A total of eight paddy soil samples representing eight parent material types from different provinces were collected from northern China to southern China during September, 2016, to exclude seasonal variation (Figure 1 and Table S1). These paddy soil samples were obtained from Heilongjiang (HLJ), Beijing (BJ), Shanxi (SX), Henan (HeN), Anhui (AH), Jiangxi (JX), Chongqing (CQ) and Hunan (HuN), respectively. The types of soils were indicated in Figure 1. The paddy field all had a history of rice cultivation for more than 15 years. At each site, triplicate surface soil samples (depth, 0 to 15 cm) were randomly obtained by using a sterile hand trowel. Collected samples were placed in sterile plastic bags, sealed, and transported to the laboratory on ice. Plant residues and other materials, such as stones and obvious macrofauna, were removed before each replicate from each field was homogenized through a 2 mm mesh. Aliquots (5 g wet weight) of soil from each site were stored at −80˚C prior to subsequent DNA extraction and molecular analysis. The subsamples were stored at 4˚C for chemical determination and potential nitrification activity (PNA). PNA was processed immediately after arrival.

Chemical characteristics including pH, soil water content (SWC), NH 4 + -N, NO 3 − -N, soil organic matter (SOM) and total nitrogen (TN) were determined in this study. The soil pH was determined using a Mettler Toledo 320-S pH meter (Mettler-Toledo Instruments Co.Ltd., Shanghai, China) with a water-to-soil ratio of 2.5. Water content was determined gravimetrically by drying the soil at 105˚C for 6 h. NH 4 + -N and NO 3 − -N were extracted from soil samples with 2 M KCl, and the contents were determined using a Skalar San Plus segmented flow analyzer (Skalar Inc., Breda, The Netherlands). SOM was detected using a total carbon analyzer (TOC-V CPH, Shimadzu, Japan). TN was determined by the Kjeldahl digestion method. All analyses were performed in triplicate for each soil sample. These parameters were then calculated based on oven-dried soil weight.

The PNA is the maximum capacity of nitrifying microorganisms to transform NH 4 + -N to NO 3 − -N in a soil population. Changes in nitrification potential provide quantitative information on how nitrifying communities behave in various environmental conditions and reflect the in situ nitrification rates. Nitrification potential was determined according to the shaken-slurry method [26] . PNAs were measured with fresh soils from each site by mixing each of triplicate. 15 g (wet weight) of soil from each sample was mixed with 100 ml of 1.5 mM ammonium sulfate. After incubation for 0, 2, 6, 12 and 24 h at 25˚C, 10 ml slurry samples were centrifuged and the supernatant was filtered through a 0.45 μm pore size membrane and stored at −20˚C until analysis. NO x − -N content in the supernatant was immediately analyzed. NO x − -N concentration increased linearly, and nitrification potential (μg NO x − -N∙g⁻¹∙h⁻¹) was calculated from the rate of increase in NO x − -N concentration over time in the slurry using linear regression.

The total DNA in soil samples was extracted and purified from 0.5 g (dry weight) of soil using a FastDNA spin kit for soil (MP Biomedicals, Cleveland, OH, USA), according to the manufacturer’s instructions. The quantity and quality of DNA extracts were assayed using a NanoDrop ND-1000 UV-visible light spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA). The soil DNA was stored at −20˚C for further analysis.

Real-time quantitative PCR (qPCR) was performed in triplicate on a CFX96 Optical RealTime detection system (Bio-Rad Laboratories, Inc., Hercules, CA, USA) to determine the copy number of amoA and 16S rRNA genes in soil DNA extracts. The number of amoA gene copies per g of soil (dry weight) was measured using primers amoA-1F and amoA-2R [27] targeting amoA gene of AOB, and primers amoA-AF and amoA-AR [28] targeting amoA gene of AOA. A universal 16S rRNA gene qPCR was conducted to assess the total copy number of bacterial communities by using the primer 515/907. The primers and PCR/qPCR thermal conditions were detailed in Table S2 in the supplemental material.

The real-time quantitative PCR standard was generated using plasmid DNA from one representative clone containing bacterial or archaeal amoA, and 16S rRNA gene. A dilution series of standard template over eight orders of magnitude per assay was used. In addition, a series dilution of soil DNA extract was also used to assess whether a PCR inhibition by humic substance co-extraction occurred in the paddy soils. The amoA gene copies decreased proportionally with the diluted soil DNA template concentrations. A 10-fold diluted soil DNA was used for subsequent analysis. qPCR was performed in a reaction mixture containing 1 × SYBR® Premix Ex Taq™ (TaKaRa Biotech, Dalian, China), 0.5 mM of each primer, and approximately 1.0 - 5.0 ng DNA template. Amplification efficiencies ranged from 91% to 103%, with R² values of 0.993 to 0.999. Melting curve analysis and standard agarose gel electrophoresis were always performed at the end of a PCR run to verify the amplification specificity.

AOB and AOA composition were analyzed by a DGGE of amoA genes using the D-Code system (Bio-Rad Laboratories, Hercules, CA), as described previously [6] . PCR reaction was carried out in a thermal cycler (Bio-Rad Laboratories, Hercules, CA). The thermal condition was listed in Table S2. The PCR products were checked using 1.2% agarose gel to ascertain the specificity of bacterial amoA and archaeal amoA gene amplification with a size of ~489 bp and ~629 bp, respectively. The 150 to 200 ng PCR amplicons of bacterial amoA were separated using 8% (wt/vol) polyacrylamide [acrylamide-bisacrylamide (37.5:1)] gels with a denaturing gradient of 30% - 60% (100% denaturant contains 7 M urea and 40% formamide), and the 150 to 200 ng PCR amplicons of archaeal amoA were separated using 6% gels with a denaturing gradient of 30% - 50%. A 1 mm thick gel was poured from bottom to top using a gradient former and peristaltic pump at a speed of 4.5 ml∙min⁻¹. A 5.0 ml stacking gel containing no denaturants was subsequently added on top before polymerization to insert a comb to make 24 wells. An electrophoresis separation of bacterial amoA and archaeal amoA gene amplicons was carried out using a 0.5 × Tris-acetate-EDTA buffer at 80 V for 17 h at 60˚C. The gels were stained with 1:10000 (v/v) SYBR Green I for 30 min, and then scanned with a Molecular Imager FX using Quantity One software package (Bio-Rad Laboratories, Hercules, CA, USA). Dominant bands in the DGGE fingerprints were excised and reamplified. The reamplified PCR products of the DGGE bands were cloned using pEASY-T3 vector (TransGenBiotech, Beijing, China) according to the instructions of the manufacturer. The clones containing the exact bacterial amoA and archaeal amoA gene insert from each DGGE band were sequenced (Invitrogen, Shanghai, China). DNAMAN software was used to check the clone sequences manually for subsequent analyses. Phylogenetic analysis was performed using the Molecular Evolutionary Genetics Analysis (MEGA 7.0) software package. The sequences of basic tree were from known AOB and AOA cultures, and fosmid clones of the amoA genes.

The concentration of NO 3 − -N, NH 4 + , SOM, TN, PNA and amoA gene copy numbers were compared through multiple sample comparisons using one-way ANOVA analysis followed by Student-Newman-Keulstest to check for quantitative variance between different samples. The two-way ANOVA analysis was performed to examine the effect of different soil characteristics and soil types on PNA. Pairwise comparisons between treatments were determined with two-tailed Student’s t-tests. All analyses were conducted using SPSS version 20.0 (IBM, Armonk, NY, USA). Pearson’s correlation analyses were performed to assess the relationships among PNA, soil properties and the abundances of AOA and AOB. Redundancy analysis (RDA) with the Monte Carlo permutation’s test (499 permutations) was carried out to determine if the AOA and AOB community structures were correlated with PNA and soil properties using Canoco (version 4.5).

The nucleotide sequences were deposited at the GenBank with Accession Numbers MG000601-MG000613 and MG000592-MG000600 for the bacterial amoA and archaeal amoA genes in this study, respectively.

Soil properties including SOM, TN, inorganic nitrogen concentration, soil water content, pH were listed in Table 1. The soil samples were collected from different latitude from 21 m to 931 m. Soil samples from BJ, SX, HeN, AH, and CQ had similar pH values from 7.72 to 7.99, which was classified as alkaline soil. Acidic soil samples from HLJ, JX, and HuN had relatively lower pH values of 5.37, 5.38 and 6.10, respectively. The water content of the soil samples ranged from 10.82% to 29.37%. There was significantly difference in SOM and TN content of these soil samples from different sampling sites (P < 0.05), ranging from 21.34 to 49.11 g∙kg⁻¹ and 0.68 to 1.23 g∙kg⁻¹, respectively, with the highest content in HLJ sample. SX sample had the highest NO 3 − -N and NH 4 + -N content, while CQ sample had the lowest NO 3 − -N content and JX had the lowest NH 4 + -N content. Soil NH₃ concentrations were estimated to be varied 0.41 to 553.18 μM. The NH₃ content was lowest in the acidic soil sample JX and was highest in the alkaline soil sample AH. This is consistent with the principle that NH₃ is ionized exponentially to NO 3 − with decreasing pH [29] . The PNA ranged from 0.43 to 3.57 μg NO x − -N∙g⁻¹∙h⁻¹ (Table 1) across all paddy soils. Sample from AH was found to have the high PNA. As revealed by Pearson correlation analysis, PNA had significant relationship with soil NH₃ concentration (r = 0.82, P < 0.01) in Table 2.

Bacterial 16S rRNA and amoA gene were quantified by qPCR to investigate the variation in population sizes of microbial and nitrifying communities (Figure 1S and Figure 2). As shown in Figure 2, the copy numbers of archaeal and bacterial amoA genes ranged from 3.58 × 10⁹ to 6.54 × 10¹⁰, and from 3.35 × 10⁶ to 5.09 × 10⁸ copies∙g⁻¹ d.w.s in these paddy soils, respectively. AOA abundance was

^aMeans of three replicates per treatment are presented with standard deviation; ^bSOM and TN denote soil organic matte and total nitrogen, respectively. ^cpH was measured with water-soil ratio of 2.5. ^dSoil NH₃ concentration was estimated on the basis of soil water content, ammonium concentration, pH and adjusted for temperature (pKa of 9.245 at 25˚C); ^ePNA, potential nitrification activity.

*Correlation is significant at the 0.05 level; **Correlation is significant at the 0.05 level.

generally higher that AOB abundance in paddy soils. The ratio of archaeal to bacterial amoA gene copy numbers was varied from 12.8 to 2400.0. The abundance of bacterial 16S rRNA genes ranged from 3.67 × 10¹² to 2.45 × 10¹³ copies∙g⁻¹ d.w.s in different paddy soil samples (Figure S1).

Correlation analysis was performed to reveal the influence of environmental parameters on the nitrifying communities and PNA (Table 2). There was a significant positive correlation between PNA and AOA abundance (r = 0.83, P < 0.01), whereas no relationship was observed between PNA and AOB abundance (r = −0.16, P > 0.05). The PNA was also significantly correlated with soil NH₃ content (r = 0.82, P < 0.01). A significant positive correlation was found between AOB population size and SOM (r = 0.65, P < 0.05) and TN (r = 0.64, P < 0.05). In addition, the abundance bacterial 16S rRNA gene was positively correlated with soil NO 3 − -N content.

The structures of AOA and AOB communities were revealed by fingerprinting analysis of bacterial and archaeal amoA genes in all paddy soils (Figure 3(a) and Figure 4(a)). Highly reproducible DGGE fingerprints were obtained in the triplicate of each soil sample. The diversity of AOA and AOB were indicated by Shannon’s diversity index (H). richness (S) and Evenness (Eh), which were calculated by the relative abundance of various DGGE fingerprints in different soil samples (Table 3). The H and S index of AOA communities in HLJ, BJ and SX samples were clearly higher, when compared with the corresponding index of AOB, whereas the opposite trend was observed for the remaining soil samples. The highest H, S and Eh index of AOA communities was found in BJ soil, while the lowest was found in AH soil.

Distinct different DGGE fingerprints of archaeal amoA genes were observed among the different paddy soils. AOA communities in paddy soils were dominated by the DGGE band-A3, which were affiliated with the 54d9-like cluster of the soil group 1.1b lineage (Figure 3 and Figure 5). The majority of the DGGE bands for AOA were affiliated with the soil group 1.1b lineage except for the DGGE band-A2, which was placed within the Nitrosotaleav devanaterra-like cluster of the marine group 1.1a-associated lineage. The DGGE band-A5 and A6 showed higher intensities in the SX, HuN and HLJ samples, although they were faint in other soils. The DGGE band-A8 and A9 were observed in six soil samples except for AH and JX. The result was confirmed by cluster analysis which revealed that AH and JX samples were clearly separate from other samples with an 89% similarity (Figure 3(b)). DGGE band-A7 was only detected in HLJ.

Note: Different lowercase letters in the same column mean significant difference (P < 0.05)

Distinct AOB communities were also observed in the paddy soils. Phylogenetic analysis showed that eight of the thirteen DGGE bands were related with the Nitrosospira cluster lineage, whereas the remaining 5 DGGE bands were placed within Nitrosomonas cluster 7 (Figure 4(a) and Figure 6). The dominant AOB DGGE bands B11, B12 and B13 with high-intensity were all clearly present in all soil samples, which were affiliated with the Nitrosospira cluster 3. Additionally, the DGGE band B3 and B5 belonging to Nitrosomonas europaea-like cluster seemed obviously higher in BJ, HeN and HLJ samples, which were also grouped into one cluster with a similarity of 55% (Figure 4 and Figure 6). The AOB communities in SX and AH samples reached 87% of the similarity. The high intensity of DGGE band-B2 affiliated with the Nitrosomonas cluster 7 was only detected in HuN sample.

Pearson correlation analyses were performed to reveal the impact of soil properties on the diversity index of nitrifying communities and PNA (Table 4). The PNA was significantly negatively correlated with the diversity of AOB (r = 0.81, P < 0.01). The Richness E of AOA and AOB communities was positively correlated with soil SOM (AOA, r = 0.66, P < 0.05 and AOB, r = 0.64, P < 0.05) and TN (AOA, r = 0.63, P < 0.05 and AOB, r = 0.66, P < 0.05). All other diversity indexes did not show a significant correlation with the soil properties in paddy soils.

RDA was conducted to determine the correlation of soil properties with community structures of AOA and AOB (Figure 7). Partial RDAs based on Monte Carlo permutation (n = 499) only retained the significant parameters in the models. The first and second axes accounted for 49.0% and 18.0% of the total variation in AOA community structure, respectively (Figure 7(a)). Regarding

*Indicate significant level at p < 0.05 level. **Indicate significant level at p < 0.01 level.

the AOA community, the most influential constraining variable 40% was TN (F = 4.04, P = 0.002) (Figure 7(a) and Table 5). The first and second axes accounted for 35.6% and 29.5% of the total variation in AOB community structure (Figure 7(b)). No significant correlation between AOB community structure and soil properties was found.