Compensatory increases of atypical PKCι and conventional, but not novel PKCs, in conditional PKMζ-KO mouse hippocampus.

(A, B) Immunoblots of hippocampal extracts from Camk2a-CreERT2Prkczfl/flmice receiving tamoxifen (2 mg/200 µl i.p., 5 daily doses) to activate Cre recombinase selectively in excitatory neurons. Mice are sacrificed 7 days after the last dose. Left, representative immunoblots with Mr markers. Right, mean ± SEM. Significance by two sample Student t tests with Bonferroni correction denoted by *; not significant, n.s.; statistics in Figure 1 — table supplement 1A,B. Tamoxifen is a partial PKC antagonist and may still be present after a week (O’Brian et al., 1985); therefore, WT mice that also received tamoxifen are non-transgenic controls (NTC). (A) PKMζ decreases and PKCι increases in PKMζ-cKO mice. (B) Conventional PKCs increase and novel PKCs do not change. Actin loading controls shown in Figure 1 — figure supplement 2.

Compensatory increases of PKCι during spatial memory in conditional PKMζ-KO mouse hippocampus.

(A) Left, schematic of active place avoidance training apparatus with a slowly rotating arena containing a nonrotating shock zone sector (shown in red). Visual cues located on the walls of the room are needed to avoid the shock zone. Right, experimental protocol. PKMζ is genetically ablated in Camk2a-CreERT2Prkczfl/fl mice. Cre is activated using 4-OH tamoxifen (OH-TAM, 2 mg/200 µl i.p., 3 doses every other day). Control mice receive vehicle injections. Active place avoidance training begins 3 weeks later, and 1 week after training memory retention is tested in the absence of shock followed by sacrifice and immunohistochemistry. (B) Immunohistochemistry shows ζ-cKO reduces PKMζ and increases PKCι in CA1 st. pyramidale (p) and radiatum (r), but not lacunosum-moleculare (lm) 1 week after training. Left above, PKMζ expression decreases in cell bodies and dendritic compartments of the PKMζ-cKO. Left below, PKCι expression increases in cell bodies as well as in dendritic compartments where it is ordinarily expressed at low levels. DAPI staining of nuclei shown in blue. Bar = 50 µm. Right, mean ± SEM. Student t tests with Bonferroni corrections compared differences in PKMζ and PKCι expression separately in strata of CA1 (Figure 2 — table supplement 1). (C) Compensatory spatial memory in ζ-cKO. Left, representative paths during first 10-min of pretraining, training trial 3, and 1-day memory retention. Right, mean ± SEM. Two-way ANOVA (treatment X training) revealed a significant effect of training (F1.662, 9.972 = 41.93, P < 0.0001), but not an effect of treatment (F1, 6 < 0.001, P = 1.0) or their interaction (F2, 12 = 0.48, P = 0.6). Further comparisons using Bonferroni-corrected tests revealed significant differences between pretraining and training (Trial 3) (P = 0.002) and pretraining and retention (P = 0.001), but no differences between training (Trial 3) and retention (P = 0.1). Further comparison revealed significant differences between pretraining and training (Trial 3) in both vehicle and 4-OH tamoxifen groups (P = 0.02 and P = 0.01, respectively) and between pretraining and retention in both vehicle and 4-OH tamoxifen groups (P = 0.03 and P = 0.05, respectively), confirming the treatment groups did not behave differently.

Compensatory increases of PKCι during hippocampal late-LTP maintenance in Prkcifl/fl-Prkcz-/-mice.

(A) Left, schematic of experimental protocol shows AAV expressing Cre by cytomegalovirus (CMV) promoter injected into ipsilateral hippocampus of a Prkcifl/fl-Prkcz-/-mouse, and control AAV expressing eGFP injected into contralateral hippocampus. Hippocampal slices are prepared 3 weeks later. Right, representative images of PKCι-immunohistochemistry from adjacent slices in AAV-eGFP-injected (ζ-KO) hippocampus show PKCι persistently increases 3 h post-tetanization (top row), and low, unchanging levels of PKCι in the AAV-Cre-injected (ι/ζ-dKO) hippocampus (bottom row). White boxes show st. radiatum regions of interest. (B) Mean ± SEM. The two-way ANOVA reveals the main effects of treatment (AAV-Cre [ι/ζ-dKO] vs. AAV-eGFP [ζ-KO], F1,68 = 83.58, P < 0.00001, η2p = 0.55), and stimulation (HFS vs. test, F1,68 = 20.47, P = 0.00003, η2p = 0.23), and an interaction of treatment X stimulation (F1,68 = 9.09, P = 0.004, η2p = 0.12). Post-hoc analysis confirms that, compared to the ζ-KO control group, the intensity of PKCι immunoreactivity was significantly decreased in ι/ζ-dKO (P’s < 0.002 for both control and LTP in ι/ζ-dKO), and increased in ζ-KO after HFS (P = 0.00011, ζ-KO, n’s = 16, ι/ζ-dKO, n’s = 20). Intensity of PKCι immunoreactivity did not change in the ι/ζ-dKO between the control and HFS groups (P = 0.3). Bar = 100 µm.

ι/ζ-dKO hippocampus shows transient LTP, but not persistent LTP.

(A) Late-LTP is absent in ι/ζ-dKO hippocampus. Above left inset, schematic of intrahippocampal injections of AAC-Cre recombinase and AAV-eGFP. Middle inset, representative fEPSPs correspond to numbered times in time-course below. Below, filled red circles, AAV expressing Cre and HFS with 2 tetanic trains; open red circles, test stimulation of a second synaptic pathway within the hippocampal slice. HFS tetani shown at arrows. Open black circles, AAV expressing eGFP by CMV promoter with HFS; open grey circles, with test stimulation. Three-way mixed-design ANOVA reveals main effects of treatment (hippocampal injections of AAV-Cre [ι/ζ-dKO] vs. AAV-eGFP [ζ-KO], F1,20 = 8.45, P = 0.0009, η2p = 0.30), and stimulation (HFS vs. test stimulation, F1,20 = 5.90, P = 0.025, η2p = 0.23), as well as a 3-way interaction among treatment X stimulation X time (5-min average of pre-HFS and 3-h post-HFS, F1,20 = 12.68, P = 0.002, η2p = 0.39). Post-hoc analysis confirms established LTP is not maintained in ι/ζ-dKO 3 h after HFS when compared to pre-HFS basal responses (P = 0.7). Post-hoc analysis also confirms the control hippocampus maintains established LTP (P = 0.0002). Test stimulation was unaffected by AAV-Cre or AAV-eGFP injections (P = 0.9 and P = 0.7, respectively, n’s = 6). Right inset, ι/ζ-dKO by CaMKIIα promoter expression of Cre eliminates late-LTP. Three-way mixed-designed ANOVA reveals interaction between treatment (ζ-KO vs. ι/ζ-dKO) and stimulation (HFS vs. test stimulation, F1,14 = 6.62, P = 0.02, η2p = 0.32), and a 3-way interaction among treatment, stimulation, and time (5 min pre-HFS and 3 h post-HFS, F1,14 = 8.56, P = 0.01, η2p = 0.38). Post-hoc analysis confirms that compared to pre-HFS basal responses, LTP is not maintained in ι/ζ-dKO hippocampus 3 h post-HFS (P = 0.8) and is maintained in the control hippocampus (P = 0.003). Test stimulation was unaffected by AAV-Cre or AAV-eGFP injections (P = 0.4 and P = 0.9, respectively). ι/ζ-dKO HFS, n = 5; ι/ζ-dKO test, n = 4; ζ-KO HFS, n = 5; ζ-KO test, n = 4. (B) LTP does not persist in ι/ζ-dKO mice after stronger afferent stimulation with 4 tetanic trains. ANOVA with repeated measurements reveals main effects of time (5 min pre-HFS, 20 min post-HFS, and 3 h post-HFS, F2,14 = 20.51, P < 0.0001, η2p = 0.75). Post-hoc analysis confirms that early-LTP is established in both ι/ζ-dKO and control groups (5 min pre-HFS vs. 20 min post-HFS, P = 0.005 and 0.002, respectively), and no difference between these two groups at 20 min post-HFS (P = 0.6). However, LTP in ι/ζ-dKO did not persist 3 h (5 min pre-HFS vs. 3 h post-HFS, P = 0.4), whereas LTP is intact in control. (P = 0.008). ι/ζ-dKO, n = 5; control, n = 4.

Impaired long-term memory and intact short-term memory for spatial information in mice with bilateral hippocampal ι/ζ-dKO.

(A) Experimental protocol. Prkcifl/fl-Prkcz-/- mice are injected bilaterally in hippocampus with AAV-Cre (ι/ζ-dKO) or AAV-eGFP (ζ-KO, control), and 3 weeks later they received pretraining and, after 1 day, a single 30 min trial repeated daily for a total of 3 trials. Long-term retention is tested without shock 1 day after the last training trial. (B) ι/ζ-dKO does not affect short-term memory in the first training trial. Left, ι/ζ-dKO does not affect short-term memory as assessed by the time to enter the shock zone for the first 8 times (all animals had up to at least 8 entries in trial 1). ANOVA with repeated measurements finds the main effects of training (pretraining and trial 1, F1,28 = 35.19, P < 0.00001, η2p = 0.56) indicating trial 1 learning, time to entry (the 1st to 8th entry within a trial, F7,196 = 145.80, P < 0.00001, η2p = 0.84), and their interaction (F7,196 = 37.68, P < 0.00001, η2p = 0.57). However, there is no group effect (AAV-eGFP- and AAV-Cre-injected, F1,28 = 0.19, P = 0.7, η2p = 0.007) nor interaction with either training or time to each entry (F’s < 0.66, P’s> 0.60, η2p’s < 0.02). Right, ι/ζ-dKO does not affect short-term memory as assessed by maximum avoidance time during the first training trial. The contrast analysis reveals that the increases of maximum avoidance time from pretraining to trial 1 are not different between AAV-eGFP-injected and AAV-Cre-injected groups (t14 = 1.91, P = 0.08, d = 1.91). Paired t-tests reveal trial 1 is greater than pretraining in each genotype (t’s > 3.10, P’s < 0.018, Cohen’s d’s > 1.62), indicating both groups of mice successfully established short-term memory. In contrast, the improvement of maximum avoidance time from trial 1 to trial 3 are different between the groups (t14 = 2.93, P = 0.01, d = 2.88), suggesting the two groups performed differently between daily training sessions when between-day memory influences avoidance. In addition, the ANOVA with repeated measurement discovers no group effect (AAV-eGFP-injected vs. AAV-Cre-injected, F1,14 = 0.56, P = 0.47, η2p = 0.04), but significant effects of trial (F3,42 = 30.37, P < 0.0001, η2p = 0.68), and interaction (F3,42 = 2.93, P = 0.04, η2p = 0.17). Post-hoc tests confirm that the maximum avoidance time in trial 1 is not different between the two groups (P = 0.14). The AAV-eGFP-injected group improved their performance between trial 1 and trial 3 (P = 0.0002), whereas the AAV-Cre showed no improvement (P = 0.2; n’s = 8). These data indicate no differences in short-term memory between AAV-eGFP- and AAV-Cre-injected groups, but only the AAV-Cre-injected group failed to improve between daily trials, suggesting inability to retain avoidance memory across days. (C) PKCι gene ablation impairs long-term memory in Prkcifl/fl-Prkcz-/-mice. Left, representative paths during 10-min of pretraining, at end of training trial 3, and 1-day memory retention. Right, mean ± SEM. The ANOVA with repeated measurement finds main effects of group (AAV-eGFP vs. AAV-Cre, F1,14 = 10.53, P = 0.006, η2p = 0.43) and training phase (pretraining, trial 3 of training, retention, F2,28 = 7.65, P = 0.002, η2p = 0.35). Post-hoc analysis reveals that the mice with AAV-Cre-injected ι/ζ-dKO hippocampus perform poorer during the memory retention test, compared to AAV-eGFP-injected littermates (P = 0.02). The mice with ι/ζ-dKO hippocampus show no difference between the memory retention test and pretraining trial (P = 0.9), whereas the AAV-eGFP-injected mice show long-term memory is maintained (P = 0.02; n’s = 8). In addition, pretraining vs. training trial 3 was significantly different in ζ-KO (P = 0.006), but not in ι/ζ-dKO (P = 0.4).

ζ-cKO increases activation-loop-phosphorylation state of atypical PKCι and conventional PKCs, but not novel PKCε or CaMKIIα autophosphorylation.

(A) Left, above, representative immunoblots show increases in activation-loop phosphorylated PKCι (p-PKCι) and conventional-PKCs (p-cPKC) in PKMζ-cKO mice, compared to non-transgenic controls (NTC). Red, phospho-PKCs; green, total PKCs from the same samples. Mr’s shown at right. Below, p-PKCε, recognized by its higher Mr, does not change. Right, mean ± SEM (B) Levels of total CaMKIIα and T286-autophosphorylated CaMKIIα do not change in ζ-cKO hippocampus. Statistics in Figure 1 — table supplement 2.

Actin loading controls for immunoblots shown in Figure 1B.

The PKMζ and actin lanes from columns 3 and 4 are from 3 adjacent lanes shown in Fig. 1A.

Data used in the statistical analysis of experiments in Figure 4 and Figure 4 — figure supplement 2B.

(A) Figure 4A, (B) Figure 4A, insert, (C) Figure 4 — figure supplement 2B, (D) Figure 4B.

Effects of AAV expression of Cre-recombinase by the CaMKIIα promoter in Prkcifl/fl-Prkcz-/-and Prkcifl/fl-Prkcz+/+ mice.

(A) Representative immunohistochemistry of AAV expressing Cre-recombinase by CaMKIIα promoter in Prkcifl/fl-Prkcz-/-mice shows loss of PKCι in ι/ζ-cKO hippocampus and compensatory increase in PKCι during LTP maintenance in control eGFP-injected hippocampus. Left, schematic of sites of injection; right, PKCι immunohistochemistry. DAPI stains nuclei in CA1. Bar = 100 µm. (B) ι-cKO (AAV with CaMKIIα promoter expressing Cre-recombinase in hippocampus of Prkcifl/fl-Prkcz+/+ mice) shows compensated LTP, as previously described (Sheng et al., 2017). Loss of early-LTP is compensated in the PKCι-cKO as in earlier reports. The repeated measurement ANOVA reveals the main effect of LTP (5-min average of pre-HFS, 20-min, and 2-h post-HFS, F2,6 = 14.03, P = 0.005, η2p = 0.82) in ι-cKO mice. Post-hoc tests confirm that LTP was established at 20 min post-tetanization (5-min pre-HFS vs. 20-min post-HFS, P = 0.006) and maintained for 2 h (5-min pre-HFS vs. 120-min post-HFS, P = 0.009); n = 4.

Statistics for data presented in (A) Figure 1A and (B) Figure 1B.

Significant differences with Bonferroni correction are in bold.

Statistics for data presented in (A) Figure 1 — figure supplement 1A and (B) Figure 1 — figure supplement 1B.

Significant differences with Bonferroni correction are in bold.

Statistics for data presented in Figure 2B.

Significant differences with Bonferroni correction are in bold.