NMDAR activation promotes rapid translocation of aCaMKII::GFP to synapses, causing AMPAR trapping at 1 min (only synapses with CaMKII translocation)
tCaMKII (active prion) promotes immobilization of endogenous GluR1 (containing) AMPARs (both synaptic and extrasynaptic), and to a much lesser extent GluA2 (containing) AMPARs.
CaMKII direct phosphorylation of AMPARs unnecessary for synaptic trapping
GluA1 - SAP97 interaction unnecessary for CaMKII-dependent synaptic trapping
Stargazin increased tCaMKII-mediated trapping of recombinant GluA1 (homomeric), but tCaMKII had no effect on mobility of recombinant GluA2 (homomeric)
Stargazin phosphorylation (by tCaMKII) is necessary for GluA1 trapping; blocking phosphorylation caused AMPAR mobility to significantly increase.
intriguing finding: GluA1 subunit-specific effect of CaMKII, where it immobilizes recombinant GluA1 but not GluA2 homomeric AMPARs.
findings consistent with specific role of GluA1 in activity-dependent trafficking - but Stargazin can bind all subunits??
findings raise possibility that during LTP, CaMKII activation triggers both classical LTP and PPD. Interesting that LTP is frequently accompanied by PPD (opposite of PPF: paired-pulse facilitation)
The Ca(2+)/calmodulin-dependent protein kinase II (CaMKII) is critically required for the synaptic recruitment of AMPA-type glutamate receptors (AMPARs) during both development and plasticity. However, the underlying mechanism is unknown. Using single-particle tracking of AMPARs, we show that CaMKII activation and postsynaptic translocation induce the synaptic trapping of AMPARs diffusing in the membrane. AMPAR immobilization requires both phosphorylation of the auxiliary subunit Stargazin and its binding to PDZ domain scaffolds. It does not depend on the PDZ binding domain of GluA1 AMPAR subunit nor its phosphorylation at Ser831. Finally, CaMKII-dependent AMPAR immobilization regulates short-term plasticity. Thus, NMDA-dependent Ca(2+) influx in the post-synapse triggers a CaMKII- and Stargazin-dependent decrease in AMPAR diffusional exchange at synapses that controls synaptic function.
The principal recruitment mechanisms anticipated is that CaMKII promotes the trapping at the postsynaptic density (PSD) of laterally diffusing AMPARs. Here's why: First, NMDAR activation causes the rapid translocation of CaMKII from dendritic compartments to activated synapses. Second, following NMDAR activation, CaMKII can remain at postsynaptic sites for prolonged periods of time, through binding to several PSD proteins, including the NMDAR. Third, CaMKII bound to the NMDAR remains active independent of Ca2+/CaM.
To this end, we either activated or inhibited CaMKII using a number of genetic, pharmacological, and physiological approaches while simultaneously tracking the mobility of surface AMPARs imaged via luminescent semiconductor quantum dots (QDs) precoupled to specific antibodies against AMPAR subunits (GluA1 or GluA2, corresponding to GluR1 and GluR2
Our results indicate that CaMKII activation stops the diffusion of surface AMPARs at synaptic sites.
Furthermore, we show that this novel function of CaMKII is mediated by phosphorylation of stargazin and binding of its C-terminus to PDZ domain scaffold proteins such as PSD95
Postsynaptic Translocation of CaMKII Promotes the Diffusional Trapping of AMPARs at Synapses
Using fluorescence microscopy on cultured hippocampal neurons and NMDAR stimulation with glutamate and glycine (Glu/Gly), we simultaneously monitored (1) the translocation of aCaMKII::GFP (2) and the surface mobility of AMPARs using quantum dots precoupled to a GluA1 antibody (QD-GluA1) and separately the GluA2 antibody (QD-GluA2)
As previously shown, NMDAR activation promoted the rapid translocation of aCaMKII::GFP to synaptic sites (marked by Homer1C::DsRed)
In most cases, AMPARs were completely immobilized during the 1 min posttranslocation recording period, an effect that was only observed at synapses where CaMKII translocated; they were not diffusionally trapped either at extrasynaptic sites or synapses without translocated CaMKII
To investigate whether AMPAR immobilization was a direct consequence of CaMKII postsynaptic translocation, we first overexpressed aCaMKII:: GFP carrying a mutation (CaMKII::GFP I205K) that suppresses its ability to translocate postsynaptically by disrupting its binding to the NMDAR
The Glu/Gly treatment promoted neither CaMKII::GFP I205K translocation nor synaptic immobilization of AMPARs
Next, we investigated whether the catalytic activity of CaMKII is required for the synaptic trapping of AMPARs. We overexpressed aCaMKII::GFP carrying a mutation known to disrupt its kinase activity but not its ability to translocate to synapses (CaMKII::GFP K42R)
Although the Glu/Gly treatment caused the synaptic translocation of CaMKII::GFP K42R (data not shown), AMPARs were not trapped at synapses enriched with this catalytically inactive mutant
Altogether, these findings suggest that CaMKII translocation promotes the diffusional trapping of AMPARs by the phosphorylation of specific targets in the PSD.
High-Frequency Stimulation Promotes AMPAR Immobilization through CaMKII Activation
We previously showed that high-frequency neuronal stimulation (HFS; 50 Hz) induces a rapid NMDAR and Ca2+-dependent AMPAR immobilization
We examined whether this process depended on CaMKII activation by stimulating a small population of neurons using a field bipolar electrode (Figure 3D) in the absence or presence of KN93.
HFS stimulation induced a strong immobilization of AMPARs in neurons treated with either the vehicle or the inactive analog KN92 (Figures 3E–3G). As a measure of CaMKII activation, we found that HFS also promoted a strong synaptic translocation of CaMKII. Thus, patterns of synaptic activity (spontaneous or HFS-mediated) necessary for the activation of NMDARs and CaMKII induced the immobilization of AMPARs.
Constitutively Active tCaMKII Promotes a Strong Immobilization of AMPARs
We thus tracked the mobility of endogenous GluA1- containing AMPARs, 16–24 hr after transfection of tCaMKII:: GFP.
We found that tCaMKII produced a strong reduction in the diffusion of both synaptic and extrasynaptic AMPARs
It is unclear why AMPARs are immobilized at extrasynaptic sites in presence of tCaMKII, perhaps extrasynaptic scaffolding proteins
Though to a lesser extent, tCaMKII also promoted the immobilization of AMPARs containing the GluA2 subunit
Thus, bypassing NMDAR activation by directly overexpressing a constitutively active form of CaMKII also leads to the diffusional trapping of AMPARs. This strongly supports the notion that CaMKII is a negative regulator of AMPAR lateral mobility.
CaMKII Phosphorylation of AMPARs Is Not Necessary for Immobilization
We found that tCaMKII equally promoted the immobilization of HA-GluA1 S831A, suggesting that CaMKII regulates AMPAR mobility by phosphorylating substrates other than GluA1.
Another important CaMKII substrate is SAP97, a scaffolding protein known to interact with GluA1 and to be recruited to synapses upon phosphorylation
to explore whether GluA1 binding to SAP97 was necessary for AMPAR immobilization, we coexpressed tCaMKII with HAGluA1 lacking the PDZ-binding domain (HA-GluA1D7). We found that HA-GluA1D7 mobility was still strongly reduced by tCaMKII
These results suggest that the interaction between GluA1 and SAP97 is not necessary for CaMKII-dependent immobilization of AMPARs.
Stargazin Mediates the Effects of CaMKII on AMPAR Mobility
Expression of WT Stargazin increased the CaMKII-mediated immobilization of recombinant HA-GluA1 to levels similar to those observed with endogenous AMPARs (Figure 5C) and increased the immobile fraction of AMPARs to significant levels.
we found that Stargazin itself had no effect on AMPAR diffusion. Also, we found that tCaMKII had no effect on the mobility of recombinant GluA2 either in the absence or presence of Stargazin suggesting a subunit-specific effect of CaMKII.
To investigate the possible implication of Stargazin phosphorylation, we coexpressed tCaMKII with Stargazin Ser9Ala (S9A), mutated at the nine putative CaMKII/PKC phosphorylation sites.
We tracked the mobility of endogenous AMPARs (QDGluA1) and found that tCaMKII was no longer able to immobilize them. In fact, AMPAR mobility was significantly increased.
In addition, Stargazin S9A had no significant effect on AMPAR mobility when expressed alone, though there was a tendency to increase receptor mobility
To determine whether Stargazin phosphorylation was sufficient to immobilize AMPARs, we overexpressed a Stargazin phosphomimetic mutant (STG S9D) alone and found that it promoted a strong immobilization of QD-GluA1
Since our results suggest that CaMKII is directly stabilizing Stargazin (and only indirectly AMPARs), a prediction is that tCaMKII should promote the diffusional trapping of Stargazin itself. To test this hypothesis, we coexpressed tCaMKII and Stargazin tagged extracellularly with HA (HA-Stargazin), and tracked the surface mobility of HA-Stargazin using QD-HA. We found that tCaMKII caused a robust immobilization of HA-Stargazin, but not of HA-Stargazin S9A, confirming the critical role of CaMKII phosphorylation in the diffusional trapping of Stargazin
Role of CaMKII in the Synaptic Recruitment of AMPARs
An intriguing finding of our study is that active CaMKII promotes the immobilization of both synaptic and extrasynaptic AMPARs.
An intriguing finding of our study is the apparent GluA1 subunit-specific effect of CaMKII. Although CaMKII triggers the immobilization of both GluA1 and GluA2 containing endogenous AMPARs, it immobilizes recombinant GluA1 but not GluA2 homomeric AMPARs.
Although this finding is consistent with the specific role of GluA1 in the activity-dependent recruitment of AMPARs, it is at odd with the fact that Stargazin can bind all subunits
It is unlikely that the CaMKII-mediated immobilization of AMPARs corresponds to a universal mechanism for LTP. For instance, LTP at the dentate gyrus is independent of CaMKII activity. Also, CaMKII activity is not necessary for LTP early in development at the CA1 region. Further studies will be necessary to determine whether other kinases known to be important for LTP induction, such as PKA, PI3-K, PKC, and MAPK, also trigger AMPAR immobilization.
Our findings thus raise the possibility that during LTP, CaMKII activation triggers both classical LTP and PPD. It is interesting to note that LTP is frequently accompanied by a decrease in paired-pulse facilitation (PPF)