Evaluation of Brachypodium distachyon L-Tyrosine Decarboxylase Using L-Tyrosine Over-Producing Saccharomyces cerevisiae

To demonstrate that herbaceous biomass is a versatile gene resource, we focused on the model plant Brachypodium distachyon, and screened the B. distachyon for homologs of tyrosine decarboxylase (TDC), which is involved in the modification of aromatic compounds. A total of 5 candidate genes were identified in cDNA libraries of B. distachyon and were introduced into Saccharomyces cerevisiae to evaluate TDC expression and tyramine production. It is suggested that two TDCs encoded in the transcripts Bradi2g51120.1 and Bradi2g51170.1 have L-tyrosine decarboxylation activity. Bradi2g51170.1 was introduced into the L-tyrosine over-producing strain of S. cerevisiae that was constructed by the introduction of mutant genes that promote deregulated feedback inhibition. The amount of tyramine produced by the resulting transformant was 6.6-fold higher (approximately 200 mg/L) than the control strain, indicating that B. distachyon TDC effectively converts L-tyrosine to tyramine. Our results suggest that B. distachyon possesses enzymes that are capable of modifying aromatic residues, and that S. cerevisiae is a suitable host for the production of L-tyrosine derivatives.


Introduction
Plants produce various kinds of compounds containing aromatic residues via secondary metabolite pathways, such as the phenylpropanoid biosynthesis pathway [1][2][3][4][5][6][7]. Although a number of plant genes involved in the modification of aromatic residues have been identified, the majority of plant genomes have not been sequenced due to their large sizes compared to those of microbes and are expected to contain numerous novel and unidentified genes.
Brachypodium distachyon is a model plant for cereal crops, such as barley and wheat, and is often used for biological characterization of grass biomass due to its short life cycle, small size, simple transformation procedure and small genome size [8]. Recently, full-length cDNA libraries of B. distachyon were constructed and have been made publically available [9]. However, there are few reports concerning the characterization and application of genes and proteins derived from B. distachyon.
The yeast S. cerevisiae has been widely studied and is commonly used as a model eukaryote. Various heterogeneous genes have been functionally characterized using S. cerevisiae as a host strain [10][11][12][13]. Genetically modified S. cerevisiae strains have also been used in the fermentation industry to produce various compounds, including fuels and organic acids [13,14]. S. cerevisiae has also been used as host for the biosynthesis of aromatic compounds. For example, Kim et al. [15] reported 2-phenylethanol production via the Ehrlich pathway, and Vannelli et al. [16] demonstrated p-hydroxycinnamic acid production using a cytochrome P-450-expressing strain of S. cerevisiae. Koopman et al. successfully produced flavonoid naringenin using genetically engineered S. cerevisiae [17].
The shikimate pathway is a metabolic route for the biosynthesis of aromatic amino acid in microorganisms. The first reaction in this pathway involves the stereo-specific condensation of erythrose-4-phosphate (E4P) and phosphoenolpyruvate (PEP) to 3-deoxy-D-heptulosonate-7-phosphate (DAHP) in a reaction catalyzed by DAHP synthase (Fig 1) [18]. In S. cerevisiae, DAHP synthase is encoded by the ARO3 and ARO4 genes, and the corresponding proteins, ARO3 and ARO4, are strongly regulated by L-phenylalanine and L-tyrosine, respectively, which are produced in this pathway [19]. According to a report by Helmstaedt et al. [20], a single serine-to-alanine substitution in ARO4 at position 195 impairs L-tyrosine sensitivity, leading to deregulation of ARO4. The conversion of chorismate to phenylpyruvate (PPA) by chorismate mutase is another step regulating aromatic amino acid productivity in this pathway [18]. S. cerevisiae chorismate mutase is encoded by ARO7 and its activity is inhibited by L-tyrosine and L-tryptophan; however, the substitution of glycine with serine at position 141 generates L-tyrosine-insensitive ARO7 [19]. Although these findings indicate that enzymes involve in amino acid biosynthesis in S. cerevisiae can be improved through genetic modification, only a few reports have described the application of the in S. cerevisiae biosynthesis pathway for aromatic amino acids for chemical production [15,16,21].
To demonstrate that the genomes of herbaceous biomass such as B. distachyon is a versatile and useful resource for genes involved in the production of aromatic compounds, here, we searched the B. distachyon genome for genes encoding L-tyrosine decarboxylase (TDC), which is involved in alkaloid biosynthesis. In E. coli, tyramine production pathway was previously reported, and TDC gene derived from Lactobacillus brevis JCM1170 was used for tyramine production in that report [22]. Several genes annotated as TDC encoding were identified by screening B. distachyon cDNA libraries and were then evaluated using S. cerevisiae as a host. TDC-expressing S. cerevisiae successfully converted L-tyrosine to tyramine, which is the decarboxylation product of L-tyrosine. By increasing L-tyrosine availability, tyramine productivity by the recombinant S. cerevisiae strain expressing TDC derived from B. distachyon was 6.6-fold higher than that of the control strain.

Plasmid construction and yeast transformation
Polymerase chain reactions (PCR) were performed using PrimeSTAR HS (Takara Bio, Shiga, Japan) and the primer pairs listed in Table 1. PCR cycle conditions were as follows: 98°C for 1 minute followed by 30 cycles of 98°C for 15s, 68°C for 30s, and 72°C for 90s. Plasmids for transformation of S. cerevisiae were constructed by PCR amplifying the identified gene fragments encoding TDC homologs using Bradi1g28960.1, Bradi2g51120.1, Bradi2g51170.1, Bra-di3g14750.1, or Bradi3g14780.1 as a template with the appropriate primer pairs. Each gene was identified using GRAMENE (http://www.gramene.org/) (Brachypodium.org is also available (http://www.brachypodium.org/)). PCR cycle conditions were as follows: 98°C for 1 minute followed by 30 cycles of 98°C for 15s, 68°C for 30s, and 72°C for 90s. Each amplified fragment was introduced into the NheI or SalI, and XmaI sites of pGK422 [21], generating plasmids pGK422-tdc60, pGK422-tdc20, pGK422-tdc70, pGK422-tdc50, or pGK422-tdc80. δ-integrative plasmids were constructed by PCR amplifying the gene fragment encoding LEU2 from pRS405 DNA [22] with LEU2d(F)_InF and LEU2d(R)_InF. The obtained fragment was introduced into the XhoI sites of pδU [23], which contained URA3 as a selective marker, using an In-Fusion HD Cloning kit (Takara Bio), generating the plasmid pδL. A gene fragment containing the PGK1 promoter region was amplified by PCR using pGK422 as a template with the appropriate   primer pair and was then introduced into the PstI and BamHI sites of pδU and pδL using an In-Fusion HD Cloning kit, generating pδU-PGK and pδL-PGK, respectively. The synthetic gene fragments ARO4 fbr and ARO7 fbr were obtained from a commercial source (Invitrogen, San Diego, CA) (see S1 File). A gene fragment encoding ARO4 fbr was PCR amplified using ARO4 fbr as a template with ARO4 fbr _Fw and ARO4 fbr _Rv, and was then introduced into the BamHI sites of pδU-PGK using an In-Fusion HD Cloning kit, generating pδU-ARO4 fbr . The synthetic ARO7 fbr gene fragment was directly introduced into the BamHI sites of pδL-PGK using an In-Fusion HD Cloning kit, generating pδL-ARO7 fbr . Plasmids were transformed into S. cerevisiae using lithium acetate method [24,25], and the resulting transformants are listed in Table 1. The transformants with the highest tyramine or L-tyrosine productivity were selected and used in subsequent experiments.

Culture conditions
A single colony of each S. cerevisiae transformant was inoculated into a test tube containing 5 mL synthetic dextrose (SD) medium containing 2% glucose without adenine, uracil, or leucine as preculture. To evaluate tyramine or L-tyrosine productivity, preculture broth was seeded into 5 mL SD medium containing 2% glucose to give an initial OD 600 value of 0.1. Test tubes were incubated at 30°C for 72 h with agitation at 180 rpm.

Analytical methods
The concentration of ethanol and glucose in the culture supernatant was measured using a BF-5 biosensor (Oji Scientific Instruments, Hyogo, Japan).
For estimation of produced L-tyrosine and tyramine, GC-MS was carried out using a GCMS-QP2010 Ultra (Shimadzu, Kyoto, Japan) equipped with a CP-Sil 8 CB-MS capillary column (30 m x 0.25 mm x 0.25 μm; Agilent). Helium was used as carrier gas to maintain a flow rate of 2.1 ml/min. The injection volume was 1 μl with a split ratio of 1:10. The oven temperature was initially held at 150°C for 5 min, raised to 300°C at 10°C/min, and further maintained at 300°C for 5 min. The total running time was 25 min. The other settings were as follows: 250°C interface temperature, 200°C ion source temperature, and electron impact ionization (EI) at 70 eV. Dried residues of tyramine and tyrosine were derivatized for 60 min at 80°C in 50 μL N-(tert-butyldimethylsilyl)-N-methyl-trifluoroacetamide (MTBSTFA) and 50 μL N, Ndimethylformamide prior to analysis [26,27]. Cycloleucine was used as the internal standard.

Quantification of integrated copy numbers by real-time PCR
The integrated copy number of each recombinant strain was quantified using real-time PCR. Template genomic DNA was isolated from yeast cells cultivated in SD medium for 72 h at 30°C using a GenTLE precipitation carrier (Takara Bio) following the manufacturer's protocol.
The two sets of PCR primers used to detect ARO4 and ARO4 fbr , and ARO7 and ARO7 fbr listed in Table 1. Quantitative real-time PCR was performed using an ABI PRISM 7000 Sequence Detection System (Applied Biosystems, Foster City, CA) with thunderbird SYBR qPCR Mix (Toyobo, Osaka, Japan). The normalized gene copy number was calculated by the relative quantification method with the PGK1 gene as the housekeeping gene.

Results
Cloning and functional expression of the gene encoding B. distachyon Ltyrosine decarboxylase in S. cerevisiae The B. distachyon genome was screened for genes homologous to TDC genes derived from A. thaliana using GRAMENE, and 5 candidate TDC genes (Bradi1g28960.1, Bradi2g51120.1, Bra-di2g51170.1, Bradi3g14750.1, and Bradi3g14780.1) were identified. After cloning each candidate gene into multi-copy vector pGK422, the resulting TDC expression vectors were individually introduced into S. cerevisiae YPH499. Each transformant was cultured in SD medium, and the culture supernatant was analyzed by GC-MS. A specific peak derived from tyramine-tyramine-2TBDMS derivatives (m/z = 144) was observed at approximately 17.2 min in GC-MS spectra of the culture supernatants of YPH499/p422tdc20 and YPH499/p422tdc70, but was not detected in the culture supernatants of the control strain, YPH499/p422, or those of YPH499/p422tdc60, YPH499/p422tdc50, and YPH499/p422tdc80 (data not shown). YPH499/p422tdc20 and YPH499/p422tdc70 produced 20 and 25 mg/L tyramine, respectively, in medium containing 2% glucose as the carbon source. The results of these analysis demonstrated that the B. distachyon transcripts Bradi2g51120.1 and Bradi2g51170.1 encoded a gene encoding TDC.

Construction of a L-tyrosine over-producing S. cerevisiae strain
To increase tyramine productivity in S. cerevisiae, we attempted to construct a strain that overproduces L-tyrosine by introduction of the enzymes, ARO4 and ARO7, which regulate L-tyrosine biosynthesis in S. cerevisiae [19,20], into YPH499. After the construction of YPH499/ δUARO4 fbr , the gene encoding ARO7 fbr was introduced into that transformant. Both ARO4 and ARO7 were integrated into the genome of YPH499 using the δ-integration method [24]. YPH499/δU/δL, YPH499/δU/δLARO7 fbr , YPH499/δUARO4 fbr /δL and YPH499/δUAR-O4 fbr /δLARO7 fbr were cultured in SD medium containing 2% glucose, and the culture supernatants were analyzed by GC-MS to quantify the amount of L-tyrosine produced after 72 h cultivation (Fig 2A). A total of 0.80, 1.99 and 4.84 mg/L tyrosine was found in the culture supernatant of YPH499/δU/δLARO7 fbr , YPH499/δUARO4 fbr /δL and YPH499/δUARO4 fbr /δLAR-O7 fbr , respectively, whereas no tyrosine was detected in the culture supernatant of the control strain YPH499/δU/δL. The copy number of ARO4 fbr integrated into the genome of YPH499/ δUARO4 fbr /δL and YPH499/δUARO4 fbr /δLARO7 fbr , which originated from YPH499/δUAR-O4 fbr /δL, was estimated by real-time PCR to be 2 in all strains. In contrast, the copy number of ARO7 fbr in the genome of YPH499/δUARO4 fbr /δLARO7 fbr was estimated to be approximately 20, whereas that of YPH499/δU/δLARO7 fbr was approximately 9 (Fig 2(B)).

Discussion
Plants accumulate large numbers of compounds that contain aromatic residues, such as phenylpropanoids, flavonoids, coumarins, and alkaloids, via secondary biosynthesis pathways [1][2][3][28][29][30]. The structural diversity of aromatic compounds produced in plants is realized through sets of enzyme superfamilies, such as oxygenases, ligases, and decarboxylases [6]. For example, (S)-norcoclaurine, which is an intermediate of the benzylisoquinoline alkaloid  biosynthetic pathway, is synthesized from two molecules containing L-tyrosine modified with hydroxyl groups in the benzene ring through reactions catalyzed by aromatic amino acid decarboxylase and monooxygenase [3]. L-tyrosine derivatives can be converted to various compounds due to the hydroxyl group at the para position, and categorized into important parts in aromatic compounds. As various types of enzymes capable of modifying aromatic residues are found in plants, an increasing number of enzymes involved in the synthesis aromatic compounds will be identified as the genomes sequences of more plants become available.
Recently, the complete genome of B. distachyon was sequenced and used to construct fulllength cDNA libraries [9]. To demonstrate that B. distachyon is a useful gene resource, we here focused on the B. distachyon genome for homologs of TDC, which catalyzes the decarboxylation of L-tyrosine and is involved in the production of aromatic compounds [31,32]. It is suggestive that the transcripts Bradi2g51120.1 and Bradi2g51170.1 encode enzymes with Ltyrosine decarboxylation activity, and the corresponding genes were identified as novel TDC genes of B. distachyon.
The activity of B. distachyon TDC was further evaluated by constructing an L-tyrosine overproducing strain of S. cerevisiae. In the biosynthesis pathway of L-tyrosine in S. cerevisiae, ARO4 and ARO7 (ARO4 fbr and ARO7 fbr ) are key enzymes that regulate L-tyrosine productivity and are subject to feedback inhibition by the produced L-tyrosine [19,20]. Here, genes encoding L-tyrosine-insensitive ARO4 and ARO7 mutants were introduced into the genome of S. cerevisiae YPH499 strain using the δ-integration method. Helmstaedt et al. reported L-tyrosine-insensitive ARO4 fbr [20], whereas ARO7 fbr was previously constructed by Luttik et al. [19]. As shown in Fig  2(A), the L-tyrosine productivity of YPH499/δUARO4 fbr /δL and YPH499/δU/δLARO7 fbr was higher than that of YPH499/δU/δL. Quantitative real-time PCR analysis revealed that 2 copies of ARO4 fbr were introduced into the genome of YPH499/δUARO4 fbr /δL, whereas approximately 10 copies of ARO7 fbr genes were integrated into the YPH499/δU/δLARO7 fbr genome (Fig 2(B)). Together, these findings indicate that ARO4 fbr enhances L-tyrosine productivity more efficiently than ARO7 fbr (Fig 2(A)). This result may be attributed to the low availability of intracellular chorismate in YPH499/δU/δLARO7 fbr compared to that in ARO4 fbr -expressing strains. ARO4 fbr catalyzes the specific condensation of E4P and PEP into chorismate in the first step of the shikimate pathway, and the subsequent dislocation reaction is catalyzed by ARO7 fbr (Fig 1). As YPH499/ δU/δLARO7 fbr expresses L-tyrosine sensitive ARO4, the formation of chorismate is strongly regulated by the produced L-tyrosine, which would therefore limit the available chorismate in this strain. Consistent with this speculation, the amount L-tyrosine produced by YPH499/δUAR-O4 fbr /δLARO7 fbr reached 4.84 mg/L in the culture supernatant, whereas YPH499/δU/δL did not produce L-tyrosine at detectable levels. We also investigated the correlation between L-tyrosine productivity and the copy number of ARO4 fbr or ARO7 fbr . Although the copy number of ARO4 fbr affected L-tyrosine productivity in the case of ARO4 fbr (See S2 File), L-tyrosine productivity wasn't proportional to the copy number of ARO7 fbr in the case of ARO7 fbr (See S3 File).
The TDC homolog of B. distachyon encoded by Bradi2g51170.1 was functionally characterized by introduction into YPH499/δUARO4 fbr /δLARO7 fbr . YPH499/δUARO4 fbr /δLARO7 fbr / tdc70 was cultured using SD medium containing 2% glucose as the carbon source. As shown in Fig 3(D), 200 mg/L tyramine was produced by YPH499/δUARO4 fbr /δLARO7 fbr /tdc70, a level that was 6.6-fold higher than that of YPH499/δU/δL/tdc70 as the control strain. Based on the cell density, and protein and amino acid compositions of S. cerevisiae, we estimated the flux distribution rates of L-tyrosine into tyramine and biomass in each transformant after 72 h of cultivation [33,34]. With increasing L-tyrosine productivity, the ratio of L-tyrosine distributed into tyramine was increased ( Table 2). As L-tyrosine was not detected in the culture supernatant of the tyramine-producing strains, free L-tyrosine was thought to be completely converted to tyramine. These findings indicate that one of the rate-limiting steps of tyramine production remains L-tyrosine availability. Thus, the TDC encoded by Bradi2g51170.1 may be a promising enzyme for the microbial production of aromatic compounds. We also attempted to express a candidate TDC derived from A. thaliana in S. cerevisiae; however, A. thaliana TDC could not be expressed using our expression system (data not shown). As shown in Fig 3(A)-3(C), the cell growth, glucose consumption, and ethanol production rates of YPH499/δUARO4 fbr /δLARO7 fbr /tdc70 and YPH499/δUARO4 fbr /δL/tdc70 were higher than those of YPH499/δU/δL/tdc70 and YPH499/δU/ δLARO7 fbr /tdc70. These results may be attributed to the greater carbon flux in the glycolysis pathway resulting from the expression of ARO4 fbr , which promotes the condensation of PEP and E4P. In this study, we transformed pGK422-tdc70 into two different ARO4 fbr -and ARO7 fbrexpressing backgrounds. As a result, tyramine productivity, cell growth rates, glucose consumption rate, and ethanol production rates were almost the same among them (See S4 File). Using YPH499/δUARO4 fbr as the parent strain, TDC20 was also evaluated and compared to TDC70. Tyramine productivity of TDC70 was slightly higher than that of TDC20 (See S5 File).
In conclusion, we screened the genome of B. distachyon for genes encoding TDC, which is an enzyme involved in the modification of aromatic compounds, and identified two putative genes encoding TDC using S. cerevisiae as a host strain. This result implies that B. distachyon has high potential as a genetic resource for the microbial production of aromatic compounds. Although aromatic compounds have reportedly been produced using S. cerevisiae, the yield of L-tyrosine derivatives, such as alkaloids, was very low [21]. We speculate that the L-tyrosine over-producing strain constructed here may be applicable to the production of L-tyrosine derivatives with complicated structures.
Supporting Information S1 File. The nucleotide sequences of synthetic ARO4 fbr and ARO7 fbr genes (Under lines indicate open reading frame, capital letters indicate the nucleotide sequences substituted in order to deregulate feedback inhibition, and italic characters indicate flag-tag sequence). (DOCX) S2 File. Correlation between L-tyrosine productivity and the copy number of ARO4 fbr . YPH499/δUARO4 fbr /δL (Y; YPH499 (control), 1; colony 6, 2; colony 8, 3; colony 9).1 copy number of ARO4 fbr was integrated into the genome of colony 6 and 8, whereas 2 were colony 9, which was adopted for further experiments in this study. (DOCX) S3 File. Correlation between L-tyrosine productivity and the copy number of ARO7 fbr . Results of YPH499/δUARO4 fbr /δLARO7 fbr (Y; YPH499/δUARO4 fbr /δL (control), 1; colony 3, 2; colony 5, 3; colony 18, 4; colony adopted in this study). Gray bar indicates L-tyrosine productivity per OD 600 , and black bar indicates normalized integrated copy number of ARO7 and ARO7 fbr . (DOCX) Table 2. Flux distribution of L-tyrosine produced in each transformant to tyramine and biomass (all produced L-tyrosine was considered to be converted to tyramine except for the proportion incorporated into biomass). *A flux value to tyrosine building biomass was determined from OD 600 values and its conversion coefficient to dry cell weight (0.25 g-DCW/L/OD 600 ) by using the composition ratio of L-tyrosine in biomass [33,34]. The flux was estimated as tyrosine concentration of culture (mmol/L). Flux distributions between tyramine and biomass from tyrosine were estimated from each concentration.