Improved poly-γ-glutamic acid production in Bacillus amyloliquefaciens by modular pathway engineering

Highlights

• Improve γ-PGA production by a systematic modular pathway engineering strategy.

• The by-product pathways were blocked to increase γ-PGA production and purity.

• The degrading enzyme genes were deleted to improve γ-PGA production.

• The synthetic small regulatory RNAs were used for B. amyloliquefaciens engineering.

• γ-PGA production increased 2.91-fold in shake flask and 5.34-fold in 5-L fermenter.

Abstract

A Bacillus amyloliquefaciens strain with enhanced γ-PGA production was constructed by metabolically engineering its γ-PGA synthesis-related metabolic networks: by-products synthesis, γ-PGA degradation, glutamate precursor synthesis, γ-PGA synthesis and autoinducer synthesis. The genes involved in by-products synthesis were firstly deleted from the starting NK-1 strain. The obtained NK-E7 strain with deletions of the epsA-O (responsible for extracellular polysaccharide synthesis), sac (responsible for levan synthesis), lps (responsible for lipopolysaccharide synthesis) and pta (encoding phosphotransacetylase) genes, showed increased γ-PGA purity and slight increase of γ-PGA titer from 3.8 to 4.15 g/L. The γ-PGA degrading genes pgdS (encoding poly-gamma-glutamate depolymerase) and cwlO (encoding cell wall hydrolase) were further deleted. The obtained NK-E10 strain showed further increased γ-PGA production from 4.15 to 9.18 g/L. The autoinducer AI-2 synthetase gene luxS was deleted in NK-E10 strain and the resulting NK-E11 strain showed comparable γ-PGA titer to NK-E10 (from 9.18 to 9.54 g/L). In addition, we overexpressed the pgsBCA genes (encoding γ-PGA synthetase) in NK-E11 strain; however, the overexpression of these genes led to a decrease in γ-PGA production. Finally, the rocG gene (encoding glutamate dehydrogenase) and the glnA gene (glutamine synthetase) were repressed by the expression of synthetic small regulatory RNAs in NK-E11 strain. The rocG-repressed NK-anti-rocG strain exhibited the highest γ-PGA titer (11.04 g/L), which was 2.91-fold higher than that of the NK-1 strain. Fed-batch cultivation of the NK-anti-rocG strain resulted in a final γ-PGA titer of 20.3 g/L, which was 5.34-fold higher than that of the NK-1 strain in shaking flasks. This work is the first report of a systematically metabolic engineering approach that significantly enhanced γ-PGA production in a B. amyloliquefaciens strain. The engineering strategies explored here are also useful for engineering cell factories for the production of γ-PGA or of other valuable metabolites.

Introduction

Poly-γ-glutamic acid (γ-PGA) is an important, naturally occurring polyamide consisting of d/l-glutamate monomers (Ashiuchi and Misono, 2002). Unlike typical peptide linkages, the amide linkages in γ-PGA are formed between the α-amino group and the γ-carboxyl group (Kunioka, 1997). γ-PGA exhibits many favorable features such as biodegradable, water soluble, edible and non-toxic to humans and the environment. Therefore, it has been widely used in fields of foods, medicines, cosmetics and agriculture (Shih and Van, 2001) and many unique applications, such as a sustained release material and drug carrier (Li, 2002, Liang et al., 2006), curable biological adhesive, biodegradable fibers (Richard and Margaritis, 2001), and highly water absorbable hydrogels (Park et al., 2001).

γ-PGA can be produced by bacteria, archaea and eukaryotes; however, it is mainly produced by Bacillales order (Candela and Fouet, 2006). γ-PGA-producing strains are classified as either glutamate-dependent strains or glutamate-independent strains (Shih and Van, 2001). Glutamate-dependent strains require their fermentation medium to be supplemented with glutamate. The added glutamate may account for up to 50% of the raw material costs (Zhang et al., 2012), making the strains uneconomical for commercial-scale production. Glutamate-independent strains are preferable for industrial production because of their low cost and simplified fermentation process (Cao et al., 2011). However, the majority of the glutamate-independent strains produce less γ-PGA than the glutamate-dependent strains. Therefore, the construction of a glutamate-independent strain with high γ-PGA yield is important for industrial applications.

Several metabolic engineering strategies have been used to improve γ-PGA production. Heterologous expression of the Vitreoscilla hemoglobin VHb to alleviate oxygen limitation in the later stages of γ-PGA fermentation, the final engineered strains exhibited enhanced cell density as well as enhanced γ-PGA production (Su et al., 2010, Zhang et al., 2013). A mutant strain derived from Bacillus amyloliquefaciens C06 with its epsA gene deleted was deficient in biofilm production and exhibited an increase in γ-PGA production from 3.2 to 6.8 g/L (Liu et al., 2011). Other studies have focused on increasing γ-PGA production by knocking out the genes coding for enzymes that degrade γ-PGA. Certain peptidoglycan hydrolases, such as LytE, LytF, CwlS and CwlO, can degrade γ-PGA (Smith et al., 2000, Bisicchia et al., 2007). Mitsui et al. (2011) investigated the effects of single deletions of the lytE, lytF, cwlS and cwlO genes on γ-PGA production in B. subtilis (natto) and found that only the ΔcwlO strain exhibited increased γ-PGA production. They also studied the effects of single deletions of pgdS and ggt (encoding the γ-glutamyltranspeptidase) genes and found that these deletions had no effect on γ-PGA production. Scoffone et al. (2013) were the first to study multiple deletions of genes for γ-PGA-degrading enzymes. A mutant strain with both of the pgdS and ggt genes deleted had doubled γ-PGA yield compared with the wild-type strain. We also studied the effects of multiple deletions of the pgdS, ggt and cwlO genes on γ-PGA production in the B. amyloliquefaciens NK-1 strain (Feng et al., 2014a). In contrast to the results of Scoffone et al’s, we found that double deletion of the pgdS and ggt genes had no effect on γ-PGA production; however, the strain with double deletion of the pgdS and cwlO genes exhibited a 93% increase in γ-PGA production.

Although many studies have been conducted on this topic, most of them have focused on only one synthetic bottleneck in every round of cellular engineering, and no systematic studies of the effects of modifying multiple synthetic bottlenecks on γ-PGA production in a single strain have been performed. In this study, we used modular pathway engineering to simultaneously optimize the entire biosynthesis pathways, and fine-tune the synthetic pathways and balance the metabolism in the glutamate-independent B. amyloliquefaciens NK-1 strain (Feng et al., 2014a). A schematic of this engineering approach is shown in Fig. 1. We aimed to improve γ-PGA production by carrying out the following five tasks: (1) block the pathways for by-product synthesis by knocking out four genes associated with bacterial polysaccharides: the eps cluster, the sac cluster, glyc (responsible for glycogen synthesis) and lps, as well as two genes that are associated with the two micromolecular products lactate and acetate, ldh and pta; (2) delete the genes pgdS and cwlO conding for γ-PGA-degrading enzymes; (3) delete the cellular autoinducer AI-2 synthetase gene luxS to make the strain tolerate environmental stress; (4) overexpress the pgsBCA genes by inserting a P₄₃ promoter upstream of the cluster; (5) use synthetic small RNAs (sRNAs) to repress the expression of rocG and glnA genes to increase the amount of intracellular glutamate. The finally obtained NK-anti-rocG strain could produce γ-PGA titers of 11.04 g/L in flask and 20.3 g/L in a 5 L fermenter, which were 2.91-fold and 5.34-fold higher than that obtained from the control NK-1 strain, respectively.