Functional and Structural Characterization of aCation-dependent O-Methyltransferase from theCyanobacterium Synechocystis sp. Strain PCC 6803*□S

Received for publication, March 11, 2008, and in revised form, May 22, 2008 Published, JBC Papers in Press, May 23, 2008, DOI 10.1074/jbc.M801943200

Jakub Grzegorz Kopycki‡, Milton T. Stubbs‡, Wolfgang Brandt§, Martin Hagemann¶, Andrea Porzel§,Jurgen Schmidt§, Willibald Schliemann�, Meinhart H. Zenk**, and Thomas Vogt�1

From the ‡Department of Physical Biotechnology, Institute of Biochemistry and Biotechnology, Martin-Luther-UniversityHalle-Wittenberg, Halle (Saale) 06099, Germany, the Departments of §Bioorganic Chemistry and �Secondary Metabolism,Leibniz Institute of Plant Biochemistry, Halle (Saale) 06120, Germany, the ¶Institute of Biological Sciences, University of Rostock,Rostock 18051, Germany, and the **Donald Danforth Plant Science Center, St. Louis, Missouri 63132

The coding sequence of the cyanobacterium Synechocystis sp.strain PCC 6803 slr0095 gene was cloned and functionallyexpressed in Escherichia coli. The corresponding enzyme wasclassified as a cation- and S-adenosyl-L-methionine-dependentO-methyltransferase (SynOMT), consistent with considerableamino acid sequence identities to eukaryotic O-methyltrans-ferases (OMTs). The substrate specificity of SynOMTwas simi-lar with those of plant and mammalian CCoAOMT-like pro-teins accepting a variety of hydroxycinnamic acids andflavonoids as substrates. In contrast to the known mammalianand plant enzymes, which exclusively methylate the meta-hy-droxyl position of aromatic di- and trihydroxy systems, Syn-OMT also methylates the para-position of hydroxycinnamicacids like 5-hydroxyferulic and 3,4,5-trihydroxycinnamic acid,resulting in the formation of novel compounds. The x-ray struc-ture of SynOMT indicates that the active site allows for twoalternative orientations of the hydroxylated substrates in com-parison to the active sites of animal and plant enzymes, consist-ent with the observed preferred para-methylation and positionpromiscuity. Lys3 close to the N terminus of the recombinantprotein appears to play a key role in the activity of the enzyme.The possible implications of these results with respect to modi-fications of precursors of polymers like lignin are discussed.

Methylation by S-adenosyl-L-methionine (AdoMet)-de-pendent O-methyltransferases (OMTs)2 (EC 2.1.1) is a com-mon modification in secondary product biosynthesis (1). Site-

specific O-methylation modulates the physiological propertiesand the chemical reactivity of phenolic compounds and rendersthem more hydrophobic. Cation-dependent OMTs constitutea small group of lowmolecularmass (23 to 27 kDa) enzymes (2).In mammals, these enzymes play important roles in the modi-fication of catechol neurotransmitters in the brain ormay inac-tivate potentially bioactive metabolites like quercetin in theliver and kidney (3, 4). They are therefore referred to as catecholOMTs (COMT) and have been investigated as potential targetsto cure degenerate brain diseases (5). In plants, caffeoyl-coen-zymeAO-methyltransferases (CCoAOMTs), named after theirpreferred in vitro substrate, in conjunction with a second groupof cation-independent caffeic acid OMTs, are crucial for deter-mining the structural integrity of lignin in plant vascular tissues(6, 7). Specific subtypes of CCoAOMT-like proteins also meth-ylate, besides caffeoyl-CoA, other phenylpropanoids, preferen-tially flavonoids, with vicinal dihydroxy groups (2). The three-dimensional structures of eukaryotic animal and plant OMTsknown so far are quite similar despite their otherwise lowsequence identities, irrespective of the involvement of bivalentcations and substrates. Structural data obtained so far reveal aconserved AdoMet-binding site in all OMTs from the animaland plant kingdom (8–10). The methyl transfer mechanismproceeds via an Sn2-like transition state and a cation-facilitateddeprotonation of one of the two hydroxyl groups.Atomic structures of the cation-dependent catechol OMT

from rat (11) and the two groups of plant CCoAOMTs (10, 12)have been solved. Although the structural details and the sub-strate preferences of each enzyme subtype vary, their corre-sponding position specificity toward the meta-position in aro-matic structures is preferred. In plants, methylation is directedonly toward themeta-position, whereas in mammals this pref-erence is less stringent in case of themembrane-bound isoform(4). This scenario appears to be different in prokaryotes. In allprokaryotes analyzed so far and deposited in the databases(with the exception of archaebacteria), at least one member ofCCoAOMT-like proteins is present. In a recent report, theproduct of the SafC gene, a catecholOMT-like protein involvedin methylation of the isoquinoline derivative saframycin fromthe prokaryote Myxococcus xanthus, was described as methyl-ating the 4�-hydroxyl position of L-dihydroxyphenylalanine andof other catechols, such as caffeic acidwith a ratio of 3:1 in favor

* The costs of publication of this article were defrayed in part by the paymentof page charges. This article must therefore be hereby marked “advertise-ment” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

□S The on-line version of this article (available at http://www.jbc.org) containssupplemental Fig. S1.

The atomic coordinates and structure factors (code 3CBG) have been deposited inthe Protein Data Bank, Research Collaboratory for Structural Bioinformatics,Rutgers University, New Brunswick, NJ (http://www.rcsb.org/).

1 To whom correspondence should be addressed. Tel.: 49-345-5582-1530;Fax: 49-345-5582-1509; E-mail: tvogt@ipb-halle.de.

2 The abbreviations used are: OMT, O-methyltransferase; AdoMet, S-adeno-syl-L-methionine; AdoHcy, S-adenosyl-L-homocysteine; CCoAOMT, caf-feoyl coenzyme A O-methyltransferase; SynOMT, CCoAOMT-like proteinfrom Synechocystis sp. strain PCC 6803; PFOMT, phenylpropane and fla-vonoid O-methyltransferase from M. crystallinum; RP-HPLC, reverse phase-high performance liquid chromatography; GC-MS, gas chromatography-mass spectrometry.

THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 283, NO. 30, pp. 20888 –20896, July 25, 2008© 2008 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in the U.S.A.

20888 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 283 • NUMBER 30 • JULY 25, 2008

by Thom

as Vogt on A

ugust 8, 2008 w

ww

.jbc.orgD

ownloaded from

http://www.jbc.org/cgi/content/full/M801943200/DC1

Supplemental Material can be found at:

of the para-position (13). The structure of another prokaryoticprotein from the human pathogen Leptospira interrogans wasalso described, although neither the cation dependence nor anyfunctional data are known (14).In this article we report cloning, functional expression, enzy-

matic characterization, and x-ray structure elucidation of aCCoAOMT-like enzyme from the cyanobacterium Synechocys-tis sp. strain PCC 6803. In contrast to all known eukaryoticenzymes, this OMT displays a promiscuous position specificityfavoring the para-hydroxyl group of polyhydroxylated aro-matic phenolics for the attachment of the methyl moiety. Dataare discussed with respect to structure differences of the activesite, evolution, and application of this new enzyme activity inthe design of aromatic mono- and polymers.

EXPERIMENTAL PROCEDURES

Growth of Synechocystis sp. Strain PCC 6803 and DNAIsolation—The glucose-tolerant strain of Synechocystis sp. PCC6803was obtained fromN.Murata (National Institute for BasicBiology, Okazaki, Japan) and grown photoautotrophically inbatch cultures at 29 °C under continuous illumination of 130�mol of photons s�1m�2 (warm light, OsramL58W32/3) withbubbling of air enriched with CO2 (5%) in BG11medium at pH8.0. High molecular DNA was obtained by lysozyme action onthe cells, purification with cetylammonium bromide, followedby phenol/chloroform treatments as described (15).Isolation of the slr0095 DNA Sequence and Functional

Expression—The slr0095 coding sequence was amplified byPCR using Pfu polymerase (Promega) with the following prim-ers: N-terminal primer 5�-CGGGATCCATGGGTAAGGGC-ATCA-3�; C-terminal reverse primer 5�-CTCAAGCTTGTG-CTATTTTTTGAGTGCCA-3�. Subsequently the fragmentwas cut with BamHI and HindIII, ligated into the multicloningsite of the pQE30 expression vector (Qiagen), and transformedinto chemically competent Escherichia coliM15rep cells (Qia-gen). Inducing the bacteria with 1 mM isopropyl 1-thio-�-D-galactopyranoside at exponential growth and harvesting 3 hafter induction lead to accumulation of aHis-tagged fusion pro-teinwith a yield of about 1mg of soluble Slr0095 protein/liter ofE. coli culture. The protein was purified by metal affinity chro-matography using the Talon matrix (Clontech), checked forpurity by SDS-PAGE, and rebuffered into 20 mM Tris/HCl, pH7.5, by ultrafiltration (Millipore Waters). For determination ofthe molecular mass, size exclusion chromatography was per-formed on Superdex 200 (GE Healthcare) in 20 mM Tris/HCl,150 mM NaCl, pH 7.5. Crystallization trials and all assays wereperformed with fresh protein preparations, because the activeenzymewas unstable and could be stored for prolonged periodsat a concentration of 1 mg/ml only with the inclusion of 10%PEG 400. Even then, after 2 months at �20 °C a 20% loss ofactivity was observed.Structural Characterization of the Cation-dependent

SynOMT—Purified recombinant SynOMTwas set up for crys-tallizationwith a standard set of buffers. Crystals were obtainedfrom 0.2 M MgCl2, 0.1 M Tris/HCl, pH 8.5, 30% polyethyleneglycol 4000 (w/v). Diffraction data were collected on a MSCRigaku Raxis IV�� diffractometer at a wavelength of 1.54 Å.The crystals diffracted to 2.0 Å and belong to the space group

P3121 with cell constants of a � b � 57.62 Å and c � 119.83 Å.Integration and scaling of the diffraction data were performedusing the HKL2000 software package (16). Analysis of theintensity distribution indicated merohedral twinning, with atwinning fraction of 0.284 (twinning operator �h,�k,l) deter-mined usingCNS (17). The structure of SynOMTwas solved bymolecular replacement with the monomer of M. crystallinumPFOMT (12) (PDB ID code 3C3Y) using the program PHASER(18), part of the CCP4 suite (19), revealing one molecule in theasymmetric unit. Manual rebuilding of the model was per-formed with the use of the programs O (20) and Coot (21). Thestructure was refined using CNS to an r � 0.152 and twinnedRfree � 0.215. Data statistics are given in Table 4. The structurehas been assigned the PDB ID code 3CBG. The bootstrappedphylogenetic cladogram was constructed using the TREECONsoftware tools (22).Activity Tests and Kinetic Data—Enzyme assays were per-

formed 100 mM potassium phosphate buffer, pH 7.5, and 10%glycerol with 10 �M substrate (dissolved in 30% dimethyl sulf-oxide), 0.5–2 �g of total protein, and 400 �MAdoMet in a totalvolume of 50�l. Assays were incubated at 30 °C for 60 to 3600 s(dependent on the protein and substrate tested) and stopped bythe addition of 20 �l of 7% trichloroacetic acid in 50% acetoni-trile/water. The reaction products were analyzed by HPLC asdescribed (2). Caffeoyl glucose was prepared as described pre-viously (23) from caffeic acid and UDP-glucose with the puri-fied recombinant sinapic acid glucosyltransferase. Caffeoyl-CoA was prepared based on published methods (24, 25).Caffeic acid was obtained from Serva, quercetagetin from

Extrasynthese, and quercetin from Roth, respectively. Tricetinwas a generous gift fromRagai Ibrahim (Montreal, Canada). Allsubstrates and reaction products were analyzed by RP-HPLCon aNucleosil 5-�mC18 column (50mm length� 4mm, innerdiameter; Macherey & Nagel), as described previously (2).Compounds were analyzed with linear gradients from 10% B(acetonitrile) inA (1.5% aqueous phosphoric acid) to 70%B inA(for phenolics), from 5%B to 50%B inA (for free acids andCoAesters), from 5%B to 30% B (for glucose esters), and from 20% Bin A to 80% B in A (for flavonoids) in 4 min at a flow rate of 1ml/min. UV detection of flavonoids, catechol, coumarins, andhydroxycinnamic acid esters was performed between 200 and400 nm. Identification and quantification was achieved withreference compounds from our institute collection. For Kmdetermination of methyl group acceptors, acceptor concentra-tions were chosen between 2 and 40 �M, whereas AdoMet waskept constant at 1.5 mM. The Km and Vmax values were calcu-lated from Lineweaver-Burk plots (26). All enzyme assays wererecorded from at least two different experiments each intriplicate.Identification of Methylation Products of 3,4,5-Trihydroxy-

cinnamic Acid—SynOMT reactions were up scaled in 50-mlFalcon tubes at 37 °C with up to 5mg of purified enzyme. Reac-tions were run for 60min, centrifuged at 10,000 � g for 10min,and the supernatants were applied directly to C18-reversedphase solid phase extraction columns (Millipore Waters),washed with 5 column volumes deionized water, and the com-pounds were eluted with MeOH. The reaction products werefurther purified by RP-HPLC on a Nucleosil 5-�mC18 column

O-Methyltransferase from Synechocystis

JULY 25, 2008 • VOLUME 283 • NUMBER 30 JOURNAL OF BIOLOGICAL CHEMISTRY 20889

by Thom

as Vogt on A

ugust 8, 2008 w

ww

.jbc.orgD

ownloaded from

at 3 ml/min (250 � 8 mm, inner diameter; Macherey & Nagel)with a linear gradient from 10% B (MeOH) in A (0.2% aqueousacetic acid) to 60% B in A over 30 min. Fractions from severalruns were combined, concentrated, and subjected to NMRspectroscopy and mass spectrometry.

1H and homo- and heteronuclear two-dimensional NMRspectra from the major products dissolved in CD3OD wererecorded on a Inova 500NMR spectrometer (Varian) operatingat 499.81 MHz using a 3-mm microsample inverse detectionprobe. Chemical shifts are referenced to internal TMS (� � 0ppm, 1H) and CD3OD (� � 49.0 ppm, 13C), respectively (Table1). Negative ion electrospray (ESI-MS) high-resolution massspectra were obtained from a BRUKERApex III Fourier Trans-form ion cyclotron resonance mass spectrometer equippedwith an external APOLLO electrospray ion source: trihydroxy-cinnamic acid,m/z 195.02981 ([M-H]�, calculated for C9H7O5

�

195.02990) and 3,5-dihydroxy 4-O-methylcinnamic acid, m/z209.04577 ([M-H]�, calculated for C10H9O5

� 209.04555). Sig-nals corresponding to 3-hydroxy 4,5-di-O-methylcinnamicacid could be observed neither in positive nor negative mode.GC-MS measurements of the trimethylsilyl derivatives of

vanillic acid and isovanillic acid were performed on a Voyager/Trace GC 2000 (Therm CE Instruments) under the followingconditions: 70 eV EI, source temperature 200 °C; columnDF-5MS (J&W, 30 m � 0.25 mm; 0.25 �m film thickness);injection temperature 250 °C, interface temperature 300 °C;carrier gas helium, flow rate 1.0 ml/min, constant flow mode;splitless injection, column temperature program: 60 °C for 1min raised to 300 °C at a rate of 10 °C/min and hold on 300 °Cfor 5min. Vanillic acid trimethylsilylether: retention time 15.91min, EIMS (m/z, rel. int.(%): 312 [M�] (30), 297 [M-CH3]� (58),282 (26), 267 (56), 253 (29), 223 (32), 193 (14), 165 (10), 137 (14),126 (17), 73 (100). Isovanillic acid trimethylester: retention time15.97 min, EIMS (m/z, rel. int. in%): 312 [M�] (36), 297[M-CH3]� (76), 282 (31), 267 (58), 253 (34), 223 (36), 193 (15),165 (9), 137 (13), 126 (13), 73 (100).Gas chromatography (GC)/time-of-flight (TOF)-mass spec-

trometry (MS) measurements of several hydroxycinnamic acidderivatives and corresponding enzymatically methylated prod-ucts were performed with an 6890 series gas chromatograph(Agilent) equipped with an autosampler 7683 series and aMicroMass GCT with a TOF detector (Waters) as recentlydescribed (27) (see Table 2).Screening andDocking of Natural Products—Apharmacoph-

ore model of the active site was generated using MOE (28)based on the x-ray structure of SynOMT including distinctbinding modes of ferulic and isoferulic acids. The modelincludes the three hydrophilic binding sites related to the phe-nolic hydroxyl or methoxy groups, the center of the aromate,and a hydrophilic area defined by the carboxylic acid group ofthe ligands. All amino acid residues of the active site were usedto define an excluded volume, so that no spatial overlap of anyligand with this region was allowed. This model was used toscreen theMOE (28) and KEGG (29) databases for compoundscontaining three-dimensional structures with low energy con-formations. Highly significant hits were docked to the activesite using PLANTS (30).

RESULTS

Functional Characterization of Recombinant slr0095Protein—Slr0095 represents the only gene identified in theSynechocystis sp. strain PCC 6803 genome coding for a proteinwith a significant identity to cation-dependentOMTs fromani-mals and plants. The genes in the immediate neighborhood ofslr0095, a putative heat shock protein and a sulfate transporter,gave no definite hint to the potential function of the gene in vivo(Fig. 1). Similar genes are found in a large number of prokary-otic genomes. Based on sequence identity searches in variousdatabases, more than 200 genes from different prokaryotic or-ganisms were found when complete or partial sequence motifsof plant CCoAOMT were used as in silico probes. Sequenceidentity to plant CCoAOMTs was much higher (up to 42% atthe protein level) than to mammalian COMTs or bacterialCCoAOMT-like proteins (data not shown). A small subset andclustering of eukaryotic and prokaryotic sequences is displayedin Fig. 2. Although the prokaryotic sequences of CCoAOMT-like proteins, including Synechocystis slr0095, are distinct fromplant sequences and appear to cluster together, the low boot-strap values do not allow the phylogenetic relationships amongthe prokaryotic sequences and non-plant eukaryotic sequencesto be clearly defined.The His-tagged and metal affinity purified SynOMT dis-

played amolecular mass of about 25 kDa on SDS-PAGE, corre-sponding to the predictedmass of themonomeric unit based onthe amino acid sequence (Fig. 3). Using size exclusion chroma-tography under non-denaturing conditions, a dimeric structurewith a calculatedmolecularmass of about 50 kDa appears likely(data not shown). In this respect the cyanobacterial SynOMTresembles the dimeric plant proteins rather than the animalCOMTs, generally described as monomeric (10–12).Characterization of Enzymatic Activities and Kinetic

Measurements—Plant and mammalian cation-dependentOMTs in principle methylate the same type of substrates, i.e.aromatic compounds with vicinal di- or trihydroxy systems.The observed sequence identity and the presence of the char-acteristic AdoMet-binding domains suggested that the cya-nobacterial OMTs would use the same or a similar set of com-pounds in vitro. SynOMT displayed activity against a variety ofsubstrates (Table 3). Cation dependence was verified by inclu-

FIGURE 1. Part of the genetic map of the Synechocystis sp. strain PCC6803. A detailed view of a putative operon including the slr0095 locus. Thefunction of the protein encoded by slr0092 is unknown, slr0093 encodes aheat-shock protein, and slr0096 encodes a sulfate transporter (44),respectively.

O-Methyltransferase from Synechocystis

20890 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 283 • NUMBER 30 • JULY 25, 2008

by Thom

as Vogt on A

ugust 8, 2008 w

ww

.jbc.orgD

ownloaded from

sion or absence of EDTA in the buffers. Using 3,4,5-trihydroxy-cinnamic acid and 5-hydroxyferulic acid as substrates, a novelproduct profile was observed by HPLC (Fig. 4A). 5-Hydroxy-ferulic acid was methylated to sinapic acid and an additionalunknown compound, whereas 3,4,5-trihydroxycinnamic acidand sinapic acid gave two peaks that could not be correlated toknown standards by HPLC coupled to a diode-array UV-detec-

tion system (Fig. 4A). Absorbancemaxima of the newly found prod-ucts in the near UV were at 308 nm,inconsistent with any hydroxycin-namic acid derivatives described sofar. The products of 3,4,5-trihy-droxycinnamic acid methylationcatalyzed by a typical plantCCoAOMT-like protein, PFOMTfrom Mesembryanthemum crys-tallinum (2), 5-hydroxyferulic andsinapic acid, display characteristicabsorbance maxima between 320and 325 nm (Fig. 4B). The lack ofproduct formation by SynOMTwith p-coumaric, ferulic, orsinapic acids ruled out that thetransfer of the methyl group couldbe directed toward the carboxylgroup of the phenylpropanoids,initially considered as a plausiblealternative.For identification of the novel

major product, NMR spectroscopyand ESI-MS data as well asGC/TOFMS data were combinedwith the UV data and expectedspecificities of cation-dependentOMTs (Tables 1 and 2). Combined

use of one- and two-dimensional NMR experiments resulted inan unambiguous assignment of all 1H and 13C NMR signals of3,5-dihydroxy-4-methoxycinnamic acid (with the exception ofthe 13C signal of –COOH, for sensitivity reasons). Only a 4-O-methylation is in accordance with the isochronous NMR sig-nals of positions 2/6 and 3/5. Furthermore, the para-position ofthe O-methyl group is proved by the high-field shift of C-4 (�13C 139.0 ppm), caused by the two-electron releasing hydroxylsubstituents in ortho-position to C-4. In addition, no NOEenhancement is observed between the signal of the methoxylgroup and the aromatic protons. These data are only consistentassuming an unusual preference of this enzyme for the para-position, leading to the formation of 3,5-dihydroxy-4-methoxy-cinnamic acid (P1 in Fig. 5). The simultaneous appearance oftwo dimethylated compounds in addition to 3,5-dihydroxy-4-methoxycinnamic acid in the HPLC analysis was confirmed byGC-MS data (Table 2, Figs. 4A and 5A). One compound wasidentified as 3,5-dimethoxy-4-hydroxycinnamic acid (sinapicacid) according to spectral properties and retention times inHPLC as well as GC-MS. The second compound with identicalspectral properties to the para-methylated 3,4,5-trihydroxy-cinnamic acid showed different LC and GC retention timescompared with sinapic acid (Table 2, Figs. 4B and 5B) and,corroborated by GC-MS data were thus identified as 5-hy-droxy-3,4-dimethoxycinnamic acid. Consequently, incubationof SynOMT with 5-hydroxyferulic acid, consistent with previ-ous data, resulted in the formation of 5-hydroxy-3,4-dime-thoxycinnamic acid as well as sinapic acid as expected (Fig. 5A).This corroborates the observed specificity of CCoAOMT-like

FIGURE 2. Bootstrapped cladogram of selected CCoAOMTs and CCoAOMT-like proteins from pro- andeukaryotes as produced by TREECON (22). The neighbor joining tree was based on 1000 bootstrap trials.Notice that all bootstrap values below 500 do not allow phylogenetic relationships to be clearly defined. Thecluster of six plant proteins marked as CCoAOMTs is specific for caffeoyl-coenzyme A methylation in ligninmonomer biosynthesis. Gene products of Arabidopsis CCoAOMTs are listed based on the corresponding geneidentifiers. Crystallized proteins are underlined. The mammalian catechol OMT from rat was used as an out-group. NCBI data base accession numbers used are: BAA10567 (Synechocystis PCC 6803 slr 0095); NP_710596 (L.interrogans); Q00719 (S. mycarofaciens); BAB76778 (Nostoc sp. PCC 7120); ZP_00108749 (Nostoc punctiforme2);ZP_00111674 (Nostoc punctiforme1); AAM33748 (Dictyostelium discoideum); AAC44130 (M. xanthus);AAB61680 (Stellaria longipes); AAN61972 (M. crystallinum PFOMT); AAM64800 (Arabidopsis thaliana At4g 26220AOMT1); AJ242980 (Zea mays CCoAOMT); AAD02050 (Pinus taeda CCoAOMT); CAA12198 (Populus trichocarpaCCoAOMT); AAT40111 (Ammi majus CCoAOMT), Q40313 (M. sativa CCoAOMT); AY057554, A. thalianaAt1g34050), P22734 (rat catechol OMT).

FIGURE 3. SDS-PAGE demonstrating the purification of SynOMT. Lane 1,crude E. coli extract 4 h after isopropyl 1-thio-�-D-galactopyranoside induc-tion; lane 2, affinity column run-through; lane 3, wash fraction; lane 4, purifiedSynOMT fraction; lane 5, molecular mass ladder. The position of the SynOMT ismarked with an arrow.

O-Methyltransferase from Synechocystis

JULY 25, 2008 • VOLUME 283 • NUMBER 30 JOURNAL OF BIOLOGICAL CHEMISTRY 20891

by Thom

as Vogt on A

ugust 8, 2008 w

ww

.jbc.orgD

ownloaded from

proteins for vicinal dihydroxy groups, but also indicates anapparent lack of position specificity. If a 3,4,5-trihydroxycin-namic acid molecule is methylated at the para-position, it doesnot undergo any further methylations, due to the absence ofadditional vicinal dihydroxy groups resulting in the majormonomethylated product (Fig. 5A). In a second scenario the3,4,5-trihydroxy-cinnamic acid is initially methylated at themeta-position, which is further methylated either at the para-or the secondmeta-position, resulting in a 50% ratio of 3,4- and3,5-di-O-methylated products, respectively (Fig. 5A).When tri-hydroxylated flavonoids like the 3,3�,4�,5,5�,7-hexa-hydroxyfla-vone (myricetin) and the 3�,4�,5,5�,7-pentahydroxyflavone

(tricetin) were used as substrates, the enzyme displayed ahigher position specificity, resulting only in the single 4�-O-methylated product (data not shown). This observation again isin contrast to the results for the corresponding plant enzymes,where two products were always observed with myricetin,tricetin, and 3,4,5-trihydroxycinnamic acid, consistent withmethylation in the twometa-positions only (2).Kinetic data of the array of substrates tested indicate that the

trihydroxylated flavone tricetin is the best substrate (Table 3).Caffeoylglucose and especially 3,4-dihydroxybenzoic acid showa reduced affinity toward the enzyme, although their catalyticefficiencies in vitro are fairly similar to the best transformedsubstrates. In the case of 3,4-dihydroxybenzoic acid, the forma-tion of 3-methoxy-4-hydroxybenzoic acid (vanillic acid) takesplace at a 10-fold higher rate than the formation of 4-methoxy-3-hydroxybenzoic acid (isovanillic acid) as calculated from thepeak areas in the GC-MS. It should be noted, however, that thesubstrates used in this study represent either synthetic com-pounds (3,4,5-trihydroxycinnamic acid) or are of plant originand that the in vivo substrates of the SynOMT remain to beidentified.Structural Characterization of the Cation-dependent

SynOMT—Crystals of SynOMT diffracting up to 2.0 Å (Table4) were obtained in the presence of Mg2�, AdoMet, and caffeicacid. Inspection of the electron density clearly revealed thepresence of AdoHcy and twomethylation products, ferulic acidand isoferulic acid, in the active site of the enzyme (Fig. 6). Theoverall structure of SynOMT is that of a compact dimer whosefold closely resembles that of the plant CCoAOMTs. Asdescribed for Medicago sativa CCoAOMT (10) and M. crys-tallinum PFOMT (12), cyanobacterial SynOMT possesses thehighly conserved �/� Rossmann fold of the AdoMet-depend-ent OMTs (8, 11), with the N-terminal helix contributing sig-nificantly to the dimer interface. SynOMTdiffers from its plantcounterparts in the disposition of the N terminus and in theconformation of the “insertion loop.” These regions, that arealso among the most divergent regions of the divalent metalOMTs (supplemental Fig. S1), have been implicated in sub-strate specificity, in particular for caffeoyl-CoA (12).The residues Asp143, Asp169, and Asp170, responsible for

metal coordination, are conserved among theMg2�-dependentOMTs, as is the presence of a mediating water molecule thatmakes hydrogen bonds to the carbonyl oxygen ofMet42 and theamide nitrogen of Ile44. The latter residue abuts the carboxylategroup of AdoHcy. Interestingly, the equivalent of Ile44 is a thre-onine residue in the plant enzymes (supplemental Fig. S1),whoseO� is involved in awater-mediated hydrogen bond to theMg2� ion as well as a direct hydrogen bond to the AdoMet/AdoHcy carboxylate moiety, resulting in a slight displacementof the cofactor C� atom. As this residue is aliphatic in the bac-terial enzymes, the small displacement as well as changes in theelectronic characteristics of the metal ion could cause subtledifferences in substrate specificity. Otherwise, cofactor bindingis conserved among all the metal-dependent OMTs studied todate, involving hydrogen bonds to Gly68, Asp92, Ala121, andAsp145. The two products, ferulic and isoferulic acid, methyl-ated in the meta- and para-positions, respectively, are clearlydelineated in the electron density (Fig. 6B). The propenoic acid

FIGURE 4. HPLC pattern of SynOMT-enzyme assays. Assays measured at318 nm illustrate the methylation profiles (left) and UV spectra of products(right) obtained with the substrate 3,4,5-trihydroxycinnamic acid either bythe Slr0095 protein from Synechocystis sp. strain PCC 6803 (A) or the PFOMTfrom M. crystallinum (B) (2). S, substrate; P1, 3,5-dihydroxy-4-methoxycin-namic acid; P2, 3,5-dimethoxy-4-hydroxycinnamic acid (sinapic acid); P3,5-hydroxy-3,4-dimethoxycinnamic acid; P4, 3-methoxy-4,5-dihydroxycin-namic acid (5-hydroxyferulic acid).

TABLE 11H and 13C NMR data of 3,5-dihydroxy-4-methoxycinnamic acid

Position � 1H, multiplicity, relative integral intensity � 13Ca

ppm/J Hz ppm1 132.42/6 6.588, s, 2H 108.63/5 152.14 139.04-OCH3 3.824, s, 3H 60.87 7.448, d (15.9), 1H 146.58 6.246, d (15.9), 1H 118.09 NDb

a Chemical shifts of HSQC and HMBC correlation peaks.b ND, not determined.

O-Methyltransferase from Synechocystis

20892 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 283 • NUMBER 30 • JULY 25, 2008

by Thom

as Vogt on A

ugust 8, 2008 w

ww

.jbc.orgD

ownloaded from

moieties of the two products coincide; although the carboxylatemoiety approaches the side chains of His174 and Lys176 (thelatter conserved in the SafC gene product of M. xanthus, sup-plemental Fig. S1), it is essentially exposed to the solvent. Thearomatic rings of both products are sandwiched betweenthe side chains of Trp173 and Met42. The para-methoxylatedcatechol ring of the isoferulic acid product occupies the activesite in a substrate-like manner, both hydroxyl groups are in

contact with the catalytic magnesium atom, with the methylgroup juxtaposing theAdoHcy thiol group and in the neighbor-hood of absolutely conserved Lys146, which could act as a cata-lytic base for a nucleophilic methyl transfer reaction as sug-gested for ratCOMT(11). Themeta-substituted ferulic acid, onthe other hand, demonstrates an unproductive binding mode,with the para-hydroxy group directed toward the cofactor thioland the meta-methylated function pointing away from themetal ion. No specific stabilizing interactions for this latterbinding mode are observed; a corresponding arrangement offeruloyl-CoA was also observed for CCoAOMT (10).Previously unobserved in the plant CCoAOMTs, the

N-terminal residues of SynOMT reach into the active site,with the amino group of Lys3 closely approaching the mag-nesium ion and the meta-hydroxyl group of isoferulic acid.

FIGURE 5. Reaction schemes comparing product formation with 3,4,5-trihydroxycinnamic acid as substrate. A, SynOMT from Synechocystis sp.strain PCC6803 (SynOMT) and B, PFOMT from M. crystallinum. S, substrate3,4,5-trihydroxycinnamic acid; P1, 3,5-dihydroxy-4-methoxycinnamic acid;P2, sinapic acid; P3, 5-hydroxy-3,4-dimethoxycinnamic acid; P4, 5-hydroxy-ferulic acid. In the case of SynOMT, P4, the precursor for P2 and P3, couldnever be detected by HPLC and is therefore illustrated in reduced intensity.

TABLE 2GC/TOFMS data of derivatized hydroxycinnamic acid standards and products of 3,4,5-trihydroxycinnamic acid methylation catalyzed bySynOMT

IdentifierRt Found Calculated for Compound

cis transmin m/z m/z

C18H32O4Si3 Caffeic acid-3TMS1117 30.67 32.94 396.1589 396.1608

C16H26O4Si2 Ferulic acid-2TMS1118 29.67 32.32 338.1372 338.1370

C21H40O5Si4 3,4,5-Trihydroxycinnamic acid-4TMS1121 33.07 34.92 484.1994 484.1953

C19H34O5Si3 5-Hydroxyferulic acid-3TMS1119 32.65 34.92 426.1749 426.1714

C17H28O5Si2 Sinapic acid-2TMS1120 31.92 34.54 368.1492 368.1475

C19H34O5Si3 3,5-Dihydroxy-4-methoxycinnamic acid-3TMS1129 31.57 33.70 426.1741 426.1714

C17H28O5Si2 5-Hydroxy-3,4-dimethoxycinnamic acid-2TMS1130 31.20 33.57 368.1347 368.1475

TABLE 3Kinetic parameters of the recombinant CCoAOMT-like SynOMT fromSynechocystis sp. strain PCC 6803

Substrates Km kcat kcat/Km

�M liter/s M�1 s�1

5-Hydroxyferulic acid 74.0 0.02500 340Caffeic acid 69.3 0.00551 80Caffeoyl-CoA 32.9 0.02450 740Caffeoylglucose 106 0.00951 903,4,5-Trihydroxycinnamic acid 20.7 0.01040 500Tricetin (3�,4�,5,5�,7-pentahydroxyflavone) 5.00 0.00885 17803,4-Dihydroxybenzoic acid 215 0.00802 37

TABLE 4Crystallographic data, phasing, and refinement statistics for SynOMTThe values in parentheses represent the values for the highest resolution shell.

DatasetData collectionWavelength (Å) 1.5418Resolution (Å) 30-2.0Total reflections 140,633Unique reflections 16,207Completeness (%) 100 (99.9)II/�(I)a 12.4 (2.6)Rsym

a,b 0.124 (0.703)Redundancya 8.7 (6.9)

RefinementRefinement reflections 16,164Twinned Rc 0.152Twinned Rfree

d 0.215Protein atoms 1,703Ligand atoms 55Water molecules 124Root mean square deviation bond lengthse (Å) 0.0063Root mean square deviation bond anglese (o) 1.24�B-factor� protein (Å2) 28.9Twinning fraction 0.28Twinning operator -h, -k, lRamachandran plot 91.9% core

7.0% allowed1.1% generously allowed

a The values in parentheses represent the values for the highest resolution shell.bRsym � ��Ih � �In��/�ih, where �Ih� is the average intensity over symmetry-equiva-lent reflections.

c r � ��Fobs� � �Fcalc�/��fobs�.dRfree is calculated as R with 9.2% of the data (1485 reflections) excluded from therefinement.

e The rootmean square deviation for bonds and angles are the rmd from ideal values.

O-Methyltransferase from Synechocystis

JULY 25, 2008 • VOLUME 283 • NUMBER 30 JOURNAL OF BIOLOGICAL CHEMISTRY 20893

by Thom

as Vogt on A

ugust 8, 2008 w

ww

.jbc.orgD

ownloaded from

As small rearrangements of the side chain could allow directparticipation in the catalytic reaction, Lys3 was mutated toalanine. The resultant enzyme showed no activity towardcaffeic acid conjugates or trihydroxylated substrates. A verylow activity could be determined only in the case of 5-hy-droxyferulic acid and caffeic acid (0.7 and 0.6% as comparedwith the wild type protein) upon prolonged incubationtimes, indicating that Lys3 is absolutely required for enzy-matic activity with hydroxylated substrates.To identify structural features of unknown in vivo substrates

of SynOMT, two databases were screened for compounds con-sistent with the docking arrangement of the ligands in the x-raystructure of neighboring phenolic hydroxyl groups complexedby a magnesium allowing for a catalytic methylation of onehydroxyl group via AdoMet (Table 5). From more than 1000putative ligands in the MOE and KEGG databases about 100showed higher affinities than caffeic or ferulic acid. Best scoreswere obtained for plant-derived hydrolysable tannins laevigatinD and pentagalloylglucose (31, 32) both with several aromatictrihydroxy groups (Table 5). A search for putative substrates in

Synechocystis-related prokaryotes revealed the antimitoticcryptophycins, 16-memberedmacrolides from the cyanobacte-riumNostoc (33). Interestingly, cryptophycin 1 contains an aro-matic methoxy group in the para-position and a clorine in themeta-position. The compound showed a higher in silico affinityfor the active site of SynOMT than the enzymatically testedhydroxycinnamic acid derivatives (Table 5).

DISCUSSION

SynOMT from Synechocystis sp. strain PCC 6803 is the firstcation- and AdoMet-dependent O-methyltransferase to becharacterized and crystallized from a prokaryote. Although thenatural substrate of this enzyme is currently unknown, the sub-strate specificity for vicinal dihydroxy groups of phenolics iscomparable with the currently known eukaryotic proteins. Theposition specificity is strikingly different, with promiscuitytoward both meta- and para-positions never observed for anyof the eukaryotic enzymes. A comparable specificity has beendescribed for the prokaryotic SafC gene product fromM. xan-thus when tested with caffeic acid, although the natural sub-strate for the latter enzyme appears to be L-dihydroxypheny-lalanine, a biosynthetic precursor of the isoquinoline alkaloidsaframycin (13). The precise structural features of the Synecho-cystis protein that contribute to this unexpected phenomenonare as yet unclear. One possibility is provided by side chains ofHis174 and Lys176 in the neighborhood of the propenoic acidmoiety, equivalent to Asn and Ser in the plant CCoAOMTs,respectively (supplemental Fig. S1). Curiously, two distinctbinding modes for the monomethylated caffeic acid products,ferulic acid and isoferulic acids, were found. Whereas the posi-tioning of the aromatic ring oxygen atoms of isoferulic acidsuggest a substrate-like interaction with the active site magne-sium and the AdoMet cofactor, ferulic acid binds with only oneoxygen approaching the metal ion.Although theN-terminal lysine, Lys3,must be involved in the

catalytic mechanism, the almost complete abrogation of activ-ity on mutation to alanine suggests a more fundamental rolethan determining position specificity. N-terminal residues havealready been implicated in position specificity determination inthe plant enzymes (34). Loss of two N-terminal lysine residuesin the plant enzyme PFOMT results in a sequential and partialloss of activity (2, 10, 12), whereas the presence of an N-termi-nal His tag influences substrate specificity (12). The individualcontribution ofN-terminal lysine residues in the plant enzymes

FIGURE 6. Structureof SynOMT. A, overall structure of the dimer in schematicrepresentation. The (crystallographically related) monomers are shown in yel-low and green, with the N-terminal peptides and insertion loops highlighted.Selected residues at the active site, together with the cofactor AdoMet, theproducts ferulic and isoferulic acid, are shown in stick representation,whereas the magnesium ions are shown as magenta spheres. B, close up of theactive site, together with the experimental 2Fo � Fc electron density for thepara-methylated product isoferulic acid (pink, in a substrate-like bindingmode) and the meta-methylated ferulic acid (light green).

TABLE 5Docking results of selected compounds obtained by screening theBeilstein (29) and KEGG (30) databasesBest hits of each database (laevigatinD and 1,2,3,4,6-pentagalloyl-�-D-glucose)weredocked to the three-dimensional structure of SynOMT and compared to scores forhydroxycinnamic acid and cryptophycin 1, a substrate candidate from Nostoc spec.Most negative scores relate to highest expected affinity.

ID Compound Fitness scorea

1 Laevigatin D �134.52 1,2,3,4,6-Pentagalloyl-�-D-glucose �132.63 Caffeoyl-CoA �98.34 Cryptophycin 1 �81.25 Caffeic acid �74.46 Trihydroxycinnamic acid �73.87 Ferulic acid �68.6

a The scores are crude indications for the affinity of the ligands to the enzyme.

O-Methyltransferase from Synechocystis

20894 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 283 • NUMBER 30 • JULY 25, 2008

by Thom

as Vogt on A

ugust 8, 2008 w

ww

.jbc.orgD

ownloaded from

cannot be evaluated in detail due to a lack of structural data.Sequence alignments of the dimeric metal-dependent OMTs(supplemental Fig. S1) reveal lysine andhistidine residues in theN-terminal regions of each of these enzymes, suggesting thatminor differences in position specificity could reside here.Interestingly, the dimeric structure of the OMT of L. interro-gans (14), inwhich theN termini are domain-swapped, suggeststhat theN-terminal residues could switch from the active site ofone monomer to the other, allowing for a wider variety ofpotential products as observed for PFOMT and SynOMT. Theproposed monomeric cation-dependent mammalian COMTslack any lysine residues in the N-terminal region (5).In contrast to hydroxycinnamic acids, flavonoids with para-

methylated B-rings are observed throughout the plant kingdom(35), although methylation is sometimes performed by cation-independent enzymes with strict structural requirements. Forexample, a flavonoid-specific OMT from Catharanthus roseusmethylates the 4�-position only when the neighboring 3�-me-thoxy group of the flavonoid B-ring is present, as in case of theisorhamnetin or homoeriodictyol (36). Isorhamnetin is not asubstrate of SynOMT, as it lacks the required vicinal dihydroxysystem. A similar cation-independent OMT, cloned from riceand functionally expressed inE. coli, performs sequentialmeth-ylation of three B-ring hydroxyl groups of the flavone tricetin invitro, including the para-position (37). In contrast to the plantproteins, the para-position is an initial target of the prokaryoticenzyme, at least in the case of the phenylpropanoid esters, fol-lowed by the subsequent attack of either the 3�-OH or 5�-OHgroup.With trihydroxylated flavonoids, tricetin andmyricetin,these subsequent methylation reactions are missing and only asingle product is observed with SynOMT.Although 3,4,5-trihydroxycinnamic acid has so far not been

found to occur in the plant kingdom, 5-hydroxyferulic acid, thealdehyde, alcohol, or the corresponding caffeic acid derivativescould provide potential in vivo substrates for this new type ofenzyme. It is probably not that surprising that para-methylatedintermediates of lignin biosynthesis are unknown. Current lig-nin models favor radical-radical rather than enzyme-mediatedcoupling with an unsubstituted para-hydroxyl group requiredfor resonance stabilization to promote cross-linking of differ-ent monomer units (38). Para-methylated (oligo)lignols woulddestabilize radical formation resulting in reduced cross-linkingand a different lignin structure. An enzymatic activity such asthat described here for SynOMT could therefore be detrimen-tal to lignin formation. Tremendous efforts have been under-taken to manipulate the composition of lignin in various plantspecies with the focus on methylating enzymes (7, 39, 40) orthose involved in earlier biosynthetic steps, e.g. cinnamyl alco-hol dehydrogenase and cinnamoyl-CoA reductase (41). Expres-sion of enzymes targeting para-hydroxy groups in the appro-priate gymnosperm or angiosperm background may allow forthe production of novel types of lignin with altered propertiessuch as changes in the degree or mode of polymerization thatmight effect digestibility by ruminants (42).With the set of substrates tested in this system, themetabolic

function of this enzyme in cyanobacteria remains unclear. Theconserved presence of at least one gene locus in most pro-karyotes points to an important in vivo function for these orga-

nisms including Synechocystis. Sequence identities to the SafCgene product fromM. xanthusmethylating the saframycin pre-cursor L-dihydroxyphenylalanine (13) as well as to the mdmCgene in Streptomyces mycarofaciens, proposed to encode the4-O-methyltransferase for the 16-membered lactone ring of themacrolide antibiotic midecamycin (43) point to a possible rolein the biosynthesis of complex secondary metabolites. In con-trast to the bisquinone structure of saframycin, the midecamy-cin molecule gives no indication of any aromatic vicinal dihy-droxy system, therefore the proposedmethylation step remainsquestionable. SynOMT is also not part of any recognizable genecluster as is the case of S. mycarofaciens and the closest neigh-bors do not give any hint to its possible function (Fig. 1). A rolein methylation of a macrolide or an aromatic precursor of aquinoid metabolite is plausible. Macrolides like the crytophy-cins from Nostoc sp. (34) may also be present in Synechocystisand could provide an initial basis for detailed metabolite profil-ing analyses.The essential function of the slr0095 gene product might be

further resolved by knock-out mutations and homologousrecombination, feeding of labeled precursors, and moleculardocking studies. Cloning and functional expression of othermembers of the prokaryotic CCoAOMT-like proteins mayhold further surprises as far as substrate and position specific-ities toward natural compounds are concerned and theenzymes might be used as specific organic catalysts in yet diffi-cult to achieve chemical modifications.

Acknowledgments—We thank Ragai Ibrahim (Concordia University,Montreal) for the generous gift of tricetin, Norio Murata (NationalInstitute for Basic Biology, Okazaki, Japan) for the glucose-tolerantstrain of Synechocystis sp. PCC 6803, RegineHerbst-Irmer (Universityof Gottingen) for helpful comments on crystal twinning, and PiotrNeumann (Martin-Luther-University Halle-Wittenberg) for adviceduring structure solution and refinement.

REFERENCES1. Ibrahim, R. K., andMuzac, I. (2000) in Recent Advances in Phytochemistry,

Evolution of Metabolic Pathways (Romeo, J. T., Ibrahim, R. K., Varin, L.,and De Luca, V., eds) Vol. 34, pp. 349–384, Pergamon Press, Amsterdam

2. Ibdah, M., Zhang, X.-H., Schmidt, J., and Vogt, T. (2003) J. Biol. Chem.278, 43961–43972

3. Mannisto, P. T., and Kaakkola, S. (1999) Pharmacol. Rev. 51, 593–6284. Zhu, B. T. (2002) Curr. Drug Metab. 3, 321–3495. Vidgren, J. (1998) Adv. Pharmacol. 42, 328–3306. Pakusch, A. E., Kneusel, R. E., and Matern, U. (1989) Arch. Biochem. Bio-

phys. 271, 488–4947. Marita, J. M., Ralph, J., Hatfield, R. D., Guo, D., Chen, F., and Dixon, R. A.

(2003) Phytochemistry 62, 53–658. Schluckebier, G., O’Gara, M., Saenger, W., and Cheng, X. (1995) J. Mol.

Biol. 247, 16–209. Zubieta, C., Kota, P., Ferrer, J. L., Dixon, R. A., and Noel, J. P. (2002) Plant

Cell 14, 1265–127710. Ferrer, J. L., Zubieta, C., Dixon, R. A., and Noel, J. P. (2005) Plant Physiol.

137, 1009–101711. Vidgren, J., Svensson, L. A., and Liljas, A. (1994) Nature 368, 354–35812. Kopycki, J. G., Rauh, D., Chumanevich, A. A., Neumann, P., Vogt, T., and

Stubbs, M. T. (2008) J. Mol. Biol. 378, 154–16413. Nelson, J. T., Lee, J., Sims, J. W., and Schmidt, E. W. (2007) Appl. Envi-

ronm. Microbiol. 73, 3575–358014. Hou, X.,Wang, Y., Zhou, Z., Bao, S., Lin, Y., and Gong,W. (2007) J. Struct.

O-Methyltransferase from Synechocystis

JULY 25, 2008 • VOLUME 283 • NUMBER 30 JOURNAL OF BIOLOGICAL CHEMISTRY 20895

by Thom

as Vogt on A

ugust 8, 2008 w

ww

.jbc.orgD

ownloaded from

Biol. 159, 523–52815. Hagemann, M., Schoor, A., Jeanjean, R., Zuther, E., and Joset, F. (1997) J.

Bacteriol. 179, 1727–173316. Otwinowski, Z., and Minor, W. (1997)Methods Enzymol. 276, 307–32617. Brunger, A. T., Adams, P. D., Clore, G.M., Delano,W. L., Gros, P., Grosse-

Kunstleve, R. W., Jiang, J. S., Kuszewski, J., Nilges, M., Pannu, N. S., Read,R. J., Rice, L. M., Simonson, T., andWarren, G. L. (1998) Acta Crystallogr.Sect. D Biol. Crystallogr. 54, 905–921

18. McCoy, A. J., Grosse-Kunstleve, R. W., Adams, P. D., Winn, M. D., Sto-roni, L. C., and Read, R. J. (2007) J. Appl. Crystallogr. 40, 658–674

19. The CCP4 Suite (1994) Acta Crystallogr. Sect. D Biol. Crystallogr. 50,760–763

20. Jones, T. A., Zou, J. Y., Cowan, S. W., and Kjeldgaard, M. (1993) ActaCrystallogr. Sect. A 49, 148–157

21. Emsley, P., and Cowtan, K. (2004) Acta Crystallogr. Sect. D Biol. Crystal-logr. 60, 2126–2132

22. Van de Peer, Y., and De Wachter, R. (1994) Comput. Appl. Biosci. 10,569–570

23. Milkowski, C., Baumert, A., and Strack, D. (2000) Planta 211, 883–88624. Stockigt, J., and Zenk, M. H. (1975) Z. Naturforsch. 30, 352–35825. Strack, D., Keller, H., and Weissenbock, G. (1987) J. Plant Physiol. 131,

61–7326. Lineweaver, H., and Burk, D. (1934) J. Am. Chem. Soc. 56, 658–66627. Schliemann, W., Ammer, C., and Strack, D. (2008) Phytochemistry 69,

112–14628. MOE (Molecular Operating Environment) (2007) Chemical Computing

Group Inc., Montreal, Canada

29. Kanehisa, M., and Goto, S. (2000) Nucleic Acids Res. 28, 27–3030. Korb, O., Stutzle, T., and Exner, T. E. (2007) Swarm Intell. 1, 115–13431. Yoshida, T., Jin, Z.-X., andOkua, T. (1991) Phytochemistry 39, 2247–275232. Niemetz, R., and Gross, G. G. (2005) Phytochemistry 66, 2001–201133. Eggen, M., and Georg, G. I. (2002)Med. Res. Rev. 22, 85–10134. Vogt, T. (2004) FEBS Lett. 561, 159–16235. Wollenweber, E., and Jay,M. (1988) inThe Flavonoids (Harborne, J. B., ed)

pp. 283–302, Chapman and Hall, London36. Schroder, G., Wehinger, E., Lukacin, R., Wellmann, F., Seefelder, Schwab,

W., and Schroder, J. (2004) Phytochemistry 65, 1085–109437. Zhou, J.M., Gold, N.M.,Martin, V. J.,Wollenweber, E., and Ibrahim, R. K.

(2006) Biochim. Biophys. Acta 1760, 1115–112438. Morreel, K., Ralph, J., Kim,H., Lu, F., Goeminne, G., Ralph, S.,Messens, E.,

and Boerjan, W. (2004) Plant Physiol. 136, 3537–354939. Anterola, A. M., and Lewis, N. G. (2002) Phytochemistry 61, 221–29440. Chen, F., Reddy,M. S., Temple, S., Jackson, L., Shadle, G., and Dixon, R. A.

(2006) Plant J. 48, 113–12441. Chabannes,M., Barakate, A., Lapierre, C.,Marita, J.M., Ralph, J., Pean,M.,

Danoun, S., Halpin, C., Grima-Pettenati, J., and Boudet, A.M. (2001)PlantJ. 28, 257–270

42. Guo, D., Chen, F., Wheeler, J., Winder, J., Selman, S., Peterson, M., andDixon, R. A. (2001) Transgenic Res. 10, 457–464

43. Hara, O., and Hutchinson, C. R. (1992) J. Bacteriol. 174, 5141–514444. Kaneko, T., Sato, S., Kotani, H., Tanaka, A., Asamizu, E., Nakamura, Y.,

Miyajima, N., Hirosawa, M., Suigiura, M., Sasamoto, S., Kimura, T., Ho-souchi, T.,Matsuno, A.,Muraki, A., Nakazaki, N., Naruo, K., Okumura, S.,Shimpo, S., Takeuchi, C., and Tabata, S. (1996) DNA Res. 3, 109–136

O-Methyltransferase from Synechocystis

20896 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 283 • NUMBER 30 • JULY 25, 2008

by Thom

as Vogt on A

ugust 8, 2008 w

ww

.jbc.orgD

ownloaded from