The continuing search: low–copy nuclear sequences for lower–level plant molecular phylogenetic studies

Abstract

A recent survey by Alvarez & Wendel (2003) revealed that during a five-year period over 65% of all molecular phylogenetic studies of plants at the generic level or lower included sequences of the internal transcribed spacer region of nuclear ribosomal DNA (ITS), and more than one third of the studies used this region exclusively for reconstructing phylogeny. Why the immense popularity of ITS among plant systematists? ITS sequences have achieved popularity for the same two reasons that any method or approach becomes popular in plant systematics: it is easy to generate data and the data seem to work. There are universal primers (with rare exceptions) for the ITS region (Baldwin, 1992), and therefore, it is easy for systematists to generate data rapidly because they do not have to design primers. Also, because there are many copies of the ITS repeat per genome, it is relatively easy to amplify the region. The ITS region has worked in the sense that it has proven sufficiently variable to resolve relationships in many different plant groups at the generic level or lower, which are the levels at which many plant systematists conduct research. Several studies have also demonstrated the value of ITS sequences at higher taxonomic levels (Hershkovitz & Lewis, 1996; Hershkovitz & Zimmer, 1996; Goertzen & al., 2003). If the ITS region has served plant systematists so well, why even look for another region to sequence? It is beyond the scope of this column to discuss the potential limitations of the ITS region, and we refer readers to the insightful and comprehensive critique of these limitations provided by Alvarez & Wendel (2003). We hasten to add, however, that while there certainly are limitations in using ITS sequences for phylogenetic reconstruction, it remains a genomic region that has proven and likely will continue to prove immensely useful to plant systematists. After all, without the insights that sequences of the ITS region have provided, we would be limited in our ability to judge the utility of other nuclear sequences that we will discuss here. For example, Tank & Sang (2001) discuss the importance of having a well-established phylogeny for Paeonia (based on morphology and molecular data, including the ITS region) as a framework for judging the utility of their sequences of GPAT (Table 1). On the other hand, it is becoming increasingly clear that some of the advantages of the ITS region may also make this region problematic. For example, universal primers may facilitate PCR amplification of contaminants. Although most of the copies of ITS per genome are identical, there may be rare copies in an individual that are not amplified due to selective amplification of only one copy. Concerted evolution often homogenizes copy types, and together, selective amplification and concerted evolution could obscure processes (e.g., hybridization) in the history of an organism. Another compelling reason to search for alternative low-copy nuclear sequences is that, despite the utility of ITS sequences for resolving relationships within genera, there are many examples, too numerous to cite, where resolution is poor because of the relatively short sequence length of the ITS region and the lack of sufficient variation. Thus, including additional nuclear gene sequence data has the potential to provide a more robust estimate of phylogenetic relationships. Another motivation for seeking additional nuclear sequences is that any inferred organismal phylogeny becomes much more compelling when it is based on more than one region because it becomes more than just a single gene tree (Cronn & al., 2003). However, few nuclear genes exist as single copies in most plant species, which complicates generating DNA sequence data from the nuclear genome. Generating sequence data from multiple gene regions in the chloroplast is easily accomplished because many different regions have been well characterized (e.g., matK, trnL-trnF, psbA-trnH, atpB) and most of these regions have universal primers available for amplification and sequencing. Employing sequences of multiple cpDNA gene regions often helps to resolve relationships with more support, but since the chloroplast genome is generally inherited as a single unit and lacks recombination, phylogenies based on multiple cpDNA regions effectively represent a single gene tree (Doyle, 1992). Given that it is desirable to include low-copy nuclear