Differential Involvement of Kinase Activity of Ca²⁺/Calmodulin-Dependent Protein Kinase IIα in Hippocampus- and Amygdala-Dependent Memory Revealed by Kinase-Dead Knock-In Mouse

Representative immunoblots showing the CaMKIIα, CaMKIIβ and phospho-T286-CaMKIIα levels in the hippocampus and amygdala of the kinase-dead CaMKIIα (K42R)-KI mouse as compared to the wild-type mouse. The amounts of protein used were 2, 4, and 8 μg for the detection of CaMKIIα, CaMKIIβ, and phospho-T286-CaMKIIα (P-CaMKIIα), respectively, from hippocampal or amygdala homogenates. Autoradiography of duplicated samples from a pair of wild-type (WT) and kinase-dead CaMKIIα (K42R)-KI mice in the same experimental group used for quantitative immunoblot analyses are shown. See also Table 1.

Intact visually guided memory and impaired spatial memory in the Morris water maze tasks in the kinase-dead CaMKIIα (K42R)-KI mouse. A, Escape latencies in the visible platform trials, followed by the hidden platform trials in the water maze tasks in the kinase-dead CaMKIIα (K42R)-KI mouse as compared to the wild-type (WT) mouse. In the visible platform trials (left), both genotypes steadily decreased escape latencies and there was no significant difference between the genotypes. Both genotypes showed significantly shorter latency in the second and later blocks of trials than in the first block (labeled with gray vs black colors, p < 0.05, Bonferroni’s post hoc test). In the following hidden platform trials (right), latency differed significantly between the genotypes: CaMKIIα (K42R)-KI mice took longer time to reach the platform than wild-type mice did; *p < 0.05, **p < 0.01, ***p < 0.001 between the genotypes, Bonferroni’s post hoc test. Open circle, wild-type mice, n = 12; closed circle, CaMKIIα (K42R)-KI mice, n = 12. Note that the platform location was fixed for each mouse throughout the course of training. B, Probe trials performed on days 3 and 6 after training trials of the day (after 6 and 12 blocks of trials, respectively). Percentage time in each quadrant (left) and the number of center crossings of hypothetical platform locations (right) are shown; **p < 0.01, ***p < 0.001, ****p < 0.0001, Tukey’s post hoc test. Closed column, target quadrant; hatched column, left quadrant; open column, opposite quadrant; dotted column, right quadrant. C, Averaged swimming traces of each genotype for 60 s in the probe trials performed on days 3 and 6. The color indicates the average time spent at a certain location (1 × 1 cm square) per animal. The platform location was aligned in the right upper quadrant by rotation and indicated by a black-lined circle. D, Swimming speed measured for 60 s in the probe trials performed on days 3 and 6. There was no significant difference between the genotypes. Open column, wild-type mice; closed column, CaMKIIα (K42R)-KI mice.

No context discrimination, whereas partially preserved cued fear memory after fear conditioning in the kinase-dead CaMKIIα (K42R)-KI mouse. A, Fear conditioned memory after a single-stimulation protocol with a 30-s tone that coterminated with a 0.3-mA footshock (1CS-US). Contextual fear memory (Context) was tested in the same conditioning chamber 24 h after conditioning. Cued fear memory (Cued) was tested in a different context without (Pre-tone) or with tone (Tone) 48 h after conditioning. Ordinate indicates percentage freezing time for the period of 180 s, and the line in each column expresses the mean. Left, Wild-type (WT) mice showed significantly longer freezing time in the contextual chamber than in the cued chamber without tone, demonstrating context-dependent fear memory. In addition, they showed significantly longer freezing time in the cued chamber with tone than without tone, demonstrating tone-dependent fear memory; n = 15. Right, CaMKIIα (K42R)-KI mice showed virtually no freezing in the contextual chamber, as well as in the cued chamber without tone, and there was no significant difference in freezing time between the two conditions. On the other hand, they showed rather small, but significant freezing in the cued chamber with tone, indicating that cued fear memory was formed at least to a certain extent; n = 14. B, Fear conditioned memory after a stronger single-stimulation protocol with a 0.7-mA footshock (1CS-US-strong). Basically, similar results were observed as in A. WT, n = 16 (left). K42R, n = 14 (right). C, Fear conditioned memory after a repeated-stimulation protocol with three pairings of a tone and a 0.3-mA footshock (3CS-US). Left, Wild-type mice showed both context-dependent and tone-dependent freezing; n = 16. Right, CaMKIIα (K42R)-KI mice showed increased freezing in all of the three conditions tested, but they still did not show context-dependent freezing, whereas showed tone-dependent freezing, revealing context discrimination deficits; n = 15. D, Fear conditioned memory after a repeated stimulation protocol with five pairings of a tone and a 0.3-mA footshock (5CS-US). Left, Wild-type mice again showed both context-dependent and tone-dependent freezing; n = 15. Right, CaMKIIα (K42R)-KI mice showed generally increased freezing, and there was no significant difference in percentage freezing time between the three conditions tested, revealing severe context discrimination deficits, leading to generalized fear; n = 14; **p < 0.01, ***p < 0.001, ****p < 0.0001, Bonferroni’s post hoc test.

No fear response to tone alone without tone-footshock association, and long-lasting cued fear memory in the kinase-dead CaMKIIα (K42R)-KI mouse. A, Fear response after repeated exposure to tone alone for three times without footshocks (3CS alone). Both wild-type (WT; left) and CaMKIIα (K42R)-KI mice (right) showed almost no freezing in the three conditions tested, demonstrating the specificity of fear response to tone-footshock association in both genotypes (Fig. 4). WT, n = 13. K42R, n = 13. B, Long-term memory examined four weeks after strong 1CS-US conditioning. The mice in Figure 4B were retested four weeks later. Left, Wild-type mice still showed context-dependent and tone-dependent freezing, demonstrating the acquired fear memory was long-lasting; n = 16. Right, CaMKIIα (K42R)-KI mice also retained tone-dependent freezing, demonstrating the acquired fear memory was long-lasting; n = 14. C, Long-term memory examined four weeks after 3CS-US conditioning. The mice in Figure 4C were retested four weeks later. Here again, both wild-type (left) and CaMKIIα (K42R)-KI mice (right) retained once acquired fear memory. WT, n = 16. K42R, n = 15. D, Long-term memory examined four weeks after 5CS-US conditioning. The mice in Figure 4D were retested four weeks later. Left, Wild-type mice still showed tone-dependent freezing, but not context-dependent freezing; n = 15. Right, CaMKIIα (K42R)-KI mice revealed tone-dependent freezing this time, indicating that cued fear memory had been acquired from the beginning after 5CS-US conditioning, but had been masked due to generalized fear when tested soon after conditioning (Fig. 4D, right). However, when generally increased fear subsided in four weeks, tone-dependent freezing became apparent; n = 14; **p < 0.01, ***p < 0.001, ****p < 0.0001, Bonferroni’s post hoc test.