The initial soil sample was collected from the surface layer in Ninghe, Tianjin, China (39°25′N, 117°29′E). This fluvo-aquic soil is representative of the North China region. Its physicochemical properties were characterized as follows: total nitrogen (TN) = 0.91 g kg⁻¹, total phosphorus (TP) = 0.63 g kg⁻¹, available potassium (AK) = 0.34 g kg⁻¹, available phosphorus (AP) = 41.89 mg kg⁻¹, pH = 8.43, and electrical conductivity (EC) = 0.28 mS cm⁻¹. After being transported to the greenhouse at the Agro-Environmental Protection Institute in Tianjin, the soil was mixed and screened to eliminate large stones and other solid impurities. HU was sourced from a urine-diverting toilet in the male restroom at the Institute’s dormitory. The HU was collected, containerized, and stored at 25°C for a duration of 3 months. Subsequent analysis revealed its physicochemical properties as: TN = 3.62 g L⁻¹, TP = 0.26 g L⁻¹, TK = 1.18 g L⁻¹, pH = 9.18, EC = 34 mS cm⁻¹, and chemical oxygen demand (COD) = 12.40 g L⁻¹, sodium ions (Na⁺) = 1.35 g L⁻¹, chloride ions (Cl⁻) = 2.23 g L⁻¹.

Pots of uniform size (23 cm in height and 20 cm in diameter) were filled with 5 kg of the homogenized and sieved soil and categorized into five treatment groups: C, T1, T2, T3, and T4. Each group received a different volume of HU: 0, 65, 130, 260, and 520 mL for C, T1, T2, T3, and T4, respectively. The HU application rates were primarily based on its nitrogen content, calculated by referencing the nitrogen fertilizer usage for protected horticultural crops as reported by Zhang et al (Zhang et al., 2018). Treatments were replicated three times. Thorough mixing of the soil and HU was achieved using a mechanical stirrer. To standardize soil moisture across treatments, additional water was added to bring the combined volume of HU and tap water in each pot to 2 L. After a 2-day equilibration period, five pakchoi seeds were sown into each pot. Following germination, thinning was performed to retain only three seedlings per pot. All pots were consistently irrigated to maintain optimal soil moisture levels for plant growth.

After a 40-day growth period, pakchoi plants were harvested (with biomass data presented in Supplementary Figure S1 ), and the top 0- to 10-cm layer of soil was collected from each pot. Five soil cores from each pot were combined into a single sample, whereas plant debris and other contaminants were removed. These soil samples were then temporarily stored in sterilized, sealable bags within an incubator and promptly transported to the laboratory. Each consolidated sample from the pots was subsequently divided into three sub-samples. The first sub-sample was thoroughly mixed, sifted through a 2-mm sieve, placed in a sterile 10-mL centrifuge tube, and sent to Shanghai Majorbio Bio-Pharm Technology Co., Ltd., for microbial diversity analysis. The second sub-sample was used to assess the soil’s physicochemical properties and enzymatic activities after being ground and sieved. The third sub-sample was utilized to determine the concentrations of soil ammonium nitrogen (NH₄ ⁺-N) and nitrate nitrogen (NO₃ ⁻-N).

All data were conducted for verify normality and homoscedasticity of variance using SPSS 21.0 Statistics software (IBM Corporation, NY, USA) (Pu et al., 2022). For the data that conformed to the normal distribution and homogeneity of variance test, analysis of variance was conducted again to determine the significant differences among the treatments (Huang et al., 2019). The bar graphs were created using Origin 7.0. The “vegan” package was used for calculating alpha diversity and richness of microorganisms (Oksanen et al., 2017). The statistical method of bacterial diversity index difference was tested by Kruskal–Wallis H. The abundance differences between groups were analyzed using the R (v. 3.3.1) for statistics and mapping. The significance difference between groups was tested by ANCOM difference test (QIIME2) (Nearing et al., 2022). Non-metric multidimensional scaling (NMDS) was carried out using the Bray–Curtis dissimilarity distance to analyze microbial community diversity. Analysis of similarities (ANOSIM) was employed for the significance of separations measured under different treatments (Huang et al., 2019). A model of multivariate analysis of variance was constructed using distance-based redundancy analysis (RDA) (Francioli et al., 2016). ANOSIM, RDA, and Mantel tests were performed using R (v.4.0.3).

The application of HU to soil induced variations in several parameters such as soil TN, TP, AP, AK, C/N, pH, and EC as shown in Table 1 . The content of NH₄ ⁺-N remained statistically consistent across all treatment groups. TN and TP levels were highest in the T4 treatment, showing increases of 46.59% and 20.90%, respectively, compared to the C. AP and AK also experienced enhancements at various rates with higher volumes of HU, reaching their maximum in T4, which were 377% and 143% higher, respectively, than in C. The C/N ratio decreased under all treatments with T3 exhibiting the most significant reduction, 29.34% lower than C. Similarly, pH values decreased, with T3 showing the most substantial drop from 8.47 in the C to 7.66. EC was greatest in T4, showing an increment of 840%. Moreover, SOM levels slightly increased with greater HU addition.

As the volume of HU increased, the enzymatic activities of INV, URE, and CAT initially rose and subsequently declined ( Figure 1 ), whereas AKP activity showed a consistent decrease. Increasing the HU volume from 65 mL to 260 mL (T1 to T3) resulted in a linear increase in INV activity, with increments of 7.30% in T1 and peaking at 58.75% in T3, compared to the control (C, 0 mL). However, further increasing the HU volume to 520 mL (T4) led to a reduction in INV activity increment to 39.17% ( Figure 1A , Supplementary Figure S2A ). Trends in URE activity paralleled those observed for INV activity ( Supplementary Figure S2B ). Relative to C, URE activities in T1, T2, and T4 increased by 7.10%, 17.66%, and 0.93%, respectively, with T3 exhibiting the highest URE activity at 275.48 μg d⁻¹ g⁻¹, marking a 47.77% increase ( Figure 1A ). CAT activity reached its maximum in T1 (50.81 μmol d⁻¹ g⁻¹), which was 16.62% higher than C, but diminished in subsequent treatments (T1–T4) reaching its lowest level in T4 at 37.89 μmol d⁻¹ g⁻¹ ( Figure 1C ). Moreover, AKP activity exhibited a linear decline from C to T4 ( Supplementary Figure S2D ), with decreases of 6.40%, 9.45%, 34.25%, and 64.76% across T1 to T4, respectively, compared to C ( Figure 1D ).

Venn diagrams are commonly utilized to delineate the shared and unique operational taxonomic units (OTUs) across multiple samples. In our study, subsequent to stringent quality filtering, we amassed 630,747 high-fidelity 16S rRNA gene sequences across 15 samples, with individual samples contributing between 31,858 and 54,312 sequences. Post-chimera removal, non-redundant sequences were clustered into OTUs (excluding singletons) based on a 97% similarity criterion. From this process, a total of 4,356 OTUs were identified. Cluster analysis revealed that the bacterial communities were distributed among 33 phyla, 93 classes, and 827 genera. Additionally, a core set of 2,039 OTUs was present across all samples, whereas individual treatments T1, T2, T3, and T4 contained 3,498, 3,454, 3,219, and 2,902 OTUs, respectively ( Figure 2 ).

Figure 3 illustrates that, in comparison to the C, the diversity in other groups decreased variably with increasing HU levels. Both the Shannon–Wiener diversity index and the Evenness index recorded the lowest values in T4, at 6.0815 and 0.7973, respectively. These represent decreases of 8.47% and 5.70% below the C values. The richness index initially increased slightly before declining as the HU addition was augmented. The peak increase in this index was observed in T1, where it reached 3,356.51, marking a 0.74% increase over the C value of 3331.68. A linear regression model was used to explore the linear relationship between the effects of HU on soil quality and bacterial diversity. It is mainly reflected that HU content has a significant positive correlation with SOM (P<0.001) and TN (P<0.05) content, and a significant negative correlation with C/N (P<0.05) and Shannon (P<0.001) ( Supplementary Figure S3 ).

The distributions of soil bacterial abundances reveal that, at the phylum level ( Figure 4A ), the three predominant bacterial phyla were Pseudomonadota (28.79%), Actinomycetota (24.42%), and Chloroflexi (13.79%). These principal groups were accompanied by lesser abundances of Gemmatimonadota (7.21%), Bacteroidota (3.28%), Bacillota (2.98%), and Rokubacteriota (2.20%). In each experimental group, Pseudomonadota abundances were prominently dominant and exhibited an increase concurrent with higher HU additions. Specifically, compared to the control group (C), Pseudomonadota abundances rose by 9.81%, 5.43%, and 23.19% in T1, T2, and T3, respectively, with the most significant surge observed in T4, showing a 40.19% increase over C. Conversely, Actinomycetota displayed a decrease across all treatment groups (T1–T4) correlating with increased HU volumes, with declines of 6.82%, 9.72%, 20.30%, and 20.32% in T1–T4, respectively, relative to C. Similarly, Chloroflexi abundance trended downward, with T4 registering the lowest percentage of 10.50%, which is 30.37% lower than C. Acidobacteriota also showed a decreasing pattern across the treatments, with T4 presenting the minimal proportion of 5.69%, marking a substantial decrease of 64.24% in comparison to the C group. In contrast, with the escalation of HU concentration, both Gemmatimonadota and Bacteroidota phyla demonstrated increments in their richness. Gemmatimonadota richness increased by 15.99%, 27.71%, and 26.64% in T1, T2, and T4, respectively, when compared to C, with T3 experiencing a notable surge in richness (69.80% higher than C). Likewise, Bacteroidota saw increments of 4.05%, 7.21%, and 61.71% in T1, T2, and T3, respectively, with T4 showcasing the most considerable augmentation in richness.

At the genus level, as shown in Figure 4B , predominant bacteria with a relative abundance exceeding 1.0% included unclassified Subgroup_6 members (7.52%), Arthrobacter (4.60%), unclassified JG30-KF-CM45 members (3.41%), unclassified Gemmatimonadaceae members (3.32%), and Sphingomonas (3.20%). Relative to C, the abundance of Sphingomonas increased between 1.10% and 47.25%, reaching a peak of 4.02% in the T3 treatment. Among the top 20 genera, such as Gaiella (1.49%), Bacillus (1.47%), and Blastococcus (1.02%), significant impacts were observed due to HU treatment. With the administration of HU, both Gaiella and Blastococcus showed marked declines. At the 520-mL application rate in T4, the relative abundances of Gaiella and Blastococcus decreased to 1.12% and 0.97%, representing reductions of 42.86% and 15.65% compared to C, respectively. In contrast, Bacillus reached a relative abundance of 3.37% in T4, which is 324% higher than that observed in C. ANCOM difference test analysis shows that g:Eoetvoesia, g:Truepera, g:Luteimonas, and g:norank_f:B1_7BS were significantly different among groups ( Figure 5 ), with W of 823, 821, 800, and 781, respectively ( Supplementary Table S5 ).

The β-diversity of the soil bacterial communities was assessed using NMDS and ANOSIM. The NMDS analysis, visualized in Figure 6 , reveals variations in the composition of the soil bacterial communities across different treatments (C, T1, T2, and T3) in comparison to T4 based on the abundance of OTUs. Notably, significant differences were observed in the structure of the bacterial communities among C, T1, T2, T3, and T4 treatments (R = 0.6533, P = 0.001), as illustrated by the clear separation and independence of samples in the NMDS plot. The distributions observed within the plot—a leftward skew in C, T1, and T2, versus a rightward skew in T4-highlight the substantial impact of HU on the structure of soil bacterial communities.

Table 2 presents the relationships between the relative abundances of the top 10 most abundant bacterial phyla and the enzymatic activities of four different soil enzymes. Gemmatimonadota were found to be strongly and statistically significantly correlated with INV and URE activities, whereas Pseudomonadota, Bacillota, and Planctomycetota displayed a pronounced positive association with AKP. Notably, Gemmatimonadota and Pseudomonadota exhibited marked negative correlations, in contrast to the positive correlations for Bacillota and Planctomycetota. In addition, Actinomycetota were negatively correlated with INV while showing a substantial positive association with AKP. The Chloroflexi and Acidobacteriota groups were strongly positively correlated with CAT and AKP activities. Conversely, Bacteroidota showed a negative relationship with CAT activity and were strongly inversely correlated with AKP. Lastly, Rokubacteriota had significant strong negative correlations with both CAT and AKP activities.

RDA of bacterial community structures correlated with soil physicochemical properties, as depicted in Figure 7 , revealed that Axis1 and Axis2 explained 63.18% and 8.59% of the variation in soil bacterial communities, respectively. Collectively, both axes accounted for 71.77% of the overall variation. Mantel test outcomes demonstrated that TN (R² = 0.7524, P = 0.001), NH₄ ⁺-N (R ² = 0.8027, P = 0.001), and SOM (R² = 0.8069, P = 0.005) were the primary physicochemical parameters significantly affecting the bacterial community structure.

The application of HU led to a decrease in soil pH, which can be attributed to the nitrification of ammonium ions within the soil. This process involves the conversion of ammonium salts into nitrites, releasing two protons and subsequently lowering the soil pH (Schönning, 2001). Despite varying volumes of HU being employed, the soil’s NH₄ ⁺-N levels remained unchanged. This stability is likely due to the nitrogen within the HU transitioning from urea to ammonia, urine-N would be lost to NH₃ evaporating (Rumeau et al., 2023), and a similar phenomenon also exists in the use of liquid ammonium fertilizers (Powlson and Dawson, 2022). Another possible reason is the leaching after nitrification due to high urinary N application; after NH₄ ⁺-N is converted to NO₃ ⁻ through nitrification, NO₃ ⁻ may be carried with the water to the deeper layers of the soil due to leaching (Ramírez-Sandoval et al., 2022), which is consistent with the results of our study ( Supplementary Table S6 ). Conversely, an increase in soil EC was observed with higher volumes of HU, attributed to the enhanced presence of conductive ions such as chloride ions in the HU. Consequently, as the volume of HU applied escalated, so did the concentration of ions, resulting in an increased soil EC (Kassa et al., 2018). The addition of HU resulted in significant increases in soil TN and TP, indicative of the high nitrogen and phosphorus content of HU. Therefore, extensive application of HU as a fertilizer substantially elevates the soil’s nitrogen and phosphorus levels (Bonzi et al., 2011). SOM also saw a linearly increment with increased HU applications ( Supplementary Figure S3A ), potentially contributing to the rise in soil TN levels. Additionally, soil AK content experienced increments at various degrees with the escalation of HU volume, a phenomenon likely tied to HU’s high potassium concentration. Thus, the application of greater volumes of HU correspondingly increased the soil potassium level. Moreover, soil NO₃ ⁻-N content demonstrated a gradual decline with an increase in HU volume, as the lowered soil pH from HU application likely expedited the denitrification rate of NO₃ ⁻-N (Pradhan et al., 2009).

Soil enzymes are instrumental in the cycling of nitrogen, carbon, and phosphorus in the soil and in activating the plant’s defense mechanisms (Fan et al., 2024; Zhu et al., 2024; Yan et al., 2025). (Tao et al., 2015) noted that the application of organic fertilizers, as opposed to chemical ones, significantly enhanced soil enzyme activities such as INV and URE. HU, as a nutrient-rich organic fertilizer, has been shown to increase the activities of soil INV and URE with increasing application volume (Krause and Rotter, 2018). Consistently, our results demonstrate pronounced enhancements in the activities of INV and URE, alongside significant correlations with changes in AP, AK, SOM, C/N, and EC ( Supplementary Table S1 ). Furthermore, SOM is considered a critical substrate for enzyme synthesis and plays a vital role in stimulating soil enzyme activity (Wei et al., 2015). Such dynamics underscore the close association between soil nutrient status and enzyme activities (Demisie et al., 2014). In the present study, the application of HU corresponded with a linearly reduction in soil C/N ratios ( Supplementary Figure S3C ), aligning with observations by (Deng et al., 2020) Under certain circumstances, the soil C/N ratio can serve as an indicator of SOM quality (Qin et al., 2023). A significant negative correlation was observed between soil INV and URE activities and the C/N ratio (p < 0.01) ( Supplementary Table S1 ), consistent with findings by Li et al (Li et al., 2017). We noted that the C/N ratio decreased progressively at treatments T1, T2, and T3 but surged dramatically at T4, which was associated with a suppression of soil INV and URE activities ( Table 1 ; Figures 1A, B ), this might be that adding too much urine reduces the abundance of Gemmatimonadota, resulting in the inhibition of urease activity ( Table 2 ), which is consistent with the previous research results (Cheng et al., 2020). Additionally, previous studies indicated a positive correlation between Acidobacteriota and AKP activities, and a negative correlation between CAT activities and Bacillota (Ren et al., 2021). Soil AKP activity was found to be inhibited, likely due to increased urine content which subsequently reduced the soil pH. AKP is highly sensitive to soil acidity, and its activity is diminished under low pH conditions. Overall, both the soil physicochemical properties and microbial communities were shown to significantly influence soil enzyme activities.

Moreover, the germination and seedling phases are particularly susceptible during the course of plant development, with salinity serving as a significant environmental constraint on growth, as documented by (Tang et al., 2015) In our study, application rates of HU at 52 mL kg⁻¹ and 104 mL kg⁻¹ (T3 and T4 treatments) inhibited normal germination and growth of pakchoi ( Supplementary Figure S1 ). A contributing factor may be the observed sharp increase in soil electrical conductivity to levels between 1.78 and 3.29 mS cm⁻¹, which represents an elevation of 78% to 229% compared to the T2 treatment. This heightened soil salinity can exert detrimental osmotic pressure, impeding plant development (Acosta-Motos et al., 2017). Furthermore, the inherent sodium and chloride content of HU may also play a role; as application rates rise, the subsequent accumulation of these ions could disrupt nutrient balance across cellular membranes and ultimately prove toxic to the pakchoi plants (Negrao et al., 2017). Consequently, meticulous regulation of HU application rates is imperative in future agricultural practices to circumvent its potential phytotoxic effects and safeguard crop yields.

The Shannon–Wiener diversity index (S index) serves as a metric for assessing microbial diversity, where a greater value of the S index signifies enhanced community diversity. In the present study, elevated levels of HU application were correlated with a consistent reduction in the S index during pakchoi cultivation, suggesting that HU application attenuates soil bacterial diversity. This attenuation may be attributable to the increase in soil conductivity and salinity induced by HU, as soil microorganisms are acutely sensitive to shifts in salinity within their habitat. Such increases in salinity can impose osmotic stress upon soil microbes, thereby diminishing microbial diversity (Chen et al., 2022). Concurrently, the Evenness and Richness indices exhibited a generally declining trend as shown in Figure 3 , implying that adding HU to the soil during the cultivation of pakchoi reduces bacterial richness. Prior research indicates that soil microbial life is considerably influenced by pH levels (Geisseler and Scow, 2014). In our investigation, the rise in soil EC and the drop in pH associated with greater volumes of HU might be contributing factors to the observed alterations in soil bacterial diversity and community composition (Wu et al., 2020).

Analysis of community structure demonstrated that the abundances of Actinomycetota, Chloroflexi, and Acidobacteriota decreased to varying extents. Given that Acidobacteriota predominantly adopt an oligotrophic lifestyle while HU is nutrient-rich (Ding et al., 2018), it is suggested that HU application could enhance soil fertility and suppress the growth of these bacteria. Conversely, the relative abundances of Pseudomonadota, Bacteroidota, and Gemmatimonadota were observed to increase alongside rising HU volumes. This trend may be attributable to the higher soil fertility requirements of these bacteria, which were met through HU application, thus providing an optimal growth milieu (Guo et al., 2018). Additionally, Chloroflexi have the capability to convert nitrites into nitrates (Daims et al., 2015), whereas Pseudomonadota, Bacteroidota, and Gemmatimonadota possess denitrification capabilities (Ren, 2018). These observations suggest that HU application skewed nitrogen cycling toward reduced nitrification but enhanced denitrification, aligning with the observed declines in NO₃ ⁻-N within soil physicochemical attributes. Furthermore, genera such as Sphingomonas and Bacillus were noted to elevate the abundance of genes linked to nitrogen metabolism and potentially reduce ammonia nitrogen consumption by influencing nitrogen transport (Sun et al., 2020; Yang et al., 2021). Sphingomonas was also recognized for its capacity to degrade a broad spectrum of harmful compounds (Gatheru et al., 2017). This might be attributable to HU application elevating the soil’s ammonia-nitrogen levels, subsequently increasing Sphingomonas and Bacillus populations, a finding corroborated by (Wang et al., 2022) Moreover, belonging to Actinomycetota, Gaiella was mentioned for its role in soil nitrate assimilation and in facilitating the uptake of amino acids and other nutrients by plants (Leite et al., 2021). These predominant genera are posited as significant ecosystem contributors, notably in soils treated with HU (Roy et al., 2022). Consequently, HU application has a direct impact on soil physicochemical characteristics and nutrient content, inducing environmental shifts that influence soil bacterial communities.

Soil enzymes, predominantly synthesized by microorganisms with minor contributions from plants and other soil biota, have been found to play pivotal roles in the nutrient cycling and metabolism within soil ecosystems (Liu et al., 2022). The production of enzymes, which often varies with shifts in microbial community composition, can exert different influences on soil processes. (Baldrian et al., 2010; Qu et al., 2025) reported that soil microorganisms—and, consequently, enzymatic activities—are subject to regulation by a constellation of complex factors that govern microbial community dynamics, with microbial biomass and enzyme activity often showing significant correlation within the upper tens of centimeters of the soil profile (Baldrian et al., 2010). Correlation analyses revealed that the relative abundance of Gemmatimonadota was strongly correlated with INV and URE activities, whereas the relative abundances of Pseudomonadota, Bacillota, and Planctomycetota were closely associated with AKP activity. The former phyla exhibited notably strong negative correlations, whereas the latter displayed significantly positive correlations. These patterns may arise from the enhancement of SOM due to HU amendments, which, in turn, increase bacterial abundance and alter community structure (Jiao et al., 2013). The findings underscore the varied and intricate interconnections between soil enzymatic activities and bacterial communities, underscoring the need for more comprehensive investigation into the specific relationships between distinct bacterial phyla and soil enzymes.

Excessive application of HU for irrigation could potentially lead to soil salinization and eventually promote land erosion (Kassa et al., 2018; Chapman, 1992). Concurrently, risks associated with ammonia volatilization and pathogen contamination persist throughout the urine reuse process (Martin et al., 2022). To address these concerns, it is recommended to either dilute urine or pre-treat HU using appropriate chemical agents such as calcium hydroxide and magnesium hydroxide to convert it into a stable and useful solid form of fertilizer. These chemicals are readily available and cost-effective. Additionally, the introduction of calcium and magnesium can markedly diminish pathogen levels due to the elevated pH values in HU (Kishor et al., 2020). Furthermore, advanced oxidation processes including ozone treatment and electrochemical oxidation have proven effective in sterilization and in decomposing micropollutants such as pharmaceuticals, antibiotics, and hormones in HU (Krishnan et al., 2021).