Evaluation of Efficient Non-reducing Enzymatic and Chemical Ligation Strategies for Complex Disulfide-Rich Peptides

Abstract

Double-knotted peptides identified in venoms and synthetic bivalent peptide constructs targeting ion channels are emerging tools for the study of ion channel pharmacology and physiology. These highly complex and disulfide-rich peptides contain two individual cystine knots, each comprising six cysteines and three disulfide bonds. Until now, native double-knotted peptides, such as Hi1a and DkTx, have only been isolated from venom or produced recombinantly, whereas engineered double-knotted peptides have successfully been produced through enzymatic ligation using sortase A to form a seamless amide bond at the ligation site between two knotted toxins, and by alkyne/azide click chemistry, joining two peptide knots via a triazole linkage. To further pursue these double-knotted peptides as pharmacological tools or probes for therapeutically relevant ion channels, we sought to identify a robust methodology resulting in a high yield product that lends itself to rapid production and facile mutational studies. In this study, we evaluated the ligation efficiency of enzymatic (sortase A5°, butelase 1, wild-type OaAEP 1, C247A-OaAEP 1, and peptiligase) and mild chemical approaches (α-ketoacid-hydroxylamine, KAHA) for forming a native amide bond linking the toxins while maintaining the native disulfide connectivity of each pre-folded peptide. We used two Na_V1.7 inhibitors: PaurTx3, a spider-derived gating modifier peptide, and KIIIA, a small cone snail-derived pore blocker peptide, which have previously been shown to increase affinity and inhibitory potency on hNa_V1.7 when ligated together. Correctly folded peptides were successfully ligated in varying yields, without disulfide bond shuffling or reduction, with sortase A5° being the most efficient, resulting in 60% ligation conversion within 15 min. In addition, electrophysiology studies demonstrated that for these two peptides, the amino acid composition of the linker did not affect the activity of the double-knotted peptides. This study demonstrates the powerful application of enzymes in efficiently ligating complex disulfide-rich peptides, paving the way for facile production of double-knotted peptides.

Figure 1.

Scheme showing enzymatic and chemical ligation approaches examined in this study. The recognition sequences of the enzymes are shown as single letter amino acid residue codes. The N-terminal peptide is shown in blue, and the C-terminal peptide is shown in orange. (A) SrtA, (B) OaAEP 1, (C) butelase 1, (D) peptiligase, and (E) KAHA ligation.

Figure 2.

Enzymatic and chemical ligation approaches of PaurTx3 and KIIIA. RP-HPLC chromatograms of PaurTx3 and KIIIA starting material (top) and ligated products (bottom) following completion of the ligation reaction. (A) Ligation of PaurTx3[NGL] and [G₈]KIIIA using wt-OaAEP 1. (B) Ligation of PaurTx3[NGL] and [GVG₆]KIIIA using wt-OaAEP 1. (C) Ligation of PaurTx3[NGL] and [GVG₆]KIIIA using C247A-OaAEP 1. (D) Ligation of PaurTx3[NHV] and [GVG₆]KIIIA using butelase 1. (E) KAHA ligation between PaurTx3-ka and [OprG₈]KIIIA. (F) Ligation of PaurTx3[LPATGG] and [G₅]KIIIA using SrtA5° (SrtA5° data adapted from Tran et al. and included for comparison purposes).

Figure 3.

Enzymatic and chemical ligation products following PaurTx3 and KIIIA ligation. Analytical RP-HPLC traces and observed masses (ESI-MS) of purified bivalent peptides. (A) Product from wt-OaAEP 1 ligation using PaurTx3[NGL] and [G₈]KIIIA. (B) Product from wt-OaAEP 1 (identical product from C247A-OaAEP 1, not shown) and butelase 1 ligation using PaurTx3[NGL] and PaurTx3[NHV], respectively, and [GVG₆]KIIIA. (C) Product from KAHA ligation using PaurTx3-ka and [OprG₈]KIIIA. (D) Secondary Hα chemical shifts for P[NG₈]K, P[NGVG₆]K, and P[^hSG₈]K, overlaid with published KIIIA (from Khoo et al.), and native PaurTx3 (from Agwa et al.). The linker region was omitted for clarity.

Figure 4.

Pharmacology at hNa_V1.7 and stability of bivalent peptides. (A) Representative current trace before and after the addition of 100 nM P[NG₈]K. (B) Representative current trace before and after the addition of 100 nM P[NG₈]K. (C) Representative current trace before and after the addition of 100 nM P[^hSG₈]K. (D) Concentration–response curves of bivalent peptides on hNa_V1.7 expressed in HEK293 cells assessed using automated whole cell patch-clamp electrophysiology (n = 3). (E) Serum stability of P[NGVG₆]K (ligated with wt-OaAEP 1, C247A-OaAEP 1, or butelase 1), P[^hSG₈]K (KAHA), and P[LPATG₅]K (ligated with SrtA5°) in human serum over 24 h. The linear linker sequence KYQILPATG₅ (SrtA5° linker sequence) was used as the positive control. P[LPATG₅]K and linear linker sequence data are adapted from Tran et al. for comparison purposes.