Data were processed using the SPSS statistical software (16.0). One-way ANOVA, Duncan’s multiple range test, and t-test were used to determine the significant differences in variables such as soil enzyme activities and PLFAs between treatments. Data were presented as mean ± standard deviation, and the significance level was set at a p-value of <0.05. Principal component analysis (PCA) was performed using the Ade4 package in R statistical software (4.2.3) to evaluate the relationships between soil microbial attributes in the earthworm–plant system. The vegan package in R was used for redundancy analysis (RDA) to explore the variation in contributions of microbial attributes to Pb bioavailability and accumulation in B. campestris. The raw data were standardized to eliminate the effects of different data dimensions. In the RDA, soil microbial attributes were treated as explanatory variables, whereas soil available Pb (DTPA-Pb), plant biomass, and plant Pb (concentration and accumulation) were selected as response variables. The Mantel correlation test was then used to evaluate the strength of the correlations between explanatory variables and response variables and to find out other interactions among them.

The total soil microbial biomass, which was measured as the total content of PLFAs, ranged from 12.5 to 22.1 nmol·g⁻¹ soil across the treatments (Table 1). The maximum biomass was observed at 100 mg·kg⁻¹ Pb. At 0 mg·kg⁻¹ Pb, the presence of earthworms significantly reduced the total content of PLFAs, by 30.9% and 29.3% in SPE and SPA, respectively (p < 0.05). When the soil was contaminated with Pb, the total content of PLFAs in SPE increased significantly, whereas that in SP did not change much, and that in SPA only increased significantly at 100 and 500 mg·kg⁻¹ Pb. For the same Pb contamination level, the total content of PLFAs was higher in SPE than in SPA, with the difference being significant at 100 and 1,000 mg·kg⁻¹ Pb.

Generally, bacterial biomass (G⁺, G⁻, and CB) exhibited the same variation trend as the total content of PLFAs. In contrast, fungal and ACT biomasses showed different variation patterns. At 0 mg·kg⁻¹ Pb, the presence of either earthworm species (A. aspergillum or E. fetida) significantly reduced ACT biomass, whereas only E. fetida significantly decreased fungal biomass compared to SP. In the Pb contamination treatments, ACT biomass was lower in SPA than in SPE, and the difference was statistically significant at 100 and 1,000 mg·kg⁻¹ Pb. Both earthworm species increased fungal biomass at 100 and 500 mg·kg⁻¹ Pb compared to SP. Noticeably, a significant difference between SPE and SPA was found at 1,000 mg·kg⁻¹ Pb, where the fungal biomass in SPE was significantly higher (Table 1).

In this study, the proportions of specific soil microbial community biomass in total biomass revealed that the G⁺ bacteria had the highest relative abundance, ranging from 33% to 39%, whereas CB had the lowest relative abundance, ranging from 8% to 9% across the treatments. The relative abundances of the microbial communities followed the order of G⁺ > G⁻ > ACT > F > CB (Table 1).

The G⁺/G⁻ ratio ranged from 1.00 to 1.53 across the Pb levels (Table 1), with significantly higher values in SP compared to SPA or SPE except at 100 mg·kg⁻¹ Pb. For the SPE treatment, the G⁺/G⁻ ratio decreased with increasing Pb concentration in soil, with significant changes observed at 500 and 1,000 mg·kg⁻¹ Pb. In contrast, the G⁺/G⁻ ratio did not change consistently with increasing Pb concentration for the SPA treatment, with a significant difference only observed between 0 and 100 mg·kg⁻¹ Pb (p < 0.05). The t-test showed that the G⁺/G⁻ ratio was significantly higher in SPA than in SPE at 500 and 1,000 mg·kg⁻¹ Pb (Figure 1A).

The fungal to bacterial biomass (F/B) ratio ranged from 0.13 to 0.17, with higher values in SPA and SPE than in SP (Table 1). This indicates that the presence of earthworms tended to induce changes in the microbial community, and fungi seemed to benefit more from earthworm activities compared to bacteria. The t-test showed that at 500 mg·kg⁻¹ Pb, the F/B ratio was significantly higher in SPA than in SPE (Figure 1B).

Soil enzyme activities related to the cycling of two biologically essential elements, C and N, were studied at different levels of soil Pb contamination with or without the presence of earthworms. FDA hydrolysis, a measure of overall soil enzymatic activities, ranged from 6.70 to 223.4 μg Fluo kg⁻¹ soil h⁻¹, with the highest value observed in SP at 0 mg·kg⁻¹ Pb and the lowest value also in SP but at 1000 mg·kg⁻¹ Pb (Table 2). Generally, as the Pb concentration increased, FDA hydrolysis tended to decrease. For the same Pb contamination level, FDA hydrolysis was generally higher in SPA than in SPE, and the difference was significant at 500 mg·kg⁻¹ Pb (p < 0.05). In general, the activities of β-glu and NAG were not affected by Pb contamination. NAG activity was stimulated by the presence of A. aspergillum at 100 and 500 mg·kg⁻¹ Pb and by E. fetida at 100 and 1,000 mg·kg⁻¹ Pb, whereas β-glu activity was generally inhibited by the presence of earthworms except in SPA at 500 mg·kg⁻¹ Pb (Table 2).

PCA analysis revealed a clear separation of the soil microbial attributes between treatments (Figure 2). The first axis, which accounted for 51.6% of the explained inertia, indicated the effects of Pb contamination on soil microbial PLFAs (Figure 2C). Both the total content of PLFAs and the biomass of each microbial consortium were increased in the Pb-contaminated soil with the presence of E. fetida (p = 0.001). The second axis, representing 17.1% of the explained inertia, described microbial activities and changes in microbial community structure (F/B and G⁺/G⁻; Figure 2B; p = 0.012). The presence of A. aspergillum led to significant changes in microbial structure and FDA hydrolysis. Importantly, E. fetida inoculation was correlated with microbial attributes, suggesting its significant role in influencing soil microbial dynamics. The interspecies comparison of earthworms revealed the different effects of E. fetida and A. aspergillum on microbial attributes, which in turn affected Pb uptake by B. campestris in different ways.

The Pb uptake by B. campestris in this study was examined by focusing on the following parameters: soil available Pb (DTPA-Pb), plant biomass, and plant Pb (concentration and accumulation). Earthworm inoculation led to a Pb increase in soil availability compared to SP. Different effects on plant biomass were observed, with E. fetida promoting and A. aspergillum inhibiting B. campestris growth. As a result of the higher plant biomass, the SPE treatment exhibited a higher Pb accumulation in B. campestris at 100, 500, and 1,000 mg·kg⁻¹ Pb. In contrast, the SPA treatment exhibited a higher Pb concentration in B. campestris, especially at 500 and 1,000 mg·kg⁻¹ Pb (Supplementary Table S1). Earthworm introduction had an effect on the physio-chemical properties of the Pb-contaminated soil, including pH, oxidation–reduction potential (Eh), total N, available N, organic C, dissolved organic C, and cation exchange capacity (Supplementary Table S2). The soil property changes highlighted the significant influence of earthworms on soil biogeochemical processes and Pb migration in the food chain.

The survival rate of A. aspergillum was consistently high, 81.3%, 87.5%, 100%, and 87.5% at 0, 100, 500, and 1,000 mg kg⁻¹ Pb, respectively. The survival rates of E. fetida were 84.0%, 46.1%, 52.6%, and 62.3% at 0, 100, 500, and 1,000 mg kg⁻¹ Pb, respectively. The consistently higher survival rate of A. aspergillum at all Pb levels demonstrated the greater tolerance of A. aspergillum to Pb contamination compared to E. fetida (Tibihenda et al., 2022).

The RDA analysis showed that axes 1 and 2 explained 85.0% and 12.2% of the total variation in response variables, respectively (Figure 3A). The soil microbial attributes as explanatory variables significantly explained 52.5% of the variation in response variables (R² = 0.53, Figure 3B). Specifically, soil microbial biomass and community structure explained 43.0% of the observed variation in response variables, whereas microbial activities explained 9.51%. The G⁻ bacteria contributed 23.2% of the total variation in response variables, significantly higher than any other explanatory variable. FDA hydrolysis had a notably high contribution of 6.58%, making it the most prominent microbial activity variable. The ranking order of the top three explanatory variables based on their contributions to the total variation in response variables was as follows: G⁻ bacteria (23.2%, p = 0.001^***) > FDA hydrolysis (6.58%, p = 0.019^*) > G⁺ bacteria (5.90%, p = 0.031^*; Figure 3B). At the high Pb contamination levels (500 and 1,000 mg·kg⁻¹), SPE showed strong linear correlations with many explanatory variables and response variables such as plant biomass, DTPA-Pb, and plant Pb accumulation, whereas SPA was correlated with F/B ratio, plant Pb concentration, and DTPA-Pb. In contrast, at low Pb contamination levels (0 and 100 mg·kg⁻¹), all earthworm treatments (SP, SPA, and SPE) were associated with FDA hydrolysis and the G⁺/G⁻ ratio. This implies that the increase in the G⁺/G⁻ ratio was negatively correlated with microbial attributes and inhibited Pb uptake.

The Mantel correlation test was utilized to examine the correlations between the explanatory and response variables (Figure 4). The bacterial communities, such as the G⁺ bacteria, G⁻ bacteria, and CB, were significantly positively correlated with DTPA-Pb (p < 0.01), plant biomass (p < 0.001), Pb accumulation (p < 0.01), and Pb concentration (p < 0.05). ACT biomass was positively correlated with DTPA-Pb (p < 0.001), Pb accumulation (p < 0.01), plant biomass (p < 0.05), and Pb concentration (p < 0.05). Fungal biomass was positively correlated with DTPA-Pb only (p < 0.05). Total PLFAs and total bacteria were significantly positively correlated with all four response variables. Moreover, the G⁺/G⁻ ratio was significantly correlated with Pb accumulation (p < 0.05), whereas the F/B ratio did not show a significant effect on the response variables. FDA hydrolysis was significantly correlated with DTPA-Pb (p < 0.01). Remarkably, β-glu and NAG activities were significantly correlated with soil microbes. For instance, the G⁻ bacteria, CB, and fungi were correlated with NAG activity. Total PLFAs and total bacteria were significantly correlated with the constituent microbial communities and NAG activity. Some explanatory variables did not contribute significantly to the total variation in response variables but exhibited significant correlations in the Mantel test. These results support the finding of RDA and provide convincing proof that soil microbial attributes, particularly microbial biomass and community, affect Pb uptake by plants with the driving force of earthworms.

The earthworms had a greater influence on the microbial characteristics than on the chemical properties of the Pb-contaminated soil (Supplementary Figure S1). This finding aligns with previous research indicating that the behavioral and feeding traits of earthworms could lead to soil microbial community composition changes as well as soil enzyme activity changes (Omosigho et al., 2022).

At 0 mg·kg⁻¹ Pb, the presence of earthworms dramatically reduced the total microbial biomass, which was due to the earthworms feeding on the microbes (Table 1; Edwards and Arancon, 2022). The microbes ingested by earthworms may be killed by the microbicidal substances in the gastrointestinal fluid or may not be able to survive the anaerobic conditions in the earthworm gut (Nechitaylo et al., 2010; Gómez-Brandón et al., 2011; Thakur et al., 2022). Although both G⁺ and G⁻ bacterial biomasses were decreased by earthworm presence (Table 1; Figure 1A), the latter was less negatively influenced. This may be attributed not only to the presence of amino acids, mucosal saccharides, and soluble organic C but also to the neutral pH and anaerobic conditions in the earthworm gut, which are more favorable for the survival and growth of G⁻ bacteria (Liu et al., 2018; Hu et al., 2020; Liu P. et al., 2020).

Soil Pb contamination deleteriously affects enzyme activities, microbial diversity, and microbial abundance (Xiao et al., 2020a; Sun et al., 2023). Pb contamination did not significantly affect the total microbial biomass in SP (Table 1; Figure 2), suggesting that the microbiome in soil may have developed effective Pb resistance and tolerance mechanisms (Mao et al., 2022), resulting in resilience in maintaining microbial biomass under Pb stress. Interestingly, the total microbial biomass was increased in SPE at 100, 500, and 1,000 mg·kg⁻¹ Pb (Table 1). This may be because when exposed to high levels of Pb, E. fetida secretes a diverse range of substances such as defensive peptides called drilodefensins and other metabolites for Pb detoxification (Gudeta et al., 2023). These metabolites nourish the soil microbes, leading to an increase in total microbial biomass (Table 1). In addition, E. fetida prefers high-quality organic matter, which is a good energy source for soil microbes and thereby promotes a high microbial biomass (Lin et al., 2016; Sun et al., 2019; Yadav and Singh, 2023). Another reason for the increased microbial biomass that should not be ignored is the death of E. fetida due to Pb toxicity (Tibihenda et al., 2022), leading to less microbes being ingested and killed in the earthworm gut. The higher G⁺/G⁻ ratios in SPA compared to SPE at 100, 500, and 1,000 mg·kg⁻¹ Pb indicate that the presence of A. aspergillum might be more favorable for the G⁺ bacteria compared to E. fetida. The predominance of G⁺ bacteria can also be credited to their exceptional survival in Pb-contaminated environments (Biswas et al., 2021; Borozan et al., 2021). Earthworms generally promote soil fungal and ACT biomasses through their physical and biochemical activities (Medina-Sauza et al., 2019; Song et al., 2020). In contrast, Pb contamination has inhibitory effects on soil fungal and ACT biomasses (Xiao et al., 2020a,b). As a result, the combined effect of earthworm presence and Pb contamination led to complex changes in soil fungal and ACT biomasses.

Microbial activities in soil are biological fingerprints and reliable indicators of Pb toxicity (Luo et al., 2018). In contrast to the study of Yang et al. (2014), this study demonstrated a significant negative correlation between FDA hydrolysis and soil Pb contamination level in SP (Table 2; Figure 2). A high Pb level in the soil would lead to the denaturation of enzymes such as esterases, proteases, and lipases involved in FDA hydrolysis (Patle et al., 2018; Ramana et al., 2021; Li et al., 2022). The different feeding habits of different earthworm ecotypes can have different influences on microbial activity (Boughattas et al., 2019). This may be the reason why the presence of A. aspergillum led to a higher FDA hydrolysis activity than that of E. fetida. The toxic effect of Pb on soil enzyme activity is primarily due to its binding to the active sites of enzyme proteins (Dick, 1997; Utobo and Tewari, 2015). This suggests that the interface between Pb and enzyme active sites or substrates may be limited even at high soil Pb concentrations, resulting in minimal inhibition of enzyme activities responsible for soil N (NAG) and C (β-glu) transformation. Since a majority of enzymes are released by soil microbes (Wang et al., 2023), the presence and activity of the latter have an important effect on the activities of the former. In this study, soil enzyme activities, notably NAG activity, were enhanced by both earthworm species (Table 2), which was also found in a previous study (Jusselme et al., 2013), indicating the potential of earthworms to stimulate microbial activities in Pb-contaminated soils. In summary, the findings confirm our hypothesis that each earthworm ecotype has a unique effect on soil enzyme activities, microbial biomass, and community composition in Pb-contaminated soils due to its distinct lifestyle and feeding preference.

Earthworms influence soil microbial functioning, mitigate the toxic effect of metals on microbial population, structure, and diversity (Šrut et al., 2019; Ren et al., 2021), and promote metal transfer from soil to plant. The RDA results revealed that the soil microbial attributes significantly influenced Pb uptake by B. campestris, with a contribution of 52.5%, much higher than the 17% reported by Xiao et al. (2020b). Explaining 43% of the total variation in Pb uptake, soil microbial biomass and community structure had a greater influence on Pb transfer from soil to plant compared to microbial activity (Figure 3B). The increase in soil microbial biomass due to earthworm activities suggests a likely more diverse soil microbial community with a more diverse genetic pool for more diverse soil functions (Tardy et al., 2014; Li et al., 2018) to influence Pb uptake by plants. Pb-accumulating microbes might serve as a potential Pb reservoir for plant Pb uptake (Pham et al., 2022; Rajivgandhi et al., 2022). The G⁻ bacteria in the drilosphere and rhizosphere might play an important role in the Pb uptake by B. campestris. They promote Pb uptake under the synergistic effects of plant root exudates and earthworms (Das and Osborne, 2018; WeiXie et al., 2022). G⁻ bacteria release growth regulators, siderophores, exopolysaccharides, and organic acids, which change Pb bioavailability and boost plant growth by reducing Pb phytotoxic effects (Jusselme et al., 2015a; Alexandra et al., 2022). Although G⁺ bacteria were abundant in the Pb-contaminated soil and expected to have a major influence on Pb uptake, they only accounted for 5.90% of the total variation in Pb uptake. The possible reason could be the lack of functioning G⁺ bacterial strains that can influence Pb availability. Earthworm introduction altered fungal abundance, which was positively correlated with soil DTPA-Pb (Figure 4), emphasizing the noteworthy symbiotic colonization of plant roots by fungi to enhance soil Pb bioavailability (Manzoor et al., 2019). The correlations between fungal biomass and enzyme activities (Figure 4) indicate that soil fungi probably secrete enzymes in the rhizosphere and the drilosphere (Frey, 2019). As a result, enzymatic breakdown of organic and inorganic Pb-bearing substrates occurs, and DTPA-Pb increases, which explains the high Pb concentration in B. campestris (Supplementary Table S1; Figure 3A). The direct and indirect influences of ACT on Pb uptake by vegetables have rarely been reported. In this study, the earthworms had an effect on ACT biomass, and the latter was correlated with Pb uptake (Figure 4), which was consistent with previous studies (Ali et al., 2017; Liu X. et al., 2020). ACT are well-known to secrete various growth-promoting substances and secondary metabolites, such as metal chelators and organic acids, which promote plant growth and solubilize and mobilize metals, including Pb (Taj and Rajkumar, 2016; Shanthi, 2021). Therefore, ACT played a crucial role in promoting Pb uptake by B. campestris. This study provides evidence and new insights into the microbial mechanism of Pb uptake by B. campestris driven by earthworms via influencing soil microbial biomass and community structure. Comprehensive metagenomics analysis can be employed in future studies to identify the microbial strains promoting Pb uptake by plants.

The increased Pb accumulation in B. campestris was mainly a result of plant growth promotion enhanced by E. fetida (Figure 3A). Though biomass increase would lead to Pb concentration dilution (known as the growth dilution effect), the latter would stimulate Pb uptake, eventually resulting in higher Pb accumulation in B. campestris. Therefore, earthworms play an important role in promoting metal bioaccumulation and plant tolerance to heavy metal stress in contaminated soils (Mahohi and Raiesi, 2021). The decreased plant biomass in SPA led to higher Pb concentration compared to SPE. Moreover, some studies have demonstrated that plant Pb uptake is related to plant biomass, soil microbial biomass, and community structure (Jusselme et al., 2012, 2015b), which is consistent with our findings in the present study. E. fetida played a role in promoting Pb transfer in the vegetable food chain system by decreasing the G⁺/G⁻ ratio (Table 1), which was associated with Pb uptake increase (Figure 3A). This finding supports the well-established idea that a change in microbial community structure is always accompanied by a change in microbial activity and community function (Price-Christenson et al., 2020). Furthermore, it was found that SPA exhibited high F/B ratios, DTPA-Pb, and plant Pb concentrations (Figures 1B, 3A). This indicates that A. aspergillum introduction tended to shift the microbial community toward being fungi-dominated, which can promote Pb transfer from the soil into the vegetable food chain by improving Pb bioavailability. These findings highlight the potential influence of earthworms on soil ecosystems and the entry of metals into the food chain.