Table 1 lists all the bacterial strains used in this study. E. coli DH5α was used as the cloning host, and E. coli AB2834 was used as the metabolite production host. All E. coli strains were grown in Luria-Bertani (LB) medium at 30 or 37°C with the appropriate antibiotics. Small-scale cultivation and fed-batch fermentation for anthranilate production were conducted, as described previously (Lee et al., 2021). For small-scale cultivation in a 24-well culture plate, a single colony was inoculated in 1.3 ml LB medium at 30°C for 15 h. The culture broth was inoculated with 1% (v/v) in the same medium at 30°C for 15 h. The secondary culture broth was inoculated in 1.3 ml of E. coli production medium (EPM) at 30°C for 4 days. The miniature cultivation was performed using a humidity chamber set to 80% humidity and 200 rpm.

Table 1 lists the constructed plasmids, and Supplementary Table 1 presents all primer pairs used in this study. The CRISPR/Cas system was utilized for targeted gene editing. The two-plasmid system, in which the cas9 gene and a targeting N₂₀ sequence directing it to the gene loci of interest, were separated in the pCas and pTargetF, respectively. N₂₀ sequences followed by the PAM were designed by the web tool, CHOPCHOP.¹ The homologous DNA fragments to the upstream and downstream regions of the target gene were amplified with primer sets. The fragment, including the N₂₀ sequence and guide RNA scaffold, was also amplified. These fragments were cloned into SpeI/HindIII-digested pTargetF based on the In-Fusion Cloning method (TaKaRa, Japan). For gene overexpression, PJ23119 or PoppA promoter sequence and target gene amplified with the specific primers were cloned with homologous DNA fragments and N₂₀ sequence into pTargetF simultaneously. Genome editing and plasmid curing were performed, as described previously (Jiang and Zhang, 2016). The transformants were verified by colony PCR and DNA sequencing. TransStart® FastPfu Fly DNA polymerase (Transgen Biotech., China) and SapphireAmp® Fast PCR Master Mix (TaKaRa, Japan) were used to amplify the target-specific fragments and colony PCR, respectively.

For fed-batch fermentation, the single colony was inoculated in 5 ml LBG (25 g/L LB broth and 20 g/L glucose) medium at 30°C, 200 rpm for 12 h. The primary culture broth was inoculated with 1% (v/v) in 20 ml same medium at 30°C for 6 h. The secondary culture broth was then inoculated with 1% (v/v) in a 7 L fermenter. The production culture was performed using the modified PB4-md5 medium. The modified PB4-MD5 medium included the following: 30 g/L glucose, 10 g/L glycerol, 37.125 g/L yeast extract, 5.25 g/L KH₂PO₄, 1 g/L MgSO₄∙7H₂O, 2 ml/L trace metals, and 200 μg/L thiamine hydrochloride. The feeding medium was composed of 600 g/L glucose, 100 g/L yeast extract, 20 g/L MgSO₄∙7H₂O, and 5 ml/L trace metal. Phosphate was not added to the feeding medium to regulate cell growth. The feeding medium was supplied using a peristaltic pump when glucose was depleted. The pH was adjusted with 10 N NaOH or 3 M HCL at 7.0, and the DO level was maintained above 40% by controlling the rpm, aeration, and feeding rates (Lee et al., 2021).

The cultured broth was centrifuged at 15,000 RPM for 5 min. The supernatant was filtered using a 0.22 μm syringe filter (Rainbow PVDF membrane filter) and stored at −20°C. The concentrations of anthranilate and the shikimate pathway metabolites were determined by high-performance liquid chromatography (HPLC) using a Zorbax SB-Aq column (4.6 × 250 mm, Agilent, USA). The mobile phase was 0.1% trifluoroacetic acid (TFA) in 30% MeOH, and the flow rate was 0.5 ml/min. Anthranilate and shikimate pathway metabolites were detected at 330 and 214 nm, respectively.

The shikimate over-producing strain (Inha215) obtained in a previous study (Lee et al., 2019) was an engineered E. coli cell factory strain designed to maximize the accumulation of shikimate by deleting six genes (tyrR, ptsG, pykA, shiA, aroL, and aroK) from the chromosome and by overexpressing eight genes (aroB, aroD, aroF, aroG, galP, ppsA, aroE, and tktA; Table 1). An attempt was made to re-activate the shikimate pathway by compensating for the shikimate kinase genes (aroL and aroK) involved in the conversion of shikimate to shikimate-3-phosphate. Using the CRISPR/Cas9 system, two genes, aroL and aroK, were complemented to be expressed under their own-promoter regulation. Unlike the shikimate over-producing Inha 215 strain, which could not synthesize the essential aromatic amino acids, this complemented strain was activated by shikimate kinases and grew well in minimal medium (Supplementary Figures 1A,B).

Anthranilate is a metabolite produced by converting chorismate by an anthranilate synthase (TrpEG) in the shikimate pathway (Figure 1), which is then converted to an anthranilate-phosphate precursor by anthranilate phosphoribosyltransferase, TrpD (Xie et al., 2003; Merino et al., 2008; Balderas-Hernández et al., 2009). Therefore, by deleting trpD in the presence of aroL and aroK, anthranilate was accumulated successfully without other metabolic processes (Supplementary Figures 1C,D). In addition, there are several enzymes involved in bypass routes for chorismate to be converted to other metabolites, including bifunctional chorismate mutase/prephenate dehydratase (encoded by pheA and tyrA), anthranilate synthase (encoded by trpEG), ρ-aminobenzoate synthase (encoded by pabAB), isochorismate synthase (encoded by menF and entC), chorismate lyase (encoded by ubiC) (Figure 1, Dosselaere and Vanderleyden, 2001; Sprenger, 2007; Noda et al., 2016). The genes involved in the bypass routes mentioned above were removed sequentially to maximize the accumulation of chorismite as much as possible. Based on the shikimate over-producing Inha215 strain, Inha250 with the pheA and tyrA genes removed, Inha251 with the pabA and tyrA gene removed, Inha252 with the entC gene removed, and Inha254 strain with the pheA, tyrA, pabA, entC, and ubiC gene removed were generated sequentially (Supplementary Figures 1C,D). Finally, approximately 18.7 mg/L of anthranilate was produced into the Inha254 strain (Supplementary Figure 1E). The Inha255 strain, which also removed menF encoding isochorismate synthase in E. coli, was also developed, but unlike the Inha254 strain, anthranilate production was not carried out (Supplementary Figure 1E).

Five key shikimate pathway genes, such as trpR, aroE, tktA, aroL, and aroK, were engineered further to maximize anthranilate production using the Inha254 strain constructed above. First, a well-known shikimate pathway transcriptional regulator gene, trpR, was deleted from the Inha254 chromosome, and the aroE was integrated into the Inha254 chromosome under the control of strong oppA promoter (Supplementary Figures 2, 3; Gu et al., 2016). Second, to increase the erythrose-4-phosphate (E4P) pool, one of the essential precursors of the shikimate pathway, the tktA involved in the pentose phosphate pathway was over-expressed under the control of the strong promoter J23119 (Supplementary Figure 4; Shen et al., 2012; Lee et al., 2021). Third, the aroL and arolK genes were coupled translationally and expressed under the control of the strong promoter J23119 (Supplementary Figure 5). The strain called Inha256, which was optimized for these five genes (trpR, arooE, tktA, and aroL&aroK) produced 56.8 mg/L of anthranilate, which is approximately three times higher production yield than that of Inha254 (Figure 2).

Moreover, to minimize the chorismite bypass pathways, menF encoding 2-succinyl-5-enolpyruvyl-6-hydroxy-3-cyclohexene-1-carbosylate synthase and pabB encoding ρ-aminobenzoate synthase subunit involved in the chorismite metabolism to either isochorismate and p-aminobenzoate, respectively (Supplementary Figure 6). In addition, two genes, aroA and aroC, were translationally coupled and overexpressed under the control of strong oppA promoter (Supplementary Figure 7). Approximately 74.6 mg/L of anthranilate was produced in the strain described above (called Inha 257; Figure 2).

Although the anthranilate production yield of the Inha257 strain increased 1.3 fold compared to the Inha256 strain, the cell growth of Inha257 also increased 1.1 fold, suggesting that the redesign of the four genes (menD, fabB, aroA, and aroC) attempted in the Inha257 had a positive effect on cell growth rather than anthranilate production optimization (Figure 2). Occasionally, the inha257 strain was inconsistent in both cell growth and anthranilate productivity compared to the Inha256 strain (data not shown), so that the subsequent fed-batch experiment was conducted using the Inha256 strain.

Cell growth increased rapidly after 8 h of incubation. The stationary phase was reached after 18 h in a 7-L fed-batch fermentation of Inha256 strain (Figure 3). Subsequently, cell growth slightly increased after 27 h of culture with a lower feeding rate, but showed constant cell growth until 72 h when the culture was terminated. The production yield of anthranilate in the Inha256 strain gradually increased after 10 h of culture and reached the maximum after 60 h of culture, producing approximately 4 g/L in the recombinant E. coli (Figure 3). Interesting, however, 18 g/L of anthranilate was once observed depending on the culture and feeding conditions (Supplementary Figure 8).