A project undertaken at the Royal Botanic Gardens and Domain Trust, Sydney,
and supervised by Dr Peter Wilson. Research associate: Ms Margaret Heslewood
The genus Xanthostemon (Myrtaceae) comprises at least 48 species, occurring in Australia, Papua New Guinea, Solomon Islands, Indonesia, Philippines and New Caledonia. Xanthostemon is an almost exclusively tropical genus, with only one of the Australian species, X, oppositifolius, occurring in subtropical areas. Many species have spectacular flowers, and some are grown as ornamentals. An Australian species, X. chrysanthus, the golden penda, and a red-flowered Philippine species, X. verdugonianus, are the most commonly cultivated species, in Australia and Asia respectively.
The genus is a member of the tribe Xanthostemoneae along with two small, segregates that have been recognised in New Caledonia: Pleurocalyptus and Purpureostemon. New Caledonia is a centre of diversification in the tribe with at least 19 species of Xanthostemon occurring there, but Australia has species representing four quite diverse groups based on fruit and placentation structure. Three of the latter groups were considered to be possible generic segregates and an earlier study, using traditional sequencing approaches, was undertaken on the Xanthostemoneae to test both the published and the hypothetical genera. This study was hampered by paralogy due to apparent duplication of a portion of the chloroplast (CP) genome in some Australian Xanthostemon species.
Figure 1. Xanthostemon verticillatus – North Queensland species with white flowers and verticillate leaves (Photo: P.G. Wilson); X. petiolatus – a white-flowered species with opposite leaves from Indonesia (Photo: W. Mustaquim)
Aims
This project set out to identify the source of the paralogy. The two loci we had sequenced happened to be close to one boundary of the Inverted Repeat (IR), and our hypothesis was that the paralogous copy might be part of the IR. Our intention was to test this hypothesis using a Next Generation Sequencing (NGS) approach, identify the paralogue(s) and design new primers to refine our existing 55-taxon dataset with a view to resolving species relationships and clarifying the status of the three named genera in the tribe.
A second aim was to test morphotypes in two broadly-defined species, Xanthostemon umbrosus and X. paradoxus, to determine if segregate taxa would be supported. Field collections were made from populations of both species in the Northern Territory to adequately sample the geographic variation.
Figure 2. Xanthostemon youngii – a red-flowered species from Cape York (© Centre for Australian National Biodiversity Research, 1996); X. verdugonianus – a red-flowered species from the Philippines, cultivated, Singapore Botanic Gardens (Photo: P.G. Wilson)
Findings
The NGS component was problematic. We had trouble extracting DNA of high quality and concentration regardless of methods trialled, and even outsourcing the process still returned disappointingly weak DNA of modest quality. We eventually achieved some success, but only recently and the data is yet to be analysed.
In light of these problems, we initiated our alternative approach to the paralogy issue by sequencing two other regions of the plastid genome, both distant from the IR boundary. However, even though sequences from these extra regions seemed to be free of the paralogy that plagued the loci originally chosen for this group, variability was very low. It became clear that individual plastid markers were very unlikely to generate enough phylogenetic signal to produce a robust phylogeny in this group. This appears to lessen the potential value of complete chloroplast genomes to resolve the questions posed at the outset of this study, but that work will be continued to determine the utility of this type of full plastid data for phylogenetic reconstruction.
As well as the new plastid loci, we sequenced two nuclear regions: ITS1 & 2 and ETS, for our additional samples to complement the existing dataset. We now have a 101-taxon dataset for nuclear loci. Analysis of all current data has allowed us to reach some definite conclusions: 1) we found no clear evidence from our data to support segregation of the observed morphotypes within Xanthostemon paradoxus; 2) we did find support for a clade comprising samples of small-fruited, shrubby plants of X. umbrosus from sandstone areas of Arnhem Land. We also found support for a grouping of samples from the Iron Range area of Cape York, from populations of trees with larger fruits; and 3) at the genus level, our analyses supported relationships consistent with observed morphotypes. Samples from specimens referred to the broadly defined Xanthostemon umbrosus are grouped with other species with white flowers and opposite leaves (_X. verticillatus_and X. xerophilus from Australia and X. petiolatus and X. verus from Indonesia – Fig. 1). Secondly, species pairs that share distinctive flower and fruit characters (X. eucalyptoides + X. crenulatus and X. psidioides + X. graniticus) are confirmed as sister taxa. Thirdly, and somewhat surprisingly, the non-New Caledonian red-flowered species from the Philippines (X. verdugonianus), Solomon Islands (X. melanoxylon) and Australia (X. youngii) are supported as a monophyletic group (Fig. 2). On the other hand, our molecular data do not support recognition of Pleurocalyptus as a separate genus; two sampled taxa are embedded within the New Caledonian Xanthostemon clade, and, at this stage, our data do not allow us to make a definitive statement regarding the status of the other segregate genus, Purpureostemon.