Molecular cloning and functional expression ofO-methyltransferases common to isoquinoline alkaloid and phenylpropanoid biosynthesis


*For correspondence (fax +49 89 5902611;


In cell suspension cultures of the meadow rueThalictrum tuberosum, biosynthesis of the anti-microbial alkaloid berberine can be induced by addition of methyl jasmonate to the culture medium. The activities of the four methyltransferases involved in the formation of berberine from L-tyrosine are increased in response to elicitor addition. Partial clones generated by RT–PCR with methyltransferase-specific primers were used as hybridization probes to isolate four cDNAs encodingO-methyltransferases from a cDNA library prepared from poly(A)+ RNA isolated from methyl jasmonate-induced cell suspension cultures ofT. tuberosum. RNA gel blot hybridization indicated that the transcripts for the methyltransferases accumulated in response to addition of methyl jasmonate to the cell culture medium. The cDNAs were functionally expressed inSpodoptera frugiperdaSf9 cells and were shown to have varying and broad substrate specificities. A difference of a single amino acid residue between two of the enzymes was sufficient to alter the substrate specificity. The four cDNAs were expressed either as four homodimers or as six heterodimers by co-infection with all possible combinations of the four recombinant baculoviruses. These 10 isoforms thus produced displayed distinct substrate specificities and in some cases co-infection with two different recombinant baculoviruses led to theO-methylation of new substrates. The substrates that wereO-methylated varied in structural complexity from simple catechols to phenylpropanoids, tetrahydrobenzylisoquinoline, protoberberine and tetrahydrophenethylisoquinoline alkaloids, suggesting that some biosynthetic enzymes may be common to both phenylpropanoid and alkaloid anabolism.


The biosynthesis of the anti-microbial alkaloid berberine in plant cell culture has been well investigated at the enzyme level ( Kutchan 1998;Rueffer & Zenk 1994). Thirteen enzymes are necessary for the conversion of two molecules of the primary metabolite l-tyrosine to one molecule of berberine. In cell suspension cultures of the meadow rue Thalictrum tuberosum, berberine biosynthesis is induced by the addition of 50 μm methyl jasmonate to the culture medium. The increase in biosynthesis is due to increases in the activities of four methyltransferases that lie along the biosynthetic pathway to berberine: norcoclaurine 6-O-methyltransferase [EC2.1.1.128] ( Rueffer et al. 1983 ;Sato et al. 1994 ), (S)-coclaurine N-methyltransferase [EC2.1.1.115] ( Frenzel & Zenk 1990a), 3′-hydroxy-N-(S)-methylcoclaurine 4′-O-methyltransferase [EC2.1.1.116] ( Frenzel & Zenk 1990b) and (S)-scoulerine 9-O-methyltransferase [EC2.1.1.117] ( Muemmler et al. 1985 ;Galneder & Zenk 1990) ( Fig. 1). This regulation of the genes encoding methyltransferases in various berberine-producing species is in contrast to the regulation of benzo[c]phenanthridine alkaloid biosynthesis in cell cultures such as those from the California poppy Eschscholzia californica ( Kutchan & Zenk 1993). In E. californica cell cultures, genes encoding the oxidoreductases of macarpine biosynthesis, such as the berberine bridge enyzme and the cytochromes P-450, are transcriptionally activated in response to elicitor treatment, but the methyltransferase enzyme activities remain at a constant level ( Dittrich & Kutchan 1991;Haider et al. 1997 ;Kutchan 1993;Pauli & Kutchan 1998). Since simple analogies concerning biogenic regulation apparently cannot be drawn between the different classes of isoquinoline alkaloids, a better understanding of the regulation of berberine alkaloid biosynthesis requires isolation of the genes encoding the methyltransferases.

Figure 1.

Biosynthesis of berberine.

This proposed biosynthetic pathway that leads to the alkaloid berberine in T. tuberosum displays the positions of the four methyltransferases that were a target of this study. CYP80B1, the cytochrome P-450-dependent monooxygenase (S)-N-methylcoclaurine 3′-hydroxylase ( Pauli & Kutchan 1998); BBE, the berberine bridge enzyme ( Dittrich & Kutchan 1991).

The first plant methyltransferase for which a cDNA was isolated was caffeoyl-CoA 3-O-methyltransferase, an enzyme involved in the disease resistance response in parsley cell cultures ( Schmitt et al. 1991 ). Since then, the genes encoding multiple plant O-methyltransferases have been identified, mostly from phenylpropanoid metabolism (reviewed in Ibrahim et al. 1998 ). Comparison of the amino acid sequences derived from translation of the nucleotide sequences reveals up to five conserved regions at the C-termini of the polypeptides. These consensus sequences should allow for the direct cloning of methyltransferase-encoding plant cDNAs in a manner analogous to the approach that has been so successful with cytochromes P-450 ( Holton & Lester 1996).

Functional heterologous expression of plant methyltransferases has been achieved in Escherichia coli for cDNAs of phenylpropanoid and of alkaloid biosynthesis. When monitored, the substrate specificity of the recombinant enzymes was found to be high, as for caffeoyl-CoA 3-O-methyltransferase of tobacco which methylates caffeoyl-CoA and 5-hydroxyferuloyl-CoA, but not the corresponding free acids ( Martz et al. 1998 ). A broader substrate specificity was observed for a flavonoid 7-O-methyltransferase from barley leaves expressed in E. coli which methylated apigenin, naringenin, luteolin, kaempherol, eriodictyol and quercetin ( Christensen et al. 1998 ). The two alkaloid biosynthetic methyltransferase cDNAs isolated to date, putrescine N-methyltransferase of nicotine biosynthesis in tobacco ( Hibi et al. 1994 ) and scoulerine 9-O-methyltransferase of berberine biosynthesis in Coptis japonica ( Takeshita et al. 1995 ), were functionally expressed, but the substrate specificities were not reported.

The current work was undertaken in order to establish direct molecular genetic techniques for the isolation of methyltransferases of alkaloid biosynthesis that would bypass lengthy enzyme isolation procedures and thereby allow alkaloid biosynthetic genes to be more quickly isolated and identified. The methyltransferases of berberine biosynthesis in T. tuberosum were chosen due to their important role in the regulation of alkaloid biosynthesis in that species. In addition, a facile heterologous expression system was established in Spodoptera frugiperda Sf9 cells using a baculovirus expression vector for the functional over-expression of alkaloid biosynthetic methyltransferases such that an analysis of the physical characteristics and substrate specificities of these enzymes could be undertaken.

From the results presented herein, it has been determined that the plant cell has a tremendous capacity to O-methylate low-molecular-weight substances that contain a catechol moiety. Multiple O-methyltransferase isoforms are present that can combine, at least in a heterologous system, either as homodimers or as heterodimers to form enzymes with distinct patterns of substrate specificity. In addition, several of the isoforms appear to catalyse O-methylations common to both isoquinoline alkaloid and phenylpropanoid biosynthesis.


Identification of methyl jasmonate-inducible methyltransferase partial cDNAs

RNA was isolated from T. tuberosum cell suspension cultures 18 h after the addition of methyl jasmonate to the culture medium to a final concentration of 50 μm. This time point after elicitor addition was based upon elicitation time-course results obtained with the four methyltransferase enzyme activities of berberine biosynthesis (S. Frick and T.M. Kutchan, unpublished data). The enzyme activities typically began to increase at 18 h and reached maxima 36 h after elicitor addition. The methyltransferase RNA transcripts were estimated to reach maxima at 18 h. In comparison, berberine bridge enzyme activity reached a maximum at 18 h and by 36 h had returned to baseline levels. This observation was consistent with that noted for elicitor-treated E. californica cell cultures ( Dittrich & Kutchan 1991;Kutchan 1993;Kutchan & Zenk 1993). The T. tuberosum RNA isolated at 18 h was then used as template for RT–PCR using the following oligodeoxynucleotides as amplification primers. The primer sequences were based on conserved amino acid sequences in plant methyltransferases as partly suggested in Dumas et al. (1992 ).

inline image

In this manner, a mixture of PCR products of either 108 bp (primer pair MTsI/MTasII) or 216 bp (primer pair MTsI/MTasIV) in length was obtained. This mixture was resolved by subcloning into pGEM-T and each cDNA clone was then sequenced. Six different DNA sequences that encoded methyltransferases were thus identified. Oligodeoxynucleotides 30 nucleotides in length were designed over regions of maximal sequence variation within these six DNA sequences, and were used as hybridization probes for RNA gel blot analysis. Three of these oligodeoxynucleotides hybridized to RNA that accumulated in response to methyl jasmonate addition to the cell cultures ( Fig. 2). Screening of a λ-ZAP cDNA library, prepared from the same RNA from T. tuberosum as used for RT–PCR, with these three oligodeoxynucleotides resulted in the isolation of four full length cDNAs, MT1, MT2, MT3 and MT4.

Figure 2.

Graphic representation of an RNA gel blot prepared from methyl jasmonate-induced cell suspension cultures of T. tuberosum.

RNA was isolated at 0, 3, 6, 9, 12, 18 and 24 h after addition of elicitor to the culture medium. The blot was hybridized to five different methyltransferase partial cDNAs generated by PCR. Methyltransferase oligonucleotide 1 (•), methyltransferase oligonucleotide 2 (▵) methyltransferase oligonucleotide 4 (○), methyltransferase oligonucleotide 5 (□) and methyltransferase oligonucleotide 6 (▴).

Heterologous expression and identification of the methyltransferases

The nucleotide sequences of the four cDNAs were determined, the 5′- and 3′-flanking sequences were removed by generating only the reading frames by PCR, and the cDNAs were introduced into the baculovirus vector pFastBac1 for functional expression in S. frugiperda Sf9 cells. The insect cell culture medium and cells were tested for enzyme activity 3–4 days after infection with recombinant baculovirus using the four potential substrates for berberine biosynthesis (R,S)-norcoclaurine (S)-coclaurine, 3′-hydroxy-N-(S)-methylcoclaurine and (S)-scoulerine. Cells and medium from cultures infected with pFastBac/MT1 or pFastBac/MT2 were found to O-methylate (R,S)-norcoclaurine at the 6-hydroxy position to coclaurine, the first methylation of isoquinoline alkaloid formation ( Fig. 1). Most of the enzyme activity was associated with the medium (67%) while less (33%) was within the cells. Cells and medium from cultures infected with pFastBac/MT3 or pFastBac/MT4 did not catalyse O-methylation of (R,S)-norcoclaurine, but all four methyltransferases could transfer methyl groups to catechol and caffeic acid. Since a majority of the enzyme activity was secreted into the medium, the Sf9 cells were subsequently transferred to a serum-free medium (Sf-900 II SFM, Gibco BRL) for one passage prior to viral infection for ease of purification of the recombinant proteins. This was an approach that had worked successfully for both strictosidine synthase and the berberine bridge enzyme ( Kutchan et al. 1994 ).

Purification and characterization of the recombinant methyltransferases

The two O-methyltransferases that transform (R,S)-norcoclaurine were purified to homogeneity by a facile three-step procedure involving ammonium sulphate precipitation followed by dialysis, ion exchange chromatography and finally affinity chromatography. A typical purification starting from 530 ml insect cell culture medium which resulted in 0.9 mg pure enzyme in 40% overall yield is shown in Table 1. The production of O-methyltransferase in Sf9 cells ranged from 6–8 mg l–1. Gel electrophoretic analysis of aliquots of the protein solution from each of the purification steps is shown in Fig. 3. The recombinant O-methyltransferase was shown by size exclusion column chromatography to have a molecular mass of 74 kDa (data not shown) and the denatured protein migrated as a single 40 kDa band on an SDS polyacrylamide gel ( Fig. 3). Since translation of the nucleotide sequence of the O-methyltransferases also results in polypeptides of a molecular mass between 39 and 40 kDa, the active enzymes most likely have a dimeric structure and were renamed as OMT II;Thatu;1.1, OMT II;Thatu;2.2, OMT II;Thatu;3.3 and OMT II;Thatu;4.4 to reflect the catechol O-methylating activity, the source species T. tuberosum and the homodimeric tertiary structure. The physical characteristics of purified OMT II;1.1 and OMT II;2.2 are shown in Table 2. Catechol, caffeic acid and (R,S)-norcoclaurine were used as substrates that represented the major classes of compounds transformed by these particular O-methyltransferases. The apparent Km values determined with OMT II;1.1 for caffeic acid and (R,S)-norcoclaurine were similar (0.6 and 0.9 m m, respectively), suggesting that both the phenylpropanoid and the alkaloid could be substrates in vivo. Co-operativity was observed for O-methylation of these two substrates by OMT II;1.1 ( Table 2). This corroborates the result obtained by size exclusion column chromatography and suggests a dimeric structure for these O-methyltransferases. OMT II;2.2 was demonstrated by isoelectric focusing to have a pI of 5.6 with a minor protein band focusing at 5.4.

Table 1.  Purification of recombinant OMT II;2.2 from S. frugiperda Sf9 cells a
Purification stepTotal protein (mg)Total activity (pkatal)Specific activity (pkatal mg–1) Purification factor (fold)Yield (%)
  • a

    530 ml Sf9 cell culture medium.

Cell culture medium1061861.81100
Ammonium sulphate precipitation/dialysis24.51857.64.399
DEAE Sephacel10.8125126.667
SAH-Sepharose 4B0.974824740
Figure 3.

Purification of OMT II;2.2.

SDS–PAGE of an aliquot of each step in the purification procedure of heterologously expressed T. tuberosum O-methyltransferase OMT II;2.2 from Sf9 cells. Lane 1, uninfected Sf9 cells; lane 2, medium of Sf9 cells infected with recombinant baculovirus; lane 3, same sample as lane 2 after concentration; lane 4, DEAE Sephacel column flow-through; lane 5, DEAE Sephacel eluate; lane 6, SAH-Sepharose 4B eluate. Protein was detected by silver stain. Arrow, position of OMT II;2.2.

Table 2.  Characteristics of the purified O-methyltransferases
 OMT II;1.1OMT II;2.2
 CatecholCaffeic acid(R,S)-Norcoclaurine CatecholCaffeic acid
  1. a 67 m m Tris–HCl;b67 m m potassium phosphate;cn.d., not determined;dthe Michaelis–Menten plots for the substrates catechol, caffeic acid and (R,S)-norcoclaurine were sigmoidal, indicating co-operative substrate binding.

Temperature optimum (°C)3535403550
pH optimum 8.0 a 6.0 b 7.5 a 8.0 a 6.0 b
Vmax (μkatal) n.d. c1.
Km substrate (m m) n.d.
Km AdoMet (m m) n.d.0.250.1250.040.17
Hill coefficient dn.d.

Nomenclature and substrate specificity of the methyltransferases

Translation of the nucleotide sequences of the four methyltransferase cDNAs yielded very similar polypeptides. The four proteins were 93.2–99.7% identical to one another at the amino acid level, representing differences in 1–25 amino acids residues ( Fig. 4). The 5′ and 3′-untranslated regions were unique for each of the cDNAs. The proteins showed relatively high sequence similarity to other plant O-methyltransferases, for example 71% similarity of OMT II;1.1 compared to an o-diphenol O-methyltransferase from Capsicum annuum (accession number u83789), but relatively low sequence similarity to mammalian catechol O-methyltransferase (17% similarity compared to the rat liver enzyme ( Salminen et al. 1990 )). Of these highly similar T. tuberosum proteins, only two, however, were able to O-methylate (R,S)-norcoclaurine. In order to investigate the substrate specificity of all isoforms, a series of 38 potential substrates were tested with the four heterologously expressed methyltransferases.

Figure 4.

Amino acid sequence comparison of OMT II;Thatu;1, OMT II;Thatu;2, OMT II;Thatu;3, OMT II;Thatu;4 and OMT II;Thatu;5.

The complete amino acid sequence derived by translation of the nucleotide sequence of OMT II;Thatu;4 is shown. Only the variant amino acid residues from the remaining four sequences are indicated.

The four enzymes were all found to be O-methyltransferases with overlapping, but not identical, substrate specificity. Since these O-methyltransferases were determined to be dimeric in structure, the question was posed whether the enzymes expressed as heterodimers would have different substrate specificity from those enzymes expressed as homodimers. To achieve at least partial expression as heterodimers, the four cDNAs were expressed in all possible combinations of two different recombinant baculoviruses in Sf9 cells. Co-infection of insect cell cultures had already been successfully used for the co-expression of cytochrome P-450 and cytochrome P-450 reductase ( Pauli & Kutchan 1998;Rosco et al. 1997 ).

The same 38 substrates were then tested with 10 potentially different methyltransferases that represented four homodimers denoted OMT II;1.1, OMT II;2.2, OMT II;3.3 and OMT II;4.4 and the six heterodimers potentially resulting from co-infection that were denoted OMT II;1.2 (indicating co-infection of a recombinant baculovirus containing the Omt II;1 cDNA and a second recombinant baculovirus containing the Omt II;2 cDNA), OMT II;1.3, OMT II;1.4, OMT II;2.3, OMT II;2.4 and OMT II;3.4. The types of compounds tested as potential substrates varied widely in structural complexity and class. The smallest molecules tested were simple catechols such as catechol, guajacol and protocatechuic acid ( Table 3). The homo- or heterodimeric O-methyltransferase that catalysed this maximal conversion is indicated using the above-mentioned nomenclature. The most efficient methylation was achieved with intermediates from phenylpropanoid metabolism such as caffeic acid ( Table 4), although p-coumaric acid, ferulic acid and sinapic acid also served as less efficient substrates. Surprisingly, in contrast with catechol O-methyltransferases of mammalian origin, l-dopa and dopamine were not methylated. The structurally closely related molecules l-tyrosine, tyramine and 3,4-dihydroxyphenylacetic acid were also not methylated. A very broad substrate specificity was observed within the tetrahydrobenzylisoquinoline class of substances tested ( Table 5). As was observed for simple catechols and phenylpropanoids, the O-methyltransferases were capable of methylating more than one free phenolic hydroxyl group, in this case apparently only on the isoquinoline moiety, with preferential methylation at the 6-position. In the protoberberine class of alkaloids, (R,S)-3-O-demethylcheilanthifoline, but not (S)-scoulerine or (R,S)-tetrahydrocolumbamine, was methylated ( Table 6). The berberine alkaloids columbamine and jatrorrhizine also did not serve as substrate. The tetrahydrophenethylisoquinolines (S)-6-O-demethylautumnaline and (R)-6-O-demethylautumnaline, containing an additional carbon atom between the isoquinoline and benzyl moieties, were methylated ( Table 7).

Table 3.  Catechols tested as OMT II substrates Thumbnail image of
Table 4.  Phenylpropanoids tested as OMT II substrates Thumbnail image of
Table 5.  Tetrahydrobenzylisoquinolines tested as OMT II substrates Thumbnail image of
Table 6.  Protoberberines and berberines tested as OMT II substrates Thumbnail image of
Table 7.  Tetrahydrophenethylisoquinolines tested as OMT II substrates Thumbnail image of

The very broad substrate specificity that was found for these four O-methyltransferases from T. tuberosum was very surprising and was observed for the first time for an enzyme involved in plant alkaloid biosynthesis. Table 8 summarizes the breadth and degree of O-methylation observed with the 38 substrates tested at 167 μm with all 10 potential isoforms. The numerical values are expressed as specific activity normalized to the 3-O-methylation of 167 μm caffeic acid by OMT II;1.3 (27 pkat mg–1 protein). Several combinations of recombinant virus resulted in additional molecules being methylated, suggesting that the O-methyltransferases are indeed dimers and can also be expressed as heterodimers. New methylating activity was observed for OMT II;1.3 and OMT II;1.4 using protochatechuic acid as a substrate, for OMT II;1.2 and OMT II;3.4 using (S)-coclaurine, for OMT II;1.2, OMT II;1.3, OMT II;1.4, OMT II;2.3 and OMT II;2.4 using (R,S)-6-O-methylnorlaudanosoline, and for OMT II;1.2, OMT II;1.3, OMT II;2.3, OMT II;2.4 and OMT II;3.4 with (S)-norreticuline as substrate. These enzyme activities were not observed when the O-methyltransferases were expressed as homodimers. Caffeoyl-CoA was also tested as a substrate with pure OMT II;1.1 and OMT II;2.2 and was methylated. These enzyme assays with the CoA derivative could only be effectively carried out with purified enzyme due to the high thioesterase activity associated with the insect cells and medium.

Table 8.  Comparison of specific activities of 10 O-methyltransferase isoforms with various substrates a
SubstrateOMT II;1.1OMT II;1.2OMT II;1.3OMT II;1.4OMT II;2.2OMT II;2.3OMT II;2.4OMT II;3.3OMT II;3.4OMT II;4.4
  1. a Each value is the mean result of 4–6 assays from two separate expression experiments and is given relative to the 3-O-methylation of caffeic acid by OMT II;1.3 (100% = 27 pkat mg –1± 0.1%). The variation in the assays from a single infection ranged from ± 0.1–20%. The variation between infections was higher due to normal variations in the insect cell cultures over a period of 1 year. The reproducibility of substrate specificity was 100%. The same qualitative methylations were observed each time that the methyltransferases were tested. The concentration of substrate was held constant at 167 μm due to the limiting quantity and low solubility of many of the alkaloids used herein. The enzyme preparation used was serum-free insect cell culture medium containing secreted heterologous O-methyltransferase. Bold text indicates new methylations arising from heterodimer expression by co-infection of Sf9 cells with two different O-methyltransferase encoding cDNAs. Italic text indicates apparent subunit poisoning. OMT, O-methyltransferase. Substrate 8, l-tyrosine; 9, tyramine; 10, l-dopa; 11, dopamine; 12, 3,4-dihydroxyphenylacetic acid; other substrate structures are given in Tables 3–7. Substrates 8–12, 21–23, 27, 31, 32 and 34–36 were not methylated by any of the isoforms and are therefore not included in this table.


A low level of methylation was observed with uninfected control Sf9 cells only with (R,S)-4′-O-methylnorlaudanosoline (R,S)-6-O-methylnorlaudanosoline, (S)-norlaudanosoline and (R)-and (S)-norreticuline as substrates. These methylation values varied from 5–7% of the relative specific activity and were subtracted from all assays involving these substrates with heterologously expressed O-methyltransferase.


Previous work on the biosynthetic enzymes of various classes of alkaloids suggested that these enzymes are highly substrate-specific and are dedicated to alkaloid biosynthesis. An investigation of the substrate specificity of any enzyme may be most clearly addressed when working with a heterologously expressed clone since the purity of a native enzyme can always be questioned. Heterologous expression of cDNAs of isoquinoline alkaloid biosynthesis has indicated that, in certain cases, the substrate specificity is very high, such as for the cytochrome P-450-dependent mono-oxygenase (S)-N-methylcoclaurine 3′-hydroxlyase ( Pauli & Kutchan 1998), for which only one substrate, (S)-N-methylcoclaurine, could be identified, or is somewhat permissive, as for the berberine bridge enzyme which will accept a series of isoquinoline alkaloids of related structure as substrate ( Kutchan & Dittrich 1995). We observe herein for the first time that an enzyme that catalyses a transformation of alkaloid biosynthesis may also be involved in a non-alkaloidal secondary metabolic pathway, that of phenylpropanoid anabolism.

The plant O-methyltransferases reported in this work have surprisingly broad substrate specificity ranging from simple catechols to phenylpropanoids and isoquinoline alkaloids of relatively complex structure. These observations are in contrast to that reported in the literature for O-methyltransferases of (S)-reticuline biosynthesis from Argemone platyceras ( Rueffer et al. 1983 ) and Coptis japonica ( Sato et al. 1994 ) plant cell suspension cultures for which the methylation of catechol and caffeic acid was not detected. Other plant methyltransferases that have been characterized, either as purified enzymes or as heterologously expressed cDNAs, have demonstrated the ability to methylate several substrates but such varied structures as reported herein have not yet been investigated (summarized in Ibrahim et al. 1998 ). Some of the most thoroughly characterized plant O-methyltransferases are those that methylate caffeic acid or caffeoyl-CoA to the corresponding ferulate. It has been observed that, depending on the plant species, these enzymes either methylate only caffeoyl-CoA ( Schmitt et al. 1991 ) or caffeic acid and caffeoyl-CoA ( Li et al. 1997 ). The O-methyltransferases from T. tuberosum are capable of transforming both caffeic acid and caffeoyl-CoA in addition to the alkaloidal substrates. An observation of this type would make sense with respect to the defence responses of a plant cell when treated with an elicitor such as methyl jasmonate. The accumulation of toxic alkaloids and phenylpropanoids as well as a reinforcement of the plant cell wall in response to elicitor treatment are well documented in the literature (reviewed in Dixon & Paiva 1995;Kutchan 1998). This up-regulation of defence substance biosynthesis could be viewed as more economical when carried out by enzymes common to more than one pathway. Consistent with this view, it has recently been reported that a single O-methyltransferase from Chrysosplenium americanum can methylate both flavonoid and phenylpropanoid compounds ( Gauthier et al. 1998 ).

We have chosen herein to adopt the nomenclature suggested by Joshi & Chiang (1998). The T. tuberosum O-methyltransferases apparently belong to the OMT II group based upon occurrence of conserved motifs, the absence of a requirement for divalent cations such as Mg2+, and utilization of a variety of substrates ( Joshi & Chiang 1998). Among T. tuberosum O-methyltransferases, the amino acid sequences of OMT II;1.1, which O-methylates (R,S)-norcoclaurine, and OMT II;4.4, which does not carry out this alkaloid methylation, differ by only one amino acid (tyrosine versus cysteine, respectively, at position 21). This indicates a very close evolutionary relationship between alkaloid and phenylpropanoid biosynthetic O-methyltransferases and suggests that these probable allelic variants could have arisen by gene duplication. Southern analysis indicates that at least six alleles may be present in the T. tuberosum genome (data not shown). As suggested by Northern analyses, these methyltransferase genes are coordinately regulated upon induction with methyl jasmonate. Although the magnitude of induction between the alleles varies, the time course of induction is the same, reaching a maximum at 6–9 h after addition of methyl jasmonate to the cell culture medium. This implies that all isoforms should be simultaneously present in the cell. We do not yet know whether the O-methyltransferases can occur in planta as homodimers, heterodimers or a mixture of both as appears to be the case when the cDNAs were heterologously expressed in insect cell culture. Under conditions of elicitation, the RNA transcripts are probably collectively present. This would, in principal, allow for various combinations of monomers to occur. Given the four closely related isoforms observed from T. tuberosum that were apparently combined in this study into 10 functional enzymes of distinct substrate specificity (a fifth isoform encoding cDNA (Omt II;Thatu;5) was also isolated but could not be stably introduced into the baculovirus vector for analysis), the methylating capacity appears to be quite large and could represent a form of combinatorial biochemistry that naturally occurs within the plant cell.

Experimental procedures


Cell suspension cultures of T. tuberosum were provided by the departmental cell culture laboratory. Cultures were routinely grown in 1 litre conical flasks containing 400 ml of Linsmaier–Skoog medium ( Linsmaier & Skoog 1965) over 7 days at 23°C on a gyratory shaker (100 rev min–1) in diffuse light (750 lux). Elicitation of T. tuberosum cell suspension cultures was achieved by the aseptic addition of methyl jasmonate to the culture medium to a final concentration of 50 μm.

The isoquinoline alkaloid substrates used in the enzyme assays were from the departmental collection. Tyramine was purchased from Aldrich, catechol, 3,4-dihydroxyphenylacetic acid, l-dopa and sinapic acid from Fluka, p-coumaric acid, dopamine and guajacol from Sigma, l-tyrosine from Merck, and caffeic, ferulic and protochatechuic acids from Roth.

Isolation of partial and full-length cDNAs

Partial cDNAs encoding methyltransferases from T. tuberosum were generated by PCR using cDNA produced by reverse transcription of total RNA isolated from 2-day-old suspension cultured cells to which methyl jasmonate had been added to a final concentration of 50 μm and incubated for 18 h. DNA amplification was performed under the following conditions: 5 min at 94°C, then 30 cycles of 94°C for 30 sec; 40°C for 60 sec, 72°C for 60 sec. At the end of the last cycle, the reaction mixtures were cooled to 4°C. The amplified DNA was then resolved by agarose gel electrophoresis, the band of approximately the correct size (108 bp or 216 bp, depending on the primers used) was isolated and subcloned into pGEM-T (Promega) prior to nucleotide sequence determination.

cDNA clones encoding the methyl jasmonate-inducible T. tuberosum methyltransferases were isolated by screening of a cDNA library prepared in λ-ZAP II (Stratagene) using [γ-32P]ATP-labelled oligodeoxynucleotides, the sequences of which were based on the partial clones generated by PCR, as hybridization probes.

Heterologous expression of O-methyltransferases

The four T. tuberosum cDNA clones that yielded positive results through a third screening were converted to pBluescript SK– by excision. After determination of the nucleotide sequence on both strands, the full-length open reading frame, free of the 5′- and 3′-flanking sequences, was generated by PCR using Pfu DNA polymerase and was ligated into BamHI/XhoI-digested pFastBac1 (Life Technologies). The four vector constructions thus generated, pFastBac/MT1, pFastBac/MT2, pFastBac/MT3 and pFastBac/MT4, were transposed into baculovirus DNA in the E. coli strain DH10BAC (Life Technologies) and then transfected into S. frugiperda Sf9 cells according to the manufacturer's instructions. The insect cells were propagated and the recombinant virus was amplified according to ( Kutchan et al. 1994 ;Pauli & Kutchan 1998).

Enzyme assays

The standard enzyme assay used was 10 μmol Tris–HCl pH 8.0, 5 μmol sodium ascorbate, 10 nmol AdoMet, 50 000 d.p.m. [methyl-3H]AdoMet (80 Ci mmol–1, 2.9 TBq mmol–1), 25 nmol substrate and approximately 10 μg pure protein or 40–80 μg crude extract protein in a total volume of 150 μl, incubated at 35°C. The linear range was determined for the various protein preparations used and the incubation time was held within this range. The reaction was quenched by the addition of either 200 μl 1 m sodium carbonate or 200 μl 1 m ammonium acetate and was extracted with 400 μl ethyl acetate. A 300 μl aliquot of the organic phase was removed and the amount of radioactivity present was determined with a scintillation counter. All enzyme assays were performed at least in triplicate from at least two independent insect cell culture infections. Values presented represent the mean of these repetitive assays.

Purification of T. tuberosum O-methyltransferases from Sf9 cells

After the infection of 750 ml suspension grown insect cells had proceeded for 4–6 days at 27°C and 140 rev min–1, the cells were removed by centrifugation under sterile conditions at 1200 g for 10 min at 4°C. All subsequent steps were performed at 4°C. The pellet was discarded and the medium was brought to 100 m m potassium phosphate buffer pH 8.0, 20 m mβ-mercaptoethanol and 10% (v/v) glycerol. This solution was slowly brought to 70% saturation with (NH4)2SO4 over 30 min. After the (NH4)2SO4 was completely dissolved, the suspension was stirred for 1 h and then centrifuged at 13 000 g for 90 min. The protein pellet was gently resuspended and dissolved in 25–30 ml of standard buffer (20 m m potassium phosphate buffer pH 8.0, 20 m mβ-mercaptoethanol and 10% (v/v) glycerol). Insoluble matter was removed by centrifugation at 800 g for 10 min. The clear protein solution thus obtained was then dialysed overnight against 5 l of standard buffer. This dialysate was applied directly onto a DEAE Sephacel (Pharmacia) column (18 × 65 mm) that had been equilibrated with standard buffer at a flow rate of 1 ml min–1. The flow-through was continuously monitored at 280 nm and the effluent was discarded. The loaded column was then washed with standard buffer until the absorbance returned to background levels. The bound protein was eluted from the column with standard buffer containing 0.5 m KCl at a flow rate of 1 ml min–1. The fractions that contained active enzyme were pooled and concentrated to 10–15 ml with an Amicon filter (10 kDa cut-off). The protein solution was passed over a PD-10 column (desalting column with Sephadex G-25, Pharmacia) and then applied onto an S-adenosyl- l-homocysteine–Sepharose 4B affinity column (18 × 65 mm) prepared according to Sharma & Brown (1978). The affinity column was equilibrated with standard buffer at a flow rate of 0.25 ml min–1 and the protein solution loaded onto the column at this same flow rate. The loaded column was then washed with standard buffer until the absorbance returned to the background levels. The bound protein was eluted from the column with standard buffer containing 2 m KCl at a flow rate of 0.5 ml min–1. The fractions that contained active enzyme were pooled and the protein solution was passed over a PD-10 column to remove the excess KCl. Molecular mass determination for the native O-methyltransferase was performed in standard buffer on an AcA 34 Ultrogel (Serva) size exclusion column (2.4 × 86 cm, flow rate 0.3 ml min–1) using cytochrome c (12.4 kDa), equine myoglobin (17.8 kDa), chymotrypsinogen A (25 kDa), egg albumin (45 kDa), bovine serum albumin (67 kDa) and rabbit aldolase (160 kDa) as molecular mass markers.

General methods

Total RNA was isolated and RNA gels were run and blotted as previously described ( Pauli & Kutchan 1998). cDNA clones were labelled by random-primed labelling with [α-32P]dCTP and oligodeoxynucleotides were end-labelled with [γ-32P]ATP. Levels of hybridized RNA on Northern blots were evaluated with a Raytest BAS-1500 phosphorimager. The entire nucleotide sequence on both DNA strands of full-length cDNA clones in pBluescript SK– was determined by dideoxy cycle sequencing using internal DNA sequences for the design of deoxyoligonucleotides as sequencing primers. Caffeoyl-CoA was synthesized according to Zenk et al. (1980 ).


We thank Ms Kerstin Andersson and Dr Friedrich Lottspeich, Martinsried, for the isoelectric focusing measurements. This work was supported by SFB 369 of the Deutsche Forschungsgemeinschaft, Bonn, and Fonds der Chemischen Industrie, Frankfurt.