In Saccharomyces cerevisiae, three homologous, differentially compartmentalized isozymes of malate dehydrogenase (MDH) catalyze the NAD(H)-dependent interconversion of malate and oxaloacetate: mitochondrial MDH1, cytosolic MDH2, and peroxisomal MDH3.  During growth of yeast on fermentable carbon sources all three MDH isozymes are subject to catabolite repression, a transcriptional mechanism for reducing levels of enzymes that interfere with glycolysis.  MDH2 is additionally subject to catabolite inactivation, a phenomenon involving phosphorylation and rapid degradation.

      Physical associations among related enzymes have been described for several metabolic pathways, and there is a large body of evidence supporting interactions among enzymes within the tricarboxylic acid (TCA) cycle.  The evidence is particularly compelling for interactions between mitochondrial malate dehydrogenase and citrate synthase in several mammalian systems.  We hypothesized the existence of distinct sets of enzyme interactions for all the members of the malate dehydrogenase isozyme family, and utilized the yeast two-hybrid system to look for these interactions among the yeast enzymes.

     Two-hybrid tests confirmed self-association of MDH1 and MDH2 which function in vivo as homodimers.  Yeast mitochondrial citrate synthase (CIT1) and aspartate aminotransferase (AAT1) were tested for interaction with MDH1, but we were unable to detect an interaction of MDH1 with either of these two enzymes by two-hybrid analysis.
For MDH2 we chose the gluconeogenic enzymes fructose – 1,6 – bisphosphatase (FBPase) and phosphoenolpyruvate carboxykinase (PEPCK) as candidates for interaction.  MDH2 supplies oxaloacetate for gluconeogenesis, and a yeast mutant lacking this enzyme is unable to grow with ethanol or acetate as a carbon source.  Two-hybrid tests detected an interaction of MDH2 with both gluconeogenic enzymes; however, no interaction was detectable between these enzymes and a truncated mutant form of MDH2 (DMDH2).  The DMDH2 mutant enzyme, which lacks 12 residues at the N-terminus, was previously shown to be resistant to degradation during catabolite inactivation.  We also discovered a new phenotype for strains expressing this mutant enzyme in lieu of the authentic enzyme, an inability to grow in minimal medium with ethanol or acetate as a carbon source.  Kinetic analyses of purified enzymes revealed no differences between the authentic and truncated forms of MDH2 that could account for this growth impairment.

      We confirmed the interactions observed in the two-hybrid system by equilibrium gel filtration and surface plasmon resonance analyses (SPR) using affinity tagged versions of the enzymes.  SPR data provided binding stoichiometries and equilibrium constants, and also confirmed reduced interactions between DMDH2 and PEPCK or FBPase relative to those observed for MDH2.  We also detected an interaction not observed by yeast two hybrid methods in this system, an association between the gluconeogenic enzymes PEPCK and FBPase.

      We report here the first evidence for a physical complex of gluconeogenic enzymes.  The metabolic importance of these interactions among gluconeogenic enzymes is evident by the observed growth defect of cells expressing a functional mutant form of MDH2 impaired in its association with enzymes of this complex.