Insights into the Ecological Diversification of the Hymenochaetales based on Comparative Genomics and Phylogenomics With an Emphasis on Coltricia

Abstract To elucidate the genomic traits of ecological diversification in the Hymenochaetales, we sequenced 15 new genomes, with attention to ectomycorrhizal (EcM) Coltricia species. Together with published data, 32 genomes, including 31 Hymenochaetales and one outgroup, were comparatively analyzed in total. Compared with those of parasitic and saprophytic members, EcM species have significantly reduced number of plant cell wall degrading enzyme genes, and expanded transposable elements, genome sizes, small secreted proteins, and secreted proteases. EcM species still retain some of secreted carbohydrate-active enzymes (CAZymes) and have lost the key secreted CAZymes to degrade lignin and cellulose, while possess a strong capacity to degrade a microbial cell wall containing chitin and peptidoglycan. There were no significant differences in secreted CAZymes between fungi growing on gymnosperms and angiosperms, suggesting that the secreted CAZymes in the Hymenochaetales evolved before differentiation of host trees into gymnosperms and angiosperms. Nevertheless, parasitic and saprophytic species of the Hymenochaetales are very similar in many genome features, which reflect their close phylogenetic relationships both being white rot fungi. Phylogenomic and molecular clock analyses showed that the EcM genus Coltricia formed a clade located at the base of the Hymenochaetaceae and divergence time later than saprophytic species. And Coltricia remains one to two genes of AA2 family. These indicate that the ancestors of Coltricia appear to have originated from saprophytic ancestor with the ability to cause a white rot. This study provides new genomic data for EcM species and insights into the ecological diversification within the Hymenochaetales based on comparative genomics and phylogenomics analyses.


Introduction
Fungi are distributed worldwide in all ecosystems. It has been estimated that 2.2-3.8 species exist but only 140,000 have been described (Hawksworth and Lücking 2017;Naranjo-Ortiz and Gabaldón 2019;Wang et al. 2019). In forest ecosystems, fungi, such as wood decomposers, soil or litter saprotrophs, and ectomycorrhizal species, play a crucial role in the fluxes of nutrients, especially carbon (Heimann and Reichstein 2008;Floudas et al. 2012;Wu et al. 2022a). Among them, wood-decomposing fungi are classified as white rot and brown rot according to their decay modes (Krah et al. 2018). Furthermore, it has been inferred that ectomycorrhizal (EcM) form a symbiotic association with about 60% of trees on earth (Steidinger et al. 2019), mainly with species of Pinaceae, Fagaceae, Betulaceae, and Myrtaceae, which are widely distributed in various forest ecosystems (Whitham et al. 2008).
The Hymenochaetales, the core group of wood-inhabiting fungi, is a species-rich order in the Agaricomycetes, Basidiomycota, consisting of more than 1,100 species with a worldwide distribution (https://www. catalogueoflife.org; accessed on November 14, 2022). Species of Hymenochaetales are primarily saprophytic causing a white rot decay . Some, however, exhibit different life modes as pathogens, including important tree pathogens like Porodaedalea, Onnia, Coniferiporia, and Sanghuangporus (Wu et al. 2019Zhao et al. 2022b), or ectomycorrhizal symbionts, especially the genus Coltricia (Tedersoo et al. 2007;Wu et al. 2022b). Danielson (1984) identified the Coltricia perennis was EcM associates of pine. Tedersoo et al. (2007) morphologically illustrated four species of Coltricia formed ectomycorrhizae with Vateriopsis seychellarum, Intsia bijuga, and Eucalyptus robusta. And Korotkin et al. (2018) proved four samples of the genus Coltricia as EcM. With the genus Coltriciella treated as a synonym of Coltricia, 35 species were accepted, while many species grown on the ground of forests or decayed wood and lack of detailed studies in their lifestyle (Bian et al. 2022;Wu et al. 2022b). In addition, the genome of Coltricia has not been sequenced and explaining the symbiotic evolution.
With the reduction of genome sequencing costs and the rapidly increasing numbers of fungal genome data, more studies have focused on the innovation and evolution of fungal life modes using omics data analyses (Floudas et al. 2012;Sipos et al. 2017; Naranjo-Ortiz and Gabaldón 2019; Lebreton et al. 2021;Lofgren et al. 2021;Sun et al. 2022;Wu et al. 2022c). Comparative genomic analyses have suggested that white rot fungi differ from brown rot fungi by their CAZymes repertoire (Floudas et al. 2012). Furthermore, glycoside hydrolase (GH6 and GH7 families) and lytic polysaccharide monooxygenase (AA9 family) genes are more abundant in white rot fungi than in brown rot fungi, and class II lignin-modifying POD (AA2 family) are usually totally lost in brown rot fungi (Floudas et al. 2012;Kohler et al. 2015;Krah et al. 2018). Moreover, ectomycorrhizal fungi, such as some species of Boletales and Russulales, have less plant cell wall degrading enzymes (PCWDEs) compared with those of ancestral wood decomposers, as well as many lineage-specific genes concerned with the degradation of soil organic material (Kohler et al. 2015;Lebreton et al. 2021;Lofgren et al. 2021;Wu et al. 2022c). However, only a few Hymenochaetales genomes have been published until now, for instance only 18 genomes are currently available in the NCBI database (https://www.ncbi.nlm.nih.gov/ genome, accessed on December 7, 2022). The few genomics studies on Hymenochaetales mostly addressed their pathogenicity, phylogeny, mitochondrial genomes, and medicinal value (Floudas et al. 2012;Chung et al. 2017;Lee et al. 2019;Caballero et al. 2020;Jiang et al. 2021;Zhao et al. 2022a), while their genomic features underlying different ecological types, including ectomycorrhizal, parasitic, and saprophytic, remain underexplored.
Here, we reveal the genome traits of ectomycorrhizal, parasitic, and saprophytic species within the Hymenochaetales using comparative genomics, mainly focusing on the secreted proteins and PCWDEs repertoires, as well as a reconstruction of their phylogenomic relationships and divergence time based on single-copy orthologous genes.

Genome Features
In the present study, 15 genomes of 11 species within Hymenochaetaceae, that is seven Coltricia, two Onnia, one Phellinus, two Porodaedalea, one Pseudoinonotus, and two Sanghuangporus, were newly sequenced and assembled (table 1, figs. 1 and 2 and supplementary File S1, Supplementary Material online).
In the genus Coltricia, the assembled genome sizes ranged from 72.2 Mb (C. perennis Dai 23736) to 150.9 Mb (Coltricia weii Dai 23719) with a GC content of 45.88% and 49.98%, respectively for Dai 23736 and Dai 23719. Between 11,778 and 42,579 predicted protein-coding gene models were predicted, respectively, which showed a rapidly evolving genome size and gene number.
In the genus Onnia, the assembled genome size of Onnia himalayana Dai 22620 was 31.7 Mb, and of Onnia tomentosa Dai 23682, 37.7 Mb, with a GC content of 49.65% and 49.63%, respectively, and 9,402 and 10,231 proteincoding gene models were predicted, respectively.
The assembled genome size of Phellinus monticola Dai 22944A was 35.4 Mb with a GC content of 49.08% and 10,443 protein-coding gene models were predicted.
In  fig. 2b), suggesting that genomes captured most of the protein-coding gene space. In addition, the genomes of the ectomycorrhizal species are significantly larger than the parasitic and saprophytic species in assembly size and number of proteincoding gene models, while no significant differences were observed between parasitic and saprophytic species ( fig. 3a and b; P < 0.01).

Phylogenomic Relationships
A maximum likelihood (ML) phylogenomic analysis of the 32 genomes was carried out under the tentatively best substitution model PROTGAMMAIJTTF (supplementary Files S2 and S3, Supplementary Material online). The generated Maximum Clade Credibility (MCC) tree suggests that the ectomycorrhizal genus Coltricia forms a clade located at the base of the Hymenochaetaceae ( fig. 2a), and the divergence time estimation suggested that Coltricia occurred at a mean stem age of 105.9 Mya, most extant species of Hymenochaetales diversified at no more than 20 Mya. However, parasitic and saprophytic species of Hymenochaetaceae did not form independent clades. It showed that two parasitic/saprophytic species, Coniferiporia sulphurascens and F. viticola, clustered with saprophytic species in Hymenochaetaceae, which is in agreement with previous phylogenetic studies (Dai 2010;Wu et al. 2022b Most notably, the ectomycorrhizal genus Coltricia has the largest genomes with significantly higher TE contents compared with those of parasitic and saprophytic genera in the Hymenochaetales ( fig. 3c; P < 0.01), while there were no significant differences (P > 0.05) in the average TE contents between parasitic and saprophytic fungi Secreted CAZymes play an important role in wood degradation, and our analyses show that the average number of secreted CAZymes in ectomycorrhizal species is significantly less than that in parasitic and saprophytic species (91.7 ± 25.6 vs. 165.4 ± 24.2 and 201.4 ± 58.5, respectively; figs. 3e and 4c). Among these secreted CAZymes (supplementary fig. S1, Supplementary Material online), ectomycorrhizal species are significantly different from parasitic and saprophytic species regarding the gene number of auxiliary activities (AAs), carbohydrate-binding modules (CBMs), glycoside hydrolases (GHs), and glycosyl transferases, while less difference is found in carbohydrate esterases (CEs).

PCWDEs Loss in Ectomycorrhizal Coltricia
In this study, a total of 59 secreted CAZymes families, including eight AAs families, five CBMs families, three CEs families, 40 GHs families, and three polysaccharide lyase (PLs) families, is annotated as secreted plant, fungal, or bacterial cell wall degrading enzymes (BCWDE) ( fig. 5 and supplementary File S4, Supplementary Material online). The gene copy number for secreted PCWDEs of ectomycorrhizal species showed striking differences to the parasitic and saprophytic species (figs. 2i and 5; P < 0.01). The  . 4c). We classified the gene families involved in secreted cell wall degradation enzymes into 17 main categories as described in previous studies (Sipos et al. 2017;Wu et al. 2022c). Compared with those of parasitic and saprophytic species, ectomycorrhizal species have almost lost the capacity to degrade cellulose, hemicellulose, and pectin ( fig.  4c, fig. 5 and supplementary File S4, Supplementary Material online). For example, Coltricia species have completely lost glycoside hydrolase (GH6 and GH7 families), and cellulose-binding motif (CBM1 family, core cellulose-acting CAZymes attached to PCWDEs to mediate the targeting of enzymes to cellulose; Martínez et al. 2018;Wu et al. 2022c), which, however, are over-represented in parasitic and saprophytic species. From 11 to 21, secreted PCWDEs were predicted in the ectomycorrhizal Coltricia species, indicating a limited capacity to degrade plant cell walls. Concerning lignin degradation, the ectomycorrhizal species had one to two genes of AA2 family (class II ligninmodifying PODs) annotated, while parasitic and saprophytic species had 8 to 24 genes for the AA2 family. However, the ectomycorrhizal species have more genes to degrade chitin (CBM50 and GH18 families) and peptidoglycan (GH23 family), indicating that they have a stronger ability to degrade fungal and bacterial cell walls.
On the other hand, a total of 68 CAZyme families (including secreted and nonsecreted) are classified as cell wall degrading enzymes, and 164-445, 236-353, and 204-368 genes were annotated in ectomycorrhizal, parasitic, and saprophytic species, respectively (supplementary figs.

The Relationship Between Hymenochaetales Species and Host Trees
We analyzed the secreted CAZyme families of cellulose and lignin in Hymenochaetales species, which grown on gymnosperms or angiosperms, including AA9, AA16, CBM1, GH5_5, GH6, GH7, GH12, GH 45, GH131, and AA2 (supplementary fig. S4, Supplementary Material online). AA2 family plays a major role in the degradation of lignin content in plant cell walls. Although relatively more genes of AA2 families in species growing on gymnosperms were detected than in those growing on angiosperms, this difference was not significant (supplementary fig. S4a, Supplementary Material online; P > 0.05).
The results showed that no significant difference is observed in the numbers of genes of nine gene families involved in cellulose degradation (supplementary fig. S4b-l, Supplementary Material online; P > 0.05). Principal component analyses (PCA; fig. 6) of CAZymes (secreted and nonsecreted) suggest that ectomycorrhizal species are significantly different from parasitic and saprophytic species, which is probably related to the ecology of the fungi. The type of host plant (whether angiosperms or gymnosperms) has a less influence, except for Rickenella species associated with bryophyte ( fig. 6). In addition, the PCA analyses of host plants with CAZymes (secreted and nonsecreted) were preformed, the results showed the limited patterns observed that no significant differences to parasitic-only and saprotrophic-only species grown on angiosperms and gymnosperms (supplementary fig. S5, Supplementary Material online).

Rickenella fibula and Rickenella mellea as Special Saprophytic Fungi
Rickenella fibula and R. mellea are very similar to saprotrophic species in the Hymenochaetales in terms of genome size, gene models, TE contents, secreted proteins, and cell wall degrading enzymes (table 1, supplementary Files S3 and S4 and S5, Supplementary Material online) but are significantly different from species of the ectomycorrhizal genus Coltricia. The gene copy numbers of secreted PWCDEs in R. fibula and R. mellea are 186 and 181, respectively ( fig. 4), especially mainly concentrated in the AA1_1, AA2, AA7, AA9, CBM1, CE12, GH5_5, GH15, GH16, GH18, GH28, GH43, and PL14_4 families, suggesting that both R. fibula and R. mellea have a powerful capacity for degrading plant cell walls. And R. fibula and R. mellea have a single gene copy of secreted GH32 family, which indicates they have the ability to degrade sucrose. Korotkin et al. suggested R. fibula appear to have multiple trophic modes and most likely maintaining a commensal endophytic relationship with its moss host. (Korotkin 2017;Korotkin et al. 2018). Considering the ancestors were saprotrophic, and R. fibula and R. mellea have a large number of PCWDEs, inferring that they are as special saprophytic fungi.

Discussion
In the fungal kingdom, ectomycorrhizal fungi have evolved independently in 78-82 fungal lineages that comprise 251-256 genera (Tedersoo and Smith 2013;Martin et al. 2016). Phylogenetic analysis of 8,400 species within Agaricomycetes suggested 36 ectomycorrhizal origins (Sánchez-García et al. 2020), among which many EcM are located in the orders Agaricales, Boletales, Cantharellales, and Russulales, while only a few are found in the Hymenochaetales, that is, Coltricia species (Sánchez-García et al. 2020;Hackel et al. 2022;Wu et al. 2022bWu et al. , 2022c. Comparative genomics indicated that the ancestors of EcM were ecologically diverse, including brown rot, white rot, and soil saprotrophs (Kohler et al. 2015;Miyauchi et al. 2020;Lebreton et al. 2021;Wu et al. 2022c). Here, phylogenomic and molecular clock analyses based on 31 Hymenochaetales genomes showed that ectomycorrhizal Coltricia is located at the base of the Hymenochaetaceae and divergence time of Coltricia later than the saprotrophic species. Interestingly, brown rot and their EcM decedents, such as ectomycorrhizal Boletales and Atheliales/Amylocorticiales, lack the secreted AA2 family (Wu et al. 2022c), while a single to two gene copies of AA2 family is predicted in Coltricia, inferring that Coltricia may have originated from saprotrophic ancestor with white rot.
A distinguishing character of EcM related to white rot and brown rot fungi is the loss of several families of PCWDEs, especially those acting on cellulose and lignin (such as AA2, GH6, and GH7 families). Unique array of PCWDEs for the ectomycorrhizal fungi was found, such as GH5 endoglucanases with a CBM1 cellulose-binding motif, pectinases (GH28 family), oxidoreductases, and laccases (AA1 and AA9 families; Kohler et al. 2015;Martin et al. 2016;Lebreton et al. 2021;Sun et al. 2022;Wu et al. 2022c). In this study, we found that the gene copy number for secreted PCWDEs in Coltricia species is also dramatically reduced, as reported for ectomycorrhizal orders Boletales, Russulales, Thelephorales, and family Amanitaceae (Hess et al. 2018;Miyauchi et al. 2020;Lofgren et al. 2021;Looney et al. 2021;Wu et al. 2022c). Coltricia species completely lost the CBM1, GH6, and GH7 families, which is a significant difference to those of the parasitic and saprophytic relatives.
Additional genome traits, such as the content in transposable elements (TEs), SSPs have also been investigated in EcM (Hess et al. 2018;Miyauchi et al. 2020;Lofgren et al. 2021;Looney et al. 2021;Wu et al. 2022c). Our results show that TE and SSP content in the genome of ectomycorrhizal Coltricia species are also enriched. In addition, rapidly dynamic genomes with higher gene models are observed in Coltricia, similar to the genus Suillus (Lofgren et al. 2021). However, more genomes within the Hymenochaetales should be sequenced and analyzed to assist in studying evolutionary diversification because of the richness of species in the order.
The ecological groups (saprophytic and parasitic) and the corresponding host trees (gymnosperms and angiosperms) of Hymenochaetales species have highly similar genome features, including PCWDEs, secreted CAZymes and TEs, and are not divided into two independent groups based on their genome comparison. Despite differences in cellulose and lignin profiles between gymnosperms and angiosperms (Cornwell et al. 2009;Thakur and Thakur 2014), we revealed no significant differences in secreted CAZymes of Hymenochaetales species growing on the different host trees, inferred that the formation of secreted CAZymes of Hymenochaetales is earlier than the differentiation of the host trees and there are other recognition mechanisms for fungi allowing growth on either gymnosperms or angiosperms. Molecular dating analyses have suggested that the divergence times of Hymenochaetales and Basidiomycota are 167-259 Mya and more than 400 Mya, respectively (He et al 2019;Varga et al. 2019;Zhao et al. 2022b). In contrast, the host trees, viz., gymnosperms and angiosperms, can be traced back to the Carboniferous period 300-350 Mya (Won and Renner 2006;Clarke et al. 2011;Magallón et al. 2013;Wang and Ran 2014), which later than the origin FIG. 5-Distribution of classical secreted CAZymes in 31 Hymenochaetales genomes in three ecological groups. The CAZyme genes identified correspond to eight AAs, five CBMs, three CEs, 40 glycoside hydrolase (GHs), and three PLs families. The potential substrates are cellulose, hemicellulose, lignin, and pectin for plant cell walls; chitin, glucan, and mannan for fungal cell walls, and peptidoglycan for bacterial cell walls. Cel, cellulose; Hem, hemicellulose; Pec, pectin; Phe, phenols; Lig, lignin; PCW, partial plant cell wall degradation; Cut, cutin; Glu, glucan; Chi, chitin; Man, mannan; Pep, peptidoglycan; Sta, starch; Gly, glycogen; Tre, trehalose; Suc, sucrose. of Basidiomycota. In addition, some studies have suggested that many factors, such as Pi transporters, chitotetraose, receptor-like kinase, lipochitooligosaccharides, and SSPs, could be involved in the formation of an ectomycorrhizal life mode (Plett et al. 2011;Becquer et al. 2018;Cope et al. 2019;Pellegrin et al. 2019;Zhang et al. 2021), and these may be involved in the recognition mechanism to grow on gymnosperms or angiosperms.  To conclude, the 15 new genomes of Hymenochaetaceae were sequenced in this study, including important ectomycorrhizal Coltricia, provided valuable data for phylogenomic and genomic analyses to explain evolutionary innovations in the order Hymenochaetales.

Phylogenomic and Divergence Time Analyses
A total of 32 genomes, including 31 Hymenochaetales and one genome of Agaricus bisporus as outgroup, were used to reconstruct phylogenomic relationships based on singlecopy orthologous genes (table 1). These genes were found using OrthoFinder v2.5.4 (Emms and Kelly 2019) and aligned using MAFFT v7 (Katoh and Standley 2013); those alignments that covered less than 50 amino acids or poorly aligned were excluded. A ML phylogenomic tree was reconstructed by RAxML v8.1.12 (Stamatakis 2014) with 100 bootstrap replications. The best substitution model was estimated using ModelTest-NG v0.1.7 (Darriba et al. 2020). The divergence time was estimated using r8s v1.71 (Sanderson 2003) based on the single-copy orthologous genes. The calibration of Hymenochaetales was 167 Mya with the 130 Mya of minage and 180 Mya of max-age (Varga et al. 2019), and the penalized likelihood method was selected. Finally, the ML and MCC trees were viewed with FigTree v1.4.4 (http://tree.bio. ed.ac.uk/software/figtree; accessed on May 1, 2020).

Supplementary material
Supplementary data are available at Genome Biology and Evolution online (http://www.gbe.oxfordjournals.org/).