Phylogenetic Position and taxonomic disPosition of turraea breviflora ( meliaceae ) , a hitherto enigmatic sPecies

For decades, the rare Malesian shrublet hitherto called Turraea breviflora, was the only species of the pantropical family Meliaceae which could not confidently be placed in a genus. Previous morphological re-investigation led to the exclusion of the species from Turraea and suggested possibly close links with Munronia. In this study, parsimony, maximum likelihood and Bayesian analyses of DNA sequence data from plastid rbcL and nuclear ITS rDNA were used to infer the phylogenetic position of T. breviflora and its affinities to other representatives of the family. Phylogenetic analyses support Mabberley’s (1995a) proposal that the concept of the genus Munronia be expanded to accommodate it. The new combination Munronia breviflora (ridl.) Mabb. & Muellner is therefore made.


INTroDUcTIoN
Turraea breviflora ridl.(Meliaceae; Fig. 1) is a rather inconspicuous suckering shrublet now restricted to Peninsular Malaysia.It seems that it was first collected by H.N. ridley at Serangoon, Singapore in December 1897, but the plant has not been recorded from that island since; subsequently it was collected in Selangor and Johor in what is now Peninsular Malaysia.
When ridley described the plant as a new species in 1922, he provisionally referred it to Turraea L., a genus otherwise unknown from the peninsula, or, indeed from the wetter parts of Malesia in general (Mabberley 1995a, b).Since that time it has been collected on only a handful of other occasions and is still known from only a few localities in rather open hill Dipterocarp forest, often associated with bertam, Eugeissona tristis, and from limestone rocks.Paradoxically, it is known from some of the most well-worked sites such as Bukit Lagong and Ulu Gombak in Selangor, yet has never been found in fruit, and the number of flowering collections even from those sites is small.No doubt it is frequently overlooked, as it grows amongst tree seedlings and saplings which it greatly resembles.It is like no other Meliaceae and, in the field, sterile material may be readily confused with species of Icacinaceae or other families.Drawing by rosemary Wise, reproduced with permission from Flora Malesiana Series I, Vol. 12 / part 1 (1995) 28, f. 2. Pennington & Styles (1975) noted that the relationships of the species were uncertain, and suggested it had "almost equal affinities with both Turraea and Munronia".In the early 1980s, when preparing an account for Flora Malesiana Mabberley (1995a: 29), who had collected the plant in the field in 1974, examined herbarium material closely and made a critical morphological analysis, leading him to exclude Turraea breviflora from genus Turraea because of its habit, indumentum, corolla aestivation, disk and the usual number of ovules in each locule.He pointed out that it resembled the genus Munronia Wight, as then understood, in its indumentum, but differed from that in its habit, valvate corolla, the usual number of ovules in each locule and possible dioecy.He thus concluded that it either represented an undescribed genus as ridley had suspected, or that the concept of Munronia had to be enlarged.Today, it is the only species of the pantropical family Meliaceae which cannot confidently be placed in a genus.The writing by the second author of an account of the family Meliaceae for Flora of Peninsular Malaysia has prompted a re-examination of the plant, using molecular methods to augment the earlier morphological analysis, in an attempt to resolve its taxonomic placement.
In this study, therefore, we performed maximum parsimony (MP), maximum likelihood (ML) and Bayesian analyses of DNA sequence data from plastid rbcL and the internal transcribed spacers (ITS) of nuclear ribosomal DNA (nrDNA), defined as the unit containing the ITS1 spacer, 5.8S rrNA gene, and ITS2 spacer, to infer the phylogenetic position of Turraea breviflora and its affinities to other representatives of the family.Based on 46 species of Meliaceae, including a representative sampling for the two subfamilies, all tribes and almost all genera (44 out of c. 50), we focused on the following questions: -What is the topological position of T. breviflora in the family phylogenetic trees derived from MP, ML and Bayesian analyses?-What are the closest relatives of T. breviflora?-Do the results support either of Mabberley's (1995a) taxonomic proposals?

Sequence editing and alignment
Editing and assembly of the complementary strands were carried out with Seq-Man TM II version 5.07 (Lasergene, DNASTAr, Inc., Madison, WI, USA).Alignment of rbcL sequences was performed by eye.Alignment of ITS sequences was performed following Muellner et al. (2008a, b).New sequences have been deposited in GenBank under the accession numbers EU621669-621670, FJ194495/FJ194496 and FJ194497 (http://www.ncbi.nlm.nih.gov/).

Phylogenetic analysis
MP analyses of the rbcL dataset were performed using PAUP* 4.0b10 (Swofford 2002).Substitutions at each nucleotide position were treated as independent, unordered, multi-state characters of equal weight (Fitch parsimony;Fitch 1971).Heuristic searches were carried out using 1000 random additions of taxa, tree bisection-reconnection (TBr) branch swapping, and the option MulTrees (keeping multiple, shortest trees), but holding only ten trees per replicate to reduce time spent in swapping on large numbers of trees.After 1000 replicates, we then used the shortest trees found as starting trees for a swapping-to-completion search (but with a tree limit of 10 000).robustness of clades was estimated by bootstrapping (Felsenstein 1985) with 1000 replicates, using simple sequence addition, TBr branch swapping, and MulTrees, again holding ten trees per replicate.We consider 75-84% bootstrap values moderate support and 85-100% strong support.MP analyses of the ITS dataset and the combined rbcL/ITS dataset were conducted as described above, except for setting no constraint on the number of trees held per replicate during heuristic searches.
ML analyses were carried out with rAxML version 2.2.1 (Stamatakis 2006; http:// icwww.epfl.ch/~stamatak/index-Dateien/Page443.htm) and PAUP* 4.0b10 (Swofford 2002), and Bayesian analyses were performed with MrBayes version 3.1.2(ronquist & Huelsenbeck 2003; http://mrbayes.csit.fsu.edu/).The substitution models employed in these analyses were found using Modeltest version 3.06 (Posada & crandall 1998 shape parameter alpha to model rate heterogeneity (GTr + I + G).For the Bayesian analyses, model parameters were estimated directly during two parallel runs, using four simultaneous chains and 2 000 000 cycles, sampling one tree every 100 generations.Trees that preceded stabilization of the likelihood value were excluded, and the remaining ones were used to calculate posterior probabilities via the construction of a majority rule consensus tree in PAUP.For the ML searches with rAxML we employed the GTr + G model, using 25 rate categories (instead of four as used in the Bayesian analyses). rESULTS The aligned rbcL matrix consisted of 1387 characters.For the rbcL matrix, 277 (20%) positions were variable and 166 (12%) were potentially parsimony informative.With the limits imposed as described above, the parsimony search produced 10 000 most parsimonious trees of 552 steps with a consistency index (cI) = 0.53 and a retention index (rI) = 0.81.The aligned ITS matrix consisted of 785 characters; 480 (61%) positions were variable and 366 (47%) were potentially parsimony informative.The parsimony search produced 10 most parsimonious trees of 2072 steps with cI = 0.41 and rI = 0.46.The combined rbcL/ITS matrix consisted of 2172 characters; 671 (31%) positions were variable and 464 (21%) were potentially parsimony informative.The parsimony search produced 2 most parsimonious trees of 2453 steps with cI = 0.41 and rI = 0.50.
Figures 2 and 3 show results of the ML and Bayesian analyses.Turraea breviflora is phylogenetically closest to the two representatives of the genus Munronia (Fig. 2, 3; also true for the single rbcL and ITS ML and Bayesian trees, trees not shown).Accessions of T. breviflora and two species of Munronia appear in a clade (Fig. 2, 3).The monophyly of the group is very weakly supported in the Bayesian analysis of the rbcL dataset (59% posterior probability, PP, tree not shown), and receives strong support in the combined rbcL/ITS analysis (100% PP, Fig. 3).Although T. breviflora and the two species of Munronia are monophyletic in the combined MP strict consensus tree, the clade does not receive > 50% bootstrap percentage; the latter is also true for the single rbcL and ITS analyses (trees not shown).In all three ML analyses, T. breviflora and Munronia appear in one clade (Fig. 2, and trees not shown).
DIScUSSIoN Mabberley (1995a) argued, on morphological grounds, that, although Turraea breviflora belongs in Turraeeae, it should be excluded from Turraea and is most closely related to Munronia; this study 1) confirms the position of T. breviflora in a clade comprising representatives of Turraeeae and Trichilieae; 2) provides evidence that it is phylogenetically closest to Munronia; 3) suggests it indeed be referred to Munronia.
It should be noted here, however, that in order to reach a robust phylogenetic appreciation of Turraeeae and Trichilieae, sampling of additional taxa on species level and the collection of additional DNA data will be necessary.Mabberley's (1995a) proposal that the concept of the genus Munronia be expanded, is vindicated by our new findings, so that an amended description of the genus is provided below.
conservation status -As Munronia breviflora has not been collected in its southern sites (Johor, Singapore) for over 70 years and has never been found in fruit anywhere, while its extant populations are very restricted in Selangor, it must be considered very rare and threatened, though Bukit Lagong populations are inside a forest reserve.

Table 1 .
Voucher information and GenBank accession numbers for samples used in this study (in alphabetical order).Voucher specimens are deposited in the following herbaria: ANBG = Australian National Botanic Gardens; Fr = Herbarium Senckenbergianum, Senckenberg research Institute; Gc = University of Ghana; KEP FrI = Forestry Institute Malaysia; K = royal Botanic Gardens, Kew; L = Nationaal Herbarium Nederland, Leiden University branch; LBc = Limbe Botanic Garden; NcU = University of North carolina; WU = University of Vienna.