Enhanced cortico-amygdala efficacy and suppressed fear in absence of Rap1

Abstract

Auditory fear conditioning, a model for fear learning, is thought to be mediated by synaptic changes in the cortical and thalamic inputs to the lateral amygdala (LA); however, the specific roles of both pathways are still debated. Here, we report that a CaMKII-alpha-Cre-mediated knock-out (KO) of the rap1a and rap1b genes impaired synaptic plasticity and increased basal synaptic transmission in the cortical but not thalamic input to the LA via presynaptic changes: increases in glutamate release probability and the number of glutamate quanta released by a single action potential. Moreover, KO mice with alterations in the cortico-LA pathway had impaired fear learning, which could be rescued by training with a more aversive unconditional stimulus. These results suggest that Rap1-mediated suppression of synaptic transmission enables plasticity in the cortico-amygdala pathway, which is required for fear learning with a moderately aversive unconditional stimulus.

Figure 1.

Generation of mice with conditional KO of rap1a and rap1b genes. A, Targeting of rap1a. 1, Targeted locus. 2, Floxed rap1a allele after deletion of the TK-NEO cassette by transient expression of Cre recombinase. 3, The rap1a allele after deletion of coding exons 2 and 3. Rap1a encoding exons 2 and 3 (ex2, ex3), the TK-NEO cassette, and an 0.8 kb BamHI/KpnI fragment used as a probe are shown. BamHI fragments used for identification of mutated rap1a alleles on Southern blot are marked with the horizontal arrows. 4, Southern blot containing DNA from the ES clones after homologous recombination. +/+, WT clone with the WT alleles (14.4 kb); t/+, clone with a targeted rap1a allele (8.3 kb). 5, Southern blot after Cre-mediated excision. f/+, Clone with a floxed allele (10.1 kb); 0/+, clones with a KO allele (9 kb). B, Targeting of rap1b. 1, Targeted locus. 2, Floxed rap1b allele after deletion of the TK-NEO cassette by transient expression of Cre recombinase. 3, The rap1b allele after deletion of the coding exon 1 (ex1). The TK-NEO cassette and 0.5 kb HindIII/HindIII and 1.3 kb BglII/EcoRI fragments used as probes are shown. EcoRI and HindIII fragments used for identification of mutated rap1b alleles on Southern blot are marked with the horizontal arrows. 4, Southern blot containing DNA from the ES clones after homologous recombination. +/+, WT clone with WT alleles (10 kb); t/+, clone with a targeted rap1b allele (11 kb). 5, Southern blot after Cre-mediated excision. 0/+, Clone with the WT (0.5 kb) and KO allele (2.2 kb); +/+, WT contaminating clone; f/+, clone with a floxed allele (2.5 kb). Restriction endonuclease sites ApaI (A), BamHI (B), BglII (Bg), EcoRI (R), HindIII (H), KpnI (K), NcoI (N), NheI (Nh), SalI (S), and XhoI (X) are shown in parentheses when destroyed by cloning. LoxP sites are indicated by triangles. The regions represented with thick lines correspond to the targeting constructs.

Figure 2.

More efficient deletion of Rap1 genes in cortex compared with thalamus. A, In situ hybridization with antisense rap1a (left) and rap1b (right) probes. B, Comparison of excision efficiencies of the floxed rap1a (left) and rap1b (right) alleles by Cre-recombinase in the cortex (c) and thalamus (t) using semiquantitative PCR. From left to right, Representative agarose gels, length of amplification products in base pairs (bp) corresponding to the floxed and excised alleles, positions of PCR primers sequences (a1, a2, and a3 for rap1a and b1 and b2 for rap1b) on genomic loci, and relative amounts of the PCR products representing excised alleles. *p < 0.05. C, Left, Representative Western blot showing Rap1 total protein expression in the cortex and thalamus; the position of molecular weight standards (numbers indicate kilodaltons) are shown on the left. Arrows point to a 22 kDa Rap1 band (Rap1) and a nonspecific band (ns). Right, Quantification of KO efficiency in the cortex and thalamus by measuring a decrease in the intensity of the Rap1 protein bands from KO relative to WT samples run on same gels. *p < 0.05. D, Dissection scheme for the preparation of genomic DNA and protein extracts.

Figure 3.

Impaired STD-LTP is associated with elevated P_r in the cortical inputs to LA in Rap1 KO mice. WT (W) and KO (K) slices. A, STD-LTP protocol. B, LTP in cortical input to LA. Insets represent EPSP before (1) and after (2) induction. Calibration: 10 mV, 10 ms. C, PPR before and after LTP induction in cortical input to LA. **p < 0.01. D, CV of EPSP slopes before and after LTP induction in cortical input to LA. *p < 0.05. E, Plot of CV⁻² against percentage LTP in slices from WT mice (r = 0.647; n = 11). F, The progressive blockade of NMDA currents by MK-801 during cortical input stimulation. Each point represents averaged normalized currents from eight WT and seven KO neurons.

Figure 4.

Rap1 KO enhances the efficacy of cortico-LA synapses without altering their postsynaptic properties. A, Input–output curves in the cortical (C) and thalamic (T) inputs to LA. The averaged EPSP slopes are plotted as a function of stimulus intensity. B, Example of EPSCs evoked at holding potentials ranging from −70 to +50 mV at a step of 20 mV. Time points for determination of AMPAR and NMDAR currents are shown with dashed lines. C, D, Averaged current–voltage plots of the AMPAR (C) and NMDAR (D) currents in the WT (filled circles) and KO (open circles) slices. E, Cumulative distributions of the amplitudes of evoked asynchronous EPSCs from seven WT (solid line) and eight KO (dashed line) cells.

Figure 5.

Rap1 KO causes multiquantal release in the cortical input to LA. A, Examples of EPSCs evoked by minimal stimulation in the WT (W) and KO (K) neurons. Calibration: 20 pA, 5 ms. B, Averaged EPSC amplitude of successful responses to minimal stimulation. Diamonds show mean values from individual neurons, and boxes show averaged responses for each genotype. *p < 0.05. C, D, Representative traces of paired EPSCs at an interval of 50 ms before and after γ-DGG application in WT (W) and KO (K) slices. Calibration: 50 pA, 10 ms. E, Summary results showing an increase in the PPR after γ-DGG application in KO but not WT slices. *p < 0.05.

Figure 6.

Rap1 KO mice require more aversive US for fear learning. A, Conditioning with the 0.5 s US. Freezing before CS onset (pretraining), freezing during the first CS (CS1), and change in freezing (percentage CS freezing − percentage pre-CS freezing) at 3 and 48 h are shown. *p < 0.05. B, Rescue of fear learning with the 1 s US. Same abbreviations as in A. C, Pain thresholds. Thresholds of current intensities for flinching and running/jumping are shown. *p < 0.05.