Journal of Neurocytology 29, 19–30, 2000 A comparison of synaptic protein localization in hippocampal mossy fiber terminals and neurosecretory endings of the neurohypophysis using the cryo-immunogold technique LIXIA ZHANG 1 , WALTER VOLKNANDT 1 , ECKART D. GUNDELFINGER 2 and HERBERT ZIMMERMANN 1 ∗ 1 Biozentrum der J.W. Goethe-Universit¨ at, AK Neurochemie, Marie-Curie-Str. 9, D-60439 Frankfurt am Main, Germany; 2 Leibniz-Institut f ¨ ur Neurobiologie, Brenneckestr. 6, D-39118 Magdeburg, Germany Received 28 September 1999; revised 21 January 2000; accepted 28 January 2000 Summary In central synapses synaptic vesicle docking and exocytosis occurs at morphologically specialized sites (active zones) and requires the interaction of specific proteins in the formation of a SNARE complex. In contrast, neurosecretory terminals lack active zones. Using the cryo-immunogold technique we analyzed the localization of synaptic vesicle proteins and of proteins of the docking complex at active zones. This was compared to the localization of the identical proteins in neurosecretory terminals. In addition we compared the vesicular and granular localization of the proteins investigated. Synaptic vesicles in rat hippocampal mossy fiber synapses and microvesicles in the neurosecretory terminals of the neurohypophysis contained in common the proteins VAMP II (a v-SNARE), SV2, rab3A, and N-type Ca 2+ channels. Only minor immunolabeling for these proteins was observed at neurosecretory granules. These results support the notion of a close functional identity of microvesicles from neurosecretory endings of the neurohypophysis and of synaptic vesicles. The vesicular pool of N-type Ca 2+ channels may serve their stimulation-induced translocation into the plasma membrane. We find increased labeling for VAMP II, SNAP-25, N-type Ca 2+ channels and of rab3A at the active zones of mossy fiber synapses. Labeling at release sites is by far highest for Bassoon, a high molecular weight protein of the active zone. The labeling pattern implies an association of Bassoon with presynaptic dense projections. Bassoon is absent from neurosecretory terminals and VAMP II, SNAP-25, rab3A, and N-type Ca 2+ channels reveal a scattered distribution over the plasma membrane. The competence of the presynaptic active zone for selective vesicle docking may not primarily result from its contents in SNARE proteins but rather from the preformation of presynaptic dense projections as structural guides for vesicle exocytosis. Introduction The release of synaptic messenger substances from neurons is mediated by regulated exocytosis from vesicular organelles. Low-molecular weight messen- gers are thought to be stored in and released from small and electron-lucent vesicles whereas secretion of co-released peptides involves granules with an electron-dense core (Zimmermann, 1993). In contrast to synapses releasing fast neurotransmitters such as acetylcholine or glutamate that contain mainly electron-lucent synaptic vesicles and only few peptide- storing granules, neurosecretory endings of the poste- rior hypophysis contain large amounts of granules with an electron-dense core and a diameter of 170–200 nm. In ∗ To whom correspondence should be addressed. addition these terminals contain a considerable num- ber of electron-lucent vesicles (approximately 50 nm in diameter) that are indistinguishable in morphologi- cal terms from synaptic vesicles and are referred to as microvesicles. Under normal physiological conditions the populations of granules and microvesicles within the neurosecretory varicosities accumulate into sepa- rate clusters. Electrical stimulation dissolves the clus- ters resulting in a scattered distribution of microvesicles in the neurosecretory terminals (Hayashi et al., 1994). Whereas neurosecretory granules are involved in ex- ocytotic peptide release the functional role of the mi- crovesicles is less clear. They have been implicated in 0300–4864 C  2000 Kluwer Academic Publishers