Bacterial expression and characterization of the mitochondrial outer membrane channel. Effects of n-terminal modifications

Several forms of the voltage-dependent anion-selective channel (VDAC) have been expressed at high yield in Escherichia coli. Full-length constructs of the proteins of Neurospora crassa and Saccharomyces cerevisiae (ncVDAC and scVDAC) have been made with 20-residue-long, thrombin-cleavable, His6-containing N-terminal extensions. ncVDAC purified from bacteria or mitochondria displays a far-UV CD spectrum (in 1% lauryl dimethylamine oxide at pH 6-8) similar to that of bacterial porins, indicating extensive beta-sheet structure. Under the same conditions, the CD spectrum of bacterially expressed scVDAC indicates lower beta-sheet content, albeit higher than that of mitochondrial scVDAC under the same conditions. In phospholipid bilayers, the bacterially expressed proteins (with or without N-terminal extensions) form typical VDAC-like channels with stable, large conductance open states (4-4.5 nanosiemens in 1 M KCl) and voltage-dependent transitions to a predominant substate (about 2 nanosiemens). A variant of scVDAC missing the first eight residues and having no N-terminal extension also has been expressed in E. coli. The truncated protein has a CD spectrum similar to that of mitochondrial scVDAC, but its channel activity is abnormal, exhibiting an unstable open state and rapid transitions between multiple subconductance levels.     

The membrane topology of the amino-terminal domain of the red cell calcium pump

A systematic study of the membrane-associated regions in the plasma membrane Ca2+ pump of erythrocytes has been performed by hydrophobic photolabeling. Purified Ca2+ pump was labeled with 3-(trifluoromethyl)-3-(m-[125I]iodophenyl)-diazirine ([125I]TID), a generic photoactivatable hydrophobic probe. These results were compared with the enzyme labeled with a strictly membrane-bound probe, [3H]bis-phosphatidylethanolamine (trifluoromethyl) phenyldiazirine. A significant light-dependent labeling of an M(r) 135,000-140,000 peptide, corresponding to the full Ca2+ pump, was observed with both probes. After proteolysis of the pump labeled with each probe and isolation of fragments by SDS-PAGE, a common pattern of labeled peptides was observed. Similarly, labeling of the Ca2+ pump with [125I]TID, either in isolated red blood cell membranes or after the enzyme was purified, yields a similar pattern of labeled peptides. Taken together, these results validate the use of either probe to study the lipid interface of the membrane-embedded region of this protein, and sustain the notion that the conformation of the pump is maintained throughout the procedures of solubilization, affinity purification, and reconstitution into proteoliposomes. In this work, we put special emphasis on a detailed analysis of the N-terminal domain of the Ca2+ pump. A labeled peptide of M(r) 40,000 belonging to this region was purified and further digested with V8 protease. The specific incorporation of [125I]TID to proteolytic fragments pertaining to the amino-terminal region indicates the existence of two transmembrane stretches in this domain. A theoretical analysis based on the amino acid sequence 1-322 predicts two segments with high probability of membrane insertion, in agreement with the experimental data. Each segment shows a periodicity pattern of hydrophobicity and variability compatible with alpha-helical structure. These results strongly suggest the existence of a transmembrane helical hairpin motif near the N-terminus of the Ca2+ pump.     

Reciprocated matrix metalloproteinase activation: a process performed by interstitial collagenase and progelatinase A

Gelatinase A, a member of the matrix metalloproteinase (MMP) family, is secreted possessing an 80 amino acid N-terminal propeptide that must be removed in order to generate the active enzyme. Purified progelatinase A was activated to 38% of maximum by a 6 h incubation at 37 degrees C with equimolar concentrations of trypsin-activated interstitial collagenase (another MMP). The increase in activity was accompanied by cleavage of the M(r) 72,000 progelatinase A to the M(r) 66,000 active enzyme that has Y81 as its N-terminus. At low concentrations, progelatinase A was processed via an inactive intermediate, suggesting that its activation is a biphasic process. This was confirmed by the action of collagenase on proE375-->A (a mutant of progelatinase A that cannot become active) because, in this instance, only an M(r) 68,000 species with L38 as the N-terminus was produced. The remaining propeptide amino acids to Y81 could be readily removed by added active gelatinase A, indicating that collagenase works by generating an intermediate that is susceptible to autolytic activation. Although relatively slow, the rate of activation could be increased approximately 10-fold by the addition of 100 micrograms/mL heparin. This binds to the C-terminal domain of collagenase and progelatinase A and presumably acts as a template that positions the reactants close to one another. Collagenase activated by trypsin retains 8 or 14 amino acids of its propeptide.(ABSTRACT TRUNCATED AT 250 WORDS)     

Optical activity and conformation of cobra neurotoxin

Cobra neurotoxin from Formosan cobra (Naja naja atra) venom is a compact globular protein having an intrinsic viscosity of 4.5 mL/g. The protein is stable in 7.5 M urea but can be denatured in 4.1 M guanidine hydrochloride or at elevated temperature (above 70 degrees C). Its conformation remains virtually the same in solvents of lower polarity than water such as 1,2-ethanediol or a mixed solvent of 1-propanol-1,2-ethanediol-water (5:1:1 by volume). The circular dichroism spectrum is "atypical" in water in that the peptide chromophores show a small negative circular dichroic (CD) band at 215 nm, a large positive one at 199 nm, and another large negative one below 190 nm. The CD pattern resembles to some extent that of a beta form but differs in both positions and magnitudes from the latter. It agrees qualitatively with the theoretical calculations of the reverse beta bends, suggesting that cobra toxin contains a considerable amount of beta turns and possibly a mixture of beta form and beta turns.     

Nicotinic acetylcholine receptor from rat brain. Solubilization, partial purification, and characterization

Nicotinic acetylcholine receptor protein (nAChR) has been solubilized from rat cerebral cortices by extracting a crude membrane fraction with the nonionic detergent Triton X-100 (polyoxyethylene-p-t-octylphenol). The solubilized nAChR was partially purified by affinity chromatography (Naja naja siamensis alpha-toxin affinity arm, linked to Sepharose 4B) and characterized by binding of 125I-labeled alpha-bungarotoxin. The reaction of labeled toxin and nAChR appears to be second order with a rate constant (k1) equal to 0.38 X 10(5) M-1 S-1 at 20 degrees. The toxin-nAChR complex dissociates with a dissociation rate constant (k-1) of 1.23 X 10(-5) S-1 at 20 degrees (t 1/2 = 15.6 h). The kinetically determined dissociation constant (Kd) for the complex is 3.24 X 10(-10) M. A variety of cholinergic ligands were studied for their ability to inhibit binding of labeled toxin. The results indicate that the brain receptor is indeed nicotinic. The s20, w and v of the toxin-nAChR complex in 0.1% Triton were determined by velocity sedimentation in D2O and H2O sucrose gradients. The values are 12.9 S and 0.80 cm3 g-1. The Stokes radius of the complex determined by gel filtration equals 7.5 nm. The Mr of the complex calculated from the hydrodynamic parameters, and corrected for bound detergent, equals 357,000.     

Comparative characterization of a wild type and transmembrane domain-deleted fatty acid amide hydrolase: identification of the transmembrane domain as a site for oligomerization

Fatty acid amide hydrolase (FAAH) is an integral membrane protein responsible for the hydrolysis of a number of primary and secondary fatty acid amides, including the neuromodulatory compounds anandamide and oleamide. Analysis of FAAH's primary sequence reveals the presence of a single predicted transmembrane domain at the extreme N-terminus of the enzyme. A mutant form of the rat FAAH protein lacking this N-terminal transmembrane domain (DeltaTM-FAAH) was generated and, like wild type FAAH (WT-FAAH), was found to be tightly associated with membranes when expressed in COS-7 cells. Recombinant forms of WT- and DeltaTM-FAAH expressed and purified from Escherichia coli exhibited essentially identical enzymatic properties which were also similar to those of the native enzyme from rat liver. Analysis of the oligomerization states of WT- and DeltaTM-FAAH by chemical cross-linking, sedimentation velocity analytical ultracentrifugation, and size exclusion chromatography indicated that both enzymes were oligomeric when membrane-bound and after solubilization. However, WT-FAAH consistently behaved as a larger oligomer than DeltaTM-FAAH. Additionally, SDS-PAGE analysis of the recombinant proteins identified the presence of SDS-resistant oligomers for WT-FAAH, but not for DeltaTM-FAAH. Self-association through FAAH's transmembrane domain was further demonstrated by a FAAH transmembrane domain-GST fusion protein which formed SDS-resistant dimers and large oligomeric assemblies in solution.     

Mature carnitine palmitoyltransferase I retains the N-terminus of the nascent protein in rat liver

Carnitine palmitoyltransferase I was isolated from octylglucoside extracts of rat liver mitochondrial outer membranes. This native enzyme was digested proteolytically with V8 protease. Five major peptides were obtained all of which were found in the amino acid sequence predicted from the full-length cDNA sequence of the protein. One peptide was found to correspond to the extreme N-terminus of the deduced amino acid sequence. Therefore, the mature protein retains the N-terminus of the nascent protein after import into the mitochondrial membrane. Knowledge of the identity of the N-terminus of the mature protein allows a reappraisal of the role of the two main. N-terminal hydrophobic domains of the protein and of the possible topology of the protein within the membrane.     

Identification, cDNA sequence and deduced amino acid sequence of the mitochondrial Rieske iron-sulfur protein from the green alga Chlamydomonas reinhardtii. Implications for protein targeting and subunit interaction

Specific oligonucleotide probes were used to isolate a cDNA clone for the mitochondrial Rieske iron-sulfur protein of the green alga Chlamydomonas reinhardtii. The protein is synthesized as a longer precursor with a cleavable N-terminal presequence of 54 amino acids but without a C-terminal extension. Comparison of the predicted secondary structure of this N-terminal sequence with that of the targeting signal of the chloroplast Rieske protein from C. reinhardtii [de Vitry (1994) J. Biol. Chem. 269, 7603-7609] indicates that, although they both have the potential to form amphiphilic alpha helices, the mito-chondrial presequence may form a more hydrophobic helix that could penetrate deeper into the membrane. The N-terminal part of the mature mitochondrial Rieske protein is characterized by a long, strongly hydrophilic N-terminal domain and by a positive charge in the middle of the hydrophobic stretch that is presumed to interact with the bc1 complex. Thus, the protein from C. reinhardtii differs from the Rieske proteins from mammals or fungi.     

Roughness of the globular protein surface]

The area and volume of the approximating ellipsoids taken from low resolution X-ray data have been calculated for 65 globular proteins. It has been shown that the dependence of these values on the protein molecular mass (M) coincides with those for even isometric bodies. This indicates that the asymmetry of globular proteins does not grow with the increase of their sizes. At the same time the 0.73 slope of the log-log dependence of the accessible surface area (A(s)) on the protein molecular mass differing from the value of 0.67 for even isometric bodies was observed (Miller S. et al., J. Mol. Biol. 1987. V.196. P.641). This can be explained by peculiarities of the protein surface. The method of molecule shape recovery by spherical harmonics has shown that the domain organization of protein molecule cannot explain the observed difference. Therefore the more detailed analysis of protein surface structure would be necessary.     

Dihydrolipoamide dehydrogenase in the trypanosoma subgenus, trypanozoon

The enzyme dihydrolipoamide dehydrogenase has been discovered and characterised in four salivarian trypanosomes of the subgenus trypanozoon: Trypanosoma brucei brucei, T. b. gambiense, T. b. rhodesiense, and Trypanosoma evansi. The three T. brucei species, which have insect procyclic forms biochemically distinct from their mammalian bloodstream forms, express dihydrolipoamide dehydrogenase in both cell types, but have higher levels in the procyclic forms. Determination of Michaelis constants for the enzyme from each of the three T. brucei species did not reveal any significant kinetic differences between the bloodstream and procyclic enzymes. On Western blots, antibodies raised against dihydrolipoamide dehydrogenase from the stereorarian trypanosome, Trypanosoma cruzi, cross-react strongly with the dihydrolipoamide dehydrogenase from all three T. brucei species; by this method, the relative molecular masses of their dihydrolipoamide dehydrogenases are indistinguishable. Dihydrolipoamide dehydrogenase was purified from both the bloodstream and the procyclic forms of T. b. brucei, and the N-terminal have been sequenced. These sequences are identical to the derived protein sequence of the cloned gene (Else et al., Eur. J. Biochem. 212 (1993) 423-429), but have a nine amino acid N-terminal truncation, giving an N-terminus equivalent to that of T. cruzi dihydrolipoamide dehydrogenase. The T. b. brucei dihydrolipoamide dehydrogenase gene has been expressed in Escherichia coli and the resultant protein purified; its N-terminus is processed in a similar fashion to that in the trypanosome, but with reduced specificity.     

Multiubiquitin chain binding and protein degradation are mediated by distinct domains within the 26 S proteasome subunit Mcb1

The 26 S proteasome is a multisubunit proteolytic complex responsible for degrading eukaryotic proteins targeted by ubiquitin modification. Substrate recognition by the complex is presumed to be mediated by one or more common receptor(s) with affinity for multiubiquitin chains, especially those internally linked through lysine 48. We have identified previously a candidate for one such receptor from diverse species, designated here as Mcb1 for Multiubiquitin chain-binding protein, based on its ability to bind Lys48-linked multiubiquitin chains and its location within the 26 S proteasome complex. Even though Mcb1 is likely not the only receptor in yeast, it is necessary for conferring resistance to amino acid analogs and for degrading a subset of ubiquitin pathway substrates such as ubiquitin-Pro-beta-galactosidase (Ub-Pro-beta-gal) (van Nocker, S., Sadis, S., Rubin, D.M., Glickman, M., Fu, H., Coux, O., Wefes, I., Finley, D., and Vierstra, R. D. (1996) Mol. Cell. Biol. 16, 6020-28). To further define the role of Mcb1 in substrate recognition by the 26 S proteasome, a structure/function analysis of various deletion and site-directed mutants of yeast and Arabidopsis Mcb1 was performed. From these studies, we identified a single stretch of conserved hydrophobic amino acids (LAM/LALRL/V (ScMcb1 228-234 and At-Mcb1 226-232)) within the C-terminal half of each polypeptide that is necessary for interaction with Lys48-linked multiubiquitin chains. Unexpectedly, this domain was not essential for either Ub-Pro-beta-gal degradation or conferring resistance to amino acid analogs. The domain responsible for these two activities was mapped to a conserved region near the N terminus. Yeast and Arabidopsis Mcb1 derivatives containing an intact multiubiquitin-binding site but missing the N-terminal region failed to promote Ub-Pro-beta-gal degradation and even accentuated the sensitivity of the yeast delta mcb1 strain to amino acid analogs. This hypersensitivity was not caused by a gross defect in 26 S proteasome assembly as mutants missing either the N-terminal domain or the multiubiquitin chain-binding site could still associate with 26 S proteasome and generate a complex indistinguishable in size from that present in wild-type yeast. Together, these data indicate that residues near the N terminus, and not the multiubiquitin chain-binding site, are most critical for Mcb1 function in vivo.     

Protein-nucleic acid recognition: simulation of base and "model" amino acids complexes in DMSO by the Monte Carlo method

Phosphatidylinositol-specific phospholipase C (PI-PLC) from Bacillus thuringiensis catalyzes the hydrolysis of phosphatidylinositol (PI) in discrete steps: (i) an intramolecular phosphotransferase reaction to form inositol 1,2-(cyclic)-phosphate (cIP), followed by (ii) a cyclic phosphodiesterase activity that converts cIP to inositol 1-phosphate. Water-soluble cIP was used as the substrate to study the cyclic phosphodiesterase activity and interfacial behavior of PI-PLC. Different detergent micelles and phospholipid vesicles were used to examine if "interfacial activation" of the enzyme could occur toward a soluble substrate. Almost all detergents examined activated the enzyme at least 2-fold, with PC species yielding the largest increases in PI-PLC specific activity. Kinetic parameters were measured in the absence and presence of several representative detergents (e.g., Triton X-100 and diheptanoylphosphatidylcholine (diC7PC)). Gel filtration experiments showed that, under these conditions, the cIP did not partition to any measurable extent with these detergent micelles. The concentration at which half the maximum activation was observed occurred near the detergent CMC. Both Km and Vmax were altered by the presence of a surface: Km decreased to different degrees depending on the detergent, while Vmax increased substantially. The Km for cIP was 90 mM without detergent and decreased to 29 mM with diC7PC micelles added; Vmax increased almost 7-fold in the presence of diC7PC micelles. The enzyme efficiency (Vmax/Km) in the presence of diC7PC increased more than 21-fold, but it was still 20-fold lower than initial phosphotransferase activity for monomeric dihexanoylphosphatidylinositol. The poor efficiency of the cyclic phosphodiesterase activity is largely due to substrate binding affinity. The dependence of rate on substrate concentration exhibits cooperative behavior, especially without detergent. This cooperativity is discussed in terms of protein aggregation and ligand binding sites on the enzyme.     

Gastro-intestinal protein loss in late survivors of Fontan surgery and other congenital heart disease

A 2,048-bp nucleotide sequence containing a gene coding for an enzyme that degraded guar gum from Bacillus circulans K-1 was identified by polymerase chain reaction walking. This G-gene consisted of 1,551 nucleotides coding for a protein with Mr 55,242. The enzyme was overexpressed in Escherichia coli JM109 cells by the cloning the G-gene downstream of the lac Z promoter of pUC19. The molecular mass of recombinant G-enzyme estimated by SDS-PAGE was 62 KDa, close to that from strain K-1. Analysis of the recombinant enzyme showed GalNAc, Xyl, GlcNAc, Man, Glc, and Gal to account for 1.7%, 14.4%, 6.1%, 3.2%, 54.2%, and 10.4%, respectively, of the total monosaccharides. Polyacrylamide gel electrophoresis of this enzyme with staining gave a red band. The results suggested that the sugars accounted for the differences in the molecular masses. The recombinant enzyme had two kinds of N-terminal sequences, Thr-Met-Ile-Thr-Pro-Ser-Phe-Ala-Ser-Gly-Phe-Tyr-Val-Ile and Ile-Thr-Pro-Ser-Phe-Ala-Ser-Gly-Phe-Tyr-Val-Ile-Gly-Thr. Comparison of these sequences with the deduced N-terminal sequence coded for the G-gene showed that the amino acid, first Met, of the lac Z gene or the next residues Thr-Met in the recombinant enzyme were absent in the native enzyme. Methionines near and at the N-terminus of the mature protein probably were digested by methionine aminopeptidases of E. coli after translation. The properties of recombinant G-enzyme were similar to those of the enzyme from K-1 cells.     

Expression of equine morbillivirus (EMV) matrix and fusion proteins and their evaluation as diagnostic reagents

Full-length cDNA clones coding for the matrix (M) and fusion (F) proteins of equine morbillivirus (EMV) were isolated by RT-PCR, and expressed in Escherichia coli using two different expression systems. Western blot analysis indicated that the M and F proteins, expressed either by itself or as fusion proteins with glutathione S-transferase (GST), were insoluble and degraded after expression. Analysis of the degradation pattern of recombinant M protein suggested that the N-terminus of the matrix protein might be more stable and antigenic than the C-terminal region. Therefore a third system was used to express a truncated M protein, composed of the N-terminal amino acid residues 1-197, with a (His)6-tag attached at the N-terminus. This recombinant protein [(His)6-Mtr], was stable but was also insoluble. After one-step affinity purification under denaturing conditions, (His)6-Mtr was used to monitor the antibody response to EMV infection by Western blot and ELISA. We obtained a 100% correlation between Western blot and virus neutralisation testing although the number of positive sera available for testing was very limited, which included seven horse, two rabbit and one human sera.     

Kinetics of lysozyme refolding: structural characterization of a non-specifically collapsed state using time-resolved X-ray scattering

We report time-resolved small angle X-ray scattering (SAXS) studies of the structural characteristics of the collapsed state of lysozyme from henegg white (HEL) obtained on initiating refolding by rapidly changing solvent conditions from 8 M to 1.1 M urea at pH 2.9. At this reduced pH the lifetime, of about one second, of the non-specifically collapsed ensemble is considerably prolonged relative to its value at pH 5.2. The SAXS studies are combined with time resolved measurements of tryptophan fluorescence and of the rate of formation of native molecules using interrupted refolding experiments. We observe large burst phase changes in intrinsic tryptophan fluorescence and in the radius of gyration (Rg) which is reduced from 22 A in the fully unfolded state to approximately 19 to 20 A. Subsequent decrease of the Rg to the value for native lysozyme (15 A) follows the time course of formation of native molecules. Single exponential fits to the singular value decomposition (SVD) components of the SAXS data allow reconstruction of the SAXS profile at early time points of refolding. The results of this analysis suggest a globular shape of the collapsed state. A similar fit to the forward scattering amplitude, I(0), suggests that the collapsed state has a solvent accessible surface area which is considerably increased relative to that of the native protein. These results show directly that the non-specifically collapsed state formed during the burst phase in lysozyme refolding indeed represents a molecular compaction and a change in shape from a fully denatured random coil state (albeit restricted by disulfide bonds) to an ensemble of globular conformations which, however, have not yet formed a solvent-protected hydrophobic core.     

Molecular cloning and nucleotide sequence of the genomic DNA for 1,2-alpha-D-mannosidase gene, msdC from Penicillium citrinum

A gene encoding 1,2-alpha-D-mannosidase (EC 3.2.1.113) was cloned from Penicillium citrinum genomic DNA using the polymerase chain reaction (PCR). The coding region of the gene, msdC, occupied 1737 bp and was separated into four exons by three introns. The predicted protein consisted of 511 amino acid residues with M(r) 56,569. Penicillium enzyme had a hydrophobic signal peptide at the N-terminal region as did mammalian membrane-bound alpha-mannosidases, but in this case a proteolytic cleavage occurred at Lys-35-Ser-36 to remove the signal sequence during cell growth. Parts of amino acid sequences were similar to those of mammalian Golgi alpha-mannosidase IA and IB, but the sequence around the aspartic acid residue which interacted with 1-deoxymannojirimycin (Yoshida et al. (1994) Biochem. J. 303, 97-103) was unique in Penicillium enzyme.     

Expression of an olfactory receptor in Escherichia coli: purification, reconstitution, and ligand binding

An olfactory receptor has been expressed in bacterial cells as a fusion protein with glutathione S-transferase (GST). Overexpression of receptor protein yielding about 10% of the cell protein was achieved with mutants lacking the N-terminus and the first transmembrane region or with mutants carrying three positively charged residues in the first intracellular loop. The overexpressed fusion protein accumulated in inclusion bodies and could be solubilized in detergent. It was purified by metal chelation chromatography based on a C-terminal 6-histidine tag, and the GST portion was removed after proteolytic cleavage. The purified receptor was reconstituted into lipid vesicles and specific binding of odor ligands was shown by photoaffinity labeling and tryptophan fluorescence measurements. Thus, for the first time, an odorant receptor/ligand pair becomes available in large amounts for biophysical and screening studies.     

Effect of D57N mutation on membrane activity and molecular unfolding of cobra cardiotoxin

Cobra cardiotoxins (CTXs) are able to adopt a three-fingered beta-strand structure with continuous hydrophobic patch that is capable of interacting with zwitterionic phospholipid bilayer. In addition to the four disulfide bonds that form the rigid core of CTXs, Asp57 near the C-terminus interacts electrostatically with Lys2 near the N-terminus (Chiang et al. 1996. Biochemistry. 35:9177-9186). We indicate herein, using circular dichroism and the time-resolved polarized tryptophan fluorescence measurement, that Asp57 to Asn57 (D57N) mutation perturbs the structure of CTX molecules at neutral pH. The structural stability of the D57N mutant was found to be lower, as evidenced by the reduced effective concentration of the 2,2,2-trifluoethanol (TFE)-induced beta-sheet to alpha-helix transition. Interestingly, the single mutation also allows a greater degree of molecular unfolding, because the rotational correlation time of the TFE-induced unfolding intermediate is larger for the D57N mutant. It is suggested that the electrostatic interaction between N- and C-termini also contributes to the formation of the functionally important continuous hydrophobic stretch on the distant end of CTX molecules, because both the binding to anilinonaphthalene fluorescent probe and the interaction with phospholipid bilayer were also reduced for D57N mutant. The result emphasizes the importance of the hydrophobic amino acid residues near the tip of loop 3 as a continuous part of the three-fingered beta-strand CTX molecule and indicates how a distant electrostatic interaction might be involved. It is also implicated that electrostatic interaction plays a role in expanding the radius of gyration of the folding/unfolding intermediate of proteins.     

Renal prostaglandins and thromboxane A2 lack a functional significance in the genesis of protein-induced glomerular hyperfiltration in human renal disease

Ribulose 1,5-bisphosphate carboxylase/oxygenase (RubisCO) was purified from an obligately autotrophic hydrogen-oxidizing bacterium, Hydrogenovibrio marinus MH-110. The protein has a M(r) value of approximately 110,000, and is composed of two identical subunits of 55,000. To our knowledge, the existence of L2-form RubisCO in a chemolithoautotrophic bacterium is first reported in this paper. The N-terminal amino acid sequence determination of the purified enzyme showed high homology with those of the L2-form RubisCO of Rhodospirillum rubrum and the Lx-form RubisCO from Rhodobacter sphaeroides.