Genetic diversity of blue-flowered Scilla species as determined by random amplified polymorphic DNA

The genus Scilla, which can be either narrowly or broadly defined, is placed in the family Hyacinthaceae (Dahlgren et a/. 1985). The exact number of species of Scilla represented in the southern African flora is uncertain. Baker (1897) recognised 56 species (in a broadly defined genus), whereas Obermeyer in Arnold and De Wet (1993) recognised four (in a narrowly defined genus), based largely upon Jessop's taxonomic revision of Scilla, Schizocarphus and Ledebouria (Jessop 1970). One of these is Scilla natalensis Planch. Reduced to synonymy in this species are S. kraussii Baker and S. dracomontana Hilliard & B.L.Burtt. The opinions expressed by Jessop (1970) are not shared by all taxonomists. Although S. natalensis (sensu stricto), S. kraussii and S. dracomontana are all blue-flowered, they differ from one another in habit and vegetative morphology (Figure 1 and Table 1 ). Griffioen and Edwards (1994) having analysed gross morphology, anatomy, micro-morphology of seed, pollen and leaf epidermal pattern. palynology and cytology concluded that the three blue-flowered taxa are distinct, an opinion followed by Pooley (1998). Molecular markers which reveal extensive polymorphism at the DNA level are powerful tools in the identification of plant species. One of the commonly used molecular markers is random amplified polymorphic DNA (RAPD) (Williams et at. 1990). RAPDs has been used to confirm somatic hybrids (Xu et at. 1993, Takemori et at. 1994), genetic properties of micropropagated plantlets (Rani et at. 1995), geographic variation of plants (Nakai et a/. 1996, Rajaseger et a/. 1999) and to evaluate the genetic relationship between plant species (Rath et a/. 1998, Sastad eta/. 1999). suggest that the three blue-flowered Scilla taxa are distinct species. Analyses of the RAPD products of the hybrid and the parent species revealed shared inherited polymorphism. The results indicate that RAPD markers can be applied successfully for the identification of economically important Scilla species.


Introduction
The genus Scilla, which can be either narrowly or broadly defined, is placed in the family Hyacinthaceae (Dahlgren et a/. 1985). The exact number of species of Scilla represented in the southern African flora is uncertain. Baker (1897) recognised 56 species (in a broadly defined genus), whereas Obermeyer in Arnold and De Wet (1993) recognised four (in a narrowly defined genus), based largely upon Jessop's taxonomic revision of Scilla, Schizocarphus and Ledebouria (Jessop 1970). One of these is Scilla natalensis Planch. Reduced to synonymy in this species are S. kraussii Baker and S. dracomontana Hilliard & B.L.Burtt.
The opinions expressed by Jessop (1970) are not shared by all taxonomists. Although S. natalensis (sensu stricto), S. kraussii and S. dracomontana are all blue-flowered, they differ from one another in habit and vegetative morphology ( Figure 1 and Table 1 ). Griffioen and Edwards (1994) having analysed gross morphology, anatomy, micro-morphology of seed, pollen and leaf epidermal pattern. palynology and cytology concluded that the three blue-flowered taxa are distinct, an opinion followed by Pooley (1998).
Molecular markers which reveal extensive polymorphism at the DNA level are powerful tools in the identification of plant species. One of the commonly used molecular markers is random amplified polymorphic DNA (RAPD) (Williams et at. 1990). RAPDs has been used to confirm somatic hybrids (Xu et at. 1993, Takemori et at. 1994, genetic properties of micropropagated plantlets (Rani et at. 1995), geographic variation of plants (Nakai et a/. 1996, Rajaseger et a/. 1999 and to evaluate the genetic relationship between plant species (Rath et a/. 1998, Sastad eta/. 1999.
suggest that the three blue-flowered Scilla taxa are distinct species. Analyses of the RAPD products of the hybrid and the parent species revealed shared inherited polymorphism. The results indicate that RAPD markers can be applied successfully for the identification of economically important Scilla species.
In South Africa the bulbs of S. natafensis are used extensively as a traditional medicine. Homo iso-flavanones have been isolated from the bulbs of all three blue-flowered Scilla species (Crouch et at. 1999). Van Wyk et a/. (1997) suggested that the wound-healing, antiseptic and anti-inflammatory properties of these bulbs are possibly due to the presence of these homo iso-flavanones. McCartan and Van Staden (1998) developed a method for the micropropagation of the large S. natalensis, which could facilitate conservation by cultivation, alleviating pressure on natural resources. Since an increasing amount of research is focusing on the traditional uses and benefits of these plants as well as their horticultural potential, it is necessary to resolve the taxonomic uncertainty surrounding the number of infrageneric taxa in genus Scilla s. str. The aim of this study was to obtain conclusive data concerning the genetic relationship of the three blue-flowered species and an artificial hybrid (created by Mr R Roth) using RAPD techniques.

Material and Methods
Samples from fully expanded leaves of Scilla natafensis, S. kraussii, and S. dracomontana, originally collected from their natural habitats in KwaZulu-Natal and now growing in the University of Natal Botanical Gardens, and an artificial hybrid, S. natalensis X S. kraussii, were collected in triplicate. Leaves were washed, surface-sterilised and frozen with liquid nitrogen and then stored at -70°C for subsequent genomic DNA isolation.
Genomic DNA was isolated according to Pich and  (1 OOmg), was quickly frozen in liquid nitrogen, powdered in a mortar and pestle, and 6001JI extraction buffer (500mM NaCI, 50mM Tris HCI pH 8.0; 50mM EDTA; 1% 2-mercaptoethanol) added. The mixture was thawed on ice and 2601JI ice cold PVP stock solution added, whereafter 17.6mg solid SDS was added. The mixture was incubated for 1Om in at 65uC. Eighty-six IJI 5M potassium acetate were added followed by 30m in incubation on ice and centrifugation (1 0 OOOg) for 1 Omin. The supernatant was t ransferred into a new test-tube and mixed with 0.6 volume of iso-propanol and incubated on ice for 1Om in. After another centrifugation (1 0 OOOg) for 1Om in the supernatant was discarded and the pellet dried under vacuum. DNA was dissolved in 5001JI TE buffer and extracted once with 1 volume of phenol/chloroform/iso-amylalcohol (25:24:1). After centrifugation (1 OOOg for 10min at 4°C), the aqueous phase was transferred and DNA was precipitated in 1 volume iso-propanol. The DNA pellet was washed in 70% ethanol, dried under vacuum and dissolved in TE buffer and the DNA further purified by treating with RNase (1 01Jgl', 37"C, 30min) and proteinase K (121-Jgl '. 3JDC, 30min), respectively. This was followed by a phenol/chloroform extraction step. The DNA was precipitated by adding 2.5 volumes of pre-cooled absolute ethanol and centrifuged ( 1 0 OOOg) for 1Om in at 4°C. The pellet was dissolved in 501-JI TE buffer and the DNA was quantified using a CARY 50 CONC UV-Visible Spectrophotometer. DNA was diluted to 200ng1JI ' for RAPD analyses. RAPD anaylses were carried out with varying concentrations of MgCh (1.0, 1.5, 2.0, 2.5, 3.0, 3.5, 4.0mM) and DNA template (1 00, 150, 200, 250, 300, 400ng) to optimise the PCR conditions. Taq (2.5 units), dNTPs (0.4mM) and primers (11JM) were used.
Twenty-four random primers, each consisting of ten nucleotides (OPA, OPB), were purchased from Operon Technologies, Inc. (Alameda, CA, USA). Seven primers tha t gave reproducible results in three independent DNA extractions were then chosen for further analyses. Their sequences (5' to 3') are as follows: OPA8: GTGACGTAGG; OPA14: TCTGTGCTGG; OPA16: AGCCAGCGAA; OPA18: AGGTGACCGT; OPA 19: CAAACGTCGG ; OPB1: GTTTCGCTCC; OPB7: GGTGACGCAG. The amplification products together with a size DNA marker (1 OObp DNA ladder, Pharmacia) were separated on 1.6% (w/v) agarose gels at 70 volts for 3h. The gels were stained with 0.51Jg IJI ' ethidium bromide and photographed under UV light. The RAPD bands were scored as discrete variables, using 1 to indicate presence and 0 to indicate absence of a band. A pairwise difference matrix between genotypes was computed using the simple matching coefficient. A phenogram was generated using the unweighted pair-group method with arithmetical averages as described by Sneath and Sokal (1973).

Results and Discussion
High quality DNA was obtained from all three Scilla species and the hybrid. In the PCR reactions best results were obtained with concentrations of 200ng DNA template and 1.5mM MgCI2. Initially 24 random primers were used in the screening process. Seven primers generated reproducible RAPD profiles (i.e. each primer generated at least two identical RAPD profiles). In all, the seven primers generated 266 reproducible bands ranging from 150 to 2 000 bp (Table 2). Representative RAPD profiles generated wi th primers OPA8, OPA14, OPA17 and OP87 are shown in Figure 2.
Despite exhibiting polymorphisms, S. natalensis, S. kraussii and S. dracomontana yielded several common bands. However, some very distinct bands were also observed, providing a useful tool to distinguish among the Scilla species and the hybrid. S. dracomontana could be differen-     (Table 3). Data obtained from all seven primers, used for all plant material analysed, were used for cluster analysis, and a phenogram constructed (Figure 3). Analysis of the RAPD profiles indicated that the blue-flowered species can be divided into two broad groups at the level of 0.53. The first group includes S. natalensis and S. kraussii at the level of 0.69. S. natalensis is a tall form and S. kraussii a middlesized form (Table 1 ). The bands at 200, 800, 900 and 1 600bp were common to the two species. They could be distinguished from each other by the presence of the 300, 480, 580 and 1 400bp in S. natalensis in the RAPD profiles generated by primer OPA8 (Table 3). The second group includes S. dracomontana only which is a very small form. These results support the notion that the three species are distinct taxa (Griffioen andEdwards 1994, Pooley 1998).
Cluster analysis of S. natalensis and S. kraussii with the hybrid ( S. natalensis X kraussil) indicated that the hybrid is closely (0.76) associated with S. natalensis, and less closely with S. kraussii (Figure 3). RAPD profiles of the two parents with the seven random primers tested were distinct from those of the hybrid. One set of representative profiles : Phenogram generated for blue-flowered Scilla species and an artificial hybrid indicating genetic relationships and percentage similarity. All RAPD bands (266 in total) scored from seven random primers were used lor the calculations generated by primer OPB7 is shown in Figure 2. Scilla natalensis, the hybrid (S. nata/ensis X kraussil) and S. kraussii exhibited one common band (1300bp), which might explain the success of hybridisation between the two species. The bands at 200bp and 500bp were specific for the hybrid and S. natalensls only. Similarly, the band at 800bp was specific for S. kraussii and the hybrid. Our observations indicate that S. natafensis is the female parent of the hybrid.
We are acutely aware of the limitations involved in basing taxonomic conclusions on a small sample size of the taxa concerned, as is the case in this study. It is nevertheless hoped that this data, combined with existing macromorphological and ecological evidence, will contribute towards the acceptance of the three blue-flowered Scilla taxa as three distinct species.