Rubinstein-Taybi syndrome is inherited as an autosomal dominant trait. However, the occurrence is sporadic in the large majority of cases (~99%), with mutation occurring de novo in the family. In most families, the index case is the only member with the disease. However, cases of moderately affected relatives by somatic mosaicism have been reported [2,82,83,84] up to familial forms transmitted by one affected parent [2,4,85,86,87], confirming the clinical heterogeneity of the syndrome.

Historically, the location of the first gene involved in RSTS at 16p13.3 was identified by Imaizumi et al. in 1991 [88,89] and confirmed in 1992 by the works of Lacombe et al. [81] and Tommerup et al. [90]. Then, in 1995, Petrij et al. [12] identified this gene as CREBBP which encodes the cAMP response element-binding protein (CREB) binding protein. Initially, this protein was given this name because it was described as a partner of the CREB transcription factor [91]. Ten years later, mutations were identified in a CREBBP paralog gene, EP300 as an alternative cause of RSTS [13]. EP300 encodes the p300 protein that was originally described as a factor interacting with the EA1 protein of adenovirus type 5 [92,93].

The syndrome has been subdivided into type 1 associated with the CREBBP mutation spectrum (RSTS1; OMIM #180849) and type 2 associated with the EP300 mutation spectrum (RSTS2; OMIM #613684). The frequency of abnormalities in the responsible CREBBP gene is approximately 55–75% of cases [2,12,13,37,65,94,95,96,97,98,99]. To date, 500 pathogenic variants in this gene have been referenced as causing RSTS1 (55 of which are unpublished) based on the HGMDPro variant and LOVD databases (analyzed on 27 April 2021) [100,101] (Tables S1 and S3). The mutational spectrum includes 80.2% point mutations of which 55.2% are truncating mutations, 9.2% splicing mutations, and 16.8% missense mutations, 18.8% correspond to large rearrangements [100,101] (Figure 2A). There are no real hot spot mutations in CREBBP with a mutational spectrum distributed along the 31 exons. However, some recurrent mutations have been described and it is noted that about 52% of the reported missense mutations are located in the lysine acetyltranferase (KAT domain) [99]. An exception to this is the presence of an unstable region of CREBBP located between introns 1 and 2, characterized by a high frequency of repeated or palindromic sequences resulting in recurrent rearrangements in this region [37,102,103,104]. The presence of these heterozygous mutations or microrearrangements suggests a haploinsufficiency mechanism leading to the developmental abnormalities observed in the syndrome.

Abnormalities in the EP300 gene are responsible for about 8–11% of cases [2,13,17,18,19,20,21,22,44,99,105,106,107]. To date, 118 pathogenic variants in this gene have been referenced as causing RSTS2 (eight of which are unpublished) based on HGMDPro variant and LOVD databases (on 27 April 2021) [100,101] (Tables S2 and S3). The mutational spectrum includes 84.7% point mutations of which 69.5% are truncating mutations, 5.1% splicing mutations, and 13.6% missense mutations for 11.8% large rearrangements [100,101] (Figure 2B). Like CREBBP, there is no hot spot mutation in EP300, with only four pathogenic variants referenced more than twice in the databases: three in the catalytic domain and one in exon 2 [100,101]. In contrast, almost all of the predicted pathogenic missense mutations of EP300-associated RSTS are located in the KAT domain. Only three patients with RSTS have been reported in the literature with a missense mutation in EP300 out of KAT domain. However, each of these mutations was inherited from a healthy parent, making the pathogenic involvement of these variants difficult.

CREBBP and EP300 contain 31 exons and span approximately 155 kilobases (kb) and 87 kb, respectively [18,107,108]. The CBP (or KAT3A) and p300 (or KAT3B) proteins are paralogous transcriptional coactivators with intrinsic KAT activity. They are the only two members of the KAT3 family due to their low sequence homology to other acetyltransferases in the human genome [109] (Figure 3). They share a similar structure with different functional domains, including different types of protein–protein interaction motifs, and have high sequence identity (58% overall). On the N-terminal side, there is a nuclear receptor interaction domain (NRID or RID), that can bind to PXXP motifs. There are three cysteine-histidine-rich regions (C/H1 to C/H3) involved in protein–protein interactions. The C/H1 and C/H3 domains contain zinc finger transcriptional adapters (TAZ1 and TAZ2), and the C/H3 domain also contains a ZZ zinc finger domain and interacts with the EA1 oncoprotein. C/H2 contains a homeodomain (PHD). The catalytic domain has been highly conserved during evolution with 86% of similarity between CBP and p300. It includes a KAT domain and flanking regions (Bromodomain, C/H2, and C/H3 regions) [110,111]. Other non-catalytic domains show a high sequence similarity. The KIX domain allows the interaction of CREB, specifically at phosphorylated residue 133 (Ser133) of the CREB protein, with other transcription factors, and finally, a bromodomain (BD) links acetylated lysines [112]. On the C-terminal side of p300/CBP, there is an interferon-binding transactivation domain (IBiD), which contains a nuclear binding coactivator domain (NCBD) and a glutamine-rich domain, followed by a proline-containing PxP motif [113,114] (Figure 3).

These two proteins are able to interact with the basic transcription factors, TATA-binding protein (TBP) and TFIIB, and can form a complex with RNApolII [114,115,116]. These interactions allow KAT3 enzymes to play a crucial role in transcription initiation. They are also cofactors for oncoproteins, for viral protein processing (e.g., E1A), or for tumor suppressor proteins (e.g., p53 or BRCA1) [117]. Thus, CBP and p300 promote transcription in two ways: on one hand, they act as a bridge, linking transcription factors binding DNA to the transcription machinery, and on the other hand, by acetylation of histones, they create a chromatin environment favoring gene expression (Figure 4). This histone acetylation will play a key role in transcriptional activation by two distinct molecular mechanisms. Firstly, acetylation will partially neutralize the positive charge of histones, which will weaken their interaction with DNA. This will allow the opening of the chromatin structure and thus facilitate the access of transcription factors to their recognition elements [118]. Secondly, acetylated lysines can create a specific signal for regulatory factors or chromatin remodeling complexes to target a specific region. In contrast, histone deacetylation mediated by histone deacetylase (HDAC) will be associated with the repression of gene expression [119]. Apart from transcription, p300 and CBP act indirectly in other nuclear processes through their interaction with several proteins (which they often acetylate) involved in DNA replication and repair [120,121]. They have also been implicated in the regulation of cell cycle progression through interactions with cyclin E and cyclin-dependent kinase 260, as well as in intranuclear transport through acetylation of importin-α [122].