The substrate would be destabilized, lowering the barrier to rotation around the amide bond. This proposed stretching action is consistent with the twisted-amide mechanism, providing a more detailed description of how the isomerization might proceed. ?Stereoisomer (R,R,R)-2, with the ketone carbonyl carbon 4.4 A from the proposed Cys113-S nucleophile, and the S–C = O angle of 102u, had the lowest energy of the three stereoisomers (Table 1). The angle of 102u is close to the optimum angle for nucleophilic addition, i.e. close to the Burgi-Dunitz angle of 107u [27]. Despite ?this, the inhibition results suggest that covalent modification, i.e. suicide inhibition, of Pin1 does not occur. Ketones 1 and rac-2 were designed as tetrahedral-intermediate analogues based on the nucleophilic-addition mechanism; they do not appear to behave as such. The IC50 values are in the range of substrate analogue inhibitors. These results argue against the proposed nucleophilicaddition mechanism for Pin1 [14].
Stereochemical effects on inhibition
The stereochemistry affected the inhibition, since the racemate rac-2 was about 4-fold more potent than diastereomer 1. Molecular modeling provides insight into the stereochemical preferences of the Pin1 active site. The relative (not absolute) energies of the three models can be compared because they are all stereoisomers bound into the same Pin1 active site (Table 1). These inhibitors are substituted with tryptamine, comparable to our ground-state alkene isostere inhibitor 5 with an IC50 value of 25 mM (Figure 2) [13], and with Ac and naphthylethylamine comparable to our a-ketoamide inhibitors 6, with IC50 values of 100 and 200 mM [14]. The Pin1-(S,R,R)-1 complex, with an intermediate energy, corresponds to the native L-Ser-L-Pro configuration, yet it had very poor inhibition (260 mM), comparable to the similarly substituted a-ketoamides 6 [14]. The Pin1(S,S,S)-2 complex, which corresponds to the L-Ser-D-Pro configuration, had the highest energy of the three, while Pin1-(R,R,R)2, corresponding to a D-Ser-L-Pro configuration had the lowest energy. This is consistent with the D-Thr-L-Pip in the most potent peptide inhibitors of Pin1 [12,32]. We expect that (R,R,R)-2 isomer would be more potent than the IC50 value of 61 mM for rac-2 indicates, and (S,S,S)-2 is likely to be less potent than 61 mM, because the IC50 value represents a weighted average of the two.
The most potent that either enantiomer could possibly be
Insights into the Pin1 enzymatic mechanism
To better understand the mechanism of Pin1 PPIase activity, each of the three stereoisomers was docked into the Pin1 active site (Figure 5). Curiously, in each case the inhibitor minimized to a conformation with a trans diaxially substituted cyclohexyl ring. Attempts to force a trans diequatorial conformation on the starting structure resulted in conversion to either a twist boat or a diaxial conformation again. Clearly, the preferred conformation of these cyclohexyl substrate analogues in the Pin1 active site is diaxial. In the crystal structures of intermediates (1S,3R,4R)-11 and rac-11, the cyclohexyl rings were in the diequatorial chair conformation (Figure 4), which are likely to be the low-energy, solution-phase conformations as well. These inhibitors would thus undergo an unfavorable diequatorial to diaxial conformational change in order to bind to the Pin1 active site. We hypothesize that the binding interactions of the enzyme with the phosphate and the aromatic group are strong enough to stretch the cyclohexyl rings into the less stable diaxial conformation upon binding (Figure 6). The difference in the distances between diequatorial and diaxial carbonyl groups on a cyclohex?ane ring was 0.86 A, an elongation of the structure.Table 1. Comparison of cyclohexyl ketone inhibitor-Pin1 complex molecular models.is 30 mM if the other was not an inhibitor at all. This is highly unlikely, but it serves to show that these ketone inhibitors behave as substrate analogues.
Conclusions
Three stereoisomeric ketone analogues of Pin1 substrates were synthesized, modeled, and assayed as Pin1 inhibitors. Molecular modeling shows that the inhibitors have a preference for transdiaxial-cyclohexane conformations upon binding to Pin1. This led us to propose a stretching mechanism to attain pyramidalization of the prolyl nitrogen, consistent with the preferred twisted-amide mechanism [25]. The molecular models of the three stereoisomers in the active site of Pin1 confirmed the stereochemical preferences of Pin1 for inhibitors seen in other inhibitors [12,14,27,32]. We attribute the weaker binding of these inhibitors to a combination of: (1) the conformational change required for binding, and (2) the inability of these ketones to act as electrophilic acceptors for the Pin1 Cys113 thiol. The weak inhibition of the ketones, and the correspondingly stronger inhibition by similarly substituted reduced amide inhibitors [27], provides evidence against the nucleophilic addition mechanism for Pin1.
Materials and Methods Synthesis
Unless otherwise indicated, all reactions were carried out under dry N2 in flame-dried glassware. THF was distilled from Nabenzophenone, and CH2Cl2 was dried by passage through dry alumina. Anhydrous DMF (99.8%), MeOH, and DIEA were used directly from sealed bottles. Brine (NaCl), Na2S2O3, NaHCO3, and NH4Cl refer to saturated aqueous solutions, and HCl refers to a 1 N aqueous solution, unless otherwise noted. Flash chromatography was performed on 230?00 mesh silica gel with reagent grade solvents. Analytical HPLC were obtained on a 4.6650 mm C18 column with 10% CH3CN/H2O for 3 min followed by a 10% to 90% CH3CN/H2O gradient over 6 min unless otherwise noted. HPLC results are reported as retention time, integrated % purity. 1H, 13C, and 31P NMR spectra were obtained at ambient temperature in CDCl3, unless otherwise noted. Chemical shifts are reported in parts per million (ppm) downfield from tetramethylsilane (TMS). Data are reported as follows: chemical shift, multiplicity: singlet (s), doublet (d), triplet (t), multiplet (m), broad singlet (br s), coupling constants J in Hz, and integration. HPLC chromatograms for compounds 1 and rac-2, 1H (500
