Content uploaded by Felix Hampe
Author content
All content in this area was uploaded by Felix Hampe on Jun 03, 2019
Content may be subject to copyright.
Into and out of the tropics: global diversification patterns
in a hyperdiverse clade of ectomycorrhizal fungi
BRIAN P. LOONEY,* MARTIN RYBERG,†FELIX HAMPE,‡MARISOL S
ANCHEZ-GARC
IA* and
P. BRANDON MATHENY*
*Department of Ecology and Evolutionary Biology, University of Tennessee, 332 Hesler Biology Building, Knoxville, TN 37996-1610,
USA, †Department of Organismal Biology, Uppsala University, Evolutionsbiologiskt Centrum, Norbyv. 18D, 75236 Uppsala,
Sweden, ‡Department of Biology, Gent University, K.L. Ledeganckstraat 35, 9000 Gent, Belgium
Abstract
Ectomycorrhizal (ECM) fungi, symbiotic mutualists of many dominant tree and shrub
species, exhibit a biogeographic pattern counter to the established latitudinal diversity
gradient of most macroflora and fauna. However, an evolutionary basis for this pattern
has not been explicitly tested in a diverse lineage. In this study, we reconstructed a
mega-phylogeny of a cosmopolitan and hyperdiverse genus of ECM fungi, Russula,
sampling from annotated collections and utilizing publically available sequences depos-
ited in GenBank. Metadata from molecular operational taxonomic unit cluster sets were
examined to infer the distribution and plant association of the genus. This allowed us
to test for differences in patterns of diversification between tropical and extratropical
taxa, as well as how their associations with different plant lineages may be a driver of
diversification. Results show that Russula is most species-rich at temperate latitudes
and ancestral state reconstruction shows that the genus initially diversified in temperate
areas. Migration into and out of the tropics characterizes the early evolution of the
genus, and these transitions have been frequent since this time. We propose the ‘gener-
alized diversification rate’ hypothesis to explain the reversed latitudinal diversity gradi-
ent pattern in Russula as we detect a higher net diversification rate in extratropical
lineages. Patterns of diversification with plant associates support host switching and
host expansion as driving diversification, with a higher diversification rate in lineages
associated with Pinaceae and frequent transitions to association with angiosperms.
Keywords: fungi, latitudinal diversity gradient, phylogeography, phyloinformatics, Russulaceae
Received 28 July 2015; revision received 10 November 2015; accepted 22 November 2015
Introduction
A long established global pattern of biogeography pro-
posed for macroorganisms is the latitudinal diversity
gradient (LDG), observed by early naturalists and cor-
roborated over several centuries in numerous studies
(Von Humboldt 1807; Hillebrand 2004). This pattern
has been supported for all major groups of macroflora
and fauna including plants, amphibians, mammals,
birds, reptiles, and marine and terrestrial invertebrates
(Hillebrand 2004). Alternatively, microbes have tradi-
tionally been considered to follow the ‘everything is
everywhere, but the environment selects’ model,
although some heterogeneity has been shown for cer-
tain groups (Baas-Becking 1934; Fontaneto et al. 2008).
At the interface of these two global distribution patterns
are fungi, which have traditionally been considered to
follow the microbial model but more recently been
found to be highly geographically segregated (Taylor
et al. 2006). Due to this intermediate position that fungi
have traditionally held, biogeographic patterns of fungi
have been poorly understood and have received less
attention (Lumbsch et al. 2008; Tedersoo et al. 2012).
Given recent advances in molecular methods for detect-
ing species from environmental samples, it is now
much more feasible to investigate patterns in their
global distribution (Tedersoo et al. 2014a).
Correspondence: Brian P. Looney, Fax: 865-974-3067;
E-mail: blooney@vols.utk.edu
©2015 John Wiley & Sons Ltd
Molecular Ecology (2016) 25, 630–647 doi: 10.1111/mec.13506
Recent studies have demonstrated that ectomycor-
rhizal (ECM) fungi exhibit a biogeographic pattern
counter to the LDG, where ECM fungal diversity
increases away from the tropics and towards the tem-
perate/boreal interface (Tedersoo et al. 2012, 2014a).
ECM fungi are obligate symbionts with plant roots of
primarily tree and shrub species, whereby the fungus
provides water and nutrients (viz, nitrogen and phos-
phorus) to the plant in exchange for photosynthates
(Alexopoulos et al. 1996). This symbiosis is necessary
for these fungi to complete their life cycle, and it is also
critical for their plant partners as this symbiosis pro-
vides a competitive advantage (Perry et al. 1989). There
are an estimated 25 000 ECM species worldwide, and
this biotrophic association has evolved independently
some 80 times primarily in the Ascomycota and Basid-
iomycota (Rinaldi et al. 2008; Tedersoo & Smith 2013).
ECM symbiosis with only those plant lineages that
allow ECM colonization makes ECM fungi an ideal
guild to investigate global biogeographic patterns, as
their biogeography and diversification patterns are
probably heavily influenced by the distribution, disper-
sion and diversification patterns of their plant partners
(Hoeksema 2010).
An initial meta-analysis by Tedersoo et al. (2012)
showed a reversal of the LDG in ECM fungi by analysis
of metadata from numerous fungal communities, and
following studies have highlighted potential ecological
drivers of this pattern. This work demonstrated that
ECM fungal richness peaks between 4000 and 4500 km
from the equator (36°–40.5°N/S). Edaphic, climatic and
biotic factors were tested in a multivariate model as
predictors for ECM species richness, of which several
were significant, including mean annual temperature,
mean annual precipitation, anthropogenic disturbance,
soil texture and ECM plant family. Soil volume had a
positive correlation with ECM species richness, which
has been proposed as a possible ecological driver for a
reversed LDG by allowing more stratification for niche
space (Peay et al. 2010; Smith et al. 2011; Kennedy et al.
2012). For a lineage level analysis, ECM plant family
explained 34% of the variation in ECM fungal commu-
nities, indicating that species tend to be segregated by
ECM plant partner (Tedersoo et al. 2012). A follow-up
study utilized a standardized sampling approach while
collecting environmental soil samples from 365 sites
around the globe (Tedersoo et al. 2014a). This latter
study confirmed the reversed LDG trend in ECM fungi
but upheld the standard LDG pattern for saprotrophic,
parasitic and pathogenic fungi (Tedersoo et al. 2014a).
The strongest predictors explaining global ECM rich-
ness in this analysis included the ratio of ECM plant
abundance relative to non-ECM plant abundance in a
community, total ECM plant species richness and soil
pH. While these studies have focused on identifying
ecological factors that predict ECM species richness,
evolutionary mechanisms that might help explain how
these factors contribute to the reversed LDG pattern
have been largely overlooked (Kennedy et al. 2012).
The ability to model the evolutionary dynamics
underlying the LDG has resulted in a number of testable
hypotheses that could be applied to ECM fungi. The
‘tropical conservatism hypothesis’ has been proposed as
a general explanation of the LDG, where lineages have a
tropical origin, and the conserved environmental niches
of these organisms restrict dispersal to the extratropics,
most likely due to climatic restraints (Latham & Ricklefs
1993; Wiens & Donoghue 2004). The ‘out of the tropics’
hypothesis proposes that the tropics can act simultane-
ously as a museum and a cradle for these lineages,
where dispersal events to the extratropics are frequent
yet the lineages will still concurrently persist and diver-
sify in the tropics (Jablonski et al. 2006). The ‘diversifica-
tion rate hypothesis’ proposes that a higher net
diversification rate in the tropics is driving the LDG,
whether due to a higher rate of molecular evolution,
stable climatic conditions over evolutionary time, or
periods of tropical expansion in the evolutionary past
(Rohde 1992; Jansson et al. 2013). These three hypotheses
have been proposed as a nested hierarchy, with the
‘tropical conservatism hypothesis’ being the most restric-
tive (Kerkhoff et al. 2014). S
anchez-Ram
ırez et al. (2015a)
recently tested for an evolutionary pattern to explain the
reversed LDG in a clade of the ECM genus Amanita and
found that temperate lineages have a higher speciation
rate. This study seeks to further test for these patterns in
a hyperdiverse genus of ECM fungi.
Russula is the largest genus in the order Russulales
comprising some 750–900 described species (Kirk et al.
2008; Buyck & Atri 2011). Russula can, therefore, be con-
sidered the second most taxonomically diverse genus of
ECM fungi after the genus Cortinarius (Kirk et al. 2008).
The genus is a dominant ECM lineage in tropical, tem-
perate, boreal and tundra ECM communities (Singer
1986; Buyck et al. 1996; Geml et al. 2009). The Russu-
laceae has also been hypothesized to have a tropical
origin (Buyck et al. 1996), which according to estab-
lished biogeographic hypotheses (Wiens & Donoghue
2004; Jablonski et al. 2006; Jansson et al. 2013), would
suggest the family should be most diverse in the trop-
ics. Indeed, Tedersoo & Nara (2010) found the/russula-
lactarius lineage to be more diverse in tropical forests;
however, this conclusion was tentative as statistical
support was lacking. Members of the genus Russula are
ecologically diverse as they associate with every major
ECM plant lineage (Singer 1986), are host to myco-
heterotrophic members of Ericaceae and Orchidaceae
(Kennedy et al. 2011a), and occasionally have a
©2015 John Wiley & Sons Ltd
DIVERSIFICATION PATTERNS IN RUSSULA 631
gasteroid fruit body morphology which makes up a sig-
nificant proportion of the diet of many small mammals
(Lebel & Tonkin 2007). Phylogenetic relationships
within the genus have been proposed (Eberhardt 2002;
Miller & Buyck 2002; Buyck et al. 2008), but taxon
sampling and gene sampling have been sparse to date.
The first major phylogenetic treatment of the genus
identified six major clades using a single molecular
marker. Not unexpectedly, internodal support was lack-
ing for most higher-level relationships (Miller & Buyck
2002). A later multigene analysis of the family Russu-
laceae resolved four genera, but taxon sampling was
not adequate to resolve major clades within Russula
(Buyck et al. 2008). Because a multigene treatment with
sufficient taxon sampling is unavailable, a more robust
phylogenetic framework for the group is necessary to
investigate the history of their diversification.
The objectives of this study are to: (i) produce a
robust phylogeny of the genus Russula as a basis to
investigate its patterns of diversification; (ii) utilize clus-
tering of global sampling and metadata associated with
DNA sequences of Russula to resolve its global distribu-
tion and ECM plant associations; (iii) use ancestral state
reconstruction methods to infer the evolutionary history
of its biogeography and plant association; and (iv) com-
pare biogeographic models to infer rates of diversifica-
tion and transitions in biogeographic states and plant
associations. By examining the history of diversification
of a large genus of ECM fungi, we seek to understand
what general evolutionary patterns exist and whether
co-evolution or host switching might be driving this
pattern at a large scale.
Materials and methods
Taxon sampling, DNA sequencing and phylogenetic
analyses of the core data set
Vouchered specimens from North America and Europe
were sequenced to infer a multigene phylogeny of Rus-
sula (Table S1, Supporting information). To ensure sam-
pling of wide phylogenetic diversity, type species of
major infrageneric groups were targeted from three of
the most relevant infrageneric classification systems
proposed for Russula (Romagnesi 1967; Singer 1986; Sar-
nari 1998). Full morphological descriptions with colour
notes (Kornerup & Wanscher 1967) were made for iden-
tification of all specimens. Specimens were dehydrated
and deposited at the TENN and GENT herbaria
[herbarium abbreviations per Thiers (continuously
updated)].
DNA extraction and PCR protocols followed that of
Birkebak et al. (2013). Four loci were targeted for infra-
generic clade-level resolution including two nrDNA
regions (nuclear ribosomal large subunit (LSU) and
internal transcribed spacers (ITS) and two single-copy
genes (rpb1 and rpb2, which encode the largest and sec-
ond largest subunits of RNA polymerase II, respec-
tively). We refer to this alignment as the ‘core data set’.
The following primer pairs were used for amplification:
ITS using ITS1F–ITS4 (White et al. 1990; Gardes & Bruns
1993); LSU using LR0R–LR5 (Vilgalys & Hester 1990);
rpb2 using b6F–b7.1R (Matheny 2005); and rpb1 using
gAf–fCr (Matheny et al. 2002) with int2F and int2.1R as
internal sequencing primers. Sequences were assembled
using SEQUENCHER 4.9 (Gene Codes, Ann Arbor, MI,
USA).
Alignments incorporating multilocus data from previ-
ous systematic studies (Buyck et al. 2008; Van de Putte
et al. 2012; Looney 2015) were constructed separately
for each gene region using MAFFT 6.717 (Katoh & Toh
2008) using the L-INS-i algorithm and manually
adjusted in MACCLADE 4.08 (Maddison & Maddison
2005). Intergene conflict was investigated by inferring
phylogenies for each locus using RAXMLGUI 1.2 (Sta-
matakis et al. 2008; Silvestro & Michalak 2012) and man-
ually inspecting topologies to ensure that the same
major groupings were recovered. Data sets were then
concatenated in SEAVIEW 4.3.0 (Gouy et al. 2010) to con-
struct a supermatrix alignment. Regions of the ITS data
set with ambiguous site alignments were excluded (sites
100–112, 269–284, 302–319, 550–562, 816–904). PARTITION-
FINDER 1.0.1 (Lanfear et al. 2012) determined the optimal
evolutionary models and partition scheme for a
partitioned analysis for both the core data set and
mega-phylogeny. The alignment for the core data set is
available online at Dryad Digital Depository (http://
datadryad.org).
A multigene phylogeny was inferred using RAXMLGUI
1.2 (Stamatakis et al. 2008; Silvestro & Michalak 2012)
executing 1000 rapid ML bootstraps replicates
(Table S2, Supporting information). For further assess-
ing clade support, MRBAYES 3.2 (Ronquist et al. 2012) was
used for a Bayesian analysis of 1 000 000 generations
using default priors until the standard deviation of split
frequencies reached below 0.01. Outgroups were
selected from the remaining three genera in the family
Russulaceae: Lactifluus deceptivus,Lactarius lignyotus and
Multifurca zonaria. Bootstrap values >70% and posterior
probabilities >0.95 are considered as evidence for
strongly supported relationships.
Clustering analyses of environmental sequences and
metadata acquisition
All putative ITS sequences of the Russula clade, includ-
ing the genus Russula and associated sequestrate genera
Macowanites, Cystangium, Gymnomyces and Martellia,
©2015 John Wiley & Sons Ltd
632 B. P. LOONEY ET AL.
were extracted from GenBank using the bioinformatics
program emerencia (Ryberg et al. 2009). To ensure ade-
quate statistical power for diversification analyses and
minimize the effects of low taxon sample size and high
character state bias, we assembled a data set including
>300 species using traits representing a minimum of
10% of the sampling, as suggested by Davis et al.
(2013). Sequences were screened for chimeric assembly
using a chimera checker (Nilsson et al. 2010) and manu-
ally pruned if sequence quality was low, indicated by
either long strings of ambiguous nucleotides or having
>50% missing data. This data set is hereafter referred to
as the ‘GenBank data set.’
Two rounds of clustering analyses were performed
on the GenBank data set to define molecular operational
taxonomic units (MOTUs): one using cd-hit (Li & God-
zik 2006) with a 99% identity and 80% coverage thresh-
old and a second using CLUSTERTREE 1.0 (deposited in
Dryad) with a 0.02 branch length cut-off using phyloge-
nies inferred in FASTTREE (Price et al. 2009). FASTTREE was
also used to visualize alignment quality and identify
dubious sequences for exclusion based on extremely
long branches. Representative sequences from each
cluster were selected based on greatest sequence cover-
age, lowest number of polymorphic sites and whether
they were identified to species. Representative
sequences were aligned in MAFFT and then manually
edited in MACCLADE. Due to the size of the data set and
the variability of the region across the Russula clade,
ClustalW was used to automatically align specific
regions using SEAVIEW.
Biogeographic and ECM plant associate data were
extracted from GenBank using a custom Perl script and
by manually reading through primary literature. Bio-
geographic coding for tropical vs. extratropical used the
latitudinal cut-off of the 23.5°parallels, and regional
coding was performed by continent with the Middle
East partitioned as the Eurasian territories from the
Arabian Peninsula north through Turkey and east
through Iran due to this region’s intermediate position
between Europe and the majority of Asia. ECM plant
associates were inferred if the plant associate was
reported in GenBank, a sequence was derived from a
known root tip, or if the sample was reported from a
monodominant forest (i.e. oak forest, well-described
hardwood forest with no potential Pinaceae hosts, pine
plot, etc.). The plant associates for clusters were used to
determine the maximum level of host specificity of
MOTUs supported by global sampling then coded as
Pinaceae, angiosperm or generalist (i.e. associating with
both Pinaceae and angiosperms) associates for the gen-
eral data set. For a more refined analysis, MOTUs were
also coded by ECM plant family, with generalist MOTU
clusters coded as angiosperm or generalist as necessary.
Mega-phylogeny and BEAST analyses
The core data set was used as a backbone topology in
RAXML to preserve higher-level relationships after merg-
ing the multilocus date with the GenBank data set.
Additional gene sampling from clustered GenBank data
(i.e. LSU, rpb1 and rpb2) was incorporated into the
supermatrix to estimate a mega-phylogeny. This was
accomplished by aligning and concatenating associated
sequences of LSU, rpb1 and rpb2 from any sequence of
the same cluster set/MOTU (Smith et al. 2009). Using a
backbone topology in RAXML allowed environmental
MOTUs to be added to the starting tree using a maxi-
mum parsimony (MP) criterion. The tree was then opti-
mized under normal ML parameters. The constrained
mega-phylogeny was then ultrametricized using the
Powell algorithm for nonparametric rate smoothing
implemented in R8S1.7 (Sanderson 2003). The core data
set was then excluded from the analysis to prevent
taxon redundancies using the drop.tip function in the
‘ape’ package in R(Paradis et al. 2004).
To infer the crown ages of Russula and its major
clades, the core data set was aligned with previously
published multigene data sets of Russulaceae (Buyck
et al. 2008; Van de Putte et al. 2012; Looney 2015) and
outgroups through the AFTOL project (aftol.org). A
chronogram of Russulaceae was inferred from three
independent runs in BEAST 2 with 50 000 000 genera-
tions and a burn-in of the first 10% of trees generated
so that all ESS values exceeded 200 (Table S3, Support-
ing information). Secondary calibrations were taken
from Floudas et al. (2012) using normally distributed
mean age estimates of Russulales, Boletales, Agaricales,
Agaricomycetidae and the ancestral node of all three
orders.
Ancestral state reconstructions and diversification
analyses
Ancestral state reconstruction was performed using
MP and ML approaches in MESQUITE 2.75 (Maddison &
Maddison 2001) and Bayesian estimation in BAYESTRAITS
V2 (Pagel et al. 2004). Significance in the ML and
Bayesian analyses was determined by comparison of
the negative log-likelihood of the character states with
a difference threshold of 2. To test whether biogeo-
graphical range or plant association p is conserved in
clades, the distributions of the traits on the phylogeny
were tested for phylogenetic conservatism using PHY-
LOCOM 4.2 (Webb et al. 2008). Mann–Whitney U-tests
were performed under different assumed sampling
biases by incrementally reducing the biased state ages
by 10% for geography and plant association character
sets to test for differences in mean ages using the
©2015 John Wiley & Sons Ltd
DIVERSIFICATION PATTERNS IN RUSSULA 633
‘STATS’ package version 3.2.1 in R. Diversification rates
associated with geography and plant association were
analysed using the binary state speciation and extinc-
tion (BiSSE) model, the BiSSE–node enhanced state
shift (BiSSE-ness) model for detecting cladogenetic
shifts associated with character states, and the geo-
graphic state speciation and extinction (GeoSSE)
model, a variant of the BiSSE model that allows spe-
cies to occupy both binary states simultaneously (i.e.
widespread or generalist). SSE analyses were imple-
mented in the Rpackage ‘DIVERSITREE’ (Maddison et al.
2007; FitzJohn et al. 2009; Goldberg et al. 2011; Magnu-
son-Ford & Otto 2012). Maximum-likelihood outputs
from the models were tested and compared using the
‘anova’ function in R, and parameter estimates were
found using a Markov chain Monte Carlo (mcmc)
method using 1000 steps. To test for the effects of
sampling bias for character states, 10 iterations of the
mcmc analysis were performed with assumed sam-
pling biases at 10% increments for 1000 steps imple-
mented in the ‘DIVERSITREE’ package in R(Table S4,
Supporting information). Finally, a BAMM approach
(Rabosky et al. 2014) for trait independent analysis of
diversification rate shifts was employed using the
BAMMtools package in Rto minimize the problem
highlighted by Rabosky & Goldberg (2015) in which a
single shift in diversification rate in a single diverse
clade can bias estimates for that trait throughout the
entire tree (Table S5, Supporting information).
Results
Analysis of clustered MOTUs suggests both
biogeographic distribution and plant association are
phylogenetically overdispersed
A total of 3510 ITS sequences of Russula were extracted
from GenBank, with 3337 sequences resulting from a
search for ‘Russula’ and 173 from searches for ‘Macow-
anites’, ‘Cystangium’, ‘Gymnomyces’ and ‘Martellia’. A
total of 162 sequences were excluded from the analyses
due to low sequence quality, low coverage or as chi-
meric sequences. From the initial total, 21.6% of the
sequences were already identified to species while
78.4% represented unidentified or environmental
sequences from soil or root sampling.
Clustering analyses resulted in a phylogenetic tree
with 1064 MOTUs (Fig. 1). Of these, 202 were unique
to the tropics, 844 were extratropical and 18 were
found in both areas. An analysis of phylogenetic con-
servatism of geographic states across the mega-phylo-
geny showed that biogeographic states are
phylogenetically overdispersed with a net relatedness
index (NRI) of 2.1046 (P=0.017). At a continental
scale, North America had the greatest number of rep-
resented MOTUs at 441, with Europe and Asia also
having a high number of MOTUs at 295 and 225,
respectively (Fig. 2). Most tropical MOTUs (105 or
51%) were sampled from Africa. Over 93% of MOTUs
Angiosperm
Pinaceae
Generalist
Host association
NRI = –2.141*
Tropics
Extratropics
Cosmopolitan
Bioeography
NRI = –2.1046*
Crown
Russula
Compacta
Heterophylla
Nigricans
Archaea
Farinipes
Delica
Clades
Fig. 1 Maximum parsimony ancestral reconstructions of geography (left) and host association (right) along an ultrametric mega-phy-
logeny of environmental Russula MOTUs inferred using r8s. Major clades are designated by colour. Geographical and host tree meta-
data associated with MOTU clusters are designated with coloured lines at the tips, with equivocal tips inferred from the analysis.
Areas are coded dark blue for extratropical distribution, green for tropical distribution and orange for cosmopolitan distribution.
Plant association data are coded red for angiosperm association, aqua for Pinaceae association and black for generalist. Net related-
ness indices produced in phylocom indicate phylogenetic overdispersion for both character sets.
©2015 John Wiley & Sons Ltd
634 B. P. LOONEY ET AL.
were recovered as endemic to a single continent, with
the most range overlap detected between North
America and Europe, which shared 62 MOTUs.
For ECM plant associate data, 158 of the MOTUs
were recovered as associates of Pinaceae, 443 MOTUs
as angiosperm associates, 60 as generalists and 403
were equivocal with no metadata available. An analy-
sis of phylogenetic conservatism of plant associate
states across the mega-phylogeny shows these states
are also phylogenetically overdispersed with a NRI of
2.141 (P=0.016), indicating that plant association is
highly labile within clades. Russula MOTUs were
recovered from 16 different plant families, with 25%
of MOTUs associated with only Fagaceae, 24% with
Pinaceae, 12% as angiosperm generalists and 9% as
generalists (angiosperm and Pinaceae). Other notable
ECM plant families include the Fabaceae and Diptero-
carpaceae, which make up a large proportion of tropi-
cal plant associates, and the Myrtaceae that comprise
many of the south temperate associates in Australia.
Ancillary ecological roles were investigated, and 179
MOTUs (17%) were recovered as hosts for orchids or
achlorophyllous members of the Ericaceae (Table S6,
Supporting information). Fifty-three MOTUs (5%)
were recovered with a gasteroid morphology, and not
all of the members in these clusters shared this
morphology.
Ancestral reconstruction and molecular clock methods
suggest an extratropical origin of Russula associated
with angiosperms during the Palaeogene
The ancestral range of Russula was resolved with statis-
tical support from ML and Bayesian inference as extrat-
ropical (Table 1). The delica, nigricans, archaea and
farinipes clades were all resolved by ML as most likely
having an extratropical origin with statistical support.
The ancestor of heterophylla, russula, compacta and
crown clade was ambiguous according to ML analysis,
but it and all subtending major clades except the rus-
sula clade were resolved as tropical by MP analysis.
The compacta clade had the highest likelihood support
for a tropical ancestry. In a multistate reconstruction
separating north and south temperate, as well as
Neotropical and palaeotropical MOTUs, we could not
reject a palaeotropical origin for Russula. However, sup-
port was much higher for a north temperate origin. This
was true for most of the major clades except for delica,
farinipes, archaea and russula clades, which were all
significantly supported as having a north temperate ori-
gin. The four-state parsimony reconstruction agrees
with the binary model, where tropical groups origi-
nated in the palaeotropics.
Ancestral plant association was reconstructed as
ambiguous between angiosperm and Pinaceae under
370
8
10
72
104
203
226
441
13
11
74
105
225
295
> 20
6 - 20
1 - 5
Fig. 2 Map projection of the global distribution of Russula MOTUs. The areas of circles are scaled by the number of MOTUs relative
to total MOTUs recovered. Top numbers represent number of endemic MOTUs. Bottom numbers indicate total number of MOTUs.
Lines represent number of overlapping distributions for widespread MOTUs.
©2015 John Wiley & Sons Ltd
DIVERSIFICATION PATTERNS IN RUSSULA 635
ML and Bayesian analyses, but all major clades were
inferred as having an angiosperm association according
to MP. A multistate reconstruction of major ECM plant
families refuted an ancestral association with Myrtaceae
for Russula and some individual major clades, yet the
plant association reconstruction was ambiguous for all
other families. MP reconstruction of plant family associ-
ation supported either an ancestral association between
Pinaceae or Fagaceae for all temperate clades except the
russula clade, inferred as Fagaceae or Fabaceae. Faga-
ceae was inferred as the ancestral association for tropi-
cal clades under parsimony.
Using secondary time calibrations, Russula split from
Lactarius and Multifurca ca. 55 (41–60) million years
(MY) ago with a crown age of 44 (33–55) MY (Table S3,
Supporting information). Of the eight major clades,
heterophylla was inferred as the oldest group at 42 MY,
with compacta second oldest at 37 MY old (Table 1).
The youngest major clades inferred were the delica, far-
inipes, russula and crown clades, all around 30 MY old.
Comparisons using a Mann–Whitney U-test of the taxon
age for biogeographic ranges showed that tropical taxa,
on average, are significantly older, with an average age
of 7.8 MY, compared to extratropical taxa with an aver-
age age of 3.3 MY (Fig. 3). Accounting for potential
taxon sampling biases from the north temperate zone,
this effect holds true if our sampling misses <2 tropical
species for every one extratropical species (50% bias).
Angiosperm associates, with an mean age of 5.2 MY
old, were found to be, on average, significantly older
than Pinaceae associates with an average of 2.5 MY.
This effect holds true if sampling our sampling misses
<1.25 angiosperm associates for every Pinaceae associate
(20% bias).
State speciation–extinction models suggest higher rates
of diversification for Russula in the extratropics and in
association with Pinaceae
Model testing for GeoSSE, BiSSE and BiSSE-ness in an
ANOVA framework showed significant support for the
full model in four of the five data sets (Table 2). The
model best supported for the GeoSSE geography data
set was one that constrained transition rates between
character states to be equal, indicating that dispersal
between the tropics and extratropics is bidirectional.
The full models for all other analyses were supported
over models constraining speciation and extinction rates
to be equal and pure-birth models, demonstrating that
for all data sets diversification patterns differ between
character states and extinction rates should be esti-
mated. For both GeoSSE data sets, the best models were
supported over models that constrained combined
states to zero, indicating that speciation rates for
widespread and host generalist taxa should be esti-
mated. For BiSSE analyses, the full model was sup-
ported over models constraining transition rates as
equal, demonstrating that rates of biogeographical and
plant associate expansion or restriction are unidirec-
tional. Finally, the BiSSE-ness analysis for plant associa-
tion found the full cladogenic model supported over an
anagenesis model of diversification, indicating that host
switches are driving cladogenic events.
ML estimates of the best model were used as starting
values for Bayesian inference of model parameters
(Table 3). Rates of diversification were found to be sig-
nificantly higher in extratropical lineages than tropical
lineages, with extratropical lineages having a positive
diversification rate and tropical lineages having a mean
estimate of a negative rate, although we cannot reject a
neutral diversification rate (Fig. 4A). Diversification rate
estimates for host specificity support a higher diversifi-
cation rate with Pinaceae-associated taxa over angios-
perm-associated taxa, with angiosperm-associated taxa
having a negative diversification rate. However, we
were not able to reject Pinaceae-associated MOTUs
with a neutral diversification rate (Fig. 4B). Transitions
from Pinaceae association to angiosperm association
are estimated to occur at rates 15.3 times higher than
from angiosperm to Pinaceae. Diversification rate esti-
mates for biogeographic range indicates that wide-
spread taxa are diversifying at the same rate as those
restricted to either the tropics or extratropics (Fig. 4C).
Transition rates, however, are more biased towards
range contraction at rates 3.5 times higher than range
expansions. Diversification rate estimates for host speci-
ficity indicate that host generalists are diversifying
much faster than host specialists, with host specialists
having a negative diversification rate (Fig. 4D). Transi-
tion rates, however, are much more biased towards
host specialization with rates being 5.6 times higher
than range expansion events. These findings hold true
under moderate taxon sampling biases (Table S4, Sup-
porting information).
Discussion
Into and out of the tropics
Russula is among the most taxonomically diverse ECM
lineages in the tropics (Buyck et al. 1996; Tedersoo &
Nara 2010). However, ancestral area analyses support
that the genus is an ancestrally temperate group. In
addition, diversification rate analyses support a higher
net rate of diversification among taxa in extratropical
regions. This suggests a complex biogeographic history
for Russula, and likely most other ECM lineages, which
falls counter to the predictions of established
©2015 John Wiley & Sons Ltd
636 B. P. LOONEY ET AL.
Table 1 Crown ages and ancestral character states reconstructed for Russula and major clades
Clades
Age
Geography binary Host binary Geography 4-state Host family 6-state
MY
Geog
PP
PP
state
Geog
ML
ML
state
Geog
MP
Host
PP
PP
state
Host
ML
ML
state
Host
MP ML ML state MP ML ML state MP
Root 43.96 0.88* Temp 0.89* Temp Temp 0.50 Equi 0.50 Equi Angi 0.78/0.21* Ntem/Ptro Ntem 0.55 Fag Pin,Fab
Except del 43.52 0.79 Temp 0.89* Temp Temp 0.50 Equi 0.50 Equi Angi 0.78/0.22* Ntem/Ptro Ntem 0.54 Fag Pin,Fab
Het/rus/com/cro 43.18 0.64 Trop 0.88 Temp Trop 0.50 Equi 0.50 Equi Angi 0.48/0.52* Ntem/Ptro Ptro 0.53 Fag Fab
Nig
arc/far
42.78 0.88* Temp 0.90* Temp Temp 0.50 Equi 0.50 Equi Angi 0.88/0.12* Ntem/Ptro Ntem 0.49 Fag Pin,Fab
Het 42.17 0.50 Trop 0.86 Temp Trop 0.50 Equi 0.50 Equi Angi 0.37/0.64* Ntem/Ptro Ptro 0.5 Fag Fab
Rus/com/cro 42.14 0.65 Trop 0.84 Temp Trop 0.50 Equi 0.51 Angi Angi 0.45/0.55* Ntem/Ptro Ptro 0.49 Fag Fab
Nig/arc 37.88 0.87 Temp 0.93* Temp Temp 0.50 Equi 0.50 Equi Angi 0.95* Ntem Ntem 0.5 Equi Pin,Fab
Com 37.03 0.82 Trop 0.53 Temp Trop 0.50 Equi 0.50 Equi Angi 0.12/0.87* Ntem/Ptro Ptro 0.48 Fab Fab
Nig 36.65 0.80 Temp 0.93* Temp Temp 0.50 Equi 0.51 Pina Angi 0.98* Ntem Ntem 0.5 Equi Pin,Fab
Arc 33.73 0.69 Temp 0.90* Temp Temp 0.50 Equi 0.50 Equi Angi 0.97* Ntem Ntem 0.5 Equi Pin,Fab
Rus/cro 33.25 0.72 Temp 0.82 Temp Trop 0.50 Equi 0.50 Equi Angi 0.50/0.50* Ntem/Ptro Ptro 0.5 Equi Fab
Del 31.44 0.80 Temp 0.89* Temp Temp 0.50 Equi 0.50 Equi Angi 0.95* Ntem Ntem 0.5 Equi Pin,Fab
Far 30.57 0.83 Temp 0.91* Temp Temp 0.50 Equi 0.52 Angi Angi 0.97* Ntem Ntem 0.32 Fag Pin,Fab
Rus 30.29 0.85 Temp 0.86 Temp Temp 0.50 Equi 0.51 Angi Angi 0.91* Ntem Ntem 0.39 Fag Fab,Fag
Cro 29.94 0.64 Trop 0.72 Temp Trop 0.50 Equi 0.54 Angi Angi 0.44/0.55* Ntem/Ptro Ptro 0.57 Fag Fab
archaea clade (arc); compacta clade (com); crown clade (cro); farinipes clade (far); heterophylla clade (het); russula clade (rus); Temperate (Temp); Tropical (Trop); Angiosperm
associate (Angi); Pinaceae associate (Pina); Equivocal (Equi); nigricans (nig); delica (del); Geography 4-state, north temperate (Ntem), neotropics (Ntro), palaeotropics (Ptro) and
south temperate (Stem); Host family 6-state, Pinaceae (Pin), Betulaceae (Bet), Dipterocarpaceae (Dip), Fabaceae (Fab), Fagaceae (Fag) and Myrtaceae (Myr).
Bold indicates states that have a higher likelihood when multiple states are found significant.
*Significance based on the difference of –LnLik >2.
©2015 John Wiley & Sons Ltd
DIVERSIFICATION PATTERNS IN RUSSULA 637
biogeographic hypotheses (Wiens & Donoghue 2004;
Jablonski et al. 2006; Jansson et al. 2013).
The most recent common ancestor of Russula was prob-
ably an angiosperm associate that began to diversify ca.
40 MY ago during the Eocene in North temperate regions
of Eurasia and/or North America. The late Eocene
marked the beginning of transition to icehouse Earth con-
ditions where, despite large fluctuations in CO
2
levels,
Antarctic ice began to form and global climates began the
period of cooling leading to modern conditions (Lear
et al. 2008). It has been suggested that diversification of
ECM fungi was facilitated by an expanded niche space
caused by cooling climates (Bruns et al. 1998; Ryberg &
Matheny 2012), and Russula is a group that appears to be
well adapted to temperate climates and able to occupy
these novel niches. The early history of the group shows
the divergence of the delica, farinipes, archaea and nigri-
cans clades occur in the north temperate zone. There is
evidence for switches to the tropics in the ancestors of the
heterophylla, compacta, and crown clades with a major
reversal back to the extratropics in the most recent com-
mon ancestor of the russula clade. Since this early history,
transitions between the tropics and extratropics have
been frequent events in the evolutionary history of Rus-
sula with at least 47 independent shifts to the tropics and
a comparable number of shifts to the extratropics. Only
the compacta clade is composed of more tropical taxa
than extratropical taxa, thus representing the only major
tropical clade in Russula based on current sampling.
Several ECM clades have been hypothesized as tropi-
cal in origin (Matheny et al. 2009; Kennedy et al. 2012;
TEMPERATE TROPICAL
0 10203040
Age (millions of years)
PINACEAE ANGIOSPERM
0 10203040
Age (millions of years)
P ≤ 0.001
P ≤ 0.001
Fig. 3 Boxplot comparing average taxon age based on terminal
branch lengths of taxa from a secondarily time-calibrated
mega-phylogeny with ranges in the tropics or extratropics
(Top) and host association with Pinaceae or angiosperms
(Bottom). P-values resulted from nonparametric Mann–Whit-
ney U-tests.
Table 2 Model comparisons for BiSSE, BiSSE-ness and GeoSSE
analyses
Df lnLik AIC ChiSq Pr(>|Chi|)
GeoSSE geography models (Tropical vs. Extratropical)
full 7 1000.7 1987.3 NA NA
no.sAB 6 978.8 1945.5 43.8 0***
eq.div 5 934.2 1858.4 133.0 0***
no.mu 5 742.0 1474.1 517.3 0***
eq.trans 6 1000.3 1988.7 0.7 0.4
GeoSSE plant association models (Angiosperm vs. Pinaceae)
full 7 278.1 542.2 NA NA
no.sAB 6 276.1 540.1 4.0 0*
eq.div 5 248.0 485.9 60.3 0***
no.mu 5 66.3 122.5 423.6 0***
BiSSE geography models (Endemic vs. Widespread)
full 6 1121.9 2231.9 NA NA
eq.trans 5 1081.4 2152.9 81.0 0***
eq.div 4 1008.4 2008.8 227.0 0***
no.mu 4 957.7 1907.5 328.4 0***
BiSSE plant association models (Specific vs. Generalist)
full 6 561.0 1109.9 NA NA
eq.trans 5 480.7 951.5 160.5 0***
eq.div 4 397.2 786.5 327.4 0***
no.mu 4 512.9 1017.8 96.2 0***
BiSSE-ness plant association models (Angiosperm vs. Pinaceae)
full 10 623.9 1227.7 NA NA
no.trans 9 622.8 1227.5 2.2 0.1***
eq.div 8 432.8 849.6 382.1 0***
no.mu 9 618.1 1218.2 11.5 0***
no.pc 8 581.6 1147.2 84.6 0***
model with all parameters (full); model with transition rates
constrained as equal (eq.trans); model with no dual-state spe-
ciation (no.sAB); model with diversification constrained as
equal (eq.div); model with extinction constrained to 0 (no.mu);
model with no cladogenic diversification (no.pc).
*=P≤0.05; ** = P≤0.005; *** = P≤0.0005.
©2015 John Wiley & Sons Ltd
638 B. P. LOONEY ET AL.
Wilson et al. 2012; S
anchez-Ram
ırez et al. 2015b). The
palaeotropics, in particular, have been recognized as the
ancestral origin of Inocybaceae, Amanita sect. Caesareae,
and most ECM clades of Sclerodermatineae (Matheny
et al. 2009; Wilson et al. 2012; S
anchez-Ram
ırez et al.
2015b). The ECM Sebacinaceae is the only major ECM
lineage that has been shown explicitly to have a north
temperate origin (Tedersoo et al. 2014b). No ECM lin-
eages have yet been found endemic to the Neotropics,
nor have any groups been reconstructed with a
Neotropical origin. South temperate taxa in the family
Inocybaceae are largely derived from north temperate
progenitors, and Neotropical taxa have been shown to
have immigrated from elsewhere (Matheny et al. 2009).
Two lineages, Austropaxillus and ECM Hysterangiales,
have been inferred as having south temperate origins
(Hosaka et al. 2008; Skrede et al. 2011). While Russula
has been inferred as having a north temperate origin,
the family Russulaceae may have its origins in the trop-
ics, given that Lactifluus, an ECM genus of over 120 spe-
cies that has been hypothesized as the sister clade to
the rest of Russulaceae, is largely a tropical clade (Ver-
beken et al. 2011). In this case, Russula would represent
a major clade that diversified outside of its ancestral
range to a greater extent than the other major clades of
the ECM lineage (i.e. Russulaceae), similar to what has
been found in some other ECM lineages (Matheny et al.
2009; Kennedy et al. 2012).
Although ECM clades vary greatly in age, significant
diversification episodes have coincided with specific
geologic periods during the evolution of these groups.
The oldest ECM lineage (Tuberaceae, Ascomycota) is a
cosmopolitan group ca. 160 million years old (Bonito
et al. 2013) originating in the late Jurassic, while Aus-
tropaxillus (order Boletales) has been identified as a
young ECM lineage with a mean age of 22 million years
(Skrede et al. 2011). We recovered the ECM lineage Rus-
sulaceae to have a mean age of 76 million years, origi-
nating during the late Cretaceous, which is consistent
with ages of several ECM clades of Agaricales (Ryberg
& Matheny 2012). The crown age of Russula (44 MY)
during the Eocene corresponds with ages of many of
the major clades of Tuberaceae (30–54 MY) as well as
major clades within the ECM Sebacinaceae (30–45 MY)
(Bonito et al. 2013; Tedersoo et al. 2014b). Russula, there-
fore, conforms with an emerging pattern in which the
origin of ECM association is ancient, in this case ECM
evolving in the ancestor of Russulaceae during the late
Cretaceous, but diversification of the major extant
clades has occurred much more recently in the Eocene,
during which the global climate began cooling and tem-
perate conditions expanded.
Higher diversification rates in the extratropics explain
a reversal of the latitudinal diversity gradient (LDG)
For many of the proposed explanations of the LDG pat-
tern, biological justifications could also apply to groups
originating outside the tropics. The ‘biogeographical
conservatism hypothesis’ has been proposed as an alter-
native to the ‘tropical conservatism hypothesis’, which
suggests that thermal or climatic tolerances may restrict
groups to certain environmental niches regardless of
whether they originate in the tropics (Pyron & Burbrink
2009). As an alternative to the ‘out of the tropics’
model, an ‘into the tropics’ model would suggest that
lineages outside the tropics are not dispersal limited in
regard to the tropics but those lineages can continue to
diversify alongside endemic extratropical lineages for
an overall greater accumulation of species. Some pro-
cesses proposed for the ‘diversification rate hypothesis’
could also apply to groups with an extratropical ances-
try. This includes an accelerated rate of molecular
evolution, relatively stable climatic conditions, or an
expanded niche space due to biotic and abiotic factors.
However, the extratropics cannot be said to have seen
gross expansions compared with modern conditions
considering the relatively constant cooling trend of glo-
bal climates. For an explanation of this pattern
applied to nontropical groups as well, we propose the
Table 3 Maximum-likelihood estimates of parameters for the best model for BiSSE and GeoSSE analyses
sA sB sAB xA xB dAB dBA DA DB TB->A T ratio
GeoSSE Geog 28.9 6.6 63.8 27.2 8.2 0.9 0.9 1.7 1.6 0.4 1
GeoSSE Plant 11.0 33.3 14.2 16.1 35.0 1.4 20.8 5.1 1.7 19.5 15.3
BiSSE Geog 36.3 2.7 N/A 34.1 0 2.1 7.2 2.2 2.7 5.1 3.5
BiSSE Host 0.5 74.9 N/A 7.2 40.2 19.4 109 6.7 34.7 89.6 5.6
speciation rate A (sA); speciation rate B (sB); speciation rate for dual-state (sAB); extinction rate A (xA); extinction rate B (xB); disper-
sal rate from A to B (dAB); dispersal rate from B to A (dBA); net diversification rate A (DA); net diversification rate B (DB); transition
rate from B to A (TB->A); transition rate B to A divided by A (T ratio); GeoSSE Geog, Extratropical (A); Tropical (B); GeoSSE Plant,
Angiosperm (A); Pinaceae (B); BiSSE Geog, Endemic (A), Widespread (B); BiSSE Plant, Host Specific (A); Generalist (B).
©2015 John Wiley & Sons Ltd
DIVERSIFICATION PATTERNS IN RUSSULA 639
‘generalized diversification rate’ hypothesis, which
states that patterns of diversity can be explained by
regional abiotic or biotic factors that promote an
increased diversification rate regardless of the biogeo-
graphic origin of a group or dispersability into or out of
the region.
Diversification patterns in Russula support the ‘gener-
alized diversification rate’ hypothesis as an explanation
of the reversed LDG. Lineages of Russula in the extrat-
ropics exhibit a higher rate of net diversification as they
transition into and out of the tropics at relatively equal
rates. A pattern of phylogenetic niche conservatism has
been proposed as good support for the ‘tropical conser-
vatism hypothesis’, where we should expect tropical
lineages to disperse infrequently into the extratropics,
thus allowing tropical clades to diversify or persist as
long branches (Crisp & Cook 2012). We do see the trop-
ics acting as a museum with tropical taxa having a
much lower extinction rate and higher average species
ages (Fig. 3). However, we found the distribution of
tropical MOTUs to be phylogenetically overdispersed,
indicating that transitions have occurred into and out of
the tropics frequently during the last 40 MY. Addition-
ally, an ‘into the tropics’ model can be rejected as tran-
sition rates between the tropics and extratropics were
found to be equal in Russula. Diversification patterns in
Russula agree with the findings of S
anchez-Ram
ırez
et al. (2015a) that extratropical ECM taxa have a higher
speciation rate than tropical taxa; however, extinction
was indicated as a significant variable in our models for
−3 −2 −1 0 1 2 3
0.0 0.2 0.4 0.6 0.8 1.0 1.2
ExtT
Trop
T > E
−5 0 5 10 15 20 25
0.0 0.1 0.2 0.3 0.4 0.5
Angi
Pina
P > A
020406080
0.00 0.10 0.20 0.30
123456
0.0 0.2 0.4 0.6 0.8 1.0
Diversification/transition rate (per my)
Ende
Wide
Rest
Spec
Gen
Rest
Diversification/transition rate (per my)
Posterior probability distribution
Posterior probability distribution
(A) (B)
(C) (D)
Fig. 4 Posterior probability density means and standard error representing relative diversification (k-l) and dispersal rates for geo-
graphic and host state-specific models for Russula. (A) Estimates for extratropical MOTUs (light blue) and tropical MOTUs (orange)
with differential transition rates from tropics to extratropics (grey) for the best-supported equal transition GeoSSE model. (B) Esti-
mates for angiosperm-associated MOTUs (red) and Pinaceae-associated MOTUs (blue) with differential transition rates from Pinaceae
association to angiosperm association (grey) for the best-supported full-parameter GeoSSE model. (C) Estimates for endemic MOTUs
(green) and widespread MOTUs (purple) with differential transition rates of contraction from widespread to endemic (grey) for the
best-supported full-parameter BiSSE model. (D) Estimates for host-specific MOTUs (yellow) and host generalist MOTUs (pink) with
differential transition rates of restriction from host generalist to specialist (grey) for the best-supported full-parameter BiSSE model.
©2015 John Wiley & Sons Ltd
640 B. P. LOONEY ET AL.
Russula diversification. Given these trends, extinction in
tropical environments may be driven by an unavailabil-
ity of abundant niche space from fewer soil horizons,
more fragmented host distributions, and a lack of
community partitioning due to a lower host lineage
diversity (Tedersoo & Nara 2010).
Tedersoo et al. (2012) suggested that clade age might
explain why ECM fungi are more species rich at temper-
ate latitudes than in the tropics. If this is correct, then
temperate lineages should be older and more diverse
than tropical lineages. Kennedy et al. (2012) found no
support for the ‘clade age’ hypothesis in the ECM genus
Clavulina, which was found to be tropical in origin and
containing several derived temperate lineages. One of
these temperate lineages was found to be diversifying at
nearly 2.5 times the rate elsewhere in the tree. With a
north temperate origin, Russula provides a good test for
the ‘clade age’ hypothesis. Diversification patterns in
Russula reject the ‘clade age’ hypothesis and support an
overall higher diversification rate for extratropical taxa
as a generalized pattern, even when major clades are not
restricted to the tropics or extratropics. The ‘clade age’
hypothesis is also confounded as a generalizable pattern
for ECM fungi by the paucity of evidence for temperate
origins for a majority of diverse ECM lineages.
In the extratropics, Russula is characterized by high
speciation and extinction rates, indicating a high species
turnover evident by the low average age of extratropical
taxa. This finding is consistent with the prediction of
Buyck et al. (1996) that temperate ECM fungi may expe-
rience higher competition due to exposure to ‘foreign
invaders’, whereas the tropics act like a museum
because of the relative isolation from competition. We
find some evidence for latitudinal optima described by
S
anchez-Ram
ırez et al. (2015a) for Amanita sect. Cae-
sareae.InRussula, there is a much higher rate of transi-
tion to either the tropics or extratropics rather than
range expansion to both. This could indicate that the
subtropics represent a barrier for dispersal and that dif-
ferent adaptations are required for surviving in tropical
vs. extratropical habitats and mycorrhizal communities.
Host switching is an important driver of diversification
in Russula
To explain the reversal of the LDG in ECM fungi,
increased ECM plant diversity in temperate regions was
proposed as a driving evolutionary force, but neither
codiversification nor host switching has been investi-
gated in this context (Kennedy et al. 2012; Tedersoo
et al. 2012; P~
olme et al. 2013). The diversification of
major clades in Russula corresponds to the time of
diversification for major ECM plant lineages, including
Fagaceae, Betulaceae, Salicaceae, Malvaceae,Cistaceae
and Dipterocarpaceae (Bell et al. 2010). This is consistent
with the hypothesis that codiversification with hosts or
host switching may have been an important driver of
diversification for ECM fungi. Evolution of ECM plant
diversity makes sense as a driver for the reversed LDG
pattern in ECM fungi as several diverse ECM plant lin-
eages (e.g. Myrtaceae, Fagaceae and Pinaceae) have
their diversity centres outside the tropics (Pryor 1959;
Richardson 2000; Nixon 2006). ECM plant lineage asso-
ciation was found to be conserved in major clades in
the Agaricales, such as Cortinarius,Hygrophorus and Ino-
cybaceae, but conservation was not found in others
(Ryberg & Matheny 2012). If codiversification is an
important driver for ECM fungal diversity, then we
should expect Russula clades to be host-restricted to
particular plant lineages genera or families. We find
that plant association in Russula is not conserved by
plant lineage at the family level, evidenced by phyloge-
netic overdispersion and the lack of signal for inferring
ancestral plant associations. We also find support for a
model showing that host switching is driving clado-
genic events over an anagenic model of host diversifica-
tion. With these analyses combined, there is strong
evidence that host switching is an important driver for
diversification in Russula and is more plausible than a
codiversification scenario of diversification.
Although it has not been found to be an important
driver for diversification of Russula, there is some
evidence that codiversification may be an important
process for select ECM fungi and for ECM plant lin-
eages in general. ECM plants comprise select lineages
of Gnetaceae, Pinaceae and numerous lineages of
angiosperms, including members of Betulaceae, Dipte-
rocarpaceae, Fabaceae, Fagaceae, Juglandaceae, Myr-
taceae, Nothofagaceae and Salicaceae (Brundrett 2009)
among others. Many fungal lineages containing ECM
fungi, including Russulales, have been found to be
younger than the diversification of angiosperms (Hib-
bett & Matheny 2009). Consistent with these findings,
ancestral plant associates for a number of ECM lin-
eages, now including Russula, have been inferred as
angiosperm (Matheny et al. 2009; Ryberg & Matheny
2011; Wilson et al. 2012; Bonito et al. 2013). Only the
ECM Sebacinaceae has been recovered as having an
ancestral association with Pinaceae (Tedersoo et al.
2014b). Studies of the ECM plant genus Alnus have
found historical distributions consistent with their asso-
ciates, giving strong support for codispersal for this
plant lineage with their associates (Kennedy et al.
2011b; P~
olme et al. 2013). Another ECM plant group
that shows a strong signal of association and, poten-
tially, codiversification with its fungal partners is Pinus,
whose species are nearly ubiquitous with the ECM
genera Suillus and Rhizopogon (Bruns et al. 2002). Stud-
©2015 John Wiley & Sons Ltd
DIVERSIFICATION PATTERNS IN RUSSULA 641
ies looking at codiversification from the perspective of
species-rich ECM plant lineages, such as Quercus or
Eucalyptus, have not been attempted. Nonetheless, if
ECM fungi are codiversifying with their plant associ-
ates, this may be an important process for diversifica-
tion of ECM plants as particular host-specific fungal
associates may be necessary partners for those plant
lineages, whereas host switching may be a primary
process by which most ECM fungal lineages diversify.
A surprising result from the GeoSSE model compar-
ison of plant association was that MOTUs associated
with Pinaceae have higher speciation rates than the
ones associated with angiosperms. In this case, we can
see an evolutionary source–sink dynamics, where a
majority of species initially evolve as associates with
Pinaceae but preferentially switch to angiosperm hosts
where they either expand their host range or go extinct.
A potential mechanism to explain this pattern would be
orogenesis events that can act like a species pump simi-
lar to glacial refugia (Sedano & Burns 2010; Wang et al.
2012). Many Pinaceae species are montane and will
probably track elevational gradients as mountains are
uplifted. These events are ideal for populations not able
to track this migration due to dispersal limitation or
thermal tolerances to become isolated and either speci-
ate or switch to an angiosperm host. Populations that
are able to track Pinaceae associates may have opportu-
nities to host switch and speciate with other members
of Pinaceae in different life zones or community types
(i.e. pine to spruce dominant community) (Tang &
Ohsawa 1997). It is also probably that climate fluctua-
tions over geological time create this effect at the tem-
perate–boreal interface (Sandel et al. 2011). Russula
generalists that associated with both Pinaceae and
angiosperms have a higher diversification rate than
more host-specific species, which also indicates that
host switching or expansion may be more important
drivers than co-evolution with the plant associate.
Again, host specificity is characterized by an evolution-
ary source–sink dynamic, where speciation occurs with
generalist species, but their host ranges are frequently
restricted, which may increase extinction rates.
Potential for additional drivers of diversification
An important criticism of trait-based diversification anal-
yses broached by Rabosky & Goldberg (2015) is that a
hidden trait or traits may be driving diversification pat-
terns that, by chance, may be correlated with the trait
being tested. This criticism is not a concern for our latitu-
dinal assessment, as we are interested in analysing a pat-
tern explicitly to discover the evolutionary process. This
criticism is relevant when considering whether ECM plant
associate lineage is driving ECM fungal diversification,
but this issue is more a problem of interpretation than
any flaw in the models. ECM plant associations have
been proposed as potential drivers of the reversed LDG
pattern, and our results are consistent with this. How-
ever, this is not to say that other associated factors may
not be more important at other spatial scales, including
root stratification (Kalliokoski et al. 2010), mycorrhizal
root signalling (Felten et al. 2009), ability to associate
with arbuscular mycorrhizae (Kennedy et al. 2011a,b), or
even something external to the associate such as commu-
nity type (McGuire et al. 2013) or stratification of the soil
(Rosling et al. 2003). A final possibility is that key adap-
tations of the fungi may be playing a role in diversifica-
tion with different plant host lineages, as adaptive
radiations in fungi have been shown to be driven by a
combination of environmental opportunity and pheno-
logical adaptations to take advantage of that opportunity
(Gaya et al. 2015). For Russula, this may include adapta-
tions to labile characters such as changes in spore mor-
phology in response to changing environments such as
temperature and moisture for differential dispersability
and germination, pigmentation of the pileus cuticle as
protection against radiation or to attract animal dispersal
vectors (Eberhardt 2002), different suites of oxidative
enzymes for accessing nutrients in recalcitrant plant mat-
ter or expansions in small secreted proteins used in root
colonization (Kohler et al. 2015). By identifying traits that
support a pattern of diversification, we can develop
additional hypotheses to test for a ‘smoking gun’ trait, if
one exists.
Sampling and methodological considerations
Using a total data approach, we were able to achieve
maximal global sampling of Russula; however, there are
some caveats and biases inherent to this approach. The
total number of recovered Russula MOTUs (1064)
exceeds the number of currently accepted species in the
genus (750–900 spp.) indicating that numerous novel
species of Russula have not been formally described. The
majority of the GenBank studies evaluated here origi-
nated in North America or Europe, which have the
highest number of MOTUs. We recovered 441 MOTUs
in North America, which is near the total number of
species reported from both the USA (419 spp.) and Mex-
ico (66 spp.) (Kong et al. 2002; Buyck 2007). A high num-
ber of MOTUs (62) are shared between North America
and Europe, which closely agrees with the number of
species described from Europe that are also reported in
North America (87 spp.) (Buyck 2007). Although sam-
pling bias towards the extratropics was anticipated and
accounted for in our diversification analyses (Table S4,
Supporting information), this bias may not be as pro-
nounced given 1) the smaller land mass with available
©2015 John Wiley & Sons Ltd
642 B. P. LOONEY ET AL.
ECM habitat; 2) the lack of ECM plant richness in the
tropics; and 3) the number of tropical Russula taxa
described compared to MOTUs recovered from molecu-
lar sampling efforts. The recovered number of MOTUs
for tropical Africa (105) closely approximates the num-
ber of described species (129–165 spp.) (Buyck et al.
1996; Verbeken & Buyck 2002). The total MOTUs recov-
ered for the Neotropics (45) also matches well the num-
ber of described taxa from the region (42 spp.) (Buyck
et al. 1996). Species estimates for tropical Asia are more
difficult to obtain due to the application of traditional
European names to species from this area, but any bias
towards the extratropics in this region is probably offset
by a lack of sampling from the temperate Himalayan
region of south China where we should expect a high
diversity coinciding with a high number of ECM plant
lineages (Das et al. 2010). An assessment of Russula
diversity for tropical Asia should be an objective for
future studies. For south temperate sampling, we recov-
ered 74 MOTUs from Australia and New Zealand,
which exceeds the number of species described from
this region, given that the largest study in the genus
from this region describes 33 species (McNabb 1973). A
few disjunct distributions of MOTUs are probably
explained by local introductions from pine plantations
(Dickie et al. 2010). Six MOTUs were recovered as hav-
ing a holarctic distribution throughout North America,
Europe and Asia, three of which were independently
sampled by fourteen different GenBank studies
(Table S6, Supporting information).
Accounting for almost half of all of the MOTUs for
which it was possible to retrieve ECM plant associate
data, host preference for Russula strongly favours the
Fagaceae (165) and Pinaceae (157). This is not surprising
given that these families are the most species diverse
ECM plant lineages in north temperate regions (Pryor
1959; Nixon 2006). Russula sequences were detected
from 16 different plant families. Plant families where
Russula was not detected but where we might expect to
find Russula include Gnetaceae, Casuarinaceae and Cis-
taceae, which are mostly south temperate or tropical
lineages (Brundrett 2009; Tedersoo & P~
olme 2012). Ted-
ersoo et al. (2014b) hypothesized groups that associate
with more plant lineages should be older, but this is
not the case with Russula, a relatively young group that
associates with nearly all known ECM plant lineages.
We used a robust, multigene phylogeny as a guide
tree for the Genbank data set mega-phylogeny due to
the variability of the ITS region from which most of the
environmental data were based, which allowed the
conservation of higher-level relationships. The final
ultrametric topology was therefore dependent on rela-
tionships inferred based on the phylogeny of the core
data set, where some nodes were not supported by
bootstrapping or posterior probability. However, these
clades were resolved in both maximum likelihood and
Bayesian inference, giving some confidence for the
topology. Taxon sampling for the core data set was also
biased towards North American and European taxa due
to reliance on the major classification systems, which
are based on those regions.
Cluster sets were considered regardless of cluster size,
as excluding singleton cluster sets would reduce sam-
pling beyond the necessary limits for SSE models. The
calculated average of sequences per cluster set was 2.6,
with 79% of sequences coded as extratropical, 12% as
tropical and 9% as widespread. Given these sample sizes,
we can certainly infer presence data for all geography
and some hosts, but we cannot be certain that the full
geographic range or host is being captured for MOTUs.
There is a stronger bias for tropical samples being under-
sampled, where fewer studies have been conducted and
many MOTUs were only detected once. There must also
be a bias towards recovering more host specialists, as
there must be at least two sequences in the cluster set
with conflicting hosts to be considered a generalist. It is
probably that some MOTUs that could be considered
generalists were not coded as such because none of their
other hosts were sampled within their geographic range.
Given these limitations, only potential geographic disper-
sal and host switches can be tested. Also, as there is no
consistent sampling strategy for GenBank sequences,
there may be biases in our ability to detect rare taxa from
locations that have only had sampling done from fruit
body or root collections. Our approach, however, was
able to achieve much greater sampling than would be
possible without a worldwide network of sampling
researchers and sampling sites, and we propose that
efforts should continue to report metadata for sequence
data submitted to online data repositories and support
databases for global sampling data such as UNITE
(https://unite.ut.ee/), GBIF (http://www.gbif.org/) and
fungimap (http://fungimap.org.au/).
Conclusions
Investigation of diversification patterns in fungi is chal-
lenging given the immense diversity of these groups,
their cosmopolitan distributions and the necessity to
approximate complete global taxon sampling across a
phylogeny. Utilizing available sequence data from vari-
ous environmental sources can help mitigate these
challenges by allowing for a more complete assessment
of global diversity and more accurate estimation of
evolutionary patterns. Using state-specific diversification
models, we found strong support for the ‘generalized
diversification rate’ hypothesis as an evolutionary
process accounting for high extratropical diversity in
©2015 John Wiley & Sons Ltd
DIVERSIFICATION PATTERNS IN RUSSULA 643
Russula. Application of these models to other lineages of
fungi may confirm our findings as a generalizable pat-
tern. We also found evidence that host switching is an
important driver in Russula diversification, allowing us
to generate new hypotheses about trait-driven diversifi-
cation in fungi. For example, a finer-scale analysis com-
paring diversification between taxa from lowland
tropical forests and montane tropical forests or between
specific ECM plant lineages may indicate, counter to
most other guilds of fungi (Tedersoo et al. 2014a), that
climate effects are less important than host effects for
ECM fungi. Also, while this and other studies have
focused on evolutionary dynamics at the tropical inter-
face, the boreal–temperate interface has also been high-
lighted as an important biogeographic boundary, where
ECM begins to drop off northward as part of a unimodal
distribution (Tedersoo et al. 2012, 2014a;). Future studies
in Russula should examine this relationship to determine
whether the same evolutionary or ecological forces are
governing this pattern, especially considering that boreal
forests are composed of a higher density of ECM plants
than most temperate forests, while temperate systems
can contain a higher ECM plant species richness.
Acknowledgements
This work was financially supported by National Science Foun-
dation awards DEB-0949517 to PBM and a Doctoral Disserta-
tion Improvement Grant DEB-1501293 to BPL and PBM.
Research funding was also received from the University of
Tennessee through the Chancellor’s Award Fund. The Myco-
logical Society of America’s Mentor Travel grant is gratefully
acknowledged to finance travel to present and receive feedback
concerning this research. The authors would also like to thank
the Gulf States Mycological Society and Cumberland Mycologi-
cal Society for the opportunity to collect specimens for this
study. We would like to thank Slavom
ır Adam
c
ık (Slovak
Academy of Sciences) and Steven Miller (University of Wyom-
ing) for technical training in identification of Russula species.
We thank the curators and staff from herbaria at the University
of Gent (GENT) and the University of Tennessee (TENN) for
access to collections.
References
Alexopoulos C, Mims C, Blackwell M (1996) Introductory Mycol-
ogy. John Wiley & Sons, New York City, New York.
Baas-Becking LGM (1934) Geobiologie of Inleiding tot de Milieu-
kunde. WP Van Stockum & Zoon, The Hague, The Netherlands.
Bell CD, Soltis DE, Soltis PS (2010) The age and diversification
of the angiosperms re-revisited. American Journal of Botany,
97, 1296–1303.
Birkebak JM, Mayor JR, Ryberg M, Matheny PB (2013) A sys-
tematic, morphological and ecological overview of the
Clavariaceae (Agaricales). Mycologia,105, 896–911.
Bonito G, Smith ME, Nowak M et al. (2013) Historical biogeog-
raphy and diversification of truffles in the Tuberaceae and
their newly identified southern hemisphere sister lineage.
PLoS ONE,8, e52765.
Brundrett MC (2009) Mycorrhizal associations and other means
of nutrition of vascular plants: understanding the global
diversity of host plants by resolving conflicting information
and developing reliable means of diagnosis. Plant and Soil,
320,37–77.
Bruns TD, Szaro TM, Gardes M et al. (1998) A sequence data-
base for the identification of ectomycorrhizal basidiomycetes
by phylogenetic analysis. Molecular Ecology,7, 257–272.
Bruns TD, Bidartondo MI, Taylor DL (2002) Host specificity in
ecomycorrhizal communities: what do the exceptions tell us?
Integrative & Comparative Biology,42, 352–359.
Buyck B (2007) A new initiative towards the study of Russula
in the eastern USA. Pagine Di Micologia,27,81–86.
Buyck B, Atri NS (2011) A Russula (Basidiomycota, Russulales)
with an unprecedented hymenophore configuration from
northwest Himalaya (India). Cryptogamie, Mycologie,32, 185–
190.
Buyck B, Thoen D, Watling R (1996) Ectomycorrhizal fungi of
the Guinea-Congo Region. Proceedings of the Royal Society of
Edinburgh,104B, 313–333.
Buyck B, Hofstetter V, Eberhardt U, Verbeken A, Kauff F
(2008) Walking the thin line between Russula and Lactarius:
the dilemma of Russula subsect. Ochricompactae. Fungal
Diversity,28,15–40.
Crisp MD, Cook LG (2012) Phylogenetic niche conservatism:
what are the underlying evolutionary and ecological causes?
New Phytologist,196, 681–694.
Das K, Putte K, Buyck B (2010) New or interesting Russula from
Sikkim Himalaya (India). Cryptogamie, Mycologie,31, 373–387.
Davis MP, Midford PE, Maddison W (2013) Exploring power
and parameter estimation of the BiSSE method for analyzing
species diversification. BMC Evolutionary Biology,13,1–11.
Dickie IA, Bolstridge N, Cooper JA, Peltzer DA (2010) Co-inva-
sion by Pinus and its mycorrhizal fungi. New Phytologist,187,
475–484.
Eberhardt U (2002) Molecular kinship analyses of the agaricoid
Russulaceae: correspondence with mycorrhizal anatomy and
sporocarp features in the genus Russula.Mycological Progress,
1, 201–223.
Felten J, Kohler A, Morin E et al. (2009) The ectomycorrhizal
fungus Laccaria bicolor stimulates lateral root formation in
poplar and Arabidopsis through auxin transport and signal-
ing. Plant Physiology,151, 1991–2005.
FitzJohn RG, Maddison WP, Otto SP (2009) Estimating trait-
dependent speciation and extinction rates from incompletely
resolved phylogenies. Systematic Biology,58, 595–611.
Floudas D, Binder M, Riley R et al. (2012) The Paleozoic origin
of enzymatic lignin decomposition reconstructed from 31
fungal genomes. Science,336, 1715–1719.
Fontaneto D, Barraclough TG, Chen K, Ricci C, Herniou EA
(2008) Molecular evidence for broad-scale distributions in
bdelloid rotifers: everything is not everywhere but most
things are very widespread. Molecular Ecology,17, 3136–3146.
Gardes M, Bruns TD (1993) ITS primers with enhanced speci-
ficity for basidiomycetes-application to the identification of
mycorrhizae and rusts. Molecular Ecology,2, 113–118.
Gaya E, Fern
andez-Brime S, Vargas R et al. (2015) The adaptive
radiation of lichen-forming Teloschistaceae is associated with
sunscreening pigments and a bark-to-rock substrate shift.
©2015 John Wiley & Sons Ltd
644 B. P. LOONEY ET AL.
Proceedings of the National Academy of Sciences USA,122,
11600–11605.
Geml J, Laursen GA, Timling I et al. (2009) Molecular phyloge-
netic biodiversity assessment of arctic and boreal ectomycor-
rhizal Lactarius Pers. (Russulales; Basidiomycota) in Alaska,
based on soil and sporocarp DNA. Molecular Ecology,18,
2213–2227.
Goldberg EE, Lancaster LT, Ree RH (2011) Phylogenetic infer-
ence of reciprocal effects between geographic range evolu-
tion and diversification. Systematic Biology,60, 451–465.
Gouy M, Guindon S, Gascuel O (2010) SeaView version 4: a
multiplatform graphical user interface for sequence align-
ment and phylogenetic tree building. Molecular Biology and
Evolution,27, 221–224.
Hibbett DS, Matheny PB (2009) The relative ages of ectomycor-
rhizal mushrooms and their plant hosts estimated using Baye-
sian relaxed molecular clock analyses. BMC Biology,7,1–13.
Hillebrand H (2004) On the generality of the latitudinal diver-
sity gradient. The American Naturalist,163, 192–211.
Hoeksema JD (2010) Ongoing coevolution in mycorrhizal inter-
actions. New Phytologist,187, 286–300.
Hosaka K, Castellano MA, Spatafora JW (2008) Biogeography
of Hysterangiales (Phallomycetidae, Basidiomycota). Myco-
logical Research,112, 448–462.
Jablonski D, Roy K, Valentine JW (2006) Out of the tropics:
evolutionary dynamics of the latitudinal diversity gradient.
Science,314, 102–106.
Jansson R, Rodr
ıguez-Casta~
neda G, Harding LE (2013) What
can multiple phylogenies say about the latitudinal diversity
gradient? A new look at the tropical conservatism, out of the
tropics, and diversification rate hypotheses. Evolution,67,
1741–1755.
Kalliokoski T, Pennanen T, Nygren P, Siev€
anen R, Helmisaari
R-S (2010) Belowground interspecific competition in mixed
boreal forests: fine root and ectomycorrhiza characteristics
along stand developmental stage and soil fertility gradients.
Plant and Soil,330,73–89.
Katoh K, Toh H (2008) Recent developments in the MAFFT
multiple sequence alignment program. Briefings in Bioinfor-
matics,9, 286–298.
Kennedy AH, Taylor DL, Watson LE (2011a) Mycorrhizal
specificity in the fully mycoheterotrophic Hexalectris Raf.
(Orchidaceae: Epidendroideae). Molecular Ecology,20, 1303–
1316.
Kennedy PG, Garibay-Orijel R, Higgins LM, Angeles-Arguiz R
(2011b) Ectomycorrhizal fungi in Mexican Alnus forests sup-
port the host co-migration hypothesis and continental-scale
patterns in phylogeography. Mycorrhiza,21, 559–568.
Kennedy PG, Matheny P, Ryberg K et al. (2012) Scaling up:
examining the macroecology of ectomycorrhizal fungi. Molec-
ular Ecology,21, 4151–4154.
Kerkhoff AJ, Moriarty PE, Weiser MD (2014) The latitudinal
species richness gradient in New World woody angiosperms
is consistent with the tropical conservatism hypothesis.
Proceedings of the National Academy of Sciences USA,111,
8125–8130.
Kirk PM, Cannon PF, Stalpers JA (2008) The Dictionary of the
Fungi. CABI Publishing, Wallingford, UK.
Kohler A, Kuo A, Nagy LG et al. (2015) Convergent losses of
decay mechanisms and rapid turnover of symbiosis genes in
mycorrhizal mutualists. Nature Genetics,47, 410–415.
Kong A, Montoya A, Estrada-Torres A (2002) Russula herrerae,
a new species with marginal veil from Mexico. Mycologia,94,
290–296.
Kornerup A, Wanscher JH (1967) Methuen Handbook of Colour,
2nd edn., Methuen & Co Ltd, London.
Lanfear R, Calcott B, Ho SY, Guindon S (2012) PartitionFinder:
combined selection of partitioning schemes and substitution
models for phylogenetic analyses. Molecular Biology and Evo-
lution,29, 1695–1701.
Latham RE, Ricklefs RE (1993) Continental comparisons of
temperate-zone tree species diversity. In: Species Diversity in
Ecological Communities: Historical and Geographical Perspectives
(eds Ricklefs RE, Schluter D), pp. 294–314. University of Chi-
cago Press, Chicago, Illinois.
Lear CH, Bailey TR, Pearson PN, Coxall HK, Rosenthal Y
(2008) Cooling and ice growth across the Eocene-Oligocene
transition. Geology,36, 251–254.
Lebel T, Tonkin JE (2007) Australasian species of Macowanites
are sequestrate species of Russula (Russulaceae, Basidiomy-
cota). Australian Systematic Botany,20, 355–381.
Li W, Godzik A (2006) Cd-hit: a fast program for clustering
and comparing large sets of protein or nucleotide sequences.
Bioinformatics,22, 1658–1659.
Looney BP (2015) Molecular annotation of type specimens of
Russula species described by W.A. Murrill from the southeast
United States. Mycotaxon,129, 255–268.
Lumbsch HT, Buchanan PK, May TW, Mueller GM (2008) Phy-
logeography and biogeography of fungi. Mycological Research,
112, 423–424.
Maddison WP, Maddison D (2001) Mesquite: a modular system
for evolutionary analysis. Version 2.74. Available from http://
mesquiteproject.org.
Maddison D, Maddison WP (2005) MacClade 4: Analysis of
Phylogeny and Character Evolution. Version 4.08a. Sinauer
Associates Inc., Sunderland, Connecticut.
Maddison WP, Midford PE, Otto SP (2007) Estimating a binary
character’s effect on speciation and extinction. Systematic
Biology,56, 701–710.
Magnuson-Ford K, Otto SP (2012) Linking the investigations of
character evolution and species diversification. The American
Naturalist,180, 225–245.
Matheny PB (2005) Improving phylogenetic inference of mush-
rooms with RPB1 and RPB2 nucleotide sequences (Inocybe;
Agaricales). Molecular Phylogenetics and Evolution,35,1–20.
Matheny PB, Liu YJ, Ammirati JF, Hall BD (2002) Using RPB1
sequences to improve phylogenetic inference among mush-
rooms (Inocybe, Agaricales). American Journal of Botany,89,
688–698.
Matheny PB, Aime MC, Bougher NL et al. (2009) Out of the
Palaeotropics? Historical biogeography and diversification of
the cosmopolitan ectomycorrhizal mushroom family Inocy-
baceae. Journal of Biogeography,36, 577–592.
McGuire KL, Allison SD, Fierer N, Treseder KK (2013) Ectomy-
corrhizal-dominated boreal and tropical forests have distinct
fungal communities, but analogous spatial patterns across
soil horizons. PLoS ONE,8, e68278.
McNabb R (1973) Russulaceae of New Zealand 2. Russula Pers.
ex SF Gray. New Zealand Journal of Botany,11, 673–730.
Miller SL, Buyck B (2002) Molecular phylogeny of the genus
Russula in Europe with a comparison of modern infrageneric
classifications. Mycological Research,106, 259–276.
©2015 John Wiley & Sons Ltd
DIVERSIFICATION PATTERNS IN RUSSULA 645
Nilsson RH, Abarenkov K, Veldre V et al. (2010) An open
source chimera checker for the fungal ITS region. Molecular
Ecology Resources,10, 1076–1081.
Nixon K (2006) Global and neotropical distribution and diver-
sity of oak (genus Quercus) and oak forests. In: Ecology and
Conservation of Neotropical Montane Oak Forests (ed. Kappelle
M), pp. 3–13. Springer, Heidelberg, Germany.
Pagel M, Meade A, Barker D (2004) Bayesian estimation of
ancestral character states on phylogenies. Systematic Biology,
53, 673–684.
Paradis E, Claude J, Strimmer K (2004) APE: analyses of phylo-
genetics and evolution in R language. Bioinformatics,20, 289–
290.
Peay KG, Kennedy PG, Davies SJ, Tan S, Bruns TD (2010)
Potential link between plant and fungal distributions in a
dipterocarp rainforest: community and phylogenetic struc-
ture of tropical ectomycorrhizal fungi across a plant and soil
ecotone. New Phytologist,185, 529–542.
Perry D, Margolis H, Choquette C, Molina R, Trappe J (1989)
Ectomycorrhizal mediation of competition between conifer-
ous tree species. New Phytologist,112, 501–511.
P~
olme S, Bahram M, Yamanaka T et al. (2013) Biogeography of
ectomycorrhizal fungi associated with alders (Alnus spp.) in
relation to biotic and abiotic variables at the global scale.
New Phytologist,198, 1239–1249.
Price MN, Dehal PS, Arkin AP (2009) FastTree: computing
large minimum evolution trees with profiles instead of a
distance matrix. Molecular Biology and Evolution,26, 1641–
1650.
Pryor L (1959) Species distribution and association in Eucalyp-
tus. In: Biogeography and Ecology in Australia (eds Keast A,
Crocker RL, Christian CS), pp. 461–471. Springer Science,
Dordrecht, The Netherlands.
Pyron RH, Burbrink FT (2009) Can the tropical conservatism
hypothesis explain temperate species richness patterns? An
inverse latitudinal biodiversity gradient in the New World
snake tribe Lampropeltini. Global Ecology and Biogeography,
18, 406–415.
Rabosky DL, Goldberg EE (2015) Model inadequacy and mis-
taken inferences of trait-dependent speciation. Systematic
Biology,64, 340–355.
Rabosky DL, Grundler M, Anderson C et al. (2014) BAMM-
tools: an R package for the analysis of evolutionary dynam-
ics on phylogenetic trees. Methods in Ecology and Evolution,5,
701–707.
Richardson DM (2000) Ecology and Biogeography of Pinus. Cam-
bridge University Press, Cambridge, UK.
Rinaldi A, Comandini O, Kuyper TW (2008) Ectomycorrhizal
fungal diversity: separating the wheat from the chaff. Fungal
Diversity,33,1–45.
Rohde K (1992) Latitudinal gradients in species diversity: the
search for the primary cause. Oikos,65, 514–527.
Romagnesi H (1967) Les Russules D’Europe et D’Afrique du Nord.
Bordas, Paris, France.
Ronquist F, Teslenko M, van der Mark P et al. (2012) MrBayes
3.2: efficient Bayesian phylogenetic inference and model
choice across a large model space. Systematic Biology,61,
539–542.
Rosling A, Landeweert R, Lindahl B et al. (2003) Vertical distri-
bution of ectomycorrhizal fungal taxa in a podzol soil pro-
file. New Phytologist,159, 775–783.
Ryberg M, Matheny PB (2011) Dealing with incomplete taxon
sampling and diversification of a large clade of mushroom-
forming fungi. Evolution,65, 1862–1878.
Ryberg M, Matheny PB (2012) Asynchronous origins of
ectomycorrhizal clades of Agaricales. Proceedings of the
Royal Society London Series B: Biological Sciences,279, 2003–
2011.
Ryberg M, Kristiansson E, Sj€
okvist E, Nilsson RH (2009) An
outlook on the fungal internal transcribed spacer sequences
in GenBank and the introduction of a web-based tool for
the exploration of fungal diversity. New Phytologist,181,
471–477.
S
anchez-Ram
ırez S, Etienne RS, Moncalvo JM (2015a) High
speciation rate at temperate latitudes explains unusual diver-
sity gradients in a clade of ectomycorrhizal fungi. Evolution,
69, 2196–2209.
S
anchez-Ram
ırez S, Tulloss RE, Amalfi M, Moncalvo JM
(2015b) Palaeotropical origins, boreotropical distribution and
increased rates of diversification in a clade of edible ectomy-
corrhizal mushrooms (Amanita section Caesareae). Journal of
Biogeography,42, 351–363.
Sandel B, Arge L, Dalsgaard B et al. (2011) The influence of
Late Quaternary climate-change velocity on species ende-
mism. Science,334, 660–664.
Sanderson MJ (2003) r8s: inferring absolute rates of molecular
evolution and divergence times in the absence of a molecular
clock. Bioinformatics,19, 301–302.
Sarnari M (1998) Monographia Illustrata del Genere Russula in
Europa. Associazioni Micologica Bresadola, Trento, Italy.
Sedano RE, Burns KJ (2010) Are the Northern Andes a species
pump for Neotropical birds? Phylogenetics and biogeogra-
phy of a clade of Neotropical tanagers (Aves: Thraupini).
Journal of Biogeography,37, 325–343.
Silvestro D, Michalak I (2012) raxmlGUI: a graphical front-end
for RAxML. Organisms Diversity & Evolution,12, 335–337.
Singer R (1986) The Agaricales in Modern Taxonomy. Koeltz Sci-
entific Books, Koenigstein, Germany.
Skrede I, Engh IB, Binder M et al. (2011) Evolutionary history
of Serpulaceae (Basidiomycota): molecular phylogeny, histor-
ical biogeography and evidence for a single transition of
nutritional mode. BMC Evolutionary Biology,11,1–13.
Smith SA, Beaulieu JM, Donoghue MJ (2009) Mega-phylogeny
approach for comparative biology: an alternative to supertree
and supermatrix approaches. BMC Evolutionary Biology,9,1–
12.
Smith ME, Henkel TW, Aime MC et al. (2011) Ectomycorrhizal
fungal diversity and community structure on three co-occur-
ring leguminous canopy tree species in a Neotropical rain-
forest. New Phytologist,192, 699–712.
Stamatakis A, Hoover P, Rougemont J (2008) A rapid bootstrap
algorithm for the RAxML Web servers. Systematic Biology,57,
758–771.
Tang CQ, Ohsawa M (1997) Zonal transition of evergreen,
deciduous, and coniferous forests along the altitudinal gradi-
ent on a humid subtropical mountain, Mt. Emei, Sichuan,
China. Plant Ecology,133,63–78.
Taylor JW, Turner E, Townsend JP, Dettman JR, Jacobson D
(2006) Eukaryotic microbes, species recognition and the geo-
graphic limits of species: examples from the kingdom Fungi.
Philosophical Transactions of the Royal Society B: Biological
Sciences,361, 1947–1963.
©2015 John Wiley & Sons Ltd
646 B. P. LOONEY ET AL.
Tedersoo L, Nara K (2010) General latitudinal gradient of bio-
diversity is reversed in ectomycorrhizal fungi. New Phytolo-
gist,185, 351–354.
Tedersoo L, P~
olme S (2012) Infrageneric variation in partner
specificity: multiple ectomycorrhizal symbionts associate
with Gnetum gnemon (Gnetophyta) in Papua New Guinea.
Mycorrhiza,22, 663–668.
Tedersoo L, Smith ME (2013) Lineages of ectomycorrhizal
fungi revisited: foraging strategies and novel lineages
revealed by sequences from belowground. Fungal Biology
Reviews,27,83–99.
Tedersoo L, Bahram M, Toots M et al. (2012) Towards global
patterns in the diversity and community structure of ectomy-
corrhizal fungi. Molecular Ecology,21, 4160–4170.
Tedersoo L, Bahram M, P~
olme S et al. (2014a) Global diversity
and geography of soil fungi. Science,346,1–10.
Tedersoo L, Bahram M, Ryberg M et al. (2014b) Global bio-
geography of the ectomycorrhizal/sebacina lineage (Fungi,
Sebacinales) as revealed from comparative phylogenetic anal-
yses. Molecular Ecology,23, 4168–4183.
Thiers B (continuously updated) Index herbariorum: a global
directory of public herbaria and associated staff. New York
Botanical Garden’s virtual herbarium. http://sweet-
gum.nybg.org/ih.
Van de Putte K, Nuytinck J, Das K, Verbeken A (2012) Expos-
ing hidden diversity by concordant genealogies and mor-
phology—a study of the Lactifluus volemus (Russulales)
species complex in Sikkim Himalaya (India). Fungal Diver-
sity,55, 171–194.
Verbeken A, Buyck B (2002) Diversity and ecology of tropical
ectomycorrhizal fungi in Africa. In: Tropical Mycology. Vol. 1
Macromycetes (eds Watling R, Frankland JC, Ainsworth M
et al.), pp. 11–24. CABI Publishing, Wallingford, UK.
Verbeken A, Nuytinck J, Buyck B (2011) New combinations in
Lactifluus. 1. L. subgenera Edules,Lactariopsis, and Russulopsis.
Mycotaxon,118, 447–453.
Vilgalys R, Hester M (1990) Rapid genetic identification and
mapping of enzymatically amplified ribosomal DNA from
several Cryptococcus species. Journal of Bacteriology,172, 4238–
4246.
Von Humboldt A (1807) Essai sur la g
eographie des plantes:
accompagn
e d’un tableau physique des r
egions
equinoxiales,
fond
e sur des mesures ex
ecut
ees, depuis le dixi
eme degr
ede
latitude bor
eale jusqu’au dixi
eme degr
e de latitude australe,
pendant les ann
ees 1799, 1800, 1801, 1802 et 1803.
Wang L, Schneider H, Zhang X-C, Xiang Q-P (2012) The rise of
the Himalaya enforced the diversification of SE Asian ferns
by altering the monsoon regimes. BMC Plant Biology,12,1–9.
Webb CO, Ackerly DD, Kembel SW (2008) Phylocom: software
for the analysis of phylogenetic community structure and
trait evolution. Bioinformatics,24, 2098–2100.
White TJ, Bruns T, Lee S, Taylor J (1990) Amplification and
direct sequencing of fungal ribosomal RNA genes for phylo-
genetics. In: PCR Protocols: A Guide to Methods and Applica-
tions (eds Innis MA, Gelfand DH, Sninsky JJ, White TJ), pp.
315–322. Academic Press, New York City, New York.
Wiens JJ, Donoghue MJ (2004) Historical biogeography, ecol-
ogy and species richness. Trends in Ecology & Evolution,19,
639–644.
Wilson AW, Binder M, Hibbett DS (2012) Diversity and
evolution of ectomycorrhizal host associations in the
Sclerodermatineae (Boletales, Basidiomycota). New Phytolo-
gist,194, 1079–1095.
B.L., B.M. and M.R. jointly conceived of the study; B.L.
generated all sequence data; F.H. provided European
specimens and assisted B.L. with species identification;
B.L. analysed the data with input from B.M. and M.R.;
M.S. wrote Perl scripts for data manipulation; and B.L.
wrote the manuscript with input from all authors.
Data accessibility
All annotated sequences generated for this study are
deposited in GenBank with accession numbers given in
bold in Table S1 (Supporting information). Molecular
alignments, phylogenies and the CLUSTERTREE file pack-
age are deposited in Dryad: doi:10.5061/dryad.gn4p4.
All other relevant data are available through Supple-
mentary files.
Supporting information
Additional supporting information may be found in the online ver-
sion of this article.
Table S1 GenBank accession numbers of DNA sequences gen-
erated for this study.
Table S2 Phylogenetic relationships of Russula inferred from
nuclear ribosomal and single-copy (ITS, nrLSU, rpb1 and rpb2)
sequences derived from a maximum-likelihood analysis.
Table S3 Chronogram of Russulaceae inferred in BEAST 2. The
95% HPD Posterior probabilities >0.9 are reported.
Table S4 Taxon sampling bias runs at increments of 10% sam-
pling bias for 1000 MCMC generations each for the four data
sets depicted in Fig. 4: GeoSSE tropical vs. extratropical geog-
raphy (A), GeoSSE angiosperm vs. Pinaceae association (B),
BiSSE binary tropical/temperate endemicity vs. widespread
(C), and BiSSE binary angiosperm/Pinaceae specificity vs. gen-
eralist association.
Table S5 BAMM analysis showing (A) the top nine shift con-
figurations of the most credible shift configuration set. Red cir-
cles represent increases in diversification rate, while blue
circles represent slowdowns in diversification. The size of the
circle indicates how significant the shift is. The f values indi-
cate what proportion of the confidence can be assigned to that
particular scenario; and (B) a circle phylogeny of the best shift
configuration.
Table S6 Final MOTU clusters with associated metadata for
mycoheterotrophic parasitism, gastroid morphology, plant
association and geographic distribution.
©2015 John Wiley & Sons Ltd
DIVERSIFICATION PATTERNS IN RUSSULA 647
