Table of Contents Author Guidelines Submit a Manuscript
Neural Plasticity
Volume 2017, Article ID 9601046, 13 pages
https://doi.org/10.1155/2017/9601046
Research Article

Activation State-Dependent Substrate Gating in Ca2+/Calmodulin-Dependent Protein Kinase II

Biochemistry and Molecular Biology, Stark Neuroscience Research Institute, Indiana University School of Medicine, Indianapolis, IN 46202, USA

Correspondence should be addressed to A. Hudmon; ude.eudrup@nomduha

Received 15 July 2017; Accepted 23 October 2017; Published 17 December 2017

Academic Editor: Yi Zhou

Copyright © 2017 D. E. Johnson and A. Hudmon. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Calcium/calmodulin-dependent protein kinase II (CaMKII) is highly concentrated in the brain where its activation by the Ca2+ sensor CaM, multivalent structure, and complex autoregulatory features make it an ideal translator of Ca2+ signals created by different patterns of neuronal activity. We provide direct evidence that graded levels of kinase activity and extent of T287 (T286 α isoform) autophosphorylation drive changes in catalytic output and substrate selectivity. The catalytic domains of CaMKII phosphorylate purified PSDs much more effectively when tethered together in the holoenzyme versus individual subunits. Using multisubstrate SPOT arrays, high-affinity substrates are preferentially phosphorylated with limited subunit activity per holoenzyme, whereas multiple subunits or maximal subunit activation is required for intermediate- and low-affinity, weak substrates, respectively. Using a monomeric form of CaMKII to control T287 autophosphorylation, we demonstrate that increased Ca2+/CaM-dependent activity for all substrates tested, with the extent of weak, low-affinity substrate phosphorylation governed by the extent of T287 autophosphorylation. Our data suggest T287 autophosphorylation regulates substrate gating, an intrinsic property of the catalytic domain, which is amplified within the multivalent architecture of the CaMKII holoenzyme.

1. Introduction

Long-term potentiation (LTP) is a long-lasting enhancement of excitatory postsynaptic currents that many believe to be a cellular correlate of learning and memory [1]. While LTP can be observed in multiple areas of the mammalian brain (as well as in lower organisms), the Schaffer collateral/CA3 synapses onto CA1 pyramidal neurons of the hippocampus are a classic example whereby brief patterns of high-frequency activity favor LTP to enhance neuronal connectivity [2]. The opposing change in synaptic plasticity, long-term depression (LTD), is produced by low firing frequencies [2]. As might be expected, activity-dependent forms of plasticity like LTP and LTD both require the second messenger calcium (Ca2+). While LTP has been historically ascribed to kinase activity and LTD to dephosphorylation and phosphatase activity [3], a common effector system capable of decoding Ca2+ spike frequency into different functional outputs would be strategically positioned to regulate both LTP and LTD. The Ca2+/calmodulin- (CaM-) dependent protein kinase II (CaMKII) is an ideal regulator of synaptic plasticity because it has complex autoregulatory features that make it ideally suited for regulating activity-dependent plasticity [47].

In mammals, CaMKII family is composed of four closely related isoforms (α, β, γ, and δ) where it regulates diverse substrates in cellular processes ranging from metabolism and cell cycle control to Ca2+ homoeostasis and excitability [8]. Thus, like other multifunctional serine/threonine protein kinases (e.g., PKA and PKC), CaMKII phosphorylates many different cellular substrates in different subcellular compartments to globally coordinate cellular function. CaM is a ubiquitous Ca2+ sensor that activates CaMKII by binding to a target domain that disinhibits the autoregulatory domain (ARD) to allow substrate binding to the catalytic surface (Figures 1(a), 1(b), and 1(c)).

Figure 1: Molecular structure and schematic models for CaMKII autoregulation and activation. (a–c) Structural representations of CaMKII’s catalytic domain (green), autoregulatory domain (ARD) (blue), and hub domain (gray). (a) Illustration of CaMKII holoenzyme with detailed view of the catalytic domain. Adapted X-ray crystal structure shows monomeric human CaMKIIδ11–309 (PDB ID: 2VN9; [11]) in the autoinhibited state with the ARD (blue) and modified to include ATP in the active site (PDB ID: 1ATP; [71]). (b) hCaMKIIδ11–335 activated (PDB ID: 2WEL; [11]) by Ca2+ (red spheres) and calmodulin (CaM) (magenta) and exposing T287 (white) for autophosphorylation. (c) Linear schematic of CaMKII holoenzyme (CaMKIIholo) and monomeric CaMKII (CaMKIIm, or CaMKII1–317). Expanded view shows the ARD (blue) with subdivisions R1 (T287 containing region) and R2/R3 (CaM-binding region) [72].

The unique multivalent structure of CaMKII permits complex forms of autoregulation [9, 10]. The CaMKII holoenzyme is comprised of 12–14 subunits [1113] that permits coincident CaM binding to support an intraholoenzyme-intersubunit autophosphorylation [14, 15] (T286 α isoform and T287 β, γ, and δ isoforms) reaction (for convenience, we use T287 throughout the paper). While CaMKII is known to undergo autophosphorylation on multiple sites [16], many of which have unknown functions, the T287 autophosphorylation site within the ARD leads to enhanced CaM binding affinity in the presence of Ca2+ [17] and Ca2+/CaM-independent activity (autonomous) in low Ca2+ [18, 19]. The unique structure and autoregulation make CaMKII an ideal sensor to encode Ca2+-spike frequency into T287 autophosphorylation; a process influenced by Ca2+-spike duration and amplitude [20], phosphatase activity [21], alternative spliced linker length, and CaM availability [22, 23].

While CaMKII is present throughout the body, it is highly enriched at neuronal connections. In fact, CaMKII is one of the most abundant proteins in the postsynaptic density (PSD) [2426], a cytoskeletal organelle located just underneath the spine plasma membrane that localizes the postsynaptic signaling machinery with presynaptic release sites. While CaMKII appears to have multiple functions in synaptic plasticity, phosphorylation of GluA1 receptors to regulate their trafficking and signaling appears to be critical for LTP [7]. CaMKII’s role in LTP and Ca2+ signaling in the postsynaptic compartment is also linked to its ability to undergo translocation to NMDA receptors and undergo autophosphorylation [5, 27, 28], processes that might serve to restrict the kinase to its targets within the PSD. However, like the second messenger Ca2+, CaMKII is also implicated in LTD [29, 30]. And while the role of CaMKII in postsynaptic LTD has not been as intensively studied as its role in LTP [4], CaMKII was recently shown to differentially phosphorylate two different phosphoacceptor sites on GluA1 in LTP versus LTD [31]. Because LTP is promoted by strong stimuli versus weak for LTD, one plausible explanation is that graded levels of CaMKII activity and/or autophosphorylation promote differential substrate phosphorylation to induce the opposing forms of synaptic plasticity. However, recent studies have revealed little or no impact of autophosphorylation on the catalytic output of CaMKII [32, 33]. Using purified PSDs and immobilized peptide arrays to create a well-defined multisubstrate system, we test the hypothesis that the catalytic output and substrate selectivity of CaMKII are regulated by its activation state. Our data support a model whereby T287 autophosphorylation regulates substrate gating, an intrinsic property of the catalytic domain, which is amplified within the multivalent architecture of the CaMKII holoenzyme.

2. Results

CaMKII translocation and phosphorylation of PSD proteins are believed to play an important role in postsynaptic modifications supporting LTP [47]. We determined whether PSD phosphorylation by CaMKII induced by saturating Ca2+/CaM is equivalent or different to monomeric CaMKII catalytic activity. Conceptually, one might expect that a similar number of activated T287 autophosphorylated subunits (T286 in alpha isoform) would produce similar levels of substrate phosphorylation within the PSD regardless of whether the catalytic subunits are within a holoenzyme or monomeric. While all four CaMKII isoforms exhibit a high degree of sequence identity in their catalytic and regulatory domains, we selected the human δCaMKII isoform for its enhanced stability in our kinase assays. While the association domain hub was deleted from monomeric CaMKII (1-317) (Figures 1(a), 1(b), and 1(c)), this recombinant kinase, like the native holoenzyme, retains its Ca2+/CaM dependence and undergoes T287 autophosphorylation as an intersubunit reaction mechanism which we accomplished in a prereaction (see Materials and Methods). Postsynaptic densities were acquired from adult rats (male and female Sprague-Dawley) using differential centrifugation and detergent extraction to obtain the PSD Triton-insoluble fraction (see Materials and Methods) [34, 35]. The PSD preparation displayed a characteristic size and ultrastructure (data not shown) and SDS-PAGE separation and a Coomassie blue staining revealed the complex protein profile characteristic of PSD preparations.

Our intent was to focus on characterizing PSD phosphorylation from exogenously added CaMKII; therefore, several steps were taken to accomplish this. First, rat brains were extracted and snap frozen in liquid nitrogen as rapidly as possible (~90 secs) after animal sacrifice to reduce CaMKII translocation to the PSDs in response to decapitation [36]. Second, to minimize autonomous activity of endogenous PSD CaMKII, isolated PSDs were dephosphorylated overnight using a PP1α catalytic fragment. The 6X-His-tagged phosphatase was removed the next day using immobilized metal affinity chromatography. Third, the allosteric small molecule CaMKII inhibitor, KN93, was included in the PSD phosphorylation reaction to inhibit endogenous CaMKII from undergoing T287 autophosphorylation with addition of Ca2+/CaM (as KN93 prevents Ca2+/CaM binding/activation but does not inhibit activated T287 autophosphorylated CaMKII). Fourth, phosphatase and protease inhibitor cocktails were used in the PSD phosphorylation reactions to disrupt dephosphorylation and proteolysis reactions. Finally, exogenous CaMKII was preautophosphorylated at T287 with thiol-ATP (ATPγS) to limit potential dephosphorylation and to minimize subsequent incorporation of 32P in exogenous CaMKII autophosphorylation sites.

Postsynaptic densities were phosphorylated (plus [γ-32P-] ATP) with T287 autophosphorylated holoenzyme (CaMKIIholo+P) or monomer (CaMKIIm+P) in the presence of Ca2+/CaM. Similar levels of catalytic activity were applied for both forms of the autophosphorylated CaMKII (see Materials and Methods). To control for endogenous Ca2+/CaM-dependent or Ca2+/CaM-independent kinase activity, PSDs were also phosphorylated in the presence of Ca2+/CaM or EGTA, respectively. As described in detail in Materials and Methods, PSD phosphorylation was quantified by phosphorimaging of the 32P incorporation following SDS-PAGE separation of the proteins (Figures 2(a) and 2(b)). The corresponding phosphorylation data (autoradiography of 32P incorporated bands) (Figures 2(c) and 2(d)) was scaled to the total PSD protein in a given sample based on the area under the curve (AUC) as described in Materials and Methods. Under these conditions, the addition of Ca2+/CaM alone produced less than a 2-fold increase in the total phosphorylation compared to other Ca2+/CaM-independent kinases (the EGTA condition; Figure 2(d)). The inclusion of exogenous CaMKIIm+P produced ~2-fold increase in the phosphorylation compared to Ca2+/CaM-activated PSDs alone, suggesting that PSD substrates were available to exogenously added kinase (Figure 2(d)). The CaMKIIholo+P produced a 6.4-fold increase in phosphorylation compared to Ca2+/CaM-activated PSDs (Figure 2(d)). More importantly, compared to the T287 autophosphorylated monomer, there was a 4.4-fold increase in total phosphorylation when corrected for endogenous Ca2+/CaM-stimulated phosphorylation (Figure 2(d)) even though the holoenzyme might be expected to have less accessibility to substrates in the PSD compared to the monomer because of their size difference (~750 kDa versus 35 kDa, resp.).

Figure 2: Maximally active CaMKII holoenzymes enhance the phosphorylation of PSD proteins over monomeric CaMKII. (a–d) Isolated rat PSDs were phosphorylated in low (EGTA) or high calcium (Ca2+/CaM) to assess endogenous, copurified kinases. PSDs were also phosphorylated with the addition of equal molar catalytic subunits of T287 autophosphorylated CaMKII multimer (CaMKIIholo+P = diamond) or monomer (CaMKIIm+P = circle). (a) Coomassie-stained gel illustrating the PSD profile and the purified proteins used in the reaction. Square indicates CaM. (b) Phosphorimaging of the SDS-PAGE showing 32P-phosphoprotein bands. (c) Phosphorylation profile scaled to total Coomassie-stained protein. Molecular weight markers indicated along the scaled axis (pixels) representing the scaled lane of the SDS-PAGE. (d) Total phosphorylation for each condition assessed using the area under the curve (AUC) as arbitrary intensity units. (e) Quantitation of Ca2+-dependent (Ca2+/CaM) and Ca2+-independent (EGTA) enzymatic activity (katal/mol) of CaMKII measured via soluble peptide assays with radioactive (γ32P) ATP. CaMKII, monomer or multimer, was autophosphorylated at T287 prior to the PSD phosphorylation (and activity measurement) (see Materials and Methods) (). Error bars denote ± s.d.

The amount of CaMKII enzymatic activity (both Ca2+/CaM stimulated and autonomous) added to the PSD phosphorylation reaction was determined both before and after the experiment via a soluble kinase assay using the CaMKII-specific peptide AC-2 (see Materials and Methods). While the catalytic activity of the monomeric kinase was greater than the holoenzyme at the initiation of the PSD phosphorylation reaction, the level of activity at the end of the PSD phosphorylation was similar between monomeric and multimeric CaMKII (Figure 2(e)). While it is unclear why a loss of monomeric activity is observed at the end of the PSD reaction, no loss of enzymatic activity is observed for either CaMKIIholo+P or CaMKIIm+P in the absence of PSD protein (data not shown). Thus, while the enzymatic activity of monomeric CaMKII is favored or equal (start or end of the reaction) to that of the holoenzyme CaMKII, substrate phosphorylation is higher for CaMKII when multiple subunits are activated within the holoenzyme for a diffusion-restricted environment like the PSD.

Specificity for the switch-like behavior in CaMKII critical for LTP [3] may arise from this enzyme’s ability to translate Ca2+-spike frequency produced by brief patterns of high-frequency stimulation into graded levels of activity and T287 autophosphorylation [20]. Thus, we tested whether regulating the extent of activation within the CaMKII holoenzyme regulates substrate selectivity. However, endogenous effectors, scaffolding, and other regulatory proteins within the PSD could complicate our analysis; therefore, we created a multiple substrate SPOT array in order to explore how graded levels of CaMKII activity influence substrate utilization (see Materials and Methods). An obvious advantage of this immobilized peptide assay is that diverse substrates can be exposed to identical conditions without high-affinity substrates dominating mixed soluble substrate reactions [37]. We selected four well-studied CaMKII substrates studied with varying affinities as determined by standard solution kinetics—a high-affinity substrate (GluN2BS1303, Km = 4.6 ± 1.1 μM) and three lower-affinity, weak substrates (Syntide, Km = 43.5 ± 2.3 μM; VimentinS83 and GluA1S849, Km > 1 mM). All peptide sequences are denoted in Materials and Methods.

To create graded levels of CaMKII activity and T286 autophosphorylation in conditions of saturating Ca2+/CaM, we took advantage of the basal, inhibitory (Ca2+/CaM-independent) autophosphorylation (i.e., capping) described previously [38]. Specifically, naïve multimeric CaMKII is exposed to Mg2+/ATP for varying times in the absence of Ca2+/CaM. This is a relatively slow intramolecular autophosphorylation reaction (complete in 15–20 mins) within the CaM-binding domain (T306 in CaMKIIδ) that blocks subsequent Ca2+/CaM binding/activation [19, 3840]. The degree of inactivation is measured by exposing the Mg2+/ATP-treated CaMKII to saturating Ca2+/CaM and AC-2 in a soluble kinase assay (see Materials and Methods). We expressed maximal catalytic activity as a function of the number of subunits in the holoenzyme (the human CaMKIIδ holoenzyme is tetradecameric [11]). This allowed for the determination of the molar catalytic activity associated with one catalytic subunit. Using this metric, the average number of activatable subunits per holoenzyme was calculated for each inactivation timepoint shown in Figure 3(a) (see Materials and Methods). Thus, by normalizing our treated or “capped” holoenzymes to naïve untreated CaMKII (14/14 subunits), we calculated that residual bulk kinase activity in solution as represented on average by holoenzymes with 1/14 or 5/14 subunits is available for Ca2+/CaM activation.

Figure 3: Enhanced phosphorylation of intermediate and weak SPOT substrates with increasing number of active subunits per holoenzyme. (a) Time-dependent inactivation of CaMKII by basal autophosphorylation. Ca2+/CaM-stimulated activity of CaMKIIholo expressed in molar terms (katal/mol) and measured via standard soluble peptide (AC-2) assays with radioactive (γ32P) ATP. Enzymatic activity was measured following preinactivation reactions of various times (see Materials and Methods); mean ± s.d. (). By calculating the average enzymatic activity per subunit in a fully activatable tetradecameric holoenzyme (i.e., 0 min of inactivation), the average number of activated subunits per holoenzyme was determined for each condition. (b, c) Phosphorylation ([32P] phosphate incorporation) of immobilized CaMKII substrates (GluN2B, Syntide, GluA1, and Vimentin synthesized via SPOTs) by CaMKIIholo exhibiting 1, 5, or 14 activated subunits per holoenzyme. Substrates ranked from the lowest to the highest Km. (b) Raw SPOT phosphorylation (PSL/mm2) represented as mean ± s.d. (). (c) SPOT phosphorylation normalized by the number of subunits activated per holoenzyme and represented as mean ± s.d. (). One-way ANOVA of log-normalized data with Holm-Šídák posttest compared to GluN2B within each group (, ).

Using this capping pretreatment strategy to limit the number of Ca2+/CaM activatable subunits per holoenzyme, we observed that one active subunit per holoenzyme (CaMKIIholo1/14) displays preferential phosphorylation for the high-affinity substrate GluN2B (Figure 3(b)). In Figure 3(c), the phosphorylation data was normalized to account for the different levels of “starting” kinase activity produced by the capping reaction in order to better evaluate substrate selectivity profiles. Unlike GluN2B, increasing the number of activatable subunits per holoenzyme significantly increases relative phosphorylation of Syntide (intermediate substrate affinity) and GluA1 and Vimentin (weak substrates) (Figure 3(c)). While minimal changes in GluN2B phosphorylation were observed, a 20-fold increase in Syntide phosphorylation by submaximally activated CaMKIIholo5/14 was measured (Figure 3(c)). Maximally activated CaMKIIholo14/14 produced the greatest increase (12-fold) in phosphorylation of the weakest substrates, GluA1 and Vimentin (Figure 3(c)). These data indicate that intermediate- and low-affinity substrates require multiple activated subunits within the holoenzyme for phosphorylation. In contrast, high-affinity substrates are preferentially phosphorylated when the number of subunits per holoenzyme is limited. The reduced phosphorylation of high-affinity substrates seen in Figure 3(c) may result from substrates like NR2B possessing T-Site (i.e., targeting mode) characteristics [41], which allow substrates mimicking the ARD of CaMKII to bind in a mode that is biochemically distinct from the substrate binding pocket [42]. These stable, high-affinity interactions may punish substrate utilization via avidity effects as discussed further below.

We elected to use the SPOTs immobilized peptide approach as it has the obvious advantage that multiple defined substrates can be examined simultaneously without having the high-affinity substrates dominate the diffusion-limited solution assay. However, it is still important for these interfacial reactions to obey specific requirements for any enzyme analysis, including a nonsaturating reaction (Figure 4). We observed that less than 2% of the available peptide for each substrate (as compared to a 24-hour reaction for GluN2B, Syntide, GluA1, and Vimentin) was phosphorylated by CaMKIIholo+P in our standard reactions (Figure 4(a)). While we did not extend the phosphorylation reactions longer than 24 hrs, this time point is likely at equilibrium as the extent of phosphorylation follows an expected ranking based on substrate affinity (GluN2B > Syntide > GluA1/Vimentin). Furthermore, we observed minimal differences between maximally activated (14/14 subunits) holoenzymes at 60 nM versus 5 nM (Figures 3(b) and 4(a)—inset, resp.), indicating that while the extent of phosphorylation is sensitive to kinase concentration, changes in the substrate phosphorylation profile are not. Thus, for our standard SPOT reactions (4 mins), changes in the phosphorylation profile are not due to differential substrate availability.

Figure 4: Substrate availability and reaction linearity for SPOT phosphorylation reactions. (a–c) Substrate phosphorylation profiles ([32P] phosphate incorporation of SPOT peptides). Phosphorylation time course of CaMKIIholo+P (a) or CaMKIIm+P (b) with an expanded view of 4 min reaction (inset); mean ± s.d. (). (c) Phosphorylation by CaMKIIm+P comparing a normal 4 min SPOT reaction (c-1) to reactions with only radioactive secondary 4 min reactions with (c-2) or without (c-3) additional kinase in the secondary reaction; mean ± s.d. ().

One caveat of using SPOT arrays for enzyme analysis is that they can be negatively influenced by avidity effects created by multivalent contacts [43]. For example, the binding affinity of a protein with its immobilized peptide target increased by over 1000-fold when a dimeric protein is compared to a monomeric variant [44]. Consistent with avidity effects contaminating CaMKII phosphorylation of the high-affinity substrates in the SPOT assays, we observed that phosphorylation for the high- (GluN2B) and intermediate-affinity (Syntide) substrates was suppressed as the number of activated subunits within the holoenzyme increased from 5/14 to 14/14. This becomes readily apparent when CaMKII substrate phosphorylation is normalized to the number of activated subunits in the capping experiments (Figure 3(c)). This avidity effect was not seen with monomeric CaMKII, as substrate phosphorylation profile for CaMKIIm+P follows the expected profile (i.e., higher affinity, greater substrate phosphorylation (Figure 4(b))). Again, linearity for the monomer CaMKII reaction rates was supported by the observation that in our standard SPOT reactions, less than 2% of the available peptide for each substrate (as compared to a 24-hour reaction for GluN2B, Syntide, GluA1, and Vimentin) was phosphorylated by CaMKIIm+P (Figure 4(b)). Conclusive support for a linear kinase-substrate reaction is supported by a pulse-chase style reaction in which nearly identical levels of 32P incorporation are obtained when a secondary reaction occurs on the same SPOT membrane immediately after an initial nonradioactive reaction (Figure 4(c)—reaction 1 versus 2). A modification of this assay whereby CaMKIIm+P is applied to the SPOT membrane in the absence of ATP32 reveals that bound kinase continues to form product over time when the membrane is washed and ATP32 is added during the chase without additional kinase (Figure 3(c)). In total, our standard SPOT reactions are not limited by substrate availability nor are they outside of a linear reaction range. However, because the reduced substrate phosphorylation of higher affinity substrates with maximally activated CaMKIIholo+P is likely contaminated by multivalent avidity effects, we employed the monomeric kinase to directly evaluate the contribution of T287 autophosphorylation on substrate selectivity.

CaMKIIm autophosphorylation at T287 is a well-controlled, second-order reaction under saturating Ca2+/CaM conditions [15]. Like with CaMKIIholo+P, the percentage of autophosphorylated subunits can be detected using the high-affinity substrate AC-2 (Figure 5(a)) which exhibits a 1 : 1 correlation in the number of T287 autophosphorylated subunits and maximal autonomous activity ([Ca2+/CaM-independent]/[Ca2+/CaM-stimulated]) [45]. Therefore, we generated populations of CaMKIIm containing varying ratios of CaMKIIm−P and CaMKIIm+P using temperature and time to vary the extent of autophosphorylated subunits (see Materials and Methods) (Figures 5(a) and 5(b)). Importantly, for each condition (i.e., extent of autophosphorylated CaMKII), the number of CaMKIIm+P subunits generated in the prereaction does not change during our standard SPOT assays, as the level of autonomous activity measured in a solution assay using the AC-2 substrate is the same pre- and post-SPOTs (Figure 5(c)). We observed a direct relationship between the extent of substrate phosphorylation and the percentage of T287 autophosphorylated subunits in the SPOT assay (Figure 5(d)). However, intermediate- and low-affinity substrates (Figure 5(d)) displayed the greatest increases in their phosphorylation in response to T287 autophosphorylation. Interestingly, the phosphorylation profile of the non-T287 autophosphorylated CaMKIIm (Figure 5(b)) was highly similar to the minimally activated CaMKIIholo (one subunit/holoenzyme condition in Figure 3(b), suggesting that inactivation (via T306 capping) can allow holoenzymes with limited activatable subunits to behave as non-T287-autophosphorylated monomers. Overall, our data support the hypothesis that the extent of activation and T287 autophosphorylation within the CaMKII holoenzyme functions to broaden the substrate specificity of CaMKII.

Figure 5: CaMKII substrate utilization and selectivity are differentially altered by T287 autophosphorylation. (a) Schematic representation of various conditions used to preautophosphorylate CaMKIIm prior to Ca2+/CaM-stimulated substrate phosphorylation reactions on SPOT arrays (see Materials and Methods). Essentially, time and/or temperature was altered to generate different extents of T287 autophosphorylation (i.e., varying ratios of CaMKIIm−P/CaMKIIm+P) as graphically illustrated in (b). (c) As a readout of T287 autophosphorylation, the % autonomy (i.e., [Ca2+/CaM-independent]/[Ca2+/CaM-stimulated] activity) was measured in separate soluble kinase assays using the high-affinity CaMKII substrate AC-2. Autonomous activity was measured immediately after the prereaction (pre-SPOTs) and following the SPOT reaction (post-SPOTs); mean ± s.e.m. (). (d) Ca2+/CaM-stimulated SPOT substrate phosphorylation for various T287 autophosphorylation conditions; mean ± s.d. (). Results represent one-way ANOVA of log-normalized data with Holm-Šídák posttest compared to 0% autonomy. ; ; .

3. Discussion

Like other multifunctional protein kinases, CaMKII exhibits broad substrate specificity, targeting substrates involved in carbohydrate, amino acid and lipid metabolism, neurotransmitter synthesis and release, ion channels, receptors, transcription and translation, cytoskeletal organization and dynamics, cell cycle control, and Ca2+ homoeostasis [8]. In the nervous system, the role of CaMKII in synaptic plasticity has been intensively studied from both its role as a structural protein and as kinase [47]. In fact, while a number of different proteins and effector pathways have been implicated in regulating LTP over the years, CaMKII is considered to be a key mediator of this plasticity [46]. CaMKII is best known for signaling in the postsynaptic neuron for regulating the macromolecular signaling complex known as the PSD, a cytoskeletal organelle important for postsynaptic function. CaMKII translocation to the PSD is thought to be a critical feature of its function [4751], a process that presumably functions to sequester and localize CaMKII activated with postsynaptic activity. The NMDA-R complex is a critical targeting molecule for CaMKII localization to the PSD and a substrate [27]. CaMKII phosphorylation of S1303 on the GluN2B subunit of the NMDA-R [52] reduces desensitization of the channel and hypothesized to serve as a feed-forward mechanism to sustain Ca2+ entry during learning and memory [53]. In addition to NMDA-R targeting, densin-180 and alpha-actinin also serve to recruit CaMKII to the PSD, possibly to simultaneously interact with CaMKII via multivalent contacts to nucleate and create a defined signaling complex [54] for the >40 CaMKII substrates identified at the PSD [5557]. Thus, CaMKII translocation to the PSD appears to involve complex interactions that function to restrict CaMKII signaling to active postsynaptic sites. However, once targeted to the PSD, it is unknown how different substrates are selected to create specific functional outputs, including opposing forms of synaptic plasticity LTP and LTD. Thus, the purpose of our study was to determine how graded activity and T287 autophosphorylation impact the output of CaMKII, substrate phosphorylation, and selectivity.

We reasoned that for graded levels of CaMKII activation to be important for its signaling, then there should be an intrinsic advantage for multiple subunits to be activated within the holoenzyme versus a similar catalytic level of individual or monomeric CaMKII molecules in phosphorylating PSD substrates. Our studies revealed that while we began our phosphorylation reactions with ~2-fold monomeric CaMKII activity (as measured via a soluble assay) compared to multimeric CaMKII, the native form displays a catalytic advantage at phosphorylating purified PSDs in vitro. Thus, while the early work by Cline and colleagues using viral expression of monomeric constitutively active CaMKII (T286D) [5860] clearly shows that monomeric CaMKII modulates axonal growth, glutamatergic synapse maturation and stabilization of dendritic arbors, our work demonstrates that the native holoenzyme has an advantage over the monomer in PSD phosphorylation in vitro. We reasoned that the diffusion-restricted substrate environment like the PSD with many different CaMKII substrates (and targeting proteins) takes advantage of the multivalent structure of CaMKII [54], a process that could also take advantage of graded activity within the same CaMKII holoenzyme. Multivalent interactions between CaMKII and its targeting proteins and substrates [54] may contribute to the catalytic advantage observed for native CaMKII over monomeric CaMKII. Future studies are required to explore the interplay between autophosphorylation-dependent substrate gating and the contribution of known PSD targeting proteins (e.g., NR2B, actin, and actinin) in this enhanced PSD phosphorylation by autophosphorylated CaMKII.

We explored this idea directly using multisubstrate SPOT membranes with peptide substrates synthesized from well-known CaMKII substrates. The four model substrates we selected differed in affinity (GluN2BS1303 > Syntide > VimentinS83 and GluA1S849), possessed divergent phosphorylation motifs, and have been previously characterized in vivo and in vitro as protein and peptide substrates. Unlike the other three substrates, the GluN2BS1303 is a classic example of a T-site-interacting substrate (i.e., mimics CaMKII’s ARD) which allows for this substrate to form high-affinity stable interactions to influence targeting [42] and catalytic activity [41, 61]. We elected to use the multisubstrate SPOT array approach because like isolated PSDs, we can explore CaMKII phosphorylation of diverse substrates simultaneously in a diffusion-restricted environment; however, unlike the PSD, the SPOT arrays avoid the intrinsic complexity produced by endogenous CaMKII, scaffolding, and regulatory proteins, factors that can be envisioned to influence our enzymatic analysis of substrate specificity.

To create graded levels of CaMKII activity autophosphorylation, we considered the following. First, the intraholoenzyme-intermolecular T287 autophosphorylation reaction is rapid (i.e., seconds) under saturating Ca2+/CaM conditions [15]. Second, while limiting Ca2+/CaM used to restrict the activation process, it will ultimately produce a heterogeneous mixture of activation states in CaMKII (±T287 autophosphorylation, ±T306 autophosphorylation, and/or ±Ca2+/CaM activation) that will be difficult to interpret. Third, creating mixed holoenzymes of dead or T287A mutants in our insect expression system has an obvious limitation in that quantitatively controlling the mutant to wild-type subunit ratios within individual holoenzymes will be difficult. Fourth, inhibitory autophosphorylation is used by CaMKII in vivo to create “capped” or Ca2+/CaM-insensitive subunits [4]. Thus, we elected to create graded levels of CaMKII activation by using this intrinsic property of CaMKII whereby inhibitory (Ca2+/CaM-independent) basal autophosphorylation inactivates CaMKII subunits [19, 3840]. We created holoenzyme populations with varying ratios of capped subunits (nonresponsive to Ca2+/CaM and inactive) by incubating naïve multimeric CaMKII with Mg2+/ATP. This intramolecular subunit reaction [62] occurs in the absence of Ca2+/CaM as the autophosphorylation event leading to capping occurs within the CaM-binding domain (T306 in δ/β/γ CaMKII and T305 in α CaMKII) [38]. Using this strategy, we created CaMKII holoenzymes with graded levels of activity in the presence of saturating Ca2+/CaM to reveal dramatic shifts in the substrate specificity of the kinase. Higher affinity substrates appear to be preferentially phosphorylated when limited subunits are available for activation per holoenzyme (1/14). However, intermediate and weak substrates become utilized as multiple subunits are available for activation within the holoenzyme (5/15 and 14/14). This analysis requires the assumption that the capping protocol limiting subunit availability to Ca2+/CaM is stochastic at the microscopic level (i.e., individual subunit). The substrate phosphorylation profile for the CaMKII multimer (1/14) looks very similar to monomeric CaMKII in the absence of T287 autophosphorylation, suggesting that at the macroscopic level (bulk holoenzymes in solution), the theoretical 1/14 condition functionally approximates a single non-T287 autophosphorylated state. While we did have success in creating CaMKII with graded levels of activity (under saturating Ca2+/CaM), this protocol does not have sufficient control to separate different activation states (± T287 autophosphorylation in the capped CaMKII holoenzymes). Therefore, we employed monomeric CaMKII to explore the specific question of whether T287 autophosphorylation impacts substrate selectivity.

The role of autophosphorylation in regulating substrate accessibility was examined more recently and found to have mild [32] or no effect [33] on substrate phosphorylation. While differences in our conclusions may be attributable to different species of CaMKII (human versus rat) and isoforms (α versus δ), we do not favor these possibilities as the catalytic domain of all four CaMKII isoforms is highly conserved [9]. Historically, early structure-function studies demonstrated a lag in substrate catalysis that was relieved by CaMKII autophosphorylation [63]. Our substrate reactions used preautophosphorylated CaMKII and sufficiently dilute monomeric CaMKII in order to minimize this caveat. Our studies with monomer and native multimeric CaMKII indicate that graded levels of activation and/or T287 autophosphorylation are important determinants of intermediate and weak substrate phosphorylation. While high-affinity substrate phosphorylation is also enhanced by T287 autophosphorylation in CaMKIIm+P, the effect was not seen in native CaMKII due to avidity effects. Whether this CaMKIIholo+P phenomenon is simply an artifact of peptide density generated in our SPOT membranes or operates in complex substrate environments like the PSD requires additional studies. However, our studies suggest that high-affinity substrates do not have the same requirement for T287 autophosphorylation as intermediate- or low-affinity substrates. In total, our data consistent with PSD substrates are tuned for phosphorylation by the multivalent structure of CaMKII and T287 autophosphorylation specifically broadens substrate specificity by enhancing intermediate and weak substrate utilization.

4. Conclusions

The unique structure and autoregulation make CaMKII an ideal sensor to encode Ca2+-spike frequency into graded levels of activity and T287 autophosphorylation within the holoenzyme. While this process is influenced by Ca2+-spike duration and amplitude [20] as well as phosphatase activity [21], alternative spliced linker length and CaM availability [22, 23], how the number of activated subunits and/or extent of T287 autophosphorylation elicits different functional outputs is unknown. Our studies indicate that the activation state of CaMKII, and specifically T287 autophosphorylation, is critical drivers of substrate selectivity. Furthermore, our studies suggest a mechanistic explanation for how CaMKII signaling could regulate even opposing functional outputs like LTP and LTD via alterations in substrate selectivity. Future studies are required to determine if graded activity and autophosphorylation operates within complex heterogeneous compartments like the PSD. Furthermore, while our highly reduced biochemical studies suggest that the substrate gating effect of T287 autophosphorylation is an intrinsic property of the catalytic domain (and independent of the holoenzyme), further studies are required to illuminate how this mechanism differentially regulates substrate accessibility.

5. Materials and Methods

5.1. Expression and Purification of CaMKII and Calmodulin

Recombinant human CaMKIIδ (NCBI RefSeq: NP_742113.1) with an N-terminal 6xHN tag was integrated into a baculoviral construct (BacPAK9-6xHN), amplified in Sf9 insect cells (expression systems), and expressed in Hi5 (T. ni; (Expression Systems)) insect cells [64]. Site-directed mutagenesis was used to generate monomeric hCaMKIIδ1–317 (i.e., CaMKIIm) (by truncation through the addition of stop codon at aa318). Kinases were purified under reducing conditions by affinity chromatography (NiNTA resin) followed by size exclusion chromatography (Sephacryl S-400 [holoenzyme] or S-300 [CaMKIIm]) using an Äkta Purifier (Amersham). SDS-PAGE of the purified proteins revealed a single band with purities > 98%. Recombinant calmodulin was expressed and purified in E.coli as described previously [12, 65] via boiling, ammonium sulfate precipitation, and phenyl-sepharose affinity chromatography.

5.2. Postsynaptic Density Purification

PSDs were acquired from male/female adult rats (Sprague-Dawley) using previously defined procedures involving differential centrifugation with various sucrose gradients followed by Triton X-100 extraction and collection of the Triton-insoluble fraction [34, 35]. Several steps were taken to limit and reduce the Ca2+/CaM-stimulated and autonomous activity of endogenous CaMKII within the PSDs. Rat brains were extracted and snap frozen in liquid nitrogen within 90 secs after sacrificing the animal, as CaMKII is known to activate and translocate to the PSDs within minutes in response to cell death [36].

5.3. Postsynaptic Density Phosphorylation

The PSDs were dephosphorylated to prevent autonomous activity of endogenous CaMKII (resulting from prior T287 autophosphorylation), as well as to reduce the phosphorylation state of endogenous substrates. This was accomplished with 1.5 μM PP1α (6xHis purified) overnight at 4°C or at RT for 2 hours with shaking in the presence of 50 mM HEPES pH 7.4, 100 mM NaCl, 1 mM MnCl2, 0.015 Brij 35, 2.5 mM DTT, 10 μM KN-93, and 2x calbiochem protease inhibitor cocktail set V. Phosphorylation reactions were carried out on dephosphorylated PSDs in the presence of 2x phosphatase inhibitor cocktail (Calbiochem) as well as the small molecule CaMKII inhibitor KN93 which only inhibits the autoinhibited kinase (naïve). PSD phosphorylation reactions were carried out on ~50 μg of dephosphorylated PSDs (see above) in the presence of 20 mM HEPES pH 7.4, 100 mM NaCl, 10 mM MgCl2, 0.5 mM CaCl2, 5 μM CaM, 100 μM cold ATP, 120 μCi/ml [γ-32P]-ATP, and 350 nM CaMKII (per subunit). For reactions involving T287 autophosphorylated CaMKII, a prereaction was performed in the presence of 20 mM HEPES pH 7.4, 100 mM NaCl, 10 mM MgCl2, 0.5 mM CaCl2, 5 μM CaM, 500 μM ATPγS, and 3.5 μM kinase for 10 min at 30°C (for monomer) or on ice (for the holoenzyme). PSD phosphorylation reactions (60 μl total) were incubated at room temperature for 4 min unless otherwise noted. Reactions were terminated with the addition of 500 μl of 100 mM sodium phosphate pH 7.0, 1 M NaCl, and 10 mM EDTA) followed by centrifugation to obtain a PSD pellet. Pellet was three times in the same termination buffer. The PSD pellet was resuspended in 2x LDS sample buffer and resuspended using a horn sonicator. SDS-PAGE was performed followed by Coomassie staining and drying of the gel. Total protein was assessed by densitometry of the Coomassie signal while the extent of radioactive phosphate incorporation was quantified using a Fujifilm phosphorimager and profiles measured as photostimulated luminescence (PSL/mm2). The densitometric intensity data was converted to area under the curve (AUC) to determine to protein or total phosphorylation.

5.4. Soluble Peptide Assays

Soluble peptide substrates (generally 15mers) were obtained (Biopeptide Co. Inc.) and utilized in standard radioactive activity assays. In typical Ca2+/CaM-dependent reactions, 50-200 μM substrate was used in the presence of 50 mM HEPES pH 7.4, 100 mM NaCl, 10 mM MgCl2, 0.2 mM CaCl2, 1 μM CaM, 100 μM cold ATP, 60 μCi/ml [γ-32P] ATP, and 5–10 nM CaMKII (per subunit). In Ca2+/CaM-independent reactions, 5 mM EGTA (or BAPTA) was used to chelate free Ca2+. This constitutive (i.e., autonomous) activity in the absence of Ca2+ was expressed, a percentage of the Ca2+/CaM-dependent activity. Reactions were carried out at 30°C for 1 minute and quenched by spotting onto Whatman Grade P81 ion exchange chromatography paper (Whatman, GE Healthcare, Piscataway, NJ). The filter papers were washed with 75 mM phosphoric acid 3 times for 5 minutes each and subsequently quantified in a scintillation counter (Beckman) via the Cerenkov counting method [66, 67]. AC-2 (KKALRRQEtVDAL), a CaMKII substrate derived from the CaMKII T287 autophosphorylation site was used for standard soluble peptide substrate reactions as described previously [64, 68].

5.5. Holoenzyme Inactivation

CaMKIIholo (600 nM) was exposed to inhibitory autophosphorylation (T306/7 capping) for 0, 5, or 15 minutes in the presence of 50 mM HEPES pH 7.4, 100 mM NaCl, 10 mM MgCl2, and 500 μM ATP. Inactivated states of CaMKII were then diluted to 60 nM for SPOT substrate phosphorylation reactions (described below). For all states of preinactivation (0, 5, 8, or 15 min), the SPOT phosphorylation was normalized to the number of active subunits per holoenzyme (assuming each holoenzyme is tetradecameric). The number of active subunits per holoenzyme was assessed by measuring the Ca2+/CaM-dependent activity of each inactivation condition in standard soluble peptide substrate phosphorylation assays using AC-2 peptide. These activity assays were performed at the beginning and end (4 min) of the SPOT phosphorylation assay, their values averaged together. The activity from naïve CaMKIIholo (i.e., 0 min preinactivation) was considered to maximal (i.e., all 14 subunits activated) and was thus used to calculate the average enzymatic activity per subunit.

5.6. Preautophosphorylation Reactions

For reactions involving T287 autophosphorylated CaMKII, a prereaction was performed in the presence of 20 mM HEPES pH 7.4, 100 mM NaCl, 10 mM MgCl2, 0.5 mM CaCl2, 5 μM CaM, 500 μM cold ATP, and 500 nM kinase for 10 minutes on ice. To achieve varying numbers/percentages of T287 autophosphorylated subunits (%AutoP), similar prereactions were used with 50 nM kinase incubation times of 0, 5, or 20 minutes on ice, or 10 minutes at 30°C (Figure 5; preconditions 1–4, resp.).

5.7. In Vitro SPOT Phosphorylation Assay

Peptides (generally 15mers) were synthesized using the SPOT method [69, 70] onto β-alanine derivatized cellulose membranes via a MultiPep synthesizer (Intavis AG, Cologne, Germany). The peptide blots were blocked with 5% BSA in 50 mM HEPES pH 7.4, 100 mM NaCl, 1 mM EDTA, and 0.02% NP-40 for 30 minutes followed by three washes in 100 mM Tris–HCl pH 7.4. Membranes were subjected to a kinase phosphorylation assay in the presence of 50 mM HEPES pH 7.4, 100 mM NaCl, 10 mM MgCl2, 0.2 mM CaCl2, 1 μM CaM, 0.2 mg/ml BSA, 1 mM DTT, 100 μM cold ATP, 6–12 μCi/ml (γ-32P) ATP, and 5–10 nM CaMKII (per subunit). The reactions were incubated at room temperature for 4 minutes unless otherwise noted, terminated with three washes (100 mM sodium phosphate pH 7.0, 1 M NaCl, and 10 mM EDTA), and dried. The extent of radioactive phosphate incorporation was quantified using a Fujifilm phosphoimager and expressed as photostimulated luminescence (PSL/mm2) for a 1.5 mm × 1.5 mm circle centered on each spot. Three to four replicate spots were averaged for each peptide for a given condition with error bars indicating s.d., unless otherwise noted. GluN2BS1303 = RNKLRRQHsYDTFVD; Syntide = PLARTLsVAGLPGKK; GluA1S849 = GFCLIPQQsINEAIR; VimentinS83 = PGVRLLQDsVDFSLA.

5.8. Data and Statistical Analysis

One-way analysis of variance (ANOVA) with subsequent Holm-Šídák posttest was performed on log transformed to account for mean-dependent variances, compared to control (#). Statistical analyses were performed using SigmaPlot 12.5. Statistical significance is denoted by the number of symbols for different values (e.g., , , and ).

Abbreviations

CaMKII:Calcium/calmodulin-dependent protein kinase II
AUC:Area under the curve
PSD:Postsynaptic density.

Disclosure

A. Hudmon’s present address is Department of Medicinal Chemistry and Molecular Pharmacology, College of Pharmacy, Purdue University, West Lafayette, IN 47907, USA. D. E. Johnson present address is AIT Bioscience, 7840 Innovation Blvd, Indianapolis, IN 46278, USA.

Conflicts of Interest

The authors declare no conflicts of interest. The authors declare no competing financial interests.

Acknowledgments

The authors thank N. M. Ashpole, J. M. Bradshaw, H. Schulman, M. C. Sousa, and M. N. Waxham for valuable suggestions and insightful comments during various iterations of this manuscript. This work was supported by the National Institutes of Health Grant NS078171 to A. Hudmon.

References

  1. J. D. Sweatt, “Neural plasticity and behavior - sixty years of conceptual advances,” Journal of Neurochemistry, vol. 139, Supplement 2, pp. 179–199, 2016. View at Publisher · View at Google Scholar · View at Scopus
  2. P. K. Stanton, “LTD, LTP, and the sliding threshold for long-term synaptic plasticity,” Hippocampus, vol. 6, no. 1, pp. 35–42, 1996. View at Publisher · View at Google Scholar
  3. J. Lisman, “The CaM kinase II hypothesis for the storage of synaptic memory,” Trends in Neurosciences, vol. 17, no. 10, pp. 406–412, 1994. View at Publisher · View at Google Scholar · View at Scopus
  4. S. J. Coultrap and K. U. Bayer, “CaMKII regulation in information processing and storage,” Trends in Neurosciences, vol. 35, no. 10, pp. 607–618, 2012. View at Publisher · View at Google Scholar · View at Scopus
  5. B. C. Shonesy, N. Jalan-Sakrikar, V. S. Cavener, and R. J. Colbran, “CaMKII: a molecular substrate for synaptic plasticity and memory,” Progress in Molecular Biology and Translational Science, vol. 122, pp. 61–87, 2014. View at Publisher · View at Google Scholar · View at Scopus
  6. J. W. Hell, “CaMKII: claiming center stage in postsynaptic function and organization,” Neuron, vol. 81, no. 2, pp. 249–265, 2014. View at Publisher · View at Google Scholar · View at Scopus
  7. B. E. Herring and R. A. Nicoll, “Long-term potentiation: from CaMKII to AMPA receptor trafficking,” Annual Review of Physiology, vol. 78, no. 1, pp. 351–365, 2016. View at Publisher · View at Google Scholar · View at Scopus
  8. A. Hudmon and H. Schulman, “Structure-function of the multifunctional Ca2+/calmodulin-dependent protein kinase II,” Biochemical Journal, vol. 364, no. 3, pp. 593–611, 2002. View at Publisher · View at Google Scholar · View at Scopus
  9. A. Hudmon and H. Schulman, “Neuronal CA2+/calmodulin-dependent protein kinase II: the role of structure and autoregulation in cellular function,” Annual Review of Biochemistry, vol. 71, no. 1, pp. 473–510, 2002. View at Publisher · View at Google Scholar · View at Scopus
  10. J. Lisman, H. Schulman, and H. Cline, “The molecular basis of CaMKII function in synaptic and behavioural memory,” Nature Reviews Neuroscience, vol. 3, no. 3, pp. 175–190, 2002. View at Publisher · View at Google Scholar · View at Scopus
  11. P. Rellos, A. C. W. Pike, F. H. Niesen et al., “Structure of the CaMKIIdelta/calmodulin complex reveals the molecular mechanism of CaMKII kinase activation,” PLoS Biology, vol. 8, no. 7, article e1000426, 2010. View at Publisher · View at Google Scholar · View at Scopus
  12. T. R. Gaertner, S. J. Kolodziej, D. Wang et al., “Comparative analyses of the three-dimensional structures and enzymatic properties of alpha, beta, gamma and delta isoforms of Ca2+−calmodulin-dependent protein kinase II,” Journal of Biological Chemistry, vol. 279, no. 13, pp. 12484–12494, 2004. View at Publisher · View at Google Scholar · View at Scopus
  13. L. H. Chao, M. M. Stratton, I. H. Lee et al., “A mechanism for tunable autoinhibition in the structure of a human Ca2+/calmodulin- dependent kinase II holoenzyme,” Cell, vol. 146, no. 5, pp. 732–745, 2011. View at Publisher · View at Google Scholar · View at Scopus
  14. R. C. Rich and H. Schulman, “Substrate-directed function of calmodulin in autophosphorylation of Ca2+/calmodulin-dependent protein kinase II,” Journal of Biological Chemistry, vol. 273, no. 43, pp. 28424–28429, 1998. View at Publisher · View at Google Scholar · View at Scopus
  15. P. I. Hanson, T. Meyer, L. Stryer, and H. Schulman, “Dual role of calmodulin in autophosphorylation of multifunctional cam kinase may underlie decoding of calcium signals,” Neuron, vol. 12, no. 5, pp. 943–956, 1994. View at Publisher · View at Google Scholar · View at Scopus
  16. A. J. Baucum 2nd, B. C. Shonesy, K. L. Rose, and R. J. Colbran, “Quantitative proteomics analysis of CaMKII phosphorylation and the CaMKII interactome in the mouse forebrain,” ACS Chemical Neuroscience, vol. 6, no. 4, pp. 615–631, 2015. View at Publisher · View at Google Scholar · View at Scopus
  17. T. Meyer, P. I. Hanson, L. Stryer, and H. Schulman, “Calmodulin trapping by calcium-calmodulin-dependent protein kinase,” Science, vol. 256, no. 5060, pp. 1199–1202, 1992. View at Publisher · View at Google Scholar
  18. C. M. Schworer, R. J. Colbran, J. R. Keefer, and T. R. Soderling, “Ca2+/calmodulin-dependent protein kinase II. Identification of a regulatory autophosphorylation site adjacent to the inhibitory and calmodulin-binding domains,” The Journal of Biological Chemistry, vol. 263, no. 27, pp. 13486–13489, 1988. View at Google Scholar
  19. L. L. Lou and H. Schulman, “Distinct autophosphorylation sites sequentially produce autonomy and inhibition of the multifunctional Ca2+/calmodulin-dependent protein kinase,” Journal of Neuroscience, vol. 9, no. 6, pp. 2020–2032, 1989. View at Google Scholar
  20. P. De Koninck and H. Schulman, “Sensitivity of CaM kinase II to the frequency of Ca2+ oscillations,” Science, vol. 279, no. 5348, pp. 227–230, 1998. View at Publisher · View at Google Scholar · View at Scopus
  21. R. J. Colbran, “Protein phosphatases and calcium/calmodulin-dependent protein kinase II-dependent synaptic plasticity,” Journal of Neuroscience, vol. 24, no. 39, pp. 8404–8409, 2004. View at Publisher · View at Google Scholar · View at Scopus
  22. A. Persechini and P. M. Stemmer, “Calmodulin is a limiting factor in the cell,” Trends in Cardiovascular Medicine, vol. 12, no. 1, pp. 32–37, 2002. View at Publisher · View at Google Scholar · View at Scopus
  23. R. P. Estep, K. A. Alexander, and D. R. Storm, “Regulation of free calmodulin levels in neurons by neuromodulin: relationship to neuronal growth and regeneration,” Current Topics in Cellular Regulation, vol. 31, pp. 161–180, 1989. View at Publisher · View at Google Scholar
  24. D. J. Grab, R. K. Carlin, and P. Siekevitz, “Function of a calmodulin in postsynaptic densities. II. Presence of a calmodulin-activatable protein kinase activity,” The Journal of Cell Biology, vol. 89, no. 3, pp. 440–448, 1981. View at Publisher · View at Google Scholar
  25. P. T. Kelly and P. R. Montgomery, “Subcellular localization of the 52,000 molecular weight major postsynaptic density protein,” Brain Research, vol. 233, no. 2, pp. 265–286, 1982. View at Publisher · View at Google Scholar · View at Scopus
  26. P. T. Kelly, T. L. McGuinness, and P. Greengard, “Evidence that the major postsynaptic density protein is a component of a Ca2+/calmodulin-dependent protein kinase,” Proceedings of the National Academy of Sciences of the United States of America, vol. 81, no. 3, pp. 945–949, 1984. View at Publisher · View at Google Scholar
  27. M. Sanhueza and J. Lisman, “The CaMKII/NMDAR complex as a molecular memory,” Molecular Brain, vol. 6, no. 1, p. 10, 2013. View at Publisher · View at Google Scholar · View at Scopus
  28. K. Kim, T. Saneyoshi, T. Hosokawa, K. Okamoto, and Y. Hayashi, “Interplay of enzymatic and structural functions of CaMKII in long-term potentiation,” Journal of Neurochemistry, vol. 139, no. 6, pp. 959–972, 2016. View at Publisher · View at Google Scholar · View at Scopus
  29. B. G. Mockett, D. Guevremont, M. Wutte, S. R. Hulme, J. M. Williams, and W. C. Abraham, “Calcium/calmodulin-dependent protein kinase II mediates group I metabotropic glutamate receptor-dependent protein synthesis and long-term depression in rat hippocampus,” Journal of Neuroscience, vol. 31, no. 20, pp. 7380–7391, 2011. View at Publisher · View at Google Scholar · View at Scopus
  30. K. C. Marsden, J. B. Beattie, J. Friedenthal, and R. C. Carroll, “NMDA receptor activation potentiates inhibitory transmission through GABA receptor-associated protein-dependent exocytosis of GABAA receptors,” Journal of Neuroscience, vol. 27, no. 52, pp. 14326–14337, 2007. View at Publisher · View at Google Scholar · View at Scopus
  31. S. J. Coultrap, R. K. Freund, H. O’Leary et al., “Autonomous CaMKII mediates both LTP and LTD using a mechanism for differential substrate site selection,” Cell Reports, vol. 6, no. 3, pp. 431–437, 2014. View at Publisher · View at Google Scholar · View at Scopus
  32. S. J. Coultrap, K. Barcomb, and K. U. Bayer, “A significant but rather mild contribution of T286 autophosphorylation to Ca2+/CaM-stimulated CaMKII activity,” PLoS One, vol. 7, no. 5, article e37176, 2012. View at Publisher · View at Google Scholar · View at Scopus
  33. S. J. Coultrap, I. Buard, J. R. Kulbe, M. L. Dell'Acqua, and K. U. Bayer, “CaMKII autonomy is substrate-dependent and further stimulated by Ca2+/calmodulin,” Journal of Biological Chemistry, vol. 285, no. 23, pp. 17930–17937, 2010. View at Publisher · View at Google Scholar · View at Scopus
  34. M. D. Ehlers, “Activity level controls postsynaptic composition and signaling via the ubiquitin-proteasome system,” Nature Neuroscience, vol. 6, no. 3, pp. 231–242, 2003. View at Publisher · View at Google Scholar · View at Scopus
  35. M. T. Swulius, Y. Kubota, A. Forest, and M. N. Waxham, “Structure and composition of the postsynaptic density during development,” The Journal of Comparative Neurology, vol. 518, no. 20, pp. 4243–4260, 2010. View at Publisher · View at Google Scholar · View at Scopus
  36. T. Suzuki, K. Okumura-Noji, R. Tanaka, and T. Tada, “Rapid translocation of cytosolic Ca2+/calmodulin-dependent protein-kinase-II into postsynaptic density after decapitation,” Journal of Neurochemistry, vol. 63, no. 4, pp. 1529–1537, 1994. View at Publisher · View at Google Scholar · View at Scopus
  37. J. A. Ubersax and J. E. Ferrell Jr., “Mechanisms of specificity in protein phosphorylation,” Nature Reviews. Molecular Cell Biology, vol. 8, no. 7, pp. 530–541, 2007. View at Publisher · View at Google Scholar · View at Scopus
  38. R. J. Colbran, “Inactivation of Ca2+/calmodulin-dependent protein kinase II by basal autophosphorylation,” Journal of Biological Chemistry, vol. 268, no. 10, pp. 7163–7170, 1993. View at Google Scholar
  39. R. J. Colbran and T. R. Soderling, “Calcium/calmodulin-independent autophosphorylation sites of calcium/calmodulin-dependent protein kinase II. Studies on the effect of phosphorylation of threonine 305/306 and serine 314 on calmodulin binding using synthetic peptides,” Journal of Biological Chemistry, vol. 265, no. 19, pp. 11213–11219, 1990. View at Google Scholar
  40. B. L. Patton, S. G. Miller, and M. B. Kennedy, “Activation of type II calcium/calmodulin-dependent protein kinase by Ca2+/calmodulin is inhibited by autophosphorylation of threonine within the calmodulin-binding domain,” Journal of Biological Chemistry, vol. 265, no. 19, pp. 11204–11212, 1990. View at Google Scholar
  41. K. U. Bayer, P. de Koninck, A. S. Leonard, J. W. Hell, and H. Schulman, “Interaction with the NMDA receptor locks CaMKII in an active conformation,” Nature, vol. 411, no. 6839, pp. 801–805, 2001. View at Publisher · View at Google Scholar · View at Scopus
  42. S. Strack, R. B. McNeill, and R. J. Colbran, “Mechanism and regulation of calcium/calmodulin-dependent protein kinase II targeting to the NR2B subunit of the N-methyl-D-aspartate receptor,” Journal of Biological Chemistry, vol. 275, no. 31, pp. 23798–23806, 2000. View at Publisher · View at Google Scholar · View at Scopus
  43. X. Liao, J. Su, and M. Mrksich, “An adaptor domain-mediated autocatalytic interfacial kinase reaction,” Chemistry, vol. 15, no. 45, pp. 12303–12309, 2009. View at Publisher · View at Google Scholar · View at Scopus
  44. J. L. Naffin, Y. Han, H. J. Olivos, M. M. Reddy, T. Sun, and T. Kodadek, “Immobilized peptides as high-affinity capture agents for self-associating proteins,” Chemistry & Biology, vol. 10, no. 3, pp. 251–259, 2003. View at Publisher · View at Google Scholar · View at Scopus
  45. A. Ikeda, S. Okuno, and H. Fujisawa, “Studies on the generation of Ca2+/calmodulin-independent activity of calmodulin-dependent protein kinase II by autophosphorylation. Autothiophosphorylation of the enzyme,” Journal of Biological Chemistry, vol. 266, no. 18, pp. 11582–11588, 1991. View at Google Scholar
  46. J. R. Sanes and J. W. Lichtman, “Can molecules explain long-term potentiation?” Nature Neuroscience, vol. 2, no. 7, pp. 597–604, 1999. View at Publisher · View at Google Scholar · View at Scopus
  47. K. Shen and T. Meyer, “Dynamic control of CaMKII translocation and localization in hippocampal neurons by NMDA receptor stimulation,” Science, vol. 284, no. 5411, pp. 162–167, 1999. View at Publisher · View at Google Scholar · View at Scopus
  48. K. Shen, M. N. Teruel, J. H. Connor, S. Shenolikar, and T. Meyer, “Molecular memory by reversible translocation of calcium/calmodulin-dependent protein kinase II,” Nature Neuroscience, vol. 3, no. 9, pp. 881–886, 2000. View at Publisher · View at Google Scholar · View at Scopus
  49. M. R. Gleason, S. Higashijima, J. Dallman, K. Liu, G. Mandel, and J. R. Fetcho, “Translocation of CaM kinase II to synaptic sites in vivo,” Nature Neuroscience, vol. 6, no. 3, pp. 217-218, 2003. View at Publisher · View at Google Scholar · View at Scopus
  50. N. Otmakhov, J. H. Tao-Cheng, S. Carpenter et al., “Persistent accumulation of calcium/calmodulin-dependent protein kinase II in dendritic spines after induction of NMDA receptor-dependent chemical long-term potentiation,” Journal of Neuroscience, vol. 24, no. 42, pp. 9324–9331, 2004. View at Publisher · View at Google Scholar · View at Scopus
  51. A. R. Halt, R. F. Dallapiazza, Y. Zhou et al., “CaMKII binding to GluN2B is critical during memory consolidation,” The EMBO Journal, vol. 31, no. 5, pp. 1203–1216, 2012. View at Publisher · View at Google Scholar · View at Scopus
  52. R. V. Omkumar, M. J. Kiely, A. J. Rosenstein, K. T. Min, and M. B. Kennedy, “Identification of a phosphorylation site for calcium/calmodulindependent protein kinase II in the NR2B subunit of the N-methyl-D-aspartate receptor,” Journal of Biological Chemistry, vol. 271, no. 49, pp. 31670–31678, 1996. View at Publisher · View at Google Scholar · View at Scopus
  53. S. J. Tavalin and R. J. Colbran, “CaMKII-mediated phosphorylation of GluN2B regulates recombinant NMDA receptor currents in a chloride-dependent manner,” Molecular and Cellular Neurosciences, vol. 79, pp. 45–52, 2017. View at Publisher · View at Google Scholar
  54. A. J. Robison, M. A. Bass, Y. Jiao et al., “Multivalent interactions of calcium/calmodulin-dependent protein kinase II with the postsynaptic density proteins NR2B, densin-180, and alpha-actinin-2,” Journal of Biological Chemistry, vol. 280, no. 42, pp. 35329–35336, 2005. View at Publisher · View at Google Scholar · View at Scopus
  55. H. Jaffe, L. Vinade, and A. Dosemeci, “Identification of novel phosphorylation sites on postsynaptic density proteins,” Biochemical and Biophysical Research Communications, vol. 321, no. 1, pp. 210–218, 2004. View at Publisher · View at Google Scholar · View at Scopus
  56. Y. Yoshimura, C. Aoi, and T. Yamauchi, “Investigation of protein substrates of Ca(2+)/calmodulin-dependent protein kinase II translocated to the postsynaptic density,” Brain Research. Molecular Brain Research, vol. 81, no. 1-2, pp. 118–128, 2000. View at Publisher · View at Google Scholar · View at Scopus
  57. Y. Yoshimura, T. Shinkawa, M. Taoka, K. Kobayashi, T. Isobe, and T. Yamauchi, “Identification of protein substrates of Ca(2+)/calmodulin-dependent protein kinase II in the postsynaptic density by protein sequencing and mass spectrometry,” Biochemical and Biophysical Research Communications, vol. 290, no. 3, pp. 948–954, 2002. View at Publisher · View at Google Scholar · View at Scopus
  58. D. J. Zou and H. T. Cline, “Expression of constitutively active CaMKII in target tissue modifies presynaptic axon arbor growth,” Neuron, vol. 16, no. 3, pp. 529–539, 1996. View at Publisher · View at Google Scholar · View at Scopus
  59. G. Wu, R. Malinow, and H. T. Cline, “Maturation of a central glutamatergic synapse,” Science, vol. 274, no. 5289, pp. 972–976, 1996. View at Publisher · View at Google Scholar · View at Scopus
  60. G. Y. Wu and H. T. Cline, “Stabilization of dendritic arbor structure in vivo by CaMKII,” Science, vol. 279, no. 5348, pp. 222–226, 1998. View at Publisher · View at Google Scholar · View at Scopus
  61. M. Praseeda, M. Mayadevi, and R. V. Omkumar, “Interaction of peptide substrate outside the active site influences catalysis by CaMKII,” Biochemical and Biophysical Research Communications, vol. 313, no. 4, pp. 845–849, 2004. View at Publisher · View at Google Scholar · View at Scopus
  62. S. Mukherji and T. R. Soderling, “Regulation of Ca2+/calmodulin-dependent protein kinase II by inter- and intrasubunit-catalyzed autophosphorylations,” Journal of Biological Chemistry, vol. 269, no. 19, pp. 13744–13747, 1994. View at Google Scholar
  63. A. P. Kwiatkowski, D. J. Shell, and M. M. King, “The role of autophosphorylation in activation of the type II calmodulin-dependent protein kinase,” Journal of Biological Chemistry, vol. 263, no. 14, pp. 6484–6486, 1988. View at Google Scholar
  64. N. M. Ashpole and A. Hudmon, “Excitotoxic neuroprotection and vulnerability with CaMKII inhibition,” Molecular and Cellular Neurosciences, vol. 46, no. 4, pp. 720–730, 2011. View at Publisher · View at Google Scholar · View at Scopus
  65. S. I. Singla, A. Hudmon, J. M. Goldberg, J. L. Smith, and H. Schulman, “Molecular characterization of calmodulin trapping by calcium/calmodulin-dependent protein kinase II,” Journal of Biological Chemistry, vol. 276, no. 31, pp. 29353–29360, 2001. View at Publisher · View at Google Scholar · View at Scopus
  66. A. Hudmon, J. Aronowski, S. J. Kolb, and M. N. Waxham, “Inactivation and self-association of Ca2+/calmodulin-dependent protein kinase II during autophosphorylation,” Journal of Biological Chemistry, vol. 271, no. 15, pp. 8800–8808, 1996. View at Publisher · View at Google Scholar
  67. R. Roskoski Jr., “Assays of protein kinase,” Methods in Enzymology, vol. 99, pp. 3–6, 1983. View at Publisher · View at Google Scholar
  68. M. L. Bangaru, J. Meng, D. J. Kaiser et al., “Differential expression of CaMKII isoforms and overall kinase activity in rat dorsal root ganglia after injury,” Neuroscience, vol. 300, pp. 116–127, 2015. View at Publisher · View at Google Scholar · View at Scopus
  69. R. Frank, “The SPOT-synthesis technique. Synthetic peptide arrays on membrane supports—principles and applications,” Journal of Immunological Methods, vol. 267, no. 1, pp. 13–26, 2002. View at Publisher · View at Google Scholar · View at Scopus
  70. R. Frank and H. Overwin, “SPOT synthesis. Epitope analysis with arrays of synthetic peptides prepared on cellulose membranes,” Methods in Molecular Biology, vol. 66, pp. 149–169, 1996. View at Publisher · View at Google Scholar
  71. J. Zheng, E. A. Trafny, D. R. Knighton et al., “2.2 A refined crystal structure of the catalytic subunit of cAMP-dependent protein kinase complexed with MnATP and a peptide inhibitor,” Acta Crystallographica Section D Biological Crystallography, vol. 49, no. 3, pp. 362–365, 1993. View at Publisher · View at Google Scholar
  72. L. H. Chao, P. Pellicena, S. Deindl, L. A. Barclay, H. Schulman, and J. Kuriyan, “Intersubunit capture of regulatory segments is a component of cooperative CaMKII activation,” Nature Structural & Molecular Biology, vol. 17, no. 3, pp. 264–272, 2010. View at Publisher · View at Google Scholar · View at Scopus