Piccolo Promotes Vesicle Replenishment at a Fast Central Auditory Synapse

Abstract

Piccolo and Bassoon are the two largest cytomatrix of the active zone (CAZ) proteins involved in scaffolding and regulating neurotransmitter release at presynaptic active zones (AZs), but have long been discussed as being functionally redundant. We employed genetic manipulation to bring forth and segregate the role of Piccolo from that of Bassoon at central auditory synapses of the cochlear nucleus-the endbulbs of Held. These synapses specialize in high frequency synaptic transmission, ideally poised to reveal even subtle deficits in the regulation of neurotransmitter release upon molecular perturbation. Combining semi-quantitative immunohistochemistry, electron microscopy, and in vitro and in vivo electrophysiology we first studied signal transmission in Piccolo-deficient mice. Our analysis was not confounded by a cochlear deficit, as a short isoform of Piccolo ("Piccolino") present at the upstream ribbon synapses of cochlear inner hair cells (IHC), is unaffected by the mutation. Disruption of Piccolo increased the abundance of Bassoon at the AZs of endbulbs, while that of RIM1 was reduced and other CAZ proteins remained unaltered. Presynaptic fiber stimulation revealed smaller amplitude of the evoked excitatory postsynaptic currents (eEPSC), while eEPSC kinetics as well as miniature EPSCs (mEPSCs) remained unchanged. Cumulative analysis of eEPSC trains indicated that the reduced eEPSC amplitude of Piccolo-deficient endbulb synapses is primarily due to a reduced readily releasable pool (RRP) of synaptic vesicles (SV), as was corroborated by a reduction of vesicles at the AZ found on an ultrastructural level. Release probability seemed largely unaltered. Recovery from short-term depression was slowed. We then performed a physiological analysis of endbulb synapses from mice which, in addition to Piccolo deficiency, lacked one functional allele of the Bassoon gene. Analysis of the double-mutant endbulbs revealed an increase in release probability, while the synapses still exhibited the reduced RRP, and the impairment in SV replenishment was exacerbated. We propose additive roles of Piccolo and Bassoon in SV replenishment which in turn influences the organization and size of the RRP, and an additional role of Bassoon in regulation of release probability.

Keywords: Bassoon; cochlear nucleus; endbulb of Held; readily releasable pool; short-term depression.

Figure 1

Selective expression of Piccolo (Aczonin) in central synapses. (A) Domain structure of Piccolo (dark green line) and its shorter isoform Piccolino (light green line). Magenta lines illustrate the position of antigenic peptides used to raise Piccolo antibodies employed in this study. Antibody #1 binding to the C-terminus, selectively identifies Piccolo, while antibody #2 binding to the central region recognizes both Piccolo and Piccolino. Colored asterisks denote binding domains unique to Piccolo (not present in Bassoon) with their respective binding partners written in the same color (see “Introduction” section for description). (B) Scheme of the site of investigation (not drawn to scale): the endbulbs of Held are formed by the auditory nerve fibers (central neurites of SGN) on bushy cells (BCs) of the anterior ventral cochlear nucleus (aVCN; red box). SGNs receive their input from ribbon-type active zones (AZs) of inner hair cells (IHCs, blue box) of the organ of Corti. Spherical bushy cells (SBC) and gobular bushy cells (GBC) receive different numbers of endbulbs. SGNs form bouton-like synapses on stellate cells (SC). (C) Preservation of Piccolino in the organ of Corti of PicMut mice: Maximal projection of confocal images show immunofluorescent puncta of Piccolino (C, antibody #2) in otoferlin (Otof)-labeled IHCs of the organ of Corti, while no Piccolo signal (C’, antibody #1) is found. (D) Reduced expression of Piccolo in PicMut: Maximum projection of confocal image stack of a bushy cell in PicWT (D) and PicMut (D’) labeled for Piccolo (antibody #1), Vglut1 (excitatory synapses) and Gephyrin (inhibitory synapses). (E) Reduced fluorescence intensity of Piccolo (antibody #1) puncta at the BCs of aVCN, at both the endbulb AZs and the inhibitory AZs (inhib. syn.), in PicMut (N = 3; n = 8) mice as compared to PicWT (N = 3; n = 14) mice as obtained in maximum projections of confocal images. ***p-value < 0.001, **p-value < 0.01, Wilcoxon rank sum test. (F) Unaltered fluorescence intensity of Vglut1 (F, staining excitatory synapses) and Gephyrin (F’, staining inhibitory synapses) in the confocal images used to analyze molecular composition of cytomatrix of the AZ (CAZ) proteins at BCs in the aVCN (n.s. p-value ≥ 0.05, Wilcoxon signed rank test in F, paired Student’s t-test in F’). Data information: Box and whisker plots present median, lower/upper quartiles and 10–90th percentiles. Bar plot represents mean ± SEM (error bars). N, number of animals; n, number of BCs. All scale bars −5 μm.

Figure 2

Auditory brainstem response (ABR) indicates preserved cochlear function but alteration in signal propagation at the cochlear nucleus in PicMut mice. (A) Grand averages (line) ± SEM (shaded area) of ABR waveform responses to 80 dB click stimuli at a stimulus rate of 20 Hz of PicWT (N = 7) and PicMut (N = 6) mice at P21–23. N is the number of animals. (B) Comparison of ABR wave amplitudes (calculated as the amplitude difference between peak of the wave and the following trough). In PicMut, wave I (compound action potential of spiral ganglion) has unaltered amplitude, but wave II thought to arise from activity in the cochlear nucleus demonstrates reduced amplitude. Interestingly, the later waves arising from downstream stations in the auditory pathway seem unaltered. Bar plots represent mean ± SEM (error bars). n.s. p-value ≥ 0.05, **p-value < 0.01, Student’s t-test.

Figure 3

Number of endbulbs and endbulb AZs per bushy cell in aVCN. (A) Confocal section of a bushy cell in PicWT labeled with Bassoon (Bsn; AZ marker), Calretinin (endbulbs of Held) and Vgat (inhibitory presynaptic terminals). (B) Number of endbulbs converging onto a bushy cell was quantified by visually tracing and counting Calretinin-stained endbulbs. PicWT (N = 2; n = 12) and PicMut (N = 2; n = 9) receive comparable number of endbulbs (n.s. p-value ≥ 0.05, Student’s t-test). (C) Confocal section of a bushy cell in PicWT labeled with RIM2 (AZ marker), Vglut1 (excitatory synapses) and Gephyrin (Geph, inhibitory synapses). (D) Number of endbulb AZs (approximated from the # of excitatory AZs) per bushy cell quantified by subtracting the number of inhibitory AZs (AZ marker puncta juxtaposed with Gephyrin) from the total number of AZ marker puncta. Endbulb AZ number in PicWT (N = 9; n = 43) and PicMut (N = 9; n = 47) was comparable (n.s. p-value ≥ 0.05, Wilcoxon rank sum test). Data information: Box and whisker plot presents median, lower/upper quartiles and 10–90th percentiles. Bar plot represents mean ± SEM (error bars). N, number of animals; n, number of BCs.

Figure 4

Altered molecular composition at aVCN synapses in PicMut. (A,C,E,G) Maximal projection of confocal image stacks of BCs in PicWT (left) and PicMut (right). Slices were immunolabeled for different CAZ proteins: Bassoon (A), RIM1 (C), RIM2 (E) and Munc13-1 (G) and co-stained for Vglut1 (excitatory synapses) and Gephyrin (inhibitory synapses). (B,D,F,H) Quantification of fluorescence intensity of CAZ proteins at endbulbs and (inhibitory synapses) of BCs: Bassoon fluorescence intensity (B) was significantly increased at AZs of both endbulbs and inhibitory synapses in the mutant. RIM1 (D) fluorescence intensity was significantly lower at the endbulb AZs in mutant but only tended to be lower at inhibitory AZs. RIM2 (F) fluorescence intensity tended to be reduced at all AZs, but this reached significance only at inhibitory AZs. Munc13-1 (H) fluorescence intensity tended to be slightly increased, which reached significance only at inhibitory AZs. Data information: N, number of animals; n, number of BCs. All scale bars −5 μm. All data presented as box and whisker plots (median, lower/upper quartiles, 10–90th percentiles). Statistical significance between groups was determined by either unpaired Student’s t-test (in case of normally distributed data with comparable variances between the groups) or Wilcoxon rank sum test (when data distribution did not satisfy the criteria). Normality of distribution was tested with Jarque-Bera test and variances were compared with F-test. *p-value < 0.05, ***p-value < 0.001. PicWT and PicMut samples were strictly treated in parallel and images were acquired in parallel, using same laser power, gain and microscope settings.

Figure 5

Ultrastructural analysis reveals reduced synaptic vesicle (SV) complement at PicMut AZ. (A,B) Representative electron micrographs of PicWT (A) and PicMut (B) endbulb synapse active zones (AZs) demonstrating reduced SVs in the mutant AZ. Red arrows demarcate the boundaries of the AZ. Yellow arrow heads mark a region of the PicMut AZ almost devoid of SVs. (C) Unaltered postsynaptic density (PSD) length in PicMut endbulb AZs. Box and whisker plots present grand median, lower/upper quartiles, 10–90th percentiles. n.s.—p-value ≥ 0.05, Wilcoxon’s rank sum test. Each data point represents the PSD length of individual endbulb AZs. (D) Mean SV number per 100 nm of the PSD length significantly reduced at the PicMut endbulb AZs. Data represented as mean ± SEM (**p-value < 0.01, Student’s t-test). Each data point represents the mean SV number per 100 nm of PSD for each endbulb AZ imaged. (E) Mean SV number within 0–10, 10–20, 20–30, 40–50, till 90–100 nm of the AZ perpendicular to the AZ membrane into the presynaptic cytosol normalized to PSD length. Normality of distribution was tested with Jarque-Bera test and variances were compared with F-test. Non-normally distributed data are presented as box and whisker plots, lower/upper quartiles, 10–90th percentile; *p-value < 0.05, **p-value < 0.01, ***p-value < 0.001 and were tested by Wilcoxon rank sum test. Normally distributed data were tested by Student’s t-test and are presented as bar plot and error bars represent SEM. (F) Mean SV number within 40 nm of the AZ membrane (perpendicular to the AZ membrane into the presynaptic cytosol) normalized to the length of the PSD reduced in PicMut endbulb AZs. Box and whisker plots present grand median, lower/upper quartiles, 10–90th percentiles. ***p-value < 0.001, Wilcoxon rank sum test. Each data point represents the SV number per 100 nm of PSD within 40 nm for each endbulb AZ imaged. PicWT (N = 3; n = 32) in black and PicMut (N = 2; n = 33) in red (N, number of animals; n, number of AZs).

Figure 6

Miniature EPSC amplitude and kinetics preserved in PicMut synapses. (A,B) Images of bushy cells (BC, A) and stellate cells (SC, B) filled with fluorescent dye Alexa 488 and fixed after the recording, illustrating typical BC morphology, spherical with one primary dendrite ending in a dense bush-like dendritic tree and typical SC morphology, asymmetrical with multiple far-ranging dendrites branching off in different directions (B). All four cells were recorded and imaged at different times and the images assembled together for presentation. (C) Representative traces of mEPSC: Continuous recording (left) and average (dark bold line) of all mEPSC events of the representative cell (light thin lines; right) for PicWT (C) and PicMut (C’). (D–H) Analysis of mEPSC: mEPSC amplitude (D), decay time (E), full-width at half-maximum (FWHM; F) rise time (G) and frequency (H) remained unaltered. Each data point represents the mean estimate of a given BC. Normally distributed data presented as mean (grand average of the means of all BCs) ± SEM (D–F,H; n.s.—p-value ≥ 0.05, Student’s t-test). Non-normally distributed data presented as box and whisker plots (grand median (of the means of all BCs), lower/upper quartiles, 10–90th percentiles; (G) n.s.—p-value ≥ 0.05, Wilcoxon rank sum test). PicWT N = 16; n = 22, PicMut N = 16; n = 24 (N, number of animals; n, number of BCs).

Figure 7

Reduced evoked EPSC amplitude in Piccolo-deficient endbulb of Held synapses. (A) Average traces of evoked EPSC (eEPSC) in PicWT (black) and PicMut (red) showing reduced eEPSC amplitude in the mutant. Inset: Average PicMut eEPSC trace scaled to the peak of the average wildtype trace demonstrating unaltered eEPSC kinetics in the mutant. Positive peak at onset of trace reflects the stimulation artifact. (B) Reduced eEPSC amplitude in PicMut (N = 19; n = 34) as compared to PicWT (N = 21; n = 33). Each data point represents the mean estimate of a given BC, box and whisker plots present grand median (of the means of all BCs), lower/upper quartiles, 10–90th percentiles). **p-value < 0.01, Wilcoxon rank sum test. (C–F) eEPSC kinetics: rise time (C), full-width at half-maximum (FWHM; D) and decay time (E), and synaptic delay (F) were not significantly altered between the two genotypes. Non-normally distributed data presented as box and whisker plots (grand median (of the means of all BCs), lower/upper quartiles, 10–90th percentiles; (C,D,F) n.s. —p-value ≥ 0.05, Wilcoxon rank sum test). Normally distributed data presented as mean (grand average of the means of all BCs) ± SEM (E; n.s.—p-value ≥ 0.05, Student’s t-test). PicWT N = 19; n = 28, PicMut N = 19; n = 28 (N, number of animals; n, number of BCs).

Figure 8

Analyzing vesicle pool dynamics during high-frequency stimulation at Piccolo-deficient endbulb of Held synapses. (A) Average traces of EPSCs evoked in response to 100 Hz train stimulation, recorded from PicWT (A) and PicMut (A’) endbulb synapses, illustrating characteristic fast kinetics and short-term depression of bushy cell EPSCs, which remain preserved in the mutant. Inset shows the last EPSC of the train and spontaneous activity (mEPSC events) for 100 ms after the cessation of train stimulus, demonstrating comparable asynchronous activity in PicWT (black, left) and PicMut (red, right) (A”; PicWT N = 7; n = 9, PicMut N = 6; n = 10) p-value ≥ 0.05, Student’s t-test. (B,B’) Normalized EPSC amplitude plotted against stimulus number demonstrates comparable short-term depression in PicWT (black) and PicMut (red) in response to high-frequency train stimulation at 100 Hz (B) and 200 Hz (B’). (C,C’) To estimate the readily releasable pool size (RRP), replenishment rate and release probability (P_r) using the Schneggenburger-Meyer-Neher (SMN) method, EPSCs from trains were plotted cumulatively against stimulus number and the linear fit (solid gray line) to the last ten steady-state amplitudes was back-extrapolated (dotted gray line) to the y-axis for 100 Hz (C) and 200 Hz (C’). For quantitative analysis refer to Table 1. (D,D’) To estimate the RRP and P_r using the Elmqvist and Quastel (EQ) method, absolute EPSC amplitudes from trains were plotted against cumulative amplitudes of the all EPSCs preceding the corresponding EPSC and, the linear fit (solid blue line) to the first 3–5 data points was forward-extrapolated (dotted blue line) to the x-axis for 100 Hz (D) and 200 Hz (D’). For quantitative analysis refer to Table 1. For 100 Hz: PicWT N = 19; n = 28, PicMut N = 19; n = 28. For 200 Hz: PicWT N = 13; n = 21, PicMut N = 13; n = 19. N, number of animals; n, number of BCs.

Figure 9

Recovery from short-term depression is slowed at Piccolo-deficient endbulb of Held synapses. (A) Representative traces of PicWT (upper panel, black) and PicMut (lower panel, red) endbulb synapses to illustrate the recovery protocol. Following a conditioning 100 Hz train of 50 stimuli, recovery from short-term depression was assessed by single test pulses presented after (in ms) 25, 50, 75, 100, 250, 500 (further in s) 1, 2, 4, 6, 10, 12 and 16. Inset shows the responses to the first 5 stimuli in detail. (B) Recovery plotted as mean (± SEM) EPSC amplitude in response to test pulses normalized to the first EPSC amplitude of the conditioning train. Dashed lines are double exponential fits. The time constants (τ) are provided on the graph, amplitude ratios of the two recovery components (fast/slow) were 0.15 and 0.55 for PicWT and PicMut respectively. Inset shows the first five responses in detail. **p-value < 0.01, *p-value < 0.05. Statistical significance between groups was determined by either unpaired Student’s t-test (in case of normally distributed data with comparable variances between the groups) or Wilcoxon rank sum test (when data distribution did not satisfy the criteria). Normality of distribution was tested with Jarque-Bera test and variances were compared with F-test. PicWT N = 15; n = 10–18, PicMut N = 25; n = 12–24. N, number of animals; n, number of BCs.

Figure 10

Impairment in recovery from short-term depression further aggravated at PicBsn endbulb of Held synapses. (A) Representative traces of PicWT (upper panel, black) and PicBsn (lower panel, green) endbulb synapses to illustrate the recovery protocol. Following a conditioning 100 Hz train of 50 stimuli, recovery from short-term depression was assessed by single test pulses presented after (in ms) 25, 50, 75, 100, 250, 500 (further in s) 1, 2, 4, 6, 10, 12 and 16. Inset shows the responses to the first 5 stimuli in detail. (B) Recovery plotted as mean (± SEM) EPSC amplitude in response to test pulses normalized to the first EPSC amplitude of the conditioning train. Dashed lines are double exponential fits. The time constants (τ) are provided on the graph, amplitude ratios of the two recovery components (fast/slow) were 0.15 and 0.67 for PicWT and PicBsn respectively. PicMut recovery trace shown in gray for comparison. Inset shows the first five responses in detail. Comparing PicWT and PicBsn recoveries, ***p-value < 0.001, **p-value < 0.01, *p-value < 0.05. Statistical significance between groups was determined by either unpaired Student’s t-test (in case of normally distributed data with comparable variances between the groups) or Wilcoxon rank sum test (when data distribution did not satisfy the criteria). Normality of distribution was tested with Jarque-Bera test and variances were compared with F-test. PicWT N = 15; n = 10–18, PicBsn N = 5; n = 10. N, number of animals; n, number of BCs.

Figure 11

Miniature EPSC amplitude and kinetics preserved in PicBsn endbulb synapses. (A) Representative traces of mEPSC: Continuous recording (left) and average (dark bold line) of all mEPSC events of the representative cell (light thin lines) (right) for PicWT (A) and PicBsn (A’). (B) Analysis of mEPSC: mEPSC amplitude (B), decay time (C), full-width at half-maximum (FWHM; D), rise time (E) and frequency (F) remain unaltered. Each data point represents the mean estimate of a given BC. Normally distributed data presented as mean (grand average of the means of all BCs) ± SEM (B–E; n.s. p-value ≥0.05, Student’s t-test). Non-normally distributed data presented as box and whisker plots (grand median (of the means of all BCs), lower/upper quartiles, 10–90th percentiles; (F) n.s.—p-value ≥ 0.05, Wilcoxon rank sum test). PicWT N = 16; n = 22, PicBsn N = 5; n = 13. N, number of animals; n, number of BCs.

Figure 12

Unaltered evoked EPSC amplitude in PicBsn mutants as “pseudo-rescue” of PicMut phenotype. (A) Average traces of evoked EPSC (eEPSC) in PicWT (black) and PicBsn (green) showing unaltered eEPSC amplitude in the PicBsn mice. Inset: Average PicBsn eEPSC trace scaled to the peak of the average wildtype trace demonstrating unaltered eEPSC kinetics in the mutant. Positive peak at onset of trace reflects the stimulation artifact. (B) Unaltered eEPSC amplitude in PicBsn (N = 6; n = 17) as compared to PicWT (N = 21; n = 33). Data presented as mean (grand average of the means of all BCs) ± SEM (n.s.—p-value ≥ 0.05, Student’s t-test). (C–F) eEPSC kinetics: rise time (C), full-width at half-maximum (FWHM; D) and decay time (E), and synaptic delay (F) were not significantly altered in PicBsn as compared to PicWT. Non-normally distributed data presented as box and whisker plots (grand median (of the means of all BCs), lower/upper quartiles, 10–90th percentiles; (C,D,F) n.s. p-value ≥ 0.05, Wilcoxon rank sum test). Normally distributed data presented as mean (grand average of the means of all BCs) ± SEM (E; n.s. p-value ≥ 0.05, Student’s t-test). PicWT N = 19; n = 28, PicBsn N = 6; n = 17. N, number of animals; n, number of BCs.

Figure 13

Analyzing vesicle pool dynamics during high-frequency stimulation at PicBsn endbulb of Held synapses. (A) Average traces of EPSCs evoked in response to 100 Hz train stimulation, recorded from PicWT (A) and PicBsn (A’) endbulb synapses, illustrating characteristic fast kinetics and short-term depression of bushy cell EPSCs, which remain preserved in the mutant. Inset shows the last EPSC of the train and spontaneous activity (mEPSC events) for 100 ms after the cessation of train stimulus, demonstrating a trend towards increased asynchronous activity in PicBsn (green, right) as compared to PicWT (black, left) (A”; PicWT N = 7; n = 9, PicBsn N = 4; n = 13) p-value = 0.06, Student’s t-test. (B,B’) Normalized EPSC amplitude plotted against stimulus number demonstrates greater short-term depression in PicBsn (green) as compared to PicWT (black) in response to high-frequency train stimulation at 100 Hz (B) and 200 Hz (B’). (C,C’) To estimate the readily releasable pool size (RRP), replenishment rate and release probability (P_r) using the SMN method, EPSCs from trains were plotted cumulatively against stimulus number and the linear fit (solid gray line) to the last ten steady-state amplitudes was back-extrapolated (dotted gray line) to the y-axis for 100 Hz (C) and 200 Hz (C’). For quantitative analysis refer to Table 2. (D,D’) To estimate the RRP and P_r using the EQ method, absolute EPSC amplitudes from trains were plotted against cumulative amplitudes of the all EPSCs preceding the corresponding EPSC, and the linear fit (solid blue line) to the first 3–5 data points was forward-extrapolated (dotted blue line) to the x-axis for 100 Hz (D) and 200 Hz (D’). For quantitative analysis refer to Table 2. For 100 Hz: PicWT N = 19; n = 28, PicBsn N = 6; n = 17. For 200 Hz: PicWT N = 13; n = 21, PicBsn N = 5; n = 14. N, number of animals; n, number of BCs.