Identification of collected macrofungi
Identification of the 280 fungal samples was conducted based on morphological features of the fruiting bodies, with reference to microscopic features and/or sequence information from the ITS region if necessary. The samples belonged to 186 species, 100 genera, 53 families, 18 orders, 7 classes, and 2 phyla (Table S2). In addition, four samples comprised novel species that have been officially described [23, 38, 39] (Table S1 and Figure S1). Furthermore, 17 samples comprised potential novel species that require further validation (Table S1 and Figure S1).
Annotation performances of the four classifiers
To enable more accurate and efficient identification of large fungal samples, four annotation classifiers (C1–C4) were generated by training on four ITS sequence databases. The four classifiers were used to classify and annotate the ASVs from all samples, and the annotation performances were compared based on their accuracy and confidence of host annotation. Comparison with traditional morphological species identification methods indicated that classifier C3 had the highest annotation accuracy rate (at the species level) of 54%, with 150 of the 280 samples exhibiting consistent identification as with traditional methods (see Fig. 1). At the genus level, classifiers C3 and C4 exhibited the highest accuracy (94.29%), with 264 samples exhibiting consistent identification as with traditional methods (Fig. 1). Comparative analysis of the host ASV annotation confidence values was conducted to evaluate the annotation performances of the four classifiers. Classifier C4 exhibited the best annotation performance (Fig. 1). Specifically, the C4 classifier achieved a confidence level of 95% or higher for 1,358 ASVs among the total 1,756 host ASVs, comprising an accuracy of 77.33%. Based on the above results, classifier C4 exhibited the best annotation performance in terms of accuracy and confidence, consequently making it most suitable for use in the annotation and classification of macrofungi ASVs.
Distribution of fungicolous fungi in macrofungal samples
The diversity of fungicolous fungi inhabiting sporocarps of 280 fungal host species collected from the Qinling Mountains was evaluated using a metabarcoding approach. A total of 769 species of fungicolous fungi were identified belonging to 573 genera, 225 families, 99 orders, 36 classes, and 15 phyla. Fungicolous fungal species and genera with relative abundances < 1% were excluded. Among fungicolous fungi with relative abundances > 1% in each sample, the most widely distributed species were Simplicillium cylindrosporum, followed by Tyromyces chioneus, Stereum hirsutum, Laccaria amethystina, and Neofavolus alveolaris, which were identified in 21.07%, 17.50%, 12.14%, 11.79%, and 11.43% of the samples, respectively (see Fig. 2a). The most widely distributed genera were Mycena, Gymnopus, Psathyrella, Simplicillium, and Russula that were found in 43.57%, 28.57%, 23.57%, 21.07%, and 20.36% of the samples, respectively (see Fig. 2b).
Diversity and composition of fungicolous fungal communities varied with growth months and host fungi habitats
To identify the environmental variables that were associated with the diversity and composition of fungicolous fungi, the alpha and beta diversities of fungicolous fungal communities were analyzed and compared between hosts growing in different environments, from different collection months, habitats, altitudes, and growth substrates. Significant differences in Chao1 and Shannon index values were observed between samples collected in different months or from different habitats (Fig. 2). Host fungi collected in August exhibited higher diversity (for both Chao1 and Shannon index values) than those collected in other months (Figs. 2c and 2d, p < 0.001). Hosts from coniferous forests also exhibited higher fungicolous fungal diversity than fungi from other habitats (Figs. 2e and 2f, p < 0.001). In addition, small differences were observed for both Chao1 and Shannon index values of fungicolous fungi for host species from different altitude areas or growth substrates (Figure S2).
Community compositional variation in fungicolous fungi was evaluated between samples of different categories using NMDS analysis based on Bray–Curtis dissimilarities of fungicolous fungal communities (Figs. 2, and S3). Each sample was distinct from the others, and a high degree of variation was observed between samples. Moreover, significant differences in fungicolous fungal community composition were observed between macrofungi sampled in different months or from different habitats (Fig. 2, p = 0.001). However, these differences were not clear for host macrofungi collected from different altitudes and with different growth substrates (Figure S3 and Table S5). These results were overall consistent with Shannon and Chao1 index value variation.
LEfSe was used to identify individual fungicolous fungi species or genera that exhibited statistically different relative abundances between host fungi sampled in different months or from different habitats. A total of 20 species and 22 genera exhibited significantly different enrichment among sampling months (Figure S4), while 18 species and 15 genera exhibited significantly different enrichment between samples from different habitats (Figure S4).
The distribution of fungicolous fungi in host species collected in different months or from different habitats was also evaluated. The most widely distributed fungicolous fungal species were Tubaria praestans (March-April), Laccaria amethystine (May-June), Simplicillium cylindrosporum (July), Agaricus moelleroides (August), Stereum hirsutum (September), and Hypholoma fasciculare (October-November) that were identified in 56.41%, 50.00%, 38.30%, 39.34%, 30.49%, and 40.00% of samples, respectively (Table 1). The most widely distributed genera were Psathyrella (March-April), Entoloma (May-June), Russula (July), Mycena (August), Mycena (September), and Gymnopus (October-November) that were identified in 61.54%, 58.33%, 53.19%, 63.93%, 43.90%, and 73.33% of host macrofungi, respectively (Table 1). Among the fungicolous fungi with relative abundances > 1% in each sample, the most widely distributed species in host samples collected from the scrub, mixed, hardwood, and coniferous forest habitats were Tubaria praestans, Stereum hirsutum, Tyromyces chioneus, and Agaricus moelleroides that were identified in 70.00%, 22.90%, 35.78%, and 43.33% of hosts, respectively (Table 2). At the genus level, the most widely distributed fungicolous fungi were Tubaria, Mycena, Mycena, and Agaricus that were identified in 70.00%, 32.06%, 52.30%, and 76.67% of host species, respectively (Table 2). Taken together, these results indicate that the diversity and composition of fungicolous fungi varied with the month of collection and the habitats of host fungi.
Table 1
Most widely distributed fungicolous fungal species and genera in macrofungal samples collected in different months.
Months | Widely distributed species | Widely distributed genera |
Name | Proportion of samples (%) | Name | Proportion of samples (%) |
March-April | Tubaria praestans | 56.41 | Psathyrella | 61.54 |
May-June | Laccaria amethystina | 50.00 | Entoloma | 58.33 |
July | Simplicillium cylindrosporum | 38.30 | Russula | 53.19 |
August | Agaricus moelleroides | 39.34 | Mycena | 63.93 |
September | Stereum hirsutum | 30.49 | Mycena | 43.90 |
October-November | Hypholoma fasciculare | 40.00 | Gymnopus | 73.33 |
Source data are provided as a source data file.
Table 2
Most widely distributed fungicolous fungal species and genera in macrofungal samples collected from different habitats.
Habitat | Widely distributed species | Widely distributed genera |
Name | Proportion of samples (%) | Name | Proportion of samples (%) |
Scrub forest | Tubaria praestans | 70.00 | Tubaria | 70.00 |
Mixed forest | Stereum hirsutum | 22.90 | Mycena | 32.06 |
Hardwood forest | Tyromyces chioneus | 35.78 | Mycena | 52.30 |
Coniferous forest | Agaricus moelleroides | 76.67 | Agaricus | 76.67 |
Source data are provided as a source data file.
Trophic modes of host macrofungi and fungicolous fungi
The lifestyle of fungicolous fungi were categorized into three trophic modes and compared. The proportion of saprophytic fungicolous fungi in samples decreased on average from March to July, while the proportion of symbiotic fungi gradually increased (p < 0.0001, Table 3 and Fig. 3). The proportion of saprophytic fungi gradually increased from July to November, while the proportion of symbiotic fungi decreased over the same period (p < 0.0001, Table 3 and Fig. 3). Compared with host macrofungi growing in other forests, the host macrofungi growing in mixed forests hosted fewer saprophytic fungicolous fungi and more symbiotic fungi (p < 0.0001, Table 3 and Fig. 3). In addition, host species growing on wood harbored more saprophytic fungicolous fungi and fewer symbiotic fungi compared with the hosts growing on the ground (p < 0.0001, Table 3 and Fig. 3).
Table 3
Trophic modes of fungicolous fungi from different sample types.
Categories | Saprotroph | Pathotroph | Symbiotroph | Others |
Month | March-April | 0.53 ± 0.11 *** | 0.05 ± 0.05 * | 0.03 ± 0.03 *** | 0.39 ± 0.08 |
May-June | 0.47 ± 0.1 *** | 0.06 ± 0.05 * | 0.06 ± 0.07 *** | 0.41 ± 0.08 |
July | 0.28 ± 0.11 *** | 0.07 ± 0.06 * | 0.25 ± 0.14 *** | 0.4 ± 0.09 |
August | 0.34 ± 0.1 *** | 0.09 ± 0.08 * | 0.15 ± 0.09 *** | 0.42 ± 0.08 |
September | 0.35 ± 0.13 *** | 0.08 ± 0.07 * | 0.15 ± 0.09 *** | 0.42 ± 0.09 |
October-November | 0.39 ± 0.05 *** | 0.07 ± 0.04 * | 0.13 ± 0.06 *** | 0.41 ± 0.06 |
Habitat | Coniferous forest | 0.39 ± 0.05 *** | 0.07 ± 0.03 | 0.11 ± 0.03 *** | 0.44 ± 0.06 * |
Hardwood forest | 0.44 ± 0.15 *** | 0.07 ± 0.07 | 0.1 ± 0.1 *** | 0.39 ± 0.09 * |
Mingled forest | 0.32 ± 0.11 *** | 0.08 ± 0.07 | 0.18 ± 0.12 *** | 0.42 ± 0.08 * |
Scrub forest | 0.46 ± 0.13 *** | 0.1 ± 0.09 | 0.03 ± 0.06 *** | 0.41 ± 0.06 * |
Substrate | Wood | 0.42 ± 0.15 *** | 0.08 ± 0.08 | 0.09 ± 0.08 *** | 0.41 ± 0.1 |
Ground | 0.35 ± 0.12 *** | 0.07 ± 0.05 | 0.17 ± 0.12 *** | 0.41 ± 0.08 |
Altitude | Intermediate | 0.38 ± 0.15 | 0.08 ± 0.07 | 0.14 ± 0.13 | 0.41 ± 0.09 |
High | 0.38 ± 0.11 | 0.07 ± 0.05 | 0.13 ± 0.08 | 0.41 ± 0.08 |
Data mean the average proportions of fungicolous fungi with different trophic modes at ASV level in all samples. Statistical significance was analyzed using Student’s t-tests (based on substrate and altitude) or one-way ANOVA tests (month and habitat). ***: p < 0.001; *: p < 0.05. Values are means ± SD. Source data are provided as a source data file.
The trophic modes of macrofungal hosts were also enumerated. Among the 280 hosts, 147 were saprophytic fungi, 2 were pathotrophs, and 75 were symbiotrophs (Table 4). Among the 111 host macrofungi growing on wood, 79 were saprophytic, accounting for 71.2% of the total, while only five were symbiotic (4.5% of the total; Table 4). Among the 169 hosts growing on the ground, 68 were saprophytic and 70 were symbiotic, accounting for 40.2% and 41.4% of the totals, respectively (Table 4).
Table 4
Trophic modes of host species.
Trophic modes | Number of host species |
Grown on ground | Grown on wood | Total |
Saprotroph | 68 | 79 | 147 |
Pathotroph | 1 | 1 | 2 |
Symbiotroph | 70 | 5 | 75 |
Others | 30 | 26 | 56 |
Source data are provided as a source data file.
Carbohydrate-active enzyme diversity of host macrofungi and fungicolous fungi
The annotated protein sequences of 216 host or fungicolous fungi, comprising 96 saprophytic fungi, 58 parasitic fungi, and 62 symbiotic fungi, were annotated against the CAZyme database using the dbCAN platform. The results from all three models were consistent. Overall, saprophytic fungi possessed higher abundances of CAZymes than parasitic and symbiotic fungi (p < 0.0001, Fig. 4). Saprophytic fungi encoded greater abundances of glycoside hydrolases (GHs), auxiliary activities (AAs), carbohydrate esterases (CEs), polysaccharide lyases (PLs), and carbohydrate-binding modules (CBMs), as well as fewer glycosyltransferases (GTs) (p < 0.05, Fig. 4).