«Mistletoe phylogenetics: Current relationships gained from analysis of DNA sequences Daniel L. Nickrent Department of Plant Biology and Center for ...»
Mistletoe phylogenetics: Current relationships
gained from analysis of DNA sequences
Daniel L. Nickrent
Department of Plant Biology and Center for Systematic Biology
Southern Illinois University, Carbondale,IL 62901-6509
Western International Forest Disease Work Conference
August 14-18, Kona, Hawaiii
Citation: Nickrent, D. L. 2002. Mistletoe phylogenetics: Current relationships gained
from analysis of DNA sequences. Pp. 48-57 in: Proceedings of the Western International
Forest Disease Work Conference, August 14-18, 2000. Waikoloa, Hawai’i. 253 pp.
Introduction The purpose of this paper to briefly summarize some of the advancements made in our understanding of evolutionary relationships of mistletoes. It is no exaggeration to say that the past decade has witnessed a virtual revolution in phylogenetic investigation, owing mainly to the application of molecular methodologies and advancements in data analysis techniques. These powerful approaches have provided a source of data, independent of morphology, that have been used to address questions in green plant evolution, including parasitic plants like mistletoes.
How Many Mistletoes?
The term mistletoe is not strictly a taxonomic term but refers to shrubby branch parasites that occur in four families of the sandalwood order (Santalales). Of the total 90 genera and 1306 species of mistletoes (Table 1), Loranthaceae has the majority of genera (83%) and species (69%). Another 26% of the total mistletoe species are found in Viscaceae, mainly because of the speciose genera Phoradendron and Viscum.
How Many Times Did Mistletoes Evolve?
As indicated in Table 1, molecular data have been instrumental in shaping our understanding of phylogenetic relationships in Santalales. At present, complete nuclear small-subunit (18S) rDNA and chloroplast rbcL sequences have been obtained from representatives of all mistletoe groups. Analyses of these genes separately and in combination give the same general topology for the order (Fig. 1). Olacaceae are basalmost and, although paraphyletic, this assemblage is sister to the remaining members of Santalales. This position is in agreement with traditional concepts and fits with the observation that Olacaceae is the only family in the order with parasitic and nonparasitic members. The two-gene tree (Fig. 1, no. 1) also shows that Schoepfia is not closely related to other Olacaceae but is sister to Misodendrum, a mistletoe that parasitizing southern hemisphere beech trees (Nothofagus) in Chile and Argentina. Indeed, if this 1 phylogeny is correct, Misodendrum may represent the oldest genus that evolved the mistletoe habit.
As can be deduced from the topology of the tree in Figure 1, the common ancestor of Misodendrum and Loranthaceae may have been an aerial parasite. Two factors argue against this interpretation. First, Misodendron is sister to Schoepfia (a root parasite), not the Loranthaceae clade. Second, three genera of Loranthaceae are root parasites (Atkinsonia, Gaiadendron, and Nuytsia), and it is generally thought that these plants are primitive in the family. At present sequences for both genes from all three genera are not available to test this hypothesis, however, it seems unlikely that “reversions” to root parasitism would have occurred in these three genera after the evolution of aerial parasitism. Once aerial parasitism had evolved in Loranthaceae, it appears that this habit was selectively advantageous given the explosive adaptive radiation of genera and species that subsequently ensued in the southern hemisphere. Therefore, the molecular results obtained to date suggest an independent evolution of the mistletoe habit in Loranthaceae (Fig. 1, no. 2) and Misodendraceae.
Within the sandalwood order, the next occurrences of the mistletoe habit are in Santalaceae. The number of times this evolutionary event happened cannot be stated precisely at present because molecular data are incomplete for all relevant genera. One possible case of independent evolution of aerial parasitism involves the three neotropical genera traditionally classified in Eremolepidaceae (Fig. 1, no. 3). Both 18S rDNA and rbcL sequences for all three genera have been obtained and upon analysis they emerge within a clade composed of Exocarpos and Omphacomeria (tribe Anthoboleae of Santalaceae). The three genera are not monophyletic, hence there is no support from the molecules for maintaining a separate family Eremolepidaceae. Indeed the entire “family” Santalaceae represents a series of grades which will require further work to arrive at an evolutionarily sound circumscription. Another possibly unique evolution of the mistletoe habit occurred within tribe Amphorogyneae and involves four genera: Dendromyza (including Cladomyza), Dendrotrophe, Dufrenoya (including Hylomyza), and Phacellaria (Fig. 1, no. 4). Recent work by Jill Macklin (unpublished, pers. com.) using nuclear ITS rDNA sequences indicates these aerial parasites are monophyletic (evolved once from a common ancestor) and that the basalmost genus is Dendrotrophe. Interestingly, the genus Phacellaria is an obligate parasite on other mistletoes (usually Loranthaceae). At present, sequences for rbcL are lacking for the two genera shown in Fig. 1 (Dendrotrophe and Dufrenoya), hence their position on this tree was determined from the 18S rDNA sequence alone. Efforts are being made to include a full complement of sequences for these and other Amphorogyneae because resolving their position in Santalaceae may help shed light upon the evolutionary origin of Viscaceae.
Emerging from the series of Santalaceae grades is a strongly-supported clade that represents the family Viscaceae (Fig. 1, no. 5). This group of seven genera is clearly monophyletic and support (using bootstrap resampling techniques) is high (100%) using 18S rDNA, rbcL or combinations of the two molecules. As previously demonstrated empirically (Nickrent and Starr 1994), rates of molecular evolution in Viscaceae are higher than other families in the order. Surprisingly, despite these high rates, 2 relationships within the family are remain unresolved. Phoradendron is clearly sister to Dendrophthora and Korthasella sister to Ginalloa, however these are the only wellsupported clades within the family. It is likely that more sequences from genes with fast evolutionary rates are required to tease apart the intergeneric relationships in Viscaceae.
Infrageneric Relationships in Viscaceae
At present, all mistletoe infrageneric molecular phylogenetic studies have been in Viscaceae. Molecular phylogenetic analyses have been published on three of the seven genera: Arceuthobium (Nickrent et al. 1994), Korthalsella (Molvray et al. 1999), and Phoradendron (Ashworth 2000). The following gives a brief summary of the results from these studies as well as preliminary information on Viscum.
Arceuthobium. Interspecific relationships in dwarf mistletoes have been addressed using a variety of biochemical methodologies beginning with flavonoids (Crawford and Hawksworth 1979) and continuing with isozymes (Nickrent 1986) and DNA sequencing (Nickrent et al. 1994). The taxonomic and systematic results from the latter two studies were summarized in Nickrent (1996). The original ITS study of Arceuthobium included sequences from 22 taxa, however, since that time sequences from an additional six New World taxa have been obtained. Given that these sequences are from Central American taxa, their relationships are discussed in Mathiasen et al. (2000) and will not be repeated here. In general, the topology of the dwarf mistletoe ITS tree is fully congruent with the one published in 1994, but greater resolution has been obtained by adding other taxa. If section Pusilla were reduced to a Series, the New World dwarf mistletoes would comprise five Sections: Arceuthobium, Americana, Penda, Campylopoda, and Vaginata.
The latter section would then be divided into fives Series: Globosa, Minuta, Pusilla, Rubra, and Vaginata. Nearly complete taxon sampling has been achieved for New World Arceuthobium, hence it is likely that the tree topology will remain stable.
Sequences from seven members of Section Campylopoda are known and all are very similar or identical to those of the four taxa represented on the tree. Therefore, it is unlikely that additional sequences from this Section will alter the results. Only two
Mexican taxa remain to be added to achieve complete sampling for the New World taxa:
A. oaxacanum (Series Rubra) and A. yecorense (Series Stricta). Continuing phylogenetic work with Arceuthobium involves obtaining sequences of the remaining Old World taxa (A. juniperi-procerae, A. azoricum, A. chinense, A. minutissimum, A. pini, A.
sichuanense, and A. tibetense). Attempts are being made to extract genomic DNA from herbarium specimens with some success.
Korthalsella. The study by Molvray et al. (1999) used nuclear ITS rDNA and chloroplast trnL-F spacer sequences from 25 populations of Korthasella collected from across its range. Parsimony analyses of each data set separately and in combination were highly congruent. A species from Queensland Australia (K. papuana) was sister to the remaining taxa and these taxa were further resolved as two subclades with either differentiated or undifferentiated inflorescence branches. These results indicate that a classification based upon morphology (Danser 1937) is not supported by genetic data.
Moreover, plants on different hosts that are genetically closely related can have markedly 3 different morphologies (internode shapes). This prompted the authors to propose host influence on the morphology of the parasite, an issue visited by workers looking at other viscaceous genera such as Arceuthobium (Gill 1935) and Viscum (Meyer von Freyhold 1987).
Phoradendron. The genus Phoradendron, comprising at least 100 species of New World mistletoes, is closely related to Dendrophthora. Indeed, a single morphological character defines the two genera: one anther locule for Dendrophthora and two for Phoradendron.
The monophyly of these genera has been questioned based upon molecular evidence (Nickrent and Duff 1996). More recently, a detailed molecular phylogenetic analysis of these two genera was conducted using nuclear ITS and 26S rDNA sequences (Ashworth 2000). One taxon, D. guatemalensis, was tentatively included in this genus despite the absence of male flowers. More recently (Kuijt, pers. com.), bilocular anthers have been observed in this taxon, hence it will be transferred to Phoradendron (P. navicularis (Standley) Kuijt). Thus, five Dendrophthora and 35 Phoradendron taxa were analyzed with parsimony using santalaceous genera as outgroups. Three major clades were identified: clade A, a morphologically heterogeneous one containing all five Dendrophthora species plus P. crassifolium, P. piperoides, and P. sulfuratum; clade B, containing seven Phoradendron species typically with biseriate inflorescences and one pair of basal cataphylls; and clade C containing the remaining 25 Phoradendron taxa (with the exception of P. californicum) that have biseriate or triseriate inflorescences and that generally lack basal cataphylls. As with Arceuthobium and Korthalsella, this study has demonstrated how morphological characters can be unreliable indicators of phylogenetic relationships. Although all five Dendrophthora species were resolved in clade A, this clade also contained three Phoradendron species, supporting the previous suggestion that neither genus is monophyletic.
Viscum. The utility of ITS sequences for resolving phylogenetic relationships in Arceuthobium, Korthalsella and Phoradendron prompted an investigation of its utility in the more speciose Old World genus Viscum (Nickrent, unpublished). ITS sequences for five species of Viscum were generated: V. album, V. cruciatum (both Europe), V. capense, V. continuum (both South Africa), and V. triflorum (Kenya).
Substitutional and insertion/deletion mutations were quite high, even in relatively conserved regions such as the 5.8S rDNA (1% of the sites polymorphic). Indels in ITS-1 and -2 prevented unambiguous alignment of the South African and European species. For this reason, variable domains of 26S rDNA were examined (portions of D2, D7 and D8 resulting in ca. 1 kb of sequence). In contrast to ITS, the 26S rDNA alignment was easily accomplished with essentially no ambiguous regions. Using the same species as above, ca. 100 potentially phylogenetically informative sites were obtained. A collaboration with V. Ashworth is currently underway to further explore 26S rDNA sequencing in Viscum.
Dwarf Mistletoe Coevolution with Their Hosts?
The incredible proliferation of molecular phylogenetic studies may prompt one to ask “how can these results be used?” One interesting topic that can be explored when working with parasitic organisms is whether the evolutionary histories of the two sets of 4 organisms are related in any way. For example, it has been observed among animals that host speciation and divergence is accompanied by speciation of its parasites. This process has been called the Fahrenholtz Rule (Nobel and Nobel 1976) and was defined by Price (1980) as: “common ancestors of present-day parasites were themselves parasites of the common ancestors of present day hosts.” Does the Fahrenholtz Rule apply to parasitic angiosperms? A requirement for answering this question is the presence of relatively detailed phylogenies for and sufficient numbers of both the host and the parasite species. Moreover, there must exist some degree of host specificity because parasites exhibiting very broad host ranges are eliminated a priori.
Dwarf mistletoes (Arceuthobium) are good candidates for determining whether host tracking is taking place because numerous host and parasite combinations can be seen in nature. Several molecular phylogenies of the genus Pinus have been published (Krupkin et al. 1996, Liston et al. 1999, Wang et al. 1999), however, none of these approached complete sampling of the 100 or so species in the genus. Sampling in the Wang et al.
(1999) paper was aimed at Eurasian pines whereas Liston et al. (1999) utilized all recognized subsections of the genus. When only species common to both studies are considered, the trees are congruent with respect to major clades, thus genes from the chloroplast and nucleus appear to be tracking the same phylogenetic signal.
Because of its broad subsectional sampling, the nuclear ribosomal ITS tree of Liston et al.