Plant Cell Rep
DOI 10.1007/s00299-015-1790-0

Metabolic engineering of 2-phenylethanol pathway producing
fragrance chemical and reducing lignin in Arabidopsis
Guang Qi1
• Dian Wang1,3
• Li Yu1
• Xianfeng Tang1,3
• Guohua Chai1
•
Guo He1
• Wenxuan Ma1
• Shengying Li1
• Yingzhen Kong2
• Chunxiang Fu1
•
Gongke Zhou1
Received: 11 February 2015 / Revised: 25 March 2015 / Accepted: 31 March 2015
 Springer-Verlag Berlin Heidelberg 2015

Abstract
Key message Two 2-phenylethanol biosynthetic pathways were constructed into Arabidopsis; 2-phenylethanol biosynthesis led to reduced rate of lignin biosynthesis and increased cellulose-to-glucose conversion in the transgenic plants.
Lignin is the second most abundant biopolymer on the planet with importance for various agro-industrial activities. The presence of lignin in cell walls, however, impedes biofuel production from lignocellulosic biomass.
The phenylpropanoid pathway is responsible for the biosynthesis of lignin and other phenolic metabolites such as 2-phenylethanol. As one of the most used fragrance
chemicals, 2-phenylethanol is synthesized in plants from Lphenylalanine which is the first specific intermediate towards lignin biosynthesis. Thus, it is interesting to prove
the concept that the phenylpropanoid pathway can be modulated for reduction of lignin as well as production of natural value-added compounds. Here we conferred two
2-phenylethanol biosynthetic pathways constructed from plants and Saccharomyces cerevisiae into Arabidopsis. As anticipated, 2-phenylethanol was accumulated in transgenic plants. Moreover, the transformants showed 12–14 % reduction in lignin content and 9–13 % increase in cellulose content. Consequently, the glucose yield from cell wall hydrolysis was increased from 37.4 % in wild type to 49.9–52.1 % in transgenic plants with hot water pretreatment. The transgenic plants had normal development and even enhanced growth relative to the wild type.
Our results indicate that the shunt of L-phenylalanine flux to the artificially constructed 2-phenylethanol biosynthetic pathway most likely reduced the rate of lignin biosynthesis in Arabidopsis.

Keywords
Biofuel  Fragrance chemical  Lignin
biosynthesis  Phenylpropanoid pathway  2-Phenylethanol

Introduction
Lignocellulosic biomass represents the most abundantly available renewable materials on Earth for the pulping and paper-making, ruminant animal feed and biofuel production (Li et al. 2008). It consists of cell wall that makes up more than 80 % of plant dry matter biomass (Ragauskas et al. 2014). Plant cell wall, mainly containing cellulose, hemicellulose and lignin, is highly recalcitrant to chemical or biological degradation due to its rigid and compact structure (Chen and Dixon 2007; Eudes et al. 2014). Lignin is a complex branched polymer of phenolic alcohols that plays an important role in cell wall structure reinforcement, mechanical support, water transport, and plant defense against biotic and abiotic stress (Campbell and Sederoff 1996; Douglas 1996; Boerjan et al. 2003). Lignin content and composition, however, have been recognized for its negative impact on a wide range of industrial applications, such as pulp, livestock forage, and bioethanol production (Chen and Dixon 2007; Li et al. 2010). Currently, monolignol biosynthesis has been relatively well understood in plants. The monolignols are formed from L-phenylalanine via the phenylpropanoid pathway. Enzymes are known that can catalyze these reactions, involving phenylalanine ammonia lyase (PAL), hydroxylases, O-methyltransferases and reductase(Vanholme et al. 2010; Van Acker et al. 2013). Genetic manipulation of lignin is a promising strategy to reduce cell wall recalcitrance and therefore increase saccharification of lignocellulosic biomass (Bonawitz and Chapple 2013; Bonawitz et al. 2014). Previous studies have shown that downregulation of lignin biosynthetic genes in spruce, poplar, tobacco, alfalfa, switchgrass, and ryegrass can successfully reduce lignin biosynthesis and magnificently increase pulping efficiency, forage digestibility and bioethanol production (Sewalt et al. 1997a, b; Guo et al. 2001a, b; Reddy et al. 2005; Jouanin et al. 2000; Sarath et al. 2008; Fu et al. 2011a; Samuel et al. 2014; Louie et al. 2010; Tu et al. 2010). Other studies have indicated that overexpression of MYB transcription factor Atmyb4 or its homologs in Arabidopsis, tobacco and switchgrass leads to strong suppression of lignin biosynthesis (Shen et al. 2012). The above traditional strategies for lignin manipulation focus on the identification and regulation of lignin genes or their transcription factors. The available targets employed for lignin modification, however, are limited due to the number of genes known in monolignol biosynthetic pathways. Particularly, the genes that can cause a substantial decrease in lignin biosynthesis without major visible defects in plant growth are not sufficient for the purpose of commercial production of low lignin biomaterials (Bonawitz et al. 2014). Thus, the major challenge in current lignin bioengineering is identification of numerous novel targets or reconstruction of new pathways to partially alter the substantial carbon flux into lignin pathway (Li et al. 2008; Bonawitz and Chapple 2013). Previous studies have shown that disruption of S-adenosyl-L-methionine synthetases (SAMS) or methylenetetrahydrofolate reductase (MTHFR)
in plants can affect the biosynthesis of SAM, the methyl donor consumed by two O-methyltransferases in lignin biosynthetic pathway, and therefore significantly reduce
lignin content (Shen et al. 2002; Tang et al. 2014). Another promising structure-based protein engineering approach indicates that expression of an engineered monolignol 4-Omethyltransferase created by iterative saturation mutagenesis in Arabidopsis can result in etherealization of the para-hydroxyls of lignin monomeric precursors, and therefore lead to depression of lignin biosynthesis and improvement of cell wall saccharification (Zhang et al. 2012). Other studies have suggested that lignin polymerization can be reduced through the overproduction of sidechain- truncated lignin monomers achieved by expressing a bacterial hydroxycinnamoyl-CoA hydratase-lyase (HCHL) in lignifying tissues of Arabidopsis inflorescence stems and improve saccharification (Eudes et al. 2012). 2-PE is one of the most used flavor principles with a pleasant rose-like odor. Several plants such as rose, carnation, hyacinth, and jasmine are capable of producing natural 2-PE. However, those plant tissues usually contain trace amounts of 2-PE, except rose flower (Rusanov et al. 2005). Thus, the majority source of 2-PE currently in use is synthesized by chemical means. Although there is no difference between the synthetic 2-PE and the natural one, the increasing demand for natural flavors makes biotechnological production of 2-PE an interesting option. 2-PE is a general metabolite of microbial fermentation. Previous studies have shown that microorganisms can convert Lphenylalanine (L-phe) to 2-PE in their culture via the Ehrlich pathway (Hazelwood et al. 2008). Three enzymes, transaminase, decarboxylase, and dehydrogenase, are known in this route which is by transamination of L-phe to phenylpyruvate, followed by decarboxylation to phenylacetaldehyde and reduction to 2-PE. 2-PE biosynthetic pathway in plants, by contrast, is yet to be clearly elucidated. Three plausible pathways are proposed for 2-PE biosynthesis in a variety of different plant species (Tieman et al. 2007). The first pathway consists of an enzyme of CYP79 family responsible for the oxidative decarboxylation of L-phe to produce phenylacetaldoxime. Phenylacetaldoxime is successively hydrolyzed to yield phenylacetaldehyde (PAld) which is reduced to 2-PE by an alcohol dehydrogenase (ADH) or PAld reductase (PAR). PAR has been identified in tomato and is designated as LePAR1 (Tieman et al. 2007). The second pathway was first found in tomato fruits, where an aromatic L-amino acid decarboxylase (AADC) converts L-phe to 2-phenylethylamine (2PNH2). A monoamine oxidase (MAO) further catalyzes the conversion of 2PNH2 to pAld which is then transformed to 2-PE by PAR. The third pathway involves a direct conversion of L-phe to PAld by a bifunctional phenylacetaldehyde synthase (PAAS) which is a petunia AADC reported by Kaminaga et al. (2006), and 2-PE is formed from PAld by PAR. The aim of our present work was to reduce carbon flux to lignin biosynthesis by introducing 2-PE biosynthetic
pathway into Arabidopsis. The enzymes in novel pathways were recruited from Saccharomyces cerevisiae, tomato, and petunia. Among them, ARO9 (transaminase,
Hazelwood et al. 2008) and PAAS will, respectively, compete with PAL for L-phe, the initial monolignol precursor. We expect that the altered carbon flux towards
lignin pathway would reduce lignin biosynthesis. Thus, lignin content, 2-PE and its derivatives were determined in transgenic Arabidopsis plants. Furthermore, we detected
cellulose content, matrix polysaccharide composition, and cell wall saccharification efficiency to study the impact of the reconstructed 2-PE pathway on other cell wall components and biomass recalcitrance of transgenic Arabidopsis plants.

Materials and methods
Plant materials and growth conditions Arabidopsis thaliana ecotype Columbia (Col-0) was used in this study. Arabidopsis plants were grown in greenhouse under 16 h light/8 h dark at 22 C with 65 % relative humidity. Seeds were sterilized before sown on half-strength MS medium. After stratification at 4 C for 2 days, Arabidopsis
seeds were germinated at 22 C. Tomato (S. lycopersicum cv. M82) and petunia (P. hybrida, cv. Mitchell Diploid) plants were grown in greenhouse with day/night
temperatures of 22/17 C under 16 h light/8 h dark. Tomato mature fruits and petunia flowers were harvested and immediately frozen in liquid nitrogen and stored at
-80^。C.

Chemicals and reagents
Taq DNA polymerase and all restriction enzymes were purchased from MDbio (Taiwan) and New England Biolabs (USA). TRIzol reagent for RNA isolation was from Invitrogen (USA). RNase-free DNase and the kits used for cDNA synthesis and RT-PCR were from Thermo Fisher (USA) and TransGen (Beijing, China). The kits used for molecular cloning were from Takara (Japan) or Thermo Fisher (USA). Oligo nucleotides synthesis and DNA sequencing were performed by Sunnybio (Shanghai, China).
The other chemicals used for molecular biology and phytochemistry analysis were purchased from Sigma-Aldrich USA).

Strains and plasmids
The Gateway entry vector, pEN-L4-2-L3, and the plant destination vectors, pK7m34GW2-8m21GW3 and pK7m34GW2-8m21GW3-9m56GW4 were purchased from VIB/Gent (Belgium). The Gateway entry vector pDONR P5-P6 was purchased from Invitrogen (USA). Agrobacterium strain GV3101 was used for plant transformation. S. cerevisiae strain was used for gene cloning.
Gene cloning and vector construction
The ARO9, ARO10, and ADH2 genes were amplified from the genomic DNA of S. cerevisiae using gene-specific
primers as follows: ARO9 (forward, 50-ATGACTGCTGGTTCTGCCCCC-30; reverse, 50-TCAACTTTTATAGTTGTCAAAAAAT-30), ARO10 (forward, 50   ATGGCACCTGTTACAATTGAAAAG-30; reverse, 50-CTATTTTTTATTTCTTTTAAGTGC-30), and ADH2 (forward, 50-ATGTCTATTCCAGAAACTCAAAAAG-30; reverse, 50-TT ATTTAGAAGTGTCAACAACGTATC-30). The amplified fragments were ligated to the Gateway entry vector pDONR P5-P6, pEN-L4-2-L3, and pGWC-T, sequenced, and then transferred into the Gateway binary vector pK7m34GW2-8m21GW3-9m56GW4 using the Gateway recombination system (Invitrogen) (Karimi et al. 2007).
The PAAS and LePAR1 genes were isolated from the cDNAs of petals of P. hybrida (cv. Mitchell) and mature fruits of S. lycopersicum using the gene-specific primers as
follows: PAAS (forward, 50-ATGGATACTATCAAAATCAACCCAG-30; reverse, 50-CTACGCATTCAGCATCATAGTTG-30) and LePAR1 (forward, 50-ATG AGTGTGACAGCGAAAACAGTG-30; reverse, 50-TTACATAGAAGATGAACCTCCAAA-30). The amplified fragmentsof PAAS and LePAR1 were ligated to pEN-L4-2-L3,
and pGWC-T, respectively, and transferred into the Gateway binary vector pK7m34GW2-8m21GW3.
RNA isolation and RT-PCR
Total RNA extraction and RT-PCR was conducted as described previously (Qi et al. 2013). Briefly, 7-week-old Arabidopsis stems were collected and extracted with TRIzol reagent (Invitrogen) according to manufacturer’s instructions. RNA was digested with DNase I (Sigma) to remove genomic DNA contamination, and the first-strand
cDNA was reverse-transcribed with total RNA (2 lg) using RevertAid First-Strand cDNA Synthesis Kit (Thermo Fisher) and oligo-dT primers. Beacon Designer v7.0 (Premier
Biosoft International) was used to design the genespecific primers as follows: ARO9 (forward, 50-TGC CCGTGTCATCCGTTTGG-30; reverse, 50-AAGTTGGACTCAGCCATTGCCTTT- 30), ARO10 (forward, 50-CCCTGGTGATGTTGTCGTTTGTGAAA-30; reverse, 50-ATTGATGTGAGCGTTTGAGTGGTCTTG-30), ADH2
(forward, 50-GTTCAAGCCGCTCACATTCCTCAA-30; reverse,50-TAGACCACCAGCAGCACCAGAA-30), PAAS(forward, 50-CTCAGAAATTTCATAAGAAGC-30; reverse,
50-ATCATAGTTGCATGGTTTCGAA-30) and LePAR1(forward, 50-TCCTCTTTTGGGTGGGTTAACGT-30;reverse, 50-CTCCTTTGATACTTGATAATTTTG-30).The expression of the AtACTIN2 gene was used as aninternal control.
Histochemistry assay
7-week-old Arabidopsis basal stems were cut and fixed with 4 % paraformaldehyde at 4 C overnight. After fixation, the tissues were dehydrated in a graded ethanol series,and embedded in paraplast as described previously (Chaiet al. 2014). The paraplast-embedded stems were sectioned to a thickness of 10 lm using a Leica RM 2235 microtome(Leica). The dewaxed and rehydrated sections were incubated for 5 min in the solution of Phloroglucinol (Sigma) in 20 % HCl and rinsed with water (Pomar et al. 2002). All sections were observed at bright field with an Olympus BX-51 microscope equipped with an OLYMPUS DP26 digital camera and OLYMPUS DP2-BSW software.
Cell wall residues preparation The inflorescence stems of 10-week-old mature senesced tissues were collected 3 cm above the base for cell wall residues (CWRs). The senesced stems were harvested and lyophilized, and the dried materials were then grinded in ball mill (Retsch). The ground-well stem materials were thoroughly washed with chloroform:methanol (2:1), 100 % methanol, 50 % methanol, and MiliQ water, and then dried in vacuum machine (Fu et al. 2011b). De-starching was performed by treating CWR with pullulanase M3 (0.5 U mg-1, Megazyme) and a-amylase (0.75 U mg-1, Sigma) in0.1 M NaOAc buffer (pH 5.0) overnight (Li et al. 2009).Lignin analysis
Total lignin content was determined by the AcBr method (Foster et al. 2010). Briefly, dried-well CWR samples werereacted with freshly prepared acetyl bromide reagent at
50 C for 4 h. After centrifugation at 3500g for 15 min, the upper layer was quantitatively transferred and reacted with2 mol/L NaOH and 0.5 mol/L hydroxylamine. The samples were diluted with acetic acid, and the absorptions at 280 nm were determined with a NanoDrop ND-1000 spectrophotometer (Thermo Scientific). AcBr lignin content was calculated by means of the Bouguer–Lambert– Beer law in five biological duplicates.Cellulose content assay Cellulose content was determined using the method as described previously (Foster et al. 2010). Briefly, CWR was hydrolyzed by trifluoroacetic acid (TFA) at 120 C for 120 min. The TFA resistant materials were treated with Updegraff reagent (acetic acid: nitric acid: water, 8:1:2, v/v) at 100 C for 30 min, and the resulting pellets were completely hydrolyzed using 67 % H2SO4 (v/v). The released glucose was measured using a glucose assay kit (Cayman Chemical, MI) with a dehydration factor of 0.9. Matrix polysaccharide composition analysis
Matrix polysaccharide composition analysis was performed with TFA-hydrolyzed materials as described previously (Yu et al. 2010). The released monosaccharides were
derived by 1-phenyl-3-methyl-5-pyrazolone (PMP), and the derivatives were analyzed by high-performance liquid chromatography (HPLC).Cell wall pretreatment and saccharification Pretreatments and saccharification of CWR of 10-week-old senesced Arabidopsis stems were performed as described previously with minor modifications (Van Acker et al.2013). Ball-milled CWR of senesced stems (50 mg) was incubated in glass culture tubes containing 2 mL water at 30 C for 30 min and autoclaved at 120 C for 1 h. Saccharification was initiated by the addition of 1.5 mL of 100 mM citrate buffer at pH 4.8, 0.5 % w/w cellulase complex NS50013 and 0.5 % w/w glucosidase NS50010 (Novozymes, Bagsværd, Denmark). After 24 h of incubation at 50 C with 100 rpm shaking, the samples were centrifuged at 15,000g for 10 min, and 100 lL of the supernatant was collected for glucose measurement using a glucose assay kit (Cayman Chemical, MI). Quantification of 2-phenylethanol in the transgenic Arabidopsis plants Rosette leaves and 7-week-old stems were collected and frozen individually in liquid nitrogen, and then grinded in ball mill (Retsch) and kept at -80 C. For the 2-PE analyses, samples were extracted using methyl-tert-butyl ether with 0.5 mM benzyl methyl ether as internal standard, and the extracts were individually analyzed with an HP 6890 Series GC System equipped with a RESTEK-5Sil-MS column using the method as described previously (Tieman et al. 2007). 2-PE was quantified based on m/z 122 and 91 extracted ion traces and areas normalized to benzyl methyl ether peak area and quantified using external calibration with authentic 2-PE standard. Each chemical analysis data point is the average of five independent transgenic lines. Phenolics profiling analysis Phenolics profiling analysis was determined using the
method adapted from Fu et al. (2011b). Briefly, samples were extracted with methanol:water (8:2, v/v) containing 0.5 mM naringenin as internal standard and analyzed using liquid chromatography electrospray ionization mass spectrometry (LC–ESI–MS/MS). An Agilent 1290 Infinity LC coupled to a Bruker Esquire Ion-trap Mass Spectrometer equipped with an electrospray ionization source (ESI) system (Agilent Technologies, Palo Alto, CA) was employed. Mass determination was conducted by ESI in negative ion polarity. Mass spectra were recorded over the range 50–2200 m/z. Statistical analysis Triplicate samples were collected for each transgenic line. The mean values were used for statistical analyses. Data from each trait were subjected to one-way ANOVA. The significance of treatments was tested at the P\0.05 level.

Results
Introduction of 2-PE biosynthetic pathway into Arabidopsis plants 2-PE biosynthesis from L-phenylalanine is involved in different pathways in plants and S. cerevisiae. To assess the relative efficiency of each pathway for 2-PE production in Arabidopsis, both pathways were reconstructed and introduced into Arabidopsis, respectively. In S. cerevisiae, the Ehrlich pathway for 2-PE biosynthesis consists of transaminase, 2-keto-acid decarboxylase, and alcohol dehydrogenase (Fig. 1a). Three corresponding genes (ARO9, ARO10, and ADH2) that encoded these enzymes were selected and amplified from the genomic DNA of S. cerevisiae, and constructed in the binary vector pK7m34GW2- 8m21GW3-9m56GW4 (Karimi et al. 2007; Hazelwood et al. 2008), which contains three cassettes with different promoters and terminators that work well in plants (Fig. 1b). The 2-PE biosynthetic pathway in plants requires at least two genes (PAAS and PAR) (Fig. 1a). Accordingly, the PAAS gene was isolated from P. hybrida (cv. Mitchell) petals, and PAR was amplified from S. lycopersicum (LePAR1, Sakai et al. 2007) mature fruits. The two geneswere constructed into the binary vector pK7m34GW2- 8m21GW3, which has two cassettes with different promoters and terminators (Fig. 1b). Arabidopsis plants were transformed with the two vectors, respectively. Independent T1 kanamycin-resistant plants were screened for the insertion of genes of the two pathways using genomic PCR. RT-PCR analysis further revealed high expression levels of the recruited genes in transgenic Arabidopsis plants (Fig. 2b). Five homozygous lines containing ARO9/ARO10/ADH2 or PAAS/LePAR1 were separately selected for further analysis.

Fig. 1 Recruitment of enzymes for construction of 2-phenylethanol biosynthetic pathways in Arabidopsis. a Proposed scheme of routes for 2-phenylethanol biosynthesis. The enzymes ARO9, ARO10, and ADH2 highlighted in purple were recruited from Saccharomyces cerevisiae. The enzymes PAAS and PAR highlighted in red were from P. hybrida and S. lycopersicum, respectively. ARO9, aromaticaminotransferase II; ARO10, 2-keto-acid decarboxylase; ADH2, alcohol dehydrogenase-2; PAAS, phenylacetaldehyde synthase; PAR, phenylacetaldehyde reductase. Phenylalanine ammonia lyase (PAL), cinnamate 4-hydroxylase (C4H) and 4-coumarate coenzyme
A:ligase (4CL) are the enzymes directing carbon flux towards lignin biosynthesis. b Design of the binary vector constructs containing 2-phenylethanol biosynthetic genes isolated from Saccharomyces cerevisiae and plants. Two proposed 2-phenylethanol biosynthetic pathways were recruited, and the genes were inserted into the
Gateway destination vector pK7m34GW2-8m21GW3 and pK7m34GW2-8m21GW3-9m56GW4, respectively. LB, the left border of T-DNA; RB, the right border of T-DNA; NPT II, the kanamycin resistance gene; P35S and t35S, Cauliflower Mosaic Virus promoter and terminator sequences; PROLD, Agrobacterium rhizogenes promoter sequence; tOCS, Agrobacterium tumefaciens octopine synthase terminator sequence; pCSVMV, Cassava vein mosaic virus promoter sequence; tg7, Agrobacterium tumefaciens g7 terminator



更多详细信息请参考
Plant Cell Rep
DOI 10.1007/s00299-015-1790-0