This Week in The Journal

Diacylglycerol Levels Influence Forgetting

Mary Arai, Itsuki Kurokawa, Hoshinosuke Arakane, Tomohiro Kitazono, and Takeshi Ishihara

(see pages 8039–8053)

Most memories are useful for a limited time. Phone numbers can often be forgotten after the number is dialed; a to-do item can be forgotten after it is completed; and many chemical reactions can be forgotten after you pass Organic Chemistry. In some cases, memories that seem to have been forgotten are in fact retained, but suppressed or not retrievable. Yet it is likely that many memories degrade or are erased over time. The molecular mechanisms underlying such forgetting are beginning to be revealed in studies in invertebrates, including Caenorhabditis elegans.

C. elegans worms are attracted to diacetyl, which tastes like butter, but they stop responding to the odor if they are exposed to it for a prolonged period in the absence of food. This adaptation is accompanied by a decrease in diacetyl-induced calcium responses in AWA sensory neurons. After 4 h without diacetyl exposure, however, neuronal and behavioral responses resume; this loss of adaptation is considered a form of forgetting. But worms with a null mutation in the multifunctional protein tir-1 fail to resume responding to diacetyl after a 4 h recovery period, and previous work has suggested that tir-1/JNK signaling within AWC sensory neurons is required to forget adaptation to diacetyl.

To identify additional molecular contributors to forgetting, Arai et al. performed an unbiased genetic screen for mutations that rescued postadaptation chemotaxis in tir-1 mutants. One such mutation disrupted function of diacylglycerol kinase-1 (dgk-1), a protein that converts diacylglycerol (DAG) to phosphatidic acid. This mutation is expected to increase DAG levels. Two other mutations expected to increase DAG levels—a gain-of-function mutation in the G_qα protein Egl-30 and a loss-of-function mutation in the G_oα protein goa-1—also rescued postadaptation chemotaxis in tir-1 mutants, as did a chemical analog of DAG. In contrast, expressing a gain-of-function mutant form of goa-1 in AWC neurons prolonged adaptation.

These results suggest that DAG levels in AWC neurons influence forgetting of diacetyl adaptation memories. Specifically, low levels prolong adaptation, and high levels promote forgetting. Notably, however, altering DAG levels in AWC neurons did not always affect AWA responses to diacetyl. Therefore, future work should clarify how AWC DAG levels influence chemotaxis toward diacetyl after adaptation.

• Download figure
• Open in new tab
• Download powerpoint

Both wild-type (top) and Ank3-deficient (bottom) motor axons (blue) form NMJs in which synaptic vesicle protein 2 (red) is apposed to muscle acetylcholine receptors (green). But the NMJs formed by Ank3-deficient axons are more fragmented and less compact than normal. See Teliska et al. for details.

Axon Initial Segment Promotes Axon Regeneration

Lindsay H. Teliska, Irene Dalla Costa, Ozlem Sert, Jeffery L. Twiss, and Matthew N. Rasband

(see pages 8054–8065)

The axon initial segment (AIS) is a specialized structure containing a high concentration of voltage-gated sodium and potassium channels enmeshed in a scaffold of actin and spectrin filaments that is anchored to microtubules by ankyrin G. The AIS not only facilitates action potential generation, but also acts as a barrier that restricts the transport of proteins, mRNAs, and organelles into the axon. Because the AIS is often affected by neuronal injury, Teliska et al. asked how loss of this structure affects the function and regeneration of motor and sensory axons in mice.

In the absence of ankyrin G, axons do not form an AIS. Surprisingly, however, knocking out Ank3 (which encodes ankyrin G) in motor axons did not affect the ability of mice to stay on a rotating rod, suggesting that the AIS is not essential for action potential generation in motor axons. Nevertheless, loss of Ank3 disrupted development and/or maintenance of neuromuscular junctions (NMJs): these were more fragmented and less compact when axons lacked Ank3. In addition, levels of mRNAs encoding three growth-related and/or plasticity-related proteins were lower in Ank3-deficient axons than in controls. And knocking out Ank3 slowed reinnervation of muscles after peripheral nerve crush: muscles were mostly reinnervated within 21 d of injury in control mice, but similar reinnervation required 35 d in Ank3-deficient mice. Muscle reinnervation was slowed to a lesser extent when ankyrin G levels were reduced by ∼50% by knocking out β4 spectrin in motor neurons. In contrast, reinnervation of the skin was not disrupted when Ank3 was knocked out in sensory axons.

These results suggest that the AIS is unnecessary for spike generation in motor axons, but it promotes trafficking of some axonal mRNAs. Deficient trafficking likely contributes to the impaired development of NMJs and the slower regeneration of motor axons after injury. In contrast, the AIS does not appear to affect regeneration of sensory neurons. The authors attribute this difference to the fact that motor neurons are multipolar, whereas sensory neurons are pseudo-unipolar. Future work should investigate this possibly by examining the distribution of axonal and dendritic proteins and mRNAs in sensory axons, as well as motor axons, before and after nerve crush.