Limits...
FRET-FLIM investigation of PSD95-NMDA receptor interaction in dendritic spines; control by calpain, CaMKII and Src family kinase.

Doré K, Labrecque S, Tardif C, De Koninck P - PLoS ONE (2014)

Bottom Line: We used a FRET-FLIM approach in developing cultured rat hippocampal neurons expressing fluorescently tagged NMDA receptor (NMDAR) and PSD95, two essential proteins in synaptic plasticity, to examine the regulation of their interaction.The activity of both CaMKII and calpain were essential for this effect in both developmental stages.Finally, we found that calpain inhibition reduced spine growth that was caused by NMDAR activity, supporting the hypothesis that PSD95-NMDAR separation is implicated in synaptic remodeling.

View Article: PubMed Central - PubMed

Affiliation: Institut Universitaire en Santé Mentale de Québec, Université Laval, Québec, QC, Canada.

ABSTRACT
Little is known about the changes in protein interactions inside synapses during synaptic remodeling, as their live monitoring in spines has been limited. We used a FRET-FLIM approach in developing cultured rat hippocampal neurons expressing fluorescently tagged NMDA receptor (NMDAR) and PSD95, two essential proteins in synaptic plasticity, to examine the regulation of their interaction. NMDAR stimulation caused a transient decrease in FRET between the NMDAR and PSD95 in spines of young and mature neurons. The activity of both CaMKII and calpain were essential for this effect in both developmental stages. Meanwhile, inhibition of Src family kinase (SFK) had opposing impacts on this decrease in FRET in young versus mature neurons. Our data suggest concerted roles for CaMKII, SFK and calpain activity in regulating activity-dependent separation of PSD95 from GluN2A or GluN2B. Finally, we found that calpain inhibition reduced spine growth that was caused by NMDAR activity, supporting the hypothesis that PSD95-NMDAR separation is implicated in synaptic remodeling.

Show MeSH

Related in: MedlinePlus

Proposed mechanism of NMDAR-PSD95 dissociation during NMDAR activation in young neurons (A) versus mature neurons (B).(A) In young neurons expressing more GluN2B than GluN2A, the PDZ2 domain of PSD95 supports its binding to GluN2B c-terminus [15]. The phosphorylation by CaMKII of S73 in the PDZ1 domain of PSD95, following NMDAR activation and Ca2+ influx, would repel this portion of PSD95 from the receptor, possibly where another binding site for PSD95 on GluN2B was recently identified (aa 1149–1157; [31]. This may then reduce the protective effect of PSD95 against calpain-mediated cleavage of the ctail of GluN2B [6], leading to the separation of PSD95 from the receptor. This scenario reconciles our data showing that inhibiting CaMKII (KN93) prevents the separation of PSD95 from the receptor, but that S73D-PSD95 can still bind GluN2B-containing NMDARs [16]. (B) In mature neurons, PSD95 is mainly bound to GluN2A-containing receptors, via two binding sites on the receptor: i) the c-terminal site binding to PDZ2 and ii) another site (aa 1382–1389) binding to the SH3 domain of PSD95 [31]. The phosphorylation of S73-PSD95 alone (S73D mutation) was sufficient to prevent the PSD95-NMDAR interaction in mature synapses, but CaMKII was unable to trigger the complex separation in the presence of a calpain inhibitor, suggesting that a calpain-sensitive PSD component (spectrin is a well know calpain substrate [34], [38], [39] and is thus shown merely as an example) must be removed to allow CaMKII to access S73 on PSD95.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4230936&req=5

pone-0112170-g006: Proposed mechanism of NMDAR-PSD95 dissociation during NMDAR activation in young neurons (A) versus mature neurons (B).(A) In young neurons expressing more GluN2B than GluN2A, the PDZ2 domain of PSD95 supports its binding to GluN2B c-terminus [15]. The phosphorylation by CaMKII of S73 in the PDZ1 domain of PSD95, following NMDAR activation and Ca2+ influx, would repel this portion of PSD95 from the receptor, possibly where another binding site for PSD95 on GluN2B was recently identified (aa 1149–1157; [31]. This may then reduce the protective effect of PSD95 against calpain-mediated cleavage of the ctail of GluN2B [6], leading to the separation of PSD95 from the receptor. This scenario reconciles our data showing that inhibiting CaMKII (KN93) prevents the separation of PSD95 from the receptor, but that S73D-PSD95 can still bind GluN2B-containing NMDARs [16]. (B) In mature neurons, PSD95 is mainly bound to GluN2A-containing receptors, via two binding sites on the receptor: i) the c-terminal site binding to PDZ2 and ii) another site (aa 1382–1389) binding to the SH3 domain of PSD95 [31]. The phosphorylation of S73-PSD95 alone (S73D mutation) was sufficient to prevent the PSD95-NMDAR interaction in mature synapses, but CaMKII was unable to trigger the complex separation in the presence of a calpain inhibitor, suggesting that a calpain-sensitive PSD component (spectrin is a well know calpain substrate [34], [38], [39] and is thus shown merely as an example) must be removed to allow CaMKII to access S73 on PSD95.

Mentions: Deciphering the relationship between the three classes of enzymes in regulating the activity-dependent separation of the PSD95-NMDAR complex may require further investigation. Our results suggest that calpain activity is necessary at both immature (DIV7) and mature (DIV21) synapses. CaMKII activity is also required at both stages. However, if the implication of CaMKII involves PSD95 phosphorylation, our results with PSD95-S73D would argue that it is not sufficient in immature synapses, but maybe sufficient in mature synapses. Meanwhile PSD95-S73A did not dissociate upon NMDAR stimulation at either developmental stage. Finally, while calpain activity appears to be required for PSD95-NMDAR dissociation, it is not sufficient, since CaMKII inhibition with KN93 prevented the process. How can these results be interpreted by a concerted CaMKII and calpain-dependent process? Clues may come from the facts that i) there is a developmental change in GluN2A/2B ratio in these neurons; ii) PSD95 was shown to protect the c-terminal domains of GluN2 against calpain in HEK cells [6], [17]; iii) Previous results suggest that GluN2B is more prone to calpain cleavage compared to GluN2A [14]; iv) calpain may cleave additional substrates at synapses, other than GluN2 (e.g. spectrin). In young synapses, dominated by GluN2B, phosphorylation of PSD95 by CaMKII should not prevent their interaction [16]. However, the phosphorylation may modify the conformation of PSD95 in a way that it would no longer protect GluN2B from calpain cleavage. Meanwhile at mature synapses, dominated by GluN2A, phosphorylation of PSD95 by CaMKII may be sufficient to prevent the complex formation [16]; the role of calpain might then be to remove proteins bound to the NMDAR complex, such as spectrin [34], [38], [39] to enable the large CaMKII enzyme to access PSD95 for phosphorylation. We present in Figure 6 a working model, based on this above interpretation, of the regulation of the NMDAR-PSD95 complex separation by calpain and CaMKII, following NMDAR activation.


FRET-FLIM investigation of PSD95-NMDA receptor interaction in dendritic spines; control by calpain, CaMKII and Src family kinase.

Doré K, Labrecque S, Tardif C, De Koninck P - PLoS ONE (2014)

Proposed mechanism of NMDAR-PSD95 dissociation during NMDAR activation in young neurons (A) versus mature neurons (B).(A) In young neurons expressing more GluN2B than GluN2A, the PDZ2 domain of PSD95 supports its binding to GluN2B c-terminus [15]. The phosphorylation by CaMKII of S73 in the PDZ1 domain of PSD95, following NMDAR activation and Ca2+ influx, would repel this portion of PSD95 from the receptor, possibly where another binding site for PSD95 on GluN2B was recently identified (aa 1149–1157; [31]. This may then reduce the protective effect of PSD95 against calpain-mediated cleavage of the ctail of GluN2B [6], leading to the separation of PSD95 from the receptor. This scenario reconciles our data showing that inhibiting CaMKII (KN93) prevents the separation of PSD95 from the receptor, but that S73D-PSD95 can still bind GluN2B-containing NMDARs [16]. (B) In mature neurons, PSD95 is mainly bound to GluN2A-containing receptors, via two binding sites on the receptor: i) the c-terminal site binding to PDZ2 and ii) another site (aa 1382–1389) binding to the SH3 domain of PSD95 [31]. The phosphorylation of S73-PSD95 alone (S73D mutation) was sufficient to prevent the PSD95-NMDAR interaction in mature synapses, but CaMKII was unable to trigger the complex separation in the presence of a calpain inhibitor, suggesting that a calpain-sensitive PSD component (spectrin is a well know calpain substrate [34], [38], [39] and is thus shown merely as an example) must be removed to allow CaMKII to access S73 on PSD95.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4230936&req=5

pone-0112170-g006: Proposed mechanism of NMDAR-PSD95 dissociation during NMDAR activation in young neurons (A) versus mature neurons (B).(A) In young neurons expressing more GluN2B than GluN2A, the PDZ2 domain of PSD95 supports its binding to GluN2B c-terminus [15]. The phosphorylation by CaMKII of S73 in the PDZ1 domain of PSD95, following NMDAR activation and Ca2+ influx, would repel this portion of PSD95 from the receptor, possibly where another binding site for PSD95 on GluN2B was recently identified (aa 1149–1157; [31]. This may then reduce the protective effect of PSD95 against calpain-mediated cleavage of the ctail of GluN2B [6], leading to the separation of PSD95 from the receptor. This scenario reconciles our data showing that inhibiting CaMKII (KN93) prevents the separation of PSD95 from the receptor, but that S73D-PSD95 can still bind GluN2B-containing NMDARs [16]. (B) In mature neurons, PSD95 is mainly bound to GluN2A-containing receptors, via two binding sites on the receptor: i) the c-terminal site binding to PDZ2 and ii) another site (aa 1382–1389) binding to the SH3 domain of PSD95 [31]. The phosphorylation of S73-PSD95 alone (S73D mutation) was sufficient to prevent the PSD95-NMDAR interaction in mature synapses, but CaMKII was unable to trigger the complex separation in the presence of a calpain inhibitor, suggesting that a calpain-sensitive PSD component (spectrin is a well know calpain substrate [34], [38], [39] and is thus shown merely as an example) must be removed to allow CaMKII to access S73 on PSD95.
Mentions: Deciphering the relationship between the three classes of enzymes in regulating the activity-dependent separation of the PSD95-NMDAR complex may require further investigation. Our results suggest that calpain activity is necessary at both immature (DIV7) and mature (DIV21) synapses. CaMKII activity is also required at both stages. However, if the implication of CaMKII involves PSD95 phosphorylation, our results with PSD95-S73D would argue that it is not sufficient in immature synapses, but maybe sufficient in mature synapses. Meanwhile PSD95-S73A did not dissociate upon NMDAR stimulation at either developmental stage. Finally, while calpain activity appears to be required for PSD95-NMDAR dissociation, it is not sufficient, since CaMKII inhibition with KN93 prevented the process. How can these results be interpreted by a concerted CaMKII and calpain-dependent process? Clues may come from the facts that i) there is a developmental change in GluN2A/2B ratio in these neurons; ii) PSD95 was shown to protect the c-terminal domains of GluN2 against calpain in HEK cells [6], [17]; iii) Previous results suggest that GluN2B is more prone to calpain cleavage compared to GluN2A [14]; iv) calpain may cleave additional substrates at synapses, other than GluN2 (e.g. spectrin). In young synapses, dominated by GluN2B, phosphorylation of PSD95 by CaMKII should not prevent their interaction [16]. However, the phosphorylation may modify the conformation of PSD95 in a way that it would no longer protect GluN2B from calpain cleavage. Meanwhile at mature synapses, dominated by GluN2A, phosphorylation of PSD95 by CaMKII may be sufficient to prevent the complex formation [16]; the role of calpain might then be to remove proteins bound to the NMDAR complex, such as spectrin [34], [38], [39] to enable the large CaMKII enzyme to access PSD95 for phosphorylation. We present in Figure 6 a working model, based on this above interpretation, of the regulation of the NMDAR-PSD95 complex separation by calpain and CaMKII, following NMDAR activation.

Bottom Line: We used a FRET-FLIM approach in developing cultured rat hippocampal neurons expressing fluorescently tagged NMDA receptor (NMDAR) and PSD95, two essential proteins in synaptic plasticity, to examine the regulation of their interaction.The activity of both CaMKII and calpain were essential for this effect in both developmental stages.Finally, we found that calpain inhibition reduced spine growth that was caused by NMDAR activity, supporting the hypothesis that PSD95-NMDAR separation is implicated in synaptic remodeling.

View Article: PubMed Central - PubMed

Affiliation: Institut Universitaire en Santé Mentale de Québec, Université Laval, Québec, QC, Canada.

ABSTRACT
Little is known about the changes in protein interactions inside synapses during synaptic remodeling, as their live monitoring in spines has been limited. We used a FRET-FLIM approach in developing cultured rat hippocampal neurons expressing fluorescently tagged NMDA receptor (NMDAR) and PSD95, two essential proteins in synaptic plasticity, to examine the regulation of their interaction. NMDAR stimulation caused a transient decrease in FRET between the NMDAR and PSD95 in spines of young and mature neurons. The activity of both CaMKII and calpain were essential for this effect in both developmental stages. Meanwhile, inhibition of Src family kinase (SFK) had opposing impacts on this decrease in FRET in young versus mature neurons. Our data suggest concerted roles for CaMKII, SFK and calpain activity in regulating activity-dependent separation of PSD95 from GluN2A or GluN2B. Finally, we found that calpain inhibition reduced spine growth that was caused by NMDAR activity, supporting the hypothesis that PSD95-NMDAR separation is implicated in synaptic remodeling.

Show MeSH
Related in: MedlinePlus