Antibody-linked polymerase assay on protein blots: a novel method for identifying polymerases following SDS-polyacrylamide gel electrophoresis

We describe a method for correlating polymerase activity with a particular polypeptide band in an SDS-polyacrylamide gel which does not require renaturation of the SDS-denatured enzyme. The method involves the following steps: (i) transfer of proteins from an SDS-polyacrylamide gel onto nitrocellulose; (ii) incubation with excess antiserum raised against a partially purified polymerase preparation to link one Fab site of an antibody molecule to the denatured enzyme on the nitrocellulose; (iii) binding of native polymerase to the other Fab site of the antibody molecule in the immune complex to generate a specific polymerase 'sandwich'; (iv) assaying of the nitrocellulose filter for antibody-linked native polymerase activity using an appropriate template and a radioactive substrate followed by treatment with trichloroacetic acid to precipitate in situ the radioactive product. The essential feature of this method is that the use of both non-specific anti-polymerase serum and a partially purified enzyme preparation is sufficient to allow identification of a specific protein following SDS-polyacrylamide gel electrophoresis. This antibody-linked polymerase assay has been developed to identify a 130,000-dalton RNA-dependent RNA polymerase from cowpea leaves. Possible applications of this type of assay as a tool for identifying a wide variety of proteins are discussed.     

Renewable enzyme reactors based on beds of artificial gel antibodies

A novel approach is described for the synthesis of beds for enzyme reactors. The method is based on the use of artificial antibodies in the form of polyacrylamide gel particles with diameters around 0.1–0.3 mm. These gel particles mimic protein antibodies, raised in experimental animals, in the sense that they selectively recognize and adsorb only the protein present during the preparation of the “antibodies”. The gel antibodies have several advantages over conventional protein antibodies, which can be taken advantage of in the design of enzyme reactors; for instance, if upon prolonged use the immobilized enzyme loses its activity it can easily be replaced by an active enzyme, which is not possible when the enzyme is immobilized via a conventional protein antibody (a new bed with immobilized protein antibodies must be prepared); and equally or more remarkable: the enzyme can be applied in the form of a non-purified extract since the selectivity of the artificial gel antibodies is so high that they will “fish-out” the enzyme, but no other proteins in the extract. In addition, no preconcentration of the enzyme solution is required prior to the immobilization, since the enzyme is enriched at the top of the column upon the application. These unique properties make enzyme reactors based on artificial gel antibodies very attractive, also in process chromatography. The potential application range of the artificial gel antibodies is enormous since the same method for their synthesis can be used independent of the structure and the size of the “antigen”; for instance, renewable biosensors based on gel antibodies for the selective detection of protein biomarkers, as well as pathogenic viruses, bacteria, and spores (for instance Anthrax) should not be difficult to design.     

Purification and characterization of neuraminidase from Clostridium perfringens.

Clostridium perfringens cells were cultivated on a large scale using an automatic system. Neuraminidase secreted by the cells into the culture medium was purified 380 000-fold by: precipitation with ammonium sulfate between 50 and 85% saturation, filtration on Sephadex G-75, electrophoresis on polyacrylamide gel, and by isoelectric focusing. Three enzyme fractions with different migration rates were obtained by preparative disc electrophoresis in polyacrylamide gel, and five fractions with isoelectric points between pH 4.7 and 5.4 were observed after isoelectric focusing. This microheterogeneity disappeared after denaturation of the enzyme in 0.1% sodium dodecylsulfate or 8M urea. The isoelectric point of the denatured enzyme corresponded to pH 4.3. All enzyme fractions were identical with regard to their immunological and kinetic properties; they had the same molecular weights. The origin of the different "conformers" of neuraminidase is discussed. The existence of genuine isoenzymes could largely be excluded. The yield of neuraminidase was 65%, which corresponded to about 10 mg of pure enzyme from 100 l of culture medium. The enzyme was free of protease and various other glycosidase activities. The neuraminidase preparation appeared not to be contaminated by other proteins as judged by electrophoretic analysis using either the native enzyme or the enzyme denatured by sodium dodecylsulfate or urea; ultracentrifugation; chromatography on Sephadex G-200; and immunological methods. The molecular weights of the native or denatured enzyme were found to be in the range between 60 000 and 69 000 (on an average 63 750) using four independent methods. The existence of subunits of neuraminidase was excluded. The neuraminidase exhibited a spec. act. of 580 or 615 U/mg protein with glycopeptides from edible birds' nests or sialyllactose, respectively, as substrates. Additional kinetic properties and the UV-absorption spectrum of the enzyme are described.     

Immunochemical comparison of cardiac glycoside-sensitive (lamb) and -insensitive (rat) kidney (Na+ + K+)-ATPase

The immunological cross-reactivity of the ouabain-sensitive lamb kidney and the ouabain-insensitive rat kidney (Na+ + K+)-ATPase (EC 3.6.1.37) was examined using polyclonal and monoclonal antibodies. Studies using rabbit antisera prepared against both the lamb kidney and rat kidney holoenzymes showed the existence of substantial antigenic differences as well as similarities between the holoenzymes and the respective denatured alpha and beta subunits of these two enzymes. Quantitation of the extent of cross-reactivity using holoenzyme-directed antibodies showed a 40-60% cross-reactivity. In addition, rabbit antisera monospecific to the purified, denatured alpha and beta subunits of the lamb kidney enzyme showed about a 50% cross-reactivity towards the respective subunit of the rat enzyme. In contrast to the cross-reactivity observed using the polyclonal antibodies, six monoclonal antibodies specific for the alpha subunit of the lamb holoenzyme exhibited no cross-reactivity with the rat holoenzyme. Four of these monoclonal antibodies, however, showed substantial cross-reactivity with rat alpha subunit as resolved by SDS-polyacrylamide gel electrophoresis. A fifth antibody did not bind to the denatured alpha subunit of either the lamb or the rat enzyme. Another monoclonal antibody (M7-PB-E9), which is specific for an epitope previously implicated in the regulation of both ATP and ouabain binding to (Na+ + K+)-ATPase (Ball, W.J., Jr. (1984) Biochemistry 2275-2281) was found to bind to the denatured lamb alpha but not to the rat alpha. This antibody has identified a region of the lamb alpha that has an altered amino acid sequence in the ouabain-insensitive rat enzyme. These immunological studies indicate that there are substantial antigenic differences between the lamb and rat kidney (Na+ + K+)-ATPases. The majority of these antigenic differences appear to be due to variations in the tertiary structures rather than to variations in the primary structures of the alpha subunits.     

Purification and properties of NADP(+)-dependent isocitrate dehydrogenase from the corpus luteum

Cytoplasmic NADP(+)-dependent isocitrate dehydrogenase (isocitrate: NADP+ oxidoreductase (decarboxylating), EC 1.1.1.42) was purified 290-fold from the 15,000 x g supernatant fraction of porcine corpora lutea. The major purification step was by anion-exchange chromatography with an FPLC mono P column. Enzyme lability was overcome by including Mg2+, DL-isocitrate, dithiothreitol and glycerol in the elution buffers. The molecular weight of the denatured enzyme was found to be 48,000 by SDS-polyacrylamide gel electrophoresis. The Stokes' radius was estimated to be 3.7 nm by gel filtration and the isoelectric point was 4.8 as determined by chromatofocusing. The purified enzyme had a specific activity of 57.8 units/mg and a broad optimal pH for activity from 7.5 to 9.0. The Km for the substrates DL-isocitrate and NADP+ were 13 and 12 microM, respectively. Polyclonal antibodies were raised against the purified enzyme. Protein (Western) blotting showed an immunological similarity between the cytoplasmic enzyme of the ovary, liver, adrenal gland and heart. A difference was demonstrated between the ovarian enzyme and the heart mitochondrial enzyme. The substrate turnover number and Mr of the ovarian enzyme were similar to those found for the enzyme from the liver and adrenal gland.     

Biochemical and immunological characterization of the secreted forms of human neutrophil gelatinase

Human neutrophils contain a neutral metalloproteinase which degrades denatured collagens and potentiates the action of interstitial collagenase. This gelatinase is rapidly secreted from neutrophils stimulated with phorbol myristate acetate. The secreted enzyme has been purified by a combination of chromatography on DEAE-cellulose and gelatin-Sepharose. The purified enzyme was latent and had a specific activity of 24,000 units. Estimated molecular weight obtained by gel filtration was 150,000-180,000. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of the purified enzyme showed three bands with relative molecular weights of 225,000, 130,000, and 92,000. Electrophoresis in the presence of a reducing agent revealed a single band of Mr = 92,000. All the proteins seen on the unreduced gel were found to contain proteolytic activity against gelatin and native type V collagen. Polyclonal antibodies were prepared against the Mr = 130,000 and 92,000 proteins. When analyzed by immunoblotting, both antibodies recognized all three proteins. Furthermore, the identical three proteins were identified by the antibodies when crude culture medium was immunoblotted. The purified enzyme was inhibited by EDTA and 1,10-phenanthroline but not by serine or thiol proteinase inhibitors, suggesting that the enzyme is a metalloendoproteinase. The enzyme had little or no activity against common protein substrates such as bovine serum albumin or casein. Native type I collagen was not cleaved under conditions where native type V collagen was extensively degraded.     

A new neomycin phosphotransferase II solid phase assay in combination with polyacrylamide sodium dodecylsulphate gel electrophoresis

A new general method for the determination of neomycin phosphotransferase (NPT) II (EC 2.7.1.95) activity in cell extracts after separation in SDS-polyacrylamide gels is described. The enzymatic activity of NPT II is restored after SDS-polyacrylamide gel electrophoresis by incubating the gel for 3 h (20 mM Tris-HCl buffer, pH 7.4). The enzymatic activity is determined by in situ phosphorylation of aminoglycoside antibiotics bound to solid supports and brought into direct contact with the gel surface. A novel, mechanically stable, negatively charged matrix was synthesized for use in this solid phase enzyme assay and compared to phosphocellulose and carboxymethylcellulose paper. This new method allows the easy and exact determination of the molecular weight of any fusion protein with NPT II by assaying the position of the enzymatic activity in the gel and a consecutive immunological reaction following protein transfer onto nitrocellulose membranes.     

Physico-chemical and immunological properties and partial amino acid sequencing of a new metalloprotease: endoprotease Thr-N

Previous studies have described the isolation of a new metalloprotease with a strict specificity for the amide bonds of peptide substrates having a threonine residue at the P1' position [Biochem. Biophys. Res. Commun. 256 (1999) 307]. The present work reports the physico-chemical properties of the enzyme which enable the optimal conditions for the digestion of proteins by the protease to be determined. At pH 8.2 and up to 37 degrees C, the enzyme possesses a good proteolytic activity and is stable for at least 12 h. The protease is sensitive to detergents and dithiol-reducing agents so that these chemicals must be eliminated after treatment of the protein substrate when this needs to be denatured and reduced before its hydrolysis by the enzyme. An increase in the enzymatic activity is observed in the presence of urea up to a 2.0 M concentration, beyond which the activity decreases. The enzyme can also be used in the presence of organic solvents such as acetonitrile, isopropanol or dioxane (10%, v/v) without loss of activity. Studies performed with antibodies raised against the purified endoprotease Thr-N indicated the absence of cross-immunoinactivation and cross-immunoprecipitation with all tested proteases. Also, no homology of sequence was found with the proteases indexed in the databases. Thus, our results show that endoprotease Thr-N not only represents an original protease by its unique specificity but also by its immunological and molecular properties.     

Antibodies to Pseudomonas testosteroni 3-oxosteroid Δ4-Δ5-isomerase

Antibodies to Pseudomonas testosteroni isomerase have been generated in rabbits immunized with the isomerase. Complexes between the isomerase and antibodies were characterized by double immunodiffusion and by enzymatic assays. The titre of antibodies was determined. The enzymatic activity was totally abolished with increasing antibody concentrations. Antibody-isomerase complexes were inactive. Total immunological identity was observed between the unpurified extract and the pure enzyme either free or bound to a competitive inhibitor. However, the denatured enzyme did not interact with antibodies. Antigenic identy could not be detected between the bacterial antibodies and the bovine adrenocortical isomerase from microsomes.     

Neutral aminopeptidase: a potential marker enzyme of the amphibian urinary bladder epithelial cell apical membrane

Antidiuretic hormone induces, in the apical plasma membrane of amphibian urinary bladder epithelial cells, the exocytotic insertion of intramembranous particle aggregates that probably contain water channels. Purification of the apical membrane is a way to characterize the aggregates. The isolation of such purified membranous fractions involves the use of specific exogenous or endogenous markers. One of them could be the neutral aminopeptidase (AP), whose activity was detected in urinary bladder. Enrichment in AP activity was observed in plasma membrane preparations compared to cell homogenates (X2.7). However, a large part of the enzyme activity was also recovered in the soluble fraction of the preparation, suggesting large proteolysis of the protein. The enzyme presents a low optimal pH (6.4) and a high specificity for proline-p-nitroanilide as compared to the AP present in kidneys and intestines. To localize the protein in the amphibian bladder epithelium, an immunological approach was necessary due to the low activity of the enzyme in this tissue. The low enzymatic activity also prevented the purification of sufficient amounts of the urinary bladder AP as antigen, and we prepared antibodies against purified AP from frog or toad kidneys where the activity is 60 times higher than in the bladder. The serum specificity was verified by spot immunodetection, Western blot, inhibition capacity of antibodies, and immunoadsorption on a solid support with the renal enzyme. The sera were found to be able to react with native as well as denatured forms of the kidney enzyme. Antibodies cross-reacted with several peptides of low molecular weight (40-60 kDa) from urinary bladder plasma membrane proteins (Western blot).(ABSTRACT TRUNCATED AT 250 WORDS)     

Immunological comparison of the starch branching enzymes from potato tubers and maize kernels

Starch branching enzyme was purified from potato (Solanum tuberosum L.) tubers as a single species of 79 kilodaltons and specific antibodies were prepared against both the native enzyme and against the gel-purified, denatured enzyme. The activity of potato branching enzyme could only be neutralized by antinative potato branching enzyme, whereas both types of antibodies reacted with denatured potato branching enzyme. Starch branching enzymes were also isolated from maize (Zea mays L.) kernels. All of the denatured forms of the maize enzyme reacted with antidenatured potato branching enzyme, whereas recognition by antinative potato branching enzyme was limited to maize branching enzymes I and IIb. Antibodies directed against the denatured potato enzyme were unable to neutralize the activity of any of the maize branching enzymes. Antinative potato branching enzyme fully inhibited the activity of maize branching enzyme I; the neutralized maize enzyme was identified as a 82 kilodalton protein. It is concluded that potato branching enzyme (M_r = 79,000) shares a high degree of similarity with maize branching enzyme I (M_r = 82,000), in the native as well as the denatured form. Cross-reactivity between potato branching enzyme and the other forms of maize branching enzyme was observed only after denaturation, which suggests mutual sequence similarities between these species.     

Pyrophosphate:fructose-6-phosphate 1-phosphotransferase from barley seedlings. Isolation, subunit composition and kinetic Characterization

Pyrophosphate:fructose-6-phosphate I-phosphotransferase (PFP: EC 2.7.1.90) was purified 260-fold from leaves of etiolated barley seedlings. The purified enzyme consisted of two subunits, with apparent molecular masses of 65 (α) and 60 (β) kDa. Polyclonal antibodies were raised against the denatured PFP protein eluted from an SDS-polyacrylamide gel. The antibodies recognized both denatured and native PFP. Western blots of crude extracts showed that the activity of PFP in barley leaves is correlated to the amount of PFP protein, and that both the α- and the β-subunits are present in near stoichiometric amounts in all investigated tissues. The apparent molecular mass of the boloenzyme. as determined by gel filtration chromatography, was dependent on the presence of pyrophosphate. In absence of pyrophosphate. barley PFP elutes as a heterotetramer whereas it elutes as a heterooctamer in the presence of 20 m M pyrophosphate. Pure PFP obtained by gel filtration chromatography in the presence of 20 m M pyropnosphaie reached a specific activity of 28 U mg⁻¹. Barley PFP was characterized with respect 10 kinetic properties in the forward direction (use of PP₁) and in the reverse direction (formation of PP₁). The affinity for the activator Fru-2.6-P₂: was very high, with an estimated K₃ of 2.8 n M when PFP activity was assayed in the forward direction.     

Identification of cell surface dipeptidylpeptidase IV in human fibroblasts.

An antigen with dipeptidylpeptidase IV activity was identified at the surface of normal human fibroblasts. Hydrophobic interaction electrophoresis in phenyl-Sepharose revealed that the enzyme contained a hydrophobic domain, while lactoperoxidase-catalysed iodination with 125I of living cells indicated that the protein was located at the cell surface. Crossed immunoelectrophoresis with specific antibodies of acid-extracted or papain-treated cells showed a shift of the dipeptidylpeptidase IV peak to a faster mobility. The molecular properties of the fibroblast enzyme were clearly different from those described for dipeptidylpeptidase IV from other tissues and species. Fibroblast dipeptidylpeptidase IV contained two different disulphide-linked subunits, of apparent Mr values 125000 and 135000 (denatured and reduced). In gel filtration, an Mr of about 400000 was observed for the unreduced molecule. The enzymic properties of fibroblast dipeptidylpeptidase IV were very similar to those of the well-characterized pig kidney enzyme. Activity towards glycyl-L-prolyl-beta-naphthylamide was inhibited 50% by 0.023 mM-di-isopropylphosphorofluoridate. L-Alanyl-L-alanyl-beta-naphthylamide was hydrolysed ten times more slowly than glycyl-L-prolyl-beta-naphthylamide.     

Effects of Proline Analogues on the Formation of Alkaline Phosphatase in Escherichia coli

Two of the four proline analogues tested for their effect on the formation and activity of Escherichia coli alkaline phosphatase were able to substitute for proline in protein synthesis in a proline auxotroph. One of these, 3,4-dehydroproline, effectively replaced proline and led to formation of an active enzyme under conditions where no proline was present in the polypeptides. Substitution of azetidine-2-carboxylate for proline prevented active enzyme formation, producing instead altered monomeric forms of the alkaline phosphatase. These were detected with antibodies specific to denatured forms of the enzyme, and they were also characterized, together with cellular proteins, by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Alkaline phosphatase, as well as several other proteins, is localized exterior to the bacterial cell cytoplasm in the periplasmic space. In the presence of azetidine-2-carboxylate, a substantial number of these periplasmic proteins retain their specific site of localization, and the denatured subunits of alkaline phosphatase were only detected in the periplasmic fraction of the cell. Thus, secretion of these proteins does not appear to require a high degree of specificity in the native structure of the polypeptide chain. The analogues 4-allohydroxyproline and 4-thiazolidine carboxylate were unable to substitute for proline in protein synthesis but they inhibited growth of E. coli.     

The antigenic structure of the acetylcholine receptor from Torpedo californica

The immunological structure of the acetylcholine receptor (AChR) from the electric organ of Torpedo californica was studied using a large number of monoclonal antibodies which were initially selected for their abilities to bind to intact AChRs. The monoclonal antibodies were tested for their ability to bind to denatured AChR subunits labeled with 125I. Antibodies derived from rats immunized with individual denatured subunits or a mixture of subunits of Torpedo AChR reacted well in the assay. A much smaller proportion of antibodies derived from rats immunized with native Torpedo AChR or native AChR from Electrophorus electricus electric organ, bovine muscle, or human muscle reacted with denatured subunits of Torpedo AChR. Many monoclonal antibodies reacted with more than one subunit, but they always reacted best with the subunit used for immunization. Those monoclonal antibodies that bound to intact subunits were mapped more precisely by their ability to bind characteristic fragments of each subunit generated by proteolysis with Staphylococcal V8 protease. These fragments were analyzed by SDS polyacrylamide gel electrophoresis, and monoclonal antibodies that precipitated the same fragment pattern were placed in groups. By this method, we define a minimum of 28 determinants on Torpedo AChR.     

Monoclonal antibodies against various forms of the adenylyl cyclase catalytic subunit and associated proteins

Monoclonal antibodies against partially purified adenylyl cyclase from bovine brain cortex were raised in mice. Three types of antibody were obtained. Type 1 was specific for the calmodulin-sensitive enzyme. Type II also recognized this enzyme, but recognized the calmodulin-insensitive enzymes from a variety of species and tissues as well. Type I antibodies precipitated their antigens in both the native and denatured forms, while type II strongly favored the denatured forms. Type III antibodies precipitated adenylyl cyclase activity, but as shown by Western blot analysis, were directed against 38-kDa and 45-kDa glycoproteins. The 38-kDa protein was identified as synaptophysin.     

Monoclonal antibodies against human beta-glucocerebrosidase

Monoclonal antibodies were obtained against the membrane-bound lysosomal enzyme beta-glucocerebrosidase (acid beta-glucosidase), which is deficient in Gaucher's disease. BALB/c mice were immunized with homogeneous enzyme protein extracted from a sodium dodecyl sulphate/polyacrylamide gel. The mice were subsequently hyperimmunized with partially purified enzyme prior to fusion of spleen cells with myeloma cells. After fusion, 32 primary hybrid cell populations were obtained which continued to produce antibodies against beta-glucocerebrosidase after prolonged time of culture. All antibodies reacted with both native and denatured enzyme. Four primary cell populations were subcloned and the antibodies produced were characterized. The antibodies were all four of the IgG1 subclass. Three of these antibodies bind to protein A whereas one does not. The results of binding assays indicated that three of the antibodies react with the same antigenic domain (epitope 1), but the fourth with a different one (epitope 2). Probably two antigenic determinants are present in epitope 1 since one of the antibodies with specificity for epitope 1 is inactivated after iodination by the chloramine-T procedure whereas a second one is not.     

Production and Characterization of Monoclonal Antibodies against Aspartate Aminotransferase-P1 from Lupin Root Nodules

Six hybridoma clones were obtained that secreted monoclonal antibodies against the aspartate aminotransferase-P1 (AAT-P1) isoenzyme from root nodules of Lupinus angustifolius [L.] cv Uniharvest. This enzyme is found constitutively in the plant cytosol fraction. The monoclonal antibodies produced were all of the immunoglobulin G1 class, recognized two distinct epitopes on the protein, and represented the major paratopes found in the immunoglobulin fraction of sera taken from mice and rabbits immunized with the pure AAT-P1 protein. One of these epitopes was unique to lupin nodule AAT-P1. The other epitope was shown to be present on enzyme from lupin bean, white clover and tobacco leaves, lupin roots and nodules, and potato tubers. Both epitopes were recognized by the appropriate monoclonal antibodies in both their native and denatured forms. None of the monoclonal antibodies produced reacted with Rhizobium lupini NZP2257, Escherichia coli extracts, or with the inducible aspartate aminotransferase-P2 (AAT-P2) isoform also found in root nodules. A sandwich enzyme-linked immunosorbent assay utilizing two monoclonal antibodies recognizing the two distinct epitopes was developed and was capable of quantitating AAT-P1 in plant extracts. The limit of detection of AAT-P1 was less than 15 pg/mL and AAT-P1 protein could be quantified in the range 80 to 1000 pg/mL. Using this assay, AAT-P1 protein was shown to remain relatively constant during nodule development. Use of an AAT-P2-specific monoclonal antibody that inhibits the enzyme activity of this isoform enabled the direct determination of AAT-P1 enzyme activity in nodule extracts. Using these assays, specific activities of the individual isoforms were calculated; that of the AAT-P1 isoform was shown to be 7.5-fold higher than that of the AAT-P2 isoform.     

Purification of a DNA nicking-closing enzyme from mouse L cells.

A DNA nicking-closing enzyme has been purified from the nuclei of mouse L cells to 90% homogeneity. The denatured and reduced form of the enzyme has a molecular weight of 68,000 which is in agreement with the molecular weight of the native enzyme as determined by gel filtration and by sucrose sedimentation velocity assuming the protein is globular. Therefore, the active form of the enzyme is a monopolypeptide. Its isoelectric point is pH 4.2 +/- 0.2. The nicking-closing activity does not require a cofactor and does not involve any sulfhydryl group. The enzyme requires 0.2 M NaCl and pH in the range of 6.5-7.5 for optimal activity.     

Isolation of serine:glyoxylate aminotransferase from cucumber cotyledons

Serine:glyoxylate aminotransferase, a marker enzyme for leaf peroxisomes, has been purified to homogeneity from cucumber cotyledons (Cucumis sativus cv Improved Long Green). The isolation procedure involved precipitation with polyethyleneimine, a two-step ammonium sulfate fractionation (35 to 45%), gel filtration on Ultrogel AcA 34, and ion exchange chromatography on diethylaminoethyl-cellulose, first in the presence of pyridoxal-5-phosphate, and then in its absence. The enzyme was purified approximately 690-fold to a final specific activity of 34.4 units per milligram. Electrophoresis of the purified enzyme on sodium dodecyl sulfate-polyacrylamide gels revealed two polypeptide bands with apparent molecular weights of approximately 47,000 and 45,000. Both polypeptides coeluted with enzyme activity under all chromatographic conditions investigated, both were localized to the peroxisome, and both accumulated in cotyledons as enzyme activity increased during development. The two polypeptides appear not to be structurally related, since they showed little immunological cross-reactivity and gave rise to different peptide fragments when subjected to partial proteolytic digestion. Antiserum raised against either the denatured enzyme or the 45,000-dalton polypeptide did not react with any other polypeptides present in a crude cotyledonary homogenate. The purified enzyme also had alanine:glyoxylate aminotransferase activity, but was about twice as active with serine as the amino donor.