A recombinant approach for stapled peptide discovery yields inhibitors of the RAD51 recombinase

Stapling is a macrocyclisation method that connects amino acid side chains of a peptide to improve its pharmacological properties. We describe an approach for stapled peptide preparation and biochemical evaluation that combines recombinant expression of fusion constructs of target peptides and cysteine-reactive divinyl-heteroaryl chemistry as an alternative to solid-phase synthesis. We then employ this workflow to prepare and evaluate BRC-repeat-derived inhibitors of the RAD51 recombinase, showing that a diverse range of secondary structure elements in the BRC repeat can be stapled without compromising binding and function. Using X-ray crystallography, we elucidate the atomic-level features of the staple moieties. We then demonstrate that BRC-repeat-derived stapled peptides can disrupt RAD51 function in cells following ionising radiation treatment.


Materials
Synthesised by Sigma-Aldrich.See Table S2.
Purified proteins and peptides

GB1-peptide fusions This work
Un-tagged BRC repeat peptides This work Fluorescein-BRC4 Gift from Dr Laurens Lindenburg, previously reported in [1] Human RAD51 This work HumRadA22 protein This work, previously described in [2] Lysozyme Sigma-Aldrich described previously [3] DVT linker Spring lab, University of Cambridge, described previously [4] TCEP Melford T26500 Small-scale preparation of GB1-fused stapled peptides Peptide-coding DNA was designed using the DNAWorks online application, [5] optimising codon usage for E. coli expression.DNA oligonucleotides for the assembly of these fragments were generated using the same software and 15-20 nt linkers were appended to the 5' ends of the outermost forward and reverse oligos for sequence and ligation independent cloning (SLIC).Oligonucleotides used for assembly of the fragments are shown in Table S2.
The inserts were then synthesised by assembly PCR using Phusion DNA polymerase (New England Biolabs) with standard reaction conditions.Each assembly PCR reaction contained 1 μM of the outermost oligos and 0.02 μM of each internal oligo.Inserts were then purified by gel extraction using the GeneJet gel extraction kit (ThermoScientific).Sequence and ligationindependent cloning (SLIC) was used to clone the inserts into the pPEPT1 vector (Addgene #195001, Figure S1) digested with the BsaI restriction enzyme.Using BsaI allows the peptide to be cloned in a seamless fashion as cleaves outside of its recognition site.10 µl SLIC reactions contained purified insert at 2-10 ng/μl and digested vector at 4-20 ng/μl in 1x NEB Buffer 2.1 (New England Biolabs).0.6 U of T4 DNA polymerase (New England Biolabs, M0203S) was added to each reaction and incubated for 1-5 min at RT, after which either dGTP or dCTP was added to a final concentration of 10 mM to stop exonuclease activity.Reaction was further incubated for 1 minute at RT, after which it was heated for 5 minutes at 65 °C in a PCR cycler to deactivate the T4 polymerase, after which the PCR tubes were left at room temperature for 10-20 minutes for the complementary resected ends of the insert and vector to anneal.Reaction mixtures were then used to transform 50 μl of chemicompetent T7Express E. coli cells by heat shock and transformants were plated on LB agar plates supplemented with ampicillin (100 μg/ml).Individual colonies were transferred to a replica plate and presence of insert was determined by colony PCR using the forward primer used for the assembly PCR and T7 terminator primer (GCTAGTTATTGCTCAGCGG).For expression, clones carrying an insert were used to inoculate 10 ml 2xYT bacterial cultures supplemented with ampicillin (100 μg/ml).The cultures were grown in a 50 ml centrifuge tube at 37 °C overnight with the tube lid slightly unscrewed and fixed with tape to ensure aeriation.Next day, protein expression was induced by the addition of IPTG (400 μM) for three hours at 37°C, after which cells were harvested by centrifugation.Cell pellets were then either frozen for future use or used directly for purification of peptides.Correct inserts were confirmed later by Sanger sequencing with the T7 terminator primer.
Lysate was incubated for 10 min at room temperature on a rotating mixer.Lysates were spun down in a 2 ml tube on a bench-top centrifuge at 15 000 g for 10 min and supernatant collected by aspiration.200 μl of 50% (v/v) slurry of Ni-NTA agarose resin (Cube Biotech, !31103) was washed twice with 1 ml of MilliQ water and resuspended in 200 μl PBS.The resin was mixed with the soluble lysate and incubated on a rotating mixer for 5 min at room temperature, after which it was applied in two portions to a 0.5 ml micro-spin chromatography column and centrifugated for 1 min at 1000 xg to remove flow-through.Same centrifuge settings were also used for subsequent wash and elution steps.The resin was washed with a total of 1 ml of PBS + 20 mM imidazole containing 1 mM TCEP, followed by a 0.5 ml wash using the same buffer without any reducing agent.The second wash step is essential for the removal of any residual TCEP that can form undesired side-products upon reaction with the divinyl-heteroaryl linker.
The GB1-BRC repeat was eluted with 0.5 ml PBS + 200 mM imidazole, and the elution immediately used for subsequent stapling reactions.
The eluted sample was split into two 250 μl parts. 2 mM DVT or DVP linker solution in DMSO was gradually titrated into the stapling reaction to achieve pseudo-dilution conditions.Different linker titration schemes were initially trialled and are shown in Figure S2.Most optimal linker titration was observed for reaction h (Figure S2) and was used for subsequent preps.At the same time, an identical volume of DMSO control without any linker was added to the other 250 μl peptide solution.1 mM TCEP was added to the control reaction but not the stapling reaction to maintain free sulfhydryl groups in the control peptides.At the end of the titration, reactions were quenched with 2 mM DTT.

Preparation of stapled peptides in un-tagged form
Peptides were cloned in an identical fashion to the GB1-peptide-His8 constructs, except different expression vectors, pOP3BT and pEXP-GB1, were used (Addgene #112603 and #112565, respectively).The vectors contain an N-terminal instead of a C-terminal His-tag.
T7Express E. coli cells carrying the plasmids expressing GB1-fused BRC repeat were plated directly from glycerol stocks of sequence-verified clones onto LB agar supplemented with ampicillin (100 μg/ml) and grown overnight at 37 °C.Next day, cells were scraped to inoculate separate flasks containing 1 L of 2x YT medium supplemented with 100 μg/mL ampicillin.
Cultures were grown at 37 °C until OD600 of ~1, after which expression was induced with 400 μM IPTG for 3 h.Cells were resuspended in 25 mL of IMAC buffer A (50 mM Tris-HCl pH=8.0, 150 mM NaCl, 20 mM imidazole) and frozen.Later, cells were thawed and supplemented with DNase I (100 μL, 2 mg/mL) and AEBSF (1 mM), and lysed on an Emulsiflex C5 homogenizer (Avestin) or by sonication.Cell lysate was centrifuged at 40000 xg for 30 min and supernatant collected.GB1-BRC lysate was loaded on a 3 mL Ni-NTA agarose matrix (Cube Biotech), after which column matrix was washed with 10 column volumes Nickel Buffer A. GB1-BRC repeat was eluted with 12 ml nickel buffer B (50 mM Tris-HCl pH 8.0, 150 mM NaCl, 200 mM imidazole).The eluent was buffer exchanged back into nickel buffer A on a PD-10 desalting column (Cytiva).Buffer exchanged GB1-BRC fusion (~18 ml) was incubated with 100 μL of 2 mg/ml TEV protease overnight at 4°C.The GB1 tag was then removed from the solution by a second Ni-NTA affinity step, collecting the flowthrough that contains the BRC peptide.The flow-through was acidified with HCl to pH 2-4 and acetonitrile was added to 10%, after which the solution was centrifuged at 10000 xg for 15 min and supernatant collected.The acidified flow-through was then applied to an ACE C8 300 4.6 x 250 mm semi-prep RP-HPLC column equilibrated with 10 % MeCN + 0.1% TFA and peptides were eluted with a 20 column volume gradient to 90 % MeCN + 0.1% TFA.BRC repeat peptides typically elute at 20-40% of the gradient.Peak fractions corresponding to the cleaved peptide were pooled and diluted 5x in PBS + 10 mM EDTA in a 50 ml centrifuge tube.
A small stirrer bar was added to the tube, which was then placed on a magnetic stirrer.A syringe was filled with 20 mM linker in DMSO which was then gradually added to the mixture by piercing the centrifuge tube lid.To maintain pseudo-dilution conditions, linker was injected in 50 μl increments every two minutes, to a final concentration of 2 mM, ensuring at least 2x stoichiometric excess of linker over peptide.The reaction mixture was then quenched with 5 mM TCEP, filtered through a 0.45 μm filter and acidified with HCl to pH ~3.Stapled peptide was then purified by a on an ACE C18 300 4.6 x 250 mm semi-prep RP-HPLC column with a 0-100% gradient of A: 10 % MeCN + 0.1% TFA, B: 90 % MeCN + 0.1% TFA.Peak fractions containing the desired product were pooled and dried under vacuum.

Purification of HumRadA22
HumRadA22 is an archaeal RadA mutant with surface residues exchanged for the human Rad51 sequence and can be used as a bona fide mimic of monomeric human Rad51.
The protein was prepared as described previously. [2]rification of full-length human RAD51 oligomerises on the heparin matrix, which acts as a DNA mimic, and dissociates from the MBP-BRC4 fusion, which is removed in flow-through and wash steps.Column was washed with 8 CV Heparin-A buffer, after which the protein was eluted with a 20 CV, 0-100% linear gradient of Heparin-B buffer (20 mM HEPES pH 7.4, 1 M NaCl, 1 mM EDTA, 1 mM TCEP).
RAD51 was concentrated, flash-frozen with liquid nitrogen and stored for future use.

Fluorescence polarisation competition assay
FP competition assay was a modified version of a protocol described previously. [2]uorescein-labelled HsBRC4 probe for this assay was kindly provided by Dr Laurens KD values were estimated from the fitted IC50 parameters using a previously reported equation. [6]sothermal titration calorimetry Peptides were resuspended in MilliQ water to 10 times the desired final concentration.
This was then diluted 10x with the ITC buffer to obtain the final titrant solution (20 mM CHES pH 9.5, 150 mM NaCl, 0.1% Tween-20).HumRadA22 was buffer-exchanged on a NAP-5 desalting column (Cytiva) into ITC buffer and protein concentration was adjusted to 10:9 of the desired final value.One ninth volume of MilliQ water was added to the solution to bring the protein concentration to the desired final value, while maintaining identical buffer:MilliQ volume proportions in both the syringe and the cell.ITC was carried out using a Microcal ITC200 or Malvern PEAQ ITC instruments at 25°C with a 5.00 μCal reference power DP value, stirring speed of 500-750 rpm, 2 sec filter period.Injection spacing, speed and volume, cell/syringe concentrations as well as the number of injections were adjusted for each peptide and its binding properties.ITC data were fitted using a single-site binding model using the Microcal ITC data analysis program in the Origin 7.0 package.Data points affected by baseline spikes were omitted from the analysis.

Circular dichroism spectroscopy
Dried peptides were dissolved in MilliQ water to 0.3 mg/ml, and then two-fold diluted in 20 mM sodium phosphate, pH 7.4, giving a final solution of 0.15 mg/ml peptide in 10 mM sodium phosphate.CD spectra of selected peptides were recorded on an AVIV 410 circular dichroism spectropolarimeter using a 1 mm path length quartz cuvette.Measurements were done at 25 °C, with a 185-260 nm range, 1 nm bandwidth, 5 s averaging time and 0.3 s settling time.Spectra were prepared as smoothed average of three scans and normalised against blank solvent.

Electrophoretic mobility shift assay (EMSA)
The ability of linear and stapled BRC repeat peptides to dissociate RAD51-ssDNA nucleofilament was evaluated using an electrophoretic mobility shift assay (EMSA).RAD51 DNA-binding reactions (40 μl) were set up in 50 mM HEPES pH 7.4, 150 mM NaCl, 10 mM MgAc2, 2 mM CaCl2, 1 mM TCEP, 1 mM ATP. 5 μM full-length human RAD51 was incubated with varying concentrations of BRC repeats for 10 min at room temperature, followed by the addition of 100 nM fluorescently labelled FAM-dT60 oligonucleotide, and further incubation at 37 °C for 10 min.Control reactions were set up with free FAM-dT60 probe and FAM-dT60 Pantelejevs, T. Materials and methods 120 + 5 μM RAD51.10 μl of reactions were then loaded on a 1xTBE non-denaturing acrylamide gel (5%) and run at 100 V for 1:30 h at 4 °C.The gel was directly visualized on a Typhoon FLA 9000 imager (GE Healthcare) using FAM channels.

X-ray crystallography
Stapled peptide complexes were re-constituted from purified peptides and HumRadA22.Peptides were added at a 1.5 stoichiometric excess to HumRadA22 in its sizeexclusion buffer, to a final concentration of 0.75 and 0.5 mM for the peptide and protein, respectively.ADP and MgCl2 were added to the protein solutions in for some of the complexes (see Table S1).Crystallisation screening was done in 96-well MRC plates using the sittingdrop vapor diffusion technique and a variety of commercial crystallisation screens.A Mosquito liquid handling robot (TTP Labtech) was used to dispense protein and reservoir solutions in sub-microlitre volumes.Typically the two sitting drops contained 200 or 400 nl of protein solution and 200 nl of crystallisation solution, while the reservoir contained 80 μl of crystallisation solution.Plates were stored at 17 °C in a RockImager crystallisation hotel (Formulatrix) and imaged regularly.Crystal hits were flash-frozen in liquid nitrogen using cryoloops.Additional cryoprotectant was not added before freezing of crystal hits.Diffraction data were collected on Diamond Light Source (Harwell, UK) MX beamlines.Full native datasets with goniometer sweeps of at least 180° were collected to ensure completeness of diffraction data.Molecular replacement phasing method was used with the apo HumRadA22 structure (PDB: 5KDD) as a search model.Molecular replacement was done with Phaser. [7]e structures were refined without BRC repeats first and the peptides were built into the clearly visible electron density manually.Manual refinement was done in Coot [8] and automated refinement with phenix.refine [7]and autoBUSTER. [9]Crystallisation conditions, as well as data collection and refinement statistics, are provided in Table S1.The coordinates have been deposited in the Protein Data Bank under accession codes 8C3J (SP2), 8BR9 (SP24) and 8C3N (SP30).
After 24 hours, medium was replaced by fresh medium containing or not SP30 or SP31 or L31 (40 μM).Cells were incubated for an hour before they were irradiated (3 Gy) and allowed to recover for 3 hours before being fixed.Non-irradiated cells treated in a similar way were used as a control.RAD51 foci were detected following a protocol previously described.[  Optimisation of the small-scale stapling reaction SP31 in vitro binding data to HumRadA22

Lindenburg(
Hollfelder group, Department of Biochemistry, University of Cambridge).Black 384-well flat-bottom microplates (Corning, 3821) were used with a 40 μl final reaction volume in all measurements.Following buffer conditions were used: 20 mM CHES pH 9.5, 150 mM NaCl, 0.1% BSA, 0.1% Tween-20.Each reaction contained 100 nM HumRadA22 and 10 nM BRC4-fluorescein.Two-fold serial dilutions of stapled peptides were added to the reactions.A free probe control reaction containing only 10 nM BRC4-fluorescein was used to calibrate gain and focal height.FP measurements were performed on Pherastar FX or Clariostar Plus (BMG Labtech) plate readers equipped with an FP 485-520-520 optic modules.Binding curves were fitted using the four-parameter logistic model with a variable Hill slope using Prism software (Graphpad).Regression fitting was performed using the least squares optimisation algorithm.

Figure S1 .
Figure S1.Map of the pPEPT1 plasmid.(A) circular plasmid map of the pPEPT1 vector with key features and unique restrictions sites in the fusion part and in the multiple cloning site.(B) Focused view of the DNA and protein features of the part where the peptide-encoding sequences are inserted.

Figure S2 .
Figure S2.Optimisation of small-scale stapling reaction.Lower-case letters a-i represent different conditions.For each condition, the volumes (µl) of linker addition are depicted at appropriate time-points under the yellow circles, are which is relative to the volume for more visual interpretation of the experiments.pH and presence of TCEP are also indicated.ESI-MS mass spectra for each condition are shown below.

Figure S3 .
Figure S3.(A) ITC and (B) fluorescence polarisation competition measurements demonstrating that poly-arginine containing SP31 similarly to SP30 has high in vitro binding affinity for HumRadA22.

Figure S6 . 4 SP12Figure S17 .Figure S19 .
Figure S6.(A) Quantification of RAD51 foci in U2OS cells from five independent experiments in the presence of 40 μM L31 or vehicle (control).(B) Representative IF images depicting RAD51 signal.Representative cells shown in Figure S5A are indicated with an arrow and they were chosen as the median RAD51 foci value for each of the images.Foci counts determined automatically using SImA software are provided next to each cell.Scale bar 20 μm.(C) Dot plot depicting the number of RAD51 foci per cell for one of the independent biological experiments.Red bars represent the mean values.
Cells were plated from a glycerol stock on LB agar supplemented with kanamycin (25 μg/mL) and chloramphenicol (34 μg/mL), and grown overnight at 37°C.Next day, cells were scraped and used to inoculate 1 L of 2xYT medium supplemented with same antibiotics.Cells were grown at 37°C with shaking at 200 RPM until an OD600 = 0.6, afer which they were Full-length HsRAD51 was prepared based on a protocol developed at the lab of Prof Luca Pellegrini (Department of Biochemistry, University of Cambridge).E. coli BL21(DE3) Rosetta2 cells (Novagen) carrying a pRSF-Duet plasmid co-expressing wild-type HsRAD51 and a BRC4 sequence fused to an N-terminal His-MBP tag were kindly provided by Dr Joseph Maman.CV Ni-A300 buffer, after which MBP-BRC4:RAD51 complex was co-eluted with buffer Ni-B-300 (50 mM Tris-HCl, pH 8.0, 300 mM NaCl, 200 mM imidazole, 1 mM TCEP).The sample was then diluted with Heparin-A buffer (20 mM HEPES pH 7.4, 50 mM NaCl, 1 mM EDTA, 1 mM TCEP) and loaded on a HiTrap Heparin HP 5 ml column.During this step, RAD51

Table S1 .
Crystallographic data collection and refinement.Values in parentheses are for the high-resolution cell.
1] In summary, after being washed with PBS, cells were fixed with paraformaldehyde, 4% in PBS

Table S2 .
Oligonucleotides used for the cloning of peptide constructs.