JXB Advance Access originally published online on May 19, 2006
Journal of Experimental Botany 2006 57(8):1603-1620; doi:10.1093/jxb/erj202
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
RESEARCH PAPER |
NAD(P) synthesis and pyridine nucleotide cycling in plants and their potential importance in stress conditions*
Institut de Biotechnologie des Plantes, Université de Paris XI, F-91405 Orsay, France
To whom correspondence should be addressed. E-mail: noctor{at}ibp.u-psud.fr
Received 1 March 2006; Accepted 17 March 2006
 |
Abstract |
|---|
Top
Abstract
IntroductionPyridine nucleotides are key redox carriers in the soluble phase of all living cells, and both NAD and NADP play crucial roles in pro-oxidant and antioxidant metabolism. Recent data also suggest a number of non-redox mechanisms by which these nucleotides could influence cell function. In cases where these mechanisms involve NAD(P) consumption, resynthesis must occur to maintain nucleotide pools. Important information on the pathways involved in NAD synthesis in plants is beginning to appear, but many outstanding questions remain. This work provides an overview of the current state of knowledge on NAD synthesis pathways in plants and other organisms, analyses plant sequences for the first two enzymes of the de novo synthesis of NAD, proposes a preliminary model for the intracellular distribution of NAD synthesis, presents plant homologues of recently identified yeast mitochondrial NAD transporters, and discusses factors likely to be important in the regulation of NAD synthesis and contents in plants, with particular reference to stress conditions.
Key words: L-Aspartate oxidase, compartmentation, NAD transporter, NAD signalling, nicotinamide, nicotinate, quinolinate synthase
|
Introduction |
|---|
NAD and its derivative NADP are ubiquitous in living cells (Penfound and Foster, 1996; Kurnasov et al., 2003). They mediate hundreds of redox reactions, and thus impact on virtually every metabolic pathway in the cell. Pyridine nucleotides are also key players in signalling through reactive oxygen species (ROS) since they are crucial in the regulation of both ROS-producing and ROS-consuming systems in plants (Møller, 2001; Apel and Hirt, 2004; Mittler et al., 2004; Foyer and Noctor, 2005). Chloroplast ROS production is influenced by NADP+:NADPH ratios, mitochondrial NAD(P) status is important in determining ROS formation by the respiratory electron transport chain, and NAD(P)H oxidases are key players in the generation of ROS at the plasmalemma. Similarly, ROS processing partly depends on ascorbate and glutathione, and redox flux through these pools is maintained by NAD(P)H. Pyridine nucleotides are involved in many other defence and signalling reactions, for example, production of nitric oxide and metabolism of reactive lipid derivatives (Crawford and Guo, 2005; Mano et al., 2005). Thus, as well as their roles as cofactors in numerous energy-producing and biosynthetic reactions, NAD and NADP are crucial in redox signalling linked to stress and development.
Recent studies show that NAD and NADP play important roles in cell signalling in animals, plants, and fungi (Hunt et al., 2004; Ziegler, 2005). NAD and NADP are the respective substrates for the production of the calcium agonists cyclic ADP-ribose (cADP-R; Sánchez et al., 2004) and nicotinate adenine dinucleotide phosphate (NaADP; Navazio et al., 2000; Lee, 2005). Transfer of ADP-ribose units from NAD to target proteins is an important post-translational modification, catalysed by poly(ADP-ribose)polymerases (PARPs) and mono(ADP-ribosyl)transferases (Ying et al., 2003; Graziani et al., 2005). The yeast Silent Information Regulator 2 (Sir2) protein has been shown to be an NAD-dependent deacetylase (Imai et al., 2000; Lin et al., 2000; Berger et al., 2004; Ziegler, 2005). Although these components and phenomena are better characterized in other systems than in plants, many of the products and the enzymes that produce them are also found in plants (Hunt et al., 2004). Since some of these NAD(P)-consuming reactions are important in stress conditions and/or are stress inducible, NAD concentration could be influential in plant responses to environmental challenge. Exposure of plant cells to oxidative stress produces decreases in overall NAD contents, linked to PARP activity (Amor et al., 1998; De Block et al., 2005). Recent studies of mutants provide evidence that changes in NAD status can alter photosynthesis and plant stress responses (Dutilleul et al., 2003a, b; Hayashi et al., 2005), and suggest that NAD content could be a powerful modulator of metabolic integration (Dutilleul et al., 2005). Much remains to be discovered about the factors controlling pyridine nucleotide synthesis and recycling in plants.
|
NAD recycling and associated reactions |
|---|
The most studied route to NAD in mammals is via nicotinamide and nicotinate. Together, these compounds are called ‘niacins’ and were identified by their ability to prevent pellagra in dogs fed a diet poor in tryptophan, which is a precursor in the de novo synthesis of the pyridine ring in animals (see below). It was subsequently shown that niacins can be converted to NAD via nicotinate phosphoribosyltransferase (NaPRT), nicotinate mononucleotide adenylyltransferase (NaMNAT), and NAD synthetase (Preiss and Handler, 1958). Recently, the structure of a yeast NaPRT has been determined at a resolution of 1.75 Å (Chappie et al., 2005). In animals, which lack nicotinamide deamidase activity (Chappie et al., 2005), NAD can also be produced via nicotinamide phosphoribosyltransferase (NPRT) activity, followed by adenylylation of NMN. This allows incorporation of nicotinamide into NAD in two steps rather than the four-step pathway via nicotinate (Fig. 1, orange background).
|
Labelling studies in plants suggest that nicotinamide released from NAD is first deamidated by nicotinamidases to nicotinate (Ryrie and Scott, 1969; Ashihara et al., 2005; Zheng et al., 2005) which is then converted to NaMN by NaPRT (Fig. 1). This view is supported by the presence in Arabidopsis and other plants of genes predicted to encode NaPRT, NaMNAT, and NAD synthetase (Hunt et al., 2004; Katoh and Hashimoto, 2004). Consistent with this, homologues of mammalian NPRT gene sequences appear to be absent from plants (Katoh and Hashimoto, 2004) and NPRT activity has not been detected in extracts of tobacco, mungbean, or white spruce cells (Wagner et al., 1986; Ashihara et al., 2005; Zheng et al., 2005).
Results of a detailed study of the fate of 14C-labelled nicotinamide in coffee and cocao (Zheng et al., 2004) also suggest that nicotinate is the main precursor of pyridine nucleotides, although NMN was also labelled in significant amounts, probably as a breakdown product of NAD through NAD pyrophosphatase activity (Fig. 1). In view of these results, together with data gathered from labelling studies in white spruce and mungbean (Ashihara et al., 2005; Zheng et al., 2005), three possible routes were suggested from NAD to nicotinamide. The first involves production of nicotinamide in a single step, with ADP-R (or derivatives) as the other product. This route would result from the activities of enzymes such as NAD glycohydrolase, PARP, SIRTUINS, and ADP-ribosyl cyclase (Fig. 1). A second pathway would proceed via NMN and involve two steps, catalysed by NAD pyrophosphatase and NMN nucleosidase (Zheng et al., 2004). In the third pathway, NMN is indirectly converted to nicotinamide via nicotinamide riboside, and overall the three steps involve NAD pyrophosphatase, 5' nucleotidase, and nicotinamide nucleoside phosphorylase or nucleosidase (Ashihara et al., 2005; Zheng et al., 2005). According to this scenario, full turnover of the pyridine cycling pathway (from NAD to NAD) could involve five, six, or seven steps. The final four steps would be common to all these recycling pathways and involve formation of nicotinate from nicotinamide, conversion of nicotinate to NaMN, adenylylation of NaMN to NaAD (desamidoNAD), and amidation of NaAD to produce NAD (Fig. 1).
Given the presence of NMN in plant tissues, further work is required to confirm the substrate specificities of plant N(a)MNAT activities, and to establish whether these enzymes are specific to NaMN. In mammals, three adenylyltransferases have been characterized and all are able to use NaMN and NMN (Magni et al., 2004). This probably reflects the dual role of mammalian N(a)MNAT in both the de novo pathway (production of NaAD from NaMN) and the recycling pathway (production of NAD from NMN; Fig. 1, orange background, top). In bacteria, the adenylyltransferase shows a clear preference for NaMN, a compound which in these organisms is common to both de novo synthesis and the recycling pathway from nicotinamide by the successive action of the pncA and pncB gene products (Fig. 1). The final step of NAD synthesis is catalysed by NAD synthetases, probably using glutamine, although in some organisms isoforms are able to use ammonia under physiological conditions (Magni et al., 2004).
In addition to synthesis of pyridine mononucleotides (NMN and NaMN) by phosphoribosyltransferase activities, a third path to NAD+ has been described in species of bacteria, yeast, and mammals (Bieganowski and Brenner, 2004). This route involves synthesis of NMN by phosphorylation of nicotinamide riboside (Fig. 1). An elegant vitamin screen using yeast mutants indicated the presence of nicotinamide riboside in milk (Bieganowski and Brenner, 2004). As discussed above, nicotinamide riboside might be formed in mungbean from NMN (Zheng et al., 2005). A BLAST search of several plant genome databases using bacterial, yeast, and mammalian sequences for nicotinamide riboside kinase revealed no fully homologous sequences, although homologous domains can be identified among the uridine kinase family.
In addition to nicotinamide riboside, plants contain nicotinate riboside. Indeed, labelling studies in coffee have provided some evidence that this compound is an intermediate in pyridine cycling and that NaMN can be formed from nicotinate through two paths (Zheng et al., 2004). As well as direct phosphoribose transfer catalysed by NaPRT, a second path is possible through successive ribosylation and phosphorylation (Fig. 1, orange background, bottom). These reactions are reversible and could also represent a route from NaMN to nicotinate (Ashihara et al., 2005). If they operate in the direction of NaMN synthesis, it would add yet another step to the recycling pathway in plants. In this case, recycling of the pyridine ring from NAD to NAD could potentially consist of eight steps rather than the five-step pathway described in bacteria (Fig. 1, orange background, cf. shorter route traced by the solid arrows with the longer path including dotted arrows via the two ribosides).
Methylation of nicotinamide and nicotinate produces N-methyl nicotinamide (MNam) and N-methyl nicotinate (trigonelline). In mammals, MNam is the initial product of a degradation pathway in which pyridone products are subsequently formed by MNam oxidation (Fig. 1, grey background). Urinary excretion of both MNam and pyridone products is stimulated when the diet is rich in nicotinamide or nicotinate (Creeke and Seal, 2005, and references therein). MNam has been detected in plants (Waller et al., 1966) while trigonelline is known to be formed in many types of plants, notably legumes, and may have hormonal-type functions and be important as a store of nicotinate (Minorsky, 2002; Zheng et al., 2004, 2005). In addition to MNam and trigonelline, other plant compounds could act as conjugated stores of nicotinamide and nicotinate, including N-glucosylated forms and nicotinate riboside (Berglund et al., 1996; Zheng et al., 2004). In coffee and cocoa, most of the label from [14C]nicotinamide was incorporated into trigonelline, with only minor labelling of nicotinate riboside and nicotinate glucoside (Zheng et al., 2004). In contrast, in tea, the major labelled compound was nicotinate glucoside, with no detectable label found in trigonelline. These observations show that the type of nicotinate conjugate formed is species dependent (Zheng et al., 2004).
As well as trigonelline, other pyridine-based alkaloids include ricinine, nicotine, and anabasine (Fig. 1, blue background). Production of these compounds is enhanced by stress, particularly wounding, and they are effective feeding deterrents. Much of the information that has been gathered on the regulation of NAD-linked metabolism in plants has come from investigations of pyridine alkaloid synthesis rather than NAD(P) synthesis per se (Waller et al., 1966; Frost et al., 1967; Mann and Byerrum, 1974a, b; Wagner et al., 1986; Sinclair et al., 2000; Cane et al., 2004; Zheng et al., 2004, 2005). Labelling studies have shown that nicotinate, nicotinamide, quinolinate, NaAD, and NAD are all effective sources of the pyridine ring for nicotine and ricinine synthesis (Waller et al., 1966; Frost et al., 1967). These data emphasize the close relationship between NAD metabolism and alkaloid production in these species. The observation that quinolinate and NAD are as effective as nicotinate in supporting production of pyridine alkaloids (Waller et al., 1966; Frost et al., 1967; Zheng et al., 2005) suggests that, at least in those species examined, the pyridine recycling pathway is capable of relatively rapid turnover under many conditions.
|
NADP(H) synthesis and utilization |
|---|
NADP is produced from NAD by NAD(H) kinase. At least one NAD kinase in tobacco is a calmodulin-binding protein (Harding et al., 1997) and activity in wild tomato may be subject to regulation by thiol/disulphide exchange (Delumeau et al., 2000). NAD kinase activity could be located in the chloroplast, cytosol, and mitochondria (Gallais et al., 2000a; Turner et al., 2004; Chai et al., 2005). Two NAD+-specific Arabidopsis NAD kinases, named NADK1 and NADK2, have been cloned and characterized biochemically (Turner et al., 2004, 2005). Compared with NADK1, NADK2 has a large N-terminal extension (Turner et al., 2004), probably involved in regulation by calmodulin (Harding et al., 1997). Total extractable NAD kinase was shown to be highly dependent on calmodulin in Arabidopsis roots and shoots throughout development (Turner et al., 2004). In contrast to NADK1 and NADK2, a third characterized Arabidopsis isoform, NADK3, shows a strong preference for NADH over NAD (60-fold lower Km for NADH; Turner et al., 2005). The yeast homologue of Arabidopsis NADK3 is a mitochondrial enzyme, whereas the plant enzyme appears to be located in the cytosol (Turner et al., 2005). In plants, mitochondrial NADP could be at least partly imported from the cytosol (Bykova and Møller, 2001). If NADK3 is a cytosolic enzyme, some in vivo activity with NAD+ cannot be completely discounted in view of the high total NAD+/NADH ratio in this compartment (10–30: Igamberdiev and Gardeström, 2003). Indeed, the physiologically relevant ‘free’ NAD+/NADH ratio (nucleotides not bound to protein at any given moment) may be as high as 600–1000 in the cytosol of both plant and mammalian cells (Heineke et al., 1991; Zhang et al., 2002). Other authors have reported that Arabidopsis NADK2 appears to localize to the chloroplast and that an insertional mutant for the enzyme shows stunted growth and stress hypersensitivity, associated with a minor decrease in total tissue NADPH (Chai et al., 2005). NADP+ phosphatase activity, which converts NADP+ to NAD+, has also been reported in plants (Gallais et al., 2000b), and the mitochondrial NAD(H) and NADP(H) pools are redox-linked via transhydrogenases (Bykova et al., 1999).
As well as its redox cofactor function, NADP is the precursor for the calcium channel activator NaADP (Lee, 2005). NaADP has been shown to be active in plants (Navazio et al., 2000) and can be synthesized through base exchange of the nicotinamide moiety of NADP with free nicotinate, a reaction catalysed by ADP-ribosyl cyclases under certain conditions (Fig. 1, green background, left). NaADP concentrations could also be regulated by specific calcium-dependent phosphatases, which convert it to NaAD (Berridge et al., 2002), or nucleotide pyrophosphatases, which produce NaMN and AMP-2'-phosphate (Evans et al., 2005; Fig. 1). Glycohydrolases can function to degrade NADP to nicotinamide and ADP-R 2'-phosphate or cADP-R 2'-phosphate. The best-characterized ADP-ribosyl cyclase in mammals is the membrane-bound glycoprotein CD38 (Evans et al., 2005). This protein can convert extracellular or intracellular NAD to cADP-R, hydrolyse cADP-R to ADP-R, and catalyse synthesis of NaADP via the base exchange reaction with nicotinate (Fig. 1, green and grey backgrounds). Although no genes have been identified on the basis of sequence homology, low activities of both ADP-ribosyl cyclase and cADP-R hydrolase have been detected in Arabidopsis, and cADP-R has been shown to accumulate transiently in response to abscisic acid (Sánchez et al., 2004). In human platelets, it has also been reported that CD38 can cleave NADP to ADP-R 2'-phosphate (Torti et al., 1999). Like the reactions catalysed by PARP and SIRTUINS, many of these NAD and NADP cleavage reactions produce nicotinamide (Fig. 1, green background).
|
De novo synthesis of nicotinate mononucleotide (NaMN) |
|---|
As well as producing NaMN from nicotinate obtained from the environment or formed during the recycling pathway, almost all organisms appear to be able to produce this NAD precursor via synthesis of the pyridine ring in the form of quinolinate (pyridine 2,3-dicarboxylate; Fig. 1, purple background). NaMN is then produced by quinolinate phosphoribosyltransferase (QPRT; EC 2.4.2.19 [EC] ), encoded by nadC in Escherichia coli (Fig. 1). Characterization of the enzyme purified from castor bean endosperm showed that Km values for quinolinate and 5-phosphoribosyl-1-pyrophosphate were 12 and 45 µM, respectively (Mann and Byerrum, 1974a). Tobacco QPRT cDNA sequences suggested that the protein is probably located in the mitochondria, although a shorter transcript might produce a cytosol-directed protein (Sinclair et al., 2000). Computer analysis (http://urgi.infobiogen.fr/predotar) of sequences encoded by the single QPRT gene in Arabidopsis (Fig. 1) predicts two transcripts, the longer of which probably encodes a mitochondrial import sequence. Quinolinate-dependent synthesis of NaMN therefore probably occurs in both cytosolic and mitochondrial compartments.
In animals and fungi, quinolinate is formed through a multistep pathway from tryptophan. Recently, gene sequences for several enzymes of this pathway have also been found in certain pathogenic bacteria (Kurnasov et al., 2003), and some anaerobic species have been reported to make quinolinate by acetylation of N-formyl-L-aspartate (Scott et al., 1969; Griffith et al., 1975). However, by far the best characterized prokaryotic pathway for quinolinate synthesis is from aspartate and triose phosphate (Griffith et al., 1975; Flachmann et al., 1988; Kurnasov et al., 2003). This pathway involves two enzymatic steps, catalysed by quinolinate synthase, a term which has been used to refer to the two enzymes functioning together or specifically to designate the second enzyme. The first enzyme, L-aspartate oxidase (or quinolinate synthase ‘B’), is encoded by nadB, while the second, quinolinate synthase (or quinolinate synthase ‘A’), is encoded by nadA (Fig. 1). Close association of the two enzymes might enable the intermediate, iminoaspartate, an unstable compound that spontaneously hydrolyses to oxaloacetate, to be channelled from the first enzyme to the second. In E. coli, the nadB and nadA genes are arranged in a single operon together with the nadC gene encoding QPRT (Griffith et al., 1975).
|
Quinolinate synthesis in plants |
|---|
In mammals, in addition to the well-known neuroactivity of nicotine, the endogenously produced NAD precursor quinolinate acts as an agonist at N-methyl-D-aspartate (NMDA) receptors, which are classed among the glutamate receptor family (Stone and Addae, 2002). Quinolinate accumulation in the brain has been implicated in several neurodegenerative disorders (Fukuoka et al., 2002). The Arabidopsis genome contains numerous loci that encode subunits of glutamate receptors. Although much remains to be discovered concerning the functions of these proteins, plant NMDA receptor-like proteins share certain characteristics of their animal counterparts (Dubos et al., 2003). It is an intriguing possibility that, in addition to NAD(P)-linked signalling functions associated with redox control, production of NaADP and cADP-ribose, and PARP and SIRTUIN activities, other NAD(P) precursors or metabolites such as quinolinate might act as information transmitters in plants.
To date, there have been very few studies of the enzymes that produce quinolinate in plants, and it has often been stated or implied that there is a strict prokaryotic/eukaryotic division between quinolinate synthesis pathways (Garavaglia et al., 2003; Bieganowski and Brenner, 2004) and thus that plants use the tryptophan pathway (Hunt et al., 2004). Indeed, the tryptophan pathway has been considered an important route in several plants based on some labelling studies (reviewed by Arditti and Tarr, 1979). However, several investigations reported a lack of evidence for quinolinate synthesis from tryptophan, for example, in maize and tobacco (Henderson et al., 1958), and other authors have thus favoured the aspartate pathway while noting that no enzymes necessary to produce quinolinate had been found in plants (Mann and Byerrum, 1974b). Subsequently, it was reported that L-aspartate oxidase activity could be detected in cotton callus cell extracts (Hosokawa et al., 1983) and, more recently, analyses of gene sequences have reported the presence in plant species of the bacterial pathway of quinolinate synthesis from aspartate and triose phosphate (Kurnasov et al., 2003; Katoh and Hashimoto, 2004). Indeed, the Arabidopsis genome contains, as well as a single gene predicted to encode QPRT, two sequences that show significant homology to the two bacterial enzymes involved in this pathway (Fig. 1). In contrast, the Arabidopsis genome does not appear to contain gene sequences that show significant homology to tryptophan pathway enzymes. Other authors have noted the presence of genes encoding enzymes of both the aspartate and the tryptophan pathway in rice (Katoh and Hashimoto, 2004), but sequences associated with the latter pathway have been eliminated by subsequent annotation updates of the genome. Likewise, no sequences of significant homology are found in poplar, Medicago, or Physcomitrella databases. Moreover, a preliminary report has described the cloning of Arabidopsis homologues for nadA and nadB, their ability to rescue E. coli mutants deficient in the respective enzymes, and the presence in plastids of the encoded proteins (Katoh et al., 2005). These data, together with the apparent absence of genes encoding enzymes involved in other possible routes to quinolinate, suggest that plants obtain this key precursor via the aspartate pathway used by many bacteria (Fig. 1, purple background).
|
An nadB homologue in plants |
|---|
L-Aspartate oxidase (EC 1.4.3.16 [EC] ) catalyses the FAD-dependent oxidation of L-aspartate to iminoaspartate (Fig. 1). FAD is regenerated by either molecular oxygen or fumarate. A single sequence is annotated L-aspartate oxidase in Arabidopsis, and homologues are found in several other plant species. Selected plant and bacterial protein sequences were used to construct a phylogenetic tree that reveals a very close relationship between proteins from angiosperm species, with the moss sequence being classed slightly apart (Fig. 2A). By contrast, the algal protein sequence is somewhat intermediate between the bacterial and plant proteins (Fig. 2A). Bacterial sequences cluster less tightly than do the plant sequences, probably because the bacteria for which sequences were analysed are found in very distinct taxonomic groups.
|
A sequence alignment was performed to identify conserved amino acid residues in L-aspartate oxidases from different species (Fig. 2B). The proteins all contain two large domains, one aspartate oxidase domain in the N-terminal two-thirds of the protein and a succinate dehydrogenase domain in the C-terminal part (Mattevi et al., 1999). Earlier studies of the cotton enzyme reported the presence of a permanent cofactor with an apparent molecular mass of 1050 (Hosokawa et al., 1983). This cofactor could be FAD, despite the latter's smaller molecular mass, as this nucleotide has been shown to be essential to the bacterial enzyme. Recent crystallization of the E. coli enzyme allied with site-directed mutagenesis allowed identification of amino acid residues involved in the FAD- and succinate-binding sites (Mattevi et al., 1999; Bossi et al., 2001; Tedeschi et al., 2001). It is noteworthy that these amino acids are strictly conserved between plant and bacterial species (Fig. 2B), consistent with the key role of the residues in the catalytic properties of the enzyme. Though not shown in Fig. 2B, an analysis of plant sequences suggests the presence of a chloroplast transit peptide in L-aspartate oxidase from Arabidopsis, rice, poplar, and Medicago.
Certain organisms can oxidize aspartate to iminoaspartate using enzymes other than L-aspartate oxidase. In Thermogata maritima, the nadB gene encodes an NAD-dependent aspartate dehydrogenase (EC 1.4.1.–), an enzyme with a very different primary protein sequence from L-aspartate oxidase (Yang et al., 2003). In mammals, a D-aspartate oxidase (EC 1.4.3.1 [EC] ) synthesizes iminoaspartate, and in vitro this enzyme can replace L-aspartate oxidase in the quinolinate synthase complex (Nasu et al., 1982a). The physiological role of this enzyme is not in NAD synthesis but in regulating the concentration of D-aspartate, a compound present in the brain and some secretory organs, by converting it to iminoaspartate which then decomposes to oxaloacetate (Wang et al., 2000; Wolosker et al., 2000).
|
An nadA homologue in plants |
|---|
Iminoaspartate produced by L-aspartate oxidase is combined with dihydroxyacetone phosphate to produce quinolinate, a reaction catalysed by quinolinate synthase (also called quinolinate synthase ‘A’). Sequences homologous to the bacterial proteins were found in the Arabidopsis, rice, poplar, Medicago, and Physcomitrella databases, but for Chlamydomonas only a partial sequence is available. For the purposes of sequence alignment, the analysis was restricted to this part of the protein. Construction of a phylogenetic tree for the protein sequences generates a clear distinction between plants and bacteria (Fig. 3A). The five plant sequences are closely related to each other and also to the Chlamydomonas sequence. Bacterial sequences form a cluster in which E. coli and Synechocystis are most closely related. Plant and bacterial sequences appear more distinct from each other than for L-aspartate oxidase.
|
Alignment of plant sequences with the partial Chlamydomonas sequence indicates the presence of several highly conserved amino acid residues (Fig. 3B). The three-dimensional structure of the crystallized quinolinate synthase from Pyrococcus horikoshii shows a triangular architecture in which conserved amino acids determine three structurally homologous domains, termed ‘three-fold repeat of three-layer sandwich’ (Sakuraba et al., 2005). Comparison of complete bacterial, rice, and Arabidopsis sequences with the P. horikoshii sequence suggests that all quinolinate synthases share this unique triangular architecture, although the plant sequences contain an additional C-terminal extension.
Recent studies demonstrate that E. coli quinolinate synthase harbours a [4Fe–4S] cluster (Cicchillo et al., 2005; Ollagnier de Choudens et al., 2005). The sensitivity of this cofactor to oxidation probably explains why quinolinate synthase is the site of oxygen poisoning of pyridine nucleotide synthesis in anaerobic bacteria (Gardner and Fridovich, 1991; Draczynska-Lusiak and Brown, 1992; Ollagnier de Choudens et al., 2005). The characteristic binding motif for [4Fe–4S] clusters is Cys-w-x-Cys-y-z-Cys, and this is found in the C-terminal region of most quinolinate synthases from bacteria (Ollagnier de Choudens et al., 2005), including Synechocystis. No homologous sequence is found in the plant quinolinate synthase. The catalytic mechanism and cofactors for the plant enzyme remain to be identified.
Although less clear for other available plant quinolinate synthase sequences, the Arabidopsis and poplar enzymes are predicted to contain an N-terminal plastid targeting peptide. Together with the chloroplastic localization of L-aspartate oxidase, this suggests that quinolinate synthesis probably occurs in this organelle. As iminoaspartate is unstable, L-aspartate oxidase and quinolinate synthase could occur in close association, although no conclusive evidence has yet been found in bacteria for the formation of a complex (Sakuraba et al., 2005).
|
Compartmentation and transport in NAD synthesis and metabolism |
|---|
Total NAD(H) concentrations in leaf cell mitochondria and cytosol are
2 and 0.6 mM, respectively, while total NADP(H) pools are of the order of 0.3 mM in both compartments (Igamberdiev and Gardeström, 2003). NAD(H) and NADP(H) pools of isolated chloroplasts have both been calculated to be 0.4 mM (Takahama et al., 1981). Because of their key roles in metabolism in these compartments and in others such as the peroxisomes, NAD and NADP show very rapid redox turnover during processes such as photosynthesis, photorespiration, and respiration. Numerous small metabolite exchange systems are able to shuttle redox equivalents rapidly across chloroplast and mitochondrial membranes and thus link NAD(P) pools in the different compartments (Raghavendra and Padmasree, 2003; Scheibe et al., 2005). These involve triose phosphate/3-phosphoglycerate, malate/oxaloacetate, and malate/aspartate exchange (Day and Wiskich, 1981; Journet et al., 1981; Heineke et al., 1991). Glutamate/2-oxoglutarate shuttle systems, which are key in nitrogen assimilation and ammonia reassimilation (Woo et al., 1987), also potentially involve exchange of redox equivalents. Exchange systems also operate to link peroxisomal and cytosol NAD pools (Reumann et al., 1994). As well as indirect redox transfer between the mitochondrial matrix and the cytosol, evidence has recently been presented that a cytosolic glycerol-3-phosphate dehydrogenase (G3PDH) operates with an externally oriented mitochondrial inner membrane G3PDH:ubiquinone reductase to deliver electrons to the Arabidopsis mitochondrial electron transport chain, and this system appears to contribute significantly to cellular redox homeostasis (Shen et al., 2006). Other externally oriented ‘alternative’ dehydrogenases located in the inner mitochondrial membrane may also allow oxidation of NADH or NADPH produced in the cytosol or intermembrane space (Giegé et al., 2003; Rasmusson et al., 2004). In comparison with the redox exchange between compartments, little is known about the transport of NAD and NADP themselves. As discussed above, available data suggest that the first three steps of the de novo pathway of NAD synthesis involve production of quinolinate in the chloroplast and NaMN formation in the cytosol and mitochondrion (Fig. 4). In Arabidopsis, only a single gene is predicted to encode each of the enzymes necessary for the subsequent adenylylation and amidation steps. Neither sequence encodes an apparent transit peptide sequence. In mammals, three N(a)MNAT isoforms have been characterized: one is found in the nucleus, another outside the nucleus in an undefined location, and the third in both the cytosol and mitochondria (Magni et al., 2004). A green fluorescent protein (GFP)–yeast NAD synthetase fusion construct showed a diffuse localization in both the cytosol and the nucleus (Suda et al., 2003). Taken together, available data suggest that neither plastids nor mitochondria are involved in the last two steps of NAD synthesis, though this remains to be confirmed (Fig. 4). Based on this view, there would appear to be a minimum requirement for plastid inner membrane transporters able to export quinolinate and import NAD, as well as for mitochondrial import of NAD and quinolinate for mitochondrial QPRT activity (Sinclair et al., 2000). Both NAD+ and NADP+ have been shown to cross plant mitochondrial membranes (Neuburger and Douce, 1983; Bykova and Møller, 2001), but very little genetic and molecular information is available on NAD transporters.
|
As they are relatively large, phosphorylated metabolites, pyridine nucleotides are usually considered to be exclusively intracellular, with virtually no movement between cells. However, mammalian plasma contains low concentrations of NAD, which are important in signalling through membrane-bound ADP-ribosyl cyclases such as CD38 that could act both to produce cADP-R on the extracellular face of the membrane and to transport it into the cell (Evans et al., 2005). Recently, a plasma membrane NAD transporter has been shown to operate in a pathogenic intracellular bacterium (Haferkamp et al., 2004). This strain is unable to synthesize NAD and instead takes it up from the host cell, probably in exchange for ADP (Haferkamp et al., 2004). Other bacteria, such as Salmonella typhimurium, are able to take up exogenous NMN via the product of the pnuC gene, particularly under conditions of low intracellular NAD+ (Zhu et al., 1991; Penfound and Foster, 1999), where uptake activity could be stimulated by the regulatory protein NadR (see below). Very recently, two genes encoding mitochondrial NAD+ transporters have been identified in yeast where, as in plants, the final steps of NAD synthesis appear to occur outside the mitochondria (Todisco et al., 2006). Biochemical characterization of one of the gene products, NDT1, showed high specificity for NAD+ (Km = 0.38 mM), with virtually no activity with NADH, NADP+, or NADPH, but some activity with other nucleotides, including NaAD (Todisco et al., 2006). The protein appears to catalyse unidirectional transport at low rates and exchange with other nucleotides at faster rates (Todisco et al., 2006). Genome searches reveal that homologues of the two yeast mitochondrial transporters exist in taxonomically diverse organisms, including Arabidopsis, poplar, and rice (Fig. 5). The two Arabidopsis proteins have 30–31% identity with yeast NDT1 and the sequences are currently annotated adenine nucleotide translocator/mitochondrial substrate carrier, with a predicted inner mitochondrial membrane location. It remains to be seen whether similar proteins exist in the plastid envelope, and whether mitochondrial import of NAD occurs in exchange for other nucleotides such as NaMN, whose fate after production by mitochondrial QPRT is unclear (Fig. 4). Since NAD(H) kinases appear to exist in multiple compartments (Gallais et al., 2000a; Turner et al., 2004, 2005; Chai et al., 2005), the requirement for NADP transport is yet to be established.
|
As well as mitochondrial and chloroplast transporters, important questions remain concerning nuclear NAD metabolism and nicotinamide recycling, and their relationship to cytoplasmic processes. Imaging analysis of NAD(H) suggests that the free nuclear NAD+:NADH ratio is similar to cytosolic values (Zhang et al., 2002). If pyridine nucleotides pass readily through nuclear membrane pores, then changes in cellular NAD redox state or total NAD pools could impact directly on transcription in animal and yeast cells via redox regulation of the co-repressor CtBP (Zhang et al., 2002) or substrate availability for Sir2 and related proteins (Imai et al., 2000; Lin et al., 2000). Interestingly, the last two enzymes of NAD synthesis are present in yeast and mammalian nuclei (Suda et al., 2003; Magni et al., 2004), and significant resynthesis of NAD could occur within this compartment.
Intercellular transport is another issue. In tobacco, nicotine is translocated from roots to shoots, and root synthesis and shoot accumulation are stimulated by stresses such as wounding (Cane et al., 2005, and references cited therein). Ricinine is found in the phloem sap of castor bean (Holfelder et al., 1998), and trigonelline probably moves between organs in developing mungbean seedlings (Zheng et al., 2005). Whether any of the other intermediates and precursors shown in Fig. 1 move significantly between cells or organs under physiological conditions is unclear.
|
Stress-induced changes in NAD status |
|---|
Given the centrality of pyridine nucleotides to cell function, it seems likely that contents are highly regulated. Operation of NAD-consuming reactions means that regulation will be necessary to ensure homeostatic replenishment of NAD pools. Furthermore, adjustment of NAD pools could play an important part in environmentally induced ‘dynamic’ signalling, in which the cell is able to sense and respond to changes in concentrations of NAD and/or precursors or metabolites (Zhang et al., 2002).
Several NAD-consuming reactions could be important in stress conditions and/or signalling in interactions with ROS and other redox components. Most of these reactions release nicotinamide, which can be recycled to NAD by a salvage pathway (Fig. 1). One important NAD-consuming enzyme induced by stress is PARP. This nuclear enzyme transfers ADP-ribose units from NAD+ to residues on target proteins, including itself, releasing nicotinamide and ultimately forming long poly(ADP-ribose) chains that are important in regulating processes such as DNA repair (Berglund et al., 1996; Graziani et al., 2005). Hyperactivation of PARP-1 by ROS or nitric oxide has been shown to deplete NAD pools in mammalian and plant cells, and decreasing PARP inhibition can diminish death in both types of cell (Heller et al., 1995; Amor et al., 1998). Recently, it has been shown that stress-induced PARP activation causes NAD+ and ATP depletion, and that these effects and the attendant cell death are attenuated in lines in which PARP expression is diminished by RNA interference (RNAi) (De Block et al., 2005).
As well as producing changes in NAD availability, adjustments in NAD metabolism could impact on stress signalling through changes in metabolic intermediates such as quinolinate or nicotinamide, which are known to be neuroactive compounds in animals (Stone and Addae, 2002; Magni et al., 2004). Quinolinate accumulation in the brain has been implicated in several neurodegenerative disorders (Fukuoka et al., 2002, and references therein). Nicotinamide is known to be an inhibitor of nicotinamide-generating enzymes such as PARP and Sir2 (Magni et al., 2004; De Block et al., 2005), and high concentrations have been shown to influence defence processes in plants (Berglund et al., 1993, 1996). Besides PARP, other enzymes can modify the biological activity of proteins by transferring a single ADP-ribose moiety from NAD to target amino acids. An Arabidopsis mutant, rcd1, which shows hypersensitive responses to extracellular superoxide, contains a PARP-type NAD-binding signature that could reflect the mono(ADP-ribose) transferase function (Ahlfors et al., 2004).
Several literature studies show that NAD pools are plastic in plants, and can undergo significant changes in response to the environment. Powdery mildew infection of barley leaves was reported to be associated with increased NAD contents, although no evidence was obtained for increases in synthesis (Ryrie and Scott, 1969). It is evident from Fig. 1 that NAD contents could be modified by changes at multiple levels, including degradation and recycling efficiency, as well as de novo synthesis. Within this context, key points relate to the removal of metabolites from the recycling pathway for pyridine alkaloid synthesis, degradation of nicotinamide, or formation of nicotinate or nicotinamide conjugates (Fig. 1). Enzymes able to decarboxylate and oxidize both pyridine monocarboxylates (nicotinate, isonicotinate, and picolinate) and dicarboxylates (quinolinate and dipicolinate) to hydroxylated products are found in bacteria (Uchida et al., 2003). Another question concerns stress-induced effects on the NAD:NADP ratio. In cultured tobacco cells expressing a hyperactive calmodulin, increased activity of NAD kinase, a calmodulin-activated enzyme, was associated with a significant increase in the NADP+:NAD ratio (Harding et al., 1997). Interestingly, NAD kinase from Bacillus subtilis is allosterically activated by quinolinate (Garavaglia et al., 2003), and this could play some role in the co-regulation of the de novo synthesis of NAD and NADP.
Further evidence pointing to a role for the concentration of NAD and/or associated metabolites in influencing stress resistance comes from analysis of tobacco lines deficient in mitochondrial complex I. These plants respire through alternative respiratory dehydrogenases. Growth is slowed but leaf respiration is increased, as is resistance to both abiotic and biotic stress (Dutilleul et al., 2003a, b). Enhanced resistance is accompanied by a 2-fold increase in leaf contents of NAD+ and NADH, without an appreciable increase in NADP+, NADPH, H2O2, ascorbate, or glutathione (Dutilleul et al., 2003b, 2005). From a physiological point of view, increases in NADH in these plants can perhaps be rationalized in terms of the relatively high Km NADH of the alternative matrix-orientated mitochondrial NADH dehydrogenase, which in this line is likely to be important as a functional replacement of complex I. The biochemical and molecular processes responsible for increases in NAD in the cytoplasmic male sterile (CMS) mutant remain unidentified, and further work is required to determine the extent to which the observed changes in NAD status in these plants contribute to their enhanced resistance to stress (Dutilleul et al., 2003b).
The extent to which the redox state of NAD and NADP pools changes in response to the environment is still a matter of some debate. Concerning the chloroplast, it is often assumed that increasing light causes progressive over-reduction at the reducing side of photosystem I (PSI), decreasing the availability of NADP+, the principal oxidant of ferredoxin, and so favouring diversion of electrons to oxygen to produce ROS. However, the stromal redox state is rather constant in response to changes in irradiance, though less so when metabolism strongly limits electron transport, for example, at low temperatures or if both CO2 and O2 are artificially decreased (Harbinson et al., 1990). Direct measurements of fractionated pea leaf protoplasts showed that cytosolic NADP redox states were relatively independent of light, dark, or CO2 availability (Igamberdiev and Gardeström, 2003). However, cytosolic NAD and both mitochondrial NAD and NADP pools were more reduced in the light, particularly under conditions favouring photorespiration (Igamberdiev and Gardeström, 2003). It was proposed that these changes are an important factor in mediating light-induced redirection of leaf metabolism, particularly via inhibition of the mitochondrial NAD-ICDH caused by more reduced NAD pools (Igamberdiev and Gardeström, 2003). Such phenomena could be at least partly responsible for some of the changes in metabolite profiles observed in the tobacco CMS mutant, where increases in total leaf NAD+ and NADH are associated with enhanced nitrate reduction and an increased citrate:2-oxoglutarate ratio despite enhanced extractable NAD-ICDH activity (Dutilleul et al., 2005). However, as for almost all other similar work, pyridine nucleotide measurements in these studies represent total nucleotide pools, i.e. they include nucleotides that are bound to proteins as well as the more biochemically relevant ‘free’ pools. Recent analyses of isolated potato tuber mitochondria in short-term experiments have led to the conclusion that free mitochondrial NADH is maintained at a more constant value than total NAD+ and NAD, at least in organelles isolated from non-photosynthetic cells (Kasimova et al., 2006). Several other recent studies using leaf systems indicate that changes in cytosolic and mitochondrial NAD status are important in the light–dark regulation of nitrate reduction and respiratory pathways associated with nitrogen assimilation (Kaiser et al., 2002; Igamberdiev and Gardeström, 2003; Rachmilevitch et al., 2004; Dutilleul et al., 2005).
Notwithstanding the attention paid to chloroplastic events using non-invasive techniques such as chlorophyll fluorescence, the impact of stress on NAD(P) pools and redox states in the different compartments remains unresolved. NAD(P)H oxidases are among the key players in ROS signalling, and PARP (and perhaps other NAD-consuming enzymes) can be activated by stress. In this context, key questions relate, first, to possible links between NAD degradation/(re)synthesis and redox states of NAD(P) and, secondly, to what extent NAD(P) pools in different compartments are autonomous or interdependent. Given that NAD(P)H oxidases use cytosolic reductant and several NAD-consuming enzymes are located in the nucleus, the cytosol and nucleus could be the key locations for NAD-linked signalling. Further information is required to establish whether chloroplast and mitochondrial pools might be relatively insulated from potentially larger fluctuations in cytosolic NAD. Within this context, the location of the different reactions of the pyridine cycling pathway is an important consideration.
|
Regulation of NAD contents |
|---|
What are the mechanisms that could act to regulate NAD synthesis in plants? In bacteria, in which NAD metabolism involves similar enzyme activities to those found in plants, regulation occurs at transcriptional and post-translational levels. The bacterial nadA, nadB, and nadC genes are grouped in the nad operon under the control of a single promoter (Penfound and Foster, 1996). In S. typhimurium, transcription of nadA, nadB, and pncB (encoding NaPRT) is repressed by the DNA-binding function of the repressor protein NadR (Penfound and Foster, 1999). Binding of NadR to the nadA operator is highly dependent on NAD+, and uninfluenced by physiological concentrations of quinolinate, nicotinate, nicotinamide, NaMN, NMN, NaAD, NADH, or NADP+ (Penfound and Foster, 1999). NadR is an intriguing protein for which two additional activities have been described. First, as well as the DNA-binding repressor function that is favoured in NAD+-replete cells, NadR could act in NAD+-deficient cells as a transport facilitator in conjunction with the pnuC gene product, encoding a NMN transporter protein (Penfound and Foster, 1999). Secondly, in addition to repressor and transport functions, NadR also has dual nicotinamide riboside kinase/NMNAT activity. These activities have been described for NadR from several bact