Robust performance of a live bacterial therapeutic chassis lacking the colibactin gene cluster

Abstract

E. coli Nissle (EcN) is a non-pathogenic probiotic bacterium of the Enterobacteriaceae family that has been used for over a century to promote general gut health. Despite the history of safe usage of EcN, concerns have been raised regarding the presence of the pks gene cluster, encoding the genotoxin colibactin, due to its association with colorectal cancer. Here, we sought to determine the effect of pks island removal on the in vitro and in vivo robustness and activity of EcN and EcN-derived strains. A deletion of the pks island (Δpks) was constructed in wild type and engineered strains of EcN using lambda red recombineering. Mass spectrometric measurement of N-myristoyl-D-asparagine, released during colibactin maturation, confirmed that the pks deletion abrogated colibactin production. Growth curves were comparable between Δpks strains and their isogenic parents, and wild type EcN displayed no competitive advantage to the Δpks strain in mixed culture. Deletion of pks also had no effect on the activity of strains engineered to degrade phenylalanine (SYNB1618 and SYNB1934) or oxalate (SYNB8802). Furthermore, 1:1 mixed dosing of wild type and Δpks EcN in preclinical mouse and nonhuman primate models demonstrated no competitive disadvantage for the Δpks strain with regards to transit time or colonization. Importantly, there was no significant difference on in vivo strain performance between the clinical-stage strain SYNB1934 and its isogenic Δpks variant with regards to recovery of the quantitative strain-specific biomarkers d5- trans-cinnamic acid, and d5-hippuric acid. Taken together, these data support that the pks island is dispensable for Synthetic Biotic fitness and activity in vivo and that its removal from engineered strains of EcN will not have a deleterious effect on strain efficacy.