Molecular Basis of S100 Proteins Interacting with the P53 Homologs P63 and P73

Oncogene (2010) 29, 2024–2035 & 2010 Macmillan Publishers Limited All rights reserved 0950-9232/10 $32.00 www.nature.com/onc ORIGINAL ARTICLE Molecular basis of S100 proteins interacting with the p53 homologs p63 and p73

J van Dieck, T Brandt, DP Teufel, DB Veprintsev, AC Joerger and AR Fersht

MRC Centre for Protein Engineering, Hills Road, Cambridge, UK

S100 proteins modulate p53 activity by interacting with its binding proteins in vertebrates (Donato, 2003). Their tetramerization (p53TET, residues 325–355) and trans- expression has been implicated with neurological activation (residues 1–57) domains. In this study, we disorders, cardiac and inflammatory diseases (Ilg et al., characterized biophysically the binding of S100A1, 1996; Mrak and Griffinbc, 2001; Donato, 2003; Dean S100A2, S100A4, S100A6 and S100B to homologous et al., 2006; Sedaghat and Notopoulos, 2008). Several domains of p63 and p73 in vitro by fluorescence S100 proteins have been shown to be valuable bio- anisotropy, analytical ultracentrifugation and analytical markers for cancerous diseases (Salama et al., 2008). gel filtration. We found that S100A1, S100A2, S100A4, Further, S100 proteins have been linked to processes of S100A6 and S100B proteins bound different p63 and p73 neuronal development, cell survival and differentiation tetramerization domain variants and naturally occurring (Chan et al., 2003; Buckiova and Syka, 2009). Similarly isoforms with varying affinities in a calcium-dependent to calmodulin, the S100 proteins undergo a conforma- manner. Additional interactions were observed with tional change on Ca2 þ binding. S100 proteins interact peptides derived from the p63 and p73 N-terminal with other proteins in a calcium-dependent as well as a transactivation domains. Importantly, S100 proteins calcium-independent manner (Zimmer et al., 2003; bound p63 and p73 with different affinities in their Santamaria-Kisiel et al., 2006). A common feature of different oligomeric states, similarly to the differential S100 proteins is their ability to bind to the tumor modes of binding to p53. On the basis of our data, we suppressor p53. We previously discovered that S100 hypothesize that S100 proteins regulate the oligomeriza- proteins generally bind to the tetramerization domain tion state of all three p53 family members and their (residues 325–355) and the N-terminal transactivation isoforms, with a potential physiological relevance in domain (residues 1–57) of p53. In addition, a subset of developmental and disease-related processes. The regula- the protein family binds to the negative regulatory tion of the p53 family by S100 is complicated and depends domain (367–393) of the tumor suppressor. Further, on the target preference of each individual S100 protein, S100 proteins have been shown to potentially activate as the concentration of the proteins and calcium, as well as well as inhibit p53 (Baudier et al., 1992; Scotto et al., the splicing variation of p63 or p73. Our results outlining 1998, 1999; Rustandi et al., 2000; Grigorian et al., 2001; the complexity of the interaction should be considered Lin et al., 2001, 2004; Fernandez-Fernandez et al., 2005, when studying the functional effects of S100 proteins in 2008; Mueller et al., 2005; Salama et al., 2008; Slomnicki their biological context. et al., 2009; van Dieck et al., 2009). Oncogene (2010) 29, 2024–2035; doi:10.1038/onc.2009.490; The tumor suppressor p53 has two homologs, p63 and published online 8 February 2010 p73. Unlike p53, which is mutated in more than 50% of all tumors (Joerger and Fersht, 2007), p63 and p73 are Keywords: S100; p63; p73; tumor suppressor; protein– rarely mutated in cancers. Nevertheless, overexpression protein interaction of p63 and p73 is observed in several types of cancer (Hibi et al., 2000; Yamaguchi et al., 2000; Moll and Slade, 2004; Deyoung and Ellisen, 2007; Finlan and Hupp, 2007). Further, p63 and p73 are involved in Introduction several developmental processes (Danilova et al., 2008). Structurally, the three p53 family members share The family of S100 proteins is a highly conserved group conserved transactivation, DNA-binding and oligo- of more than 20 members of small acidic calcium- merization (TET) domains, but differ greatly in their C-terminal region. p63 and p73 full-length proteins have an extended C-terminus including a SAM-domain, Correspondence: Professor AR Fersht, MRC Centre for Protein which could be involved in oligomerization, protein– Engineering, Cambridge University, Hills Road, Cambridge, Cambs protein and protein–RNA interactions, and a transcrip- CB2 0QH, UK. E-mail: [email protected] tion inhibitory domain, which may interact with the p63 Received 6 August 2009; revised 16 October 2009; accepted 27 October or p73 transactivation domain to inhibit transcriptional 2009; published online 8 February 2010 activity (Serber et al., 2002; Scoumanne et al., 2005). Interaction of S100 proteins with p63 and p73 J van Dieck et al 2025 Several splice-isoforms have been identified including Table 1 Binding of S100 proteins to the tetramerization domain of (TA-) or lacking (DN-) the transactivation domain, the p53 family and with a full ( a) or truncated C-terminus (p63b-g, À Protein Kd (mM) p73b-Z) (Scoumanne et al., 2005). Overall, more than 50 isoforms have been identified for the p53 family p63TET p73TET p73TET-long p53TET a (Rosenbluth and Pietenpol, 2008), some of them with S100A1 0.15±0.04 n.q. n.q. 9.2±0.2 similar but others even with opposing functions, implying S100A2 0.11±0.04 0.25±0.01 n.q. 2.1±0.1 a complicated regulatory network of the p53 family based S100A4 0.09±0.03 0.02±0.01 0.21±0.05 17±1.0 on splicing. p63 and p73 share great sequence identity S100A6 0.12±0.02 0.07±0.04 0.48±0.14 12±0.2 with p53, especially in the DNA-binding domain (B 60% S100B 2.80±0.60 n.q. n.q. 112±3.0 identity) (Melino et al., 2003), the transactivation domain B B Abbreviation: n.q., not quantifiable. ( 30% identity) and the tetramerization domain ( 40% aTaken from (Fernandez-Fernandez et al., 2005, 2008). identity with p53, B70% identity with each other) (Scoumanne et al., 2005; Vilgelm et al., 2008). On the basis of the sequence homology between the the weaker the binding to S100 proteins. When the tetramerization domain of p53, p63 and p73, we studied concentration of labelled peptide was much higher than whether the members of the S100 family also interact with the Kd, it was possible to perform a stoichiometric p63 and p73 in vitro by fluorescence anisotropy, analytical titration. Titration of S100A4 to 0.5 mM p63TET size-exclusion chromatography (SEC) and analytical ultra- revealed binding of one dimer S100A4 per monomer centrifugation (AUC). We found that binding to the p63TET (Figure 1b). The same stoichiometry was tetramerization domains and the transactivation domains observed in similar experiments for S100A1, S100A2 ofthep53familyareacommonfeatureofS100proteins. and S100A6 (data not shown). The binding of S100 The binding is calcium dependent and influenced by the proteins to p53 is dependent on calcium (Fernandez- oligomerization state of the p53 family. Interestingly, the Fernandez et al., 2005, 2008; van Dieck et al., 2009). affinity of S100A1, S100A2, S100A4, S100A6 and S100B is Consequently, we analyzed if S100 proteins can still significantly higher for the tetramerization domains of p63 bind p63TET in the absence of calcium and in the and p73 than for p53. We show that S100 proteins presence of 1 mM ethylene-diaminetetraacetic acid to differentially bind tetrameric, dimeric and monomeric p63 complex any traces of bivalent cations. We found that and p73, suggesting a common regulatory function for all S100 proteins were not able to bind p63TET in the the p53 family members. The calcium-dependent interac- presence of ethylene-diaminetetraacetic acid (Figure 1c). tion of S100 proteins with p63 and p73 could be For the direct titrations of S100 proteins to p63TET, we physiologically relevant in developmental as well as used a buffer of 10 mM CaCl2, which is a large excess of disease-related processes. calcium, to induce the active conformation of S100 proteins. We performed calcium titrations to monitor how much calcium is needed to induce the binding of Results S100 proteins to p63TET (Figure 1d, Table 2). For that purpose, we titrated calcium to a solution of p63TET S100 proteins bind the p63 tetramerization domain in and S100 proteins at concentrations far above the Kd in different oligomeric states and influence oligomerization a buffer without calcium. Consequently, the binding of To quantify the binding of S100 proteins to the S100 to p63TET was solely dependent on the amount of tetramerization domain of p63, we overexpressed calcium titrated to the cuvette. We found that far less human S100A1, S100A2, S100A4, S100A6 and S100B than 10 mM CaCl2 was sufficient to induce the interac- in Escherichia Coli and purified them. All five tion, measuring EC50 values ranging from B5 mM for proteins interact with the tetramerization domain of S100A4 to B 280 mM for S100B. p53 in fluorescence anisotropy experiments (Fernandez- As the concentration of p63TET in the fluorescence Fernandez et al., 2005, 2008). A homologous p63- anisotropy experiments was very low, probably contain- derived peptide, corresponding to the canonical p53 ing tetrameric p63TET as well as lower oligomeric tetramerization domain region, p63TET(359–390) was forms, we performed analytical SEC with a recombinant chemically synthesized with an N-terminal fluorescein p63TET construct at high mM concentrations, in which label. S100 proteins were titrated into 0.1 mM of p63TET p63TET could be expected to be tetrameric. Indeed, in physiological ionic strength buffer. We measured p63TET eluted as a tetramer as determined by multi- dissociation constants (Kd) in the range of 100 nM for angle light scattering. The S100 proteins eluted as S100A1, S100A2, S100A4 and S100A6, and 2.8 mM for dimers (Table 3). A mixture of S100A1, S100A2, S100B (Table 1). The observed affinities were signifi- S100A4 or S100A6 and p63TET with a ratio of 2:1 cantly higher than for the homologous p53TET peptide resulted in a single peak with a shift in retention time (tr) (Fernandez-Fernandez et al., 2005, 2008). Similarly to from that of S100 alone (Figure 2). According to the p53TET, the affinity to p63TET was highly dependent multi-angle light scattering detection, the MW corre- on the concentration of the labelled peptide used in the sponding to the peaks of the complexes increased, with anisotropy titration experiments, as illustrated by the the maximum apparent MW increase corresponding to binding of S100A4 to 0.1, 0.5, 1 and 2 mM p63TET the mass of a p63TET monomer (Table 3). This means (Figure 1a). The more TET used in the experiments, that the tetramer of p63TET was disrupted through the

Oncogene Interaction of S100 proteins with p63 and p73 J van Dieck et al 2026

Figure 1 Fluorescence anisotropy titration of S100A4 to p63TET(359–390). The higher the concentration of ﬂuorescein-labelled p63TET peptide (0.1, 0.5, 1 and 2 mM) used in anisotropy studies in physiological ionic strength (PIS) buffer, the higher the relative amount of tetrameric p63TET and the lower the afﬁnity to S100A4 (a). The stoichiometry of binding was determined by titrating S100A4 into 0.5 mM labelled p63TET (b). In a buffer containing 1 mM ethylene-diaminetetraacetic acid (EDTA) instead of 10 mM CaCl2, the binding of S100A4 to p63TET is completely inhibited (c). Calcium activated the binding when titrated to S100A4 and p63TET in a buffer without calcium (d).

Table 2 Ca2 þ activation of binding of S100 proteins to p63TET ing the overall structure of the tetramer, in addition to the canonical p53 tetramerization domain motif (Joerger et al., Protein EC (mM) 50 2009). This additional C-terminal helix is conserved in p63 S100A1 46.6±7.5 and stabilizes the p63 tetramer (Joerger et al., 2009). S100A2 48.7±0.7 Consequently, we assumed that this helical region also S100A4 5.4±1.1 influences the interaction of p63 and p73 with S100 S100A6 19.5±2.5 S100B 277.7±56.1 proteins. As we were not able to purify the p63 tetramerization domain variant (residues 356–411) containing this additional structural element after labelling with binding of S100 proteins to the monomer. In contrast, fluorescein, we decided to study the interaction of S100 S100B did not interact tightly with p63TET: the elution proteins to naturally occurring isoforms containing the full- profile showed two separate peaks for S100B length tetramerization domain instead. We, therefore, (trB17.5 min) and p63TET (trB20 min), and no shift performed a recombinant protein expression screening of in tr for S100B. In addition, the MW for the elution peak a series of p63 and p73 isoforms (Brandt et al., 2009) and of S100B did not change on addition of p63TET were able to express DNp63b and DNp73b in amounts (Table 3). The results were confirmed by analytical large enough to perform biophysical experiments. SEC, in which the eluted peaks for S100 proteins were DNp63b is a naturally occurring splice isoform collected and analyzed by sodium dodecyl sulfate– without the N-terminal transactivation domain and the polyacrylamide gel electrophoresis. S100A1, S100A2, C-terminal SAM domain. It is the most effective DN- S100A4 and S100A6 co-eluted with p63TET but S100B splice isoform of p63 for activation of gene expression did not (data not shown). (Boldrup et al., 2009). We labelled DNp63b with a C- Simultaneously performed structural studies on the terminal FlAsH-Tag to perform fluorescence-monitored tetramerization domain of p73 revealed that it contains a AUC. The sedimentation velocity ultracentrifugation of C-terminal helical extension, which is essential for stabiliz- DNp63b-FlAsH resulted in three peaks corresponding

Oncogene Interaction of S100 proteins with p63 and p73 J van Dieck et al 2027 Table 3 Complex formation of S100 proteins and p63TET determined by SEC-MALS (kDa) S100A1 S100A2 S100A4 S100A6 S100B p63TET

Mw 21.1±0.8 22.9±0.3 24.6±0.2 20.0±0.2 21.7±0.1 14.9±1.1 M(theo.)a 21.4 22.0 23.5 20.4 21.4 16.8 Mw þ p63TET 22.8±0.3 26.0±0.3 27.0±1.8 24.6±0.7 21.6±0.6 M(theo.)b 25.3 26.2 27.7 24.6 25.9

Abbreviations: SEC, size-exclusion chromatography; MALS, multi-angle light scattering. a Theoretical Mw of S100 dimer and p63TET tetramer determined by MALDI-TOF. b Theoretical Mw of a complex of one S100 dimer with one p63TET monomer. to monomeric, dimeric and tetrameric protein with We also studied the interactions of S100 proteins with sedimentation coefficients of B1.2, B1.8 and B3.0 S. the splice isoform DNp73b-FlAsH in fluorescence- On addition of 20 mM S100 protein, the sedimentation monitored AUC experiments. As for DNp63b-FlAsH, profiles changed. An increase in sedimentation coeffi- we obtained three individual peaks for DNp73b-FlAsH cient for the labelled DNp63b-FlAsH species would in the sedimentation experiments, with sedimentation indicate a higher molecular mass, thus formation of a coefficients of B1.4, B2.0 and B3.1 S corresponding to complex with S100 proteins. In the presence of S100A1, monomeric, dimeric and tetrameric DNp73b-FlAsH. The the peak for the p63 tetramer had a lower intensity than largest changes in sedimentation coefficient for tetra- that of the peak for dimeric p63, indicating a disruption meric DNp73b-FlAsH was observed for S100A2 of the DNp63b-FlAsH tetramer into dimers. In the (B4.0 S) and S100A6 (B3.7 S), while smaller shifts were profiles for S100A2 and S100B, a fourth peak appeared observed for S100A1 (B3.3 S), S100A4 (B3.5 S) and with a sedimentation coefficient of B4.0 S, indicating S100B (B3.3 S), indicating a weaker binding of these binding of these S100 proteins to tetrameric DNp63b- three proteins to tetrameric DNp73b-FlAsH. We also FlAsH. In the presence of S100A6, an additional peak observed a small shift in sedimentation coefficient for of B3.5 S appeared in the form of a broad shoulder of dimeric DNp73b-FlAsH on addition of S100A1, the tetramer peak at B3.0 S, also indicating binding of S100A4, S100A6 and S100B to B2.2 S, indicating a S100A6 to tetrameric DNp63b. For all S100 proteins, complex of these S100 proteins with dimeric DNp73b- the peaks for the DNp63b-FlAsH monomer and dimer FlAsH. We did not detect any sedimentation shifts shifted to B1.4 S and B2.2 S, indicating binding of S100 caused by a binding event to monomeric p73 (Figure 5a). proteins to p63 monomers and dimers (Figure 3a). Analytical SEC for DNp73b revealed a co-elution with We then confirmed the binding of S100 proteins to S100A2 and S100A6, the two proteins that gave the tetrameric DNp63b in SEC experiments. S100A2, largest shifts for tetrameric p73 in fluorescence AUC. In S100A6 and S100B co-eluted with DNp63b whereas contrast, S100A1, S100A2, S100A4 and S100B did not S100A1 and S100A4 did not (Figure 3b). co-elute with DNp73b, suggesting that there was no stable complex formation of these S100 proteins with tetrameric DNp73b (Figure 5b). S100 proteins interact with the tetramerization domain of p73 in different oligomeric states We performed fluorescence anisotropy titration with S100 proteins interact with the transactivation domains S100 proteins with recombinant p73TET(351–383), of p63 and p73 which was fluorescein labelled at the N-terminus. As We previously showed that S100 proteins interact with for p63TET, we found Kd values in the low nanomolar the transactivation domain of p53 in vitro (van Dieck range for S100A2, S100A4 and S100B (Table 1). It was et al., 2009). Therefore, we synthesized and purified not possible to obtain Kd values for S100A1 and S100B N-terminal peptides of p63 (p63N; residues 5–35) and because of a very low overall change in anisotropy for p73 (p73N; residues 10–40) with a C-terminal Lys- these proteins (data not shown). We then expressed the methoxycoumarin label to study in fluorescence aniso- recombinant p73TET-long (residues 351–399) compris- tropy titration experiments whether S100 proteins ing the full-length p73 tetramerization domain (Joerger interact with the transactivation domains of the p53 et al., 2009) and labelled it with fluorescein at the homologs as well. For p63N, we observed high-affinity N-terminus to perform further fluorescence anisotropy binding, with Kd values in the nanomolar range for titrations. The overall change in the anisotropy experi- S100A1, S100A2, S100A4 and S100A6, whereas S100B ments was very low. Nevertheless, it was possible to bound B100 times more weakly (Table 4 and Figure 6). obtain binding curves for S100A4 and S100A6 (Fig- We also observed binding of S100 proteins to p73N, ure 4). The affinity of p73TET-long was B7 times lower but the affinity was up to three orders of magnitude for S100A6 and B10 times lower for S100A4 than the weaker than for p63N, with Kd values between 25 and values measured for p73TET (Table 1) but still more 50 mM for S100A1, S100A2, S100A4 and S100A6. than 20 times higher than for p53TET. Owing to the low We did not detect an interaction between S100B and change in anisotropy, we could not determine a binding p73N (Table 4 and Figure 6). No binding was observed curve for S100A1, S100A2 and S100B. The binding to for p63N and p73N in a buffer without calcium (data p73TET was calcium dependent (data not shown). not shown).

Oncogene Interaction of S100 proteins with p63 and p73 J van Dieck et al 2028

Figure 2 Analytical SEC of S100 proteins with p63TET(359–390). 100 ml of S100 protein (50–150 mM) were injected in the absence (black) or presence of p63TET (red) in a ratio of one dimer S100 per monomer p63TET. In contrast to all other S100 proteins, the tr of S100B did not change on addition of p63TET.

Discussion suggesting that the p53 homologs could be the preferred targets. In particular, S100A1, S100A2, S100A4 and In this study, we showed that S100 proteins interact with S100A6 showed a higher affinity for p63TET and the p53 homologs p63 and p73 in vitro. We characterized p73TET than for p53(293–393). In contrast, S100B the binding biophysically by fluorescence anisotropy was the strongest binder to p53(293–393) (Fernandez- titrations, analytical SEC and AUC. We observed a Fernandez et al., 2005, 2008) but the weakest binder to higher affinity of S100A proteins for the tetramerization p63TET (Table 1), implying that p53 is a preferred domains of p63/p73 than for that of p53 (Table 1), target of S100B. Further, we noticed that the binding of

Oncogene Interaction of S100 proteins with p63 and p73 J van Dieck et al 2029

Figure 3 Binding of S100 proteins to DNp63b. Analytical AUC of 200 nM DNp63b-FlAsH in the absence or presence of 20 mM S100 protein. Differences in the sedimentation proﬁles imply differential binding to monomeric, dimeric and tetrameric DNp63b by S100 proteins (a). Co-elution of tetrameric DNp63b with S100 proteins in analytical SEC (b); 1: DNp63b þ S100A1, 2: DNp63b þ S100A2, 3: DNp63b þ S100A4, 4: DNp63b þ S100A6, 5: DNp63b þ S100B.

Oncogene Interaction of S100 proteins with p63 and p73 J van Dieck et al 2030

Figure 4 Fluorescence anisotropy titration of S100 proteins to p73TET. Experiments were performed in physiological ionic strength (PIS) buffer with 0.1 mM of ﬂuorescein-labelled p73TET(351–388) or p73TET-long(351–399).

S100 proteins to p63/p73 depended on the length of the S100B with the DNA-binding domain or the linker construct used, and we observed different binding region in DNp63b. properties for individual S100 proteins. Interestingly, presence of the additional C-terminal S100A1 disrupted tetrameric p63TET in analytical helix in p73TET-long, reduced the afﬁnity to S100 SEC experiments (Figure 2 and Table 3) and also proteins compared with p73TET (Table 1). Again, this disrupted tetrameric DNp63b-FlAsH into dimers (Fig- could be the result of shifts in the equilibrium between ure 3). We could not detect binding of S100A1 to tetramers, dimers and monomers or differences in the tetrameric DNp63b or DNp73b in analytical SEC surface complementarity of the different species. We experiments (Figures 3 and 5), but we observed a small could not detect binding of S100 proteins to monomeric change in the sedimentation coefﬁcient for tetrameric DNp73b-FlAsH in the sedimentation velocity AUC DNp63b in AUC experiments on addition of S100A1, experiments. The inability of S100 proteins to interact indicative of weak binding. The results imply that with monomers of longer p73 variants comprising the S100A1 binds tightly to lower oligomeric species of full tetramerization domain, could be the reason for the p63/p73 but only weakly to tetrameric species compared very low overall anisotropy shift observed in titration with other S100 proteins. We previously reported similar experiments with labelled p73TET-long compared with results for S100A1 and p53, in which we showed that p73TET (Figure 4), because the change of anisotropy is S100A1 bound tightly to the monomeric mutant L344P dependent on the mass difference of bound and of p53(293–393) but not to tetrameric full-length p53 in unbound state of the labelled component. Further, we analytical SEC (van Dieck et al., 2009). S100A2 and could not detect binding of S100B to tetrameric S100A6 bound monomeric, dimeric as well as tetrameric DNp73b, which supports our theory that its primary p63 and dimeric and tetrameric p73 in AUC experiments target within the family is p53 and not its homologs. (Figures 3 and 5). We previously discovered similar We previously suggested a binding model for the binding properties for p53 (van Dieck et al., 2009). interactions of S100 proteins with p53 based on in vitro Interestingly, we observed that S100B bound relatively and in vivo data, in which the effect of S100 proteins on weakly to p63TET (Table 1), and did not form a stable p53 was highly dependent on the concentration of each complex with a p63TET monomer in analytical SEC, component in the cell (van Dieck et al., 2009). The unlike S100A1, S100A2, S100A4 and S100A6 (Figure 2). model also seems to be valid for p63 and p73, because By contrast, S100B was able to bind the tetrameric S100 proteins bind p63 and p73 in their different splice-isoform DNp63b in AUC and SEC experiments. oligomeric states. As S100 proteins potentially bind The DNp63b isoform contains the full tetramerization the TET domains as well as the N-terminus of p63 and domain, including the additional C-terminal helix, p73, the effect on different isoforms of the p53 family which stabilizes the tetrameric state, whereas p63TET may also vary. For example, all DN splice isoforms of more readily dissociates into lower oligomeric species p63 and p73 can still interact with S100 proteins through (Joerger et al., 2009). Hence, differences in S100B their TET domain, with either an inhibitory effect binding could be caused by differences in the oligome- disrupting the tetramer or a stimulating effect by rization equilibria or in the molecular surface properties binding to the intact tetramer. This effect of differential of the full-length p63 tetramerization domain because binding has already been suggested for p53 (van Dieck of the C-terminal helix, in analogy to the structures of et al., 2009). In addition, the binding of S100 to TA- different p73 tetramerization domain variants (Joerger isoforms could also differ depending on the splice- et al., 2009). There is, of course, also the possibility that variation at the C-terminus of p63 and p73. The the observed differences are caused by interaction of C-terminal inhibitory domain of p63 is able to bind to

Oncogene Interaction of S100 proteins with p63 and p73 J van Dieck et al 2031

Figure 5 Binding of S100 proteins to DNp73b. Analytical AUC of 100 nM DNp73b-FlAsH in the absence or presence of 40 mM S100 protein. Differences in the sedimentation proﬁles imply differential binding to dimeric and tetrameric DNp73b by S100 proteins (a). Co-elution of tetrameric DNp73b with S100 proteins in analytical SEC (b); 1: DNp73b þ S100A1, 2: DNp73b þ S100A2, 3: DNp73b þ S100A4, 4: DNp73b þ S100A6, 5: DNp73b þ S100B. its N-terminal transactivation domain, and thus may and g variant. Although the C-terminal SAM domain is interfere with the binding to S100 proteins. Conse- thought to be monomeric in p63 and p73, it has been quently, the binding of the a form may differ from the b shown to be involved in homodimerization in other

Oncogene Interaction of S100 proteins with p63 and p73 J van Dieck et al 2032 Table 4 Binding of S100 proteins to p63/p73 N-terminal peptides in proteins to the TET domain, altering the effect on the fluorescence anisotropy titrations overall oligomerization state. The interaction network is further complicated by the fact that different splice- Protein Kd (mM) isoforms (for example, DN- and TA-isoforms) can p63 N-term (5–35) p73 N-Term (10–40) heterotetramerize (Deyoung and Ellisen, 2007), provid- ing even more diverse interaction combinations for S100 S100A1 0.09±0.02 27.03±9.35 S100A2 0.10±0.02 28.20±1.92 proteins. On the basis of the binding properties of S100 S100A4 0.14±0.03 49.97±6.17 proteins to individual domains of p63 and p73 portrayed S100A6a 0.02±0.005 25.35±3.82 here, it would be interesting to study the effects of S100 S100Bb 13.7±2.76 n.q. proteins on all of the B 50 isoforms of the p53 family as well as their known heterotetrameric combinations. But, Abbreviation: n.q., not quantifiable. aMeasured at 0.1 mM peptide concentration. given the complexity of possible interactions, it is clear bMeasured at 0.5 mM peptide concentration. that the binding of S100 proteins will be difficult to study in vivo. Transactivation assays with overexpressed (and therefore mostly tetrameric) proteins will most likely only detect binding of S100 proteins to tetrameric p63 and p73 but not binding to lower oligomeric forms, which are also present at physiological concentrations of the proteins. Knock-out of a particular S100 protein in animal models may not necessarily show conclusive effects because of possible overlapping functions of the S100 family members in their binding to the p53 family. The interactions of S100 proteins with p63 and p73 are likely to be physiologically relevant. The interaction of several S100 proteins with p53 has been firmly established in vitro and in vivo in several studies (Baudier et al., 1992; Scotto et al., 1998, 1999; Grigorian et al., 2001; Lin et al., 2001, 2004; Fernandez-Fernandez et al., 2005, 2008; Mueller et al., 2005; Slomnicki et al.,2009; van Dieck et al., 2009). It is, therefore, reasonable to assume that the interactions with the paralogs p63 and p73 also occur in vivo. S100 proteins have been shown to be involved in differentiation processes and are usually expressed in a tissue-specific manner, and it is established that p63 and p73 are also involved in developmental processes (Fano et al., 1995; Zimmer et al., 1995; Ilg et al., 1996; Mills et al., 1999; Yang et al., 1999, 2000; Chan et al., 2003; Donato, 2003; Buckiova and Syka, 2009; Vigliano et al., 2009). S100A2 has been shown to be a target gene of p63 and p73 (Lapi et al., 2006; Kirschner et al., 2008). An interaction of the two proteins with S100A2 could be part of a feedback loop mechanism. Further, this interaction may be relevant in the development of cancer, because S100A2 and p63 have been linked to squamous cell carcinomas (Hibi et al., 2000; Lauriola et al., 2000; Yamaguchi et al., 2000; Pelosi et al.,2002; Pruneri et al., 2002; Dotto and Glusac, 2006). The interaction of S100 proteins with the p63 and p73 tetramerization domain could constitute a fundamental regulatory mechanism, because almost all known splice isoforms of the p53 family contain a functional tetramerization domain, which could interact with Figure 6 Binding of S100 proteins in fluorescence anisotropy titrations to p63/p73 N-terminal peptides. S100 proteins with the members of the S100 protein family in developmental exception of S100B bind to p63N(3–35) with a nanomolar affinity as well as disease-related processes. As S100 proteins (a). The binding to p73N(10–40) has a lower affinity (b). interact with the p53 family in a calcium-dependent manner, our findings may contribute to elucidate, how calcium as a second messenger is involved in proteins involved in development (Deyoung and Ellisen, developmental processes (Langenbacher and Chen, 2007). Consequently, splice isoforms with or without 2008; Whitaker, 2008), linking calcium signalling to this domain may respond differently to binding of S100 p63/p73 through the S100 protein family.

Oncogene Interaction of S100 proteins with p63 and p73 J van Dieck et al 2033 Materials and methods described earlier (Fernandez-Fernandez et al., 2005, 2008; van Dieck et al., 2009). The concentration of labelled peptide was Plasmids and protein expression usually 0.1 mM for p63TET and p73TET peptides, and 0.2 mM for The expression and purification of recombinant human S100A1, p63N and p73N unless stated otherwise. The titration data were S100A2, S100A4, S100A6 and S100B in Escherichia coli C41 cells fitted to a quadratic fitting equation for a single-site binding were as described, using the same pRSET plasmids (Fernandez- model, when necessary with the addition of a term accounting Fernandez et al., 2005, 2008). The complementary DNA of for linear drift (Teufel et al., 2007). Each experiment was p63TET, p73TET and p73TET-long were cloned into pRSET performed at least three times, including measurements with vectors with an additional N-terminal His-tag, a lipoyl domain different protein and peptide batches. Calcium titrations were and a thrombin cleavage site. The recombinant proteins were performed with 10 mM S100 protein (100 mM for S100B) and expressed in E. coli BL21 cells and purified through a metal 0.1 mM fluorescein-labelled p63TET in 25 mM Tris pH 7.4, (Nickel) affinity chromatography step. The His-tag and the lipoyl 99.2 mM NaCl and 1 mM dithiothreitol in the cuvette, and CaCl2 domain were cut off with thrombin (Sigma, Dorset, UK) ranging from 250 mM to 100 mM in the same buffer was used as overnight, and the p63 or p73 proteins were further purified titrant. The concentration of calcium at 50% of maximal using a second Ni-affinity column, anion exchange chromato- excitation (EC50) was determined with either a sigmoidal fit or a graphy and gel filtration (Superdex 75, GE Healthcare, Little single-site fit, depending on the shape of the activation curve. Chalfont, UK). DNp63b and DNp73b with and without a tetracysteine FlAsH-Tag motif were cloned according to Analytical ultracentrifugation the InFusion cloning system protocol (Clontech, Saint-Germain- Fluorescence-monitored AUC sedimentation velocity experi- en-Laye, France) into a pET24aHLTEV vector containing an ments were performed in a Beckman-Coulter Optima XL-1 N-terminal 6xHis tag, followed by a lipoyl domain and a TEV (Beckman-Coulter, Brea, CA, USA) with an Aviv fluorescence protease cleavage site. The proteins were expressed in E. coli B834 detection system (MacGregor et al., 2004). Proteins were buffer cells at 18 1C for 16–20 h and purified by Ni-affinity chromato- exchanged into 20 mM Tris 7.4, 150 mM NaCl, 10 mM CaCl , graphy and gel filtration. 2 10% glycerol, 15 mM b-mercaptoethanol and 0.2% bovine serum albumin. The concentration of DNp63b-FlAsH was 200 nM and Peptide synthesis 100 nM for DNp73b-FlAsH in the experiments. The concentra- The fluorescein-labelled p63TET(359–390) was synthesized tion of S100 proteins was 20 or 40 mM. The sample volume was according to a solid-phase Fmoc/tBu strategy on a rinkamide 100 ml. The experiments were performed as described earlier resin. Before deprotection and cleavage from the resin, the (Rajagopalan et al., 2008; van Dieck et al.,2009),andthe peptides were labelled with 50 and 60 carboxyfluorescein sedimentation coefficients were determined using the SedFit succimidyl ester at the N-terminus. The p63TET(359–390) software (www.analyticalultracentrifugation.com) (Schuck, peptide was then high performance liquid chromatography- 2000). purified on a preparative C18 column with a gradient from 20– 80% solvent B (solvent A: 0.1% TFA in water, solvent B: Analytical SEC 0.1% TFA, 5% water in acetonitrile). N-terminal peptides Analytical SEC with recombinant p63TET was performed on p63N(5–35) and p73N(10–40) were synthesized with a Lys- a GE Healthcare SuperdexTm 75 analytical gelfiltration column methoxycoumarin at the C-terminus and purified as described under the same conditions as described earlier (van Dieck et al. earlier (Burge , 2009). et al., 2009). Proteins were buffer exchanged in physiological ionic strength buffer and 100 ml of protein in concentrations Protein labelling between 50 and 250 mM were injected. The eluted peaks were p73TET and p73TET-long were labelled at the N-terminus either collected with a fraction collector and analyzed by SDS– with an equimolar amount of 50 (and 60) carboxyfluorescein polyacrylamide gel electrophoresis or the weight-averaged succimidyl ester (Molecular Probes, Eugene, OR, USA). The molar mass (MW) was determined on-line by multi-angle fluorescent dye was dissolved in dimethylsulfoxide (B40 mM), light scattering (Wyatt Technology, Santa Barbara, CA, USA) and an equimolar amount was added to the protein in a buffer (van Dieck et al., 2009). SEC with DNp63b and DNp73b of 20 mM potassium phosphate pH 7.0, 20 mM NaCl. The were performed on a SuperoseTm 6 10/300GL column (GE labelling reaction was monitored by matrix-assisted laser Healthcare) with a flow of 0.7 ml/min in a buffer of 20 mM desorption/ionization–time of flight and quenched after 2–3 h Tris 7.4, 150 mM NaCl, 10 mM CaCl2 10% glycerol, 1 mM with 1 M TRIS pH 7.0 buffer in excess. The peptide was then dithiothreitol. 100 mlof10mM DNp63b or DNp73b were purified on an analytical C18-high performance liquid injected in the presence of excess S100 protein (50–250 mM). chromatography column with a gradient from 40–65% solvent The eluted DNp63b or DNp73b was collected with a fraction B. FlAsH-tagged DNp63b and DNp73b were labelled with 1.5 collector, concentrated with a centrifugational filter and equivalents of FlAsH-EDT2 (Invitrogen, Paisley, UK) at 8 1C analyzed by SDS–polyacrylamide gel electrophoresis to detect for 2.5 h. co-elution with S100 proteins.

Fluorescence anisotropy The experiments were carried out in a physiological ionic Conflict of interest strength buffer (PIS) containing 25 mM Tris pH 7.4, 10 mM CaCl2, 99.2 mM NaCl and 1 mM dithiothreitol or 25 mM Tris pH The authors declare no conflict of interest. 7.4, o0.1 mM CaCl2,1mM ethylene-diaminetetraacetic acid, 99.2 mM NaCl and 1 mM dithiothreitol (PIS-Ca). Fluorescence anisotropy titrations were performed on a Perkin-Elmer LS-55 Acknowledgements (Perkin Elmer, Waltham, MA, USA) and a Cary Eclipse Varian fluorescence spectrophotometer (Varian, Walnut, CA, USA) We thank Karoly von Glos for technical assistance with equipped with a Hamilton microlab M dispenser (Hamilton peptide synthesis and Dr Maria R Fernandez-Fernandez for Robotics, Birmingham, UK) under the same conditions as critical reading of the paper.

Oncogene Interaction of S100 proteins with p63 and p73 J van Dieck et al 2034 References

Baudier J, Delphin C, Grunwald D, Khochbin S, Lawrence JJ. (1992). p53 homologues during keratinocyte differentiation. Oncogene 25: Characterization of the tumor suppressor protein p53 as a protein 3628–3637. kinase C substrate and a S100b-binding protein. Proc Natl Acad Sci Lauriola L, Michetti F, Maggiano N, Galli J, Cadoni G, Schafer BW USA 89: 11627–11631. et al. (2000). Prognostic significance of the Ca(2+) binding protein Boldrup L, Coates PJ, Gu X, Nylander K. (2009). DeltaNp63 S100A2 in laryngeal squamous-cell carcinoma. Int J Cancer 89: isoforms differentially regulate gene expression in squamous cell 345–349. carcinoma: identification of Cox-2 as a novel p63 target. J Pathol Lin J, Blake M, Tang C, Zimmer D, Rustandi RR, Weber DJ et al. 218: 428–436. (2001). Inhibition of p53 transcriptional activity by the S100B Brandt T, Petrovich M, Joerger AC, Veprintsev DB. (2009). calcium-binding protein. J Biol Chem 276: 35037–35041. Conservation of DNA-binding specificity and oligomerisation Lin J, Yang Q, Yan Z, Markowitz J, Wilder PT, Carrier F et al. (2004). properties within the p53 family. BMC Genomics 10: 628. Inhibiting S100B restores p53 levels in primary malignant melano- Buckiova D, Syka J. (2009). Calbindin and S100 protein expression in ma cancer cells. J Biol Chem 279: 34071–34077. the developing inner ear in mice. J Comp Neurol 513: 469–482. MacGregor IK, Anderson AL, Laue TM. (2004). Fluorescence Burge S, Teufel DP, Townsley FM, Freund SM, Bycroft M, Fersht detection for the XLI analytical ultracentrifuge. Biophys Chem AR. (2009). Molecular basis of the interactions between the p73 N 108: 165–185. terminus and p300: effects on transactivation and modulation by Melino G, Lu X, Gasco M, Crook T, Knight RA. (2003). Functional phosphorylation. Proc Natl Acad Sci USA 106: 3142–3147. regulation of p73 and p63: development and cancer. Trends Biochem Chan WY, Xia CL, Dong DC, Heizmann CW, Yew DT. (2003). Sci 28: 663–670. Differential expression of S100 proteins in the developing Mills AA, Zheng B, Wang XJ, Vogel H, Roop DR, Bradley A. (1999). human hippocampus and temporal cortex. Microsc Res Tech 60: p63 is a p53 homologue required for limb and epidermal 600–613. morphogenesis. Nature 398: 708–713. Danilova N, Sakamoto KM, Lin S. (2008). p53 family in development. Moll UM, Slade N. (2004). p63 and p73: roles in development and Mech Dev 125: 919–931. tumor formation. Mol Cancer Res 2: 371–386. Dean B, Gray L, Scarr E. (2006). Regionally specific changes in levels Mrak RE, Griffinbc WS. (2001). The role of activated astrocytes and of cortical S100beta in bipolar 1 disorder but not schizophrenia. of the neurotrophic cytokine S100B in the pathogenesis of Aust N Z J Psychiatry 40: 217–224. Alzheimer’s disease. Neurobiol Aging 22: 915–922. Deyoung MP, Ellisen LW. (2007). p63 and p73 in human cancer: Mueller A, Schafer BW, Ferrari S, Weibel M, Makek M, Hochli M defining the network. Oncogene 26: 5169–5183. et al. (2005). The calcium-binding protein S100A2 interacts with Donato R. (2003). Intracellular and extracellular roles of S100 p53 and modulates its transcriptional activity. J Biol Chem 280: proteins. Microsc Res Tech 60: 540–551. 29186–29193. Dotto JE, Glusac EJ. (2006). p63 is a useful marker for cutaneous Pelosi G, Pasini F, Olsen Stenholm C, Pastorino U, Maisonneuve P, spindle cell squamous cell carcinoma. J Cutan Pathol 33: 413–417. Sonzogni A et al. (2002). p63 immunoreactivity in lung cancer: yet Fano G, Biocca S, Fulle S, Mariggio MA, Belia S, Calissano P. (1995). another player in the development of squamous cell carcinomas? The S-100: a protein family in search of a function. Prog Neurobiol J Pathol 198: 100–109. 46: 71–82. Pruneri G, Pignataro L, Manzotti M, Carboni N, Ronchetti D, Neri A Fernandez-Fernandez MR, Rutherford TJ, Fersht AR. (2008). et al. (2002). p63 in laryngeal squamous cell carcinoma: evidence for Members of the S100 family bind p53 in two distinct ways. Protein a role of TA-p63 down-regulation in tumorigenesis and lack of Sci 17: 1663–1670. prognostic implications of p63 immunoreactivity. Lab Invest 82: Fernandez-Fernandez MR, Veprintsev DB, Fersht AR. (2005). 1327–1334. Proteins of the S100 family regulate the oligomerization of p53 Rajagopalan S, Jaulent AM, Wells M, Veprintsev DB, tumor suppressor. Proc Natl Acad Sci USA 102: 4735–4740. Fersht AR. (2008). 14–3–3 activation of DNA binding of p53 Finlan LE, Hupp TR. (2007). p63: the phantom of the tumor by enhancing its association into tetramers. Nucleic Acids Res 36: suppressor. Cell Cycle 6: 1062–1071. 5983–5991. Grigorian M, Andresen S, Tulchinsky E, Kriajevska M, Carlberg C, Rosenbluth JM, Pietenpol JA. (2008). The jury is in: p73 is a tumor Kruse C et al. (2001). Tumor suppressor p53 protein is a new target suppressor after all. Genes Dev 22: 2591–2595. for the metastasis-associated Mts1/S100A4 protein: functional Rustandi RR, Baldisseri DM, Weber DJ. (2000). Structure of the consequences of their interaction. J Biol Chem 276: 22699–22708. negative regulatory domain of p53 bound to S100B(betabeta). Hibi K, Trink B, Patturajan M, Westra WH, Caballero OL, Hill DE Nat Struct Biol 7: 570–574. et al. (2000). AIS is an oncogene amplified in squamous cell Salama I, Malone PS, Mihaimeed F, Jones JL. (2008). A review of the carcinoma. Proc Natl Acad Sci USA 97: 5462–5467. S100 proteins in cancer. Eur J Surg Oncol 34: 357–364. Ilg EC, Schafer BW, Heizmann CW. (1996). Expression pattern of Santamaria-Kisiel L, Rintala-Dempsey AC, Shaw GS. (2006). S100 calcium-binding proteins in human tumors. Int J Cancer 68: Calcium-dependent and -independent interactions of the S100 325–332. protein family. Biochem J 396: 201–214. Joerger AC, Fersht AR. (2007). Structure-function-rescue: the diverse Schuck P. (2000). Size-distribution analysis of macromolecules by nature of common p53 cancer mutants. Oncogene 26: 2226–2242. sedimentation velocity ultracentrifugation and lamm equation Joerger AC, Rajagopalan S, Natan E, Veprintsev DB, Robinson CV, modeling. Biophys J 78: 1606–1619. Fersht AR. (2009). Structural evolution of p53, p63, and p73: Scotto C, Deloulme JC, Rousseau D, Chambaz E, Baudier J. (1998). implication for heterotetramer formation. Proc Natl Acad Sci USA Calcium and S100B regulation of p53-dependent cell growth arrest 106: 17705–17710. and apoptosis. Mol Cell Biol 18: 4272–4281. Kirschner RD, Sanger K, Muller GA, Engeland K. (2008). Transcrip- Scotto C, Delphin C, Deloulme JC, Baudier J. (1999). Concerted tional activation of the tumor suppressor and differentiation regulation of wild-type p53 nuclear accumulation and activation by gene S100A2 by a novel p63-binding site. Nucleic Acids Res 36: S100B and calcium-dependent protein kinase C. Mol Cell Biol 19: 2969–2980. 7168–7180. Langenbacher A, Chen JN. (2008). Calcium signaling: a common Scoumanne A, Harms KL, Chen X. (2005). Structural basis for thread in vertebrate left-right axis development. Dev Dyn 237: gene activation by p53 family members. Cancer Biol Ther 4: 3491–3496. 1178–1185. Lapi E, Iovino A, Fontemaggi G, Soliera AR, Iacovelli S, Sacchi A Sedaghat F, Notopoulos A. (2008). S100 protein family and its et al. (2006). S100A2 gene is a direct transcriptional target of application in clinical practice. Hippokratia 12: 198–204.

Oncogene Interaction of S100 proteins with p63 and p73 J van Dieck et al 2035 Serber Z, Lai HC, Yang A, Ou HD, Sigal MS, Kelly AE et al. (2002). Whitaker M. (2008). Calcium signalling in early embryos. Philos Trans A C-terminal inhibitory domain controls the activity of p63 by an R Soc Lond B Biol Sci 363: 1401–1418. intramolecular mechanism. Mol Cell Biol 22: 8601–8611. Yamaguchi K, Wu L, Caballero OL, Hibi K, Trink B, Resto V et al. Slomnicki LP, Nawrot B, Lesniak W. (2009). S100A6 binds p53 and (2000). Frequent gain of the p40/p51/p63 gene locus in primary affects its activity. Int J Biochem Cell Biol 41: 784–790. head and neck squamous cell carcinoma. Int J Cancer 86: Teufel DP, Freund SM, Bycroft M, Fersht AR. (2007). Four domains 684–689. of p300 each bind tightly to a sequence spanning both transactiva- Yang A, Schweitzer R, Sun D, Kaghad M, Walker N, Bronson RT tion subdomains of p53. Proc Natl Acad Sci USA 104: 7009–7014. et al. (1999). p63 is essential for regenerative proliferation in limb, van Dieck J, Fernandez-Fernandez MR, Veprintsev DB, Fersht AR. craniofacial and epithelial development. Nature 398: 714–718. (2009). Modulation of the oligomerization state of p53 by Yang A, Walker N, Bronson R, Kaghad M, Oosterwegel M, Bonnin J differential binding of proteins of the S100 family to p53 monomers et al. (2000). p73-deficient mice have neurological, pheromonal and and tetramers. J Biol Chem 284: 13804–13811. inflammatory defects but lack spontaneous tumours. Nature 404: Vigliano FA, Bermudez R, Nieto JM, Quiroga MI. (2009). Develop- 99–103. ment of rodlet cells in the gut of turbot (Psetta maxima L.): Zimmer DB, Cornwall EH, Landar A, Song W. (1995). The S100 relationship between their morphology and S100 protein immuno- protein family: history, function, and expression. Brain Res Bull 37: reactivity. Fish Shellfish Immunol 26: 146–153. 417–429. Vilgelm A, El-Rifai W, Zaika A. (2008). Therapeutic prospects for p73 Zimmer DB, Wright Sadosky P, Weber DJ. (2003). Molecular and p63: rising from the shadow of p53. Drug Resist Update 11: mechanisms of S100-target protein interactions. Microsc Res Tech 152–163. 60: 552–559.

Oncogene