Clathrin-mediated endocytosis near active zones in snake motor boutons

Abstract

We have used the activity-dependent probe FM1-43 with electron microscopy (EM) to examine endocytosis at the vertebrate nerve-muscle synapse. Preparations were fixed after very brief neural stimulation at reduced temperature, and internalized FM1-43 was photoconverted into an electron-dense reaction product. To locate the reaction product, we reconstructed computer renderings of individual terminal boutons from serial EM sections. Most of the reaction product was seen in 40-60 nm vesicles. All of the labeled vesicles were clathrin-coated, and 92% of them were located within 300 nm of the plasma membrane, suggesting that they had undergone little processing after retrieval from their endocytic sites. The vesicles (and by inference the sites) were not dispersed randomly near the plane of the membrane but instead were clustered significantly near active zones. Additional reaction product was found within putative macropinosomes; these appeared to form from deep membrane invaginations near active zones. Thus two mechanisms of endocytosis were evident after brief stimulation. Endocytosis near active zones is consistent with the existence of local exo/endocytic cycling pools. This mechanism also might serve to maintain alignment of active zones with postsynaptic folds during periods of activity when vesicular and plasma membranes are interchanged.

Fig. 1.

Electron micrographs (65 nm tissue sections) comparing two activity-dependent endocytic probes. A, Horseradish peroxidase (HRP) provided satisfactory contrast to visualize active zones (AZs, white arrows) and vesicles labeled with reaction product (LVs). All LVs were clathrin-coated (black arrowheads), but some coated vesicles were not labeled (white arrowhead). Coated pits (black arrows) and a deep membrane invagination (asterisk) are also present in this section. Note the affinity of HRP for the plasma membrane, particularly coated pits and AZs. B, Photoconversion reaction product of FM1-43 (without post-staining). LVs were clathrin-coated (black arrowheads). Coated pits (black arrow) were visible, but putative AZs (white arrow) exhibited low contrast without post-staining (compare Fig.2A), possibly because FM1-43 rinsed more easily from the AZ densities than did HRP. Note LVs in B and coated pit in A that were close to the Schwann cell (SC; see Results). m, Mitochondrion. Scale bar, 500 nm.

Fig. 2.

Size distribution of LVs. A, Virtually all LVs were clathrin-coated (black arrowheads). Note LV and coated pit (black arrow) near AZ (white arrow;asterisk marks center of postjunctional fold). B, C, Examples of LVs that were larger than the unlabeled vesicles surrounding them. D, Large coated pit near AZ (white arrow) is possible precursor to “doublet” LVs, as shown in B and C.E, Distribution of vesicle sizes in 57 sections from one animal, showing substantial contribution of larger LVs to population. Effective diameters were calculated as measured perimeters/π. Shown are FM1-43 preparations with post-staining. Scale bar, 100 nm.

Fig. 3.

Endosomes and deep membrane invaginations in FM1-43 preparations. A, Two views of a deep membrane invagination, rendered from eight EM sections. Black arrow points to budding vesicle. B, One of five contiguous EM sections from rendering in A, showing continuity with synaptic cleft (arrowhead); note clathrin coat on bud (arrow). C, Double membrane invaginations (white arrowheads) near AZ. The invaginations and their openings to the cleft were seen in six contiguous EM sections. D, Labeled fully internalized endosome (left) and unlabeled fully internalized endosome (arrowhead). Note coated LVs that may have budded from the endosomes. E, Unlabeled endosome containing clathrin-coated buds (arrows). Scale bars:A, 250 nm; B–E, 500 nm.

Fig. 4.

Two budding endosomes in HRP preparation. Stereo view is from stack of four EM sections. SC, Schwann cell. Scale bars, 500 nm.

Fig. 5.

Endocytosis occurs near AZs in snake motor boutons. A, Rendering of bouton portion from 31 serial EM sections. Stereo pair shows postjunctional folds of muscle fiber atleft (blue). AZs (red) lie on presynaptic membrane. LVs (white), shown as 50 nm spheres, were found predominantly near the presynaptic membrane. Also shown are four endosomes and two deep membrane invaginations (gray). B, C, Projections of fold centers (FCs; blue squares) AZs (redsquares), and LVs (white spheres) onto a plane corresponding to that of the presynaptic membrane (see Results).B, AZs appeared near (or in direct apposition to) FCs. Some fold regions were not occupied by AZs. C, LVs were found clustered near AZs (and therefore near folds as well). D, E, Locations of coated pits (yellow spheres) and deep membrane invaginations (gray; arrows) relative to AZs in two rendered boutons. Bouton inD is the same as that in A–C. Scale bars: A, 1 μm; B–E, 1 μm.

Fig. 6.

Clusters of LVs appear as punctate hot spots of FM1-43 staining at light level. A, Confocal photomicrograph of FM1-43 uptake by snake motor terminal in response to 150 stimuli delivered at 5 Hz. B, Magnified view ofbracketed region in A showing large, bright, irregular spots plus small circular spots in background.C, Region of rendered projection from EM as in Figure5C (same stimulation as the terminal inA). LVs with actual diameter of ∼50 nm are depicted as 200 nm partially transparent shaded disks to emulate diffraction-limited detail in light image (see Results).D, Contrast in C was adjusted manually to match the image in B best. Note similarity of large bright spots (overlapping vesicles) and small background spots (individual vesicles) to structures in B. Scale bars:A, 2 μm; B–D, 500 nm.