Glomus, the Largest Genus of the Arbuscular Mycorrhizal Fungi (Glomales), Is Nonmonophyletic

Molecular Phylogenetics and Evolution Vol. 21, No. 2, November, pp. 190 –197, 2001 doi:10.1006/mpev.2001.1007, available online at http://www.idealibrary.com on

Glomus, the Largest Genus of the Arbuscular Mycorrhizal Fungi (Glomales), Is Nonmonophyletic Daniel Schwarzott,* ,1 Chris Walker,† and Arthur Schu¨ßler* *Institute of Botany, Technische Universita¨t Darmstadt, Schnittspahnstrasse 10, D-64287 Darmstadt, Germany; and †School of Conservation Sciences, Bournemouth University, Talbot Campus, Fern Barrow, Poole, Dorset BH12 5BB, United Kingdom Received February 12, 2001; revised May 7, 2001

Arbuscular mycorrhizal (AM) fungi form a widespread and ecologically important symbiosis with plants in the land ecosystem. The phylogeny of the largest presently accepted genus, Glomus, of the monogeneric family Glomaceae (Glomales; AM fungi) was analyzed. Phylogenetic trees were computed from nearly full-length SSU rRNA gene sequences of 30 isolates, and show that “Glomus” is not monophyletic. Even after the very recent separation of Archaeospora and Paraglomus from “Glomus,” the genus further separates into two suprageneric clades. One of them diverges further into two subclades, differing by phylogenetic distances equivalent to family level. The other, comprising Glomus versiforme, G. spurcum, and a species morphologically similar to G. etunicatum, is not closely related to the Glomaceae, but clusters together with the Acaulosporaceae and Gigasporaceae in a monophyletic clade. Based on the molecular evidence, a new family, separate from the Glomaceae, is required to accommodate this group of organisms, initially named Diversisporaceae fam. ined. The current taxonomic concept of the recently erected family Archaeosporaceae also requires future emendation, because Geosiphon pyriformis (Geosiphonaceae) renders Archaeospora, the sole genus formally included in this family, paraphyletic. The suborders Gigasporineae and Glominaeae are not congruent with the natural phylogeny of the AM fungi. Our data necessitate a general reexamination of the generic concepts within the Glomales. In addition to the new family structure hypothesized herein, establishment of at least three new genera will be necessary in the future. © 2001 Academic Press Key Words: arbuscular mycorrhiza; Glomales; Glomus; phylogeny; SSU rRNA; taxonomy.

INTRODUCTION Glomalean fungi are considered to have originated more than 450 million years ago (Redecker et al., 2000a) and arbuscular mycorrhiza-like symbioses with the early 1

To whom correspondence should be addressed. Fax: 0049-6151166049. E-mail: [email protected]. 1055-7903/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.

land plants are postulated from the discovery of structures morphologically identical to arbuscules in 400-Myrold Aglaophyton fossils (Remy et al., 1994; Taylor et al., 1995). The majority of extant vascular plant species form mycorrhizas with this important fungal group. It thus seems evident that plants and arbuscular mycorrhizal (AM) fungi coevolved at least since the earliest colonization of terrestrial habitats by plants. The presently accepted phylogenetic classification of the order Glomales, which contains the AM fungi, was based on a cladistic analysis (Morton and Benny, 1990) of morphological characteristics (Gerdemann and Trappe, 1974; Walker, 1983). The very early origin of the AM fungi was not taken into account in that classification. Recently, the taxonomic concept was modified based on molecular evidence of the SSU rRNA gene (Morton and Redecker, 2001). Two new families (Archaeosporaceae, Morton and Redecker and Paraglomaceae Morton & Redecker) were erected to contain ancestral lineages (Redecker et al., 2000b) within the order Glomales. Both families include fungi with Glomus-like spore morphology and are monogeneric, with three species in the former (Archaeospora trappei (Ames and Linderman) Morton and Redecker, Ar. leptoticha (Schenck and Smith) Morton and Redecker and Ar. gerdemannii (Schenck and Smith) Morton and Redecker) and two in the latter (Paraglomus occultum (Walker) Morton and Redecker and P. brasilianum (Spain and Miranda) Morton and Redecker). We present a more detailed molecular analysis of the “genus” Glomus sensu Morton and Benny and a discussion of the taxonomy and phylogeny of the Glomales. Our analyses of the SSU rRNA gene show that the present taxonomic concept does not reflect the natural relationships within the AM fungi. MATERIALS AND METHODS Origin of Fungi Spores of AM fungi were obtained from the culture collections of the BEG (La Banque Europe´enne des Glomales, INRA, Dijon Cedex, France; http://www.

190

191

Glomus (GLOMALES) IS NONMONOPHYLETIC

ukc.ac.uk/beg/), INVAM (International Collection of Arbuscular and Vesicular–Arbuscular Mycorrhizal Fungi, Morgantown, USA; http://invam.caf.wvu.edu/), and the personal culture collection of one of the present authors (C. Walker). Species and cultures (Table 1) were chosen as far as possible to provide a wide representation of the different developmental and morphological types (Walker, 1992). These included several cultures derived from the type culture of their particular species (EX-TYPE material, Greuter et al., 2000). Specimens from INVAM cultures were received as spores in a sandy substrate from open-pot cultures (Morton et al., 1993). Those from the culture collection of C. Walker were from sealed systems, precluding cross-contamination by glomalean fungi (Walker and Vestberg, 1994). Except for Glomus luteum Kennedy, Stutz and Morton and G. proliferum Dalpe´ & Declerck species names and authorities of AM fungi are after Walker and Trappe (1993). Voucher specimens (Walker, 1979) were kept of all fungi. They can be obtained from Walker at the address above. Additional information about the cultures used in this study is available at http://amf-phylogeny.com and will be updated if further details become available. Isolation of Fungal Spores and DNA Extraction, Cloning, and Sequencing Samples for DNA extraction were removed by centrifugation and sugar floatation (Walker et al., 1982) or by swirling with water and decanting through a fine sieve. Individual spores were selected under a dissecting microscope and placed in an Eppendorf vial containing distilled water. Each single spore was carefully checked to ensure it appeared fresh and undamaged, and to verify that there was no obvious contamination from other fungi. Normally, 10 spores were individually processed in this way for each fungus and stored at 4°C pending processing. For detailed procedure on single-spore cleaning, DNA isolation, and nested PCR conditions and primers, cloning, and sequencing of the near full-length SSU rRNA gene, see the methods paper of Schwarzott and Schu¨ßler (2001). Phylogenetic Analysis Thirty-five new sequences from cloned near-fulllength SSU rRNA genes from 30 glomalean isolates were analyzed. Several shorter fragments were sequenced for proofing purposes and to construct nearfull-length consensus sequences. In total, 25 Glomus cultures belonging to 21 described and four undescribed Glomus species were analyzed. Additionally, both presently described Paraglomus, two Archaeospora species, and Geosiphon pyriformis were sequenced again to obtain more full-length sequences representative for these “ancestral” lineages. Published sequences (see Table 1) were taken from the EMBL database (http://srs.ebi.ac.uk/).

Sequence alignment was performed manually with the program ALIGN (freeware: http://domix␪.tripod. com), taking the secondary structure into account (Van de Peer et al., 1998). All phylogenetic trees were computed with PHYLIP 3.573 (Felsenstein, 1989). Input order of species was randomized and analyses were bootstrapped to estimate robustness of tree structure. Phylogenetic trees were computed by the neighborjoining method (Saitou and Nei, 1987; Nei et al., 1995) with Kimura parameters (Kimura, 1980) and the parsimony method (Felsenstein, 1983). For a comparison of computation methods, see Kumar and Gadkar (2000). Trees were viewed with the program TreeView (freeware: http://taxonomy.zoology.gla.ac.uk/rod/treeview. html), saved in vector format, and edited with Micrographx Windows Draw 5.0. Two data sets were analyzed, one of sequences from all fungal higher taxa (Fig. 1) and the other comprising only glomalean sequences, with Mortierella and Endogone as outgroups (Fig. 2). The former allowed the analysis of 1597 and the latter of 1680 aligned sites. The alignment is deposited at the EMBL database (http://www3.ebi.ac.uk/Services/align/listali.html) under accession number ALIGN_000124. Links to the software used and alignments with secondary structure information are available at http://www.geosiphon.de. RESULTS Phylogeny of the Glomales In the phylogenetic analyses, the Glomales—including Geosiphon pyriformis (see Schu¨ßler and Kluge, 2001)—represent a monophyletic clade (Fig. 1, see also Gehrig et al., 1996; Redecker et al., 2000b; Schu¨ßler, 1999; Tanabe et al., 2000). In this sense, the Glomales, including the ancient lineages, are supported to be monophyletic by bootstrap values of 84% (neighbor joining) and 80% (parsimony) in the analyses comprising other fungal taxa (Fig. 1). If the Paraglomaceae are excluded from the analyses, the value rises to 93%; if the Archaeosporaceae are excluded, it drops to 65% (both neighbor-joining, not shown). The inclusion of 100 new Zygomycetes sequences (neighbor-joining analysis including the Paraglomaceae, not shown) leads to a rise of the bootstrap support for the monophyletic group of the Glomales to 94%. This confirms its monophyletic origin. Within the Glomales, the clade comprising the Glomaceae, Acaulosporaceae, and Gigasporaceae is separated from the more ancient lineages (Archaeosporaceae and Paraglomaceae) by bootstrap values of ⱖ99% in all analyses. The phylogenetic analyses resulted in trees showing highly supported clades at a level we propose here to represent families.

TABLE 1 Sequences Used in the Present Study Species

Isolate-code(s) and/or voucher-no./culture-no. a

Culturing origin(originator)

Supplier of sequenced culture (if known); notes

BEG15 W3294/Att263-15; BEG20 BEG31; W1843/Att79-3 BEG23 BEG14

M (Jakobsen) b N (Hayman) M (Vestberg) M (Gryndler) N (Rosendahl)

INRA; from Denmark Walker; “Rothamsted isolate”; from GB Walker; from Finland INRA; from Czech Republic INRA from Denmark

G. claroideum G. manihotis/clarum d G. coremioides

subculture of W2537/Att599-0 BR147B-8; W3163/Att72-1 “Biorize”

S (Vandenkoornhuyse) N (Ming Lin) N (Blal)

G. G. G. G. G.

coronatum etunicatum etunicatum fasciculatum fragilistratum

W3153/Att143-5; COG1 UT316 W2423/Att382-16 BEG53 W3238/Att112-6

S (McGee) N (Wood) M (Walker) S (Furlan) M (Jakobsen)

G. G. G. G. G.

geosporum intraradices lamellosum lamellosum luteum

BEG11; W992/Att191-1 DAOM197198 W3161/Att672-13 W3160/Att244-13 SA101-3; W3090/Att676-0

S (Dodd) N (Parent) S (Vestberg) S (Vestberg) N (Talukdar)

G. manihotis G. manihotis G. manihotis/clarum d G. microaggregatum G. mosseae G. mosseae G. mosseae G. mosseae G. mosseae G. mosseae G. mosseae G. proliferum G. spurcum G. sinuosum G. versiforme G. verruculosum G. vesiculiferum G. viscosum G. sp. morph1 G. sp. morph2 G. sp. “clustered” G. sp. G. sp. G. sp. A. rugosa A. spinosa E. colombiana E. “contigua” Ar. leptoticha Ar. leptoticha Ar. trappei Ar. trappei Ar. trappei P. brasilianum P. occultum P. occultum P. occultum Gi. albida Gi. decipiens

W3224/Att575-9 FL879-3; W3181/Att575-25 BR212 DAOM215235 BEG12 FL156B DAOM221475 DAOM212595 DAOM198394 BEG25 BEG69 DAOM226389/MUCL41827 W3239/Att246-4 MD126 BEG47 W3295/Att298-6 none BEG27; W3207/Att179-8 WUM3; W2940/Att15-5 WUM3; W2939/Att15-5 W3234/Att13-7 UY110.6.10; W3347/Att565-7 UY110.6.9; W3349/Att565-11 DAOM225952 WV949 WV860 FL356 WV201 NC176 MAFF520055 NB112 W3179/Att186-1 AU219; WUM19(1) W3086/Att260-4; BR105 IA702-3 HA771 CL700 FL927 BEG45

F (Howeler) F (Howeler) N (Sturmer) N (Dalpe´ and Mitrow) M (Mosse) N (Schenck) N N (Dalpe´ and Mitrow) N (Furlan) S (Dodd) S (Leyval) F (Rise`de) M (Pfeiffer) N (Chabot) M (Daniels) S (Blaszkowski) N (Chabot) ST (Giovannetti) M (Abbott) M (Abbott) SC (Walker) SC (Merryweather) SC (Merryweather) N (Vandenkoornhuyse) N (Dant) N N N (Morton) N S (Murakoshi and Siato) N (Klopatek) ST (Schweiger) N (Gazei) M (Spain) N (Klopfenstein) N (Koske) N N (Perez) M (Mercer)

Gi. gigantea Gi. margarita S. castanea S. cerradensis S. heterogama S. heterogama S. pellucida S. projecturata Ge. pyriformis

WV932 DAOM194757 BEG1 MAFF520056 BR154-5 WV858B WV873 W3254/Att697-0 W3619/GEO1

N N (Menge) M (Gianinazzi-Pearson) S (Saito) N (Ming Lin) N (Morton) N (Morton) ST (Kramadibrata) S (Mollenhauer)

From Germany; partial sequence Bioplanta, Inc. (Brazil) Societe´ Biorize (Dijon, France); from Ivory Coast Walker; from Australia INVAM Walker; from Scotland INRA; from Canada Walker; EX-HOLOTYPE; from Denmark Walker and INRA; from GB Piche´; from Canada Walker; from Iceland Walker; EX-HOLOTYPE; from Canada INVAM; formerly also “G. clarum NT4” Walker; EX-HOLOTYPE (CIAT-C-1-1) INVAM; EX-HOLOTYPE (CIAT C-1-1) INVAM (isolate lost); from Brazil From Canada; partial sequence INRA; Rothamsted “Yellow Vacuolate” INVAM; formerly WV156; from USA From Canada; partial sequence From Canada; partial sequence From Canada; partial sequence INVAM; partial sequence; from GB INVAM; partial sequence; from France EX-HOLOTYPE; from Guadeloupe Walker; EX-HOLOTYPE; from USA INVAM formerly S. sinuosa; from USA INRA and Torino; EX-HOLOTYPE Walker; EX-HOLOTYPE; from Poland Chabot, from Canada (Chabot) Walker; EX-HOLOTYPE Walker; from Australia Walker; from Australia Walker; from India Walker; from GB Walker; from GB From Germany; partial sequence INVAM, formerly WV935 INVAM INVAM, formerly WV877 INVAM, formerly WV796 INVAM From Japan INVAM; from Namibia Walker; from Australia INVAM; from Australia Walker; EX-HOLOTYPE; from Brazil INVAM; from USA INVAM; from Hawaii INVAM INVAM, formerly WV1034; from USA INRA; partial sequence; from Australia INVAM From California (originally) INRA; EX-HOLOTYPE; from France From Japan INVAM; from Brazil INVAM, formerly WV929; from USA INVAM Walker; EX-HOLOTYPE Schu¨ßler; from Germany

G. G. G. G. G.

caledonium caledonium claroideum claroideum claroideum

Accession-number(s) c Y17653; AJ301854 Y17635; AJ301853 Y17641; AJ276079 AJ276080; Y17642 AJ301851; AJ301852; AJ276075; Y17636 AF139732 AJ276084 AJ249715 AJ276086 Z14008; Y17639 Y17644; AJ276076; AJ301860&63 Y17640 AJ276085 Y17643; AJ132664; AJ245637 X58725; AJ301859 AJ276083 AJ276087 U36591; AJ276089; Y17645 Y17648 Y17638; U36590 U36592 U96144 U31995; U96139 Z14007 U96145 U96143 U96142 U96140 U96141 AF213462 AJ276077; AJ276078; Y17650&49 AJ133706 Y17651; AJ132666; X86687; AJ276088 AJ301858 L20824 Y17652 AJ301864 AJ301865 AJ301855 AJ301857 AJ301856 AF139733 Z14005 Z14004 Z14006 Z14011 AJ006466; AJ301861 AB015052 AJ006800 Y17634 AJ006801 AJ301862 AJ276081; AJ276082 AJ006799 AJ006798 Z14009 U96146 Z14010 X58726 U31997; AF038590 AB041344; AB041345 U36593 Z14013 Z14012 AJ242729 X86686; AJ276074; AJ132665; Y15904, Y15905; Y17831

Note. Sequences from near-full-length clones delivered by the present study are printed in bold letters and underlined. a DAOM-no. (Dept. Agriculture, Ottawa, Mycology, Canada), herbarium voucher no.; W-no./Att-no. (collection of Chris Walker, Great Britain), voucher no./culture no.; MAFF-no. (“Ministry of Agriculture, Forestry, and Fisheries, Japan), culture identity no.; BEG-no., culture identity no.; for INVAM culture identities, see Morton et al. (1993). b N, no details; F, root fragments; M, multispore; S, single spore; SC, spore cluster; ST, soil trap, c More than one accession no. means that a consensus sequence of those was used for analyses. d G. manihotis and G. clarum are suggested by Morton to be synonymous. 192

Glomus (GLOMALES) IS NONMONOPHYLETIC

193

FIG. 1. Phylogenetic tree of the AM fungi (Glomales). The family ranking proposed in this study is indicated by shaded ovals. Distances are derived from a nonbootstrapped neighbor-joining analysis of 1597 sites. Neighbor-joining and Parsimony analyses (1000 bootstraps each) result in the same consensus tree topology, supporting the monophyly of the Glomales with 84%–94%. Thick lines indicate bootstrap support of at least 90% in both analyses.

The Genus Glomus To better resolve the phylogeny within the Glomales, analyses of the glomalean sequences only, with Endogone and Mortierella as outgroup organisms, were performed (Fig. 2). This allowed the analysis of a 83sites longer alignment, compared to that also comprising the other fungal taxa (Fig. 1). The genus Glomus sensu Morton & Redecker represents by far the largest genus within the Glomales and is placed in the monogeneric family Glomaceae. Most of the species we examined from the Glomaceae sensu Morton & Redecker separate naturally into two major clades, shown as group A and B, each supported by a bootstrap value of 100%. The monophyly of both groups is supported by bootstrap values of about 80%. The clade “Glomus group A” (GlGrA; following the terminology of Schu¨ßler et al., 2001) comprises two subclades (GlGrAa and -b), each highly supported (bootstrap values 98 –100%). The first (GlGrAa) contains G. geosporum, G. mosseae, G. fragilistratum, G. caledonium, G. verruculosum, G. sp. WUM3, G. sp. DAOM225952, and G. coronatum. The other (GlGrAb) encompasses G. intraradices, G. fasciculatum, G. proliferum, G. coremioides, G. sinuosum, G. vesiculiferum,

G. clarum, and G. manihotis. The type specimen of G. manihotis clearly is related to the latter subclade of GlGrA and clusters together with G. clarum. Both species have been suggested to be synonymous. A published sequence from another supposed G. manihotis isolate (BR212) clusters with GlGrB. This isolate could not be reinvestigated because it is lost at INVAM. A fungus (G. sp. W3347), as yet unidentified but with spore morphology similar to G. macrocarpum, forms an “outlier,” probably as a member of a third subclade (GlGrAc), though no close relatives have yet been sequenced. It is poorly supported as more closely related to subclade Aa then to Ab (bootstrap support 65%). The other clade, Glomus group B (GlGrB), is composed of the species G. claroideum, G. lamellosum, G. manihotis (and probably G. clarum, whether synonymous or not), G. etunicatum, G. viscosum, G. luteum, and two undescribed species. Three of the “Glomus” species analyzed fall into a clade (GlGrC; Fig. 2) separate from all other members of the “genus” Glomus. These species, G. spurcum, G. versiforme, and a species conforming to the description of G. etunicatum, share few morphological characters except that they are easily distinguishable, by spore

194

¨ ßLER SCHWARZOTT, WALKER, AND SCHU

FIG. 2. Phylogenetic neighbor-joining consensus tree of the AM fungi (Glomales). The family ranking proposed in this study is indicated by arrowheads. Neighbor-joining (values of 10000 bootstraps above branches) and Parsimony analyses (values of 1000 bootstraps below branches) of 1680 sites in alignment both result in the same main tree topology. Thick lines indicate bootstrap support of at least 90% in both analyses. Where no parsimony bootstrap values are given, the topology differs from the neighbor-joining tree shown or values are below 60%. Clades with support below 60% in both analyses are collapsed to polytomies. Distances are derived from a nonbootstrapped neighbor-joining analysis.

development and morphology, from members of genera other than Glomus sensu lato. They are part of a monophyletic clade shared with the Gigasporaceae and Acaulosporaceae (bootstrap support 91 and 90% in the neighbor-joining and parsimony analysis, respectively), with a moderately (66 and 82%) supported closer relationship with the Acaulosporaceae.

DISCUSSION This study mainly considers one large group of the AM fungi, the genus Glomus sensu Morton & Redecker. It is currently placed in the monogeneric family Glomaceae, and represents by far the largest genus within the Glomales. Walker (1992) drew attention to the fact that, even by classical criteria, this genus may

Glomus (GLOMALES) IS NONMONOPHYLETIC

be polyphyletic and that the taxonomic position of some Glomus species in the current broad interpretation of the genus is unclear. Our studies presented here clearly confirm this. Despite the separation of the Paraglomaceae and Archaeosporaceae (Morton and Redecker, 2001), the fungi presently classified within the Glomaceae are still separated by phylogenetic distances at least as large as between other families within the Glomales. One of the three clades (GlGrC) clusters with other families, rendering the Glomaceae, presently given monophyletic family status, as polyphyletic or paraphyletic. Our data can be used to subdivide Glomus sensu Morton & Redecker into several groups that are phylogenetically consistent but require further investigation at the generic level and below. If seen from the view of a natural system based on phylogenetic distances, the Glomaceae should be divided into three families representing GlGrA, GlGrB, and GlGrC. Because the former two “families” seem to be monophyletic, no urgent need exists to change their classification pending a deeper understanding of the phylogenetic relationships within the Glomales. But for the monophyletic clade GlGrC, a change in the taxonomic classification of the fungi within it is inevitable. Nevertheless, the only evidence for its existence at present is from our analysis of SSU rRNA gene, and although we hypothesize the existence of a family, we consider it prudent to await more data before its formal creation (see below). The “Glomus mosseae Group” (GlGrA) and “Glomus etunicatum Group” (GlGrB) In one of the first analyses of AM fungal SSU rRNA gene sequences, Simon et al. (1993) found unexpectedly large genetic distances among the genus Glomus. He suggested that a species identified as G. etunicatum isolate UT316 (formerly INVAM147), because of its position distinct from other Glomus species, might be placed in a separate family. However, only four species of Glomus were used in that study. The authors consequently did not make formal emendations to the family structure. Investigations of a 550-bp fragment of the SSU rRNA gene (Schu¨ßler et al., 2001) recently showed that G. etunicatum UT316 belongs to a clade (GlGrB) comprising several well-characterized Glomus species. These had not been previously recognized as a separate group, very distantly related to the other Glomus clades, GlGrA and GlGrC. From the improved phylogenetic resolution allowed by the nearly complete SSU rRNA sequences, we now show that GlGrB is a large, monophyletic clade containing several well-characterized Glomus species. In turn, it forms a sister group to GlGrA, containing most other of the members of Glomus sensu lato so far analyzed. GlGrA comprises two or three subclades (a– c). The suggested synonymy of G. manihotis and G. clarum is

195

supported when only the fungi clustering in GlGrA are considered. However, G. manihotis (clarum) BR212, sequenced in earlier studies, clusters within GlGrB. Because this isolate is lost (Morton, personal communication), we could not proof its characters in culture and the question about synonymy remains open. G. sp. W3347 probably represent a member of a third subclade (c) within GlGrA, because it does not cluster firmly with one of the other subgroups (a and b). The phylogenetic distance between the “Glomus” clades GlGrA and GlGrB is larger then between the families Gigasporaceae and Acaulosporaceae. From the molecular data, the classification in higher taxa of glomalean fungi will require emendation by the erection of a new family to accommodate the species clustering in one of these groups. However, because this emendation is dependent on the yet unknown position of the generic type species, Glomus microcarpum, further evidence must be found before this can be done. A Putative New Family, the “Diversisporaceae” fam. ined. (GlGrC) Beside several indications of new taxonomic concepts within the AM fungi, as discussed in this study, one needed change is abundantly evident. Species in the clade GlGrC, comprising G. versiforme, G. spurcum, and a fungus (W2423) that resembles G. etunicatum, do not belong to the Glomaceae. This can only be resolved at higher taxonomic levels by removing them from the Glomaceae. GlGrC probably represents a sister group to the Acaulosporaceae, but this topology is supported only by bootstrap values of 66/82% (NJ/parsimony; Fig. 2), and by 76/86% in the analysis comprising all higher fungal taxa (Fig. 1). This moderately supported topology renders Glomus to be probably polyphyletic. Even if GlGrC would represent—in an alternative topology—the first branch within the highly supported Gigasporaceae/Acaulosporaceae/ GlGrC clade, it renders Glomus to be paraphyletic. Based on the phylogenetic data, a new family is suggested here to accommodate the species in GlGrC, the Diversisporaceae fam. ined. At the molecular level, the “Diversisporaceae” can be characterized by several specific sequence signatures, e.g., “GTGAGATGRRTCTCTACCTTC” (corresponding to homologous position 677 of the Saccharomyces cerevisiae SSU rRNA sequence J01353). Although it may be possible to define it primarily from the rDNA sequence fragment, it is morphologically so disparate that it seems best at present to reserve judgement until more evidence is forthcoming. To render the main “Glomus” clade monophyletic, it will be necessary to recognize at least one new genus within the Diversisporaceae fam. ined. However, despite the phylogenetic proximity to the Acaulosporaceae, the species in GlGrC are not known to share characters with Acaulospora, except for several SSU rRNA sequence signatures, and

196

¨ ßLER SCHWARZOTT, WALKER, AND SCHU

there is no obvious similarity in spore morphology. Thus, at present, there are no clear characteristics to use in defining this clade or its members. Consequently it is not formally erected pending the discovery of appropriate characters (e.g., ultrastructure) for a formal description. Work is in progress to identify suitable characteristics to use in defining the genera within the suggested new family. Further Phylogenetic Implications The species concept and isolate definitions within the glomalean fungi are controversial and these fungi are not known to possess sexual reproduction. More data are clearly needed to improve the classification of the Glomales at the generic level. Nevertheless, higher taxonomic groupings can be clearly separated by molecular analyses, and it is possible to reach conclusions on how to achieve a natural systematics, as recently also demonstrated for other fungal taxa which were based on morphological characters not reflecting the natural phylogeny (e.g., O’Donnel et al., 2001). A monophyletic origin of the Glomales was indicated by several molecular studies (Gehrig et al., 1996; Redecker et al., 2000b; Schu¨ßler, 1999; Tanabe et al., 2000). However, Morton (2000) questioned the molecular data and, based on morphological and biochemical data, suggested the Glomales are paraphyletic, with the suborders Glominaeae (containing the Glomaceae and Acaulosporaceae) and Gigasporineae (containing the Gigasporaceae) as two independent lineages arising from separate ancestors. The molecular studies cited above consistently contradict this, and the monophyletic origin of the Glomales, including the ancestral lineages, is verified by the molecular evidence shown here. The suborders Gigasporineae and Glominaeae are not congruent with the natural phylogeny of the AM fungi, because the Acaulosporaceae and Gigasporaceae (presently placed in different suborders) belong to a monophyletic clade (including GlGrC), which itself shares common ancestry with the remaining “Glomaceae” (GlGrA and GlGrB). The family and higher ranking concept within the Glomales is now becoming clear at the molecular level, though there is also conflict with the existing scheme (Morton and Benny, 1990; Morton and Redecker, 2001). The three presently known main lineages within the Glomales are represented by (1) the Glomaceae–Acaulosporaceae–Gigasporaceaea grouping, (2) the Paraglomaceae; and (3) the Archaeosporaceae– Geosiphonaceae grouping. It is not yet possible to draw final conclusions about the branching order between these lineages. The Paraglomaceae and Archaeosporaceae–Geosiphonaceae clades cluster together with very weak bootstrap below 60% (collapsed to polytomy in Fig. 2). However, if the Paraglomus clade is excluded from the analyses, the bootstrap support for a monophyletic origin of the remaining glomalean fungi rises

significantly. This means that the Paraglomaceae do not cluster as robustly within the Glomales as the other clades, which indicates that this clade probably represents the most ancient group within the Glomales. Nevertheless, such conclusions must be interpreted with care because species sampling can influence the lower bootstrap values considerably, and future data from more glomalean organisms may answer this question. The Archaeosporaceae does not represent a valid family concept, when phylogenetic relationships are considered. Our analyses show that Geosiphon pyriformis (Geosiphonaceae) unquestionably falls into the Archaeosporaceae, though this fact was rejected in its protologue (Morton and Redecker, 2001; Redecker et al., 2000b). This calls into question the use of arbuscule formation as a fundamental taxonomic character (synapomorphy) for the order Glomales. On present evidence, it is not possible to confirm if arbuscule formation represents a convergent character (homoplasy) or a shared ancestral character (symplesiomorphy). Certainly, the apparent (but not provable) lack of AM formation by the Geosiphon fungus does not validate its exclusion from the taxonomic grouping. With respect to natural relationships, which must be the basis for a completely acceptable taxonomic concept, Archaeospora and Archaeosporaceae sensu Morton & Redecker, as currently constituted, are paraphyletic and cannot be sustained. The situation could easiest be resolved by a three-family concept as indicated in Figs. 1 and 2 or, alternatively, by a two-family concept, whereas one has to comprise Ar. leptotichum as well as Geosiphon. More detailed work is needed before the taxonomy and systematics of the AM fungi and their close relatives are truly understood. This paper contributes to the necessary reorganization by identifying the natural groupings at family level. Investigations are now in progress to link morphological and molecular data to a generally more natural taxonomic concept of these organisms, which also has an important impact on the interpretation of physiological, molecular, genetic, and biodiversity studies. ACKNOWLEDGMENTS Funding for this work was provided by the Deutsche Forschungsgemeinschaft (SCHU1203). Thanks to all those who provided cultures, isolates, and sequences.

REFERENCES Felsenstein, J. (1989). PHYLIP: Phylogeny Inference Package (version 3.2). Cladistics 5: 164 –166. Felsenstein, J. (1983). Parsimony in systematics: Biological and statistical issues. Annu. Rev. Ecol. Syst. 14: 313–333. Gehrig, H., Schu¨ßler, A., and Kluge, M. (1996). Geosiphon pyriforme,

Glomus (GLOMALES) IS NONMONOPHYLETIC a fungus forming endocytobiosis with Nostoc (Cyanobacteria), is an ancestral member of the Glomales: Evidence by SSU rRNA analysis. J. Mol. Evol. 43: 71– 81. Gerdemann, J. W., and Trappe, J. M. (1974). The Endogonaceae in the Pacific Northwest. Mycolog. Mem. 5: 1–76. Greuter, W., McNeill, J., Barrie, F. R., Burdet, H. M., Demoulin, V., Filgueiras, T. S., Nicolson, D. H., Silva, P. C., Skog, J. E., Trehane, P., Turland, N. J., and Hawksworth, D. L. (2000). “International Code of Botanical Nomenclature (Saint Louis Code).” Koeltz Scientific Books, Germany. Kimura, M. (1980). A simple model for estimating evolutionary rates of base substitutions through comparative studies of nucleotide sequences. J. Mol. Evol. 16: 111–120. Kumar, S., and Gadkar, S. R. (2000). Efficiency of the neighbourjoining method in reconstructing deep and shallow evolutionary relationships in large phylogenies. J. Mol. Evol. 51: 544 –553. Morton, J. B. (2000). Evolution of endophytism in arbuscular mycorrhizal fungi of Glomales. In “Microbial Endophytes” (C. W. Bacon, and J. H. White, Eds.), pp. 121–140. Dekker, New York. Morton, J. B., and Benny, G. L. (1990). Revised classification of arbuscular mycorrhizal fungi (Zygomycetes): A new order, Glomales, two new suborders, Glomineae and Gigasporineae, and two new families, Acaulosporaceae and Gigasporaceae, with an emendation of Glomaceae. Mycotaxon 37: 471– 491. Morton, J. B., Bentivenga, S. P., and Wheeler, W. W. (1993). Germ plasm in the international collection of arbuscular and vesicular– arbuscular mycorrhizal fungi (INVAM) and procedures for culture development, documentation, and storage. Mycotaxon 48: 491– 528. Morton, J. B., and Redecker, D. (2001). Two new families of Glomales, Archaeosporaceae and Paraglomaceae, with two new genera Archaeospora and Paraglomus, based on concordant molecular and morphological characters. Mycologia 93: 181–195. Nei, M., Takezaki, N., and Sitnikova, T. (1995). Assessing molecular phylogenies. Science 267: 253–255. O’Donnel, K., Lutzoni, F. M., Ward, T. J., and Benny, G. L. (2001). Evolutionary relationships among mucoralean fungi (Zygomycota): Evidence for family polyphyly on a large scale. Mycologia 93: 286 –296. Redecker, D., Kodner, R., and Graham, L. E. (2000a). Glomalean fungi from the Ordovician. Science 289: 1920 –1921. Redecker, D., Morton, J. B., and Bruns, T. D. (2000b). Ancestral lineages of arbuscular mycorrhizal fungi. Mol. Phylogenet. Evol. 14: 276 –284. Remy, W., Taylor, T. N., Hass, H., and Kerp, H. (1994). Four hundred-million-year-old vesicular arbuscular mycorrhizae. Proc. Natl. Acad. Sci. USA 91: 11841–11643.

197

Saitou, N., and Nei, M. (1987). The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4: 406 – 425. Schu¨ßler, A. (1999). Glomales SSU rRNA gene diversity. New Phytol. 144: 205–207. Schu¨ßler, A., Gehrig, H., Schwarzott, D., and Walker, C. (2001). Analysis of partial Glomales SSU rRNA genes: Implications for primer design and phylogeny. Mycol. Res. 105: 5–15. Schu¨ßler, A., and Kluge, M. (2001). Geosiphon pyriforme, an endocytosymbiosis between fungus and cyanobacteria, and its meaning as a model system for arbuscular mycorrhizal research. In “The Mycota IX—Fungal Associations” (B. Hock, Ed.), pp. 151–161. Springer-Verlag, Berlin/Heidelberg. Schwarzott, D., and Schu¨ßler, A. (2001). A simple and reliable method for SSU rRNA gene DNA-extraction, amplification, and cloning from single AM fungal spores. Mycorrhiza 10: 203–207. Simon, L., Bousquet, J., Le´vesque, R. C., and Lalonde, M. (1993). Origin and diversification of endomycorrhizal fungi and coincidence with vascular land plants. Nature (London) 363: 67– 69. Tanabe, Y., O’Donnell, K., Saikawa, M., and Sugiyama, J. (2000). Molecular phylogeny of parasitic Zygomycota (Dimargaritales, Zoopagales) based on nuclear small subunit ribosomal DNA sequences. Mol. Phylogenet. Evol. 16: 253–262. Taylor, T. N., Remy, W., Hass, H., and Kerp, H. (1995). Fossil arbuscular mycorrhizae from the Early Devonian. Mycologia 87: 560 –573. van de Peer, Y., Caers, A., De Rijk, P., and De Waechter, R. (1998). Database on the structure of small subunit RNA. Nucleic Acids Res. 26: 179 –182. Walker, C. (1992). Systematics and taxonomy of the arbuscular endomycorrhizal fungi (Glomales)—A possible way forward. Agronomie 12: 887– 897. Walker, C. (1983). Taxonomic concepts in the Endogonaceae: Spore wall characteristics in species descriptions. Mycotaxon 18: 443– 455. Walker, C. (1979). The mycorrhizast and the herbarium: The preservation of specimens from VA mycorrhizal studies. In “Program and Abstracts of the 4th North American Conference on Mycorrhizas.” Fort Collins, Colorado. Walker, C., Mize, C. W., and McNabb Jr., H. S. (1982). Populations of endogonaceous fungi at two locations in central Iowa. Can J. Bot. 60: 2518 –2529. Walker, C., and Trappe, J. M. (1993). Names and epithets in the Glomales and Endogonales. Mycol. Res. 97: 339 –344. Walker, C., and Vestberg, M. (1994). A simple and inexpensive method for producing and maintaining closed pot cultures of arbuscular mycorrhizal fungi. Agricult. Sci. Finland 3: 233–240.