15
Phylogenetic relationships of Pestalotiopsis and allied genera inferred from ribosomal DNA sequences and morphological characters Rajesh Jeewon, a, * Edward C.Y. Liew, b and Kevin D. Hyde a a 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 School of Land, Water and Crop Sciences, McMillan Building A05, The University of Sydney, NSW 2006, Australia Received 21 August 2001; received in revised form 8 November 2001 Abstract The taxonomy of the coelomycetous fungus Pestalotiopsis and other closely related genera based on morphological characters has been equivocal. To gain insight in the phylogenetic relationships of Pestalotiopsis and its allies, part of the large subunit (28S) ribosomal DNA region was examined and compared with existing morphological information. Phylogenetic analyses were con- ducted using parsimony, distance, and likelihood criteria. Results of the analyses showed that Bartalinia, Pestalotiopsis, Seima- tosporium, and Seiridium represent distinct monophyletic groups with high bootstrap support. However, Truncatella species are paraphyletic with Bartalinia, sharing a common ancestor. Pestalotia species sequenced clustered together with Pestalotiopsis. These genera should be recognized as distinct natural groups except for Monochaetia and Discosia, which need to be further resolved. Tree topologies are generally in concordance with previous morphological hypotheses, most notably the placement of all Pestalotia species, except the type P. pezizoides, in Pestalotiopsis. Well-supported clades corresponding to groupings based on conidial morphology were resolved and the relative importance of morphological characters for generic delimitation is discussed. Molecular data also provide further evidence to support the association of these coelomycetes with the Amphisphaeriaceae. Ó 2002 Elsevier Science (USA). All rights reserved. Keywords: Pestalotia; Coelomycetes; Amphisphaeriaceae; Ribosomal DNA 1. Introduction The genus Pestalotiopsis Steyaert is a heterogeneous group of coelomycetous fungi where species are defined primarily based on conidial characteristics including size, septation, pigmentation, and presence or absence of appendages (Nag Rag, 1993; Steyeart, 1949; Sutton, 1980). Other fungal genera such as Bartalinia Tassi, Discosia Libert, Monochaetia (Saccardo) Allescher, Pestalotia de Notaris, Seimatosporium Corda, Seiridium Nees: Fries, and Truncatella Steyeart, however, also possess morphological characters very similar to those of Pestalotiopsis, resulting in considerable ambiguity and confusion in intergeneric classification of these fungi. Most species in these genera have morphological characters that overlap in many respects. These char- acters include number of median cells, which may or may not be pigmented, presence of apical and basal appendages, and hyaline apical or basal cells. The clas- sification, validity, and delimitation of these genera have been problematic and have been resolved differently by various authors (Guba, 1961; Nag Rag, 1993; Steyeart, 1949; Sutton, 1980). Guba (1929, 1961) adopted a broad generic concept by synonymizing Pestalotiopsis and Truncatella with Pestalotia. He divided the genus Pestalotia into three sections designated Quadriloculatae, Quinqueloculatae, and Sexloculatae for three-septate, four-septate, and five-septate spores, respectively. The same subdivisions were applied to Monochaetia (species having only one apical appendage), which was treated as congeneric with Seiridium. The main feature on which he relied for his system of classification was the number of apical ap- pendages and he stated that the distinction between Molecular Phylogenetics and Evolution 25 (2002) 378–392 MOLECULAR PHYLOGENETICS AND EVOLUTION www.academicpress.com * Corresponding author. Fax: +852-251-760-82. E-mail addresses: [email protected] (R. Jeewon), kdhyde@ hkucc.hku.hk (K.D. Hyde). 1055-7903/02/$ - see front matter Ó 2002 Elsevier Science (USA). All rights reserved. PII:S1055-7903(02)00422-0

Phylogenetic relationships of Pestalotiopsis and allied genera inferred from ribosomal DNA sequences and morphological characters

Embed Size (px)

Citation preview

Phylogenetic relationships of Pestalotiopsis and allied genera inferredfrom ribosomal DNA sequences and morphological characters

Rajesh Jeewon,a,* Edward C.Y. Liew,b and Kevin D. Hydea

a Centre for Research in Fungal Diversity, Department of Ecology and Biodiversity, The University of Hong Kong, Pokfulam Rd,

Hong Kong, SAR, People�s Republic of Chinab School of Land, Water and Crop Sciences, McMillan Building A05, The University of Sydney, NSW 2006, Australia

Received 21 August 2001; received in revised form 8 November 2001

Abstract

The taxonomy of the coelomycetous fungus Pestalotiopsis and other closely related genera based on morphological characters

has been equivocal. To gain insight in the phylogenetic relationships of Pestalotiopsis and its allies, part of the large subunit (28S)

ribosomal DNA region was examined and compared with existing morphological information. Phylogenetic analyses were con-

ducted using parsimony, distance, and likelihood criteria. Results of the analyses showed that Bartalinia, Pestalotiopsis, Seima-

tosporium, and Seiridium represent distinct monophyletic groups with high bootstrap support. However, Truncatella species are

paraphyletic with Bartalinia, sharing a common ancestor. Pestalotia species sequenced clustered together with Pestalotiopsis. These

genera should be recognized as distinct natural groups except forMonochaetia and Discosia, which need to be further resolved. Tree

topologies are generally in concordance with previous morphological hypotheses, most notably the placement of all Pestalotia

species, except the type P. pezizoides, in Pestalotiopsis. Well-supported clades corresponding to groupings based on conidial

morphology were resolved and the relative importance of morphological characters for generic delimitation is discussed. Molecular

data also provide further evidence to support the association of these coelomycetes with the Amphisphaeriaceae.

� 2002 Elsevier Science (USA). All rights reserved.

Keywords: Pestalotia; Coelomycetes; Amphisphaeriaceae; Ribosomal DNA

1. Introduction

The genus Pestalotiopsis Steyaert is a heterogeneous

group of coelomycetous fungi where species are defined

primarily based on conidial characteristics including

size, septation, pigmentation, and presence or absence of

appendages (Nag Rag, 1993; Steyeart, 1949; Sutton,

1980). Other fungal genera such as Bartalinia Tassi,

Discosia Libert, Monochaetia (Saccardo) Allescher,Pestalotia de Notaris, Seimatosporium Corda, Seiridium

Nees: Fries, and Truncatella Steyeart, however, also

possess morphological characters very similar to those

of Pestalotiopsis, resulting in considerable ambiguity

and confusion in intergeneric classification of these

fungi. Most species in these genera have morphological

characters that overlap in many respects. These char-acters include number of median cells, which may or

may not be pigmented, presence of apical and basal

appendages, and hyaline apical or basal cells. The clas-

sification, validity, and delimitation of these genera have

been problematic and have been resolved differently by

various authors (Guba, 1961; Nag Rag, 1993; Steyeart,

1949; Sutton, 1980).

Guba (1929, 1961) adopted a broad generic conceptby synonymizing Pestalotiopsis and Truncatella with

Pestalotia. He divided the genus Pestalotia into three

sections designated Quadriloculatae, Quinqueloculatae,

and Sexloculatae for three-septate, four-septate, and

five-septate spores, respectively. The same subdivisions

were applied to Monochaetia (species having only one

apical appendage), which was treated as congeneric with

Seiridium. The main feature on which he relied for hissystem of classification was the number of apical ap-

pendages and he stated that the distinction between

Molecular Phylogenetics and Evolution 25 (2002) 378–392

MOLECULARPHYLOGENETICSANDEVOLUTION

www.academicpress.com

*Corresponding author. Fax: +852-251-760-82.

E-mail addresses: [email protected] (R. Jeewon), kdhyde@

hkucc.hku.hk (K.D. Hyde).

1055-7903/02/$ - see front matter � 2002 Elsevier Science (USA). All rights reserved.

PII: S1055 -7903 (02 )00422-0

Pestalotia and Monochaetia rested primarily upon thenumber of apical appendages arising from the apical

cell.

This broad generic concept of Pestalotia was highly

criticized by Steyeart (1949) who erected two new gen-

era, Pestalotiopsis (for the group of fungi known as

Pestalotia section Quinqueloculatae) and Truncatella

(for the group of fungi known as Pestalotia section

Quadriloculatae). He also discarded the genus Mono-

chaetia and reassigned the species in Pestalotia into

these two new genera. Steyeart (1949) preferred to keep

the genus Pestalotia monotypic, which was represented

by Pestalotia pezizoides. This species is morphologically

distinct from all Pestalotiopsis species, as it possesses

cupulate conidiomata and distoseptate median cells. By

keeping Pestalotia monotypic, he divided the genus

Pestalotiopsis and Truncatella into different sectionsbased on the number of apical appendages. All the

species bearing single apical appendages (e.g., Mono-

chaetia) were not given generic status and were trans-

ferred to the section Monosetulatae of Pestalotiopsis

and Truncatella, thereby contradicting Guba�s treat-ment. Steyeart (1953a,b, 1961, 1963) published further

evidence in support of his contradictions despite major

criticisms by Guba (1955, 1956). Guba (1955) disagreedwith Steyeart (1949) and added his objections to Steye-

art�s treatment with substantial reasons and argued thatall of the new generic designations proposed by Steyeart

(1949) are related to synonymy. He further explained

that the presence of a single apical appendage in

Monochaetia species clearly sets them apart from Pest-

alotia and that Monochaetia merits generic status.

Steyeart (1956) pointed out that it was unwise to givegeneric status to Monochaetia based only on one mor-

phological character (i.e., the presence of a single apical

appendage), especially considering that Pestalotia, Pes-

talotiopsis, and Truncatella, which differ from each other

by a set of characters, were not given generic status.

However Guba (1961) and Steyeart (1949) did not

consider other genera such as Bartalinia, Discosia, and

Seimatosporium, which have close affinities to Pestal-

otiopsis.

Arx (1981) treated Bartalinia as congeneric to Se-

imatosporium but such a synonymy was not accepted by

Nag Rag (1993) who, on morphological grounds, ar-

gued that these two genera are distinct anamorphic

genera. The taxonomic history and complexity of Pes-

talotiopsis and its allies have been debated for over half a

century (Arx, 1981; Guba, 1955; Roberts and Swart,1980; Steyeart, 1949; Sutton, 1969, 1980) and more re-

cently by Nag Rag (1993). All previous studies on these

genera relied heavily on morphological characters as the

main criteria for generic delimitation without taking

into consideration phylogenetic relationships. Following

Steyeart�s concept, it has been necessary to reassign allfour-septate conidia (three median cells) to Pestalotiop-

sis, a view supported by Sutton (1969, 1980), but NagRag (1993) preferred to adopt a wider generic concept of

Pestalotiopsis by including three-septate (two median

cells) forms originally assigned to Truncatella. Sutton

(1969) discussed the validity of Pestalotiopsis and its

relationships with Pestalotia, Monochaetia, and Seiri-

dium. He favored Steyeart�s treatment but expanded thegeneric concepts of Monochaetia and Seiridium to in-

clude species with single apical appendages. The pres-ence of a single apical appendage was thought to be a

hallmark for Monochaetia and Seiridium, but the latter

merited generic status as the conidia are distoseptate.

whereas those of Monochaetia are euseptate (Nag Rag,

1993; Sutton, 1969). It has been recently shown that

conidium wall structure inMonochaetia and Seiridium is

different, the latter having thick-walled median cells,

whereas the former possesses a less elaborate structurewith thinner-walled and lighter-colored median cells

(Roberts and Swart, 1980).

Griffiths and Swart (1974a,b) investigated the devel-

opment of conidia in two species of Pestalotiopsis and in

P. pezizoides by means of electron microscopy in an

attempt to establish the affinities of these genera with

other members of Monochaetia and Seimatosporium.

Their results were congruent with that of Sutton (1969)that the difference in conidial wall structure with a more

elaborate zonation in P. pezizoides clearly distinguishes

it from Pestalotiopsis. Nag Rag (1993) suggested that

species accepted by Guba (1961) in Monochaetia and

Pestalotia section Sexloculatae may well belong to Se-

iridium and that a large number of taxa which properly

belong in Pestalotiopsis still remain in Pestalotia.

An important step toward a more natural classifica-tion of these fungi was taken by Sutton (1980) where

conidiogenesis was given more weighting as a taxonomic

character. However, all these genera exhibit similar ap-

pendage morphogenesis (Nag Rag, 1993) and conidio-

genesis (Sutton, 1961, 1980). Other developmental

studies including conidium ontogeny in Pestalotiopsis

neglecta (Jones, 1977) and conidiomatal development in

Pestalotiopsis (Watanabe et al., 1998) and Bartalinia

(Roux and Warmelo, 1990) have so far failed to eluci-

date relationships among these genera. Recently, Mor-

gan et al. (1998) explored the ability of artificial neural

networks to identify Monochaetia, Pestalotiopsis, and

Truncatella, but the results were not convincing enough

to assess the validity of these genera. Despite the fact

that these genera are currently being treated as distinct

taxa, the wide generic concept proposed by Guba (1961)has never been tested on other grounds except conidial

morphology.

Given the considerable taxonomic confusion of these

genera, this study was undertaken to address several

questions: (i) Do these genera represent natural groups?;

(ii) Which morphological characters are phylogeneti-

cally significant and are therefore useful for generic de-

R. Jeewon et al. / Molecular Phylogenetics and Evolution 25 (2002) 378–392 379

lineation?; and (iii) Are phylogenies based on molecularcharacters concordant with traditional morphology-

based classification schemes?

2. Materials and methods

2.1. Cultures and DNA extraction

Our taxonomic sampling included a total of 32 spe-

cies representing seven genera. Species names, accession

numbers, and geographical origin of the isolates are

listed in Table 1. For each isolate, single-spore cultures

were plated on potato dextrose agar and incubated at

25 �C for 10–20 days. Spores were examined under the

microscope to verify the authenticity of the fungi under

investigation. Some isolates were given different treat-ments to induce sporulation. Living cultures have been

deposited at The University of Hong Kong Culture

Collection. Genomic nucleic acids were extracted from

fresh fungal mycelia following a modified protocol of

Doyle and Doyle (1987). Briefly, mycelia were scraped

off from the surface of the plate, and ground with

200mg of sterilized quartz sand and 600 ll of 2� CTAB

extraction buffer (2% w/v CTAB, 100mM Tris–HCl,1.4M NaCl, 20mM EDTA, pH 8) in a 1.5-ml Eppen-

dorf tube. The whole contents were incubated at 60 �C ina water bath for 40min with occasional swirling. The

solution was then extracted two or three times with an

equal volume of phenol:chloroform (1:1) at 14,000g for

30min until no interface was visible. The upper aqueous

phase containing the DNA was precipitated by addition

of 2.5 volumes of absolute ethanol and kept at )20 �Covernight. The precipitated DNA was then washed with

70% ethanol, dried under vacuum, suspended in TE

buffer (1mM EDTA, 10mM Tris–HCl, pH 8), and

treated with RNase (1mg ml�1).

2.2. Amplification of genomic DNA

A fragment of DNA, spanning approximately 900 bpof the 50 end of the 28S ribosomal gene, was symmetricallyamplified using one pair of primers LROR (50-ACCCGCTGAACTTAAGC-30; Vilgalys and Hester, 1990)and LR5 (50-TCCTGAGGGAAACTTCG-30; Vilgalysand Hester, 1990). Amplification was performed using 2–

3 ll of genomic DNA in a standard 50-ll PCR mixture

(25mM MgCl2, 10� Mg-free buffer, 2.5 lM dNTPs,

1.5 lM primers, and 1 unit of Taq Polymerase) under thefollowing thermal conditions: 94 �C for 3min, 94 �C for50 s, 30 cycles of 94 �C for 50 s, 50 �C for 1min, and 72 �Cfor 1.5min, with a final extension step of 72 �C for 10min.Amplified products were visualized on 1% agarose gel

electrophoresis (stained with ethidium bromide) under

UV light to check for size and purity. Negative control

reactions omitting DNA were included in all sets of am-

plifications to monitor for potential contamination byexogenous DNA. PCR products were purified using mi-

nicolumns, purification resin, and buffer according to the

manufacturer�s protocol (Wizard PCR Preps DNA Pu-

rification System).

2.3. DNA sequencing and alignment

Both strands of the purified products were directly se-quenced in an automated sequencer (ALF Express,

Pharmacia-Biotech, Piscataway, NJ) following the man-

ufacturer�s protocol. Four sequencing primers were used:LROR (50-ACCCGCTGAACTTAAGC-30; Vilgalys andHester, 1990) and LR3R (50-GTCTTGAAACACGGACC30: Vilgalys and Hester, 1990) for sequences read-ing in the 50–30 direction and LR5 (50-TCCTGAGGGAAACTTCG: Vilgalys and Hester, 1990) and LR3 (50-CCGTGTTTCAAGACGGG: Vilgalys and Hester, 1990)

for sequences reading in the 30–50 direction. In addition tothe ingroup used, Pleospora herbarum var herbarum and

Dothidea sambuci were also sequenced and used as the

outgroup taxa to test potential monophyly and root the

cladograms. Reference sequences from different fungal

orders obtained from GenBank were also included in the

analysis. Sequences from each strain were assembled us-ing ALF software and SeqPup (Gilbert, 1996) to obtain

the entire sequence flanked by primers LR5 and LROR.

Initial alignments of the partial 28S rDNA from different

isolates were performed using the multiple alignment

program Clustal X (Thomson et al., 1997). Alignments

were checked and then manually edited where necessary.

2.4. Phylogenetic analysis

Maximum-parsimony (MP). Phylogenetic analyses

were performed using PAUP* 4.0b8 (Swofford, 2001).

Parsimony trees were obtained using heuristic searches

only because of the large data set. To increase the

probability of finding all most parsimonious trees,

searches were implemented using the random sequence

addition option and the tree bisection–reconnection(TBR) branch-swapping algorithm. Each search was

repeated 10 times from different random starting points

using the stepwise addition option. Single-position gaps

were treated as missing data. For phylogenetic analysis

each homologous sequence position was treated as a

discrete character with four possible unordered states

(A, G, C, or T), and equally weighted parsimony (with a

transition:transversion ratio of 1:1) was included in theparsimony analysis. A series of minor analyses under

different conditions (different transition:transversion

ratios and treating gaps as missing or fifth state) was

carried out to test the phylogenetic relationships among

the taxa and to determine the most reliable parameters

giving the best trees for subsequent analyses. Branch

support of the trees resulting from maximum-parsimony

380 R. Jeewon et al. / Molecular Phylogenetics and Evolution 25 (2002) 378–392

was assessed by bootstrapping (Felsenstein, 1985; San-

derson, 1989). The bootstrap analysis was performed

with 1000 replications using the heuristic search option

as described above to estimate the reliability of inferred

monophyletic groups. Consistency index (CI), retentionindex (RI), rescaled consistency index (RC), and ho-

moplasy index (HI) were calculated for all parsimony

trees.

In addition to MP with changes among character

states having unequal weights, alternative using analyses

were conducted based on the maximum-likelihood and

distance methods using PAUP*.

Maximum-likelihood (ML). For ML analyses, we

estimated parameters and trees using an iterative ap-

proach. A strict consensus tree was selected from an

earlier MP analysis as the starting tree for the maxi-

mum-likelihood analyses. A tree generated under thedistance criterion was also used as a starting tree for

maximum-likelihood. Two models of nucleotide substi-

tution, the HKY (Hasegawa et al., 1985) and F84

(Felsenstein, 1984) were used. The gamma model of site

rate variation was used with no enforcement of a mo-

lecular clock. Transition:transversion ratios and base

frequencies were estimated and initial branch lengths

Table 1

Fungal strains used in the study and their accession numbers, hosts, and locality

Taxon Source of culturesa Host/geographic origin GenBank Accession No. (LSU/ITS)

Ingroup

Bartalinia robillardoidesc BRIP 14180 Macrotyloma daltonii, Australia AF382366/AF405301

Bartalinia biscofiae HKUCC 6534 Unidentified dead leaf, Hong Kong AF382367

Bartalinia lateripes HKUCC 6654 Unidentified dead leaf, Hong Kong AF382368

Bartalinia laurinac HKUCC 6537 Unidentified dead leaf, Hong Kong AF382369/AF405302

Discosia sp.c HKUCC 6626 Unidentified dead leaf, Hong Kong AF382381/AF405303

Discostroma sp. HKUCC 1004 Unidentified terrestrial wood, Hong Kong AF382380

Lepteutypa cupressi IMI 052255 Cupressus forbesii, Kenya AF382379

Monochaetia monochaeta CBS 199.82 Quercus pubescens, Italy AF382370

Monochaetia karsteniic ICMP 10669 Camellia sp., New Zealand AF382371/AF405300

Pestalotia photiniae ICMP 10737 Photinia sp., New Zealand AF382363

Pestalotia sp. 1 ICMP 3062 Prunus domestica, New Zealand AF382364

Pestalotia vaccinii ICMP 5446 Vaccinium sp., New Zealand AF382362

Pestalotia sp. 2 ICMP 5476 Acca sellowiana, New Zealand AF382365

Pestalotia palmarumc ATCC 10085 Coconut Palm, India AF382361/AF009818

Pestalotiopsis maculansc CBS 322.76 Camellia sp., France AF382354/AF405296

Pestalotiopsis sp.c HKUCC 7982 Protea neriifolia, S. Africa AF382355/AF405297

Pestalotiopsis bilicia HKUCC 7983 Leucospermum sp., S. Africa AF382356

Pestalotiopsis versicolorc BRIP 14534 Psidium guajava, Australia AF382357/AF405298

Pestalotiopsis funeraec ICMP 7314 Cupressocyparis leylandii, New Zealand AF382358/AF405299

Pestalotiopsis sp. EN 8c HKUCC 7984 Scaevola hainanensis, Hong Kong AF382359/AF405294

Pestalotiopsis sp. EN 10c HKUCC 7985 Scaevola hainanensis, Hong Kong AF382360/AF405295

Seimalosporium grevillaec ICMP 10981 Protea sp., S. Africa AF382372/AF405304

Seimatosporium leptospermi ICMP 11845 Leptospermum scoparium, New Zealand AF382373

Seimatosporium vaccinii ICMP 7003 Vaccinium ashei Reade, New Zealand AF382374

Seimatosporium sp. HKUCC 7986 Leucospermum sp., S. Africa AF382375

Seiridium cardinale 1 CBS 172.56 NAb AF382376

Seiridium cardinale 2c ICMP 7323 Cupressocyparis leylandii, New Zealand AF382377/AF405305

Seiridium cupressi FABI, CMW 5596 Cupressus sempervirens, S. Africa AF382378

Truncatella angustatac ICMP 7062 Malus X domestica, New Zealand AF382383/AF405306

Truncatella conorum piceae ICMP 11213 Cedrus deodara, New Zealand AF382384

Truncatella laurocerasi ICMP 11214 Prunus persica, New Zealand AF382385

Truncatella sp. HKUCC 7987 Leucospermum sp., S. Africa AF382382

Outgroup

Xylaria hypoxylonc ATCC 42768 NA U47841/AF201711

Diaporthe phaseolorum NA U47830

Hypocrea schweinitzii NA U47833

Ophiostoma piliferum NA U47837

Pleospora herbarum var

herbarumc

CBS 191.86 Medicago saliva, India AF382386/AB026165

Dothidea sambuci CBS 198.58 Acer pseudoplatanus, Switzerland AF382387

aATCC, American Type Culture Collection; BRIP, Queensland Department of Primary Industries Plant Pathology Herbarium; CBS, Centra-

albureau voor Schimmelcultures; FABI, Forestry and Agricultural Biotechnology Institute; HKUCC, The University of Hong Kong Culture

Collection; ICMP, International Collection of Microorganisms from Plants.bNA, information not available.c Strains used in combined ITS/LSU analysis.

R. Jeewon et al. / Molecular Phylogenetics and Evolution 25 (2002) 378–392 381

were obtained using Rogers–Swofford approximationmethods. Using these initial estimates of substitution

rates and kinds a heuristic search with TBR branch

swapping was used to find a maximum-likelihood tree.

Neighbor-joining (NJ). In addition to MP and ML,

phenetic trees were constructed from distance matrix

values by the neighbor-joining method (Saitou and Nei,

1987) to reflect DNA sequence similarity. Based on the

assumptions about rates and nucleotide substitutionmodels, NJ trees were constructed under a variety of

distance measures including HKY (Hasegawa et al.,

1985), K2P (Kimura, 1980), JC (Jukes and Cantor,

1969), LogDet (Lockhart et al., 1994), and GTR (Lan-

ave et al., 1984; Rodriquez et al., 1990). One-thousand

neighbor-joining bootstrap replicates were performed

on the 28S data set. Gaps were treated as missing data

and all characters were given equal weight.Phylogenetic analysis and evaluating congruence of the

combined data set. Phylogenetic analyses (MP and ML)

based on a combined data set of the complete ITS 1,

5.8S, ITS 2, and partial LSU sequences were also con-

ducted. The combined data set consisted of 1575 bases

for 14 Pestalotiopsis isolates with Xylaria hypoxylon and

Pleospora herbarum as the outgroups. To find as many

equally parsimonious trees as possible, we used 10 rep-licated heuristic searches with random addition of taxa,

and support for clade stability was estimated using

nonparametric bootstrapping with 1000 replicates. The

partition homogeneity test (Cunningham, 1997; Farris

et al., 1995), as implemented in PAUP*, was used to

examine data for conflicting hierarchic signals and to

evaluate congruence of the combined data set. Most

parsimonious trees, using the same parameters as de-scribed above for MP analyses, were used to construct

trees for the combined data set. ML and NJ trees were

also generated by the same procedure as described

above and compared with MP trees.

Estimation of topological differences between trees.

Kishino–Hasegawa tests (Kishino and Hasegawa, 1989)

and Templeton tests (Templeton, 1983), as implemented

in PAUP*, were performed to determine whether thetrees inferred from the different tree-building methods

were significantly different. Trees were viewed in Tree-

view (Page, 1996). The nucleotide sequences reported in

this paper have been deposited in GenBank.

3. Results

3.1. Phylogenetic analyses of the large subunit (LSU)

data set

Maximum-parsimony (MP): The final LSU rDNA

alignment included 885 characters of which 178 (20.1%)

were parsimony-informative sites. The maximum-parsi-

mony analysis of the LSU data set with alignment gaps

treated as missing data with no differential weighting oftransitions against tranversions, and with random-addi-

tion sequence and TBR branch swapping yielded 10 most

parsimonious trees of tree length (TL) 573 steps with CI,

RI, RC, and HI of 0.682, 0.792, 0.541, and 0.318, re-

spectively. The strict consensus of the equally most par-

simonious tree is shown in Fig. 1. When a weighted

parsimony (transition:transversion ratios of 2:1 and 1:5)

was applied to the same data set, the tree topologywas thesame but with longer tree lengths of 810 and 619.5 steps,

respectively. Gaps treated as a fifth character with

weighted or unweighted parsimony had no effect on tree

topology (data not shown). Separate heuristic searches

implementing different addition sequences and swapping

strategies gave identical results. Bootstrapping with 1000

replicates provided good support for all the terminal

clades, and all the branches received more than 60%bootstrap support except for the node comprising

Monochaetia monochaeta (51%) (Fig. 1). All the genera

form distinct monophyletic clades with high bootstrap

values ð> 90%Þ except Truncatella, which appears to beparaphyletic. Pestalotiopsis and Pestalotia together form

a putative monophyletic group supported by a 94%

bootstrap value (Fig. 1).

To assess tree outputs, we divided the ingroup taxa inthe cladogram into seven clades (A–G). Generally, each

genus clustered separately from the rest except for Pest-

alotia and Discosia. Pestalotia clustered together and in-

termingled with Pestalotiopsis in one clade (clade A) with

a high bootstrap support (94%). Clade A can also be

further subdivided into two subclades, the top one ap-

pearing in 89% of the bootstrap replicates, whereas the

second appeared in 71%.M. monochaeta (Clade B) formsamonophyletic clade as the sister group to the preceeding

taxa of Pestalotiospsis but the confidence for the support

is quite low (51% bootstrap value). Five of the seven

clades have highbootstrap values, ð> 93%) but two clades

(B and G) have significantly lower bootstrap values (51%

and 88%, respectively). Clades C and D, also monophy-

letic, include Seiridium isolates together with Lepteutypa

cupressi (appears monophyletic in 93% of the 1000bootstrap replicates) and Bartalinia isolates (appears

monophyletic in 99% of the 1000 bootstrap replicates),

respectively. Conversely, clade E, which comprises only

Truncatella species, appears to be paraphyletic with re-

spect to the other ingroup. Interestingly, Truncatella

species appear to be very closely related to Bartalinia

species and this is supportedby abootstrap value of 100%.

Clade F contains Seimatosporium species with Discost-

roma and forms a distinct monophyletic group with a

bootstrap value of 97%. Sequential deletion of reference

taxa or use of different taxa as reference taxawas observed

to have no effect on the topology of the different phylog-

enies. By removing the loculoascomycetes (P. herbarum

var herbarum and D. sambuci) and members from the

other different orders (Diaporthe phaesolorum, Hypocrea

382 R. Jeewon et al. / Molecular Phylogenetics and Evolution 25 (2002) 378–392

schweinitzii, and Ophiostoma piliferum) except X. hypox-

ylon, it was found that the data set produced the same

topologies.

Maximum-likelihood (ML). Based on the likelihood

analyses under the HKY model with an estimated shape

parameter of 0.2487 and estimated transition:transver-

Fig. 1. Strict consensus of 10 equally most parsimonious trees generated from MP analysis of partial 28S rDNA gene sequences. Numbers above the

branches indicate bootstrap values from an anaylsis with 1000 replicates. Schematic representations on the righthand side show the significant

morphological characters traditionally used to distinguish these genera. Letters A–G represent each distinct genus. Designated outgroups are

Pleospora herbarum var herbarum and Dothidea sambuci.

R. Jeewon et al. / Molecular Phylogenetics and Evolution 25 (2002) 378–392 383

sion ratio of 1.6168, a ML tree of )loglikelihood4131.5538 was obtained (Fig. 2). Estimated base fre-

quencies were as follows: A ¼ 0:228, C ¼ 0:24,G ¼ 0:284, and T ¼ 0:248. The ML tree derived from

the HKY and F84 (Felsenstein, 1984) models were

identical and perfectly matched the MP tree topology.

Under the F84 model of nucleotide substitution, the

topology of the tree with respect to the ingroup was thesame except that X. hypoxylon did not form a separate

clade; instead, it clustered together with D. phaseolorum

and H. schweinitzii (results not shown).

Neighbor-joining (NJ). The dendograms generated

were similar under different models (HKY, K2P, JC,

and GTR) and the data confirmed the monophyly of the

Fig. 2. Cladogram generated from a ML search from partial 28S rDNA sequence data with the HKY substitution model (length¼ 573; CI¼ 0.682;Ln-likelihood¼)4141.1252). Designated outgroups are Pleospora herbarum var herbarum and Dothidea sambuci. Letters A–G represent each distinctgenus.

384 R. Jeewon et al. / Molecular Phylogenetics and Evolution 25 (2002) 378–392

different genera. Minimum-evolutionary distance anal-

ysis under the HKY model resulted in a single best tree

with slightly higher bootstrap values than those of the

other models and with a )loglikelihood of 4156.9816(Fig. 3). This tree is topologically identical to those

obtained in the MP and ML analyses except that M.

monochaeta positions itself as a sister group to the Se-

iridium clade. The estimate of phylogeny based on a

rooted tree indicates high levels of confidence ð> 80%Þ

for internal branches uniting species from the different

genera. NJ analysis confirmed the monophyly for the

different genera under investigation except for Mono-

chaetia. All models used for NJ analysis suggest that the

actual taxonomic status and the phylogenetic relation-

ships of M. monochaeta with its allies are doubtful. TheNJ trees positioned M. monochaeta as a basal, inter-

mediate lineage between Seiridium and Bartalinia but

this did not receive much bootstrap support (55%).

Fig. 3. Neighbor-joining tree generated from partial 28S rDNA gene sequences under the HKY substitution model. Values above branching nodes

indicate bootstrap support obtained from bootstrap analysis with 1000 replicates. Designated outgroups are Pleospora herbarum var herbarum and

Dothidea sambuci. Letters A–G represent clades for members of each genus.

R. Jeewon et al. / Molecular Phylogenetics and Evolution 25 (2002) 378–392 385

Contradictorily, the MP tree placedM. monochaeta as asister group of Pestalotiopsis with a bootstrap value of

51%.

3.2. Phylogenetic analyses and congruence of the com-

bined data set

The partition homogeneity test resulted in a P value

of 0.02, indicating that the combined data set was con-gruent and combinable. The phylogram generated by a

weighted MP analysis (transition:transversion ratio of

1.5:1) and treating gaps as missing data helps to resolve

intergeneric relationships. Two trees with 242 parsimo-

ny-informative sites were obtained (TL ¼ 1021:25,CI ¼ 0:786, RI ¼ 0:756, RC ¼ 0:590, HI ¼ 0:214, Lnlikelihood¼)6308.6580) (Fig. 4). This combined anal-ysis of the two rDNA regions generated a well-resolvedand strongly supported phylogeny that is topologically

congruent with the other trees. The results clearly sup-

port that all the genera are monophyletic as they form

distinct groupings with high levels of confidence (Fig. 4).

Discosia clusters together with Seimatosporium as in the

MP, ML, and NJ trees. An unweighted MP analysis

treating gaps as new states yielded one most parsimo-

nious tree with similar tree topology (results not shown).The ML tree obtained by analyzing the combined data

set turned out to be congruent with the MP tree (data

not shown).

3.3. Comparison between topologies of MP, ML, and NJ

trees

Both Kishino–Hasegawa and nonparametric tests

(Templeton test) used to evaluate trees showed that the

NJ tree is significantly different from the MP and ML

trees and is therefore rejected (Table 2).

4. Discussion

Steyeart (1949) established the genera Pestalotiopsis

and Truncatella to accommodate some species previ-

ously disposed of in Pestalotia and Monochaetia. He

retained Pestalotia as monotypic and discarded Mono-

chaetia. In contrast, Guba (1961) adopted a wide generic

concept of the genus Pestalotia by synonymizing Pes-

talotiopsis and Truncatella to Pestalotia and synony-

mizing Seiridium toMonochaetia. Arx (1981) treated the

genus Bartalinia as a synonym of Seimatosporium but,

based on morphology, Nag Rag (1993) disagreed be-

cause these two anamorphic genera are quite distinct.

Whether the proposed synonym of Arx (1981) is valid

was highly contentious. The main aim of this study was

to test the morphology-based hypotheses of Steyeart(1949) and Guba (1961) using molecular data.

Based on our current data, the wide generic concept

proposed by Guba (1961) and the treatment of Bartalinia

Fig. 4. Phylogenetic tree of 16 taxa estimated from a combined data set of partial nuclear LSU rDNA, ITS, and 5.8S rDNA sequences under the

maximum-parsimony optimality criterion. Pleospora herbarum is the designated outgroup. Bootstrap values generated from 1000 replicates are

shown above the branches.

386 R. Jeewon et al. / Molecular Phylogenetics and Evolution 25 (2002) 378–392

and Truncatella as synonyms of Seimatosporium and

Pestalotia, respectively (Arx, 1981), are rejected. How-

ever, findings are partially congruent with Steyeart�sview, which is in agreement with the commonly accepted

taxonomic classification (Nag Rag, 1993; Sutton, 1961,

1980). The traditional morphological characters used to

define the genera have been shown on the cladogram

inferred from analysis of DNA sequences (Fig. 1). Spe-cial emphasis has been placed on the different morpho-

logical characters that are typical of each genus. In

essence, topologies of all trees support recognizing

Bartalinia, Pestalotiopsis, Seimatosporium, Seiridium,

and Truncatella as natural groups. While our data do

not reject Discosia and Monochaetia as natural and

distinct genera, further studies are required to support

this.

4.1. Monophyly of Pestalotiopsis and Pestalotia

Our current results show that species of Pestalotiopsis

and Pestalotia under investigation are closely related as

they cluster together in a single monophyletic clade with

high bootstrap support. All these species are character-

ized by fusiform or slightly curved conidia bearing four-to five-euseptate median cells that are pigmented, with

apical appendages that arise as tubular extensions from

the apical cell, and a centric basal appendage arising

endogenously from the basal cell. The Pestalotia species

examined differ slightly in morphology from those of

Pestalotiopsis in that the spores possess median cells that

have slightly thicker walls and are doliform in shape,

mostly verrucose in ornamentation, and guttulate. Theyalso have a slower growth rate on synthetic media as

compared to Pestalotiopsis species. Most probably, the

fact that these median cells have thicker walls and ap-

pear somehow distoseptate might have confused some

mycologists during identification. As indicated in our

results and previous morphological hypotheses (Nag

Rag, 1993; Sutton, 1969, 1980) these species should be

Pestalotiopsis species as the median cells are not disto-septate as in Pestalotia. It is also clear that Pestalotia

species not having distoseptate conidia should be

transferred to Pestalotiopsis, as proposed by Steyeart

(1949), unless they possess other distinctive featurescharacteristics of other genera.

The taxonomic position of Pestalotiopsis has been

controversial (Guba, 1961; Steyeart, 1949). Steyeart

(1949) proposed that all Pestalotia species should be

transferred to Pestalotiopsis and that Pestalotia should

be monotypic with P. pezizoides, while Guba (1961)

advocated that Pestalotiopsis should be reduced to

synonymy. The molecular data are concordant withSteyeart�s treatment as all the Pestalotia species se-

quenced clustered together with Pestalotiopsis species.

P. pezizoides could not be included in our study as no

culture was available. Various attempts were made to

extract and amplify DNA from the dried specimens but

all our attempts were unsuccessful and yielded highly

damaged genomic DNA probably because the dried

material was too old. However the dried specimen wasexamined microscopically and it was found that it differs

from Pestalotiopsis in that it has got four-distoseptate

median cells with large lumens. Therefore it would not

be surprising that all Pestalotia species except P. pezi-

zoides properly belong to Pestalotiopsis.

One oddity in this monophyletic group is the clus-

tering ofMonochaetia karstenii within the Pestalotiopsis

clade. This culture was obtained from ICMP and mi-croscopic examination of the spores from the culture

revealed that this species actually produced spores with

a single apical appendage (characteristic of the genus

Monochaetia). Based on morphological similarities such

as apical appendage arising as a tubular extension from

the apical cell and median cells having thin walls, M.

karstenii was synonymized with Pestalotiopsis karstenii

by Nag Rag (1988, 1993). Based on molecular evidenceprovided here, we support this synonymy.

4.2. Monophyly of Seiridium

The presence of four-distoseptate median cells that

are dark brown and that of a single attenuated apical

appendage are believed to characterize species of Seiri-

dium. Guba (1961), Shoemaker et al. (1966), and Sutton(1969, 1975) have discussed the generic synonymy of

Seiridium. While Guba (1961) synonymized Seiridium

with Monochaetia by transferring all the Seiridium spe-

cies to the section Sexloculatae of Monochaetia, such

conservative treatment was not accepted by other

workers (Nag Rag, 1993; Roberts and Swart, 1980;

Sutton, 1969, 1980). The Seiridium isolates sequenced

form a monophyletic group in all our analyses. None ofthe trees obtained from our analyses support that Seir-

idium should be synonymized with Monochaetia. Even

though the bootstrap support for this relationship ap-

pears to be quite low, they are clearly distinct genera.

Therefore Guba�s synonymy of Seiridium is not vali-

dated. The presence of a single appendage is not a un-

ique feature characterizing Seiridium as a monophyletic

Table 2

Results of the Kishino–Hasegawa and templeton tests for the trees

generated by different optimality criteria

MP tree

(Fig. 1)

ML tree

(Fig. 2)

NJ tree

(Fig. 3)

Tree length (steps) 573 573 588

Consistency index 0.682 0.682 0.665

)Ln Likelihood 4141.2883 4131.5538 4156.9816

Kishino–Hasegawa test Pa¼ 0.1574 Best Pa¼ 0.0006Templeton test Pa¼ 0.5 Best Pa¼ 0.0005a Probability of getting a more extreme, t value under the null hy-

pothesis of no difference between the two trees (two-tailed test) with

significance at P < 0:05.

R. Jeewon et al. / Molecular Phylogenetics and Evolution 25 (2002) 378–392 387

group, as this character is also found inM. monochaeta,M. karstenii, and Seimatosporium species. Based on

morphological observations of the teleomorphs, Samu-

els et al. (1987) suggested a close affinity between Seir-

idium and Pestalotiopsis. This close relationship is

confirmed by our analyses as Seiridium forms the sister

group to Pestalotiopsis. The distoseptate feature of the

spore is, however, of particular phylogenetic relevance

in generic delimitation. It is noteworthy to point outhere that P. pezizoides is also characterized by disto-

septate conidia. Its affinities to Seiridium, however, are

yet to be resolved.

4.3. Monophyly of Bartalinia and Seimatosporium

Molecular results obtained here offer robust evidence

to support the distinct monophyletic status of Bartaliniaand Seimatosporium. There have been debates whether

to give Bartalinia generic status or treat it as congeneric

with Seimatosporium (Arx, 1981; Morgan-Jones et al.,

1972; Nag Rag, 1993) because Bartalinia species are

similar to Pestalotiopsis and Seimatosporium in having

euseptate median cells, two or three apical appendages,

and either exogenous or excentric basal appendages.

DNA analyses herein place the genus Bartalinia as amonophyletic group, distinct from Seimatosporium. This

finding does not corroborate the synonymy of Bartalinia

to Seimatosporium as proposed by Arx (1981), but is in

agreement with Nag Rag�s (1993) concept that Bartali-nia and Seimatosporium are distinct anamorphic genera.

Even though species from both genera possess basal

appendages that are excentric, this character is probably

evolutionarily convergent. Bartalinia is different fromSeimatosporium in that median cells are almost hyaline

or very pale brown (not pigmented in other genera),

with apical appendages arising from a particular locus

above the apical cell (not separated by a septum). Se-

imatosporium, on the other hand, is characterized by

having pigmented median cells, a mixture of appen-

daged and nonappendaged conidia with single apical

appendages, and excentric basal appendages.At the species level, Bartalinia laurina and B. biscofiae

cluster together in all trees as both of these species have

two hyaline median cells, whereas B. robillardoides and

B. lateripes have three hyaline median cells. This sug-

gests that the number of median cells may be phyloge-

netically important and a diagnostic character for the

segregation of species within Bartalinia but not for ge-

neric circumscription.

4.4. Paraphyly of Truncatella

Current results indicate that the genus Truncatella is

paraphyletic with Bartalinia sharing a common ances-

tor. Truncatella is characterized by having spores with

two brown or dark brown concolorous median cells

with thick walls and mostly irregularly branched apicalappendages. Although the results suggest that the

Truncatella species sampled are not monophyletic, they

do not support the previous morphological hypothesis

proposed by Guba (1961) that Truncatella is congeneric

with Pestalotiopsis.

Truncatella species appear to be very closely related

to Bartalinia species and this is inconsistent with the

morphology based hypotheses that Truncatella is relatedto Pestalotiopsis (Guba, 1961). The treatment proposed

by Arx (1981), that Bartalinia and Pestalotia should be

considered a synonym of Truncatella, is also refuted

here, as these two genera are quite distinct (based on our

study and on morphological characters; Nag Rag,

1993). Bartalinia is characterized by having only hyaline

median cells with axial apical appendages, which branch

once, whereas median cells are pigmented in Truncatellaand apical appendages arise usually irregularly from an

apical crest.

Our molecular results are also informative for clari-

fying relationships at the species level. Truncatella sp.

and T. conorum piceae clustered away from T. angustata

(type species) and T. laucocerasi. Truncatella sp. and T.

conorum piceae also possess two median cells but differ

from the other two species in terms of appendagecharacters. T. angustata and T. laucocerasi are charac-

terized by the presence of one or more than one irreg-

ularly and dichotomously branched apical appendages

at the apical cell with no basal appendages. Truncatella

sp. and T. conorum piceae possess only two unbranched

those similar to apical appendages found in Pestaloti-

opsis species. In addition, they possess basal append-

ages, which are absent in the other two species ofTruncatella studied. These unusual features separate

them from T. angustata and T. laucocerasi and are useful

in species delineation.

Another important conclusion from our molecular

analysis is that the wide generic concept of Pestalotiopsis

proposed by Nag Rag (1993) to accommodate species

with two and three median cells might be purely artificial.

Results herein do not support this hypothesis becauseTruncatella sp. and T. conorum piceae did not cluster with

Pestalotiopsis. However, our molecular study provides

good support for the classification established by Steyeart

(1949)who erected the genusTruncatella to accommodate

species with two median cells.

Interestingly, Truncatella sp. was originally classified

as a Seimatosporium species because of the presence of

excentric and eccentric basal appendages of the spore.However, our molecular study reveals that this species

should be a Truncatella species because the spores had

consistently two median cells with crenulated edges. The

presence of two median cells is most phylogenetically

significant at this taxonomic level and could be used to

distinguish between Truncatella species and other gen-

era.

388 R. Jeewon et al. / Molecular Phylogenetics and Evolution 25 (2002) 378–392

4.5. Phylogenetic status of Discosia

The grouping of Discosia as a sister taxon to Seima-

tosporium indicates close phylogenetic affinity but,

whether Discosia needs to be synonymized to Seima-

tosporium is uncertain. A larger sample size is required.

The single species used clustered separately from all of

the Seimatosporium species in ML and MP analyses and

this relationship is supported by a high bootstrap con-fidence. Nag Rag (1993) considered the genus Seima-

tosporium distinct from Discosia as it contains a mixture

of appendaged and nonappendaged conidia with ex-

centric basal appendages (not arising from the center)

and pigmented median cells. In contrast, Discosia is

characterized by basal appendages that are inserted at

the end cells on the concave side of the conidia and

median cells that are almost hyaline. Analyses of thecombined LSU and ITS data provide further evidence to

justify that Discosia is closely related to Seimatosporium.

Morphologically, there are some resemblances between

these genera and this raises the concern whether some of

the species in Discosia need to be reexamined and

transferred to Seimatosporium. The location of the co-

nidial septa and a rounded apical cell is quite a char-

acteristic morphological feature with a considerabletaxonomic value in delimiting these two genera (Vanev,

1991) but it would be premature to discuss a specific

taxonomic relationship given that only one species of

Discosia was included in this study. We are left with the

need to assess the phylogenetic relationships of Discosia

with Seimatosporium and their morphological synapor-

morphies as data available at present appear equivocal.

4.6. Phylogenetic status of Monochaetia

This genus shares similar characters with Pestaloti-

opsis except that the species possess only one apical

appendage arising from the apical cell. Results obtained

are incongruent with previous morphological hypothe-

ses on two major points. First, molecular analyses based

on MP and ML produced trees that place M. mono-

chaeta as the sister taxon to Pestalotiopsis and not

within Pestalotiopsis as proposed by Steyeart (1956). He

proposed that all the Monochaetia species should be

allocated in the section Monosetulae of Pestalotiopsis

but this is not supported herein. It has been observed

that M. monochaeta is structurally very similar to Pes-

talotiopsis (Griffiths and Swart, 1974a), emphasizing the

close relationships between these two genera. Sutton(1969) advocated that conidial wall structure should be

used as an additional morphological character to sup-

port Steyeart�s ideas, but our results indicate that such acharacter is not suitable for generic differentiation be-

tween Pestalotiopsis, Monochaetia, and Seiridium. Sec-

ond, the NJ analysis did not resolve the problem by

placing M. monochaeta as a sister taxon to Seiridium

instead of Pestalotiopsis. In any case,M. monochaeta didnot fit in either the Pestalotiopsis or the Seiridium clades.

This finding concurs with the morphological evidence

obtained by Roberts and Swart (1980) who observed

that conidial wall structures of Seiridium were structur-

ally different from those of M. monochaeta. We are not

surprised by the association between Monochaetia and

Seiridium (Fig. 3). Although Monochaetia was consid-

ered to be closely related to Pestalotiopsis (Steyeart,1949) due to the euseptate nature of the three median

cells, it has only a single apical appendage, unlike those

of Pestalotiopsis which have two to four apical ap-

pendages. Guba (1961) inferred a close relationship be-

tween Seiridium and Monochaetia based on

morphology, a connection supported by the results ob-

tained from the NJ analysis. However the presence of a

single appendage is also a common phenomenon in M.

karstenii, Seimatosporium, Seiridium, and Truncatella.

Our findings indicate that presence of a single append-

age may be a character that has evolved more than once

among these genera and therefore may not be phyloge-

netically significant nor useful for generic delimitation.

Steyeart (1956) postulated that Monochaetia should not

be given generic status based solely on the presence of a

single apical appendage, as opposed to Guba (1955) whostated that Monochaetia merits generic status. In an-

other study, it was observed that the acervulus mor-

phology of another species of Monochaetia, M. lutea,

has affinities with Seiridium. There are therefore mor-

phological grounds to suspect that the position of M.

monochaeta as revealed by molecular analysis in the NJ

analysis may reflect the correct phylogeny. Our study

supports neither Steyeart�s nor Guba�s views.Placement of M. monochaeta is problematic and lar-

gely unresolved in this study because of (i) an inconsis-

tency in the grouping pattern of M. monochaeta

obtained with the different methods of analysis and (ii)

the relatively low corresponding bootstrap values. The

phylogenetic relationships of Monochaetia among its

allies need to be established in future studies. Further-

more it remains unclear whether the presence of a singleapical appendage is significant enough to give Mono-

chaetia generic status as favored by Guba (1955, 1956,

1961). Our molecular data also do not favor Steyeart�streatment (1949) that the genus Monochaetia should be

discarded and all its species reallocated to the section

Monosetulae of Pestalotiopsis.

4.7. Affinities with the amphisphaeriaceae

These anamorphic coelomycetous fungi have been

shown to have close affinities with the ascomycetous

family Amphisphaeriaceae. Some have been frequently

reported to produce their sexual states in culture (Barr,

1975, 1990; Boesewinkel, 1983; M€uuller and Shoemaker,1965; Samuels et al., 1987; Shoemaker and M€uuller, 1964,

R. Jeewon et al. / Molecular Phylogenetics and Evolution 25 (2002) 378–392 389

1965; Shoemaker and Simpson, 1981; Swart, 1973; Zhuet al., 1991). Based on morphological studies, Kang et

al. (1999) considered confining the Amphisphaericeae to

those fungi producing Pestalotiopsis-like anamorphs. In

addition, molecular studies based on the ITS region of

the rDNA have revealed that Amphisphaeria umbrina,

Discostroma tosta, Lepteutypa cupressi, Pestalosphaeria

elaeidis (telemorph of Pestalotiopsis), and Pestalotia

palmarum are closely related (Kang et al., 1998). Anadditional yet equally important aspect addressed in this

study is the anamorph–teleomorph connections of

Seridium species with L. cupressi and of Seismatospo-

rium with Discostroma. These connections have so far

been based only on cultural characters (Boesewinkel,

1983; M€uuller and Shoemaker, 1965; Okane et al., 1995;Shoemaker and M€uuller, 1964, 1965; Swart, 1973). Theclustering of Discostroma sp. with Seimatosporium vac-

inii and of L. cupressi with Seridium cupressi in all trees

obtained clearly supports the previous anamorph–tele-

omorph connections and provides unambiguous evi-

dence that these anamorphic fungi have close affinities

with the Amphisphaericeae.

5. Conclusion

The current study does not support the taxonomic

treatment of Guba (1961). Phylogenetic analyses of the

rDNA sequences is generally in agreement with the

morphological hypotheses proposed by Steyeart (1949)

and Nag Rag (1993). The partial LSU sequences to-

gether with existing morphological data have provided

valuable insights in the understanding of the naturalrelationships at the intergeneric level among these coel-

omycetous fungi. Our results provide a better phyloge-

netic interpretation of morphological characters and

their utility in determining generic delineations. Analysis

of molecular data combined with morphological data

resolves many disputes not resolved by morphology

alone. Useful characters include pigmentation, septate

nature of median cells, and position of appendages withrespect to the apical and basal cells. Pestalotiopsis is

characterized by spores having mostly four-euseptate

and pigmented median cells with two to four apical

appendages arising as tubular extensions from the apical

cell and a centric basal appendage; Seridium spores

contain five- to six-distoseptate median cells; Bartalinia

is characterized by spores having almost hyaline median

cells with apical appendages arising from a particularlocus and not separated by a septum; Truncatella have

two pigmented median cells and Seimatosporium have

two or three pigmented median cells with a single apical

appendage and basal appendages that are excentric.

Morphological characters such as number of median

cells, number and presence of apical appendages, and

presence of excentric basal appendages have presumably

undergone convergent evolution and are of limited usein delineating these genera. A taxonomic key to sum-

marize the delineating morphological characters for

these genera is provided in Fig. 5.

From our study, there is little support to give Mono-

chaetia generic rank, although this genus may putatively

be characterized by the presence of a single apical ap-

pendage. The relationships between Seimatosporium and

Fig. 5. Dichotomous key to genera based on phylogenetically significant morphological characters.

390 R. Jeewon et al. / Molecular Phylogenetics and Evolution 25 (2002) 378–392

Discosia are also equivocal. However, Discosia may ten-tatively be delineated by having almost hyaline median

cells with appendages inserted at the concave side of the

conidium. Since the efficiencies of methods in recon-

structing accurate phytogenies are poor under conditions

of sparse taxonomic sampling, the relationships between

Monochaetia andDiscosiawith respect to the other genera

remain obscure. For a proper resolution of their system-

atic positions further taxon sampling involving the addi-tion of more diverse taxa and the inclusion of other

genomic loci is necessary. Furthermore, more morpho-

logical studies are required, especially those at the ultra-

structural level. It will also be beneficial to include other

related taxa in future studies, e.g., Bleptosporium, Dip-

loceras, Doliomyces, Labridella, and Pestalotia with

distoseptate conida, Pestalotiopsis species with three-

septate median cells, Sarcostroma, Zetiasplozna, and te-leomorphic amphisphaericeous species.

Acknowledgments

The University of Hong Kong is thanked for pro-

viding a Postgraduate scholarship to the first author.

Thanks are extended to the University of Hong Kong

Research Grants Council for funding this research. We

also thank several colleagues, A. Aptroot, R. Fogel,

E.H.C. Mckenzie. R.G. Shivas, J.E. Taylor, and M.J.

Wingfield, for providing cultures and specimens for this

study. J.A. Simpson is acknowledged for his discussionand suggestions during the course of this research. R.

Dulymamode and G.J.D. Smith are thanked for helpful

advice. Rajeeta Jeewon is thanked for additional sup-

port. Heidi Kong and Helen Leung are thanked for

laboratory assistance.

References

Arx, Von. J.A., 1981. The Genera of Fungi Sporulating in Culture,

third ed. J. Cramer, Vaduz.

Barr, M.E., 1975. Pestalosphaeria, a new genus in the Amphisphae-

riaceae. Mycologia 67, 187–194.

Barr, M.E., 1990. Prodromus to nonlichenized, pyrenomycetous

members of class Hymenoascomycetes. Mycotaxon 39, 43–184.

Boesewinkel, H.J., 1983. New records of the three fungi causing

cypress canker in New Zealand, Seridium cupressi (Guba) comb.

nov. and Seiridium cardinale on Cupressocyparis and Seiridium

unicorne on Cryptomeria and Cupressus. Trans. Br. Mycol. Soc. 80,

544–547.

Cunningham, C.W., 1997. Is congruence between data partitions a

reliable predictor of phylogenetic accuracy? Empirically testing an

iterative procedure for choosing among phylogenetic methods.

Syst. Biol. 46, 464–478.

Doyle, J.J., Doyle, J.L., 1987. A rapid DNA isolation procedure for

small quantities of fresh leaf tissues. Phytochem. Bull. 19, 11–15.

Farris, J.S., K€aallerj€oo, M., Kluge, A.G., Bult, C., 1995. Testing

significance of incongruence. Cladistics 10, 315–319.

Felsenstein, J., 1984. Distance methods for inferring phytogenies: A

justification. Evolution 38, 16–24.

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.

Griffiths, D.A., Swart, H.J., 1974a. Conidial structure in two species of

Pestalotiopsis. Trans. Br. Mycol. Soc. 62, 295–304.

Griffiths, D.A., Swart, H.J., 1974b. Conidial structure in Pestalotia

pezizoides. Trans. Br. Mycol. Soc. 63, 169–173.

Guba, E.F., 1929. Monograph of the genus Pestalotia de Notaris. Part

1. Phytopathology 19, 191–232.

Guba, E.F., 1955.Monochaetia and Pestalotia. Mycologia 47, 920–921.

Guba, E.F., 1956. Monochaetia and Pestalotia vs. Truncatella, Pestal-

otiopsis and Pestalotia. Ann. Microb. Enzymol. Milan 7, 74–76.

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.

Jones, J.P., 1977. The ultrastructure of conidium ontogeny in

Pestalotiopsis neglecta. Can. J. Bot. 55, 766–771.

Jukes, T.H., Cantor, C.R., 1969. Evolution of protein molecules. In:

Monro, H.N. (Ed.), Mammalian Protein Metabolism. Academic

Press, New York, pp. 21–132.

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. Fungal Diversity 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.

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.

Lanave, C., Preparata, G., Saccone, C., Serio, G., 1984. A new method

for calculating evolutionary substitution rates. J. Mol. Evol. 20,

86–93.

Lockhart, P.J., Steel, M.A., Hendy, M.D., Penny, D., 1994. Recov-

ering evolutionary trees under a more realistic model of sequence

evolution. Mol. Biol. Evol. 11, 605–612.

Morgan-Jones, G., Nag Rag, T.R., Kendrick, B., 1972. Genera

Coelomycetarum. V. Alpakesa and Bartalinia. Can. J. Bot. 50, 877–

882.

Morgan, A., Boddy, L., Mordue, E.M., Morris, C.W., 1998. Evalu-

ation of artificial neural networks for fungal identification,

employing morphometric data from spores of Pestalotiopsis

species. Mycol. Res. 103, 975–984.

M€uuller, E., Shoemaker, R.A., 1965. The ascogenous state of Seima-

tosporium (¼Monoceras) kriegerianum on Epilobium species. Can.

J. Bot. 43, 1343–1345.

Nag Rag, T.R., 1993. Coelomycetous Anamorphs with Appendage

Bearing Conidia. Mycologue Publications, Waterloo, Ontario,

Canada.

Okane, I., Nagakiri, A., Ito, T., 1995. Discostroma tricellulaire, a new

endophytic ascomycete with a Seimatosporium anamorph isolated

from Rhodendron. Can. J. Bot. 74, 1338–1344.

Page, R.D.M., 1996. TREEVIEW: An application to display phylo-

genetic trees on personal computers. Comput. Appl. Biosci. 12,

357–358.

Roberts, D.C., Swart, H.J., 1980. Conidium wall stucture in Seiridium

and Monochaetia. Trans. Br. Mycol. Soc. 74, 289–296.

Rodriquez, R., Oliver, L., Marin, A., Medina, J.R., 1990. The general

stochastic model of nucleotide substitution. J. Theor. Biol. 142,

485–501.

R. Jeewon et al. / Molecular Phylogenetics and Evolution 25 (2002) 378–392 391

Roux, C., Warmelo, K.T.V., 1990. Conidiomata in Bartalinia robil-

lardoides. Mycol. Res. 99, 109–116.

Saitou, N., Nei, M., 1987. The neighbor-joining method: A newmethod

for reconstructing phylogenetic trees. Mol. Biol. Evol. 4, 406–425.

Samuels, G.J., M€uuller, E., Petrini, O., 1987. Studies on the Amphi-sphaeriaceae (sensu lato) 3. New species of Monographella and

Pestalosphaeria, and two new genera. Mycotaxon 28, 473–499.

Sanderson, M.J., 1989. Confidence limits on phylogenies: The boot-

strap revisited. Cladistics 5, 113–129.

Shoemaker, R.A., M€uuller, E., 1964. Generic correlations and concepts:

Cladithrium; (¼Griphosphaeria) and Seimatosporium (¼Sporoca-

dus). Can. J. Bot. 42, 403–410.

Shoemaker, R.A., M€uuller, E., 1965. Type of the pyrenomycete genera

Hymenopleella and Lepteutypa. Can. J. Bot. 43, 1457–1460.

Shoemaker, R.A., M€uuller, E., Morgan-Jones, G., 1966. Fuckel�sMassaria marginatai and Seiridium marginatum Nees ex Steudel.

Can. J. Bot. 42, 247–254.

Shoemaker, R.A., Simpson, J.A., 1981. A new species of Pestalosp-

haeria on pine with comments on the generic placement of the

anamorph. Can. J. Bot. 59, 986–991.

Steyeart, R.L., 1949. Contributions �aa l��eetude monographique de

Pestalotia de Not. et Monochaetia Sacc. (Truncatella gen. nov. et

Pestalotiopsis gen. nov.). Bull. Jard. Bot. �EEtat Bruxelles 19, 285–354.

Steyeart, R.L., 1953a. New and old species of Pestalotiopsis. Trans. Br.

Mycol. Soc. 36, 81–89.

Steyeart, R.L., 1953b. Pestalotiopsis from the Gold Coast and

Togoland. Trans. Br. Mycol. Soc. 36, 235–242.

Steyeart, R.L., 1955. Pestalotia, Pestalotiopsis et Truncatella. Bull.

Jard. Bot. �EEtat Bruxelles 25, 191–199.

Steyeart, R.L., 1956. A reply and an appeal to Professor Guba.

Mycologia 48, 767–768.

Steyeart, R.L., 1961. Type specimens of Spegazzini�s collections in thePestalotiopsis and related genera (Fungi Imperfecti: Melanconi-

ales). Darwinia (Buenos Aires) 12, 157–190.

Steyeart, R.L., 1963. Complimentary informations concerning Pestal-

otiopsis guepinii (Desmazieres) Steyeart and its designation of its

lectotype. Bull. Jard. Bot. �EEtat Bruxelles 33, 369–373.

Sutton, B.C., 1961. Coelomycetes. I. Mycol. Pap. 80, 1–16.

Sutton, B.C., 1969. Forest Microfungi. III. The heterogeneity of

Pestalotia de Not. Section Sexloculatae Klebahn sensu Guba. Can.

J. Bot. 47, 2083–2094.

Sutton, B.C., 1975. Coelomycetes. V. Coryneum. Mycol. Pap. 138, 1–

224.

Sutton, B.C., 1980. The Coelomycetes: Fungi Imperfecti with Pycnidia.

Acervular and Stromata. Commonwealth Mycological Institute,

Kew, Surrey, England.

Swart, H.J., 1973. The fungi causing cypress canker. Trans. Br. Mycol.

Soc. 61, 71–82.

Swofford, D.L., 2001. PAUP*: Phylogenetic Analysis Using Parsi-

mony (*and other methods), ver. 4.0b8. Sinauer, Sunderland,

MA.

Templeton, A.R., 1983. Phylogenetic inference from restriction endo-

nuclease 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.

Vanev, S.G., 1991. Species conception and sections delimitation of

genus Discosia. Mycotaxon 41, 387–396.

Vilgalys, R., Hester, M., 1990. Rapid genetic identification and

mapping of enzymatically amplified ribosomal DNA from several

Cryptococcus species. J. Bacterial. 172, 4238–4246.

Watanabe, K., Doi, Y., Kobayashi, T., 1998. Conidiomatal develop-

ment of Pestalotiopsis guepinii and P. neglecta on leaves of

Gardenia jasminoides. Mycoscience 39, 71–75.

Zhu, P., Ge, Q., Xu, T., 1991. The perfect stage of Pestalosphaeria

from China. Mycotaxon 50, 129–140.

392 R. Jeewon et al. / Molecular Phylogenetics and Evolution 25 (2002) 378–392