Walker and Schüßler (2002)
Arbuscular mycorrhizal (AM) fungi are soil-borne fungi that establish an obligately mutualistic symbiosis with many plant species. They are characterized by transient dichotomously-branched arbuscules that form inside cortical cells of plant roots from intraradical branch hyphae. Spores appear to be obligately asexual, forming outside or, less frequently, within plant roots singly or in loose aggregates. In some cases they are produced in compact and highly organized sporocarps. Fungi in some families form their spores in ways not found in any other fungal group, and species in all families synthesize unique and often diverse subcellular structural components of the spore wall and (in some groups) flexible inner "germinal" walls. These morphological properties attest to an ancient evolutionary departure from other fungal lineages. The closest nonmycorrhizal relative appears to be Geosiphon pyriforme, a symbiont that hosts a Nostoc cyanobacterium.
At the present time, provisionally, Glomales is grouped in the phylum Zygomycota. Arbuscular fungi originally were placed in the family Endogonaceae of the order Mucorales because of a superficial resemblance between spores of Endogone species and those of sporocarpic Glomus species (Gerdemann and Trappe, 1974). This resemblance was not a valid grouping criterion since Endogone spp. form sexual zygospores while AM fungi produce only asexual spores. Later, they were grouped in Glomales to recognize their monophyletic origin and unique evolutionary trajectory (Morton and Benny, 1990). Molecular evidence (as small ribosomal subunit RNA sequences) complemented morphological evidence to affirm this group is a distinct lineage and is a putative sister group to Basdiomycota and Ascomycota (the figure at right is from Schüßler et al. (2001). This phylum is ancient and has had substantial time to evolve. Arbuscular fungi have been found in 400-million-year-old fossils from the Devonian Period (Taylor et al., 1995).
The history of members of Glomales is long and convoluted. Early on, member species usually were placed in Endogone, which was erected in 1809 by the German mycologist, Link. In 1844, the genus Glomus was described by the Tulasne brothers (Tulasne and Tulasne, 1844), then comprised of two species, G. microcarpum and G. macrocarpum. When first described, Glomus was considered by the Tulasnes to be closely allied to Endogone, which was placed in the Tuberaceae by Fries (thanks to Jim Trappe for pointing this out). The genus Sclerocystis was described by Berkeley and Broome in 1875. Thaxter revised the family Endogoneae (note spelling) in 1922, placing all members of Glomus in the genus Endogone, while maintaining the genus Sclerocystis. Bucholtz (1922) placed the Endogonaceae in the Mucorales, due to the affinities of Endogone with members of the Mortierellaceae. The Endogonaceae were placed in their own order, Endogonales, by Moreau (1953), which was later validated by Benjamin (1979).
The next landmark publication came in 1974, when Jim Gerdemann and Jim Trappe resurrected Glomus in the Endogonaceae and moved several species from Endogone to Glomus. They also described the genera Acaulospora and Gigaspora. Ames and Schneider (1979) described the genus Entrophospora, and Walker and Sanders (1986) split Scutellospora from Gigaspora sensu lato. The present-day family Glomaceae was formally erected by Pirozynski and Dalpé in 1989. The order Glomales was erected by Morton and Benny (1990), along with two suborders Glomineae and Gigasporineae, and two other families Acaulosporaceae and Gigasporaceae. The genus, Glomites was erected by Taylor et al. (1995) to describe fossil fungi that closely resemble modern-day Glomus species.
Over the years, current members of Glomales (in parentheses) have been placed in the following genera: Ackermannia (Sclerocystis); Endogone (Glomus, Gigaspora, Scutellospora); Paurocotylis (Glomus); Sphaerocreas (Glomus, Sclerocystis); Stigeosporium (Glomus); Stigmatella (Glomus); Rhizophagites (Glomus); Rhizophagus (Glomus); Xenomyces (Sclerocystis).
Members of other groups (in parentheses) which were once thought to have affinities to AM fungi, and thus placed in the Endogonaceae include the following: Glaziella (Glaziellaceae, Pezizales); Complexipes (affinities unknown, currently classified as a hyphomycete); Modicella (Mortierellaceae, Mucorales).
The separation between Glomus and Sclerocystis became controversial in the early 1990's. Almeida and Schenck placed all Sclerocystis species in Glomus with the exception of S. coremioides. Wu (1993) resisted this change and opted for a return to Gerdemann and Trappe's (1974) classification scheme. Wu's rationale for genus-level separation, based on ontogeny of sporogenesis and sporocarp formation, was not convincing. More recently, Redecker et al. (2000b) transferred S. coremioides to Glomus with the discovery that 18S rDNA sequences placed the species well within the Glomus clade.
Two new families were erected by Morton and Redecker (2000) based on some atypical morphological characters, striking immunological and fatty acid distance, and 18S rDNA sequence divergence. Two dimorphic sister species with similar ontogenetic sequences, Acaulospora gerdemannii and Glomus gerdemannii (sensu lato), together with Acaulospora trappei (sensu lato) were transferred to Archaeosporaceae. Two ancestral species previously classified in Glomus, G. occultum and G. brasilianum (sensu lato), were grouped in a sister family, Paraglomaceae. The 18S sequence data complement other data to affirm that Glomus is not monophyletic, but likely consists of at least two divergent lineages (Schwarzott et al., 2001).
The 18S rRNA sequences provide some enlightening information that clarifies some of the vagueness that morphology simply cannot resolve. However, there is a huge conflict between the proposed classification and the one in place, and that is the grouping of Acaulosporaceae with Gigasporaceae in Diversisporales. There is not just complete divergence in the process of spore formation and traits of inner wall development, but also in mycorrhizal structures, timing of their formation, infectivity of those structures, fatty acid profiles, cell wall chemistry, and even timing and abundance of transcript synthesis during spore germination. On this website, we are skeptical of these relationships from rRNA data alone, and resolution will require congruence with other genes. If that occurs, then morphological and behavioral patterns are creating a misleading picture that can be explained, but by reoccurrence of traits and convergent evolution.