Effects of Different Land Use Types on Active Autotrophic Ammonia and Nitrite Oxidizers in Cinnamon Soils

ABSTRACT

Land use types with different disturbance gradients show many variations in soil properties, but the effects of different land use types on soil nitrifying communities and their ecological implications remain poorly understood. Using ¹³CO₂-DNA-based stable isotope probing (DNA-SIP), we examined the relative importance and active community composition of ammonia-oxidizing archaea (AOA) and bacteria (AOB) and nitrite-oxidizing bacteria (NOB) in soils under three land use types, forest, cropland, and greenhouse vegetable soil, representing three interference gradients. Soil net nitrification rate was in the order forest soil > cropland soil > greenhouse vegetable soil. DNA-SIP showed that active AOA outcompeted AOB in the forest soil, whereas AOB outperformed AOA in the cropland and greenhouse vegetable soils. Cropland soil was richer in NOB than in AOA and AOB. Phylogenetic analysis revealed that ammonia oxidation in the forest soil was predominantly catalyzed by the AOA Nitrosocosmicus franklandus cluster within the group 1.1b lineage. The ¹³C-labeled AOB were overwhelmingly dominated by Nitrosospira cluster 3 in the cropland soil. The active AOB Nitrosococcus watsonii lineage was observed in the greenhouse vegetable soil, and it played an important role in nitrification. Active NOB communities were closely affiliated with Nitrospira in the forest and cropland soils, and with Nitrolancea and Nitrococcus in the greenhouse vegetable soil. Canonical correlation analysis showed that soil pH and organic matter content significantly affected the active nitrifier community composition. These results suggest that land use types with different disturbance gradients alter the distribution of active nitrifier communities by affecting soil physicochemical properties.

IMPORTANCE Nitrification plays an important role in the soil N cycle, and land use management has a profound effect on soil nitrifiers. It is unclear how different gradients of land use affect active ammonia-oxidizing archaea and bacteria and nitrite-oxidizing bacteria. Our research is significant because we determined the response of nitrifiers to human disturbance, which will greatly improve our understanding of the niche of nitrifiers and the differences in their physiology.

INTRODUCTION

Nitrification is an important step in the soil N cycle. It is related not only to the transformation of NH₄⁺-N released by mineralization or exogenous NH₄⁺-N in the soil N cycle but also to the loss of soil N. The processes involved in nitrification are related to environmental issues such as soil acidification, aquatic eutrophication, and N₂O emission (1). Therefore, nitrification has always been a research focus in relation to N transformation. Ammonia oxidation is the first rate-limiting step of nitrification. Ammonia-oxidizing bacteria (AOB) and archaea (AOA) are considered the primary contributors to ammonia oxidation (2). Nitrite-oxidizing bacteria (NOB) catalyze the second oxidation step of the nitrification process, oxidizing nitrite to nitrate. Since functional groups such as AOA, AOB, and NOB have different response characteristics to environmental factors (such as oxygen and ammonia), their activities in different soil types and land use types are expected to vary.

Comparative genomic analysis showed that AOA and AOB may have significant differences in physiological characteristics and metabolic pathways (3). These differences imply that soil pH, soil N conditions, organic matter, and other environmental factors may determine the presence of different active ammonia oxidizing populations (AOA versus AOB) in different soil environments. Some studies have shown that acidic soils are suitable for AOA growth (4–8) and that neutral and alkaline soils are suitable for AOB growth (6, 9, 10). However, the relative contribution of AOA and AOB to soil ammonia oxidation remains controversial. Some studies have shown that AOB dominate ammonia-oxidizing communities in acidic soils (11–13), whereas AOA are dominant in neutral and alkaline soils (14, 15). These results suggested that, in addition to pH, other factors may have an important influence on the activity of different ammonia-oxidizing communities. For example, the availability of ammonia substrates and content of organic matter are also important factors affecting the relative competitive advantages of AOA and AOB (16–18). Wang et al. (19) found that soil ammonium concentration determines the communities of active AOA and AOB. In cultivated N-rich soils, AOA and AOB jointly drive ammonia oxidization, whereas in noncultivated N-limited soils, only AOA drive this process. Compared to AOA, the AOB communities and activity levels were more responsive to ammonium fertilization (20).

Soil organic matter (SOM) can stimulate AOA for heterotrophic or mixotrophic growth; however, the ammonia generated from the mineralization of organic matter can promote an increase in AOA abundance (17). NOB are also affected by environmental factors. R-strategists, represented by Nitrobacter, prefer high substrate concentrations (high N levels) and have a lower substrate affinity, whereas K-strategists, represented by Nitrospira, generally have a higher substrate affinity (21, 22). Therefore, fertilization can stimulate the growth of Nitrobacter-like NOB but not the growth of Nitrospira-like NOB (21, 23). Meng (24) found that Nitrospira abundance increases with an increase in soil pH and that the community structure of Nitrospira and Nitrobacter also changes with the changes in pH. Therefore, soil nitrifiers are complexly affected by a variety of environmental factors. Since the functions of ammonia and nitrite oxidation are spatially dependent (25, 26) and nitrite oxidation has been shown to be more sensitive than ammonia oxidation in disturbed soils (27, 28), it is important to concurrently assess the responses of ammonia and nitrite oxidizer communities to environmental disturbances. Clarifying the relative contributions to the nitrification of AOA, AOB, and NOB in different soil environments and active groups will help us understand the dynamic mechanisms of the global soil N cycle under climate change and human disturbance.

Different land use and management practices can have significantly different effects on the soil environment, leading to changes in soil physicochemical properties and microbial activities, which can further cause a decline in land ecological functions, including soil acidification (29, 30). In China, large areas of natural land have been converted into agricultural land in the past 60 years, and the intensity of agricultural land use has increased. This is especially noticeable in greenhouse vegetable land, which is characterized by intensive fertilization and high planting density, among other characteristics (31). Changes in land use have significantly affected soil physicochemical properties. For example, the conversion of forests to agricultural land has significantly reduced soil organic C and total N content (32). The application of a large amount of N fertilizers leads to the acidification of cropland and greenhouse vegetable soils (33). Land use conversion may change the community structure of ammonia oxidizers and the relative importance of AOA and AOB functional groups (11, 34). However, most previous studies on related topics have focused on acidic soils, and little attention has been paid to changes in NOB. Few studies have investigated how active nitrifiers respond to different disturbance degrees in originally alkaline soils.

The soils under different land use may have differences in main soil properties, such as pH and N status. These differences in soil conditions may affect the abundance and growth of AOB, AOA, and NOB. We hypothesized that (i) different land use patterns would change the active nitrifier community and that (ii) AOA would have a greater contribution to ammonia oxidation in forest soil with less disturbance and rich organic matter content, and the influence of active AOB would surpass that of AOA in agricultural land owing to the application of N fertilizer. To investigate these differences, soils from three common land use types, namely, forest, cropland, and greenhouse vegetable soil, were selected to represent three different interference levels (forest < cropland < greenhouse vegetable soil). Combined with high-throughput sequencing, DNA-based stable isotope probing (DNA-SIP) was used to investigate the active nitrifiers under these three land use types.

RESULTS

Soil physicochemical properties under different land use types.

The soil physicochemical properties are shown in Table 1. Under different land use types, the soil pH and the SOM, total nitrogen (TN), NH₄⁺-N, and NO₃⁻-N contents varied significantly (Table 1, P < 0.05). The forest soil was alkaline (pH 7.83), and its pH was in the range of values common for cinnamon soils. The agricultural soil was very acidic, and the pH value of the greenhouse vegetable soil was the lowest (5.32). The forest soil had the highest SOM and TN content, whereas these values were the lowest in cropland soil. The ammonium and nitrate concentrations in greenhouse vegetable soil were significantly higher than those of the forest and cropland soils, and the nitrate concentration in cropland soil was significantly higher than that of the forest soil.

TABLE 1.

Physicochemical properties of cinnamon soil under different land use types^a

Land use type	Mean pH ± the SD	Mean g·kg−1 ± the SD		Mean mg·kg−1 ± the SD
		Soil organic matter	Total N	NH4+-N	NO3−-N
Forest	7.83 ± 0.11C	39.8 ± 1.58C	2.27 ± 0.07B	0.31 ± 0.08A	46.3 ± 11.2A
Cropland	6.54 ± 0.01B	16.6 ± 2.40A	1.18 ± 0.20A	0.98 ± 0.40A	125 ± 5.99B
Greenhouse	5.32 ± 0.07A	29.1 ± 2.59B	2.10 ± 0.18B	6.35 ± 2.55B	265 ± 40.3C

^aDifferent letters in the same column indicate significant differences at P < 0.05 according to Tukey’s HSD tests.

Effect of land use types on soil nitrification activity.

The nitrite concentration was below the detection limit, and soil nitrification activity was estimated by measuring the changes in nitrate concentration in ¹³CO₂ plus [¹³C]urea and ¹²CO₂ plus [¹²C]urea microcosms before and after the 56-day incubation. Compared with that on day 0, the nitrate concentrations of the soil under ¹³CO₂-labeled and ¹²CO₂ control microcosms increased significantly on day 56 of incubation, but there was no significant difference in soil nitrate concentrations between these two treatments (Fig. 1a). The nitrate concentration of the soil treated with C₂H₂ (¹³CO₂+C₂H₂) was significantly lower than that of the soil in the ¹³CO₂-labeled and ¹²CO₂ control treatments, indicating that acetylene application completely inhibited autotrophic nitrification. During the 56-day incubation, the net nitrification rates of forest, cropland, and greenhouse vegetable soils were 14.5, 13.2, and 11.0 μg NO₃^–-N g⁻¹ soil day⁻¹, respectively (see Table S1 in the supplemental material). This was consistent with the accumulation trend of ammonium after cultivation (Fig. 1b). The ammonium concentrations of forest, cropland, and greenhouse vegetable soils were 0.15, 38.3, and 88.3 μg NH₄⁺-N g⁻¹ soil, respectively.

FIG 1.

Changes in concentrations of NO₃⁻-N (a), NH₄⁺-N (b), and copy numbers of ammonia-oxidizing archaea (AOA) (c) and ammonia-oxidizing bacteria (AOB) (d) amoA genes among the forest soil, cropland soil, and greenhouse vegetable soil in soil microcosms incubated with ¹²CO₂,¹³CO₂, or ¹³CO₂+C₂H₂ for 56 days. Error bars represent the standard errors of the mean of triplicate microcosms. Different letters above the columns indicate significant differences (P < 0.05) based on ANOVA.

Abundance and composition of nitrifying populations.

The population sizes of AOA and AOB were assessed by quantitative PCR before and after the 56-day incubation (Fig. 1c and d). Without the addition of C₂H₂, the abundance of the archaeal amoA gene in the forest soil increased significantly (P < 0.05) from 4.77 × 10⁸ on day 0 of incubation to 9.38 × 10⁸ on day 56 representing a 1.97-fold increase. However, from days 0 to 56 of incubation, the abundance of the archaeal amoA gene declined from 1.68 × 10⁷ to 4.87 × 10⁶ and from 2.02 × 10⁷ to 2.26 × 10⁶ in the cropland and greenhouse vegetable soils, respectively. The abundance of the bacterial amoA gene in the forest soil increased significantly (P < 0.05) from 4.99 × 10⁶ on day 0 of incubation to 4.43 × 10⁷ after 56 days, representing an 8.88-fold increase. In the same period, its abundance increased slightly (P > 0.05) in the cropland and greenhouse vegetable soils (Fig. 1d). The addition of C₂H₂ completely inhibited the growth of AOA and AOB.

MiSeq sequencing of the 16S rRNA gene at the community level was performed on days 0 and 56 in the labeled microcosms. Approximately 726,564 high-quality sequence reads were obtained (see Table S2). AOA, AOB, and NOB were selected from the reads for further analysis. The nitrifying community occupies a small part of the soil microbial population. Compared to that on day 0, the relative abundance of AOA in the forest soil on day 56 increased from 2.10% to 2.77%, but its relative abundance decreased in the cropland and greenhouse vegetable soils (Fig. 2a). The relative abundance of AOB increased from 0.02, 0.07, and 0.15% on day 0 to 0.26, 0.13, and 0.22% on day 56 in the forest, cropland, and greenhouse vegetable soils, respectively, representing 13.0-, 1.86-, and 1.47-fold increases, respectively (Fig. 2b). The NOB population was also strongly stimulated by 2.53-, 2.48-, and 1.81-fold in the forest, cropland, and greenhouse vegetable soils, respectively (Fig. 2c).

FIG 2.

Changes in relative abundances of AOA (a), AOB (b), and NOB (c) among different land use types after incubation for 56 days. The ratios were calculated based on the calculations of targeted 16S rRNA gene reads to total microbial 16S rRNA gene reads in each microcosm. The error bars represent the standard errors of the mean values of triplicate microcosms. Different letters above the columns indicate significant differences (P < 0.05) based on ANOVA.

SIP analysis of active ammonia- and nitrite-oxidizing populations.

Isopycnic gradient centrifugation of total DNA extracts from DNA-SIP microcosms was performed to identify which nitrifiers were labeled with ¹³CO₂ after 56 days of incubation. The ultracentrifugation yielded 15 fractions, and the buoyancy density ranged from 1.68 to 1.78 g ml⁻¹ from the top to the bottom of the centrifuge tube. qPCR was performed to determine the archaeal and bacterial abundances of the amoA gene in the 15 fractions. The highest copy numbers of AOA and AOB amoA genes treated with ¹²CO₂ and ¹³CO₂+C₂H₂ were observed in the lighter fractions with buoyant densities of <1.72 g ml⁻¹ (Fig. 3). However, in the ¹³CO₂-labeled microcosms of the forest soil, although the archaeal amoA genes peaked in the lighter fractions, a small number of archaeal amoA gene copies was detected in the heavy fraction (>1.72 g ml⁻¹) (Fig. 3a), suggesting partial labeling of some AOA. In cropland and greenhouse vegetable soils, the archaeal amoA genes peaked in the heavy fraction with buoyant densities of ∼1.73 g ml⁻¹ (Fig. 3c and e).

FIG 3.

Distribution of relative abundances of archaeal (a, c, and e) and bacterial (b, d, and f) amoA genes across the entire buoyant density gradient of the DNA fractions from soil microcosms incubated with ¹²CO₂,¹³CO₂, or ¹³CO₂+C₂H₂ for 56 days. AOA and AOB, ammonia-oxidizing archaea and ammonia-oxidizing bacteria, respectively. Error bars represent the standard errors of the mean of three repetitions.

In contrast, the highest copy number of AOB amoA genes was observed in the heavy fractions with buoyant densities of >1.72 g ml⁻¹ in all three soils of the ³CO₂-labeled microcosms (Fig. 3b, d, and f). Labeled amoA genes did not occur in the microcosms amended with ¹³CO₂+C₂H₂, implying that ¹³CO₂ assimilation by AOA and AOB was dependent on ammonia oxidation. The percentages of archaeal amoA gene copy numbers in heavy DNA in forest, cropland, and greenhouse vegetable soils (>1.72 g ml⁻¹) were 38.5, 67.9, and 74.6%, respectively, and the percentages of AOB were 91.2, 76.2, and 93.1%, respectively (see Table S1). Although the percentage of labeled AOA cells was low, it still exceeded that of AOB owing to the higher abundance of AOA in the forest soil (9.38 × 10⁸). However, the number of labeled AOB cells in the cropland and greenhouse vegetable soils was higher than that of labeled AOA cells during the nitrification process (Fig. 3; see also Table S1).

High-throughput sequencing of the total 16S rRNA genes further revealed the proportions of AOA, AOB, and NOB nitrifiers in the heavy fractions. The percentage of AOA 16S rRNA genes in all autotrophic nitrifiers in the forest soil (49.4%) was much higher than that of AOB16S rRNA genes (9.40%), indicating that AOA dominated ammonia oxidation in this soil. In vegetable soils, the percentage of AOB 16S rRNA gene reads was 41.6%, which was much higher than that of AOA 16S rRNA genes (1.84%), implying that AOB dominated ammonia oxidation in this soil. In cropland soil, the percentage of AOB 16S rRNA genes was 1.32 times that of AOA16S rRNA genes, but it only accounted for 2.81% of the nitrifiers. In this soil, the percentage of NOB was much higher than that of AOA and AOB, reaching 97.2% (see Fig. S1).

Phylogenetic analysis of AOA (see Fig. S2), AOB (see Fig. S3), and NOB (see Fig. S4) in day 0 soil and of the heavy fractions from the ¹³CO₂ treatments was carried out. Of the archaeal amoA genes in the forest soil, 57.8% was phylogenetically affiliated with the 54d9 cluster. The Nitrosocosmicus franklandus cluster comprised 69.3% of the archaeal amoA genes in the cropland soil, whereas 79.7% of the archaeal amoA genes were affiliated with Nitrososphaera viennensis in the vegetable soil on day 0. The AOA community on day 0 was significantly different among the three soils (F = 9.580, R² = 0.726, P = 0.003). Approximately 63.0 and 22.9% of the labeled AOA genes in the forest soil belonged to the Nitrosocosmicus franklandus and 54d9 clusters, respectively. Approximately 71.9 and 28.2% of the labeled AOA genes in the cropland soil belonged to the Nitrosocosmicus franklandus and Nitrososphaera viennensis clusters, respectively. Members belonging to the Nitrososphaera viennensis cluster dominated the AOA communities, accounting for 83.3% of the ¹³C-labeled archaeal 16S rRNA genes in the vegetable soil (Fig. 4a).

FIG 4.

Percentage of nitrifying phylotypes of AOA (a), AOB (b), and NOB (c) in forest, cropland, greenhouse vegetable soils. Day 0 refers to soils prior to incubation, and HF indicates the heavy DNA fraction from the ¹³CO₂ treatments. OTU, operational taxonomic unit. The error bars in the columns represent the standard errors of the mean of triplicate microcosms.

Approximately 53.3 and 46.7% of the AOB genes in the forest soil on day 0 belonged to the Nitrosomonas oligotropha lineage and Nitrosospira cluster 3, respectively. The AOB in the cropland soil were solely affiliated with Nitrosospira cluster 3, whereas the AOB in the vegetable soil on day 0 were phylogenetically most closely affiliated with Nitrosospira cluster 3 (82.9%) and Nitrosococcus watsonii lineage (17.1%) (Fig. 4b). The AOB community on day 0 was significantly different among the three soils (F = 7.665, R² = 0.719, P = 0.003). SIP incubation indicated that the Nitrosomonas oligotropha lineage and Nitrosospira cluster 3 accounted for 46.0 and 23.2%, respectively, of the ¹³C-labeled AOB in the forest soil. Intriguingly, Nitrosomonas communis, which was low in abundance on day 0, accounted for 30.8% of the ¹³C-labeled AOB in the forest soil. Further, 94.8% of the active AOB belonged to the Nitrosospira cluster 3 in the cropland soil. Approximately 44.6 and 54.6% of the active AOB genes in the vegetable soil belonged to the Nitrosospira cluster 3 and Nitrosococcus watsonii lineage, respectively.

With respect to NOB, all 16S rRNA gene sequences in the forest and cropland soils exclusively belonged to the genus Nitrospira. Nitrospira japonica dominated the NOB in the forest soil on day 0 (65.88%). Nitrospira moscoviensis accounted for 58.5% of the NOB in the cropland soil, and Nitrolancea hollandica dominated the NOB in the vegetable soil on day 0, accounting for 97.1% (Fig. 4c). The NOB communities on day 0 were significantly different among the three soils (F = 69.424, R² = 0.959, P = 0.006). The NOB labeled with ¹³C in the forest soil were Nitrospira marina, Nitrospira defluvii, Nitrospira moscoviensis, and Nitrospira japonica, which accounted for 34.6, 26.1, 10.6, and 28.7%, respectively. Nitrospira moscoviensis accounted for 82.1% of the ¹³C-labeled NOB in the cropland soil. In the vegetable soil, Nitrolancea hollandica and Nitrococcus mobilis accounted for 70.7 and 28.7% of the active NOB communities, respectively. Adonis analysis of variance of AOA, AOB, and NOB in the three soils showed no significant difference among communities in heavy fractions from microcosms and in day 0 soil (see Table S3).

Correlating soil properties to nitrifiers.

The potential relationship between nitrifiers and soil properties on day 0 can be inferred through canonical correspondence analysis (CCA). The results of this analysis showed that the active nitrifying community structure was significantly affected by SOM and pH, explaining 86.1% of the variance in active nitrifying communities (Fig. 5, F = 18.626, P = 0.001). The first CCA axis explained 65.1% of the variance in the active nitrifying phylotypes under in situ conditions (F = 11.215, P = 0.001), and the second CCA axis explained 21.0% of the variance (F = 3.989, P = 0.065). Pearson’s correlation analyses showed that pH was significantly positively correlated with the Shannon-Wiener diversity index (R = 0.843, P = 0.004) and operational taxonomic unit (OTU) richness index (R = 0.982, P = 2.67 × 10⁻⁶).

FIG 5.

Canonical correspondence analysis (CCA) between nitrifying populations and physicochemical properties of cinnamon soils under different land use types. The phylotypes of the nitrifiers were grouped based on the OTU taxa of 16S rRNA sequencing. A1 to A9 indicate the AOA phylotypes, B1 to B3 indicate the AOB phylotypes, and N1 to N14 indicate the NOB phylotypes, as shown in Fig. S2, S3, and S4 in the supplemental material. SOM, soil organic matter.

DISCUSSION

Relative importance of AOA and AOB in ammonia oxidation under different land use types.

DNA-SIP can directly connect active lineage with activities defined in complex soil environment, and it has been applied to identify active nitrifiers (4–15, 35). In the DNA-SIP-based experiments, the nitrification activity was stimulated after the addition of urea, but previous studies have demonstrated that there was no significant phylogenetic divergence in microcosms treated with water and urea (35). The ratio of labeled AOA/AOB cells largely reflected the relative importance of AOA and AOB in active soil nitrification (15). Quantitative analysis of the ¹³C-labeled amoA gene showed that, under three different levels of interference gradients, the AOA/AOB ratios in the forest, cropland, and greenhouse vegetable soils were 22.3, 0.34, and 0.19, respectively (see Table S1). Furthermore, the sequencing results of the total 16S rRNA genes of the ¹³C-DNA in the heavy fraction showed that the AOA/AOB ratios in the three soils were 5.27, 0.75, and 0.04, respectively (see Fig. S1). These results consistently showed the predominant role of AOA in the nitrification in forest soil and the predominant role of AOB in cropland and greenhouse vegetable soils. We expected that different long-term land use practices would affect the relative contribution of AOA and AOB to soil ammonia oxidation in the present study. However, because the impact of land use on the soil nitrification process is very complex, the results of this study contradicted those of many previous reports. For example, previous studies have shown that AOB in forest soils dominate ammonia oxidation (12), whereas AOA are sometimes the main contributors to nitrification in farmland and greenhouse vegetable soils (7, 8). Thus, the question arises as to which environmental factors affected the relative importance of AOA and AOB in the present study.

Our study and many previous reports have shown that soil pH is one of the key factors affecting the nitrifier community (Fig. 5) (36, 37). Previous studies have shown that AOB mostly play a predominant role in alkaline or neutral soils (6, 9, 10) and that AOA play a predominant role in acidic soils (4, 5, 8). We found that the forest soil dominated by AOA was alkaline (pH 7.83), whereas the greenhouse vegetable soil dominated by AOB was acidic (pH 5.32); in addition, AOB were slightly dominant in neutral farmland soil (pH 6.54). In contrast, some previous studies have shown that AOA are dominant in alkaline (pH 7.5 and 8.23) agricultural soil (14, 15) and that AOB are dominant in acidic forest soil and paddy soil (12, 13). Therefore, pH may not always affect the determinants of AOA and AOB differentiation in a complex soil environment (13–15). Studies on acidic red soil have shown that AOA are more important than AOB in terms of nitrification (38). Dai et al. (11) believed that AOA dominated nitrification in acidic forest soil and paddy soil and that AOB dominated nitrification in acidic upland soil.

Other factors, such as SOM content and ammonium concentration, also affect the relative importance of AOA and AOB (17, 19). In accordance with the findings of Dai et al. (11), we found that AOA were more prominent in forest soils rich in SOM and devoid of any N fertilization, whereas AOB were more prominent in upland agricultural soils/vegetable soils with low SOM content but with N fertilization (Table 1). This may be because abundant SOM promotes the growth of AOA, whereas the availability of ammonia produced at low but steady rates through SOM mineralization under natural conditions is conducive to the growth of AOA, which have a high affinity for ammonia (17, 39). In contrast, the findings of Zhong et al. (40), corresponding to those in present study, also showed that AOB were dominant in intensively fertilized greenhouse vegetable soil. This may be related to the high N content because of high level of fertilization in greenhouse vegetable fields (18, 19). In addition, some studies have shown that AOA were dominant in both acidic and alkaline paddy soils (11, 15), suggesting that other factors such as soil moisture, oxygen, and temperature may also affect the community structure of nitrifiers.

Active ammonia oxidizers under different land use types.

The results of our study showed that the active AOA in the three soils were affiliated with group 1.1b. In a previous study, surveys at global, regional, and local scales strongly suggested that the proportion of group 1.1b in AOA was high in nonacidic soil, and it was considered that, compared to other AOA groups, this AOA group had better adaption to relatively higher pH values (37, 41). In the present study, the active AOA in acidic greenhouse vegetable soil (pH 5.32) was also affiliated with group 1.1b. The results further showed that 63.0% of active AOA in forest soil belonged to the Nitrosocosmicus franklandus cluster in group 1.1b. Nitrosocosmicus franklandus was first reported in 2016. The pH range of its habitat is 6.0 to 8.5, and it belongs to the group of autotrophic microorganisms and cannot directly utilize organic matter (42), but organic matter can improve the activity of some members of this genus (43, 44). Forest soils have a high level of organic matter (Table 1), which may be the reason for the large growth of the Nitrosocosmicus franklandus cluster. Although AOA showed low relative abundance in cropland soils, the Nitrosocosmicus franklandus cluster was still the main participant in ammonia oxidation, possibly because Nitrosocosmicus is able to adapt to environments with high ammonium concentrations (42, 43) and grow in fertilized cropland soil. Compared to that in the forest soil, the Nitrososphaera viennensis cluster in the cropland and greenhouse vegetable soils accounted for a higher percentage in the active AOA group (cropland and vegetable soils versus forest soil), and it was dominant in the active AOA group of the vegetable soil. This result was inconsistent with the conclusions of previous studies, in which the active AOA group in acidic soil was affiliated with the marine group 1.1a-associated lineage (4, 35). However, other recent studies have found that archaeal nitrification in acidic soil was phylogenetically closely associated with Nitrososphaera viennensis EN76, and its activity may be related to the tolerance to acidic environment rather than to the high affinity for ammonia (e.g., in alkaline forest soil) (5, 6). This conclusion was supported by the present study.

Only the Nitrosomonas oligotropha lineage of AOB was detected in the forest soil before incubation, but after 56 days of incubation with urea (without C₂H₂), the Nitrosomonas communis lineage was detected as well. The reason for this was that Nitrosomonas oligotropha has a high affinity for ammonia, which is conducive to its growth in forest soils with very low ammonia concentrations, whereas Nitrosomonas communis has a low affinity for ammonia (45) and can grow rapidly after the addition of urea to the soil. The same situation can be seen in aquatic environments. The AOB members within the Nitrosomonas communis lineage are only present in ammonium-rich stream biofilms, whereas Nitrosomonas oligotropha lineage members were found in both ammonium-poor and ammonium-rich stream biofilms (46). The ammonia oxidation process in cropland soil is dominated by AOB, and the active group is mainly Nitrosospira cluster 3, which may be stimulated by fertilization. Bruns et al. (47) indicated that Nitrosospira cluster 3 became the dominant group in fertilized versus nonfertilized soils. Chu et al. (48) also observed that the application of N fertilizers gave a competitive advantage to Nitrosospira cluster 3. Nitrosospira cluster 3 was also observed in the forest soil in the present study, indicating that this group exists in both high-N and low-N environments (48).

Interestingly, we observed that 55.45% of active AOB in acidic greenhouse vegetable soil belonged to the Nitrosococcus watsonii lineage. It was previously thought that this lineage was mainly distributed in marine environments and salt lakes with high salt concentrations (500 to 700 mM) and an optimal pH of 7.6 to 8 (49). Recently, the numerically dominant Nitrosococcus watsonii lineage was observed in neutral and alkaline grassland soils (50). The findings of the present study further expanded the understanding of the active habitat of Nitrosococcus watsonii lineage, ranging from a neutral and alkaline high-salt environment to the acidic low-salt environment of cinnamon soil. Further research is needed to verify this hypothesis.

Active nitrite oxidizers under different land use conditions.

The sequencing of the 16S rRNA genes of total DNA indicated that the relative abundance of NOB in ¹³C-labeled microcosms after incubation for 56 days increased significantly compared to that at the beginning of incubation, implying that NOB plays a very important role in nitrification (Fig. 2). The percentage of NOB in cropland soil was the highest, accounting for 97.2% of nitrifiers and heavily outnumbering AOA and AOB (see Fig. S1). It should be noted that most NOB belong to the genus Nitrospira, which may comprise many members of the comammox group (51). Clearly, we could not confirm that OTU7 affiliates to comammox Nitrospira based on the sequencing of total 16S rRNA genes (52, 53). Thus, metagenomic evidence should be obtained to accurately identify whether there active comammox Nitrospira is abundant in cropland soil.

Furthermore, our results showed that the NOB community composition was significantly different among the three types of soils (P < 0.05). Nitrospira members, which have strong affinity for substrates and are K-strategists, were dominant in the forest and farmland soil. However, the composition of Nitrospira differed between the two soils (Fig. 4c). Previous studies have found that the genus Nitrospira has a high diversity and that there are great differences in cell morphology, cell size, optimum growth temperature, utilization of organic matter, and responses to nitrite concentration among different Nitrospira sublineages (54–56). For example, pure cultured Nitrospira marina is an obligate chemolithotroph, but organic matter can promote its growth (54), whereas Nitrospira defluvii (57) and Nitrospira japonica (56) belong to the mixotrophic type, which can utilize organic compounds. The three above-mentioned species can grow in forest soil with high SOM content. Nitrospira moscoviensis shows relatively better growth in cropland soil than in other soils, possibly because the excessive concentration of nitrate in vegetable soils and the abundant SOM in forest soil are not conducive to its growth and competitive advantage (58). In the present study, in the greenhouse vegetable soil, 28.7% of NOB belonged to Nitrococcus mobilis, and 70.7% belonged to Nitrolancea hollandica. They tended to be R-strategists which prefer high N environments and have low substrate affinity (59). Therefore, these two NOB were dominant in the high N environment of the greenhouse vegetable soil (Table 1), whereas Nitrospira was dominant in the forest soil with no N amendment and cropland soil with low N application.

Correlating soil properties to nitrifiers.

The above discussion showed that the community composition of active nitrifiers may be affected by many factors, including pH value, SOM content, and ammonia N concentration. The CCA showed that pH and SOM were the key factors shaping the phylogenetic distribution of active nitrifying communities (Fig. 5). Moreover, pH had a significant positive correlation with the Shannon-Wiener diversity index and OTU richness index. Previous studies have suggested that pH may affect the chemical form, concentration, and availability of nitrification substrates and may directly stimulate or inhibit the activity of specific nitrifiers (36, 60). They have also proposed that pH is an important factor affecting the niche differentiation of AOA and AOB (37). Moreover, considering that the functions of ammonia and nitrite oxidation are spatially dependent (25, 26), it is likely that NOB community composition is also affected by soil pH.

The reason AOA dominated nitrification in the forest soil may be the high SOM content and low ammonia concentration in the soil. In fact, in previous studies, it was also observed that root exudates (61) and litter (62) promoted significant growth of AOA. This is because some AOA are mixotrophic and can utilize organic matter (17), and the ammonia released by the mineralization of SOM also provides a substrate for the nitrification of AOA (63). In contrast, AOB have a low affinity for ammonia, and the slow release of ammonia during SOM mineralization is not sufficient to support its large-scale growth (16); thus, in the present study, AOB were dominant in N-fertilized cropland and vegetable soils. In addition to being affected by ammonia oxidizers, NOB are also affected by SOM. For example, some members of Nitrospira and Nitrobacter have been found to be stimulated by simple organic compounds (54, 64), whereas some Nitrobacter strains can grow completely heterotrophically (65).

MATERIALS AND METHODS

Site description and soil sampling.

Soil samples were collected in Yinan County, Shandong Province, northern China (35°36′28″N, 118°22′4″E). This region is characterized by a warm temperate continental monsoon climate, with a mean annual temperature of 13.7°C and mean annual precipitation of 856 mm. The soil type is cinnamon soil, which originates from limestone. Three adjacent but different land use types, 1 km away from each other, were selected for this study: forest, cropland, and greenhouse vegetable field soils. The forest site was planted with plantation forests (∼20 years old) dominated by Populus tomentosa Carr. The cropland site was managed with long-term traditional wheat (Triticum aestivum L.)–maize (Zea mays L.) rotation cropping; at the time of soil sampling, wheat was in the maturing stage. This site was fertilized with 350 to 450 kg N ha⁻¹. In the greenhouse vegetable field, cucumber (Cucumis sativus L.) has been the main cultivated crop for more than 10 years. This field was annually fertilized with 800 to 900 kg N ha⁻¹.

Soil samples were collected in May 2017. At each sampling location, three replicates of soil samples were collected. Each replicate was arranged into five randomized blocks (5-cm diameter × 20-cm length) which were then mixed. The samples were stored on ice in an ice box and transported to the laboratory immediately after sampling. After removing plant residues and stones, the sampled soils were passed through a 2-mm mesh. The sieved soil samples from each replicate plot were subdivided into two subsamples; one part was stored at 4°C for DNA-SIP microcosm experiments, and the other was air-dried for analyses of physicochemical properties.

Soil pH was measured using a mixture of water and soil at a ratio of 2.5. SOM content was measured using the dichromate oxidation method (66), and TN content was determined using the Kjeldahl method (67). Ammonium and nitrate were extracted with 2.0 M KCl (1:5), and their contents were determined using a segmented flow analyzer (SAN Plus; Skalar, Inc., Breda, The Netherlands).

DNA-SIP microcosms.

Microcosms for DNA-SIP were constructed as described in previous studies (7, 10). For the soil of each land use type, three SIP treatments were prepared with six replicates each: ¹³CO₂ microcosms, ¹²CO₂ microcosms, and ¹³CO₂+C₂H₂ (100 Pa) microcosms. For each treatment, fresh soil, equivalent to 6.0 g dws (dry weight soil), at 60% maximum water-holding capacity, was incubated in a 120-ml serum bottle capped with black butyl stoppers for 56 days at 28°C in the dark. Six-milliliter volumes of ¹²CO₂ or ¹³CO₂ (99 atom% ¹³C; Shanghai Engineering Research Center of Stable Isotope, Shanghai, China) were injected into the bottles with or without 100 Pa C₂H₂ on a weekly basis. The ¹³CO₂ and ¹³CO₂+C₂H₂ treatments were amended with 100 mg [¹³C]urea-N g⁻¹ dws (atom% ¹³C; Shanghai Engineering Research Center of Stable Isotope), whereas the ¹²CO₂ treatment was amended with 100 mg [¹²C]urea-N g⁻¹ dws on a weekly basis. A 14-day preincubation period was adopted at 40% of the maximum water-holding capacity before the incubation of the SIP microcosms. Destructive sampling was performed for each treatment on days 0 and 56. Three of the six replicates were used for determining the pH value and soil moisture content, and ∼2.0 g of fresh soil was collected from the remaining three replicates and stored immediately at −80°C for molecular analysis. The remaining soil was used for measuring soil ammonium, nitrite, and nitrate contents.

DNA extraction and SIP gradient fractionation.

The extraction of soil DNA was carried out using a FastDNA Spin kit for Soil (MP Biomedicals, Cleveland, OH). The quality and purity of the DNA were detected by 1% agarose gel electrophoresis and a NanoDrop ND-1000 UV-Vis spectrophotometer (NanoDrop Technologies, Wilmington, DE). The obtained DNA was stored at −20°C until further use.

SIP fractionation was carried out as described in Wang et al. (7). Approximately 3.0 μg of the DNA extract was mixed with CsCl solution, and the initial buoyancy density of CsCl solution was adjusted to 1.725 g ml⁻¹. The mixture was placed in a 5.1-ml polyallomer ultracentrifuge tube and centrifuged in an NTV-100 vertical rotor (Beckman Coulter, Brea, CA) at 55,000 × g for 44 h at 20°C.

Fifteen DNA fractions (333.3 ml) were collected starting from the bottom to the top of the tube by displacing them with sterile water using an NE-1000 single syringe pump (New Era Pump Systems Inc., Farmingdale, NY). The refractive index of each layer was measured using an AR200 digital handheld refractometer (Reichert, Inc., Buffalo, NY). The fractionated DNA was purified with polyethylene glycol-6000 and dissolved in 30 ml of Tris-EDTA buffer (pH 8.0).

Quantitative PCR.

The abundances of the archaeal and bacterial amoA genes before and after the 56-day incubation period were quantified by real-time quantitative PCR (qPCR) on a CFX96 optical real-time detection system (Bio-Rad Laboratories, Inc., Hercules, CA). The amoA gene abundance in the DNA-SIP fractions obtained by ultracentrifugation was used to estimate the ¹³CO₂-labeled ammonia oxidizers. The primer pairs Arch-amoAF/Arch-amoAR (68) and amoA-1F/amoA-2R (69) were used to amplify archaeal and bacterial amoA genes, respectively. The reaction was carried out in a 20-μl qPCR system, which included 10 μl of 2× SYBR Premix Ex PCR buffer (TaKaRa, Dalian, China), 0.5 μM concentrations of upstream and downstream primers, and 1 μl of DNA template. The real-time PCR conditions were as follows: 95°C for 3 min; 40 cycles of 95°C for 10 s, 55°C for 30 s, and 72°C for 30 s; followed by plate readings at 83°C. Melting curve analysis was performed to confirm the specificity of amplification products by measuring fluorescence continuously as the temperature increased from 65 to 95°C. qPCR amplification was carried out in three biological replicates with three technical replicates. Plasmid DNA from representative clones containing archaeal or bacterial amoA genes was used as the template to generate the standard curve. A dilution series of standard template with 3.24 × 10² to 3.24 × 10⁸ and 1.77 × 10² to 1.77 × 10⁸ per assay was used for archaeal amoA gene and bacterial amoA gene, respectively. The amplification efficiency was 92.9 to 96.5%, with R² values of 0.996 to 0.999. In addition, a serial dilution of the DNA templates was used to assess whether the PCR was inhibited during amplification.

Sequencing and phylogenetic analysis.

MiSeq sequencing of 16S rRNA genes in the V4-V5 regions was performed using the universal primers 515F-907R. The total DNA extracts from microcosms with ¹³CO₂ and ¹³CO₂+C₂H₂ on days 0 and 56 and the DNA recovered from the heavy buoyant density (>1.72 g ml⁻¹) were measured using the Illumina MiSeq platform (Majorbio Bio-pharm Technology Co., Ltd., Shanghai, China).

The sequence was merged with FLASH (70), and the obtained sequence was quantitatively analyzed using the QIIME software for microbiome analysis (71). Sequences with quality scores below 20 or length below 200 bp were eliminated, and only the sequences that overlapped by more than 10 bp were used for further analysis. OTUs were clustered at 97% sequence identity. The RDP classifier was used to classify and analyze the 16S rRNA gene sequences in the Silva 16S rRNA database (release 123), with a confidence threshold of 70%. AOA (Thaumarchaeota at the phylum level), AOB (Nitrosomonas, Nitrosococcus, and Nitrosospira), and NOB (Nitrobacter, Nitrospira, Nitrotoga, Nitrolancea, Nitrococcus, Nitrospina, and “Candidatus Nitromaritima”) were screened from the reads. A phylogenetic tree was constructed with representative sequences of 16S rRNA with MEGA version 4.0 using the neighbor-joining method and maximum composite likelihood model with 1,000 replicates to obtain bootstrap values (72).

Statistical analysis.

Variance among different treatments was evaluated using one-way analysis of variance (ANOVA), and Tukey’s post hoc tests were performed for multiple comparisons. All analyses were conducted using SPSS 17.0 (IBM, Armonk, NY). P values of <0.05 were considered statistically significant. CCA was used to identify the relationship between the nitrifier population and physicochemical factors. The nitrifier population was characterized by OTUs, and the physicochemical factors included pH, SOM, TN, ammonium concentrations, and nitrate concentrations. CCA was carried out in CANOCO (version 4.5 for Windows; PRI, Wageningen, The Netherlands). Correlations between the physicochemical factors that significantly affected the nitrifier population and Shannon’s index and richness were analyzed using Pearson’s correlation analyses, which were carried out using SPSS 17.0. The nitrifying community differences based on Bray-Curtis dissimilarity matrices were analyzed using the Adonis command in R language (73).

Data availability.

Nucleotide sequences have been uploaded to NCBI Sequence Read Archive (SRA) database under accession number PRJNA659889 for the 16S rRNA genes derived from the DNA-SIP experiment.