Sulfono-γ-AApeptides as Helical Mimetics: Crystal Structures and Applications

CONSPECTUS:

Foldamers have defined and predictable structures, improved resistance to proteolytic degradation, enhanced chemical diversity, and are versatile in their mimicry of biological molecules, making them promising candidates in biomedical and material applications. However, as natural macromolecules exhibit endless folding structures and functions, the exploration of the applications of foldamers remains crucial. As such, it is imperative to continue to discover unnatural foldameric architectures with new frameworks and molecular scaffolds. To this end, we recently developed a new class of peptidomimetics termed γ-AApeptides”, oligomers of γ-substituted-N-acylated-N-aminoethyl amino acids, which are inspired by the chiral peptide nucleic acid backbone. To date γ-AApeptides have been shown to be resistant to proteolytic degradation and possess limitless potential to introduce chemically diverse functional groups, demonstrating promise in biomedical and material sciences. However, the structures of γ-AApeptides were initially unknown, rendering their rational design for the mimicry of a protein helical domain impossible in the beginning, which limited their potential development. To our delight, in the past few years, we have obtained a series of crystal structures of helical sulfono-γ-AApeptides, a subclass of γ-AApeptides. The single-crystal X-ray crystallography indicates that sulfono-γ-AApeptides fold into unprecedented and well-defined helices with unique helical parameters. On the basis of the well-established size, shape, and folding conformation, the design of sulfono-γ-AApeptide-based foldamers opens a new avenue for the development of alternative unnatural peptidomimetics for their potential applications in chemistry, biology, medicine, materials science, and so on.

In this Account, we will outline our journey on sulfono-γ-AApeptides and their application as helical mimetics. We will first briefly introduce the design and synthetic strategy of sulfono-γ-AApeptides and then describe the crystal structures of helical sulfono-γ-AApeptides, including left-handed homogeneous sulfono-γ-AApeptides, right-handed 1:1 α/sulfono-γ-AA peptide hybrids, and right-handed 2:1 α/sulfono-γ-AA peptide hybrids. After that, we will illustrate the potential of helical sulfono-γ-AApeptides for biological applications such as the disruption of medicinally relevant protein-protein interactions (PPIs) of BCL9-β-catenin and p53-MDM2/MDMX as well as the mimicry of glucagon-like peptide 1 (GLP-1). In addition, we also exemplify their potential application in material science. We expect that this Account will shed light on the structure-based design and function of helical sulfono-γ-AApeptides, which can provide a new and alternative way to explore and generate novel foldamers with distinctive structural and functional properties.

Graphical Abstract

1. INTRODUCTION

Foldamers^5–7 have recently received considerable research interest, owing to their remarkable performance in various promising applications, e.g., molecular recognition, catalysis, supramolecular design, and rational design of biomimetic structures. Since 2011, we reported a class of peptidomimetics termed “γ-AApeptides” inspired by the backbone of the chiral peptide nucleic acid (PNA)^8,9 backbone (Figure 1). They are named as such because they are oligomers of γ-substituted-N-acylated-N-aminoethyl amino acids. We demonstrated that γ-AApeptides are resistant to proteolytic degradation and possess limitless potential to introduce chemically diverse functional groups and show promise in biomedical applications.^10,11 However, the crystal structures of γ-AApeptides were initially unknown, rendering it impossible for the rational design to mimic the helical domain of proteins, thereby modulating biologically relevant PPIs.

Figure 1.

Chemical structure of α-peptide, chiral PNA, l-γ-AApeptide, l-sulfono-γ-AApeptide, d-sulfono-γ-AApeptide, 1:1 α/l-sulfono-γ-AApeptide, and 2:1 α/d-sulfono-γ-AApeptide.

To our delight, in the past few years, we successfully determined the X-ray crystal structures of a series of helical sulfono-γ-AApeptides and their derivatives, including left-handed homogeneous sulfono-γ-AApeptides (Figure 1),¹ right-handed 1:1 α/sulfono-γ-AA peptide hybrids (Figure 1),^2,12,13 and right-handed 2:1 α/d-sulfono-γ-AA peptide hybrids (Figure 1).¹⁴ All of the oligomers fold into unprecedented well-defined and robust helices with unique helical parameters, stabilized by both intramolecular hydrogen bonding and the curved nature of sulfonamido groups. On the basis of the helical structure of sulfono-γ-AApeptides, we exemplified the promising potential of sulfono-γ-AApeptides for protein surface recognition such as disruption of medicinally relevant protein-protein interactions (PPIs) of BCL9-β-catenin³ and p53-MDM2/MDMX¹⁵ and mimicry of glucagon-like peptide 1 (GLP-1) for protein recognition,⁴ as well as applications in material science.^13,16

2. DESIGN AND SYNTHESIS OF SULFONO-γ-AAPEPTIDES

Our interest in sulfono-γ-AApeptides, a subclass of γ-AApeptides, was stimulated by investigations of the role of side chains in canonical γ-AApeptides. As shown in Figure 1, the replacement of carboxyl side chains with sulfonyl side chains leads to the generation of sulfono-γ-AA-peptides.¹ We envisioned that the tertiary sulfonamido moieties may be sufficiently hindered to induce intrinsic curvature into the sulfono-γ-AApeptide backbone.¹ They are also devoid of cis- or trans-rotamers observed in tertiary-amide-containing canonical γ-AApeptides, thereby making the whole molecular skeleton more rigid and structurally defined. In addition, half of the side chains in sulfono-γ-AApeptide are chiral, and protons are still present in the second amide moieties for potential intramolecular hydrogen bonding. These intrinsic features prompted us to speculate that sulfono-γ-AApeptides may exhibit folding propensities and form well-defined secondary structures.¹ The homogeneous sulfono-γ-AApeptides could be designed and synthesized by assembling the desired sulfono-γ-AApeptide building blocks (Scheme 1)¹⁷ on the solid phase (Scheme 2A).¹⁸ In the meantime, we also set out to develop sulfono-γ-AApeptides with heterogeneous backbones (e.g., containing both sulfono-γ-AA residues and canonical amino acid residues), as they could further increase the availability of molecular frameworks, three-dimensional structures, and functions.⁶ Additionally, compared with homogeneous sulfono-γ-AApeptides, the presence of α-amino acid residues in the heterogeneous backbone could contribute to the folding propensity with more amide hydrogens for potential intramolecular hydrogen bonding. As such, efforts were also directed to develop heterogeneous backbones by combining the alternative l-α-amino acid and l-sulfono-γ-AA amino acid residues in a 1:1 repeat pattern^2,12,13 or alternative l-α-amino acid/d-sulfono-γ-AA residues in the 2:1 repeat pattern (Figure 1, Scheme 2B and 2C).¹⁴

Scheme 1.

General Synthetic Route for the Preparation of Sulfono-γ-AApeptide Building Blocks

Scheme 2.

General Synthetic Route for the Preparation of Sulfono-γ-AApeptides

3. THE CRYSTAL STRUCTURES OF γ-AAPEPTIDES

3.1. Homogeneous Sulfono-γ-AApeptides

A panel of homogeneous sulfono-γ-AA peptidic oligomers were first prepared for structural and spectroscopic evaluation (Figure 2).¹ The single crystals of 1a, 3a, 4b, or 6a were obtained, and their structures were successfully determined by X-ray crystallography (Figure 3). It is intriguing that these oligomers form identical left-handed helices (Figure 3A) in a 4₁₄-helix pattern (Figure 3B). These helical sulfono-γ-AApeptides, with a radius of 2.8 Å and a pitch of 5.1 Å, have exactly four side chains per turn that are aligned perfectly on the top of one another, leading to an unprecedented ordered C₂-symmetric helix. Notably, 6a with only methyl side chains formed the same left-handed helix with the same structural parameters as those of the other sequences (Figure 3H), indicating that sulfono-γ-AApeptides can form the identical left-handed helices regardless of their side-chain identity. It is intriguing that sulfono-γ-AApeptides demonstrated a higher helical stability than α-helices in solution, which is in line with our supposition that the helicity of sulfono-γ-AApeptides is stabilized by both intramolecular hydrogen bonding and curvature of sulfonamido moieties on the molecular backbone, leading to enhanced folding propensity.

Figure 2.

Homogeneous sulfono-γ-AA peptidic oligomers prepared for structural and spectroscopic evaluation in the study. Reproduced with permission from ref 1. Copyright 2020 John Wiley& Sons, Inc.

Figure 3.

(A) Side and top views of the crystal structure of 1a. Hydrogen bonding is shown in red. (B) The intramolecular 14-hydrogen-bonding pattern of 1a detected in the crystal structure. (C) Crystal packing of 1a viewed perpendicular to and down the helix axis. (D-G) Comparison of the crystal structures of 1a, 3a, 4b, and 6a. (H) Sequence structure of oligomer 6a. Reproduced with permission from ref 1. Copyright 2020 John Wiley& Sons, Inc.

3.2. 1:1 l-α/l-Sulfono-γ-AA Heterogeneous Peptides

Next we attempted to investigate the structures of a series of 1:1 α/sulfono-γ-AA heterogeneous peptides.^2,12 We first solved the single crystal of the 7:8 type monomer 7 with an alternative methyl-sulfono-γ-AApeptide incorporating a chlorobenzene sulfonyl group and α-alanine in a 1:1 repeat pattern (Figure 4).²

Figure 4.

(A) Sequence structure of monomer 7 and the 13-atom-hydrogen-bonding pattern. (B) Sequence structure of dimer 8. (C) Crystal structure of monomer 7 stabilized by intramolecular hydrogen bond (magentas dashed line in inset). (D) Crystal packing model of 7. The disordered acetonitriles are excluded from the crystal lattice of 7. (E) Cartoon representation of X-ray crystal structures from mono helix 7 to the covalent-bonded zippered dimer 8. Dashed red lines highlight the intramolecular hydrogen bond in dimer 8. Reproduced with permission from ref 2. Copyright 2020 American Chemical Society.

The crystal of 7 reveals a defined hydrogen-bonded right-handed 4₁₃-helix with a virtually identical helical pitch of 5.34 Å and a radius of 3.05 Å (Figure 4C). There are also exactly four sides chains per helical turn like the homogeneous sulfono-γ-AApeptides, and all side chains are almost perpendicular to the helical axis and pointing away from the peptide axis, which lead to a pseudo-4-fold symmetry of a windmill shape on the top view.

The single crystal of 8, via dimerization of monomer 7 at the third sulfono side chain using terephthaloyl as the linker, was subsequently obtained.² An analysis of the results of the single-crystal X-ray diffraction reveals that a right-handed helical scaffold formed with the same helical pitch and radius as that of monomer 7 (Figure 4E). More intriguingly, the dimer formed a stable zipper-like tertiary structure, with an angle of 80° between two helical strands.

In order to substantiate the universality of 7’s structural form, several new 1:1 l-α/l-sulfono-γ-AA heterogeneous peptides with different lengths and side chains were investigated (Figure 5).¹² Luckily, we were able to solve the single crystal of the 8:9 type foldamer 9 and 9’s analogues 10, 11, and 12 (Figure 5). They all displayed the same backbone folding propensity. Our findings also demonstrate that halogen bonding exists in the 1:1 l-α/ l-sulfono-γ-AA heterogeneous peptides containing halogen atoms in the crystal lattices. Such a combination of halogen bonding and intermolecular hydrophobic interactions may enable the 3D supramolecular assembly.

Figure 5.

(A) Sequence structures of foldamers 9, 10, 11, and 12. (B) Crystal structure of foldamer 9, as a helical representation. (C) Infinite 1D chain formed by the head-to-tail assembly of foldamer 9 in terms of both intramolecular and intermolecular hydrogen bonding. (D) Crystal packing of 9, including both antiparallel and perpendicular helices. (E) Crystal structure of foldamer 10. 3D supramolecular network of foldamer 10. (F) comparison of intermolecular halogen interactions between foldamer 9 (a), foldamer 10 (b), foldamer 11 (c), and foldamer 12 without halogenated side chains (d). Nonpolar hydrogens are omitted for clarity. Reproduced with permission from ref 12. Copyright 2020 John Wiley& Sons, Inc.

3.3. 2:1 l-α/d-Sulfono-γ-AA Heterogeneous Peptides

To explore the impact of d-sulfono-γ-AApeptides on the formation of helical secondary structures, we recently reported the X-ray crystal structures of heterogeneous peptides (13, 14, 15, and 16) consisting of a 2:1 pattern of an alternative acid/l-amino d-sulfono-γ-AA residue (Figure 6).¹⁴

Figure 6.

(A) 2:1 α/d-sulfono-γ-AA peptidic oligomers prepared for structural and spectroscopic evaluation in this study. (B) Side views of single-crystals 13, 14, 15, and 16. Hydrogen bonding is shown in cyan. (C) Top views of single-crystals 13, 14, 15, and 16 along the helix axis. (D) Structure of crystal 14 packing along the peptide axis; the intermolecular hydrogen-bonding pattern is shown in the inset for clarity. (E) Cartoon representation of 14 shown in an oval to further clarify the helix. (F) Crystal packing of oligomer 13 viewed perpendicular and then down to the helical axis. (G) 16–16-14-hydrogen-bonding pattern detected in the crystal structure of 14. Reproduced with permission from ref 14. Copyright 2020 American Chemical Society.

Interestingly, these crystals reveal a similar right-handed 4.5_16–14 helix with a virtually unanimous helical radius of 2.6 Å and a helical pitch of 5.1 Å, with 4.5 residues per turn. This is beyond our expectation because helical d-peptides are known to adopt a left-handed conformation.

4. THE APPLICATION OF HELICAL SULFONO-γ-AAPEPTIDES

4.1. Modulation of Protein–Protein Interactions

Having obtained the crystal structures of γ-AApeptides, we set out to study their capability to mimic the helical domain of proteins and modulate the protein-protein interactions (PPIs).^3,4,15

4.1.1. Disruption of p53-MDM2 PPIs.

The p53-MDM2 PPI is the classic PPI that has been recognized as the testing ground for newly designed foldamers.¹⁹ As shown in Figure 7, there are three critical residues, Phe 19, Trp 23, and Leu 26, in the helical domain of p53. They are on the same face of p53 helix, by which p53 binds to the cleft of MDM2 and deeply insert into the MDM2 hydrophobic pocket. Therefore, the molecules that could reproduce the functionalities of those three hot spots are expected to bind to MDM2 and thus disrupt the p53-MDM2 interactions.^19–23

Figure 7.

(A) The interaction of p53 with the crystal structure of MDM2 (PDB: 1YCR). p53 is shown as the cartoon whereas MDM2 is shown as the surface representation. (B) The chemical structure of sulfono-γ-AApeptides. a and b denote the chiral side chain and the sulfonamido side chain from the building block, respectively. (C) The crystal structure of a sulfono-γ-AApeptide. (D) Top view of panel C. (E and F) The schematic representation of the distribution of side chains from sulfono-γ-AApeptides. (E) Side view and (F) top view of the helical wheel. Reproduced with permission from ref 15. Copyright 2020 American Chemical Society.

4.1.1.1. Left-Handed Homogeneous Sulfono-γ-AApeptides.

As the helical pitch of sulfono-γ-AApeptides (5.1 Å) is virtually similar to that of the α-peptide (5.4 Å), we speculated that sulfono-γ-AApeptides could mimic one helical face of the α-helix in the proteins and thereby modulate α-helix-mediated PPIs (Figure 7). A series of left-handed homogeneous sulfono-γ-AApeptides (Table 1) were then designed to inhibit p53-MDM2 PPIs. The first sequence γ-AApeptide 17 was obtained by fixing the key side chains of Phe 19, Trp 23, and Leu 26 at positions 2a, 4a, and 6a, respectively, with the activity (K_d = 98 nM) being twice as strong than that of p53 (K_d = 208 nM). One missing key group mimics Trp 23 at the 4a position, leading to the loss of the binding affinity to MDM2 (γ-AApeptide 18). The γ-AApeptide 19 with a cyclobutylmethyl group at the 6a position was identified as the most potent compound (K_d and IC₅₀ values were 26 nM and 0.891 μM, respectively). More structural information of the γ-AApeptide 19 interaction with MDM2 was collected from the nuclear magnetic resonance (NMR) spectroscopy (Figure 8A–D), which coherently indicated that γ-AApeptide 19 interacted with MDM2 at the binding site virtually the same as that of the MDM2-p53 PPI. It is intriguing that CD studies show that these sulfono-γ-AApeptides are already helically structured in aqueous solution in the absence of MDM2, in contrast to the random coiled p53 peptide,^24,25 which may be accountable for their stronger binding affinity to MDM2 than p53. It should be noted that these left-handed sulfono-γ-AApeptides were completely resistant to proteolytic degradation, compared to the instability of p53 (Figure 8E).

Table 1.

Structures of Sulfono-γ-AApeptides Investigated for the Disruption p53–MDM2 Interaction^a

Peptide	Sequence	Kd(nM)	IC50(μM)
p53	QETFSDLWKLLPEN	208	4.61
Nutlin			0.6
17		98	3.95
18		>5000	–
19		26	0.891

^aThe side chains mimicking Phe19, Trp23, and Leu26 in p53 are shown in blue. Reproduced with permission from ref 15. Copyright 2020 American Chemical Society.

Figure 8.

(A-D) Chemical shift mapping of 19 binding to MDM2. (A) Overlay of ¹⁵N heteronuclear single quantum coherence (HSQC) spectra of MDM2 before (blue resonances) and after (red resonances) the addition of 19. HSQC spectra were collected with a 2-fold stoichiometric excess of 19. (B) Average chemical shift changes, in parts per million (ppm), for the amide proton and nitrogen resonances in MDM2 p53BD residues binding to 19. (C and D) Surface image of the MDM2 p53BD structure. (E) Analytic high-performance liquid chromatography (HPLC) trace of p53 and 19 before and after incubation with Pronase (0.1 mg/mL) in 100 mM, pH 7.8 ammonium bicarbonate buffer at 37 °C. Reproduced with permission from ref 15. Copyright 2020 American Chemical Society.

4.1.1.2. Right-Handed Homogeneous d-Sulfono-γ-AApeptides.

As enantiomers of left-handed sulfono-γ-AApeptides, d-sulfono-γ-AApeptides are expected to be right-handed helical foldamers possessing an α-peptide-like folding conformation. As a result, d-sulfono-γ-AApeptides could also be promising candidates for designing α-helical mimetics (Figure 9).

Figure 9.

(A-F) The chemical and crystal structures of the α-peptides (A and B), chemical and crystal structures of homogeneous l-sulfono-γ-AApeptides (C and D), and chemical and modeled structures of homogeneous d-sulfono-γ-AApeptides (E and F). (G and H) Schematic representation of the distribution of side chains from homogeneous d-sulfono-γ-AApeptides based on computational modeling. (G) Side view and (H) top view of the helical wheel. Reproduced with permission from ref 26. Copyright 2020 American Chemical Society.

To demonstrate the feasibility of d-sulfono-γ-AApeptides for the mimicry of helical domain of proteins, we still chose the p53-MDM2 PPI as the model system.²⁶ As shown in Figure 9, like the l-sulfono-γ-AApeptides, the 2a, 4a, and 6a positions in γ-AApeptide 20 were on the same face of the helical scaffold, mimicking the key side chains Phe19, Trp23, and Leu 26 of p53. The determined K_d value of the γ-AApeptides 20 was 220 nM, which was similar to that of p53 and MDM2 (Figure 10A). With further development, γ-AApeptide 21 was found to be most potent (K_d = 27.5 nM). Computational modeling further suggests that d-sulfono-γ-AApeptides are ideal for the mimicry of the α-helix (Figure 10B–D) due to their same helical handedness and pitch.

Figure 10.

(A) Structures of d-sulfono-γ-AApeptides investigated for the disruption the p53-MDM2 interaction. The side chains mimicking Phe19, Trp23, and Leu26 in p53 are shown in blue. (B-D) The crystal structure of the interaction of p53 with MDM2 (PDB: 1YCR) (B), modeling of the lead homogeneous l-sulfono-γ-AApeptide (C), and the designed homogeneous d-sulfono-γ-AApeptide 21 (D) interaction with MDM2. p53 and the homogeneous d-sulfono-γ-AApeptide are shown as a magenta cartoon, the homogeneous l-sulfono-γ-AApeptide is shown as a green cartoon, and MDM2 is shown as a gray cartoon. Reproduced with permission from ref 26. Copyright 2020 American Chemical Society.

4.1.2. Disruption of BCL9-β-Catenin PPIs.

The Wnt-β-catenin signaling pathway plays an important role in embryonic development and tissue homeostasis, as well as several types of human cancers, such as colon cancer, breast cancer, melanoma, and prostate cancer.^27,28 As a central intermediary of signaling, β-catenin regulates the cell cycle and apoptosis by controlling the expression of several key genes. In fact, the transcriptional activation of the Wnt/γ-catenin signaling pathway depends on the formation of β-catenin super complexes involving BCL9 or BCL9-like (B9L) and the transcription factor T cell factor (Tcf)/lymphocyte enhanced binding factor (Lef) family. Thus, new anticancer drugs could be developed by inhibiting the Wnt/ β-catenin signal transduction via the disruption of BCL9-β-catenin PPIs.

The crystal structure of the β-catenin-BCL9-TCF-4 ternary complex shows that the α-helix region of BCL9 interacts with the binding pocket of β-catenin (Figure 11A and 11B).²⁹ The key residues R359, L363, L366, I369, and L373 are located on the same side of the BCL9 helix structure, forming hydrophilic and hydrophobic contacts to bind the β-catenin surface. Despite its clear mechanism of action, it is still a challenge to design effective inhibitors that can enter cells to block their protein–protein interactions. This is mainly due to the interaction between BCL9 and β-catenin, which is a helical segment of BCL9 with approximately 25 residues.^30–32 Although there are reports for the use of small molecules and peptide inhibitors to disrupt β-catenin-BCL9 protein-protein interactions,^30–32 there are few examples of unnatural peptidomimetic inhibitors.

Figure 11.

(A and B) The α-helical HD2 domain of BCL9, which directly engages a surface groove of β-catenin, provided the template for structural stabilization by hydrocarbon stapling (PDB: 2GL7). (A) Cartoon representation of the residues of BCL9 (red), which are critical for binding to β-catenin, are shown as sticks. (B) BCL9 is shown in as a stick model, and β-catenin is represented with a surface model. (C-F) The schematic representation of the distribution of side chains from sulfono-γ-AApeptides. (C) Side and (D) top view of the helical wheel. (E) The position map of critical residues of the BCL9 helix. (F) The position map of side chains of sulfono-γ-AApeptides that are designed to mimic the residues in panel E. (G) The BCL9 peptide 22 and lead compounds 23–25. Reproduced with permission from ref 3. Copyright 2020 United States National Academy of Sciences.

4.1.2.1. Left-Handed Homogeneous Sulfono-γ-AApeptides.

Previous studies have shown that the α-helix HD2 domain of BCL9 can directly interact with the surface groove of β-catenin.³² However, the BCL9 helix is much longer than p53, posing a more significant challenge for the design of helical mimetics. We designed a series of left-handed homogeneous sulfono-γ-AApeptides to mimic the helix structure of BCL9, hoping to disrupt β-catenin-BCL9 protein-protein interactions.³ As shown in Figure 11C and 11D, the chiral side chains 2a, 4a, 6a, 8a, and 10a are on the same surface of the spiral skeleton of sulfono-γ-AApeptides. Therefore, this face was chosen to mimic those key residues of the regular BCL9 helical domain. The position of these residues (Figure 11E) on the α-helical scaffold shows that, except for I369, R359, L363, L366, and L373 are almost on the same line. A careful comparison of the structure of the helical skeleton and the BCL9 peptide reveals that 8b instead of 8a can be mimicked by I369 (Figure 11F). As 8b is a sulfonyl side chain, we assumed that methylsulfonyl is sufficient because the sulfonyl group protrudes more than the chiral side chain on the helical sulfono-γ-AApeptide.

On the basis of our design, a series of unnatural helical sulfono-γ-AApeptides was synthesized to mimic the α-helical region of natural BCL9.³ These unnatural helical peptidomimetics could effectively and specifically inhibit cancer-related β-catenin-BCL9 PPIs. Cell-based studies showed that sulfono-γ-AApeptides have cell permeability (Figure 12A) and could effectively inhibit the growth of cancer cells and activate Wnt/β-catenin signaling. TOPFlash/FOPFlash luciferase report analyses showed that sulfono-γ-AApeptides could selectively inhibit the transactivation of the Wnt/β-catenin signaling pathway. Protein pull-down and co-immunoprecipitation (co-IP) experiments showed that these sulfono-γ-AApeptides could inhibit β-catenin–BCL9 protein-protein interactions in cells by binding to β-catenin (Figure 12B–C). It is worth noting that these sulfono-γ-AApeptides are completely resistant to pronase hydrolysis.

Figure 12.

(A) Confocal fluorescence microscopy images of SW480 cells treated with 1 μM and 10 μM of the FITC-labeled peptide 22 and sulfono-γ-AApeptides 23–25 for 2 h (magnification, 630×). (B) The SW480 cell lysate was incubated with 24-biotin or 25-biotin, followed by streptavidin pull-down experiments. (C) Co-immunoprecipitation (co-IP) experiments to evaluate the disruption of the β-catenin-BCL9 PPI by 25 in Wnt/ β-catenin hyperactive cancer cells. Reproduced with permission from ref 3. Copyright 2020 United States National Academy of Sciences.

4.2. Mimicry of GLP-1 Peptide

Glucagon-like peptide 1 receptor (GLP-1R)^33–35 belongs to the B family of GPCRs, and its glucagon helical peptide ligand GLP-1 analogues³⁵ are expected to be a candidate drug for the treatment of type 2 diabetes and obesity. However, due to the rapid degradation by proteases, the half-life of GLP-1 is very short.³⁶ The stabilization of GLP-1 is essential for the development of drugs to treat diabetes. Side chain cross-linking strategies have been used to stabilize the conformation and metabolism of GLP-1.³⁷ However, unexpected contact may occur between the crosslinked chain and GLP-1R, and the degree of proteolytic stabilization may be limited. Helical mimetics have also been used as an alternative strategy to develop proteolytically stable GLP-1R agonists.^38,39 However, there is no report on the use of the entire unnatural skeleton to mimic GLP-1. This is because unnatural peptidomimetics mimicking long α-helices are very challenging due to the difference in helicity between α-helix and helical foldamers.

4.2.1. Left-Handed Homogeneous Sulfono-γ-AA-peptides.

We were intrigued whether our homogeneous sulfono-γ-AApeptides could be used to mimic the very long helical peptide GLP-1 and could functionally form effective GLP-1R agonists. An analysis of the GLP-1-GLP-1R interaction revealed that multiple key residues on GLP-1 are tightly bound to GLP-1R.^34,40 In short, GLP-1 participates simultaneously with its N-terminal domain and C-terminal domain to interact with GLP-1R (Figure 13C). In the N-terminal domain, the interaction of the polar residues H7, E9, T11, T13, and S17 with the seven transmembrane domains (7TMD) of the receptor is most critical (Figure 13D). The truncated N-terminal domain GLP-1 peptide derivatives are still agonists of the nM concentration of GLP-1R in the cAMP analysis.⁴¹ In contrast, the hydrophobic residues F28 and L32 dominate the interaction between the C-terminal α-helix of GLP-1 and the extracellular domain (ECD) of GLP-1R. A closer look at the key residues on GLP-1 (Figures 13D and 13E) shows that H7 and T11 are on the same plane (X) of the helix; F28 and L32 on the Y plane; and E9, T13, and T11 on the Z Up plane (arbitrarily specify X, Y, or Z).

Figure 13.

(A and B) Schematic representation of the distribution of side chains from the sulfono-γ-AApeptide in panel F. (A) Side view and (B) top view of the helical wheel. (C) GLP-1 binds to GLP-1R (PDB: 5VAI). GLP-1 (7–36) is shown in blue, and GLP-1R is represented as a green cartoon. (D) The helical domain of GLP-1 with critical residues are presented as sticks. (E) Design of sulfono-γ-AApeptide 27, with side chains mimicking the important residues in panel B. The helix was built on the crystal structure. X, Y, and Z was designated to indicate the face of the residues on the helix. (F) The Structures and agonist activities of GLP-1(7–36) 26 and lead sulfono-γ-AApeptide 27. (G) Analytic HPLC traces of GLP-1(7–36) 26 and lead sulfono-γ-AApeptide 27 before and after incubation with pronase (0.1 mg/mL). (H) The serum stability of 26 and 27 at 37 °C for 24 h. (I and J) Pharmacodynamics of the GLP-1 mimic peptide 27 in mice. A single dose of peptides was intraperitoneally administered into mice 1 h before the oral glucose tolerance test (OGTT) (2 g/kg glucose). (I) Blood glucose concentrations were monitored for up to 120 min after oral glucose challenge. (J) The average area under the curve (AUC) was calculated from OGTT data. Results show the mean ± the standard error of the mean (SEM) of six mice per treatment group; *P < 0.05 versus vehicle; t test. Reproduced with permission from ref 4. Copyright 2020 the American Association for the Advancement of Science.

On the basis of these preliminary analyses, we designed the helical sulfono-γ-AApeptides that may mimic multiple faces of the native GLP-1 helix.⁴ As shown in Figure 13A and B, the chiral side chain and the sulfonyl side are perfectly distributed on the four sides of the helical scaffold, which can be used to mimic the key residues of the whole GLP-1 helical domain interacting with GLP-1R. A further comparison showed that side chains 1a and 3a can mimic residues on the X-face of GLP-1, and 11b and 13b may replicate the function of residues on the Y-face of GLP-1. We further speculated that 2a, 4a, and 6a can capture the functions of E9, T13, and S17. On the basis of this design strategy, we synthesized a series of sulfono-γ-AApeptides that could structurally and functionally mimic residues on multiple faces of the GLP-1 α-helix domain. These unnatural helical peptidomimetics show effective GLP-1R agonistic activity in cell-based assays and oral glucose tolerance tests in vivo (Figure 13I and J). They also exhibited excellent proteolytic stability (Figure 13G and H), thereby enhancing their potential for biological applications.

4.3. Helical Sulfono-γ-AApeptides with Both Aggregation-Induced Emission and Circularly Polarized Luminescence Properties

4.3.1. Left-Handed Homogeneous Sulfono-γ-AApeptides.

Motivated by the folding structure, we also investigated the material applications of these foldameric sulfono-γ-AApeptides. It is known that the AIE (aggregation-induced emission) phenomenon was induced by the restricted intramolecular rotation (RIR) process, i.e., upon aggregation, the rotation of the AIEgens (tetraphenylethylene (TPE)) was confined by the constrained chemical environment, thus the activation of the radiative decay was stimulated over the nonradiative decay path-ways.⁴² To this end, four TPE-sulfono-γ-AApeptides 28–31 were designed and synthesized (Figure 14), with the assumption that, as the rotation of the TPE would be restricted by the rigid left-handed helical structure, the peptides would be emissive at both the solution and aggregation states.¹⁶ To our delight, all the peptides 28–31 have shown fluorescence at each stage of the PBS percentage as we anticipated (Figure 14). It was believed that there was a combined action of the helical scaffold restriction and the aggregation-induced emission enhancement (AIEE).^43,44 Interestingly, the circularly polarized luminescence (CPL) signals were also observed in these peptides with a relatively large dissymmetry factor (glum).

Figure 14.

Chemical and crystal structures of the sulfono-γ-AApeptides 28–31. Reproduced with permission from ref 16. Copyright 2020 John Wiley & Sons, Inc.

4.3.2. Right-Handed 1:1 l-α/l-Sulfono-γ-AA Heterogeneous Peptides.

We also incorporated the TPE into the right-handed 1:1 α/sulfono-γ-AA hybrid peptides and studied their structure and properties.¹³ In the crystal structure of the TPE-α/sulfono-γ-AApeptide 32, the right-handed peptide revealed a 13-atom hydrogen-bonding pattern, which is in good agreement with the parameters of the above-mentioned 4₁₃ structures (Figure 15).

Figure 15.

Chemical and crystal structure of TPE-α/sulfono-γ-AApeptide 32. (A) Chemical structure and the 13-atom hydrogen-bonding pattern. (B) Crystal structure of the bonding pattern. (C) Helical cartoon of the crystal structure. (D) Crystal packing of 1 along the peptide axis. (E) Cartoon structure of structure in panel D. (F) Packing mode of the crystal. Reproduced with permission from ref 13. Copyright 2020 American Chemical Society.

The TPE-α/sulfono-γ-AApeptide 32 was dissolved and was fluorescent in pure water (good solvent). As the percentage of PBS buffer (poor solvent) was gradually increased to 99%, the quantum yield did not significantly change, implying that, instead of AIE, the fluorescence was induced from the restriction on TPE bond rotation. In addition, 32 also generated CPL signals with a large glum data ca. 1.2 × 10⁻² due to the chiral helical scaffold.

5. FUTURE PERSPECTIVE OF SULFONO-γ-AAPEPTIDES

The availability of a crystal structure has made it possible to elucidate the relationship in helical sulfono-γ-AApeptides and rationally design functional ones in different biological systems. The classic p53-MDM2 PPI initially validated our design rationale to mimic one face of the α-helix. We then chose to interrogate the more challenging PPI target β-catenin-BCL9, involving residues on the very long BCL9 helix at the interface. The success prompted us to undertake an even more intricate task, mimicry of GLP-1, which engages residues from multiple faces of the α-helical scaffold. These studies manifested that sulfono-γ-AApeptides could play a role in the modulation of biological relevant PPIs due to their ability to mimic residues on the multiple faces of a long α-helix.

Although these foldameric sulfono-γ-AApeptides are promising, further exploration is demanded for this new class of foldamers. First, more protein targets should be studied to demonstrate the principle governing the structure function relationship of helical sulfono-γ-AApeptides. Since sulfono-γ-AApeptides, 1:1 α/ sulfono-γ-AApeptides, and d-sulfono-γ-AApeptides all demonstrated their ability to mimic α-helix regardless of their helical handedness, it is intriguing to study their structure–function relationship for all the PPI targets in the future, which may enhance the successful rate of identifying potent molecular candidates. Second, to improve the potency, hybrid peptides instead of homogeneous ones may be explored.⁶ Another research endeavor would be obtaining cocrystal structures of sulfono-γ-AApeptides with protein targets. Although we have the crystal structures of sulfono-γ-AApeptides alone, and we also have NMR evidence to back up the binding mode of sulfono-γ-AApeptides toward protein targets, it would be imperative to directly visualize how sulfono-γ-AApeptides interact with protein residues on the protein binding sites. In addition, based on our current results, it seems that sulfono-γ-AApeptides are much more cell permeable than canonical peptides, which is observed for the modulation of β-catenin-BCL9 PPIs,³ possibly due to their more pronounced helical conformation and the less availability of amide bonds. However, it must be admitted that the cell permeability is not only related to the molecular scaffold and folding conformation but also related to the side chain functional groups on the backbone. Of course, in the case of when the poor permeability of sulfono-γ-AApeptides is observed, backbone stapling^45,46 and conjugation with cell-penetrating peptides^47–49 could still be employed.

We are optimistic that cocrystal structures and new applications of sulfono-γ-AApeptides will be discovered through rational design and synthesis. Those findings will in turn shed light on sulfono-γ-AApeptides for their applications in material and biomedical sciences both fundamentally and practically.

Biographies

Peng Sang is a postdoctoral associate at the University of South Florida. His research interests include the exploration of folding structures of γ-AApeptides and the development of novel sulfono-γ-AApeptide helical foldamers potentially targeting cancer and diabetes.

Yan Shi is a postdoctoral associate at the University of South Florida. Her current research interests focus on the synthesis and combinatorial screening of γ-AApeptides and antimicrobial peptide-mimetic polymers.

Bo Huang is a postdoctoral associate at the University of South Florida. His research interest lies in the rational design of biomimetic catalysts based on sulfono-γ-AApeptides.

Songyi Xue is a graduate student at the University of South Florida. His research interest is the synthesis and combinatorial screening of γ-AApeptides.

Timothy Odom is a graduate student at University of South Florida. His research interest is the design and synthesis of antimicrobial γ-AApeptides.

Jianfeng Cai is a preeminent professor in the Department of Chemistry at the University of South Florida. His research group focuses on the development and application of AApeptide-based peptidomimetics.