olza : quelques exemples de recherche en france Erucic acid rapeseed : 1 . Prospects of improvements

In the current context of boosting production of high erucic acid rapeseed, because of the wide range of its industrial applications, this literature review is designed to provide a general overview of available varieties, current knowledge of plant improvement and paths of developing research to increase competitiveness of varieties with high erucic acid content. A limited market dominated by a few companies, cropping burdens of high erucic acid rapeseed varieties among the majority “00” varieties and the still low erucic acid content in rapeseed, explains the reduced and uncompetitive varietal offers. To improve this situation, new varieties could be developed, thanks to the classical methods of selection and biotechnology.


Introduction
Several plant species in the Brassicaceae family produce erucic acid.The criterion for selecting plant species for erucic acid production are: enough agronomic productivity and a high enough seed erucic acid content.Among the Brassicaceae species the main ones are Crambe (Crambe abyssinica), the Ethiopian mustard (Brassica carinata) and oilseed rape (Brassica napus).Nevertheless the main interest in rapeseed is due to its higher agronomic productivity when compared with others species.Rapeseed oil "00" with low erucic acid and glucosinolates content has an average composition of 8% saturated fatty acids, 61% monounsaturated fatty acids and 9% polyunsaturated fatty acids.High erucic acid rapeseed (HEAR) contains around 50% of C22:1.For its technical use, Correspondence: fine@cetiom.fr it is desirable to increase the 22:1 content and to decrease the eicosenoic acid (20:1), the polyunsaturated fatty acid (PUFA, 18:2 + 18:3), and saturated fatty acid (16:0 + 18:0) content.
Erucic acid (cis-13-docosenoic acid, C22:1) is a long chain polyunsaturated fatty acid with 22 carbon atoms, with a double bound at the cis-13 position of the carbon chain.High erucic acid rapeseed (HEAR) oil is of interest for industrial purposes because erucic acid (22:1) and its derivatives are important renewable raw materials for the oleochemical industry.Its chemical formula is as follows: Before the newer low erucic acid varieties were developed, the proportion of erucic acid was 45 to 50% of fatty acids in rapeseed oil.
Erucic acid is created by elongating oleoyl-CoA (adding two carbon units to the carbon chain) with malonyl-coA.The enzyme acyl-CoA elongase, which is responsible for erucic acid synthesis, is located in the endoplasmic reticulum.The elongation process studied in Lunaria annua consists of the following steps: oleoyl-CoA and malonyl-CoA condensation and creation of 3-ketoacyl-CoA; -3-ketoacyl CoA reduction to 3-hydroxyacyl-CoA; -3-hydroxyacyl-CoA dehydration and 2,3 transénoyl-CoA reduction.
The authors here present the first paper which deals with the prospects of improvements in erucic acid rapeseed.A second paper on variety supply, agronomic management and quality will then be presented in the OCL volume.
Improving HEAR rapeseed varieties is a difficult task which has to integrate precise knowledge of plant and seed metabolism and combine several approaches and technologies.Among the possibilities for plant breeders, two main areas are being explored in order to produce rapeseed with very high erucic acid content: the conventional selection methods, and biotechnology.

Conventional selection methods
A specific feature of Brassica napus is to have two genomes to form an amphidiploid species.Figure 2 shows the relationships between the genomes of the different Brassica species.This specific feature provideses opportunities as well as complexities for breeders.Resynthetized amphidiploids from B. rapa and B. oleracea is a possible strategy.Nevertheless, classical quantitative genetics has to deal with the presence of two closely related copies of a given gene for each trait, one on the A genome and the other on the C genome.

Comprehension of genetic determinism
When an interesting character is identified, it takes years to transfer it into the chosen material.During the last few years, marker assisted selection (RFLP probe, RAPD and microsatellites) has played a significant role in effective plant selection (Friedt and Lühs, 1995;Paterson et al., 1991;Szewc-Mc Fadden et al., 1996;Quiros et al., 1994;Wang et al., 1994;Waugh and Powell, 1992).
The two genes controlling erucic acid synthesis in rapeseed have been mapped on hybrid groups which are different from RFLP markers on doubled haploid (DH) lines derived by crossing "Mansholts Hamburger Raps" X "Samourai" (Friedt and Lühs, 1995;Uzunova et al., 1995).In the same population, three quantitative trait loci (QTL) for seed oil content have been mapped, two of which mapped to the positions of the erucic acid genes (Uzunova et al., 1995).Zhao et al. (2008) identified eight QTL for erucic acid content in the segregating doubled haploid (DH) population derived from a cross between two high erucic acid rapeseed cultivars.One QTL colocated with one of the two erucic acid genes (FAE1) and the remaining seven QTL were not related to the FAE1 gene, but contributed to the 22:1 content.
Other genetic studies have revealed that erucic acid synthesis is controlled by two major gene loci E1 (Bn-FAE1.1)and E2 (Bn-FAE1.2),which have additive effects (Harvey and Downey, 1964;Lühs et al., 1999;Stefansson, 1983).The two genes Bn-FAE1.1 and Bn-FAE1.2 mapped in B. napus representing the parental species B. rapa (A-genome) and B. oleracea (C-genome) Fatty Acid Elongase (FAE1) genes, showed polymorphisms (Fourmann et al,. 1998;Nath, 2008).Jönsson (1977), and Pourdad and Sachan (2003) reported that in rapeseed (B.napus), 22:1 content is controlled by alleles at one, one or two and two loci leading to 5-10%, 10-35% and more than 35% of 22:1, respectively.The two elongation steps from oleoyl-CoA to 22:1 are each controlled by alleles of two loci (Harvey and Downey, 1964;Stefansson, 1983).Ecke et al. (1995) and Jourdren et al. (1996) mapped the two loci determining erucic acid content in rapeseed populations using both random fragment length polymorphism (RFLP) and random amplified polymorphic DNA (RAPD) markers.Assignment of the two loci to independent linkage groups was confirmed by a QTL approach by Thormann et al. (1996).However, the two loci do not contribute equally to erucic acid content.Each locus can have multiple alleles and leads to erucic acid contents ranging from 0.1% to about 60%.At least five alleles govern the erucic acid proportions in Brassica, including; e, Ea, Eb, Ec and Ed.Therefore, levels of erucic acid can be fixed at a large number of values ranging from < 1% to > 60% (Jönsson, 1977).
In winter rapeseed varieties, a single allele contributes to the production of approximately 16 to 17% C22:1 (Lühs and Friedt, 1994, 1995, 1997).Erucic acid biosynthesis involves elongation of oleic acid to eicosenoic acid and then to erucic acid.The elongation is the result of two cycles of a four-step process, in which 18:1-CoA and 20:1-CoA are used as substrates.In rapeseed, erucic acid biosynthesisis is controlled by the expression and the specificity of ß-ketoacyl-CoA synthase (KCS) (the enzyme responsible for the fatty acid elongation of oleic acid (18:1) to eicosenoic acid (20:1) and then to erucic acid), which is a gene similar to the FAE1 (fatty acid elongase) gene present in Arabidopsis thaliana (James et al., 1995;Lühs and Friedt, 1997).It is believed that the initial reaction of the four step elongation process is the rate-limiting step in the biosynthesis of erucic acid (Cassagne et al., 1994).Crossing of conventional HEAR to rapeseed with reduced contents of linoleic acid (18:2) and linolenic acid (18:3) did result in recombinant high erucic low polyunsaturated fatty acid (HELP) F3-plants which, however, did not show an increased 22:1 content compared to the parental HEAR genotype.This indicated that the β-ketoacyl-CoA synthase (KCS; fae1 gene) activity may be limiting.The fae1 gene encoding the KCS enzyme, has been cloned from a range of plant species and has been overexpressed under the control of a seed specific promoter in HEAR.However, only very minor increases in 22:1 content were reported (Han et al., 2001;Katavic et al., 2001).Even in combination with the expression of the Ld-LPAAT gene from Limnanthes douglasii, no substantial increase in the 22:1 content has been found (Han et al., 2001).
Other bottlenecks in the pathway could be the availability of oleic acid for elongation to eicosenoic and erucic acids.In order to determine whether the availability of oleic acid can be a limiting factor, Sasongko and Möllers (2005) crossed HEAR (cv.Maplus) with high oleic acid rapeseed (HOAR) (Schierholt et al., 2001) to recombine the genes for high 22:1 with those for high 18:1 (i.e.low content of polyunsaturated fatty acids).However, the recombinant line HELP did not show a significant change in 22:1 acid content, indicating that in this material the ß-ketoacyl-CoA synthase (KCS) activity may be limiting.Nath (2008) showed a large variation in erucic acid content ranging from 44 to 72% with a mean of 58.8% in the F 2 plants (F 3 seeds) and from 50% to 72% with the mean of 64.8% in the F 3 plants (F 4 seeds) obtained by crossing transgenic HEAR winter rapeseed line (361.2B)and non-transgenic high erucic and low polyunsaturated winter rapeseed line 6575-1 HELP to produce F 1 plants which were selfed to produce F 1 , F 2 and F 3 plants.The mean for 22:1 exhibited in the F3-population (F4seeds) was 6% more compared to the mean of F2-population, indicating a response to selection.This compares favourably with the 63.2% of the transgenic parent 361.2B and 49.6% of the non-transgenic 6575-1 HELP (High Erucic and Low Polyunsaturated fatty acid) parents.The best F3-lines had a PUFA content of only 5 to 6%, which is about 10% lower than the PUFA content of the parent 361.2B.Results from regression analysis of the F2-population indicated that reduction in PUFA content by 10% led to a 6.3% increase in erucic acid content.The 72% erucic acid content achieved in the study marks a major breakthrough in breeding high erucic acid rapeseed.The trierucin content varied between 0 and 24.5% with a mean of 12.6%, respectively.In the segregating F 2 population, there were two groups; one group which lacked trierucin because it lacked Limnanthes douglasii lysophosphatidic acid acyltransferase (Ld-LPAAT) gene in the individuals of the F2 population.The group with trierucin showed a continuous variation as one would expect for a polygenic trait.Erucic acid content varied from 45% to 57% for the zero trierucin plants and in the trierucin group it varied from 48% to 72% in the segregating F2 population.Path coefficient analysis to partition the correlations of the different factors into direct and indirect effects showed little direct effect of trierucin content on erucic acid (0.06) content.Strong negative direct effects of oleic acid and PUFA content on erucic acid content (-0.58 and -0.73) were observed.Trierucin content also showed larger pos-itive indirect effect via 18:1, PUFA and 20:1 content on erucic acid than direct effect.Eicosenoic acid showed indirect negative effect via oleic acid as much as direct effect on erucic acid content (Nath, 2008).This information and material will be valuable for future approaches to increase erucic acid content in rapeseed beyond the levels currently obtained.Nath (2008) also showed that the erucic acid content varied from 35-59% in the segregating DH-lines, 49-52% in 6575-1 HELP parental lines and, 46-48% in transgenic resynthesised high 22:1 rapeseed line TNKAT.TNKAT, a resynthesised transgenic winter rapeseed line (RS306) was derived from an interspecific cross between Yellow Sarson (Brassica rapa) for A-genome (FAE1.1)and cauliflower (Brassica oleracea sp.capitata) cv.Super Regama for C-genome (fae1.2) as outlined by Lühs and Friedt (1994).RS306 was used for Agrobacterium mediated transformation to produce the TNKAT line carrying a single transgene copy (chimeric Bn-FAE1.1 with Ld-LPAAT where both were under the control of the seed specific napin promoter (Han et al., 2001).
Previous work on a B. juncea RFLP map (Cheung et al., 1998) identified two QTL on linkage groups 7 and 4 which explained approximately 85% of the phenotypic variance of C22:1.A candidate-gene approach has helped locate two genes encoding the elongase enzyme within the confidence interval of the two QTLs.These results are like those obtained by a similar approach on another amphidiploid obtained by integrating the genomes of B. rapa and B. napus (Ecke et al., 1995;Jourden et al., 1996) using common markers, thus suggesting that the elongase genes are present in the genome of B. rapa.

Use of doubled haploid lines
The use of classical methods for rapeseed breeding (genealogical selection) to change the fatty acid composition would take at least 10-12 years.A faster method would include the application of microspore culture techniques to produce DH populations (Friedt and Lühs, 1995;Kontowski and Friedt, 1994;Lühs andFriedt, 1994, 1995;Nath et al., 2007;Thierfelder et al., 1993).Since the development of the isolated microspore culture technique for the production of homozygous DH lines in oilseed rape (Brassica napus L.) by Lichter (1982), this method has gained considerable importance in rapeseed breeding programmes.The method has been optimised variously (Chen et al., 1994;Iqbal et al., 1994;Möllers et al., 1994;Zaki and Dickinson, 1991).
Crosses were made between cell lines derived from cultivars of winter rapeseed relatively rich in erucic acid.Parental and F1 progeny lines were used as donors for microspore culture and as a source of microspore derived embryos (MDEs) which are cultured to produce haploid plants.The haploid plants were vernalized and treated with colchicine to double the chromosomes to produce DH lines.This method provided varieties with high levels of erucic acid (up to 60%) (Friedt and Lühs, 1995;Jourdren et al., 1996).Furthermore, Nath et al. (2007) developed a simple protocol for simultaneously extracting lipids (for oil quality analysis) and DNA (for marker-assisted selection) from single cotyledons dissected from MDEs of Brassica napus, suitable for fatty acid analysis, PCR amplification and regeneration of the rest of the embryo, allowing selection of valuable MDE genotypes at an early stage in segregating MDE populations; derived by crossing the homozygous transgenic resynthesised high erucic acid rapeseed line RS306, carrying a single T-DNA with two chimeric genes (Han et al., 2001), with the high erucic acid winter rapeseed line 6575-1 (Sasongko and Möllers, 2005) segregating for two linked transgenes, fae1 and plsC, affecting the fatty acid composition.The early identification of 50% MDE individual plants carrying the desired transgenes, along with high expression of the trait, allows for early selection for plantlet regeneration, and reduces the probability of the inclusion of undesirable genotypes and exclusion of desirable recombinant genotypes among the total number of regenerated MDEs.Hence, marker-assisted selection (MAS) at the in vitro stage would screen a larger population of MDEs and reduce greenhouse costs.Unequivocal selection for absence of, intermediate and high erucic acid (Albrecht et al., 1995) and oleic acid (Möllers et al., 2000) was possible in segregating MDE populations by determining the fatty acid composition by gas liquid chromatography (GLC) from single cotyledons dissected from MDEs.However, the application of this early in vitro selection system is limited to those traits that can be rapidly and cost effectively analysed and for which a close correlation between the MDEs and the seeds from the regenerated plants has been shown.Furthermore, phenotypic results may be confounded by MDE genotype x environment interactions, i.e., in vitro culture conditions may differently affect the fatty acid composition of storage lipids in individual MDE genotypes.Such shortcomings are overcome if MAS is applied concomitantly.The erucic acid content ranged from 14.1 to 42.7% in the MDE genotypes where both plsC (from Limnanthes douglasii, accession no: X83266) and fae1 (accession no : AF274759) genes were expressed and 14.7 to 30% in the MDE genotypes where only plsC was expressed.However, results from fatty acid analysis did not reveal a pronounced effect of the transgene on the erucic acid content of the MDE and were not correlated with the erucic acid contents of the seeds obtained from the corresponding DH plants in the green house.The mean erucic content of 54 DH lines with the transgene was 2.3% lower than the 36 DH lines lacking the transgene.A strong negative correlation was found between erucic acid and polyunsaturated fatty acid content (18:2 + 18:3; rs = −0.40**).This result indicates that the ectopic FAE1.1 gene may not be functional (Nath, 2008).Futhermore, it was observed that dividing the DH population into half according to their PUFA (polyunsaturated fatty acid) content revealed that a 8.1% reduction in PUFA content (mean = 11.4%) in the DH lines resulted in an increase in erucic acid content by 3.7% when compared to the DH lines with high PUFA content (mean = 19.5%).The best DH line (IV-10-F-6) had 59.1% erucic acid in the seed oil.This was 9% more than the higher erucic acid parent 6575-1 HELP (Nath, 2008).Although the best three DH lines selected for high 22:1 were transgenic, it seems that transgene (fae1) cannot increase the 22:1 content but interacting with other factors like low PUFA genes play a vital role in increasing 22:1 proportions by reducing PUFA content.
The fatty acid composition of the MDE and the mature seeds being substantially identical, the immature embryos were used to study the composition of the storage compounds (lipids, proteins) and more specifically, the biosynthesis of triacylglycerols (TAG) containing high erucic acid proportions.Because of their high enzymatic activity, homogenates of immature embryos are able to make TAG.It has been demonstrated that they can synthetize trierucin in the presence of 1,2-dierucylglycerol and erucyl-CoA (Taylor et al., 1990(Taylor et al., , 1991(Taylor et al., , 1992;;Friedt and Lühs, 1995).The technique of the immature embryos can be used as a method of early sorting, coupled with the one of DHs.Among the early sorting techniques, it is also possible to make measurements on the half-cotyledons.
The range of the erucic acid content recorded by Nath (2008) in 90 DH lines derived from the cross between TNKAT × 6575-1 HELP was 34.6 to 59.1% (mean = 47.1) and the heritability was 0.88 whereas the oil content range was 28.3 to 50.7% (mean = 40.5)and the heritability of oil content was 0.82, respectively.The oil and erucic acid content of the parental line TNKAT were 40.8% and 46.1% and that of the parental line 6575-1 were 46.4% and 50.4%, respectively.

New synthesis of Brassica napus from B. rapa and B. oleracea
Brassica napus is a natural amphidiploid species, resulting from the spontaneous hybridization of B. rapa (A-genome) and B. oleracea (C-genome).But it turns out that the genetic variability of B. napus is lesser than that of B. rapa and B. oleracea, which are highly polymorphic.This wide variability of the parents can be exploited by manipulation and re-synthesis of B. napus from the original parents.The new variety of B. napus obtained is genetically intermediate between the two parents, but very different from the natural forms of B. napus (Becker et al., 1995;Chen and Heneen, 1989;Engqvist and Becker, 1994;Friedt and Lühs, 1995;Kräling, 1987;Lydiate et al., 1993;Olsson, 1986;Song et al., 1993;Thierfelder et al., 1993).
Currently, erucic acid content in B. napus is limited to 55%, the average content ranging mostly between 45 and 50% (Friedt and Lühs, 1995;Malher and Auld, 1988;Lühs andFriedt, 1994, 1995).However, the variation range of erucic acid in both parents is greater, ranging from 30.1 to 61.4% in B. rapa and from 28.2 to 63.4% in B. oleracea.Biotechnology helped in creating a new amphidiploid from parents rich in erucic acid (about 55-60%).New somatic hybrid rapeseed was created by the fusion of protoplasts of Brassica oleracea var botrytis and Brassica rapa var oleifera, which were selected for the high erucic acid content in their seed oil.One of the regenerated plants contained 57.4% erucic acid.The fatty acid composition of plants in the R1 generation was stable and was coupled with increasing female fertility (Heath and Earle, 1995).
Genetic studies on the newly synthesized material showed that the genes responsible for erucic acid synthesis had an allelic contribution ranging from 16 to 17% relative to the total content of erucic acid.However, using this technique, limits the synthesis of trierucin (Lühs andFriedt, 1994, 1995).
D303, page 4 of 10 A. Sanyal et al.: OCL 2015, 22(3) D303 Some cauliflowers genotypes were able to esterify the erucic acid at the sn-2-position of the triacylglycerols (Taylor et al., 1994;Friedt et Lühs, 1995).However, trierucin was not found in the seed oil in a mutant form of B. oleracea.According to Downey and Taylor (1996), previous work by the same team (Taylor et al., 1995) produced cauliflowers with 60% erucic acid in their seeds, including 20-25% in the sn-2-position.B. oleracea × B. rapa and B. napus × B. oleracea crosses are expected to result in high erucic acid varieties with low glucosinolate content and about 75% C22:1, respectively.
Another promising method for introducing the desired traits is the fusion of protoplasts of plants of the crucifer family with wild and cultivated varieties of B. napus or B. juncea.Thus, fertile hybrids between B. napus and Arabidopsis thaliana (Forsberg et al., 1994;Friedt and Lühs, 1995) or between B. napus and Thlaspi perfoliatum (Fahleson et al., 1994;Friedt and Lühs, 1995) were obtained.However, this method is limited by several factors: the problem of sustainability of plants, fewer fertile plants, new obtained material far from being cultivable (and therefore requiring series of backcrosses, i.e. crossing back the interspecific hybrid with the cultivated species, in order to regain its traits, along with selection to retain the advantageous traits).

Trierucin synthesis (trierucylglycerol)
Although the main component of triacylglycerol in rapeseed is erucic acid, the long fatty acid chain can be attached to the sn-1 and sn-3-positions of glycerol, but is excluded from the sn-2-position.Therefore, we do not find any trierucin in rapeseed oil and the theoretical limit of the erucic acid content in rapeseed is 66%.Several studies suggested that two biochemical steps are critical for improvement of erucic acid production in rapeseed: membrane bound fatty-acid elongation and lysophosphatidic acid acyltransferase (LPAAT) activity leading to the biosynthesis of trierucin (Bernerth and Frentzen, 1990;Creach et al., 1993;Friedt and Lühs, 1995;Taylor et al., 1993).
Thus, the two preferential targets for genetic research are: acyl-CoA elongase and lysophosphatidic acyltransferase.

The lysophospatidic acid acyltransferase (LPAAT)
In recent years, efforts related to cloning acyltransferases genes focused on erucyl-CoA preferring acyltransferase (LPAAT) were made.The first gene of LPAAT, plsC from E. coli, was cloned from a complementary homologous DNA, of a mutant strain of E. coli, which was deficient in the activity of LPA-AT (Coleman, 1990(Coleman, , 1992;;Frentzen, 1998).The gene of an erucoyl-CoA preferring sn-2 acyltransferase from Limnanthes douglasii (Ld-LPAAT) has been successfully cloned and overexpressed in rapeseed (Brough et al., 1996;Brown et al., 1995;Friedt and Lühs, 1998;Hanke et al., 1995;Lassner et al., 1995;Frentzen, 1998).However, the overall proportions of 22:1 in the seed oil did not increase.Zou et al. (1997) have confirmed that the yeast (Saccharomyces cerevisiae) SLC1-1 gene encodes sn-2-acyltransferase capable of acylating sn-1-oleoyl-lysophosphatidic acid using a range of acyl-CoA thioesters, including 22:1-CoA.However, neither the meadowfoam nor the yeast-LPAAT transgene approach was successful in achieving high trierucin content in HEAR B. napus seed oil.The failure to significantly increase the 22:1 level by engineering LPAAT could be due to a limitation in the acyl-CoA pool in the cytosol, which is required to support high levels of trierucin synthesis (Lühs et al., 1999;Sasongko and Möllers, 2005).This hypothesis was supported by an increase in the levels of 22:1 (48 to 53%) in transgenic "Hero" plants which expressed the yeast fae1 when compared with 43% of 22:1 in the wild-type control lines (Katavic et al., 2000).Similar results were obtained with the expression of Arabidopsis and B. napus fae1 in rapeseed (Han et al., 2001;Katavic et al., 2001;Wilmer et al., 2003).Thus, these studies show that the proportions of 22:1 in rapeseed oil is limited by both 22:1 synthesis and its subsequent incorporation into TAG (Katavic et al., 2000).HEAR oil could eventually be produced by combining these and other genetic modifications.Weier et al. (1997) suggested that the level of trierucin depends not only on the activity of the introduced sn-2-acyltransferase but also on other biosynthesis or incorporation steps.It is possible that the levels of erucoyl-CoA in the seed acyl-CoA pool may be too low to allow high levels of trierucin biosynthesis.If this is the case, then overexpression of genes regulating very long chain fatty acids (VLCFAs) biosynthesis may be required to boost very long-chain acyl-CoA availability for incorporation into seed triacylglycerols (TAGs).Analysis of the cDNA encoding the microsomal LPAAT revealed two distinct classes of genes, A and B (Frentzen, 1998;James et al., 1995).The class A microsomal LPAATs defined by Frentzen (1998) possess substrate preferences for C18:1-CoA typical of enzymes involved in membrane lipid synthesis and are ubiquitously expressed in the plant.In contrast, individual members of the class B LPAATs display preferences for distinct, unusual saturated or unsaturated acyl groups and are normally expressed in storage organs.Although class B LPAATs have been exploited to alter the stereochemical composition of rapeseed (Brassica napus) oil to permit the incorporation of modified fatty acids at sn-2 (Knutzon et al., 1999;Lassner et al., 1995), a significant increase in the total amount of unusual fatty acid was not accomplished by the expression of the class B LPAATs alone.In contrast, the transformation of rapeseed and Arabidopsis thaliana with a yeast gene encoding a variant LPAAT, SLC1-1, capable of accepting very long chain fatty acyl (VLCFA)-CoA substrates resulted in an increase in the total VLCFAs and, unexpectedly, in total oil content (Zou et al., 1997).
In a study by Nath (2008), seed oil from the non-transgenic parent (6575-1 HELP) predominantly contained oleic acid at the sn-2-position (73.3%), while very long chain fatty acid (22:1) was detectable in trace amounts only.On the other hand, the oil from the transgenic parents (TNKAT and 361.2B) and the selected best DH and F3 lines (F4-seeds) contained higher amounts of 22:1 and correspondingly lower proportions of 18:1, respectively.The highest amount of 22:1 at the sn-2-position was found in the F3 line (III-G-7), followed by the DH line (IV-10-F-6) with values of 65.3% and 40.3%, respectively.These sn-2 compositions of the transgenic seed oils correlated with the 22:1-CoA specificity of the expression of the Ld-LPAAT gene from L. douglasii.This result is in agreement with the observation of Weier et al. (1997), andHan et al. (2001).Hence, lipid analyses revealed that the introduced Ld-LPAAT gene effectively competes with the endogenous rapeseed enzyme and preferentially incorporates 22:1 into the sn-2-position of the glycerol backbone.However, considerable amounts of oleic acid were also detected at the sn-2position, indicating that endogenous Bn-LPAAT activity may be limiting for achieving higher erucic acid content at the sn-2 of Ld-LPAAT overexpressing rapeseed lines.
Molecular biologists are trying to regulate the biosynthesis of trierucin by replacing rapeseed LPAAT with corresponding enzymes known for its affinity for erucyl-CoA as substrate from appropriate donors like Limnanthes douglasii and Limnanthes alba (Friedt and Lühs, 1995;Taylor et al. 1993;Wolter et al., 1991Wolter et al., , 1995)).Limnanthes seed oil consists of more than 90% of VLCFAs (Lassner et al., 1995;Miller et al., 1964).The sn-2-position of Limnanthes douglasii, has a higher content of C22:1 relative to the total oil composition (Lassner et al., 1995;Phillips et al., 1971).Thus, cDNA encoding 1-acylglycerol-3-phosphate acyltransferase in L. douglasii was introduced into rape.Trierucin was absent in the control plants and was found at a level of 0.4% and 2.8% in two transgenic plants.The fatty acid analysis did not find erucic acid in the sn-2-position in the control plants, but the two transgenic plants had 9% and 28.3% of erucic acid in the sn-2-position.These results showed that the gene encoding 1-acylglycerol-3phosphate acyltransferase in L. douglasii may be functional in rapeseed and allow the incorporation of erucic acid at the sn-2-position of the triacylglycerols in rapeseed (Brough et al., 1996).
The same strategy has been used by Calgene (Lassner et al., 1995).The cDNA encoding LPAAT had been isolated from seeds of Limnanthes alba and was expressed in the seeds of transgenic rapeseed.Erucic acid was seen in the sn-2position of the glycerols in the transgenic plants but was not seen in the control plants.The LPAAT gene obtained from L. alba incorporated erucic acid in the sn-2-position of the rapeseed triglyceride.The highest level of erucic acid in the sn-2position observed in the oil of the transgenic plants is 22.3%.However, these manipulations do not significantly increase the total concentration of erucic acid content which was 40% in the control plants, and a little over 50% in transgenic plants.Recently, the genes encoding the elongase enzyme have been isolated and used to increase the erucic acid content of rapeseed oil (James et al., 1995;Lassner et al., 1996).The expression of these genes, in addition to the use of LPAAT from Limnanthes, should allow a sharp increase in levels of erucic acid in the rapeseed cultivars.
The LPAAT gene in E. coli has a broad specificity for acyl-CoA and may use VLCFAs.The oil in the transformed plants (Erox) with the gene from L. douglasii contains about 40% of C22:1, but less than 1% of C20:1 in the sn-2-position, whereas the transformed plants (Erox) with the E. coli gene have ten times less C22:1 and ten times more C20:1 in the sn-2-position (Weier et al., 1998).These results suggest a kind of specificity or preference of E. coli acyltransferase for C20:1 with respect to C 22:1.The increase in the total amount of C22:1 was not much: 42% for the control and 33-50% for the different Erox transformants, the amounts of C20:1 + C22:1 rang-ing from 46.2 to 60.4% against 54.8 in the unmodified Erox control plant.The authors hope to increase the proportions of C22:1 at the sn-2-position at the expense of C20:1 by changing the elongase activities.
However, as it has already been pointed out that the high levels of erucic acid in the sn-2-position in the oil of transgenic plants containing the gene LPAAT from Limnanthes does not increase the total concentration of C22:1 (Brough et al., 1996;Frentzen, 1998;Lassner et al., 1995;Weier et al., 1997).It is therefore likely that in the transformed erucic acid rich rapeseed, the total amount of C22:1 is limited predominantly by the enzyme activities for the production of C22:1-CoA, especially by the activity of β-ketoacyl-CoA synthase (Frentzen, 1998;Lassner et al., 1996;Millar and Kunst, 1997).Recently, this hypothesis has been confirmed by the development of transgenic erucic acid rich rapeseed, expressing both the LPA-AT from L. alba and β-ketoacyl-CoA synthase (Frentzen, 1998).The expression of the additional β-ketoacyl-CoA synthase has caused an increase in the total concentration of erucic acid by about 45-60% and an increase in trierucin content in the oil of transgenic plants (Tab.7).
Finally, new perspectives have been provided by Zou's team (Frentzen, 1998;Zou et al., 1997).They developed transgenic rapeseed with high erucic acid content expressing a variant LPAAT yeast (Saccharomyces cerevisiae), encoded by the gene SLC1-1 (Frentzen, 1998;Nagiec et al., 1993), which can use a wide range of acyl-CoA thioesters with long and very long chains and also seems to have low specificity for acyl acceptors (Frentzen, 1998;Zou et al., 1997).However, it appears that the expression of the LPAAT from yeast does not only affect the stereochemical composition of the oil of the transgenic plants, but also the proportions of C22:1 and oil.Thus, the oil content increases from 34% of the dry weight in the seeds of the control plants to 41% in some transgenic plants, and the concentration of erucic acid ranges from 45% to 56%, respectively.Furthermore, the erucic acid concentration in the sn-2-position of the transgenic plants is low and does not exceed 4%, respectively.Therefore, the increase in the content of C22:1 cannot be attributed only to the sn-2position, but to the concomitant increase in the concentrations of erucic acid at the sn-1-position and sn-3-positions.These findings, as well as the increase in the oil content, highlight the effect of the yeast LPAAT which allows an increase in the flow of fatty acids in the triglycerides biosynthetic pathway.The use of acyl-CoA thioesters by all acyltransferases involved in the biosynthesis of triglycerides seems to stimulate the biosynthesis of VLCFAs probably by removing feedback inhibition (Frentzen, 1998;Zou et al., 1997).The results obtained with the yeast LPAAT should be attributed to the specific properties of this enzyme.Thus, the low specificity of acyl acceptors of yeast LPAAT might be the decisive property that would differentiate it from the LPAAT genes obtained from other species.This hypothesis needs to be confirmed by further research (Tab.1).
Recently, Nath et al. (2009) achieved around 72% erucic acid content and also reported that further increases in erucic acid content can be expected from progress in reducing the contents of the remaining fatty acids, mainly oleic acid, polyunsaturated fatty acids and eicosenoic acid.
Table 1.Modified erucic acid lines expressing a foreign acyl-transferase gene allowing to bind an erucic acid in sn-2-position of triglycerol (Frentzen, 1998)

The elongase
If binding of erucic acid at the sn-2-position is required to exceed the theoretical threshold of 66% C22:1, it has been observed that the increase in erucic acid content also requires the optimization of the transformation of fatty acids into C16 to C18 or longer chain C20 or C22.The regulation stage of the enzymatic activity has been studied more than the biochemical or molecular aspects.This is partly due to the complexity of the enzyme system and its regulations.The structure and functioning of the acyl-CoA elongase complex is poorly understood because of the difficulty in purifying functional membrane proteins to homogeneity.The acyl-CoA elongase complex has been partially purified from developing rapeseed embryos and has resulted in the enrichment of four proteins between 54 and 67 kDa in size (Creach and Lessire, 1993).The β-ketoacyl-CoA synthase (KCS) was purified from jojoba embryos by Lassner et al. (1996).The corresponding cDNA, homologous to the Arabidopsis Fatty Acid Elongation1 (FAE1) gene (James et al., 1995), was used to transform rapeseed plants.Subsequent KCS activity in developing embryos of low erucic acid rapeseed (LEAR) plants resulted in an enrichment of VLCFAs (up to 33.5% by weight) in the seed oil, thereby demonstrating that KCS activity had been restored.
Additional transgenic approaches to increase erucic acid (i) ATP-citrate lyase (ACL): The acetyl-CoA pool required for new fatty acid biosynthesis is primarily generated by the plastidic isoform of the pyruvate dehydrogenase complex (Ke et al., 2000).Fatty acid elongation is a cytosolic process, and cytosolic ATP-citrate lyase (ACL) generates the required acetyl-CoA precursor.The temporal distribution of ATP-citrate lyase (ACL) activity in developing seeds of rapeseed closely paralleled both that of acetyl-CoA carboxylase (ACCase) in the cytosol and the overall rate of lipid biosynthesis (Fatland et al., 2002).In the cytosol, acetyl-CoA can be carboxylated by ACCase to form malonyl-CoA and hence is converted to long chain fatty acids.Therefore, increased activity of ATP-citrate lyase (ACL) gene will help to produce more acetyl-CoA in the cytosol from mitochondria, which might produce the necessary malonyl-CoA for long chain fatty acid biosynthesis.
(ii) Cytosolic acetyl-CoA carboxylase (ACCase): In the cytosol, plant fatty acids are synthesized by the action of acetyl-CoA carboxylase (ACCase) which converts acetyl-CoA to malonyl-CoA.Cytosolic acetyl-CoA is metabolized via one of three mechanisms: carboxylation, condensation, or acetylation.In the cytosol, acetyl-CoA can be carboxylated by acetyl-CoA carboxylase to form malonyl-CoA (Fatland et al., 2002).Cytosolic malonyl-CoA is required for the biosynthesis of long chain fatty acids like erucic acid (22:1).Therefore, overexpression of cytoplasmic acetyl-CoA carboxylase (ACCase) and ATP-citrate lyase in transgenic high erucic and low polyunsaturated fatty acid (HELP) rapeseed lines could help to further increase erucic acid content in the seed oil of rapeseed.
(iii) Brassica napus lysophosphatidic acid acyltransferase (Bn-LPAAT) antisense: During glycerolipid synthesis, three distinct acyltransferases are responsible for the sequential transfer of acyl groups from acyl thioesters to the glycerol backbone forming TAG (Ohlrogge and Browse, 1995).Among them lysophosphatidic acid acyltransferase (LPAAT) catalyzes the second acylation reaction so that the central position (sn-2) in the biosynthesis of the various glycerolipids, is formed.Therefore, LPAAT substrate specificities are decisive for establishing the fatty acid pattern of TAG.In conventional rapeseed the microsomal LPAAT has a pronounced specificity for 18:1 over other fatty acids.Therefore, the microsomal pathway results in the formation of glycerolipids in which the sn-2-position is specifically esterified with oleic acid (Frentzen, 1998).The enzyme activity of the endogenous rapeseed LPAAT (Bn-LPAAT) competes with the activity of the erucoyl-CoA specific Ld-LPAAT from Limnanthes douglasii.Down-regulation of endogenous Bn-LPAAT gene by antisense technique or mutation could help to increase 22:1 content at sn-2-position as well as in the seed oil (Nath, 2008).
Work to purify and characterize the enzyme complexes, acting in response to the carbon chain elongation, are being carried out under a French project supported by CETIOM and ONIDOL and funded by MENESR.Rustica and Serasem are also participating in these projects.The different research teams are INRA Versailles and INRA Le Rheu, University of Bordeaux 2, Laboratory of Plant Physiology and Molecular Biology of University of Perpignan, University Pierre and Marie Curie of Paris.The Bordeaux team characterized the acyl-CoA elongase as a multi-membranes enzyme complex including at least two elongases, one responsible for the formation of C20:1 and the other C22:1 (Creach et al., 1993).The Laboratory of Plant Physiology and Molecular Biology of Perpignan (Roscoe and Delseny, 1997) works on the isolation of genes encoding the elongase which are necessary to assemble of triacylglycerols enzymes.Studies have characterized the structure and expression of the gene FAE1 (elongase) rape.In collaboration with INRA Rennes, they showed that the two cDNA homologous to FAE1 from immature embryos were linked to the E1 locus associated with the variation in erucic acid content in rapeseed (Jourden et al., 1996).A series of truncated FAE1 sequences were used to express portions of the FAE1 protein in E. coli for functional studies and to isolate cDNA encoding the sub-units of the elongase complex.
Ongoing studies on elongase showed more complex issues than the ones initially planned, because of its different forms and features as a membrane enzyme.The results of this work is in progress and still confidential.

Coming varieties and conclusion
Progress in improving agronomic traits and erucic acid content in oil will provide the rapeseed varieties rich in erucic acid: -Higher yields than the erucic rapeseed varieties currently available: close to those of conventional rapeseed yields, against the current yields which are lower by 15%, respectively.-Lower glucosinolate content, with a generalization of varieties +0; -Richer erucic acid content (50-55%), pending very high erucic acid content varieties (80% minimum), obtained by genetic modification that might appear on the market in the medium to long term (Anonymous, 1998).
However, the successful production of erucic acid from rapeseed will depend on its competitiveness, when compared with products derived from other fatty acids (oleic and stearic) and also in the interest in rapeseed, when compared to other crucifers such as mustard, turnip, and crambe.Through the pursuit of agricultural work to improve yield, Crambe abyssinica seems to be the most competitive alternative plant (Merrien, 1997).