Fertilization enhances rice productivity by promoting phosphorus uptake and altering soil microbiota

Fertilization can enhance crop yield and improve soil health. However, its effects on nutrient uptake, soil microbiota, and rice yield remain unclear. Herein, we designed a double-cropping system with different fertilization treatments to determine their contributions to paddy soil agroecosystem and rice yield. Soil samples were collected before planting the early rice, after harvesting early rice, and after harvesting the late rice. Soil physicochemical properties, and rice yield, and rice nutrient (total nitrogen, phosphorus, and potassium contents) were determined. Variation of soil microbiota were also determined by high-throughput sequencing. We found that soil potassium content significantly improved during the planting process (P < 0.05), while rice phosphorus displayed significant variation under fertilization (P < 0.05). Notably, late rice yield was significantly higher than early rice yield (P < 0.001) after fertilization treatment. Furthermore, rice yield was positively correlated with available soil phosphorus (P < 0.05), indicating that fertilization promoted phosphorus uptake. Organic fertilizer altered the soil microbiota and increased Chloroflexi phyla abundance, while organic fertilizer combined with a compound microbial agent increased the diversity of soil microbial communities. A partial least squares path model revealed that fertilizer treatment directly positively affected rice yield by influencing phosphorus uptake and Shannon index (P < 0.01). Collectively, this study demonstrates that organic fertilization with compound microbial agents can stabilize soil nutrients, increase soil microbial diversity, and improve rice yield, thereby offering a guide for enhancing fertilizer utilization and improving agroecosystems.


Introduction
Along with wheat and corn, rice is one of the three major staple crops grown in China, with its cultivation area and production accounting for over 30% and approximately 40% of the total national grain planting area and production, respectively [1].The use of chemical fertilizers affects crop yields [2], however, excessive or imbalanced application of these fertilizers has led to significant soil-related issues, such as nutrients loss [3], and soil degradation [4].
In contrast, organic fertilizers, compared to chemical fertilizers, encompass a wider array of nutrients essential for crop growth and development.They achieve this by improving the physical and chemical properties of soil and augmenting soil nutrient content [5].Furthermore, organic fertilizers can reduce soil bulk density, improve soil aggregate stability [6], and enhance soil organic carbon, porosity and water holding capacity [7].Consequently, they are extensively utilized in paddy soil to enhance rice yield.The substitution of chemical fertilizers with organic counterparts not only augments soil organic matter content but also curtails soil compaction [8].Furthermore, reducing nitrogen fertilizer usage while increasing base fertilizer application, such as urea, calcium superphosphate and potassium chloride, has been shown to have no adverse effects on rice yield and can simultaneously mitigate soil nitrogen loss [9].
In agroecosystems, soil microorganisms are also considered important soil nutrient recycling drivers [10,11].Microbial agents include specific living microorganisms that enhance crop growth and yield through microbial metabolism [12,13].The application of microbial agents with organic fertilizers can improve soil fertility [14] and assist in nutrient uptake by crops; therefore, they are important for reducing the application of chemical fertilizers and improving soil ecology [15].Furthermore, microbial agents containing functional bacteria can increase soil phosphatase activity, thereby enhancing the available phosphorous and potassium contents in the soil [16].
With respect to rice research to date has primarily focused on the impact of organic fertilizer application, nitrogen reduction, and various other methods on rice yield.However, there has been limited characterization of how different fertilization methods affect soil nutrient levels and rice yield.Additionally, soil microorganisms play essential role in nutrient cycling, such as carbon and nitrogen cycle, and that soil biodiversity is a vital contributor to ecosystem functioning [17], yet, the influence of soil microbiota on rice productivity and plant nutrition in the context of fertilization remains poorly understood.
To gain further insights, in this study, we conducted double-season rice cultivation which is a common planting pattern in south-China, and assessed the effects of the following distinct treatments: no fertilization (CK), conventional fertilization (CT), conventional fertilizer reduction with organic fertilizer (OF), and organic fertilizer combined with a compound microbial agent (M-OF).Our analysis aimed to determine the following: (1) the effects of these treatments on soil and rice nutrition; (2) their impact on soil microbiota and rice yields; and (3) the relationships among fertilization, plant and soil nutrition, and rice yields.We hypothesized that: (1) organic fertilizer and organic fertilizer combined with a compound microbial agent would contribute to enhancing soil nutrients; (2) organic fertilizer combined with a compound microbial agent would have more effect on the soil microbiota, thereby improving rice yields.By addressing these aspects, our findings in this study will contribute to developing a theoretical framework for enhancing rice yield stability and optimizing fertilizer utilization.

Site description and experimental design
The study area was located in Heping Village, Lingxi Town, Wenzhou City, Zhejiang Province (27°30′N, 120°21′ E), China.The field experiment was established in 2017.The cropping pattern was double-season rice, and the soil type was clay.The average annual precipitation is 1753.9mm and the average annual temperature is 16.1℃.Four different treatments were established (Table S1), including no fertilization (CK); conventional fertilizer with urea (N by mass fraction = 46%), calcium superphosphate (P 2 O 5 by mass fraction = 12%), and potassium chloride (K 2 O by mass fraction = 62%) (CT); 30% reduction of the conventional fertilizer treatment with added organic fertilizer (OF); and 30% reduction of the conventional fertilizer treatment combined with organic fertilizer and microbial agents (M-OF).Each plot measured 33 m 2 (500 × 667 cm), with concrete poured between the plots to avoid disturbance.Each treatment was performed in triplicate, and using of randomized complete block design (figure S1).
The rice (Oryza sativa L.) used in the study, was transplanted into the experimental plot on April 13, 2022, following the fertilization of the paddy soil on April 12, 2022.The harvest for the early rice took place on July 18, 2022, while the Late rice was transplanted to the experimental plot on July 19, 2022, and subsequently harvested on October 26, 2022.The transplanting density for both early and late rice was maintained at 330,000 plants per hectare, with identical fertilization rates for both early and late season rice crops.

Sampling and chemical measurements
Soil samples were collected (from 0-20 cm depth) before planting the early rice (11 April 2022), labelled as Before, after harvesting the early rice (18 July 2022), labelled as Middle, and after harvesting the late rice (26 October 2022), labelled as After.
The soil physicochemical properties were determined using the methods described by You et al [18].Soil pH values were measured using a standard pH meter (HT-PHJ 962-2018) at a soil-water ratio of 1:2.5, whereas soil moisture content was measured on the basis of mass loss after drying at 105 °C for 24 h.Soil total nitrogen (TN), total phosphorus (TP) and total potassium (TK) were determined using Kjeldahl method, alkali fusion molybdenum-antimony resistance spectrophotometry and sodium hydroxide melting and flame photometry, respectively.Available phosphrous (AP) and available potassium (AK) were determined using molybdenumantimony colorimetry and flame photometry, respectively.Soil organic carbon (SOC) was determined using the potassium dichromate method, and soil organic matter (SOM) was calculated by multiplying the obtained SOC value by a constant value of 1.724.
Rice yield were calculated by weighing the rice in each plot after harvesting.After digestion of the rice grains (after harvesting) with H 2 SO 4 -H 2 O 2 ; total nitrogen (TN), total phosphorus (TP), and total potassium (TK) were determined using the Kjeldahl method, vanadium-molybdenum yellow colorimetric method, and flame photometry, respectively [19,20].

DNA extraction and high-throughput sequencing
Soil DNA was extracted from fresh soil (0.5 g) using E.Z.N.A™ Mag-Bind Soil DNA Kit (Thermo Fisher Scientific, USA) following the manufacturer's instructions.
The bacterial 16 S rRNA genes targeting primer pairs 338 F (ACTCCTACGGGAGGCAGCA)/ 806 R (GGACTACHVGGGTWTCTAAT) in V3-V4 regions were used for sequence analysis via the Illumina Miseq platform at Shanghai Sangon Biotechnology Co., Ltd.(Shanghai, China) [21].The amplification conditions were as follows: an initial denaturation at 98 °C for 60 s, 60 °C for 30 s and 12 cycles at 72 °C for 30 s, followed by 5 min at 72 °C.To assess the bacterial alpha diversity in the soil, we utilized several metrics, including the Shannon index, Chao1, abundance-based coverage estimators (ACE), and Simpson index [22].These metrics were employed in conjunction with the Silva database to analyze variations in the composition of soil microbiota.Additionally, Operational taxonomic Units (OTUs) were assigned using the Quantitative Insights into Microbial Ecology (QIIME) platform, with similarity set at 97% sequence similarity.

Statistical analysis
Data were calculated as mean and standard deviation using Excel 2021.Principal Component Analysis (PCA) and analysis of variance was carried out to detect significant differences using the SPSS 23.0 software (IBM Corp., Armonk, NY, USA).Statistical significance was defined by 95% confidence intervals with a p value smaller than 0.05 (two-tailed).Bar graphs were plotted using Origin 2021 (OriginLab, Northampton, MA, USA), and R (version 6.0) was used to generate correlation scatter plots and the partial least squares path model (PLS-PM) in the ggplot2 and plspm packages, respectively.PLS-PM evaluation was assessed according to the methods described in a previous study [18].

Changes in soil chemical properties
Under different fertilization treatments, the pH remained stable, ranging from 5.25 to 6.44 (table 1).There was a significant increase in soil moisture during the early and late rice harvest periods (designated 'Middle' and 'After,' respectively), particularly in the Middle-CK, After-CK and After-OF, and After-M-OF treatments (table 1; P < 0.05).
Although there were no significant differences in soil nutrients between the CK, CT, OF, and M-OF treatments, soil organic matter (SOM) showed a significant drop during the double-cropping planting process (figure 1; P < 0.05).The TP showed a similar trend (P < 0.05), whereas AP reduced in the middle of the process before seeing a significant rise in late rice harvest.During this process, the potassium content also increased, especially the AK content (P < 0.001).

Effects on crop yield and rice nutrients
The potassium nutrient content in the paddy soil increased in the different fertilizer treatments compared to the CK group; however, the TK content in rice grains leveled off, ranging 3.55-4.17g/kg.The TN in the different fertilizer treatments ranged 15.2-16.5 g kg −1 and showed a slight increase compared with that in the CK group (figure 2).
The TP content in rice grains significantly differed under different fertilizer applications.There was a significant increase in TP in the CT treatment group (figure 2; P < 0.05) compared with that in the CK group and a highly significant decrease in the OF and M-OF treatment groups compared with that in the CT group (P < 0.01).
The change in rice nutrients revealed that the CT treatment exhibited greater benefits in terms of nitrogen, phosphorus, and potassium uptake by the rice grains when compared to the other fertilizer treatments.Furthermore, it effectively promoted phosphorus uptake by the rice plants.The data represent the means ± standard deviation of three independent replicates, and the values provided are mean ± standard deviation.Abbreviations: CK (no fertilizer treatment), CT (conventional fertilizer treatment), OF (organic fertilizer treatment), and M-OF (organic fertilizer treatment with microbial agents).pH (soil acidity/alkalinity), SOM (soil organic matter, g/kg), TN (total nitrogen, g/kg), TK (total potassium, g/kg), TP (total phosphorus, g/kg), AK (available potassium, mg/kg), AP (available phosphorus, mg/kg).Before (before the double-cropping planting process), Middle (in the middle of the double-cropping planting process), and After (after the double-cropping planting process).* P < 0.05, compared with Before period.
Organic fertilizers significantly affected rice yield (figure 3).The OF treatment produced the highest yield of early rice, and the M-OF treatment produced the highest yield of late rice.
The OF and M-OF treatments showed significant increases in early rice yield compared to the CK treatment (P < 0.05).The CT, OF, and M-OF treatments showed highly significant increases in late rice yield compared to the CK treatment (P < 0.01).
In conclusion, our findings demonstrate that the various fertilizer applications within the double-season rice cropping pattern exert a substantial influence on the yield of the double-season rice crop.Notably, fertilization proves to be significantly more effective in enhancing the yield of late rice as compared to early rice.It's worth (conventional fertilizer treatment), OF (organic fertilizer treatment), and M-OF (organic fertilizer treatment with microbial agents).pH (soil acidity/alkalinity), SOM (soil organic matter, g/kg), TN (total nitrogen, g/kg), TK (total potassium, g/kg), TP (total phosphorus, g/kg), AK (available potassium, mg/kg), AP (available phosphorus, mg/kg).Before (before the double-cropping planting process), Middle (in the middle of the double-cropping planting process), and After (after the double-cropping planting process).noting that although the CT treatment involved a higher quantity of nutrients, treatments involving organic fertilizers (OF and M-OF) yielded more positive effects on rice yield.
In addition, diversity indicators of the soil revealed significant differences under M-OF and OF treatments (P < 0.05; figure 4).Especially, M-OF treatment increased soil microbial diversity, while OF treatment decreased the Chao1, ACE, and Shannon indices (figures 4(b)-(e)).However, PCA analysis revealed a clustering of CK, CT and M-OF treatments with respect to soil microbial community composition, while OF treatment clustered separately in the ordination plot (figure 4(f)), suggesting that organic fertilization addition without microbial agent could decline the soil microbial diversity.

Linkages between fertilizer treatments, nutrient uptake, and rice yield
The relationship between soil nutrients and rice yield was determined using scatter plots (figure S2).A significant positive correlation was found between rice yield and AP (r = 0.785, P < 0.05; figure S2f), whereas no significant correlation was observed with soil nutrients in the other treatments.Given that different fertilizations had little effect on soil nutrients (table 1) but significantly affected rice yield (figure 3), a partial least squares path model (PLS-PM) was constructed to further verify the relationships among fertilization, nutrient output, soil microbiota, and rice production.
The PLS-PM analysis revealed that fertilizer application can impact rice yield by influencing the phosphorus nutrient content and the diversity of the soil community.More specifically, the PLS-PM indicated a notably positive effect of fertilization on rice yield (as shown in figure 5; P < 0.01) and elucidated the pathway through which this effect contributes to the enhancement of rice yield.
We found that fertilizer treatment has a positive influence on soil AP (P < 0.01), whereas it had a direct negative effect on rice TP and microbial community diversity (Shannon index).However, rice TP and Shannon index were observed to have highly significant effects on rice yield (P < 0.01).Thus, although fertilization was characterized an indirect negative effect on rice grains, the fertilizer treatment had an overall enhancement effect on rice productivity mainly by influencing phosphorus content and microbial diversity (figure 5(a)).

Discussion
While the effects of different fertilizer application patterns on soil fertility enhancement and plant yield have been established, the link between fertilizer reduction through organic fertilizer or the combination of organic fertilizer with microbial agents and their impact on soil nutrition and rice yield has remained unclear.Our findings emphasize the significance of fertilizer application as a pivotal factor that influences alterations in soil phosphorus content.Furthermore, we established a strong correlation between rice yield, soil phosphorus content, and fertilizer application.
In this study, the rice yield was positively correlated with soil AP (figure S2; P < 0.05), PLS-PM further verified that fertilizer application was positively associated with AP (figure 5; P < 0.01).The absence of a significant correlation between yield, TN, and TK implies that the oversupply of nitrogen and potassium in the soil, combined with the limited presence of available nitrogen and available potassium, are not conducive to plant uptake [23,24], and thereby cannot significantly increase crop yield.Additionally, a similar lack of response in nutrient uptake has been observed in intensive crops like potatoes and sugarcane, which could be attributed to the inherent properties of the soil [25,26].Although there were not significant differences among CT, OF and M-OF in soil phosphorus content, interestingly, CT treatment significantly increased the uptake of total phosphorus in rice (figure 2; P < 0.01).Nobile et al [27] demonstrated that the application of organic fertilizer increased phosphorus absorption and promoted the change from moderately active organic phosphorus to active organic phosphorus.Wang et al [28] suggested that phosphatic fertilization improved the available P content in paddy soil, which is consistent with the results of this study.These results indicate that conventional fertilizers, which contained more phosphorus content, may make a more contribution to crop nutrients than organic fertilization [29].
Additionally, although the fertilizer application rate of CT in this study was higher than that of OF and M-OF, we found that the rice yield was significantly improved under OF and M-OF treatments compared with that under CK (figure 3), indicating that reduced fertilizers did not affect soil nutrient intake but improved nutrient utilization.Furthermore, the effect of M-OF on the yield of late rice was greater than that of OF (figure 3; P < 0.05), which may attribute to compound microbial agent under M-OF treatment potential to enhance soil quality and promote yields [30].
Moreover, we found that the Shannon index plays a crucial role in rice yield, as it displayed a significant positive effect on rice yield (figure 5).Previous studies have suggested that organic fertilization greatly affects soil microbiota structure and their activities in nutrient cycling [31,32] and can adjust soil microbial composition and enhance soil enzyme activity [33].In this study, the abundance of the bacterial phylum Chloroflexi increased under organic fertilization, while the alpha diversity, indicated by Chao1, ACE, and Shannon indices, decreased.Conversely, M-OF significantly enhanced the soil microbial diversity (P < 0.05), although soil microbial composition did not largely vary (figure 4).Generally, the Shannon index reflects the variation degree among the soil community [34].The Shannon index is usually higher for microbial communities with a larger number of species [35].Numerous studies have demonstrated the positive impact of microbial biodiversity on crop growth and production.For instance, Yu et al [36] revealed that both diazotrophic alpha-and beta-diversity have a positive influence on wheat yield.Besides, correlation heamap revealed that Proteobacteriaand Gemmatimonadtes was positively correlated with soil TP and TN respectively(P < 0.05).Additionally, M-OF treatment consisting of Bacillus subtilis, Bacillus gelatinoides, and Bacillus amylolyticus, exhibits phosphatesolubilizing functions [37], facilitating the transformation of soil phosphorus and enhancing plant phosphorus uptake [27].Given the significant role of soil microorganisms in soil ecology, our findings suggest that microbial agents may contribute to the stabilization of soil fertility and enhance nutrient utilization by crops, ultimately leading to improved rice yield.
Additionally, in our study, we observed that Soil Organic Matter (SOM) levels did not increase significantly under the OF and M-OF treatments, as indicated in table 1. SOM represents the primary constituent of the soil organic-mineral complex and serves as the reservoir for soil nutrients.Therefore, changes in organic matter content can be indicative of alterations in soil fertility [38].Mi et al [39] demonstrated that long-term application of organic fertilizer significantly increased SOM content.Mockeviciene et al [40] concluded that SOM, alkaline nitrogen, fast-acting phosphorus, and fast-acting potassium content were increased under the application of complex microbial fertilizer.Yin et al [41] suggested that the application of organic fertilizer and bio-organic fertilizer in combination significantly increased SOM content.This is inconsistent with the results of this study, which may be attributed to the washout by rainfall [42] and the shorter application cycle [2].

Conclusions
The double-season rice patterns positively impacted the soil's TK, AK, and AP contents under various fertilization conditions (P < 0.05).Notably, The CT treatment led to a significant increase in the TP content of rice grains (P < 0.05), while both OF and M-OF treatments also substantially improved rice yield (P < 0.05), indicated organic fertilzaiton enhance nutrient utilization.Moreover, among these treatments, the M-OF treatment triggered an increase in soil microbial diversity, facilitated soil fertility stabilization and thereby exhibiting pronounced effect in improving rice yield.The PLS-PM model further verified that fertilizer application had a significant direct positive effect on rice yield (P < 0.05) and phosphorus nutrition (especially in soil AP), and Shannon index is also positively associated with the rice yield (figure 5).These findings underscore the continued significance of fertilizer application rates as a limiting factor for achieving high rice yields.Meanwhile, other rice yield parameters like panicles, panicle lengths, filled panicles and unfilled ones which can also effected yield differences.Consequently, further research is warranted to investigate the effects of phosphorus content within fertilizer application regimes and the proportions of organic fertilizers combined with microbial compounds on soil ecosystems, taking into consideration a broader spectrum of environmental factors, such as temperatures and light intensity.Additionally, considering the pivotal role played by soil microbiota in the soil nutrient cycle, there is a need for a deeper exploration of the relationship between organic fertilizers and microbial agents concerning soil enzyme activity and the structure of the soil microbial community in paddy fields.The insights obtained from this study provide valuable guidance for enhancing fertilizer utilization and optimizing agroecosystems.

Figure 2 .
Figure 2. Total nitrogen (TN), total phosphorus (TP), and total potassium (TK) content with different fertilization treatments in rice nutrients.

Figure 3 .
Figure 3. Yield of double-cropping rice under different fertilization treatments.

Figure 5 .
Figure 5. Partial least squares path model (PLS-PM) of the (a) fertilizer treatments affecting rice yield and (b) the standardized total effect of the fertilizer treatments on rice yield.

Table 1 .
Changes in soil physicochemical properties under the double-cropping system.