Cyclic Octapeptides Composed of Two Glutathione Units Outperform the Monomer in Lead Detoxification

Abstract A rationally‐designed scaffold of cyclic octapeptides composed of two units of the natural tripeptide glutathione (GSH) was optimized to strongly and selectively capture toxic lead ions (Pb(II)). Using state‐of‐the‐art computational tools, a list of eleven plausible peptides was shortened to five analogs based on their calculated affinity to Pb(II) ions. We then synthesized and investigated them for their abilities to recover Pb‐poisoned human cells. A clear pattern was observed from the in vitro detoxification results, indicating the importance of cavity size and polar moieties to enhance metal capturing. These, together with the apparent benefit of cyclizing the peptides, improved the detoxification of the two lead peptides by approximately two folds compared to GSH and the benchmark chelating agents against Pb poisoning. Moreover, the two peptides did not show any toxicity and, therefore, were thoroughly investigated to determine their potential as next‐generation remedies for Pb poisoning.


Computational details
All DFT calculations were carried out using the Turbomole 7.5.1 program. [1] Geometries of all structures were optimized employing the dispersion-corrected BP86 functional [BP86-D3(BJ)] [2][3][4][5] and the def2-TZVP [6] basis set (for Pb, this basis set includes the Stuttgart-Dresden effective core potentials) in an implicit solvent with the dielectric constant of εr = 80 corresponding to water -employing the COSMO solvation model [7] as implemented in Turbomole 7.5.1 program.
Subsequent vibrational frequency calculations were performed at the same level of theory for all calculated structures in vacuo after reoptimization (in vacuo). All equilibrium structures had all Hessian eigenvalues positive, so all stationary points were confirmed to be genuine minima on the potential energy surface (PES).
According to previous work, [8] the free energy value corresponding to a particular structure/molecule can be conveniently expressed as (Equation S1): where Eel is the electrostatic potential energy of the molecule in vacuo (gas-phase molecular The ΔGsolv was obtained using Klamt's conductor-like screening model for the realistic solvation method (COSMO-RS). [9] COSMO-RS calculations were carried out using cosmotherm21 software with the parameter file "BP_TZVPD_FINE_C30_2101.ctd" and the recommended protocol: BP86-D3(BJ)/def2-TZVPD single point calculations in vacuo (on top of the in vacuo geometries) and in an ideal conductor (ε = ∞) for the solvent geometry, followed by the COSMO-RS (cosmotherm21) calculations in the target solvent (water). Throughout, FINE cavities ($cosmo_isorad keyword) were used to increase numerical precision. Finally, a correction of (1.9•Δn) kcal mol -1 (corresponding to the difference between the concentration of the ideal gas at 298 K and 1 atm and its 1 mol L -1 concentration; Δn is the change in the number of moles in the reaction) has been applied in order that the computed values refer to 1 mol L -1 standard state.

Definition of a complexation energy
The following definition of the complexation Gibbs free energy, ΔGcomp (which relates to Kd as ΔGcomp = RTlnKd) has been used throughout (Equation S2): The "reaction" used is the following (Equation S3): where 5 is the optimal hydration number for Pb(II) (the lone pair on Pb(II) leads to the hemidirected hydration and the n = 5 to be the optimal coordination number); m denotes the number of the first-sphere, n the number of second-sphere waters, and q the number of protons dissociated upon Pb(II) binding.
6 This approach was previously tested [8] and compared with an alternative approach involving To find the global minimum of the peptide X in solution, we employed the Maestro/MD-LLMOD program (Schrodinger, Inc.) with default settings to obtain the initial set of ~200 conformers for each peptide, which were further processed by the QM//COSMO-RS protocol described above.
Thus, we expect that we located the global minimum for each peptide X.
7 Figure S1. Calculated lowest-energy structures of [Pb-X(H2O)m]n{H2O} where X is all eleven examined peptides and m + n = 4. Hydrogen atoms and four water molecules, placed in the first or the second coordination spheres, were omitted for clarity.

Calculations of complexation of Zn(II) and Cu(II) by peptide 5
The calculations of the complexation Gibbs free energy was carried out in the same manner like in the case of Pb(II). As a reference state, we used the hexahydrate of Zn(II) and pentahydrate of Cu(II), so n = 6 for Zn(II) and n = 5 for Cu(II) in equation S3. In both cases, four water molecules were used as a model of the first hydration sphere (m+n = 4). In case of Zn(II), no protons were released during the complexation (q = 0); one proton was released during the complexation of Cu(II) (q = 1).
8 Table S2. Computed complexation Gibbs free energy values for the complexation of peptide 5 with Zn(II) and Cu(II)

Coupling of the first amino acid
The resin was washed with dry CH2Cl2, and a mixture of Fmoc-Xaa-OH (3 equiv.) and Hünig's base (6 equiv.) in dry CH2Cl2 was added. The resin was agitated for 2 h and then washed with DMF (x3) and CH2Cl2 (x3).

Fmoc deprotection
A solution of 20% piperidine in DMF was added to the resin, and the suspension was agitated for 10 min. The procedure was repeated twice. The resin was then washed with DMF (x3) and CH2Cl2 (x3).

Amino acid couplings
Fmoc-Xaa-OH (3 equiv.) and HATU (2.7 equiv.) were dissolved in DMF (0.8 M), and Hünig's base (6 equiv.) was added. After 1 min, the mixture was added to the amino-functionalized resin that was suspended in a minimal amount of DMF. The mixture was agitated for 60 min and washed with DMF (x3) and CH2Cl2 (x3). Couplings were monitored by test-cleavage of a small portion of resin and subsequent LCMS analysis.   The plates were incubated at 37 °C, and 5% CO2 for additional 23 h, after which the medium was removed, the wells were washed with H2O, and 50 µL of crystal violet solution [10,11] (500 mg crystal violet powder in 20 mL CH3OH and 80 mL H2O) were added to each well. The plates were gently shaken (50 rpm) for 20 min. The plates were then washed with H2O until no more unbound dye was observed and allowed to dry overnight. 100 µL of CH3OH were added to each well, and the plates were gently shaken (50 rpm) for 20 min, after which their absorbance at 560 nm was recorded. 16 The recovery values of each concentration of a peptide were calculated according to Equation

S4
:  performed in triplicates. The plates were incubated at 37 °C, and 5% CO2 for 23 h, after which the medium was removed, the wells were washed with H2O, and 50 µL of crystal violet solution [10] (500 mg crystal violet powder in 20 mL CH3OH and 80 mL H2O) were added to each well. The plates were gently shaken (50 rpm) for 20 min. The plates were then washed with H2O until no more unbound dye was observed and allowed to dry overnight. 100 µL of CH3OH were added to each well, and the plates were gently shaken (50 rpm) for 20 min, after which their absorbance at 560 nm was recorded.  Fluorescence was then measured at an excitation of 490 nm and an emission of 520 nm.

ROS production determination in the absence of Pb(II) ions
96-well plates were prepared such that every well contained 5,000 cells in 50 µL medium, and the cells were allowed to adhere overnight.
To all wells but the positive control wells, 5 µL of 8 or GSH (12 mM) were added to reach final concentrations of 1 mM. To the positive control wells, 5 µL of H2O were added.
The plates were incubated at 37 °C, and 5% CO2 for additional 23 h, after which the medium was removed, the wells were washed with PBS, and 50 µL of PBS were added to each well. 6.75 µL of           was titrated to a peptide solution (1.8 mL) at concentrations according to Table S9.    Sample preparation for ICP-MS analysis After respective incubation times, the supernatants were homogenized by gently pipetting up and down and 50 µL were transferred and digested with 100 µL HNO3 (70%) for 1 h in a 15 mL tube. 42 µL of Lu2O3 (25 mg L -1 in 2% HNO3) were added as internal standard, followed by the addition of 3.808 mL H2O. Pb in the supernatant was then quantified using ICP-MS.
Since the cells are adherent, residual supernatant was removed from each well and the cells were washed with RPMI (3x200 µL). To each well, 100 µL of HNO3 (70%) were added and the cells were digested for 1 h at rt. 50 µL of the resulting solution were taken and further 100 µL of HNO3 (70%) were added. 42 µL of Lu2O3 (25 mg L -1 in 2% HNO3) were added as internal standard, followed by the addition of 3.808 mL H2O. Pb in the cells was then quantified using ICP-MS.
For each condition three independent experiments were conducted.