Soil microbial respiration is regulated by stoichiometric imbalances:Evidence from a humidity gradient case

Jiwei LI,Jiangbo XIE,Jianzhao WU,Yongxing CUI,Lingbo DONG,Yulin LIU,Xuying HAI,Yan LI,Zhouping SHANGGUAN,2,Kaibo WANG,Changhui PENG and Lei DENG,2,

1State Key Laboratory of Soil Erosion and Dryland Farming on the Loess Plateau,College of Soil and Water Conservation Science and Engineering(Institute of Soil and Water Conservation),Northwest A&F University,Yangling 712100(China)

2Institute of Soil and Water Conservation,Chinese Academy of Sciences and Ministry of Water Resources,Yangling 712100(China)

3State Key Laboratory of Subtropical Silviculture,Zhejiang A&F University,Lin’an 311300(China)

4Sino-French Institute for Earth System Science,College of Urban and Environmental Sciences,Peking University,Beijing 100871(China)

5University of Chinese Academy of Sciences,Beijing 100049(China)

6State Key Laboratory of Loess and Quaternary Geology,Institute of Earth Environment,Chinese Academy of Sciences,Xi’an 710061(China)

7Center of CEF/ESCER,Department of Biological Science,University of Quebec at Montreal,Montreal H3C 3P8(Canada)

ABSTRACT Humidity not only affects soil microbial respiration(SMR)directly,but,indirectly by regulating the availability of soil water and nutrients.However,the patterns of direct and indirect effects of humidity on SMR over large precipitation gradients remain unclear,limiting our understanding of the effects of precipitation changes on soil C cycle.Here,we investigated the relationships among humidity,soil nutrients,and SMR by identifying stoichiometric imbalances,microbial elemental homeostasis,and microbial C use efficiency along a precipitation gradient at a continental scale.The relationship between SMR and humidity index(HI)corresponded to a Richard’s curve with an inflection point threshold value of approximately 0.7.Soil microbial respiration increased with increasing humidity in drier areas(HI<0.7),but tended to balance above this threshold.Increasing humidity exacerbated C:P and N:P imbalances across the selected gradient.Severe N and P limitations in soil microbial communities were observed in drier areas,while soil microbes suffered from aggravated P limitation as the humidity increased in wetter areas(HI>0.7).Soil microbial communities regulated their enzyme production to maintain a strong stoichiometric homeostasis in drier areas;enzyme production,microbial biomass,and threshold elemental ratios were non-homeostatic under P limitation in wetter areas,which further contributed to the increase in SMR.Our results identified a moisture constraint on SMR in drier areas and highlighted the importance of nutrient(especially for P)limitations induced by humidity in regulating SMR in wetter areas.Understanding the modulation of SMR via soil enzyme activity may improve the prediction of soil C budget under future global climate change.

Key Words: carbon use efficiency,ecological stoichiometry,microorganisms,nutrient limitations,precipitation,soil enzyme activities,stoichiometric homeostasis

INTRODUCTION

Soil microbial respiration(SMR)modulates the release of carbon (C) from terrestrial ecosystems into the atmosphere at a rate of 35—70 Pg C year-1(Solomonet al.,2009; Yaoet al., 2020).Thus, alterations in SMR can have a considerable impact on the variation in the atmospheric CO2concentration(Denget al.,2020).Further,by determining soil moisture content,humidity can influence the stoichiometry of C,nitrogen(N),and phosphorous(P),thereby affecting SMR rates(Liuet al.,2005;Heimann and Reichstein,2008;Nielsen and Ball,2015).Stoichiometric imbalances(SIs)are defined as the mismatch between resources and microbial decomposers(Mooshammeret al.,2014a;Tianet al.,2015).Because SIs can limit microbial metabolic activity(Mooshammeret al.,2014b)and eventually modify C allocation for microbial growth and exacerbate carbon dioxide (CO2) release into the atmosphere (Chenet al.,2019;Shaoet al.,2022),investigating the pathways through which soil microbial communities cope with SIs and regulate SMR by increasing humidity is necessary to understand the rates of global C emissions.

The physiological responses of soil microbes to SIs at different humidity levels vary biogeographically(Mooshammeret al., 2014a).For example, in arid and semi-arid climates,water may be the key factor limiting soil microbial activity (Denget al., 2021), rather than nutrient stoichiometry(Heimann and Reichstein,2008);however,resource nutrient availability directly determines the occurrence and magnitude of a SI(Mooshammeret al.,2014a).Soil microbial communities cope with SIs by redistributing nutrients to maintain growth(Yuanet al.,2019).Meanwhile,in temperate and subtropical forests,high humidity levels can accelerate the formation of soil organic matter(SOM)(Butenschoenet al.,2011;Nielsen and Ball,2015).In turn,the abundance of C in SOM means that microbial metabolism is limited by the availability of N and P,resulting in severe SIs(Mooshammeret al., 2014b; Liet al., 2019).In addition, excessive moisture may restrict oxygen (O2) diffusion into the soil,suppressing microbial activity and CO2production(Bradfordet al.,2019).In such cases,soil microbes may release redundant elements and then“go to waste”viarespiration to adapt to SIs(Mooshammeret al.,2014a).However,the mechanism whereby SMR responds to humidity-induced SIs remains unclear.

Soil microbial communities undergo various physiological adaptations to fine-tune the use they make of resources under conditions of a SI(Manzoniet al.,2012;Faninet al.,2013; Mooshammeret al., 2014a), consequently influencing C storage and release(Heimann and Reichstein,2008;Debnathet al.,2020).One such adaptation strategy is the regulation of microbial biomass according to the specific available C:N:P ratio by altering microbial abundance and community composition or storing elements in excess(Sterner and Elser 2002; Faninet al., 2013; Tapia-Torreset al.,2015).Therefore,the microbial community persists in a nonhomeostatic state(Yuanet al.,2019).According to resource allocation theory, microbes may increase the synthesis of extracellular enzymes to preferentially metabolize the limited resources (Faninet al., 2016; Peng and Wang, 2016;Cuiet al.,2018)and improve mobilization of the substrates comprising limited nutrients(Yuanet al.,2019).Microbes can also more efficiently retain the elements required for growth and metabolism, while releasing excess elements(Sinsabaughet al.,2013;Chenet al.,2019);therefore,they can regulate their biomass to match available resources by adjusting their use (Sinsabaughet al., 2016; Yuanet al.,2019).The threshold element ratio(TER)is defined as the limit of biological tolerance to the C:nutrient element ratio(Mooshammeret al., 2014b; Cuiet al., 2020).If the resource ratio exceeds this threshold value,the microorganism is limited by the nutrient element; otherwise, it is limited by C(Mooshammeret al.,2014b;Yuanet al.,2019).The TER integrates strategies for adjusting microbial biomass,extracellular enzymes, and resource use efficiency to the substrate C:N:P level(Scottet al.,2012;Sinsabaughet al.,2016).Soil microbial communities are particularly efficient in acquiring limited nutrients by altering their abundance and community composition(Mooshammeret al.,2014a;Liet al.,2019;Yuanet al.,2019).Therefore,understanding the physiological adaptation of soil microbial communities to humidity-induced SIs can help elucidate the potential consequences of humidity on soil C flux.

Biogeographic variations in soil- and climate-driven vegetation types provide a unique opportunity to investigate the pathways related to microbial adaptations to SIs.In this study, we aimed to determine i) how changes in nutrient stoichiometry affect soil microbial communities along a gradient of increasing humidity and ii)how soil microbial communities adjust their physiology as a conglomerate to cope with SIs that affect SMR along the transect.

MATERIALS AND METHODS

Study sites

This study was conducted in 36different zonal vegetation ecosystems along a 3 000-km northwest-southeast transect in China(30?11′—42?19′N,93?04′—119?17′E),ranging from arid to humid.Humidity index(HI,defined as the ratio of annual precipitation to potential evapotranspiration),mean annual temperature(MAT),and mean annual precipitation(MAP)are 0.07—1.7,2.5—15.5?C,and 41—1 554 mm,respectively.Climatic data,including MAT,MAP,and HI,were collected from the China Meteorological Background Dataset(Xu and Zhang, 2017)using ArcGIS version 10.2(ESRI,Redlands,USA).The study sites are located in Zhejiang,Anhui,Henan,Shaanxi,Gansu,and Xinjiang Provinces(Fig.1),where the vegetation type exhibits a transition from subtropical forests to temperate grasslands to deserts(xeromorphic shrubs) in China.In most sampling sites, anthropogenic disturbances were totally absent(e.g.,fertilizer application,grazing,tillage,or fire)under natural conditions.

Fig.1 Geographic distribution of the 36sampling sites along the 3 000-km humidity gradient transect across Zhejiang(ZJ),Anhui(AH),Henan(HN),Shaanxi(SX),Gansu(GS),and Xinjiang(XJ)provinces of China.

Soil sampling

Soil sampling was conducted from late July to August 2018,i.e.,the peak period of plant growth.At each site,we established six plots for forests(20 m×20 m)or six plots each for temperate grasslands and xeromorphic shrublands(5 m×5 m)representing the dominant plant species.For each plot,20 core samples were collected from the 0—10 cm topsoil layer, after removing plant litter.Composite soil samples were transported to the laboratory in a portable cooler box.One subsample was used to determine soil physicochemical properties after air-drying at 25?C, and the other subsample was used to estimate microbial biomass content and soil extracellular enzyme activity at 4?C,over two weeks.

Soil physical and chemical analyses

Soil pH,bulk density(BD),organic carbon(SOC),total nitrogen(TN),dissolved organic carbon(DOC),available N(SAN),available phosphorus(SAP),and total phosphorus(TP)were determined using the standard methods described by Cuiet al.(2020).

Soil microbial analyses

Soil microbial biomass C and N contents were analyzed using the chloroform fumigation extraction method,and microbial biomass P was measured using the fumigation extraction method (Brookeset al.1982).The activities of C-acquiring enzymes,β-1,4-glucosidase(BG)andβ-D-cellobiosidase(CBH),N-acquiring enzymes,L-leucine aminopeptidase(LAP)andβ-1,4-N-acetyl glucosaminidase(NAG),and P-acquiring enzyme,phosphatase(AP),were assayed as described previously (Denget al., 2019).The activities of soil extracellular enzymes are expressed in units ofμmol h-1g-1soil.Soil ecoenzymatic stoichiometries for C:N,C:P,and N:P are expressed as ln(BG+CBH):ln(NAG+LAP),ln(BG+CBH):lnAP,and ln(NAG+LAP):lnAP,respectively(Faninet al.,2016).

We placed 100 g of fresh soil from each site in a 500-mL glass culture flask covered with plastic parafilm to maintain gas exchange and moisture resistance.The samples were adjusted to 60% field capacity and incubated at 25?C.After a 14-d pre-incubation period,internal SMR was relatively stable (Wanget al., 2014).Following the pre-incubation,we used CO2-free compressed air to remove residual CO2and immediately collected the gas as a blank.After airtight incubation for 4 h,the gas in the flask was collected using a gas-tight syringe and injected into a gas chromatograph(Agilent 7890 A,Agilent,USA)to determine the cumulative amount of CO2produced.

Data analysis

Stoichiometric(C:P,C:N,or N:P)imbalances between resources and microbes were determined as soil DOC:SAP(DOC:SAN or SAN:SAP)ratio over microbial biomass C:P(C:N or N:P) ratio along the selected humidity gradient(Mooshammeret al.,2014b).

To evaluate the response of stoichiometric homeostasis(H′)of soil microbial communities,we used the following equation:

wheremrepresents the slope of the natural logarithm(ln)regression between resource C:N, C:P, or N:P (based on DOC:SAN,DOC:SAP,or SAN:SAP ratios in this study)and microbial biomass C:N,C:P,or N:P.Soil microbes exhibit strong stoichiometric homeostasis when the regression slope is not significant(P>0.05)orH′exceeds 1 and weak or no homeostasis whenH′is approximately 1(Cuiet al.,2018;Sterner and Elser,2002).

We calculated TERs to determine the elements(i.e.,Cvs.N or P) limiting metabolic activity in response to SIs at different levels of humidity (Sinsabaughet al., 2013;Mooshammeret al., 2014b).The TERs for C:N and C:P(TERC:Nand TERC:P,respectively)were calculated using the following equations(Sinsabaughet al.,2013;Yuanet al.,2019):

wheren0andp0are the dimensionless normalization constants for N and P,respectively,and are related to the intercepts in the regression analyses for BG+CBHvs.NAG+LAP and BG+CBHvs.AP,respectively,BC:NandBC:Pare the ratios of microbial biomass C:N and C:P,respectively,andEC:NandEC:Pare the ratios of soil ecoenzymatic C:N and C:P, respectively.In this study, TERC:Nand TERC:Pwere applied to express the microbial resource(i.e.,Cvs.N or P)limitation,comparing to resource ratios(DOC:SAN or DOC:SAP).Soil microbes are limited by N or P if the DOC:SAN or DOC:SAP ratio is greater than the value calculated for TER(Sinsabaughet al.,2016;Cuiet al.,2020);otherwise,C is the limiting factor(Yuanet al.,2019).

Microbial C use efficiency(CUE)was calculated using the following equations according to a stoichiometric model(Sinsabaughet al.,2013).

whereSC:NorSC:Preflects the degree to which soil enzymes modulate the formation of microbial biomass according to the available resources(Sinsabaughet al.,2016),KC:NandKC:Pare the half-saturation constants(0.5),LC:NandLC:Prepresent the resource C:N and C:P ratios,respectively,and CUEmaxrepresents the upper limit of microbial growth efficiency,which was set at 0.60,according to Sinsabaughet al.(2016).

According to the stoichiometric model,microbial N use efficiency(NUE)and P use efficiency(PUE)were calculated using the following equations (Sinsabaughet al., 2013,2016):

Statistical analysis

We determined that the relationship between HI and SMR corresponded well with the Richard’s growth functiony=y0×{1+d×exp[-k×(x-xi)]}-1/d, wherey0represents the saturation value of biological growth,dis the curve shape parameter,kis the growth rate constant,andxiis the inflection point of HI(approximately 0.7).Therefore,an HI of 0.7 was used as the threshold for all regression analyses.One-way analysis of variance (ANOVA) was used to test the significance of changes in SMR and soil characteristics according to vegetation type.Stoichiometric homeostasis between the biomass C:N:P of soil microbial communities and their resources was calculated using linear regression analysis.A two-tailed Pearson rank-order correlation test was used to examine the relationships among SMR,microbial biomass C:N:P,ecoenzymatic C:N:P,and TERs.The partial least squares path model(PLS-PM)was applied to clarify both indirect and direct impacts of humidity on SIs,microbial biomass C:N:P,ecoenzymatic C:N:P,and TERs using the R software.

RESULTS

Changes in SMR

The relationship between SMR and HI was well described by a Richard’s curve along the humidity gradient(Fig.2a).For our study region, the inflection point of the curve along the humidity gradient had a HI value of approximately 0.7 (Fig.2a), above which no further increase in SMR was observed.Soil microbial respiration increased with increasing humidity in drier areas(HI<0.7).Meanwhile,in wetter areas(HI>0.7),SMR increased slightly,although it remained mostly unchanged.The results showed a significant 2.1-fold increase in SMR in moving from deserts to forests(Fig.2b).

Variation in SIs and soil microbial features

Humidity did not influence the C:N imbalance between soil microbes and their resources across the entire transect(Fig.3a), but forest and grassland exhibited significantly higher C:N imbalances than deserts(Fig.S1, see Supplementary Material for Fig.S1).The N:P and C:P imbalances increased significantly with increasing HI;that is,the C:P and N:P imbalances between soil microbes and resources increased while moving from deserts to grasslands to forests(Fig.3b,c).Among the three vegetation types,forests showed the highest imbalance values of N:P(1.97)and C:P(1.38)(Fig.S1).

Soil microbial biomass C:N:P varied between wetter and drier areas.The microbial biomass C:N remained relatively constant in drier areas(Fig.3d),but increased with increasing humidity in wetter areas.In turn, the microbial biomass C:P and N:P significantly decreased by 75.3%and 19.3%,respectively,as humidity increased(Fig.3e,f)in drier areas.In contrast,the same ratios increased significantly by 524.2%and 343.6%,respectively,with increasing humidity in wetter areas(Fig.3e,f).Moreover,forests showed higher values of microbial biomass C:N than grasslands and deserts,and grasslands registered the lowest microbial biomass C:P and N:P(Fig.S1).No significant correlations(P>0.05)between microbial biomass and soil nutrient stoichiometries were found in drier areas(Fig.4),indicating strong stoichiometric homeostasis at the community level.Conversely,there were strong positive relationships between soil and microbial biomass C:P(R2=0.90,P<0.001)and N:P(R2=0.81,P<0.001)stoichiometries in wetter areas,demonstrating non-homeostasis(Fig.4).

Fig.2 Patterns of soil microbial respiration(SMR)across the selected humidity gradient(a)and SMR data and boxplot in vegetation types(b).In a,the bars indicate the standard errors of means(n=6)for each site,and the vertical dashed line indicates a humidity index(HI)value of approximately 0.7.In b,different lowercase letters indicate significant differences among the different vegetation types(P <0.05).

Fig.3 Patterns of stoichiometric imbalances between soil microbes and their resources(a—c)as well as soil microbial biomass C:N:P(d—f)across vegetation types along the selected humidity gradient.The vertical dashed line indicates a humidity index(HI)value of approximately 0.7.Vertical bars are the standard errors of means for each site(n=6).C:NMicrobe,C:PMicrobe,and N:PMicrobe are the microbial biomass C:N,C:P,and N:P,respectively.

Soil ecoenzymatic stoichiometry and TERs

Soil ecoenzymatic C:N,C:P,and N:P showed different trends with increasing humidity (Fig.5a—c).In drier areas,the ecoenzymatic C:N decreased significantly(Fig.5a),whereas the ecoenzymatic C:P and N:P increased with increasing humidity(Fig.5b,c).In contrast,in wetter areas,the ecoenzymatic C:N increased by 12.0%(Fig.5a),whereas ecoenzymatic C:P and N:P decreased significantly with increasing humidity(Fig.5b,c).Grasslands registered lower ecoenzymatic C:N and higher C:P and N:P than forests and deserts, while forests registered lower ecoenzymatic C:P and N:P than the other ecosystem types(Fig.S2, see Supplementary Material for Fig.S2).We did not observe any correlation between TERC:Nand increasing humidity(Fig.5d).Although the TERC:Pratio remained relatively steady as humidity increased in drier areas,it increased with increasing humidity in wetter areas(Fig.5e).Furthermore,TERC:Pincreased significantly with increasing humidity in moving from deserts to forests(Fig.S2).

Potential pathways for regulating SMR

The partial least squares path model(PLS-PM)demonstrated that humidity had a direct positive effect on SIs,and SIs showed significant positive effects on soil ecoenzymatic stoichiometry in drier areas(Fig.6).However,SIs did not affect soil microbial biomass stoichiometry or TERs.Interestingly,the SIs induced by humidity mediated SMRviaa change in soil ecoenzymatic stoichiometry and not in soil microbial biomass stoichiometry or TERs in drier areas.The total standardized effects of HI,SIs,soil ecoenzymatic stoichiometry, soil microbial biomass stoichiometry, and TER on SMR were 0.14,0.33,0.35,-0.22,and-0.31,respectively(Fig.6),confirming that SI and soil ecoenzymatic stoichiometry play a pivotal role in determining SMR.

In wetter areas,increasing humidity negatively affected SIs(Fig.6),whereas SIs showed significant positive effects on soil ecoenzymatic stoichiometry.In contrast with drier areas, soil ecoenzymatic stoichiometry had strong negative effects on microbial biomass stoichiometry and TERs.Moreover, microbial biomass stoichiometry had a strong positive effect on TERs and a significant negative effect on SMR.Intriguingly, soil ecoenzymatic and microbial biomass stoichiometries had positive and negative effects on SMR,respectively.Thus,humidity-induced SIs may regulate microbial physiology by altering soil ecoenzymatic stoichiometry,thereby affecting SMR.

Fig.4 Soil microbial community homeostasis that is reflected by the stoichiometric relationships between soil microbial biomass and resources across vegetation types in drier and wetter areas along the selected humidity gradient.Bars are the standard errors of means for each site(n =6).C:NMicrobe,C:PMicrobe,and N:PMicrobe are the soil microbial biomass C:N,C:P,and N:P,respectively,and C:NResource,C:PResource,and N:PResource are the soil C:N,C:P,and N:P,respectively,based on soil dissolved organic C,available N,and available P.

Fig.5 Variations of C-,N-and P-acquiring enzymatic stoichiometry(a—c)and threshold elemental ratios(TERs)for C:N and C:P(TERC:N and TERC:P,respectively;d and e)of soil microbes across vegetation types along the selected humidity gradient.The vertical dashed line indicates a humidity index(HI)value of approximately 0.7.Vertical bars are the standard errors of means for each site(n=6).BG=β-1,4-glucosidase activity;CBH=β-D-cellobiosidase activity;NAG=N-acetyl-β-D-glucosaminidase activity;LAP=leucine aminopeptidase activity;AP=phosphatase activity;Ecoenzymatic C:N=ln(BG+CBH):ln(NAG+LAP);Ecoenzymatic C:P=ln(BG+CBH):lnAP;Ecoenzymatic N:P=ln(NAG+LAP):ln(NAG+LAP).

DISCUSSION

Soil microbial-nutrient limitation and SIs

Fig.6 A partial least squares path model(PLS-PM)used for evaluating the effects of humidity-induced stoichiometric imbalances(SI)on the direct and indirect regulatory pathways of soil microbial respiration(SMR)in drier and wetter areas along the selected humidity gradient,as well as the standardized effects of related drivers derived from the PLS-PM.The red and green solid lines indicate significant positive and negative coefficients(P <0.05),and the black solid and dashed lines indicate non-significant relationships with positive and negative coefficients,respectively.HI=humidity index;EES=ecoenzymatic stoichiometry;MBS=microbial biomass stoichiometry;TER=threshold elemental ratio.The asterisks*,**,and***indicate significant different at P <0.05,P <0.01 and P <0.001,respectively.

Several studies have shown that high microbial biomass C:N and C:P ratios may increase microbial nutrient demand(Takritiet al., 2018; Liet al., 2019).We found N:P and C:P imbalances between soil microbes and their resources increased with increasing humidity in moving from deserts to grasslands and to forests (Fig.3b, c); however, we detected no relationship between C:N imbalance and humidity(Fig.3a),which suggests that the C:N imbalance remained relatively constant despite changes in soil moisture(Fig.S3,see Supplementary Material for Fig.S3).The strong C:P and N:P imbalances in the microbial community indicated that soil P availability and cycling were more affected by humidity than N across the entire transect.This may be attributed to the fact that rock weathering is the major pathway for producing fresh P(Liet al.,2014;Waringet al.,2014;Denget al.,2019),whereas C and N are derived from plant litter and N deposition as the vegetation develops from deserts to grasslands to forests(Liet al.,2019).

Low humidity restricts both plant growth and C input into soil (Mooshammeret al., 2014b; Yuanet al., 2019).Further,in drier areas,microbial C limitation decreased with increasing humidity(Fig.S4,see Supplementary Material for Fig.S4).Indeed,soil microbial metabolism was severely suppressed by limited nutrient and water availability in areas with lower soil moisture (Fig.S3).In contrast, plants in wetter areas provide sufficient litter to sustain soil microbial growth(Tapia-Torreset al.,2015;Zechmeister-Boltensternet al.,2015),which is consistent with the increases in DOC and SOC observed in this study(Figs.S5 and S6,see Supplementary Material for Figs.S5 and S6).Additionally,we observed a widespread nutrient limitation in drier areas across the transect(Fig.S4),which is consistent with soil P content being the major determinant of microbial structure and function in drier regions(Fenget al.,2019).Thus,in wetter areas, soil microbial communities exhibited an increasing P limitation(Fig.S4),as evidenced by the increasing microbial biomass C:P and N:P ratios with humidity(Fig.3e,f)and the reduced ecoenzymatic C:P and N:P ratios(Fig.5b,c).This was mainly due to increased P leaching in soils experiencing higher levels of precipitation(Cuiet al.,2018).Another reason may be the decline in TP and SAP contents in soils(Figs.S5—S7,see Supplementary Material for Fig.S7).Hence,soil microbes tend to mineralize more phosphates to meet their growth requirements by secreting more extracellular enzymes upon P depletion(Waringet al.,2014;Cuiet al.,2020).

Potential adaptation strategies of soil microbial communities toSIs

Soil microbes adapt to different mechanisms underlying SIs under different climatic conditions.We observed strong stoichiometric homeostasis of soil microorganisms in drier areas,which was supported by the non-significant relationships between soil microbial biomass C:N:P and available resource stoichiometry(Fig.4).This suggests that soil microbes are inherently constrained to SI despite severe C and nutrient limitations(Mooshammeret al.,2014b;Peng and Wang 2016).Soil microbes tend to increase the synthesis of soil extracellular enzymes for limited nutrient elements to overcome a C or N/P limitation,which allows microbial communities to invest energy in acclimating to arid and oligotrophic habitats(Waringet al.,2014;Tapia-Torreset al.,2015).Across the drier areas of the transect,soil microbes were restricted by C and/or N/P limitation,consistent with the findings in arid-oligotrophic areas(Tapia-Torreset al.,2015).Despite water deficit and mineral element limitation in arid and semi-arid regions, soil microbes can maintain strict homeostasis to process low N and P resources(Sterner and Elser,2002;Scottet al.,2012)by optimizing the allocation of C and N/P in enzyme synthesis(Scottet al.,2012;Zechmeister-Boltensternet al., 2015; Yuanet al., 2019).Furthermore,homeostatic regulation to adapt to SIs is closely associated with the modulation of C or N/P use efficiency(Waringet al., 2014; Liet al., 2023).In this study, soil microbial CUE showed no significant variation in drier areas(Fig.S8, see Supplementary Material for Fig.S8), which is inconsistent with previous observations of exponentially enhanced CUE with increasing aridity(Fenget al.,2019;Liet al.,2022).This inconsistency is likely related to soil microbes prioritizing their energy investment in acquiring limited resources rather than supporting growth (Tapia-Torreset al.,2015;Liet al.,2019).Therefore,compensatory responses of soil microbial investment towards extracellular enzymes allow microbes to maintain community homeostasis in the face of SIs caused by increasing humidity.

Soil microbial communities showed non-homeostasis of C:P and N:P in wetter areas, in contrast to the strict stoichiometric homeostasis observed in drier areas(Fig.4).The regulation of microbial biomass according to variations in substrate stoichiometry can reduce the severity of SIs in soil microbes (Scottet al., 2012; Yuanet al., 2019).This non-homeostatic physiological adjustment is likely an adaptive strategy (Scottet al., 2012; Faninet al., 2013;Mooshammeret al.,2014b).In addition to non-homeostatic regulation,soil microbes increase their investment in ecoenzyme expression as P limitation gradually increases(Takritiet al,2018;Liet al.,2019).Along the wetter areas of the transect, soil ecoenzymatic stoichiometry was associated with lower ecoenzymatic N:P and C:P ratios(Fig.5b,c)with increasing humidity,indicating that soil microbes produce more P-acquiring enzymes than other enzymes to address P limitation in wetter areas(Yuanet al.,2019).With regard to microbial element use efficiency,microbial CUE decreased,but microbial PUE increased in wetter areas with increasing humidity, which might also be explained by an increased investment in P-acquiring enzymes (HI> 0.7, Fig.S8).Interestingly,SMR increased slightly and then reached equilibrium(Fig.2a),implying that increasing humidity promotes more plant litter for soil microbial growth;thus,soil microbial communities did not increase their investment in C but rather in other nutrients, either N or P (Kaiseret al.,2014; Yuanet al., 2019; Liet al., 2022).The increased microbial biomass C:N with increasing humidity(Fig.3d)indicates a higher ratio of fungi to bacteria (Zhanget al.,2005;Huanget al.,2015).Fungal adaptation to a SI can be realized indirectly through the production of extracellular enzymes,and fungi may be the predominant source of extracellular enzymes in forests, especially in ectomycorrhizal forests(Yuanet al.,2019).Altogether,our results indicate that soil microbes can adjust their physiological pathways to compensate for SIs that occur as a result of increasing humidity.This adaptive strategy of soil microbes also verifies the applicability of the resource allocation theory(Yuanet al.,2019).

Relative effects of the potential adaptation strategies on SMR

The trend in SMR changed along the humidity gradient,with the most obvious shifts observed at an approximate HI threshold value of 0.7,indicating that microbes rely on different physiological adaptations to balance microbial C assimilation.Soil microbial respiration is tightly controlled by resource quality,climatic factors,and microbial community structure(Bradfordet al.,2019;Yuanet al.,2019).However,few researchers have investigated the changes in SMR induced by soil microbial physiological-adaptive pathways to humidity-induced SIs(Yuanet al.,2019).The results of the PLS-PM analysis showed that humidity had a significant direct influence on SIs in both drier and wetter areas(Fig.6).Similarly, a previous study reported that N addition had significant effects on SMRviathe alteration of SIs(Yuanet al.,2019).However,this study revealed that SIs were more likely to directly affect SMR by altering the ecoenzymatic stoichiometry in both drier and wetter areas(Fig.6).On one hand, in drier areas, the increase in SMR with increasing humidity was directly linked to changes in soil ecoenzymatic stoichiometry rather than to microbial biomass stoichiometry or TERs (Fig.6), and significant correlations (P< 0.01)between SMR and soil ecoenzymatic stoichiometry were also observed(Fig.S9,see Supplementary Material for Fig.S9).Soil microflora is affected by water and nutrient limitations in arid and semi-arid areas (Peng and Wang, 2016).Nevertheless,soil microbes can maintain strict stoichiometric homeostasis to obtain or reassign relatively scarce resources by optimizing the allocation of C, N, and P toward enzymatic synthesis in response to SIs,rather than diverting any resources towards their growth(Mooshammeret al.,2014b),as evidenced by the relatively stable microbial CUE in drier areas(Fig.S8),along with an increase in SMR.The obvious increase in soil ecoenzymatic C:P and N:P ratios(Fig.5a—c)with increasing humidity suggests that the humidity-induced increase in SMR can be explained by the reallocation of nutrients to enzymatic synthesis in drier areas.

On the other hand,in wetter areas,an increase in SMR was linked to alterations in microbial biomass stoichiometry, ecoenzymatic stoichiometry, and TERs (Fig.6).Interestingly, microbial biomass stoichiometry and TERs were greatly influenced by adjustment of SIs in ecoenzymatic stoichiometry(Fig.6).These results indicate that soil microorganisms reduce the energy input committed to C-and N-acquiring enzymes and maintained phosphatase production instead.Extracellular enzymes produced by the microbial community decompose plant residues into their active components,thereby facilitating their absorption and assimilation(Mooshammeret al.,2014b;Zechmeister-Boltensternet al.,2015;Denget al.,2019).We observed high P-invertase activity in wetter areas,which reflected the lack of TP and SAP in wetter areas (Figs.S5 and S6).To meet their P demand, microbes secrete more P-acquiring enzymes to mineralize organic P in the soil(Mooshammeret al.,2014b;Waringet al., 2014).As there was sufficient moisture in the wetter areas (Fig.S3), microbial communities tended to be non-homeostatic in their adaptation of microbial biomass C:N:P to reduce the P limitation induced by SIs across the humidity gradient.This adjustment may be reflected in nutrient-element utilization efficiency(Kaiseret al.,2014;Liet al.,2022),as evidenced by the indirect regulation of SMR through TER(Fig.6).

Variations in SMR affect microbial nutrient-element use efficiency,as microbial CUE may be reduced by C investment in enzyme production(Sinsabaughet al.,2013;Yuanet al.,2019; Cuiet al., 2020).Although we observed a decline in microbial CUE and an increase in microbial PUE,SMR increased slightly,although it remained mostly unchanged in wetter areas(Fig.2a),probably because soil microbes can reduce TERs through continued intracorporeal recycling of limited nutrients(Kaiseret al.,2014;Mooshammeret al.,2014b;Yuanet al.,2019).Hence,microbes may meet their stoichiometric demands with resources rather than by discharging excess C through respiration(Kaiseret al.,2014).Interestingly,TERC:Pdid not decrease in wetter areas with increasing humidity(Fig.5e),which is likely related to the trade-offbetween microbial CUE and PUE(Fig.S8).Under severe P limitation,microbes can increase microbial PUE and reduce microbial CUE to ensure a balance between biomass production and nutrient acquisition,thus adjusting SMR accordingly.In addition,the predominant microbial life history(k-selected orr-selected)induced by SI shifts may favor different growth strategies characterized by various microbial CUEs,thereby affecting SMR(Scottet al.,2012;Mooshammeret al.,2014a;Yuanet al.,2019).Hence,the combined effects of ecoenzymatic stoichiometry,microbial biomass stoichiometry,and TERs caused soil microbes to invest more C in their biomass.However, these processes were primarily caused by the response of soil ecoenzymatic stoichiometry to humidity-induced SIs,as further supported by the significant correlations between SMR and soil ecoenzymatic stoichiometry (Fig.S9).Our findings highlight the importance of ecoenzymatic stoichiometry in various pathways related to SMR regulation in wetter areas.

CONCLUSIONS

Microbial responses to SIs occurred predominantly at an HI threshold value of approximately 0.7.The humidity threshold (HI = 0.7) was also a characteristic inflection point in moving from grasslands to forests.A turning point in microbial metabolism was observed near this threshold along the transect.Soil microbes regulate their physiology in response to SIs, which in turn affect SMR.Humidityinduced SIs influenced SMR by altering the soil ecoenzymatic stoichiometry in drier areas(HI<0.7).Meanwhile,in wetter areas(HI>0.7),SIs led to broad changes in ecoenzymatic stoichiometry,microbial biomass stoichiometry,and TERs.

ACKNOWLEDGEMENTS

This study was sponsored by the National Natural Science Foundation of China(Nos.42277471 and 42307578),the Strategic Priority Research Program of the Chinese Academy of Sciences(Nos.XDB40000000 and XDA23070201),the Postdoctoral Research Funds of the Shaanxi Province,China(2023BSHYDZZ76),the Open Grant for State Key Laboratory of Loess and Quaternary Geology, the Institute of Earth Environment,Chinese Academy of Sciences(SKLLOG2230),the Fundamental Research Funds for the Central Universities,China(2023HHZX002),and the Special Support Plan of Young Talents Project of Shaanxi Province and National Forestry and Grassland Administration in China(No.20201326015).

SUPPLEMENTARY INFORMATION

Supplementary material for this article can be found in the online version.