Cellular/Molecular
Compound Fusion at Ribbon Synapses
Gary Matthews and Peter Sterling
(see pages 5403–5411)
How ribbon synapses in photoreceptors and retinal bipolar cells can release vesicles extremely rapidly has long been a mystery. Diffusion of undocked vesicles to the plasma membrane is too slow to account for the speed of release. Matthews and Sterling now provide electron-microscopic evidence that compound fusion—in which vesicles fuse to each other before fusing with the plasma membrane—occurs at ribbon synapses of goldfish bipolar cells. Stimulated ribbon synapses had significantly fewer vesicles tethered to the ribbon than those in which exocytosis was blocked. Although most vesicles had similar sizes in stimulated and unstimulated synapses, vesicles up to three times larger appeared after stimulation. The largest of these occurred at the base of the ribbon, adjacent to the plasma membrane, whereas normal-sized vesicles predominated at the top of the ribbon. Invaginations of the plasma membrane near ribbon synapses also appeared after stimulation, and like vesicles, these were tethered to the ribbon.
Stereo-pair electron micrographs of a synaptic ribbon fixed during repetitive stimulation. Large membranous structures (black arrow) are connected to the ribbon by electron-dense filaments (white arrowhead). See the article by Matthews and Sterling for details.
Development/Plasticity/Repair
Depolarizing GABA Currents, NMDA Receptors, and Synaptogenesis
Doris D. Wang and Arnold R. Kriegstein
(see pages 5547–5558)
In the developing neocortex, GABAergic synapses form before glutamatergic synapses, but GABA is depolarizing due to elevated internal chloride. GABA-mediated depolarization regulates subsequent development of both GABAergic and glutamatergic synapses. To elucidate how GABAergic depolarization affects synaptic development, Wang and Kriegstein made GABA hyperpolarizing in embryonic mice by blocking the chloride transporter NKCC1 in utero. As shown previously, this disrupted postnatal AMPA transmission and reduced dendrite length and branching. At 1–3 postnatal weeks, the frequency of spontaneous GABAergic and AMPAergic postsynaptic currents was reduced in Nkcc1-knockdown mice compared to controls. Furthermore, dendritic spine density decreased and spine length increased in knockdown mice, suggesting fewer glutamatergic synapses were present. Synapse development in mutant mice was rescued by cotransfection of a voltage-independent NMDA receptor (NMDAR), and knockdown of an NMDAR subunit replicated the effects of NKCC1 knockdown. These data suggest that depolarizing GABA currents enable NMDAR activation, which, in turn, is necessary for AMPA synapse formation.
Behavioral/Systems/Cognitive
Opioid Attenuation of Human Fear Conditioning
Falk Eippert, Ulrike Bingel, Eszter Schoell, Juliana Yacubian, and Christian Büchel
(see pages 5465–5472)
When pain is paired with an innocuous stimulus, animals learn to fear the innocuous stimulus. But pain also triggers release of endogenous opioids, which dampen pain—does this affect fear conditioning? It has been hypothesized that opioids provide negative feedback that attenuates painful stimuli and thereby tempers fear conditioning. If so, opioid antagonists should enhance conditioning. This week, Eippert et al. provide evidence supporting this hypothesis in humans. The authors monitored fear conditioning using a reaction time task, functional magnetic resonance imaging, skin conductance, and subjective pain ratings in subjects treated with an opioid antagonist or saline. Antagonist-treated subjects demonstrated stronger conditioning (faster reaction times), while controls showed more habituation in brain regions involved in pain processing. Furthermore, brain regions involved in descending pain control (rostral anterior cingulate cortex, amygdala, and periaqueductal gray) were deactivated when the conditioned stimulus was presented to controls, suggesting inhibition of these regions may be responsible for attenuating conditioning.
Neurobiology of Disease
Interconnections of Medium Spiny Neurons
Stefano Taverna, Ema Ilijic, and D. James Surmeier
(see pages 5504–5512)
The two main types of medium spiny neurons (MSNs) in the striatum—those that project directly to the substantia nigra and those that project indirectly via the globus pallidus—have opposing effects on thalamic outputs and behavior, but are known to be interconnected. This week, Taverna et al. present a systematic analysis of these GABAergic interconnections, revealing differences between MSN types. Recording from pairs of neurons, they found that when neurons were connected, it was always unidirectionally: no pairs were reciprocally connected. In addition, MSNs that expressed D1 dopamine receptors (D1 MSNs, mostly striatonigral) synapsed primarily onto other D1 MSNs, whereas those that expressed D2 receptors (D2 MSNs, mainly striatopallidal) synapsed onto both types. Furthermore, when the presynaptic cell was a D2 MSN, PSPs were larger than when the presynaptic cell was a D1 MSN, regardless of the postsynaptic cell type. Depletion of dopamine (a model of Parkinson's disease) greatly reduced collateral connections of all types.
