Vitaminas complejo b

View Article Online / Journal Homepage / Table of Contents for this issue

REVIEW

www.rsc.org/npr | Natural Product Reports

Roles of vitamins B5, B8, B9, B12 and molybdenum cofactor at cellular and organismal levels†

Published on 20 August 2007. Downloaded by Harvard University on 19/11/2013 20:46:55.

Fabrice R´ebeill´e,*a St´ephane Ravanel,a Andr´ee Marquet,b Ralf R. Mendel,c Alison G. Smithd and Martin J. Warrene Received (in Cambridge, UK) 14th June 2007 First published as an Advance Article on the web 20th August 2007 DOI: 10.1039/b703104c Covering: 1984 to 2007 Many efforts have been made in recent decades to understand how coenzymes, including vitamins, are synthesised in organisms. In the present review, we describe the most recent findings about the biological roles of five coenzymes: folate (vitamin B9), pantothenate (vitamin B5), cobalamin (vitamin B12), biotin (vitamin B8) and molybdenum cofactor (Moco). In the first part, we will emphasise their biological functions, including the specific roles found in some organisms. In the second part we will present some nutritional aspects and potential strategies to enhance the cofactor contents in organisms of interest.

1 2 2.1 2.1.1 2.1.2 2.1.3 2.1.4 2.1.5 2.2 2.2.1 2.2.2 3 3.1 3.2 3.3 4 5

Introduction Biological functions Main functions found in all organisms Nucleic acid synthesis: the role of folate The methylation cycle: the roles of folate and cobalamin Fatty acid biosynthesis and gluconeogenesis: the roles of biotin and pantothenate Redox reactions: the role of Moco Other metabolic functions for folate, biotin and cobalamin The main differences among eukaryotic organisms Compartmentalisation Specific needs in some eukaryotes Nutritional aspects Effects of deficiency on human health Main dietary sources Strategies for enhancement Conclusion: compartmentalisation, a challenging area References

a Laboratoire de Physiologie Cellulaire V´eg´etale, UMR5168, Universit´e Joseph Fourier-CNRS-CEA-INRA, Institut de Recherche en Technologies et Sciences du Vivant, CEA-Grenoble, 17 rue des Martyrs, F-38054, Grenoble, Cedex 9, France. E-mail: frebeille@cea.fr; Fax: +33 438-78-50-91; Tel: +33 438-78-44-93 b Department of Chemistry, Universit´e Pierre et Marie Curie, UMR CNRS 7613, 75252, Paris, France. E-mail: marquet@ccr.jussieu.fr c Department of Plant Biology, Technical University of Braunschweig, 38106, Braunschweig, Germany. E-mail: r.mendel@tu-bs.de d Department of Plant Sciences, University of Cambridge, Cambridge, CB2 3EA, UK. E-mail: as25@cam.ac.uk e Department of Biochemistry, University of Kent, Canterbury, UK. E-mail: m.j.warren@kent.ac.uk † This paper was published as part of a themed issue on vitamins and cofactors.

This journal is © The Royal Society of Chemistry 2007

1 Introduction Cofactors are small molecules (at least compared to the size of a protein) that facilitate an enzyme to catalyze a reaction.

These ‘chemical tools’ can be inorganic (metal ions or clusters) or organic (coenzymes) and are generally involved in group transfer or redox reactions. They can act as co-substrates or be permanently associated with the structure of the enzyme (prosthetic groups).

A large number of these coenzymes are derived from vitamins.

Vitamins, by definition, are dietary substances required for good health and normal development of animals. Most of them are only synthesised by microorganisms and plants. During the course of animal evolution, the ability to biosynthesise these compounds has been lost and, instead, elaborate uptake mechanisms have been developed. As many vitamins are only required in trace quantities, their biosynthesis is normally strictly controlled and the enzymes involved are produced in vanishingly small amounts. This is why it has been extremely difficult to elucidate their complete biosynthetic pathways, and it still remains the case that many steps within the biosynthesis of vitamins are poorly understood (see the review by Webb and Smith in this issue).

Because they are essential in all organisms and are required in a number of biological processes, vitamins are of considerable interest in terms of what they do and how they are made. In the post-genomic era there now exist opportunities to understand fully how these compounds are synthesised and what their whole cellular functions are. These functions can be quite complex because one particular vitamin may have various metabolic and chemical roles. In addition, this role may fluctuate from one organism to another depending on the presence of specific metabolisms (for example photosynthesis in plants). Increasing our knowledge concerning their synthesis and function is a prerequisite to develop new strategies for health and/or wealth creation, including improvement of food quality, design of new antibiotics targeting vitamin biosynthesis, and engineering synthesis of new compounds.

Nat. Prod. Rep., 2007, 24, 949–962 | 949

View Article Online

Published on 20 August 2007. Downloaded by Harvard University on 19/11/2013 20:46:55.

Dr Fabrice R´ebeill´e is Director of Research at the Commissariat a` l’Energie Atomique (CEA), Grenoble, France. He graduated from the University of Grenoble where he obtained a Diploma in Pharmacology in 1978 and a PhD in 1983. After a post-doc in Prof. M. D. Hatch’s lab in CSIRO, Canberra, he made a career within the CEA and was also Professor of Biochemistry at the University of Grenoble for four years.

His main research activities are focused on plant metabolism, including phosphate metabolism, photosynthesis, photorespiration, and more recently, folate biosynthesis and C1 metabolism.

Dr St´ephane Ravanel is Associate Professor at the University of Grenoble, France.

He studied biochemistry and molecular biology at the universities of Lyon and Grenoble, and then completed a PhD in plant biology in 1995. After a post-doc in Prof. Rochaix’s lab at the University of Geneva, he joined the University of Grenoble and the lab of Plant Cell Physiology in 1998. His present research focuses on the molecular, biochemical and regulatory aspects of folate and one-carbon metabolism in plants.

St´ephane Ravanel Fabrice R´ebeill´e

Here, we will describe, as examples, the role of five coenzymes: folate (vitamin B9), pantothenate (vitamin B5), cobalamin (vitamin B12), biotin (vitamin B8) and molybdenum cofactor (Moco).

Firstly, their biological functions will be emphasised, including the specific roles found in some organisms. In the second part we will present some nutritional aspects concerning these coenzymes and potential strategies to enhance the cofactor contents in organisms of interest.

2 2.1

Biological functions Main functions found in all organisms

Cofactors are required in almost all important metabolic pathways. Because they are specialised in certain types of reaction, one particular cofactor can be involved in several pathways and, conversely, several cofactors can be required in one particular pathway. The following section considers the main areas of metabolism in which the above five coenzymes are involved.

2.1.1 Nucleic acid synthesis: the role of folate. The syntheses of both purine and pyrimidine nucleotides require a folate coenzyme. Folates are involved in ‘one-carbon’ unit (C1 unit) transfer reactions, also called ‘C1 metabolism’. Folate is a generic term that represents a family of molecules (Fig. 1) deriving from tetrahydrofolate (5,6,7,8-tetrahydropteroylpolyglutamate, THF).

Chemically, these folate molecules are composed of a pterin ring, a p-aminobenzoic acid (pABA) unit and a polyglutamate chain with a variable number (1 to 14) of glutamate residues.1 Their role is to transport and donate C1 units. The C1 units transported by the vitamin arise essentially from the reaction catalyzed by serine hydroxymethyltransferase (SHMT) that converts serine into glycine.2 Formate is also a potential, although minor, source of C1 units.3 Once attached to the THF body, these C1 units can be reduced or oxidised, from methyl (the most reduced), via methylene, to formyl or methenyl (the most oxidised).4 Depending 950 | Nat. Prod. Rep., 2007, 24, 949–962

Fig. 1 Chemical structure of THF and major reactions of C1 metabolism.

THF is substituted at the N-5 and/or N-10 positions by C1 units having various oxidation states. There are generally between 4 and 8 glutamate residues. Serine and, to a lesser extent, formate, are the sources of C1 units.

The highest flux of C1 units occurs through methionine synthesis to sustain AdoMet turnover and all the methylation reactions, which explains why 5-methyltetrahydrofolate is the dominant folate species.

on the nature of the C1 unit carried, the folate coenzyme will be involved in various pathways (summarised in Fig. 1). Thus, the folate pool is a complex mixture of related molecules differing in the oxidation state of the pterin ring (di- or tetrahydrofolate), in the oxidation state of the C1 unit carried and in the length of This journal is © The Royal Society of Chemistry 2007

Published on 20 August 2007. Downloaded by Harvard University on 19/11/2013 20:46:55.

View Article Online

the glutamate chain. Only the most reduced form of the cofactor (tetrahydrofolate) can transport these C1 units.

Synthesis of the purine ring provides AMP and GMP bases for DNA and RNA strands, as well as for coenzymes such as NAD(P)+ , FAD, CoA and S-adenosylmethionine (AdoMet). In plants, these nucleotides are also precursors for purine alkaloids and the cytokinin hormones. The pathway for their synthesis is similar in plants, animals and micro-organisms.5 It is a complex process requiring 13 steps from ribose 5-phosphate. The fourth and tenth steps, respectively catalyzed by glycinamide ribonucleotide transformylase and aminoimidazole carboxamide ribonucleotide transformylase, involve the addition of a C1 unit. This C1 unit is provided by the 10-formyltetrahydrofolate derivative (Fig. 2a), inserting the carbons which become C-2 and C-8 of the purine ring.

involved in the de novo synthesis of THF in folate-autotroph organisms.6 2.1.2 The methylation cycle: the roles of folate and cobalamin.

Methionine (Met) is an essential amino acid not only required for the synthesis of protein but also for the formation of Sadenosylmethionine (AdoMet), the universal methyl donor and a key element in the ‘methylation cycle’ (Fig. 3a).7,8 In fact, 80% of the free Met present in the cell is used in this cycle, whose function is continuously to supply with AdoMet the dozens of methyltransferase enzymes present in all cells. These methyltransferases, as the name suggests, transfer the methyl group from AdoMet to a very large range of substrates for the synthesis of numerous compounds, including chlorophyll, lipids, lignin, hormones and vitamins.4 They are also involved in a wide range of functions such as regulation of gene expression (methylation of DNA and histones)9,10 or regulation and repair of proteins (rubisco enzyme, myelin basic protein etc.)11,12 The central step in this cycle is the methylation of homocysteine (Hcy), a reaction catalyzed by methionine synthase and responsible for the continual regeneration of Met (Fig. 3a). In this reaction, the C1 unit transferred to Hcy arises from 5-methyltetrahydrofolate,13 making this folate derivative the de facto source of all methyl groups. Interestingly, there are two types of methionine synthase whose activities depend or not on the presence of a cobalamin cofactor. The cobalamin-independent form of the enzyme (MetE) is found in plants and fungi, the cobalamin-dependent type (MetH) is found in animals14 (see below), and both types exist in enterobacteria and certain algae. Thus, this second class of enzyme (MetH) requires two coenzymes: folate and cobalamin.

Fig. 2 Role of folates in nucleotide synthesis. A: reactions involved in purine synthesis: 1) GAR transformylase; 2) AICAR transformylase. THF generated during these reactions is recycled back to 10-CHO-THF or 5,10-CH2 -THF, as shown in Fig. 1. B: reactions involved in thymidylate synthesis: 3) thymidylate synthase; in this reaction, 5,10-CH2 -THF is both a C1 unit donor and a reducing agent; 4) dihydrofolate reductase; 5) SHMT. In each case the C1 unit added is indicated by a box. GAR, glycinamide ribonucleotide; FGAR, formyl glycinamide ribonucleotide; AICAR, aminoimidazole carboxamide ribonucleotide; FAICAR, formyl aminoimidazole carboxamide ribonucleotide; 10-CHO-THF, 10-formyltetrahydrofolate; 5,10-CH2 -THF, 5,10-methylenetetrahydrofolate.

The synthesis of thymidylate is also closely linked to C1 metabolism. Indeed, the enzyme that converts the uracil base into the thymine base, to form the pyrimidine (dTMP) found uniquely in DNA, uses the folate vitamin with the C1 unit attached as 5,10methylenetetrahydrofolate (Fig. 2b). Whereas THF is regenerated during purine synthesis, it is dihydrofolate (DHF) that is formed during dTMP synthesis. The latter is converted back to THF by the important enzyme dihydrofolate reductase, an enzyme also This journal is © The Royal Society of Chemistry 2007

Fig. 3 The methylation cycle. A, overview of the methylation cycle; 1) methionine synthase (either MetE or MetH); 2) AdoMet synthetase; 3) methyl transferases; 4) S-adenosylhomocysteine hydrolase; 5) SHMT; 6) methylenetetrahydrofolate reductase; serine is the source of the C1 units required to methylate X; 5,10-CH2 -THF, 5,10-methylenetetrahydrofolate; 5-CH3 -THF, 5-methyltetrahydrofolate; AdoHcy, S-adenosylhomocysteine. B, the intermediary role of cobalamin in the B12-dependent methionine synthase (MetH).

Nat. Prod. Rep., 2007, 24, 949–962 | 951

Published on 20 August 2007. Downloaded by Harvard University on 19/11/2013 20:46:55.

View Article Online

The B12-dependent methionine synthase (MetH) catalyses the transfer of a methyl group from methyltetrahydrofolate to Hcy.15 In essence, this is the same reaction as catalysed by the B12independent enzyme (MetE), and both enzymes face the same challenge of transferring the methyl group from a very poor leaving group, the tertiary amine of methyltetrahydrofolate, to the sulfur of Hcy.16 The B12-dependent enzyme differs from the independent enzyme in that it uses cobalamin as an intermediary methyl group carrier (Fig. 3b), whereas the B12-independent enzyme catalyses direct transfer of the methyl group. The employment of cobalamin appears to make the process much more favourable, as the B12 enzyme has significantly a higher (by two orders of magnitude) kcat value.17 Methionine synthase (MetH) is a large modular enzyme consisting of 4 distinct regions representing (i) Hcy-binding, (ii) methyltetrahydrofolate-binding, (iii) cobalamin-binding and (iv) AdoMet-binding domains.15 The structure of the full length protein has not been determined but the topology of the individual domains has been elucidated and this, coupled with a large amount of mechanistic work, has allowed a comprehensive overview of the catalytic cycle of the enzyme to be resolved. The enzyme operates by binding Hcy to a specific site in domain (i) that is defined, in part, by the presence of an essential zinc ion, itself ligated to the region via three cysteine residues. Methyl cobalamin is formed by the transfer of a methyl group from methyltetrahydrofolate bound to domain (ii) to Co(I) cobalamin bound in domain (iii). This transfer is facilitated by the strong nucleophilicity of the Co(I) form of cobalamin. Domain (iii) housing the Co(III) methylcobalamin then interacts with the Hcy binding domain (i) to generate methionine and Co(I) cobalamin. To facilitate the passage and transfer of the methyl group between the various substitutents of the catalytic cycle, the enzyme has to undergo large-scale conformational changes.18 Moreover, in the presence of oxygen, approximately once every 200 turnovers, the cobalt ion in cobalamin becomes oxidised to the Co(II) form, which is catalytically inactive. The enzyme is able to reactivate itself through its AdoMet binding domain (iv). Through an interaction with a flavoprotein, which in E. coli is flavodoxin and in humans is methionine synthase reductase, the cobalt ion is reduced back to the Co(I) form. This is rapidly methylated by AdoMet to generate methylcobalamin and allows the catalytic cycle of the enzyme to be restored.15 2.1.3 Fatty acid biosynthesis and gluconeogenesis: the roles of biotin and pantothenate. Fatty acid biosynthesis is a repeated series of reactions that incorporate acetyl moieties (two-carbon units) of acetyl-CoA into an acyl group to form a 16- or 18carbon-long chain. Both biotin and pantothenate are essential as cofactors for the enzymes involved in this process (Fig. 4), acetylCoA carboxylase (ACCase), a biotin-dependent enzyme, and fatty acid synthase (FAS). FAS catalyzes a set of repetitive reactions including condensation of the two-carbon units with the growing fatty acyl chain, then reduction, dehydration and reduction again to obtain a fully reduced acyl group. FAS is a complex system working either as a multifunctional enzyme characterised by large subunits (animal, yeast) or as individual proteins functioning much like a metabolic pathway (plants, most bacteria). In both cases, the FAS system requires an essential protein cofactor: the acyl-carrier protein (ACP), though in the multifunctional 952 | Nat. Prod. Rep., 2007, 24, 949–962

Fig. 4 The role of pantothenate and biotin in fatty acid synthesis. A, Schematic representation of the fatty acid synthesis pathway showing where the two vitamins are required; B, reaction catalyzed by the acetyl-CoA carboxylase (ACCase) showing the role of biotin in the ATP-dependent activation of CO2 (in the form of HCO3 − ) and its transfer to acetyl-CoA; C, the structure of CoA and ACP showing the pantothenate unit (boxed) of the molecules.

This journal is © The Royal Society of Chemistry 2007

Published on 20 August 2007. Downloaded by Harvard University on 19/11/2013 20:46:55.

View Article Online

enzyme the ACP is incorporated into the enzyme as one of its domains.

ACP is a small protein of about 80 amino acids to which is bound a 4 -phosphopantetheine moiety, derived from pantothenate (Fig. 4c). This group is also found in coenzyme A (CoA), and one role of the cofactor is to solubilise hydrophobic acyl groups.

Moreover, the thioester linkage between the thiol group of 4 phosphopantetheine and the acyl group provides a good leaving group for the formation of C–C bonds. As well as its role in fatty acid synthesis, CoA plays an important role in key metabolic reactions such as the TCA cycle and secondary metabolism including lignin, flavonoid and terpenoid biosynthesis. In E. coli, 4% of the known enzymes have been shown to require CoA.19 4 Phosphopantetheine is also the prosthetic group of polyketide and non-ribosomal peptide synthases,20 involved in the biosynthesis of many antibiotics, toxins and pigments.

Biotin on the other hand has a much narrower role as the cofactor for a limited number of enzymes in central metabolism (see below for other functions). In all cases it is covalently bound to its partner enzymes, and serves as a CO2 carrier between bicarbonate and the acceptor substrate (Fig. 4b). NCarboxybiotin is first produced from HCO3 − and ATP, and this activated form of CO2 is then transferred to the substrate, in this case acetyl-CoA, to produce malonyl-CoA, the elongating unit.

Beside ACCase, there are three other metabolically important carboxylases already widely reviewed.21 The first one is the pyruvate carboxylase, which catalyses an important step in gluconeogenesis by converting pyruvate into oxaloacetate. This reaction also serves to replenish the oxaloacetate pool available for the tricarboxylic acid cycle. The second one is the propionyl-CoA carboxylase that allows propionate (arising from the catabolism of odd-chain fatty acids and branched-chain amino acids) to enter the citric acid cycle, by converting propionyl-CoA into methylmalonylCoA. Methylmalonyl-CoA is then transformed into succinate by two further steps. Finally, 3-methylcrotonyl-CoA carboxylase converts 3-methylcrotonyl-CoA, an intermediate in the catabolism of leucine, into 3-methylglutaconyl-CoA. This molecule eventually results in acetoacetate and acetyl-CoA formation.

2.1.4 Redox reactions: the role of Moco. Oxido-reduction reactions are of ultimate importance in cell biology because most of the free-energy available in living organisms relies on this type of reaction. As a matter of fact, a very large variety of substrates are either reduced or oxidised in almost all aspects of cellular metabolism. These reactions involve various cofactors such as metal ions, iron–sulfur clusters, hemes, glutathiones, nicotinamides (from vitamin B3, also called niacin or vitamin PP), flavins (from riboflavin, vitamin B2), ascorbate (vitamin C) or Moco. It is not possible to describe here the role of all these molecules, and we will concentrate on the molybdenum cofactor (Moco), to which less attention has been paid up to now. There are more than 50 molybdenum-containing enzymes known, most of them of bacterial origin, that participate in essential redox reactions in the global nitrogen-cycle (involving the molybdenum enzymes nitrogenase, nitrate reductase, nitrite oxidase), the sulfur-cycle (involving sulfite oxidase and DMSOreductase) and the carbon-cycle (involving CO-dehydrogenase, aldehyde oxidases and formate dehydrogenase).22 With the exception of bacterial nitrogenase, molybdenum as a catalytically active This journal is © The Royal Society of Chemistry 2007

metal is complexed by a unique tricyclic pterin, molybdopterin,23 to produce the molybdenum cofactor, and only in this form is it inserted into its diverse target proteins. During catalysis, it shuttles between oxidation state +4 and +6, and for several enzymes the reaction mechanism has been worked out in detail.24 The task of the pterin moiety, however, is less defined. Clearly, the pterin scaffold has to position the catalytic metal correctly within the active site, thereby controlling its redox behaviour. It has also become clear that the pterin moiety participates with its ring system in the electron transfer to or from the Mo atom.

Yet, it appears that the cofactor does not participate directly in catalysis. Obviously, it plays a more indirect role in modulating the reduction potential and reactivity of the molybdenum center. The pterin, with its several possible reduction states as well as different structural conformations, could also be important for channelling electrons to other prosthetic groups.

2.1.5 Other metabolic functions for folate, biotin and cobalamin. Beside their role in C1 metabolism, folate derivatives act also as chromophores. This property arises from the aromatic nature of the pterin and p-aminobenzoyl rings. This function is exploited by a certain class of enzymes named photolyases and involved in DNA repair following UV-B light damage.25,26 UV-B light (280–310 nm) induces covalent bonds between two adjacent pyrimidines. These pyrimidine dimers have deleterious effects, including inhibition of replication and transcription, followed by growth arrest and cell death. The removal of these covalent bonds is catalyzed by photolyases, a class of enzymes using the energy of UV-A/blue light (350–600 nm), and two cofactors: a 5,10-methenyltetrahydrofolate and a flavin cofactor FAD, derived from riboflavin (vitamin B2). The folate chromophore functions as a photoantenna which absorbs the light energy and transfers the resultant excitation to the FAD cofactor that, in turn, breaks the undesired covalent bond.25,26 The search for other functions of biotin is presently a very active field.27 Early studies had shown that biotin influences the expression of some genes,28 and now more than 2000 human genes that depend on biotin have been identified. This effect is mediated by different cell signals or transcription factors: biotinyl-AMP, NF-jB, Sp1 and Sp3, receptor tyrosine kinases.

The best documented area is the biotinylation of histones. It is now established that biotin is covalently bound to histones through amide bonds with distinct lysine residues. This post-translational modification, regulated by the biotin level, influences the structure of chromatin, and hence gene expression. The biotinylation of histones appears to play a role in cell proliferation, gene silencing, and the cellular response to DNA repair. This reaction is catalyzed by biotinidase and holocarboxylase synthetase, as illustrated in Fig. 6.

Vitamin B12- or cobalamin-dependent enzymes are required in three broad classes of reactions: (i) B12-dependent isomerases, (ii) B12-dependent methyltransferases and (iii) B12-dependent Nat. Prod. Rep., 2007, 24, 949–962 | 953

Published on 20 August 2007. Downloaded by Harvard University on 19/11/2013 20:46:55.

View Article Online

reductive dehalogenases.29 In the B12-dependent isomerases, the biological form of cobalamin is adenosylcobalamin, the coenzyme form of B12.30 Here, the properties of the weak cobalt– carbon bond are exploited through homolytic cleavage, generating a 5 -deoxyadeonsyl radical and cob(II)alamin. The radical is used to promote a variety of complex 1,2-rearrangements that are largely associated with anaerobic fermentative processes.

These include the diol dehydratases of propanediol and glycerol metabolism, ethanolamine ammonia lyase and the amino mutases that function in the fermentation of glutamate, lysine, leucine and ornithine.31 This class also includes the reactions associated with the methylmalonyl-CoA mutase and the B12-dependent ribonucleotide reductase. There appear to be two subclasses of isomerases that differ fundamentally on how the coenzyme binds to the enzyme. In class I enzymes, the bond between the lower base of B12, the dimethylbenzimidazole (DMB) group and the cobalt ion is replaced by a link with an imidazole side chain of a specific histidine residue, and thus the coenzyme is said to bind in a baseoff conformation. In class II enzymes, the DMB link is retained, and the coenzyme is said to bind in a base-on conformation.29 Methylcobalamin is associated with methylation reactions, as exemplified and described above with methionine synthase, and thus this form of cobalamin plays an important role in amino acid metabolism as well as in one-carbon metabolism.29 Apart from methionine synthase, B12-dependent methyltranferases play essential roles in anaerobic microbiology through participation in methanogenesis32 and acetogenesis, where the enzymes are able to accept methyl groups from a broad range of donors and pass them onto specific receptors.29 In the final class of B12-dependent reaction, B12 is associated with the reductive dehalogenation of aromatic and aliphatic chlorinated organics. Although it appears that the reductive dehalogenation enzymes are mechanistically quite distinct from either the isomerase or the methyltransferases, the role of cobalamin in the dehalogenation process has yet to be fully elucidated.29 2.2

The main differences among eukaryotic organisms

2.2.1 Compartmentalisation. In eukaryotic cells, metabolic pathways may be split or shared between several compartments.

Compartmentalisation is not necessarily the same in all eukaryotes. For example, plant cells have plastids, a unique compartment having important biosynthetic functions, which makes the distribution and cellular trafficking of numerous metabolites even more complex in these organisms. a) Folate and C1 metabolism. Only bacteria, plants and lower eukaryotes (yeast, protozoa) have a complete biosynthetic pathway for THF. In plants, this pathway involves the cytosol for the synthesis of pterin, the plastids for the synthesis of paminobenzoate and the mitochondria for the assembly of the different parts of the molecule33 (see also the review by Webb and Smith in this issue). In yeast and protozoa the situation is not so clear, but some of the proteins that are present in the mitochondria of plants appear also to be associated with mitochondria in yeast.34 In all eukaryotes, folate and C1 metabolism are compartmentalised between the cytosol and organelles.35 In yeast and animals, the C1 units required to sustain C1 metabolism are produced from serine or formate in the cytosol as well as in the mitochondria.

However, the syntheses of purines and dTMP are mostly present 954 | Nat. Prod. Rep., 2007, 24, 949–962

in the cytosol where the methylation cycle is also exclusively located. Thus, most C1 units produced in the mitochondria are exported to the cytosol to sustain these activities. In plants, the situation appears more complex for several reasons. Firstly, plants have plastids that also contain SHMT, the enzyme involved in the generation of C1 units.4 Secondly, whereas the methylation cycle appears to be exclusively located in the cytosol, as it is in other organisms, nucleotide synthesis in plants is also located in organelles: purines are essentially synthesised in the plastids,36 whereas dTMP is produced in the mitochondria, the plastids and possibly the cytosol, as suggested by the presence of thymidylate synthase in these compartments.33 b) Biotin. Biotin synthesis in plants and bacteria appears to follow a similar pathway from pimeloyl-CoA37 (see review by Webb and Smith in this issue), but in plants the first three enzymes are cytosolic and the terminal enzyme, biotin synthase, is mitochondrial. The Arabidopsis BIO2 gene encoding one of the subunits of this enzyme has sequence similarity with the bacterial enzyme, but appears to encode an N-terminal extension not present in the latter. This extra sequence acts as a mitochondrial targeting peptide. Interestingly, when a truncated BIO2 construct was introduced into the Arabidopsis bio2 mutant, so that the protein was cytosolic, it was unable to complement the mutant, even when plants were fed with the substrate dethiobiotin.38 It is likely that one or more mitochondrial proteins are necessary for biotin synthase activity, since the enzyme appears to be a large multisubunit complex. It may also reflect the need for correct assembly of the Fe–S centre on the enzyme, discussed in more detail in the reviews by Marquet et al. and Mendel et al. in this issue. However, in other eukaryotic organisms, such as yeast, the situation appears different. Saccharomyces cerevisiae apparently contains only the three last enzymes of the pathway, namely diaminopelargonic acid aminotransferase, dethiobiotin synthase and biotin synthase,39 whereas Schizosaccharomyces pombe contains only the biotin synthase gene.40 As in plants, the biotin synthase activity is associated with the mitochondrial compartment. Because of these truncated pathways, these two organisms are auxotrophic for biotin, and they have developed specific transporters allowing growth in the presence of the appropriate biotin intermediates. c) Pantothenate. The first enzyme of pantothenate biosynthesis, ketopantoate hydroxymethyltransferase (KPHMT), uses 5,10-methylenetetrahydrofolate as cofactor. The enzyme from Arabidopsis was found to be synthesised with an N-terminal extension with the characteristics of a mitochondrial targeting sequence, and the terminal part of folate biosynthesis is also located in this compartment. The location of KPHMT was confirmed as mitochondrial using GFP fusion experiments.41 In the same paper, pantothenate synthetase, the final enzyme of the pathway, was shown to be cytosolic, confirming earlier work that found no evidence for targeting sequences in this protein.42 These results imply that either ketopantoate or pantoate must be transported out of the mitochondrion for the final step, but to date nothing is known about possible transporters. In contrast, in the yeast S. cerevisiae, both KPHMT and pantothenate synthetase have putative mitochondrial targeting sequences at the N-terminus.42 d) Moco. In plants and humans, molybdenum cofactor and the enzymes that use it are synthesised in the cytoplasm. As some of This journal is © The Royal Society of Chemistry 2007

View Article Online

Published on 20 August 2007. Downloaded by Harvard University on 19/11/2013 20:46:55.

the Moco-dependent enzymes are localised in the mitochondria or the peroxisomes,23 they are transported after synthesis to their final cellular compartments.43 2.2.2 Specific needs in some eukaryotes. Some organisms have particular metabolic functions (for example, autotrophic organisms, such as plants, have developed several metabolisms specifically related to the photosynthetic activity) and might have specific needs for one cofactor or another. a) Folate and C1 metabolism. In plants there is a specific pathway associated with photosynthetic carbon assimilation and connected to C1 metabolism: the photorespiratory pathway.44 This pathway is initiated at the level of ribulose 1,5-bisphosphate carboxylase/oxygenase (Rubisco). This is a bifunctional chloroplastic enzyme that catalyses both the carboxylation and oxygenation of ribulose 1,5-bisphosphate. The carboxylation reaction leads to the production of two molecules of 3-phosphoglycerate, whereas the oxygenation reaction leads to one molecule of 3-phosphoglycerate and one molecule of 2-phosphoglycolate. This oxygenation reaction is the primary event of a metabolic pathway called photorespiration because it is associated with the uptake of O2 and the evolution of CO2 . The wasteful nature of the process has resulted in the evolution, in certain plant lineages, of so-called C4 -metabolism, which avoids the oxygenation reaction of Rubisco. In leaves of C3 -plants where photorespiration occurs, there is a recycling of two molecules of 2-phosphoglycolate into one molecule of 3phosphoglycerate involving three different organelles: the chloroplast, the peroxisome and the mitochondrion. The key steps of this pathway take place in mitochondria (Fig. 5) where glycine is either oxidised or converted into serine by two folate-dependent reactions that are intimately coupled. These two reactions are catalyzed

Fig. 5 Schematic representation of the photorespiratory pathway emphasising the role of the folate dependent reactions within the mitochondria.

1) GDC: glycine decarboxylase complex; this complex is constituted of 4 subunits, the P, H, T and L proteins. The T-protein is the folate-dependent protein that catalyzes the transfer of the C-2 of glycine to THF, leading to the release of NH3 and the formation of 5,10-CH2 -THF. 2) SHMT; SHMT catalyzes the conversion of a second molecule of glycine and 5,10-CH2 -THF into THF and serine.

This journal is © The Royal Society of Chemistry 2007

by the glycine decarboxylase complex (GDC) and SHMT.45 In these reactions, GDC catalyzes the oxidative cleavage of one molecule of glycine, providing the 5,10-methylenetetrahydrofolate required for the conversion of a second molecule of glycine into serine by the reverse activity of SHMT. These two enzymatic systems exist in trace amounts in the mitochondria from non photosynthetic organisms, where they are involved in glycine catabolism. In leaves of C3 -plants, however, where high photorespiratory activity requires high GDC and SHMT activities, these enzymes represent 40% of the soluble mitochondrial proteins. This is possibly why the folate concentration in these mitochondria is higher than in the other cellular compartments.46 It is also possible that, in C3 -plants, glycine represents a potential source of C1 units. b) Cobalamin and methionine synthase. Vitamin B12 is unusual in that its synthesis is confined to prokaryotic organisms.47 Humans require this vitamin as a cofactor for methionine synthase (MetH) and methylmalonyl-CoA mutase, involved in the catabolism of odd-chain fatty acids. Higher plants do not make (or require) vitamin B12 because they have the alternative, cobalamin-independent methionine synthase called MetE. The absence of vitamin B12 in higher plants means that, unlike other vitamins, the major source of vitamin B12 in our diets is from animal-derived products. Thus people who follow strict vegetarian regimes can easily become deficient. A particularly rich dietary source of this vitamin is seaweed or macroalgae, such as nori (Porphyra yezoensis), which is commonly used to wrap sushi.

Recent work on the metabolism of cobalamin in algae has shown that like all other eukaryotic organisms, algae are not able to synthesise this vitamin de novo.48 Instead, over half of all algal species are like animals, in that they have a requirement for an external source of B12, which is needed as a cofactor for MetH.48 Algae that do not need exogenous cobalamin contain MetE, like higher plants. In some cases, such as the model green alga Chlamydomonas reinhardtii, genes for both isoforms are present, and the alga can use MetH if cobalamin is present, in which case the MetE gene is turned off.48 This is analogous to the situation in E. coli, where in the absence of a supply of cobalamin MetE accumulates to 3% of cellular protein, demonstrating that the cobalamin-dependent enzyme MetH is more efficient.17 Although the dinoflagellate Phaeodactylum tricornutum has been reported to contain methylmalonyl-CoA mutase, it does not require vitamin B12 for growth, indicating that this enzyme is unlikely to be the reason for the widespread auxotrophy.

Intriguingly, the levels of free cobalamin in the environment are insufficient to support the growth of auxotrophic algae, and evidence has been obtained that bacteria can supply the vitamin directly in exchange for fixed carbon, in an apparently symbiotic interaction. This may explain the observation that there is no phylogenetic relationship between those algae that require cobalamin, indicating that the loss of the MetE gene occurred multiple times during algal evolution, which in turn implies the absence of selection pressure to retain the gene. A ready supply of micronutrients from close association with bacteria is further borne out by the observation that both thiamin and biotin auxotrophy are found within the algal kingdom, although at a lower frequency.49 In these cases, however, it appears that the requirement for the vitamin has arisen as a result of the loss of one or more genes for biosynthetic enzymes.

Nat. Prod. Rep., 2007, 24, 949–962 | 955

View Article Online

Published on 20 August 2007. Downloaded by Harvard University on 19/11/2013 20:46:55.

3

Nutritional aspects

The diet must provide all the micronutrients required to alleviate nutritional disorders and to promote good health. The recommended dietary allowances (RDAs) indicate for each micronutrient the minimal intake needed to avoid nutritional disorders, although higher intakes might be required for optimal health.

Conversely, for some vitamins, excess intake can be deleterious, as is known for vitamin A and pyridoxal (vitamin B6). Requirements fluctuate widely depending on the micronutrient, with the recommended daily intakes ranging from a few lg (cobalamin) to several mg (pantothenate). Deficiencies among the population are not only found in developing countries but also in developed countries where bad food habits lead to suboptimal intakes.

Indeed, micronutrient levels vary widely depending on the food source, and good diets are made of multiple sources. In the following sections the main dietary sources and the effects of a deficiency on human health are described.

3.1

exceptional affinity for biotin), individuals maintained on total parental nutrition, and those receiving long-term anticonvulsant therapy.

Inherited deficiencies have also been recognised, the most frequent ones being a deficiency in biotinidase (the occurrence of which is 1 in 60 000) and holocarboxylase synthetase (HCS).56 These two proteins are involved in the recycling of biotin (see Fig. 6). Although these inherited deficiencies are rather rare, they can have severe, even fatal, consequences if not treated. They are mainly associated with a modification of the biotinidase and HCS activities, either because the K M for biotin is increased or because the availability of free biotin is decreased.

Effects of deficiency on human health

a) Folate. Folate deficiency is one of the most prevalent vitamin deficiencies worldwide and leads to a number of serious diseases.

Deficiency of folate would be expected to produce a reduction of the cell capacity to synthesise DNA and thus to maintain a normal rate of cell division. Indeed, a poor folate status is often correlated with a high cellular dUMP/dTMP ratio due to limiting supply of 5,10-methylenetetrahydrofolate and a decrease of dTMP synthesis.

The increased dUMP/dTMP ratio results in higher incorporation of dUTP in DNA, which generates point mutations, singleand double-strand DNA breaks, and ultimately chromosomal breakage.50 These damages are risk factors for a number of cancers such as colorectal, breast, pancreatic, bronchial, and cervical cancer, as well as leukaemia. Cells undergoing rapid division, such as those of the bone marrow, are likely to be more affected. From this point of view, one of the most obvious consequences of folate deficiency is megaloblastic anemia, which probably results from apoptosis of erythroblasts.51 It has also been shown that some neural tube defects, such as spina bifida, occurring in the early period of embryogenesis, are caused by a shortage of folate.52 Inefficient synthesis of methionine may also have several repercussions. The first one is the accumulation of homocysteine, the precursor of methionine. It is generally accepted that high plasma levels of homocysteine are a primary cause of higher risk for coronary and cardiovascular diseases.53 Likewise, high levels of homocysteine in plasma are a risk factor for dementia and Alzheimer’s disease.54 Secondly, a low level of methionine results in insufficient amounts of AdoMet available for all the methyltransferase-catalyzed reactions. Although the resulting problems are less well defined, they might include neuropathy due to impaired myelin biosynthesis. Also, hypomethylation of DNA, especially hypomethylation of gene promoter regions, may alter gene expression resulting, for example, in elevated expression of some oncogenes.50 b) Biotin. The biotin requirement of most organisms is low and no severe biotin deficiency has been observed in humans, except in cases of genetic diseases.55 The well-known ‘nutritional’ biotin deficiency concerns people eating a diet very rich in raw eggwhite (which contains a large amount of avidin, a protein with an 956 | Nat. Prod. Rep., 2007, 24, 949–962

Fig. 6 The biotin cycle. Biocytin is formed from the degradation of biotinylated proteins, such as holocarboxylases, histones etc. Biotin is released from biocytin by biotinidase. Depending on the pH, the biotinidase activity results either in its own biotinylation or in the production of free biotin. Either free biotin or biotin from biotinylated biotinidase are later used to biotinylate new proteins, through reactions catalysed by holocarboxylase synthetases.

The main consequence of a biotin deficiency is a multicarboxylase deficiency, with expected metabolic consequences such as organic aciduria. Many other signs, such as dermatitis, conjunctivitis, ataxia, different neurological disorders, developmental delay, etc., are often associated with biotin deficiency, although it is difficult to establish a direct correlation. These signs remain mysterious at the biochemical level, and it is likely that proteins other than carboxylases are affected. From this point of view, it has been shown recently that biotin controls the expression of various proteins.28 The physiological and biochemical consequences of biotin deficiency are thus far from being completely understood. c) Cobalamin. Cobalamin deficiency is associated with pernicious anaemia, which is a form of anaemia characterised by defective production of erythrocytes and the presence of megaloblasts in the bone marrow. The condition is sometimes associated with neurological disorders. A lack of B12 is also thought to be an independent risk factor associated with neural tube defects in unborn babies, a commonality shared with folate deficiency.

The link between folate and B12 deficiency is likely explained by their shared responsibilities in methionine metabolism. A lack of B12 affects the two human enzymes that require it, namely methionine synthase and methylmalonyl-CoA mutase, and gives rise to elevated levels of homocysteine and methylmalonic acid, respectively. The high levels of the latter allow B12 deficiency to be differentiated from simply a lack of folate.

This journal is © The Royal Society of Chemistry 2007

Published on 20 August 2007. Downloaded by Harvard University on 19/11/2013 20:46:55.

View Article Online

The major reason for B12 deficiency is associated with problems with the B12 absorption system. B12 is absorbed from the intestine by binding of the cobalamin to a secreted glycoprotein called intrinsic factor. The bound complex interacts with a specific receptor on the mucosa of the ileum, where it is internalised. B12 is released and transferred to another protein called transcobalamin, which is responsible for transport of the vitamin around the body. Some congenital forms of pernicious anaemia are due to defects in this uptake system, most notably with intrinsic factor.57 More significantly, some people, especially the elderly, no longer produce intrinsic factor, due to atrophy of the mucosa of the stomach, which appears to be brought about by autoimmune factors.58 However, the deficiency in B12 can be easily treated by intramuscular injections of vitamin B12 or by very large oral doses of cobalamin. d) Pantothenate. The word pantothenate is derived from the Greek word pantos, meaning everywhere, due to the presence of the vitamin in most foodstuffs,59 a fact that is supported by a lack of pantothenate deficiencies reported in nature. However, experimental induction of dietary pantothenate deficiencies, by treatment with x-methylpantothenate, a metabolic antagonist, has been shown to produce varying symptoms in different organisms.

For example, animals exhibit depression, sleep disturbances, personality changes, cardiac instabilities, and neurological disorders such as ‘burning feet’ syndrome, amongst others. Pantothenate deficiency observed in chicks was shown to cause an outbreak of dermatitis around the eyes and beak, and resulted in the vitamin being referred to as chick anti-dermatitis factor because treatment with pantothenate alleviated the problem.60 Due to the absence of pantothenate deficiencies in the general population, there is no minimum dietary requirement as an adequate amount of pantothenate can be obtained by eating a balanced diet (www.foodstandards.gov.uk/multimedia/pdfs/panto.pdf). e) Moco. A mutation in the biosynthetic pathway of Moco has dramatic consequences for the cell because pleiotropically all enzymes needing Mo lose their activity at the same time. In humans, a combined deficiency of Mo-enzymes was first described

by Duran et al.61 Babies born with this defect show feeding difficulties, severe and progressive neurological abnormalities, and dysmorphic features of the brain and head. So far, disease-causing mutations have been identified in three of the four known Mocobiosynthetic genes in humans: mocs1, mocs2 and gephyrin.62 The clinical symptoms may result from the deficiency of sulfite oxidase that protects the organism, in particular the brain, from elevated levels of toxic sulfite.63 No therapy is currently available to cure the symptoms of this disease. Moco deficiency cannot be treated by supplementation with the cofactor. Moco is extremely unstable outside the protecting environment of an apo-Mo-enzyme; its half-life is only a few minutes in aqueous solution at neutral pH.64 In addition, no chemical synthesis of Moco or any of its intermediates has been successful so far, which hampers its large-scale production for therapeutic use. However, very recently, a model has been developed that could lead to a cure of Moco-deficiency. Genetic analyses of patients showed that most of them had defects in the first step of Moco biosynthesis, i.e. the conversion of GTP to cyclic pyranopterin monophosphate, pPMP.65 The idea was to treat patients of this class with the missing intermediate pPMP because the steps subsequent to pPMP formation are not affected This journal is © The Royal Society of Chemistry 2007

by the mutation and should allow the synthesis of Moco. pPMP is more stable than Moco itself and has an identical structure in all organisms. Thus, pPMP was overproduced in the bacterium E. coli and purified. MOCS1 knockout-mice with a block in the first step of Moco biosynthesis were created bearing a genetic defect identical to the human patients.66 Similar to humans, heterozygous mice displayed no symptoms, but homozygous Moco-deficient animals displayed symptoms resembling those of the human deficiency state and died within ten days after birth. Due to the mutation, no molybdopterin or active Moco was detectable, and consequently all Mo-enzyme activities were absent.

Repeated injections of pPMP into MOCS1-deficient mice resulted in a dose-dependent extension of life span.67 Molybdopterin (MPT) levels and Mo-enzyme activities were partially restored.

Stopping pPMP treatment at any time resulted in a progressive reduction of MPT levels and Mo-enzyme activities, and death of the animal 10–15 days after receiving the last injection. Injection of pPMP into these mice every second day normalised their symptoms, and they reached adolescence and were fertile.67 It remains to be seen whether delayed onset of the described therapy will still allow reversal of neurological damage. As a next step, scaling up of pPMP production is in progress in order to have sufficient amounts available for clinical trials.

3.2 Main dietary sources a) Folate. The recommended dietary allowance for folates is currently 400 lg for an adult and increases to 600 lg for pregnant women (Food and Nutrition Information Center, 2004). Although folate is quite abundant in liver, which plays an important role in folate metabolism and storage in animals, plant food is by far the biggest contributor to the folate intake of adults (Table 1).

However, levels vary considerably depending on the plant species and the nature of the tissues.

Folate synthesis in plants is tightly controlled and fluctuates depending on the metabolic requirements. This implies that the folate content will vary from one tissue to another and as a function of plant development. Generally, folate is more abundant in actively dividing tissues, such as embryos and meristematic tissues. These observations fit well with a high activity of nucleotide synthesis in rapidly dividing tissues and the utilisation of 5,10-methyleneand 10-formyltetrahydrofolate for the synthesis of thymidylate and purines. This is probably why the folate content in the embryo part of the seed increases further during germination and is often much higher than in cotyledons.46 Folate concentration also increases in leaves during development, and this accumulation was correlated with the build-up of the photosynthetic apparatus. The relationship between folate accumulation in leaves and photosynthesis is not yet understood. Part of the folate synthesised in leaves might contribute to the photorespiratory process in mitochondria, but most of it accumulates in a ‘cytosolic fraction’.46 In the cytosol, methyl-THF, the predominant derivative, is likely required for Met synthesis and turnover of AdoMet, thus suggesting a high C1 metabolic activity associated with photosynthesis. In any case, the relationship between photosynthesis and C1 metabolism probably explains why green leafy vegetables are a good source of folate.

By contrast, roots (such as carrots), storage organs (potatoes) and most fruits are poor sources (Table 1). Finally, it must be kept in mind that the folate content will depend on the way the food is Nat. Prod. Rep., 2007, 24, 949–962 | 957

View Article Online

Published on 20 August 2007. Downloaded by Harvard University on 19/11/2013 20:46:55.

Table 1 Amounts of folate, biotin, pantothenate and cobalamin found in a range of food. Values are given in lg per 100 g food portion. Data for folate, pantothenate and cobalamin were adapted from the United States of Agriculture National Nutrient Database for Standard Reference (www.nal.usda.gov/fnic/foodcomp/search). Data for biotin were adapted from Staggs et al.107 Food

Folate

Biotin

Pantothenate

Cobalamin

Spinach, raw Broccoli, raw Carrots, raw Potatoes, baked Orange Tomatoes Lentils, cooked Rice, white, cooked Cow’s milk Beef, cooked Haddock, cooked Liver (beef), cooked Egg, whole, hard-boiled

193 62 19 9 30 15 181 97 5 7 12.9 260 44

0.7 0.95 0.6(canned) 0.2 0.05 0.7 — — 0.1 4.5(hamburger) 0.7 41.6 21.4

70 570 273 519 250 88 630 390 362 676 150 6940 1400

0 0 0 0 0 0 0 0 0.44 1.76 1.39 83 1.10

stored, processed or cooked, the different forms of folate differing in their susceptibility to loss (for a review see ref. 68). b) Biotin. The recommended daily allowance of biotin is 150 lg (EU) and 300 lg (US). Biotin deficiency is rare because these relatively low requirements are generally covered by a common diet. Biotin present in food (Table 1) is covalently linked to proteins. During the digestive process, these proteins are degraded by peptidase, and the final steps of these proteolyses lead to biocytin, i.e. biotinyl-lysine or biotinyl-lysyl peptides. Biotin is liberated from these products by biotinidase, a specific enzyme present in the pancreatic juice. Biotin is then transported in blood, either as free biotin (20%) or linked to the biotinidase (80%). Once in the cell, biotin is activated as biotinyl-AMP before being attached to apocarboxylases (or other proteins such as histones), a reaction catalyzed by holocarboxylase synthetase. This set of reactions (degradation of the holoenzyme by peptidases, biotinidase activity and attachment to new carboxylases, see Fig. 6) also exists within cells. Such a cellular biotin cycle explains the low requirement for biotin from foodstuffs, especially taking into account that some biotin is also produced by the intestinal flora. c) Cobalamin. B12 is required in very small amounts by humans.

The RDA is around 2 lg per day. Dairy products are good sources of B12, as are meat and eggs.69 Vegetables do not contain B12, as higher plants neither make nor require it for their metabolism.

Consequently, those who adhere to strict vegetarian diets are prone to becoming B12-deficient, and dietary deficiency of B12 due to vegetarianism is increasing. One way to counteract this is for these people to eat certain types of macroalgae (seaweed) that are rich in B12 (see above). Prokaryotes are the world’s suppliers of vitamin B12, and many enteric bacteria are able to produce cobalamin.

Although ruminants are able to absorb the B12 produced by their enteric bacteria, the human B12 uptake system does not allow the uptake of the B12 produced in the lower intestine. Thus the vast majority of B12 is obtained by humans from their diet. d) Pantothenate. Pantothenate is found ubiquitously in food including meat, vegetables, mushrooms, fish and eggs, and therefore there is no known dietary deficiency. Moreover, there is no RDA for pantothenate, and it is generally considered to be safe to consume in large quantities. Indeed this has led to various therapeutic treatments with large doses of the vitamin, including for acne and in facilitating weight loss.

958 | Nat. Prod. Rep., 2007, 24, 949–962

e) Moco. Molybdenum in the form of molybdate is abundant in water and food, and hence there are no nutritional shortages considered. The recommended dietary allowance for molybdenum is currently 50 lg for an adult, and in most countries the amount of molybdenum taken up by adults widely exceeds this allowance.

3.3 Strategies for enhancement a) Folate. The fact that folate levels vary greatly in different plant species (Table 1), implies that there is a natural potential for folate enhancement. Different strategies can be followed to achieve the goal of folate biofortification. These strategies have already been discussed in several reviews.33,70 In summary, they can be divided into two main groups: exploiting the natural variation in folate levels and metabolic engineering.

The first approach relies on a selection process based on molecular mapping techniques: the quantitative trait loci (QTL) responsible for folate accumulation can be identified and used in molecular-marker-assisted breeding programs. However, it must be kept in mind that natural variations existing between different varieties of one particular plant species are unlikely to be very large. Nevertheless, for plants that intrinsically have a low folate level, such as rice, an increment factor of ten, at least, should be reached to obtain a truly improved crop. A second difficulty is that such an approach requires a simple, fast, reliable and preferably high-throughput procedure for folate determination.

Several methods of folate analysis in plants have been established (reviewed in ref. 71) but, up till now, none of them meet these requirements.

The metabolic engineering approach implies genetic modifications of plant folate metabolism. These modifications could target the folate biosynthetic pathway itself, the reactions affecting the stability and storage, and the reactions involved in the breakdown and recycling of the cofactor.

Concerning the biosynthetic pathway, a promising route is the simultaneous enhancement of the synthesis of the two main folate precursors, the H2 pterin and the p-aminobenzoic acid.

This hypothesis has recently been confirmed in two independent attempts to engineer pterin biosynthesis. These two attempts aimed to express GTP cyclohydrolase I (GTPCHI), the first enzyme involved in pterin synthesis. In the first experiment, the authors expressed in Arabidopsis the GTPCHI from E. coli, an This journal is © The Royal Society of Chemistry 2007

Published on 20 August 2007. Downloaded by Harvard University on 19/11/2013 20:46:55.

View Article Online

enzyme that is not subject to metabolic regulation,72 whereas in the second experiment the authors expressed in the tomato the mammalian enzyme, which is predicted to escape feedback control in plants.73 In both experiments the transgenic lines contained several hundred times more pterin but only two to four times more folate, indicating that GTPCHI was indeed limiting but that other factors are also regulating folate synthesis. Evidence that synthesis of p-aminobenzoic acid is another limiting step for folate accumulation was shown by the latter group: in their attempt to engineer the pterin branch of folate synthesis, they observed that the p-aminobenzoic level in fruits was depleted following GTPCHI expression and pterin accumulation. In fact, when both aminodeoxychorismate synthase, the first step in p-aminobenzoic synthesis, and GTPCHI were co-expressed in tomato fruits, the transgenic fruits contained 20 times more folate than the control.74 Folate derivatives are rather unstable molecules that undergo spontaneous oxidative degradation. However, the stability of the cofactor is considerably increased when bound to folate-dependent proteins. Thus, stabilisation of folate can possibly be achieved by over-expressing folate-binding proteins (FBP) in plant cells, such as the mammalian FBP found in milk.75 Other proteins having a high affinity for folate, for instance the T-protein of GDC,76 could also be tested. These strategies would not only stabilise folate but also create a folate sink, resulting presumably in a stimulation of synthesis. Re-routing of C1 metabolism towards accumulation of the most stable folate derivatives may be yet another approach to stabilise the folate pool. The most stable form of folate is 5-formyltetrahydrofolate. No function in C1 metabolism has been assigned to this molecule so far: it is thought to play a role as a storage form or to have a regulatory function, inhibiting some enzymes of C1 metabolism, such as mitochondrial SHMT. The only enzyme that uses 5-formyltetrahydrofolate is 5formyltetrahydrofolate cycloligase (5-FCL), which catalyzes the ATP-dependent reverse conversion of 5-formyltetrahydrofolate to 5,10-methenyltetrahydrofolate.77 Recently, an Arabidopsis 5-FCL knockout mutant has been characterised,78 but only a two-fold increase was observed in the folate pool.

Reducing the rate of folate catabolism might also lead to higher folate accumulation. Indeed, it was recently shown that plants can have high folate-breakdown rates, approximately 10% per day.

This breakdown involves oxidative cleavage of the molecule, giving p-aminobenzoylglutamate and pterin. However, most of these degradation products are recycled as folate precursors,79 and it is not certain that engineering the catabolic pathway alone would be sufficient to significantly increase the folate concentration. b) Pantothenate. The overexpression of elements of the pantothenate pathway as a mechanism for increased vitamin production has been explored in several organisms. Enhanced expression of pantothenate to some extent was achieved by the overexpression of three of the four enzymes: in E. coli K12 ketopantoate reductase overexpression leads to increases in cellular pantothenate levels,80 whilst in Salmonella enterica serovar Typhimurium overexpression of ketopantoate hydroxymethyltransferase is sufficient to increase cellular pantothenate levels.81 In plants, the supply of b-alanine, which in these organisms is via a different pathway, seems to be a limiting factor. Overexpression of the bacterial aspartate a-decarboxylase in tobacco leaves resulted in an increase in both b-alanine and pantothenate levels.82 Strains of E. coli that accumulate up to 65 g l−1 of pantothenate in culture have been This journal is © The Royal Society of Chemistry 2007

isolated. These strains, which produce pantothenate upon supplementation with b-alanine, were generated by UV-induced random mutagenesis and identification of strains resistant to several potential antimetabolites including salicylate, a-ketoisovalerate and b-hydroxyaspartate.83 Overexpression of the first enzyme in branched chain amino acid biosynthesis, acetohydroxyacid synthase isozyme II, led to a further increase in pantothenate production.84 The most comprehensive study of the pantothenate pathway in a single organism for the purpose of pathway engineering has been in the commercially important Corynebacterium glutamicum.

Early work with this organism demonstrated that the supply of b-alanine via aspartate a-decarboxylase was the limiting factor in pantothenate accumulation.85 Overexpression of the native C. glutamicum protein leads to pantothenate production equivalent to that observed when supplying the product of the enzyme, b-alanine. Sahm and Eggeling enhanced the flux through the other branch of the biosynthetic pathway by overexpressing genes for valine biosynthesis (ilvBNCD) in tandem with the C. glutamicum panBC operon in a strain unable to synthesise isoleucine (ilvA).86 In the presence of exogenous b-alanine this strain was able to accumulate up to 1 g l−1 of pantothenate. The third step in the biosynthetic pathway (ketopantoate reductase) is solely encoded by ilvC in this organism, and so the tandem overexpression of the valine biosynthetic pathway also enhanced the rate of this transformation.87 A metabolic network analysis of this overproduction strain demonstrated that the flux from the branch point between valine and pantothenate biosynthesis was 10-fold more favourable for valine biosynthesis. Chassagnole and colleagues attempted to increase the proportion of pantothenate over valine by the use of nitrogen-limiting conditions, but this led to the production of a range of non-nitrogenous compounds and accumulation of glycine.88,89 This suggests that the limiting factor might be the regeneration of 5,10-methyleneTHF for ketopantoate synthesis. More recently, the addition of an ilvE mutation, to prevent valine biosynthesis, together with multiple copies of the panBC operon on an expression vector, have been used to increase the pantothenate levels further.90 As yet, however, the levels achieved do not match those observed in the E. coli overproduction strains. c) Cobalamin. Cobalamin is one of the few vitamins that is produced commercially by bacterial fermentation.91 This is because the chemical synthesis of the vitamin is far too complex to contemplate for industry. There has thus been some significant research into investigating the molecular genetics and biochemistry of cobalamin biosynthesis with a view towards enhancing the production of B12 by bacteria. Indeed, it was this ˆ approach developed by the Rhone-Poulenc company that led to the elucidation of the aerobic biosynthetic pathway for vitamin B12,92 and which allowed the company to take out several patents on maximising cobalamin synthesis. ˆ The research undertaken by Rhone-Poulenc gave rise to a production strain of Pseudomonas denitrificans, which was generated using the acquired information on cobalamin biosynthesis.

Over a two year period, this resulted in an increase in B12 in P. denitrificans of approximately 100-fold. There are over thirty genes involved in cobalamin biosynthesis and the detailed patents suggest that at least 10 of these genes have had their copy number increased. For instance, it was noted that increasing the Nat. Prod. Rep., 2007, 24, 949–962 | 959

Published on 20 August 2007. Downloaded by Harvard University on 19/11/2013 20:46:55.

View Article Online

cobF–cobM gene cluster increased cobalamin production by 30%, whereas amplification of cobA and cobE resulted in a further 20% enhancement. Moreover, it is likely that gene transcription and translation elements were modified to enhance the production of the pathway enzymes. The researchers also looked to engineer the pathway by replacing one of the key regulatory enzymes of the pathway that displayed substrate inhibition with a variant that no longer had this unwanted characteristic. Thus enhanced flux along the pathway could be achieved by replacing the uroporphyrinogen methyltransferase with a variant from a methanogen. One of the limiting factors for cobalamin production is the synthesis of the ˆ lower axial ligand, dimethylbenzimidazole (DMB). The RhonePoulenc researchers found that overproducing a protein called BluB gave rise to significantly enhanced DMB production and increased the yield of B12.91 Only recently has it been shown that BluB is able to catalyse a remarkable oxygen-dependent transformation of reduced flavin into DMB.93 This combined approach of basic science and applied biotechnology allowed ˆ Rhone-Poulenc to produce a strain that accounts for about 80% of the world’s production of cobalamin.

4 Conclusion: compartmentalisation, a challenging area Over the course of this last decade, much progress has been made in understanding how coenzymes are synthesised. Indeed, most of the pathways involved in the biosynthesis of these molecules are now established in bacteria, lower eukaryotes and plants, although the mechanisms of a fair number of the required reactions remain to be understood in detail. Identification of the genes involved in the various steps of these metabolic routes open new possibilities of manipulating (engineering) the coenzyme content of various organisms, either for chemical production of these molecules in bacteria or to improve nutrition by modifying levels in plants.

However, the situation in eukaryotes is more complex than in bacteria because of the subcellular compartmentalisation of the pathways, which imposes the need to take into account the transport of coenzymes, intermediates and precursors from one compartment to another.

As a matter of fact, once synthesised in a given compartment, coenzymes have to be transported to the other cellular territories where they are required. In plants, all the cellular compartments are more or less involved in vitamin biosynthesis. Some of these syntheses are restricted to one compartment, but it is remarkable that vitamin biosyntheses in plants are often split between different compartments. So, pyridoxal phosphate (vitamin B6) is synthesised in two steps, exclusively located in the cytosolic compartment.94 Riboflavin (vitamin B2) synthesis is probably restricted to the plastids,95 and this also holds true for tocopherols (vitamin E)96 and carotenoids (pro-vitaminA).97 However, as mentioned above, the first step in pantothenate (vitamin B5) biosynthesis is located within mitochondria whereas the next steps are in the cytosol.98 In contrast, the first steps in biotin (vitamin B8) synthesis are in the cytosolic compartment, but the last one is associated with mitochondria,99 and this also holds true for ascorbate (vitamin C).100 As also described here, the synthesis of folate (vitamin B9) is shared between cytosol, chloroplast and mitochondria.101 Nicotinate or nicotinamide (also known as niacin 960 | Nat. Prod. Rep., 2007, 24, 949–962

or vitamin PP or vitamin B3) are the precursors of NAD(P) synthesis in animals. They are also involved, in plants, in the synthesis of pyridine alkaloids (ricinine, nicotine, trigonelline).

The pathways of NAD(P) synthesis and recycling in plants are starting to emerge and implicate the plastids, the cytosol and possibly the mitochondria.102 Thiamine pyrophosphate (vitamin B1) synthesis is poorly known in the plant kingdom. What is known is that thiamine pyrophosphate is synthesised through two different branches, one for the synthesis of the pyrimidine moiety and another for the synthesis of the thiazole moiety. It is not clear whether the thiazole moiety is synthesised in chloroplasts or mitochondria,103 but it has been clearly shown that the enzyme catalyzing the condensation of thiazole phosphate and pyrimidine pyrophosphate to produce thiamine monophosphate is exclusively located in the chloroplasts.104 Taking these new data into account, it is obvious that transport of coenzymes from one compartment to another might play an important role in the regulation of the pathways. Transport systems are always difficult to study, and this is particularly true when the carriers are present in very low amounts, as is presumably the case for those involved in the transport of coenzymes. With the exception of two folate transporters recently identified, both located in the envelope of chloroplasts,105,106 carriers involved in vitamin transport are not known in plants. As pointed out above, these transport systems are potential limiting steps for vitamin synthesis, distribution and storage, and will probably be important keys in most of the strategies used to engineer vitamin content in plants.

5 References 1 E. A. Cossins, Can. J. Bot., 2000, 78, 691–708.

2 J. M. Mouillon, S. Aubert, J. Bourguignon, E. Gout, R. Douce and F. R´ebeill´e, Plant J., 1999, 20, 197–205.

3 A. U. Igamberdiev, N. V. Bykova and L. A. Kleczkowski, Plant Physiol.

Biochem., 1999, 37, 503–513.

4 S. Ravanel, R. Douce and F. R´ebeill´e, ‘The uniqueness of tetrahydrofolate synthesis and one-carbon metabolism in plants’, in Plant Mitochondria: From Genome To Function, ed. D. A. Day, A. H. Miller and J. Whelan, Kluwer Academic, Dordrecht, 2004, pp. 277– 292.

5 P. M. Smith and C. A. Atkins, Plant Physiol., 2002, 128, 793–802.

6 M. Neuburger, F. R´ebeill´e, A. Jourdain, S. Nakamura and R. Douce, J. Biol. Chem., 1996, 271, 9466–9472.

7 A. D. Hanson and S. Roje, Annu. Rev. Plant Physiol. Plant Mol. Biol., 2001, 52, 119–137.

8 S. Ravanel, M. A. Block, P. Rippert, S. Jabrin, G. Curien, F. R´ebeill´e and R. Douce, J. Biol. Chem., 2004, 279, 22548–22557.

9 N. J. Krogan, M. Kim, A. Tong, A. Golshani, G. Cagney, V. Canadien, D. P. Richards, B. K. Beattie, A. Emili, C. Boone, A. Shilatifard, S. Buratowski and J. Greenblatt, Mol. Cell. Biol., 2003, 23, 4207– 4218.

10 R. A. Martienssen and E. J. Richards, Curr. Opin. Genet. Dev., 1995, 5, 234–242.

11 R. Surtees, J. Leonard and S. Austin, Lancet, 1991, 338, 1550–1554.

12 T. Tolstykh, J. Lee, S. Vafai and J. B. Stock, EMBO J., 2000, 19, 5682–5691.

13 R. G. Matthews, ‘Methionine biosynthesis’, in Folates and Pterins, ed.

R. L. Blakley and S. J. Benkovic, Wiley Interscience, New York, 1984, pp. 497–553, ,.

14 M. T. Croft, A. D. Lawrence, E. Raux-Deery, M. J. Warren and A. G. Smith, Nature, 2005, 438, 90–93.

15 R. G. Matthews, Acc. Chem. Res., 2001, 34, 681–689.

16 R. Pejchal and M. L. Ludwig, PLOS Biol., 2005, 3, 254–265.

17 J. C. Gonzalez, R. V. Banerjee, S. Huang, J. S. Sumner and R. G. Matthews, Biochemistry, 1992, 31, 6045–6056.

This journal is © The Royal Society of Chemistry 2007

Published on 20 August 2007. Downloaded by Harvard University on 19/11/2013 20:46:55.

View Article Online 18 J. C. Evans, D. P. Huddler, M. T. Hilgers, G. Romanchuk, R. G. Matthews and M. L. Ludwig, Proc. Natl. Acad. Sci. U. S. A., 2004, 101, 3729–3736.

19 T. P. Begley, C. Kinsland and E. Strauss, Vitam. Horm., 2001, 61, 157–171.

20 D. E. Cane and C. T. Walsh, Chem. Biol., 1999, 6, R319–R325.

21 H. G. Wood and R. E. Barden, Annu. Rev. Biochem., 1977, 46, 385– 413.

22 E. I. Stiefel, Met. Ions Biol. Syst., 2002, 39, 1–29.

23 G. Schwarz and R. R. Mendel, Annu. Rev. Plant Biol., 2006, 57, 623– 647.

24 R. Hille, Arch. Biochem. Biophys., 2005, 433, 107–116.

25 C. L. Thompson and A. Sancar, Oncogene, 2002, 21, 9043–9056.

26 C. M. Bray and C. E. West, New Phytol., 2005, 168, 511–528.

27 J. Zempleni, Annu. Rev. Nutr., 2005, 25, 175–196.

28 K. Dakshinamurti and S. Litvak, J. Biol. Chem., 1970, 245, 5600–5605.

29 R. Banerjee and S. W. Ragsdale, Annu. Rev. Biochem., 2003, 72, 209– 247.

30 E. N. G. Marsh and C. L. Drennan, Curr. Opin. Chem. Biol., 2001, 5, 499–505.

31 J. R. Roth, J. G. Lawrence and T. A. Bobik, Annu. Rev. Microbiol., 1996, 50, 137–181.

32 R. K. Thauer, Microbiology (Reading, U. K.), 1998, 144, 2377–2406.

33 F. Rebeille, S. Ravanel, S. Jabrin, R. Douce, S. Storozhenko and D. Van Der Straeten, Physiol. Plant., 2006, 126, 330–342.

34 U. Guldener, G. J. Koehler, C. Haussmann, A. Bacher, J. Kricke, D. Becher and J. H. Hegemann, Mol. Biol. Cell, 2004, 15, 3811–3828.

35 K. E. Christensen and R. E. MacKenzie, Bioessays, 2006, 28, 595–605.

36 R. Boldt and R. Zrenner, Physiol. Plant, 2003, 117, 297–304.

37 V. Pinon, S. Ravanel, R. Douce and C. Alban, Plant Physiol., 2005, 139, 1666–1676.

38 N. Arnal, C. Alban, M. Quadrado, O. Grandjean and H. Mireau, Plant Mol. Biol., 2006, 62, 471–479.

39 V. Phalip, I. Kuhn, Y. Lemoine and J. M. Jeltsch, Gene, 1999, 232, 43–51.

40 V. Phalip, Y. Lemoine and J. M. Jeltsch, Curr. Microbiol., 1999, 39, 348–0350.

41 H. H. Ottenhof, J. L. Ashurst, H. M. Whitney, S. A. Saldanha, F. Schmitzberger, H. S. Gweon, T. L. Blundell, C. Abell and A. G. Smith, Plant J., 2004, 37, 61–72.

42 U. Genschel, C. A. Powell, C. Abell and A. G. Smith, Biochem. J., 1999, 341, 669–678.

43 K. Nowak, N. Luniak, C. Witt, Y. Wustefeld, A. Wachter, R. R. Mendel and R. Hansch, Plant Cell Physiol., 2004, 45, 1889–1894.

44 J. Bourguignon, F. Rebeille and R. Douce, ‘Serine and glycine metabolism in higher plants’, in Plant Amino Acids, ed. B. K. Singh, Marcel Dekker, New York, 1999, pp. 111–146.

45 R. Douce, J. Bourguignon, M. Neuburger and F. R´ebeill´e, Trends Plant Sci., 2001, 6, 167–176.

46 S. Jabrin, S. Ravanel, B. Gambonnet, R. Douce and F. R´ebeill´e, Plant Physiol., 2003, 131, 1431–1439.

47 M. J. Warren, E. Raux, H. L. Schubert and J. C. Escalante-Semerena, Nat. Prod. Rep., 2002, 19, 390–412.

48 M. T. Croft, A. D. Lawrence, E. Raux-Deery, M. J. Warren and A. G. Smith, Nature, 2005, 438, 90–93.

49 M. T. Croft, M. J. Warren and A. G. Smith, Eukaryotic Cell, 2006, 5, 1175–1183.

50 Y. I. Kim, J. Nutr. Biochem., 1999, 10, 66–88.

51 G. M. Li, S. R. Presnell and L. Gu, J. Nutr. Biochem., 2003, 14, 568– 575.

52 J. Geisel, J. Perinat. Neonat. Nurs., 2003, 17, 268–279.

53 P. J. Stover, Nutr. Rev., 2004, 62, S3–S12.

54 S. Seshadri, A. Beiser, J. Selhub, P. F. Jacques, I. H. Rosenberg, R. B. D’Agostino, P. W. Wilson and P. A. Wolf, N. Engl. J. Med., 2002, 346, 476–483.

55 A. Marquet, B. T. Bui and D. Florentin, Vitam. Horm., 2001, 61, 51–101.

56 B. Wolf, J. Nutr. Biochem., 2005, 16, 441–445.

57 W. A. Fenton and L. E. Rosenberg, ‘Inherited disorders of cobalamin transport and metabolism’, in The Metabolic and Molecular Bases of Inherited Disease, ed. C. R. Scriver, A. L. Beaudet, W. S. Sly and D. Valle, McGraw–Hill, New York, 1995, pp. 3129–3149.

58 I. Chanarin, Br. J. Haematol., 2000, 111, 407–415.

59 J. B. Tarr, T. Tamura and E. L. R. Stokstad, Am. J. Clin. Nutr., 1981, 34, 1328–1337.

This journal is © The Royal Society of Chemistry 2007

60 C. M. Smith and W. O. Song, J. Nutr. Biochem., 1996, 7, 312–321.

61 M. Duran, F. A. Beemer, C. van de Heiden, J. Korteland, P. K. de Bree, M. Brink, S. K. Wadman and I. Lombeck, J. Inherited Metab.

Dis., 1978, 1, 175–178.

62 J. Reiss and J. L. Johnson, Hum. Mutat., 2003, 21, 569–576.

63 J. L. Johnson and M. Duran, in The Metabolic and Molecular Bases of Inherited Disease, ed. C. Scriver , A. L. Beaudet, W. S. Sly and D. Valle, McGraw–Hill, New York, 2001, pp. 3163–3177.

64 S. Kramer, R. V. Hageman and K. V. Rajagopalan, Arch. Biochem.

Biophys., 1984, 233, 821–829.

65 J. Reiss, Hum. Genet., 2000, 106, 157–163.

66 H. J. Lee, I. M. Adham, G. Schwarz, M. Kneussel, J. O. Sass, W. Engel and J. Reiss, Hum. Mol. Genet., 2002, 11, 3309–3317.

67 G. Schwarz, J. A. Santamaria-Araujo, S. Wolf, H. J. Lee, I. M. Adham, H. J. Grone, H. Schwegler, J. O. Sass, T. Otte, P. Hanzelmann, R. R. Mendel, W. Engel and J. Reiss, Hum. Mol. Genet., 2004, 13, 1249– 1255.

68 J. Scott, F. R´ebeill´e and J. Fletcher, J. Sci. Food Agric., 2000, 80, 795–824.

69 S. P. Stabler and R. H. Allen, Annu. Rev. Nutr., 2004, 24, 299–326.

70 G. J. C. Basset, E. P. Quinlivan, J. F. Gregory and A. D. Hanson, Crop Sci., 2005, 45, 449–453.

71 S. Storozhenko, S. Ravanel, G.-F. Zhang, F. R´ebeill´e, W. Lambert and D. Van Der Straeten, Trends Food Sci. Technol., 2005, 16, 271–281.

72 T. Hossain, I. Rosenberg, J. Selhub, G. Kishore, R. Beachy and K. Schubert, Proc. Natl. Acad. Sci. U. S. A., 2004, 101, 5158–5163.

73 R. Diaz de la Garza, E. P. Quinlivan, S. M. Klaus, G. J. Basset, J. F. Gregory, 3rd and A. D. Hanson, Proc. Natl. Acad. Sci. U. S. A., 2004, 101, 13720–13725.

74 R. Diaz De La Garza, J. Gregory III and A. D. Hanson, Proc. Natl.

Acad. Sci. U. S. A., 2007, 104, 4218–4222.

75 M. L. Jones and P. F. Nixon, J. Nutr., 2002, 132, 2690–2694.

76 F. Rebeille, M. Neuburger and R. Douce, Biochem. J., 1994, 302(1), 223–228.

77 S. Roje, M. T. Janave, M. J. Ziemak and A. D. Hanson, J. Biol. Chem., 2002, 277, 42748–42754.

78 A. Goyer, E. Collakova, R. Diaz de la Garza, E. P. Quinlivan, J. Williamson, J. F. Gregory, 3rd, Y. Shachar-Hill and A. D. Hanson, J. Biol. Chem., 2005, 280, 26137–26142.

79 G. Orsomando, G. G. Bozzo, R. D. de la Garza, G. J. Basset, E. P. Quinlivan, V. Naponelli, F. Rebeille, S. Ravanel, J. F. Gregory and A. D. Hanson, Plant J., 2006, 46, 426–435.

80 F. Elischewski, A. Puhler and J. Kalinowski, J. Biotechnol., 1999, 75, 135–146.

81 A. Rubio and D. M. Downs, J. Bacteriol., 2002, 184, 2827–2832.

82 W. M. Fouad and B. Rathinasabapathi, Plant Mol. Biol., 2006, 60, 495–505.

83 Y. Hikichi, T. Moriya, I. Nogami, H. Miki and T. Yamaguchi, Patent No. EP-A 0 590 857, 1994.

84 T. Moriya, Y. Hikichi, Y. Moriya, T. Yamaguchi, Patent No. WO 97/10340, 1997.

85 N. Dusch, A. Puhler and J. Kalinowski, Appl. Environ. Microbiol., 1999, 65, 1530–1539.

86 H. Sahm and L. Eggeling, Appl. Environ. Microbiol., 1999, 65, 1973– 1979.

87 M. Merkamm, C. Chassagnole, N. D. Lindley and A. Guyonvarch, J. Biotechnol., 2003, 104, 253–260.

88 C. Chassagnole, F. Letisse, A. Diano and N. D. Lindley, Mol. Biol.

Rep., 2002, 29, 129–134.

89 C. Chassagnole, A. Diano, F. Letisse and N. D. Lindley, J. Biotechnol., 2003, 104, 261–272.

90 A. T. Huser, C. Chassagnole, N. D. Lindle, M. Merkamm, A. Guyonvarch, V. Elisakova, M. Patek, J. Kalinowski, I. Brune, A. Puhler and A. Tauch, Appl. Environ. Microbiol., 2005, 71, 3255–3268.

91 J. H. Martens, H. Barg, M. J. Warren and D. Jahn, Appl. Microbiol.

Biotechnol., 2002, 58, 275–285.

92 F. Blanche, B. Cameron, J. Crouzet, L. Debussche, D. Thibaut, M. Vuilhorgne, F. J. Leeper and A. R. Battersby, Angew. Chem., Int. Ed. Engl., 1995, 34, 383–411.

93 M. E. Taga, N. A. Larsen, A. R. Howard-Jones, C. T. Walsh and G. C. Walker, Nature, 2007, 446, 449–453.

94 M. Tambasco-Studart, O. Titiz, T. Raschle, G. Forster, N. Amrhein and T. B. Fitzpatrick, Proc. Natl. Acad. Sci. U. S. A., 2005, 102, 13687–13692.

95 F. J. Sandoval and S. Roje, J. Biol. Chem., 2005, 280, 38337–38345.

Nat. Prod. Rep., 2007, 24, 949–962 | 961

Published on 20 August 2007. Downloaded by Harvard University on 19/11/2013 20:46:55.

View Article Online 96 P. Rippert, C. Scimemi, M. Dubald and M. Matringe, Plant Physiol., 2004, 134, 92–100.

97 H. van den Berg, R. Faulks, H. F. Granado, J. Hirschberg, B. Olmedilla, G. Sandmann, S. Southon and W. Stahl, J. Sci. Food Agric., 2000, 80, 880–912.

98 K. M. Coxon, E. Chakauya, H. H. Ottenhof, H. M. Whitney, T. L. Blundell, C. Abell and A. G. Smith, Biochem. Soc. Trans., 2005, 33, 743–746.

99 A. Picciocchi, R. Douce and C. Alban, J. Biol. Chem., 2003, 278, 24966–24975.

100 C. G. Bartoli, G. M. Pastori and C. H. Foyer, Plant Physiol., 2000, 123, 335–343.

101 F. Rebeille, S. Ravanel, S. Jabrin, R. Douce, S. Storozhenko and D. Van Der Straeten, Physiol. Plant., 2006, 126, 330–342.

962 | Nat. Prod. Rep., 2007, 24, 949–962

102 G. Noctor, G. Queval and B. Gakiere, J. Exp. Bot., 2006, 57, 1603– 1620.

103 S. M. Chabregas, D. D. Luche, M. A. Van Sluys, C. F. M. Menck and M. C. Silva, J. Cell Sci., 2003, 116, 285–291.

104 I. Ajjawi, Y. Tsegaye and D. Shintani, Arch. Biochem. Biophys., 2007, 459, 107–114.

105 M. Bedhomme, M. Hoffmann, E. A. McCarthy, B. Gambonnet, R. G. Moran, F. Rebeille and S. Ravanel, J. Biol. Chem., 2005.

106 S. M. J. Klaus, E. R. S. Kunji, G. G. Bozzo, A. Noiriel, R. D. de la Garza, G. J. C. Basset, S. Ravanel, F. Rebeille, J. F. Gregory and A. D. Hanson, J. Biol. Chem., 2005, 280, 38457– 38463.

107 C. G. Staggs, W. M. Sealey, B. J. McCabe, A. M. Teague and D. M. Mock, J. Food Compos. Anal., 2004, 17, 767–776.

This journal is © The Royal Society of Chemistry 2007