenopus oocytes (Jorgensen et al., 1997; Kunzelmann et al., 1997). With ICACC inactivated, IMCU can now be isolated without EGTA injection (Fig. 4 B) or extensive Cl− dialysis (Fig. S2), maneuvers that often make oocyte membranes unstable. Control experiments show that IMCU is absent in cells expressing nonfunctional E215A or G353W mutants (Fig. 5 C). In TEVC experiments below, these low-ICACC oocytes were used to further characterize uniporter properties.
We first titrated Ca2+ to obtain the dose response of IMCU, yielding a Km of 22 mM (Fig. 6). Increasing Ru360 concentration in the presence of 20 mM Ca2+ progressively inhibits the uniporter with a half maximal inhibitory concentration (IC50) of 4.2 nM (Fig. 7, A and B). Recovery of Ca2+ currents after Ru360 withdrawal follows a slow time course with a time constant of 28 ± 5 s (Fig. 7 C). IMCU shows a divalent cation conductivity of Ca2+ > Sr2+ >> Ba2+, Mn2+, and Mg2+ (Fig. 8). In the presence of 10 mM Ca2+, IMCU is inhibited by Mn2+ with an IC50 of 410 µM (Fig. 9, A and B). It is also slightly reduced by 9 ± 3% by 10 mM Mg2+ (unpublished data), a result consistent with experiments using mitochondrial Ca2+ uptake (Bragadin et al., 1979) or mitoplast patch clamp (Kirichok et al., 2004).
Although TEVC is highly efficient in measuring whole-cell currents, it has certain limitations, including difficulties in controlling intracellular solutions and incompatibility with single-channel recordings. To circumvent these problems, we used the patch clamp to record hME in excised outside-out patches. In these experiments, 5 mM EGTA and EDTA were added to the pipette solution to eliminate ICACC, and 100 mM CaCl2 was placed in the bath to elicit IMCU. Of 45 patches tested, 4 exhibited macroscopic Ca2+ currents of >5 pA strongly inhibited by extracellularly applied Ru360 (200 nM), with removal of Ru360 yielding an unblock time constant of 17 ± 4 s (Fig. 10, A and B), in good agreement with the kinetics seen in TEVC. Moreover, 11 patches showed Ru360-sensitive currents of 1–5 pA, allowing observation of single-channel openings (Fig. 10, C and D). At a holding membrane potential of −80 mV, the uniporter has a single-channel amplitude of ∼0.3 pA. Ru360 does not affect the current amplitude, but greatly shortens the channel open time. None of these electrical phenomena were observed in 30 outside-out patches excised from uninjected oocytes or in 15 patches from oocytes expressing the G353W mutant. Altogether, these results demonstrate the suitability of the Xenopus oocyte system for more extensive patch clamp analysis of the uniporter in the future.
In the current study, we expressed the Ca2+-conducting TM region of the human uniporter in the form of an MCU-EMRE fusion construct in Xenopus oocyte plasma membranes. TEVC and patch clamp recordings show that these proteins produce IMCU, which is fully inhibited by Ru360 from the extracellular side. As Ru360 is known to block MCU from the mitochondrial IMS (Cao et al., 2017), we infer that hME adopts a homogeneous IMS-side-out orientation in oocyte membranes. Further characterization of IMCU reveals a Ca2+ dose response, I–V relation, single-channel amplitude, and divalent cation selectivity remarkably similar to observations made using the mitoplast patch clamp (Kirichok et al., 2004; Table 1). In addition, inhibition of IMCU by Mn2+ is a feature commonly seen in other Ca2+-selective channels (Sather and McCleskey, 2003). Finally, IMCU is fully abolished by mutations known to eliminate uniporter function in mitochondria: E215A (E264A in MCU), which destroys a Ca2+ site in the MCU pore (Baughman et al., 2011; Cao et al., 2017), and G353W (G81W in EMRE), which disrupts TM interactions between MCU and EMRE (Tsai et al., 2016). Collectively, these results unambiguously rule out electrical artifacts, and demonstrate that the human uniporter exhibits similar ion-conducting behaviors in mitochondrial and Xenopus oocyte membranes.
Three issues about our method are discussed here. First, we refrained from using the most well-known inhibitor of the uniporter, ruthenium red (Rigoni and Deana, 1986), as this hexavalent compound can nonspecifically inhibit several Ca2+ transport proteins, such as ryanodine receptors (Ma, 1993) and voltage-gated Ca2+ channels (Cibulsky and Sather, 1999). Ru360 is a divalent ruthenium amine complex that specifically and potently inhibits the uniporter (Ying et al., 1991; Matlib et al., 1998). Second, the use of a tandem hME construct neither assumes nor implies an MCU-EMRE interacting stoichiometry of 1:1—the number of EMRE subunits associated with MCU is currently unknown. The purpose of using this construct is to ensure that both MCU and EMRE travel to the plasma membrane to assemble into a functional channel. We cannot perform recordings by coinjecting MCU and EMRE cRNAs into oocytes, as MCU fails to travel to plasma membranes without EMRE fused to its C terminus. Third, the regulatory MICU subunits are entirely omitted in this study. It might seem plausible to incubate purified MICU proteins with Xenopus oocytes to form a full uniporter complex, but we were reluctant to do so because the oligomerization status of MICUs is the subject of current uncertainty: these subunits have been shown to form homodimers (Wang et al., 2014; Li et al., 2016; Kamer et al., 2017), homohexamers (Wang et al., 2014), and heterodimers with or without an intersubunit disulfide (Patron et al., 2014; Petrungaro et al., 2015; Kamer et al., 2017). It is unclear which of these can properly assemble with the Ca2+-conducting subunits to gate the uniporter. Adding MICUs to our electrical analysis would be premature before these conflicts are resolved.
In the last few years, there have been reports describing single-channel recordings of the uniporter using reconstituted planar lipid bilayers (De Stefani et al., 2011; Patron et al., 2014; Wu et al., 2018). These recordings are problematic for several reasons. First, size-exclusion chromatography has not been used to test if uniporter proteins are denatured during detergent solubilization, a very common issue in membrane-protein purification. Second, macroscopic recordings have not been reported in these studies, making it difficult to judge if observed single channels exhibit representative uniporter behaviors. This raises serious concerns as electrical properties described in these reports are vastly different from those obtained using the mitoplast patch clamp (Kirichok et al., 2004). Third, none of these recordings show reversible block by ruthenium red or Ru360. In our hands, ruthenium compounds can occasionally perturb planar bilayers, leading to irreversible loss of channel activities.
Results here also have implications in two controversies in the field. First, the TM orientation of EMRE is currently in dispute regarding whether the N terminus of this single-pass IMM protein faces IMS or the matrix (Tsai et al., 2016; Vais et al., 2016; Yamamoto et al., 2016). That the hME construct can form functional channels is a powerful demonstration that EMRE’s N terminus is located in the matrix where MCU’s C-terminal region also resides. Second, based on electrical recordings in reconstituted bilayers, it was proposed that human MCU by itself is sufficient to conduct Ca2+ (De Stefani et al., 2011; Patron et al., 2014). This argument has been undermined by mitochondrial assays showing that EMRE is also required for uniporter-mediated mitochondrial Ca2+ uptake (Sancak et al., 2013; Kovács-Bogdán et al., 2014; Tsai et al., 2016), and is unambiguously rejected here with recordings showing that the G353W mutation eliminates Ca2+ currents of hME (Fig. 5 C).
In summary, we have established an efficient Xenopus oocyte system to record electrical behaviors of the human uniporter. This achievement now permits quantitative work in the future using TEVC or patch clamp electrophysiology to shed light on fundamental mechanisms underlying the crucial roles of the uniporter in mammalian pathophysiology.
We thank Dr. Willem Laursen and Dr. Gonzalo Budelli for technical assistance with TEVC, and Dr. Paul Garrity and Dr. David Clapham for lending us some electrophysiological equipment. We particularly thank Dr. Christopher Miller for constructive suggestions and unconditional support during the development of this project.
C.-W. Tsai is supported by National Institutes of Health grant R01-GM-107023.
The authors declare no competing financial interests.
Author contributions: Both C.-W. Tsai and M.-F. Tsai performed experiments and analyzed data. C.-W. Tsai prepared the figures, and M.-F. Tsai wrote the manuscript.
Richard W. Aldrich served as editor.