The transcarboxylase domain of pyruvate carboxylase is essential for assembly of the peroxisomal flavoenzyme alcohol oxidase

Pyruvate carboxylase (Pyc1p) has multiple functions in methylotrophic yeast species. Besides its function as an enzyme, Pyc1p is required for assembly of peroxisomal alcohol oxidase (AO). Hence, Pyc1p-deficient cells share aspartate auxotrophy (Asp-) with a defect in growth on methanol as sole carbon source (Mut-). To identify regions in Hansenula polymorpha Pyc1p that are required for the function of HpPyc1p in AO assembly, a series of random mutations was generated in the HpPYC1 gene by transposon mutagenesis. Upon introduction of 18 mutant genes into the H. polymorpha PYC1 deletion strain (pyc1), four different phenotypes were obtained, namely Asp- Mut-, Asp- Mut+, Asp+ Mut-, and Asp+ Mut+. One mutant showed an Asp+ Mut- phenotype. This mutant produced HpPyc1p containing a pentapeptide insertion in the region that links the conserved N-terminal biotin carboxylation domain (BC) with the central transcarboxylation (TC) domain. Three mutants that were Asp- Mut- contained insertions in the TC domain, suggesting that this domain is important for both functions of Pyc1p. Analysis of a series of constructed C-terminal and N-terminal truncated versions of HpPyc1p showed that the TC domain of Pyc1p, including the region linking this domain to the BC domain, is essential for AO assembly.     

Function of Escherichia coli biotin carboxylase requires catalytic activity of both subunits of the homodimer

Biotin carboxylase catalyzes the ATP-dependent carboxylation of biotin and is one component of the multienzyme complex acetyl-CoA carboxylase that catalyzes the first committed step in fatty acid synthesis. The Escherichia coli biotin carboxylase is readily isolated from the other components of the acetyl-CoA carboxylase complex such that enzymatic activity is retained. The three-dimensional structure of biotin carboxylase, determined by x-ray crystallography, demonstrated that the enzyme is a homodimer consisting of two active sites in which each subunit contains a complete active site. To understand how each subunit contributes to the overall function of biotin carboxylase, we made hybrid molecules in which one subunit had a wild-type active site, and the other subunit contained an active site mutation known to significantly affect the activity of the enzyme. One of the two genes encoded a poly-histidine tag at its N terminus, whereas the other gene had an N-terminal FLAG epitope tag. The two genes were assembled into a mini-operon that was induced to give high level expression of both enzymes. "Hybrid" dimers composed of one subunit with a wild-type active site and a second subunit having a mutant active site were obtained by sequential chromatographic steps on columns of immobilized nickel chelate and anti-FLAG affinity matrices. In vitro kinetic studies of biotin carboxylase dimers in which both subunits were wild type revealed that the presence of the N-terminal tags did not alter the activity of the enzyme. However, kinetic assays of hybrid dimer biotin carboxylase molecules in which one subunit had an active site mutation (R292A, N290A, K238Q, or E288K) and the other subunit had a wild-type active site resulted in 39-, 28-, 94-, and 285-fold decreases in the activity of these enzymes, respectively. The dominant negative effects of these mutant subunits were also detected in vivo by monitoring the rate of fatty acid biosynthesis by [(14)C]acetate labeling of cellular lipids. Expression of the mutant biotin carboxylase genes from an inducible arabinose promoter resulted in a significantly reduced rate of fatty acid synthesis relative to the same strain that expressed the wild type gene. Thus, both the in vitro and in vivo data indicate that both subunits of biotin carboxylase are required for activity and that the two subunits must be in communication during enzyme function.     

Novel insights into the biotin carboxylase domain reactions of pyruvate carboxylase from Rhizobium etli

The catalytic mechanism of the MgATP-dependent carboxylation of biotin in the biotin carboxylase domain of pyruvate carboxylase from R. etli (RePC) is common to the biotin-dependent carboxylases. The current site-directed mutagenesis study has clarified the catalytic functions of several residues proposed to be pivotal in MgATP-binding and cleavage (Glu218 and Lys245), HCO(3)(-) deprotonation (Glu305 and Arg301), and biotin enolization (Arg353). The E218A mutant was inactive for any reaction involving the BC domain and the E218Q mutant exhibited a 75-fold decrease in k(cat) for both pyruvate carboxylation and the full reverse reaction. The E305A mutant also showed a 75- and 80-fold decrease in k(cat) for both pyruvate carboxylation and the full reverse reaction, respectively. While Glu305 appears to be the active site base which deprotonates HCO(3)(-), Lys245, Glu218, and Arg301 are proposed to contribute to catalysis through substrate binding interactions. The reactions of the biotin carboxylase and carboxyl transferase domains were uncoupled in the R353M-catalyzed reactions, indicating that Arg353 may not only facilitate the formation of the biotin enolate but also assist in coordinating catalysis between the two spatially distinct active sites. The 2.5- and 4-fold increase in k(cat) for the full reverse reaction with the R353K and R353M mutants, respectively, suggests that mutation of Arg353 allows carboxybiotin increased access to the biotin carboxylase domain active site. The proposed chemical mechanism is initiated by the deprotonation of HCO(3)(-) by Glu305 and concurrent nucleophilic attack on the γ-phosphate of MgATP. The trianionic carboxyphosphate intermediate formed reversibly decomposes in the active site to CO(2) and PO(4)(3-). PO(4)(3-) then acts as the base to deprotonate the tethered biotin at the N(1)-position. Stabilized by interactions between the ureido oxygen and Arg353, the biotin-enolate reacts with CO(2) to give carboxybiotin. The formation of a distinct salt bridge between Arg353 and Glu248 is proposed to aid in partially precluding carboxybiotin from reentering the biotin carboxylase active site, thus preventing its premature decarboxylation prior to the binding of a carboxyl acceptor in the carboxyl transferase domain.     

Kinetic characterization of yeast pyruvate carboxylase isozyme Pyc1 and the Pyc1 mutant, C249A

The yeast Pyc1 isoform of pyruvate carboxylase has been further characterized and shown to differ from the Pyc2 isoform in its K(a) for K(+) activation. Pyc1 differs from chicken liver pyruvate carboxylase in the lack of effect of acetyl-CoA on ADP phosphorylation by carbamoyl phosphate, which may be a result of differences in the loci of action of the effector between the two enzymes. Solvent D(2)O isotope effects have been measured with Pyc1 on the full pyruvate carboxylation reaction, the ATPase reaction in the absence of pyruvate, and the carbamoyl phosphate-ADP phosphorylation reaction for the first time for pyruvate carboxylase. Proton inventories indicate that the measured isotope effects are due to a single proton transfer step in the reaction. The inverse isotope effects observed in all reactions suggest that the proton transfer step converts the enzyme from an inactive to an active form. Kinetic measurements on the C249A mutant enzyme suggest that C249 is involved in the binding and action of enzyme activators K(+) and acetyl-CoA. C249 is not involved in ATP binding as was observed for the corresponding residue in the biotin carboxylase subunit of Escherichia coli acetyl-CoA carboxylase, nor is it directly responsible for the measured inverse (D)(k(cat)/K(m)) isotope effects. The size of the inverse isotope effects indicates that they may result from formation of a low-barrier hydrogen bond. Modification of the wild type and C249A mutant with o-phthalaldehyde suggests that C249 is involved in isoindole formation but that the modification of this residue is not directly responsible for the accompanying major loss of enzyme activity.     

Allosteric regulation of the biotin-dependent enzyme pyruvate carboxylase by acetyl-CoA

The activity of the biotin-dependent enzyme pyruvate carboxylase from many organisms is highly regulated by the allosteric activator acetyl-CoA. A number of X-ray crystallographic structures of the native pyruvate carboxylase tetramer are now available for the enzyme from Rhizobium etli and Staphylococcus aureus. Although all of these structures show that intersubunit catalysis occurs, in the case of the R. etli enzyme, only two of the four subunits have the allosteric activator bound to them and are optimally configured for catalysis of the overall reaction. However, it is apparent that acetyl-CoA binding does not induce the observed asymmetrical tetramer conformation and it is likely that, under normal reaction conditions, all of the subunits have acetyl-CoA bound to them. Thus the activation of the enzyme by acetyl-CoA involves more subtle structural effects, one of which may be to facilitate the correct positioning of Arg353 and biotin in the biotin carboxylase domain active site, thereby promoting biotin carboxylation and, at the same time, preventing abortive decarboxylation of carboxybiotin. It is also apparent from the crystal structures that there are allosteric interactions induced by acetyl-CoA binding in the pair of subunits not optimally configured for catalysis of the overall reaction.     

Identification of the catalytic residues involved in the carboxyl transfer of pyruvate carboxylase

To clarify the mechanism of carboxyl transfer from carboxylbiotin to pyruvate, the following conserved amino acid residues present in the carboxyl transferase domain of Bacillus thermodenitrificans pyruvate carboxylase were converted to homologous amino acids: Asp543, Glu576, Glu592, Asp649, Lys712, Asp713, and Asp762. The carboxylase activity of the resulting mutants, D543E, E576D, E576Q, E592Q, D649N, K712R, K712Q, D713E, D713N, D762E, and D762N, was generally less than that of the wild type from mutation, but it decreased the most to 5% or even less than that of the wild type with D543E, D576Q, D649N, K712R, and K712Q. The decrease in activity observed for Asp543, Asp649, and Lys712 mutants was not for structural reasons because their structures seemed to remain intact as assessed by gel filtration and circular dichroism. On the basis of these data, a mechanism is proposed where Lys712 and Asp543 serve as the key acid and base catalyst, respectively.     

Kinetic characterization of yeast pyruvate carboxylase isozyme pyc1

Yeast (Saccharomyces cerevisiae) is unusual in being the only organism thus far identified as having two genes for pyruvate carboxylase. The expression of the two isozymes Pyc1 and Pyc2 appears to be differentially regulated, and since both are expressed cytoplasmically, this suggests that they have different properties. To the present, little has been done to characterize these isozymes, and almost all of the published kinetic information on yeast pyruvate carboxylase comes from measurements of enzyme prepared from bakers' yeast which is likely to be a mixture of both isozymes. Here we have measured basic kinetic parameters for Pyc1 and found that the K(a) of this isozyme for acetyl CoA is in the order of 8-10-fold higher than previously recorded, suggesting that Pyc1 and Pyc2 may be differentially regulated by this effector. Pyc1 is highly dependent on the presence of acetyl CoA for activity and in this respect is similar to chicken liver pyruvate carboxylase. However, unlike the chicken liver enzyme, the quaternary structure of the enzyme is quite stable in the absence of acetyl CoA, and the major locus of action of this effector appears to lie outside of the stimulation of the biotin carboxylation reaction.     

Biotinoyl domain of human acetyl-CoA carboxylase: Structural insights into the carboxyl transfer mechanism

Acetyl-CoA carboxylase (ACC) catalyzes the first step in fatty acid biosynthesis: the synthesis of malonyl-CoA from acetyl-CoA. As essential regulators of fatty acid biosynthesis and metabolism, ACCs are regarded as therapeutic targets for the treatment of metabolic diseases such as obesity. In ACC, the biotinoyl domain performs a critical function by transferring an activated carboxyl group from the biotin carboxylase domain to the carboxyl transferase domain, followed by carboxyl transfer to malonyl-CoA. Despite the intensive research on this enzyme, only the bacterial and yeast ACC structures are currently available. To explore the mechanism of ACC holoenzyme function, we determined the structure of the biotinoyl domain of human ACC2 and analyzed its characteristics and interaction with the biotin ligase, BirA using NMR spectroscopy. The 3D structure of the hACC2 biotinoyl domain has a similar folding topology to the earlier determined domains from E. coli and P. shermanii. However, the local structures near the biotinylation sites have notable differences that include the geometry of the consensus "Met-Lys-Met" (MKM) motif and the absence of "thumb" structure in the hACC2 biotinoyl domain. Observations of the NMR signals upon the biotinylation indicate that the biotin group of hACC2 does not affect the structure of the biotinoyl domain, while the biotin group for E. coli ACC interacts directly with the thumb residues that are not present in the hACC2 structure. These results imply that, in the E. coli ACC reaction, the biotin moiety carrying the carboxyl group from BC to CT can pause at the thumb of the BCCP domain. The human biotinoyl domain, however, lacks the thumb structure and does not have additional noncovalent interactions with the biotin moiety; thus, the flexible motion of the biotinylated lysine residue must underlie the "swinging arm" motion. The chemical shift perturbation and the cross saturation experiments of the human ACC2 holo-biotinoyl upon the addition of the biotin ligase (BirA) showed the interaction surface near the MKM motif, the two glutamic acids (Glu 926, Glu 953), and the positively charged residues (several lysine and arginine residues). This study provides insight into the mechanism of ACC holoenzyme function and supports the swinging arm model in human ACCs.     

Activation and inhibition of pyruvate carboxylase from Rhizobium etli

While crystallographic structures of the R. etli pyruvate carboxylase (PC) holoenzyme revealed the location and probable positioning of the essential activator, Mg(2+), and nonessential activator, acetyl-CoA, an understanding of how they affect catalysis remains unclear. The current steady-state kinetic investigation indicates that both acetyl-CoA and Mg(2+) assist in coupling the MgATP-dependent carboxylation of biotin in the biotin carboxylase (BC) domain with pyruvate carboxylation in the carboxyl transferase (CT) domain. Initial velocity plots of free Mg(2+) vs pyruvate were nonlinear at low concentrations of Mg(2+) and a nearly complete loss of coupling between the BC and CT domain reactions was observed in the absence of acetyl-CoA. Increasing concentrations of free Mg(2+) also resulted in a decrease in the K(a) for acetyl-CoA. Acetyl phosphate was determined to be a suitable phosphoryl donor for the catalytic phosphorylation of MgADP, while phosphonoacetate inhibited both the phosphorylation of MgADP by carbamoyl phosphate (K(i) = 0.026 mM) and pyruvate carboxylation (K(i) = 2.5 mM). In conjunction with crystal structures of T882A R. etli PC mutant cocrystallized with phosphonoacetate and MgADP, computational docking studies suggest that phosphonoacetate could coordinate to one of two Mg(2+) metal centers in the BC domain active site. Based on the pH profiles, inhibition studies, and initial velocity patterns, possible mechanisms for the activation, regulation, and coordination of catalysis between the two spatially distinct active sites in pyruvate carboxylase from R. etli by acetyl-CoA and Mg(2+) are described.     

Behavior of the ATP grasp domain of biotin carboxylase monomers and dimers studied using molecular dynamics simulations

The enzyme biotin carboxylase (BC) uses adenosine triphosphate (ATP) to carboxylate biotin and is involved in fatty acid synthesis. Structural evidence suggests that the B domain of BC undergoes a large hinge motion of ～45° when binding and releasing substrates. Escherichia coli BC can function as a natural homodimer and as a mutant monomer. Using molecular dynamics simulations, we evaluate the free energy profile along a closure angle of the B domain of E. coli BC for three cases: a monomer without bound Mg(2)ATP, a monomer with bound Mg(2)ATP, and a homodimer with bound Mg(2)ATP in one subunit. The simulation results show that a closed state is the most probable for the monomer with or without bound Mg(2)ATP. For the dimer with Mg(2)ATP in one of its subunits, communication between the two subunits was observed. Specifically, in the dimer, the opening of the subunit without Mg(2)ATP caused the other subunit to open, and hysteresis was observed upon reclosing it. The most stable state of the dimer is one in which the B domain of both subunits is closed; however, the open state for the B domain without Mg(2)ATP is only approximately 2k(B)T higher in free energy than the closed state. A simple diffusion model indicates that the mean times for opening and closing of the B domain in the monomer with and without Mg(2)ATP are much smaller than the overall reaction time, which is on the order of seconds.     

The biotin carboxylase-biotin carboxyl carrier protein complex of Escherichia coli acetyl-CoA carboxylase

Escherichia coli acetyl-CoA carboxylase (ACC) is composed of four different protein molecules. These proteins form a large but very unstable complex. Hints of a sub-complex between the biotin carboxylase (BC) and biotin carboxyl carrier protein (BCCP) subunits have been reported in the literature, but the complex was not isolated and thus the protein stoichiometry could not be determined. We report isolation of the BC.BCCP complex. By use of affinity chromatography using two different affinity tags it was shown that the complex consists of a two BCCP molecules per BC molecule. The molar ratio in the complex is the same as the ratio of the subunit proteins synthesized in vivo. We conclude that the complex consists of a dimer of BC plus four BCCP molecules instead of the 2BC.2BCCP complex previously assumed. This subunit ratio allows two conflicting models of the ACC mechanism to be rectified. We also report that the N-terminal 30 or so residues of BCCP are responsible for the interaction of BCCP with BC and that the BC.BCCP complex is a substrate for biotinylation in vitro.     

Differential regulation of the yeast isozymes of pyruvate carboxylase and the locus of action of acetyl CoA

Unlike other eukaryotes studied to date, yeast has two genes for pyruvate carboxylase coding for very similar, but not identical, isozymes (Pyc1 and Pyc2), both of which are located in the cytoplasm. We have found that there are marked differences in the kinetic properties of the isozymes potentially leading to differential regulation of Pyc1 and Pyc2 activity by both activators and substrates. For example, Pyc2 is only activated 3.7-fold by acetyl CoA, and 9.6-fold by NH(4)(+), whilst the figures for Pyc1 are 16 and 14.6-fold, respectively. Pyc1 and Pyc2 display different allosteric properties with respect to acetyl CoA activation and aspartate inhibition, with Pyc1 showing a higher degree of cooperativity than Pyc2, even in the absence of aspartate. We have investigated the locus of action in the amino acid sequence of the isozymes of this activator by measuring its regulation of various chimeric constructs of the two isozymes. In this way, we conclude that the main locus of action of acetyl CoA lies in the N-terminal half of the enzyme, within the biotin-carboxylation domain, between amino acids 99 and 478 of Pyc1.     

Biochemical evidence for diverse etiologies in biotin-responsive multiple carboxylase deficiency

Biotin-responsive multiple carboxylase deficiency can be categorized by clinical criteria into a neonatal-onset disorder and a distinct syndrome of infantile onset. Pedigrees in each instance are consistent with autosomal recessive inheritance. For a neonatal-onset proband, the sensitivity to relative biotin deprivation and the rapid clinical response to biotin supplementation are reflected by in vitro studies. Specific activities of biotin-dependent pyruvate carboxylase, propionyl CoA carboxylase, and 3-methylcrotonyl CoA carboxylase are 0.8 to 16% of mean control values after growth of fibroblasts in intermediate and very low biotin concentrations. Following relative biotin depletion, pyruvate carboxylase activity returns to normal after only 14 hr of growth in biotin-supplemented medium. In contrast, carboxylase activities in fibroblasts of an infantile-onset proband remain normal at very low biotin concentrations, even when avidin is added to the growth medium. The clinical heterogeneity, taken together with the distinct responses of cultured skin fibroblasts to biotin deprivation in vitro, probably reflect fundamentally different etiologies for the two categories of biotin-responsive multiple carboxylase deficiency.This work was supported by USPHS Grants GM28838 and AM25884.     

Modeling and numerical simulation of biotin carboxylase kinetics: implications for half-sites reactivity

Biotin carboxylase catalyzes the ATP-dependent carboxylation of biotin and is one component of the multienzyme complex acetyl-CoA carboxylase that catalyzes the first committed step in fatty acid synthesis in all organisms. In Escherichia coli, biotin carboxylase exists as a homodimer where each subunit contains a complete active site. In a previous study (Janiyani, K., Bordelon, T., Waldrop, G.L., Cronan Jr., J.E., 2001. J. Biol. Chem. 276, 29864-29870), hybrid dimers were constructed where one subunit was wild-type and the other contained an active site mutation that reduced activity at least 100-fold. The activity of the hybrid dimers was only slightly greater than the activity of the mutant homodimers and far less than the expected 50% activity for completely independent active sites. Thus, there is communication between the two subunits of biotin carboxylase. The dominant negative effect of the mutations on the wild-type active site was interpreted as alternating catalytic cycles of the active sites in the homodimer. In order to test the hypothesis of oscillating catalytic cycles, mathematical modeling and numerical simulations of the kinetics of wild-type, hybrid dimers, and mutant homodimers of biotin carboxylase were performed. Numerical simulations of biotin carboxylase kinetics were the most similar to the experimental data when an oscillating active site model was used. In contrast, alternative models where the active sites were independent did not agree with the experimental data. Thus, the numerical simulations of the proposed kinetic model support the hypothesis that the two active sites of biotin carboxylase alternate their catalytic cycles.     

The multisubunit acetyl-CoA carboxylase is strongly associated with the chloroplast envelope through non-ionic interactions to the carboxyltransferase subunits

The committed step for de novo fatty acid biosynthesis is the carboxylation of acetyl-CoA catalyzed by acetyl-CoA carboxylase (ACCase). Plastidial ACCase from most plants is a multisubunit complex composed of multiple copies of four different polypeptides, biotin carboxyl carrier protein (BCCP), biotin carboxylase (BC), and carboxyltransferase (alpha-CT and beta-CT). Immunoblot analyses revealed these four proteins were mostly (69% of total) associated with a 17,000 g insoluble fraction from lysed pea chloroplasts. Under the same conditions only 8% of ribulose-1,5-bisphosphate carboxylase was associated with this insoluble fraction. BCCP and biotin carboxylase BC subunits freely dissociated from 17 kg insoluble fractions under high ionic strength conditions, whereas alpha-CT and beta-CT subunits remained tightly associated. Both CT subunits were highly enriched in envelope versus stroma and thylakoid preparations whereas BC and BCCP subunits were predominantly stromal-localized due to partial dissociation. Rapid solubilization of intact chloroplasts with Triton X-100 followed by centrifugation at 30 kg resulted in a pellet that was up to 8-fold enriched in ACCase activity and 21-fold enriched in BC activity. Triton-insoluble 30 kg pellets were reduced in lipid and chlorophyll content but enriched in chloroplast DNA due to the isolation of nucleoid particles. However, ACCase was not directly associated with nucleoids since enzymatic digestion of DNA or RNA had no effect on the association with Triton-insoluble matter. The amount of Triton-insoluble ACCase was similar in chloroplasts isolated from dark- or light-adapted leaves suggesting transitory starch granules were also not involved in this association. It is proposed that ACCase is associated with envelope membranes through interactions with an unidentified integral membrane protein.     

Antisense expression and overexpression of biotin carboxylase in tobacco leaves.

The plastid acetyl-coenzyme A carboxylase (ACCase) catalyzes the first committed step of fatty acid synthesis and in most plants is present as a heteromeric complex of at least four different protein subunits: the biotin carboxylase (BC), the biotin carboxyl carrier protein, and the alpha and beta subunits of the carboxyltransferase. To gain insight into the subunit organization of this heteromeric enzyme complex and to further evaluate the role of ACCase in regulating fatty acid synthesis, BC expression was altered in transgenic plants. Tobacco (Nicotiana tabacum) was transformed with antisense-expression and overexpression tobacco BC constructs, which resulted in the generation of plants with BC levels ranging from 20 to 500% of wild-type levels. Tobacco plants containing elevated or moderate decreases in leaf BC were phenotypically indistinguishable from wild-type plants. However, plants with less than 25% of wild-type BC levels showed severely retarded growth when grown under low-light conditions and a 26% lower leaf fatty acid content than wild-type plants. A comparison of leaf BC and biotin carboxyl carrier protein levels in plants with elevated and decreased BC expression revealed that these two subunits of the plastid ACCase are not maintained in a strict stoichiometric ratio.     

By-product formation during exposure of respiring Saccharomyces cerevisiae cultures to excess glucose is not caused by a limited capacity of pyruvate carboxylase

Upon exposure to excess glucose, respiring cultures of Saccharomyces cerevisiae produce substantial amounts of ethanol and acetate. A possible role of a limited anaplerotic capacity in this process was investigated by overexpressing pyruvate carboxylase and by replacing it with a heterologous enzyme (Escherichia coli phosphoenolpyruvate carboxylase). Compared to the wild-type, neither the pyruvate carboxylase (Pyc)-overexpressing nor the transgenic strain exhibited reduced by-product formation after glucose pulses to aerobic glucose-limited chemostat cultures. An increased intracellular malate concentration was observed in the two engineered strains. It is concluded that by-product formation in S. cerevisiae is not caused by a limited anaplerotic capacity.     

Pyruvate carboxylase in Leptosphaeria michotü. Isolation, properties, intracellular localization and cyclic variations of its activity in relation to polypeptide level

Pyruvate carboxylase (EC 6.4.1.1) was obtained from the fungus Leptosphaeria michotü (West) Sacc. and enriched 543-fold by a 5-step purification procedure as an a4-β4 tetramer of M_r 440000, composedof a M_r 60000 α-subunit, containing bound biotin, and a M_r 50000 β-subunit. The enzyme was active from pH 6.5 to 12.0, with a maximum between pH 8.0 and 8.5. Its specific activity was 125nkat (mg protein)⁻¹: it was not affected by acetyl CoA. A rabbit antiserum raised against the yeast pyruvate carboxylase was specifically reactive against the α-subunits of the L. michotü enzyme. The enzyme was localized into the cytosol by gold-labelled streptavidin and immunogold staining of thin sections of Lowicryl-K4M-embedded colonies. Pyruvate carboxylase and acetylCoA carboxylase in L. michotü had synchronous activity rhythms at constant temperature and in darkness; these rhythms were suppressed by cycloheximide or avidin supply. The pyruvate carboxylase level was quantified along the activity rhythm by gel electrophoresis using ³⁵S-streptavidin. and by enzyme-linked immunosorbent assay (ELISA) using serum against the yeast pyruvate carboxylase. The cyclic variations of pyruvate carboxylase activity were correlated with cyclic variations in the enzyme level. Suppression of pyruvate and acetyl CoA carboxylase activities by avidin had a no important effect on the transaminase rhythms of L. michotü .