APP下载

Responses of soil phosphorus pools accompanied with carbon composition and microorganism changes to phosphorus-input reduction in paddy soils

2021-03-17JiahuiYUANLeiWANGHaoCHENGuangleiCHENShenqiangWANGXuZHAOandYuWANG

Pedosphere 2021年1期

Jiahui YUANLei WANG∗Hao CHENGuanglei CHENShenqiang WANGXu ZHAO and Yu WANG∗

1State Key Laboratory of Soil and Sustainable Agriculture,Changshu National Agro-Ecosystem Observation and Research Station,Institute of Soil Science,Chinese Academy of Sciences,Nanjing 210008(China)

2University of Chinese Academy of Sciences,Beijing 100049(China)

3Organic Food Development Center,Nanjing Institute of Environmental Sciences,Ministry of Environmental Protection,Nanjing 210042(China)

4College of Resources and Environmental Sciences,Nanjing Agricultural University,Nanjing 210095(China)

(Received May 11,2020;revised September 18,2020)

ABSTRACT In rice-wheat rotation systems,changes in soil phosphorus(P)pools and microorganisms in rice-growing seasons have been studied;however,further investigations are required to test whether these indexes exhibit different responses in wheat-growing seasons.Additionally,such studies need to include potential variations in soil carbon(C)structure and microbial community composition.In this study,a long-term rice-wheat rotation P-input reduction experiment was conducted to observe the variations in soil P pools and C composition in the 7th wheat season and to investigate the responses of soil enzyme activity and microbial communities.Four P fertilization treatments were included in the experiment,i.e.,P application for rice season only(PR),for wheat season only(PW),and for both rice and wheat seasons(PR+W)and no P application in either season(Pzero).Compared with PR+W treatment,Pzero treatment significantly decreased(P<0.05)labile and stable P pools.Different P fertilization regimes altered soil microbial community composition and enzyme activity,whereas C composition did not vary.However,PW treatment resulted in relatively more O-alkyl-C than PR treatment and the highest number of microorganisms.Besides,the higher ratios of fungi/bacteria and Gram-positive bactetia/Gram-negative bactetia were related to labile C pools,particularly O-alkyl-C,as opposed to recalcitrant C.Our results clarified the status of soil P pools,C chemistry,and the response of microorganisms under dry-farming conditions in the P input-reduced rice-wheat rotation system.

Key Words: 13C nuclear magnetic resonance,dry farming,enzyme activity,microbial community composition,phospholipid fatty acid,phosphorus fractionation,wheat season

INTRODUCTION

Phosphorus(P)is a pivotal limiting nutrient for crop yield and quality.The application of organic and chemical fertilizers is a primary method of increasing soil fertility and crop yield in farmland.However,improper application of fertilizers can exacerbate P loss from soil,increasing eutrophication risk(Fischeret al.,2017;Watts and Way,2019).Therefore,appropriate fertilization methods are important for the sustainable development of agriculture.Our previous studies aimed to propose a feasible scheme for reducing the application rates of phosphate fertilizer in the rice-wheat rotation farmland system of Taihu Lake region,and revealed that P application only in wheat season had no significant effect on crop yield(Wanget al.,2016a,b).On one hand,it could ensure that the soil can provide adequate labile and moderately labile P;on the other hand,the activities and functions of soil microorganisms are also adequate in facilitating the transformation of P in wheat soil(Wanget al.,2016a,b;Yuanet al.,2018).However,under dry-farming conditions,the responses of soil P pools and the potential accompanying shifts in soil microbial communities remain poorly understood.

The characteristics and activity of soil microbial communities are sensitive to fluctuations in soil moisture(Liet al.,2018).P transformation and transport in wheat season aligned with the variations in soil microbial community composition(Tenget al.,2018).Additionally,soil microorganisms also play a crucial role in decomposition and mineralization of different inorganic P(Pi)and organic P(Po)proportions,respectively,through chemical and biological mechanisms,which result in different levels of P availability(Jinget al.,2017).Extracellular enzymes play an irreplaceable part in biogeochemical cycling and solubilization of soil Pi,as well as formation and mineralization of Po(Richardson and Simpson,2011;Lianget al.,2017),which both directly influence P availability and crop P uptake.In addition,phosphatase promotes the hydrolysis and mineralization of specific P compounds.For instance,it catalyzes the hydrolysis of diester and monoester(Tianet al.,2016).Soil arbuscular mycorrhizal fungi(AMF)have been elucidated as key organisms participating in transformation of P under different P fertilization treatments(Wanget al.,2016b).

Soil organic carbon(SOC)is an important controller of soil microbial communities in paddy fields(Liet al.,2018;Liuet al.,2019;Weiet al.,2019).Meanwhile,organic and chemical fertilizer application alters biogeochemical processes that influence SOC accumulation and decomposition(Mackenzieet al.,2002;Vitouseket al.,2010;Wanget al.,2017).Additionally,the balance between soil carbon(C)decomposition and formation influences the variations in soil C storage forms(Yeet al.,2018).Microbial community composition and activity,aggregate stability,mineral availability,and soil acidity influence soil C dynamics(Wanget al.,2014;Jianet al.,2016;Yuet al.,2017;Averill and Waring,2018).Organic fertilizers have been reported to elevate the quantity and recalcitrance of SOC,whereas mineral fertilizers have relatively weak effects(Liet al.,2014;Shresthaet al.,2015;Luoet al.,2019).Wanget al.(2017)demonstrated that organic and chemical fertilization can significantly increase the SOC content,but chemical fertilization treatment does not change the C structure.Fertilizer application could also directly affect the metabolic capacity of soil bacterial communities,which could decelerate recalcitrant C decomposition,and potentially result in advanced soil C sequestration rates(Ramirezet al.,2012).

However,the considerable variability in the chemical nature of soil C,based on different forms and compounds present in the soil,influences microbial regulation of C cycling activities(Baldocket al.,2004;Nget al.,2014).Moreover,the complex forms of soil C influence the compositions,capacities,and functions of soil microorganisms(Bowleset al.,2014).Soil microbes are the major decomposers of soil C and are invaluable regulators of nutrient cycling activities(Ltzowet al.,2006).Therefore,both the amounts and forms of soil C influence soil microorganisms that exploit the different forms of soil C(Nget al.,2014).Previous studies have shown that Gram-negative bacteria mainly exploit simple C groups(Holding,1960),resulting in higher amounts of recalcitrant C,particularly soil aryl-C.Additionally,a suite of soil enzymes have been confirmed to be connected with transformation,mineralization,and degradation of soil organic matter(Shi,2010;Sinsabaugh,2010).Using13C nuclear magnetic resonance(NMR)spectroscopy,Wanget al.(2017)demonstrated that changes in enzyme activity were closely associated with soil C proportions,particularly with alkyl-C and ketone-C,as opposed to C quantity.Similarly,the addition of different crop residues could alter the chemical forms and decomposition dynamics of SOC(Xu Y Het al.,2017),resulting in variations of soil microbial diversity and community structure(Linet al.,2018).Information on the effects of different fertilization methods remains scarce,particularly for P on soil C cycling rates,and in addition to how such fertilization activities influence microbial species and their functions.Microbial function is the linkage between soil C chemistry and P pools.In arable soils,Jinget al.(2017)observed an inverse relationship between P deficiency and C retention,and a positive relationship between labile C and P,which could alter soil P availability through pool cycling driven by specific P-solubilizing taxonomic groups.Revealing the status of soil P availability and C chemistry,as well as the potential underlying microbial mechanisms,is essential in addressing the numerous puzzles associated with the uniqueness of P cycling dynamics(Reedet al.,2015).

Therefore,based on a long-term rice-wheat rotation Pinput reduction field experiment in Taihu Lake region,the changes in different P pools and the variations in microbiological response were investigated during the 7th wheat season under different P fertilization treatments,and the response of C structure and its influences on microbial community and enzyme activity dynamics were supplemented.Our objectives were:i)to observe the status of various soil P pools;ii)to investigate whether different phosphate fertilizers influence soil C structure;and iii)to characterize the variations in microbial species and functions affected by soil P speciation and C composition.

MATERIALS AND METHODS

Field experiment

The field experiment was carried out at Yixing Agro-Environment Research Base,the National Agroecosystem Observatory and Research Station of Changshu(119°54′E,31°16′N),Jiangsu Province of China.The study site is located near Taihu Lake,and has a mild average yearly temperature of 15.7°C and an average annual rainfall of 1 054 mm.The region practices a typical rice-wheat cropping system,and the experiments were set up in June 2010 when the rice-growing season began.The initial soil physical and chemical parameters have been listed previously(Zhuet al.,2016),and the details of the experiments have been introduced in our previous research(Wanget al.,2016b;Yuanet al.,2018).Briefly,the four different P fertilization systems included P application in wheat season only(PW),in rice season only(PR),and in both wheat and rice seasons(PR+W,current farming practice)and no P application in either season(Pzero).The dose of P fertilizer used per season was 40 kg P2O5ha−1.The urea fertilization rate was 240 kg N ha−1,30% of the urea was applied basally,40% as top dressing in tillering stage,and the remanent 30% applied in jointing stage.The basal application rate of the potassium fertilizer was 60 kg K2O ha−1.The field was randomly divided into 12 plots for the four treatments.Soil samples were collected during the 7th year of wheat harvest season and then air-dried,followed by sieving(2 mm)for further use.

Sequential extraction of soil P

The Tiessen and Moir(1993)sequential extraction method modified by Wanget al.(2016b)was used to process the soils sampled in the 7th year in the wheat harvest season.Air-dried soil samples(0.5 g)were processed sequentially to obtain different P fractionations as follows:addition of deionized water with one resin strip(Sinopharm Chemical Reagent Co.,Ltd,China)(resin-P),0.5 mol L−1NaHCO3(pH 8.5)(NaHCO3-P),0.1 mol L−1NaOH(NaOH-P),1 mol L−1HCl(HCl-P),and H2SO4followed by digestion with H2O2at 360°C(residual-P).The soil samples were subjected to reciprocating shaking at 180 r min−1for 16 h after the addition of each solution,followed by centrifuging for 10 min at 8 000 r min−1at 0°C and filtering through Whatman No.42 filter paper.The NaHCO3and NaOH extractions were divided into two portions to quantify contents of total P and Pi.Soil Pi was measured directly from one aliquot of the extracts,whereas total P was measured after the digestion of another aliquot with 10 mL 0.9 mol L−1H2SO4and 0.5 g ammonium persulphate at 330°C.Soil Po was calculated based on the difference between total P and Pi.Ultraviolet spectrometer(UV 2500,Shimadzu,Japan)was used to measure the P concentrations in supernatants and digested sample based on the ascorbic acid molybdenum blue method(Murphy and Riley,1962).In addition,the above P forms were divided into three categories:labile P(LP,including resin-P,NaHCO3-Pi,and NaHCO3-Po),moderately labile P(MLP,including NaOH-Pi and NaOH-Po),and stable P(STP,including HCl-P and residual P)(Tiessen and Moir,1993;Wanget al.,2016b).

Phospholipid fatty acid(PLFA)analysis

Soil microbial community composition under different P treatments was analyzed by measuring soil PLFA concentrations.Chloroform-methanol-citrate buffer mixture(1:2:0.8,volume/volume/volume;0.15 mol L−1,pH 4.0)was used to extract the PLFAs from the lyophilized soil samples(Frostegrdet al.,1993;Wanget al.,2016b).The obtained fatty acid methyl esters were isolated and distinguished on a 7890 N gas chromatograph(Agilent,USA)equipped with a MIDI Sherlock microbial identification system(version 4.5,MIDI,USA).Chloroform,acetone,and methanol were added to the middle of the SupelcleanTMsolid phase extraction tubes that had been washed with acetone and chloroform to dissolve the glycolipid.Following saponification,methylation,and extraction,100µL chromatographically puren-hexane was used to dissolve the sample for further test.Based on 19:0 fatty acid methyl ester ratios,the concentration of each PLFA was calculated and expressed asµg g−1.

Soil fungi,Gram-positive bacteria(G+),Gram-negative bacteria(G−),AMF,and actinomycetes were identified based on their specific PLFA biomarkers.In detail,they included the sum of i14:0,i15:0,a15:0,i16:0,i17:0,and a17:0 for G+;16:1w9c,16:1w7c,17:1w8c,18:1w5c,18:1w7c,cy17:0,and cy19:0 for G−;10 Me16:0,10 Me17:0,and 10 Me18:0 for actinomycetes;18:2w6c and 18:3w6c for fungi;and 16:1w5c for AMF(Frostegrdet al.,1993;Olsson,1999;Zelles,1999;Wanget al.,2016b).To assess changes in microbial community composition caused by different fertilizers,the ratios of fungi/bacteria(F/B)(bacteria is the sum of G+,G−,15:0,and 17:0)and G+/G−were further calculated(Elfstrandet al.,2007;Wanget al.,2017).

Enzyme activity

A series of hydrolases and oxidoreductases associated with soil P and C cycles were assessed to quantify the functional capacities of microbes.Soil acid(ACP)and alkaline phosphatase(ALP)activities were determined as described by Tabatabai(1994).The activity ofβ-glucosidase(β-GC)was calculated by quantifyingp-nitrophenol released from the samples using a colorimetric method(Eivazi and Tabatabai,1988).Soil sucrase(SC)and urease(UE)activities were analyzed based on sodium thiosulfate titration and the hypochlorite-alkaline phenol method as described by Guan(1986).Nitrate reductase(NR)activity was determined by a colorimetric method(Guan,1986)and cellulase(CL)activity according to Deng and Tabatabai(1994).Catalase(CAT)activity was analyzed by measuring the amount of O2evolved following H2O2decomposition(Johnson and Temple,1964),whereas polyphenol oxidase(PPO)activity was analyzed according to the method of Montgomery and Sgarbieri(1975).The enzyme activities were presented asµmol g−1d−1.In addition,the geometric means of the measured enzyme activities(GMea)and hydrolase(GMh)and oxidoreductase activities(GMor)(Garca-Ruizet al.,2008;Wanget al.,2017)were calculated as follows:

13C NMR spectroscopy

The13C NMR spectroscopy was employed to analyze soil C composition.All 12 soil samples from the wheatharvest stage were analyzed.Soils were processed 8 times with 50 mL 10% HF followed by 4 times with H2O2.Subsamples were packed in a 4-mm zirconia rotor and spun at 6 kHz using the Bruker Avance III 400 spectrometer(Bruker BioSpin,Germany)with a resonance frequency of 100.625 MHz for13C.To decouple the pulse program,we combined 1 000µs of contact time,0.5 s of relaxation delay,a ramp-contact,and a small phase incremental alternation with 64 steps(SPINAL-64)(Wanget al.,2017).Eight soil C components were estimated by calculating corresponding chemical shift values of the spectra obtained according to Shresthaet al.(2015).Primary chemical compositions of C were expressed as follows:alkyl-C(0-45 ppm),N-alkyl-C(45-60 ppm),O-alkyl-C(60-90 ppm),di-O-alkyl-C(90-110 ppm),aryl-C(110-145 ppm),O-aryl-C(145-165 ppm),carbonyl-C(165-190 ppm),and ketone-C(190-220 ppm).All manual integration procedures were completed using the MestReNova software.The ratios of soil alky-C/O-alkyl-C and aryl-C/O-alkyl-C were considered as the indicators of decomposition rates(Shresthaet al.,2015;Wanget al.,2017).

Statistical analysis

Analysis of the influence of four fertilization methods on soil parameters was performed using the Duncan method(P=0.05)and one-way analysis of variance by SPSS 17.0.Redundancy analysis(RDA)was conducted using CANOCO 4.5 to further explore relationships between C composition and PLFAs coupled with soil microbial enzyme activity,and between P pools and PLFAs combined with soil microbial enzyme activity.The multivariate regression tree(MRT)was constructed using R 2.15.2 to simultaneously explore nonlinear correlations within the measured parameters,which may not have been captured by RDA.

RESULTS

Soil P fractionation changes

The differences in soil P fractionations in the four fertilization treatments after the 7th(2017)wheat harvest are shown in Fig.1.In three soil P categories,the LP pool accounted for 5.8%-16.3% of the total P concentration,the MLP pool ranged from 15.6% to 21.2%,and the STP pool yielded the highest P concentration with a 71.9%-74.1% total P range.The mean soil total P ranged from 275.7 to 347.3 mg kg−1.Compared to PR+W treatment,Pzerotreatment resulted in lower(P<0.05)LP and STP concentrations,particularly for resin-P,NaHCO3-Pi,and HCl-P,whereas no significant differences were found for the treatments of P fertilization just for single cropping season(i.e.,PR and PW).

Phospholipid fatty acids and soil enzyme activity

Fig.1 Sequentially-extracted P fractionations in the paddy soil during the 7th wheat season under four P fertilization treatments,including P application in wheat season only(PW),in rice season only(PR),and in both rice and wheat seasons(PR+W,current farming practice)and no P application in either season(Pzero).Sequentially-extracted P fractionations included resin-extractable P(resin-P),NaHCO3-extractable P(NaHCO3-P),NaOH-extractable P(NaOH-P),HCl-extractable P(HCl-P),and residual P.The NaHCO3-P and NaOH-P are subdivided into inorganic P(NaHCO3-Pi and NaOH-Pi)and organic P(NaHCO3-Po and NaOH-Po).These P fractionations are divided into three categories:labile P(LP,including resin-P,NaHCO3-Pi,and NaHCO3-Po),moderately labile P(MLP,including NaOH-Pi and NaOH-Po),and stable P(STP,including HCl-P and residual P).Vertical bars indicate standard deviations of the means(n=3).Bars with the same letter(s)are not significantly different among P treatments at P<0.05.

Soil microbial community composition and their functions under the four P application regimes were examined using 19 PLFA biomarkers in the 7th wheat season(Fig.2a).Total microbial PLFA concentrations were measured within the range of 58.11-264.1µg g−1.As illustrated in Fig.2,PW treatment yielded the highest total relative concentrations of all the PLFAs,which were somewhat similar to the levels in PR+W treatment.In addition,the total relative abundances under the treatments of P fertilization in wheat season(PW and PR+W)were significantly greater than those under the treatments for no P fertilizer applied in wheat season(PR and Pzero).Further as listed in Table I,different P fertilization regimes affected soil F/B ratio values significantly(P<0.05)as follows:the ratio for PW treatment was the highest(0.181-0.250),whereas 0.113-0.165,0.073-0.090,and 0.079-0.100 were the ratios for the PR+W,Pzero,and PR treatments,respectively.Similar results obtained for the G+/G−ratio are presented in Table I:PW treatment increased the G+/G−ratio.

The changes of soil enzyme activity are illustrated in Fig.2b.After 7 years of rice-wheat rotation,P fertilization significantly reduced(P<0.05)soil NR activity,although it facilitated soil UE and SC activities(Fig.2b).The CL and PPO were the two most active enzymes among all the tested enzymes,and they exhibited the highest values in PW treatment.In addition,based on the geometric mean of the measured soil enzyme activity,no P fertilizer application in rice season(Pzeroand PW treatments)induced higher(P<0.05)GMor value,but it did not influence the GMea and GMh values(Table I).

Relationships of enzyme activity and PLFAs with P fractionations

The relationships of P fractionations with soil microbial community composition and enzyme activities were explored using RDA analysis(Fig.3).The two soil microbial parameters including microbial community composition and enzyme activities accounted for 80.2% and 95.6%,respectively,of soil P variations.Variations in P fractionation explained more changes in soil enzyme activity than in PLFAs.Specifically,resin-P explained a higher variation in soil microbial community composition when compared with other P species(Fig.3a).Among all the PLFAs tested,AMF was positively and highly correlated with LP and STP,especially with resin-P and HCl-P(Fig.3c).In addition,soil hydrolase activity contributed to more variations,particularly for PPO and CL(Fig.3b),whereas the variations in P fractionation were more related to the variation in soil hydrolase activity than that in oxidoreductase activity(Fig.3d).

Relative abundances of C components

The relative abundances of eight C components are illustrated in Fig.4.Until the 7th wheat harvest season,the four P fertilization treatments generally had no significant impacts on the relative abundances of different C groups(Fig.4).Nevertheless,PW treatment enhanced soil O-alkyl-C content,which represents a crucial form of labile C.In addition,the alkyl-C/O-alkyl-C and aryl-C/O-alkyl-C ratios were the lowest in PW treatment at 1.004±0.033 and 0.345±0.017,respectively(Table II).

Linking enzyme activity and PLFAs to C composition

The relationships between soil C components and soil microbial community composition combined with enzymeactivity were further explored using MRT and RDA analyses(Fig.5).In the ordination space of relative abundances of C functional forms and PLFAs,more than 80% of the constraint variance was explained by the RDA axis(Fig.5a).Soil microbial community composition under PW treatment was highly positively linked to the relative abundance of O-alkyl-C,the most influential parameter,although it had an inverse relationship with alkyl-C.

Fig.3 Redundancy analysis(RDA)for the relationships between soil P fractionations and soil microbial community composition(a and c)and between soil P fractionations and enzyme activities(b and d)during the 7th wheat season under four P fertilization treatments,including P application in wheat season only(PW),in rice season only(PR),and in both rice and wheat seasons(PR+W)and no P application in either season(Pzero).Affected factors are denoted by red arrows,and affecting factors by blue arrows.See Fig.1 for the detailed descriptions of the abbreviations of soil P fractionations and Fig.2 for those of soil microbial community composition and enzyme activities.

In the MRT analysis,based on the relative abundance of soil O-alkyl-C,soil PLFAs could be divided into two groups,with higher contents under PW treatment,and 43.69% of the variance in soil microbial community composition was accounted for by the PW arm(Fig.5b).The Pzerotreatment was characterized by lower alkyl-C content,which explained 18.62% of the variance in the second arm.Soil microbial community composition under PR+W treatment was associated with lower ketone-C(related abundance<1.0%)compared to the other treatments.Moreover,the MRT analysis facilitated the determination of how much of the variation in the parameters was attributed to various C groups.18:1w7c,an indicator for G−,accounted for 17.5% of the variance in microbial community composition(Table III).

Correlations among C group and enzyme activity were analyzed using RDA(Fig.5c).The first axis explained 79.2%of the varieties in enzyme activities,and the second axis explained 8.5%.Similar to the results for the PLFAs,the relative abundance of soil O-alkyl-C was the most influential factor in all soil C forms,and was positively associated with enzyme activity in the samples obtained from PW treatment.

Fig.4 Relative abundances of eight C components,estimated by calculating chemical shift values of 13C cross polarization/magic angle spinningnuclear magnetic resonance spectra,in the paddy soil during the 7th wheat season under four P fertilization treatments,including P application in wheat season only(PW),in rice season only(PR),and in both rice and wheat seasons(PR+W)and no P application in either season(Pzero).Vertical bars indicate standard deviations of the means(n=3).Bars with the same letter(s)are not significantly different among P treatments at P<0.05.

Soil enzyme activity were divided into two groups by the MRT analysis,which were based on soil O-alkyl-C relative abundance,with lower O-alkyl-C related to P fertilization for rice season(PR and PR+W treatments)(Fig.5d).An additional split was attributed to the relative di-O-alkyl-C contents.MRT explained 82.5% of total variation in soil enzyme activity,of which 36.62% was explained by the first split.PPO accounted for 81.0% of the variance in enzyme activity.70.5% of the variance was explained by the tree,whereas 26.0% of this variance was explained by the first split(Table IV).

DISCUSSION

Soil P pool and C quality responses to P fertilization regimes

Soil P sequence fractionation and13C NMR spectroscopy were employed to illustrate the alterations in soil P pools and C chemical forms under four fertilization regimes in wheat-growing season.After 7 years of rice-wheat cropping rotation,we observed that at the 7th wheat harvest season,P fertilization for single season(PR and PW)could still guarantee the level of soil P pools when compared with PR+W treatment,whereas P reduction practice did not significantly influence soil C structure.Notably,the relative amount of O-alkyl-C under PW treatment was significantly greater than those in other treatments.O-alkyl-C,a major component in carbohydrates,cellulose,and hemicellulose,represents an essential and relatively unstable C pool(Baldocket al.,2004).The results indicated that P fertilization in wheat season only facilitated the provision of labile C.Moreover,among four P regimes,PW treatment had the lowest alkyl-C/O-alkyl-C and aryl-C/O-alkyl-C ratios(P>0.05),which are indicators of chemical decomposition rates in soil organic matter(Shresthaet al.,2015).Contrary to our findings,Xu J Set al.(2017)reported that P omission in a maize-wheat rotation system impaired the accumulation and facilitated decomposition of SOC,with a decrease mostly in O-alkyl-C and an upswing with alkyl-C.Unlike chemical fertilizers,organic fertilizers have a greater impact on soil C structure.For instance,Pisaniet al.(2016)showed that 20 years of input manipulations accelerated the processing of labile SOC and sequestration of recalcitrant C,which could be attributed to the variations in soil PLFAs.In addition,the disparate forms of C could be attributed to different durations of fertilizer application.Long-term organic fertilization resulted in a greater alkyl-C/O-alkyl-C ratio,whereas chemical fertilization reduced this ratio.However,different results were obtained following short-term fertilizer application(Panettieriet al.,2014;Wanget al.,2017).

TABLE IIIVariances of soil microbial phospholipid fatty acid(PLFA)concentration explained by a multivariate regression tree analysis,based on relative abundance(RA)of soil O-alkyl-C

Fig.5 Redundancy analysis(RDA)and multivariate regression tree for soil microbial community composition(a and b)and enzyme activities(c and d)during the 7th wheat season under four P fertilization treatments,including P application in wheat season only(PW),in rice season only(PR),and in both rice and wheat seasons(PR+W)and no P application in either season(Pzero).The red arrows represent C composition.RA is the relative abundance.

TABLE IVVariances of soil enzyme activity explained by a multivariate regression tree analysis based on relative abundance(RA)of soil O-alkyl-C

Interactions among PLFAs,enzyme activity,and P fractionations

Fertilizer application is beneficial to microbial biomass,bacteria,fungi,and soil enzyme activity.The results are in conjunction with the observations of other studies(Liet al.,2015;Zhuet al.,2017).In our present study,soil microbial community composition and enzyme activities were the highest under PW treatment.Significant differences in F/B ratio were found between PW and PR+W treatments,which suggested greater susceptibility of soil fungal communities to variations in P pools compared with bacterial communities.Previous studies have revealed that P fertilization could increase soil microbial community composition largely through the alteration of F/B ratio,which could result in enhanced polyphenol oxidase activity(Liet al.,2015).Polyphenol oxidase activity could stimulate the decomposition of lignin and supply more cellulose substrates for cellulase(Liuetal.,2012;Liet al.,2015;Luoet al.,2019).Moreover,PW treatment elevated the abundance of G+compared with PR+W and other P reduction treatments(Pzeroand PR).Resin-P is a form of P with the highest correlation with soil microbial community composition(Fig.4).With regard to soil enzyme activity,lack of P application in the rice season(Pzeroand PW)enhanced oxidoreductase activity and resulted in higher(P<0.05)GMor value.In addition,phosphate fertilizer application increased hydrolase activity compared with no application.Conversely,Wanget al.(2017)reported that long-term use of chemical fertilizers relatively reduced hydrolase activity and enhanced oxidoreductase activity,however,long-term addition of organic fertilizer resulted in an opposite trend.

Fertilizer application mainly augmented the levels of soil LP and STP.Among all the PLFAs tested,AMF was positively and highly associated with LP and STP(Fig.3c),indicating that AMF play a critical role in crop P uptake(Richardsonet al.,2011;Wanget al.,2016a).Particularly in P-deficient soils,diverse AMF develop mutualistic symbioses with crop roots,which could facilitate the transfer of immobile P resources to plant rhizosphere(Mäderet al.,2000;Richardsonet al.,2011;Liet al.,2015).Higher levels of bacteria and fungi also play an important role in liberating orthophosphate by releasing organic anions and phosphatases,thereby influencing the forms and distributions of soil P(Wakelinet al.,2004;Taurianet al.,2010).This may contribute to the similar levels of LP in PW treatment and the treatments with P fertilization in rice season(PR and PR+W).

More than 90% of changes in soil P can be correlated with enzyme activity,particularly soil hydrolase enzyme activity(Fig.3d).Soil enzymes have a greater influence on STP than LP and MLP,owing to its mineralization decomposition(Meenaet al.,2018).Moreover,the synchronous enhancement of enzyme activities,fungi,G+,and G−under PW treatment also indicated an increase in microbial biomass P,which represented potential available P(Jinget al.,2017).As mentioned previously,variations in PLFAs and enzyme activity are predominantly related to soil C forms,especially for O-alkyl-C.It can facilitate soil P availability by mediating the structure,composition,and function of microorganism,as the method that stimulates the proliferation and turnover of soil microorganism will be beneficial to cycling and utilization of soil P(Richardsonet al.,2011).

Response of PLFAs and enzyme activity to C forms

The patterns in soil PLFAs and enzyme activity had a large influence by soil C forms identified through RDA and MRT analyses.The result highlights the significance of linking the nature of soil C with species and functions of soil microbial community.Inclusion of PLFAs,enzymes,and C forms revealed that soil O-alkyl-C was positively and robustly linked to RDA ordination(Fig.5a,c).The di-fferences in C composition could be explained by shifts in enzyme activities more than by microbial community composition,since high proportions of the relationships were accounted for by enzymes.Based on the relative abundance of O-alkyl-C,the MRT analysis demonstrated that the microbial community composition in PW treatment was directly separated from those in the other treatments(Fig.5b),which was also highly correlated with bacterial PLFA markers such as 16:1w7c,18:1w7c,i15:0,and a15:0,and fungal PLFA marker 18:2w6c(Table III).Wanget al.(2017)also observed an increase with organic fertilization for 11 years to chemical fertilization in bacterial PLFA markers such as i14:0,18:1w7c,and i15:0 and actinomycetous PLFA maker 10 Me16:0.Coinciding with the findings of Bthet al.(1992),the number of actinomycetous and bacterial PLFA markers such as 10 Me18:0 and i14:0 was higher in alkaline soil than other microbial communities.For analysis of soil enzyme activities,O-alkyl-C combined with di-O-alkyl-C distinguished the four P fertilization regimes(Fig.5d)and polyphenol oxidase activity was the most dominant factor influencing the variation(Table IV).Polyphenol oxidase activity has the capacity to degrade more recalcitrant aromatic compounds,which could minimize the toxicity of phenolic molecules(Sinsabaugh,2010).Fungi and bacteria are key producers of polyphenol oxidase(Kirk and Farrell,1987).Therefore,the results exhibited a synchronous enhancement of polyphenol oxidase,fungi,G+,and G−under PW treatment.Besides,the result of significantly higher O-alkyl-C under PW treatment than PR treatment may contribute to the differences in abundances of fungi and bacteria and PPO activity between the two treatments.The significance of interactions and/or linkages in soil C chemistry and soil microbial community composition and functions was corroborated by the differences in the RDA and MRT patterns.

Bacteria G+and G−have unique functions in soil C accumulation and decomposition.A higher G+/G−ratio facilitates not only the formation of recalcitrant organic C but also the generation of more assimilation products,which favors fungal and actinobacterial activities(Ponder and Tadros,2002;Hillet al.,2008;Zhanget al.,2013).For our study,a greater G+/G−ratio for PW than PR is consistent with a higher proportion of soil labile C pools(especially for O-alkyl-C)rather than recalcitrant C forms as reported by Wanget al.(2017).Although bacteria decompose soil labile organic C,abundant organic C sources facilitate the growth of bacteria.In addition to soil PLFA dynamics,soil enzyme patterns have been elucidated,which could reflect not only the availability of soil nutrients but also their diversity and microbial activity(Sinsabaughet al.,2008).In this study,the relatively low alkyl-C/O-alkyl-C and aryl-C/O-alkyl-C ratios under PW treatment enhanced soil oxidoreductase activity,including its consumption of soil labile C.This behavior conflicted with the results reported by Wanget al.(2017),who found that during a 11-year period of chemical fertilizer application,the alkyl-C/O-alkyl-C and aryl-C/Oalkyl-C ratios remained the same,indicating no significant variations in soil enzyme activity.

CONCLUSIONS

The annual addition of mineral phosphate fertilizer for 7 years significantly enhanced the concentrations of LP and STP,whereas there was no significant effect on MLP.The RDA analysis further revealed that both enzyme activity and microbial community composition interacted with soil P fractions,particularly enzyme activity.We also observed that the four P fertilization regimes had no significant impact on soil C structure,based on the13C NMR results.Nevertheless,PW treatment resulted in a relatively high O-alkyl-C content.In addition,soil C composition influenced enzyme activity and microbial community composition,and the relative abundance of soil O-alkyl-C was the most influential factor in all soil C forms.Hence,our results put forward a new viewpoint for understanding different soil P pools and C forms combined with variations of enzyme activity and microbial community composition in dry-farming paddy soils under P-input reduction.Further research can be focused on the effect of exogenous labile C material on soil P availability in the agricultural system.

ACKNOWLEDGEMENT

This work was funded by the National Natural Science Foundation of China(No.41671304)and the National Key Research and Development Program of China(No.2017YFD0200206).


登录APP查看全文