ELECTROPHORETIC PATTERN OF SEED PROTEINS IN TRIFOLIUM L. AND ITS TAXONOMIC IMPLICATIONS

The taxonomic delimitations of 61 taxa of the genus Trifolium L. belonging to presently accepted five sections, namely Lotoidea, Mistyllus, Vesicaria, Chronosemium and Trifolium are evaluated, based on numerical analysis of their electrophoretic seed protein profiles. The dendrogram, resulted from the hierarchical cluster analysis of SDS-PAGE profiles of seed proteins conform, with some restrictions, to the present splitting of the genus Trifolium into the sections but not into the subsections and series. Introduction The genus Trifolium L. (Clover) is one of the important genera of Papilionoideae of the Leguminosae with agricultural value. It contains 237 species and represented in all continents (Zohary, 1972b). The Mediterranean region and its adjacent countries are one of the main centres of distribution of Trifolium species, and also the centre of domestication and breeding of the cultivated species (Zohary and Heller, 1984). Several taxonomic treatments were made by botanists to divide the genus into natural groups. Linnaeus (1753) divided the genus into five groups, some of which were later accepted as sections. Seringe (1825) proposed the genus with seven sections. Presl (1832) splitted the genus into nine new genera and all of these genera are retained today as sections. Lojacono (1883) distinguished two subgenera within the genus and divided the first subgenus into 11 sections and the second one into only two sections. Boissier (1873) reduced the number of sections to seven. Hossain (1961) divided the genus into eight subgenera. Another approach was adopted by Zohary and Heller (1984), who recognized eight sections for the genus. The first and largest section is tentatively divided into nine subsections and 13 series. Based on morphological characters alone, it is difficult to distinguish the subordinate taxa of the genus Trifolium from one another because they have overlapping variations in terms of the major delimiting morphological and biological characters. The importance of electrophoretic evidence in plant systematics has been discussed in detail by mamy workers (Boulter and Derbyshire, 1971; Gottlieb, 1977; Ghareeb et al., 1999; Kamel, 2005). Electrophoretic profiles of seed proteins have been used in different systematic studies (Badr et al., 2000; Zecevic et al., 2000). In Leguminosae many studies have been carried out based on the electrophoresis of seed proteins (Hussein and George, 2002; Hussein et al., 2005). Electrophoretic patterns of total seed proteins as revealed by polyacrylamide gel electrophoresis (PAGE) with sodium dodecyl sulphate (SDS) have been successfully used to resolve the taxonomic and evolutionary problems of some plant species (Ladizinsky and Hymowitz, 1979; Potokina et al., 2000; Ghafoor and Arshad, 2008; Ayten et al., 2009). Badr (1995) and Nikolic et al. (2010) studied the electrophoretic seed profiles of some taxa of the genus Trifolium. Recently the phylogeny of the genus Trifolium was studied based on DNA sequencing (Ellison et al., 2006). Corresponding author. Email: nmgtadrous@yahoo.com Flora and Phytotaxonomy Research Department, Horticultural Research Institute, Agriculture Research Center, Cairo-Egypt.


Introduction
The genus Trifolium L. (Clover) is one of the important genera of Papilionoideae of the Leguminosae with agricultural value.It contains 237 species and represented in all continents (Zohary, 1972b).The Mediterranean region and its adjacent countries are one of the main centres of distribution of Trifolium species, and also the centre of domestication and breeding of the cultivated species (Zohary and Heller, 1984).
Several taxonomic treatments were made by botanists to divide the genus into natural groups.Linnaeus (1753) divided the genus into five groups, some of which were later accepted as sections.Seringe (1825) proposed the genus with seven sections.Presl (1832) splitted the genus into nine new genera and all of these genera are retained today as sections.Lojacono (1883) distinguished two subgenera within the genus and divided the first subgenus into 11 sections and the second one into only two sections.Boissier (1873) reduced the number of sections to seven.Hossain (1961) divided the genus into eight subgenera.Another approach was adopted by Zohary and Heller (1984), who recognized eight sections for the genus.The first and largest section is tentatively divided into nine subsections and 13 series.Based on morphological characters alone, it is difficult to distinguish the subordinate taxa of the genus Trifolium from one another because they have overlapping variations in terms of the major delimiting morphological and biological characters.
The importance of electrophoretic evidence in plant systematics has been discussed in detail by mamy workers (Boulter and Derbyshire, 1971;Gottlieb, 1977;Ghareeb et al., 1999;Kamel, 2005).Electrophoretic profiles of seed proteins have been used in different systematic studies (Badr et al., 2000;Zecevic et al., 2000).In Leguminosae many studies have been carried out based on the electrophoresis of seed proteins (Hussein and George, 2002;Hussein et al., 2005).Electrophoretic patterns of total seed proteins as revealed by polyacrylamide gel electrophoresis (PAGE) with sodium dodecyl sulphate (SDS) have been successfully used to resolve the taxonomic and evolutionary problems of some plant species (Ladizinsky and Hymowitz, 1979;Potokina et al., 2000;Ghafoor and Arshad, 2008;Ayten et al., 2009).Badr (1995) and Nikolic et al. (2010) studied the electrophoretic seed profiles of some taxa of the genus Trifolium.Recently the phylogeny of the genus Trifolium was studied based on DNA sequencing (Ellison et al., 2006).
In the present study, the taxonomic delimitations of 61 taxa of Trifolium are re-assessed based on the data resulted from SDS-PAGE profiles of their seed proteins.

Materials and Methods
In the present study, 61 taxa of Trifolium have been investigated.Sources of the seeds directly used for protein extraction are given in Table 1.To extract the seed proteins, 0.5 g of mature seeds ground to meal using a mortar and pestle.The meals were homogenized with 0.5 ml of Tris-HCl buffer containing 2% SDS and 10% sucrose at pH 6.8 for overnight at 4ºC.The slurry was centrifuged at 9000 rpm for 6 min.The supernatant (protein extract) was taken for loading on 12.5% polyacrylamide gel.Protein samples (20 µl) including loading dye were loaded in the stacking gel.Electrophoresis was carried out under non-reducing conditions in 12.5% polyacrylamide gel.The assay was carried out by an electric supply of 15 mA for 30 min, and then raised to 25 mA for 5-6 h, using a protein marker with low molecular weights.Gels were then stained in Coomassie brilliant blue for 16 h at room temperature, distained and photographed.The bands produced by each sample were counted.The similarity coefficient between the species based on comparisons of their SDS-PAGE profiles was calculated by Jaccard ' s coefficient using the SPSS program (version 10.1).
The data obtained from the seed protein banding patterns, each species, were subjected to the numerical analysis.The presence or absence of each of the bands (coded as 1 and 0 respectively) was treated as a binary character in a data matrix.The OUTs (Operational Taxonomic Units), produced from the analysis of SDS-PAGE profiles of seed proteins, collected from the investigated taxa of Trifolium, resulted in a dendrogram and it was compared with the current taxonomic treatments of the genus Trifolium.

Results and Discussion
The banding patterns of Trifolium taxa are shown in Figure 1.The seed protein profiles of examined taxa illustrated that bands in between marker weight 116KDs and 55KDs are homogenous in comparison to bands in between 50KDs and 14KDs.The relationships among the taxa of Trifolium are presented in Figure 2. The dendrogram resulted from the hierarchical cluster analysis of SDS-PAGE profiles of seed proteins of 61 Trifolium taxa conform, with some restrictions, to the splitting of this genus into sections, but not with the sub-sectional arrangement under the section Lotoidea and section Trifolium considered by Zohary and Heller (1984).
The dendrogram shows that the investigated taxa of Trifolium are split into two major clusters.The first major cluster includes 20 taxa belonging to section Trifolium and the second major cluster includes 41 taxa belonging to four sections, viz., Lotoidea, Mistyllus, Vesicaria and Chronosemium.Within the first major cluster, the taxa are divided into two clusters.The first one included T. alexandrinum, T. caudatum and T. canescens in which T. alexandrinum was delimited leaving T. caudatum and T. canescens as a group.In the second cluster, the taxa are divided into two groups.The first group includes T. arvense, T. bocconei, T. cherleri and T. incarnatum.The second group includes 12 taxa of section Trifolium.Trifolium ligusticum represents the subsection Phleoidea.The similarity between the taxa belonging to section Trifolium ranged from 36.4% to 100%.Zohary and Heller (1984) showed that section Trifolium ranks second in the number of species, after section Lotoidea, which is consistent with the results of this study.It is heterogeneous in appearance but have several distinctive proteins banding pattern after SDS-PAGE.But their agreement in splitting of section Trifolium into 17 small and natural clusters by Zohary (1971Zohary ( , 1972a, b), b), regarded as subsections does not conform to the results of this study (Table 1, Fig. 2).The grouping of T. caudatum and T. canescens, as well as, the high similarity  (95.7%) between them support their position in subsection Ochroleuca.On the other hand, T. alexandrinum show low similarity (36.4%) with T. apertum, although the obtained results, in the present work, referred that both the two species still delimited under the same section Trifolium, it may be claimed that the inclusion of them in the same subsection Alexandrina is inconsistent and needs further investigation.Three species T. cherleri, T. hirtum and T. lappaceum representing subsection Lappacea are distant from one another.This result implies that it may be better to treat them under separate subsections.Although T. angustifolium, T. purpureum var.desvauxii and T. dichroanthum belonging to subsection Angustifolia, T. angustifolium and T. purpureum var.desvauxii grouped together but T. dichroanthum grouped with T. clypeatum showing similarity (87.0%).Among the three taxa T. heldreichianum, T. medium var.medium and T. medium var.sarosiense comprising the subsection Intermedia, the two varieties of T. medium shows no difference with each other with a similarity of 100% and T. heldreichianum differs from these two varieties with a similarity of 52.05%.The present data show that the taxonomic delimitations in section Trifolium requires reconsideration and the number of its subsections as proposed by Zohary (1971Zohary ( , 1972a, b), b), should be reduced.The second major cluster comprising of four sections (Lotoidea, Mistyllus, Vesicaria and Chronosemium) is divided into two large clusters.One includes 30 taxa belonging to the section Lotoidea and other includes 11 taxa belonging to the sections Mistyllus, Vesicaria and Chronosemium.Within the large cluster of section Lotoidea the taxa combine variously and form six similarity groups as described below.The pairs of taxa T. tembense and T. thalii, T. mattirolianum and T. polystachyum, and T. isthmocarpum and T. masaiense are consequently segregated as separate groups.The remaining taxa of the section are separated into three groups.The first group is formed by T. africanum, T. amabile, T. ambiguum, T. bilineatum, T. burchellianum, T. burchellianum var.johanstonii, T. cernum, T. decorum, T. glomeratum and T. hybridum.The second group comprised T. nigrescans ssp.nigrescens and T. suffocatum and the third group is formed by the remaining 12 taxa of the section Lotoidea.These groupings of taxa also show that the members included in the subsections Loxospermum, Ochreata, Lotoidea, Platystylium and Calycospatha or that included in the series Lotoidea, Platystylium and Micrantheum by Zohary and Heller (1984) do not belong to these subsections or series (Table 1, Fig. 2).The similarity between the taxa belonging to this section ranged from 34.3% to 100%.Among these taxa, T. ruppellianum var.lianruppeelum and T. repens shows no difference respectively with T. ruppellianum var.lanceolatum and T. repens var.giganteum, rather a similarity of 100%.Trifolium semipilosum and T. semipilosum var.glabrescens presented the same similarity (100%).Trifolim michelianum differs from T. michelianum var.balansae with a similarity of 96.4%.These results show that the nine subsections and 13 series recognized in section Lotoidea by Zohary and Heller (1984) based on morphological characters should be reconsidered.Their view to consider this section as the most primitive group of the genus should be justified by its robust phylogeny.George and Hussein (2002) separated tribe Ononidea based on chromosome study, as well as the seed proteins analysis of 10 taxa of tribe Trifolieae.Badr (1995) illustrated that, on the basis of seed protein electrophoresis, section Lotoidea appears as a heterogenous group in which species relationship requires reconsideration.
The large cluster formed by 11 taxa following the section Lotoidea is segregated into three groups, one of which including T. quartinianum, T. spumosum, T. teudneri and T. xerocephalum is consistent with the section Mistyllus recognized by Zohary and Heller (1984).The unique structure of the symmetrically vesicular calyx and the persistent corolla, the manifestly bracteolate flowers and 2-4 seeded pod dehiscing suturally, sharply delimits this section from the others (Zohary and Heller, 1984).The other two groups that include the taxa of sections Vesicaria and Chronosemium and share the similarities between 53.8% and 76.2% do not completely conform to these sections, as recognized by Zohary and Heller (1984).The two taxa T. resupinatum and T. tomentosum belonging to the section Vesicaria group with the taxa of section Chronosemium (Table 1, Fig. 2) which is inconsistent with Zohary and Heller (1984).based on the SDS-PAGE of seed protein characters under non-reducing conditions; numbered as in Table 1.

Fig. 2 .
Fig. 2. Dendrogram illustrating the average taxonomic distance (dissimilarity) between the Trifolium taxa studied,based on the SDS-PAGE of seed protein characters under non-reducing conditions; numbered as in Table1.