Soil Bacteria to Regulate Phoebe bournei Seedling Growth and Sustainable Soil Utilization under NPK Fertilization

Soil bacteria play a key role in the plant–soil system and can regulate the growth of Phoebe bournei seedlings under fertilization. However, there are few reports on how soil bacteria respond to fertilization and regulate seedling growth. This study adopted the “3414” field fertilization experiment, combined with soil microbial sequencing, nutrient contents, and biomass measurement, to explore the changes of soil chemical properties and bacterial structure under different NPK fertilization conditions and to establish the coupling relationship between soil bacteria, soil nutrients, and plant growth. The results showed that NPK fertilization decreased soil pH; increased soil N, P, and K content; reduced bacterial diversity and abundance; promoted the growth of dominant bacterial species; and enhanced Phoebe bournei seedlings’ soil N, P, and K elements. NPK fertilization promoted Proteobacteria growth, especially of three genera (Methylobacterium, Sphingobium, and Acinetobacter) and Actinobacteria, while it decreased Acidobacteria and Chloroflexi. By reducing the ratio of N to K and increasing P, NPK fertilization can slow soil acidification, promote bacterial reproduction, maintain P. bournei seedlings’ soil ecological stability, and balance the seedlings’ growth and sustainable soil utilization. AD3, Pseudomonas, and Rhodanobacter can be used as the marker species for N, P, and K fertilization, respectively, while Methylobacterium, Brevundimonas, Acinetobacter, and Sphingobium can be used as indicator species for soil pH and soil N, P, and K content changes, respectively. These results provided a theoretical basis and technical guidance for the effective fertilization and cultivation of robust P. bournei seedlings.


Introduction
Soil microorganisms, the main driving force of soil material circulation, have a direct relationship with soil fertility and play an essential role in determining soil properties and affecting plant growth [1][2][3]. Bacteria account for about 70~90% of the soil microbial population [4,5] and participate in a variety of soil physical and chemical processes, such as decomposition of soil organic matter, formation of humus, and transformation and circulation of nutrients. Soil bacteria not only perform soil material transformation but also store plant nutrient elements and can be used as the measurement index of soil quality [6][7][8][9]. At the same time, bacteria are sensitive to environmental changes and can also be used as an early indicator of soil changes [10]. Thus, studying soil bacteria is particularly important.
Soil bacteria play an important role in the material circulation and energy flow of the plant-soil system [11]. Changes in the soil environment are expected to lead to a change in bacterial adaptability as well as an effect on its community structure [12]. Fertilization Means ± SD with different letters indicate a significant difference among the 14 treatments, as determined by Duncan's multiple range test (p < 0.05). R means the difference between the maximum and minimum values at different levels for single-factor, two-factor, and three-factor effects of N, P, and K fertilizer. R N is the difference at 4 N levels (T2, T3, T6, and T11), R P at 4 P levels (T4, T5, T6, and T7), R K at 4 K levels (T8, T9, T6, and T10), R NP at 8 NP levels (T2, T3, T4, T5, T6, T7, T11, and T12), R NK at 8 NK levels (T2, T3, T6, T8, T9, T10, T11, and T13), R PK at 8 PK levels (T4, T5, T6, T7, T8, T9, T10, and T14), and R NPK at 14 NPK levels (T1-T14).

Effect of NPK Fertilization on Phoebe bournei Seedlings' Soil Bacterial Diversity
Different NPK fertilization conditions significantly affected soil microbial Chao1, Simpson, and Shannon indices (Table 2). Chao1, Simpson, and Shannon indices of the control (T1) were higher than all fertilization treatments, indicating that fertilization reduced Phoebe bournei seedlings' soil microbial diversity, with T10 showing the lowest values for the three indices, indicating that soil bacteria diversity was inhibited by a high-K fertilizer level. The three (Chao1, Simpson, and Shannon) diversity indices for the three-factor fertilizer (N, P, and K) were the highest, followed by the two-factor and single-factor fertilizers, with P having the least effect on soil diversity indices. Means ± SD with different letters indicate a significant difference among the 14 treatments, as determined by Duncan's multiple range test (p < 0.05). R means the difference between the maximum and minimum values at different levels for single-factor, two-factor, and three-factor effects of N, P, and K fertilizer. R N is the difference at 4 N levels (T2, T3, T6, and T11), R P at 4 P levels (T4, T5, T6, and T7), R K at 4 K levels (T8, T9, T6, and T10), R NP at 8 NP levels (T2, T3, T4, T5, T6, T7, T11, and T12), R NK at 8 NK levels (T2, T3, T6, T8, T9, T10, T11, and T13), R PK at 8 PK levels (T4, T5, T6, T7, T8, T9, T10, and T14), and R NPK at 14 NPK levels (T1-T14).

Differential Groups of Phoebe bournei Seedlings' Soil Bacteria under NPK Fertilization
The Anosim significance test showed that N, P, and K three-, two-, and single-factor fertilization treatments had significantly different impacts on soil bacterial community structure, with the following order: NPK, N, P, NP, NK, PK, and K fertilization ( Figure 2). Anosim significance test of community structure difference between Phoebe bournei seedlings' different fertilization groups. R-value ranges between −1 and 1 (values >0 indicate a significant difference between groups).

Network Association Analysis of Phoebe bournei Seedlings' Soil Microbial Community under NPK Fertilization
Network modularity segmentation of bacterial communities in Phoebe bournei seedlings soil was performed under the various fertilization conditions ( Table 3). The modularity index was greater than 0.7 for the seven NPK fertilization combinations, indicating that the community had a modular structure. Number of nodes, number of edges, and closeness centrality in bacterial community were the highest under NPK, NP, and NK, followed by N, PK, K, and P, fertilization. Transitivity coefficient and average path length in the bacterial ecosystem under P was the highest, followed by K, PK, the remaining three fertilization conditions (NPK, NP, and NK), and N fertilization, indicating that single-factor fertilization was the lowest in network stability, followed by two-factor fertilization; the highest stability was with three-factor fertilization. Under single-factor fertilization, N was the most stable network, followed by K and P; this could be due to the improved diversification of three-factor fertilization, leading to enhanced interaction and stability among bacteria. Number of nodes: the number of nodes in the target network; number of edges: the number of edges in the target network; closeness centrality: the ratio of the average distance between the target node and all other nodes; average path length: the sum of all the short paths in the network; transitivity: the probability of connecting the target node and the adjacent node; modularity: the modularity index.
The network association among microbial members was visualized using the R graph package, and nine key phylum species were found in different NPK fertilization treatments, including Proteobacteria, Deinococcy-Thermus, Chloroflexi, Actinobacteria, Acidobacteria, Nitrospirae, Bacteroidetes, Gematimonadetes, and WPS-2 ( Figure 6). . Scatter plots of soil bacterial marker species of Phoebe bournei seedlings under NPK fertilization treatments. The node in the network can be divided into four parts using Zi and Pi values, namely, peripherals, connectors, module hubs, and network hubs. Peripherals represent some specialists in microbial networks. Module hubs and connectors represent species that are close to generalists. Network hubs represent super-generalists among the microbial networks.

Driving Factors of Phoebe bournei Seedling Growth under NPK Fertilization
PLS-PM results showed that the growth of Phoebe bournei seedlings was significantly affected by NPK accumulation, which was driven by three aspects, namely, the significant and positive effect of NPK fertilizers, the positive effect of soil chemical properties, and the negative effect of bacteria growth. According to the absolute value of the correlation coefficient, the factors influencing P. bournei seedling growth were soil bacteria, soil chemical character, and NPK fertilization, indicating soil bacteria was the most important factor for P. bournei seedling growth (Figure 7). Further correlation analysis of soil dominant phylum and genera bacteria indicated that Methylobacterium, Brevundimonas, Sphingobium, and Acinetobacter had significant or highly significant negative correlation with Actinobacteria, Chloroflexi and Acidobacteria and significant or highly significant positive correlation with Proteobacteria, which indicated that Methylobacterium, Brevundimonas, Sphingobium, and Acinetobacter were the key species regulating the soil ecology of Phoebe bournei seedling soil (Table 4).  Correlation analysis of soil key bacteria and chemical traits showed that pH was negatively correlated with Methylobacterium, Brevundimonas, Sphingobium, and Acinetobacter abundance, among which significant correlation with observed for Methylobacterium and Acinetobacter. Soil NPK contents were positively and significantly correlated with the abundance of four key bacteria (N content with Brevundimonas and Acinetobacter, P content with Acinetobacter, and K content with Sphingobium) ( Table 5).

NPK Fertilization Regulation of Phoebe bournei Seedlings' Soil Chemical Characters
Fertilization is the main influencing factor for soil properties, causing changes in soil nutrients and pH [28]. NPK fertilization reduced P. bournei seedlings' soil pH, with the largest change under N and the least under P; these results are consistent with previous observations [29,30]. NPK fertilization can reduce soil acidification by decreasing and increasing the proportion of N and P, respectively. NPK fertilization significantly increased P. bournei seedlings' soil N and K contents, possibly due to promoting N-and K-rich microorganism growth. Soil P content can be significantly increased only when appropriate NPK application is applied, especially by reducing the K ratio. K fertilizer increased soil K content and the amount of soil H + due to the increased competition between the binding points of H + and K + on the plasma membrane, thus enhancing P activation capacity [31][32][33] and, consequently, increasing P. bournei seedlings' P uptake, resulting in reduced soil P content. Hence, soil P content in T8 (N 2 P 2 K 0 ), T9 (N 2 P 2 K 1 ), and T6 (N 2 P 2 K 2 ) was higher than that of T1 (N 0 P 0 K 0 ), with T8 representing the highest soil P content among the 14 fertilization treatments. Multiple range analysis showed that the three-factor NPK fertilization had the greatest influence on soil NPK content, followed by double-factor and single-factor fertilization, indicating that NPK could improve the sustainable soil productivity of P. bournei seedlings. P appeared to be the key factor in regulating P. bournei seedlings' stem growth [34]. Therefore, reducing the K fertilization proportion in NPK fertilizer combinations can improve the seedlings' P uptake.

NPK Fertilization Regulation of Phoebe bournei Seedlings' Soil Bacterial Diversity
Soil bacterial diversity is closely related to soil functions. Protecting soil bacterial diversity can effectively ensure the existence of a balanced and stable soil ecosystem [13][14][15]. Fertilization level, type, duration, and application method can individually and/or in concert change soil physical and chemical properties and, subsequently, affect soil bacteria species and diversity [17,35]. In the present experiment, NPK fertilizer acidized the soil and increased its nutrient content, resulting in the enrichment of some soil bacterial growth along with a reduction in their diversity; these results support previous observations [17][18][19]. Soil bacterial diversity under three-factor fertilization was higher than two-and single-factor fertilization, indicating that three-factor fertilization provided the most balanced nutrients for P. bournei seedlings' soil bacteria and satisfied the growth and metabolism requirements of more bacteria types, a keeping high diversity level. At the same time and as previously reported, P fertilizer had the least effect on soil bacteria diversity [36]. Therefore, reducing the ratio of N and K and increasing P can maintain P. bournei seedlings' soil ecological stability under NPK fertilization.

Soil Bacteria Response to Phoebe bournei Seedlings' NPK Fertilization
NPK fertilization significantly affected P. bournei seedlings' soil bacterial community composition. The abundance and network correlation analyses of soil bacteria indicated that Proteobacteria, Actinobacteria, Chloroflexi, and Acidobacteria were the dominant and core phylum species, playing an important role in soil material circulation and ecological environment balance [37,38]. NPK fertilization significantly increased soil fertility and the available N and P contents in the soil, resulting in increased Proteobacteria abundance, as affected by the available N, and decreased Acidobacteria abundance, mainly due to poor soil environment [39,40]. Urea, as the N source in our experiment, was conducive to increasing P. bournei seedlings' nitrite absorption and reducing the nitrite content in the soil, thus decreasing Chloroflexi abundance, which oxidizes nitrites into nitrates, increasing autotrophic nitrite bacteria growth (e.g., the best Chloroflexi growth was in T2, which did not include N fertilizer) [41,42]. Actinobacteria are known to be beneficial bacteria that produce agricultural antibiotics and fight plant diseases [43]. Similar to most NPK fertilization results, high N levels resulted in reduced Actinobacteria abundance; however, an appropriate NPK ratio can promote Actinobacteria growth, reaching maximum abundance in T14 (N 2 P 1 K 1 ), probably because NPK fertilizer provides material and energy in relatively arid red soil [40,44]. Cupriavidus, Methylobacterium, Brevundimonas, Sphingobium, Acinetobacter, and Sphingomonas were the dominant genera in Phoebe bournei seedling soil, and they all belong to Proteobacteria, which indicates that Proteobacteria have strong adaptability to NPK fertilization. Methylobacterium, Sphingobium, and Acinetobacter can fully absorb soil nutrients, providing NPK fertilization as substances and energy to promote their reproduction. However, Brevundimona propagated best under the application of high-level P fertilizer (T7, N 2 P 3 K 2 ), indicating that it had a high phosphorus-dissolving function [45]. Sphingomonas grew best without K fertilizer (T8, N 2 P 2 K 0 ), indicating that it had a weak potassium-releasing capacity [46].
Soil bacteria can reflect changes in soil in time, such as soil nutrients and pH value, and can temporarily reflect the quality of soil [18,19,28,47]. With a relative abundance of more than 1%, Bacteroidetes was only used as the marker phyla species for NPK threefactor fertilization, while there was no marker phyla species for two-and single-factor fertilization, indicating that the phyla species could not be used as the marker species for NPK fertilization. Under NPK three-factor fertilization, Rhodanobacter, Sphingononas, and AD3 were used as the marker genera species, with a relative abundance of more than 1%. Under two-factor fertilization, AD3, Pseudomonas, and Chujaibacter were used as the marker genera species, with a relative abundance of more than 1% for N and P fertilization; AD3, Pseudomonas, Cupriavidus, and Rhodanobacter for N and K fertilization; and Thermus, AD3, Sphingomonas, and Rhodanobacter for P and K fertilization. Under singlefactor fertilization, AD3 was the marker bacteria for N, with relative abundance over 1%; Thermus and Pseudomonas for P fertilizers; and Rhodanobacter for K fertilizer. These indicate that the genera species can be used as the marker species for NPK fertilization.

Soil Bacteria Regulate P. bournei Seedling Growth and Soil Development
Soil microorganisms play an important role in soil nutrient cycling and the soil-plant growth system [11]. Plants, soil, and microorganisms are not independent in soil material circulation and energy exchange processes but interact with each other in a synergetic manner. NPK fertilization significantly increased soil N, P, and K contents, directly promoting P. bournei seedling growth and indirectly improving soil bacteria growth. On the other hand, NPK fertilization directly and significantly inhibited soil bacteria growth; however, it did promote dominant bacteria growth by enriching nutrient elements, promoting N, P, and K element storage in the soil bank, and providing nutrients for subsequent seedling growth. Hence, soil bacteria was the most important factor for P. bournei seedling growth and the sustainable utilization of soil. Methylobacterium, Brevundimonas, Sphingobium, and Acinetobacter are the key species for regulating soil fertilization. Methylobacterium and Acinetobacte had a significant negative correlation with soil pH, Brevundimonas a significant positive correlation with soil N content, Sphingobium with soil K content, and Acinetobacter with soil N and P contents, which indicated that Methylobacterium and Acinetobacter could be used as indicator species of soil pH, Brevundimonas as an indicator of soil N content, Acinetobacter as an indicator of soil P content, and Sphingobium as an indicator of soil K content. Therefore, improving Brevundimonas, Sphingobobacter, and Acinetobacter abundance can increase N, P, and K contents in the soil bank and sustainably provide nutrients for P. bournei seedling growth.

Site Description
The experiment site was at a field nursery of the Fujian Agriculture and Forestry University, China (119 • 23 E, 26 • 09 N). The annual average temperature is 19.6 • C, the minimum temperature is −2.5, and the maximum temperature is 42.3 • C. Effective accumulated temperature ≥10 • C is 5880 • C, with 326 days >0 • C. Annual precipitation is 1490 mm, and annual average humidity is 77% [48].

Materials
Well-grown, well-developed buds and one-year-old Phoebe bournei bareroot seedlings were selected for the experimental population in May 2018. The average height of the seedlings was 20.2 ± 0.9 cm, and the average diameter was 2.3 ± 1.0 mm. Seedlings had total N, P, and K contents of 1.219, 1.555, and 14.022 g·Kg −1 , respectively, at the start of the experiment [48].

Experimental Design
The test followed the "3414" fertilizer experimental design, which is set as three factors of N, P, and K [27,49]. Every factor has 0, 1, 2, 3 fertilization levels, respectively. Level 0 is the control (no fertilizer). Level 2 (medium level) is the common fertilizer rates, which were 0.532, 0.133, and 0.356 g·plant −1 for N, P 2 O 5 , and K 2 O, respectively [50,51]. Level 1 (low level) and Level 3 (high level) are 0.5 and 1.5 times Level 2, respectively. The specific rates of fertilizer were calculated in terms of N, P 2 O 5 , and K 2 O ( Table 6). The "3414" fertilizer experiment has 14 treatment combinations and can analyze the single-factor, two-factor, and three-factor interaction effects of NPK fertilization. In the single-factor effect treatments, when the "2" level was fixed by P and K fertilizers, the 0, 1, 2, 3 levels of N fertilizer were T2, T3, T6, and T11, respectively. When the "2" level was fixed by N and K fertilizers, the 0, 1, 2, 3 levels of P fertilizer were T4, T5, T6, and T7, respectively. When the "2" level was fixed by N and K fertilizers, the 0, 1, 2, 3 levels of K fertilizer were T8, T9, T6, and T10, respectively. In the two-factor interaction treatments, when K fertilizer was fixed at the "2" level, the N-P interactions had 8 treatments, which were T2, T3, T4, T5, T6, T7, T11, and T12. When the "2" level was fixed by P fertilizer, the N-K interaction had 8 treatments, which were T2, T3, T6, T8, T9, T10, T11, and T13. When the N fertilizer was fixed at the "2" level, the P-K interaction had 8 treatments, which were T4, T5, T6, T7, T8, T9, T10, and T14. The three-factor interaction of N × P × K was the T1-T14 treatments. The experiment had 14 treatments, with each treatment replicated three times, and 1260 seedlings, with 30 seedlings at each treatment replication. Random arrangement (complete randomized design) was used for the fertilizer replications to minimize any environmental effects [48]. Table 6. The "3414" fertilization experiment rates [48].

Experimental Management
This field experiment began in March 2018 and ended in December 2018. P fertilizer was used as the base fertilizer, and N fertilizer and K fertilizer were applied in different stages under the annual growth characteristics of P. bournei seedlings [52]. The applied fertilization regime was April-25% N and 20% K, June-35% N and 25% K, August-25% N and 35% K, and October-15% N and 20% K through a liquid application of fertilizer. Additionally, the concentration of fertilizer liquid was 0.05%. P. bournei seedlings need a light intensity equivalent of 75% natural light, and a 2.7 m high shed was equipped in the nursery. Seedlings were watered on a varying schedule determined by the empirical method of topsoil coloration, resulting in watering at~7 days during March to April, 15 days during May to July, not watered in August,~10 days during September to November, and~15 days in December, to maintain the desired soil water status of~75% field capacity [48].

Sample Collection and Processing
Nine seedlings per treatment were harvested after growth cessation in December 2018. The intact root systems were dug up and gently shaken to harvest the loose soil on the root surface, which were the soil samples. Soil from 9 plants were mixed in each treatment and divided into two parts, and one part was placed in a 5 mL sterile centrifuge tube and stored in a −80 • C refrigerator for soil bacterial sequencing use. The other part was air-dried and screened for chemical indicators.

Soil and Seedling Measurement Parameters
The sampled seedlings were washed with water and dried in the shade, fixed at 105°C for 15 min, and dried at 75 • C to achieve a stable weight. The samples' dry biomass was measured with 0.001 g accuracy by electronic scales (AL204, Mettler-Toledo, Melbourne, Australia) [48]. Seedlings were prepared for analytical sampling by grinding with a plant crusher and then processing the material through a 0.5 mm plastic sieve. Total N, P, and K of seedlings were dissolved with the H 2 SO 4 -HClO 4 method, soil total P and K with the NaOH solution-melting method, and soil total N with concentrated sulfuric acid and mixed catalyst (K 2 SO 4 : CuSO 4 = 10:1). N, P, and K contents were determined by the Kjeldahl method (ATN-300, Hongji, Shanghai, China) [53], the molybdenum-antimony colorimetric method (UV-2600A, Unicom, Shanghai, China) [54], and the atomic absorption spectrophotometer method (AA7002, Dongxi, Beijing, China) [55], respectively. Each indicator was repeated three times. Plant N (P, K) accumulations (mg·plant −1 ) were obtained by multiplying dry biomass (g·plant −1 ) by N (P, K) content (g·Kg −1 ). Plant biomass and the N, P, and K accumulation of each fertilizer combination are listed in Table 7.

Soil Bacterial Sequencing
Soil samples of 0.5 g were used to extract soil DNA using the Fast DNA ® Spin Kit for Soil (MP Biomedicals, Carlsbad, CA, USA). The extracted DNA products were tested by agarose gel electrophoresis, and the qualified DNA samples stored at −80 • C until further use. PCR conditions were 30 cycles at 98 • C pre-denaturation for 1 min, 98 • C denaturation for 10 s, 50 • C annealing for 30 s, 72 • C extending for 60 s, and finally extending 72 • C for another 5 min. PCR amplifications of the V3-V4 highly variable region of the bacterial 16S rRNA gene were performed with primers 338F (5 -ACTCCTACGGGAGCAG-3 ) and 806R (5 -GGACTACHVGGGTWTCTAAT-3 ). After amplification, the recovered products were quantified by Quanti Fluor TM fluorometer, denatured by NaOH into single chains, connected to sequencing joints, and sequenced on PE251 mode of Hiseq2500; a sequencing library was constructed according to Illumina instructions. The 16S rRNA sequence bases were analyzed using the Quantitative Insights into Microbial Ecology (QIIME) platform. Sequences with quality lower than Q20 and lengths less than 200 bp were removed and matched with samples according to the barcode sequence. Amplified primers and joint sequences were removed by Cutadapt and TrimMomatic software, and the Usearch algorithm was used to cluster the sequences according to 97% similarity and to remove chimeric operational taxonomic units (OTUs). Samples were divided into different OTUs; among each of them, that the sequence with the highest abundance was selected as the representative sequence.

Data Analysis
The annotated species in the Silva (Release132, http://www.arb-silva.de) database were selected (10 September 2020), and cluster comparisons were performed on all valid sequences to get OTUs with a >97% sequence similarity level. In this experiment, on average, each sample obtained 593 OTUs, belonging to 23 phyla, 51 classes, 107 orders, 155 families, and 208 genera. Mothur software was used to analyze the abundance and Alpha and Beta diversity of OTUs in each treatment soil. Alpha diversity was assessed by Chao1, Shannon, and Simpson indices. Chao1, the estimated number of OTUs, is defined as chao1 = S obs + F1(F1−1)/2(F2 + 1) where S obs are the number of OTUs observed, and F1 and F2 are the counts of singletons and doubletons, respectively [56,57].
Shannon's index is defined as where s is the number of OTUs, and p i is the proportion of the community represented by OUT i [56,57]. Simpson's index is defined as where p i is the proportion of the community represented by OUT i [56,58]. Additionally, Beta diversity was analyzed by the Anosim test, which was carried out on soil bacteria of all experimental treatments using R language [57,58]. Lefse1.0 software was used to further compare the composition differences of bacterial species among the different treatments and to find biomarker species [58]. The partial least squares path mode (PLS-PM) was used to analyze the relationships between fertilization, soil microorganisms, soil nutrition, and plant nutrition and growth by R language [59]. The results were drawn by Adobe Photoshop CC. Analysis of variance (ANOVA) was used to determine whether soil chemical characteristics and Alpha diversity indices of soil bacterial were significant in all treatments. Then Duncan's multiple comparisons (α = 0.05) were used to analyze whether there were significant differences between the treatments by SPSS 22.0 (Chicago, IL, USA). Excel 2016 and SPSS 22.0 (Chicago, IL, USA) were used to conduct the range analyses, which were the differences between the maximum and minimum values at different levels for the single-factor, two-factor, and three-factor effects of N, P, and K fertilizer. The correlation analysis between dominant phyla bacteria and dominant genera bacteria and key bacteria and soil chemical characteristics was performed by SPSS 22.0 (Chicago, IL, USA) for the three-factor interaction effect of N, P, and K fertilizer.

Conclusions
NPK fertilization reduced Phoebe bournei seedlings' soil pH, with the largest effect under N and the least under P for single-factor fertilization. By reducing the rate of N and increasing P, soil acidification can be controlled with minimal pH reduction. NPK fertilization significantly increased soil's N and K contents and decreased bacterial diversity. With proper NPK application, by reducing the N and K ratios and increasing P, not only can P. bournei seedlings be provided with balanced nutrients, but various bacteria can meet their growth and metabolism needs while keeping ecological stability and maintaining seedling growth and a soil sustainable utilization balance. Proteobacteria, Actinobacteria, Chloroflexi, and Acidobacteria were the dominant and core phylum species, while Cupriavidus, Methylobacterium, Brevundimonas, Sphingobium, Acinetobacter, and Sphingomonas were the dominant genus species in Phoebe bournei seedlings soil. NPK fertilization significantly affected soil bacterial community composition, increasing Proteobacteria, Methylobacterium, Sphingobium, and Acinetobacter abundance, decreasing Acidobacteria, and promoting Actinomycetes and Chloroflexi under appropriate NPK ratios, Brevundimonas under high P level, and Sphingomonas in the absence of K. NPK fertilization significantly increased P. bournei seedlings' soil N, P, K contents and indirectly promoted soil bacteria growth. However, it directly and significantly inhibited soil bacteria growth while promoting dominant bacterial species growth (which enrich nutrient elements) and N, P and K element storage in the soil bank (which provide the nutrients needed for subsequent seedling growth). Soil bacteria was the most important factor for P. bournei seedling growth and the sustainable utilization of soil. Methylobacterium, Brevundimonas, Sphingobium, and Acinetobacter were the key species in P. bournei seedlings' soil under fertilization, with increasing Brevundimonas, Sphingobium, and Acinetobacter abundance inducing increased N, P, and K contents in the soil bank, which sustainably provided the nutrients needed for seedling growth. AD3 can be used as the marker species under N fertilizer application, Pseudomonas under K fertilizer application, and Rhodanobacter under K fertilizer application; Methylobacterium and Acinetobacter for mentoring soil pH; Brevundimonas for soil N content; Sphingobium for soil K content; and Acinetobacter for soil P content change.

Data Availability Statement:
The datasets used and analysed during the current study could be available from the corresponding author on reasonable request.