Genetic engineering of Pseudomonas chlororaphis GP72 for the enhanced production of 2-Hydroxyphenazine

Background The biocontrol strain Pseudomonas chlororaphis GP72 isolated from the green pepper rhizosphere synthesizes three antifungal phenazine compounds, 2-Hydroxyphenazine (2-OH-PHZ), 2-hydroxy-phenazine-1-carboxylic acid (2-OH-PCA) and phenazine-1-carboxylic acid (PCA). PCA has been a commercialized antifungal pesticide registered as “Shenqinmycin” in China since 2011. It is found that 2-OH-PHZ shows stronger fungistatic and bacteriostatic activity to some pathogens than PCA. 2-OH-PHZ could be developed as a potential antifungal pesticide. But the yield of 2-OH-PHZ generally is quite low, such as P. chlororaphis GP72, the production of 2-OH-PHZ by the wide-type strain is only 4.5 mg/L, it is necessary to enhance the yield of 2-OH-PHZ for its application in agriculture. Results Different strategies were used to improve the yield of 2-OH-PHZ: knocking out the negative regulatory genes, enhancing the shikimate pathway, deleting the competing pathways of 2-OH-PHZ synthesis based on chorismate, and improving the activity of PhzO which catalyzes the conversion of PCA to 2-OH-PHZ, although the last two strategies did not give us satisfactory results. In this study, four negative regulatory genes (pykF, rpeA, rsmE and lon) were firstly knocked out of the strain GP72 genome stepwise. The yield of 2-OH-PHZ improved more than 60 folds and increased from 4.5 to about 300 mg/L. Then six key genes (ppsA, tktA, phzC, aroB, aroD and aroE) selected from the gluconeogenesis, pentose phosphate and shikimate pathways which used to enhance the shikimate pathway were overexpressed to improve the production of 2-OH-PHZ. At last a genetically engineered strain that increased the 2-OH-PHZ production by 99-fold to 450.4 mg/L was obtained. Conclusions The 2-OH-PHZ production of P. chlororaphis GP72 was greatly improved through disruption of four negative regulatory genes and overexpression of six key genes, and it is shown that P. chlororaphis GP72 could be modified as a potential cell factory to produce 2-OH-PHZ and other phenazine biopesticides by genetic and metabolic engineering. Electronic supplementary material The online version of this article (doi:10.1186/s12934-016-0529-0) contains supplementary material, which is available to authorized users.


Open Access
Microbial Cell Factories is based on its effectiveness against specific phytopathogens and minimal toxicity toward humans, animals, and the environment [4]. In previous studies, however, 2-OH-PHZ showed stronger fungistatic and bacteriostatic activity than PCA toward some pathogens, such as Gaeumannomyces. graminis var. tritici which cause the take-all disease of wheat [5]. The take-all disease is one of the most important root diseases of wheat worldwide, and no resistant cultivars or effective chemical fungicides are available, it is important to develop biological agents to control the disease [6][7][8][9]. Therefore, it is necessary to enhance the yield of 2-OH-PHZ for its application in agriculture. Although it may be practical to synthesize phenazines by chemical methods, the yield is low [10], and toxic byproducts such as lead oxide, aniline, o-phenylenediamine, and azobenzoate are produced [11]. So the biocatalytic synthesis of 2-OH-PHZ using recombinant microorganisms provides an attractive alternative. Until now, 2-OH-PHZ has been produced primarily by P. chlororaphis with the exception of the strain P. aurantiaca PB-St2 [12]. P. chlororaphis GP72 is a 2-OH-PHZ producing strain, although the production by the wild type is only 4.5 mg/L [1], it is possible to improve the 2-OH-PHZ production largely for its industrial production by genetic engineering and regulation.
The mechanism of the phenazine biosynthesis pathway has already been elucidated. A gene cluster phzABC-DEFG facilitates phenazine biosynthesis in Pseudomonas [13]. The enzymes coded by phzABCDEFG convert chorismate, the end product of the shikimate pathway, into PCA, which is the synthesis substrate of the phenazine derivatives [13,14]. In recent years, many studies on the production of shikimic acid (SA), a very important intermediate of the shikimate pathway, which is used as substrate in the synthesis of oseltamivir phosphate (Tamiflu) by chemical methods, have been reported [15]. To enhance the production of SA in recombinant microorganisms, several strategies have been developed to enhance the shikimate pathway. The principal purpose of these strategies is to increase the availability of phosphoenolpyruvate (PEP) and erythrose 4-phosphate (E4P), the direct precursors of the shikimate pathway, through genetic alterations that redistribut the metabolic fluxes in the central metabolism [16]. Other methods include improved channeling of carbon toward SA through overexpression of shikimate dehydrogenase, transketolase, DHQ synthase, feedback-resistant DAHP synthases, and DHQ dehydrogenase (coded by aroE, tktA, aroB, aroF-GH fbr and aroD) [17]. Besides that, strains lacking the pyruvate kinase (coded by pykA and pykF) and strains overexpressing PEP synthase (coded by ppsA) have also been evaluated as means of increasing the intracellular availability of PEP [18][19][20]. All of these strategies may be used in enhancing the production of phenazines for the shikimate pathway was necessary in synthesis of phenazines.
In addition, in Pseudomonas protegens, the Lon protease, which is an ATP-dependent protease, reduces the stability of the GacA protein and stymies the expression of the Gac/Rsm signal transduction system [21]. The GacS/GacA two-component system, which is conserved in many Gram-negative bacterias, stimulates production of phenazines and other secondary metabolites in P. chlororaphis 30-84 [22]. According to Wang et al. [22], overexpression of the gene rsmE, one of the RNA-binding proteins, resulted in decreased phenazine production in P. chlororaphis . Also, in our previous research, a negative regulatory gene rpeA was insertionally mutated to construct strain GP72AN, which resulted in a fivefold increase of 2-OH-PHZ (24.6 mg/L). All of these negative regulators may be used to enhance the production of phenazines.
The aim of this study was to construct a genetically engineered P. chlororaphis GP72 strain to significantly increase 2-OH-PHZ production. Two aspects were investigated to achieve this goal: (1) the stepwise disruption of four negative regulatory genes (pykF, rpeA, rsmE and lon), and (2) the overexpression of six key genes (ppsA, tktA, phzC, aroB, aroD and aroE) selected from different pathways (Fig. 1). This work provides an efficient way for enhancing 2-OH-PHZ production.

Disruption of four negative regulatory genes to enhance 2-OH-PHZ production
In order to obtain a high yield of 2-OH-PHZ, we firstly chose to inactivate the gene pykF of GP72, and obtained the mutant strain GP72Δpyk (Figs. 2, 3). After fermentation and analyses by HPLC, the 2-OH-PHZ production of GP72Δpyk had a 6.7-fold increase compared to that of GP72, it increased from 4.5 to 35 mg/L (Figs. 2a, 3). And our results suggest that deletion of this gene had little effect on bacterial growth (Fig. 2c).
In this study, the strain GP72Δpyk was constructed with the non-scar deletion method (Additional file 1: Fig.  S1). This is different from the insertional mutagenesis method that constructed GP72AN. GP72AN is a chromosomally inactivated GP72 mutant which was modified by inserting the resistance cassettes of gentamycin as a selective marker [1]. But insertional mutagenesis limited the multiplicity of genetic modification: P. chlororaphis GP72 is sensitive to only a few antibiotics [2]. Obviously, insertional mutagenesis could not be applied to inactivate the regulatory regions of GP72. Therefore, in this study, all of the mutant strains were constructed using the nonscar deletion strategy.
In our previous research, a negative regulator RpeA (coded by rpeA) was insertionally mutated to construct GP72AN, which resulted in a fivefold increase of 2-OH-PHZ [1]. In this work, we also knocked out rpeA from the GP72Δpyk genome to construct GP72ND-1 with the non-scar deletion strategy. Similar to the result of insertional mutagenesis, the 2-OH-PHZ production of GP72ND-1 had a 5.8-fold increase and reached 239.8 mg/L (Figs. 2a, 3).
Most of bacterial intracellular proteolysis is initiated by four energy-dependent proteases: FtsH, HslUV, Lon, and the Clp family. Lon is key to over half of the energydependent proteolysis in E. coli [27]. According to a previous report, in Pseudomonas protegens, the Lon protease can impair the stability of GacA protein and reduce the expression of Gac/Rsm signal transduction pathway [21]. A lon gene knockout mutant has been previously shown to increase the production of antibiotics in P. protegens Pf-5 [28]. Importantly, the Gac/Rsm system is conserved in many Gram-negative bacterias, and it activates secondary metabolite production, including production of phenazines in P. chlororaphis 30-84 [22]. Thus, lon may participate in the regulation of phenazine synthesis. The gene lon was also found in P. chlororaphis GP72. We obtained the mutant strain GP72Δlon after the gene lon was knocked out of GP72, and the yield of 2-OH-PHZ had a 6.0-fold increase (from 4.5 to 31.7 mg/L) (Fig. 2a). In order to gain more 2-OH-PHZ, the gene lon was also knocked out from GP72ND-2 to construct GP72ND-3. After fermentation, the yield of 2-OH-PHZ increased from 273.5 to 300.5 mg/L (Figs. 2a and 3).

Enhanced 2-OH-PHZ production by overexpressing six key genes
Our results proved that the disruption of negative genes is a valid strategy for enhancing phenazines production. Compared with the disruption of negative genes, the overexpression of key genes is another effective strategy used frequently to enhance the production of biological products. As mentioned before, the disruption of pykF improved the production of 2-OH-PHZ by diverting Fig. 2 Growth curves, 2-OH-PHZ and PHZ production of the GP72 mutant derivative strains. a 2-OH-PHZ production. b Phenazine production. c Growth curves. The error bars indicate standard deviations from triplicate experiments more metabolic flux into the shikimate pathway. This suggested that enhancing the shikimate pathway may be a valid strategy to enhance the production of phenazines in GP72. In order to enhance the shikimate pathway of P. chlororaphis GP72, six key genes (ppsA, tktA, phzC, aroB, aroD and aroE) selected from the gluconeogenesis, pentose phosphate and shikimate pathways were chosen to be overexpressed. Because up to six genes would be simultaneously overexpressed, only a few kinds of antibiotics would be sensitive towards P. chlororaphis GP72 and be chosed to use [2]. To solve this problem, we chose a kind of modular vector-the BglBrick plasmid which has been recently widely used [19]. Firstly, six genes were constrcted into pBbB5K which is a BglBric plasmid. Then six genes were constructed into pbBb5K step by step, to build a fusion plasmid following the modular principle (Additional file 2: Fig. S2). To test the effect of overexpression, these pBbB5K plasmids were transformed into the strain GP72ND-3 by elec-transformation. The fermentation results showed that when single genes were overexpressed, the effect was not obvious (Fig. 4a, b). But when two or more adjacent genes were expressed at the same time, not only did 2-OH-PHZ production increase, but the production of other phenazines (such as PCA, 2-OH-PCA) also increased (Fig. 4c, d). When the fusion plasmid pBbB5K-aroE-aroD-aroB-phzC-tktA-ppsA was overexpressed in GP72ND-3, 2-OH-PHZ production of 450.4 mg/L (which increased 99-fold more than that of wide type GP72) and phenazine derivatives production of 1520 mg/L was obtained in the fermentation experiments (Figs. 3, 4). This result suggested that the overexpression of the corresponding genes was effective for enhancing 2-OH-PHZ production.

Discussion
2-OH-PHZ, produced mainly by P. chlororaphis, is a potential antifungal pesticide, which has a broad application prospect. P. chlororaphis GP72 could produce 2-OH-PHZ, but its yield is only 4.5 mg/L. Therefore, it is necessary to enhance the 2-OH-PHZ production of GP72 for its application in agriculture. Here we tried four different strategies to improve the yield of 2-OH-PHZ, knocking out the negative regulatory genes, enhancing the shikimate pathway, deleting the competing pathways based on chorismate, and improving the activity of PhzO, although the last two strategies did not give us satisfactory results.
In P. chlororaphis, the biosynthesis of phenazines has been traced to the shikimate pathway. This pathway begins with DAHP-synthase-mediated condensation of When PEP and E4P accumulate, it may channel the metabolic flow to biosynthetic routes of shikimate and chorismate, which are key metabolic precursors of phenazines. PEP is a key intermediate of the central metabolism, it serves as a precursor in several biosynthetic pathways [29]. Pyruvate synthesis from PEP catalyzed by pyruvate kinase (coded by pyk) is one of the main PEP consuming reactions [20]. Therefore weakening or blocking the conversion of PEP to pyruvate is a key strategy to enhance PEP availability (Fig. 1). According to previous research in E. coli, two pyruvate kinases were detected, which were coded by pykA and pykF respectively [30]. In contrast to the disruption of both pyk genes, a single inactivation of either the pykA or pykF genes may increase the quantity of PEP available for DAHP synthesis. This would not impede synthesis of pyruvate or the flux of pyruvate to acetyl-CoA [18]. It has been shown in E. coli that both pykA and pykF have active roles in pyruvate biosynthesis and PykF displayed the higher enzyme activity when compared to PykA in the wild type strain [30]. Therefore, PykF is thought to produce the main Pyk activity in E. coli [20]. Similar to E. coli, P. chlororaphis GP72 also contains two pyruvate kinases, which were coded by pykA and pykF. In order to increase the availability of the PEP, the gene pykF of GP72 was selected to knock out. And the 2-OH-PHZ production had a 6.7-fold increase (Figs. 2a, 3). TCST systems help prokaryotes interact with their environments through both sensing and response, coordinating many of cellular pathways. Pseudomonas species contain a large number of TCST systems, for example, there are 91 in P. fluorescens Pf-5, 127 in P. aeruginosa PAO1, and 51 in P. chlororaphis 30-84 [24]. In P. chlororaphis GP72, a large number of TCST systems also have been discovered. Huang et al. [1] have reported that RpeA negatively controlled the production of 2-OH-PHZ in P. chlororaphis GP72. In this study, three genes (rpeA, rsmE and lon), which are all related to the TCST systems, were knocked out of the strain GP72Δpyk genome stepwise. After delection of these three genes, the 2-OH-PHZ production increased from 35 to 300.5 mg/L. According to previous reports on P. chlororaphis 30-84, P. chlororaphis PCL1391 and our results, we propose a simple regulatory model of GP72 to illustrate the role of phenazine regulators modulated by TCST systems [1,22,24,31] (Fig. 5). In this simple model, GP72 is similar to the strains P. chlororaphis PCL1391 and 30-84, Pip promotes the production of phenazine by increasing the expression of phzI and phzR. The expression of Pip is regulated by the sigma factor rpoS, which is itself regulated by GacA, and by the RpeA/RpeB TCST system. Each of the RpeA/ RpeB and RpoS regulates phenazine production and both of them use pip as a regulatory intermediate. However, GacS/GacA and RpeA/RpeB do not regulate phenazines hierarchically. The GacS/GacA system also regulates the synthesis of phenazines through the noncoding RNAs (ncRNAs) pathway (including rsmX, rsmY, and rsmZ). GacA is involved in the positive controls of noncoding RNAs expression, which titrate the translational repressors RsmA and RsmE. And only RsmE, not RsmA, is involved in the negative regulation of phenazines. Lon protease (coden by lon) also regulates the phenazine synthesis through negatively affecting the stability of GacA protein.
In this work, we tried to improve the 2-OH-PHZ production of GP72 by enhanceing the shikimate pathway. According to previous research, several successful strategies have been used to enhance the shikimate pathway in E. coli strains, for example, by increasing the availability of direct precursors PEP and E4P, improving DAHP synthesis in the shikimate pathway, enhancing the metabolic flow through the biosynthetic pathway by impeding allosteric and transcriptional regulation, and identifying and interfering with rate-limiting enzymatic reactions [32]. Disruption of pyruvate kinases is not the only way which could increase the availability of the direct precursors of shikimate pathway. Several other strategies have been reported. On the one hand, high PEP availability was achieved by improving the recycling of PYR to PEP by overexpression of gene ppsA (code PEP synthetase) [33]. On the other hand, high E4P availability could be achieved by the overexpression of gene coding for a transketolase (tktA) [34]. So ppsA (code PEP synthetase) and tktA (code transketolase) in GP72 were selected for overexpression to enhance shikimate pathway. Further increases in carbon flux through the shikimate pathway were realized by removing the allosteric and transcriptional regions and by relieving impeding enzymatic reactions [35]. The reactions catalyzed by DHQ dehydratase (coded by aroD), quinate/shikimate dehydrogenase (coded by aroE), dehydroquinic acid (DHQ) synthase (encoded by aroB) and DHAP synthetase (coded by aroFGH in E. coli) have been reported as limiting steps in the shikimate pathway [36]. Therefore, four key genes of GP72 (phzC, aroB, aroD, aroE which code DAHP synthase, DHQ synthase, DHQ dehydratase and SA dehydrogenase respectively) were also selected to enhance the shikimate pathway of GP72 (Fig. 1). Thus six key genes (ppsA, tktA, phzC, aroB, aroD and aroE) selected from different pathways were overexpressed to enhance the skimate pathway. According to previous report of Juminaga et al. [19], genes located near the promoter usually show much higher levels of induction than those distant from the promoter. The reverse arrangement of genes in the operon was found to benefit from the formation of the desired product through higher concentrations of protein in the latter part of the pathway. Therefore, we followed this method, in which the gene order in the fusion plasmid was aroE-aroD-aroB-phzC-tktA-ppsA, and the reverse order of the expression pathway (Fig. 6). When the plasmid pBbB5K-aroE-aroD-aroB-phzC-tktA-ppsA was overexpressed in GP72ND-3, the 2-OH-PHZ production increased to 450.4 mg/L (which was near 100-fold of that in the wide type GP72) and phenazine derivatives production also increased to 1520 mg/L.
Preventing the loss of carbon flow towards competing pathways is a successful strategy to enhance the  5 Proposed model for the regulation of phenazine biosynthesis by the TCST system in GP72. Solid straight arrows point to genes that are positively regulated. Blunt lines point to genes that are negatively affected. A dashed arrow indicates an unknown or as-yet uncharacterized regulatory pathway. In GP72, the sensors GacS and RpeB are activated by a putative environmental factor. Lon protease negatively affects GacA by controlling its protein stability. GacA positively controls the expression of rsmX, rsmY, and rsmZ, which in turn activates phenazine production by titrating the translation suppressor RsmE. In the absence of RpeA, RpeB is possibly over-phosphorylated by small phospho-donors (PD), resulting in the increased expression of the pip, phzR/phzI and the phenazine biosynthetic genes production of phenazines [37]. P. chlororaphis GP72 synthesizes phenazines with the enzymes encoded by the phz cluster using chorismate as a precursor (Fig. 1). GP72 contains at least four additional chorismate-utilizing pathways for the synthesis of folate, tryptophan, coenzyme Q, tyrosine and phenylalanine. The phenazine biosynthetic pathway must compete with these pathways for chorismate, which is a central intermediate precursor.
In order to enhance the production of phenazines, we tried to block these chorismate-utilizing pathways. Thus, pabB/pabC encoding para-aminobenzoate synthase, trpE encoding anthranilate synthase, and pheA encoding chorismate mutase/prephenate dehydratase were all deleted from GP72. The attempt was not successful because of the bad growth of these mutants (unpublished data), although a successful precedent was obtained in Pseudomonas aeruginosa PA1201 [37]. Therefore, in this work the genetic modification of GP72 mainly focused on the regulatory genes.
The phenazines produced in P. chlororaphis GP72 include 2-OH-PHZ, 2-OH-PCA and PCA [1]. According to our previous research on P. chlororaphis GP72, 2-OH-PHZ was derived from PCA, through the action of the enzyme, PhzO [1,3]. Differing from other P. chlororaphis strains, PCA could not be completely converted to 2-OH-PHZ by PhzO in the GP72 strain because of the high production of PCA [1][2][3]. Previous studies have shown the production of 2-OH-PCA in P. chlororaphis GP72AN to be relatively low, only 10-20 % of PCA [3]. Improving the efficiency of PhzO, which converts PCA to 2-OH-PHZ is also a strategy for enhancing 2-OH-PHZ production, although the mechanism to improve the activity of PhzO is still unclear. In this study, we analyzed the codon usage of phzO. Many rare codons were found to exist in this gene. To improve the expression of phzO, rare codons of the phzO were optimized, yielding phzO op , the codon-optimized variant of phzO. The phzO op was introduced into the GP72 and replaced the phzO gene successfully, generating corresponding strains. However, according to the fermentation result, optimization of the phzO gene did not improve the efficiency of PCA to 2-OH-PHZ conversion (unpublished data). This result suggested that the rare codon usage of phzO may not be the reason why PCA cannot be converted to 2-OH-PHZ completely. Currently, a study on promoting the conversion of PCA to 2-OH-PHZ is in progress.
In this work, four regulatory genes pykF, rpeA, rsmE and lon were firstly knocked out of the strain GP72 genome step by step. Then, up to six genes of ppsA, tktA, phzC, aroB, aroD and aroE were constructed into one BglBrick vector-pBbB5K and overexpressed to enhance the production of phenazine derivatives. A strain producing 450.4 mg/L of 2-OH-PHZ and 1520 mg/L of phenazine derivatives was obtained. This study laid a good foundation for the future industrial production and agricultural application of biopesticide phenazines.

Conclusions
In this work, four regulatory genes pykF, rpeA, rsmE and lon were firstly knocked out of the strain GP72 genome step by step. Then, ppsA, tktA, phzC, aroB, aroD and aroE six genes were constructed into one BglBrick vector-pBbB5K and overexpressed to enhance the production of phenazine derivatives. A strain producing 450.4 mg/L of 2-OH-PHZ and 1520 mg/L of phenazine derivatives was obtained. This study laid a good foundation for the future industrial production and agricultural application of biopesticide phenazines.

Construction of the non-scar deletion mutant strains
Two pairs of primers, lonF1 (EcoRI)-lonR1 and lonF2-lonR2 (XbaI), were designed to delete the lon sequence Fig. 6 The BglBrick plasmid of pBbB5K-aroE-aroD-aroB-phzC-tktA-ppsA, which was overexpressed to enhance the production of 2-OH-PHZ. The gene order in the plasmid was aroE, aroD, aroB, phzC, tktA and ppsA, which is a reverse of the expression pathway from P. chlororaphis GP72. Then PCR was used to amplify a 700 bp DNA fragment covering the upstream area of the lon gene and an 820 bp DNA fragment covering the downsteam of the lon gene. Overlap PCR was performed to align the two fragments, which shared a 16 bp homologous region, as shown in Additional  Fig. S1. The 1.5 kb DNA fragment that had not been modified with scars was digested with the restriction enzymes XbalI and EcoRI and cloned into the plasmid pK18mobsacB. The products were then transferred to E. coli S17-1 (λpir) and then moved to P. chlororaphis GP72 to generate the mutant GP72Δlon through biparental mating. Single-crossover clones were selected on plates containing 50 μg/mL Kn and 100 μg/mL Ap. Then double-crossover clones were selected on plates containing 15 % sucrose. These were confirmed by PCR analysis and sequencing. The simple sketch for constructing a non-scar deletion GP72Δlon is shown in Additional file 1: Fig. S1. In similar ways, the pykF, rpeA and rsmE non-scar deleted mutants were constructed in their corresponding strains.

DNA manipulations for overexpression
All genes were PCR-amplified from the genomic DNA of P. chlororaphis GP72 and extended by using the 5′ terminal sequence 5′-AAAGGAGGCCATCC-3′ and endonuclease restriction sites located at the 5′ and 3′ ends [19]. All of these genes were cloned into T vector pMD19-T or pEASY-Blunt and transformed into E. coli.
The plasmids used for overexpression were constructed following the Bglbrick standard, and the plasmid pBbB5K-GFP was used as the backbone. All XhoI, BglII, BamHI, and EcoRI restriction sites were removed from all genes cloned into BglBrick plasmids (pBbB5K). Because the genes aroD, tktA and ppsA contain 1-3 of these restriction sites respectively, point mutation was used to remove these restriction sites. Point mutations were conducted using a Quick-Changes ™ Site-Directed Mutagenesis Kit following the manufacturer's instructions (La Jolla, CA, USA). The ORFs were then PCR-amplified using primers that extended the 5′ ends with EcoRIXXBglII, the consensus 5′-AGGAGG-3′ ribosome binding site (RBS) and a spacer sequence, 5′-CCATCC-3′ and the 3′ ends with BamHIXXXhoI. Like the other ORFs in this study, before cloning, all PCR fragments were digested with XhoI and EcoRI then inserted into the plasmid from the 5′ end to the 3′ end, replacing the gene GFP which contained in the plasmid originally [19].
In brief, six brick plasmids, which contain pBbB5K-aroB, pBbB5K-aroD, pBbB5K-aroE, pBbB5K-phzC, pBbB5K-tktA and pBbB5K-ppsA respectively, were first constructed. Subsequently, the overexpression plasmids which contained two, three or six genes were constructed. A simplified scheme of the steps used in the construction of these plasmids is shown in Additional file 2: Fig. S2.
All of the strains which held plasmids for over-expression were generated by transforming the plasmid into corresponding competent cells with elec-transformation (Bio-Rad, Hercules, USA).

Fermentation processing
Pseudomonas chlororaphis GP72 and its derivative strains stored in a −80 °C freezer were activated at 28 °C for 12-24 h in King's B agar media with the corresponding antibiotics. Selection of single colonies from Petri plates was performed, which were then used to inoculate approximately 5 mL of King's B broth (supplemented with the corresponding antibiotics) in 50 mL flasks. Cultures were then incubated at 28 °C with 180 rpm of shaking overnight. Portions of these cultures were then inoculated into 250 mL baffled flasks containing 60 mL King's B broth with corresponding antibiotics to achieve an initial OD600 of 0.02. Then the fermentation process was initiated. For induction of the overexpression strains, 50 μM IPTG (isopropyl-β-dthiogalactopyranoside) was added to the culture after 12 h of incubation time. After 24-120 h growth at 28 °C and 180 rpm, cultures were collected for the measurement of phenazine compounds and OD600. Triplicate experiments were carried out for each fermentation test.

Quantification of phenazine compounds
In order to quantify the phenazine compounds, the fermentation broth was firstly adjusted to pH 2.0 with 6 M HCl. Then the fermentation broth was extracted with three volumes of ethyl acetate with vigorous shaking. The organic layer was collected and mixed with a 1/10 volume of distilled water. The mixture was then shaken rigorously. Finally, the organic phase containing phenazine compounds was evaporated under vacuum pressure. The phenazine compounds were dissolved in methanol for further analysis. HPLC analysis was performed on the phenazine compounds (2-OH-PCA, 2-OH-PHZ, and PCA) using a 1260 Infinity HPLC apparatus (Agilent Technologies Group, Santa Clara, USA) with a C-18 reverse phase column and UV detector (Agilent, Santa Clara, USA) as described previously [1].

Nucleotide sequence accession numbers
The nucleotide sequence of the genes used in this work were deposited in the GenBank, National Center for Biotechnology Information, http://www.ncbi.nlm.nih.gov/ nucleotide. They correspond to the genome shotgun sequence of P. chlororaphis strain GP72, more information was shown in Additional file 4: Table S2.