Development of two sequence-specific PCR markers linked to the le gene that reduces pod shattering in narrow-leafed Lupin ( Lupinus angustifolius L . )

Wild types of narrow-leaf lupin (Lupinus angustifolius L.) have seed pods that shatter upon maturity, leading to the loss of their seeds before or during the harvest process. Two recessive genes have been incorporated into domesticated cultivars of this species to maximize harvest-ability of the produce. One of these genes is called lentus (le). Two microsatellite – anchored fragment length polymorphism (MFLP) candidate markers were identified as closely linked to the le gene in a recombinant inbred line (RIL) population derived from a domesticated x wild type cross. The candidate MFLP markers were isolated from the gel, re-amplified by PCR, cloned and sequenced. The MFLP polymorphisms were converted into sequence-specific PCR-based markers. Linkage analysis by MapManager indicated that one of the markers, LeM1, was 2.6 centiMorgans (cM) and the other, LeM2, was 1.3 cM from the gene, with both being on the same side. The correlation between the marker genotype and the plant phenotype for the le gene is 95% for the Australian cultivars, and approximately 36% on wild types tested. These markers may be useful in marker assisted selection for the le gene when introgressing wild material into lupin breeding programs.


Introduction
An essential characteristic of a modern grain crop is the ability to retain its seeds long enough to allow mechanical harvesting at full maturity.In the genus Lupinus, this characteristic has long been present in two species -Lupinus albus L., an old crop species of the Mediterranean region, and L. mutabilis (sweet) which has been cultivated in the Peruvian Andes for over 2000 years (Gladstones, 1967).
Narrow-leaf lupin (Lupinus angustifolius L.) is the most widely cultivated grain legume crop in Australia.Originally introduced into Western Australia (W.A.) towards the end of the 19 th century (Gladstones, 1994) as a green manure crop, these plants were essentially wild.Wild types of L. angustifolius have seed pods that shatter upon maturity, making harvesting very difficult.Research aimed at finding plants with non-shattering pods was reported to have begun in Germany in 1929 (Hanelt, 1960).They were successful in selecting a strain that had a reduced pod shatter (von Sengbusch and Zimmermann, 1937) apparently related to a large reduction in the thick-ness of the pod wall (Atebekova, 1958).Domestication of this species in W.A. began in 1960 with the discovery of two natural mutant (recessive) genes, lentus and tardus for reduced pod-shattering (Gladstone, 1967) and, in 1961 a gene for nil-vernalisation requirement and early flowering (Gladstones, 1969).
One of the genes for reduced pod-shatter, known as tardus (ta), affects the sclerenchyma strips of the dorsal and ventral pod seams, fusing the two halves to such an extent that separation of the two halves is greatly impeded.It was considered analogous to a gene previously found in L. luteus.The second gene lentus (le) modified the orientation of the sclerified endocarp of the pod, resulting in a reduction of torsional forces upon drying, and hence reduced pod shatter.This gene was also associated with a reduction in the thickness of the pod wall, but not to the extent of the previously found strain of von Sengbusch and Zimmerman (1937), leading to the conclusion that the controlling genes were not the same (Gladstones, 1967).This modification was associated with a change in internal pod pigmentation that gave the immature pods a purplish tinge and the inside surface of mature pods a bright yellowish-brown color (Gladstones, 1967).In addition, this trait is also associated with development of a reddish pigmentation within the stem of plants older than two months.The first Australian cultivar to carry both these genes was Uniharvest, released in 1971. Apart from Fest (released 1973) all Australian and many European cultivars released since then carry both of these genes.
Since field based selection for the le gene requires that the F 2 progeny be maintained for at least 8 weeks from sowing to ensure that this trait is retained, it would be more efficient to use molecular markers that could be used to evaluate progeny within the first weeks of growth.Recently, two molecular genetic maps have been produced in L. angustifolius on which le has been mapped (Boersma et al., 2005;Nelson et al., 2006).No marker was located within 5 centiMorgans (cM) of the le gene on the map of Nelson et al. (2006).Conversely, several markers on the map of Boersma et al. (2005) were closely linked to the le gene.However, the MFLP markers in the map of Boersma et al. (2005) are not implementable in practical lupin breeding as the multiple bands produced by their primer combinations may cause confusion, makes it impossible to multiplex several markers for high efficiency and, is too tedious and expensive to be implemented in a lupin breeding program for screening large numbers of plants.The aim of this research is to develop simple, sequence-specific PCR based markers linked to the le gene desirable for marker assisted selection (MAS) in lupin breeding.

Plant materials and phenotyping of le gene
A population of F 8 recombinant inbred lines (RILs) of a domestic x wild type (DxW) cross of L. angustifolius was previously developed by the Department of Agriculture and Food, Western Australia (DAFWA) using as parents lines P27255 (wild) and 83A:476 (domesticated).Part of the resulting RIL population has previously been used to produce linkage maps for L. angustifolius (Boersma et al., 2005;Nelson et al., 2006).
The full population consisting of two parents and 115 RILs, was sown in a screen-house during the winter growing season with 25-30 plants of each RIL grown in 1.5 meter rows.Plants were either rated as LeLe (wild, shattering; pigment absent) or lele (domesticated, reduced shattering; pigment present) based on the method described by Gladstones (1967).

Marker development
The parents and 89 RILs were previously subject to MFLP tests involving 10 SSR-anchor primers each in combination with 16 MseI-primers (Boersma et al., 2005).Two MFLP markers having the best correlation to the le phenotyping data were selected.DNA fragments from candidate MFLP markers were isolated from MFLP gels, re-amplified by PCR, ligated into plasmids (pGEM-T Easy Vector, Promega) and cloned into E. coli according to the manufacturer's instructions.Plasmid DNA with MFLP fragment in-serts were isolated from E. coli, and sequenced using the BigDye Terminator system (Applied Biosystems).To ensure the accuracy of DNA sequencing, at least five clones of E. coli were used to prepare the plasmid DNA containing the inserts, and sequenced separately.A sequence-specific primer was designed near the MseI-end internal to the SSR-MseI fragment.(Yang et al., 2002(Yang et al., , 2004;;You et al., 2005).Primers were designed so that the annealing temperature was approximately 54 °C based on calculations using the nearest-neighbor model (http://www.sigmaaldrich.com).

Testing of converted markers
Testing of the converted markers was achieved by PCR, using the sequence-specific primer in combination with the SSR anchor-primer from which the original MFLP polymorphism was produced (Yang et al., 2001(Yang et al., , 2002)).Marker DNA fragments were amplified in a10 μL PCR consisting of 1.5 μL of the template DNA (approximately 100 ng), 0.5 unit of Taq polymerase (Fisher Biotec, Perth), 5 pmol of each primers, 67 mM Tris-HCl (pH 8.8), 2 mM MgCl 2 , 16.6 mM (NH 4 ) 2 SO 4 , 0.45% Triton X-100, 4 μg gelatin, and 0.2 mM dNTPs.The sequence-specific primers were labeled with γ-33 P based on a previously reported method (Yang et al., 2001(Yang et al., , 2002(Yang et al., , 2004)).PCR was performed on a thermocycler (Hybaid DNA Express) with each cycle comprising 30s at 94 °C, 30s at the annealing temperature (see below), and 1 min at 72 °C.The annealing temperature of the first cycle was 60 °C, and decreased 0.7 °C in each subsequent cycle until the temperature reached 54 °C.The final 25 cycles used an annealing temperature of 54 °C.The selective PCR products were separated on a 5% polyacrylamide denaturing sequencing gel (7 M urea) using a 38 x 50 cm (0.4 mm in thickness) Sequi-Gen GT sequencing cell (Bio-Rad).After electrophoresis at 55W for about 3h, the gel was dried on a gel drier (Model 583, Bio-Rad).Marker bands were detected by autoradiography (Yang et al., 2002;You et al., 2004) with overnight exposure of the X-ray film to the dried gel.

Confirmation of linkage
The two converted sequence -specific markers were tested on all 115 F 8 RILs of the population derived from the DxW cross.The marker score data and le phenotype data of the 115 F 8 RILs were merged and analyzed using MapManager (Manly et al., 2001) to determine the genetic linkage between the markers and the le gene.
Further testing of the markers was carried out on all 23 Australian historical and current cultivars and on 36 landrace accessions from the Australian Lupin Collection, to examine correlation of their phenotype with the le markers' genotypes.Landrace accessions were selected on the basis of their geographic origins.

Phenotyping of le gene
Phenotyping on the individual plants from each of the 115 RILs revealed that 55 RIL lines were homozygous for le (reduced shatter), and 60 lines were homozygous for Le (shattering).The segregation of RILs with the lele genotype to plants with the LeLe genotype fits the expected 1:1 ratio (χ2 = 0.217, P = 0.641) for a single gene.

DNA sequencing of candidate MFLP markers
Among the markers originated from 153 sets of MFLPs during a mapping study (Boersma et al., 2005), two dominant markers designated as DAWA323.150 and DAWA468.290 were identified as candidate markers for development as sequence-specific markers tagging the le gene.
Marker DAWA468.290 was also present in plants with lele genotype but absent in plants with LeLe genotype (Figure 2).This marker had not been shown in the map (Boersma et al., 2005) as it was also calculated to be 1.2 cM from the le gene, and therefore co-segregated with marker DAW323.150.DNA sequencing revealed that marker DAWA468.290 is a 284 bp fragment including the anchor primer sequence MF42 (3-GTCTAACAACAACAACA AC-5') and the sequence of primer MseI-CCA (5'-GATGA GTCCTGAGTAACCA-3') (Table 2).
Marker LeM1: Based on the DNA sequence of MFLP marker DAWA323.150, a sequence-specific primer LeMS1 (3'-TTAACGAACCTACCATTTG-5') was designed near the MseI end of the original MFLP marker (Table 1).The dominant MFLP marker was successfully converted into a dominant simple PCR based marker by  using the sequence-specific primer LeBMS1 in combination with SSR-anchor primer MF51.F 8 RILs with the genotype lele developed the 126 bp marker band, while plants with genotype LeLe did not develop this band (Figure 3).We designated this sequence-specific dominant marker linked to the le gene as LeM1.
Marker LeM2: A sequence-specific primer LeMS2 (3'-AGAAAAAGATGAATGCACG-5') was designed near the MseI end based on the sequence of MFLP marker DAWA468.290(Table 2).The marker was successfully converted into a simple PCR based marker by using primer LeMS2 in combination with SSR-anchor primer MF42.Plants with genotype lele produced a 204 bp marker band, but the marker fragment is absent in plants with LeLe genotype (Figure 4).We designated this marker as LeM2.

Confirmation of linkage
Markers LeM1 and LeM2 were tested on the parents and 115 RILs of the DxW population.Marker results and RIL genotypes showed imperfect correlation, with LeM1 having six differences and, LeM2 having three differences between the marker result and RIL genotypes.The three RILs which showed inconsistence between phenotype and marker genotype of LeM2 also showed the inconsistency with LeM1.Map distances were calculated by MapManager (Manly et al., 2001) to be 1.3 cM (LeM2) and 2.6 cM (LeM1) from the gene (Figure 5).
Among the 23 Australian cultivars, only the first released cultivar Uniwhite, carries the shattering alleles LeLe, and all the other 22 cultivars have the reduced-shattering genotype lele.The 22 cultivars having the lele genotype were correctly scored by the markers.However, the marker genotype and Le phenotype of Uniwhite did not match (Table 3).
The Australian lupin collection holds many accessions of wild and landrace L. angustifolius collected or received from a range of locations.Most of these accessions have been rated as either 'shedding' (shattering; no le or ta), 'reduced shedding' (carrying at least one gene reducing pod shatter) or 'non-shedding' (presumably carrying at least 2 genes for the reduction of pod shatter).
Of the 36 landraces, marker LeM1 designated only six accessions as Le ('shedding' (shattering)) and LeM2 designated seven as Le, despite all but one of them having previously been rated as Le ('shedding', Australian Lupin Collection database, 2005).Only one of the accessions was designated as Le by both markers.Similarly, both markers  Primer MseI-CCA (5'-GATGAGTCCTGAGTAACCA-3').
designated as le the one accession listed as 'non-shedding', received from Belarus (Table 4).Accessions correctly designated as Le or le were collected (or received) from eight countries including Belarus, France, Greece, Italy, Morocco, Portugal, Spain, Syria.Accessions from the other eight source countries including Cyprus, Chile, Germany, India, Israel, Turkey, USA and the USSR were incorrectly genotyped by the two markers.In total, 13 out of 36 (36%) land-race accessions showed a positive correlation between marker score and pod-shatter genotypes for one or both markers.

Discussion
We have successfully developed two sequence-specific PCR markers linked to the Le gene in L. angustifolius in this study.Both markers are simple PCR based, are costeffective, and could be used for MAS in lupin breeding.Previously Boersma et al., (2005) had mapped marker DAWA323.150 from which LeM1 was derived as being only 1.2 cM from the gene based on 89 RILs.This distance has now been re-calculated as being 2.6 cM from the gene on the basis of a population of 115 RILs, including the 89 RILs previously used by them.LeM2 which had not previously been placed on the map has been positioned mid-way between the Le gene and LeM1 at 1.3 cM.The genetic distance between the marker and the gene determines the ac-Markers for reduced pod shattering in Lupin 627    For MAS, it would be ideal to develop so-called 'perfect Markers' where the DNA fragments of the markers are actually on the gene of interest, in which case the linkage between the markers and the gene is almost unbreakable (Ellis et al., 2002).However, perfect markers are very difficult to develop.Most of the molecular markers being implemented in practical plant breeding programs are 'imperfect' where certain genetic distances exist between the markers and the gene of interest in the chromosomes (Staub and Serquen, 1996;Gupta et al., 1999).The two markers linked to the Le gene in lupin reported in this paper, are like-wise imperfect.Over long periods of time and in the crossing in plant breeding programs, genetic recombination may occur between the marker locus and the gene locus, and consequently plants not having the target gene may carry the marker DNA sequence, and vice-versa (Sharp et al., 2001, You et al., 2005).In MAS, a molecular marker can be used to screen the progeny from a cross only if the marker is polymorphic among the parents so that the desirable allele can be distinguished from the undesirable ones (Eagles et al., 2001).Fortunately, all the modern Australian cultivars having the lele genotype show the target bands for markers LeM1 and LeM2.However, testing of these markers on land-races that were fully shattering (having the LeLe,TaTa genotype) resulted in only 13 of the 36 landrace accessions correlating the marker bands with genotype.Eleven of those 13 accessions were correctly identified by either one or the other of the two markers, indicative of the degree of crossing over that has occurred between the marker and the le gene in the wild, even though it only occurs at an estimated 1.3-2.6% of the time in any given cross.The results suggest that these markers may be of use in MAS to screen the progeny of approximately one third of wild accessions when they are crossed with modern Australian cultivars.This study although not exhaustive, also suggests that these markers would be more likely useful when testing progeny of seed material originating from Greece, Morocco (source of DxW parent P27255), Portugal, Spain and Syria, although seed received from other sources including collections (e.g.France, Belarus) should not be ruled out.
When markers LeM1 and LeM2 are used to screen the F 2 progenies from DxW crosses, they enable breeders to eliminate the approximately 25% of plants which are homozygous for the LeLe allele.However, since both markers are dominant, they cannot differentiate homozygous lele F 2 plants from the heterozygous Lele plants.Therefore, further selection for the lele genotype may be necessary in later generations of the breeding cycle.

Figure 2 -
Figure 2 -MFLP fingerprinting on 20 RILs in Lupinus angustifolius generated by SSR-anchor primer MF42 in combination with primer MseI-CCA.RILs portrayed in this figure are the same as in Figure 1.Arrow indicates the candidate MFLP marker linked to the le allele at 284 bp.

Figure 4 -
Figure 4 -Screening molecular marker LeM2 on the parents and 28 F 8 derived RILs of Lupinus angustifolius using PCR with sequence-specific primer LeMS2 (3'-GAAAAAGATGAATGCACG-5') and SSR anchor primer MF42.The length of the band fragment produced here is 204 bp.RILs portrayed in this figure are the same as in Figure 3.

Figure 5 -
Figure 5 -Genetic linkage of the two molecular markers LeM1 and LeM2 and the reduced-pod-shatter gene lentus (le) of Lupinus angustifolius, as analyzed by MapManager.Distances between adjacent points given in centiMorgans (on the left side) between adjacent points on the linkage map are non-cumulative.

Table 1 -
DNA sequence of the MFLP marker DAWA323.150showing primers giving rise to sequence-specific marker LeM1.

Table 3 -
Correlation of marker score and Le phenotype of 23 Australian cultivars of Lupinus angustifolius. of marker assisted selection when the marker is used in the breeding program.Since markers LeM1 and LeM2 are both on the same side of the le gene, and LeM2 is closer to the gene than LeM1, it would be appropriate to use marker LeM2 in MAS if both markers are applicable. curacy

Table 4 -
Correlation of marker score and Le phenotype of 36 Wild and Landrace accessions of L. angustifolius from the Western Australian Lupin Collection.Both LeM1 and LeM2 are dominant markers, where presence of the marker band is linked to the le allele.Accessions for which the Le phenotype correlates with either or both of the two markers are presented in bold.+ = band present.-= band absent.