Entorhinal Neuron Classes Targeted by Ventral Hippocampus
Andrei Rozov, Märt Rannap, Franziska Lorenz, Azat Nasretdinov, Andreas Draguhn, et al.
(see pages 8413–8425)
Much of the communication between the hippocampus and the rest of the brain is transmitted through the entorhinal cortex. Neurons in entorhinal cortical layer 2 (L2) and L3 provide input to the hippocampus, while L5 neurons receive hippocampal output. Recent work revealed that L5 comprises two molecularly and functionally distinct sublayers, 5a and 5b. Surprisingly, axons from the dorsal hippocampus appeared to target only L5b excitatory neurons, which do not project beyond the entorhinal cortex (Sürmeli et al., 2015, Neuron, 88:1040). Subsequent work using transsynaptic tracers suggested that L5b neurons provide input to L5a neurons, which in turn project to other parts of the telencephalon. Electrophysiological recordings by Rozov et al. indicate a different connectivity pattern, however.
Naturally occurring sharp-wave-ripple events arising in the CA1 region of the intermediate/ventral hippocampus evoked compound postsynaptic potentials (PSPs) in excitatory neurons and fast-spiking interneurons in both L5a and L5b. In all three cell types, PSPs persisted when extracellular divalent cation levels were increased to minimize polysynaptic transmission. Responses in fast-spiking interneurons were most tightly coupled with sharp-wave-ripple activity while responses in L5b excitatory neurons were least tightly coupled.
Connections between CA1 and L5 neurons were confirmed by photostimulation of channelrhodopsin-expressing pyramidal cells in either dorsal or intermediate/ventral hippocampus. Whereas photostimulation of ventral hippocampus evoked EPSPs of similar amplitude in L5a and L5b excitatory neurons, however, EPSPs evoked by photostimulation of dorsal hippocampus were ∼6× smaller in L5a than in L5b neurons. Furthermore, photostimulation of dorsal hippocampus did not evoke EPSPs in all recorded neurons in L5a.
Contrary to previous suggestions, few functional connections were detected between excitatory L5a and L5b neurons. Fast-spiking interneurons often received input from excitatory L5a neurons and, to a lesser extent, from L5b neurons, however. Finally, fast-spiking interneurons provided inhibitory input to L5a and L5b excitatory neurons with similar probability.
These results suggest that projections from CA1 to the medial entorhinal cortex target both L5a neurons, which transmit information to other brain areas, and L5b neurons, which transmit the information locally. Contrary to previous suggestions, interactions between L5a and L5b neurons may occur primarily via inhibitory interneurons. Thus, the entorhinal cortex appears to divide information from hippocampus into parallel streams.
Two simultaneously recorded neurons (green) in L5a of entorhinal cortex. Ctip2 immunolabeling (red) marks L5b. See Rozov et al. for details.
Rapid Pool Refilling after Multivesicular Release
Yujin Kim, Unghwi Lee, Chunghon Choi, and Sunghoe Chang
(see pages 8426–8437)
Presynaptic boutons contain numerous synaptic vesicles distributed in multiple pools. The small readily releasable pool (RRP) includes vesicles docked at the active zone; these can fuse rapidly with the synaptic membrane when an action potential occurs. After a vesicle is released, a new vesicle must be recruited to the RRP to take its place; otherwise, the RRP might be depleted, causing a pause in synaptic communication. Vesicles can be recruited to the RRP from either the recycling pool—vesicles formed when components of recently released vesicles are retrieved—or the reserve pool, which lies deeper within the synaptic bouton. Which of these is the main source for replenishing the RRP has been debated.
Kim et al. hypothesized that the RRP may be replenished from different pools, depending on how rapidly the RRP is being depleted. Testing this hypothesis took several steps. First, the authors expressed a fluorescent glutamate reporter, iGluSnFRpre, in the presynaptic terminals of cultured hippocampal neurons so they could detect release of single vesicles. Next, they confirmed that they could manipulate the number of vesicles released with each action potential—and thus the rate of RRP depletion—by varying extracellular calcium levels: release of a single vesicle occurred when calcium levels were low, and multivesicular release occurred when levels were high. The authors then monitored the rate of RRP refilling by monitoring the recovery of vesicle release after RRP depletion.
When calcium levels were high, RRP recovery had fast and slow components, suggesting that vesicles were recruited from two sources. The fast component was prevented by blocking either endocytosis or neurotransmitter transport into vesicles, suggesting that it depended on recruitment from the recycling pool. Finally, fitting the data with a quantitative model indicated that differences in recovery kinetics in normal and elevated calcium were explained by differences in the proportion of vesicles recruited to the RRP from the recycling pool.
These results suggest that the RRP can be refilled rapidly by reusing newly endocytosed vesicles, and that the extent to which this happens varies depending on calcium levels and/or the rate of vesicle release. Therefore, neurons may alter the ability of synapses to sustain multivesicular release by changing how vesicles are recruited to the RRP.
Footnotes
This Week in The Journal was written by Teresa Esch, Ph.D.
