The study was carried out at Palmerillas Experimental Station (Cajamar Foundation), located in the province of Almería (geographical coordinates: 36° 48′N, 2° 3′W; elevation: 155 m). Two adjacent greenhouses with similar soils and comparable crop histories were selected (Figure 1). Further baseline soil parameters are available in Supplementary Material A.

In each of these greenhouses, two different fertilization systems were established (Supplementary Material B). The greenhouse under conventional management (C) received inorganic fertilization applied through fertigation. In contrast, the greenhouse with ecological intensification practices (EI) was fertilized by incorporating crop residues from the previous year (Solanum melongena L.), which were mixed with organic amendments (manure). The detailed description of the management systems, as well as the soil total organic carbon (TOC) and total nitrogen (TN) contents of the amendments used (crop residues and manure), can be found in Hernández Maqueda et al. (2024).

Sampling was carried out after 6 years of incorporating organic amendments in the greenhouse under ecological intensification management (I.E.,). The crop at the time of sampling was eggplant (Solanum melongena).

The soil sampling was carried out at the end of the growing season, once retired the crop (May 2021). In each greenhouse three drip lines were chosen randomly and, in each of them, a dripper was chosen randomly too. Soil samples were taken at 0–15 cm below the drip line. Once taken, nine soil samples per greenhouse (3 replicates x 3 positions) were stored field-moist in sealed polythene bags at 4 °C until required. A sample aliquot was stored in polythene containers at −80 °C for DNA extraction and molecular biology analysis.

Field moist samples were disaggregated and sieved through a 2 mm screen.

Total organic carbon was measured on fine ground soil by wet oxidation method according to Mingorance et al. (2007). Water soluble organic carbon (WSOC) was measured in a 1:2 soil:water suspension (Scaglia and Adani, 2009). Total Nitrogen (TN) was measured using an Elementar Rapid N exceed elemental analyzer. Soil ammonium (NH₄ ⁺) and nitrate (NO₃ ⁻) contents were extracted with 2 mol L⁻¹ KCl according to Mulvaney (1996). The analysis of NH₄ ⁺ and NO₃ ⁻were carried out within 3 days after sampling. Electrical conductivity (EC, dSm⁻¹) was measured in a 1:5 soil: water suspension (v/v), with a Crison 522 conductivity meter (UNE 77308:2001), and pH was measured in a 1:5 soil:water suspension (v/v) with a Crison basic 20 pH-meter (UNE-ISO 10390:2012). Dissolved cations were measured in 1:5 soil water suspension (v/v): Ca⁺⁺ and Mg⁺⁺ by atomic absorption spectroscopy. Na⁺and K⁺by flame photometry. Finally, the percentage of resistant soil aggregates (RA) was calculated using a Wet Sieving Apparatus (Eijkelkamp).

DNA was extracted in duplicate from 500 mg of each individual soil sample employing the PowerSoil™DNA Extraction Kit (MoBio Laboratories Inc., Carlsbad, CA, USA) according to the manufacturer’s instructions. DNA concentration and purity were estimated spectroscopically using Nanodrop 2000™ Spectrophotometer (Thermo Fisher, Wilmington, Delaware USA). The quantification of bacterial and fungi abundance was performed by Real-Time Quantitative Polimerase Chain Reaction (qPCR) using a 7900HT Fast Real Time PCR System (Applied Biosystems, Foster City, California USA.). PCR reactions were performed in triplicate with a total volume of 20 μL containing 5 μL × iTaq™ universal SYBR® Green supermix (BioRad, Los Angeles, California USA), 2 μL of each primer (10 μM) and 10 ng of template DNA. PCR conditions were, an initial denaturation step at 95 °C for 10 min, followed by 40 cycles of 95 °C for 15 s and 60 °C for 1 min. Specific primers from conserved regions of the bacteria 16S rRNA gene (Steven et al., 2014) and fungi ITS1-5.8S (Delgado-Delgado-Baquerizo et al., 2016) were used in order to perform the quantitative PCR analysis. Results were expressed as number of copies of bacterial 16S rRNA genes or ITS, per Gram of soil. The same protocol was followed to obtain the abundance of the genes involved in the denitrification pathway (nirK, nirS, nosZ1 and nosZ2). The primers employed, and the specific real-time PCR conditions for this genes are described by Hernández Maqueda et al. (2024), Hernández Maqueda et al. (2025). Detailed qPCR conditions are described as Supplementary Material C,D.

Statistical analyses were performed using the open-source software R version 4.5.1 (R Core Team, 2022). The illustrations were generated with the ggplot2 package. A linear mixed model (LMM) was employed to examine the influence of management on both the physical, chemical and biological variables included in the study by means of the packages lme4 and lmerTest. To test the assumptions of parametric statistical analyses, all variables were transformed to a natural logarithm, when necessary, to achieve normality and homogeneity of variance, and then examined using the Shapiro-Wilk (S-W) and Bartlett’s tests, respectively. The model, following the “REML” specification, utilized residual (restricted) maximum likelihood estimation. The fixed effects included EI application versus control, and dripline was considered a random effect (RE), capturing random deviations for each sample from the overall fixed intercept. The goodness of fit of the best model was assessed using the Bayesian Information Criterion (BIC). Correlations were calculated with the package Hmisc. The relationships between the DNA abundances and the physicochemicals variables were assessed by meaning of a Spearman’s Rho nonparametric test. Significant correlations (p < 0.05) were adjusted using the False Discovery Rate (FDR). Graphics were performed with the library corrplot.

According to the linear mixed model (LMM), statistically significant differences were detected between the two management systems in relation to the soil’s physical and chemical properties, with the exception of the ammonium ion (NH₄ ⁺). With regard to the biological properties analyzed, a substantial increase was observed in fungal DNA abundance, as well as in the copy numbers of the nosZ1 and nosZ2 genes involved in the reduction of N₂O to N₂ (Table 1).

When analyzing the proportional increase of each variable in soil managed with EI compared to conventionally managed soil (Figure 2), we found that the physicochemical parameter with the greatest increase is electrical conductivity (EC), which is 7.4 fold higher than under conventional management.

The analyzed organic carbon fractions (TOC and WSOC) more than doubled under EI management. Total nitrogen (TN) increased 1.5-fold, while ammonium (NH₄ ⁺) remained at similar levels, and nitrate (NO₃ ⁻) nearly doubled compared with conventional management. Consequently, the C to N ratio increased 1.5-fold. Among the soluble cations, sodium (Na⁺) increased four-fold, potassium (K⁺) three-fold, and magnesium (Mg²⁺) 1.7-fold, whereas calcium (Ca²⁺) decreased by 34%. Additionally, soil pH declined by approximately 9%, averaging a decrease of 0.7 units, and the percentage of soil resistant aggregates (SRA), an indicator of soil structural stability, increased roughly 1.5-fold.

Bacterial abundance measured by the 16S gene remained similar in both treatments (1.65 vs. 1.33 log₁₀ copies g^-1 dry soil). However, a substantial increase was observed in the number of ITS2 gene copies used to detect fungal abundance (2.11 vs. 0.27 log₁₀ copies g^-1 dry soil). Consequently, the Fungi/Bacteria ratio also increased threefold.

All genes analyzed to evaluate the denitrification process increased in EI-managed soil, with the increase being statistically significant for both variants of the nitrous oxide reductase genes (nosZ1 and nosZ2).

The correlation analysis based on Spearman’s Rho between the abundances of various genes and the different soil parameters revealed the following main findings (Figure 3):

TOC correlated positively (Rho = 0.700, p = 0.008) with the abundance of ITS2;

Real-time PCR analyses showed substantial increases in biological indicators following EI management, indicating an overall enhancement of microbial presence. Both fungi and bacteria are key decomposers in soil, though they differ in the types of carbon substrates they utilize (Lauber et al., 2008). In this study, both management systems received equivalent amounts of nitrogen, though in different forms: the conventional system received 100% as mineral nitrogen, while only 40% of the nitrogen applied in the EI system was in soluble form.

Despite this, bacterial abundance did not significantly differ between treatments. In contrast, fungal abundance was markedly higher under EI management. This is likely due to the incorporation of lignified eggplant crop residues and mature compost, which favor fungal proliferation (Ciaccia et al., 2019; Denef et al., 2001). Consequently, the fungal-to-bacterial ratio increased under EI, consistent with observations by Malik et al. (2016). Intensive fertilization, as commonly practiced in conventional systems, is known to reduce the fungal to bacterial ratio by favoring bacterial dominance (de Vries et al., 2006; Parfitt et al., 2012). Additionally, the Fungal to Bacteria ratio is strongly influenced by the carbon-to-nitrogen (C: N) ratio (Xiao et al., 2022; Waring et al., 2013). In our study, the C to N ratio rose from 5.98 to 9.1 under EI management, potentially contributing to the higher fungal abundance observed.

Under EI management, we observed increases in the copy numbers of key denitrification genes, particularly nosZ1 and nosZ2 (nitrous oxide reductases). This finding aligns with previous studies indicating that organic nitrogen inputs promote denitrifier communities (You et al., 2022; Ouyang et al., 2018). Nevertheless, gene abundance is not solely dependent on fertilization type. Recent research has highlighted the importance of pH and TOC as major drivers of microbial abundance and diversity (Dong et al., 2023; Xiao et al., 2021). Other factors such as crop developmental stage, soil moisture, and salinity also influence denitrification genes abundance, particularly in Mediterranean greenhouse systems (Hernández Maqueda et al., 2024; Hernández Maqueda et al., 2025).

A notable finding in both treatments was that the ratio of N₂O-consuming (nos genes) to N₂O-producing (nir genes) organisms was greater than 2, suggesting a potential to minimize NO₃ ⁻ accumulation and N₂O formation. However, caution is warranted when interpreting this indicator, as denitrification is only one of several microbial pathways that contribute to N₂O emissions (Dong et al., 2023).

The observed microbial shifts under EI management suggest that fungi play an increasingly central role in carbon cycling and potentially in denitrification processes. While a direct correlation between fungal abundance (measured via ITS gene copies) and denitrification genes was not detected in this study, the concurrent increases in ITS, nosZ1, and nosZ2 gene copy numbers under EI management point to a possible association. This highlights a need for further research into the functional roles of fungal communities in denitrification. Recent work has emphasized that the role of fungi in nitrogen cycling has been underestimated (Aldossari and Ishii, 2021), particularly in systems where fungi dominate decomposition processes, such as those amended with compost and crop residues.

Despite the observed improvements in soil organic matter, microbial abundances, and structural stability (as indicated by increased soil resistant aggregates), EI management was also associated with a remarkable rise in soil electrical conductivity (EC). This suggests a potential salinization risk. The use of eggplant residues, which are high in mineral salts, and compost derived from greenhouse crop residues contributed significantly to the increase in soil salinity. In this case study, Mg²⁺, Na⁺, and K⁺ increased considerably, whereas Ca²⁺ decreased. A possible explanation for the reduction in Ca²⁺ is that the presence of organic acids derived from the incorporation of crop residues may have promoted the leaching of this cation, as observed in the study by Jones et al. (2012).

To mitigate this risk, it is advisable to reconsider the types of organic amendments used. Specifically, avoiding high-salt biomass inputs and exploring alternative combinations of fresh and mature organic matter may help maintain soil quality without increasing salinity. This approach would also reduce the need for salt leaching, a practice that may have adverse effects on groundwater quality.