Calcium-dependent Interaction of S100B with the C-terminal Domain of the Tumor Suppressor p53*

In vitro, the S100B protein interacts with baculovirus recombinant p53 protein and protects p53 from thermal denaturation. This effect is isoform-specific and is not observed with S100A1, S100A6, or calmodulin. Using truncated p53 proteins in the N-terminal (p531–320) and C-terminal (p5373–393) domains, we localized the S100B-binding region to the C-terminal region of p53. We have confirmed a calcium-dependent interaction of the S100B with a synthetic peptide corresponding to the C-terminal region of p53 (residues 319–393 in human p53) using plasmon resonance experiments on a BIAcore system. In the presence of calcium, the equilibrium affinity of the S100B for the C-terminal region of p53 immobilized on the sensor chip was 24 ± 10 nm. To narrow down the region within p53 involved in S100B binding, two synthetic peptides, O1357–381 (residues 357–381 in mouse p53) and YF-O2320–346 (residues 320–346 in mouse p53), covering the C-terminal region of p53 were compared for their interaction with purified S100B. Only YF-O2 peptide interacts with S100B with high affinity. The YF-O2 motif is a critical determinant for the thermostability of p53 and also corresponds to a domain responsible for cytoplasmic sequestration of p53. Our results may explain the rescue of nuclear wild type p53 activities by S100B in fibroblast cell lines expressing the temperature-sensitive p53val135 mutant at the nonpermissive temperature.

The S100 family is one of the largest subfamily of calciumbinding proteins that are thought to play roles in mediating calcium signals during cell growth, differentiation, and motility (reviewed in Ref. 1). These proteins are characterized by highly conserved helix-loop-helix calcium-binding domains, known as EF-hand motifs. The S100B is found in astroglial cells in the central nervous system but also in a number of tissues outside the nervous system, including adipose tissue, testis, skin, and lymphocytes (2). In cultured cells, synthesis of S100B is tightly regulated and is maximum in the G 1 phase of the cell cycle (3)(4)(5). The S100B protein is a noncovalent homodimer formed by the association of two S100-␤ subunits (6,7). Calcium bind-ing induces conformational changes in the protein structure (8 -10), destabilizing the S100B quaternary structure and allowing interaction with target proteins (7)(8)(9)(10)(11)(12). For example, the Ca 2ϩ -dependent binding of S100B to microtubules (13), GFAP (14), Ndr protein kinase (15), and p53 (16) has been demonstrated. The identification of S100B target proteins is essential to better understand the mechanisms underlying S100B functions. Like calmodulin, S100B is likely to regulate multiple target proteins.
The tumor suppressor p53 protein (17) is a putative intracellular target for the S100B protein. In vitro, S100B binds to p53 and inhibits p53 aggregation and phosphorylation by PKC (16). 1 S100B also binds to a peptide derived from the extreme C-terminal end of p53 (12,18). A functional interaction between S100B and p53 was recently demonstrated in p53negative mouse embryo fibroblasts by sequential transfection with the S100B and temperature-sensitive p53val135 genes (5). In mouse embryo fibroblast cells expressing a low level of p53val135, S100B cooperates with calcium in triggering p53val135-dependent cell growth arrest and death at the nonpermissive temperature (37.5°C). Activation of wild type p53 functions by S100B correlated with nuclear accumulation of p53val135 under a wild type conformation. S100B modulation of p53val135 functions was confirmed in the rat embryo fibroblast (REF) cell line clone 6, which is transformed by oncogenic Ha-ras and overexpression of p53val135 (19). Ectopic expression of S100B in clone 6 cells reverts transformed phenotypes and restores wild type p53 activities at 37.5°C (5). We have hypothesized that S100B could either inactivate the mutant p53val135 species and/or favor folding of the p53val135 under a wild type conformation (5).
The functional activity of the p53val135 mutant is associated with its conformational flexibility (20). At the nonpermissive temperature (37.5°C), the temperature-sensitive p53val135 mutant folds into an inactive mutant conformational form, detectable by the monoclonal antibody PAb240. At the permissive temperature, the p53val135 protein adopts a wild type conformation detectable by the monoclonal antibody PAb246 and is functionally active (i.e. able to specifically bind DNA and to activate transcription). The conformational flexibility that characterizes the temperature-sensitive p53val135 mutant is not unique to this mutation and could be representative of the conformational flexibility that also characterizes wild type p53 (20,21). In vivo, conformational shift between the wild type and mutant conformations (recognized by the monoclonal antibodies PAb246 and PAb240, respectively) has been observed during embryonic differentiation of embryonic stem cells (22). In vitro, wild type conformation of Escherichia coli expressed recombinant p53 can be allosterically regulated for stabilization by two regulatory sites found within the N and C termini of the molecule to which ligands may bind (23). A major challenge now is to characterize the domain on p53 associated with conformational shift and to identify the cellular co-factors that may govern the conformational switch controlling p53 activity (20).
In this report, we show that the S100B protein interacts with baculovirus recombinant p53 and protects p53 from thermal denaturation. We also identify a high affinity S100B-binding domain within p53 called YF-O2 (residues 320 -346 in mouse p53). This domain constitutes a critical determinant for the stability and oligomerization status of p53 (24 -26). It corresponds to the minimal region of the C terminus that confers thermal instability to p53 (25) and also corresponds to a domain responsible for cytoplasmic sequestration of p53 (27).

MATERIALS AND METHODS
Peptides-The p53 319 -393 peptide was synthesized as described previously (28). The peptide was dissolved in 20 mM HEPES-NaOH, pH 7.5, at the concentration of 98 M. Other peptides were synthesized by the solid phase method, using Fmoc (fluoren-9-ylmethoxycarbonyl) chemistry on an Applied Biosystems 430A peptide synthesizer. The O1 peptide, which encompasses the PKC phosphorylation site domain, corresponds to residues 357-381 within mouse p53: RAHSSYLKT-KKGQSTSRHKKTMVKK. The peptide was dissolved in 20 mM Tris-HCl, pH 7.5, at the concentration of 100 -200 M. YF-O2 peptide has the following sequence: YFTLKIRGRKRFEMFRELNEALELKD and corresponds to residues 321-346 within mouse p53. W-O2 peptide is a mutated version of YF-O2, and Tyr 321 and Phe 322 were both substituted with Trp residues. YF-O2 and W-O2 peptides were dissolved in 10 mM HEPES-HCl, pH 2.5, at the concentration of 200 -400 M. Peptide stock solutions were stored at Ϫ20°C.
Proteins-S100B, S100A1, and calmodulin were purified from bovine brain (10). Recombinant Human p53 was expressed in SF9 cells and purified to homogeneity by S100B affinity chromatography as described previously (16). To express truncated human p53, corresponding cDNA fragments were generated from full-length human p53 by polymerase chain reaction (oligonucleotide sequences available upon request). After verification of DNA sequences, c-DNA fragments were inserted into pBlueBac III vector (Invitrogen). Expression in insect cells and interaction with S100B-Sepharose column were as described (16).
Electrophoretic Mobility Shift Assays-The method used was identical to that previously published (29). To study the effect of S100B and other related calcium-binding proteins on p53 DNA binding activity, purified p53 was first incubated in the presence or in the absence of calcium-binding proteins for 15 min at 30°C in 20 mM Tris-HCl buffer, pH 7.5, 110 mM KCl, 5 mM dithiothreitol, and 1 mM CaCl 2 . The proteins were then mixed with 1 ng of labeled DNA, further incubated at 4°C for 10 min prior to electrophoresis onto a 5% nondenaturing polyacrylamide gel, and run in buffer containing 10 mM Hepes, 10 mM Tris, pH 8.0, and 0.5 mM CaCl 2 .
Real Time Surface Plasmon Resonance Recording-Real time binding experiments were performed on a BIAcore biosensor system (Pharmacia Biosensor AB, Uppsala, Sweden). All experiments were performed at 25°C. p53 319 -393 peptide (35 l of 1 M peptide in 10 mM sodium acetate, pH 3.5) was directly coupled through its amino groups to the sensor surface activated by N-hydroxy-succinimide and N-ethyl-NЈ-(dimethylaminopropyl)carbodiimide according to the manufacturer's instructions. The remaining reactive groups were then inactivated with 1 mM ethanolamine. For control experiments, sensor surface was treated as above in the absence of peptide. All interaction experiments were done using a running buffer containing 20 mM Tris-HCl, pH 7.5, 120 mM NaCl, 1 mM CaCl 2 . Between injections, the sensor chip was washed with a buffer containing 20 mM Tris-HCl, pH 7.5, 120 mM NaCl, 2 mM EDTA. Sensorgrams were analyzed using BIAevaluation 3.0 program (Pharmacia Biosensor AB) and kinetic constants were obtained by fittings curves to a single-site binding model (A ϩ B ϭ AB).
Protein Kinase C Activity Assay and PKC Inhibitor Assay-The PKC assays were performed as described previously (30). To analyze the effect of S100B and other calcium-binding proteins on the inhibitor capacity of YF-O2 or W-O2 peptides, the following procedure was used. 10 l of diluted inhibitor peptides in H 2 O (1:4000 -1:50 dilution of stock solution) were mixed in Eppendorf tubes with increasing concentration of calcium-binding proteins in 30 l of 50 mM Tris-HCl buffer, pH 8.0, 0.5 mM CaCl 2 . 10 l of PKC was subsequently added to each tube to reach a final concentration of 3 nM, and the protein mixture was vortexmixed. The phosphorylation reaction was immediately initiated by the addition of 20 l of a mixture containing O1 peptide used as substrate at 8 M final concentration plus MgCl 2 , phosphatidylserine, dioleoylglycerol, and [␥-32 P]ATP at final concentrations used in standard PKC assay (30).
Native Gel Electrophoresis-Gels containing 13% acrylamide, 0.375 M Tris-HCl, pH 8.3, and 0.5 mM CaCl 2 or 1 mM EGTA were prepared without stacking gel. W-O2 peptide and S100B were incubated in 40 mM Tris-HCl, pH 8.0, 0.2 M NaCl, 5 mM dithiothreitol, 15% glycerol, and 0.5 mM CaCl 2 or 1 mM EGTA for 5 min on ice prior to electrophoresis. The gels were run at 20 mA. The electrode buffer consisted of 25 mM Tris, 192 mM glycine, and 0.5 mM CaCl 2 or 1 mM EGTA. The gels were stained with Coomassie Blue.
Fluorescence Measurements-The fluorescence measurements were performed with a Perkin-Elmer spectrofluorometer in 25 mM Tris-HCl, pH 7.5, 0.12 M NaCl containing either 0.1 mM CaCl 2 or 2 mM EGTA. For titration experiments, excitation was at 280 nm, and emission was monitored at 340 nm.

S100B Protects p53 from Thermal Denaturation-Heating
baculovirus recombinant p53 at 30°C for 15 min leads to selfaggregation (16) and permanent loss of DNA binding activity ( Fig. 1A; see also Ref. 23). S100B protein, which prevents p53 from temperature-dependent aggregation (16), counteracts the temperature-dependent inhibition of p53 DNA binding activity (Fig. 1B, lanes 1 and 2). The S100B effect is specific because calmodulin, S100A1 and S100A6 have no effect (Fig. 1B, lanes [3][4][5]. The dose response experiment shown in Fig. 1C reveals that the S100B concentrations required to protect the DNA binding activity of p53 from thermal inhibition are similar to those required to protect p53 from temperature-dependent aggregation (16). S100B was only able to activate p53 for DNA binding if preincubated with p53 at 30°C. S100B was not able to activate p53 for DNA binding when incubated at 4°C, and it did not reactivate DNA binding activity of denatured p53 (data not shown). Thus the effect of S100B is limited to protection of  5-8). The proteins were then mixed with 1 ng of labeled DNA and further incubated at 4°C for 10 min prior to electrophoresis. B, S100B counteracts the temperature-dependent inhibition of p53-DNA binding. p53 (0.1 g) was incubated for 15 min at 30°C in the absence (lane 1) or in the presence of 1 M of S100B (lane 2), calmodulin (lane 3), S100A1 (lane 4), or S100A6 (lane 5). The proteins were then mixed with 1 ng of labeled DNA and further incubated at 4°C for 10 min prior to electrophoresis. C, S100B titration of maximal stabilization of the DNA binding activity of p53 at 30°C. 1 M p53 was first incubated for 15 min at 30°C in the absence (lane 1) or in the presence of increasing amount of S100B (lanes 2-6) as indicated. The proteins were mixed with 1 ng of labeled DNA and 20 ng of unlabeled DNA. p53 from thermal denaturation. In that respect, S100B behaves differently from PAb421. PAb421 binds to the extreme C terminus of p53 to activate p53 DNA binding but does not protect p53 from thermal denaturation (23).
Mapping of the Minimal S100B-binding Domain on p53-Recombinant human p53 (p53 1-393 ) expressed in SF9 cells accumulates in the cytoplasm and in the nuclei of infected cells. Both cytoplasmic and nuclear recombinant human p53 bind to S100B-Sepharose column in the presence of calcium and are eluted with buffer containing EGTA ( Fig. 2A). We next compared the binding to S100B-Sepharose of human p53 fragments that lacks the C-terminal (p53 1-320 ) (Fig. 2B) or Nterminal (p53 73-393 ) domain (Fig. 2C). p53 1-320 that lacks nuclear localization signals (31) accumulates within the cytoplasm of infected cells, whereas p53 73-393 is predominantly present within the cell nuclei. Cell extracts enriched with each of these p53 fragments were loaded onto a S100B-Sepharose column, and the calcium-dependent binding of the p53 fragments to the column was analyzed by SDS-polyacrylamide gel electrophoresis, followed by Western blot analysis using monoclonal antibodies directed against N-terminal or C-terminal epitopes (Fig. 2, B and C). p53 1-320 that lacks the C-terminal domain does not bind to S100B and was recovered in the flow-through fractions (Fig. 2B). Only p53 73-393 efficiently binds to the S100B column in the presence of calcium and is eluted with buffer containing EGTA (Fig. 2C). Hence, the S100B-binding region is localized to the C-terminal region of the p53 protein (residues 320 -393) (Fig. 3).
The molecular interaction between S100B and the C-terminal region of p53 was then analyzed using surface plasmon resonance detection (Fig. 4). A synthetic peptide corresponding to human p53 319 -393 was coupled to the biosensor chip through amino functions. Fig. 4A shows a typical sensorgram representing the real time interaction between purified Ca 2ϩ -bound S100B and the p53 319 -393 peptide. The S100B-p53 319 -393 interaction is strictly dependent on the presence of Ca 2ϩ and completely reversed by addition of EDTA (Fig. 4A). We controlled that there is absolutely no interaction of S100B with a biosensor ship with no peptide immobilized and neutralized with ethanolamine (Fig. 4A). The reversibility of the S100B-p53 319 -393 interaction allows recycling the biosensor chip with EDTA containing buffer and to perform a dose-dependent titration of the interaction utilizing the same biosensor chip (Fig.  4B). At low S100B concentrations (10 -100 nM) the kinetics of interactions could be fitted with a single interaction site model and an apparent equilibrium affinity of S100B for p53 319 -393 of 25 Ϯ 10 nM. At higher S100B concentrations (100 -900 nM), kinetics of interaction could barely be fitted with a single site model may be due to titration of lower affinity sites (see "Discussion"). The specificity of the high affinity Ca 2ϩ -dependent interaction between p53 319 -393 and S100B was confirmed by comparison with two other related EF-hand calcium-binding proteins, S100A1 and calmodulin (Fig. 4C, inset). Neither S100A1 nor calmodulin was able to interact with the p53 319 -393-coated sensorship (Fig. 4C).
Molecular Mapping of the High Affinity S100B-binding Site on p53-To define in further detail the high affinity S100Bbinding region within p53 319 -393 , two synthetic peptides corresponding to putative S100B-binding sites within the C-terminal domain of p53 were synthesized and assessed for their capacity to interact with purified S100B (Fig. 3).
The first peptide called O1 357-381 (residues 357-381 in mouse p53) shows similarities with the S100B and calmodulinbinding domain of the MARCKS protein (16). Like its cognate peptide on MARCKS, O1 peptide is phosphorylated by PKC in vitro on multiple sites (30). Because binding of S100B to p53 inhibits p53 phosphorylation by PKC (16), we analyzed the effect of S100B on PKC-mediated phosphorylation of O1. As shown in Fig. 5, addition of S100B (O1 357-381 :S100B molar ratio 0.5) had no effect on the phosphorylation of O1 357-381 by PKC. On the contrary , monoclonal antibody PAb421 that binds to O1 357-381 inhibited phosphorylation by PKC. PAb242, which does not bind to O1 357-381 , had no effect on its phosphorylation. These results suggest that the high affinity S100-binding domain within p53 is distinct from O1 357-381 . In further support of that conclusion, O1 peptide at the concentration of 1 M did not antagonize S100B binding to the p53 319 -393 -coated sensorship (not shown).
The second peptide, YF-O2 320 -346 (residues 320 -346 in  FIG. 3. The C-terminal region of p53. Amino acid sequence of the C-terminal region of human p53 involved in binding to S100B protein. This region is sufficient to transform primary fibroblasts in cooperation with ras. The minimal domain required for cell transformation is in bold (44). The oligomerization (28) and cytoplasmic sequestration domain is boxed (27). The putative PKC phosphorylation sites and PAb421 epitope are indicated (30). The mouse YF-O2 and O1 peptide sequences used in this study are also presented. mouse p53) have sequence, structural, and functional homologies with the basic helix I domain on myogenic basic helix-loophelix proteins, which is also capable of interacting with S100A1 or calmodulin (30, 32-34). We next analyzed the interaction of S100B with the YF-O2 320 -346 peptide and its tryptophan derivative, W-O2. In a competition assay, W-O2 is capable of antagonizing S100B binding to the p53 319 -393 -coated sensorship (Fig. 4A). A physical interaction of S100B with W-O2 peptide was then demonstrated by fluorescence spectroscopy, taking advantage of the fact that the S100B protein has no Trp residue (Fig. 6, A and B). Incubation of S100B with W-O2 in the presence of 0.1 mM CaCl 2 resulted in a 25% decrease in Trp fluorescence of the W-O2 peptide with a shift of the maximum emission from 345 to 340 nm (Fig. 6A). We next performed a dose-dependent titration of the fluorescence intensity changes of the W-O2 peptide at concentrations (1 M and 2 M) far above the apparent K d determined with surface plasmon resonance. Maximum change in Trp fluorescence intensity was observed at a molar ratio of 0.5 mol S100B/mol W-O2, suggesting a stoichiometric binding of one ␤ subunit of S100B protein/W-O2 peptide molecule (Fig. 6B). In the presence of EGTA, significant decrease in Trp fluorescence intensity can also be observed (Fig. 6B). These changes are probably representative of low affinity and nonspecific interactions. Native gel electrophoresis was also used to monitor complex formation between S100B and W-O2 (Fig. 6C). Using that method, low affinity Ca 2ϩindependent interaction could not be detected. Only high affinity Ca 2ϩ -dependent complex formation between S100B and W-O2 was observed, as indicated by an upward shift of the S100B band. Note that in the presence of Ca 2ϩ , the change in electrophoretic mobility of S100B is due to conformational changes induced upon Ca 2ϩ binding (32).
YF-O2 320 -346 corresponds to a putative PKC-binding domain on p53. The YF-O2 320 -346 peptide binds to the catalytic site of calcium-dependent PKC (PKC ␣, ␤, and ␥) with affinity in the 50 -100 nM range and inhibits kinase activities (30). As previously reported, the interaction of YF-O2 320 -346 with PKC results in a total inhibition of O1 357-381 peptide phosphorylation (Fig. 7A). The tryptophan derivative W-O2 was as potent as YF-O2 320 -346 in inhibiting O1 357-381 peptide phosphorylation (Fig. 7A). In the presence of S100B, inhibition of O1 357-381 peptide phosphorylation is abrogated, suggesting that S100B competes with PKC for binding on YF-O2 and W-O2 peptides (Fig. 7A). As expected, the half-maximum effect of S100B was dependent on the W-O2 peptide concentrations used in the assay (Fig. 7B). In all conditions, the maximum were reached for 0.5 mol S100B added per mol peptide, confirming a stoichiometric binding of one S100-␤ subunit per W-O2 or YF-O2 peptide molecule. Calmodulin or S100A6 counteracted W-O2and YF-O2-dependent inhibition of PKC activity only at high concentrations (Fig. 7A). It has been proposed that YF-O2 domain might serve to target PKC on p53 and that it could be an important potency determinant in p53 phosphorylation by PKC (30). The ability of S100B to prevent YF-O2-PKC interac- FIG. 4. Characterization of the interaction of purified S100B with p53 319 -393 using surface plasmon resonance detection. p53 319 -393 peptide was covalently immobilized on the sensor surface as described under "Materials and Methods." Running buffer contains 20 mM Tris-HCl, pH 7.5, 120 mM NaCl, 1 mM CaCl 2 . The different sensorgrams shown here have been obtained on the same sensor surface except for the control. Three different sensor surfaces coated with p53 319 -393 peptide gave similar results with respect to their interaction with the different proteins. A, S100B (90 nM in running buffer) was injected. At the end of injection, the sensor surface was washed with running buffer and then with a buffer containing 2 mM EDTA. Because of the high resonance signal induced by EDTA, the dissociation of S100B is only visible after the end of the injection of the EDTA buffer. S100B (90 nM in running buffer) was preincubated with WF-02 peptide (1 M) prior to injection (S100B-W-O2). Control corresponds to the injection of S100B on a sensor surface treated in the absence of p53 319 -393 peptide. B, concentration dependence interaction of S100B with p53 319 -393 . Increasing concentration of S100B was injected in running buffer on the sensor surface coated with 11.25 nM (a), 30 nM (b), 45 nM (c), 60 nM (d), or 90 nM (e) of p53 319 -393 peptide. Control is as in A. C, the interaction of S100B (90 nM) with p53 319 -393 immobilized on the sensor chip surface is compared with S100A1 (90 nM) and calmodulin (90 nM). Insert shows SDS-polyacrylamide gel electrophoresis and Coomassie Blue staining of the purified bovine brain S100A1 (lane 1), S100B (lane 2), and calmodulin (lane 3) used in that study. The asterisks indicate position of covalent S100A1 and S100B dimers. tion might explain the inhibitory effect of S100B on full-length p53 phosphorylation by PKC (16). DISCUSSION In the first part of this study, we have confirmed that baculovirus recombinant wild type p53 is highly sensitive to thermal denaturation (16). Temperature-dependent denaturation of p53 is linked to protein aggregation (16) and compromised DNA binding activity of p53 ( Fig. 2A). The interaction of S100B with recombinant wild type p53 protects p53 from thermal denaturation (16; Fig. 2, B and C). Secondly, we have confirmed a specific and strict calcium-dependent interaction of the S100B with a synthetic peptide derived from the C-terminal region of p53 (residues 319 -393 in human p53) by plasmon resonance measurement experiments. We have also identified the YF-O2 peptide as part of a high affinity S100B-binding site within the C-terminal region of p53. Structural studies have shown that the YF-O2 sequence is a critical domain for p53 thermostability (24,25). It is likely that interaction of S100B with YF-O2 peptide on p53 is directly responsible for the increased resistance of p53 to thermal denaturation. Like S100B, the E. coli heat shock protein DnaK has been shown to bind to the C terminus of p53 and protect p53 from thermal denaturation (23,26). S100B is the first mammalian protein to share this property with DnaK.
The YF-O2 peptide on p53 is involved in the dimerization and in subsequent tetramerization of p53 dimers (28,(35)(36)(37). The K d value found for the tetramer-monomer transition of the C-terminal domain of p53 was determined to be 1-10 M (28,35). Because of the high K d value, it has been proposed that in normal, undamaged cells, when the cytoplasmic p53 concentration is very low, p53 may be largely monomeric (35). However in its monomeric state, the p53 molecule is thermodynamically unstable relative to the tetramer (24,25). The high equilibrium affinity of the S100B for the C-terminal domain immobilized on the sensor chip (K d 24 Ϯ 10 nM) suggests that S100B could transiently interact with cytoplasmic p53 monomer to stabilize the p53 in a native conformation prior to nuclear translocation. Once p53 has translocated to the cell nuclei, it may be stabilized through its interactions with DNA targets as a tetramer. The idea that S100B could be involved in conformational modulation and stabilization of p53 monomer into a native (wild type) conformation is supported by in vivo observations (5). In mouse embryo fibroblast cells expressing low levels of the temperature-sensitive p53val135 mutant, S100B cooperates with calcium in stabilization and activation of the wild type p53val135 conformational species at the nonpermissive temperature (37.5°C) (5). The YF-O2 sequence on p53 also corresponds to a "cytoplasmic sequestration domain" (27). Hence, one can envision that in response to intracellular calcium elevation, S100B could also dissociate the interaction between p53 and cytoplasmic anchoring proteins to favor p53 nuclear translocation. It is noteworthy that in mouse embryo fibroblast cells expressing S100B and the p53val135, calcium-mediated stabilization of the p53val135 under a wild type conformation is also associated with wild type p53val135 nuclear accumulation (5). The involvement of calcium signaling and S100B in activation of wild type p53 probably not only concern the p53val135 mutant but could be of more general occurrence. Intracellular calcium elevation is induced in physiological stimulation known to activate p53 functions, including cell contact (38) and hypoxia associated with tumor formation (39,40). S100B, which is normally expressed at low levels in proliferating glial cells or in peripheral tissues, is also strongly induced by such stimulations. For examples, in C6 glial cells, S100B synthesis correlates with G 1 phase growth arrest at confluence (5). In cardiac myocytes, S100B is induced in response to hypoxia (41). This last observation also suggests that overexpression of S100B generally observed in brain tumors and peripheral tumors (42,43) could be linked to hypoxia during tumor formation (39). It is tempting to speculate that in normal cells and also in tumor cells S100B induction could also be associated with p53 activation.
Finally, the high affinity S100B-binding domain within p53 that we identified also corresponds to the minimal transforming domain on mutant p53 (Ref. 44 and Fig. 3). Overexpression of the transforming C-terminal p53 mini-protein harboring YF-O2 domain cooperates with the ras oncogene in cell transformation in the absence of endogenous wild type p53 (45). This suggests that YF-O2-mediated transformation operates via stable interaction and inactivation of p53 target proteins. Be-  1 (lanes 3), and 2 (lane 4). The arrowhead on the right points to S100␤-W-O2 complex.
FIG. 7. S100B counteracts YF-O2 and W-O2 inhibition of PKC activity. A, inhibition of PKC activity by YF-O2 and W-O2 and release of this inhibition by S100B, S100A6, and calmodulin. PKC activity was measured using O1 peptide as substrate. B, titration curves of S100Bdependent release of PKC in the presence of 0.6 M (), 3.3 M (q), or 26 M (ࡗ) of W-O2 peptide. cause ectopic expression of S100B in a REF cell line overexpressing ras and the p53val135 is able to revert transformed phenotypes of the cells, it would be worth investigating whether S100B could also dissociate interactions between YF-O2 and target proteins.
The K d value found for the interaction between S100B and the p53 319 -393 peptide (24 nM) is in the same range as that reported for the interaction between the S100A1 and the basic helix-loop-helix peptide of MyoD (20 nM) (32)(33)(34). A conserved amino acid sequence exists within the YF-O2 sequence of p53 ( 344 LNEALELK 351 ) that is implicated in S100B binding and the helix I motif within myogenic transcription factor MyoD ( 124 VNEAFETLK 132 ) involved in S100A1 binding. These two motifs are involved in the regulation of p53 and MyoD cellular localization (27,46). The identification of a putative S100 target consensus motif within p53 and MyoD with functional homologies suggests that regulation of nuclear translocation of transcription factors associated with negative cell growth regulation may be a general feature of S100 proteins function. It is interesting to note that interaction of S100B and S100A1 with an other nuclear protein, the Ndr protein kinase, has recently been characterized (15). It would be interesting to study the possible contribution of S100 proteins in Ndr protein kinase nuclear translocation. Despite the strong amino acid sequence homology between the YF-O2 sequence on p53 and the helix I motif on MyoD, the interaction between p53 319 -393 peptide and S100B is isoform-specific (Fig. 4C). Such specificity suggests that in addition to the consensus sequence, other amino acids adjacent to that motif could be responsible for the specificity of interaction between S100B and p53. Weber's group (12,18) recently reported that S100B binds to a basic peptide derived from the extreme C-terminal basic tail of human p53 (residues 367-388), which corresponds to the O1 peptide. The high dissociation constant for this interaction (K d ϭ 20 M) (12,18) would explain why we have not been able to detect interaction of this peptide with S100B (Fig. 5). In our experimental conditions, the low S100B concentration used (4 M) was probably not sufficient to reveal low affinity interaction. It is nevertheless possible that O1 domain contribute to strengthen the interaction of S100B dimer with the full-length C-terminal region of p53. The possible implication of O1 in S100B binding to p53 319 -393 peptide may explain why analysis of the kinetics of interactions between S100B and p53 319 -393 immobilized on the sensor chip cannot be analyzed with a single interaction site model at high S100B concentration. Although O1 peptide is not involved in p53 dimerization and tetramerization (24,28,35,36), this domain is nevertheless required for formation of higher oligomeric aggregates (37) and is engaged in transient interchain interaction with another domain within the protein tetramer (35,47). It is likely that, in addition to its effect on wild type p53 stabilization, S100B could also affect an important protein interaction at the C-terminal domain involving O1 domain.