Purification of Mitochondrial Glutamate Dehydrogenase from Dark

Plant Physiol. (1 996) 11 2: 1357-1 364

Purification of Mitochondrial Glutamate Dehydrogenase from Dark-Grown Soybean Seedlings

Frank 1. Turano*, Ralph Dashner’, Abha Upadhyaya, and Charles R. Caldwell United States Department of Agriculture, Agricultura1 Research Service, Climate Stress Laboratory,

Beltsville, Maryland 20705

Proteins in extracts from cotyledons, hypocotyls, and roots of 5-d-old, dark-grown soybean (Glycine max L. Merr. cv Williams) seedlings were separated by polyacrylamide gel electrophoresis. Three isoforms of glutamate dehydrogenase (CDH) were resolved and visualized i n gels stained for C D H activity. Two isoforms with high electrophoretic mobility, C D H l and CDH2, were in protein extracts from cotyledons and a third isoform with the lowest electrophoretic mobility, CDH3, was identified in protein extracts from root and hypocotyls. Subcellular fractionation of dark-grown soybean tissues demonstrated that CDH3 was associated with intact mitochondria. CDH3 was purified to homogeneity, as determined by native and sodium dodecyl sulfate-polyacrylamide gels. The isoenzyme was composed of a single 42-kD subunit. The pH optima for the reductive amination and the oxidative deamination reactions were 8.0 and 9.3, respectively. At any given pH, C D H activity was 12- to 50-fold higher in the direction of reductive amination than in the direction of the oxidative deamination reaction. CDH3 had a cofactor preference for NAD(H) over NADP(H). The apparent Michaelis constant values for a-ketoglutarate, ammonium, and NADH at pH 8.0 were 3.6, 35.5, and 0.07 mM, respectively. The apparent Michaelis constant values for glutamate and NAD were 15.8 and 0.1 O mM at p H 9.3, respectively. To our knowledge, this i s the first biochemical and physical characterization of a purified mitochondrial NAD(H)-dependent C D H isoenzyme from soybean.

GDH (EC 1.4.1.24.) has been the subject of numerous investigations (for review, see Stewart et al., 1980; Srivas- tava and Singh, 1987, and refs. therein), and it has been found in abundance in practically a11 plants investigated to date. GDH catalyzes the reversible reductive amination of a-ketoglutarate to form glutamate in the presence of the cofactor NAD(P)H. The enzyme exists as numerous isoen- zymic forms (isoforms) that may differ in their cofactor preference and intracellular locations. In general, the NAD(H)-specific enzyme has been identified in the mitochondria, and the NADP(H)-specific enzyme has been identified in the chloroplast. The number and relative pro- portions of the different GDH isoforms can vary in plant tissues, depending on developmental and environmental conditions.

Present address: National Institutes of Health, National Insti- tute of Arthritis and Musculoskeletal and Skin Diseases, Building 6, Room 324, MSC 2745, 9000 Rockville Pike, Bethesda, MD 20705.

* Corresponding author; e-mail fturano8asrr.arsusda.gov; fax 1-301-504 -7521.

Based on their electrophoretic mobilities, three GDH isoforms, GDH1, GDH2, and GDH3, were identified in extracts of developing soybean (Glycine max) seeds obtained 35 d after flowering (McKenzie et al., 1981; McKenzie and Lees, 1981). The isoforms had different intracellular locations and cofactor affinities, but the isoforms had similar pH optima, relative molecular masses, and specificities for a-ketoglutarate and glutamate. McKenzie and Lees (1981) demonstrated that GDH activity from root extracts was associated with intact mitochondria and that the mitochondrial GDH co-migrated in a native polyacrylamide gel with GDH3 isolated from developing seeds, thus suggesting that the seed GDH3 was associated with mitochondria. Other researchers have also demonstrated mitochondrial GDH activities (King and Yung-Fan Wu, 1971; Duke and Ham, 1976). McKenzie and Lees (1981) demonstrated that GDH2 was associated with the chloroplast. Shailendra and Shargool (1991) have also reported a plastid GDH in soybean.

The results from the experiments conducted with soybean suggest that there are distinct mitochondrial and plastid GDH isoforms with different affinities for the same cofactors and substrates, and it is likely that these isoforms are encoded by distinct genes. Small, multigene families are common among enzymes associated with nitrogen metabolism. Severa1 gene families have been identified and characterized for Gln synthetase (Cullimore et al., 1984; Coruzzi et al., 1988), glutamate synthase (Sakakibara et al., 1991; Gregerson et al., 1993), aspartate aminotransferase (Udvardi and Kahn, 1991; Gantt et al., 1992; Turano et al., 1992), and Asn synthetase (Tsai and Coruzzi, 1990). Fur- thermore, the presence of more than one GDH gene in plants is supported by results from genetic and molecular studies in maize (Pryor, 1979; Goodman et al., 1980; Sakak- ibara et al., 1995) and Arabidopsis (Cammaerts and Jacobs, 1983), which suggest that the GDH isoforms were composed of two distinct polypeptides or were encoded by two distinct loci. Recently, two distinct Arabidopsis expressed- sequence tags with high homology to bacterial and animal GDH genes were identified, thus providing conclusive ev- idence for the existence of multiple GDH genes in plants (Newman et al., 1994).

Despite detailed information on the physical and biochemical properties of the enzyme, relatively little is

Abbreviations: Cat, catalase; GDH, glutamate dehydrogenase; TPI, triose phosphate isomerase.

1357

1358 Turano et al. Plant Physiol. Vol. 11 2, 1996

known about the molecular regulation and gene structure of GDH in plants. Furthermore, the physiological role of GDH in plants is not fully understood. The lack of understanding of the physiological role of GDH in plants can be partially attributed to more interest in the other enzymes associated with nitrogen assimilation, contradictory results from different investigations, and, until recently, the lack of molecular tools (antibodies, mutants, and cDNA probes) for analyses. In this report, a GDH isoform identified in the roots and hypocotyls of dark-grown soybean seedlings was purified to homogeneity and the pure enzyme was physi- cally and biochemically characterized. This study is part of a broader effort in the laboratory to understand the cellular, developmental, and environmental factors that control GDH isoenzyme activation and / or gene expression in soybean. Fundamental biochemical <studies are required to gain a basic understanding of the cellular factors, i.e. pH, substrate, and cofactor availability, that control enzyme activity. Furthermore, such studies can provide a means of developing the molecular tools for detailed analyses of the molecular regulation and gene structure of GDH in the future.

MATERIALS A N D METHODS

Plant Material and Crude Protein Extractions '

Soybean (Glycine mux L. Merr. cv Williams) seeds were germinated on paper towels moistened with deionized water in the dark for 5 d at 25°C under aseptic conditions. After 5 d, cotyledons, hypocotyls, and roots were separated and protein samples were extracted from each of the or- gans. The samples (200 mg) were ground in 400 pL of buffer containing 40 mM Tris-HC1, pH 7.2,l mM EDTA, 5% (v/v) glycerol, and 0.01 mg/mL bromphenol blue. Triton X-100 was added to each sample to a final concentration of 0.05% (v/v) to disrupt organelles. The samples were incubated on ice for 30 min. Debris was removed from the sample by centrifugation at 13,OOOg for 10 min.

Protein Determinations and Enzymatic Activity Assays

Protein concentrations were determined using a Bio- Rad' or Pierce protein assay kit. The amination and deamination reactions were used to determine GDH activity and cofactor preference as described by Loulakakis and RoubelakislAngelakis (1990b). However, due to the high background in crude protein preparations and lower activity associated with the deamination reaction, the amination reaction was used routinely to determine activity and the amount of GDH loaded onto gels. In the amination reaction, GDH activity was determined in the presence of .100 mM NH,Cl or 50 mM (NH,),SO, 13 .mM a-ketoglutarate, 0.25 mM NADH, and 1 mM CaCI, in 100 mM Tris-HC1 or 100 mM KPO,, pH 8.0, and the decrease in A340 was'ye- corded for 1 min. In the deamination reaction, GDH activ-

Mention of trademark, proprietary product, or vendor does not constitute a guarantee or warranty of the product by the U.S. Department of Agriculture and does not imply its approval to, the exclusion of other products or vendors that may be suitable.

ity was determined in the presence of 35 mM glutamate, 0.25 mM NAD, and 1 mM CaCI, in 100 mM Tris-HC1 or 100 mM KPO,, pH 9.3, and the increase in A,,, was recorded for 1 min. One unit of GDH activity is defined as the reduction or oxidation of 1 micromole of coenzyme (NAD / NADP or NADH/NADPH, respectively) per min at 30°C. . To determine cofactor specificity, NADP(H) replaced NAD(H) at the same concentrations as described above.

Gel Electrophoresis, C e l Staining Procedures, and Western Blot Analysis

During the purification procedure, proteins were routinely separated by native PAGE and SDS-PAGE on the Phast (Pharmacia) system to determine the relative purity of GDH. Gradient gels (8-25 or 4-15%, polyacrylamide) or homogeneous gels (7.5% polyacrylamide) were used for nondenaturing (native) gels, and gradient gels (4-15% polyacrylamide) or homogeneous gels (12.5 or 7.5% polyacrylamide) were used for denaturing (SDS) gels. Proteins were visualized by silver stain on the Phast system as described by the manufacturer. Protein bands containing GDH activity were visualized in native polyacrylamide gels by incubation in 100 AM Tris-HC1, pH 9.3, with 50 mM glutamate, 0.5 'mM NAD, 0.25 mM nitroblue tetrazolium, and 0.1 mM phenazine methosulfate for 15 to 60 min at 37°C (Hartmann et al., 1973), Identical gels were incubated in the GDH stain solution minus glutamate as controls. In some instances, NADP replaced NXD in the GDH stain solution to identify distinci isoforms that may use a different cofactor or those isoforms that may utilize either cofactor. To identify different GDH isoforms in various dark- grown tissues or to purify GDH3 by preparative gel electrophoresis, proteins were separated in a native 6% polyacrylamide gel as described by Laemmli (1970), except. that SDS was omitted and the acry1amide:BIS ratio was 125:l. In both cases; gels were stained for GDH activity as described above. Pure GDH3 was resolved in a SDS 7.5% polyacrylamide gel as described by Laemmli (1970).

Western blot ' analysis was conducted as described by Turano et al. (1990). Proteins were separated by SDS-PAGE in 7.5% gels as described by Laemmli (1970). Rabbit serum raised against grape leaf NADH-GDH was provided by Loulakakis and Roubelakis-Angelakis (1990a).

Enriched Mitochondrial Fraction

Mitochondria were isolated from 10 g of 5-d-old roots and hypocotyls. The tissues 'were homogenized with a Polytron (Brinkmann) in two volumes of mitochondrial extraction buffer containing 0.3 M mannitol, 3 mM p- mercaptoethanol, 1 mM EDTA, 0.1% BSA (w/v), and 0.6% (w/v) insoluble PVP in 30 mM Mops, pH 7.5, per gram fresh weight. BSA was omitted from the buffer when the preparation was used for purification of GDH3. The ho- mogenate was filtered through one layer of Miracloth (Cal- biochem). Debris was removed from the resulting suspension by centrifugation at 500g. Mitochondria were concentrated by centrifugation at 12,0008. The resulting

Purification of Mitochondrial Soybean Glutamate Dehydrogenase 1359

crude mitochondria! pellet was used for subcellular local-ization or for purification of GDH3.

samples could be frozen until needed with a minimal lossof GDH activity.

Subcellular Localization

The mitochondrial pellet was washed twice in 5 mL ofmitochondrial resuspension buffer containing 0.3 M man-nitol, 1 mM EDTA, 1 mM MgCl2, 25 mM Hepes, pH 7.8, 0.1%(w/v) BSA, and 0.05% (v/v) /3-mercaptoethanol. The mi-tochondrial suspension (5 mL) was layered onto a discon-tinuous Percoll gradient (21, 26, 32, 47, and 60%). Aftercentrifugation at 45,000g for 45 min at 4°C, three distinctorganelle bands were removed and assayed for GDH ac-tivity and specific organelle marker enzymes. The mito-chondrial marker enzyme Cyt c oxidase (Smith, 1955), theplastid marker enzyme TPI (Miflin, 1974), and the peroxi-somal/glyxosomal marker enzyme Cat (Beers and Sizer,1952) were assayed as previously described by Turano etal. (1991).

Protein Purification

The crude mitochondrial pellet from 80 to 100 g of tissuewas washed as described above and resuspended in mito-chondrial lysis buffer containing 10 mM Tris-HCl, pH 7.5, 1mM EDTA, and 0.1% (v/v) Tween 20. The suspension waspassed through a tuberculin syringe (22-gauge) two tothree times and then subjected to a freeze-thaw cycle toensure maximum solubilization of membrane proteins. Thesolubilized mitochondrial suspension was clarified by cen-trifugation. The pellet was washed twice in a minimum of100 mM Tris-HCl, pH 7.5, and the supernatants werepooled.

The supernatant was applied to an Affi-Gel Blue (Bio-Rad) column (3-mL bed volume) preequilibrated in washbuffer containing 10 mM Tris-HCl, pH 7.5, 1% (v/v) glyc-erol, 100 ju,M CaCl2, 100 |u,M EDTA, and 150 ;uM /3-mercap-toethanol. The column was washed with 5 bed volumes ofwash buffer before eluting the bound proteins with 5 bedvolumes of 10 mM NAD and 500 mM glutamate in concen-trated (10 X) wash buffer. The affinity chromatographicprocedures were performed at room temperature. Theeluted sample was dialyzed overnight against 1 L of 5 mMTris-HCl, pH 7.5, with 1% (v/v) glycerol at 4°C and wasconcentrated in a Speed-Vac (Savant, Farmingdale, NY). A400-juL sample, representing approximately 100 units ofGDH activity, was loaded onto a 10 cm X 8 cm Mini-Gel(Bio-Rad) with a native 6% polyacrylamide resolving gel.The gel was run for 3 h at 10 mA. The sides of the gel werecut off and stained for GDH activity. After realigning thegel, a section of the unstained gel that corresponded to theGDH activity was removed. The gel fragment was choppedinto fine pieces, resuspended in 10 mL of 10 mM Tris-Glybuffer, pH 8.5, and loaded into dialysis tubing. The dialysistubing was placed in a horizontal electrophoresis appara-tus filled with the same Tris-Gly buffer, and the proteinwas electroeluted from the macerated gel by running theapparatus at 10 mA for 3 h at 4°C. The electroeluted samplewas dialyzed and concentrated as described above. The

RESULTS

Organ-Specific Accumulation

Proteins extracted from cotyledons, hypocotyls, androots of 5-d-old dark-grown soybeans were separated byPAGE. Three isoforms were visualized in gels stained spe-cifically for GDH activity (Fig. 1). Two isoforms with rela-tively high electrophoretic mobilities, GDH1 and GDH2,were identified in extracts from cotyledons. One isoformwith the lowest electrophoretic mobility, GDH3, was iden-tified in hypocotyl and root preparations.

Subcellular Localization of GDH3

Since only GDH3 activity was identified in hypocotylsand roots of dark-grown soybean seedlings, these tissueswere used as a source of material for the subcellular local-ization of the isoenzyme. Mitochondria were separatedfrom plastids and peroxisomes by centrifugation on a dis-continuous Percoll gradient as described above. After cen-trifugation, the three distinct organelle bands were col-lected from the gradient. Each sample was assayed forNADH- and NADPH-dependent GDH, Cyt c oxidase, TPI,and Cat activity. Much of the NADH-dependent GDHactivity (66%) was observed in a band located between the32/47% Percoll interface, which migrated with 73% of theCyt c oxidase activity (Table I). A small amount (16%) ofNADH-dependent GDH activity was associated with theorganelle band located between the 26 and 32% Percollinterface, which contained 81% of the TPI activity. Only 2%of the NADH-dependent GDH activity was associated withthe organelle band located between the 47 and 60% Percollinterface, which contained 81% of the Cat activity. Thesedata indicate that GDH3 was primarily associated with Cytc oxidase activity in intact mitochondria. Similar resultswere obtained with Sue gradients (data not shown), inwhich the mitochondrial fraction contained approximately60% of the total NADH-dependent GDH activity applied to

C H R

GDH2-GDH1-

-GDH3

Figure 1. GDH isoforms in germinating soybean. Protein extractsfrom cotyledons (C), hypocotyls (H), and roots (R) of 5-d-old dark-grown soybeans were separated by PAGE in a native 6% gel. Eachlane contains 10 units of GDH activity. Three isoforms, designatedGDH1, GDH2, and GDH3, were visualized in gels stained for GDHactivity. The isoforms are indicated via their anodal migration.


Table 1. Distribution o f soybean GDH3 and the marker enzymes Cyt c oxidase (Cyt c ox), TPI, and Cat in bands obtained from a Percoll step-gradient

The results presented in this table are representative of three separate experiments.

C D H Marker Enzymes Sample

NADH NADPH Cyt c o x TPI Cat

sh activity (units) 1 O0 (40) 1 O0 (225) Enriched mitochondriab 1 O0 (300)” 1 O0 (45) 1 O0 (45)

26/32% Percoll 16 13 1 1 81 6 32/47% Percoll 66 70 73 6 7 47/60% Percoll 2 n.d.‘ 5 1 60

a The value in the parentheses represents the total activity for each enzyme in the enriched mitochondrial preparation that was loaded onto the gradient. A preparation enriched for mitochondria was obtained by differential centrifugation (see ”Materials and Methods”). The enriched mitochondrial preparation (equivalent to 300 units of NADH-GDH activity) was loaded on a discontinuous Percoll gradient. One unit of G D H activitv eauals the oxidation of 1 micromole of NADH or NADPH Der minute at 30°C. n.d.. Not detectable. - , ,

the gradient, 11% of the GDH activity was associated with the plastid fraction, and 1% of the GDH activity was associated with the glyoxosomal/ peroxisomal fraction.

Purification and Characterization of GDH3

A protocol that included the isolation of mitochondria, affinity chromatography, and preparative PAGE was de- vised for the purification of GDH3. This procedure resulted in a 1134-fold purification and 11% recovery of GDH3 (Table 11). GDH3 eluted from a preparative PAGE ap- peared as a single band on silver- and activity-stained native polyacrylamide gels (Fig. 2A, lanes 4 and 7, respectively). The relative molecular weight on a native gradient 8 to 25% gradient polyacrylamide gel was 280,000 ? 20,000. Similar results were obtained on native homogeneous gels composed of 7.5 or 12% polyacrylamide (data not shown). When the sample was analyzed in a SDS 4 to 15% polyacrylamide gel, only one band of protein at 42 kD was present (Fig. 2B, lane 3); these results also suggest that GDH3 was purified to homogeneity. When the same protein preparation was separated by SDS-PAGE in a 7.5% polyacrylamide gel, a single polypeptide was visible after staining the gel with silver (Fig. 2C, lane 2). Western blot analysis of crude protein preparations (Fig. 2C, lane 3) and pure GDH3 (Fig. 2C, lane 4) separated in a 7.5% SDS-PAGE identified a single polypeptide (42 kD) that cross-reacted with rabbit serum raised against grape leaf NADH-GDH (Loulakakis and Roubelakis-Angelakis, 1990a), suggesting that the single polypeptide that constituted pure GDH3 was not an artifact obtained during purification.

Purified protein preparations were used to determine pH optima, cofactor specificities, and K , values. The optimum pH for the amination reaction was determined to be 8.0, but the optimum pH for the deamination reaction was 9.3 (Fig. 3). At any pH tested, the estimates of enzyme activity were 12 to 55 times higher in assays that favored the amination reaction compared with activity in the deamination reaction.

To determine cofactor preference, GDH assays were conducted in both directions using NAD(H) and NADP(H) at the two pH optima (Table 111). GDH3 had a preference for NADH over NADPH, which could be expressed as a ratio of 7.2 and 8.9 at pH 8.0 and 9.3, respectively. Likewise, there was a preference for an NAD over NADP (ratio 17 at pH 9.3) by the enzyme. Furthermore, estimates of GDH3 activity measured by the amination reaction w e r e about 35 times higher with NADH at pH 8.0 and 17 times higher at pH 9.3 than estimates of GDH activity using NAD in the corresponding deamination reactions. A similar trend was observed with estimates of GDH activity using NADPH, in which the estimates of activity were at least 33 times higher compared with estimates of activity using NADP.

An estimation of K, values was conducted (Table IV). The apparent K, values for each of the substrates were affected by pH. The apparent K , values for a-ketoglutarate, ammonium, and NADH at pH 8.0 were 3.6, 35.5, and 0.07 mM, respectively; but at pH 9.3 the apparent K , values for those substrates were 12.3, 20.1, and 0.06 mM, respectively. The apparent K, values for glutamate and NAD were 15.8 and 0.1 mM at pH 9.3 and 7.2, and 0.12 mM at pH 8.0, respectively.

Table II. Purification o f CDH3 from sovbean roots and hvpocovls grown in the dark

Purification Step NADH-GDH

Activity Protein Specific Activity Recovery Purification

unifsa Crude 381 Crude mitochondria preparation 31 6 Affi-Gel Blue chromatography 194 Electroelution 41

mg un its/Íng % -fold 1002 0.38 1 O0 -

53 5.96 83 15.7 2.5 77.6 51 204 0.095 43 1 11 1134

b

a One unit of C D H activity equals the oxidation of 1 micromole of NADH per minute at 30°C. GDH activity was determined in the presence of 50 miu (NH,),SO,, 13 mM a-ketoglutarate, 0.25 mM NADH, and 1 mM CaCI, in 100 mM Tris-HCI, p H 8.0. -, Not applicable.

Purification of Mitochondrial Soybean Glutamate Dehydrogenase 1361

1 2 3 4 5 6 7

B 1 2 3

1 2 3 4

Figure 2. Electrophoretic analyses of GDH3 by native and SDS-PAGE. A, Enzyme preparations containing GDH3 (0.5 units/lane)were separated by PAGE in a native 4 to 15% gradient gel. Theproteins in lanes 1 through 4 were stained with silver. Proteins inlanes 5 through 7 were stained for NADH-GDH activity. Molecularmass markers (lane 1) are as follows: ferritin (440 kD), catalase (232kD), lactate dehydrogenase (140 kD), and BSA (66 kD). Proteins fromcrude mitochondrial preparations are in lanes 2 and 5, Affi-GelBlue-purified preparations are in lanes 3 and 6, and gel/electroelutedpurified GDH are in lanes 4 and 7. B, Enzyme preparations contain-ing GDH3 (0.5 units/lane) were separated by SDS-PAGE in a 4 to15% gradient gel. Protein samples from crude mitochondria (lane 1),Affi-Gel Blue chromatography (lane 2), and gel/electroelution (lane3) preparations were visualized by silver stain. Molecular mass mark-ers (lane 4) are as follows: phosphorylase b (94 kD), BSA (67 kD),ovalbumin (43 kD), carbonic anhydrase (30 kD), trypsin inhibitor(20.1 kD), and a-lactalbumin (14.4 kD). C, Enzyme preparationscontaining GDH3 (10 units/lane) were separated by SDS-PAGE in a7.5% homogeneous gel. The molecular mass markers (lane 1) are asdescribed above. Pure (gel/electroeluted) GDH3 (lane 2) was visual-ized by silver stain. One polypeptide (42 kD) cross-reacted withrabbit serum raised against grape leaf NADH-GDH (Loulakakis andRoubelakis-Angelakis, 1990a) in the crude (lane 3) and pure (lane 4)preparations of GDH3.

DISCUSSION

In this investigation protein extracts from germinatingsoybeans contained three GDH isoforms that were resolvedby native PAGE and were designated GDH1, GDH2, andGDH3, according to their electrophoretic mobilities towardthe anode. The isoforms exhibited organ-specific activity:GDH1 and GDH2 were localized in protein extracts fromcotyledons, whereas GDH3 was identified in both hypoco-tyls and roots. In earlier studies by McKenzie et al. (1981)and McKenzie and Lees (1981), three GDH isoforms withdistinct electrophoretic mobilities were identified in pro-tein extracts from developing soybean seeds. In their stud-ies GDH1 was detected in developing seeds and not inroots or leaves, GDH2 was identified in leaves and devel-oping seeds, and GDH3 was identified in roots and devel-oping seeds. Their results, as well as those presented here,demonstrate the organ-specific activity of different GDHisoforms. A summary of the results from the combinedstudies suggest that (a) GDH1, the fastest migrating iso-form, is specific to the cotyledonary tissue in both devel-oping seeds and germinating seedlings, (b) GDH2 is asso-ciated with both cotyledons and leaves, and (c) the slowestmigrating isoform, GDH3, is in roots and hypocotyls.

In this study the intracellular location of GDH3 fromhypocotyl and root extracts was determined using discon-tinuous Percoll gradients. The isoform was associated withintact mitochondria. Other investigators (Duke and Ham,1976) have identified a mitochondrial GDH isoform fromsoybean roots, but the specific isoform was not determinedby electrophoretic mobility. McKenzie and Lees (1981)demonstrated that GDH3 was localized in root mitochon-dria. A mitochondrial isoform was identified in the coty-ledons of 11-d-old soybeans (King and Yung-Fan Wu,1971), but again, the specific isoform was not determinedby electrophoretic mobility. In this study we were unableto definitively determine the cellular locations of GDH1and GDH2 due to difficulties associated with the clearseparation of organelles from cotyledons on Percoll or Suegradients. Similar complications were reported previously

10 11 12

PH

Figure 3. Determination of pH optima for soybean GDH3. Theoptimum pH for the amination reaction (NADH-dependent) wasdetermined to be 8.0 and the optimum pH for the deaminationreaction (NAD-dependent) was determined to be 9.3. All data areexpressed as relative amount of activity versus NADH-GDH activityat pH 8.0. (See "Materials and Methods" for specific details of thereaction conditions.)


Table 111. Cofactor specificity of soybean GDH3

separate experiments. The results presented in this table are representative of three

Sample GDH Activity GDH Activity

GDH3 NADH NADPH NADHI NADPH

NAD NADP NADI NADP

unitP pH 8.0 345 48 7.2 10 n.d.b -' pH 9.3 294 33 8.9 17 1 17

a One uni t of GDH activity equals the reduction or oxidation of 1 n.d., N o t detecta- micromole of coenzyme per minute at 30°C.

ble. -, Not applicable.

by McKenzie and Lees (1981). Furthermore, McKenzie and Lees (1981) were not able to localize GDH2 and GDH3 directly from extracts of developing seeds. The localization of GDH2 and GDH3 were determined by analogy from other plant parts, namely leaves and roots, respectively.

Initial estimates of the relative molecular mass were determined by native PAGE on a series of homogeneous and gradient gels (F.J. Turano and R. Dashner, unpublished results). The relative molecular mass of the native protein was 280 2 20 kD. These data are similar to physical data reported for a GDH isoform isolated from the plastids of soybean cell culture (Shargool and Jain, 1989); the relative molecular weight was 263 kD and it was composed of 41-kD subunits. Moreover, these data are consistent with those from other plant GDHs. In general, plant GDH isoforms have relative molecular weights ranging from 208 to 300 kD and are composed of four subunits of about 60 kD or six subunits of about 45 kD (Stewart et al., 1980; Srivas- tava and Singh, 1987). GDH enzymes are usually composed of homogeneous monomers. Recently, however, NADH-GDH from grape was shown to be composed of two different polypeptides (Loulakakis and Roubelakis- Angelakis, 1991). The two polypeptides (about 42 kD each, designated a and P ) were clearly resolved by 7.5% SDS- PAGE. However, a11 of our results suggest that soybean GDH3 is composed of a 42-kD subunit and is most likely a hexameric protein.

In this study pure preparations of GDH3 were separated by SDS-PAGE in a 7.5% polyacrylamide gel, conditions that resolved two 42-kD polypeptides in grape (Loulakakis and Roubelakis-Angelakis, 1991), but only one polypeptide was apparent. We observed similar results using different gel systems (Phast [Pharmacia], Protean I and Mini-Protean I1 [Bio-Rad]) and a wide range of polyacrylamide concentrations using homogeneous (5, 5.5, 6, 7.0, up to 12%) and gradient gels (4-15%, 10-15%, and 8-25%). Western blot analysis of crude protein preparations from soybean roots and hypocotyls and pure GDH3 separated by 7.5% SDS- PAGE identified a single 42-kD polypeptide that cross- reacted with rabbit serum raised against grape leaf NADH- GDH (Loulakakis and Roubelakis-Angelakis, 1990a). These data suggest that GDH3 was composed of a single polypeptide that was not an artifact obtained during purification of GDH3 or due to preferential purification of one polypeptide. Likewise, our preliminary results from two-

dimensional electrophoresis suggest that GDH3 is composed of a single polypeptide (F.J. Turano, unpublished results). In grape, the ratio of the a and P polypeptides was shown to change when calli were maintained on different nitrogen sources (Loulakakis and Roubelakis-Angelakis, 1991). The cr subunit was almost exclusively present in calli grown on medium containing 10 mM NH,Cl, but the P subunit was predominant in protein extracts from calli maintained on 20 mM KN03. If soybean GDH3 were similar to the grape isoenzymes, i.e. composed of two distinct polypeptides, the presence of a single polypeptide in this experiment could be due to the absence of either nitrogen source during germination of the seeds. Therefore, the existence of a second GDH3 polypeptide cannot be ruled out at this time.

An alternative explanation for the appearance of one band in pure preparations of GDH3 by SDS-PAGE could be the inability to resolve two polypeptides of very similar or identical molecular weight. This hypothesis provides a plausible explanation for why soybean GDH does not resolve into numerous distinct bands on native polyacrylamide gels, as do grape (Loulakakis and Roubelakis-An- gelakis, 1990b, 1991) or Arabidopsis (Cammaerts and Jacobs, 1983, 1985; F.J. Turano, unpublished results) NADH-GDH isoenzymes. Grape NADH-GDH isoenzymes are composed of two subunits of different molecular weight that can be resolved by SDS-PAGE. The two subunits combined in different ratios to form hexameric com- plexes that can be resolved into seven isoenzymes by native PAGE (Loulakakis and Roubelakis-Angelakis, 1991). Similarly, two nonallelic genes are responsible for the formation of seven isoenzymes in Arabidopsis (Cammaerts and Jacobs, 1983).

GDH3 had two pH optima, one at 8.0 for the amination reaction and the other for the deamination reaction at 9.3. The existence of two pH optima is not uncommon for GDH isoenzymes (Stewart et al., 1980); similar findings have been observed with GDH from grape leaves (Loulakakis and Roubelakis-Angelakis, 1990) and soybean cotyledons (King and Yung-Fan Wu, 1971). Estimates of enzyme activity were 12 to 55 times higher in assays that favored the amination reaction compared with activity in the deamination reaction at the same pH.

GDH3 had a preference for NAD(H) over NADP(H); the NADH-to-NADPH ratio was slightly greater than 7. Our results are consistent with results from studies of other plant mitochondrial GDH isoforms, which usually show a preference for NAD(H) (Stewart et al., 1980; Srivastava and

Table IV. Apparent K,,, values ofsoybean GDH3 with different substrates at pH 8.0 or 9.3

The results presented in this table are representative of three separate experiments.

"H Amination Reaction

Deamination Reaction

CY-KG NH4 NADH Clu NAD

mM

8.0 3.6 35.5 0.070 7.2 0.1 20 9.3 12.3 20.1 0.060 15.8 0.100

1363 . Purification of Mitochondrial Soybean Glutamate Dehydrogenase

Sirigh, 1987). In addition, these findings are in agreement with results from western blot analysis (Fig. 2C), in which GDH3 cross-reacted with rabbit serum raised against grape leaf NADH-GDH (Loulakakis and Roubelakis-Angelakis, 1990a). The rabbit serum raised against grape leaf NADH- GDH does not cross-react with NADPH-GDH from soybean or Arabidopsis (F.J. Turano, unpublished results). McKenzie and Lees (1981) demonstrated that NADPH- dependent activity of GDH3 was less than 5% of the NADH-dependent activity, which relates to a NADH-to- NADPH ratio of greater than 20. The difference between the results from their study and this study may reflect iscfenzymic variants among different soybean varieties or variations in the experimental procedures.

The K , values for aLketoglutarate, ammonium, and NADH at the optimal pH 8.0 for the amination reaction were 3.6, 35.5, and 0.07 mM; respectively. The K , for glutamate and NAD at the optimal pH 9.3 for the deamination reaction were 15.8 and 0.10 mM, respectively. These values are similar to those observed for other plant NADH-GDH isoforms (Stewart et al., 1980; Srivastava and Singh, 1987). When estimates of the K , for a-ketoglutarate, ammonium, and NADH were conducted at pH 9.3, the optimal pH for the deamination reaction, the values were 12.3, 20.1, and 0.06 mM, respectively. Estimates of the K , for glutamate

. and NAD were 7.2 and 0.12 mM at pH 8.0, the optimal pH for the deamination reaction. In both cases the most dra- matic changes in apparent K, values were for the substrates a-ketoglutarate, ammonium, and glutamate, and not the cofactors NADH and NAD. These results suggest that the interaction between pH and available substrate could control catabolic or metagolic NAD(H)-GDH activity in the mitochondria.

Results from recent studie? contain data to support both a catabolic and a metabolic role of NAD(H)-GDH in plant nitrogen metabolism. The specific cellular factors that control NAD(H)-GDH activity remain elusive, but the general environmental factors are more defined. NAD-GDH activity may play a complementary role to Gln synthetase/ glutamate synthase in the reassimilation of excess ammo- nia released during stress conditions or during specific developmental stages (Yamaya et al., 1986; Rhodes et al., 1989). However, Robinson et al. (1991, 1992) demonstrated that glutamate oxidation may provide carbon skeletons, in the form of a-ketoglutarate, to the tri’carboxylic acid cycle in cells exposed to carbon-limited conditions. A better understanding of how these pH and substrate availability and other cellular factors (i.e. Ca2+, tricarboxylic acids, nucle- otides, etc.) interact to affect GDH activity may give greater insight into the physiological role of GDH in plants. How- ever, it is important to determine the location of GDH in the mitochondria and obtain accurate estimates of mitochondrial pH and substrate concentrations before consid- ering how these factors control GDH activity in vivo.

ACKNOWLEDCMENTS

The authors would like to thank Drs. Melinda Martin, Benjamin F. Matthews, H. David Husic, and Michael J. Muhitch for critica1

reyiew of the manuscript; and Drs. Perry Kregan (U.S. Department of qgriculture/ Agricultura1 Research Service, Soybean and Alfalfa Research Laboratory, Beltsville, MD) and William Kenworthy (Agronomy Department, University of Maryland, College Park, MD) for providing a supply of Williams soybean seed.

Received June 11, 1996; accepted August 7, 1996. Cbpyright Clearance Center: 0032-0889/96/ 112/ 1357/08.

LITERATURE ClTED

Beers PF, Sizer IW (1952) A spectrophotometric for measuring the breakdown of hydrogen peroxide by catalase. J Biol Chem 195:

Cammaerts D, Jacobs M (1983) A study of the polymorphism and the genetic control of the glutamate dehydrogenase isoenzymes in Arabidopsis thaliana. Plant Sci Lett 31: 65-73

Cammaerts D, Jacobs M (1985) A study of the role of glutamate dehydrogenase in nitrogen metabolism of Arabidopsis thaliana. Planta 163: 517-526

Coruzzi GM, Edwards JW, Tingey SV, Tsai FY, Walker EL (1988) Glutamine synthetase molecular evolution of an eclectic multigene family. Zn R Goldberg, ed, The Molecular Basis of Plant Development, Vol92. AR Liss, New York, pp 1-22

Cullimore JV, Gebhardt C, Saarelainen R, Miflin BJ, Idler KB, Baker RF (1984) Glutamine synthetase of Phaseolus vulgaris L.: organ-specific expression of a multigene family. J Mo1 Appl Genet 2: 589-599

Duke SH, Ham GE (1976) The effect of nitrogen addition on N,-fixation and on glutamate dehydrogenase and glutamate synthase activities in nodules and roots of soybean inoculated with various strains of Rhizobium japonicum. Plant Cell Physiol

Gantt JS, Larson RJ, Farnham MW, Pathirana SM, Miller SS, Vance CP (1992) Aspartate aminotransferase in effective and ineffective alfalfa nodules. Cloning of a cDNA and determination of enzyme activity, protein and mRNA levels. Plant Physiol

Goodman MM, Stuber CA, Newton K, Weissinger HH (1980) Linkage relationships of 19 loci in maize. Genetics 96: 697-710

Gregerson RB, Miller SS, Twary SN, Gantt JS, Vance CP (1993) Molecular characterization of NADH-dependent glutamate synthase from alfalfa nodules. Plant Cell 5: 215-226

Hartmann T, Nagel M, Ilert HJ (1973) Organ-specific multiple forms of glutamate dehydrogenase in Medicago sativa. Planta

King J, Yung-Fan Wu W (1971) Partia1 purification and kinetic properties of glutamic dehydrogenase from soybean cotyledons. Phytochemistry 10: 915-928

Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227: 680-685

Loulakakis KA, Roubelakis-Angelakis KA (1990a) Immunochar- acterization of NADH-glutamate dehydrogenase from Vi t i s vinijera L. Plant Physiol 94: 109-113

Loulakakis KA, Roubelakis-Angelakis KA (1990b) Intracellular localization and properties of NAD(H)-glutamate dehydrogenase from Vi t i s vinijera L.: purification and characterization of the major leaf isoenzyme. J Exp Bot 41: 1223-1230

Loulakakis KA, Roubelakis-Angelakis KA (1991) Plant NAD(H)- glutamate dehydrogenase consists of two ‘subunit polypeptides and their participation in the seven isoenzymes occurs in an ordered ratio. Plant Physiol 97: 104-111

McKenzie EA, Copeland L, Lees EM (1981) Glutamate dehydrogenase activity in developing soybean seed: kinetic properties of three forms of the enzyme. Arch Biochem Biophys 212: 298-305

McKenzie EA, Lees EM (1981) Glutamate dehydrogenase activity in developing soybean seed: isolation and characterization of three forms of the enzyme. Arch Biochem Biophys 212: 290-297

Miflin B (1974) The localization of nitrite reductase and other

133-140

17: 1037-1044

98: 868-878

111: 119-128


enzymes related to amino acid biosynthesis in the plastids of root and leaves. Plant Physiol 54: 550-555

Newman T, de Bruijn FJ, Green P, Keegstra K, Kende H, McIn- tosh L, Ohlrogge J, Raikhel N, Somerville S, Thomashow M, Reze1 E, Somerville C (1994) Genes galore. A summary of the methods for accessing results from large-scale partia1 sequenc- ing of anonymous Ambidopsis cDNA clones. Plant Physiol 106:

Pryor AJ (1979) Mapping of glutamate dehydrogenase (Gdh) on chromosome 1,20.1 recombination units dista1 to Adhl . Maize Genetics Cooperative News Letter 53: 25-26

Rhodes D, Brunk DG, Magalhaes JR (1989) Assimilation of am- monia by glutamate dehydrogenase? In JE Poulton, JT Romeo, EE Conn, eds, Plant Nitrogen Metabolism, Vol 5. Plenum Press, New York, pp 191-206

Robinson SA, Slade AP, Fox GG, Phillips R, Ratcliffe G, Stewart GR (1991) The role of glutamate dehydrogenase in plant nitrogen metabolism. Plant Physiol 95: 509-516

Robinson SA, Stewart GR, Phillips R (1992) Regulation of glutamate dehydrogenase activity in relation of carbon limitation and protein catabolism in carrot cell cultures. Plant Physiol98 1190- 1195

Sakakibara H, Fujii K, Sugiyama T (1995) Isolation and characterization of a cDNA that encodes maize glutamate dehydrogenase. Plant Cell Physiol 36: 789-797

Sakakibara H, Watanabe M, Hase T, Sugiyama T (1991) Molec- ular cloning and characterization of complementary DNA encoding for ferredoxin-dependent glutamate synthase in maize leaf. J Biol Chem 266: 2028-2035

Shailendra BK, Shargool PD (1991) A plastidal localization and origin of L-glutamate dehydrogenase in a soybean culture. Plant Physiol 95: 258-263

1241-1255

Shargool PD, Jain JC (1989) Purification and immunological properties of an NAD(H) dependent glutamate dehydrogenase from soybean cells (Glycine mux L.). Plant Sci 60: 173-179

Smith L (1955) Cytochromes a, al , a2, and a3. Methods Enzymol2:

Srivastava HS, Singh RP (1987) The role and regulation of L- glutamate dehydrogenase activity in higher plants. Phytochem- istry 26: 597-610

Stewart GR, Mann AF, Fentem PA (1980) Enzymes of glutamate formation: glutamate dehydrogenase, glutamine synthetase and glutamate synthase. In BJ Miflin, ed, The Biochemistry of Plants, Vol 5. Academic Press, New York, pp 271-327

Tsai FY, Coruzzi GM (1990) Dark-induced and organ-specific expression of two asparagine synthetase genes in Pisum sutivum.

Turano FJ, Jordan RL, Matthews BF (1990) Immunological characterization of in vitro forms of homoserine dehydrogenase from carrot suspension cultures. Plant Physiol 92: 395-400

Turano FJ, Weisemann JM, Matthews BF (1992) Identification and expression of a cDNA clone encoding aspartate aminotransferase in carrot. Plant Physio1100: 374-381

Turano FJ, Wilson BJ, Matthews BF (1991) Rapid purification and thermostability of the cytoplasmic aspartate aminotransferase from carrot suspension culture. Plant Physiol 9 7 606-612

Udvardi MK, Kahn ML (1991) Isolation and analysis of a cDNA clone that encodes an alfalfa (Medicago sativa) aspartate aminotransferase. Mo1 Gen Genet 231: 97-105

Yamaya T, Oaks A, Rhodes D, Matsumoto H (1986) Synthesis of [15N]glutamate from [I5N]H, and [15N]glycine by mitochondria isolated from pea and corn shoots. Plant Physiol 81: 754-757

732-740

EMBO J 9: 323-332

Documents

Purification of Mitochondrial Glutamate Dehydrogenase from Dark