Metabolism of Homocysteine-thiolactone in Plants*

Editing of the amino acid homocysteine (Hcy) by certain aminoacyl-tRNA synthetases results in the formation of an intramolecular thioester, Hcy-thiolactone. Here we show that the plant yellow lupin, Lupinus luteus, has the ability to synthesize Hcy-thiolactone. The inhibition of methylation of Hcy to methionine by the anitifolate drug aminopterin results in greatly enhanced synthesis of Hcy-thiolactone by L. luteus plants. Methionine inhibits the synthesis of Hcy-thiolactone in L. luteus, suggesting involvement of methionyl-tRNA synthetase. Consistent with this suggestion is our finding that the plant Oryza sativa methionyl-tRNA synthetase, expressed in Escherichia coli, catalyzes conversion of Hcy to Hcy-thiolactone. We also show that Hcy is a component of L. luteus proteins, most likely due to facile reaction of Hcy-thiolactone with protein amino groups. In addition, L. luteus possesses constitutively expressed, highly specific Hcy-thiolactone-hydrolyzing enzyme. Thus, Hcy-thiolactone and Hcy bound to protein by an amide (or peptide) linkage (Hcy-N-protein) are significant components of plant Hcy metabolism.

Homocysteine (Hcy) 1 -thiolactone, a cyclic thioester of Hcy, was discovered by serendipity almost 70 years ago as a byproduct of the digestion of methionine with hydriodic acid, a procedure used then for the determination of protein methionine (1). The discovery of an error editing reaction of aminoacyl-tRNA synthetases, in which Hcy is converted to Hcythiolactone, highlighted the biological significance of Hcy-thiolactone (2). Hcy-thiolactone is synthesized by methionyl-tRNA synthetase (MetRS) in bacterial (3)(4)(5)(6), yeast (6 -8), and mammalian, including human, cells (9 -17). Isoleucyl-tRNA synthase and leucyl-tRNA synthase, in addition to MetRS, synthesize Hcy-thiolactone from exogenous Hcy, at least in bacteria (5). Hcy-thiolactone forms in a two-step reaction driven by the hydrolysis of ATP (2). In the first step, MetRS catalyzes reaction of Hcy with ATP, which leads to the formation of a MetRS-bound homocysteinyl adenylate.
MetRS ϩ Hcy ϩ ATPNMetRS ⅐ Hcy ϳ AMP ϩ PP i REACTION 1 In the second step, MetRS catalyzes the reaction of the side chain thiolate of Hcy, which displaces the AMP moiety from the activated carboxyl group of Hcy; Hcy-thiolactone is a product of this reaction (Reaction 2).
Although its role in cell physiology is largely unknown, Hcythiolactone has been suggested to be a positive effector of the stationary phase response in Escherichia coli (21) and is also likely to be involved in the regulation of methionine synthase gene expression in E. coli (6).
In humans, Hcy-thiolactone is likely to play a role in cardiovascular disease due to its ability to form Hcy-N-protein, which leads to protein damage (10 -17, 20). A protein component of high-density lipoproteins, Hcy-thiolactonase/paraoxonase, detoxifies Hcy-thiolactone, thereby minimizing formation of Hcy-N-protein in humans (22,23).
Whether Hcy-thiolactone is present and how it is metabolized in plants was unknown. Here we report that Hcy-thiolactone and Hcy-N-protein are components of Hcy metabolism in yellow lupine (Lupinus luteus). We also show that Hcy-thiolactone is synthesized by rice (Oryza sativa) methionyl-tRNA synthetase and degraded by a highly specific yellow lupine Hcythiolactone hydrolase. Lupine Seed Germination and 35 S Labeling Conditions-Yellow lupine (L. luteus, var. Juno) seeds were germinated at 21°C on cellulose paper towels soaked with deionized sterile water. On the 6 th day, the roots were removed, and the seedlings were transferred into 50-ml Falcon tubes (5 seedlings/tube) containing 1 ml of sterile water or 25 M aminopterin (Sigma-Aldrich) in water. Hypocotyls tips were immersed in the liquid medium. After 48 h, the seedlings were transferred for 12 h to fresh tubes containing 1.5 M [ 35 S]Met or [ 35 S]Hcy (15 Ci in 0.5 ml of sterile water). 35 S-amino acids were used as tracers to facilitate monitoring of Hcy-thiolactone during purification; quantification was by measurements of A 240 (see below). Antifolate drugs, aminopterin, sulfonamide, or trimethoprim (from Sigma-Aldrich), were included as indicated under "Results." The hypocotyls and coltyledons were then collected separately and frozen at Ϫ20°C.

MATERIALS AND METHODS
Preparation of L. luteus Extracts-Yellow lupine hypocotyls or cotyledons (ϳ1 g) were ground up at 0°C with 1 ml of 50 mM potassium phosphate buffer, pH 7.5, using a mortar and pestle. The extracts were centrifuged at 30,000 ϫ g using a JA25.50 rotor in a Beckman-Coulter J2 centrifuge (15 min, 2°C).
Determination of Hcy-N-protein-Proteins, extracted from hypocotyls of 6-day-old yellow lupine seedlings, were treated with 5 mM DTT for 5 min at room temperature and precipitated with 80% ethanol at 0°C to remove free Hcy. The plant protein was dissolved in phosphatebuffered saline containing 5 mM DTT and precipitated with 80% ethanol. The cycle of DTT treatment and ethanol precipitation was repeated four more times. This procedure removed Ͼ99% total Hcy from plant protein extracts. Samples of DTT-treated protein were diluted to 0.1 ml with 25 mM DTT and transferred to glass ampoules (1-ml volume) containing 0.1 ml of 12 N HCl. The ampoules were sealed under vacuum, and the samples were hydrolyzed at 120°C for 1 h. This procedure quantitatively converted Hcy-N-protein into Hcy-thiolactone. After hydrolysis, samples were lyophilized and dissolved in 10 l of water, and 3.3-l aliquots were subjected to two-dimensional TLC on 6.7 ϫ 5-cm cellulose plates (Analtech) as described previously (26). [ 35 S]Hcy-thiolactone, localized on TLC plates by autoradiography using Kodak BioMax x-ray film, was extracted with water (60 l) and finally purified and quantified by cation exchange HPLC (19,26). To determine Hcy-N-protein relative to protein methionine in lupine plants, radiolabeled spots corresponding to Hcy-thiolactone and methionine were cut out from a duplicate set of TLC separations and quantified by using a Beckman LS6500 scintillation counter.
Determination of Total Hcy-The principle of the procedure involves the conversion of Hcy to Hcy-thiolactone, which is then quantified by HPLC (19). Plant extracts were treated with 5 mM DTT to convert disulfide-bound forms of Hcy to free reduced Hcy, deproteinized by ultrafiltration through Millipore 10-kDa cut-off membranes at 4°C. The ultra-filtrate (50 l) was lyophilized on a SpeedVac concentrator and dissolved in 6 l of 50 mM DTT, and Hcy was converted to Hcythiolactone by treatment with 3 l of 6 M HCl for 30 min at 100°C. After lyophylization, samples were dissolved in 50 l of water and subjected to HPLC. HPLC Chromatography-HPLC analyses were carried out using a cation exchange PolySULFOETHYL Aspartamide column (2.1 ϫ 200 mm, 5 , 300 Å) from PolyLC, Inc. and System Gold Noveau HPLC instrumentation from Beckman-Coulter as described previously (19,26). Solution A (10 mM mono-sodium phosphate) and solution B (200 mM NaCl in 10 mM mono-sodium phosphate) were used as solvents. After application of sample, the column was eluted with a linear gradient from 50% to 100% solution B for 5 min, followed by 100% solution B for 0.5 min and 2-min reequilibration with 50% solution B.
The effluent was monitored at multiple wavelengths, including A 240 , the UV absorption maximum of Hcy-thiolactone (⑀ ϭ 3,500 M Ϫ1 cm Ϫ1 ) (5,15). For each sample, the identity of the eluted material as Hcythiolactone was confirmed by its co-migration with an authentic Hcythiolactone, by its characteristic absorbance spectrum with a maximum at A 240 , and by its sensitivity to lupine Hcy-thiolactonase or NaOH. The detection limit was 5 pmol of Hcy-thiolactone.
Spectrophotometric assays were used in substrate specificity studies with nonradiolabeled substrates (all from Sigma). Hydrolysis of Hcythiolactones was determined from the decrease of their characteristic UV absorption at ϭ 240 nM (⑀ ϭ 3,500 M Ϫ1 cm Ϫ1 ) (22). Hydrolysis of phenyl acetate and p-nitrophenyl acetate was determined spectrophotometrically using ⑀ ϭ 1,300 M Ϫ1 cm Ϫ1 at 270 nm for phenol and ⑀ ϭ 13,000 M Ϫ1 cm Ϫ1 at 412 nm for p-nitrophenol, respectively. Hydrolysis of diethyl p-nitrophenyl phosphate (paraoxon) was measured spectrophotometrically using ⑀ ϭ 13,000 M Ϫ1 cm Ϫ1 at 412 nm for p-nitrophenol (22).
In experiments in which utilization of other (thio)esters (10 mM) by Hcy-thiolactonase was tested, potential substrates and products were separated by TLC and visualized by staining with ninhydrin or under UV. With all potential substrate-product pairs, complete separation was achieved on cellulose plates (Analtech) using 1-butanol/acetic acid/ water (4:1:1, v/v) as a solvent. Complete separation of acetyl-S-CoA (Sigma) and Met-S-CoA (prepared as described in Ref. 27) thioesters from free CoA-SH was achieved on polyethyleneimine-cellulose plates (Sigma) using 1.2 M LiCl as a solvent.
Yellow lupine seed (L. luteus, var. Juno) meal (100 g) was extracted with 300 ml of Buffer A. Protein (9,030 mg) in crude extract, obtained by centrifugation at 20,000 ϫ g for 30 min, was fractionated with ammonium sulfate. Protein (2,320 mg) precipitated between 0 -35% ammonium sulfate saturation was collected by centrifugation, dissolved in 5 ml of Buffer B, and extensively dialyzed against Buffer B. Dialysate was clarified by centrifugation and applied on DEAE-Sephacel column. The column was washed with 5 volumes of Buffer B and eluted with a KCl gradient in Buffer B. Protein fractions with Hcy-thiolactonase activity (60 mg), eluting at 0.3-0.35 M KCl, were concentrated by ammonium sulfate precipitation, dissolved in 5 ml of Buffer C, and further purified by gel filtration on a Superdex 200 column equilibrated with Buffer C. The enzyme eluted from the gel filtration column as a protein with a molecular mass of 55 kDa. Active fractions (0.7 mg of protein) were applied on a hydroxylapatite column equilibrated with Buffer A and eluted with a gradient 10 -100 mM phosphate, pH 6.8, in Buffer A. Active fractions, eluted at 75 mM phosphate, were concentrated, dialyzed against Buffer A, and stored at Ϫ20°C. The Hcy-thiolactonase preparation (0.2 mg of protein) had a specific activity of 536 mol/mg/h and was purified 25,000-fold. The purified enzyme preparation showed several protein bands on SDS-PAGE.

Synthesis of Hcy-thiolactone in Yellow Lupine Seedlings
Increases upon Depletion of Tetrahydrofolate-In plants, Hcy is synthesized de novo from sulfate and also as a by-product of cellular methylation reactions ( Fig. 1) (28,29). Three pathways of further Hcy metabolism are utilized to different extents by living organisms: methylation to methionine, trans-sulfuration to cysteine, and conversion to Hcy-thiolactone. In plants, Hcy is further metabolized by methylation to methionine by a methyltetrahydofolate-dependent methionine synthase (Fig. 1, MS) (29) or by S-methyl-methionine-dependent Hcy S-methyltransferase (29 -31). Trans-sulfuration of Hcy to cysteine, present in fungi and mammals, is absent in plants (29). MetRS-dependent metabolism of Hcy to Hcy-thiolactone, present in bacteria, yeast, and mammalian cells (15)(16)(17), was not known to be present in plants.
To determine whether metabolism of Hcy to Hcy-thiolactone occurs in plants, yellow lupine seedlings were examined for the presence of Hcy-thiolactone and Hcy-thiolactone hydrolase. Before extraction of Hcy-thiolactone, 6-day-old seedlings were maintained for additional 60 h on water in the absence and presence of the antifolate drug aminopterin (25 M), an inhibitor of eukaryotic dihydrofolate reductase enzymes (32). To facilitate purification of Hcy-thiolactone from plant tissues, yellow lupine seedlings were metabolically labeled with radiotracers [ 35 S]Met or [ 35 S]Hcy for 12 h before harvesting. As shown in Figs. 2B and 3, Hcy-thiolactone was present in hypocotyls of yellow lupine seedlings. Although exogenous 35 Samino acids were taken up by seedlings and metabolized to [ 35 S]Hcy-thiolactone intracellularly (Fig. 2B), their contribution to total Hcy-thiolactone, measured by A 240 (Fig. 3) synthesis, was Ͻ0.1%. Lupine tissue concentrations of Hcy-thiolactone were 49.5 M and Ͻ0.6 M in the presence and absence of aminopterin, respectively (Table I). The presence of up to 1 mM exogenous [ 35 S]Hcy in culture medium did not increase Hcythiolactone synthesis by the plant seedlings in the presence or absence of aminopterin (not shown). Treatment of seedlings with aminopterin also increased the plant tissue total Hcy level from 4.3 M in the absence of aminopterin to 245 M in the presence of aminopterin. Hcy-thiolactone represented ϳ20% of total Hcy concentration in hypocotyls of yellow lupine seedlings grown in the presence of aminopterin. The growth of 6-day-old lupine seedlings was reduced in the presence of aminopterin.
Other antifolates, such as trimethoprim (0.1 mM) or sulfonamide (5 mM), did not reduce growth and did not affect Hcythiolactone or total Hcy levels in the plants. This suggests that trimethoprim, an inhibitor of bacterial dihydrofolate reductase enzymes (33), does not inhibit plant mitochondrial dihydrofolate reductase. The observation that Hcy or Hcy-thiolactone levels do not increase in the presence of sulfonamide, an inhibitor of de novo folate synthesis (32), suggests that endogenous methyltetrahydrofolate pools in lupine seedlings are not significantly depleted during growth. This suggestion is consistent with a study of one carbon fluxes in Arabidopsis thaliana, which indicated that cellular folate pools have a relatively long half-life in this plant (33).
Because of its mostly neutral character under physiological pH (18), Hcy-thiolactone is expected to diffuse out from lupine seedlings. Indeed, we have found that Hcy-thiolactone was excreted from seedlings grown in the presence of aminopterin. When yellow lupine seedlings were labeled with [ 35 S]Hcy in the presence of aminopterin, the amount of [ 35 S]Hcy-thiolactone excreted into medium in which seedlings were maintained represented 30% of the amount formed in hypocotyls (data not shown). In the absence of plant seedlings, no [ 35 S]Hcy-thiolactone was formed in the aminopterin-containing medium.
MetRS Is Involved in the Synthesis of Hcy-thiolactone in Plants-To determine whether plant MetRS metabolizes Hcy to Hcy-thiolactone, rice MetRS was expressed in E. coli BL21 harboring pET/MOs⌬C, a plasmid bearing the rice MetRS gene under the control of the lac promotor (25). The rate of Hcythiolactone synthesis in the culture of E. coli BL21/pET/ MOs⌬C increased about 3-fold upon induction with isopropyl-␤-D-thiogalactopyranoside (Fig. 4), which indicates that rice MetRS catalyzes the synthesis of Hcy-thiolactone. Supplementation of growth medium with methionine resulted in inhibition of Hcy-thiolactone synthesis in these cultures, as expected (Fig. 4).
To determine whether MetRS is involved in Hcy-thiolactone synthesis in plants in vivo, yellow lupine seedlings were maintained on aminopterin in the presence of increasing concentrations of methionine. In the presence of 0.1 and 1 mM methio- nine, the plant tissue levels of Hcy-thiolactone dropped to 60% and 6%, respectively, of the levels observed in the absence of methionine (Table II). The inhibition by methionine is consistent with the involvement of MetRS in the synthesis of Hcythiolactone in lupine seedlings.
Hcy Is Present in Yellow Lupine Proteins-To determine whether Hcy is present in plant proteins, proteins from yellow lupine seedlings were extracted and depleted of free and disulfide forms of Hcy by treatments with DTT and precipitation with ethanol. Lupine proteins were then hydrolyzed with HCl in the presence of DTT. Under these conditions, Hcy, linked to protein by amide linkage (Hcy-N-protein), is converted to Hcythiolactone (10,11,26). As shown in Table I (Table I).
When seedlings were labeled with [ 35 S]Hcy in the presence of aminopterin, the conversion of [ 35 S]Hcy to [ 35 S]Met-protein was inhibited 85%, compared with the conversion in the absence of aminopterin (Table III) (Table III).
Hcy-thiolacone Hydrolase Metabolizes Hcy-thiolactone in Yellow Lupine Plants-To determine whether plants have the ability to metabolize Hcy-thiolactone, yellow lupine seedlings were maintained on 0.75 M [ 35 S]Hcy-thiolactone in water. The seedlings metabolized 60% and 100% [ 35 S]Hcy-thiolactone after 7 and 24 h, respectively. Analysis of plant extracts showed that [ 35 S]Met was a major metabolite derived from [ 35 S]Hcythiolactone after 24 h (data not shown). In the absence of yellow lupine seedlings, Hcy-thiolactone was stable under the experimental conditions utilized (half-life Ͼ 3 days). Because Hcythiolactone is unlikely to be metabolized without ring opening, its fast metabolism suggests that Hcy-thiolactone-hydrolyzing enzyme is present in plants.
Indeed, when crude extracts from yellow lupine seeds were incubated with [ 35 S]Hcy-thiolactone, it was hydrolyzed to [ 35 S]Hcy at a rate of 0.21 mol/mg/h. This level of Hcy-thiolactone-hydrolyzing activity is 3.7-fold higher than the level present in human serum (22). Lupine Hcy-thiolactonase activity, measured in extracts from cotyledons, did not change significantly after seed germination and growth up to 6 days (data not shown).
The Hcy-thiolactonase activity, precipitated from crude extracts of yellow lupine seed meal with 35% ammonium sulfate, was further purified by anion exchange chromatography on DEAE-Sephacel, gel exclusion chromatography on Superdex, and absorption chromatography on hydroxylapatite. At all steps of purification procedure, a single peak of Hcy-thiolactonase activity was observed, suggesting that a single enzyme was responsible for Hcy-thiolactone hydrolysis in yellow lupine. The specific activity of the purified Hcy-thiolactonase preparation, 536 mol/mg/h, was 25,000-fold greater than that measured in crude extracts. The plant Hcy-thiolactonase preparation exhibited 7-fold higher specific activity than pure human Hcy-thiolactonase.
Examination of the substrate specificity showed that, in addition to L-Hcy-thiolactone, the purified enzyme also hydrolyzed ␣-aminoacyl esters and thioesters (Table IV). For example, thioesters of methionine, such as Met-S-CoA and Met-S-DTT, and methionine methyl esters were hydrolyzed. Esters of other ␣-amino acids, such as methyl esters of alanine, cysteine, phenylalanine, tryptophan, and lysine, were also hydrolyzed. D-Hcy-thiolactone and D-forms of ␣-aminoacyl esters were hydrolyzed up to 20-fold less efficiently than the L-forms. L-Homoserine-lactone was also a substrate. However, N-acetyl-D,L-Hcy-thiolactone was not hydrolyzed. Esters of ␤-amino acids, such as ␤-Ala methyl ester, esters and thioesters of acetic acid, such as O-acetyl-L-serine and acetyl-S-CoA, and ␥-methyl ester of glutamic acid were not hydrolyzed. In contrast to human Hcy-thiolactonase/paraoxonase, the plant enzyme did not hydrolyze non-natural aryl esters, such as phenyl acetate and p-nitrophenyl acetate, or the organophosphate paraoxon (Table IV).
The plant Hcy-thiolactone-hydrolyzing enzyme eluted from a Superdex gel filtration column as a 55-kDa protein. The enzyme exhibited a broad pH optimum, from pH 6 to pH 8, and did not require calcium or any other divalent cation for activity. The K m value for L-Hcy-thiolactone was 45 mM. Taken together, these data indicate that the plant enzyme is a novel Hcy-thiolactonase, fundamentally different from human Hcy-thiolactonase/paraoxonase. DISCUSSION This work demonstrates a novel aspect of Hcy metabolism in plants: synthesis and degradation of Hcy-thiolactone in the plant yellow lupine (Fig. 1). In the synthetic pathway, Hcy is converted to Hcy-thiolactone by the plant MetRS. In the degradation pathway, Hcy-thiolactone is hydrolyzed to Hcy by a unique plant Hcy-thiolactone hydrolase.
Hcy-thiolactone reacts easily with protein lysine residues under physiological conditions (20). This reaction is responsible for the presence of Hcy in endothelial cell proteins (11)(12)(13)(14)(15) and, most likely, in human blood proteins (26). Our data demonstrate that Hcy is also present in yellow lupine proteins. When methylation of Hcy to methionine synthase was inhibited by the antifolate drug aminopterin, Hcy-N-protein became a major metabolite of Hcy in yellow lupine seedlings. The presence of Hcy-N-protein in mammals (26) and plants suggests that Hcy-N-protein is likely to be a component of Hcy metabolism in multicellular organisms.
Incorporation of Hcy into protein mediated by Hcy-thiolactone is known to result in protein damage (14 -17, 20). Because of this, the ability to detoxify Hcy-thiolactone is essential for biological integrity, particularly in multicellular organisms. Indeed, specific Hcy-thiolactone hydrolase/paraoxonase, tightly associated with high density lipoprotein, exists in mammals, including humans (22,23). Our present work shows that yellow lupine plants possess a novel Hcy-thiolactone-hydrolyzing enzyme. The plant Hcy-thiolactonase is different from the human Hcy-thiolactone-hydrolyzing enzyme. For example, the plant Hcy-thiolactonase does not require calcium for activity, whereas the human Hcy-thiolactonase/paraoxonase does (22,23). Although both enzymes hydrolyze Hcy-thiolactone, they differ in their ability to hydrolyze other (thio)esters. For example, whereas the plant enzyme hydrolyzes ␣-aminoacyl (thio)- Assays were carried out for 1 h at 30°C in 10-l reaction mixtures containing 10 mM indicated compound, 0.1 M potassium phosphate buffer, pH 7.2, and yellow lupine Hcy-thiolactonase. Reaction products were analyzed by TLC on cellulose or polyethyleneimine-cellulose plates. esters, the human enzyme does not (Table IV). On the other hand the plant enzyme does not hydrolyze phenyl and p-nitrophenyl esters of acetic acid or the organophosphate paraoxon. These artificial esters are very good substrates of the human enzyme (Table IV).
In conclusion, our findings show that two novel pathways of Hcy metabolism are utilized by the plant L. luteus: 1) metabolic conversion of Hcy to Hcy-thiolactone, a fundamental editing reaction in protein synthesis, which appears to be conserved in all living organisms; and 2) hydrolysis of Hcy-thiolactone to Hcy, which thus far has been documented in multicellular organisms such as mammals and plants.