The use of plant, bacterial and fungal resources in soil food webs of ecto- and arbuscular mycorrhiza-dominated deciduous forests
Data files
Jul 17, 2025 version files 211.20 KB
-
training_data.xlsx
30.32 KB
-
consumer_MixSIAR_4EAA_norm.xlsx
23.52 KB
-
TP_measurements.xlsx
123.52 KB
-
Training_4EAA_MixSIAR.xlsx
11.60 KB
-
AA_discrimination_4EAA.xlsx
11.18 KB
-
README.md
11.05 KB
Abstract
Soil food webs, driven by complex interactions among plants, microbes, and invertebrates, are crucial for carbon and nutrient cycling in forest ecosystems. Over the last decade, it has become evident that forests dominated by arbuscular mycorrhizal fungi (AMF) or ectomycorrhizal fungi (EMF) differ in their litter chemistry and microbial community composition, leading to different carbon and nutrient cycling. Still, the role of soil invertebrates in soil food webs of AMF- and EMF-dominated forests remains undescribed. Here, we tested whether trophic positions (TP) and basal resources of nine soil invertebrate groups (Araneae, Chilopoda, Collembola, Diplopoda, Diptera, Isopoda, Lumbricina, Mesostigmata, and Oribatida) are different in AMF- (maple and/or ash) and EMF-dominated (beech and/or oak) deciduous forests. We addressed this question using compound-specific stable isotope analysis of carbon and nitrogen in amino acids (CSIA-AA) – a novel method that allows precise estimation of the TP and basal resource use (bacteria, fungi, plants) of consumers in field conditions. Plant- and microbial-derived resources of soil invertebrates generally varied little between AMF and EMF-dominated forests. We only found a slight tendency of soil invertebrates in AMF-dominated forests relying more on fungal-derived resources (ca. 2% difference) and more on bacteria-derived resources in EMF-dominated systems (ca. 2% difference). Most of the variation in resource use was explained by the taxonomic identity of invertebrates, independently of the forest type and location. Mesofauna, particularly Collembola and Oribatida, were key consumers of fungi (relative contribution 49-72.1%), while macrofauna (except earthworms) relied more on plant resources (relative contribution 42.1-58.9%), suggesting that different size compartments of soil food webs predominantly use different basal resources. Earthworms and Myriapoda (Chilopoda and Diplopoda) had the highest proportions of bacteria-derived essential amino acids (relative contribution 23.2-35.2%). TPs of taxa remained largely consistent across mycorrhizal types (except for an elevated TP of Isopoda in EMF-dominated forests), suggesting similar vertical organisation of soil food webs across the studied forests. Overall, our study shows that TPs and basal resource use of soil fauna vary little across temperate deciduous forests, being mainly explained by the faunal taxonomic identity.
Dataset DOI: 10.5061/dryad.0rxwdbsd1
Description of the data and file structure
Data was collected to analyze the use of different basal resources and trophic position of nine soil invertebrate species using compound-specific isotope analysis of amino acids
Datasets included:
(1) AA_discrimination_4EAA
(2) consumer_MixSIAR-4EAA_norm
(3) Training_4EAA_MixSIAR
(4) training_data
(5) TP_measurements
Files and variables
(1) AA_discrimination_4EAA.xlsx
Description: We assume direct routing, so all discrimination factors were set to 0 (see Manlick & Newsome, 2022). A file is needed for mixing models.
Variables:
source - Basal resources of bacteria, fungi, and plants
MeanLeu - Mean discrimination factor for leucine
SDLeu - Standard deviations of the discrimination for leucine
MeanMet - Mean discrimination factor for methionine
SDMet - Standard deviations of the discrimination for methionine
MeanPhe - Mean discrimination factor for phenylalanine
SDPhe - Standard deviations of the discrimination for phenylalanine
MeanThr - Mean discrimination factor for threonine
SDThr - Standard deviations of the discrimination for threonine
(2) consumer_MixSIAR_4EAA_norm.xlsx
Description: δ13C values normalized via the central mean of eAAs of essential amino acids from soil animal communities to run the mixing models.
Variables:
SampleID - Running number used during the extraction of amino acids
AnimalGroup - Names of animal groups
Site - Site ID
MycType - Associated mycorrhizal type (AMF: arbuscular mycorrhiza; EMF: ectomycorrhiza)
Leu - Mean-standardized δ13C values of leucine (Leu)
Met - Mean-standardized δ13C values of methionine (Met)
Phe - Mean-standardized δ13C values of phenylalanine (Phe)
Thr - Mean-standardized δ13C values of threonine (Thr)
(3) Training_4EAA_MixSIAR.xlsx
Description: δ13C values of training data normalized via the central mean of eAAs for basal resources of bacteria, fungi, and plants.
Variables:
source - Organism origin of δ13C values
MeanLeu - Mean δ13C values of leucine measured in bacteria, fungi, and plants
SDLeu - Standard deviations of δ13C values of leucine measured in bacteria, fungi and plants
MeanMet - Mean δ13C values of methionine measured in bacteria, fungi, and plants
SDMet - Standard deviations of δ13C values of methionine measured in bacteria, fungi and plants
MeanPhe - Mean δ13C values of phenylalanine measured in bacteria, fungi, and plants
SDPhe - Standard deviations of δ13C values of phenylalanine measured in bacteria, fungi and plants
MeanThr - Mean δ13C values of threonine measured in bacteria, fungi, and plants
SDThr - Standard deviations of δ13C values of threonine measured in bacteria, fungi and plants
n - Number of bacteria, fungi, and plants measured
(4) training_data.xlsx
Description: δ13C values of amino acids from basal resources of fungi, plants, and bacteria used as training data for the stable isotope fingerprinting (see Methods).
Variables:
Identifier.1 - Unique identifier for each sample. Data labelled with OBS are site-specific litter data derived for this study. All other training data stem from Larsen et al. (2013, 2016), Pollierer et al. (2020), and Junggebauer et al. (2025)
Ile - δ13C values of isoleucine (Ile)
Leu - δ13C values of leucine (Leu)
Met - δ13C values of leucine (Leu)
Phe - δ13C values of leucine (Leu)
Thr - δ13C values of leucine (Leu)
Val - δ13C values of valine (Val)
Plot - Identifier for the sampling site
Month - The type of organism amino acids were extracted from
Species - Organisms from which amino acids were extracted. Note that algae, coniferous trees, herbaceous plants, and mycorrhiza were excluded from the analysis.
(5) TP_measurements.xlsx
Description: δ15N values of amino acids measured using a Thermo Trace gas-chromatograph connected through a GP interface to a DeltaPlus mass spectrometer (Thermo, Bremen, Germany), which was equipped with an Agilent J&W VF-35ms GC column (30m x 0.32mm, film thickness 1 µm). Nitrogen isotope ratio of AAs was expressed relative to atmospheric N by normalizing measured values (vs. reference gas) using scales derived from known δ15N values (versus atmospheric N) in AAs of derivatized standard mixtures. The data in this file was used to produce Figure 3.
Variables:
SampleID - Running number used during the extraction of amino acids
AnimalGroup - Names of animal groups
Site - Site ID
MycType - Associated mycorrhizal type (AMF: arbuscular mycorrhiza; EMF: ectomycorrhiza)
glx - Mean stable nitrogen isotope ratio of glutamic acid (glx)
SDglx - Standard deviation of isotope ratio of glx
phe - Mean stable nitrogen isotope ratio of phenylalanine (phe)
SDphe - Standard deviation of isotope ratio of phe
σ² glx - Squared mean 15N/14N glx
σ² phe - Squared mean 15N/14N phe
β - Mean β-values measured from site-specific leaf litter samples
{} - (Mean 15N/14N glx(Mean 15N/14N glx-Mean 15N/14N phe+ß))², used to calculate the uncertainty of TP
σ TP - Uncertainty of TP using the trophic discrimination factor (TDF = 7.6) as follows: (1/7.6)² ~ 0.0173: SQRT(0.0173σ² glx + 0.0173σ² phe + 0.0173σ² ß + {}*(σ²TDF)).
σ² TP - Squared uncertainty of TP
TPNew - Trophic position, see method section for formula
Code/software
(copy from “Data analyses” part in the manuscript)
All statistical analyses (Table 1) were performed in R v.4.3.0 (R Core Team, 2023). To analyse the effect of soil invertebrate group and mycorrhizal type on the use of different basal resources (bacteria, fungi, plants), we ran a multivariate analysis of variance (MANOVA) on the linear discriminant analysis (LDA) classification, which was based on the δ13C values of the eAAs including Leu, Phe, Met, and Thr (Figure S1). To estimate the relative contribution of bacteria, fungi, and plants to the diet of consumers, Bayesian mixing models were used based on eAAs δ¹³C values centred to the mean value of all eAAs in a sample (“MixSIAR” package; Stock et al., 2018). The normalized δ¹³C values of Ile and Val in consumers fell outside the range observed in resources and thus beyond the bounds of the mixing space, leading to their exclusion from the mixing model (Figure S2). These deviations may result from analytical challenges, such as co-elution with nearby peaks (Ile) and instability in early chromatographic regions (Val) - or from metabolic processes, as both amino acids share similar pathways (Vane et al. 2025). Instead, we focused on the four eAAs: Leu, Thr, Phe, and Met. For the classification of basal resource contribution, we used training data from Larsen et al., (2009, 2016) and Pollierer et al., (2020). We further supplemented the training data for plants with site-specific litter values from our experimental sites and freshly fallen beech leaves from different regions across Germany (Junggebauer et al., 2025). While bacterial training data included only lab-derived, and plant training data only field-derived endmembers, fungal training data included both lab- and field-derived fungal samples to cover the full variation of fungal signatures (see sensitivity analysis, Figure S4). Statistical inferences regarding the estimated relative contributions of resources between AMF- and EMF-dominated forests were based on the Bayesian approach followed by visual interpretation of the mixing model output and its credible intervals (Figure S3). Therefore, the observed patterns should be regarded as tentative descriptive trends rather than statistically significant effects.
TP of soil invertebrates was derived from δ¹⁵N values of Glu and Phe *averaged across the triplicate measurements for each sample, with β representing the specific and averaged differences in δ¹⁵N of *Glu and Phe in primary producers. For each TP calculation, we used a site-specific β value to better address the issue of potential background variability (Steffan et al., 2013). For the trophic discrimination factor (TDF), we used the value of 7.6 ± 1.2‰, following the recommendations by Chikaraishi et al. (2014). The TP of each animal group was calculated separately per location and mycorrhizal type using the following equation:
The relationships between TP, animal group and mycorrhizal type were analysed using linear mixed-effect models (*lmer *function of the “lme4” package, Bates et al., 2015) with animal identity and mycorrhizal type as predictors, and site nested within location as random effect. To assess the impact of mycorrhizal type on the TP of soil invertebrate groups, we conducted contrasts to examine differences in their TP between the two mycorrhizal types using the “emmeans” package (Lenth, 2024). The contrasts represented the differences in estimated marginal means of TP for the different animal groups within AMF- and EMF- dominated systems. Mean values and standard deviations are reported in the main text, with a p-value of 0.05 set as the threshold for statistical significance.
(Information added to Dryad)
More detailed information about the Bayesian model can be found here: https://github.com/brianstock/MixSIAR
Access information
Training data were derived from the following sources (see manuscript for details):
-
Larsen, T., Lee Taylor, D., Leigh, M. B., & O’Brien, D. M. (2009). Stable isotope fingerprinting: A novel method for identifying plant, fungal, or bacterial origins of amino acids. Ecology, 90(12), 3526–3535. https://doi.org/10.1890/08-1695.1
- Larsen, T., Pollierer, M. M., Holmstrup, M., D’Annibale, A., Maraldo, K., Andersen, N., & Eriksen, J. (2016). Substantial nutritional contribution of bacterial amino acids to earthworms and enchytraeids: A case study from organic grasslands. Soil Biology and Biochemistry, 99, 21–27. https://doi.org/10.1016/j.soilbio.2016.03.018
- Pollierer, M. M., Scheu, S., & Tiunov, A. V. (2020). Isotope analyses of amino acids in fungi and fungal feeding Diptera larvae allow differentiating ectomycorrhizal and saprotrophic fungi-based food chains. Functional Ecology, 34(11), 2375–2388. https://doi.org/10.1111/1365-2435.13654
- Junggebauer, A., Jüds, M., Klarner, B., Dyckmans, J., Pollierer, M.M., Scheu, S. (2025). Fungal energy channeling sustains soil animal communities across forest types and regions. Ecology Letters, 28(5), e70122. https://doi.org/10.1111/ele.70122