Macrocyclic Modalities Combining Peptide Epitopes and Natural Product Fragments

“Hot loop” protein segments have variable structure and conformation and contribute crucially to protein–protein interactions. We describe a new hot loop mimicking modality, termed PepNats, in which natural product (NP)-inspired structures are incorporated as conformation-determining and -restricting structural elements into macrocyclic hot loop-derived peptides. Macrocyclic PepNats representing hot loops of inducible nitric oxide synthase (iNOS) and human agouti-related protein (AGRP) were synthesized on solid support employing macrocyclization by imine formation and subsequent stereoselective 1,3-dipolar cycloaddition as key steps. PepNats derived from the iNOS DINNN hot loop and the AGRP RFF hot spot sequence yielded novel and potent ligands of the SPRY domain-containing SOCS box protein 2 (SPSB2) that binds to iNOS, and selective ligands for AGRP-binding melanocortin (MC) receptors. NP-inspired fragment absolute configuration determines the conformation of the peptide part responsible for binding. These results demonstrate that combination of NP-inspired scaffolds with peptidic epitopes enables identification of novel hot loop mimics with conformationally constrained and biologically relevant structure.


1) Supplementary Figures and Tables
It can be noted that the independent enantiopure synthesis of K1-minor was challenging and a small amount of the desired PepNat was formed (middle HPLC profile). This reveals the robustness of the imine/cycloaddition strategy on resin allowing to isolated mg amount of both diastereomers. The NMRs were acquired using 700 MHz in CD3OD at room temperature. The comparison demonstrates that the major (top) and minor (bottom) diastereomers of PepNat K1 have different chemical shifts. Therefore, the two diastereomers adopt two distinct conformations in solution, as demonstrated for PepNat H1 (for details see Section 6).
PepNat K1-minor conformers that complied with those distance restraints were subjected to solvent explicit 10 ns MD simulations, their trajectories clustered and the most populated clusters for each conformation selected. MSpin NOE Fitter algorithm was utilized to select which cluster agreed best with the NMR experimental data.
Example sensorgrams for selected PepNats D1−D4, D7, D8, D10, D12 and D18 interacting with immobilized hSPSB2 using SPR (red). Each PepNat was injected at 0.020 µM, 0.078 µM, 0.313 µM, 1.25 µM and 5 µM concentration in quick succession followed by a longer dissociation time. A binding kinetic model was fitted to each experimental trace (black). All experiments were conducted in triplicates. Example concentration response curves for selected PepNats in MC1R, MC3R, MC4R, MC5R radioligand binding assays and MC1R, MC3R, MC4R cAMP assays. Responses were normalized to reference agonist NDP-α-MSH and presented as % effect of control, with full competition in binding defined as -100% and full agonism in cAMP assays defined as 100%. Data were fitted using a four-parameter logistic fit and derived IC50 and EC50 values are reported in table S5. Curve fits were removed for compounds not reaching an upper plateau in a cAMP assay or if considered inactive in a binding assay. Threshold to define a compound as active was set to -30% in binding assays and 20% in cAMP assays respectively. Data are from at least three independent experiments and error bars represent ± Standard Error of the Mean (SEM).  The ALFPGF sequence was bound to solid support using Rink Amide low loading resin (loading 0.36 mmol/g). * The conversion was obtained by integration of the product and remaining starting material using analytical RP-HPLC-MS, after reduction of the imine using NaBH3CN (20 eq.) and subsequent TFA cleavage from the solid support.
To date, the literature is lacking examples of intramolecular imine macrocyclization on solid support. Only few examples are available either intramolecular imine formation in solution [1][2][3] or intermolecular imine formation on solid support. 3,4 According to the screening conditions depicted in the table above, trimethyl orthoformate afforded good compatibility with the resin and the subsequent cycloaddition conditions (entry 4). The full conversion to the desired imine was assessed by selective reduction of the Schiff base using sodium cyanoborohydride and subsequent cleavage from the resin followed by analytical RP-HPLC-MS. The peptide epitope sequence is bound to solid support using Rink Amide LL Resin (loading 0.28 to 0.40 mmol.g -1 ). The epitope is represented using single capital letter code amino acid. * The conversion was determined by integration of the product and SM using analytical RP-HPLC profile at 210 nm. ** The diastereomeric ratio was determined by integration of the HPLC profile at 210 nm. n.d. not determined, a complex mixture of diastereomer was observed using optimized analytical RP-HPLC-MS. † complex LC-MS profile was observed with various diastereomers and consequent amount of side products. ‡ complex mixture of diastereomers ( 4).
After screening catalyst systems (see table above), lithium bromide led to cleaner crude profile, demonstrated good reliability with the different dipolarophiles and gave  99% conversion to the desired cyclic ALFPGF and DINNN PepNats on solid support (entries 8-11 and 13). Lithium bromide proved to be essential for the cycloaddition on resin, as in its absence the cycloaddition could not reach completion (entries 12 and 14). Moreover, avoiding the use of heavy metal containing catalyst is of major advantages to develop a robust method to create PepNats collection readily available for biological evaluation. The peptide epitope sequence is bound to solid support using Rink Amide LL Resin (loading 0.28 to 0.40 mmol.g -1 ). The epitope is represented as single capital letter code amino acid with D-amino acid indicated by lower case letter. mixt.: mixture of diastereomers. * reverse diastereoselectivity observed with minor* = (2R, 3S, 4R, 5S) diastereomer, major* = (2S, 3R, 4S, 5R) diastereomer (for details see Experimental Section). ** The cycloaddition diastereomeric ratio (d.r.) is depicted as major:minor diastereomers unless more than two diastereomers were obtained. The d.r. was determined from the crude product by integration of the HPLC profile at 210 nm recorded using either the analytical RP-HPLC-MS (I) or the Optimized Analytical RP-HPLC-MS (II) (for details see Experimental Section) † Isolated mass and overall isolated yield of the major diastereomer (2R, 3S, 4R, 5S) PepNat from the starting unfunctionalized resin, unless otherwise noted (for details see Experimental Section). ‡ Purity was determined by integration of the product peak on the HPLC profile at 210 nm (for details see Experimental Section). § PepNat isolated as mixture of diastereomers (for the d.r. of the isolated product see Experimental Section). # The isolated diastereomer contains the (2S, 3R, 4S, 5R) stereocenters due to reverse selectivity of the cycloaddition due to steric hinderance. The DINNN epitope sequence is bound to solid support using Rink Amide LL Resin (loading 0.28 to 0.40 mmol.g -1 ). The epitope is represented as single capital letter code amino acid with D-amino acid indicated by lower case letter. * reverse diastereoselectivity observed with minor* = (2R, 3S, 4R, 5S) diastereomer (for details see Experimental Section). ** The cycloaddition diastereomeric ratio is depicted as major:minor diastereomers unless more than two diastereomers were obtained. The d.r. was determined from the crude product by integration of the HPLC profile at 210 nm recorded using either the analytical RP-HPLC-MS (I) or the Optimized Analytical RP-HPLC-MS (II) (for details see Experimental Section). † Isolated mass of the major diastereomer unless otherwise noted, after purification by RP-HPLC-MS (for details see Experimental Section). ‡ Overall isolated yield of the major diastereomer from the starting unfunctionalized resin. § The purity was determined by integration of the product peak on the 210 nm HPLC profile. # The PepNats product was isolated as mixture of diastereomers, the diastereomeric ratio of the product can be found in the Experimental Section. The disulfide bridge decapeptide 28 showed differences in binding on the human receptors compared to the mouse receptors reported in the literature. 5 The affinity differences between the two species is likely due to the low similarity (76%) between the human and the mouse melanocortin receptors. 6 In the functional assay, the cyclic decapeptide showed weak partial agonistic activity at hMC1R. Agonist effect could not be observed on hMC3R and hMC4R. Additionally, modification of Phe 112 and Phe 113 to their D-analogues reduced the selectivity profile of the disulfide bridge decapeptide. However, selective and almost full agonistic activity at hMC1R was observed (29, EC50 = 0.14 M  0.01 M; % effect = 84). This result confirms the hypothesis that the conformation adopted by the two phenylalanine hot spots can increase the agonistic activity at the MC1R.

2) A) Chemical Synthesis Materials and Methods
All reactants, reagents and commercially available compounds 6, 7, 9 and 10 were purchased MeCN + 0.1% TFA was used at a flow rate of 6 mL/min. The gradient system was adjusted according to the elution profile of the crude product. The relevant fractions containing the desired product were collected, analyzed by RP-HPLC-MS and accordingly pooled and lyophilized.

HRMS were obtained on a Thermo Scientific LTQ Fleet Orbitrap mass spectrometer coupled to an
Accela HPLC-System (HPLC column: Hypersyl GOLD, 50 mm x 1 mm, particle size 1.9 μm) equipped with an electron spray ionization source (ESI). The mass detection was recorded in the range of 150 to 2000 Da.

Surface Plasmon Resonance (SPR)
All SPR experiments were conducted on a Biacore T200, S200 or 8K instrument (GE Healthcare) at 20 C using HBS-P+ as running buffer (GE Healthcare). A CM5 sensor chip (GE Healthcare) was conditioned three times at 60 s using 50 mM NaOH, 1 M NaCl prior to standard NHS/EDC activation (e.g. amine coupling in the Biacore instruction manual) of the sensor surface for 420 s. (NM_005912) and MC5 (NM_005913) were cloned into the pcDNA4/TO vector (ThermoFisher). In the constructs, a consensus Kozak sequence (GCCACC) was incorporated immediately before the start ATG. DNA was amplified, isolated and sequenced using standard techniques.

Melanocortin receptor radioligand binding and cAMP assays.
Binding and cAMP assays were performed as described by Durek et al. 7  PepNats, literature compounds and references were tested in ten-point concentration response (½ log serial dilution) with 100, 10 or 1 µM as final start concentrations. Raw data output was analyzed in Screener software (Genedata AG) and pEC50 and pIC50 values calculated using a fourparameter logistic fit.

Automated Solid Phase Peptide Synthesis (RP_01)
Linear peptide precursors were synthesized with a SyroXP solid-phase peptide synthesizer using
Step 2: To a solution of 4-carboxybenzaldehyde (38.3 mg, 255.0 mol, 3 eq.) in DMF (3 mL) was added HATU (97.0 mg, 255.0 mol, 3 eq.) followed by DIPEA (118 L, 0.68 mmol, 8 eq.). The resulting mixture was stirred for 5 min before adding it to a suspension of Mtt deprotected peptide resin from step 1 in DMF (2 mL). The suspension was shaken at room temperature for 2.5 h. The solution was removed by vacuum filtration and the resin was washed with DCM (3  6 mL), DMF (3  6 mL) and DCM (3  6 mL). The desired aldehyde containing peptide resin 3 was quickly dried under vacuum filtration with a flow of argon.
Test cleavage for the aldehyde functionalized peptide 3: To assess the completion of the aldehyde coupling, a small portion of the functionalized resin (ca.

Fmoc deprotection (RP_03)
The Fmoc protecting group was removed by treating the aldehyde containing resin 3 (250.0 mg, 85.0 mol, loading 0.34 mmol/g) with a solution of 20% piperidine in DMF (6 mL) for 4 min. The solution was removed by vacuum filtration and the resin was washed with DMF (3  6 mL). The deprotection was repeated two additional times for 4 min each. The resin was finally washed with DMF (4  6 mL) and DCM (4  6 mL) and directly used into the next step without further modification.
The resulting suspension was shaken from 16 to 48 h at room temperature. The solution was removed by vacuum filtration and the resin was washed with DCM (3  6 mL), DMF (3  6 mL) and DCM (3  6 mL). The desired PepNat cycloadduct resin 5 then quickly dried under vacuum filtration with a flow of argon.
Test cleavage for the PepNat cycloadduct resin 5: The test cleavage was performed with a small portion of the resin from step 2 (ca. Purification by semi-preparative RP-HPLC-MS: The crude PepNat was dissolved in DMSO (ca. 1 mL) and purified by semi-preparative RP-HPLC-MS.
Selected pure fractions were combined and lyophilized to afford the desired PepNat (A−P). The reported yield corresponds to the overall isolated yield after purification by semi-preparative RP-HPLC from the starting unfunctionalized Rink Amide resin, through a total of six steps after the SPPS linear precursor synthesis on resin, unless otherwise noted.

Representative Procedure for the Synthesis of Lower Loaded Rink Amide Resin
Step 1: Commercially available Rink Amide Low Loading resin (602 mg, 0.22 mmol, loading 0.36 mmol/g)) was treated with 20% piperidine in DMF (6 mL) for 4 min. The treatment was repeated 2 additional times. The resin was finally washed with DMF (3  8 mL) and DCM (3  8 mL) then quickly dry under a flow of argon. Step

Resin Loading Quantification by UV Absorbance
A sample of the desired dry resin (ca. 10 mg) was treated with 20% piperidine in DMF (10 mL) for 30 min. The filtrate was collected by filtration. The mean absorbance of the dibenzofulvenepiperidine adduct contained in the filtrate was determined at 301 nm with a NanoDrop spectrophotometer (Thermo Fisher Scientific, Nanodrop 2000c, Quartz cuvette with optical path length = 1 cm) as an average of three independent measurements. The initial concentration was calculated using the Lambert-Beer's law equation with the molar extinction coefficient of 8021 l.mol -1 .cm -1 . The final loading of the resin was deduced from the calculated initial concentration and reported in Figure 3a.

3) B) Synthesis and Characterization of Linear Peptide Precursors 2 on Solid Support
The following linear peptides 2 were synthesized according to the representative procedure (RP_01). The successful formation of the desired linear peptide was confirmed by test cleavage as described above followed by analytical RP-HPLC-MS (I) analysis using method (I-A

3) C) Synthesis and Characterization of Aldehyde Containing Linear Peptides 3 on Solid Support
The aldehyde functionalized peptides 3 were prepared from the respective linear peptide precursors 2 using Mtt deprotection followed by amide coupling of the 4-carboxybenzaldehyde as described in RP_02. The completion was confirmed by test cleavage with a solution of TFA/H2O (95:5) as specify in the representative procedure followed by analytical RP-HPLC-MS (I) analysis using method (I-A). The
The yellow precipitate was filtered off, dried under reduce pressure. The resulting solid was purified by recrystallization from EtOH (11 mL) to afford the title product (1.10 g, 66% yield) as a yellow crystalline solid and E-isomer (E/Z ≥ 95:5).

H1-major
The titled fused di-pyrrolidine-GLGF PepNat was prepared from supported aldehyde peptide resin

4) C) Independent Racemic Synthesis in Solution Used for the LC-MS Comparison Depicted in
Supplementary Figure S2 and S3 The independent racemic synthesis of the endo cycloadduct products 19a-racemic and 19bracemic was performed using the procedure described in Section 4)A) for the cycloadducts 19a

Linear peptides 22, 23 and disulfide bridged peptide 24
The acetylated peptide 22 and 23 were synthesized according to the representative procedure

27
Step 1: The Fmoc protecting group was removed by treating the aldehyde containing resin 3d-1 (300.0 mg, 35.0 mol, calculated loading 0.12 mmol/g, starting unfunctionalized loading 0.26 mmol/g) with a solution of 20% piperidine in DMF (8 mL) for 4 min. The solution was removed by vacuum filtration and the resin was washed with DMF (3  8 mL). The deprotection was repeated two additional times for 4 min each. The resin was finally washed with DMF (4  8 mL) and DCM (4  8 mL) and directly used into the next step without further modification.
Step 2: A solution of 5% AcOH in DMF (5 mL) was added to the free amine peptide resin form step 1 in DMF (4 mL). The suspension was shaken for 1 h at room temperature then NaBH3CN (43.4 mg, 20 eq.) was added in one lump to the suspension. The suspension was shaken for 16 h at room

Synthesis and Characterization of Linear Peptide Precursors on Solid Support
The following linear peptides were synthesized according to the representative procedure (RP_01).
The successful formation of the desired linear peptide was confirmed by test cleavage as described in the representative procedure section followed by analytical RP-HPLC-MS (I) analysis using method (I-B
Step 2: The free amine resin from step 1 was cleaved from the Rink Amide Resin LL using a solution of TFA/TIS/H2O (4 mL, 95:2.5:2.5, v/v/v) by shaking for 1.5 h at room temperature. The filtrate was collected and cooled Et2O (40 mL) was added on ice. The slurry was kept at 0 °C for 30 min. The resulting white precipitate was then centrifuged (4000 rpm, 10 min, 4 °C). The supernatant was removed, and the residual white solid was dissolved in H2O (40 mL) and freeze-dried to afford the unbound and unprotected peptide (37.6 mg) as a white fluffy solid.
Step 3: The crude solid from step 2 was dissolved in DMSO (1.72 mL) and acetic acid (0.86 mL) then added dropwise to a solution of DMSO (0.86 mL) and water (14.7 mL). The resulting colorless reaction mixture was stirred for 48 h at room temperature. The reaction mixture was then diluted with additional water and freeze dried overnight to afford the crude disulfide bridge product. The crude product was purified by semi-preparative RP-HPLC using a step gradient of 10% to 70% of MeCN + 0.1% TFA in H2O + 0.1% TFA in 30 min at a flow rate of 6 mL/min to give the desired disulfide bridge product (12.3 mg, 25% yield) as a white fluffy solid with  99% purity. The analytical

6) A) Conformational Analysis by NMR of PepNats H1 and K1
All NMR spectra were recorded on a Bruker 700 MHz. NMR data was acquired in CD3OH and chemical shifts (δ values), given in parts per million (ppm), referenced to the CD3OH residual solvent signal (3.31 ppm). To identify which NH groups were buried or forming intramolecular hydrogen bonds, NMR spectra were also run in fully deuterated CD3OD (single endpoint, no rates were calculated). For the structural assignment of the peptides 1D 1 H, 2D COSY, 1 H-13 C HSQC/HMBC spectra were acquired using the standard pulse sequences available in TopSpin 3.5 and 4.0 (Bruker GmbH). Conformational analysis was carried out using the 2D 1 H, 1 H ROESY with 1048 and 512 data points in F2 and F1 respectively, a spectral width of 14 ppm, mixing time of 300 ms and a relaxation delay of 3 s. To determine the relative nOe intensities for pairs of spins, the extraction of F2-slices in the 2D ROESY at the F1-chemical shift of each resonance was carried out and the signals then integrated. 12 To improve accuracy, the PANIC method (Peak Amplitude Normalisation for improved cross-relaxation) was applied, 13,14 where the nOe intensities were normalized relative to the diagonal for each slice. Correction factors were then applied to compensate for the number of spins in each environment (corrected integral). For each molecule the integral for two protons at a known distance was used as reference to calibrate the other interproton distances in the molecule using the equation below: where is the intensity of the NOE between I and S (S being the inverted spin) and −6 is the internuclear distance between I and S.
The sampling of the conformational space for the peptides was carried out with Maestro Macrocycle Sampling algorithm using the OPLS3 force field (version 11.6.013, Schrödinger) with an energy threshold of 25 Kcal/mol to allow a full exploration of the rotation around the peptidic bonds. The NMR restraints were collected in the form of NOEs and JNH-H couplings. The resulting conformers were filtered to a reduced set of conformations that fulfilled key long-range NOEs in the macrocycle. To avoid being too restrictive and missing possible conformers complying with the NMR data, the filter for those distances was set to an upper limit of 5. The Fitter algorithm fits NMR data to ensembles of conformers rather than a priori assuming the presence of a single, low energy conformation in solution. The H-bond information was used for further refinement at the end of the analysis and it was introduced as distance restraints (equivalent to NOEs) between the NH and the H-bond acceptor in the peptide. For each NH potentially forming a H-bond according to the NMR data in CD3OD (ie, not exchanging with the solvent) all the possible interactions with nearby acceptors were considered in the calculation, and only those that complied with the rest of the NMR data were kept as distance restraints. complied with those distance restraints were subjected to 10 ns MD simulations, and the most populated cluster from each of them selected for further analysis with MSpin. The cluster below showed the best fit with the experimental NMR data (NOEs, J couplings and H-bonds). The 1611 conformers generated in the conformational sampling were filtered using three longrange constraints: (75,78) -(30,31), (150,151,152) -(119,120,121) and (119,120,121) -(50,51)

Solution conformation of compound H1-major derived using NMR and computational data
with an upper limit of 5.5 Å. 35 conformers that complied with those distance restraints were subjected to 10 ns MD simulations, and the most populated cluster from each of them selected for further analysis with MSpin. The cluster below showed the best fit with the experimental NMR data (NOEs, J couplings and H-bonds).

K1-major
PepNat K1-major showed two species in solution with a 80:20 ratio (measured by 1 H integration). nOe peaks (blue, different color of diagonal) and exchange peaks (red, same color of diagonal) are visible in the ROESY spectrum. The exchange peaks connect the NMR signals from the major and the minor species. Therefore, these two species are not chemically different, but the same chemical entity present in solution in two conformations in slow exchange.