The study area was located in the Yellow River Delta Dongying City, Shandong Province (118°44’2”E, 38°1’18”N), which is the temperate semi-humid continental monsoon climate with annual precipitation of 550-640 mm and an average annual temperature of 12.6°C. The dominant plants are Phragmites communis, S. salsa, Aeluropus sinensis, and Tamarix chinensis. The terrain is relatively flat, mostly sandy, and muddy soil, which is easy to compact and has poor ventilation.

A 1m×1m sample square was set up within a single plant, and a 5m buffer strip was set up between the sample squares for N and P fertilizer addition treatment in May 2021. Five N-P addition ratio treatments CK (0 kg N ha^-1, 0 kg P ha^-1), N0P (0 kg N ha^-1, 10 kg P ha^-1), N1P (MN, 50 kg N ha^-1, 10 kg P ha^-1), N2P (HN, 150 kg N ha^-1, 10 kg P ha^-1), N3P (HN, 450 kg N ha^-1, 10 kg P ha^-1), and each treatment consisted of three repetitions. The selection of the N-P addition ratio was referred to a previous study (Liu et al., 2019). Previous studies have found that urea addition has a stronger effect on soil N cycling than ammonia and nitrate N addition (Li et al., 2019a; Wang et al., 2022b). Therefore, urea was used as N fertilizer. KH₂PO₄ was used as P fertilizer by referring to Li et al. (2019a). Urea and KH₂PO₄was first dissolved in pure water and then sprayed evenly on the target area, respectively. CK treatment was sprayed with pure water.

Sample collection was conducted in July 2021, approximately 3 months after the N and P addition. The rhizosphere soil was carefully collected from soil attached to the root system (0-0.5 cm from the root) of all plants. The bulk soil was collected from the soil at a depth of 15cm on the ground without plant growth. A total of 30 soil samples were collected, and then each soil sample was divided into two parts after removing impurities. One part was naturally dried for the determination of soil physicochemical properties. Another part was stored at -80°C for soil microbial analysis.

Soil organic matter (SOM) content was measured by the potassium dichromate volumetric method. Total N (TN) and total carbon (TC) content were measured using an elemental analyzer (Vario EL III, Elementar, Germany). Soil pH and salt were measured by pH meter and conductivity meter, respectively. Total P (TP) content was determined using the molybdenum blue method with an ultraviolet-visible spectrophotometer, and the available P (AP) content was measured by NaHCO₃ extraction-molybdenum antimony anti-colorimetry.

Total genomic DNA from samples was extracted using CTAB method by OMEGA DNA kit, and the purity and concentration of DNA were detected by 1% agarose gel electrophoresis, and the samples were diluted to 1 ng/µl with sterile water in a centrifuge tube. The V3-V4 region of the bacterial 16S rRNA gene was amplified with specific primers 341F and 806R. The products were detected by 2% agarose gel electrophoresis, and purified by using AxyPrep DNA Gel Extraction Kit. The MiSeqPE300 platform of Illumina company (Wekemo Tech Group Co., Ltd. Shenzhen China.) was used for high-throughput sequencing. Qiime2 processed the original DNA data, and the obtained effective sequences were clustered by OTU for further analysis.

Two-way ANOVA followed by LSD multiple-comparison test was used to analyze the effects of rhizosphere and N-P addition ratio on soil characteristics, indices of community diversity, and abundance of dominant taxa. The Mantel test was used to determine the effect of soil characteristics on bacterial community diversity, dominant bacterial phyla, and dominant bacterial genera groups. PERMANOVA (per=999) was conducted to test the effect of different N-P addition ratio treatments on bacterial community composition. Pearson analysis was used to correlate soil characteristics with bacterial communities.

All data were processed based on R software (4.1.0), and Two-way ANOVA was performed using the basic command aov. LSD was performed using the agricolae and multcomp package. Mantel test was performed using the linkET package. PERMANOVA and PCoA analyses were performed using the vegan package. The correlation analysis was performed using the corrplot package. Image processing was performed using the ggplot2 package.

The results of the α-diversity analysis of the rhizosphere and bulk soil bacterial community in S. salsa under different N-P addition ratio additions were shown in Tables 1 , 2 , and it was observed that a-diversity indices were significantly different between rhizosphere and bulk soil. N-P addition ratio had a significant effect on Chao1, Observed OTUs and Shannon index (P<0.05), Chao1 and Observed OTUs index were also significantly affected by the interaction of N-P addition ratio and Soil (P<0.05). N and P addition significantly increased the richness of the bulk soil bacterial community, the bacterial evenness was reduced in the high N-P addition ratio treatment.

The PCoA result showed that the first principal components accounted for 78% of the data matrix variance ( Figure 1 ), and the explanations were 56.9% and 21.1% for PCoA1 and PCoA2, respectively. The results show that the rhizosphere and bulk soil were clearly separated, and a significant difference in the distance was found after the PERMANOVA test ( Table 3 , R² = 0.364, P=0.001). The rhizosphere bacterial community was more concentrated, while the bulk soil bacterial community was more dispersed, indicating that the N-P addition ratio had less effect on the rhizosphere than bulk soil bacterial community. PERMANOVA test ( Table 3 , R² = 0.637 P=0.001) indicates N-P addition ratio has a significant effect on bacterial community composition in bulk soil.

The Proteobacteria was stable in the rhizosphere and bulk soil with the highest relative abundance ( Figure 2 ), followed by the Bacteroidetes, Actinobacteria, Gemmatimonadetes, Acidobacteria and Chloroflexi phylum. The relative abundance of Bacteroidetes and Verrucomicrobia was significantly higher in the rhizosphere than in bulk soil (P<0.01, P<0.001), while the Actinobacteria, Acidobacteria and Gemmatimonadetes were significantly lower than in bulk soil (P<0.001, P<0.001, P<0.01). Two-way ANOVA ( Table 4 ) showed that the Actinobacteria, Acidobacteria and Gemmatimonadetes relative abundance differed significantly between rhizosphere and bulk soil. In rhizosphere soil, the N-P addition ratio had small effects on the top five bacterial phyla. In contrast, in bulk soil, N and P addition significantly increased the relative abundance of the Gemmatimonadetes and Bacteroidetes ( Figure 3 ). The difference is that the Bacteroidetes increased with increasing N-P addition ratio and the Gemmatimonadetes in contrast. N and P addition reduced the relative abundance of the Acidobacteria. The highest relative abundance of the Acidobacteria was observed under the N0P treatment, while it decreased with the co-addition of N and P.

The relative abundance of rhizosphere and bulk soil with relative abundance of bacterial genera greater than 1% is shown in Figure 4 . The number of abundant genera in the rhizosphere and bulk soil was 6 and 8, respectively, where Marinobacter, Subgroup_10, bacterium_YC_LK_LKJ36 and Woeseia were present in both soils. In rhizosphere soil, the relative abundance of bacterium_YC_LK_LKJ36, Marinobacter and Saccharospirillum were highest in the N0P treatment, while the relative abundance of all three decreased with increasing N-P addition ratio. In contrast, the Salinimicrobium was lowest in the N0P treatment, while N and P addition increased its relative abundance, however, the relative abundance decreased with increasing N-P addition ratio. In bulk soil, N and P addition decreased the relative abundance of MND1 and Marinobacterium and increased the relative abundance of bacterium_YC_LK_LKJ36. The relative abundance of Marinobacter was lower at low N-P addition ratio and higher at the high N-P addition ratio.

Soil environmental factors of rhizosphere and bulk soil added with different N-P addition ratios are shown in Figure 5 . All environmental factors except C: N were significantly different between the rhizosphere and bulk soil ( Figure 5 , Table 5 , P<0.05). The pH and TP content of rhizosphere soil were lower than those of bulk soil, while the C: N, C: P, N: P and salt, SOM, TC and TN were higher than those of bulk soil. Under CK and N0P treatments, AP showed that rhizosphere soil was larger than bulk soil, while the opposite was found under N and P addition.

In the rhizosphere soil, the AP, pH, SOM, TC, and TN content showed an increase followed by a decrease with increasing N-P addition ratio, with the highest in the N0P treatment and the lowest in the N3P treatment. In bulk soil, pH decreased with increasing N-P addition ratio. N and P addition significantly increased soil AP, SOM, TC, TN, and TP (P<0.05) and decreased soil N:P, C: N, and N:P (P<0.05). SOM content was highest in the N0P and N2P treatment and decreased in the N3P treatment, but all were significantly higher than the CK (P<0.05).

The α-diversity index, the top 8 bacterial phyla, and the genera with relative abundance greater than 1% are used as the bacterial community in the Mantel test ( Figure 6 ). In rhizosphere soil, soil environmental factors had no effects on bacterial α-diversity and bacterial community phylum, while TP, C:P, and N:P had significant effects (P<0.05) on bacterial enrichment genus. The bacterial community is strongly influenced by environmental factors in bulk soil, with high contributions from SOM and pH.

The correlation analysis between soil environmental factors and bacterial community ( Figure 7 ) showed that N: P, salt, SOM, TC, and TN content were positively correlated with bacterial α-diversity index (P<0.001), while Chao1 and observed OTUs were negatively correlated with pH (P<0.05). The Bacteroidetes were positively correlated with salt, SOM, TN, and TC, while the Actinobacteria were the opposite of them. The woeseia was most strongly correlated with soil environmental factors.

Plant root secretion of mucus, shed cells and tissues, and soluble products of ruptured cells all contribute to the TC and SOM of rhizosphere soil (Zhao et al., 2022). In addition, sugars and amino acids secreted by plant roots increase N availability in the soil (Zhao et al., 2022). In this study, the rhizosphere soil contained higher TN, SOM, and TC content compared to the bulk soil, which is evidence that the plant rhizosphere effect is at work. Soil pH was in the range of 7.21 to 7.56, showing weak alkalinity, exhibiting less rhizosphere than bulk soil, while salt content was higher in the rhizosphere than in the bulk soil. S. salsa was a salt aggregating plant, which can gather salt into its body without damage, thus the rhizosphere soil has higher salt content. Previous studies have found that S. salsa is useful in lowering soil pH, and this was again demonstrated in our study.

Both rhizosphere and bulk soil pH gradually decreased with the increase of the N-P addition ratio. Previous research (Huang et al., 2016; Sha et al., 2021) has found that N and P addition leads to soil acidification and pH decrease, which is consistent with the result of this study. As the N-P input ratio increased, the SOM content increased and then decreased. Previous studies have found that low levels of N and P can enhance SOC storage in the nutrient-limited ecosystem (Pascault et al., 2013; Jian et al., 2016). However higher N-P addition ratio may accelerate SOC mineralization. Li et al. (2022a) found that there is a threshold value of the N-P addition ratio between 15:1 and 45:1, which once exceeded, causes the loss of SOC, which is consistent with our study.

At the CK and N0P treatment, the AP content showed higher in the rhizosphere than in bulk soil. This may be due to the root secretion in the rhizosphere soil, which promoted the P activation and availability (Zhao et al., 2022). This may be caused by plant uptake of P. The addition of N reduces the availability of soil P (Yang et al., 2015). Zhu et al. (2016) found that unbalanced input of N and P could potentially alter nutrient availability and enhance P limitation, which is consistent with our study.

We found that the bacterial community α-diversity was significantly higher in rhizosphere soil than in bulk soil, which is consistent with the previous study. This is due to the presence of the plant root system, some organic compounds are deposited into the soil, providing more carbon and N source for the bacterial community, thus promoting bacterial growth and increasing diversity (Ran et al., 2021). In addition, healthy soil environment promote the growth of bacteria. According to the analysis of soil environmental factors, the higher SOM and TN content of rhizosphere soil can provide sufficient nutrients for bacterial growth.

There was no significant effect of N-P ratio addition on the rhizosphere bacterial community diversity. Li et al. (2019b) studied the effect of N addition on the diversity of rhizosphere soil microbial community in different soil types and found that N addition had minimal effects on bacterial community diversity, which is consistent with our study. Previous studies found that the decrease in bacterial diversity due to N addition was mainly caused by the decrease in soil pH (Liu et al., 2020b). In the bulk soil of this study, although pH also decreased significantly as mentioned in previous studies, bacterial community richness increased and bacterial diversity first increased and then decreased with increasing N-P addition ratio. This may be due to the low or medium N-P addition ratio treatment where the pH did not reach the threshold that limits the α-diversity of the bacterial community, while the lower pH under the high N-P addition ratio treatment limited the bacterial growth. The effect of fertilizer supply on microbial diversity was influenced by changes in soil SOC under both original soil conditions and the fertilizer application time, while variations in soil pH had less effect (Wang et al., 2018a). Yu et al. (2016) used structural equation modeling to propose that soil SOC is generally considered as a key factor affecting soil microorganisms. In our study, we found that N and P addition significantly increased soil SOM and decreased at high N-P addition ratio. The Mantel test again proved that SOM was the most important factor affecting the α-diversity of bulk soil bacteria. The rhizosphere soil SOM was high in S. salsa, so bacterial diversity did not change significantly between treatments.

Soil bacterial community composition is related to soil nature and the interaction between plants and microorganisms (Weand et al., 2010). PCoA analysis and PERMANOVA test based on Bray-Curtis distance showed a large difference in bacterial community composition between rhizosphere and bulk soil, which is consistent with previous studies. Liu (Liu et al., 2020a) studied the rhizosphere and bulk soil bacterial community of S. salsa in two habitats of the YRD and suggested that root secretion and apoplast play a key role in shaping the bacterial community composition.

The N-P addition ratio increase due to increased N input had insignificant effects on rhizosphere bacterial community composition, which is consistent with the previous study (Li et al., 2019b). However, the N-P addition ratio significantly altered the bacterial community composition of bulk soil. In general, the lower β-diversity of the rhizosphere bacterial community indicated that it was more stable than bulk soil (Tian et al., 2022). In terms of the ecological building process, the stochastic process contributed more to the rhizosphere than the bulk soil bacterial community, where the difference in SOM content might be the reason for the difference in nutrient addition to the bacterial community construction process (Tian et al., 2022). Both the SOM initial content and the degree of SOM content variation in response to nutrient addition influenced the relative contribution of different ecological construction processes (Feng et al., 2018). Specifically, when soil SOM content was high, nutrient addition homogenized the bacterial community composition. In contrast, when SOM initial content was low or SOM response increased significantly with nutrient addition, nutrient addition changes bacterial community composition (Feng et al., 2018). In this study, bacterial community composition was not significantly altered due to the high SOM content of rhizosphere soil. In addition, microorganisms and plants evolved synergistically and showed high host specificity (Li et al., 2019b), which may also be a key factor contributing to the insignificant effect of the N-P addition ratio on the structure of the inter-rooted microbial community.

At the phylum level, the highest Proteobacteria relative abundance was found in both soils, which is consistent with previous studies (Liu et al., 2020a), which indicates that Proteobacteria have excellent adaptability to the environment. Bacteroidetes and Actinobacteria are eutrophic taxa. We predict that these two taxa will have relatively high abundance in rhizosphere soil, while oligotrophic taxa represented by Acidobacteria will have high relative abundance in bulk soil (Kanokratana et al., 2011; Fierer et al., 2012; Redding et al., 2016; Ren et al., 2017; Sha et al., 2021). The rhizosphere soil had high soil nutrients, which stimulated the growth of the Bacteroidetes, while limiting the Actinobacteria, which further supports the bacterial nutrient survival strategy. However, in contrast to Bacteroidetes, Actinobacteria had a lower relative abundance in the rhizosphere soil, which may be due to the difference in salt tolerance between them. In the current study, the Actinobacteria and the Bacteroidetes were significantly negatively and positively correlated with salt, respectively, which is consistent with the previous study (Zhao et al., 2020). Bacteroidetes have better salt tolerance and resistance, while Actinobacteria are less salt tolerant or more susceptible to the negative effect of high salt (Zhao et al., 2020). Despite the high nutrient content in rhizosphere soil, high salinity reduces the Actinobacteria competitiveness to compete for nutrients with Bacteroidetes.

With an increasing N-P addition ratio, the bulk soil Gemmatimonadetes and Acidobacteria relative abundance gradually decreased, while Bacteroidetes relative abundance increased, which further reflected the trophic strategy of bacteria. The elevated N-P addition ratio increased the availability of utilizing C sources in the soil, which promoted and inhibited the development and reproduction of the phylum Bacteroidetes and Acidobacteria, respectively. Gemmatimonadetes have a strong denitrification ability with strong denitrification ability (Guo et al., 2017). However, the present study result is contrary to the previous study. The previous study showed that the abundance of Gemmatimonadetes increased with increasing N application in water-N integrated environmental conditions (Guo et al., 2017), therefore, we speculate that Gemmatimonadetes is more sensitive to soil moisture, while in terrestrial, the trend of Gemmatimonadetes decreases with increasing N-P addition ratio due to insufficient moisture.

At the genus level, Marinobacter, Subgroup_10, bacterium_YC_LK_LKJ36, and Woeseia had high abundance in both soils. Marinobacter, Subgroup_10, and Woeseia have salt-tolerant bacteria, indicating that soil salinity has a strong shaping effect on bacterial genera (Li et al., 2017). The Marinobacter, a common denitrifying bacterium in the Yellow River Delta, is capable of denitrification using nitrate or nitrite as electron acceptors (Chi et al., 2021; Zhou et al., 2021). In this study, the relative abundance of Marinobacter was increased in bulk soil, indicating that the imbalanced input of the N-P ratio enhances bacterial denitrification.