Strong and Weak Synapses

The diversity of synaptic properties is well illustrated by phasic and tonic synapses at crustacean neuromuscular junctions.  Tonic boutons release almost no transmitter to single action potentials, but facilitate tremendously to repetitive stimulation; phasic boutons release many quanta to single action potentials, and depress to repetitive stimulation.  The number and structure of active zones in the two motor neuron terminals are similar, as is Ca²⁺ influx per active zone.  We have used photolysis of caged Ca²⁺ combined with [Ca²⁺]_i measurement and focal recording of transmitter release from single boutons to show that phasic synapses are much more sensitive than tonic synapses to a rapid rise in presynaptic [Ca²⁺]_i.  Theoretical analysis and computer simulations of molecular reaction schemes shows that the differences in response of the synapses to steps of [Ca²⁺]_i, single action potentials, and trains of action potentials can all be explained by assuming that at phasic synapses all the docked vesicles are fully primed and immediately releasable, while at tonic synapses almost all docked vesicles are unprimed, but become primed by the residual Ca²⁺ accumulating during a train of spikes (Millar et al., 2005).

We have recently produced a comprehensive model of the presynaptic processes controlling transmitter release at phasic and tonic synapses.  The model has calcium acting on distinct molecular targets at different locations to mobilize reserve vesicles to docking sites, to prime docked vesicles to enable fast release, and to trigger exocytosis of docked and primed vesicles.  Release sites remain refractory for a short time after releasing a docked vesicle, while docking, priming, and fusion machinery is reassembled.  The model has provisions for calcium entering presynaptic terminals through voltage-dependent calcium channels during action potentials and diffusing to target sites with binding to mobile and immobile buffers.  Model simulations accounted for the kinetics of presynaptic presynaptic [Ca²⁺]_i measurements, synchronous release (EJPs), and asynchronous release (increased mEJP frequency) to trains of action potentials, the dynamics of both facilitation and depression, effects on spike-evoked release of exogenous and photolyzable calcium buffers, release to and recovery from prolonged action potentials, and responses to step rises in presynaptic [Ca²⁺]_i (Pan and Zucker, 2009).

Regulation of Transmitter Release by Epac and HCN Channels

Serotonin activates a presynaptic adenylyl cyclase to produce cAMP, which in turn activates hyperpolarization and cyclic nucleotide (HCN)-activated ion channels and exchange protein activated by cAMP (Epac) in motor nerve terminals, leading to a short-lasting enhancement of transmitter release to action potentials. The enhancement is not accompanied by increasing resting [Ca²⁺]_i or Ca²⁺ influx in action potentials, but requires an intact actin cytoskeleton (Beaumont and Zucker, 2000).  High levels of presynaptic activity cause a long-term facilitation (LTF) in glutamate release to action potentials, due to HCN channel activation by the electrogenic Na⁺/K⁺ exchanger resulting from excessive Na⁺ loading during tetanic stimulation; the synaptic enhancement involves Ca²⁺ influx during the tetanus and local presynaptic protein synthesis, turnover of the actin cytoskeleton, and a variety of kinases and phosphatases that probably regulate protein synthesis (Beaumont et al., 2001; Beaumont et al., 2002; Zhong et al., 2004; Zhong and Zucker, 2004; Zhong et al., 2005).

We have extended these findings to Drosophila neuromuscular junctions (Cheung et al., 2005) permitting the use of genetic techniques in further analysis.  Current experiments are focused on identifying the effector protein(s) activated by Epac, using directed mutations and RNAi knockdown of candidate targets.  We are also using a genetically targeted fluorescence resonance energy transfer (FRET)-based sensor of cAMP activity (cAMPs) to characterize the spatio-temporal profile of presynaptic cAMP changes during serotonin action.   Future goals include using similar techniques to follow changes in cytoskeleton structure during serotonergic enhancement and LTF, and to track the spatio-temporal profile of presynaptic protein synthesis by injecting mRNA of fluorescently tagged proteins.

PTP at crayfish NMJs

Post-tetanic potentiation (PTP) is a form of homosynaptic plasticity following repetitive activity in which the efficacy of synaptic transmission grows over a period of several minutes of activity and decays slowly (t » 2 min) afterwards. We have previously shown that PTP is caused by a post-tetanic plateau of residual Ca²⁺ -- resulting from leakage of Ca²⁺ from tetanically loaded mitochondria and reverse mode operation of the plasma membrane Na⁺/Ca²⁺ exchange pump (Tang and Zucker, 1997; Zhong et al., 2001). We have also detected increases in intracellular Ca²⁺ concentration ([Ca²⁺]_i) transients in crayfish motor nerve terminals to brief AP trains following induction of PTP and LTF. We recently discovered (Minami et al., 2007) that this is due to increases in Ca²⁺ influx through Na⁺/Ca²⁺ exchangers acting in reverse mode during action potentials, and we demonstrated that this regulation of a Ca²⁺ pump by neural activity is unrelated to PTP and LTF.

SNARE interactions by FRET in secretion and recovery

Exocytosis of vesicles from the presynaptic terminal is mediated by a set of vesicle and plasma membrane associated proteins, collectively termed the SNARE complex.  interactions between proteins that comprise this molecular machine (synaptobrevin, SNAP-25 and syntaxin) are critical in the  fusion of vesicles, making them important factors in regulating secretion and recovery.  we are using fluorescence resonance energy transfer (FRET)-based recording methods to examine interactions between specific CFP and YFP N-terminal-tagged SNARE protein pairs before, during, and after vesicle exocytosis in hippocampal neurons.  We have developed a novel multichannel imaging system, and detected FRET signals corresponding to the assembly of SNAREs prior to secretion, conformational changes in the SNAREs accompanying exocytosis, and the disassembly of SNAREs following their dispersion to the margins of active zones and prior to endocytosis.

We have observed dispersal of the plasma membrane SNARE proteins SNAP-25 and syntaxin, as has been reported for vesiclular VAMP, beginning immediately after vesicles fuse on stimulation, and preceding the loss of FRET signalling SNARE disassembly.  This may represent dispersion of intact SNAREs untethered after vesicle fusion before they are disassembled at active zone edges.

We have also detected interactions between the C-terminal labeled proteins VAMP and syntaxin.   Vesicle fusion leads to a trans-cis transformation of the SNAREs bridging docked vesicles and the plasma membrane, and the apposition of the C-terminal labels on vesicle fusion induces new FRET between them.  A resting FRET is also present, due to orphan VAMP left behind from prior activity, still assembled with syntaxin in SNAREs.