**3. Results**

#### *3.1. Environmental Factors*

The coverage in the area within 100 m from the highway reduced more significantly than the control (400 m) at three sites (Figure 2B). The coverage at 20 m decreased by 64.44% (0.20), 67.24% (0.24) and 64.86% (0.33) in 4000 m, 4600 m and 5200 m sites, respectively. The aboveground biomass was negatively correlated with distance in three sites. The aboveground biomass decreased significantly within 100 m, 50 m, and 100 m from the road, respectively (Figure 2A). The plant Shannon index decreased significantly in 4000 m and 5200 m sites within 20 m from the highway and decreased significantly at the 4600 m site within 50 m from the highway. For the plots 20 m away from the highway, the values decreased by 28.14% (1.62), 14.05% (1.56), and 28.80% (1.21) compared with the control (400 m) in 4000 m, 4600 m, and 5200 m sites, respectively (Figure 2C). The Simpson index (Figure 2D) of plants also shows similar changes in the areas close to the road (50 m, 20 m, and 5 m plots). Compared with the control (400 m), the plants' heights were significantly higher. The increase ranges were 561.17% (35.59 cm), 250.46% (16.15 cm) and 59.38% (8.54 cm).

**Figure 2.** *Cont*.

**Figure 2.** Comparison of (**A**) above-ground biomass, (**B**) plant cover degree, (**C**) plant simpson and (**D**) above-ground biomass between treatments at 5, 20, 50, 100 and 400 m from the curb G109 highways at altitudes of 4000 m (L), 4600 m (M) and 5200 m (H).

The soil moisture (SM) was significantly reduced by 58.63%, 38.45%, and 73.04% within 100 m along the highway at 20 m plots from the highway in 4000 m, 4600 m, and 5200 m sites, respectively (Figure 3A). In the 5200 m site, soil pH significantly decreased by 6.26%, 6.07%, 3.47%, and 2.11% at 5 m, 20 m, 50 m, and 100 m from the road, respectively (Figure 3B). At the same time, smaller changes occurred at 4000 m and 4600 m, with only significant decreases of 2.75% and 1.29% at 5 m from the road, respectively. The soil temperature in the areas 100 m and 50 m away from the highway decreased significantly compared with the control, 4000 m site and 5200 m site, respectively (Figure 3C). Soil compaction only changed significantly at 5200 m and decreased by 9. 14% (1715.75 Pa) and 5.69% (1780.75 Pa) at 5 m and 20 m away from the road, respectively (Figure 3D).

SOC is negatively related to the distance from the road. This trend is most obvious at 5200 m (Figure 4A). Roads significantly reduced soil organic carbon by 33.85% (.84 g·kg<sup>−</sup>1), 36.43% (8.49 g·kg−1), and 32.72% (8.99 g·kg−1) at plots 5 m, 50 m, and 100 m away from the road, respectively. Soil organic carbon content decreased by 70.19% (8.35 g·kg−1) and 66.01% (9.52 g·kg<sup>−</sup>1) at 5 m and 20 m plots from the road, respectively, compared with the control (400 m) at 5200 m. Soil total nitrogen content decreased by 67. 48%, 60.98%, 19.86% and 21.12% at 5 m, 20 m, 50 m, and 100 m away from the highway, respectively (Figure 4B). In the 4600 m site, soil total nitrogen content significantly decreased by 27.17% at 20 m from the road. At the same time, the change of soil total nitrogen content was not significant at the 4000 m site. The change trend of alkali-hydrolyzable nitrogen (AN) was the same as TN in the 5200 m site (Figure 4C). The difference is that angle in which the content of soil AN decreased significantly became within 50 m from the road. On the contrary, at the site of 4600 m, the content of soil AN at 5 m plot away from the road was significantly increased by 16.36% (67.02 mg·kg−1) compared with the control (400 m). At the same time, at the site of 4000 m, the content of soil AP (Figure 4D) at the plot of 20 m from the road was significantly increased by 30.34% (80.00 mg·kg−1) compared with the control (400 m). In 5200 m site, the content of soil available phosphorus increased by 119.04% (4.80 mg·kg<sup>−</sup>1) and 198.97% (6.56 mg·kg−1) at 5 m and 20 m distance respectively. The content of soil available phosphorus increased at the distance of 50 m (56.15%) and 20 m (169.42%) from the highway, and at the distance of 4000 m and 4600 m, respectively. The change of soil AK with distance was not significant (Figure 4E).

**Figure 3.** Comparison of soil: (**A**) soil moisture content, (**B**) pH, (**C**) soil temperature and (**D**) soil density between treatments at 5, 20, 50, 100, and 400 m from the curb G109 highways at altitudes of 4000 m (L), 4600 m (M), and 5200 m (H).

The Contents of Cu, Zn, and Cd increased significantly in the 4000 m site, and the increasing areas were mainly concentrated in the area 20 m away from the road (Figure 5). The three heavy metals contents increased by 17.18%, 14.73%, and 31.58% at 20 m from the road, compared with the control(400 m in 4000 m sites).The total lead content in the soil increased by 41.92% and 63.57% at 5 m and 20 m, respectively, in the 4600 m sites.

#### *3.2. Bacterial α-Diversity*

Alpha Diversity Index is used to reflect species diversity. Observed Species and Chao1 are used to indicate the actual number of OTUs in the community. The larger the Shannon Wiener and Simpson values are, the higher the community diversity is and the more uniform the individual distribution is. The Good's Coverage index reflects the sequencing depth, and the closer the index is to 1, it indicates that the sequencing depth has covered all species in the sample. The larger the value of the Phylogenetic Diversity index, indicates that the species that make up the biome are further apart in evolution. Specaccum accumulation curves determine whether the amount of sequenced data completely covers the species on the total sample. After pruning and quality filtering, 98,532 high-quality readings were obtained from a total of 60 samples from three altitudes, and the amount of clean tags data after quality control ranged from 4802 to 52,373. The data size of the valid tags (used for analysis) ranged from 11,648 to 42,192, with an average length of 426.58 to 432.68 bp.

**Figure 4.** Comparison of soil: (**A**) organic carbon (SOC), (**B**) total nitrogen (TN), (**C**) alkali-hydrolyzable nitrogen (AN), (**D**) available phosphorus (AP), and (**E**) available potassium (AK) between treatments at 5, 20, 50, 100, and 400 m from the curb G109 highways at altitudes of 4000 m (L), 4600 m (M), and 5200 m (H).

**Figure 5.** Concentrations of Cu (**A**), Zn (**B**), Pb (**C**), and Cd (**D**) in the 0–20 cm soil in sites 5, 20, 50, 100, and 400 m away from the G109 highway at altitudes of 4000 m (L), 4600 m (M), and 5200 m (H). The dotted line in the figure represents the background value of heavy metals in the Qinghai-Tibet Plateau.

The number of OTUs obtained from all treated soil samples ranged from 786 to 1987, and the species coverage (Good's Coverage index) was more than 90%. Based on a genetic distance of 3%, the sparse curve tends to saturate the plateau, indicating that the sequencing depth is sufficient (Figure 6D). The species accumulation curve showed that the sample size was sufficient for subsequent data analysis (Figure 6E).

The impact of highway traffic on bacterial α-diversity was mainly reflected in the 5200 m site, and the OTUs of soil bacteria significantly increased by 12.78% (1713) and 11.67% (1696) in the 5 m and 20 m plots in the 5200 m sites, respectively. Consistently, the bacterial Shannon index also increased significantly by 4.40% (9.40) and 3.86% (9.35) at 5 m and 20 m from the road, respectively, at 5200 m. In the 4600 m site, the bacterial CHAO1 index significantly increased by 8.93% (2901.93) at the 5 m site. On the contrary, the number of bacterial OTUs and Shannon index decreased significantly by 13.35% (1544.7) and 7.61% (8.56), respectively at 50 m from the road in 4000 m site (Figure 6).

Pearson correlation analysis showed that the environmental factors, especially altitude, SOC, AK, TN, AN, SM, pH, Density, Cu, and Simpson were significantly correlated with the bacterial α diversity index. OUT number and CHAO were negatively correlated with the PD whole tree index. In addition, there was a significant positive correlation between altitude and bacterial Simpson index. Cu content was significantly negatively correlated with Simpson index and Good's Coverage index. The Simpson index of plants was negatively correlated with that of bacteria (Figure 6D).

**Figure 6.** Box Diagram of (**A**) bacterial operational taxonomic unit (OTU), (**B**) bacterial Shannon and (**C**) bacterial CHAO1 in the 0–20 cm soil depths at 5, 20, 50, 100, and 400 m from G109 highway at altitudes of 4000 m (L), 4600 m, (M) and 5200 m (H). Pearson correlation heatmaps are based on bacterial alpha diversity estimators and environmental factors (**D**). \* Means *p* < 0.05 for significance test; \*\* means *p* < 0.01 for significance test; \*\*\* means *p* < 0.001 for significance test.

## *3.3. Bacterial Community Structure*

The relative abundance of phylum-level bacterial communities is shown in Figure 7. The major phyla in all samples were *Proteobacteria* (20.83–40.02%), *Actinobacteria* (15.28–31.09%), *Bacteroidetes* (13.16–27.99%), *Firmicutes* (9.72–27.99%), *Gemmatimonadetes* (1.61–8.10%), *Acidobacteria* (1.39–5.06%), *Nitrospirae* (0.15–1.13%), *Chloroflexi* (0.13–0.86%), *Cyanobacteria* (0.15–1.08%) and *Chlorobi* (0.13–0.47%), with a total of more than 98% in each sample (Table S1).

**Figure 7.** Chart of the relative abundance of the different levels of bacteria at the phylum level in the 0–20 cm depths soil at 5, 20, 50, 100, and 400 m from the curb G109 highways at altitudes of 4000 m, 4600 m, and 5200 m.

The *Proteobacteria* abundance decreased by 30.65% (28.03%) at the 50 m site than the control (400 m). The *Actinomycetes* abundance significantly decreased by 47.07% (20 m plot, 15.28%) in the 4000 m and 4600 m sites than the control. Conversely, the abundance of *Actinomycetes* significantly increased by 87.71% (29.03%) and 98.43% (28.03%) at 5 m and 20 m, respectively. *Bacteroides* abundance in the 20 m plot significantly increased by 75.92% (23.15%) in the 4000 m site compared with the control, while in the 5 m plot it significantly decreased by 34.82% (19.08%) in the 5200 m site compared with the control. The same results occurred with acid bacilli. At the 4000 m and 4600 m sites in the highway disturbance area, the abundance of acid bacilli at the 20 m and 50 m plots showed a significant decrease than control. However, the abundance of *Acidobacteria* significantly increased by 88.16% (2.61%) and 138.27% (3.31%) in the 5 m and 100 m sites, respectively. In addition, *Firmicutes* abundance in the 5 m and 20 m plots decreased significantly by 59.50% (10.61%) and 42.23% (15.13%) compared with the control in the 5200 m site. The abundance of *Nitrospirillum*, *Chlorospirillum*, *Cyanobacteria* and *Chlorobacteria* increased significantly in different degrees in the highway disturbance area.

In the genus level, the abundance of bacteria was analyzed by LEfSe in each site at each elevation, and the results are shown in Figure 8. Road traffic significantly increased the number of 4, 8, and 14 taxa of soil bacterial communities within 100 m (4000 m, 4600 m, and 5200 m), and significantly decreased the number of 5, 2, and 4 taxa of soil bacterial communities. *Flavisolibacter*, *Gemmatimonas*, *Microvirga*, *Nocardioides,* and *Rubrobacter* are enriched at sites of 4600 m and 5200 m in the disturbed area of the highway. *Crossiella* and *Gaiella* are enriched in the disturbed area at 4000 m and 5200 m sites respectively. At the same time, *Barnesiella* and *Blastococcus* were the main bacterial genera enriched in the unaffected area. *Flavisolibacter* abundance increased significantly at 5 m and 20 m sites at 4600 m and 5200 m sites, respectively. In addition, *Flavisolibacter* is also the genus with the highest importance index in 5200 m site in the random forest analysis. *Alistipes* and

*Barnesiella* increased in abundance in the undisturbed area at the 5200 m site. In the 4600 m site, *Gemmatimonas*, *Rubrobacter*, and *Microvirga* all increased in abundance close to the road, consistent with the 5200 m site. *Pseudomonas* was the most important genus in random forest analysis at the 4000 m site, and its abundance increased significantly at the 20 m site, and the content of soil alkali-hydrolysable nitrogen also increased significantly at the 20 m site. *Marmoricola* abundance was lower in the highway disturbance area compared to the control in the 4000 m site. *Nocardioides* were enriched at 5 m in 4600 m and 5200 m sites, in contrast to 4000 m in the control (400 m) group. In random forest analysis, the top 10 genera were *Skermanella*, *Prevotella* 9, *Rubrobacter*, *Gaiella*, *Blastococcus*, *Crossiella*, *Flavisolibacter*, *Faecalibacterium*, *Pseudomonas,* and *Microvirga* (Figure 8A).

**Figure 8.** Linear discriminant analysis (LDA) of soil bacterial community in 3 sites of (**A**) 4000 m, (**B**) 4600 m, and (**C**) 5200 m. Rarefaction curve (**D**) and species accumulation curves (**E**) of the 60 soil samples.
