B CBP andp300

CBP (CREB binding protein) and p300 were probably the most extensively characterized transcriptional coactivators even before their HAT activity was first identified (Goodman and Smolik, 2000). CBP was initially identified as a protein that selectively binds the phosphorylated active form of the CREB transcription factor (thus the name CBP) (Chrivia et al., 1993) while p300 was purified as a cellular protein targeted by the El A oncoprotein (Yee and Branton, 1985; Harlow et al, 1986). Despite being identified in different ways, CBP and p300 are closely related proteins and are often referred to as CBP/p300 (Arany et al, 1994). Although both p300 and CBP have potent histone acetyltransferase activity, their relation to other HATs is a distant one, having only very limited homology with the GNAT superfamily (Martinez-Balbas et al., 1998). CBP and p300 interact with a large number of other proteins through at least four protein-protein interaction domains and, similar to the GCN5 HAT, a bromodomain (Sterner and Berger, 2000). In fact, they possess the ability to act as physical scaffolds for the transcription machinery as another mechanism of activation in addition to their HAT activity. As such, they are essential cofactors for more than 40 different transcription factors (Kalkhoven, 2004).

Unlike GCN5 and PCAF, there is no evidence that p300 or CBP must stably exist in a large protein complex in order to function. However, CBP and p300 are regulated by several posttranslational modifications. CBP and p300 are both phosphoproteins and subject to regulation by several kinases. For example, CBP can be activated by calcium/calmodulin-dependent kinases in neuronal signaling and survival (Chawla et al., 1998; Hu et al, 1999). In addition, protein kinase A, cdk2, and p42/44 MAP kinase have also been reported to upregulate CBP (Kalkhoven, 2004; Ait-Si-Ali et al., 1998; Ait-Si-Ali et al, 1999) and p300 HAT activity while protein kinase C(8) has been shown to reduce it (Yuan et al, 2002). CBP and p300 are also regulated at the level of protein stability. They are subject to ubiquitination and degraded by the proteasome (Poizat et al, 2000; Lonard et al, 2000). p300 is progressively ubiquitinated and degraded during retinoic acid-induced differentiation (Iwao et al., 1999), suggesting that the ubiquitination of CBP and p300 may play a regulatory role during specific cellular conditions. Interestingly, degradation of CBP accompanies cell death triggered by the accumulation of mutant huntingtin, an expanded poly-glutamine protein (Jiang et al., 2003). CBP is also cleaved by caspases under apoptotic conditions in neurons (Rouaux et al., 2003), and p300 is subject to sumoylation, which converts p300 into a transcriptional repressor (Girdwood et al., 2003).

Given its functional interaction with a large set of important transcriptional factors, it is not surprising that ablation of CBP or p300 leads to embryonic lethality (Yao et al., 1998). Interestingly, mice with heterozygous mutations for both CBP and p300 also die as embryos, supporting the idea that CBP and p300 share certain functions during development (Yao et al., 1998). However, further genetic studies reveal that CBP and p300 are not identical and have distinct functions in vivo. For example, heterozygous mutation of CBP in humans is the cause of Rubinstein-Taybi syndrome (RTS), a genetic disorder characterized by mental retardation and skeletal malformations (Rouaux et al., 2004). A similar phenotype is seen in mice heterozygous for CBP mutations (Tanaka et al., 1997; Oike et al., 1999). Although mice with heterozygous mutations of p300 show reduced viability similar to the CBP heterozygous mutant mice, they do not develop an RTS-like syndrome (Yao et al., 1998).

In addition to developmental defects, both p300 and CBP appear to be involved in tumor formation. CBP heterozygous mutant mice are predisposed to hematological malignancy (Kung et al., 2000). In those tumors, the wild type CBP allele is lost, supporting a function of CBP as a tumor suppressor. Indeed, a recent study showed that inactivation of both CBP alleles by conditional knockout leads to T cell lymphoma in mice (Kang-Decker et al., 2004). In fact, human RTS patients are also prone to developing tumors (Goodman and Smolik, 2000). In contrast, heterozygous mutation of p300 did not predispose mice to tumor formation. However, rare p300 genetic alterations have been reported in solid tumors in humans (Muraoka et al., 1996; Gayther et al., 2000). Interestingly, in one gastric tumor, the wild type p300 allele is lost while the other allele harbors a point mutation that inactivates p300 HAT activity. These results suggest that HAT activity is important for the function of p300 and probably CBP as tumor suppressors.

Both CBP and p300 are also involved in hematological malignancy through aberrant chromosomal translocation. Chromosomal translocations of both CBP and p300 are common in acute myeloid leukemias

(AML) (reviewed in (Yang, 2004a)). Fusion proteins are formed with the MYST acetyltransferases MOZ and MORF, as well as with the methyltransferase MLL (mixed lineage leukemia). These fusions retain the HAT domain of CBP or p300 and often the HAT domain of MOZ and MORF as well. These fusions most likely lead to mistargeting and misregulation of HAT activity leading to aberrant gene expression that contributes to the leukemic phenotype. It is also worth noting both CBP and p300 appear to have a role in hematopoietic stem cells, though self-renewal requires the former and differentiation the latter (Rebel et al., 2002). Although the exact mechanism underlying the role of CBP and p300 in hematological malignancy remains to be established, it is clear that CBP and p300 can function both as tumor suppressors and oncogenes depending on the molecular and cellular context.

CBP is also implicated in the complex pathogenesis of neurodegenerative diseases. CBP loss of function has been seen in both models and patient samples from several such diseases, including Alzheimer's disease, Huntington's disease, amyotrophic lateral sclerosis, spinocerebellar ataxias, and other poly-glutamine (polyQ) diseases (Rouaux et al., 2003; Steffan et al., 2000; Takahashi et al., 2002; Hughes, 2002). PolyQ diseases typically involve aggregation of an expanded poly-glutamine protein into insoluble masses that can also sequester other native polyQ-containing proteins. CBP has a tract of 18 glutamines, and it has been shown to colocalize with several different polyQ aggregates (McCampbell et al., 2000; Yamada et al., 2001). In addition, a mutant huntingtin protein can also repress CBP and p300 HAT activity (Nucifora et al., 2001). CBP is also cleaved and degraded by caspase-6 in a cellular model of neurodegeneration (Rouaux et al., 2004). Strikingly, overexpression of CBP can reduce cell death in both culture models and a fly model of neurodegeneration (McCampbell et al., 2000; Taylor et al., 2003). This loss of CBP function can also be rescued by inhibition of HDACs in Drosophila and mice (see below). Therefore, the regulation of acetylation plays an important role in neurodegeneration, and this data suggests a new method of therapy for these diseases.

C: The MYST Proteins

MYST is an acronym from the four founding members of the HAT family: MOZ, Ybf2 (also called Sas3), Sas2, and TIP60 (Reifsnyder et al., 1996). Additional members include Esal, MOF, HBOl, and MORF (Sterner and Berger, 2000). While not as well-characterized as GCN5 or CBP, the MYST proteins have several intriguing properties. They all share a highly homologous MYST domain that is required for HAT activity. This domain contains a zinc finger (except Esal which has a zinc finger-like structure) (Yang, 2004a) and an acetyl co-A binding motif similar to the GNAT motif (Neuwald and Landsman, 1997). Despite the shared domains, MYST proteins are quite divergent in their association. Most of the MYST members are part of their own distinct protein complexes. Yeast Esal, for example, forms complexes called NuA3 and NuA4, which among other proteins contain TRRAP, a component of the SAGA complex (Allard el al., 1999; Grant et al., 1997). The Tip60 complex has been purified and also found to contain TRRAP and other shared proteins as well as unique components such as p400, a SWI2/SNF2 subunit (Ikura et al., 2000). Tip60 may also be able to translocate between the nucleus and cytoplasm, where it associates with the endothelin receptor (Lee et al., 2001).

Given the range of complexes, it is perhaps not surprising that the MYST proteins have divergent functions as well. Esal was first shown to be required for proper cell cycle progression, and later it was implicated in DNA repair as well (Clarke et al., 1999; Bird et al., 2002). The original MYST member Sas2 is involved in transcriptional silencing and opposes the action of Sir2 (see below). Sas2 helps establish the boundary between euchromatin and heterochromatin (Kimura et al., 2002; Suka et al., 2002). Interestingly, it seems to have opposite effects on silencing at different loci. Sas2 inhibits silencing at the HMR yeast mating type locus while promoting silencing at the HML locus (Ehrenhofer-Murray et al., 1997), indicating the complexity of silencing regulation. Sas3 was originally found to be involved in silencing as a weaker version of Sas2 (Sterner and Berger, 2000), however, it also regulates transcription elongation (John et al., 2000). MOZ and the related MORF appear to be linked to specific developmental processes since they act as coactivators for the Runxl and Runx2 transcription factors (Pelletier et al., 2002; Bristow and Shore, 2003). This agrees with data that reduction in levels of Querkopf, the murine MORF, leads to defects in osteogenesis and neurogenesis (Thomas et al., 2000). MOZ and MORF also contain transcription repression as well as activation domains, suggesting that they may have repressive activity in certain conditions (Champagne et al., 1999). In the same manner, the functional consequence of Tip60 activity appears to depend on the context. For example, it acts as a coactivator for NF-kB (Baek et al., 2002) and c-Myc (Frank et al., 2003), and as a corepressor of STAT3

(Xiao et al., 2003). Further, it has been implicated in DNA repair, specifically chromatin remodeling involving H2AX (Kusch et al., 2004).

MOF is a Drosophila MYST protein that plays a role in dosage compensation (Sterner and Berger, 2000). It provides a definitive picture of how histone acetylation can be important in a specific biological process. In flies, dosage compensation is achieved by doubling the expression of X-linked genes in males, rather than by female X-inactivation as is generally the case in humans. It was first noted that the X chromosomes of male flies were hyperacetylated compared to autosomes and female X chromosomes, and that this hyperacetylation occurred specifically on lysine 16 of histone H4 (Turner et al., 1992). Later, MOF (males absent on the first) was identified, and MOF mutant flies were characterized as lacking this acetylation on K-16 (Hilfiker et al., 1997). A complex called MSL (male-specific lethality) had already been observed to associate at the hyperacetylated sites (Bone et al., 1994), and MOF was shown to be a part of that complex (Smith et al., 2000). In vitro, MOF strongly acetylates H4, and a partially purified MSL complex specifically acetylated K-16 on H4 in vitro (Smith et al., 2000), thus completing the link between dosage compensation and MOF.

Dysregulation of MYST proteins is also implicated in disease pathways. MOZ and MORF are frequently involved in chromosomal rearrangements in leukemia. As discussed above, CBP and p300 form fusion proteins with MOZ and MORF, which usually retain the HAT domains of both proteins (Yang, 2004a). These fusions have been shown to leukemogenic (e.g. (Lavau et al., 2000)). MOZ has also been observed in two similar fusions with TIF2, a transcription-related protein (Carapeti et al., 1998; Carapeti et al., 1998). TIF2 can bind CBP, and it appears that the recruitment of CBP to the fusion protein is responsible for its oncogenic potential (Deguchi et al., 2003). MYST proteins can be involved in other disease processes as well. Tip60, which stands for Tat interacting protein of 60kD, specifically interacts with the HIV transactivator protein Tat (Kamine et al., 1996). Tat is able to inhibit the HAT activity of Tip60, and it is postulated that Tat thus blocks Tip60 function at gene targets, such as superoxide dismutase (Mn-SOD) (Creaven et al, 1999). MYST proteins are less characterized than PCAF, CBP, and p300; however, what is known about this HAT family suggests that they have many important biological functions.

A number of different HATs have been identified. An interesting question is why it was necessary for nature to evolve so many different types of HATs. Several specific HATs, most prominently p300 and CBP, appear only in higher organisms and are not present in yeast. One possibility is that the demand is much greater in higher eukaryotes for processing and integrating multiple signals for regulation of transcription. The fact that p300 and CBP can interact with a plethora of transcription factors have led to the proposal that they can function as transcriptional integrators that allow communication among many transcriptional networks. It is also possible that different HATs could have different substrate specificities that dictate their functions. For instance, in the context of the SAGA complex, GCN5 primarily acetylates histone H3 and not others (Sterner and Berger, 2000). Acetyltransferases can also target proteins other than histones (see below).

0 0

Post a comment