Responses of Nitrogen-Fixing Bacteria Communities to Elevation, Season, and Slope Aspect Variations in Subtropical Forests of Yunnan, China

: Nitrogen-ﬁxing bacteria play a signiﬁcant role in tropical forest ecosystems. However, little is known about the comprehensive effects of altitude gradient (1000–2600 m), seasons (October, January, April, and July), and slope aspects (east and west) on the abundance and diversity of nitrogen-ﬁxing bacteria in subtropical forest. Q-PCR and PCR-DGGE methods were performed to explore the abundance and diversity of nitrogen-ﬁxing bacteria, respectively, in the Ailao Mountain subtropical forest. Our results showed that the abundance of nitrogen-ﬁxing bacteria was highest in October and December, whereas it was lowest in April and July. Moreover, there was no difference in the total number of soil nitrogen-ﬁxing bacteria on the eastern and western slopes. The diversity of soil nitrogen-ﬁxing bacteria is higher at low and medium altitudes, but lower at high and medium altitudes with increasing altitude, and similar variation in the eastern and western slopes as well. Moreover, the most inﬂuential factors affecting the abundance of nitrogen-ﬁxing bacteria was NH 4+ -N and herbal coverage, while those most affecting the diversity of nitrogen-ﬁxing bacteria were NH 4+ -N and NO 3 − -N. In addition, permutational multivariate analysis demonstrated that the season had the greatest effects on the abundance of nitrogen-ﬁxing, whereas altitude had the greatest effects on the diversity of nitrogen-ﬁxing bacteria. These ﬁndings provide evidence that the variation in nitrogen-ﬁxing bacteria is affected by multiple factors (altitudes, seasons and slope aspects) in the subtropical forests of Yunnan, China.


Introduction
Biological nitrogen fixation (BNF) is an important process for maintaining the nitrogen supply in the terrestrial ecosystem and one of the principal pathways of nitrogen cycling in the ecosystem [1,2]. Diazotrophs reduce atmospheric nitrogen, via BNF, into ammonia for plant growth needs. The nifH gene is the best biomarker for assessing the community structure of nitrogen-fixing microorganisms [3]. Previous studies have shown that the abundance of nifH gene differs greatly in different ecosystems (e.g., forests [3][4][5], ocean [6][7][8], farmland [9], freshwater lakes [10,11], grassland [12], facility gardening [13][14][15], and other special habitats [16]), and the nitrogen-fixing bacteria in forest soil systems are the most abundant, followed by farmland and horticultural facility soil systems. Moreover, there were differences in the number of copies of nifH genes in different forest systems. Some studies have found that the number of nifH gene copies in subtropical forests [4] is greater than that in rainforests [17]. In addition, diverse vegetation types also affect the number of nitrogen-fixing bacteria in subtropical forest soils, such as subtropical mixed forests [18] The study was conducted on Ailao Mountain, which is located in the Xujiaba area of the Ailao Mountain National Nature Reserve, Yunnan, China (1200-2600 m, 24 • 00 N-24 • 44 , 100 • 54 -101 • 29 ). The annual precipitation in this area is 1931.1 mm, the sunshine duration is 1239 h, and the frost-free period is 180 d. The dry and rainy seasons were distinct due to the precipitation during the rainy season (May-October), accounting for approximately 85% of the annual precipitation. The monthly temperatures of the hottest (July) and coldest (January) months were 16.4 • C and 5.4 • C, respectively. Vegetation was dominated by Fagaceae, followed by Lauraceae, Magnoliaceae, and Theaceae, and the soil type was montane brownish yellow soil.
The sample plots were set up on the eastern and western slopes of Ailao Mountain at an altitude of 1000-2600 m, with 10 m × 10 m sample plots set up at an altitude interval of 200 m (Table 1). Forty-two sample plots, with seven sample plots in each slope direction, were set up and repeated three times. The sampling period was from October 2015 to July 2016 during the coldest month (January), dry season (April), hottest month (July), and rainy season (October). Five topsoil cores (0-20 cm) were collected from each sample plot after removing the current plant litter, small stones, and other residues, and mixed as a composite sample. All samples were packed in plastic bags and transported to the laboratory in containers of ice, passed through a 0.25 mm soil sieve, and stored at −80 • C for further analysis.

Soil Analysis and DNA Extraction
The soil water content was measured by drying the soil at 105 • C to a constant weight. The soil pH was measured using a soil-to-water ratio of 1:5 (5 g of soil and 25 mL of water were mixed and left at 22 ± 3 • C for 15 min) [35]. The total N, total P, NH 4 + , and NO 3 − were measured using a SmartChem 200 Analyzer using standard methods. The total K was analysed by atomic emission spectrometry on an AA-6300C flame photometer. The soil organic-matter content was determined using the K 2 Cr 2 O 7 oxidation method. The total soil DNA was extracted and purified using a MoBio Power Soil DNA extraction kit (MoBio, Carlsbad, CA, USA) with 0.20 g of soil. Three replicate DNA extractions per sample were used, and they were pooled to obtain sufficient DNA quantity and were stored in a refrigerator at −80 • C.

Quantitative Real-Time PCR
Absolute quantification of nifH genes was performed using the LightCycler ® 480 II System (Roche, Basel, Switzerland). PCR amplification was performed using the primer pair polF (5 -GCTGCGAYCCSAARGCBGACTC-3 ) and AQER (5 -GACGATGTAGATYTCCTG-3 ) targeting the nifH gene. The total volume for each reaction was 20 µL, including 1 µL DNA, 0.5 µL forward and reverse primers, and 10 µL SYBR. The specificity of the product was confirmed by melting curve analysis (65-95 • C, 0.5 • C per read, 5 s hold). The plasmid concentration was determined using a spectrophotometer (AA-6300C, Japan), and the copy number of the nifH gene was directly calculated from the extracted plasmid DNA concentration. Plasmids with known plasmid gene copy numbers were consecutively diluted 10 times to obtain a standard curve of more than seven orders of magnitude (6.54 × 10 3 to 6.54 × 10 9 copies) each time. A high amplification efficiency of 101.1-103.4%, R 2 = 0.99, and slope from −3.2-−3.5 was obtained. The threshold cycle (Ct) values obtained for each sample were compared with the standard curve to determine the initial copy number of the target gene. Each experiment was repeated three times.

PCR and DGGE Community Fingerprints
The extracted DNA was amplified using primers polF-GC/AQER. The total reaction volume used was 50 µL, which contained 2 µL of template DNA, 1 µL of forward and reverse primers, 25 µL of Premix ExTaq TM (Takara, Dalian, China), and 21 µL of ddH 2 O. The PCR program was carried out at 94 • C for 5 min (94 • C for 45 s, 57 • C for 45 s, and 72 • C for 1 min) for 35 cycles, and 72 • C for 10 min. PCR products were electrophoresed on a 1.2% agarose gel in 1% TAE and observed under UV light using nucleic acid dyes. DGGE analysis was performed using 8% polyacrylamide 100% demonized formamide gel and 8% polyacrylamide 0% demonised formamide gel, with a denaturation gradient of 40-60%. The DGGE gel was run in 1× TAE at 60 • C and 50 V for 12 h. After electrophoresis, the gel was stained with silver nitrate, placed in a lamp box, and imaged using a digital camera.

Statistical Analysis
The DGGE gel images were analysed using Quantity One 4.6 software. SPSS 22.0 was used for variance and correlation analyses. Canno 4.5 software was used for principal component analysis (PCA) and canonical correspondence analysis (CCA). The R programming language was used for heat map analysis, and the sequences were analysed using the BLAST program in GenBank to find the closest sequence matches. MEGA 7 was used to conduct a phylogenetic analysis. The sequences obtained were deposited in GenBank under accession nos. OK513414-OK513437.

Vegetation Conditions and Soil Physical-Chemical Characteristics
It was found that the vegetation at different elevations in Ailao Mountain was a secondary vegetation type in which the low-altitude vegetation composition was mainly arbor layer + herb layer and the high altitude was mainly the arbor layer + shrub layer + herb layer; moreover, the herb layer at 2600 m above sea level was mainly frozen soil moss. The coverage of trees in different seasons first increased and then decreased with the elevation, on the contrary, the shrub coverage decreased with the elevation.
The pH of the east-slope soil did not change significantly, whereas the pH of the west-slope soil decreased with increasing altitude ( Table 2). The soil water and organic matter content increased monotonously, whereas the total N, total P, NH 4 + -N, and NO 3 − -N increased and then decreased as the altitude gradient increased. The total K of the soil on the eastern slope showed an evident downward trend with the altitude gradient, and there was no distinct change on the western slope. In addition, the altitude gradient was positively correlated with the soil water content, organic matter, total N, NH 4 + -N, NO 3 − -N (p < 0.01), and total P (p < 0.05) and negatively correlated with the pH value and total K (p < 0.01).

Quantitative Real-Time PCR Assay
The nifH gene abundance of nitrogen-fixing bacteria in the soil ranged from 8.87 × 10 5 to 8.27 × 10 6 copies/g (Figure 1). For the east-slope soils, as the elevation gradient increased, the abundance of nitrogen-fixing bacteria increased gradually in October, whereas in April and July, it decreased (p > 0.05), while no significant differences were observed in January. For the west-slope soils, as the elevation gradient increased, the abundance of nitrogenfixing bacteria increased gradually in October but decreased in April and July. In different seasons, the abundance of nitrogen-fixing bacteria on the eastern and western slopes was similar in the order of winter > autumn > spring = summer. Moreover, the abundance of nitrogen-fixing bacteria in autumn on the western slope was significantly higher than that on the eastern slope (p < 0.01). There was no difference in the total number of soil nitrogen-fixing bacteria on the eastern and western slopes. In addition, the total amount of soil nitrogen-fixing bacteria was negatively correlated with the herbal coverage (p < −0.369) and significantly positively correlated with soil NH 4 + content. A, B, C and D: Oct., Jan., Apr., Jul. respectively; E and W: East and West, respectively.

Quantitative Real-Time PCR Assay
The nifH gene abundance of nitrogen-fixing bacteria in the soil ranged from 8.87 × 10 5 to 8.27 × 10 6 copies/g (Figure 1). For the east-slope soils, as the elevation gradient increased, the abundance of nitrogen-fixing bacteria increased gradually in October, whereas in April and July, it decreased (p > 0.05), while no significant differences were observed in January. For the west-slope soils, as the elevation gradient increased, the abundance of nitrogen-fixing bacteria increased gradually in October but decreased in April and July. In different seasons, the abundance of nitrogen-fixing bacteria on the eastern and western slopes was similar in the order of winter > autumn > spring = summer. Moreover, the abundance of nitrogen-fixing bacteria in autumn on the western slope was significantly higher than that on the eastern slope (p < 0.01). There was no difference in the total number of soil nitrogen-fixing bacteria on the eastern and western slopes. In addition, the total amount of soil nitrogen-fixing bacteria was negatively correlated with the herbal coverage (p < −0.369) and significantly positively correlated with soil NH4 + content.

DGGE Analysis
Denature-gradient gel electrophoresis analysis results of nifH PCR products from the Ailao Mountain nitrogen-fixing bacteria is shown in Figure 2a. Significant differences were found in the soil nitrogen-fixing bacterial community structure at different altitudes and seasons. For different altitudes, the diversity of soil nitrogen-fixing bacteria first decreased and then increased with increasing altitudes. Higher diversity was observed at low altitudes and high altitudes, and lower diversity at moderate altitudes. In different seasons, the diversity of soil nitrogen-fixing bacteria showed a trend of first increasing and then decreasing. In addition, the diversity of soil nitrogen-fixing bacteria in winter (October and January) and summer (July) was greater than that in April. As shown in Figure 2b, the diversity indices of nifH bacteria on the eastern and western slopes were similar, and the seven altitude gradients were grouped into three subgroups: H5/6, H3/4, and H1/2/7.
The relationship between the soil parameters and diversity of soil nitrogen-fixing bacteria was studied using redundancy analysis (RDA). RDA results (Figure 2c) showed that NH 4 + -N, NO 3 − -N, and K had an important effect on the soil nifH community in sequence, whereas less significant effects were observed for organic matter and soil moisture content. In addition, the total number of soil nitrogen-fixing bacteria had no correlation with the herbal (p = −0.244) and arbor coverage (p = 0.163). similar, and the seven altitude gradients were grouped into three subgroups: H5/6, H3/4, and H1/2/7.
The relationship between the soil parameters and diversity of soil nitrogen-fixing bacteria was studied using redundancy analysis (RDA). RDA results (Figure 2c) showed that NH4 + -N, NO3 − -N, and K had an important effect on the soil nifH community in sequence, whereas less significant effects were observed for organic matter and soil moisture content. In addition, the total number of soil nitrogen-fixing bacteria had no correlation with the herbal (p = −0.244) and arbor coverage (p = 0.163).

Permutational Multivariate Analysis
The permutational multivariate analysis showed that the slope direction, season, and altitude had significant effects on the copy number of nitrogen-fixing bacteria, with the season having the greatest effect (η 2 = 0.991), followed by the altitude and slope direction. These were also significant effects on the diversity of nitrogen-fixing bacteria (Shannon index), with the altitude having the greatest impact (η 2 = 0.960), followed by the season

Permutational Multivariate Analysis
The permutational multivariate analysis showed that the slope direction, season, and altitude had significant effects on the copy number of nitrogen-fixing bacteria, with the season having the greatest effect (η 2 = 0.991), followed by the altitude and slope direction. These were also significant effects on the diversity of nitrogen-fixing bacteria (Shannon index), with the altitude having the greatest impact (η 2 = 0.960), followed by the season and slope direction (Table 3). Partial eta square (η 2 ) is used to quantify the effect. The significance was examined by F-test based on sequential sum of square from 999 permutations of the nifH copies (Shannon index). Significant levels: ** p < 0.01, *** p < 0.001.

Sequences Analysis
To obtain the DGGE map results for the nifH gene of soil nitrogen-fixing microorganisms at different altitudes, the main bands were selected and sequenced. Twenty-four different genes with similar sequences were obtained and blasted in the NCBI database. The results presented that the obtained band sequences belonged to uncultured bacterium nitrogenase genes. Comparing the similarity of the nifH gene sequences with those of known species, the similarity was 91-99% (Table 4). All the nifH gene sequences were grouped into six clusters. The majority of the sequenced clones were not closely related to any known cultivated nitrogen-fixing bacteria (Figure 3).

Discussion
The permutational multivariate analysis showed that the slope direction, season, and altitude had significant effects on the copy number of nitrogen-fixing bacteria, with the season having the greatest effect (η 2 = 0.991), followed by the altitude and slope direction. In this study, the abundance of nitrogen-fixing bacteria on the eastern and western slopes was similar in the four seasons in the order of winter > autumn > spring = summer. The abundance of nitrogen-fixing bacteria was highest in winter, which is inconsistent with some research results [41][42][43]. Although some studies have shown that the abundance of nitrogen-fixing bacteria in subtropical forests is the highest in summer [41,42], others have reached the opposite conclusion [43]. Comparable results have also been observed in different ecosystems such as farmland [31,37], lakes [7,32], and grasslands [33,39]. Moreover, the abundance of nitrogen-fixing bacteria varies with altitude gradient in different seasons. As the elevation gradient increased, the abundance of nitrogen-fixing bacteria increased gradually in October but decreased in April and July. The results of our study are inconsistent with the results of several other studies [21][22][23], although some studies found hump-shaped [23] or monotonically decreasing or increasing [21] patterns of nitrogen-fixing bacterial abundance with increasing elevation. These results may be due to different vegetation types, soil conditions, climate conditions, and ecosystems. Our results also suggest that plant coverage is significantly negatively correlated with nitrogen-fixing bacterial abundance, especially herb coverage (r = −0.369, p < 0.005). Compared to other soil microorganisms, bacteria rely more on plant litter as a substrate [46]. Therefore, the opposite change characteristics were observed with herb coverage. It was found that the levels of C and N influenced the activity and distribution of nitrogen fixation bacteria [26]. In our study, nitrogen-fixing bacterial abundance was significantly positively correlated with the soil NH4 + content, which indicated that the NH4 + content might be the key factor influencing the nitrogen-fixing bacterial abundance. NH4 + -N rather than NO3 − -N was the main form of inorganic nitrogen in the subtropical forest soil of Ailao Mountain, which may be because soil acidity inhibits nitrification, resulting in a low NO3 − -N content. It has been reported that low available nitrogen could provide selective advantages to diazotrophs, thus enhancing the diazotrophic count to a level that could be sustained with other available nutrient resources [30].

Discussion
The permutational multivariate analysis showed that the slope direction, season, and altitude had significant effects on the copy number of nitrogen-fixing bacteria, with the season having the greatest effect (η 2 = 0.991), followed by the altitude and slope direction. In this study, the abundance of nitrogen-fixing bacteria on the eastern and western slopes was similar in the four seasons in the order of winter > autumn > spring = summer. The abundance of nitrogen-fixing bacteria was highest in winter, which is inconsistent with some research results [41][42][43]. Although some studies have shown that the abundance of nitrogen-fixing bacteria in subtropical forests is the highest in summer [41,42], others have reached the opposite conclusion [43]. Comparable results have also been observed in different ecosystems such as farmland [31,37], lakes [7,32], and grasslands [33,39]. Moreover, the abundance of nitrogen-fixing bacteria varies with altitude gradient in different seasons. As the elevation gradient increased, the abundance of nitrogen-fixing bacteria increased gradually in October but decreased in April and July. The results of our study are inconsistent with the results of several other studies [21][22][23], although some studies found hump-shaped [23] or monotonically decreasing or increasing [21] patterns of nitrogen-fixing bacterial abundance with increasing elevation. These results may be due to different vegetation types, soil conditions, climate conditions, and ecosystems. Our results also suggest that plant coverage is significantly negatively correlated with nitrogen-fixing bacterial abundance, especially herb coverage (r = −0.369, p < 0.005). Compared to other soil microorganisms, bacteria rely more on plant litter as a substrate [46]. Therefore, the opposite change characteristics were observed with herb coverage. It was found that the levels of C and N influenced the activity and distribution of nitrogen fixation bacteria [26]. In our study, nitrogen-fixing bacterial abundance was significantly positively correlated with the soil NH 4 + content, which indicated that the NH 4 + content might be the key factor influencing the nitrogen-fixing bacterial abundance. NH 4 + -N rather than NO 3 − -N was the main form of inorganic nitrogen in the subtropical forest soil of Ailao Mountain, which may be because soil acidity inhibits nitrification, resulting in a low NO 3 − -N content. It has been reported that low available nitrogen could provide selective advantages to diazotrophs, thus enhancing the diazotrophic count to a level that could be sustained with other available nutrient resources [30].
Altitude had the greatest impact on the diversity of nitrogen-fixing bacteria (Shannon index) (η 2 = 0.960), followed by the season and slope direction. With the increase in altitude, the diversity of soil nitrogen-fixing bacteria was higher at low (H1, H2) and medium altitudes (H3, H4), but lower at high medium altitudes (H5, H6). The observed diversity pattern was similar to that of nitrogen-fixing bacteria in Gongga Mountain, which showed that the diversity of nitrogen-fixing bacteria did not change significantly in the range of 1800-2800 m but decreased with an increase in altitude in the range of 2800-4100 m [24]. This might be due to the fact that under relatively mild temperature and humidity conditions, high plant diversity in low-and medium-altitude areas would maintain high biodiversity [47]. However, the correlation analysis showed that there was no correlation between the diversity of nitrogen-fixing bacteria and vegetation. This may be because the scale of the altitude gradient selected by our experimental plot was not sufficiently large. Remarkably, the diversity of diazotrophs at the highest elevation (H7) was also high. A possible explanation is that altitude changes have significant effects on abiotic factors such as temperature, precipitation, and soil properties, which could shape nitrogen-fixing bacterial communities. Moreover, a higher diversity of soil nitrogen-fixing bacteria was found in autumn (October) and winter (January), which indicated that low temperature conditions could increase the diversity of nitrogen-fixing bacteria in the Ailao Mountain soil. All nifH gene sequences obtained in our study belonged to uncultured bacteria. A previous study showed that diverse functional free-living psychrotrophic diazotrophs were isolated only from high-altitude Gangotri soils [25]. This implies that more psychrotrophic diazotrophs may exist in the Ailao Mountain soil. Furthermore, high plant species richness may provide more abundant and diverse root exudates and leaf litter into soils to support higher diazotrophic diversity [48,49]. However, there was no significant correlation between the nitrogen-fixing bacteria diversity index and tree or herb coverage in our study. The most influential factors in the soil nifH community were NH 4 + -N, NO 3 − -N, and K. It has been reported that higher soil TC and TN contents can result in a higher diazotrophic diversity [24]. In addition, soil TC and TN can regulate microbial nitrogen and carbon use efficiencies and maintain microbial element homeostasis [50]. As mentioned above, the functional traits of soil microbes may be influenced more by evolutionary history than by environmental variables, even across a broad environmental gradient [51].
In some environments (Israel's arid Mediterranean and Gannan alpine meadow) [52,53], the structure and composition of soil microbial communities are affected by the slope direction; however, none of the studies have evaluated the effect of slope aspects on soil microbial communities in forests. Although the influence of the slope direction is second only to that of the altitude and season in our study, it is evident that the abundance and diversity of nitrogen-fixing bacteria were affected by the different slopes of Ailao Mountain. Part of this variation can be explained by changes in environmental indicators, since the spatial variation in slope direction is an important determinant of the community composition, species distribution, and ecosystem process [54][55][56]. In different seasons, the abundance of nitrogen-fixing bacteria on the eastern and western slopes was similar in the order of winter > autumn > spring = summer. In particular, the abundance of nitrogenfixing bacteria in autumn on the western slope was significantly higher than that on the eastern slope (p < 0.01). It may be due to the huge mountain body of Ailao Mountain being the main reason for the significant difference in regional environmental factors and vegetation variation on both of its sides [57]. Furthermore, there was no difference in the total abundance of soil nitrogen-fixing bacteria on the eastern and western slopes, while the diversity indices of nifH bacteria on the eastern and western slopes with elevation variation were similar. The effect of changes in soil moisture, near-surface air temperature, soil temperature, species distribution, and vegetation composition on the slope aspect gradient were similar to those in longitude and altitude gradients on a large scale [58]. Hence, as a topographic factor, the slope aspect is an indirect environmental factor that leads to differences in the composition and structural characteristics of plant species, affects the survival and growth of plants, and further affects the abundance and composition of soil nitrogen-fixing bacteria.

Conclusions
Our results showed that the abundance and diversity of soil nitrogen-fixing bacterial communities in subtropical forests are affected by the elevation, seasonal variation, and slope aspect. The abundance of nitrogen-fixing bacteria was highest in winter and increased gradually in October but decreased in April and July as the elevation gradient increased. In different seasons, the abundance of nitrogen-fixing bacteria on the eastern and western slopes was similar in the order of winter > autumn > spring = summer. Moreover, there was no difference in the total number of soil nitrogen-fixing bacteria on the eastern and western slopes. The results also demonstrated that the diversity of soil nitrogen-fixing bacteria was higher at low (H1, H2) and medium altitudes (H3, H4) but lower at high medium altitudes (H5, H6). Furthermore, the diversity indices of nifH bacteria on the eastern and western slopes with elevation variation were similar. The abundance of nitrogen-fixing bacteria was significantly affected by herb coverage and NH 4 + -N and the most influential factors in the soil nifH community were NH 4 + -N, NO 3 − -N, and K. In addition, permutational multivariate analysis demonstrated that the abundance of nitrogen-fixing was most affected by season, whereas the diversity of nitrogen-fixing bacteria was most affected by altitude.