at SciVerse ScienceDirect

Soil Biology & Biochemistry 55 (2012) 85e92

Contents lists available

Soil Biology & Biochemistry

journal homepage: www.elsevier .com/locate/soi lb io

Modeling the effects of temperature and moisture on soil enzyme activity: Linkinglaboratory assays to continuous field data

J. Megan Steinweg a,b,*, Jeffrey S. Dukes c,d, Matthew D. Wallenstein a,b,e

aNatural Resource Ecology Laboratory, Colorado State University, Fort Collins, CO 80523-1499, USAbGraduate Degree Program in Ecology, Colorado State University, Fort Collins, CO 80523-1878, USAcDepartment of Forestry and Natural Resources, Purdue University, West Lafayette, IN 47907, USAdDepartment of Biological Sciences, Purdue University, West Lafayette, IN 47907, USAeDepartment of Ecosystem Science and Sustainability, Colorado State University, Fort Collins, CO 80523, USA

a r t i c l e i n f o

Article history:Received 29 February 2012Received in revised form14 June 2012Accepted 19 June 2012Available online 4 July 2012

Keywords:b-glucosidaseMoisture thresholdTemperature sensitivityIn situ conditionsDroughtEnzyme assay methods

* Corresponding author. Present address: Oak Ridge2008, MS 6038, Oak Ridge, TN 37831-6038, USA. Tel.:576 8646.

E-mail addresses: steinwegjm@ornl.gov (J.M. Ste(J.S. Dukes), matthew.wallenstein@colostate.edu (M.D

0038-0717/$ e see front matter � 2012 Elsevier Ltd.http://dx.doi.org/10.1016/j.soilbio.2012.06.015

a b s t r a c t

Although potential enzyme activity measurements have a long history of use as an indicator of microbialactivity, current methods do not provide accurate estimates of in situ activity. In the field, diffusion ratestypically limit the rate at which enzymes can pair with substrates. However, the common laboratorypractice of creating soil slurries removes all diffusion constraints. In addition, temperature stronglyaffects in situ enzyme activities, but is rarely considered in enzyme assays. To address these limitations,we developed a new protocol to measure the moisture and temperature sensitivity of enzyme activities.We incorporated sensitivity data obtained using this protocol into a model to estimate the effects oftemperature and moisture on in situ b-glucosidase enzyme activity, recognizing that other factors such assubstrate concentrations and diffusion constraints also affect in situ enzyme activities.

Soil samples were collected from the Boston-Area Climate Experiment every two weeks over a 10-week period to track enzyme dynamics as field temperature and moisture changed. Precipitationinputs to an old-field were manipulated to produce drought (50% ambient precipitation), ambient, andwet (150% ambient precipitation) treatments. Temperature sensitivity of b-glucosidase was determinedby assaying for the enzyme in soil slurries at three different temperatures (15, 25 and 35 �C). Moisturesensitivity was determined by exposing soils to different moisture levels in the lab and adding substrateto homogenized dry or moist soils instead of slurries. Temperature sensitivity was calculated as Q10 andmoisture sensitivity was calculated using a linear regression for each field treatment at each samplecollection date.

Moisture sensitivity varied significantly among the five sample dates and treatments, whereastemperature sensitivity remained stable. At almost every time point, b-glucosidase activity respondedmore strongly to increased moisture in soils of drought plots than in soils of ambient and wet plots. Weestimated in situ b-glucosidase activity in the fall using the temperature and moisture sensitivities.Estimates that used only temperature or only moisture sensitivity suggested that ambient plots had thehighest activity, followed by wet and then drought plots. Estimates based on both temperature andmoisture suggested that b-glucosidase activity responded primarily to changes in temperature, exceptwhen soils were dry, with water potentials below �1 MPa. These results demonstrate that low soilmoisture can strongly limit in situ enzyme activity in soils, negating any positive effect of warming. Thisstudy provides a template for parsing out the role of specific abiotic drivers on in situ enzyme activities,which could lead to the explicit incorporation of enzymes in biogeochemical models, improving upon theability of current models to predict rates of biogeochemical processes in dynamic environments.

� 2012 Elsevier Ltd. All rights reserved.

National Laboratory, PO Boxþ1 865 241 3776; fax: þ1 865

inweg), jsdukes@purdue.edu. Wallenstein).

All rights reserved.

1. Introduction

Current soil organic matter (SOM)models (e.g. DAYCENT, RothC)are able to reproduce large-scale changes in carbon dynamicsunder most conditions (Coleman and Jenkinson, 1999; Parton et al.,1998, 1987; Schimel et al., 1997) without explicitly modeling

J.M. Steinweg et al. / Soil Biology & Biochemistry 55 (2012) 85e9286

microbial community dynamics (Andren and Balandreau, 1999).However, this is often not the case in highly variable environments,which may require more mechanistic models that include enzymecatalysis and substrate utilization efficiency parameters (Allisonet al., 2010; Lawrence et al., 2009; Li et al., 2010). These parame-ters may be important to include in models because microbes oftenrespond to changing environmental conditions through alteredresource allocation and changes in community composition, whichcan affect enzyme kinetics and the aggregate physiology ofmicrobial communities (Collins et al., 2008; Fierer et al., 2003;Stone et al., 2012). Enzymatic depolymerization of biopolymers isthe rate-limiting step in decomposition of unprotected SOM(Bengtson and Bengtsson, 2007; Conant et al., 2011; Schimel andBennett, 2004). A few studies have explicitly incorporatedenzymes into models and have demonstrated the potential toimprove predictions of litter decomposition rates and microbialresponses in dynamic systems when enzyme pools and kinetics areexplicitly modeled (Davidson et al., 2012; Lawrence et al., 2009;Moorhead and Sinsabaugh, 2000; Schimel and Weintraub, 2003).However, these models have also exposed a fundamental gap in ourknowledge of abiotic controls on the rates of enzyme activitiesunder field conditions, and a disconnect between what is beingmeasured in laboratory assays and model needs (Wallenstein andWeintraub, 2008).

Enzyme activity has been assayed in soils for over sixty years(Skujins, 1976) and used as a descriptor of soil quality, an indicatorof substrate use and nutrient cycling, and to provide a mechanisticunderstanding of decomposition in natural and disturbed systems(Bandick and Dick, 1999; Dilly and Nannipieri, 1998; Nannipieri,1994; Sinsabaugh et al., 2009). There have been significantchanges in the methodology used to assess activity, but currentmethods still only provide estimates of potential enzyme activityand do not measure parameters that drive in situ activity (Drobník,1961; Skujins, 1976; Wallenstein and Weintraub, 2008). Twocharacteristics of current enzyme assays that hamper our ability tobetter estimate in situ activity are the lack of diffusion limitationsand the use of a single assay temperature (German et al., 2011b;Wallenstein and Weintraub, 2008).

Almost all contemporary enzyme assays are performed in slurry(Dick, 2011; Saiya-Cork et al., 2002) to ensure adequate homogeni-zation of added substrates and consistent estimation of activity ratesover time.However,mostupland-soils are rarelywater saturated, andeven when saturated, substrates are not well mixed. The slurrydecreases the substrate diffusion limitation observed in most soilsresulting in higher activities than would be expected in the field(Geisseler and Horwath, 2009; Geisseler et al., 2011). In addition,substrates are not homogenously distributed throughout soil (EttemaandWardle, 2002) as in the slurry, again leading to anoverestimationof activity. Diffusion constraints on substrates, enzymes, or both canhave a large impact on in situ activity, especially under droughtconditions where diffusion is limited (Koch, 1990).

In early soil enzyme protocols, assay temperatures were veryhigh (e.g. 40 �C) to optimize activity, but were outside thetemperature range experienced in most soils (Skujins, 1976). Morerecently, many studies have measured activity at temperaturescloser to those experienced by soils in the field. However, enzymeactivity is typically measured only at one temperature. Enzymaticreactions, like all other chemical reactions, are sensitive totemperature (Trasar-Cepeda et al., 2007; Wallenstein et al., 2011,2009). The temperature sensitivity of enzymes has previously beenfound to vary seasonally in several ecosystems (Brzostek and Finzi,2012; Wallenstein et al., 2009). Field soil temperatures can changedrastically over the course of a day or year, and the use of one assaytemperature does not provide enough information on thetemperature sensitivity of soil enzyme activities to predict in situ

rates (McClaugherty and Linkins, 1990). Temperature sensitivity ofenzyme reactions can be tempered by diffusion limitations in thefield, resulting in a lack of enzyme temperature dependence(Brzostek and Finzi, 2012; Davidson and Janssens, 2006). Thustemperature and moisture need to be considered together whentrying to understand field activity.

We evaluated the independent and combined effects oftemperature and moisture on b-glucosidase activity at the Boston-Area Climate Experiment (BACE). Utilizing data generated from twodifferent enzyme assay methods in combination with field soilmoisture and temperature data, we modeled the joint effects oftemperature andmoisture on b-glucosidase activity, to predict howthese factors affect in situ activity. b-glucosidase was chosen as themodel enzyme because it is involved in the final step of cellulosebreakdown and is produced by a variety of microorganisms.Enzyme assayswere performed at three different temperatures andmultiple soil moistures to assess changes in temperature andmoisture sensitivity of b-glucosidase activity in soils collected everytwo weeks during Fall 2009 from the BACE soil moisture manipu-lation plots. We hypothesized that soil moisture was the dominantcontrol on enzyme activity, and that the activity of existingenzymes would increase as diffusion limitations were alleviated.

2. Methods

2.1. Study site

Soils were obtained from the BACE, which is located in an old-field ecosystem in Waltham, Massachusetts, at the University ofMassachusetts’ Suburban Experiment Station (42� 230 300N, 71� 120

5200 W). Mean annual temperature and precipitation in the area is9.5 �C and 1194 mm yr�1 (Hoeppner and Dukes, 2012). Soils wereclassified as mesic Typic Dystrudepts; the upper 30 cm consists ofloam soils (45% sand, 46% silt, and 9% clay), with a pH of 5.5. The sitewas previously an apple orchard, and has harbored old-fieldvegetation for over 40 years. Current vegetation includes 42species of grasses and forbs, which are primarily non-native(Hoeppner and Dukes, 2012).

2.2. Precipitation manipulation

We used soil from nine plots in the BACE; all plots experiencedambient temperatures, but plots were divided among three precip-itation treatments, with three replicates per treatment. Precipitationtreatments were designated as ambient, “wet” (150% of ambientprecipitation during the growing season), and “drought” (50% ofambient precipitation during the growing season). Precipitationwascontrolled by clear partial roofs in the drought plots, with half of theincoming precipitation diverted to wet plots during the growingseason (MayeOctober). During the colder months of the year,NovembereApril, drought plots were maintained, but additionalwater was not added to the wet plots. Drought treatments began inJanuary 2007 and wet treatments in June 2008. Thus, by the fall of2009, soils had been exposed to drought treatments for 2.6 years andto wet treatments for 1.1 years. During the non-freezingmonths, soilmoisture was measured weekly using time-domain reflectometry(waveguides were installed across 0e10 and 0e30 cm depths).Dataloggers recorded soil temperature near the center of each plotevery 30 min throughout the year, as measured by linear tempera-ture sensors positioned at 2 and 10 cm below the surface.

2.3. Soil sampling and processing

Soils were collected every twoweeks fromAugust 2009 throughOctober 2009 (Fig. 1). Three cores (5 cm diameter) were collected

230 240 250 260 270 280 290 300

0

5

10

15

20

25

30

in

s

itu

a

ve

ra

ge

s

oil te

mp

era

tu

re

(

o

C)

day of year, 2009

230 240 250 260 270 280 290 300

0

5

10

15

20

25

30

35

a

b

in

s

itu

%

v

olu

me

tric

s

oil m

ois

tu

re

drought

ambient

wet

Fig. 1. In situ (a) volumetric soil moisture and (b) daily average temperature fromAugust 2009 through October 2009 for precipitation manipulation plots. Lines indicatesoil sample collection dates.

J.M. Steinweg et al. / Soil Biology & Biochemistry 55 (2012) 85e92 87

from each plot at 0e5 cm depths. Cores were packaged on ice andshipped to the laboratory overnight, where the soils from each plotwere sieved (2 mm), separated from rocks and roots, homogenized,and frozen at �10 �C until analysis.

Soil moisture was measured on all soil samples by weighing 5 gof field-moist soil, drying the soils for 24 h at 60 �C, and thenreweighing. Soil water potential was measured on soil samplesusing a Decagon WP-4 potentiometer. Water potential wasmeasured on field-moist and air-dried samples. In addition, 0, 100,200 or 300 mL of water was added to dried samples and 0, 100, or200 mL was added to the moist samples to measure the range ofwater potentials that were used in the enzyme moisture sensitivityassay.

2.4. Enzyme assays

2.4.1. Temperature sensitivity assaysEnzyme assays were performed on samples from all plots at

each collection date, using two different methods to assesstemperature and moisture sensitivity. Temperature sensitivity ofsamples was assayed using a protocol modified from Saiya-Corket al. (2002). The temperature sensitivity of b-glucosidase wasassessed by performing assays in deep-well 96-well plates at threetemperatures: 15, 25 and 35 �C. Each column of 8 wells on eachplate corresponded to one soil sample. One additional plate was

used to create a standard curve for each sample at 25 �C. Thereference standard for b-glucosidase activity was 4-methylumbelliferone (MUB). The standard curve plates hada column for each soil slurry sample and a different concentrationof the MUB standard in each of the slurry wells (0, 2.5, 5, 10, 25, 50and 100 mM).

A 2.75 g subsample from each soil core was warmed to roomtemperature (w20 �C) and homogenized with 50 mM sodiumacetate buffer (pH 5.5) for 1 min in a Waring laboratory gradeblender on high to make a soil slurry. Each column on the 96-welldeep-well plates corresponded to one sample. After homogeniza-tion, 800 mL of soil slurrywas aliquoted into each of the eightwells ofone column on all four plates, one plate for each of the three assaytemperatures and one for the standard curve at 25 �C. Following theaddition of twelve samples into their respective columns, 200 mL of200 mM MUB-b-D glucopyranoside substrate was added.

The plates were incubated for different lengths of time depend-ing on incubation temperature, 1.5 h at 35 �C, 3 h at 25 �C, and 6 h at15 �C. Three temperatures were used to estimate two Q10 values incase there was a significant difference between Q10s at 15e25 �C and25e35 �C. Following incubation, the plates were centrifuged for3min at 350 g. Afterwards, 250 mL of supernatant from eachwell wasplaced into the corresponding well on a 96-well black plate. A TecanInfinite M500 spectrofluorometer was used to measure fluorescencewith wavelengths set at 365 nm and 450 nm for excitation andemission, respectively. The plates with the standards were used tocalculate a linear standard curve and determine b-glucosidaseactivity for each sample as nmol g�1 dry soil h�1.

2.4.2. Moisture sensitivity assaysThemoisture sensitivity assays were performed using soils from

the field precipitation treatment plots with additional soil moisturemanipulations in the laboratory. Two subsamples, each w20 gwhen field-moist, were taken from each frozen soil sample. Onesubsample was kept at field moisture (hereafter called moist)overnight, while the other was allowed to dry overnight (hereaftercalled dried) at room temperature, both about 20 �C. The followingday 2 g of soil from each group, moist and dried, were weighed intoseven different scintillation vials. Next, between 0 and 300 mL DIwater was added to alter the soil moisture in “dried” samples and0e200 mL DI water was added to alter the soil moisture in “moist”samples. Water was added in small drops and then the sample wasstirred with a small spatula for 5e10 s to distribute water moreevenly throughout the sample. Immediately following wateraddition, 250 mL of 591 mM MUB-b-D glucopyranoside substratewas added to samples at 20 �C and stirred for 5 s. A greaterconcentration of substrate was added than in the temperaturesensitivity assays because the lack of diffusion limited enzymeactivity resulting in very rapid loss of activity in preliminaryexperiments. Lower concentrations of substrate made it difficultto measure the maximum rate of enzyme activity on themicroplate reader. Eight minutes after substrate addition, 31 mLof 50 mM sodium acetate (pH 5.5) was added and the samplewas vortexed for about 5 s. After vortexing, 800 mL of slurry fromeach sample was added to three wells in a deep-well plate andcentrifuged for 3 min at 350 g. Following centrifugation, 250 mL ofeach sample was transferred to a black 96-well plate and fluores-cence was measured as previously described. Standard curves werealso created using MUB standard at 0, 5, 25, 50 and 100 mM.Preliminary trials indicated that soils dried down and then rewet-ted to their original fieldmoisture content produced activity similarto field moist samples that had not been dried. In addition,preliminary assays indicated that 8 min provided an adequateincubation period to assess enzyme activity before substrate limi-tation occurred in the moisture sensitivity assays.

230 240 250 260 270 280 290 300

2.0

2.5

3.0

3.5

4.0

4.5

5.0

te

mp

era

tu

re

s

en

sitiv

ity

(Q

)

day of year, 2009

drought

ambient

wet

Fig. 2. Temperature sensitivity of b-glucosidase activity during Fall 2009 calculatedfrom lab activity at 15 �C and 25 �C enzyme assays. Points are the average Q10 collectedfrom sample dates with standard error bars, n ¼ 3.

J.M. Steinweg et al. / Soil Biology & Biochemistry 55 (2012) 85e9288

2.5. Calculations and statistics

The temperature sensitivity of b-glucosidase activity wasassessed on samples incubated in the lab at 15, 25 and 35 �C andcalculated as Q10:

Q10 ¼ activity at T

�10

T2�T1

�

2activity at T1

(1)

where T2 is the incubation temperature 10 �C greater than T1. TheQ10 values were calculated from the 15 and 25 �C assays. Therewerefive dates resulting in 15 Q10 values, one for each of the three fieldprecipitation treatments. For the predictive model, Q10 values wereinterpolated between dates to obtain a Q10 for each day.

Moisture sensitivity for b-glucosidase activity was estimatedusing separate linear regressions of enzyme activity measured in thelab for each date and treatment. PROC GLM (SAS Institute, Cary, NC)was used to determine if there were significant differences amongtreatments and dates in moisture sensitivity slope and intercept.Precipitation treatments significantly affected slope and intercept ateach date, so we used different slopes and intercepts to estimateactivity at each date for each precipitation treatment. Only five dateswere assayed, so the slopes and intercepts were interpolated for thedays between sample dates for each field precipitation treatment forthe model. In addition, we interpolated between weekly field soilmoisture measurements to obtain daily soil moisture values fromAugust 24, 2009 to October 22, 2009. PROC CORR was used to esti-mate the correlation coefficient and significance of the correlationbetween enzyme activity and soil moisture.

The effect of temperature on predicted in situ b-glucosidaseactivity was estimated using the Q10 values estimated from labtemperature incubations in conjunctionwith field soil temperature.We estimated temperature-based in situ activity using thefollowing equation from Wallenstein et al. (2009):

in situ activity from field temperature ¼ R25*Q

�t�2510

�

10 (2)

where R25 is the b-glucosidase activity at 25 �C, Q10 is derived fromthe 15 and 25 �C assays, and t is the in situ temperature. Q10 did notsignificantly vary by treatment (PROC GLM, SAS Institute, Cary, NC),but instead of averaging over treatments we chose to use theestimated Q10 from each treatment.

We also modeled the effect of soil moisture on in situ enzymeactivity using a linear regression:

in situ activity from field moisture ¼ msmxsm þ bsm (3)

where m and b are the slope and intercept, respectively, deter-mined from enzyme assays conducted at a range of soil moisturesand x is the field soil moisture. A linear regression model was usedbecause the response of enzyme activity to moisture in the lab waslinear across the range of soil moistures we used.

To estimate the effects of both temperature and moisture on insitu b-glucosidase activity we combined the Equations (2) and (3).Equation (2) was modified to estimate activity using both fieldvariables, by replacing R25 in Equation (2) with y ¼ mx þ b fromEquation (3), as shown in Equation (4):

in situ activity from field temperature and moisture

¼ ðmSM*xSMþbSMÞ*Q�t�2010

�

10 (4)

where the variables are the same as above. Multiple comparisonsANOVAs using PROC GLM were made for model predictions todetermine if average predicted activity from each precipitationtreatmentover all dateswas significantlydifferentbasedon themodelused, i.e. the temperature, moisture, and temperature � moisturemodels.

3. Results

Precipitation treatments altered soil moisture substantially,with soil moisture in drought treatments being on average 50% ofambient moisture at 0e10 cm in 2009 (Fig. 1a). There was no effectof additional water on the soil moisture of wet plots compared toambient. Volumetric soil moisture increased from the end ofAugust 2009 to the end of October 2009, while soil temperaturedeclined by about 17 �C during the same period (Fig. 1b). Droughtplots tended to be slightly warmer than ambient and wet plots butthe difference was minimal; about 0.25 �C higher on average.

Temperature sensitivity of b-glucosidase activity did not differsignificantly among precipitation treatments or sample dates. Thetemperature sensitivity of b-glucosidase activity (Q10) in all plots onaverage was about 3 throughout the sampling period (Fig. 2).

Laboratory manipulation of soil moisture resulted in a widerange of soil water potentials, from �4 MPa in the dried samplesto �0.10 MPa in field-moist soils amended with 200 mL water(Fig. 3). At every sampling period, except September 10 and October22, 2009, b-glucosidase activity in the drought treatment was moresensitive to soil moisture than activity in ambient or wet plots, withslope values 2e5 times higher (Table 1; Fig. 4). Drought plotsdemonstrated a significant positive correlation between soilmoisture and enzyme activity for every sample date exceptSeptember 10, 2009. Enzyme activity in ambient soils was nega-tively correlated with soil moisture on the first two sample dates,but the correlation became positive for the two October sampledates. In wet plots, enzyme activity was positively correlated withsoil moisture except on September 25, 2009. Therewere differences(P < 0.05) in slope and intercept values between drought and othertreatments at each date except October 22, 2009, when ambientand drought plots had similar responses.

Temperature-based in situ b-glucosidase activity estimatesdeclined as field temperatures dropped from August to October(Fig. 5a). Drought plots had the lowest estimated activity despite

driest dried+200uL field moist wettest

-4

-2

0

so

il w

ate

r p

ote

ntia

l (M

Pa

)

lab moisture manipulation

drought field treatment

ambient field treatment

wet field treatment

Fig. 3. Water potential in soils from drought, ambient and wet treated plots onSeptember 10, 2009, DOY 253 following laboratory moisture manipulations via dryingand water additions, n ¼ 1. The term “driest” on the x-axis refers to the air driedsubsample, with no water or substrate solution addition. Moving from left to right, thenext three points are the air dried sample with additions of 100, 200 and 300 mL water,respectively. The “field moist” point is the thawed soil sample with no moisturemanipulation. The two points following the “field moist” label have 100 and 200 mLwater added, with the 200 mL water addition termed “wettest”.

J.M. Steinweg et al. / Soil Biology & Biochemistry 55 (2012) 85e92 89

having, on average, slightly higher plot temperatures because oflow potential activity measured in the lab. Moisture-based in situactivity estimates suggested increasing enzyme activity as soilmoisture increased over the course of the season (Fig. 5b). Therewas a dry down in all plots around day 250, and estimated enzymeactivity immediately began to decline in drought and ambienttreatments. Soil moisture increased immediately following the drydown event, but enzyme activity did not increase in ambient- ordrought-treated plots until after day 270. In situ activity estimatesin the wet plots increased over the course of the sample period andwas not as sensitive to changes in soil moisture.

For ambient and wet precipitation treatments, estimatedactivity averaged over the whole season was highest in themoisture-based model, followed by the combined temperature-moisture model and then the temperature-based model (Table 2).In drought treatments the moisture-based and temperature-basedmodels had similar average estimated activity over the whole

Table 1Slope and intercept values determined from linear regressions of moisture sensi-tivity assays for each field treatment at each date. Bold lettering indicates significantregression, P < 0.05.

Date Field treatment Slope Intercept Pearsons’s R

24-Aug-09 Drought 15.43 L91.41 0.73Ambient �2.571 263.07 �0.22Wet 3.193 41.32 0.27

10-Sep-09 Drought 3.472 119.6 0.09Ambient �2.069 209.3 �0.17Wet 6.702 19.53 0.48

25-Sep-09 Drought 10.37 4.406 0.55Ambient 2.488 76.50 0.25Wet �0.2223 181.9 �0.01

6-Oct-09 Drought 14.32 L8.926 0.52Ambient 5.998 173.1 0.57Wet 0.3521 234.8 0.04

22-Oct-09 Drought 5.988 53.68 0.46Ambient 6.386 37.60 0.75Wet 2.410 107.94 0.35

season, but had very different trends in activity over the course ofthe season (Fig. 5c). Using both field soil moisture and temperatureto estimate in situ activity resulted in average estimated activitiessignificantly different from the temperature-based and moisture-based models in all precipitation treatments (P < 0.05; Table 2).In the drought plots, the predicted activity trend in the combinedmodel followed the moisture-based model throughout the season.In ambient and wet plots, estimated activity trends from thecombinedmodel were similar to those from the temperature-basedmodel throughout the sampling period.

4. Discussion

Given the critical role of extracellular enzymes in decomposi-tion and nutrient cycling, our limited ability to predict enzymeactivities is an important constraint in understanding feedbacksbetween climate change and biogeochemical cycles. Predictingsoil enzyme activities in dynamic environments is challenging dueto our limited understanding of enzyme production, residencetime and turnover (Burns, 1982; Wallenstein and Weintraub,2008). In this study, we demonstrated the importance of consid-ering the sensitivity of enzyme activity to both temperature andmoisture. The lack of a consistent effect of experimental climatechange on enzyme activity at the BACE (unpublished data) maskedimportant changes in the moisture sensitivity of enzyme activitythat appear to control changes in enzyme activities under fieldconditions.

By measuring b-glucosidase activity in soils at different soilmoistures, we observed a strong positive effect of moisture onactivity in the drought soils and occasionally in the ambient andwet soils. These assays were very short, and therefore detected theactivity of the extant enzyme pool rather than that of newlyproduced enzymes. As lab soil moisture increased, the extant poolof functional enzymes was able to interact more freely withsubstrates as diffusion constraints were alleviated. The mainte-nance of an enzyme pool during drought at the beginning of theFall could have been due to either increased production (Harderand Dijkhuizen, 1983) or reduced enzyme turnover due toenzyme stabilization (George et al., 2007). If substrate limitedenzyme activity, due to either diffusion constrains or lowsubstrate quantity (German et al., 2011a, 2011b), the nutrientrequirements to produce enzymes could have exceeded the netincrease in nutrient availability due to enzyme activities (Allisonand Vitousek, 2005). Increased enzyme production would nothave been favored in this scenario, because nutrient mineraliza-tion would have declined at low moistures and nutrient redistri-butionwould have been diffusion-limited, but a decline in enzymeturnover could have accounted for the relatively large pool ofenzymes (indicated by potential activities) detected under dryconditions.

As soils dried, enzymes may have become adsorbed to clays andtannins, rendering them less susceptible to proteolytic breakdownbut able to maintain reactivity (Ensminger and Gieseking, 1942;Kandeler, 1990). The results of this study indicate that low soilmoisture can limit b-glucosidase activity, particularly when soilwater potential falls below�1MPa, indicated by reduced estimatedenzyme activity in the drought plots following a dry down eventaround September 10, 2009. Our results contrast with those ofLawrence et al. (2009) who suggested that enzyme activity isinsensitive to moisture because activity can continue in water filmsduring drought. The steep increase in activity as diffusion limita-tions were eliminated in the lab indicated that soil moisture in thedrought plots constrained the activity of the available enzyme pool.In ambient and wet plots, field soil moisture stayed above 15% forthe duration of the study and diffusion limitations may not have

drought

ambient

wet

0 10 20 30 40 50 60

0

100

200

300

400

500

a b

dc

e

nm

ol a

ctiv

ity

g

d

ry

s

oil-1

h

-1

0 10 20 30 40 50 60

0

100

200

300

400

500

0 10 20 30 40 50 60

0

100

200

300

400

500

600

700

800

nm

ol a

ctiv

ity

g

d

ry

s

oil-1

h

-1

0 10 20 30 40 50 60

0

100

200

300

400

500

600

700

800

% volumetric soil moisture

0 10 20 30 40 50 60

0

100

200

300

400

500

nm

ol a

ctiv

ity

g

d

ry

s

oil-1

h

-1

% volumetric soil moisture

DOY 236 DOY 253

DOY 268 DOY 279

DOY 295

Fig. 4. Potential b-glucosidase activity and respective significant (P < 0.05) linear regressions for each precipitation treatment at different lab induced soil moistures at each sampledate, (a) Day of year (DOY) 236: August 24, 2009, (b) DOY 253: September 10, 2009, (c) DOY268: September 25, 2009, (d) DOY 279: October 6, 2009 and (e) DOY 295: October 22,2009.

J.M. Steinweg et al. / Soil Biology & Biochemistry 55 (2012) 85e9290

been imposed. As a result, predicted activity in ambient and wetplots did not consistently show a strong positive correlation withincreasing field moisture. Despite the presence of an availableenzyme pool in drought plots, a concurrent decline in soil respi-ration at BACE (Suseela et al., 2012) indicated a moisture limitationcontrol on enzyme activity and/or diffusion of soluble substrates formicrobial assimilation and metabolism.

Enzyme assays conducted at multiple temperatures confirmedthat temperature is a strong driver of enzyme activity rates, as hasbeen previously demonstrated in other systems (German et al., 2012;

Stone et al., 2012; Trasar-Cepedaet al., 2007;Wallenstein et al., 2009).Despite the positive response of enzymes to increasing temperaturein the lab assays, there was no measurable change in the enzymetemperature sensitivity over the course of the sample dates as othershavemeasured (Wallenstein et al., 2009),whichwas unexpected dueto a 17 �Cdecline in soil temperature over the sample period. Changesin temperature sensitivity could have been due to production ofisoenzymes with temperature sensitivities that differ from those ofthe isoenzymes they are replacing within the enzyme pool(Wallenstein and Weintraub, 2008). The replacement of an enzyme

230 240 250 260 270 280 290 300

0

100

200

300

400

b

c

ae

stim

ate

d a

ctiv

ity

b

y in

s

itu

s

oil te

mp

era

tu

re

(n

mo

l a

ctiv

ity

g

-1

d

ry

s

oil h

-1

)

drought

ambient

wet

230 240 250 260 270 280 290 300

0

100

200

300

400

es

tim

ate

d a

ctiv

ity

b

y in

s

itu

s

oil m

ois

tu

re

(n

mo

l a

ctiv

ity

g

-1

d

ry

s

oil h

-1

)

230 240 250 260 270 280 290 300

0

100

200

300

400

es

tim

ate

d in

s

itu

a

ctiv

ity

(n

mo

l a

ctiv

ity

g

-1

d

ry

s

oil h

-1

)

day of year, 2009

Fig. 5. Modeled effects of (a) field soil temperature and lab measured temperaturesensitivity, (b) field soil moisture and moisture sensitivity, (c) combined temperatureand moisture sensitivity on b-glucosidase activity.

Table 2Average estimated in situ b-glucosidase activity during Fall 2009 by field treatmentfor each model. Letters indicate whether estimated in situ activity estimate averagesare significantly different between models for each precipitation treatment,P < 0.05.

Moisture Temperature Moisture � temperature

Drought 97.72 � 4.66a 96.59 � 3.77a 84.10 � 3.18b

Ambient 211.1 � 7.46a 129.2 � 7.33b 173.7 � 3.41c

Wet 167.6 � 5.67a 123.0 � 3.58b 142.0 � 2.79c

J.M. Steinweg et al. / Soil Biology & Biochemistry 55 (2012) 85e92 91

pool with enzymes that had different temperature optima may havetaken longer than the 60-day period of observation.

The temperature-based model estimated similar enzymeactivity trends to the temperature � moisture model in ambientand wet plots, indicating that temperature was an importantcontroller of activity when moisture was not limiting. The esti-mated activity from the temperature model in the ambient plotswas lower than the moisture model because temperature wasconstant at 20 �C in the moisture model, whereas in the temper-ature model it was often below 20 �C. Also, despite the lack ofa significant difference in temperature sensitivity across moisturetreatments, the ambient plots had lower temperature sensitivitiesfor three of five dates resulting in lower estimated activity. Astrong effect of moisture on estimated activity in the drought plotswas apparent in the first half of fall, when drought plots wereextremely dry. The lack of moisture overwhelmed any enzymetemperature response in the temperature � moisture model, sothat the moisture � temperature model resulted in lower esti-mated activity than the temperature-based or moisture-basedmodels individually. Later in the season when soil water poten-tial increased, temperature did influence estimated activity in thedrought plots, but was not as dominant a control as it was onambient and wet plot enzyme activity.

Predicted enzyme activities averaged over the whole season forthe moisture- and temperature-based models individually werequite different in the ambient and wet plots. The moisture-basedmodel estimated higher activity than the temperature-basedmodel, likely due to the higher substrate concentration in theassay and the assumption of a constant 20 �C temperature in themoisture-based model. In general, the methodology used toascertain moisture and temperature sensitivities likely over-estimated in situ enzyme activity because of high substrateconcentrations compared to in situ conditions. Despite these limi-tations, the combined models provided a more comprehensive andrealistic view of soil enzyme dynamics, because moisture andtemperature can both strongly affect these dynamics.

5. Conclusion

The three models developed here were the first attempts toestimate enzyme activity using field soil moisture and tempera-ture data. Current enzyme methodology provides information onpotential activity, but relationships between enzymes andecosystem processes have remained poorly understood due to theconstraints of laboratory methodology (German et al., 2011b). Asclimate change increases the variability of precipitation and thefrequency of droughts, diffusion constraints on enzyme activitiesmay emerge as an important control on soil C cycling in moreecosystems. Temperature sensitivity of enzyme activity hasreceived increasing attention as of late (McClaugherty and Linkins,1990; Trasar-Cepeda et al., 2007; Wallenstein et al., 2009), butthere is still a lack of understanding regarding moisture sensitivity(Wallenstein and Weintraub, 2008). Our study demonstrates thatboth moisture and temperature can strongly control enzymeactivity; we found temperature was the dominant control whensoil moisture was not limiting. Since both of these variablesinfluence in situ enzyme activities, and warming can often inducesoil drying in the field, it is important to understand the interac-tive effects of these factors. Our results suggest that soil dryingcould mitigate the positive effect of warming on enzyme activity.This research provides new insights into climate controls on in situenzyme activity, adding to an increasingly sophisticated under-standing of how microbial responses to climate will affect soil Cand nutrient cycling (Wallenstein and Hall, 2012; Wallensteinet al., 2012).

J.M. Steinweg et al. / Soil Biology & Biochemistry 55 (2012) 85e9292

Acknowledgments

We thank Carol Goranson for sampling and shipping soils andmaintaining the BACE. This research was supported by grants toMDW and JSD from the U.S. Department of Energy’s Office ofScience (BER), through the Northeastern Regional Center of theNational Institute for Climate Change Research, and from NSF (DEBto JSD).

References

Allison, S.D., Vitousek, P.M., 2005. Responses of extracellular enzymes to simple andcomplex nutrient inputs. Soil Biology & Biochemistry 37, 937e944.

Allison, S.D., Wallenstein, M.D., Bradford, M.A., 2010. Soil-carbon response towarming dependent on microbial physiology. Nature Geoscience 3, 336e340.

Andren, O., Balandreau, J., 1999. Biodiversity and soil functioning e from black boxto can of worms? Applied Soil Ecology 13, 105e108.

Bandick, A.K., Dick, R.P., 1999. Field management effects on soil enzyme activities.Soil Biology & Biochemistry 31, 1471e1479.

Bengtson, P., Bengtsson, G., 2007. Rapid turnover of DOC in temperate forestsaccounts for increased CO2 production at elevated temperatures. EcologyLetters 10, 783e790.

Brzostek, E.R., Finzi, A.C., 2012. Seasonal variation in the temperature sensitivity ofproteolytic enzyme activity in temperate forest soils. Journal of GeophysicalResearch 117, G01018.

Burns, R.G., 1982. Enzyme-activity in soil e location and a possible role in microbialecology. Soil Biology & Biochemistry 14, 423e427.

Coleman, K., Jenkinson, D., 1999. ROTHC-26.3. A Model for the Turnover of Carbonin Soil. IACR-Rothamsted, Harpenden.

Collins, S.L., Sinsabaugh, R.L., Crenshaw, C., Green, L., Porras-Alfaro, A., Stursova, M.,Zeglin, L.H., 2008. Pulse dynamics and microbial processes in aridland ecosys-tems. Journal of Ecology 96, 413e420.

Conant, R.T., Ryan, M.G., Ågren, G.I., Birge, H.E., Davidson, E.A., Eliasson, P.E.,Evans, S.E., Frey, S.D., Giardina, C.P., Hopkins, F.M., Hyvönen, R.,Kirschbaum, M.U.F., Lavallee, J.M., Leifeld, J., Parton, W.J., Steinweg, J.M.,Wallenstein, M.D., Wetterstedt, M.J.Å., Bradford, M.A., 2011. Temperature andsoil organic matter decomposition rates e synthesis of current knowledge anda way forward. Global Change Biology 17, 3392e3404.

Davidson, E.A., Janssens, I.A., 2006. Temperature sensitivity of soil carbon decom-position and feedbacks to climate change. Nature 440, 165e173.

Davidson, E.A., Samanta, S., Caramori, S.S., Savage, K., 2012. The Dual Arrhenius andMichaeliseMenten kinetics model for decomposition of soil organic matter athourly to seasonal time scales. Global Change Biology 18, 371e384.

Dick, W.A., 2011. Development of a soil enzyme reaction assay. In: Dick, R.P. (Ed.),Methods of Soil Enzymology. Soil Science Society of America, Madison, WI,pp. 71e84.

Dilly, O., Nannipieri, P., 1998. Intracellular and extracellular enzyme activity in soilwith reference to elemental cycling. Zeitschrift Fur Pflanzenernahrung UndBodenkunde 161, 243e248.

Drobník, J., 1961. On the rôle of toluene in the measurement of the activity of soilenzymes. Plant and Soil 14, 94e95.

Ensminger, L.E., Gieseking, J.E., 1942. Resistance of clay-adsorbed proteins toproteolytic hydrolysis. Soil Science 53, 205e209.

Ettema, C.H., Wardle, D.A., 2002. Spatial soil ecology. Trends in Ecology & Evolution17, 177e183.

Fierer, N., Schimel, J.P., Holden, P.A., 2003. Influence of drying-rewetting frequencyon soil bacterial community structure. Microbial Ecology 45, 63e71.

Geisseler, D., Horwath, W.R., 2009. Relationship between carbon and nitrogenavailability and extracellular enzyme activities in soil. Pedobiologia 53, 87e98.

Geisseler, D., Horwath,W.R., Scow, K.M., 2011. Soilmoisture and plant residue additioninteract in their effect on extracellular enzyme activity. Pedobiologia 54, 71e78.

George, T.S., Simpson, R.J., Gregory, P.J., Richardson, A.E., 2007. Differential interactionof Aspergillus niger and Peniophora lycii phytases with soil particles affects thehydrolysis of inositol phosphates. Soil Biology & Biochemistry 39, 793e803.

German, D.P., Chacon, S.S., Allison, S.D., 2011a. Substrate concentration and enzymeallocation can affect rates of microbial decomposition. Ecology 92, 1471e1480.

German, D.P., Weintraub, M.N., Grandy, A.S., Lauber, C.L., Rinkes, Z.L., Allison, S.D.,2011b. Optimization of hydrolytic and oxidative enzyme methods for ecosystemstudies. Soil Biology & Biochemistry 43, 1387e1397.

German, D.P., Marcelo, K.R.B., Stone, M.M., Allison, S.D., 2012. The Michae-liseMenten kinetics of soil extracellular enzymes in response to temperature:a cross-latitudinal study. Global Change Biology 18, 1468e1479.

Harder, W., Dijkhuizen, L., 1983. Physiological responses to nutrient limitation.Annual Review of Microbiology 37, 1e23.

Hoeppner, S.S., Dukes, J.S., 2012. Interactive responses of old-field plant growth andcomposition to warming and precipitation. Global Change Biology.

Kandeler, E., 1990. Characterization of free and adsorbed phosphatases in soils.Biology and Fertility of Soils 9, 199e202.

Koch, A.L., 1990. Diffusion- the crucial process in many aspects of the biology ofbacteria. Advances in Microbial Ecology 11, 37e70.

Lawrence, C.R., Neff, J.C., Schimel, J.P., 2009. Does adding microbial mechanisms ofdecomposition improve soil organic matter models? A comparison of fourmodels using data from a pulsed rewetting experiment. Soil Biology &Biochemistry 41, 1923e1934.

Li, X., Miller, A.E., Meixner, T., Schimel, J.P., Melack, J.M., Sickman, J.O., 2010. Addingan empirical factor to better represent the rewetting pulse mechanism in a soilbiogeochemical model. Geoderma 159, 440e451.

McClaugherty, C.A., Linkins, A.E., 1990. Temperature responses of enzymes in 2forest soils. Soil Biology & Biochemistry 22, 29e33.

Moorhead, D.L., Sinsabaugh, R.L., 2000. Simulated patterns of litter decay predictpatterns of extracellular enzyme activities. Applied Soil Ecology 14, 71e79.

Nannipieri, P., 1994. The potential use of soil enzymes as indicator of productivity,sustainability, and pollution. In: Pankhurst, C.E., Doube, B.M., Gupta, V.V.S.R.,Grace, P.R. (Eds.), Soil Biota: Management in Sustainable Farming Systems.CSIRO, Adelaide, pp. 238e244.

Parton, W.J., Schimel, D.S., Cole, C.V., Ojima, D.S., 1987. Analysis of factors controllingsoil organic-matter levels in Great-Plains Grasslands. Soil Science Society ofAmerica Journal 51, 1173e1179.

Parton, W.J., Hartman, M., Ojima, D., Schimel, D., 1998. DAYCENT and its land surfacesubmodel: description and testing. Global and Planetary Change 19, 35e48.

Saiya-Cork, K.R., Sinsabaugh, R.L., Zak, D.R., 2002. The effects of long term nitrogendeposition on extracellular enzyme activity in an Acer saccharum forest soil. SoilBiology & Biochemistry 34, 1309e1315.

Schimel, J.P., Bennett, J., 2004. Nitrogen mineralization: challenges of a changingparadigm. Ecology 85, 591e602.

Schimel, J.P., Weintraub, M.N., 2003. The implications of exoenzyme activity onmicrobial carbon and nitrogen limitation in soil: a theoretical model. SoilBiology & Biochemistry 35, 549e563.

Schimel, D.S., Emanuel, W., Rizzo, B., Smith, T., Woodward, F.I., Fisher, H.,Kittel, T.G.F., McKeown, R., Painter, T., Rosenbloom, N., Ojima, D.S., Parton, W.J.,Kicklighter, D.W., McGuire, A.D., Melillo, J.M., Pan, Y., Haxeltine, A., Prentice, C.,Sitch, S., Hibbard, K., Nemani, R., Pierce, L., Running, S., Borchers, J., Chaney, J.,Neilson, R., Braswell, B.H., 1997. Continental scale variability in ecosystemprocesses: models, data, and the role of disturbance. Ecological Monographs 67,251e271.

Sinsabaugh, R.L., Hill, B.H., Shah, J.J.F., 2009. Ecoenzymatic stoichiometry ofmicrobial organic nutrient acquisition in soil and sediment. Nature 462. 795-U117.

Skujins, J., 1976. Extracellular enzymes in soil. Critical Reviews in Microbiology 4,383e421.

Stone, M.M., Weiss, M.S., Goodale, C.L., Adams, M.B., Fernandez, I.J.,German, D.P., Allison, S.D., 2012. Temperature sensitivity of soil enzymekinetics under N-fertilization in two temperate forests. Global ChangeBiology 18, 1173e1184.

Suseela, V., Conant, R.T., Wallenstein, M.D., Dukes, J.S., 2012. Effects of soil moistureon the temperature sensitivity of heterotrophic respiration vary seasonally inan old-field climate change experiment. Global Change Biology 18, 336e348.

Trasar-Cepeda, C., Gil-Sotres, F., Leiros, M.C., 2007. Thermodynamic parameters ofenzymes in grassland soils from Galicia, NW Spain. Soil Biology & Biochemistry39, 311e319.

Wallenstein, M., Hall, E., 2012. A trait-based framework for predicting when andwhere microbial adaptation to climate change will affect ecosystem func-tioning. Biogeochemistry 109, 35e47.

Wallenstein, M.D., Weintraub, M.N., 2008. Emerging tools for measuring andmodeling the in situ activity of soil extracellular enzymes. Soil Biology &Biochemistry 40, 2098e2106.

Wallenstein, M.D., McMahon, S.K., Schimel, J.P., 2009. Seasonal variation in enzymeactivities and temperature sensitivities in Arctic tundra soils. Global ChangeBiology 15, 1631e1639.

Wallenstein, M.D., Allison, S.D., Ernakovich, J., Steinweg, J.M., Sinsabaugh, R.L., 2011.Controls on the temperature sensitivity of soil enzymes: a key driver of in situenzyme activity rates. In: Shukla, G., Varma, A. (Eds.), Soil Enzymology.Springer-Verlag, Berlin, pp. 245e258.

Wallenstein, M.D., Haddix, M.L., Lee, D.D., Conant, R.T., Paul, E.A., 2012. A litter-slurry technique elucidates the key role of enzyme production and microbialdynamics in temperature sensitivity of organic matter decomposition. SoilBiology and Biochemistry 47, 18e26.