Fungi are the agents primarily responsible for the transformation of plant-derived carbon in terrestrial ecosystems. However, little is known of their responses to the seasonal changes in resource availability in deciduous forests, including photosynthate allocation below ground and seasonal inputs of fresh litter.
Vertical stratification of and seasonal changes in fungal abundance, activity and community composition were investigated in the litter, organic and upper mineral soils of a temperate Quercus petraea forest using ergosterol and extracellular enzyme assays and amplicon 454-pyrosequencing of the rDNA-ITS region.
Fungal activity, biomass and diversity decreased substantially with soil depth. The highest enzyme activities were detected in winter, especially in litter, where these activities were followed by a peak in fungal biomass during spring. The litter community exhibited more profound seasonal changes than did the community in the deeper horizons. In the litter, saprotrophic genera reached their seasonal maxima in autumn, but summer typically saw the highest abundance of ectomycorrhizal taxa. Although the composition of the litter community changes over the course of the year, the mineral soil shows changes in biomass.
The fungal community is affected by season. Litter decomposition and phytosynthate allocation represent important factors contributing to the observed variations.
Temperate forests are one of the major biomes on earth, covering an area of 570 million hectares and thus playing an important role in the global carbon (C) budget (FAO & JRC, 2012). In forest ecosystems, C enters the soil in the form of plant litter (Berg & McClaugherty, 2003), through the belowground allocation of C fixed by plant photosynthesis (Högberg et al., 2010) and as a dead fungal and animal material. Fungi play the primary role in regulating the flow of C through all of these pathways. Saprotrophic fungi decompose organic matter as a result of their ability to produce a wide range of extracellular enzymes (Steffen et al., 2007; Baldrian et al., 2011), which allows them to efficiently attack the recalcitrant lignocellulose matrix that other organisms are unable to decompose (Boer et al., 2005). Mycorrhizal fungi, as obligate symbionts, acquire access to the C compounds derived from the photosynthates of their host plants (Hobbie, 2006) in exchange for soil-derived nutrients (van der Heijden & Horton, 2009), and also contribute directly to the C enrichment of soils by mediating the belowground allocation of C from plant roots to soil (Clemmensen et al., 2013).
In temperate deciduous forests, as a consequence of the input of new litter and its transformation, it is possible to recognize three distinct compartments in the soil profile: the litter (L horizon), containing organic matter derived from dead plant biomass almost exclusively; the organic (or humic) H horizon, representing a mixture of processed plant-derived organic matter and soil components; and the mineral soil horizon, with a lower content of organic matter, originating both from the decomposition of organic matter and exudation from the abundant tree roots. If invertebrate mixing is limited, the age of the litter-derived organic material increases with soil depth as decomposition progresses, and this is accompanied by changes in its chemical composition, leading to increasing recalcitrance and the formation of humic compounds (Šnajdr et al., 2008). The vertical distribution of the fungal community in boreal and temperate forests has been demonstrated to reflect soil stratification: saprotrophic taxa are more abundant close to the surface of the forest floor where most C is mineralized, whereas mycorrhizal fungi increase in abundance with soil depth, where they mobilize nitrogen to be supplied to the roots of plants (O'Brien et al., 2005; Lindahl et al., 2007).
Observations from diverse forest soils suggest that environmental factors, such as temperature, water availability and substrate quality, may be important factors affecting microbial community composition (Aponte et al., 2010; Kaiser et al., 2010; Landesman & Dighton, 2011; Kuffner et al., 2012). Temperate deciduous forests are characterized by the photosynthetic activity of trees during the vegetative period and a short period of litterfall in autumn, when fresh litter with easily available nutrients accumulates on the forest floor (Šnajdr et al., 2011). These seasonal processes then underlie the seasonality of soil C allocation and its availability to the soil biota (Högberg et al., 2010; Kaiser et al., 2010). Belowground C allocation via plant roots exhibits several-fold seasonal differences with a maximum during the late vegetative season (Högberg et al., 2010).
Seasonal variations in fungal communities have been widely studied; however, the methods used have been unable to sufficiently characterize fungal community structure or the authors have focused only on a particular soil horizon or group of fungi. Previous studies have been based mainly on traditional approaches, such as enzyme assays or the assessment of microbial biomass (Berg et al., 1998; Björk et al., 2008; Šnajdr et al., 2008; Baldrian et al., 2013a). There have thus far been several reports concerning specific functional groups of fungi (Rosling et al., 2003; Koide et al., 2007; Courty et al., 2008, 2010a,b) or studies limited to particular soil or litter horizons (Jumpponen et al., 2009; Dumbrell et al., 2011; Davey et al., 2012; Coince et al., 2013). However, a thorough knowledge of seasonal influences on fungal communities in temperate deciduous forest soil with respect to vertical stratification is missing. Understanding of the seasonal dynamics of the fungal community and its functioning in forest soil ecosystems is necessary for the prediction of its response to global changes, considering soils as a sink of CO2. Moreover, even though detailed descriptions of fungal communities by soil horizons exist for some ecosystems (Baldrian et al., 2012; Uroz et al., 2013), it is unclear how representative these studies might be based on a single sampling.
In this study, an analysis of fungal community composition, abundance and the activity of extracellular enzymes was performed for the upper horizons of a deciduous Quercus petraea forest. The primary goal of this study was to describe the seasonal variations in the fungal community composition in the context of changing resource availability throughout the year. To achieve this, sampling was performed in spring (early May) shortly after leaf appearance, in summer (July), together with the highest temperatures and high photosynthetic production, in autumn (October) in the middle of the litterfall period and in winter (February; Fig. 1). In addition to changes in tree productivity, the amount and quality of litter changed with the input of freshly fallen leaves (c. 4 t ha−1; J. Voříšková, unpublished results), whose composition supports fast decomposition (October) to a litter horizon depleted of easily decomposable compounds (July). We hypothesized that the structure of the fungal community would reflect the availability of nutrients in the soil profile horizons. Based on a previous study, where considerable temporal shifts in fungal community structure were observed during the decomposition of oak litter, we anticipated similar changes in the litter horizon because the last year's litter represents a considerable percentage of the total litter mass (Voříšková & Baldrian, 2013). In the deeper horizons, we expected a shift from a high relative abundance of ectomycorrhizal (ECM) taxa during the vegetative season to a high proportion of saprotrophic taxa in the absence of root C allocation, because our previous study showed that saprotrophic taxa are more metabolically active during the period in which photosynthesis does not occur (Baldrian et al., 2012). We also intended to answer questions concerning the suitability of the one-time surveys that are frequently reported for the description of fungal community composition within an ecosystem.
Materials and Methods
Study site and sample collection
The study site was an oak (Q. petraea (Matt.) Liebl) forest in the Xaverovský Háj Natural Reserve, near Prague, Czech Republic (50°5′38″N, 14°36′48″E). The site has been studied previously with respect to decomposition-related extracellular enzymes in the forest topsoil (Šnajdr et al., 2008; Baldrian et al., 2010, 2013a), as well as the decomposition of litter and associated changes in fungal community composition (Voříšková & Baldrian, 2013). The soil was an acidic cambisol with developed L, H, Ah and A horizons. Sampling of the topsoil was performed in the spring (9 May, c. 2 wk after the emergence of leaves), summer (29 July), autumn (28 October, during the late phase of litterfall) and winter (19 February; Fig. 1). Soil samples were collected in four defined plots (10 m2, c. 100 m from each other) of the sampling site. Six soil cores (4.5 cm in diameter) were collected at each sampling plot and were divided into L horizon (c. 0.5–1 cm, 3.5), H horizon (c. 1–3 cm) and Ah soil horizon (upper portion, up to a depth of 5 cm). Samples of the L horizon were cut into c. 0.25-cm2 pieces, and the soil samples were sieved using a 2-mm sieve. The resulting material was combined to yield a composite sample from each horizon and plot. Subsamples for chemical analyses, quantification of microbial biomass and DNA extraction were frozen and stored at −45°C; subsamples for enzyme analysis were stored at 4°C.
Enzyme assays were performed within 48 h on samples extracted using 160 mM phosphate buffer, pH 7 and desalted using Sephadex columns, as described previously (Šnajdr et al., 2008). Briefly, laccase was assayed using the oxidation of 2,2′-azinobis-3-ethylbenzothiazoline-6-sulfonic acid, manganese peroxidase using 3-methyl-2-benzothiazolinone hydrazone and 3,3-dimethylaminobenzoic acid in the presence of manganese and hydrogen peroxide, endocellulase and endoxylanase using azo-dyed carboxymethyl cellulose and birchwood xylan, and the activities of all other enzymes using p-nitrophenyl-based substrates. All enzyme assays were performed at pH 5, except for laccase, where the pH of the buffer was 4.5. One unit of enzyme activity was defined as the amount of enzyme forming 1 mmol of reaction product per minute.
Dry mass content was measured after drying at 85°C, organic matter content after burning at 650°C, and pH was measured in distilled water (1 : 10). C and nitrogen (N) contents were measured using an elemental analyser.
Total ergosterol was extracted with 10% KOH in methanol and analysed by high-performance liquid chromatography (HPLC) (Šnajdr et al., 2008) using a method modified from Bååth (2001).
454-Pyrosequencing of fungal internal transcribed spacer (ITS)
Total genomic DNA was extracted from 300 mg of soil material using a modified SK method according to Sagova-Mareckova et al. (2008). The primers ITS1/ITS4 (White et al., 1990) were used to amplify the ITS1 region, the 5.8S ribosomal DNA and the ITS2 region of the fungal ribosomal DNA. The primer pair used in this study is especially suitable for the analysis of Ascomycota, Basidiomycota, Mucoromycotina and Mortierellomycotina, whereas it may be biased against some members of the Glomeromycota and Chytridiomycota. A two-step PCR amplification using composite primers containing multiplex identifiers (Baldrian et al., 2012) was performed to obtain amplicon libraries for 454-pyrosequencing. In the first step, each of three independent 25-μl reactions per DNA sample contained 2.5 μl of 10 × polymerase buffer, 1 μl of each primer (0.01 mM), 0.5 μl of PCR Nucleotide Mix (10 mM) and 0.25 μl of polymerase (2 U μl−1; Pfu DNA polymerase : OmniTaq DNA polymerase, 1 : 24). The cycling conditions were 94°C for 5 min; 35 cycles of 94°C for 1 min, 60°C for 1 min and 70°C for 1 min; followed by 70°C for 10 min. Pooled PCR products were purified using a MinElute PCR Purification Kit (Qiagen, Hilden, Germany). The product of the first PCR was used as a template for the second PCR. In the second step, one 50-μl reaction per DNA sample contained 5 μl of 10 × polymerase buffer, 1.5 μl of dimethylsulfoxide (DMSO) for PCR, 0.4 μl of forward fusion primer (ITS1, tag sequence, 454-specific sequence), 0.4 μl of reverse fusion primer (ITS4, 454-specific sequence), 1 μl of PCR Nucleotide Mix, 1.5 μl of polymerase (2 U μl−1; Pfu DNA polymerase : Dynazyme DNA polymerase, 1 : 24) and 100 ng of template DNA. The cycling conditions were 94°C for 5 min; 10 cycles of 94°C for 1 min, 62°C for 1 min and 72°C for 1 min; followed by 70°C for 10 min. PCR products were purified using Agencourt AMPure XP (Beckman Coulter, Beverly, MA, USA). The concentration of PCR products was quantified using the Qubit 2.0 Fluorometer (Life Technologies, Carlsbad, CA, USA), and an equimolar mix of PCR products from all samples was prepared. The mixture of PCR products was separated by electrophoresis and gel purified using the Wizard SV Gel and PCR Clean-Up System (Promega, Madison, WI, USA), followed by purification using Agencourt AMPure XP and a MinElute PCR Purification Kit to remove primer dimers. The amplicons were subjected to sequencing on a GS Junior 454-pyrosequencer (Roche, Basel, Switzerland).
The pyrosequencing data were processed using the pipeline SEED with respect to the proposed procedures of standardized data analysis (Nilsson et al., 2011; Větrovský & Baldrian, 2013). Pyrosequencing noise reduction was performed using the Denoiser 0.851 (Reeder & Knight, 2010), and chimeric sequences were detected using UCHIME (Edgar et al., 2011) and deleted. The sequences were shortened to 380 bases and clustered using Usearch (Edgar, 2010) at a 97% similarity level. Consensus sequences were constructed for each cluster, and the operational taxonomic units (OTUs) were constructed by clustering these consensus sequences at 97% identity (Lundberg et al., 2012). The abundance data reported in this paper are based on this dataset of sequence abundances and should be taken as proxies of taxon abundances only with caution (Lindahl et al., 2013). Closest hits were identified using the PlutoF pipeline (Tedersoo et al., 2010); non-fungal sequences (< 1%) were disregarded. Sequence data have been deposited in the MGRAST public database (http://metagenomics.anl.gov/, dataset number 4524551.3).
Diversity and statistical analysis
The Shannon–Wiener index and the amount of the most abundant OTUs that represented 80% of all sequences were used as diversity estimates, providing combined information on species richness and evenness at particular sampling depths. These estimates were calculated for a dataset containing 1800 randomly chosen sequences from each sample. As the fungal communities at individual plots differed and because these among-plot differences might have hidden the differences among seasons, comparisons of seasonal abundances were performed after normalization of the abundances of each fungal taxon using the mean abundance in the particular plot and horizon during all seasons. Because the majority of taxa were represented by a very small number of reads, and because such read counts were demonstrated not to be technically reproducible (Lundberg et al., 2012), only taxa with higher relative abundances (≥ 0.5% in ≥ 5 samples) were tested for seasonal variations in abundance. The plot-normalized abundances of these measurable taxa were also subjected to a principal component analysis (PCA) together with environmental variables. The Jaccard Index (JI) calculated for all OTUs with relative abundances ≥ 0.5% in at least one sample was used as a measure of community similarity (Koleff et al., 2003). The JI is calculated as A/(A + B + C), where A is the number of species found in both of the samples, and B and C represent the number of species unique to either of the two samples analysed. JI ranges from zero (no species shared) to unity (all species shared). The pipeline SEED (Větrovský & Baldrian, 2013) was used for data pre-processing and diversity calculations, and Statistica 7 (Statsoft, Tulsa, OK, USA) was used for statistical analyses. A one-way analysis of variance with Fisher's least significant difference post hoc test was used to analyse the statistical significance of differences among groups of samples. Differences with P < 0.05 were regarded as statistically significant.
Soil properties, activity of extracellular enzymes and fungal biomass
The soil properties changed substantially with soil depth: the organic matter content decreased from 82% in the L to 42% and 16% in the H and Ah horizons, respectively, and the N content showed a similar inter-horizon trend. With depth, the soil dry mass content increased and the soil pH decreased. With decreasing organic matter content in the soil, the activity of extracellular enzymes also decreased, being 4–40× lower in the Ah horizon than in the L horizon. Fungal biomass in the H horizon was 6× and, in the Ah horizon, 23× lower than in the litter; the fungal biomass content per gram of organic matter decreased with depth from 188 μg g−1 in the litter to 76 μg g−1 and 50 μg g−1 in the H and Ah horizons, respectively (Fig. 2).
Fungal biomass was similar among the seasons in the H horizon, ranging between 26 and 36 μg g−1 soil dry mass. In the L horizon, the ergosterol content was highest in spring (272 μg g−1, compared with the minimum of 128 μg g−1 in the autumn). In the Ah horizon, the ergosterol content was increased significantly in summer (14 μg g−1), whereas it was only between 5 and 6 μg g−1 during the winter and spring (Fig. 1). The majority of the enzymes studied showed their highest activity in winter, especially in the L and H horizons. The exception was laccase, whose activity was highest in summer (Fig. 1).
Fungal community composition
In total, 213 339 raw sequences were obtained from 454-pyrosequencing, 135 830 of which remained for analysis after quality filtering, de-noising and the removal of short and chimeric sequences and sequences not belonging to fungi (the latter accounted for < 1% of the total). An average of 2830 sequences was obtained (minimum of 1803) per sample. All of the sequences clustered into 8264 OTUs (including 5730 singletons) at a 97% similarity threshold (Supporting Information Table S1). Fungal diversity, expressed as the Shannon–Wiener index calculated at 1800 sequences/sample, decreased from the L (4.51 ± 0.49) to the H (4.05 ± 0.36, P = 0.003) horizon, and from the H to the Ah horizon (3.43 ± 0.40, P < 0.0001). Seasonal differences were not observed, except in the Ah horizon, where summer communities were marginally more diverse than winter communities (P = 0.06). Community evenness, expressed as the number of the most abundant OTUs that represented 80% of all of the sequences in each sample, also decreased significantly with soil depth and did not show seasonal variations (Fig. 1).
The fungal communities at each plot and season were more similar between the H and Ah horizons (mean JI = 0.513) than between the L and H horizons (mean JI = 0.451; P = 0.013 that the similarity expressed as the JI between L/H and H/Ah samples from the same season and plot is the same). The L horizon communities were more similar among plots for each season than were the samples from the two deeper horizons (mean JI for L of 0.488, H of 0.402 and Ah of 0.379; P < 0.001 that the JI for each season across plots in L is the same as in H and Ah), and were also more similar for each plot across seasons (JI = 0.466, 0.422 and 0.397; P < 0.017 that the JI for each plot across seasons in L is the same as in H and Ah).
The overall fungal community was dominated by sequences assigned to the Basidiomycota (58%) and Ascomycota (27%). The Mucoromycotina were represented in 7.9% of all sequences, and fungi from the Glomeromycota and Chitridiomycota comprised 3.4% and 2.2% of all sequences, respectively. Sequences from the Ascomycota and Basidiomycota demonstrated comparable counts in the L horizon – 42% and 48%. With soil depth, the abundances of the ascomycetous sequences decreased and those of the basidiomycetous fungi increased to reach 15% and 71%, respectively, in the Ah horizon (Fig. 2). The most abundant fungal orders were the basidiomycetous Russulales (25%), Agaricales (11%) and Tremellales (8.7%). Members of the orders Agaricales, Helotiales and Tremellales dominated in samples from the L horizon. The abundances of the Agaricales and Tremellales did not change significantly with increasing soil depth, whereas the abundance of the Helotiales decreased. By contrast, the H and Ah horizons contained more sequences belonging to the ECM Russulales, their relative abundance in the Ah horizon being six times higher than in the L horizon.
In total, 757 fungal genera were identified as being the best hits for the OTUs for the whole dataset. The most abundant fungal genera in the litter horizon were the saprotrophic Mycena, Sistotrema and Cryptoccocus, whereas the deeper horizons were enriched with fungi belonging to the ECM genera Russula and Lactarius. The genus Russula represented 18% of all the sequences in the H horizon and 31% in the Ah horizon; the most abundant OTUs across all samples also belonged to Russula (OTU133, 100% similarity to R. atropurpurea) and Lactarius (OTU002, 98% similarity to L. quietus; Table S1).
Fungi in the litter were apparently more influenced by seasonal effects than were those in the deeper horizons: 59% of the abundant fungal genera (among them, eight of the top 10) showed statistically significant differences in their seasonal abundance, compared with 29% and 32% in the H and Ah horizons, respectively (Table S2).
The seasonal differences among the relative abundances of fungal genera in the L horizon were profound: the saprotrophic genus Mycena was represented by only 0.5% of the sequences in winter, whereas it represented 16% in spring, Mycosphaerella represented 0.04% in summer but 8% in autumn, and Naevala represented 0.01% in summer and 3.5% in winter. For most ECM fungi, low relative abundances were recorded in spring (Russula, 1%; Lactarius, 0.3%; Amanita, 0.05%), whereas high abundances were recorded in summer (Russula, 7%; Lactarius, 6%; Amanita, 3.4%; Table S2). These seasonal changes were also demonstrated in the PCA in which samples from winter clustered separately from those of the other seasons (Fig. 3). In the autumn, when fresh litter accumulated on the forest floor, the saprotrophic genera Mycosphaerella, Mucor, Geomyces, Umbelopsis and Lachnellula reached their seasonal maxima. The highest activity of most enzymes was recorded in winter, as well as the highest C : N ratio. This finding was accompanied by the highest abundances of the saprotrophic genera Cryptococcus, Rhodotorula, Naevala, Fulvoflamma and Kriegeria. The other saprotrophic genera Mycena, Cladophialophora and Meliniomyces were abundant during the spring, the season with the highest fungal biomass in the litter. Finally, summer (the driest season, with high laccase activity) typically demonstrated the highest abundances for all ECM taxa (Russula, Lactarius, Tomentella, Amanita and Hygrocybe, except for Xerocomus, which was highest in autumn). Not surprisingly, the proportion of sequences for the ECM fungi was highest in summer at 28.6%, which was significantly higher than that in spring (8.5%).
In the H horizon, PCA showed a clear separation between winter and spring samples along the first axis. Winter was characterized by high activities of most extracellular enzymes and by the lowest proportion of ECM fungi (35%), whereas spring exhibited the highest proportion of ECM sequences (58%). In the Ah horizon, PCA separated summer samples from those of spring and winter. The proportion of ECM did not show significant fluctuations, but the proportions of individual ECM fungi varied seasonally. The highest activities of several enzymes, but the lowest fungal biomass, were observed in winter (Fig. S1).
Forest soils represent an environment that exhibits distinct and sharp vertical stratification. The ultimate cause is likely to be the decrease in organic matter with soil depth as a result of the accumulation of litter on the soil surface and its gradual decomposition, together with temperature and moisture content differences. The decrease in soil organic matter content is accompanied by a decrease in microbial biomass and in the rates of microbial processes, such as respiration and the activities of extracellular enzymes (Agnelli et al., 2004; Šnajdr et al., 2008; Baldrian et al., 2013a). In this study, soil organic matter decreased by a factor of five between the L and Ah horizons. The fact that fungal biomass in the Ah horizon was 23× lower than that in the L horizon probably reflects changes in the quality of the organic matter. Taking into account the ergosterol/fungal biomass ratio of 3.8 mg g−1 fungal biomass (Baldrian et al., 2013b), fungal biomass might represent as much as 6.1% of the organic matter in the litter horizon; in the H and Ah horizons, this biomass would be less, namely 2% and 1.2%, respectively. Previous studies have also shown that the C : N ratio decreases with soil depth in certain soils (Baldrian & Štursová, 2011; Yang & Luo, 2011). During the process of decomposition, C from freshly fallen litter is released as CO2 and, if the N is retained, its relative proportion increases. For example, fresh Q. petraea litter has a C : N ratio of 25, compared with 13–17 after in vitro degradation by saprotrophic fungi (Steffen et al., 2007). Here, we observed the highest C : N ratio in the Ah horizon, a fact that might support the importance of the allocation of C from tree roots into deeper soil, as proposed by Clemmensen et al. (2013), who demonstrated that 70% of soil C was root derived and was thus allocated from plants into the soil by mycorrhizal fungi. Alternatively, the increased C : N ratio in mineral soil might be caused by the depletion of N in the bulk soil as a result of its allocation to plants by ECM fungi.
The fungal community structure differed substantially among the three horizons studied. Litter-associated communities exhibited higher similarity among sampling plots than did soil communities. Within each plot, the communities of the deeper H and Ah horizons were more similar to each other than to the L horizon community, mainly as a result of the comparable abundances of ECM taxa. Interestingly, our study showed a substantial, significant decrease in community diversity: the amount of OTUs representing 80% of the fungal community was 90 in the L horizon, 51 in the H horizon and 25 in the Ah horizon (Fig. 1). Similar reductions in fungal diversity have been demonstrated previously in prairie soils, but a reduction in the Shannon index was only observed over several tens of centimetres of soil depth (Jumpponen et al., 2010). In the forest ecosystems, the higher chemical heterogeneity of nutrient sources in the litter horizon (composed of material of various ages) might be the prerequisite for high diversity; the cause of the higher diversity in the H horizon than in the Ah horizon might be the co-presence of organic substrates and mycorrhizal tree roots in the former. However, previous papers reporting on other forest ecosystems did not show differences in fungal diversity among horizons (O'Brien et al., 2005; Baldrian et al., 2012).
With respect to the abundance of the major functional groups of fungi, our results are in accordance with previous studies from boreal and temperate forest soils in that the relative proportion of ECM taxa (and thus the Basidiomycota) increases with soil depth (Lindahl et al., 2007; Edwards & Zak, 2010; Baldrian et al., 2012; Clemmensen et al., 2013). However, interestingly, if we consider the absolute amounts of fungal biomass, the highest ECM biomass per gram of soil dry mass, despite its lower proportion, was present in the L horizon. The supply of C from trees to the ECM fungi thus contributes substantially to the formation of fungal biomass in the litter. The fact that the ECM biomass in litter is produced from root-supplied C and is not obtained from saprotrophic litter transformation is supported by the substantial increase in the proportion of ECM in the litter from spring to summer. Summer would then also be the season with the highest total ECM biomass when considering the whole soil profile.
Despite the increasing evidence that enzyme activity shows seasonal variation (Wittmann et al., 2004; Baldrian et al., 2013a), as does the C allocation below ground (Ekblad et al., 2013), the influence of these factors on soil fungal communities has never been addressed in sufficient detail. Several studies have focused on seasonal variations in ECM fungi (Buée et al., 2005; Koide et al., 2007; Courty et al., 2008); moreover, Parrent & Vilgalys (2007) showed their seasonality across several years. However, little information is thus far available on the seasonality of entire fungal communities (Schadt et al., 2003; Baldrian et al., 2013a). In this study, the activity of extracellular enzymes showed significant seasonal variation, with the highest activity for most enzymes being detected in winter in all horizons, but being most pronounced in the litter (Fig. 1). This result is not surprising because the fresh litter shed in late autumn contains easily degradable compounds, and its decomposition is rapid over the whole winter period (Šnajdr et al., 2011). Consistent with the high enzyme activities over the winter, fungal biomass in litter in the spring increased, with a relative increase in the proportion of non-mycorrhizal taxa (Fig. 1).
The litter horizon also exhibited the highest proportion of fungal genera showing seasonal variation. This finding is in agreement with a previous study from a Mediterranean forest, where the extent of the seasonal variations in microbial communities also decreased with soil depth as a result of higher seasonal variations in environmental conditions (Andreetta et al., 2012). The changes in the fungal community in the litter were profound. The saprotrophic fungal genera Mycena, Mycosphaerella and Naevala showed 30×, 200× and 350× differences in abundance among seasons. This result partly reflects succession on fresh litter: Mycosphaerella, which peaked in autumn, is typical of senescent and freshly fallen oak leaves (Voříšková & Baldrian, 2013), and the other autumn fungal genera, for example, Mucor, Umbelopsis and Lachnellula, also belong to taxa that grow rapidly in the nutrient-rich environment of fresh litter (Osono, 2006). In addition, the fungi that increased in winter, Naevala, Rhodotorula and Cryptococcus, are typical saprotrophs. These genera were found to be associated with litter decomposition c. 4 months after abscission (Voříšková & Baldrian, 2013), and thus seem to be supported nutritionally by the last year's litter, which represents, in the ecosystem studied, c. 40% of the total mass of the L horizon (data not shown). The summer was characterized by a dramatic increase in ECM abundance: compared with spring, the abundance of Amanita increased 68×, of Lactarius 20× and of Russula 7× (Table S2), and the relative abundance of ECM fungi increased from 9% to 29% (Fig. 1). The increased abundance of ECM fungi in late summer or autumn has also been reported previously from boreal forests (Wallander et al., 2001; Högberg et al., 2010; Davey et al., 2012).
In the H and Ah horizons, seasonal differences in abundance were recorded for 30% of the dominant taxa. Both horizons exhibited higher enzyme activity during winter. Although, in the H horizon, this might partly have been the result of the priming effect of nutrients leached from litter, the fact that enzyme activity also increased in the Ah horizon may indicate instead the switch from the use of root-supplied photosynthates to the decomposition of organic matter. This switch was able to maintain a comparable fungal biomass content in spring as well as in winter. Contrary to our expectations, the relative proportion of saprotrophic fungi did not increase during winter–spring as a result of this switch to decomposition. One of the possible explanations could be the temporal switch of certain ECM taxa to a saprotrophic lifestyle, allowing them to preserve their biomass. Although the saprotrophic abilities of ECM fungi are still debated (Ekblad et al., 2013), there is growing evidence derived from enzymatic analyses and in vitro experiments that they are involved in the decomposition of litter (Courty et al., 2010a,b; Rineau et al., 2012). This explanation would be in agreement with observations that the enzymatic activities of ECM root tips increased before bud break in oak trees (Courty et al., 2007) and that litter-derived C is accumulated by oak roots via ECM mycelia (Bréda et al., 2013). Recent evidence indicates that members of the genus Russula contain genes for both exocellulases and ligninolytic peroxidases (Bödeker et al., 2009; Štursová et al., 2012; Voříšková & Baldrian, 2013). Their temporary saprotrophy is quite possible because the relative abundance of this genus during winter was less reduced than that of the other ECM fungi. The fungal biomass content in the Ah horizon increases approximately threefold from spring to summer, which corresponds to the expected increase in photosynthate allocation below ground. The unexpected observation that the proportion of ECM fungi does not increase during this season might indicate that root-supplied C can be used by both ECM fungi and by saprotrophs in the soil.
In this study, we have demonstrated that the fungal community in a temperate forest soil is very dynamic, showing significant seasonal changes in activity, biomass content, composition and relative abundance of different fungal groups. The results showed that the litter community exhibited seasonal changes in composition, whereas the mineral soil responded rather by changes in fungal abundance. ECM and saprotrophic fungi were indicated as the major players in this respect. Both litter decomposition and photosynthate allocation represent important factors that contribute to the observed seasonal changes. The study was limited to the upper part of the soil profile, which may represent a limitation, because it is well established that fungal communities are stratified even much more deeply below ground (Rosling et al., 2003). Conclusions with regard to the composition of the whole fungal community can be made, however, as deeper in the soil fungal biomass continues to decrease rapidly (Šnajdr et al., 2008) and the bulk of the community is thus contained within the depth analysed. To achieve a deeper understanding of the seasonal transitions in fungal community functions, it would be necessary, however, to complement the current data with a functional analysis of metatranscriptomes or metaproteomes that would answer important questions about seasonal changes in the physiology of individual taxa, including the extent of mycorrhizal saprotrophy. This study also shows that our understanding of the fungal community composition in those ecosystems in which environmental factors show seasonal variation is limited if this phenomenon is not considered.
This work was supported by the Grant Agency of Charles University (445111), by the Czech Science Foundation (P504/12/0709) and by the research concept of the Institute of Microbiology of the Academy of Sciences of the Czech Republic, v.v.i. (RVO61388971). The authors would also like to acknowledge the Editor and the Reviewers of this paper for their valuable comments that helped to improve the original manuscript.