Systematics Project Phylogenetic investigations of the African Restionaceae using rbcL data Carl Morrow Botany Honours 1994 Supervisor: Mr. Nigel Barker BOLUS L.IBRARY C24 0004 7771 11111111111 U II ABSTRACT DNA sequences for the rbcL gene were determined for four representatives of the African Restionaceae. These data were analysed along with six rbcL sequences (representing the rest of the Poales with Carex as the outgroup) from GenBank. Both parsimony analysis and distance measures were used. The topologies of the resultant trees were similar with the Restionaceae being a well supported group. Conflict in genus groupings, within the Restionaceae, was noticed when the trees from the morphological data set and the rbcL data set were compared. Reasons for this conflict are discussed. Times of divergence between different taxa were calculated and these compared closely to the ages published elsewhere. 1 ' INTRODUCTION The recent developments in the field of molecular biology have made it possible for systematists to collect phylogenetic information from a wide range of plant species. An example of this information gathering can be seen in Chase et al. (1993). Here, a collaboration of 42 scientists made it possible to create a data set containing the DNA sequences of the rbcL gene from 499 species of seed plants. The size of the data set meant that it was impossible to exhaustively analyse the information and so the result of the study was a first approximation of the phylogeny of the seed plants. At the end of the paper the authors encourage other workers to take up the challenge and perform more restricted rbcL based studies on smaller groups of plants so that, ultimately, a better approximation of the phylogeny can be obtained. The project presented here is performing this important function for the African Restionaceae. THE AFRICAN RESTIONACEAE The Restionaceae are a group of evergreen, rush-like plants with erect photosynthetic culms and leaves reduced to sheaths (Linder, 1984). All of the species are wind pollinated and, with a few exceptions, the flowers are dioecious. The flowers are small and aggregated into spikelets. The family is predominantly distributed in the southern hemisphere. In Africa, there are approximately 320 species of Restionaceae grouped into 19 genera (Linder, 1984; Linder pers. comm.) The African Restionaceae are primarily restricted to the South Western Cape where they can dominate the vegetation of a particular area (Linder, 1984; Linder, 199lb). There are about 100 species of the Restionaceae in Australia, 3 in New Zealand, 1 in Malaysia and South-east Asia and one in Chile (Linder, 1984). This distribution is shown Figure 1. The African Restionaceae are an old group of plants. The antiquity of the group was demonstrated by the discovery of restiod pollens 60 - 65 million years old (Linder, 1991b). Recently, the taxonomy of this group along with its allied families has received considerable attention, although knowledge about biological aspects of the family is poor (Linder, 199lb). It is generally accepted that the African Restionaceae are a monophyletic group within the family (Linder, 1986; Linder 199lb). The affinities of this family with related groups are still somewhat unclear. The Restionaceae have been placed in the order 2 \ . '\ '\ \ \ \ I ' ' ' ,\ . \ '\ I I I \ . \ : ·· .. I i c-.----- ·- Figure 1: The worldwide distribution of the Restionaceae. The number of genera and species in each area are indicated on the map with single numbers indication only one genus present. (from Linder, 1991h) the Restionales along with the Ecdeiocoleaceae, Anarthriaceae, Centrolepidaceae, Joinvilleaceae and Flagellariaceae. Dahlgren and Clifford (1982)(cited in Linder, 1991 b) contend that the order Restionales is the sister to the Poales and that these two groups should be combined. It is generally agreed that the Poaceae and the Restionaceae are related to one another but the precise positioning of the various sister groups is not clear (Linder, 199lb). Kellogg and Linder (in press) analysed the relationships within the Poales by combining data from 6 data sets (Morphology, cpDNA structure, rbcL, rpoC2, cpDNA restriction analysis and ribosomal RNA). Analysis of this combined data set revealed the following: the Centrolepidaceae may be derived from within the Restionaceae, the Restionaceae and Ecdeiocoleaceae are a monophyletic group and the position of the Anarthriaceae in relation to the other families has yet to be convincingly resolved (Figure 2). The complex of families comprising the Poales are quite distinct from the Cyperales. For this reason, Carex hostiana (a member of the Cyperales) has been used as the outgroup in the analyses that were performed in this study. The taxonomy of the Restionaceae has been difficult due to the shortage of convenient macro-morphological characters (Linder, 1984). A phylogeny of the Restionaceae was 3 Oryza e Nardus JPo Avena Aegl/ops Eragrostls a. Danthonla A Phragmltes A 93 Centotheceae c. Neurachne J Pennlsetum I'll 53 100 Andropogoninae Zea Bambusa e 80 Joinvllleaceae Hopklnsla/Lyglnla Elegia , Austral Restlolds 90 Centrolepidaceae Anarthriaceae Ecdelocoleaceae Flagellariaceae Typha Juncus Carex Figure 2: The strict consensus tree showing the phylogeny of the Poales based on the 6 data sets outlined above. (from Kellogg and Linder, in pr~) ascertained (Linder, 1984) and subsequent to this initial publication other characters were studied and added to the data set. An updated phylogeny was published (Linder, 1991•) and this is the phylogeny (Figure 3) that is used as the basis for the discussion of the results achieved in this study. AIM OF THIS PROJECT The existing phylogeny for the African Restionaceae (see Figure 3) is weak at nodes 2,3,5 and 7. This is due the fact that there are not many macro-morphological characters for the Restionaceae (Linder 1984) The aim of this project was to construct a phylogeny of the African Restionaceae using rbcL data. The resultant tree was then compared to the existing phylogeny. The data generated in this study is also a useful addition to the established rbcL sequence data base. USING MOLECULAR DATA IN SYSTEMATICS The use of nucleotide sequence data in systematics has received much attention. A key advantage of nucleotide sequences is the 4 distinct states (G, A, Tor C) that each character can hold. The separateness of the characters means that they can be coded simply and unambiguously. This simplicity can be problematic in that if two sequences are the same 4 E II :l a :l fl II E -c E I. :l II fl I. II II d :l 0 c fl II )i c .• m 0 :l I. II a a. : - :l I. a.= a. E 0 - 0 : m II -:l u 0 - - 0 - d 0 0 II 0 c :- : -II= II u •0 0 I. 0 a u (,) I. E c I. I. (,) -II c a. 0 'ti a c 0 0 = 0 0 m 0 -'ti fl m 'ti )i a. )i c .• E 'ti - I. fl c ,Q 0 fl fl 0 - : 0 c 'ti= : 0 )= ~ II c 0 ) 'ti fl II c a. I. a (,) a - fl : 0 II -c : : )i c c a )i G - II = = ~ - (.) UJ 0 ( Q, a:= (.) I- a: z I ( ~ (.) I (.) i 16 24 9 16 32 22 2 1 l Figure 3: The consensus cladogram for the African Restionaceae. The large numbers are labelling the nodes while the smaller numbers relate to the characters presented in the original paper. (from Linder, 1991") at a particular position one cannot be certain that they have always been the same. Reversals of state are undetectable. That is to say: a change from an A to a G and back to an A is undetectable. To alleviate this problem, one has to use long stretches of sequence so that one can sense the average change over time and a sufficiently conserved study molecule is chosen so that none of the codon positions are saturated. I 5 Other advantages of molecular data are as follows (Penny et al., 1990): Sequences evolve at different rates and so they can be specifically chosen to allow for the resolution of a particular historical event. Along with this, sequences often have a wide scope or domain meaning they occur in a wide range of organisms which makes it possible to analyse .the relationships between widely diverged lineages. With the correct choice of study molecule, one can be reasonably sure that phylogenetically useful characters will be obtained from the organisms in a particular study. Finally, Penny et al. (1990) state that the cost per character from a DNA sequence is relatively low and it is getting progressively cheaper. This makes sequence data an attractive option in phylogenetic studies but the limitations of this data must also be taken into account. There are numerous problems in weighting transitions (purine to purine or pyrimidine to pyrimidine changes), transversions (purine to pyrimidine or vice versa) and whether or not there is such a thing a neutral changes in the codons considering the possibility of transfer RNA (tRNA) biases (Clegg, 1993; Chase et al., 1993). West and Faith (1990) report that transversions are rarer than transitions and so, during certain analyses, the data should be coded as purines or pyrimidines without specifying the actual base in order to prevent dilution effects by the number of transitions. This is undesirable because one loses half the information content of the sequence one has and Penny et al. (1990) point out the importance of maintaining the maximum amount of information possible. Nucleotide sequences generate a large number of characters that can be used in the phylogeny reconstruction. The combination of the morphological (non-molecular) and molecular data has caused systematists to consider the informativeness of the different characters. Morphological characters generally encompass the product of a series of genetically controlled steps that ultimately results in the structure being examined.