[English
Start Page]
[Deutsche
Startseite]
[Email
Feedback] |
(Last modification: 07.March 2010)
Small Molecule O-Methyltransferases (OMTs) in Plants
Overview
S-adenosyl-L-methionine (SAM) dependent O-methyltransferases (OMTs) methylating small molecules in plants are involved in the biosynthesis of lignin, flavonoids, alkaloids, and many other plant secondary products, and often the methylations are essential in determining specific physiological functions of the molecules. The increasing number of reactions, substrates, and proteins necessitated very early some classification. In the beginning, it was reasonable to base it on substrate specificities, see for example Hermann et al., 1987; Pellegrini et al., 1993; Jaeck et al., 1996; Maury et al., 1999. With the emergence of DNA sequence data, and thus the increasing availability of protein sequences, the next step was to include primary protein sequences, i.e. to consider protein relationships and sequence motifs rather than only substrate specificities, see e.g. Ibrahim, 1997; Ibrahim et al., 1998; Joshi and Chiang, 1998. With the emergence of protein 3D-structures and much more detailed insights into structure and function, it was logical to use these also in the classification, and to include the powerful possibilities of protein modelling. Such system was proposed by Noel and coworkers (Noel et al., 2003). I myself found it quite attractive, and therefore I will give a brief description here, including some examples and references, often on work reported after 2003. The authors defined three types:
This is by far the largest group, a protein superfamily with an astonishing diversity of substrates. It contains for example enzymes methylating hydroxyl groups in phenylpropanoids (e.g. in the lignin pathway), eugenol, chavicol, flavonoids, isoflavonoids, alkaloids, coumarins, orcinols, stilbenes, and so on. Often overlooked, but noteworthy: several of these enzymes have been shown to be active also with phenylpropanoid CoA-esters, i.e. the substrates typical for type 2 OMTs (see below). The common characteristics of type 1 proteins are that they are homodimeric with subunit sizes of about 38-43 kDa, and that they do not require divalent cations for activity (in contrast to type 2 OMTs, see below). The range of possible substrates may be much larger than we presently realize since many of the family members are still only identified by sequence similarities. Often changes in one or a few amino acids are sufficient for pronounced alterations in substrate specificity; there are excellent examples for that in the literature, but I will not discuss them in detail here. Several of these enzymes have been crystallized, e.g. the chalcone O-methyltransferase (ChOMT), the isoflavone O-methyltransferase (IOMT), and the caffeic acid/5-hydroxyferulic acid 3/5-O-methyltransferase (COMT) from alfalfa (Medicago sativa) (Zubieta et al., 2001; Zubieta et al., 2002) as well as the hydroxyisoflavanone 4'-O-methyltransferase (HI4'OMT) from barrel medic (Medicago truncatula) (Liu et al., 2006). All of this provided a structural basis for understanding OMT architecture, substrate selectivity and catalytic mechanisms. Nevertheless it is still difficult to predict the physiological substrates of newly identified family members based upon structural models alone. Moreover, several enzymes possess broad substrate specificities in vitro that do not permit easy conclusions about in vivo substrates; the subfamily of the so-called caffeic acid OMTs (COMTs) and closely related proteins gives some good examples for that: more....
It has long been taken for granted that the proteins of this superfamily are
specific for hydroxyl groups as methyl group acceptors, but even that is not
correct: the type 1 OMT protein family includes a
S-methyltransferase
(SMT) from Catharanthus roseus (more...), and
two N-methyltransferases
(NMT):
This protein family is much smaller than the type 1 group, and less diverse in sequence and gene number. All of these proteins appear to be specific for CoA-esters of phenylpropanoids, and their main role seems to be in lignin biosynthesis. However, the specificity is not absolute: there are examples that members of this family are also active with flavonoids and/or phenylpropanoid conjugates (Ibdah et al., 2003; Kim et al., 2006; So et al., 2007; Lee et al., 2008), and there is at least one example with activity for only flavonols and anthocyanins (Hugueney et al., 2009); this protein carried out sequential 3'- and 5'-methylations. The proteins are also homodimers, but the subunit sizes are smaller (23-28 kDa), and they typically require a divalent cation (e.g. Mg2+) to mediate the deprotonation of the acceptor hydroxyl group prior to the methyl group transfer. One of these proteins has been crystallized, i.e. the alfalfa caffeoyl coenzyme A 3-O-methyltransferase (Ferrer et al., 2005); it shows high structural similarity to the animal catechol OMT that also require Mg2+ (Vidgren et al., 1994).
The large majority of these proteins are carboxyl methyltransferases. Examples are: Benzoic acid / salicylic acid carboxyl methyltransferases from Clarkia breweri (Ross et al., 1999), Stephanotis floribunda and Nicotiana suaveolens (Pott et al., 2002; Pott et al., 2004), Arabidcpsis thaliana (Chen et al., 2003), snap dragon (Antirrhinum majus) (Murfitt et al., 2000; Dudareva et al., 2000), Atropa belladonna (Fukami et al., 2002), and rice (Oryza sativa) (Xu et al., 2006; Koo et al., 2007). Also to this family belong the jasmonic acid carboxyl methyltransferase (JMT) from Arabidopsis thaliana (Seo et al., 2001) and hot pepper (Capsicum annuum L.) (Min et al., 2005), the IAA carboxyl OMTs from Arabidopsis thaliana and other plants (Qin et al., 2005; Zhao et al., 2007; Zhao et al., 2008), the cinnamate/p-coumarate carboxyl methyltransferases (CCMTs) from basil (Ocimum basilicum) (Kapteyn et al., 2007), the loganic acid O-methyltransferase involved in secologanin biosynthesis in Catharanthus roseus (Murata et al., 2008), the Arabidopsis thaliana GAMT1 and GAMT2 genes encoding enzymes that catalyze the formation of the methyl esters of gibberellins (GAs) (Varbanova et al., 2007), and an Arabidopsis thaliana methyltransferase capable of methylating farnesoic acid (Yang et al., 2006). Interestingly, this family also contains some members which are not carboxyl OMTs, but N-methyltransferases (NMT): enzymes carrying out methylations in the biosynthesis of caffeine (Kato et al., 2000; Ogawa et al., 2001; Uefuji et al., 2003; Mizuno et al., 2003a; Mizuno et al., 2003b; Kato and Mizuno, 2004; Yoneyama et al., 2006). This protein family has also been called the 'SABATH' family, named from the abbreviations of the most well known members up to 2003: Salicylic Acid carboxyl OMT, Benzoic Acid carboxyl OMT, and THeobromine N-methyltransferase (one of the enzymes in caffeine biosynthesis) (D'Auria et al., 2003). If you did not know that term yet: You better get used to it because it has been used very often in the last years. 3D-Structures are also available, e.g. from the salicylic acid carboxyl methyltransferase (SAMT) from Clarkia breweri (Zubieta et al., 2003), and the indole-3-acetic acid (IAA) methyltransferase from Arabidopsis thaliana (Zhao et al., 2008). XMT and DXMT, the enzymes from Coffea canephora that catalyse the three independent N-methyl transfer reactions in the caffeine-biosynthesis pathway were also both crystallized (McCarthy et al., 2007a; McCarthy and McCarthy, 2007b). As might have been expected from the sequence similarities, their overall structure is nearly identical to that found in the carboxyl OMTs.
Four years later, the issue of the nomenclature was raised again (Lam et al., 2007). Apparently, these authors were not aware of the proposal outlined above (the Noel publication from 2003 is not cited and not discussed). The nomenclature is different in several points, and thus some confusion will probably be inevitable in the future. In brief: based on their sequence alignment analysis, the authors distinguish two lineages, A and B. Lineage A contains the enzymes methylating CoA-esters (A1, also called group I; Noel = type 2, Joshi = Pl-OMT I) and the carboxyl OMTs (A2, Noel = type 3, Joshi = unknown at that time). The N-methyltransferases in caffeine biosynthesis are not considered in this analysis. Lineage B contains the remaining OMTs (also called group II; Noel = type 1; Joshi = Pl-OMT II), and is also divided into two subgroups (B1 and B2). In several figures of supplementary data, the authors provide among other results some attempts to identify in the protein sequences residues that might be characteristic for position specific methylations in flavonoids: A or B ring, or meta and para positions.
Conclusions It is a bit difficult to compare the two nomenclature proposals. They differ in the importance assigned to structural analysis (Noel) and sequence motifs (Joshi, Lam). I myself spent years staring at primary protein sequences and trying to see significances in certain motifs (not only with OMTs, but even more so with the type III polyketide synthases that are discussed in some detail in our website: more...). This was the thing to do until about 8-10 years ago (i.e. 2001, when the first 3D-structures were published); it was pretty useful, but one has to admit also that the success was limited in many ways. This changed with the availability of 3D-structures and the insights they provided into substrate selectivity and reaction mechanisms; in particular the possibilities to model new proteins on the basis of known structures promised easier access to understand related enzymes. However, predicting the functions of new members of the protein families (e.g. substrate preferences) remains difficult, and in most (if not all) cases the structural studies/models simply explained something found before by the biochemical functional identification. I am not aware that modelling led to successful prediction of the physiological substrates of an OMT, and the same is true for type III PKS. Nevertheless, my conviction is that this is a much better help to understand the complexity of these enzymes than looking at sequence motifs, and therefore I would tend to favor the Noel nomenclature. Whatever your preferences are: do us all a favor and cite the reference if you use any of these different terminologies! It also should be noticed that there are automated servers doing some good modelling (e.g. at http://swissmodel.expasy.org/), but the more sophisticated stuff (e.g. docking of substrates) requires hardware and software that might not be available (or affordable) to everybody.
File History:
. |
|
|
