Fauna-microbe diversity coupling lost in agricultural soils: Implications from the bacteria hidden in earthworm gut

Abstract

Earthworms are critical for supporting soil health and microbial diversity and simultaneously maintaining a highly diverse gut microbiome. The earthworm is predominantly vulnerable to physical disturbance, yet how changes in earthworm diversity influence the richness and ecological network of soil-gut microbiomes in response to anthropogenic disturbance is virtually unknown. Here, we investigated the richness of earthworms, and their connection with the diversity of the soil-gut microbiome using a large-scale survey covering paired agricultural and natural sites. Our results showed that earthworm diversity was positively correlated with soil and gut bacterial diversity across sites. However, the connection between soil bacterial and earthworm diversity is lost in agricultural ecosystems. We further show that earthworm richness supported greater modularity in microbial networks, being both positively correlated with the richness of earthworm gut bacteria in both land-use types. Together, we provided the first empirical evidence that agricultural practices can break the fundamental links between soil bacterial and earthworm diversity, and further identify an unreported consistent connection between the diversity of earthworms and the modularity of microbial networks in natural and managed ecosystems. These findings emphasize the primary roles of earthworms in supporting soil biodiversity and point to the wider contributions of the soil animal-microbe interactions in preserving the whole soil biodiversity in anthropogenically disturbed ecosystems.

Graphical abstract

1. Introduction

Anthropogenic disturbances account for more than 50% of the human impacts on global biodiversity decline, with land-use changes ranking as the greatest driver of soil biota [1]. Given that the soil ecosystem underpins human well-being by providing food and fibers, it has been a particular issue to concern the effects of land-use changes on soil biota [2]. Earthworms – the soil ecosystem engineers – are emblematic of soil-dwelling macrofauna that play a key role in maintaining terrestrial ecosystem functions and services [3]. They are also highly vulnerable to environmental changes and disturbances, making them promising bioindicators, for instance, to assess the effects of land use on soil ecosystems [4]. Indeed, it is demonstrated that the diversity, biomass, and function of earthworms could be negatively affected by agriculture. However, since earthworms can regulate the soil microbiota implicitly or explicitly through feeding and burrowing [5], independently assessing changes in earthworm or the soil microbial communities is insufficient to reveal the land-use effects on earthworm-microbe co-occurring soil ecosystems.

Earthworms carry a vast microbial diversity in their guts, which is mainly derived from the soil microbial communities they ingest and discharge [6], making the earthworm gut an important regulator of the soil microbial community assembly and further the soil functions and ecosystem services [7]. For example, keystone taxa detected by greater connectivities within the network of earthworm gut microbiota contributed to carbon metabolisms and potentially the carbon cycling in soil [8]. Therefore, any changes in the earthworm community can generate cascade impacts on soil microbial communities. For example, the land-use change reduced the richness of decomposers at higher trophic levels, which further altered the stabilization of soil biota [9,10]. However, the underlying processes and mechanisms of the interactions between earthworms and soil microbes are elusive. Moreover, the land-use change could make the exploration of the relationships more difficult, as land use not only directly affects the earthworm and microbial communities but also shapes their interactions. Understanding the effects of agriculture on earthworms, microbes, and their interactions is not only critical for developing strategies to engineer soil biodiversity for agricultural sustainability, but it is also highly relevant to predicting the responses of soil biodiversity to agriculture in the future [11].

Gut microbiota provides a new way to think about the effects of land-use change on the whole soil microbial community. Firstly, the microbial taxa could be different between guts and soils [12]. Assessing differences between bulk soil microbial communities, therefore, provides comprehensive information on the impacts of agriculture on soil microbes inhabited in different microhabitats [13]. In addition, the identity, diversity, and biomass of earthworms could also mirror the agricultural activities on microbial community composition in the earthworm gut, as the gut microbiota is sheltered from the intestine of earthworms. With the movement of earthworms, the gut microbes can be migrated to a relatively far place, and the selective feeding of earthworms can also pose a selection on soil microbial community composition [14]. Thus, the reduction of earthworm diversity derived from land-use changes (e.g., agriculture) would induce a loss of gut microbial diversity, reflecting harm in the exchange of microbes between soil and earthworm gut.

Intra-group interactions between microbial species are important drivers in maintaining microbial diversity [15]. The cross-domain network is an emerging means to elucidate the interactions of the complex and diverse microbial communities [16]. The animal host is closely associated with their affiliated microbiome, with the latter being essential to the host's health and adaptability [17,18]. The potential increase of the module of a network might benefit the interactions between soil biota as well as the efficiency of soil nutrient cycling [19,20]. The co-occurrence patterns illustrated by a network capture important information on keystone species in a community [21], although they are not equal to real ecological interactions [22]. The network topology emerges as a reliable reflection of microbial community dynamics [23]. Among them, the modularity of a network is a key feature for inferring the complexity of networks [24], as microbial activities including metabolic and regulatory networks within organisms to biofilm formation, and food-web interactions are modular [25]. A more modular network is expected to support a larger number of microbial clusters including taxa strongly co-occurring with each other. For instance, the modularity of root microbiota shifted in response to tolerate the pathogen disturbance [26]. Further, biodiversity increases with network modularity [19], which provides knowledge to understand the assembly of microbial communities. Soil-foraging animals were reported to alter the co-occurrence of soil microbial communities in arid ecosystems [27]. It remains unclear, however, whether microbial network architectures are associated with the reduction of earthworm biodiversity in agricultural systems.

Here, we aimed to understand the patterns of the diversity, composition, and interactions of three interconnected biological components i.e., earthworms, their gut microbiome, and soil microbiomes, and to explore their responses to anthropogenic land-use change. For that, we conducted a field survey to investigate the distribution and composition of earthworms, gut, and soil microbiomes covering both natural and agricultural ecosystems from China over a large spatial scale. We hypothesized that (1) due to the reduction of earthworms [28], agricultural practices may result in a loose connection of microbial communities in the soil and earthworm gut; (2) the sites with greater earthworm richness might exhibit more soil microbial diversity and simultaneously a larger degree of network modularity due to greater microbial interactions [19].

2. Materials and methods

2.1. Soil and earthworm sampling

We selected the Yangtze Plain as our research area as it is an important and traditional farming area in China. Ten sites that vary in climate and soil conditions were chosen. These sites extend for about 546.9 km from east to west and 657.1 km from north to south. Four natural and four agricultural plots were established within each site, resulting in 80 plots in total (Fig. S1). Agricultural sites were those used for wheat (Triticum aestivum L.) cropping continuously for more than 20 years until the sampling time, and natural sites were permanently covered by natural forests, e.g. Populus spp. and Pinus spp. The plots that are of similar habitats from one site were separated by a minimum distance of 3 km. We obtained soil and earthworm samples from these plots in May 2020. At each sampling site, eight soil cores (6.25 cm diameter × 20 cm depth) were collected and mixed as a composite sample. Simultaneously, earthworm individuals were collected by hand-sorting equipped with the electrical method. In brief, 1 m² was measured on the bare soil surface with vegetation removed. The 2 steel electrodes were pushed into the soil to at least 15 cm depth and connected to an AC generator (model RWF1, Elektrotechnik Schuller, Germany), which was connected to a 12 V battery. The electric current and voltage from the generator were increased stepwise (every four minutes) from 0.2 A, 46 V to 1 A, 230 V, and emerging earthworms on the soil surface were collected within 20 min [29]. The earthworms were kept in soil when transported to the lab and then preserved in anhydrous alcohol at 4 °C before identification.

2.2. Soil and climate properties

The soil texture, pH, moisture, organic carbon (SOC), and total nitrogen (TN) of soils were measured as soil abiotic variables. A subsample soil (10 g) sieved through a 2.0 mm sieve was fractionated into sand (particle size, 20–2000 μm), silt (2–20 μm), and clay (< 2 μm) using the ultrasonic energy method [30]. All results of particle size analysis were expressed as the percentage of the total weight of the oven-dried soil. Soil pH was measured using a pH meter (PHS-3C., Shanghai Leici) after shaking (soil-water suspension, 1:5 w/v) for 30 min. The soil moisture content was determined using 10 g of fresh soil dried at 105 °C for 48 h to constant weight. SOC and TN were measured using an elemental analyzer (Elemental Analyzer System Vario Macro Cube, Langenselbold, Germany). For each sampling site, we extracted mean annual precipitation (MAP) and temperature (MAT) from WorldClim version 2.1 (http://www.worldclim.org/) at a spatial resolution of 30 s [31].

2.3. Earthworm identification and gut content isolation

For each earthworm specimen, the tail muscle tissue with 2–3 segments (∼25 mg) was scissored and used for total genomic DNA extraction using the E.Z.N.A. Mollusc DNA Kit (Omega Bio-tek, Norcross, GA, USA). Mitochondrial DNA (mtDNA) cytochrome c oxidase subunits I (COI) gene were amplified, which was the standard DNA barcode for earthworm species identification [32]. The primers and PCR conditions were described in ref. [33]. The morphologic characteristics of each specimen were detected with a stereo microscope according to ref. [34] and identified to species level complemented with DNA barcoding. Simultaneously, the gut content of each earthworm specimen was isolated by dissection. An incision was made longitudinally along the body wall and the digestive tracts taken behind the clitellum to the anus were collected in a 2.0 ml Eppendorf tube and stored at −20 °C before use.

2.4. DNA extraction and Illumina sequencing

Total genomic DNA of the gut content from earthworm individuals as well as soils in the present study was extracted using the FastDNA Spin Kit for Soil and the FastPrep Instrument (MP Biomedicals, Santa Ana, CA, USA) following the manufacturer's instructions. Before sequencing preparation, the quality and quantity of the extracted DNA were certified with 1% agarose gel electrophoresis and a Nanodrop-2000 spectrophotometer (NanoDrop Technologies Inc. Wilmington, DE, USA), respectively. The V4 hypervariable region of the bacterial 16S rRNA gene was amplified with the primers (FWD: 5′-GTGYCAGCMGCCGCGGTAA-3′; REV: 5′-GGACTACNVGGGTWTCTAAT-3′) and the amplification protocols designed in the EMP protocol (https://earthmicrobiome.org/protocols-and-standards/16s/). The PCR amplification was performed as follows: initial denaturation at 95 ℃ for 3 min, followed by 27 cycles of denaturing at 95 ℃ for 30 s, annealing at 55 ℃ for 30 s, and extension at 72 ℃ for 30 s, and a single extension at 72 ℃ for 10 min, and end at 4 ℃. The PCR mixtures contain 5 × TransStart FastPfu buffer 4 μL, 2.5 mM dNTPs 2 μL, forward primer (5 μM) 0.8 μL, reverse primer (5 μM) 0.8 μL, TransStart FastPfu DNA Polymerase 0.4 μL, template DNA 10 ng, and finally ddH₂O up to 20 μL. The PCR reactions were performed in triplicate and with a negative control replacing template DNA with ddH₂O. The PCR products were then extracted from 2% agarose gel and purified using the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, USA) according to the manufacturer's instructions and quantified using Quantus™ Fluorometer (Promega, USA). Purified amplicons were pooled in equimolar and sequenced on an Illumina MiSeq PE 300 platform (Illumina, San Diego, USA). The raw sequences were deposited in NCBI Sequence Read Archive under the accession number PRJNA400302.

2.5. Sequence data processing

Sequence data were processed using the UPARSE-UNOISE3 pipeline implemented in USEARCH v11, which showed the best balance between resolution and specificity in 16S rRNA amplicon data analysis among current analysis pipelines [35]. Paired-end reads were merged using the command -fastq_mergepairs with the parameter of 10 maximum differences (-fastq_maxdiffs 10) in sequences. For quality filtering, we used a maxee value of 1.0, which indicated that on average a maximum of 1.0 nucleotides were incorrectly assigned in every sequence using the command -fastq_filter. Then, sequences were dereplicated and singleton sequences were removed before phylotype determinations using the command -fastx_uniques. After that, denoised representative sequences were obtained using the command -unoise3. For further phylogenetic analysis, a phylogenetic tree was constructed for the representative sequences firstly by multiple sequences aligning using MUSCLE v5, and then a Maximum Likelihood tree was generated using FastTree2 with default settings [36]. The representative sequences were annotated according to the SILVA v138 database [37] using the command -syntax and an ASV-level table defining the reads in each sample was constructed using the command -otutab. The table was resampled to a minimum number of 15,700 sequences per sample for both soil and gut microbiota dataset in further downstream statistical analysis.

2.6. Statistical analysis

Statistical analyses were performed in R 4.1.2 otherwise stated elsewhere [38]. The richness of earthworms, soil microbes, and gut microbes was estimated using the command specnumber implemented in the vegan package [39]. For the analyses of land-use type on the richness of earthworms, soil microbes and gut microbes, we used linear and generalized linear mixed-effects models (LMMs), fit with maximum likelihood (lme4 package; [40]), by setting the land-use type as the fixed effect, the sampling site as the random effect. To test the relationships between the richness of earthworms, soil microbes, and gut microbes, we fitted a linear and non-linear regression model to these factors using the function lm implemented in the base package. The phylogenetic trees were visualized using the function ggtree implemented in the package ggtree [41]. To determine the bacterial taxa that are significantly different by relative abundances between soils and the earthworm gut, i.e., taxa that are enriched or depleted in the earthworm guts, the aldex function implemented in the ALDEx2 package was used [42]. Taxa with differences tested by both Wilcoxon's rank-sum tests and a Benjamini–Hochberg false discovery rate (FDR) correction with significantly (P < 0.05) different relative abundances were retained [43].

To precisely quantify interactions between earthworms and microbes, we applied bipartite networks, which were derived from physical science and demonstrated to understand plant-microbe interactions [44,45]. For cross-domain network construction, we concatenated the species table of earthworms, and the ASV-level table of both soil and gut microbes and then calculated pairwise correlations using the script sparcc implemented in mother v.1.35.0 [46]. We chose the SparCC (Sparse Correlations for Compositional data) method for network construction, as this algorithm is particularly designed to deal with compositional datasets with lots of zeroes, such as microbial community composition data [47]. Significant (P < 0.01) and strong (R > 0.7) correlations and ASVs with an abundance of more than 0.1% were considered in the visualization using the Fruchterman–Reingold layout algorithm implemented in the igraph package [48]. Topological properties including the number of nodes and edges, average degrees, i.e. the mean number of edges of each node, and modularity, were used to measure the complexity of networks.

The stability of networks including robustness, vulnerability, and compositional stability was calculated according to the pipeline suggested in [23] with some modifications. Briefly, the robustness of a network was defined as the proportion of the remaining species in this network after random (Random robustness) or targeted (Targeted robustness) removal of nodes. For simulations of random robustness, nodes in the cross-domain networks were randomly removed. For targeted robustness, we sequentially removed the earthworm nodes to quantify the effects of earthworms on the stability of cross-domain networks. The vulnerability of a network is indicated by the maximal vulnerability of nodes in the network, which measures how fast the shortest path decreased in the network. The compositional stability was defined as the differences between networks of earthworm richness.

3. Results

3.1. Earthworm, soil, and gut bacterial richness

In total, 16 earthworm species were found in the present study. At each site, one to four earthworm species were found, and one individual of each species was used for gut content analysis. Agricultural management significantly decreased the richness of earthworms (Wilcoxon test, p = 0.033; Fig. 1a), but increased the bacterial richness in the earthworm guts (Wilcoxon test, p = 0.002; Fig. 1c). The community composition of earthworms, soil microbiota, and the earthworm gut microbiota changed with land-use types, with the microbiota of both soil and earthworm gut getting more homogenous with the increase of earthworm richness (Figs. 1d–f and S2).

Fig. 1.

Agricultural activities alter the α- and β-diversity of soil earthworms, soil bacteria, and earthworm gut bacteria. Richness of earthworms (a), soil bacteria (b), and earthworm gut bacteria (c) and community composition of earthworms (d), soil bacteria (e), and earthworm gut bacteria (f). Data in the barplot are shown as mean ± standard deviations. In d–f, ns, not significant; *, P < 0.05;**, P < 0.01; ***, P < 0.001, which were derived from the results of permutational multivariate analysis of variance using distance matrices; The ellipse corresponds to the 95% confidence interval.

The soil bacterial richness was significantly correlated with earthworm richness in natural land-use types but not for those in agricultural land-use types (Fig. 2a). In the gut, the bacterial richness was significantly correlated with earthworm richness for both agricultural and natural land-use types (R² 0.55 and 0.49 for agricultural and natural land-use types, respectively; Fig. 2b). The richness of shared bacteria between earthworm gut and soil was significantly correlated with earthworm richness for both agricultural and natural land-use types (Fig. 2c).

Fig. 2.

Relationships between bacterial richness and earthworm richness in agricultural and natural land-use types. The relationships between earthworm richness and the richness of soil (a) and gut (b) bacteria. The richness of bacteria shared by gut and soil (c) refers to the bacteria that are occurred simultaneously in the gut and soil. The R² and P values from linear regressions are shown. Dashed and solid regression lines represent non-significant (P > 0.05) and significant (P < 0.05) relationships.

The richness of specific bacteria in both the soil and earthworm gut was significantly correlated with earthworm richness in natural land-use types but not for those in agricultural land-use types (Fig. S3a and b). The richness of specific bacteria in both the earthworm gut and soil was greater than the shared bacteria (Fig. S3c and d), with the agricultural management resulting in a higher change fold. Compared with soils, there were more bacteria enriched than those depleted in the gut of earthworms (739 vs 445 taxa, respectively; Fig. S4). The enriched taxa were mostly affiliated with Acidobacteria, Firmicutes, and Planctomycetes, while the depleted taxa were mostly from the phyla Proteobacteria, Acidobacteria, and Gemmatimonadetes (Fig. S4).

Mantel tests revealed a higher effect of climate factor, i.e., mean annual temperature, was associated with the community composition of soil bacteria in both agricultural and natural sites (Fig. S5a). In the agricultural sites, the contribution of geographic distance ranked as the second important driver, whereas in the natural sites, the soil pH was more important, followed by the geographic distance of sampling sites. The earthworm phylogenetic distance was the primary factor associated with the community composition of microbiota in the earthworm gut (Fig. S5b), with the soil pH also related to the community composition of gut microbiota in the agricultural sites.

3.2. Cross-domain networks

The networks were more complex with the increasing of earthworm richness in both agricultural and natural land-use types, as the degree and modularity of the networks generally increased (Fig. 3a). The number of interactions between earthworms and soil microbes increased with earthworm richness in both agricultural and natural land-use types (Figs. S4 and S6). In the gut, the bacterial richness was significantly correlated with the network modularity for both agricultural and natural land-use types (R² = 0.58 and 0.51 for agricultural and natural land-use types, respectively; Fig. 3b). The soil bacterial richness was significantly correlated with the network modularity in natural land-use types but not for those in agricultural land-use types (Fig. 3c).

Fig. 3.

Modular networks and the relationships between network modularity and bacteria richness. Cross-domain co-occurrence networks (a) of soil bacteria, gut bacteria, and earthworms in agricultural and natural land-use types. Relationships between bacterial richness and network modularity for earthworm gut (b) and soil (c) in agricultural and natural land-use types.

The random robustness of cross-domain networks increased with earthworm richness in agricultural land-use types but decreased in natural land-use types (Fig. 4a). The targeted robustness of networks generally decreased with earthworm richness in both natural and agricultural land-use types (Fig. 4b). The vulnerability of networks decreased with earthworm richness for both agricultural and natural land-use types (Fig. 4c). Compositional stability generally increased with earthworm richness, with the coefficient in agricultural being greater than that in natural land-use types (Fig. 4d).

Fig. 4.

Network stability in agricultural and natural land-use types. (a) Robustness measured as the proportion of taxa remained with 50% of the taxa randomly removed from each of the networks. (b) Robustness is measured as the proportion of earthworm taxa removed from each of the networks. In (a) and (b), error bars correspond to the standard deviations of 100 times of simulations. Significant comparisons (two-sided t-test) between agricultural and natural habitats are indicated by ***P <  0.001. (c) Network vulnerability measured by maximum node vulnerability in each network. (d) Compositional stability of the networks. The adjusted R² and P values from linear regressions are shown.

4. Discussion

4.1. Earthworms enrich bacteria in their guts

Our study provides novel evidence that the diversity of earthworms is critical for supporting the biodiversity of the soil-gut microbiome and the modularity of microbial networks. However, we also showed that agricultural practices can break the natural connection between earthworms and soil bacterial richness, with unexplored consequences for soil function. Both soil bacterial richness and soil texture in the present study were not differed between agricultural and natural ecosystems (Fig. S8), suggesting that the bacterial richness might be more determined by the soil pH and climatic variables [49,50]. This result suggests that anthropogenic derived agricultural activities are among the overwhelming drivers of soil animal biodiversity [51], although the total bacterial diversity in agricultural sites was not different from those in natural sites, which might be due to the reason that earthworms enriched bacteria in their gut. The earthworm guts can be viewed as bioreactors, in which various biogeochemical processes are going on simultaneously. The metabolites, involving carbon, nitrogen, and phosphorus transformation, were fundamentally expressed by gut microbes to support their basic metabolism and potentially some of those are required by the earthworms [52]. In our study, we demonstrated that the main driver for the variation in the community composition of the gut microbiota was the earthworm phylogenetic distance. Thus, we infer that the earthworm gut microbiota might be assembled by the processes to provide the host with substantial metabolites that are required for earthworm fitness [53]. At the same time, this effect was amplified in agricultural soils, which was positively correlated with earthworm richness. These results suggest that more earthworm diversity leads to a more versatile microbiota in both the soil and gut, being more actively communicated. Although the total bacterial richness in agricultural sites was not different from those in natural sites, the bacterial richness in the earthworm gut was higher in agricultural than in natural sites. This might be due to the reason that more diverse bacteria formed a synergistic relationship with earthworms in agricultural sites. This was embodied in our results that the change fold was significantly higher in agricultural sites than in natural sites (Fig. S3). Moreover, we admit that some bacteria might be depleted during the earthworm gut passage, but we found that the number of bacteria that were enriched (739) in the earthworm gut was 1.7 times higher than those that were depleted (445). Thus, the earthworm gut can be viewed as a ‘moving castle’ in protecting the bacterial biodiversity in the soil. However, it is unclear why there were more diverse bacteria formed a synergistic relationship with earthworms in agricultural sites than in natural sites, more detailed information might be achieved through further investigations [54].

4.2. Network modularity sustains the soil microbial diversity

The biotic interactions are essential for maintaining soil biodiversity [55]. We found that cross-domain networks including earthworms and microbiota were more complex and stable in soils with more earthworm richness. As the earthworms are well-known ecosystem engineers that regulate soil biota [56], we here found that earthworm diversity is important for the preservation of cross-domain interactions between earthworms and soil microbiota in both agricultural and natural land-use types. This might be explained that the amounts of compulsory resources, such as available nitrogen that are needed by soil microbes, could be elevated by earthworms with more diversity [57]. Meanwhile, earthworm diversity is a determinant in regulating plant litter decay [58], the process that provides more various substances for the whole soil biota. Thus, we could anticipate the important roles of cross-domain interactions in regulating the whole soil food webs by both top-down and bottom-up processes [59]. We found that more taxa were enriched than those that were depleted in the earthworm gut, which explains the greater richness in the earthworm gut than in the soil. The gut is assumed as a “hot spot” for certain bacterial taxa [60], which are selected by the unique anoxic and humus earthworm gut environment. The selection process might reflect the cross-domain interactions between microbes and earthworms. Typically, the selected microbes might be useful for the host's fitness [61]. Adversely, this might result in sustaining soil microbial diversity in the earthworm gut. Moreover, we found that the complexity of networks was similar in agricultural and natural land-use types when the earthworm richness was the greatest (four taxa in the present study). This result suggests that cross-domain interactions recovered by increasing the earthworm richness in agricultural land-use types. The impacts of the abundance and community composition of earthworms should not be ignored when considering the earthworm effects on soil microbiome. It is also documented that earthworm effects on soil microbial community and the ecosystem functions derived from the microbiota are greatly dependent on earthworm ecotypes [62]. Further, a global meta-analysis found that the earthworm density was positively correlated with litter decomposition [58].

4.3. Conservation strategies in agricultural ecosystems

In agricultural land-use types, the number of interactions between earthworms and gut microbiota increased with earthworm diversity. While in natural land-use types, the number decreased with earthworm diversity, suggesting that the gut microbiota of earthworms in agricultural land-use types are more dependent on the earthworms. The earthworm gut serves as an important repository of microbial biodiversity [7]. In agricultural land-use types, the number of interactions between earthworms and gut microbiota increased with earthworm diversity. In natural land-use types, however, the number of interactions between earthworms and gut microbiota decreased with earthworm richness. This might be due to the lower input in natural ecosystems compared with agricultural ecosystems. Thus, the maintaining effects of earthworms on soil microbiota are more important in natural ecosystems, as the earthworm gut might serve as an important repository of microbial biodiversity. Thus, protecting the diversity of soil animals, such as earthworms, is a realistic strategy for conserving microbial diversity in agricultural ecosystems. Further, we found that the stability of the cross-domain networks increased with the earthworm diversity in agricultural land-use types. This might indicate that besides protecting the interaction diversity, increasing earthworm diversity is also central to increasing the stability of biotic networks in agricultural ecosystems. In the earthworm gut, microbial interactions are widespread and responsible for nutrient cycling [63]. It explains the functional potentials of earthworm gut microbiota related to the metabolic process in soils [64]. With the diverse microbial taxa, it is assumed that the earthworm gut is a pool of metabolism genes encoded by the abundant bacteria, especially for agricultural ecosystems. The main driver of the gut community composition was the earthworm phylogenetic distance, suggesting that different features of the earthworm gut host different bacterial taxa that are suitable and valuable to be protected in the earthworm gut [65]. Thus, the earthworm diversity is vital for the microbiota to recover the metabolism functions, in the light of the food-web implications. Together, these results highlight the necessity of protecting earthworm diversity in conserving soil microbial functions in agricultural ecosystems.

5. Conclusion

The earthworm guts might serve as “Moving Castles”, protecting a diverse of microbes in the earthworm gut. Notably, the earthworms, which have a large body and are sensitive to agricultural activities themselves, are important in both agricultural and natural habitats to conserve microbial diversity in soil. These processes include not only inhabiting more diverse microbial taxa in their gut but also supporting more modules within the microbial community. We, therefore, highlight that more attention should be paid to earthworm diversity in terms of conserving microbial diversity in terrestrial ecosystems.

Declaration of competing interest

The authors declare that they have no conflicts of interest in this work.

Biography

Manqiang Liu is the head of Soil Ecology Lab, College of Resources and Environmental Sciences, Nanjing Agricultural University. His group is interested in soil ecology, which focuses on the processes belowground involving 1. linking structure and function of soil food web across scales; 2. bridging plant and soil organisms based on functional traits; 3. harnessing biodiversity to maintain soil health via Nature-based Solutions.