Impact of the Warhead of Dipeptidyl Keto Michael Acceptors on the Inhibition Mechanism of Cysteine Protease Cathepsin L

Cathepsin L (CatL) is a lysosomal cysteine protease whose activity has been related to several human pathologies. However, although preclinical trials using CatL inhibitors were promising, clinical trials have been unsuccessful up to now. We are presenting a study of two designed dipeptidyl keto Michael acceptor potential inhibitors of CatL with either a keto vinyl ester or a keto vinyl sulfone (KVS) warhead. The compounds were synthesized and experimentally assayed in vitro, and their inhibition molecular mechanism was explored based on molecular dynamics simulations at the density functional theory/molecular mechanics level. The results confirm that both compounds inhibit CatL in the nanomolar range and show a time-dependent inhibition. Interestingly, despite both presenting almost equivalent equilibrium constants for the reversible formation of the noncovalent enzyme/inhibitor complex, differences are observed in the chemical step corresponding to the enzyme–inhibitor covalent bond formation, results that are mirrored by the computer simulations. Theoretically determined kinetic and thermodynamic results, which are in very good agreement with the experiments, afford a detailed explanation of the relevance of the different structural features of both compounds having a significant impact on enzyme inhibition. The unprecedented binding interactions of both inhibitors in the P1′ site of CatL represent valuable information for the design of inhibitors. In particular, the peptidyl KVS can be used as a starting lead compound in the development of drugs with medical applications for the treatment of cancerous pathologies since sulfone warheads have previously shown promising cell stability compared to other functions such as carboxylic esters. Future improvements can be guided by the atomistic description of the enzyme–inhibitor interactions established along the inhibition reaction derived from computer simulations.


Figure S1 .
Figure S1.Time evolution of the RMSD (in Å) computed with all the protein atoms and only the protein backbone along the classical MD for a) compound 1, and b) compound 2.

Figure S2 .Figure S3 .
Figure S2.Comparative analysis between the averaged inter-atomic distances computed through the long MM MD simulations of the non-covalent reactant complex (in blue) and the selected final structure employed for the QM/MM calculations (in green) for a) compound 1; and b) compound 2.

Figure S4 .
Figure S4.Analysis of structures generated during 1µs MD simulations a) RMSD of the position of all atoms of compound 1 b) Evolution of the distance between the Sγ sulfur of the catalytic cysteine and the Cα carbon of the double bond of compound 1 c) RMSD of the position of all atoms of compound 2 along the MD d) Distance between the sulfur of the reactive cysteine and the Cα carbon of the double bond of compound 2.All RMSD and distances are in Å.

Figure S5 .Figure S7 .
Figure S5.Key distances evolution (in Å) along the last 500 ps of the classical MD for a) compound 1; and compound 2 b).

Figure S9 .
Figure S9.DFT/MM Free energy profiles for the formation of the activated ion-pair, E:I (+/-) , from the non-covalent complex E:I, computed by means of the free energy perturbation method at M06-2X/MM level of theory for compound 1 (a) and compound 2 (b).

Figure S10 .
Figure S10.QM/MM Potential Energy Surfaces, PESs, of the attack of the Sγ C25 sulfur atom to: a) Cα atom of the double bond of compound 1; b) Cα atom of the double bond of compound 2; c) Carbonyl of LEU3 of compound 1; d) Cβ atom of the double bond of compound 1; e) Carbonyl of LEU3 of compound; and 2 f) Cβ atom of the double bond of compound 2. PESs on panels a and b were computed at DFT:AM1/MM level, while those on panels c-f were computed at DFT/MM level.

Figure S11 .
Figure S11.Detail of the active site after the attack of the Sγ C25 sulfur atom to a) Carbonyl of LEU3 of compound 1 b) Cβ atom of the double bond of compound 1 c) Carbonyl of LEU3 of compound 2 d) Cβ atom of the double bond of compound 2.Structures optimized at M06-2X/MM level.

Figure S12 .
Figure S12.DFT/MM Free energy profiles for the Michael addition reaction from the ion pair E:I (+/-) computed by means of the free energy perturbation method at M06-2X/MM level of theory for compound 1 (a) and compound 2 (b).

Table S1 .
Compound 1 (top) and 2 (bottom) were divided in four parts (see Figure4of main text).The WAR, WAS and CBZ parts were parametrized with Gaff forcefield, and the atoms were named as it is shown.The rest was parametrized with AMBER general force field for amino acids.Atom types, charges and parameters for bonding and non-bonding interactions obtained for WAR, CBZ and WAS of both inhibitors computed using Antechamber software.

Table S4 .
Key interatomic distances for all the states appearing along the Michael addition mechanism for compound 2. All states have been optimized at M06-2X/MM level of theory and the distances are given in Å.

Table S5 .
Charges in a.u of the most relevant heavy atoms along the mechanism of reaction computed at the M06-2X/MM level of theory for compound 1.

Table S6 .
Charges in a.u of the most relevant heavy atoms along the mechanism of reaction computed at the M06-2X/MM level of theory for compound 2.