Lichen yeasts show structured host choice, not thallus dependence

Amplicon sequencing of 169 lichen specimens revealed that cystobasidiomycete yeasts were present in about 35% of the samples, exposing previously undocumented diversity among Arthoniales but showing no general association between yeast occurrence and thallus architecture. The study recovered 21 operational taxonomic units, or OTUs, from Arthoniales lichens, with eight OTUs assigned to three new lineages. Although the fungi showed preferences for particular lichen families, genera and species, their occurrence was uneven and was not explained consistently by growth form or the presence of a cortex.

The samples were collected primarily from Arthoniomycetes in Belgium and Italy during 2022 and 2023, with additional archived material from France, Portugal, Crete, Madeira, the Canary Islands, Cabo Verde, Madagascar, Guadeloupe and Curaçao. Of the 169 specimens, 133 belonged to Arthoniomycetes and 36 to other classes. The collection deliberately included crustose, foliose and fruticose lichens, with corticate and ecorticate forms growing on bark or rock.

This design was important because Arthoniales contain lineages that repeatedly acquired or lost a cortex and evolved different thallus forms, making the order a useful system for testing whether lichen-associated fungi contribute to structural development. Researchers extracted DNA from each specimen, amplified the internal transcribed spacer region of the fungal ribosomal RNA gene and used Oxford Nanopore MinION sequencing to characterize the fungal communities. General primers were chosen instead of Cystobasidiomycetes-specific primers so that the full fungal microbiome could be screened, although the reported analysis focused on cystobasidiomycetes.

Reads were quality filtered, assembled into consensus sequences and compared with GenBank. The resulting worldwide analysis combined 79 sequences generated in the study with 2,692 selected GenBank sequences, producing a data set of 2,771 sequences used to infer a maximum-likelihood phylogeny. Researchers used similarity thresholds of 97%, 98% and 99% to identify candidate OTUs, then curated those groups using phylogenetic relationships, available type sequences and taxonomic information.

They ultimately distinguished 376 consensus OTUs, 137 of which contained lichen-derived sequences. Among those, 23 contained Arthoniales sequences, with 21 newly detected in the study and two taken from earlier GenBank records from Enterographa pallidella and Opegrapha vulgata. Six consensus groups remained paraphyletic.

The authors cautioned that the ITS-only phylogeny showed some apparent non-monophyly, including within Cyphobasidiales and Cystobasidiales, that may reflect limitations of the marker rather than the accepted classification based on broader genomic data. Cystobasidiomycetes were detected in 60 specimens, including 47 Arthoniales lichens and 13 lichens from other groups, and were assigned to 39 OTUs. Individual lichen specimens contained one to three OTUs, while read abundance ranged from 10 to 105,854.

Some specimens carried only a single OTU, whereas two held three apiece. One Usnea esperantiana sample produced more than 100,000 cystobasidiomycete reads, and four other specimens yielded between 3,000 and 10,000. These differences show that detection of a fungal group is not equivalent to its regular or uniform presence within a lichen species.

The global comparison demonstrated a pronounced separation between sequences recovered from lichens and those obtained from other sources. Lichen-associated OTUs occurred mainly in Cyphobasidiales and Microsporomycetaceae, with smaller numbers in Cystobasidiales, Buckleyzymales, Symmetrosporaceae and groups of uncertain classification; none appeared in Erythrobasidiales or Sakaguchiales. Most of the 137 lichen-associated OTUs also occurred in Lecanorales, whereas 17% included Arthoniales sequences and 10% included Caliciales.

The authors noted that some mixed groups probably represent true generalists, but others may result from airborne cells or environmental fungi deposited on a lichen surface and detected because the thalli were not cleaned before analysis. Network analyses indicated that the relationships were neither random nor uniformly host-specific. Some OTUs occurred with mycobionts from several lichen orders, while others were selective at order, family, genus or species level.

Cyphobasidium hypogymniicola and Occultifur pisutianus were associated predominantly with Lecanorales, while “Cyphobasidiales 8” was associated with 18 Arthoniales sequences, nearly all from the family Roccellaceae. Other OTUs were concentrated on particular genera, including Bryoria, Cladonia, Opegrapha, Syncesia and Evernia. The reciprocal pattern was also evident: some lichenized fungi consistently carried one group of cystobasidiomycetes, while others were linked to several.

The researchers therefore described the overall pattern as structured but flexible partner choice, with preferred hosts and occasional secondary associations rather than strict specificity or complete generalism. Thallus architecture did not emerge as a consistent explanation. OTUs in the “Cyphobasidiales 3–8” group were recovered from lichens with several forms, including Enterographa, Dendrographa, Dirina, Roccella and Roccellographa, and these lineages collectively included crustose, corticate, fruticose and ecorticate hosts.

Some OTUs appeared restricted to particular morphologies, including those associated with fruticose Bryoria and Cladonia or crustose Opegrapha and Syncesia, but rarefaction curves indicated that sampling was insufficient to determine whether those patterns were genuine. Moreover, even a restricted association would not demonstrate that a yeast produces cortex tissue. The absence of a detectable relationship supports a lack of evidence for a general cortex-building role, rather than proving that every lineage lacks specialized functions.

Variation partitioning attributed 17% of the variance across the full data set to host identity and 10% to host and substrate considered together, while geography accounted for 6%. Climate and substrate alone contributed approximately 1% and 0%, respectively, and 38% remained unexplained. In analyses restricted to lichen-associated fungi, host identity became more influential as the taxonomic resolution increased, explaining 24% of variation in one analysis and 33% in the Cyphobasidiales analysis.

Geographic effects ranged from 5% to 19%, depending on the data set. These environmental results remain uncertain because the records were concentrated in continental and oceanic climatic zones and on bark, with only 13% of OTUs from rock and very few from foliicolous lichens or tropical environments. Sampling and detection methods further constrain the conclusions.

Rarefaction curves did not reach full coverage for Arthoniales or other lichen orders, suggesting that many more associations remain to be discovered. The authors also note that the surfaces of the thalli were not cleaned before DNA extraction, creating a risk that airborne or surface-associated cells would be mistaken for internal lichen associates. Differences in thallus thickness, from very thin structures such as those of Opegrapha to thicker structures such as those of Dirina, may also affect detection probability.

The study therefore expands the known diversity of Arthoniales-associated cystobasidiomycetes and supports structured host associations, but it does not establish mutualistic or structural functions. Broader geographic, substrate and method comparisons, including tropical, saxicolous, terricolous and foliicolous lichens, will be needed to test the limits of the pattern.

Subject of Research: Host-associated diversity and ecological distribution of cystobasidiomycete yeasts in Arthoniales lichens

Article Title: Amplicon sequencing reveals host-associated lineages of Cystobasidiomycetes ( Basidiomycota , Pucciniomycotina ) in Arthoniales ( Ascomycota , Arthoniomycetes ), but no association with thallus architecture

Article References: Borgato, L., Asselman, P., Ertz, D., Amalfi, M., & Verbeken, A. (2026). Amplicon sequencing reveals host-associated lineages of Cystobasidiomycetes (Basidiomycota, Pucciniomycotina) in Arthoniales (Ascomycota, Arthoniomycetes), but no association with thallus architecture. IMA Fungus, 17, Article e201170. https://doi.org/10.3897/imafungus.17.201170

Image Credits: AI Generated

DOI: 10.3897/imafungus.17.201170

Keywords: Cystobasidiomycetes, Arthoniales, lichen microbiome, amplicon sequencing, ITS phylogeny, host specificity, thallus architecture, cortex, fungal ecology, Oxford Nanopore, Amplicon, sequencing