Mitochondrial flashes mediated by optic atrophy 1 (OPA1) fusion protein are bioenergetic responses to stochastic drops in mitochondrial membrane potential (m) whose origin is unclear. to the expression of eight variants that differ in their ability to trigger membrane Telatinib fusion and in their resistance to apoptosis 18, 19, 20. These variants can be constitutively cleaved by the IMM peptidase OMA1 and the and in intact heart and skeletal muscle 29, 34 and shown to require a functional respiratory chain 28, 29, but their mechanistic basis remains unclear. The m drops that trigger mitoflashes are thought to reflect transient opening of the mitochondrial permeability transition pore (mPTP), but the impact of mPTP modulators on mitoflashes varies depending on the cell or tissue tested 29, 34, 35. Earlier studies proposed that m fluctuations reflect local Ca2+ or Na+ elevations 36, 37, 38, 39, opening of mPTP 36, 38, 40, 41, coupling of m to the ATP synthase 42, or switching between active and inactive states of oxidative phosphorylation 43. Hyperosmotic or oxidative stress and ROS\dependent apoptosis are consistently associated with increased mitoflash frequency, defining mitoflashes as quantitative bioindicators of cellular dysfunctions 35, 44, 45, 46. In mitoflash frequency correlates with life span and worms with low mitoflash activity live significantly longer Pllp than those with high mitoflash activity, highlighting the physiological relevance of these events 47. In this study, we sought to clarify the mechanism of mitoflash generation by tracking the transfer of protons across the IMM. Using different fluorescent pH sensors targeted to mitochondrial compartments, we consistently observed matrix alkalinization transients, establishing the pH nature of the flashes, which we renamed mitopHlashes. However, we failed to detect pH changes in the cristae or in the intermembrane space (IMS) during drops in m. We further show that OPA1 is not required for stochastic m Telatinib fluctuations but allows efficient generation of mitopHlashes, that L\OPA1 isoforms can restore flashing independently of mitochondrial fusion, and that loss of OPA1\mediated mitopHlash/m coupling is an early marker of apoptosis. Results Mitoflashes are transient matrix alkalinization events The mitoflash probe cpYFP is pH\sensitive and does not respond to superoxide 31, 33 but the notion that mitoflashes reflect bursts of superoxide still persists 29, 47, 48, 49. To confirm the pH nature of mitoflashes, we used two pH\sensitive proteins structurally distinct from cpYFP: pHred and the super\ecliptic pHluorin (spHluorin). pHred is a ratiometric pH indicator derived from the red fluorescent protein mKeima 50, and spHluorin is a non\ratiometric pH probe derived from GFP 51. We targeted these probes to the mitochondrial matrix by fusing pHred to the signal sequence of the cytochrome c oxidase subunit VIII (Cox8) and spHluorin to the C\terminus of the matrix soluble protein peptidase (MPP) or to a subunit of the CV facing the matrix (CV?) (Rieger B = 29) and in 70% Telatinib of cells expressing MPP\spHluorin (= 57) or CV?\spHluorin (= 47), indicating that the three probes report dynamic changes in matrix pH. We then tested whether mitoflashes could be recorded with these matrix pH sensors. Antiparallel fluorescence transients occurring spontaneously and randomly in space were readily detected with Cox8\pHred, reporting matrix pH flashes phenocopying the mitoflashes (Fig ?(Fig1G1G and Movie EV1). Reversible fluorescence increases corresponding to transient alkalinization transients were detected in single mitochondria of cells expressing spHluorin fused to CV? and MPP. Simultaneous recording of pH and m using these probes and tetramethyl rhodamine methyl ester (TMRM) confirmed that the matrix pH flashes always occurred concomitantly with drops in m without any lag observed between the initiation phase of these two signals (Fig ?(Fig1H1H and I, and Movie EV2). The decay phases of pH and m transients did not always mirror each other as mitochondrial potential sometimes exhibited a delayed recovery. The kinetics pH flashes reported by the new probes and by the previously validated matrix pH probe mito\sypHer were comparable, the probes reporting events of similar duration (Cox8\pHred: 7.68 1.79 s, CV?\spHluorin: 8.2 1.98 s, MPP\spHluorin: 7.35 1.26 s, and mito\sypHer: 8.6 0.6 s) and time to peak (Cox8\pHred: 3.06 1.12 s, CV?\spHluorin: 1.91 0.83 s, MPP\spHluorin: 1.70 0.44 s, and mito\sypHer: 1.63 0.6 s). The flashing area measured with all the sensors was comparable and drastically increased upon enforced mitochondrial fusion with a dominant negative mutant of dynamin\related protein 1 (DRP1K38A, data not shown),.