Abstract

The low linolenic acid Brassica napus accession 'Stellar' (3% C18:3) was previously obtained through chemical mutagenesis of a conventional type high C18:3 (around 10%) rapeseed. QTL analysis of the cross 'Stellar' (low C18:3) x 'Drakkar' (high C18:3) revealed that the 'C18:3 content' trait was governed by two loci, L1 and L2, with additive effects (Jourdren et al. 1996a). A desaturase fad3 gene localized in B. napus by interspecific amplification with Arabidopsis thaliana primers was found to be linked to the L1 locus (Jourdren et al. 1996b). In this work, the second B. napus fad3 gene was found to be linked to the second locus L2. We have detected one mutation for each Bn-fad3 gene. These mutations were specific to 'Stellar' in comparison with wild type accessions. A specific Restriction-site Generating PCR (RG-PCR) test was focused for each mutation. With this test, a panel of thirty B. napus accessions was screened. The presence of the mutations was only detected in 'Stellar' and 'Rak-1' accessions, originated from the same mutagenesis event. The validation of the two Bn-fad3 candidate genes for the L1 and L2 QTL are discussed. The application of the codominant PCR test in breeding programs will allow significant improvement of the efficiency in breeding low C18:3 rapeseed varieties.

Introduction

Conventional breeding, chemical mutagenesis and molecular techniques have been used to modify oil quality in crops like sunflower, soya, rapeseed (Miguel and Browse, 1995). Relatively high levels of α-linolenic acid (around 10%) are present in rapeseed oil, conferring flavor instability (Röbbelen and Nitsch, 1975). The ideal rapeseed oil for human diet uses would contain less than 3% of C18:3 (Scarth et al. 1988). The goal was then to reduce the C18:3. For this, at first, the Canadian variety ‘Oro’ was subjected to chemical mutagen experiment and a mutant line, 'M11', with a low C18:3 trait (3%), was obtained (Röbbelen and Nitsch, 1975). Backcrossing the 'M11' line by the Canadian variety ‘Regent’ had resulted in the development of the low C18:3-high (C18:1 + C18:2) variety ‘Stellar’ (Scarth et al. 1988).

In this ‘Stellar’ variety, low C18:3 trait has been shown to be controlled by two genes with additive effects named L1 and L2 (Jourdren et al. 1996a). The continuous phenotypic distribution of seed C18:3 level in a segregating progeny does not allow to clearly identify the genotypic classes (Jourdren et al. 1996a). Because of this continuous distribution and of the low heritability in artificial conditions, it would be useful to have codominant molecular markers to screen homozygous and heterozygous genotypes in backcross breeding programs rather than to apply gas-liquid chromatography. Two QTLs corresponding to L1 and L2 loci were detected in a doubled haploid progeny derived from the cross 'Stellar' x 'Drakkar' (Jourdren et al. 1996a). RAPD markers used for this study were dominant and were not strictly linked to the L1 and L2 loci. So, we decided to test candidate genes to generate codominant PCR-based markers for the ‘low C18:3 level' trait. Primers chosen on Arabidopsis thaliana fad3 gene (Arondel et al. 1992) allowed us to amplify the B. napus fad3 locus corresponding to B. rapa genome (Bn-fad3A) which was shown to be completely linked to the L1 locus (Jourdren et al. 1996b). In the present work, we have looked first for a genetic linkage between the B. napus fad3 gene corresponding to the C genome (Bn-fad3C) and the L2 locus, and second for a 'Stellar' specific mutated allelic form of A and C Bn-fad3 genes which could be easily detected by a PCR-based test. The validation of the two fad3 candidate genes for L1 and L2 QTLs is discussed, and a new diagnostic-assisted breeding program is presented.

Material and methods

Plant material. Doubled haploid plants were generated from the cross ‘Stellar’ (3% C18:3) x ‘Drakkar’ (9% C18:3) as described in Jourdren et al. (1996a). Twenty eight other accessions (listed in table 1) were tested in this study including ‘Rak-1’, another low C18:3 rapeseed expected to be originated from the same 'M11' mutant as ‘Stellar’.

DNA extractions. DNA extractions were performed as described in Jourdren et al. (1996b).

PCR-based diagnostic. A 'Restriction site Generating-PCR' test (Halliassos et al. 1989; RG-PCR, Gasparini et al. 1992) was choosen for optimal detection of the two mutations using PCR amplification, restriction endonuclease digest and agarose gel electrophoresis. Localization of the mutations in the fad3 sequences, PCR conditions and primers used for PCR are confidential data which can be obtained from INRA through a 'know-how' secret agreement.

Cloning PCR products and PCR amplification of plasmid inserts. Cloning PCR products were performed as described in Barret et al. (1998b). PCR amplification of plasmid inserts was described in Barret et al. (1998a).

Genetic analysis. Genetic analysis were performed as described in Jourdren et al. (1996b).

Results and discussion

The A and C Bn-fad3 genes mapped the two L1 and L2 loci controlling linolenic acid content in the seeds of rapeseed. Two mutations that could involve genes disruption have been identified and a specific codominant PCR test for detecting each point mutation is presented.

RG-PCR for optimal detection of L1 and L2 'Stellar' mutation

In order to obtain a codominant marker without the presence of a pre-existing restriction site simple to use, we decided to create a restriction site with direct mutagenesis using the PCR primer as a template (Halliassos et al. 1989). After PCR amplification with primers specific of the A or C Bn-fad3 gene, restriction endonuclease digest allow us to detect the Bn-fad3A 'Stellar' mutation (figure 1) or the Bn-fad3C 'Stellar' mutation (figure 2).

Figure 1 : Agarose gel electrophoresis (inverted ethydium bromide coloration) of three allelic forms of the Bn-fad3A gene (A) and two allelic forms of the Bn-fad3C gene (B) detected by RG-PCR test. The a1 and b1 allelic forms are specific to the ‘Stellar’ accession.

The two allelic forms a1 and b1 for these two loci specific to ‘Stellar’ and ‘Rak-1’ accessions, are expected to be originated from the same mutagenesis event (Table 1). These two RG-PCR tests are simple to use and have a codominant status.

Table 1 : Allelic forms of the 30 B. napus accessions tested.

Locus L1 / Locus L2	Accessions
a1 / b1	Stellar, Rak-1
a2 / b2	Accord, Cesar, Darmor-bzh, Drakkar, Hobson, Regent, Tanto, Westar, Yudal
a3 / b2	Bienvenu, Ceres, Chine 32, Darmor, Diadem, Envol, Eurol, Gaspard, Genkaï, Liratop, Maxol, Norin 9, R26, R51, Samouraï, Tapidor, Vivol, Wesbrook, Wesway

The two Bn-fad3 genes are good candidates for controlling linolenic acid level in B. napus seed

The rapeseed accession 'Oro' has been submitted to EMS mutagenesis and selected for low C18:3 level. As the resultant accession 'Stellar' was 'high C18:1 + C18:2' and 'low C18:3', Bn-fad3 genes which are converting C18:2 onto C18:3 could have been the mutagenesis target.

The two QTL controlling linolenic acid content in the seeds on the cross 'Stellar' x 'Drakkar' were located exactly at Bn-fad3A and Bn-fad3C loci. Then, these two genes are good candidates for controlling the linolenic acid level in this cross. Only two bands were amplified in rapeseed with the two A. thaliana fad3 primers (one corresponding to B. oleracea gene and one for B. rapa gene) and no sequence disruption was observed in all the wild type exonic sequences analyzed, by comparison with the functional A. thaliana gene previously published (Arondel et al. 1992). So, we can expect that the two fad3 genes that we have analyzed are the two rapeseed fad3 functional genes. Genetic evidences (the two fad3 loci were located just upon the peak of the two QTL) and molecular evidences (one mutation for each Bn-fad3 'Stellar' gene) were in concordance with the hypothesis of the implication of these two genes in the 'low C18:3 level' trait of 'Stellar'.

Even if strong evidences were in concordance with our hypothesis, additional experiment should be performed for confirmation. At first, we are not sure that the mutations that we have identified really disrupt the protein activity. Another mutation elsewhere on the gene or in regulator sequences could be responsible of A or C Bn-fad3 gene disruption. Secondly, even if mutation implication were checked, we are not sure that we are not looking at a pseudogene and that the functional Bn-fad3 gene is not located just nearby our gene. Functional complementation of 'Stellar' by genetic transformation with constructions including our wild type A or C Bn-fad3 genes could allow us to check for their respective functionality.

Application in breeding programs

The RG-PCR test presented here allows us to specifically detect the two mutations conferring the 'low C18:3 level' trait to 'Stellar' accession.

The RG-PCR test is simple to use with a minimal quantity of equipment (thermocycler and agarose gel electrophoresis equipment) and is relatively cheap (about 1.5-2 US$ for the genotyping of one plant for A and C Bn-fad3 mutations). Its codominant status makes it possible to select plants homozygous for the A and C mutations in segregating populations. Since 'C18:3 level' is highly influenced by environmental conditions and not always reproducible between greenhouse and field conditions, breeding programs are conducted in two steps in field conditions with one generation per year. The first step is to create near isogenic lines or hybrid through backcross and the second will be to transfer the 'low C18:3' trait in populations to develop recurrent breeding. Since molecular markers are not influenced by the environment, RG-PCR diagnostic assisted selection will allow us to make 'low C18:3 level' selection in greenhouse, with a four fold acceleration of breeding programs (four years instead of sixteen, figure 3 A and B).

Figure 3 : A. Breeding program for ‘low C18:3’ winter rapeseed with GC analyses of the trait. Sixteen generations must be conducted in field conditions to obtain B3F4 plants, which corresponds to sixteen years of experimentation and analyses. B. Diagnostic assisted breeding program. Seven generations in greenhouse are necessary to obtain B3F4 plants, which corresponds approximately to four years.

References

1. Arondel V, Lemieux B, Hwang I, Gibson S, Goodman HM, Somerville CR (1992) Map-based cloning of a gene controlling omega-3 fatty acid desaturation in Arabidopsis. Science 258, 1353-1354.

2. Barret P, Delourme R, Renard M, Domergue F, Lessire R, Delseny M, Roscoe TJ (1998a) A rapeseed fae1 gene is linked to the E1 locus associated with variation in the content of erucic acid. Theor Appl Genet 96, 177-186.

3. Barret P, Delourme R, Foisset N, Renard M (1998b) Development of a SCAR (Sequence Characterised Amplified Region) marker for molecular tagging of the dwarf BREIZH (Bzh) gene in Brassica napus L. Theor Appl Genet 97 , 828-833.

4. Gasparini P, Bonizzato A, Dognini M, Pignetti PF (1992) Restriction site generating-polymerase chain reaction (RG-PCR) for the probeless detection of hidden genetic variation: application to the study of some common cystic fibrosis mutations. Mol. Cell. Probes 6, 1-7.

5. Haliassos A, Chomel JC, Tesson L, Baudis M, Kruh J, Kaplan JC and Kitzis A (1989) Modification of enzymatically amplified DNA for the detection of point mutations. Nucleic Acid Res. 17, 3606.

6. Jourdren C, Barret P, Horvais R, Delourme R, Renard M (1996a) Identification of RAPD markers linked to linolenic acid genes in rapeseed. Euphytica 90, 351-357.

7. Jourdren C, Barret P, Brunel D, Delourme R, Renard M (1996b) Specific molecular marker of the genes controlling linolenic acid content in rapeseed. Theor Appl Genet 93, 512-518.

8. Miquel FM, Browse JA (1995) Molecular biology of oilseed modification. INFORM 6, 108-111.

9. Röbbelen G, Nitsch A (1975) Genetical and physiological investigations on mutants for polyenoic fatty acids in rapeseed (B. napus L.) I. Selection and description of new mutants. Z Pflanzenzuechtg 75, 93-105.

10. Scarth R, Mc Vetty PBE, Stefansson BR (1988) Stellar low linolenic-high linoleic acid summer rape. Can J Plant Sci 68, 509-511.