Threonine-insensitive Homoserine Dehydrogenase from Soybean

Aspartate kinase (AK) and homoserine dehydrogenase (HSD) function as key regulatory enzymes at branch points in the aspartate amino acid pathway and are feedback-inhibited by threonine. In plants the biochemical features of AK and bifunctional AK-HSD enzymes have been characterized, but the molecular properties of the monofunctional HSD remain unexamined. To investigate the role of HSD, we have cloned the cDNA and gene encoding the monofunctional HSD (GmHSD) from soybean. Using heterologously expressed and purified GmHSD, initial velocity and product inhibition studies support an ordered bi bi kinetic mechanism in which nicotinamide cofactor binds first and leaves last in the reaction sequence. Threonine inhibition of GmHSD occurs at concentrations (Ki = 160–240 mm) more than 1000-fold above physiological levels. This is in contrast to the two AK-HSD isoforms in soybean that are sensitive to threonine inhibition (Ki∼150 μm). In addition, GmHSD is not inhibited by other aspartate-derived amino acids. The ratio of threonine-resistant to threonine-sensitive HSD activity in soybean tissues varies and likely reflects different demands for amino acid biosynthesis. This is the first cloning and detailed biochemical characterization of a monofunctional feedback-insensitive HSD from any plant. Threonine-resistant HSD offers a useful biotechnology tool for manipulating the aspartate amino acid pathway to increase threonine and methionine production in plants for improved nutritional content.

In plants and microorganisms, aspartate is the common precursor of the essential amino acids lysine, threonine, methionine, and isoleucine (1)(2). Because this branching set of pathways is critical for central metabolism and not found in mammals, the enzymes of the aspartate family of amino acid biosynthesis present targets for the development of anti-microbial agents and herbicides (1)(2)(3) and for improving the nutritional value of food crops (4 -7). In the initial step of the pathway (1), aspartate kinase (AK) 7 catalyzes the phosphorylation of aspartate (Fig. 1). Conversion of aspartyl phosphate to L-aspartate-␤-semialdehyde (Asa) leads to the first branch point of the pathway, as Asa can be metabolized to either lysine or homoserine (Hse). Homoserine dehydrogenase (HSD) catalyzes the conversion of Asa to Hse in the third step of the pathway. Subsequent metabolism of Hse yields methionine and threonine. Both AK and HSD are highly regulated by feedback inhibition through threonine to control flux through the various branches of the pathway (1)(2). The regulation of these enzymes by threonine depends on their structural organization in different organisms.
AK and HSD function as either monofunctional or bifunctional enzymes in bacteria, yeast, and plants. In Escherichia coli, AK occurs as either a threonine-resistant monofunctional enzyme or as a threonine-sensitive bifunctional AK-HSD (8 -10). HSD operates either as part of a bifunctional feedback responsive AK-HSD in Gram-negative bacteria such as E. coli or as a monofunctional protein in Gram-positive bacteria and yeast (11,12). In the Gram-positive bacteria Corynebacterium glutamicum and Bacillus subtilis, threonine inhibits the monofunctional HSD, but the enzyme from yeast (Saccharomyces cerevisiae) is resistant to the amino acid (13,14). Differences in the C-terminal region in the yeast HSD compared with the bacteria enzyme prevent inhibition by threonine (1,14). Plants utilize a similar complement of enzymes for the synthesis of other amino acids derived from aspartate. A family of genes encodes three monofunctional AKs in Arabidopsis thaliana (thale cress) (15,16). To date, the only HSD cloned from plants, including A. thaliana, maize (Zea mays), and soybean (Glycine max), are those fused with AK as a bifunctional enzyme (17)(18)(19)(20)(21). The AK-HSD isoforms from Arabidopsis and maize are sensitive to inhibition by threonine (17)(18)(19)21), but the enzymes from soybean remain biochemically uncharacterized (20). In both bacterial and plant AK-HSD, the linker region between each enzyme domain is responsible for the regulatory effect of threonine (13,21). Interestingly, plants also appear to use monofunctional HSD, as the activity of a threonine-resistant HSD in maize extracts was reported (22,23), but no molecular details described.
To investigate the role of a threonine-resistant HSD in the aspartate amino acid pathway of plants, we have cloned the cDNA and gene encoding the enzyme (GmHSD) from soybean. Biochemical analysis of GmHSD using initial velocity and product inhibition studies suggest that the enzyme follows an ordered bi bi kinetic mechanism. Importantly, GmHSD is not inhibited by physiological concentrations of threonine. In contrast, kinetic analysis of the two previously uncharacterized GmAK-HSDs demonstrates that threonine tightly regulates each isoform. The ratio of threonine-resistant to threoninesensitive HSD in soybean tissues varies and may have a role in meeting amino acid demands in different tissues. This work is the first reported isolation and molecular characterization of a monofunctional HSD from any plant. The isolation of a threonine-resistant HSD from soybean may provide a useful biotechnology tool for manipulating the aspartate amino acid pathway to increase threonine and methionine production in plants for improved nutritional content.

EXPERIMENTAL PROCEDURES
Materials-Five days after sowing, seedlings of soybean cv. Williams 82 were used to isolate RNA. Preparation of Asa was performed by ozonolysis of L-allyl glycine (24). The ozonolysis product was purified using a cation exchange resin (Dowex 50WX8 -400; column: 1.5 ϫ 10 cm). Immediately before use, the product was neutralized with NaHCO 3 . All other chemical reagents were purchased from Sigma.
Isolation of GmHSD cDNA and Genomic Clones-Total RNA from seedling hypocotyls was isolated through a lithium chloride precipitation procedure (25). Primers (forward, 5Ј-dGCT-GGTCTTCCTGTGATAGCATCA-3Ј; reverse, 5Ј-dCC-ATTTGAAGCTGCCTTCTCAACTC-3Ј) were synthesized based on sequence information obtained from an EST clone (GenBank TM accession number BM308805) for use as probes to screen cDNA libraries. A 394-bp fragment was amplified by reverse transcriptase-PCR (Stratagene). This fragment was cloned into a sequencing vector pGEMT-EASY (Promega). After confirming the sequence, the fragment was purified by agarose gel electrophoresis and radiolabeled with [ 32 P]dCTP using a random labeling kit (Takara Bio). The probe was used to screen a seedling cDNA library (courtesy of Dr. Joseph Polacco, University of Missouri). Two positive clones were isolated after colony hybridization with a radiolabeled probe following the standard method for three times (26). One of the recovered clones was sequenced and showed a full-length cDNA (deposited in GenBank TM accession number as DQ172918.1; supplemental Fig. 1).
The same probe used to screen the cDNA library was used for screening a commercially available soybean genomic library (Stratagene). Seven positive clones were isolated through colony hybridization with a radiolabeled probe. Hybridization was performed three consecutive times (26). DNA restriction followed by Southern analysis of one of the seven clones revealed hybridization with two HindIII inserts (3.2 and 4.2 kb) and one EcoRI insert (0.6 kb). These inserts were gel-isolated, purified, and subcloned into an intermediate vector pBlueScript before sequencing.
Southern Blot Analysis-Soybean genomic DNA was isolated from the leaves of the Williams 82 cultivar using the cetyltrimethyl ammonium bromide method (27). After quantification, 8 g of DNA was digested along with different restriction enzymes overnight at 37°C. The digested DNA was separated on a 0.8% agarose gel and transferred to nylon membranes by a capillary transfer using 0.4 M of NaOH. The same DNA fragment used to screen the cDNA library was labeled with [ 32 P]dCTP using the Ladderman labeling kit (PanVera) then employed as a probe for Southern analysis. After 6 h of prehybridization, overnight hybridization was done at 65°C using 6ϫ SSPE buffer (1ϫ SSPE is 100 mM NaCl, 10 mM Na 2 PO 4 , and 1 mM EDTA), 5ϫ Denhardt's solution, 0.5% (w/v) SDS, and 50 g ml Ϫ1 salmon sperm DNA. After hybridization, the membranes were washed 3 times in wash solution I (2ϫ SSPE and 0.5% (w/v) SDS) for 10 min at room temperature before 2 washes in wash solution II (0.1ϫ SSPE and 0.1% (w/v) SDS) at 65°C for 30 min. After washing, the nylon membrane was exposed to x-ray film overnight at Ϫ80°C.
Protein Expression and Purification-The coding region of GmHSD was amplified by PCR, and the resulting product was digested with NdeI and XhoI and ligated into the corresponding restriction sites of pET-28a (Novagen). The pET28a-GmHSD expression construct was transformed into E. coli Rosetta(DE3) (Novagen) cells. Transformed cells were grown at 37°C in Terrific broth containing 50 g ml Ϫ1 kanamycin and 34 g ml Ϫ1 chloramphenicol until A 600nm ϭ 0.8 -1.2. After induction with 1 mM isopropyl 1-thio-␤-D-galactopyranoside, the cultures were grown for 6 -8 h at 18°C. Cells were pelleted by centrifugation (10,000 ϫ g; 10 min) and resuspended in lysis buffer (50 mM Tris (pH 8.0), 500 mM NaCl, 25 mM imidazole, 10% (v/v) glycerol, and 1% (v/v) Tween 20). After sonication and centrifugation (48,000 ϫ g; 60 min), the supernatant was passed over a Ni 2ϩ -nitriloacetic acid-agarose (Qiagen) column. The column was washed with lysis buffer without Tween 20, and the His-tagged protein eluted with elution buffer (wash buffer containing 250 mM imidazole). Size-exclusion chromatography used a Superdex-200 16/20 fast protein liquid chromatography column (Amersham Biosciences) equilibrated with 50 mM Tris (pH 8.0) and 250 mM NaCl. Protein concentration was determined using Bradford reagent (Bio-Rad) with bovine serum albumin as the standard. Purified protein was stored at Ϫ80°C.
HSD Assays-A standard reaction mixture (0.5 ml; 25°C) for spectrophotometrically monitoring GmHSD activity contained 0.1 M Tris (pH 8.0), 5 mM Asa, and 2 mM NAD(P)H in the forward reaction or 0.1 M Tris (pH 8.0), 50 mM Hse, and 2 mM NAD(P) ϩ in the reverse reaction. Initial velocities were determined by following the change in A 340 nm . For determination of steady-state kinetic parameters, reaction rates were measured under standard assay conditions with a matrix of varied substrate concentrations. Curve fitting in SigmaPlot (Systat Software, Inc.) was used to model the kinetic data to the equation where v is the initial velocity, V max is the maximum velocity, K A and K B are the K m values for substrates A and B, respectively, and K ia is the substrate inhibition constant. The same assays were used to monitor activity of monofunctional and bifunctional HSD.
For product and feedback inhibition assays, enzymatic activity was measured in reactions under standard assay conditions containing inhibitor and varied concentrations of either substrate. The fitting analysis of data used SigmaPlot to fit data to the equations for either competitive Assays of HSD activity in extracts of soybean tissues were performed using the standard assay system. Tissue extracts were prepared in 100 mM Tris (pH 8.5), 2 mM dithiothreitol, 1 mM EDTA. Homogenates were centrifuged at 4°C, and the supernatant was collected. Assays contained 10 -100 g of total protein.
Expression and Purification of GmAK and GmAK-HSD-Soybean AK (GmAK; GenBank TM accession number AAD41796.1) was amplified from the seedling cDNA library and cloned into pCR-Blunt-II TOPO vector. Sequencing confirmed the fidelity of the clone. GmAK was subcloned into pET-28a for bacterial expression using Nhe1 and NotI sites. The two bifunctional AK-HSDs in soybean (GmAK-HSD-1, GenBank TM accession numbers AAC05983; GmAK-HSD-2, GenBank TM accession number AAC05981 (20)) were also amplified from the same library, cloned into pCR-Blunt-II TOPO vector, sequenced, and subcloned into pET-28a using NheI and NotI sites. Protein expression and purification of His-tagged GmAK, GmAK-HSD-1, and GmAK-HSD-2 used the same procedure as described for GmHSD.
AK Assays-AK activity of GmAK, GmAK-HSD-1, and GmAK-HSD-2 was determined spectrophotometrically at 25°C by measuring the rate of ADP formation using a coupled assay with pyruvate kinase and lactate dehydrogenase (28). A standard reaction mixture (0.5 ml) contained 100 mM Tris (pH 8.0), 100 mM KCl, 2.5 mM ATP, 10 mM aspartate, 5 mM MgCl 2 , 2 mM sodium phosphoenolpyruvate, 0.2 mM NADH, 5 units of pyruvate kinase, and 10 units of lactate dehydrogenase. Kinetic parameters and inhibition constants were determined using the equations described above.

RESULTS
Cloning and Genome Analysis of GmHSD-Screening a soybean seedling cDNA library with an EST sequence (accession number BM308805) as the probe revealed a full-length cDNA clone. The putative cDNA clone was sequenced, and analysis of the sequence revealed a 1472-bp DNA insert containing a single open reading frame of 1128 bp (deposited in GenBank TM accession number as DQ172918.1; supplemental Fig. 1). The cDNA has a 90-bp 5Ј-untranslated region (UTR) and a 251-bp 3Ј-UTR. The predicted open reading frame encodes a protein of 376 amino acids with a calculated molecular mass of 40.6 kDa and predicted isoelectric point of 6.19. BLAST analysis showed that this sequence shares 40 and 25% amino acid sequence identity with the monofunctional HSD from S. cerevisiae and B. subtilis, respectively.
To determine the gene organization of GmHSD, sequencing of two HindIII (3.2 and 4.2 kb) fragments and one EcoRI (0.6 kb) fragment from 1 of the 7 clones obtained from genomic library screening revealed the entire coding region of GmHSD (deposited in GenBank TM accession number as DQ788566.1). Of the 7997 bp sequenced, the cDNA of GmHSD begins at position 1461 and ends at position 7763. The start codon of the cDNA maps to position 1537, and the stop codon is located at 7760 in the genomic sequence. The nucleotide sequence of the genomic clone along with the deduced amino acid sequence from the full-length DNA, was used to determine the position and span of introns and exons ( Fig. 2A). The GmHSD coding sequence is interrupted by 11 introns that vary in length from 107 to 1637 bp. Southern analysis with six different restriction enzymes was carried out to determine the copy number of HSD in the soybean genome (Fig. 2B). Except for EcoRI, strong hybridization was seen for one band, whereas the other bands showed weak hybridization. The obtained results indicate that GmHSD is encoded by a single low copy gene but that hybridization to other HSD-related genes occurs most likely the AK-HSD isoforms in soybean (20).
Expression and Purification of GmHSD-To verify that the isolated cDNA encoded a functional HSD, recombinant protein was overexpressed in E. coli Rosetta(DE3) as a hexahistidinetagged fusion protein and purified to apparent homogeneity using Ni 2ϩ -affinity and size-exclusion chromatography (Fig. 3). By SDS-PAGE analysis, the purified protein migrated with a monomer weight of ϳ40 kDa, which corresponds to the predicted molecular weight (40.6 kDa). GmHSD eluted from the size-exclusion column as a 70-kDa species representing a dimer.
Kinetic Analysis of GmHSD-Purified protein was first analyzed for HSD and AK activity. GmHSD displayed a specific activity of 13 mol min Ϫ1 mg of protein Ϫ1 using Asa and NADH as substrates and exhibited no detectable AK activity. A complete steady-state kinetic analysis of the forward and reverse reactions of the enzyme was performed to understand Soybean Homoserine Dehydrogenase the equilibrium of this metabolic step and to determine nucleotide cofactor specificity. Kinetic parameters for GmHSD were determined for the forward (Asa/NADH and Asa/NADPH) and reverse (Hse/NAD ϩ and Hse/NADP ϩ ) reactions. Initial velocity data were generated by varying both substrates and fitting to the equation for a sequential bi bi mechanism ( Fig. 4 and Table 1). Comparison of catalytic efficiencies for the forward and reverse reactions shows that the equilibrium favors production of Hse using either nucleotide cofactor. Although GmHSD displays a 1.6-fold preference for NADPH as the cofactor in the forward reaction, both cofactors are efficiently used for catalysis. In the reverse reaction NADP ϩ is favored nearly 4-fold as the cofactor.
Kinetic Mechanism of GmHSD-The reciprocal plots of Asa versus NAD(P)H and Hse versus NAD(P) ϩ (Fig. 4) showed intersecting patterns, suggesting a sequential bi bi kinetic mechanism rather than a ping-pong type of mechanism. To distinguish between ordered and random bi bi mechanisms, product inhibition studies were performed. For these experiments, NADPH and NADP ϩ were used as the nucleotide cofactors because these are preferred by GmHSD. Reactions were performed in the presence of a product inhibitor with one substrate at varied concentrations and the other at an unsaturating concentration. In the forward reaction Hse was a noncompetitive inhibitor versus both Asa and NADPH, and NADP ϩ showed non-competitive and competitive inhibition patterns versus Asa and NADPH, respectively ( Table 2). Similar product inhibition patterns were observed for the reverse reaction. Asa was a noncompetitive inhibitor versus both substrates, and NADPH displayed non-competitive inhibition versus Hse but was competitive versus NADP ϩ . The competitive inhibition observed in each direction with the nicotinamide cofactors indicates that they bind to the same form of the enzyme in an ordered mechanism.
Feedback Inhibition of GmHSD by Aspartate-branch Amino Acids-HSD activity in plants is reported to be sensitive to inhibition by downstream products of the aspartatederived amino acid pathway. To examine the effect of the aspartatederived amino acids (aspartate, lysine, methionine, isoleucine, and threonine) on GmHSD, enzyme assays were performed in the presence of each amino acid. Isoleucine (100 mM) had no effect on GmHSD activity. Aspartate, lysine, and methionine showed slight inhibition (i.e. less than 10% at 100 mM). Threonine (100 mM) inhibited GmHSD ϳ40%. Detailed analysis of threonine inhibition of GmHSD (Fig. 5) showed the amino acid to be a competitive inhibitor versus Asa (K i ϭ 163 Ϯ 30 mM) and a non-competitive inhibitor versus NADPH (K i ϭ 239 Ϯ 36 mM). Importantly, the concentrations of threonine (and other aspartate-derived products) are significantly higher than physiological levels, suggesting that GmHSD is insensitive to feedback inhibition.
HSD Activity in Soybean-To determine the physiologic level of threonine inhibition of HSD activity in vivo, we measured enzymatic activity with and without exogenous threonine in seed, root, stem, and leaf tissues from soybean (Fig. 6). Measurement of HSD activity showed the highest activity in mature leaves and stems with the lowest activity in roots. The addition of threonine to the extracts inhibited HSD activity in all tissues but to varying degrees. Protein extracts from root and stem tissues showed an approximate 50% decrease in HSD activity after threonine addition. HSD activity in leaf tissues (both young and mature) was reduced to ϳ70% that of the original activity. Seed protein extracts displayed a slight 5% decrease in HSD activity. These results suggest that the monofunctional HSD is most active in seeds followed by roots and stems and then leaves. Moreover, the differences in HSD activity indicate the presence of threonine-sensitive HSD, most likely bifunctional AK-HSD proteins, in these tissues.
Kinetic Analysis of GmAK and GmAK-HSD-To examine the threonine inhibition of proteins encoded by reported sequences for the monofunctional AK (GenBank TM accession number AAD41796.1) and bifunctional AK-HSD (GenBank TM accession numbers AAC05983 and AAC05981; (20)) in soybean, each of these enzymes was cloned, expressed, and purified for kinetic characterization and analysis of threonine inhibition.
In both the forward and reverse HSD reactions, the bifunctional GmAK-HSD proteins displayed similar steady-state kinetic parameters with 2-3-fold differences in k cat /K m values (Table 3). In general, the catalytic efficiencies of the bifunctional and monofunctional HSD proteins were similar in the forward reaction (Tables 1 and 3), although GmHSD showed a consistent preference for NADPH as cofactor. The K m for Asa with NADPH as co-substrate is 12-20-fold higher in both of the GmAK-HSD than GmHSD. In the reverse reaction the monofunctional GmHSD is generally more efficient than the bifunctional AK-HSD proteins, which show a roughly 10-fold higher K m value for NAD ϩ . The AK activity of GmAK and both GmAK-HSD showed that the monofunctional protein was catalytically more efficient than the fused proteins but with a slightly higher K m for aspartate ( Table 4).
The effect of threonine on GmAK and each of the GmAK-HSDs was examined. GmAK displayed a K 0.5 ϭ 1.2 Ϯ 0.1 mM for threonine inhibition. AK activity in the bifunctional enzymes was inhibited by threonine, with K 0.5 ϭ 23 Ϯ 2 M and K 0.5 ϭ 25 Ϯ 1 M for GmAK-HSD-1 and GmAK-HSD-2, respectively. Using the forward reaction substrates (2.5 mM Asa and 50 M NADPH), the HSD activity in GmAK-HSD-1 and GmAK-HSD-2 showed K 0.5 values for threonine of 159 Ϯ 9 and 140 Ϯ 17 M, respectively. The inhibition constants for the bifunctional enzymes are nearly 1000-fold lower than those determined for GmHSD and similar to those reported for AK-HSD in Arabidopsis and maize (17)(18)(19).

DISCUSSION
In plants and microbes the synthesis of four essential amino acids occurs through the aspartate amino acid pathway. Within this pathway, AK and HSD are critical regulatory targets and vary in their structural organization. HSDs exist either as bifunctional enzymes in Gram-negative bacteria or as monofunctional enzymes in Gram-positive bacteria and yeast (11)(12)(13)(14). To date, the only HSD activity to be cloned and characterized in the plant kingdom has been as part of bifunctional AK-HSD (17)(18)(19)(20)(21). Here we describe the first molecular and kinetic analysis of a threonine-resistant HSD from soybean, demonstrating that plants, unlike bacteria and yeast, are unique in using both monofunctional and bifunctional HSD.
Sequence analysis of GmHSD shows that the monofunctional enzyme shares 20 -40% sequence identity with bacterial  . Initial velocity data for GmHSD. Enzyme assays were performed as described under "Experimental Procedures." Initial velocity data are shown as symbols in the reciprocal plots. All assays were performed in triplicate with less than 5% error between measurements. Lines represent the global fit of data to a sequential bi bi kinetic mechanism. A, shown is a double-reciprocal and yeast HSD and less than 40% identity with the HSD domain of the bifunctional enzymes from Arabidopsis. In soybean, Southern analysis indicates strong hybridization to one gene and much weaker hybridization with other related forms (Fig.  2). Together with the two previously isolated cDNAs coding for bifunctional AK-HSD from soybean (20), there are three HSDencoding genes in the G. max genome. Importantly, the presence of a monofunctional HSD in plants is not limited to soybean. A BLAST search of available databases reveals that multiple plants, including Medicago truncatula (barrel clover; 85%), Populus trichocarpa (California poplar; 76%), Ricinus communis (castor; 75%), Vitis vinifera (grape; 70%), A. thaliana (69%), Z. mays (maize; 66%), Sorghum bicolor (66%), and Oryza sativa (rice; 66%), contain sequences sharing high identity with GmHSD that are in many cases annotated as either unknown or hypothetical proteins or even incorrectly as AK. This suggests that monofunctional HSD in plants are more prevalent than previously described (22)(23).
To understand the biochemical properties of GmHSD, a detailed kinetic analysis of the enzyme was performed. For examining the kinetics of the forward and reverse reactions catalyzed by GmHSD, we used both NAD(H) and NADP(H) as co-substrates. Previously, the E. coli HSD was only analyzed in the forward direction with NADPH, and the yeast enzyme was analyzed using NAD(H) in each direction (10 -11). Likewise, kinetic studies of Arabidopsis AK-HSD have primarily relied on using the reverse reaction substrates Hse and NADP ϩ (21,29). Based on comparison of k cat /K m values for GmHSD, the forward reaction is favored by 50 -200-fold over the reverse reaction (Table 1). Moreover, use of the phosphorylated nucleotides is only modestly favored. Further experiments determined the kinetic mechanism of the enzyme.
Using a combination of initial velocity and product inhibition experiments ( Table 2) in both reaction directions, our data suggest an ordered bi bi mechanism for GmHSD. The intersect-  . HSD activity in soybean tissues. HSD activity was measured using Asa and NADPH under standard assay conditions as described under "Experimental Procedures" using protein extracts from seed (developmental stage 4), roots, stems, young leaves, and mature leaves. Activity measurements were performed in the absence (white) and presence (black) of exogenous 10 mM threonine. Bars represent S.E. (n ϭ 3).  ing patterns observed in the initial velocity studies (Fig. 4) eliminate a possible ping-pong mechanism, and the determined inhibition patterns (Table 2) are consistent with a reaction sequence in which nucleotide cofactor binds first and is released last (Fig. 7). The competitive inhibition pattern of product inhibition by nicotinamide cofactors in each direction indicates that the inhibitor (NADP ϩ or NADPH) and the varied substrate (NADPH or NADP ϩ ) both bind to free enzyme ( Table 2). The non-competitive patterns show that Asa or Hse bind as the second substrate ( Table 2). The kinetic mechanism for GmHSD is similar to that reported for the yeast enzyme (11). In contrast to the plant and yeast HSD, the E. coli enzyme displays either a fully ordered (10) or a preferred order random mechanism (30, 31) depending on pH of the assay conditions. Additional experiments using either isotope exchange or substrate dependence of kinetic isotope effects could conclusively distinguish between the random and ordered models (32,33). In contrast to the well characterized AK-HSD isoforms from plants (17)(18)(19)(20)(21), GmHSD is resistant to feedback inhibition by threonine (Fig. 5). The amino acid can act as an inhibitor of GmHSD but with K i values greater than 100 mM higher than the 300 M free threonine concentration in plants (34). For the plant and bacterial AK-HSD, the linker region between the two enzyme domains is critical for feedback regulation by threonine and for modulation of activity by other aspartate pathway amino acids (13,21). The lack of this region in GmHSD corresponds with resistance to threonine inhibition. Moreover, a major difference between the HSD of yeast and plants versus the enzyme in Gram-positive bacteria is resistance to feedback inhibition by threonine (1,14). The monofunctional enzymes from Corynebacterium and Bacillus have longer C-terminal regions implicated in feedback inhibition by threonine (35). A single amino acid change in the HSD of Corynebacterium was shown to be critical for inhibition by threonine (35). This residue is conserved across the HSD from Gram-positive bacteria but is absent in yeast because of a shorter C-terminal region (14). Similarly, GmHSD lacks the C-terminal extension, and biochemical analysis demonstrates that this enzyme is insensitive to inhibition by threonine at physiological levels.
Although the aspartate amino acid pathway is of major interest for metabolic engineering efforts aimed at enhancing essential amino acid content and nutritional value (4,36), the physiological role of threonine-insensitive HSD is not well understood (37) and represents a little explored feature of this pathway in plants. Here we show that threonine-sensitive and threonine-insensitive HSD activity in soybean differs by tissue type (Fig. 6). This difference results from the soybean expressing both monofunctional HSD and bifunctional AK-HSD, which are strongly regulated by threonine levels (Table 3). In leaf, stem, and roots the percent of HSD resistant to threonine is comparable with the activity seen in maize, in which threoninesensitive HSD accounts for ϳ50% of HSD activity (19). Earlier in vivo studies indicate that pathway activity changes in response to environmental conditions, including pH and illumination and spatial and temporal differences (19,38). For example, in soybean seeds amino acid biosynthesis occurs before production of seed storage proteins (6,39,40), and the differences in the response of HSD isoforms to threonine may be a mechanism for maintaining production of amino acids derived from aspartate during early seed development. Ultimately, the levels of threonine-sensitive and -insensitive HSD in plant tissues likely reflect a need to tailor metabolism in FIGURE 7. Proposed kinetic mechanism of GmHSD. Based on the kinetic analysis of GmHSD, NAD(P)H binds first followed by Asa. This leads to catalysis and release of Hse followed by release of NAD(P) ϩ .

TABLE 3
Kinetic parameters for HSD activity of GmAK-HSD-1 and GmAK-HSD-2 response to metabolic demands. It appears that plants, unlike microbes, require a full complement of monofunctional HSD, monofunctional AK, and bifunctional AK-HSD to integrate multiple metabolic inputs for modulating the aspartate amino acid pathway. Further studies aimed at understanding the interplay of these proteins in plants are required.
The monofunctional HSD of plants may provide a useful biotechnology tool. Typically, control steps in pathways can be bypassed by inactivating the regulatory feature or by isolating inhibition-resistant forms of the target enzyme. Overexpression of threonine-insensitive HSD or reduction of its expression by RNAi or other methods may alter the flux through the branch point that balances between production of lysine versus threonine and methionine.