Genes and Enzymes Involved in Caffeic Acid Biosynthesis in the Actinomycete Saccharothrix espanaensis
http://www.100md.com
《细菌学杂志》
Pharmazeutische Biologie und Biotechnologie, Institut fur Pharmazeutische Wissenschaften, Albert-Ludwigs-Universitt Freiburg im Breisgau, Stefan-Meier-Strae 19, 79104 Freiburg, Germany,Institut fur Pharmazeutische Biotechnologie, Universitt des Saarlandes, Postfach 151150, 66041 Saarbrucken, Germany,Combinature Biopharm AG, Robert-Rssle-Strae 10, 13125 Berlin, Germany
ABSTRACT
The saccharomicins A and B, produced by the actinomycete Saccharothrix espanaensis, are oligosaccharide antibiotics. They consist of 17 monosaccharide units and the unique aglycon N-(m,p-dihydroxycinnamoyl)taurine. To investigate candidate genes responsible for the formation of trans-m,p-dihydroxycinnamic acid (caffeic acid) as part of the saccharomicin aglycon, gene expression experiments were carried out in Streptomyces fradiae XKS. It is shown that the biosynthetic pathway for trans-caffeic acid proceeds from L-tyrosine via trans-p-coumaric acid directly to trans-caffeic acid, since heterologous expression of sam8, encoding a tyrosine ammonia-lyase, led to the production of trans-p-hydroxycinnamic acid (coumaric acid), and coexpression of sam8 and sam5, the latter encoding a 4-coumarate 3-hydroxylase, led to the production of trans-m,p-dihydroxycinnamic acid. This is not in accordance with the general phenylpropanoid pathway in plants, where trans-p-coumaric acid is first activated before the 3-hydroxylation of its ring takes place.
INTRODUCTION
Saccharothrix is a genus of gram-positive bacteria belonging to the well-known order Actinomycetales. Most agents used at present for the treatment of bacterial infections were discovered in members of the Actinomycetales. Saccharothrix espanaensis produces the two heptadecaglycoside antibiotics saccharomicins A and B, which represent a new class of antibiotics (15, 18). They exhibit potent antibacterial activity both in vitro and in vivo against multiply-resistant strains of Staphylococcus aureus as well as vancomycin-resistant enterococci (25). The saccharomicins consist of an oligosaccharide portion and the intriguing aglycon N-(m,p-dihydroxycinnamoyl)taurine (Fig. 1), in which caffeic acid is linked to the amino sulfonic acid taurine via an amide bond.
Enzymes belonging to the group of ammonia-lyases catalyze the conversion of -amino acids into ,-unsaturated acids by elimination of ammonia. Ubiquitous in plants and fungi, phenylalanine ammonia-lyase (PAL) (EC 4.3.1.5) catalyzes the nonoxidative deamination of the primary amino acid L-phenylalanine to trans-cinnamic acid (trans-cinnamate), which is the first reaction of the so-called general phenylpropanoid pathway in plants (8). Phenylpropanoids include several important natural product classes, for example, flavonoids, lignins, and coumarins. In monocotyledons, PAL utilizes L-tyrosine in addition to L-phenylalanine (resulting in trans-p-coumaric acid), whereas the enzyme from dicotyledons converts only L-phenylalanine sufficiently. Both PAL and tyrosine ammonia-lyase (TAL) activity are very rare in bacteria (17, 30).
The next reactions of the three-step general phenylpropanoid pathway are catalyzed by the enzymes trans-cinnamate 4-monooxygenase (also called cinnamate 4-hydroxylase; EC 1.14.13.11), leading to trans-p-coumaric acid (trans-4-coumarate), and 4-coumarate-coenzyme A (CoA) ligase (EC 6.2.1.12), leading to 4-coumaroyl-CoA (8).
In this report, we describe the cloning and identification of two genes from S. espanaensis which are involved in caffeic acid biosynthesis. Heterologous expression of sam8, encoding a TAL, led to the production of trans-p-coumaric acid. Coexpression of sam8 and sam5, encoding a 4-coumarate 3-hydroxylase (Coum3H), led to the production of trans-caffeic acid. An enzyme assay performed with recombinant Sam8 verified unambiguously that L-tyrosine (and not L-phenylalanine) is its natural substrate. This is the first report of a Coum3H and the first report of a TAL in the Actinomycetales.
MATERIALS AND METHODS
Strains, vectors, culture conditions, and DNA manipulation. S. espanaensis was obtained from DSMZ (Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH, Braunschweig, Germany; DSM 44229). Escherichia coli XL1-Blue MRF (Stratagene) and the E. coli cloning vector pBluescript SK(–) (Stratagene) were used for subcloning. pUWL201 was a gift from U. Wehmeier and W. Piepersberg (Department of Chemical Microbiology, University of Wuppertal, Germany) (7). The construction of the PKSII mutant strain S. fradiae XKS has been described previously (28).
Protoplast formation, transformation, and regeneration of protoplasts from S. fradiae XKS were carried out by standard procedures. Isolation of plasmid DNA from E. coli, restriction, and ligation were performed according to the protocols of the manufacturers of the kits and enzymes (Promega, QIAGEN).
Construction of a genomic library of S. espanaensis and screening. An S. espanaensis genomic cosmid library was constructed in E. coli DH5 using pOJ436 as cosmid vector. For DNA extraction, mycelium was embedded in agarose and then the DNA was partially digested and isolated, yielding fragments with an average size greater than 35 kb. Robotically produced high-density colony arrays (Hybond N+; Amersham Biosciences) were utilized for the screening of 2,304 cosmid clones with a strain-specific sugar biosynthesis probe (2,3-dehydratase) according to standard nonradioactive hybridization procedures. The primers for the amplification of a gene-specific 2,3-dehydratase probe were 5'-CAGGCSACSWSSAACTACAC-3' and 5'-SWRGAASCGSCCSCCCTCCTC-3'.
DNA sequencing and computer-assisted sequence analysis. Nucleotide sequences were determined at 4base lab GmbH (Reutlingen, Germany) and at SEQLBAB (Sequence Laboratories Gttingen GmbH, Gttingen, Germany) by using either standard primers (M13 universal and reverse, T3, and T7) or customized, internal primers. Computer-assisted sequence analysis was done with DNASIS software (version 2.1; Hitachi Software Engineering). Artemis software, created by the Wellcome Trust Sanger Institute (Cambridge, England) and available at the website http://www.sanger.ac.uk/Software/Artemis/, was used to identify potential protein-coding regions. Database comparison was performed with the BLAST search tools on the server of the National Center for Biotechnology Information, National Library of Medicine, NIH (http://www.ncbi.nlm.nih.gov/).
Production of trans-p-coumaric acid in S. fradiae XKS. Gene sam8 was amplified with Pfu polymerase (Promega) from cosmid 1B11 using primers 5'-CCGCTGGGAATTCTGCGGCACGG-3' (EcoRI restriction site underlined) and 5'-GCCCTCGTTCTAGACCCGTGCGG-3' (XbaI restriction site underlined), resulting in a 1,639-bp product. The fragment was cloned into the EcoRI-XbaI cloning site of the vector pBluescript SK(–) to yield pMB1, whose sequence was confirmed by sequencing. The pMB1 insert was then transferred into the EcoRI-XbaI cloning site of the expression vector pUWL201, which contains the erythromycin resistance promoter ermE to create pMB3. The nonmethylating E. coli strain ET12567 was used to obtain unmethylated DNA. The unmethylated plasmid pMB3 and vector pUWL201 (negative control) were then introduced into S. fradiae XKS by protoplast transformation.
The strain was grown on R2YE agar plates at 28°C for 16 h. (All of the following incubations were carried out at 28°C.) The plates were then overlaid with NB soft agar containing thiostrepton (concentration on the plate, 50 μg ml–1) and incubated for another 4 days. For the selection of transformants, single colonies were transferred onto thiostrepton-containing HA agar plates (30 μg ml–1). After 3 days of incubation, several flasks of HA liquid medium (100 ml) with a thiostrepton concentration of 15 μg ml–1 were inoculated with colonies from these plates and put on a rotary shaker (180 rpm). Culture samples (each 1 ml) were taken over a period of 5 days (24, 48, 72, 96, and 120 h after inoculation). The samples were extracted with ethyl acetate (each 1 ml), vacuum dried, and dissolved in acetonitrile/water (30/70 [vol/vol]; 250 μl).
HPLC/ESI-MS analysis of fermented cultures. High-performance liquid chromatography (HPLC)/electrospray ionization mass spectrometry (ESI-MS) was performed on an Agilent 1100 Series LC/MSD system equipped with an Agilent Zorbax SB-C18 column (5-μm particle size, 4.6 by 150 mm) maintained at 23°C and an Agilent Zorbax SB-C18 precolumn (5-μm particle size, 4.5 by 12.5 mm). Each extract was analyzed at a flow rate of 0.5 ml min–1 with solvent A (99.5% H2O, 0.5% acetic acid) and solvent B (acetonitrile), using the following gradient: 0 to 3 min, 10% B; 3 to 6 min, linear to 20% B; 6 to 9 min, linear to 30% B; 9 to 12 min, linear to 50% B; 12 to 18 min, linear to 95% B. The UV detection wavelength was 310 nm, and the MS scan range was 100 to 300 m/z.
Production of trans-caffeic acid in S. fradiae XKS. Gene sam5 was amplified with Pfu polymerase (Promega) from cosmid 1B11 using primers 5'-CCGCGTTCTAGACCAAGCTTCACCTCAGC-3' (HindIII restriction site underlined) and 5'-GCGCGGGAATTCATCGGGTGC-3' (EcoRI restriction site underlined), resulting in a 1,662-bp product. The fragment was cloned into the HindIII-EcoRI cloning site of the vector pBluescript SK(–) to yield pMB11, whose sequence was confirmed by sequencing. The pMB11 insert was then transferred into the HindIII-EcoRI site of pMB3 to create pMB9. For the construction of pMB10, the pMB11 insert was put into the HindIII-EcoRI site of pUWL201. All of the following procedures were carried out as described for pMB3 (see above). In the case of the expression experiment with pMB10, the flasks of HA liquid medium (100 ml) were supplemented with coumaric acid: p-coumaric acid was added at concentrations of 25, 18.75, 12.5, and 6.25 μg ml–1.
Isolation of trans-caffeic acid from fermented cultures. As a preculture, 100 ml of HA liquid medium with a thiostrepton concentration of 15 μg ml–1 was inoculated with one of the transformants obtained after heterologous expression carried out with pMB9 (see above). After 3 days of incubation, 10 Erlenmeyer flasks each containing 100 ml of HA liquid medium (supplemented with thiostrepton, 15 μg ml–1) were inoculated with 250 μl of the preculture and put on a rotary shaker (180 rpm). After 3 days of incubation, the 10 cultures were merged and then harvested by centrifugation. After adjusting the pH to 3.2, the supernatant (approximately 900 ml) was extracted with 900 ml of ethyl acetate, using a separating funnel. The aqueous phase was extracted again with 900 ml of ethyl acetate. The merged organic phase (1,800 ml) was washed twice with 180 ml of distilled H2O and subsequently with 180 ml of a saturated sodium chloride solution. A couple of spoonfuls of anhydrous sodium sulfate were added to remove further water, before the ethyl acetate was vacuum dried. The residue (dried crude extract) was then dissolved in 3 ml of solvent (dichloromethane-ethyl acetate-acetic acid, 50:50:0.5 [vol/vol/vol]).
A glass column (diameter, 30 mm), filled with a suspension of 7.5 g of silica gel 60 (Merck, Darmstadt, Germany) in the solvent mentioned above, was loaded with the sample. All 10 collected eluate fractions (the volume was approximately 70 ml each) were vacuum dried and then dissolved in 2 ml of solvent. To determine the fraction containing trans-caffeic acid, 20 μl of each fraction was put on a silica gel 60 F254 thin-layer chromatography plate (Merck). Commercially available trans-caffeic acid was used as a reference.
Preparative HPLC. Preparative HPLC was performed on a system consisting of a Waters Delta 600 pump unit, a Waters 600 controller, a Waters 2487 Dual absorbance detector, a Waters Fraction Collector III, and a Waters In-Line Degasser AF. The system was equipped with a Waters XTerra Prep MS C18 column (5-μm particle size, 7.8 by 150 mm) and an XTerra VP-1 precolumn. All injections were carried out at room temperature. The injected extract (20 μl per injection) was analyzed at a flow rate of 2.0 ml min–1 with solvent A (99.5% H2O, 0.5% acetic acid) and solvent B (99.5% acetonitrile, 0.5% acetic acid), using the following gradient: 0 to 3 min, 10% B; 3 to 9 min, linear to 30% B; 9 to 12 min, linear to 50% B; 12 to 18 min, linear to 95% B. The UV detection wavelength was 310 nm. The fraction collector was set at 15 s of collection time per fraction. The fraction containing trans-caffeic acid was determined by its UV spectrum, with an Amersham Biosciences Ultrospec 2100 pro photometer.
NMR characterization of trans-caffeic acid. All 25 fractions containing trans-caffeic acid were merged. After vacuum drying and subsequent lyophilization, the dried extract was dissolved in methanol-d4 (CD3OD). The nuclear magnetic resonance (NMR) data for both the sample and commercially available trans-caffeic acid were recorded by Varian Unity-300 equipment at 300 MHz. The sample's 1H-NMR data are summarized as follows: 1H-NMR (300 MHz, CD3OD) : 6.24 (1H, d, J = 15 Hz; C2-H); 6.80 (1H, d, J = 8.5 Hz; C5'-H); 6.95 (1H, dd, J = 8.5 Hz; C6'-H); 7.05 (1H, d, J = 2.1 Hz; C2'-H); 7.54 (1H, d, J = 15 Hz; C3-H).
Production of recombinant TAL. The coding sequence of Sam8 was PCR amplified with cloned Pfu polymerase (Stratagene) from plasmid pMB1 using primers TALsx_startBgl (5'-GACGCAGGAGATCTGTGGAACGTCAGGC-3', BglII restriction site underlined) and TALsx_stopEco (5'-TCATCCGAGAATTCTCCTTCCCGTCTG-3', EcoRI restriction site underlined) and ligated into the pGEX6p1 vector (Amersham Biosciences) via the BamHI and EcoRI restriction sites to yield plasmid pGEX6p1-TALsx. The expression plasmid was sequenced to verify the correctness of the PCR amplification of the 1,523-bp insert and transferred into E. coli BL21(DE3) cells. Production cultures of 0.6-liter volume were inoculated 1:100 with an overnight culture and grown to an optical density at 600 nm of 0.7 at 30°C (200 rpm) in a 3-liter Erlenmeyer flask using 2YT medium that contained 100 μg ml–1 ampicillin. After induction of protein expression with 0.1 mM isopropyl--D-thiogalactopyranoside, cells were incubated at 30°C (200 rpm) for another 2 h. The cells were then harvested by centrifugation at 4°C (5,000 rpm), resuspended in 12 ml of phosphate-buffered saline buffer (140 mM NaCl-2.7 mM KCl-10 mM Na2HPO4-1.8 mM KH2PO4, pH 7.3), and passed twice through a French pressure cell. Cell debris was sedimented by centrifugation, and the supernatant was loaded onto a glutathione Sepharose 4B gel matrix. Washing and elution of fusion protein with reduced glutathione or, alternatively, on-column cleavage with PreScission protease was carried out according to the manufacturer's protocol (GST Gene Fusion System Handbook; Amersham Biosciences).
TAL enzyme activity assay. Enzyme activity assays contained 20 μg of purified protein and were performed in 1-ml volumes using 0.1 M Tris buffer (pH 8.8) at 30°C and a substrate concentration of 2 mM (Tyr) or 10 mM (Phe, His). Identification of products was carried out via reverse-phase HPLC separation (solvent delivery and DAD detection system from Dionex). For this purpose, 15 μl of acetic acid was added to each assay mixture, followed by extraction with 1 ml of ethyl acetate. After removal of the solvent in vaccuo, samples were dissolved in 100 μl of methanol and loaded onto a Nucleodur C18 column (Macherey-Nagel; 5-μm particle size, 125 by 2 mm). The solvent system consisted of H2O (A) and acetonitrile (B), each containing 0.1% formic acid. Using a gradient of 15 min running from 1% to 99% solvent B, separation of products was observed at the following retention times (detection wavelength given in parentheses): coumaric acid, 7.8 min (310 nm); cinnamic acid, 10.0 min (295 nm). Identification of products was also confirmed by LC-MS methods (Bruker HCTplus coupled to an Agilent 1100 series HPLC system; Nucleodur C18 column [3-μm particle size, 125 by 2 mm]; same HPLC conditions as above; detection in positive ionization mode).
For the determination of kinetic parameters, 2 μg of purified protein was incubated for 30 min at 30°C in 600 μl of Tris buffer (0.1 M, pH 8.8) containing 0.01 to 0.04 mM L-tyrosine or 0.5 to 10 mM L-phenylalanine (in the latter case, 4 μg of enzyme was used, with a 2-h incubation time). Reaction progress was monitored photometrically at 310 nm (tyrosine as substrate) or 290 nm (phenylalanine as substrate) with an AnalytikJena Specord 50 UV/Vis photometer. The pH dependency of initial rates of formation was assayed using 0.1 M Tris buffer at pH 7.0, 8.0, and 8.8 and 50 mM CHES buffer at pH 8.8, 9.2, and 10. For the calculation of product concentrations and conversion into initial velocities, a calibration curve was generated by using serial dilutions of commercially available trans-cinnamic acid and trans-p-coumaric acid in 0.1 M Tris buffer. Nonlinear regression and curve fitting was carried out with the SigmaPlot enzyme kinetics module (Systat Software Inc.). All deduced parameters are based on measurements in three replicates.
To investigate the formation of -tyrosine, a 50-μl aliquot from an assay mixture containing 20 μg of purified Sam8 was subjected to derivatization with OPA-NAC reagent (3). The reagent was prepared by addition of 15 mg of o-phthaldialdehyde and 15 mg of N-acetyl-L-cysteine (Fluka) to 0.5 ml of ethanol and subsequent dilution with 11 ml of 0.4 M borate buffer, pH 10.2. The sample aliquot was mixed with 200 μl of reagent for 10 min at room temperature prior to HPLC analysis. The mobile phase consisted of 50 mM sodium phosphate buffer, pH 6.5 (solvent A), and a mixture of phosphate buffer and methanol (35:65) containing 5% tetrahydrofuran (solvent B). Using a linear gradient from 35% B to 40% B in 24 min on a Nucleodur C18 column (5-μm particle size, 125 by 2 mm), separation of the diastereomeric OPA-NAC derivatives of commercially available (R/S)--tyrosine could be achieved. The obtained retention times were as follows: L--tyrosine, 16.5 min; (R)--tyrosine, 12 min; (S)--tyrosine, 15 min. UV/visual detection was at 330 nm.
GenBank accession number. The GenBank accession number of the DNA sequence reported in this paper is DQ357071.
RESULTS
Cloning and sequencing of a locus containing caffeic acid biosynthetic genes. In order to identify possible genes involved in caffeic acid biosynthesis, an S. espanaensis genomic cosmid library was generated in E. coli. Since caffeic acid is a saccharomicin substructure, it was considered likely that genes for its biosynthesis adjoin other genes of the saccharomicin biosynthetic gene cluster. As saccharomicin contains several 2,6-dideoxyhexoses, a dNDP-glucose 2,3-dehydratase gene was expected to be involved in its biosynthesis. Consequently, the cosmid library was screened with a 2,3-dehydratase gene probe, resulting in 34 positively hybridizing cosmids.
Restriction analysis of the isolated cosmids showed that most cosmids contained DNA overlapping with DNA from other cosmids. By comparing the BamHI fragment pattern, as a result, 22 cosmids could be combined into three subgroups; a cosmid map was created for each of these subgroups. Four cosmid inserts, at least one insert from each subgroup, were subcloned using the restriction endonuclease BamHI. In the case of cosmid 2L12, random sequencing of subclones revealed significant sequence similarity to deoxysugar biosynthetic genes, glycosyltransferase genes, and bacterial ammonia-lyase genes.
Restriction mapping also indicated that cosmid 2L12 did not contain the entire expected ammonia-lyase gene. To detect cosmids possibly containing the whole gene sequence, further screening of the cosmid library was performed, using a DNA fragment as a gene probe consisting of DNA subcloned from the respective end of cosmid 2L12. This resulted in the isolation of the overlapping cosmid 1B11. Finally, a 14-kb DNA region located on both cosmids was sequenced.
Sequence analysis. Analysis of the final contiguous 14-kb DNA sequence revealed the presence of eight open reading frames, named sam1 to sam8. Table 1 lists the putative catalytic functions of the proteins deduced from sam8, sam5, and sam7, which are thought to be involved in the biosynthesis of the saccharomicin aglycon. The deduced protein of sam8 is similar to bacterial histidine ammonia-lyases (HALs) and as well to plant PALs. Sam5 strongly resembles phenol hydroxylases from various bacteria. The product of sam7 shows similarity to acyl-CoA synthetases. Sam6 resembles bacterial dNDP-glucose synthases. The deduced proteins of sam1 to sam4 show similarity to proteins of unknown functions.
Amino acid sequence alignments of several members of the ammonia-lyase enzyme group revealed that the protein encoded by sam8 was more similar to bacterial HALs in size and sequence than to eukaryotic PALs. A conserved Ala-Ser-Gly segment which is typical for ammonia-lyases can also be found in sam8 (4, 24). Furthermore, several active-site amino acid residues proposed for HAL after elucidation of the X-ray structure of HutH, a HAL from Pseudomonas putida, have analogous residues in Sam8 (22, 24).
Heterologous expression experiments. The pUWL201-based plasmids pMB3, containing sam8, and pMB9, containing sam8 and sam5, were heterologously expressed in S. fradiae XKS (28), a PKSII mutant strain from the urdamycin A producer S. fradiae Tu2717, which has shown to be a very convenient host for expression experiments. Organic extracts of the transformants were analyzed for the production of trans-cinnamic acid, trans-p-coumaric acid, and trans-caffeic acid, respectively, by HPLC/ESI-MS analysis using the corresponding commercially available reference standards. The expressions resulted in the production of trans-p-coumaric acid (Mr = 164.16) and trans-caffeic acid (Mr = 180.16) (Fig. 2). Neither of these compounds was detectable in S. fradiae XKS transformed with the expression vector pUWL201. Cinnamic acid, the product of the PAL reaction, could not be detected in any extract of the transformants. Caffeic acid was also produced after heterologous expression of pMB10 (pUWL201-based plasmid containing sam5) and addition of coumaric acid to the transformant cultures (data not shown).
These results verified that the gene product of sam8 is a TAL and the gene product of sam5 is a Coum3H.
Isolation of trans-caffeic acid from fermented cultures and NMR characterization. The organic extract of one of the transformants was purified by silica gel column chromatography. The eluate fraction containing caffeic acid, as proven by TLC, was then used to isolate trans-caffeic acid by preparative HPLC. As expected, 1H-NMR spectroscopy proved that the sample contained trans-caffeic acid. Apart from the trans-caffeic acid (signals were assigned by comparison with the signals of a solution of commercially available trans-caffeic acid), supposable cis isomers and truxillic acid derivatives likely being generated during the extraction process could also be detected. These known cis-trans isomerizations of trans-p-coumaric acid, trans-caffeic acid, and its derivatives under light exposure have been reported by several authors (9, 11). The same sample and the reference were subsequently characterized by H,H- correlation NMR spectroscopy analysis, whereby the presence of trans-caffeic acid could again be verified.
Production of recombinant TAL and enzyme assay. Sam8 was overproduced in E. coli as an N-terminal fusion to glutathione S-transferase (GST), linked by the recognition site for PreScission protease. After purification of the 80-kDa fusion protein and cleavage from the GST tag by protease treatment, the purified protein (calculated Mr, 55,084 Da) was obtained in a yield of approximately 2.5 mg liter–1 of culture (Fig. 3A).
Ammonia-lyase activity was assayed in analogy to published procedures (4, 5). The substrates tested were L-tyrosine, L-phenylalanine, and L-histidine at concentrations up to 10 mM. While no significant activity with histidine was observed, incubation with tyrosine or phenylalanine resulted in the production of coumaric acid and cinnamic acid, respectively. Identity of these products was verified by HPLC analysis in comparison with authentic reference standards.
The determination of kinetic parameters for both enzymatic activities of Sam8 was carried out by a spectrophotometric assay. Initial velocities of product formation were fitted to the Michaelis-Menten equation by the direct nonlinear regression method. The calculated Km values are 15.5 ± 0.3 μM for tyrosine and 2.86 ± 0.3 mM for phenylalanine. They resemble the published values for TAL from Rhodobacter capsulatus (17) and are also strikingly similar to the Km values for PAL from Petroselinum crispum, though with a reversed substrate preference (Table 2). The kcat values for Sam8 are somewhat lower than the turnover numbers for the other known enzymes listed in Table 2 that exhibit TAL or PAL activity; they are calculated to be 0.015 s–1 for tyrosine and 0.0038 s–1 for phenylalanine. It can be concluded by comparison of kcat/Km ratios for the two substrates that the catalytic efficiency for the conversion of tyrosine is approximately 750 times higher than for the substrate phenylalanine. Therefore, tyrosine, and not phenylalanine, should be regarded the primary substrate of Sam8, and the enzyme shows an even more pronounced preference for tyrosine than the TAL enzyme from R. capsulatus. The rate of product formation of Sam8 decreases at pH values lower than 8.5 but remains at a relatively high level at higher pH values (Fig. 3B). Since the amino acid sequence of Sam8 shows significant similarity to the aminomutase enzyme SgcC4 from S. globisporus (5, 6), the enzyme was also assayed for its ability to convert L--tyrosine into -tyrosine. However, no such activity was detected.
DISCUSSION
In several bacteria and animals, HAL (EC 4.3.1.3) catalyzes the first step in the degradation of histidine to glutamate, which is the conversion of L-histidine to trans-urocanic acid. Whereas HALs are fairly common in bacteria, PALs were long thought to be absent from bacteria (29, 30, 31). The first bacterial TAL was isolated from R. capsulatus (17); it was shown to be 150 times more catalytically efficient toward L-tyrosine than L-phenylalanine as the substrate.
It was initially considered likely that a PAL is involved in the first step of the proposed biosynthetic pathway for N-(m,p-dihydroxycinnamoyl)taurine (Fig. 4), a caffeic acid derivative. Only one example of a prokaryotic PAL, named EncP, had been known when the experiments reported here were started (30). The higher similarity of Sam8 in size and sequence to prokaryotic HALs in comparison to eukaryotic PALs was not surprising, since the same differences had been observed for EncP.
Based on amino acid sequence alignment of HAL and PAL from P. crispum, PAL mutagenesis was performed on amino acid residues that were identical or similar to the active-site residues in HAL to gain insight into the importance of these residues for substrate binding or catalysis (21). Most of the active-site residues in HAL have analogous residues in the PAL protein and Sam8. However, the alignment depicted in Fig. 5 reveals that distinct differences exist in some regions. The highly conserved E414 in HAL is stringently replaced by Q500 in all PAL enzymes known to date, giving evidence that Sam8 is most likely not a HAL (Fig. 5c). Furthermore, H38 is replaced by the hydrophobic L138 in Photorhabdus luminescens PAL, P. crispum PAL, and R. capsulatus TAL, as well as in Sam8 (Fig. 5a); it has been pointed out that this may help to better accommodate the aromatic ring of the substrate phenylalanine in the binding pocket of PAL (20). The first three-dimensional structure of PAL, from the fungus Rhodosporidium toruloides, was solved recently (4). The positively charged side chain of K468 in R. toruloides PAL, which is also found in P. luminescens PAL and P. crispum PAL, is proposed to play an important role in recognizing the carboxyl group of the substrate. It is supposed that M382 in HAL (as well as in R. capsulatus TAL and S. maritimus PAL) has the same substrate-directing function (4). Interestingly, Sam8 has an alanine residue in this position (Fig. 5b), and the same substitution is found in the S. globisporus TAM protein. This tyrosine aminomutase catalyzes the conversion of L-tyrosine to (S)--tyrosine, which proceeds via a trans-p-coumarate intermediate; it also displays significant TAL activity (5, 6). While recombinant Sam8 did not exhibit any tyrosine aminomutase activity in in vitro assays, it may be reasoned that the A468 mutation accounts for its rather low kcat value, being "only" 10 times higher than the TAL side activity of the S. globisporus TAM enzyme and incredibly below the turnover number of the R. capsulatus TAL protein (Table 2).
To prove our proposed biosynthetic pathway for the aglycon assembly of the saccharomicins, we decided to heterologously express sam8 in an appropriate host, S. fradiae XKS. The production of p-coumarate (Fig. 4) in the host and the results of the biochemical characterization show unambiguously that the 510-amino-acid protein Sam8 acts as a TAL under physiological conditions. This is the first example of a TAL in actinomycetes.
Since the starting point of the reaction catalyzed by Sam8 is L-tyrosine and not L-phenylalanine, 4-hydroxylation, the next step in plant phenylpropanoid biosynthesis, was thought to be superfluous. The second step in saccharomicin aglycon biosynthesis was expected rather to be the conversion of p-coumarate to caffeate, catalyzed by a Coum3H (Fig. 4). By our second expression experiment, we were able to show that Sam5 indeed catalyzes the conversion of the TAL reaction's product, trans-p-coumaric acid, to trans-caffeic acid, though assigned to be, as presumed, a Coum3H. To our knowledge, this is the first report of such a hydroxylase in bacteria.
Discussions about the order in which ring substitutions (hydroxylations and methylations) of trans-p-coumarate can occur during flavonoid biosynthesis in plants are still being conducted. Two different hypotheses have been proposed (8): (i) coumaric acid becomes first substituted (e.g., 3-hydroxylation, resulting in caffeic acid), before further transformation into metabolically active esters occurs, or (ii) vice versa, coumaric acid is first transformed in an active ester and this ester is subsequently substituted. Reflecting these two hypotheses, several types of enzymes have been suggested for the 3-hydroxylation resulting in trans-caffeic acid derivatives (2, 13, 14, 26, 27). Two reports describe proteins belonging to the cytochrome P450-dependent enzymes that can catalyze the 3-hydroxylation of shikimate and quinate esters of trans-p-coumarate to their caffeoyl esters (11, 16). More recent works from three different groups include the characterization of CYP98A3 from Arabidopsis thaliana, a 3-monooxygenase of the phenylpropanoid metabolism that catalyzes also the 3-hydroxylation of p-coumaroyl shikimate and p-coumaroyl quinate (10, 19, 23). Based on these results, the conclusion was drawn that trans-p-coumaric acid cannot be directly converted to trans-caffeic acid (12).
We cannot exclude the possibility that p-coumarate is first activated by a hypothetical enzyme in the host S. fradiae XKS before the 3-hydroxylation of its ring occurs. However, since the heterologous expression of sam5 in the presence of coumaric acid also led to the production of trans-caffeic acid, our assumption that trans-p-coumaric acid is directly transferred to trans-caffeic acid is further substantiated.
In most cases, the phenylpropanoid pathway in plants leads from 4-coumarate to 4-coumaroyl-CoA, catalyzed by a 4-coumarate-CoA ligase. Other substituted cinnamates, such as caffeate (Fig. 4) and ferulate (3-methoxy-4-hydroxycinnamate), can also be transferred into the corresponding CoA thiol esters. As in the pathway in plants, Sam7 is thought to be a caffeoyl-CoA ligase, catalyzing the ligation of caffeate to CoA, leading to the corresponding thioester caffeoyl-CoA.
ACKNOWLEDGMENTS
We thank Gisela Grabellus for excellent technical assistance, V. Brecht for accomplishing the NMR analysis, and W. Heller (GSF Forschungszentrum fur Umwelt und Gesundheit, Neuherberg, Germany) for helpful discussion.
REFERENCES
Appert, C., E. Logemann, K. Hahlbrock, J. Schmid, and N. Amrhein. 1994. Structural and catalytic properties of the four phenylalanine ammonia-lyase isoenzymes from parsley. Eur. J. Biochem. 225:491-499.
Boniwell, J. M., and V. S. Butt. 1986. Flavin nucleotide-dependent 3-hydroxylation of 4-hydroxyphenylpropanoid carboxylic acids by particulate preparations from potato-tubers. Z. Naturforsch. C 41:56-60.
Buck, R. H., and K. Krummen. 1987. High-performance liquid-chromatographic determination of enantiomeric amino acids and amino alcohols after derivatization with o-phthaldialdehyde and various chiral mercaptans. J. Chromatogr. 387:255-265.
Calabrese, J. C., D. B. Jordan, A. Boodhoo, S. Sariaslani, and T. Vannelli. 2004. Crystal structure of phenylalanine ammonia lyase: multiple helix dipoles implicated in catalysis. Biochemistry 43:11403-11416.
Christenson, S. D., W. Liu, M. D. Toney, and B. Shen. 2003. A novel 4-methylideneimidazole-5-one-containing tyrosine aminomutase in enediyne antitumor antibiotic C-1027 biosynthesis. J. Am. Chem. Soc. 125:6062-6063.
Christenson, S. D., W. M. Wu, M. A. Spies, B. Shen, and M. D. Toney. 2003. Kinetic analysis of the 4-methylideneimidazole-5-one-containing tyrosine aminomutase in enediyne antitumor antibiotic C-1027 biosynthesis. Biochemistry 42:12708-12718.
Doumith, M., P. Weingarten, U. F. Wehmeier, K. Salah-Bey, B. Benhamou, C. Capdevila, J. M. Michel, W. Piepersberg, and M. C. Raynal. 2000. Analysis of genes involved in 6-deoxyhexose biosynthesis and transfer in Saccharopolyspora erythraea. Mol. Gen. Genet. 264:477-485.
Forkmann, G., and W. Heller. 1999. Biosynthesis of flavonoids, p. 713-748. In Sir Derek Barton and Koji Nakanishi (ed.), Comprehensive natural products chemistry. Elsevier Science Ltd., Oxford, United Kingdom.
Franke, R., M. R. Hemm, J. W. Denault, M. O. Ruegger, J. M. Humphreys, and C. Chapple. 2002. Changes in secondary metabolism and deposition of an unusual lignin in the ref8 mutant of Arabidopsis. Plant J. 30:47-59.
Franke, R., J. M. Humphreys, M. R. Hemm, J. W. Denault, M. O. Ruegger, J. C. Cusumano, and C. Chapple. 2002. The Arabidopsis REF8 gene encodes the 3-hydroxylase of phenylpropanoid metabolism. Plant J. 30:33-45.
Heller, W., and T. Kuhnl. 1985. Elicitor induction of a microsomal 5-O-(4-coumaroyl)shikimate 3'-hydroxylase in parsley cell suspension cultures. Arch. Biochem. Biophys. 241:453-460.
Humphreys, J. M., and C. Chapple. 2002. Rewriting the lignin roadmap. Curr. Opin. Plant Biol. 5:224-229.
Kneusel, R. E., U. Matern, and K. Nicolay. 1989. Formation of trans-caffeoyl-CoA from trans-4-coumaroyl-CoA by Zn2+-dependent enzymes in cultured plant cells and its activation by an elicitor-induced pH shift. Arch. Biochem. Biophys. 269:455-462.
Kojima, M., and W. Takeuchi. 1989. Detection and characterization of p-coumaric acid hydroxylase in mung bean, Vigna mungo, seedlings. J. Biochem. 105:265-270.
Kong, F. M., N. Zhao, M. M. Siegel, K. Janota, J. S. Ashcroft, F. E. Koehn, D. B. Borders, and G. T. Carter. 1998. Saccharomicins, novel heptadecaglycoside antibiotics effective against multidrug-resistant bacteria. J. Am. Chem. Soc. 120:13301-13311.
Kuhnl, T., U. Koch, W. Heller, and E. Wellmann. 1987. Chlorogenic acid biosynthesis—characterization of a light-induced microsomal 5-O-(4-coumaroyl)-D-quinate/shikimate 3'-hydroxylase from carrot (Daucus carota L.) cell suspension cultures. Arch. Biochem. Biophys. 258:226-232.
Kyndt, J. A., T. E. Meyer, M. A. Cusanovich, and J. J. Van Beeumen. 2002. Characterization of a bacterial tyrosine ammonia lyase, a biosynthetic enzyme for the photoactive yellow protein. FEBS Lett. 512:240-244.
Labeda, D. P., and R. M. Kroppenstedt. 2000. Phylogenetic analysis of Saccharothrix and related taxa: proposal for Actinosynnemataceae fam. nov. Int. J. Syst. Evol. Microbiol. 50:331-336.
Nair, R. B., Q. Xia, C. J. Kartha, E. Kurylo, R. N. Hirji, R. Datla, and G. Selvaraj. 2002. Arabidopsis CYP98A3 mediating aromatic 3-hydroxylation. Developmental regulation of the gene, and expression in yeast. Plant Physiol. 130:210-220.
Poppe, L., and J. Retey. 2005. Friedel-Crafts-type mechanism for the enzymatic elimination of ammonia from histidine and phenylalanine. Angew. Chem. Int. Ed. Engl. 44:3668-3688.
Rther, D., L. Poppe, G. Morlock, S. Viergutz, and J. Retey. 2002. An active site homology model of phenylalanine ammonia-lyase from Petroselinum crispum. Eur. J. Biochem. 269:3065-3075.
Rther, R., L. Poppe, S. Viergutz, B. Langer, and J. Retey. 2001. Characterization of the active site of histidine ammonia-lyase from Pseudomonas putida. Eur. J. Biochem. 268:6011-6019.
Schoch, G., S. Goepfert, M. Morant, A. Hehn, D. Meyer, P. Ullmann, and D. Werck-Reichhart. 2001. CYP98A3 from Arabidopsis thaliana is a 3'-hydroxylase of phenolic esters, a missing link in the phenylpropanoid pathway. J. Biol. Chem. 276:36566-36574.
Schwede, T. F., J. Retey, and G. E. Schulz. 1999. Crystal structure of histidine ammonia-lyase revealing a novel polypeptide modification as the catalytic electrophile. Biochemistry 38:5355-5361.
Singh, M. P., P. J. Petersen, W. J. Weiss, F. Kong, and M. Greenstein. 2000. Saccharomicins, novel heptadecaglycoside antibiotics produced by Saccharothrix espanaensis: antibacterial and mechanistic activities. Antimicrob. Agents Chemother. 44:2154-2159.
Stafford, H. A., and S. Dresler. 1972. 4-Hydroxycinnamic acid hydroxylase and polyphenolase—activities in Sorghum vulgare. Plant Physiol. 49:590-595.
Tanaka, M., and M. Kojima. 1991. Purification and characterization of p-coumaroyl-D-glucose hydroxylase of sweet potato (Ipomoea batatas) roots. Arch. Biochem. Biophys. 284:151-157.
von Mulert, U., A. Luzhetskyy, C. Hofmann, A. Mayer, and A. Bechthold. 2004. Expression of the landomycin biosynthetic gene cluster in a PKS mutant of Streptomyces fradiae is dependent on the coexpression of a putative transcriptional activator gene. FEMS Microbiol. Lett. 230:91-97.
Williams, J. S., M. Thomas, and D. J. Clarke. 2005. The gene stlA encodes a phenylalanine ammonia-lyase that is involved in the production of a stilbene antibiotic in Photorhabdus luminescens TT01. Microbiology 151:2543-2550.
Xiang, L. K., and B. S. Moore. 2002. Inactivation, complementation, and heterologous expression of encP, a novel bacterial phenylalanine ammonia-lyase gene. J. Biol. Chem. 277:32505-32509.
Xiang, L. K., and B. S. Moore. 2005. Biochemical characterization of a prokaryotic phenylalanine ammonia lyase. J. Bacteriol. 187:4286-4289.(Martin Berner, Daniel Kru)
ABSTRACT
The saccharomicins A and B, produced by the actinomycete Saccharothrix espanaensis, are oligosaccharide antibiotics. They consist of 17 monosaccharide units and the unique aglycon N-(m,p-dihydroxycinnamoyl)taurine. To investigate candidate genes responsible for the formation of trans-m,p-dihydroxycinnamic acid (caffeic acid) as part of the saccharomicin aglycon, gene expression experiments were carried out in Streptomyces fradiae XKS. It is shown that the biosynthetic pathway for trans-caffeic acid proceeds from L-tyrosine via trans-p-coumaric acid directly to trans-caffeic acid, since heterologous expression of sam8, encoding a tyrosine ammonia-lyase, led to the production of trans-p-hydroxycinnamic acid (coumaric acid), and coexpression of sam8 and sam5, the latter encoding a 4-coumarate 3-hydroxylase, led to the production of trans-m,p-dihydroxycinnamic acid. This is not in accordance with the general phenylpropanoid pathway in plants, where trans-p-coumaric acid is first activated before the 3-hydroxylation of its ring takes place.
INTRODUCTION
Saccharothrix is a genus of gram-positive bacteria belonging to the well-known order Actinomycetales. Most agents used at present for the treatment of bacterial infections were discovered in members of the Actinomycetales. Saccharothrix espanaensis produces the two heptadecaglycoside antibiotics saccharomicins A and B, which represent a new class of antibiotics (15, 18). They exhibit potent antibacterial activity both in vitro and in vivo against multiply-resistant strains of Staphylococcus aureus as well as vancomycin-resistant enterococci (25). The saccharomicins consist of an oligosaccharide portion and the intriguing aglycon N-(m,p-dihydroxycinnamoyl)taurine (Fig. 1), in which caffeic acid is linked to the amino sulfonic acid taurine via an amide bond.
Enzymes belonging to the group of ammonia-lyases catalyze the conversion of -amino acids into ,-unsaturated acids by elimination of ammonia. Ubiquitous in plants and fungi, phenylalanine ammonia-lyase (PAL) (EC 4.3.1.5) catalyzes the nonoxidative deamination of the primary amino acid L-phenylalanine to trans-cinnamic acid (trans-cinnamate), which is the first reaction of the so-called general phenylpropanoid pathway in plants (8). Phenylpropanoids include several important natural product classes, for example, flavonoids, lignins, and coumarins. In monocotyledons, PAL utilizes L-tyrosine in addition to L-phenylalanine (resulting in trans-p-coumaric acid), whereas the enzyme from dicotyledons converts only L-phenylalanine sufficiently. Both PAL and tyrosine ammonia-lyase (TAL) activity are very rare in bacteria (17, 30).
The next reactions of the three-step general phenylpropanoid pathway are catalyzed by the enzymes trans-cinnamate 4-monooxygenase (also called cinnamate 4-hydroxylase; EC 1.14.13.11), leading to trans-p-coumaric acid (trans-4-coumarate), and 4-coumarate-coenzyme A (CoA) ligase (EC 6.2.1.12), leading to 4-coumaroyl-CoA (8).
In this report, we describe the cloning and identification of two genes from S. espanaensis which are involved in caffeic acid biosynthesis. Heterologous expression of sam8, encoding a TAL, led to the production of trans-p-coumaric acid. Coexpression of sam8 and sam5, encoding a 4-coumarate 3-hydroxylase (Coum3H), led to the production of trans-caffeic acid. An enzyme assay performed with recombinant Sam8 verified unambiguously that L-tyrosine (and not L-phenylalanine) is its natural substrate. This is the first report of a Coum3H and the first report of a TAL in the Actinomycetales.
MATERIALS AND METHODS
Strains, vectors, culture conditions, and DNA manipulation. S. espanaensis was obtained from DSMZ (Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH, Braunschweig, Germany; DSM 44229). Escherichia coli XL1-Blue MRF (Stratagene) and the E. coli cloning vector pBluescript SK(–) (Stratagene) were used for subcloning. pUWL201 was a gift from U. Wehmeier and W. Piepersberg (Department of Chemical Microbiology, University of Wuppertal, Germany) (7). The construction of the PKSII mutant strain S. fradiae XKS has been described previously (28).
Protoplast formation, transformation, and regeneration of protoplasts from S. fradiae XKS were carried out by standard procedures. Isolation of plasmid DNA from E. coli, restriction, and ligation were performed according to the protocols of the manufacturers of the kits and enzymes (Promega, QIAGEN).
Construction of a genomic library of S. espanaensis and screening. An S. espanaensis genomic cosmid library was constructed in E. coli DH5 using pOJ436 as cosmid vector. For DNA extraction, mycelium was embedded in agarose and then the DNA was partially digested and isolated, yielding fragments with an average size greater than 35 kb. Robotically produced high-density colony arrays (Hybond N+; Amersham Biosciences) were utilized for the screening of 2,304 cosmid clones with a strain-specific sugar biosynthesis probe (2,3-dehydratase) according to standard nonradioactive hybridization procedures. The primers for the amplification of a gene-specific 2,3-dehydratase probe were 5'-CAGGCSACSWSSAACTACAC-3' and 5'-SWRGAASCGSCCSCCCTCCTC-3'.
DNA sequencing and computer-assisted sequence analysis. Nucleotide sequences were determined at 4base lab GmbH (Reutlingen, Germany) and at SEQLBAB (Sequence Laboratories Gttingen GmbH, Gttingen, Germany) by using either standard primers (M13 universal and reverse, T3, and T7) or customized, internal primers. Computer-assisted sequence analysis was done with DNASIS software (version 2.1; Hitachi Software Engineering). Artemis software, created by the Wellcome Trust Sanger Institute (Cambridge, England) and available at the website http://www.sanger.ac.uk/Software/Artemis/, was used to identify potential protein-coding regions. Database comparison was performed with the BLAST search tools on the server of the National Center for Biotechnology Information, National Library of Medicine, NIH (http://www.ncbi.nlm.nih.gov/).
Production of trans-p-coumaric acid in S. fradiae XKS. Gene sam8 was amplified with Pfu polymerase (Promega) from cosmid 1B11 using primers 5'-CCGCTGGGAATTCTGCGGCACGG-3' (EcoRI restriction site underlined) and 5'-GCCCTCGTTCTAGACCCGTGCGG-3' (XbaI restriction site underlined), resulting in a 1,639-bp product. The fragment was cloned into the EcoRI-XbaI cloning site of the vector pBluescript SK(–) to yield pMB1, whose sequence was confirmed by sequencing. The pMB1 insert was then transferred into the EcoRI-XbaI cloning site of the expression vector pUWL201, which contains the erythromycin resistance promoter ermE to create pMB3. The nonmethylating E. coli strain ET12567 was used to obtain unmethylated DNA. The unmethylated plasmid pMB3 and vector pUWL201 (negative control) were then introduced into S. fradiae XKS by protoplast transformation.
The strain was grown on R2YE agar plates at 28°C for 16 h. (All of the following incubations were carried out at 28°C.) The plates were then overlaid with NB soft agar containing thiostrepton (concentration on the plate, 50 μg ml–1) and incubated for another 4 days. For the selection of transformants, single colonies were transferred onto thiostrepton-containing HA agar plates (30 μg ml–1). After 3 days of incubation, several flasks of HA liquid medium (100 ml) with a thiostrepton concentration of 15 μg ml–1 were inoculated with colonies from these plates and put on a rotary shaker (180 rpm). Culture samples (each 1 ml) were taken over a period of 5 days (24, 48, 72, 96, and 120 h after inoculation). The samples were extracted with ethyl acetate (each 1 ml), vacuum dried, and dissolved in acetonitrile/water (30/70 [vol/vol]; 250 μl).
HPLC/ESI-MS analysis of fermented cultures. High-performance liquid chromatography (HPLC)/electrospray ionization mass spectrometry (ESI-MS) was performed on an Agilent 1100 Series LC/MSD system equipped with an Agilent Zorbax SB-C18 column (5-μm particle size, 4.6 by 150 mm) maintained at 23°C and an Agilent Zorbax SB-C18 precolumn (5-μm particle size, 4.5 by 12.5 mm). Each extract was analyzed at a flow rate of 0.5 ml min–1 with solvent A (99.5% H2O, 0.5% acetic acid) and solvent B (acetonitrile), using the following gradient: 0 to 3 min, 10% B; 3 to 6 min, linear to 20% B; 6 to 9 min, linear to 30% B; 9 to 12 min, linear to 50% B; 12 to 18 min, linear to 95% B. The UV detection wavelength was 310 nm, and the MS scan range was 100 to 300 m/z.
Production of trans-caffeic acid in S. fradiae XKS. Gene sam5 was amplified with Pfu polymerase (Promega) from cosmid 1B11 using primers 5'-CCGCGTTCTAGACCAAGCTTCACCTCAGC-3' (HindIII restriction site underlined) and 5'-GCGCGGGAATTCATCGGGTGC-3' (EcoRI restriction site underlined), resulting in a 1,662-bp product. The fragment was cloned into the HindIII-EcoRI cloning site of the vector pBluescript SK(–) to yield pMB11, whose sequence was confirmed by sequencing. The pMB11 insert was then transferred into the HindIII-EcoRI site of pMB3 to create pMB9. For the construction of pMB10, the pMB11 insert was put into the HindIII-EcoRI site of pUWL201. All of the following procedures were carried out as described for pMB3 (see above). In the case of the expression experiment with pMB10, the flasks of HA liquid medium (100 ml) were supplemented with coumaric acid: p-coumaric acid was added at concentrations of 25, 18.75, 12.5, and 6.25 μg ml–1.
Isolation of trans-caffeic acid from fermented cultures. As a preculture, 100 ml of HA liquid medium with a thiostrepton concentration of 15 μg ml–1 was inoculated with one of the transformants obtained after heterologous expression carried out with pMB9 (see above). After 3 days of incubation, 10 Erlenmeyer flasks each containing 100 ml of HA liquid medium (supplemented with thiostrepton, 15 μg ml–1) were inoculated with 250 μl of the preculture and put on a rotary shaker (180 rpm). After 3 days of incubation, the 10 cultures were merged and then harvested by centrifugation. After adjusting the pH to 3.2, the supernatant (approximately 900 ml) was extracted with 900 ml of ethyl acetate, using a separating funnel. The aqueous phase was extracted again with 900 ml of ethyl acetate. The merged organic phase (1,800 ml) was washed twice with 180 ml of distilled H2O and subsequently with 180 ml of a saturated sodium chloride solution. A couple of spoonfuls of anhydrous sodium sulfate were added to remove further water, before the ethyl acetate was vacuum dried. The residue (dried crude extract) was then dissolved in 3 ml of solvent (dichloromethane-ethyl acetate-acetic acid, 50:50:0.5 [vol/vol/vol]).
A glass column (diameter, 30 mm), filled with a suspension of 7.5 g of silica gel 60 (Merck, Darmstadt, Germany) in the solvent mentioned above, was loaded with the sample. All 10 collected eluate fractions (the volume was approximately 70 ml each) were vacuum dried and then dissolved in 2 ml of solvent. To determine the fraction containing trans-caffeic acid, 20 μl of each fraction was put on a silica gel 60 F254 thin-layer chromatography plate (Merck). Commercially available trans-caffeic acid was used as a reference.
Preparative HPLC. Preparative HPLC was performed on a system consisting of a Waters Delta 600 pump unit, a Waters 600 controller, a Waters 2487 Dual absorbance detector, a Waters Fraction Collector III, and a Waters In-Line Degasser AF. The system was equipped with a Waters XTerra Prep MS C18 column (5-μm particle size, 7.8 by 150 mm) and an XTerra VP-1 precolumn. All injections were carried out at room temperature. The injected extract (20 μl per injection) was analyzed at a flow rate of 2.0 ml min–1 with solvent A (99.5% H2O, 0.5% acetic acid) and solvent B (99.5% acetonitrile, 0.5% acetic acid), using the following gradient: 0 to 3 min, 10% B; 3 to 9 min, linear to 30% B; 9 to 12 min, linear to 50% B; 12 to 18 min, linear to 95% B. The UV detection wavelength was 310 nm. The fraction collector was set at 15 s of collection time per fraction. The fraction containing trans-caffeic acid was determined by its UV spectrum, with an Amersham Biosciences Ultrospec 2100 pro photometer.
NMR characterization of trans-caffeic acid. All 25 fractions containing trans-caffeic acid were merged. After vacuum drying and subsequent lyophilization, the dried extract was dissolved in methanol-d4 (CD3OD). The nuclear magnetic resonance (NMR) data for both the sample and commercially available trans-caffeic acid were recorded by Varian Unity-300 equipment at 300 MHz. The sample's 1H-NMR data are summarized as follows: 1H-NMR (300 MHz, CD3OD) : 6.24 (1H, d, J = 15 Hz; C2-H); 6.80 (1H, d, J = 8.5 Hz; C5'-H); 6.95 (1H, dd, J = 8.5 Hz; C6'-H); 7.05 (1H, d, J = 2.1 Hz; C2'-H); 7.54 (1H, d, J = 15 Hz; C3-H).
Production of recombinant TAL. The coding sequence of Sam8 was PCR amplified with cloned Pfu polymerase (Stratagene) from plasmid pMB1 using primers TALsx_startBgl (5'-GACGCAGGAGATCTGTGGAACGTCAGGC-3', BglII restriction site underlined) and TALsx_stopEco (5'-TCATCCGAGAATTCTCCTTCCCGTCTG-3', EcoRI restriction site underlined) and ligated into the pGEX6p1 vector (Amersham Biosciences) via the BamHI and EcoRI restriction sites to yield plasmid pGEX6p1-TALsx. The expression plasmid was sequenced to verify the correctness of the PCR amplification of the 1,523-bp insert and transferred into E. coli BL21(DE3) cells. Production cultures of 0.6-liter volume were inoculated 1:100 with an overnight culture and grown to an optical density at 600 nm of 0.7 at 30°C (200 rpm) in a 3-liter Erlenmeyer flask using 2YT medium that contained 100 μg ml–1 ampicillin. After induction of protein expression with 0.1 mM isopropyl--D-thiogalactopyranoside, cells were incubated at 30°C (200 rpm) for another 2 h. The cells were then harvested by centrifugation at 4°C (5,000 rpm), resuspended in 12 ml of phosphate-buffered saline buffer (140 mM NaCl-2.7 mM KCl-10 mM Na2HPO4-1.8 mM KH2PO4, pH 7.3), and passed twice through a French pressure cell. Cell debris was sedimented by centrifugation, and the supernatant was loaded onto a glutathione Sepharose 4B gel matrix. Washing and elution of fusion protein with reduced glutathione or, alternatively, on-column cleavage with PreScission protease was carried out according to the manufacturer's protocol (GST Gene Fusion System Handbook; Amersham Biosciences).
TAL enzyme activity assay. Enzyme activity assays contained 20 μg of purified protein and were performed in 1-ml volumes using 0.1 M Tris buffer (pH 8.8) at 30°C and a substrate concentration of 2 mM (Tyr) or 10 mM (Phe, His). Identification of products was carried out via reverse-phase HPLC separation (solvent delivery and DAD detection system from Dionex). For this purpose, 15 μl of acetic acid was added to each assay mixture, followed by extraction with 1 ml of ethyl acetate. After removal of the solvent in vaccuo, samples were dissolved in 100 μl of methanol and loaded onto a Nucleodur C18 column (Macherey-Nagel; 5-μm particle size, 125 by 2 mm). The solvent system consisted of H2O (A) and acetonitrile (B), each containing 0.1% formic acid. Using a gradient of 15 min running from 1% to 99% solvent B, separation of products was observed at the following retention times (detection wavelength given in parentheses): coumaric acid, 7.8 min (310 nm); cinnamic acid, 10.0 min (295 nm). Identification of products was also confirmed by LC-MS methods (Bruker HCTplus coupled to an Agilent 1100 series HPLC system; Nucleodur C18 column [3-μm particle size, 125 by 2 mm]; same HPLC conditions as above; detection in positive ionization mode).
For the determination of kinetic parameters, 2 μg of purified protein was incubated for 30 min at 30°C in 600 μl of Tris buffer (0.1 M, pH 8.8) containing 0.01 to 0.04 mM L-tyrosine or 0.5 to 10 mM L-phenylalanine (in the latter case, 4 μg of enzyme was used, with a 2-h incubation time). Reaction progress was monitored photometrically at 310 nm (tyrosine as substrate) or 290 nm (phenylalanine as substrate) with an AnalytikJena Specord 50 UV/Vis photometer. The pH dependency of initial rates of formation was assayed using 0.1 M Tris buffer at pH 7.0, 8.0, and 8.8 and 50 mM CHES buffer at pH 8.8, 9.2, and 10. For the calculation of product concentrations and conversion into initial velocities, a calibration curve was generated by using serial dilutions of commercially available trans-cinnamic acid and trans-p-coumaric acid in 0.1 M Tris buffer. Nonlinear regression and curve fitting was carried out with the SigmaPlot enzyme kinetics module (Systat Software Inc.). All deduced parameters are based on measurements in three replicates.
To investigate the formation of -tyrosine, a 50-μl aliquot from an assay mixture containing 20 μg of purified Sam8 was subjected to derivatization with OPA-NAC reagent (3). The reagent was prepared by addition of 15 mg of o-phthaldialdehyde and 15 mg of N-acetyl-L-cysteine (Fluka) to 0.5 ml of ethanol and subsequent dilution with 11 ml of 0.4 M borate buffer, pH 10.2. The sample aliquot was mixed with 200 μl of reagent for 10 min at room temperature prior to HPLC analysis. The mobile phase consisted of 50 mM sodium phosphate buffer, pH 6.5 (solvent A), and a mixture of phosphate buffer and methanol (35:65) containing 5% tetrahydrofuran (solvent B). Using a linear gradient from 35% B to 40% B in 24 min on a Nucleodur C18 column (5-μm particle size, 125 by 2 mm), separation of the diastereomeric OPA-NAC derivatives of commercially available (R/S)--tyrosine could be achieved. The obtained retention times were as follows: L--tyrosine, 16.5 min; (R)--tyrosine, 12 min; (S)--tyrosine, 15 min. UV/visual detection was at 330 nm.
GenBank accession number. The GenBank accession number of the DNA sequence reported in this paper is DQ357071.
RESULTS
Cloning and sequencing of a locus containing caffeic acid biosynthetic genes. In order to identify possible genes involved in caffeic acid biosynthesis, an S. espanaensis genomic cosmid library was generated in E. coli. Since caffeic acid is a saccharomicin substructure, it was considered likely that genes for its biosynthesis adjoin other genes of the saccharomicin biosynthetic gene cluster. As saccharomicin contains several 2,6-dideoxyhexoses, a dNDP-glucose 2,3-dehydratase gene was expected to be involved in its biosynthesis. Consequently, the cosmid library was screened with a 2,3-dehydratase gene probe, resulting in 34 positively hybridizing cosmids.
Restriction analysis of the isolated cosmids showed that most cosmids contained DNA overlapping with DNA from other cosmids. By comparing the BamHI fragment pattern, as a result, 22 cosmids could be combined into three subgroups; a cosmid map was created for each of these subgroups. Four cosmid inserts, at least one insert from each subgroup, were subcloned using the restriction endonuclease BamHI. In the case of cosmid 2L12, random sequencing of subclones revealed significant sequence similarity to deoxysugar biosynthetic genes, glycosyltransferase genes, and bacterial ammonia-lyase genes.
Restriction mapping also indicated that cosmid 2L12 did not contain the entire expected ammonia-lyase gene. To detect cosmids possibly containing the whole gene sequence, further screening of the cosmid library was performed, using a DNA fragment as a gene probe consisting of DNA subcloned from the respective end of cosmid 2L12. This resulted in the isolation of the overlapping cosmid 1B11. Finally, a 14-kb DNA region located on both cosmids was sequenced.
Sequence analysis. Analysis of the final contiguous 14-kb DNA sequence revealed the presence of eight open reading frames, named sam1 to sam8. Table 1 lists the putative catalytic functions of the proteins deduced from sam8, sam5, and sam7, which are thought to be involved in the biosynthesis of the saccharomicin aglycon. The deduced protein of sam8 is similar to bacterial histidine ammonia-lyases (HALs) and as well to plant PALs. Sam5 strongly resembles phenol hydroxylases from various bacteria. The product of sam7 shows similarity to acyl-CoA synthetases. Sam6 resembles bacterial dNDP-glucose synthases. The deduced proteins of sam1 to sam4 show similarity to proteins of unknown functions.
Amino acid sequence alignments of several members of the ammonia-lyase enzyme group revealed that the protein encoded by sam8 was more similar to bacterial HALs in size and sequence than to eukaryotic PALs. A conserved Ala-Ser-Gly segment which is typical for ammonia-lyases can also be found in sam8 (4, 24). Furthermore, several active-site amino acid residues proposed for HAL after elucidation of the X-ray structure of HutH, a HAL from Pseudomonas putida, have analogous residues in Sam8 (22, 24).
Heterologous expression experiments. The pUWL201-based plasmids pMB3, containing sam8, and pMB9, containing sam8 and sam5, were heterologously expressed in S. fradiae XKS (28), a PKSII mutant strain from the urdamycin A producer S. fradiae Tu2717, which has shown to be a very convenient host for expression experiments. Organic extracts of the transformants were analyzed for the production of trans-cinnamic acid, trans-p-coumaric acid, and trans-caffeic acid, respectively, by HPLC/ESI-MS analysis using the corresponding commercially available reference standards. The expressions resulted in the production of trans-p-coumaric acid (Mr = 164.16) and trans-caffeic acid (Mr = 180.16) (Fig. 2). Neither of these compounds was detectable in S. fradiae XKS transformed with the expression vector pUWL201. Cinnamic acid, the product of the PAL reaction, could not be detected in any extract of the transformants. Caffeic acid was also produced after heterologous expression of pMB10 (pUWL201-based plasmid containing sam5) and addition of coumaric acid to the transformant cultures (data not shown).
These results verified that the gene product of sam8 is a TAL and the gene product of sam5 is a Coum3H.
Isolation of trans-caffeic acid from fermented cultures and NMR characterization. The organic extract of one of the transformants was purified by silica gel column chromatography. The eluate fraction containing caffeic acid, as proven by TLC, was then used to isolate trans-caffeic acid by preparative HPLC. As expected, 1H-NMR spectroscopy proved that the sample contained trans-caffeic acid. Apart from the trans-caffeic acid (signals were assigned by comparison with the signals of a solution of commercially available trans-caffeic acid), supposable cis isomers and truxillic acid derivatives likely being generated during the extraction process could also be detected. These known cis-trans isomerizations of trans-p-coumaric acid, trans-caffeic acid, and its derivatives under light exposure have been reported by several authors (9, 11). The same sample and the reference were subsequently characterized by H,H- correlation NMR spectroscopy analysis, whereby the presence of trans-caffeic acid could again be verified.
Production of recombinant TAL and enzyme assay. Sam8 was overproduced in E. coli as an N-terminal fusion to glutathione S-transferase (GST), linked by the recognition site for PreScission protease. After purification of the 80-kDa fusion protein and cleavage from the GST tag by protease treatment, the purified protein (calculated Mr, 55,084 Da) was obtained in a yield of approximately 2.5 mg liter–1 of culture (Fig. 3A).
Ammonia-lyase activity was assayed in analogy to published procedures (4, 5). The substrates tested were L-tyrosine, L-phenylalanine, and L-histidine at concentrations up to 10 mM. While no significant activity with histidine was observed, incubation with tyrosine or phenylalanine resulted in the production of coumaric acid and cinnamic acid, respectively. Identity of these products was verified by HPLC analysis in comparison with authentic reference standards.
The determination of kinetic parameters for both enzymatic activities of Sam8 was carried out by a spectrophotometric assay. Initial velocities of product formation were fitted to the Michaelis-Menten equation by the direct nonlinear regression method. The calculated Km values are 15.5 ± 0.3 μM for tyrosine and 2.86 ± 0.3 mM for phenylalanine. They resemble the published values for TAL from Rhodobacter capsulatus (17) and are also strikingly similar to the Km values for PAL from Petroselinum crispum, though with a reversed substrate preference (Table 2). The kcat values for Sam8 are somewhat lower than the turnover numbers for the other known enzymes listed in Table 2 that exhibit TAL or PAL activity; they are calculated to be 0.015 s–1 for tyrosine and 0.0038 s–1 for phenylalanine. It can be concluded by comparison of kcat/Km ratios for the two substrates that the catalytic efficiency for the conversion of tyrosine is approximately 750 times higher than for the substrate phenylalanine. Therefore, tyrosine, and not phenylalanine, should be regarded the primary substrate of Sam8, and the enzyme shows an even more pronounced preference for tyrosine than the TAL enzyme from R. capsulatus. The rate of product formation of Sam8 decreases at pH values lower than 8.5 but remains at a relatively high level at higher pH values (Fig. 3B). Since the amino acid sequence of Sam8 shows significant similarity to the aminomutase enzyme SgcC4 from S. globisporus (5, 6), the enzyme was also assayed for its ability to convert L--tyrosine into -tyrosine. However, no such activity was detected.
DISCUSSION
In several bacteria and animals, HAL (EC 4.3.1.3) catalyzes the first step in the degradation of histidine to glutamate, which is the conversion of L-histidine to trans-urocanic acid. Whereas HALs are fairly common in bacteria, PALs were long thought to be absent from bacteria (29, 30, 31). The first bacterial TAL was isolated from R. capsulatus (17); it was shown to be 150 times more catalytically efficient toward L-tyrosine than L-phenylalanine as the substrate.
It was initially considered likely that a PAL is involved in the first step of the proposed biosynthetic pathway for N-(m,p-dihydroxycinnamoyl)taurine (Fig. 4), a caffeic acid derivative. Only one example of a prokaryotic PAL, named EncP, had been known when the experiments reported here were started (30). The higher similarity of Sam8 in size and sequence to prokaryotic HALs in comparison to eukaryotic PALs was not surprising, since the same differences had been observed for EncP.
Based on amino acid sequence alignment of HAL and PAL from P. crispum, PAL mutagenesis was performed on amino acid residues that were identical or similar to the active-site residues in HAL to gain insight into the importance of these residues for substrate binding or catalysis (21). Most of the active-site residues in HAL have analogous residues in the PAL protein and Sam8. However, the alignment depicted in Fig. 5 reveals that distinct differences exist in some regions. The highly conserved E414 in HAL is stringently replaced by Q500 in all PAL enzymes known to date, giving evidence that Sam8 is most likely not a HAL (Fig. 5c). Furthermore, H38 is replaced by the hydrophobic L138 in Photorhabdus luminescens PAL, P. crispum PAL, and R. capsulatus TAL, as well as in Sam8 (Fig. 5a); it has been pointed out that this may help to better accommodate the aromatic ring of the substrate phenylalanine in the binding pocket of PAL (20). The first three-dimensional structure of PAL, from the fungus Rhodosporidium toruloides, was solved recently (4). The positively charged side chain of K468 in R. toruloides PAL, which is also found in P. luminescens PAL and P. crispum PAL, is proposed to play an important role in recognizing the carboxyl group of the substrate. It is supposed that M382 in HAL (as well as in R. capsulatus TAL and S. maritimus PAL) has the same substrate-directing function (4). Interestingly, Sam8 has an alanine residue in this position (Fig. 5b), and the same substitution is found in the S. globisporus TAM protein. This tyrosine aminomutase catalyzes the conversion of L-tyrosine to (S)--tyrosine, which proceeds via a trans-p-coumarate intermediate; it also displays significant TAL activity (5, 6). While recombinant Sam8 did not exhibit any tyrosine aminomutase activity in in vitro assays, it may be reasoned that the A468 mutation accounts for its rather low kcat value, being "only" 10 times higher than the TAL side activity of the S. globisporus TAM enzyme and incredibly below the turnover number of the R. capsulatus TAL protein (Table 2).
To prove our proposed biosynthetic pathway for the aglycon assembly of the saccharomicins, we decided to heterologously express sam8 in an appropriate host, S. fradiae XKS. The production of p-coumarate (Fig. 4) in the host and the results of the biochemical characterization show unambiguously that the 510-amino-acid protein Sam8 acts as a TAL under physiological conditions. This is the first example of a TAL in actinomycetes.
Since the starting point of the reaction catalyzed by Sam8 is L-tyrosine and not L-phenylalanine, 4-hydroxylation, the next step in plant phenylpropanoid biosynthesis, was thought to be superfluous. The second step in saccharomicin aglycon biosynthesis was expected rather to be the conversion of p-coumarate to caffeate, catalyzed by a Coum3H (Fig. 4). By our second expression experiment, we were able to show that Sam5 indeed catalyzes the conversion of the TAL reaction's product, trans-p-coumaric acid, to trans-caffeic acid, though assigned to be, as presumed, a Coum3H. To our knowledge, this is the first report of such a hydroxylase in bacteria.
Discussions about the order in which ring substitutions (hydroxylations and methylations) of trans-p-coumarate can occur during flavonoid biosynthesis in plants are still being conducted. Two different hypotheses have been proposed (8): (i) coumaric acid becomes first substituted (e.g., 3-hydroxylation, resulting in caffeic acid), before further transformation into metabolically active esters occurs, or (ii) vice versa, coumaric acid is first transformed in an active ester and this ester is subsequently substituted. Reflecting these two hypotheses, several types of enzymes have been suggested for the 3-hydroxylation resulting in trans-caffeic acid derivatives (2, 13, 14, 26, 27). Two reports describe proteins belonging to the cytochrome P450-dependent enzymes that can catalyze the 3-hydroxylation of shikimate and quinate esters of trans-p-coumarate to their caffeoyl esters (11, 16). More recent works from three different groups include the characterization of CYP98A3 from Arabidopsis thaliana, a 3-monooxygenase of the phenylpropanoid metabolism that catalyzes also the 3-hydroxylation of p-coumaroyl shikimate and p-coumaroyl quinate (10, 19, 23). Based on these results, the conclusion was drawn that trans-p-coumaric acid cannot be directly converted to trans-caffeic acid (12).
We cannot exclude the possibility that p-coumarate is first activated by a hypothetical enzyme in the host S. fradiae XKS before the 3-hydroxylation of its ring occurs. However, since the heterologous expression of sam5 in the presence of coumaric acid also led to the production of trans-caffeic acid, our assumption that trans-p-coumaric acid is directly transferred to trans-caffeic acid is further substantiated.
In most cases, the phenylpropanoid pathway in plants leads from 4-coumarate to 4-coumaroyl-CoA, catalyzed by a 4-coumarate-CoA ligase. Other substituted cinnamates, such as caffeate (Fig. 4) and ferulate (3-methoxy-4-hydroxycinnamate), can also be transferred into the corresponding CoA thiol esters. As in the pathway in plants, Sam7 is thought to be a caffeoyl-CoA ligase, catalyzing the ligation of caffeate to CoA, leading to the corresponding thioester caffeoyl-CoA.
ACKNOWLEDGMENTS
We thank Gisela Grabellus for excellent technical assistance, V. Brecht for accomplishing the NMR analysis, and W. Heller (GSF Forschungszentrum fur Umwelt und Gesundheit, Neuherberg, Germany) for helpful discussion.
REFERENCES
Appert, C., E. Logemann, K. Hahlbrock, J. Schmid, and N. Amrhein. 1994. Structural and catalytic properties of the four phenylalanine ammonia-lyase isoenzymes from parsley. Eur. J. Biochem. 225:491-499.
Boniwell, J. M., and V. S. Butt. 1986. Flavin nucleotide-dependent 3-hydroxylation of 4-hydroxyphenylpropanoid carboxylic acids by particulate preparations from potato-tubers. Z. Naturforsch. C 41:56-60.
Buck, R. H., and K. Krummen. 1987. High-performance liquid-chromatographic determination of enantiomeric amino acids and amino alcohols after derivatization with o-phthaldialdehyde and various chiral mercaptans. J. Chromatogr. 387:255-265.
Calabrese, J. C., D. B. Jordan, A. Boodhoo, S. Sariaslani, and T. Vannelli. 2004. Crystal structure of phenylalanine ammonia lyase: multiple helix dipoles implicated in catalysis. Biochemistry 43:11403-11416.
Christenson, S. D., W. Liu, M. D. Toney, and B. Shen. 2003. A novel 4-methylideneimidazole-5-one-containing tyrosine aminomutase in enediyne antitumor antibiotic C-1027 biosynthesis. J. Am. Chem. Soc. 125:6062-6063.
Christenson, S. D., W. M. Wu, M. A. Spies, B. Shen, and M. D. Toney. 2003. Kinetic analysis of the 4-methylideneimidazole-5-one-containing tyrosine aminomutase in enediyne antitumor antibiotic C-1027 biosynthesis. Biochemistry 42:12708-12718.
Doumith, M., P. Weingarten, U. F. Wehmeier, K. Salah-Bey, B. Benhamou, C. Capdevila, J. M. Michel, W. Piepersberg, and M. C. Raynal. 2000. Analysis of genes involved in 6-deoxyhexose biosynthesis and transfer in Saccharopolyspora erythraea. Mol. Gen. Genet. 264:477-485.
Forkmann, G., and W. Heller. 1999. Biosynthesis of flavonoids, p. 713-748. In Sir Derek Barton and Koji Nakanishi (ed.), Comprehensive natural products chemistry. Elsevier Science Ltd., Oxford, United Kingdom.
Franke, R., M. R. Hemm, J. W. Denault, M. O. Ruegger, J. M. Humphreys, and C. Chapple. 2002. Changes in secondary metabolism and deposition of an unusual lignin in the ref8 mutant of Arabidopsis. Plant J. 30:47-59.
Franke, R., J. M. Humphreys, M. R. Hemm, J. W. Denault, M. O. Ruegger, J. C. Cusumano, and C. Chapple. 2002. The Arabidopsis REF8 gene encodes the 3-hydroxylase of phenylpropanoid metabolism. Plant J. 30:33-45.
Heller, W., and T. Kuhnl. 1985. Elicitor induction of a microsomal 5-O-(4-coumaroyl)shikimate 3'-hydroxylase in parsley cell suspension cultures. Arch. Biochem. Biophys. 241:453-460.
Humphreys, J. M., and C. Chapple. 2002. Rewriting the lignin roadmap. Curr. Opin. Plant Biol. 5:224-229.
Kneusel, R. E., U. Matern, and K. Nicolay. 1989. Formation of trans-caffeoyl-CoA from trans-4-coumaroyl-CoA by Zn2+-dependent enzymes in cultured plant cells and its activation by an elicitor-induced pH shift. Arch. Biochem. Biophys. 269:455-462.
Kojima, M., and W. Takeuchi. 1989. Detection and characterization of p-coumaric acid hydroxylase in mung bean, Vigna mungo, seedlings. J. Biochem. 105:265-270.
Kong, F. M., N. Zhao, M. M. Siegel, K. Janota, J. S. Ashcroft, F. E. Koehn, D. B. Borders, and G. T. Carter. 1998. Saccharomicins, novel heptadecaglycoside antibiotics effective against multidrug-resistant bacteria. J. Am. Chem. Soc. 120:13301-13311.
Kuhnl, T., U. Koch, W. Heller, and E. Wellmann. 1987. Chlorogenic acid biosynthesis—characterization of a light-induced microsomal 5-O-(4-coumaroyl)-D-quinate/shikimate 3'-hydroxylase from carrot (Daucus carota L.) cell suspension cultures. Arch. Biochem. Biophys. 258:226-232.
Kyndt, J. A., T. E. Meyer, M. A. Cusanovich, and J. J. Van Beeumen. 2002. Characterization of a bacterial tyrosine ammonia lyase, a biosynthetic enzyme for the photoactive yellow protein. FEBS Lett. 512:240-244.
Labeda, D. P., and R. M. Kroppenstedt. 2000. Phylogenetic analysis of Saccharothrix and related taxa: proposal for Actinosynnemataceae fam. nov. Int. J. Syst. Evol. Microbiol. 50:331-336.
Nair, R. B., Q. Xia, C. J. Kartha, E. Kurylo, R. N. Hirji, R. Datla, and G. Selvaraj. 2002. Arabidopsis CYP98A3 mediating aromatic 3-hydroxylation. Developmental regulation of the gene, and expression in yeast. Plant Physiol. 130:210-220.
Poppe, L., and J. Retey. 2005. Friedel-Crafts-type mechanism for the enzymatic elimination of ammonia from histidine and phenylalanine. Angew. Chem. Int. Ed. Engl. 44:3668-3688.
Rther, D., L. Poppe, G. Morlock, S. Viergutz, and J. Retey. 2002. An active site homology model of phenylalanine ammonia-lyase from Petroselinum crispum. Eur. J. Biochem. 269:3065-3075.
Rther, R., L. Poppe, S. Viergutz, B. Langer, and J. Retey. 2001. Characterization of the active site of histidine ammonia-lyase from Pseudomonas putida. Eur. J. Biochem. 268:6011-6019.
Schoch, G., S. Goepfert, M. Morant, A. Hehn, D. Meyer, P. Ullmann, and D. Werck-Reichhart. 2001. CYP98A3 from Arabidopsis thaliana is a 3'-hydroxylase of phenolic esters, a missing link in the phenylpropanoid pathway. J. Biol. Chem. 276:36566-36574.
Schwede, T. F., J. Retey, and G. E. Schulz. 1999. Crystal structure of histidine ammonia-lyase revealing a novel polypeptide modification as the catalytic electrophile. Biochemistry 38:5355-5361.
Singh, M. P., P. J. Petersen, W. J. Weiss, F. Kong, and M. Greenstein. 2000. Saccharomicins, novel heptadecaglycoside antibiotics produced by Saccharothrix espanaensis: antibacterial and mechanistic activities. Antimicrob. Agents Chemother. 44:2154-2159.
Stafford, H. A., and S. Dresler. 1972. 4-Hydroxycinnamic acid hydroxylase and polyphenolase—activities in Sorghum vulgare. Plant Physiol. 49:590-595.
Tanaka, M., and M. Kojima. 1991. Purification and characterization of p-coumaroyl-D-glucose hydroxylase of sweet potato (Ipomoea batatas) roots. Arch. Biochem. Biophys. 284:151-157.
von Mulert, U., A. Luzhetskyy, C. Hofmann, A. Mayer, and A. Bechthold. 2004. Expression of the landomycin biosynthetic gene cluster in a PKS mutant of Streptomyces fradiae is dependent on the coexpression of a putative transcriptional activator gene. FEMS Microbiol. Lett. 230:91-97.
Williams, J. S., M. Thomas, and D. J. Clarke. 2005. The gene stlA encodes a phenylalanine ammonia-lyase that is involved in the production of a stilbene antibiotic in Photorhabdus luminescens TT01. Microbiology 151:2543-2550.
Xiang, L. K., and B. S. Moore. 2002. Inactivation, complementation, and heterologous expression of encP, a novel bacterial phenylalanine ammonia-lyase gene. J. Biol. Chem. 277:32505-32509.
Xiang, L. K., and B. S. Moore. 2005. Biochemical characterization of a prokaryotic phenylalanine ammonia lyase. J. Bacteriol. 187:4286-4289.(Martin Berner, Daniel Kru)