βCaMKII plays a nonenzymatic role in hippocampal synaptic plasticity and learning by targeting αCaMKII to synapses

Abstract

The calcium/calmodulin-dependent kinase type II (CaMKII) holoenzyme of the forebrain predominantly consists of heteromeric complexes of the αCaMKII and βCaMKII isoforms. Yet, in contrast to αCaMKII, the role of βCaMKII in hippocampal synaptic plasticity and learning has not been investigated. Here, we compare two targeted Camk2b mouse mutants to study the role of βCaMKII in hippocampal function. Using a Camk2b(-/-) mutant, in which βCaMKII is absent, we show that both hippocampal-dependent learning and Schaffer collateral-CA1 long-term potentiation (LTP) are highly dependent upon the presence of βCaMKII. We further show that βCaMKII is required for proper targeting of αCaMKII to the synapse, indicating that βCaMKII regulates the distribution of αCaMKII between the synaptic pool and the adjacent dendritic shaft. In contrast, localization of αCaMKII, hippocampal synaptic plasticity and learning were unaffected in the Camk2b(A303R) mutant, in which the calcium/calmodulin-dependent activation of βCaMKII is prevented, while the F-actin binding and bundling property is preserved. This indicates that the calcium/calmodulin-dependent kinase activity of βCaMKII is fully dispensable for hippocampal learning, LTP, and targeting of αCaMKII, but implies a critical role for the F-actin binding and bundling properties of βCaMKII in synaptic function. Together, our data provide compelling support for a model of CaMKII function in which αCaMKII and βCaMKII act in concert, but with distinct functions, to regulate hippocampal synaptic plasticity and learning.

Figure 1.

Morphological and molecular analysis of the Camk2b^−/− mice. a, Immunocytochemistry analysis using αCaMKII- and βCaMKII-specific antibodies shows complete absence of βCaMKII in the hippocampus, with no apparent change in αCaMKII expression. b, Western blot analysis using αCaMKII- and βCaMKII-specific antibodies reveals no change in the levels of αCaMKII protein or in the levels of αCaMKII-T286 phosphorylation. In contrast, βCaMKII protein and βCaMKII-T287 phosphorylation are completely absent. c, Thionin staining shows no apparent morphological change in the hippocampus of Camk2b^−/− mice compared with wild-type mice. d, Quantification of Golgi analysis of the hippocampal pyramidal cells does not reveal any difference in spine density.

Figure 2.

Impaired αCaMKII targeting to dendritic spines in Camk2b^−/− mice. a, Representative images of cultured wild-type hippocampal neurons (top), and an enlargement of a dendritic segment (bottom), labeled with αCaMKII (green) and bassoon (red), and their colocalization (merged). b, Representative images of cultured Camk2b^−/− hippocampal neurons (top), and an enlargement of a dendritic segment (bottom), labeled with αCaMKII (green) and bassoon (red), and their colocalization (merged), showing reduced targeting of αCaMKII to synapses in Camk2b^−/− hippocampal neurons. c, Representative images of wild-type (top row) and Camk2b^−/− (bottom row) hippocampal dendritic segments, labeled with αCaMKII (green) and GKAP (red), and their colocalization (merged) (left) and labeled with αCaMKII (green) and F-actin (phalloidin staining) (red), and their colocalization (merged) (right), showing reduced targeting of αCaMKII to spines in Camk2b^−/− hippocampal neurons. d, Western blot shows comparable ratio of αCaMKII (α) to βCaMKII (β) in E18-derived wild-type cultured neurons on DIV14 compared with a lysate from an adult wild-type mouse.

Figure 3.

Hippocampal plasticity and learning in Camk2b^−/− mice. a, Camk2b^−/− mice show normal synaptic transmission (wild-type, n = 74; Camk2b^−/−, n = 63). b, Camk2b^−/− mice show impaired 100 Hz LTP (wild-type, n = 27; Camk2b^−/−, n = 21). c, Camk2b^−/− mice show normal 200 Hz LTP (wild-type, n = 14; Camk2b^−/−, n = 12). d, The 200 Hz LTP is dependent on αCaMKII, as Camk2a^−/− mice show impaired 200 Hz LTP (wild-type, n = 12; Camk2a^−/−, n = 18). e, Impaired contextual fear conditioning in Camk2b^−/− mice. Percentage of time spent freezing during training before the footshock (Pre) and 24 h after conditioning (Post) indicates reduced postshock freezing, while preshock freezing is normal (wild-type, n = 10; Camk2b^−/−, n = 10). Error bars represent SEM; asterisks indicate a significant difference between mutant and control.

Figure 4.

Generation of Camk2b^A303R mice. a, Schematic diagram for the generation of the Camk2b^A303R mutants. Genome, Wild-type Camk2b locus with the exons around Ala303 depicted as black boxes; Construct, targeting construct used for introducing the mutation in Ala303. The asterisk in exon 12 indicates the mutation in Ala303. The LoxP sites flanking the neomycin gene are depicted as triangles. The DTA was cloned in the construct for positive selection. Recombined, Mutant Camk2b^A303R locus after homologous recombination in ES cells; Floxed, mutant Camk2b^A303R locus after Cre recombination. b, Sequence of Ala303 in exon 12 showing the specific mutation made to induce Ala303Arg. c, Immunocytochemistry staining using an antibody specific for βCaMKII shows no difference in βCaMKII staining in the hippocampus of Camk2b^A303R mice (bottom) compared with wild-type mice (top). d, Western blot analysis using antibodies specific for αCaMKII and βCaMKII shows no difference in the levels of αCaMKII and βCaMKII in the hippocampus of the Camk2b^A303R mice. Western blot analysis using an antibody specific for detecting the phosphorylation levels of αCaMKII-T286 and βCaMKII-T287 reveals a slight reduction of Thr286 phosphorylation, whereas the Thr287 phosphorylation is completely absent in the hippocampus of the Camk2b^A303R mice. e, Thionin staining showed no gross morphological difference in the brains of Camk2b^A303R mice.

Figure 5.

Normal distribution of αCaMKII in cultured neurons of Camk2b^A303R mice. Representative images of cultured wild-type and Camk2b^A303R hippocampal neurons labeled with αCaMKII (green) and bassoon (red), and their colocalization (merged) showing normal distribution of αCaMKII in Camk2b^A303R hippocampal neurons. Inserts show an enlargement of a dendritic segment with two spines.

Figure 6.

Normal hippocampal synaptic plasticity and learning of Camk2b^A303R mice. a, Camk2b^A303R mice show normal synaptic transmission (wild-type, n = 68; Camk2b^A303R, n = 58). b, Camk2b^A303R mice show normal 100 Hz LTP (wild-type, n = 28; Camk2b^A303R, n = 19). c, Camk2b^A303R mice show normal contextual fear conditioning. Percentage of time spent freezing during training before the footshock (Pre) and 24 h after conditioning (Post), showing normal learning in Camk2b^A303R mice (wild-type, n = 14; Camk2b^A303R, n = 6). Error bars represent SEM.