Response of soil microbial communities to alpine meadow degradation severity levels in the Qinghai-Tibet Plateau

Soil microbial community structure is an effective indicator to reflect changes in soil quality. Little is known about the effect of alpine meadow degradation on the soil bacterial and fungal community. In this study, we used the Illumina MiSeq sequencing method to analyze the microbial community structure of alpine meadow soil in five different degradation levels (i.e., non-degraded (ND), slightly degraded (LD), moderately degraded (MD), severely degraded (SD), and very severely degraded (VD)) in the Qinghai-Tibet Plateau. Proteobacteria, Actinobacteria, and Acidobacteria were the mainly bacterial phyla in meadow soil across all five degradation levels investigated. Basidiomycota was the mainly fungal phylum in ND; however, we found a shift from Basidiomycota to Ascomycota with an increase (severity) in degradation level. The overall proportion of Cortinariaceae exhibited high fungal variability, and reads were highest in ND (62.80%). Heatmaps of bacterial genera and fungal families showed a two-cluster sample division on a genus/family level: (1) an ND and LD group and (2) an SD, VD, and MD group. Redundancy analysis (RDA) showed that 79.7%and 71.3% of the variance in bacterial and fungal composition, respectively, could be explained by soil nutrient conditions (soil organic carbon, total nitrogen, and moisture) and plant properties (below-ground biomass). Our results indicate that meadow degradation affects both plant and soil properties and consequently drives changes in soil microbial community structure.


46
The Qinghai-Tibet Plateau (QTP) has an area over 250×10 4 km 2 within China and is 47 biodiversity maintenance [5], and soil carbon (C) sequestration [6]. 54 Furthermore, the lower air temperature and higher altitude make alpine meadows 55 more sensitive to global warming. Thus, these ecosystems are considered as good 56 indicators of environmental change [2]. In recent decades, frequent reports on the

93
Study area 94 The study area was located at the source of the Yangtze River on the QTP (34°49′N, 95 92°55′E; 4635 m a.s.l.) (Fig 1)     respectively, compared to LD (Table 1). Plant AGB in MD, SD, and VD significantly 160 (P <0.05) decreased by 47.09%, 51.60%, and 84.14%, respectively, compared to LD 161 (Table 1). Plant BGB in SD and VD also significantly decreased by 60.43% and 162 67.12% compared to ND and by 59.67% and 66.49% in SD and compared to LD, 163 respectively ( Table 1). The ND and LD sites not exhibited a significant differ in plant 164 coverage, AGB, or BGB (Table 1). There was no significant change in species 165 richness (SR) and plant height throughout the whole degradation process (Table 1). 166 For plant group, sedge biomass in SD and VD significantly decreased compared to 167 LD, but we found no significant change in the biomass/proportion of graminoids and 168 forbs as degradation severity increased (Table 2). 169 Soil TN, SOC, NO 3 --N, NH 4 + -N, and SM content decreased significantly in MD, 170 SD, and VD compared to ND, but BD exhibited the opposite response pattern (Table   171 3). In MD, there were significant decreases in soil TN, SOC, NO 3 --N, NH 4 + -N, and 172 SM by 62.20%, 67.75%, 48.40%, 47.94%, and 43.57%, respectively, compared to ND 173 (Table 3) respectively, compared to ND (Table 3). However, soil BD increased significantly (P 178 < 0.05) by 37.14%, 28.57%, and 47.14% in MD, SD, and VD, respectively, compared 179 to ND (Table 3). We found no significant difference in SOC, TN, NO 3 --N, NH 4 + -N, 180 SM, and BD between ND and LD (Table 3). for all samples at a 3% genetic distance. The progression of rarefaction curves 189 (99.01%-99.26%; Good's coverage) was very close for all samples (S1 Table). in the samples, for which LD had the highest one and ND the lowest (S1 Table).

217
The fungal communities were assigned to 8 phyla, 78 families, and 138 genera.

218
Ascomycota was the most dominant division, comprising 34.52% (339) of the OTUs 219 and 44.43% of the total reads (Fig 3A). Basidiomycota was the second largest division, VD had the highest read value (Fig 3A). 230 On a family level, we found abundant unidentified fungal sequences in each 231 sample. The total relative abundance of unidentified fungi in ND, LD, MD, SD, and 232 VD were 9.65%, 40.06%, 46.40%, 43.89%, and 37.19%, respectively (Fig 3B). showed a two-sample cluster division. The first cluster was the ND and LD group, 246 and the second cluster was the SD and VD group, which first clustered together 247 before clustering with MD, resulting in the second SD, VD, and MD group. Results 248 from PCA also showed that the bacterial communities of SD and VD grouped to the 249 left of the graph along the PC1 axis, accounting for 41.08% of total variation, whereas 250 ND and LD grouped along the PC2 axis, with a total variance of 31.92% (Fig 5A).

251
For fungi, the heatmap (Fig 4B) was based on the top 20 fungal families. The figure   252 shows a two cluster sample division on a family rank level: ND and LD into one 253 group, and SD, VD, and MD into another. The PCA score plot (Fig 5B) Table).  Table). significantly decreased the soil nutrient status (Table 3). This was due to a decrease in  Table) demonstrates that The bacterial phyla in meadow soil investigated in this study exhibited low variability 316 in the different samples (Fig 2A); however, the composition of fungal phyla 317 significantly changed under conditions of degradation severity (Fig 3A), indicating 318 that fungal communities are more sensitive to degradation than bacterial communities.

319
It was also reported that biotic and abiotic factors have a greater influence on fungi  For the bacterial groups, the overall proportion of Proteobacteria generally 341 decreased while Actinobacteria generally increased in MD and VD compared to ND 342 and LD (Fig 2A). This could have been because many members of Proteobacteria 343 (particularly α-proteobacteria; Fig 2B)  In conclusion, this study, using the high throughput Illumina MiSeq sequencing