Four alanine aminotransferases (AlaATs) are expressed in Medicago truncatula. In adult plants, two genes encoding mitochondrial isoforms m-AlaAT and alanine–glyoxylate aminotransferase (AGT), catalysing, respectively, reversible reactions of alanine/oxoglutarate↔glutamate/pyruvate and alanine/glyoxylate↔glycine/pyruvate, were expressed in roots, stems, and leaves. A gene encoding a cytosolic (c-AlaAT) isoform, catalysing the same reaction as m-AlaAT, was expressed specifically in leaves, while a gene encoding an isoform involved in branched chain amino acid metabolism was expressed in stems and roots. In young seedlings, only m-AlaAT and AGT were expressed in embryo axes. In hypoxic embryo axes, the amounts of transcript and putative protein of m-AlaAT (EC 188.8.131.52) increased while those of AGT (EC 184.108.40.206) decreased and in vivo enzyme activities changed as revealed by [15N]alanine and [15N]glutamate labelling. Under hypoxia, m-AlaAT catalysed only alanine synthesis while glutamate synthesis using alanine as amino donor was inhibited. As a result, alanine accumulated as the major amino acid in hypoxic seedlings instead of asparagine, in agreement with the involvement of the fermentative AlaAT pathway in hypoxia tolerance. Regulation of m-AlaAT at both the transcriptional and post-translational levels allowed for an increase in gene expression and orientation of the activity of the product of its transcription towards alanine synthesis under hypoxia. Labelling experiments showed that glycine synthesis occurred at the expense of either alanine or glutamate as amino donor, indicating that a glutamate–glyoxylate aminotransferase was operating together with AGT in Medicago truncatula seedlings. Both enzymes seemed to be inhibited by hypoxia, resulting in a very low amount of glycine in hypoxic seedlings.
open reading frame
rapid amplification of cDNA ends
Germination and post-germination are crucial phases of the plant life cycle. During these early phases of development, the capacity of seedlings to colonize the soil uniformly and rapidly is critical to vigour and performance of the future plant. Flooding is one of the adverse environmental factors that severely harm germination, seedling establishment, and plant development (Subbaiah and Sachs, 2003). Gases diffuse approximately 10 000 times more slowly in water than in air (Jackson, 1985); as a consequence, when soils are saturated with water, the flux of oxygen into plants becomes too slow to support respiration, resulting in energy deficits and, eventually, death of cells and tissues in non-adapted plants (Jackson and Armstrong, 1999; Gout et al., 2001).
Plants have evolved inducible developmental and metabolic mechanisms to adapt to low oxygen stress conditions that determine their sensitivity/tolerance to flooding. Adaptation to long-term submergence is frequently associated with developmental changes such as root aerenchyma formation, internode and petiole elongation, adventitious root development, and alteration of root porosity, morphology, and depth. However, the initial cellular response to decreased oxygen availability is promotion of anaerobic metabolism of pyruvate in both tolerant and intolerant species. In hypoxic/anoxic tissues, the pyruvate content increased and glycolytic (glyceraldehyde-3-phosphate dehydrogenase) and fermentative enzymes [pyruvate decarboxylase (PDC), alcohol dehydrogenase (ADH) and lactate dehydrogenase (LDH)] are induced as a consequence of the need for increased glycolysis to compensate for the lower ATP yield due to the inactivation of oxidative phosphorylation (Saglio et al., 1999; Sato et al., 2002). Fermentative products, i.e. acetaldehyde, ethanol, and lactate, accumulate, allowing for the regeneration of NAD+ from NADH. Regeneration of NAD by the fermentative enzymes ADH and LDH is vital for hypoxia/anoxia tolerance because, in the absence of NAD+, glycolysis ceases (Ismond et al., 2003; Kursteiner et al., 2003). Limitations of lactate accumulation and better regulation of cytoplasmic pH are also part of the strategy that confer tolerance to hypoxia and anoxia. In Medicago truncatula seedlings, it was found that the fermentative activity of alanine aminotransferase (AlaAT) contributed to the control of cytoplasmic pH under hypoxia/anoxia (Ricoult et al., 2005) because (i) synthesis of alanine was accompanied by an increase in γ-aminobutyric acid (GABA) and both amino acids counteract the acidic effect of lactate (Reggiani et al., 2000); and (ii) the alanine synthesis pathway competes with the lactate fermentation pathway for pyruvate, contributing to a limitation of cytoplasmic acidification by lactate. Besides this role, alanine synthesis saves C3 skeletons that would otherwise go through the ethanolic fermentative pathway, thus avoiding a shortage in carbon availability (Ricoult et al., 2005). Furthermore the limitation of the functioning of the ethanol fermentation pathway due to the competition with the alanine pathway limits the synthesis of acetaldehyde, a very toxic compound that constitutes protein adducts (Ismond et al., 2003).
AlaAT, an enzyme widely distributed in animals, plants, yeast, and bacteria, catalyses the reversible reaction of interconversion of alanine and 2-oxoglutarate to pyruvate and glutamate. The roles of AlaAT isoforms have been thoroughly studied in animals, showing different roles for the two isoforms. Recent findings showed that in mammalian liver, starvation increased specifically the activity of the cytosolic (c-AlaAT) isoenzyme but not the mitochondrial isoenzyme (m-AlaAT) (Vedavathi et al., 2004). A higher activity and affinity of liver c-AlaAT for alanine allowed for efficient alanine catalysis in fish species dependent more on amino acids for energy production via gluconeogenesis (Srivastava et al., 2004). In the same fish species, the reaction catalysed by m-AlaAT has been found to be instrumental in the transport of crucial metabolites across the mitochondrial membrane, thereby channelling alanine into general metabolic pathways (Srivastava et al., 2004). In plants, AlaAT gene induction and alanine accumulation have been shown to be involved in at least four roles: (i) in the C4 plant Panicum miliaceum, c-AlaAT-2 was associated with the synthesis of alanine in leaves for the transfer of C3 units (pyruvate) from mesophyll to bundle sheath (Son and Sugiyama, 1992); (ii) a peroxisomal AlaAT with alanine–glyoxylate transaminase activity for glycine synthesis was associated with photorespiration (Liepman and Olsen, 2001, 2003; Fukao et al., 2002; Igarashi et al., 2003); (iii) an AlaAT lacking glyoxylate transamination activity in starchy endosperm of seeds was associated with the conversion of alanine to pyruvate for gluconeogenesis in relation to N utilization of reserves (Sakagishi, 1995); and (iv) in several species, i.e. barley (Hordeum vulgare), corn (Zea mays), Panicum miliaceum, and M. truncatula (Good and Muench, 1993; Muench and Good, 1994; Orzechowski et al., 1999; Ricoult et al., 2005), AlaAT gene expression and the AlaAT fermentative reaction were stimulated in hypoxic or anoxic tissues (Bray et al., 2002).
In a previous study it was shown that the adaptive response of M. truncatula to anoxia/hypoxia stress was not limited to glycolysis and LDH/ADH fermentative reactions. Evidence was presented showing the necessary involvement of the fermentative AlaAT reaction in the adaptation of germinating seeds and seedlings of M. truncatula to low oxygen availability. Hence when alanine synthesis was impaired, germination and seedling development failed under anoxia (Ricoult et al., 2005). To further this investigation, in the present work, the expression of the AlaAT multigene family was characterized in M. truncatula, and the involvement of AlaAT isogene(s) in hypoxia tolerance of young seedlings was determined. Additionally, changes in alanine metabolism under low oxygen availability were studied by the means of [15N]alanine and [15N]glutamate labelling.
Materials and methods
Seed germination under normoxia and hypoxia
Seeds of M. truncatula Gaertn. (cv. Paraggio) were germinated in Petri dishes (diameter 9 cm) on Whatman paper soaked with 4 ml of deionized water, and maintained in a growth chamber in the light at 23 °C (control). Hypoxia was imposed by completely submerging seeds and seedlings in 50 ml of flooding water; this treatment represented a progressive depletion of oxygen (Subbaiah and Sachs, 2003). Gases diffuse approximately 10 000 times more slowly in water than in air, rendering submergence stressful because it inhibits the entry of atmospheric oxygen into plant tissue (Jackson, 1985).
Three replicates of 50 seeds per treatment were used for germination and post-germination growth. For biochemical and molecular analyses, germinated seeds and seedlings were sampled at various times throughout germination and post-germination growth. For all samples, the seed coat, albumen, and cotyledons were removed and embryo axes were collected separately and frozen in liquid nitrogen before being stored at −80 °C. Embryo axes sampled after germination (21 h in the present conditions) corresponded to developed radicles and hypocotyls.
[15N]Glutamate and [15N]alanine labelling experiment
For 15N labelling experiments, seeds were germinated as described before on Petrti dishes (control) or hypoxic medium with added 5 mM 15N-labelled amino acids (99% 15N, Euriso-top, Saarbrüken, Germany). At sampling time, embryo axes were collected, rinsed twice with deionized water, frozen in liquid nitrogen, and stored at −80 °C for further metabolite analysis.
Amino acid extraction and purification
Total amino acids from embryo axes were extracted in 96% ethanol for 1 h at 4 °C. After centrifugation, the ethanol fraction was removed and the same process was then repeated with deionized water. The ethanol and water fractions were combined and stored at −20 °C.
After evaporation of the extract under vacuum, organic residues were dissolved in distilled water and extracted with the same volume of chloroform. After centrifugation, an aqueous phase containing amino acids was vacuum dried. Then, amino acids were redissolved in distilled water and passed through a 2 ml column of Dowex 50WX8 200–400 mesh, H form resin (Supelco, Bellefonte, PA, USA). Columns were washed three times with 1 ml of water, and amino acids were eluted with 5 ml of 6 M NH4OH. After evaporation of the extract under vacuum, amino acids were resuspended in deionized water of high-performance liquid chromatography (HPLC) quality.
Detection and quantification of amino acids by HPLC
The amino acids were determined as q-phthaldialdehyde derivatives on a C-18 column using 32-Karat software (Beckman-Coulter, Fullerton, CA, USA). The internal standard was GABA and the gradient was produced using two eluents: 50 mM sodium acetate buffer (pH 5.9) with 200 ml l−1 methanol and 100% methanol.
Detection of labelled amino acids by gas chromatography–mass spectrometry (GC–MS)
Samples of purified amino acids were vacuum dried and derivatized by N-methyl-N-(tert-butyldimethylsilyl)-trifluoroacetamid (Alltech, Templemars, France) for 30 min at 90 °C as described in Glevarec et al. (2004). The BDMS (butyldimethylsilyl) derivatives obtained were analysed by GC–MS with a Fisons MD800 quadrupole GC–MS system in order to determine incorporation of 15N into amino acids. A sample of 0.4 μl was injected into the gas chromatograph fitted with an on-column injector, and a 30 M methylpolysiloxane, 0.25 μm film thickness, fused silica capillary column. Helium was used as a carrier gas with a flow rate of 1 ml min−1, and the oven was temperature programmed from 60 °C for 1 min, +30 °C min−1 to 120 °C, and from 120 to 260 °C at 8 °C min−1. Mass spectra were acquired with an electron energy of 70 eV, and the mass range scanned from mass to charge ratios of 50–650 with a total scan time of 0.9 s. Incorporation of 15N (atom % excess) was calculated after integrating the areas obtained for fragment ions for both labelled and unlabelled amino acids. BDMS derivatives allow the determination of the total amount of 15N incorporated into amino acids and the proportion of the amide that is singly and doubly labelled.
Protein extraction and AlaAT detection by SDS–PAGE and immunobloting
Soluble proteins were extracted from frozen plant material in Tris–HCl buffer (25 mM, pH 7.6) supplemented with 1 mM MgCl2, 1 mM EDTA, 1 μl ml−1 2-mercaptoethanol, and polyvinylpolypyrrolidone (Sigma). After denaturation, protein samples were separated on an SDS–polyacrylamide gel [10% (w/v) polyacrylamide]. The proteins were then electrophoretically transferred onto a PVDF membrane (Bio-Rad, Hercules, CA, USA). Antiserum raised in rabbits against AlaAT of barley (Muench and Good, 1994) was used to detect AlaAT protein. Protein–antibody complexes were located using peroxidase-conjugated goat anti-rabbit immunoglobulin G.
Cloning of the 5′ and 3′ ends of AlaAT cDNA
A 100 ng aliquot of poly(A)+ RNA was purified from total RNA from M. truncatula embryo axes after 37 h of imbibition with the PolyATract® System 1000 (Promega, Madison, Wisconsin, USA). The 5′ and 3′ ends of the AlaAT cDNA were obtained with a reverse primer (5′-CGTTGTCACCATATCCCATGGATCG-3′) and a forward primer (5′- GGTATTTTGTGTCCCATCCCCCAGT-3′), respectively, with the SMART™ RACE cDNA Amplification Kit (Clontech, Palo Alto, CA, USA). Primers were designed from the expressed sequence tag (EST) MtC00229 sequence (annotated as probable alanine aminotransferase) from the Toulouse M. truncatula database (http://medicago.toulouse.inra.fr/Mt/EST/).
RNA extraction and reverse transcription
Total RNA from embryo axes was extracted using TRIzol Reagent (Invitrogen, Breda, The Netherlands) according to the manufacturer's protocol. A 2 μg sample of total RNA was reverse transcribed for 1 h at 37 °C, using 200 U of M-MLV reverse transcriptase (Promega) and 2 μg of pd(N)6 random hexamer (Amersham Biosciences, Freiburg, Germany) in the presence of 40 U of recombinant RNasin ribonuclease inhibitor (Promega), and in a final volume of 50 μl. Genomic DNA was removed by purifying the first strands using a QIAquick PCR Purification Kit (Qiagen, German Town, MD, USA).
Real-time PCR and SYBR Green detection
PCR was performed on a light cycler ABI Prism 7000 SDS (Applied Biosystems, Foster City, CA, USA) with the SYBR Green PCR Master Mix (Applied Biosystems) according to the manufacturer's protocol. Each reaction was performed on 5 μl of a 1:2 (v/v) dilution of the first cDNA strands, synthesized as described above with 0.3 μM of each primer in a total reaction of 20 μl. The reactions were incubated for 2 min at 50 °C and 10 min at 95 °C, followed by 60 cycles of 15 s at 95 °C and 1 min at 60 °C. The specificity of the PCR amplification procedures was checked with a heat dissociation protocol (from 65 °C to 95 °C) after the final cycle of the PCR. Each reaction was done in triplicate and the corresponding Ct values were determined.
Amplified fragments (70–100 bp) were cloned into pGEM T Easy vector (Promega). The plasmids were diluted several times to generate templates ranging from 105 to 103 copies used for standard curves for the estimation of copy numbers. The results are expressed as copy number of cDNA in 5 μl of first strands.
The sequences of AlaAT isogenes were used to design primer sets for the amplification of cDNAs: m-AlaAT forward primer, CCGCACCTGATGCCTTCTAT, reverse primer, AAAATGCCATGTGCCAGGAA; c-AlaAT forward primer, CCGCACCTGATGCCTTCTAT, reverse primer, AAAATGCCATGTGCCAGGAA; AGT forward primer, CCGCACCTGATGCCTTCTAT, reverse primer, AAAATGCCATGTGCCAGGAA; and AlaAT of class IV forward primer, CCGCACCTGATGCCTTCTAT, reverse primer, AAAATGCCATGTGCCAGGAA.
Cloning of AlaAT cDNA and in silico identification of the AlaAT multigene family in M. truncatula
Total RNAs of the embryo axis of germinating M. truncatula (48 h of imbibition) were reverse transcribed and selective amplification of AlaAT full-length cDNA was performed using the 3′–5′ RACE (rapid amplification of cDNA ends) PCR technique. Only one clone was obtained; the amino acid sequence deduced from its open reading frame (ORF) revealed that it encodes a mature protein of 524 amino acids. Using this coding sequence as a BLASTN query for TIGR's MtGI and INRA-Toulouse (http:medicago.toulouse. inra.fr Mt EST) data bank, four sequences of cDNAs encoding AlaAT isogenes were identified. Two of the four cDNAs code for two AlaAT proteins (EC 220.127.116.11) belonging to class I and II of pyridoxal phosphate aminotransferase. Both enzymes catalyse the reversible transamination L-glutamate+pyruvate↔L-alanine+2-oxoglutarate. A third cDNA codes for an AlaAT protein (EC 18.104.22.168) belonging to class III of pyridoxal phosphate aminotransferase. The enzyme catalyses the reversible transamination of glycine, L-alanine+glyoxylate↔pyruvate+glycine; it is called alanine–glyoxylate aminotransferase (AGT). A fourth cDNA codes for an AlaAT protein (EC 22.214.171.124 and EC 126.96.36.199) that is an alanine/branched chain amino acid aminotransferase belonging to class IV of pyridoxal phosphate aminotransferase. The enzyme catalyses the reversible transamination of valine and leucine, L-valine+pyruvate↔3-methyl-2-oxobutanoate+L-alanine and L-leucine+2-oxoglutarate↔4-methyl-2-oxopentanoate+L-glutamate.
In silico translation of the ORFs showed that the two AlaAT genes of class I and II, AlaAT of class III and AlaAT of class IV of pyridoxal phosphate aminotransferase encode proteins of 524, 481, 481, and 344 amino acid residues, respectively. Alignment of the deduced amino acid sequences was done by Multalin (Corpet, 1988) (Fig. 1). The highest percentage identity (44.5%) was found between the two AlaAT proteins of class I and II of pyridoxal phosphate aminotransferase. Identity between AlaAT proteins of class III and IV was low (18.1%); they also showed low identities with AlaAT proteins of class I and II (18.4–21%). PSORT algorithm [http://psort.nibb.ac.jp/form.html (Nakai and Horton, 1999)] and Pence Proteome Analysis for subcellular localization of proteins [http://www.cs.ualberta.ca/∼bioinfo/PA/Sub/ (Lu et al., 2004)] allowed the prediction of a mitochondrial localization for the AlaAT isoform of 524 amino acid residues belonging to class I and II (P = 0.82 and 1.0, respectively); it will be named putative m-AlaAT. A cytosolic localization was predicted for the AlaAT isoform of 481 amino acid residues belonging to class I and II (P=0.65 and 0.99; respectively); it will be named putative c-AlaAT. For AGT, a peroxisomal localization was predicted by the PSORT algorithm (P=0.56) and a mitochondrial localization was predicted by Pence Proteome Analysis (P=0.99). A chloroplastic localization was predicted for the AlaAT isoform with branched chain amino acid transaminase activity (P=0.99) by Pence Proteome Analysis, and a peroxisomal localization was predicted by the PSORT algorithm P=0.49). The sequence identified by RACE PCR in the M. truncatula embryo axis corresponded to AlaAT cDNA coding for m-AlaAT.
Organ specificity of AlaAT isogene expression in M. truncatula
In order to check for developmental regulation and organ specificity of AlaAT isogenes, their expression was quantified by quantitative RT–PCR in embryo axes, roots, stems, and leaves of adult plants. Four pairs of PCR primers were designed in the 3′- or 5′-untranslated region for specific amplification of each AlaAT isogene reverse-transcribed RNA. The results in Fig. 2 show that the putative mitochondrial AlaAT (m-AlaAT) is expressed in all organs of adult plants and the embryo axis. Putative cytosolic AlaAT (c-AlaAT) was specifically expressed in leaves of adult plants. AGT was expressed mainly in the embryo axis, and stems and roots of adult plants, and AlaAT with branched chain amino acid transaminase activity was specifically expressed in stems and roots of adult plants (Fig. 2).
Effect of hypoxia on the expression of AlaAT isogenes in M. truncatula under normoxia and hypoxia
It has been shown previously (Ricoult et al., 2005) that hypoxia resulted in a significant increase in the expression of fermentative genes encoding ADH, LDH, and AlaAT in the M. truncatula embryo axis throughout germination and post-germination growth. However, alignment of AlaAT multigene family sequences showed that AlaAT expression that has been measured did not correspond to a specific isogene because the pair of PCR primers for quantitative RT–PCR was designed in a cDNA region conserved in the four AlaAT isogenes.
In the present work, the response of AlaAT to hypoxia was investigated further in M. truncatula seedlings by the quantification of the expression of each AlaAT isogene under hypoxia (Fig. 3). For this purpose, M. truncatula seedlings imbibed for 24 h in control (normoxia) conditions were either maintained under normoxia or transferred for 5, 12, and 24 h to hypoxia and expression of AlaAT isogenes was quantified by quantitative RT–PCR using primers specific for each isogene (Fig. 3). The experiment has been repeated three times, each corresponding to a biological repetition. Only the results of one biological repetition with three technical repetitions are presented in Fig. 3. The onset of hypoxic metabolism in young seedlings has been monitored by quantification of the expression of genes encoding the fermentative enzymes ADH and LDH (data not shown). After 12 h under hypoxia, expression of ADH and LDH increased significantly compared with the normoxic control, indicating that under these experimental conditions young seedlings shifted from normoxic to hypoxic metabolism.
Expression of putative m-AlaAT was boosted by hypoxia; it was four times higher under hypoxia compared with under normoxia at 12 h and 24 h of treatment. Expression of AGT was inhibited by hypoxia, being roughly twice lower under hypoxia compared with under normoxia (Fig. 3). Expression of genes encoding putative cAlaAT and AlaAT of class IV of pyridoxal phosphate transaminase was, as expected, very low in the control condition and was not stimulated by hypoxia (Fig. 3).
Effect of hypoxia on AlaAT proteins determined by immunobloting
The amount of AlaAT proteins was monitored by immunoblotting after protein separation by SDS–PAGE in embryo axes developing under normoxia and hypoxia. Only two bands probably corresponding to two AlaAT proteins were detected. From the size of the detected proteins, it is suggested that they corresponded to proteins encoded by the two major AlaAT isogenes expressed in embryo axes, m-AlaAT (57.9 kDa) and AGT (53 kDa). It is interesting to note that at 24 h of treatment, in agreement with gene expression data, the amount of putative m-AlaAT was higher in hypoxic embryo axes than in the control, while the amount of putative AGT was higher in the control than in hypoxic embryo axes (Fig. 4).
Effect of hypoxia on alanine metabolic pathways analysed by 15N labelling
By using an in vivo [15N]glutamate labelling technique for the assessment of AlaAT activity, it has been shown previously (Ricoult et al., 2005) that alanine accumulated more in anoxic and hypoxic seedlings compared with the normoxic control. Given the new findings on AlaAT isoforms, the aim of the present experiment was to study alanine metabolism more thoroughly by comparing the fate of the substrates of the two enzymes, m-AlaAT and AGT, expressed in embryo axes under either hypoxia or normoxia. For this purpose, seeds were germinated and seedlings developed in normoxia and hypoxia; [15N]glutamate or [15N]alanine (99% atom excess) was added at a final concentration of 5 mM. Embryo axes were sampled at 30 h and 50 h of imbibition, labelled amino acids were analysed by GC–MS (Fig. 5), and amino acids were quantified by HPLC (Fig. 6). The results of labelling are expressed as a percentage of 15N abundance in each amino acid.
The [15N]glutamate labelling experiment showed that alanine synthesis activity in vivo, using glutamate as amino donor (glutamate–pyruvate transaminase), was higher under hypoxia than under normoxia. At 50 h of imbibition, the 15N abundance in alanine was almost four times higher in hypoxic embryo axes than in the control (Fig. 5a). The 15N abundance in GABA was also higher in hypoxic embryo axes than in the control, while the 15N abundance in glycine was similar in both treatments, indicating that the glutamate–glyoxylate transaminase activity was not affected by hypoxia (Fig. 5a). The 15N abundance in serine was very low in both treatments, although in hypoxia treatment it was double that in normoxia treatment (Fig. 5a).
The [15N]alanine labelling experiment showed that the pathways utilizing alanine in vivo through alanine–oxoglutarate transaminase or alanine–glyoxylate transaminase for glutamate or glycine synthesis, respectively, were inhibited under hypoxia. While at 50 h of imbibition, under normoxia, the 15N abundance in glutamate, GABA, glycine, and serine was significantly high, no labelled glycine or glutamate was detected in hypoxic embryo axes. GABA and serine, derived from glutamate and glycine, respectively, were either not labelled or labelled less in hypoxic embryo axes compared with the control (Fig. 5b).
Amino acids have been quantified in order to determine whether differences in 15N abundance in amino acids were accompanied by differences in the amounts of these amino acids. In Fig. 6, the amounts of glutamate, alanine, glycine, and serine at 50 h of imbibition are shown to support labelling data pertaining to AlaAT and AGT activities. The amounts of asparagine are shown for the purpose of comparison as this is the most abundant amino acid in M. truncatula under control conditions. When embryo axes were fed [15N]glutamate, alanine accumulated as the major amino acid, supporting labelling data showing that AlaAT activity using glutamate as amino donor was stimulated by hypoxia. Glycine and serine accumulated in both normoxic and hypoxic embryo axes, supporting labelling data suggesting that glutamate–glyoxylate transaminase activity might have been operational under normoxia as well as hypoxia (Fig. 6a). When embryo axes were fed [15N]alanine, the amounts of glutamate, glycine, and serine were lower under hypoxia than in the control (Fig. 6b). It is particularly interesting to note that neither glutamate nor glycine were detected in embryo axes at 50 h of imbibition, in agreement with the absence of [15N]glycine and [15N]glutamate in the same embryo axes, supporting the idea that AlaAT and AGT activities using alanine as amino donor were inhibited by hypoxia.
Alanine amino transferase multigene family in M. truncatula
AlaATs belong to a pyridoxal phosphate multigene family widely distributed in animals, plants, algae, yeast, and bacteria, with isoforms localized in various organs and cell compartments, i.e. cytosol, mitochondria, and peroxisomes (Vedavathi et al., 2004). The enzymes catalyse transamination reactions using several amino donor:acceptor combinations: alanine:2-oxoglutarate, glutamate:pyruvate, alanine:glyoxylate, serine:glyoxylate, and serine:pyruvate (Sakagishi, 1995; Ward et al., 2000; Liepman and Olsen, 2001, 2003; Srivastava et al., 2004; Vedavathi et al., 2004).
The present work is the first to report on the expression of AlaAT isogenes in the model legume M. truncatula. Only the gene encoding the putative mitochondrial isoform (m-AlaAT) was expressed, at very similar levels, in all analysed organs of adult plants (root, stem, and leaf) and in embryo axes of young seedlings. Conversely, the gene encoding the putative cytosolic isoform (c-AlaAT) was expressed only in leaves of adult plants. The results show that although m-AlaAT and c-AlaAT catalyse the same reversible reaction of alanine:2-oxoglutarate and glutamate:pyruvate transamination, they might fulfil different roles in planta in relation to their organ and subcellular localizations. These two isoforms of AlaAT have been more thoroughly characterized in mammals than in plants. They are expressed in various organs: brain, kidney, skeletal and cardiac muscles, and particularly in liver where it appeared that the cytosolic isoform was dedicated to alanine catalysis allowing for energy production from amino acids through the gluconeogenesis pathway (Srivastava et al., 2004; Vedavathi et al., 2004), while the mitochondrial isoform appeared to be dedicated to alanine synthesis and transport across the mitochondrial membrane (Srivastava et al., 2004).
The gene encoding AGT was expressed in all organs of adult plants with, however, a very low level of expression in the leaf compared with that in root and stem. Two algorithms have been used for subcellular localization of this enzyme, the PSORT algorithm predicted a peroxisomal localization (P=0.56), while Pence Proteom Analysis predicted a mitochondrial localization (P=0.99). To our knowledge, available data in the literature assigned a peroxisomal localization to AGT with a photorespiratory role; however, comparison of M. truncatula AGT protein with sequences of Arabidopsis thaliana AGT proteins showed the best percentage identity with mitochondrial isoforms, 80.5% identity with AGT2 (NP568064) and 60% identity with AGT3 (NP181374), while the identity with the peroxisomal AGT1 (NP849951) was very low (18.4%). Taken together, the results support a mitochondrial localization of M. truncatula AGT protein where it should be involved in the glyoxylate cycle and glycine synthesis. The gene encoding the AlaAT isoform with branched chain amino acid aminotransferase activity was expressed only in stems and roots of adult plants. It showed a very low level of expression, 4–6 times lower than the level of expression of the other three isogenes.
Molecular and biochemical aspects of alanine metabolism as affected by hypoxia in M. truncatula seedlings
AlaAT isogenes expressed in the M. truncatula embryo axis, i.e. m-AlaAT and AGT, were differentially regulated by hypoxia. While the levels of transcript and protein of m-AlaAT increased under hypoxia, levels of transcript and protein of AGT decreased. It is concluded that the formerly observed increase in AlaAT gene expression under hypoxia stress (Ricoult et al., 2005) in embryo axis was accounted for by the increase in putative mAlaAT isogene and that in the multigene family of AlaATs, m-AlaAT is the one involved in the adaptive response of embryo axis to hypoxia. Although AlaAT induction by hypoxia has been reported by several authors (Good and Muench, 1993; Muench and Good, 1994; Orzechowski et al., 1999; Ricoult et al., 2005), this is the first time this induction is ascribed to a specific isoform that is very probably located in mitochondria.
Changes in gene expression and protein synthesis were accompanied by changes in in vivo enzyme activities and alanine metabolism as revealed by 15N labelling experiments. Feeding the embryo axis [15N]glutamate or [15N]alanine under normoxia showed that m-AlaAT catalysed a reversible reaction allowing for synthesis of alanine with glutamate as the amino donor and synthesis of glutamate with alanine as the amino donor. The same experiment showed that glycine synthesis occurred at the expense of either glutamate or alanine, indicating that beside AGT, a glutamate–glyoxylate transaminase was also operating. Under hypoxia, however, feeding the embryo axis either [15N]glutamate or [15N]alanine showed that m-AlaAT activity was directed towards alanine synthesis using glutamate as the amino donor (Fig. 5a), while the reaction of glutamate synthesis using alanine as the amino donor was inhibited (Fig. 5b). The results indicate that the putative m-AlaAT isoform is regulated at both transcriptional and post-translational levels. A double level of regulation by hypoxia allowed for an increase in gene expression and the orientation of the activity of the product of its transcription towards alanine synthesis. As a result, labelled alanine was four times higher in hypoxic embryo axes than in the controls (Fig. 5a) and alanine accumulated as the major amino acid instead of asparagine (see also Ricoult et al., 2005). By competing with ethanolic fermentation for pyruvate, alanine synthesis saves C3 skeletons, thus avoiding a shortage in carbon availability, and limits accumulation of acetaldehyde, a toxic compound. Also, an increase in alanine synthesis, by competing with lactic fermentation for pyruvate, intervenes in cytosolic pH regulation. Furthermore, synthesis of alanine arising from the decarboxylation of malate to pyruvate as a result of malic enzyme activation by hypoxia along with decarboxylation of glutamate to GABA are proton-consuming reactions (Carroll et al., 1994; Edwards et al., 1998; Bouché and Fromm, 2004; Ricoult et al., 2005). AGT was inhibited at the transcriptional level by hypoxia; as a result the amount of the protein revealed by immunoblot was lower in hypoxic embryo axes than in the control. Furthermore, the in vivo enzyme activity of AGT in the direction of glycine synthesis was inhibited by hypoxia, as shown by the fact that almost no labelled glycine was detected in embryo axes when they were fed [15N]alanine and the amount of glycine was dramatically lower in hypoxic embryo axes than in the controls (Fig. 6b). The total absence of glycine (both labelled and unlabelled) in hypoxic embryo axes fed [15N]alanine meant that glutamate–glyoxylate transaminase did not compensate for the lack of AGT activity probably because glutamate was competitively recruited to alanine and GABA synthesis pathways catalysed by m-AlaAT and GDC (GABA decarboxylase). Hence, when embryo axes were fed 5 mM [15N]glutamate, glycine was synthesized in hypoxic embryo axes (Fig. 6).
This work has been funded by grants from the EU FP6 project FOOD-CT-2004-506223, ‘Grain Legumes Integrated Project’ (GLIP), and Contrat Etat Région pays de la Loire/France CER-Semences. The authors are grateful to Dr Allen Good for the gift of AlaAT antibodies.