Tackling fungal diversity in lichen symbioses: molecular and morphological data recognize new lineages in Chaetothyriales (Eurotiomycetes, Ascomycota)

Lichens have been reappraised as self-sustaining and long-living ecosystems in which a multiplicity of microorganisms are housed, in addition to the main symbiotic partners. Lichen-associated microfungi can frequently occur cryptically, and their species diversity has recently been more fully elucidated by DNA metabarcoding studies and culture isolations. These lichen-associated fungi represent a wide array of major lineages in ascomycetes and basidiomycetes, including both filamentous and yeast species. Thanks to culture isolations, the morphology of a subset of the lichen-associated microfungal diversity has been studied. Metabarcoding analyses have shown high diversity of ascomycetous lichen-associated fungi in the two cosmopolitan rock-inhabiting lichens – Rhizoplaca melanophthalma and Tephromela atra – and many of these taxa were successfully isolated in culture. Based on DNA sequence data and morphological analyses, two new lineages within Chaetothyriales are here recognized. Both occur in lichens from dry habitats and are described here as the new species Cladophialophora endolichena Cometto, de Hoog, Muggia and Paracladophialophora lichenicola Cometto, de Hoog, Muggia. Other strains are placed in Pleostigmataceae, Trichomeriaceae, Pleosporales, Mycosphaerellales, Coniochaetales and Hypocreales, further filling gaps of knowledge of the high fungal diversity residing in lichen thalli.


Introduction
Lichens have evolved as a life form that develops a particular housing morphology through the interactions of a biotrophic fungus -the mycobiont -with one or more phototrophic organisms -the photobiont (Hawksworth and Honegger 1994). However, in contrast to a simple partnership between the myco-and photobionts, lichens have been reappraised as self-sustaining and long living ecosystems (Hawksworth and Grube 2020), in which a multiplicity of other microorganisms -including other filamentous and yeast microfungi, microalgae and bacteria -are housed Moya et al. 2017;Fernández-Mendoza et al. 2017;Banchi et al. 2018;Molins et al. 2018;Muggia and Grube 2018). The potential functional roles of these complementary microorganisms are a widely discussed subject of research, and they have not been clarified yet (e.g., Grube et al. 2009Grube et al. , 2015Moya et al. 2017;Molins et al. 2018;Muggia and Grube 2018;Spribille 2018;Hawksworth and Grube 2020). Furthermore, the overall diversity of these lichen inhabitants/ co-symbionts is still largely unknown. Ongoing research has shed light on the geographic and, only in part, ecological distributions of certain groups of microfungi and microalgae associated to lichens (Wang et al. 2016;Williams et al. 2017).
Microfungi identified in/on lichen symbioses were initially discovered in the early nineteenth century and have been the 1 3 53 Page 2 of 26 focus of a wide range of studies. They are commonly known as "lichenicolous fungi" for over a century, and more than 2300 species are formally recognized (Diederich et al. 2018). These lichen-associated microfungi are mainly represented by ascomycetes (95%), while only a small fraction appears to be basidiomycetous (Diederich 1996;Lawrey et al. 2007). Lichenicolous fungi usually develop symptoms of infections on their host lichen thalli, but many are asymptomatic, occurring mostly as resting spores or hyphal fragments in other lichen species (Arnold et al. 2009;U'Ren et al. 2010U'Ren et al. , 2012U'Ren et al. , 2014Muggia et al. 2016Muggia et al. , 2017Fernández-Mendoza et al. 2017;Banchi et al. 2018;Hafellner 2018). Most of the symptomatically occurring lichenicolous fungi show some level of host-specificity -and sometimes even dependency, on their lichen host (Lawrey and Diederich 2003;Hafellner 2018). However, the specificity observed for the lichenicolous fungi, has been shown in only a few cases for some of the cryptically occurring taxa (Smith et al. 2020). Furthermore, some studies have shown that certain abiotic factors, such as climate (U'Ren et al. 2012), seasonality, light exposure (Beck et al. 2014), altitude (Wang et al. 2016) and geographic distance (Zhang et al. 2015) may be crucial in shaping the lichenicolous fungal diversity of lichen thalli (Harutyunyan et al. 2008;Arnold et al. 2009;U'Ren et al. 2010;Lagarde et al. 2018;Yoshino et al. 2020;Cometto et al. 2022).
The lichenicolous fungi which cryptically occur in lichens belong mainly to the classes Eurotiomycetes and Dothideomycetes and are represented by filamentous or yeast-like melanised taxa, which are closely related to the polyphyletic lineages of rock-inhabiting fungi (RIF) and black yeasts Ruibal et al. 2009;Gostinčar et al. 2012;Quan et al. 2020). These have frequently been reported from epilithic lichens (Harutyunyan et al. 2008;Muggia et al. 2016Muggia et al. , 2017Muggia et al. , 2021Muggia and Grube 2018;Quan et al. 2020). Only a few taxa, such as lichenicolous Phoma species in Phaeosphaeriaceae (Dothideomycetes), were reported both cryptically and symptomatically occurring in epilithic, epiphytic and soil inhabiting lichens (Lawrey et al. 2012;Muggia et al. 2016Muggia et al. , 2017. Recently two lineages within Eurotiomycetes have been formally recognized from Alpine epilithic lichens as the family Pleostigmataceae and the new genus and species Melanina gunde-cimermaniae (Muggia et al. 2021). Interestingly, lichenicolous basidiomycetes yeast species that have been described from thalli of the lichen-forming mycobiont genera Cladonia, Lecanora and Letharia (Zamora et al. 2011;Millanes et al. 2011Millanes et al. , 2016Millanes et al. , 2021Černajová and Škaloud 2019;Tuovinen et al. 2019Tuovinen et al. , 2021 have been isolated recently from the epilithic lichens Rhizoplaca melanophthalma and Tephromela atra , suggesting that certain taxa of cryptic lichenicolous fungi can be unexpectedly widespread in lichens. In the present study, we deepen our investigation into the diversity of culturable, cryptically occurring lichenicolous ascomycetes from thalli of R. melanophthalma and T. atra, as these lichens revealed to be important sources for lichenicolous fungi (Muggia et al. 2016;Smith et al. 2020;Cometto et al. 2022). Here we report on the successful isolation of 131 ascomycetous strains, 39 of which represent two new lineages within Chaetothyriales and are formally described as new taxa, while the others belong to already known genera and families of Dothideomycetes, Eurotiomycetes and Sordariomycetes. All together, these and earlier results highlight that certain species of cryptically occurring lichenicolous fungi are more frequent than previously thought and may have their realized ecological niches in lichen thalli.

Sampling
Lichens representing the R. melanophthalma aggregate (Leavitt et al. 2011) have an umbilicate thallus (attached at a single point), whereas lichens representing the T. atra group (Muggia et al. 2014) build a crustose thallus composed of adjacent areoles. Both lichens are characterized by a worldwide distribution and occur at different ecological conditions and elevations. Also, the mycobiont and photobiont diversity of both lichens has been extensively investigated previously (Muggia et al. , 2010(Muggia et al. , 2014Leavitt et al. 2011Leavitt et al. , 2016De Carolis et al. 2022). For the present study, lichen samples were collected in 23 different localities on different rock types (i.e., schist-arenaria, siliceous, acidic, granitic and calcareous rocks) and at altitudes ranging from 550 to 5100 m above sea level (a.s.l.). The sampling was performed in North America (Utah, Nevada and Idaho), South America (Argentina and Chile), Europe (Spain), Mauritius and Tasmania. In total 20 populations of R. melanophthalma and five populations of T. atra were analysed (Supplementary  Table S1). All the lichen samples were deposited at the herbarium of the University of Trieste (TSB).

Culture isolation
Fungal isolation was performed from four thalli for each population of R. melanophthalma and T. atra following the protocol of Yamamoto et al. (2002). Approximately 2 mm 2 fragments of lichen thalli were dissected with a sterile razor blade. For R. melanophthalma, one marginal lobe and one apothecium were taken, while for T. atra, one marginal areole and one apothecium. The fragments were washed three times for 15 minutes with sterile water, followed by 30 minutes of washing with 500 μl of Tween80 diluted 1:10, and a final washing step of 15 minutes for three times with sterile water. The clean fragments were ground in sterile water under the hood and tiny thallus fragments were picked with a sterile bamboo stick and transferred into agar tubes. Six different media were used to promote the growth of as many different fungi as possible: Lilly and Barnett (LB, Lilly and Barnett 1951), Trebouxia medium (TM, Ahmadjian 1987), Potato Dextrose agar (PDA, ApplChem A5828), Sabouraud's glucose agar base medium (SAB, Pagano et al. 1958), Dichloran/Glycerol agar (DG18, Hocking and Pitt 1980) and Malt Yeast-extract (MY, Lilly and Barnett 1951). Two replicates for each medium were inoculated for a total of 12 inocula from each lichen individual, and incubated in growing chamber (17 °C, 20 μmol × photons m −2 × s −1 , with a light/ dark cycle of 14/10 h). When the inocula developed into a mycelium mass of about 5 mm size (after about three to six months), they were sub-cultured into Petri plates, on the same medium of the original tube.

Molecular analyses: DNA extraction, PCR amplification and sequencing
Small parts of the cultured fungal colonies were taken and put into 1.5 ml reaction tubes, containing three sterile tungsten beads for homogenization, frozen and ground using a TissueLyserII (Retsch). The DNA extractions were performed following the CTAB protocol of Cubero et al. (1999), with minor adjustments. The identity of all fungal strains was studied with sequences of the nuclear internal transcribed spacers (nucITS) and 5.8S rDNA ribosomal gene amplified with the primers ITS1F (Bruns and Gardes 1993) and ITS4 (White et al. 1990). If ITS sequences were identical (99%-100% identity) among strains sharing the same origin -i.e., isolated from the same lichen thallus, or from thalli coming from the same population -for only a single strain the D1/D2 domain of the 28S nuclear large ribosomal subunit (nucLSU) was further amplified with the primers LR0R and LR5 (Vilgalys and Hester 1990; http:// www. biolo gy. duke. edu/ fungi/ mycol ab/ prime rs. htm). Polymerase chain reactions (PCR) were prepared for a 25 μl final volume containing 5 μl DNA, 12.5 μl of AccuStart II PCR ToughMix, 0.4 μl for each of the 10 μM primers. Negative control reactions were always used to check for potential contamination. PCR amplifications followed the same conditions of previous studies (Muggia et al. 2017;Cometto et al. 2022). All the amplicons were checked for their quality and size by 1% agarose gel electrophoresis stained with Green Safe Gel (Sigma-Aldrich) and purified using Mag-Bind® Normalizer Kit (Omega bio-tek). Clean amplicons were sent for Sanger sequencing to Macrogen Europe (The Netherlands).

Phylogenetic analyses
The identity of the newly generated sequences was checked with sequences available in the GenBank database by BLAST similarity search (Altschul et al. 1990). Taxa that matched our sequences with an identity value higher than 97% were selected for the phylogenetic analyses. As our sequences showed high similarity with representatives of the classes Eurotiomycetes (particularly the order Chaetothyriales), Dothideomycetes and Sordariomycetes, three separate datasets representing each individual fungal class were prepared (Supplementary Table S2 -S4). The total dataset included the widest spectrum of taxon diversity, as also genera and families closely related to our sequences were selected from previous studies. In particular, the Eurotiomycetes dataset was based on Harutyunyan et al. (2008), Muggia et al. (2016Muggia et al. ( , 2019Muggia et al. ( , 2021 and Quan et al. (2020); that of Dothideomycetes was based on Muggia et al. (2016) and Ametrano et al. (2019); that of Sordariomycetes was based on Muggia et al. (2016). Sequence alignments for each locus (nucITS and nucLSU) and for each fungal class (Eurotiomycetes, Dothideomycetes, Sordariomycetes) were prepared with MAFFT v.7 (Katoh and Standley 2013) using the g-ins-I alignment strategy. Ambiguously aligned positions and introns were removed from the alignments using Trimmomatic (Bolger et al. 2014). Single locus phylogenies were inferred with Maximum Likelihood (ML) and Bayesian Inference (BI) approaches. RAxML v.8.2 (Stamatakis 2014) was used for the ML analysis applying GTRGAMMA substitution model and 1000 bootstrap pseudoreplicates. The BI analysis was carried out with the program BEAST v.2.6.7 (Bouckaert et al. 2014) running GTRGAMMA substitution model for 100 million generations. The results were analysed using the program Tracer v1.7.2 (Rambaut et al. 2018) to check the runs for convergence (burn-in = 10%). TreeAnnotator (included in the BEAST package) was used to summarize the trees in a consensus tree representing the posterior distribution, the first 10% of data were discarded as burn-in. After checking the phylogenetic concordance of the nucITS and nucLSU loci, they were concatenated with MEGA (Kumar et al. 2018) and then analysed with both RAxML and BEAST with the same settings of the single locus analyses.
To better clarify the phylogenetic placement of two potentially new lineages recognized within Chaetothyriales we performed a further analysis based on a more extended taxon sampling of Herpotrichiellaceae and Paracladophialophoraceae (Supplementary Table S5), selecting six species of Cyphellophoraceae as outgroups (Cyphellophora eucalypti, C. laciniata, C. olivacea, C. pluriseptata, C. sessilis and Phialophora attae), and running the phylogenetic analysis using RAxML and BEAST approaches with the same settings explained above.
The phylogenetic trees were visualized using ITOL (Letunic and Bork 2019).

Morphological analysis
Morphological and anatomical characters of 10-month to one-year old cultured fungal strains were analysed with light microscopy considering the following characters: form of growth, melanisation, thickness and branching of the hyphae. A tiny part of the colony was removed by a sterile inoculation loop and mounted in water. Digital photos at light microscope were taken with a Zeiss AXIO Imager A2 coupled to a Thorlabs digital camera. The photos were adjusted for colour saturation and sharpness with Adobe Photoshop 7.0 (Adobe System Incorporated, San Jose, CA, USA) and photo-tables were assembled using Inkskape (www. inksc ape. org).

Culture isolation
A total of 131 fungal strains were isolated and characterized from 80 R. melanophthalma and 20 T. atra samples (Fig. 1a, b and Table 1; Supplementary Table S1): 65 strains belong to Eurotiomycetes (order Chaetothyriales), 53 strains to Dothideomycetes and 13 strains to Sordariomycetes. The Eurotiomycetes (order Chaetothyriales) strains were isolated from 20 samples of R. melanophthalma collected in 10 localities across South America (Argentina), North America (Utah) and Europe (Spain) and from eight samples T. atra collected in five localities across South America (Chile), Tasmania, Mauritius and Europe (Spain). The Dothideomycetes strains were isolated from 20 samples of R. melanophthalma collected in 13 localities across South America (Argentina and Chile), North America (Utah, Idaho and Nevada) and from seven samples T. atra collected in three localities in South America (Chile) and Europe (Spain). The Sordariomycetes strains were isolated from four samples of R. melanophthalma collected in two localities in South America (Argentina) and North America (Utah) and from two samples of T. atra collected in two localities in Tasmania and Mauritius. No fungal strains were obtained from R. melanophthalma thalli coming from three populations in Argentina above 3600 m a.s.l. and from one population in Spain at 2080 m a.s.l.. Only six strains belonging to Phaeosphaeriaceae sp., Elasticomyces sp., Knufia sp., Pleostigmataceae sp. and Hyalotiella sp. were isolated from lichen thalli collected over 3000 m a.s.l. in South America. Six different Eurotiomycetes taxa (identified as Pleostigma sp., Muelerella sp. and two new lineages in Chaetothyriales) and one Dothideomycetes taxon (belonging to Phaeosphaeriaceae) were isolated from both R. melanophthalma and T. atra.      Eurotiomycetes (order Chaetothyriales) strains were unable to grow on DG18 media, while all the other fungal strains grew on the six different culture media ( Fig. 1c and Table 1). Eurotiomycetes mainly grew on MY media; Dothideomycetes on LBM and DG18 media; Sordariomycetes on SAB and MY media. In general, MY and LBM were the most suitable media for the isolation of ascomycetes taxa, as 49% of the isolates grew well on them ( Fig. 1c and Table 1).

Phylogenetic and morphological analysis
A total of 131 new nucITS and 68 new nucLSU fungal sequences were generated (Table 1). Phylogenetic analyses were performed individually for each taxonomic class -Eurotiomycetes, Dothideomycetes and Sordariomycetes -using the concatenated two-locus datasets (Supplementary  Material Tables S2-S5). Maximum Likelihood and Bayesian phylogenetic inference were highly concordant; most of the clades were supported and topologically congruent with previous studies (Harutyunyan et al. 2008;Ametrano et al. 2019;Muggia et al. 2016Muggia et al. , 2019Muggia et al. , 2021Quan et al. 2020).
The 'Muellerella + Lichenodiplis' clade (sensu Muggia et al. 2015Muggia et al. , 2019 and the family Epibryaceae were found as the most basal lineages in Chaetothyriales. In this clade we found one strain (L3774) isolated from T. atra collected in Tasmania, and the strain Chaetothyriales sp. Pet 5a (Vasse et al. 2017). The strain L3774 was characterized by heavy melanized elliptical and elongated cells (4 × 8 μm) often constricted at the septa and with rare ramification (Fig. 3e) and conidia-like cells (4 up to 7 μm diameter; Fig. 3f).
In the Melanina gunde-cimermaniae clade, five (L2604, L2605, L2618, L2628 and L2876) isolates from R. melanophthalma collected in South America (Argentina) and two isolates from T. atra collected in Europe (Spain) were found together with other two specimen Capronia sp. 97003b and Capronia sp. 97003a. Our isolates were characterized by a dark grey to black mycelium (Fig. 3s) composed by toruloid hyphae and filaments of conidia (4 up to 10 μm diameter; Figs. 3t-v).  (Wijayawardene et al. 2020) and to a monophyletic lineage that was identified here for the first time and we referred to it as the new species "Paracladophialophora lichenicola" (see below and Fig. 4). The two isolates L3233 and L3784 were characterized by heavy or slight melanized hyphae in which cylindrical and rectangular cells (4 × 10-15 μm; Fig. 3ac) were intercalated to spherical cells (5 to 15 μm diameter; Fig. 3ad).
The remaining 20 strains, isolated from both R. melanophthalma collected in North and South America (Utah and Argentina, respectively) and from T. atra collected in Tasmania, formed a separate monophyletic lineage with further Cladophialophora strains (Figs. 2 and 4) isolated from lichens in previous studies (Harutyunyan et al. 2008;Muggia et al. 2016Muggia et al. , 2017Muggia et al. , 2021, i.e. Cladophialophora sp. Sh8, A1044 and A1069. We recognize this second new lineage as the new species Cladophialophora endolichena (see below). The phylogenetic position of the two new species Paracladophialophora lichenicola and Cladophialophora endolichena has been further confirmed by the extended analyses of Fig. 4.
The detailed morphological descriptions of both Paracladophialophora lichenicola and Cladophialophora endolichena are presented below in the Taxonomy section.
Dothideomycetes (Figs. 5, 6 and Table 1; Supplementary Material Table S4) -Fifty-three new strains were found in Dothideomycetes. The strain L3077, isolated from R. melanophthalma collected in North America (Idaho), was placed on an own branch basal to Dothideomycetes, as well as Lichenothelia papilliformis (Ametrano et al. 2019), Lichenostigmatales sp. A930 (Muggia et al. 2016) and Fungal sp. TRN529 (Ruibal et al. 2009). L3077 was morphologically similar to Lichenostigmatales sp. A930, with a yeastlike black mycelium (Fig. 6a), budding hyphae and melanized cells (3 up to 20 μm diameter; Figs. 6b, c). Most strains belonged mainly to the orders Pleosporales and Mycosphaerellales. Twenty-four strains were placed in Pleosporales and belonged to five family level lineages highly supported and fully resolved. In particular, the strain L2869, isolated from R. melanophthalma collected in South America (Argentina), was placed in Lophiotremataceae and was characterized by hyaline hyphae (3 μm diameter; not shown). The strain L3091, isolated from T. atra collected in Tasmania, was placed in Teichosporaceae and was closely related to an uncultured fungus B3_1986 (found by Vázquez-Nion et al. 2016) and characterized by a grey mycelium with brown margin with septate hyphae (4 × 15 μm; not shown). The strain L2868, isolated from T. atra from Tasmania, was placed in the well-supported clade of Paraphaeosphaeria michotii and it was characterized by hyaline hyphae (2 μm diameter; not shown).
The strain L3036, isolated from R. melanophthalma collected in North America, was nested in a clade with the Pleosporales sp. The strain L3021, isolated from T. atra collected in Tasmania, was placed in the family Didymellaceae, here represented by Phoma herbarum, Ampelomyces sp. and Didymella spp. This strain was characterized by a white to grey mycelium (Fig. 6d) with slight melanized hyphae (5 × 15 μm; Fig. 6e).
The strain L3078, isolated from R. melanophthalma collected in North America was placed in the family Pleosporaceae next to Pleospora spp. and Comoclathris lini. This strain was characterized by a grey mycelium with brown margin and septate hyphae (4 × 15 μm; not shown).
Twenty-two strains were found in the order Mycosphaerellales (Abdollahzadeh et al. 2020) and belonged to Mycosphaerellaceae and Teratosphaeriaceae. The strain L3747 isolated from R. melanophthalma collected in South America (Argentina) was genetically identical to Ramularia vizellae in Mycosphaerellaceae. Within the Teratosphaeriaceae, instead, the newly isolated strains were placed in four separated clades. The strain L2879, isolated from R. melanophthalma collected in South America (Argentina) was close related to Teratosphaeriaceae sp. CPC 12419 (Crous et al. 2008) and closely related to Saxomyces penninicus and Teratosphaeria parva. This strain was characterized by a dark black mycelium (Fig. 6p) composed of melanized and hyaline hyphae with rectangular cells (4 × 10 μm) from which ramification started. Filaments of isodiametric, conidia-like cells (7 μm diameter) with apical cell developing into hyphae were observed (Fig. 6q). The strain L3239, isolated from R. melanophthalma collected in North America (Utah) was placed in a supported clade together with three unknown fungi labelled as sp. agrD231, agrD244 and agrD242 (Peršoh and Rambold 2012). This strain was characterized by a blackish mycelium composed by heavy melanized hyphae with rectangular cell (5 × 15 μm) from which ramification started (data not shown). The strain L3270, isolated from R. melanophthalma collected in South America (Argentina) was close related to three unidentified fungi Dothideomycetes sp. AK1125 (U'Ren et al. 2012), Dothideomycetes sp. PIMO_109 and fungal sp. PIMO_21 (Larkin et al. 2012). Lastly, 14 strains, isolated from R. melanophthalma collected in South America (Argentina and Chile) and North America (Utah and Idaho), built a separate clade together with Elasticomyces elasticus and four still undetermined fungal samples [i.e. Dothideomycetes sp. s_C03_05.ab (Amend et al. 2010), Dothideomycetes sp. PIMO_446 (Larkin et al. 2012) and two uncultured fungi (127_NA4_P32_L9 and FunN4_01B; Timling et al. 2014;Nemergut et al. 2008)]. These strains likely correspond to Elasticomyces elasticus and were characterized by blackish to greenish mycelia with irregular margin (Fig. 6r) and heavy melanized hyphae composed by cylindrical and rectangular cells (5 × 8-12 μm) with ramification (Figs. 6s, t).
Six strains (strains L2856, L2857, L2890, L3049, L3776 and L3817), isolated from T. atra collected in Europe Maximum Likelihood analysis based on the concatenated nuclear ITS-LSU dataset; branches in bold denote RAxML bootstrap support ≥75%; Bayesian posterior probabilities ≥0.8 are reported above branches. Newly obtained sequences are in bold and reported in the same line when they were isolated from the same lichen thallus. Symbols and colours indicate the different lichen hosts and the geographic origin from where the strains were isolated, respectively. The newly identified lineages are highlighted in grey ◂ (Spain), were placed together with two Dothideomycetes sp., i.e. A931 and A552, isolated from lichens (Muggia et al. 2016) in a clade closely related to Venturiales, Lichenocloniales and Abrothallales. These strains had a pink mycelium with regular margin (Fig. 6m) composed by hyaline hyphae built by rectangular cells (2 × 15 μm) from which branches generated (Figs. 6n, o).
Mycobank: MB 848887. Etymology: residing inside lichens. Holotype: L3074, cultured strain, preserved in metabolically inactive state in MY medium (September 2020, date at which they were first identified in culture isolation), isolated from the thallus of R. melanophthalma (L2638).

Discussion
The culture approach applied in the present study captured a great diversity of the microfungi associated with two cosmopolitan, epilithic lichens -R. melanophthalma and T. atra -collected in the broad range of their distribution and including sites characterized by harsh environmental conditions. Here, we detected both already known fungal lineages and two new lineages, i.e. the species Cladophialophora endolichena sp. nov. and Paracladophialophora lichenicola sp. nov., which seem to be recurrently associated to lichen thalli. Indeed, many strains belonging to the two new lineages were previously detected in local communities of alpine (Fleischhacker et al. 2015;Muggia et al. 2016Muggia et al. , 2017 and Mediterranean lichens (Harutyunyan et al. 2008). Although fungi can adapt to diverse environments and develop different lifestyles and evolutionary strategies, some lineages seem to preferentially reside in lichen thalli and might develop a certain specificity to their hosts or, more in general, to the micro conditions that the symbiotic lichen thalli offer.
Phylogenetic relationships of the 131 new fungal isolates were inferred, among which 65 belong to Eurotiomycetes, 53 to Dothideomycetes and 13 to Sordariomycetes. Furthermore, isolates representing most of the phylogenetic lineages were morphologically characterized. We found that these microfungi grew on six different culture media, and the percentage of their growth success corresponded well to that reported by Muggia et al. (2017). We obtained fungal isolates from almost all the lichen samples, with the exception of those thalli of R. melanophthalma collected in three localities in the Argentinian Andes at altitudes ranging from 3600 to 5100 m a.s.l., and in Spain at 2080 m a.s.l.. The fact that no fungal strains could be isolated from these lichen thalli may be explained by the selective constraints in which the lichen grew, that were not simulated in the culture conditions. Thus, the culture conditions applied seem to have not favoured in vitro the development of mycelia, particularly of those fungi that would be more extremophilic. In fact, the inocula were incubated at 17 °C and constant humidity in growth chambers, while the original thalli experienced harsher conditions of drought stress and radiation in the environments. An alternative explanation may be found in the fact that the original lichen thalli presented a lower and different mycobiome diversity in the DNA metabarcoding analyses , which would likely support the lack of culturable strains within the species-poor pool of fungi. Extremophilic fungi, indeed, grow extremely slowly in vitro, if not at all, and their isolation often demands specific requirements and extended amounts of incubation time (Urzı and De Leo 2001;Selbmann et al. 2014).
Our data show that most of the isolated strains are members of the Eurotiomycetes. Within this class, the majority of microfungi belong to families in Chaetothyriales, an Fig. 7 Phylogenetic inference of Sordariomycetes: Maximum Likelihood analysis based on the concatenated nuclear ITS-LSU dataset; branches in bold denote RAxML bootstrap support ≥75%; Bayesian posterior probabilities ≥0.8 are reported above branches. Newly obtained sequences are in bold and reported in the same line when they were isolated from the same lichen thallus. Symbols and colours indicate the different lichen hosts and the geographic origin from where the strains were isolated, respectively order which also includes many saprophytic and opportunistic pathogens on humans and cold-blooded vertebrates (de Hoog et al. 2011;Teixeira et al. 2017;Quan et al. 2020). The newly isolated strains share the characteristic traits of the melanised polyextremotolerant fungi already described for rock-inhabiting fungi, pathogens and lichen-associated lineages (Gostinčar et al. 2012(Gostinčar et al. , 2018. Indeed, three strains represent Pleostigma species, one Muellerella and another seven correspond to the recently described species Melanina gunde-cimermaniae. Muellerella is a well-known genus of symptomatic lichenicolous fungi which was successfully isolated from ascospores and conidia but occurs also cryptically in lichens (Muggia et al. 2015(Muggia et al. , 2019(Muggia et al. , 2021. The strain L3774 which is in the clade 'Muellerella + Lichenodiplis' sensu Muggia et al. (2015Muggia et al. ( , 2019 was indeed isolated from a thallus of T. atra (which is the original lichen host species of Muellerella atricola and Lichenodiplis lecanorae; Atienza et al. 2009;Muggia et al. 2015Muggia et al. , 2019, but morphological inspections confirmed that this thallus is devoid of any perithecia or sporodochia, thus supporting the cryptic occurrence of the Muellerella fungus. The genera Pleostigma -together with the corresponding family Pleostigmataceae -and Melanina were recently described by Muggia et al. (2021) to allocate black fungal strains that were isolated from alpine lichens (i.e. Aspicilia caesiocinerea, A. simoensis, Aspicilia sp., Lecanora intricata, L. polytropa, Lecidea lapicida, Lecidea sp., Rhizocarpon geographicum, Schaereria fuscocinerea and Umbilicaria cylindrica; Muggia et al. 2021) and a few fungi that were identified by Ruibal et al. (2009) from calcareous rocks in the Mediterranean basin. Finding here both Pleostigma spp. and Melanina also in thalli collected at diverse altitudes and across a worldwide geographic range supports their specific endolichenic lifestyle and likely their ubiquitous distribution in lichens.
The most important outcome is that the present research conducted on such a broad scale allowed us to recognize additionally two new lineages of lichen-associated fungi in Chaetothyriales. Our phylogenetic inferences are topologically  Crous et al. (2016Crous et al. ( , 2018, Quan et al. (2020) and Wijayawardene et al. (2020). The new specie level lineage of Paracladophialophora lichenicola sp. nov. is included in Paracladophialophoraceae, being closely related to Paracladophialophora cyperacearum and P. carceris (Fig. 2, S1). Paracladophialophora lichenicola sp. nov. is well represented by multiple strains isolated here form both R. melanophthalma and T. atra (n. 15), but also previously from other lichens collected in the Mediterranean basin and Armenia (i.e., Caloplaca gomerana, L. stigmatea, Protoparmeliopsis muralis, Umbilicaria virginis and U. vellea;Harutyunyan et al. 2008). Members of this lineage occur at different altitudes and on different rock substrates (basalt, calcareous and siliceous). Harutyunyan et al. (2008) assigned the first isolated strains to the genus Rhinocladiella because at that time only a few sequences were available for comparison. Later Rhinocladiella was included in Herpotrichiellaceae (Teixeira et al. 2017), while the still undescribed strains provisionally maintained this genus name. Also at that time, Harutyunyan et al. (2008) suggested that these strains could be facultative lichen colonisers because of the absence of a specific lichen host and the infection symptoms. However, our results strengthen the idea that Paracladophialophora lichenicola has developed a certain preference for the lichen-associated lifestyle, thus presenting additional data to support its formal recognition.
The second new lineage -here described as the new species Cladophialophora endolichena -is placed within Herpotrichiellaceae, and its topology is mostly congruent with that of Muggia et al. (2017Muggia et al. ( , 2021. In addition to our 26 strains isolated from both R. melanophthalma and T. atra, this new clade includes other Cladophialophora samples previously isolated from other lichen species (Gyalolechia fulgida, L. polytropa, P. muralis and Rusavskia elegans; Harutyunyan et al. 2008;Muggia et al. 2016Muggia et al. , 2017. Herpotrichiellaceae is the largest family in Chaetothyriales which includes ecologically very diverse fungi (e.g., human opportunists, rock-inhabiting fungi and lichenicolous fungi; Crous et al. 2007;Quan et al. 2020) and our results are not unexpected -the lichen-associated lifestyle is ancestral in this order (Quan et al. 2020, Quan et al. 2023. Interestingly, the lichen associated Herpotrichiellaceae are mainly found to be likely Cladophialophora species, a genus known to be involved in the aromatic hydrocarbon degradation (Prenafeta-Boldú et al. 2006;Badali et al. 2008) and supposed to be able also to take benefits from the secondary metabolites found in lichens (Harutyunyan et al. 2008). The strains of Cladophialophora endolichena are strictly lichen-associated fungi, while the closest related species are C. chaetospira, C. nyingchiensis and C. tengchongensis, the latter two recently described as microcolonial melanised rock-inhabiting fungi (Sun et al. 2020).
Significant diversity of lichen-associated microfungi is also found in Dothideomycetes, although here the isolated strains are phylogenetically more heterogeneous. In Dothideomycetes, our samples were related to fungi with different lifestyles (Schoch et al. 2009;Hyde et al. 2013;Wijayawardene et al. 2014), as well as lichenicolous fungi such as Phoma caloplacae and P. cladonicola (Lawrey et al. 2012). Interestingly, most of the isolates retrieved in Dothideomycetes are related to fungi previously isolated from lichens or to rockinhabiting fungi from extreme environments, such as Elasticomyces elasticus or Saxomyces penninicus (Selbmann et al. 2008(Selbmann et al. , 2013Ruibal et al. 2011;Muggia et al. 2016Muggia et al. , 2017. All the new strains placed in Teratosphaeriaceae were isolated from R. melanophthalma collected in relatively high altitude (about 2000-4300 m a.s.l.) in Argentina, Chile and Utah, allowing a hypothesis that their choice to live endolichenically enhances protection from the high irradiation and the continuous fluctuation of temperature that characterize high-altitude mountain environments. However, one exception is the strain L3239, isolated from R. melanophthalma which (in Teratosphaeriaceae) forms a clade together with fungi isolated from lichens in the association of Letharietum vulpinae (Peršoh and Rambold 2012). Four strains are nested with Phoma cladoniicola, P. caloplacae and many other Phoma samples isolated from alpine epilithic lichens (Muggia et al. 2016). This result strengthens the hypothesis of Lawrey et al. (2012), who suggested that the same Phoma species can be isolated from a variety of lichens, in contrast to a previous assumption that Phoma species were highly selective for their hosts (Hawksworth and Cole 2004;Diederich et al. 2007).
In Sordariomycetes, some new microfungal strains also correspond to previously isolated fungi from lichens, in particular within Coniochaetales, Hypocreales and Xylariales. Five strains isolated from T. atra collected in Tasmania and Mauritius are related to Tolypocladium sp. and Elaphocordyceps sp. isolated from other lichen species (Jiang et al. 2015;Yang et al. 2022). Interestingly, Tolypocladium is a genus of fungicolous fungi (Sun et al. 2019), and its presence in lichens may hint to his parasitic behaviour towards the lichen mycobiont. Furthermore, we also recovered two strains related to Microcera physciae, recently isolated and described from the lichen Physcia tenella (Crous et al. 2021).
In conclusion, lichens are long-lived symbiotic systems and serve as suitable niches for many cryptically occurring fungi, some of which have likely specialized to them, finding in these systems a kind of protection and realized niche. These fungi do not germinate further and do not exit the lichen thalli when environmental conditions are too harsh and unfavourable, remaining even undetectable by morphological inspections. However, most of these fungi, that do not readily grow in the natural habitat, may start to grow when isolated as axenic culture under suitable culture conditions. The remaining unculturable/uncultivated fraction of lichenicolous fungi seems to demand further efforts to be evidenced and morphologically studied.