Efficient synthesis of tyrosol from L-tyrosine via heterologous Ehrlich pathway

Xiaobo Ruan,Sheng Zhang,Wei Song,Jia Liu,Xiulai Chen,Liming Liu,Jing Wu

1 School of Pharmaceutical Science,Jiangnan University,Wuxi 214122,China

2 State Key Laboratory of Food Science and Technology,Jiangnan University,Wuxi 214122,China

3 Tianrui Chemical Co.,Ltd,Department of Chemistry,Quzhou 324400,China

Keywords:Tyrosol L-tyrosine Ehrlich pathway Enzyme cascade In situ product removal

ABSTRACT For the efficient conversion of L-tyrosine (L-Tyr)to tyrosol,which is an aromatic compound widely used in the pharmaceutical and chemical industries,a novel four-enzyme cascade pathway based on the Ehrlich pathway of Saccharomyces cerevisiae was designed and reconstructed in Escherichia coli.Then,the expression levels of the relevant enzymes were coordinated using a modular approach and gene duplication after the identification of the pyruvate decarboxylase from Candida tropicalis (CtPDC) as the rate-limiting enzymatic step. In situ product removal (ISPR) strategy with XAD4 resins was explored to avoid product inhibition and further improve tyrosol yield.As a result,the titer and conversion rate of tyrosol obtained were 35.7 g·L-1 and 93.6%,respectively,in a 3-L bioreactor.Results presented here provide a potential enzymatic process for industrial production of tyrosol from cheap amino acids.

1.Introduction

Tyrosol(2-(4-hydroxyphenyl)ethanol),a natural phenolic compound,possesses antioxidant and anti-inflammatory activities [1].Hydroxytyrosol and salidroside,derivatives of tyrosol,are vital antioxidant and cardiovascular drugs widely used in the pharmaceutical industry [1,2].In addition,tyrosol can also be used as a food additive to improve flavor,such as in sake [3] and wine [4].Therefore,tyrosol is a promising chemical widely used in pharmaceutical and other industries.

Three methods have been developed to produce tyrosol,including plant extraction,chemical synthesis,and microbial synthesis.Tyrosol is present in plants such as olive andRhodiola rosea[5].However,the tyrosol content of virgin olive oil is only 40-180 mg·(kg oil)-1[2].Despite its natural abundance,the cost of extraction for high-purity tyrosol from natural sources is very high [6].For chemical synthesis,tyrosol can be produced from precursors such as phenol,phenylethanol [7],4-hydroxyphenylacetic acid [8],andp-hydroxyphenylamine [9] via one or more chemical reactions.However,these methods have some drawbacks such as expensive substrates,complicated purification,or low yields.In recent years,the biosynthesis of tyrosol has received increasing attention because of its low cost,high production efficiency,and environmental friendliness [5].

Tyrosol can be produced by microbial fermentation and enzymatic conversion.For microbial fermentation,S.cerevisiaecan directly biosynthesize tyrosol via the Ehrlich pathway[10].Recent studies reported increased tyrosol as high as (927.68 ± 25.26)mg·L-1in shaking flasks and 9.90 g·L-1in a 5-L bioreactor via the metabolic engineering of the tyrosol biosynthesis pathway [11-13].ForE.colistrain,three natural pathways have been introduced as follows:(a)The tyrosine decarboxylase-tyramine oxidase(TDCTYO) pathway [14];(b) Ehrlich pathway [15];(c) the aromatic aldehyde synthase (AAS) pathway [16] (Fig.1).For the TDC-TYO pathway,TDC fromPapaver somniferumand TYO fromMicrococcus luteuswere co-expressed in strain JW1380 to obtain the engineered strain and achieve tyrosol production at 69.08 mg·L-1with 1% (w/v) glucose in 48 h [14].For the Ehrlich pathway,the engineering ofE.colifor tyrosol production has been explored by introducing the phenylpyruvate decarboxylase gene (ARO10) fromS.cerevisiaeintoE.coli,with the consequent production of 573.36-1508.7 mg·L-1and 3.9 g·L-1tyrosol from glucose in shaking flasks and a 5-L fermenter,respectively [17-19].For the AAS pathway,AAS fromPetroselinum crispumwas introduced intoE.coliΔfeaBΔtyrRΔpheA,and 531 mg·L-1tyrosol was produced in 36 h[1].On the other hand,enzymatic conversion provides an efficient way to produce tyrosol from L-Tyr.For example,when ARO10 and the aromatic amino acid aminotransferase gene (ARO8) fromS.cerevisiaewere co-overexpressed inE.coli,the recombinant strain could convert 1812 mg·L-1tyrosine to 1203 mg·L-1tyrosol [17].Moreover,when L-amino acid deaminase (LAAD),ARO10,and phenylacetaldehyde reductase (PAR) were co-expressed inE.coli,0.65 g·L-1tyrosol was synthesized from 0.9 g·L-1L-Tyr[20].Unfortunately,the low titer and productivity of tyrosol have significantly limited its industrial application.Therefore,it is still challenging to develop an efficient tyrosol synthesis method.

Fig.1.Biosynthetic pathway for the production of tyrosol.(a)In the TDC-TYO pathway,tyrosol can be synthesized from tyrosine through decarboxylation,deamination and reduction by tyrosine decarboxylase,tyramine oxidase and alcohol dehydrogenase.(b) In the Ehrlich pathway,tyrosol can be synthesized from tyrosine through transamination,decarboxylation and reduction by aromatic amino acid transferase,pyruvate decarboxylase and alcohol dehydrogenase.(c)In the AAS pathway,tyrosine can be directly converted to tyrosol by aromatic aldehyde synthetase and alcohol dehydrogenase.

In this study,a cascade pathway was designed by mimicking the Ehrlich pathway that could transform L-Tyr to tyrosol.After identifying the rate-limiting enzymatic step in the multi-enzyme reaction,the cascade pathway was optimized via a modular assembly and duplication ofCtPDC inE.coli.The optimizedE.colistrain was used to convert L-Tyr to tyrosol with high efficiency viain situproduct removal (ISPR) strategy.

2.Materials and Methods

2.1.Strains,plasmids,and media

The expression plasmid pET-28a(+) and the host strainEscherichia coliBL21 (DE3) were purchased from Novegen (Madison,WI,USA).The pACYCDuet-1 (pACYC),pCDFDuet-1 (pCDF),pETDuet-1(pET)and pRSFDuet-1(pRSF)expression plasmids were purchased from Novagen (Darmstadt,Germany).All chemicals and reagents were pure and obtained commercially.Molecular biology reagents such as the restriction enzymes,PrimeSTAR,DNA polymerase and polymerase chain reactions (PCR) reagents were supplied by TaKaRa (Dalian,China).Cultivation for gene cloning,plasmid construction and inoculum preparation were performed in Luria-Bertani(LB)broth or agar plate.The recombinant strains were cultured in Terrific Broth (TB) medium for enzyme expression.Chloramphenicol (34 μg·ml-1),streptomycin (34 μg·ml-1),kanamycin(50 μg·ml-1)and ampicillin(100 μg·ml-1)were added to the medium as required.

2.2.Construction of the co-expressed strains

The biosynthesis pathway of tyrosol constructed in this study is shown in Fig.2.Genes of LAAD,PDC,ADH6 and GDH were amplified fromProteus mirabilis(GenBank Accession No.U35383),C.tropicalis(GenBank Accession No.XM_002549483),S.cerevisiae(GenBank Accession No.NM_001182831) andBacillus megaterium(GenBank Accession No.LK055286),respectively.The genes were amplified by PCR,and plasmids (pET-28a(+),pACYC,pCDF,pET,and pRSF) used in this study were extracted by the plasmid miniprep kit (Sangon Biotech).Main primers used for constructing coexpressed strains are listed in Table A1.The genes were ligated to the plasmids using ClonExpress Entry One Step Cloning Kit(Vazyme Biotech Co.,Ltd.,Nanjing,China).Further transformation of constructed plasmid(s) intoE.coliBL21(DE3) to express the genes.The plasmids and strains constructed in this study are shown in Table 1.

Table 1 Strains and plasmids used in this study

Fig.2. In vitro design and reconstruction of tyrosol biosynthesis pathway from L-tyrosine.LAAD:L-amino acid deaminase,PDC:pyruvate decarboxylase,ADH:alcohol dehydrogenase and GDH:glucose dehydrogenase.

2.3.Cultivation of recombinant E.coli strains

RecombinantE.colistrains were inoculated into 30 ml of liquid LB medium with appropriate antibiotics and cultured at 200 rpm and 37°C for 10 h.The inoculated strain was then transferred into a 500 ml baffled flask with 100 ml TB medium (24 g·L-1yeast extract,12 g·L-1tryptone,4 g·L-1glycerol,2.31 g·L-1KH2PO4,and 16.43 g·L-1K2HPO4) containing appropriate antibiotics.When the optical density at 600 nm (OD600) of the culture reached 0.6-0.8,isopropyl-β-D-thiogalactopyranoside (IPTG) was added to a final concentration of 0.4 mmol·L-1to induce the protein expression at 25°C and 200 r·min-1.After incubation for 12 h,cells were harvested by centrifugation (6000g,10 min,4 °C).

2.4.Analytical methods

The titer of tyrosol was determined by high-performance liquid chromatography(HPLC,Dionex UltiMate 3000).The analytical conditions of HPLC were as follows:Dionex UltiMate 3000 VWD;Agilent Zorbax SB-C18 column,5 μm,4.6 × 250 mm;column temperature,25°C;detection wavelength,222 nm;injection quantity,10 μl;and flow rate,0.5 ml·min-1.The mobile phase A was 100%methanol containing 0.1%trifluoroacetic acid,and the mobile phase B was water containing 0.1%trifluoroacetic acid.The methanol concentration was increased from 16% to 45% for 20 min,and then decreased from 45%to 16%for 5 min.Under these conditions,the retention times of L-Tyr,4-hydroxyphenylpyruvate (4-HPP)and tyrosol were 11.9,15.8 and 18.8 min,respectively.

2.5.Enzyme expression and purification

Enzymes were individually overexpressed and purified fromE.coliBL21(DE3)with the pET28a(+)plasmid.Cells were harvested by centrifugation at 6000gfor 10 min and resuspended in Binding buffer (pH 8.0).The cell suspensions were sonicated by high pressure cell crusher(Union-Biotech,Shanghai,China)at 4°C and centrifuged at 12,000gfor 20 min.The recombinant enzymes were purified by an AKTA pure system (GE Healthcare Life Science,USA) with a nickel-affinity column.The purification process was carried out as previously described[21].Purified enzyme solutions were concentrated using proper ultrafiltration devices according to their molecular weight,and testedviasodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE).All operations were performed at 4 °C when necessary.

2.6.Enzyme activity assays

The activity of L-amino acid deaminase fromP.mirabilis(PmLAAD) was assayed by detecting the increase of tyrosol when excessCtPDC,alcohol dehydrogenase 6 fromS.cerevisiae(ScADH6),glucose dehydrogenase fromB.megaterium(BmGDH) and L-Tyr were added.The enzyme activity ofCtPDC was determined by adding excess 4-HPP,NADPH,andScADH6 and detecting the increase in the tyrosol concentration.In the determination of the activity ofScADH6,excess 4-HPP,NADPH,andCtPDC were added to detect the increase of the tyrosol concentration.Excess glucose and NADP+were added when assaying the activity of GDH,and then the increase of NADPH was measured at 340 nm.One unit of enzyme activity was defined as the amount of enzyme required for the increase of 1 μmol of tyrosol (NADPH) in 1 min [22].The activity of the recombinant strains expressing single enzyme was measured using purified enzyme,while the activity of the coexpressed strains was measured using whole-cell catalysts.Values and error bars represent the mean values and standard deviations of biological triplicates.

2.7.Production of tyrosol from tyrosine

Cells were collected by centrifugation at 6000gfor 10 min after induction.The conversion experiments were carried out in a 100-ml shake flask with 10 ml working volume and a 3-L bioreactor with 1 L working volume.The reaction system was 10 ml of phosphate buffer(200 mmol·L-1,pH 8.0)containing 20 g·L-1whole-cell catalysts (wet cells),25 g·L-1L-Tyr,2 mmol·L-1thiamine pyrophosphate chloride (TPP),5 mmol·L-1MgSO4·7H2O,0.5 mmol·L-1NADP+and 25 g·L-1glucose.The reactions were carried out in 100-ml erlenmeyer flasks at 30 °C and 200 r·min-1for 36 h and terminated by 2 mol·L-1NaOH.To optimize the reaction conditions,different temperatures and pHs were explored to optimize the temperature and pH of the whole-cell catalyst.Then,the biotransformations of L-Tyr to tyrosol were conducted withE.coli07-2 in the presence of XAD4 resins under optimal conditions.The samples were analyzed by HPLC and gas chromatography (GC) to quantify the concentration of tyrosol.

2.8.Selection of adsorbent resins

Five adsorbents including D101,HP20,XAD1180,XAD4 and DA201 were selected to adsorb tyrosol.Before use,the resins were soaked in ethanol to remove impurities,and then washed with distilled water to remove the solvent.The adsorption ability for tyrosol and L-Tyr of these adsorbents was measured as follows:10 g·L-1each of tyrosol and L-Tyr were dissolved in 20 ml phosphate buffer(200 mmol·L-1,pH 7.5)in the 250-ml flasks.The resin(5%)was added into the flask and shaken at 200 r·min-1for 2 h at 25°C.All experiments were performed in triplicate.Concentrations of tyrosol and L-Tyr in the aqueous phase were analyzed by HPLC.

2.9.Adsorption of tyrosol by XAD4

To establish the optimum amount of XAD4 for the biotransformation,20 g·L-1each of tyrosol and L-Tyr were dissolved in phosphate buffer (200 mmol·L-1,pH 7.5).XAD4 was added to 2%,4%,6%,8%,10% and 12%,respectively.Adsorption was carried out at 200 r·min-1for 2 h at 25 °C.All experiments were performed in triplicate.Concentrations of tyrosol and L-Tyr in the aqueous phase were analyzed by HPLC.

2.10.Desorption of tyrosol

The resin on which tyrosol had been adsorbed was collected and eluted with ethanol(1:5,w/v).Desorption was performed with shaking at 200 r·min-1for 2 h at 25 °C,and tyrosol was fully released after desorping three times.The ethanol eluates were combined and titers of tyrosol were analyzed by HPLC or GC.After desorption,the resins were soaked in ethanol and washed with distilled water sufficiently to remove the solvent,and stored for the next biotransformation.

Fig.3.Optimization of activity ratio of LAAD:PDC:ADH:GDH in vitro.(a)SDS-PAGE analysis of the purified protein.(b)Analysis of the in vitro reconstructed system with GCMS.(c)Effect of different activity ratios of LAAD:PDC on the 4-HPAA production.(d)Effect of different activity ratios of ADH:GDH on tyrosol production.(e)Effect of different activity ratios of Module 1 to Module 2 on tyrosol production.

3.Results

3.1.Cascade design and in vitro reconstruction of the tyrosol biosynthesis pathway

As illustrated in Fig.1b,through the Ehrlich pathway,L-tyrosine was converted to 4-HPP by aromatic amino acid transaminase(AAT,EC 2.6.1.57),and then,4-HPP was subsequently catalyzed to 4-hydroxyphenylacetaldehyde (4-HPAA) by pyruvate decarboxylase(PDC,EC 4.1.1.1)and reduced to tyrosol by alcohol dehydrogenase(ADH,EC 1.1.1.2).Among these reactions,the first transaminate reaction of L-Tyr to 4-HPP could be replaced by deamination catalyzed by LAAD (EC 1.4.3.2) [23].The third reaction depends on NAD(P)H;therefore,glucose dehydrogenase (GDH,EC 1.1.1.47) [24] was introduced as a cofactor regeneration system to provide enough NADPH (Fig.2).

To reconstruct the enzyme cascadein vitro,PmLAAD,CtPDC,ScADH6 andBmGDH were selected according to their specific enzyme activities (Table A2 and A3).Next,the four selected genes were amplified,overexpressed,and purified (Fig.3(a)).To confirm the feasibility of the designed cascade,these four enzymes at a molar ratio of 1:1:1:1 was combined with 5 mmol·L-1L-Tyr.After 4 h,the formation of the final product,2.76 mmol·L-1tyrosol,was confirmed using mass spectroscopy (MS) and nuclear magnetic resonance (Fig.3(b),A1),which demonstrated that the designed cascade usingPmLAAD,CtPDC,ScADH6 andBmGDH was feasible to convert L-Tyr to tyrosol.

Furthermore,the kinetic parameters ofPmLAAD,CtPDC,ScADH6,andBmGDH were determined using purified enzymes,and the results are listed in Table 2.Among them,BmGDH showed the highestkcat/Km(7.45 min-1·mmol·L),with the lowestkcat/Km(0.37 min-1·mmol·L)detected forCtPDC.This result demonstrated thatCtPDC was the rate-limiting step for efficient tyrosol production.Based on thekcat/Kmvalue of these four enzymes,the enzyme level was optimized using a modular approach wherePmLAAD andCtPDC were combined as module 1 to convert L-Tyr to 4-HPAA.Module 2,which consisted ofScADH6 andBmGDH,catalyzed 4-HPAA conversion to tyrosol.To identify the potential bottleneck in the two modules,the contribution of a single enzyme was examined by fixing the concentration of one enzyme while changing that of the other in the module.In module 1,the titer and yield of 4-HPAA gradually increased to 3.52 g·L-1and 93.7%,respectively,with the increasing ratio ofPmLAAD toCtPDC from 1:0.5 to 1:3.0(Fig.3(c))with 5 g·L-1L-tyrosine and 0.5 U·ml-1PmLAAD.However,if the ratio was under 1:1.5,the intermediate,4-HPP,was detected in the transformation broth,demonstrating thatCtPDCactivity in this module should be at least 1.5 times higher than that ofPmLAAD.Using the same method,the highest titer (3.44 g·L-1)and yield (90.1%) of tyrosol were obtained when the ratio ofScADH6 toBmGDH was 1.0:1-5.0:1,with 3.76 g·L-1of 4-HPAA and 0.5 of U·ml-1ScADH6 (Fig.3(d)).To balance the pathway between the two modules,we combined different ratios of modules 1 and 2 (from 1:0.5 to 1:5) with 5 g·L-1of L-Tyr (Fig.3(e)),and found that the optimal ratio ofPmLAAD:CtPDC:ScADH6:BmGDH was 1:1.5:1:1.Under the optimal conditions,the concentration of enzymes was the lowest,without any 4-HPP accumulation and the titer of tyrosol was the highest (3.55 g·L-1).

Table 2 Kinetic constants of PmLAAD, CtPDC, ScADH6 and BmGDH

Fig.4.Construction and optimization of the co-expressed strain.(a) Engineering of recombinant E.coli strains expressing PmLAAD, CtPDC, ScADH6 and BmGDH using two modules (PmLAAD-CtPDC and ScADH6-BmGDH) with different plasmids.(b) Construction of the plasmids used for strain 07.(c) SDS-PAGE analysis of strain E.coli 07 from cell-free extracts.(d)Production of tyrosol upon biotransformation of L-Tyr(25 g·L-1)using 12 constructed recombinant E.coli strains(20 g·L-1)constructed in a mixture of phosphate buffer(200 mmol·L-1,pH 8.0,containing 2 mmol·L-1 thiamine pyrophosphate chloride(TPP),5 mmol·L-1 MgSO4·7H2O,0.5 mmol·L-1 NADP+,and 25 g·L-1 glucose)for 36 h.(e) The enzyme activity in recombinant strains with duplicated CtPDC.

Fig.5.Effects of temperature,pH,substrate concentration and whole-cell biocatalysts concentration,on the production of strain E.coli 07-2.(a) Effects of different temperatures on tyrosol titer.(b)Effects of different pH values on tyrosol titer.(c)Effects of different L-Tyr concentrations on tyrosol titer.(d)Effects of different whole-cell biocatalyst concentrations on tyrosol titer.All experiments were carried out in 10 ml of 200 mmol·L-1 phosphate buffer containing(2 mmol·L-1 TPP,5 mmol·L-1 MgSO4·7H2O and 0.5 mmol·L-1 NADP+) at 200 r·min-1 for 36 h.

3.2.In vivo construction in E.coli and optimization of the tyrosol biosynthesis pathway

Plasmids with different copy numbers have been used to balance the expression of enzymes in a multi-enzymatic cascade[25,26].To improve the efficiency of cascade reactionin vivo,PmLAAD,CtPDC,ScADH6,andBmGDH were overexpressed at different expression levels.RecombinantE.colistrains were constructed using a modular approach.Four plasmids with different copy numbers were used to optimize the co-expression:Pm-LAAD-CtPDC module andScADH6-BmGDH module were constructed in pACYC,pCDF,pET,and pRSF,respectively(Table 1).The copy number of pACYC,pCDF,pET,and pRSF plasmids are 10,20,40,and 100,which could be classified as low,medium and high copy numbers[3].Then,fourPmLAAD-CtPDC plasmids and fourSc-ADH6-BmGDH plasmids were combined and transformed intoE.coliBL21 to obtain 12E.colistrains (Fig.4(a,b)),each coexpressingPmLAAD,CtPDC,ScADH6,andBmGDH for one-pot synthesis of tyrosol from L-Tyr(Fig.4(c)).As shown in Fig.4(d),all 12 strains successfully produced tyrosol in 36 h.Tyrosol concentration varied with the expression levels of bothPmLAAD-CtPDC module(L-Tyr to 4-HPAA) andScADH6-BmGDH module (4-HPAA to tyrosol) due to the difference on the plasmid copy number.Among them,strainE.coli07 with the medium copy number ofPm-LAAD-CtPDC module and low copy number ofScADH6-BmGDH module achieved the highest tyrosol titer 13.97 g·L-1with 25 g·L-1L-tyrosine as the substrate.However,the conversion rate was only up to 73.3%,and further,the accumulation of intermediate 4-HPP was increased to 6.6 g·L-1(Fig.A2).This was because the specific activities ofPmLAAD,CtPDC,ScADH6 andBmGDH inE.coli07 were(0.83±0.05)U·ml-1,(0.66±0.03)U·ml-1,(0.91±0.05)U·ml-1and(2.32 ± 0.08) U·ml-1,respectively,and the corresponding ratio of enzyme activity was 1:0.8:1.1:2.8,which was not consistent with the result of thein vitroreconstruction.Therefore,it was necessary to control the activity ratio ofPmLAAD toCtPDC between 1:1.5 and 1:3.0,to decrease the accumulation of 4-HPP.A duplicateCtPDC was first cloned into the plasmid pET-PmLAAD-CtPDC,giving strainE.coli07-2 (Figs.A3,A4).The specific activities ofCtPDC andPmLAAD were (1.05 ± 0.06) U·ml-1and (0.69 ± 0.05) U·ml-1(Fig.4(e)) inE.coli07-2,which led to a ratio of 1:1.5:1.2:3 forPmLAAD,CtPDC,ScADH6,BmGDH,and the titer of was increased to 17.42 g·L-1(25 g·L-1L-tyrosine) with 91.4% conversion rate,without any 4-HPP detected in the transformation broth.

Table A1 Primers used for gene cloning and plasmid construction.

Table A2 The enzyme activity of PDC from different origins.

Table A3 The enzyme activity of ADH from S.cerevisiae and E.coli

3.3.Optimizing the transformation conditions for tyrosol production

The effect of conditions on tyrosol production was observed in 100-ml shaking flasks and shown in Fig.5.It was found the following:(1) Tyrosol titer increased to 20.63 g·L-1with 90.2% conversion rate and 30 g·L-1L-Tyr at 25°C(Fig.5(a));(2)the highest tyrosol titer (21.84 g·L-1) and conversion rate (95.5%) were detected at pH 7.5 and tyrosol titer decreased when pH ＜7.5 or pH ＞8.0(Fig.5(b));(3) when L-Tyr concentration increased from 30 to 50 g·L-1,the titer of tyrosol was still 21.8 g·L-1with the conversion rate decreasing to 57.2%,and about 1.33 to 21.25 g·L-1of the residual L-Tyr was detected in the conversion broth (Fig.5(c));(4) tyrosol titer did not increase notably by further increasing the concentration of the whole-cell biocatalysts from 20 to 30 g·L-1,and the optimal concentration for tyrosol production was determined to be 20 g·L-1(Fig.5(d)).

3.4.Minimizing product inhibition using in situ product adsorption

Increasing L-Tyr and whole-cell biocatalyst concentration did not increase the tyrosol titer,which might be because of the inhibited cell activity by tyrosol.Therefore,5-25 g·L-1of exogenous tyrosol was added at the beginning of the conversion to investigate the effect of product inhibition on tyrosol titer(Fig.6).It was found that the titer of tyrosol decreased with an increase in the amount of exogenous tyrosol,and no tyrosol was produced when the amount of exogenous tyrosol reached 22 g·L-1.These results indicated that tyrosol was toxic toE.coli07-2,and higher concentrations of exogenous tyrosol significantly decreased the reaction rate.Therefore,removal of tyrosol is an efficient way to further increase the titer of tyrosol.

Effect of different adsorbents(5%,resins D101,HP20,XAD1180,XAD4,DA201) on the adsorption of tyrosol was tested using a phosphate buffer containing 10 g·L-1tyrosol and 10 g·L-1L-Tyr.Among them,XAD4 resins could adsorbed 85% of tyrosol after 2 h (Fig.7(a)).The effect of the amount of XAD4 resins (2%-12%)on tyrosol adsorption was determined,and the result is illustrated in Fig.7(b).It was found that 10% and 12% XAD4 resins showed almost the same adsorption with approximately 17.5 g·L-1tyrosol being adsorbed.Then,with the addition of 20%XAD4 resins to the system with 40 g·L-1L-Tyr for 36 h,29.1 g·L-1tyrosol with 95.4%conversion rate was obtained (Fig.7(c),A5),which was 7.28 g·L-1and 23.8% higher than that of the control (without the addition of XAD4 resins).

The biotransformation process was scaled up to 3-L scale,when 20 g·L-1of whole-cell biocatalysts were added to 1-L of phosphate buffer(200 mmol·L-1,pH 7.5,containing 50 g·L-1L-Tyr,2 mmol·L-1TPP,5 mmol·L-1MgSO4·7H2O,0.5 mmol·L-1NADP+,50 g·L-1glucose)with 25%XAD4 resins at 25°C for 36 h,and the titer and conversion rate of tyrosol was 35.7 g·L-1and 93.6%,respectively(Fig.7(d),13.87 g·L-1and 36.4% higher than the control (without resins addition).

Fig.6.Identification of the product inhibition of strain E.coli 07-2.The conversion of tyrosol upon biotransformation of L-Tyr (30 g·L-1) using E.coli 07-2 wet cells(20 g·L-1) with the addition of different initial tyrosol concentrations (0-25 g·L-1)for 36 h.

Fig.7.Minimzing product inhibition production via adsorption using resin.(a) The concentration of tyrosol and L-Tyr in the aqueous phase after adsorption of tyrosol(10 g·L-1)and L-Tyr(10 g·L-1)with different adsorbents(5%)in phosphate buffer(200 mmol·L-1,pH 7.5)at 200 r·min-1 and 25°C for 2 h.(b)The concentration of tyrosol and L-Tyr in the aqueous phase after adsorption of tyrosol(20 g·L-1)and L-Tyr(20 g·L-1)with different addtion of XAD4 resins(2%-12%)in phosphate buffer(200 mmol·L-1,pH 7.5)at 200 r·min-1 and 25°C for 2 h.(c)The concentration of tyrosol upon biotransformation of L-Tyr(40 g·L-1)using E.coli 07-2 whole-cell biocatalysts(20 g·L-1)and(20%)XAD4 resins in a mixture of phosphate buffer(200 mmol·L-1,pH 7.5,2 mmol·L-1 TPP,5 mmol·L-1 MgSO4·7H2O,0.5 mmol·L-1 NADP+,40 g·L-1 glucose)for 36 h at 200 r·min-1 and 25 °C.(d) The concentration of tyrosol upon biotransformation of 1-L of L-Tyr (50 g·L-1) using E.coli 07-2 whole-cell biocatalysts (20 g·L-1)and (25%) XAD4 resins in a mixture of phosphate buffer (200 mmol·L-1,pH 7.5,2 mmol·L-1 TPP,5 mmol·L-1 MgSO4·7H2O,0.5 mmol·L-1 NADP+,50 g·L-1 glucose) for 36 h at 200 r·min-1 and 25 °C.

4.Discussion

This study aimed to develop an efficient tyrosol production process from L-Tyr.To achieve this,a four-enzyme cascade includingPmLAAD,CtPDC,ScADH6,andBmGDH,was designed and constructedin E.coli07-2.The cascade pathway was optimized using modular assembly and gene duplication,and the titer and conversion rate of tyrosol were 21.84 g·L-1and 95.5%,respectively.The titer and conversion rate further increased to 35.7 g·L-1and 93.6%through ISPR with the addition 25%XAD4 resins.Results presented here provide a potential enzymatic process for the industrial-scale production of tyrosol from cheap amino acids.

In this study,a synthetic pathway for tyrosolviathe coexpression ofPmLAAD,CtPDC,ScADH6,andBmGDH inE.coli07-2 was established and a high titer and conversion rate of tyrosol were achieved.To the best of our knowledge,this is the highest yield of biotechnological tyrosol production achieved till date.We usedE.colito mimic the Ehrlich pathway fromS.cerevisiaeto produce tyrosol and construct a cofactor regeneration system to provide an economical and efficient pathway.During the past decades,several biotechnological methods have been developed for tyrosol production.S.cerevisiaewas engineered using a pushpull-restrain strategy by disrupting pyruvate decarboxylase(PDC1),prephenate dehydratase(PHA2),and anthranilate synthase(TRP3),along with the introduction ofPcAAS,chorismate mutase/prephenate dehydrogenase mutant (EcTyrAM53I/A354V),and phosphoketolase (Xfpk),which revealed a tyrosol titer of (927.68 ± 25.26)mg·L-1in a shaking flask and 8.37 g·L-1in a 5-L bioreactor for 192 h [12].InE.coli,five differentS.cerevisiae ARO10* (E.colicodon-optimized) were separately integrated intoE.coliΔfeaBΔpheAΔtyrBΔtyrR to get the mutant strain YMG5A*R [19],with sugar as the carbon source,which increased tyrosol production to 1.50 g·L-1in shaking flasks and 3.9 g·L-1in a 5-L fermenter.However,further increase of tyrosol titer was difficult,because of the long tyrosol synthesis pathway inE.coli,which makes it difficult to achieve a gene expression balance.Researchers have also investigated enzymatic conversion with L-tyrosine as the substrate,where tyrosol was successfully synthesizedviatwo established enzymatic cascades of ARO8-ARO10-ADH [17] and LAADARO10-PAR [20],and the titer of tyrosol reached 1.20 g·L-1(87.1%) and 0.65 g·L-1(94%),respectively.

In this study,a modular approach was used to assemblePmLAAD,CtPDC,ScADH6,andBmGDH inE.coli07-2,resulting in the enzyme activity ratio of 1:0.8:1.1:2.8.Then,the ratio was changed to 1:1.5:1.2:3 through doubling theCtPDC expression.As a result,the accumulation of the intermediate,4-HPP,was eliminated.In the multi-enzyme cascade reaction,it has been proved that balancing pathway enzyme activities is critical for pathway efficiency [28].If the activity of each enzyme is not wellcoordinated,the rate-limiting steps will appear,which usually leads to the accumulation of intermediates and reduces the efficiency of the pathway.A modular approach is widely used to balance the expression of multiple enzymes by assembling plasmids with different copy numbers [26].In one study,a modular approach was used to design a cascade to co-express five enzymes using four plasmids with different copy numbers,and thisE.colistrain (RE) achieved the highest production of 2-phenylethanol[27].On the other hand,the gene duplication strategy could be used to resolve the accumulation of the intermediate in cascade reaction.For example,when increasing the copy number of the rate-limiting enzyme panD to five in the β-alanine synthesis cascade,the titer of the intermediate L-aspartic acid was decreased to less than 2 g·L-1and β-alanine production was enhanced by 27.7% [29].

In this study,the high concentration of tyrosol in the transformation broth significantly inhibited the biotransformation of theE.coli07-2 strain (Fig.6).An ISPR technology,with the addition of 25%XAD4 resins,was used to continuously adsorb tyrosol from the transformation broth,so that a titer and conversion rate of 35.7 g·L-1and 93.6%,respectively,were achieved in a 3-L fermenter,and these values were higher than those of the control(21.83 g·L-1and 57.2%,without resin addition).Tyrosol,a phenylethanol compound,which has a variety of biological activities,could damage cell membranes to increase membrane fluidity,leading to a decrease in cell activity [30,31].As a result,this could decrease the biotransformation efficiency.There are several strategies developed to increase the tolerance of products,such as alcohol,2-phenylethanol,and butanol [32].On the one hand,mutagenesis and genetic engineering have been used to modify strains to tolerate these products.Overexpression of four endogenous genes ofS.cerevisiae(INO1,DOG1,HAL1andMSN2) can increase ethanol tolerance;INO1-overexpressing strain elicited the highest ethanol tolerance and the volumetric productivity of ethanol was 1.30 g·L-1·h-1in the fermentation experiments performed with 300 g·L-1glucose,as previously reported,which was 35% higher than that of the control strain [33].On the other hand,a more direct method is the ISPR technology,which is widely used in fermentation and biotransformation systems to decrease the toxicity of products.ISPR technology with the addition of 7%HZ818 resin was used to adsorb 2-phenylethanol in a biotransformation system so that the titer and conversion rate of tyrosol were 6.6 g·L-1and 74.4%,respectively,and the product concentration was improved by 66.2% compared to the biotransformation without the addition of adsorbent resin [34].Moreover,when 50 g·L-1Dowex Optipore SD-2 resin was added ton-butanol fermentation byClostridium acetobutylicumATCC 824,2.22%n-butanol production was achieved,which was above the inhibitory threshold ofC.acetobutylicumATCC 824,nearly twice as much of traditional,single-phase fermentation [35].

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This work was financially supported by the Fundamental Research Funds for the Central Universities(JUSRP21915),National Natural Science Foundation of China (22008089,21878126),Provincial Natural Science Foundation of Jiangsu Province(BK20200622),the key technologies Research&Development Program of Jiangsu Province(BE2018623),and the National First-Class Discipline Program of Light Industry Technology and Engineering(LITE2018-20).

Appendix A

Fig.A1.NMR spectra of tyrosol.(a) 1H NMR.(b) 13C NMR.

Fig.A2.HPLC chromatography of an Agilent Zorbax SB-C18 column with a UV detector.The retention time of L-Tyr,4-HPP and tyrosol were 11.9,15.8 and 18.8 min,respectively.(a) Analysis of mixture by standard samples.(b) Analysis of reaction miture containing 25 g·L-1 L-Tyr and 20 g·L-1 E.coli 07(wet cells)after reaction for 36 h.

Fig.A3.Construction of CtPDC gene duplication strain.

Fig.A4.SDS-PAGE of E.coli strain 07 and E.coli strain 07-2.Lanes:M,protein marker;1,supernatant of E.coli strain 07 cell-free extract;2,supernatant of E.coli strain 07-2 cell-free extract.

Fig.A5.HPLC chromatography of an Agilent Zorbax SB-C18 column with a UV detector.(a) Analysis of sample of aqueous phase after reaction.(b) Analysis of sample of ethanol eluates after desorption the XAD4 resins.