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

Metabolic Engineering ∎ (∎∎∎∎) ∎∎∎–∎∎∎

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Q1 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

Contents lists available at ScienceDirect

Metabolic Engineering journal homepage: www.elsevier.com/locate/ymben

Improved poly-γ-glutamic acid production in Bacillus amyloliquefaciens by modular pathway engineering Jun Feng a,b, Yanyan Gu a, Yufen Quan a, Mingfeng Cao c, Weixia Gao a, Wei Zhang a, Shufang Wang b,n, Chao Yang a, Cunjiang Song a,nn a

Key Laboratory of Molecular Microbiology and Technology for Ministry of Education, Nankai University, Tianjin 300071, China State Key Laboratory of Medicinal Chemical Biology, Nankai University, 94 Weijin Road, Tianjin 300071, China c Department of Chemical and Biological Engineering, Iowa State University, Ames, IA 50011, United States b

art ic l e i nf o

a b s t r a c t

Article history: Received 1 July 2015 Received in revised form 10 September 2015 Accepted 11 September 2015

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 byproducts 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 productiong of γ-PGA or of other valuable metabolites. & 2015 International Metabolic Engineering Society. Published by Elsevier Inc.

Keywords: Poly-γ-glutamic acid γ-PGA degrading-enzymes Autoinducer AI-2 Synthetic small regulatory RNAs Modular pathway engineering

1. 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 n

Corresponding author. Fax: þ 86 22 23503753. Corresponding author. Fax: þ 86 22 23503866. E-mail addresses: [email protected] (S. Wang), [email protected] (C. Song). nn

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

http://dx.doi.org/10.1016/j.ymben.2015.09.011 1096-7176/& 2015 International Metabolic Engineering Society. Published by Elsevier Inc.

Please cite this article as: Feng, J., et al., Improved poly-γ-glutamic acid production in Bacillus amyloliquefaciens by modular pathway engineering. Metab. Eng. (2015), http://dx.doi.org/10.1016/j.ymben.2015.09.011i

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87

2

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 Q3 66

J. Feng et al. / Metabolic Engineering ∎ (∎∎∎∎) ∎∎∎–∎∎∎

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 glutamateindependent 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

Fig. 1. Schematic of modular engineering approach in Bacillus amyloliquefaciens NK-1 strain. The red marks indicate the genes modified in the optimized pathway. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Please cite this article as: Feng, J., et al., Improved poly-γ-glutamic acid production in Bacillus amyloliquefaciens by modular pathway engineering. Metab. Eng. (2015), http://dx.doi.org/10.1016/j.ymben.2015.09.011i

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132

J. Feng et al. / Metabolic Engineering ∎ (∎∎∎∎) ∎∎∎–∎∎∎

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 Q4 53 54 55 56 57 58 59 60 61 62 63 64 65 66

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.

2. Materials and methods 2.1. Microorganisms, plasmids and cultivation conditions All of the strains and plasmids used in this work are listed in Table 1. The B. amyloliquefaciens NK-1 strain, a derivative of B. amyloliquefaciens LL3 strain (Cao et al., 2011) with the deletion of its endogenous plasmid and the upp gene (Feng et al., 2014a), was used as the initial strain. All of the B. amyloliquefaciens and Escherichia coli strains were grown at 37 °C in Luria-Bertani (LB) medium for routine strain construction and maintenance. E. coli DH5α was used for plasmid propagation and transformation. E. coli GM2163 was used for plasmid demethylation. For γ-PGA production, B. amyloliquefaciens was cultured in γ-PGA fermentation medium and processed following our previously reported protocols (Feng et al., 2013). P5 medium is an optimized γ-PGA fermentation medium with the concentrations of K2HPO4/KH2PO4 adjusted to160 mM/120 mM. When required, antibiotics were used at the following concentrations: 100 μg/mL ampicillin, 5 μg/mL chloramphenicol and 20 μg/mL tetracycline. The concentration of 5-fluorouracil used in this study was 100 μg/mL. 2.2. DNA manipulation and plasmid construction To construct the gene deletion vectors, the temperature-sensitive p-KSU plasmid with a upp expression cassette was used, which was derived from the pKSV7 plasmid (Smith and Youngman, 1992). The plasmids were constructed as previously reported (Feng et al., 2014a). All of the primers used in this work are listed in Table S1. To construct the P43 promoter insertion plasmid, the upstream and downstream fragments were amplified by PrimeSTAR HS DNA polymerase (Takara Bio, Japan) using primers HJ-SF/HJ-SR and HJXF/HJ-XR, respectively. The P43 promoter was amplified from the B. subtilis 168 genome using primers P43-F and P43-R. The three DNA fragments were joined by overlap-PCR. The generated fragment was digested with Sal I and BamH I, and ligated into p-KSU vector digested with the same enzymes, generating the insertion plasmid pKSV7-spx. 2.3. Strain constructions To carry out multiple gene deletions or insertions on a single strain, a marker-less gene deletion method was used to construct the gene knockout mutants (Keller et al., 2009; Zhang et al., 2014a, 2014b). The demethylated gene deletion vectors from E. coli GM2163 were treated with BamH I methyltransferase (NEB) prior to be transformed into target strains by electroporation. Cells were incubated at 42 °C for 24 h on LB agar plates with chloramphenicol. Single colonies were picked and primers N-SS/N-XX (N represents relative gene name) were used to identify the single-cross clones by PCR. The selected single-cross clones were then incubated in LB medium, supplemented with 5-fluorouracil at 42 °C for 24 h. The cells were diluted for 105 times and spread on the LB agar plates with 5-fluorouracil. Single colonies were picked and primers N-SS/ N-XX were used to identify the gene deletion clones by PCR. The P43 promoter insertion process was the same as that for gene deletion.

3

2.4. Design and expression of synthetic sRNAs We developed the two synthetic sRNAs (anti-glnA sRNA and anti-rocG sRNA) composed of three parts: a target mRNA-binding sequence, a scaffold sequence, and a transcriptional terminator sequence (Fig. 2A and B). 24bp reverse complement sequences of the glnA gene and rocG gene from the start codon (ATG) were used as the mRNA-binding sequences. A MicC sRNA scaffold was used; it was discovered in E. coli and was reported having a single target gene (Na et al., 2013). The B. subtilis spoVG gene transcription terminator was used to terminate transcription (de Hoon et al., 2005). The Hfq protein is known to be recruited by MicC sRNA and makes the sRNA more efficient at targeting the mRNA (Na et al., 2013). Thus, the hfq gene was co-expressed with the synthetic sRNAs. The promoter PBSP25 (from B. subtilis 168) with anti-glnA sRNA (Xu et al., 2013), the promoter Pupp (from B. subtilis 168) with antirocG sRNA and the promoter Phbs with the hfq gene were synthesized by GenScript (Nanjing, China) (Micka et al., 1991). The DNA fragments PBSP25-anti-glnA sRNA and Phbs-hfq were joined by overlap-PCR. The generated fragment was restricted by Sph I and Bgl II and ligated into the pWH1520 plasmid to generate the pWH1520-anti-glnA plasmid. The DNA fragments Pupp-anti-rocG sRNA and Phbs-hfq were joined by overlap-PCR and ligated into the pWH1520 plasmid with the same enzymes to generate the pWH1520-anti-rocG plasmid. The DNA fragment PBSP25-anti-glnA sRNA and Pupp-anti-rocG-Phbs-hfq were also joined by overlap-PCR and ligated into the pWH1520 plasmid to generate the pWH1520anti-glnA-rocG plasmid (Fig. 2C). The plasmids pWH1520-anti-glnA, pWH1520-anti-rocG and pWH1520-anti-glnA-rocG were firstly introduced into E. coli GM2163 and then transformed into the B. amyloliquefaciens NKE11 strain by electroporation after treatment with BamH I methyltransferase. LB agar plates with 20 μg/mL tetracycline were used to select the positive colonies. The strains harboring pWH1520-anti-glnA, pWH1520-anti-rocG and pWH1520-antiglnA-rocG plasmids were designated as B. amyloliquefaciens NKanti-glnA, B. amyloliquefaciens NK-anti-rocG and B. amyloliquefaciens NK-anti-glnA-rocG, respectively. 2.5. Production of γ-PGA by fed-batch culture in a 5 L fermenter For upscaled cultivation, 100 mL seed cultures of B. amyloliquefaciens NK-anti-rocG were cultivated in P5 medium in 500 mL shake flasks for 30 h. Three hundred milliliter seed cultures were then transferred to a 5 L fermenter system (Bailun Biological Technology Co. Ltd., China) containing 3 L of P5 medium. The pH, agitation speed and sterile air rate were maintained at 7.0, 600 rpm, and 2 vvm, respectively. Ammonia was used to adjust the pH value, moreover it served as a way for nitrogen source supplement. Sucrose was added to the setpoint value (50 g/L) at 24, 36 and 48 h, respectively. The fermentation was carried out at 37 °C for 60 h and the samples were withdrawn at 6 h intervals for further analysis. 2.6. Western blot analysis of the PgsB expression The NK-E11 and NK-E12 strains were cultured in P5 medium for 30 h. Three hundred milliliter cultures were collected, and the cells were pelleted by centrifugation (8000 rpm for 20 min). The cell pellets were washed three times with PBS buffer (pH 7.0) and then adjusted to the same optical density (OD) by PBS buffer. The cells were then broken by a sonicator (600 W for 30 min with cycles of 3 s sonication followed by 3 s pause). The broken cells were then analyzed by Western blot.

Please cite this article as: Feng, J., et al., Improved poly-γ-glutamic acid production in Bacillus amyloliquefaciens by modular pathway engineering. Metab. Eng. (2015), http://dx.doi.org/10.1016/j.ymben.2015.09.011i

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132

J. Feng et al. / Metabolic Engineering ∎ (∎∎∎∎) ∎∎∎–∎∎∎

4

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

Fig. 2. Synthetic sRNAs design and expression. (A) Mechanism of translation repression by sRNA. (B) Structures of anti-glnA sRNA (a) and anti-rocG sRNA (b). The synthetic sRNAs are composed of three parts: a target mRNA-binding sequence, a scaffold sequence and a transcriptional terminator sequence. (C) designs for synthetic sRNAs and hfq gene expression. The synthetic sRNAs were put under the promoters from B. subtilis 168 (PBSP25 and Pupp) and expressed by plasmid pWH1520. The hfq gene was also coexpressed with the synthetic sRNAs under the promoter of Phbs.

Protein samples were separated by a 12% SDS–PAGE gel. The PageRuler Prestained Protein Ladder (Thermo Scientific) was used as the protein marker in SDS–PAGE. Western blot analysis was conducted as previously described by Renart et al. (1979). The PgsB expression levels were determined by using a rabbit anti-PgsB polypeptide (IEKRRHQKNIDALPC) antibody (Genscript, Nanjing, China) as the primary antibody and the goat anti-rabbit IgG-HRP as the secondary antibody (Santa Cruz Biotechnology). Antigen/ antibody complexes were detected by chemiluminescence HRP Substrate using the Immobilon Western kit (Millipore).

2.8. Biofilm formation Pellicle formation analysis was carried out by a method modified from that previously reported (Feng et al., 2014b). The NK-1, NK-E1, NK-E2, NK-E3, NK-E4, and NK-E5 strains were cultured on LB agar plates for 18 h. Single colonies were picked and incubated in LB broth to an OD600 of 1.0. Ten microliters of each culture were introduced to 10 mL of MSgg broth (Branda et al., 2006) contained within a six-well microtiter dish. The dishes were incubated at 30 °C for 72 h without agitation, and the formed pellicles were photographed by a digital camera (Canon, Tokyo, Japan).

2.7. Enzyme assays

2.9. Analytical procedures

Glutamate dehydrogenase (GDH) activity was measured using a Glutamate Dehydrogenase Activity Colorimetric Assay Kit (BioVision, USA) following the manufacturer’s instructions. The NK-E11, NK-anti-glnA, NK-anti-rocG and NK-anti-glnA-rocG strains were harvested by centrifugation at 8000 rpm (4 °C) for 20 min after 30 h of cultivation in P5 medium. The cell pellets were washed three times with assay buffer (0.05 M Tris, 2 mM MgSO4, 2 mM DTT, 0.4 M sucrose, adjusted to pH 8.0 with HCl) and were then resuspended in 5 mL of the same buffer. The cells were then borken by sonicator on ice as described above. The broken cells were centrifugated at 6000 rpm, 4 °C for 10 min, and the supernatants were used for GDH assays. One unit (U) of GDH activity was defined as the amount of enzyme that generated one micromole of NADH per minute under assay conditions. The Glutamine synthetase (GS) activity was measured using a Glutamine Synthetase Assay Kit (Solarbio, China) following the manufacturer’s directions. The supernatants used for the GDH assay were also used to measure GS activity. One unit (U) of the GS activity was defined as the amount of enzyme that resulted in 0.01 absorbance change (at 540 nm) per minute under one milliliter assay condition.

The optical density of the assay cultures was measured using a Shimadzu UV-1800 spectrophotometer (Kyoto, Japan). γ-PGA was purified using a previously described method (Goto and Kunioka, 1992). The concentrations of acetate in the cultures were measured by high-performance liquid chromatography (HPLC) with a BDS 5 u column (Alltech, USA) and a UVIS 200 detector (Alltech, USA). (NH4)H2PO4 (0.04 mol/L, pH 2.6) was used as the mobile phase with a flow rate of 1 mL/min at 40°C. The concentrations of lactate in the cultures were measured using a lactate analyzer (SBA-40D, Biology Institute of Shandong Academy of Sciences, China). The residual concentration of sucrose was analyzed using a previously reported method (Li et al., 2003). The molecular weight of the produced γ-PGA was determined using a gel permeation chromatography (GPC) system, according to a previously described method (Cao et al. 2011). The relative viscosity of the cultures was measured by a Brookfield Digital Rheometer (model DV-I, USA) fitted with spindle code S00, at a shear rate of 2.5 rpm and at 25°C. The polysaccharide content (%) in the γ-PGA products was measured by the phenol-sulfuric acid method (DuBois et al., 1951), and the purity of γ-PGA (%) was defined as 1-polysaccharide content (%). Real-time quantitative PCR (qRT-PCR) was performed to test pgsB gene expression as previously reported (Feng et al., 2014a).

Please cite this article as: Feng, J., et al., Improved poly-γ-glutamic acid production in Bacillus amyloliquefaciens by modular pathway engineering. Metab. Eng. (2015), http://dx.doi.org/10.1016/j.ymben.2015.09.011i

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132

J. Feng et al. / Metabolic Engineering ∎ (∎∎∎∎) ∎∎∎–∎∎∎

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

3. Results and discussion 3.1. Effects of polysaccharide genes deletion on γ-PGA production Bacteria can produce wide variety of polysaccharides such as curdlan (Shih et al., 2009), alginate (Celik et al., 2008), xanthan (Kalogiannis et al., 2003), levan (Shih et al., 2010), glycogen (Kiel et al., 1994) and any others. The biosynthesis of polysaccharides needs a large amount of energy which will distribute the metabiolic flux used for target products synthesis. Moreover, most of the polysaccarides can be precipitated by ethanol, which share the similar extraction process with γ-PGA, therefore they possibly decrease the production and quality of γ-PGA. The whole genome of B. amyloliquefaciens NK-1 (a derivative of B. amyloliquefaciens LL3) strain has been sequenced (Geng et al., 2011) and many genes and clusters have been found related to the synthesis of polysaccharide such as: eps cluster (responsible EPS synthesis); sac cluster (responsible for levan synthesis); glyc gene (responsible for glycogen synthesis) and lps (responsible for lipopolisaccharide synthesis). To obtain high purity and production of γ-PGA product, we aimed to delete these genes. The eps cluster, sac cluster, glyc gene and lps gene were deleted in the NK-1 strain in sequence and the constructed strains were designated as NK-E1, NK-E2, NK-E3 and NK-E4. The fermentation results were shown in Table 2. Contrast to our prediction, ploysacchride genes deletion did not significant affect the γ-PGA production and only the NK-E1 strain showed a silightly increase in γ-PGA production. The NK-E3 and NK-E4 strains even showed γ-PGA production decrease after deletion of glyc gene. To exclude the effect of glyc gene deletion, we further deleted the lps gene in the NK-E2 strain and resulted in the NK-E5 strain. However, the NK-E5 strain still did not show favorable freature for γ-PGA production and its γ-PGA production was comparable with the NK-1 strain. Although the gene deletions had little effects on γ-PGA production, the purity of γ-PGA products significantly increased (Table 2). γ-PGA product from the NK-E5 strian exhibited the highest purity of 95.2%, which was significantly higher than that of the NK-1 strain (78.6%). From the GPC results in Fig S1, we observed that the impruties in the NK-1 strain (around 20 min) were almost disappeared in NK-E5 strain, which indicated that the low molecular weight impurities in γ-PGA products were probably polysaccharides and the deletion of polysaccharide genes blocked the polysaccharides synthesis thereafter resulted in the increase of γ-PGA purity. The viscosities of the cluture broths were also substantially increased (Table 2). The broth viscosities in NK-E1 and NK-E2 strains increased 81% and 259% after deleting the eps cluster and sac cluster compared with that in NK-1 strain (34.4 cP), respectively. After deletion the glyc gene, the broth viscosity in NK-E3 (107.2 cP) strain showed a slight decrease compared with the NKE2 strain (123.5 cP). The highest broth viscosity was observed in NK-E5 strain, which was 7.8-fold higher than that of the NK-1 strain. The viscosity in the broth is mostly affected by the γ-PGA titer, purity and molecular weight. As shown in Table 2 and Fig S1, the γ-PGA molecular weights of the six strains are all around 40 kDa and there has no significant change between these results (P 40.05). The γ-PGA molecular weights of the six strains are comparable and the γ-PGA production is not increased which confirms that the viscosity increase is mostly related to increase of γ-PGA purity. 3.2. Effects of polysaccharide genes deletion on biofilm formation Some bacterial exopolysaccharides are contributed to the biofilm formation (Dogsa et al., 2013). Branda et al. (2006) have reported that EPS and TasA are two major components of the

5

extracellular matrix in Bacillus strains, which join the cell chains together to form highly organized pellicle architecture; and the single deletion of eps cluster can result in incomplete biofilm formation. Dogsa et al. (2013) found that levan is also a part of the biofilm matrix and it contributes to biofilm robustness. We investigated the biofilm formation between NK-1, NK-E1, NK-E2, NK-E3, NK-E4 and NK-E5 strains and the results are shown in Fig S2. As expected the NK-1 strain could form a complete biofilm and the eps cluster deletion NK-E1 strain could only form incomplete biofilm. However, the NK-E2 strain with further deletion of the sac cluster could not form a pellicle. The following NK-E3, NK-E4 and NK-E5 strains deficient with EPS and levan synthesis also could not form a pellicle. These results indicate that levan is also important to biofilm formation in our strain and it combines with EPS to make the biofilm complete. 3.3. Effects of ldh and pta genes deletion on γ-PGA production Bacillus strains can produce small molecular by-products under fermentative condition. Our previous work indicates that the NK-1 strain can produce small molecular by-products such as: lactate, acetate, ethanol and butanediol, which will distribute a large amount of carbon source and energy used for target product synthesis. Furthermore, the accumulation of acidic small molecules of lactate and acetate is also toxic to cell growth (Liu et al., 2014). To improve γ-PGA production, we blocked the pathways for lactate and acetate synthesis. The ldh gene and pta gene were single deleted and double deleted in the NK-E5 strain and constructed the NK-E6 (Δldh), NK-E7 (Δpta) and NK-E8 (Δldh and Δpta) strains. The fermentation results were shown in Fig. 3. Although the defect of acetate synthesis slightly inhibited the cell growth, the NK-E7 strain showed a 11% improvement in γ-PGA production (Fig. 3a and b) and the γ-PGA purity was 94.8%. The γ-PGA molecular weights of NK-E6, NK-E7 and NK-E8 strains were 384. 4 kDa, 415.6 kDa and 401.2 kDa, respectively, and they were comparable with that of the NK-E5 strain (391.2 kDa). The deletion of pta and ldh genes has no effect on their γ-PGA molecular weights. The production of actate in these strains were also determined (Fig. 3c). The NK-E5 strain and NK-E6 strain could produce about 2.26 g/L and 1.58 g/L acetate, respectively; and the pta gene deletion NK-E7 and NK-E8 strains showed significant decrease in acetate production. However, trace amount of acetate could still be detected at end stage of fermentation. Some alternative pathways relate to acetate production contribute to its synthesis such as the aldehyde dehydrogenase can catalyze acetaldehyde to the acetate. Unexpected, there are results contrast to a previously reported work. Liu et al. (2014) had reported that the deletion of ldh gene in B. subtilis BSGN3 not only increased the DCW but also the production of N-acetylglucosamine. However, in this study, the ldh gene deletion NK-E6 and NK-E8 strains showed significant decrease in cell growth compared with NK-E5 strain (Fig. 3a); and the γ-PGA production was also significantly decreased about 81% and 40% compared with NK-E5 strain, respectively (Fig. 3b). Table 2 Comparison of γ-PGA fermentation results between NK-1 and mutant strains. Strains γ-PGA titer (g/ L) NK-1 NK-E1 NK-E2 NK-E3 NK-E4 NK-E5

3.8 7 0.11 3.94 7 0.09 3.25 7 0.03 2.647 0.31 2.96 7 0.47 3.62 7 0.18

Molecular weight (kDa)

Viscosity (cP) Purity (%)

428.6 7 27.8 414.17 27.0 377.5 7 24.4 380.9 7 17.7 376.5 7 24.3 391.2 7 27.9

34.4 7 2.5 62.3 7 4.5 123.5 7 16.8 142.3 7 11.2 107.2 7 39.3 268.77 57.0

78.6 7 4.1 83.2 7 5.3 91.5 7 2.1 94.4 7 0.6 93.9 7 1.1 95.2 7 0.9

Please cite this article as: Feng, J., et al., Improved poly-γ-glutamic acid production in Bacillus amyloliquefaciens by modular pathway engineering. Metab. Eng. (2015), http://dx.doi.org/10.1016/j.ymben.2015.09.011i

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132

6

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 Q5 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

J. Feng et al. / Metabolic Engineering ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Similar phenomenon was also reported by Romero et al. (2007), who observed cell growth defect in B. subtilis WB700 strain after disruption of the ldh gene. As the lactate is produced by the reduction of pyruvate through the LDH, which will simultaneous result in the oxidation of NADH (Cruz-Ramos et al., 2000). Thus, the inactivation of LDH probably leads to the disruption of NADH/ NAD þ redox balance thereafter inhibits cells growth. As shown in Fig. 3d, the ldh gene deletion strains could still produce lactate, which might related to existence of the alternative pathway for lactate synthesis. 3.4. Effects of cwlO and pgdS genes deletion on γ-PGA production Certain works focused on enhancing of γ-PGA production by knockout its degrading enzyme genes and many favorable results had been reported. In our previous work, the effects of γ-PGA degrading genes deletions on its production was investigated in NK-1 strain and found that the pgdS and cwlO genes double deletion NK-pc strain showed the highest 93% increase in γ-PGA production (Feng et al., 2014a). Thus, we determined to delete the two genes in the NK-E7 strain. The cwlO gene deletion NK-E9 strain showed 39.8% increased in γ-PGA titer (5.8 g/L) compared with the NK-E7 strain (4.15 g/L). The cwlO and pgdS double deletion NK-E10 strain showed 52.3% increase in γ-PGA titer (6.32 g/L) compared with the NK-E7 strain (Table S2). Although increased, γ-PGA production in NK-E10 strain was still not that high. As the increase of γ-PGA production and purity will affect the pH condition of the fermentation broth, we supposed that the low pH condition might be harmful to cell thereafter inhibited the γ-PGA production. To verify our speculation, we measured the pH values of strains NK-1, NK-E5, NK-E7, NK-E9 and NK-E10. As shown in Table S2, the pH value indeed decreased with the increase of γ-PGA production. To deal with this problem, we optimized the ratio of K2HPO4/KH2PO4 to obtain a more appropriate buffer system. The results showed that when the ratio of K2HPO4/KH2PO4 was 160 mM/120 mM, the NK-E10 strain exhibited the highest γ-PGA titer (9.18 g/L), which was 45.3% higher than that in previous γ-PGA fermentation medium. The final pH value of NK-E10 in the optimized fermentation broth (P5 medium) was 6.1, which was higher than that in the γ-PGA fermentation

medium (5.49). These results indicated that low pH in the NK-E10 fermentation broth could influence the increase of γ-PGA production and the optimized buffer system can improve the fermentation condition thereafter increase the γ-PGA production. The γ-PGA molecular weights of NK-E9 and NK-E10 strains were also determined and they were 452.7 kDa and 439.2 kDa, respectively. The molecular weights of these two strains were all higher than that of the NK-E7 strain for the control (415.6 kDa). These results were consistent with our previous report, which found that the deletion of cwlO gene would lead to γ-PGA molecular weight increase (Feng et al., 2014a). 3.5. Effects of luxS gene deletion on γ-PGA production Quorum sensing is a cell-to-cell signaling process. It permits bacteria to control their gene expression and exhibit synchronizing activities that are only productive at a high population density (Xavier and Basseler, 2005). This process accomplishes according to the small chemical signals-autoinducers (Xavier et al., 2005). It has shown that autoinducer AI-2 exhibits regulating behavior during pre-stationary growth and communicates cell density, growth rate and metabolic potential of the environment (Surette and Bassler, 1998; Surette and Bassler, 1999). Huo et al. (2011) deleted the luxS gene in E. coli YH19 strain and the generated strain could tolerate the stress caused by the increased fuel production and showed increased production of isobutanol. To determine the effects of AI-2 on γ-PGA production, the luxS gene was deleted in the NK-E10 strain. The resulting NK-E11 strain showed a 11% increase in dry cell weight and its γ-PGA titer also slightly increased from 9.18 g/L to 9.54 g/L comparable with the NK-E10 strain. Its molecular weight was 440.5 kDa, which was comparable with that of the NK-E10 strain (439.2 kDa). It seems that the behavior of autoinducer regulator AI-2 has little effect on γ-PGA production as well as its molecular weight. 3.6. Effects of pgsBCA overexpression on γ-PGA production Some strategies overexpress the genes via introducing an expression plasmid carrying the target genes. However, these overexpression strains are not stable and the plasmid might insert

Fig. 3. Fermentation results from NK-E5, NK-E6, NK-E7 and NK-E8 strains. (A) Time curves of cell growth of NK-E5 and the mutant strains. (B) γ-PGA fermentation results of NK-E5 and the mutant strains. Acetic acid (C) and lactic acid (D) production in NK-E5 and the mutant strains. Values represent means 7 SD.

Please cite this article as: Feng, J., et al., Improved poly-γ-glutamic acid production in Bacillus amyloliquefaciens by modular pathway engineering. Metab. Eng. (2015), http://dx.doi.org/10.1016/j.ymben.2015.09.011i

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132

J. Feng et al. / Metabolic Engineering ∎ (∎∎∎∎) ∎∎∎–∎∎∎

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

into the chromosome through homologous recombination. In this study, the pgsBCA genes were overexpressed by inserting a strong promoter P43 from B. subtilis 168 into the upstream of the pgs cluster and resulting the NK-E12 strain (Fig. S3a). This is the first report that studies the effect of pgsBCA overexpression on the production of γ-PGA.The transcription levels of the pgsB gene (the first gene of the pgs cluster) were measured by RT-qPCR. The results showed that the transcription level of pgsB in the NK-E12 strain were 27-fold higher than that in NK-E11 strain. Western blot results showed that the PgsB expression level was also increased in the NK-E12 strain compared the control strain (Fig. S3b). As the pgsB, pgsC, pgsA genes are controlled by the same promoter, thus we concluded that the γ-PGA synthetase (PgsBCA) was overexpressed in the NK-E12 strain. In contrast to our speculation, the increased expression of PgsBCA did not lead to the increase of γ-PGA production. The γ-PGA production decreased 15.6% compared to NK-E11 strain. Bacterial membrane is important for substance transportation, cell growth and other metabolic activity. The PgsBCA expression level is strictly regulated by the cell global performance (Ohsawa et al., 2009). The increased amount of PgsBCA in NK-12 strain might disrupt the cell balance or other membrane-associated metabolic activity thereafter result in γ-PGA production decrease. 3.7. sRNAs design and their effect on γ-PGA production Glutamate is the only precursor for γ-PGA synthesis, blocking the pathways for glutamic acid usage seems to be a way for glutamate accumulation. Glutamate in vivo can be degraded by RocG and GudB to produce 2-oxoglutarate (Belitsky and Sonenshein, 1998). RocG is enzymatically active and mostly conributes to the glutamate degradation (Zeigler et al., 2008). Glutamate can also be used to produce glutamine by GlnA for the cell to assimilate ammonium (Kloosterman et al., 2006). As glutamate plays an important role in cell growth (such as: transamination reactions) and it is also a most important intersection linking carbon to nirogen metabolism (Oh et al., 2007; Gunka and Commichau, 2012), thus it is not wise to directly delete the genes related for glutamate usage. The trans-action sRNA can be exploited for fine flux control (Na et al., 2013). Metabolic engineering of bacterial strains based on synthetic sRNAs has been successfully used in the E. coli strain for tyrosine and cadaverine production improvement (Na et al., 2013). This strategy was also been successfully used in B. subtilis for the first time for N-acetylglucosamine production by Liu’s group (2014). The sRNA related engineering strategy has many advantages. The most outstanding feature is that it can dynamically repress the target gene expression without shutting down it. Thus it can be used for the study of metabolic key genes. Therefore, we used synthetic sRNAs to repress the expression of rocG and glnA genes to decrease the competitive pathways of glutamate usage. The plasmids with sRNAs were transported into the NK-E11 strain (Fig. 2) and the fermentation results are shown in Fig. 4. The GDH activities from NK-anti-rocG and NK-anti-glnA-rocG strains were repressed to 0.04 mU/mg and 0.10 mU/mg, which were about 6.1% and 14.3% of the control NK-E11 strain (Fig. 4a). The GS activities from NK-anti-glnA and NK-anti-glnA-rocG strains were repressed to 0.39 U/mg and 0.38 U/mg, which were about 74.5% and 72.6% of the control NK-E11 strain (Fig. 4b). The gene repression efficiencies in anti-glnA sRNA and anti-rocG sRNA coexpression strain were comparable with the strains with only one sRNA was expressed. This indicated that introduction of two copies of sRNAs did not exert metabolic burden on the host cells. Although anti-glnA sRNA and anti-rocG sRNA have the same MicC scaffold, their repression efficiencies were different. The repression efficiency of anti-glnA sRNA was lower than that of anti-

7

rocG sRNA and previously reproted sRNA repression efficiency in B. subtilis (40%) (Liu et al., 2014). These datas indicate that the MicC derivative sRNA can function in Bacillus species, however, its repression efficiency seems unstable like it is in E. coli (90%) (Na et al., 2013). A more systematic work on the relationship between snythetic sRNA structure and the target gene repression efficiency must be studied in the future. We further investigated the sRNAs effects on γ-PGA production. As shown in Fig. 4c. The anti-glnA sRNA expression strains NKanti-glnA and NK-anti-glnA-rocG showed γ-PGA titers decreased of 55.9% and 44.3% to 4.11 g/L and 5.19 g/L, respectively, compared with the control NK-E11 strain (9.32 g/L). The anti-rocG sRNA expression strain NK-anti-rocG showed γ-PGA titer increased about 18.5% to 11.04 g/L compared with the control NK-E11 strain. These results indicated that the full function of GS was important for γ-PGA production. Commichau and Stülke (2008) pointed out that GS is a trigger enzyme that is active in metabolism and gene expression controlling; thus we hypothesize that the inhibition of glnA gene expression would disrupt the normal metabolic of bacteria thereafter lead to the decrease of γ-PGA production in the NK-anti-glnA and NK-anti-glnA-rocG strains. Zhang et al. (2014a, 2014b) have demonstrated that the rocG gene deletion has no effect on γ-PGA production. However, in this study the inhibition of rocG gene expression resulted in the increase of γ-PGA production. This difference must benefit to the partially expression of rocG gene. Previous works demonstrated that the cryptic gudB gene will be rapidly activated and produce enzymatically active GDH in the background of rocG gene mutant (Belitsky and Sonenshein, 1998; Gunka et al., 2012). We speculate that the rocG gene deletion probably induces the expression of GudB to maintain the GDH activity in the bacterial cell, therefore its deletion has little effect on the amount of intracellular glutamate and the production of γ-PGA; however, in the NK-anti-rocG strain, the low expression level of rocG makes sure the gudB gene remains inactivated and the low activity of GDH reduces the degradation of glutamate thereafter increase the production of γ-PGA. The γ-PGA molecular weights of NK-anti-glnA, NK-anti-rocG and NK-antiglnA-rocG strain were 452.9 kDa, 440.1 kDa and 443.0 kDa, respectively, which was comparable with that of the control NKE11 strain (440.5 kDa). It seems that the amount of intracellular glutamate precursor does not affect the molecular weight of γ-PGA. 3.8. Fed-batch culture for γ-PGA production by B. amyloliquefaciens NK-anti-rocG strain To evaluate the performance of the γ-PGA overproducing strain NK-anti-rocG in a more stable and comfortable condition, fedbatch culture was carried out in a 5 L up-scaled system. The carbon source (sucrose) was supplemented to the initial concentration (50 g/L) at 24, 36 and 48 h, respectively. And the ammonia was used for pH modulation as well as for nitrogen source supplement. As shown in Fig. 4d, the cells enterred to stationary phase after 36 h cultivation and the γ-PGA production increased with the cultivation process. The highest titer was obtained at 54 h and the NK-anti-rocG strain could produce 20.3 g/L γ-PGA, which was 5.34-fold higher than that obtained from NK-1 strain in flask. The γ-PGA fermentation broth is sticky and oxygen is always the limiting factor at the end stage of fermentation. In order to improve this condition, we chose the maximum sterile air rate of our 5 L fermenter (2 vvm) in this work. However, the DO still rapidly decreased and reached to 1-2% after 24 h of cultivation. The γ-PGA production will be further enhanced if DO limitation is solved. Providing higher rate of sterile air or heterologous expression of the vgb gene (encodes Vitreoscilla hemoglobin) in the mutant strain might be two effective ways to solve this

Please cite this article as: Feng, J., et al., Improved poly-γ-glutamic acid production in Bacillus amyloliquefaciens by modular pathway engineering. Metab. Eng. (2015), http://dx.doi.org/10.1016/j.ymben.2015.09.011i

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132

J. Feng et al. / Metabolic Engineering ∎ (∎∎∎∎) ∎∎∎–∎∎∎

8

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 Q7 Q6 63 64 65 66

Fig. 4. The GDH activities (A) and GS activities (B) in NK-E11 and the mutant strains after 30 h cultivation in P5 medium. γ-PGA fermentation results of NK-E11 and the mutant strains (C). Time curves of process parameters in a 5 L fermenter of the NK-anti-rocG strain (D). Values represent means7 SD.

problem in the future (Su et al., 2010). Its γ-PGA purity slightly decresed to 90.1% and this probably because the higher γ-PGA yield adhered cells or impurities in the medium therefore decreased the γ-PGA purity. Its γ-PGA molecular weight was 446.7 kDa, which is comparable with that obtained from the flask culture (440.1 kDa).

4. Conclusion In this study, we improved the γ-PGA production in a glutamate-independent Bacillus amyloliquefaciens NK-1 strain from the following mentioned aspects: block the byproduct synthetic patheways, delete the γ-PGA degrading enzyme genes, block the cell autoinducer synthetic pathway and inhibit the usage of the γ-PGA synthetic precusor. The finally obtained NK-anti-rocG strain could produce 11.04 g/L and 20.3 g/L γ-PGA in flask and in a 5 L fermenter, respectively, which was 2.91-fold and 5.34-fold higher than that from the NK-1 strain in flask. This study firstly exhibited a systematic modular pathway engineering method for γ-PGA production improvement; moreover, we showed that the synthetic sRNAs originated from E. coli would be a atternative way for metabolic engineering in B. amyloliquefaciens. Although the γPGA production in NK-anti-rocG strain is significantly increased, it is still too low to meet the industrial demand. More works must be done to improve the γ-PGA production in the future.

Acknowledgments This work was supported by National key Basic Research Program of China (“973”-Program) 2012CB725204, National High Technology Research and Development Program of China (“863”Program) 2012AA021505, National Natural Science Foundation of

China, China Grant Nos. 31070039, 31170030 and 51073081, Project of Tianjin, China (13JCZDJC27800 and 13JCYBJC24900) and the Ph.D. Candidate Research Innovation Fund of Nankai University.

Appendix A. Supplementary material Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.ymben.2015.09.011.

References Ashiuchi, M., Misono, H., 2002. Biochemistry and molecular genetics of poly-γglutamate synthesis. Appl. Biochem. Biotechnol. 59, 9–14. Belitsky, B.R., Sonenshein, A.L., 1998. Role and regulation of Bacillus subtilis glutamate dehydrogenase genes. J. Bacteriol. 180, 6298–6305. Bisicchia, P., Noone, D., Lioliou, E., Howell, A., Quigley, S., Jensen, T., Jarmer, H., Devine, K.M., 2007. The essential YycFG two-component system controls cell wall metabolism in Bacillus subtilis. Mol. Microbiol. 65, 180–200. Branda, S.S., Chu, F., Kearns, D.B., Loslck, R., Kolter, R.A., 2006. Major protein component of the Bacillus subtilis biofilm matrix. Mol. Microbiol. 59, 1229–1238. Candela, T., Fouet, A., 2006. Poly-gamma-glutamate in bacteria. Mol. Microbiol. 60, 1091–1098. Cao, M.F., Geng, W.T., Liu, L., Song, C.J., Xie, H., Guo, W.B., Jin, Y.H., Wang, S.F., 2011. Glutamic acid independent production of poly-γ-glutamic acid by Bacillus amyloliquefaciens LL3 and cloning of pgsBCA genes. Bioresour. Technol. 102, 4251–4257. Celik, G.Y., Aslim, B., Beyatli, Y., 2008. Characterization and production of the exopolysaccharide (EPS) from Pseudomonas aeruginosa G1 and Pseudomonas putida G12 strains. Carbohydr. Polym. 73, 178–182. Commichau, F.M., Stülke, J., 2008. Trigger enzymes: bifunctional proteins active in metabolism and in controlling gene expression. Mol. Microbiol. 67, 692–702. Cruz-Ramos, H., Hoffmann, T., Marino, M., Nedjari, H., Presecan-Siedel, E., Dreesen, O., Glaser, P., Jahn, D., 2000. Fermentative metabolism of Bacillus subtilis: physiology and regulation of gene expression. J. Bacteriol. 182, 3072–3080. de Hoon, M.J., Makita, Y., Nakai, K., Miyano, S., 2005. Prediction of transcriptional terminators in Bacillus subtilis and related species. PLoS Comput. Biol. 1, e25. Dogsa, I., Brloznik, M., Stopar, D., Mandic-Mulec, I., 2013. Exopolymer diversity and the role of levan in Bacillus subtilis biofilms. PLoS One 8, e62044.

Please cite this article as: Feng, J., et al., Improved poly-γ-glutamic acid production in Bacillus amyloliquefaciens by modular pathway engineering. Metab. Eng. (2015), http://dx.doi.org/10.1016/j.ymben.2015.09.011i

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132

J. Feng et al. / Metabolic Engineering ∎ (∎∎∎∎) ∎∎∎–∎∎∎

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

DuBois, M., Gilles, K.A., Hamilton, J.K., Rebers, P.A., Smith, F., 1951. Colorimetric method for determination of sugars and related substances. Anal. Chem. 28, 350–356. Feng, J., Gu, Y.Y., Wang, J.Q., Song, C.J., Yang, C., Xie, H., Zhang, W., Wang, S.F., 2013. Curing the plasmid pMC1 from the poly(γ-glutamic acid) producing Bacillus amyloliquefaciens LL3 strain using plasmid incompatibility. Appl. Biochem. Biotechnol. 171, 532–542. Feng, J., Gao, W.X., Gu, Y.Y., Zhang, W., Cao, M.F., Song, C.J., Zhang, P., Sun, M., Yang, C., Wang, S.F., 2014a. Functions of poly-gamma-glutamic acid (γ-PGA) degradation genes in γ-PGA synthesis and cell morphology maintenance. Appl. Microbiol. Biotechnol. 98, 6397–6407. Feng, J., Gu, Y.Y., Sun, Y., Han, L.F., Yang, C., Zhang, W., Cao, M.F., Song, C.J., Gao, W.X., Wang, S.F., 2014b. Metabolic engineering of Bacillus amyloliquefaciens for polygamma-glutamic acid (γ-PGA) overproduction. Microb. Biotechnol. 7, 446–455. Geng, W.T., Cao, M.F., Song, C.J., Xie, H., Liu, L., Yang, C., Feng, J., Zhang, W., Jin, Y.H., Du, Y., Wang, S.F., 2011. Complete genome sequence of Bacillus amyloliquefaciens LL3, which exhibits glutamic acid-independent production of poly-γglutamic acid. J. Bacteriol. 193, 3393–3394. Goto, A., Kunioka, M., 1992. Biosynthesis and hydrolysis of poly(γ-glutamic acid) from Bacillus subtilis IFO3335. Biosci. Biotechnol. Biochem. 56, 1031–1035. Gunka, K., Tholen, S., Gerwig, J., Herzberg, C., Stülke, J., Commichau, F.M., 2012. A high-frequency mutation in Bacillus subtilis: requirements for the decryptification of the gudB glutamate dehydrogenase gene. J. Bacteriol. 1036, 1036–1044. Gunka, K., Commichau, F.M., 2012. Control of glutamate homeostasis in Bacillus subtilis: a complex interplay between ammonium assimilation, glutamate biosynthesis and degradation. Mol. Microbiol. 85, 213–224. Huo, Y.X., Cho, K.M., Rivera, J.G.L., Monte, E., Shen, C.R., Yan, Y.J., Liao, J.C., 2011. Conversion of proteins into biofuels by engineering nitrogen flux. Nat. Biotechnol. 29, 346–351. Kalogiannis, S., Iakovidou, G., Liakopoulou-Kyriakides, M., Kyriakidis, D.A., Skaracis, G.N., 2003. Optimization of xanthan gum production by Xanthomonas campestris grown in molasses. Process Biochem. 39, 249–256. Keller, K.L., Bender, K.S., Wall, J.D., 2009. Development of a markerless genetic exchange system for Desulfovibrio vulgaris hildenborough and its use in generating a strain with increased transformation efficiency. Appl. Environ. Microb. 75, 7682–7691. Kloosterman, T.G., Hendriksen, W.T., Bijlsma, J.J., Bootsma, H.J., van Hijum, S.A., Kok, J., Hermans, P.W., Kuipers, O.P., 2006. Regulation of glutamine and glutamate metabolism by GlnR and GlnA in Streptococcus pneumoniae. J. Biol. Chem. 281, 25097–25109. Kiel, J.A., Boels, J.M., Beldman, G., Venema, G., 1994. Glycogen in Bacillus subtilis: molecular characterization of an oeperon encoding enzymes involved in glycogen biosynthesis and degradation. Mol. Microbiol. 11, 203–218. Kunioka, M., 1997. Biosynthesis and chemical reactions of poly(amino acid)s from microorganisms. Appl. Microbiol. Biotechnol. 47, 469–475. Li, C., 2002. Poly(L-glutamic acid)–anticancer drug conjugates. Adv. Drug Deliv. Rev. 54, 695–713. Li, L.J., Kong, H.X., Wu, S.H., 2003. A new UV-Spectrophotometric method for the determination of sucrose. J. Anal. Sci. 19, 367–369. Liang, H.F., Chen, C.T., Chen, S.C., Kulkarni, A.R., Chiu, Y.L., Chen, M.C., Sung, H.W., 2006. Paclitaxel-loaded poly(γ-glutamic acid)-poly(lactide) nanoparticles as a targeted drug delivery system for the treatment of liver cancer. Biomaterials 27, 2051–2059. Liu, J., Ma, X., Wang, Y., Liu, F., Qia, J.Q., Li, X.Z., Gao, X.W., Zhou, T., 2011. Depressed biofilm production in Bacillus amyloliquefaciens C06 causes γ-polyglutamic acid (γ-PGA) overproduction. Curr. Microbiol. 62, 235–241. Liu, Y.F., Zhu, Y.Q., Li, J.H., Shin, H.D., Chen, R.R., Du, G.C., Liu, L., Chen, J., 2014. Modular pathway engineering of Bacillus subtilis for improved Nacetylglucosamine production. Metab. Eng. 23, 42–52. Micka, B., Groch, N., Heinemann, U., Marahiel, M.A., 1991. Molecular cloning, nucleotide sequence, and characterization of the Bacillus subtilis gene encoding the DNA-binding protein HBsu. J. Bacteriol. 173, 3191–3198. Mitsui, N., Murasawa, H., Sekiguchi, J., 2011. Disruption of the cell wall lytic enzyme CwlO affects the amount and molecular size of poly-γ-glutamic acid produced by Bacillus subtilis (natto). J. Gen. Appl. Microbiol. 57, 35–43.

9

Na, D., Yoo, S.M., Chung, H., Park, H., Park, J.H., Lee, S.Y., 2013. Metabolic engineering of Escherichia coli using synthetic small regulatory RNAs. Nat. Biotechnol. 31, 170–174. Oh, Y.K., Palsson, B., Park, S.M., Schilling, C.H., Mahadevan, R., 2007. Genome-scale reconstruction of metabolic network in Bacillus subtilis based on high throughput phenotyping and gene essentiality data. J. Biol. Chem. 282, 28791–28799. Ohsawa, T., Tsukahara, K., Ogura, M., 2009. Bacillus subtilis response regulator DegU is a direct activator of pgsB transcription involve in γ-poly-glutamic acid synthesis. Biosci. Biotechnol. Biochem. 73, 2096–2102. Park, Y.J., Liang, J., Yang, Z., Yang, V.C., 2001. Controlled release of clot-dissolving tissue-type plasmmogen activator from a poly(L-glutamic acid) semiinterpenetrating polymer network hydrogel. J. Control. Release 74, 243–247. Renart, J., Reiser, J., Stark, G.R., 1979. Transfer of proteins from gels to diazobenzyloxymethyl-paper and detection with antisera: a method for studying antibody specificity and antigen structure. Proc. Natl. Acad. Sci. USA 76, 3116–3120. Richard, A., Margaritis, A., 2001. Poly (glutamic acid) for biomedical applications. Crit. Rev. Biotechnol. 21, 219–232. Romero, S., Merino, E., Bolívar, F., Gosset, G., Martinez, A., 2007. Metabolic Engineering of Bacillus subtilis for ethanol production: lactate dehydrogenase plays a key role in fermentative metabolism. Appl. Environ. Microbiol. 73, 5190–5198. Scoffone, V., Dondi, D., Biino, G., Borghese, G., Pasini, D., Galizzi, A., Calvio, C., 2013. Knockout of pgdS and ggt genes improves γ-PGA yield in B. subtilis. Biotechnol. Bioeng. 110, 2006–2012. Shih, I.L., Van, Y.T., 2001. The production of poly(γ-glutamic acid) from microorganism and its various applications. Bioresour. Technol. 79, 207–225. Shih, I.L., Yu, J.Y., Hsieh, C., Wu, J.Y., 2009. Production and characterization of curdlan by Agrobacterium sp. Biochem. Eng. J. 43, 33–40. Shih, I.L., Chen, L.D., Wu, J.Y., 2010. Levan production using Bacillus subtilis natto cells immobilized on alginate. Carbohydr. Polym. 82, 111–117. Smith, K., Youngman, P., 1992. Use a new integrational vector to investigate compartment-specific expression of the Bacillus subtilis spoIIM gene. Biochimie 74, 705–711. Smith, T.J., Blackman, S.A., Foster, S.J., 2000. Autolysins of Bacillus subtilis: multiple enzymes with multiple functions. Microbiology 146, 249–262. Surette, M.G., Bassler, B.L., 1998. Quorum sensing in Escherichia coli and Salmonella typhimurium. Proc. Natl. Acad. Sci. USA 95, 7046–7050. Surette, M.G., Bassler, B.L., 1999. Regulation of autoinducer production in Salmonella typhimurium. Mol. Microbiol. 31, 585–595. Su, Y.S., Li, X., Liu, Q.Z., Hou, Z.W., Zhu, X.Q., Guo, X.P., Ling, P.X., 2010. Improved poly-γ-glutamic acid production by chromosomal integration of the Vitreoscilla hemoglobin gene (vgb) in Bacillus subtilis. Bioresour. Technol. 101, 4733–4736. Xavier, K.B., Basseler, B.L., 2005. Regulation of uptake and processing of the quorum-sensing autoinducer AI-2 in Escherichia coli. J. Bacteriol. 187, 238–248. Xu, D.Q., Li, R.K., Lin, J., Yan, F., Ye, X.Y., 2013. Clone and functional verification of promoters from Bacillus subtilis. J. Fuzhou Univ. 41, 391–396. Zeigler, D.R., Prágai, Z., Rodriguez, S., Chevreux, B., Muffler, A., Albert, T., Bai, R., Wyss, M., Perkins, J.B., 2008. The origins of 168, W23, and other Bacillus subtilis legacy strains. J. Bacteriol. 190, 6983–6995. Zhang, D., Feng, X.H., Zhou, Z., Zhang, Y., Xu, H., 2012. Economical production of poly(γ-glutamic acid) using untreated cane molasses and monosodium glutamate waste liquor by Bacillus subtilis NX-2. Bioresour. Technol. 114, 583–588. Zhang, W., Xie, H., He, Y., Feng, J., Gao, W.X., Gu, Y.Y., Wang, S.F., Song, C.J., 2013. Chromosome integration of the Vitreoscilla hemoglobin gene (vgb) mediated by temperature-sensitive plasmid enhances γ-PGA production in Bacillus amyloliquefaciens. FEMS Microbiol. Lett. 343, 127–134. Zhang, W., Gao, W.X., Feng, J., Zhang, C., He, Y.L., Cao, M.F., Li, Q., Sun, Y., Yang, C., Song, C.J., Wang, S.F., 2014a. A markerless gene replacement method for B. amyloliquefaciens LL3 and its use in genome reduction and improvement of poly-γ-glutamic acid production. Appl. Microbiol. Biotechnol. 98, 8963–8973. Zhang, W., He, Y.L., Gao, W.X., Feng, J., Cao, M.F., Yang, C., Song, C.J., Wang, S.F., 2014b. Deletion of genes involved in glutamate metabolism to improve polygamma-glutamic acid production in B. amyloliquefaciens LL3. J. Ind. Microbiol. Biotechnol. 42, 297–305.

Please cite this article as: Feng, J., et al., Improved poly-γ-glutamic acid production in Bacillus amyloliquefaciens by modular pathway engineering. Metab. Eng. (2015), http://dx.doi.org/10.1016/j.ymben.2015.09.011i

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132