CaMKIIβ binding to stable F-actin in vivo regulates F-actin filament stability

CaMKIIβ, but Not CaMKIIα, Localized to Prominent F-actin-Rich Structures in Cortical Neurons.

To investigate the mechanisms regulating CaMKIIβ association with F-actin in cortical neurons, we first examined CaMKIIβ localization in cortical neurons. CaMKIIβ was detected in the cytoplasm and processes of dissociated cortical neurons, but remarkably, CaMKIIβ was enriched in prominent elongated structures that were present primarily at the cell body (Fig. 1A). These discrete CaMKIIβ-enriched structures were detected exclusively in MAP2 immunostained neurons and strongly labeled by phalloidin. In fact, phalloidin labeling was eightfold greater in these structures than the soma [supporting information (SI) Fig. S1]. These phalloidin-labeled structures resembled bundled F-actin like that seen in filopodia. We term these F-actin rich protrusions “microspikes.”

Along with CaMKIIβ, CaMKIIα predominates in the brain. However, CaMKIIα and CaMKIIβ were differentially expressed in the developing cortex and cortical cultures (Fig. 1 C and D). CaMKIIβ expression was detected in E18 neocortex, but CaMKIIα was not detected until postnatal day 7 (P7). Cortical cultures expressed detectable levels of CaMKIIβ at 4 days in vitro (DIV), while CaMKIIα was not detected until 12 DIV. This temporal regulation in vitro is consistent with previous reports from hippocampal neurons (4). To determine if other isoforms of CaMKII are enriched in microspikes, we immunostained with CaMKII isoform antibodies. As expected, CaMKIIα specific antibodies did not detect CaMKIIα in cultured neurons at 4 DIV (data not shown). Antibodies that recognize CaMKIIα, γ, and δ isoforms detected CaMKII in the cytoplasm, indicating that CaMKII (γ and δ) are expressed in cortical neurons. However, confocal images showed that these other isoforms were not present in the microspikes (Fig. 1B).

CaMKIIβ Binding to F-actin Required an F-actin Binding Domain and Was Regulated by CaMKIIα.

Like endogenous CaMKIIβ, a GFP-tagged CaMKIIβ (GFP-CaMKIIβ), when expressed in neurons, localized in the cytoplasm and processes of neurons but was highly enriched in microspikes (Fig. 1 A, E, and H and Fig. 2A). This heterogeneous distribution, because of the enrichment of GFP-CaMKIIβ in microspikes where F-actin is enriched eightfold, demonstrates strong colocalization with F-actin and suggests a specific interaction with F-actin. To selectively examine the role of the putative F-actin binding and association domains to GFP-CaMKIIβ localization in the microspikes, we first examined localization of GFP-CaMKIIα, a GFP-CaMKIIβ mutant that lacks the F-actin binding domain (β-Δ355–392), a mutant consisting of only the putative F-actin binding domain (β-355–393), and a truncated mutant that lacks (β-1–370) and one that contains (β-394–544) the association domain. GFP-CaMKIIα, CaMKIIβ-Δ355–392, CaMKIIβ-394–544, and CaMKIIβ-355–544, but not CaMKIIβ-355–393 or CaMKIIβ-1–370, showed biased colocalization with F-actin and endogenous CaMKIIβ in the microspikes (see Fig. 1 F–H and Fig. S1). These data indicate that both the F-actin binding and the association domains contribute to localization to the microspikes. Although GFP-CaMKIIα, CaMKIIβ-Δ355–392, and CaMKIIβ-394–544 localized to microspikes, they were significantly less enriched in the microspikes than WT GFP-CaMKIIβ.

To examine the association of CaMKIIβ specifically with F-actin in the microspikes, GFP-CaMKIIβ was photobleached at a single microspike in neurons at 5 DIV (see Fig. 2 A and B). This FRAP analyses indicated that GFP-CaMKIIβ recovered more slowly than GFP alone. Detection of a slowed recovery after photobleaching suggests the presence of a binding partner that slowed down the mobility of GFP-CaMKIIβ. We next performed FRAP analyses to determine if GFP-CaMKIIα and CaMKIIβ-Δ355–392 interacted specifically with F-actin. In addition to identifying binding specificity, FRAP analyses with CaMKIIα and CaMKIIβ-Δ355–392 served as appropriate size controls for CaMKIIβ. GFP-CaMKIIβ-Δ355–392 recovered significantly faster than CaMKIIβ and CaMKIIα (Fig. 2 C and D). In addition, GFP-CaMKIIβ-Δ355–392 and CaMKIIα reached full recovery twofold faster than WT GFP-CaMKIIβ (see Fig. 2D). Together these data indicate that the putative F-actin binding domain is important, but not sufficient, for binding in F-actin-rich microspikes.

In cortical cultures, CaMKIIα expression was not detected until 12 DIV, indicating that our FRAP analyses at 5 DIV were not influenced by endogenous CaMKIIα expression. However, CaMKIIα and CaMKIIβ are both expressed later in development and in the adult. GFP-CaMKIIα, but not a CaMKIIβ mutant that lacks the association domain (β-1–370), colocalized with endogenous CaMKIIβ in microspikes (see Fig. 1 F and H). CaMKIIα and β form hetero-oligomers in vivo and in vitro CaMKIIβ subunits in the CaMKII hetero-oligomer were sufficient to localize CaMKIIα to F-actin (4, 11). However, whether CaMKIIα alters CaMKIIβ binding strength with F-actin is not known. Expression of GFP-CaMKIIβ and untagged CaMKIIα at 1:2 molar ratio significantly increased the recovery rate of GFP-CaMKIIβ (Fig. 2 E and F). This suggests that the isoform composition of the CaMKII hetero-oligomer determines the binding strength of CaMKII to F-actin.

CaMKIIβ-Bound Stable F-actin Is Significantly Stronger than Dynamic Actin.

Cellular F-actin is rarely stable but continuously undergoes polymerization and depolymerization. Understanding the interactions of two proteins is complicated when both are dynamic. To examine specific interactions between two proteins, one protein can be immobilized and FRAP performed on the potential interacting protein (12). We immobilized actin with jasplakinolide, which completely inhibited the recovery of GFP-actin, and cytochalasin D, which slowed GFP-actin recovery (13). CaMKIIβ and F-actin remained colocalized after jasplakinolide and cytochalasin D treatments. In fact, GFP-CaMKIIβ recovered significantly slower after actin was immobilized than it did in control conditions (Fig. 3). Furthermore, CaMKIIβ fully recovered in control conditions, but not when F-actin was stabilized (see Fig. 3 A and D). Even after 500 s, CaMKIIβ did not fully recover. A partial recovery of the fluorescence and the appearance of an unrecoverable fraction is an indication that binding is tight. An unrecoverable fraction of CaMKIIβ after actin stabilizing treatments, but not in control conditions, indicates CaMKIIβ binds stable F-actin more tightly than dynamic F-actin. GFP-CaMKIIα also recovered more slowly from stable F-actin; however, CaMKIIα reached full recovery by 500 s, whereas CaMKIIβ did not (see Fig. 3 B and D). These data indicate that CaMKIIβ bound strongly to stable F-actin, whereas CaMKIIα did not.

Ca2⁺/Calmodulin Binding and Enzymatic Activation Disrupted CaMKIIβ Binding to Stable F-actin.

To examine if the molecular events that regulate CaMKII activity also regulate CaMKIIβ interaction with stable F-actin, we immobilized actin with jasplakinolide and disrupted Ca2⁺/calmodulin binding (Ca2⁺/CaM). Ca2⁺/CaM binding to CaMKII precedes and is required for subsequent phosphorylation and kinase activity (14–19). However, in in vitro assays Ca2⁺/CaM binding competes with F-actin-CaMKIIβ interaction, resulting in the dissociation of CaMKIIβ from actin filaments (4, 10, 20). Neurons were treated with either KN93, an inhibitor of CaMKII that competes for Ca2⁺/CaM binding, or KN92, an inactive analog of KN93. KN93 treatment dissociated GFP-CaMKIIβ from stabilized actin filaments while KN92 had no effect (Fig. 4 A and B). The fraction of CaMKIIβ that did not recover from binding stable F-actin was reversed by KN93 treatment (see Fig. 4B). KN93 binds to the calmodulin binding site of CaMKIIβ and this binding may mimic Ca2⁺/CaM binding and also compete with F-actin binding (17). To selectively examine the role of Ca2⁺/CaM binding to CaMKIIβ-F-actin interaction, we imaged the recovery of a Ca2⁺/CaM-binding mutant, CaMKIIβ-A303R (Fig. 4 C and D). GFP-CaMKIIβ-A303R did not recover differently than WT CaMKIIβ from stabilized F-actin. These data indicate that lack of Ca2⁺/CaM binding to CaMKIIβ does not enhance or attenuate CaMKIIβ binding to F-actin. Together with the KN93 treatment results, these data suggest that binding of Ca2⁺/CaM to the enzyme decreased the binding of CaMKIIβ to F-actin. Like CaMKIIβ-A303R, a kinase inactive CaMKIIβ mutant, GFP-CaMKIIβ-K43R, did not recover differently from WT CaMKIIβ (see Fig. 4 C and D), indicating that CaMKIIβ binding to actin is not dependent upon catalytic activity. Furthermore, CaMKIIβ-A303R and CaMKIIβ-K43R localization to microspikes was not different from WT CaMKIIβ (see Fig. 1H).

All cells that expressed CaMKIIβ were also immunostained with phospho-CaMKII-specific antibodies. Furthermore, CaMKIIβ and phospho-CaMKII were remarkably colocalized in the microspikes, signifying that the CaMKIIβ in microspikes was phosphorylated at the key active residue T287 (Fig. 5 A and B). This indicates that the CaMKIIβ oligomeric enzyme was partially activated in basal conditions. To examine if CaMKIIβ binding to F-actin in microspikes is regulated by degree of oligomer activation, we treated the neurons with 50-mM KCl to induce calcium influx and robustly activate the CaMKIIβ oligomeric enzyme. Endogenous CaMKIIβ was partially phosphorylated in basal conditions, but after KCl stimulation, CaMKIIβ was highly phosphorylated (Fig. 5C). When in this elevated activation state, CaMKIIβ dissociated from the F-actin-rich microspikes to the cytoplasm (Fig. 5D). In fact, CaMKIIβ enrichment in microspikes (ratio of microspike to soma) was 2.283 +/− 0.164 under control conditions, but became 1.046 +/− 0.103 after KCl stimulation. This dissociation was rapid and occurred within 1 min of KCl stimulation (data not shown). To determine the degree to which activation of CaMKIIβ altered binding to actin, we examined the localization and binding dynamics of constitutively active CaMKIIβ, CaMKIIβ-T287D. GFP-CaMKIIβ-T287D localized with F-actin in microspikes but not highly, as did WT GFP-CaMKIIβ (see Fig. 1H and Fig. 5E). FRAP analyses of GFP-CaMKIIβ-T287D showed a significantly faster recovery from stable F-actin than WT CaMKIIβ (Fig. 5 F and G). Together, these data indicate that it is inactive CaMKIIβ that binds F-actin and suggests that, as the ratio of active to inactive subunits increases, CaMKIIβ binding to F-actin decreases.

Binding of CaMKIIβ to F-actin-Mediated Microspike Stability.

To establish if CaMKIIβ played a role in microspike formation or stability, we determined the prevalence of the microspikes in GFP-CaMKIIβ-transfected neurons. CaMKIIβ expression increased the percentage of cells that contained microspikes fourfold (Fig. 5H), whereas CaMKIIα did not (see Fig. 2G). In fact, coexpression of CaMKIIα with GFP-CaMKIIβ reversed the CaMKIIβ-induced increase in microspike-containing cells (see Fig. 2G). CaMKIIα also reduced CaMKIIβ-F-actin binding; therefore, we examined the data to see if CaMKIIβ promoted microspikes via binding to F-actin. CaMKIIβ mutants (β-A303R and β-K43R) that bound F-actin-like WT CaMKIIβ also increased microspike-containing cells fourfold (see Fig. 5H). Consistent with binding being the key event, when CaMKIIβ-F-actin binding was weakened, microspike-containing cells were not increased. CaMKIIβ-T287D bound F-actin more weakly than WT CaMKIIβ and fewer β-T287D transfected neurons contained microspikes (see Fig. 5 F and H). Activation and dissociation of endogenous CaMKIIβ from F-actin also decreased the number of neurons with microspikes (Fig. 5I). The kinetics of endogenous CaMKIIβ activation and dissociation from the microspikes were rapid after stimulation, precluding further clarification of the relationship between CaMKIIβ dissociation and microspikes. Together, these data suggest that it is the binding ability of CaMKIIβ to F-actin that is important for the formation and maintenance of microspikes.

Cell Culture.

Dissociated cortical cultures were established as in Redmond et al. (27). Briefly, cortical neurons from embryonic day 18 rats were plated onto polylysine and laminin-coated coverslips at 0.25 × 10⁶ cells per well for immunostaining and 3 × 10⁶ cells per 60-mm dish for biochemical assays, and maintained in Basal Medium Eagle, 5% FBS, 1% N₂ supplement, and 2-mM Glutamine. Cells were transfected by a modified calcium phosphate procedure at 3 DIV as in Redmond et al. (27).

Immunocytochemistry.

Cultures were fixed with 4% paraformaldehyde and 4% sucrose in PBS. Cultures were blocked in 3% BSA, 0.3% Triton X-100, 0.02% sodium azide in PBS, and immunostained for CaMKIIβ (1:250; monoclonal CBβ-1, Zymed), pCaMKII (1:500; Santa Cruz), CaMKII (1:500; monoclonal 6G9, Stressgen), MAP2 (1:2000; Sigma), GABA (1:2000; Sigma), and GFP (1:1000; Molecular Probes). F-actin was labeled with Alexa 488 or 568 conjugated phalloidin (1:25; Molecular Probes) to identify the microspike structures. Cell nuclei were stained with Hoechst. Images were acquired on a LSM510Meta Zeiss confocal microscope.

To quantify microspike/soma ratio and cell counts, images were acquired from two to four independent experiments using an Axiovert200 Zeiss fluorescent microscope with 40× (cell counts) or 63× (ratio) objective. Analyses were performed blind on GFP and phalloidin labeled cells using Improvision software. The mean intensity of GFP in three randomly chosen microspikes from one cell and the soma were measured. The ratio is (mean intensity in the microspike)/(mean intensity in the soma). The ratio value for the three microspikes was averaged to generate one value per cell for 18 to 62 cells per construct.

Molecular Cloning.

Cortices were isolated from P3 rats, total RNA prepared using TRIzol Reagent (Invitrogen), and cDNA synthesized with random primers and SuperScriptIII reverse transcriptase (Invitrogen). CaMKIIβ was subcloned into pEGFP-C3 (Clontech) with CaMKIIβ specific primers, 5′XhoI (AGACTCGAGATGGCCACCACGGTGACCTGCA) and 3′ EcoRI (ACGGAATTCTCACTGCAGTGGGGCCACTGGA). The sequence of CaMKIIβ cDNA was confirmed with Gene ID: 24245. CaMKIIβ mutant constructs were generated by PCR and subcloned into pEGFP. Additional details are provided in the SI Materials and Methods. CaMKIIα was supplied by Richard Maurer and subcloned into pEGFP-C3 vector to generate GFP-CaMKIIα.

Immunoblotting.

E18 cortical cultures were harvested at 4, 5, 8, 12, 14 and 18 DIV. Cortices from E18, P3, P7 and adult rats were homogenized with lysis buffer containing TBS, 1% Triton X-100, 10% glycerol, 1-mM PMSF, 1-μg/ml leupeptin, and 500-μM Na₃VO₄ on ice. Protein concentration was determined by BCA assay. Of the total protein, 10 μg from each DIV or cortex was loaded onto SDS/PAGE and transferred to nitrocellulose. Blots were incubated in blocking buffer as follows: 5% BSA in TBST (0.1% Triton X-100) at 4°C overnight (CaMKIIβ, 1:1000); 5% BSA in TBST at 25°C for 1 h (CaMKIIα, 1:1000; pCaMKII, 1:1000); 5% nonfat milk in TBST at 25°C for 1 h (GAPDH, 1:10,000), and probed with antibodies as indicated above in the respective blocking solution at 4°C overnight. HRP-conjugated secondary antibodies were diluted in blocking buffer and visualized by chemiluminescence (Pierce). E18 cortical neurons were stimulated with 50-mM KCl at 4 DIV for 20 min. Cells were lysed with lysis buffer, lysates incubated with protein A/G (Pierce) and immunoprecipitated with CaMKIIβ antibodies. Immunoprecipitates were analyzed by Western blots as above.

FRAP.

E18 cortical neurons were cultured on 35-mm glass bottom dishes (MatTek) at 0.75 × 10⁶ cells/dish and imaged at 5 DIV. Before imaging, media was replaced with Culture External Base containing 2-mM MgCl₂, 2-mM CaCl₂, 150-mM NaCl, 2.5-mM KCl, 10-mM glucose, and 10-mM NaHEPES. Cells were maintained in Culture External Base at room temperature during imaging for a maximum of 2 h. In the FRAP experiments, images were taken with 63× objective at 16× zoom. Images were captured every 1 s for 500 s. In Fig. 2, however, images were taken every 393.21 ms. A circular ROI for photobleaching was chosen with radius of 0.43 μm. Cells were photobleached after 10 frames of imaging with 8 iterations (10 iterations for GFP alone) of maximal excitation power. Background fluorescence, F_bkgd was determined in an unbleached area with similar initial fluorescence as the bleached ROI. F(t)_ROI was normalized to background fluorescence with an equation of F(t)_norm = (F(t)_ROI/F_bkgd. The first time-point after the bleach was set to t = 0. The normalized fluorescence of the frame immediately before photobleaching, F(−1) _norm, was set as 1. The fluorescence at other time points were normalized to F(−1)_norm to generate the final fluorescence value, F(t)_final = F(t)_norm/F(−1)_norm. Final fluorescence was plotted over time to generate the fluorescence recovery curve. When F(t)_final reached 1, it was considered to have completely recovered. The time-point at which fluorescence first reached 1 was determined as the full-recovery time. The half-time of the recovery (t_1/2) was determined by the time required to reach 50% of final recovered fluorescence, F_1/2 = (F(∞)−F (0))/2. The unrecoverable fraction (UF) represents the immobile fraction of molecules in the ROI and was determined by UF = (1−F(∞))/(1−F (0)). To stabilize actin filaments, cells were incubated with 10-μg/ml cytochalasin-D for 30 min or 1-μM jasplakinolide for 1 h before imaging. Then, 10-μM KN92 or KN93 were added with jasplakinolide 1 h before imaging. All data were analyzed by Microsoft Office Excel 2003 and Origin 6.1 software. Results are reported as mean ± SEM. Student's t test was used to determine statistical significance.