Land-Use and Depth-Dependent Assembly of Soil Microbiomes Shapes Ecological Functions, Interaction Networks, and Pathogen Communities Across Crop and Orchard Systems

1. Introduction

Maximizing agricultural production is a primary objective for farmers today. However, it is crucial to achieve this goal while prioritizing human health, which relies on the nutritional quality of food. Modern food production heavily relies on the extensive use of fertilizers, various tillage systems, and other treatments, all of which significantly impact soil health. Studies indicate that different forms of soil treatment have a direct impact on soil health and productivity, influencing changes in its microbial composition [1]. Soil microorganisms play a crucial role in ecosystems by cycling nutrients, stabilizing soil, and aiding plant growth. They enhance disease resistance, regulate soil pH, boost soil organic carbon levels, and compete effectively against pathogens [2,3].

Soil microbial composition is primarily influenced by crops that are planted. For instance, orchards show a significant difference in soil composition compared to croplands. According to a study by Cheng et al. [1], orchard-planted soils have higher microbial diversity and impact the preservation of soil structure, especially in its surface layers However, the continuous cultivation of fruit trees as monocultures has been shown to cause replant syndrome issues. This disease, resulting from the repeated planting of the same crops in the same field, has led to reduced yields, hindered growth and development of plants, and affected their nutritional composition [4]. This occurs because the pathogenic members of the soil community become more robust, increasing their resistance to agrochemicals, and acting synergistically to cause such problems [5]. Monocultures decrease microbial diversity and increase pathogen resistance to various agrochemicals. This resistance can be transferred through gene transfer, directly impacting soil productivity, and leading to the development of resistant strains. Soils planted with monocultures or fruit trees have higher levels of antibiotic resistance due to the more favorable conditions for microorganisms in these undisturbed environments, facilitating their easier spread [6]. Cover crops such as alfalfa, buckwheat, and vetch are an environmentally friendly solution with desirable outcomes in supporting soil health. They promote the development of beneficial microorganisms, including those that solubilize phosphate, fix nitrogen, and others, thereby contributing to the enhancement of soil health and the promotion of microbiological growth [7]. Crop rotation has also demonstrated its significant impact on carbon sequestration and the formation of soil structure, which are essential for supporting plant growth and enhancing nutrient availability [8]. The practice of rotating different crops in agricultural soil disrupts the life cycles of pathogenic microorganisms by creating less favorable conditions for their growth and proliferation. This aids to break the cycle of pests, diseases, and pathogens that may build up when the same crop is grown repeatedly in the same soil. By introducing crop rotation, the need for agrochemicals application is reduced, thereby supporting sustainable agriculture practices.

Soil microbial composition is influenced by soil properties as well. The topography of agricultural land directly impacts microbial composition through the nutrients provided to plants and the available water content. For instance, fungal levels are decreased in lower topographic areas compared to higher ones because fungi adapt more easily to challenging environments with limited nutrient availability [9]. Due to less favorable temperatures, nutrient stress and variations in soil characteristics, the microbial community, especially bacteria, thrives more abundantly at lower elevations and demonstrates a higher capability for decomposing organic carbon [10]. Moreover, such differences are obvious in different soil layers, where deeper layers are related with more restrict living conditions for microbes. This impacts a higher microbial diversity in higher layers of soil compared to deeper layers [11,12]. Such differences affect the physico-chemical parameters of soil, shaping the adaptation of microorganisms in specific soils.

In this study, we investigated the soil microbiome across contrasting agricultural land-use systems in the Hajdúnánás region of Hungary, including crop fields, orchards, alfalfa, and uncultivated soils. Using shotgun metagenomic sequencing combined with soil physicochemical analyses, we aimed to assess how land use, soil depth, and environmental conditions influence microbial diversity, community composition, ecological functions, microbial interactions, and pathogen distribution. Specifically, this study aimed to: (i) characterize the diversity and composition of bacterial and fungal communities across contrasting land-use systems and soil depths; (ii) evaluate the relationships between microbial communities and soil physicochemical properties; (iii) investigate the functional potential and interaction networks of soil microorganisms; and (iv) assess the distribution, prevalence, and environmental drivers of plant-associated bacterial and fungal pathogens. By integrating taxonomic, functional, network, and pathogen analyses, this study provides a comprehensive assessment of the ecological mechanisms governing soil microbiome assembly and functioning in agricultural ecosystems.

2. Materials and Methods

2.1. Sampling

The experiment was conducted in the Hajdúnánás region, Hungary (N 47°51′, E 21°24′), including six agricultural management systems representing different land-use types. These included two crop production systems (feed corn and sweet corn), two orchard systems (quince and walnut), one alfalfa-covered area, and one unplanted control area. The control site was managed using shallow loosening and ploughing practices but was not cultivated during the study period. Soil sampling was performed across all sites to evaluate the effects of land use, soil depth, and topographic position on soil microbial communities and soil properties. Samples were collected from both high- and low-altitude locations within each site and from three depth intervals: 0–30 cm, 31–60 cm, and 61–90 cm. Multiple subsamples were collected from each sampling point and combined to obtain representative composite samples. Following collection, soil samples were transferred into sterile Falcon tubes, transported under controlled conditions, and stored at −80 °C until further laboratory analyses.

Soil physicochemical analyses were performed according to standard laboratory procedures at the University of Debrecen. The investigated parameters included soil pH (KCl), consistency (KA), water-soluble salts (WSS), calcium carbonate (CaCO₃), humus content, available phosphorus (P₂O₅), available potassium (K₂O), and nitrate concentration. These measurements were used to characterize the chemical and physical properties of the soils and to evaluate their relationships with microbial community composition and functional profiles.

2.2. DNA Extraction and Shotgun Sequencing

Microbial genomic DNA was recovered from soil samples using the DNeasy® PowerSoil® Pro Kit (Qiagen, Germany), following the approach reported by Remenyik et al. [13]. Prior to extraction, soil suspensions were processed to concentrate microbial cells, and the resulting pellets were subjected to mechanical and chemical lysis. Cell disruption was performed using a MagNA Lyser Instrument (Roche Applied Sciences, Germany), ensuring efficient release of nucleic acids from diverse soil microorganisms. Purified DNA was obtained according to the manufacturer’s protocol and subsequently evaluated for concentration and purity using a Qubit® 4.0 Fluorometer (Thermo Fisher Scientific, USA). Only DNA samples meeting the required quality criteria (OD260/280 ratio between 1.8 and 2.0) and a concentration of at least 10 ng μL⁻¹ were selected for downstream analysis. In cases where these thresholds were not achieved, DNA extraction was repeated to obtain material suitable for high-throughput sequencing.

Sequencing libraries were prepared from purified metagenomic DNA through fragmentation and adapter ligation. Whole-metagenome shotgun sequencing was carried out by Novogene Co. Ltd. (China) on an Illumina NovaSeq 6000 platform (Illumina, USA), generating paired-end reads of 150 bp. To ensure robust characterization of both taxonomic diversity and functional gene content, a sequencing depth of approximately 20 million reads per sample was achieved.

2.3. Statistical Analysis

All data processing, statistical analyses, and visualizations were performed in RStudio (v2025.09.0+387, R Foundation for Statistical Computing, Vienna, Austria). Species-level bacterial and fungal abundance tables obtained from shotgun metagenomic sequencing were filtered to retain only reliably identified taxa.

Soil physicochemical properties, including pH, Arany plasticity index, water-stable soil structure (WSS), CaCO₃, humus, P₂O₅, K₂O, and nitrate concentration, were compared among land-use systems using Kruskal–Wallis tests followed by Dunn’s post hoc comparisons. Microbial alpha diversity was evaluated using the Chao1 richness index, while relationships between diversity and soil properties were assessed using Spearman correlation analysis. Microbial community composition was analyzed using Bray–Curtis dissimilarities, non-metric multidimensional scaling (NMDS), and PERMANOVA. Both presence/absence and abundance-based approaches were used to evaluate shared taxa, dominant species, and differences among land-use systems, soil depths, and environmental gradients. Bacterial and fungal community composition was further examined through land-use-specific comparisons and species overlap analyses.

To investigate ecological functions, taxa were assigned to twelve literature-curated functional groups related to nutrient cycling, plant growth promotion, environmental adaptation, decomposition, and remediation processes. Functional community analyses were performed using only topsoil samples (0–30 cm depth). For each functional group, Bray–Curtis dissimilarity matrices were calculated from species-level abundance data and visualized using non-metric multidimensional scaling (NMDS). Overall differences among land-use systems were evaluated using permutational multivariate analysis of variance (PERMANOVA; 999 permutations). Pairwise comparisons between land-use systems were subsequently performed using pairwise PERMANOVA, with p-values adjusted using the Benjamini–Hochberg false discovery rate (FDR) procedure. Mean pairwise Bray–Curtis dissimilarities among land-use systems were additionally visualized as bubble plots, where bubble size and colour intensity were proportional to community dissimilarity, and statistically significant pairwise comparisons (adjusted p < 0.05) were indicated by an asterisk. Depth-related variation in microbial functional potential was further evaluated using shotgun metagenomic functional profiles. Oxidative stress-related KEGG orthologs were extracted and grouped into five literature-based categories: ROS Detoxification, Redox Homeostasis, Oxidative Damage Repair, Stress Regulation, and Antioxidant Metabolite Production (Supplementary Materials S1) [14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40]. Functional differences among soil depths were evaluated using Bray–Curtis dissimilarities and Principal Coordinate Analysis (PCoA), while statistical significance was assessed using PERMANOVA and Wilcoxon rank-sum tests. To identify oxidative stress-related biomarkers associated with topsoil and subsoil communities, Linear Discriminant Analysis Effect Size (LEfSe) was performed using relative abundances of normalized KEGG functions. Significantly enriched KEGG orthologs (adjusted p < 0.05) were manually classified into five literature-based oxidative stress-related functional categories. For visualization, significant biomarkers were ordered according to functional category and descending LDA score and displayed using a circular lollipop plot. The relative contribution of Topsoil- and Subsoil-enriched biomarkers within each functional category was additionally summarized using donut charts showing the proportion and total number of significant KEGG orthologs assigned to each category. In addition, selected KEGG pathways related to carbon, nitrogen, phosphorus, sulfur, and iron metabolism were compared between topsoil and subsoil using Wilcoxon rank-sum tests. The complete LEfSe output is provided in Supplementary Materials (S2).

Microbial dominance structure was evaluated using abundance-based classification of dominant and rare taxa. Microbial interactions were investigated through co-occurrence network analyses and keystone taxa identification. Network topology metrics, including node connectivity and modularity, were used to compare microbial interaction patterns among land-use systems and soil depths.

Plant-associated bacterial and fungal pathogens were identified using literature-curated pathogen lists developed separately for crop (feed corn and sweet corn) [41,42,43,44,45,46,47,48,49,50] and orchard (quince and walnut) systems [51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71]. The complete pathogen lists are provided in the Supplementary Materials (S3). Pathogen prevalence, composition, abundance patterns, vertical distribution, and relationships with soil physicochemical properties were evaluated using prevalence analyses, abundance profiling, and Pearson correlation analysis.

Statistical significance was defined at p < 0.05. Data manipulation was performed using the dplyr, tidyr, and stringr packages; ecological analyses were conducted using vegan and Hmisc; network analyses were performed using NetCoMi; statistical testing was carried out using rstatix; LEfSe analyses were performed using microbiomeMarker. Graphical visualizations, including circular lollipop plots, donut charts, heatmaps, and Venn diagrams, were generated using ggplot2, patchwork, pheatmap, and VennDiagram.

3. Results

3.1. Spatial Variability of Soil Physicochemical Properties

The physicochemical characterization of soils revealed generally similar environmental conditions across the evaluated land-use systems, although several nutrient-related parameters exhibited noticeable variation (Figure 1). Soil pH ranged from approximately 5.0 to 8.2, with median values between 5.6 and 7.2, and did not differ significantly among treatments (p = 0.72). Likewise, the Arany plasticity index (36–51; p = 0.271), WSS (0.028–0.051%; p = 0.896), CaCO₃ content (0–1.2%; p = 0.838), humus content (2.0–4.1%; p = 0.45), and nitrate concentration (15–72 mg kg⁻¹; p = 0.0966) remained relatively consistent across land-use categories.

In contrast, greater variability was observed for the plant-available nutrient fractions. Available phosphorus (P₂O₅) ranged from approximately 70 to 330 mg kg⁻¹ and showed a marginally significant land-use effect (p = 0.0709). Pairwise comparisons indicated higher P₂O₅ concentrations in quince soils than in sweet corn and walnut soils (adjusted p = 0.048). Available potassium (K₂O) exhibited the strongest land-use response (p = 0.0335), with the highest concentrations recorded in quince soils (approximately 450–580 mg kg⁻¹). Dunn’s post hoc test further revealed significantly higher K₂O values in quince soils compared with sweet corn and walnut soils (adjusted p = 0.03).

3.2. Soil Alpha Diversity

Given the relatively limited variation observed in most soil physicochemical properties, microbial alpha diversity also remained largely stable across the evaluated land-use systems (Figure 2). Chao1 richness values ranged from approximately 18,000 to 26,500 across all samples, indicating a consistently diverse microbial community throughout the study area.

No significant differences in microbial richness were detected among crop types within the 0–30 cm soil layer (p = 0.91; Figure 2a), although slightly higher median richness values were observed in quince, walnut, and sweet corn soils compared with feed corn and alfalfa soils. Similarly, richness did not differ between high- and low-fertility soils (p = 0.64; Figure 2c) or between crop and orchard systems (p = 0.60; Figure 2d), suggesting that land-use category alone exerted only a limited influence on species richness.

In contrast, soil depth emerged as an important driver of alpha diversity (Figure 2b). The highest Chao1 values were recorded in the 0–30 cm layer, with median richness reaching approximately 22,500 species, whereas richness gradually declined with increasing depth. Pairwise comparisons confirmed significantly higher richness in the topsoil compared with the 61–90 cm layer (p < 0.05), indicating a reduction in taxonomic richness along the soil profile.

To further explore potential environmental controls on microbial richness, correlations between soil properties and Chao1 values were examined (Figure 2e). Overall, relationships were weak and predominantly positive. The strongest association was observed for CaCO₃ (ρ = 0.30), followed by WSS (ρ = 0.19) and K₂O (ρ = 0.13). Weaker positive correlations were detected for pH (ρ = 0.11), Arany plasticity index (ρ = 0.08), nitrate (ρ = 0.06), and P₂O₅ (ρ = 0.04), whereas humus content showed virtually no relationship with richness (ρ ≈ 0.00). However, none of these correlations reached statistical significance (p > 0.05).

3.3. Community Structure and Beta Diversity Patterns

Although microbial richness showed only limited variation among most land-use categories, beta-diversity analyses revealed clear differences in microbial community composition (Figure 3). The species-level NMDS ordination based on bacterial and fungal taxa identified soil depth as the primary factor structuring microbial assemblages across the study area.

Samples from the three soil layers formed distinct clusters in ordination space, and PERMANOVA confirmed a significant effect of depth on community composition (R² = 0.233, p = 0.001). Importantly, multivariate dispersion did not differ among depth groups (p = 0.547), indicating that the observed separation primarily reflected differences in community composition rather than differences in within-group variability. Together, these results suggest that vertical soil stratification exerts a strong influence on microbial community assembly.

Land-use type also contributed to variation in microbial composition, although its effects were less pronounced than those associated with soil depth. Crop type explained a relatively large proportion of the observed variation (R² = 0.225) and the ordination indicated a tendency toward crop-specific microbial assemblages, particularly among cultivated systems. However, this effect was only marginally significant (p = 0.085).

In contrast, soil fertility level accounted for only a small fraction of community variation (R² = 0.037) and did not significantly influence microbial composition (p = 0.130). Plant system (crop fields versus orchards) explained a more modest proportion of the total variation (R² = 0.092) but yielded a significant PERMANOVA result (p = 0.048), suggesting differences in community structure between these production systems. Nevertheless, significant heterogeneity of multivariate dispersion was also detected (p = 0.025), indicating that part of the observed separation may be attributable to differences in within-group variability.

3.4. Land-Use Effects on Bacterial and Fungal Assemblages

Land-use systems influenced both the composition and distribution of bacterial and fungal communities in topsoil (Figure 4). Coverage-based analyses revealed the presence of a substantial shared microbial core among all land-use systems, particularly for bacteria. A total of 4,997 bacterial species were detected across alfalfa, not cultivated, crop, and orchard soils (Figure 4a), representing the dominant fraction of the bacterial community. Nevertheless, each land-use system also harbored unique bacterial taxa, with crops containing the highest number of unique species (1,574), followed by not cultivated soils (482), orchards (466), and alfalfa soils (260).

Fungal communities exhibited a smaller shared core and a greater proportion of unique taxa (Figure 4b). Only 226 fungal species were common to all land-use systems, whereas unique species represented a substantially larger component of the fungal assemblage. Crops again contained the highest number of unique fungal species (177), followed by orchards (118), not cultivated soils (44), and alfalfa soils (23). These patterns indicate greater taxonomic turnover among fungal communities than among bacterial communities.

The differences observed in species occurrence were further reflected in overall community composition. NMDS ordinations demonstrated significant separation among land-use systems for both bacterial and fungal assemblages. Bacterial communities differed significantly among land-use categories (PERMANOVA, R² = 0.139, p = 0.012; Figure 4c), with land use explaining approximately 14% of the variation in community structure. Fungal communities exhibited an even stronger response (R² = 0.160, p = 0.001; Figure 4d), indicating that fungal assemblages were slightly more sensitive to land-use differences than bacterial assemblages.

The abundance patterns of dominant shared species further highlighted these compositional differences (Figure 4e–f). Among bacteria, members of the genera Sphingomonas, Skermanella, Flavisolibacter, and Gemmatirosa displayed contrasting abundance profiles across land-use systems. Similarly, dominant fungal species such as Mortierella alpina, Alternaria alternata, Fusarium oxysporum, Rhizophagus irregularis, and Ustilago maydis varied substantially among treatments, indicating that even widely distributed taxa responded differently to management and environmental conditions.

Analysis of unique taxa provided additional evidence for land-use-specific microbial signatures (Figure 4g–h). Crop soils were characterized by highly abundant unique bacterial species, including Sulfuriferula thiophila and Paraburkholderia lava, whereas orchard soils contained characteristic taxa such as Asticcacaulis tiandongensis. Likewise, fungal communities exhibited distinct land-use-associated species, with Hanseniaspora uvarum dominating crop-specific assemblages and Penicillium griseofulvum representing the most abundant orchard-specific fungal species.

3.5. Microbial Dominance Structure and Community Evenness

To determine whether the observed taxonomic and functional differences among land-use systems were driven by a small number of dominant taxa or reflected broader community-wide shifts, species were classified into above-average and below-average groups based on treatment-specific mean relative abundance thresholds.

Across all land-use systems, a relatively small proportion of taxa accounted for most microbial abundance (Figure 5). In bacterial communities, above-average taxa contributed approximately 87–90% of total abundance, whereas below-average taxa accounted for only 10–13% (Figure 5a). A similar pattern was observed for fungi, where dominant taxa represented approximately 84–88% of total abundance (Figure 5b). In contrast, richness patterns showed the opposite trend. Above-average bacterial taxa represented only 10–15% of detected species, while below-average taxa accounted for approximately 85–90% of total richness (Figure 5c). Comparable distributions were observed for fungal communities (Figure 5d), indicating that most species were rare and contributed only marginally to overall abundance.

Despite the compositional differences identified in previous analyses, dominance structure remained remarkably consistent across land-use systems. Only minor variation was observed in the relative contribution of dominant and rare taxa, suggesting that community turnover was primarily associated with changes in taxonomic identity rather than major shifts in abundance distribution.

This pattern was further supported by Pielou’s evenness values. Bacterial communities exhibited relatively uniform evenness across all systems, ranging from approximately 0.66 to 0.73 (Figure 5e). Fungal communities showed slightly greater variability, with median values ranging from approximately 0.62 to 0.75 (Figure 5f), although differences among land-use systems remained modest overall.

The ridge-density analyses revealed a clear separation between dominance classes (Figure 5g–h). In both bacterial and fungal communities, above-average taxa consistently displayed lower Pielou evenness values than below-average taxa, reflecting the disproportionate contribution of a limited number of dominant species to total abundance. This pattern was particularly pronounced for fungi, where the distributions of the two dominance classes showed minimal overlap.

3.6. Microbial Network Organization

Although land-use systems exhibited broadly similar dominance structures and levels of community evenness, co-occurrence network analyses revealed clear differences in the organization and connectivity of microbial communities (Figure 6). Network density varied considerably among management systems, indicating differences in the extent of potential microbial interactions and community integration.

Alfalfa soils exhibited the highest network density for both bacterial (De = 0.806) and fungal communities (De = 0.808), reflecting highly interconnected microbial assemblages with numerous co-occurrence relationships among taxa. In contrast, crop soils displayed the lowest fungal network density (De = 0.384), suggesting a less connected and more fragmented fungal community structure. Not cultivated soils also maintained relatively high connectivity for both bacteria (De = 0.748) and fungi (De = 0.667), whereas orchard soils exhibited intermediate density values (bacteria: De = 0.686; fungi: De = 0.590).

Compared with density, differences in modularity were less pronounced. Bacterial modularity ranged from 0.006 in orchards to 0.051 in crop soils, while fungal modularity varied only slightly between 0.021 and 0.026. Overall, all networks were characterized by relatively low modularity values, indicating limited subdivision into strongly separated modules.

The network topologies visually reflected these patterns. Alfalfa and not cultivated soils displayed densely connected bacterial and fungal communities with numerous links among taxa, whereas crop soils were characterized by reduced connectivity, particularly within fungal communities. Orchard soils occupied an intermediate position, maintaining relatively connected bacterial networks while supporting fungal communities that were less integrated than those observed in alfalfa soils.

Given that soil depth emerged as the strongest driver of microbial community composition, co-occurrence networks were further examined across the three soil layers to determine whether vertical stratification also influenced microbial interaction patterns (Figure 7). Clear depth-related changes were observed in both bacterial and fungal networks. For bacteria, network density increased progressively from 0.625 in the 0–30 cm layer to 0.689 at 31–60 cm and 0.737 at 61–90 cm (Figure 7d), indicating increasingly interconnected microbial associations with depth. At the same time, bacterial modularity declined from 0.051 in the topsoil to 0.002 and 0.006 in the deeper layers, suggesting reduced subdivision into distinct network modules.

Fungal communities exhibited a similar but more pronounced response. Network density increased substantially from 0.384 in the 0–30 cm layer to 0.631 at 31–60 cm and remained high at 0.624 in the deepest layer. The largest shift occurred between the topsoil and subsoil horizons, indicating a marked increase in fungal connectivity below 30 cm depth. In contrast, fungal modularity remained consistently low, decreasing only slightly from 0.026 in the surface layer to 0.021 and 0.015 at greater depths.

The network visualizations supported these quantitative patterns. Fungal communities in the topsoil formed relatively sparse interaction networks, whereas both subsoil layers exhibited denser and more integrated structures. Bacterial communities displayed a comparable, although less pronounced, increase in connectivity with depth. Across all soil layers, modularity values remained low, indicating limited compartmentalization of microbial interactions throughout the soil profile.

3.7. Land-Use Effects on Microbial Functional Composition

To facilitate ecological interpretation, the twelve literature-curated microbial functions were grouped into four broader ecosystem-function categories: Nutrient Cycling and Plant Nutrition, Plant–Microbe Interactions, Environmental Adaptation and Remediation, and Organic Matter Turnover and Carbon Cycling, following the classification framework previously developed and described by Gashi et al. [72].

Non-metric multidimensional scaling (NMDS) based on Bray–Curtis dissimilarities revealed that the composition of predicted microbial functional groups varied among land-use systems, although the degree of separation differed according to functional category (Figure 8). Within the Nutrient Cycling and Plant Nutrition category, phosphorus solubilization (PERMANOVA, p = 0.004, Figure 8b), potassium solubilization (p = 0.001, Figure 8c), and plant growth promotion (p = 0.004, Figure 8d) showed significant overall differences among land-use systems, whereas nitrogen fixation did not (p = 0.092, Figure 8a), indicating that nitrogen-fixing communities remained comparatively conserved despite partial visual separation. Pairwise PERMANOVA further showed significant differences for phosphorus solubilization, potassium solubilization, and plant growth promotion after Benjamini–Hochberg correction, whereas no pairwise differences were detected for nitrogen fixation. Among Plant–Microbe Interactions, siderophore production (p = 0.002, Figure 8f) and mycorrhizal symbiosis (p = 0.007, Figure 8g) exhibited significant land-use-dependent clustering, while production of phytohormones showed only moderate overall separation (p = 0.097, Figure 8e). Pairwise analyses confirmed significant differences for siderophore production, mycorrhizal symbiosis, and production of phytohormones between selected land-use systems. Within the Environmental Adaptation and Remediation category, hydrocarbon degradation displayed one of the strongest separations among land-use systems (p = 0.001, Figure 8i), whereas bioremediation (p = 0.162, Figure 8h) and antibiotic resistance (p = 0.251, Figure 8j) exhibited substantial overlap among groups, indicating relatively stable functional composition across vegetation types. Consistently, pairwise PERMANOVA identified significant differences only for hydrocarbon degradation, while bioremediation and antibiotic resistance remained non-significant after multiple-testing correction. Similarly, functions related to Organic Matter Turnover and Carbon Cycling differed significantly among land-use systems, as both organic matter decomposition (p = 0.001, Figure 8k) and methane oxidizers (p = 0.006, Figure 8l) exhibited distinct ordination patterns, particularly separating orchard-associated communities from the remaining land-use systems. Pairwise comparisons further revealed the highest number of significant differences for organic matter decomposition, whereas methane oxidizers showed a more limited but still significant differentiation among selected land-use systems.

3.8. Depth-Dependent Variation in Microbial Functional Potential

3.8.1. Oxidative Stress-Related Functional Profiles

Oxidative stress-related microbial functions play a central role in protecting soil microorganisms from reactive oxygen species (ROS), maintaining intracellular redox balance, repairing oxidative damage, and regulating cellular responses to environmental stress. To facilitate ecological interpretation, oxidative stress-associated KEGG functions were grouped into five categories: ROS Detoxification, Redox Homeostasis, Oxidative Damage Repair, Stress Regulation, and Antioxidant Metabolite Production (Figure 9). The analysis revealed a clear vertical stratification of oxidative stress functions throughout the soil profile. When all oxidative stress-related KEGGs were considered together, microbial communities were significantly separated according to depth (R² = 0.403, p = 0.001), and topsoil exhibited a significantly greater overall oxidative stress functional potential than subsoil (p = 0.0358) (Figure 9a).

Among the individual categories, Stress Regulation showed the strongest depth-related structuring (R² = 0.548, p = 0.001) (Figure 9d), followed by Redox Homeostasis (R² = 0.471, p = 0.001) (Figure 9e) and ROS Detoxification (R² = 0.396, p = 0.001) (Figure 9b). Moreover, the functional potential of ROS Detoxification was substantially higher in topsoil than in subsoil (p = 5.1 × 10⁻⁶). Similarly, Redox Homeostasis functions were significantly enriched in topsoil (p = 0.00025) (Figure 9e). Stress Regulation functions also showed significantly greater potential in topsoil (p = 0.026).

In contrast, Oxidative Damage Repair displayed significant compositional differences across depths (R² = 0.361, p = 0.001) but no significant difference in total functional potential between topsoil and subsoil (p = 0.36) (Figure 9c). Likewise, Antioxidant Metabolite Production exhibited clear depth-dependent community structuring (R² = 0.384, p = 0.001), yet its overall functional potential remained comparable between soil layers (p = 0.71) (Figure 9f).

3.8.2. LEfSe Biomarkers of Oxidative Stress Functions

To identify the oxidative stress-related functions driving the functional differentiation between topsoil and subsoil microbial communities, LEfSe analysis was performed using oxidative stress-associated KEGG orthologs (Figure 10). A total of 117 significantly enriched KEGG biomarkers were identified, revealing clear depth-dependent functional specialization (Figure 10a). Topsoil was primarily characterized by biomarkers involved in redox balance, ROS detoxification, and cellular stress responses, including thioredoxin reductase (K00384), superoxide dismutase (K03781), and several DNA repair and stress-regulation functions. In contrast, subsoil communities were predominantly enriched in biomarkers associated with antioxidant metabolite production, including K00549, K01011, K03833, K12960, K21148, K00245, and K03636.

Grouping the significant biomarkers into five oxidative stress-related functional categories further emphasized these depth-dependent patterns (Figure 10b). Antioxidant Metabolite Production represented the largest functional group (n = 54), with 64.8% of biomarkers enriched in subsoil and 35.2% in topsoil. Oxidative Damage Repair contained 16 significant biomarkers and exhibited an equal distribution between topsoil and subsoil (50% each), suggesting that DNA and protein repair mechanisms are equally important across both soil horizons. In contrast, ROS Detoxification showed the strongest topsoil preference, with 83.3% of significant biomarkers enriched in topsoil and only 16.7% in subsoil. Likewise, Redox Homeostasis was predominantly represented by topsoil biomarkers (70%) compared with subsoil (30%), while Stress Regulation also favored topsoil (60%) over subsoil (40%).

3.8.3. Nutrient Cycling and Metabolic Pathways

Depth-related functional shifts were explored with selected KEGG pathways associated with carbon, nitrogen, phosphorus, sulfur, and iron metabolism were compared between topsoil and subsoil (Figure 11). The results indicate that topsoil harbors a greater functional potential for several key metabolic processes. The biosynthesis of siderophore group nonribosomal peptides was significantly enriched in topsoil (p = 0.021). Similarly, phosphonate and phosphinate metabolism showed significantly higher abundance in topsoil (p = 0.015). Among all pathways examined, starch and sucrose metabolism exhibited the strongest difference between soil layers (p = 0.0012). Inositol phosphate metabolism (p = 0.058) and sulfur metabolism (p = 0.054) displayed similar trends, with higher relative abundances in topsoil, although the differences remained slightly above the conventional significance threshold. In contrast, nitrogen metabolism did not differ significantly between topsoil and subsoil (p = 0.19).

3.9. Plant Pathogen Communities

3.9.1. Overall Pathogen Prevalence Across Agricultural Systems

To provide an initial overview of pathogen occurrence across agricultural systems, pathogen prevalence was examined separately for crop and orchard soils (Figure 12). Mean prevalence values were calculated from the proportion of samples in which each pathogen was detected, allowing comparison of the overall distribution of bacterial and fungal pathogens among host systems.

Among crop systems, bacterial pathogens exhibited relatively high prevalence in both feed corn and sweet corn soils. Sweet corn showed a higher mean bacterial pathogen prevalence (72.5%) than feed corn (62.5%), despite harboring fewer detected bacterial pathogen species (15 vs. 18). This pattern suggests that bacterial pathogens in sweet corn were distributed more consistently across samples, whereas feed corn supported a broader pathogen diversity with less uniform occurrence. However, the difference between crop types was not statistically significant (Wilcoxon test, p = 0.546).

A similar trend was observed for fungal pathogens. Sweet corn displayed a slightly higher mean prevalence (36.8%) than feed corn (32.8%) and contained a greater number of detected fungal pathogen species (18 vs. 16). Nevertheless, prevalence distributions did not differ significantly between the two crop types (p = 1.000).

Within orchard systems, bacterial pathogen prevalence was nearly identical between quince and walnut soils, with both systems exhibiting a mean prevalence of 77.8%. Although quince contained a larger number of detected bacterial pathogen species (15) than walnut (12), prevalence patterns remained highly similar (p = 0.892), indicating that bacterial pathogens were equally widespread across both orchard types.

The largest contrast was observed for fungal pathogens in orchard soils. Quince exhibited a substantially higher mean fungal pathogen prevalence (71.4%) than walnut (54.2%), despite containing slightly fewer pathogen species (7 vs. 8). Although this difference was not statistically significant (p = 0.288), it suggests that fungal pathogens were detected more consistently across quince samples.

3.10. Distribution of Plant Pathogens Across Agricultural Systems

A substantial proportion of the curated pathogen lists was detected within the studied soils, indicating that both crop and orchard systems harbored diverse communities of potential plant pathogens (Figure 13). Bacterial pathogen assemblages in crop soils were dominated by members of the genera Xanthomonas, Pseudomonas, Ralstonia, Clavibacter, and Xylella. In contrast, orchard soils were characterized primarily by Pseudomonas marginalis, Burkholderia glumae, Xylella fastidiosa, and several Xanthomonas species. Fungal pathogen communities displayed even clearer differences between production systems. Crop soils were dominated by Fusarium-associated pathogens, including Fusarium oxysporum, F. graminearum, F. proliferatum, F. verticillioides, and F. solani, whereas orchard soils were characterized by Alternaria alternata, Epicoccum nigrum, Fusarium oxysporum, Monilinia fructigena, and Verticillium dahliae.

Detected pathogen diversity was relatively evenly distributed among host types within each production system. Among bacterial pathogens, feed corn contributed 55.9% of the crop-associated species, while sweet corn accounted for 44.1%. A similar pattern was observed in orchards, where quince and walnut contributed 55.6% and 44.4% of the detected bacterial pathogen diversity, respectively. Fungal communities followed the same general trend, although sweet corn (54.5%) and walnut (53.3%) contributed slightly larger proportions of fungal pathogen diversity than feed corn (45.5%) and quince (46.7%).

Despite the relatively balanced species distributions, pathogen abundance was strongly concentrated in a limited number of taxa. Within crop systems, bacterial pathogen communities were dominated by Xanthomonas-related species, particularly Xanthomonas oryzae, X. citri, X. campestris, and X. translucens, together with Pseudomonas syringae. Orchard bacterial communities were largely dominated by Micrococcus luteus and Pantoea agglomerans, which together accounted for most of the bacterial pathogen abundance. For fungal pathogens, crop soils were overwhelmingly dominated by Ustilago maydis and several Fusarium species, whereas orchard soils were characterized by the strong dominance of Monilinia fructigena, accompanied by Alternaria alternata and Fusarium oxysporum.

Comparisons among individual host systems revealed additional differences in pathogen structure. Feed corn and sweet corn supported largely similar bacterial pathogen assemblages, although the relative contribution of individual Xanthomonas species varied between crops. Greater variation was observed within fungal communities, where sweet corn contained a larger contribution of Fusarium species than feed corn. Orchard systems also differed in their dominant pathogens. Walnut soils showed a stronger dominance of Monilinia fructigena, whereas quince soils supported a more diverse fungal pathogen assemblage involving Alternaria, Fusarium, and Rhizoctonia species. Similar host-specific shifts were observed among bacterial pathogens, where the relative abundance of individual taxa varied between quince and walnut despite the presence of a broadly shared pathogen pool.

3.11. Species-Level Pathogen Prevalence Patterns

Species-level prevalence analyses revealed that a relatively small number of pathogens occurred consistently across most samples, forming a core pathogen community within each production system (Figure 14). In contrast to abundance-based analyses, prevalence reflects the frequency with which individual pathogens were detected and therefore identifies taxa that are widespread regardless of their relative abundance.

Crop-associated bacterial pathogens exhibited particularly high prevalence (Figure 14a). Xylella fastidiosa, Xanthomonas translucens, Xanthomonas oryzae, Xanthomonas citri, Xanthomonas campestris, Ralstonia solanacearum, Pseudomonas syringae, and Clavibacter michiganensis were present in all feed corn and sweet corn samples, indicating a highly stable bacterial pathogen core. At the opposite end of the spectrum, Dickeya dadantii, Xanthomonas cucurbitae, and Rathayibacter tritici occurred only sporadically, with prevalence values ranging from 8% to 17%.

A similarly high prevalence was observed among orchard bacterial pathogens (Figure 14b). Xanthomonas campestris, Xanthomonas arboricola, Pseudomonas viridiflava, Pseudomonas syringae, Pantoea agglomerans, and Micrococcus luteus were detected in every quince and walnut sample. Other pathogens, including Xylella fastidiosa, Pseudomonas savastanoi, and Burkholderia glumae, occurred in all quince samples but only two-thirds of walnut samples, suggesting moderate host-specific variation in distribution.

Among fungal pathogens, Fusarium oxysporum was the most consistently detected species in crop systems, occurring in all feed corn and sweet corn samples (Figure 14c). Ustilago maydis also showed high prevalence, being present in 83% of feed corn and all sweet corn samples. Several additional Fusarium species, including F. solani, F. equiseti, and F. proliferatum, exhibited moderate to high prevalence, whereas Eutypa lata, Fusarium sporotrichioides, and Phaeoacremonium chlamydospora were encountered only occasionally. These patterns indicate that crop-associated fungal pathogen communities were largely structured around a core group of Fusarium-related taxa.

Orchard fungal communities displayed a somewhat different prevalence structure (Figure 14d). Fusarium oxysporum and Alternaria alternata were detected in all quince and walnut samples, making them the most consistently occurring fungal pathogens within orchard soils. Rhizoctonia solani was also highly prevalent, occurring in every quince sample and in 67% of walnut samples. In contrast, Monilinia fructigena and Phaeoacremonium minimum were restricted to walnut soils, whereas Verticillium dahliae was detected exclusively in quince soils, highlighting host-specific differences in pathogen distribution.

3.12. Depth-Related Distribution of Pathogens

The widespread occurrence of several bacterial and fungal pathogens prompted a further examination of their vertical distribution within the soil profile. By comparing pathogen abundance across the 0–30, 31–60, and 61–90 cm soil layers, it was possible to assess whether pathogens were concentrated in surface soils or persisted throughout deeper horizons (Figure 15).

Bacterial pathogens displayed relatively broad depth distributions, with several taxa occurring across all three soil layers (Figure 15a). Xanthomonas oryzae, Pseudomonas syringae, Ralstonia solanacearum, Xanthomonas citri, and Clavibacter michiganensis maintained measurable abundance throughout the soil profile, indicating persistence beyond the surface horizon. Among these taxa, Xanthomonas oryzae exhibited the highest abundance, particularly within the 0–30 cm layer. In contrast, Ralstonia pseudosolanacearum, Pseudomonas cichorii, and Xanthomonas cucurbitae showed more restricted distributions and occurred only within selected depth intervals at comparatively low abundance.

Fungal pathogens exhibited stronger depth-related patterns than bacterial pathogens (Figure 15b). The highest abundance values were generally concentrated in the topsoil, where Fusarium oxysporum and Ustilago maydis emerged as the dominant fungal pathogens. Several Fusarium species, including F. solani, F. proliferatum, F. poae, and F. equiseti, were detected across multiple soil layers, demonstrating their ability to persist throughout the soil profile. Other fungal pathogens displayed more distinct depth preferences. Verticillium longisporum, Phaeomoniella chlamydospora, and Macrophomina phaseolina were primarily associated with the upper soil layer, whereas Claviceps purpurea and Bipolaris sorokiniana were more frequently detected in deeper horizons.

3.13. Environmental Drivers of Pathogen Distribution

The widespread occurrence of several pathogens across crop and orchard systems prompted an examination of their relationships with soil physicochemical properties (Figure 16). Overall, fungal pathogens in crop systems exhibited the strongest and most frequent associations with soil variables, whereas pathogens in orchard soils generally showed weaker and predominantly non-significant responses.

Among crop-associated fungal pathogens (Figure 16a), soil pH emerged as the most influential environmental factor. Several Fusarium species exhibited significant negative correlations with pH, including Fusarium oxysporum (r = −0.68, p < 0.001), Fusarium solani (r = −0.61, p < 0.01), Fusarium equiseti (r = −0.51, p < 0.05), and Fusarium graminearum (r = −0.45, p < 0.05). Similar negative relationships were observed for Ustilago maydis (r = −0.67, p < 0.001) and Gaeumannomyces tritici (r = −0.58, p < 0.01). In contrast, Rhizoctonia solani was positively associated with available phosphorus (P₂O₅) (r = 0.53, p < 0.01), while U. maydis also exhibited a negative relationship with CaCO₃ content (r = −0.49, p < 0.05).

Relationships between orchard fungal pathogens and soil properties were considerably weaker (Figure 16b). Although several taxa displayed moderate positive or negative trends, none of the observed correlations reached statistical significance, suggesting that the measured soil parameters explained only a limited proportion of variation in fungal pathogen abundance within orchard soils.

Crop-associated bacterial pathogens generally exhibited weaker responses than fungal pathogens (Figure 16c). Most correlations were low to moderate and non-significant. The clearest associations were observed for Pseudomonas syringae, which showed positive correlations with pH (r ≈ 0.49, p < 0.05) and CaCO₃ (r ≈ 0.61, p < 0.01). Other bacterial pathogens displayed only limited relationships with the measured soil variables.

A similar pattern was observed within orchard bacterial communities (Figure 16d). Most associations were weak and non-significant, although Xanthomonas arboricola exhibited a strong negative correlation with pH (r ≈ −0.81, p < 0.001) and a negative association with CaCO₃ (r ≈ −0.69, p < 0.01). Pseudomonas syringae also showed a weaker negative relationship with pH (p < 0.05). Positive associations with nutrient-related variables such as nitrate, P₂O₅, and K₂O were observed for several taxa, although these relationships were not statistically significant.

4. Discussion

The results of this study demonstrate that soil microbial communities are shaped by the combined effects of vertical soil stratification and land-use history, although these factors influence different aspects of community organization. Among all evaluated variables, soil depth emerged as the strongest determinant of microbial community composition, highlighting the importance of vertical environmental gradients in structuring belowground biodiversity.

The strong influence of depth is consistent with the concept of environmental filtering, whereby soil horizons represent distinct ecological habitats characterized by differences in carbon availability, root density, oxygen concentration, moisture dynamics, and nutrient accessibility. Surface soils receive continuous inputs of plant residues and root exudates that create diverse ecological niches and support highly heterogeneous microbial communities. In contrast, deeper soil layers are generally characterized by lower carbon availability and more stable environmental conditions, favoring a smaller subset of microorganisms adapted to resource limitation [73]. Consequently, the observed decline in richness with depth likely reflects not only reduced resource availability but also increasing environmental specialization. The pronounced separation of microbial communities among soil layers further suggests that depth-related patterns are driven primarily by species replacement rather than simple richness loss, indicating that deeper horizons harbor distinct microbial assemblages adapted to specific ecological conditions [74].

While depth represented the dominant environmental filter, land-use systems exerted a strong influence on community composition without substantially affecting microbial richness. Such a pattern is frequently observed in soil ecosystems and reflects the high degree of functional redundancy within microbial communities. Multiple taxa often perform similar ecological functions, allowing species turnover to occur without major changes in overall diversity or ecosystem functioning [75]. This interpretation is supported by the presence of a large shared bacterial core across all land-use systems, suggesting that a considerable proportion of the soil microbiome remains relatively stable despite differences in management practices [76].

At the same time, the identification of numerous unique bacterial and fungal taxa indicates that individual land-use systems provide distinct ecological niches that support specialized microbial assemblages [77]. The particularly high number of unique taxa detected in crop soils may reflect the greater environmental heterogeneity associated with annual cultivation, fertilization practices, seasonal vegetation turnover, and recurrent soil disturbance [77].

Fungal communities appeared to respond more strongly to land-use differences than bacterial communities, a pattern commonly reported in agricultural ecosystems. Unlike bacteria, fungi rely heavily on plant-derived carbon inputs and frequently establish close associations with roots, litter, and decomposing plant residues [78]. As a result, differences among alfalfa, crop, orchard, and uncultivated systems are likely to create contrasting resource environments that selectively favor distinct fungal assemblages. Bacterial communities, although responsive to environmental change, generally exhibit greater metabolic flexibility and are therefore capable of persisting across a wider range of environmental conditions [79]. This difference in ecological strategy may explain why fungal communities exhibited stronger compositional turnover than bacterial communities across the studied land-use systems.

The observed functional differentiation indicates that the effects of land use extend beyond taxonomic composition and ultimately influence ecological processes linked to nutrient cycling, plant productivity, and carbon turnover [80,81]. However, the strength of this response varied among functional groups. Functions associated with phosphorus and potassium solubilization, plant growth promotion, mycorrhizal symbiosis, hydrocarbon degradation, and organic matter decomposition exhibited the clearest land-use-specific differentiation, whereas nitrogen fixation, bioremediation, and antibiotic resistance remained comparatively conserved after correction for multiple comparisons. These findings suggest that microbial functions directly linked to nutrient acquisition and carbon turnover are more responsive to long-term changes in vegetation and management practices than broadly distributed ecological functions [82]. The particularly strong differentiation observed for organic matter decomposition, which exhibited the highest number of significant pairwise comparisons among all evaluated functional groups, further highlights the importance of organic matter quality in shaping microbial communities Although microbial communities exhibit a considerable degree of functional redundancy, certain ecosystem processes remain strongly dependent on community composition. Glassman et al. [83] demonstrated that shifts in microbial composition can directly alter decomposition rates and litter chemistry, suggesting that different vegetation types support microbiomes with distinct functional capacities for degrading cellulose, lignin, and other organic compounds. Therefore, the functional divergence observed among land-use systems likely reflects differences in both resource quality and long-term vegetation inputs.

Despite substantial taxonomic and functional turnover, microbial dominance patterns remained remarkably stable across land-use systems. In all cases, a relatively small fraction of taxa accounted for the majority of microbial abundance, whereas most species occurred at low abundance. This pattern is characteristic of complex microbial ecosystems and reflects the existence of a stable abundance hierarchy. Dominant taxa are often ecological generalists capable of efficiently exploiting available resources, while rare taxa constitute a large reservoir of genetic and functional diversity that contributes to ecosystem resilience. Similar patterns were reported by Gschwend et al. [84], who showed that the relationship between abundant and rare taxa remains surprisingly consistent across contrasting land-use systems despite ongoing species turnover.

Although dominance structure remained largely unchanged, network analyses revealed pronounced differences in the organization of microbial interactions. This apparent contrast suggests that land-use systems influence how microorganisms interact rather than how abundance is distributed among taxa. The highly connected bacterial and fungal networks observed in alfalfa soils likely result from the continuous root inputs and reduced disturbance associated with perennial vegetation. Long-term carbon inputs can promote cooperative interactions and increase network stability, leading to more integrated microbial communities. Conversely, the reduced connectivity observed in crop soils may reflect periodic disturbance, seasonal vegetation turnover, and fluctuating resource availability [85].

Depth-related network responses revealed an equally interesting pattern. Although deeper soil layers contained fewer species, the remaining microorganisms formed increasingly interconnected networks. One possible explanation is that resource limitation strengthens microbial dependencies, promoting cooperative interactions, cross-feeding relationships, and metabolic complementarity among the taxa that persist in deeper horizons. Similar mechanisms have been proposed for oligotrophic environments, where community stability relies more heavily on interspecific interactions than on taxonomic diversity itself. However, the direction of depth-related network responses is not universal. Guseva et al. [86] reported lower connectivity in forest subsoils, suggesting that the effects of depth depend strongly on ecosystem type, vegetation cover, and local environmental conditions.

Soil depth strongly influenced the oxidative stress response strategies of microbial communities. The enrichment of ROS detoxification, redox homeostasis, and stress-regulation functions in topsoil suggests that microorganisms in surface horizons are exposed to greater oxidative stress due to higher oxygen availability, stronger environmental fluctuations, and intensive root activity. According to Pett-Ridge and Firestone [87], microbes in the surface layers of tropical soils, exposed to strong environmental fluctuations, develop physiological tolerance mechanisms and antioxidant enzymes to cope with redox stress. In contrast, the predominance of antioxidant metabolite production biomarkers in subsoil indicates a shift toward more stable and long-term protective mechanisms. The enrichment of thioredoxin reductase and superoxide dismutase further supports the importance of active oxidative stress management in topsoil communities.

Depth also affected nutrient-related metabolic pathways. The higher abundance of siderophore biosynthesis, phosphorus-acquisition pathways, and carbohydrate metabolism in topsoil likely reflects greater microbial competition for nutrients and increased availability of plant-derived substrates. According to Huang et al. [88], soil depth significantly influences the composition of soil organic matter, with topsoil characterized by a higher abundance of carbohydrates and a greater availability of plant-derived substrates, such as plant residues and root exudates, compared with subsoil. These findings suggest that soil depth selects for distinct microbial functional strategies, with topsoil communities adapted for rapid stress response and resource acquisition, whereas subsoil communities appear better adapted to relatively stable and resource-limited conditions.

Pathogen communities provided an additional perspective on how environmental conditions shape microbial assemblages. The widespread occurrence of several bacterial and fungal pathogens suggests the existence of stable pathogen cores within both crop and orchard systems. However, prevalence alone should not be interpreted as disease risk because pathogen activity depends on interactions among host plants, environmental conditions, and surrounding microbial communities.

Among the evaluated soil properties, pH emerged as the most consistent predictor of pathogen abundance. The strong negative associations observed between pH and several Fusarium species support the view that soil acidity acts as a major ecological filter regulating pathogen persistence and competitiveness. This interpretation is consistent with the findings of Li et al. [89], who identified soil pH as one of the strongest predictors of Fusarium root rot severity and demonstrated that soil acidification reduces the natural suppressive capacity of the soil microbiome. In contrast, nutrient-related variables exhibited comparatively weak relationships with pathogen abundance, indicating that pathogen distribution is influenced more strongly by environmental filtering and biological interactions than by nutrient availability alone.

The weaker environmental responses observed among orchard pathogens may reflect the greater ecological stability of perennial systems. Compared with annually cultivated fields, orchard soils experience lower levels of physical disturbance and maintain long-term plant–microbe associations that can buffer short-term fluctuations in environmental conditions. Similar patterns have been reported in perennial fruit systems, where management history and landscape context often exert stronger influences on microbial community assembly than individual soil properties [90]. Together, these findings suggest that pathogen communities are governed by a combination of abiotic filtering, host specificity, and microbial interactions, with soil pH representing one of the most important environmental drivers identified in the present study.

Collectively, these findings indicate that soil depth acts as the primary environmental filter shaping microbial community assembly, whereas land-use systems primarily influence community composition, ecological functions, and microbial interactions. Although richness and dominance structure remained relatively stable, substantial turnover was observed in taxonomic composition, functional potential, and network organization. The stronger response of fungal communities compared with bacterial communities highlights the importance of vegetation-driven resource availability in structuring soil microbiomes. Together, these results demonstrate that both vertical stratification and land-use history contribute to the development of functionally distinct soil microbial communities with potential implications for nutrient cycling, ecosystem stability, and plant health.

5. Conclusions

Understanding how agricultural management and environmental gradients shape soil microbial communities is essential for maintaining soil health and ecosystem sustainability. The results highlight the importance of both soil depth and land-use system in shaping agricultural soil microbiomes, although their effects operate at different levels of community organization. Soil depth emerged as the dominant factor influencing microbial richness, community composition, and network structure, whereas land-use systems primarily affected taxonomic turnover, ecological functions, microbial interactions, and pathogen distribution. Despite relatively stable alpha diversity and dominance patterns, substantial differences were observed in community composition, functional potential, and network organization. Fungal communities responded more strongly to land-use changes than bacterial communities, and shifts in microbial composition were accompanied by changes in nutrient cycling, plant-growth-related functions, and decomposition processes. In addition, soil depth was associated with distinct oxidative stress response strategies, with topsoil communities exhibiting greater potential for ROS detoxification, redox homeostasis, and stress regulation, while subsoil communities were characterized by a higher representation of antioxidant metabolite production functions. Network analyses further revealed greater connectivity in perennial systems, particularly alfalfa soils, while pathogen analyses identified soil pH as one of the most important factors associated with pathogen distribution. Overall, these findings demonstrate that agricultural management influences not only microbial diversity but also the ecological functions and interactions that underpin soil health and ecosystem functioning. Future studies should combine multi-omics approaches with broader geographical sampling to better understand the mechanisms governing soil microbiome assembly and resilience under different management practices. Such approaches may also help elucidate how beneficial soil microbiomes influence crop nutritional quality and the accumulation of health-promoting bioactive compounds, thereby supporting the development of sustainable agricultural systems for future medical food production.