The Remediation Mechanism of Soil Atrazine Contamination by Vermicompost: A Metagenomic Perspective

TL;DR

This study explores how vermicompost helps break down atrazine in soil by analyzing microbial communities and key genes involved in the process.

Contribution

The study identifies specific microbial genera and functional genes enriched by vermicompost that enhance atrazine degradation.

Findings

Vermicompost significantly restructured soil microbial communities, increasing Proteobacteria and Bacteroidetes abundance.

SnV2 showed the highest atrazine degradation (2.94%) and enriched Bauldia for dechlorination.

60-day vermicompost (SnV2) upregulated key genes like trzN and atzB, demonstrating superior remediation performance.

Abstract

Atrazine persistence poses serious environmental threats. This study used metagenomics and qPCR to elucidate the remediation mechanism of vermicompost in atrazine degradation pathways. Seven treatments were established: unsterilized soil (CKn); sterilized soil amended with 45 (SsV1), 60 (SsV2), and 75 (SsV3) days of vermicompost; and unsterilized soil with the same vermicompost (SnV1, SnV2 and SnV3). Vermicompost significantly restructured soil microbial communities. SsV1 exhibited the highest Proteobacteria abundance (51.38%), while SsV3 markedly increased Bacteroidetes abundance (10.34%). Functional annotation revealed that vermicompost enriched carbohydrate metabolism-related COG units and upregulated CAZymes (e.g., CE1 and CE10 families), providing energy support for degrading microbial communities. Regarding metabolic pathways, SnV2 exhibited the highest atrazine degradation abundance (2.94%), significantly enriching Bauldia (4.84 RPKM) for dechlorination. During cyanuric acid ring-opening, SnV3 significantly enriched Pseudorhodoplanes (12.14 RPKM). During terminal mineralization, SsV2 increased Caenimonas abundance (15.25 RPKM) and introduced the exogenous genus Pseudorhodoplanes (7.78 RPKM). qPCR confirmed SnV2’s trzN (day 20) and atzB (day 40) reached 9.03 × 104 and 6.95 × 107 copies/g, respectively. These findings indicate vermicompost accelerated atrazine mineralization by enriching degradative microbial communities and promoting key functional gene expression, with 60-day vermicompost demonstrating superior performance. This study provides a robust theoretical framework for remediating atrazine-contaminated soil by vermicompost.

Figures

Full text

1. Introduction

Atrazine, a triazine herbicide widely used in agricultural production, has garnered significant attention due to its persistence, mobility, and potential toxicity to non-target organisms. It exhibits a half-life ranging from 41 to 231 days [1]. Owing to its low soil adsorption and moderate water solubility, atrazine can contaminate both farmland and water bodies (groundwater and surface water), reaching concentrations as high as 30 μg·L^−1^ [2]. Studies have detected atrazine in 32% of U.S. water bodies, with a mean concentration of 0.17 μg·L^−1^ [3]. In the lower Ganges basin (WBB) of West Bengal, India, atrazine concentrations range from 0.95 to 3.93 μg·L^−1^, exceeding maximum permissible limits by a factor of 46 [4]. Atrazine and its metabolites have been detected in Brazilian drinking water treatment plants at 2–2744 ng·L^−1^ [5], with a 100% detection rate. In the Mid-Atlantic region, atrazine was detected in over 99% of water samples near agricultural land (max. 1.9 μg·L^−1^), positioning it as a dominant pollutant [6]. Concurrently, atrazine exerts ecotoxicological effects on soil enzymatic and microbial activities. High atrazine concentrations significantly inhibit soil enzymatic activity [7] and disturb the structure of bacterial communities, resulting in soil compaction and diminished fertility [8,9]. Prolonged atrazine application leads to a gradual decline in soil urease, sucrase, cellulase, and catalase activities. Consequently, microbial nutrient consumption decreases, leading to a reduction in soil microbial populations [10]. Thus, atrazine residues are ubiquitous in soil and groundwater systems, posing long-term threats to ecological security and human health [11,12]. Additionally, atrazine exposure in human cumulus granulosa cells interferes with steroidogenesis and ovulation, thereby adversely affecting human reproduction [13]. Furthermore, it can induce dysregulation and abnormal expression of human mRNA, increasing the risk of cancer, neurological disorders, and vascular development impairments [14]. Given its environmental prevalence and multifaceted toxicity, developing efficient remediation strategies to eliminate residues and mitigate ecological risks is imperative.

Among existing soil remediation technologies, bioremediation is widely employed for atrazine mitigation due to its environmental friendliness, low cost, and potential for complete pollutant mineralization [15,16]. Microbial atrazine degradation is mediated via a complex metabolic network (Figure 1; KEGG pathway 00791) [17]. This network outlines the primary biochemical pathways facilitating complete mineralization, comprising three initial pathways that converge at cyanuric acid to facilitate final mineralization. These pathways are predominantly bacterially driven and proceed efficiently under aerobic conditions, differing primarily in their initial reactions, key enzymes, and associated genes. The hydrolysis pathway (dominant) initiates with dechlorination catalyzed by chlorohydrolases encoded by the atzA or trzN genes, yielding non-herbicidal hydroxyatrazine (HYA) [18]. This route is the most extensively studied and exhibits the highest mineralization efficiency. Subsequent steps are sequentially catalyzed by enzymes encoded by atzB, atzC, atzD, atzE, and atzF. Representative strains, such as Pseudomonas sp. ADP, possess the complete atzA–F gene cluster, illustrating the capability of a single organism to mineralize atrazine completely [19]. In contrast, deethylation and deisopropylation (minor routes) are catalyzed by distinct systems, including cytochrome P450 monooxygenases, producing deethylatrazine (DEA) and deisopropylatrazine (DIA). Although these intermediates retain toxicity, they can be further metabolized by other microorganisms, ultimately entering the mainstream cyanuric acid mineralization pathway [20]. These routes often supplement the core hydrolysis pathway, particularly in environments lacking atzA/trzN genes. In natural soils, complete mineralization is rarely accomplished by a single strain; instead, it typically requires a multispecies cooperative metabolic relay. This mechanism primarily involves metabolic complementation, where one microorganism performs upstream dealkylation and the resulting products are further degraded by organisms carrying the atzABC gene cluster. Furthermore, functional division occurs as different microorganisms become enriched at specific stages, achieving efficient mineralization through community-level metabolic networks. This synergy is crucial for ensuring complete pollutant removal and preventing the accumulation of toxic intermediates.

Numerous bioremediation studies have aimed to enhance in situ soil remediation efficiency. Currently, utilizing organic fertilizers to stimulate the metabolic activity of indigenous microorganisms is a research priority [21,22]. Virk et al. (2025) reported that organic amendments effectively mitigate the negative impacts of pesticides by stimulating soil carbon-cycle enzymes and improving biochemical properties, thereby indirectly promoting degrader activity and restoring ecological functions [23]. Li et al. (2022) [22] demonstrated that sheep manure compost significantly improves atrazine degradation efficiency. The underlying mechanism involves organic fertilizers providing labile carbon sources, enhancing soil physicochemical properties, and activating indigenous degraders . Among various organic amendments, vermicompost exhibits remediation properties superior to those of conventional compost. Vermicompost is a specialized fertilizer produced via the synergy between earthworms and microorganisms during organic waste conversion. It possesses a larger specific surface area, more stable humus, and greater microbial diversity [24,25]. Research indicates that vermicompost offers significant advantages for the degradation of tetracycline and triazine herbicides [26,27]. Castillo-Díaz et al. (2016) [28] found that vermicompost significantly regulates imidacloprid adsorption–desorption in soil. By reshaping bacterial communities and enhancing dehydrogenase activity, vermicompost mitigates pesticide toxicity and promotes pollutant degradation. Lin et al. (2016) [29] demonstrated that earthworm activity activates indigenous pentachlorophenol degradation by augmenting substrate availability. Zhang et al. (2023) [30] observed that vermicompost enhances atrazine adsorption by increasing organic carbon and humic fractions, thereby reducing migration risks. Vermicompost is enriched with microbial taxa possessing degradative potential, including Proteobacteria, Actinobacteria, and Firmicutes. Within these groups, typical degraders such as Pseudomonas and Arthrobacter participate in atrazine metabolism via hydrolytic and oxidative pathways [31,32,33]. Functional analysis by Zhang et al. (2025) [34] indicates that vermicompost increases the relative abundance of the atrazine-degradation genes atzB and atzF, further confirming its potential for enhanced degradation.

However, despite the significant potential of bioremediation for atrazine removal, current research into the underlying mechanisms still faces numerous bottlenecks. Existing studies predominantly focus on degradation rates or the abundance of specific genes, lacking a systematic understanding of metabolic network succession within microbial communities during remediation. Regarding atrazine remediation using vermicompost, there is an urgent need to explore how complex microbial communities synergistically facilitate full-pathway mineralization and to identify potential auxiliary metabolic mechanisms. To address these gaps, in-depth studies using metagenomics are essential. Metagenomics enables the culture-independent analysis of microbial functional profiles at the whole-genome level. This approach identifies key functional groups and systematically maps functional profiles, including carbohydrate-active enzymes (CAZymes), thereby addressing the mechanistic depth lacking in current studies. Consequently, this study employed vermicompost at different maturation stages (45, 60, and 75 days) and integrated metagenomics with qPCR to reveal the molecular mechanisms by which vermicompost enhances atrazine degradation in soil. The microbial molecular analysis in this study extends prior research on vermicompost-mediated atrazine degradation. Previous work demonstrated that vermicompost significantly enhances atrazine removal by promoting initial reactions such as dechlorination, which leads to the accumulation of intermediate metabolites, including HYA, DEA and DIA [34]. Building on these findings, this study employs metagenomics and quantitative PCR (qPCR) to elucidate the microbial functional regulatory networks driving this transformation process. This integrated approach provides a deeper understanding of how different vermicompost maturation stages influence the synergistic mineralization of atrazine and its intermediates. The specific objectives of this study are to: (1) clarify the dynamic regulatory effects of vermicompost at different maturation stages on soil microbial community structure and diversity, focusing on the enrichment characteristics of core genera associated with atrazine degradation; (2) elucidate the regulatory mechanisms of vermicompost on atrazine metabolic pathways at the functional gene level, including key enzymes (e.g., trzN, atzB, and atzC) and auxiliary energy metabolism pathways; and (3) evaluate the differential impacts of vermicompost produced at different composting stages on soil atrazine degradation capacity to optimize application strategies and provide a robust theoretical basis for the precision bioremediation of atrazine-contaminated soils.

2. Materials and Methods

2.1. Experimental Materials

Analytical-grade atrazine standard (purity ≥ 99.5%) was procured from Shanghai Yien Chemical Technology Co., Ltd. (Shanghai, China). All other chemical reagents were of analytical grade. Soil samples were collected from the surface layer (0–20 cm) of farmland at the experimental base of Jilin Agricultural University (43°26′ N, 125°05′ E), Changchun City, Jilin Province. No prior atrazine contamination was detected in the soil. The collected soil was air-dried naturally and passed through a 2 mm sieve for subsequent use. Vermicompost was prepared by mixing maize stover and cattle dung in equal proportions. Destructive sampling was performed at 45, 60, and 75 days of composting. Samples from distinct batches were thoroughly homogenized and represented various maturation stages, which were subsequently stored at −80 °C until further analysis. High-performance liquid chromatography (HPLC) analysis revealed no detectable atrazine or its major metabolites in the initial soil and vermicompost (detection limit: 0.01 mg/kg), confirming they were uncontaminated baseline media. The physicochemical properties of the soil and vermicompost (e.g., pH, organic matter, and total nitrogen) were previously characterized (Table S1) [34]. The baseline soil properties included a pH of 5.92, total organic carbon (TOC) of 14.87 g/kg, and total nitrogen (TN) of 1.73 g/kg. Corresponding properties for the 45, 60, and 75-day vermicompost batches were: pH (7.61, 7.78, and 7.56), TOC (396.57, 381.62, and 372.23 g/kg), and TN (15.85, 18.23, and 19.24 g/kg), respectively.

2.2. Atrazine Degradation Experiment

The atrazine biodegradation experiment consisted of seven experimental groups (Table 1). A portion of the original soil was sterilized at 121 °C for 3 h to prepare the sterilized soil medium. Atrazine solution (100 mg·L^−1^ in methanol) was added to 1500 g of soil to achieve an initial concentration of 10 mg·kg^−1^ (dry weight); the mixture was subsequently evaporated at room temperature for 24 h to ensure complete methanol removal. Based on typical environmental levels (3–6 mg·kg^−1^) and concentrations used in similar studies (10 mg·kg^−1^) [27], the contamination level in this study was set at 10 mg·kg^−1^. In accordance with typical agricultural application rates (30–60 m^3^·ha^−1^) [35], 50 mg of vermicompost was incorporated per gram of soil. Sterile distilled water was added to maintain the soil at 60% of its maximum water-holding capacity. Biodegradation experiments were carried out in the dark at 25 °C for 40 d. Destructive sampling was conducted at 0, 10, 20, 30, and 40 d for qPCR quantification, with metagenomic analysis performed on the day 40 samples. The sampling time points for qPCR analysis (0, 10, 20, 30, and 40 d) were selected to capture the dynamic expression patterns of key functional genes throughout the degradation period, following the atrazine degradation kinetics established in our previous study [34]. Metagenomic analysis was performed exclusively on terminal samples (day 40) to characterize the stabilized microbial communities that mediated the final degradation process. This endpoint approach is commonly utilized in remediation studies to identify core functional taxa and pathways enriched under sustained selection pressure, thereby minimizing transient fluctuations observed during earlier phases. All experimental treatments were conducted in triplicate (n = 3) to ensure statistical significance and reliability.

2.3. Metagenomic Analysis

2.3.1. DNA Extraction, Library Construction, and Sequencing

Total genomic DNA was extracted from 0.5 g of soil using the TIANamp Soil DNA Kit (Tiangen Biotech, Beijing, China) following the manufacturer’s instructions. DNA integrity and concentration were verified via 1% agarose gel electrophoresis and a Qubit fluorometer (Thermo Fisher Scientific, Waltham, MA, USA). Sequencing libraries were constructed using the Illumina TruSeq Nano DNA LT Library Preparation Kit (Illumina, San Diego, CA, USA). Briefly, high-quality DNA was randomly fragmented to an average size of ~350 bp using a Covaris M220 focused-ultrasonicator (Covaris, Woburn, MA, USA). The fragmented DNA underwent end-repair, adenylation, and ligation of Illumina paired-end adapters. Fragments with ligated adapters were selectively enriched via PCR amplification. Library quality was assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). Qualified libraries were sequenced on the Illumina HiSeq 4000 platform (Illumina, San Diego, CA, USA) using the HiSeq 3000/4000 PE Cluster Kit and SBS Reagent Kits (Illumina, San Diego, CA, USA), generating 150 bp paired-end reads.

2.3.2. Sequence Data Processing and Quality Control

Raw sequencing reads were demultiplexed according to their unique barcodes. Adapter sequences and low-quality bases were trimmed using FastP (v0.23.2) with default parameters. Subsequently, Sickle (v1.33) was employed to remove reads with an average quality score below 20 or a length shorter than 50 bp. This process yielded high-quality clean data for downstream analysis. An average of approximately 45 million clean read pairs (~6.8 Gb of clean data) were obtained per sample.

2.3.3. Bioinformatic Analysis

De novo assembly of quality-filtered reads was performed using MEGAHIT (v1.2.9) with the “--meta-sensitive” preset parameter. Contigs with a length ≥ 300 bp were retained for subsequent analysis. Open reading frames (ORFs) were predicted from the assembled contigs using Prodigal (v2.6.3) in metagenomic mode. A non-redundant gene catalog was constructed by clustering all predicted amino acid sequences with CD-HIT (v4.8.1) using 95% sequence identity and 90% coverage thresholds. The abundance of each gene in each sample was calculated by mapping clean reads back to the non-redundant gene catalog using Bowtie2 (v2.4.5) and normalized as Reads Per Kilobase per Million mapped reads (RPKM).

For taxonomic annotation, non-redundant gene sequences were aligned against the NCBI NR database (2023-10) using DIAMOND (v2.1.6) with an e-value cutoff of 1 × 10^−5^. Functional annotations were conducted by aligning sequences against the following databases: eggNOG (v5.0) for COG categories, KEGG (Release 107.0) for metabolic pathways, and the CAZy database (dbCAN3) for carbohydrate-active enzymes, using DIAMOND with an e-value threshold of 1 × 10^−5^. The relative abundances of taxonomic groups and functional units were calculated based on the highest-scoring annotation for each gene. Detailed statistical results regarding raw data, quality control, assembly, and gene prediction are provided in Supplementary Tables S2–S5.

2.4. Real-Time PCR Quantification (qPCR) of Atrazine Degradation Genes

Real-time PCR analysis of atrazine degradation genes was conducted by Shanghai Qiyin Biotechnology Co., Ltd. (Shanghai, China). Soil genomic DNA was extracted using the TIANAMP Soil DNA Kit (Tiangen Biotech, Beijing, China). Samples were analyzed on the StepOnePlus™ Real-Time Fluorescent Quantitative PCR System (Thermo Fisher Scientific, Waltham, MA, USA) using the TB Green™ Premix Ex Taq™ II (Tli RNaseH Plus) Reagent Kit (Takara, Bio, Kusatsu, Shiga, Japan; Code No. RR820A). The primer sequences used for qPCR are shown in Table 2. qPCR reaction system was as follows: TB Green™ Premix Ex Taq™ II (TaKaRa Bio, Kusatsu, Shiga, Japan) 5 μL, Primer F 0.4 μL, Primer R 0.4 μL, ROX Reference Dye 0.2 μL, DNA Sample 1 μL and ddH2O 3 μL. The samples were run on the StepOnePlusTM Real-Time PCR Systems (Thermo Fisher Scientific, Waltham, MA, USA). The reaction conditions were as follows: 95 °C, 30 s, one cycle, 95 °C for 20 s, 55 °C for 30 s, 72 °C for 30 s, 40 cycles. qPCR of each gene was performed in triplicate. Melting curve analysis and agarose gel electrophoresis confirmed the specificity of qPCR reaction.

2.5. Data Statistical Analysis

M Metagenomic data analysis was performed using the Majorbiocloud Platform (https://cloud.majorbio.com, accessed on 5 March 2024). Linear discriminant analysis effect size (LEfSe) analysis (http://huttenhower.sph.harvard.edu/LEfSe, accessed on 5 March 2024) was utilized to identify microbial taxa with significant abundance differences across groups from phylum to genus levels, applying an LDA score > 3.8 and p < 0.05 as the significance thresholds. Statistical analyses were conducted using SPSS v. 25.0 (IBM Corp., Armonk, NY, USA, accessed on 4 March 2024). Differences in measured parameters between treatment groups were assessed via two-way analysis of variance (ANOVA) at a 95% confidence level. Multiple comparisons were performed using Duncan’s multiple range test (p < 0.05). All results are expressed as the mean ± standard deviation (SD).

3. Results and Analysis

3.1. NR Species Annotation

Based on high-quality metagenomic sequence data (see Section 2.3 and Supplementary Tables S2–S5 for detailed outputs), species annotation analysis was performed to characterize the microbial composition of each sample. As shown in Figure 2a, at the domain level, bacteria exhibited the highest abundance across all treatments, followed by Archaea. The relative abundance of Archaea was higher in unsterilized soil than in sterilized soil amended with vermicompost. Figure 2b shows that, at the phylum level, Proteobacteria had the highest relative abundance, followed by Actinobacteria. The dominant phyla in CKn were Proteobacteria, Actinobacteria, Chloroflexi, Acidobacteria, and Bacteroidetes. The dominant phyla in SsV1, SsV2, and SsV3 were Proteobacteria, Actinobacteria, Bacteroidetes, Firmicutes, and Verrucomicrobia. The dominant phyla in SnV1, SnV2, and SnV3 were Proteobacteria, Actinobacteria, Bacteroidetes, Acidobacteria, and Verrucomicrobia. Vermicompost application increased the relative abundance of Proteobacteria, Bacteroidetes, Verrucomicrobia, and Firmicutes in the soil. Among these, Proteobacteria exhibited the highest relative abundance (51.38%) in SsV1, representing an 11.11 percentage-point increase relative to CKn. The relative abundances of Bacteroidetes, Verrucomicrobia, and Firmicutes in SsV3 reached 10.34%, 5.66%, and 6.25%, respectively, corresponding to increases of 5.85, 1.63, and 4.85 percentage points compared with CKn. At the genus level (Figure 2c), the dominant genera in CKn were unclassified_p_Actinobacteria, Anaeromyxobacter, unclassified_c_Actinomycetia, unclassified_p_Acidobacteria, and unclassified_p_Chloroflexi. The dominant genera in SsV1, SsV2, and SsV3 were Anaeromyxobacter, unclassified_p_Verrucomicrobia, unclassified_p_Acidobacteria, Azoarcus, and Azotobacter. The dominant bacterial genera in SnV1, SnV2, and SnV3 were unclassified_p_Actinobacteria, unclassified_p_Acidobacteria, unclassified_c_Actinomycetia, unclassified_o_Myxococcales, and unclassified_p_Proteobacteria. Combining the genus-level heatmap (Figure 2d), the heatmap clustering analysis revealed distinct separation between the microbial communities in the CKn and vermicompost treatment groups, indicating that vermicompost significantly altered the soil microbial community structure. Following vermicompost application, the relative abundances of Azotobacter, unclassified_p_Proteobacteria, Hyphomicrobium, Devosia, Cellulomonas, Pseudoxanthomonas, Mycobacterium, Microbacterium, and Mycolicibacterium were markedly higher than in the CKn treatment.

Principal coordinate analysis (PCoA) was employed to compare bacterial community composition among treatment groups at the genus level, with β-diversity serving as the metric for between-community variation (Figure 2e). Based on Bray–Curtis dissimilarity, the first two principal coordinates explained 78.23% of the community variance. The bacterial community structure of the CKn treatment, sterilized soil amended with vermicompost, and unsterilized soil amended with vermicompost was distinct. Furthermore, the bacterial community structures among the vermicompost treatments were clearly separated, indicating differences in community structure among the groups. Overall, vermicompost application altered the soil bacterial community structure.

LEfSe uses linear discriminant analysis of taxonomic composition to identify taxa that differ among treatments and to screen for potential biomarker microorganisms. As shown in Figure 3, all LDA scores exceed 3.8, indicating marked differences in taxonomic composition. At the genus level, biomarkers in CKn included Anaeromyxobacter, Aquabacterium, Methlobacillus, Nocardioides, Ramlibacter, unclassified_c_Actinomycetia, unclassified_c_Deltaproteobacteria, unclassified_p_Acidobacteria, unclassified_p_Actinobacteria, and unclassified_p_Chloroflexi. In SsV1, biomarkers included Azoarcus, Cellulomonas, Hydrogenophaga, and Mycolicibacterium. In SsV2, biomarkers included Actinotalea, Azotobacter, unclassified_o_Bacteroidales, and unclassified_p_Verrucomicrobia. In SnV2, biomarkers included Hyphomicrobium and unclassified_p_Proteobacteria. In SnV3, biomarkers included Mesorhizobium, Microbacterium, Pseudoxanthomonas, and unclassified_o_Solirubrobacterales.

3.2. COG Function Annotation

Differences in microbial functional profiles among samples reflect variations in microbial physiological and metabolic activities. By aligning non-redundant gene sequences to the EggNOG orthologous group database, gene functional annotations were generated and analyzed. The corresponding COG functions were classified into four major categories: metabolism, information storage and processing, cellular processes and signaling, and uncharacterized functions. As shown in Figure 4, CKn showed comparatively high relative abundances of amino acid transport and metabolism, signal transduction mechanisms, energy production and conversion, and coenzyme transport and metabolism. Vermicompost application was associated with higher relative abundances of carbohydrate transport and metabolism, transcription, inorganic ion transport and metabolism, lipid transport and metabolism, replication, recombination, and repair, defense mechanisms, posttranslational modification, protein turnover, chaperones, secondary metabolite biosynthesis, transport, and catabolism, intracellular trafficking, secretion, and vesicular transport, cell cycle control, cell division, chromosome partitioning, and cell motility. The relative abundances of cell wall/membrane/envelope biogenesis, ribosomal structure and biogenesis, inorganic ion transport and metabolism, intracellular trafficking, secretion, and vesicular transport, cell cycle control, cell division, chromosome partitioning, and cell motility were higher in SsV1, SsV2, and SsV3 than in the other groups.

3.3. CAZy Carbohydrate-Active Enzyme Annotation

The overall distribution of CAZy-encoded genes across treatment groups was analyzed (Figure 5). Figure 5 shows the 20 most abundant carbohydrate-active enzyme families. The dominant CAZy classes across treatment groups were glycosyltransferases (GTs), carbohydrate esterases (CEs), glycoside hydrolases (GHs), and auxiliary activities (AAs). Among these families, GT41 exhibited the highest proportion (annotated as peptide β-N-acetylglucosamine transferase and peptide N-β-glucosyltransferase), followed by GT4 (annotated as sucrose synthase), with the highest relative abundances of 7.04% and 6.44% in the CKn group, respectively. Following vermicompost application, the relative abundances of CE1 (acetylglucosaminidase), CE10 (arylesterase), GH94 (cellobiose phosphorylase), and GH24 (lysozyme) increased. Compared with CKn, CE1 (acetylglucosaminidase), GH2 (β-galactosidase), and GH94 (cellobiose phosphorylase) showed higher relative abundances in SsV1, SsV2, and SsV3. Meanwhile, CE9 (N-acetylglucosamine-6-deacetylase), GH3 (β-glucosidase), GT51 (murein polymerase), and AA3_2 (AA3 subfamily) were enriched in SnV1, SnV2, and SnV3. GT41 (β-N-acetylglucosamine transferase, peptide N-β-glucosyltransferase), AA7 (oligoglucose oxidase), CE4 (acetylmucopolysaccharide esterase), GH23 (G-type lysozyme), and AA1 (laccase/dihydroxybenzene oxidoreductase/iron oxidase) decreased after vermicompost application.

3.4. KEGG Functional Annotation

The primary KEGG functional categories were grouped into six major classes: Metabolism, Cellular Processes, Human Diseases, Organismal Systems, Genetic Information Processing, and Environmental Information Processing. Figure 6a presents a heatmap of the relative abundances of KEGG level-2 pathways. Pathways with relatively high abundances included carbohydrate metabolism, amino acid metabolism, and energy metabolism. Among these, carbohydrate metabolism and energy metabolism in CKn reached the highest relative abundances (11.2% and 5.7%, respectively), whereas amino acid metabolism in SsV1 showed the highest relative abundance (8.4%). Within the KEGG level-2 category “Xenobiotics biodegradation and metabolism,” the 16 most abundant degradation pathways were selected for differential analysis (Figure 6b). The most abundant pathways included benzoate degradation, drug metabolism—other enzymes, and aminobenzoate degradation. In SnV1, SnV2, and SnV3, pathways including drug metabolism—other enzymes, benzoate degradation, chlorocyclohexane and chloroalkane/chloroalkene degradation, drug metabolism—cytochrome P450, metabolism of xenobiotics by cytochrome P450, caprolactam degradation, atrazine degradation, and steroid degradation were higher than in CKn. In SsV1, SsV2, and SsV3, the relative abundances of cytochrome P450-mediated metabolism of exogenous drugs, caprolactam degradation, toluene degradation, atrazine degradation, and steroid degradation were higher than in CKn. In the atrazine degradation pathway, the relative abundance of atrazine degradation genes in the vermicompost-amended groups was higher than in CKn. Among these groups, SnV2 showed the highest relative abundance of atrazine degradation genes (2.94%), followed by SsV1 (2.71%). These values represent increases of 1.25 and 1.02 percentage points, respectively, relative to the CKn group. Overall, vermicompost application was associated with higher relative abundances of genes involved in atrazine degradation in soil. Moreover, the heatmap indicates that vermicompost application was associated with higher relative abundances of steroid degradation, cytochrome P450-mediated metabolism of exogenous drugs, and caprolactam degradation.

The gene abundance map of atrazine degradation pathway enzymes annotated by KEGG is shown in Figure 7a. Combined with the analysis results of the gene abundance pie chart for enzymes at each step of the atrazine degradation pathway (Figure 7b), the degradation metabolic pathway of atrazine in all treatments of this experiment was primarily the first atrazine chlorohydrolase pathway. Following vermicompost application, functional genes annotated by soil cyanuric acid aminohydrolase (3.5.2.15) (atzD), biuret hydrolase (3.5.2.15) (atzE), the functional gene atzF annotated by the urea hydroxylase (3.5.1.54), the functional genes uca, ureA, ureB, ureC, and ureAB annotated by the urease (3.5.1.5), and the functional gene URE annotated by the biuret hydrolase (3.5.1.54) urease (3.5.1.5) annotated functional genes URE, ureA, ureB, ureC, ureAB, and biuret hydrolase (3.5.1.84) annotated functional gene BiuH were all higher than in the CKn group. The functional abundance of atzB (annotated functional genes for hydroxyatrazine deethylaminohydrolase (3.5.4.43)), ureA, ureB, ureC, and atzF was relatively high in SsV1, SsV2, and SsV3. Therefore, the degradation of atrazine in vermicompost exhibited relatively high functional abundance. atzB, atzC (annotated functional genes for deisopropyl hydrolase (3.5.4.42)), and atzD exhibited the highest abundance in SsV2, while trzN (annotated functional genes for deisopropyl hydrolase (3.5.4.42)), atzF, uca, and ureC were most abundant in SsV1. URE, atzE, ureA, and ureB showed the highest abundance in SsV3. ureAB and DUR1 (annotated functional genes for urease carboxylase (6.3.4.6) and urea hydrolase (3.5.1.54)) were most abundant in SnV2.

3.5. Identification of Key Microorganisms in Atrazine Degradation Pathways and Their Functional Gene Contributions

Based on metagenomic functional annotation and host-tracing analysis of atrazine-degrading genes, this study identified core microbial genera that harbor and contribute key functional genes at each step of the complete mineralization pathway (Figure 8; Table S6). The analysis revealed that atrazine mineralization involved eight key enzymatic steps across all soils; however, the microbial hosts of the corresponding functional genes differed among treatments.

During the initial dechlorination step (Atrazine → Hydroxyatrazine), vermicompost application markedly altered the community of microbes harboring trzN and related genes. In SnV2, the abundance of gene sequences assigned to Bauldia increased from 0 RPKM in CKn to 4.84 RPKM, suggesting that Bauldia was a key host for initial dechlorination genes under 60-day vermicompost amendment. Similarly, in sterilized soil amended with 60-day vermicompost (SsV2), gene sequences traced to Mycolicibacterium rose from 0 to 3.27 RPKM, consistent with the introduction of exogenous taxa carrying early degradation genes. Gene sequences originating from Arthrobacter remained abundant across all groups, peaking at 6.41 RPKM in SnV3, supporting its persistent role as a host of early detoxification genes. At midstream stages of the pathway, different treatments favored distinct functional gene hosts. During cyanuric acid-to-biuret conversion, SnV3 showed the highest total gene abundance (17.03 RPKM), with sequences primarily assigned to Pseudorhodoplanes (12.14 RPKM). In the subsequent biuret-to-allophanate step, SsV2 showed the highest total host-traced gene abundance (61.20 RPKM), driven mainly by sequences associated with Panacagrimonas (11.09 RPKM) and Caenimonas (15.25 RPKM). Pseudorhodoplanes (9.50 RPKM) and Mycobacterium (4.05 RPKM) also contributed substantially, suggesting complementary functional hosting. During terminal mineralization (Allophanate → CO_2_), the abundance of gene sequences linked to Mycolicibacterium in SsV2 increased from 0.30 RPKM in CKn to 11.59 RPKM. Similarly, gene sequences assigned to Devosia peaked at 12.76 RPKM in SsV1 and were enriched in SnV2 (11.34 RPKM vs. 0.26 RPKM in CKn), supporting a potential role in the final mineralization steps.

3.6. Effects of Vermicompost on the Absolute Abundance of Atrazine Degradation Genes trzN, atzB, and atzC

The abundances of the atrazine degradation genes trzN, atzB, and atzC are shown in Figure 9. On days 10 and 20, trzN abundance in the vermicompost-amended groups was higher than that in CKn. On day 20, SnV2 showed the highest trzN abundance (9.03 × 10^4^ copies g^−1^), representing a 32-fold increase relative to CKn. On days 30 and 40, trzN abundance in SsV1, SnV2, and SnV3 was higher than that in CKn. Among these, SnV3 showed the highest trzN abundance, reaching 1.09 × 10^6^ and 4.17 × 10^5^ copies g^−1^ on days 30 and 40, respectively, corresponding to 4.9- and 1.6-fold increases relative to CKn. atzB abundance in the vermicompost-amended groups remained higher than that in CKn throughout the 0–40-day period. Except for SnV2, atzB abundance across treatments followed an initial decline, a subsequent increase, and a final decline. atzB abundance on day 20 was lower than that on day 10; however, SnV2 showed the highest atzB abundance on day 20 (3.25 × 10^7^ copies g^−1^). On day 30, SnV3 had the highest atzB abundance (8.75 × 10^7^ copies g^−1^), corresponding to a 4.1-fold increase relative to CKn. On day 40, SnV2 had the highest atzB abundance (6.95 × 10^7^ copies g^−1^), corresponding to a 5.0-fold increase relative to CKn. Except for SsV3, vermicompost-amended treatments showed higher atzC abundance than CKn from days 0 to 20. On day 30, SsV1, SsV2, SnV2, and SnV3 all showed higher atzC abundance than CKn, and on day 40, SnV2 and SnV3 remained higher than CKn. Among these, SnV3 showed peak atzC abundance on days 30 and 40, reaching 4.63 × 10^6^ and 2.67 × 10^6^ copies g^−1^, respectively, corresponding to 3.1- and 1.1-fold increases relative to CKn.

4. Discussion

This study systematically revealed the multi-level mechanisms by which vermicompost remediates atrazine-contaminated soil using quantitative PCR (qPCR) and metagenomic analysis. The results indicate that vermicompost promotes atrazine degradation by reshaping microbial community structure and, more importantly, by enriching microorganisms that harbor key functional genes and increasing the abundance and/or expression of degradation-related genes.

In the NR-based taxonomic annotation, vermicompost application significantly altered the soil microbial composition, notably increasing the relative abundances of phyla such as Proteobacteria, Bacteroidetes, and Firmicutes, which are widely recognized for pollutant-degrading capabilities [38,39,40,41]. For instance, the relative abundance of Proteobacteria reached 51.38% in SsV1, representing an 11.11 percentage-point increase relative to CKn. This shift was consistent with elevated copy numbers of degradation genes quantified by qPCR: trzN in SnV2 reached 9.03 × 10^4^ copies g^−1^ on day 20 (32-fold higher than CKn), and atzB and atzC in SnV3 reached 8.75 × 10^7^ and 4.63 × 10^6^ copies g^−1^ on day 30, respectively. These results indicate that, although vermicompost from different composting periods exhibited variation in its effects on microbial communities and functional genes, the treatments collectively showed an overall trend toward increasing the abundance of degradation-related microbial populations and increasing the abundance and/or expression of key genes. Importantly, host-tracing analysis of atrazine-degradation genes linked these genetic changes to specific microbial hosts. For example, the increase in trzN abundance in SnV2 corresponded to a rise in gene sequences assigned to Bauldia (4.84 RPKM), suggesting that this genus is a major host for genes involved in initial dechlorination. Similarly, the enrichment of Mycolicibacterium in SsV2 (3.27 RPKM) was consistent with the introduction of exogenous taxa carrying early-stage degradation genes. At the genus level, vermicompost significantly increased the relative abundances of Azotobacter, Devosia, Mycobacterium, and Microbacterium. Among these genera, Mycobacterium harbors functional genes such as atzB, atzC, and atzD, which participate in atrazine dealkylation and dechlorination. Mycobacterium plays a key role in atrazine metabolism and can degrade other triazine herbicides [42,43]. Microbacterium can degrade the organophosphorus pesticide chlorpyrifos, potentially by modulating the activities of cytochrome P450, glutathione S-transferase, catalase, and superoxide dismutase [44]. Devosia has been identified as a key microorganism in the co-metabolic or synergistic degradation of total petroleum hydrocarbons [45].

Metagenomic functional annotation (COG and CAZy) provided insights into how vermicompost creates a favorable microenvironment for degradative microorganisms. COG analysis showed that vermicompost was associated with higher relative abundances of functional categories related to carbohydrate transport and metabolism (G), lipid metabolism (I), secondary metabolite biosynthesis (Q), and secretion/vesicular transport. These functional shifts may facilitate organic matter decomposition, increase the availability of assimilable carbon and energy, and potentially support the synthesis and secretion of degradation enzymes. Correspondingly, CAZy analysis indicated that vermicompost was associated with increased abundances of enzyme families such as CE1 (acetylglucosaminyl esterase), CE10 (aromatic esterase), and GH94 (cellobiose phosphorylase). Although not directly involved in atrazine cleavage, these enzymes can contribute to the breakdown of complex soil organic compounds (e.g., lignin-derived substrates) into more readily available carbon sources, thereby supplying energy to the degrading community. This enhanced basal metabolism likely helps explain the sustained high abundances and/or expression of degradation genes during the later incubation phase (days 30–40) and aligns with the mechanism proposed by Virk et al. [23], in which organic amendments indirectly promote pesticide degradation by stimulating general microbial activity.

KEGG pathway analysis further characterized the effect of vermicompost on atrazine degradation routes. The relative abundance of genes assigned to the “Atrazine degradation” pathway within the “Xenobiotics biodegradation and metabolism” subcategory was higher in vermicompost-treated soils, peaking at 2.94% in SnV2 (an increase of 1.25 percentage points relative to CKn). Consistent with this pattern, vermicompost-treated soils showed a higher relative abundance of the atrazine degradation pathway within “Xenobiotics biodegradation and metabolism,” with SnV2 reaching 2.94% (+1.25 percentage points vs. the control; Figure 6b). Analysis of pathway-associated gene abundances (RPKM) indicated that vermicompost-treated soils showed higher abundances of downstream genes involved in atrazine degradation and mineralization (e.g., atzD, atzE, atzF, and uca). The enzymes encoded by these genes catalyze the conversion of cyanuric acid to NH_3_ and CO_2_, a critical step toward complete atrazine mineralization that may reduce the accumulation of toxic intermediates [46]. Notably, enzymes encoded by atzD, atzE, and atzF have been classified as glycoside hydrolases (GHs) [47]. Vermicompost-treated soils (Figure 5) exhibited higher GH abundance, supporting an association between broader soil metabolic potential and the abundance of key genes involved in atrazine degradation and dissipation pathways. Overall, downstream gene abundances were higher in vermicompost-treated groups than in CKn, with SnV2 showing the highest abundances of urease-related genes ureAB and DUR1. Host-tracing analysis indicated that these downstream genes were predominantly associated with genera such as Pseudorhodoplanes, Mycobacterium, and Devosia, which were consistently enriched in vermicompost treatments. This genetic structuring may help maintain the integrity of the degradation chain, providing a molecular basis for more complete atrazine detoxification in the environment [48,49].

Integrating gene–host attribution data across the entire degradation pathway provided a dynamic, function-resolved view of the microbial community. The 60-day vermicompost treatment (SnV2 and SsV2) showed particularly efficient functional coupling from upstream dechlorination to downstream mineralization. For example, at the initial step, SnV2 enriched Bauldia as a major host of trzN, whereas at the terminal mineralization step, Devosia and Mycolicibacterium were key hosts of allophanate- and urea-hydrolyzing genes, with gene-associated abundances increasing by more than an order of magnitude relative to CKn. This synchronized enrichment of gene-carrying microorganisms across consecutive metabolic steps may facilitate rapid channeling of atrazine and its intermediates toward complete mineralization, thereby reducing the accumulation of toxic metabolites such as deethylatrazine (DEA) and deisopropylatrazine (DIA) and lowering associated environmental risks.

Overall, vermicompost composted for 60 days exhibited the most consistent performance across multiple indicators: in unsterilized soil (SnV2), it resulted in the highest relative abundance of Proteobacteria, the greatest relative abundance of atrazine-degradation pathway genes (2.94%), and the highest copy numbers of trzN (day 20) and atzB (day 40). This optimal effect may be attributed to the balanced microbial community succession and metabolite profile achieved at this maturation stage: shorter composting (45 days) may not fully establish functionally important taxa, whereas longer composting (75 days) may deplete labile organic matter, thereby diminishing the stimulatory effect on degradation. This study provides a scientific rationale and technical guidance for designing vermicompost-based bioremediation strategies in atrazine-contaminated soils. Future research should prioritize field-scale validation and in situ implementation. Coupled models should be developed by integrating intermediate-metabolite flux data with core responsive genera identified in this study (e.g., Mycolicibacterium and Devosia) to enable molecular diagnostic tools for rapid field assessment of remediation efficacy. Furthermore, in-depth investigation of the long-term effects of vermicompost on native degradation-gene pools in farmland and their sustained contribution to microbial ecosystem stability will facilitate the development of durable, stable ecological remediation systems based on stimulation of endogenous potential.

5. Conclusions

This study systematically reveals the multidimensional mechanism by which vermicompost promotes efficient atrazine degradation through synergistic regulation of microbial composition and gene expression. Vermicompost enriched microbial communities associated with pollutant degradation, including Proteobacteria, Bacteroidetes, and Firmicutes, while increasing the abundance of functional bacterial genera such as Azotobacter, Devosia, Mycobacterium, and Microbacterium. Functional analysis indicates that vermicompost significantly enhances key processes in soil, including carbohydrate metabolism, secondary metabolite synthesis, and intracellular transport, providing the material and energy foundation for the degradation process. Vermicompost promoted the expression of functional genes associated with atrazine downstream mineralization (e.g., atzD, atzE, atzF), driving the complete conversion of atrazine into NH_3_ and CO_2_. Simultaneously, vermicompost demonstrated sustained degradation potential. The trzN gene rapidly responded in SnV2, reaching an expression level of 9.03 × 10^4^ copies·g^−1^ at an early stage (20 days). Later (40 days), atzB maintained high expression at 6.95 × 10^7^ copies·g^−1^, indicating sustained mineralization capacity. Collectively, these findings demonstrate that 60-day vermicompost exhibits well-coordinated microbial functions, enrichment of key genes, and sustained degradation capacity. This provides theoretical basis and technical reference for remediating atrazine contamination in agricultural fields.