Interaction of L-canaline with ornithine aminotransferase of the tobacco hornworm, Manduca sexta (Sphingidae).

Ornithine aminotransferase (L-ornithine:2-oxo-acid aminotransferase (EC 2.6.1.13)) has been purified to homogeneity from last instar larvae of the tobacco hornworm, Manduca sexta (Sphingidae). This enzyme is a 144,000-Da tetramer constructed from 36,000-Da protomeric units. It has a high aspartate/asparagine and glutamate/glutamine content and 2 cysteine residues/subunit. All 8 cysteine residues can react with N-ethylmaleimide to inactivate the enzyme. Maintenance of the enzyme in the presence of 2-mercaptoethanol and dithiothreitol maximizes enzymatic activity and improves storage conditions, presumably by protecting these sulfhydryl groups. The apparent Km values for L-ornithine and 2-oxoglutaric acid are 2.3 and 3.2 mM, respectively. The turnover number is 2.0 +/- 0.1 mumol min-1 mumol-1. L-Canaline (L-2-amino-4-(aminooxy)butyric acid) is a potent ornithine aminotransferase inhibitor. Reaction of the enzyme with L-[U-14C]canaline produces an enzyme-bound, covalently linked, radiolabeled canaline-pyridoxal phosphate oxime. The L-[U-14C]canaline-pyridoxal phosphate oxime has been isolated from canaline-treated enzyme. Dialysis of canaline-inactivated ornithine aminotransferase against free pyridoxal phosphate slowly reactivates the enzyme as the oxime is replaced by pyridoxal phosphate. Analysis of L-[U-14C]canaline binding to ornithine aminotransferase reveals the presence of 4 mol of pyridoxal phosphate/mol of enzyme.

Kinetic analysis of glutamic dehydrogenase from larval M. sexta reveals that the K, for ammonia exceeds 400 mM, a value well above its physiological concentration (2). Glutamine is not a nitrogen donor to this insect dehydrogenase. Thus, this reaction pathway probably contributes little to glutamate biosynthesis. Glutamic acid production from pro- line is important in such tissues as adult flight muscles where proline consumption is coupled to ATP formation (3). In the larva, however, glutamate formation from proline is limited since the latter amino acid is not a major component of M. sextu hemolymph (4). Analysis of M. sextu fails to disclose significant larval glutaminase activity, thereby precluding such deamination as an important source of glutamic acid. Glutamic acid can be supplied by catabolism of hemolymph storage proteins, but Kramer et al. (5) found that the major function of these macromolecules is to provide aromatic amino acids. Larval aminotransferases other than ornithine aminotransferase (L-ornithine:2-oxo-acid aminotransferase (EC 2.6.1.13)) (OAT)' that are able to use 2-oxoglutarate can provide glutamic acid, but such larval pathways have not been evaluated. Thus, the reaction fostered by OAT, while not the sole reaction, is nevertheless an important means for glutamic acid formation in insects such as M. sex&.
Insects produce ample ornithine through arginase-directed hydrolysis of L-arginine; this enzyme is distributed widely among these arthropods (6). Since insects lack a functional Krebs-Henseleit ornithine/urea cycle (7,8), ornithine is not lost through arginine biosynthesis. In spite of the importance of OAT, this protein has not been purified previously from an insect.
L-Canaline, the L-2-amino-4-(aminooxy)butyric acid structural analog of L-ornithine, is a potent antimetabolite of higher plants (9). Indirect evidence, such as the shift in absorption spectrum that occurs when canaline reacts with an appropriate protein, suggests that this nonprotein amino acid reacts with the pyridoxal phosphate moiety of vitamin Be-containing proteins to form a stable, covalently linked complex (10,11).
The experiments detailed in this study were designed to purify and characterize OAT and to analyze its interaction with L-canaline.

Materials
Terminal instar tobacco hornworm larvae were reared as described previously (12). Unless otherwise indicated, all biochemical reagents were purchased from Sigma.

RESULTS
Characterization of M. sexta OAT:Mass-The molecular mass of the OAT subunit was obtained by SDS-polyacrylamide gel electrophoresis utilizing proteins of known mass as the standard. This procedure yielded a single band with an apparent minimum subunit molecular mass of 36,000 Da.
The molecular mass of M. sexta OAT was determined to be 148,000 f 5,000 Da by polyacrylamide gel electrophoresis under nondenaturing conditions. Molecular mass evaluation by Sephadex G-ZOO chromatography yielded a value of 152,000 f 6,000 Da. It appears that this insect OAT exists as a tetramer composed of four 36,000-Da subunits.
Amino Acid Analysis-Automated amino acid analysis of OAT disclosed its high glutamate/glutamine content. Aspartate/asparagine are the next most abundant amino acids. This protein also has numerous basic and hydrophobic amino acids. Automated amino acid analysis revealed 8 cysteine residues,   (Table II). Kinetic Parameters-Kinetic analysis of OAT indicated that the apparent K,,, values for L-ornithine and 2-oxoglutaric acid are 2.3 X 10m3 and 3.2 X 10d3 M, respectively (Fig, 3). The highest specific activity OAT gave a turnover number of 2.0 f 0.1 rmol/min/pmol OAT.  Canaline-dependent inactivation of OAT was determined as described under "Canaline Inhibition Studies." 0, control; A, 0, 0, and A, lo-', 10e7, 10m6, and 10m5 M canaline, respectively.
content of OAT was determined by reaction with L-[U-W] canaline and evaluation of the radiolabeled protein as described. These determinations established that M. senta OAT contained 4 mol of pyridoxal phosphate/m01 of enzyme.

OAT-Canaline
Interaction-Treatment of OAT with canaline caused a rapid and precipitous loss of activity (Fig. 4). At 10e6 M canaline, for example, nearly three-fourths of the enzymatic activity was lost after a 5-min exposure at 21 "C. Treatment with as little as lo-' M canaline manifested a significant reduction in activity. This is an interesting observation since when OAT was exposed to lo-' M canaline, there was a lOOO-fold excess of free pyridoxal phosphate relative to canaline in the reaction mixture (pyridoxal phosphate is required to protect the enzyme). Other workers (11,25) have noted the ability of canaline to react preferentially with the pyridoxal phosphate moiety of rat liver OAT in the presence

Ornithine
Aminotransferase-L-Canaline Interaction of a large excess of free pyridoxal phosphate. Evaluation of the second-order rate constant for the formation of the canaline-pyridoxal phosphate oxime disclosed that the reaction rate was more than 16 times greater for OAT in the complete buffer system as compared to free pyridoxal phosphate (Table  III). The rate constant fell from 4380 M-' min-' for the complete buffer system to 2060 M-' min-' when both pyridoxal phosphate and mercaptans were absent. In the presence of buffered mercaptans, the rate constant was 3215 M-' min-' as compared to 2895 M-' min-' for buffered pyridoxal phosphate (Table III).
Formation of Canaline-Pyridoxal Phosphate Oxime-To establish that canaline reacted with the pyridoxal phosphate moiety of OAT, the enzyme was treated with [14C]canaline, and the radiolabeled canaline-pyridoxal phosphate oxime was released by dialysis against deionized water. The radiolabeled oxime was analyzed by automated amino acid analysis as described. The ['4C]canaline-pyridoxal phosphate oxime possessed a column retention time of 87 min. The retention time for the chemically prepared canaline-pyridoxal phosphate oxime is also 87 min. Analysis for Y! in the spent ninhydrin, following the method of Rosenthal (26), indicated that virtually all of the dialysate 14C eluted with the canaline-pyridoxal phosphate oxime. Thus, it was possible to isolate the ['4C]canaline-pyridoxal phosphate oxime created by treating OAT with ['4C]canaline.
The canaline-pyridoxal phosphate complex was dissolved in 'HZ0 and analyzed by NMR using tetramethylsilane as an internal standard. Such analysis disclosed the anticipated downfield shift of the (Y, /3, and y canaline protons (6 = 0.12, 0.2, and 0.7 ppm, respectively).
The resonance signal at 8.9 ppm, due to the methine proton of the Schiff-base linkage between canaline and pyridoxal phosphate, was readily discernible.
Further identification of the canaline-pyridoxal phosphate oxime was achieved by discovery of its appreciable fluorescence after excitation at 380 nm (Fig. 5) Fig. 5. These analyses revealed that the emission pattern of the OAT-derived radiolabeled canaline-pyridoxal phosphate complex was identical to that of the chemically prepared material. Under these experimental conditions, neither canaline nor pyridoxal phosphate possessed detectable fluorescence over the range of 400-600 nm.
Sulfhydryl Inactivation-The essentiality of the sulfhydryl groups of M. sexta OAT was demonstrated by reaction with N-ethylmaleimide.
As shown in Fig. 6, this reagent elicited a  concentration-dependent inactivation of OAT. Amino acid analysis and titration of OAT with 5,5'-dithiobis-(2-nitrobenzoic acid) following the method of Ellman (27) indicated that the tetramer contained 8 cysteines, presumably Z/subunit. Treatment of OAT with 6 M guanidine, prior to reaction with the Ellman reagent, failed to increase the cysteine titer. Thus, all of the cysteines were exposed and readily available to this reagent.

Ornithine
aminotransferase isolated from larval M. sexta exists in a single discernible form: a 144,000-Da tetramer composed of four 36,000-Da monomers. This contrasts with rat liver OAT as studied by Sanada et al. (28). The mammalian There is much homology as well as significant differences in the primary structure of M. sextu OAT and those of the rat liver and kidney proteins. Glutamic acid and leucine are the most abundant residues in the mammalian enzymes, whereas histidine, methionine, and tryptophan are the least common (30). The cysteine content of liver and kidney OATS is 16 and 14, respectively.
Whereas glutamic acid is also the most abundant residue in the insect protein, the leucine level is low and is exceeded by the isoleucnie content. Histidine and tryptophan are limited, but the methionine content is higher, exceeding that of serine and phenylalanine. The thiol group content is about half of that for the mammalian enzymes (Table II).
[Y!]Canaline provides an effective tool for determining the pyridoxal phosphate content of this insect enzyme. '%-Labeled canaline reacts stoichiometrically with the pyridoxal phosphate of OAT to form a stable, radiolabeled oxime that can be removed by simple dialysis.
[14C]Canaline-protein binding analysis may provide a sensitive and convenient radiometric assay for determining the pyridoxal phosphate content of vitamin B,-containing proteins.
Reaction of canaline with OAT to form a canaline-pyridoxal phosphate oxime rapidly inactivates OAT. This oxime can be removed by dialysis to create an apoenzyme that can react with free pyridoxal phosphate to regenerate a catalytically competent macromolecule (Table IV). These experiments provide the first direct chemical evidence that canaline inactivates pyridoxal phosphate-containing proteins by forming an oxime with the cofactor that block its catalytic activity.
Analysis of the rate constant for the reaction of canaline with OAT suggests that when the enzyme is protected by pyridoxal phosphate and mercaptans, thereby most effectively stabilizing the native conformation, it is most reactive with canaline (Table III). Interestingly, 2-mercaptoethanol-and dithiothreitol-protected OATS in buffer yield a more reactive OAT than when the enzyme is surrounded by free pyridoxal phosphate. OAT rapidly loses its enzymatic activity if maintained under mercaptan-free conditions. The importance of stabilizing the cysteinyl residues is revealed by the dramatic loss in activity upon treatment with N-ethylmaleimide. The sensitivity of M. sexta OAT to this reagent undoubtedly results from the fact that all of the cysteines are exposed at the surface and are able to react readily with N-ethylmaleimide.