This Week in The Journal

Clustering Kv2.1 at the Axon Initial Segment

Camilla Stampe Jensen, Shoji Watanabe, Jeroen Ingrid Stas, Jessica Klaphaak, Ayaka Yamane, et al.

Integration and transmission of information in neural circuits relies on the precise localization of ion channels, neurotransmitter receptors, and related proteins in specific neuronal compartments, such as dendritic spines, shafts, and axon initial segments (AISs). Multiple mechanisms work together to ensure proper localization of transmembrane proteins. Some newly synthesized dendritic proteins, for example, are segregated into specific vesicles that are excluded from axons. In contrast, some axonal proteins appear in dendrites, but are rapidly endocytosed and trafficked to the axon, possibly by axon-specific kinesin motors. Whereas some transmembrane proteins are inserted at specific locations, others are inserted randomly and diffuse until they become anchored to molecular scaffolds. How a given protein is localized depends on amino acid motifs that determine which vesicle, motor, or scaffold proteins it interacts with (Bentley and Banker 2016 Nat Rev Neurosci 17:611).

Kv2.1 (green) is clustered in the proximal dendrites (labeled with MAP2 antibodies; blue) and the AIS (marked by ankyrin-G expression; red, arrowheads). See Jensen et al. for details.

Kv2.1 voltage-gated potassium channels are clustered in the AIS, soma, and proximal dendrites of neurons. Clustering in somatodendritic compartments depends on a proximal-restriction-and-clustering (PRC) motif in the C-terminus of the protein. Although point mutations in the PRC motif disrupt Kv2.1 clustering in dendrites, clustering at the AIS persists, indicating different mechanisms are used to cluster the channels in the two domains. Because mutant channels are inserted (albeit not clustered) in somatodendritic membranes, Jensen et al. asked whether diffusion and anchoring of Kv2.1 mediates clustering at the AIS. Experiments involving fluorescence recovery after photobleaching suggested this was not the case. Moreover, preventing vesicular trafficking from the Golgi—which eliminated insertion into somatodendritic domains—did not disrupt Kv2.1 clustering at the AIS. This suggests that Kv2.1 channels destined for the AIS travel via an unusual pathway that bypasses the Golgi. Mutation analyses revealed that entry into this Golgi-bypass route required a second sorting motif downstream of the PRC. This motif contains two putative phosphorylation sites, and mutating these sites caused Kv2.1 to disperse uniformly in axonal and dendritic membranes.

These results suggest that clustering of Kv2.1 channels at the AIS requires phosphorylation, and therefore might be regulated by neural activity. Such regulation could contribute to homeostatic plasticity: increasing AIS Kv2.1 levels could raise spike thresholds when neurons become hyperactive. Future studies should test this possibility.

Role of Basolateral Amygdala in Risky Choices

Caitlin A. Orsini, Caesar M. Hernandez, Sarthak Singhal, Kyle B. Kelly, Charles J. Frazier, et al.

(see pages 11537–11548)

To make good decisions, one must accurately assess the magnitude and probability of costs and benefits of each option. Several brain areas contribute to such assessments, including the basolateral amygdala (BLA). The BLA contains multiple populations of neurons that encode the valence and magnitude of aversive or rewarding events, as well as the cues that predict these events. Therefore, blocking BLA activity might be expected to impair decision making. Indeed, previous work showed that BLA lesions increased the likelihood that rats would choose a large reward despite the risk of foot shock (Orsini et al. 2015 J Neurosci 35:1368). New work by this group demonstrates that the effect of BLA inactivation depends on when the inactivation occurs.

Rats expressing halorhodopsin in BLA neurons were trained to press two levers, one that delivered a small reward (a safe option), and another that delivered a large reward sometimes accompanied by a foot shock (a risky option). After rats learned the task and consistently chose the risky lever on some proportion of trials, the authors used laser illumination to inhibit BLA neurons during specific task epochs. Inhibiting BLA neurons while rats decided which lever to press significantly decreased the number of risky choices. In contrast, when neurons were inhibited during receipt of a large reward accompanied by shock, rats became more likely to choose the risky option, even though they exhibited normal sensitivity to the shock. Inhibiting BLA neurons during delivery of the safe option or when a large reward was delivered without shock had no effect on choices.

These results confirm that the BLA contributes to multiple aspects of cost/benefit analysis. The authors suggest that activity during deliberation is required for rats to assess the relative magnitude (i.e., incentive salience) of two positive rewards; thus when BLA was inhibited, choices were driven by relative risk. In contrast, inhibition during receipt of shock likely prevented this shock from influencing subsequent choices. Future work should test these hypotheses and determine which subpopulations of BLA neurons contribute to each effect.