Memory  >   The General Mechanism of Memory Extinction and Retrieval

The term "memory extinction" has been definded in Chapter 10. In short, it refers to the inhibition of memory retrieval. Mounting evidence suggests that protein kinase A (PKA) plays key roles in both extinction and retrieval:

  1. Direct infusions of PKA inhibitor into the amygdala induce more rapid extinction (Koh et al., 2002).
  2. Inhibition of PKA facilitated extinction of both recent and remote contextual fear memories. (Isiegas et al., 2006).
  3. Disruption of PKA anchoring to AKAP79/150 promotes extinction of contextual fear memories (Nijholt et al., 2008).
  4. The activators of PKA block fear extinction (Corcoran et al., 2015).
  5. The GluN2B-mediated PKA inhibition prevents retrieval of remote memory (Corcoran et al., 2013).
  6. Norepinephrine promotes retrieval via the stimulation of β1-adrenergic receptors, the production of cAMP, and the activation of PKA (Zhang et al., 2013).
  7. Blockade of the M1 type muscarinic acetylcholine receptors (M1-AChRs) impairs the retrieval of well-trained memory (Soma et al., 2014).

Chapter 12 proposes that memory extinction could be fundamentally caused by the binding between GluN2B and the CABT complex which consists of a CRMP2 monomer and a tubulin heterodimer. Chapter 13 shows how the GluN2B-CABT binding is regulated by PKA on the basis of structural data. This chapter will focus on physiological processes that lead to binding/unbinding between GluN2B and CABT.

From NMDAR Extinction to Memory Extinction

NMDAR Extinction refers to the closed state of the NMDA receptor (NMDAR) due to the blockade by the CABT complex, which consists of a CRMP2 monomer and a tubulin heterodimer. It is a form of "NMDAR desensitization", defined as the reduction of NMDA currents in the continuous presence of glutamate. There are several mechanisms that may cause NMDAR desensitization: Ca2+-dependent, glycine-dependent, glycine-independent, etc. (Thomas et al., 2006; Krupp et al., 2002; Tavalin and Colbran, 2017; Glasgow et al., 2017). To distinguish from other forms of desensitization, the NMDAR blocked by CABT will be called "NMDAR extinction", because it may give rise to the macroscopic memory extinction.

A memory is stored in a large population of memory units. Each memory unit may span an entire dendritic branch which contains many NMDARs. For memory engram cells, neuronal firing depends largely on NMDA spikes and plateaus. Thus, the macroscopic memory state can be considered as an ensemble of the microscopic states of these NMDARs, and memory extinction could result from a substantial fraction of these NMDARs in the extinction state. Since the CABT complex does not bind to GluN2A, the GluN2A-containing NMDAR does not have the extinction state. Therefore, memory extinction should depend on the GluN2B/GluN2A ratio in the dendritic branches that encode the memory. A high GluN2B/GluN2A ratio is prone to memory extinction which is not an "all-or-none" phenomenon. Rather, it has various degrees. The more NMDARs in the extinction state, the harder the memory will be retrieved.

The Extinction Mechanism

The CABT complex binds to the C-terminal domain (CTD) of the GluN2B subunit. Their binding depends on the phosphorylation state of GluN2B at S1166 which is a target of protein kinase A (PKA). Loss of this single phosphorylation site abolishes PKA-dependent potentiation of NMDAR Ca2+ permeation, synaptic currents, and Ca2+ rises in dendritic spines (Murphy et al., 2014). The structural basis for PKA to regulate CABT-GluN2B binding (hence, NMDAR extinction) is described in Chapter 13. It is concluded that S1166 phosphorylation by PKA may prevent CABT-GluN2B binding.

In the resting state, GluN2B-NMDARs are phosphorylated (presumably at S1166) by basally active PKA (Raman et al., 1996) which is anchored to GluN2B-NMDARs via AKAP79/150 (Chapter 18). During learning, the Ca2+ influx through NMDARs may activate calcineurin (CaN) to dissociate AKAP79/150 from PSD-95, accompanied by F-actin reorganization (Gomez et al., 2002). This structural change causes AKAP79/150, together with PKA, to move away from GluN2B, consequently attenuating PKA's capacity to phosphorylate S1166. Dephosphorylation at S1166 may result in CTD conformational switch from a folded to unfolded structure. Only the unfolded conformation may bind to CABT (Chapter 13). In the mean time, F-actin reorganization could bring CABT to bind with the unfolded CTD, thereby leading to NMDAR extinction (see Chapter 20).

Once activated, CaN may dissociate from AKAP79/150 and move toward the nucleus. CaN has been demonstrated to activate the transcription factor, nuclear factor of activated T-cells (NFAT), resulting in memory extinction (de la Fuente et al., 2011). Brain-derived neurotrophic factor (BDNF) is a key target of NFAT (Groth et al., 2007; Vashishta et al., 2009). It can also be upregulated by the ERK-CREB pathway (Wang et al., 2012). As described in Chapter 22, BDNF plays a crucial role in memory extinction by increasing the localization of GluN2B-NMDARs to the dendritic membrane.

The action of CaN causes only mild structural changes in the spine, leading to NMDAR extinction. AKAP79/150 remains bound to the dendritic membrane (via palmitoylation) and likely to re-associate with PSD-95 when the Ca2+ concentration (and thus CaN activity) decreases. Additional activation of CaMKII is required to remove AKAP79/150 out of the spine, resulting in long-term depression (LTD) and spine shrinkage (Woolfrey et al., 2018). Moreover, phospholipase C (PLC) is essential for the release of cofilin from PIP2. Cofilin is an actin depolymerization factor that plays a key role in actin remodeling and spine morphology.

The Retrieval Mechanism

Memory retrieval is a process that reactivates extinguished memory. If memory extinction is caused by CABT-GluN2B binding, then memory retrieval should result from CABT-GluN2B unbinding. As mentioned above, phosphorylation of GluN2B at S1166 by PKA may prevent CABT-GluN2B binding, thereby promoting memory retrieval. The activity of PKA depends on the binding of cyclic AMP (cAMP), whose level is regulated by adenylyl cyclase (AC) that catalyzes the conversion from ATP to cAMP. The activity of AC is under the regulation of various pathways. For instance, norepinephrine may bind to the β1 adrenergic receptors, triggering G-protein-coupled signaling to enhance AC activity (Zhang et al., 2013). In addition to this pathway, the AC subtype 1 (AC1) and 8 (AC8) can also be activated by calcium-calmodulin to enhance PKA activity (Wong et al., 1999; Wang et al., 2003). Hence, AC1 and AC8 provide a shortcut for Ca2+ to increase PKA activity. This pathway is very important as PKA is anchored to GluN2B via AKAP79/150 (Chapter 18), permitting the Ca2+ influx through GluN2B-containing NMDARs to efficiently activate PKA and enhance long-term potentiation (LTP). This also explains why AC1 and AC8 contribute to LTP (Chen et al., 2014; Yamanaka et al., 2017).


The following two figures summarize the proposed mechanisms for memory extinction and retrieval.


Figure 19-1. The proposed mechanism for NMDAR extinction and recovery.
(1) In the resting state, S1166 of GluN2B is phosphorylated by PKA, preventing CABT-GluN2B binding.
(2) The Ca2+ influx through GluN2B-containing NMDARs may activate calcineurin to dissociate AKAP79/150 from PSD-95, accompanied by F-actin reorganization. This should attenuate PKA's capacity to phosphorylate S1166.
(3) S1166 dephosphorylation allows CABT-GluN2B binding, resulting in NMDAR extinction.
(4) Neurotransmitters, such as norepinephrine and acetylcholine, may trigger signaling cascades to activate PKA, resulting in S1166 phosphorylation, consequently leading to the dissociation of CABT from NMDARs.


Figure 19-2. The signaling cascades induced by extinction training.
(1) Extinction training activates CaN and ERK signaling cascades.
(2) In the cytoplasm, CaN may activate NFAT, resulting in translocation to the nucleus where BDNF could be upregulated.
(3) ERK can also upregulate BDNF via the ERK-CREB pathway.
(4) BDNF promotes GluN2B-NMDAR trafficking to the dendritic membrane, thereby enhancing memory extinction.


Author: Frank Lee
First published: November, 2017
Last updated: June, 2018