Metabolic engineering of Escherichia coli for production of 2‐Phenylethylacetate from L‐phenylalanine

Abstract In order to meet the need of consumer preferences for natural flavor compounds, microbial synthesis method has become a very attractive alternative to the chemical production. The 2‐phenylethanol (2‐PE) and its ester 2‐phenylethylacetate (2‐PEAc) are two extremely important flavor compounds with a rose‐like odor. In recent years, Escherichia coli and yeast have been metabolically engineered to produce 2‐PE. However, a metabolic engineering approach for 2‐PEAc production is rare. Here, we designed and expressed a 2‐PEAc biosynthetic pathway in E. coli. This pathway comprised four steps: aminotransferase (ARO8) for transamination of L‐phenylalanine to phenylpyruvate, 2‐keto acid decarboxylase KDC for the decarboxylation of the phenylpyruvate to phenylacetaldehyde, aldehyde reductase YjgB for the reduction of phenylacetaldehyde to 2‐PE, alcohol acetyltransferase ATF1 for the esterification of 2‐PE to 2‐PEAc. Using the engineered E. coli strain for shake flasks cultivation with 1 g/L L‐phenylalanine, we achieved co‐production of 268 mg/L 2‐PEAc and 277 mg/L 2‐PE. Our results suggest that approximately 65% of L‐phenylalanine was utilized toward 2‐PEAc and 2‐PE biosynthesis and thus demonstrate potential industrial applicability of this microbial platform.

Although the chemically synthesized 2-PEAc has strong advantages in cost and yield, consumers prefer more natural flavor compounds in the field of food and cosmetics (Etschmann, Bluemke, Sell, & Schrader, 2002). However, production of natural 2-PEAc by extraction from the essential oils of flowers and plants is costly and labor-intensive (Etschmann et al., 2002). Therefore, the production of 2-PEAc using engineered microorganisms can be an attractive alternative to the traditional low-yielding and costly extraction and distillation processes.

| Enzymes and DNA kits
Phusionpolymerase and T4 ligase were purchased from New England Biolabs. Plasmid mini kits, PCR purification kits and gel extraction kits were ordered from Fermentas (Burlington, Canada).
The XbaI-XhoI fragment of YjgB from pPG32 was inserted into NheI and XhoI sites of pPG30 to give pPG34. The XbaI-XhoI fragment of ADH6 from pPG31 was inserted into NheI and XhoI sites of pPG30 to give pPG35.
The XbaI-XhoI fragment of ATF1 from pPG33 was inserted into NheI and XhoI sites of pPG34 to give pPG36. ARO8 (GenBank EWH18548.1) gene from S. cerevisiae YPH499 was amplified by PCR using primers ARO8-XbaI, and ARO8-XhoI, and inserted into NheI and XhoI sites of pPG36 to give pPG37. The sequences of all primers used in PCRs are listed in

| Cell transformation
Escherichia coli strain MG1655 competent cells were transformed using plasmids pPG30-37. Cells were selected on LB plates containing kanamycin (50 mg/L).

| Shake flask cultures
Recombinant strains of E. coli were streaked onto LB agar plates with antibiotics (50 mg/L kanamycin) and incubated at 37°C for 12-20 hr.
Single colonies were picked and inoculated into 5 ml LB medium in 20-ml flasks. Flasks were incubated at 37°C in a rotary shaker at 200 rpm for 12 hr. Cells were collected by centrifugation at 4,000 g for 1 min, resuspended in 100 ml modified M9 medium as previously described by Guo, Zhu, Deng, & Liu (2014), and shaken as above.
When the OD 600 reached 0.8, IPTG was added to a final concentration of 0.1 mmol/L. Samples were taken at 28 hr after fermentation and analyzed by GC/MS. F I G U R E 1 Engineered pathways for production of 2-PEAc from L-phenylalanine. Aminotransferase ARO8 for transamination of L-phenylalanine to phenylpyruvate, 2-keto acid decarboxylase KDC for the decarboxylation of the phenylpyruvate to phenylacetaldehyde and aldehyde reductase YjgB for the reduction of phenylacetaldehyde to 2-PE. 2-PEAc were produced from 2-PE and acetyl-CoA through the expression of alcohol acetyltransferase ATF1 that catalyzes the esterification of 2-PE and acetyl-CoA

| Construction of 2-PE biosynthetic pathway in Escherichia coli
The gene KDC which encode 2-keto acid decarboxylase was amplified from S. cerevisiae and introduced to E. coli MG1655 strain to yield The recombinant and the control strains were cultured in a modified M9 media and the 2-PE production was analyzed by GC-MS ( Figure S1). No 2-PE was detected in the negative control strain E. coli.
In contrast, 85.7 ± 4.34 mg/L of 2-PE was formed in the recombinant strains (Table 3). This result suggests that there is an endogenous phenylacetaldehyde reductase in E. coli.
YjgB has been identified as aldehyde reductase in E. coli and can accept a broadly range of various aldehydes as substrates for the various alcohols production (Guo, Hong, & Xun, 2015;. To improve the production of 2-PE, YjgB was amplified and co-expressed with KDC in E. coli. The overexpression of the YjgB gene resulted in an approximately 110% increase in 2-PE production up to 180.9 ± 4.23 mg/L (Table 3). An aldehyde reductase ADH6 gene from S. cerevisiae, which shows high activity for the various shortchain alcohols production (Larroy et al. 2002), was also co-expressed with KDC and resulted in a small increase of 2-PE (Table 3). This result suggests that phenylacetaldehyde is not a good substrate for ADH6.  Figure S1). Induction of the engineered pathway in recombinant E. coli in a modified M9 media resulted in the accumulation of 2-PEAc with a titer of 53.7 ± 2.83 mg/L (Table 3).

| Increase 2-PEAc production by adding the precursor substrate phenylpyruvate
Phenylpyruvate is the precursor substrate for the biosynthesis of 2-PEAc. Therefore, we hypothesized that increasing phenylpyruvate availability would improve 2-PEAc production. In this study, we evaluated  Figure S2). This results suggested that phenylpyruvate availability is a bottleneck for the biosynthesis of 2-PEAc.

| The biosynthesis of 2-PEAc from L-phenylalanine
Although the precursor substrate phenylpyruvate can be de novo synthesized from glucose, the efficiency is quite low because of feedback regulations in many branched metabolic pathways (Manuel et al., 2009). Another strategy is transamination of L-phenylalanine to phenylpyruvate by aminotransferase. Several groups have demonstrated the biosynthesis of 2-PE from L-phenylalanine with a high titer by transamination (Kim, Cho, et al., 2014;Yin et al., 2015).
The gene ARO8 which encode aminotransferase was amplified from S. cerevisiae and introduced to 2-PEAc-producing strain. In this study, we evaluated the biosynthesis of 2-PEAc from L-phenylalanine (the range from 0.5 to 2 g/L) in engineered strain MG1655/pDG37.
The titer of 2-PE and 2-PEAc is presented in Figure S3.  (Table 4). A large amount of 2-PE failed to be effectively converted to 2-PEAc, suggesting that alcohol acetyltransferase ATF1 is a another bottleneck for the biosynthesis of 2-PEAc ( Figure S3).

| DISCUSSION
Flavor compounds are in high demand and widely used as an additive in food and cosmetics industry. The production of natural flavor compounds has recently attracted a great deal of interest and represents a challenging target for synthetic biology research (Kuo et al., 2014).
In order to meet the need of consumer preferences for natural flavor compounds, microbial synthesis method has become a very attractive alternative to the chemical production.  (Kang et al., 2014). However, a metabolic engineering approach for 2-PEAc production is rare.
Several enzymatic methods have been reported for the preparation of 2-PEAc by lipase (Kuo, Liu, Chen, Chang, & Shieh, 2013;Kuo et al., 2012). However, there are some drawbacks by using enzymes in bioprocesses such as the need of long and complicated steps for enzyme isolation and purification. In this study, a fermentative route was created using an engineered E. coli for biosynthesis of 2-PEAc from L-phenylalanine based on ATF1. The results suggest that approximately 65% of L-phenylalanine was utilized toward 2-PEAc and 2-PE biosynthesis and thus demonstrate potential industrial applicability of this microbial platform.
GC/MS analysis of extracts from the strain of our study revealed that a large number of 2-PE failed to be effectively converted to 2-PEAc, suggesting that ATF1 is a bottleneck for further improvement of 2-PEAc biosynthesis. Thus, improving of the catalytic activity of ATF1 through protein engineering or identification a better alcohol acetyltransferase to replace ATF1 in the future is a promising strategy for enhancing 2-PEAc production in genetically engineered microorganisms.

ACKNOWLEDGMENTS
This study was funded by National Natural Science Foundation of China (81460312), and Foundation of Jiangxi Educational Committee (GJJ150991).
T A B L E 4 2-PE and 2-PEAc production in engineered strain MG1655/pDG37 (co-expression of aro8, kdc, yjgB and atf1) with modified M9 medium containing 0.5, 1.0, or 2.0 g/L of Lphenylalanine in shake flasks for 28 hr

Production (mg/L)
The concentration of L-phenylalanine 0.5 g/L 1 g/L 2 g/L All experiments were performed in triplicate and error bars show S.D.