Arbuscular mycorrhizal fungal communities shift their foraging functional traits in response to grassland fertilisation

Terrestrial plants obtain carbon (C) largely via photosynthesis but acquire other resources, including nitrogen (N) and phosphorus (P), through their roots and fungal mycorrhizal symbionts. As these resources are not uniformly distributed throughout the soil, roots have evolved a foraging ability to locate patches with higher nutrient availability (Cahill & Mcnickle, 2011), and the hyphae of mycorrhizal fungi also forage for nutrients (Chen et al., 2018). Here, we determine how the average foraging traits of a community of arbuscular mycorrhizal fungi (AMF) in managed grassland adjust to long-term fertilisation, and we present novel findings based on combining results from two field experiments. Plant ecologists characterise resource foraging by roots using three parameters: foraging precision is the ability to preferentially grow into nutrient-rich patches; foraging speed shows how quickly roots arrive at a new patch; and foraging scale is the size of soil volume readily available for root exploitation (Chen et al., 2018). The ability to target nutrient-rich patches is important for plants in nutrient-poor soil (Fitter et al., 1991), and its utility diminishes as the patch size grows (Wang et al., 2022), reaching a state in which the soil has a high, mostly homogeneous nutrient availability, such as in fertilised agricultural land. Foraging precision of roots is partly phylogenetically conserved (Kembel & Cahill Jr., 2005; Wang et al., 2022), and forbs have a higher precision than C3 and C4 grasses (Johnson & Biondini, 2001; Wang et al., 2022). In a global plant trait framework, variation in plant foraging traits mostly contributes to trait gradients of leaf/root conservation and plant size (Weigelt et al., 2021). Another important gradient in this framework is the ‘root collaboration gradient’, which shows roots' suitability for symbiosis with mycorrhizal fungi (Weigelt et al., 2021). Mycorrhizal fungi occur in the roots of most vascular plant species, and the most widespread among them are AMF, which form symbioses with an estimated 72% of species (Brundrett & Tedersoo, 2018). Similar to roots, mycorrhizal fungi respond to spatial heterogeneity in soil nutrients and water availability, but compared to roots, they are better able to cope with small-scale heterogeneity and compete successfully with other soil microbes (Tibbet, 2000; Cavagnaro et al., 2005; Cahill & Mcnickle, 2011). The ability of AMF hyphae to target nutrient-rich patches has been demonstrated in glasshouse experiments (Gavito & Olsson, 2003; Cavagnaro et al., 2005), but evidence from field studies is scarce and limited to entire communities of AMF and ectomycorrhizal symbionts in trees (Chen et al., 2016, 2018), with no results reported for individual fungal taxa. Thus, the foraging traits of mycorrhizal fungi should be further studied. Van der Heijden & Scheublin (2007) suggested that the functional traits of individual AMF species cannot be measured in the field, and there has been very limited progress in this area of study. Therefore, we know little about how the foraging traits of AMF vary across taxa in the field and how this variation is affected by environmental conditions. Our research group previously proposed a method for measuring two hyphal foraging traits (precision and speed) and estimated them for a large subset of AMF taxa present in a managed grassland (Šmilauer et al., 2020; Košnar et al., 2025). In the same grassland, we examined the effects of fertilisation and the long-term removal of host plant functional groups on AMF communities (Šmilauer et al., 2025b). Using two complementary datasets collected from that long-term field experiment, this study aimed to determine how the average foraging precision and speed of AMF communities shift due to fertilisation. We also compared AMF foraging traits between the functional groups of host plant species (grasses and dicotyledonous forbs) and among plots with prevailing grasses, prevailing forbs, or their mixture. The data synthesised in this study were collected on a long-term research site located near Zvíkov village, Czech Republic (48°59′20″N, 14°36′28″E), at 500 m a.s.l. The vegetation is a managed oligotrophic grassland, in which no fertiliser has been applied in the past 40 yr. Macronutrient availability is low to medium (total N, 0.38% w/w; N in NO3− form, 0.81 mg kg−1 DW soil; N in NH4+ form, 19.2 mg kg−1 DW soil; P in PO43− form, 7.3 mg kg−1 DW soil), and the soil (Cambisol type) is acidic (pH 5.6), with a mostly sandy texture. There are 85 vascular plant species recorded in this site, and among them, 16 abundant forb and grass species with AM symbiosis were used in this study (Supporting Information Table S1), representing c. 67% of the aboveground biomass of the grassland. Data from two field experiments (published elsewhere) were combined. The removal experiment manipulated the vegetation by selectively weeding the plant functional groups (dicotyledonous forbs or grasses) for 20+ yr, and this removal was combined with fertilisation in the 4 yr preceding data collection. We characterised the AMF community and the effects of experimental manipulation by sampling 16 plant species (dataset D1) and soil cores (dataset D2) from all experimental plots, and published this research in Šmilauer et al. (2025b). A short-term experiment examined the response of AMF hyphae to vacant experimental soil patches with manipulated nutrient availability and estimated two hyphal foraging traits (Košnar et al., 2025). Combining the information obtained from original soil in places where ingrowth patches were established with the data from ingrowth patches after 6 or 12 wk of exposure, we estimated foraging speed, and by comparing nutrient-enriched and control patches, we estimated foraging precision. That study was performed in a nutrient-poor, nonfertilised area of the study site. In this study, we combined the results from both experiments: using the information on AMF community changes with plant community modification and fertilisation (datasets D1 and D2) and knowing foraging traits of a large set of AMF taxa, we calculated community-weighted means (CWMs) of foraging traits (using relative abundances of AMF taxa). Here we examine the responses of trait CWMs to experimental treatments. Additional details about the experiments and data collection are provided in Methods S1. The DNA isolated from plant roots and rhizosphere soil was amplicon-sequenced based on a semi-nested polymerase chain reaction amplification of a 550-bp fragment of the SSU rDNA, using the NS31–AML2 primer pair, followed by the WANDA–AML2 primer pair. The resulting pools of barcoded AMF amplicons were analysed using Illumina paired-end 2 × 300 bp sequencing. After the resulting DNA sequences were quality filtered, the identified AMF sequences were assigned to virtual taxa (VTXs) using the MaarjAM database (Öpik et al., 2010), and up to 400 sequences were chosen for individual samples to achieve a comparable sampling effort. Additional details about the molecular analyses can be found in Methods S2. We evaluated the change in the community averages of foraging traits using linear mixed effect models. For dataset D1 (sampled individual host plants of 16 species), we estimated the effects of fertilisation, host plant functional group, and their interaction as fixed effect terms. Random effects reflected the experimental design and variation among individual host species. For dataset D2 (sampled roots and soil at the plant community level), we tested the effects of fertilisation, plot type (forb-dominated, grass-dominated, or with both functional groups), and their interaction jointly for root and soil samples and separately for these sample types. Additional details about estimating foraging traits, calculating their community means, and statistical modelling are presented in Methods S3. Across datasets D1 and D2, originating from the long-term removal experiment, we identified 108 AMF VTXs, the majority of which were assigned to the Glomeraceae family, but six other AMF families were also present (Table S2). From the ingrowth patch experiment (Košnar et al., 2025), we obtained foraging speed estimates for 48 VTXs, representing 97.8% of the total read count in D1 and 96.4% in D2, and foraging precision estimates for 28 VTXs, representing 89.2% of the read count in D1 and 75.8% in D2. Fig. S1 summarizes the distribution of these proportions across the main categories of experimental treatments. With such a high coverage and no extreme trait values, the estimated trait averages were assumed to be unbiased (Pakeman & Quested, 2007). The distribution of foraging speed and precision estimates for AMF VTXs is shown in Fig. S2. The effects of fertilisation on the average AMF traits are summarised in the upper part of Fig. 1, with additional details in Tables S3 and S4. The foraging speed increased significantly due to fertilisation in the D1 dataset and both for root and soil samples in the D2 dataset. Foraging precision declined with fertilisation, with the estimate larger in soil compared to roots. The directions of both changes (Figs 1 and 2) matched our expectation that fertilisation creates a homogeneous, nutrient-rich soil environment in which ruderal AMF taxa with faster hyphal growth (Chagnon et al., 2013) but lower foraging precision are favoured. This suggests that sufficient N and P availability to plants suppresses AMF taxa, which are effectively exploring nutrient resources through extensive extraradical hyphae, due to a lower C provision by the host (Johnson, 2010; Konvalinková et al., 2017). When comparing the size of fertilisation effects between the roots of sampled plant species and the roots from cores (Fig. 1), the weaker effects on foraging speed estimated for roots from cores were likely caused by a large turnover of plant species present in the cores and possible bias due to the selection of dominant plant species for individual plant sampling. The large effect of fertilisation on foraging precision in soil samples (Fig. 1; Table S4) suggests more extensive changes in the abundance of extraradical hyphae; AMF taxa with effective acquisition of soil nutrients are known to invest more C into extraradical than into intraradical hyphae (Chagnon et al., 2013). Fertilisation also decreased (Šmilauer et al., 2025b) the average AMF colonisation of roots (by 58%) and AMF α-diversity (by 14%). Therefore, the increase in foraging speed with fertilisation reported here may be partly explained by the loss of the VTXs with lower foraging speed and the decrease in foraging precision by the disappearance of taxa with higher foraging precision. Under longer-term fertilisation, both AMF and their host coadapt their symbiotic relationship to higher nutrient availability (Johnson, 2010). This adaptation may be reflected in modified foraging trait values, but our results cannot show such changes, as we quantified the traits (Košnar et al., 2025) in a community without long-term fertilisation, so we do not have the data needed to account for adaptive changes of the traits in our analyses. In dataset D1, the estimated difference of AMF foraging speed between grasses and forbs is large and significant (Fig. 1). Symbionts of grasses seem to rely more on quick arrival to the unoccupied space (Table S3), similar to the roots of their hosts (Šmilauerová & Šmilauer, 2010). This supports the suggestion of Chagnon et al. (2013), who posited that the life strategies of host plants and their symbionts might be correlated. Forbs and grasses showed no differences in the response of AMF foraging speed and precision to fertilisation (Table S3). We found large, significant variation in both AMF foraging traits among individual host plant species within their functional groups (Table S3; Fig. S3). Root-based AMF had significantly lower foraging precision in grass-dominated plots than in mixed plots, but this trend was not present in the soil samples (Fig. 1) or supported by dataset D1, where the effect of functional group on foraging precision was absent (Fig. 1). Our results showed for a managed grassland that the compositional change of AMF communities caused by fertilisation resulted in a substantial shift of foraging strategies. This was complemented by differences in AMF foraging traits between two important plant functional groups – grasses and forbs – but the AMF traits of both host groups responded similar to fertilisation. Our study demonstrates the feasibility and informativeness of quantifying functional traits of mycorrhizal symbionts under field conditions. This research was funded by a grant from the Czech Science Foundation (GAČR 22-09594S). All authors are thankful to Dr Eva Holá for molecular analyses and to Blanka Divišová for processing field samples. The authors are also thankful to the reviewers and the manuscript editor for their careful review and many good suggestions. The manuscript text was revised by Proof-reading-service.com Ltd. (Cambridge, UK). None declared. PS, MS and JK designed the study. PS and MS performed the field work and participated in the processing of the plant and soil material in the ecological laboratory. JK coordinated the analyses in the molecular laboratory and developed novel bioinformatic approaches. PS and JK analysed the data and wrote the manuscript. All authors contributed to data interpretation and approved the manuscript. Data from Illumina sequencing and all the additional datasets used in this manuscript are accessible on zenodo.org platform (Šmilauer et al., 2025a). DNA sequences are available on GenBank under the accessions PQ763233–PQ763354. Fig. S1 Coverage of AMF community composition by foraging traits. Fig. S2 Distribution of foraging traits of virtual AMF taxa. Fig. S3 Distribution of average foraging traits for individual host plant species. Methods S1 Details of experimental design and data sampling. Methods S2 Details of molecular analyses. Methods S3 Details of data manipulation and statistical analyses. Table S1 Studied host plant species (in D1 dataset). Table S2 List of AMF virtual taxa in the removal – fertilisation dataset. Table S3 Effects of fertilisation and host plant on foraging traits in dataset D1. Table S4 Effects of fertilisation and plot type on foraging traits in dataset D2. Please note: Wiley is not responsible for the content or functionality of any Supporting Information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article. The New Phytologist Foundation remains neutral with regard to jurisdictional claims in maps and in any institutional affiliations.