Mycologia, 98(6), 2006, pp. 885–895.

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

Phylogeny of the Glomeromycota (arbuscular mycorrhizal fungi): recent

developments and new gene markers

Dirk Redecker1 symbionts. Within root cells arbuscular mycorrhizal

Philipp Raab (AM) fungi form typical tree-like structures, the
Institute of Botany, University of Basel, Hebelstaße 1, arbuscules (FIG. 1B), or hyphal coils. Some also
CH-4056 Basel, Switzerland produce storage organs, termed vesicles (FIG. 1C).
Although the different fungal genera were reported
to show some differences in the morphology of their
Abstract: The fungal symbionts of arbuscular mycor- intraradical structures (Merryweather and Fitter
rhiza form a monophyletic group in the true Fungi, 1998) it is generally not feasible to identify species
the phylum Glomeromycota. Fewer than 200 de- with these characters.
scribed species currently are included in this group. Glomeromycotan fungi produce relatively large
The only member of this clade known to form (40–800 mm) spores with layered walls, containing
a different type of symbiosis is Geosiphon pyriformis, several hundreds to thousands of nuclei (Becard and
which associates with cyanobacteria. Because none of Pfeffer 1993). Spores may be formed singly (FIG. 1A,
these fungi has been cultivated without their plant D), in clusters or aggregated in so-called sporocarps
hosts or cyanobacterial partners, progress in obtain- (FIG. 1E, Gerdemann and Trappe 1974). Similar to
ing multigene phylogenies has been slow and the most Zygomycota, the hyphae of glomeromycotan
nuclear-encoded ribosomal RNA genes have re- fungi lack regular septation.
mained the only widely accessible molecular markers. The phylum Glomeromycota comprises about 200
rDNA phylogenies have revealed considerable poly- described morphospecies that traditionally have been
phyly of some glomeromycotan genera that has been distinguished by features of the spore wall. The way
used to reassess taxonomic concepts. Environmental
the spore is formed on the hypha (‘‘mode of spore
studies using phylogenetic methods for molecular
formation’’) has been important to circumscribe
identification have recovered an amazing diversity of
genera and families, and the layered structure of the
unknown phylotypes, suggesting considerable cryptic
spore walls is used to distinguish species (Morton
species diversity. Protein gene sequences that have
1988). Walker (1983) established the concept of
become available recently have challenged the rDNA-
‘‘murographs’’ to describe and compare the layered
supported sister group relationship of the Glomer-
structure of the spore walls more easily. Morton
omycota with Asco/Basidiomycota. However the
(1995) and Stürmer and Morton (1997, 1999) in-
number of taxa analyzed with these new markers is
cluded considerations of the spore development to
still too small to provide a comprehensive picture of
group these wall components hierarchically into
intraphylum relationships. We use nuclear-encoded
complexes linked by ontogeny.
rDNA and rpb1 protein gene sequences to reassess
There is no evidence that the Glomeromycota
the phylogeny of the Glomeromycota and discuss
reproduce sexually. A study reporting the formation
possible implications.
of sexual zygospores by Gigaspora (Tommerup and
Key words: arbuscular mycorrhiza, evolutionary
Sivasithamparam 1990) was not confirmed. Studies
history, ribosomal RNA genes, RNA polymerase
using molecular marker genes have detected no
genes
genetic recombination or only low levels (Kuhn et al
2001). Therefore it is assumed generally that the
INTRODUCTION spores are formed asexually. There are conflicting
Arbuscular mycorrhiza is a widespread mutualistic reports on the question of whether the nuclei in the
symbiosis between land plants and fungi of the mycelium and spores of one organism are genetically
phylum Glomeromycota. The association is essential identical (Hijri and Sanders 2005, Kuhn et al 2001,
for plant ecosystem function because the great Pawlowska and Taylor 2004).
majority of plant species depend on it for mineral Glomeromycotan fungi are obligate symbionts.
nutrient uptake. This task is efficiently performed by Progress in obtaining new marker genes has been
the extensive extraradical mycelium of the fungal slow because their spores usually contain numerous
other microorganisms, including fungi from other
Accepted for publication 27 September 2006. phyla (Hijri et al 2002). Axenic fungal biomass can be
1
Corresponding author. E-mail: [email protected] obtained only from cultures on transformed plant

885
886 MYCOLOGIA

FIG. 1. Some characteristic morphological features of glomeromycotan fungi. A. Colonized roots of Plantago media with
hyphae and spores of Glomus clarum. B. Arbuscule of Glomus mosseae stained with chlorazol black. C. Vesicle of Glomus
mosseae. D. Spore of Glomus sp. S328 showing the hyphal attachment. E. Section of a sporocarp of Glomus sinuosum with
spores grouped around a hyphal plexus and covered by a layer of hyphae. F. Spore of Scutellospora cerradensis, showing
bulbous sporogenous cell and inner flexible walls with germination shield (arrow). Inset: germination shield of S. scutata in
face view. G. Germinating spore of Gigaspora decipiens with sporogenous cell, warty germination layer and germination hypha.
H. Spore of Acaulospora denticulata with tooth-like wall ornamentations and inner germinal walls. Spores in D, E, F, G and H
were embedded in polyvinylalcohol lactoglycerol and in F, G and H they were cracked under the cover slip. Images courtesy of
Kerstin Wex (B, C), Fritz Oehl (F) and the American Society for the Advancement of Science (D). Bars 5 100 mm (A, E, F, G),
50 mm (D, H), 5 mm (B, C).

roots, but only a small number of species are available Before 1974 most arbuscular mycorrhizal fungi
in culture. were in the genus Endogone until Gerdemann and
The presence of multiple, slightly differing variants Trappe (1974) placed them in four different genera
of the nuclear-encoded ribosomal RNA genes (rDNA) in the order Endogonales (Glomus, Sclerocystis,
within single spores is well established (Lanfranco et Gigaspora, Acaulospora). Morton and Benny (1990)
al 1999, Sanders et al 1995). A similar phenomenon established a new order ‘‘Glomales’’ in the Zygomy-
was reported for some protein genes (Kuhn et al cota, comprising six genera. Since then evidence has
2001) but not others (Stukenbrock and Rosendahl accumulated supporting the view that arbuscular
2005). The rDNA heterogeneity causes problems mycorrhizal fungi are distinct from other Zygomy-
when closely related species or isolates of the same cota. They do not seem to form characteristic
species are to be distinguished. It was shown recently zygospores, and in all cases when the nutritional
that the mitochondrial large ribosomal subunit gene mode has been elucidated they form mutualistic
does not show this variation (Raab et al 2005), symbioses. Based on their rDNA phylogeny AM fungi
suggesting that mitochondrial genes might be useful are the sister group of Asco- and Basidiomycota and
as molecular markers in the future. not monophyletic with any part of the Zygomycota.
Due to the problems outlined above there is Therefore the ‘‘Glomales’’ was raised to the rank of
currently no molecular species concept for glomer- a phylum Glomeromycota (Schüßler et al 2001). In
omycotan fungi. Nevertheless molecular markers have the same study the grammatically incorrect order
proven to be highly useful to characterize the diversity name ‘‘Glomales’’ was corrected to ‘‘Glomerales’’
of AM fungi in the field and have revealed an and several new orders were established. It must be
unexpectedly high diversity of phylotypes in some emphasized however that ‘‘Glomales’’ sensu Morton
settings. Some of these studies indicate that the and Benny (1990) is synonymous with the Glomer-
number of 200 described morphospecies might be omycota sensu Schüßler et al (2001) and ‘‘Glomer-
a strong underestimation of the true diversity of the ales’’ sensu Schüßler et al (2001) comprises a smaller
Glomeromycota (Husband et al 2002, Vandenkoorn- subset of these taxa.
huyse et al 2002). In general molecular phylogenies have shown that
 REDECKER AND RAAB: PHYLOGENY OF GLOMEROMYCOTA 887

glomeromycotan diversity at the phylum and genus tion process. The flexible walls are not naturally
level is much higher than expected through micro- pigmented but certain sublayers often can be
scopic observation of spore morphology (Redecker et stained with Melzer’s reagent. Gigaspora germi-
al 2000b, Schwarzott et al 2001). Some of the nates through the spore wall after a papillate
morphological characters that were used previously layer has formed on the inside of the spore wall
to delimit genera and families probably have evolved around the point of penetration (FIG. 1G). Note
multiple times independently. that this warty layer structure also has been
Ten genera of the Glomeromycota currently are referred to as ‘‘germinal wall’’ by some authors
distinguished: (Spain et al 1989), a usage of the term we will
not employ here.
(i) Glomus is the largest genus in the phylum,
(iv The diagnostic feature that was used for the
with more than 70 morphospecies. Spores,
and family Acaulosporaceae, comprising the genera
typically with layered wall structure, are formed
v) Acaulospora and Entrophospora, is the formation
by budding from a hyphal tip. The sporogenic
hypha (or ‘‘subtending hypha’’) often remains of spores next to a ‘‘sporiferous saccule’’. This
attached to the mature spore (FIG. 1D). The saccule collapses during spore maturation and
spores germinate through this hyphal attach- eventually disappears. Flexible inner walls (ger-
ment or the remains of it. This glomoid mode minal walls) are found in all members of the
of spore formation is symplesiomorphic and family. During spore germination a ‘‘germina-
occurs in several distinct lineages, namely tion orb’’ is produced on the inner walls,
Glomus, Paraglomus, Archaeospora, Pacispora, a membraneous structure that is instrumental
Diversispora and Geosiphon. Some of these in penetrating the outer spore wall.
genera were separated from Glomus based on The two genera of the Acaulosporaceae are
molecular phylogenetic data. Clades of Glomus distinguished by the position of the sporiferous
are distinguished as Glomus group A, B and C saccule. It is produced laterally in Acaulospora
(Schwarzott et al 2001). Groups A and B form and formed within the subtending hypha in
a monophyletic clade (FIG. 2A). The largest Entrophospora. The Entrophospora mode of spore
diversity of all Glomus lineages has been found formation occurs in several different clades.
in group A, which dominates AM fungal One, exemplified by E. colombiana, groups in
communities in many field settings (Helgason a clade with Acaulospora (FIG. 2A). E. schenckii is
et al 1998, Öpik et al 2003, Vandenkoornhuyse closely related to Archaeospora trappei (Redecker
et al 2002) and Morton unpubl). The phylogenetic position
Some species forming Glomus-type spores in of E. infrequens, the type species of the genus, is
complex sporocarps (FIG. 1E) were placed pre- unclear because rDNA sequences from several
viously in the genus Sclerocystis. Almeida and unrelated glomeromycotan lineages were re-
Schenck (1990) transferred all but one of these ported to occur within its spores (Rodriguez et
species to Glomus, based on morphological al 2001).
considerations. Molecular analysis of the last (vi) The genus Pacispora was established recently
remaining species of this genus (the type species for AM fungal species forming spores in the
S. coremioides) showed that it is nested within same way Glomus typically does but having
a clade of other well characterized Glomus flexible inner walls and a germination orb
species, therefore this species also was trans- (Oehl and Sieverding 2004). It was reported to
ferred to Glomus (Redecker et al 2000c). be basal to the Gigasporaceae in rDNA phylog-
(ii Gigaspora and Scutellospora are closely related enies (Walker et al 2004). The genus name
and genera in the family Gigasporaceae. Spores are Gerdemannia published for the same group only
iii) formed on a bulbous sporogeneous cell and a few weeks later is an illegitimate name based
germinate through a newly formed opening in on the publication date (Walker and Schüßler
the spore wall. The two genera do not form 2004).
vesicles within roots, and the extraradical myce- (vii) Another clade of Glomus species, referred to as
lium bears so-called auxiliary cells of unknown Glomus group C, is more closely related to the
function. In contrast to Gigaspora, species of Acaulosporaceae than to Glomus groups A and
Scutellospora possess flexible inner spore walls, B, based on rDNA phylogenies (Schwarzott et al
which are present permanently in mature spores 2001). One species of it has been described in
(‘‘germinal walls’’ and a ‘‘germination shield’’, a new genus Diversispora as D. spurca, mainly
FIG. 1F). These inner germinal walls and the based on ribosomal sequence signatures (Walk-
germination shield are involved in the germina- er and Schüßler 2004).
888 MYCOLOGIA
 REDECKER AND RAAB: PHYLOGENY OF GLOMEROMYCOTA 889

(viii) The genus Archaeospora was established for some at all in the standard procedure using acidic
AM fungi deeply divergent within the phylum in stains (trypan blue, direct blue) which is
rDNA phylogenies (Sawaki et al 1998, Redecker commonly used to view mycorrhizal coloniza-
et al 2000b, Morton and Redecker 2001). A. tion. The basal position of Archaeospora and
leptoticha and A. gerdemannii are dimorphic, Paraglomus is supported by unique fatty acids
producing both glomoid and acaulosporoid not found in other glomeromycotan fungi
spores. A glomoid morph of A. trappei was (Graham et al 1995).
reported only recently (Spain 2003). Most
Above the genus level the family Glomaceae was
isolates produce both types of spores at the
erected by Pirozynski and Dalpé (1989). Gigaspor-
same time. The acaulosporoid spores of Archae- aceae and Acaulosporaceae were established by
ospora show several characters distinct from Morton and Benny (1990). Morton and Redecker
Acaulospora. The sporiferous saccule is separat- (2001) described new families Archaeosporaceae and
ed from the hyphae by a short stalk (a Paraglomaceae to comprise deeply divergent lineages
‘‘pedicel’’), the flexible inner walls are much of AM fungi. Schüßler et al (2001) divided the new
thicker in A. leptoticha and A. gerdemannii, do phylum Glomeromycota into the orders Glomerales,
not stain with Melzer’s reagent and do not Diversisporales, Archaeosporales and Paraglomerales.
appear to be involved in the germination Together with the corresponding new genus Diversis-
process. pora, the family Diversisporaceae was erected (Walker
(ix) Geosiphon pyriformis is the only member of the and Schüßler, 2004).
phylum that is known to engage in a different In contrast to evolutionary trends in spore mor-
type of symbiosis (Schüßler et al 1994). It forms phology, differences among glomeromycotan
an endocytosis containing the cyanobacterium lineages with regard to symbiotic behavior or other
Nostoc punctiforme, harboring these photobionts properties have been more difficult to address.
in fungal bladders up to 2 mm large. It forms Different preferences with regard to nutrient trans-
glomoid spores and first was identified as a basal location and increase of root pathogen resistance
relative of AM fungi by rDNA phylogeny (Gehrig have been reported (Klironomos 2000). The foraging
et al 1996). The exact phylogenetic relationship of the extraradical mycelium has been compared
of Geosiphon relative to Paraglomus, Archaeospora among some species of Glomus (Jansa et al 2005). Soil
and the clade with the previously known three chemistry, agricultural practice and other parameters
families was not resolved well in earlier studies were shown to have an influence on certain sets of
(Redecker et al 2000b). An updated sequence of taxa (Helgason et al 1998, Hijri et al 2006). Ecological
Geosiphon in the databases now places this preferences have been reported in several molecular
fungus closer to Archaeospora leptoticha/gerde- field studies of AMF diversity. For instance legume
mannii (Redecker 2002, Schwarzott et al 2001). roots and nodules appear to contain different AM
(x) Paraglomus species form small, hyaline spores fungal taxa than other plants under the same
that do not show light microscopic characters conditions (Scheublin et al 2004).
that would distinguish them from those of Differences in hyphal architecture and growth
Glomus species. However rDNA phylogenies patterns have been observed between Glomeraceae
clearly showed that Paraglomus is not related to and Gigasporaceae. Gigasporaceae can colonize only
other Glomus species and is basal to the phylum roots from germinating spores and do not form
(Redecker et al 2000b). Paraglomus and Archae- anastomoses (cross-links) among hyphae. If their
ospora share the characteristic that their intra- hyphae are injured the main hypha is repaired.
radical structures consistently stain weakly or not Glomeraceae are able to colonize roots from myceli-

FIG. 2. Phylogenetic trees of Glomeromycota and other fungi. A. Bayesian tree based on 1539 bp of small subunit rDNA
sequences. Neurospora crassa was used as outgroup. B. Neighbor joining tree derived from maximum likelihood distances,
based on 1419 bp of small subunit rDNA sequences. The glomeromycotan clade in the tree was replaced by tree A, which had
an extended taxon sampling within the Glomeromycota, minus outgroup. Asterisk indicates clades that were not recovered in
Bayesian analysis. In particular the Mucorales grouped with Asco/Basidiomycota in the Bayesian tree. In both A and B the
trees had an identical overall topology with trees obtained by maximum likelihood analysis. C. Bayesian tree obtained from
protein sequences of rpb1. Numbers on the nodes indicate Bayesian posterior probabilities/bootstrap values from 1000
replicates of neighbor joining (A, B) and Bayesian posterior probabilities/bootstrap values of 100 replicates of parsimony
bootstrap with a PAM250-derived step matrix (C). Boldface ‘‘100’’ indicates both values were 100%.
890 MYCOLOGIA

um fragments or colonized root pieces; they form exclusion of the first 10% of trees from the first stage of
extensive anastomoses in their mycelium and they the run (burn-in). A maximum likelihood heuristic search
repair injured hyphae by forming a network of was performed in PAUP*4.0b10 (Swofford 2001), as well
anastomoses instead of repairing the main hyphal as distance analysis by neighbor joining and 1000 rep-
axis (de la Providencia et al 2005). licates of bootstrap analysis. Distances for neighbor joining
were obtained by the Kimura 3-parameter method. The
The first molecular phylogeny of AM fungi was
proportion of invariable characters and the gamma shape
reported by Simon et al (1993), using ribosomal small parameter for this analysis were determined with Modeltest.
subunit (SSU) sequences. These authors addressed Dataset B comprised 1419 bp from representatives of all
the phylogenetic relationships among the three fungal phyla as well as Baeria nivea and Ichthyophonus
families known at that time and attempted to date irregularis as outgroups. This dataset was analyzed to
their divergence by a molecular clock analysis. elucidate the phylogenetic position of the Glomeromycota
Morphological characters and fatty acid methyl ester among other fungal clades. It was analyzed by the same
profiles were evaluated phylogenetically by Benti- methods as described above. A total of 4 859 000 genera-
venga and Morton (1996). Most of the later phyloge- tions of Bayesian analysis were performed, trees were
netic studies used SSU sequences, which are still the sampled every 500 generations and 10% were discarded as
only gene with a broad taxon sample in the group. burn-in.
Marker genes other than ribosomal RNA became A protein dataset of 444 amino acids was compiled by
translating nucleotide sequences into peptide sequences.
available in the past few years, but their impact on
Apparent stop codons, most likely the result of sequencing
classification has been limited because some of the
errors, were treated as missing data. In a few instances
results were difficult to interpret. For instance a possible frame-shift in adjacent amino acids was caused,
phylogenies based on alpha and beta tubulin gene thus these amino acids were replaced by X. The amino acid
sequences were obscured by multiple paralogues dataset was analyzed by these methods: (i) parsimony
(Corradi et al 2004). A multigene phylogeny of fungi analysis in PAUP* with 1000 replicates of bootstrapping,
including basal lineages used nuclear ribosomal using a step matrix based on the PAM250 similarity matrix
small, large subunit and 5.8R subunit, tef, rpb1 and adapted for PAUP by R.K. Kuzoff. (ii) Bayesian analysis over
rpb2 sequences and placed the Glomeromycota as 1 000 000 generations using a mixed model. Trees were
a sister group of Asco- and Basidiomycota ( James et al sampled every 500 generations and a 50% consensus was
2006). calculated after discarding the first 10% of trees.
The aim of this article is to provide an overview of
previous studies on glomeromycotan phylogeny and RESULTS AND DISCUSSION
discuss newly available data in the context of the
AFTOL project. The phylogenetic relationship of the Glomeromycota to
other fungal phyla.—The Glomeromycota is sup-
ported consistently as a monophyletic group in
MATERIALS AND METHODS
phylogenetic analyses of rDNA and protein genes
Ribosomal SSU sequences used here were obtained from (FIG. 2B, Berbee and Taylor 2000, Schüßler et al 2001,
the public databases. rpb1 sequences were obtained by the Helgason et al 2003, James et al 2006). Ribosomal
procedure outlined (supplementary material) (database RNA analyses place these species as a sister group of
accession numbers AM284973-AM284982) or originate Asco- and Basidiomycota, although not always with
from the AFTOL database (http://ocid.nacse.org/
strong support. This ‘‘symbiomycota’’ clade (Tehler
research/aftol/data.php). The sequences were aligned in
PAUP*4.0 b10 (Swofford 2001). Alignments and trees were et al 2003) would be characterized by the ability to
submitted to TreeBASE under the study accession number form mutualistic symbioses with plants or algae that is
S1614 and the matrix accession numbers M2898 and not normally encountered in fungi outside this clade
M2899. (an exception is ectomycorrhiza-forming species of
Two datasets of ribosomal SSU sequences were used. the zygomycete Endogone, Warcup 1990). A large
Parameters for maximum likelihood analysis and Bayesian proportion of land plant-associated species generally
analysis were estimated with MrModeltest (Nylander 2004) are found in this clade. Most fungal groups in the
or Modeltest 3.5 (Posada 2004). The GTR+I+G model was ‘‘symbiomycota’’ also frequently form hyphal anasto-
determined to be appropriate for both nucleotide datasets.
moses, a trait that is rare or absent in most lineages of
After exclusion of regions of ambiguous alignment
Zygo- and Chytridiomycota. However it also is obvious
dataset A comprised 1539 positions of a representative set
of glomeromycotan taxa and Neurospora crassa as outgroup.
that some lineages of Zygomycota show strong
Bayesian analysis was performed with four chains over evidence of long-branch attraction in rRNA phyloge-
800 000 generations in MrBayes 3.1.1 (Ronquist and nies (FIG. 1B), which may exaggerate their separation
Huelsenbeck 2003). Trees were sampled every 200 genera- from Mortierellales, Endogonales, chytrids and possi-
tions and a 50% consensus tree was constructed after the bly Glomeromycota.
 REDECKER AND RAAB: PHYLOGENY OF GLOMEROMYCOTA 891

Evidence for a concurrent origin of this clade with major groups within the Glomeromycota is depicted
early land plants 500 000 000–400 000 000 y ago was (FIG. 2A). This gene and the adjacent highly vari-
presented, based on the phylogeny and fossil findings able internal transcribed spacers still offer the broad-
(Redecker et al 2000a), whereas other studies est range of taxa, including many unidentified
suggested that the Glomeromycota lineage could be species from environmental studies. The database of
older than land plants (1 000 000 000–1 200 000 000 y ribosomal large subunit sequences obtained from
ago) and might have associated with algae (Heckman spores has grown considerably in the past few years
et al 2001). However a radiation of major glomer- but has been used less frequently in field studies.
omycotan lineages long before land plants seems The intraspecies variability of nuclear ribosomal
unlikely considering that all major lineages in the genes complicates distinguishing closely related spe-
phylum comprise AM symbionts. An independent cies but is not an issue at the generic or family level.
adoption of the mycorrhizal way of living by all of One study reporting the occurrence of strongly
these lineages does not seem very parsimonious. diverging sequences within single spores (Hijri et al
Ribosomal SSU phylogenies (FIG. 2A) also indicate 1999) has been shown to be due to contamination by
that Geosiphon pyriformis, the only known glomero- Ascomycota (Redecker et al 1999). Other reports
mycotan fungus living in symbiosis with cyanobac- suggesting a sequence continuum among morpho-
teria, is derived from early mycorrhizal lineages. In logically well separated species (Clapp et al 2001)
SSU phylogenetic trees Geosiphon is closely related have been criticized and the results might be due to
to Archaeospora (FIG. 2A), and in some phylogenies problems with the culturing of the fungi (Schüßler et
it groups even between the two major clades of this al 2003).
genus (Schwarzott et al 2001). In the rpb1 phylogeny Protein genes are available for relatively few taxa,
Geosiphon branches off earlier than Paraglomus but key taxa are still missing from all datasets. In
(FIG. 2C). To conclusively address this relationship addition analyses of tubulin genes showed strong
with rpb1, data are needed from Archaeospora, the problems with paralogous gene copies found in the
AM fungal lineage that is most closely related to Glomeromycota (Corradi et al 2004). Therefore it still
Geosiphon. Up to now no protein gene sequences is not feasible to address comprehensively the
have been reported from Archaeospora. In addition intraphylum phylogeny using protein genes and only
there is still some possibility that Geosiphon forms the SSU phylogeny will be discussed here.
AM in addition to its unique cyanobacterial symbiosis The genus Glomus as it was defined before any
and that its mycorrhizae have not been detected yet. molecular data were available (e.g. Morton and Benny
It is interesting to note on the other hand that the 1990) later was shown to be polyphyletic (Redecker et
mycorrhizal status of a large proportion of described al 2000b, Schwarzott et al 2001, Walker et al 2004).
glomeromycotan morphospecies in fact has not been The polyphyly was resolved by separating several
confirmed because these fungi have never been lineages into the new genera Archaeospora, Paraglo-
obtained in single-species culture. If saprophytes mus, Diversispora and Pacispora (Morton and Redec-
occur in the Glomeromycota it can be assumed that ker 2001, Walker et al 2004, Walker and Schüßler
they are derived from mycorrhizal ancestors. 2004), but only one species of Glomus group C (G.
The rpb1 phylogeny presents strong support for spurcum) has been transferred to the new genus
a monophyletic Glomeromycota but not for the Diversispora.
symbiomycotan clade. This finding is in agreement The descriptions of both Paraglomus and Diversis-
with analyses of elongation factor and actin genes. pora have presented problems because their separa-
The three protein genes indicate a relationship of the tion from Glomus is well supported in rDNA
Glomeromycota with zygomycotan lineages, in partic- phylogenies (and biochemical characters in Paraglo-
ular the Mortierellales (FIG. 2C, Helgason et al 2003). mus) but morphological characters to distinguish
In contrast, based on tubulin sequences, the Glomer- them are scarce. The mycorrhizae of Paraglomus and
omycota group with the chytrids. The tubulin Archaeospora share the consistently weak staining
phylogenies however show a strong separation be- behavior with acidic dyes, whereas intraradical struc-
tween fast-evolving (Basidio-, Ascomycota, Ento- tures of Glomus group C/Diversispora occasionally
mophthorales, Mucorales and others) and slowly stain faintly. All Paraglomus species currently are
evolving lineages (Glomero-, Chytridiomycota), which known produce small hyaline spores, whereas Glomus
can be expected to cause problems in phylogenetic group C members show a variety of spore morphol-
reconstruction. ogies, even including an isolate fitting the morpho-
logical description of Glomus etunicatum, a species
Phylogenetic relationships within the Glomeromycota.— that normally is thought to belong to Glomus group B
A SSU-based phylogeny of representatives of the (Schwarzott et al 2001). Walker and Schüßler (2004)
892 MYCOLOGIA

transferred Glomus spurcum to the new genus assessed yet with this gene. A combined Bayesian
Diversispora (grammatically incorrectly as ‘‘Diversis- analysis of elongation factor and actin supported the
pora spurcum’’), using an inner flexible wall as alternative grouping of Acaulospora laevis and Glomus
a distinguishing character. The presence of this caledonium but contained several other inconsisten-
feature in G. spurcum however is controversial cies (Helgason et al 2003). The value of elongation
(Kennedy et al 1999) and it might not be present in factor sequences to elucidate deep phylogenies re-
other members of the clade. cently was questioned because of saturation among
There is some evidence that one isolate of Acaulos- deep fungal lineages (Tanabe et al 2004). Alpha and
pora (W3424) groups basally to Entrophospora colombi- beta tubulin sequences yielded conflicting results on
ana in the monophyletic Acaulosporaceae (Walker this question (Corradi et al 2004) possibly due to
et al 2004), but this is not supported in all analyses problems with paralogues.
(FIG. 2 A). Therefore the monophyly of Acaulospora The presence of flexible germinal walls staining
as currently defined is not strongly rejected. with Melzer’s reagent is a possible synapomorphy
It has been disputed whether Scutellospora or uniting Gigasporaceae, Pacisporaceae and Acaulos-
Gigaspora are derived genera within the family Giga- poraceae in the Diversisporales. Within the Diversis-
sporaceae. Gigaspora spores have the same wall porales clade the germination shield in Scutellospora,
structure as early development stages of Scutellospora, the papillate germination wall layer in Gigaspora and
suggesting a classical Haeckelian recapitulation of the germination orb in the Acaulosporaceae and
phylogeny during ontogeny (Bentivenga et al 1997, Pacispora could be homologous structures. The
Morton 1995). In molecular SSU phylogenies Giga- discovery of Pacispora, a genus that shows glomoid
spora appears to be a derived monophyletic clade spore formation but has germinal walls and a germi-
within the Gigasporaceae. The genus is nested between nation orb, further supports this hypothesis as
Scutellospora species, rendering the latter paraphyletic. a possible morphological intermediate. Consequently
Among Gigaspora species the ribosomal RNA sequence it appears as an early divergent lineage in the
variation is low. However phylogenetic analyses do not Diversisporales clade (FIG. 2A). In the phylogeny
offer strong support for a paraphyletic genus Scutello- presented by Walker et al (2004) Pacispora groups
spora (FIG. 2A) even when more species of Gigaspora together with the Gigasporaceae, whereas it is closer
are included (Schwarzott et al 2001). A derived position to Acaulosporaceae/Glomus group C in FIG. 2A.
of Gigaspora would imply the loss of the complex The Gigasporaceae shows some unique character-
flexible inner walls and the germination shield during istics in their growth and symbiotic behavior (see
the evolution of the genus and a reduction in introduction) and a characteristic mode of spore
complexity of the wall structure. The papillate wall formation. These features were suggested to set apart
layer appearing immediately before germination then this group from all other glomeromycotan fungi, but
could be a relic of the germination shield. they are equally compatible with phylogenies support-
Morton and Benny (1990) divided the order ing this family as a derived clade of the Diversisporales.
Glomales into the suborders Glomineae (comprising In this context Glomus group C either should be
Sclerocystis, Glomus, Entrophospora and Acaulospora) expected to be a basal member of the Diversisporales
and Gigasporineae (Gigaspora and Scutellospora). The arising before the development of germinal walls or
Glomineae was characterized by the formation of a secondary loss of germinal walls might have
vesicles, whereas members of Gigasporineae form occurred in this group. Most SSU analyses place
auxiliary cells. This concept could not be substantiat- Glomus group C close to the Acaulosporaceae, which
ed by SSU molecular phylogenies. Acaulospora would be most consistent with the second alternative.
appendicula (now Archaeospora leptoticha), which was Overall the evolutionary trends within the Glomer-
thought to be a transitional species between Acaulo- omycota suggest that the glomoid type of spore
spora and Glomus (Morton and Benny, 1990) later was formation is a symplesiomorphy. This is supported
shown to be in a distinct basal clade (Redecker et al by the earliest known fossils of Glomeromycota, which
2000b). Simon et al (1993) presented phylogenetic are glomoid spores (Redecker et al 2002). The
data supporting an alternative concept uniting acaulosporoid and entrophosporoid types of spore
Acaulosporaceae and Gigasporaceae. formation also arose early in organisms that most
SSU phylogenies consistently support the Diversis- likely were dimorphic.
porales clade, comprising Acaulosporaceae, Gigaspor-
aceae, Diversisporaceae and Pacisporaceae (FIG. 2A;
CONCLUSION
Schüßler et al 2001, Walker and Schüßler 2004).
Because no rpb1 sequences of Acaulospora, Pacispora Although several lines of currently available evidence
or Glomus group C are available the clade cannot be support the ‘‘symbiomycota’’ concept of a relation-
 REDECKER AND RAAB: PHYLOGENY OF GLOMEROMYCOTA 893

ship of Asco-, Basido- and Glomeromycota the a fungus forming endocytobiosis with Nostoc (Cyano-
conflicting results from different genes should incite bacteria) is an ancestral member of the glomales:
further analysis. Analyses using different methods of evidence by SSU rRNA analysis. J Mol Evol 43:71–81.
phylogenetic analysis and a broad spectrum of genes Gerdemann JW, Trappe JM. 1974. Endogonaceae in the
Pacific Northwest. Mycol Mem 5:1–76.
clearly are needed. The same is true for the
Graham JH, Hodge NC, Morton JB. 1995. Fatty acid methyl
intraphylum relationships, where evolutionary ten-
ester profiles for characterization of glomalean fungi
dencies among the clades can be resolved only with and their endomycorrhizae. Appl Environ Microbiol
a more complete sampling of key taxa with multiple 61:58–64.
genes. Heckman DS, Geiser DM, Eidell BR, Stauffer RL, Kardos
NL, Hedges SB. 2001. Molecular evidence for the early
colonization of land by fungi and plants. Science 293:
ACKNOWLEDGMENTS 1129–1133.
The authors thank Tim James and Arthur Schüßler for Helgason T, Daniell TJ, Husband R, Fitter AH, Young JPW.
providing unpublished sequences from the AFTOL project, 1998. Ploughing up the wood-wide web? Nature 394:
Fritz Oehl and Kerstin Wex for providing images, the 431.
Botanical Institute at the University of Basel for generous ———, Watson IJ, Young JPW. 2003. Phylogeny of the
support, Zuzana Sykorova for critically reading the manu- Glomerales and Diversisporales (Fungi: Glomeromy-
script and the Swiss National Science Foundation for cota) from actin and elongation factor 1-alpha
financial support through grants 3100 066688.01 and 3100 sequences. FEMS Microbiol Lett 229:127–132.
A0109466.1 to D.R. We also acknowledge support from NSF Hijri I, Sykorova Z, Oehl F, Ineichen K, Mäder P, Wiemken
0090301, Research Coordination Network: a phylogeny for A, Redecker D. 2006. Communities of arbuscular
kingdom Fungi to M. Blackwell, J.W. Spatafora and J.W. mycorrhizal fungi in arable soils are not necessarily
Taylor. low in diversity. Mol Ecol 15:2277–2289.
Hijri M, Hosny M, van Tuinen D, Dulieu H. 1999.
Intraspecific ITS polymorphism in Scutellospora casta-
LITERATURE CITED nea (Glomales, Zygomycota) is structured within
multinucleate spores. Fung Genet Biol 26:141–151.
Almeida RT, Schenck NC. 1990. A revision of the genus ———, Redecker D, Macdonald-Comber Petetot JA, Voigt
Sclerocystis (Glomaceae, Glomales). Mycologia 82:703– K, Wöstemeyer J, Sanders IR. 2002. Identification and
714. isolation of two Ascomycete fungi from spores of the
Bécard G, Pfeffer PE. 1993. Status of nuclear division in arbuscular mycorrhizal fungus Scutellospora. Appl
arbuscular mycorrhizal fungi during in-vitro develop- Environ Microbiol 68:4567–4573.
ment. Protoplasma 174:62–68. ———, Sanders IR. 2005. Low gene copy number shows
Bentivenga SP, Bever JD, Morton JB. 1997. Genetic variation that arbuscular mycorrhizal fungi inherit genetically
of morphological characters within a single isolate of different nuclei. Nature 433:160–163.
the endomycorrhizal fungus Glomus clarum (Gloma- Husband R, Herre EA, Turner SL, Gallery R, Young
ceae). Amer J Bot 84:1211–1216. JPW. 2002. Molecular diversity of arbuscular mycorrhi-
———, Morton JB. 1996. Congruence of fatty acid methyl zal fungi and patterns of host association over time
ester profiles and morphological characters of arbus- and space in a tropical forest. Mol Ecol 11:2669–
cular mycorrhizal fungi in Gigasporaceae. Proc Natl 2678.
Acad Sci USA 93:5659–5662. James TY, Kauff F, Schoch C, Matheny PB, Hofstetter V, Cox
Berbee ML, Taylor JW. 2000. Fungal molecular evolution: C, Celio G, Gueidan C, Fraker E, Miadlikowska J,
gene trees and geologic time. In: McLaughlin DJ, Lumbsch HT, Rauhut A, Reeb V, Arnold AE, Amtoft A,
McLaughlin EG, Lemke PA, eds. The Mycota. New Stajich JE, Hosaka K, Sung G-H, Johnson D, O’Rourke
York: Springer Verlag. p 229–246. B, Crockett M, Binder M, Curtis JM, Slot JC, Wang Z,
Clapp JP, Rodriguez A, Dodd JC. 2001. Inter- and intra- Wilson AW, Schüßler A, Longcore JE, O’Donnell K,
isolate rRNA large subunit variation in Glomus Mozley-Standridge S, Porter D, Letcher PM, Powell MJ,
coronatum spores. New Phyto 149:539–554. Taylor JW, White MM, Griffith GW, Davies DR, Humber
Corradi N, Hijri M, Fumagalli L, Sanders IR. 2004. RA, Morton JB, Sugiyama J, Rossman AY, Rogers JD,
Arbuscular mycorrhizal fungi (Glomeromycota) har- Pfister DH, Hewitt D, Hansen K, Hambleton S,
bour ancient fungal tubulin genes that resemble those Shoemaker RA, Kohlmeyer J, Volkmann-Kohlmeyer B,
of the chytrids (Chytridiomycota). Fungal Genet Biol Spotts RA, Serdani M, Crous PW, Hughes KW,
41:1037–1045. Matsuura K, Langer E, Langer G, Untereiner WA,
de la Providencia IE, de Souza FA, Fernandez F, Delmas NS, Lücking R, Büdel B, Geiser DM, Aptroot A, Diederich
Declerck S. 2005. Arbuscular mycorrhizal fungi reveal P, Schmitt I, Schultz M, Yahr R, Hibbett D, Lutzoni F,
distinct patterns of anastomosis formation and hyphal McLaughlin D, Spatafora J, Vilgalys R. 2006. Recon-
healing mechanisms between different phylogenic structing the early evolution of the fungi using a six
groups. New Phytol 165:261–271. gene phylogeny. Nature 443:818–822.
Gehrig H, Schüßler A, Kluge M. 1996. Geosiphon pyriforme, Jansa J, Mozafar A, Frossard E. 2005. Phosphorus acquisition
894 MYCOLOGIA

strategies within arbuscular mycorrhizal fungal com- ———, Hijri M, Dulieu H, Sanders IR. 1999. Phylogenetic
munity of a single field site. Plant Soil 276:163–176. analysis of a dataset of fungal 5.8S rDNA sequences
Kennedy LJ, Stutz JC, Morton JB. 1999. Glomus eburneum shows that highly divergent copies of internal tran-
and G. luteum, two new species of arbuscular mycor- scribed spacers reported from Scutellospora castanea are
rhizal fungi, with emendation of G. spurcum. Mycologia of Ascomycete origin. Fung Genet Biol 28:238–244.
91:1083–1093. ———, Kodner R, Graham LE. 2000a. Glomalean fungi
Klironomos JN. 2000. Host-specificity and functional di- from the Ordovician. Science 289:1920–1921.
versity among arbuscular mycorrhizal fungi. In: Bell ———, ———, ———. 2002. Palaeoglomus grayi from the
CR, Brylinsky M, Johnson-Green P, eds. Microbial Ordovician. Mycotaxon 84:33–37.
biosystems: new frontiers, 8th International Symposium ———, Morton JB, Bruns TD. 2000b. Ancestral lineages of
on Microbial Ecology. Halifax, Canada: Atlantic C Soc arbuscular mycorrhizal fungi (Glomales). Mol Phylo-
Microb Ecol. p 845–851. gen Evol 14:276–284.
Kuhn G, Hijri M, Sanders IR. 2001. Evidence for the ———, ———, ———. 2000c. Molecular phylogeny of the
evolution of multiple genomes in arbuscular mycorrhi- arbuscular mycorrhizal fungi Glomus sinuosum and
zal fungi. Nature 414:745–748. Sclerocystis coremioides. Mycologia 92:282–285.
Lanfranco L, Delpero M, Bonfante P. 1999. Intrasporal Rodriguez A, Dougall T, Dodd JC, Clapp JC. 2001. The large
variability of ribosomal sequences in the endomycor- subunit ribosomal RNA genes of Entrophospora infre-
rhizal fungus Gigaspora margarita. Mol Ecol 8:37–45. quens comprise sequences related to two different
Merryweather J, Fitter A. 1998. The arbuscular mycorrhizal glomalean families. New Phytol 152:159–167.
fungi of Hyacinthoides non-scripta: I. Diversity of fungal Rokas A, Krüger D, Carroll SB. 2005. Animal evolution and
taxa. New Phytol 138:117–129. the molecular signature of radiations compressed in
Morton JB. 1988. Taxonomy of VA mycorrhizal fungi: time. Science 310:1933–1938.
classification, nomenclature and identification. Myco- Ronquist F, Huelsenbeck JP. 2003. MrBayes 3: Bayesian
taxon 32:267–324. phylogenetic inference under mixed models. Bioinfor-
———. 1995. Taxonomic and phylogenetic divergence matics 19:1572–1574.
among five Scutellospora species based on comparative Sanders IR, Alt M, Groppe K, Boller T, Wiemken A. 1995.
developmental sequences. Mycologia 87:127–137. Identification of ribosomal DNA polymorphisms
———, Benny GL. 1990. Revised classification of arbuscular among and within spores of the Glomales: application
mycorrhizal fungi (Zygomycetes): a new order Glo- to studies on the genetic diversity of arbuscular
males and Gigasporineae and two new families Acau- mycorrhizal fungal communities. New Phytol 130:419–
losporaceae and Gigasporaceae with an emendation of 427.
Glomaceae. Mycotaxon 37:471–491. Sawaki H, Sugawara K, Saito M. 1998. Phylogenetic position
———, Redecker D. 2001. Two new families of Glomales, of an arbuscular mycorrhizal fungus, Acaulospora
Archaeosporaceae and Paraglomaceae, with two new gerdemannii, and its synanamorph, Glomus leptotichum,
genera Archaeospora and Paraglomus, based on concor- based upon 18s rRNA gene sequence. Mycoscience 39:
dant molecular and morphological characters. Mycolo- 477–480.
gia 93:181–195. Scheublin TR, Ridgway KP, Young JPW, van der Heijden
Nylander JAA. 2004. MrModeltest 2.0 Evolutionary Biology MGA. 2004. Nonlegumes, legumes and root nodules
Centre, Uppsala University: Uppsala, Sweden. harbor different arbuscular mycorrhizal fungal com-
Oehl F, Sieverding E. 2004. Pacispora, a new vesicular munities. Appl Environ Microbiol 70:6240–6246.
arbuscular mycorrhizal fungal genus in the Glomer- Schüßler A, Mollenhauer D, Schnepf E, Kluge M. 1994.
omycetes. J App Bot 78:72–82. Geosiphon pyriforme, an endosymbiotic association of
Öpik M, Moora M, Liira J, Koljalg U, Zobel M, Sen R. 2003. fungus and cyanobacteria—the spore structure resem-
Divergent arbuscular mycorrhizal fungal communities bles that of arbuscular mycorrhizal (AM) fungi. Bot
colonize roots of Pulsatilla spp. in boreal Scots pine Acta 107:36–45.
forest and grassland soils. New Phytol 160:581–593. ———, Schwarzott D, Walker C. 2003. Glomeromycota
Pawlowska T, Taylor JW. 2004. Organization of genetic rRNA genes—the diversity of myths? Mycorrhiza 13:
variation in individuals of arbuscular mycorrhizal fungi. 233–236.
Nature 427:733–737. ———, ———, ———. 2001. A new fungal phylum, the
Pirozynski KA, Dalpé Y. 1989. Geological history of the Glomeromycota: phylogeny and evolution. Mycol Res
Glomaceae, with particular reference to mycorrhizal 105:1413–1421.
symbiosis. Symbiosis 7:1–36. Schwarzott D, Walker C, Schüßler A. 2001. Glomus, the
Posada D. 2004. Modeltest 3.5 Facultad de Biologia, largest genus of the arbuscular mycorrhizal fungi
Universidad de Vigo: Vigo, Spain. (Glomales), is non-monophyletic. Mol Phylogen Evol
Raab P, Brennwald A, Redecker D. 2005. Mitochondrial 21:190–197.
large ribosomal subunit sequences are homogeneous Simon L, Bousquet J, Levesque RC, Lalonde M. 1993.
within isolates of Glomus (arbuscular mycorrhizal fungi, Origin and diversification of endomycorrhizal fungi
Glomeromycota). Mycol Res 109:1315–1322. and coincidence with vascular land plants. Nature 363:
Redecker D. 2002. Molecular identification and phylogeny 67–69.
of arbuscular mycorrhizal fungi. Plant Soil 244:67–73. Spain JL. 2003. Emendation of Archaeospora and of its
 REDECKER AND RAAB: PHYLOGENY OF GLOMEROMYCOTA 895

type species, Archaeospora trappei. Mycotaxon 87:109– netic tree from 1551 ribosomal sequences of chitinous
112. fungi. Mycol Res 107:901–916.
———, Sieverding E, Schenck NC. 1989. Gigaspora Tommerup IC, Sivasithamparam K. 1990. Zygospores and
ramisporophora: a new species with novel sporophores asexual spores of Gigaspora decipiens, an arbuscular
from Brazil. Mycotaxon 34:667–677. mycorrhizal fungus. Mycol Res 94:897–900.
Stukenbrock EH, Rosendahl S. 2005. Development Vandenkoornhuyse P, Husband R, Daniell TJ, Watson IJ,
and amplification of multiple co-dominant genetic Duck JM, Fitter AH, Young JPW. 2002. Arbuscular
markers from single spores of arbuscular mycorrhizal mycorrhizal community composition associated with
fungi by nested multiplex PCR. Fung Genet Biol 42:73– two plant species in a grassland ecosystem. Mol Ecol 11:
80. 1555–1564.
Stürmer SL, Morton JB. 1997. Developmental patterns Walker C. 1983. Taxonomic concepts in the Endogonaceae:
defining morphological characters in spores of four spore wall characteristics in species descriptions.
species in Glomus. Mycologia 89:72–81. Mycotaxon 18:443–455.
———, ———. 1999. Taxonomic reinterpretation of ———. 1992. Systematics and taxonomy of the arbuscular
morphological characters in Acaulosporaceae based mycorrhizal fungi (Glomales)—a possible way forward.
on developmental patterns. Mycologia 91:849–857. Agronomie 12:887–897.
Swofford DL. 2001. PAUP*: phylogenetic analysis using ———, Blaszkowski JDS, Schüßler A. 2004. Gerdemannia
parsimony (*and other methods) Sinauer Associates: gen. nov., a genus separated from Glomus, and
Sunderland, Massachusetts. Gerdemanniaceae fam. nov., a new family in the
Tanabe Y, Saikawa M, Watanabe MM, Sugiyama J. 2004. Glomeromycota. Mycol Res 108:707–718.
Molecular phylogeny of Zygomycota based on EF-1 ———, Schüßler A. 2004. Nomenclatural clarifications and
alpha and RPB1 sequences: limitations and utility of new taxa in the Glomeromycota. Mycol Res 108:981–982.
alternative markers to rDNA. Mol Phylogen Evol 30: Warcup JH. 1990. Taxonomy, culture and mycorrhizal
438–449. associations of some zygosporic Endogonaceae. Mycol
Tehler A, Little D, Farris J. 2003. The full-length phyloge- Res 94:173–178.