Reversible transdominant inhibition of a metabolic pathway. In vivo evidence of interaction between two sequential tricarboxylic acid cycle enzymes in yeast.

The enzymes of the Krebs tricarboxylic acid cycle in mitochondria are proposed to form a supramolecular complex, in which there is channeling of intermediates between enzyme active sites. While interactions have been demonstrated in vitro between most of the sequential tricarboxylic acid cycle enzymes, no direct evidence has been obtained in vivo for such interactions. We have isolated, in the Saccharomyces cerevisiae gene encoding the tricarboxylic acid cycle enzyme citrate synthase Cit1p, an "assembly mutation," i.e. a mutation that causes a tricarboxylic acid cycle deficiency without affecting the citrate synthase activity. We have shown that a 15-amino acid peptide from wild type Cit1p encompassing the mutation point inhibits the tricarboxylic acid cycle in a dominant manner, and that the inhibitory phenotype is overcome by a co-overexpression of Mdh1p, the mitochondrial malate dehydrogenase. These data provide the first direct in vivo evidence of interaction between two sequential tricarboxylic acid cycle enzymes, Cit1p and Mdh1p, and indicate that the characterization of assembly mutations by the reversible transdominant inhibition method may be a powerful way to study multienzyme complexes in their physiological context.

Macromolecular compartmentation and channeling.

One of the accepted characterizations of the living state is that it is complex to an extraordinary degree. Since our current understanding of the living condition is minimal and fragmentary, it is not surprising that our first descriptions are simplistic. However, in certain areas of metabolism, especially those that have been amenable to experimentation for the longest period of time, the simplistic explanations have been the most difficult to revise. For example, current texts of general biochemistry still view metabolism as occurring by a series of independent enzymes dispersed in a uniform aqueous environment. This notion has been shown to be deeply flawed by both experimental and theoretical considerations. Thus, there is ample evidence that, in many metabolic pathways, specific interactions between sequential enzymes occur as static and/or dynamic complexes. In addition, reversible interactions of enzymes with structural proteins and membranes is a common occurrence. The interactions of enzymes give rise to a higher level of complexity that must be accounted for when one wishes to understand the regulation of metabolism. One of the phenomena that occurs because of sequential enzyme interactions is the process of channeling. This article discusses enzyme interactions and channeling and summarizes experimental and theoretical results from a few well-studied examples.

Metabolic effects of mislocalized mitochondrial and peroxisomal citrate synthases in yeast Saccharomyces cerevisiae.

Genes CIT1 and CIT2 from Saccharomyces cerevisiae encode mitochondrial and peroxisomal citrate synthases involved in the Krebs tricarboxylic acid (TCA) cycle and glyoxylate pathway, respectively. A Deltacit1 mutant does not grow on acetate, despite the presence of Cit2p that could, in principle, bypass the resulting block in the TCA cycle. To elucidate this absence of cross-complementation, we have examined the ability of Cit1p to function in the cytosol, and that of Cit2p to function in mitochondria. A cytosolically localized form of Cit1p was also incompetent for restoration of growth of a Deltacit1 strain on acetate, suggesting that mitochondrial localization of Cit1p is essential for its function in the TCA cycle. Cit2p was able, when mislocalized in mitochondria, to restore a wild-type phenotype in a strain lacking Cit1p. We have purified these two isoenzymes as well as mitochondrial malate dehydrogenase, Mdh1p, and have shown that Cit2p was also able to mimic Cit1p in its in vitro interaction with Mdh1p. Models of Cit1p and Cit2p structures generated on the basis of that of pig citrate synthase indicate very high structural and electrostatic surface potential similarities between the two yeast isozymes. Altogether, these data indicate that metabolic functions may require structural as well as catalytic roles for the enzymes.

Electrostatic channeling of oxaloacetate in a fusion protein of porcine citrate synthase and porcine mitochondrial malate dehydrogenase.

Mitochondrial malate dehydrogenase and citrate synthase are sequential enzymes in the Krebs tricarboxylic acid cycle. We have shown [Lindbladh, C., Rault, M., Hagglund, C., Small, W. C., Mosbach, K., Bülow, L., Evans, C., and Srere, P.A (1994) Biochemistry 33, 11692-11698] that a fusion protein of yeast mitochondrial citrate synthase and yeast mitochondrial malate dehydrogenase channels oxaloacetate between the active sites. A Brownian dynamics simulation model of porcine mitochondrial enzymes of citrate synthase and malate dehydrogenase was used [Elcock, A. H., and McCammon, A. M. (1996) Biochemistry 35, 12652-12658], showing that a positive electrostatic surface potential between the active sites of the fusion protein could account for the channeling of oxaloacetate we observed with the yeast fusion protein. Since the data were established with a yeast fusion protein and the model was with porcine fusion protein, we have now prepared and studied the porcine fusion protein. The channeling of the oxaloacetate intermediate was the same for the porcine fusion protein as it was for the yeast fusion protein. This channeling behavior is eliminated at high ionic strength. A fusion protein of porcine citrate synthase and porcine cytosolic malate dehydrogenase does not exhibit any channeling of oxaloacetate. A model of the fusion protein with the cytosolic malate dehydrogenase shows no clear positive electrostatic potential surface between the two active sites, thus distinguishing it from the fusion protein with the mitochondrial malate dehydrogenase. These results establish the electrostatic nature of channeling in mitochondrial fusion proteins.

Interaction between citrate synthase and malate dehydrogenase. Substrate channeling of oxaloacetate.

The interactions between pig heart citrate synthase and mitochondrial malate dehydrogenase or cytosolic malate dehydrogenase were studied using the frontal analysis method of gel filtration and by precipitation in polyethylene glycol. This method showed that an interaction between citrate synthase and mitochondrial malate dehydrogenase occurred but no interaction between citrate synthase and cytosolic malate dehydrogenase. Channeling of oxaloacetate in the malate dehydrogenase and citrate synthase-coupled systems was tested using polyethylene glycol precipitates of citrate synthase and mitochondrial malate dehydrogenase, and citrate synthase and cytosolic malate dehydrogenase. The effectiveness of large amounts of aspartate aminotransferase and oxaloacetate decarboxylase, as competing enzymes for the intermediate oxaloacetate, was examined. Aspartate aminotransferase and oxaloacetate decarboxylase were less effective competitors for oxaloacetate when precipitated citrate synthase and mitochondrial malate dehydrogenase in polyethylene glycol was used at low ionic strength compared with free enzymes in the absence of polyethylene glycol or with a co-precipitate of citrate synthase and cytosolic malate dehydrogenase. Substrate channeling of oxaloacetate with citrate synthase-mitochondrial malate dehydrogenase precipitate was inefficient at high ionic strength. These effects could be explained through electrostatic interactions of mitochondrial but not cytosolic malate dehydrogenase with citrate synthase.

Model of a quinary structure between Krebs TCA cycle enzymes: a model for the metabolon.

The enzymes which are responsible for catalyzing sequential reactions in several metabolic pathways have been proposed to be highly organized in supramolecular complexes termed metabolons. However, the in situ existence of these weak complexes is difficult to demonstrate because many of them are dissociated during isolation due to dilution effects. Consequently, the metabolon concept is subject to controversy. A model system consisting of genetically prepared bienzymatic fusion proteins has been used to immobilize sequential metabolic enzymes in close proximity and to demonstrate possible kinetic advantages of metabolons. These experiments use the sequential Krebs TCA cycle enzymes from yeast mitochondrial malate dehydrogenase (MDH), citrate synthase (CS), and aconitase (ACO). Using the porcine high-definition structures of these three enzymes, we have performed computer-modeling studies in order to understand how the molecules may interact. Among the thousands of docking orientations we have tried, one was found to respond to the structural and experimental constraints from the results obtained with the yeast fusion proteins. Interestingly, this quinary structure model shows substantial interacting surface areas with spatial and electrostatic complementarities which make the complex thermodynamically stable. This structure also contains an unbroken electrostatically favorable channel connecting the active sites of ACO and CS, as well as the one previously reported between CS and MDH active sites. Charged amino acids which could be involved in interactions stabilizing the complex have been identified. This model will be used as the basis for further experimental work on the structure of the Krebs TCA cycle metabolon.

Metabolic consequences of enzyme interactions.

The fact that enzyme complexes, stable, quasi-stable, and dynamic, exist in cells can no longer be ignored. Experimental evidence done with a variety of techniques has demonstrated these interactions in vitro and in vivo. There is scarcely a single known metabolic pathway in which no interactions of any of its enzymes exist (see reference 27 for a list of these). Such interactions are not only ubiquitous throughout metabolism, but they exist in all cell types, procaryote and eucaryote. In many of these systems the advantages of and regulatory power of enzyme-enzyme or enzyme-structural protein interactions has been amply demonstrated. The more difficult task is to assess accurately quantitative aspects of a system that varies between the solid, gel, and aqueous state. It is clear that the metabolic paradigm of soluble randomly dispersed metabolic elements in cells must be replaced, and new theoretical and experimental approaches introduced into this vital area of biological research.

Adenosine triphosphate citrate lyase mediates hypocitraturia in rats.

Chronic metabolic acidosis increases proximal tubular citrate uptake and metabolism. The present study addressed the effect of chronic metabolic acidosis on a cytosolic enzyme of citrate metabolism, ATP citrate lyase. Chronic metabolic acidosis caused hypocitraturia in rats and increased renal cortical ATP citrate lyase activity by 67% after 7 d. Renal cortical ATP citrate lyase protein abundance increased by 29% after 3 d and by 141% after 7 d of acid diet. No significant change in mRNA abundance could be detected. Hypokalemia, which causes only intracellular acidosis, caused hypocitraturia and increased renal cortical ATP citrate lyase activity by 28%. Conversely, the hypercitraturia of chronic alkali feeding was associated with no change in ATP citrate lyase activity. Inhibition of ATP citrate lyase with the competitive inhibitor, 4S-hydroxycitrate, significantly abated hypocitraturia and increased urinary citrate excretion fourfold in chronic metabolic acidosis and threefold in K+-depletion. In summary, the hypocitraturia of chronic metabolic acidosis is associated with an increase in ATP citrate lyase activity and protein abundance, and is partly reversed by inhibition of this enzyme. These results suggest an important role for ATP citrate lyase in proximal tubular citrate metabolism.

Active site mutants of pig citrate synthase: effects of mutations on the enzyme catalytic and structural properties.

We examined the catalytic efficiency of 18 pig citrate synthase mutants. The residues mutated were selected according to two criteria: the conservation of that residue in all known citrate synthase sequences, and the importance of that residue in substrate-amino acid interactions suggested by the extensive crystal structure information on the enzyme and its complexes. Several changes were made at certain residues to probe the effects of size, hydrogen bonding, and charge on the kinetics of the enzyme. The mutations, as expected, affected the kcats and Kms for OAA and acetyl-CoA to varying degrees. The catalytic efficiency of each of the mutants was determined by the kcat/Km for the individual substrates, OAA and acetyl-CoA. All mutations affected kcat. There was only one mutant, Asp327 Asn, in which the Kms primarily were affected. Most mutations affected both kcat and Km and included the following: His274Gly, His274Arg, Asp375Gly, Asp375Asn, Asp375Glu, Asp375Gln, His320Gly, His320Gln, His320Asn, His320Arg, Arg401His, Gly275Val, and Gly275Ala. The mutations, Arg401Gly, Arg401Lys, His235Gln, and Asn242Glu, had smaller effects on kcat and Km. The CS mutant Arg401Lys exhibited a modestly improved kcat/Km for both substrates compared to the nonmutant enzyme. X-ray crystallographic studies at 2.7 A resolution of one of the mutants, His274Gly, have been undertaken. The mutant enzyme crystallizes in an "open" conformation essentially isomorphous to wild type. The refined model has good geometry and a crystallographic R factor of 0.187 for 11 441 reflections observed between 6.0 and 2.7 A resolution. The refined model revealed a localized relaxation of the structure to relieve strain imposed by a high-energy main and side chain conformation of His274 in the nonmutant, but otherwise the mutation does not result in major structural alterations. Preliminary electrostatic calculations provide support for the concept that the transition state in the rate-limiting step of the citrate synthase catalyzed reaction may be an "enolized" version of acetyl-CoA that is neither neutral nor fully negatively charged and that a possible role for the catalytically essential His274 is to stabilize this by charge delocalization mediated by a hydrogen bond. These results provide the basis for further studies of the effects of these changes on the several reactive intermediates, activated substrates, and transition states which may occur along the reaction coordinate for this type of Claisen enzyme.

Catalytic strategy of citrate synthase: subunit interactions revealed as a consequence of a single amino acid change in the oxaloacetate binding site.

The active site of pig heart citrate synthase contains a histidine residue (H320) which interacts with the carbonyl oxygen of oxaloacetate and is implicated in substrate activation through carbonyl bond polarization, a major catalytic strategy of the enzyme. We report here the effects on the catalytic mechanism of changing this important residue to glycine. H320G shows modest impairment in substrate Michaelis constants [(7-16)-fold] and a large decrease in catalysis (600-fold). For the native enzyme, the chemical intermediate, citryl-CoA, is both hydrolyzed and converted back to reactants, oxaloacetate and acetyl-CoA. In the mutant, citryl-CoA is only hydrolyzed, indicating a major defect in the condensation reaction. As monitored by the carbonyl carbon's chemical shift, the extent of oxaloacetate carbonyl polarization is decreased in all binary and ternary complexes. As indicated by the lack of rapid H320G--oxaloacetate catalysis of the exchange of the methyl protons of acetyl-CoA or the pro-S-methylene proton of propionyl-CoA, the activation of acetyl-CoA is also faulty. Reflecting this defect in acetyl-CoA activation, the carboxyl chemical shift of H320G-bound carboxymethyl-CoA (a transition-state analog of the neutral enol intermediate) fails to decrease on formation of the H3020G-oxaloacetate-carboxymethyl-CoA ternary complex. Progress curves and steady-state data with H320G using citryl-CoA as substrate show unusual properties: substrate inhibition and accelerating progress curves. Either one of two models with subunit cooperativity [Monod, J., Wyman, J., & Changeux, J.-P. (1965) J. Mol. Biol. 12, 88; Koshland, D. E., Jr., Nemethy, G., & Filmer, D. (1966) Biochemistry 5, 365] quantitatively accounts for both the initial velocity data and the individual progress curves. The concentrations of all enzyme forms and complexes are assumed to rapidly reach their equilibrium values compared to the rate of substrate turnover. The native enzyme also behaves according to models for subunit cooperativity with citryl-CoA as substrate. However, the rates of formation/dissociation and reaction of complexes are kinetically significant. Comparisons of the values of kinetic constants between the native and mutants enzymes lead us to conclude that the mutant less readily undergoes a conformation change required for efficient activation of substrates.

Enzymatic and metabolic studies on retrograde regulation mutants of yeast.

Two nuclear genes, RTG1 and RTG2, which sense the functional state of yeast mitochondria, have been described recently. Yeast strains with null alleles of either of these two genes (delta rtg1, delta rtg2) cannot grow on acetate as the sole carbon source and are auxotrophic for glutamate and aspartate. We report here a series of metabolic experiments and enzyme activity measurements that were made in an attempt to determine the reason for the acetate- phenotype and the glutamate/aspartate auxotrophy. Decreases in the activities (approximately 50%) in mitochondrial citrate synthase (CS1), acetyl-CoA synthetase, NAD isocitrate dehydrogenase, and pyruvate carboxylase were noted. When CS1 was overexpressed in the delta rtg1 and delta rtg2 mutants, these strains could grow on acetate but were still auxotrophic for glutamate/aspartate. We propose that, in the mutant strain, CS1 activity becomes limiting for efficient acetate utilization, but that other complex metabolic interactions are affected, limiting production of intermediates that would allow synthesis of glutamic and aspartic acids.

Azotobacter vinelandii citrate synthase.

We have purified the citrate synthase from Azotobacter vinelandii and have determined that the size of the subunit is 48,000 Da and the structure of the holoenzyme is a hexamer. This contrasts with earlier estimates that indicate a 58,000 Da subunit and a tetrameric structure. In addition, the enzyme is allosteric with a Hill coefficient of 1.5 and is inhibited by NADH. The Hill coefficient is changed to about 1 by high ionic strength and AMP. The enzyme is thus similar to the citrate synthases of many other Gram-negative, facultative, anaerobic organisms. In addition, the amino acid sequence of about 100 residues has been determined and found to be highly similar to the sequence of Pseudomonas aeruginosa citrate synthase.

Cooperation between enzyme and transporter in the inner mitochondrial membrane of yeast. Requirement for mitochondrial citrate synthase for citrate and malate transport in Saccharomyces cerevisiae.

We have characterized 1,2,3-benzenetricarboxylic acid-sensitive, mersalyl-insensitive citrate uptake by mitochondria from two strains of Saccharomyces cerevisiae by describing the time course, Km and Vmax values, pH dependence, and response to inhibitors. In unloaded mitochondria from PSY142 CS1- cells, a mutant that lacks mitochondrial citrate synthase, both citrate uptake and efflux were reduced 7- and 8-fold, respectively, compared with the parental strain. No malate uptake was detectable in mitochondria from CS1- cells, while in the parental strain, uptake was 5.4 nmol/min/mg of protein. In contrast, mutations in peroxisomal citrate synthase (CS2-) or in other tricarboxylic acid cycle enzymes did not result in changes in mitochondrial citrate transport, suggesting a specific functional role for mitochondrial citrate synthase in citrate transport. More important, liposomes containing protein extracts from CS1- mitochondria showed the same citrate and malate transport rates as liposomes made from protein extracts of parental strain mitochondria. Thus, an apparently normal amount of both the citrate transporter and the dicarboxylate carrier is present in CS1- mitochondria, but both function abnormally in undisrupted mitochondria. We suggest that cooperation between the citrate transporter and mitochondrial citrate synthase is necessary for normal function of the transporter.

Metabolic studies on Saccharomyces cerevisiae containing fused citrate synthase/malate dehydrogenase.

We have constructed two different fusion proteins consisting of the C-terminal end of CS1 fused in-frame to the N-terminal end of MDH1 and HSA, respectively. The fusion proteins were expressed in mutants of Saccharomyces cerevisiae in which CS1 and MDH1 had been deleted and the phenotypes of the transformants characterized. The results show that the fusion proteins are transported into the mitochondria and that they restore the ability for the yeast mutants CS1-, MDH1-, and CS1-/MDH1- to grow on acetate. Determination of CS1 activity in isolated mitochondria showed a 10-fold increase for the strain that expressed native CS1, relative to the parental. In the transformant with CS1/MDH1 fusion protein, parental levels of CS1 were observed, while one-fifth this amount was observed for the strain expressing the CS1/HSA conjugate. Oxygen consumption studies on isolated mitochondria did not show any significant differences between parental-type yeast and the strains expressing the different fusion proteins or native CS1. [3(-13)C]Propionate was used to study the Krebs TCA cycle metabolism of yeast cells containing CS1/MDH1 fusion constructs. The 13C NMR study was performed in respiratory-competent parental yeast cells and using the genetically engineered yeast cells consisting of CS1- mutants expressing native CS1 and the fusion proteins CS1/MDH1 and CS1/HSA, respectively. [3(-13)C]Propionate is believed to be metabolized to [2(-13)C]succinyl-CoA before it enters the TCA cycle in the mitochondria. This metabolite is then oxidized through two symmetrical intermediates, succinate and fumarate, followed by conversion to malate, oxalacetate, and other metabolites such as alanine.(ABSTRACT TRUNCATED AT 250 WORDS)

Preparation and kinetic characterization of a fusion protein of yeast mitochondrial citrate synthase and malate dehydrogenase.

We have expressed the DNA of the fusion of CS1 to MDH1 in Escherichia coli gltA-. The fusion protein (CS1/MDH1) is the C-terminus of CS1 linked in-frame to the N-terminus of MDH1 with a short linker of glycyl-seryl-glycyl. The fusion protein produced was isolated and purified. Gel filtration studies indicated that CS1/MDH1 had a M(r) of approximately 170,000. Western blotting analysis with SDS gel indicated a M(r) of approximately 90,000-95,000 (theoretical M(r) = 87,000). This is the expected M(r) for the fusion protein subunit. The kinetics of CS1 and MDH1 activities of the fusion protein were compared to those of the free enzymes. In addition, the effect of AAT reaction, as a competitor for the intermediate OAA of the coupled MDH-CS reaction, was examined. It was observed that AAT was a less effective competitor for OAA when the CS1/MDH1 fusion protein is used than when the separate enzymes are employed. In addition, the transient time for the coupled reaction sequence was less for the fusion protein than for the free enzymes.