Mouse chromaffin cells have two populations of dense core vesicles

Abstract

The quantal hypothesis states that neurotransmitter is released in discrete packages, quanta, thought to represent the neurotransmitter content of individual vesicles. If true, then vesicle size should influence quantal size. Although chromaffin cells are generally thought to have a single population of secretory vesicles, our electron microscopy analysis suggested two populations as the size distribution was best described as the sum of two Gaussians. The average volume difference was fivefold. To test whether this difference in volume affected quantal size, neurotransmitter release from permeabilized cells exposed to 100 microM Ca2+ was measured with amperometry. Quantal content was bimodally distributed with both large and small events; the distribution of vesicle sizes predicted by amperometry was extremely similar to those measured with electron microscopy. In addition, each population of events exhibited distinct release kinetics. These results suggest that chromaffin cells have two populations of dense core vesicles (DCV) with unique secretory properties and which may represent two distinct synthetic pathways for DCV biogenesis or alternatively they may represent different stages of biosynthesis.

FIG. 1

A quantitative analysis of dense core vesicle (DCV) size reveals a striking heterogeneity within individual cells. A: random sections from cells that possessed multiple DCVs and well-fixed cellular constituents (as shown here) were selected for analysis (filled arrows highlight DCVs at the cell membrane and open arrowheads point to mitochondria.). B: typically the larger DCVs (open arrows) possessed excess membrane giving a halo appearance around the vesicle’s dense core, and the cores of smaller vesicles (filled arrow) often appeared to have a tightly associated membrane. C: plots the size distribution for the corrected vesicle diameter (see Experimental procedures), which was measured from an individual cell. There appear to be two populations of vesicles centered at 187 and 330 nm, which have been fit with the sum of 2 Gaussian distributions.

FIG. 2

Electron microscopic (EM) serial sections confirm the existence of 2 populations of vesicles. A: shows a section from a chromaffin cell enriched in DCVs. The image in A is 1 of 5 consecutive sections (60-nm-thick sections). The circled regions are presented at higher magnification in sections (B1) and (C2). B, 1–3: 2 large DCVs that span multiple sections. B2: the vesicles at their widest diameter (filled arrows). B1 and B3: the vesicles at less than their greatest diameter (open arrows). C, 1–3: DCVs that are smaller than those presented in B. C2: the smaller vesicles at their widest diameter (filled arrows). C1 and C3: the vesicles at a reduced diameter (open arrows), and the lower vesicle is completely absent in C3. D–F: vesicle size measurements made from the cell presented above are plotted as the widest, apparent and corrected diameters. D: plots the vesicle diameter measured from the same sections used in D, but this time all DCV profiles were measured in each section to yield the apparent diameter. The distribution is described as the sum of small and large vesicles with diameters of 126 and 197 nm, respectively. E: plots the corrected vesicle diameter, made by correcting for sectioning errors (see RESULTS) (Parsons et al. 1995). The corrected distribution is fit best as the sum of 2 Gaussians, with diameters centered at 154 and 235 nm. F: the distribution of widest diameters is best fit as the sum of 2 Gaussians, with small vesicles centered at 134 nm and the larger vesicles centered at 233 nm.

FIG. 3

Quantitative analysis of quantal size reveals 2 populations of vesicles within individual cells. A: shown here is a representative 2-min amperometric trace with multiple events. Cells were permeabilized with digitonin (20 μM) and then exposed to Ca²⁺ (100 μM). On average there were ~20 events/min detected at the electrode. B: the amperometric events from A are shown on an expanded time scale. Each amperometric event has been aligned such that they all begin at the same time. The events differ considerably in amplitude and overall shape. C: the area under each of the amperometric events, like those shown in B, was integrated to obtain the quantal size (Q). Plotted in C is a frequency histogram for Q^1/3 collected from a single cell. The data are best fit as the sum of 2 Gaussian distributions. The 1st group of events is centered at 0.41 pC^1/3. The 2nd group is centered at 0.74 pC^1/3.

FIG. 4

Measurements from multiple cells revealed a strong correlation between vesicle size and quantal size. A: the corrected vesicle size distribution. The results represent data from 11 cells (1 section per cell, total of 2,390 vesicle profiles; mean: 227.0 nm), and the apparent vesicle diameters were transformed to corrected diameters using the algorithm of Parsons et al. (1995) (see Experimental procedures). The resulting distribution is best described as the sum of 2 Gaussians (r² = 0.99; single Gaussian: r² = 0.93; for additional comparison of models, see Table 2), which represent small and large vesicles centered at 179 and 310 nm, respectively. B: the cumulative Q^1/3 distribution measured from 27 cells (totaling 3,847 events; mean: 0.322 pC^1/3). The quantal size distribution is fit best as the sum of 2 Gaussians (r² = 0.99; single Gaussian: r² = 0.93; also see Table 2). The small and large Q^1/3 distributions are centered at 0.40 and 0.69 pC^1/3, respectively. The vesicle size and quantal size distributions appear very similar in overall shape, and the ratio of small and large distributions reflects a fivefold difference in volume and quantal size (see Table 1). The data used for each figure was pooled from multiple experiments. To ensure each experiment was not distinct from the total pooled values, data from each experiment was compared with the total pooled values using a 1-way ANOVA test. None of the individual experiments was found to be significantly different (alpha = 0.05) from the pooled data, and the average P value was 0.514.

FIG. 5

Ensembles of small events have different kinetics than do those of large events. A: groups of ~50 amperometric events were averaged in 7 ranges (in pC^1/3): 0.20–0.25, 0.25–0.30, 0.3–0.4, 0.4–0.5, 0.5–0.6, 0.6–0.7, 0.7–0.8, and 0.8–0.9. I₁: the same data on expanded abscissa and ordinate. I₂: 1 small event group (denoted by **) superimposed with a large event group (denoted by *), after scaling them to the same amplitude.

FIG. 6

Kinetic properties are distinct for small and large quantal sizes. A: the relationship between the amplitude and Q of individual events. The data are best fit by a composite of 2 separate linear processes. The fits for small and large events were made using data in regions where overlap was minimal. Small events could be fit with a line that exhibits a modest slope compared with that used to fit large events. B: analysis of individual spikes reveals the rise time (○) and half-width (■) are separable into 2 components. Both rise time and half-width increase in value and follow one another closely over the range where small events predominate, from 0.2 to 0.4 pC^1/3. Rise time and half-width decline slightly in the 0.6- to 0.85-pC^1/3 range and then become more dispersed at larger values. Smaller events tend to have a slower rate of transmitter release relative to large events as seen in the rapid rise in amplitude for the large events. Data are presented as means ± SE.

FIG. 7

Quantal size distributions made with data from nicotine-stimulated cells were similar to those from digitonin-permeabilized cells. A: Q^1/3 distribution for a single nicotine-stimulated cell. This cell was exposed to 10 μM nicotine for 3 min. B: Q^1/3 distribution for a digitonin-permeabilized cell. This cell was permeabilized with 20 μM digitonin and then exposed to 100 μM Ca²⁺ for 2 min. C: a cumulative Q^1/3 distribution from 6 nicotine-stimulated cells. D: a cumulative Q^1/3 distribution from 9 digitonin-permeabilized cells. The nicotine and digitonin experiments were carried out in the same group of cells. In all cases, the distributions were best fit as the sum of 2 Gaussians. Please note that even though the quantal sizes were similar, there were small kinetic differences observed for amperometric events obtained with the 2 stimulation techniques.