Glycine decarboxylase (GDC, also named glycine-cleavage-system or glycine dehydrogenase) is a multi-protein complex that occurs in all organisms, prokaryotes and eukaryotes. GDC, together with serine hydroxymethyltransferase (SHMT), is responsible for the inter-conversion of glycine and serine, an essential and ubiquitous step of primary metabolism. In Escherichia coli, 15% of all carbon atoms assimilated from glucose are estimated to pass through the glycine–serine pathway (Wilson et al., 1993). In eukaryotes, GDC is present exclusively in the mitochondria, whereas isoforms of SHMT also occur in the cytosol and, in plants, in plastids. The term ‘glycine–serine interconversion’ might suggest that the central importance of this pathway is just the synthesis of serine from glycine and vice versa. However, in both directions of the concerted reaction of GDC and SHMT, tetrahydrofolate (THF) becomes N5,N10-methylenated making these reactions the most important source of active one-carbon-units for a number of biosynthetic processes such as the biosynthesis of methionine, pyrimidines, and purines (Fig. 1). Glycine and serine itself are precursors for chlorophyll, glutathione, tryptophan, phosphatidylcholine and related phospholipids, and ethanolamine. The role of GDC in all organisms is to interconnect the metabolism of one-, two-, and three-carbon compounds (reviewed by Kikuchi, 1973; Oliver, 1994; Cossins, 2000; Hanson and Roje, 2001; Douce et al., 2001). It is therefore not surprising, that a malfunction of GDC results in serious metabolic consequences. Humans, for example, can suffer from non-ketotic hyperglycinemia, an inherited and incurable disease with devastating and often lethal symptoms (Kure et al., 1997). Plants are not able to perform oxygenic photosynthesis without GDC or SHMT and, with reduced activities of these enzymes, will usually show severe growth retardation (Somerville, 2001; Wingler et al., 1997; Heineke et al., 2001).
View larger version (30K):
[in this window]
[in a new window]
Fig. 1. Schematic presentation of the glycine–serine interconversion and its connection to one-carbon metabolism in different subcellular compartments. Circles P, T, H, and L represent the four protein components of glycine decarboxylase and circle S represents serine hydroxymethyltransferase (Cossins, 2000; Ravanel et al., 2001).
Compared with other organisms, the photorespiratory pathway of plants provides a novel role for both GDC and SHMT. In plants, GDC and SHMT are integral components of primary metabolism not only in the context of ‘house-keeping’ glycine–serine interconversion as discussed above. Their additional function in plants is the breakdown of glycine that originates, after several enzymatic reactions, from the oxygenase reaction of Rubisco (Bowes et al., 1971; Tolbert, 1973). By this side reaction of oxygenic photosynthesis, 2-phosphoglycolate is produced and, by the action of ten different enzymes including GDC and SHMT, is subsequently recycled as 3-phosphoglycerate to the Calvin cycle. The contributing enzymes are localized in three different organelles, chloroplasts, peroxisomes, and mitochondria. In C3 plants, if grown under illumination in ambient air, glycine synthesis occurs at very high rates and requires a high capacity for mitochondrial glycine oxidation. In fact, glycine is the preferred substrate of mitochondria and becomes very rapidly oxidized (Day et al., 1985; Krömer and Heldt, 1991) leading to relatively low glycine concentrations in leaves (Leidreiter et al., 1995).
GDC, under unstressed conditions, represents the sole source of photorespiratory CO2 and NH3 and functions as an important link between photorespiration and other metabolic pathways such as nitrate and ammonia assimilation. Much of the earlier work on photorespiration was directed towards attempts to reduce the massive net CO2 losses that occur in C3 plants especially in warm environments. From research conducted over the past 20 years, it is now clear that attempts to abolish or even reduce photorespiration by reducing the activity of individual enzymes of the photorespiratory pathway, except ribulose-1,5-bisphosphate oxygenase, will not lead to improved plant performance.
What then can be the purpose of continuing attempts to manipulate glycine decarboxylation genetically? Firstly, it appears that regulatory interactions exist between photorespiration and photosynthesis triggered by metabolite levels. The nature of these interactions is not well understood. Secondly, the glycine–serine interconversion, by providing one-carbon units, is directly related to many biosynthetic processes outside the photorespiratory pathway. Finally, in photosynthesizing organs of C3 plants, GDC is the major source of internally generated CO2 and, as will be discussed in more detail later, may influence CO2 concentration gradients within leaves.
Several excellent recent reviews cover different aspects of the biochemistry and enzymology of glycine decarboxylation and its relation to plant metabolism (for example Douce et al., 2001; Mouillon et al., 1999; Hanson and Roje, 2001). In this review, these aspects will only be discussed briefly, instead the focus will be on the underlying genetics and on the results obtained with mutants and transgenic plants. As stated above, GDC closely co-operates with SHMT both during the photorespiratory decarboxylation of glycine and the supply of one-carbon units for other biosynthetic processes. Therefore, both GDC and SHMT will be covered in this survey.
Protein components and reactions of the glycine–serine interconversion
Top
Abstract
Introduction
Protein components and reactions...
Possible contributions of...
Genetic manipulation of glycine...
Conclusions
References
The general course of the individual reactions is well known from the work of several groups over many years (Kikuchi, 1973; Oliver, 1994; Bourguignon et al., 1988; Walker and Oliver, 1986a). More details of the involved catalytic mechanisms can be expected from crystallographic data in the near future. Strongly simplified, the course of the reactions in the context of the photorespiratory pathway can be described by the following equations:
GDC:
Glycine + NAD+ + THF -> Methylene-THF + CO2 + NH3 + NADH
SHMT:
Glycine + Methylene-THF + H2O -> Serine + THF
GDC/SHMT:
2 Glycine + NAD+ -> Serine + CO2 + NH3 + NADH
GDC comprises four protein components (Fig. 1). All four individual proteins, which have been designated P, T, H, and L protein, are nuclear encoded and targeted into the mitochondrial matrix.
P protein (EC 1.4.4.2)
P protein, a pyridoxal-5-phosphate containing homodimer of about 200 kDa, is the actual glycine decarboxylating subunit. P protein has also been identified as the binding protein of a host-specific toxin, victorin (Wolpert et al., 1994). The product of the P protein-catalysed decarboxylation of glycine is CO2 and not bicarbonate (Sarojini and Oliver, 1983). The remaining amino methylene moiety is transferred to the distal sulphur atom of the oxidized lipoamide arm of H protein (Douce et al., 2001).
H protein
H protein, a 14 kDa lipoamide (5[3-(1,2) dithiolanyl] pentanoic acid) containing non-enzyme protein, interacts as a co-substrate with all three enzyme proteins of the complex. The three-dimensional structures of all forms of H protein have been resolved (Pares et al., 1994, 1995; Cohen-Addad et al., 1995; Macherel et al., 1996; Faure et al., 2000; reviewed in Douce et al., 2001). Lipoylation of H protein is catalysed by a lipoate–protein ligase (Wada et al., 2001a) and occurs after import of the apoprotein into the mitochondria (Fujiwara et al., 1990) where lipoic acid is synthesized from fatty acid precursors (Wada et al., 1997). Once aminomethylated, the lipoate arm becomes locked within a cleft at the surface of the H protein and released only by interaction with T protein which induces a change in the overall conformation of the H protein (Douce and Neuburger, 1999). In some plants, tissue-specific alternative splicing results in two H proteins with or without an N-terminal extension of two amino acids. The possible effects of this extension onto the H protein’s properties are not yet known (Kopriva et al., 1995a, 1996a).
T protein (E.C. 2.1.2.10)
T protein, a 45 kDa monomeric aminomethyl transferase, needs THF and H protein as co-substrates. One of the conserved domains of T protein shows significant similarity to a domain of formyltetrahydrofolate synthetase from both prokaryotes and eukaryotes suggesting that T protein is not as unique as generally thought (Kopriva et al., 1995b). T protein takes over the aminomethylene group for further processing. The methylene group becomes transferred to tetrahydrofolate resulting in the synthesis of N5,N10-methylene tetrahydrofolate (CH2-THF) and NH3 is released. During these reactions, the lipoamide arm of H protein becomes full reduced and, to be ready for the next cycle, needs to be re-oxidized.
L protein (EC 1.8.1.4)
This reoxidation is achieved by the L protein (dihydrolipoamide dehydrogenase, LPD). L protein is present as a homodimer of about 100 kDa containing FAD as a co-enzyme. During the oxidation of reduced H protein, FAD is reduced to FADH2 which, in turn, becomes immediately reoxidized by NAD+ resulting in the synthesis of one NADH per decarboxylated glycine. The three-dimensional structure of L protein has been resolved (Faure et al., 2000).
L protein is a component not only of GDC but, as the so-called E3 subunit, also of {alpha}-ketoacid dehydrogenase complexes, namely pyruvate dehydrogenase, {alpha}-ketoglutarate dehydrogenase and the branched chain {alpha}-ketoacid dehydrogenase complex (Luethy et al., 1996). By contrast with pea, where it was reported that mitochondrial L protein is encoded by a single gene and shared between {alpha}-ketoacid dehydrogenase complexes and GDC (Turner and Ireland, 1992; Bourguignon et al., 1992, 1996), two genes encoding mitochondrial L protein (mtLPD1 and mtLPD2) have been reported for Arabidopsis thaliana. mtLPD1, seems to provide L protein for GDC whereas the mtLPD2 gene product mainly interacts with {alpha}-ketoacid dehydrogenases (Lutziger and Oliver, 2001). However, from the high sequence identity of 92%, the authors conclude that both L proteins can work in either multienzyme complex. In a more recent analysis of the mRNA and subunit protein levels of the pea leaf mitochondrial pyruvate dehydrogenase complex it was shown that, in sharp contrast to all other subunits, the activity of the E3 subunit (L protein) was highest in mature, fully expanded leaves, reflecting its role as a component of GDC (Luethy et al., 2001). Pea chloroplasts contain a lipoamide dehydrogenase that is different from the mitochondrial isoenzyme (Conner et al., 1996). Similarly, two plastidic LPD genes were identified in Arabidopsis thaliana that are only 33% identical to their mitochondrial counterparts (Lutziger and Oliver, 2000). Apparently, the plastidic LPD is part of the plastidic pyruvate dehydrogenase. There is experimental evidence that LPD is present in soybean nodules, too, and that this LPD is identical to ferric leghaemoglobin reductase-2 (Moran et al., 2002).
Molecular interactions between GDC components
In green leaves, GDC can be present in concentrations of up to 200 mg ml–1 (Oliver, 1994; Douce et al., 1994). The ratio of the protein subunits has been roughly estimated as 4P:27H:9T:2L (Oliver et al., 1990). It is not yet well understood how the GDC subunits interact with one another. They are probably able spontaneously to assemble within the mitochondrial matrix as can be concluded from their behaviour in vitro at protein concentrations above 0.25 mg ml–1 with the H protein possibly building a kind of central core (Oliver et al., 1990; Oliver, 1994) or the ‘structural and mechanistic heart’ of the complex (Douce et al., 2001). Structure–function relationships of and between the individual subunits are now becoming clearer from crystallographic data for some of the respective proteins and the analysis of their interaction by nuclear magnetic resonance studies (Faure et al., 2000; Neuburger et al., 2000; Pares et al., 1995; Douce et al., 2001). Several lines of evidence strongly suggest that, except the catalytic interaction with the lipoyl arm, there is no apparent molecular recognition and interaction between L protein and the reduced H protein. It is assumed that the main role of H protein could be to maintain the hydrophobic lipoate in a state that is freely accessible to the catalytic site of the L protein (Faure et al., 2000; Neuburger et al., 2000). As far as is known, no crystallographic data are available for the T protein and for the P protein.
Corresponding cDNAs and genes have been cloned and analysed over the last ten years by several groups and from different plant sources. More recently, sequences of GDC genes became available from genome and full-length cDNA sequencing projects for a vast number of organisms. Some of the genes and their expression behaviour have been analysed in more detail (Macherel et al., 1992; Srinivasan and Oliver, 1995; Kopriva et al., 1995a; Bauwe et al., 1995; Vauclare et al., 1998). For several genes encoding GDC subunits, induction by light has been observed (Walker and Oliver, 1986b; Kim et al., 1991; Macherel et al., 1990; Turner et al., 1992b; Vauclare et al., 1998; Ma et al., 2001). In the case of H protein and SHMT, negative effects of methyljasmonate on the transcript levels were reported (Schenk et al., 2000).
SHMT (EC 2.1.2.1)
SHMT (also named glycine hydroxymethyltransferase), a tetramer of pyridoxal-5-phosphate containing 53 kDa subunits, catalyses the reversible conversion of serine and THF to glycine and N5,N10-methylene THF (Schirch, 1982; Mouillon et al., 1999). In photosynthetic cells, by their high photorespiratory production of glycine, the mitochondrial SHMT reaction flows in the reverse direction, i.e. towards the synthesis of serine.
During the photorespiratory decarboxylation of glycine in plants, a high mitochondrial activity of SHMT is needed not only to synthesize serine but also permanently to recycle the methylenated THF to THF for its reuse in the GDC reaction. It was shown that CH2-THF is not perfectly channelled between T protein and SHMT and that high CH2-THF/THF rates prevail during steady-state glycine oxidation in mitochondrial matrix extracts (Rebeille et al., 1994).
By contrast with animal cells, which need an external supply of folate (Appling, 1991), plant cells are able to synthesize folate in their mitochondria. Plant mitochondria contain 100–150-fold more THF than chloroplasts (Neuburger et al., 1996; Ravanel et al., 2001). The cytosolic concentrations have not yet been estimated. The mitochondrial CH2-THF/THF pool does not equilibrate with the cytosolic or plastidic pools (Bourguignon et al., 1988; Mouillon et al., 1999). Therefore, it is not regarded as a direct major source of one-carbon units for biosynthetic reactions outside the mitochondria (Mouillon et al., 1999).
SHMT is present not only in mitochondria but in at least two other intracellular compartments, the cytosol and the chloroplasts (Turner et al., 1992a; Besson et al., 1995). The photorespiratory cycle is thus able, via export of serine, to provide one-carbon units for use in biosynthetic pathways outside of the mitochondria. It is assumed that cytosolic SHMT represents the major source of one-carbon units for biosynthetic reactions within the cell including chloroplasts and cytosol (Appling, 1991; Mouillon et al., 1999). CH2-THF itself can be converted to methyl-, methenyl- and formyl-THF thus providing one-carbon units for a number of different biosynthetic reactions, such as the biosynthesis of methionine, purines, pyrimidines, and lipids, not only in plants but in all organisms (Cossins and Chen, 1997; Hanson et al., 2000; Hanson and Roje, 2001).
Possible contributions of glycine decarboxylase for the evolution of C4 plants
Top
Abstract
Introduction
Protein components and reactions...
Possible contributions of...
Genetic manipulation of glycine...
Conclusions
References
The majority of C4 plants evolved about six to eight million years ago under conditions of relatively low atmospheric CO2 concentrations that, by favouring energetically wasteful photorespiratory processes, increase the so-called Rubisco penalty (Edwards et al., 2001). C4 photosynthesis evolved polyphyletically and differs from the ancestral C3 photosynthesis in a number of features.
The major achievement of C4 plants relative to C3 plants is the presence of a highly efficient CO2 concentrating mechanism, the C4 cycle, leading to CO2 levels within the bundle-sheath of C4 plant leaves in excess of 20 times atmospheric concentrations (Hatch, 1987; Kellog, 1999). Besides other effects, this results in a suppression of primary photorespiration (rates of internal CO2 generation by decarboxylation of glycine) by greatly reduced synthesis of phosphoglycolate, the initial substrate of the photorespiratory carbon oxidation cycle. Usually, C4 plants show a specialized leaf anatomy, ‘Kranz’ anatomy (Haberlandt, 1914), with two distinctive and co-operating types of photosynthetic cells, namely mesophyll and bundle-sheath cells. Very much like the enzymes of the photosynthetic carbon reduction cycle and the decarboxylating enzymes of the C4 cycle, GDC is present in the bundle-sheath but not in the mesophyll of C4 plant leaves (Ohnishi and Kanai, 1983). Photorespiratory CO2 is therefore released only within the bundle-sheath and becomes efficiently recaptured. Collectively, these related biochemical and cell-biological aspects of C4 photosynthesis result in the high CO2 assimilation rates of C4 plants, even under conditions of low stomatal conductance (for a recent comprehensive treatise see Sage and Monson, 1999).
Several recent reports provide evidence that C4 photosynthesis does not necessarily require Kranz anatomy. This has been shown for two species of the Chenopodiaceae family, namely Borszczowia aralocaspica and Bienertia cycloptera (Voznesenskaya et al., 2001b, 2002; reviewed in Sage, 2002). In these two succulent halophytic plants, C4 photosynthesis is accomplished by the separation of two types of chloroplasts and other organelles between the two opposite ends (B. aralocaspica) or between two concentric cytoplasmic layers (Bi. cycloptera) of the individual chlorenchymatic cells. Chloroplasts in the distal (B. aralocaspica) or outer (Bi. cycloptera) cytosolic layer, respectively, in contrast to the more proximally or centrally located chloroplasts, for example, lack grana and do not accumulate starch but contain most of the pyruvate orthophosphate dikinase. Moreover, mitochondria were found exclusively within the central cytoplasmic layer of Bi. cycloptera. It is not yet clear whether this type of C4 photosynthesis is related to the evolution of the ‘classical’ Kranz-type of C4 photosynthesis in the Chenopodiaceae or, alternatively, represents a separate ‘non-classical’ type of C4 photosynthesis. It is important to note that this single-cell type of C4 photosynthesis is not typical for the large number of C4 plants present in this family and has not yet been found within other plant families. Hence, single-cell C4 photosynthesis as found in B. aralocaspica and Bi. cycloptera could possibly be a relatively rare adaptation to salty habitats.
Apparently, C4 photosynthesis must have evolved step-by-step by the successive modification of many genes. In genera of several families, species have been identified that possess no or no fully developed C4 cycle (Rawsthorne and Bauwe, 1998). Probably not all these C3–C4 intermediate plants can be regarded as derived from extinct predecessors of C4 plants. However at least in Flaveria, a genus that includes a relatively broad range of species with varying degrees of C3/C4 photosynthesis, phylogenetic studies strongly suggest that C3–C4 intermediate representatives can be regarded as being derived from the extinct evolutionary links between C3 and C4 Flaveria species (Kopriva et al., 1996b).
One of the most characteristic features of C3–C4 intermediate plants, relative to C3 plants, are high reassimilation rates for photorespiratory CO2 leading to greatly reduced rates of apparent photorespiration (Holbrook et al., 1985; Bauwe et al., 1987). On a biochemical level, leaves of C3–C4 intermediate plants contain relatively high concentrations of glycine (Holaday and Chollet, 1984). Significant progress has been made in the explanation of the underlying molecular and cell-biological events, but they are still far from being fully understood (Rawsthorne, 1992; Rawsthorne and Bauwe, 1998). According to current knowledge, both the mesophyll and the bundle-sheath of C3–C4 intermediate plant leaves contain functionally complete carbon reduction cycles. In contrast to the bundle-sheath cells, however, that contain the full enzyme set of the photorespiratory cycle the mesophyll mitochondria of C3–C4 intermediate plants lack at least one of the GDC subunits rendering the enzyme inactive (Hylton et al., 1988; Morgan et al., 1993). It was concluded that most of the photorespiratory glycine produced in the mesophyll of C3–C4 intermediate plants moves to the bundle sheath, where it can be decarboxylated.
Interestingly, at least one C3–C4 intermediate plant, Salsola arbusculiformis, has been identified within the Chenopodiaceae (Voznesenskaya et al., 2001a). The leaf anatomy of this plant, in contrast to the Salsoloid Kranz leaf anatomy that is typical for Salsola C4 species, was described as being Kranz-like. The authors did not specifically examine the intercellular distribution of GDC in this species but found that a very high fraction of bundle-sheath cell volume is occupied by mitochondria (50% relative to the respective chloroplast volume). This suggests the possibility of a similar role for GDC in S. arbusculiformis as in C3–C4 intermediate plants from other families.
Except GDC, there is no other enzyme which is specifically confined to the bundle-sheath of C3–C4 plants. As already mentioned, GDC is the major source of CO2 internally generated from photorespiratory processes. Estimates concerning the rate of photorespiratory CO2 release in C3 plants vary, depending on the method used for determination, from about 25% to about 100% of net photosynthesis (Zelitch, 1979; Peterson, 1983; Pärnik and Keerberg, 1995). Despite these uncertainties it can be stated that, as a general rule, rates of photosynthesis and photorespiration are of the same order of magnitude in C3 plants. In the mesophyll of C3 plant leaves, photorespiration moves freshly assimilated carbon from the chloroplasts into the mitochondria where it is released as photorespiratory CO2. Due to corresponding high carbon fluxes this process can be compared with a carbon concentrating mechanism that is futile because the inlet (chloroplasts) and the outlet (mitochondria) are present within the same cell.
The situation is different in C3–C4 intermediate plants where photorespiratory glycine is produced with high rates both in the mesophyll and in the bundle-sheath, but can be decarboxylated only by the mitochondria of the bundle-sheath. It is therefore tempting to speculate that these combined features may result in elevated CO2 concentrations within the bundle-sheath. This hypothesis has been tested by quantitative 14CO2 labelling experiments with leaves of several Flaveria species designed to permit the determination of in vivo carboxylation/oxygenation ratios of ribulose-1,5-bisphosphate. These data indicate that the C3–C4 intermediate species Flaveria anomala has about a 2-fold increased carboxylation/oxygenation ratio of ribulose-1,5-bisphosphate ratio relative to the C3 plant Flaveria cronquistii. Because there are no significant differences between these two species in their in vitro affinity of Rubisco to CO2 and O2 (Bauwe, 1984) it was concluded that Rubisco operates under an approximately doubled mean CO2 concentration in leaves of the C3–C4 intermediate plant (Bassüner, 1985; U Bauwe and O Keerberg, unpublished data).
From all the findings discussed above it is most likely that the photorespiratory cycle of C3–C4 intermediate plants, by the exclusive presence of GDC in the bundle-sheath and by using glycine as the vehicle, is capable of transporting large amounts of freshly assimilated carbon from mesophyll chloroplasts (the ‘pump’s’ inlet) to bundle-sheath mitochondria (the ‘pump’s’ outlet) where it is released as photorespiratory CO2 leading to elevated CO2 concentrations within the bundle-sheath. This glycine-to-serine conversion possibly provided one of the biochemical starting points for the evolution of C4 plants.
There are many other open questions related to the evolution of C4 from C3 via C3–C4 intermediate plants. For example, some characteristics of C4-photosynthesis have been reported for the cells surrounding the vascular bundles in stems and petioles of C3 plants like tobacco or celery (Hibberd and Quick, 2002). It must also be mentioned that the possible effects of a relocation of GDC for the evolution of C4 plants are controversial (Monson, 1999; Edwards et al., 2001). It cannot be ruled out that, in different families, quite different evolutionary scenarios led to the evolution of C4 plants.
In the authors’ opinion, the detailed analysis of the physiological and adaptive implications caused by the presence of a primary CO2 concentrating mechanism driven by GDC in the context of the photorespiratory cycle is an important key for a better understanding of the evolution of C4 photosynthesis. To test this hypothesis further it appears as an intriguing task to attempt a relocation of GDC in a C3 plant. Such experiments require at least two prerequisites. Firstly, a mutant that does not contain endogenous GDC and, secondly, genes encoding GDC subunits under the control of bundle-sheath specific promoters to supplement the mutant with a functional photorespiratory cycle. Appropriate GDC genes have been cloned and characterized from C3–C4 intermediate and C4 Flaveria species (Chu, 1996; Chu et al., 1998; Nan et al., 1998; Nan and Bauwe, 1998; Cossu, 1997; Cossu and Bauwe, 1998). The current situation with respect to available GDC defective mutants will be discussed below.
Genetic manipulation of glycine decarboxylation
Top
Abstract
Introduction
Protein components and reactions...
Possible contributions of...
Genetic manipulation of glycine...
Conclusions
References
Historically, three programmes for the identification of photorespiratory mutants in chemically mutagenized seed sets were performed. The first mutant screen was devized for Arabidopsis thaliana. The analysis of corresponding mutants was very fruitful for a short time, but has not received very much attention during the last decade (Somerville and Ogren, 1982a; Somerville, 1984, 2001). A second mutant screen was performed with barley (Kendall et al., 1983; Blackwell et al., 1988). The analysis of these mutants has continued over the years (Wingler et al., 2000). The third programme was directed towards the C4 plant Amaranthus edulis (Dever et al., 1995; Wingler et al., 1999). Notably, immunocytochemical studies with this plant indicate that the cell-specific biosynthesis of a number of photosynthetic and photorespiratory enzymes in C4 plant leaves is more complex than has been thought previously (Bailey et al., 2000). In addition, there were some initial reports on the use of mutagenized tobacco callus cultures (Berlyn, 1978; Zelitch and Berlyn, 1982) but these studies apparently have not been continued. More recently, antisense plants with reduced contents of GDC subunits and SHMT were studied (Heineke et al., 2001; Winzer et al., 2001; Bauwe et al., 1999).
Barley mutants obtained by chemical mutagenesis
Two mutants of barley, LaPr 85/55 and LaPr 87/30, were isolated that grow well in 0.7% CO2 but accumulate glycine 5–10-fold relative to wild-type levels and show reduced levels of glutamate and alanine in combination with rapid senescence when exposed to air (Blackwell et al., 1990). These mutants behaved differently insofar that LaPr 85/55 was able to metabolize much more 14C-glycine into sugars than LaPr 87/30 after 2 h (70% and 4%, respectively). SHMT activity was not affected, however, the data corresponded well with reduced GDC activities (measured via the glycine–bicarbonate exchange reaction that needs H and P protein, but no T or L protein) with 70% wild-type activity with LaPr 85/55 and only 14% with LaPr 87/30, respectively. Protein blotting showed severely reduced levels of P and H protein (10% remaining) and a slight reduction in T protein (50%). Supply of 40 mM serine through the xylem stream was able to at least partially (70%) restore wild-type 14CO2 fixation rates for both mutants. A mutation in a glycine transporter was suggested for LaPr 85/55 (Blackwell et al., 1990) and a reduction in H protein down to 1% relative to wild-type has been shown for homozygous LaPr 87/30 plants (Wingler et al., 1997).
From a more detailed analysis of LaPr 87/30, including heterozygote lines, it was concluded that the biosynthesis and activity of GDC biosynthesis in vivo is determined by the biosynthesis of H protein. More specifically, P protein content in LaPr 87/30 heterozygous lines was reduced by 25% but GDC activity increased linearly with increasing H protein content. The authors also suggested that photorespiratory carbon flux is not restricted by GDC activity (Wingler et al., 1997, 2000).
Studies with LaPr 87/30 on a cellular level revealed significant changes in the redox status of the cells such as over-reduction and over-energization of chloroplasts (Igamberdiev et al., 2001a). Surprisingly, these studies also revealed a rate of glycine oxidation both in leaf cuttings and in intact mitochondria of 30–40% relative to the wild type. However, the authors could not exclude that this effect was due to growth of the plants under low light which might result in lower GDC levels in wild-type plants. In addition, the level of alternative oxidase was reduced. It was also shown that 13C/12C isotope fractionation is higher in LaPr 87/30 relative to the wild type (Igamberdiev et al., 2001b).
The TIGR Barley Gene Index (
http://www.tigr.org/tdb/hvgi/index.html), at the time of this writing, includes two entries for tentative consensus sequences (TC) corresponding to H protein genes, TC8419 (green leaf preference) and TC8850 (root and caryopsis preference). The strong metabolic effects, as described above, suggest that the gene corresponding to TC8419 is affected in LaPr 87/30. It also indicates that the second gene is not able to take over the tasks, most likely because of its preferential expression in non-photosynthetic organs.
Arabidopsis thaliana mutants obtained by chemical and insertional mutagenesis
About 20 years ago, the use of A. thaliana in a genetic approach to resolve controversial ideas about the mechanism of photorespiration led to the identification of a number of mutants with defects in enzymes of the photorespiratory cycle (for a historical view compare Somerville, 2001). Unfortunately, as mentioned above, following their initial characterization (Somerville and Ogren, 1982b, 1981), not much effort has been put into a more detailed analysis of mutants where genes encoding GDC subunits or SHMT were affected.
The availability of the complete genome nucleotide sequence of A. thaliana (The Arabidopsis Genome Initiative, 2000) revealed the existence of small multi-gene families for all GDC components, except T protein which is encoded by a single gene, and SHMT (Table 1). This knowledge opens new opportunities for a closer investigation of the genetics and transcriptional regulation of corresponding genes, for example, during the ontogenetic development of A. thaliana.
View this table:
[in this window]
[in a new window]
Table 1. Summary of genes encoding GDC subunits or SHMT in A. thaliana (The Arabidopsis Genome Initiative, 2000) Designation of SHM1–SHM5 corresponds to the proposal by McClung et al. (2000). Direct experimental evidence for the predicted subcellular localization is not available.
P protein is encoded by two genes, AtGDP1 and AtGDP2. The derived proteins are 90% identical to each other. Two loci, gld1 (originally named glyD) and gld2, have been identified by chemical mutagenesis (Somerville and Ogren, 1982b; Artus et al., 1994). The major characteristics of gld1 were high accumulation of glycine under normal air, no decline in glycine concentrations during a following dark period, reduced rate of photosynthesis, no glycine oxidation by isolated mitochondria, and no glycine–bicarbonate carbon exchange. By the osmotic-swelling technique, no indication could be found for an impaired glycine transport into mitochondria (Somerville and Ogren, 1982b). The affected locus was mapped to chromosome 2 about 40 cM from the er-py region (Artus et al., 1994). This chromosome harbours one gene encoding P protein, AtGDP2, and two genes encoding H protein, AtGDH1 and AtGDH2, however, the mapped position of gld1 does not correspond with any of these loci (Fig. 2).
View larger version (9K):
[in this window]
[in a new window]
Fig. 2. Approximate positions of genes encoding GDC protein components and SHMT on Arabidopsis thaliana chromosomes 1 to 5.
Very similar to gld1, mutation of gld2 reduced glycine–bicarbonate exchange rates by 70–80% and glycine oxidation by isolated mitochondria by more than 90%. The affected locus was mapped to chromosome 5 at a distance of about 21 cM from tt3 (Artus et al., 1994). From their biochemical data and under the assumption that gld1 represents a GDP locus, the authors hypothesized that the gld2 mutation most likely represents a defect in the H or T protein or in glycine transport into the mitochondria (Artus et al., 1994). However, the nucleotide sequence of chromosome 5 does not contain a gene for a functional GDC subunit. These data support the idea that glycine transport instead of GDC biosynthesis could be affected as it has already been suggested as a possible alternative by Artus et al. (1994). Unfortunately, knowledge about glycine transport into the mitochondria is very limited. Although 20 years ago it was suggested that glycine/serine antiporters might reside in the inner mitochondrial membrane (Walker et al., 1982) such transporters have not yet been identified (Oliver, 1994; Laloi, 1999).
These data suggest that, most likely, neither gld1 nor gld2 represent genes encoding GDC components. At least theoretically, similar metabolic effects as observed with gld1 and gld2 could be induced, for example, by mutation of the lipoate–protein ligase that is required for the lipoylation of H protein at the {epsilon}-amino group of a lysine residue. In A. thaliana, both a mitochondrial (LIP2, At1g04640, Wada et al., 2001a) and a plastidic form (LIP2p, At4g31050, Wada et al., 2001b) have been cloned and characterized. However, due to their chromosomal location, these genes are clearly no candidates for the loci defined by mutations gld1 and gld2.
L protein is encoded by four genes in A. thaliana, two of each encoding plastidic (Lutziger and Oliver, 2000) and mitochondrial lipoamide dehydrogenases (Lutziger and Oliver, 2001). Although the genes encoding mitochondrial proteins are expressed in all organs, the isologue genes show distinctly different expression patterns both with respect to their organ preference and their response to light. An insertional knockout mutant for AtLPD2 did not show any apparent morphological phenotypic change. By contrast to the unchanged CO2 release from pyruvate, a 25% reduction in 14CO2 release from [1-14C]glycine was observed. It was suggested that the two proteins, once in the mitochondrial matrix, are interchangeable among the different multienzyme complexes of GDC and {alpha}-ketoacid dehydrogenases (Lutziger and Oliver, 2001).
Knowledge about H and T proteins in Arabidopsis is much more limited. With three members, H-protein is the only GDC subunit that is encoded by a multigene family. In addition, a pseudogene exists on chromosome 5 (F26C17). Notably, the AtGDH1 and AtGDH3 encoded proteins are 92% identical to each other but only about 60% identical to the homologue protein encoded by AtGDH2. In promoter studies and other experiments with AtGDH1, transcriptional activation by light was shown (Srinivasan and Oliver, 1992).
T protein is the only GDC subunit that is encoded by a single-copy gene in A. thaliana. This singular occurrence could indicate a central role of T protein in the regulation of GDC biosynthesis and might explain the, as yet unsuccessful, search for insertional mutants for this gene in this laboratory (Ü Kolukisaoglu and H Bauwe, unpublished data).
An inspection of the A. thaliana genome sequence reveals the presence of seven SHM genes in A. thaliana, AtSHM1 to AtSHM7 (Table 1; Fig. 2). For reasons of conformity, the designation of genes encoding SHMT as SHM will be adopted (instead of STM) as suggested by McClung et al. (2000).
Recent studies have shown that AtSHM1 expression is high in leaves with light inducibility, suggesting that SHM1 encodes a photorespiratory SHMT, and circadian oscillations in transcript abundance. Similar to AtSHM1, the expression of AtSHM2 is strongly induced by light in leaves, but not in roots. AtSHM4 is expressed with low abundance only in roots and in flowers. This gene does not show a light response but, like AtSHM1, shows circadian oscillations as well (McClung et al., 2000; Ho et al., 1999). Using a positional cloning approach, an A. thaliana SHM1 mutant has been identified (Renné et al., 2001). This mutant is unable to grow under ambient conditions, but can be recovered under 1500 ppm CO2. Biochemical data are not yet available.
Meanwhile, the complete cDNA sequence of AtSHM3 is available from the RAFL project (Seki et al., 2002). From these new data and the correction of the deduced N-terminus, a plastidic targeting appears as more likely than the formerly assumed cytosolic localization (McClung et al., 2000). The proteins encoded by AtSHM6 and AtSHM7 differ from those encoded by AtSHM1-AtSHM5 by long N-terminal extensions of more than 100 amino acids. According to PSORT (
http://psort.nibb.ac.jp) and TargetP (
http://genome.cbs.dtu.dk/services/TargetP/) these proteins are candidates for becoming targeted to the nucleus.
Three allelic A. thaliana putative SHM mutants have been isolated following chemical mutagenesis in Ogren’s laboratory (Somerville and Ogren, 1981). They showed severe growth retardation under ambient air conditions and, like the other photorespiratory mutants, plants had to be grown under an elevated CO2 concentration. Total SHMT activity in leaves was about 15% relative to the wild-type total and the mutants were shown to be deficient both in glycine decarboxylation and in the conversion of glycine to serine. The authors concluded that these mutants do not possess any mitochondrial SHMT at all. However, this conclusion could not be supported by more recent data which indicate that the level of SHM1 transcripts is unaltered in the stm mutant (Beckmann et al., 1997). The locus affected in this mutant (Nottingham Stock Centre N8010) has been mapped to chromosome 5 (A Weber, personal communication). More recent fine mapping data support this result but, surprisingly, indicate that the stm locus is probably not related to SHM2 or to any other of the seven SHM genes in A. thaliana (Schilling et al., 2001). This supports the view that the stm mutation might affect a locus that is required for SHMT activity, but is distinct from SHM loci encoding SHMT protein (McClung et al., 2000).
Taken together, the mutant data discussed above, especially those related to P and L protein of GDC and those related to SHMT in A. thaliana, raise several questions. First of all, the two loci gld and smt are probably only indirectly related to the proper function of GDC and SHMT. Nevertheless, the quite massive effects of the respective mutations indicate important, but as yet unknown, functions in glycine–serine metabolism. Secondly, P protein, mitochondrial L protein and mitochondrial SHMT are all encoded by two genes in A. thaliana. It is not known, whether the respective genes are equally important or, alternatively, whether they serve different functions in different organs or developmental contexts.
The recent publication of a draft sequence of the rice genome (Yu et al., 2002; Goff et al., 2002) opened the opportunity to compare the number and structure of GDC and SHMT encoding genes, respectively, in a monocotyledonous plant. Searches for GDC and SHMT encoding sequences in the genome of rice and comparisons to their orthologues in A. thaliana revealed two important insights. First, the genomic structure of these genes and the deduced protein structures are very similar in both species. Second, and perhaps more important, the number of rice homologues to the A. thaliana GDC and SHMT genes seems to be roughly equal. For instance, only one orthologue of AtGDT, a single copy gene in A. thaliana, was found in the rice genome. There are also seven rice OsSHM genes with exon–intron boundaries identical to those found in the different AtSHM genes (data not shown). Due to the preliminary character of the rice genome sequence these data just represent estimations. However, it can be already concluded that the information about content and organization of GDC and SHMT genes extracted from the A. thaliana genome is transferable to a cereal. On the basis of this knowledge from two distantly related model plants it is likely that a similar organization of photorespiratory genes exists in other angiosperms, too.
Plants with reduced content of GDC subunits or SHMT by antisense approaches
Mutational approaches provide the possibility to study the function of individual genes as soon as the mutagenized locus is known. By contrast, antisense or RNAi-based approaches allow the evaluation of more general effects. Such an approach is useful if no mutants are available, for example, due to their lethality, or if the intended studies concern multigene families.
Transgenic potato plants with about 60–70% less P protein relative to wild-type potato plants and a corresponding decrease in the ability of leaf mitochondria to decarboxylate glycine were indistinguishable from wild-type plants when grown under 800 ppm CO2 (Heineke et al., 2001; Winzer et al., 2001). When grown under ambient CO2 and moderate light, there were no clear phenotypic changes, except the early senescence of older leaves. Photosynthetic and growth rates were reduced, but the plants were viable under ambient air and produced tubers. Glycine concentrations, especially in fully expanded leaves, were elevated by up to about 100-fold during illumination. Notably, nearly all of the glycine accumulated during the day in leaves of the antisense potato plants was metabolized during the following night. This was accompanied by distinctly increased levels of serine at the end of the night.
Similarly, leaves of transgenic potato plants with severely reduced amounts of SHMT contained up to 100-fold elevated levels of glycine relative to the wild type. Photosynthesis rates were reduced and the degree of this reduction was correlated with glycine levels, i.e. with the reduction in SHMT activity. These negative effects on growth were greatly elevated by higher light intensity. Two lines were unable to grow in ambient air even under moderate light intensity but could be recovered in 2000 µl l–1 CO2 (Bauwe et al., 1999).
Collectively, the data obtained with transgenic plants suggest that the photosynthetic–photorespiratory metabolism of potato plants responds flexibly to limited changes in the capacity of leaves to decarboxylate glycine. GDC seemingly operates far below substrate saturation in wild-type plants held under ‘normal’ conditions. This provides the opportunity to respond rapidly to enhanced rates of photorespiration as they occur during increased temperatures or under conditions of stomatal closure during periods of insufficient water supply. Under such circumstances, perhaps much like transgenic plants with moderately reduced GDC activity, GDC operates under higher saturation with glycine, thus achieving a similar steady-state throughput as during normal photosynthesis. At least in potato, GDC exerts high control over the level of glycine, but only low control over the flux rates through the interconnected cycles of photosynthesis and photorespiration (Heineke et al., 2001).
Conclusions
Top
Abstract
Introduction
Protein components and reactions...
Possible contributions of...
Genetic manipulation of glycine...
Conclusions
References
The glycine–serine interconversion, catalysed by GDC and SHMT, is an important reaction of primary metabolism in all organisms including plants. Quite generally, this reaction provides one-carbon units for many biosynthetic reactions. In plants, in addition to this general role in metabolism, it is an integral part of the photorespiratory metabolic pathway in which glycine is produced with high rates from Calvin cycle intermediates and converted into serine within the mitochondria. Large amounts of photorespiratory CO2 are produced by this plant-specific pathway. Several lines of evidence suggest that this latter process, by relocation of GDC from one leaf-cell type (the mesophyll) to another (the bundle-sheath) contributed to the evolution of C4 plants. Although this hypothesis is controversial (compare Edwards et al., 2001), it is regarded as most likely that changes in the intercellular distribution of GDC are capable of significantly influencing the concentration of CO2 within the respective cells. If this is true, they will unavoidably modify the relative rates of carbon flux into the photosynthetic carbon reduction cycle and the photorespiratory carbon oxidation cycle thus influencing the efficiency of photosynthesis. Is this influence on the overall efficiency of photosynthesis very minor or is it perhaps of greater significance? In light of the progress made with the analysis of genes and mutants related to photorespiratory processes and the cloning of GDC genes from C3–C4 intermediate plants this question now can be targeted.
----------------------
in short: everyone on this thread, except for me, is a mega-flaming queer.
Please Scroll Down to See Forums Below 
someone who I'm friends with too.
