Brain activity regulates loose coupling between mitochondrial and cytosolic Ca2+ transients

Abstract

Mitochondrial calcium ([Ca²⁺]_mito) dynamics plays vital roles in regulating fundamental cellular and organellar functions including bioenergetics. However, neuronal [Ca²⁺]_mito dynamics in vivo and its regulation by brain activity are largely unknown. By performing two-photon Ca²⁺ imaging in the primary motor (M1) and visual cortexes (V1) of awake behaving mice, we find that discrete [Ca²⁺]_mito transients occur synchronously over somatic and dendritic mitochondrial network, and couple with cytosolic calcium ([Ca²⁺]_cyto) transients in a probabilistic, rather than deterministic manner. The amplitude, duration, and frequency of [Ca²⁺]_cyto transients constitute important determinants of the coupling, and the coupling fidelity is greatly increased during treadmill running (in M1 neurons) and visual stimulation (in V1 neurons). Moreover, Ca²⁺/calmodulin kinase II is mechanistically involved in modulating the dynamic coupling process. Thus, activity-dependent dynamic [Ca²⁺]_mito-to-[Ca²⁺]_cyto coupling affords an important mechanism whereby [Ca²⁺]_mito decodes brain activity for the regulation of mitochondrial bioenergetics to meet fluctuating neuronal energy demands as well as for neuronal information processing.

Fig. 1

Mitochondrial morphology in the mouse M1 pyramidal neurons. Data were obtained with automated tape-collecting ultramicrotome scanning electron microscopy (ATUM-SEM). a Gallery view of SEM slices of the dendritic mitochondria. Yellow and green arrows indicate the mitochondria nanotunneling in the dendrite highlighted in pink. Scale bar: 1 μm. b Volume reconstruction of mitochondria (blue) in the dendrite (semi-transparent white) as shown by the pink area in (a). Yellow and green arrows indicate the mitochondria nanotunneling. c Image of a cell body of pyramidal neuron in M1. Scale bar: 5 μm. d 3D volume reconstruction of mitochondria in the mouse M1 neuron soma (semi-transparent white). n = 291 mitochondria identified are shown in different colors. Scale bar: 5 μm. e Statistics of mitochondrial length in dendrites and somas. n = 35 dendritic and 2465 somatic mitochondria. f Volume densities of mitochondria in dendrites and somas. n = 6 dendrites and 5 somas. Data are presented as mean ± SEM. *P < 0.05. **P < 0.01. ***P < 0.0001, unpaired T-test. Source data are provided as a Source Data file.

Fig. 2

Imaging of [Ca²⁺]_cyto and [Ca²⁺]_mito transients in cortical neurons in awake mice. a Schematic showing experimental timeline for virus injection, head holder mounting, open-skull surgery, and two-photon imaging. Genetically-encoded Ca²⁺ indicators, mito-GCaMP6f and jRGECO1a, were used to report cytosolic and mitochondrial Ca²⁺ activity, respectively. b Confocal imaging of primary motor cortex neurons labeled by mtGCaMP6f and jRGECO1a in brain slice. Most of the labeled cells were in L2/3 (~150–350 μm deep from pial). Scale bars: 50 μm. c, d Representative 2PM images of double labeling of L1 dendrites at 25 μm depth (c) and a L2/3 soma at 300 μm depth (d). Scale bars: 20 μm. e, f Representative examples of pairs of [Ca²⁺]_cyto and [Ca²⁺]_mito transients in L1 dendrites (e) and L2/3 somas (f). Upper panels: time courses of fluorescence changes. Lower panels: corresponding dual channel images at three time points marked by arrows. Note that [Ca²⁺]_mito transients display highly variable kinetics that can be mostly categorized into regular (upper panel of e), plateaued (lower panels), and staircase groups (upper panel of f). Scale bars: 10 μm (e) and 20 μm (f). g Amplitudes of [Ca²⁺]_cyto and [Ca²⁺]_mito transients in dendrites (n = 29 [Ca²⁺]_cyto events and n = 29 [Ca²⁺]_mito events from 6 mice) and somas (n = 29 [Ca²⁺]_cyto events and n = 22 [Ca²⁺]_mito events from 8 mice). h, i Distribution of duration of [Ca²⁺]_cyto and [Ca²⁺]_mito transients in dendrites (h) and somas (i) (h: n = 6 mice; i: n = 8 mice.). j Frequencies of spontaneous [Ca²⁺]_cyto and [Ca²⁺]_mito transients in dendrites (n = 9 [Ca²⁺]_cyto events and n = 21 [Ca²⁺]_mito events from 6 mice) and somas (n = 8 [Ca²⁺]_cyto events and n = 25 [Ca²⁺]_mito events from 8 mice). Data are presented as mean ± SEM. *P < 0.05. **P < 0.01. ***P < 0.0001, unpaired T-test. Source data are provided as a Source Data file.

Fig. 3

Dynamic coupling of [Ca²⁺]_mito to [Ca²⁺]_cyto transients in neurons in vivo. a Schematic showing [Ca²⁺]_cyto transients with or without coupled [Ca²⁺]_mito transients. N_cyto or N_mito: number of cytosolic or mitochondrial Ca²⁺ transients observed. Coupling fidelity is defined as N_mito/N_cyto × 100%. b Averaged [Ca²⁺]_cyto transients in dendrites (upper panel) and somas (lower panel). 10–29 traces were used for each group. c, d Correlation of amplitudes of [Ca²⁺]_cyto and [Ca²⁺]_mito transients for coupled events in dendrites (c) and somas (d). (c: n = 23 paired events from 6 mice; d: n = 17 paired events from 8 mice). e Coupling fidelity in dendrites and somas (n = 47 from 6 mice in dendrites, n = 23 from 8 mice in soma). Data are presented as mean ± SEM. *P < 0.05. **P < 0.01. ***P < 0.0001, unpaired T-test. n.s.: no significance. Source data are provided as a Source Data file.

Fig. 4

Enhanced [Ca²⁺]_mito-to-[Ca²⁺]_cyto coupling amidst exercise-elicited neuronal activity. a Schematic of transcranial two-photon imaging in the forelimb motor cortex of awake, head-restrained mice before, during and after running on treadmill. b, c Heatmaps of [Ca²⁺]_cyto and [Ca²⁺]_mito activity in dendrites and somas under different conditions. Vertical dashed lines partition records obtained in the pre-run, run, and post-run stages. Each row represents Ca²⁺ activity from a dendrite (b, n = 48 events) or a soma (c, n = 49 events). Pairs of [Ca²⁺]_cyto(upper panels) and [Ca²⁺]_mito activities (lower panels) were sorted by time of peak amplitude of [Ca²⁺]_cyto activity. Scale bars: 20 s. d Averaged [Ca²⁺]_cyto transients in dendrites (upper panel) and somas (lower panel) prior to and during treadmill running. 52–60 traces were used for each group. e Amplitudes of [Ca²⁺]_cyto and [Ca²⁺]_mito transients in dendrites (n = 18–55 dendrites from 6 mice) and somas (n = 21–60 somas from 7 mice) during pre-run and run stages. f, g Histogram of duration of [Ca²⁺]_cyto and [Ca²⁺]_mito transients in dendrites (f) and somas (g) during pre-run and run stages. (f: n = 66 dendrites from 6 mice; g: n = 67 somas from 7 mice. Mann–Whitney U-Test). h, i Exercise increased [Ca²⁺]_cyto frequency in dendrites (h, n = 7–26 from 6 mice) and somas (i, n = 11–17 from 7 mice). j Exercise increased [Ca²⁺]_mito-to-[Ca²⁺]_cyto coupling fidelity (n = 47–49 dendrites and 47–48 somas from 6–7 mice,). Data are presented as mean ± SEM. *P < 0.05. **P < 0.01. ***P < 0.0001, by one-way ANOVA and multiple comparisons. Source data are provided as a Source Data file.

Fig. 5

Temporal analysis of [Ca²⁺]_mito and [Ca²⁺]_cyto coupling. a Schematic of eight selected time windows centered on the onset of [Ca²⁺]_mito transient. Latency: the delay of the onset of a coupled [Ca²⁺]_mito-to-[Ca²⁺]_cyto event. b, c Statistics of the coupling latency in dendrites (b, n = 48 events) and somas (c, n = 49 events). d–i Distribution of the number (d, e), peak amplitude (f, g), and duration of [Ca²⁺]_cyto transients (h, i) over different time windows in dendrites (d, f, h, n = 42–60 dendrites from 6 mice) and somas (e, f, i, n = 40–60 somas from 7 mice). Data are presented as mean ± SEM. *P < 0.05. **P < 0.01. ***P < 0.0001, by one-way ANOVA and multiple comparisons. n.s.:no signicance.  Source data are provided as a Source Data file.

Fig. 6

Effects of CaMKII inhibition on [Ca²⁺]_mito-to-[Ca²⁺]_cyto coupling. a Heatmaps of [Ca²⁺]_cyto and [Ca²⁺]_mito activities in dendrites and somas with KN93 and KN92 treatment. Vertical dashed lines partition records obtained in the pre-run, run, and post-run stages. Each row represents Ca²⁺  activity from a dendrite (55 dendrites in KN93 group; 48 dendrites in KN92 group) or a soma (48 somas in KN93 group; 48 somas in KN92 group). Pairs of [Ca²⁺]_cyto and [Ca²⁺]_mito activities were sorted by time of maximum response of [Ca²⁺]_cyto activity. Scale bars: 20 s. b–e Effect of CaMKII inhibitors on [Ca²⁺]_cyto and [Ca²⁺]_mito transient frequency before, during and after running on treadmill in dendrites (b, d) and somas (c, e). n = 4–5 mice. 100 μM KN62, 100 μM KN93, 100 μM KN92 (inactive analog of KN93) in ACSF, or ACSF were administrated on the brain with flapped skull for 20 min before imaging. f, g Effect of CaMKII inhibitors on [Ca²⁺]cyto amplitude in dendrites (f) and somas (g) before and during running. n = 4–5 mice. h, i CaMKII inhibition decreased coupling fidelity in dendrites (h) and somas (i). Data are presented as mean ± SEM. *P < 0.05. **P < 0.01. ***P < 0.0001, by one-way ANOVA and multiple comparisons. Source data are provided as a Source Data file.