PCR Applications in Fungal 7 Phylogeny S. Takamatsu Laboratory of Plant Pathology, Faculty of Bioresources, Mie University, Tsu 514, Japan 7.1 Introduction Historically, the phylogeny of fungi has been inferred from various phenotypic characters, e.g. morphological, physiological and developmental characters, and/or chemical components such as secondary metabolites. The recent development of molecular techniques has enabled the derivation of fungal phylogenies based on the analyses of proteins or nucleic acids, e.g. isozyme analysis, DNA–DNA hybridization, electrophoretic karyo- typing, restriction fragment length polymorphism (RFLP), and DNA sequencing. However, these molecular techniques have been difficult to apply to obligate parasites from which only small amounts of fungal materials are available, or to rare taxa which are available only as herbarium specimens. The polymerase chain reaction (PCR) developed in the mid 1980s (Saiki et al., 1985; Mullis et al., 1986; Mullis and Faloona, 1987) made possible the amplification of particular nucleotide sequences from very small amounts of starting materials. In addition, relatively impure DNA extracts can be used for PCR. PCR-based techniques have enabled the construction of molecular phylogenies for obligate parasites and rare taxa, and as PCR techniques have led to simplified molecular procedures, they have also increased the availability of phylogenetic studies for many fungi. This chapter reviews some of the recent advances that have been made through the application of PCR methods to phylogenetic studies of fungi. © CAB INTERNATIONAL 1998. Applications of PCR in Mycology (eds P.D. Bridge, D.K. Arora, C.A. Reddy and R.P. Elander) 125 126 S. Takamatsu 7.2 Sequence Analysis from Small Samples 7.2.1 DNA extraction A number of methods have been developed for the rapid isolation of fungal DNA (Biel and Parrish, 1986; Zolan and Pukkila, 1986; Taylor and Natvig, 1987; Lee et al., 1988; Moller ¨ et al., 1992). Most of these are applicable to molecular phylogenetic analysis when large amounts of fungal materials can be obtained. However, a small amount of starting material is required for PCR-based sequence analysis, and sufficient single-copy genomic sequences can be routinely amplified from 1 µg of total DNA (Saiki et al., 1988). Lee and Taylor (1990) simplified an earlier method for total DNA isolation and used it for the amplification of multicopy genes from small quantities of fungal materials (Lee et al., 1988). They reported that rDNA sequences could be successfully amplified from a single spore of Neurospora tetrasperma. In molecular phylogenetic analysis, sequence data are often required from herbarium collections, especially when they concern rare taxa. Bruns et al. (1990) applied Lee and Taylor’s method to 35 collections of dried fungal materials stored under a variety of conditions for 1–50 years. They successfully amplified the mitochondrial large subunit rRNA gene from all of the specimens, and 16 specimens were subsequently sequenced. Walsh et al. (1991) developed a method utilizing a chelating resin for extracting DNA from forensic-type samples for PCR. The method was simple, rapid, did not include organic solvents and did not require multiple tube transfers. This method has proved applicable to the extraction of DNA from small amounts of fungal materials, and Hirata and Takamatsu (1996) have used the method to obtain the rDNA internal transcribed spacer (ITS) regions from several obligately parasitic powdery mildew fungi. 7.2.2 PCR amplification A variety of oligonucleotide primers have been designed for the amplification of rDNA sequences from a wide range of fungal materials (e.g. White et al., 1990). Amplification using these primer sets, followed by direct sequencing has considerably shortened the time required to perform phylogenetic studies. In addition to universal primers, many kinds of primer sets for the amplification of rDNA have been developed for particular groups of fungi; two examples of this are lichenized fungi (Gargas and Taylor, 1992) and slime moulds (Rusk et al., 1995). In most cases simple thermal cycling parameters have proved sufficient to amplify rDNA sequences; however, two sequential reactions using nested primers can improve the efficiency of amplification from small quantities of DNA (Hirata and Takamatsu, 1996). Applications in Fungal Phylogeny 127 7.2.3 Sequence analysis Two methods are in general use for sequencing PCR-amplified DNA fragments. One procedure involves cloning. Although cloning procedures have become more straightforward, considerable effort is still required especially when obtaining sequence data from large numbers of individuals. Another disadvantage of the cloning methods is that mutations can be introduced by the DNA polymerase during synthesis. Taq DNA polymerase lacks editing functions and incorporates an incorrect nucleotide at a rate of 2 × 10–4 nucleotides per PCR cycle. This rate of misincorporation translates into an overall error frequency of one per 400 bp after 30 cycles of amplification (Saiki et al., 1988). However, conditions can be optimized to reduce this error frequency to less than one substituion in 15,000 bp (Gelfand and White, 1990). As a result a consensus sequence, based on several cloned PCR products, has to be determined for each sample in order to distinguish mutations present in the original sequence from any random misincorpora- tion of nucleotides introduced by the polymerase during PCR. The second procedure involves direct sequencing of PCR product without cloning. This generates a consensus sequence at each nucleotide and minimizes possible errors from misincorporation. Since the procedure does not involve the cloning of the PCR product, the time required for comparative sequencing studies is significantly reduced. As a result, direct sequencing is being widely used for phylogenetic studies based on DNA sequences (Lott et al., 1993; Morales et al., 1993; Zambino and Szabo, 1993; Saenz et al., 1994; Nishida and Sugiyama, 1994). 7.2.4 Construction of phylogenetic trees Many computer packages have been developed for phylogenetic studies, and six of the more widely used are described here. CLUSTAL W The program CLUSTAL W (Thompson et al., 1994) has a multiple sequence alignment function that works well with large numbers of either nucleic acid or protein sequences. The program also constructs phylogenetic trees by the neighbour-joining method of Saitou and Nei (1980), but does not draw the tree. Confidence intervals on the tree are calculated using a bootstrap procedure similar to that described by Felsenstein (1985). MACCLADE MACCLADE (Maddison and Maddison, 1992) is a program for analysing character evolution. It has a graphical spreadsheet editor for entering and editing data matrices, a tree window for viewing and manipulating trees, charting facilities, and other features. Although MACCLADE does not have a sophisticated search algorithm to find the best tree, it can be complemented by PAUP (see below) as the two packages are compatible. 128 S. Takamatsu MEGA MEGA (molecular evolutionary genetics analysis; Kumar et al., 1993) is a phylogeny inference package that utilizes various distance methods and parsimony. Various other statistics are included for the analysis of DNA, RNA and protein sequence data. PAU P PAUP version 3 (Swofford, 1993) is probably the most sophisticated program for inferring phylogenies using parsimony analysis, with many options and good compatibility with MACCLADE. Version 3 itself is currently not available, but the new program, PAUP version 4, will be released in the summer of 1998. Version 4 will include parsimony, distance matrix, invar- iants, and maximum likelihood methods. PHYLIP PHYLIP (Felsenstein, 1989) includes programs for parsimony, distance matrix, and maximum likelihood methods for a variety of data types, including molecular sequences (DNA and protein), gene frequencies, continuous characters, discrete (0, 1) characters and restriction sites. It also includes programs for plotting phylogenies and consensus trees. PUZZLE PUZZLE is a package that uses maximum likelihood to reconstruct phyloge- netic trees from sequence data. It implements a fast tree search algorithm (quartet puzzling; Strimmer and von Haeseler, 1996) that allows analysis of large data sets and automatically assigns estimations of support to each internal branch. Rate heterogeneity is incorporated in all models of substitu- tion. All model parameters including rate heterogeneity can be estimated from the data by maximum likelihoods. PUZZLE also computes pairwise maximum likelihood distances as well as branch lengths for user specified trees, and offers a novel method, likelihood mapping, to investigate the support of internal branches without computing an overall tree. 7.3 Methods for Phylogenetic Analysis 7.3.1 DNA sequence analysis DNA sequence analysis has become the most useful method for inferring phylogenetic relationships between organisms. Bruns et al. (1990) described the benefits of using DNA sequences for phylogenetic analysis as being ‘the large number of characters compared can substantially increase the resolving power’. One can also observe the mode of sequence variation, that is, whether a change is a transversion or transition, silent or selected, and one can measure the degree of nucleotide bias; results from different laboratories can be Applications in Fungal Phylogeny 129 compared directly, and the publication of sequences and their deposition in electronic