Plant Physiol, March 2002, Vol. 128, pp. 793-802

UPDATE ON PURINE BIOSYNTHESIS
Purine Biosynthesis. Big in Cell Division, Even Bigger in Nitrogen Assimilation

Botany Department, The University of Western Australia, 35 Stirling Highway, Crawley, Western Australia 6009, Australia

	INTRODUCTION

TOP INTRODUCTION PURINE SYNTHESIS AS A... A SELECT FEW LEGUMES... AND IT IS NODULE... WHY ASSIMILATE N AS... THE DE NOVO PURINE... WE NEED TO KNOW... IS ARABIDOPSIS HELPFUL? THE PURINE PATHWAY HAS... CONCLUSION LITERATURE CITED

Synthesis of the purine ring is a central metabolic function of all cells. The products, AMP and GMP, provide purine bases for DNA and RNA, as well as for a number of essential coenzymes (NAD, NADP, FAD, and coenzyme A) and signaling molecules (e.g. cAMP; Fig. 1). ATP serves as the energy source for many chemical reactions. In addition, in plants, the nucleotides are the precursors for purine alkaloids, and for the adenine moiety of cytokinin plant growth regulators (Fig. 1). Despite the essential functions for purines, salvage pathways, which retrieve the purine ring after nucleic acid or coenzyme breakdown, recycle nucleotides to meet day-to-day needs. Thus, the requirement to synthesize new purines in differentiated cells is small. It is only when DNA is replicated that de novo synthesis of the purine ring is required and so, although present in most tissues, the activity of the metabolic pathway is relatively low.

View larger version (82K): [in this window] [in a new window]   Figure 1.   Schematic diagram for the pathway of de novo purine biosynthesis (the box at the top of the figure) and further metabolism of purines in plant cells. The first committed step of the purine pathway, catalyzed by phosphoribosylpyrophosphate amidotransferase (PRAT), is one of a number of possible metabolic fates for both substrates (phosphoribosylpyrophosphate [PRPP] and Gln [gln]). The final product of the pathway, inosine monophosphate (IMP), is the first purine nucleotide and provides the basic purine ring structure for the other purine nucleotides, xanthosine monophosphate (XMP), AMP, and GMP. The last two of these are the building blocks of DNA and RNA, whereas AMP is the precursor for the cytokinin group of plant growth regulators and a number of important coenzymes. GTP and ATP participate in the energy metabolism of the cell. Although there are a number of routes that can generate the purine bases from IMP, in legume nodules the preferred route is through IMP dehydrogenase (IMPDH). Both xanthosine and xanthine serve as precursors for the purine alkaloids (theobromine and caffeine) and their further oxidation yields the ureides, allantoin and allantoic acid. Enzyme names are abbreviated as follows: glycinamide ribonucleotide synthetase (GARS), glycinamide ribonucleotide transformylase (GART), formylglycinamide ribonucleotide amidotransferase (FGARAT), aminoimidazole ribonucleotide synthetase (AIRS), aminoimidazole ribonucleotide carboxylase (AIRC), succinoaminoimidazolecarboximide ribonucleotide synthetase (SAICARS), adenylosuccinate-AMP lyase (ASAL), aminoimidazolecarboximide ribonucleotide transformylase (AICART), inosine monophosphate cyclohydrolase (IMPCH), and xanthine dehydrogenase/xanthine oxidase (XDH/XO). Substrate names are abbreviated as follows: phosphoribosylamine (PRA), glycinamide ribonucleotide (GAR), formylglycinamide ribonucleotide (FGAR), formylglycinamidine ribonucleotide (FGAM), aminoimidazole ribonucleotide (AIR), carboximideaminoimidazole ribonucleotide (CAIR), succinoaminoimidazolecarboximide ribonucleotide (SAICAR), aminoimidazolecarboximide ribonucleotide (AICAR), formylaminoimidazolecarboximide ribonucleotide (FAICAR), XMP, GMP, Gln (gln), Gly (gly), N10-formyl tetrahydrofolate (N10-formyl THF), Asp (asp), and inorganic phosphate (Pi). Genes encoding the enzymes of the purine pathway are shown in parentheses beneath the enzyme abbreviation.

	PURINE SYNTHESIS AS A PATHWAY FOR N ASSIMILATION?

TOP INTRODUCTION PURINE SYNTHESIS AS A... A SELECT FEW LEGUMES... AND IT IS NODULE... WHY ASSIMILATE N AS... THE DE NOVO PURINE... WE NEED TO KNOW... IS ARABIDOPSIS HELPFUL? THE PURINE PATHWAY HAS... CONCLUSION LITERATURE CITED

The purine pathway also functions in pathways that are different and distinct from these "housekeeping roles." It is employed in specialized tissues to assimilate and detoxify NH₃. In the liver of birds and terrestrial reptiles, purines synthesized de novo are oxidized to uric acid and excreted as the main form of waste N. Plants do not engage in N excretion but rather store it. In members of the Aceraceae, Plantanaceae, and Borraginaceae, N storage is accomplished as purines are oxidized to ureides (allantoin and allantoic acid). In such plants, ureides are the dominant forms of stored N in stems and underground organs (Reinbothe and Mothes, 1962), accounting for as much as one-half the plant's N in some species, such as comfrey (Symphytum offinalis). However, it is in nodules of tropical legumes, such as soybean (Glycine max) and cowpea (Vigna unguiculata), that the pathway plays a dominant role in primary N metabolism (Atkins and Smith, 2000; Fig. 1). Within cells infected with rhizobia, bacterial nitrogenase activity leads to secretion of fixed N, principally as NH₃ or NH₄⁺ (Fig. 2). Although assimilated initially through the amide group of Gln (by Gln synthetase), almost all fixed N is subsequently incorporated through the purine pathway to form IMP, and finally ureides. In these species, ureides are translocated in xylem from nodules and provide the major supply of N for the plant's nutrition (Fig. 2). To accommodate this flux of fixed N in nodules, activity of enzymes in the purine pathway is enhanced considerably compared with other tissues, including active meristems (Atkins and Smith, 2000). For this reason, nodules have been exploited as the tissue of choice in which to study the enzymology of purine biosynthesis in plants.

View larger version (64K): [in this window] [in a new window]   Figure 2.   Two types of cell in the central infected tissue zone of nodules are required for the synthesis of ureides from fixed N. The enlarged infected cells (IC) contain as many as 50,000 rhizobia (Bact), differentiated into bacteroids and enclosed in groups within membrane (peribacteroid membrane) vesicles of plant origin. These bacteria express nitrogenase and export ammonia to the host cell cytosol where it is assimilated as the amide group of Gln (gln). Gln, together with other amino acids, is used by both plastids (P) and mitochondria (M) to form purines. These plant organelles are localized at the periphery of the infected cell and are concentrated particularly adjacent to gas spaces. The purines are oxidized within the infected cell cytosol to urate that is transferred to uninfected cells (UC) of the central zone and further oxidized within enlarged microbodies (Mb) to allantoin and allantoic acid (ureides). Ureides are exported from the nodule in xylem to provide the majority of the N for the plant's nutrition.

	A SELECT FEW LEGUMES CHOOSE THE PURINE PATHWAY

TOP INTRODUCTION PURINE SYNTHESIS AS A... A SELECT FEW LEGUMES... AND IT IS NODULE... WHY ASSIMILATE N AS... THE DE NOVO PURINE... WE NEED TO KNOW... IS ARABIDOPSIS HELPFUL? THE PURINE PATHWAY HAS... CONCLUSION LITERATURE CITED

The formation and translocation of fixed N as ureides is not a feature of all nodulated legumes but is restricted, almost exclusively, to species of the tribes Phaseoleae, Desmodieae, and Indigofereae (Atkins, 1991) within the Phaseoloid group (Doyle et al., 2000). Species in the Phaseoleae and Desmodieae form "desmodioid" nodules. These are typically more or less spherical; as for soybean or cowpea, they develop from infection through root hairs, have a central infected tissue zone that is interspersed with uninfected cells (Fig. 2), and are determinate (Corby, 1988). However, desmodioid nodules are also a feature of the Loteae, a temperate tribe that is taxonomically distant from the Phaseoloid group (Doyle et al., 2000) and does not assimilate fixed N as ureides. Among the Indigofereae, only Cyamopsis tetragonoloba has been reported to translocate significant levels of ureide in xylem (Atkins, 1991) and its nodules are not desmodioid, but indeterminate.

The ureide pathway occurs in determinant nodules where the N-fixing tissue zone includes both infected and uninfected cells (Fig. 2), an arrangement that apparently facilitates the function of uricase, the enzyme catalyzing the ultimate step of purine oxidation (Fig. 1). Uricase requires molecular O₂, but has an apparent K_m (O₂) that is extremely unfavorable (30 µM; Rainbird and Atkins, 1981) in nodules where the average pO₂ in infected cells is maintained at 10 to 60 nM in solution (Kuzma et al., 1993). Thus, it is not surprising that uricase expression is confined to greatly enlarged microbodies in interstitial, uninfected cells of the N-fixing zone (Fig. 2). Here, pO₂ is predicted to be higher and adequate to support urate oxidation at rates commensurate with those of N₂ fixation (Thumfort et al., 1999).

Species of other tribes in the Papilionoideae assimilate fixed N as Gln and Asn and these, especially Asn, are translocated out of nodules. Data for the Caesalpinoideae and Mimosoideae are neither extensive nor well documented (Atkins, 1991). There have been reports of significant ureide translocation in species outside the Phaseoloid group (Atkins, 1991), but a number of these have been because of analytical artifacts (for references, see Atkins and Smith, 2000). Similarly, there have been reports of low levels of ureide in xylem exudate or in extracts of nodules from species that assimilate fixed N principally as Asn (Brown and Walsh, 1994). Thus, in Robinia pseudoacacia, 3% of xylem-N was ureide (Atkins et al., 1991) and in Lotus japonicus, Medicago sativa, and Medicago truncatula (Cheng et al., 2000), levels less than 10% compared with Asn-N have been reported. In these species, uninfected cells differentiate in the central zone and express uricase (Atkins et al., 1991; Tajima et al., 2000). Thus, it seems likely that small amounts of ureide are formed from nucleic acid breakdown accompanying senescence, in much the same way as ureide accumulates transiently in the cotyledons of germinating seeds (Fujihara and Yamaguchi, 1978).

	AND IT IS NODULE SPECIFIC, TOO!

TOP INTRODUCTION PURINE SYNTHESIS AS A... A SELECT FEW LEGUMES... AND IT IS NODULE... WHY ASSIMILATE N AS... THE DE NOVO PURINE... WE NEED TO KNOW... IS ARABIDOPSIS HELPFUL? THE PURINE PATHWAY HAS... CONCLUSION LITERATURE CITED

Ureide synthesis in nodulated legumes has attracted interest in part because many of the species that show this metabolism are important crops (soybean, cowpea, common bean [Phaseolus vulgaris], and mung bean [Vigna radiata]), but also because utilization of this pathway for N assimilation is greatly enhanced in nodules. Roots and other tissues of soybean or cowpea assimilate soil mineral N (NO₃ or NH₄⁺) into the amides Gln and Asn (Atkins and Smith, 2000) and these are the translocated forms of N in both xylem and phloem. Thus, elevated expression of de novo purine synthesis is a specific metabolic feature of the symbiosis in ureide-forming legumes. In fact, the unique association of ureide synthesis with nodules has proven sufficiently specific that an assay for xylem-borne N-solutes has been developed as the basis of a practical field method to estimate relative proportions of fixed and soil N in soybean.

	WHY ASSIMILATE N AS PURINES?

TOP INTRODUCTION PURINE SYNTHESIS AS A... A SELECT FEW LEGUMES... AND IT IS NODULE... WHY ASSIMILATE N AS... THE DE NOVO PURINE... WE NEED TO KNOW... IS ARABIDOPSIS HELPFUL? THE PURINE PATHWAY HAS... CONCLUSION LITERATURE CITED

Purine/ureide synthesis would seem to be an extraordinarily complex mechanism for NH₃ assimilation, requiring enhanced expression of 20 separate enzymes as compared with just four required for Asn synthesis. Furthermore, two of the steps involve one-C additions as formyl-tetrahydrofolate derivatives (Fig. 1), requiring enhanced activity of enzymes for their synthesis also. Thus, it is reasonable to ask: why purine/ureides in some species and not in others and what advantages/disadvantages might accrue as a result of this trait?

Despite the biochemical complexity of the pathway, the "cost" in terms of ATP and reductant expended per N assimilated is not much different from that required for Asn. Ureides are made up of four C and four N atoms (Fig. 1), having a C:N ratio of 1.0 compared with 2.5 and 2.0 for Gln and Asn, respectively. Thus, less C is required for translocation of fixed N as ureide. Theoretical considerations of the likely C to N economy of nodules of amide-forming versus ureide-forming species indicate that this "saving" of translocated C could be significant. An attempt to compare experimentally the C/N economy of an amide former (white lupine [Lupinus albus]) with a ureide former (cowpea) indicated a significantly lower level of C consumption and loss as CO₂ by nodules of the ureide former (Layzell et al., 1979). However, the two symbioses also differed in levels of phosphoenolpyruvate carboxylase and rates of H₂ evolution as well as the proportional distribution of infected and cortical tissue. These, as well as ureide versus amide synthesis, were also likely to affect the relative C costs of fixation. There are no other data available that approach this question, and attempts to isolate mutants of ureide formers that assimilate fixed N as amides have not been successful. As a consequence, whether one strategy for NH₃ assimilation is more or less costly remains unresolved.

There is a growing body of evidence indicating that pathways of C and N assimilation are coordinately regulated and that C to N signaling networks in plants may be a central feature influencing growth and development (Coruzzi and Zhou, 2001). However, whether this extends to the "sensing" of the C to N ratio per se through specific N solutes in sources and sinks and in translocation pathways that link them is not known.

Despite the putative economy of C use in translocating N in the form of ureides, ureides are not as suitable as Asn or Gln as a source of N for amino acid and protein synthesis. Their metabolism in the shoot involves release of all N as NH₃, with the need for reassimilation. Similarly, most of the ureide-C is released as CO₂. As a consequence, little xylem-borne ureide is transferred as such to phloem in leaves but rather its N is reassimilated to Asn and Gln, and these constitute the principal N solutes of the assimilate stream in ureide-forming legumes (Peoples et al., 1985).

	THE DE NOVO PURINE PATHWAY

TOP INTRODUCTION PURINE SYNTHESIS AS A... A SELECT FEW LEGUMES... AND IT IS NODULE... WHY ASSIMILATE N AS... THE DE NOVO PURINE... WE NEED TO KNOW... IS ARABIDOPSIS HELPFUL? THE PURINE PATHWAY HAS... CONCLUSION LITERATURE CITED

Apart from the need for uninfected cells in the nodule central zone so that uricase function is secure, the most significant nodule-specific feature of ureide synthesis is enhanced expression of genes encoding the enzymes in the purine pathway. Recent studies have concentrated on the molecular aspects of gene expression and regulation in relation to N metabolism.

Most of the initial work on the enzymology of the purine pathway in plants was based on information gained from other organisms. The pathway, first characterized in prokaryotes and avian liver, involves 10 enzymatic steps (Fig. 1) to synthesize IMP (the first product with a complete purine ring) from PRPP (Zalkin and Dixon, 1992). AMP and GMP are subsequently formed from IMP. Further research showed that this model holds in all other organisms, but the organization of the enzymes differs among them. In E. coli, nine genes encode the 10 enzyme activities. These genes are termed the pur genes. (The nomenclature for the pur genes has changed over time. In some cases, the genes have been called by the enzyme name and in others they are termed pur genes. The bacterial genes are denoted by letters, but for ease of understanding, we use a numerical nomenclature for plants, e.g. pur1 encodes the enzyme catalyzing the first step of the pathway. To distinguish genes from different plants the initials of species name are used as a prefix to the gene name, e.g. the pur1 gene from cowpea is designated Vupur1.) In E. coli, eight are monofunctional polypeptides, and one bifunctional polypeptide catalyzes the 10 enzymic reactions (Fig. 3). The pathway organization in plants is the same, and in this respect, plants are more similar to prokaryotes than to other eukaryotes (Fig. 3), in which all but three steps are catalyzed by either bifunctional or trifunctional enzymes (Zalkin and Dixon, 1992). The functional domains of these multifunctional proteins are similar to those of their monfunctional counterparts in bacteria and plants, suggesting that during evolution of the eukaryotic animal cell prokaryotic pur genes were condensed.

View larger version (12K): [in this window] [in a new window]   Figure 3.   Schematic diagram showing the structural organization of enzymes of the de novo purine biosynthesis pathway from a range of organisms. In E. coli, each enzyme is monofunctional, except aminoimidazolecarboximide ribonucleotide transformylase/inosine monophosphate cyclohydrolase, which is bifunctional. AIRC has two subunits encoded by the purE and purK genes in E. coli (Zalkin and Dixon, 1992). The organization in plants is similar to that in E. coli except that each enzyme has an N-terminal organelle targeting sequence and the two domains of AIRC are encoded by a single gene (pur6; Senecoff and Meagher, 1993; Chapman et al., 1994; Arabidopsis Genome Inititiave, 2000). In yeast (Saccharomyces cerevisiae), AIRC has a similar structure to that in plants (but no targeting sequence) and GARS (step 2) and AIRS (step 5) form a bifunctional polypeptide encoded by a single gene. All other enzymes are monofunctional. In both Drosophila melanogaster and the vertebrates, a trifunctional polypeptide (with activities for GARS, GART, and AIRS) encoded by a single gene catalyzes steps 2, 3, and 5. In Drosophila melanogaster, this trifunctional polypeptide has a duplication of the AIRS domain. A monofunctional GARS protein that is derived by alternative splicing of the GARS/AIRS/GART gene is also expressed in insects and vertebrates. In these organisms, AIRC and succinoaminoimidazolecarboximide ribonucleotide synthetase (steps 6 and 7) form a bifunctional polypeptide that includes only the purE-like domain of E. coli (Zalkin and Dixon, 1992).

	WE NEED TO KNOW MORE ABOUT REGULATION OF THE PURINE PATHWAY

TOP INTRODUCTION PURINE SYNTHESIS AS A... A SELECT FEW LEGUMES... AND IT IS NODULE... WHY ASSIMILATE N AS... THE DE NOVO PURINE... WE NEED TO KNOW... IS ARABIDOPSIS HELPFUL? THE PURINE PATHWAY HAS... CONCLUSION LITERATURE CITED

Most work on regulation of purine biosynthesis in plants has focused on its role in N assimilation. Because expression of some of the pur genes is significantly increased in nodules after N fixation begins (Smith et al., 1998), a question that has received substantial attention is whether the products of N assimilation are involved in regulating the pathway. The connection between products of N fixation and activity of the purine pathway was first noted in studies where nodulated root systems were transiently exposed to an atmosphere of 80% (v/v) Ar:20% (v/v) O₂ (Atkins et al., 1984). Ar replaces N₂ in air and, although the nodules remain physiologically normal, they do not fix N₂ and have no product of fixation to assimilate. After 3 d of exposure to Ar:O₂, activity of the purine biosynthesis pathway (measured as the ability of a nodule extract to synthesize IMP when provided with pathway substrates) was reduced to a fraction of that in control plants. Analysis of intermediates of the pathway during assay suggested a block after the synthesis of formylglycinamide ribonucleotide (step 3, Fig. 1), but because assays were not available for the component enzymes and none of the pur genes had been cloned at that time, the mechanism of regulation could not be determined. A later study showed activity of AIR synthetase was significantly reduced after 3 h of exposure to Ar:O₂ and was likely to result from a reduction in transcription of the pur5 gene. These results suggest that pur5 may be transcriptionally regulated by products of N₂ fixation (Atkins and Smith, 2000). Other data supporting a role for products of N assimilation in regulation of pur gene expression come from experiments where N solutes were applied to root systems of soybean. Expression of pur1 was increased 7-fold when Gln (Kim et al., 1995) was added to the culture medium, but pur2 and pur3 expression was unaffected (Schnorr et al., 1996). Treatment with NH₄NO₃ did not affect expression of the genes. This result for pur1 in roots conflicts with those from early Ar:O₂ experiments where the pathway appeared to be blocked after step 3 of the pathway (i.e. after the step catalyzed by PRAT, which is encoded by pur1) suggesting that PRAT activity was unaffected by the block in N fixation. What these results do tell us, however, is that there is not coordinated regulation of the pathway genes, although the products of N assimilation are likely to regulate part of the pathway.

There is a noteworthy link between N₂ fixation and the assimilation of fixed N. When purine biosynthesis is blocked by allopurinol (an inhibitor of xanthine dehydrogenase), fixed N is not assimilated via alternative pathways, such as those that form Asn. N₂ fixation is inhibited and the nodules begin to senesce after 24 h (Atkins et al., 1988). Similarly, where ureide synthesis was blocked by antisense expression of uricase (activity reduced by 80%), the transgenic plants showed symptoms of N deficiency (Lee et al., 1993). These results indicate that N₂ fixation is inhibited under conditions where ureide synthesis is limited, but the nature of the connection is not clear.

	IS ARABIDOPSIS HELPFUL?

TOP INTRODUCTION PURINE SYNTHESIS AS A... A SELECT FEW LEGUMES... AND IT IS NODULE... WHY ASSIMILATE N AS... THE DE NOVO PURINE... WE NEED TO KNOW... IS ARABIDOPSIS HELPFUL? THE PURINE PATHWAY HAS... CONCLUSION LITERATURE CITED

It is only recently, with the sequencing of the Arabidopsis genome (Arabidopsis Genome Initiative, 2000), that sequences for all genes encoding the purine biosynthesis enzymes became available and, as a result, that the structure of the plant enzymes could be deduced. Upstream sequences of the genes are also available and this will speed research on regulation of the pathway for its "housekeeping roles." Each enzyme in Arabidopsis, except PRAT ( pur1), is encoded by a single gene. Atpur1,1 (AtATase1) and Atpur1,2 (AtATase2) are differentially expressed (Ito et al., 1994; although both are expressed in flower buds, the tissue from which the cDNAs were isolated) with Atpur1,1 expressed in roots and flowers and Atpur1,2 expressed in flowers, strongly in leaves but not in roots. To date, there is no evidence that Atpur1,3 is expressed, but because cDNAs have only been isolated from flowers, this is not surprising. The presence of more than one copy of pur1 in Arabidopsis suggests that regulation of the pathway in different tissues may be affected through regulation of transcription of pur1, and as a consequence the enzyme it encodes, PRAT.

Array studies with Arabidopsis also provide a valuable source of data that can be analyzed to give insights into regulation of the pathway. In a recent report (Girke et al., 2000), expression of pur5 (encoding AIR synthetase) was approximately 19-fold higher in seeds than in roots or leaves, whereas expression of other pur genes present in the arrays (e.g. pur1 and pur9) was either reduced or the same. It is not yet possible to assess the significance of this observation, but perhaps the pathway is regulated at this step through transcriptional control of pur5. Assays of the pathway activity in seeds would be necessary to validate this idea but it is consistent with the transcriptional regulation of pur5 in cowpea nodules exposed to Ar:O₂ (Atkins and Smith, 2000).

Analysis of the upstream regions of the pur genes in Arabidopsis using the PLACE database (http://www.dna.affrc.go.jp/htdocs/PLACE/; Higo et al., 1999) showed that most contained a motif, which in tomato (Lycopersicon esculentum) controls circadian regulation of the Lhc gene (Piechulla et al., 1998). However, in an array study by Schaffer et al. (2001), five of the pur genes (pur1,1; pur4; pur5; pur6; and pur9) did not show a pattern of expression consistent with circadian regulation.

Work done with Arabidopsis pur7, in which the gene promoter was fused to GUS and expression studied throughout development, showed that expression was highest in actively dividing meristematic tissues, as might be expected (Senecoff et al., 1996). Auxin treatment also increased expression from the pur7 promoter, but whether this was a direct result of auxin on gene expression or an indirect result of the stimulation of cell division/elongation by auxin was not addressed (Senecoff et al., 1996). However, analysis of upstream regions of pur genes using PLACE identified an auxin-responsive element (Baumann et al., 1999) in the promoter of pur7 and also in the promoters of many of the other pur genes (pur1,2; pur1,3; pur2; pur4; pur8; and pur9). A second auxin responsive element (Xu et al., 1997) was identified in the upstream regions of pur1,3 and pur4.

The significance of these elements in relation to plant growth and development is not known and functional analysis of the pur gene promoters will be required to validate any predictions derived from sequence analysis of upstream regions. However, with information provided by the genome sequence, the ease of transformation of Arabidopsis and other tools available for this model species, the analysis should not be too difficult.

	THE PURINE PATHWAY HAS AN UNUSUAL LOCATION IN NODULES

TOP INTRODUCTION PURINE SYNTHESIS AS A... A SELECT FEW LEGUMES... AND IT IS NODULE... WHY ASSIMILATE N AS... THE DE NOVO PURINE... WE NEED TO KNOW... IS ARABIDOPSIS HELPFUL? THE PURINE PATHWAY HAS... CONCLUSION LITERATURE CITED

The localization of the purine biosynthesis pathway in plants is different to that of all other organisms in that it is organelle based (Fig. 2). Synthesis of purines in animals and bacteria appears to be cytosolic (Gooljarsingh et al., 2001). Early studies (for review, see Atkins and Smith, 2000), where organelles from either cowpea or soybean nodules were fractionated, showed that the pathway in plants was localized to plastids. However, a later study (Atkins et al., 1997) using more effective fractionation methods on both Suc and Percoll density gradients, found that both plastids and mitochondria from cowpea nodules were capable of IMP synthesis from R5P or PRPP. In addition, activities of a number of pathway enzymes were detected in both organelles.

Localization of the purine biosynthesis enzymes in two organelles within the same cell raises a number of questions about their intracellular targeting. First, does dual localization occur in all plants and all plant tissues, or is the localization to both organelles a strategy to increase purine biosynthesis in nodules where the flux of purines is much greater than in other tissues? Because of the low rates of purine biosynthesis in tissues other than nodules, this question has proved difficult to answer. Plastid transit peptides and mitochondrial presequences are generally found at the N terminus of a protein (Fig. 3) and are cleaved to produce a mature protein once the final location is reached. The targeting sequences that direct proteins to plastids and mitochondria proteins share some features in that they have a similar amino acid composition, but are generally different in their secondary structure. Table I shows an analysis of the putative targeting sequence of all plant purine enzymes for which genes have been cloned using targeting prediction software, TargetP (Emanuelsson et al., 2000), and Predotar (http://www.inra.fr/Internet/Produits/Predotar/). All but two are predicted to be targeted to plastids (and in both cases TargetP and Predotar differ in their predictions). This might be considered as evidence that dual localization occurs only in nodules. However, none of the targeting sequences for the enzymes from cowpea predict mitochondrial localization and yet the enzymes are present in mitochondria. Given that genes for eight purine enzymes from Phaseoloid legumes have been cloned, it seems unlikely that none of the genes encoding unique mitochondrial isoforms has been identified. This suggests that targeting sequences for purine enzymes in ureide-forming legumes may have unique features, or that import into mitochondria in nodule cells may be different to that in other tissues.

View this table: [in this window] [in a new window]   Table I.   Characteristics of N-terminal extensions of plant purine biosynthetic enzymes The deduced amino acid sequences of de novo purine biosynthetic enzymes from different plant species were aligned with those of E. coli using the "Pileup" program of Genetics Computer Group (Madison, WI). Only data from full-length clones were used. The length of the N-terminal extensions of the plant sequences when compared with the E. coli sequence was calculated (E). Sequences were also analyzed with the TargetP (version 1.01, Emanuelsson et al., 2000) and Predotar (version 0.5, http://www.inra.fr/Internet/Produits/Predotar/) software to determine their predicted subcellular localisations and targeting sequence lengths. Multiple sequences from the same species are distinguished by a numerical suffix. AICART refers to the bifunctional enzyme AICART-IMPCH. P, Precursor; E, extension; L, subcellular localization predicted by TargetP; RC, reliability class (1 most reliable, 5 least reliable); TP, targeting peptide length predicted by TargetP; At, Arabidopsis; Gm, soybean; Nt, Nicotiana tabacum; Vu, cowpea; P, plastid; M, mitochondrion; O, other.

Second, are there two sets of enzymes, one targeted to plastids and the second to mitochondria (as is the case for many proteins localized to different subcellular locations), and if so, are these enzymes products of the same or different genes? There is little evidence for multiple copies of the pur genes (pur2, Schnorr et al., 1996; pur5, Smith et al., 1998). To date, Atpur1 and Gmpur3 are the only cases for which multiple genes have been described. Alternative splicing or translation from different start codons can result in two different proteins being produced as the result of transcription from a single gene (for review, see Danpure, 1995). In addition, a number of proteins have recently been identified that are directed to plastids and mitochondria by the same targeting sequence. These include glutathione reductase from pea (Pisum sativum; Creissen et al., 1995), and ferrochelatase-I (Chow et al., 1997) and methionyl-tRNA synthetase (Menand et al., 1998) from Arabidopsis.

The best studied of the purine enzymes in terms of localization is AIR synthetase from cowpea. All the AIR synthetase cDNAs isolated to date have the same sequence, and Southern analysis indicates there is only a single copy of the gene (Smith et al., 1998). The putative targeting sequence is predicted to target the protein to plastids (Table I) but is not typical of a plastid transit peptide. The plastid and mitochondrial isozymes have been purified from cowpea nodules. Although the proteins from the two organelles differ in the point where they are processed to produce the mature protein, the actual NH₂-terminal amino acid sequences are the same (Goggin, 2001) and could be derived from the cDNA isolated by Smith et al. (1998). However, in vitro import assays with the protein derived from the cDNA indicated import into chloroplasts but not into mitochondria (Goggin, 2001).

In the case of the third enzyme of the pathway, glycinamide ribonucleotide transformylase, two cDNAs encoding the enzyme have been isolated from soybean nodules (Schnorr et al., 1996). These encode proteins that differ only in their N-terminal amino acid sequence. It is tempting to speculate that one would be targeted to plastids and the other to mitochondria. Both are predicted to be localized to plastids by TargetP but the shorter protein (GmGART2, Table I) is predicted as mitochondrial by Predotar. In cowpea, three cDNAs encoding the same protein have been described (GenBank accession no. AAA75367), each has two in-frame ATGs and it is possible that translation is initiated at both. The longer protein, which shares an almost identical targeting sequence with the longer soybean protein, is predicted to form an amphipathic -helix at its N terminus, a structure characteristic of mitochondrial-targeted proteins. However, Target P and Predotar predict plastid localization. The shorter protein is predicted to be targeted to mitochondria by Target P but to be targeted to neither organelle by Predotar. In in vitro import experiments with chloroplasts and mitochondria, neither of these proteins were imported (Goggin, 2001). There is obviously a need for more research to elucidate how the purine enzymes are targeted to mitochondria.

	CONCLUSION

TOP INTRODUCTION PURINE SYNTHESIS AS A... A SELECT FEW LEGUMES... AND IT IS NODULE... WHY ASSIMILATE N AS... THE DE NOVO PURINE... WE NEED TO KNOW... IS ARABIDOPSIS HELPFUL? THE PURINE PATHWAY HAS... CONCLUSION LITERATURE CITED

Despite the central and indispensable roles for purine biosynthesis in all cells, there is a lack of understanding of the regulation of expression of pur genes or for localization of pathway enzymes in plants. Why purine biosynthesis and ureide translocation occurs as the means to assimilate fixed-N in a small group of legumes remains a mystery. It is clear that the host plant does not require this metabolic solution for the assimilation of NH₃ or for translocation of N as ureides, the amides serving the purpose equally well when the source of N is the soil. Perhaps the key lies in an interaction in the specific association that develops between this group of legumes and their symbiotic rhizobia partners. The fact that purine synthesis is also involved in generating ureides for N storage in some species suggests that perhaps a more extensive investigation of pathways for N assimilation in a wider range of plants will reveal this complex mechanism to be more common than is presently appreciated.

	FOOTNOTES

Received October 10, 2001; accepted December 7, 2001.

^* Corresponding author; e-mail pmsmith{at}cyllene.uwa.edu.au; fax 61-08-9380-1001.

www.plantphysiol.org/cgi/doi/10.1104/pp.010912.

	LITERATURE CITED

TOP INTRODUCTION PURINE SYNTHESIS AS A... A SELECT FEW LEGUMES... AND IT IS NODULE... WHY ASSIMILATE N AS... THE DE NOVO PURINE... WE NEED TO KNOW... IS ARABIDOPSIS HELPFUL? THE PURINE PATHWAY HAS... CONCLUSION LITERATURE CITED

• Arabidopsis Genome Inititiave (2000) Analysis of the genome sequence of the flowering plant Arabidopsis thaliana. Nature 408: 796-815[CrossRef][Medline]
• Atkins CA (1991) Ammonia assimilation and export of nitrogen from the legume nodule. In MJ Dilworth, AR Glenn, eds, Biology and Biochemistry of Nitrogen Fixation. Elsevier, Amsterdam, pp 293-319
• Atkins CA, Pate JS, Shelp BJ (1984) Effects of short-term N₂ deficiency on N metabolism in legume nodules. Plant Physiol 76: 705-710[Abstract/Free Full Text]
• Atkins CA, Sanford PJ, Storer PJ, Pate JS (1988) Inhibition of nodule functioning in cowpea by a xanthine oxidoreductase inhibitor, allopurinol. Plant Physiol 88: 1229-1234[Abstract/Free Full Text]
• Atkins CA, Smith PMC (2000) Ureide synthesis in legume nodules. In EJ Triplett, ed, Prokaryotic Nitrogen Fixation: A Model System for the Analysis of a Biological Process. Horizon Scientific Press, Wymondham, Norfolk, UK, pp 559-587
• Atkins CA, Smith PMC, Storer PJ (1997) Reexamination of the intracellular localization of de novo purine synthesis in cowpea nodules. Plant Physiol 113: 127-135[Abstract]
• Atkins CA, Storer PJ, Young EB (1991) Translocation of nitrogen and expression of nodule-specific uricase (nodulin-35) in Robinia pseudoacacia. Physiol Plant 83: 483-491[CrossRef]
• Baumann K, De Paolis A, Costantino P, Gualberti G (1999) The DNA binding site of the Dof protein NtBBF1 is essential for tissue-specific and auxin-regulated expression of the rolB oncogene in plants. Plant Cell 11: 323-333[Abstract/Free Full Text]
• Brown SM, Walsh KB (1994) Anatomy of the legume nodule cortex with respect to nodule permeability. Aust J Plant Physiol 21: 49-68
• Chapman KA, Delauney AJ, Kim JH, Verma DPS (1994) Structural organization of de novo purine biosynthesis enzymes in plants: 5-aminoimidazole ribonucleotide carboxylase and 5-aminoimidazole-4-N-succinocarboxamide ribonucleotide synthetase cDNAs from Vigna aconitifolia. Plant Mol Biol 24: 389-395[Medline]
• Cheng XG, Nomura M, Takane K, Kouchi H, Tajima S (2000) Expression of two uricase (Nodulin-35) genes in a non-ureide type legume, Medicago sativa. Plant Cell Physiol 41: 104-109
• Chow KS, Singh DP, Roper JM, Smith AG (1997) A single precursor protein for ferrochelatase-I from Arabidopsis is imported in vitro to both chloroplasts and mitochondria. J Biol Chem 272: 27565-27571[Abstract/Free Full Text]
• Corby HDL (1988) Types of rhizobial nodule and their distribution among the Leguminoseae. Kirkia 13: 53-123
• Coruzzi GM, Zhou L (2001) Carbon and nitrogen sensing and signaling in plants: emerging matrix effects. Curr Opinion Plant Biol 4: 247-253[CrossRef][ISI][Medline]
• Creissen G, Reynolds H, Xue Y, Mullineaux P (1995) Simultaneous targeting of pea glutathione reductase and of a bacterial fusion protein to chloroplasts and mitochondria in transgenic tobacco. Plant J 8: 167-175[CrossRef][ISI][Medline]
• Danpure CJ (1995) How can the products of a single gene be localized to more than one intracellular compartment? Biochem Biophys Res Comm 203: 866-873
• Doyle JJ, Chappill JA, Bailey CD, Kajita T (2000) Towards a comprehensive phylogeny of legumes: evidence from RBCL sequences and non-molecular data. Adv Legume Syst 9: 1-20
• Emanuelsson O, Nielsen H, Brunak S, von Heijne G (2000) Predicting subcellular localization of proteins based on their N-terminal amino acid sequence. J Mol Biol 300: 1005-1016[CrossRef][ISI][Medline]
• Fujihara S, Yamaguchi M (1978) Effects of allopurinol [4-hydroxyprazolo(3, 4-d)pyrimidine] on the metabolism of allantoin in soybean plants. Plant Physiol 62: 134-138[Abstract/Free Full Text]
• Girke T, Todd J, Ruuska S, White J, Benning C, Ohlrogge J (2000) Microarray analysis of developing Arabidopsis seeds. Plant Physiol 124: 1570-1581[Abstract/Free Full Text]
• Goggin DE (2001) Studies on the Localization and Physical Characteristics of Some Enzymes of de Novo Purine Biosynthesis in the Root Nodules of the Tropical Legume Vigna unguiculata L. (Walp). PhD thesis. The University of Western Australia
• Gooljarsingh LT, Ramcharan J, Gilroy S, Benkovic SJ (2001) Localization of GAR transformylase in Escherichia coli and mammalian cells. Proc Natl Acad Sci USA 98: 6565-6570[Abstract/Free Full Text]
• Higo K, Ugawa Y, Iwamoto M, Korenaga T (1999) Plant cis-acting regulatory DNA elements (PLACE) database. Nucleic Acids Res 27: 297-300[Abstract/Free Full Text]
• Ito T, Shiraishi H, Okada K, Shimura Y (1994) Two amidophosphoribosyltransferase genes of Arabidopsis thaliana expressed in different organs. Plant Mol Biol 26: 529-533[CrossRef][ISI][Medline]
• Kim JH, Delauney AJ, Verma DPS (1995) Control of de novo purine biosynthesis genes in ureide-producing legumes: induction of glutamine phosphoribosylpyrophosphate amidotransferase gene and characterization of its cDNA from soybean and Vigna. Plant J 7: 77-86[CrossRef][Medline]
• Kuzma MM, Hunt S, Layzell DB (1993) Role of oxygen in the limitation and inhibition of nitrogenase activity and respiration rate in individual soybean nodules. Plant Physiol 101: 161-169[Abstract]
• Layzell DB, Rainbird RM, Atkins CA, Pate JS (1979) Economy of photosynthate use in nitrogen-fixing legume nodules. Plant Physiol 64: 888-891[Abstract/Free Full Text]
• Lee N-G, Stein B, Suzuki H, Verma DPS (1993) Expression of antisense nodulin-35 RNA in Vigna aconitifolia transgenic root nodules retards peroxisome development and affects nitrogen availability to the plant. Plant J 3: 599-606[CrossRef][ISI][Medline]
• Menand B, Maréchal-Drouard L, Sakamoto W, Dietrich A, Winz H (1998) A single gene of chloroplast origin codes for mitochondrial and chloroplastic methionyl-tRNA synthetase in Arabidopsis thaliana. Proc Natl Acad Sci USA 95: 11014-11019[Abstract/Free Full Text]
• Peoples MB, Pate JS, Atkins CA (1985) The effect of nitrogen source on transport and metabolism of nitrogen in fruiting plants of cowpea (Vigna unguiculata (L.) Walp.). J Exp Bot 36: 567-582[Abstract/Free Full Text]
• Piechulla B, Merforth N, Rudolph B (1998) Identification of tomato Lhc promoter regions necessary for circadian expression. Plant Mol Biol 38: 655-662[CrossRef][ISI][Medline]
• Rainbird RM, Atkins CA (1981) Purification and some properties of urate oxidase from nitrogen-fixing nodules of cowpea. Biochim Biophys Acta 659: 132-140[Medline]
• Reinbothe H, Mothes K (1962) Urea, ureides, and guanidines in plants. Ann Rev Plant Physiol 13: 129-151
• Schaffer R, Landgraf J, Accerbi M, Simon VV, Larson M, Wisman (2001) Microarray analysis of diurnal and circadian-regulated genes in Arabidopsis. Plant Cell 13: 113-123[Abstract/Free Full Text]
• Schnorr KM, Laloue M, Hirel B (1996) Isolation of cDNAs encoding two purine biosynthetic enzymes of soybeanand expression of the corresponding transcipts in the roots and root nodules. Plant Mol Biol 32: 751-757[Medline]
• Schnorr KM, Nygaard P, Laloue M (1994) Molecular characterization of Arabidopsis thaliana cDNAs encoding three purine biosynthetic enzymes. Plant J 6: 113-121[CrossRef][Medline]
• Senecoff JF, McKinney EC, Meagher RB (1996) De Novo Purine Synthesis in Arabidopsis thaliana: II. The PUR7 gene encoding 5' phosphoribosyl-4-(N-succinocarboxamide)-5-aminoimidazole synthetase is expressed in rapidly dividing tissues. Plant Physiol 112: 905-917[Abstract]
• Senecoff J, Meagher RB (1993) Isolating the Arabidopsis thaliana genes for de novo purine synthesis by suppression of Escherichia coli mutants: I. 5'-Phosphoribosyl- 5-aminoimidazole synthetase. Plant Physiol 102: 387-399[Abstract]
• Smith PMC, Mann AJ, Goggin DE, Atkins CA (1998) AIR synthetase in Cowpea Nodules: a single gene product targeted to two organelles? Plant Mol Biol 36: 811-820[Medline]
• Tajima S, Takane K, Nomura M, Kouchi H (2000) Symbiotic nitrogen fixation at the late stage of nodule formation in Lotus japonicus and other legume plants. J Plant Res 113: 467-473[CrossRef]
• Thumfort PP, Layzell DB, Atkins CA (1999) Diffusion and reaction of oxygen in the central tissue of ureide-producing legume nodules. Plant Cell Environ 22: 1351-1363[CrossRef]
• Xu N, Hagen G, Guilfoyle T (1997) Multiple auxin response modules in the soybean SAUR 15A promoter. Plant Sci 126: 193-201[CrossRef]
• Zalkin H, Dixon JE (1992) De novo purine nucleotide biosynthesis. Prog Nucleic Acids Res 42: 259-287[ISI][Medline]