Mitochondrial clearance of cytosolic Ca(2+) in stimulated lizard motor nerve terminals proceeds without progressive elevation of mitochondrial matrix [Ca(2+)]

Abstract

This study used fluorescent indicator dyes to measure changes in cytosolic and mitochondrial [Ca(2+)] produced by physiological stimulation of lizard motor nerve terminals. During repetitive action potential discharge at 10-50 Hz, the increase in average cytosolic [Ca(2+)] reached plateau at levels that increased with increasing stimulus frequency. This stabilization of cytosolic [Ca(2+)] was caused mainly by mitochondrial Ca(2+) uptake, because drugs that depolarize mitochondria greatly increased the stimulation-induced elevation of cytosolic [Ca(2+)], whereas blockers of other Ca(2+) clearance routes had little effect. Surprisingly, during this sustained Ca(2+) uptake the free [Ca(2+)] in the mitochondrial matrix never exceeded a plateau level of approximately 1 microM, regardless of stimulation frequency or pattern. When stimulation ceased, matrix [Ca(2+)] decreased over a slow ( approximately 10 min) time course consisting of an initial plateau followed by a return to baseline. These measurements demonstrate that sustained mitochondrial Ca(2+) uptake is not invariably accompanied by progressive elevation of matrix free [Ca(2+)]. Both the plateau of matrix free [Ca(2+)] during stimulation and its complex decay after stimulation could be accounted for by a model incorporating reversible formation of an insoluble Ca salt. This mechanism allows mitochondria to sequester large amounts of Ca(2+) while maintaining matrix free [Ca(2+)] at levels sufficient to activate Ca(2+)-dependent mitochondrial dehydrogenases, but below levels that activate the permeability transition pore.

Fig. 1.

Changes in fluorescence of Ca²⁺indicator dyes in cytosol and mitochondria and end-plate potentials (EPPs) during repetitive stimulation of motor nerve terminals.A, ΔF/F for cytosolic OG-5N (a–c) and mitochondrial rhod-2 (d–f) recorded simultaneously during stimulation at 50 Hz (a, d) and 25 Hz (b, e), and with an intermittent pattern (c, f; 1 sec at 50 Hz alternating with 1 sec rest; 0.533 sec/image). Top horizontal barsindicate duration of stimulation. B, EPPs recorded from a different end-plate during a 50 Hz, 10 sec stimulus train like that in A, a. Top trace shows all EPPs on a slow time scale; bottom trace shows first 10 (left) and last 10 (right) EPPs sampled on a faster time scale. The ends of the muscle fiber were cut to depolarize the resting potential to approximately −40 mV; this procedure minimized contractions, enabling recordings at normal quantal content without use of nicotinic antagonists. C, Changes in net fluorescence of mitochondrial rhod-2 (net F mit_Rhod-2) in a different terminal. Left record shows two superimposed 50 Hz, 10 sec trains separated by a 10 min rest. Right record shows fluorescence in the presence of 5 μm ionomycin, first in saline containing no added Ca²⁺ and 2 mm BAPTA, then in normal 2 mm Ca²⁺ saline. Note the different time scales for the stimulation and ionomycin data. In this preparation the cytoplasm of the underlying muscle fiber was cleared by cutting muscle fiber ends in trypsin, as described in Materials and Methods.

Fig. 2.

Stimulation-induced fluorescence transients recorded with OG-5N loaded into mitochondria or cytosol.A, B, OG-5N net fluorescence in mitochondria (A) and cytosol (B) during 50 Hz stimulation in control saline, 15 min after application of oligomycin (5 μg/ml) and 5 min after addition of antimycin A1 (2 μm). Photographs show OG-5N fluorescence (green) superimposed on transmitted-light images (white arrowhead indicates motor terminal; red arrowheads indicate axon;arrows connected by dotted lines indicate muscle fiber in B, and “ghost” of cleared muscle fiber in A). In A, 2 mmphosphate was present in all solutions. C, Changes in net fluorescence of mitochondrial OG-5N in a different terminal.Left records show superimposed responses to two 50 Hz, 10 sec trains. Right records show response to changes in bath [Ca²⁺] when membranes were permeabilized with digitonin (5 μm). The preparation in digitonin was initially washed with an intracellular-like saline containing 150 mm K-gluconate, 2 mm Na-pyruvate, 2 mm Na-lactate, and 2 mm BAPTA, and then washed with a similar saline containing 2 mmCa²⁺ and no BAPTA. The large increase in net fluorescence after Ca²⁺ addition was followed by loss of fluorescence, probably caused by loss of dye from mitochondria (digitonin-induced permeabilization of mitochondrial inner membrane and/or opening of the mitochondrial permeability transition pore).D, ΔF/F for mitochondrial OG-5N in a different terminal stimulated at 50 Hz for 10 sec, 25 Hz for 20 sec, and 10 Hz for 25 sec. Trains were separated by 20 min rest intervals. Muscle fibers were cleared in A,C, and D. Imaging at 0.533 sec/image inA, C, D; 0.266 sec/image in B.

Fig. 3.

Effect of stimulation (A) and antimycin A1 (B) on mitochondrial membrane potential in a motor terminal, measured using JC-1. A,Open and filled circles show that stimulation (two 50 Hz, 10 sec trains) produced no detectable depolarization. B, Antimycin A1 (2 μm) depolarized mitochondria, but oligomycin alone (5 μg/ml) did not. Note the different time scales in A andB. The stimulus trains in A were delivered during the interval labeled control inB. Each point in B is thegreen/red JC-1 emissions ratio averaged from 40 images collected at 0.533 sec/image over a period of 21.3 sec.

Fig. 4.

Effect of blocking various Ca²⁺transport mechanisms on stimulation-induced ΔF/F transients. A, Superimposed mitochondrial OG-5N transients produced by 50 Hz stimulation in control saline (○) and 15 min after equimolar substitution of Li⁺ for Na⁺ (●) to block the plasma membrane Na⁺/Ca²⁺exchanger. (Li⁺ passes through voltage-gated Na⁺ channels and thus supports action potential propagation.) B, Superimposed mitochondrial rhod-2 transients produced by 50 Hz stimulation in control saline (○) and after addition of cyclopiazonic acid (CPA, 20 μm, 120 min) and caffeine (10 mm, 50 min) to deplete endoplasmic reticular Ca²⁺ stores (●).C, Cytosolic OG-5N transients produced by intermittent stimulation (50 Hz for 1 sec alternating with 2 sec rest) in control saline, 33 min after application of CPA (25 μm), 1 min after addition of CCCP (1 μm) to depolarize mitochondria, and 20 min after washout of both drugs. D, Cytosolic OG-5N transients produced by 50 Hz stimulation for 2 sec in control saline, 30 min after changing bath pH to 10 to inhibit the plasma membrane Ca²⁺ ATPase, and 9 min after application of CCCP (1 μm). In A–D, successive stimulation periods were separated by ≥15 min rest periods. Imaging is at 1.066 sec/image in A and B, 0.533 sec/image in C, and 0.266 sec/image in D. The muscle in A was cleared.

Fig. 5.

Effects of CGP 37157 (A) and post-train application of CCCP (B) on cytosolic OG-5N ΔF/F transients. A,a, Superimposed responses to 50 Hz, 10 sec trains before and ≥20 min after addition of CGP 37157 (●; 50 μm, prepared from 10 mm stock in DMSO). Each trace is an average of four repetitions separated by 15 min rest periods.A, b, Top graph plots average steady-state ΔF/F during the train as a function of train duration. Bottom graph plots post-train time integral of ΔF/F, measured starting at the inflection point separating fast and slow decay phases on a semilogarithmic plot and ending 50 sec after cessation of stimulation (n = 2–4 for control trains, 4 for 500 stimuli CGP train, and 1 for all other CGP trains; error bars indicate ±2 SEM). B, Superimposed traces plot ΔF/F produced in another terminal by applying CCCP (2 μm; ●) after 50 Hz trains lasting 2 sec (B, a), or 10 sec (B,b). CCCP was applied over the interval indicated by thetop horizontal bars, using a fast perfusion system that exchanged the solution in the experimental chamber. Imaging at 1.066 images/sec in both A and B.

Fig. 6.

Changes in mitochondrial (rhod-2,A–C) and cytosolic (OG-5N, D) net fluorescence during and after steady and intermittent stimulation.Vertical lines above traces mark the onset of each train, not its duration. A, Transient evoked by 50 Hz, 10 sec train administered after a 25 min rest period. Imaging was at 2.1 sec/image for first 50 images, slowing thereafter to one image every 30 sec, to measure the prolonged decay without continuous laser illumination. B, Transient evoked in another terminal by nine short trains (50 Hz, 1 sec alternating with 2 sec rest intervals) preceded by a 20 min rest period. Images were sampled at 0.533 sec/image during the first 40 sec; the same imaging rate was used after the indicated 6 and 4 min periods during which the preparation was not illuminated. The fluorescence increase evoked by the second train was larger than that produced by the first, possibly because of partial saturation of matrix buffers. C, Transient produced by trains like that in A (50 Hz, 10 sec) delivered every 5 min. This was the first stimulation applied to this terminal. Fluorescence was measured at 17.066 sec/image during the first five trains. During the next ∼15 min the same stimulation pattern continued without laser illumination. Laser illumination resumed for the illustrated five additional trains. D, Cytosolic transients in a different terminal stimulated with the same pattern as in C and plotted on the same time scale. This was the first stimulation applied to this terminal. The last train in the series is displayed on an expanded time scale at right. For each train, fluorescence was measured at 0.533 sec/image for 42 sec. The muscle inB was cleared.

Fig. 7.

Simulations of stimulation-induced changes in cytosolic and mitochondrial free [Ca²⁺] based on a model that includes reversible precipitation of a Ca phosphate salt (hydroxyapatite) in the mitochondrial matrix.A, Diagram of processes simulated in the model, including Ca²⁺ influx (I_Ca) and extrusion through the plasma membrane, Ca²⁺ uptake by the mitochondrial uniporter, Ca²⁺ extrusion via the mitochondrial Na⁺/Ca²⁺ exchanger, Ca²⁺ buffering by cytosolic (Bc) and mitochondrial (Bm) buffers, and formation of a Ca-phosphate precipitate. The concentration of free mitochondrial inorganic phosphate [(Pi) in the form of HPO₄^2-] was maintained at 500 μm by a Pi transporter (except in C,c,f). Rates and concentrations used in the simulation are described below.B, Changes in cytosolic (a–c) and mitochondrial (d–f) [Ca²⁺] during and after 50 Hz stimulation and during 25 Hz stimulation. Note the slower time scale in c and f.C, Changes in cytosolic (a–c) and mitochondrial (d–f) Ca²⁺during intermittent stimulation (1 sec at 50 Hz alternating with a 2 sec rest) under control conditions (a,d), after reducing uniporter activity by a factor of 1000 (b, e) or after eliminating Pi from the mitochondrial matrix (c, f). Note the altered ordinates in b and f. Simulations were conducted using ModelMaker (Cherwell Scientific, Oxford, UK). The rate of Ca²⁺ entry across the plasma membrane during 50 Hz stimulation was 50 μm/sec [calculated from the measured increase in average cytosolic [Ca²⁺] produced by a single action potential (20 nm; David et al., 1997) using an assumed ratio of 50 for bound/free Ca in cytosol]. Ca²⁺ extrusion across the plasma membrane, Ca²⁺ uptake via the mitochondrial uniporter, and mitochondrial extrusion of Ca²⁺ were all assumed to follow Michaelis-Menten kinetics (plasma membrane extrusion V_max = 10 μm/S, K_m = 0.1 μm (first order); uniporterV_max = 500 μm/S,K_m = 1 μm (third order dependence on cytosolic [Ca²⁺]); mitochondrial Ca²⁺ extrusion V_max = 12.5 μm/S, K_m = 3 μm (first order). Total buffers and bufferK_d were 50 and 1 μm for cytosol and 5000 and 5 μm for mitochondrial matrix, respectively, with the ratio of cytosolic volume/mitochondrial volume = 10. Hydroxyapatite exhib-its variable stoichiometry and complex solubility behavior; the empirical ion product [Ca²⁺ ][HPO₄^2-] = 400 μm² was used to define saturation at pH >7 (Neuman and Neuman, 1958, calculated from their Fig. II-5). When this apparent solubility product was exceeded, precipitation was assumed to occur at a rate proportional to the relative supersaturation of the solution (Nancollas et al., 1989). Solvation was assumed to occur at a slow constant rate. Values for which estimates were not available in the literature were adjusted to yield plateau values of cytosolic and mitochondrial free [Ca²⁺] similar to those measured during 50 Hz stimulation.