A transcription factor-binding domain of the coactivator CBP is essential for long-term memory and the expression of specific target genes

Abstract

Transcriptional activation is a key process required for long-term memory formation. Recently, the transcriptional coactivator CREB-binding protein (CBP) was shown to be critical for hippocampus-dependent long-term memory and hippocampal synaptic plasticity. As a coactivator with intrinsic histone acetyltransferase activity, CBP interacts with numerous transcription factors and contains multiple functional domains. Currently, it is not known which transcription factor-binding domain of CBP is essential for memory storage. Using mice that carry inactivating mutations in the CREB-binding (KIX) domain of the coactivator CBP (CBPKIX/KIX mice), we show that the KIX domain is required for long-term memory storage. These results are the first to identify an in vivo function for the KIX domain of CBP in the brain, and they suggest that KIX-interacting transcription factors recruit CBP histone acetyltransferase activity during long-term memory storage. One such KIX-interacting factor is the transcription factor CREB. Using quantitative real-time RT-PCR, we find that the expression of specific CREB target genes is reduced in the hippocampi of CBPKIX/KIX mice during memory consolidation. The recruitment of the transcriptional coactivator CBP via the KIX domain thus imparts target gene-dependent selectivity to CREB-driven transcriptional regulation, thereby activating genes required for the long-term storage of hippocampus-dependent memory.

Figure 1.

Schematic diagram of CREB-binding protein (CBP) showing different functional domains. (A) The endogenous allele of CBP generates a protein product with multiple domains, including several protein–protein interaction motifs as well as the histone acetyltransferase domain. (B) The CBP^KIX allele generates a protein product carrying three point mutations in the CREB-binding domain (KIX). These mutations block the interaction between phosphorylated CREB and CBP.

Figure 2.

CBP and CBP^KIX are expressed similarly in wild-type and CBP^KIX/KIX littermate mice, respectively. (A) A representative section from a wild-type mouse shows normal distribution of CBP protein in all brain regions examined (Hip, Str, Cx, and Amy shown in left column). DAPI staining is shown in the right column. (B) A representative section from a CBP^KIX/KIX homozygous knock-in mouse shows similar expression of CBP^KIX protein in the same brain regions as CBP in wild-type sections (Hip, Str, Cx, and Amy shown in left column). DAPI staining is shown in the right column. Hip, hippocampus; Str, striatum; Cx, cortex; Amy, amygdala.

Figure 3.

CBP^KIX/KIX mice exhibit impaired long-term memory for contextual fear conditioning but normal memory for cued fear conditioning. (A) Training and 24-h long-term memory test for contextual fear conditioning. CBP^KIX/KIX mice (n = 8) exhibited a significant decrease in freezing in a 24-h retention test, performed in the conditioned context, as compared with wild-type littermates (n = 8). (B) Training and 1-h short-term memory test for contextual fear conditioning. CBP^KIX/KIX mice (n = 6) and wild-type littermates (n = 6) showed similar levels of freezing in a 1-h retention test, performed in the conditioned context. (C) Training and 24-h long-term memory test for cued fear conditioning. No differences in freezing behavior were observed between CBP^KIX/KIX mice (n = 8) and wild-type littermates (n = 10) during training or the 24-h retention test. The 24-h test was performed in a novel context. Values are mean ±SEM. *P < 0.05.

Figure 4.

Impaired long-term memory in the novel object recognition task in CBP^KIX/KIX mice. (A) Wild-type mice (n = 8) show similar preference for a novel object in a 1-h retention test as compared with CBP^KIX/KIX (n = 6) mice. (B) In contrast, wild-type mice (n = 11) show significantly higher preference for a novel object in 24-h retention test than CBP^KIX/KIX (n = 10) mice. Values are mean ±SEM. *P < 0.05.

Figure 5.

Differential regulation of CREB-target genes in the hippocampus of CBP^KIX/KIX mice. (A) Samples for quantitative real-time RT-PCR (qRT-PCR) were prepared from the hippocampi of CBP^KIX/KIX mice (n = 4) and wild-type (n = 3) littermates that were subject to handling (home cage). qRT-PCR was performed in triplicate. Expression levels were normalized to the house keeping genes Actg1 and Tuba4. Levels of expression for CBP^KIX/KIX mice are presented as fold change from wild-type levels, which were set to 1.0. Only Fosb had significantly reduced expression in home cage samples. (B) Samples were prepared from CBP^KIX/KIX mice (n = 3) and wild-type (n = 3) littermates that were subject to contextual fear conditioning (fear cond). At 4 h after conditioning, mice were sacrificed, and samples were prepared for qRT-PCR. Levels of expression for CBP^KIX/KIX mice and wild-type littermates are presented as a fold change from wild-type (home cage) levels, which were set to 1.0. Dusp1, Fosb, Bdnf4, and Icer had significantly reduced expression. *P < 0.05.