The first glance into the Glomus genome: an ancient asexual scandal with meiosis?


It has not yet been possible to assemble the full genome of an arbuscular mycorrhizal fungus (AMF), but in this issue of New Phytologist, a consortium of laboratories has published an analysis of the transcriptome from Glomus intraradices (Tisserant et al., pp. 755–769). The transcriptome of this fascinating universal root symbiont showed many features to be expected from an obligate plant symbiont, such as the lack of expression of hydrolytic enzymes and induced expression of genes coding for membrane transporters. Unexpectedly, genes related to meioses were also found raising the possibility that these apparently asexual fungi are able to perform a meiotic cycle.

‘Unexpectedly, genes related to meioses were also found raising the possibility that these apparently asexual fungi are able to perform a meiotic cycle.’

A biotrophic transcriptome

The transcriptome showed several characteristics which have been seen in true obligate biotrophs that rely on the host’s metabolism, such as the loss of thiamine biosynthesis and the lack of invertase. An important consequence of the latter point cannot simply be that the fungus can only access the host plant’s glucose. In addition, no genes for plant cell wall degrading enzymes were detected. Interestingly, GH5 and GH7 endoglucanases were found, but this was interpreted as a way of removing cellubioses that would trigger the plant immune defense system. Likewise, this can be seen as an adaption to the mutualistic life style of the fungus.

A number of genes that code for enzymes involved in nitrogen and sulfur metabolism were also detected. This is in contrast to the true biotrophs, such as rusts, and is probably linked to the ability of the fungus to take up nutrients from the soil. This is interesting as it might suggest that either the ancestors of AMF were saprobic soil fungi, or the genes have evolved parallel to genes in other soil fungi. Future phylogenetic studies of these genes will undoubtedly throw light on this question.

Sequence polymorphism

Sequence polymorphism was also detected in the transcriptome of Glomus intraradices. The number of SNPs (single nucleotide polymorphisms) varied, with the highest number in a ribosomal protein (15 SNPs) and in a Ras-related Rab (11 SNPs). The polymorphism indicated that similar gene variants were functionally active raising questions of whether the polymorphism is allelic or whether it reflects repetitions in the genome. The SNPs were not analyzed further, but future studies on how the polymorphism might have originated, might reveal interesting insights into the genetics and evolutionary biology of this fungus.

Meiosis related proteins

The meiotic genes detected in the expressed sequence tag (EST) library included HOP2, MND1 and MER3 that are known to function in meiosis (Malik et al., 2008). If these were the only meiosis-related genes found, it could be postulated that the genes have evolved a new function in asexual AMF, but a new study has recently identified several meiosis-related genes in four other AMF, making it highly likely that all these meiosis-related genes have retained ancestral functions (Halary et al., 2011).

The key meiotic recombinase SPO11 was not found by Tisserant et al. but was detected in other species (Halary et al., 2011). It is not known whether this gene is actually absent in the isolate of Glomus intraradices, or whether it was not expressed at the studied developmental stage because the fungus was not involved in the meiotic process. From other fungal model organisms it is known that the expression level of SPO11 depends on the meiotic stage, reaching a peak early in meiosis, but the transcript profiles were also seen to differ between species (Burns et al., 2010). Given the derived position in Eumycetes phylogeny, the finding of meiosis genes in G. intraradices strongly suggests that meiosis was the ancestral state in the Glomeromycota.

Another indication of a sexual cycle in Glomus was the finding of genes with similarity to the sexP and sexM genes from Phycomyces blakesleeannus. These genes encode a high-mobility group domain protein that could be ancestral to the mating-type (MAT) loci in Dikarya (Idnurm et al., 2008). Further studies of these genes might not only reveal interesting functions in Glomeromycota, but also clarify evolutionary questions regarding sex determination in fungi.

The evolutionary advantage of sex

There are a considerable number of reviews and commentaries dealing with the subject of sexuality vs asexuality and discussion of the evolutionary advantage of sex (see e.g. Barton & Charlesworth, 1998; Sun & Heitman, 2011). The central idea is that recombination increases the genetic diversity offered to natural selection by breaking down the negative correlation between favorable variants at different loci. Perhaps just as importantly, recombination involving mutated genes can lower the mutational load by removing mutated alleles from the genome.

The hypothesis that sex increases adaptation in a changing environment, biotically or abiotically (Red Queen Hypothesis), is difficult to test experimentally, but experimental evidence exists that yeast populations under stress benefit from recombination (Goddard et al., 2005). It can, of course, be argued that AMF occur in a very stable niche, as the plant cell environment where the fungus obtains its nutrients can be considered a relatively unchanging environment. In this respect sex should be disadvantageous as recombination might break beneficial linkages between symbiotic genes adapted to this particular environment. However, as a mutualist you might have to change fast to catch up with the evolution of your partner, for example, to avoid cheaters.

Recombination and meiosis have also been proposed as the only way organisms can reduce the mutation load. By recombining genes with high mutation loads that are subsequently lost, the organism can purge mutations in the genome, whereas asexual organisms cannot accumulate such mutations (Müller’s Ratchet). This has been used to argue that asexual lineages can only be short lived, and that ancient asexuality is an evolutionary scandal (Maynard Smith, 1986).

Is Glomus clonal?

Why do we consider AMF to be asexual? The first indication is, of course, that no sexual structures are observed, and the spores of AMF do not resemble any known sexual structures in fungi. Another indication of asexuality is the clonal population structure observed in the field. Only a few studies have been carried out, and these studies do show a strong linkage between alleles from different loci (Stukenbrock & Rosendahl, 2005; Croll & Sanders, 2009). In both cases the population genetic structure deviated from what would be expected from a sexual population. However, the clonal structure is interpreted from the phylogenetic congruence between the loci, but this congruence can also be obtained if the loci are sampled from noninterbreeding sexual individuals. This could be the case if what are considered as individuals of, for example, Glomus mosseae, in fact represented discrete populations. As the criteria for recognizing species of AMF are not trivial (Rosendahl, 2008), the linkage of alleles in the field cannot rule out meiosis and recombination in AMF.

Is Glomus an ancient asexual scandal?

Whereas it is easy to confirm asexuality either by direct observation of asexual structures (e.g. conidia or sporangia) or by population genetic detection of linkage disequilibrium, it is more complicated to assess when, or if, the organisms have lost sexuality (Judson & Normark, 1996). Even if the population structure is clonal, we cannot conclude that it is ancient. Consider a coalescent model where the time to the last common ancestor is 2Ne (where Ne is the effective population size). If all loci are linked as in an asexual population the coalescent time will be shorter than under neutral expectation, and linkage disequilibrium will be lost in a few generations. Moreover, the fungi of interest might have undergone a recent population expansion in which Ne at the outset could have been very low (Rosendahl et al., 2009), and it is thus difficult to conclude that an observed clonal population structure is ancient.

Several proposed ancient asexuals have been studied. The most well known scandal is the 50 Myr-old bdelloid rotifers studied by Mark Welch & Meselson (2000). One of the signatures of ancient asexuality is the lack of allelic congruence at a given locus due to the lack of the homogenizing effect of meiosis: accumulation of mutation entails a divergence between alleles that grows with time. This has been named the Meselson effect, and although it is only valid for diploid organisms it could also be applied for AMF where the genetic polymorphism among nuclei in spores can be interpreted as extreme heterozygosity, and thus an indication of ancient asexuality (Hijri et al., 2001).

There might not be a contradiction between the finding of meiosis-related genes and the clonal population structure observed in the field, as the fungi might not be ancient asexuals. The reason AMF have been interpreted as ancient asexuals could simply be due to the fact that phylogenetic analyses place them as basal to the Dikarya, combined with the finding of structures resembling extant AMF in fossilized rhizomes (Redecker et al., 2000). We should, however, not forget that we have no evidence that the ancestors of extant AMF were asexual or even that they were mutualists.

Conclusion: is Glomus sexual or obligate asexual?

The authors behind the Glomus transcriptome (Tisserant et al.) conclude that the fungus has the genes for meiosis. But can it be concluded from the expression of the genes that the fungus undergoes meiosis? The expression of the genes is definitely necessary for meiosis, and if they were not functioning they would most likely have been loaded with mutations. However, it remains debatable whether the genes could have a function during mitosis if the genes were involved in mitotic recombination. However the direct observation of linkage disequilibrium in the field cannot be used to declare the fungi to be obligate asexual, and the evidence of them being ancient asexual seems to be even more vague. The published transcriptome is beyond doubt an important contribution to the understanding of the reproductive biology of AMF, but we still do not know where we are on the ‘long hard road’ to understanding the genetics and evolution of these fungi.


Many thanks to Marc-André Selosse and Olaf Nielsen for helpful comments and discussions.