Genetic diversity and determination of the optimum number of RAPD markers in lettuce ( Lactuca sativa L . )

Twenty lettuce accessions from the UENF Germplasm Collection were assessed for genetic diversity using RAPD markers. Multivariate statistical techniques were used, such as grouping analyses, using the Tocher and Single Linkage methods. Analysis of 55 polymorphic markers, obtained with 25 primers, showed that there was sufficient variability in the material studied to be exploited in future breeding programs. The ‘BGH 4060’ and ‘Grand Rapids’ accessions were the most dissimilar, while ‘BGH 4325’ and ‘BGH 4326' were the most similar. There was high agreement between ecogeographic origin and molecular similarity. It was further found by the stress statistic (Kruskal, 1964) and by the correlation between number of markers and increase in ideal formation of groups that 50 polymorphic markers is the optimum number for genetic diversity study in the assessed


Introduction
Lettuce, which originated in the Mediterranean basin, spread with the Roman conquests to France, England and the rest of Europe.With the discovery of the New World in 1492, it spread rapidly to the Americas and was introduced in Brazil in 1647 (Ryder, 1986).
Currently, it is a popular vegetable throughout the world and one of the most important in economic terms.It is part of the Brazilian diet, and it is the most important leafy vegetable (Alvarenga, 1999).It has a high pro-vitamin A content, 400 U.I. in 100 grams of green leaves, about four times the content found in tomatoes (Sonnenberg, 1985).
In spite of being the second most commercialized leafy vegetable in the State of Rio de Janeiro, Brazil, according to Ceasa (wholesale vegetable market) data of 1998, lettuce has not been investigated for genetic diversity; thus, the producers use a reduced number of genotypes which compete for the vulnerability of the varieties planted.
Thus, any research that can contribute to widen the knowledge on the species genetic diversity is necessary and justifiable.The DNA makers of the RAPD type are a valuable option, because they consist of relatively simple procedures, with the advantages of not being subject to environmental influence or epistatic interactions and offering high sensitivity in detecting variability.
This study was carried out to investigate the genetic-molecular diversity among 20 L. sativa accessions from the Germplasm Collection at the Center of Agricultural Sciences and Technology (CCTA) at the Universidade Estadual do Norte Fluminense (UENF), Rio de Janeiro, Brazil, using the RAPD markers technique.Furthermore, by the RAPD technique, the optimum number of polymorphic markers recommendable for a safe analysis of the genetic diversity in lettuce was determined to cut costs in the development of future studies.

Genetic materials
Twenty lettuce (Lactuca sativa L.) accessions were assessed from the Germplasm Collection at the Center of Agricultural Sciences and Technology (CCTA) at the Universidade Estadual do Norte Fluminense (UENF), state of Rio de Janeiro and are described in Table 1.Seedlings were produced in extruded polystyrene trays with drainage holes, previously filled with organic-vegetable substrate.

DNA extraction
Healthy leaf samples were removed in bulk from twenty plants per accession, seven days after planting, while still in the extruded polystyrene tray, and immediately placed in plastic bags, properly identified by labels and submerged in liquid nitrogen.The samples thus obtained were taken to the Plant Genetic Breeding Laboratory (LMGV) of the CCTA/UENF and kept at -85 º C in an ultrafreezer.The methodology, according to Doyle and Doyle (1990), was used to extract the DNA with modifications implemented by the LMGV.
The amplification reactions were performed according to Willians et al. (1990), with modifications.For this, each reaction contained 12.05 µL distilled water, 2.50 µL buffer 10 X (Tris HCl a 10 mM, pH 8.3; KCl at 50 mM); 2.00 µL MgCl 2 ; 1.25 µL dNTPs; 0.20 µL of Taq polimerase enzyme; 2.0 µL of one primer; and 5.0 µL DNA.After amplification, the DNA fragments underwent electrophoresis in 1.4% agarose gel.The electrophoresis was performed by submerging the gel in TAE buffer (Tris base, sodium acetate, 0.5 M EDTA and distilled water) 0.5X at 100 volts for 3h, using as standard the size of the lambda bacteriophage DNA, from cleavages with the restriction enzymes Bam Hl, Eco Rl and Hind III.
After electrophoresis, the gels were stained for 30 minutes in ethyl bromide solution (75 µL of ethyl bromide / liter of distilled water) and photographed under ultraviolet light, using the Stratagene 'Eagle Eye' photodocumentation system.

Statistical analysis of RAPD markers
The polymorphic RAPD markers were used to make a matrix of binary data, by attributing 1 to the presence and 0 to the absence of a determined band.The distance among the pairs of accessions was calculated based on the Arithmetic Complement of the Jaccard Index (Amaral Júnior and Thiébaut, 1999) expressed by: where a represents the number of DNA fragments, codified with 1 (positive agreement) common to both the individuals and b and c registered the number of DNA fragments, in which both the individuals disagree, respectively represented by 1-0 and 0-1.
The matrix of binary distances was used to group the genotypes by the Tocher and Single Linkage methods using the GENES computer program (Cruz, 1997).In the Tocher grouping, the first group was formed by the pair with the lowest distance value (C ij ).From then on, new groups were formed adopting the criterion of the mean intragroup distance being lower than any intergroup distance.The group identification process by the Single Linkage Method was performed by successive identifications of the closest genotypes, starting from the most similar pair until a dendrogram was established (Cruz and Regazzi, 1997).
The optimum number of RAPD markers was also determined for studying the dissimilarity of the genotypes assessed by correlation among the number of markers and the increase in the groups ideal formation, and also by the stress statistic (Kruskal, 1964), based on the mean from 20 markers combinations, sampled 20 times, from the fifteenth polymorphic.The GQMOL computer program was used, developed by the Genetic Sector at the Universidade Federal de Viçosa, Minas Gerais, Brazil.

Polymorphic fragments
Table 2 shows that there was 67.03% polymorphism among the amplified fragments; in a total of 82 developed bands, 55 were polymorphic.
Genotype 11 (BGH 4059) was one of the most divergent and formed group VI alone.Also, genotype 5 (BGH 502) alone made up group V.
It is interesting to point out that there was high agreement in this grouping between the ecogeographic origin and molecular similarity.The latter consideration is supported by the fact that the great majority of the genotype pairs from the same locality was not disassociated in distinct groups.For example: introduction 2 (BGH 292) and 3 (BGH 303), 6 (BGH 2471) and 7 (BGH 2517), 8 (BGH 2625) and 9 (BGH 2629), and 14 (BGH 4325) and 15 (BGH 4326), derived, respectively from Sarzedo (MG, Brazil), Copenhagen (Denmark), Valence (France) and Ipameri (GO, Brazil), which were present in group I.   Five groups were established by the Single Linkage method, shown in Figure 1, taking as base the discrepant changes of level in the dendrogram, where group I united the genotypes belonging to groups I and II by the Tocher method.Genotype 5 (BGH 502) formed the second dendrogramic group, while in Tocher it was present on group V. Genotypes 4 (BGH 410) and 16 (Grand Rapids) formed the third group in both methods.The fourth group by the Single Linkage method was composed Acta Scientiarum: Agronomy Maringá, v. 25, n. 1, p. 1-5, 2003 by genotype 11 (BGH 4059) and the fifth group, made up of the genotypes 13 (BGH 4064) and 17 (Maravilha de Verão).These groups were equivalent, respectively, to groups VI and IV by the Tocher method.Thus, the results obtained by the Single Linkage method were very similar to those by the Tocher method.
The perfect genetic similarity between genotypes 19 (Regina) and 20 (Repolhuda Todo Ano) was not expected as these materials are phenotypically divergent in the literature, and it can be quoted for example, that Regina is classified as a classic example of a looseleaf variety while, as the name indicates, the 'Repolhuda Todo Ano' is a classic example of the butterhead type.However, during the course of this study, the 'Regina' variety plants morphoagronomically assessed showed the cabbage formation in the field, which called attention at the time.The most plausible explanation for this is a possible error made in the introduction of the seeds of this material in the UENF Germplasm Collection, raising the hypothesis that the genotypes 19 (Regina) and 20 (Repolhuda Todo Ano) may constitute identical genomes.
Six of the 25 primers used in the present study, OPC 05, OPC 11, OPI 11, OPN 12, OPO 04 and OPW 09 were also used by Kesseli et al. (1994) to construct a genetic linking map in lettuce.These primers, in the study by Kesseli et al. (1994) amplified seven distinct regions of the genome, of which five were present in different linkage groups.Considering that the authors identified eight large linkage groups (>70 cM) it may be supposed comparatively that the present study covered a large part of the lettuce genome.

Recommendable number of polymorphic markers
The mean of the correlation among the number of markers and the increase in the ideal formation of groups (Table 4) show mean values varying from 0.6575 and 1.000, which occurred, respectively, when 15 and 55 markers were sampled.When 0.05 was adopted as the cutting point for the stress statistic, it was initially concluded that, in the present study, the use of 45 to 51 primers would be sufficient to express the variability present.However, for the samples of 45, 46, 47, 48 and 49 primers, at least one sample presented stress values, for the estimate of the mean over 0.05, and therefore it was not considered as a potential optimum number of markers to detect the variability in the studied population.But 50 was the number of markers which did not present any sample values higher than 0.05 for the composition of the estimated mean of statistical stress, and therefore it Genotypes 1/ Acta Scientiarum: Agronomy Maringá, v. 25, n. 1, p. 1-5, 2003 was considered the optimum number of markers to investigate the genetic diversity in the studied material.For this magnitude of markers, the mean correlation was 0.9762, which, because it was close to 1.0000, portrays the good precision of the estimates obtained with the primers used, which will cut costs in future researches.

Table 1 .
Lettuce accessions with their respective identification and origin

Table 2 .
Oligonucleotids primers used and their respective base sequences and number of the fragments associated to them

Table 3 .
Grouping by the Tocher method of 20 Lactuca sativa genotypes based on 55 polymorphic RAPD fragments

Table 4 .
Estimates of the mean values of the correlation among the number of markers and the increase in the ideal formation of groups and mean of the stress coefficient estimate