Taxonomic considerations among and within some Egyptian taxa of Capparis and related genera (Capparaceae) as revealed by RAPD fingerprinting

Taxonomic considerations among and within some Egyptian taxa of Capparis and related genera (Capparaceae) as revealed by RAPD fingerprinting.This investigation was carried out to assess the taxonomic relationships among eight taxa of the Egyptian members of Capparaceae based on random amplified polymorphic DNA markers, and to compare the results with those obtained from morphological studies. A total of 46 bands were scored for three RAPD primers corresponding to an average of 15.3 bands per primer. The three primers (A03, A07 and A09) revealed eight polymorphic RAPD markers among the studied taxa ranging in size from 200 bp to 1000 bp. Jaccard’s coefficient of similarity varied from 0.28 to 0.84, indicative of high level of genetic variation among the genotypes studied. UPGMA cluster analysis indicated three distinct clusters, one comprised Cleome amblyocarpa and Gynandropsis gynandra, while another included two clusters at 0.74 phenon line; one for Capparis decidua, and the other for Capparis sinaica and all varieties of Capparis spinosa. The four varieties of Capparis spinosa were segregated at 0.84 phenon line. However, one of these varieties was more closely related to Capparis sinaica than to the other three varieties of C. spinosa. The RAPD analysis reported here confirms previous studies based on morphological markers.


INTRODUCTION
The family Capparaceae Juss. is a fairly large family (45 genera and 675 species), mainly subtropical, being most conspicuous in tropical seasonally dry habitats and with diversity in floral structure (Mabberley, 1997). The family is sometimes divided into eigth subfamilies and four tribes (Pax & Hoffman, 1936), or into two subfamilies: Capparoideae and Cleomoideae (Jafri, 1974). Actually, the two major subfamilies are quite distinct and have been elevated to family status by some authors (Airy Shaw, 1965;Hutchinson, 1967). Daniel & Sabnis (1977) determined flavonoids and phenolic acids in seven members of the Capparaceae and in five members of the Cleomaceae. The results supported the proposal (made on anatomical grounds) that Cleomaceae could be recognized as a separate family. Recently, this proposal was also supported by Kamel et al. (2009). The type genus of each of the subfamilies is by far the largest and houses the majority of the species: Capparis L. has about 250 species and Cleome L. has 200 species. This imbalance suggests that plants with extreme morphological traits may have been segregated into smaller genera, making the larger genera paraphyletic.
Capparis is a polymorphic genus and is distributed in the tropical and subtropical regions of the old and new world. According to Iltis (2001), there are about 110 Capparis taxa in the old world. The study of the reproductive characters in Capparis is problematic due to the difficulty of preserving the flowers (Hedge & Lamond, 1970). In Egypt, Täckholm (1974) recognized six species of Capparis, whereas Boulos (1999) recognized three species and four varieties.
Cleome has nine species in Egypt. Unresolved problems are still in need of further studies concerning Cleome gynandra L., which has been treated in the recent past as belonging to a separate genus Gynandropsis, and it was so treated in Graham's manuscript, and in Täckholm (1974), Jafri (1977) and Boulos (1999). Iltis (1957) gave convincing reasons for restoring this species to Cleome, and his treatment was later followed by Ridley (1967), Stewart (1972) and Thulin (1993).
During this study, we realized that morphological variability has led to much confusion in distinguishing species using the diagnostic characters proposed by different authors (Zohary, 1960;Davis, 1965;Jacobs, 1965;Al-Gohary, 1982;Higton & Akeroyd, 1991;Heywood, 1993;Fici & Gianguzzi, 1997;El-Karemy, 2001). Species identification is hard or even impossible when only vegetative parts are present, which is often the case during collection. Additional information about the genotype of plants is very much needed to resolve taxonomic problems in these genera. Because the genotype is not influenced by environmental factors, evolution of closely related taxa can be investigated from an objective point of view with molecular techniques (Hillis, 1987). In addition, some molecular marker assays, e.g., the use of random amplified polymorphic DNA (RAPD), allow the detection of DNA polymorphisms by randomly amplifying multiple regions of the genome by polymerase chain reaction (PCR) using single arbitrary primers designed independently of target DNA sequence (Welsh & McClelland, 1990;Hadry et al., 1992;Williams et al., 1993;Karp et al., 1996). Therefore, it has been extensively used as a genetic marker for estimating genetic, taxonomic, and phylogenetic relationships of plants and animals (Williams et al., 1990;Wachira et al., 1995;Kapteyn & Simon, 2002;Belaj et al., 2003;Deshwall et al., 2005). The method does not require any prior characterization of the genome to be analyzed unlike universal PCR where sequence information is a prerequisite for designing primers. RAPD analysis requires only a small amount of genomic DNA and can produce high levels of polymorphism and may facilitate more effective diversity analysis in plants (Williams et al., 1990). The technique is therefore simple, fast, and efficient, and requires little tissue for assays. RAPD has proved to be a good marker to assay and evaluate the genetic relationships among species and even among populations and individuals of the same species Warburton & Bliss, 1996).
Publised studies on Capparaceae based on molecular markers are scarce. Application of RAPD technique has been previously performed for the conservation of isolated populations of the extensively grazed range plant Capparis decidua in Saudi Arabia (Abdel-Mawgood et al., 2006). Inocencio et al. (2005) used genetic fingerprinting technique (AFLP) to determine the relationships among species of Capparis. Genetic distances, based on AFLP data were estimated for 45 accessions of Capparis species from Spain, Morocco and Syria. The results of this analysis supported the differentiation of four of the five taxa involved.
To our knowledge, there is no published information on the use of RAPD-PCR markers for the characterization of genetic relationships of the Capparis species that we study in this contribution. Therefore, the present study was undertaken to assess taxonomic relationships and divergence within and among the species of the genus Capparis from Egypt using RAPD markers, and to compare the results with those obtained from morphological studies.
A secondary aim was to determine whether Gynandropsis must be considered an independent genus or a synonym of Cleome

MATERIAL AND METHODS
A total number of eight taxa from different populations from Mersa Matruh were collected along the western Mediterranean coast, Sinai Peninsula and Ismailia region (Table 1). Ten individuals per population were included in the study. Young, healthy leaves were frozen or dried in plastic bags with silica gel until extracted in the Molecular laboratory, at Cairo University (Egypt). At least three independent leaf samples were collected for each species, in order to account for any artifactual amplification. Further descriptions of morphological traits and flowering were noted in separate data sheets for all individuals. The nomenclature follows Zohary (1966), Täckholm (1974), Thulin (1993) and Boulos (1999). A voucher of each species used in this study was deposited in Cairo University Herbarium (CAI). Initially, total genomic DNA extraction was carried out by SDS-proteinase K treatment (Brown, 1991) but this yielded poor quality DNA, as indicated by the brownish pellets. Subsequently, a modified version of the CTAB DNA extraction protocol described by Brown (1991) yielded high-quality genomic DNA.
Because the RAPD-PCR technology is sensitive to changes in experimental parameters, a total of ten primers were initially screened against ten individuals selected from every taxon. Experiments were carried out with varying concentrations of magnesium chloride (1.5, 2.0, 2.5, 3.0, 3.5, and 4.0 mM) and DNA template (5, 10, 15, 20, and 25 ng), in order to optimize the PCR conditions. The length of the denaturation stage of the amplification was also examined. When trying to optimize annealing temperatures, we ran test reactions at 34°C, 35°C, 36°C, and 37°C. The decamer primers could be clearly amplified at 34°C. A subset of 10 primers for further analyses was based on the following criteria: (i) consistent, strong amplification products, and (ii) production of uniform, reproducible fragments between replicate PCRs. An initial screening resulted in selection of three decamer oligonucleotides from the "A" RAPD primer kit that produced clear and reproducible amplification product, and estimated fragment size (bp), e.g. OPA03700 stands for the 700 bp marker generated by primer 03, kit A, from Operon Technologies Inc. (Alameda, CA, USA).
Amplifications were performed in 50-µl reaction volumes containing 1unit of Taq polymerase BioChain, 10 mM Tris-HCl (pH 9.0), 25 mM KCl, 4 mM MgCl 2 , 10 mM of each dNTP, 1µM each of random primer and 25 ng of template DNA. The negative control consisted of all reagents, except the DNA template in the reaction mixture. The cycling regime for the reaction was as follows:  (Sambrook et al., 1989), and the banding patterns were compared. RAPD bands were discerned from the agarose gel. Any fragment thought to be artifact, based on the controls or those too difficult to score with certainly, were not included in the data set and only distinct reproducible, well-resolved fragments were scored (1) for presence and (0) for absence of bands and assembled into a data matrix (Table 2). The genetic similarity between different pairs of individuals was estimated according to Jaccard's coefficient. A dendrogram following the unweighted pair group arthimetic average method UPGMA algorithm in the Multi Variate Statistic Package (MVSP) for windows version 3.13 (Kovach, 1999) was generated with the Jaccard's coefficient based on the entire marker generated (Rohlf, 1992).

RESULTS AND DISCUSSION
A necessary precondition for any RAPD analysis is the establishment of PCR conditions that ensure reliable and reproducible results (Ramser et al., 1996). Various parameters likely to affect PCR amplification were optimized. However, only data from optimum amplification conditions in our experiments are presented here.
To investigate the intra-varietal polymorphism among the studied taxa, the different RAPD profiles within the different species and varieties were compared. Figures 1 and 2 represented RAPD profiles obtained from primers A03 and A07, respectively. A total of 46 bands were scored for the three RAPD primers corresponding to an average of 15.3 bands per primer. The three primers revealed eight polymorphic RAPD markers among the studied taxa, ranging in size from 200 bp to 1,000 bp, and one monomorphic marker at A07 of 750 bp (Table 2). Within Capparis, some fragments were shared by two species, and that was clearly observed for Capparis sinaica and C. decidua (A03, 1000 bp). Some bands that were present in different genera were recorded: a band at A03 (750 bp) was present in all studied taxa except in Cleome amblyocarpa; a band at A03 (200 bp) was present in all species of Capparis; and a band at A07 (1000 bp) was present in the genus Capparis and in Cleome amblyocarpa. Species-specific RAPD markers suitable for discriminating the studied taxa of Capparis and their allied genera were also detected.  The polymorphism in presence/absence of RAPD fragments was used to construct a dendogram (Fig. 3) based on Jaccard's similarity coefficient. In the present analysis of eight taxa, two major clusters were identified. The first included Cleome amblyocarpa and Gynandropsis gynandra, while the latter included two clusters; one for Capparis decidua, while the other for C. sinaica and all varieties of C. spinosa.

Primer Nucleotide Marker code sequence (5'-3') (bp) CP SP CP OV CP DS CP DC CP OR CP SN CL AM GY GY
Cluster analysis revealed also that C. spinosa var. inermis was closer to C. sinaica than to the remaining three varieties of C. spinosa. These results supported the earlier taxonomic studies and numerical analyses (Abd El-Ghani et al., 2007;Kamel et al., 2009Kamel et al., , 2010. Thus, C. spinosa var. inermis could be treated as an independent species, C. orientalis, as suggested by some authors (Täckholm, 1974;Al-Gohary, 1982;Inocencio et al., Boulos (1999). Thus, a DNA based diagnostic assay like RAPD is able to identify genotypes directly and can therefore be of help to mitigate complications arising from earlier morphological studies. Additional future studies should focus on the taxonomic and genetic relationships among Capparoideae and Cleomoideae, not only in Egypt but in a wider geographic scale.
RAPD data does not contradict the recommendation of considering Gynandropsis gynandra within Cleome as Cleome gynandra (Kamel et al., 2010), because it is grouped with Cleome amblyocarpa. However, further studies including other species of Cleome should be undertaken to confirm these results.
Similar studies are important in detailing the level of variation and relationships within and between the species in order to plan future domestication trails and to manage properly the wild species collections which are kept in gen banks. In conclusion, we recommend that RAPD analyses are used as additional tool in the study of members of the Capparaceae.