INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

Chloroplasts of the higher plants and green algae contain NADP-dependent malate dehydrogenase (NADP-MDH).¹ The enzyme is completely inactive in the dark and redox-activated in the light through an oxidoreduction chain consisting of ferredoxin, ferredoxin-thioredoxin reductase, and thioredoxin. In C3 plants NADP-MDH is a part of the malate valve that transfers excess reducing equivalents from chloroplast to cytosol and serves therefore to balance the stromal ATP/NADPH ratio (1, 2).

In the last few years the body of evidence has grown to suggest that the plastidic malate valve functions not only in the light but also in the dark. Incubation of spinach chloroplasts and red pepper fruit chromoplasts with dihydroxyacetonephosphate led to significant formation of 3-phosphoglycerate in the dark only in the presence of the oxaloacetate (OAA) in the incubation media (3). Incubation of chloroplasts isolated from CAM-induced Mesembryanthemum crystallinum with OAA in the dark provoked considerable rates of malate synthesis (4). The rate of starch degradation, which is dependent in the isolated plastids on exogenously supplied ATP, was stimulated by the addition of OAA, and the synthesis of malate was accompanied with nearly stoichiometric synthesis of 3-phosphoglycerate. An essentially similar situation was observed during fatty acid and glycerolipid synthesis from acetate in isolated pea root plastids (5), which is strictly dependent on exogenously supplied ATP. When provided with dihydroxyacetonephosphate and inorganic phosphate, the plastids could generate ATP for fatty acid synthesis themselves but only when OAA was added to the incubation medium.

Redox-modulated NADP-MDH is completely inactivated in the dark and is altogether absent from root plastids (6). This implies that one more, NAD-dependent malate dehydrogenase isoenzyme, which is responsible for the observations summarized above, must be associated with plastids. The hypothesis of the existence of an NAD-dependent MDH in chloroplasts is not new. In the seventies, in reviews concerning the plastidic malate-oxaloacetate shuttle (7, 8) considerable disagreement existed about whether NADH or NADPH participates in the reduction of OAA in the illuminated chloroplasts. However, findings of NAD-MDH associated with chloroplasts (9) were later dismissed as an artifact caused by a high degree of contamination in chloroplasts isolated by contemporary methods (10). The discussion about the plastidic NAD-MDH came to an end temporarily with the discovery that the low activities of NADP-MDH can be dramatically increased in vivo by light and in vitro by various thiols (11-13). For the last 20 years it has been widely accepted that chloroplasts contain only one, strictly NADP-dependent MDH isoform, because plastidic NAD-MDH was never purified and any definitive proof of its existence was lacking. The elusiveness of the plastidic NAD-MDH can be explained with the following reasons.

First, NAD-dependent MDHs (EC 1.1.1.37) are present in mitochondria, microbodies, and cytosol (for a review see Ref. 14). Each of the three plant cell compartments contains, in turn, multiple molecular isoforms of the enzyme, their total number varying from five to nine and more in different plants (15, 16).

Second, all eukaryotic NAD-MDHs are homodimers with a subunit molecular mass of 33-36 kDa. Being compared, their sequences fall into two groups: "mitochondrial" and "cytosolic", with an amino acid identity of above 50% inside each group and only 20-25% between them. These conserved amino acids are concentrated in the regions of the polypeptide chain that are responsible for dinucleotide binding, catalysis, and helix formation (17). Therefore, despite the low primary structural similarity, NAD-MDHs from different compartments of plant cell have nearly identical spatial structure and catalytic properties.

Third, although the NAD-MDH activity was repeatedly reported in isolated plastids of several plants (18-20), the measured activity was very low. For example in the chloroplasts from Chenopodium rubrum it comprised only 2.6% of the total cellular enzyme activity (21). Because the isolated plastids were never free from at least a few percent of mitochondrial and/or peroxisomal contamination one could not be sure that the registered NAD-MDH activity is actually stroma-localized.

To resolve these problems we combined improved methods of chloroplast isolation with a molecular approach. Here we present unequivocal evidence for the existence of NAD-dependent MDH in spinach and Arabidopsis chloroplasts and report the cDNA sequence encoding the Arabidopsis enzyme.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

Plant Material-- Plants of Arabidopsis thaliana L. (Heyhn) ecotype Columbia and Spinacia oleracea L. var. U. S. hybrid 424 were grown at 23 °C with a photoperiod of 16 h of light/8 h of dark. Arabidopsis was grown in soil, and spinach was grown hydroponically according to Ref. 22.

Preparation of Crude Extracts from Spinach and Arabidopsis Leaves-- Fresh leaves were powdered under liquid nitrogen. The powder was mixed with 10 mM Tris-HCl (pH 8.0), 1 mM EDTA, 0.1 mM Pefabloc SC (Serva, Heidelberg, Germany), 0.6% polyvinylpolypyrrolidone. The suspension was sonified twice for 30 s on ice and spun (10 min, 23000 × g) to remove insoluble fraction.

Isolation of Chloroplasts from Spinach and Arabidopsis-- Spinach chloroplasts were isolated as in Ref. 23. Chloroplasts from Arabidopsis were prepared from protoplasts according to Ref. 24. The chloroplast extraction medium (medium C) contained 0.3 M sorbitol, 20 mM Tricine (pH 8.4), 10 mM EDTA, and 0.1% BSA. To remove remaining contamination with other cell compartments, the method was supplemented by centrifugation through a continuous density gradient (50% Percoll in medium C, 30 min at 1000 × g). Intact chloroplasts appeared as a narrow band near the bottom of the tube and were washed twice by dilution (medium C without BSA) and centrifugation (1500 × g for 90 s including acceleration and deceleration). The final pellet was resuspended in a small volume of buffer A (10 mM Tris-HCl (pH 8.0), 1 mM EDTA, 0.1 mM Pefabloc SC), and chloroplasts were disrupted by sonication. Membranes were removed by centrifugation for 20 min at 24000 × g, and the resulting supernatant was used for subsequent examinations.

Isolation of Mitochondria from Arabidopsis-- Fresh leaves were homogenized with chilled medium I (0.3 M sucrose, 30 mM MOPS-NaOH (pH 7.5), 10 mM EDTA, 5 mM glycine, 4 mM ascorbate, 0.6% polyvinylpolypyrrolidone, and 0.2% BSA). The homogenate was filtered, and the filtrate was spun twice for 25 min at 3500 × g. Mitochondria were collected from the supernatant by centrifugation (20 min, 15000 × g) and resuspended in 3-4 ml of medium II (medium I without ascorbate and polyvinylpolypyrrolidone). The suspension was layered on a continuous density gradient (30 ml of 33% Percoll in medium II) and spun for 30 min at 35000 × g in a swing bucket rotor. Only the lowest 10% of the gradient, containing fumarase activity, was collected. This fraction was diluted with medium II without BSA and centrifuged (10 min at 23000 × g). The mitochondria were resuspended in buffer A and broken by sonication, and the soluble fraction was used for subsequent examinations.

Analysis of Purity of Isolated Organelles-- To check the purity of isolated organelles, enzyme activities specific for different cell compartments (NADP-GAPDH (EC 1.2.1.13) for chloroplasts, fumarase (EC 4.2.1.2) for mitochondria, 2-hydroxypyruvate reductase (EC 1.1.1.81) for peroxisomes, and UDP-glucose pyrophosphorylase (EC 2.7.7.9) for cytosol) were measured in crude extract (CE) and in the chloroplast (CP) and mitochondrial preparations. The fractions of total enzyme activities in the chloroplast preparation were calculated according to the following formula: [NADP-GAPDH (CE)/enzyme (CE)] × [enzyme (CP)/NADP-GAPDH (CP)] × 100. The relative recoveries of enzyme activities in the mitochondrial preparation were calculated accordingly with fumarase instead of NADP-GAPDH.

Enzyme Assay and Chlorophyll and Protein Determination-- Enzyme activities of NAD-MDH and NADP-MDH (25), NADP-GAPDH (26), fumarase (27), 2-hydroxypyruvate reductase (28), and UDP-glucose pyrophosphorylase (29) were measured at 25 °C. To test whether the activity of plastidic NAD-MDH is affected by reduction, the sample was preincubated under reducing conditions as described for NADP-MDH. Chlorophyll was determined according to Ref. 30. Protein was determined according to Ref. 31 using crystalline BSA as a standard.

SDS-Polyacrylamide Gel Electrophoresis and Isoelectric Focusing-- SDS-polyacrylamide gel electrophoresis was performed as described (32). The acrylamide concentrations in the resolving gel and in the stacking gel were 12.5 and 2.5% respectively. ¹⁴C-Labeled molecular-mass standards were obtained from Sigma (Deisenhofen, Germany).

In Vitro Translation and Targeting Experiments-- cDNAs coding for different NAD-MDH isoenzymes inserted in pZL-1 plasmid between SalI and NotI cloning sites were obtained from ABRC DNA Stock Center. The plasmids were used for coupled transcription/translation with a TNT Coupled Reticulocyte Lysate System (Promega, Heidelberg, Germany) in the presence of [³⁵S]methionine according to the instructions of the manufacturer. In vitro synthesized ³⁵S-labeled precursor proteins were used for import experiments into isolated spinach chloroplasts as described (34). After pretreatment with thermolysin the chloroplasts were recollected and fractionated into envelope membranes, thylakoids, and stroma as in Ref. 35. The samples were subsequently analyzed by SDS-polyacrylamide gel electrophoresis and autoradiography.

Northern Blot Analysis-- Total RNA was isolated from Arabidopsis leaves and roots according to Ref. 36. Poly(A)⁺ RNA was isolated from total RNA with Oligotex-dT mRNA mini-kit (Qiagen, Hilden, Germany). Denaturing gel electrophoresis of RNA was carried out in a 1.5% agarose gel containing formaldehyde (37), and RNA was blotted onto Nylon membranes (Amersham, Little Chalfont, UK).

Expression of A22 in Escherichia coli-- The region of A22 encoding the predicted mature MDH was amplified by Pfu polymerase in polymerase chain reaction using the sense primer A22N (5'-CAAAATCCATATGTCG TAC AAA GTA GCT GTT C-3') and the antisense primer A22C (5'-ATGTTATCAGCTGGATCC CTA GTT AGC TGC-3'). Start and stop codons are shown in bold, and restriction sites NdeI and BamHI are underlined. The polymerase chain reaction product was digested with NdeI and BamHI and cloned into the corresponding restriction sites of the vector pET-21a⁺ (Novagen/AGS, Heidelberg, Germany). The resulting plasmid, pET-A22, was transformed into BL21 (DE3) pLysS E. coli cells. Synthesis of recombinant protein and preparation of crude bacterial extracts were performed as in Ref. 38. Cells containing the original plasmid (pET-21a⁺) without an insert were treated accordingly as a control.

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

Demonstration of NAD-MDH Activity in the Extremely Pure Chloroplasts from Spinach and Arabidopsis-- To obtain the plastid preparation free from contamination with other cell compartments, we isolated spinach chloroplasts according to Ref. 23. Isolated chloroplasts were virtually free from contamination with cytosol, mitochondria, and peroxisomes, as the activities of the marker enzymes UDP-glucose pyrophosphorylase, fumarase, and 2-hydroxypyruvate reductase were below the limits of detection. Additionally, we developed a method for preparation of extremely pure chloroplasts from A. thaliana leaves (Table I). Chloroplasts from both plants contained significant activities of NAD-dependent MDH (2.7 units/mg chlorophyll for spinach and 1.7 units/mg chlorophyll for Arabidopsis), which were in the same range as NADP-dependent activities measured under reducing conditions (3.8 units/mg chlorophyll for spinach and 2.1 units/mg chlorophyll for Arabidopsis). No NADP-MDH activity could be detected under oxidizing conditions, and in turn, NAD-MDH activity did not change significantly upon incubation with dithiothreitol. Similarly to the results obtained with chloroplasts from C. rubrum (21), extremely pure chloroplasts from spinach and Arabidopsis contained less then 2% of the total cellular NAD-MDH activity (Table I). As previously mentioned, all known plant NAD-MDHs are very similar in terms of their structure and catalytic properties, which should make the task of purification of the new enzyme isoform quite unpromising.

Identification of the cDNA Clone Coding for the Plastidic NAD-MDH-- To find a cDNA clone coding for plastidic NAD-MDH (pNAD-MDH) we extracted from the EST Arabidopsis data base all putative NAD-MDH clones divided between three known classes by BLAST sequence alignment program and analyzed them closely. Most sequences in the Arabidopsis EST data base are from the oriented PRL1 and PRL2 libraries and contain on average 375 bp from the 5'-end of the cDNA (39). All clones placed in the cytosolic group contained initiating ATG codon in the same position as cytosolic MDH from M. crystallinum (40). Therefore, they were of no interest to us, because we were looking for a chloroplast protein, which must possess an N-terminal transit peptide. Plant mitochondrial NAD-MDHs (mNAD-MDHs) start with easily recognizable mitochondrial signal sequences of 19-26 amino acids, which were also found in putative Arabidopsis mNAD-MDH clones. One such clone, 132N21T7, designated A20, was requested from the ARBC DNA stock center and completely sequenced. Microbody NAD-MDH (mbNAD-MDH) belongs to a small group of microbody proteins, which are also directed into the organelle by cleavable presequences on their N-terminal ends, termed PTS2 (peroxisome-targeting signal 2) (41). The PTS2 contain two consensus motifs, RL-X5-HL and SXLXXAXCXA, essential for targeting into the organelle and for subsequent processing of the N-terminal region, respectively. They are also present in the typical representative of Arabidopsis mbNAD-MDH clones, 226H24T7, designated A33, which was also completely sequenced.

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Alignment of the N-terminal parts of precursor proteins for NAD-MDHs from Arabidopsis plastids (A22), mitochondria (A20), and microbodies (A33). The putative N termini of the mature proteins are marked by asterisks. Identical amino acids are marked by black background. Gaps introduced during alignment are indicated by dots.

A22 in Vitro Translation Product Is Imported into Isolated Spinach Chloroplasts-- To verify that the A22 clone encodes a chloroplast protein, we performed an in vitro translation reaction using a T7 reticulocyte coupled transcription/translation system. A [³⁵S]methionine-labeled polypeptide of the expected size of approximately 43 kDa was obtained. Additional bands of lower molecular masses also appeared, which may represent polypeptides initiated at internal translation sites or products of premature termination of translation. The obtained translation mix was incubated with intact isolated spinach chloroplasts in the light and in the dark. After fractionation of the chloroplasts into stroma, thylakoids and envelope membranes, the labeled product, processed to a mature form of approximately 34 kDa, was found exclusively in the stroma fraction of illuminated chloroplasts (Fig. 2). A parallel control experiment was performed with the A33 clone, coding for mbNAD-MDH. As expected, in vitro synthesized A33 precursor protein was not imported into the chloroplasts either in the light or in the dark (Fig. 2).

View larger version (73K): [in this window] [in a new window]   Fig. 2.   Import of A22 and A33 in vitro translation products into isolated spinach chloroplasts. Polypeptides encoded by the clones A22 and A33 were synthesized via coupled in vitro transcription/translation assay in the presence of [35S]methionine (lanes 1). The precursors were incubated with isolated chloroplasts either in the light (lanes 2) or in the dark (lanes 3). The chloroplasts were recollected, and the stroma fractions were recovered and analyzed by SDS-polyacrylamide gel electrophoresis and autoradiography.

Northern Blot Analysis of the Organ-specific Transcription of the Plastidic NAD-MDH-- As described in the Introduction, the presence of NAD-specific malate dehydrogenase is necessary to explain both OAA-dependent 3-phosphoglycerate production in "dark" chloroplasts and OAA and dihydroxyacetonephosphate substitution for ATP in driving fatty acid and glycerolipid biosynthesis in pea root plastids. To analyze A22 gene expression, RNA isolated from Arabidopsis roots and leaves was probed with a ³²P-labeled fragment of an A22 clone, corresponding to the 5'-untranslated region and to the part coding for the transit peptide. This segment of the sequence was chosen to avoid false signals with mitochondrial and microbody MDH transcripts. Obviously A22 mRNA is not very abundant in both these tissues, because we failed to obtain distinct signals even when 80 µg of total RNA were applied per lane (not shown). For this reason the experiment was repeated with the poly(A)⁺ RNA fraction. Fig. 3 shows that A22 is equally transcribed in leaves and in roots, indicating that the same plastidic NAD-MDH is present both in chloroplasts and in root plastids.

View larger version (72K): [in this window] [in a new window]   Fig. 3.   Analysis of A22 mRNA levels in leaf and root tissues of Arabidopsis. Poly(A)+-RNA (7 µg) prepared from leaves and roots of Arabidopsis was subjected to Northern analysis (b) as described under "Experimental Procedures." After hybridization with 32P-labeled DNA corresponding to the 5'-untranslated region and to the transit-peptide-coding region of A22, the blot was washed three times for 30 min with 0.5 × SSC at 65 °C. Standard filters with dotted serial dilutions of the clones A22 and A33 were developed with the blot for control of unspecific hybridization (a).

Sequence Analysis of Arabidopsis cDNA Clones Coding for NAD-dependent Malate Dehydrogenases from Plastids, Microbodies, and Mitochondria-- The complete nucleotide sequence of a A22 cDNA clone encoding plastidic NAD-MDH is shown in Fig. 4. It contains a 1209-bp-long open reading frame, the 48-bp-long 5'-flanking region, and the 160-bp-long 3'-flanking region. The open reading frame corresponds to a polypeptide of 403 amino acids with predicted molecular mass of 42.4 kDa. All NAD-MDHs sequenced so far, start 6-13 amino acid residues prior to the dinucleotide binding motif GAAGGIG. Obviously this also holds true for the plastidic isoenzyme encoded by A22, because the labeled precursor protein was shortened to approximately 34 kDa upon import into the chloroplast stroma. The cleavage site for the stromal processing peptidase follows for most chloroplast proteins a semi-conserved motif (I/V)X(A/C)A (44). Indeed, the start of the NADP-dependent chloroplast MDH isoform, (I/V)XCS, fits almost exactly the proposed consensus. The most likely motif found in A22 among 20 amino acids upstream of the NAD(H)-binding site is INAS. Consequently, the mature protein was supposed to begin with Ser⁸².

View larger version (60K): [in this window] [in a new window]   Fig. 4.   Nucleotide and deduced amino acid sequences of Arabidopsis plastidic NAD-MDH (A22). The 5'- and 3'-untranslated regions are shown in italics. The C-terminal extension of the protein sequence not found in mitochondrial (A20) and microbody (A33) NAD-MDHs, as well as the putative transit peptide are underlined. Amino acids conserved in all three organellar isoenzymes are shaded. The sequence was submitted to the EMBL Nucleotide Sequence Data Base and assigned the accession number Y13987.

Expression of A22 in E. coli: Analysis of Arabidopsis NAD-MDH Isoforms by Isoelectric Focusing-- For the heterologous expression of pNAD-MDH in E. coli, the pET-A22 plasmid was constructed using a pET-21a⁺ cloning vector as described under "Experimental Procedures." 3 h after the E. coli cells carrying pET-A22 were induced with isopropyl--D-thiogalactopyranoside, a crude bacterial extract demonstrated NAD-MDH activity of 430 units/mg protein.

View larger version (75K): [in this window] [in a new window]   Fig. 5.   Comparison of NAD-MDH isoenzymes from Arabidopsis and of recombinant pNAD-MDH by isoelectric focusing. Arabidopsis mitochondria (lane 1), chloroplasts (lane 2), and crude leaf extract (lane 4) as well as crude extract of E. coli cells over-expressing recombinant plastidic NAD-MDH (lane 3) were analyzed by isoelectric focusing as described under "Experimental Procedures." NAD-MDH isoenzymes were detected by specific activity staining and attributed to their putative compartments, cytosol (cy), mitochondria (m), plastids (p), and microbodies (mb). Isoelectric points of pI standard proteins are shown on the left.

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

The cDNA clone A22 was extracted from the Arabidopsis EST data base during the search for the novel plastidic malate dehydrogenase isoform. The data base contains information about the first 350-400 bp of anonymous cDNA clones from a library composed of mRNA from different plant tissues. The following lines of evidence form a proof that the A22 clone encodes the plastidic enzyme isoform rather than one of the mitochondrial or peroxisomal enzyme isoforms: First, length and amino acid content of the predicted presequence (residues 2-81), including the serine-to-arginine ratio, point to a plastidic localization of the enzyme. For higher plant proteins this ratio has been shown to be able to discriminate between 90% of mitochondrial and chloroplast targeting peptides (43). Second, in vitro synthesized precursor protein was imported into the stroma of isolated spinach chloroplasts. Targeting experiments provide a reliable method for addressing a problem of intracellular localization of proteins. Where higher plants are concerned, mistargeting of plant mitochondrial proteins into the chloroplasts has never been, to our knowledge, reported in the literature. Third, the recombinant protein exhibits the same isoelectric point of 5.35 as NAD-MDH from stroma of highly pure Arabidopsis chloroplasts. In accordance with the isoenzyme distribution pattern in other plant species (14, 16, 47), NAD-MDH from Arabidopsis mitochondria and microbodies are characterized by a more acidic and more basic pI, respectively.

Plant mitochondria and microbodies are known to possess multiple molecular isoforms of NAD-MDH. For example, three genes located on three different chromosomes code for mNAD-MDH in corn (15). At least three slightly different cDNA clones encoding mNAD-MDH are found in the Arabidopsis EST data base; their exact number cannot be determined without complete sequencing of the corresponding clones. Even very distantly related eukaryotic NAD-MDHs are able to form catalytically active mixed dimers (48). When found in the same cell compartment, three different monomers should form three homodimers and three mixed dimers, all of them possibly having disparate isoelectric points. This can explain why NAD-MDH activity in isolated mitochondria from Arabidopsis was not resolved into distinct bands. Mitochondrial MDH from mammalian sources was shown to form complexes with aspartate aminotransferase, fumarase, and citrate synthase and to be attached to the mitochondrial membrane either directly or with the help of unidentified binding proteins (49, 50). Supposing that such multienzyme complexes occur also in plant mitochondria, it could be another reason for the observed behavior of mNAD-MDH upon IEF. In contrast to the situation with mitochondria, NAD-MDH activity in isolated chloroplasts from Arabidopsis and spinach focuses as one sharp band. Most probably, this organelle possesses only the molecular isoform of NAD-MDH described in this work. The results of Northern analysis indicate that at least in Arabidopsis the same or a very similar gene is transcribed in the roots as well.

NAD-MDHs found in three different organelles of plant cell: mitochondria, microbodies, and plastids, arose clearly via triplication of a pre-existing mitochondrial gene because they demonstrate high amino acid similarity of about 80%. The NADP-dependent, redox-regulated chloroplast MDH arose from the pre-existing nuclear gene of the eukaryotic host during the evolution of the plant cell (40). Therefore, the example of both plastidic MDHs contradicts the widespread opinion that nuclear-encoded chloroplast proteins generically originate from the cyanobacterial ancestors and is consistent with the findings of mosaic origin of higher plant Calvin cycle enzymes (51).

Until now the chloroplast malate valve was defined as a mechanism to balance the NADPH/ATP ratio in illuminated chloroplasts. NADPH and ATP are generated in the light reactions of photosynthesis by coupled electron transport. Light-activated NADP-MDH transfers excessive electrons from NADPH to OAA to form malate, which is subsequently transported into the cytosol. In all plastid types but the illuminated chloroplasts, the production of ATP and NADPH is not coupled. NADPH is produced from hexoses in the oxidative pentose-phosphate pathway and is required in numerous reactions of dark metabolism such as nitrite assimilation (52), amino acid synthesis, etc. In chloroplasts the key enzyme of the oxidative pentose-phosphate pathway, glucose-6-phosphate dehydrogenase, is inactive under light conditions and is activated in the dark via reversible oxidation (38, 53).

In dark plastids ATP can be imported from the cytosol via an ATP/ADP translocator (54). However, under physiological conditions the calculated rates of ATP import (8) are not sufficient to account even for the observed rates of starch degradation alone (55). Therefore, ATP must be generated inside the plastids from dihydroxyacetonephosphate by the coupled action of GAPDH and phosphoglycerate kinase (5, 56). Chloroplast NAD(P)-GAPDH is redox-regulated. The 150-kDa light form of it uses light-generated NADPH to reduce 1,3-bis-phosphoglycerate. In contrast, the oxidized 600-kDa form of the enzyme, which occurs in the dark, prefers NAD as a cosubstrate (3). In heterotrophic plastids, such as red pepper chromoplasts, NAD(P)-GAPDH is replaced by a strictly NAD-dependent isoform, which was recently isolated and cloned.² Therefore, the production of ATP in all plastid types, with the exception of illuminated chloroplasts, is coupled to the generation of reducing equivalents in the form of NADH. Consequently, NAD-linked MDH is required to operate the malate valve under these conditions.

As a matter of fact, there is no reason to suppose that plastidic NAD-MDH does not participate in the metabolism of illuminated chloroplasts as well. Measurements of NAD-MDH activity in isolated chloroplasts and in partially purified preparations of the recombinant enzyme (not shown) indicate that the enzyme is not subjected to reductive inactivation in the light as is the plastidic isoform of glucose-6-phosphate dehydrogenase. In this connection, other NAD-dependent processes in the plastids must be considered, among them fatty acid biosynthesis. It is well known that these organelles are the main sites of de novo fatty acid synthesis in plants (57). Because the fatty acid precursor, acetyl-CoA, cannot cross the plastid membrane, it must be generated inside the organelle (58). In chloroplasts of some plants, acetyl-CoA is produced from pyruvate by an NAD-dependent pyruvate dehydrogenase complex rather than by an acetyl-CoA synthetase reaction. The pNAD-MDH can play a role in regeneration of NAD required for the reaction of plastidic pyruvate dehydrogenase complex. Malate formed by pNAD-MDH can enter the pathway as a substrate after it is decarboxylated to pyruvate by NADP-linked malic enzyme. In leucoplasts from developing castor bean endosperm, malate is supposed to be an in vivo precursor for fatty acid synthesis (59). At any rate the action of both enzymes must be closely coordinated, because pyruvate dehydrogenase complex is strongly inhibited by the increase in the NADH/NAD ratio, and the pNAD-MDH is the enzyme maintaining this ratio.

In summary, the novel plastidic MDH isoform may play an important role in different aspects of plastid metabolism. Further characterization of pNAD-MDH at the level of biochemical and transcriptional regulation as well as the analysis of transgenic plants will help to understand better the physiological role of the new enzyme.