Novel Malonamidases in Bradyrhizobium japonicum PURIFICATION, CHARACTERIZATION, AND IMMUNOLOGICAL COMPARISON*

Three novel malonamidases (Ela, Elb, and E2) occur- ring constitutively in Bradyrhizobium japonicum were purified to electrophoretic homogeneity. They were found to catalyze one or more of the following three types of reactions: malonyl transfer to hydroxylamine (reaction l), hydroxaminolysis of malonamate (reaction 2), and hydrolysis of malonamate (reaction 3). The mo- lecular sizes of Ela, Elb, and E2 were 126,107, and 103 m a , respectively, and they were each composed of two identical subunits. The PIS of Ela and Elb, 5.5 and 5.0 respectively, were similar, but that of E2 was 7.2. Opti- mum pH values varied with the type of reactions catalyzed, but among the enzymes they were found to be similar. The affinity of E2 for malonamate was about 30-and 70-fold higher than that of Ela and Elb, respec- tively. Acetate and propionate inhibited Ela activity competitively, whereas malonate inhibited E2 activity noncompetitively. The amino acid composition and N-terminal amino acid sequence of the three enzymes were found to be different. These enzymes were also immuno-logically different. Ela was found to form a malonyl-enzyme intermediate during the catalysis through the isolation of [14Clmalonyl-enzyme with gel filtration and through isotope exchange experiments with [‘*O]ma- lonate. These malonamidases may play a role

min at 37 "C. After the reaction, the malonohydroxamate formed was determined by the same method described previously for malonyl-CoA synthetase (11). The second method was based on hydroxaminolysis of malonamate (Reaction 2). Reaction mixtures containing (in pmol) Tris-HCI, pH 8.0, 50; malonamate 5; neutralized NHzOH 100; a n enzyme and water in a total volume on 0.5 ml were incubated for 30 min at 37 "C. The malonohydroxamate was determined by the method described above. The third method was based on the determination of the concentration of ammonium ion generated from hydrolysis of malonamate (Reaction 3), which was made by Nessler's method (12) or by phenol-hypochlorite reaction (13). One unit of the enzyme is defined to be the amount of the enzyme which catalyzes the formation of 1 pmol of malonohydroxamate or ammonium iodminute at 37 "C.
Growth of Bacteria-B. japonicum USDA 110 was used for enzyme purification. The organism was maintained of YM (0.05% potassium phosphate, pH 6.8,0.02% MgSO, 7H20, 1% mannitol, 0.02% NaCl, and 0.1% yeast extract) medium plate containing 2% Bacto-agar. Cells were grown in GYP medium (1% glucose instead of mannitol in YM medium) a t 30 "C with vigorous shaking in a 500-ml baffled-flask for 6 days. Glucose was separately sterilized and added to the medium.
Preparation of Crude Extract-All purification procedures were performed at 4 "C. B. japonicum organisms (30 g wet weight) harvested and kept frozen a t -70 "C were thawed and resuspended in 75 ml of buffer A (20 mM Tris-HC1, pH 7.4, containing 15% glycerol). The preparation of crude extract followed by the first AfEGel blue column chromatography was performed by the method described previously with a slight modification (14). The adsorbed proteins were sequentially eluted by the application of a 400-ml linear gradient of NaCl (0-2 M) in buffer A and a 300-ml stepwise gradient of 2 M NaCl in buffer B (20 mM Tris-HC1, pH 7.4). Fractions (4 ml each) were collected, and the adivities of malonamidases were assayed. Chromatography yielded two different malonamidase activity peaks. One malonamidase was identified in the breakthrough fractions (designated as malonamidase El), and another malonamidase was eluted with 2 M NaCl in buffer B (designated as malonamidase E2). Fractions containing malonamidase El and E2 were pooled separately and subjected to subsequent purification procedures.
Purification of Malonamidase El-The breakthrough fractions obtained from first AfE-Gel blue were refractionated by ammonium sulfate precipitations. E l was precipitated in the protein fractions obtained between 40-75% saturation of ammonium sulfate. The protein precipitate was collected by centrifugation at 20,000 x g for 30 min and was dissolved in about 15 ml of buffer C (10 mM Tris-HC1 buffer, pH 7.4, containing 10 mM MgClz and 15% glycerol) and was dialyzed against 2 liters of the same buffer overnight. The insoluble materials were removed by centrifugation. The enzyme solution obtained from the previous step was applied to Sephadex G-100 column (4 x 86 cm) preequilibrated with buffer C. Fractions (3 ml each) were collected and assayed for E l activity, which was also found right after turbid breakthrough fractions. The fractions containing El activity were mixed and loaded onto a DEAE-Sephacel column (3 x 4.5 cm) equilibrated previously with buffer C. After loading was completed, unadsorbed materials were washed with about 150 ml of buffer C. Adsorbed proteins were consecutively eluted with two gradients, the first being a 200-ml linear gradient of sodium malonate (0-0.04 M) in buffer C and the second a 100-ml linear gradient of NaCl (0-0.4 M) in buffer C. Fractions (3 ml each) were collected and assayed for El activity. Two E l enzymes were separated by differential elution from this column. One E l was eluted a t about 15 ~IM sodium malonate (designated as Ela) and another E l was eluted a t about 150 mM NaCl (designated as Elb). Fractions containing E l a and E l b activity were pooled separately and dialyzed against buffer D (10 m~ potassium phosphate, pH 6.5, containing 15% glycerol). The dialyzed solution containing E l a activity was applied to an Mi-Gel blue column (2 x 4 cm) pre-equilibrated with buffer D. After washing the column with -50 ml of buffer D. Ela was eluted by the application of a 200-ml linear gradient (0-15 m~) of ATP in buffer D. Ela was eluted at 5-7 mM ATP. The fractions containing E l a activity from the previous step were pooled and slowly loaded onto a Red-120 agarose column (1.6 x 4 cm) pre-equilibrated with buffer D (about 9 ml/min). After loading was completed, the unadsorbed materials were washed one after another with 50 ml of buffer D and 30 ml of buffer D containing 10 m~ ATP. Ela was eluted by the application of a 200-ml linear gradient (0-0.3 M) of sodium malonate in buffer D. The fractions containing E l a were pooled, dialyzed against buffer A, and concentrated. This E l a solution was stored a t 4 "C and used for all subsequent experiments.
The enzyme solution containing E l b activity obtained from DEAE-Sephacel was loaded onto a hydroxyapatite column (3 x 1 cm) pre-equilibrated with buffer D. E l b was eluted during the washing with buffer D. The fractions containing E l b were pooled, dialyzed against buffer A, and concentrated. This E l b solution was stored at 4 "C and used for all subsequent experiments. Purification of Malonamidase E2-ZnS04 was added to the enzyme solution containing E 2 activity obtained from first Ml-Gel blue to make a final concentration of 0.5 m~. This solution was applied to phenyl-Sepharose CL-4B (2.8 x 4 cm) equilibrated with buffer B containing 2 M NaCl and 0.5 m~ ZnS04. After loading was completed, the unadsorbed materials were washed with 100 ml of buffer B containing 2 M NaCl and 2 mM ZnSO, followed with 50 ml of buffer B containing 2 M NaCl and 2 m~ EDTA. E2 was eluted by the application of a 300-ml linear descending gradient (2.0-0 M) of NaCl in buffer B containing 2 mM EDTA.
Fractions containing the enzyme activity were pooled and dialyzed twice against 2 liters of buffer B. The dialyzed solution containing E 2 activity was loaded onto a DEAE-Sephacel (2.8 x 2.5 cm) column preequilibrated with buffer B. After washing the column with the same buffer, E2 was eluted by the application of a 200-ml linear gradient (0-0.5 M) of NaCl in the buffer B. E 2 was eluted at 0.2-0.25 M NaCl. The fractions containing the E2 activity from the previous step were pooled and applied to Red-120 agarose (1.8 x 3.5 cm) pre-equilibrated with buffer B. As soon as the loading was completed, the adsorbed proteins were eluted by the application of a linear gradient of NaCl (0-2.0 M) in buffer B. E2 was eluted a t 1.7-1.9 M NaC1. Fractions containing the enzyme activity were pooled, dialyzed against buffer B, and concentrated. This E2 solution was stored a t 4 "C and used for all subsequent experiments.
Activity Staining-Native PAGE was performed as described above. The gel was soaked in 0.1 M MOPS, pH 6.8, buffer for 10 min with gentle agitation. The gel was transferred to a reaction mixture containing (pmol) MOPS buffer, pH 6.8,50; sodium malonate, 60 (in the case ofE2, malonamate, 5); neutralized NH,OH, 100; and water in a total volume of 10 ml. After incubation for 1 h a t 37 "C, the gel was fixed in 10% trichloroacetic acid solution for 10 min. After discarding the trichloroacetic acid solution, the gel was immersed in the same developing solution described in the assay procedure above. The dark-brown colored band, which appeared within 10-20 min as a result of malonamidase reaction, remained for about 30 min, and diffused out. The developed gel was photographed immediately and stained with Coomassie Blue for the demonstration of a corresponding protein band.
Protein Determination-The protein content of the extract and the fractions from all steps of purification were determined using the Sigma bicinchoninic acid protein assay kit with bovine serum albumin as standard.
Preparation of Substrates-a-Ketoglutaramate and a-ketosuccinamate were prepared by the oxidation of the corresponding amino acids with L-amino acid oxidase according to the method of Meister et al. (16). Malonamic acid was prepared by the method of Rig0 et al. (17) with a slight modification. 1,1,1,3,3,3-Hexamethyldisilazane (25 g, 154.9 mmol) was added via a syringe through a septum cap to a solution of Meldrum's acid (22.3 g, 154.9 mmol) in 200 ml of anhydrous CH,Cl, with stirring under nitrogen at room temperature. After 2 h, absolute methanol (20 ml) was slowly added to the reaction mixture. The mixture was allowed to stand in a -20 "C freezer overnight. The resulting acid precipitate was filtered and washed with CH2C12, then with ether. The solvents were removed under reduced pressure. To obtain pure malonamic acid, the Dowex l-X8 anion-exchange chromatography was performed using a linear gradient of 0-5 N formic acid. The fractions containing malonamic acid were identified with E2, pooled, and stored in a lyophilized state. The purity and structure of synthetic malonamic acid were confirmed by GC-MS as described below. ['ROIMalonate was synthesized by the reaction of malonyl dichloride with HzLsO. H,180 (0.5 g) and malonyl dichloride (1.3 ml) were mixed in a round bottom flask (molar ratio 2:l). After 1 min, malonic acid remained as a solid product.
It was dissolved in 20 ml of ethyl acetate and decolorized with Norit. 20 ml of benzene was added, and the mixture was kept at -20 "C for 12 h.
[lsOIMalonic acid was crystallized and dried under reduced pressure.
[lsOIMalonic acid was methylated with ethereal diazomethane and analyzed by GC-MS to be [1801malonic acid containing 50% 180.
Product Identification-Malonohydroxamate formed by the enzyme reaction was identified by thin layer chromatography with 1-butanol/ ethanoywater (2:2:1) as a developing solvent. The thin layer chromatography was performed by the ascending method, using Silica Gel 60-F precoated plates. The location of hydroxamate on thin layer chromatography plate was monitored by spraying 5% FeCl, solution in 95% alcoholic 0.1 N HCI.
Amino Acid Composition and N-terminal Amino Acid Sequence-Amino acid analysis of the malonamidases was performed in a Waters Pic0 Tag system amino acid analyzer. Samples for analysis were hydrolyzed with 200 pl of 6 N HCl at 110 "C for 24 h in an evacuated sealed tube. Half-cystine was determined as cysteic acid after performic acid oxidation. The total amount of tryptophan was estimated by methanosulfonic acid hydrolysis (18). N-terminal amino acid sequence analysis was performed in a model 477A Pulse-Liquid Protein Sequencer (Applied Biosystems, Inc.) with 1 nmol of the protein which was electroblotted onto a polyvinylidene difluoride membrane according to the method of Matsudaira (19).
Isotope Exchange Experiment-The reaction mixture contained 160 m~ [1801malonate, 0.1 M MOPS buffer, pH 6.8, and 32 units of purified E l a in a final volume of 0.5 ml. In the control reaction mixture, the enzyme was omitted. The reaction mixtures were incubated for 13 days at 37 "C. During the incubation, the reaction mixtures were centrifuged, and 32 units of fresh enzyme were added to them every 2 days. After the incubation, the reaction mixtures were loaded onto 500 pl of Dowex 1-X8 resin in a 1.5-ml Eppendorf tube and were washed with 1 ml of water five times. The bound compounds were eluted with 1 ml of 5 N formic acid twice. The combined eluents were freeze-dried. The compounds were methylated with ethereal diazomethane and analyzed by GC-MS as described below.
GC-MS Spectrometry-Mass spectrometry was carried out with a Hewlett-Packard GC-MS model 5988 fitted with a HP-5 column (25 m x 0.2 mm, Hewlett-Packard Co.). Helium flow was 1 ml x min" and electron energy was 70 eV. Samples of 1-3 p1 in methanol were introduced by direct injection through a septum. A solvent control was run between each sample to ensure that the injection port and column were free of the previous sample. Malonic acid dimethyl ester was passed through the column with a retention time of 6.2 min under the following program: 10 min at 70 "C; 10 "C x min" to 250 "C; 3 min at 250 "C.
Formation of Malonyl-EIa-The reaction mixture prepared for the isolation of malonyl-&la contained 5 pmol of MOPS buffer, pH 6.8, 0.8 pmol of sodium malonate, 0.2 pmol of [2-l4C1malonate, and 0.14 nmol of purified E l a in a final volume of 100 pl. After the mixture was incubated for 1 h at 37 "C, the mixture was chilled and immediately passed over a Sephadex G-25 column (1.0 x 35 cm), equilibrated with 20 m~ Tris-HC1 buffer, pH 7.4, containing 20 m~ malonate and 15% glycerol at 4 "C. The reaction mixtures prepared for bond characterization of malonyl-Ela contained 50 pmol of MOPS buffer, pH 6.8, 0.1 pmol of [l-14C]malonate, and 0.47 nmol of purified E l a in a final volume of 0.5 ml. The control sample contained the denatured enzyme at the same condition. After the mixtures were incubated for 2 days at 37 "C, the enzymes from the reaction mixtures were recovered by filtration through a Centricon-30 membrane filter. The remaining [l-14Clmalonate was removed by washing five times with 500 pl of 20 m~ Tris-HC1 buffer, pH 7.4. The isolated [14Clmalonyl-Ela was subsequently characterized.
Immunoblot Analysis-Antiserum to the purified E2 was prepared in New Zealand White rabbit. The purified E2 (300 pg in 1 ml of 20 m~ Tris-HC1, pH 7.4) was emulsified with an equal volume of complete Freund's adjuvant and was injected subcutaneously at multiple sites over the back of rabbit. m e r the initial dose, a booster infection of the purified enzyme (200 pg of protein) in 20 m M Tris-HC1 buffer, pH 7.4 (1 ml) and incomplete adjuvant (1 ml) was given twice subcutaneously at 2-week intervals in a similar manner. The formation of antibodies was confirmed by double immunodiffusion according to Ouchterlony and Nilsson (20). The rabbit was bled 12 days after the last infection. IgG fraction was purified by ammonium sulfate precipitation (40% saturation) and DEAE-Sephacel chromatography (21), dialyzed sufficiently against Tris-buffered saline (TBS), and used for subsequent experiments. The protein concentration was 7.5 mgjml. The purified malonamidases were incubated with anti-E2 antibody (diluted serially with TBS) for 1 h at 37 "C, followed by 12 h at 4 "C. 20 pl of Pansorbin cells (20% suspension in TBS after prewashing) was added, and incubation was continued on ice for 1 h. Precipitates were removed by centrifugation, and aliquots of supernatant fraction were taken for the determination of any remaining activity of Ela, Elb, or E2. Proteins were separated by SDS-PAGE and then electrotransferred to nitrocellulose filters in a solution containing 48 m~ Tris-HC1, 39 m~ glycine, 20% methanol, and 0.037% SDS according to Sambrook et al. (22). Immunodetection was carried out with the antibodies prepared against purified E2 as the first antibody (1:500 dilution), alkaline phosphate-conjugated goat anti-rabbit antibody as the second antibody, andp-nitroblue tetrazolium and 5-bromo-3-indolyl phosphate as the substrate, according to protoblot immunoscreening system protocol.

B. japonicurn Cells and Their Nodule Bacteroids
Novel enzymatic activity which catalyzes the formation of malonohydroxamate from malonate and hydroxylamine was discovered in the cell-free extract of free-living B. japonicum.

This enzyme was expressed constitutively in cells grown on
Brown's minimal medium (23) supplemented with a variety of carbon sources, such as glucose, arabinose, glycerol, acetate, succinate, malonate, histidine, YM, and GYP (the specific activities (in unit/mg) of the enzyme in the extract of cells were 0.024, 0.02, 0.017, 0.016, 0.02, 0.021, 0.014, 0.017, and 0.022, respectively). In addition, the cell-free extract of B. japonicum contained another enzyme activity which catalyzes the formation of malonohydroxamate from malonamate, synthetic amide derivative of malonate, and hydroxylamine. This activity was also expressed constitutively in YM and GYP medium. The enzymes were fractionated by Aff-Gel blue chromatography at pH 7.4. The breakthrough fractions contained enzyme activity with malonate or malonamate. The enzyme corresponding to this activity was designated as malonamidase El. However, the fractions obtained from 2 M NaCl stepwise elution showed activity only with malonamate. The enzyme corresponding to this activity was designated as malonamidase E2. El was divided into two types of enzyme which were eluted separately from DEAE-Sephacel chromatography. The enzymes obtained by malonate and NaCl elution were designated as Ela and Elb, respectively. The malonamidase activities of these were measured in the plant cytosol and bacteroids of soybean nodule infected with B. juponicurn and the physiological roles of these enzymes were proposed (1).

Purification of Malonamidases
The crude extract prepared from cells grown in GYP medium was subjected to batch treatment with AfE-Gel blue at pH 7.4. E l did not bind to Mi-Gel blue in this condition, whereas E2 did. El in the breakthrough fraction was precipitated by ammonium sulfate between 40-75% saturation. E l was further purified by the combination of gel filtration with Sephadex G-100 and ion exchange with DEAE-Sephacel. El was divided into Ela and E l b types through DEAE-Sephacel. Although Ela a n d E l b were separately fractionated with a linear gradient of NaCl (0-0.5 M ) (data not shown), they were separated more effectively by the differential elution with malonate and NaCl. After chromatography with DEAE-Sephacel, 85% o f E l activity was recovered in Ela, while 15% of that was recovered in Elb. Ela obtained from the previous step was purified by the combination of blue and red dye affinity chromatography at pH 6.5. Ela was eluted from blue gel under the buffer D containing 7 mM ATP, but was bound to red gel in the same condition. The Ela from red gel was purified -450-fold to a specific activity of 10.4 unit.mg" with a recovery of about 14% (Table I)  " Sum of E l a and E l b activity.
* The values determined into disregard the E l b activity.
" The values included the activity by malonamidase El.
ND, not determined. cific activity of 1.9 unikmg" in comparison to the enzyme preparation by DEAE-Sephacel. The E2 which bound to MI-Gel blue at pH 7.4 was eluted with 2 M NaCl stepwise elution. E2 was purified about 900-fold to a specific activity of 405 unit.mg" with a recovery of 6.4% by successive combination of phenyl-Sepharose, DEAE-Sephacel, and red-gel chromatography (Table I). Using 0.5 mM ZnS04 E2 could bind to phenyl-Sepharose and was eluted with a descending gradient of NaCl under the presence of 2 mM EDTA. This step was definitely attributed to the purification of E2. The purity of the malonamidase preparations resulting from the final purification step of each enzyme was examined by SDS-PAGE. All of the isolated malonamidases were electrophoretically homogeneous (Fig. 1). PAGE of the purified enzyme showed a protein band which was coincident with the band from activity staining (data not shown).

Reactions Catalyzed by Malonamidases
Malonamidases catalyze all or some of the three different reactions described under "Experimental Procedures." The Determined by a pore gradient PAGE. Determined by SDS-PAGE. e Determined by using a Phast Gel isoelectrofocusing (PI 3-9) system " ND, not determined. (Pharmacia).
The amount of ammonium ion produced was determined by nesslerization as described under "Experimental Procedures." fThe tyrosine and tryptophan content of each subunit is shown in Table V. These extinction coefficients were calculated based on these results. relative rate of the reactions (1, 2, and 3) for each enzyme varied: 2 > 3 >1 for E l a , 3 > 2 > 1 for Elb, and 2 >> 3 for E2. E2 did not show any activity with malonate. All of the above characteristics distinguish the three malonamidases. Meister et al.
(2) had reported that dicarboxylate w-amidase catalyzed Reactions 1,2, and 3 with broad specificity, but that the order of the reaction rates was 2 > 3 > 1. Ela showed a similar activity pattern with that of dicarboxylate w-amidase.
Characterization of Malonamidases Molecular Weight and pZ-The molecular weights of Ela, Elb, and E2 were determined by native gradient PAGE to be 126,000,107,000, and 103,000, respectively. SDS-PAGE of Ela, Elb, and E2 revealed only one protein band. The enzyme had subunit molecular weights of 58,900, 51,600, and 47,800, respectively, indicating that these malonamidases are homodimeric enzymes composed of two identical subunits (Fig. 1).
The PI values of native Ela, Elb, and E2 were determined by a gel isoelectric focusing to be 5.5, 5.0, and 7.2, respectively ( Table 11). Enzyme Stability and Optimum pH-The purified E l and E2 in buffer A and B, respectively, showed no detectable loss of enzyme activity at 4 "C for 1 month, and their activities were maintained after repeated freezing (-20 "C) and thawing. However, in phosphate buffer, E2 showed complete loss of its activity within 1 day. The optimum pH for a n enzyme varied depending on the reaction it catalyzed (Table 11). Ela activity for Reaction 1 depended on the buffer used; MOPS buffer produced the highest activity. Considering the concentration of unionized form (R-COOH) of malonic acid (<15%) above pH 6.8, the sharp pH optimum of Ela activity for Reaction 1 indicates that the unionized form of malonic acid may be the true substrate. This may be the reason for the high substrate concentration required for Reaction 1. The above result corresponds to those of previous studies done with other bacterial w-amidase (4).
Substrate Specificity-The substrate specificity of malonamidases was examined for three different reactions using a variety of carboxylic acids and their amide derivatives as substrates (Table 111). When methyl-malonate was used as a substrate instead of malonate in the malonyl transfer to hydroxylamine (Reaction l ) , methyl-malonohydroxamate was synthesized at a rate of 55 and 123% by E l a and Elb, respectively. However, no other carboxylic acids examined (acetate, succinate, oxalate, propionate, citrate, isocitrate, glutarate, and malate) were converted into hydroxamic acid. In addition, in hydroxaminolysis of malonamate (Reaction 2) and hydrolysis of malonamate (Reaction 31, all malonamidases were highly specific for malonamate. Meister et al. (2) had reported that dicarboxylate w-amidase uses a-ketoglutaramate, glutaramate, succinamate, a-ketosuccinamate, succinate, or glutarate, but not malonate or malonamate as their substrates. These results indicate that novel malonamidases are clearly distinct from dicarboxylate w-amidase.
Product Identification-The formation of malonohydroxamate from malonate (or malonamate) and hydroxylamine by malonamidases was confirmed by the isolation of these products using silica gel thin layer chromatography (RF = 0.27).
Malonamate was hydrolyzed by malonamidase into malonate and ammonia. An attempt to detect the formation of malonamate from malonate and ammonium chloride with Ela or E l b failed in contrast to the ability of w-amidase to catalyze amide formation (3). However, malonamate formation by E l a or E l b may be possible in bacteroids which keep a high concentration of ammonia.
Kinetic Parameters-With increasing concentration of substrates, the rate of malonohydroxamate formation and malonamate hydrolysis increased, and typical Michaelis-Menten saturation kinetics were obtained. The double-reciprocal plots were linear, and from these plots the kinetic parameters in the three different reactions were determined for malonamidases (Table IV). However, the kinetic parameters for E l b and E2 in malonamate hydrolysis could not be determined because E l b showed a saturation pattern below the detection range of the assay and because E2 catalyzed the hydrolytic reaction of malonamate too slowly. From KCatIKm, it was found that the substrate affinities of Ela and E l b for malonate were similar and that the substrate affinity of E2 for malonamate was about 30and 70-fold higher than those of Ela and Elb, respectively. Amino Acid Compositions and N-terminal Amino Acid Sequences-The amino acid compositions and the N-terminal sequences of malonamidases are presented in Table V. There about 554,422, and 440 residuedmonomer ofEla, Elb, and E2, respectively. The amino acid compositions of the three malonamidases are different. Since Elb has an acidic PI (5.0) and contains 29 basic residues (His, k g , and Lys), the major fraction of the 37 Asx and Glx residues must represent aspartic and glutamic acids rather the amide derivatives. On the other hand, since Ela contains 33 basic residues, the major fraction of the 117 Asx and Glx residues must represent the amide derivatives. In addition, since E l a and E2 contain a number of    tryptophan residues, their molar extinction coeffkients were high (Table 11). Microsequencing allowed the identification of the first 15 amino acids from the N terminus of the three malonamidases (Table V). However, in the case of E l b , four amino acids of the first 15 amino acids could not be identified, even after repeated attempts. The three malonamidases are clearly distinct from one another in sequences in the N-terminal region and their entire amino acid compositions, suggesting that they may be products from different genes. Inhibition-Ela was competitively inhibited by acetate and propionate with Ki, 8.3 and 22.4 mM, respectively. However, compounds such as succinate, malate, citrate, etc., which are structurally similar to but larger than malonate had no effect on Ela activity. On the other hand, E2 was inhibited noncompetitively by malonate with a K, of 0.18 mM, indicating that E2 has another binding site for malonate which may be related to the regulation of the in vivo function of E2. However, acetamide and succinamate had no effect on E2 activity. Ela was inhibited by TPCK, diethylpyrocarbonate, and diisopropyl fluorophosphate, whereas pyridoxal-5'-phosphate and thiol-specific compounds (iodoacetamide, p-chloromercuriphenylsulfo- nate, N-ethylmaleiimide, and 5,5'-dinitro-bis(2-nitrobenzoic acid) did not show detectable inhibition. When E l a was treated with TPCK (0.5 mM) for 60 min, with diethylpyrocarbonate (2 mM) for 30 min, or with diisopropylfluorophosphate (4 mM) for 30 min, the enzyme activity remained at about 50, 31, or 65%, respectively. These results indicate that histidine and serine, but not cysteine, residues may be involved in the enzyme catalysis. The catalytic action of Ela may be similar to that of a class of serine proteases rather than that of dicarboxylate w-amidases which contain the essential thiol group for its catalytic activity (3).

Formation of Malonyl Enzyme
The fact that Ela catalyzes the direct formation of malonohydroxamate from malonate and hydroxylamine suggests that the enzyme itself activates the carboxyl group of malonate by a functional group in its active site. The activation could be through an acyl-enzyme covalent intermediate. The first approach to the identification of the acyl-enzyme intermediate was through an the isolation of [14Clmalonyl-Ela by gel filtration (Fig. 2). The excluded peak of [l4CImalonate was co-chromatographed with the protein peak. Radioactivity retained in the protein was released by incubation with hydroxylamine, indicating that the radioactivity was due to [14C]malonate. The second approach to the identification of the intermediate was through a n isotope exchange experiment (Reaction 4). Ela was incubated in the presence of [lsOlmalonate (50% of the oxygen was labeled with l80). The malonate in the reaction mixture was isolated, and the l8O content was analyzed by GC-MS (Fig. 3). 74% of malonate with two l80 was entirely converted to [1601malonate, while 23% of that was converted to malonate with one "0. This result clearly shows that malonyl-enzyme was formed. The bond character of the malonyl-Ela intermediate was determined using the ['4Clmalonyl-Ela which had been isolated from the reaction mixture containing [1-14C] malonate. The molar ratio of malonyl group/enzyme was found to be 1.77, indicating that two malonyl groups are bound to one molecule of Ela. Since Ela is a homodimer, the result suggests that each subunit of the enzyme is malonylated. The Elabound malonate was rather stable upon application of heat (100 "C, 5 min) and strong alkali (final concentration, 0.1 N NaCI) than that of trichloroacetic acid (final concentration, 10%). At present, the functional group in the enzyme which activates the carboxyl group of malonate has not yet been clearly identified.

Immunological Characterization between Three
Malonamidases When immunoprecipitation experiments with the three malonamidases were performed with rabbit antiserum prepared against malonamidase E2, only the antigen itself (E21 precipitated. No cross-reactivity with the other enzymes, Ela and Elb, occurred. These results were also confirmed by the Ouchterlony immunodiffusion test and immunoblot analysis (Fig. 4). These results clearly indicate that E2 does not share any immunological identity with Ela and E l b in spite of the similarity in the type of reactions that they catalyze. Until now, however, the similarity between the immunological identities of Ela and E l b h a s not yet been examined. DISCUSSION In this report we describe the purification to homogeneity of novel malonamidase Ela, Elb, and E2 from B. juponicum and some of their properties. These enzymes can be classified as +amidase based on the pattern of the reactions catalyzed; however, their substrate specificity for malonate and/or malonamate clearly distinguishes them from dicarboxylate +amidase reported previously. These enzymes were found to catalyze all or some of the following three different reactions (Reaction 1, malonyl transfer to hydroxylamine; Reaction 2, hydroxaminolysis of malonamate; Reaction 3, hydrolysis of ma-Ionamate). An interesting point was that, although Ela and E l b acted on the same substrates, they displayed different relative rates of catalysis. The rate of hydrolysis activity of malonamate (Reaction 3) of E l b was 2.5-fold higher than the rate of its hydroxaminolysis activity of malonamate (Reaction 2), in contrast to the corresponding relative rates of Ela (Table  11). Furthermore, these malonamidases were shown to be clearly distinct enzymes in molecular structures. This result implies that Ela and E l b must slightly differ from each other in the protein structural aspects of their active sites, whereas the functional groups involved in catalysis are similar.
E l a was found to form a malonyl-enzyme intermediate during catalysis through the isolation of [14C]malonyl-Ela and the isotope exchange experiment. Although the nature of the functional group of the enzyme has not yet been elucidated, the inhibition study indicates that the functional group could be either an imidazole or a hydroxyl group. The formation of acylenzyme intermediate in dicarboxylate +amidase has been also proposed by kinetic experiment (24). The thiol group of cysteine residues was proposed to be involved in the catalysis by dicarboxylate w-amidase (25). However, various thiol-directed reagents did not affect the activity of malonamidase. Even though a malonyl-enzyme intermediate formed, the attempt to synthesize malonamate directly from NH; instead of hydroxylamine in a test tube failed. In uiuo, however, malonamate is very likely to be synthesized by the proper local concentration of "NH3" or NH3 donor-like glutamine. Although the formation of acyl-enzyme intermediate was only confirmed in Ela catalysis, E l b properly follows the same mechanism as Ela.
It has been proposed before that one of the biological roles of malonamidases is as a malonamate shuttle in symbiotic nitrogen metabolism (1). Under symbiotic conditions the plant keeps a high concentration of malonate in its nodule (14 mM) (9,(25)(26)(27). Considering the fact that nodules are packed with bacteroids, which cannot keep malonate in high concentration, the actual concentration of malonate in nodule plant cells must be much higher than the reported concentration. As reported previously, malonate in plant cell seems to be transported into bacteroids by diffusion (28). Recently, it has also been reported that the dicarboxylate carrier on a peribacteroid membrane recognizes malonate (ICm, 620 2 260 VM) (29). Since malonate is the most abundant organic acid in plant cell cytosol, the dicarboxylate carrier may preferentially transport it into a peribacteroid unit. As a consequence, malonate seems to be transported first into peribacteroid units by a dicarboxylate carrier, followed by its diffusion into bacteroid cells. Since malonate is a competitive inhibitor of succinate dehydrogenase (K, 17 pd, bacteroids would have t o expel it. This is indirectly supported by the fact that the malonate in bacteroids amounted to only -3% of the total malonate in nodules (27). There are two possible routes for malonate metabolism in bacteroids: one is to malonyl-CoA by malonyl-CoA synthetase (14) and the other is to malonamate. Even though the attempt to synthesize malonamate by Ela or E l b catalysis in a test tube failed, this reaction is very likely to occur this reaction in uivo. Ela forms a malonyl-enzyme intermediate which nucleophile, NH3, generated by nitrogenase, may attack easily. We have the evidence that malonamate exists in a cell-free extract of soybean nodule.' Another hypothesis cannot be ruled out. That is, some bacteroidal membrane contain ammonia monooxygenase which catalyzes the formation of hydroxylamine from ammonia (30,31). It is this hydroxylamine instead of ammonia which may be the substrate of malonamidase, resulting in the formation of malonohydroxamate. E2 has very high hydroxaminolysis activity with malonamate. Its role is not clear, but it may possibly be Y. S. Kim and S. W. Kang, unpublished data. involved in amidolysis or transmalonylation of some acceptor molecules in the periplasmic space or in the peribacteroid space. There is some evidence that E2 may be located in the periplasmic space and may be secreted into peribacteroid space.2 However, it is not clear whether E2-type enzyme in plant cell reported previously (1) originated from B. juponicum bacteroids.
Although it has been proposed that transfer of ammonia from the nitrogen-furing bacteroid to the host cytosol in soybean root nodules occurs by the simple diffusion of NH; across the bacteroid and peribacteroid membranes (32-341, the research carried out until now has been done only with isolated bacteroids, and little is actually known about the in uivo transport of fixed nitrogen from bacteroids to plant cells. Hence, it is necessary to investigate malonamidases in all of their genetic and physiological aspects in order to define more clearly both of their roles in the self-protection against malonate and ammonia transport to the plant cell.