Overall characteristics of fungal communities
The fungal community structure associated with P. densiflora was determined for 78 root samples and 75 soil samples from 16 locations in South Korea after removing samples with insufficient sequence reads (Figure 1A; Table 1). In total, 11,150,431 rRNA sequence reads were obtained after quality filtering, and 10,712,967 reads after filtering singletons and non-fungal sequences. The average length of sequence reads was 386.01 ± 48.18 base pairs. We found 12,927 OTUs after post-clustering, and 2,206 OTUs remained after filtering rare OTUs (≤10 reads or ≤5 samples) and normalization. From each microhabitat, 1 831 OTUs and 1 928 OTUs were found in the roots and soil, respectively. On average, 195.86 ± 53.13 OTUs and 222.76 ± 71.42 OTUs were found in each root and soil sample, respectively. Good's coverage ranged from 0.978 to 0.998 with an average of 0.989 ± 0.003, indicating that our sequencing depth was sufficient to represent most of the fungal OTUs (Additional file 1: Table S1).
Fungal community composition
The composition of belowground fungi varied between the root and soil microhabitats. At the genus level, Tomentella, Geminibasidium, Suillus, Oidiodendron, and Cortinarius were the most abundant taxa (Figure 1B, Additional file 2: Table S2). Among the ectomycorrhizal genera, Suillus and Phellodon showed higher relative abundance in roots than in soil. In contrast, the relative abundance of Tomentella, Oidiodendron, and Cortinarius was higher in the soil than in root. Meanwhile, non-ectomycorrhizal fungi also showed different relative abundances according to the microhabitats. The genera Mycena, Trichoderma, and Basidiodendron were more abundant in roots than in soil, whereas Gemnibasidium, Umbelopsis, and Oidiodendron were more abundant in soil (Additional file 2: Table S2). Among the plant pathogenic fungi, Venturia was mostly found in root microhabitats but in relatively low abundance compared to other guilds. The relative abundance of Geminibasidium, Umbelopsis, and Mycena showed significant differences across microhabitats according to LefSe analysis (Figure 1B, Additional file 3: Figure S1).
The composition of belowground fungal communities also changed across altitudinal gradients. Suillus was more abundant in root samples collected at high altitudes, whereas Geminibasidium was more abundant in soil samples from low altitudes. Several genera, such as Mortierella and Tricholoma, showed opposite altitudinal patterns between root and soil samples (Fig. 1B, Additional file 2: Table S2). According to the results of LefSe analysis, in both microhabitats, Trichoderma sp. (OTU 31) was found at high altitudes, whereas Geminibasidium and Mycena were significantly abundant at low altitudes. Among ectomycorrhizal fungi, Astraeus and Cortinarius were the discriminating taxa of root and soil samples respectively from low altitudes (Additional file 4-5: Figure S2-3).
In terms of ecological function, saprotrophs and ECM were the major guilds in both microhabitats. Saprotroph (p < 0.01), endophytes (p < 0.001), and saprotroph + plant pathogens (p < 0.001) were more abundant in roots, whereas the ECM (p < 0.001) and ericoid mycorrhiza (p < 0.001) were more abundant in the soil (Fig 2A). The relative abundance of saprotroph + endophytes was not significantly different between the microhabitats. Relative abundance of Saprotroph + endophyte guild was higher in the soil than in the root at low altitudes, and the pattern was opposite at high altitudes (p < 0.05). Similarly, the relative abundance of ECM was significantly higher in the soil than in the roots only at high altitudes (Fig. 2B, p < 0.001).
Change in fungal diversity across microhabitats and altitudinal gradients
We found that microhabitat type had a significant effect on the community composition of belowground fungal communities (Fig. 3, Table 2, Additional file 6: Table S3). However, their influence differed according to the ecological groups. According to the results of the PERMANOVA test, the effects of microhabitat were higher in non-ectomycorrhizal fungi (R2 = 0.083) than in ectomycorrhizal fungi (R2 = 0.018) or the entire fungal communities (R2 = 0.048). UPGMA analysis showed that the composition of fungal communities between the two microhabitats was much more similar in ECM fungal communities than in non-ECM fungal communities (Additional file 7: Figure S4).
Altitude had a significant effect on community composition, but the effect was different between ECM and non-ECM fungal communities (Fig. 3, Additional file 6: Table S3). The community of non-ECM fungi was more strongly influenced by altitude than that of ECM fungi, regardless of the microhabitat type (Additional file 6: Table S3). Similar to the results of the PERMANOVA test, those of the Mantel test revealed that altitude exhibited significant distance–decay patterns with fungal communities. The negative correlation between altitudinal difference and fungal community dissimilarity was stronger in the soil than in the roots in all groups (Additional file 8: Figure S5). In contrast, the similarity of fungal communities was more negatively correlated with geographic distances in roots than in soil (Additional file 9: Figure S6). Overall, both microhabitat and altitude were significant factors that affected the fungal community composition. In most microhabitats, non-ectomycorrhizal fungal communities were more affected by altitudinal differences than the total fungal or ectomycorrhizal fungal communities.
The relationship of the alpha-diversity indices of fungal communities associated with P. densiflora and altitude varied according to microhabitats and ecological groups (Fig. 4). In roots, significant negative relationships were found between alpha-diversity indices and altitude in both groups. However, in soil, different patterns were observed between the ECM and non-ECM communities. Although a significant negative correlation between altitude and alpha-diversity of ectomycorrhizal and overall fungal communities was detected, non-ECM communities exhibited no significant relationships with altitude.
Table 2 PERMANOVA results for fungal community structure*
|
Variable
|
df
|
R2
|
p-value
|
Overall fungal community (ECM + non-ECM)
|
Altitude
|
1
|
0.025
|
0.001***
|
Microhabitats
|
1
|
0.048
|
0.001***
|
ECM fungal community
|
Altitude
|
1
|
0.018
|
0.001***
|
Microhabitats
|
1
|
0.018
|
0.001***
|
Non-ECM fungal community
|
Altitude
|
1
|
0.031
|
0.001***
|
Microhabitats
|
1
|
0.083
|
0.001***
|
*calculated based on Jaccard distance and relative abundance at the OTU-level with 999 permutations (*p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001).