MICU1 and MICU2 finely tune the mitochondrial Ca2+ uniporter by exerting opposite effects on MCU activity

Abstract

Mitochondrial calcium accumulation was recently shown to depend on a complex composed of an inner-membrane channel (MCU and MCUb) and regulatory subunits (MICU1, MCUR1, and EMRE). A fundamental property of MCU is low activity at resting cytosolic Ca(2+) concentrations, preventing deleterious Ca(2+) cycling and organelle overload. Here we demonstrate that these properties are ensured by a regulatory heterodimer composed of two proteins with opposite effects, MICU1 and MICU2, which, both in purified lipid bilayers and in intact cells, stimulate and inhibit MCU activity, respectively. Both MICU1 and MICU2 are regulated by calcium through their EF-hand domains, thus accounting for the sigmoidal response of MCU to [Ca(2+)] in situ and allowing tight physiological control. At low [Ca(2+)], the dominant effect of MICU2 largely shuts down MCU activity; at higher [Ca(2+)], the stimulatory effect of MICU1 allows the prompt response of mitochondria to Ca(2+) signals generated in the cytoplasm.

Graphical abstract

Figure 1

MICU1 and MICU2 Interact in the Mitochondrial Intermembrane Space (A) HeLa cells were transfected with MICU1-HA and/or MICU2-Flag. After 24 hr the cells were fixed and immunocytochemistry was performed with αHA, αFlag, or αTOM20 antibodies followed by incubation with Alexa 488 (pseudocolored green) and Alexa 555 (pseudocolored red) -conjugated secondary antibodies. The scale bars represent 10 μm. (B) HeLa cells were transfected with GFP and mCherry, MCU-GFP and MCU-mCherry, MICU1-GFP and MICU1-mCherry, MICU2-GFP and MICU2-mCherry, MICU2-GFP and MICU1-mCherry, and MICU1-GFP and MICU2-mCherry and analyzed after 24 hr. Images of donors and acceptors were taken before and after photobleaching of the indicated region (white box). FRET was calculated as detailed in Experimental Procedures. The histogram bar diagram shows FRET efficiency of the indicated donor and acceptor pairs. Data are presented as mean ± SD. Descriptive statistics can be found in Table S1. ^∗p < 0.05 compared to control. The scale bars represent 10 μm. (C) The same amount of mitoplasts isolated from mouse liver was subjected to the indicated proteolytic treatment, processed for electrophoresis and western blot, and probed for MICU1 and MCU.

Figure 2

MICU1 and MICU2 Dimerization (A) HeLa cells were transfected with the indicated constructs or siRNAs. HeLa cells were harvested after 48 hr of transfection, and total protein was extracted and subjected to western blotting analysis with αMICU2 and αMICU1 antibodies. SDS-PAGE was performed in the presence or absence of DTT as indicated. (B) Homo- or heterodimer formation of MICU1, MICU2, MICU1^C465A, and MICU2^C410A after overexpression. HeLa cells were harvested after 48 hr of transfection, and total protein was extracted and subjected to western blotting analysis with αFlag and αHA antibodies. SDS-PAGE was performed in the absence of DTT. (C) HeLa cells were transfected with the indicated constructs. Flag-tagged MICU2 was immunoprecipitated from whole-cell lysate with a specific αFlag-agarose-conjugated antibody. The precipitated proteins were immunoblotted with αMCU and αFlag antibodies. SDS-PAGE was performed in the absence of DTT. See also Figures S1 and S2.

Figure 3

MICU1 Acts as an Activator of MCU (A and B) In vitro expressed MCU activity was recorded in sodium-gluconate containing 5 mM EDTA (10 pM free calculated [Ca²⁺]) at −40 mV V_cis. Addition of MICU1 to either the cis or the trans side did not cause significant change in channel behavior nor an increase in open probability. (A) Representative current traces. (B) Amplitude histograms were obtained by analyses of 60 s long traces before and after addition of MICU1 to the cis side. Gaussian fits of the multipeak histogram (red), baseline current (yellow), and MCU current (green) were obtained using the Origin 7.5 program set (n = 3). (C and D) In vitro expressed MCU was incorporated into a sodium-gluconate medium containing 1 μM Ca²⁺-gluconate and activity was monitored at −120 mV. Channel activity in this medium was detectable with much lower unitary current and open probability compared to the medium containing sodium and 5 mM EDTA. Left: activity recorded before addition of MICU1. Right: activity following addition (arrowhead in open probability histogram) of MICU1. (C) Representative traces of the channel activity. (D) Open probability as a function of time was monitored and measured for the indicated times. See also Figure S3.

Figure 4

MICU1 Regulates Both the Threshold and Activation of MCU (A) [Ca²⁺]_mt measurements in intact HeLa cells overexpressing MICU1 and challenged with maximal histamine stimulation. (B) [Ca²⁺]_mt measurements in control and MICU1-silenced HeLa cells. (C) Resting mitochondrial calcium levels in control and MICU1-silenced HeLa cells evaluated through ratiometric imaging of the mitochondrial targeted GCaMP6m. (D) [Ca²⁺]_mt measurements in control and MICU1-silenced permeabilized HeLa cells perfused with 400 nM buffered [Ca²⁺]. Data are presented as mean ± SD. Descriptive statistics can be found in Table S1. ^∗p < 0.05 compared to control.

Figure 5

MICU2 Acts as an Inhibitor of MCU Activity (A) [Ca²⁺]_mt measurements in control and MICU2-silenced HeLa cells and challenged with maximal histamine stimulation. (B) [Ca²⁺]_mt measurements in intact HeLa control and MICU2-overexpressing cells. (C and D) In vitro expressed MCU activity was recorded in sodium-gluconate containing 5 mM EDTA (10 pM free calculated [Ca²⁺]) at −40 mV V_cis. Upper panels show representative traces, whereas lower panels contain amplitude histograms that were obtained by analyses of 100 s long traces before (C) and after (D) addition of MICU2 to the cis side. A similar effect was observed in another four experiments with a decrease of the open probability by ≥70% upon addition of MICU2. Data are presented as mean ± SD. Descriptive statistics can be found in Table S1. ^∗p < 0.05 compared to control. See also Figures S4 and S5.

Figure 6

Ca²⁺-Dependent Functional Effects of MICU1 and MICU2 (A) [Ca²⁺]_mt measurements in intact HeLa cells transfected with the indicated constructs or siRNA and challenged with maximal histamine stimulation. (B) [Ca²⁺]_mt measurements in intact HeLa cells overexpressing MICU1 and/or MICU2 and challenged with maximal histamine stimulation. (C) Resting mitochondrial calcium levels in control, MCU-, MCU- and MICU1-, or MCU- and MICU1- and MICU2-overexpressing HeLa cells, evaluated through ratiometric imaging of the mitochondrial targeted GCaMP6m. (D) [Ca²⁺]_mt measurements in intact HeLa cells overexpressing MICU1^EFmut and/or MICU2^EFmut and challenged with maximal histamine stimulation. Data are presented as mean ± SD. Descriptive statistics can be found in Table S1. ^∗p < 0.05 compared to control. See also Figure S6.

Figure 7

Schematic Representation of the Proposed Model In resting conditions (upper left), MICU1-MICU2 heterodimers act as the MCU gatekeeper, thanks to the prevailing inhibitory effect of MICU2; increases in calcium concentration (upper right) induce a conformational change in the whole dimer that releases MICU2-dependent inhibition and triggers MICU1-mediated enhancement of MCU channeling activity. Moreover, the removal of MICU1 with the consequent disappearance of MICU2 (lower left) leads to the loss of both gatekeeping at resting and cooperativity triggered by activation of calcium signaling. On the other hand, the selective removal of MICU2 (lower right) causes the loss of the gatekeeping mechanism but induces the formation of the MICU1-MICU1 homodimer that enhances mitochondrial calcium uptake after cellular stimulation. IMM, inner mitochondrial membrane.