Deciphering the Synthetic and Refolding Strategy of a Cysteine-Rich Domain in the Tumor Necrosis Factor Receptor (TNF-R) for Racemic Crystallography Analysis and d-Peptide Ligand Discovery

Many cell-surface receptors are promising targets for chemical synthesis because of their critical roles in disease development. This synthetic approach enables investigations by racemic protein crystallography and ligand discovery by mirror-image methodologies. However, due to their complex nature, the chemical synthesis of a receptor can be a significant challenge. Here, we describe the chemical synthesis and folding of a central, cysteine-rich domain of the cell-surface receptor tumor necrosis factor 1 which is integral to binding of the cytokine TNF-α, namely, TNFR-1 CRD2. Racemic protein crystallography at 1.4 Å confirmed that the native binding conformation was preserved, and TNFR-1 CRD2 maintained its capacity to bind to TNF-α (KD ≈ 7 nM). Encouraged by this discovery, we carried out mirror-image phage display using the enantiomeric receptor mimic and identified a d-peptide ligand for TNFR-1 CRD2 (KD = 1 μM). This work demonstrated that cysteine-rich domains, including the central domains, can be chemically synthesized and used as mimics for investigations.


Optimization of buffer pH and temperature
Reduced TNFR-1 CRD2 (0.5 mg/mL -measured by UV absorbance 1 ) was dissolved in 200 μL of denaturation buffer in a 2 mL centrifuge tube ( i & iii.6 M Gn•HCl, 0.1 M Tris, 6 mM glutathione disulfide, 60 mM glutathione, pH 8.5 ; ii & iv.6 M Gn•HCl, 0.1 M phosphate, 6 mM glutathione disulfide, 60 mM glutathione, pH 6.5 ).Each reaction mixture i-iv was diluted five-fold with 800 μL renaturation buffer ( i & iii.0.1 M Tris, pH 8.5 ; ii & iv.0.1 M phosphate, pH 6.5).Folding reactions proceeded for 96 hours, either at room temperature (i-ii) or at 4 °C (iii-iv).100 μL of reaction was removed, filtered through glass wool, and analyzed by LCMS.The refolding yield for each condition was estimated by the relative integration of the peak shifted up field by 1 min in the HPLC chromatogram at 210 nm (Table S7.2).
Condition ii produced the highest yield (33%) and moved onto the next round of optimization below.

Optimization of reactant concentrations
Reduced TNFR-1 CRD2 (0.5 mg/mL i-ii, or 2 mg/mL iii-v -measured by UV absorbance 1 ) was dissolved in 200 μL of denaturation buffer in a 2 mL centrifuge tube (6 M Gn•HCl, 0.1 M phosphate, pH 6.5 containing ; i, iii & v. 6 mM glutathione disulfide, 120 mM glutathione ; ii & iv. 25 mM glutathione disulfide, 50 mM glutathione).Each reaction mixture i-v was diluted five-fold with 800 μL renaturation buffer (0.1 M phosphate, pH 6.5).Folding reactions proceeded for 96 hours at room temperature.100 μL of reaction was removed, filtered through glass wool, and analyzed by LCMS.The refolding yield for each condition was estimated by the relative integration of the peak shifted up field by 1 min in the HPLC chromatogram at 210 nm (Table S7.2).Neither condition produced in improvement in refolding yield (<21%).Higher TNRCD2 concentrations were desirable as this would result in smaller reaction volumes during scale-up preparations.However, folding at 0.4 mg/mL (iii-v) gave significantly reduced yield (5-13%).Therefore, the best condition (ii) from the previous round was used for TNRCD2 refolding.

Component
Oxidation method Buffer pH and temperature Concentrations

S2. Racemic protein crystallography of TNFR-1 CRD2
Figure S2.1:Circular dichroism spectrum of L-TNFR-1 CRD2 in ddH 2 O (40 M), recorded from 250 nm to 190 nm at 20 C using an Applied Photophysics Chirascan CD spectrometer.Each measurement was performed in triplicate using a sample cell with a 0.1 mm path, 1 nm bandwidth and 0.5 s per point.CD spectra of the solvents were subtracted from the CD spectra of the proteins to eliminate interference.The units of ellipticity are expressed as the mean residue ellipticity Table S2.1:X-ray data collection, processing, and refinement statistics.

Amino acid sequence
.1: Sequence consensus of the enriched peptide library following the second biopanning experiment, using a CX 4 WLGX 2 C library format.

TCPB solid-phase oxidation
To obtain high purity samples of the cyclic peptide conformer for kinetic analysis, and access the remaining conformer (disulfide 1-11), a solid-phase synthetic approach was devised.The procedure utilizes the differential reactivity of iodine towards orthogonally protected cysteine residues, based on a previously reported procedure in solution-phase. 2The method was optimized here for application on the solid-phase.

Materials and instruments
Unless otherwise stated, chemicals and solvents were purchased from commercial suppliers (Sigma Aldrich, Fluorochem, Acros Organics, Alfa Aesar, Cambridge Bioreagents and Fisher Scientific) and used without further purification.HPLC grade (>99.8%)dimethylformamide was used for peptide synthesis.Rink amide ProTide resin was purchased from CEM and 2-Cl-trt Fmoc-hydrazine resin was prepared as previously described. 3LCMS data was obtained using an Agilent Infinity 1260 II HPLC system fitted with an on-line Agilent 6120 quadruple ESI-MS.Preparative HPLC was carried out using a Shimadzu Nexera preparative HPLC system.All HPLC systems used UV analyte detection at 210 nm and 280 nm.Automated SPPS was performed using a Liberty Blue microwave peptide synthesiser (CEM corp.).Peptide lyophilization was carried out by flash freezing the sample in liquid N2 and drying on a Christ Alpha 2-4 LDplus freeze-dryer.The pH measurements were conducted using a Mettler Toledo FiveEasy Plus pH meter fitted with an FP20-Micro glass electrode.For measurements in 6 M Gdn•HCl, the measured pH value was assumed to be 0.8 units lower than the actual value. 4Circular dichroism (CD) spectra were collected using an Applied Photophysics Chirascan CD spectrometer.Protein crystallization screening was conducted using a Douglas Instruments Oryx 4 crystallization robot.Grating coupled interferometry experiments were conducted using a Creoptix WAVEsystem.

Automated SPPS of ligation fragments
Automated SPPS was conducted at a 50 μmol scale using modified CEM CarboMax coupling cycles. 5oc-amino acid stock solutions, oxyma and DIC were used at 0.2 M concentration.Reactions were stirred by N 2 bubbling for 2 seconds on, 3 seconds off.
Fmoc deprotection: Piperidine in DMF (3mL, 20% v/v) was delivered to the reaction vessel.Microwave heating proceeded as follows: 0 W 20±5 °C for 5 s, 100 W 78±2 °C for 20s, 60 W 88±2 °C for 10s, 20 W 90±1 °C for 60s.The resin was then washed with DMF (4 x 2 mL).For residues 88-93 of the N-terminal cysteine peptides, the deprotection was conducted at room temperature by delivery of piperidine in DMF (3mL, 20% v/v) for 3 minutes.The resin was drained and the deprotection repeated for 7 minutes.The resin was then washed with DMF (4 x 2 mL).
The resin was then washed once with DMF (2 mL).
For single couplings, reactions proceeded for a total coupling time of 4 mins (X = 225 s).For double couplings (Arg, Cys and Val), the coupling was repeated.Due to high cost of diastereomeric D-Ile and its slower coupling rate (β-branched), a single 8 min coupling (X = 465 s) was implemented.

Peptide hydrazide preparation
Peptides were assembled by either manual or automated SPPS onto a 2-chlorotityl Fmoc-hydrazine resin 3 and subject to peptide cleavage.Because during automated SPPS, the mildly acidic oxyma (pK a 4.60) can cause premature release of the peptide from a 2-Cl-(Trt) resin at 90 °C, DIPEA (20 μM) was added to the oxyma solution to minimize premature cleavage and increase yields of peptide hydrazide. 5terminal biotinylated peptide linker for TNRCD2 Rink amide ProTide resin (0.1 mmol, 0.19 mmol/g, CEM) was swollen in 50%/50% v/v DMF/DCM for 10 mins in an SPE column and drained.The Fmoc-protecting group was removed by addition of 20% piperidine in DMF (3 mL) to the resin for 2 x 5 mins.The resin was washed five times with DMF (3 mL).segment (Cys88-Asn116) was assembled onto the linker using automated SPPS.

Peptide cleavage
The resin was washed with DMF (3×3 mL), DCM (3×3 mL) and Et 2 O (3×3 mL).Cleavage cocktail was added to the resin and allowed to stir for 120 mins at room temperature.
Two cleavage cocktail variations were used: Cleavage reagent B containing 8.75 mL TFA, 0.25 mL triisopropylsilane, 0.5 g phenol and 0.5 mL water was used for TCPB-E peptides following solid-phase oxidation.
Cleavage reagent K containing 8.25 mL TFA, 0.25 mL EDT, 0.5 mL H 2 O, 0.5 mL thioanisole and 0.5 g phenol was used for all other peptides.
The cleavage mixture was drained from the SPE column into a 50 mL centrifuge tube and the mixture was concentrated under a stream of N 2 to <3 mL.The peptide was precipitated using ice cold Et 2 O and collected by centrifugation at 3500 RCF.The crude peptide was triturated twice with Et 2 O, dissolved in 1% acetic acid (20 mL) and lyophilized.Crude peptides were analysed by LCMS and purified using preparative HPLC.

Peptide LCMS analysis
Peptide samples were prepared at 0.1 mg/mL using 0.1% TFA in water and passed through a 0.22 μM nylon filter.Unless otherwise stated, samples (10 μL) were eluted with reversed mobile phase A (water + 0.1% formic acid) and B (acetonitrile + 0.1% formic acid) at 0.3 mL/min over a RP-C18 column (ACE, 2.1 mm x 100 mm, 110 Å, 3 μm) at 40 °C.A 5-70% gradient of A/B was applied over 30 minutes and analyte was detected using a UV detector at 210 nm and 280 nm, and positive electrospray ionisation mass spectrometry (ESI+ MS).ESI+ mass spectra are reported as the integrated spectra for the duration of the major peak in each UV210-nm chromatogram.
Peptide preparative HPLC Samples were passed through a 0.22 μM nylon filter.3-10 mL of sample was eluted with reversed mobile phase A (water + 0.1% TFA) and B (acetonitrile + 0.1% TFA) at 18 mL/min over a RP-C18 column (Shimpack GIST, 20 mm x 150 mm, 100 Å, 5μm) at room temperature.A 20-60% gradient of A/B was applied over 40 minutes and analyte was detected using a photodiode array detector at 210 nm and 280 nm.Sample fractions were collected using an automated fraction collector, their identities were confirmed by LCMS and the fractions containing the target peptide were combined and lyophilized

Figure S4. 2 :
Figure S4.2:Solution-phase oxidation of TCPB-E with cystine (left) or glutathione disulfide (right) for deduction of active bacteriophage conformation.See below for identification of peaks by disulfide bond mapping experiments.Peak corresponding to dimerized product with three disulfide bonds indicate by asterisk (*).

Figure S4. 3 :
Figure S4.3:LCMS disulfide bond mapping of isolated TCPB-E peak ii, oxidized with cystine.Lyophilized peptide was digested with chymotypsin (0.1 mg/mL) in phosphate buffer ( 0.1 M, pH 6) at room temperature for 6 hours.HPLC trace of undigested peptide is shown in top left, and HPLC trace and corresponding ESI+ MS of major digested peak is shown in bottom left.Identity of the major digested fragment and the deduced, undigested conformer are shown to the right.No digestion was observed between Phe2 and His3 after 6 hours.

Figure S4. 4 :
Figure S4.4:LCMS disulfide bond mapping of isolated TCPB-E peak i, oxidized with cystine.Lyophilized peptide was digested with chymotypsin (0.1 mg/mL) in phosphate buffer ( 0.1 M, pH 6) at room temperature for 6 hours.HPLC trace of undigested peptide is shown in top left, and HPLC trace and corresponding ESI+ MS of major digested peaks are shown in bottom left.Identity of the major digested fragments and the deduced, undigested conformer are shown to the right.

Figure S4. 5 :
Figure S4.5:LCMS disulfide bond mapping of isolated TCPB-E peak ii, oxidized with glutathione disulfide.Lyophilized peptide was digested with chymotypsin (0.1 mg/mL) in phosphate buffer ( 0.1 M, pH 6) at room temperature for 6 hours.HPLC trace of undigested peptide is shown in top left, and HPLC trace and corresponding ESI+ MS of major digested peak is shown in bottom left.Identity of the major digested fragment and the deduced, undigested conformer are shown to the right.No digestion was observed between Phe2 and His3 after 6 hours.

Figure S4. 6 :
Figure S4.6:LCMS disulfide bond mapping of isolated TCPB-E peak i, oxidized with glutathione disulfide.Lyophilized peptide was digested with chymotypsin (0.1 mg/mL) in phosphate buffer ( 0.1 M, pH 6) at room temperature for 12 hours.HPLC trace of undigested peptide is shown in top left, and HPLC trace and corresponding ESI+ MS of major digested peaks are shown in bottom left.Identity of the major digested fragments and the deduced, undigested conformer are shown to the right.

Figure S4. 7 :
Figure S4.7:Grating-coupled interferometry (GCI) analysis of TCPB-E isomers binding to D-TNFR-1 CRD2.Repeated analyte pulses of increasing duration (RAPID) were passed over the sensor surface, with fixed concentrations of each peptide conformer shown above.Sensograms shown are blank subtracted (target flow cell -flow cell with no target).The highest sensor response with respect to concentration was observed for TCPB-E-C1.
Briefly, in non-polar solvents, Iodine selectively removes Cys(Trt) groups and oxidizes cysteines to form disulfide bonds, with Cys(Acm) left intact.In the more polar DMF, Cys(Acm) is removed and enables formation of the second disulfide bond.This method circumvents the use of the commonly employed Cys(Mmt) and N-chlorosuccinimide, due to the high cost of the Fmoc-D-Cys(Mmt)-OH building block.TCPB sequence was assembled by automated SPPS (See S5) on a low-loading PEG-PS rink amide resin (CEM, 0.19 mmol/g) to minimize unwanted intermolecular disulfide bond formation.For TCPB-E-C1, Fmoc-Cys(Trt)-OH was used at position 1, and Fmoc-Cys(Acm)-OH used at positions 4 and 11.For TCPB-E-C4, Fmoc-Cys(Trt)-OH was used at position 4, and Fmoc-Cys(Acm)-OH used at positions 1 and 11.First, the intermolecular disulfide bond with cysteine was formed.Boc-Cys(Trt)-OH (2 equiv., 60 mM) was dissolved in 50% TFE in DCM and added to the peptide resin.Then, an equal volume of Iodine (1 equiv., 30 mM) in 50% TFE in DCM was also added to the resin, and the slurry stirred for 10 mins at room temperature.The resin was washed with DCM (3 mL) and the reaction was repeated.The resin was washed once with DCM (3 mL) and three times with DMF (3 x 3 mL).For intramolecular disulfide bond formation, ten equivalents of Iodine in DMF (0.5 M) was added to the resin, and the slurry stirred for 60 mins at room temperature.The resin was then washed three times with DMF (3 × 3 mL), once with 1 M aq.ascorbic acid (3 mL), three times with water (3 × 3 mL) and three times with DMF (3 × 3 mL).The peptide was cleaved from the resin and isolated by preparative HPLC.

Figure S4. 8 :
Figure S4.8:Illustrative representation of a solid-phase synthetic route to TCPB-E peptide conformers.An orthogonal protection scheme was implemented using cysteine acetamidomethyl (acm) or trityl (trt).Disulfide bond formation was controlled by the relative reactivity of Iodine towards protected cysteine residues in non-polar (trt-trt) and polar (acm-acm) solvents, based on a modified procedure.2

Figure S4. 9 :
Figure S4.9:Comparison of HPLC retention times of TCPB-E-C1 and -C4 prepared on solid-phase with the solution-phase oxidation of TCPB-E with cystine, reinforcing the results of LCMS disulfide bond mapping.Small quantities of TCPB-E-C4 appear to have been formed in solution-phase but was insufficient for isolation, likely suggesting that this conformer is thermodynamically unfavorable.

Figure S4. 10 : 1 Figure S5. 1 :
Figure S4.10:Grating-coupled interferometry (GCI) analysis of TCPB-E-C1 and -C4 isomers binding to D-TNFR-1 CRD2.A single cycle of association and dissociation were passed over the sensor surface, with 50 μM of each peptide conformer.Sensograms shown are blank subtracted (target flow cell -flow cell with no target).The highest sensor response with was observed for TCPB-E-C1, with little binding observed for TCPB-E-C4.The TCPB-E-C1 conformer was thus identified as the active form of the TCPB peptide.

2
equiv. of Fmoc-Lys(Mtt)-OH (200 mM), 1.95 equiv. of HBTU (195 mM), 2 equiv. of HOBt (200 mM) and 4 equiv. of DIPEA (400 mM) was dissolved in DMF and mixed for 0.5 min.The coupling mixture was transferred to the resin and allowed coupling to proceed for 30 mins at room temperature.The remainder of the flexible linker containing Gly-D-Ser-Gly-D-Ser-Gly was assembled using manual SPPS procedure (Section 5.3.1), with the N-terminal Fmoc group left in place.The lysine side chain protecting group, 4-methyltrityl (Mtt) was removed using 1% TFA in DCM through 14 flow washes (3 mL each), monitored qualitatively by the intense yellow colour of the Mtt-OH group.The resin was washed three times with DCM (3 mL) and three times with DMF (3mL).Biotin-N-hydroxysuccinimde ester (0.2 mmol) was dissolved in DMF (10 mL) and added to the resin along with DIPEA (0.4 mmol).The resin slurry was heated by microwave irradiation in the Liberty Blue peptide synthesizer.Microwave heating proceeded as follows: 15 W 75±2 °C for 15 s, 30 W 90±1 °C for 225 s.The resin was then washed once with DMF (2 mL) and the biotinylation reaction was repeated.The remainder of the TNFR-1 CRD2

Table S3 .1: Results
of bacteriophage biopanning experiment 1 with the sequence format CX 9 C.

Table S3 . 2 :
Next generation sequencing of enriched CX 9 C bacteriophage library following screening experiment 1.A total of 20282 sequences were identified, with sequences in abundance >40 listed below, in decreasing order.

Table S3 . 4 :
.3: Results of bacteriophage biopanning experiment 2 with the sequence format CX 4 WLGX 2 C.Enrichment factor for each round is defined as the amount of phage eluted from the target well containing D-TNFR-1 CRD2 divided by the amount of phage eluted from the target well without immobilized target (streptavidin/neutravidin only).Sequencing of enriched CX 4 WLGX 2 C bacteriophage library following screening experiment 2. A total of 22 monoclones were randomly selected, with sequences in listed below, in decreasing order of abundance.