Vesicular ATPase inserted into the plasma membrane of motor terminals by exocytosis alkalinizes cytosolic pH and facilitates endocytosis

Abstract

Key components of vesicular neurotransmitter release, such as Ca(2+) influx and membrane recycling, are affected by cytosolic pH. We measured the pH-sensitive fluorescence of Yellow Fluorescent Protein transgenically expressed in mouse motor nerve terminals, and report that Ca(2+) influx elicited by action potential trains (12.5-100 Hz) evokes a biphasic pH change: a brief acidification (∼ 13 nM average peak increase in [H(+)]), followed by a prolonged alkalinization (∼ 30 nM peak decrease in [H(+)]) that outlasts the stimulation train. The alkalinization is selectively eliminated by blocking vesicular exocytosis with botulinum neurotoxins, and is prolonged by the endocytosis-inhibitor dynasore. Blocking H(+) pumping by vesicular H(+)-ATPase (with folimycin or bafilomycin) suppresses stimulation-induced alkalinization and reduces endocytotic uptake of FM1-43. These results suggest that H(+)-ATPase, known to transfer cytosolic H(+) into prefused vesicles, continues to extrude cytosolic H(+) after being exocytotically incorporated into the plasma membrane. The resulting cytosolic alkalinization may facilitate vesicular endocytosis.

Figure 1

Biphasic stimulation-induced changes in YFP fluorescence in motor terminals and preterminal axon. A, Fluorescence micrograph of YFP-expressing motor terminal. B, Fluorescence images of this terminal were divided by prestimulation (rest) images (average of 20), and the resulting F/F_rest images are displayed before stimulation (a), at the onset of 50 Hz stimulation (b), immediately following the end of stimulation (c), and ~80 s following the end of stimulation (d). C, Time course of fluorescence changes averaged over the entire area of this terminal, and calculated pH and [H⁺] changes. D, Average time course from 18 terminals (thick line) with 95% confidence intervals (dashed lines), plotted using the same scales as in C. The best-fit (r²=0.99) single exponent for the decay of the alkalinizing component had a time constant of 73 s (half-time 52 s). E, Another YFP-expressing motor terminal, arbitrarily subdivided into two regions (a and b) plus a region of preterminal axon (c). F, Time course of F/F_rest changes in each of the regions outlined in E. Images in B are the average of 4 images taken at the indicated time. F/F_rest plots in C and F were averaged from 5 and 8 repetitions, respectively; vertical lines superimposed on the traces indicate SEM. Preparations in Figures 1-6 were bathed in physiological saline containing 24 mM HCO₃⁻, gassed with 5% CO₂/95% O₂, bath pH 7.3; muscle contractions were blocked with d-tubocurarine. Calculations in C assumed a resting pH of 6.9, [H⁺] = 126 nM; calibration and calculations are described in Supplemental Information (Supplemental Experimental Procedures, Figure S1).

Figure 2

Stimulation-induced changes in YFP fluorescence (and [H⁺]) require Ca²⁺ entry into motor terminals. A. Responses before, during (10-20 min exposure) and 30-60 min after replacing bath Ca²⁺ with Mg²⁺. B, Responses before and following addition of 100 μM Cd²⁺ (20-60 min exposure), which blocks most Ca²⁺ channels. C, Responses before, during (20-40 min exposure) and 90-170 min after washout of ω-agatoxin GIVA (0.5 μM), which blocks P/Q-type Ca²⁺ channels. Records in A-C came from different terminals; each plotted response is the average of 4-7 repetitions in one terminal. Each plotted result is representative of experiments repeated in 4-5 terminals (3-4 animals).

Figure 3

The stimulation-induced alkalinizing phase is selectively blocked by inhibiting exocytosis (A) and by inhibiting vesicular H⁺-ATPase (vATPase, B). A, Responses before and after exposure to 20 nM BoNT A (left) or 14.5 nM BoNT E (right) for 80-140 min. B, Responses before and after exposure to 1 μM folimycin (left) or 1 μM bafilomycin (right) for 1-4 hr. Folimycin produced no significant change in resting fluorescence. Ca, averaged effects of these drugs and of vesamicol on the magnitudes of the early acidifying response (measured 3-4 s following stimulation onset) and the late alkalinizing response (measured 3-4 s following cessation of stimulation). Vesamicol (7 μM) exposure was 3 hr (n=4 terminals). Results with both BoNTs were averaged together, as were results with foli- and bafilo-mycin (fol/baf). * indicates significant difference from control (p<0.01, one-way ANOVA followed by Dunnett’s multiple comparison test). Cb, Diagram indicating sites of action of inhibitors summarized in Ca. VAChT = vesicular acetylcholine transporter. Da, fluorescence changes averaged from control (blue) and alkalinization-blocked (red) terminals, including SEM. Db, Averaged data from Da converted to Δ[H⁺]. Subtraction of alkalinization-blocked from control record yields the net contribution due to vesicular release (dashed black line).

Figure 4

The alkalinizing component is larger and decays more slowly following longer stimulation trains, and is inhibited by dynasore. A and B plot the fractional change in YFP fluorescence from a terminal stimulated with short and long 50 Hz trains (200 and 1000 stimuli, demarcated by vertical lines). Red lines in B show the measured parameters: peak acidification (measured 3-4 s after stimulation onset), peak alkalinization (5 s after stimulation stopped), and t_1/2, the time required for alkalinization to decay to half its peak amplitude. C,D and E plot these parameters for 7 terminals, each stimulated with long and short trains, with points from the same terminal connected with a line (each point was averaged from 4-8 train repetitions). C, The number of stimuli had no effect on peak acidification. D, peak alkalinization was larger following the long train (0.028 ±0.0048 SEM vs 0.016 ± 0.003). E, t_1/2 of the alkalinizing component was longer after the long train (91.2 ± 15.4 s, range 40-150 s vs. 45.8 ± 4.5 sec, range 30-61 sec). *p<0.05, ** p<0.005, paired t-test, two-tailed. F plots stimulation-induced (50 Hz, 20 s) changes in YFP fluorescence before and 30-50 min after bath application of 70 μM dynasore. Each plot is the average (± SEM) of 12 trains (3 terminals, 2 animals). The decays of the alkalinizing components were fitted with single exponentials (R²>0.98), yielding time constants of 56 s (53-59 95% confidence interval) and 290 s (275-307 95% confidence interval) for control and dynasore, respectively. The corresponding t_1/2 were 39 and 201 s.

Figure 5

Inhibition of vesicular H⁺-ATPase decreases stimulation-induced loading of FM1-43. A, Pairs (FM1-43 and α-bungarotoxin [BgTx]) of micrographs from a levator auris muscle incubated in FM1-43 without stimulation (left pair), and from another muscle whose nerve was stimulated in FM1-43 (right pair). The loading protocol is diagrammed in the inset of part B. Stimulated preparations were incubated with FM1-43fx (a fixable analog of FM1-43, 3 μM) and stimulated with 6 trains (each 50 Hz for 10 s with 10 s inter-train interval). FM1-43 was left in the bath for 15 min after stimulation ended and then washed out for 30 min. When applied, folimycin (2 μM) was present from 120 min before stimulation until 15 min after stimulation. Preparations were then fixed and post-stained with α-bungarotoxin Alexa Fluor 594 conjugate to label endplate acetylcholine receptors (see Experimental Procedures). Confocal Z stacks acquired from multiple regions of the preparation were used to construct pairs of maximal Z projections at FM1-43 and Alexa Fluor excitation/emission wavelengths (see Experimental Procedures). Regions of interest were drawn around endplates in the BgTx image and used to measure the total fluorescence intensity of the corresponding terminals in the FM1-43 image (expressed in arbitrary fluorescence units, a.u.). This method allowed non-biased identification of terminal regions, regardless of the amount of FM1-43 uptake. B, Bars show mean (± 95% confidence interval, n indicates number of terminals) of FM1-43 fluorescence in the indicated conditions. *** = p<0.001; Kruskal-Wallis ANOVA (non-parametric) and Dunn’s multiple comparison post test. C, Cumulative histograms of FM1-43 fluorescence for the terminals analyzed in B (bin size=100 a.u.).

Figure 6

Time courses of Δ[Ca²⁺] and acidifying Δ[H⁺] responses to stimulation. Changes in cytosolic [H⁺] came from the “vesicular contribution blocked” record in Figure 3Db, after inversion and scaling to facilitate comparison of Δ[Ca²⁺] and Δ[H⁺] time courses. Averaged changes in cytosolic [Ca²⁺] (normalized to resting [Ca²⁺]) were measured in 5 terminals from the fluorescence of Oregon Green 488 BAPTA 1 (OG-1) injected ionophoretically into the internodal axon (see Experimental Procedures). Insets at right show responses at the beginning and end of stimulation on an expanded time scale. [H⁺] changes developed more slowly than [Ca²⁺] changes.

Figure 7

The magnitude of the acidifying component of the stimulation-induced Δ[H⁺] response to stimulation is increased by substitution of HEPES for the standard HCO₃⁻/CO₂ pH buffer (A), by inhibition of the plasma membrane Na⁺/H⁺ exchanger with 0.5 mM amiloride (A), and by inhibition of the HCO₃⁻/CO₂ buffering system with 2 μM acetazolamide (B). Records in A were obtained sequentially from the same terminal. Addition of folimycin to the physiological control solution caused the acidifying component to peak at a later time and a greater magnitude. For records in HEPES and amiloride the preparation was gassed with 100% O₂ rather than the standard 5% CO₂/95% O₂. Symbols next to the traces indicate the peak amplitude (± SEM) of the response averaged from the indicated number of terminals exposed to the same condition. The difference between these averages was statistically significant (one-way ANOVA followed by Bonferroni’s multiple comparison test, p<0.05). Records in B are averages of 5-8 responses from a single terminal; similar results were obtained in 2 additional terminals.

Figure 8

Schematic diagram linking measured stimulation-induced changes in cytosolic pH (top panel) with hypothesized underlying changes in H⁺ and Ca²⁺ fluxes across the plasma membrane via a Na⁺/HCO₃⁻ cotransporter (NBC), a Na⁺/H⁺ exchanger (NHE), voltage-dependent Ca²⁺ channels (VDCC), the plasma membrane Ca²⁺-ATPase (PMCA, which imports H⁺), and vesicular H⁺-ATPase (vATPase) inserted into the plasma membrane by exocytosis. (HCO₃⁻ import via NBC is equivalent to H⁺ extrusion.) Boxes represent the nerve terminal plasma membrane. The height of the gray bar represents resting concentrations. The size of letters for H⁺ and Ca²⁺ represents the concentration. A, At resting steady-state, cytosolic [Ca²⁺] is maintained at ~0.1 μM by Ca²⁺ extruders (PMCA and Na⁺/Ca²⁺ exchanger [not shown]), which offset the inward leak of Ca²⁺. Cytosolic [H⁺] (pH ~7) is maintained by acid extruders (e.g. NHE, NBC, drawn in part A only) which offset H⁺ buildup from plasma membrane leak and metabolic sources. B, Action potential depolarization opens VDCC. The consequent elevation of cytosolic [Ca²⁺] has opposing effects on cytosolic [H⁺]: (i) acidification via increased operation of the PMCA and by Ca²⁺ displacement of H⁺ from intracellular binding sites; and (ii) alkalinization initiated by vATPase incorporation into the plasma membrane. Due to the relatively small number of fused vesicles at this stage, vATPase-mediated alkalinization is still insufficient to offset Ca²⁺-induced acidification. C, Cytosolic [Ca²⁺], and thus Ca²⁺-induced acidification, reach a plateau after 2-3 s of stimulation, but the number of fused vesicles, and thus vATPase-mediated H⁺ extrusion, continue to increase, becoming sufficiently large to produce a net cytosolic alkalinization which is maintained until the end of the stimulation. D, As stimulation stops, VDCCs close and cytosolic [Ca²⁺] declines to near resting levels within 3-5 s. Since endocytotic retrieval of vATPase proceeds at a slower rate, vATPases remaining in the plasma membrane continue to extrude H⁺, un-opposed by Ca²⁺-induced acidification. This produces a fast alkalinizing “jump” at the end of stimulation. Endocytotic retrieval of vATPase from the plasma membrane eventually restores cytosolic pH to its resting value.