Genome minimization of a Burkholderia bacterial host

Abstract

Genome minimization, including the deletion of endogenous gene clusters that encode natural products, is a common strategy to improve the yield of heterologous products. We have been interested in developing Burkholderia sp. FERM BP-3421 as an alternative bacterial host. Instead of indiscriminately deleting gene clusters, which may have deleterious effects, we guided our efforts using transcriptomics data from production cultures. The genome of FERM BP-3421 is subdivided into two chromosomes and two plasmids. The top transcribed gene clusters were those encoding polyketide-nonribosomal peptide spliceostatins on plasmid p1 and nonribosomal peptide selethramide on chromosome 1. Deletion of the spliceostatin cluster had been shown to improve titers of the ribosomal peptide capistruin, whereas we showed that deletion of the selethramide cluster had no effect on capistruin titers. We next targeted the two endogenous plasmids using a CRISPR-Cas12a strategy, resulting in an 11 % reduction in genome size. The plasmid cured strains showed improved growth and 20–40 % increased production of capistruin depending on whether one or both plasmids were deleted. However, deletion of p2 alone negatively affected the heterologous production of two distinct polyketide-nonribosomal peptides. The p2⁻ strain produced only 5–23 % of the glidobactin A and megapolipeptin A titers compared to the wild type, respectively, whereas titers were restored to wild type levels in the p1⁻ p2⁻ strain. The observation that p2 appears to contain functions that support polyketide-nonribosomal peptide biosynthesis was unexpected and sets the stage for future studies aimed at identifying these functions and further enabling engineering efforts that may be widely applicable to other strains.