Structure-guided design of CPPC-paired disulfide-rich peptide libraries for ligand and drug discovery

Peptides constrained through multiple disulfides (or disulfide-rich peptides, DRPs) have been an emerging frontier for ligand and drug discovery. Such peptides have the potential to combine the binding capability of biologics with the stability and bioavailability of smaller molecules. However, DRPs with stable three-dimensional (3D) structures are usually of natural origin or engineered from natural ones. Here, we report the discovery and identification of CPPC (cysteine–proline–proline–cysteine) motif-directed DRPs with stable 3D structures (i.e., CPPC–DRPs). A range of new CPPC–DRPs were designed or selected from either random or structure–convergent peptide libraries. Thus, for the first time we revealed that the CPPC–DRPs can maintain diverse 3D structures by taking advantage of constraints from unique dimeric CPPC mini-loops, including irregular structures and regular α-helix and β-sheet folds. New CPPC–DRPs that can specifically bind the receptors (CD28) on the cell surface were also successfully discovered and identified using our DRP-discovery platform. Overall, this study provides the basis for accessing an unconventional peptide structure space previously inaccessible by natural DRPs and computational designs, inspiring the development of new peptide ligands and therapeutics.


List of Supplementary Material
Experimental section Table S1. Information of phage libraries Table S2. Information of peptides Recombinant plasmids used for library construction were transformed into TG1 E. coli using electroporation (Bio-Rad). Peptides were synthesized by solid phase peptide synthesis (SPPS) using CEM Liberty Blue TM automated microwave peptide synthesizer and analyzed by high performance liquid chromatography (HPLC) (SHIMADZU system). Bruker autoflex max MALDI-TOF mass spectrometry was used for identifying peptides. Peptides were quantified using HITACHI U-3900H UV/Vis spectrometer. The concentrations of proteins or nucleic acid were determined by Nanodrop (Thermo Fisher Scientific). Fluorescence anisotropy was measured in a 96 well flat-bottom OptiPlate black plate by Infinite® 200 PRO multimode microplate readers (TECAN). NMR experiments were recorded at 298 K on Bruker AVANCE III 850 MHz equipped with a cryogenic triple-resonance probe.

Library construction
All peptide libraries were constructed as described previously 1 . Oligonucleotides (chapter 2.1) coding peptides were co-incubated with the reverse primer (5'−CACCGGCGCACCTTGCGGCCGC−3') in extended system containing Klenow fragment. The two double-stranded products of DNA and phagemid vector (pCantab 5E) were then digested with Sfi I (8 h, 50℃) and Not I (8 h, 37℃). For each library, the purified Sfi I/Not I-digested DNA fragments were ligated into the phagemid and electroporated into the E. coli TG1 cells. After that, the library capacity was evaluated by counting the clone number of gradient diluted plates (Table S2 shows exogenous peptides that were displayed on the surface of phages).

Biotinylation of protein
Protein was biotinylated at a concentration of 1 μM with a 5-fold molar excess of Sulfo-NHS-LC-biotin in PBS, pH 7.4 at room temperature for 30 min. The biotinylated protein was purified on a HiTraq TM 5 mL using PBS, pH 7.4.

Phage screening
Phages encoded exogenous peptides were recovered from 2-YT medium supernatant with PEG/NaCl precipitation and PBS re-suspension. The phage titer was then measured, and the phages were blocked in blocking buffer for 30 min. The streptavidin magnetic beads (or neutravidin-coated magnetic beads) binding with biotinylated protein were also blocked using the blocking buffer for 120 min. The magnetic beads were incubated with the phage peptide library at room temperature for 30 min. After washing nine times with washing buffer and twice with binding buffer, the phages were eluted from magnetic beads with glycine-hydrochloric acid solution (50 mM glycine, pH 2.2). The phage amplification and titer measurement were conducted to make a preparation for the next round of panning. Phage titer was determined after each round of panning to monitor the results of screening. After three or four rounds of panning, the enriched phages were sequenced using next-generation sequencing.

Sequence identification by next-generation sequencing (NGS)
Phage vectors extracted from E. coli TG1 were stored as glycerol stocks. The DNA was isolated with a commercial plasmid purification kit (TIANGene ® Tianjin, China). Phage vector DNA was amplified by PCR. The PCR reaction contained final concentrations of 250 μM dNTPs, 100 nM primer, 100 ng phage vector as temple, 1 unit Taq polymerase and 10 μL of 10×Taq buffer. The mixture solution was filled up to 100 μL with water. The following program was used: initial denaturation for 60 s at 95℃, 12 cycles of from 2% agarose gel (containing 0.005% v/v of 4S Red Plus Nucleic Acid Stain) used a commercial agarose gel purification kit (E.Z.N.A ® Gel Extraction Kit, OMEGA BIO-TEK, USA). Then, Novogene Co., Ltd. was commissioned for sequence identification. The data were processed and analyzed using MatLab scripts 2, 3 .

Synthesis of peptides
All peptides were synthesized on a CEM Liberty Blue automated microwave peptide synthesizer by standard Fmoc solid-phase chemistry on a Rink Amide MBHA resin (0.025 mmol scale). Coupling reactions were performed with a Fmoc-AA-OH (5 eq., 0.2 M solution in DMF), DIC (5 eq., 0.25 M solution in DMF) and Oxyma Pure (5 eq., 1.0 M solution in DMF). Fmoc groups were removed using a 20% (v/v) solution of piperidine in DMF. Following the synthesis, peptidyl resins were cleaved by treating with a mixture of TFA/EDT/thioanisole/Phenol/H2O (87.5/2.5/5/2.5/2.5 v/v, 5 mL) for 4 h at 37℃. The resin was removed by filtration and the peptides were precipitated with cold diethyl ether (30 mL), collected by centrifugation (6000 rpm at 4℃, 4 min). The precipitated peptides were resuspended and washed three times with diethyl ether (30 mL each time). Then, the crude peptides were purified by preparative HPLC.
The peptides isolated from preparative HPLC can be purified again to ensure the purity (>95%).

Oxidative folding of peptides
In a typical experiment, reduced peptide (50 μM) was dissolved in 100 mM phosphate buffer (500 μL, pH 7.4) containing 6 M Gu· HCl, 0.5 mM GSSG and 10% DMSO (v/v). The reaction mixture was stirred for 12 h at 37℃. After the reaction is complete, the oxidized peptide was further injected directly in HPLC using a gradient of 15-80% acetonitrile (0.1% TFA) and water (0.1% TFA) over 60 min with a C-18 column. After analyzed by MALDI-TOF MS, the oxidized peptide was purified and lyophilized to yield a white powder.

Circular dichroism (CD) spectra characterization
Circular dichroism (CD) wavelength and temperature scans were recorded for thermal denaturation experiments. The drp4 was prepared at 61 μM final concentration in pure water. Wavelength scans from 190 nm to 260 nm were recorded at 25℃, 45℃, 75℃ and 95℃, and temperature were carried out at a heating rate of 5℃/min in 1 mm path length cuvettes.

Fluorescence polarization (FP) assay
(a) Fluorescence polarization binding assay. The binding of oxidized FITC-drp8 to CD28 was measured by adding dilutions of CD28 to a fixed concentration of the fluorescent peptide in PBS (pH 7.4). The solutions were added to wells of a 96-well plate to reach a total assay volume of 125 μL. The final concentrations of oxidized FITC-drp8 were 25 nM and the CD28 ranged from 10 nM to 800 nM. After 10 min incubation at room temperature, the fluorescence anisotropies were measured using a plate reader (Eex= 485 nm, Eem=535 nm). All of the measurements were recorded in triplicate. In the fluorescence polarization assays measuring the binding of reduced FITC-drp8 to CD28 and the binding of oxidized FITC-drp8 to CTLA-4, the fluorescence anisotropies were recorded in the same way as described above.

Surface plasmon resonance (SPR) assay
The affinity of peptides to Keap1 was measured using Biacore T200 (GE healthcare) at 25℃. At first, carried out for structure calculation. The NMR data were processed using NMRPipe/NMRDraw and analyzed using NMRFAM-SPARKY 5 . More than 95% of all NOE cross-peaks were assigned manually.
The backbone dihedral angle restraints were obtained using programs TALOS-N based on chemical shifts of backbone resonances 6 . The distance constraints were deduced from volume integration of NOE crosspeaks and carried out with ARIA2.3.2 and CNS1.21 7,8 . All of the ensemble of 15 lowest-energy structures were generated from a total of 150 structures in last iteration for each run. The structures were visualized in PyMol, and the quality of structures was analyzed using PROCHECK 9 .  Table S1 shows the information of displayed peptides and capacity of different libraries.

Laser confocal cell imaging
Oligonucleotides of Library-1: Oligonucleotides of Library-2: Oligonucleotides of Library-3:               Characterization of drp7, drp8, drp9 and FITC-drp8 Figure S20. a) drp7 used for FP assays was purified using HPLC to a purity of >95%. b) drp8 used for FP assays was purified using HPLC to a purity of >95%. c) drp9 used for FP assays was purified using HPLC to a purity of >95%. d) Mass spectrum of reduced drp7. e) Mass spectrum of oxidized drp8. f) Mass spectrum of oxidized drp9. Figure S21. a) The reduced FITC-drp8 used for FP assays was purified using HPLC to a purity of >95%.

Information of peptides
b) The oxidized FITC-drp8 used for FP assays was purified using HPLC to a purity of >95%. c) The oxidation product drp1 used for FP assays was purified using HPLC to a purity of >95%. d) Mass spectrum of reduced FITC-drp8. e) Mass spectrum of oxidized FITC-drp8.
Binding of CD80 to CD28