We also evaluated the effect of varying ionic str http
We also evaluated the effect of varying ionic strength. These experiments included previously characterized H3 (residues 1–21) and SNAIL (residues 1–9) peptides ,  to allow a comparative analysis with different ligands (Fig. 1A). As shown in Fig. 2 and listed in Table 2, it is evident that a high sensitivity to salt concentration is a feature shared by all these peptides in their association to LSD1. Specifically, binding of SNAIL1 is the least perturbed (15–20 times Kd decrease at 200 mM salt), while p53 seems most affected (75–80 times Kd decrease at 200 mM salt). Collectively, these experiments gave a clear indication on the features characterizing p53-CTD/LSD1-CoREST1 interaction: binding is mostly electrostatic, mediated by residues 379–388, and moderately affected by site-specific variations on these amino acids (Fig. 2 and Table 2).
To further probe whether the binding was due to non-specific interactions, a p53 peptide (residues 371–388) consisting of D- (rather than L-) amino BINA was used in our studies. Surprisingly, it turned out that this D-peptide does inhibit LSD1-CoREST1 equally well as the l-amino acid peptide (Table 1, Fig. 1B). This intriguing result was validated by fluorescence polarization using a SNAIL1 peptide (residues 1–9) conjugated with a specific fluorophore to assess competitive binding properties of the D- and L-p53-CTD peptides. We were able to confirm the binding constants obtained with the enzyme inhibition assays and we also verified that there is no significant difference between l-amino acid and d-amino acid ligands (Fig. 1B and Table 1).
We further analyzed the binding of the p53 peptides to LSD1 by including in our studies three mutants, which were recently discovered in patients suffering by a newly described genetic disease . Of interest, all three disease-associated mutations (E379K, D556G, and Y761H) affect charged or H-bonding side chains in the active site (Fig. 3). Although they do not perturb the protein conformation, they substantially impair catalysis, primarily by negatively affecting H3 binding. The affinity to p53-CTD turned out to be unmodified by Y761H and D556G mutations whereas charge-reverted E379K caused a significant 10-fold decrease in CTD binding (data shown in Fig. 4 and summarized in Table 3). Thus, it appears that the binding of p53-CTD to LSD1-CoREST1 is strongly affected by a localized mutation (E379K) that targets an active-site residue.
This pattern of observations resembles the case of polymyxins . These polycationic cyclic peptides comprise a cyclic heptapeptide loop with a DAB-DAB-Leu(or Phe)-Leu-DAB-DAB-Thr sequence that resembles the 381Lys-Lys-Leu-Met-Phe-Lys-Thr387 segment of p53 , . Polymyxins bind to LSD1 with much higher nanomolar affinity than p53-CTD. The tighter binding probably reflects the more elaborated and conformationally restrained cyclic structure of polymyxins compared to the highly flexible p53-CTD sequence , , . Bearing in mind these differences, it is notable that, as now found for p53-CTD, the affinity of polymyxins to LSD1 is greatly diminished by E379K . Polymyxins were found to bind at the entrance of the enzyme catalytic cleft where Glu379 is located (Fig. 3B). This region likely represents the site that electrostatically steers the H3 tail to favour substrate admission and ordered binding to the depth of the catalytic site where flavin-mediated demethylation takes place . In light of the observation that E379K specifically perturbs the binding of both the highly charged polymyxin and p53-CTD peptides, it seems that also p53-CTD exploits this negatively charged area on the active-site outer surface for its binding to LSD1 (Fig. 3).
Discussion The intrinsically disordered C-terminal domain of p53 is characterized by the presence of charge-rich sub-regions that are an ideal target for a large number of post-translational modifications by diverse enzymes, such as p300/CBP acetyltransferases, Mdm2 ubiquitin ligase and protein kinases , ,  (Fig. 1A). Consistently, alteration of the degree and combination of these modifications influences the binding of p53 to its regulators , . Furthermore, the positively charged CTD contributes to DNA binding, supposedly though electrostatic interactions between Lys, Arg and the phosphate backbone of DNA . Interestingly, a recently published work on the interaction between p53-CTD and the oncoprotein SET indicates this charge-dependency as “widespread regulatory mode” . In this context, our experiments were motivated by several published reports demonstrating a functional cross-talk between p53 and LSD1, whose molecular bases were suggested to involve a direct interaction between p53-CTD with the demethylase enzyme , , , , , . To tackle the issue of the biochemical nature of this association we have performed complementary assays which probe binding to LSD1 by enzyme activity, thermal shift, enzyme inhibition and binding assays using a 21-amino acid H3 tail substrate, and fluorescence polarization using a 9-amino acid binder. Our experiments demonstrate that LSD1-CoREST does not show a catalytic demethylase activity on p53-CTD (at least on Lys370) in vitro. On the other hand, they also coherently indicate that there is a micromolar-affinity association between the LSD1-CoREST1 and p53-CTD and that binding can be mostly ascribed to residues 379–388 of p53-CTD (Fig. 1A). Modifications of the Lys side chains on this CTD region reduce binding affinity, which is consistent with the idea that the interaction is mainly enforced by electrostatic attraction. The clear dependency of the affinity on the ionic strength further confirms this notion (Table 2).