Artigo Acesso aberto Revisado por pares

7-Methylxanthine Methyltransferase of Coffee Plants

2001; Elsevier BV; Volume: 276; Issue: 11 Linguagem: Inglês

10.1074/jbc.m009480200

ISSN

1083-351X

Autores

Mikihiro Ogawa, Yuka Herai, Nozomu Koizumi, Tomonobu Kusano, Hiroshi Sano,

Tópico(s)

Coffee research and impacts

Resumo

Caffeine is synthesized through sequential three-step methylation of xanthine derivatives at positions 7-N, 3-N, and 1-N. However, controversy exists as to the number and properties of the methyltransferases involved. Using primers designed on the basis of conserved amino acid regions of tea caffeine synthase andArabidopsis hypothetical proteins, a particular DNA fragment was amplified from an mRNA population of coffee plants. Subsequently, this fragment was used as a probe, and four independent clones were isolated from a cDNA library derived from coffee young leaves. Upon expression in Escherichia coli, one of them was found to encode a protein possessing 7-methylxanthine methyltransferase activity and was designated as CaMXMT. It consists of 378 amino acids with a relative molecular mass of 42.7 kDa and shows similarity to tea caffeine synthase (35.8%) and salicylic acid methyltransferase (34.1%). The bacterially expressed protein exhibited an optimal pH for activity ranging between 7 and 9 and methylated almost exclusively 7-methylxanthine with low activity toward paraxanthine, indicating a strict substrate specificity regarding the 3-N position of the purine ring. K mvalues were estimated to be 50 and 12 μm for 7-methylxanthine and S-adenosyl-l-methionine, respectively. Transcripts of CaMXMT could be shown to accumulate in young leaves and stems containing buds, and green fluorescent protein fusion protein assays indicated localization in cytoplasmic fractions. The results suggest that, in coffee plants, caffeine is synthesized through three independent methylation steps from xanthosine, in which CaMXMT catalyzes the second step to produce theobromineAB039725AB048792AB048793AB048794. Caffeine is synthesized through sequential three-step methylation of xanthine derivatives at positions 7-N, 3-N, and 1-N. However, controversy exists as to the number and properties of the methyltransferases involved. Using primers designed on the basis of conserved amino acid regions of tea caffeine synthase andArabidopsis hypothetical proteins, a particular DNA fragment was amplified from an mRNA population of coffee plants. Subsequently, this fragment was used as a probe, and four independent clones were isolated from a cDNA library derived from coffee young leaves. Upon expression in Escherichia coli, one of them was found to encode a protein possessing 7-methylxanthine methyltransferase activity and was designated as CaMXMT. It consists of 378 amino acids with a relative molecular mass of 42.7 kDa and shows similarity to tea caffeine synthase (35.8%) and salicylic acid methyltransferase (34.1%). The bacterially expressed protein exhibited an optimal pH for activity ranging between 7 and 9 and methylated almost exclusively 7-methylxanthine with low activity toward paraxanthine, indicating a strict substrate specificity regarding the 3-N position of the purine ring. K mvalues were estimated to be 50 and 12 μm for 7-methylxanthine and S-adenosyl-l-methionine, respectively. Transcripts of CaMXMT could be shown to accumulate in young leaves and stems containing buds, and green fluorescent protein fusion protein assays indicated localization in cytoplasmic fractions. The results suggest that, in coffee plants, caffeine is synthesized through three independent methylation steps from xanthosine, in which CaMXMT catalyzes the second step to produce theobromineAB039725AB048792AB048793AB048794. 7-methylxanthine S-adenosyl-l-methionine caffeine synthase green fluorescent protein xanthosine 7-N-methyltransferase polymerase chain reaction glutathione S-transferase high performance liquid chromatography 4-morpholineethanesulfonic acid Among more than 50,000 secondary metabolites of plants, 12,000 are alkaloids. Their physiological roles are considered to be chemical defense against invertebrate herbivores. Caffeine, a typical purine alkaloid, is found in seeds and leaves of coffee (Coffea arabica), cola (Cola nitida), maté (Ilex paraguariensis), and tea (Camellia sinensis) at concentrations up to 1 mg/1 g, dry weight (1Roberts M.F. Wink M. Alkaloids: Biochemistry, Ecology, and Medical Applications. Plenum Press, New York1998: 11Crossref Google Scholar, 2Baumann T.W. Koetz R. Morath P. Plant Cell Rep. 1983; 2: 33-35PubMed Google Scholar). It exhibits a lethal effect on tobacco horn worm (Manduca sexta) by inhibiting phosphodiesterase activity, which hydrolyzes cAMP (3Croteau R. Kutchan T.M. Lews N.G. Buchanan B.B. Gruissem W. Jones R.L. Biochemistry and Molecular Biology of Plants. American Society of Plant Physiologists, Rockville, MD2000: 1250-1318Google Scholar). The biosynthetic pathway of caffeine has been intensively studied, and it is now established that it is successively synthesized from adenine nucleotides through multiple steps catalyzed by several enzymes (4Ashihara H. Kato M. Chuang-xing Y. J. Plant Res. 1998; 111: 599-604Crossref Google Scholar, 5Ashihara H. Monteiro A.M. Gillies F.M. Crozier A. Plant Physiol. 1996; 111: 747-753Crossref PubMed Scopus (109) Google Scholar, 6Ashihara H. Crozier A. Adv. Bot. Res. 1999; 30: 117-205Crossref Scopus (126) Google Scholar). The final series of steps involves methylation of xanthosine byN-methyltransferase, yielding 7-methylxanthosine, whose ribose residue is removed by 7-methylxanthosine nucleosidase. The resulting 7-methylxanthine (7mX)1 is methylated at the 3-N-position by N-methyltransferase, producing 3,7-dimethylxanthine (theobromine), which is again methylated at the 1-N-position to give 1,3,7-trimethylxanthine (caffeine) (Fig. 1). All reactions requireS-adenosyl-l-methionine (AdoMet) as a methyl donor. Some bypass pathways, for example featuring paraxanthine, have also been suggested, but in coffee and tea plants, it was confirmed that the major pathway is through theobromine (5Ashihara H. Monteiro A.M. Gillies F.M. Crozier A. Plant Physiol. 1996; 111: 747-753Crossref PubMed Scopus (109) Google Scholar, 6Ashihara H. Crozier A. Adv. Bot. Res. 1999; 30: 117-205Crossref Scopus (126) Google Scholar). At least three N-methyltransferases are considered to contribute to this pathway; these catalyze methylation of xanthosine (the first), methylation of 7mX (the second), and methylation of theobromine (the third). Their isolation and characterization have attracted a good deal of attention, and enzymes catalyzing the second and the third steps were first identified in crude extract of tea leaves (7Suzuki T. Takahashi E. Biochem. J. 1975; 146: 87-96Crossref PubMed Scopus (68) Google Scholar). Since then a dozen surveys describing their purification and characterization in coffee and tea plants have been published (2Baumann T.W. Koetz R. Morath P. Plant Cell Rep. 1983; 2: 33-35PubMed Google Scholar,8Roberts M.F. Waller G.R. Phytochemistry. 1979; 18: 451-455Crossref Scopus (51) Google Scholar, 9Negishi O. Ozawa T. Imagawa H. Agric. Biol. Chem. 1985; 49: 887-890Crossref Scopus (1) Google Scholar, 10Mozzafera P. Wingsle G. Olsson O. Sandberg G. Phytochemistry. 1994; 37: 1577-1584Crossref Scopus (37) Google Scholar, 11Kato M. Kanehara T. Shimizu H. Suzuki T. Gillies F.M. Crozier A. Ashihara H. Physiol. Plantarum. 1996; 98: 629-636Crossref Scopus (6) Google Scholar, 12Mosli Waldhauser S.S. Gillies F.M. Crozier A. Baumann T.W. Phytochemistry. 1997; 45: 1407-1414Crossref PubMed Scopus (33) Google Scholar, 13Mosli Waldhauser S.S. Kretschmar J.A. Baumann T.W. Phytochemistry. 1997; 44: 853-859Crossref Scopus (26) Google Scholar). However, it was found that the enzymes are extremely labile, making it difficult even to distinguish each activity. Indeed, it is not clear yet whether the activities are catalyzed by independent or multifunctional proteins (2Baumann T.W. Koetz R. Morath P. Plant Cell Rep. 1983; 2: 33-35PubMed Google Scholar, 12Mosli Waldhauser S.S. Gillies F.M. Crozier A. Baumann T.W. Phytochemistry. 1997; 45: 1407-1414Crossref PubMed Scopus (33) Google Scholar). Despite such difficulties, a caffeine synthase (CS) was recently isolated successfully from tea leaves (14Kato M. Mizuno K. Fujimura T. Iwama M. Irie M. Crozier A. Ashihara H. Plant Physiol. 1999; 120: 579-586Crossref PubMed Scopus (95) Google Scholar). The enzyme has a native molecular mass of 61 kDa and exhibits 3- and 1-N-methyltransferase activities toward substrates such as 7mX, theobromine, and paraxanthine (14Kato M. Mizuno K. Fujimura T. Iwama M. Irie M. Crozier A. Ashihara H. Plant Physiol. 1999; 120: 579-586Crossref PubMed Scopus (95) Google Scholar). It was thus concluded that, at least in tea leaves, a single enzyme has dual functions in caffeine synthesis. Subsequently, the gene encoding this CS was isolated (TCS1), and the predicted amino acid sequence was found to show considerable similarity with salicylic acidO-methyltransferase (15Kato M. Mizuno K. Crozier A. Fujimura T. Ashihara H. Nature. 2000; 406: 956-957Crossref PubMed Scopus (163) Google Scholar). Whether or not a similar enzyme(s) functions in coffee plants has not been hitherto determined. Although a coffee gene encoding xanthosine methyltransferase (XMT), was reported in a patent (16Stiles, J. I., and Moisyadi, I. (February 10, 1997) International Patent WO 97/35960.Google Scholar), the details remain to be clarified. In this work, we document isolation of a gene encoding an enzyme that catalyzes methylation of 7mX from coffee plants. In contrast to tea CS, the enzyme features strict substrate specificity toward methylation only at the 3-N-position of the purine ring. It is suggested that, in coffee plants, caffeine synthesis is mediated by three methylation steps catalyzed by distinct enzymes, including the presently identified 7mX methyltransferase. Coffee plants (C. arabica L. var. caturra) were cultivated in a greenhouse. Two degenerated oligonucleotides, 5′-GGITGYDSIDSIGGICCIAAYAC-3′ (forward) and 5′-ARIYKIYYRTRRAAISWICCIGG-3′ (reverse), which correspond to the amino acid sequences of GC(A/S)(A/S)GPNT and PGSF(H/Y)(G/K)(R/N)LF, respectively, were synthesized based on conserved regions among TCS1 (Ref. 15Kato M. Mizuno K. Crozier A. Fujimura T. Ashihara H. Nature. 2000; 406: 956-957Crossref PubMed Scopus (163) Google Scholar; accession number AB031280) and two Arabidopsishypothetical proteins (Z99708 and AC008153). PCR was performed in 25 μl of reaction mixture containing C. arabica cDNA and the pair of primers mentioned above under the conditions of 94 °C for 1 min, 30 cycles of denaturation at 94 °C for 30 s, annealing at 52 °C for 30 s, and extension at 72 °C for 1 min, followed by a final extension at 72 °C for 7 min. A 255-base pair fragment was amplified, and one of the deduced amino acid sequences from its DNA sequence showed 34% identity to that of TCS1. This fragment was used to screen the C. arabica cDNA library. Total RNA was extracted by the cetyltrimethylammonium bromide method (17Hatanaka T. Choi Y.E. Kusano T. Sano H. Plant Cell Rep. 1999; 19: 106-110Crossref PubMed Scopus (45) Google Scholar) with a slight modification, and poly(A+) RNA was purified using an mRNA purification kit (Amersham Pharmacia Biotech) according to the manufacturer's instructions and converted into double-stranded cDNA using a ZAPII cDNA synthesis kit (Stratagene). The cDNA was ligated with Uni-ZAP XR vector arms and packaged using a Gigapack III kit. The titer of the library was 3 × 107 plaque-forming units. The open reading frame regions of clones 1, 6, 35, and 45 sandwiched with SmaI and NotI restriction sites were subcloned into the pGEX 4T-2 vector (PharmaciaPhandEscherichia coli JM109 cells were transformed with the resulting plasmids. When the A 600 of theE. coli cell culture reached 0.5, 1 mmisopropyl-1-thio-β-galactoside was added for production of GST fusion proteins followed by further incubation at 18 °C for 6 h. The bacterial cells were collected by centrifugation, resuspended in a sonication buffer, and disrupted by a sonicator. Fusion proteins were purified from the clear lysates as described earlier (18Ikeda Y. Koizumi N. Kusano T. Sano H. Plant Physiol. 1999; 121: 813-820Crossref PubMed Scopus (55) Google Scholar). The enzyme activity was determined by an established procedure (14Kato M. Mizuno K. Fujimura T. Iwama M. Irie M. Crozier A. Ashihara H. Plant Physiol. 1999; 120: 579-586Crossref PubMed Scopus (95) Google Scholar) with a slight modification. The reaction mixture of 100 μl containing 100 mm Tris-HCl (pH 8.3), 200 μm substrate, 4 μm S-adenosyl-l-[methyl-14C]methionine (2.15 GBq/mmol; Amersham Pharmacia Biotech), 200 μmMgCl2, and 200 ng of purified recombinant protein was incubated at 27 °C for 2 h. The reaction was terminated by the addition of 1 ml of chloroform, and the organic phase was recovered, dried at 60 °C, and dissolved in 10 μl of 50% methanol. This fraction was separated by thin layer (Silica gel 60 F254; Merck) chromatography with a solution of H2O/acetic acid/n-butyl alcohol (2:1:4, v/v/v). Radioactive images were detected with an image analyzer (Fuji BAS2000). Data from at least three replicate experiments in each case were pooled and analyzed by nonlinear least squares regression fitting to the Hill equation (Equation 1) with the Anemona program (19Hernandez A. Ruiz M.T. Bioinformatics. 1998; 14: 227-228Crossref PubMed Scopus (78) Google Scholar). v=Vmax[S]h/(K+[S]h)Equation 1 where v is the rate of reaction (rate of product formation), Vmax is maximum rate, Kis the rate constant, [S] is the substrate concentration, andh is the Hill number. A reaction mixture of 100 μl containing 100 mm Tris-HCl (pH 7.5), 200 μm substrate, 50 μm AdoMet, 200 μm MgCl2, and 200 ng of purified GST fusion protein was incubated at 27 °C for 2 h and extracted with 1 ml of chloroform. The chloroform phase was dried, resolved in 12% acetonitrile, and separated by HPLC using a column (Shodex Rspak DS-613, Showa Denko) with a flow rate of 1 ml/min of 12% acetonitrile and then monitored for absorbance at 254 nm. Total RNAs were isolated from various C. arabica tissues and reverse-transcribed by SuperScript II (Life Technologies, Inc.). The first-strand cDNAs were used as a template for reverse transcription-PCR analysis, performed as follows: 96 °C for 20 s; 30 cycles of 96 °C for 20 s, 60 °C (55 °C in case of XMT) for 30 s, and 72 °C for 30 s; followed by further extension at 72 °C for 7 min. The primers used were CaMXMT-Fw (5′-CCAGTAAGATCCCATGAACAAAT-3′), CaMXMT-RV (5′-TTATTACGAATACAAAACGACAATACC-3′), XMT-Fw (5′-AGCACATTCGGACTCTCCAG-3′), XMT-RV (5′-TACCGAGTTAAGCGATGCAC-3′), CaMTL1/2-Fw (5′-CCATTCCCCAGAATACAGCG-3′), CaMTL1/2-RV (5′-CCCCGTATCAGAAAACAAACC-3′), CaMTL3-Fw (5′-GGCTTCTCTATTGACGATGAACATAT-3′), and CaMTL3-RV (5′-CACTTATTCCTTTCCCCAACAC-3′). The CaMXMT-entire coding region fragments sandwiched with XbaI and KpnI sites were subcloned into pGFP2 (provided by Drs. Chua and Spielhofer), resulting in pCaMXMT::GFP. Thin sections of onion bulbs cut into 9-cm2 squares were biolistically bombarded as described (20Hara K. Yagi M. Kusano T. Sano H. Mol. Gen. Genet. 2000; 263: 30-37Crossref PubMed Scopus (189) Google Scholar), with gold particles (Bio-Rad) coated with the plasmids pGFP2, pCaMXMT::GFP. After bombardment, they were incubated for 12 h at 25 °C in darkness and then viewed using epifluorescence microscopy (20Hara K. Yagi M. Kusano T. Sano H. Mol. Gen. Genet. 2000; 263: 30-37Crossref PubMed Scopus (189) Google Scholar). All chemicals were purchased from Sigma unless otherwise described.S-Adenosyl-l-[methyl-14C]methionine (2.15 GBq/mmol) was purchased from Amersham Pharmacia Biotech. To isolate genes for caffeine synthase of coffee plants, a 255-base pair fragment amplified by PCR with degenerated primers was used for screening a phage library. A total of 35 randomly selected plaques hybridized to the probe were converted into phagemids. They were classified into four groups by physical mapping and by partial DNA sequencing. The longest cDNAs of each group, clones 1, 6, 35, and 45 were selected, their DNA sequences were determined, and the deduced products were aligned (Fig.2). Pairwise identities between clone 45 product and those of clones 1, 6, and 35 were 80.8, 81.3, and 84.7%, respectively. Clones 1 and 6 showed 95.8% identity with each other. The GST fusion proteins of clones 6, 35, and 45 were produced in E. coli and purified on a glutathione-Sepharose column (Fig.3 A), andN-methyltransferase activity was assayed. The product of clone 45 catalyzed conversion of 7mX to theobromine and that of paraxanthine to caffeine (Fig. 3 B). Identification of the product as theobromine was performed by high performance liquid chromatography (Fig. 3 C). The protein catalyzes methylation either of 7mX or of paraxanthine at the 3-N-position and has a 5-fold preference for 7mX as opposed to paraxanthine as the substrate (Table I). Substrate specificity of the clone 45 product is distinct from that of tea CS, which prefers paraxanthine to 7mX (14Kato M. Mizuno K. Fujimura T. Iwama M. Irie M. Crozier A. Ashihara H. Plant Physiol. 1999; 120: 579-586Crossref PubMed Scopus (95) Google Scholar). The cDNA of clone 45 was, therefore, designated as CaMXMT (C. arabica7-methylxanthine methyltransferase). The deduced amino acid sequence showed identity to TCS1 of 35.8%, to salicylic acid methyltransferase of 34.1%, and to benzoic acid carboxyl methyltransferase of 34.2% (Fig. 4). Although the products of clones 1, 6, and 35 showed high similarity to CaMXMT, they had no methyltransferase activity for the substrates tested (data not shown). These clones were designated asCaMTL1 (C.a rabicamethyltransferase-like 1, clone 1),CaMTL2 (clone 6), and CaMTL3 (clone 35), respectively.Table ISubstrate specificity of CaMXMTMaterialSubstrateSources7mX3mX1mXTbTpPxXXRXMPRecombinant CaMXMT (coffee)100NDNDNDND15NDNDNDThis studyCrude enzyme (coffee fruits)1005.71274.6175NDNDRef.8Roberts M.F. Waller G.R. Phytochemistry. 1979; 18: 451-455Crossref Scopus (51) Google ScholarNative TCS1 (tea)10017.64.226.8tr210trNDNDRef.14Kato M. Mizuno K. Fujimura T. Iwama M. Irie M. Crozier A. Ashihara H. Plant Physiol. 1999; 120: 579-586Crossref PubMed Scopus (95) Google ScholarRecombinant TCS1 (tea)1001.012.318.5<0.1230NDRef.15Kato M. Mizuno K. Crozier A. Fujimura T. Ashihara H. Nature. 2000; 406: 956-957Crossref PubMed Scopus (163) Google ScholarRelative enzyme activities of CaMXMT, a crude extract from coffee fruits, and native and recombinant caffeine synthase (TCS1) from tea were compared. Activity of each towards 7mX was set as 100, and their relative activities are shown. 7mX, 3mX, and 1mX, 7-, 3-, and 1-methylxanthine, respectively; Tb, theobromine; Tp, theopylline; Px, paraxanthine; X, xanthosine; XR, xanthine riboside; XMP, xanthosine monophosphate; ND, not detected; tr, trace. Open table in a new tab Figure 4Amino acid alignment of CaMXMT and related methyltransferases from higher plants. A, amino acid alignment: CaMXMT (this paper), TCS1 (15Kato M. Mizuno K. Crozier A. Fujimura T. Ashihara H. Nature. 2000; 406: 956-957Crossref PubMed Scopus (163) Google Scholar), salicylic acid methyltransferase (SAMT) (AAF00108), and benzoic acid carboxyl methyltransferase (BAMT) (AAF98284). Conserved residues in three out of four sequences are boxed. Motifs I and III are supposed to be AdoMet-binding sites (20Hara K. Yagi M. Kusano T. Sano H. Mol. Gen. Genet. 2000; 263: 30-37Crossref PubMed Scopus (189) Google Scholar). B, phylogenetic relationships among the enzymes.View Large Image Figure ViewerDownload Hi-res image Download (PPT) Relative enzyme activities of CaMXMT, a crude extract from coffee fruits, and native and recombinant caffeine synthase (TCS1) from tea were compared. Activity of each towards 7mX was set as 100, and their relative activities are shown. 7mX, 3mX, and 1mX, 7-, 3-, and 1-methylxanthine, respectively; Tb, theobromine; Tp, theopylline; Px, paraxanthine; X, xanthosine; XR, xanthine riboside; XMP, xanthosine monophosphate; ND, not detected; tr, trace. The optimal pH for 7mX methyltransferase activity of CaMXMT ranged between 7 and 9, with the peak at 7.5 (Fig. 5 A). The effects of 7mX and AdoMet concentrations on the reaction velocity of GST-CaMXMT protein were determined (Fig. 5 B). TheK m values for 7mX and AdoMet were 50 and 11.9 μm, respectively, and apparentVmax values were estimated to be 7.14 and 7.94 pmol of theobromine/min/μg of protein upon measurement with the variable amounts of 7mX and AdoMet, respectively. Accumulation of CaMXMTtranscripts was estimated by reverse transcription-PCR together withCaMTL1, CaMTL2, and CaMT3 in various tissues including roots, stems containing buds, old leaves, and young leaves of C. arabica (Fig.6 A). The level of transcripts for XMT, which catalyzes the conversion of xanthosine to 7-methylxanthosine, was also tested. Transcripts of CaMXMTwere detected in stems and young leaves but not in roots and old leaves, similar to the expression pattern for XMT. Transcripts of CaMTL1 and CaMTL2 were present in all tissues at high levels, whereas CaMTL3 transcripts were abundant in stems and young leaves and also in roots and old leaves at a lower level. To identify the cellular localization of CaMXMT, the cDNA fragment covering the entire coding region of CaMXMT was fused to pGFP2, and the resulting plasmid was introduced into the onion epidermal layer by a biolistic bombardment. Green fluorescence was detected in the cytoplasm (Fig. 6 B). This report documents isolation of a gene encoding 7mX methyltransferase from coffee plants and characterization of the bacterially expressed recombinant enzyme. Screening a coffee cDNA library with a probe constructed from a conserved amino acid region of TCS1 and similar sequences derived from Arabidopsisexpressed sequence tag clones, four distinct cDNA clones were isolated. The protein encoded by one of them showed 7mX methyltransferase activity when expressed as a fusion protein with GST in E. coli and was designated as CaMXMT. Proteins encoded by other clones (CaMTL1, -2, and -3) did not show any methyltransferase activity on substrates examined for CaMXMT. The deduced amino acid sequence of CaMXMT showed rather low similarity to TCS1 (35.8%) and high similarity to CaMTLs (more than 80%) (Fig. 4 B). This result indicates CaMXMT is not close in an evolutionary sense to TCS1. In other words, caffeine biosynthetic pathway in coffee and tea might have evolved independently, consistent with different catalytic properties of the enzymes involved (see below). CaMXMT also showed low similarity to salicylic acidO-methytransferase from Clarkia breweri (21Ross J.R. Nam K.H. Dı́Auria J.C. Pichersky E. Arch. Biochem. Biophys. 1999; 367: 9-16Crossref PubMed Scopus (203) Google Scholar) and bezoic acid carboxyl methyltransferase isolated from snapdragon (Antirrhinum sp.) flowers (22Dudareva N. Murfitt L.M. Mann C.J. Gorestein N. Kolosava N. Kish C.M. Bonham C. Wood K. Plant Cell. 2000; 12: 949-961Crossref PubMed Scopus (260) Google Scholar). In addition, we found several related sequences in the expressed sequence tag ofArabidopsis. Although more than 120 methyltransferases have so far been reported from various organisms (23Cheng X. Blumenthal R.M. S-Adenosylmethionine-Dependent Methyltransferases: Structure and Functions. World Scientific, Singapore1999: 3Google Scholar), methyltransferases of this type are not well characterized. Their structures appear to be unique, with little similarities to other methyltransferases, suggesting a new class. However, it has been pointed out that salicylic acid methyltransferase contains domains similar to motifs I and III found in plant O-methyltransferases (21Ross J.R. Nam K.H. Dı́Auria J.C. Pichersky E. Arch. Biochem. Biophys. 1999; 367: 9-16Crossref PubMed Scopus (203) Google Scholar). Those are proposed to be involved in AdoMet binding and conserved in salicylic acid methyltransferase, benzoic acid carboxyl methyltransferase, TCS, and CaMXMT (Fig. 4 A), although TCS1 and CaMXMT areN-methyltransferases. The motifs are also found in CaMTLs, making it highly probable that they possess methyltransferase activity, although they do not participate in caffeine biosynthesis. The major difference in amino acid sequence between CaMXMT and CaMTLs is Val159-His160-Tyr161 (VHW), which is present in TCS1 and CaMXMT but absent in CaMTLs. It is tempting to speculate that substrate specificity of this class is determined by a few particular amino acids, and further investigations with point-mutated proteins are needed to clarify this point. Despite the similar pH optimum for activity, the substrate specificities of CaMXMT and TCS1 are clearly different. Whereas both native and recombinant TCS1 equally show catalytic activity toward the 1-N- and 3-N-sites of the purine ring, CaMXMT catalyzes only 3-N-methylation (Table I). In a crude extract of coffee fruits, the capacity of 1-N-methylation of theobromine to caffeine was detected (8Roberts M.F. Waller G.R. Phytochemistry. 1979; 18: 451-455Crossref Scopus (51) Google Scholar), and we have confirmed this with crude extracts of young leaves. 2M. Ogawa, Y. Herai, N. Koizumi, T. Kusano, and H. Sano, unpublished observation. Since recombinant CaMXMT did not show any 1-N-methylation activity, it is obvious that, in coffee plants, 3-N- and 1-N-methylation is catalyzed by different enzymes. This is consistent with findings that the apparent K m for xanthine derivatives markedly differs among enzymes. Crude enzymes exhibit K m values for both 7-methylxanthine and theobromine ranging between 100 and 500 μm (13Mosli Waldhauser S.S. Kretschmar J.A. Baumann T.W. Phytochemistry. 1997; 44: 853-859Crossref Scopus (26) Google Scholar). This is also the case for purified tea CS, except that it has much higher affinity for paraxanthine, with a K m of 24 μm (14Kato M. Mizuno K. Fujimura T. Iwama M. Irie M. Crozier A. Ashihara H. Plant Physiol. 1999; 120: 579-586Crossref PubMed Scopus (95) Google Scholar). Such differential K m values suggest that, despite apparent multifunctional properties, each enzyme may be able to select its correct substrate. CaMXMT methylates predominantly 7mX with a K m of 50 μm, a much higher affinity than for any other enzymes reported. The observations suggest that enzymes involved in caffeine synthesis may possess rather strict substrate preference and that this arises from diversity in a few amino acids. The transcript accumulation profiles of CaMXMT,XMT, and CaMTLs were analyzed by reverse transcription-PCR with specific primers for each to avoid cross-hybridization between CaMXMT and CaMTLs. Transcripts of CaMXMT and XMT accumulated in young leaves and stems containing buds, suggesting that biosynthesis of caffeine occurs mainly in those tissues in coffee plants. This is consistent with the fact that theobromine and caffeine are primarily found in their buds and young leaves (5Ashihara H. Monteiro A.M. Gillies F.M. Crozier A. Plant Physiol. 1996; 111: 747-753Crossref PubMed Scopus (109) Google Scholar). It should be noted that the transcript accumulation profile of CaMTL3 is similar to that of CaMXMT and XMP, suggesting its involvement in the metabolism of caffeine-related compounds. Examination of the subcellular localization of CaMXMT using the fusion protein of CaMXMT and GFP demonstrated an existence predominantly in the cytoplasm of onion epidermal cells. The PSORT program with the deduced amino acid sequence also predicted a high possibility of cytoplasmic localization for CaMXMT.2 It can thus be concluded that caffeine biosynthesis occurs in the cytoplasm of cells in buds and young leaves. It is worthy of mention that CaMXMT may have practical applications. To cope with occasional health problems caused by caffeine, decaffeinated coffee is currently produced by chemical treatment of coffee beans. Recombinant DNA technology usingCaMXMT may remove the need for this by creating caffeineless coffee plants. Furthermore, the opposite approach may also be applicable to important crops in such a way as to produce caffeine derivatives as insect repellants. We thank Drs. H. Ashihara, M. Kato (Ochanomizu University) and T. Fujimura (Tsukuba University) for providing the plasmid pTCS1. We are also grateful to Drs. N.-H. Chua (The Rockefeller University) and P. Spielhofer (Berne University) for supplying the plasmid pGFP2. We also thank Dr. M. Moore (Intermal) for critical reading of the manuscript.

Referência(s)
Altmetric
PlumX