Synthesis and stability studies of constrained peptide–antimony bicycles

Introduction

Peptide-based drugs occupy a growing niche in the therapeutic landscape.^1–3 A wide range of chemical transformations to modify peptides contributes to their success.⁴ At the forefront of this are constrained peptides. Their conformational freedom is restricted into cyclic, stapled or multicyclic topologies by one or more key modifications.^5–8 This generally improves the pharmaceutical properties of the peptide, such as prolonged metabolic stability, increased binding affinities due to pre-organisation and, in some instances, enhanced cell permeability.^5,9 Although constrained peptides have the potential to match the high binding affinity and selectivity of antibodies, they remain accessible through chemical synthesis.^5,10,11 Owing to these qualities, bioactive peptide macrocycles are considered highly promising next-generation drugs.¹²

The identification of targeted peptide macrocycles can be facilitated by genetically encoded libraries and display screenings.^13,14 Resulting hit compounds can be further refined through conventional medicinal chemistry where certain properties such as binding affinity or cell membrane permeability are further improved.^15–17 With increasing uptake of these drug discovery techniques, the next wave of therapeutic peptide macrocycles is expected to originate from such encoded libraries.^18,19 It is therefore a priority in peptide research to innovate peptide modification and cyclisation with a focus on biocompatibility for display screenings.²⁰ As multicyclic peptides with enhanced structural constraints emerge as promising drugs, an increasing number of bicyclic peptide syntheses are appearing in the literature.^7,21–24

The pronounced reactivity of cysteine residues is frequently exploited for peptide (multi)cyclisation.^25,26 For instance, alkylating agents such as 1,3,5-tris(bromomethyl)benzene (TBMB) are widely used to generate bicyclic peptide libraries on phage.²⁷ Our group has described an alternative class of peptide bicycles scaffolded around a single thiophilic element.²⁸ We pioneered the synthesis of bioactive and cell-permeable peptide–bismuth bicycles, where bismuth(III) is covalently bound to three cysteine residues of the peptide.^28,29 Owing to the minimalistic ‘scaffold’ of a single atom, the constraint of the bicyclic peptide is enhanced compared to conventional scaffolding agents.²⁸ This single-element strategy has recently been expanded to arsenic(III).³⁰ Additionally, the compatibility of both bismuth(III) and arsenic(III) modification techniques with phage display screenings was successfully established.^30,31 We hypothesised that antimony, located in group 15 between arsenic and bismuth, is similarly suitable for peptide bicyclisation. Hence, this study describes the first synthesis, characterisation and stability evaluation of Sb^III-centred bicyclic peptides.

Results and discussion

Antimony trichloride (SbCl₃) from a concentrated dimethyl sulfoxide (DMSO) stock served as the reagent to access Sb^III-centred bicyclic peptides. The reagent was added to a peptide comprising three cysteine residues (1a) in aqueous buffer at physiological pH (Fig. 1a). Liquid chromatography–mass spectrometry (LCMS) revealed the presence of the resulting bicyclic peptide 1b with ≥95% conversion (Fig. 1b, Table 1). High-resolution mass spectrometry (HRMS) confirmed the identity of the peptide–antimony bicycle. Its distinctive metalloid core exhibits a characteristic isotope pattern (Fig. 1c). We further expanded this technique to various triple-cysteine peptides 2a–7a – covering hydrophobic, polar, acidic and basic amino acids – and converted them into their respective bicyclic peptide (Table 1). LCMS analysis confirmed ≥95% presence of the bicyclic peptides after exposure to SbCl₃ (Table 1).

Fig. 1.

Synthesis of model peptide–antimony bicycle 1b. (a) Reaction scheme. (b) High-performance liquid chromatography (HPLC) trace. (c) Simulated isotope pattern (grey) and high-resolution positive electrospray ionisation (ESI⁺) mass spectrum (blue) of the [M + Na]⁺ adduct.

Next, we assessed the stability of the peptide–antimony bicycles in the presence of chelators and thiol reagents (Fig. 2, Supplementary Tables S3–S5). Specifically, we conducted competition experiments with ethylenediaminetetraacetic acid (EDTA) and glutathione (GSH). The evaluated bicyclic peptides 1b and 2b exhibited similar trends in the presence of the hexadentate metal chelator EDTA. At a five-fold excess of EDTA, ~90% of the peptide–antimony bicycles remained intact. Even at a 25-fold excess of EDTA, complete dissociation of Sb^III from the peptides was not observed, as the fraction of peptide–antimony bicycles remained at 22 and 49% for 1b and 2b respectively (Fig. 2a), suggesting an equilibrium between the Sb(peptide) and Sb(EDTA) complexes with a stronger stability constant of the peptide–antimony complex. To assess the stability of our bicycles in a natural reducing environment, we determined their stability in the presence of GSH, the most abundant intracellular thiol.³² We anticipated that binding of three thiols from cysteine residues to Sb^III may be affected by competition with other thiol-containing compounds such as GSH. For 1b, we observed increased decomposition of the bicycle with increasing equivalents of GSH, resulting in 40% intact peptide bicycle at a 25-fold excess of GSH. Bicycle 2b was found to be more resistant. Even at the highest analysed GSH concentration (25 equiv., 12.5 mM) only ~20% decomposition was observed (Fig. 2b). Considering that the GSH concentration in most cells is ~2 mM, this result suggests that peptide–antimony bicycles are sufficiently guarded.³³

Fig. 2.

Stability of peptide–antimony bicycles 1b (orange square) and 2b (green circle) in competition with (a) ethylenediaminetetraacetic acid (EDTA) and (b) glutathione (GSH). Bicycles (0.5 mM) were incubated with increasing equivalents (0.5, 5, 10 and 25) of EDTA (a) or GSH (b) for 1 h at room temperature to assess stability.

Like antimony(III), the higher pnictogen bismuth(III) has been studied for its medicinal properties.³⁴ The high affinity of bismuth(III) towards thiolates has been previously demonstrated for small molecules as well as peptides.^{28–31,35,36} Therefore, we were keen to assess whether bicycles 1b and 2b would withstand increasing concentrations of bismuth(III) from gastrodenol (bismuth tripotassium dicitrate) (Fig. 3). Remarkably, we observed that only one equivalent of gastrodenol is sufficient to replace >80% of antimony(III) in 1b and 2b by bismuth(III) (Fig. 3a, c). A further increase in the gastrodenol concentration led to >90% conversion into the peptide–bismuth bicycles (1c and 2c respectively, Table 1, Fig. 3b, d). We attribute this observation to the higher thiophilicity of bismuth(III) in comparison to that of antimony(III),³⁷ resulting in a larger stability constant of the Bi(SR)₃ complex. As the bismuth–antimony competition experiments were conducted within minutes of incubation, we note that the kinetics of peptide ligand exchanges between both metals are rather fast.

Fig. 3.

Conversion of peptide–Sb^III bicycles into peptide–Bi^III bicycles using gastrodenol in aqueous buffer. Formation of peptide–Bi^III bicycles 1c (b) and 2c (d) with increasing equivalents of gastrodenol. HPLC (254 nm) overlay and mass spectra of (a) 1b (orange) and 1c (maroon), and (c) 2b (green) and 2c (dark green).

Conclusion

Our study introduces peptide–antimony bicycles as an additional class of single-element centred bicyclic peptides. This expands the repertoire of single atom ‘scaffolds’ for constrained peptide synthesis from arsenic(III) and bismuth(III) to antimony(III). These metal(loid)–peptide ‘scaffolds’ exhibit unique properties due to the combination of the stability and structural diversity of constrained peptides coupled with the unique characteristics of the metal(loid) in the centre. In situ generated Sb^III-centred bicycles are considerably resistant to degradation in the presence of glutathione at physiological concentrations or EDTA. The substitution of antimony by bismuth in competition experiments supports the case of peptide–bismuth bicycles as the most promising class. As opposed to arsenic and antimony, bismuth is not generally known for its toxicity.^38,39 In the future, it could be intriguing to investigate how these peptides affect target specificity and bioavailability. Overall, our study underscores the untapped potential of elements rarely used in chemical biology, which we use here to advance the synthesis of peptide macrocycles.

Experimental

Supplementary material

Supplementary material is available online.