Synapses are specialized intercellular junctions in which cell adhesion molecules connect the presynaptic machinery for neurotransmitter release to the postsynaptic machinery for receptor signalling. Neurotransmitter release requires the presynaptic co-assembly of Ca2+ channels with the secretory apparatus, but little is known about how synaptic components are organized. Alpha-neurexins, a family of >1,000 presynaptic cell-surface proteins encoded by three genes, link the pre- and postsynaptic compartments of synapses by binding extracellularly to postsynaptic cell adhesion molecules and intracellularly to presynaptic PDZ domain proteins. Using triple-knockout mice, we show that alpha-neurexins are not required for synapse formation, but are essential for Ca2+-triggered neurotransmitter release. Neurotransmitter release is impaired because synaptic Ca2+ channel function is markedly reduced, although the number of cell-surface Ca2+ channels appears normal. These data suggest that alpha-neurexins organize presynaptic terminals by functionally coupling Ca2+ channels to the presynaptic machinery.

␣-Neurexins are neuron-specific cell-surface molecules that are essential for the functional organization of presynaptic Ca 2؉ channels and release sites. We have now examined postsynaptic glutamate receptor function in ␣-neurexin knockout (KO) mice by using whole-cell recordings in cultured neocortical slices. Unexpectedly, we find that ␣-neurexins are required for normal activity of N-methyl-D-aspartate (NMDA)-but not ␣-amino-3-hydroxy-5-methyl-4-isoxyzolepropionic acid (AMPA)-type glutamate receptors. In ␣-neurexin-deficient mice, the ratio of NMDA-to AMPAreceptor currents, recorded as evoked synaptic responses, was diminished Ϸ50%. Furthermore, the NMDA-receptor-dependent component of spontaneous synaptic miniature responses was reduced Ϸ50%, whereas the AMPA-receptor-dependent component was unaffected. No alterations in the levels of NMDA-or AMPA-receptor proteins were detected. These results suggest that ␣-neurexins are required to maintain normal postsynaptic NMDAreceptor function. The decrease in NMDA-receptor activity in ␣-neurexin-deficient synapses could be due to a transsynaptic effect on the postsynaptic neuron (i.e., ␣-neurexins on the presynaptic inputs guide postsynaptic NMDA-receptor function) or to a cell-autonomous postsynaptic effect of ␣-neurexins on NMDAreceptor activity. To distinguish between these two possibilities, we cocultured WT GFP-labeled neurons with neocortical slices from ␣-neurexin-deficient or control mice. No difference was found between WT neurons innervated by inputs that contained or lacked ␣-neurexins, indicating that the absence of presynaptic ␣-neurexins alone does not depress postsynaptic NMDA-receptor function. Our data suggest that, in addition to the previously described presynaptic impairments, loss of ␣-neurexins induces postsynaptic changes by a cell-autonomous mechanism.T he physiological properties of excitatory synapses of a set of neurons, or even the same neuron, can differ remarkably (1-4). Multiple pre-and postsynaptic processes differentially regulate the strength of synapses. For example, in cortical synapses, the properties of presynaptic release sites formed by a single neuron can be differentially modulated by the postsynaptic target neuron (5). At least in part, this regulation appears to act on the amount of presynaptic Ca 2ϩ influx that is induced by an action potential. Conversely, postsynaptic glutamate receptors are modulated by synaptic activity (6, 7). Both of the two principal types of synaptic glutamate receptors, N-methyl-Daspartate (NMDA) and ␣-amino-3-hydroxy-5-methyl-4-isoxyzolepropionic acid (AMPA) receptors, undergo such usedependent changes (7-9). The two receptor types are coordinately up-or down-regulated when the total strength of the synaptic inputs into a neuron is held constant during synaptic scaling (9). In contrast, only AMPA receptors are selectively increased during NMDA-receptor-dependent long-term potentiation in the CA1 region of the hippocampus (7). Viewed together, synapses can thus be considered as dynamic units in wh...

Central glutamatergic synapses are thought to initially form as immature, so-called silent synapses showing exclusively N-methyl-d-aspartate receptor-mediated synaptic transmission. Postsynaptic insertion of alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors during further development leads to a conversion into functional, mature synapses. Here, we tested the hypothesis that, according to the "inside first-outside last" pattern of neocortical layer formation and synaptogenesis, pyramidal cells in the superficial layers might show a higher fraction of silent synapses compared with pyramidal cells in the deep layers. We performed an electrophysiological analysis of glutamatergic synapses in acute rat visual cortex slices during postnatal development. In layer VI pyramidal neurons the incidence of silent synapses was high during the first postnatal week and strongly declined during further development. Surprisingly, in superficial cortical plate pyramidal neurons (immature layers II/III), the fraction of silent synapses was initially very low and increased up to the second postnatal week. Thereafter, a similar decline as found in layer VI pyramidal neurons was observed. Thus the developmental regulation of silent synapses was clearly different in pyramidal neurons from different neocortical layers. The almost complete absence of silent synapses at early stages in layer II/III pyramidal neurons indicates that an initially formed subset of synapses is constitutively functional. This might be important to enable spontaneous activity and latter activity-dependent maturation of synapses.