Genetic engineering of Pseudomonas chlororaphis Lzh-T5 to enhance production of trans-2,3-dihydro-3-hydroxyanthranilic acid

Trans-2,3-dihydro-3-hydroxyanthranilic acid (DHHA) is a cyclic β-amino acid used for the synthesis of non-natural peptides and chiral materials. And it is an intermediate product of phenazine production in Pseudomonas spp. Lzh-T5 is a P. chlororaphis strain isolated from tomato rhizosphere found in China. It can synthesize three antifungal phenazine compounds. Disruption the phzF gene of P. chlororaphis Lzh-T5 results in DHHA accumulation. Several strategies were used to improve production of DHHA: enhancing the shikimate pathway by overexpression, knocking out negative regulatory genes, and adding metal ions to the medium. In this study, three regulatory genes (psrA, pykF, and rpeA) were disrupted in the genome of P. chlororaphis Lzh-T5, yielding 5.52 g/L of DHHA. When six key genes selected from the shikimate, pentose phosphate, and gluconeogenesis pathways were overexpressed, the yield of DHHA increased to 7.89 g/L. Lastly, a different concentration of Fe3+ was added to the medium for DHHA fermentation. This genetically engineered strain increased the DHHA production to 10.45 g/L. According to our result, P. chlororaphis Lzh-T5 could be modified as a microbial factory to produce DHHA. This study laid a good foundation for the future industrial production and application of DHHA.

www.nature.com/scientificreports/ Lzh-T5 is a P. chlororaphis strain isolated from the tomato rhizosphere located in China. It has a phenazine biosynthesis cluster phzABCDEFG, and could produce phenazine-1-carboxylic acid (PCA) and other phenazine derivatives 15 . In this study, phzF of P. chlororaphis Lzh-T5 was disrupted causing DHHA accumulation. Three negative regulatory genes (pykF, psrA, and rpeA) were stepwise disrupted in P. chlororaphis Lzh-T5. The yield of DHHA increased from 2.15g/L to 5.52 g/L. To improve DHHA production, key genes were overexpressed by BglBrick vectors from the shikimate, pentose phosphate, and gluconeogenesis pathways of Lzh-T5, obtaining 7.89 g/L DHHA. The effect of Fe 3+ on DHHA production was investigated. A final DHHA yield of 10.45 g/L was obtained.

Methods
Microorganisms, growth conditions, and plasmids. All plasmids, primers, and microorganisms are listed in Table 1 and Supplementary Material 1: Table S1. All Escherichia coli strains were cultured in LB medium at 37 °C, and all P. chlororaphis strains were cultured in KB medium at 30 °C. Kanamycin and Ampicillin were used. All details are detailed in our previous research 16 . DNA manipulation. A no-scar deletion method was used in the genome of P. chlororaphis Lzh-T5 and its derivative strain. In order to create LDA-1 by the interruption of phzF in Lzh-T5, two pairs of primers (phzF-A (EcoRI)-phzF-B and phzF-C-phzF-D (XbaI)) were designed. Upstream (740 bp) and downstream (793 bp) of phzF were first amplified by PCR. A 1515 bp fragment fusion was amplified by overlap PCR. The fusion fragment was constructed into pK18mobsacB, creating the recombinant plasmid pK18-phzF.
The pK18-phzF plasmid was transferred to E. coli S17-1 (λpir) by heat shock transformation. Then biparental mating between E. coli S17-1 and P. chlororaphis Lzh-T5 generated the mutant LDA-1 strain. Single-crossover and double-crossover clones were selected stepwise. To ensure accuracy, PCR analysis and sequencing were used to confirm the deletion. Detailed steps are listed in our previous research 16 . Similarly, rpeA, pykF, and psrA were disrupted in their corresponding strains.   17 . The plasmid pBbB5K-GFP was used as the backbone to overexpress six key genes. Recombinant plasmids used in this study were constructed following the methods described in our previous research 16 . In brief, six Brick plasmids containing aroB, aroD, aroE, phzC, tktA, and ppsA, respectively, were first constructed. Point mutations were made in order to remove restriction sites (EcoRI, HindIII, BamHI, and BglII) in aroD, tktA, and ppsA. Then, a complicated overexpression plasmid pBbB5K-aroE-aroD-aroB-phzC-tktA-ppsA was constructed following the BglBrick standard.
Quantitative RT-PCR. Quantitative RT-PCR was used to detect the transcriptional changes of related genes of different strains. And the rpoD which is one of housekeeping gene in P. chlororaphis was selected as the internal reference gene. The measurement methods of Quantitative RT-PCR are following our previous work 18 . And the fold change for mRNAs was calculated by the 2 −ΔΔCt method 19 .
Fermentation processing. All P. chlororaphis strains were stored in an ultra-low temperature refrigerator.
Strains were activated in KB petriplate with Ampicillin at 30 °C for 12-24 h before fermentation. Single colonies were inoculated in a 50 mL flask which contains 5 mL of KB medium and incubated overnight. Portions of seed bacteria were then inoculated in 50 mL KB medium in a 250 mL baffled flask and inoculation initial OD600 was 0.03. The gene expression of BglBrick plasmids was induced by isopropyl-β-d-thiogalactopyranoside (IPTG). 50 μM IPTG was used during an incubation of 12 h. After growing at 30 °C and centrifuged at 200 rpm for 24-72 h, the fermentation broth was collected to measure phenazine compounds and OD600. All experiments were performed in triplicate, and data were averaged and reported as mean ± standard deviation. After growing at 30 °C and 200 centrifuged at rpm for 24-72 h, the culture was collected for measurement of phenazine compounds and OD600. All fermentation experiments were performed in triplicate, and the data were averaged and reported as mean ± standard deviation. Measurement, purification, and quantification of DHHA from fermentation broth. Determination and purification of DHHA from fermentation broth followed the methods described in previous research 20 . Briefly, the supernatant of the fermentation broth was collected by centrifuging at 12,000g for 15 min. Then, it was analyzed by liquid chromatography-mass spectrometry (LC-MS) after processing through a 0.22 μm polyvinylidene difluoride syringe filter. LC-MS was performed on the Agilent HPLC1290-MS6230 system (Agilent Technologies, Santa Clara, CA) by an Agilent Extend C18 column (50 mm × 2.1 mm, 1.8 μm). It was eluted with methanol/0.1% formic acid (50:50, v/v) at a 0.15 mL/min flow rate. The DHHA sample was analyzed by mass spectrometry in the positive ion detection mode.
In order to purify DHHA, Shimadzu Inert Sustain phenyl column (20 × 250 mm, 15 μm) was used in HPLC (Shimadzu LC8A, Shimadzu, Kyoto, Japan). It was eluted with water/methanol (90:10, v/v) with a flow rate of www.nature.com/scientificreports/ 2 mL/min at a scanning wavelength of 278 nm. We collected the peak containing DHHA and dried it by vacuum freezing. The crystals of DHHA were obtained at room temperature after dissolving them in hot ethanol (65 °C). Quantification of DHHA was described in our previous work 18 . Briefly, the fermentation broth was centrifuged at 11,000 rpm for 3 min and supernatant collected. The samples were analysis by HPLC (Agilent 1260, USA) with a Shimadzu Inert Sustain phenyl column (4.6 × 250 mm, 5 μm) to determine the amount of DHHA. It was eluted with 0.1% formic acid/methanol (85:15, v/v) with a flow rate of 1 mL/min at a scanning wavelength of 278 nm. A DHHA standard curve was used to infer the DHHA content by plotting the concentration of the DHHA standard solution on the abscissa and the corresponding absorbance (peak area) on the vertical. Superoxide dismutase activity measurement. In the Fe 3+ addition experiments, superoxide dismutase (SOD) activities of different strains were measured. The methods of measure are following the previous literature 21,22 .

Results
Disruption of phenazine synthesis in Lzh-T5. According to the research from Blankenfeldt, DHHA undergoes an isomerization reaction facilitated by PhzF and converts to 6-amino-5-oxocyclohex-2-ene-1-carboxylic acid in Pseudomonas spp 23 . In order to accumulate DHHA, phzF was disrupted in P. chlororaphis Lzh-T5, and the strain P. chlororaphis LDA-1 was got. As shown in Fig. 2, after 48 h of culture, the P. chlororaphis LDA-1 strain on an agar plate turns milky; however, the colony color of Lzh-T5 remains orange. This suggests that the phenazine derivatives could not be synthesized in LDA-1. After fermentation and analysis by HPLC-UV, phenazines (including 2-hydroxyphenazine, PCA, and 2-OH-PCA) disappeared in the broth of P. chlororaphis LDA-1 (Fig. 3).
Similar to previous research, a new absorption peak appears in the HPLC-UV at a wavelength of 278 nm, which is the maximum absorption wavelength of DHHA (Fig. S1) 20 . According to the analysis of LC-MS, the mass-to-charge ratio (m/z) of the compound was 156.0658 for [C 7 H 9 NO 3 + H] + (the mass of DHHA is 155.0655; Fig. S1). When phzF is overexpressed in the LDA-1 strain, production of phenazine was recovered (Fig. 3). According to these results, it suggests that disruption of phzF in Lzh-T5 causes accumulation of DHHA. Using HPLC-UV analysis, the DHHA yield of LDA-1 reached 2.15 g/L in 48 h (Fig. 4a). Our results suggest that disruption of phzF had little effect on Lzh-T5 cell growth (Fig. 4b).
Knockout of negative regulatory genes to boost DHHA production. According to previous research, interruption the pykF (which encode pyruvate kinase) enhances the production of 2-hydroxyphenazine 16 . In this study, pykF from LDA-1 was initially chosen to inactivate, so we obtained a mutant strain LDA-2 (Fig. 5). After analysis of the fermentation broth by HPLC-UV, the DHHA yield of strain LDA-2 increased from 2.15 g/L to 4.17 g/L (Fig. 5a). The growth condition of the strain was detected to have little effect after deletion of pykF (Fig. 5b).
PsrA, a sigma regulator, was first reported in P. chlororaphis PCL1391 24 . According to Chin-A-Woeng et al., PsrA played a negative regulatory role in the production of the antifungal metabolite PCA in P. chlororaphis PCL1391 24 . Similar results were obtained from psrA disruption in P. chlororaphis HT66 (Peng et al. 2018) 25 . According to our research, the sigma regulator psrA also exists in P. chlororaphis Lzh-T5. The strain LDA-3 was obtained after the gene psrA was disrupted in LDA-2, and the production of DHHA increased from 4.17 g/L to 4.92 g/L (Fig. 5a).
RpeA, a negative regulator of Phenazine, was mutated by insertion in P. chlororaphis GP72 and results in a production increase of 2-hydroxyphenazine 16,26 . RpeA is part of the two-component signal transduction system  27,28 . In this study, rpeA was disrupted in the LDA-3 genome to construct LDA-4. Similar to the insertional mutagenesis of P. chlororaphis GP72, DHHA yield of LDA-4 increased from 4.92 g/L to 5.52 g/L (Fig. 5a). After knocking out the negative regulatory genes, quantitative RT-PCR results showed that the transcript level of genes phzD and phzE in the derivative strains, which is key gene for DHHA synthesis, has increased significantly (Fig. S2).
Enhanced DHHA production by key gene overexpression. We disrupted pykF to improve the yield of DHHA from 2.15 g/L to 4.17 g/L by diverting more metabolic flux into the shikimate pathway from other pathways (Fig. 5a). This indicates that enhancing the lead synthesis pathway, we could improve the yield of DHHA in P. chlororaphis Lzh-T5. Compared with knocking out negative regulatory genes, gene overexpression is another effective strategy often used to increase the yield of biologic products in microorganisms. According to previous research, overexpression of key genes in the shikimate pathway enhanced the yield of 2-OH-PHZ 16 .
In order to increase the production of DHHA, aroB, aroD, aroE, phzC, tktA, and ppsA from Lzh-T5 were overexpressed from the shikimate, pentose phosphate, and gluconeogenesis pathways. We used a previously employed kind of modular vector, the BglBrick plasmid 29 . A recombinant plasmid containing six genes, pBbB5K-aroE-aroD-aroB-phzC-tktA-ppsA, was constructed. Strain LDA-5 was created after transformation into the strain  www.nature.com/scientificreports/ LDA-4 by electrotransformation. After fermentation, the DHHA production of LDA-5 increased to 7.89 g/L after 48 h (Fig. 5a). This indicates that overexpression of key genes is an effective strategy to enhance the production of DHHA. Quantitative RT-PCR results showed that the transcript level of six genes overexpressed in the derivative strain LDA-5 has increased (Fig. S3).
Enhanced DHHA production with Fe 3+ . Environmental factors have important effects on secondary metabolite production in Pseudomonas strains, especial ion concentration in the medium 30,31 . There is no universal medium suitable for Pseudomonas strains which can produce phenazines due to different nutritional requirements 30 . According to previous research, the DHHA production has a 30% increase after adding 3 mM of Fe 3+20 . To improve the production of DHHA, the effect of different concentrations of Fe 3+ in the medium was investigated. After fermentation, DHHA production was detected by HPLC-UV. Low concentration of iron ions promoted DHHA production. High concentrations of iron ions inhibited DHHA production (Fig. 6). Different from our previous research, in the fermentation of LDA-5, the optimum concentration which enhanced the production of DHHA was 2 mM. We obtained a maximum DHHA yield of 10.45 g/L with 2 mM Fe 3+ (Fig. 6a).

Discussion
Pseudomonas spp is a class of microorganism that exists widely in the environment 32 . It has strong adaptability and often produces resistant substances. Among these substances, phenazine derivatives are typical secondary metabolites, such as PCA, Pyocyanin (PYO), 2-Hydroxyphenazine and phenazine-1-carboxamide (PCN) 25,33 . Phenazine derivatives are produced by the phz gene clusters (phzABCDEFG) found in several Pseudomonas spp (including P. chlororaphis, P. fluorescence, and P. aeruginosa) [34][35][36] . In P. chlororaphis, phzC catalyzes the first reaction of the shikimic acid pathway, by catalysis of E4P and PEP to synthesize DAHP 13 . PhzE, the first key enzyme in the phenazine synthesis pathway which catalyzes the last product of the shikimate pathway into  www.nature.com/scientificreports/ 2-amino-4-deoxy branched acid 37 . ADIC converts to DHHA and pyruvate by the isobranse enzyme PhzD 23 . It is then epimerized with diaminopimelate enzyme (diaminopimelate epimerase (DapF)) PhzF to 6-amino-5-oxocyclohex-2-ene-1-carboxylic acid (AOCHC) 38 . Two molecules of AOCHC are converted to hexahydrophenazine-1,6-dicarboxylic acid (HHPDC) catalyzed by the dimer PhzAB 23 . HHPDC spontaneously undergoes oxidative decarboxylation to form tetrahydrophenazine-1-carboxylic acid (THPCA). THPCA is catalyzed by PhzG to form 5,10-dihydro-phenazine-1-carboxylic acid (DHCCA) 39 , DHCCA eventually undergoes self-oxidation in the air to form PCA. While DHHA is an important intermediate product of phenazines, the production of phenazines is quite low. For example, in wild type P. chlororaphis GP72, the yield of 2-OH-PHZ is 4.5 mg/L, and the higher phenazine (PCN) production in wild type is about 400 mg/L in P. chlororaphis HT66 25,26 . The phenazine production strains are also the potential DHHA producers. P. aeruginosa is an opportunistic pathogenic bacterium, while P. chlororaphis is not. P. chlororaphis was selected as a candidate for DHHA production 16 . According to previous research, P. chlororaphis GP72 could accumulate 1.92 g/L DHHA with phzF disruption 20 . Lzh-T5 is a P. chlororaphis strain isolated from the tomato rhizosphere found in China. It has the phenazine biosynthesis cluster phzABCDEFG, and can produce phenazine derivatives 15 . In this study, phzF was disrupted from the genome of P. chlororaphis Lzh-T5 making the strain P. chlororaphis LDA-1 which has 2.15 g/L DHHA accumulation. Enhancing the shikimate pathway is an effective strategy for aromatic compound production in microorganisms 20 . The direct precursors of the shikimate pathway are PEP and E4P. The PEP and E4P supply could be enhanced by an overexpression of PEP synthase encoded by ppsA and transketolase encoded by tktA 1,40 . In addition, inactive pykF, which encodes pyruvate kinase, can increase PEP 41,42 . Further carbon flux of the shikimate pathway impedes enzymatic reactions and removes the allosteric and transcriptional regions 20,43 . Quinate/shikimate dehydrogenase, dehydroquinic acid synthase, dehydroquinic acid dehydratase, and DHAP synthetase, encoded by aroE, aroB, aroD, and phzC, respectively, were reported as limiting steps in the shikimate pathway 44 . Our previous studies showed that disruption of pyruvate kinase (encoded by pykF) and coexpression of ppsA, tktA, aroB, aroD, aroE, and phzC in GP72 increased phenazine production 20 . In this study, the disruption of pykF and six gene tktA, ppsA, aroB, aroD, aroE, and phzC coexpression in LDA-1 increased the DHHA production to 7.89 g/L.
TCST systems exist in Pseudomonas spp that help them adapt to the environment by coordinating cellular pathways to interact with the environment 28 . Different TCST systems were found in P. chlororaphis Lzh-T5. Among these TCST systems, GacS/GacA is one of the most researched. GacS/GacA was the earliest TCST system which could enhance bioactive secondary metabolite production in Pseudomonas spp 45,46 . In P. chlororaphis HT66, GacA positively regulates the expression of psrA. psrA negatively controls the expression of rpoS and the expression of phenazine 25 . In this study, psrA was found in P. chlororaphis Lzh-T5. Interruption of psrA increased the production of DHHA from 4.17 g/L to 4.92 g/L, this suggests that psrA negatively controls the production of DHHA. RpeA/RpeB is a TCST system found widely in Pseudomonas spp 28 . According to the research of Whistler et al., Pip improved the production of PCA by enhancing the expression of phzR and phzI. Sigma factor rpoS regulated the expression of pip, which itself is regulated by the RpeA/RpeB TCST system. RpeB was inhibited by RpeA, so an rpeA mutant strain enhanced the production of phenazine 28 . Similar results were observed from P. chlororaphis GP72 and P. chlororaphis HT66 16,25 . The rpeA/rpeB TCST system was found in P. chlororaphis Lzh-T5, interruption of rpeA results in the increase in DHHA from 4.92 g/L to 5.52 g/L (Fig. 5a). This result suggests that, similar with P. chlororaphis 30-84, rpeA negatively impacts the production of DHHA.
Metal ions are an important factor which could affect the production of secondary metabolites of Pseudomonas spp 30 . Fe 3 + is one of the important ions, and according to the research of Shtark et al., it plays a positive role in the production of phenazines in P. chlororaphis SPB1217. They hypothesize that Fe 3+ activates some dependent superoxide dismutases and those superoxide dismutases promotes phenazine production by removing reactive oxygen species or bacterial metabolites that suppress enzymes involved in the synthesis pathway of phenazines 30 . Similar results were reported in P. aeruginosa PCL1391 and P. fluorescens 2-79 47,48 . Fe 3+ also had a positive effect on DHHA production 20 . And our results shown that the SOD activity of the strain has increased after Fe 3+ adding to the medium (Fig. S4). In this study, adding different concentration of Fe 3+ had different effects on DHHA production in P. chlororaphis LDA-5. Low concentrations of Fe 3+ promoted DHHA production, while high concentrations of Fe 3+ inhibited DHHA production (Fig. 6). Adding 2 mM of Fe 3+ had a positive effect on DHHA production, and our maximum DHHA yield was 10.45 g/L in this study.
In conclusion, phzF of P. chlororaphis Lzh-T5 was disrupted to construct the strain P. chlororaphis LDA-1 which had DHHA accumulation. Then, three negative regulatory genes (pykF, psrA and rpeA) were disrupted stepwise in P. chlororaphis LDA-1. The production of DHHA increased from 2.15 g/L to 5.52 g/L. Next, different key genes selected from the shikimate, pentose phosphate, and gluconeogenesis pathways of Lzh-T5, were overexpressed by BglBrick vectors. The production of DHHA increased from 5.52 g/L to 7.89 g/L. The effect of adding Fe 3+ on DHHA production was investigated in a strain (LDA-5) with a resulting DHHA yield of 10.45 g/L.

Data availability
All data generated or analyzed during this study are included in this published article (and its Additional file). www.nature.com/scientificreports/