Integrated management of crop residue and nutrient enhances new carbon formation by regulating microbial taxa and enzymes

WU Hong-liang ,CAI An-dong ,XING Ting-ting ,HUAI Sheng-chang ,ZHU Ping ,HAN Xiao-zeng ,XU Ming-gang,LU Chang-ai

1 Institute of Agricultural Resources and Regional Planning,Chinese Academy of Agricultural Sciences/National Engineering Laboratory for Improving Quality of Arable Land,Beijing 100081,P.R.China

2 Institute of Environment and Sustainable Development in Agriculture,Chinese Academy of Agricultural Sciences,Beijing 100081,P.R.China

3 Institute of Agricultural Resources and Environment,Jilin Academy of Agricultural Sciences,Changchun 130033,P.R.China

4 Northeast Institute of Geography and Agroecology,Chinese Academy of Sciences,Changchun 130102,P.R.China

5 College of Resources and Environment,Shanxi Agricultural University,Taiyuan 030031,P.R.China

Abstract Although returning crop residue to fields is a recommended measure for improving soil carbon (C) stocks in agroecosystems,the response of newly formed soil C (NFC) to the integrated supply of residue and nutrients and the microbial mechanisms involved in NFC are not fully understood. Therefore,an 84-day incubation experiment was conducted to ascertain the microbial mechanisms that underpin the NFC response to inputs of residue and nitrogen (N),phosphorus (P),and sulfur (S) in two black (Phaeozem) soils from experimental plots at Gongzhuling,Jilin Province and Hailun,Heilongjiang Province,China. The results showed that adding residue alone accelerated microbial nutrient mining,which was supported by decreases of 8-16% in the ratios of C:N and C:P enzyme activities,relative to soils with nutrient inputs. The NFC amounts increased from 1 156 to 1 722 mg kg?1 in Gongzhuling soil and from 725 to 1 067 mg kg?1 in Hailun soil as the levels of nutrient supplementation increased. Boosted regression tree analysis suggested that β-glucosidase (BG),acid phosphatase (AP),microbial biomass C (MBC),and Acidobacteria accounted for 27.8,18.5,14.7,and 8.1%,respectively,of the NFC in Gongzhuling soil and for 25.9,29.5,10.1,and 13.9%,respectively,of the NFC in Hailun soil. Path analysis determined that Acidobacteria positively influenced NFC both directly and indirectly by regulating BG,AP,and MBC,in which MBC acquisition was regulated more by AP. The amount of NFC was lower in Hailun soil than in Gongzhuling soil and was directly affected by AP,indicating the importance of soil properties such as SOC and pH in determining NFC. Overall,our results reveal the response of NFC to supplementation by N,P,and S,which depends on Acidobacteria and Proteobacteria,and their investment in BG and AP in residue-amended soil.

Keywords:newly formed soil carbon,extracellular enzyme activities,gene abundance,nutrient supplementation,black soil

1.Introduction

Soil carbon (C) sequestration in agroecosystems is a primary measure for mitigating climate change,and small changes in soil C stocks can clearly affect the biogeochemical C cycle (Lal 2008;Stockmannet al.2013). Globally,crop residue incorporation in soils has been recommended in field practices with the aim of improving soil organic C (SOC) (Hanet al.2018).However,knowledge of the sustainable contribution of residue to NFC (derived from residue) is insufficient,and studies have even shown that increased residue inputs have resulted in little or no increase in SOC across soils with various fertility levels (Powlsonet al.2011;Hanet al.2018). Thus,feasible management practices are needed to cause a greater proportion of residue C to be transferred into NFC,and the underlying mechanism needs to be clarified.

Nutrient management is believed to be the key to regulating NFC with C-rich crop residue inputs (Zhaoet al.2018). Certain nutrients such as nitrogen (N),phosphorus(P),and/or sulfur (S) can enhance the humification of residue C and increase or decrease SOC (Hanet al.2016). Studies have proven that residue decomposition is related to the availability of nutrients (e.g.,N and P) (Kirkbyet al.2013;Caiet al.2018). Moreover,Kirkbyet al.(2014)provided novel evidence that integrated management of residue and nutrients would result in a greater proportion of residue C being transformed into soil C than for residue addition alone. Soil microbes control the conversion of residue C to SOC,and an understanding of the impact of added nutrients on microbial growth is key to increasing the conversion of residue C to SOC. However,the relationship between NFC and microbial activity and its abundance under integrated supplementation of nutrients is still unclear.

Inputs of nutrients and labile organic materials (LOM)to soils alter microbial investment strategies and cause higher enzyme production according to the theories of“microbial nutrient mining” and “microbial stoichiometry decomposition” (Blagodatskaya and Kuzyakov 2008;Chenet al.2014). Residue inputs with high C to nutrient ratios cause microbes to mine nutrients from native soil nutrient reserves,which may result in a decrease in C to nutrient acquisition extracellular enzyme activities (EEAs)ratios (Waringet al.2014). That is,when there is an increase in labile C input,but insufficient nutrient supply,microbes will invest in more nutrient acquisition enzymes,which target nutrient-rich SOM. Supplying sufficient nutrients while adding residue may stimulate the growth of the related microbial taxa and change the strategies of microbes to employ C,N,and P acquisition enzymes with a stoichiometric constant (Gulis and Suberkropp 2003;Heucket al.2015). Therefore,nutrient supplementation enhances the mineralization of native SOC and increases the amount of NFC that is derived from residue C (Kirkbyet al.2013).

Changes in the availability of environmental resources alter the growth rate of various microbes,mainly of copiotrophs and oligotrophs in soils (Huet al.1999).Generally,oligotrophs,such as Acidobacteria and fungi,prefer to use native SOM to meet their energy and nutrient requirements in nutrient-poor environments (Kaiseret al.2014). However,the growth rates of most microbial taxa,including Proteobacteria and Actinobacteria,increase rapidly in nutrient-rich environments (Ramirezet al.2012;Chenet al.2014). Shifts in microbial activities lead to accelerated mineralization of LOM and native SOM,which regulate the conversion of soil new C and old C(Kirkbyet al.2014). That is,the dominant microbial taxa(e.g.,copiotrophs or oligotrophs) change with nutrient availability,and their activities and abundances determine NFC accumulation and decomposition of native SOM(Chenet al.2014;Kirkbyet al.2014). However,the pathways of microbial activity strategies have not been clarified in terms of the link between NFC and microbial taxa or EEAs with the goal of identifying appropriate residue and nutrient management practices.

Black soils,classified as Phaeozems according to the FAO (2015) are widely distributed in Northeast China. In recent decades,the land in this region has been cultivated excessively and treated excessively with chemical fertilizers,which have caused serious soil degradation(Dinget al.2016). However,the large amounts of crop residue in this region have not been fully utilized. Longterm field experiments in this region have shown that the intensity of SOC sequestration under residue input is positively related to latitude and SOC density (Wang S Cet al.2018). Therefore,soils with different C levels were selected from two long-term trials to:1) identify the NFC response to the integrated inputs of residue and nutrients and 2) examine how shifts in microbial abundance and EEAs might affect NFC. Two hypotheses were proposed:1) the availability of N and P was the dominant factor influencing NFC,and the magnitude of NFC increased with increased levels of nutrient input;2) MBC was the direct source of NFC,which was regulated more by P acquisition EEAs.

2.Materials and methods

2.1.Soil collection and processing

Soils for the experiment were collected at 0-20 cm depth in October 2018 from two long-term trials with crop residue incorporation in China. The trial sites were located at (1)Gongzhuling,Jilin Province (43°30′N,124°48′E);and(2) Hailun,Heilongjiang Province (47°27′N,126°55′E)(Table 1). Residue return experiments with chemical N,P,and K fertilizer were initiated in 1990 at both sites(Appendix A). Prior to the experiments,the fields had been cultivated for at least 150 years at Gongzhuling and for 60 years at Hailun. Air-dried soils were gently crushed by hand and were then passed through a 2-mm sieve. Any recognizable gravel,debris,and plant-like materials (coarse fractions of SOM,CF-SOM,＞2 mm)were carefully removed. The samples that were sieved at 2 mm were re-sieved with a 0.4-mm sieve to separate the remaining CF-SOM from the heavy mineral fraction on top of the 0.4-mm sieve as described in Kirkbyet al.(2013). This study used the remaining soil (fine fractions of SOM) as the research object. The basic soil properties are shown in Table 1.

Soil subsamples were sieved through a 0.15-mm sieve and analyzed for C,N,and S concentrations using an elemental analyzer (Hanau,Germany). Organic P was determined by the ignition-extraction procedure(Olsen and Sommers 1982). The atom13C of the soil was analyzed with an elemental analyzer that was interfaced with a PDZ Europa 20-20 isotope ratio mass spectrometer(Sercon Ltd.,Cheshire,UK). Soil pH was determined at a 1:5 soil:water ratio using a pH electrode.

2.2.Crop residue and nutrient treatments

Five treatments were used for each soil type:control (no amendment),residue alone,residue+low nutrient addition,residue+medium nutrient addition,and residue+high nutrient addition (Table 2). To highlight the variations in NFC,maize residue was added to soil at 12.5 g kg?1soil (dry soil weight),which corresponded to a double field rate of 7.5 t ha?1at 20 cm soil depth (Gongzhuling soil had a bulk density of 1.24 t m?3). The13C-labeled maize residue was collected by harvesting mature maize residue labeled by13CO2pulse and was dried at 60°C(Anet al.2015). The maize residue was cut into pieces that were approximately 2-mm long prior to mixing with soil. Analysis showed the residue contained the following element concentration:total C,42.8%;total N,0.97%;total P,0.12%;total S,0.11% and atom13C,1.48%.

Table 1 Introduction of the study sites and the properties of the soils used in the incubation experiment

Table 2 Amounts of maize residue and nutrients (N,P,and S) added to soils in different treatments in the incubation experiment

The low nutrient addition rate was based on an idealized humification coefficient of 30% of residue C to form SOC with a C:N:P:S stoichiometry of 10 000:860:169:129 (Kirkbyet al.2013). The medium and high nutrient addition levels were set to two and three times that of the low nutrient addition level,respectively.The N,P,and S nutrients were added as ammoniumnitrate,potassium dihydrogen phosphate,and ammonium sulfate. The pH of the nutrient solution was adjusted to 7 with a 10-mol L?1sodium hydroxide solution.

The prepared residue and soil were evenly mixed on a smooth,soft plastic sheet. Then,the appropriate nutrient solution was added,or not added,to the mixture of soil and residue,which was followed by addition of distilled water to obtain 60% of field capacity.

2.3.Incubation conditions

Soil samples from both sites were treated with various levels of nutrients and were then separately incubated at 25°C in a dark incubator for 84 days. Each treatment contained 21 replicates of 40 g soil,and each replicate was placed in a plastic tube with open ends that measured 50 mm in diameter and 60 mm in height. The bottom of each tube was covered with nylon cloth (0.074 mm)to prevent spilling but which allowed gas exchange. The tubes were placed in 500-mL Mason jars with close-fitting three-way valves. Distilled water (10 mL) was spread on the bottoms of the jars to decrease water evaporation during incubation. Each tube was supported above the water by a circular plastic rack (50 mm in diameter and 30 mm high). CO2was collected regularly (at days 1,3,7,14,28,56,and 84) from three replicates of each treatment with a 50-mL syringe and measured using meteorological chromatography (GC6890,Agilent,USA).The CO2emission in each period was calculated using the following equation:

F=[(Qt-Q0)×V×M×273×1 000×24)]/[22.4×m×t×(273+T)]where F is the emission of CO2(mg CO2kg?1soil d?1);Qtand Q0are the CO2concentrations detected at room temperature (CO2/air,×10?1mol mol?1);V is the gas volume in the Mason jar (L);m is the soil sample dry weight (kg);t is the incubation time (h);M is the molar mass of CO2(g mol?1);and T is the incubation temperature.

Mixtures from each tube were spread out onto a soft,plastic sheet and were remixed after gas collection,and the soil moisture was readjusted with distilled water to maintain a constant tube weight. In addition,ambient air was repeatedly injected into the Mason jars with a 100-mL syringe before returning the tube its original location.

2.4.Sample analysis and calculation of NFC

Some of the soil samples were air-dried after completion of the incubation period. Any remaining partially degraded maize residue was removed by the dry-sieving winnowing method (Kirkbyet al.2013). Each replicate sample was passed through a 0.15-mm sieve and analyzed for C and atom13C (Appendix B) with an elemental analyzer that was interfaced to a PDZ Europa 20-20 isotope ratio mass spectrometer. We defined the new SOC that derived from the residue as NFC (mg kg?1soil) and this parameter was calculated using the following equations (Kirkbyet al.2014):

NFC=(Net-13C×100-Net-C×atom13Csoil)/(atom13Cresidueatom13Csoil)

where Net-13C (mg kg?1soil) is the net change in atom13C of sample,Net-C (mg kg?1soil) is the net change in SOC of the sample,atom13Csoil(%) is the atom13C of the initial soil,and atom13Cresidue(%) is the atom13C of the initial residue.

Net-13C=[(Csample×atom13Csample-Csoil×atom13Csoil)×1 000]/100

Net-C=(Csample-Csoil)×1 000

where Csample(g kg?1) is the SOC concentration of the sample,atom13Csample(%) is the atom13C concentration of the sample,and Csoil(g kg?1) is the SOC concentration of the initial soil.

The differences between NFC and Net-C were considered to be mineralized old C. Some fresh soil samples were analyzed for MBC at days 7,28,56,and 84 by a fumigation-extraction method (Vanceet al.1987). MBC was measured in the leachate with a C/N analyzer (multi N/C 3100,Analytik Jena AG,Germany). A conversion factor of 0.45 was applied to determine MBC(Wuet al.1990).

2.5.Soil extracellular enzyme activities (EEAs)

Five enzymes (Appendix C),including β-glucosidase (BG),cellobiohydrolase (CBH),β-N-acetylglucosaminidase(NAG),leucine aminopeptidase (LAP),and acid phosphatase (AP),were detected on days 7,28,56,and 84 by using a 96-well plate assay (Saiya-Corket al.2002). In brief,soil slurries were prepared by adding 1 g of fresh sample from each replicate to 100 mL of 50 mmol L?1sodium acetate buffer (pH=7.0) and shaken for 5 min by a magnetic stirrer. Then,200 μL of the slurries and 50 μL of sodium acetate buffer were pipetted into the sample and black wells. Fifty microliters of substrate solution and 200 μL of sodium acetate buffer were pipetted into the substrate wells. Fifty microliters of standard (MUB or AMC;Appendix C) and 200 μL of soil slurries were pipetted into the quench control wells. A total of 50 mL of standard and 200 μL of sodium acetate buffer were pipetted into the standard wells. All assay plates were incubated in the dark at 25°C for 4 h. After a 4-h incubation period,10 μL of 1 mol L?1NaOH was added to each well,and the fluorescence plates were read (excitation wavelength=365 nm and emission wavelength=450 nm) with a microplate detector (Synergy H1M,USA) to assess enzyme activities.

2.6.Soil microbial community abundance

The DNA from three replicates for each treatment was extracted from 0.25 g of soil at days 7,28,56,and 84 by a soil DNA isolation kit (TransGen Biotech,China). The abundances of the 16S V3-V4 and ITS1 genes were determined by quantitative realtime polymerase chain reaction (qRT-PCR) using an ABI7500 machine (Applied Biosystems,USA). The bacterial 16S rRNA gene was amplified using 338F(5′-GTACTCCTACGGGAGGCAGCA-3′) and 806R(5′-GTGGACTACHVGGGTWTCTAAT-3′) primers as previously described (Houet al.2017). The ITS gene was amplified using the primer sets ITS1F(5′-CTTGGTCATTTAGAGGAAGTAA-3′) and ITS1R(5′-TGCGTTCTTCATCGATGC-3′) (Liuet al.2017). The qRT-PCR system was as follows:2 μL of DNA,10 μL of 2×Master Mix (CoWin Biosciences,China),0.5 μL of 10 μmol L?1forward primer,and 0.5 μL of 10 μmol L?1reverse primer.

A 30-ng sample of purified DNA from each treatment was amplified using AP221-02 TransStart FastPfu DNA Polymerase (TransGen Biotech,China). Equal quantities of PCR products for each sample were analyzed using a MiSeq Platform (Illumina MiSep PE300/PE250,USA) to quantify certain microbial taxa (Actinobacteria,Proteobacteria,Acidobacteria,and Ascomycota). Quality control and statistical analysis of the paired-end (PE)reads were performed by Trimmomatic (ver.0.36). Bases of reads with a tail mass of 20 or less,overlap of PE reads less than 10 bp,chimeras,and unmatched sequences were then removed. After depleting barcodes and primers,clean tags were obtained by the further removal of chimeras and short sequences. For bacterial 16S rRNA,operational taxonomic units (OTUs) were defined by clustering clean tags with 97% similarity and were classified according to Silva database ver.128 using the UPARSE classifier (Quastet al.2013). For the fungal ITS region,clean tags after chimera checking were clustered at 97% similarity for OTUs according to Unite ver.7 using UPARSE (K?ljalget al.2013). Then,the absolute gene abundances of certain microbial taxa could be quantified based on the gene abundances of bacteria and fungi and the relative abundances of certain taxa.

2.7.Statistical analysis

All data were entered into and organized in Excel 2016 (Microsoft,Redmond,WA,USA). All results are reported as the mean±standard errors for the three replicates. First,one-way analysis of variance (ANOVA)and Duncan’s multiple comparisons (P＜0.05) were used to determine the significant differences among NFC,MBC,EEAs,and gene abundances under the different nutrient management practices with SPSS version 20(IBM,Chicago,IL,USA). Then,boosted regression tree(BRT) analysis was conducted to quantify the relative influences of 10 variables on the NFC using the GBM packages and code from Elithet al.(2008) in R version 3.3.3. Five variables (MBC,BG,AP,Acidobacteria,and Proteobacteria) were retained. Finally,structural equation modeling (SEM) was conducted using Amos Graphics version 21.0 (IBM,Chicago,IL,USA) to determine the direct or indirect pathways and to test whether nutrient supplementation affected NCF by regulating the gene abundances of major microbial taxa and their investment strategies in extracellular enzymes. All graphs were prepared with SigmaPlot version 12.5 (Systat Software,Chicago,IL,USA).

3.Results

3.1.Soil respiration

The CO2emissions from the control soils were relatively low (0.4-1.0 mg CO2-C kg?1soil d?1) in both soils throughout the incubation period (Fig.1). A clear increase in CO2emission rates was observed in soil with residue or with residue plus nutrient additions at the early stage(day 1 to day 28);they were 4.5 times (with a range of 2.5-6.5-fold) those of the control soils. On the 84th day of incubation,the respiration rates of both soils from each treatment tended to be consistent. The CO2emission rates of Hailun soil with residue or residue plus nutrient addition (from day 1 to day 14) were higher than those of Gongzhuling soil and ranged from 2.2-6.5 mg CO2-C kg?1soil d?1and 1.9-5.2 mg CO2-C kg?1soil d?1,respectively.

Fig.1 CO2-C efflux from soils from Gongzhuling,Jilin Province(A) and Hailun,Heilongjiang Province (B) with maize residue and nutrient addition or no addition over an 84-day incubation period. S,soil only;S_Re,soil with maize residue addition;S_Re_Lnu,soil with maize residue and low nutrient addition;S_Re_Mnu,soil with maize residue and medium nutrient addition;S_Re_Hnu,soil with maize residue and high nutrient addition.Error bars represent standard errors of the mean (n=3).

3.2.Changes in NFC

Compared with the control soils,residue input increased the NFC in both soils and caused mineralization of native SOC (Fig.2). NFC accumulation was more receptive to nutrient addition in the Gongzhuling soils than in the Hailun soils. By increasing the input of nutrients,the NFC increased significantly in Gongzhuling soil,with a high range of 1 155.9-1 722.4 mg kg?1soil. The NFC accounted for 30.4% of the residue C addition at the high nutrient input level. The NFC in Hailun soil was also enhanced by nutrient addition and ranged from 725.1 to 1 067.5 mg kg?1soil while accounting for 12.8-18.8%of the residue C addition. The inputs of residue and nutrients caused mineralization of native SOC,while more NFC was formed (Fig.2-B).

Fig.2 Effect of residue and nutrient addition or no addition on new soil-C formed (A) and old soil-C mineralized (B) at the completion of the 84-day incubation period.Values above the bars show changes in soil-C expressed as percentage of the residue-C added. GZL,Gongzhuling soil;HL,Hailun soil. S,soil only;S_Re,soil with maize residue addition;S_Re_Lnu,soil with maize residue and low nutrient addition;S_Re_Mnu,soil with maize residue and medium nutrient addition;S_Re_Hnu,soil with maize residue and high nutrient addition. Error bars represent standard errors of the mean (n=3).

3.3.Evolution of MBC,EEAs,and gene abundance

MBCSoil MBC increased first and then decreased over time across all treatments for both soils (Fig.3).Compared with the control soil,MBC significantly increased in soils with residue additions and continued to improve when combined with nutrient inputs. High nutrient inputs led to high MBC in the early period (days 7 and 28) for both soils. The MBC was higher in the soil from Hailun than in the Gongzhuling soil. However,the MBC obviously increased by nearly 1.6 times in the soil from Gongzhuling from day 7 to day 28,which was 1.1 times that for the soil from Hailun.

Fig.3 Microbial biomass carbon (mg C kg?1 soil) after 7,28,56,and 84 days of the incubation period in soils from Gongzhuling(A) and Hailun (B) with or without the input of maize residue and nutrients. S,soil only;S_Re,soil with maize residue addition;S_Re_Lnu,soil with maize residue and low nutrient addition;S_Re_Mnu,soil with maize residue and medium nutrient addition;S_Re_Hnu,soil with maize residue and high nutrient addition. Mean values with different letters are significantly different (P＜0.05). Error bars represent standard errors of the mean (n=3).

EEAsSoil EEAs exhibited a general downward trend over time in both soils (Fig.4). Compared with the control soil,the EEAs of BG+CBH,LAP+NAG,and AP in soils with residue inputs increased by 2.4-,2.9-,and 4.5-fold,respectively. The EEAs of BG and AP obviously increased with nutrient supplementation from low to high at different sampling times in both soils,and the EEAs of AP were higher than those of LAP+NAG in the Hailun soil(Appendices D and E). Soil C and N acquisition EEAs were higher in the Gongzhuling soil than in the Hailun soil,while AP in the Hailun soil exhibited stronger activities than in the Gongzhuling soil.

Fig.4 Soil extracellular enzyme activities (nmol g?1 soil h?1) across all treatments in soils from Gongzhuling,Jilin Province (A and C)and Hailun,Heilongjiang Province (B and D) at 7 and 84 days of the incubation period. BG,β-glucosidase;CBH,cellobiohydrolase;LAP,leucine aminopeptidase;NAG,β-N-acetylglucosaminidase;AP,acid phosphatase;S,soil only;S_Re,soil with maize residue addition;S_Re_Lnu,soil with maize residue and low nutrient addition;S_Re_Mnu,soil with maize residue and medium nutrient addition;S_Re_Hnu,soil with maize residue and high nutrient addition.Asterisks indicate that the enzyme activities increased significantly (P＜0.05) with nutrient supplementation from low to high. Details of extracellular enzyme activities over the complete incubation period are provided in Appendices D and E.

Higher ratios of C-to-N [(BG+CBH):(NAG+LAP)] and C-to-P [(BG+CBH):(AP)] acquisition EEAs were observed in the control soils than in the soils with other treatments in both soils (Fig.5). The ratios of C-to-N and C-to-P in the soils with residue were lower than those in soils with nutrient supplementation at days 7 and 28.

Gene abundancesThe gene abundances of Acidobacteria from each treatment increased over time,while the abundances of the other microbial taxa increased first and then slightly decreased (Fig.6).Addition of residue alone resulted in a relatively higher increase (relative to the combined supply of residue and nutrients) in the gene abundances of Acidobacteria(day 7) and Ascomycota (days 7 and 84) in both soils.As the level of nutrient input increased,the abundances of Proteobacteria (both soils) and Actinobacteria(Gongzhuling soils) increased. Higher gene abundance values for these four taxa were observed in the soil from Hailun than in the soil from Gongzhuling. Furthermore,the gene abundances of the four taxa increased far more in the Gongzhuling soil (1.2-1.8-fold) than in the Hailun soil (1.0-1.3-fold) from day 7 to day 28.

3.4.Driving factors and pathways of NFC

The BRT results explained 75% (Gongzhuling soils)and 84% (Hailun soils) of the variations in NFC (Fig.7).BG,AP,MBC,Acidobacteria,and Proteobacteria were responsible for 27.8,18.5,14.7,8.1,and 6.1% of the NFC in Gongzhuling soil,respectively,and were responsible for 25.9,29.5,10.1,13.9,and 5.0% of the NFC in Hailun soil,respectively. The SEM analysis suggested that nutrient N (P and S),MBC,and BG directly influenced the NFC in Gongzhuling soil (Path coefficients:0.52,-0.19,and -0.54) (Fig.8-A). Acidobacteria,which was significantly associated with the input of nutrient N,directly (Path coefficients:0.18) and indirectly affected NFC through MBC and BG (Path coefficients:-0.46 and-0.62). In Hailun soil,nutrient N,MBC,and AP directly affected NFC (Path coefficients:0.42,-0.21,and -0.48),while Acidobacteria affected NFC indirectly by MBC and AP (Path coefficients:-0.86 and -0.16) (Fig.8-B).Path analysis explained 80% of the NFC variations in Gongzhuling soil (R2=0.80) and 76% of the variations in Hailun soil (R2=0.76).

Fig.7 The relative influence (%) of microbial biomass carbon (MBC),β-glucosidase (BG),acid phosphatase(AP),Acidobacteria,and Proteobacteria for the boosted regression tree model of new soil-C formed in Gongzhuling,Jilin Province (A) and Hailun,Heilongjiang Province (B).BG,β-glucosidase;AP,acid phosphatase;MBC,microbial biomass carbon. The variables with a relative influence of＞3% are shown.

4.Discussion

4.1.NFC response to residue input

Our study found that crop residue addition was beneficial for the accumulation of NFC in both soils which is consistent with results of previous studies that have shown that,generally,inputs of LOM,such as crop residue,can increase overall SOC (Yanet al.2011;Hanet al.2018).The increased amount of NFC accounted for 10.7-14.9%of the total residue C input,which was consistent with the ranges mentioned in previous studies with LOM addition(Kirkbyet al.2014;Creameret al.2016). Soil disturbance and residue incorporation also refer to obtain nutrients from native SOM when they decompose LOM (Fontaineet al.2003). Moreover,inputs of C-rich residue induce nutrient limitation and stimulate microbes to mine more nutrients from the labile fractions (in both native SOM and residue) by deploying more extracellular enzymes (Waringet al.2014). In a nutrient-poor environment that is caused by adding residue,oligotrophs tend to contribute more to the decomposition of exogenous LOM due to their dominant advantage (Fanget al.2018). In the present study,this was supported by the higher gene abundances of Acidobacteria and Ascomycota in both soils when residue was added alone (relative to soils with nutrient supplementation) at the early stage of incubation (Fig.6).Furthermore,addition of residue alone led to a decrease of 8-16% in the ratios of C to N and C to P EEAs at day 7 and day 28 (relative to soils with nutrient supplementation,Fig.5). This phenomenon also indicated that residue addition could induce nutrient limitation and promote the dominant taxa to engage in stronger N and P acquisition relative to C acquisition (Waringet al.2014). Therefore,further transfer of residue C to NFC might be limited by the nutrient mining of the dominant microbial taxa (mainly Acidobacteria or Ascomycota) in soils with residue alone.

Fig.5 Carbon (C) acquisition enzyme activity to nitrogen (N) acquisition enzyme activity (upper graphs) and phosphorus (P)acquisition enzyme activity (lower graphs) ratios with or without the input of residue and nutrients in the soil from Gongzhuling,Jilin Province (A and C) and Hailun,Heilongjiang Province (B and D) after 7,28,56,and 84 days. S,soil only;S_Re,soil with maize residue addition;S_Re_Lnu,soil with maize residue and low nutrient addition;S_Re_Mnu,soil with maize residue and medium nutrient addition;S_Re_Hnu,soil with maize residue and high nutrient addition. C,N and P acquisition enzyme activity was measured by the activity of β-glucosidase (BG) and cellobiohydrolase (CBH),leucine aminopeptidase (LAP) and β-N-acetylglucosaminidase(NAG),and acid phosphatase (AP),respectively. Mean values with different letters are significantly different (P＜0.05). Error bars represent standard errors of the mean (n=3).

Fig.6 The gene abundance of major microbial taxa (×106 copy g?1 soil) across all treatments in the soils from Gongzhuling,Jilin Province (A,C,E,and G) and Hailun,Heilongjiang Province (B,D,F,and H) on days 7,28,56,and 84 of the incubation periods.S,soil only;S_Re,soil with maize residue addition;S_Re_Lnu,soil with maize residue and low nutrient addition;S_Re_Mnu,soil with maize residue and medium nutrient addition;S_Re_Hnu,soil with maize residue and high nutrient addition. Mean values with different letters are significant different (P＜0.05). Error bars represent standard errors of the mean (n=3).

4.2.NFC response to the combined supply of residue and nutrients

Compared to adding residue alone,the amount of NFC was increased by 1.2-to 2.0-fold with nutrient supplementation (Fig.2). By increasing the levels of nutrient inputs,the consumption of native SOC was exacerbated by PE,and NFC production increased due to the efficient humification of residue C based on a stable C:N:P:S stoichiometry (Kirkbyet al.2014). Compared with the addition of residue alone,mineralization of native SOC increased by 13-81%,and the amount of NFC increased by 20-103% under the combined inputs of residue and nutrients (from low to high),which is in agreement with previous studies (Kirkbyet al.2013;Weiet al.2020).Hence,the increased nutrient availability enhanced C consumption of the native SOM by microbial activity,and the corresponding microbes were stimulated to rapidly form additional new SOC by using an exogenous labile C substrate (Kirkbyet al.2013).

The NFC amounts were positively related to the nutrient supplementation levels (Fig.8),which could be attributed to the theory of microbial stoichiometry decomposition(Chenet al.2014). As the nutrient supplementation levels increased,the gene abundance values of Acidobacteria(except on the 7th day) and Proteobacteria in both soils increased by an average of 10% (4-17%) at different sampling times. Path analysis also suggested that the increased nutrient availability promoted growth of Acidobacteria and Proteobacteria,which could increase the accumulation of microbial biomass. Moreover,the ratios of native SOC that was mineralized to NFC decreased from 0.60 to 0.45 with increasing levels of nutrient supplementation,which suggested that decomposers preferred to use supplemental nutrients and/or decompose exogenous LOM for nutrients rather than use native SOM (Fanget al.2018). Thus,a greater proportion of residue C along with N,P,and S was utilized by microbes and transferred into NFC as microbial detritus,metabolites,and secretions (Miltneret al.2009;Liang and Balser 2011). The positive correlations among nutrient availability,Acidobacteria,and NFC in the SEM analysis also supported this view (Fig.8). The gene abundances of Acidobacteria increased over time in our study,consistent with the view that Acidobacteria are generally considered to dominate residue decomposition in the late stage (Banerjeeet al.2016). This phenomenon could be related to the consumption of labile residue C,while native SOC gradually became the C substrate for microbes.

Higher MBC and EEAs (e.g.,BG and AP) were observed in this study in the mixtures of soil and residue that received nutrient supplementation than in mixtures that received no nutrients (Fig.3;Appendices D and E).Rapid MBC accumulation was observed with increasing nutrient input levels,especially in the early stage,which could provide a potential pool for expanding SOC stocks(Huanget al.2019). In the samples with high nutrient inputs,soils were rich in nutrients and labile C sources,which stimulated greater BG production to enhance microbial C acquisition,which thereby contributed to a sustained accumulation of MBC (Fosteret al.2016;Stockaet al.2019). This was supported by the pathways by which nutrient availability directly affected BG and indirectly affected the variations in MBC. Hence,the increased availability of N,P,and S could meet the requirements for microbial stoichiometry decomposition,and more residue C was transferred by the decomposers into soils in the nutrient-rich environment (Weiet al.2020). The gene abundances of Acidobacteria and Proteobacteria increased as the levels of nutrient addition increased,which also supported the increased utilization of residue C by microbes. Soil EEAs exhibited a decreasing trend over time,and the stoichiometric ratios of C to P or N EEAs decreased slightly over time,which indicated that the extent of nutrient limitation gradually increased due to the activities of microbes that were acquiring residue C to form new soil C (Xuet al.2017). SEM analysis also showed that microbial C acquisition was limited more by AP EEAs,which thereby affected the generation of NFC (Fig.8). Microbial growth could face limitations of labile C resources with the reduction in labile residue C over time. Overall,this study showed that increased N,P,and S availability promoted the amount of NFC,which was manipulated by Acidobacteria and Proteobacteria,and their investment in BG and AP for microbial C acquisition from crop residue.

4.3.NFC response to soil status

Path analysis showed that nutrient inputs directly affected Acidobacteria and Proteobacteria and the strategies of investment in C and P acquisition enzymes,which thus regulated NFC in both soils (Fig.8). The gene abundance values of the major taxa and MBC were higher in Hailun soil than in Gongzhuling soil,but they were substantially improved only in Gongzhuling soil after nutrient additions(Figs.3 and 6). This result indicated that microbes were more sensitive to nutrient supplementation in Gongzhuling soil,and that exogenous C utilization was enhanced to promote their own growth (Marschneret al.2015;Tianet al.2020). The higher Path coefficients of nutrients for Acidobacteria,AP,and MBC that were observed in Gongzhuling soil also supported this view. Long-term cultivation and very large outputs of agricultural products have resulted in the declining stability of the SOC pool and continuous loss of soil C over the last 30 years.Additionally,the fields at Gongzhuling had been under crop production for at least 150 years (Yanet al.2011;Wang Xet al.2018). Long-term cultivation and soil fertility decline in soil from Gongzhuling would have led to greater demand for exogenous C by microbes to form MBC and NFC with nutrient addition (Kirkbyet al.2013;Songet al.2013),than in Hailun soil where the SOC was relatively higher with a cultivation time of～60 years.Nutrient supplementation directly and positively affected the NFC,and the total coefficients of nutrient were 0.89 and 0.58 in Gongzhuling and Hailun soils,respectively,which indicated that nutrient addition was more conducive for stimulating microbes to metabolize exogenous C in Gongzhuling soil than in Hailun soil. In addition,higher C acquisition EEAs were observed in the soils of Gongzhuling,which would also lead to more production of MBC and NFC (Jianet al.2016). Moreover,the P acquisition EEAs maintained high levels in both soils(especially in Hailun),which indicated that microbial C acquisition was regulated more by P than by N (Fatemiet al.2016). The direct relationships between AP and BG,MBC,Acidobacteria,and Proteobacteria shown in the path analysis also proved this point. Furthermore,variations in AP activities had the greatest influence and directly affected the NFC in Hailun soil,which suggested that the lower soil pH may cause P to be adsorbed on the surfaces of Fe and Al hydroxide and thereby reduce soil P availability (Zhouet al.2016).

Fig.8 Path analysis of the nutrient input,microbial biomass carbon,extracellular enzyme activities and gene abundance on newly-formed soil carbon in Gongzhuling,Jilin Province (A) and Hailun,Heilongjiang Province (B). N,inorganic nitrogen addition;Aci/ Pro,variation in gene abundance of Acidobacteria/Proteobacteria from day 7 to day 84;MBC,variation in microbial biomass carbon from day 7 to day 84;BG/AP,variation in β-glucosidase/acid phosphatase activity from day 7 to day 84;NFC,newlyformed soil carbon. Numbers next to the arrows are the standardized path coefficients. A solid-line Path indicates that the effect is significantly positive.

The release of N,P,and S from the residue and native SOM by microbial metabolism and their desorption or dissolution from mineral nutrient reserves could affect the conversion of residue C to soil C (Manzoniet al.2010;Sarkeret al.2018). Therefore,the higher bioavailability of released nutrients in the soil from Gongzhuling could be attributed to the enhanced mineralization and/or chemical mobilization from the unstable SOM reserves,which would also promote the synergistic conversion of residue C along with the activated nutrients (Brookset al.2011;Creameret al.2015). Furthermore,the total amount of native SOC that was mineralized in Gongzhuling soil was 1.3 times greater than that for Hailun soil,which confirmed this view (Fig.2). In conclusion,the NFC amounts were higher in Gongzhuling soil than in Hailun soil emphasizing the importance of certain soil properties,such as the C level and pH,in determining the amount of new C formed following residue incorporation.

5.Conclusion

This study enhances the understanding of the underlying microbial mechanisms that contribute to NFC in response to integrated inputs of crop residue and nutrients in soils with different properties (such as SOC and pH). Higher gene abundances of Acidobacteria and Ascomycota and lower ratios of C:N or C:P enzyme activities were observed in soils treated with residue alone (relative to soils treated with residue and nutrient inputs) in the early stage of incubation,which suggests that oligotrophs contributed more to residue decomposition in a nutrientpoor environment and invested in more nutrient acquisition enzymes for nutrient mining rather than C mining.As the level of nutrient supplementation increased,residue inputs caused positive NFC accumulation,which was conducted by Acidobacteria and Proteobacteria and their investment in BG and AP for microbial C acquisition from crop residue. These results suggest that nutrient supplementation could meet the demands of microbial decomposition and promote NFC based on a stable C:N:P stoichiometry. A direct relationship among AP and BG,MBC,and Acidobacteria and the pathways of BG that indirectly affected MBC through AP were observed in both soils,which indicated that P availability could be crucial for the adjustment of microbial C acquisition strategies.The amount of NFC was higher in Gongzhuling soil than in Hailun soil,which indicated that soil properties such as C level and pH,also affected NFC. This research reveals the key steps in SOC dynamics following residue incorporation in soils and provides a basis for optimizing nutrient management in agroecosystems.

Acknowledgements

We thank colleagues at the long-term monitoring sites(Gongzhuling,Jilin,China;Hailun,Heilongjiang,China)for providing us with the soil samples. This work was financially supported by the Agro-scientific Research in the Public Interest of China (201503122),the Agricultural Science and Technology Innovation Program of the Chinese Academy of Agricultural Sciences (CAASXTCX2016008),and the National Natural Science Foundation of China (41620104006).

Declaration of competing interest

The authors declare that they have no conflict of interest.

Appendicesassociated with this paper are available on http://www.ChinaAgriSci.com/V2/En/appendix.htm