Reconstruction of tyrosol synthetic pathways in Escherichia coli

Accepted Manuscript Reconstruction of tyrosol synthetic pathways in Escherichia coli

Cui Yang, Xianzhong Chen, Junzhuang Chang, Lihua Zhang, Wei Xu, Wei Shen, You Fan PII: DOI: Reference:

S1004-9541(17)31639-7 doi:10.1016/j.cjche.2018.04.020 CJCHE 1130

To appear in: Received date: Revised date: Accepted date:

23 November 2017 17 March 2018 15 April 2018

Please cite this article as: Cui Yang, Xianzhong Chen, Junzhuang Chang, Lihua Zhang, Wei Xu, Wei Shen, You Fan , Reconstruction of tyrosol synthetic pathways in Escherichia coli. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Cjche(2018), doi:10.1016/j.cjche.2018.04.020

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Biotechnology and Bioengineering

Reconstruction of tyrosol synthetic pathways in Escherichia coli☆ Cui Yang 1,2, Xianzhong Chen1,2*, Junzhuang Chang 1,2, Lihua Zhang1,2, Wei Xu, Wei

Key Laboratory of Carbohydrate Chemistry & Biotechnology, Ministry of Education,

RI

1

PT

Shen1, You Fan1,2

☆

Supported by the

NU

School of Biotechnology, Jiangnan University, Wuxi 214122, China

Priority Academic Program Development of Jiangsu Higher

MA

2

SC

Jiangnan University, Wuxi 214122, China

Education Institutions, the Fundamental Research Funds for the Central Universities

PT E

D

(JUSRP51611A, JUSRP51504), the Natural Science Foundation of Jiangsu province

CE

(BK20171138), 863 program（2013AA102101-5）and the 111 Project (No.1112-06).

AC

*Correspondence author: Xianzhong Chen: [email protected], Tel: +86-510-85918122, fax: +86-510-85918122 Address: School of Biotechnology, Jiangnan University, 1800 Lihu Avenue, Wuxi 214122, China

1

ACCEPTED MANUSCRIPT Abstract: Tyrosol is a pharmacologically active phenolic compound widely used in the medicine and chemical industries. Traditional methods of plant extraction are complicated and chemical synthesis of tyrosol is not commercially viable. In this

PT

study, a recombinant E. coli strain was constructed by overexpressing the

RI

phenylpyruvate decarboxylase ARO10 from Saccharomyces cerevisiae, which could

SC

produce tyrosol from glucose. Furthermore, genes encoding key enzymes from the competing phenylalanine and tyrosine synthesis pathways and the repression protein

NU

TyrR were eliminated, and the resulting engineered strain generated 3.57 mmol·L-1

MA

tyrosol from glucose. More significantly, codon optimization of ARO10 increased expression and tyrosol titer. Using the novel engineered strain expressing

D

codon-optimized AR10 in shake-flask culture, 8.72 mmol·L-1 tyrosol was obtained

PT E

after 48 h. Optimization of the induction conditions improved tyrosol production to 9.53 mmol·L-1 (1316.3 mg·L-1). A higher titer of tyrosol was achieved by

CE

reconstruction of tyrosol synthetic pathway in E. coli.

AC

Keywords: Tyrosol; Escherichia coli; Phenylpyruvate decarboxylase; Gene knockout; Codon optimization

2

ACCEPTED MANUSCRIPT INTRODUCTION Tyrosol (2-(4-hydroxyphenyl) ethanol) is a pharmacologically active phenolic compound that is widely distributed in nature. Tyrosol and its derivatives, hydroxytyrosol and salidroside, have antioxidant activity and cardioprotective effects

PT

[1]. Recent research found that tyrosol can protect streptozotocin-induced diabetic rats

RI

through altered glycoprotein components; further, this study can be extrapolated to

SC

humans [2]. Tyrosol also has taste-sharpening effects that play an important roles in the taste of alcoholic beverages, particularly in sake [3] and wine [4]. A recent

NU

discovery suggests that differences in taste, astringency and bitterness, characteristic

MA

of tyrosol, might derive from the tyrosol level in sake [3]. Therefore, tyrosol has recently gained considerable attention as a pharmacological chemical in industry and

D

as a taste-sharpening compound in the taste and functionalities of alcohol beverages.

PT E

With the discovery of more and more health benefits offered by tyrosol, increasing effort has been expended to increase the yield of tyrosol. Because the cost of

CE

extracting tyrosol from nature sources is very high, tyrosol used for industrial

AC

purposes is often synthesized chemically [5]. However, the chemical synthesis of tyrosol is plagued by purification problems and very low overall yield. Tyrosol occurs naturally in olive oil [6], wine [1], sake [7], and some plant tissues. The tyrosol content of virgin olive oil ranges from 40 to 180 mg·(kg oil)-1 [5]. However, extracting tyrosol from olive oil is challenging because tyrosol co-localizes with a large number of other phenolic substances in olive oil [5, 8]. There are two pathways for tyrosol synthesis in microbes and plants [8]. One is the Ehrlich pathway, in which

ACCEPTED MANUSCRIPT tyrosine is converted into tyrosol as follows: (1) transamination of tyrosine by aminotransferase to form 4-hydroxyphenylpyruvate (4HPP), (2) decarboxylation of 4HPP by pyruvate decarboxylase to form 4-hydroxyphenylacetaldehyde (4HPAAL), and (3) reduction of 4HPAAL by alcohol dehydrogenase (ADH) to form tyrosol

PT

(Fig.1) [8-10]. Another pathway involves the conversion of L-tyrosine into tyramine

RI

by tyrosine decarboxylase and transformation of the resulting tyramine into tyrosol by

SC

the consecutive actions of tyramine oxidase and alcohol dehydrogenase (ADH)[9][11] [12]. Recently, microbial production of tyrosol has been explored using engineered E.

NU

coli strains [8]. Initial attempts at microbial production yielded a low tyrosol titer, so a

MA

metabolic engineering strategy is need to increase tyrosol production to the level needed for use on an industrial scale.

D

In this study, we constructed a tyrosol-producing E. coli strain and used metabolic

PT E

engineering to improve product titer. After codon usage and induction conditions were optimized, a yield of 9.53 mmol·L-1 (1316.3 mg·L-1) was obtained. This lays the

AC

CE

foundation for the industrialization of microbial tyrosol synthesis.

ACCEPTED MANUSCRIPT HO

OH

HO

HO

O

OH HO

OH

Glucose O

Tyrosol

O O

P

O

OH

O

OH

HO

O

HO

P

OH

H OH

OH

CH2

OH

OH

Er ythrose 4-phosphate

Tyrosol Tyr R

AroG

ADH

HO

HO

O

HO

Tyr A

O

OH O

O

Chor ismate

4HPP

PheA

O

4HPAAL

Tyr B

SC

HO O OH H2N

FeaB

O

OH

HO OH

O

H2N

NU

Phenylalanine

RI

DAHP

ARO10

OH

O HO

PT

Phosphoenolpyr uvate

Tyrosine

4-hydroxyphenylacete

MA

Figure 1. The engineered tyrosol pathway in E. coli. AroG, 3-deoxy-D-arabino-heptulosonate-7-phosphate synthase; TyrR, autorepressor; DAHP,

D

3-deoxy-arabino-heptulonate7-phosphate; TyrA, chorismate mutase or prephenate

PT E

dehydrogenase; PheA, chorismate mutase and prephenate dehydratase; 4HPP, 4-hydroxyphenylpyruvate; TyrB, aromatic amino acid transaminase; ARO10,

CE

phenylpyruvate decarboxylase; 4HPAAL, 4-hydroxyphenyacetaldehyde; FeaB,

AC

phenylacetaldehyde dehydrogenase; ADH, alcohol dehydrogenase. Dashed arrows indicate feedback inhibitions. Tandem dashed arrows indicate multiple enzyme reactions. Red cross bars denote the disruption of indicated genes. MATERIALS AND METHODS Genetic material, microbial hosts and plasmids The bacterial strains and plasmids used in this study are listed in Table 1. E. coli JM109 was used as the host for plasmid construction and E. coli MG1655 and its

ACCEPTED MANUSCRIPT derivatives were used for tyrosol production. The plasmids pKK223-3-ARO10 and pKK223-3-ARO10* were used as expression plasmids. The plasmids pKD46, pKD13 and pCP20 (Table 1), which are involved in the λ Red recombinase expression system[13] and were previously stored in our lab, were used in this study.

PT

Table 1 Strains and plasmids used in this study. Relevant genotype and

Source or

CGSC 6300 Reference

CMG strains

Escherichia coli MG1655 characteristics

CMGF

CMG ΔfeaB

CMGP

CMGF ΔpheA

CMGB

CMGP ΔtyrB

RI

Plasmids and

This study

SC

This study This study

CMGB ΔtyrR

CMG0

CMG with pKK223-3

CMGA

CMG with

This study

CMGFA

CMGF with pKK223-3-ARO10

This study

CMGPA

CMGP with pKK223-3-ARO10

This study

MA

NU

CMGR

CMGBA CMGRA CMGRA*

pKD13

PT E

pCP20

This study

CMGB with pKK223-3-ARO10

This study

CMGR with pKK223-3-ARO10

This study

CMGR with pKK223-3-ARO10

This study

R Amp , helper plasmid pKK223-3-ARO10*

CGSC

R

Amp , helper plasmid

CGSC

AmpR, helper plasmid

CGSC

D

pKD46

This study

R

pKK223-3

Amp

pKK223-3-ARO

AmpR, EcoRI

This study

CGSC

pKK223-3-ARO 10

AmpR, EcoRI, HindⅢ

This study

CE

10*

Pathway and plasmid construction.

AC

The primers used to amplify DNA fragments of targeted genes are listed in Table 2. The artificial tyrosol production pathway constructed in MG1655, is shown in Fig.1. The plasmids constructed during this study were confirmed by restriction enzyme digestion (Table 1). All restriction enzymes and DNA ligase were purchased from TaKaRa (Dalian, China).

ACCEPTED MANUSCRIPT The endogenous E. coli phenylacetaldehyde dehydrogenase gene feaB (Gene ID: 945933), the chorismate mutase and prephenate dehydratase gene pheA (Gene ID: 947081), the tyrosine aminotransferase gene tyrB (Gene ID: 948563), and the aromatic amino acid biosynthesis and transport regulon transcriptional regulator gene

PT

tyrR (Gene ID: 945879) were successively knocked out of the E. coli MG1655

RI

chromosome by using the classical λ Red homologous recombination method [14].

SC

hese knockouts yielded strains CMGF, CMGP, CMGB, and CMGR, respectively. Deletion of the feaB gene is described as an example of the procedure. The feaB

NU

fragment knockout expression cassette FRT-Km-FRT, which contains a kanamycin

MA

resistance marker flanked by arms homologous with feaB, was amplified using primers NfeaBU and NfeaBD. Km-resistant strains were selected after using the λ

D

Red homologous recombination method. Insertion of the FRT-Km-FRT cassette was

PT E

confirmed using colony-direct PCR with Ex Taq DNA polymerase and locus-specific primers YfeaBU and YfeaBD (Table 2), according to methods detailed in previous

CE

studies [15]. The FLP recombination target, the (FRT)-flanked antibiotic resistance

AC

gene used for selection, was eliminated by induction of the pCP20 vector, which is a temperature-conditional plasmid that expresses FLP recombinase from a heat-inducible promoter. The resulting strain, CMGF, was verified using primers YfeaBU and YfeaBD. The other genes (pheA, tyrB, and tyrR) were deleted using similar strategies. The wild-type phenylpyruvate decarboxylase gene ARO10 was amplified from S. cerevisiae EBY100 (Invitrogen) using primers ARO10U and ARO10D (Table 2). To

ACCEPTED MANUSCRIPT improve ARO10 expression, optimization was performed by replacing codons predicted to be less frequently used in E. coli with more favored codons according to the JCat [16]. The wild-type ARO10 gene and optimized gene ARO10* (Supplemental material S1) were inserted into plasmid pKK223-3 at the EcoRI and HindIII sites and

PT

generated pKK223-3-ARO10 and pKK223-3-ARO10*, respectively. The strong

RI

promoter tac was used in the regulation of expression of desired gene.

Name

Oligonucleotide sequence (5'–3')

SC

Table 2 Oligonucleotides used in this study

Relevant characteristic

ATGACAGAGCCGCATGTAGCAGTATTAAGCCAGGTCCAACAGTTTCTCGAGTGTGGCTGGA NfeaBU

NU

GCTGCTTC

feaB upstream

TTAATACCGTACACACACCGCTTAGTTTCACACCAACCGTCCAGCCAGTATTCCGGGGATCC NfeaBD GTCGACC

feaB downstream

AGGCAACACTATGACATCGGAAAACCCGTTACTGGCGCTGCGAGAGAAAAGTGTGGCTGG

MA

NpheAU AGCTGCTTC

pheA upstream

TCAGGTTGGATCAACAGGCACTACGTTCTCACTTGGGTAACAGCCCAATAATTCCGGGGAT NpheAD CCGTCGACC

pheA downstream

GTCTGTACTACAACGAAGACGGAATTATTCCACAACTGCAAGCCGTGGCGGTGTAGGCTGG

D

NtyrBU AGCTGCTTC

tyrB upstream

CGTCGACC

PT E

TTACATCACCGCAGCAAACGCCTTTGCCACACGTTGTACATTTGCCGTATATTCCGGGGATC NtyrBD

tyrB downstream

TTCAGCAGTCTGATGGCCGAAATACGCCGTATTGCGGGTGTTACCGATGTGTGTAGGCTGG NtyrRU AGCTGCTTC

tyrR upstream

CE

AGATTGAGCGTCAACACGTTCAGACGATAATAGAGATCTTCACGGAACATATTCCGGGGAT NtyrRD

tyrR downstream

CCGTCGACC

GTCGCTGCCTTTTACTTCCT

YfeaBD

ACCGGTCTTGGACAATACAG

YpheAU

CGCGTCCTTTATATTGAGTGTATCGC

YpheAD

TGCTGCGTTTTCAGAGTGAAAGC

pheA downstream validation

YtyrBU

TGCTGCGTTTTCAGAGTGAAAGC

tyrB Upstream validation

YtyrBD

TATTTCACTGCA GGCTGGGTAG

tyrB downstream validation

YtyrRU

ATCCCATTGGGCGAATCTAC

tyrR Upstream validation

YtyrRD

ACAGCTCAGTTAACGGCATG

tyrR downstream validation

AC

YfeaBU

ARO10U

CATGCCATGGATGGCACCTGTTACAATTG

ARO10D

CGGGCGCGCGGATCCCTATTTTTTATTTCTT

Culturing of the recombinant E. coli strains.

feaB Upstream validation feaB downstream validation pheA Upstream validation

NcoI BamHI

ACCEPTED MANUSCRIPT All strains were cultivated in LB broth medium (10 g·L-1 tryptone, 5 g·L-1 yeast extract, 10 g·L-1 NaCl) and M9Y medium, which contained 1×M9 minimal salts (Na2HPO4·12H2O, 17.1 g·L-1; KH2PO4, 3.0 g·L-1; NaCl, 0.5 g·L-1; NH4Cl, 1.0 g·L-1) and 2% (w/v) glucose, and was supplemented with 0.025% (w/v) yeast extract, 5

PT

mmol·L-1 MgSO4, and 100 mg·ml-1 ampicillin.

RI

A single colony of the indicated E. coli strain was used to inoculate 20 ml liquid LB

SC

medium containing appropriate antibiotics and allowed to grow overnight at 37 ℃. The overnight culture was then diluted 1:100 with 50 ml fresh LB medium and

NU

incubated in a gyratory shaker incubator at 37 ℃ and 200 r·min-1. When the

MA

absorbance of the culture (measured at 600 nm) reached approximately 0.6–0.8, isopropyl β-D-thiogalactopyranoside (IPTG) was added to the medium to reach a final

D

concentration of 0.6 mmol·L-1. After the IPTG was added, the culture was incubated

PT E

at 30 ℃ for 8 h to induce recombinant protein expression. Subsequently, the cells were collected by centrifugation, resuspended in 50 ml M9Y medium, and cultured at

CE

30 ℃ for 48 h for the extraction of compounds and further analysis. The shake-flask

AC

experiments were conducted in triplicate. Analysis of tyrosol and byproducts from the recombinant E. coli strains. The tyrosol and byproducts, the pyruvic acid, the lactic acid and the acetic acid, content of the fermentation medium was analyzed using high-performance liquid chromatography (HPLC) with diode array detection or using HPLC with a mass spectrometric (MS) detector (HPLC-MS).

ACCEPTED MANUSCRIPT For the analysis of tyrosol: HPLC analysis was performed with an Agilent 1260 system (Agilent Technologies Inc., Santa Clara, CA) equipped with an Agela Innoval C18 column (4.6 × 250 mm; particle size, 5 mm) and a diode array spectrophotometer (Agilent Technologies). A 10 μl sample of the fermentation supernatant was applied

PT

to the column, which was eluted at room temperature with a mobile phase containing

RI

80% solvent A (0.1% methanoic acid in H2O) and 20% solvent B (methanol) at a flow

SC

rate of 1 ml·min-1. The products were detected at 276 nm. B. Under these conditions, the retention time of a tyrosol standard was 10.28 min. To quantify tyrosol in the

NU

culture medium, calibration curves were generated with a series of known

MA

concentrations of the tyrosol standard dissolved in culture medium. The R2 coefficients for the calibration curves were >0.999.

D

For the analysis of byproducts: HPLC analysis was performed with an Agilent 1260

PT E

system (Agilent Technologies Inc., Santa Clara, CA) equipped with an Bio-Rad Aminex HPX-87H column (300 mm×7.8 mm) and the detector used for the detection

CE

was RID-10A. A 10 µl sample of the fermentation supernatant was applied to the

AC

column, which was eluted at room temperature with a mobile phase was 5 mmol·L-1 H2SO4, and the flow rate was 0.5 ml·min-1, column temperature 65°C. Under these conditions, the retention time of the pyruvic acid standard was 11.18 min; the retention time of the lactic acid standard was 15.89 min; the retention time of the acetic acid standard was 17.12 min. To quantify byproducts in the culture medium, calibration curves were generated with a series of known concentrations of the

ACCEPTED MANUSCRIPT byproducts standard dissolved in culture medium. The R2 coefficients for the calibration curves were >0.999. HPLC-MS analysis was performed with a Waters Acquity UPLC system that included a BEH C18 column (2.1 × 150 mm × 1.7 μm), a Waters Acquity photodiode array

PT

detector, and a Waters MALDI Synapt Q-TOF MS detector equipped with an

RI

electrospray ionization probe. Samples were analyzed at 45 °C using a gradient

SC

elution protocol at a flow rate of 0.3 ml·min-1 as follows. For the first 0.1 min, the mobile phase consisted of 10% solvent A (acetonitrile) and 90% solvent B (0.1%

NU

formic acid in H2O). From 0.1 to 5 min, a linear gradient from 10% A and 90% B to

MA

20% B was used; from 5 to 7 min, a linear gradient from 80% A and 20% B to 100% A was used; from 7 to 7.1 min a linear gradient from 100% A to 10% A and 90% B

D

was used; and from 7.1 to 10 min, the mobile phase was maintained at 10% A and

CE

RESULTS

PT E

90% B. The products were detected at 276 nm and identified using MS.

AC

Creation of a tyrosol biosynthesis pathway in E. coli MG1655 E. coli MG1655 does not synthesize detectable amounts of tyrosol, not does E. coli MG1655 harboring the empty vector pKK223 (strain CMG0; data not shown). The tyrosol biosynthesis pathway created in E. coli MG1655 (Fig.1) begins with 4HPP, the critical precursor of tyrosine in the aromatic amino acid biosynthesis pathway. To construct a biorthogonal pathway catalyzing the conversion of 4HPP to tyrosol, the S. cerevisiae ARO10 gene, which encodes phenylpyruvate decarboxylase, was inserted

ACCEPTED MANUSCRIPT into pKK223-3 and the resulting plasmid was inserted into E. coli MG1655 to form strain CMGA. The CMGA strain could produce 0.05 mmol·L-1 tyrosol (Fig.2). The result showed that overexpression of ARO10 combined with endogenous ADHs could

PT E

D

MA

NU

SC

RI

PT

convert 4HPP into tyrosol using glucose as substrate.

CE

Figure 2. The influence of gene deletions on tyrosol production. All strains were derived from Escherichia coli MG1655. Data are averages of results for three

AC

biological replicates. Error bars represent the standard deviation of three sets of parallel averages.

Engineering aromatic amino acids pathways to improve tyrosol yield. It has been reported that 4HPAAL is reduced to tyrosol by the endogenous ADHs in E. coli (Fig.1) [12]. However, this intermediate compound could also be oxidized to

ACCEPTED MANUSCRIPT 4HPA by endogenous phenylacetaldehyde dehydrogenases (FeaB or PadA) [17]. Therefore, the feaB gene was deleted to evaluate its effect on the tyrosol production. We found that the engineered strain CMGFA (Escherichia coli MG1655ΔfeaB harboring pKK223-3-ARO10) could produce 0.52 mmol·L-1 tyrosol in the M9Y for

RI

feaB had a positive effect on the accumulation of tyrosol.

PT

48 h, which was 10-fold the amount produced by CMGA (Fig.2). Thus, knockout of

SC

Chorismate is a very important node in the biosynthesis of phenylalanine, tyrosine and tyrosine [18, 19]. To reduce carbon flux from chorismate toward phenylalanine,

NU

the pheA gene, which encodes chorismate mutase and prephenate dehydratase, was

MA

knocked out of the genome of CMGF to produce CMPG. After introducing pKK223-3-ARO10, the resulting engineered strain CMGPA was employed to

D

produce tyrosol. A yield of 2.48 mmol·L-1 tyrosol, which is 49.6-fold greater than that

PT E

obtained with the control strain CMGA, was obtained (Fig.2). Furthermore, our observations indicated that the combined deletion of feaB and pheA resulted in an

CE

obviously synergistic effect on tyrosol synthesis.

AC

To reduce tyrosine formation from the tyrosol precursor 4HPP, the tyrB gene was deleted from the genome of the double knockout strain of feaB and pheACMGP to form strain CMGB. After inserting pKK223-2-ARO10, the resulting strain CMGBA yielded 3.26 mmol·L-1 tyrosol (Fig.2), which was higher 31.5% than that obtained with CMGPA. In our opinion, cumulative elimination of competing pathways increased the efficiency of tyrosol production and increased its accumulation in the culture medium.

ACCEPTED MANUSCRIPT Effect of elimination of repression protein TyrR on tyrosol synthesis. In E. coli, tyrR encodes a repressor of the aromatic amino acid pathway. The activity of the TyrR protein is modulated by binding to one or more of the aromatic amino acids [20]. The consensus DNA sequence for TyrR-binding sites, which are referred

PT

to as TyrR boxes, is TGTAAAN6TTTACA.[20, 21] TyrR exists as a dimer in solution,

RI

but in the presence of ATP and tyrosine it can restrain the transcription of tyrB, aroP,

SC

aroL-aroM, aroF-tyrA and tyrP[21, 22]. Therefore, it seemed reasonable to hypothesize that its presence in our system had a negative effect on the tyrosol yield.

NU

To test this hypothesis, the tyrR gene was deleted from CMGB to produce strain

MA

CMGR. After introducing pKK223-3-ARO10 to form CMGRA, induction of ARO10 overexpression and incubation in M9Y for 48 h, CMGRA could produce 3.57

D

mmol·L-1 tyrosol (Fig.2), which was 9.5% higher than that of CMGBA (Fig.2). These

PT E

results indicated that tyrR knockout had a positive effect on the accumulation of

CE

tyrosol.

Moreover, the effect of gene deletion on the byproducts and cell growth were

AC

investigated. We found that deletion of feaB, pheA and tyrR genes led to the little impact on the byproducts accumulation (Fig.3), however elimination of tyrB gene could decrease the pyruvic acid content in the final broth (48 h) (Fig.3a). More importantly, deletion of tyrB gene decreased the lactic acid accumulation significantly (Fig.3b). In our opinion, due to that TyrB catalyze 4HPP to tyrosine (Fig.1), deletion of tyrB gene could reduce tyrosine production and release the feedback repression, thereby increasing the tyrosol flux from pyruvate. Also, the deletion of the indicated

ACCEPTED MANUSCRIPT genes gave the very slight effect on the acetic acid accumulation (Fig.3c). Fig.3d showed that accumulative deletion of the genes defected the growth performance. The engineered strain of CMGRA (Escherichia coli MG1655 ΔfeaB ΔpheA ΔtyrB ΔtyrR with pKK223-3-ARO10) exhibited the lowest biomass content compared the control

AC

CE

PT E

D

MA

NU

SC

RI

PT

strain and other engineered strains (Fig.3d).

Figure 3. The influence of gene deletions on byproducts production and the cell growth. Broth samples were taken at 24 h, 36 h and 48 h and pyruvic acid (a), lactic acid (b) and acetic acid (c) were detected, respectively. Cell density was also monitored (d). Error bars represent the standard deviation of three sets of parallel averages.

ACCEPTED MANUSCRIPT Enhancement of tyrosol production by codon-optimized. The ARO10 gene was cloned from S. cerevisiae. Many studies have shown that codon preference in bacteria is significantly different than that in eukaryotes. This is particularly true for E. coli and S. cerevisiae [23]. Our codon analysis showed that

PT

ARO10 possesses a number of codons that are rarely used in the E. coli genome.

RI

SDS-PAGE analysis also indicated that ARO10 was expressed at a low level (Fig.4a).

SC

Therefore, ARO10 was redesigned (optimized) to include codons preferred by E. coli, and then artificially synthesized (Supplemental material S1). The resulting gene,

NU

ARO10*, was inserted into pKK223-3 and the resulting plasmid was inserted into

MA

CMGR to generating the recombinant strain CMGRA*. SDS-PAGE analysis showed that codon optimization significantly improved the foreign gene expression level,

D

compared with expression of the wild-type gene (Fig.4a). Furthermore, tyrosol

PT E

production by the two engineered strains CMGRA and CMGRA*was compared. The tyrosol yield of CMGRA* reached 8.72 mmol·L-1, which was 2.44-fold greater than

CE

the tyrosol yield CMGRA (Fig.4b). These results demonstrate that codon optimization

AC

of ARO10 significantly improved its expression in CMGR, leading to significantly improved tyrosol production.

PT

ACCEPTED MANUSCRIPT

RI

Figure 4.Improvement of tyrosol production by codon-optimization of ARO10 gene.

SC

(a) SDS-PAGE analyses of ARO10* and ARO10 expression. M represents the

NU

molecular marker, lane 1, 2 and 3 indicate the strains CMG0 (pKK223-3), CMGA (pKK223-3-ARO10) and CMGA* (pKK223-3-ARO10*), respectively. (b) The

MA

influence of codon optimization on tyrosol production. Strain CMGRA harbors the

D

vector pKK223-3-ARO10; strain CMGRA* harbors the vector pKK223-3-ARO10*.

PT E

ARO10* is the codon-optimized version of ARO10. Data are averages of results for three biological replicates. Error bars represent the standard deviation of three sets of

CE

parallel averages.

AC

Optimization of the tyrosol production process Induction conditions were optimized to improve the efficiency of tyrosol synthesis using CMGRA*. Firstly, we evaluated the effect of the growth phase of the cells being induced on tyrosol production. The results showed that tyrosol production was greater when ARO10* expression was induced after 8 h of growth (OD600 0.6-0.8) than when expression was induced after 4, 16, or 24 h (Fig.5a). Next, we assessed the

ACCEPTED MANUSCRIPT effect of induction temperature on tyrosol production. Four different temperatures (20, 25, 30, and 37 °C) were chosen for the induction phase. The production of tyrosol by CMGRA* was greatest (9.21 mmol·L-1) when the induction was conducted at 30 °C (Fig.5b). Finally, the effect of inducer concentration on tyrosol production was

PT

investigated. IPTG was added to final concentrations of 0.1, 0.2, 0.4, 0.6, 0.8, 1.0, and

RI

1.2 mmol·L-1. Tyrosol production by CMGRA* was greatest with an IPTG

SC

concentration of 0.6 mmol·L-1 (Fig.5c). As shown in Fig.5c, the production of tyrosol from glucose by CMGRA* reached 9.53 mmol·L-1 under optimal culture conditions

NU

(induction with 0.6 mmol·L-1 IPTG at an OD600 of 0.6−0.8 and maintaining the

MA

temperature at 30 °C during the induction phase). Finally, we repeated the fermentation experiment under the optimal conditions, and assessed the relationship

D

between tyrosol yield and the bacterial OD during the fermentation process. The

PT E

experiment demonstrated that the fermentation was reproducible and the final tyrosol titer reached 9.53 mmol·L-1 (1316.3 mg·L-1) . Therefore, a higher titer of tyrosol was

AC

CE

achieved by reconstruction of tyrosol synthetic pathway in E. coli.

RI

PT

ACCEPTED MANUSCRIPT

SC

Figure 5. Determination of optimal culture conditions for the strain CMGRA*. (a)

NU

Effect of induction time on tyrosol titer at an OD600 of 0.6, an IPTG concentration of 0.2 mmol·L-1, and a temperature of 25 °C. (b) Effect of induction temperature on

MA

tyrosol production at an IPTG concentration of 0.2 mmol·L-1 and an induction time of

D

8 h. (c) Effect of IPTG concentration on tyrosol production at an OD600 of 0.6 and a

PT E

temperature of 30 °C. (d) Evaluation of tyrosol production under the optimal conditions using the recombinant strain CMGRA*. Data are the average experiments

CE

performed in triplicate. Error bars represent standard deviations from the mean. Data are averages of results for three biological replicates. Error bars represent the standard

AC

deviation of three sets of parallel averages. HPLC and HPLC-MS analysis of a sample of the material produced in this experiment confirmed that the substance was tyrosol (Fig.6a and Fig.6b). a

%

100

1: TOF MS ESA2.31e3

119.1 138.1

50

m/z 100 150 200 250 300 350 400

NU

0

273.2

RI

137.1

SC

standard-1 C8H10O2 b 20170418-18 182 (3.155)

PT

ACCEPTED MANUSCRIPT

MA

Figure 6. (a) HPLC analysis of tyrosol production in the fermentation supernatant of recombinant strain CMGRA*. CMGRA*, E. coli MG1655 (ΔfeaB ΔpheA ΔtyrB

PT E

D

ΔtyrR) harboring plasmid pKK223-3-ARO10*; CMGRA, E. coli MG1655 (ΔfeaB ΔpheA ΔtyrB ΔtyrR) harboring plasmid pKK223-3-ARO10; standard, authentic

CE

tyrosol. (b) Mass spectrum of tyrosol from the supernatant of a CMGRA* fermentation. Three biological replicates were performed and the representative data

AC

are presented.

DISCUSSION E. coli has become a promising host for the microbial production of a variety of valuable chemicals from renewable resources [24]. Engineering E. coli for tyrosol production has been explored in recent years. Two pathways for tyrosol synthesis

ACCEPTED MANUSCRIPT were evaluated: through expression of ARO10 and the indole-3-pyruvate decarboxylase gene ipdC [12, 25], and through expression of tyrosine decarboxylase and tyrosine oxidase [8]. More recently, we constructed a tyrosol-producing E. coli BL21(DE3) system that was able to effectively produce 4.15 mmol·L-1 tyrosol from

PT

glucose [9]. This engineered strain, when used as a biocatalyst, could successfully

RI

convert 10 mmol·L-1 tyrosine to 8.71 mmol·L-1 tyrosol, achieving a conversion rate of

SC

87.1%. In this study, a more efficient tyrosol-producing engineered strain was constructed. This strain, CMGRA*, was able to produce 9.53 mmol·L-1 (1316.3

NU

mg·L-1) tyrosol from glucose. To the best of our knowledge, this is the tyrosol

MA

production ever reported using E. coli as the host.

The elimination of competing pathways is useful strategy for improving bacterial

D

production of interesting chemicals [26]. In this study, we found that deleting both

PT E

pheA and feaB significantly increased tyrosol production. Deleting tyrB, which encodes tyrosine aminotransferase, further increased the tyrosol yield. Together, these

AC

65-fold.

CE

three deletions, which eliminate competing pathways, increased tyrosol production by

In contrast, deletion of tyrR, which eliminates potential repression of the tyrosol biosynthesis pathway, had a minimal effect on tyrosol production (9.1% increase). This suggests that eliminating the competing pathways lowered the concentrations of the aromatic amino acids, making repression by TyR less of an issue. Subsequently, we assessed whether the strain CMGB was tyrosine deficient. Interestingly, the triple-deletion mutant CMGB was not tyrosine deficient; however, it was

ACCEPTED MANUSCRIPT phenylalanine deficient. A plausible explanation for this is that deletion of pheA had a greater effect on phenylalanine biosynthesis than deletion of tyrB had on tyrosine biosynthesis. It seems likely that an alternative enzyme, like aspartate aminotransferase or histidine aminotransferase (AspC or HisC) may be able to

PT

partially compensate for the loss of TyrB [27].

RI

Codon optimization of ARO10 led to a significant, 2.4-fold increase in tyrosol

SC

production. Our data are clearly consistent with the well-known positive effect of codon optimization on gene expression in E. coli. In this particular instance, codon

NU

optimization also had a crucial effect on tyrosol production, suggesting that

MA

phenylpyruvate decarboxylase activity was limiting tyrosol production. Future work pathway engineering should take into consideration the issue of codon usage.

PT E

D

In another study, the indole-3-pyruvate decarboxylase gene (ipdC) was introduced into the genome of E. coli BW25113 (DE3), which overproduces tyrosine, and then the feaB and pheA genes were deleted to generate a recombinant tyrosol producer [25].

CE

The resulting strain produced 8.3 mmol·L-1 tyrosol from 1% glucose. Although the

AC

yield of tyrosol in this study was not as high as that seen with the production system described in this report, they integrated ipdC into the bacterial genome by homologous recombination. Thus, the strain was no longer subject to the growth pressure caused by the plasmid, nor did they have to add antibiotics to prevent plasmid loss. In future studies, we will pursue the integration of key enzymes into the bacterial chromosome to avoid the use of inducers and antibiotics, which should smooth the path toward the industrialization of novel biocatalysts

ACCEPTED MANUSCRIPT References [1]

Chung D, Kim S.Y, Ahn J.H. Production of three phenylethanoids, tyrosol, hydroxytyrosol, and salidroside, using plant genes expressing in Escherichia coli, Sci Rep. 7 (2017) 2578.

[2]

Chandramohan R, Saravanan S, Pari L, Beneficial effects of tyrosol on altered glycoprotein components in streptozotocin-induced diabetic rats. Pharm Biol.55 (2017) 1631-1637.

[3]

Soejima H, Tsuge K, Yoshimura T, Sawada K, Kitagaki H, Breeding of a high tyrosol-producing sake yeast by isolation of an ethanol-resistant mutant from a trp3 mutant. J.

[4]

PT

Inst Brew. 118 (2012) 264–268. Silva L.R, Andrade P.B, Valentao P, Seabra R.M, Trujillo M.E, Velazquez E, Analysis of non-coloured phenolics in red wine: Effect of Dekkera bruxellensis yeast. Food Chem. 89 [5]

RI

(2005) 185-189.

Dewapriya P, Li Y.X, Himaya S.W.A, Kim S.K. Isolation and characterization of

SC

marine-derived Mucor sp. for the fermentative production of tyrosol. Proc Biochem. 49 (2014) 1402-1408. [6]

Di B.R, Varì R, Scazzocchio B, Filesi C, Santangelo C, Giovannini C, Matarrese P, D'Archivio

NU

M, Masella R, Tyrosol, the major extra virgin olive oil compound, restored intracellular antioxidant defences in spite of its weak antioxidative effectiveness. Nutr Metab Cardiovasc Dis. 17 (2007) 535-545.

Allouche N, Damak M, Ellouz R, Sayadi S, Use of whole cells of Pseudomonas aeruginosa

MA

[7]

for synthesis of the antioxidant hydroxytyrosol via conversion of tyrosol. Appl Environ Microbiol. 148 (1992) 14-27. [8]

Satoh Y, Tajima K, Munekata M, Keasling J.D, Lee T.S, Engineering of a tyrosol-producing

D

pathway, utilizing simple sugar and the central metabolic tyrosine, in Escherichia coli. J. [9]

PT E

Agric Food Chem. 60 (2012) 979-984. Xue Y, Chen X, Yang C, Chang J, Shen W, Fan Y. Engineering Eschericha coli for enhanced tyrosol production. J. Agric Food Chem. 65 (2017) 4708-4714. [10]

Hazelwood L.A, Daran J.M, Van Maris A.J.A, Pronk J.T, Dickinson J.R, The ehrlich pathway

CE

for fusel alcohol production: a century of research on Saccharomyces cerevisiae metabolism. Appl Environ Microbiol. 74 (2008) 2259-2266. [11]

Sentheshanmuganathan S, Elsden S.R, The mechanism of the formation of tyrosol by

[12]

AC

Saccharomyces cerevisiae. J. Biochem. 69 (1958) 210-218. Bai Y, Bi H, Zhuang Y, Chang L, Tao C, Liu X, Zhang X, Tao L, Ma Y, Production of salidroside in metabolically engineered Escherichia coli. Sci Rep. 4 (2014) 6640. [13]

Datsenko K.A, Wanner B.L. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc Nati Acad Sci USA. 97 (2000) 6640-6645.

[14]

Baba T, Ara T, Hasegawa M, Takai Y, Okumura Y, Baba M, Datsenko K A, Tomita M, Wanner B.L, Mori H, Construction of Escherichia coli K-12 in-frame, single-gene knockout mutants: the Keio collection. Mol Syst Biol. 2 (2006) 2006.0008.

[15]

Baba T, Huan H.C, Datsenko K, Wanner B.L, Mori H, The Applications of Systematic In-Frame, Single-Gene Knockout Mutant Collection of Escherichia coli K-12. Methods Mol Biol. 416 (2008) 183-194.

ACCEPTED MANUSCRIPT [16]

Grote A, Hiller K, Scheer M, Münch R, Nörtemann B, Hempel D.C, Jahn D, JCat: a novel tool to adapt codon usage of a target gene to its potential expression host. Nucleic Acids Res. 33 (2005) 526-531.

[17]

Satoh Y, Tajima K, Munekata M, Keasling J.D, Lee T.S, Engineering of a Tyrosol-Producing Pathway, Utilizing Simple Sugar and the Central Metabolic Tyrosine, in Escherichia coli. J Agric Food Chem. 60 (2012) 979-984.

[18]

Smith, G.D, D.V. Roberts, A. Daday, Affinity chromatography and inhibition of chorismate mutase-prephenate dehydrogenase by derivatives of phenylalanine and tyrosine. Biochem J.. 165 (1977) 121. Hudson G.S, Davidson B.E, Nucleotide sequence and transcription of the phenylalanine and

PT

[19]

tyrosine operons of Escherichia coli K12. J. Mol Biol. 180 (1984) 1023.

Verger D, Carr P.T, Ollis D, Crystal structure of the N-terminal domain of the TyrR

RI

[20]

transcription factor responsible for gene regulation of aromatic amino acid biosynthesis and [21]

SC

transport in Escherichia coli K12. J Agric Food Chem. 60 (2012):979-984. Koyanagi T, Katayama T, Suzuki H, Kumagai H, Altered Oligomerization Properties of N316 Mutants of Escherichia coli TyrR. J Bacteriol. 190 (2008) 8238-8243. James P, Helen C, Ji Y, The TyrR regulon. Mol Microbiol. 55 (2005) 16-26.

[23]

Satapathy, S.S., et al., Discrepancy among the synonymous codons with respect to their

NU

[22]

selection as optimal codon in bacteria. DNA Res. 23 (2016) 441-449. Chen X.Z, Zhou L, Tian K.M, Kumar A, Singh S, Prior B.A, Wang Z.X, Metabolic

MA

[24]

engineering of Escherichia coli: a sustainable industrial platform for bio-based chemical production. Biotechnol Adv. 31 (2013) 1200-1223. [25]

Koma D, Yamanaka H, Moriyoshi K, Ohmoto T, Sakai K, Production of Aromatic

D

Compounds by Metabolically Engineered Escherichia coli with an Expanded Shikimate Pathway. Appl Environ Microbiol. 78 (2012) 6203-6216 Zhou L, Niu D.D, Tian K.M, Chen X.Z, Prior B.A, Shen W, Shi G Y, Singh S, Wang Z.X,

PT E

[26]

Genetically switched D-lactate production in Escherichia coli. Metab Eng. 14 (2012) 560-568. Deu E, Kirsch J,F, Cofactor-directed reversible denaturation pathways: the cofactor-stabilized Escherichia coli aspartate aminotransferase homodimer unfolds through a pathway that differs

CE

from that of the apoenzyme. Biochemistry. 46 (2007) 5819-5829.

AC

[27]