Materials and instruments
Selenol-l-cystine (SeCys) was purchased from J&K Chemical (Guangzhou, China). Reduced glutathione (GSH), oxidized glutathione (GSSG), dithiothreitol (DTT), tris (2-carboxyethyl) phosphine (TCEP), N,N-diisopropylethylamine (DIEA), ethyl cyanoglyoxylate-2-oxime (Oxyma), N,N′-diisopropylcarbodiimide (DIC), ethanedithiol (EDT), and piperidine were purchased from Energy Chemical (Shanghai, China). Acetonitrile (ACN), trifluoroacetic acid (TFA), and trypsin were purchased from Sigma-Aldrich (Beijing, China). Guanidine hydrochloride (Gn·HCl) was purchased from Sangon Biotech (Shanghai, China). 4-Mercaptophenylactic acid (MPAA) was purchased from Alfa Aesar (Beijing, China). Thioanisole was purchased from TCI (Shanghai, China). Rink-amide MBHA resins, 2-chlorotrityl chloride resins and Fmoc-protected amino acids were purchased from GL Biochem (Shanghai, China). Other conventional reagents were purchased from Sinopharm Chemical Reagent (Beijing, China). All reagents were analytical grade at least and used without further purification. Ultrapure water (18.0 MΩ/cm) was used throughout the experiments.
All peptides were synthesized on a CEM Liberty Blue automatic microwave peptide synthesizer. Waters high-performance liquid chromatography (HPLC) and SHIMADZU HPLC were used to purify and analyze peptides and proteins. Bruker Esquire 3000 plus ion trap ESI mass spectrometry, Bruker autoflex maX MALDI-TOF mass spectrometer and Bruker Impact II QqTOF mass spectrometer were used to identify peptides and proteins. HITACHI U-3900H UV–Vis spectrophotometer was used to determine the concentration of peptides and proteins. Jasco J-810 circular dichroism (CD) spectrometer was used for recording CD spectra. Proteins were purified using a SuperdexTM 75 10/300 GL column (GE Healthcare) on a fast protein liquid chromatography (FPLC). NMR experiments were performed at 298 K on Bruker AVANCE III 850 MHz equipped with a cryogenic triple-resonance probe.
Computation placement of disulfides into designed structures
A database of 30,000 example disulfide geometries was acquired from the PDB database filtering for high-resolution structures (<2.0 Å resolution). The 30,000 example disulfide geometries were utilized to generate a hash database. To improve coverage of the example geometries, all of the example disulfides from the PDB were randomly perturbed 100 times up to 5 Å on the sidechain degrees of freedoms. Once the large set of example disulfides was generated, the relative transformation from backbone-to-backbone coordinates was calculated and hashed using the same hash function in previous work. The cartesian resolution parameter was 1.0 Å and the angular resolution parameter was 15.0°. The hashed transformation was saved as a key in a dictionary and the corresponding rotamers of the cysteine residues were stored as the associated value. This database can be quickly queried to identify residue pairs in peptides and proteins that can accommodate a disulfide bond by calculating all of the hashed residue-to-residue transformations of the design peptide/protein structure and comparing those values to the keys stored in the database. When a hashed residue pair transformation is found in the database, the associated cysteine rotamers can be placed into design peptide/protein structure. The placed disulfide will be close to optimal geometry, but a minimization with the newly placed disulfide bond is necessary to optimize the structure. An implementation of this protocol with instructions for installation can be found at: https://github.com/atom-moyer/stapler.
Computational design of peptide heterodimers
Various native and non-native scaffolds were evaluated with the hash-based disulfide placement protocol. After disulfides were placed into the potential scaffolds, the structures were prepared for design by splitting the structures into hetero-dimeric structures by splitting the monomeric chains into two separate structures. The penicillamine residues were manually mutated to obey the synthetic restrictions of the Cys–Pen pairing and ensure that the bulkier Pen residue could be structurally accommodated. The sequences were optimized using Rosetta Scripts and PyRosetta with the goal of introducing favorable sidechain interactions across the heterodimer interfaces. The Rosetta protocol focuses on introducing hydrogen bonding interactions across the interface to improve specificity. A representative script and associated parameter files used to design the peptides can be found in a Supplementary file.
Peptide synthesis and purification
All peptides were synthesized using Fmoc solid-phase peptide synthesis method on a CEM Liberty Blue automatic microwave peptide synthesizer (peptide sequences were provided in Supplementary Table 1). Coupling of Fmoc-protected amino acids to the resins (0.05 mmol scale) was performed in the reaction vessel using standard coupling methods. In general, coupling reactions were performed with a 5-fold excess of Fmoc-protected amino acids, 0.5 mmol DIC, and 0.5 mmol Oxyma in DMF. After each coupling step, Fmoc groups were removed using 20% piperidine in DMF. The synthesized peptides were cleaved from the resins by treating the peptidyl resins with a cleavage cocktail (TFA/thioanisole/EDT/H2O/phenol in a volume ratio of 87.5/5.0/2.5/2.5/2.5) for 3.0 h on a shaker at 37 °C. After that, the cleaved peptides were precipitated in anhydrous ether and purified by HPLC.
Preparation of Fmoc-hydrazine resins
1.5 g of 2-chlorotrityl chloride resins (resin loading: 1.14 mmol/g) were swelled in 12 mL DCM for 20 min with protective nitrogen and cooled in an ice bath. To the swollen 2-chlorotrityl chloride resins, Fmoc-hydrazine (4 equiv. relative to the resin loading) and DIEA (10 equiv.) dissolved in 15 mL DMF and 3 mL DCM was added slowly and dropwise. After 15 h reaction at room temperature, 0.69 mL methanol was added to quench the reaction and block unreacted sites. The resins were then washed successively with DMF, methanol and anhydrous ether. The washed resins were vacuum dried for 1 h. The prepared Fmoc-hydrazine resins were determined to have a loading of 0.346 mmol/g, which were used for the synthesis of peptide hydrazides.
Native chemical ligation
The general procedure for ligation of peptides using peptide hydrazides was followed as described previously38. Peptide hydrazide was dissolved in 100 mM phosphate buffer (pH 3.0) containing 6.0 M Gn·HCl to reach a concentration of 1.0 mM. Acetyl acetone (5 equiv. relative to the peptide hydrazide) and 4-mercaptophenylacetic acid (20 equiv.) were then added to the peptide mixture, which was allowed to react for 4 h on a shaker at 37 °C, followed by purification using HPLC to obtain the peptide thioester. The obtained peptide thioester was freeze-dried and dissolved in 100 mM phosphate buffer (pH 7.4) containing 50 mM TCEP. N-Terminal Cys-containing peptides (1.3 equiv. relative to the peptide thioester) were added to the mixture, and the cocktail was stirred for 2.0 h at 37 °C. After that, the ligated peptides were purified by HPLC. hd1–TEV–hd3-A and hd1–TEV–hd3-B were synthesized using this method.
Oxidative dimerization of peptides
In a typical experiment, HPLC-purified peptide monomers with multiple Cys residues (100 μM) and multiple Pen residues (2 equiv.; 200 μM) were dissolved in phosphate buffers containing 50 μM selenol-l-cystine (SeCys). After reaction for ~24 h on a shaker at 37 °C, the formed products were characterized by HPLC and mass spectrometry. HPLC chromatograms and mass spectrometry were acquired using LabSolutions version 5.85 (Shimadzu) and flexControl version 3.4 (otofControl version 4.0.15.3248, trapControl Version 7.0; Bruker), respectively. Detailed conditions for the dimeric folding of peptides were provided in Supplementary Table 5.
Circular dichroism measurements
General method
Circular dichroism (CD) spectra were measured using a 1.0 mm path length cuvette. All peptides were dissolved in water to reach a concentration of 50 μM. CD spectra were recorded in a wavelength range of 190‒260 nm with a bandwidth of 2.0 nm, a date pitch of 1.0 nm, a response time of 8.0 s, and a scanning speed of 50 nm/min. CD spectra were acquired using Spectra Manager version 1.52.01 (JASCO).
Thermal denaturation
The thermal denaturation CD spectra were recorded in a temperature range of 25‒95 °C with a heating rate of 5 °C/min. For temperature melts, the change of ellipticity at 220 nm was monitored as the temperature increased from 25 to 95 °C in an increment of 5 °C. Other experimental conditions were consistent with those of the general method.
Chemical denaturation
Peptide heterodimers with a concentration of 50 μM were dissolved in water containing 1, 3 and 6 M Gn·HCl, respectively. Then, CD spectra were recorded using the conditions given in the general method. The baseline signals from the aqueous Gn·HCl solution were subtracted from the corresponding spectrum.
NMR characterization of peptide heterodimers
The hd1 and hd2 peptides were dissolved in H2O/D2O (90%/10%) and H2O/CD3CN (50%/50%) with a final concentration of 1.3 mM and 0.8 mM, respectively. All NMR experiments were conducted at 298 K on a Bruker AVANCE III 850 MHz spectrometer equipped with a 5 mm z-gradient 1H/13C/15N TCI cryogenic probe. Two-dimensional (2D) 1H–1H TOCSY (80 ms mixing time) and 1H–13C/15N HSQC spectra were measured for resonance assignments. 2D 1H–1H NOESY experiments with a mixing time of 200 or 300 ms were performed to extract 1H–1H distance. The phi and psi backbone torsion angles were predicted by TALOS+ using chemical shifts of HN, HA, CA, CB, and N39. The Xplor-NIH (version 2.53) program was used for the structure determination and refinement40. The intra- and inter-molecular hydrogen bond distance constraints inferred from the preliminarily calculated structures were applied in the late stage of structure determination. All NMR spectra were processed using TopSpin 3.5 and analyzed using NMRFAM-SPARKY41. 1H chemical shifts were referenced to DSS, and 13C/15N chemical shifts were referenced indirectly to DSS.
Tryptic digestion analysis of disulfide connectivity
Freeze-dried peptide heterodimers (~100 μM) were dissolved in 100 mM phosphate buffer (pH = 6.0 or 7.4), and trypsin with a final concentration of 100 μg/mL was added to the reaction solution. The reaction mixture was kept in a shaker at 37 °C for 8 h (pH = 6.0) or 2 h (pH = 7.4). Then, the digested fragments were characterized using HPLC and mass spectrometry.
Protein expression and purification
Genes encoding the protein sequences (Neo-2/15-hd1-B, MLK3-SH3-hd1-B, SUMO-hd3-B, and MLK3-SH3-hd1/hd3-B) were cloned into pET-28b(+) E. coli expression vectors (Supplementary Table 4). The recombinant plasmids were then transformed into chemically competent E. coli BL21 (DE3) cells. The cells were grown at 37 °C until A600 nm reached 0.6–0.8 in fresh LB medium, and then the protein expression was induced by adding 0.5 mM isopropyl-β-d-thiogalactopyranoside (IPTG). After ~14 h culture at 16 °C, the cells were harvested and lysed by sonication on an ice bath using a lysis buffer (100 mM Tris, 500 mM NaCl, 1 mM PMSF, pH 7.9). The soluble fraction was clarified by centrifugation at 11,000 rpm for 30 min, and then purified by Ni2+ Sepharose column with a concentration gradient of imidazole. All proteins were finally purified using FPLC and HPLC, and identified by MALDI-TOF MS (flexControl version 3.4). The purified proteins were stored at ‒80 °C for the further labeling applications.
Labeling of proteins
The purified proteins (5 or 12.5 μM) and the relevant Pen-bearing monomers modified with functional molecules (100 μM) were dissolved in buffers containing 50 μM selenol-l-cystine (SeCys). After reaction for 24 h on a shaker at 37 °C, the reaction mixture was analyzed by HPLC, SDS–PAGE, and MALDI-TOF MS. Detailed conditions for the labeling of proteins were provided in Supplementary Table 6. To further identify the protein labeling, the labeled proteins were freeze-dried and dissolved in 100 mM phosphate buffer (pH = 7.0), into which 2 µL AcTEVTM Protease (10 U/µL in AcTEVTM Protease buffer) was added. The reaction mixtures were kept on a shaker at 37 °C for 2 h. Then, the digested fragments were analyzed by HPLC (LabSolutions version 5.85) and MALDI-TOF MS (flexControl version 3.4).
Statistics and reproducibility
No sample size calculation was performed. All experiments were repeated independently at least once with similar results. All results are reproducible.
Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
