A strategy for producing isotopically labeled peptides with antimicrobial activity or with short in vivo lifetime in Escherichia coli

Abstract

Engineered Escherichia coli (E. coli) strains have been widely used to produce isotopically labeled peptides for NMR characterization on their structures and interactions. However, production of antimicrobial peptides (AMPs) by E. coli is still challenging, because AMPs are toxic to E. coli host and would lead to cell death after induction. On the other hand, expression of short peptides in E. coli host often encounter problems of the short in vivo lifetime of the peptides, which were rapidly degraded by endogenous enzymes during expression and purification steps. This report presents a practical method for overcoming these bottlenecks to enable E. coli to express AMPs and peptides that have short in vivo lifetime. This design uses the fusion of thioredoxin tags at both the N- and C-termini of the target peptides. The steric effect of the large soluble tags at both ends of the peptide reduces peptide accessibility, thereby enhancing their in vivo stability and eliminating the toxicity associated with AMPs. The approach was validated using an AMP A3K/L7K-LAH4 (K3K7) and a membrane fusion peptide (FP), which is a segment of the spike protein of SARS-CoV-2 and functions in fusing viral membranes and host cell membranes. Fusion expression of K3K7 with a thioredoxin tag only at the N-terminal resulted in high toxicity to the host cells, leading to impaired cell growth and a failure to obtain expressed fusion protein. In contrast, the fusion proteins from both termini were successfully expressed and purified. In the case of the FP, the fusion of thioredoxin at both termini significantly enhanced its stability, protecting it from enzymatic degradation during expression and purification steps. On the contrary, the FP with thioredoxin fused only at the N-terminal was found to be unstable in E. coli host strains. As stable isotope labeling on peptide is essentially important in NMR-based structure and interaction studies, we also demonstrated that the developed approach enables efficient ¹⁵N labeling for NMR studies. This strategy may also be extended to produce other challenging peptides.