Design and Synthesis of Novel Helix Mimetics Based on the Covalent H-Bond Replacement and Amide Surrogate

A novel hydrogen bond surrogate-based (HBS) α-helix mimetic was designed by the combination of covalent H-bond replacement and the use of an ether linkage to substitute an amide bond within a short peptide sequence. The new helix template could be placed in position other than the N-terminus of a short peptide, and the CD studies demonstrate that the template adopts stable conformations in aqueous buffer at exceptionally high temperatures.


Introduction
The α-helix is the most abundant secondary structural element in proteins and is found frequently at the interfaces of protein-protein interactions [1]. For example, transcriptional activator p53 contain a short α-helical sequence that mediates function by direct interaction with a receptor [2]. Stable isolated peptides with a defined short α-helical segment would be ideal inhibitors of macromolecular interactions. However, those with less than~15 amino acid residues rarely adopt a defined conformation in isolation [3][4][5][6], and often lack the ability to fold into their bioactive conformation due to an entropic penalty for folding. New techniques for stabilizing short peptide helices may aid the design of inhibitors or mimics of protein function [7][8][9][10][11][12]. The head-to-backbone cross-linking strategy for helix stabilization includes salt bridges, metal chelates, and covalent cyclization methods such as disulfide and lactam bridges [13,14], hydrocarbon stapling [15,16], and hydrogen-bond surrogate (HBS) methods [17][18][19]. P53 can act as a tumor suppressor and induce cancer cell death, and the levels of p53 can be increased by blocking the p53-MDM2 interaction and reactivate the p53 function [20]. Thus, it is a promising therapeutic strategy for the treatment of cancers by designing and developing small-molecule inhibitors of the MDM2-p53 interaction [21,22]. We discovered that the CUL4A-DDB1-ROC1-L2DTL-PCNA ubiquitin E3 ligase complex interacts with MDM2 and p53 and regulates p53 polyubiquitination and proteolysis through MDM2. We also found that both p53 and MDM2 bind to PCNA directly through a conserved PIP box in p53 and MDM2 [23]. Based on these findings and our experience obtained from medicinal chemistry [24] and total synthesis of marine natural products [25][26][27][28][29][30][31][32], we initiated a chemical biology program aiming to develop inhibitors to block p53 degradation based on the PCNA binding motif, a helix-containing PIP-box of PCNA interaction proteins.
Strategically placed covalent linkages of the type C = X − Y − N to replace the weak (i, i + 4) hydrogen bond, hydrogen-bond surrogates (HBS), have been shown to stabilize the helical conformations in short peptide sequences [17][18][19]. Modifications of the peptide backbone by the replacement of the hydrogen bond with hydrazine [17], carbon-carbon links [33], and thioether linkage [34] have been reported. This strategy provides peptides with increased target affinity and allows a stabilization of the α-helical conformation [9]. However, this approach seems limited to the N-terminal position of a short peptide, and could not be extended to our targeted system. In order to introduce the covalent H-bond replacement at an internal helical turn, we strategically employed an ether linkage to substitute an amide bond as well as replaced the corresponding (i, i + 4) hydrogen bond with a covalent ethylane bridge to afford a novel hydrogen bond surrogate-based (HBS) α-helix 1 (see Scheme 1). Notably, the new helix template (1) could be placed in position other than N-terminus of a short peptide. In connection with the previously mentioned chemical biology program, we selected the helix mimetic template 1 as a model system to explore the ability of the newly designed cyclopeptide to promote an α-helical conformation in aqueous solution.
Molecules 2023, 28, x FOR PEER REVIEW 2 of 14 helical conformations in short peptide sequences [17][18][19]. Modifications of the peptide backbone by the replacement of the hydrogen bond with hydrazine [17], carbon-carbon links [33], and thioether linkage [34] have been reported. This strategy provides peptides with increased target affinity and allows a stabilization of the α-helical conformation [9]. However, this approach seems limited to the N-terminal position of a short peptide, and could not be extended to our targeted system. In order to introduce the covalent H-bond replacement at an internal helical turn, we strategically employed an ether linkage to substitute an amide bond as well as replaced the corresponding (i, i + 4) hydrogen bond with a covalent ethylane bridge to afford a novel hydrogen bond surrogate-based (HBS) α-helix 1 (see Scheme 1). Notably, the new helix template (1) could be placed in position other than N-terminus of a short peptide. In connection with the previously mentioned chemical biology program, we selected the helix mimetic template 1 as a model system to explore the ability of the newly designed cyclopeptide to promote an α-helical conformation in aqueous solution.  The retrosynthetic analysis of the designed helix template (1) was shown in Scheme 1. The target molecule 1 could be obtained from a macrocyclization of the corresponding linear precursor 2, which should be readily available from a coupling reaction of the advanced intermediate 3 with a suitable dipeptide. Further disconnection of 3 gave rise to the protected tyrosine derivative 4 and alcohol 5, which is readily available from β-hydroxy azide 6.

Synthesis of Helix Mimetic Template
The synthesis commenced with the substrate controlled allylation of the known αhomochiral aldehyde 7 [35] (Scheme 2). Thus, the treatment of aldehyde 7 with allyltributyltin in the presence of SnCl4 at −78 °C afforded the syn- (8) and anti-homoallylic alcohols in a diastereomeric ratio of 9:1 [36,37]. Homoallylic alcohol 8 was converted into acetonide 9 in 76% overall yield via a three-step sequence including (i) oxazolidine formation with 2,2-dimethoxypropane/PTSA, (ii) the dihydroxylation of the terminal alkene with osmium tetroxide and 4-methylmorpholine N-oxide as co-oxidant, followed by sodium periodate Scheme 1. Structure of α-helix peptide and retrosynthetic analysis of helix template 1.
The retrosynthetic analysis of the designed helix template (1) was shown in Scheme 1. The target molecule 1 could be obtained from a macrocyclization of the corresponding linear precursor 2, which should be readily available from a coupling reaction of the advanced intermediate 3 with a suitable dipeptide. Further disconnection of 3 gave rise to the protected tyrosine derivative 4 and alcohol 5, which is readily available from β-hydroxy azide 6.

Synthesis of Helix Mimetic Template
The synthesis commenced with the substrate controlled allylation of the known αhomochiral aldehyde 7 [35] (Scheme 2). Thus, the treatment of aldehyde 7 with allyltributyltin in the presence of SnCl 4 at −78 • C afforded the syn-(8) and anti-homoallylic alcohols in a diastereomeric ratio of 9:1 [36,37]. Homoallylic alcohol 8 was converted into acetonide 9 in 76% overall yield via a three-step sequence including (i) oxazolidine formation with 2,2-dimethoxypropane/PTSA, (ii) the dihydroxylation of the terminal alkene with osmium tetroxide and 4-methylmorpholine N-oxide as co-oxidant, followed by sodium periodate cleavage of the diol, resulted in the corresponding aldehyde, and (iii) reduction in the resulting aldehyde with NaBH 4 in methanol. The simultaneous removal of N-Boc and isopropylidene protective groups of 9 was realized under acidic conditions (6 M HCl in THF) to afford the corresponding amino diol, which was then converted into azide diol 10 in 74% yield via a diazo-transfer reaction with triflyl azide. The regioselective protection of the primary alcohol as its TBS ether afforded the key intermediate 6 in 87% yield. With alcohol 5 at hand, attention was turned to the installation of the tyrosine units via the Fukuyama-Mitsunobu protocol and later-stage macrocyclization (Scheme 3). Thus, the treatment of sulfonamide-protected amine 4 [39][40][41] and alcohol 5 in the presence of diethylazodicarboxylate and triphenylphosphine afforded the corresponding alkylation adduct 16 in 78% yield. The nosyl group (Ns-) of 16 was then cleaved with Fukuyama's thiophenol/K2CO3/CH3CN conditions [42], giving rise to the corresponding free secondary amine 3 in 85% yield. It is known that the coupling of a secondary amine and carboxylic acid using standard peptide coupling techniques is often a low-yielding process with certain difficulties. Gratifyingly, the HATU/HOAt-promoted coupling reaction between secondary amine 3 and dipeptide acid 17a provided 2a in 85% yield. The simultaneous removal of the benzyl ester, O-Bn ether and Cbz-protecting group was achieved by the hydrogenolysis of 2a with Pd(OH)2 on carbon to produce the desired amino acid which was immediately activated by HATU/HOAT in the presence of N-methylmorpho- We then addressed the formation of ether linkage of 12 so as to generate the required precursor 5 for further coupling reaction. Considering the nature of chiral triflate, also supported by literature precedents [38], we anticipated that the displacement of the triflate 11 by alkoxy anion derived from 6 could be achieved. Unfortunately, the attempted synthesis of the requisite ether 12 via the displacement reaction proved to be frustrating. All attempts to convert 6 to 12 by the reaction of the anion derived from the former with triflate 11(a-c) [38], under a variety of reaction conditions provided no product or resulted in a much lower conversion (Scheme 2). In order to avoid more side-reaction derived from the slow intermolecular displacement process, we elected to construct the ether linkage via an intramolecular S N 2 reaction, which would also avoid any ambiguity of the stereochemistry.
Thus, the hydrogenation of the azide alcohol 6 over Pd/C gave the corresponding amine, which was condensed with (R)-2-bromopentanoic acid (13) employing EDCI to provide amide 14 in 63% yield over two steps. To our delight, upon the treatment of 14 with sodium hydride in THF, the intramolecular S N 2 reaction proceeded smoothly to furnish the desired morphorlinone 15 as a single diastereomer in 57% yield. The inversion of the C-2 stereogenic center of 13 leading to the structure of 15 with the all-S stereochemical configurations was ascertained by NMR correlation and NOESY experiments. Morphorlinone 15 was then converted into the key intermediate 5 in 66% overall yield by a three-step sequence of straightforward transformations: (i) acid cleavage of TBS ether and the hydrolysis of amide to the corresponding amino acid; (ii) protection of the resultant amino group as its Boc carbamate; and (iii) protection of the carboxyl functionality as its benzyl ester.
With alcohol 5 at hand, attention was turned to the installation of the tyrosine units via the Fukuyama-Mitsunobu protocol and later-stage macrocyclization (Scheme 3). Thus, the treatment of sulfonamide-protected amine 4 [39][40][41] and alcohol 5 in the presence of diethylazodicarboxylate and triphenylphosphine afforded the corresponding alkylation adduct 16 in 78% yield. The nosyl group (Ns-) of 16 was then cleaved with Fukuyama's thiophenol/K 2 CO 3 /CH 3 CN conditions [42], giving rise to the corresponding free secondary amine 3 in 85% yield. It is known that the coupling of a secondary amine and carboxylic acid using standard peptide coupling techniques is often a low-yielding process with certain difficulties. Gratifyingly, the HATU/HOAt-promoted coupling reaction between secondary amine 3 and dipeptide acid 17a provided 2a in 85% yield. The simultaneous removal of the benzyl ester, O-Bn ether and Cbz-protecting group was achieved by the hydrogenolysis of 2a with Pd(OH) 2 on carbon to produce the desired amino acid which was immediately activated by HATU/HOAT in the presence of N-methylmorpholine to afford cyclodepsipeptide 1a in 62% yield. The steric hinderance of amine 3 significantly deterioate the reaction rate and yield of this peptide coupling step. Among all the tested coupling reagents such as PyAOP, DEPBT. HATU, BOPCl etc., HATU/HOAT condition was found with best results. The removal of the Boc-protecting group of 1a with trifluoroacetic acid in CH 2 Cl 2 produced the more hydrophilic template 18 in 78% yield (See Supplementary Materials).

CD Spectra of Helix Templates
The solution conformation of the helix-template 1a-d was investigated by circular dichroism spectroscopy ( Figure 1). All samples were measured at 10 mM concentration in pH 7.4 phosphate buffer. The CD spectra obtained for 1a and 1c were typical for the αhelical structure, showing the characteristic two molar ellipticity minima (λ = 222, 208 nm) and an ellipticity maximum (λ = 195 nm) [43,44]. However, the CD spectra of 1b and 1d showed almost no α-helical secondary structure ( Figure 1, left panel). The loss of helicity in 1b and 1d relative to 1a and 1c indicated the stereogenic center and the substituent R1 of 1 (Scheme 1) are vital for the overall stability of helical conformation. Similar to 1a, the CD spectra of 18 also showed two negative peaks at 208 and 222 nm wave trough and a positive value at 195 nm, which almost did not change spectral intensities by adding 2,2,2trifluoroethanol ( Figure 1, right panel). Previous thermal stability studies of HBS helices have shown that the conformations of these peptides remain remarkably consistent at high temperatures [19,33,45]. The thermal stabilities of 18 were investigated by monitoring the temperature-dependent change in the intensity of the 220 nm bands in the CD spectra ( Figure 2). To our delight, we observed that the spectral lineshapes or intensities almost did not change along with a gradual increase in the temperature, even at 95 • C. Overall, the CD studies demonstrated that the template 18 adopts stable conformations. trifluoroacetic acid in CH2Cl2 produced the more hydrophilic template 18 in 78% yield (See Supplementary Materials).

CD Spectra of Helix Templates
The solution conformation of the helix-template 1a-d was investigated by circular dichroism spectroscopy ( Figure 1). All samples were measured at 10 mM concentration in pH 7.4 phosphate buffer. The CD spectra obtained for 1a and 1c were typical for the αhelical structure, showing the characteristic two molar ellipticity minima (λ = 222, 208 nm) and an ellipticity maximum (λ = 195 nm) [43,44]. However, the CD spectra of 1b and 1d showed almost no α-helical secondary structure (Figure 1, left panel). The loss of helicity in 1b and 1d relative to 1a and 1c indicated the stereogenic center and the substituent R1 of 1 (Scheme 1) are vital for the overall stability of helical conformation. Similar to 1a, the CD spectra of 18 also showed two negative peaks at 208 and 222 nm wave trough and a Previous thermal stability studies of HBS helices have shown that the conformations of these peptides remain remarkably consistent at high temperatures [19,33,45]. The thermal stabilities of 18 were investigated by monitoring the temperature-dependent change in the intensity of the 220 nm bands in the CD spectra ( Figure 2). To our delight, we observed that the spectral lineshapes or intensities almost did not change along with a gradual increase in the temperature, even at 95 °C.
Overall, the CD studies demonstrated that the template 18 adopts stable conformations.

General Experimental Details
All non-aqueous reactions were performed under an atmosphere of nitrogen or argon using oven-dried (120 °C) or flame-dried glassware under a N2 atmosphere. Commercially available reagents were used without further purification. All solvents were distilled prior to use: tetrahydrofuran (THF) from Na/benzophenone, dichloromethane (DCM), triethylamine (TEA) and dimethylformamide (DMF) were distilled from CaH2. Methanol (MeOH) was distilled under a N2 atmosphere from Mg/I2. 1 H NMR and 13 C NMR spectra were recorded in CDCl3 or MeOH-d4 on a Bruker Avance AV500 or Bruker Avance AV400 at 500 MHz (125 MHz) or 400 MHz (100 MHz), respectively. Chemical shifts are reported as δ values (ppm) referenced to either a tetramethylsilane (TMS) internal standard or the signals due to the solvent residual. Data for 1 H NMR are reported as follows: chemical shift (δ ppm), multiplicity (s = singlet, br = broad, d = doublet, t = triplet,

General Experimental Details
All non-aqueous reactions were performed under an atmosphere of nitrogen or argon using oven-dried (120 • C) or flame-dried glassware under a N 2 atmosphere. Commercially available reagents were used without further purification. All solvents were distilled prior to use: tetrahydrofuran (THF) from Na/benzophenone, dichloromethane (DCM), triethylamine (TEA) and dimethylformamide (DMF) were distilled from CaH 2 . Methanol (MeOH) was distilled under a N 2 atmosphere from Mg/I 2 . 1 H NMR and 13 C NMR spectra were recorded in CDCl 3 or MeOH-d 4 on a Bruker Avance AV500 or Bruker Avance AV400 at 500 MHz (125 MHz) or 400 MHz (100 MHz), respectively. Chemical shifts are reported as δ values (ppm) referenced to either a tetramethylsilane (TMS) internal standard or the signals due to the solvent residual. Data for 1 H NMR are reported as follows: chemical shift (δ ppm), multiplicity (s = singlet, br = broad, d = doublet, t = triplet, q = quartet, m = multiplet), coupling constant (Hz), integration. Mass spectra were measured on ABI Q-star Elite. Optical rotations were measured on a Perkin-Elmer 351 polarimeter at 589 nm with a 100 mm path length cell at 20 • C (reported as follows: concentration (c (in 1 g/100 mL), solvent). The reaction progress was checked on TLC plates. TLC was carried out using pre-coated sheets (Qingdao silica gel 60-F250, 0.2 mm) which, after development, were visualized under UV light at 254 nm, and/or staining in p-anisole, ninhydrin or phosphomolybdic acid solution followed by heating. Flash column chromatography was performed using the indicated solvents on Qingdao silica gel 60 (230-400 mesh ASTM). Yields refer to chromatographically purified compounds, unless otherwise stated.
CD spectroscopy studies: CD spectra were recorded on an Applied Photophysics Chirascan CD spectrometer equipped with a temperature controller using 0.1 cm path length cells. Experiments were performed at 20 • C using a 0.1 cm cell, at a scan speed of 100 nm/min from 250 to 186 nm. Spectral baselines were obtained under analogous conditions as that for the samples. All spectra were baseline subtracted, converted to a uniform scale of molar ellipticity, and smoothed using OriginPro 8.0 software. The samples were dissolved in 10 mM phosphate buffer, measured as pH 7.4. The raw CD data of 1a-1d and 18 were recorded in raw ellipticity units. The concentrations of 1a-1d and 18 were determined by quantitative RP-HPLC against a standard of known concentrations. All samples were between 20 and 70 µM, and the CD spectra were recorded in a range of concentrations to confirm that the sample aggregation did not occur. CD data were converted to mean the residue ellipticity, [θ] (deg·cm 2 ·dmol −1 ), using the equation where c is the sample concentration (M) and l is the cell path length (cm), n is number of amino acid residues in the peptides (n = 4).

Two-Dimensional COSY and NOESY Analysis of Compound 15
The COSY and NOESY experiments were recorded in acetone-d 6 /methanol-d 4 (2:1). The co-solvent system was used here to distinguish the signals of H a from H 2" , H c from H 1" (which were found severely overlapped in either CDCl 3 or Acetone-d 6 ). From the 1 H NMR and COSY spectra, H a was assigned to 3.80 ppm (1H, dt, J 1 = 10 Hz, J 2 = 5.2 Hz), H c were assigned to 1.69-1.71 ppm (2H, m) and H b was assigned to 4.03 ppm (1H, dd, J 1 = 8.0 Hz, J 2 = 4.4 Hz). The NOESY spectrum showed clear correlations of H a -H c and H c -H b , no direct correlation between H a and H b , and these correlation signals unambiguously proved that H b is at the equatorial position (six-membered ring's chair-confirmation), and thus the stereochemistry at C-2 was determined as 2S.

Two-Dimensional COSY and NOESY Analysis of Compound 15
The COSY and NOESY experiments were recorded in acetone-d6/methanol-d4 (2:1). The co-solvent system was used here to distinguish the signals of Ha from H2", Hc from H1" (which were found severely overlapped in either CDCl3 or Acetone-d6). From the 1 H NMR and COSY spectra, Ha was assigned to 3.80 ppm (1H, dt, J1 = 10 Hz, J2 = 5.2 Hz), Hc were assigned to 1.69-1.71 ppm (2H, m) and Hb was assigned to 4.03 ppm (1H, dd, J1 = 8.0 Hz, J2 = 4.4 Hz). The NOESY spectrum showed clear correlations of Ha-Hc and Hc-Hb, no direct correlation between Ha and Hb, and these correlation signals unambiguously proved that Hb is at the equatorial position (six-membered ring's chair-confirmation), and thus the stereochemistry at C-2 was determined as 2S.

Conclusions
The development of new methods for the helical stabilization of a linear peptide could provide additional opportunities for the discovery of peptide-based therapeutics targeting PPI. Design and synthesis of hydrogen-bond surrogates (HBSs) are one of the most successful approaches. In this paper, we developed a new helix template based on

Conclusions
The development of new methods for the helical stabilization of a linear peptide could provide additional opportunities for the discovery of peptide-based therapeutics targeting PPI. Design and synthesis of hydrogen-bond surrogates (HBSs) are one of the most successful approaches. In this paper, we developed a new helix template based on the replacement of the weak (i, i + 4) H-bond with a covalent carbon-carbon linkage; and also employed a covalent ether bond to mimic one amide-bond. In comparing the precedent hydrogen-bond surrogates in which the hydrogen bond was replaced with hydra-zine, carbon-carbon links, and thioether linkage, our new helix template could be placed in a position other than the N-terminus of a short peptide. The CD studies show that the helix template adopts stable conformations in aqueous buffer at exceptionally high temperatures. We also found that the stereogenic centers presented in the macrocycle are vital for the stability of helical conformation. The scaffold obtained has been proven as a single-turn helical mimetic. The further coupling of additional short peptide fragments and beta-strand mimetic aiming to develop inhibitors of p53 degradation based on the PIP motif of PCNA interaction proteins will be reported in due course.