260
R. E. HARRINGTON et al.
The decameric elements may be separated by as much as 21 bp without complete loss of p53 binding affinity (Waterman et aI., 1995), but functional sites,
defined as those able to activate transcriptionally a nearby reporter gene, evidently follow very closely the consensus decamer pattern with no or only very
short intervening spacers (Tokino et aI., 1994). Furthermore, the sequences of the
tetrameric elements that span the pseudo dyad in each half site assume unusual
importance in determining the binding properties of p53 (Durell et aI., 1998;
Nagaich et aI., 1998). These elements are most commonly CATG, but may also
include CAAG (Tokino et aI., 1994). It is well known that these sequences, especially the former, exhibit unusual flexibility for bending or kinking into the major
groove (McNamara et aI., 1990; Zhurkin et aI., 1991; Nagaich et aI., 1994; El Hassan and Calladine, 1998; Olson et aI., 1998). Since many architectural proteins
utilize PyPu sequence elements (Steitz, 1990; Werner et aI., 1996; Bewley et aI.,
1998), it is natural to suspect that these elements play an important role in specific DNA recognition by transcription factors, especially those with high functional multiplicities such as p53. Furthermore, the many regulatory roles for p53
and its large variety of binding sites suggest that its specificity of binding to individual DNA binding sites provides a clue about p53 function in its interactions
with other regulatory proteins.
The tetrameric nature of the p53 nucleoprotein complex is one of the aspects
of this system that is unique among known transcription factors. It is of special
interest that the p53DBD peptide alone has been found to self-assemble with high
cooperativity as a tetrameric complex when bound to a full 20 bp p53 response
element, although it exists as a monomer in the absence of cognate DNA (Balagurumoorthy et al., 1995). Thus, although the tetramerization 'domain may in part
mediate tetramerization of the wild type protein, this cannot be its sole function,
and the full role of the tetramerization domain in p53 function remains somewhat of a mystery at the present time.
2.2
Relationship of a Co-Crystal Structure of a p53 Nucleoprotein Complex
to Solution Studies
It is clear that the relationships of structure to function in p53 are extremely
important. However, at the present time, only a single crystallographic structure
of a p53 nucleoprotein complex has been reported (Cho et aI., 1994). This cocrystal structure clearly showed direct interactions between p53DBD peptides
and a cognate DNA, but since only a single p53DBD peptide was specifically
bound in the asymmetric unit, it could not directly address the role of tetramerization in specific DNA recognition. In addition, it did not predict bending of the
response element DNA in the complex, which was subsequently demonstrated
using cyclization methods (Balagurumoorthy et aI., 1995) and later confirmed by
both cyclic permutation (Nagaich et al., 1997a) and by phase sensitive detection
(Nagaich et al., 1999). These limitations suggested that an advanced model for
the complex, which would encompass these additional features, was required for
a more complete understanding of structure-function relationships in the complex.
R. E. HARRINGTON et al.
The decameric elements may be separated by as much as 21 bp without complete loss of p53 binding affinity (Waterman et aI., 1995), but functional sites,
defined as those able to activate transcriptionally a nearby reporter gene, evidently follow very closely the consensus decamer pattern with no or only very
short intervening spacers (Tokino et aI., 1994). Furthermore, the sequences of the
tetrameric elements that span the pseudo dyad in each half site assume unusual
importance in determining the binding properties of p53 (Durell et aI., 1998;
Nagaich et aI., 1998). These elements are most commonly CATG, but may also
include CAAG (Tokino et aI., 1994). It is well known that these sequences, especially the former, exhibit unusual flexibility for bending or kinking into the major
groove (McNamara et aI., 1990; Zhurkin et aI., 1991; Nagaich et aI., 1994; El Hassan and Calladine, 1998; Olson et aI., 1998). Since many architectural proteins
utilize PyPu sequence elements (Steitz, 1990; Werner et aI., 1996; Bewley et aI.,
1998), it is natural to suspect that these elements play an important role in specific DNA recognition by transcription factors, especially those with high functional multiplicities such as p53. Furthermore, the many regulatory roles for p53
and its large variety of binding sites suggest that its specificity of binding to individual DNA binding sites provides a clue about p53 function in its interactions
with other regulatory proteins.
The tetrameric nature of the p53 nucleoprotein complex is one of the aspects
of this system that is unique among known transcription factors. It is of special
interest that the p53DBD peptide alone has been found to self-assemble with high
cooperativity as a tetrameric complex when bound to a full 20 bp p53 response
element, although it exists as a monomer in the absence of cognate DNA (Balagurumoorthy et al., 1995). Thus, although the tetramerization 'domain may in part
mediate tetramerization of the wild type protein, this cannot be its sole function,
and the full role of the tetramerization domain in p53 function remains somewhat of a mystery at the present time.
2.2
Relationship of a Co-Crystal Structure of a p53 Nucleoprotein Complex
to Solution Studies
It is clear that the relationships of structure to function in p53 are extremely
important. However, at the present time, only a single crystallographic structure
of a p53 nucleoprotein complex has been reported (Cho et aI., 1994). This cocrystal structure clearly showed direct interactions between p53DBD peptides
and a cognate DNA, but since only a single p53DBD peptide was specifically
bound in the asymmetric unit, it could not directly address the role of tetramerization in specific DNA recognition. In addition, it did not predict bending of the
response element DNA in the complex, which was subsequently demonstrated
using cyclization methods (Balagurumoorthy et aI., 1995) and later confirmed by
both cyclic permutation (Nagaich et al., 1997a) and by phase sensitive detection
(Nagaich et al., 1999). These limitations suggested that an advanced model for
the complex, which would encompass these additional features, was required for
a more complete understanding of structure-function relationships in the complex.