MOLECULAR PHYLOGENETICS AND EVOLUTION Molecular Phylogenetics and Evolution 27 (2003) 372–383 www.elsevier.com/locate/ympev Phylogenetic significance of morphological characters in the taxonomy of Pestalotiopsis species Rajesh Jeewon,a,* Edward C.Y. Liew,b Jack A. Simpson,c I. John Hodgkiss,d and Kevin D. Hyded a School of Biological Sciences, King Henry Building, University of Portsmouth, Portsmouth PO1 2DY, UK School of Land, Water and Crop Sciences, McMillan Building A05, The University of Sydney, NSW 2006, Australia c Research Division, State Forests of New South Wales, P.O. Box 100, Beecroft, NSW, Australia Centre for Research in Fungal Diversity, Department of Ecology and Biodiversity, The University of Hong Kong, Pokfulam Rd, Hong Kong, SAR, PeopleÕs Republic of China b d Received 19 December 2001; revised 12 November 2002 Abstract There has been considerable disagreement regarding the relationships among Pestalotiopsis species and their delimitations. A molecular phylogenetic analysis was conducted on 32 species of Pestalotiopsis in order to evaluate the utility of morphological characters currently used in their taxonomy. Phylogenetic relationships were inferred from nucleotide sequences in the ITS regions and 5.8S gene of the rDNA under four optimality criteria: maximum parsimony, weighted parsimony, maximum likelihood, and neighbor joining. Phylogenies estimated from all analyses yielded trees of essentially similar topology and revealed 3 major groups that correspond with morphology-based classification systems. Molecular data indicated that the genus contains two distinct lineages based on pigmentation of median cells and four distinct groupings based on morphology of apical appendages. The analyses did not support reliability of other phenotypic characters of this genus, such as spore dimensions. Characters with particular phylogenetic significance are discussed in relation to the taxonomy of Pestalotiopsis. Ó 2003 Elsevier Science (USA). All rights reserved. 1. Introduction Pestalotiopsis Steyaert consists of approximately 205 described species that are easily identified by the presence of relatively fusiform conidia formed within compact acervuli (CABI Bioscience database, 2001). The conidia are usually 5-celled, with 3 brown median cells and hyaline end cells, and with two or more apical appendages arising from the apical cell. Pestalotiopsis species are ubiquitous in distribution, occurring on a wide range of substrata. Many are saprobes (Wu et al., 1982) while others are pathogenic or endophytic on living plant leaves and twigs (Bissett, 1982; Brown et al., 1998; Howard and Albregs, 1973; Hyde and Fr€ ohlich, 1995; Karaca and Erper, 2001; Rivera and Wright, 2000; Taylor et al., 2001; Tuset et al., 1999). Some of these * Corresponding author. Fax: 852-2517-6082. E-mail address: rajeshjeewon@yahoo.com (R. Jeewon). Pestalotiopsis species have gained much attention in recent years as they have been found to produce many important secondary metabolites (Li et al., 2001; Li and Strobel, 2001; Ogawa et al., 1995; Pulici et al., 1997; Strobel et al., 1996). Pestalotiopsis species are anamorphic members of the family Amphisphaeriaceae (Barr, 1975, 1990; Kang et al., 1998, 1999). Pestalotiopsis is a complex genus and consists of members difficult to classify at the species level. At present, inter-specific delineation of this genus is based on morphology of the conidia (Guba, 1961; Nag Rag, 1993), conidiogenesis (Sutton, 1980) and teleomorph association, which has been described for only a few species (Barr, 1975, 1990; Metz et al., 2000; Zhu et al., 1991). Since the establishment of the genus (Steyaert, 1949), numerous taxonomic studies have been conducted in an attempt to devise a suitable classification scheme for the different species (Guba, 1961; Nag Rag, 1993; Suto and Kobayashi, 1993; Sutton, 1980). 1055-7903/03/$ - see front matter Ó 2003 Elsevier Science (USA). All rights reserved. doi:10.1016/S1055-7903(03)00010-1 R. Jeewon et al. / Molecular Phylogenetics and Evolution 27 (2003) 372–383 Steyaert (1949) recognized Pestalotiopsis as a distinct genus, which is not congeneric to Pestalotia as proposed by Guba (1929). This was supported by a recent molecular study based on rDNA sequences (Jeewon et al., 2002). Steyaert (1949) divided the genus Pestalotiopsis into different sections based on the number of apical appendages: Monosetulatae, Bisetulatae, Trisetulatae, and Multisetulatae for species bearing 1, 2, 3, and more than 3 apical appendages, respectively. Each section was further divided into subsections based on differences in conidial shape, pigmentation of median cells, and presence or absence of spatulated apical appendages. An alternative arrangement was proposed by Guba (1961) who grouped Pestalotiopsis species into 3 major sections based on differences in pigmentation of the median cells: concolorous (for those possessing equally colored median cells), versicolorous: umber olivaceous (two upper median cells umber and lowest median cell yellow brown), versicolorous: fuliginous olivaceous (two upper median cells fuliginous, usually opaque, often swollen with a dark central band, and lowest median cell light brown). However, the main features that he relied on were morphometry of the conidia, and number and characteristics of the appendages. A total number of 258 species were described in his monograph. Pigmentation is the result of the deposition of melanin granules within the cell matrix but the origin of such pigmentation has not been established in all species except Pestalotiopsis funerea and Pestalotiopsis triseta (Griffiths and Swart, 1974). Griffiths and Swart (1974) recognized that differences in pigmentation of median cells were of some taxonomic value. This corroborated with the results of Sutton (1961), who investigated cultural differences on different media and the relative abundance of different spore types of Pestalotiopsis sydowiana present in the conidial life cycle. However, in another study carried out by Satya and Saksena (1984), pigmentation of the median cells was shown to be unreliable for differentiating certain Pestalotiopsis species. They observed that Pestalotiopsis glandicola and Pestalotiopsis versicolor var. polygoni produced spores of different color intensities in culture and on different hosts and argued that color contrast of median cells is not a dependable character. Individual species were also found to produce different spore shapes (claviform and fusiform) and were thus incongruent with SteyaertÕs system. Another difficulty in the classification of Pestalotiopsis species stems from the various degrees of cultural variation seen within a species, for such characters as growth rate, conidial morphology and fruiting structure charateristics. Dube and Bilgrami (1965) observed morphological variations in the shape, number and orientation of appendages in cultures of Pestalotiopsis darjeelingensis. A similar phenomenon was also reported by Purohit and Bilgrami (1969) who examined more than 100 pathogenic isolates of Pestalotiopsis. 373 Hughes (1953) and Kendrick (1979) pointed out that developmental features of conidia and conidiophores should be given more importance in taxonomic studies. This concept was also advocated by Sutton (1980), who suggested that a more rationale and natural classification of coelomycetous fungi would be one based on conidiogenesis. This approach however has been aimed mainly at suprageneric classification and in most cases conidiogenous structures have been very difficult to interpret (Morgan-Jones et al., 1972). Watanabe et al. (1998) investigated the conidiomatal development of Pestalotiopsis guepinii and Pestalotiospis neglecta, and found that the two species possess the same type of acervulus development, which is similar to those of Phoma richardiae and Phyllosticta harai. Morphological and developmental studies have been inadequate in establishing evolutionary relationships in Pestalotiopsis. Recently, the taxonomy of this genus was reviewed by Morgan et al. (1998) who explored the utility of artificial neural networks to identify Pestalotiopsis species. These networks have been demonstrated to be quite informative as they revealed that some morphological characters are not good either individually or in combination, and that some species are not sufficiently different to warrant species designation. These studies, however, were restricted in taxonomic sampling and did not explicitly test phylogenetic hypotheses. The systematic relationships of Pestalotiopsis species are difficult to establish as many of them have characters that overlap in many respects. While all species possess appendages, pigmented median cells and spores of similar shape, the major delimiting characters at the species level have been spore and appendage sizes in a broad range of taxa (Guba, 1961; Nag Rag, 1993; Steyaert, 1949). In addition many species have been described, renamed and synonymized based on slight differences in spore morphology from culture and host (Mordue, 1985, 1986; Nag Rag, 1985, 1986; Pal and Purkayastha, 1992; Venkatasubbaiah et al., 1991). The taxonomic affinities of Pestalotiopsis species have been equivocal, confused and hampered by differences of opinions regarding the basic criteria used in segregating species. Molecular studies have shown that Pestalotiopsis is a monophyletic genus (Jeewon et al., 2002) but relationships at the species level were not addressed. The purpose of the current study is to investigate phylogenetic relationships among Pestalotiopsis species by analysis of sequence data derived from the rDNA. The region targeted is the ITS and the 5.8S gene. The specific objectives are to elucidate how morphologically different species are phylogenetically related; to determine whether morphological-based classification schemes are congruent with phylogenies derived from molecular characters; and to test the validity of morphological characters currently used for differentiating Pestalotiopsis species. 374 R. Jeewon et al. / Molecular Phylogenetics and Evolution 27 (2003) 372–383 2. Materials and methods 2.2. DNA extraction, amplification, and sequencing 2.1. Sources of fungal strains For each culture, mycelia were scraped from the surface of the agar and used for DNA extraction following a modified protocol of Doyle and Doyle (1987). Target regions of the rDNA 5.8S gene + ITS regions were amplified symmetrically using primers ITS 4 and ITS 5 (White et al., 1990). Taq polymerase was used in the PCR to amplify approximately 650 base pairs with the following thermal cycling profile: initial denaturation of the double stranded DNA for 3 min at 94 °C, followed by 30 cycles of 1 min denaturation at 94 °C, primer annealing at 54 °C for 50 s, 1.5 min extension at 72 °C, and a final extension for 10 min at 72 °C. A small sample of each amplified product was size-verified by gel electrophoresis. PCR products were purified using the Wizard Preps DNA purification system (Promega, Madison, WI, USA). Primers ITS 2, ITS 3, ITS 4, and ITS 5 (White et al., 1990) were used to Thirty-two strains of Pestalotiopsis were selected for this study on the basis of their morphological characters. Representatives exhibiting a broad range of varying morphological characters were included. The sources of these cultures and specimens are listed in Table 1. For each strain, conidia were isolated, and single spore cultures were grown on PDA at 25 °C for 7–10 days. Strain identity was verified by microscopically examining the fruiting bodies and spores. The morphological characters (ornamentation and pigmentation of the median cells as well as length and width of the conidia and appendages) for each strain were recorded. Species identification was based on the keys provided by Guba (1961) and Steyaert (1949). Table 1 Representative strains of Pestalotiopsis used in this study, their accession numbers, hosts, and geographical origins Species Source of culturea Host/geographic origin GenBank Accession No. Pestalotiopsis adusta Pestalotiopsis aquatica Pestalotiopsis bilicia Pestalotiopsis disseminata Pestalotiopsis funerea Pestalotiopsis gracilis Pestalotiopsis karstenii Pestalotiopsis leucotho€es Pestalotiopsis maculans Pestalotiopsis microspora Pestalotiopsis neglecta Pestalotiopsis palmarum Pestalotiopsis pauciseta Pestalotiopsis rhododendri Pestalotiopsis sydowiana Pestalotiopsis theae Pestalotiopsis uvicola Pestalotiopsis versicolor 1 Pestalotiopsis versicolor 2 Pestalotiopsis vismiae Pestalotiopsis sp. 1 Pestalotiopsis sp. 2 Pestalotiopsis sp. 3 Pestalotiopsis sp. 4 Pestalotiopsis sp. 5 Pestalotiopsis sp. 6 Pestalotiopsis sp. 7 Pestalotiopsis sp. EN8 Pestalotiopsis sp. 8 Pestalotiopsis sp. EN9 Pestalotiopsis sp. 9 Pestalotiopsis sp. EN12 Seiridium cardinale ICMP 5434 HKUCC 8311 BRIP 25718 HKUCC 255 ICMP 7314 HKUCC 8320 ICMP 10669 HKUCC 8315 CBS 322.76 HKUCC 8316 HKUM 996 BRIP 25618 ICMP 11874 BRIP 25628 HKUCC 8326 HKUCC 7982 BRIP 25613 BRIP 25468 BRIP 14534 HKUCC 8328 BRIP 25446 HKUCC 8323 BRIP 25640 BRIP 25624 HKUCC 8322 BRIP 25619 HKUCC 8324 HKUCC 7984 STE-U 1755 HKUCC 8319 HKUCC 8325 HKUCC 8321 ICMP 7323 Digitalis purpurea, New Zealand Leucospermum sp., S. Africa Xanthorrhoea sp., Australia Sonneratia alba, The Philippines Cupressocyparis leylandii, New Zealand Scaevola hainanensis, Hong Kong, China Camellia sp., New Zealand Telopea sp., Hawaii, USA Camellia sp., France Aegiceras cornilatum, Hong Kong Calamus sp., Australia Palm, Australia Ulex europaeus, New Zealand Antidesma ghaesembilla, Australia Protea mellifera, S. Africa Protea mellifera, S. Africa Verticordia sp., Australia Garcia mangostana, Australia Psidium guajava, Australia Leucospermum sp., Hawaii, USA Garcia mangostana, Australia Saccharum officinarum, Hong Kong, China Callistemon sp., Australia Nepenthes khasiana, Australia Unidentified leaf, Hong Kong, China Nepenthes truncata, The Philippines Leucospermum sp., S. Africa Scaevola hainanensis, Hong Kong, China Leucospermum sp., S. Africa Scaevola hainanensis, Hong Kong, China Leucospermum sp., S. Africa Scaevola hainanensis, Hong Kong, China Cupressocyparis leylandii, New Zealand AF409955 AF409956 AF409973 AF409976 AF405299 AF409962 AF405300 AF409969 AF405296 AF409958 AF409975 AF409990 AF409972 AF409986 AF409970 AF405297 AF409994 AF409993 AF405298 AF409977 AF409984 AF409968 AF409985 AF409989 AF409992 AF409991 AF409961 AF405294 AF409980 AF409963 AF409979 AF409994 AF409995 a BRIP, Queensland Department of Primary Industries Plant Pathology Herbarium; CBS, Centraalbureau voor Schimmelcultures; HKUCC, The University of Hong Kong Culture Collection; ICMP, International Collection of Microorganisms from Plants; STE-U, University of Stellenbosch Culture Collection. R. Jeewon et al. / Molecular Phylogenetics and Evolution 27 (2003) 372–383 sequence both strands of the DNA molecule in an automated sequencer (ALF Express, Pharmacia-Biotech, Piscataway, NJ, USA) following the manufacturerÕs protocols. Nucleotide sequences reported in this paper have been deposited in GenBank and are listed in Table 1. 2.3. Sequence assembly and alignment DNA sequences obtained for each strain from each primer were inspected individually for quality and then spliced together using the ALF software. Both strands of the DNA were then assembled to produce a consensus sequence for each strain using SeqPup (Gilbert, 1996). All sequences were aligned using Clustal X with default settings (Thomson et al., 1997). Gaps were introduced to improve alignments, which were finally optimized manually. Trees were viewed in Treeview (Page, 1996). 2.4. Phylogenetic analyses Phylogenies based on the ITS and 5.8S gene data were constructed by performing heuristic searches under four optimality criteria: maximum parsimony (MP), weigthed parsimony (WP), maximum likelihood (ML), and neighbor joining (NJ). Searches were carried out using PAUP* 4.0b8 (Swofford, 2001). 2.4.1. Maximum parsimony MP analysis was performed using the heuristic search option, simple, and random addition stepwise of taxa. All characters for the datasets were coded as unordered and branch swapping was performed using the tree bisection-reconnection (TBR) swapping algorithm. The entire dataset was analyzed by treating gaps as missing data as well as fifth state and with transitions–transversions weighted equally. The dataset consisted of 531 sites of which 79 were parsimony informative. Nonparametric bootstrapping (Felsenstein, 1985; Sanderson, 1989) with 1000 replications was used to assess the confidence associated with the various clades. A maxtrees limit of 1000 trees and simple sequence addition were used in the bootstrap analyses. The outgroup taxa selected for rooting the trees were a sister group to Pestalotiopsis based on previous phylogenetic analyses (Jeewon et al., 2002). Outgroups of more distantly related genera were attempted, but these created excessive ambiguous alignment in the highly variable ITS regions. Consistency index (CI) and other tree indices were calculated for each consensus tree to give an indication of the amount of homoplasy present. 2.4.2. Weighted parsimony A weighted parsimony analysis was carried out with transitions weighted 1.5 and 2 times over transver- 375 sions. Gaps were treated as missing data or fifth state. Insertions/deletions (indels) were included in the analysis because in most of the analyzed sequences, indels were short (1–3 nucleotides) except for two regions in the ITS 1, which were 16 and 15 nucleotides, respectively. These two large indel regions were excluded from the analysis. Support for the inferred trees topologies was evaluated using 1000 bootstrap replications, as implemented in PAUP* 4.0b8 (Swofford, 2001). 2.4.3. Maximum likelihood Analyses were conducted under the likelihood criterion following an iterative search strategy. A single most parsimonious tree was used as starting tree in the ML search. The HKY and F84 models of nucleotide substitution were used with rates assumed to follow a gamma distribution with no enforcement of a molecular clock. Analyses were performed by firstly estimating the transition–transversion ratios, shape parameter of the gamma distribution and base frequencies. These estimated parameters were used in subsequent ML searches. 2.4.4. Neighbor joining For distance analysis, the dataset was analyzed under a variety of assumptions and under different nucleotide substitution models including HKY85 (Hasegawa et al., 1985) and K2P (Kimura, 1980). All characters were treated as unordered and were weighted equally. Bootstrap values were obtained from 1000 replicates. To evaluate the statistical significance of all the topologies inferred from the different optimality criteria, the Kishino and Hasegawa (1989) and Templeton (1983) tests as implemented in PAUP* 4.0b8 (Swofford, 2001) were conducted. Trees were viewed in Treeview (Page, 1996). 3. Results MP analysis with gaps treated as missing data yielded 78 equally most parsimonious trees, the strict consensus of which was 135 steps in length (CI ¼ 0.852, RI ¼ 0.970, )log likelihood ¼ 1506.4127). All MP analyses treating gaps as missing data essentially yielded trees of similar topologies and same )log likelihoods (Table 2). This dataset could not be bootstrapped as the procedure was computationally too demanding and had to be aborted after 20 replicates. Cladistic analysis employing the criterion of weighted parsimony (transitions weighted 2 times over transversions, as related to the estimated values) and treating gaps as fifth state yielded 9 trees, the strict consensus tree of which is shown in Fig. 1 (TL ¼ 214 steps, C ¼ 0.813, RI ¼ 0.969, )log likelihood ¼ 1490.6493). The likelihoods of the 3 consensus 376 R. Jeewon et al. / Molecular Phylogenetics and Evolution 27 (2003) 372–383 Table 2 Summary of the Kishino–Hasegawa and Templeton tests on the topologies obtained from parsimony analyses with various transition–transversion differential weightings TT ratio TL CI RI HI )ln L KH/Templeton testa Gapmode ¼ missing 1:1 1.5:1 2:1 135 159.5 184 0.852 0.862 0.870 0.970 0.972 0.974 0.148 0.138 0.130 1506.4127 1506.4127 1506.4127 P ¼ 0:3178=P ¼ 1:0 P ¼ 0:3178=P ¼ 1:0 P ¼ 0:3178=P ¼ 1:0 Gapmode ¼ newstate 1:1 1.5:1 2:1 203 222 241 0.847 0.829 0.813 0.975 0.972 0.969 0.153 0.171 0.187 1510.5714 1496.0495 1492.6901 P ¼ 0:1025=P ¼ 0:25 P ¼ 0:1575=P ¼ 0:5 Best TT, transition–transversion; TL, tree length; CI, consistency index; RI, retention index; HI, homoplasy index; )ln L ¼ )log likelihood. Probability of getting a more extreme T value under the null hypothesis of no difference between the two trees (two tailed test) with significance at P < 0:05. a trees obtained with gaps treated as newstate and under different transitions:transversions, however were not significantly worse (P < 0:05, Table 2). Consensus trees generated by both MP and WP analyses (transitions weighted 2 times over transversions) had essentially similar tree topologies except that one of the clades was partially unresolved when gaps were treated as missing data and transitions:transversions weighted equally. The maximum likelihood analysis using either a MP or NJ tree as starting trees and under different models of evolution gave trees of similar topologies. The HKY model of substitution with an estimated shape parameter yielded 2 trees, the consensus of which is shown in Fig. 2. This tree was identical in topology with that obtained by unweighted parsimony and treating gaps as missing data. Two ML trees similar in topology were obtained, which had the same )log likelihood score of 1490.6493. Tree length was 135, CI ¼ 0.852, RI ¼ 0.970, and HI ¼ 0.148. Estimated transition:transversion ratio was 2.0059 and base frequencies were as follows: A ¼ 0.2446, C ¼ 0.2274, G ¼ 0.2, and T ¼ 0.3287. A phylogram constructed under the Neighbor Joining criterion yielded a tree, which was essentially the same as the unweighted MP tree but less resolved (data not shown). Bootstrap support was strong for all the branches (>65%). Tree length was 155, CI ¼ 0.742, RI ¼ 0.941, HI ¼ 0.258, and )log likelihood ¼ 1543.1011. Table 3 shows the results of the Kishino–Hasegawa and Templeton tests between all alternative topologies obtained by MP, WP, ML, and NJ criteria. The ML tree was identified as the best. Results indicate that trees produced by MP, WP (treating gaps as newstate) and ML are not significantly different, whereas the NJ tree is significantly worse and is therefore rejected. Although trees from MP and WP analyses are very similar in topology, the WP tree generated by treating gaps as newstate has been selected for the purpose of inferring phylogenies among Pestalotiopsis species because the confidence of its topology could be sta- tistically assessed by bootstrapping and it shows relationships which are more resolved among the species in Subclade b (Fig. 1). All clades are supported by high bootstrap values. As shown in Figs. 1 and 2, the partition of the genus Pestalotiopsis falls into three distinct lineages (Clades X, Y, and Z). These three groups of Pestalotiopsis are all monophyletic and are supported with high bootstrap confidence of 100%. Clade X consists of species, which possess versicolorous median cells. Clade Y and Clade Z group species that characterized by brown concolorous median cells. Characters of conidial morphology are shown in Fig. 1. 3.1. Pigmentation of median cells Two clades are observed for pigmentation of versicolorous and concolorous. The versicolorous clade (Clade X) is monophyletic, while the concolorous clades (Clade Y and Z) are paraphyletic (Fig. 1). All phylogenetic analyses support the separation of species possessing versicolorous median cells from species possessing concolorous median cells as recognized by Guba (1961) and Steyaert (1949). However the grouping of Pestalotiopsis species with versicolorous median cells into two categories (umber olivaceous and fuliginous olivaceous) is not supported by our data as species possessing both types of pigmentation cluster together. For instance, Pestalotiopsis sp. 3, characterized by umber olivaceous median cells, groups together with other members in the same Subclade possessing fuliginous median cells (Subclade a). Another interesting group in the tree are the representatives of Subclade b, which are characterized by umber olivaceous median, with the exception of Pestalotiopsis sp. EN8, Pestalotiopsis sp. 4, and Pestalotiopsis pauciseta, which also possess a small number of fuliginous olivaceous median cells. Members of Subclade b are closely related to those of Subclade a but are distinguished by having umber olivaceous median cells with no dark septa separating the two upper R. Jeewon et al. / Molecular Phylogenetics and Evolution 27 (2003) 372–383 377 Fig. 1. Strict consensus of 9 most parsimonious trees based on the ITS and 5.8S dataset of 33 taxa (TL ¼ 241 steps, CI ¼ 0.813, RI ¼ 0.969, and )log likelihood ¼ 1492.6901). Tree was obtained by treating gaps as a fifth character and weighting transitions 2 times over transversions. Groups labeled X–Z are the same as in Fig. 1 and the letters (a–g) indicate the different monophyletic subclades. Morphological characters distinguishing each group pertaining to each clade are shown on the right of the cladogram. Numbers above the nodes represent the proportion of 1000 bootstrap replications. The designated outgroup was Seiridium cardinale. The asterisk (*) indicates the clades which receives less than 50% bootstrap support. 378 R. Jeewon et al. / Molecular Phylogenetics and Evolution 27 (2003) 372–383 Fig. 2. Strict consensus of 2 trees generated from a maximum likelihood analysis of the ITS and 5.8S dataset (TL ¼ 135, CI ¼ 0.852, RI ¼ 0.970, and )log likelihood ¼ 1490.6493). Letters X–Z above the branches represent the different monophyletic groups having distinct morphological characters. The asterisk (*) indicates the clade which is more resolved in Fig. 1. Designated outgroup was Seiridium cardinale. 379 R. Jeewon et al. / Molecular Phylogenetics and Evolution 27 (2003) 372–383 Table 3 Results of the Kishino–Hasegawa (KH) and Templeton tests for the trees generated by different optimality criteria Consistency index )Ln likelihood KH testc Templeton Testc MP treea WP treeb (Fig. 1) ML tree (Fig. 2) NJ tree 0.852 1506.4127 P ¼ 0:3178 P ¼ 0:3178 0.813 1492.6901 P ¼ 1:0 P ¼ 1:0 0.852 1490.6493 Best Best 0.742 1543.1011 P ¼ 0:0001 P ¼ 0:0001 a MP tree treating gaps as missing data and equal weighting. WP tree treating gaps as fifth state with transition weighted 2 times over transversions. c Probability of getting a more extreme T value under the null hypothesis of no difference between the two trees (two tailed test) with significance at P < 0:05. b median cells. During the course of our study it was found that P. pauciseta, P. sydowiana, and Pestalotiopsis leucotho€es produced a small number of spores possessing umber and fuliginous olivaceous median cells as well. Clade Y, which forms the sister group to Clade X, is characterized by species with concolorous median cells. The same applies to Clade Z, which is supported by a 100% bootstrap confidence. those possessing 3 appendages as well. In addition, we observed that P. adusta, Pestalotiopsis microspora, Pestalotiopsis sp. 7, and Pestalotiopsis sp. 9 possess mostly 2 apical appendages, but some species produced an equal number of conidia with 3 apical appendages. 3.2. Appendage morphology Clade Z forms a distinct monophyletic clade (100% bootstrap) and the relationships within this clade are consistent in all phylogenies inferred. This clade consists of species characterized by brown concolorous median cells and diverse spore sizes. Species in Group X, Y, and those in Subclades c, d, and g (Clade Z) are characterized by conidial length greater than 20 lm whereas species in Subclades e and f possess spore length of less than 20 lm. The only ambiguity is P. adusta, which possesses a conidial length of less than 20 lm but clusters with other species possessing conidial length of greater than 20 lm. It is also worth mentioning that species grouping based on conidial width is supported by the sequence data. Species having conidial width of less than 6 lm are present in Clades Y and Z only whereas Clade X includes species characterized by a conidial width of greater than 6 lm. However, conidial form did not correspond to any particular grouping in the trees. Species possessing conidia that are fusiform, fusiformelliptical, clavate, and reniform intermingle in all the clades. Different types of ornamentation of the median cells can be seen in different groups. Members of Clade X are characterized by having the two upper median cells with verruculose and verrucose ornamentation. Species of Clade Y and Z possess median cells, which have smooth ornamentation with the exception of species in Subclade g that possess spinose and foveate ornamentation. Another well-defined clade includes Pestalotiopsis theae, Pestalotiopsis sp. 5, and Pestalotiopsis sp. 6 (Clade Y), which is supported by 100% bootstrap confidence. Morphologically this monophyletic group contains taxa that are characterized by brown concolorous median cells, long fusiform conidia (greater than 25 lm in length), apical appendages with a length ranging from 25 to 40 lm and which is knobbed at the tip. All species in other clades possess apical appendages that are not knobbed. Results also show that species form distinct groupings based on the length of the apical appendages. Within Clade X, two Subclades are evident based on dimensions of apical appendages. Subclade a groups species having apical appendages of greater than 25 lm in length whereas Subclade b groups species having apical appendages of 20–25 lm in length. There are also other welldefined Subclades in Clade Z based on the dimension of apical appendages. Apical appendage length of 15–19 lm can be found in Subclades c, f, and g (except Pestalotiopsis adusta which has a length of less than 15 lm) whereas species characterized by apical appendages of less than 15 lm can be found only in Subclades d and e. There are, however no clades that correspond to particular groupings based on the dimension of basal appendages or of the presence of 2 or 3 apical appendages. Species possessing basal appendages less than or greater than 5 lm appear to intermingle in all clades. Such is also the case regarding the number of apical appendages. Species of Subclades c, d, e, and f examined in this study possess 2 as well as 3 apical appendages. Those species possessing mostly 2 appendages did not cluster with others sharing the same number of apical appendages; they intermingled with 3.3. Conidial size and shape 4. Discussion Like most coelomycetes, classification of Pestalotiopsis species based on morphological characters has 380 R. Jeewon et al. / Molecular Phylogenetics and Evolution 27 (2003) 372–383 been equivocal because of overlapping variation in size and shape of homologous structures. Major emphasis has been placed on the pigmentation of median cells and the size of appendages. In this study, the ITS regions and 5.8S gene of the rDNA of 32 isolates were analyzed to elucidate relationships within Pestalotiopsis. This molecular based cladistic analysis provides a context for reexamining the key characters that have been used in the classification of this genus. 4.1. Pigmentation of median cells Pigmentation of the median cells has been considered to be of great taxonomic importance at the species level. Sequences from the ITS and 5.8S gene analyzed in this study demonstrated strong bootstrap support for a close phylogenetic relationship among species possessing versicolorous median cells as well as those species characterized by concolorous median cells. The data are also generally concordant with species relationships proposed by Guba (1961) and Steyaert (1949) who classified all species producing versicolorous median cells under one section (Versicolores) and species characterized by concolorous median cells in another section (Concolorae). However, morphological subgroups of versicolorae as defined by Guba (1961), who organized species possessing versicolorous median cells into 2 sections according different color intensities (umber olivaceous and fuliginous olivaceous) is doubtful because species relationships between the two different color intensities were not resolved confidently. By examining conidia of P. glandicola and P. versicolor var. polygoni on the host and in culture, Purohit and Bilgrami (1969) and Satya and Saksena (1984) showed that the conidia exhibited different color intensities (concolorous as well as versicolorous) on different substrates, thus demonstrating the difficulty in using degree of pigmentation as a reliable character. They pointed out that the color of median cells is an unstable feature used in the systematics of Pestalotiopsis. Molecular data, however, is inconsistent with their findings as species producing spores with concolorous median cells and versicolorous median cells form distinct groups supported by high bootstrap values. Therefore, despite arguments in the morphological approach over the reliability of pigmentation as a taxonomic character, results here indicate that pigmentation is a sound diagnostic character for species differentiation. Current results are in agreement with Sutton (1961), who investigated cultural differences of P. sydowania on different media and observed that the species produced umber as well as fuliginous median cells. Sutton (1961) inferred that spores either: (i) undergo a natural sequence of maturation, developing from one spore type to another, or (ii) produce morphologically distinct spore types. The former conclusion was refuted by Purohit and Bilgrami (1969). Based on the results obtained herein, it can be hypothesized that some species may produce umber as well as fuliginous median cells at different stages of maturity. Pigmentation is the result of deposition of melanin granules within the cell matrix but the origin of such pigmentation has not been established except in P. funerea and P. triseta (Griffiths and Swart, 1974). Probably those species possessing versicolorous median cells contain similar pigments but this has not been investigated. 4.2. Appendage morphology Species possessing apical appendages that are knobbed clearly fall into a monophyletic clade (Clade Y). Although these species possess other characters more commonly found in Group Z, they did not cluster in that group as expected. This group, characterized by concolorous median cells forms the sister group to Clade X, which possesses versicolorous median cells. Presumably these species have lost the ability to produce versicolorous median cells and have gained the ability to produce knobbed apical appendages. The presence of a knobbed apical appendage seems to have evolved once during the evolution of Pestalotiopsis species, and hence is a phylogenetically reliable taxonomic character. SteyaertÕs and GubaÕs system that grouped all species possessing spathulate appendages in one section (spathulate) is therefore valid. The mean lengths of apical and basal appendages have been used as important taxonomic characters to delineate species. From the phylogenies generated here it can be seen that species sharing similar apical appendage length are closely related to each other (Fig. 1). Sequence data suggest the possibility that there may be four phylogenetic groups within the current morphological concept of Pestalotiopsis. These include species possessing apical appendage length: (i) less than 15 lm, (ii) 15–19 lm, (iii) 20–25 lm, and (iv) greater than 25 lm. Current molecular analyses also resolved longstanding confusion as to whether species characterized by different basal appendage dimensions are closely related. Species with basal appendages greater or less than 5 lm appear to overlap in all the clades and there is not a clear relationship among species sharing similar length of basal appendages. Therefore, length of the basal appendage as a taxonomic criterion to segregate species can be misleading. On the other hand, absence of basal appendages may be quite important for species delimitation. In this study, only 2 species (Pestalotiopsis karstenii and Pestalotiopsis maculans) are characterized by the absence of basal appendages and they cluster together with a high bootstrap support. R. Jeewon et al. / Molecular Phylogenetics and Evolution 27 (2003) 372–383 Based on our morphological observations and molecular data, it can also be inferred that the number of apical appendages is not a reliable taxonomic character to separate species within the genus Pestalotiopsis. There is no clear separation of species possessing 2 or 3 apical appendages. Many of them, for instance, Pestalotiopsis disseminata, Pestalotiopsis uvicola, and Pestalotiopsis palmarum, examined in this study possess 2 as well as 3 apical appendages. Moreover those species possessing mostly 2 appendages (for example, P. adusta, P. microspora, Pestalotiopsis sp. 7, and Pestalotiopsis sp. 9) are not phylogenetically closely related to other species sharing the same number of apical appendages. Instead they cluster according to spore and apical appendage length. It is worth mentioning that, during microscopic examination of spores, some species produced an equal number of conidia with 2 and 3 apical appendages. Results presented here corroborate the findings of Suto and Kobayashi (1993) who pointed out that it was impossible to group species based on the number of apical appendages. Furthermore, the bisetulae section in the system advocated by Steyaert (1949) to accommodate for species with 2 apical appendages is doubtful. 4.3. Conidial size and shape Traditionally, taxonomists placed much emphasis on length and width of the conidia (length–width ratio), length of appendages and ornamentation of median cells. Spore length has been used as a key character and many new species have been named based on slight differences in spore sizes (Mordue, 1985, 1986; Nag Rag, 1985, 1986; Pal and Purkayastha, 1992; Venkatasubbaiah et al., 1991). The results here, however, indicate that spore size is homoplasious. Species sharing similar conidial length did not cluster together and are therefore not closely related to each other. Similar conclusions were made by Dube and Bilgrami (1965), whose work indicated that conidial length is not a reliable taxonomic character to define Pestalotiopsis species. On the other hand, spore width can be phylogenetically informative and given high taxonomic weighting since all species in Clade X are characterized by conidial width of greater than 6 lm, whereas those in Clade Y and Z possess a conidial width of less than 6 lm. Pestalotiopsis microspora and Pestalotiopsis sp. EN12 clustered together and are closely related to P. karstenii and P. maculans. These four species share the same conidium length (less than 20 lm) whereas members of other clades are characterized by spores greater than 20 lm in length. Spore length is the only morphological difference separating species in Subclades e and f from species from other Subclades in Clade Z. Comparison of P. microspora and Pestalotiopsis sp. EN12 ITS sequences revealed a high degree of similarity (96%) and a 98% nucleotide similarity among species in Subclade f. 381 Species of Subclade d, which share 100% sequence similarity (Pestalotiopsis bicilia, Pestalotiopsis sp. 7, and Pestalotiopsis vismiae) and Subclade e (P. microspora and Pestalotiopsis sp. EN12) are phylogenetically closely related. This is consistent with the fact that these species produce spores that possess very short apical appendages, their lengths ranging from 8–15 lm. The principal separation between species in Subclade d and those in Subclade e (P. microspora and Pestalotiopsis sp. EN12) is based on spore sizes. The latter group is characterized by having spore lengths of less than 20 lm, whereas those in Subclade d possess spore lengths of greater than 20 lm. Their placement as different species is supported by the molecular data even though these species share the same appendage length. These results generally lead to the assumption that classification of Pestalotiopsis species based on conidium sizes and conidium length– width ratio as proposed by Guba (1961) and Nag Rag (1993) might be artificial. On the basis of tremendous variation in spore sizes and in light of the rDNA phylogeny, it can be inferred that conidium sizes can be taxonomically useful when given lower weighting. Molecular data does not reveal a close relationship among species characterized by similar conidial form. This suggests that conidial form is uninformative for species differentiation and results cast doubt on the grouping of Pestalotiopsis species into different groups based on the form of the conidia as outlined by Steyaert (1949). Presumably these species are dimorphic or polymorphic and exhibit different forms under different cultural conditions as suggested by Dube and Bilgrami (1965). On the other hand, it appears that ornamentation is closely related to the degree of pigmentation of median cells. The presence of verruculose ornamentation is more common in species possessing fuliginous median cells, whereas smooth ornamentation is mostly present in species characterized by concolorous median cells. The phylogenetic affinities of P. adusta (Subclade c remain ambiguous and are not dealt with in detail as it groups with members of Subclade d in the NJ tree and because of a lack of bootstrap support. The diagnostic morphological features that separate members of Subclade g (basal to the other species in group Z) are obscure, as this group possesses divergent morphologies. They possibly form a genetic species complex indistinguishable using the molecular approach here and its phylogenetic relationships remain uncertain. Another point of interest is the clustering of P. maculans and P. karstenii, which agrees with morphologybased systems. The type species of the genus, P. maculans is characterized by a spore size of 15–20 lm; 1–3 apical appendages, which are at times branched, and basal appendages that are usually absent. Similar morphologies are shared by P. karstenii except that the latter possesses only 1 apical appendage that usually forms 2–3 branches. Nag Rag (1993) and Guba (1961) recognized a close re- 382 R. Jeewon et al. / Molecular Phylogenetics and Evolution 27 (2003) 372–383 lationship between these two species on morphological grounds and host association (Camellia). Guba (1961) considered P. maculans and P. karstenii as synonyms. Such synonymy was not accepted by Nag Rag (1993) and Sutton (1980), who treated these as two distinct species. When these two species were examined under the microscope, the main difference observed was the number of appendages arising from the apical cell. Our sequencebased analyses also demonstrate that these two species are closely related (98% sequence similarity and 90% bootstrap support). It is possible that P. maculans and P. karstenii are synonyms, as the characters used to support their distinction can be interpreted as extreme reductions. 5. Conclusion Based on rDNA evidence, pigmentation of the median cells appears to be an important phylogenetic marker for species delineation within Pestalotiopsis as species possessing versicolorous median cells are distinct from those characterized by concolorous median cells. A close relationship among pigmented median cells has been suggested by numerous taxonomists (Guba, 1961; Steyaert, 1949; Sutton, 1980) and this is supported herein by molecular evidence. However, GubaÕs subsections of species possessing versicolorous median cells (umber versicolorous and fuliginous versicolorous) are doubtful. Appendage tip morphology and pigmentation of median cells are phylogenetically significant and should be given high weighting in species delineation. Conidium length, length of basal appendages, ornamentation of median cells, and spore form appear to be highly plastic morphologies. A word of caution is necessary when grouping species based on spore sizes. This character has been given too much emphasis in the past when describing new species but our results stimulate a serious reconsideration. Nevertheless, spore length could be useful when given lower taxonomic weighting. Distinction of Pestalotiopsis species based on spore sizes can be more reliable for delineating species groups rather than individual taxa. Contrarily conidial form and ornamentation of median cells appear to have undergone convergent evolution within Pestalotiopsis and are not definitive taxonomic characters. Our findings contribute to the understanding of the evolution of morphological characters. Versicolorous median cells appear to have evolved from concolorous median cells and spathulate apical appendages from nonspathulated apical appendages. Pigmentation and apical appendage tip morphology appear to be synapomorphies for Pestalotiopsis species. While not intended as a comprehensive survey of the genus Pestalotiopsis, our study is significant in evaluating the relative importance and utility of morphological characters in segregating Pestalotiopsis species. We propose that Pestalotiopsis species be delineated, in order of taxonomic weighting, on the basis of: (1) the degree of pigmentation of median cells (either brown concolorous or versicolorous), (2) apical appendage tip morphology (presence or absence of spathulation), (3) apical appendage length, and (4) spore length. Acknowledgments We are grateful to the following mycologists who provided cultures and specimens: Dr. A. Aptroot, Dr. R. Fogel, Dr. E.H.C. Mckenzie, T. Nikulin, Dr. R.G. Shivas, Dr. J.E. Taylor, and Prof. M.J. Wingfield. This research was funded by the Hong Kong Research Grants Council. Dr. R. Dulymamode and G.J.D. Smith are thanked for thoughtful reviews and support. Rajeeta Jeewon is thanked for her continuous support during the course of the study. Heidi Kong and Helen Leung are thanked for laboratory assistance. This paper represents part of a dissertation submitted by the first author in partial fulfillment of a doctoral degree. References Barr, M.E., 1975. Pestalosphaeria, a new genus in the Amphisphaeriaceae. Mycologia 67, 187–194. Barr, M.E., 1990. Prodromus to nonlichenized, pyrenomycetous members of class Hymenoascomycetes. Mycotaxon 39, 43–184. Bissett, J., 1982. Pestalotiopsis menezesiana on greenhouse plantings of Cissus rhombifolia with notes on related fungi occurring on Vitaceae. Can. J. Bot. 60, 2570–2574. Brown, K.D., Hyde, K.D., Guest, D.I., 1998. Preliminary studies on endophytic fungal communities of Musa acuminata species complex in Hong Kong and Australia. Fung. Divers. 1, 27–51. CABI Bioscience database, 2001. http://194.131.255.3/cabipages/ names/Names.asp. Doyle, J.J., Doyle, J.L., 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissues. Phytochem. Bull. 19, 11–15. Dube, H.C., Bilgrami, K.S., 1965. Variations in the conidial morphology of Pestalotiopsis darjeelingensis in culture. Curr. Sci. 34, 487. Felsenstein, J., 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39, 783–791. Gilbert, D.G., 1996. SeqPup, biosequence editor and analysis software for molecular biology Bionet. Software, July 1996. Griffiths, D.A., Swart, H.J., 1974. Conidial structure in two species of Pestalotiopsis. Trans. Br. Mycol. Soc. 62, 295–304. Guba, E.F., 1929. Monograph of the genus Pestalotia de Notaris. Part 1. Phytopathology 19, 191–232. Guba, E.F., 1961. Monograph of Pestalotia and Monochaetia. Harvard University Press, Cambridge, MA. Hasegawa, M., Kishino, K., Yano, T.A., 1985. Dating of human–ape splitting by a molecular clock of mitochondrial DNA. J. Mol. Evol. 21, 160–174. Howard, C.M., Albregs, E.E., 1973. A strawberry fruit rot caused by Pestalotia longisetula. Phytopathology 63, 862–863. Hughes, S.J., 1953. Conidiophores, conidia, and classification. Can. J. Bot. 31, 577–659. Hyde, K.D., Fr€ ohlich, J., 1995. Mycosphaerella palmicola associated with leaf spots of Cocos nucifera in Australia, Iran Jaya and Papua New Guinea. Mycol. Res. 99, 704–706. R. Jeewon et al. / Molecular Phylogenetics and Evolution 27 (2003) 372–383 Jeewon, R., Liew, E.C.Y., Hyde, K.D., 2002. Phylogenetic relationships of Pestalotiopsis and allied genera inferred from ribosomal DNA sequences and morphological characters. Mol. Phylogenet. Evol. 25, 378–392. Kang, J.C., Kong, R.Y.C., Hyde, K.D., 1998. Studies on the Amphisphaeriales I. Amphisphaeriaceae (sensu stricto) and its phylogenetic relationships inferred from 5.8S rDNA and ITS2 sequences. Fung. Divers. 1, 147–157. Kang, J.C., Hyde, K.D., Kong, R.Y.C., 1999. Studies on the Amphisphaeriales. The Amphisphaeriaceae (sensu stricto). Mycol. Res. 103, 53–64. Karaca, G.H., Erper, I., 2001. First report of Pestalotiopsis guepinii causing twig blight on hazelnut and walnut in Turkey. Plant Pathol. 50, 145. Kendrick, W.B., 1979. The Whole Fungus. National Museum of Canada, Ottawa. Kimura, M., 1980. A simple method for estimating evolutionary rate of base substitution through comparative studies of nucleotide sequences. J. Mol. Evol. 16, 111–120. Kishino, H., Hasegawa, M., 1989. Evaluation of the maximum likelihood model estimates of the evolutionary tree topologies from sequence data, and the branching order in Homonoidea. J. Mol. Evol. 29, 170–179. Li, J.Y., Harper, J.K., Grant, D.M., Tombe, B.O., Bashyal, B., Hess, W.M., Strobel, G.A., 2001. Ambuic acid, a highly functionalized cyclohexenone with antifungal activity from Pestalotiopsis spp., Monochaetia sp. Phytochemistry 56, 463–468. Li, J.Y., Strobel, G.A., 2001. Jesterone and hydroxy-jesterone antioomycete cyclohexenone epoxides from the endophytic fungus Pestalotiopsis jesteri. Phytochemistry 57, 261–265. Metz, A.M., Haddad, A., Worapong, J., Long, D.M., Ford, E.J., Hess, W.M., Strobel, G.A., 2000. Induction of the sexual stage of Pestalotiopsis microspora, a taxol-producing fungus. Microbiology 146, 2079–2089. Mordue, J.E.M., 1985. An unusual species of Pestalotiopsis:P. steyaertii. sp. nov. Trans. Br. Mycol. Soc. 85, 379–380. Mordue, J.E.M., 1986. Another unusual species of Pestalotiopsis:P. montellicoides sp.nov. Trans. Br. Mycol. Soc. 86, 665–668. Morgan-Jones, G., Nag Rag, T.R., Kendrick, W.B., 1972. Conidium ontogeny in coelomycetes. IV. Percurrently proliferating phialides. Can. J. Bot. 50, 2009–2014. Morgan, A., Boddy, L., Mordue, E.M., Morris, C.W., 1998. Evaluation of artificial neural networks for fungal identification, employing morphometric data from spores of Pestalotiopsis species. Mycol. Res. 103, 975–984. Nag Rag, T.R., 1985. Redisposals and redescriptions in the Monchaetia Seiridium, Pestalotia Pestalotiopsis complexes. I. The correct name for the type species of Pestalotiopsis. Mycotaxon 22, 43–51. Nag Rag, T.R., 1986. Redisposals and redescriptions in the MonchaetiaSeiridium, Pestalotia- Pestalotiopsis complexes. VII. Pestalotia citrini, P. maura and Pestalotiopsis uvicola. Mycotaxon 26, 211–222. Nag Rag, T.R., 1993. In: Coelomycetous Anamorphs with Appendage Bearing Conidia. Mycologue Publications, Waterloo, Ontaria, Canada, p. 1101. Ogawa, T., Ando, K., Aotani, Y., Shinoda, K., Tanaka, T., Tsukuda, E., Yoshida, M., Matsuda, Y., 1995. RES-1214-1 and Novelpeptidic endothelin type A receptor antagonists produced by Pestalotiopsis sp. J. Antibiot. 48, 1401–1406. Page, R.D.M., 1996. TREEVIEW: an application to display phylogenetic trees on personal computers. Comput. Appl. Biosci. 12, 357–358. 383 Pal, A.K., Purkayastha, R.P., 1992. New parasitic fungi from Indian mangrove. J. Mycopathol. Res. 30, 172–176. Pulici, M., Sugawara, F., Koshino, H., Okada, G., Esumi, Y., Uzawa, J., Yoshida, S., 1997. Metabolites of Pestalotiopsis spp., endophytic fungi of Taxus brevifolia. Phytochemistry 46, 313–319. Purohit, D.K., Bilgrami, K.S., 1969. Variations in the conidial morphology of genus Pestalotiopsis. Indian Phytopath. 22, 275–279. Rivera, M.C., Wright, E.R., 2000. First report of azalea petal blight caused by Pestalotiopsis guepini in Argentina. Plant Dis. 84, 100. Satya, H.N., Saksena, S.B., 1984. Some aspects of taxonomy of the genus Pestalotia I-Color intensities of intermediate cells of spores. In: Subramanian, C.V. (Ed.), Proceedings of the International Symposium on Taxonomy of Fungi. Sanderson, M.J., 1989. Confidence limits on phylogenies: the bootstrap revisited. Cladistics 5, 113–129. Strobel, G., Yang, X., Sears, J., Kramer, R., Sidhu, R.S., Hess, W.M., 1996. Taxol from Pestalotiopsis microspora of Taxus wallachiana. Microbiology 142, 435–440. Steyaert, R.L., 1949. Contributions a lÕetude monographique de Pestalotia de Not. et Monochaetia Sacc. (Truncatella gen. nov. et  tat Bruxelles 19, Pestalotiopsis gen. nov.). Bull. Jard. Bot. E 285–354. Suto, Y., Kobayashi, T., 1993. Taxonomic studies on the species of Pestalotiopsis, parasitic on conifers in Japan. Trans. Mycol. Soc. Jpn. 34, 323–344. Sutton, B.C., 1961. Coelomycetes. I.. Mycol. Pap. 80, 1–16. Sutton, B.C., 1980. In: The Coelomycetes: Fungi Imperfecti with Pycnidia, Acervular and Stromata. Commonwealth Mycological Institute, Kew, Surrey, England, p. 696. Swofford, D.L., 2001. PAUP*: Phylogenetic Analysis Using Parsimony (*and other methods), Version 4.0b8. Sinauer Associates, Sunderland, MA. Taylor, J.E., Crous, P.W., Palm, M.E., 2001. Foliar and stem fungal pathogens of Proteaceae in Hawaii. Mycotaxon 78, 449–490. Templeton, A.R., 1983. Phylogenetic inference from restriction endonuclease cleavage sites maps with particular reference to the evolution of humans and the apes. Evolution 37, 269–285. Thomson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., Higgins, D.G., 1997. The Clustal_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res. 25, 4876–4882. Tuset, J.J., Hinarejos, C., Mira, J.L., 1999. First report of leaf blight on sweet Persimmon tree by Pestalotiopsis theae in Spain. Plant Dis. 83, 1070. Venkatasubbaiah, P., Grand, L.F., Van Dyke, G.C., 1991. A new species of Pestalotiopsis on Oenothera. Mycologia 83, 511–513. Watanabe, K., Doi, Y., Kobayashi, T., 1998. Conidiomatal development of Pestalotiopsis guepinii and P. neglecta on leaves of Gardenia jasminoides. Mycoscience 39, 71–75. White, T.J., Bruns, T.L., Lee, S., Taylor, J.W., 1990. Amplication and direct sequencing of fungal ribosomal RNA genes for phylogenetics. In: Innis, M., Gelfand, D.H., Sninsky, J.J., White, J.T. (Eds.), PCR Protocols: A Guide to Methods and Applications. Academic Press, San Diego, pp. 315–322. Wu, C.G., Tseng, H.Y., Chen, Z.C., 1982. Fungi inhabiting on Schoenoplectus triqueter (L.) Palla (I). Taiwania 27, 35–38. Zhu, P., Ge, Q., Xu, T., 1991. The perfect stage of Pestalosphaeria from China. Mycotaxon 50, 129–140.