A molecular guide to the taxonomy of arbuscular mycorrhizal fungi


In this issue of New Phytologist, Krüger et al. (pp. 970–984) present an extensive molecular phylogeny of arbuscular mycorrhizal fungi (AMF). This is the basis for a radical modernization of the taxonomy and will be an important point of reference for future studies.

Sift through any soil in which plants have grown, and there is a good chance that it will contain the spores of AMF. It requires a careful eye to spot them, though, as even the largest, the ‘giant spores’ of Gigaspora gigantea, are less than half a millimetre across (Fig. 1). The other parts of the fungus are even less conspicuous: a gossamer web of hyphae spreading through the soil, and structures within plant roots that are only visible under the microscope. Unsurprisingly, the identification of AMF is challenging, but it is important because AMF form symbioses with the majority of land plants. They have major effects on plant growth in a number of ways, the best known being by aiding phosphorus uptake (Smith & Read, 2008).

Figure 1.

Spores of arbuscular mycorrhizal fungi (AMF) representing the different genera of the Glomeromycota. Row 1, Glomus, Archaeospora, Redeckera, Pacispora; row 2, Acaulospora, Rhizophagus, Claroideoglomus, Racocetra; row 3, Paraglomus, Diversispora, Sclerocystis, Scutellospora; row 4, Ambispora, Funneliformis, Gigaspora, Entrophospora. Bars, 200 μm. Figure courtesy of C. Walker.

‘It has been obvious for years, ever since molecular sequences began to accumulate, that the taxonomy of AMF was inadequate and needed radical reform.’

Putting new names to old faces

Even mycorrhizal experts are likely to be baffled on initial reading of the paper by Krüger et al., because it uses many unfamiliar names for familiar organisms. These names reflect the well-supported clades that the authors establish by the phylogenetic analysis of substantial parts of the nuclear locus encoding ribosomal RNA (Fig. 2). It has been obvious for years, ever since molecular sequences began to accumulate, that the taxonomy of AMF was inadequate and needed radical reform. All AMF form a deep, monophyletic branch within the fungi – at least, it is monophyletic if one includes Geosiphon, a fungus, so far unique, that forms a symbiosis with a cyanobacterium rather than a plant (Gehrig et al., 1996). The AMF branch is so diverse that it deserves the taxonomic rank of a phylum, the Glomeromycota (Schüßler et al., 2001). Its exact placement within the fungal tree is not yet certain, but it is clearly outside the Dikarya, which comprise the familiar ascomycetes and basidiomycetes (Lee & Young, 2009). Within the Glomeromycota there are a number of clearly defined clades that form the basis for orders and families, as documented by Krüger et al. In the classical, morphology-based taxonomy, the genus Glomus predominated, but molecular comparisons made it evident that the spore morphology regarded as typical of this genus is actually seen in fungi belonging to multiple clades in all four orders of the Glomeromycota (Schwarzott et al., 2001). While some of the more conspicuous anomalies were recognized by taxonomic revisions some years ago, this new paper finishes the job by defining clades within the Glomerales that deserve the status of genera. The result is that nearly all the species formerly known as Glomus are now in new genera. This includes the strain that is currently the target of the first genome sequencing project for the Glomeromycota (Martin et al., 2008). When the project began, it was Glomus intraradices, but now it is Rhizophagus irregularis.

Figure 2.

Phylogeny of the Glomeromycota based on small sub-unit ribosomal RNA sequences, showing the relationships among the orders and genera. Families are indicated by distinct colours. Maximum Likelihood tree rooted with ascomycete outgroups, modified from a tree provided by A. Schüßler, drawn with Fig Tree.

While Krüger et al. provide the phylogenetic evidence to support the taxonomy, the new names are not proposed in this paper. In fact, they were defined a year earlier in another work, which was published through an unusual combination of eighteenth and twenty-first century technologies. The document does not appear in a scientific journal. It was officially published by lodging paper copies in several scholarly libraries in Germany, the UK and the United States (Schüßler & Walker, 2010). Since this is hardly convenient for most readers, a copy was also made available online through a web site (http://www.lrz.de/~schuessler/amphylo/). At the time, changes in fungal taxonomy were only validly published if they were on paper, although this requirement will soon disappear (Knapp et al., 2011). The rationale for rushing out the new nomenclature so far ahead of the supporting evidence was no doubt to establish priority for the authors’ chosen scheme.

The game is not over yet

In the days before molecular techniques, the only way to identify AMF was by a careful microscopic examination of the spores. Unfortunately, spores are relatively simple structures that offer only a limited number of potential discriminating features. To make matters worse, a single AMF culture can sometimes produce spores that differ substantially in morphology and colour, and spores found in soil are often not in pristine condition so they may not match fresh reference material produced under controlled conditions. Reliable identification requires a good deal of patience and experience, and only a handful of experts worldwide can confidently identify the whole range of AMF by examining spores. Expertise in DNA is more widespread, although AMF pose some challenges here, too. The strength of the study by Krüger et al. is that the team included both morphological and molecular expertise, and they also had access to a large collection of well-characterized cultures. Altogether, they analysed sequences for 109 named species, plus 27 cultures of species as yet unnamed. While this was a major undertaking, it represents only half the described species of AMF, of which there are currently 230 according to Krüger et al. Most of the rest are not available in culture. The number of described species is almost certainly a severe underestimate of the real diversity, if one considers that > 200 000 plant species form symbioses with AMF. It is certainly true that many AMF are capable of symbiosis with a wide range of plants, and the same species can be found in diverse habitats around the world. However, molecular tools reveal that most individual plants in the field are colonized by multiple species of AMF, and it seems unlikely that, when we finally get the full picture, the number of AMF species will remain as low as one thousandth of that of their host plants.

Indeed, the known diversity of molecular sequences that appear to derive from AMF is already much greater than the diversity of available cultures. The development of suitable primers and protocols has allowed us to identify AMF symbionts within plant roots collected in the field (Helgason et al., 1998). A number of different primers and target regions have been used, and these are summarized by Krüger et al., whose own analysis spans several of these regions. The most widely used region is part of the nuclear-encoded small-subunit ribosomal RNA, and these have been collected and curated in a database that facilitates research into their diversity and distribution (Öpik et al., 2010). While some of the DNA sequences recovered from field roots correspond closely to those of named AMF cultures, many do not, although they clearly fall within the AMF clade. These phylotypes presumably represent AMF species that are probably not currently in culture, given that Krüger et al. include the great majority of available cultures in their molecular analysis. There is some evidence that ‘weedy’ AMF, found in many habitats and many host plant species, are over-represented in culture, so that the isolates that are available for study do not reflect the whole diversity of AMF. An additional complication is that an AMF culture, even if established from a single spore, commonly carries several variants of the sequences used for classification. While these are commonly closely related, there are examples of substantially different sequences in a single isolate (Clapp et al., 2002).

For all these reasons, the taxonomy and identification of AMF remain challenging. The painstaking and wide-ranging analysis by Krüger et al. provides an invaluable chart for navigating these rock-strewn waters.