Mycologia, 98(6), 2006, pp. 971–981.

# 2006 by The Mycological Society of America, Lawrence, KS 66044-8897

Molecular systematics and biological diversification of Boletales

Manfred Binder1 INTRODUCTION

David S. Hibbett
The Boletales (Agaricomycetidae) is one of the major
Clark University, Biology Department, Lasry Center for
Bioscience, 950 Main Street, Worcester, Massachusetts groups of mushroom-forming fungi that is represent-
01610-1477 ed in most forest ecosystems worldwide. This order
contains approximately 1000 described species (Kirk
et al 2001), which might be an underestimate
Abstract: Historical patterns of morphological evo- considering that larger parts of the neotropics and
lution and ecology in the Boletales are largely the paleotropics are still understudied, where Bole-
unresolved but appear to involve extensive conver- tales have reached most notable divergence. The
gence. We studied phylogenetic relationships of Boletales includes conspicuous stipitate-pileate forms
Boletales based on two datasets. The nuc-lsu dataset that mainly have tubular and sometimes lamellate
is broadly sampled and includes roughly 30% of the hymenophores or intermediates that show transitions
described species of Boletales and 51 outgroup taxa between the two types of hymenophores (Gilbert
across the Hymenomycetes. The multigene dataset 1931). The Boletales also includes gasteromycetes
(nuc-ssu, nuc-lsu, 5.8S, mt-lsu, atp6) sampled 42 key (puffball-like forms), resupinate or crust-like fungi
species of Boletales in a framework of 14 representa- that produce smooth, merulioid (wrinkled to warted),
tive Hymenomycetes. The Boletales are strongly or hydnoid (toothed) hymenophores, and a single
supported as monophyletic in our analyses on both polypore-like species, Bondarcevomyces taxi (e.g. Besl
datasets with parsimony, maximum likelihood and and Bresinsky 1997, Jarosch 2001, Larsson et al 2004).
Bayesian approaches. Six major lineages of Boletales In view of the existing diversity of fruiting body forms,
that currently are recognized on subordinal level, there has been extensive homoplasy in the evolution
Boletineae, Paxillineae, Sclerodermatineae, Suilli- of Boletales and there is no obvious morphological
neae, Tapinellineae, Coniophorineae, received varied character that unites the group (FIG. 1). Species in
support. The backbone of the Boletales was moder- Boletales pursue diverse habits, but unlike in their
ately resolved in the analyses with the nuc-lsu dataset, sister clades (Agaricales and Atheliales) white-rot
but support was strong for most major groups. saprotrophy is absent in the group (Binder et al
Nevertheless, most brown-rot producing forms were 1997). Instead, saprotrophs among Boletales have
placed as a paraphyletic grade at the base of the developed a unique mode of brown-rot called
Boletales. Analyses on the multigene dataset confirm ‘‘Coniophoraceae-rot’’ (Kämmerer et al 1985, Besl
sister group relationships among Boletales, Agaricales et al 1986) that primarily is aimed at decaying wood of
and Atheliales. Boletineae and Suillineae received the conifers. Mycorrhizal associations are established by
highest support values; Paxillineae and Scleroderma- the majority of Boletales, and host plants include
tineae were not consistently resolved as monophyletic Betulaceae, Casuarinaceae, Dipterocarpaceae, Erica-
groups. The Coniophorineae were not monophyletic ceae, Fabaceae, Fagaceae, Mimosaceae, Myrtaceae,
in any analyses. The Tapinellineae consisting of Pinaceae and Salicaceae (Newman and Reddell
morphologically diverse brown-rotting fungi forms 1987). Some Boletales are mycoparasites, a deviation
the basal group in the Boletales. We performed of either the saprotrophic or ectomycorrhizal mode
ancestral state reconstruction with BayesMultiState, that is limited to some species in Boletaceae and in
which suggested that the ancestor of the Boletales was Gomphidiaceae (Agerer 1991, Raidl 1997).
a resupinate or polyporoid saprotrophic fungus, In recent years Boletales have been studied widely
producing a brown-rot. by fungal systematists, chemists, ecologists and my-
Key words: evolution, morphology, MRCA, corrhizal biologists (e.g. Agerer 1987–1998, Arpin and
multigene analyses, nuc-lsu rDNA, nutritional modes Kühner 1977, Besl and Bresinsky 1997, Both 1993,
Gill and Steglich 1987, Moser 1983, Singer 1986,
Smith and Thiers 1971, Watling 1970). Six suborders
have been established that are thought to represent
the major lineages of Boletales: Boletineae, Paxilli-
neae, Sclerodermatineae, Suillineae, Coniophori-
Accepted for publication 16 June 2006. neae, and Tapinellineae. Remarkably, phylogenetic
1
Corresponding author. E-mail: [email protected] inferences were used rarely to improve higher-level

971
972 MYCOLOGIA

FIG. 1. Morphological diversity in Boletales. a. Bondarcevomyces taxi; b. B. taxi, pores; c. Coniophora puteana; d.
Leucogyrophana mollusca; e. Hygrophoropsis aurantiaca; f. Suillus granulatus; g. Chroogomphus vinicolor; h. Boletinellus
merulioides, hymenophore; i. Calostoma cinnabarinum; j. Scleroderma septentrionale; k. Meiorganum neocaledonicum, young
hymenophore; l. ‘‘Tylopilus’’ chromapes; m. Phylloporus centroamericanus; n. Xerocomus sp. Pictures a and b courtesy Y.-C. Dai; m
courtesy M.-A. Neves.

classifications in Boletales but had a synergistic effect menophores were placed at that time. Innovative
on traditional and interdisciplinary methods in methods shifted the focus from fruiting body mor-
general. For example, the well established study of phology to below ground characters and the recog-
chemistry of Boletales pigments and other colorless nition of rhizomorphs and substrate hyphae as
secondary metabolites helped the application of morphologically conserved characters led to the
chemotaxonomy to separate the Suillineae from description of Tapinellineae and Coniophorineae in
Boletineae (Besl and Bresinsky 1997), in which the the study of Agerer (1999). Early phylogenetic studies
large preponderance of species with tubular hy- on Boletales examined relationships between stipi-
 BINDER AND HIBBETT: EVOLUTION OF BOLETALES 973

tate-pileate and gasteroid forms and explored rate iegatus, C. marmorata, Suillus spraguei, Porphyrellus porphyr-
differences in base substitutions between nuclear and osporus; ITS for Dendrocorticium roseocarneum; and nuc-ssu
mitochondrial genes (Bruns et al 1989, Bruns and for Scleroderma hypogaeum and Suillus ochraceoroseus.
Szaro 1992). Several phylogenetic analyses with focus The nuc-lsu dataset included 301 species of Boletales,
which is roughly 30% of the described species in this order.
on systematics, using nuclear and mitochondrial
In addition, some of these species were represented by
rDNA, suggest that the Boletales is monophyletic
multiple sequences (133 in total) and 51 outgroup species
(Binder and Bresinsky 2002a, Binder et al 2005, Bruns were selected to represent the major clades of homobasi-
et al 1998, Grubisha et al 2001, Jarosch 2001, Kretzer diomycetes. Sequence data of outgroup species were
and Bruns 1999) and show that Agaricales and gathered largely from the studies of Moncalvo et al
Atheliales are sister groups of the Boletales (Hibbett (2002), Hibbett and Binder (2002) and Larsson et al
and Binder 2002, Larsson et al 2004, Binder et al (2004). Three hundred twenty-four sequences of Boletales
2005). were drawn from published studies (Binder et al 1997;
This study combines the efforts from previous Binder and Besl 2000; Binder and Bresinsky 2002a, b;
studies and provides 134 new sequences for 60 Bresinsky et al 1999; Hughey et al 2000; Grubisha et al 2001;
species. One objective was to assemble a multigene Jarosch 2001; Jarosch and Besl 2001; Peintner et al 2003).
Fifty sequences that originate from unpublished studies
dataset (nuc-ssu, nuc-lsu, 5.8S, atp6, mt-lsu) to resolve
(James et al, Carlier et al) are available from GenBank. We
sister-group relationships among Boletales, Agaricales
generated new sequences of 61 species for the nuc-lsu
and Atheliales and to test the monophyly of major dataset.
groups in the Boletales with maximum likelihood and
Bayesian methods. The second objective was to Phylogenetic analyses of the multigene dataset.—To test for
generate a most inclusive nuc-lsu dataset and to congruence among nuclear (nuc) and mitochondrial (mt)
analyze it with Bayesian methods. The results of this genes, a preliminary series of parsimony bootstrap analyses
was performed in PAUP*. Five separately estimated gene
analysis were used to estimate probabilities of
phylogenies using nucleotide data were obtained running
ancestral states of morphology and nutritional mode
1000 replicates, all characters equally weighted, 10 random
for supported nodes with BayesMultiState (Pagel et al taxon addition sequences, tree bisection reconnection
2004). (TBR) branch swapping, with MAXTREES set to 10 000. None
of the positively conflicting nodes between partitions
MATERIALS AND METHODS received bootstrap support .63%, and the data were
combined to a single dataset encompassing 3939 aligned
Taxon sampling and molecular datasets.—One hundred positions. A bootstrap analysis then was performed on the
thirty-four sequences that were newly generated for this concatenated dataset with the previously described settings.
study included three nuclear rDNA regions (nuc-ssu, nuc- The multigene dataset was analyzed further with maxi-
lsu, ITS) and two mitochondrial genes (atp 6, mt-lsu). The mum likelihood (ML) and a Bayesian approach (Metrop-
sequences have been deposited in GenBank (DQ534563– olis-coupled MCMC or MC3). Six-parameter models were
DQ534696, SUPPLEMENTARY TABLE I, II, SUPPLEMENT). For estimated as the best-fit likelihood models with Modeltest
DNA extraction protocols, PCR, cloning, sequencing and 3.06 (Posada and Crandall 2001) for all five partitions
sequence alignment, refer to APPENDIX 1 and to the studies (GTR+C+I for nuc partitions, TVM+C for atp6, and
of Bruns et al (1998), Kretzer and Bruns (1999), Binder et al TVM+C+I for mt-lsu), while there was considerable variation
(2005) and references therein. Two datasets were assem- among model parameters between nuc and mt partitions.
bled, a multigene dataset (nuc-ssu, nuc-lsu, 5.8S, atp 6, mt- To perform the ML analysis the GTR+C+I model was
lsu) with 56 terminals and a broadly sampled nuc-lsu dataset specified with proportion of invariable sites and distribution
with 485 terminals. All analyses were performed on a Linux of rates at variable sites modeled on a discrete gamma
Pro 9.2 Opteron AMD 246 cluster (Microway) unless distribution (a 5 0.4) with four rate classes. The sub-
otherwise noted. Both alignments were submitted to stitution rate matrix was set to empirical frequencies and
TreeBASE (SN2858). More information on alignment the proportion of invariable sites was estimated during the
procedures and command blocks running MrBayes v3.1.1 run. The ML analysis was started with a user-defined starting
(Ronquist and Huelsenbeck 2003) and PAUP* 4.0b10 tree obtained with neighbor joining.
(Swofford 2002) are available (APPENDIX 1). The GTR+C+I model also was specified in the MC3
The multigene dataset included 42 species sampled analysis as prior for both nuc and mt partitions, assuming
across the six suborders of Boletales and 14 outgroup equal probability for all trees and unconstrained branch
species. The studies by Bruns et al (1998) and Kretzer and length. The substitution rate matrix, transition/transversion
Bruns (1999) provided the core data for the multigene rate ratio, character state frequencies, gamma shape
dataset, which was expanded with 73 new sequences. Several parameter a and proportion of invariant sites were unlinked
genes were not amplified successfully in these species: atp6 across nuc and mt partitions and calculated independently
for Athelia arachnoidea, Austropaxillus sp., Coniophora by MrBayes. Posterior probabilities were determined twice
marmorata, Leucogyrophana mollusca, Pseudomerulius aure- by running one cold and three heated chains for 10 3 106
us; mt-lsu for Fomitiporia mediterranea, Melanogaster var- generations in parallel mode, saving trees every 100th
974 MYCOLOGIA

generation. A 50% majority rule consensus tree was used to both trait files under the ML criterion performing 10
calculate posterior probabilities including the proportion of independent optimizations per tree. All MRCA analyses
trees gathered after the convergence of likelihood scores were run on a Powermac G5 via Darwin in OS 3 10.4.3.
was reached.

Phylogenetic analyses of the nuc-lsu dataset (1071 positions).— RESULTS AND DISCUSSION
They were performed with a Bayesian MC3 approach. Two
parallel MC3 analyses were run under the GTR+C+I model Higher-level relationships of Boletales.—Recent phylog-
using four chains and an extended run time employing 50 enies using increased taxon sampling and multiple
3 106 generations, saving trees every 100th generation. gene loci (Binder et al 2005, Hibbett and Binder
Posterior probabilities were calculated as previously de- 2002, Larsson et al 2004, Matheny et al 2006)
scribed. consistently resolve a large clade that contains
Ancestral state reconstruction was performed with most Agaricales, Atheliales and Boletales—the Agaricomy-
recent common ancestor (MRCA) analysis implemented in cetidae. As in previous studies, sister relationships
BayesMultiState v1.0.2 in maximum likelihood mode (Pagel within the Agaricomycetidae remain ambiguous and
et al 2004). Ancestral state reconstructions, based on either receive varied support. The Atheliales is resolved as
parsimony or maximum likelihood, are performed fre- sister group of the Boletales in our multigene analyses
quently with a single input tree, which implies that the (FIG. 2). This relationship is supported strongly by
phylogeny is known with certainty (e.g. Hibbett 2004). Such
posterior probabilities (PP), however bootstrap sup-
an assumption usually is not warranted. In contrast, the
Bayesian approach combines probability estimates of
port (BS) is weak (,50%). A sister relationship of
ancestral traits across a statistically justified sample of trees, Atheliales and Boletales also was inferred from
which effectively factors out phylogenetic uncertainty (e.g. combined nuc-lsu and 5.8S rDNA data by Larsson et
Lutzoni et al 2001). BayesMultiState was used to estimate al (2004) without receiving statistical support. Phylo-
the probabilities of ancestral character states for morphol- genetic analyses of the nuc-lsu dataset in the present
ogy and nutritional mode at eight nodes, including the root study (SUPPLEMENTARY FIG. 1) place the Agaricales as
node of the Boletales and seven nodes within the Boletales. sister group of the Atheliales, again supported by high
Each of the eight nodes supported a group that was resolved posterior probabilities. It therefore is important to
as monophyletic in all trees recovered from the MC3 identify the basal groups in all three orders and to
analyses, except the Sclerodermatineae, which was resolved generate comprehensive multigene data to resolve
as monophyletic in 89% of the trees. relationships within the Agaricomycetidae.
To reduce the computational burden, ancestral states
were estimated with a sample of 250 trees recovered from Resolving the major clades of Boletales.—Both nuc-lsu
the stationary tree distribution of the MC3 analysis. These and multigene phylogenies support the Boletales as
250 trees represent all the unique topologies present in the a monophyletic order and this result is consistent with
set of trees sampled in the MC3 analyses and were obtained previous studies (Binder and Bresinsky 2002a, Bruns
with the tree filtering functions in PAUP*. MRCA analysis et al 1998, Grubisha et al 2001, Jarosch 2001, Kretzer
accommodates compositional heterogeneity of clades, such
and Bruns 1999). Our approach to present two
as that exhibited by the Sclerodermatineae, by estimating
disparate datasets needs to be seen as a transitional
ancestral states across trees at the shallowest node that
subtends all the species assigned to the clade. In other
stage between using the most inclusive taxon sam-
words, if a group of species assigned to a clade at the outset pling and a steadily increasing availability of multiple
of the analysis is not monophyletic on one of the trees genes sampled for the same set of species. Brown-rot
analyzed, the MRCA approach estimates the state at the producing saprotrophs were recovered as the earliest
most recent common ancestor of those species in all of the branching groups in this study and in most afore-
trees (Pagel et al 2004). Two trait input files were mentioned analyses, however, assessing branching
constructed in the PAUP* editor that coded for morphol- order among these clades proves to be difficult. The
ogy and nutritional mode (five states each) of all species in study by Kretzer and Bruns (1999) combined two
the dataset (see SUPPLEMENTARY FIGURE 1 for the coding mitochondrial loci (atp6 and mt-lsu) for 23 members
regime). Morphology was coded as: 0 5 stipitate-pileate with of the Boletales for the first time and resolved the
tubes; 1 5 stipitate-pileate with gills; 2 5 gasteroid; 3 5 Tapinellineae as basal lineage, but the position of
resupinate; 4 5 polyporoid. Intermediate states were
other brown-rotting fungi was not well supported.
accommodated with a combined state identifier (e.g. the
Inferences of the multigene data in this study (FIG. 2)
stipitate-pileate hydnoid fungus Sarcodon imbricatus was
coded as 01). Coding for the nutritional mode included: 5 show that extended taxon sampling and the addition
5 potentially ectomycorrhizal; 6 5 brown-rot saprotroph; 7 of three rDNA genes improves the overall resolution
5 white-rot saprotroph; 8 5 mycoparasitic; – 5 uncertain of major groups in Boletales but still is not answering
state (which is treated as if the trait can be any of the other all questions about sister relationships. Our results
states 5–8). The MRCA analyses were set up to reconstruct suggest that there might be as many as eight mono-
eight specified nodes (TABLE I) and were run separately for phyletic lineages in the Boletales and that the
 BINDER AND HIBBETT: EVOLUTION OF BOLETALES 975

FIG. 2. Phylogenetic relationships of Boletales inferred from the multigene dataset (nuc-ssu, nuc-lsu, 5.8S, atp6, mt-lsu)
under the ML criterion (–lnL 5 38227.648). The dataset included 3791 characters after the exclusion of ambiguously aligned
148 characters. Bootstrap frequencies .50% are shown at supported branches. Posterior probabilities 0.98–1.0 obtained in the
Bayesian analyses are indicated by bold nodes. The major clades of Boletales and the sister groups of Boletales, Agaricales and
Atheliales are indicated with brackets.
976 MYCOLOGIA

Tapinellineae and Coniophorineae form the basal spp. are morphologically similar to Boletinellus spp. in
clades. The Coniophorineae including Coniophora- the Sclerodermatineae, which contributed much to
ceae, Serpulaceae and Hygrophoropsidaceae is not taxonomic uncertainty (Binder and Bresinsky 2002a).
monophyletic and forms three independent groups. Another question regards the extent and monophyly
A sister relationship of Coniophoraceae and Hygro- of Melanogastraceae. This family includes the ecto-
phoropsidaceae is weakly supported (BS 5 58%) in mycorrhizal false-truffles Alpova and Melanogaster
the multigene analysis, however, additional Leucogyr- (Trappe 1975), which form two independent clades
ophana spp. break up this relationship in the trees in this study. The Paxillineae is resolved as sister
inferred from the nuc-lsu dataset (SUPPLEMENTARY group of the Boletineae (PP 5 1.0, BS 5 97%) in the
FIG. 1). This result is consistent with the study of multigene analyses. In contrast to previous studies
Jarosch and Besl (2001), showing that a morphologi- (Bresinsky et al 1999, Jarosch 2001) the Paxillineae
cally well characterized genus Leucogyrophana is formed either a paraphyletic or a polyphyletic group
polyphyletic and accordingly should be divided into in our analyses. Without receiving a strong phyloge-
several new genera. netic signal, the Paxillineae is sustained possibly as
The more derived groups in the Boletales compris- natural group by the production of secondary
ing Suillineae, Sclerodermatineae, Paxillineae and metabolites that are unique in the Boletales (Besl et
Boletineae form a clade with varying statistical al 1996).
support (PP 5 0.98, BS 5 67%). These four suborders The Boletineae is the most species-rich group of
include the majority of stipitate-pileate mushrooms stipitate-pileate fungi with tubular hymenophores in
with lamellate hymenophores (5 agaricoid) and the Boletales and also includes a few species with
tubular hymenophores (5 boletoid), false-truffles lamellate hymenophores and gasteroid forms. Sup-
and earthballs, and also the majority of ectomycor- port values for this suborder are high in the multi-
rhizal forms. The Suillineae receives high support gene analyses (PP 5 1.0, BS 5 100%) and Chalciporus
values (PP 5 1.0, BS 5 100%), although branching occupies the basal position in the clade. The analyses
order within the clade is not resolved with confi- on the nuc-lsu dataset (SUPPLEMENTARY FIG. 1) with
dence. The group includes gasteromycetes (Rhizopo- increased taxon sampling show that relationships
gon, Truncocolumella), agaricoid forms (Gomphidius, among genera are poorly resolved and moreover that
Chroogomphus) and boletoid fungi (Suillus). The most of the larger genera (e.g. Boletus, Tylopilus,
Sclerodermatineae received weak support in the Xerocomus) are not monophyletic.
multigene analyses (BS 5 56%) and strong support
in the analyses on the nuc-lsu dataset (PP 5 1.0). The Morphological and ecological evolution.—Important
group includes a few boletoid forms (Boletinellus, but also challenging key questions in the evolution
Phlebopus, Gyroporus) and an overwhelming diversity of mushroom-forming fungi concern the directional-
of gasteromycetes (Fischer 1899–1900). For example, ity of morphological and ecological traits and their
Scleroderma spp. have compact peridia enclosing the potential reversibility (Bruns and Shefferson 2004,
gleba, Pisolithus spp. produce spore containing Hibbett 2004, Hibbett et al 2000). The majority of
peridioles that resemble those in the Nidulariaceae species in the Boletales are thought to enter
(Agaricales), Calostoma spp. produce gelatinous- ectomycorrhizal symbioses, even though this assump-
stalked fruiting bodies with multiple peridial layers, tion is based largely on observations in the field that
Astraeus spp. resemble the earthstars in the Geas- need to be confirmed by additional evidence.
trales, Diplocystis wrightii produces individual fruiting Fortunately, there is increasing interest in document-
bodies congregated on shared stromata and Tremello- ing ectomycorrhizal fungi and collaborative projects,
gaster surinamensis produces heavy fruiting bodies such as DEEMY (http://www.deemy.de) and UNITE
with strongly gelatinized peridial layers that superfi- (http://unite.zbi.ee; Kõljalg et al 2005), have become
cially are similar to those of species in the Phallales. valuable resources for systematists and ecologists.
The genus Gyroporus is nested within the gasteromy- Other species in the Boletales are brown-rot sapro-
cetes. The Boletinellaceae (Boletinellus, Phlebopus) trophs, especially on coniferous trees, and a few
usually is resolved as basal group (Binder and species are host specific mycoparasites that attack
Bresinsky 2002a, Hughey et al 2000) but sometimes other Boletales. The Boletales also includes a great
forms an independent sister clade of the remaining diversity of fruiting body forms (FIG. 1, clavarioid and
Sclerodermatineae (e.g. Kretzer and Bruns 1999). coralloid fungi are absent in this group however) and
Family concepts in the Paxillineae are still in flux as therefore provides an excellent model to study
reflected by a proposal to adopt the Paxillaceae in character evolution on a relatively manageable scale.
a wider sense by including Gyrodontaceae and This study used multistate coding combined with a ML
Melanogastraceae (Bresinsky et al 1999). Gyrodon approach to estimate probabilities of ancestral states
 BINDER AND HIBBETT: EVOLUTION OF BOLETALES 977

optimized on eight nodes (SUPPLEMENTARY FIG. 1, been suggested by Besl et al (1986) using the pigment
TABLE I) that were resolved in the analyses of the nuc- chemistry of secondary metabolites as a comparative
lsu dataset: Tapinellineae, Coniophorineae, Serpula- marker, and their findings found strong support in
ceae, Hygrophoropsidaceae, Suillineae, Scleroderma- recent phylogenetic studies (Bresinsky et al 1999,
tineae, Boletineae and Boletales. Jarosch 2001, Jarosch and Besl 2001).
Fruiting body evolution. The ancestral morphologi- The tubular hymenophore type that is symptom-
cal form of the Boletales was estimated as either atic for Boletales occurs in Suillineae, Scleroderma-
resupinate (P 5 0.545) or polyporoid (P 5 0.366) and tineae, Paxillineae and Boletineae. The stipitate-
we interpret this result as inconclusive. Similarly, the pileate form with a tubular hymenophore is resolved
ancestral fruiting body form of the Tapinellineae, as the ancestral state of the Boletineae and
which is placed at the base of the Boletales, was Sclerodermatineae. The Paxillineae also includes
estimated as either resupinate or polyporoid, although such typical boletoid forms, but its ancestral state
the polyporoid condition received a higher probabil- was not estimated because this group was not
ity. Bondarcevomyces taxi (FIG. 1a, b) is the only known resolved as monophyletic (FIG. 2, SUPPLEMENTARY
polypore in the Boletales and its placement next to the FIG. 1; TABLE I). The clade including Boletineae,
resupinate fungus Pseudomerulius aureus in the Paxillineae and Sclerodermatineae is not strongly
Tapinellineae, which recently was discovered in the supported. Nevertheless, it is most parsimonious to
study by Larsson et al (2004), is a startling finding that infer that the common ancestor of these groups had
challenges previous views of the morphological evolu- a boletoid form. If so, then the gasteroid taxa in all
tion in this group. Morphological transformations three suborders, and the lamellate taxa in the
from resupinate to polyporoid fruiting bodies or vice Paxillineae and Boletineae, must have been derived
versa are not uncommon in other fungal groups ultimately from boletoid forms.
(Binder et al 2005), however the directionality of Gasteromycetation occurs in most lineages of
events appears to be nonuniform. The Hymenochae- Boletales except Tapinellineae, Coniophorineae and
taceae (Hymenochaetales) might serve as a good Hygrophoropsidaceae. In most cases the ancestral
example because the family includes several genera states of clades containing gasteroid forms were
(e.g. Phellinus, Fomitiporia) in which both fruiting resolved as nongasteroid. However, we obtained
body forms occur side by side and often represent a surprising result in the Suillineae, which includes
cryptic species complexes (Fischer and Binder 2004). boletoid, agaricoid and gasteroid forms (TABLE I).
Taken together, our results suggest that at least five The ancestral state of the Suillineae was supported
independent transformations from resupinate forms strongly as gasteroid, implying parallel evolution of
to stipitate-pileate forms with lamellate hymenophores boletoid and lamellate forms and secondary evolution
have occurred in the basal lineages of Boletales. of ballistospory via reversals of gasteromycetation.
Leucogyrophana olivascens and L. romellii represent This contradicts the generally accepted view that the
a clade entirely composed of resupinate fungi and they loss of ballistospory is irreversible (Hibbett et al 1997,
obviously are not closer related to stipitate-pileate Savile 1955, Thiers 1984) as well as the specific
forms. The Paxillineae were not reconstructed in the findings of Bruns et al (1989), who suggested that
MRCA analyses but include another resupinate form, Rhizopogon species are derived from within the
Hydnomerulius pinastri. suilloid clade by selection for animal dispersal and
Moving up the tree, the results of the MRCA reduction of water loss.
analyses strongly suggest that the most recent Our finding that the ancestor of the Suillineae was
common ancestors of Serpulaceae, Hygrophoropsida- gasteroid could be due to an error in phylogenetic
ceae and Coniophorineae were resupinate forms. reconstruction. Paraphyly of Rhizopogon has been
Extant resupinate forms (FIG. 1c, d) in these taxa suggested by several other phylogenetic studies
include fungi with smooth hymenophores (Conio- (Binder and Bresinsky 2002a, Grubisha et al 2001,
phora, Leucogyrophana) and merulioid hymenophores Jarosch 2001, Kretzer and Bruns 1999). The place-
(Serpula, Leucogyrophana) and multiple transitions ment of Rhizopogon in our multigene analyses
from resupinate fruiting bodies to stipitate-pileate (FIG. 2), forming an unsupported sister group of
fruiting bodies with lamellate hymenophores can be the remaining Suillineae together with Gomphidia-
inferred in all three groups. For example, ‘‘Paxillus’’ ceae, might be an artifact caused by asymmetric
gymnopus and ‘‘P.’’ chalybaeus are nested within nuclear and mitochondrial base-substitution rates in
Coniophora, Austropaxillus is the sister group of different branches of the Suillineae (Bruns and Szaro
Serpula and Leucogyrophana mollusca (FIG. 1d) forms 1992). Nevertheless, the analyses on the nuc-lsu
a clade with the false cantharelle Hygrophoropsis dataset with 25 Rhizopogon species produce a similar
aurantiaca (FIG. 1e). All these relationships have topology with a paraphyletic Rhizopogon at the base of
978 MYCOLOGIA

! 5 character present, — 5 character absent; probabilities (P 5 0–1) are represented as arithmetric means across 250 input trees. Non-significant probabilities (P ,
Suillineae (SUPPLEMENTARY FIG. 1). Of course, the

Sclerodermatineae Paxillineae Boletineae Boletales

! 0.545
! 0.366

! 0.936
position of Rhizopogon in our trees might be correct

—
8

!
!
!

!
!
and the ancestral state of the Suillineae that was
estimated with a ML approach might be an artifact
caused by the use of an inappropriate model for
! 0.709

! 0.245

! 0.938
—
—

—
fruiting body evolution (for a discussion of the use of
!

!
!
7

Markov models to understand morphological evolu-

tion see Felsenstein 2004).
Ecological evolution. The results of the MRCA
analyses suggest that brown-rot is the ancestral state
—

—
!
!
!
!

!
*

for Boletales. Brown-rot also is estimated as the

ancestral nutritional mode for Tapinellineae, Conio-
phorineae, Serpulaceae, Hygrophoropsidaceae and
Probabilities of ancestral morphological and nutritional states at eight nodes in the Boletales estimated using MRCA

therefore might have evolved a single time in the

—0.196
! 0.908

! 0.697

basal clades of Boletales. A single switch from brown-

—

—
—

—
—

0.05) are not shown. *Paxillineae were either paraphyletic or polyphyletic and the node was therefore not reconstructed.
!
6

rot to ectomycorrhiza emerges in the Serpulaceae

leading from resupinate Serpula spp. saprotrophs to
agaricoid Austropaxillus spp. and gasteroid Gymnopax-
illus spp., which are associated with Nothofagus and
Suillineae

Eucalyptus (Claridge et al 2001). Because Tapinella

! 0.990

! 0.983

and Hygrophoropsis have maintained a saprotrophic

—
—

—
—
!
!

!
5

survival system, changing the nutritional mode to

ectomycorrhizal evidently is not correlated with the
Coniophorineae Hygrophoropsidaceae

gain of agaricoid morphology. Buchwaldoboletus lig-

nicola, reportedly a brown-rot fungus, is described
growing on stumps of conifers, woody debris or needle
! 0.144

! 0.854

! 0.999

litter (Pilát 1969). At first view this lifestyle is

—

—
—
—

—
4

exceptional in the Boletineae and appears to be

a prime example of a reversal from mutualism to
saprotrophism. The results in the study of Szczepka
and Sokól (1984) suggest that B. lignicola is able to
grow only on wood that was decayed by the polypore
Phaeolus schweinitzii and thus contradict that B.
! 0.955

! 0.999

lignicola actually causes the brown-rot. These findings

—

—
—
—

—
!
3

show evidence of an ecological specialization and close

association between both fungi, however, we question
whether the ecology of B. lignicola has been re-
Serpulaceae

searched sufficiently.
! 0.907

! 0.999

Our ancestral state reconstructions suggest that

—

—
!
!

!
2

mycoparasitism in Boletales is derived from ectomy-

corrhizal forms. An example from the Suillineae
indicates that competition for food is most likely
Tapinellineae

a crucial factor that triggers parasitic interactions

! 0.226
! 0.672

! 0.999

among closely related groups. Species in Suillineae

—

—
—

—
!

associate with Pinaceae and frequently produce their

1

fruiting bodies nearby. Chroogomphus and Gomphidius

spp. are penetrating already established ectomycor-
brown-rot saprotrophic

rhizae of Suillus and Rhizopogon spp. and access

white-rot saprotrophic
stipitate-pileate, tubes
stipitate-pileate, gills

nutrition by sending haustoria into the rhizomorph

hyphae of the fungal host or even into the cortical
ectomycorrhizal
Trait

cells of the plant partner (Agerer 1987–1998, 1990,

Node

mycoparasitic
polyporoid

1991; Miller 1964; Olsson et al 2000). The parasites

resupinate
gasteroid
TABLE I.

produce clamydospores (asexual spores) inside the

ectomycorrhiza and initiate fruiting body formation
from primordia that develop on the rhizomorphs of
 BINDER AND HIBBETT: EVOLUTION OF BOLETALES 979

the exploited fungal host (Agerer 1990, 1991). The ACKNOWLEDGMENTS

nutritional mode in the Boletinellaceae is still some-
The authors are grateful to the organizers of the Deep
what elusive and appears to include generalists, root Hypha Research Coordination Network NSF 0090301 (M.
parasites and ectomycorrhizal fungi (Brundrett and Blackwell, J. Spatafora, J. Taylor) who invited us to
Kendrick 1987, Singer 1986). For example Phlebopus contribute to this issue of Mycologia. We are indebted to
tropicus forms lethal symbioses with scale insects Drs M.C. Aime, N. Arnold, V. Bandala, H. Besl, A. Bresinsky,
(Pseudococcus comstockii) that attack the roots of Citrus Y.-C. Dai, R. Halling, E. Horak, M. Jarosch, G. Lannoy, P.B.
trees (Singer 1986 p 744). This relationship is even Matheny, J. Trappe, R. Vilgalys, Z.-L. Yang and many others
more complex because it involves ants dispersing the who contributed materials. M. Pagel and A. Meade are
scale insects close to the roots (Singer 1986). The thanked for providing the program BayesMultiState. T.
Bruns and one anonymous reviewer improved the manu-
Boletineae contains a parasitic fungus that attacks the
script by providing helpful comments and suggestions. This
common earthball Scleroderma citrinum. Pseudoboletus research was supported by National Science Foundation
parasiticus is capable of forming ectomycorrhizal awards DEB-0444531 (MB and DSH, Global Boletales), DEB
associations, but this fungus is not efficient in nutrition 0228657 (DSH et al, AFTOL) and DBI-0320875 (DSH et al,
uptake (Raidl 1997). P. parasiticus enters S. citrinum Clark computer cluster).
rhizomorphs to exhaust the host fruiting body and
obtains ample nutrition to produce its own fruiting
bodies (Raidl 1997). The sister group of P. parasiticus LITERATURE CITED
in our nuc-lsu analyses is the agaricoid species Agerer R. 1987–1998. Colour Atlas of Ectomycorrhizae. 1st–
Phylloboletellus chloephorus, a rare fungus from South 11th delivery. Schwäbisch Gmünd, Germany: Einhorn
America and Mexico (Singer 1986). Bandala et al Verlag.
(2004) consider remnants of tropical deciduous forests ———. 1990. Studies on ectomycorrhizae XXIV. Ectomy-
as potential mycorrhizal partners in plantations where corrhizae of Chroogomphus helveticus and C. rutilus
P. chloephorus occurs. If it can be demonstrated that P. (Gomphidiaceae, Basidiomycetes) and their relation-
ship to those of Suillus and Rhizopogon Nova Hedwig
chloephorus parasitizes another Boletales species in
50:1–63.
a similar way as P. parasiticus does, then this would be
———. 1991. Studies on ectomycorrhizae XXXIV. Mycor-
a nice example of a phylogenetic inference predicting rhizae of Gomphidius glutinosus and of G. roseus with
the ecology of a fungus. some remarks on Gomphidiaceae (Basidiomycetes).
Nova Hedwig 53:127–170.
Conclusions.—The Boletales is a monophyletic group ———. 1999. Never change a functionally successful
of fungi and there is growing evidence that the principle: the evolution of Boletales s. l. (Hymenomy-
Atheliales is the sister group of Boletales. We resolve cetes, Basidiomycota) as seen from below ground
in our analyses eight major lineages of Boletales on features. Sendtnera 6:5–91.
suborder or family level. The Coniophorineae are Arpin N, Kühner R. 1977. Les grandes lignes de la
not monophyletic and need larger taxonomical classification des Boletales. Bull Soc Linn Lyon 46:83–
revision. The Paxillineae also are not monophyletic 108, 181–208.
but form a strongly supported clade with Boletineae. Bandala VM, Montoya L, Jarvio D. 2004. Two interesting
We therefore suggest merging both groups (Taxon- records of boletes found in coffee plantations in
eastern Mexico. Persoonia 18:365–380.
omy, SUPPLEMENT). The results of the MRCA analyses
Besl H, Bresinsky A. 1997. Chemosystematics of Suillaceae
show that the diversification of brown-rotting fungi and Gomphidiaceae (suborder Suillineae). Pl Syst Evol
poses critical events in the evolution of early 206:223–242.
Boletales, including multiple transformations from ———, ———, Kämmerer A. 1986. Chemosystematik der
resupinate to stipitate-pileate fungi. If Hydnomerulius Coniophoraceae. Z Mykol 52:277–286.
pinastri is derived from other brown-rot producing ———, Dorsch R, Fischer M. 1996. Zur verwandtschaftli-
species, then this suggests that brown-rot has a single chen Stellung der Gattung Melanogaster (Melanogas-
origin in Boletales. In addition, if this hypothesis is traceae, Basidiomycetes). Z Mycol 62:195–199.
correct, ectomycorrhizae have evolved at least twice Binder M, Besl H. 2000. 28S rDNA sequence data and
in Boletales. Mycoparasites in the Boletales represent chemotaxonomical analyses on the generic concept of
Leccinum (Boletales). A.M.B., Italy. Centro Studi
transitions from ectomycorrhizal lifestyles. The an-
Micologici, Micologia 2000:71–82.
cestral state of the nutritional mode of the Boletales
———, ———, Bresinsky A. 1997. Agaricales oder Bole-
was estimated as brown-rot and the ancestral mor- tales? Molekularbiologische Befunde zur Zuordnung
phology appears to be either polyporoid or re- einiger umstrittener Taxa. Z Mykol 63:189–196.
supinate. The analyses of the nuc-lsu dataset demon- ———, Bresinsky A. 2002a. Derivation of a polymorphic
strate the importance of this locus to identify lineage of gasteromycetes from boletoid ancestors.
phylogenetic key species. Mycologia 94:85–98.
980 MYCOLOGIA

———, ———. 2002b. Retiboletus, a new genus for a species- ———, Gilbert L-B, Donoghue MJ. 2000. Evolutionary
complex in the Boletaceae producing retipolides. instability of ectomycorrhizal symbioses in basidiomy-
Feddes Repert 113:30–40. cetes. Nature 407:506–508.
———, Hibbett DS, Larsson KH, Larsson E, Langer E, ———, Pine E, Langer E, Langer G, Donoghue MJ. 1997.
Langer G. 2005. The phylogenetic distribution of Evolution of gilled mushrooms and puffballs inferred
resupinate forms across the major clades of mush- from ribosomal DNA sequences. Proc Natl Acad Sci
room-forming fungi (homobasidiomycetes). Syst Biodiv USA 94:12002–12006.
3:113–157. Hughey BD, Adams GC, Bruns TD, Hibbett DS. 2000.
Both EE. 1993. The Boletes of North America. A compendi- Phylogeny of Calostoma, the gelatinous-stalked puffball,
um. Buffalo, New York: Buffalo Museum of Science. 436 p. based on nuclear and mitochondrial ribosomal DNA
Bresinsky A, Jarosch M, Fischer M, Schönberger I, Wittmann- sequences. Mycologia 92:94–104.
Bresinsky B. 1999. Polyphyletic relationships within Jarosch M. 2001. Zur molekularen Systematik der Boletales:
Paxillus s.l. (Basidiomycetes, Boletales): separation of Coniophorineae, Paxillineae und Suillineae. Bibl Mycol
a southern hemisphere genus. Plant Biol 1:327–333. 191:1–158.
Brundrett MC, Kendrick B. 1987. The relationship between ———, Besl H. 2001. Leucogyrophana, a polyphyletic genus
the ash bolete (Boletinellus merulioides) and an aphid of the order Boletales (Basidiomycetes). Plant Biol 3:
parasite on ash tree roots. Symbiosis 3:315–319. 443–448.
Bruns TD, Fogel R, White TJ, Palmer JD. 1989. Accelerated Kämmerer A, Besl H, Bresinsky A. 1985. Omphalotaceae
evolution of a false-truffle from a mushroom ancestor. fam. nov. und Paxillaceae, ein chemotaxonomischer
Nature 339:140–142. Vergleich zweier Pilzfamilien der Boletales. Plant Syst
———, Shefferson RP. 2004. Evolutionary studies of Evol 150:101–117.
ectomycorrhizal fungi: recent advances and future Kirk PM, Cannon PF, David JC, Stalpers JA. 2001. Ainsworth
directions. Can J Bot 82:1–11. and Bisby’s Dictionary of the Fungi. 9th ed. Cambridge,
———, Szaro TM. 1992. Rate and mode differences United Kingdom: CAB International University Press.
between nuclear and mitochondrial small-subunit
Kõljalg U, Larsson K-H, Abarenkov K, Nilsson RH, Alexander
rRNA genes in mushrooms. Mol Biol Evol 9:836–855.
IJ, Eberhardt U, Erland S, Hoiland K, Kjoller R, Larsson
———, ———, Gardes M, Cullings KW, Pan JJ, Taylor DL,
E, Pennanen T, Sen R, Taylor AFS, Tedersoo L, Vralstad
Horton TR, Kretzer AM, Garbelotto M, Li Y. 1998. A
T, Ursing BM. 2005. UNITE: a database providing web-
sequence database for the identification of ectomycor-
based methods for the molecular identification of
rhizal basidiomycetes by phylogenetic analysis. Mol
ectomycorrhizal fungi. New Phytol 166:1063–1068.
Ecol 7:257–272.
Kretzer AM, Bruns TD. 1999. Use of atp 6 in fungal
Claridge AW, Trappe JM, Castellano MA. 2001. Australasian
phylogenetics: an example from the Boletales. Mol
truffle-like fungi X. Gymnopaxillus (Basidiomycota,
Phyl Evol 13:483–492.
Austropaxillaceae). Aust Syst Bot 14:273–281.
Larsson KH, Larsson E, Kõljalg U. 2004. High phylogenetic
Felsenstein J. 2004. Inferring phylogenies. Sunderland,
diversity among corticioid homobasidomycetes. Mycol
Massachusetts: Sinauer Associates Inc. 662 p.
Res 108:983–1002.
Fischer E. 1899–1900. Lycoperdineae, Nidulariineae, Plec-
tobasidiineae (Sclerodermineae). In: Engler A, Prantl Lutzoni F, Pagel M, Reeb V. 2001. Major fungal lineages
K, eds. Die natürlichen Pflanzenfamilien nebst ihren derived from lichen-symbiotic ancestors. Nature 411:
Gattungen und wichtigeren Arten. Leipzig, Germany: 937–940.
Wilhelm Engelmann Verlag. p 313–342. Matheny PB, Curtis JM, Hofstetter V, Aime MC, Moncalvo J-
Fischer M, Binder M. 2004. Species recognition and host- M, Ge Z-W, Yang Z-L, Slot JC, Ammirati JF, Baroni TJ,
pathogen relationships: a case study in a group of Bougher NL, Hughes KW, Lodge DJ, Kerrigan RW,
lignicolous basidiomycetes, Phellinus s.l. Mycologia 96: Seidl MT, Aanen DK, DeNitis M, Daniele GM,
798–810. Desjardin DE, Kropp BR, Norvell LL, Parker A,
Gilbert JE. 1931. Les Bolets, in les livres du mycologue. Vellinga EC, Vilgalys R, Hibbett DS. 2006. Major clades
Vol. 3. Paris, 254 p. of Agaricales: a multilocus phylogenetic overview.
Gill M, Steglich W. 1987. Pigments of fungi (Macromy- Mycologia 98:984–997.
cetes). Prod Chem Org Nat Prod 51:1–317. Miller OK. 1964. Monograph of Chroogomphus (Gomphi-
Grubisha LC, Trappe JM, Molina R, Spatafora JW. 2001. diaceae). Mycologia 56:526–549.
Biology of the ectomycorrhizal genus Rhizopogon V. Moncalvo JM, Vilgalys R, Redhead SA, Johnson JE, James
Phylogenetic relationships in the Boletales inferred TY, Aime MC, Hofstetter V, Verduin SJW, Larsson E,
from LSU rDNA data. Mycologia 93:82–89. Baroni TJ, Thorn RG, Jacobsson S, Clémençon H,
Hibbett DS. 2004. Trends in morphological evolution in Miller OK. 2002. One hundred and seventeen clades of
homobasidiomycetes inferred using maximum likeli- euagarics. Mol Phyl Evol 23:357–400.
hood: a comparison of binary and multistate ap- Moser M. 1983. Die Röhrlinge und Blätterpilze (Polypora-
proaches. Syst Biol 53:889–903. les, Boletales, Agaricales, Russulales). 5th ed. In: Gams
———, Binder M. 2002. Evolution of complex fruiting-body H, ed. Kleine Kryptogamenflora IIb/2. Basidiomyce-
morphologies in homobasidiomycetes. Proc R Soc ten. Stuttgart, New York: Gustav Fischer Verlag. 432 p.
Lond B 269:1963–1969. Newman EI, Reddell P. 1987. The distribution of mycorrhi-
 BINDER AND HIBBETT: EVOLUTION OF BOLETALES 981

zas among families of vascular plants. New Phytol 106: Savile DBO. 1955. A phylogeny of the Basidiomycetes.
745–751. Can J Bot 33:60–104.
Olsson PA, Münzenberger B, Mahmood S, Erland S. 2000. Singer R. 1986. The Agaricales in modern taxonomy.
Molecular and anatomical evidence for a three-way 4th ed. Königstein, Germany: Koeltz Scientific Books.
association between Pinus sylvestris and the ectomycor- 981 p.
rhizal fungi Suillus bovinus and Gomphidius roseus. Smith AH, Thiers HD. 1971. The Boletes of Michigan.
Mycol Res 104:1372–1378. Ann Arbor, Michigan: University of Michigan Press.
Pagel M, Meade A, Barker D. 2004. Bayesian estimation of 428 p.
ancestral character states on phylogenies. Syst Biol 53: Swofford DL. 2002. PAUP*: phylogenetic analysis using
673–684. parsimony (*and other methods). Version 4.0b10.
Peintner U, Ladurner H, Simonini G. 2003. Xerocomus Sunderland, Massachusetts: Sinauer Associates.
cisalpinus sp. nov. and the delimitation of species in Szczepka MZ, Sokól S. 1984. Buchwaldoboletus lignicola (Kal-
the X. chrysenteron complex based on morphology and lenb.) Pilát and Phaeolus schweinitzii (Fr.) Pat. – das
rDNA-LSU sequences. Mycol Res 107:659–679. Problem ihres gemeinsamen Auftretens. Z Mykol 50:95–99.
Pilát A. 1969. Buchwaldoboletus genus novum Boletacearum. Thiers HD. 1984. The secotioid syndrome. Mycologia 76:
Friesia 9:217–218. 1–8.
Posada D, Crandall KA. 2001. Selecting the best-fit model of Trappe JM. 1975. A revision of the genus Alpova with notes
nucleotide substitution. Syst Biol 50:580–601. on Rhizopogon and the Melanogastraceae. Beih Nova
Raidl S. 1997. Studien zur Ontogenie an Rhizomorphen von Hedwig 51:270–309.
Ektomykorrhizen. Bibl Mycol 169:1–184. Watling R. 1970. Boletaceae: Gomphidiaceae: Paxillaceae.
Ronquist F, Huelsenbeck JP. 2003. MrBayes 3: Bayesian In: Henderson DM, Orton PD, Watling R, eds. British
phylogenetic inference under mixed models. Bioinfor- fungus flora. Agarics and Boleti I. Edinburgh, Scotland:
matics 19:1572–1574. Royal Botanical Garden. 124 p.