Changes,in,the,activities,of,key,enzymes,and,the,abundance,of,functional,genes,involved,in,nitrogen,transformation,in,rice,rhizosphere,soil,under,different,aerated,conditions

XU Chun-mei ,XlAO De-shun ,CHEN Song,CHU Guang,LlU Yuan-hui,ZHANG Xiu-fu,WANG Dan-ying

The State Key Laboratoty of Rice Biology,China National Rice Research Institute,Hangzhou 310006,P.R.China

Abstract Soil microorganisms play important roles in nitrogen transformation. The aim of this study was to characterize changes in the activity of nitrogen transformation enzymes and the abundance of nitrogen function genes in rhizosphere soil aerated using three different methods (continuous flooding (CF),continuous flooding and aeration(CFA),and alternate wetting and drying (AWD)). The abundances of amoA ammonia-oxidizing archaea (AOA) and ammonia-oxidizing bacteria (AOB),nirS,nirK,and nifH genes,and the activities of urease,protease,ammonia oxidase,nitrate reductase,and nitrite reductase were measured at the tillering (S1),heading (S2),and ripening(S3) stages. We analyzed the relationships of the aforementioned microbial activity indices,in addition to soil microbial biomass carbon (MBC) and soil microbial biomass nitrogen (MBN),with the concentration of soil nitrate and ammonium nitrogen. The abundance of nitrogen function genes and the activities of nitrogen invertase in rice rhizosphere soil were higher at S2 compared with S1 and S3 in all treatments. AWD and CFA increased the abundance of amoA and nifH genes,and the activities of urease,protease,and ammonia oxidase,and decreased the abundance of nirS and nirK genes and the activities of nitrate reductase and nitrite reductase,with the effect of AWD being particularly strong. During the entire growth period,the mean abundances of the AOA amoA,AOB amoA,and nifH genes were 2.9,5.8,and 3.0 higher in the AWD treatment than in the CF treatment,respectively,and the activities of urease,protease,and ammonia oxidase were 1.1,0.5,and 0.7 higher in the AWD treatment than in the CF treatment,respectively. The abundances of the nirS and nirK genes,and the activities of nitrate reductase and nitrite reductase were 73.6,84.8,10.3 and 36.5% lower in the AWD treatment than in the CF treatment,respectively.The abundances of the AOA amoA,AOB amoA,and nifH genes were significantly and positively correlated with the activities of urease,protease,and ammonia oxidase,and the abundances of the nirS and nirK genes were significantly positively correlated with the activities of nitrate reductase. All the above indicators were positively correlated with soil MBC and MBN. In sum,microbial activity related to nitrogen transformation in rice rhizosphere soil was highest at S2. Aeration can effectively increase the activity of most nitrogen-converting microorganisms and MBN,and thus promote soil nitrogen transformation.

Keywords: rhizosphere aeration,gene abundance,enzyme activities,soil microbial biomass carbon,soil microbial nitrogen

The nitrogen supply capacity of paddy soil is the most important factor affecting soil fertility and limiting rice production (Liuet al.2016). Regardless of whether fertilizer is applied,more than 50% of the nitrogen used by rice is derived from the soil (Sharmaet al.2021). Soil nitrogen is thus the main source of nitrogen for plants.Most of the nitrogen in the soil is organic nitrogen (＞95%),and organic nitrogen can only be absorbed and utilized by plants when it is converted into inorganic nitrogen through the activities of microorganisms and soil enzymes (Akteret al.2018). Therefore,the supply of inorganic nitrogen in the soil is one of the main factors limiting crop yields. In recent years,the application of large amounts of fertilizer to increase rice yields has resulted in a decline in rice quality and nitrogen use efficiency (Yuet al.2020),as well as the loss of nitrogen and deterioration of the ecological environment (Erismanet al.2015). New measures are needed to improve the ability of rice to assimilate soil nitrogen and mitigate the use of chemical nitrogen fertilizers.

Soil microorganisms play a key role in the biochemical process underlying nitrogen transformation in rice rhizosphere soil (Cheet al.2018). They participate in the soil nitrogen cycle through their own metabolism;organic nitrogen is also mineralized and decomposed into inorganic nitrogen through their own life activities (Turneret al.2017;Liet al.2019). The species,structure,and function of rhizosphere soil microorganisms are closely related to the decomposition and transformation of nitrogen (Kuyperset al.2018;Zilioet al.2020). Soil microorganisms are living components in the soil and are highly sensitive to various changes in soil properties. Thus,changes in soil microorganisms are indicators of changes in the ecological function of soil;soil microorganisms also play important roles in maintaining the nitrogen cycle in ecosystems (Gallowayet al.2008). Flooded (hypoxic) conditions during rice growth lead to a series of changes in the physical and chemical properties of soil,which affect the structure of the microbial community,the soil nitrogen transformation process,and thus the amount of soil available nitrogen. Aerobic cultivation can improve the nitrification potential,promote the nitrification process,accelerate the turnover of different forms of nitrogen,and facilitate the release of NH4+from fertilizer and soil organic nitrogen,so that it can be oxidized to NO3-,all of which enhance the use of soil organic nitrogen by rice (Xuet al.2020). However,the effects of aerobic cultivation on soil nitrogen transformation in rice rhizosphere soil and the mechanism of soil nitrogen absorption and utilization remain unclear. Other questions requiring clarification include how aerobic cultivation affects the distribution of key microbes involved in nitrogen transformation in rice rhizosphere soil,and how these functional microorganisms regulate the nitrogen conversion rate of rice rhizosphere soil and the supply of available nitrogen. Few studies have examined the relationships between the key processes of nitrogen transformation in rice rhizosphere soil and nitrogen uptake and utilization in rice (Chenet al.2019;Xuet al.2020),or the mechanisms underlying these processes under aerobic cultivation. In addition,the role of the key microorganisms in the transformation of nitrogen in rhizosphere soil has not been studied.

Here,we investigated the effects of rhizosphere aeration conditions on nitrogen function gene (NFG)abundances and on key nitrogen transformation processes,including nitrification,denitrification,and fixation. We hypothesized that changes in NFG abundances and soil enzyme activities would be affected by different rhizosphere aeration conditions. We also hypothesized that NFG abundances,soil enzyme activities,and nitrogen transformation processes would be affected by edaphic conditions. To test these hypotheses,we conducted an experiment in 2021 with three rhizosphere aeration treatments,continuous flooding(CF),alternate wetting and drying (AWD),and continuous flooding and aeration (CFA). We determined NFG abundances and soil enzyme activities of key microbes as well as the relationships of NFGs with different forms of nitrogen in rhizosphere soil,to determine the relationships between nitrogen transformation and NFG abundances in the rhizosphere soil under aerobic cultivation.

2.1.Site description and experimental design

Pot experiments (plastic buckets,80 cm×60 cm×80 cm)were carried out at the China National Rice Research Institute (CNRRI),Hangzhou,Zhejiang Province,China(30.3oN,120.2oE,11 m a.s.l.) from June to November in a network room. Natural illumination was provided throughout the pot experiment. Each pot was filled with 20 kg of paddy soil (0-15 cm in depth),fertilized with 1.25 g N kg-1soil,0.6 g P kg-1soil,and 1.0 g K kg-1soil. The timing of fertilization was as follows: 40% of N,100% of P,and 50% of K were applied at the start of the experiment in the form of compound fertilizer (N:P:K as 15:15:15);30% of N (in the form urea) was top-dressed at both the mid-tillering and panicle initiation stages;and 50% of K (in the form of KCl) was top-dressed at panicle initiation.

The pots were arranged in a completely randomized design consisting of three treatments: (CF,AWD,and CFA) with six replicates. In the CF treatment (i.e.,anaerobic irrigation),the rice plants were flooded with 3-5 cm of water during the entire growing period. In the AWD treatment (i.e.,aerobic irrigation,which had periodic drying and reflooding during rice growth),the rice plants were allowed to dry naturally after an initial flooding period and then were irrigated again before harvest. The rice plants in the CFA treatment were grown in a similar manner to those in the CF. Before rice seeding,a gas vent with holes was embedded in the CFA pots,and a ventilation pipe was then connected to a ventilation pump. Ventilation (controlled by a timer) was initially performed for 2 h. Thereafter,ventilation was performed for 10 min every 2 h throughout the day. The variety Xiushui 09 (conventionalJaponicarice) was used in the experiment,and it was planted as a single crop. The seeds were directly sown into the plastic buckets. After germination,the seedlings were thinned to 36 seedlings per bucket. All treatments began one week after the emergence of rice.

2.2.Sampling and analysis

SamplingBefore rice seeding,soil samples (of the 0-15 cm layer) were taken for physicochemical analyses.Rhizosphere soil samples (approximately 2 mm of soil scraped from the root surface with a thin blade) were taken from plants harvested during the main growth stages (tillering (S1),heading (S2),and ripening (S3)stages).

Each sample was divided into two parts: one half was immediately stored at -20°C for subsequent DNA extraction and determination of enzyme activities,and the other half was stored at 4°C for nitrogen analyses.

Determination of the soil O2 diffusion rateThe diffusion rate of O2dissolved in the soil was measured using an O2diffusion meter equipped with an additional guard cathode (EK070 14.36,Holland). A cylindrical platinum electrode was inserted into the soil to measure the amount of electrical current required for the reduction of all the O2at the surface. Upon reaching a steady-state,the flow of O2in the pores filled with air,and the water film that passed through the electrode surface was measured.

Soil analysisPrimary soil properties were measured before fertilization using methods as described by Lu(2000). The experimental soil contained total N,2.34 g kg-1;organic matter,30.2 g kg-1;alkalihydrolysable N,312.3 mg kg-1;available P,10.3 mg kg-1;and available K,78.6 mg kg-1and had a pH of 6.83.

Soil microbial biomass carbon (MBC) and soil microbial biomass nitrogen (MBN) were determined using a fumigation-extraction method based on the difference between carbon and nitrogen extracted with 0.5 mol L-1K2SO4(soil:extraction ratio=1:4) from chloroformfumigated and unfumigated soil samples using theKEC(0.38) andKEN(0.54) factors,respectively (Lu 2000).The dissolved organic carbon (DOC) and total nitrogen content was determined using catalytic oxidation and nondispersive infrared spectroscopy at 680°C (TOC-V CPH,Shimadzu,Japan).

NO3--N and NH4+-N were extracted with 2 mol L-1KCl by shaking at 250 r min-1for 60 min at 25°C. The NO3--N and NH4+-N concentration was measured using the indophenol blue and phenol disulphonic acid methodsviaspectrophotometry,respectively.

The activities of soil urease,protease,nitrite reductase,and ammonia oxidase were measured using appropriate appropriate extraction and analysis kits (Suzhou Keming Co.,China).

Soil DNA extraction and target gene quantification by qPCRSoil DNA was extracted from 0.25 g of soil using a MoBio PowerSoil DNA Isolation Kit (MoBio Laboratories Inc.,Carlsbad,USA) following the manufacturer’s instructions. We measured the abundance of functional genes involved in the first steps in nitrification (amoA;encoding ammonia monooxygenase in ammoniaoxidizing archaea (AOA) and ammonia-oxidizing bacteria(AOB)),in nitrogen fixation (nifH),and in denitrification(nirKandnirS). Primers and conditions for qPCR are shown in Table 1. All qPCR reactions were carried out on an MG96+real-time PCR machine (Hangzhou Langji Scientific Instrument Co.,Ltd.,China) in wells containing 10 μL of SYBR@PremixExTaq(TaKaRa,Japan),including 0.8 μL forward primer (5 μmol L-1),0.8 μLreverse primer (5 μmol L-1),7.4 μL double deionizedwater (ddH2O),and 1 μL template DNA adjusted to a final volume of 20 μL with ddH2O. Plasmids for qPCR were extracted using a TaKaRa MiniBEST Plasmid Purification Kit ver.4.0 (TaKaRa,Japan),and then the purified PCR products were cloned into the Pgem-T Easy Vector (Promega,Madison,WI,USA) and transformed intoEscherichiacoliJM109 competent cells (Promega,Madison,WI,USA). The plasmid DNA was then diluted in 10-fold increments (108-101copies) after estimating the concentrations of plasmids using a UV-Vis spectrophotometer (NanoDrop®ND-2000) to generate the standard curves. High efficiencies from 90.10 to 97.92% were obtained for all amplifications,withR2values ranging between 0.99 and 1.00. The number of gene copies per gram of dry soil in each sample was then calculated by comparing the outputs of the qPCR reactions with the calibration curves.

Table 1 Primers and thermal cycling conditions for PCR reactions

2.3.Statistical analyses

For qPCR data,the number of gene copies were performed to test significant differences between treatments using the SPSS/STAT statistical analysis package (ver.11.0,Chicago,IL,USA). Additionally,one-way analysis of variance (ANOVA) was conducted to evaluate the significance of differences between treatments. Pearson correlation coefficient (r) was used to determine the correlations between the abundance ofamoA-AOA,amoA-AOB,nirS,nirK,andnifHgenes and the activities of urease,nitrite reductase,nitrate reductase,protease,and ammonia oxidase enzymes,and MBN,MBC,and soil nitrate and ammonium nitrogen concentrations. The statistical and correlation analyses were conducted using SPSS 19.0. The figures were created with Sigma Plot 14.0.

3.1.Soil oxygen diffusion rate

The oxygen diffusion rate of rhizosphere soil differed among the treatments (CF,CFA,and AWD) (Table 2). It was significantly higher in the CFA and AWD treatments than in the CF treatment. No significant difference in the oxygen diffusion rate was observed between the CFA and AWD treatments at growth stages S1 and S2;however,at stage S3,the oxygen diffusion rate of soil was significantly higher in the AWD treatment than in the CFA treatment.The soil oxygen diffusion rate continued to decrease during the entire growth period in the CF treatment.

Table 2 Oxygen diffusion rate (μA) in the rhizosphere soil under different soil oxygen conditions at three rice growth stages1)

3.2.Soil NH4+ and NO3– concentration

The concentration of soil NO3-did not significantly differ among treatments at growth stage S1,however,significant differences among treatments were observed at stages S2 and S3 (Fig.1-A). The concentration of NO3-was significantly higher in the CFA and AWD treatments than in the CF treatment,but no significant difference was observed between the CFA and AWD treatments. The concentration of NO3-decreased in the CF treatmentthroughout the rice growth period,but increased in the AWD and CFA treatments at stage S2. The average NO3-concentrations in the CF,AWD,and CFA treatments were 1.7,2.5,and 2.6 mg kg-1,respectively. The concentration of NH4+increased in the CF treatment as growth progressed and fluctuated in the CFA and AWD treatments (Fig.1-B). The average concentration of NH4+was significantly lower in the CFA and AWD treatments (by 55.3 and 49.6%,respectively) than in the CF treatment.

Fig.1 Soil nitrate-nitrogen (NO3--N) concentration (A),and ammonium-nitrogen (NH4+-N) concentration (B) at three rice growth stages under different oxygen conditions. S1,tillering;S2,heading;S3,ripening. CF,continuous flooding;CFA,continuous flooding and aeration;AWD,alternate wetting and drying. Different letters above bars indicate statistically significant differences(P＜0.05) among the different treatments (n=6;bars are SD).

3.3.Soil MBC and MBN

Soil MBC increased with rice growth and peaked at stage S3 (Table 3). The magnitude of the increase in soil MBC varied under different soil oxygen conditions. The magnitude of increase in soil MBC was the greatest in the CFA treatment,followed by the AWD and CF treatments.Soil MBC was slightly higher in the CF treatment than in the AWD and CFA treatments at stage S1,and it was slightly higher in the AWD treatment than in the CF and CFA treatment at stage S2. No significant difference in soil MBC was observed between growth stages S1 and S2. At stage S3,soil MBC was 28.3 and 13.6% higher(P＜0.05) in the CFA treatment than in the CF and AWD treatments,respectively. Throughout the rice growth period,soil MBN was the highest in the CFA treatment,and was 2.00,1.18,and 1.89 times higher at growth stages S1,S2,and S3,respectively,in the CFA treatment than in the CF treatment. There was no significant difference in soil MBN at stage S1 between the CF and AWD treatments. MBN was the highest in all treatments at stage S2,and decreased at stage S3.

Table 3 Soil microbial biomass carbon (MBC) and soil microbial biomass nitrogen (MBN) under different soil oxygen conditions at different rice growth stages (mg kg-1)1)

3.4.Abundances of amoA genes in ammonia-oxidizing archaea (AOA) and ammonia-oxidizing bacteria (AOB)

The copy numbers ofamoAgenes in AOA and AOB in the rhizosphere under different treatments varied significantly with rice growth (Fig.2). The abundances ofamoAgenes in AOA and AOB increased in the rhizosphere soon after rice seeding,and their abundances were highest at stage S2. The abundances ofamoAgenes in AOA and AOB varied among treatments. Increases in the oxygen content in the rhizosphere (AWD and CFA treatment)decreased the AOAamoAgene abundances but increased the AOBamoAgene abundances. AOAamoAgene abundances were 45.0,1.2,and 6.5% lower in the CFA treatment and 53.4,13.1 and 13.9% lower in the AWD treatment at stages S1,S2,and S3,respectively,than in the CF treatment. The copy numbers of theamoA gene in AOB gene in the rhizosphere in the AWD treatment,which ranged from 4.1×107to 13.3×107g-1dry soil,were greater than those in the CF treatment,which ranged from 0.9×107to 9.2×107g-1dry soil. The abundance of the AOBamoAgene increased as the rhizosphere oxygen content increased,and there was a significant difference in the abundance of the AOBamoAgene between the AWD and CF treatments and between the CFA and CF treatments. There was no significantdifference in the abundance of the AOBamoAgene between the AWD and CFA treatments. The changes in the abundances of theamoAgene in AOA and AOB were similar. The abundance of AOBamoA gene peaked at stage S2,and they were both 1.4 times higher in the CFA and AWD treatments than in the CF treatment.

Fig.2 Abundance of the amoA (ammonia-oxidizing archaea(AOA) and ammonia-oxidizing bacteria (AOB)) genes in rhizosphere soil at three rice growth stages under different oxygen conditions. S1,tillering;S2,heading;S3,ripening. CF,continuous flooding;CFA,continuous flooding and aeration;AWD,alternate wetting and drying. Different letters above bars indicate statistically significant differences (P＜0.05) among the different treatments (n=6;bars are SD).

3.5.Denitrification gene abundance

The abundances of the denitrification genesnirS,andnirKwere significantly (P＜0.05) greater in the CF treatment than in the CFA and AWD treatments at most growth stages (Fig.3-A and B). The copy numbers of thenirSgenes peaked at the S2 growth stage under different rhizosphere oxygen treatments,and decreased as the rhizosphere oxygen content increased (Fig.3-A).Rhizosphere soil samples had a significantly (P＜0.05)greater number ofnirSgene copies in the CF treatment than in the CFA and AWD treatments. No significant difference in the number ofnirSgene copies was observed between the CFA and AWD treatments at the S2 and S3 stages. The number ofnirKgene copies was significantly higher in the AWD treatment than in the CFA treatment at the stages S1 and S2,but it was significantly lower in the AWD treatment than in the CFA treatment at the stage S3.

Fig.3 Abundance of nirS (A) and nirK (B) genes in rhizosphere soil at three rice growth stages under different oxygen conditions.S1,tillering;S2,heading;S3,ripening. CF,continuous flooding;CFA,continuous flooding and aeration;AWD,alternate wetting and drying. Different letters above bars indicate statistically significant differences among the different treatments for nirS and nirK (n=6;bars are SD).

3.6.Abundance of nitrogen-fixing genes

The copy number of thenifHgene,which indicates the abundance of nitrogen-fixing bacteria,was significantly greater in the rhizosphere in the CFA and AWD treatments than in the CF treatment at all rice growth stages (P＜0.05)(Fig.4). The maximum relative increase in the abundance of thenifHgene (5.19-fold) was from stage S1 to S2 in the CFA treatment;the lowest relative decrease was from stage S2 to S3 in the AWD treatment. At growth stage S3,the abundance of thenifHgene was significantly higher in the CFA treatment than in the CF and AWD treatments.

Fig.4 Abundance of the nifH gene in rhizosphere soil at three rice growth stages under different soil watering treatmentsoxygen conditions. S1,tillering;S2,heading;S3,ripening. CF,continuous flooding;CFA,continuous flooding and aeration;AWD,alternate wetting and drying. Different letters above bars indicate statistically significant differences among treatments (n=6;bars are SD).

3.7.Soil enzyme activities

The activities of rhizosphere soil enzyme activities increased with rice growth (Fig.5). At growth stages S2 and S3,urease activity was significantly higher in the AWD treatment than in the CF and CFA treatments(Fig.5-A). The only significant difference in urease activity between the CF and CFA treatments was at growth stage S3. At stages S2 and S3,urease activity was 2.9-and 2.3-fold higher in the AWD treatment,and 1.3-and 1.1-fold higher in the CFA treatment,respectively,than in the CF treatment. There were no significant differences in urease activity between the CF and AWD treatment at S1. The soil nitrite reductase activity was highest in the CF treatment,followed by the CFA treatment and the AWD treatment,at all growth stages (Fig.5-B). There were significant differences in soil nitrite reductase activity among the three treatments at the S2 and S3 stages. At S2,soil nitrite reductase activity in the AWD and CFA treatments was 0.8 and 0.6 times,respectively,that in the CF treatment (Fig.5-C). At S1,soil nitrite reductase activity was higher in the CF treatment than in the AWD and CFA treatments,but there was no significant difference in soil nitrite reductase activity between the CFA and CF treatment. No significant difference was observed in the nitrate reductase activity between the CF and CFA treatments at any rice growth stage,but it was significantly lower in the AWD treatment than in the CF and CFA treatments. Protease activity was 1.3-and 1.7-fold higher,and 1.1-and 0.2-fold higher,respectively,at the S2 and S3 growth stages in the CFA and AWD treatments than in the CF treatment (Fig.5-D). At stage S1,protease activity was 41.0% higher and 20.1% lower in the AWD and CFA treatments,respectively,than in the CF treatment. Ammonia oxidase activity was 1.4-,1.8-and 1.7-fold higher in the AWD treatment than in the CF and CFA treatments at growth stages S1,S2,and S3,respectively,and all differences between treatments were significant (Fig.5-E). Ammonia oxidase activity was higher in the CFA treatment than in the CF treatment,however,no significant difference in ammonia oxidase activity was observed at stages S1 and S3 between the CF and CFA treatments.

Fig.5 Urease (A),nitrite reductase (B),nitrate reductase (C),protease (D),and ammonia oxidase (E) activities in rhizosphere soil at three rice growth stages under different oxygen conditions. S1,tillering;S2,heading;S3,ripening. CF,continuous flooding;CFA,continuous flooding and aeration;AWD,alternate wetting and drying. Different letters above bars indicate statistically significant differences among treatments (n=6;bars are SD).

3.8.Relationships of different forms of soil nitrogen with microbial properties and enzyme activities

The abundances of nitrogen transformation genes (amoAAOA,amoAAOB,nirS,nirK,andnifH) were significantly correlated with most soil nitrogen enzyme activities(Table 4). The abundances ofamoAAOA,amoAAOB,andnifHwere significantly positively correlated with urease,protease,ammonia oxidase activities,and the NO3--N concentration and MBN content. ThenirSandnirKabundances were significantly positively correlated with nitrate reductase activity and NH4+-N concentration,and negatively correlated with the NO3--N concentration.Although the correlations of nitrogen transformation genes with the activities of nitrogen transformation enzymes and the concentration of different forms of nitrogen varied,the abundance of all the nitrogen transformation genes was positively correlated with the MBC and MBN content.

4.1.Soil microbial communities involved in nitrogen transformation

Soil nitrogen transformation includes nitrogen fixation,nitrification,and denitrification,and it is a complex process that is mainly driven by soil microorganisms (Kuyperset al.2018;Liet al.2019). Soil microorganisms are sensitive indicators of environmental change,as they respond rapidly to variation in soil type,oxygen content,and pH(Jangidet al.2008;Streitet al.2014). The rhizosphere oxygen content affects the different forms of soil nitrogen by changing the soil microecological environment,which affects the absorption and utilization of nitrogen by rice (Xuet al.2020). Molecular markers are widely used to estimate the abundance of microorganisms;for example,the AOB and AOAamoAgenes are used to estimate the abundance of microorganisms involved in nitrification (Gruber and Galloway 2008;van der Heijdenet al.2008). Nitrification is the aerobic oxidation process of microorganisms,and the soil oxygen concentration is considered the most important factor affecting nitrification(Behrendtet al.2017),as it can regulate the population dynamics,activity,and abundance of AOB and AOA in rhizosphere soil,which affects soil nitrification anddenitrification. Soil aeration is an important factor affecting the structure of microbial communities in paddy soils;however,the sensitivity of microorganisms to the oxygen content in soil varies (Zhanget al.2016). Our findings indicated that aerobic conditions (AWD and CFA treatments) increase the abundance of AOB,whereas hypoxic conditions (CF treatment) reduce the abundance of AOB. TheamoAAOB abundance was significantly higher in the AWD treatment than in the CFA treatment,which indicates that the oxygen content was high in the AWD treatment (Table 2). Some studies have shown that O2has no significant effect on the diversity ofamoAAOA in the form of both DNA and mRNA in estuarine sediment(Yanget al.2016). We found that the composition of the AOA was not affected by the oxygen concentration in the rhizosphere. Indeed,the content ofamoAAOA increased even under hypoxic treatment (Fig.1). Thus,the rhizosphere oxygen concentration mainly affected nitrification through its effect on the abundance of AOB.

Table 4 Relationships between functional genes and soil properties1)

Denitrification is a process in which NO3-and NO2-are used as microbial electron receptors and are gradually reduced to gases (NO,N2O,and N2) by various enzymes encoded bynirKandnirSgenes under anoxic or low oxygen conditions (Schimel and Bennett 2004);thenirSgene is more sensitive to external environmental factors (Levy-Boothet al.2014). We found that the abundances of thenirKandnirSgenes were higher in the CF treatment than in the CFA and AWD treatments,andnirS-type denitrifiers were more abundant thannirKtype denitrifiers in all treatments,which may stem from the oxygen or inorganic nitrogen content. This is consistent with the finding that higher NH4+concentration increases the abundance of microorganisms containingnirKornirSgenes because high NH4+concentrations can increase the supply of NO3-,which provides a direct substrate for the growth of denitrifying microorganisms (Yiet al.2015).All other environmental factors being equal,an increase in the inorganic nitrogen concentration within a certain range has a direct effect on the denitrification process.Ammonium nitrogen also promotes the nitrification process and affects the abundance and community structure of AOB and AOA. Biological nitrogen fixation is driven by a highly diverse group of microorganisms. All nitrogenfixing microorganisms possess the samenifHgene,which is widely used as a marker for studying the abundance and composition of nitrogen-fixers in various environments(Chenet al.2018). Biological nitrogen fixation plays an important role in reducing the loss of nitrogen in the soil,and it significantly changes as the amount of dissolved oxygen in the rhizosphere changes (Wanget al.2012).This is consistent with the results of our experiments,the abundance of thenifHgene was lower in the CF treatment than in the AWD and CFA treatments (Fig.4),which indicates that improved rhizosphere oxygen conditions could increase the abundance of thenifHgene and thus enhance the nitrogen fixation capacity of paddy soil.Increases in MBN also reflect increases in the soil nitrogen storage capacity and improvements in the continuous supply capacity of soil nitrogen. MBN is involved in nitrogen conversion and promotes the interaction between plants and the soil nitrogen cycle (Caoet al.2018).Rhizosphere oxygenation significantly increased the content of soil MBN,and the soil MBN was significantly positively correlated with the abundance of thenifHgene.This might be explained to some extent by the fact that the oxygenation treatment enhanced soil microbial activity and promoted the reproduction of nitrogen-fixing bacteria in soil,which is consistent with the increase innifHgene abundance in this experiment.

We found that the abundances of all the NFGs (amoA,nirS,nirK,andnifH) were higher in the rhizosphere soil at growth stage S2 than at the S1 and S3 stages. This might be explained by the fact that the full heading stage is the vigorous period of rice growth. In this period,the plant tissue developed and the root system grew in the limited space in the pot,which would have improved the rhizosphere environment. Root exudates and abscission produced by root metabolism would have increased soil carbon levels,affecting soil microbial community structure and activity,and the nitrogen cycle (Coskunet al.2017).The abundance of NFGs was affected by the different rhizosphere oxygen treatments,as rhizosphere conditions regulate the microclimate and make the chemical environment favorable for the development of rice plants,thereby affecting rice growth and development (Xuet al.2020). Aeration affects the photosynthetic rate and root activity (Zhaoet al.2010),which can alter the ecological distribution of microorganisms in the rice rhizosphere.

4.2.Activities of soil enzymes involved in nitrogen transformation

Soil enzymes are mostly derived from microorganisms,and they play an important role in soil nutrient cycling and the provision of nutrients essential to plant growth(Siet al.2018). Thus,changes in the composition and functional genes of the microbial community lead to changes in soil enzyme activity and affect the turnover of nutrient elements. The abundance of functional genes involved in nitrogen transformation was closely related to soil invertase activity (Table 4). Soil enzymes mediate the conversion of soil nitrogen (Zumstein and Helbling 2019). Soil enzyme activity (including protease and urease activity) can indicate the strength of soil nutrient transformation and the ability of the soil to supply nutrients to plant roots (Ghiloufiet al.2019;Zhanget al.2019). Many factors affect the activity of soil nitrogen conversion-related enzymes;for example,there is a positive correlation between the oxygen concentration and invertase activity (Songet al.2018). Ammonia oxidation is the first and rate-limiting step of nitrification,and ammonia oxidase is an important catalyst in ammonia oxidation,with soil nitrate being the product of nitrification (Jinet al.2009). In this study,changes in AOB abundance and the soil nitrate nitrogen concentration were consistent.The activities of nitrate reductase and nitrite reductase are closely related to denitrification (Niet al.2016),and the rhizosphere oxygen content is an important factor affecting denitrification. Improvements in the rhizosphere oxygen environment (in the CFA and AWD treatments)reduced the activities of nitrate reductase and nitrite reductase,which alters the number of denitrifying bacteria in rhizosphere soil (Fig.2). These findings can be explained by the fact that aerobic cultivation increases the amount of organic acids and soluble sugars secreted by roots (Xuet al.2013),which affects the rhizosphere soil environment,the number of soil microorganisms (Table 3),and thus soil enzyme activities (Fig.4). Aerobic cultivation improved the rhizosphere micro-ecological environment of rice and affected the activities of enzymes involved in the soil nitrogen transformation process such as urease,nitrate reductase,and protease,which affected nitrogen mineralization,ammonification,nitrification,and urea hydrolysis in the soil. Thus,aerobic cultivation promoted nitrogen mineralization and urea hydrolysis in rhizosphere soil by enhancing protease,ammonia oxidase,and urease activities,which increased the mineralization and hydrolysis of soil organic nitrogen and artificially applied urea to NO3--N and NH4+-N.

In this study,we found that aeration of the soil significantly increased the NO3--N concentration,the MBC,and MBN content,and the nitrogen invertase activity and altered the abundance of nitrification,denitrification,and nitrogenfixation genes. The number of functional genes and activity of enzymes involved in nitrogen transformation in rhizosphere soil were higher at growth stage S2. The abundances of the AOAamoA,AOBamoA,andnifHgenes and the activities of ammonia oxidase,urease,and protease in the key growth stages of rice (S1,S2,and S3)were higher in the AWD and CFA treatments than in the CF treatment;in addition,the abundances ofnirS,andnirK,and the activities of nitrate reductase and nitrite reductase were lower in the AWD and CFA treatments than in the CF treatment. Increases in soil nitrogen invertase activity and the MBN content were the main factors affecting the abundance of soil nitrogen transformation genes in the rhizosphere soil of rice after aeration. The findings of our study provide new insights into the effects of rhizosphere oxygen conditions on the biology and chemistry of nitrogen transformation in paddy soil.

Acknowledgements

This research was supported by the Key Research and Development Program of Zhejiang Province,China(2022C02008),the National Natural Science Foundation of China (31401343),the earmarked fund for China Agriculture Research System (CARS-01),and the Agricultural Science and Technology Innovation Program of Chinese Academy of Agricultural Sciences (CAASZDRW202001).

Declaration of competing interest

The authors declare that they have no conflict of interest.