Glutamine synthetase in the phloem plays a major role in controlling proline production

To inhibit expression specifically in the phloem, a 274-bp fragment of a cDNA (Gln1-5) encoding cytosolic glutamine synthetase (GS1) from tobacco was placed in the antisense orientation downstream of the cytosolic Cu/Zn superoxide dismutase promoter of Nicotiana plumbaginifolia. After Agrobacterium-mediated transformation, two transgenic N. tabacum lines exhibiting reduced levels of GS1 mRNA and GS activity in midribs, stems, and roots were obtained. Immunogold labeling experiments allowed us to verify that the GS protein content was markedly decreased in the phloem companion cells of transformed plants. Moreover, a general decrease in proline content in the transgenic plants in comparison with wild-type tobacco was observed when plants were forced to assimilate large amounts of ammonium. In contrast, no major changes in the concentration of amino acids used for nitrogen transport were apparent. A (15)NH(4)(+)-labeling kinetic over a 48-hr period confirmed that in leaves of transgenic plants, the decrease in proline production was directly related to glutamine availability. After 2 weeks of salt treatment, the transgenic plants had a pronounced stress phenotype, consisting of wilting and bleaching in the older leaves. We conclude that GS in the phloem plays a major role in regulating proline production consistent with the function of proline as a nitrogen source and as a key metabolite synthesized in response to water stress.

Simultaneous expression of NAD-dependent isocitrate dehydrogenase and other krebs cycle genes after nitrate resupply to short-term nitrogen-starved tobacco

Mitochondrial NAD-dependent (IDH) and cytosolic NADP-dependent isocitrate dehydrogenases have been considered as candidates for the production of 2-oxoglutarate required by the glutamine synthetase/glutamate synthase cycle. The increase in IDH transcripts in leaf and root tissues, induced by nitrate or NH4+ resupply to short-term N-starved tobacco (Nicotiana tabacum) plants, suggested that this enzyme could play such a role. The leaf and root steady-state mRNA levels of citrate synthase, acotinase, IDH, and glutamine synthetase were found to respond similarly to nitrate, whereas those for cytosolic NADP-dependent isocitrate dehydrogenase and fumarase responded differently. This apparent coordination occurred only at the mRNA level, since activity and protein levels of certain corresponding enzymes were not altered. Roots and leaves were not affected to the same extent either by N starvation or nitrate addition, the roots showing smaller changes in N metabolite levels. After nitrate resupply, these organs showed different response kinetics with respect to mRNA and N metabolite levels, suggesting that under such conditions nitrate assimilation was preferentially carried out in the roots. The differential effects appeared to reflect the C/N status after N starvation, the response kinetics being associated with the nitrate assimilatory capacity of each organ, signaled either by nitrate status or by metabolite(s) associated with its metabolism.

Does root glutamine synthetase control plant biomass production in lotus japonicus L.?

To investigate the contribution of root cytosolic glutamine synthetase (GS) activity in plant biomass production, two different approaches were conducted using the model legume Lotus japonicus. In the first series of experiments, it was found that overexpressing GS activity in roots of transgenic plants leads to a decrease in plant biomass production. Using (15)N labelling it was shown that this decrease is likely to be due to a lower nitrate uptake accompanied by a redistribution to the shoots of the newly absorbed nitrogen which cannot be reduced due to the lack of nitrate reductase activity in this organ. In the second series of experiments, the relationship between plant growth and root GS activity was analysed using a series of recombinant inbred lines issued from the crossing of two different Lotus ecotypes, Gifu and Funakura. It was confirmed that a negative relationship exists between root GS expression and plant biomass production in both the two parental lines and their progeny. Statistical analysis allowed it to be estimated that at least 13% of plant growth variation can be accounted for by variation in GS activity.

The HLA-G gene is expressed at a low mRNA level in different human cells and tissues.

Recently, HLA-G transgenic mice were shown to exhibit transgene transcription in several extraembryonic tissues. To determine whether HLA-G mRNAs are also expressed in other human tissues, we have undertaken Northern blot and RT-PCR assays using HLA-G locus-specific probe and primers. These studies demonstrate that the HLA-G gene is transcribed in a variety of cells and adult tissues obtained from different individuals (peripheral blood leukocytes, placenta, skin, spleen, thymus, prostate, testicle, ovary, small intestine, colon, heart, brain, lung, liver, and kidney), as well as in fetal tissues (heart, lung, liver, and kidney). The HLA-G mRNA level observed in most tissues is orders of magnitude lower than the level of classic class I genes in the same tissues. RT-PCR studies have demonstrated that alternative splicing of the HLA-G primary transcript is different from tissue to tissue and could be regulated in a tissue-specific fashion. Sequencing of keratinocyte transcripts has confirmed previous observations: (a) three different alternative splicing transcripts are produced (a full-length transcript, an mRNA lacking exon 3, and a transcript devoid of exon 3 and 4) and (b) HLA-G polymorphism is limited in the coding regions. In view of this wide HLA-G tissue distribution, a new hypothesis dealing with possible HLA-G function is proposed.

In Vitro expression of drug metabolizing enzyme activities in human adult keratinocytes under various culture conditions and their response to inducers.

In this study we analysed the expression and induction of several drug metabolizing enzymes involved in either phase I or phase II reactions, in adult human keratinocytes cultured in submerged conditions. We also evaluated the influence of confluence, subcultivation and cryopreservation on the expression of these enzymes. Besides ethoxyresorufin-O-deethylase (EROD) and glutathione S-transferase (GST) activities, which have been shown previously to be maintained in such cultures, three additional enzyme activities were measured (i.e. phenacetin deethylase, a phase I enzyme, and procainamide N-acetyltransferase and paracetamol sulfotransferase, two phase II enzymes). Post-confluent keratinocytes showed decreased activities in comparison with preconfluent cells and the different enzymes tested revealed different patterns. After confluence, some activities, such as those of procainamide N-acetyltrans-ferase, phenacetin deethylase and paracetamol sulfotransferase, showed only a slight decrease, whereas EROD and GST activities were decreased by 65 and 50%, respectively. No major differences were observed between keratinocytes in primary culture and those in second subculture. After freezing, xenobiotic metabolizing enzyme activities were only slightly reduced, if at all. Induction of EROD and GST enzymes was also analysed. Maximum EROD activity was obtained with 1 muM 3-methylcholanthrene (3-MC) and 20 muM benzanthracene (BA), in both pre-confluent and post-confluent cultures. At their optimal concentration 3-MC was a stronger inducer than BA. GST activity was slightly induced by the different compounds tested only in pre-confluent keratinocytes. In conclusion, the presence of a variety of drug metabolizing enzymes in adult human keratinocytes cultured in submerged conditions suggests that this model is suitable for investigating epidermal biotransformation of drugs and other chemicals and for determining the potential cutaneous toxicity of metabolites.

Cytotoxicity testing of beta-lactam antibiotics, non-steroidal anti-inflammatory drugs and sulfonamides in primary human keratinocyte cultures.

We have studied the cytotoxity to human keratinocytes of three main classes of drugs known to induce cutaneous adverse reactions, namely beta-lactam antibiotics, non-steroidal anti-inflammatory drugs (NSAIDs) and sulfonamides, using the neutral red uptake as an endpoint. IC(50) values were determined for 21 drugs, after 20 hr of exposure and compared with those obtained with rat hepatocytes. NSAIDs were found to be more cytotoxic than beta-lactum antibiotics to human keratinocytes. Large variations in IC(50) values were obtained between molecules of a same class, as well as between keratinocyte cultures from different donors, especially for beta-lactum antibiotics. All NSAIDs and beta-lactam antibiotics tested were more cytotoxic to rat hepatocytes (1.6- to 27-fold). Both cell types were only slightly sensitive to sulfonamides, if at all.

Expression of differentiation markers in human adult keratinocytes cultured in submerged conditions.

A number of studies have shown that human keratinocytes cultured in submerged conditions with non-delipidized serum do not express the major differentiation markers, i.e. 67 kDa keratin, ceramides, and lanosterol. However, they were mostly performed with neonatal or juvenile keratinocytes after a few passages, and not all the markers were analyzed in parallel. In this study, we compared the expression of several differentiation markers in preconfluent and postconfluent adult breast keratinocytes in primary and secondary cultures before and after cryopreservation. When primary cultures reached confluence, the 67 kDa keratin was synthesized, transglutaminase activity was increased, and, although overall lipid synthesis dropped, both lanosterol and free fatty acids contents were augmented. The same pattern was observed in postconfluent subcultures at Passage 2; however decreased overall lipid synthesis was more pronounced. Cryopreservation of keratinocytes just after isolation or after a few days in culture did not result in the loss of expression of these specific epidermic markers. Thus, adult breast keratinocytes in postconfluent submerged cultures represent an in vitro model that possesses various features of the normal epidermis, even after cryopreservation.

Genetic variation for N-remobilization and postsilking N-uptake in a set of maize recombinant inbred lines. 3. QTL detection and coincidences.

The objective of this study was to map and characterize QTLs for traits related to nitrogen utilization efficiency (NUE), grain N yield, N-remobilization and post-silking N-uptake. Furthermore, to examine whether QTLs detected with recombinant inbred lines (RILs) crossed to a tester are common to those detected with line per se evaluation, both types of evaluations were developed from the same set of RILs. The material was studied over two years at high N-input, and one year at low N-input. We used (15)N-labelling to evaluate with accuracy the proportion of N remobilized from stover to kernels and the proportion of postsilking N-uptake allocated to kernels. With 59 traits studied in three environments, 608 QTLs were detected. Using a method of QTL clustering, 72 clusters were identified, with few QTLs being specific to one environment or to the type of plant material (lines or testcross families). However, considering each trait separately, few QTLs were common to both line per se and testcross evaluation. This shows that genetic variability is expressed differently according to the type of progeny. Studies of coincidences among QTLs within the clusters showed an antagonism between N-remobilization and N-uptake in several QTL-clusters. QTLs for N-uptake, root system architecture and leaf greenness coincided positively in eight clusters. QTLs for remobilization mainly coincided in clusters with QTLs for leaf senescence. On the whole, sign of coincidences between QTLs underlined the role of a "stay-green" phenotype in favouring N-uptake capacity, and thus grain yield and N grain yield.

Partial sequences of nitrogen metabolism genes in hexaploid wheat.

Our objective was to partially sequence genes controlling nitrogen metabolism in wheat species in order to find sequence polymorphism that would enable their mapping. Primers were designed for nitrate reductase, nitrite reductase, glutamate dehydrogenase and glutamate synthase (GOGAT), and gene fragments were amplified on Triticum aestivum, T. durum, T. monococcum, T. speltoides and T. tauschii. We obtained more than 8 kb of gene sequences, mainly as coding regions (60%). Polymorphism was quantified by comparing two-by-two the three genomes of the hexaploid cultivar Arche and genomes of diploid wheat species. On average, the polymorphism rate was higher for non-coding regions, where it ranged from 1/60 to 1/23, than for coding regions (range: 1/110-1/40) except when the hexaploid D genome was compared to that of T. tauschii (1/800 and 1/816, respectively). Genome-specific primers were devised for the ferredoxin-dependent (Fd)-GOGAT gene, and they enabled the mapping of this gene on homoeologous chromosomes of group 2 using Chinese Spring deletion lines. A single nucleotide polymorphism (SNP) detected between the two hexaploid wheat cultivars Arche and Recital was used to genetically map Fd-GOGAT on chromosome 2D using a population of dihaploid lines. Fd-GOGAT-specific primers were used to estimate the SNP rate on a set of 11 hexaploid and nine Durum wheat genotypes leading to the estimate of 1 SNP/515 bp. We demonstrate that polymorphism detection enables heterologous, homeologous and even paralogous copies to be assigned, even if the elaboration of specific primer pairs is time-consuming and expensive because of the sequencing.

An approach to the genetics of nitrogen use efficiency in maize.

To study the genetic variability and the genetic basis of nitrogen (N) use efficiency in maize, a set of recombinant inbred lines crossed with a tester was studied at low input (N-) and high input (N+) for grain yield and its components, grain protein content, and post-anthesis nitrogen uptake and remobilization. Other physiological traits, such as nitrate content, nitrate reductase, glutamine synthetase (GS), and glutamate dehydrogenase activities were studied at the level of the lines. Genotypexnitrogen (GxN) interaction was significant for yield and explained by variation in kernel number. In N-, N-uptake, the nitrogen nutrition index, and GS activity in the vegetative stage were positively correlated with grain yield, whereas leaf senescence was negatively correlated. Whatever N-input, post-anthesis N-uptake was highly negatively related to N-remobilization. As a whole, genetic variability was expressed differently in N+ and N-. This was confirmed by the detection of QTLs. More QTLs were detected in N+ than in N- for traits of vegetative development, N-uptake, and grain yield and its components, whereas it was the reverse for grain protein content and N-utilization efficiency. Several coincidences between genes encoding for enzymes of N metabolism and QTLs for the traits studied were observed. In particular, coincidences in three chromosome regions of QTLs for yield and N-remobilization, QTLs for GS activity and a gene encoding cytosolic GS were observed. This may have a physiological meaning. The GS locus on chromosome 5 appears to be a good candidate gene which can, at least partially, explain the variation in nitrogen use efficiency.

Glutamine Synthetase in Rice: A COMPARATIVE STUDY OF THE ENZYMES FROM ROOTS AND LEAVES.

Chromatographic, kinetic, and regulatory properties of glutamine synthetase in rice were investigated. By DEAE-Sephacel column chromatography, two forms (glutamine synthetase 1 and glutamine synthetase 2) were identified in leaves and one form (glutamine synthetase R) was identified in roots. Purification on hydroxyapatite and gel electrophoresis showed that glutamine synthetase R was distinct from the leaf enzymes. The three isoforms were purified to similar specific activities and their properties were studied. Heat lability, pH optimum about 8, K(m) for l-glutamate of 20 millimolar, and inhibition by glucosamine 6-phosphate were the main characteristics of glutamine synthetase 2. Heat stability, pH optimum about 7.5, K(m) for l-glutamate of 2 millimolar, and no effect of glucosamine 6-phosphate differentiated glutamine synthetase 1 from glutamine synthetase 2. Glutamine synthetase R was also a labile protein but its kinetic and regulatory properties were quite similar to those of glutamine synthetase 1. These results clearly demonstrate the existence of three isoforms of glutamine synthetase in rice, two of which are located in the leaves and the third in the roots.

New artificial electron donors for in vitro assay of nitrate reductase isolated from cultured tobacco cells and other organisms.

The capacity of bromphenol blue and its analogs to act as electron donors for measurement of in vitro nitrate reductase activity from tobacco cells (Nicotiana tabacum var Techné SP 25 strain) was determined. Competitive inhibition was demonstrated to occur between NADH, the natural electron donor, and bromphenol blue, the artificial electron donor, suggesting that both donors bind to a similar active site on the enzyme. NADH-dependent or bromphenol blue-dependent nitrate reductase activity was carried out by a similar molecular weight protein exhibiting similar antigenic sites. Following ammonium sulfate precipitation, sucrose density gradient and two chromatographic steps, nitrate reductase activity from tobacco cells was purified near homogeneity using bromphenol blue as an electron donor in the absence of measurable NADH-dependent activity. The enzyme is composed of two identical subunits of 83 kilodaltons < Momega < 94 kilodaltons.

Immuno chemical characterization of nitrate reductase from spinach leaves and roots.

Nitrate reductase from spinach (Spinacia oleracea L.) leaves was purified to homogeneity and specific antibodies against the protein were raised by the immunization of rabbits. Using immunodiffusion and immunoprecipitation techniques, the enzymes from leaves and roots were compared. It was demonstrated that the two nitrate reductases were different proteins on the basis of their antigenic behaviour.

Evidence for a cytosolic-dependent light induction of chloroplastic glutamine synthetase during greening of etiolated rice leaves.

During the greening of etiolated rice leaves, total glutamine synthetase activity increases about twofold, and after 48 h the level of activity usually observed in green leaves is obtained. A density-labeling experiment with deuterium demonstrates that the increase in enzyme activity is due to a synthesis of the enzyme. The enhanced activity obtained upon greening is the result of two different phenomena: there is a fivefold increase of chloroplastic glutamine synthetase content accompanied by a concommitant decrease (twofold) of the cytosolic glutamine synthetase. The increase of chloroplastic glutamine synthetase (GS2) is only inhibited by cycloheximide and not by lincomycin. This result indicates a cytosolic synthesis of GS2. The synthesis of GS2 was confirmed by a quantification of the protein by an immunochemical method. It was demonstrated that GS2 protein content in green leaves is fivefold higher than in etiolated leaves.

Glutamine synthetase in ribulose 1,5-bisphosphate carboxylase/oxygenase deficient tobacco mutants in cell suspension culture.

In two tobacco mutants lacking ribulose, 1,5-bisphosphate carboxylase/oxygenase the amount of glutamine synthetase and its activity were determined and compared with the wild type green cells. It was shown that in these two mutants glutamine synthetase protein content was six times lower than in the wild type. This situation was comparable to that found in etiolated cells where ribulose 1,5-bisphosphate carboxylase/oxygenase was absent. These observations suggest that a common regulatory mechanism might control the dual light dependent biosynthesis of both enzymes. The results have also implications concerning the efficiency of the reassimilation of ammonia by chloroplastic glutamine synthetase during the photorespiratory process.

Glutamine synthetase and glutamate dehydrogenase isoforms in maize leaves: localization, relative proportion and their role in ammonium assimilation or nitrogen transport.

Mesophyll cells (MCs) and bundle-sheath cells (BSCs) of leaves of the C4 plant maize (Zea mays L.) were separated by cellulase digestion to determine the relative proportion of the glutamine synthetase (GS; EC 6.3.1.2) or the NADH-glutamate dehydrogenase (GDH; EC 1.4.1.2) isoforms in each cell type. The degree of cross-contamination between our MC and BSC preparations was checked by the analysis of marker proteins in each fraction. Nitrate reductase (EC 1.6.6.1) proteins (110 kDa) were found only in the MC fraction. In contrast, ferredoxin-dependent glutamate synthase (Fd-GOGAT; EC 1.4.7.1) proteins (160 kDa) were almost exclusively present in the BSC fraction. These results are consistent with the known intercellular distribution of nitrate reductase and Fd-GOGAT proteins in maize leaves and show that the cross-contamination between our MC and BSC fractions was very low. Proteins corresponding to cytosolic GS (GS-1) or plastidic GS (GS-2) were found in both the MC and BSC fractions. While equal levels of GS-1 (40 kDa) and GS-2 (44 kDa) polypeptides were present in the BSC fraction, the GS-1 protein level in the MC fraction was 1.8-fold higher than the GS-2 protein pool. Following separation of the GS isoforms by anion-exchange chromatography of MC or BSC soluble protein extracts, the relative GS-1 activity in the MC fraction was found to be higher than the relative GS-2 activity. In the BSC fraction, the relative GS-1 activity was very similar to the relative GS-2 activity. Two isoforms of GDH with apparent molecular weights of 41 kDa and 42 kDa, respectively, were detected in the BSC fraction of maize leaves. Both GDH isoenzymes appear to be absent from the MC fraction. In the BSCs, the level of the 42-kDa GDH isoform was 1.7-fold higher than the level of the 41-kDa GDH isoform. A possible role for GS-1 and GDH co-acting in the synthesis of glutamine for the transport of nitrogen is discussed.

Identification of several soybean cytosolic glutamine synthetase transcripts highly or specifically expressed in nodules: expression studies using one of the corresponding genes in transgenic Lotus corniculatus.

A DNA fragment containing sequences hybridizing to the 5' region of GS15, a gene encoding soybean cytosolic glutamine synthetase, was isolated from a soybean genomic library. Mapping and partial sequence analysis of the genomic clone revealed that it encodes a cytosolic GS gene, GS21, which is different from GS15. In parallel, a number of cDNA clones encoding cytosolic GS were isolated using the coding region of pGS20 as a probe (pGS20 is a cDNA clone which corresponds to a transcript of the GS15 gene). Two new full-length cDNAs designated pGS34 and pGS38 were isolated and sequenced. In the 5' non-coding region a strong homology was found between the two clones and the GS21 gene. However, none of these sequences were identical, which suggests that there are at least three members in this group of genes. In order to determine their relative levels of transcription, specific sequences from pGS34, pGS38 and GS21 were used in an RNAse protection assay. This experiment clearly showed that GS21 and the gene encoding pGS38 are specifically expressed in young or mature nodules, whereas the gene encoding pGS34 is highly transcribed in nodules and constitutively expressed at a lower level in other soybean organs. In order to further analyse the molecular mechanisms controlling GS21 transcription, different fragments of the promoter region were fused to the Escherichia coli reporter gene encoding beta-glucuronidase (GUS) and the constructs were introduced into Lotus corniculatus via Agrobacterium rhizogenes-mediated transformation. Analysis of GUS activity showed that the GS21 promoter-GUS constructs were expressed in the vasculature of all vegetative organs. This result is discussed in relation to species-specific metabolic and developmental characteristics of soybean and Lotus.

Nucleotide sequence of a tobacco cDNA encoding plastidic glutamine synthetase and light inducibility, organ specificity and diurnal rhythmicity in the expression of the corresponding genes of tobacco and tomato.

A full-length cDNA encoding glutamine synthetase (GS) was cloned from a lambda gt10 library of tobacco leaf RNA, and the nucleotide sequence was determined. An open reading frame accounting for a primary translation product consisting of 432 amino acids has been localized on the cDNA. The calculated molecular mass of the encoded protein is 47.2 kDa. The predicted amino acid sequence of this precursor shows higher homology to GS-2 protein sequences from other species than to a leaf GS-1 polypeptide sequence, indicating that the cDNA isolated encodes the chloroplastic isoform (GS-2) of tobacco GS. The presence of C- and N-terminal extensions which are characteristic of GS-2 proteins supports this conclusion. Genomic Southern blot analysis indicated that GS-2 is encoded by a single gene in the diploid genomes of both tomato and Nicotiana sylvestris, while two GS-2 genes are very likely present in the amphidiploid tobacco genome. Western blot analysis indicated that in etiolated and in green tomato cotyledons GS-2 subunits are represented by polypeptides of similar size, while in green tomato leaves an additional GS-2 polypeptide of higher apparent molecular weight is detectable. In contrast, tobacco GS-2 is composed of subunits of identical size in all organs examined. GS-2 transcripts and GS-2 proteins could be detected at high levels in the leaves of both tobacco or tomato. Lower amounts of GS-2 mRNA were detected in stems, corolla, and roots of tomato, but not in non-green organs of tobacco. The GS-2 transcript abundance exhibited a diurnal fluctuation in tomato leaves but not in tobacco leaves. White or red light stimulated the accumulation of GS-2 transcripts and GS-2 protein in etiolated tomato cotyledons. Far-red light cancelled this stimulation. The red light response of the GS-2 gene was reduced in etiolated seedlings of the phytochrome-deficient aurea mutant of tomato. These results indicate a phytochrome-mediated light stimulation of GS-2 gene expression during greening in tomato.

Localization of tobacco cytosolic glutamine synthetase enzymes and the corresponding transcripts shows organ- and cell-specific patterns of protein synthesis and gene expression.

The subcellular localization of glutamine synthetase in tobacco and the differential expression of two genes encoding cytosolic enzyme was investigated using both immunocytochemistry and in situ hybridization. Two full length cDNA clones each encoding cytosolic GS (Gln 1-3 and Gln 1-5) were isolated from a tobacco seeding cDNA library. A strong homology was found in the coding region of the two clones whereas the 3'- and 5'-untranslated sequences were dissimilar. In order to determine the levels of transcription, specific sequences from Gln1-3 and Gln1-5 were used in an RNAse protection assay. This experiment clearly showed that the gene encoding Gln1-3 is expressed in roots and flowers whereas the gene encoding Gln1-5 is transcribed at a high level in stems and at a lower level in roots and flowers. Immunogold labelling was used to examine the subcellular and cellular distribution of glutamine synthetase in vegetative and reproductive organs of tobacco plants. In mature leaf tissue or petals and sepals, plastidic GS was visualised only in the stroma matrix of chloroplasts and plastids. Cytosolic GS was detected in a number of vegetative or reproductive organs including leaves and flowers. In leaves cytosolic GS was preferentially located in the vascular tissue. In situ hybridization was performed using sections of tobacco organs and specific antisense RNA probes to the genes encoding Gln1-3 and Gln1-5. Gln1-5 transcripts were localised in the vascular tissues of stems and roots whereas Gln1-3 transcripts were detected in all root cells and floral organs including petals, sepals and anthers.

Isolation of cDNAs encoding two purine biosynthetic enzymes of soybean and expression of the corresponding transcripts in roots and root nodules.

Soybean nodule cDNA clones encoding glycinamide ribonucleotide (GAR) synthetase (GMpurD) and GAR transformylase (GMpurN) were isolated by complementation of corresponding Escherichia coli mutants. GAR synthetase and GAR transformylase catalyse the second and the third steps in the de novo purine biosynthesis pathway, respectively. One class of GAR synthetase and three classes of GAR transformylase cDNA clones were identified. Northern blot analysis clearly shows that these purine biosynthetic genes are highly expressed in young and mature nodules but weakly expressed in roots and leaves. Expression levels of GMpurD and GMpurN mRNAs were not enhanced when ammonia was provided to non-nodulated roots.