Soybean seeds expressing feedback-insensitive cystathionine γ-synthase exhibit a higher content of methionine

Abstract

Soybean seeds provide an excellent source of protein for human and livestock nutrition. However, their nutritional quality is hampered by a low concentration of the essential sulfur amino acid, methionine (Met). In order to study factors that regulate Met synthesis in soybean seeds, this study used the Met-insensitive form of Arabidopsis cystathionine γ-synthase (AtD-CGS), which is the first committed enzyme of Met biosynthesis. This gene was expressed under the control of a seed-specific promoter, legumin B4, and used to transform the soybean cultivar Zigongdongdou (ZD). In three transgenic lines that exhibited the highest expression level of AtD-CGS, the level of soluble Met increased significantly in developing green seeds (3.8–7-fold). These seeds also showed high levels of other amino acids. This phenomenon was more prominent in two transgenic lines, ZD24 and ZD91. The total Met content, which including Met incorporated into proteins, significantly increased in the mature dry seeds of these two transgenic lines by 1.8- and 2.3-fold, respectively. This elevation was accompanied by a higher content of other protein-incorporated amino acids, which led to significantly higher total protein content in the seeds of these two lines. However, in a third transgenic line, ZD01, the level of total Met and the level of other amino acids did not increase significantly in the mature dry seeds. This line also showed no significant change in protein levels. This suggests a positive connection between high Met content and the synthesis of other amino acids that enable the synthesis of more seed proteins.

cystathionine γ, -synthase (CGS), Glycine max (L.) Merr., methionine, nutritional quality, proteins, sulfur-containing amino acids, soybean.

Introduction

Soybean [Glycine max (L.) Merr.] is an important staple source of vegetal protein in human food because of its high quality, competitive cost, palatability, and digestibility traits. Soybean seeds have a high protein content that varies from 35 to 50% and are about 20% oil, depending on the genotype and growing conditions. As it has the highest percentage of total protein content among all protein food sources such as meat, eggs, milk, and fish (Krishnan, 2005), many countries have started to use defatted soybean as an excellent source to feed farm animals. Although soybean is an excellent source of protein, its nutritional quality suffers from a low concentration of the sulfur amino acids, methionine (Met) and cysteine (Cys). This limitation can lead to non-specific signs of protein deficiencies in humans, such as lowered resistance to disease, decreased blood proteins, and retarded mental and physical development in young children (Jez, 2008). Met deficiencies in plant-derived feeds for farm animals limit animal growth, as well as affecting the quantity and quality of the products (Tabe and Higgins, 1998; Xu et al., 1998). Met, an essential amino acid, is the main limiting sulfur amino acid, because it can be converted by animals to Cys, and thus it meets the requirements for these two amino acids. The concentration of Met and Cys in soybeans is about 1.3g per 100g of protein, which falls short of the required amount of 3.5g per 100g (Shewry, 2000; WHO, 2007).

As the content of Met limits the nutritional value of soybean, attempts to increase Met content have been made in soybean seeds by traditional breeding and selection of mutants. However, these efforts have yielded only limited success, and the plants with enhanced Met content are usually associated with a significant reduction in yield (Imsande, 2001; Takahashi et al., 2003; Krishnan, 2005). Therefore, researchers have genetically engineered soybeans (Townsend and Thomas, 1994; Dinkins et al., 2001; Kim and Krishnan, 2004), as well as other legume seeds (Altenbach et al., 1992; Hagan et al., 2003; Molvig et al., 1997), to express heterologous seed proteins rich in sulfur-containing amino acids. Although in some transgenic seeds a net increase in the sulfur amino acid content was reported, the increase was not sufficient to meet the nutritional requirements of human, livestock, and poultry. Furthermore, these increases were often accompanied by a decrease in endogenous sulfur-rich proteins or sulfur compounds (Jung et al., 1997; Tabe and Droux, 2002; Hagan et al., 2003). This suggests that the synthesis of sulfur amino acids in developing soybean seeds is not sufficient to meet the demand created by the introduction of sulfur-rich proteins (reviewed by Tabe and Higgins, 1998; Amir and Tabe, 2006; Amir, 2010; Galili and Amir, 2013).

Hence, in recent years, efforts have also been directed towards an increase in the content of the soluble sulfur amino acids through manipulation of their biosynthesis pathways. It has been reported that overexpression of feedback-insensitive forms of serine acetyl transferase and O-acetylserine (thiol) lyase, the last two enzymes in the Cys pathway in the cytosol, lead to increases of up to 32% in soluble Cys and up to 58–74% in protein-bound Cys in soybean seeds (Kim et al., 2012). The overall increase in soybean total Cys content (both free and protein bound) satisfies the recommended levels required for the optimal growth of monogastric animals. In addition, transgenic soybean plants expressing mutated forms of key enzymes in the biosynthesis pathways of three additional essential amino acids, lysine, tryptophan, and threonine show significant increases in the levels of these amino acids (Falco et al., 1995; Kita et al., 2010; Qi et al., 2011). These promising results led us to manipulate the Met biosynthesis pathway in order to increase the level of Met in soybean seeds.

Despite recent accumulated knowledge on Met biosynthesis in vegetative tissues (reviewed by Ravanel et al., 1998; Hesse et al., 2004; Amir, 2008; Jander and Joshi, 2010), there is still little knowledge regarding Met biosynthesis in seeds (reviewed by Amir et al., 2012). Biochemical and genetic evidence show that Met in seeds can be synthesized by two pathways: from the aspartate family pathway through its first committed enzyme, cystathionine γ-synthase (CGS), similar to the pathway in leaves (Hacham et al., 2002; Kim et al., 2002), and/or from S-methylmethionine (SMM), which is produced from Met in non-seed tissues and transported to the seeds where it is then converted back to Met (Fig. 1) (Ranocha et al., 2001; Gallardo et al., 2007; Lee et al., 2008). Whilst the debate about the significance of SMM and aspartate pathways in Met synthesis in seeds is still ongoing, several studies have suggested that, in legume seeds, Met is synthesized mainly through the aspartate family via CGS. Tabe and Droux (2001) demonstrated that cotyledons of developing lupin seeds were able to transfer the sulfur atom from ³⁵S-labelled sulfate into seed proteins. This demonstrated the ability of the developing cotyledons to perform all the steps of sulfur reduction and sulfur amino acid biosynthesis in a similar manner to leaves, i.e. through the aspartate family pathway via CGS activity (see Fig. 1). Met is most probably also synthesized via CGS activity in developing narbon bean seeds, as enhancing the flux of the carbon/amino skeleton into the aspartate family biosynthesis pathway by expressing seed-specific, feedback-insensitive bacterial aspartate kinase results in a significant increase in Met content in seeds (Demidov et al., 2003). However, the regulatory role of CGS in legume seeds has not yet been studied. In this study, we used the Met-insensitive form of Arabidopsis CGS (AtD-CGS) (Hacham et al., 2006) expressed in soybean seeds to examine the role of CGS in Met synthesis in soybean seeds and to reveal whether this manipulation could lead to a higher Met content.

Fig. 1.

Scheme of the aspartate family biosynthesis pathway and Met metabolism. Only some of the enzymes and metabolites are indicated. Broken arrows represent more than one enzymatic step. AK, aspartate kinase; HK, homoserine kinase; TS, threonine synthase; CGS, cystathionine γ-synthase; CBL, cystathionine β-lyase; MS, Met synthase; SAM, S-adenosylMet; SAMS, SAM synthase; SMM, S-methylMet; MMT, SAM:Met S-methyltransferase; HMT, Hcy S-methyltransferase; MGL, Met γ-lyase.

Materials and methods

Materials

All chemicals were from Sigma Aldrich except where noted.

Binary plasmid construction

To achieve seed-specific expression of AtD-CGS, the 35S promoter of cauliflower mosaic virus was replaced by the seed-specific promoter of legumin B4 (kindly provided by Ulrich Wobus, Gatersleben, Germany; Baumlein et al., 1992), in a construct used previously (Hacham et al., 2006). This was done using HindIII and NcoI restriction enzymes. The fragment containing the promoter, the cDNA encoding the AtD-CGS, and the terminator was then subcloned by HindIII and XbaI digestion into the binary plasmid pGPTV-BAR, which carries the gene for glufosinate resistance (Becker et al., 1992).

Agrobacterium preparation

Agrobacterium strain EHA105 carrying the plasmid pGPTV-BAR was prepared as described previously (Paz et al., 2005). For plant transformation, the Agrobacterium was suspended in infection medium containing 10% Gamborg’s B5 salts (Gamborg et al., 1968), 10% MS iron, 3% sucrose, 20mM MES (pH 5.4), 10% B5 vitamins, 3.3mM cysteine and 1mM dithiothreitol, 7.4 µM 6-benzylaminopurine, 0.7 µM gibberellic acid, and 200 µM acetosyringone.

Plants and transformation

Soybean cultivar Zigongdongdou (ZD) was used for Agrobacterium-mediated transformation. The cotyledonary-node method (Zhang et al., 1999; Olhoft and Somers, 2001) was adopted with modifications as described below. Soybean seeds were sterilized with chlorine and germinated for about 4 d on germination medium (pH 5.7) containing B5 salts, B5 vitamins, MS iron and 2% sucrose. The seedlings were placed at 4 °C for 1 d before the transformation, and the cotyledons were then excised and sliced according to the method of Zhang et al. (1999). The sliced cotyledons were placed immediately into the Agrobacterium suspension and inoculated for 2h with occasional agitation. Ten sliced cotyledons were then placed on a sterile filter paper covering the co-cultivation infection medium solidified with 0.8% agar. After 3 d incubation in a growth chamber (18/6h light/dark at 24 °C), the explants were washed in sterilized water and imbedded in the solid shoot induction medium containing B5 salts, MS iron, 2% sucrose, 3mM MES (pH 5.7), B5 vitamin, 7.4 µM 6-benzylaminopurine, 100mg l^–1 of timentin, 0.42 mM of cefotaxime, 0.47 mM of carbenicillin, and 10 μM of glufosinate ammonium, solidified with 0.8% agar. After 7 d, the explants were subcultured in fresh shoot induction medium with 30 μM of glufosinate ammonium. Three weeks later, the explants were transferred to shoot elongation medium (Paz et al., 2005) containing 15 μM glufosinate ammonium. The elongated shoots were collected when they were about 3–4cm tall and transferred to rooting medium comprised of B5 salts, MS iron, 2% sucrose, 3mM MES (pH 5.7), B5 vitamins, 7.4 μM of indole-3-butyric acid, and 0.8% agar. After root emergence, the T₀ plants were transferred to soil matrix (vermiculite:peat:soil, 1:1:1).

Screening the transgenic plants

Screening of the T₁ plants was done by painting half of the adaxial surface of a young leaf with 120mg l^–1 of glufosinate. Selection of T₁ plants was done by spraying with 1.3mM of glufosinate ammonium when the first trifoliate was fully expanded. The glufosinate-resistant plants were examined by PCR for insertion of the AtD-CGS gene, using sense primer 5’-AGCATGCTCGTCCGTCAGCTGAGCATTAAAGC-3’ and antisense primer 5’-CATAACGACCATACTCGAAACTC-3’. Western blot analysis was performed as described previously (Hacham et al., 2006).

Growth conditions

The transgenic T₁ lines and wild-type used for PCR and Western blot analysis were planted in late April, at the Institute of Crop Science, the Chinese Academy of Agricultural Sciences, Beijing, PR China (39° 54’N, 116° 46’E). The pot was 10 litres in size with a dimension of 29cm in height and 32.5cm diameter at the top, and contained mixed loam soil with 5g of (NH₄)₂HPO₄ in each pot as basal fertilizer. The seedlings were grown under short day condition (12/12h, light/dark) after emergence. The 12h short day was controlled by transferring the pots into and out of a dark room each day, from 7:00 a.m. to 7:00 p.m., as described by Jiang et al. (2011).

T₃ homozygous transgenic lines and wild-type plants used for the seed measurements were grown at the Experimental Farm of Nanchong Municipal Academy of Agricultural Sciences, Sichuan Province, PR China. The geographic location is 30° 49’N, 105° 58’E. The seeds were sown in July 2011. The plot was 6×7.68 m, and the row space was 0.64 m. The plants matured in November 2011.

Seeds samplings

Seeds from T₃ homozygous transgenic lines were used. Green developing seeds were harvested at the middle stage of seed filling, between R5 and R6, as described by Fehr et al. (1971). Dry seeds were collected from pods taken from the lower nodes of the plant harvest after the plants and seeds were fully matured. For measurements of soluble and total amino acids, total protein and lipids contents, four samples (each comprising a pool of 12 seeds) were taken from three different plants for each line.

Determination of amino acids and total proteins

The extraction, derivatization, and analysis of the soluble and total seed amino acids were performed by gas chromatography/mass spectrometry (GC-MS,) as previously described (Golan et al., 2005).

The determination of total proteins was done according to the Kjeldahl method (Kim et al., 2006). For each sample, 25 soybean seeds were ground into powder using an Analytical Grinder A10 (IKA). The powder (200mg) was mixed with 0.5g CuSO₄ and 5g K₂SO₄ in a burette, followed by the gentle addition of 18ml H₂SO₄. Subsequently, the mixtures were heated to 340–370 °C for 2h for digestion in a Speed Digester K-435 (Büchi). Exhausted gas was circularly cleaned using a Scrubber B-414 (Büchi). When the samples had cooled to room temperature, they were placed into a Kjeldahl Apparatus K-370 (Büchi) for the determination of total protein content. Total proteins were estimated from the total nitrogen using the equation described by Benton Jones (1991).

In addition, the level of total proteins in each single seed of soybean was determined using the non-destructive analysis method of near-infrared (NIR) transmittance spectroscopy (Tajuddin et al., 2003). Measurements were taken at room temperature in a NIR Matrix I (Bruker). Data were collected from the spectrum of 4000–12 000cm^–1. The model was built based on the partial least squares method PLS, using protein data from typical soybean cultivars as the calibration set.

Determination of lipids in seeds

For lipid determination, the Soxhlet method was used (Kim et al., 2006). Soybean seeds (2g) were ground into powder using an Analytical Grinder A10 (IKA), and the powder was placed in a dried and weighed paper wrap and then placed in the extraction cylinder of a Soxhlet apparatus with anhydrous ether and incubated overnight. The anhydrous ether was transferred into an extraction tube and fresh anhydrous ether was added to the extraction cylinder. The solvent was boiled and refluxed in the extractor at 70–80 °C for 6h. The sample wrap with the remaining sample was dried in a desiccator followed by drying in an oven at 105 °C for 2h, and finally weighed and the lipid content calculated. The results are given as the percentage of total lipids (g per 100g of dry seeds).

Statistical analysis

The data obtained from this study were analysed statistically using JMP 8 software. In this software, we used analysis of variance and Student’s t-test, as described in the text. A P value of <0.05 was considered statistically significant.

Results and discussion

Generation of transgenic soybean seeds expressing AtD-CGS

To gain more knowledge on the regulatory role of CGS in Met biosynthesis in soybean seeds, we expressed AtD-CGS under the control of the seed-specific promoter of legumin B4 (Baumlein et al., 1992). This mutated form of CGS was selected because tobacco plants overexpressing this form have a significantly higher content of Met compared with plants overexpressing the feedback-sensitive form of AtCGS (Hacham et al., 2006, 2008). In addition, overexpression of AtD-CGS in alfalfa plants resulted in a significant elevation in the levels of soluble and total Met (about 29- and twofold, respectively) (Avraham et al., 2005). Therefore, the construct harbouring AtD-CGS was used to transform the ZD soybean cultivar. Thirty glufosinate-resistant soybean plants were screened for insertion of the AtD-CGS gene by PCR. Twenty-one positive lines were found to contain the transgene. Although PCR analyses showed the presence of the heterologous gene in the DNA (Fig. 2a), the AtD-CGS enzyme was not detected by Western blot analysis in mature dry seeds using antibodies previously used to detect AtD-CGS (Fig. 2b) (Hacham et al., 2006). A possible explanation for the low expression levels of AtD-CGS is that the activity of the legumin B4 promoter is relatively low in the mature dry seeds. Therefore, the protein levels of AtD-CGS were measured in green developing seeds harvested at the middle stage of seed filling (between the R5 and R6 stages, as described by Fehr et al., 1971). At this developmental stage, most of the seed storage proteins accumulate and amino acids are synthesized (Meinke et al., 1981; Jones et al., 2010). As shown in Fig. 2c, high expression levels of AtD-CGS were detected in the seeds. These results suggested that legumin B4 is expressed during seed development, but at a much lower level in the mature dry seeds. Three transgenic soybean lines, named ZD01, ZD24, and ZD91, exhibiting the highest expression levels of AtD-CGS in their seeds were self-pollinated to obtain homozygous plants. Seeds from T₃ lines were used for further analysis.

Fig. 2.

Presence and expression levels of AtD-CGS during seed development in seeds of wild-type ZD and three transgenic soybean lines. (a) PCR analysis. (b) Western blot analysis of mature dry seeds using antibodies against AtCGS. (c) Western blot analysis of green developing seeds using antibodies against AtCGS. Coomassie blue staining was used to check equal loading.

Transgenic soybean seeds accumulate significantly higher levels of soluble Met than wild-type seeds

To assess whether the high expression level of AtD-CGS affected Met content, the levels of soluble Met in the seeds of three transgenic and the wild-type soybean lines were measured. Examination of soluble Met levels in the green developing seeds revealed that all three transgenic lines exhibited higher levels of soluble Met than the wild-type seeds. Transgenic line ZD01 had the lowest expression level of AtD-CGS (Fig. 2c) with a 3.8-fold increase in soluble Met, whereas transgenic line ZD24 had the highest expression level of AtD-CGS with about a sevenfold increase in soluble Met (Table 1; Supplementary Data at JXB online). Unexpectedly, the levels of homoserine, a metabolite upstream of Met synthesis, in the aspartate family pathway (Fig. 1) also increased (Table 1), suggesting that high levels of Met enhanced the flux towards its biosynthesis pathway by affecting enzymes upstream of homoserine. The increase in the levels of homoserine also implied that homoserine kinase (HK) had achieved its maximum activity and as a result its substrate accumulated. This further suggests that, under such conditions, HK becomes a metabolic bottleneck and could be used as a target for engineering. As overexpressing HK in potato (Rinder et al., 2008) and Arabidopsis (Lee et al., 2005) did not lead to higher levels of Met, it will be worthwhile examining whether co-expression of CGS and HK can lead to a higher Met content in soybean seeds.

Table 1.

Soluble amino acid contents of in green developing seeds of wild-type ZD soybean cultivar and transgenic soybean lines expressing the AtD-CGS gene. The quantities of amino acids were analysed using GC-MS and their levels were normalized to norleucine and calculated as nmol per g fresh weight of seeds. The data presented represent the mean ± standard deviation of four samples taken from three different plants for each line. Statistically significant changes (P <0.05, using two-way ANOVA) are identified by different letters.

Amino acid	ZD	ZD01	ZD24	ZD91
Alanine	7832±110 b	9295±1238 b	12380±726 a	10094±783 b
Valine	1522±26 b	1030±76 c	2194±183 a	1688±134 b
Serine	1059±109 c	1524±223 b	2327±177 a	2479±194 a
Leucine	881±14 b	646±40 d	1302±79 a	729±44 c
Threonine	552±48 b	489±51 b	726±67 a	711±60 a
Isoleucine	801±16 a	372±39 b	327±12 b	345±13 b
Proline	199±24 b	185±22 b	377±23 a	381±69 a
Glycine	1495±345 ab	1352±185 b	2211±611 a	1719±245 ab
Homoserine	111±11 c	263±23.9 c	421±33 b	604±60 a
Methionine	132±15 c	513±79 b	916±95 a	863±62 a
Aspartate	1008±86 c	1036±89.9 d	1557±123 a	1310±94 b
Phenylalanine	403±21 b	171±16.6 c	616±54 a	370±75 b
Glutamate	8983±635 b	9840±1085 ab	11230±955 a	9511±1101 b
Asparagine	3854±547 ab	3574±281 b	4556±479 a	4523±676 a
Lysine	185±6 b	98±11.8 c	273±41 a	109±11 c
Tyrosine	387±9.5 b	160±28 c	840±115 a	192±29 c
Tryptophan	2509±53 a	2384±273 ab	1936±140 c	2090±270 bc
Total	31914±756 c	32937±3624 bc	43890±3945 a	37418±3912 b

Recent studies have shown that a high Met content leads to elevation of branched amino acids (Hacham et al., 2008) through the activity of Met γ-lyase, the Met catabolic enzyme (Fig. 1) (Joshi and Jander, 2009). Transgenic rice plants overexpressing an enzyme in the Cys pathway that accumulated higher levels of Met in their seeds had significantly higher levels of the branched amino acids (Nguyen et al., 2012). However, in the transgenic soybean seeds, the levels of the branched amino acids were not significantly different from those in the wild-type seeds, suggesting that Met is not degraded via Met γ-lyase.

Notably, in the transgenic soybean seeds, the elevation of soluble Met content was associated with significantly increased levels of most other amino acids compared with wild-type seeds (Table 1). As a result, the levels of total soluble amino acids in these seeds increased by 3.2% in ZD01, 17.2% in ZD91, and 37.5% in ZD24 (Table 1). The results indicated that increased Met levels during seed development trigger the accumulation of the other amino acids by as-yet-unknown processes. A similar result was also obtained when Met was applied exogenously to explant medium of developing soybean cotyledons, leading to a greater than threefold elevation in the free amino acid pool (Horta and Sodek, 1997). A similar observation was also reported in soybean seeds expressing a feedback-insensitive aspartate kinase: in addition to a high content of threonine, these transgenic seeds also showed a substantial increase in other major soluble amino acid levels (up to 3.5-fold increase; Qi et al., 2011). These researchers suggested that changes in the flux of threonine metabolism trans-regulated other pathways of seed amino acid metabolism. In these high-threonine seeds, the level of soluble Met increased significantly by 3.9-fold (Qi et al., 2011). In a similar manner, Met might also be involved in elevation of the other amino acids in seeds of the transgenic soybean lines. However, further studies are required to reveal the mechanisms behind this phenomenon.

Next, the level of soluble Met was determined in mature dry seeds of the three transgenic lines. The level of soluble Met increased by about 1.5- to twofold in the three transgenic lines in comparison with the wild-type seeds (Table 2; Supplementary Data). The reduction in the soluble Met levels between the green developing and mature dry seeds suggested that, at the later stages of seed development, Met synthesis is reduced due to low activity of the legumin B4 promoter, or that the soluble Met was incorporated into seed storage proteins. Similarly, the levels of most of the other amino acids were significantly reduced between the green developing seeds and dry mature seeds, in both wild-type and transgenic lines, suggesting that they were incorporated into seed proteins (compare Tables 1 and 2). Despite this incorporation into seed proteins, the levels of total soluble amino acids in the seeds were still significantly higher in the transgenic lines compared with those in the wild type (increase of 4.7% in ZD01, 9.7% in ZD24, and 14.2% in ZD91) (Table 2).

Table 2.

Soluble amino acid content in mature dry seeds of wild-type ZD soybean cultivar and transgenic soybean lines expressing the AtD-CGS gene. The quantities of amino acids were analysed using GC-MS, and their levels were normalized to norleucine and calculated as nmol per g fresh weight of seeds. The data presented represent the mean ± standard deviation of four samples taken from three different plants for each line. Statistically significant changes (P <0.05, using two-way ANOVA) are identified by different letters.

Amino acid	ZD	ZD01	ZD24	ZD91
Alanine	506±40 a	377±210 a	495±26 a	518±42 a
Valine	382±15 a	303±225 a	377±154 a	487±38 a
Serine	171±8 a	191±11 a	195±32 a	201±21 a
Leucine	198±24 a	127±9 a	175±89 a	187±15 a
Threonine	254±4 b	320±12 ab	321±57 a	318±25 ab
Isoleucine	347±18 a	357±185 a	409±172 a	447±29 a
Proline	444±20 a	437±24 a	532±1 a	500±29 a
Glycine	463±73 b	615±54 ab	997±359 a	494±158 b
Homoserine	9±2 c	36±12 b	30±4 b	55±3 a
Methionine	132±7 c	252±3 a	222±25 ab	206±16 b
Aspartate	1824±51 a	1586±428 a	1791±77 a	1849±135 a
Phenylalanine	492±105 b	473±124 b	796±89 a	785±92 a
Glutamate	2389±200 b	3429±327 a	2765±419 b	3519±208 a
Asparagine	1960±84 c	2116±59 a	2165±70 b	2240±74 ab
Lysine	71±5 c	81±10 bC	90±2 b	108±13 a
Tyrosine	246±44 b	236±7 b	347±81 a	253±17 ab
Tryptophan	3204±210 b	2954±61 a	2963±645 b	2790±286 b
Total	13091±221 b	13710±1560 ab	14366±293 ab	14958±663 a

The transgenic seeds exhibit higher levels of total Met and amino acids

To reveal whether a higher level of soluble Met affected total Met, including Met incorporated into seed proteins, amino acid levels were measured after protein hydrolysis. The analysis revealed that the level of total Met in seeds of the transgenic lines of ZD24 and ZD91 had increased significantly (2.3- and 1.8-fold, respectively), whilst a similar increase was not observed in the ZD01 line (Table 3; Supplementary Data).

Table 3.

Total amino acid content in dry mature seeds of wild-type ZD soybean cultivar and transgenic soybean lines expressing the AtD-CGS gene. The quantities of amino acids were analysed using GC-MS after protein hydrolysis, and their levels were normalized to norleucine and calculated as nmol per g fresh weight of seeds. The data presented represent the mean ± standard deviation of four samples taken from three different plants for each line. Statistically significant changes (P <0.05, using two-way ANOVA) are identified by different letters.

Amino acid	ZD	ZD01	ZD24	ZD91
Alanine	73893±2239 ab	70989±7472 b	93057±10674 a	89548±7472 ab
Valine	32873±2531 c	31433±4969 c	90418±4007 a	72104±2805 b
Serine	57140±4666 b	59285±1552 b	119141±10250 a	100402±7175 a
Leucine	77960±5024 c	69394±6688 c	155884±6494 a	128146±4546 b
Threonine	40199±4088 b	36318±3498 b	105463±11784 a	86114±8248 a
Isoleucine	34434±2441 c	31766±4311 c	85377±3481 a	68608±2436 b
Proline	87551±5000 b	89179±13029 b	133469±4885 a	110342±3419 ab
Glycine	114052±2767 a	136722±51390 a	168503±25693 a	117635±17985 a
Methionine	9592±328 c	9630±966 c	22340±2616 a	17471±1531 b
Aspartate	118249±4533 b	119871±9650 b	179000±32434 a	178977±22703 a
Phenylalanine	43465±3542 b	48933±8821 b	100281±20883 a	74477±14618 ab
Glutamate	219395±11634 a	219586±7121 a	340893±173230 a	349490±121261 a
Lysine	74171±2781 b	77956±11802 b	136656±8133 a	114198±5693 a
Tyrosine	25064±431 bc	21915±542 c	41917±5991 a	37232±4193 ab
Total	1008038±52007 c	1112978±80724 c	1772399±170802 a	1565991±119561 b

Based on the results obtained from ZD24 and ZD91, we suggest that the level of soluble Met in the wild-type seeds limits the Met content incorporated into proteins. Similar results have been obtained in several other studies. Soybean seeds expressing genes in the biosynthesis pathway of Cys showed that high soluble Cys in seeds led to an increase in protein-bound Cys (Kim et al., 2012). In addition, expression of bacterial feedback-insensitive forms of two key enzymes in the lysine biosynthesis pathway in soybean seeds caused several hundred-fold increases in free lysine and the total lysine content increased (Falco et al., 1995). However, the ability of the increased soluble amino acid to be incorporated into proteins seems to differ among different legumes seeds. Total Cys levels in lupin seeds expressing a Cys regulatory enzyme was not altered, although the free Cys increased by 26-fold (Tabe et al., 2010). These authors suggest that the content of sulfur-rich proteins was reduced during the breeding processes of lupin, and thus they had fewer sulfur sink proteins (Tabe et al., 2010).

The results obtained from the protein hydrolysis also showed that the level of total protein-incorporated amino acids increased significantly in lines ZD24 and ZD91 (Table 3). Moreover, a positive relationship was found between Met content and the total levels of other protein-incorporated amino acids: in transgenic lines ZD24 and ZD91, the levels of total Met increased significantly (2.33- and 1.82-fold, respectively), as well as the level of total amino acids (1.76- and 1.55-fold, respectively), whilst in transgenic line ZD01, both the level of total Met and the total levels of the other amino acids were not altered significantly (Table 3).

Effect of high Met content on seed protein content

The increase in the total amount of protein-incorporated amino acids in the ZD24 and ZD91 lines suggested that these seeds had a higher protein content. To verify this, we measured the nitrogen content using the Kjeldahl method and calculated the protein levels. We found that the increase in protein-incorporated amino acids was accompanied by a significant increase in protein levels in mature dry seeds of ZD24 and ZD91 (Table 4). However, this was not the case in ZD01, which did not exhibit a higher level of protein-incorporated amino acids (Table 3). Due to the importance of total proteins in soybean nutritional value, we also measured the total protein using the non-destructive analysis method of NIR transmittance spectroscopy (Tajuddin et al., 2003). The results were similar to those obtains by the Kjeldahl method (Table 4). A similar relationship between a higher Met content and an increase in total proteins was previously reported in soybean seeds: application of exogenous Met to immature cotyledons grown in vitro increased the protein content (Holowach et al., 1984a) and enhanced the ratio of 11S (Met rich) to 7S (Met poor) storage proteins (Holowach et al., 1984b, 1986). To gain an initial impression of whether the pattern of seed storage proteins was altered in the transgenic seeds, we examined the total soluble protein components using SDS-PAGE and staining with Coomassie Brilliant Blue (Supplementary Data at JXB online). No substantial differences in the band pattern between the transgenic and non-transgenic seeds were apparent. To gain an initial indication of whether the transcript level of the storage proteins was altered in the transgenic seeds, the level of the transcript contents of a representative for 11S (glycinin G1, seven codons for Met; 1.4 mol%) and for 7S (β-conglycinin β-chain, two codons for Met; 0.46 mol%), as well as one Met-rich protein (sucrose-binding protein, 11 codons for Met; 2.2 mol%) were measured by quantitative real-time PCR (Supplementary Methods at JXB online). The results showed that, in general (and not always significantly), the levels of glycinin G1 and β-conglycinin β-chain increased in ZD24 and ZD91, whilst the level of the sucrose-binding protein was reduced (Supplementary Data). These results implied that the effects obtained by exogenous application of Met may be different from the effect caused by continuous exposure to high levels of Met occurring in the transgenic seeds during seed development.

Table 4.

Percentage of total protein (measured by the Kjeldahl and NIR methods) and lipids (measured by the Soxhlet method), in wild type (ZD) and transgenic seeds. The data presented represent the mean percentage ± standard deviation of four samples taken from three different plants for each line. Statistically significant changes (P <0.05, using two-way ANOVA) are identified by different letters.

Line	Total protein (Kjeldahl)	Total protein (NIR)	Total lipids (Soxhlet)
ZD	46.30±0.17 b	46.55±0.05 d	15.42±0.04 a
ZD01	46.94±0.06 b	46.77±0.06 c	15.21±0.05 b
ZD24	48.36±0.17 a	47.69±0.04 a	14.44±0.03 d
ZD91	48.23±0.05 a	47.46±0.21 b	14.57±0.06 c

Based on the results obtained from analysis of the transgenic seeds, we assumed that the following cascade occurs in the seeds. A high expression level of AtD-CGS leads to higher soluble Met levels. This triggers the synthesis of other amino acids during seed development (Table 1) and thus more soluble amino acids can be incorporated into storage proteins during seed maturation (Table 3). As a result, more storage proteins are synthesized and the level of total proteins increases in the seeds (Table 4). This cascade was seen in ZD24 and ZD91, which exhibited a high content of total Met. In these lines, higher levels of soluble amino acids at the green developing stage were observed (37.5 and 17.2%, respectively; Table 1), as well as higher levels of amino acids after protein hydrolysis (75 and 55%, respectively; Table 3). In these lines, the levels of total proteins increased significantly compared with the wild type (2.06 and 1.93%, respectively; Table 4). However, in ZD01 line, the level of other soluble amino acids increased by only 5.4% (Table 1), the levels of amino acids after protein hydrolysis increased by 10% (Table 3), and the level of total proteins barely increased (Table 4). This suggested that differences between ZD01 and ZD24/ZD91 were caused by the ability to increase the soluble levels of amino acids during the green developmental stage, when proteins are synthesized and accumulated (Meinke et al., 1981; Jones et al., 2010). Most probably, the levels of soluble amino acids (including Met) in ZD01 were not sufficient to support the accumulation of additional proteins and, as a result, there was only a slight increase in the total protein level (Table 4).

The elevated protein content found in the transgenic lines ZD24/ZD91 is an important trait, as increasing the protein content of soybean is one of the main goals that breeders try to achieve (Kroben and Gibbons, 1962). In addition to increasing the protein content, the Met mol% (total Met content compared with total amino acids after protein hydrolysis) increased from 0.95 mol% in the wild type to 1.11 mol% in ZD91 and to 1.26 mol% in ZD24. A similar increase was not observed in ZD01: its Met mol% was slightly reduced to 0.86 mol%. Thus, in the seeds of ZD24 and ZD91, we achieved two objectives: an increase in Met mol% and an increase in the total protein content.

The content of total lipids is altered in the transgenic seeds

It has been reported that a high protein content in soybean seeds is associated with a decrease in oil concentration (Hernandez-Sebastia et al., 2005). Thus, the changes in total protein levels in the transgenic soybean seeds may affect the levels of lipids. Analyses by the Soxhlet method revealed that the total lipid content of the transgenic seeds was slightly but significantly reduced compared with that of the wild type (Table 4). ZD24 and ZD91 showed a larger reduction in comparison with ZD01. Thus, these results are in accordance with the literature (Hernandez-Sebastia et al., 2005), showing a negative relationship between proteins and total lipids.

The transgenic lines have normal seed morphology and normal germination rate

The transgenic soybean seeds expressing the AtD-CGS gene exhibited normal seed morphology. In order to evaluate the effect of AtD-CGS expression on soybean seed germination rate, 50 transgenic seeds from each of the three transgenic lines as well as the wild-type seeds were planted in soil under growth-chamber conditions. No differences in germination rates between the transgenic lines and wild-type seeds were observed. These results suggested that, like high-threonine soybean seeds (Qi et al., 2011), but unlike high-lysine soybean seeds that showed poor germination (Falco et al., 1995), seeds with a high soluble Met content have no deleterious pleiotropic effects on soybean agronomic performance.

Final comments

The results obtained in this study suggest a link between a high Met content during seed development and the synthesis of other amino acids (through an as-yet-unknown process); thus, more soluble amino acids can be incorporated into storage proteins during seed maturation (Table 3). As a result, more storage proteins can be synthesized and the level of total proteins increases in the seeds (Table 4).

The results obtained in this study indicate that CGS plays a major role in Met synthesis in soybean seeds. Our results are in accordance with other reports from developing seeds of chickpea and narrow-leaf lupin, showing that legume seeds have the ability to synthesize Met and Cys de novo (Tabe and Droux, 2001; Chiaiese et al., 2004; Tabe et al., 2010; Kim et al., 2012). However, further studies are required to assess whether, in soybean seeds, Met can also be synthesized from SMM at the latter stages of seed development, as has been suggested for Medicago truncatula (Gallardo et al., 2007).

As the level of soluble Met in the transgenic seeds remained high (Table 2), even though most of the Met had been incorporated into proteins, we assume that other as-yet-unknown amino acid(s) become rate limiting for protein synthesis, and thus Met cannot be incorporated further into proteins. Another explanation for this phenomenon is that the capacity of Met codons in natural seed storage proteins in the soybean ZD cultivar limits the incorporation of the overproduced Met into seed storage proteins. This option was suggested previously for Cys in lupin plants (Tabe et al., 2010).

Taken all, the results obtained in this study constitute a major step towards the production of soybean seeds with higher Met levels and higher total protein levels, and thus with better nutritional value.

Abbreviations:

Abbreviations:
 
• Cys

  cysteine

 
• CGS

  cystathionine γ-synthase

 
• GC-MS

  gas chromatography/mass spectrometry

 
• HK

  homoserine kinase

 
• Met

  methionine

 
• NIR

  near-infrared

 
• SMM

  S-methylmethionine

 
• ZD

  Zigongdongdou.

Acknowledgments

The authors are grateful to Jinlu Tao, Weiwei Yao, and Xingzhi Duan (CAAS) for their technical assistance of the soybean transformation, to Gidi Baum (Migal) for his critical reading of the manuscript and to Dr Jihong Liang and Professors Wenbin Li, Xingguo Ye, Ping Wang, and Gad Galili for their valuable suggestions and comments. This work was supported by the Major Science and Technology Projects of China (2011ZX08004–003), Beijing Municipal Natural Science Foundation (5042019), and the Israel Science Foundation (ISF, 231-09).