Steric hindrance controls pyridine nucleotide specificity of a flavin‐dependent NADH:quinone oxidoreductase

The crystal structure of the NADH:quinone oxidoreductase PA1024 has been solved in complex with NAD⁺ to 2.2 Å resolution. The nicotinamide C4 is 3.6 Å from the FMN N5 atom, with a suitable orientation for facile hydride transfer. NAD⁺ binds in a folded conformation at the interface of the TIM‐barrel domain and the extended domain of the enzyme. Comparison of the enzyme‐NAD⁺ structure with that of the ligand‐free enzyme revealed a different conformation of a short loop (75–86) that is part of the NAD⁺‐binding pocket. P78, P82, and P84 provide internal rigidity to the loop, whereas Q80 serves as an active site latch that secures the NAD⁺ within the binding pocket. An interrupted helix consisting of two α‐helices connected by a small three‐residue loop binds the pyrophosphate moiety of NAD⁺. The adenine moiety of NAD⁺ appears to π–π stack with Y261. Steric constraints between the adenosine ribose of NAD⁺, P78, and Q80, control the strict specificity of the enzyme for NADH. Charged residues do not play a role in the specificity of PA1024 for the NADH substrate.     

Structure of an antibiotic-synthesizing UDP-glucuronate 4-epimerase MoeE5 in complex with substrate

The epimerase MoeE5 from Streptomyces viridosporus converts UDP-glucuronic acid (UDP-GlcA) to UDP-galacturonic acid (UDP-GalA) to provide the first sugar in synthesizing moenomycin, a potent inhibitor against bacterial peptidoglycan glycosyltransferases. The enzyme belongs to the UDP-hexose 4-epimerase family, and uses NAD⁺ as its cofactor. Here we present the complex crystal structures of MoeE5/NAD⁺/UDP-GlcA and MoeE5/NAD⁺/UDP-glucose, determined at 1.48 Å and 1.66 Å resolution. The cofactor NAD⁺ is bound to the N-terminal Rossmann-fold domain and the substrate is bound to the smaller C-terminal domain. In both crystals the C4 atom of the sugar moiety of the substrate is in close proximity to the C4 atom of the nicotinamide of NAD⁺, and the O4 atom of the sugar is also hydrogen bonded to the side chain of Tyr154, suggesting a productive binding mode. As the first complex structure of this protein family with a bound UDP-GlcA in the active site, it shows an extensive hydrogen-bond network between the enzyme and the substrate. We further built a model with the product UDP-GalA, and found that the unique Arg192 of MoeE5 might play an important role in the catalytic pathway. Consequently, MoeE5 is likely a specific epimerase for UDP-GlcA to UDP-GalA conversion, rather than a promiscuous enzyme as some other family members.     

Structural and Biochemical Study of the Mono-ADP-Ribosyltransferase Domain of SdeA,a Ubiquitylating/Deubiquitylating Enzyme from Legionella pneumophila

Conventional ubiquitylation occurs through an ATP-dependent three-enzyme cascade (E1, E2, and E3) that mediates the covalent conjugation of the C-terminus of ubiquitin to a lysine on the substrate. SdeA, which belongs to the SidE effector family of Legionella pneumophila, can transfer ubiquitin to endoplasmic reticulum-associated Rab-family GTPases in a manner independent of E1 and E2 enzymes. The novel ubiquitin-modifying enzyme SdeA utilizes NAD⁺ as a cofactor to attach ubiquitin to a serine residue of the substrate. Here, to elucidate the coupled enzymatic reaction of NAD + hydrolysis and ADP-ribosylation of ubiquitin in SdeA, we characterized the mono-ADP-ribosyltransferase domain of SdeA and show that it consists of two sub-domains termed mART-N and mART-C. The crystal structure of the mART-C domain of SdeA was also determined in free form and in complex with NAD⁺ at high resolution. Furthermore, the spatial orientations of the N-terminal deubiquitylase, phosphodiesterase, mono-ADP-ribosyltransferase, and C-terminal coiled-coil domains within the 180-kDa full-length SdeA were determined. These results provide insight into the unusual ubiquitylation mechanism of SdeA and expand our knowledge on the structure–function of mono-ADP-ribosyltransferases.     

Molecular modeling study for conformational changes of Sirtuin 2 due to substrate and inhibitor binding

Sirtuin is a member of NAD⁺-dependent deacetylase family. The structural details of Sirtuin 2 (SIRT2) complex will be very useful to discover the drug which might have beneficial effects on various diseases like cancer, diabetes, etc. Unfortunately, SIRT2 complex structure is not available yet, hence molecular docking was carried out to dock the substrate (NAD⁺ and acetylated lysine) and inhibitor (sirtinol) in the NAD⁺ binding site. The suitable binding orientation of substrate and inhibitor in the SIRT2 active site was selected and subjected to 5?ns molecular dynamics simulations to adjust the binding orientation of inhibitor and substrate as well as to identify the conformational changes in the active site. The result provides an insight about 3D SIRT2 structural details as well as the importance of F96 in deacetylation function. In addition, our simulations revealed the displacement of F96 upon substrate and inhibitor binding, inducing an extended conformation of loop3 and changing its interactions with the rest of SIRT2. We believe that our study could be helpful to gain a structural insight of SIRT2 and to design the receptor-based inhibitors.     

Biological channeling of a reactive intermediate in the bifunctional enzyme DmpFG

It has been hypothesized that the bifunctional enzyme DmpFG channels its intermediate, acetaldehyde, from one active site to the next using a buried intermolecular channel identified in the crystal structure. This channel appears to switch between an open and a closed conformation depending on whether the coenzyme NAD⁺ is present or absent. Here, we applied molecular dynamics and metadynamics to investigate channeling within DmpFG in both the presence and absence of NAD⁺. We found that substrate channeling within this enzyme is energetically feasible in the presence of NAD⁺ but was less likely in its absence. Tyr-291, a proposed control point at the channel's entry, does not appear to function as a molecular gate. Instead, it is thought to orientate the substrate 4-hydroxy-2-ketovalerate in DmpG before reaction occurs, and may function as a proton shuttle for the DmpG reaction. Three hydrophobic residues at the channel's exit appear to have an important role in controlling the entry of acetaldehyde into the DmpF active site.     

Biochemical characterisation of recombinant Streptomyces pristinaespiralis L-lysine cyclodeaminase

L-Lysine cyclodeaminase from Streptomyces pristinaespiralis was heterologously expressed in Escherichia coli, isolated to 90% purity after two purification steps and characterised. The size of the isolated recombinant enzyme was in agreement with the theoretical size calculated from the corresponding gene. We demonstrated that our preparation converts L-lysine to L-pipecolic acid (enantiomeric excess >95%) after isolating and identifying the conversion product by LC/MS, NMR and IR. This conversion followed Michaelis-Menten kinetics with a K(m) of 1.39+/-0.32 mM. The enzyme activity was maximal at pH 6.7. Reducing conditions, the presence of glycerol and in particular the presence of iron(II) significantly enhanced the L-lysine cyclodeaminase activity. Although the heat stability of the enzyme diminished significantly after 37 degrees C, the initial rate of reaction was maximal at 61 degrees C. We found no requirement for an external cofactor for full activity, although sequence data indicate NAD+ as cofactor. Upon enzyme denaturation, NAD+ release was observed, which indicates very tight binding of NAD+ to the enzyme. In parallel we developed selection and screening assays for lysine cyclodeaminase, which we adapted to microtitre plate format and validated. Among twenty-eight lysine analogues screened for turnover/binding to the enzyme, three were identified as substrates (L-ornithine, 5-hydroxylysine and L-4-thialysine), while another six (4-azalysine, L-2,4-diaminobutyric acid, 1,5-diaminopentane, N-epsilon-trifluoroacetyl-L-lysine, N-epsilon-Boc-L-lysine and N-epsilon-methyl-L-lysine) were shown to compete against L-lysine turnover without being converted by the enzyme. All substrates displayed Michaelis-Menten kinetics upon turnover by lysine cyclodeaminase. Our results indicate that the lysine cyclodeaminase from Streptomyces pristinaespiralis is a highly enantioselective enzyme at the substrate recognition and conversion levels, in both cases in favour of the l-isomer.     

On the mechanism of action of calf spleen NAD+ glycohydrolase

The hydrolysis of NAD⁺ analogs bearing different substituents in the pyridinium moiety, catalyzed by solubilized calf spleen NAD⁺ glycohydrolase, was studied. The enzyme was specific for analogs possessing a carbonyl function at position C-3. The observed maximal rates showed no simple dependence either on the leaving ability of the parent pyridines or on the observed binding energies. The analogs without carbonyl substituents were not found to be hydrolyzed under our experimental conditions; and, compared to the substrates, they all presented much higher affinities for the active site, which could not be related to specific interactions. These results indicate that the rate-limiting step of the NAD⁺ hydrolysis, which is the formation of an enzyme-ADP-ribosyl intermediate, is probably more complex than a simple chemical step, i.e., pyridinium-ribose bond breakage. At a molecular level we favor a catalytic mechanism which, through nonbonded interactions between the substrate and the active site of the enzyme, results in the destabilization of the pyridinium-ribose bond, i.e., unimolecular decomposition involving an oxocarbonium ion intermediate.     

Mapping the enzymatic active site of Pseudomonas aeruginosa exotoxin A

Pseudomonas aeruginosa exotoxin A is representative of a class of enzymes, the monoADP-ribosyl, which catalyze the covalent transfer of an ADP-ribose moiety of NAD⁺ to a target substrate. Availability of the three-dimensional structure of exotoxin A provides the opportunity for mapping substrate binding sites and suggesting which amino acid residues may be involved in catalysis. Data from several sources have been combined to develop a proposal for the NAD⁺ binding site of exotoxin A: the binding of NAD⁺ fragments adenosine, AMP, and ADP have been delineated crystallographically to 6.0, 6.0, and 2.7 Å, respectively; significant sequence homology spanning 60 residues has been found between exotoxin A and diphtheria toxin, which has the identical enzymatic activity; iodination of exotoxin A, under conditions in which only tyrosine 481 is iodinated in the enzymatic domain, abolishes ADP-ribosyl transferase activity.     

Glutamate dehydrogenase of lupin nodules: Purification and properties

Glutamate dehydrogenase, GDH (l-glutamate: NAD⁺ oxidoreductase (deaminating) EC 1.4.1.2) was purified from the plant fraction of lupin nodules and the purity of the preparation established by gel electrophoresis and electrofocusing. The purified enzyme existed as 4 charge isozymes with a MW of 270000. The subunit MW, as determined by dodecyl sulphate electrophoresis, was 45 000. On the basis of the results of the MW determinations a hexameric structure is proposed for lupin-nodule GDH. The pH optima for the enzyme were pH 8.2 for the amination reaction and pH 8.8 for the deamination reaction. GDH from lupin nodules showed a marked preference for NADH over NADPH in the amination reaction and used only NAD⁺ for the deamination reaction. Pyridoxal-5′-P and EDTA inhibited activity. The enzyme displayed Michaelis-Menten kinetics with respect to all substrates except NAD⁺. When NAD⁺ was the varied substrate, there was a deviation from Michaelis-Menten behaviour towards higher activity at high concentrations of NAD⁺.     

Detection and properties of l-histidinol dehydrogenase in wheat germ

The presence and partial characterization of the properties of l-histidinol dehydrogenase (EC 1.1.1.23), the enzyme catalysing the last step in the pathway of histidine biosynthesis, has been described in higher plants for the first time. The activity has been found in cell-free extracts from wheat germ, turnip root, radish root and squash fruit. The enzyme has been partially purified and characterized from extracts of acetone powders of wheat germ. DEAE-cellulose chromatography revealed two peaks of histidinol dehydrogenase activity. In one there was a rapid reduction of NAD⁺ in the absence of histidinol; however, the rate was stimulated by the addition of histidinol. The rate in the absence of substrate became quite low after several min and the histidinol-dependent rate was then easily observed. The second peak of activity did not reduce NAD⁺ unless l-histidinol was present in the assay mixture. The K^ms for l-histidinol and NAD⁺ were determined for this latter enzyme. The values obtained at saturating concentrations of the other substrate were l-histidinol, 8.8 μM and NAD⁺, 0.14 mM. The product of the dehydrogenase reaction was histidine as determined by paper chromatography.     

Structural and biochemical characterization of a novel aldehyde dehydrogenase encoded by the benzoate oxidation pathway in Burkholderia xenovorans LB400

The recently identified benzoate oxidation (box) pathway in Burkholderia xenovorans LB400 (LB400 hereinafter) assimilates benzoate through a unique mechanism where each intermediate is processed as a coenzyme A (CoA) thioester. A key step in this process is the conversion of 3,4-dehydroadipyl-CoA semialdehyde into its corresponding CoA acid by a novel aldehyde dehydrogenase (ALDH) (EC 1.2.1.x). The goal of this study is to characterize the biochemical and structural properties of the chromosomally encoded form of this new class of ALDHs from LB400 (ALDH_C) in order to better understand its role in benzoate degradation. To this end, we carried out kinetic studies with six structurally diverse aldehydes and nicotinamide adenine dinucleotide (phosphate) (NAD ⁺ and NADP ⁺). Our data definitively show that ALDH_C is more active in the presence of NADP ⁺ and selective for linear medium-chain to long-chain aldehydes. To elucidate the structural basis for these biochemical observations, we solved the 1.6-Å crystal structure of ALDH_C in complex with NADPH bound in the cofactor-binding pocket and an ordered fragment of a polyethylene glycol molecule bound in the substrate tunnel. These data show that cofactor selectivity is governed by a complex network of hydrogen bonds between the oxygen atoms of the 2′-phosphoryl moiety of NADP ⁺ and a threonine/lysine pair on ALDH_C. The catalytic preference of ALDH_C for linear longer-chain substrates is mediated by a deep narrow configuration of the substrate tunnel. Comparative analysis reveals that reorientation of an extended loop (Asn478-Pro490) in ALDH_C induces the constricted structure of the substrate tunnel, with the side chain of Asn478 imposing steric restrictions on branched-chain and aromatic aldehydes. Furthermore, a key glycine (Gly104) positioned at the mouth of the tunnel allows for maximum tunnel depth required to bind medium-chain to long-chain aldehydes. This study provides the first integrated biochemical and structural characterization of a box-pathway-encoded ALDH from any organism and offers insight into the catalytic role of ALDH_C in benzoate degradation.     

Comparison of the Kinetic Behavior toward Pyridine Nucleotides of NAD-Linked Dehydrogenases from Plant Mitochondria

In this article we compare the kinetic behavior toward pyridine nucleotides (NAD⁺, NADH) of NAD⁺-malic enzyme, pyruvate dehydrogenase, isocitrate dehydrogenase, α-ketoglutarate dehydrogenase, and glycine decarboxylase extracted from pea (Pisum sativum) leaf and potato (Solanum tuberosum) tuber mitochondria. NADH competitively inhibited all the studied dehydrogenases when NAD⁺ was the varied substrate. However, the NAD⁺-linked malic enzyme exhibited the weakest affinity for NAD⁺ and the lowest sensitivity for NADH. It is suggested that NAD⁺-linked malic enzyme, when fully activated, is able to raise the matricial NADH level up to the required concentration to fully engage the rotenone-resistant internal NADH-dehydrogenase, whose affinity for NADH is weaker than complex I.     

Structure and activity of the NAD(P)+‐dependent succinate semialdehyde dehydrogenase YneI from Salmonella typhimurium

Aldehyde dehydrogenases are found in all organisms and play an important role in the metabolic conversion and detoxification of endogenous and exogenous aldehydes. Genomes of many organisms including Escherichia coli and Salmonella typhimurium encode two succinate semialdehyde dehydrogenases with low sequence similarity and different cofactor preference (YneI and GabD). Here, we present the crystal structure and biochemical characterization of the NAD(P)⁺‐dependent succinate semialdehyde dehydrogenase YneI from S. typhimurium. This enzyme shows high activity and affinity toward succinate semialdehyde and exhibits substrate inhibition at concentrations of SSA higher than 0.1 mM. YneI can use both NAD⁺ and NADP⁺ as cofactors, although affinity to NAD⁺ is 10 times higher. High resolution crystal structures of YneI were solved in a free state (1.85 Å) and in complex with NAD⁺ (1.90 Å) revealing a two domain protein with the active site located in the interdomain interface. The NAD⁺ molecule is bound in the long channel with its nicotinamide ring positioned close to the side chain of the catalytic Cys268. Site‐directed mutagenesis demonstrated that this residue, as well as the conserved Trp136, Glu365, and Asp426 are important for activity of YneI, and that the conserved Lys160 contributes to the enzyme preference to NAD⁺. Our work has provided further insight into the molecular mechanisms of substrate selectivity and activity of succinate semialdehyde dehydrogenases. © 2012 Wiley Periodicals, Inc.     

Heterogeneity of active sites in recombinant betaine aldehyde dehydrogenase is modulated by potassium

Betaine aldehyde dehydrogenase (BADH EC 1.2.1.8) catalyzes the irreversible oxidation of betaine aldehyde to glycine betaine using NAD⁺ as a coenzyme. Porcine kidney BADH (pkBADH) follows a bi‐bi ordered mechanism in which NAD⁺ binds to the enzyme before the aldehyde. Previous studies showed that NAD⁺ induces complex and unusual conformational changes on pkBADH and that potassium is required to maintain its quaternary structure. The aim of this work was to analyze the structural changes in pkBADH caused by NAD⁺ binding and the role played by potassium in those changes. The pkBADH cDNA was cloned and overexpressed in Escherichia coli, and the protein was purified by affinity chromatography using a chitin matrix. The pkBADH/NAD⁺ interaction was analyzed by circular dichroism (CD) and by isothermal titration calorimetry (ITC) by titrating the enzyme with NAD⁺. The cDNA has an open reading frame of 1485 bp and encodes a protein of 494 amino acids, with a predicted molecular mass of 53.9 kDa. CD data showed that the binding of NAD⁺ to the enzyme caused changes in its secondary structure, whereas the presence of K⁺ helps maintain its α‐helix content. K⁺ increased the thermal stability of the pkBADH‐NAD⁺ complex by 5.3°C. ITC data showed that NAD⁺ binding occurs with different association constants for each active site between 37.5 and 8.6 μM. All the results support previous data in which the enzyme incubation with NAD⁺ provoked changes in reactivity, which is an indication of slow conformational rearrangements of the active site.     

Purification and properties of the NAD+-xylitol-dehydrogenase from the yeast Pichia stipitis

Cell-free extracts of the xylose fermenting yeast Pichia stipitis exhibited xylitol dehydrogenase activity with NAD⁺ and NADP⁺. During the purification step on DEAE-sephadex A-50 a NAD⁺-dependent xylitol dehydrogenase could be separated from a NADP⁺-dependent. The NAD⁺-xylitol dehydrogenase was further purified to electrophoretic homogeneity via gel and affinity chromatography. The purified enzyme was most active at pH 9 and 35°C. Its molecular weight was determined to be 63,000 dalton by Sephadex G-200 column chromatography, and that of its subunit was 32,000 dalton by sodium dodecyl sulphate polyacrylamide gel electrophoresis. From the results of substrate specificity, the enzyme should be named l-iditol:NAD⁺-5-oxidoreductase (EC 1.1.1.14, sorbitol dehydrogenase).     

Interaction of sheep liver cytosolic aldehyde dehydrogenase with quercetin,resveratrol and diethylstilbestrol

The effects of quercetin and resveratrol (substances found in red wine) on the activity of cytosolic aldehyde dehydrogenase in vitro are compared with those of the synthetic hormone diethylstilbestrol. It is proposed that quercetin inhibits the enzyme by binding competitively in both the aldehyde substrate binding-pocket and the NAD⁺-binding site, whereas resveratrol and diethylstilbestrol can only bind in the aldehyde site. When inhibition is overcome by high aldehyde and NAD⁺ concentrations (1 mM of each), the modifiers enhance the activity of the enzyme; we hypothesise that this occurs through binding to the enzyme-NADH complex and consequent acceleration of the rate of dissociation of NADH. The proposed ability of quercetin to bind in both enzyme sites is supported by gel filtration experiments with and without NAD⁺, by studies of the esterase activity of the enzyme, and by modelling the quercetin molecule into the known three-dimensional structure of the enzyme. The possibility that interaction between aldehyde dehydrogenase and quercetin may be of physiological significance is discussed.     

Subunit interactions in liver alcohol dehydrogenase: A partial alkylation study.

The transient kinetics of aldehyde reduction by NADH catalyzed by liver alcohol dehydrogenase consist of two kinetic processes. This biphasic rate behavior is consistent with a model in which one of the two identical subunits in the enzyme is inactive during the reaction at the adjacent protomer. Alternatively, enzyme heterogeneity could result in such biphasic behavior. We have prepared liver alcohol dehydrogenase containing a single major isozyme; and the transient kinetics of this purified enzyme are biphasic.Addition of two [¹⁴C]carboxymethyl groups per dimer to the two “reactive” sulfhydryl groups (Cys46) yields enzyme which is catalytically inactive toward alcohol oxidation. Alkylated enzyme, as initially isolated by gel filtration chromatography at pH 7·5, forms an NAD⁺-pyrazole complex. However, the ability to bind NAD⁺-pyrazole is rapidly lost in pH 8·75 buffer; therefore, our alkylated preparations, as isolated by chromatography at pH 8·75, are inactive toward NAD⁺-pyrazole complex formation. We have prepared partially inactivated enzyme by allowing iodoacetic acid to react with liver alcohol dehydrogenase until 50% of the NAD⁺-pyrazole binding capacity remains; under these reaction conditions one [¹⁴C]carboxymethyl group is added per dimer. This partially alkylated enzyme preparation is isolated by gel filtration and has been aged sufficiently to lose NAD⁺-pyrazole binding ability at alkylated subunits. When solutions of native liver alcohol dehydrogenase and partially alkylated liver alcohol dehydrogenase containing the same number of unmodified active sites are allowed to react with substrate under single turnover conditions, partially alkylated enzyme is only half as reactive as native enzyme. This indicates that some molecular species in partially alkylated liver alcohol dehydrogenase that react with pyrazole and NAD⁺ during the active site titration do not react with substrate. These data are consistent with a model in which a subunit adjacent to an alkylated protomer in the dimeric enzyme is inactive toward substrate. In addition, NAD⁺-pyrazole binding at the protomers adjacent to alkylated subunits is slowly lost so that 75% of the enzyme-NAD⁺-pyrazole binding capacity is lost in 50% alkylated enzyme. These data supply strong evidence for subunit interactions in liver alcohol dehydrogenase.Binding experiments performed on partially alkylated liver alcohol dehydrogenase indicate that coenzyme binding is normal at a subunit adjacent to an alkylated protomer even though active ternary complexes cannot be formed. One hypothesis consistent with these results is the unavailability of zinc for substrate binding at the active site in subunits adjacent to alkylated protomers in monoalkylated dimer.     

The role of the invariant glutamate 95 in the catalytic site of Complex I from Escherichia coli

Replacement of glutamate 95 for glutamine in the NADH- and FMN-binding NuoF subunit of E. coli Complex I decreased NADH oxidation activity 2.5-4.8 times depending on the used electron acceptor. The apparent K_m for NADH was 5.2 and 10.4 μM for the mutant and wild type, respectively. Analysis of the inhibitory effect of NAD⁺ on activity showed that the E95Q mutation caused a 2.4-fold decrease of K_i^NAD+ in comparison to the wild type enzyme. ADP-ribose, which differs from NAD⁺ by the absence of the positively charged nicotinamide moiety, is also a competitive inhibitor of NADH binding. The mutation caused a 7.5-fold decrease of K_i^ADP-ribose relative to wild type enzyme. Based on these findings we propose that the negative charge of Glu95 accelerates turnover of Complex I by electrostatic interaction with the negatively charged phosphate groups of the substrate nucleotide during operation, which facilitates release of the product NAD⁺. The E95Q mutation was also found to cause a positive shift of the midpoint redox potential of the FMN, from − 350 mV to − 310 mV, which suggests that the negative charge of Glu95 is also involved in decreasing the midpoint potential of the primary electron acceptor of Complex I.     

NAD+ -dependent malic enzyme of Rhizobium meliloti is required for symbiotic nitrogen fixation

DEAE-cellulose chromatography of extracts of free-living Rhizobium meliloti cells revealed separate NAD⁺-dependent and NADP⁺-dependent malic enzyme activities. The NAD⁺ malic enzyme exhibited more activity with NAD⁺ as cofactor, but also showed some activity with NADP⁺. The NADP⁺ malic enzyme only showed activity when NADP⁺ was supplied as cofactor. Three independent transposon-induced mutants of R. meliloti which lacked NADP⁺ malic enzyme activity (dme) but retained NADP⁺ malic enzyme activity were isolated. In an otherwise wild-type background, the dme mutations did not alter the carbon utilization phenotype; however, nodules induced by these mutants failed to fix N₂. Structurally, these nodules appeared to develop like wild-type nodules up to the stage where N₂-fixation would normally begin. These results support the proposal that NAD⁺ malic enzyme, together with pyruvate dehydrogenase, functions in the generation of acetyl-CoA required for TCA cycle function in N₂-fixing bacteroids which metabolize C₄-dicarboxylic acids supplied by the plant.