Recognition of ASF1 by Using Hydrocarbon‐Constrained Peptides

Abstract Inhibiting the histone H3–ASF1 (anti‐silencing function 1) protein–protein interaction (PPI) represents a potential approach for treating numerous cancers. As an α‐helix‐mediated PPI, constraining the key histone H3 helix (residues 118–135) is a strategy through which chemical probes might be elaborated to test this hypothesis. In this work, variant H3118–135 peptides bearing pentenylglycine residues at the i and i+4 positions were constrained by olefin metathesis. Biophysical analyses revealed that promotion of a bioactive helical conformation depends on the position at which the constraint is introduced, but that the potency of binding towards ASF1 is unaffected by the constraint and instead that enthalpy–entropy compensation occurs.

Inhibiting the histone H3-ASF1 (anti-silencing function1)p rotein-protein interaction (PPI) represents ap otential approach for treating numerousc ancers. As an a-helix-mediated PPI, constraining the key histoneH 3h elix (residues1 18-135) is a strategyt hrough which chemical probesm ight be elaborated to test this hypothesis. In this work, variant H3 118-135 peptides bearing pentenylglycine residues at the i and i+ +4p ositions were constrained by olefin metathesis. Biophysical analyses revealed that promotion of ab ioactive helical conformation dependso nt he positiona tw hich the constraint is introduced, but that the potency of bindingt owards ASF1 is unaffected by the constraint and instead that enthalpy-entropy compensation occurs.
Histone chaperones regulate the association of basic histone proteins with DNA, thereby permitting nucleosome assembly in an ordered and controlled manner. [28][29][30][31][32][33] ASF1 is ah ighly conserved histonec haperone that is involved in both histone H3-H4 handling and buffering. [34][35][36][37][38] It has been shown to play ak ey role in the developmenta nd progression of some cancers;h ence, it is ap otential target for chemicalp robes and drug discovery. [39][40][41] The interaction between ASF1 and the H3 and H4 histone proteins forms aA SF1-(H3-H4)c omplex that preventst he formationo ft he histone H3-H4 tetramer and shields H3-H4 dimers from unfavourable interactions.R e-establishmento ft he tetramer was proposed to be the key element for the formation of the nucleosome ( Figure 1A). [42] The ASF1 protein comprises ac onserved N-terminal domain of 156 amino acids, which is essential for its function in vivo, and a divergentu nstructured C-terminal domain, which is not considered necessary forf unction. [37,43] Its structure comprises an elongated b sandwich core with three a-helices in the loops between the b-strands ( Figure 1B). The contacts between H3 and ASF1 are extensive and result in ab uried surface area of 909 2 .T he histone H3 binding site is located in the concave face of ASF1 ( Figure 1B)a nd involves b-strands b3, b4a nd b6-9. [37,43,44] The main interactions occur through the C-terminal helix of H3 (residues 122-134), where the key residues Leu126 and Ile130 form ah ydrophobic clamp with the hydrophobic regiono fA SF1. Additionally,t here is an etwork of electrostatic interactions at the PPI interface, such as the salt bridge between Arg129f rom H3 and Asp54 from Asf1. [45] The ASF1-H3-H4 structure also shows extensive contacts between ASF1 and histoneH 4 [44] in two parts (not shown): the globular core of ASF1 interacts with the C-terminal tail of H4 to form as trand-swappedd imer,a nd the C-terminal tail of ASF1 binds to the histonefoldregion of histone H4.
We envisioned the C-terminal a-helix peptide of H3 as a templatef or the design of molecules ablet or ecognise ASF1. We used S-pentenylglycine rather than S-pentenylalanine, as the former is easier to synthesise and demonstrates comparable behaviour in biophysical analyses. [27] The sites to incorporate the mono-alkenyl-substituteda mino acids within the peptide sequence were selected by taking into account:1 )the requirementt oa ppropriatelyp osition the unnatural aminoa cids so as to constrain in am anner that promotesahelical conformation (i.e.,t he i and i+ +4p ositions);2 )the need to position the hydrocarbon bridge so as not to sterically occlude the "wild-type" interactions necessary for recognition.O nt his basis, two options were considered Met120/Ile124 and Asp123/Ala127. H3 118-135 ,t ogether with variantsb earing S-pentenylglycine in the identified positions were prepared by solidphase peptides ynthesis (see the Supporting Information), and the latter were crosslinkedb yo lefin metathesis to give H3 118-135(St120-124) and H3 118-135(St123-127) GCA (the GCA sequence was introducedf or future functionalisation, e.g.,c ell-penetrating sequences,f luorophores, etc. through the nucleophilic thiol of the cysteiner esidue). On-resin ring closure proceeded quantitatively in 4h.
The helical character of all three peptides was investigated by using circular dichroism (CD) in both 40 mm phosphate buffer and trifluoroethanol (TFE). In aqueous solvent, H3 118-135 and H3 118-135(St120-124) both gave CD spectra consistent with ap redominantly random-coil conformation (% helicities H3 118-135 = 15 %a nd H3 118-135(St120-124) = 20 %) ,w hereas in the presence of the helix-promoting TFE ( Figure S1 in the Supporting Information) [46][47] the CD spectra were indicative of am ore a-helicals ignature, thus indicatingt hat both possess sufficient conformational flexibility to access the helical conformation required for specific ASF1 binding. It is perhaps unsurprisingt hat constraining the peptide between residues 120 and1 24 did not promote ah elical conformation in H3 118-135(St120-124) given the observation from the H3/ASF1 NMR structure that the H3 helix is distorted/frayed at the Nterminus close to M120.I n contrast, CD analyses showed H3 118-135(St123-127) GCA to adopt a more helical conformation in aqueous solution by (% helicity = 29 %), as expected. The dataf or all three peptides in TFE (see the Supporting Information) demonstrate that each is capable of adopting ah elical conformation to ac omparable extent, and, that there is little difference between buffer and TFE for H3 118-135(St123-127) GCA;t his indicates that the sequence has intrinsically low helical propensity. (Figure 2).
In order to confirm the binding mode of the constrained peptides with ASF1, chemical-shift-perturbation studies were carriedo ut for all three peptides ( Figure S2) by using uniformly 15 Nl abelled ASF1A(1-156). The chemical-shift variation was mapped onto the protein structure of ASF1A-H3 (PDB ID:2 IIJ). All three peptides induced the highest values of chemical-shift variationa nd a" slow-exchange" regime for the ASF1 residues defining the already well characterised H3 binding site (V45-E51, V90-I97, R108-Y111,V 146-T147), [37,45] thus confirming the preservation of the specific binding mode for the constrained peptides. In addition, both H3 118-135 and H3 118-135(St120-124) exhibited chemical-shift variationso nt he opposite side of the protein surfacec orresponding to the Bdomain binding site (S59-F72); [48] these most probably correspond to nonspecific binding in the case of the histone peptide. Interestingly,c onstrained H3 118-135(St123-127) GCA induced no chemical-shiftv ariation in this region of ASF1 (Figure S2 B). This result suggests that unfolding of the helical conformation is probably required for this nonspecific binding.
The proteolytic stability of the peptides was also investigated by using trypsin and proteinase K. The unconstrained H3 118-135 was cleaved within 14 minutes by both proteases (Figure 4, Table 2a nd the Supporting Information), whereas the constrained peptidesh ad increased stabilityd epending on the positiono ft he constraint. H3 118-135(St120-124) wasl ess susceptible to cleavage by proteinase K( t 1/2 = 65.8 min). On the other   GCA showedi ncreased stability against trypsin (t 1/2 = 40.5 min). The constraint also affected the profile of cleavage sites, most notably for H3 118-135(St123-127) GCA, for which two proteinase Kc leavage sites weres uppressed by introducing the constraint. Thus, the results of proteolytic cleavage studies on constrained peptides need to be considered, as the protective effect is likely to arise not only from the enhanced helicity,t hat is, the greatest effect is observed for the constraint that does not markedly promote helicity (H3 118-135(St120-124) ).
In conclusion, we have shown that variant H3 118-135 peptides with pentenylglycine residuesa tt he i and i+ +4p ositions can be constrained by olefin metathesis to generate a peptidem ore biased towards ah elical conformation than the parentsequence, thus further broadening the scope of this unnatural amino acid for hydrocarbon "stapling". In addition, we have illustrated that am ore helical conformation (i.e.,f or H3 118-135(St123-127) GCA) does not necessarilyc orrelate with significant proteolyticp rotection or enhanced binding potency; rather where the later aspect is concerned, enthalpy-entropy compensation is observed. Nonetheless,c onstrainingp eptides has been shown to reduce nonspecific binding and to enhance ar ange of additional pharmacokinetic properties such as cellular uptake. The peptide sequence used in this work was shown to have moderate helical propensity.T hus our future studies will centre on exploiting the constraining strategy together with helix-stabilising amino acids to optimise these reagents for binding and cell permeability so as to develop chemical probesoft he H3-ASF1 interaction.