Effect of Non-Antibiotic Pollution in Farmland Soil on the Risk of Antibiotic Resistance Gene Transfer

Abstract

The widespread use of antibiotics, combined with pervasive exposure to diverse environmental media, has intensified the global challenge of antibiotic resistance. Accumulating evidence reveals that beyond direct antibiotic pressure, residual non-antibiotic chemicals—despite lacking intrinsic antibacterial activity—can significantly promote the enrichment and spread of antibiotic resistance genes (ARGs) in farmland soils through indirect mechanisms such as inducing oxidative stress, altering microbial community structure, and enhancing both vertical and horizontal gene transfer. To address this issue, the present study investigates the influence of representative non-antibiotic contaminants commonly detected in agricultural environments—including pesticides (e.g., Omethoate, imidacloprid, and atrazine), industrial pollutants (e.g., PCB138, BDE47, benzo [a] pyrene, 2,3,7,8-tetrachlorodibenzo-p-dioxin [TCDD], and benzene), plastic-associated compounds (e.g., Polyethylene trimer, phthalates, and tributyl acetylcitrate), and ingredients from personal care products (e.g., triclosan and bisphenol A)—on ARG transmission dynamics. Leveraging bioinformatics resources such as the CARD database, PDB, AlphaFold, and molecular sequence analysis tools, we identified relevant small-molecule ligands and macromolecular receptors to construct a simulation system modeling ARG transfer pathways. Molecular docking and molecular dynamics (MD) simulations were then implemented, guided by a Plackett–Burman experimental design, to systematically evaluate the impact of individual and co-occurring pollutants. The resulting data were processed using advanced analytical tools, and MD trajectories were interpreted at the molecular level across three scenarios: an unperturbed (blank) system, single-pollutant exposures, and dual-pollutant combinations. By integrating computational simulations with machine learning approaches, this work uncovers the “co-selection” effect exerted by non-antibiotic chemical residues in shaping the environmental resistome, thereby providing a mechanistic and scientific basis for comprehensive risk assessment of agricultural non-point source pollution and the development of effective soil health management and antimicrobial resistance containment strategies.

1. Introduction

The extensive use of antibiotics—coupled with their introduction into soil ecosystems via aquaculture effluents and other pathways—has intensified the emergence of microbial resistance, posing a growing threat to cultivated soil health [1]. Antibiotic resistance genes (ARGs) are now recognized as “emerging contaminants” that share similarities with conventional soil pollutants in terms of biological origin, environmental persistence, and dissemination patterns [2]; however, they exhibit distinct pollution characteristics and underlying mechanisms that differentiate them from traditional soil contaminants [3]. Moreover, soil degradation—one of the most pervasive forms of land degradation globally—exerts profound and far-reaching effects on terrestrial ecosystems and biogeochemical cycling [4,5]. Consequently, understanding the capacity of ARGs to co-occur with and be transported alongside soil particles—and developing methods for their quantification and characteristic identification—has become a critical and timely research priority with significant practical implications.

In recent years, the risk of antibiotic resistance gene (ARG) dissemination in farmland soils contaminated with non-antibiotic chemicals has drawn increasing attention [6,7]. Accumulating evidence indicates that beyond antibiotics, a range of non-antibiotic contaminants—including pesticides, endocrine-disrupting compounds, and microplastics—can substantially enhance horizontal gene transfer (HGT) of ARGs. These substances exert their influence by altering soil physicochemical conditions, reshaping microbial community composition, and modulating the activity of mobile genetic elements (MGEs), thereby accelerating the environmental spread of antibiotic resistance [8]. Moreover, such pollutants not only possess intrinsic ecotoxicological hazards but may also amplify ARG transmission among microorganisms through multiple synergistic mechanisms, including co-selection, MGE-mediated mobilization, reactive oxygen species (ROS)-triggered stress responses, and enhanced biofilm formation [9,10].

Soil biological activity plays a pivotal role in the migration, transformation, and fate of pollutants, driven largely by soil organic matter (SOM), plant-derived inputs, and microbial metabolites [11]. SOM serves as a primary source of energy and carbon for microorganisms, stimulating their metabolic functions and thereby enhancing the biodegradation of contaminants [12]. Its decomposition also liberates essential nutrients that support plant growth [13]. In turn, plant vigor and physiological status influence soil biodiversity and ecosystem functionality, indirectly modulating the biodegradation efficiency and persistence of soil pollutants [14]. Root exudates—comprising low-molecular-weight compounds such as organic acids, amino acids, sugars, and phenolics, as well as high-molecular-weight polysaccharides and proteins—are key mediators in root–microbe–soil interactions [15]. These exudates not only nourish soil microbiota and enhance microbial diversity and activity but also modify soil physicochemical properties, thereby influencing nutrient cycling and plant nutrient acquisition [16]. Specific exudate components include (i) respiratory byproducts like CO₂ and water vapor released during root respiration [17,18]; (ii) organic acids (e.g., citric and malic acid) that solubilize soil minerals, increase nutrient bioavailability, and facilitate uptake [19,20]; (iii) phytohormones such as gibberellins and auxin-like substances that regulate the growth and development of neighboring organisms, including the host plant itself [21,22]; and (iv) structural polymers like polysaccharides that stabilize soil aggregates, improve porosity, and enhance water-holding capacity [23,24]. Furthermore, the balance and composition of soil anions and cations are fundamental to both plant nutrition and soil quality [25,26,27]. These ions participate in nutrient uptake, pH buffering, water retention, and cation-exchange capacity, collectively shaping soil reactivity and biological activity [28,29].

This paper mainly simulates and quantifies the characteristics and ability of ARGs’ vertical and horizontal gene transfer in farmland soil system, combs the relationship between ARGs’ vertical and horizontal gene transfer risk in soil environment and agricultural non-antibiotic pollution, and evaluates the contribution from soil associated agricultural pollution, which has important practical significance for further strengthening the control of ARGs’ pollution migration risk in soil environment.

2. Materials and Methods

2.1. Acquisition of ARG Multi-Dimensional Transfer System Construction Elements

2.1.1. ARG Multi-Dimensional Transfer System Construction Ligand Acquisition Card Database

The Environmental Antibiotic Resistance Gene Database (eARG-DB) [30], developed in 2024 by the Institute of Urban Environment, Chinese Academy of Sciences, and collaborating institutions, is a curated resource providing absolute abundance data for antibiotic resistance genes (ARGs) across diverse environmental matrices. It compiles quantitative profiles of 290 ARGs from over 1400 samples spanning 18 Chinese provinces, encompassing habitats such as soil and water, along with data on 30 mobile genetic elements and health risk-based ARG classifications. As a key reference for identifying high-abundance, high-risk ARGs in the environment, the database offers publicly accessible absolute abundance values and detailed metadata via its official platform, serving as an indispensable benchmark for assessing the background levels, spatial distribution, and transmission potential of ARGs. The exclusion criteria of high-abundance ARGs in soil refer to the relative abundance ranking of ARGs in previous studies and the absolute abundance ranking of ARGs in the NSDC database [31]. First, only ARGs ranking in the middle (30–60%) of both are selected as candidates. Secondly, the exclusion of highly hereditary ARGs was based on the ranking of ARGs’ and MGEs’ physical co-loading frequencies in previous studies, and the top 20% ARGs were removed. Finally, we queried the candidate ARGs in the gcpathogen database [32,33] and card database [34] and selected the ARGs with an MGE category ≤ 10 in the list of associated mobile components as the final research goal.

2.1.2. Construction of Receptor Acquisition Protein Database via ARG Multi-Dimensional Transfer System

The Protein Data Bank (PDB), originally established by Brookhaven National Laboratory, is a globally authoritative repository of experimentally determined 3D structures of biological macromolecules—primarily proteins and nucleic acids—solved by X-ray crystallography, NMR spectroscopy, and cryo-electron microscopy. It provides atomic coordinates, crystallographic details, and structural metadata for over 200,000 entries, serving as a cornerstone in structural biology and drug design. Through its official platform, users can access PDB-formatted files and analytical tools, enabling insights into molecular mechanisms and facilitating structure-based therapeutic development by linking macromolecular structure to biological function [35].

AlphaFold, an AI system developed by DeepMind, predicts 3D biomolecular structures from amino acid sequences. Its 2024 release, AlphaFold 3, integrates a generative diffusion model and an enhanced Pairformer module to achieve >50% higher accuracy than traditional methods, enabling reliable prediction of both individual proteins and, for the first time, complex biomolecular assemblies. Outputs include atomic coordinates and per-residue confidence scores, with nearly all known protein structures freely accessible via the AlphaFold Protein Structure Database. Trained on data from the Protein Data Bank (PDB), AlphaFold enriches the global structural knowledge base. Its open-source code, model weights, and online tools support worldwide academic research. The system has accelerated drug discovery—e.g., elucidating PHGDH mechanisms in Alzheimer’s disease—and collaborates with Isomorphic Labs to expedite therapeutic screening for cancer and other diseases while also showing potential in agriculture and materials science. Despite challenges in modeling dynamic conformations and side-chain positioning, AlphaFold has revolutionized the pipeline from basic research to clinical and industrial applications, offering a powerful AI-driven platform for addressing major health and sustainability challenges [36].

2.2. Simulation and Construction of ARG Multi-Dimensional Transfer Risk System: Molecular Docking Technology

In this study, molecular docking was employed to characterize the multi-dimensional transfer risk system of antibiotic resistance genes (ARGs) in soil environments. Molecular docking computationally predicts the binding conformation and stability of a ligand–receptor complex by evaluating geometric, energetic, and chemical complementarity between the ligand and the active site of the receptor protein, based on principles of spatial and energetic compatibility [37]. Specifically, 3D structures of small-molecule ligands and macromolecular receptors—representing vertical and horizontal ARG transfer potential in soil—were imported into Discovery Studio 2020 (DS2020). Putative binding cavities on the receptor proteins were identified using the Find Sites from Receptor Cavities function within the Define and Edit Binding Site module, followed by manual adjustment of the cavity radius. Docking simulations were then performed using the LibDock module to model ligand binding within the enzyme active sites of the macromolecular receptors. The docking protocol used user-defined preferences with Max Hits to Save set to 1 and the transformation method set to Best (default setting) [35]. Parameters in molecular docking were set as “Docking Preferences” to “User Specified”, “Max Hit for Save” to 10, “Max Hit to Save” to 1, “Max Number of Hits” to 100, “Final Score Cutoff” to 0.5, “Max BFGS Steps” to 50, “Max Conformation Hits” to 30, “Max Start Conformations” to 1000, “Steric Fraction” to 0.1, “Final Cluster Radius” to 0.5, “Apolar SASA Cutoff” to 15, “Polar SASA Cutoff” to 5, “Surface Grid Steps” to 18, “Conformation Method” to “BEST”, “Maximum Conformations” to 255, “Energy Threshold” to 10, and “Radius” to 9.

2.3. Construction of Farmland Soil Environmental Background Simulation System: Molecular Dynamics Simulation Method

Molecular dynamics (MD) simulation is a widely used computational approach that models the time-dependent behavior of ligand–receptor complexes based on molecular force fields, with binding energy serving as a quantitative indicator of interaction strength: lower (more negative) values denote stronger binding, while higher values indicate weaker affinity [38,39,40]. In this study, MD simulations were performed using GROMACS (Berendsen Laboratory, University of Göttingen, Germany) on a Dell PowerEdge R7425 server. The system under investigation comprised the “positive codon” sequence of antibiotic resistance genes (ARGs) bound to Tn5 plasmid transposase from E. coli, chosen to probe strategies for inhibiting horizontal gene transfer (HGT). Energy minimization was carried out via the steepest descent algorithm, followed by equilibration under constant pressure (1 bar, standard atmospheric conditions). Two groups were established: a blank control (no environmental perturbations) and an experimental group incorporating combined environmental stressors and agronomic management practices. Post-equilibration trajectories of the complex were sampled for binding free energy calculations using the Molecular Mechanics/Poisson–Boltzmann Surface Area (MM/PBSA) method, which separately evaluates the energies of the complex, receptor, and ligand to derive the net binding energy.

The following formulae were used for the calculation of the binding energy:

$$G_{bind}=G_{complex}-G_{free-protein}-G_{free-ligand}$$

(1)

where the binding energy of the molecule in solution is given by the following expression:

$$G=E_{gas}-T·S_{gas}+G_{solvation}$$

(2)

The solvent binding energy comprising polar and non-polar parts was calculated using the following expression:

$$G_{solvation}=G_{polar}+G_{nonpolar}$$

(3)

where $G_{bind}$ is the binding energy of the system, $G_{complex}$ is the binding energy of the complex, $G_{free-protein}$ is the binding energy of the enzyme, $G_{free-ligand}$ is the binding energy of the molecule, $G$ is the binding energy of the molecule in the solution, $T$ is the gas-phase temperature, $E_{gas}$ is the gas-phase energy, $S_{gas}$ is the gas-phase entropy, $G_{solvation}$ is the solvation free energy, $G_{polar}$ is the polar part contained in the solvent binding energy, and $G_{nonpolar}$ is the non-polar part contained in the solvent binding energy.

2.4. Verification Scheme of the Impact of Antibiotic Chemicals on ARG Transmission: Plackett–Burman Design

The Plackett–Burman (PB) design is a screening methodology suited for experiments with a modest number of factors, wherein higher-order (≥third-order) interactions are assumed negligible. This assumption simplifies the experimental layout, reduces the required number of runs, and efficiently identifies factors exerting significant effects on the response variable [41]. The resolution of a PB design determines its capacity to distinguish main effects from interactions; notably, a resolution V design enables clear separation of all main effects and two-factor interactions, thereby allowing for robust identification of both significant primary factors and critical second-order interactions influencing the dependent variable [36]. In this study, a resolution V Plackett–Burman design was implemented using Minitab 15.1 to screen soil pollution factors that significantly impact the vertical and horizontal transfer of antibiotic resistance genes (ARGs) in farmland soils. Each factor was coded dichotomously: “0” indicated the absence (or exclusion) of a given pollutant from the ARG transfer system, while “1” denoted its inclusion (or presence) in the experimental setup [42].

In this paper, we choose the following as agricultural pollution and soil ARGs for the soil environment: organochlorine pesticides: hexachlorocyclohexane (HCHs); organophosphorus pesticides: Omethoate; neonicotinoid pesticides: imidacloprid; other pesticides: atrazine; polychlorinated biphenyls: pcb138; polybrominated biphenyls: bde47; polycyclic aromatic hydrocarbons: benzo [a] pyrene (BAP); dioxins: 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD); petroleum hydrocarbons: benzene; microplastics: Polyethylene (trimer), phthalate, and acetyl tributyl citrate; personal care products: triclosan; and industrial chemicals: bisphenol A. Non-antibiotic pollution factors of agricultural soil are affected by the risk of gene transfer.

3. Results

3.1. Construction of Multi-Dimensional Transfer Risk System of ARGs in Soil Environment

3.1.1. Screening of Ligand Small Molecules in Multi-Dimensional Transfer Risk System of ARGs in Soil Environment

Based on the screening process for the database in Deng et al.’s study (2025) [43], we screened the ARGs with the highest abundance in the soil environment, namely tet (a) (www.ncbi.nlm.nih.gov/nuccore/ab114188.1, accessed on 4 May 2025), and the ARGs with the highest horizontal transfer correlation, namely mexb (https://card.mcmaster.ca/ontology/36517, accessed on 4 May 2025).

Based on the calculation of the above two types of ARGs by the codon sequence usage frequency and usage times calculation module in the sequence manipulation Suite (SMS) [44], it can be seen that among the two types of ARG codons, ATG and TGG codons rank first in the number of simulated uses and the frequency of simulated uses. Therefore, ATG was selected as the target ligand of the ARG transfer risk system in soil environment (Figure 1).

3.1.2. Multidimensional Transfer Risk Framework for ARGs in Soil Environments: Screening of Macromolecular Receptors

DNA gyrases not only play an important role in the ARGs’ vertical transfer system but also participate in the process of DNA replication, transcription, repair, and recombination in prokaryotes [45] (drlica and Zhao, 1997). DNA helicase can introduce a negative super helix into double-stranded circular DNA in the presence of ATP, which is composed of two subunits, one of which has the function of cutting DNA to form a gap and closing the gap, and the other has the function of hydrolyzing ATP to provide the energy required for the formation of a super helix [46] (Gellert, et al., 1976). In this paper, the simulated three-dimensional protein structure of Escherichia coli DNA gyrase constructed by Ren et al. [35] is used as the receptor macromolecule of the ARG vertical transfer system in soil environment, and its structure is shown in Figure 2.

In the process of ARGs’ horizontal gene transfer, both conjugation and transformation involve the separation of related genes from chromosomes, and then the mobile genetic elements represented by plasmids carry the horizontal gene transfer. Among them, the plasmid transposase of prokaryotes is a key functional enzyme that can paste or transfer related genes [47], which plays a very important role in the process of ARGs’ horizontal gene transfer. In this paper, Tn5 plasmid transposase of Escherichia coli was selected as the receptor macromolecule in the process of ARGs’ horizontal gene transfer. The amino acid sequence of Tn5 plasmid transposase of Escherichia coli was queried and extracted from NCBI data [48], and the ARGs’ gene sequence of Escherichia coli induced multiple drug resistance protein expression [49]. In this paper, the simulated three-dimensional structure of Tn5 plasmid transposase of Escherichia coli constructed by Ren et al. [3] is used as the receptor macromolecule of ARGs’ horizontal transfer system in the soil environment, and its structure is shown in Figure 3.

3.2. Construction of Multi-Dimensional Gene Transfer Simulation System of ARGs in Soil Environment

3.2.1. Construction of Soil Environment Background Simulation System

Combing the relevant literature data [50,51,52,53,54,55,56,57,58], including representative crops, soil temperature, soil pH, root exudates, root microorganisms, root microbial exudates, cation and anion strengths, and soil regional differences, the soil molecular dynamics simulation system was constructed according to the soil background parameters shown in

Table 1. Background parameters of soil environmental molecular dynamics simulation system.

Representative crops	Corn
Soil temperature	25 °C
Soil pH	7.0
Root exudates	Glucose, flavonoids, benzoxazosin, ethylene, oxalic acid, gibberellin, and carbon dioxide
Root microorganisms	Agrobacterium, Rhizobium, and proteus
Root microbial exudates	Agrobacterium	Octopine, formic acid, and indole
	Rhizobium	Flavonoids, ammonia, and ammonium
	Proteus	Indole, ammonia, and ammonium
Cation and anion strengths	Cation	K+, Na+, Ca2+, Mg2+, Fe3+, Fe2+, NH4+, and Mn2+
	Anion	NO3−, SO42−, PO43−, CO32−, OH−, and Cl−

.

3.2.2. Construction of Soil Simulation System Based on Agricultural Soil Non-Antibiotic Pollution Scenario

Using Minitab software, the Plackett–Burman design method with resolution V was used in this paper. B (organophosphorus pesticide): Omethoate; C (neonicotinoid pesticide): imidacloprid; D (other pesticides): atrazine; E (polychlorinated biphenyls): pcb138; F (polybrominated biphenyls): bde47; G (polycyclic aromatic hydrocarbons): benzo [a] pyrene (BAP); H (dioxin): 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD); I (petroleum hydrocarbon): benzene; G (microplastic): Polyethylene (trimer), K (phthalate ester), and L (acetyl tributyl citrate); M (personal care products): triclosan; and N (industrial chemicals): bisphenol A, which were used as non-antibiotic pollution factors that have a synergistic impact on the risk of multi-dimensional gene transfer of ARGs in the soil environment, produced a total of 64 groups of molecular dynamics simulation schemes (

Table 2. Soil simulation experiment scheme under typical non-antibiotic chemical pollution scenario of farmland soil.

Transfer Type	No.	A	B	C	D	E	F	G	H	I	J	K	L	M	N
Vertical gene transfer	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1
	2	1	1	1	1	0	1	1	1	1	0	0	0	0	0
	3	1	1	1	0	1	1	0	0	0	1	1	1	0	0
	4	1	1	1	0	0	1	0	0	0	0	0	0	1	1
	5	1	1	0	1	1	0	1	0	0	1	0	0	1	1
	6	1	1	0	1	0	0	1	0	0	0	1	1	0	0
	7	1	1	0	0	1	0	0	1	1	1	0	0	0	0
	8	1	1	0	0	0	0	0	1	1	0	1	1	1	1
	9	1	0	1	1	1	0	0	1	0	0	1	0	1	0
	10	1	0	1	1	0	0	0	1	0	1	0	1	0	1
	11	1	0	1	0	1	0	1	0	1	0	1	0	0	1
	12	1	0	1	0	0	0	1	0	1	1	0	1	1	0
	13	1	0	0	1	1	1	0	0	1	0	0	1	1	0
	14	1	0	0	1	0	1	0	0	1	1	1	0	0	1
	15	1	0	0	0	1	1	1	1	0	0	0	1	0	1
	16	1	0	0	0	0	1	1	1	0	1	1	0	1	0
	17	0	1	1	1	1	0	0	0	1	0	0	1	0	1
	18	0	1	1	1	0	0	0	0	1	1	1	0	1	0
	19	0	1	1	0	1	0	1	1	0	0	0	1	1	0
	20	0	1	1	0	0	0	1	1	0	1	1	0	0	1
	21	0	1	0	1	1	1	0	1	0	0	1	0	0	1
	22	0	1	0	1	0	1	0	1	0	1	0	1	1	0
	23	0	1	0	0	1	1	1	0	1	0	1	0	1	0
	24	0	1	0	0	0	1	1	0	1	1	0	1	0	1
	25	0	0	1	1	1	1	1	0	0	1	0	0	0	0
	26	0	0	1	1	0	1	1	0	0	0	1	1	1	1
	27	0	0	1	0	1	1	0	1	1	1	0	0	1	1
	28	0	0	1	0	0	1	0	1	1	0	1	1	0	0
	29	0	0	0	1	1	0	1	1	1	1	1	1	0	0
	30	0	0	0	1	0	0	1	1	1	0	0	0	1	1
	31	0	0	0	0	1	0	0	0	0	1	1	1	1	1
	32	0	0	0	0	0	0	0	0	0	0	0	0	0	0
Horizontal gene transfer	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1
	2	1	1	1	1	0	1	1	1	1	0	0	0	0	0
	3	1	1	1	0	1	1	0	0	0	1	1	1	0	0
	4	1	1	1	0	0	1	0	0	0	0	0	0	1	1
	5	1	1	0	1	1	0	1	0	0	1	0	0	1	1
	6	1	1	0	1	0	0	1	0	0	0	1	1	0	0
	7	1	1	0	0	1	0	0	1	1	1	0	0	0	0
	8	1	1	0	0	0	0	0	1	1	0	1	1	1	1
	9	1	0	1	1	1	0	0	1	0	0	1	0	1	0
	10	1	0	1	1	0	0	0	1	0	1	0	1	0	1
	11	1	0	1	0	1	0	1	0	1	0	1	0	0	1
	12	1	0	1	0	0	0	1	0	1	1	0	1	1	0
	13	1	0	0	1	1	1	0	0	1	0	0	1	1	0
	14	1	0	0	1	0	1	0	0	1	1	1	0	0	1
	15	1	0	0	0	1	1	1	1	0	0	0	1	0	1
	16	1	0	0	0	0	1	1	1	0	1	1	0	1	0
	17	0	1	1	1	1	0	0	0	1	0	0	1	0	1
	18	0	1	1	1	0	0	0	0	1	1	1	0	1	0
	19	0	1	1	0	1	0	1	1	0	0	0	1	1	0
	20	0	1	1	0	0	0	1	1	0	1	1	0	0	1
	21	0	1	0	1	1	1	0	1	0	0	1	0	0	1
	22	0	1	0	1	0	1	0	1	0	1	0	1	1	0
	23	0	1	0	0	1	1	1	0	1	0	1	0	1	0
	24	0	1	0	0	0	1	1	0	1	1	0	1	0	1
	25	0	0	1	1	1	1	1	0	0	1	0	0	0	0
	26	0	0	1	1	0	1	1	0	0	0	1	1	1	1
	27	0	0	1	0	1	1	0	1	1	1	0	0	1	1
	28	0	0	1	0	0	1	0	1	1	0	1	1	0	0
	29	0	0	0	1	1	0	1	1	1	1	1	1	0	0
	30	0	0	0	1	0	0	1	1	1	0	0	0	1	1
	31	0	0	0	0	1	0	0	0	0	1	1	1	1	1
	32	0	0	0	0	0	0	0	0	0	0	0	0	0	0

Note: A (organochlorine pesticides): hexachlorocyclohexane (HCHs); B (organophosphorus pesticide): Omethoate; C (neonicotinoid pesticide): imidacloprid; D (other pesticides): atrazine; E (polychlorinated biphenyls): pcb138; F (polybrominated biphenyls): bde47; G (polycyclic aromatic hydrocarbons): benzo [a] pyrene (BAP); H (dioxin): 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD); I (petroleum hydrocarbon): benzene; J (microplastic): Polyethylene (trimer); K: phthalate esters (PAEs); L: acetyl tributyl citrate; M (personal care products): triclosan; N (industrial chemicals): bisphenol A.

) to characterize the synergistic impact of multi-dimensional gene transfer risk of ARGs in the soil environment and non-antibiotic pollution in the soil, so as to determine the soil non-antibiotic pollution factors that significantly affect the vertical and horizontal transfer of ARGs in the soil environment.

3.3. Impact and Analysis of Farmland Typical Non-Antibiotic Pollution on Multi-Dimensional Transfer Risk of Soil ARGs

3.3.1. Impact of Typical Non-Antibiotic Pollution in Farmland on Multi-Dimensional Transfer Risk of ARGs in Soil Environment

See

Table 3. Simulation calculation of the impact of ARGs multi-dimensional transfer risk under the typical non-antibiotic pollution scenario in farmland.

No.	Vertical Gene Transfer			Horizontal Gene Transfer
	Blank Binding Energy (kJ/mol)	Binding Energy (kJ/mol)	Change	Blank Binding Energy (kJ/mol)	Binding Energy (kJ/mol)	Change
1	−22.694	−43.028	89.60%	−47.274	−46.957	40.78%
2		−38.655	70.33%		−55.138	65.30%
3		−9.444	−58.39%		−49.275	47.72%
4		−72.870	221.10%		−44.975	34.83%
5		−62.343	174.71%		−34.626	3.81%
6		−63.701	180.70%		−34.346	2.97%
7		−71.077	213.20%		−64.976	94.80%
8		−37.160	63.74%		−63.274	89.69%
9		−52.038	129.30%		−16.195	−51.45%
10		−15.451	−31.92%		−49.059	47.08%
11		−124.481	448.52%		−59.302	77.79%
12		−23.108	1.82%		−49.651	48.85%
13		−36.270	59.82%		−37.989	13.89%
14		−22.999	1.34%		−39.873	19.54%
15		−40.199	77.13%		−49.651	48.85%
16		−50.930	124.42%		−45.970	37.82%
17		−17.068	−24.79%		−74.340	122.87%
18		−44.257	95.02%		−46.238	38.62%
19		−22.570	−0.55%		−31.828	−4.58%
20		−58.712	158.71%		−53.250	59.64%
21		−13.653	−39.84%		−76.013	127.88%
22		−14.676	−35.33%		−55.926	67.66%
23		−3.577	−84.24%		−35.859	7.50%
24		−16.902	−25.52%		−48.628	45.78%
25		−36.157	59.32%		−63.640	90.79%
26		−10.567	−53.44%		−67.496	102.35%
27		−41.357	82.24%		−79.234	137.54%
28		−65.389	188.13%		−64.556	93.54%
29		−26.420	16.42%		−31.257	−6.29%
30		−18.461	−18.65%		−34.200	2.53%
31		−48.542	113.90%		−52.880	58.53%
32 blank		−22.694	-		−47.274	-

for the simulation calculation results of the impact of farmland typical non-antibiotic pollution on the multi-dimensional transfer risk of soil ARGs.

3.3.2. Identification and Functional Classification of Key Binding Enhancement Factors

Vertical Gene Transfer

Through the systematic analysis of several influencing factors, this paper evaluated the influence of each factor on the binding performance of the molecular system, focusing on the change in average binding energy and the improvement of the proportion of “strong binding” conformation. The comprehensive results showed that there were significant differences in the regulation effect of different factors on binding strength, which could be divided into high-efficiency enhancement factor, suboptimal enhancement factor, and weak or unfavorable factors according to their influence.

HCHs, the representative of organochlorine pesticides, showed the most prominent enhancement effect. After introducing this factor into the system, the average binding energy decreased by about 25 kJ/mol, indicating that the overall binding capacity was significantly enhanced. More importantly, the proportion of “strong binding” state increased from 12.5% under the baseline condition to 37.5%, with an increase of 25%, which was the largest increase among all factors, showing its key role in stabilizing the favorable conformation. Imidacloprid, the representative of neonicotinoid pesticides, and PAEs showed similar and stable promoting effects. Both of them can reduce the average binding energy by about 19.5 kJ/mol and increase the proportion of strong binding conformations by 12.5%. The effects of pcb138 and bisphenol A on the system were relatively limited. Both of them can reduce the average binding energy by 17.5–17.8 kJ/mol, showing a slight trend of binding stabilization, but further analysis found that the proportion of “strong binding” conformation did not change significantly. The other factors showed different performances. Some factors led to the increase of average binding energy, reflecting the weakening effect on the binding process; Others have little influence on the energy distribution or conformational preference of the system.

In conclusion, hexachlorocyclohexane (HCHs) has the most significant effect in improving the binding strength and optimizing the conformation distribution, which is the core optimization direction; imidacloprid and PAEs, the representatives of neonicotinoid pesticides, can be used as effective supplementary strategies. Although pcb138 and bisphenol A have a slight stabilizing effect, their contribution to the key high-affinity state is limited.

Horizontal Gene Transfer

Through the systematic analysis of several potential influencing factors, this paper evaluated the contribution of each factor in regulating the binding strength of molecular system, focusing on the change in average binding energy and the promotion of the proportion of “strong binding” conformation. The results showed that there were significant differences in the effects of different factors on the binding performance, which could be clearly divided into three categories: significantly enhanced factors, secondary favorable factors, and unfavorable factors that weaken the binding.

Bisphenol A, bde47, and imidacloprid showed strong binding-promoting effects. When the three compounds were introduced into the system, the average binding energy decreased by about 7.3 kJ/mol (bisphenol A), 6.6 kJ/mol (bde47), and 5.9 kJ/mol (imidacloprid), respectively, showing a strong stabilizing effect. More importantly, the proportion of “strong binding” conformation increased from 18.8% of the baseline level to 31.3%, an increase of 12.5%, indicating that the three significantly increased the frequency of the high-affinity binding state. This advantage suggests that bisphenol A, bde47, and imidacloprid may effectively lock in favorable conformations by optimizing key interactions (such as hydrogen bond, π-stacking, or hydrophobic effect), which are the core structural features to improve binding affinity. Although the enhancement effects of benzene and 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) are relatively weak, they still show a clear positive contribution. Among them, benzene can reduce the average binding energy by about 4.7 kJ/mol and 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) by about 3.8 kJ/mol. Although the energy decrease was less than that of bisphenol A, bde47, and imidacloprid, it is worth noting that they also increased the proportion of strong binding conformations by 12.5%, indicating that their role is not limited to weak energy adjustment but rather can guide the system to a more favorable binding mode to a certain extent. This may be due to their moderate regulation of local conformational flexibility or solvation effects. Therefore, although the effect of benzene and 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) alone is limited, as an auxiliary optimization method, it may play a cumulative or synergistic effect when combined with other powerful factors, which has the potential for further exploration. In the lower group of factors, HCHs, Omethoate, atrazine, and PCB138 are significant. Many factors, such as polythene (trimer), tributyl acetylcitrate, triclosan, etc., lead to an increase in the average binding energy (Δ > 0), indicating that they tend to destroy or interfere with the effective interaction between a ligand and receptor, thus weakening the overall binding ability.

In conclusion, bisphenol A, bde47, and imidacloprid were identified as the core binding enhancers, which had the dual advantages of significantly reducing the binding energy and increasing the proportion of strong binding conformations; benzene and 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), as secondary favorable factors, can be used as supplementary optimization options. HCHs, Omethoate, atrazine, PCB138, Polythene (trimer), tributyl acetylcitrate, triclosan, etc., generally weaken the combination, which should be treated with caution.

3.4. Molecular Dynamics Mechanism Analysis Based on Typical Analysis Results

3.4.1. Vertical Gene Transfer

As shown in the Figure 4 and Figure 5, the types and intensity of non-covalent interactions of the three groups of systems have changed systematically. The blank system is dominated by conventional hydrogen bonds and pi anions, and the total number of bonds is about 30. After adding HCHs, the total number of interaction bonds increased to about 35, and strong interactions such as bump and charge were added, but some of the original effects were weakened; After the introduction of HCHs and imidacloprid, the total number of bonds was slightly reduced to about 34, but the bond length of the key role was generally shortened, indicating that the combination was tighter, and the structure was more optimized. The overall trend showed that HCHs triggered the network reconstruction, while imidacloprid achieved collaborative optimization and stability.

Among all kinds of non-covalent effects, there were five conventional effects in the blank system of vertical gene transfer (bond length: 2.33–2.65 Å, average ≈ 2.54 Å). After the introduction of HCHs, the number increased to eight (bond length: 1.73–3.08 Å, average ≈ 2.67 Å), and the number increased, but the average bond length increased, indicating that the action intensity decreased slightly. However, after the introduction of HCHs, the addition of C still maintained eight (bond length: 1.80–2.93 Å, with an average of ≈2.49 Å), and the average bond length was shortened by 0.18 Å, indicating that C significantly enhanced the intensity of such effects. Cardon acts on seven in the blank (bond length: 1.67–2.97 Å, average ≈ 2.37 Å). After the introduction of HCHs, the number increased to eight (bond length: 2.35–3.06 Å, average ≈ 2.71 Å), and the bond length increased significantly. After the introduction of imidacloprid, it was reduced to five (bond length: 2.43–3.70 Å, average ≈ 3.03 Å), and the effect continued to weaken. There were two effects of pi alkyl in the blank (5.02 and 5.32 μg). After the introduction of HCHs, only one (4.49 Å) remained, and the bond length was shortened, but the number was reduced. After the introduction of imidacloprid, the number increased to three (4.53, 4.87, and 5.01 Å, with an average of ≈4.80 Å), and the number and intensity tended to balance. There was one pi donor (2.60 Å) in the blank, which disappeared completely after the introduction of HCHs. After the introduction of imidacloprid, it recovered to one, but the bond length increased to 3.06 Å, and the effect became weaker. There were six pi anions in the blank (2.83–4.84 Å, average ≈ 3.96 Å), which were not recorded after the introduction of HCHs and imidacloprid, indicating that the effect was completely inhibited. Acceptor–acceptor appeared only once in the blank (2.74 Å) and disappeared and could not be recovered after the introduction of HCHs. In contrast, the effect of bump was not recorded in the blank. After the introduction of HCHs, there were three (0.97, 0.97, and 1.13 Å, with an average of ≈1.02 Å), and after the introduction of imidacloprid, it increased to four (0.94–1.14 Å, with an average of ≈1.09 Å) and maintained a very short bond length, indicating a strong van der Waals contact. Charge–charge also appeared only after the introduction of HCHs (4.16 Å), increased to three after the introduction of imidacloprid (3.59–4.45 Å, average ≈ 3.96 Å), doubled the number, and shortened the average bond length by 0.20 Å, and the electrostatic effect was significantly enhanced. There were two pi cations in the blank (3.49 and 3.60 Å, average ≈ 3.55 Å); After the introduction of HCHs, there was only one (4.87 µm), and the effect was significantly weakened; After the introduction of imidacloprid, it recovered to two (3.80 and 4.03 Å, with an average of ≈3.91 Å), and the original strength was not reached, but it was better than the introduction of HCHs. Pi-pi-t-shaped appeared only once (4.83 Å) when HCHs were introduced and disappeared after adding C. Pi cation and pi donor appeared twice (3.60 and 3.87) when adding a and completely disappeared after the introduction of imidacloprid. However, only one new pi sigma (3.95 μm) appeared after adding C, which was a new stability factor for the system.

The role of HCHs can be summarized as “activation and reconstruction”. Through the introduction of charge–charge electrostatic interaction and tight van der Waals contact (bump), HCHs significantly changed the intermolecular interaction pattern but, at the same time, destroyed the original π-anion, receptor–receptor, and other stability effects. In contrast, imidacloprid plays the role of “optimization and stability”: on the basis of the new framework constructed by HCHs, imidacloprid makes the system tend to a state of lower energy and stronger binding by shortening the key bond length, restoring partial π interactions, inhibiting unstable conformations, and introducing new functions such as pi sigma. The combination of HCHs and imidacloprid is not a simple superposition but rather a dynamic equilibrium with synergistic effects.

From the perspective of the force evolution path, the system has experienced a transition from “basic hydrogen bond and π interaction dominated” to “charge and van der Waals driven” and then to “multi-type collaborative stability”. As a reconstructing agent, HCHs, although sacrificing some of their original functions, activated stronger electrostatic and stacking effects; imidacloprid, as an optimization agent, screened and strengthened the favorable interaction and eliminated the unstable mode. This collaborative strategy of HCHs + imidacloprid significantly improved the binding efficiency, especially in bump, charge–charge, and conventional. Therefore, in molecular recognition or drug design, if the goal is to enhance the binding affinity, the combination of HCHs and imidacloprid may be better than a single modification.

In summary, the introduction of HCHs triggered the reprogramming of the intermolecular interaction network, enhanced the charge interaction and van der Waals contact, but weakened some π correlation effects. The introduction of imidacloprid after HCHs not only retained the advantages brought by HCHs but also significantly improved the overall bonding stability by enhancing key interactions, restoring partial π effects, and introducing new stability mechanisms (such as pi sigma). Therefore, HCHs and imidacloprid have a clear synergistic effect, which jointly promotes a more efficient and compact molecular recognition process, especially suitable for the system design requiring strong binding force.

3.4.2. Horizontal Gene Transfer

In the blank system, the intermolecular non-covalent interactions were dominated by conventional hydrogen bonds, salt bridges (charge–charge), and π-anions (pi anion), forming a total of 38 interaction bonds with relatively stable and diversified structures. After the introduction of bisphenol A, the total number of bonds decreased to 34, and some of the original effects (such as amide pi stacked and charge–charge) disappeared or decreased, but bump and pi cation were added; pi donor and other new interactions indicate that the system has significant interaction network reconstruction. After imidacloprid was further introduced, the total number of bonds decreased slightly to 32. Although some types of actions were weakened or eliminated, the average bond length of key actions (such as conventional and bump) was generally shortened, indicating that the bonding interface tended to be tight and optimized. As a whole, there is a collaborative regulation path of “bisphenol A causes changes and imidacloprid strengthens” (Figure 6 and Figure 7).

In the aspect of conventional hydrogen bonds, the number decreased from 10 to 9 and then to 6, but the bond length distribution showed that when imidacloprid was introduced after the introduction of bisphenol A, multiple bond lengths were lower than those in the blank group (such as 1.82 Å and 1.91 Å), indicating that the quality of remaining hydrogen bonds was higher. The number of cardon actions remained stable (5 → 5 → 6), but the bond length continued to increase (up to 3.76 µm), suggesting that the role of cardon was weakened and may be affected by steric hindrance. After the introduction of bisphenol A, the bond length of pi alkyl was shortened to 4.49 μm but completely disappeared after the introduction of imidacloprid, reflecting its sensitivity to the environment; The number of pi anion interactions fluctuated (5 → 4 → 6), but most bond lengths increased (up to 5.57 µm), indicating that the interactions existed but weakened. The positive–positive effect only appeared in the introduction of imidacloprid after the introduction of bisphenol A (4.14 Å), reflecting the establishment of a new electrostatic mode. The bump effect appeared for the first time after the introduction of bisphenol A (minimum bond length 0.97 μm) and was maintained or even slightly enhanced after the introduction of imidacloprid, which was the key contributor of high affinity. After the introduction of bisphenol A, the bond length of pi-pi-t-shaped was prolonged (4.83 Å), and after the introduction of imidacloprid, it was retracted to 4.68 Å, indicating that imidacloprid has the ability to repair. Amide pi stacked disappeared completely after the introduction of bisphenol A, indicating that the effect was highly sensitive to factor a. The pi cation and pi donor effect only existed when bisphenol A was introduced and completely disappeared after the introduction of imidacloprid, indicating that imidacloprid inhibited the recognition of such complex types.

Bisphenol A, as a network reconstruction agent, expands the possibility space of molecular recognition by disrupting the original equilibrium (such as eliminating Amide Pi stacked and weakening Charge Charge), introducing new interaction types (such as Bump, Pi Combination, Pi Denor), but at the cost of sacrificing some stability. Its function has the characteristics of “breaking rebuilding”. Imidacloprid behaves as a binding optimizer: it does not introduce a new type of action, but through fine-grained regulation of existing actions, such as shortening the length of conventional and bump bonds, it restores the number of pi anions, inhibits unstable composite actions, and introduces positives to achieve energy optimization and structural refinement of the binding interface. Under the coordination of the two, the system evolves from “many and miscellaneous” to “few and strong”, reflecting the evolution direction of efficient identification.

The number of non-covalent interactions is not the only index to determine the bonding strength, and the quality of the force (i.e., bond length) is more critical. Although the total bond number of the imidacloprid system after the introduction of bisphenol A was the least (32), its key role (such as bump bond length ≤ 1.13 Å and multiple conventional bond length < 2.0 Å) was much stronger than the other two groups. For example, the bump effect in the introduction of imidacloprid after the introduction of bisphenol A is in a very short range (0.96–1.13 Å), representing a strong van der Waals extrusion, which significantly improves the binding rigidity and stability. At the same time, imidacloprid can effectively inhibit ineffective or weakened effects (such as charge, pi cation, and pi donor), avoiding energy waste. This suggests that in molecular design, priority should be given to the construction of high-quality and high-specific interactions rather than blindly increasing the type or number of interactions.

In conclusion, bisphenol A and imidacloprid showed clear functional complementarity and dynamic synergy in the regulation of molecular recognition. Bisphenol A widened the recognition dimension by reconstructing the interaction network, and imidacloprid screened and strengthened the optimal interaction on this basis and finally achieved the optimization and stability of the binding interface. Although the total number of interaction bonds of the imidacloprid system after the introduction of bisphenol A is the lowest, it shows higher binding efficiency and potential affinity due to its stronger key role and more compact structure. This mechanism provides an important paradigm for rational design of multi-step and multi-factor molecular recognition system: firstly, the “disturbance factor” (bisphenol A) is used to activate diversity, and then the “optimization factor” (imidacloprid) is used to focus on high-value interactions, so as to realize the transition from “binding” to “strong binding”.

4. Conclusions

This paper employs molecular docking and molecular dynamics simulations to explore the potential influence of 14 representative non-antibiotic pollutants commonly detected in farmland soils on the structural stability and binding affinity of a key conjugative transfer-related protein. The results indicate that certain pollutants—particularly persistent organic pollutants such as 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), benzo [a] pyrene (BaP), and PCB138—may significantly alter intermolecular interaction types (e.g., hydrogen bonds and hydrophobic contacts) and reduce binding free energy, suggesting the potential for these compounds to modulate horizontal gene transfer processes at the molecular level. In contrast, other pollutants (e.g., benzene and atrazine) exhibit relatively minor effects under the simulated conditions. This paper emphasizes that these findings are derived from in silico analyses and reflect theoretical protein–pollutant interactions rather than direct evidence of enhanced antibiotic resistance gene (ARG) transfer in real environmental systems. Nevertheless, this paper provides a mechanistic hypothesis: specific non-antibiotic contaminants may act as indirect facilitators of ARG dissemination by perturbing the conformational dynamics or binding capacity of transfer-associated proteins. Such insights could help prioritize high-risk pollutants for future experimental validation and environmental monitoring. While this paper does not propose specific mitigation strategies, it highlights the need to consider non-antibiotic chemical stressors—beyond antibiotics—in holistic approaches to antimicrobial resistance management. This paper suggests that reducing uncontrolled inputs of persistent organic pollutants into agricultural soils may complement existing stewardship efforts and contribute to more sustainable and resilient soil health practices.