These data indicate that in the
These data indicate that, in the context of the N-terminal 42 amino acids of p53, threonine 18 is a substrate for phosphorylation by CK1 as mediated by prior phosphorylation of serine 15. To determine whether this is a property of full length p53, GST-p53 fusion proteins comprising the full length p53 protein, either WT, 15A or 15A/37A, were mock-phosphorylated or phosphorylated with DNA-PK using unlabelled ATP as phosphate donor. The data (Fig. 3) show that, in the absence of prior phosphorylation by DNA-PK, all three proteins were phosphorylated by CK1 to an equal extent (lanes 5–8). In contrast, the phosphorylation of WT p53 was stimulated by up to 50-fold when the protein had previously been phosphorylated by DNA-PK (lane 2). When the 15A or 15A/37A mutants were incubated with DNA-PK and unlabelled ATP prior to CK1 phosphorylation (lanes 3 and 4), the stimulation was considerably less effective than with the WT protein, although the levels of phosphorylation were slightly higher than those in the absence of any prior phosphorylation. As before, no phosphate was incorporated into the GST moiety. These data indicate that phosphorylation of full length p53 is stimulated by prior phosphorylation of the protein by DNA-PK. Fig. 3B shows that when the p53 had been phosphorylated by DNA-PK and subsequently by CK1, the binding to HDM2 in vitro was reduced by about 3-fold as compared with the unphosphorylated protein. Phosphorylation by DNA-PK alone or CK1 alone did not affect HDM2 binding. These data confirm the potential of threonine 18 modification for regulating the interaction of p53 with HDM2 , .
Discussion In this paper, we have explored the phosphorylation of human and murine p53 by the protein kinase CK1 in vitro. The most striking observation which is evident from our data is that prior phosphorylation of serine 15 in vitro can create a potent recognition determinant for subsequent phosphorylation by CK1 (this applies to both human and murine p53). Mutagenesis data and phosphoamino Nicotine analysis indicate that the residue phosphorylated by CK1 is threonine 18 (Fig. 2). Previous studies have established the importance of phosphorylated residues in acting as recognition determinants for phosphorylation by CK1 ,  and our data therefore provide a potential mechanism by which serine 15 phosphorylation may nucleate threonine 18 phosphorylation in the cell. This would be consistent with an established role of serine 15 phosphorylation orchestrating sequential modification of other residues in p53 . Such a role would be dependent on DNA damage-induced modification of serine 15 (which is mediated in vivo by ATM , ) and would also be consistent with the indication from yeast studies that CK1 activity is pivotal to the DNA damage response , . Threonine 18 lies within a highly conserved element of the p53 protein (Fig. 4) which mediates p53 activation, regulation of the association with MDM2 and control of the interaction with transcriptional activators , , , , , , . Moreover, recent evidence indicates that threonine 18 phosphorylation contributes significantly to abrogating the p53-MDM2 interaction and is therefore likely to be a significant factor in the p53 activation process , . Our data confirm that threonine 18 phosphorylation can partially block interaction of p53 with HDM2 (Fig. 3B). Threonine 18 modification may also be an important factor in the p53 response to tumour development . The identification of CK1 as a potential threonine 18 kinase therefore opens up a new avenue of exploration to analyse the physiological modification of this important regulatory site. While murine p53 is also a substrate for CK1 in the absence of serine 15 phosphorylation in vitro (at residues 4, 6 and 9 ), loss of the equivalent residues in human p53 does not affect the ability of CK1 to phosphorylate the protein (Fig. 1B). However, serine 20 (and possibly serine 37) may act as phosphate acceptor in the human protein in vitro. Comparison of the sequences of the first 42 amino acids of the murine and human proteins (Fig. 4) reveals differences which could account for these observations. For example, proline residues at positions 4 and 8 in the human protein may alter the secondary structure such that the kinase does not efficiently interact with p53. Similarly, a glutamic acid residue at position 17 (in both murine and human p53) may act as a recognition determinant for phosphorylation of serine 20 (we do not know whether serine 20 of murine p53 can also be phosphorylated by CK1). These differences may reflect genuine variance in the way in which p53 proteins from different species are regulated and add to a growing list of species-specific modifications of p53 (discussed by Meek ). Alternatively, when compared to the ability of CK1 to phosphorylate threonine 18 in p53-15P, which is many orders of magnitude greater than unphosphorylated p53, the physiological relevance of phosphorylation of these residues by CK1 may be questionable. This is not to say that the residues themselves are not of regulatory significance, only that the protein kinase(s) which modify these sites in vivo may well be different from CK1. A final issue concerns the isoform(s) of CK1 which can modify p53 in the cell. While previous data have indicated that the delta and epsilon isoforms are preferentially active towards murine p53 (at residues 4, 6 and 9 ), the use of a phosphorylated residue as a recognition determinant is not the property of any specific isoform(s) of CK1. It is therefore possible that other CK1 isoforms may modify threonine 18 in vivo. Future studies should resolve these issues.