To investigate differential effects of plant species and inter-specific plant–plant interactions among different secondary successional stage species, we selected Salix oritrepha as a representative middle-successional shrub specie, and late-successional species including Betula albosinensis, a broadleaf specie and Picea asperata, a conifer specie. The Salix oritrepha seedlings with a height of 40–45 cm were obtained from a local field shrub ecosystem while Betula albosinensis seedlings with a height of 45–50 cm were sourced from a nearby forest ecosystem. The Picea asperata seedlings with a height of 25–30 cm were acquired from a local nursery garden. In order to ensure uniformity in the rhizosphere soil of the all three species before the pot experiment and improve their survival rate, we first transplanted these seedlings to an adjacent Picea asperata plantation for acclimation over a period of 3 months before conducting the pot experiment.

The field pot experiment was conducted in an open field area near the Miyaoluo Natural Reserve of Lixian County, located on the eastern Tibetan Plateau, Sichuan Province, China (31°47′N, 102°42′E, 3210 m, a.s.l) with mean annual temperature 8.7°C, annual precipitation ranging from 600 to 1,100 mm, and annual evaporation ranging from 1,000 to1900mm. The soil is classified as mountain brown soil based on Chinese soil taxonomy. Soils were collected from nearby Picea asperata plantation and sieved with 4 mm mesh to eliminate roots and stones, and then were homogenized (pH 6.53, TOC 52.62 g/kg, TN 4.10 g/kg, available P 12.59 mg/kg, and available K 128.33 mg/kg). To investigate the differential effects of plant species and inter-specific plant interactions on soil bacterial communities, each two Salix oritrepha seedlings (SS), two Betula albosinensis seedlings (BB), and two Picea asperata seedlings (PP) as monocultures, and two combination of SB, SP, and BP seedlings as mixed cultures were planted 10 cm apart from each other into each plastic cylindrical pot (inner diameter 38 cm and depth 34 cm) in April 2018. Each combination of seedlings included 30 pots for applying inorganic and organic fertilizers, and control without fertilizer.

To examine the impact of fertilization on soil bacterial communities, we applied two types of fertilizer: organic and inorganic fertilizer. Each 10 randomly selected pots were used for each plant–plant interaction treatment to apply either organic fertilizer or inorganic fertilization, while 10 randomly selected pots were set up for controls without any fertilizer. Each treatment had 10 replicates. The organic fertilizer was derived from the local decomposed edible fungi residue, which was mixed with soil as a base fertilizer at the beginning of the experiment. A total of 60 pots with organic fertilizer were applied for 100 kg of organic fertilizer throughout the study period. This organic fertilizer contained water content 70%, N 3.78%, available K 3.91% and available P 0.71% of the dry mass. The inorganic fertilizer, consisting of KNO₃, (NH₄)₃PO₄, (NH₄)₂SO₄, Ca(NO₃)₂, was dissolved in water and applied to the plant roots through watering once a month from May 2018 to September 2019 with equivalent amounts provided in each pot throughout the experiment to match the N, P, K content of the organic fertilizer. We regularly measured the plant height from 2 months after the start of experiment until its end.

We sampled rhizosphere soil twice with a bore (8 cm diameter), one in August 2018 at the beginning of the experiment and the other in October 2019 at the end of the experiment, respectively. These samples were subsequently sieved through a 2 mm mesh, stored in an ice box at 4°C, and transported to laboratory for analysis. One subsample was stored in −80°C for soil bacterial communities analysis while another subsample was air dried for the soil properties analysis.

Soil pH was determined in suspending soil in water at a ratio of 1:2.5 v/v. The extraction method used for available phosphorus (AP) involved NaHCO₃ extracts, following the procedure developed by Olsen et al. (1954). For available potassium (AK), ammonium acetate was used as an extracting agent, and the measurement was conducted using a flame spectrophotometer. Detailed methodologies to determine concentrations of soil organic C (SOC), dissolve organic C (DOC), total nitrogen (TN), ammonium (NH₄⁺), nitrate (NO₃⁻), and nitrite (NO₂⁻)-N concentrations, as well as microbial biomass C (MBC) and N (MBN), were previously described in our study (Li et al., 2019).

Soil DNA was extracted from approximately 0.5 g of fresh soil sample using the OMEGA Soil DNA Kit (M5635-02) (Omega Bio-Tek, Norcross, GA, United States) following the manufacturer’s instructions. Subsequently, it was stored at −20°C until further analysis. The quantity and quality of the extracted DNAs were assessed using a NanoDrop NC2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, United States) and agarose gel electrophoresis, respectively.

PCR amplification was conducted to amplify the bacterial 16S rRNA gene V3–V4 region using forward primer 338F (5ˊ-ACTCCTACGGGAGGCAGCA-3ˊ) and reverse primer 806R (5ˊ-GGACTACHVGGGTWTCTAAT-3ˊ). The PCR reaction mixture included 5 μL of buffer (5×), 0.25 μL of Fast pfu DNA Polymerase (5 U/μl), 2 μL (2.5 mM) of dNTPs, 1 μL (10 uM) of each Forward and Reverse primer, 1 μL of DNA Template, and 14.75 μL of ddH2O. The PCR process began with an initial denaturation step at 98°C for 5 min, followed by 25 cycles consisting of denaturation for 30 s at 98°C, annealing for 30 s at 55°C, and extension for 45 s at 72°C. Finally, a 5 min extension step was performed at 72°C. The resulting PCR amplicons were purified with Vazyme VAHTSTM DNA Clean Beads (Vazyme, Nanjing, China) and quantified utilizing the Quant-iT PicoGreen dsDNA Assay Kit (Invitrogen, Carlsbad, CA, United States). After the individual quantification step, amplicons were pooled in equal amounts, and paired-end 2 × 250 bp sequencing was performed using the Illlumina NovaSeq platform with NovaSeq 6,000 SP Regent Kit (500 cycles) at Shanghai Personal Biotechnology Co., Ltd. (Shanghai, China).

The sequencing data analyses were performed with QIIME2 (Bolyen et al., 2018) with slight modification according to the official tutorials.¹ In order to minimize variations in sequencing depth across samples, an averaged and rounded rarefied ASV table was generated at an equal depth of 55,119 sequences per sample. Taxonomic classification of the ASVs was performed based on the Silva Release 132 Database.

Bioinformatics analyses were mainly performed using QIIME2 and R packages (v.4.4.0). The alpha diversity indices at ASV-level, such as Chao1 richness estimator, Observed species (Observed OTU) and Shannon diversity index, were calculated using the ASV table in QIIME2 and visualized as box plots. Principal Coordinate Analyses (PCoA) was performed visual the differences of bacterial community across samples based on the Weighted UniFrac distance metric. Permutational multivariate analysis of variance (PERMANOVA) method with 999 permutations was used to determine the effects of fertilizer and plant–plant interactions on the bacterial communities. Linear discriminant analysis effect size (LEfSe) was performed to detect differently abundant taxa across groups. Above microbial analyses were conducted by the genescloud tools, a free online platform for data analysis. ²

The growth rate of plant height, representing the annual increase rate in plant height, was calculated as the ratio of the change in plant height from the beginning to the end of the experiment relative to the initial plant height at the start of the experiment, and is expressed as a percentage annual growth rate. The SPSS software was utilized to conduct an analysis of variance (ANOVA) in order to assess the significant differences among the factors such as fertilization, plant–plant interaction, and sample time. These factors were examined for their impacts on soil properties, growth rate of plant height, bacterial taxa abundance (specifically the top 10 of phylum with highest average abundance), as well as alpha diversity indices of bacterial communities. The redundancy analysis (RDA) was used to investigate the effects of environmental factors on the bacterial communities and alpha diversity indices. Pearson’s correlation in SPSS and linear regression in Origin 21 also were used to determine the relationship between relative plant growth and the relative abundances of top 20 bacteria taxa from phylum to genus. Variation partitioning analysis (VPA) with the “varpart” function in vegan package were performed in R (v.4.4.0) to depict the shared and independent effects of soil properties and bacterial community on the growth rate of plant height.

The growth rate of plant height varied with plant species across various fertilization treatments (Figure 1A). Growth rate of shrub seedlings (S.oritrepha) exhibited the highest among the three species, while conifer seedlings (P.asperata) displayed the slowest growth rate in all the fertilization treatments. The interspecific interactions only affected the growth rates of plant height in the organic fertilizer treatment, but not in the control and inorganic fertilizer treatment (Figures 1B–D). When B.albosinensis (B) and P.asperata (P) grew with S.oritrepha (S) in mixed culture, their growth rates were higher than in singular monoculture under organic fertilizer treatment. Furthermore, in the singular monoculture SS and mixed culture (SP and SB), organic fertilizer promoted their growths of all these plants compared to control. In addition, irrespective plant species or interspecific interactions, plants in the inorganic fertilizer treatment grew faster than in the organic fertilizer treatment and control.

The concentrations of soil DOC, NH₄⁺, NO₂⁻, available K, MBC, MBN and MBC/MBN were significantly affected by plant–plant interactions, particularly under fertilization treatment (Table 1, Supplementary Table S1). The significant differences of these soil properties resulting from plant–plant interactions were affected by the fertilization. The impact of fertilization on soil properties was more than that of plant–plant interactions. Both inorganic fertilizer and organic fertilizer increased the concentrations of soil NH₄⁺, DOC, available P and available K, MBC and MBC/MBN but decreased the concentration of soil NO₂⁻. Moreover, organic fertilizer had a greater effect on those soil properties than inorganic fertilizer. Additionally, while inorganic fertilizer decreased the soil pH, and MBN, the organic fertilizer increased them (Supplementary Table S1).

The soil bacterial alpha diversity including Observed species (OTUs) and Shannon index, was significantly affected by both sample time and fertilization (p < 0.001) (Figure 2, Supplementary Table S2). However, neither plant–plant interaction nor the combined interactions of plant and fertilization had a significant influence on these diversity (p > 0.05). Regardless of the plant–plant interaction, the Observed species initially decreased in the organic fertilizer treatments compared to control but subsequently increased in the second year. The Shannon index and Observed species in inorganic treatment were lower than control but not significantly different between them. These diversities decreased with time in both control and inorganic fertilizer treatment while they increased with time in organic fertilizer treatments (Figure 2).

The PCoA showed the soil bacterial community was distinct across different fertilization treatment both in 2018 yr. and 2019 yr. (Supplementary Figure S1). The significant difference of beta diversity was detected by PERMNOVA test and the result showed that the soil bacterial community was significantly affected by both sample time (F = 9.50, p = 0.001) and fertilization (F = 22.18, p = 0.001), but was not affected by the plant–plant interactions (F = 1.07, p = 0.33) (Table 2).

The soil bacterial communities were predominantly composed of Proteobacteria, Acidobacteria, Actinobacteria, Chloroflexi, Verrucomicrobia, and Bacteroidetes at the phylum level, accounting for 90% of total bacterial abundance (Supplementary Figure S2). Different fertilizers had their specific influence on the bacterial compositions, and these influences varied with sample time (Supplementary Figure S3). Irrespective of the plant–plant interactions, inorganic fertilizer increased the relative abundance of Proteobacteria, Patescibacteria, Bacteroidetes and Gemmatimonadetes and decreased the relative abundance of Acidobacteria, Actinobacteria, Chloroflexi, Rokubacteria and Planctomycetes in the second year compared to control. Whereas, organic fertilizer enhanced the relative abundance of Chloroflexi, Planctomycetes and Bacteroidetes across both years (2018 and 2019), while it decreased the relative abundance of Actinobacteria and Verrucomicrobia in the second year compared to control (Figure 3, Supplementary Table S3).

The LEfSe analysis indicated that, there were no statistically significant variations in the abundance of any taxa across different plant species or interspecific interactions. However, the total relative abundances of bacteria at phylum level were significantly affected by plant–plant interactions under inorganic fertilizer treatment in 2019 yr. (Figure 3 and Supplementary Table S3). In singular monocultures pots, while the Proteobacteria in the SS soil were less abundant than in BB and PP soil, the Acidobacteria, Chloroflexi and Rokubacteria was on the contrary under inorganic fertilizer treatment in the second year. In the mixed cultures pots, while the relative abundance of Proteobacteria in PP soil were more than in SP soil, the relative abundance of Acidobacteria, Chloroflexi and Rokubacteria in PP soil were less than in SP soil, and there was no difference between SS and SP under inorganic fertilizer treatment in the second year.

The results of RDA analysis indicated that soil properties significantly affected the alpha diversity and compositions of the soil bacterial community (Figure 4). Among these soil properties, soil pH, TOC, TN, MBC, MBN were the most important factors influencing them. Besides, nutrients such as available K, available P, and NO₂⁻ also affected the bacterial communities (Supplementary Table S4).

The growth rate of conifer seedling (P.asperata) height exhibited a positive correlation with the relative abundances of bacteria genus Vibrionimonas (r = 0.49, p = 0.017), bacteria order Chitinophagales (r = 0.49, p = 0.016) and bacteria family Chitinophagaceae (r = 0.52, p = 0.010), but a negative correlation with bacteria family Micrococcaceae (r = −0.45, p = 0.032) (Figure 5A). And variation partitioning analysis indicated that growth rate of P.asperata was independently explained by bacterial communities (89.07%) rather than soil properties (25.41%) (Figure 5D). Similarly, the bacterial communities could independently explain the growth rate of shrub seedling (S.oritrepha) height greater compared to soil properties (70.85 and 23.23%, respectively) (Figure 5E). Whereas the growth rate of broadleaf seedling (B.albosinensis) height was affected by more bacterial taxa (Figure 5C), but the bacterial communities only independently explained its growth about 21.46% (Figure 5F).

Due to the rapid growth and extensive root systems of shrub seedling (S.oritrepha) when mixed with other plants, it was not feasible to separate the root. Therefore, our analysis of the plant–plant interaction was based on growth rate of plant height, as aboveground biomass is likely influenced by initial difference caused by the size of seedling. Although some plants modify the balance between aboveground and belowground allocation in response to competition, the majority of these adjustments are constrained, meaning that they are mainly influenced by competition on plant size (Mülle et al., 2000). Thus, growth rate of plant height will be a more suitable indictor for inflecting the influence of plant–plant interactions. We initially assumed that conifer species would be a superior competitor when grown in mixture with other species, according to the rule of secondary forest succession, which indicates that mid-successional plants are eventually replaced by late-successional plants. In this study, B.albosinensis and S.oritrepha grew faster than P.asperata. The interspecific interaction only affected the plant growth in the organic fertilizer treatment but not in the control and inorganic fertilizer treatment. Numerous studies have exhibited that plants tend to suffer competitive interactions from their neighbors (Guo et al., 2019; Cavalieri et al., 2020). However, in this study, the application of organic fertilizer resulted in the facilitation of P.asperata growth by B.albosinensis and S.oritrepha. S.oritreph promoted the growth of B.albosinensis, while P.asperata also promoted growth of S.oritreph. These findings exceed our initial hypothesis and further suggest that taking certain manual measures could make the mid-successional species promoting the growth of later successional species, thereby facilitating secondary forest succession. The positive influence of inorganic fertilizer on plant growth and the plant–plant interactions were widely studied (Yu et al., 2019; Cavalieri et al., 2020; Ma et al., 2022). In this study, Inorganic fertilizer only enhanced the growth rate of P.asperata and B.albosinensis in PB mixture and S.oritrepha in SB mixture and SS monocultures but not for P.asperata and B.albosinensis in singular monoculture compared to control. These results suggested that the effect of inorganic fertilizer on plant growth depends on the composition of the plants as different plants have different nutrients acquisition and different needs.

Changes in soil bacterial diversity and composition reflect the soil quality (Bending et al., 2004), as well as the maintaining integrity and stability of soil function (Dietrich et al., 2017). A meta-analysis investigating the effects of organic and mineral fertilizers on soil microbial diversity revealed no significant variation in the bacterial alpha diversity between NPK fertilizer treatment and control, but a 2.9% increase was observed in organic fertilizer treatment compared to control (Bebber and Richards, 2022). Similarly, Ma et al. (2022) found no significant difference in bacterial diversity in paddy soil between inorganic fertilizer and control; however, higher bacterial diversity was observed with organic fertilizer application compared to control. In consistent with these findings, our results indicate that while inorganic fertilizer (NPK) did not have significant effect on the soil bacterial alpha diversity, organic fertilizer increased it during the second year. The abundance and diversity of microbial communities are found to be closely associated with the amount of soil microbial biomass. Organic fertilizer contains not only NPK but also undigested plant matter (lignin, cellulose, hemicellulose), lipids, carbohydrates and essential nutrients (Cu, Fe, Zn, Mg), which may contribute to a higher microbial diversity according to species-energy hypothesis since organic fertilizer increase soil microbial biomass more effectively than NPK fertilizer do (Clarke and Gaston, 2006). In this study, organic fertilizer enhanced the soil organic carbon and microbial biomass but inorganic fertilizer only enhanced the soil microbial biomass and did not significantly affect the soil organic carbon. The observed increase in soil microbial biomass was more pronounced with organic fertilizer treatment compared to inorganic fertilizer treatment, which is consistent with findings from previous studies (Zhang et al., 2012; Francioli et al., 2016; Kumar et al., 2017). Furthermore, the bacterial diversity was mainly mediated by the soil properties such as soil pH, TOC, and N content, and pH was considered as the most important moderator to control the diversity (Zhou et al., 2020; Yu et al., 2021). Bebber and Richards (2022) found that the N and NPK fertilizer decreased soil pH, but manure fertilizer had a buffering effect on pH. This buffering effect, along with the presence of nutrient and organic content, is considered to be the primary advantage for microbial biodiversity. In this study, we observed a positive correlation between bacterial alpha-diversity (Observed species and Shannon index) and soil pH, TN, NH₄⁺-N, MBC, MBN, K; conversely, there was a negative correlation between bacterial alpha-diversity and soil NO₂⁻-N, and TC/TN (Supplementary Table S4). The application of inorganic fertilizer (NPK) increased soil MBC, available K and NH₄⁺-N concentration, but reduced soil MBN, pH, and finally slightly but not significantly decreased the soil bacterial diversity. On the other hand, the organic fertilizer resulted in an increase in soil properties such as pH, TOC, TN, NH₄⁺-N, MBC, MBN and available K, leading to an increase in soil bacterial diversity during the second year. These findings suggest that the impacts of different fertilizers on soil bacterial diversity depends on the integrated effects of those combined soil factors affected by them.

The addition of Nitrogen (N), phosphorus (P) and potassium (K) fertilizer directly influences bacterial composition by increasing the availability of N, P and K (Guo et al., 2019; Cheng et al., 2022). In this study, both inorganic fertilizer and organic fertilizer enhanced the concentrations of soil NH₄⁺, available P and available K. The PCoA and PERMANOVA analyses showed significant effects on the bacterial composition resulting from the addition of inorganic and organic fertilizer (Table 2). Actinobacteria, Proteobacteria, Bacteroidetes, Acidobacteria, Chloroflexi and Verrucomicrobia were found to be the most abundant bacterial groups accounting for approximately 90% of bacterial abundance. These six bacterial groups can be classified into copitroph r-strategists and oligotroph K-strategists according to their life history strategies. Actinobacteria, Proteobacteria, Bacteroidetes are generally regarded as r-strategists in the copiotrophic group, and Acidobacteria, Chloroflexi and Verrucomicrobia are regarded as K-strategists in the oligotrophic group (Xu et al., 2020; Ma et al., 2022). The addition of organic fertilizer enhanced the relative abundances of r-strategists in the copiotrophic group (Actinobacteria, Bacteroidetes and Proteobacteria), whereas inorganic fertilizer increased those of K-strategists in the oligotrophic group (Acidobacteria, Chloroflexi and Verrucomicrobia) (Xun et al., 2016). Additionally, Ma et al. (2022) found that the application of organic fertilizer increased the abundance of r-strategists in the copiotrophic group (Actinobacteria and Proteobacteria) but decreased those of K-strategists in the oligotrophic group (Acidobacteria, and Verrucomicrobia). Fierer et al. (2012) observed that an increase in soil N availability led to a higher abundance of copiotrophic bacteria taxa such as Proteobacteria and Bacteroidetes, while reducing the abundance of oligotrophic group Acidobacteria. Consistently, our study observed that compared to control, both organic and inorganic fertilizer resulted in an increased relative abundance of copiotrophic group Bacteroidetes and Proteobacteria (Alphaproteobacteria and Dongiale), whereas the oligotrophic group Acidobacteria and Verrucomicrobia abundances were decreased by organic fertilizer. In this study, some bacteria belonging to Acidobacteria like Holophagae and Thermoanaerobaculia were reduced by inorganic fertilizer (Xun et al., 2016), whereas bacterial phylum Acidobacteria were enhanced (Fierer et al., 2012). Additionally, Chloroflexi was regarded as oligotrophic group and its relative abundance was lower in the fertilized soils than control soil (Xu et al., 2020) but increased by inorganic fertilizer (Xun et al., 2016). However, in this study, organic fertilizer increased the relative abundance of Chloroflexi, whereas inorganic fertilizer decreased it compared to control during the second year. This observation can be explained by the positive correlation between the relative abundance of Chloroflexi and soil pH (Deng et al., 2018). Actinobacteria, known as a copiotrophic group, played an important role in the degradation of soil organic matters and exhibited higher relative abundance both in the organic and inorganic fertilizer soils (Francioli et al., 2016; Xu et al., 2020; Ma et al., 2022). We observed a decrease in the relative abundances of bacteria phylum Actinobacteria with the addition of both inorganic and organic fertilizer, but also observed increase in Frankiales and Micromonosporales belonging to Actinobacteria in the inorganic fertilizer treatment. Moreover, the inorganic fertilizer increased the total relative abundances of the r-strategists in copiotrophic group (Actinobacteria, Bacteroidetes and Proteobacteria) by approximately 3.53% and decreased the K-strategists in oligotrophic group (Acidobacteria, Chloroflexi and Verrucomicrobia) by 5.45%. Conversely, although organic fertilizer adjusted the composition of bacterial community only within its respective group, it had no significant effect on the total relative abundances of either copiotrophic or oligotrophic groups compared to control in this study. Notably, discussing bacterial communities’ responses to different fertilizers at the genus or order level may be more precise and effective in drawing conclusions than at the phylum level. For instance, certain findings in this study were inconsistent with previous studies conducted in agriculture soil or grassland, particularly at the bacterial phylum level. This underscores the significant role of vegetation in releasing root exudates, while also potentially being influenced by factors operating at the phylum level rather than genus or order level.

There are many studies showed that different plants have specific effects on rhizosphere soil physicochemical properties through root exudates and nutrient mineralization (Guo et al., 2019), and then the plant species significantly have great impact on the composition and diversity of soil bacterial communities both directly and indirectly (De Deyn et al., 2011; Chen et al., 2016; Wang et al., 2022; Zhang et al., 2024). Also numerous research showed that interspecific plant–plant interaction influenced soil microbial communities due to competition for nutrients or different plants release different exudates which will affect the soil bacterial community (Haichar et al., 2008; Knelman et al., 2018; Cavalieri et al., 2020). In this study, we observed significant effects of plant species across various stages of secondary succession in forest restoration as well as their inter-specific interactions on soil properties including the concentrations of DOC, NH₄⁺, AP, AK, MBC, MBN, particularly under fertilization treatments. These factors exhibited strong correlations with the diversity and composition of soil bacterial communities (Figure 4), suggesting that the plant species would have a significant impact on bacterial communities. However, neither the plant species in monocultures nor the plant–plant interactions in mixture had significant influence on the diversity and composition of soil bacterial community in the control. This result suggests that the rhizosphere soil bacterial communities diversity remains stable across secondary forest succession and factors other than plant diversity may be driving variations in soil communities. Pivato et al. (2017) found that the total bacterial abundance in the rhizosphere of inter-specific plant–plant mixtures was significantly greater than that of either individual plant species in a monoculture at low N levels, but this difference disappeared when the N levels were high in agroecosystems. In this study, the responses of bacterial community abundances in the rhizosphere of late-successional species P.asperata (P) to inorganic fertilizer was found to be the most sensitive, while the mid-successional species shrub S.oritrepha (S) exhibited the least sensitivity. This resulted in significant difference in the relative abundances of Proteobacteria, Acidobacteria, Chloroflexi and Rokubacteria between in the rhizosphere of SS and PP or BB monocultures, as well as between in the rhizosphere of PP and SP mixture under inorganic fertilizer treatment during the second year. This result further suggests that different successional species had varying response to fertilizer application.

When two plant species coexist, how does interspecific interaction select their bacterial community structure? Hortal et al. (2017) reported that when two plant species were grown together in mixed cultures, the composition of rhizoshpere bacterial community resembled that of the superior competitor in monocultures. We initially assumed that conifer species is superior competitor and their rhizoshpere soil bacterial community would resemble that of their monocultures soil when grown with other species in mixture. However, contrary to this expectation, our study indicates that the bacterial communities in SP soil are more similar to SS soil rather than PP soil, and similarly for SB soil compared to BB soil under inorganic fertilizer treatment (Figure 3). Interestingly, a similar pattern was observed between P and B interaction. This discrepancy may be attributed to differences in plant growth rate and root development. Our former results of plant growth indicated that the seedling growth and root development of shrubs are comparatively faster than broadleaf seedlings, while confer seedlings exhibit the slowest growth. Consequently, shrub roots produce a greater quantity of root exudates compared to broadleaf or conifer roots that significantly influence the composition and diversity of soil bacterial communities. Furthermore, while the organic fertilizer had similar impacts on the soil chemical properties and microbial biomass as the inorganic fertilizer, it also promoted plant growth and affected the interspecific interaction. However, it did not alter the effects of plant species or plant–plant interaction on the rhizosphere bacterial communities. These results imply that organic fertilizer not only improves soil nutrient availability but also maintains soil health and promotes the succession of plant communities compared to inorganic fertilizer due to its potential influence on nutrients cycling and organic matter decomposition, and plants disease susceptibility mediated by changes in bacterial community composition. Therefore, it is important to consider the potential effects of different fertilizer management practices on soil bacterial communities when managing soil resources for forest recovery.

It is widely acknowledged that soil properties and bacterial communities exert significant influences on plant growth (Kardol et al., 2006), suggesting that plant–plant interactions and fertilizers may affect plant growth by regulating soil properties and bacterial communities. In this study, although both the plant–plant interactions and fertilizers influenced the soil properties and the soil bacterial communities also were mediated by soil properties (Figure 4), the variance partitioning analysis (VPA) results revealed that the growth rates of coniferous seedling P.asperata and shrub seedling S.oritrpha were predominantly explained by bacterial communities (89.07 and 70.85%, respectively), rather than soil properties alone (25.41 and 23.23%, respectively) (Figures 5D,F). Moreover, the growth rate of broadleaf seedling B.albosinensis was found to be more strongly affected by soil properties compared to bacterial communities (Figure 5C). Nevertheless, certain bacterial taxa such as Patescibacteria, WPS-2, Armatimonadetes, Alphaproteobacteria, etc., exhibited significantly positive correlations with plant growth rate, whereas others including RB41, Bradyrhizobium, Chloroflexi, Rokubacteria, etc., showed negative correlations (Supplementary Table S5). Ultimately, the effects of different bacterial taxa on plant growth rate counterbalanced each other. These findings underscore the substantial impact of soil bacterial communities on plant growth. In line with our results, numerous studies investigating the effect of plant–soil feedback have consistently demonstrated a strong association between plant performances and soil biota rather than abiotic conditions (Kardol et al., 2006). Plant growth-promoting bacteria (PGPB) play a crucial role in enhancing plant growth and crop productivity in agroecosystems. Extensive research has been conducted on various plant growth promoting communities, such as Actinomycetota, Acidobacteria, and Chloroflexi, which have been proven to drive rhizosphere ecology and nutrient cycling. These communities promote plant growth and fitness through mechanisms like phosphate solubilization, secondary metabolite production, and antimicrobial synthesis (Kaushik et al., 2021). In this study, plants across different successional stages of secondary forest had specific bacterial communities for promoting their growth (Figures 5A–C). This finding offer valuable perspectives on the potential regulatory mechanisms of microbial ecology during the progression of plant community succession, underscoring the significance of microbial communities in driving forest succession. This understanding will aid in the development of more effective strategies for optimizing forest restoration practices. For example, promoting beneficial rhizosphere bacteria can improve nutrient cycling and facilitate the establishment of diverse native plant communities in degraded forests, ultimately contributing to overall ecosystem resilience in managed forests. Therefore, understanding the responses of bacterial communities to plant–plant interaction and different types of fertilizers can help us develop sustainable strategies for managing forests and maximizing their ecological benefits. Overall, this study highlight the critical role of bacteria in shaping forest ecosystems during succession. Further investigations are needed to unravel additional mechanisms employed by these bacteria as well as explore their potential application.