Responses of the extracellular enzyme activities in hardwood forest to soil temperature and

seasonality and the potential effects of climate change

Petr Baldrian*, Jaroslav Snajdr, V era Merhautová, Petra Dobiásová, Tomás Cajthaml, Vendula
Valásková

Keywords: Extracellular enzymes, Forest soil, Lignocellulose, Litter, Microbial ecology, Quercus
petraea, Seasonality, Climate change

Abstract

The activities of extracellular enzymes that participate in the decomposition of litter and organic
matter in forest soils depend on, among other factors, temperature and soil moisture content and
also reflect the quality of litter, which changes dramatically after a short litterfall period. Here, we
explored the effects of soil temperature and seasonality on the sizes of extracellular enzyme pools
and activities in a temperate hardwood forest soil with dominant Quercus petraea (cambisol,
mean annual temperature 9.3°C). We hypothesized that the most significant variation of enzyme
activity would occur in the litter, which faces greater variations in temperature, moisture content
and chemical quality during the season, which decrease with soil depth. The site exhibited
relatively large seasonal temperature differences and moderate changes in soil moisture content.
Enzyme activity, microbial biomass, soil moisture content, temperature and pH were monitored
for three years in the litter (L), organic horizon (O) and upper mineral horizon (Ah). Enzyme activity
in vitro strongly increased with temperature until 20-25 C, the highest temperatures recorded in
situ. While no significant differences in the pools of most extracellular enzymes and in the content
of microbial biomass were found among the seasons, enzyme activity typically increased during
the warm period of the year, especially in the O and Ah horizons. Approximately 63%, 64%, and
69% of total annual activity was recorded during the warm period of the year in the L, O, and Ah
horizons, respectively. Significant positive correlations were observed between soil moisture
content and fungal biomass, but not bacterial biomass, indicating a decrease of the
fungal/bacterial biomass ratio under dry conditions. The effect of moisture on enzyme activities
was not significant except for endoxylanase in the litter. If soil temperature rises as predicted due
to global climate change, enzyme activity is predicted to rise substantially in this ecosystem,
especially in winter, when decomposition is not limited by drought and fresh litter that can
decompose rapidly is present.

1. Introduction

Substantial variations in temperature or moisture during the year influence considerably the soil
processes of biomes with such climatic characteristics, including the decomposition of organic
matter. Thus, climatic factors have previously been identified as major causes of the observed
seasonal differences in decomposition rates in such environments due to alterations in the pools
of various extracellular enzymes, including laccase, polysaccharide hydrolases, phosphatase,
urease, protease and others (Bastida et al., 2008; Criquet et al., 2002; Prietzel, 2001; Wittmann et
al., 2004). The effects of temperature on respiration or on the activity of selected enzymes has
been repeatedly demonstrated (Ise and Moorcroft, 2006; Moore, 1986; Wallenstein et al., 2009).
However, in some environments such as the Mediterranean zone, where periods of high
temperature are accompanied by temporary droughts, the positive effect of temperature in the
warm period of the year is counteracted by the decrease of enzyme pools due to soil or litter
desiccation (Criquet et al., 2000, 2004; Sardans and Peñuelas, 2005). Even in the temperate zone,
soil moisture content was identified as one of the most important factors affecting the spatial
distribution of microbial biomass and extracellular enzymes (Baldrian et al., 2010 b). In temperate
forests, temperature variation during the year can be considerable, whereas the effects of drought
are usually less pronounced than in the warmer zones. However, the seasonality of the
decomposition processes may be seriously affected by the seasonal differences in belowground C
flux via plant roots (Högberg et al., 2010;Kaiser et al., 2010) and the fact that the quality of the
litter material on the soil surface changes abruptly during the litterfall season, typically restricted
to autumn, when fresh litter with a higher content of easily available nutrients and high C/N ratio
accumulates on the forestfloor (Dilly and Munch, 1996; Fioretto et al., 2000; Snajdr et al., 2011).
As a result of these phenomena, spring and summer are characterized by a decrease of litter
quality due to its ongoing decomposition with the concomitant increase of photosynthetic carbon
allocation underground via the mycelia of mycorrhizal fungi. This carbon flow ceases in autumn
along with leaf abscission, when it is replaced by the seasonal litterfall. Consequently, the relative
proportion of decomposer to symbiotic life strategies of fungi is predicted to increase during the
cold period of the year. In addition to changes in microbial community composition, the
production of several hydrolases by symbiotic ectomycorrhizal fungi is also increased (Mosca et
al., 2007). Although this theoretical model seems to reasonably predict the behavior of the
temperate forest soils, the extent of seasonal differences in enzyme activities and the abundance
of their microbial producers have not been addressed sufficiently with the majority of studies
obtained in the boreal and tundra ecosystems. Interestingly, these studies showed that
considerable decomposition rates can also be achieved during the cold period of the year and that
the warmest periods do not necessarily have the highest decomposition rates (Kahkonen et al.,
2001; Wallenstein et al., 2009; Wittmann et al., 2004). Data derived from similar studies make it
possible to predict the direction and the potential extent of changes in decomposition rates if
temperatures increase as a consequence of global climate change. The aim of this work was to
describe the seasonal variations in enzyme pools (i.e., the amount of enzyme molecules) and the
biomass of soil microorganisms in hardwood forest soils with dominant Quercus petraea and to
quantify the seasonal variation of enzyme activities calculated as enzyme pools multiplied by the
relative activity of the enzyme at the in situ temperature recorded. Litter, organic soil horizon and
mineral soil were separately analyzed because they were previously demonstrated to differ
substantially in chemical quality, microbial biomass content and community composition (Snajdr
et al., 2008). The enzyme activity and climatic data were collected monthly for three years. We
hypothesized that the most significant variation of enzyme activity would occur in the litter, which
faces greater variations in temperature, moisture content and chemical quality during the season,
which decrease with soil depth. The results obtained in this study were also used to predict the
potential increase of enzyme activity under the model scenarios of the future climate change
HadAM3H and ECHAM4/OPYC3 (Jacob et al., 2007).
2. Materials and methods

2.1. Study site and sampling

Soil and litter samples were collected in a sessile oak (Q. petraea) forest in the Xaverovský Háj
Natural Reserve near Prague, Czech Republic, a site where previous studies targeted the spatial
variability of extracellular enzyme distribution (Snajdr et al., 2008), the description of
environmental and microbial factors affecting enzyme production (Baldrian et al., 2010a,b; Snajdr
et al., 2011) and the decomposition abilities of saprotrophic fungi (Baldrian et al., 2011; Snajdr et
al., 2010; Valásková et al., 2007). The soil was an acidic cambisol with litter (L), organic horizon (O),
and the mineral horizons Ah and A Litter thickness was 0.5e1.5 cm, with average pH 4.3, 46.2% C,
1.76% N; O horizon thickness was 1.5-2.5 cm, average pH 3.7, 21.5% C, 0.56% N; Ah horizon
thickness was 6-8 cm, average pH 3.4, 3.0-14.3% C, 0.10-0.39% N. For the study of seasonal
variation of soil enzyme activities and microbial biomass content, soil cores (45 mm in diameter)
were collected monthly from September 2005 to August 2008. At each sampling date, a total of six
cores were collected from the same 16 m2 sampling plot with a litter layer on the forestfloor (no
growth of grasses). For each soil core,Lhorizon material (0.5-1.0 cm),O horizon material and Ah
horizon material were separated, and the materials from all cores were combined to yield a
composite sample of each horizon. Samples of the L horizon were cut into approximately 0.25-cm2
pieces, and the samples from the deeper soil horizons were sieved using a 2-mm sieve. The
resulting samples were used for the enzyme assays and the ergosterol and PLFA analyses. For the
analysis of temperature effects on enzyme activity, extracellular enzymes were extracted from
twelve cores sampled in late summer 2007. The composite sample combined theL,O, and Ah
material of all cores. Soil pH was measured in soil water extract (1 g soil and 10 mL deionized
water were mixed and left to stand overnight at room temperature), and the soil moisture content
was assessed by drying the soil at 85°C until a constant mass was reached. The temperature was
recorded hourly during the sampling period at the soil surface and at interfaces between the L and
O, O and Ah, and Ah and A horizons. From these data, temperatures in soil horizons were
calculated as the averages of temperature recorded immediately above and below the respective
horizon.

2.2. Enzyme extraction and assays

Enzymes were extracted from samples on the day of sample collection as previously described
(Snajdr et al., 2008), and at least two independent extractions were performed from each sample.
Homogenized samples of soil or litter material were extracted at 4°C for 2 h on an orbital shaker
(100 rpm) with 100 mM phosphate buffer, pH 7 (16:1 w/v),filtered through Whatman # 5 filter
paper and desalted using PD-10 desalting columns (Pharmacia, Sweden), according to the
supplier’s protocol, to remove inhibitory small-molecular-mass compounds. The desalted samples
were immediately used for enzyme activity analysis. Enzymes for the determination of
temperature-activity relationships were extracted from combined samples of the whole L, O and
Ah horizons with a total mass>100 g. Three independent extractions were performed. Extracts
were concentrated by ultrafiltration through a 10-kDa nitrocellulose membrane (Amicon,
Millipore) before desalting. Laccase (EC 1.10.3.2) activity was measured by monitoring the
oxidation of 2,2 0 -azinobis-3-ethylbenzothiazoline-6-sulfonic acid (ABTS) in citrate-phosphate
buffer (100 mM citrate and 200 mM phosphate; pH 5.0) at 420 nm (Bourbonnais and Paice, 1990).
Manganese peroxidase (MnP, EC 1.11.1.13) activity was assayed using a succinate-lactate buffer
(100 mM, pH 4.5) according to Bourbonnais and Paice (1990). 3-methyl-2-benzothiazolinone
hydrazone (MBTH) and 3,3-dimethylaminobenzoic acid (DMAB) were oxidatively coupled by the
enzymes, and the resulting purple indamine dye was detected spectrophotometrically at 595 nm.
The results were corrected by the activities of the samples without manganese (for MnP)ethe
addition of manganese sulfate was substituted by an equimolar amount of
ethylenediaminetetraacetate (EDTA). One unit of enzyme activity was defined as the amount of
enzyme forming 1mmol of reaction product per min. Endocellulase (EC 3.2.1.4) and endoxylanase
(EC 3.2.1.8) activities were routinely measured with azo-dyed carbohydrate substrates
(carboxymethyl cellulose and birchwood xylan, respectively) using the protocol of the supplier
(Megazyme, Ireland). The reaction mixture contained 0.2 mL 2% dyed substrate in 200 mM sodium
acetate buffer (pH 5.0) and 0.2 mL sample. The reaction mixture was incubated at 40°C for 60 min
and the reaction was stopped by adding 1 mL of ethanol, vortexing for 10 s and centrifuging at
10,000 g for 10 min (Baldrian, 2009). The amount of released dye
was measured at 595 nm, and the enzyme activity was calculated according to standard curves
correlating the dye release with the release of reducing sugars. One unit of enzyme activity was
defined as the amount of enzyme releasing 1 mmol of reducing sugars per min. Cellobiohydrolase
(EC 3.2.1.91) activity was assayed in microplates using p-nitrophenyl-b-D-cellobioside (PNPC). The
reaction mixture contained 0.16 mL 1.2 mM PNPC in 50 mM sodium acetate buffer (pH 5.0) and
0.04 mL sample. Reaction mixtures were incubated at 40°C for 60-120 min. The reaction was
stopped by adding 0.1 mL 0.5 M sodium carbonate, and the absorbance was read at 400 nm. 1,4-
b-glucosidase (EC 3.2.1.21), 1,4-b-xylosidase (EC 3.2.1.37) and 1,4-b-N-acetylglucosaminidase (EC
3.2.1.52) activities were assayed by the same method usingp-nitrophenylb-D-glucoside, p-
nitrophenyl-b-D-xyloside and pnitrophenylN-acetyl-b-D-glucosaminide, respectively.
Phosphomonoesterase (EC 3.1.3.1) was assayed using 2 g
L

1 p-nitrophenylphosphate in 50 mM sodium acetate buffer (pH 5.0), as previously described

(Baldrian, 2009). One unit of enzyme activity was defined as the amount of enzyme releasing
1mmol of p-nitrophenol per min. Spectrophotometric measurements of seasonal samples were
performed in triplicate using a microplate reader (Sunrise, Tecan) or a UV-VIS spectrophotometer
(Lambda 11, PerkineElmer) and expressed per g dry mass of the sample. These activities are
referred to as “enzyme pools” because they reflect the amount of enzymes present in the
samples. For the analysis of temperature effects on enzyme activity, assays were performed at
5°C, 10°C, 15°C, 20°C, 25°C, and 40°C for each enzyme in triplicate in a microplate incubator. To
calculate enzyme activity, the enzyme pool from each sampling was multiplied by the ratio of
enzyme activity of the corresponding enzyme at the actual temperature and at the standard assay
temperature (40°C). Actual temperatures for each sampling date were defined as the mean
temperatures recorded over the period of seven days preceding soil sampling. Enzyme activity was
also calculated per mg total PLFA as a measure of total microbial biomass in each sample, this
value being termed the specific enzyme activity.

2.3. Quantification of microbial biomass

The samples for phospholipid fatty acid (PLFA) analysis were extracted by a mixture of chloroform-
methanol-phosphate buffer (1:2:0.8) according to Bligh and Dyer (1959). Phospholipids were
separated using solid-phase extraction cartridges (LiChrolut Si 60, Merck), and the samples were
subjected to mild alkaline methanolysis as described previously (Snajdr et al., 2008). The free
methyl esters of phospholipid fatty acids were analyzed by gas chromatographyemass
spectrometry (Varian 3400; ITS-40, Finnigan). Fungal biomass was quantified based on 18:2u6,9
content (PLFAF), and bacterial biomass was quantified as a sum of i14:0, i15:0, a15:0, 16:1u7t,
16:1u9, 16:1u7, 10Me-16:0, i17:0, a17:0, cy17:0, 17:0, 10Me-17:0, 10Me-18:0 and cy19:0 (PLFAB).
The fatty acids found in both bacteria and fungi, 15:0, 16:0 and 18:1u7, were excluded from the
analysis (Tornberg et al., 2003). The relative content of individual PLFA molecules was also
calculated. The total content of all PLFA molecules (PLFAT) was used as a measure of total
microbial biomass. The fungal/bacterial biomass (F/B) ratio was calculated as PLFAF/PLFAB. Total
ergosterol was extracted and analyzed as previously described (Snajdr et al., 2008). Samples (0.5 g)
were sonicated with 3 mL 10% KOH in methanol at 70°C for 90 min. Distilled water (1 mL) was
added, and the samples were extracted three times with 2 mL cyclohexane, evaporated under
nitrogen, redissolved in methanol and analyzed isocratically using a Waters Alliance HPLC system
(Waters, USA) with methanol as a mobile phase at a flow rate of 1 mL
min

1. Ergosterol was detected by UV detection at 282 nm.

2.4. Statistics

Statistical analyses were performed using the software package Statistica 7 (StatSoft, USA). For
statistical purposes, all measurements were grouped into seasons (spring: 22.3e23.6, summer:
22.6e23.9, autumn: 22.9e23.12, and winter: 22.12e23.3). Autumn corresponded to the period of
litterfall of Q. petraea, with leaf abscission starting in late September, peaking in
October/November and continuing until mid-December. The year was also divided into a warm
period of 26 weeks (weeks 15e40), with mean weekly temperatures above the mean annual
temperature, and a cold period of 26 weeks (1e14 and 41e52). Differences among soil horizons
were tested by the Wilcoxon pair test always comparing the data from the same sampling time.
Differences between seasons were tested using one-way analysis of variance (ANOVA) followed by
the Fisher LSDpost hoctest. The correlations between individual variables were evaluated by linear
regression analyses based on the Pearson’s correlation coefficients. To study the effects of
temperature and moisture content on enzyme activities and microbial biomass, general linear
regression models (GLM, Statistica 7, StatSoft USA) were used. In these models, the percentage of
variability explained by individual factors (e.g., temperature and moisture) was calculated as a
ratio of the variability due to this individual factor and the total variability among samples. The
percentages of explained variability are only given for these factors when the effect of such factor
was statistically significant. In all cases, differences of P<0.05 were regarded as statistically
significant.

3. Results

3.1. Seasonal differences in soil temperature, enzyme production and microbial biomass content

Mean daily air temperatures at the study site varied between

6°C and 21°C, and the weekly averages ranged from

1.5°C in the winter to 18.5°C in the summer (Fig. 1). The mean annual temperature was 9.3°C. A
mean weekly air temperature below freezing was typically recorded during 2-3 weeks of the year,
and mean daily temperatures typically dropped below 0°C for 6-8 weeks each year. The amplitude
of soil temperatures decreased from the L horizon to the Ah horizon, with weekly minima and
maxima of 0.2°C and 15.7°C in the L, 1.4°C and 14.4°C in the O and 3.7°C and 11.1°C in the Ah
horizons. Litter horizon temperatures also showed high daily temperature variations, especially in
the warm months (data not shown). During the year, the lowest mean temperatures were
recorded in the winter (4.0°C in L) and the highest in the summer (14.7°C). Spring was slightly, but
insignificantly, warmer than autumn (9.6°C vs. 8.1°C in the L horizon; Fig. 2). In the L and O
horizons, soil moisture significantly correlated with temperature (P¼ 0.049 and 0.041,
respectively); however, in the Ah horizon, no such correlation was found. Mean moisture content
was, however, not different among seasons in any horizon (Table 1). There was a significant
difference in soil moisture among horizons, with Ah having the highest and the least variable
moisture content (Ah ¼ 0.250.08 g
g

1 ,O ¼ 0.390.10 g
g

1, and L ¼ 0.500.13 g
g

1). Enzyme pools and enzyme activities generally decreased from the L horizons to the Ah horizons
(Table 1, Fig. 2). The differences among horizons were significant during all seasons for most
enzymes, with the exception of Mn-peroxidase, whose activity in the O and Ah horizons was not
significantly different. Also, the content of microbial biomass significantly decreased from the L
through the O to the Ah horizon. When expressed per unit microbial biomass, specific enzyme
activities of most enzymes were significantly higher in the L horizon, but there were no significant
differences among the O and Ah horizons in this respect. As exceptions, the specific activities of
laccase and b-xylosidase did not differ among horizons in spring and summer, and the specific
Mnperoxidase activity was not different among horizons in all seasons. Pools of enzymes showed
high variation through-out the year, but only Mn-peroxidase activity had significant seasonal
differences in pools among seasons, as it was most abundant in autumn in all horizons (Table 1). In
addition, pools of cellobiohydrolase, N-acetylglucosaminidase and phosphomonoesterase
activities had seasonal maxima in autumn but only in the O horizon (Table 1). None of the
measurements of soil microbial biomass showed significant seasonal effects, except that the
PLFAB in the Ah horizon in winter was half that in summer (820 and 1660 ng
g

1, respectively; Table 1).

3.2. Climatic factors affecting seasonal variability in enzyme activity and microbial biomass

The activity of all extracellular enzymes in the soil extract increased with increasing temperature
of the reaction mixture (Fig. 3). Although the activity of most enzymes increased over the whole
range of tested temperatures, the highest activities of cellobiohydrolase and
phosphomonoesterase were measured at 25°C and not at 30°C. The Q10 values ranged from
approximately 1.4-1.5 for cellobiohydrolase and both ligninolytic enzymes, laccase and Mn-
peroxidase, to more than 2.5 for endocellulase, endoxylanase and N-acetylglucosaminidase (Table
2). The latter enzymes and b-glucosidase also showed the lowest activity at temperatures <5°C,
whereas the activities of cellobiohydrolase, laccase, phosphomonoesterase and b-xylosidase at 5°C
was relatively high (over 30% of the maximum). When enzyme activities were calculated for each
sampling date based on the pool of enzyme and its relative activity at the soil temperature during
the time of sampling, it was revealed that seasonal differences are often significant and that
enzyme activities tend to be high in the warm period and low in the cold winter (Fig. 2). Due to the
high variation in enzyme activities in the L horizon, only endocellulase, endoxylanase and N-
acetylglucosaminidase showed significant seasonal differences; all of the enzymes were more
active in summer then in winter (3-fold to 6-fold difference). In theOhorizon, activities of several
enzymes, laccase, endocellulase, b-glucosidase, b-xylosidase, N-acetylglucosaminidase, and
phosphomonoesterase, increased both in summer and autumn, whereas endoxylanase activity
was only high in summer. In the Ah horizon, the activities of laccase, b-xylosidase, N-
acetylglucosaminidase, and phosphomonoesterase were high in summer and autumn, and
endocellulase, endoxylanase andb-glucosidase activities were high in summer. Cellobiohydrolase
activity did not show any significant seasonal differences due to high variation of activity among
sampling dates. Exceptionally, the activity of Mn-peroxidase was only elevated in autumn, with 3-
fold to 4-fold higher activities in this season than during the rest of the year. When linear models
were constructed in which the variation of enzyme activities and microbial biomass content was
tested as a function of the soil temperature and moisture content (Table 3), temperature was
identified as the most important factor affecting enzyme activity, and soil moisture content was
found to be important for microbial biomass content. Soil temperature effects explained
substantial parts of the variability of certain enzymes, ranging up to 45% of the total (Table 3). Mn-
peroxidase, cellobiohydrolase and phosphomonoesterase activities were not related to
temperature, whereas endocellulase and endoxylanase activity strongly correlated with
temperature in all horizons. b-glucosidase and b-xylosidase activities increased with soil
temperature in the O and Ah horizons, that of N-acetylglucosaminidase in the L and Ah horizons
and that of laccase in the Ah horizon. Interestingly, the greatest effects of temperature on enzyme
activity were found in the Ah horizon, where the temperature variation was the least (Fig. 1). In
the L horizon, the activity of endoxylanase was significantly affected by both temperature and
moisture content; these two variables explained almost 60% of the total variability. The content of
PLFAF increased significantly with soil moisture in the drier O and Ah horizons, the moisture
content variation accounting for 30e43% of the total variability in PLFAF and the F/B ratio also
increased. In the O horizon, 850 ng
g

1 PLFAF and an F/B ratio of 0.32 were predicted for most samples (70% moisture), whereas these
values were only 200 ng
g

1 and 0.08 when soil was dry during sampling (26% moisture). In the Ah horizons, moist samples
were predicted from a linear fit to contain 280 ng
g

1 PLFAF and to have an F/B ratio of 0.22 at 50% moisture content, whereas these values were only
70 ng
g

1 PLFAF and an F/B ratio of 0.06 for dry samples with a moisture content of 15%. Because the
models of future climate development for central Europe predict temperature increases of 1-3°C
until 2050 (Jacob et al., 2007), we have tried to use the data obtained in this study to estimate the
change of actual enzyme activities under conditions of unchanged production for the situation of a
2°C temperature increase (Table 2). The predictions showed that such a temperature change
would result in a mean increase of enzyme activity, ranging 8-33% in the most affected L horizon,
9-28% in the O horizon, and 9-27% in the Ah horizon. Endocellulase and endoxylanase activities
are predicted to be the most affected. The predictions also show that although the activity in the
summer period would be less affected, enzyme activities during the winter may increase
substantially, often by more than 40%.

Fig. 1. Annual course of temperatures in a Quercus petraea forest. Panel A: Mean weekly air
temperature at soil surface (boxes) and lowest and highest mean daily temperature (vertical bars).
Panel B: Mean weekly temperatures in the middle of theL horizon (full line),Ohorizon (dashed line)
and Ah horizon (dotted line). The values were calculated from hourly temperature records from
September 2005-August 2008.

Fig. 2.Seasonal values of extracellular enzyme activities and soil temperature in a Quercus petraea
forest litter and soil. The data represent averages and standard errors. For each variable, values
with different letters are significantly different among seasons (P<0.05; ANOVA followed by
Tukeypost-hoctest).
4. Discussion

In the temperate Q. petraea forest of this study, temperature substantially varied seasonally, with
the difference of winter and summer averages above 10°C in the L horizon. High temperatures in
the L and O horizons were accompanied by a decrease of soil moisture content, but a dry season,
typical for warmer climates (Criquet et al., 2000, 2002), was never recorded. In autumn, the input
of organic matter in the form of litter doubled the mass of the litter horizon and changed its
composition. In a previous study, litter was demonstrated to differ before and after the litterfall in
its nitrogen content (1.98 0.13 versus 1.500.06), C/N ratio (21.8 0.9 versus 35.60.7) and acid-
insoluble residue (Klason lignin) content (52.5 1.7 versus 48.31.8). Fresh litter also contained
relatively more cellulose than old litter (Snajdr et al., 2011). Underground carbon allocation was
demonstrated to change dramatically over the year in various forest soils. Due to the absence of
green leaves between October and March, photosynthetic flow in the cold period of the year in
the hardwood forest is zero. Because the photosynthesis-allocated carbon is found mainly in
fungal PLFA markers (Högberg et al., 2010), the ratio of mycorrhizal to saprotrophic fungi is
supposed to decrease in winter. This might be theoretically reflected by the increase of
decomposition associated with a higher production of extracellular enzymes. Despite these
assumed changes in the soil nutrient availability, neither the soil microbial biomass content nor
the pools of most extracellular enzymes were significantly different among seasons. In previous
studies from temperate forest ecosystems, seasonal differences in enzyme pools or microbial
biomass content were occasionally reported. The pools of phosphomonoesterase, N-
mineralization and denitrification enzymes in hardwood forest soil had seasonal fluctuations, with
maxima between March and August, whereas b-glucosidase pools did not show consistent
seasonal patterns (Bohlen et al., 2001;Rastin et al., 1988). Neither arylsulfatase nor
dehydrogenase had seasonal pool changes in temperate coniferous forests (Prietzel, 2001; Rogers
and Tate Iii, 2001).

Table 1; Seasonal values of soil moisture, pH, pools of extracellular enzymes and microbial biomass
content in aQuercus petraea forest litter and soil. The data represent averages and standard
errors. For each variable, values with different letters are significantly different among seasons
(P<0.05; ANOVA followed by Tukeypost-hoctest). Pools of extracellular enzymes are expressed as
activity measured at standard temperature.

Seasonality also did not significantly affect enzyme pools in O horizons in deciduous forests in
southern Finland sampled during the vegetation season (Niemi et al., 2007). Organic matter
decomposition, represented by soil respiration, typically increases with increasing temperature
and soil moisture content although either factor alone is able to have an over-riding influence on
decomposition if conditions cross effect thresholds (Prescott, 2005, 2010). The effects of moisture
are, however, more apparent in ecosystems with severe long-term drought periods. In the litter of
Mediterranean Quercus forest, pools of laccase, phosphomonoesterase and glycosyl hydrolases
were substantially affected by moisture content, and Mn-peroxidase activity was only detected
during the moist autumn (Criquet et al., 2000, 2004, 2002). A relatively moderate drought
significantly decreased the pools of urease, protease, b-glucosidase and phosphomonoesterase in
Quercus soils (Sardans and Peñuelas, 2005). Soil moisture content in Q. petraea forests was
identified as one of the key factors affecting local microbial abundance and the pools of some
enzymes (Baldrian et al., 2010b). Here, we show that over the course of a year, moisture but not
temperature substantially affects fungal biomass content in the O and Ah horizons of soil and
reduces the F/B ratio. This is in contrast to coniferous forests, where both the bacterial counts and
fungal hyphal lengths increased with the soil moisture content (Berg et al., 1998). To quantify
enzyme activity in an environment with changing temperature, the effects of temperature on
enzyme activity must be considered. Temperature increases the rates of enzymatic reactions as
long as the stability of a particular enzyme is not affected. For example, the Q10 values for purified
fungal laccases and Mnperoxidases range from 1.5 to 2.0. Although the reaction rates of these
enzymes increase with temperature up to more than 50°C, enzyme stability is negatively affected
and the enzymes lose their activity rapidly at temperatures above 30°C(Snajdr and Baldrian, 2007).
Also, the activity of several hydrolytic and oxidative enzymes in soil or litter extracts was
demonstrated to substantially increase with temperature (Criquet et al., 1999; McClaugherty and
Linkins, 1990). The Q10 values of individual enzymes from boreal and temperate soils typically
ranged between 1.5 and 3.0 (Wallenstein et al., 2009; Stone et al., in press), roughly corresponding
to the values determined in this work. The Q10 values are enzyme-specific and likely differ among
soils and seasons. A global Q10 of 1.37 for the decomposition of organic matter was calculated in a
metaanalysis, the value being lower than those reported for enzymes.

Fig. 3.Temperature dependence of enzyme activities in soil extracts fromQuercus petraeaforest

soil. Activities are expressed as the percentage of the highest recorded activity.

The data represent averages and standard errors obtained from three independently extracted
samples.

Table 2: Q10 (fold-increase of enzyme activity with a temperature increase of 10°C) values for
extracellular enzymes from Quercus petraea forest litter and soil and an estimate of enzyme
activity increase with an increase of temperature by 2°C (as the percentage of current values). The
Q10 data are averages and standard errors of values calculated for temperatures between 5°C and
25°C. The increase in the percentage of the present enzyme activity with increasing temperature is
integrated for the whole year (annual), and the values for summer and winter (the seasons with
minimal and maximal changes for most enzymes) are given as the range.

Decomposition was predicted to increase rapidly at low temperatures but only moderately above
20°C, with an optimum around 33°C (Ise and Moorcroft, 2006). Importantly, the absolute
decomposition rates were demonstrated to be strongly controlled by litter water potential
(Moore, 1986). Here, we show that in the temperate Q. petraea forest, soil moisture content does
not represent a limitation of enzymatic activity because it does not drop to low values for long
times. The predictions of the actual rates of enzymatic reactions based on enzyme pools and
temperature-activity relationships were previously only studied in cold biomes. Temperature was
found to be the major factor driving seasonal differences in enzyme activities in the tundra.
However, summer activities were low, possibly due to N-limitation resulting in low enzyme pools
(Wallenstein et al., 2009). In a boreal coniferous forest, potential enzyme activities were predicted
to result in 7-32% of total annual activity in the cold period and 68-93% in the warm period
(Wittmann et al., 2004). Here, we show that also in the temperate forest, where temperature
differences are less pronounced, temperature still importantly affects enzyme activity, resulting in
higher activity in the warm seasons. Between 14 and 54% of total annual activity in the cold period
would be predicted for individual enzymes in the L horizon, 8-57% for the O horizon and 4-46% for
the Ah horizon, the means for each horizon being 37%, 36% and 31%, respectively. Contrary to our
hypothesis, seasonal effects were more pronounced in the O and Ah horizons than in the L
horizon, despite their lower annual temperature variation. One of the possible explanations is that
the higher stability of chemical properties in deeper horizons reduces variation in enzyme
production. The mean mass loss of last year’s litter at the site of this study was 0.038 g
month

1g

1 during the first six months after litterfall (late autumn to early spring), in the following warmer
six months, it reached 0.061 g
month

1g

1(Snajdr et al., 2011). This seems to correspond well to the 37% of the annual decomposition in
the cold period comparable to the value predicted for enzyme activity. However, at least a part of
this mass loss in the coldest months is likely due to the leaching of solubles from the litter since
the calculated enzyme activities for the winter period are very low; previously, the mass loss of
fresh litter due to extensive leaching was demonstrated to be>5% (Taylor and Parkinson, 1988).
Our theoretical prediction shows that an increase of temperature might increase enzyme activities
in soils substantially, provided that the enzyme pools remain the same and soil moisture is not
limiting. The increase of enzyme activity is predicted to be more pronounced in the winter than in
the summer. As a consequence, relatively more organic matter is likely to be decomposed during
the cold period of the year than has occurred recently. Global warming can thus ultimately lead to
greater differentiation of the cold period, with prevailing decomposition and dominance of soil
saprotrophs, and the summer, with an increased importance of carbon flow into soil from plant
photosynthates and the increased dominance of mycorrhizal fungi and other root-associated
microorganisms.

These theoretical models, however, contrast with recent results from experimentally warmed
soils. In the organic and mineral soil horizons of a temperate hardwood forest, microbial biomass
and substrate-induced respiration decreased after 12 years of heating by 5°C. Fungal biomass
decreased more than bacterial biomass, and PLFA analyses indicated changes in community
structure (Frey et al., 2008). In mountainous forests of the temperate zone, temperature increases
by 4-5°C resulted in no change or a decrease in soil microbial biomass content, enzyme activity
and carbon-use efficiency of the microbial community (Arnold et al., 1999; Schindlbacher et al.,
2011). One of the probable causes is the reduction of soil moisture content. In a boreal forest, a
temperature increase by as little as 0.5°C resulted in a 22% decrease of soil moisture content,
accompanied with the decrease of fungal and bacterial biomass content and N-
acetylglucosaminidase pools (Allison and Treseder, 2008). This study shows that, temperate
forests exhibit significant seasonal variations in enzyme activities in litter and soil. These changes,
which might ultimately result in the changes in the rates of decomposition, are not due to the
changes in microbial abundance or enzyme pools but rather result from the effects of temperature
on enzyme activity. If the global temperature changes in the future, the seasonal differences in
decomposition may be less significant due to the predicted higher increase of decomposition in
the cold period than in the warm period.