*3.7. Variation Partitioning*

The results of variation partitioning analysis (VPA) showed that environmental factors such as altitude, vegetation, soil, and heavy metals could explain 32.71% of the variation in soil bacterial community structure, and the unexplained rate was 67.29%, of which the interaction of soil, vegetation, and heavy metals could explain 12.42% (Figure 13A). Environmental factors such as altitude, vegetation, soil, and heavy metals accounted for 34.89% and 65.11% of the changes in the ecological functional structure of soil bacteria FAPROTAX, of which the soil factor alone accounted for 8.02% and the interaction of soil, vegetation, and heavy metals accounted for 14.76% (Figure 13B). Compared to the soil bacterial community structure, the coupling effect of soil and vegetation had a greater impact on the functional structure of bacteria.

**Figure 13.** Variation partitioning analysis (VPA) which showed the relative proportions of total (**A**) bacterial and (**B**) functional diversity composition variations that can be explained by different types of environmental factors. The circles show the variation explained by each group of environmental factors alone.

#### **4. Discussion**

#### *4.1. Impact of Highway on Plant and Soil Physical and Chemical Properties*

During the highway construction and operation, different degrees of impact were inevitable on the grassland soil environment and vegetation along the highway. As we can predict, the grassland vegetation in the areas close to the highway has been degraded to varying degrees. In general, results showed the closer to the highway, the lower the coverage and diversity and biomass of grassland vegetation were. This is consistent with many previous research results [47]. This damage includes the direct excavation and rolling of grassland vegetation by construction and access vehicles [11]. It may also be indirectly caused by soil erosion caused by the change of terrain on both sides of the highway [12,48]. In this study, the vegetation height was higher in areas closer to the road. In the field investigation, we found that this was because some invasive plants such as *Stipa purpurea* appeared after the original low grassland of the QTP was disturbed and degraded [49]. At the same time, heavy metal pollution caused by road traffic has further aggravated the degradation of grassland vegetation [50]. In different altitude environments, the degradation of grassland vegetation is also different, of which 4600 m is similar to 5200 m, and the degradation is most obvious in the 4000 m environment. There may be

two reasons, one is that the 4000 m section is closer to the city (Golmud), and the traffic flow is larger. On the other hand, because the vegetation type of the 4000 m site is desert grassland, the vegetation diversity is low, and the ecological environment is fragile. It is more vulnerable to the interference of environmental factors.

Consistent with previous studies, our results indicate that highway slopes affect soil nutrient content, resulting in lower pH. Our results show that the soil moisture decreased significantly in the highway disturbed area. Engineering disturbance may cause soil compaction or decomposition by destroying original vegetation, altering soil porosity, and disrupting soil aggregates [51–53]. At the same time, the destruction of vegetation in the disturbed area of the highway leads to the decrease of the fixation ability of roots to soil, and the process of soil erosion may cause the loosening of soil structure. Our results show that soil pH is sensitive to engineering disturbance, which may be related to soil parent material [54,55]. Road disturbance affects soil pH by destroying native vegetation and affecting soil pH and nitrogen levels in plant material [56]. In addition, increased soil acidification may lead to increased leaching of cationic nutrients from the soil, thereby exacerbating the deficiency of certain nutrients essential for plant growth and ultimately leading to reduced plant productivity [57]. Because of the decrease of plant litter, the activity of soil denitrifying bacteria was promoted, and the mineralized nitrogen tended to nitrify, which led to the decrease of soil pH [56]. The decrease of soil moisture content (SM) in the highway disturbance area is related to the coverage of vegetation [53,58]. A reduction in the capacity of the soil to retain water due to degradation of vegetation. The change of soil temperature is contrary to the results of previous studies, which may be related to the weather at the time of sampling, or because the change of vegetation cover affects the thermal insulation performance. This requires further study.

Our results show that the soil SOC, TN, and AK in the disturbed areas of the highway do not change significantly compared to the control areas at 4000 m and 4600 m sites. The research of Pan and Jiang on highway slopes also shows the same situation [53,59]. All study areas are located below 4000 m and include the Tibetan Plateau region. In contrast, at the 5200 m site, the SOC, TN, and AK of highway disturbed area decreased significantly in our study, suggesting that SOC, TN, and AK may be more sensitive to the impact of the highway at higher altitudes. He et al.'s study showed that SOC and TN of highway disturbed areas significantly reduced compared with control [14]. The results of this study are consistent with this. However, there are few studies on the soil SOC and N contents in the highway disturbed area at an altitude of more than 5000 m. Therefore, we analyzed the differences of soil nitrate and ammonium nitrogen contents and related bacterial abundances among the three altitudes in this study. Our study revealed that due to the low oxygen content in high altitude areas, a low-oxygen environment is more conducive to denitrification, resulting in significantly higher soil ammonium nitrogen content and significantly lower nitrate content at other altitudes (Figure 14A,B). At the same time, the contents of TN and MBN and the abundance of Nitrospirillum in the soil at the altitude of 5200 m were significantly higher than those at other altitudes (Figure 14C,D), possibly due to the differences in biomass and coverage of grasslands and human disturbance. This also led to a significantly higher abundance of bacteria associated with the soil nitrogen cycle, including nitrifying and denitrifying bacteria. This indicates that the soil nitrogen content is higher and the activity of related bacteria is more vigorous at the altitude of 5200 m, which may cause the soil N indicators more sensitive to the impact of the highway at high altitudes. The abundance of nitrifying bacteria is proportional to that of denitrifying bacteria. This result is in accordance with the law expounded by the predecessors [60].

**Figure 14.** N cycle related indicators: (**A**) total soil nitrogen content, (**B**) soil microbial nitrogen content, (**C**) soil ammonia nitrogen N-NH4, (**D**) soil nitrate N-NO3, (**E**) nitrification bacteria relative content, (**F**) the relative content of denitrification bacteria at 4000 m (L), 4600 m (M), and 5200 m (H). \* Means *p* < 0.05 for significance test; \*\* means *p* < 0.01 for significance test; \*\*\* means *p* < 0.001 for significance test.

The increase of soil available phosphorus may be affected by many factors. In alkaline soils, the content of calcium ions is usually high in alkaline or calcareous soils, and phosphate ions are easy to form calcium phosphate precipitation with calcium ions, thus reducing the content of available phosphorus. In this study, the pH value of the road disturbance area decreased, which promoted the dissolution of the P element in alkaline soil into AP and increased soil AP. Simultaneously, the abundance of some bacteria (such as *Bradyrhizobium* and *Mycobacterium*) related to phosphorus dissolution increased in the disturbed area of the highway, which may also be the reason for the significant increase of available phosphorus content. Ma et al. (2013) recorded relatively rich AP levels on the slopes of the Beijing-Chengde Phase III Expressway [61], consistent with the findings of this research. The ratio of carbon to nitrogen increased significantly in the disturbed zone.

Many heavy metals have been proved by previous researches to enter the soil on both sides of the highway from vehicle mechanical wear and fuel consumption [11,62]. Our findings are consistent with their results. The Zn content in the study area did not increase significantly and was generally lower than the background value of Tibet in China [23]. Parts of the vehicle that are galvanized or contain zinc, such as fuel tanks and tires, may be a source of zinc contamination [63]. This may be due to the fact that there is little human activity other than the normal driving of vehicles in the study section, and there is little exposure and wear of galvanized parts. The research shows that the enrichment of heavy metals in the 4000 m area is more obvious in the highway disturbance area, and the overall content is higher, which may be related to the differences in traffic flow and vegetation types. Heavy metals might be transferred to the soil along the highway through airflow or pavement runoff [64,65]. In this study, the range of significant increase of heavy metal content along the highway was mainly limited to 25–50 m from the highway, which may be related to the location of the transect in this study. This study area is selected in the no man's land and nature reserve without other human activities except roads, with few people and no grazing except for normal vehicle driving and road maintenance. Wildlife activity is also relatively low, which may lead to weak diffusion of heavy metal particles.

#### *4.2. Effects of Environmental Factors on Soil Bacterial Community Structure at Different Altitudes*

Therefore, the construction and traffic of the highway will significantly change the vegetation growth, soil physical, and chemical properties and content in grassland soil of heavy metal along the highway, thus affecting the diversity and structure of soil microbial communities. For example, Li et al. reported that the species composition of soil bacteria changed significantly during the degradation of vegetation [66]. Kang et al. studied the microbial community in an alpine wetland ecosystem and showed that human-induced pH changed microbial diversity and community structure in the upper soil layer as the main driving factors [67]. Underground coal mining has been reported to cause changes in soil conductivity and water content in sandy areas of western China, which can change the structure and diversity of soil microbial communities [68]. Research on the stressed soil near the QTH and Qinghai-Tibet Railway (QTR) proved heavy metals were important factors in the formation of bacterial community diversity [26]. In the Changbai mountain area, the disturbance of the highway was studied alongside the microbial diversity and communities in turf marsh soils, with the result that the microbial composition of turf marsh soils were mainly changed by road runoff and heavy metals emission [14]. In this study, the sequence data obtained indicate that the microbial communities differ significantly at the genus level.

In this study, the main dominant species at the phylum level are consistent with many previous studies (Figure 15A) [14,69,70]. In general, in disturbed or stressed soils, the abundance of *Proteobacteria* significantly decreased, while *Acidobacteria* abundance significantly increased compared with the undisturbed grassland [14,71]. In this study, this change was only consistent and significant in some areas, which may be related to the different degrees of environmental factors in different areas (Figure 15B). The ratio of *Proteobacteria* to *Acidobacteria* has been proved to be an indicator of changes in soil environmental conditions by several scholars [18]. This ratio between the road disturbed

area and the control area was significantly different at three sites, with a significant increase at the 4000 m site and a significant decrease at the 5200 m site, indicating that the soil environmental conditions in the road-disturbed area have indeed changed in this study. He et al.'s research shows that the significant decrease of this ratio in the grassland affected by the highway is related to the fact that the altitude of the study area is much lower than 4000 m. This can be explained by geographical location and soil properties. *Alistipes* and *Barnesiella* are derived from faeces and are more abundant in unaffected areas. *Barnesiella* is in a similar situation.

**Figure 15.** *Cont*.

**Figure 15.** Circos circular plot of the bacterial community (**A**) at the phylum level and (**B**) at the genus level.

At the genus level, the changes of different bacteria in this study showed a variety of patterns, which may be related to their characteristics. For example, *Flavisolibacter* was isolated from automotive air conditioners in previous studies [70]. Its significant enrichment occurred within 20 m from the road, which may be due to the movement of the vehicle. The *Blautia* abundance significantly increased in the control which was isolated from animals and humans [69,72]. This could be due to the fact that the vegetation in the unaffected area remained in the original state, the quality of grassland was high, and there were more wildlife activities. It could be seen that due to the impact of highway construction and operation, the activities of wildlife were reduced in the area near the highway, as were *Alistipes* and *Barnesiella*, which may also come from animal dung [73–75]. This was also confirmed by the increase in abundance in the unaffected areas. Abundance of *Microvirga*, *Skermanella* and *Crossiella* correlates with pH [76]. Due to the optimum pH (7.0) for growth, the abundance of those was negatively related to pH in the alkaline environment of the study area, which was also confirmed by the results of this research. Consistently in this study, a significant positive correlation was found between *Pseudomonas* abundance and N content [77]. Another study proved [78] that *Pseudomonas* abundance in the soil samples of the non-plant area was significantly enriched compared with the composite

rhizosphere soil samples. Consistent with previous studies, the vegetation coverage of the 20 m treatment declined significantly, while *Pseudomonas* abundance significantly increased. Liu et al. reported that the *Pseudomonas* can promote the degradation of weathered diesel-oil pollutants to some extent. The increased abundance of *Pseudomonas* in the area near the highway in this study may be related to its adaptability to diesel pollutants caused by the highway. Consistent with previous studies, *Marmoricola* abundance correlated with soil carbon content positively [8]. There was a reported significant positive correlation between *Nocardioides* abundance and soil phosphorus content [79], which is consistent with the enrichment of soil-available phosphorus in this study. At 5200 m and 4600 m, soil-available phosphorus increased significantly near the highway. In addition, different bacteria responded differently to heavy metals, and those sensitive to toxicity decreased, while resistant bacteria could adapt to environmental changes and their relative abundance increased. For example, the relative abundance of *Flavobacterium*, *Gemmatimonas*, *Terrimonas*, *Nitrospira,* and *Bacillus* had increased, while *Barnesiella* and *Blastococcus'* relative abundance decreased in metal-enriched soil [14,80,81]. Our results are consistent with these observations. *Flavobacterium* and *Pseudomonas* were more significantly enriched in the disturbed area at the height of 4000 m. Accordingly, the content of heavy metals at 4000 m is relatively high. This indicated that heavy metals had an impact on the soil bacterial community structure. *Prevotella* 9, *Blautia*, *Marmoricola,* and *Barnesiella* were significantly enriched in the control group and negatively correlated with the concentration of heavy metals, which were sensitive to heavy metals. Consistent with the results of this study, *Marmoricola* abundance was shown to be significantly negatively correlated with soil heavy metal content [82]. *Gemmatimonas*, *Microvirga*, *Massilia*, *Sphingomonas,* and *Blastococcus* were significantly enriched in the highway disturbed area, and were positively correlated with heavy metals, indicating that they were heavy metal tolerant bacterium [83–85].

#### *4.3. Changes in Soil Bacterial Functional Groups under the Influence of Highway*

FAPROTAX analysis showed that the functional groups with significant changes in different sites were mainly concentrated in bacteria related to soil nitrogen cycling. Ammonia oxidation, nitrite oxidation, nitrification and other soil nitrifying bacteria in the disturbed area from the road showed an upward trend compared with the control. Ammonia oxidation by ammonia-oxidizing bacteria is the first step in nitrification and the rate limiting step [60]. The conversion of ecosystem types may have important potential impacts on soil microorganisms involved in ammonia oxidation, and there is a significantly positively correlation between the number of soil ammonia-oxidizing bacteria and soil NH4 +-N content [86]. NH4 +-N significantly affected the population composition of soil ammonia-oxidizing bacteria in previous studies [80]. The results showed that the content of soil ammonium nitrogen increased significantly after the moderate degradation stage of alpine meadow, and the soil ammonium accumulation was obvious in the later stage of grassland degradation. Consistent with previous studies, the number of aerobic nitrite oxidation bacterial at 5 m significantly increased compared with the control in the 4600 m site, which was similar to the change trend of soil ammonium nitrogen. The increase of soil ammonium nitrogen was the main reason for the enrichment of ammonia-oxidizing functional genes. Nitrite oxidation is the second step of nitrification. At the 5200 m and 4600 m sites, the number of nitrite-oxidizing bacteria increased significantly at 5 m sites compared with the control, which may be due to the accumulation of ammonium nitrogen in the degraded grassland affected by the highway, which promoted ammonia oxidation, thus providing a material basis for the nitrite oxidation process. However, the content of nitrite in the soil was not measured in this study, so the specific reasons for this process need further study.

In the process of soil denitrification, the bacterial communities related to nitrate reduction decreased significantly in the sites within 100 m from the road, which was opposite to the change of nitrification. Nitrate reduction and nitrite ammonification are two important processes of soil denitrification. Nitrate reduction is the first step of denitrification. Nitrate-reducing bacteria decreased significantly in the disturbed area of the highway, and their abundance was positively correlated with distance. There is a positive correlation between distance and nitrate bacteria. In the road disturbed area, the content of soil nitrate decreased, the material base of nitrate-reducing bacteria decreased, and the substrate of nitrate reduction reaction decreased, so the abundance of nitrate-reducing bacteria decreased. Nitrate bacteria relative abundance significantly correlated with altitude, and the same correlation occurs between nitrate-reducing bacteria and altitude. The reason is that with the increase of altitude, the oxygen content and the relative abundance of anaerobic nitrate-reducing bacteria increase, and the nitrate reduction reaction becomes more intense, consuming nitrate, therefore the content of nitrate decreases.

In the disturbed areas of 4600 m and 5200 m, the content of nitrate decreased and the content of ammonium nitrogen increased, indicating that soil denitrification was dominant. These might be due to the deterioration of soil aeration caused by the slight degradation of grassland, the decrease of vegetation coverage, the increase of soil compaction and the compression of soil pore structure [81], and then the promotion of the growth of anaerobic denitrifying bacteria. Firstly, because of the decrease of nitrate content, nitratereducing bacteria, and the reaction product nitrite and the nitrogen-fixing bacteria, nitrous oxide denitrifying bacteria increases, and part of nitrogen in soil enters the air in the form of nitrous oxide and nitrogen. This also confirms the results of the decrease of total soil nitrogen and AN content. Nitrogen-fixing bacteria decreased, but the content of ammonium nitrogen increased, indicating that the way of denitrification to produce ammonium nitrogen may be mainly through nitrite ammonification [87]. The decrease of SM proved it could significantly increase nitrite-oxidizing bacteria abundance, which also confirmed the significant increase of nitrite-ammonifying bacteria in this study. This may also be due to the deterioration of soil aeration, which reduces the performance of soil respiration, weakens nitrogen fixation, and promotes the production of anaerobic bacteria in the soil. According to Yu et al., in farmland soil, the abundance of denitrifying bacteria and nitrogen-fixing bacteria decreased significantly due to the decrease of nitrogen content after long-term nitrogen application [88]. It was further confirmed that the growth of denitrifying bacteria was limited by the decrease of soil nitrogen content. On the contrary, the nitrate and ammonium nitrogen decreased significantly at 4000 m, which may be due to the more serious grassland degradation and the overall decrease of nitrate and ammonium nitrogen. At the same time, denitrification-related bacteria also decreased in the disturbed area and significantly reduced. Combined with the results of vegetation index, it can be seen that the degradation degree of grassland in the highway disturbance area of the 4000 m site is higher, referring to the classification standards commonly used by predecessors [86], to the extent of moderate or severe degradation. Studies have shown that the soils C and N in severely degraded grassland are significantly reduced. The abundance of bacteria related to the nitrogen cycle showed a downward trend because of the lack of material basis for growth.

In addition, the chemoheterotrophic bacterial community showed a significant upward trend within 50 m from the road, and showed a consistent rule in the three areas, as the similar trend of previous studies. Sulfide respiration bacteria in the road disturbance area showed a downward trend compared with the control, and the lowest value (50 m) was found in the moderately degraded grassland soil, which was consistent with previous studies.

#### *4.4. Relative Contribution of Vegetation, Soil, and Spatial Factors to Bacterial Community Structure in Different Altitudes*

Mantel test results showed that biomass significantly affected the soil bacterial community structure (R = 0.241, *p* = 0.001). Among the soil factors, SOC, TN, AN, and NH4 were the main driving factors for the difference of soil bacterial community structure. AK and SM were the main driving factors for the difference of soil bacterial functional groups. This study found that soil nutrients are the most critical. As the most important nutrients for organisms, SOC and TN have been proved by many studies to be significantly related

to soil microorganisms in different ecosystems [89,90]. Due to the anoxic soil conditions in the QTP, the decomposition of organic matter is slow. The lack of available nutrients, especially in areas disturbed by roads, limits the growth of bacteria. According to the traditional niche theory, the niches of many organisms may be related to available nitrogen, which causes impact on their coexistence in the ecosystem and changes the community structure [91]. Changes in TN, AN, and NH4 can lead to changes in niche size, which can alter bacterial diversity [92].

According to the results of redundancy analysis (RDA), altitude played an important role in affecting the soil bacterial community structure. The responses of soil bacterial communities to environmental factors at different altitudes were further discussed, and the main driving factors of soil bacterial communities varied with different altitudes. For example, in the 4000 m site, the content of Zn and Pb were important factors driving the change of the soil bacterial community. Consistently, the soil heavy metal content was higher at 4000 m. This shows that the heavy metal pollution is more serious in the 4000 m environment of the disturbed area. The toxicity of heavy metals leads to the change of the soil bacterial community structure. The results of the RDA analysis at different altitudes also proved that the distance from the road greatly affected the structure of soil bacterial communities. The community structure and functional groups of soil bacteria were significantly changed by the changes of grassland vegetation and soil in the disturbed area of the highway.

In general, the effects of the highway on soil bacterial community structure and functional groups in the disturbed area are multifaceted, which is the result of the joint action of vegetation and soil. The effects of soil-vegetation coupling on the bacterial community structure and function were greater than the effects of soil-vegetation coupling alone. This phenomenon has also been reflected in previous studies. This shows that the influence of the environment on the growth of soil bacteria is a very complex process. Road traffic has a significant impact on all aspects of environmental factors, and the resulting changes in soil bacterial communities need to be explained in many ways. Therefore, further research is of great need in this area.

### *4.5. Uncertainty and Perspectives*

Our study describes above-ground vegetation, soil chemistry, and bacterial communities at different elevations along the QTH. In addition, the driving mechanism of microbial community diversity and community structure changes in the highway-disturbed area of the QTP explored.

Several shortcomings of this study need to be considered. First, there is no repetition in the same altitude environment; however, the results of the study at different altitudes are generally consistent and can be verified by each other, which does not affect the reliability of our conclusions. Secondly, due to legal and policy requirements, we have no way to obtain specific traffic flow of data in the sampling zone. This might also play an important role in the change of grassland vegetation and soil bacteria. By further measuring more comprehensive indicators related to the soil nitrogen cycle (such as nitrate, nitrite, etc.), the response mechanism of bacterial functional groups related to the soil nitrogen cycle can be more clearly expounded. These problems deserve further study.

The impact of highway construction and transportation on soil microbial diversity of the ecosystem along the highway has attracted more and more attention [93–95]. A recent study found that soil nitrogen was a key factor driving changes in soil microbial biomass and enzyme activity in cutting slopes [96]. Interactions between potentially toxic substances from vehicle emissions, roadside soils, and associated biota have also recently been reviewed [97]. Soil nutrients such as TOC and TN were found to be the most important variables affecting soil bacterial diversity and community structure along the QTH in our research which was significant for the research of the ecological restoration process along the plateau highway in the future.
