• Open Access

Fortifying plants with the essential amino acids lysine and methionine to improve nutritional quality


  • Gad Galili,

    Corresponding author
    • Department of Plant Science, The Weizmann Institute of Science, Rehovot, Israel
    Search for more papers by this author
  • Rachel Amir

    Corresponding author
    1. MIGAL Research Institute, Kieyat Shmona, Israel
    2. Tel Hai College, Upper Galilee, Israel
    • Department of Plant Science, The Weizmann Institute of Science, Rehovot, Israel
    Search for more papers by this author

Correspondence (Tel +972 8 9232511 and +972 4 6953516; fax +972 8 9344181 and +972 4 6944980; email gad.galili@weizmann.ac.il,rachel@migal.org.il)


Humans, as well as farm animals, cannot synthesize a number of essential amino acids, which are critical for their survival. Hence, these organisms must obtain these essential amino acids from their diets. Cereal and legume crops, which represent the major food and feed sources for humans and livestock worldwide, possess limiting levels of some of these essential amino acids, particularly Lys and Met. Extensive efforts were made to fortify crop plants with these essential amino acids using traditional breeding and mutagenesis. However, aside from some results obtained with maize, none of these approaches was successful. Therefore, additional efforts using genetic engineering approaches concentrated on increasing the synthesis and reducing the catabolism of these essential amino acids and also on the expression of recombinant proteins enriched in them. In the present review, we discuss the basic biological aspects associated with the synthesis and accumulation of these amino acids in plants and also describe recent developments associated with the fortification of crop plants with essential amino acids by genetic engineering approaches.


Humans and farm animals are unable to synthesize a number of essential amino acids and therefore need to obtain them from their diets. In developed countries, these essential amino acids are generally provided from the dietary utilization of farm animals (particularly meat, eggs and milk) as well as from a variety of crop plants (particularly cereals and legumes) that together provide optimal levels of essential amino acids. In contrast, poor people from developing countries, whose foods are mostly derived from the major crop plants, suffer from particular deficiencies in the essential amino acids, Lys and Met. These two amino acids belong to the Asp family pathway, which is also responsible for the synthesis of two other essential amino acids, Thr and Ile (Figure 1) (Galili et al., 2005). Lys and Met are the most limiting essential amino acids in cereal and legume crops, respectively (Galili et al., 2005; Ufaz and Galili, 2008; Wenefrida et al., 2009). The limited Lys and Met contents reduce the nutritional values of these crop plants to 50–75%, compared to those of a diet possessing balanced levels of essential amino acids. This limitation in essential amino acids can lead to nonspecific signs of protein deficiencies in humans, such as lowered resistance to diseases, decreased blood proteins and retarded mental and physical development in young children. This syndrome is referred to as protein-energy malnutrition (PEM), and the World Health Organization (WHO) estimates that around 30% of the populations in the developing world suffer from this syndrome.

Figure 1.

The aspartate family pathways leading to lysine and methionine, threonine and isoleucine. Only several enzymes and metabolites are specified. Broken arrows represent more than one enzymatic step. Abbreviations: OPH, O-phosphohomoserine; ASD, aspartate-semialdehyde; DHDP, dihydrodipicolinate; AK, aspartate kinase; TS, threonine synthase; CGS, cystathionine γ-synthase; SAM, S-adenosylmethionine; SAMS, SAM synthase; MGL, methionine γ-lyase; DHDPS, dihydrodipicolinate synthase; LKR/SDH, bifunctional Lys-ketoglutarate reductase/saccharopine dehydrogenase.

Insufficient levels of Lys and Met in the major cereal and legume crops, respectively, also cause important biotechnological problems in developed countries because these two essential amino acids are important nutritional components in the feeds of livestock, which in turn contribute an extensive portion of the Lys and Met required in the diets of people living in these countries. Taken together, enhancing the levels of Lys and Met in the major crop plants is an important issue both from humanitarian and economical points of view, and such enhancements have so far been shown to be complex that are generally very difficult to achieve by classical breading. Two major reasons (although not the only ones) for this difficulty are (i) while the enhanced Lys and Met levels are generally required in seeds (particularly seeds of cereal crops), elevating their content by classical genetics generally leads to their accumulation also in vegetative tissues, exerting major deleterious effects on plant growth (Bright et al., 1982; Ghislain et al., 1995; Heremans and Jacobs, 1997), and (ii) enhanced incorporation of the overproduced Lys and Met into the seed storage protein is required to maintain high levels of these two amino acids (Ufaz and Galili, 2008). Nevertheless, the capacity of Lys and Met codons in natural seed storage proteins limits the incorporation of the overproduced amino acids into the seed storage proteins (Amir and Tabe, 2006). An alternative approach to solve these two problems is using genetic engineering because genetic engineering can utilize (i) seed-specific promoters to fortify Lys and Met specifically in the seeds, which is particularly important for cereal and legume crops, and (ii) recombinant proteins enriched in either Lys and/or Met. In the present review, we discuss a variety of efforts that were aimed to improve Lys and Met levels in plants by either classical breeding or genetic engineering approaches. We also discuss how understanding the biological processes associated with Lys and Met metabolism has assisted in designing approaches to improve the content of these essential amino acids with minimal negative effects on plant growth and productivity.

While breeding approaches are generally performed directly on the crop of interest, generally genetic engineering approaches are first tested on model plants, particularly due to the availability of efficient transformation technologies in these plants. Only latter are the desired genes used to transform crop plants. Hence, for the illustration of the chronological order of the studies associated with fortifying plants with essential amino acids, we initiate our review by discussing older genetic/breeding approaches aimed to improve Lys content in seeds of maize, one of the world's most important crops. We then continue by focusing on the more modern genetic engineering approaches to improve Lys and Met levels in plants, starting from studies using model plants and continuing with the translation of the information generated from the model plants to improve Lys content in maize seeds.

The development of maize genotypes with high Lys levels in their seeds by advanced breeding approaches

Because Lys levels are particularly poor in cereal seeds (WHO, 2007), major attempts to improve Lys levels in cereals were performed during the 20th century, using a variety of approaches. A major focus was placed on corn, a worldwide staple crop contributing over 50% of the dietary protein for human and livestock, but whose kernels generally possess suboptimal Lys level. The earlier studies using genetic approaches led to the discovery of high-Lys corn genotypes possessing a variety of mutations, particularly the high-Lys opaque2 mutation (Mertz, 1976; Mertz et al., 1964). Opaque2 maize mutants contain low levels of the Lys-poor zein seed storage proteins as well as a compensatory increase in the levels of other Lys- and Trp-rich seed proteins, compared to wild-type maize genotypes. Consumption of opaque2 seeds can significantly increase the growth of rats in feeding trials (Mertz et al., 1964), improve the nitrogen balance in adolescent boys (Korslund et al., 1977) and have similar nutritional quality to milk protein when being fed to Guatemalan children (Bressani, 1966). These results raised extensive hopes that opaque2 seeds can cure children from developing countries who suffer from the protein deficiency disease kwashiorkor (Harpstead, 1971). The success with opaque2 maize also stimulated extensive efforts to breed for similar mutants in other cereal crops (Azevedo and Arruda, 2010; Doll et al., 1974). Nevertheless, the classical opaque 2 mutants in maize and other cereal crops have never been commercially utilized because the opaque 2 mutations exerted negative effects on a variety of other important agronomical parameters, particularly seed parameters.

The hope of utilizing the opaque2 mutation commercially has been boosted at the beginning of the 1990s during which new opaque2-derived maize genotypes, called ‘quality protein maize’ (QPM), were developed (Gibbon and Larkins, 2005; Glover, 1992). The QPM genotypes generally possess normal kernel properties and comparable yields to other maize cultivars, rendering them quite nutritionally effective, particularly for young children in developing countries (Akalu et al., 2010). In comparison with Lys level of approximately 0.4% in normal maize seeds, the corresponding Lys value in seeds of maize opaque 2 and QPM lines is nearly doubled to approximately 0.8% (Gibbon and Larkins, 2005) (Table 1). Despite the commercial importance of the maize QPM lines, attempts to breed for similar genotypes in other crop species have not been successful, calling for additional approaches to improve Lys levels, particularly through genetic engineering.

Table 1. Lysine contents in organs of different mutants (M) or transgenic lines (T) compared to the wild-type progenitor lines (WT)
PlantOrganManipulationFree Lys*Total LysPhenotypeReferences
WT T Fold increase (T or M/WT)WTM or TFold increase (T or M/WT)
  1. *,†Free and total measured Lys in different laboratories and different approaches; NR, not reported; ND, not detected; mol%, % of Lys from total amino acids.

  2. Ppm, parts per million; M, mutant; T, transgene; DW, dry weight.

MaizeSeedsOpaque2 mutantNRNRNR1.74 mg/g DW2.94 mg/g DW1.7Abnormal, soft and starchyMertz et al. (1964)
MaizeSeeds‘Quality protein maize’ (QPM)/Opaque2 mutantNRNRNR2% of total amino acids3.9–4.1% of total amino acidsApproximately 2NormalGibbon and Larkins (2005)
MaizeSeedsQPMNRNRNR3 mg/g DW4.5 mg/g DW1.5Hard flinty endospermBetrán et al. (2003)
MaizeSeedsOpaque2 mutant15 ppm556 ppm37

3022 ppm;


% Lys/P

4744 ppm;


% Lys/P

1.78NDFrizzi et al. (2010)
MaizeSeedsHigh-Lys maize genotype (LY038)0.05 mg/g DW0.96 mg/g DW19.22.6 mg/g DW4.66 mg/g DW1.8NRLucas et al. (2007)
CanolaSeedsSeed-specific feedback-insensitive DHDPS and feedback-insensitive AKNRNR100Approximately 6% of total amino acids12% of total amino acids2NDFalco et al. (1995)
SoybeanSeedsSeed-specific feedback-insensitive DHDPS and feedback-insensitive AKNR34% of total amino acids100Approximately 6% of total amino acids25–33% of total amino acids4.1–5.5Wrinkled seeds, low germinationFalco et al. (1995)
RiceSeeds35S:: DHDPS1.777.54.23NRNRNRMorphologically distinguishLee et al. (2001)
RiceSeedsGluB:: DHDPS1.772.02No changeNRNRNRMorphologically distinguishLee et al. (2001)
ArabidopsisSeedsPhaseolin:: DHDPS and Lys-ketoglutarate reductase/saccharopine dehydrogenase (LKR/SDH) RNAi0.29–0.39 mol%25 mol%645.12 mol%7.28 mol%1.42Reduction in seed germinationZhu and Galili (2004)
MaizeSeedsLKR/SDH RNAi   0.03 mg/g DW0.65 mg/g DW21.6NRHoumard et al. (2007)
MaizeSeedsLKR/SDH RNAi endosperm & embryo0.04 nmol/g DW1.60 nmol/g DW40NRNRNRNDReyes et al. (2009)
MaizeSeedsFeedback-insensitive DHDPS and LKR/SDH RNAi100 ppm4000 ppm402500 ppm; 0.1 mg/g DW

6500 ppm;

4 mg/g DW

2.6NDFrizzi et al. (2008)
MaizeSeedsBoth 19- and 22 kD α–zeins RNAi15 ppm131–167 ppm8.7–11.1

3022 ppm;


% Lys/P

5054 ppm;


% Lys/P

2.1–2.2NDFrizzi et al. (2010)
MaizeSeeds22 kD zein RNAiNRNRNRNRNR18.5% elevationNDSegal et al. (2003)
MaizeSeeds19 and 22 kD zein RNAiNRNRNR2.72 mg/g DW3.65 mg/g DW1.35 elevationNDHuang et al. (2004)
MaizeSeedsSeed-specific feedback-insensitive DHDPS and α-zein RNAi43 ppm2908 ppm67.62575 ppm6160 ppm2.4NDHuang et al. (2005)
MaizeSeedsBoth 19- and 22 kD α-zeins RNAi25 ppm; 0.025 mg/g DW

70 ppm;

0.07 mg/g DW


2438 ppm;

2.83% Lys/P

5003 ppm;

5.62% Lys/P

2NDHuang et al. (2006)
MaizeSeedsGlutelin:: amaritinNRNRNR0.34% 100 g seeds0.40% 100 g seeds18%NRRascn-Cruz et al. (2004)
RiceSeeds13 kD prolamin RNAi134 nmol/g DW181 nmol/g DW1.326 (μmol/gDW)41 (μmol/gDW)1.6NDKawakatsu et al. (2010a)
RiceSeedsBip over-accumulationNRNRNR20 (μmol/gDW)57 (μmol/gDW)1.8Reduced grain weight, lower starch contentKawakatsu et al. (2010b)
MaizeSeedssb401 gene encoding a protein with high lysine contentNRNRNR

0.31 g/100 g seed

0.028 g/100 g protein

0.48 g/100 g seed

0.032 g/100 g protein

1.5NDYu et al. (2005)
TobaccoLeavesFeedback-insensitive DHDPS and the storage protein S-VSPNRNRNR4.85 mol%6.7 mol%1.4NDGuenoune et al. (2002)

Enhancing the level of free Lys in Arabidopsis seeds by modifying biosynthetic and catabolic fluxes

Lys level is of particularly major importance in the seeds because seeds provide the major nutritional source of human foods and livestock feeds worldwide. Hence, in this section, we focus on the most pronounced studies attempting to improve Lys level specifically in seeds of Arabidopsis (a model species) and maize (an important crop plant). To elucidate the regulation of Lys metabolism and the potential for enhancing Lys production in Arabidopsis seeds, Zhu and Galili (Zhu and Galili, 2003, 2004) have used a transgenic approach: A chimeric gene encoding a bacterial feedback-insensitive dihydrodipicolinate synthase (DHDPS) enzyme of Lys synthesis was expressed under the control of a seed-specific promoter in an Arabidopsis knockout mutant lacking a bifunctional Lys-ketoglutarate reductase/saccharopine dehydrogenase (LKR/SDH) enzyme of Lys catabolism (Zhu and Galili, 2003, 2004). This genotype (termed the KD genotype) led to a nearly 64-fold increase in the level of seed free Lys (Table 1). Yet, seed germination of the KD genotype was significantly retarded (Zhu and Galili, 2003, 2004). A subsequent study showed that boosting Lys synthesis and blocking its catabolism in the KD genotype had a major influence on the levels of several TCA cycle metabolites, indicating that the suboptimal germination of this mutant is due to a negative influence on the cellular energy status (Angelovici et al., 2011, 2009; Zhu and Galili, 2004). These studies have also exposed the significant contribution of both synthesis and catabolism fluxes to the accumulation of free Lys in seeds (Angelovici et al., 2011; Galili, 2011). They have also exposed a significant contribution of catabolic fluxes of amino acids of the Asp family pathway to the cellular energy homoeostasis both during seed development and germination. The metabolic link of Lys catabolism with the TCA cycle and its impact on various physiological parameters associated with cellular energy status have also been recently demonstrated experimentally (Angelovici et al., 2011; Galili, 2011).

Enhancing Lys level in the seeds of important crops by genetic engineering

The elucidation of regulatory factors associated with enhanced Lys accumulation in models plants, particularly tobacco and later Arabidopsis (Karchi et al., 1994; Zhu and Galili, 2004), provided a major trigger to the translation of this information into enhancing Lys levels in seeds of three major crops, namely, soybean, rapeseed and maize. In the case of soybean, in which the embryo constitutes the main section of the seed, the expression of a bacterial feedback-insensitive DHDPS under the control of an embryo-specific promoter caused a 100-fold increase in the accumulation of seed free Lys level (Falco et al., 1995) (Table 1). In the case of maize, in which the endosperm accounts for the majority of the seed volume, the bacterial feedback-insensitive DHDPS was expressed under two different promoters that are expressed either in the embryo or in the endosperm. Interestingly, even though the endosperm comprises the major part of the maize grain, only the expression of the bacterial feedback-insensitive DHDPS under the embryo-specific promoter caused a significant enhancement of seed soluble Lys level (Frizzi et al., 2008).

High Lys levels were also obtained when the expression level of LKR/SDH was suppressed. Suppression of these genes in the embryo led to an accumulation of 0.2 mg/g DW of soluble Lys in maize seeds, whereas the suppression in the endosperm resulted in an accumulation of 0.9 mg/g DW (Reyes et al., 2009). A synergistic effect on soluble Lys accumulation was observed when embryo and endosperm suppression were combined, reaching 1.6 mg/g DW (Reyes et al., 2009) (Table 1).

Combining the endosperm-specific DHDPS trait with the suppression of Lys catabolism, using an endosperm-specific LKR/SDH RNAi construct, significantly boosted seed soluble Lys level in maize seeds (Houmard et al., 2007). The reason for this observation is still not clear, but it can be due to the fact that the maize LKR/SDH gene is expressed specifically in the endosperm tissue of the maturing maize kernel (Azevedo and Arruda, 2010; Kemper et al., 1999). Expression of a recombinant transgene encoding a full-length bacterial feedback-insensitive DHDPS, together with a LKR/SDH RNAi sequence installed within an intron of this transgene, caused substantial improvement of free Lys level in maize seeds, amounting to 4000 ppm, compared to approximately 100 ppm in the control seeds (Frizzi et al., 2008) (Table 1).

An attempt to translate the research discussed above into commercial application has been recently reported by the Monsanto/Renessen companies who generated a high-Lys maize genotype (LY038) expressing a bacterial feedback-insensitive DHDPS in an embryo-specific manner (Dizigan et al., 2007). This genotype was further used to produce a maize LY038 × MON 810 hybrid, which also proved to be superior in broiler chicken performance, compared to the same lines that lacked the LY038 trait (Lucas et al., 2007). The LY038 maize has also been approved for commercial use in the livestock feeding industry in a number of countries. However, as indicated on the ‘GMO Compass’ Internet website, the LY038 maize was subsequently withdrawn. Based on the information existing on the internet site (http://foodfreedom.wordpress.com/2009/11/11/high-lysine-gm-maize-withdrawn-safety-concerns/), withdrawal was apparently due to concerns made by the EU regulators on the European Food Safety Agency (EFSA) GMO panel about the safety of LY038 maize for human consumption, even though LY038 maize was targeted to the feeding industry.

Attempts were also made to decrease the level of zeins using antisense and RNAi approaches. This approach was based on studies showing that in seeds of opaque2, the levels of zeins (preferentially 22 and 19 kD α-zeins) were significantly reduced, while the levels of the albumin and the globulin groups of proteins were significantly increased. Because these α-zeins are devoid of Lys and their decrease led to an increase of proteins rich in Lys, the reduction in the levels of zeins led to significantly higher levels of total Lys (Huang et al., 2004) (Table 1).

The combination of the two transgenic strategies described above, namely, reducing the 19 kD zein contents (Huang et al., 2004) and elevating the content of soluble Lys by expressing feedback-insensitive DHDPS (Frizzi et al., 2008), led to a synergistic effect resulting in maize seeds with high Lys content (Huang et al., 2005) (Table 1). Also, reduction in the content of the zein seed storage proteins of maize, using an RNAi approach in a genetic background of an opaque2 mutation, caused a considerable increase in the levels of free Lys and free Trp (another essential amino acid) as well as in the total kernel-free amino acids (Frizzi et al., 2010).

Enhancing the level of free Met in plants by modifying biosynthetic and catabolic fluxes

Traditional plant breeding methods have failed to increase the level of Met in crop plants. Attempts were also made to screen for mutants that are resistance to the Met analogue, ethionine (Green and Phillips, 1974; Reisch and Bingham, 1981), or that have the ability to grow in the presence of inhibitory concentrations of Lys and threonine (Diedrick et al., 1990). However, when the resistant calli were regenerated, the plants showed a high frequency of morphological abnormalities with significantly reduced yield (Reisch and Bingham, 1981). The development of transformation systems and molecular biology techniques made additional tools available to increase the Met content in plants.

Several approaches were examined to increase the Met level in plants by manipulating Met biosynthesis and/or catabolic pathways. To date, the most promising metabolic engineering approach leading to higher Met content with minimal perturbation to the plant phenotype was the manipulation of the first committed enzyme of the Met biosynthesis pathway, cystathionine γ-synthase (CGS). Overexpression of Arabidopsis CGS (AtCGS) in transgenic Arabidopsis led to 6.2-fold elevation of soluble Met content in leaves (Kim et al., 2002) (Table 2). Also, high levels of soluble Met were reported in leaves of potato (6.5-fold) (Di et al., 2003) and tobacco (12.8-fold) (Hacham et al., 2002, 2008) plants overexpressing this gene (Table 2). The transcript level of Arabidopsis CGS (AtCGS) is negatively regulated by the Met downstream product, S-adenosyl-Met (SAM), via a post-transcriptional mechanism (Chiba et al., 2003; Onouchi et al., 2005). Thus, it was suggested that the level of Met could not increase behind a certain threshold (Onouchi et al., 2005) that apparently differs between plant species. According to this assumption, the level of Met would increase when a Met/SAM feedback-insensitive form of AtCGS would be expressed to a similar expression level as the AtCGS. Indeed, expression of a mutated Met-feedback-insensitive AtCGS (D-AtCGS) in tobacco led to 2.9-fold higher level of soluble Met in leaves, compared to plants overexpressing the full-length AtCGS, and 37-fold higher level when compared to wild-type plants (Hacham et al., 2006). Expression of this mutated form, D-AtCGS, in alfalfa plants increased the total Met level (soluble and incorporated into protein) up to 2.1-fold, compared to the wild-type genotype (Avraham et al., 2005). However, overexpression of a potato CGS in transgenic potato plants, even though it caused a significant increase in CGS activity, did not enhance the Met level (Kreft et al., 2003), implying that various plant species differ in the stringency of their control of Met production and accumulation.

Table 2. The changes of Met contents in several tissues of various transgenic lines (T), expressing different genes under the control of the constitutive promoter 35S CaMV, compared to wild-type tissue (WT)
PlantTransgeneFree Met (nmol/gFW)Total MetPhenotypeReferences
TissueWTTFold increase (T/WT)
  1. a

    DO—Days old.

  2. b

    ND—not determined.

  3. c

    D-AtCGS—deleted form of AtCGS.

  4. d

    N-AtCGS—truncated form of AtCGS.

  5. e

    bAK—bacterial feedback-insensitive aspartate kinase.

ArabidopsisAtCGSLeaves 12-DOa402506.2NDbNormalKim et al. (2002)
ArabidopsisAtCGSLeaves 53 DOa9273NDNormalKim et al. (2002)


nmol/gFW (2.1-fold)

NormalAvraham et al. (2005)
TobaccoAtCGSLeaves59812.8NDSlightly abnormal(Hacham et al., 2008)
TobaccoN-AtCGSdLeaves5125251.1–3 mol%AbnormalHacham et al. (2008), Hacham et al. (2002)
TobaccoD-AtCGScLeaves518537NDSlightly abnormalHacham et al. (2008)
TobaccoD-AtCGSc/bAKeLeaves5866173.2NDSlightly abnormalHacham et al. (2008)
PotatoAtCGSLeaves10656.5NDNormalDi et al. (2003)
PotatoAtCGSTubers402406NDNormalDi et al. (2003)
PotatoD-AtCGSc/zeinTubers10909NDAbnormalDancs et al. (2008)
PotatoStCGSLeaves33No changeNo changeNormalKreft et al. (2003)
TobaccoAntisense SAMSLeaves125200433NDAbnormalBoerjan et al. (1994)
PotatoAntisense StTSLeaves1240240NDAbnormalZeh et al. (2001)
PotatoAntisense StTSTubers23819NDAbnormalZeh et al. (2001)
ArabidopsisAntisense AtTSLeaves 20 DOa1994149.5NDAbnormalAvraham and Amir (2005)

In addition to the regulatory role of AtCGS in Met biosynthesis, studies using feeding (Lee et al., 2005) and transgenic approaches (Karchi et al., 1993) suggest that the levels of the carbon amino precursors of Met synthesis also limit Met accumulation. Indeed, tobacco plants expressing both a bacterial feedback-insensitive aspartate kinase (AK), which causes an increased flux towards the Thr branch of the Asp family (Figure 1) (Shaul and Galili, 1992), and a mutated AtCGS that is insensitive to inhibition by SAM had 173-fold higher soluble Met contents in their leaves than the control, wild-type plants (Hacham et al., 2008) (Table 2). In plants, the level threonine (Thr), an additional essential amino acid, increased 2-fold.

Beside the efforts mentioned above, researchers also attempted to increase the level of Met by reducing the level of Thr synthase (TS), the last enzyme in the Thr biosynthesis pathway (Figure 1). These attempts were based on modelling studies suggesting that TS competes with CGS for their common carbon–amino substrate, O-phosphohomoserine (OPH) (Figure 1) (Amir et al., 2002; Curien et al., 2003). Transgenic potatoes having reduced expression of TS accumulated 240-fold higher levels of soluble Met in their leaves than did the control wild-type plants (Zeh et al., 2001).

Similarly, transgenic and mutant Arabidopsis having low TS expression level have 50-fold and 15-fold more soluble Met in their leaves, respectively (Avraham and Amir, 2005; Bartlem et al., 2000). However, the increased Met content was at the expense of Thr, whose level was significantly reduced, questioning the biotechnological suitability of this genotype.

Attempts to increase Met content also focused on the suppression of expression of the gene encoding the Met catabolic enzyme SAM synthase, the first and main enzyme of Met catabolism (Figure 1). Suppressing the expression of SAM synthase in Arabidopsis or tobacco plants led up to over 430-fold increased accumulation of Met, compared to control, wild-type plants (Boerjan et al., 1994; Goto et al., 2002; Kim et al., 2002) (Table 2). However, these transgenic plants exhibited severe abnormal morphological phenotypes. Because SAM is a precursor for a number of essential metabolites, such as the hormone ethylene, polyamines, phytosiderophores and biotin, and also because SAM is the primary biological methyl group donor (reviewed by Droux, 2004; Hesse et al., 2004; Ravanel et al., 1998; Roje, 2006), it was suggested that the abnormal phenotypes of these plants are due to the reduction in essential SAM-derived metabolites. Hence, this approach appears unsuitable for increasing soluble Met level in vegetative tissues of crop plants. Nevertheless, whether such an approach may be successful in seeds, using seed-specific promoters, has still to be elucidated. Additional attempts to increase Met content included reduced expression of the Met catabolic enzyme Met-γ-lyase. Knocking out the gene encoding for this enzyme in Arabidopsis significantly increased leaf Met content by approximately 9-fold under sulphate starvation, but did not affect Met level under normal growth conditions (Goyer et al., 2007). Because the Met catabolic enzyme Met-γ-lyase is a relatively inefficient enzyme having a relatively high Km value for Met (approximately 10 mm), using this approach to accumulate Met should be considered only in transgenic plants having higher Met content.

Enhancement of Lys and Met contents by expressing genes encoding proteins rich in these essential amino acids

An additional option to increase the content of Lys and/or Met is to enhance native sink proteins or add additional transgenic sink proteins that are rich in these amino acids. This can boost the incorporation of these two amino acids into the storage proteins (reviewed by Amir, 2008, 2010; Amir et al., 2012; Galili et al., 2005; Tabe and Droux, 2002). This approach is important because the free amino acids pool is small compared to the pool of protein-bound amino acids (Amir and Tabe, 2006). The researchers have tested four different strategies as described below:

  • 1.Expressing recombinant genes encoding synthetic proteins with higher proportions of Met and/or Lys. Although significant efforts were invested in designing synthetic genes encoding proteins rich in Met and/or Lys, this strategy faced major difficulties associated with protein stability, digestibility and toxicity (for review, see Beauregard and Hefford, ; Wenefrida et al., ).
  • 2.Expressing recombinant genes encoding seed storage proteins fortified with additional codons for these two essential amino acids. This strategy emerged with the idea to use genes encoding seed storage proteins that accumulate to high levels in seeds and enrich them with codons for Lys and/or Met. Most of these modified proteins were unstable in the plants (De Clercq et al., ; Hoffman et al., ). Nevertheless, several successes were reported. Forsyth et al., (Forsyth et al., ) and Torrent et al., (Torrent et al., ) reported that the barley chymotrypsin inhibitor 2 as well as maize γ-zein that were modified to have higher number of Lys codons stably accumulated in maize cells. Likewise, expression in rice seeds of a chimeric gene containing combined sequences of a rice glutelin and a Lys-rich protein from winged bean led to the elevation of total Lys level by more than 45%, compared with the control nontransformed genotype (Wenefrida et al., ). Several other modified proteins were also tested to increase the Met content in plants. Mainieri et al. () designed a chimeric gene encoding a Met-rich protein by fusing a section from a Met-rich maize gamma-zein with phaseolin, a major seed storage protein of bean. The new chimeric gene, called zeolin, proved to be stable in transgenic tobacco plants, but its contribution to the total Met content in the plant was not recorded. Although this strategy led to some successes, additional knowledge is required to provide better predictions to the in vivo stability of these chimeric proteins in the transgenic plants.
  • 3.Expressing seed storage proteins that are naturally rich in Met or Lys. Because some plant species possess endogenous seed storage proteins that are enriched with Lys and/or Met, genes encoding these proteins could be used as heterologous sources to fortify these amino acids in seeds or vegetative tissues of other plant species. Attempts were then directed towards exploring the existence of Lys- and Met-rich storage proteins in seeds of various plant species (reviewed by the Altenbach and Simpson, ; Galili, ; Tabe and Higgins, ). While several of these proteins were unstable in the transgenic plants, other proteins have accumulated, leading to the elevation in the levels of Lys and Met. The extent of enhancement of Lys and/or Met was dependent on the type of the proteins and the plant species to which they were introduced. For example, expression of a potato pollen-specific sb401 protein (containing 16.7% Lys) in maize grains increased both grain protein content by up to 39% as well as grain Lys content by up to 54.8% in the different transgenic lines, compared to the control nontransgenic maize (Onouchi et al., ; Yu et al., ) (Table ).

Attempts to increase Met content in seeds of crop plants through the expression of Met-rich seed storage proteins were also made. These concentrated mainly on expressing heterologous genes encoding Met-rich 2S albumin seed storage proteins. In cases where the proteins were stable in seeds of the foreign plants, Met levels were generally increased. For example, utilization of the Brazil nut 2S albumin gene led to an elevation of Met content in total soluble seed proteins of canola (33%) (Altenbach et al., 1992), tobacco (30%) (Altenbach et al., 1989) and Vicia narbonensis (60%) (Saalbach et al., 1995) (Table 3). Similarly, the level of Met increased up to 62% in transgenic rice plants expressing the sesame 2S albumin (Lee et al., 2003) and in lupin plants expressing a sunflower 2S albumin (up to 97%) (Molvig et al., 1997). Nevertheless, the Brazil nut protein and, to a lower extent, also the sunflower 2S albumin were subsequently found to be allergenic in some people, questioning their potential biotechnological use (Bartolome et al., 1997; Kelly and Hefle, 2000). Other groups of Met-rich proteins derived from maize, such as β-zein, γ-zein and/or δ-zein (not the Lys free α-zein), were found to have the ability to accumulate in different plants and tissues such alfalfa, lotus, tobacco, white clover and maize kernels (Bagga et al., 1995, 1997; Bellucci et al., 2005). Elevation in Met content (18%) was reported when β-zein (15 kD zein) was expressed in a seed-specific manner in soybean seeds (Dinkins et al., 2001) (Table 3).

Table 3. Met contents in seeds of various transgenic lines (T), expressing different Met-rich storage proteins under the control of several seed-specific promoters, compare to wild-type seeds (WT)
PlantTransgene (promoter::gene)WT T Changes (T/WT) (average)References
  1. a

    BN-Brazil nut 2S albumin.

  2. b

    SSA-sunflower 2S albumin.

  3. c

    Leg4-leguminB4 promoter.

  4. d

    BX17-wheat high molecular weight of glutenin promoter.

  5. e

    AT2S1-Arabidopsis 2S albumin gene 1.

CanolaPhaseolin::BNa2.64 mol%2.94–3.52 mol%33%Altenbach et al. (1992)
TobaccoPhaseolin::BNa3.6 mol%4.74–3.95 mol%30%Altenbach et al. (1989)
RiceGlutelin::sesame 2S albumin0.21%0.37%43%Lee et al. (2003)
LupinPea vicilin::SSAb0.551.07 (g/16gN)97%Molvig et al. (1997)
Vicia narbonensisLeB4c:: BNa0.5%1.5%66%Saalbach et al. (1995)
Vicia narbonensisLeB4c:: BNa0.5%1.5%66%Pickard et al. (1995)
RiceBx17d::SSAb10 μmol/g DW12.8Not significantHagan et al. (2003)
ChickpeaPea vicilin::SSAb1.7 mol%3.3 mol%94%Chiaiese et al. (2004)
Bean35S::BNa2 mol%2.5 mol%25%Aragão et al. (1999)
SoybeanPhaseolin:: 15 kD zein1.8 mol2.2 mol18%Dinkins et al. (2001)
Arabidopsis, Brassica napus, TobaccoAT2S1e:: modified AT2S10.2–5%De Clercq et al. (1990)
MaizeGlutelin:: amaritin0.53% 100 g seeds0.68% 100 g seeds0.15%Rascn-Cruz et al. (2004)

Nevertheless, several studies have shown that the elevation of Met content in these transgenic plants was at the expense of other endogenous sulphur-rich proteins as well as other sulphur compounds (Amir and Tabe, 2006; Hagan et al., 2003; Molvig et al., 1997; Tabe and Droux, 2002). This suggests that the level of soluble Met limits the synthesis of heterologous proteins (Amir et al., 2012; Tabe and Droux, 2002). Therefore, additional approaches are required to enrich the content of Met in crop plants.

  • 4.Suppressing genes encoding Met- and/or Lys-poor proteins. This more recent strategy was derived from earlier observations showing that (i) mutants such as opaque 2 have lower level of specific storage proteins, leading to changes in the distribution of storage proteins and elevation of Lys-rich storage proteins (Azevedo and Arruda, ), and (ii) the protein profiles of seeds are dependent on the levels of nitrogen and sulphate in the soil. In some cases, higher levels of Met were detected in seeds when the sulphur/nitrogen ratio was increased that led to the elevation of Met-rich proteins (Fujiwara et al., ; Harada et al., ; Higashi et al., ; Ohkama et al., ). Based on these observations, the researchers tried to manipulate the storage proteins to have higher Lys or Met content using RNAi approach. A significant increase in the level of Lys in maize kernels was obtained using an RNAi approach to down-regulate the levels of the Lys-poor, 22 kDa α-zeins (Segal et al., ) or 19 kDa α-zeins (Huang et al., ) (Table ). Simultaneously, down-regulating the levels of both of these two Lys-poor zeins led to an increase in Lys content in the total grain protein from 2.83% to 5.62% (Huang et al., ) or from 2.22% to 4.9% (Frizzi et al., ). Another study in rice seeds showed that over-production of a Lys-rich glutelin storage protein, caused a reduction in the levels of the Lys-poor prolamin storage proteins and a compensatory increase in the level of Lys (Sun and Liu, ). Similarly, a reduction in the 13 kD Lys-poor prolamins in rice grains caused an enhancement of the total grain Lys content by 56%, when compared to wild type (Kawakatsu et al., ). Furthermore, expression of the binding protein (BiP) containing 9.4% Lys, in the endosperm of rice seeds, led to a 1.8-fold elevation of total seed Lys content (Kawakatsu et al., ). Nevertheless, in these maize and rice lines having low content of seed storage proteins, the seed starch content was significantly decreased in a similar manner to its decrease in seeds of the opaque2 maize mutation (Mertz, ). The negative impact of seed Lys fortification on seed starch level is a major drawback, negatively impacting yields and seed quality (Kawakatsu et al., ).

Similar approaches to those described above were also employed to increase the level of Met in various crop plants. Suppressed production of the sulphur-rich RP10 protein in rice seeds caused an enhancement in the levels of the sulphur-rich prolamins, on the expense of the sulphur-poor seed prolamins (Kawakatsu et al., 2010a). Similarly, deficiencies in the Met-poor phaseolin and phytohemagglutinin seed storage proteins in common bean were associated with a respective 10% and 70% increase in seed Met and Cys levels (Yin et al., 2011). Another study in maize showed that reduced expression of the Met-poor 19 and 22 kDa α-zein storage proteins, when combined with enhanced expression of the Met-rich 10 kDa zein gene, caused a several-fold increase in the total seed Met content (Kirihara et al., 2001).

Although some of the strategies seem to be promising, additional studies are required to reveal the effects of these manipulations on the total level of essential amino acids and on other traits, such as seed morphology, seed starch, amino acids, oil contents and germination rates. Schmidt et al. (2011) performed metabolome and transcriptome analyses of soybean seeds with silenced expression of one of the seed storage proteins. Interestingly, this manipulation caused a rebalance of the protein composition, preserving seed protein content without major collateral changes. This suggests that in some cases, this strategy can be performed without major defects or pleiotropic effects. Thus, in summary, some of the above mentioned approaches have proven successful in increasing the content of Met and/or Lys in seeds. Nevertheless, additional knowledge is needed particularly on the in vivo stability of genetically engineered Lys-rich and Met-rich proteins, for optimal fortification of seeds in these essential amino acids.

Combining traits of enhanced free Lys and Met synthesis with expression of proteins rich in these amino acids

Several studies suggested that the levels of soluble Met and Lys represent limiting factors for the synthesis of Met- or Lys-rich proteins. Thus, expressing together genes that boost Lys and Met biosynthesis in combination with genes encoding proteins that are rich in Lys and Met codons appears to be a preferred way to increase the levels of these two essential amino acids. A soybean gene encoding a Lys-rich vegetative storage protein (S-VSP) was co-expressed together with the bacterial feedback-insensitive DHDPS in transgenic tobacco plants, leading to an increase in the total content of Lys (Guenoune et al., 2003). In addition, co-expression of either the barley Lys-rich proteins hordothionine 12 or barley high Lys 8 (BHL8), together with a bacterial feedback-insensitive DHDPS in maize seeds, led to an elevation of the total seed Lys content from 0.2% in the wild type to over 0.7% of the total seed dry weight (Jung and Falco, 2000).

A comparable approach was also employed to increase the Met content in transgenic alfalfa plants. Overexpression of a maize β-zein (Met-rich storage protein) together with mutated form of AtCGS of Met biosynthesis in this plant species enhanced the level of the β-zein compared to plants overexpressing the β-zein alone (Bagga et al., 2005; Golan et al., 2005). Although these results are promising, a study conducted in transgenic potatoes showed that combined constitutive expression of these two genes, although leading to a sixfold increase in Met level, caused a major abnormal phenotype of the plant canopy (Dancs et al., 2008). Morphological problems are less expected in seeds where Met could be efficiently incorporated into seed storage proteins. Demidov et al. (2003) produced a transgenic Vicia narbonensis plants that express under seed-specific promoters the Brazil nut 2S albumin storage protein together with the bacterial feedback-insensitive AK. The double transformant exhibited an additive 2.4-fold enhancement of seed Met content, compared to the wild-type seeds. Thus, the most promising approach to improve Met and/or Lys accumulation in seeds appears through combining (i) overproduction of free Met and/or Lys, (ii) enhancing the synthesis of seed proteins rich in Lys and/or Met, and (iii) suppressing the expression of genes encoding proteins that are poor in Met and Lys.

Finally, once transgenic nutritionally improved genotypes possessing high Lys and Met content in their seeds are produced, it will be important to select for genotypes that exhibit acceptable phenotypes, such as seed morphology and physiology, germination efficiency and lack of accelerated disease susceptibility, particularly when grown under field conditions.

Concluding summary

Plants, as a source of proteins and amino acids, are more economical to produce than farm animals used for the same purpose. However, most plant proteins are nutritionally incomplete for the human diet due to low levels of several essential amino acids, especially Lys and Met (Table 4). An FAO report from 1981 states that plant sources (legumes, cereals, fruits and seeds) are limiting in the content of Lys and sulphur containing amino acids (Met and cysteine), compared to animals sources (Table 5). In addition, the report of the WHO from 1985 declares that cereal proteins have a low content of Lys (1.5–4.5 mol% vs. 5.5 mol% of WHO recommendation), while the proteins in legume seeds and in most vegetables contain only 1–2 mol% sulphur amino acids (Met and cysteine) compared with the 3.5 mol% of the WHO recommendation (WHO, 2007).

Table 4. The requirements of the amino acids, Lys and Met in adults and the contents of Lys and Met in selected foods
Amino acid Requirement mg/kg body weight/day Requirement mg/g proteinb Content in Common foods (mg/kga)
BeefEggsKidney beanWhole wheatBrown riceYellow Corn
  1. a

    The data are taken from (WHO, 2007).

  2. b

    For mean nitrogen requirement of 105 mg/kg per day (0.66 g protein/kg per day).

  3. c

    Numbers in brackets are% of the specific amino acid in their total of food protein (mol%).

  4. d

    % protein by weight.

  5. e

    % calories is for the protein calories in the total of calories.

Lys304519980 (20)c9040 (15)6910 (17)3770 (8)980 (10)1420 (10)
Met10167890 (8)6840 (11)2600 (6)5270 (11)890 (9)970 (7)
Protein %d  25139142.53.3
Calories %e  313328161412
Table 5. The content of Lys and sulphur containing amino acids (Met and cysteine) in most important food groups as protein sources, from a worldwide perspective. The data is in milligrams of limiting amino acid per gram of total protein in the food sourcea
Food sourceLysSulphur amino acids
  1. a

    According to the FAO report (http://www.fao.org/docrep/005/AC854T/AC854T00.htm).

Seeds and nuts4536

The information described in this review shows that multiple genetic approaches, including engineering approaches, have been proven successful in enhancing the content of these essential amino acids in plants (Tables 1-3). However, the determination of the content of Lys and Met varies between the different publications and is presented in different units (Tables 1-3); thus, it is difficult to estimate whether the manipulations succeeded to reach the desired levels as determined by the WHO and FAO. The elevation in Lys and/or Met contents (expressed as fold changes) differs between different plants and even different tissues of a given plant (e.g. Table 1, for expression of feedback-insensitive DHDPS; and Table 2, for expression of AtCGS). In most cases when high Lys or Met content was achieved, an abnormal phenotype was observed that reduced the yield. However, in some cases, high levels of Lys or Met were detected with only slight or even no abnormal morphological phenotype. For example, alfalfa plants overexpressing the D-AtCGS have 3.6 mol% Met as recommended by the WHO (WHO, 2007), without an abnormal phenotype (Avraham et al., 2005). Also, the high-Lys corn (LY038) expressing a bacterial feedback-insensitive DHDPS in an embryo-specific manner has high Lys with a normal phenotype (Lucas et al., 2007).

These results show that additional knowledge and experiments are required to produce crop plants with higher levels of Met and Lys and with a normal phenotype. More knowledge is also required about the regulation of the metabolism of these amino acids, their influence on other processes and metabolites, the regulation of seed storage proteins and the effects of higher Met and Lys on protein accumulation. Although several studies already combined the two strategies to increase the levels of these amino acids, that is, increasing their soluble levels and expressing Lys- or Met-rich storage proteins, knowledge on these combinations and their effects on seed biology is still incomplete. The experience gained by the studies in model plants needs to be implemented towards the development of crop plants with higher levels of these amino acids. The Monsanto company recently used this experience to elevate Lys level in corn, producing high-Lys corn (LY038), which was approved for commercial production. Nevertheless, despite this approval, the commercial exploitation of this high-Lys corn yet suffers from extensive public opposition for ‘GMO’ crops.


We thank Dr. Gideon Baum for his assistance in editing the text. Our studies on amino acid metabolism were supported by grants from the Israel Science Foundation (ISF) and the Binational Agriculture Research and development (BARD). GG is an incumbent of the Charles Bronfman Chair of Plant Science.