It has been reported that 50.51% of micro (meso)-plastics in farmland soils were polypropylene (PP) (Liu et al., 2018). Furthermore, about 82% of microplastics found in soils had sizes <250 µm (Büks and Kaupenjohann, 2020), and only PP microplastics with a particle size smaller than 250 μm decrease nematode offspring (Kim et al., 2020b). Therefore, we selected PP pellets with a particle size of about 200 μm, which were purchased from the Youngling-TECH company, Beijing, China. As described by Liu et al. (2017), the density of the PP microplastic is 0.91 g/cm³, and its melting temperature and brittle temperature are 164–170°C and −35°C, respectively. The experimental soil was collected from the farmland in Ansai County, Shaanxi Province, China. The soil type of this area is sandy soil developed from the Loess parent material, with a SOC concentration of 2.68 g/kg, total nitrogen (TN) concentration of 0.13 g/kg, Olsen phosphorus (TP) concentration of 0.47 g/kg, available phosphorus (AP) concentration of 6.27 mg/kg, ammonia nitrogen (NH₄ ⁺-N) concentration of 4.93 mg/kg, nitrate-nitrogen (NO₃-N) concentration of 18.25 mg/kg, and pH of 8.65 (H₂O/soil at 1:2.5 weight/volume, w/v). Before use, the air-dried soil was sieved through a 5 mm steel sieve. The cultivar of maize was Huanong 887, which was provided by Yangling Seed Company.

This experiment was conducted in the climate-controlled chamber (AGC-Doo3N, Hangzhou, China) at the Institute of Soil and Water Conservation, Chinese Academy of Sciences, Yangling, Shaanxi province, China from July to October 2019. As reported, microplastic concentrations in soil samples collected from the industrial area were 300–67,500 mg/kg, that is 0.03%–6.7% (Fuller and Gautam, 2016). Meanwhile, a field survey reported that the concentration of plastic residues in Chinese Loess soil is 1% (w/w) (Chen, 2016). Taking the newly inputs of microplastics into soils as well as the great effort to reduce microplastic contamination into consideration, we set up the following four treatments: 1) CK, no microplastics added to the soil; 2) C1, 75 g of microplastics added to the soil (0.5%, w/w); 3) C2, 150 g of microplastics added to the soil (1%, w/w), and 4) C3, 300 g of microplastics added to the soil (2%, w/w). Each treatment had six replicates: 15 kg of soil and microplastics were thoroughly mixed before they were put into a PVC bucket with an upper diameter of 30 cm, a lower diameter of 22 cm, and a height of 30 cm. The pots were incubated in the light at 28°C (relative humidity of 80%, 300 m (photons) m⁻² s⁻¹). To avoid potential impacts of drought or water-logging stress on plant and soil biota, soil moisture was maintained at 65 ± 5% throughout the experiment. Crop diseases and insect pests were controlled according to field management measures in this region during the experiment.

During the fully ripe stage of maize, more accurately on October 22–24, 2019, the bulk soils from the topsoil layer (0–20 cm) of each pot were collected. All collected soil samples were sieved with 4 mm and mixed thoroughly, resulting in 24 composite samples. These samples were stored at 4°C before further processing.

Soil nematodes were extracted from 100 g of field-moist soil within 14 days after sampling. In brief, 100 g of soil were mixed with 200 ml of water in a 500 ml beaker and allowed to stand for 1 min. Then, the mixture was mixed well and sieved through 20/500-mesh stacked sieves to concentrate nematodes on a 500-mesh sieve. After replicating the same procedure three times, the materials on the 20-mesh sieve were discarded and those on the 500-mesh sieve were washed into a 100 ml centrifuge tube. Finally, nematodes were further separated from soils using the sucrose centrifugation method (Jenkins, 1964). The recovered nematodes were killed by heating at 60°C for 3 min and were preserved in 4% formalin solution before further analysis. The nematode individuals per sample were counted under a dissecting microscope, and the first 150 nematodes encountered in the sample were identified to the genus level where possible according to the external morphological features of the body shape, cuticle, head, papilla, and tail as well as the internal structures of the digestive system, reproductive system, excretory system and nervous system as described by Bongers (1994), Ahmad and Jairajpuri (2010) as well as Interactive Diagnostic Key to Plant Parasitic, Free-living, and Predaceous Nematodes (http://nematode.unl.edu/nemakey.htm) at a 400× or 1,000× magnification using an inverted compound microscope. If the individual recovered from a sample was less than 150, all specimens were identified. Nematode population was expressed as the individuals per 100 g of dry soil after being adjusted with soil moisture. Soil nematode genera were classified into bacterivores, fungivores, plant parasites, and omnivores–predators (Yeates et al., 1993) according to their feeding habits and assigned a c-p value (Bongers, 1990; Bongers and Bongers, 1998). The density of an individual trophic group was calculated.

Since soil biodiversity and community composition determine ecological functioning (Wagg et al., 2014), we examine the responses of soil nematode abundance, diversity, and composition to microplastic pollution first. Biodiversity indices, including taxa richness (S), Margalef’s richness (SR), Shannon–Wiener index (H′), Simpson’s index (λ), and Pielou’s evenness index (J′), were calculated based on the richness and relative abundance data. Subsequently, we compare the maturity indices (Bongers, 1990; Freckman and Ettema 1993), food web conditions (Ferris et al., 2001), and metabolic footprints (Ferris, 2010) across treatments. The maturity indices, food web conditions, and metabolic footprints were calculated by submitting the soil nematode array to NINJA (Nematode INdicator Joint Analysis) (Sieriebriennikov et al., 2014), provided at https://shiny.wur.nl/ninja.

The physicochemical properties of the soil, such as soil moisture, pH, SOC, TN, NH₄ ⁺-N, NO₃-N, TP, and AP, were determined using standard methods as described elsewhere. In brief, the soil moisture was calculated as the weight lost when baking a sample to a constant weight at 105 ± 2°C for 10 h, and the soil pH was estimated using an electronic pH meter fitted with a glass electrode (WTW pH 330, WTW, Weilheim, Germany) in soil: the water solution at a ratio of 1:2.5 (weight/volume, w/v), the SOC content was analyzed by the potassium-dichromate (K₂Cr₂O₇) oxidation method (Walkley and Black, 1934), TN was determined by the Kjeldahl method (Boström et al., 1974), the NH₄ ⁺-N content and NO₃-N content were extracted with 2 M KCl and were measured using an AA3 continuous flow auto-analyzer (AutoAnalyzer 3-aa3, Bran⁺ Luebbe, Germany), the total P content of the soil was determined using colorimetric analysis after digestion with H₂SO₄ and HClO₄ (Olsen et al., 1982; Wu et al., 2021) and the AP content was extracted with 0.5 M NaHCO₃ at pH 8.5 and measured colorimetrically using a UV spectrophotometer and the ammonium molybdate method (Wu et al., 2021).

The effects of microplastic concentration on total nematode abundance, abundance per trophic group, nematode diversity indices, maturity indices, food web indices, and metabolic footprints of soil nematodes were analyzed using the linear mixed model with the dose as the fixed factor and pot as the random factor using the “nlme” package in R (R Core Development Team, 2018). Before analysis, the normality and the homogeneity of the residuals for data were examined by the Shapiro–Wilk test or by the Kolmogorov–Smirnov test in the “stats” package. If the data do not meet the assumption of the normality and the homogeneity of the variations, the effect of microplastic additions on a given variable was examined with the Kruskal–Wallis test, and differences across microplastics concentrations were checked and versioned by the “boxplerk”. Correlations between nematode community parameters and environmental factors were analyzed by the “cor.test” function. Furthermore, taxonomic community composition differences were examined by non-metric multidimensional scaling (NMDS) based on the Bray–Curtis distances. To show the reliability of the results, non-parametric multivariate statistical tests, including ANOSIM, Adonis, and MRPP, based on different algorithms were performed simultaneously to examine statistical significance using the “vegan” package. The major environmental factors shaping soil nematode community were selected with “envfit” function in the “vegan” package. In addition, the association between individual nematode abundance and microplastic contamination was examined with the “Indicspecies” package. Finally, the direct and indirect effects of microplastics on soil nematode community were analyzed using the “PLSPM” package. All the aforementioned analyses were performed with R software (version 4.1.2.; R foundation for statistical computing).

The abundance of total nematodes at a depth of 0–20 cm of bulk soil ranged from 20 to 108 nematode specimens/100 g dry soil. The abundance of total nematodes declined significantly with the increase in the microplastic concentration (Kruskal–Wallis test: χ² = 12.76, p = 0.005; Figure 1A). The increase in the microplastic concentration exerted significant negative effects on the abundances of nematode trophic groups and free-living nematodes (Figures 1B,C,E,F), with exception of bacterivores (Kruskal–Wallis test: χ² = 12.76, p = 0.05; Figure 1D). The correlation analysis suggested that nematode abundances were correlated to the SOC content, TP content, and C/N of soils (Supplementary Figure S1). Specifically, nematode abundances were positively correlated to the soil TP content, whereas negatively correlated to the SOC content and C/N. PLS modeling indicated that the direct impact of microplastics on nematode abundance is more significant than the indirect one (Supplementary Figure S3A).

Twenty-nine genera of nematodes including thirteen plant-feeding nematodes, seven bacterial-feeding nematodes, five fungal-feeding nematodes, and four omnivorous nematodes were recovered from the soil samples. The number of nematode genera ranged from six to nineteen and decreased significantly with the increase in the exposure dose of microplastics (Kruskal–Wallis test: χ ² = 16.35, df = 3, p < 0.001; Figures 2, 3A). Moreover, the addition of microplastics greatly impacted other diversity indices [including SR (Kruskal–Wallis test: χ ² = 12.55, df = 3, p < 0.001; Figure 3B), H' (one-way ANOVA: F = 11.00, df = 3, p = 0.0002; Figure 3C), and λ (Kruskal–Wallis test: χ ² = 12.89, df = 3, p < 0.005; Figure 3D)] of soil nematode community. However, the addition of microplastics did not affect the J′ of the soil nematode community (Kruskal–Wallis test: χ ² = 4.71, df = 3, p = 0.19; Figure 3E). Further analysis suggested that nematode community diversity was correlated to the SOC content, TP content, C/N, and C/P of soils (Supplementary Figure S2). The PLS path modeling confirms that PP concentrations affect soil nematode communities both directly and indirectly (Supplementary Figure S3B).

The NMDS plot shows that soil nematode assemblages across the examined microplastic application doses cannot be separated from each other (Figure 4, stress = 0.20, R = 0.45, and p < 0.01). Additionally, the results of ANOSIM, Adonis, and MRPP confirmed that the differences in soil nematode communities across examined microplastic application doses were significant (Table 1), indicating that the microplastic exposure dose exerted significant impacts on the nematode community composition. Indicator species analysis (Supplementary Table SA2) revealed that the Aporcelaimus, Longidorus, and Thonus were abundant in soils of control and the Plectus was more abundant in soils of C1 treatment. The Rotylenchus, Helicotylenchus, and Microdorylaimus were observed in soils of CK and C1 treatments, Scutellonema were more abundant in soils of C1 and C2 treatment, and a few Cephalobus were observed in soils exposed to microplastic pollution. The “Envfit” showed that the environmental factors shaping soil nematode communities in bulk soils of maize at a depth of 0–20 cm were the soil pH, SOC content, C/N, and TP content (Table 2).

The examined microplastic concentration negatively affected ∑MI (Kruskal–Wallis test: χ ² = 14.946, df = 3, and p < 0.002) and SI (Kruskal–Wallis test: χ ² = 17.734, df = 3, and p < 0.0005) of the soil nematode community, whereas it did not affect PPI (Kruskal–Wallis test: χ ² = 0.576, df = 3, and p = 0.90) and EI of the soil nematode community (one-way ANOVA: F = 0.783, df = 3, and p = 0.517). Further analysis suggested the ∑MI and SI of the soil nematode community were positively correlated to the soil TP content, while they were negatively correlated to SOC content (Supplementary Figure S2).

The examined microplastic concentration negatively affected the composite footprint (Kruskal–Wallis test: χ ² = 16.72, df = 3, and p = 0.0008; Figure 5A), enrichment footprint (Kruskal–Wallis test: χ ² = 10.85, df = 3, and p = 0.01; Figure 5B), structure footprint (Kruskal–Wallis test: χ ² = 18.21, df = 3, and p = 0.0004; Figure 5C), herbivore (plant-feeding nematode) footprint (Kruskal–Wallis test: χ ² = 18.73, df = 3, and p = 0.0003; Figure 5D), bacterivore footprint (Kruskal–Wallis test: χ ² = 9.61, df = 3, and p = 0.02; Figure 5E), fungivore footprint (Kruskal–Wallis test: χ ² = 10.85, df = 3, and p = 0.01; Figure 5F), and omnivore footprint (Kruskal–Wallis test: χ ² = 18.35, df = 3, and p = 0.0004; Figure 5G). Further analysis suggested that nematode metabolic footprints were correlated to the SOC content, TP content, C/N and C/P of soils (Supplementary Figure S4).