Use of genetic markers in Alstroemeria

This thesis describes the results of various applications of the AFLP technique in Alstroemeria . The aim of this study was 1) to adapt the AFLP technique for Alstroemeria species which has a large genome size, 2) to study the genetic diversity of Alstroemeria species of Chilean and Brazilian origin, 3) to construct genetic linkage maps of the A. aurea genome and 4) to map in A. aurea QTLs involved in ornamentally important traits.

The AFLP technique was adapted to obtain a method that produces clear fingerprints in Alstroemeria . We used PCR primers with two selective nucleotides (= Eco RI+2/ Mse I+2) during preamplification before PCR amplification with 33P labelled primers with four selective nucleotides (Eco RI+4/ Mse I+4) in the final step (Chapter 2). It was noticed that increasing the number of intermediate pre-amplifications was not preferable, because of the increased bias due to competition between fragments and the extra labour. The GC contents of the selective nucleotides had a significant influence on the number of bands. The primer combinations with CG residues in the selective nucleotides showed fewer bands per lane (Chapter 2). This result confirmed earlier other observations that the Alstroemeria genome is AT rich.

In our study, the reproducibility of the AFLP technique for genetic analysis was verified. All bands in the fingerprints of offspring genotypes could be explained from the parental genotypes (Chapter 2). No PCR artefacts, mismatching and random priming had been detected in general, but reliable fingerprints of a species with a large genome such as Alstroemeria can be obtained with Eco RI+2/ Mse I+2 preamplification and Eco RI+4/ Mse I+4 for final amplification.

AFLP fingerprints were produced of 22 Alstroemeria species, one interspecific hybrid ( A. aurea x A. inodora ) and the distantly related species Bomarea salsilla and Leontochir ovallei as outgroup (Chapter 3). AFLP template of three accessions per species was mixed to obtain a more generalised fingerprint to represent an Alstroemeria species. Three primer combinations (E+ACCA/M+CATG, E+ACCT/M+CATC and E+AGCC/M+CACC), selected on the basis of their fingerprint quality, resulted into a data set of 272, 211 and 233 markers per primer combination.

The UPGMA dendrogram revealed three main clusters: the Chilean species, the Brazilian species and the outgroup. The principal co-ordinate plot revealed the same three groups, but additionally, the A. ligtu group was separated from the Chilean group. A. aurea was positioned between Chilean and Brazilian groups. The unique position of A. aurea suggests that other Chilean and Brazilian species may have evolved from A. aurea ecotypes (Tombolato, A.F.C., pers. comm.). A. haemantha Ruiz and Pavon was grouped with A. ligtu subsp. ligtu , A. ligtu subsp. incarnata and A. ligtu subsp. simsii . This confirms previous studies which assigned A. haemantha to the A. ligtu group. In the monography of Bayer on taxonomy of Chilean species suggested that A. haemantha and A. ligtu were synonymous names. Two species, A. umbellata and A. pelegrina , showed a genetic distance of only 0.26 GD, which is in the range of within-species genetic distances. The interspecific hybrid ( A. aurea´A. inodora ) showed a genetic distance of 0.45 GD and 0.59 GD with A. inodora and A. aurea , respectively. In the matrix of pairwise genetic distances these values were the lowest observed between the hybrid and any of the putative parental species. This example demonstrates that it seems to be feasible to identify the parental species of an interspecific hybrid on the basis of genetic distance values.

An F ₁ hybrid mapping population (N = 134) was established between two diploid A. aurea genotypes (A002 x A003; 2n = 2x = 16) in order to construct linkage maps. Over 374 polymorphic AFLP markers have been produced with 28 primer combinations (Chapter 4). Around 70 % of these polymorphic markers have been assigned to the linkage maps in either of the A002 and A003 parental map. As a result, these maps consisted of 8 and 10 linkage groups with 122 and 214 markers covering 306.3 and 605.6 cM, respectively. These differences between the two maps in terms of number of markers and total map length, indicates a different level of heterozygosity. This could have been caused by self-pollination for sexual maintenance of the accession by the breeders, leading to fixation on 50% of the genome. The two maps were integrated by using the F ₂ type of AFLP markers. The pollen colour locus was assigned on the A002-6 linkage group.

We also tested another method to handle the complexity of the large genome. Instead of adding selective nucleotides to 6-cutter template we generated AFLP template with an 8-cutter restriction enzyme Sse 8387I. Fingerprints generated with Eco RI+4/ Mse I+4 primers produced around 80 clear bands from which around 16 markers were polymorphic, whereas the fingerprints generated with Sse +2/ Mse I+3 primers produced 30 clear bands from which 9 markers were polymorphic (Chapter 4). On the one hand the simpler Sse 8387I / Mse I fingerprints were more easy to evaluate, on the other hand a higher number of useful markers could be obtained with the Eco RI/ Mse I fingerprints.

The previously established A. aurea linkage maps were used in order to map and characterise QTL for important traits, such as leaf morphology, the colour, size and shape of the flower, tepal stripe width, and productivity in terms of number of flowering stems and flowering period (Chapter 5). The majority of these traits were chosen on the basis of the UPOV list of cultivar descriptors. For all traits, except for flower openness, the offspring trait values displayed a continuous distribution, not deviating from normality. The transgressive segregations that phenotypic values of the progeny go beyond the parental values have been observed for the morphological traits of the leaves due to heterozygosity in the parents. Interval mapping and the Kruskal-Wallis test was performed to detect and to localize QTL, using separate parental data sets and non-integrated maps of A002 and A003. For interval mapping, a permutation test was used to empirically determine the significance threshold of the LOD score for each linkage group. This resulted in putative QTL with 95 % confidence threshold values ranging between LOD = 2.6 and LOD = 4.9, or in QTL with 99% confidence threshold values ranging between LOD = 3.85 and LOD = 4.81. In total 22 QTL for the traits studied were located throughout the map. The phenotypic variance explained by the QTL ranged from 11.2% to 32.2%. It was observed that many of the QTL did not reach high LOD values, or did not reach highly significant K-values in the Kruskal-Wallis test. It was also observed that QTL detected by interval mapping were not often confirmed by the Kruskal-Wallis test and vice versa. Probable explanations were discussed in Chapter 5. The overall conclusion is that, despite its large genome, the AFLP technology can be applied relatively easily in Alstroemeria for genetic and biodiversity studies.