Memory  >   Regulation of CABT-GluN2B Binding by Actin Depolymerization

Reduction of NMDA receptor (NMDAR) activity by actin depolymerization was first reported a quarter-century ago (Rosenmund and Westbrook, 1993a). The process was induced by Ca2+, but its underlying mechanism remains unclear. More recently, Mandal and Yan (2009) showed that actin depolymerization could reduce NMDAR currents by inducing NMDAR internalization. However, the reduction of NMDAR activity observed by Rosenmund and Westbrook did not seem to arise from the loss of NMDARs, but from the inhibition of channel opening (Rosenmund and Westbrook, 1993b). Similar results were obtained by using the actin-severing protein, gelsolin, suggesting that F-actin may promote prolonged opening of NMDARs (Furukawa et al., 1997).

Could the actin-mediated reduction of NMDAR activity result from NMDAR extinction? This notion is supported by the reports that actin depolymerization is implicated in memory extinction (Fischer et al., 2004; Wang et al., 2013; Trent et al., 2017). Therefore, the CABT-GluN2B binding that causes NMDAR extinction is likely regulated by actin depolymerization.

Dynamic vs. Stable F-Actin

The filamentous actin (F-actin) is made of globular actin (G-actin) monomers. It generally undergoes turnover by continuous "treadmilling" which involves the polymerization of G-actin at one end (called the barbed end) of the filament and depolymerization of F-actin at the opposite end (pointed end). Based on the turnover rate, two distinct pools of F-actin have been observed: dynamic and stable. The dynamic pool has rapid turnover, with a time constant around 40 seconds. The turnover rate for the stable pool is much slower, with a time constant of ~17 min (Honkura et al., 2008). Importantly, they are present at different locations: dynamic pool near the tip whereas stable pool close to the base of the spine head (Figure 20-1).

The CABT complex consists of a CRMP2 monomer and a tubulin heterodimer. CRMP2 has been shown to interact with both tubulin and F-actin in vitro. Colocalization of the CABT complex (consisting of CRMP2 and tubulin) with F-actin has also been observed in neurites (Tan et al., 2015; Yang et al., 2015). It is assumed that CABT complexes may bind to both dynamic and stable F-actin. However, they play distinct roles. The CABT bound to the dynamic F-actin could be involved in NMDAR extinction (discussed below) whereas the CABT bound to the stable F-actin may orchestrate neurite growth during memory consolidation (see Chapter 27).


Figure 20-1. A model for the regulation of CABT-GluN2B binding by actin depolymerization. A CABT complex may bind either F-actin or GluN2B-containing NMDARs.
(A) In the resting state, the CABT complex binds to F-actin.
(B) Upon prolonged activation, Ca2+ influx through NMDARs may activate cofilin via calcineurin. Depolymerization of the dynamic F-actin by cofilin causes CABT to bind and block GluN2B-containing NMDARs, resulting in NMDAR desensitization (extinction).

The Crucial Role of Cofilin

Cofilin is a major actin depolymerizing factor that may bind to F-actin and induce severing (Elam et al., 2013). The severed F-actin fragment then disintegrates into G-actin monomers, thereby releasing CABT which could move toward the dendritic membrane and bind with GluN2B, provided that its S1166 is dephosphorylated (Chapter 13). In the spine, cofilin may be present in the cytosol or associated with the membrane via PIP2 (Kanellos and Frame, 2016). Its activity is directly controlled by the phosphorylation status at Ser-3. Dephosphorylation of Ser-3 leads to cofilin activation. As mentioned above, the reduction of NMDAR activity observed by Rosenmund and Westbrook was induced by Ca2+, which has been shown to activate cofilin via calcineurin (Wang et al., 2005). Therefore, the Ca2+ influx through GluN2B-containing NMDAR may stimulate calcineurin (Chapter 18) to activate cytosolic cofilin, promoting CABT-GluN2B binding, thereby resulting in NMDAR extinction.

The effects of calcineurin activation alone may cause only mild F-actin reorganization, leading to NMDAR extinction. Additional activation of phospholipase C (PLC) may induce PIP2 hydrolysis to release membrane-associated cofilin into the cytosol. The large amount of cofilin in the cytosol would result in dramatic F-actin reorganization, leading to NMDAR internalization as observed by Mandal and Yan (2009). PLC activation is required for long-term depression (Horne and Dell'Acqua, 2007), which is accompanied by spine shrinkage that may arise from extensive cofilin-mediated F-actin reorganization.

Experimental Evidence

The above mechanism is consistent with the following reports:


Author: Frank Lee
First Published: April, 2018
Last updated: July, 2018