GATSBY COMPUTATIONAL NEUROSCIENCE UNIT

Circadian Modulation of Memory Via Suppression of Learning-Induced Gene Expression

Recently, we found that the circadian clock modulates two forms of long-term memory in Aplysia: non-associative long-term sensitization (LTS) and associative learning that food is inedible (LFI). Circadian modulation occurs such that the greatest amount of long-term memory is formed in phase with the animal’s activity period. In both paradigms, the circadian clock modulates the formation of long-term memory, rather than recall. One objective of our current research is to determine the cellular and molecular sites through which the circadian clock modulates long-term memory.

Using the LTS paradigm, we investigated whether the circadian clock modulated core processes necessary for memory formation in vivo by analyzing circadian regulation of LTS-induced increases in MAPK activity (P-MAPK) and ApC/EBP protein expression. LTS training during the subjective day produced greater increases in P-MAPK than training during the night. Moreover, LTS training increased ApC/EBP protein levels only when animals were trained during the subjective day. These results suggest that circadian modulation occurs at least at the level of the presynaptic sensory neuron. Furthermore, circadian modulation of LTS appears to occur, at least in part, by suppressing changes in key proteins at night. Long-term memory formation at night was “rescued” through combined facilitation of MAPK signaling and transcription, in conjunction with LTS training.

We further tested the hypothesis that the circadian clock modulates long-term memory through suppression of learning-induced gene expression using the associative LFI paradigm. LFI training at night results in no long-term memory formation when animals are tested 24 hours after training. However, we found that LFI training at night results in a long-lasting molecular memory (24 to 48 hours) that can be converted to overt long-term memory through additional training. Moreover, the conversion of the partial memory to overt memory was specific to the training procedure used and did not result from a more general activity-induced mechanism to enhance memory. Thus, although the circadian clock appears to inhibit the formation of long-term memory at night, training results in a specific partial memory at night that persists at least 24 h. Further experiments determined that protein synthesis (translation), but not transcription, was required for the molecular memory formed during the initial training session at night. The circadian clock appears to suppress learning-induced transcription during the initial training at night, but upon subsequent training sessions, necessary transcription occurs so that overt behavioral memory can be observed. Facilitating transcription during LFI training at night “rescues” long-term memory formation such that robust memory can be observed after a single training session. Thus, for both non-associative and associative learning paradigms, the circadian clock inhibits learning-induced transcription and gene expression when animals are trained out of phase with their activity cycle. Supported by NINDS, Grant NS050589 (A.E.).

BACK