Modified Peptide Inhibitors of the Keap1–Nrf2 Protein–Protein Interaction Incorporating Unnatural Amino Acids

Abstract Noncovalent inhibitors of the Keap1–Nrf2 protein–protein interaction (PPI) have therapeutic potential in a range of disease states including neurodegenerative diseases (Parkinson's and Alzheimer's diseases), chronic obstructive pulmonary disease and various inflammatory conditions. By stalling Keap1‐mediated ubiquitination of Nrf2, such compounds can enhance Nrf2 transcriptional activity and activate the expression of a range of genes with antioxidant response elements in their promoter regions. Keap1 inhibitors based on peptide and small‐molecule templates have been identified. In this paper we develop the structure–activity relationships of the peptide series and identify a group of ligands incorporating unnatural amino acids that demonstrate improved binding affinity in fluorescence polarisation, differential scanning fluorimetry and isothermal titration calorimetry assays. These modified peptides have the potential for further development into peptidomimetic chemical probes to explore the role of Nrf2 in disease and as potential lead structures for drug development.

Peptide isolation and purification: Ice-cold Et 2 O (30-35 mL) was added to the concentrated sample to aid precipitation of the crude product. The suspension was centrifuged at 2000 g for 5 min and the ether was decanted. This procedure was repeated twice. The resulting pellet was dried overnight in vacuo.

HPLC methods
Analytical reversed-phase HPLC was carried out on an XSELECT™ CSH™ C 18 column 50 x 6 mm (particle size: 2.5 μm) at a flow rate of 1.0 mL/min. The eluent system consisted of eluent A (H 2 O with 0.1% TFA, HPLC grade) and eluent B (MeCN with 0.1% TFA, HPLC grade) with the following gradient conditions: System A -Initial fixed composition 5% B to 50% B over 20 min, then increased to 95% B over 2 min, held for 2 min at 95% B, then returned to 5% B in 1 min. Total duration of gradient run was 25 min.
Preparative reversed-phase HPLC was carried out on an XSELECT™ CSM™ C 18 column 250 x 10 mm (particle size: 2.5 μm) at a flow rate of 5.0 mL/min. The eluent system consisted of eluent A (H 2 O with 0.1% TFA, HPLC grade) and eluent B (MeCN with 0.1% TFA, HPLC grade) with the following gradient conditions: System B -Initial fixed composition 5% B to 15% B over 60 min, then 15% B to 95% B over 1 min, held at 95% B for 3 min, returned to 5% B in 1 min and held for 1 min. Total duration of gradient run was 66 min.
System C -Initial fixed composition 5% B to 14% B over 45 min, then 14% B to 95% B over 1 min, held at 95% B for 3 min, returned to 5% B in 1 min and held for 1 min. Total duration of gradient run was 51 min.
System D -Initial fixed composition 5% B to 16% B over 45 min, then 16% B to 95% B over 1 min, held at 95% B for 3 min, returned to 5% B in 1 min and held for 1 min. Total duration of gradient run was 51 min.
System E -Initial fixed composition 5% B to 16% B over 48 min, then 16% B to 95% B over 1 min, held at 95% B for 3 min, returned to 5% B in 1 min and held for 1 min. Total duration of gradient run was 54 min.
System F -Initial fixed composition 5% B to 27% B over 45 min, then 27% B to 95% B over 1 min, held at 95% B for 3 min, returned to 5% B in 1 min and held for 1 min. Total duration of gradient run was 51 min.
System G -Initial fixed composition 5% B to 30% B over 46 min, then 30% B to 95% B over 1 min, held at 95% B for 3 min, returned to 5% B in 1 min and held for 1 min. Total duration of gradient run was 45 min.
System H -Initial fixed composition 5% B to 25% B over 40 min, then 25% B to 95% B over 1 min, held at 95% B for 3 min, returned to 5% B in 1 min and held for 1 min. Total duration of gradient run was 46 min. System I -Initial fixed composition 5% B to 18% B over 40 min, then 18% B to 95% B over 1 min, held at 95% B for 3 min, returned to 5% B in 1 min and held for 1 min. Total duration of gradient run was 46 min. System J -Initial fixed composition 5% B to 21% B over 41 min, then 21% B to 95% B over 1 min, held at 95% B for 3 min, returned to 5% B in 1 min and held for 1 min. Total duration of gradient run was 47 min.

Compound 6: Ac-Asp-Pro-Glu-Thr-Gly-Glu-Leu-NH-iPentyl
HMBA resin (100 mg, 0.93 mmol/g) was used to prepare the resin bound precursor Ac-Asp(OtBu)-Pro-Glu(OtBu)-Thr(tBu)-Gly-Glu(OtBu)-Leu-Wang resin according to the general procedure. The acid-sensitive side chain protecting groups of Asp, Glu and Thr were removed using a deprotection cocktail that consisted of TFA/TIS/H 2 O (95:2.5:2.5 v/v/v). The resin was swelled with THF (4 mL) for 1 h. A cleavage solution of TEA/isopentylamine/THF (5 mL, 1:5:5 v/v/v) was added and the reaction was stirred overnight at 50°C. The solution was filtered and the resin was washed first with isopentylamine/THF and then with DCM (3 x 5 mL each). The filtrate was concentrated to a volume of ~ 1 mL and the crude material precipitated by the addition of ice-cold Et 2 O. The peptide was isolated, subjected to preparative HPLC separation (system B) and lyophilised to give a white powder.

Compound 7: Ac-Asp-Pro-Glu-Thr-Gly-Glu-Leu-NH-Bn
HMBA resin (300 mg, 0.93 mmol/g) was used to prepare the resin bound precursor Ac-Asp(OtBu)-Pro-Glu(OtBu)-Thr(tBu)-Gly-Glu(OtBu)-Leu-Wang resin according to the general procedure. The acid-sensitive side chain protecting groups of Asp, Glu and Thr were removed using a deprotection cocktail that consisted of TFA/TIS/H 2 O (95:2.5:2.5 v/v/v). The resin was swelled with THF (4 mL) for 1 h. A cleavage solution of TEA/benzylamine/THF (15 mL, 1:5:5 v/v/v) was added and the reaction was stirred overnight at 50°C. The solution was filtered and the resin was washed first with benzylamine/THF and then with DCM (3 x 5 mL each). The filtrate was concentrated to a volume of ~ 1 mL and the crude material precipitated by the addition of ice-cold Et 2 O. The peptide was isolated, subjected to preparative HPLC separation (system B) and lyophilised to give a white powder.

Compound 8: Ac-Asp-Pro-Glu-Thr-Gly-Glu-Leu-TET
2-Chlorotrityl chloride resin (500 mg, 1.70 mmol/g) was used to prepare this compound according to the general methods. The crude material was subjected to preparative HPLC separation (system C) and lyophilisation to give a white powder. 1

Compound 10: Ac-Asp-Pro-Glu-Thr-Gly-Glu-Thi-OH
2-Chlorotrityl chloride resin (350 mg, 1.70 mmol/g) was used to prepare this compound according to the general methods. The crude material was subjected to preparative HPLC separation (system E) and lyophilisation to give a white powder.

Compound 12: Ac-Asp-Pro-Glu-hPhe-Gly-Glu-Leu-OH
Fmoc-Leu-Wang resin (500 mg, 0.69 mmol/g) was used to prepare this compound according to the general methods. The crude material was subjected to preparative HPLC separation (system G) and lyophilisation to give a white powder.

Compound 15: Ac-Asp-Thp-Glu-Thr-Gly-Glu-Leu-OH
2-Chlorotrityl chloride resin (500 mg, 1.42 mmol/g) was used to prepare this compound according to the general methods. The crude material was subjected to preparative HPLC separation (system K) and lyophilisation to give a white powder.

methoxy)carbonyl)amino)-3-(tritylamino)propanoic acid
The title compound was prepared using an adaptation of a reported procedure. [

Fluorescence polarisation assays
The FP assays were carried out as previously described. [6]  and away from light at RT, the plates were transferred to a PHERAstar microplate reader and the FP was recorded. All measurements were recorded in triplicate. The data were normalised to control and then fitted to a standard four-parameter logistic function using the Origin Pro software.

Differential scanning fluorimetry
The DSF assays were conducted using an adaptation of a published method. [7] Briefly, a solution of the Keap1 Kelch domain protein (5 µM) and detection dye SYPRO ® orange (5x) in DPBS at pH 7.4 was added to the wells of a MicroAmp ® Optical 96-well reaction plate containing the peptide inhibitor (10 µM, 10% final DMSO concentration, 40 µL final volume).
The plate was sealed using an optical adhesive cover and covered with aluminium foil to protect the dye from light. The plate was then transferred to a plate centrifuge and spun down briefly (200 g, 1 min, RT) to remove any bubbles and collect the solution at the bottom of the wells. Following a 30 min incubation at RT, the plate was placed into a 7500 Real time PCR machine and subjected to a standard protein melting protocol 8 using the 7500 Software v.2.0.0. The fluorescence intensity at an excitation of 465 nm and emission of 580 nm was recorded during a temperature scan from 25°C to 95°C with a temperature ramping rate of 1°C/min. All measurements were run in triplicate. The raw data were exported to MS Excel and analysis was performed using a custom script provided by the Structural Genomics Consortium Oxford. [8] The temperature range over which the protein unfolding occurred was selected and the maximum and minimum (at temperatures below the maximum intensity) fluorescence intensities within this range were determined. The processed data were fitted to the Boltzmann equation by linear regression analysis using the Origin Pro software.

Subcloning, expression and purification of Keap1 Kelch protein
The coding sequence for the Keap1 Kelch domain was subcloned into a pEt15b vector as previously described. [9] The plasmids were transformed into E. coli BL21 CodonPlus mM NaCl to further purify Keap1. The protein was then pooled and concentrated to 10 mg/mL using an Amicon ultraconcentration device (Millipore), aliquoted, flash-frozen in liquid nitrogen, and stored at −80°C for subsequent use.

Isothermal titration calorimetry experiments
The protein was initially dialysed overnight against buffer containing 25 mM HEPES, pH 7.4 and 200 mM NaCl. The protein was diluted to a concentration of 50 μM prior to ITC. Inhibitors were diluted to an appropriate concentration in dialysis buffer so that the final solution contained no more than 5% DMSO. To maintain a similar buffer composition for titrations, 5% DMSO was added to the protein solution. All peptides in this study were used at a concentration of 500 μM. There were no detectable solubility issues at these concentrations

Subcloning, expression and purification of Keap1 Kelch protein for crystallisation
The coding sequence for the Keap1 Kelch domain was subcloned into a pEt15b vector as previously described for the ITC experiments. [9] For crystallisation, the same domain was subcloned into a modified pEt15b vector with a His-SUMO tag and Ulp1 cleavage site. The plasmids were transformed into E. coli BL21 CodonPlus (Novagen) and grown in 6 L of TB (Terrific Broth) at 37°C, supplemented with 100 mg/L ampicillin, to an A600 of 1.0 and induced for 24 h with 0.5 mM IPTG (Melford) at 20°C. Cell harvesting and subsequent protein purification were performed as previously described with slight modifications. [9] First affinity chromatography was performed using a 5 mL His-Trap column (GE Healthcare, UK) with buffer containing 50 mM HEPES, pH 7.5, 300 mM NaCl and imidazole at 10 mM (lysis buffer), 30 mM (wash buffer) & 250 mM (elution buffer). For crystallisation, the purified SUMO-tagged Keap1 Kelch domain with a was pooled and cleaved overnight with Ulp1 protease supplemented with 3 mM DTT and dialysed simultaneously against buffer containing 50 mM HEPES, pH 7.4, 300 mM NaCl. Thereafter the cleaved protein was passed through the His-Trap column for the second time using the same buffers as before. The cleaved protein without the his-tag and sumo tag was captured in the flow through.
The purified proteins (with and without his-tag) was pooled and passed through a 16/60 Superdex 200 pg size exclusion column with a buffer containing 50 mM HEPES, pH 7.5, 200 mM NaCl to further purify Keap1. The protein was then pooled and concentrated to 10 mg/mL using an Amicon ultraconcentration device (Millipore), aliquoted, flash-frozen in liquid nitrogen, and stored at −80°C for subsequent use. Diffraction data for individual crystals were collected at beamlines I03 at Diamond Light Source. Data were processed using either XDS [10] or iMosflm. [11] The structure of the complex was solved by molecular replacement using the native Keap1 structure (PDB ID: 3ZGD [12] ) as a search model. The crystallographic data collection and refinement statistics details are provided in Table 1. The overall Keap1-peptide complexes are similar to the previously published structure. [12] There are small areas of positive and negative electron density in the overall electron density map that cannot be attributed to any model, which is consistent with previous structures. The underlying small percentage of twinning (3-4%) overall as determined by various data quality assessment software available.

Molecular modelling
Briefly, the modified peptide structures with appropriate side chains were prepared by editing the Keap1 bound conformation of peptide 5 (PDB Ref 6FMQ) using UCSF Chimera 1.12. [13] The images were prepared using Pymol (The PyMOL Molecular Graphics System, Version