Plants colonized land in the Ordovician period (ca. 475 MYA) together with associated filamentous fungi assumed based on fossil evidence representing ancestors of today’s arbuscular mycorrhizal (AM) fungi (1, 2). The deep evolutionary origin of the symbiosis between AM fungi and plants is further supported by the fact that the plant fungal signaling pathway that initiates AM symbiosis is shared among all living land plant lineages (3, 4, 5, 6). AM fungi are well known mutualistic plant symbionts that cannot complete their life cycle without their plant partner. These fungi provide nutrients (e.g. phosphorous, nitrogen) to their photosynthetic partners in exchange for fixed carbon and contribute to the overall plant fitness, during biotic (e.g. infections) or abiotic (e.g. salinity and drought) stresses (7, 8). In the symbiosis, carbon for nutrient exchange happens in the colonized root cells where an apoplastic compartment is formed between plant and fungal membranes (9). Here the fungi obtain their energy and carbon in the form of sugars and lipids exuded by the plant host (10, 11, 12). Genomic evidence for the obligate biotrophic nature of AM fungi include their loss of the ability to independently synthesize multidomain fatty acids (13) and thiamine (14). Instead, AM fungi depend on their host for certain fatty acids as demonstrated by failed host colonization in plants with mutated fatty acids synthase genes (10, 11). Further, the difficulties to culture AM fungi axenically has in part been attributed to their lack of a thiamine metabolic pathway (15, 16). Based on genomes from three AM fungi, Gigaspora rosea, G. margarita and Rhizophagus irregularis, a set of 39 Missing Glomeromycota Core Genes (MGCGs), including fatty acids and thiamin synthesis, was proposed by Tang et al (2016). Reanalysis with an additional Rhizophagus strain detected six of the initial MGCGs in AM fungal genomes (15). In addition to assembly quality and detection criteria, taxon sampling is expected to affect the confidence by which we can call MGCGs as uniquely missing in all AM fungi and not in its sister lineages.
The unique role of AM fungi, as symbionts of the first land plants, has been challenged by the observation that Endogone-like fungi can also form beneficial associations with early diverging plant lineages, such as liverworts (17). In addition to ectomycorrhizal associations, facultative symbionts in Endogonales have been shown to colonize early-diverging lineages of both vascular and non-vascular plants, forming intracellular structures similar to those observed in AM colonization (18). The broad host range and morphological diversity observed in Endogonales, as well as observations of dual colonization together with typical AM fungi in liverworts representing the earliest diverging lineages of land plants (19) further support the notion that Endogone-like fungi may represent the earliest diverging mycobionts that facilitated land colonization (18). Symbiotic efficiency at different atmospheric carbon dioxide concentrations has led to the hypothesis that these ancestral mycorrhizal partners were later replaced by the now ubiquitous AM fungi that diversified together with flowering plants that dominate terrestrial ecosystems today (18). While it is widely accepted that both AM fungi and symbionts in Endogonales formed mutualistic interactions with early-diverging terrestrial plants (18, 20), the evolutionary relationships between these mutualistic fungal partners remains to be further explored for a better understanding of the dawn of terrestrialization of Earth.
The ectomycorrhizal (ECM) lifestyle is a different mutualistic symbiosis between fungi and vascular plants. The ECM habit has evolved multiple times mostly from diverse saprotrophic and endophytic ancestral fungal lineages predominantly in Dikarya but also in the Mucoromycota in the case of Endogonales (21, 22, 23). Saprotrophic and pathogenic fungi have large repertoires of genes coding for Carbohydrate-Active enZymes (CAZyme), many of which are directly involved in the degradation of plant cell walls e.g Plant Cell Wall Degrading Enzymes (PCWDEs) (24, 25). Among fungal lineages in Basidiomycota, the transition from saprotrophic growth to an ECM lifestyle is associated with a loss of PCWDEs, genome size expansion as a result of increased repeat content and a diversification of small secreted proteins (SSP) (21, 26, 27). Low numbers of PCWDEs are also observed in ECM lineages of Ascomycota (27). Analysis of a subset of CAZymes (45 families) in the ECM fungal lineage Endogonales indicate that PCWDEs were also lost during this independent origin of the ECM lifestyle (28). However, it was noted that the numbers of CAZyme genes in Mucoromycota was generally low across species having plant-associations and saprotrophic lifestyle, and the reconstructed reduction in gene copy numbers was much smaller compared to that associated with the evolution of ECM lineages in Dikarya (28). Similarity, a limited repertoire of genes involved in degradation of plant cell walls are detected in genomes of AM fungi (14, 29, 30) as well as in the related Nostoc-associated Geosiphon pyriformis (30). The low CAZyme gene numbers have been attributed to the symbiotic lifestyle of AM fungi but limited access to genome data has so far prevented a comprehensive analysis of contractions and expansions of CAZyme gene families of Glomeromycota and their sister lineages. In addition, no expansion of SSPs was observed in mycorrhizal as compared to non-mycorrhizal Mucoromycota, and together with large number of species-specific SSPs, indicates that genomic signatures of symbiotic lifestyle are different in Mucoromycota compared to Basidiomycota (28).
AM fungi form a monophyletic clade but their taxonomic rank is currently fluctuating in the literature. Classified either as phylum Glomeromycota (31, 32, 33) or as the sub-phylum Glomeromycotina that together with Mortierellomycotina and Mucoromycotina, comprise the Mucoromycota (34, 35, 36). The phylum Glomeromycota was first proposed based on early phylogenetic analysis of the rDNA genes that resolved the AM fungi as a monophyletic clade sister to Dikarya (37). On the other hand, the phylogenomic analysis based on conserved orthologous genes, could later resolve two monophyletic clades among the paraphyletic Zygomycota (38). With the ambition to recognize the minimum number of monophyletic phyla, the authors proposed Mucoromycota to encompass Glomeromycotina, Mucoromycotina and Mortierellomycotina, as sister to Dikarya (38). Other authors have argued that fungi should be classified into monophyletic phyla that are also informative of divergence time (33). Irrespective of taxonomic rank, phylogenomic analyses of the kingdom fungi resolves the three lineages as monophyletic clades based on Maximum Likelihood (ML) analysis with concatenated data and coalescence methods, but the evolutionary relationships among the three lineages remain enigmatic and a hard polytomy cannot be rejected (36). It is worth noting that the placement of Glomeromycota in the fungal tree of life remains sensitive to the evolutionary model used and filtering of fast evolving sites (39).
The three lineages share morphological characters such as coenocytic hyphae and predominantly plant-based ecologies with mycorrhizal associations in two of the sister lineages (38, 40). However, AM fungi stand out as a lineage of obligate symbionts of photosynthesizing partners, mostly plants (40) but also including G. pyriformis that associates with the nitrogen-fixing cyanobacteria Nostoc (41, 42). Contrary to expectations, phylogenomic analysis across Glomeromycota did not place G. pyriformis as the sister to all other AM fungi and comparative analyses suggest that the genome signature of obligate biotrophy characteristic of the group was already present in the MRCA of the symbiotic clade (30). The taxonomic classification of AM fungi to phylum Glomeromycota remains the most frequently used in mycorrhizal literature, and based on their unique biology we adhere to this classification and treated the three lineages as separate phyla that share a common ancestor and branch as sister to the Basidibolales, Zoopagomycota (36). Here Mucoromycota corresponds to Mucoromycotina (38), which encompass the mycorrhizal lineage Endogonaceae (28) and saprotrophic genera including Mucor, Rhizopus, Umbelopsis and others. It is relevant to note that Mortierellomycota, corresponding to Mortierellomycotina (38), have also been shown to encompass species with beneficial interactions with plants as root endophytes (43, 44). However, these fungi are not known to form specialized structures within plant roots and to our knowledge nutrient for energy exchange has not been documented. The aim of this study is to address the evolutionary relationships among extant representatives of the three fungal lineages and to analyze derived genome signature of the obligate mutualistic AM fungi. To obtain a clear picture of the early evolutionary relationships among the three phyla we carefully selected a balanced taxon sampling from all three lineages, primarily including taxa known to inhabit soil environments, in order to minimize confounding effects of adaptations to different habitats. With the selected dataset we infer the evolutionary relationships among Glomeromycota, Mortierellomycota and Mucoromycota, and examine gene family evolution for CAZymes and peptidases in order to identify coarse genomic signatures associated with the symbiotic lifestyle of AM fungi and the Endogonaceae. We explore the distribution of previously identified Missing Glomeromycota Core Genes (MGCGs) across analyzed taxa and highlight interesting differences in gene content across different AM fungal lineages.