Soil microbial communities, soil nutrition, and seedling growth of a Chinese fir (Cunninghamia lanceolata (Lamb.) Hook.) plantation in response to three weed control methods

Competitive vegetation in forest stands influence seedling growth by changing soil nutrient availability. However, studies on the effects of different weed control methods on seedling growth of Chinese fir (Cunninghamia lanceolata (Lamb.) Hook.) are rare. We applied three weed control methods, comprising artificial sickle weeding (ASW), woody disc weeding (WDW), and nonwoven cloth weeding (nWCW), to explore their effect on growth of Chinese fir seedlings in a plantation in Jiangxi Province, China. The weed control methods affected the shoot height and root-collar diameter of the seedlings. The contents of sugar, glucose, triglycerides, total cholesterol, and free fatty acids in newly developed leaves were increased after ASW and nWCW treatment, and were consistent with the expression of genes associated with glucokinase, sucrose phosphate synthase, and sucrose synthase. Weeding method influenced soil properties, including pH, moisture, total nitrogen (TN), ammonium-N, nitrate-N, total phosphorus, available phosphorus, and dissolved organic carbon contents. Moisture content was the main factor that influenced the soil bacterial community and leaf nutrition. High-throughput sequencing of the bacterial 16 S rRNA gene revealed that the weeding methods affected bacterial community structure. Specifically, compared with ASW and nWCW, WDW contributed to lower soil bacterial diversity, simpler bacterial interaction, and increase in pathogenic bacteria potential. The weeding methods differ in influence on soil bacterial community structure, soil properties, and plant growth, which are potentially useful to improve the growth of Chinese fir seedlings. ASW and nWCW strategies were recommended to be applied in the practice of weed control on seedling growth of Chinese fir.

However, it is important to strike a balance between protection of soil and water properties and reduction of weed competition with trees, because these interactions restrict the growth of tree seedlings and may cause their death (Zimdahl 2007;Cierjacks et al. 2016;Peachey et al. 2017;dos Santos et al. 2019). Weeds interfere with tree growth by competing for nutrients, water, space, light, and oxygen, and may harbor pests and pathogens. Hence, effective weed management in forest nurseries are extremely important for initial conifer seedling survival and in the first 3 years after transplanting (during the establishment period) in the field (Knowe and Stein 1995;Schneider et al. 1998;Harper et al. 2005).
Application of herbicides is a common and effective strategy for weed control in plantation forestry (Newton 2006), but potentially harmful consequences are profound and long-lasting. During the establishment phase, particularly young trees, are sensitive to chemical herbicides. Many herbicides may cause severe phytotoxic injury to trees, including stunted growth, burning and abscission of needles, chlorosis, or death of the tree. Repeated applications of the same herbicide may result in development of herbicide resistance among weed species (Hill 2018). In addition to their effects on plants, herbicides may interfere with enzymatic activities, biochemical processes, microorganism diversity, and microbial functions in the soil (Sannino and Gianfreda 2001;Kinney et al. 2005;Mahia et al. 2008;Liu et al. 2020b). Many changes in the diversity and composition of the beneficial microbial community can be unfavorable to plant growth and development by reducing nutrient availability or by increasing disease incidence. Therefore, effective nonchemical weed-control strategies are required urgently.
Nonchemical methods of vegetation control, such as harvest weeds control, seed predation, development of herbicide-resistant crops, and bioherbicides, in some studies are reported to be as effective as repeated herbicide treatments (Bajwa et al. 2015). Some mechanical or manual control methods, such as mowing, may assist in preventing the development of herbicide-tolerant weed species, reduce the number of seeds produced by weeds, and significantly reduce weed competition with the trees when implemented at the optimal time point prior to weed seed development. Certain artificial mowing methods promote the growth of trees, such as loblolly pine (Pinus taeda L.), slash pine (Pinus elliottii Engelm. var. elliottii), and improve forest biomass, productivity, and tree survival (Vogel et al. 2011;Hamberg et al. 2011). However, large mowing equipment increases the risk of damage to the trees, and exclusive dependence on a hand weeding method to control weeds is labor intensive, expensive, and time consuming. Large numbers of unconventional weed management strategies, including mulching may be viable, feasible, and efficient, because these materials act as physical barriers that limit or exclude weeds and regulate soil temperature and moisture, which aids in seedling establishment and survival (Saha et al. 2020). Some mulches prevent water percolation into the soil but restrict the amount of evaporation away from the soil, include asphalt shingles or plastic sheets of various kinds (Appleton and Hill 1997). Meanwhile, some organic mulches, such as grass clippings, nut hulls, wood chips, compost, bark, sawdust, and other organic materials were also used in weeding (Appleton and Hill 1997;Chalker-Scott 2007). However, researches on using woody disc and nonwoven cloth in weeding are uncommon, especially on Chinese fir seedling.
Chinese fir (Cunninghamia lanceolata), among the most common coniferous species in subtropical regions of China, is a fast-growing native tree that has been planted for timber production for more than 1000 years (Ma et al. 2007). The species grows in diverse habitats ranging from hills to high mountains, from fertile soils to infertile sites, and from small disconnected fragments of habitat to large continuous areas. Recently, with the increasing national demand for wood products because of population expansion and economic development, the area of Chinese fir plantations has expanded rapidly. In some regions, the rotation period of plantations has been shortened to 17 years (Bi et al. 2007). However, continuous silvicultural activities are responsible for degradation of soil fertility, decrease in standing biomass, and reduce primary productivity of Chinese fir plantations (Tian et al. 2011). Given that weeds can affect tree growth, the development of measures to control weeds and maintain soil fertility to promote the growth of Chinese fir plantations are important. In this study, we applied three methods including artificial sickle weeding (ASW), woody disc weeding (WDW), and nonwoven cloth weeding (nWCW) to suppress the growth of competitive weeds around Chinese fir seedlings, and compared the effects of 1 3 Vol.: (0123456789) the weeding methods on the soil microbial community, soil properties, and growth of Chinese fir seedlings. The objective was to evaluate the feasibility of nonchemical weed control strategies in Chinese fir production and the surrounding environment, where microbial communities and soil nutrition are directly influenced by mowing and mulching. Such strategies are important for implementation of sustainable forest nursery and forestry practices.

Study area and experiment design
The experiment was established in Fengshushan Forestry Farm (29°17′N, 117°22′E), Jiangxi Province, China. The study site was located in a Chinese fir (Cunninghamia lanceolata (Lamb.) Hook.) plantation at an elevation of 43.0 m. The area experiences a typical long warm summer and short cool winter. The annual mean temperature, and mean temperatures in January and July are 17.1, 4.6, and 28.7 °C, respectively. The average length of the growing season is 264 days and mean annual precipitation is 1805 mm. The average depth of the krasnozem soil is 50 cm.
A randomized complete-block experimental design was used in this study. The planting area was firstly divided into three blocks, and each block consisted of three plots which containing at least two Chinese fir seedlings. Seven Chinese fir seedlings per weed control method, with 2.0 m × 2.0 m planting density, were randomly selected for growth measurement on January 2019. Shoot height and root-collar diameter of the Chinese fir seedlings with 20 cm and 0.5 cm were selected as the targets of cultivation. At the beginning of the experiments, all plants within at least one meter of the seedlings were artificially uprooted in the range of ASW, WDW and nWCW treatments. ASW involved cutting down all herbs, shrubs, lianes, and nontarget tree species in the experimental plots with sickles twice a year (in mid-April and mid-September of 2019). The radius of the mowing area around seedlings in ASW method were one meter. To maintain the soil fertility and nourish the seedlings of Chinese fir, cut weeds were leaved on site to decay into organic matters. WDW comprised generating a mixture mainly of sawdust, bamboo shavings, rice straw, and tree bark through high mechanical pressure, which was painted with a waterproof coating and used to cover the plots. A gap (5 cm diameter) was left in the center of each disc (40 cm diameter and 5 mm thickness). This woody mulch was applied around the base of each seedling after planting with the aim of inhibiting the growth of competitive weeds by blocking the available light. For the nWCW treatment, nonwoven black cloth prepared from recycled cotton (2 m width and 0.5 mm thickness) was used. The mesh size of nonwoven cloth was 2.0 μm, which was porous and allowed water easily to permeate through it, but penetration of sunlight was absolutely impeded and the soil temperature was effectively maintained. The nonwoven cloth was spread on the soil surface between rows of seedlings in the experimental plots to obstruct access to light by competitive weeds, and they were fixed using the pile anchoring method at the edges with the interval of one meter (Fig. 1).

Sample collection
One year after the three weeding treatments were performed, shoot height and root-collar diameter were recorded on 15 January 2020. As all these Chinese fir seedlings in this experiment were growing normally, seven newly developed leaves from the lateral branches of these trees in per treatment were sampled for physiological and gene expression analyses. An S-shaped sampling method was used to collect soil cores with a soil borer at 10 cm depths for each treatment. The soil and leaf samples were sealed in sterile plastic bags and transported on dry ice to the laboratory within 24 h, where the samples were quickly frozen using liquid nitrogen and stored at − 80 °C until subsequent analysis. The soil samples were mixed thoroughly, and then passed through a 2 mm sieve and divided into two portions. One portion was used to measure soil physicochemical properties.

Soil physicochemical analysis
Soil pH was measured using a fresh soil:water ratio of 1:5 with a pH meter. Soil KCl-extractable ammonium-N (NH 4 + -N) and nitrate-N (NO 3 − -N) were extracted using a mixture of fresh soil and 2 M KCl (1:5, w/v) with shaking at 200 rpm for 1 h, and quantified using a segmented flow analyzer (SANCC, Skalar, Breda, The Netherlands). Soil moisture content was 1 3 Vol:. (1234567890) determined by oven-drying the samples at 105 °C (Chinese Industry Standard LY/T 1213-1999). Dissolved organic carbon (DOC) was extracted with 0.5 M K 2 SO 4 and the content determined using a total organic carbon analyzer (Multi N/C 3000; Analytik, Jena, Germany; Watkins et al. 1987). Total nitrogen (TN) content and total phosphorus (TP) concentration were analyzed using semi-macro Kjeldahl digestion and the molybdenum-antimony colorimetric method followed by H 2 SO 4 digestion, respectively (Allen et al. 1974). Available phosphorus (AP) was extracted using sodium bicarbonate and analyzed using the molybdenum-blue colorimetric method after extraction with diluted H 2 SO 4 -HCl (Chinese Industry Standard LY/T 1232-2015).
Soil DNA extraction and high-throughput 16 S rRNA gene sequencing Total bacterial DNA was extracted from the twentyone soil samples using the PowerSoil DNA Isolation Kit (MO BIO Laboratories). The DNA quality and quantity were assessed with an Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA) and Qubit 2.0 Fluorometer (Life Technologies, Carlsbad, CA, USA). The purified DNA was stored at − 80 °C until further processing. The V3-V4 region of the bacterial 16 S rRNA gene was amplified using the forward primer 338 F (5′-ACT CCT ACG GGA GGC AGC A-3′) and reverse primer 806R (5′-GGA CTA CHVGGG TWT CTAAT-3′) combined with adapter sequences and barcode sequences . Each PCR amplification was performed in a 50 µL volume containing 10 µL buffer, 0.2 µL of Q5 High-Fidelity DNA Polymerase, 10 µL High GC Enhancer, 1 µL dNTPs, 10 µM of each primer, and 60 ng template DNA. The following PCR program was used: initial denaturation at 95 °C for 5 min, followed by 15 cycles at 95 °C for 1 min, 50 °C for 1 min, and 72 °C for 1 min, with a final extension at 72 °C for 7 min. The PCR products from the first-step PCR were purified using VAHTS™ DNA Clean Beads. A second-round PCR was then performed in a 40 µL reaction volume that contained 20 µL of 2× Phusion HF MM, 8 µL ddH2O, 10 µM of each primer, and 10 µL first-step PCR products. The second-step PCR program comprised an initial denaturation at 98 °C for 30 s, followed by 10 cycles at 98 °C for 10 s, 65 °C for 30 s, and 72 °C for 30 s, with a final extension at 72 °C for 5 min. Finally, all PCR products were quantified using Quant-iT™ dsDNA HS Reagent and pooled. High-throughput sequencing analysis of the bacterial rRNA genes in the purified, pooled sample was performed using an Illumina HiSeq 2500 platform (with 2 × 250 paired ends) by Biomarker Technologies Corporation (Beijing, China).

Chemical analyses
To determine the contents of sugars, glucose (GLU), triglycerides (TG), total cholesterol (TCH), and free fatty acids (FFAs), the sampled needles were defrosted and homogenized with nine times volume of pre-cooled phosphate-buffered solution, and then centrifuged at 3000 ×g at 4 °C for 10 min. The concentrations of sugars, GLU, TG, TCH, and FFA in the supernatant were determined using kits provided by Nanjing Jiancheng Institute of Biotechnology (Nanjing, China) in accordance with the manufacturer's instructions.

RNA extraction and real-time quantitative PCR
Expression of genes associated with energy metabolism in twenty-one newly developed needles were analyzed. Total RNA in the sampled needles was isolated with TRIzol Reagent (TaKaRa, Dalian, China) in accordance with the manufacturer's instructions. The PrimeScript II 1st Strand cDNA Synthesis Kit (TaKaRa) was used to perform reverse transcription. The RNA quantity and quality were assessed by determining the ratio of absorbance at 260 and 280 nm with a NanoDrop spectrophotometer (Thermo Fisher Scientific, Wilmington, DE, USA) and an Agilent 2100 Bioanalyzer (Agilent Technologies). First-strand cDNA was synthesized from l.0 µg total RNA using M-MLV reverse transcriptase (Invitrogen, Carlsbad, CA, USA). Quantitative RT-PCR (RT-qPCR) was performed using Premix Ex Taq™ with SYBR Green (TaKaRa) and an ABI 7500 Fast Real-Time PCR System (Applied Biosystems, Carlsbad, CA, USA). The qPCR protocol was as follows: the thermocycle protocol for 30 s at 95 °C, followed by 40 cycles of denaturation at 95 °C for 5 s, annealing/ extension at 60 °C for 35 s, and a final melting-curve analysis to monitor the purity of the PCR product. Primer sequences were designed and selected using Primer 5.0. The 2 −ΔΔCt method was used to calculate the abundance of mRNA. Relative expression levels of Glucokinase (GK), sucrose phosphate synthase (SPS), sucrose synthase (SS), and hexokinase (HXK) gene were standardized relative to the housekeeping gene GAPDH (Chen et al. 2020).

Data analysis
After the barcode and primer of the raw sequences were trimmed, sequence reads were spliced and filtered using FLASH v1.2.11 and Trimmomatic v0.33, respectively. Clean reads were clustered into operational taxonomic units (OTUs) at the threshold of 97% similarity after the chimeras of filtered reads were eliminated using UCHIME v8.1. Alpha diversity and beta diversity between the samples were analyzed with QIIME v1.9.1 (Caporaso et al. 2010). A species richness estimator (Chao 1), abundancebased coverage estimator (ACE) and diversity indices (Shannon and Simpson) were determined using Mothur (Schloss et al. 2009). The dissimilarity of microbial communities among samples was evaluated using principal coordinates analysis (PCoA), and a permutational multivariate analysis of variance (PER-MANOVA) was conducted, using a Bray-Curtis distance matrix calculated using the R package "vegan" and the permuted P-value was obtained with 10,000 permutations. The unweighted pair-group method with arithmetic means (UPGMA) was used to evaluate the similarity of the bacterial communities using beta-diversity data and QIIME v1.9.1.
Representative OTUs were taxonomically assigned using the RDP classifier v2.2 against the SILVA database (v115) (Wang et al. 2007). A linear discriminant analysis (LDA) effect size (LEfSe) analysis was performed online with the default parameters (minimum LDA score (log10) ≥ 4.0 and 30 bootstrap iterations to identify the highly abundant bacteria that showed dimensional patterns among the three weeding methods (Segata et al. 2011). A nonparametric factorial Kruskal-Wallis rank-sum test accompanied by the unpaired Wilcoxon rank-sum test were used to identify the taxa with the greatest differences in relative abundance. The Kyoto Encyclopedia of Genes and Genomes (KEGG) metabolic pathways enriched among the soil bacteria were analyzed using Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt) (http:// hutte nhower. sph. harva rd. edu/ galaxy/ tool_ runner? toll_ id= PICRU St_ norma lize). The relative frequencies of the function of overall KEGG Orthologs (KOs) associated with carbon metabolism, nitrogen metabolism, and the phosphorus cycle were analyzed using Dunn's paired test after the Kruskal-Wallis test. BugBase (https:// bugba se. cs. umn. edu/) was used to predict organism-level microbiome phenotypes; the significance of differences in relative abundance of phenotypes between two groups was assessed by performing pairwise Mann-Whitney-Wilcoxon tests after the Kruskal-Wallis test. Co-occurrence networks of the soil microbial communities in the different weeding groups were generated based on significant correlations (Spearman's ρ > 0.8 and P < 0.05) and were visualized using Gephi software (https:// gephi. org/). The Mantel test was performed to assess the correlation between soil properties and the top 10 bacteria at phylum level. Using CANOCO for Windows (v5.02), a canonical correspondence analysis (CCA) was performed to evaluate the soil property that best explained the variation in bacterial community structure.
For the soil properties data, the contents of organic nutrients in newly developed leaves, and the qPCR results, differences between groups were analyzed using Turkey's test after one-way analysis of variance (ANOVA) using SPSS 20.0 (IBM Corporation, Armonk, NY, USA). If data were non-normally distributed or the variances were unequal, the Kruskal-Wallis test and Dunn's test were applied for multiple comparisons. Results are expressed as the mean ± standard error of the mean (SEM).

Effects of the weed control methods on soil properties
The three weeding methods differentially affected the soil properties of the Chinese fir plantation (Table 1). Compared with the ASW, soil pH were remarkably decreased in response to the WDW and nWCW treatments (p < 0.05). The soil moisture content was significantly higher under nWCW than that under WDW and was lowest in response to ASW (p < 0.05). The total N, NH 4 + -N, and NO 3 − -N contents were increased under WDW and nWCW compared with those under ASW (p < 0.05). The AP contents in the WDW and nWCW treatments were significantly lower than that under ASW (p < 0.05). The TP content was not significantly affected by weeding method (p > 0.05). The WDW and nWCW treatments significantly increased the DOC content compared with the ASW treatment (p < 0.05).
Seedling growth under the weed control methods The shoot height and root-collar diameter of Chinese fir seedlings were measured to analyze the influence of weeding method on seedling growth and biomass. The shoot height in the nWCW treatment was significantly higher than that of the ASW and WDW groups (Fig. 2a). Similarly, the root-collar diameter was significantly higher under nWCW and showed no significant difference between the ASW and WDW treatments (Fig. 2b). These findings showed that growth of Chinese fir seedlings was superior in response to nWCW than that under ASW and WDW.
Composition of the soil microorganism community under the weed control methods

Alpha diversity
The soil bacterial communities of 21 samples from the three treatments were analyzed based on 16 S rRNA gene sequence data. In total, 1285 OTUs were delineated at the 97% similarity level of which 1256, 1229, and 1269 OTUs were detected in the ASW, WDW, and nWCW treatments, respectively. Among these OTUs, 1193 were detected in all groups, and only one and two specific OTUs occurred under ASW Table 1 Effect of three weed control methods on soil properties of a Chinese fir plantation Data are the mean ± standard error of the mean (SEM) (n = 7). Different lower-case letters within a column indicate a significant difference among weed control methods (p < 0.05). SM: Soil moisture; TN: total nitrogen; AP: available phosphorus; TP: total phosphorus; DOC: dissolved organic matter and nWCW, respectively (Fig. 3a). Among the three weed control methods, no significant difference in the Simpson index was observed (p > 0.05; Fig. 3c). The Shannon and Chao1 indices under WDW were significantly higher than those of the ASW and nWCW treatments (p < 0.05; Fig. 3b, d). These results showed that the bacterial communities in the ASW and nWCW treatments were more diverse than under WDW. The correlation coefficients between the bacterial alpha-diversity indices and soil parameters are shown in Table 2. The values for the abundancebased coverage estimator (ACE), Shannon index, and Simpson index were positively correlated with soil pH, moisture content, and AP content (p < 0.05), and negatively correlated with NH 4 + -N, TN, and DOC contents (p < 0.05).

Beta diversity
The first and second principal coordinates explained 55.90% and 15.40% of the total variation associated Fig. 2 Effect of weed control methods on growth of Chinese fir seedlings. a shoot height; b rootcollar diameter. Different lower-case letters above bars indicate a significant difference between treatments (p < 0.05). Values are the mean ± standard error of the mean. ASW: Artificial sickle weeding; WDW: woody disc weeding; nWCW: nonwoven cloth weeding Fig. 3 Structure of soil bacterial communities in each weed control treatment. a Venn diagram of soil bacterial operational taxonomic units; b Shannon index for each weed control method group; c Simpson index for each weed control method group; d Chao1 index for each weed control method group. ASW: Artificial sickle weeding; WDW: woody disc weeding; nWCW: nonwoven cloth weeding with the differences between weeding methods (Fig. 4a). PERMANOVA analysis based on Bray-Curtis distances revealed that the diversity of soil bacterial communities differed significantly among the weed control methods (R 2 = 0.275, p = 0.001). Significant differences were observed between ASW and WDW (R 2 = 0.207, p = 0.001), ASW and nWCW (R 2 = 0.176, p = 0.006), and between WDW and nWCW (R 2 = 0.265, p = 0.001), which indicated that unique bacterial communities were formed in response to the three weeding methods (Fig. 4a). In addition, hierarchical clustering using the UPGMA algorithm revealed a distinct disparity among the soil bacterial communities under the weeding methods (Fig. 4b). These results indicated that the structure of the soil bacterial communities was more similar between ASW and nWCW compared with that under WDW.

Abundance of three bacterial taxonomic groups
A LEfSe analysis of the taxa that were differentially abundant among the ASW, WDW, and nWCW treatments was performed (Fig. 6b). The cladogram revealed distinct differences in 42 taxa among the three treatments with LDA scores ≥ 4.0 (Fig. 6a). At the genus level, Candidatus_Solibacter were the most abundant in nWCW, and Ktedonobacter, Mucilaginibacter, and Leifsonia were the most abundant in WDW (Fig. 6c). The weed control method variously influenced bacterial composition at the phylum level. The phylum Gemmatimonadetes showed higher relative abundance under ASW compared with that under WDW and nWCW. Abundance of Firmicutes, Bacteroidetes, and Actinobacteria was significantly higher under WDW, and lower in the ASW and nWCW treatments. The relative abundance of Verrucomicrobia and Acidobacteria was higher in the nWCW treatment, but relatively lower under ASW and WDW (Fig. S1). Some taxa at the class, order, family, and species levels were revealed to differ distinctly in abundance among the three treatments (Fig. 6b).

Molecular ecology network analysis of soil bacterial communities
Potential bacteria-bacteria interactions were examined by constructing bacterial co-occurrence networks. Three networks were constructed based on OTU abundance (Fig. 7). Among the three weeding methods, differences in the values of network size, edges, degree, density, clustering coefficient, and modularity were observed (Table 3). With increase in number of nodes and edges, bacterial assemblages were clustered into a larger and more complex network of topological structures in the ASW and nWCW treatments compared with that under WDW, respectively. In addition, the values of average degree, graph density, and clustering coefficient were higher under ASW and nWCW than those under WDW. The modularity under WDW was larger than that of the other two treatments; the nWCW method showed the minimum modularity. However, the modularity in all three groups was higher than 0.4, which indicated that each weeding method resulted in an effective co-occurrence network. In summary, the co-occurrence networks of soil bacterial communities under ASW and nWCW were most similar and more complex than under WDW.

Relationship between bacterial community structure and soil properties
The Mantel test revealed that bacterial community structure was strongly correlated with soil properties, comprising soil moisture, NO 3 − -N, TP, DOC, NH 4 + -N, and pH (p < 0.05). The correlation coefficients showed the following trend: soil moisture > NO 3 − N > TP > pH > DOC > NH 4 + -N. The relationship of these soil properties with the dominant bacteria at phylum level was further analyzed using CCA (Fig. 8). The soil variables explained 38.0% of the bacterial community variation of which the first axis explained 29.6% and the second 8.4%. The correlations between the ten most abundant bacteria at phylum level and soil properties were analyzed (Table 4). The most dominant phyla were significantly correlated with soil moisture, whereas no bacterial phyla were correlated with NO 3 − -N and TP (p < 0.05). Gemmatimonadetes and Planctomycetes were positively correlated with pH and negatively correlated with NH 4 + -N, respectively (p < 0.05). Gemmatimonadetes and Verrucomicrobia were negatively correlated with TN, whereas Chloroflexi and Gemmatimonadetes were positively and negatively correlated with DOC, respectively (p < 0.05). Only Gemmatimonadetes was positively correlated with AP (p < 0.05).

Prediction of bacterial functional potential and phenotypes
To explore the functional roles of bacteria in the soil of the Chinese fir plantatiton, PICRUSt was used to predict  their function based on KEGG metabolic pathways and the relative frequencies of predicted functions in carbon metabolism, nitrogen metabolism, and the phosphorus cycle were analyzed. The functional discrepancy differed significantly among the weed control methods (Fig. 9a).
Under nWCW, bacterial metabolic pathways associated with methane metabolism, carbohydrate digestion and absorption, nitrogen metabolism, nitrotoluene degradation, and phosphotransferase system were significantly decreased compared with ASW and WDW (p ≤ 0.05). However, pathways concerned with carbon fixation in prokaryotes and oxidative phosphorylation were obviously reduced under WDW compared with ASW and nWCW (p ≤ 0.05). Based on 16 S rRNA gene sequence data, the phenotypic shift of the soil bacterial community was investigated using BugBase. Under WDW, the relative abundance of anaerobic and Gram-negative bacteria was distinctly decreased, whereas the relative abundance of bacteria possessing potential for pathogenic, stress tolerance, and mobile element-containing phenotypes was significantly increased (p ≤ 0.05) (Fig. 9b). The relative abundance of other bacterial phenotypic functions, such as aerobic and biofilm forming, did not differ significantly among the three weeding methods. The predictions of bacterial functional and phenotypic potential indicated that, compared with WDW, the relative frequencies and abundance of the bacterial functions and phenotypes were most similar between the ASW and nWCW treatments.

Effects of weeding methods on needle nutrient contents
One year after cultivated in three blocks, all these twenty-one Chinese fir seedlings weeding with three methods survived. The sugar content of new needles from these seedlings under nWCW were higher   than that under ASW and WDW (Fig. 10a). The content of GLU in new needles was significantly higher in nWCW than those in ASW and WDW (p < 0.05, Fig. 10b). The contents of TG and FFAs in new needles in the WDW treatment were significantly lower than those in new needles under the MSW and nWCW treatments (p < 0.05, Fig. 10b, c). The TCH content showed no significant difference among the weed control methods (Fig. 10b). Thus, the weeding methods influenced the soil properties, such as water and N contents, and the sugar and glucose contents of new needles. The correlations of bacterial alpha-diversity indices and soil properties with needle nutrient contents are summarized in Table S1. All alpha-diversity indices, comprising OTU number, ACE, Chao1, Simpson, and Shannon indices, showed a significant positive correlation with FFA content in new needles (p < 0.05). Soil moisture content was positively correlated with FFA content, and negatively correlated with DOC and TN contents (p < 0.05). The soil moisture content significantly contributed to production of sugar and GLU in new needles, whereas the TN content negatively influenced the biosynthesis of TG.

Transcription of genes associated with carbohydrate metabolism in new needles
To examine the change in energy metabolism of newly developed needles after weeding using three methods, the transcript abundance of genes associated with energy metabolism was analyzed. The Fig. 9 Predicted bacterial functions, functional potential, and phenotypes. a Relative frequencies of functional genes of bacteria predicted using PICRUST based on KEGG pathways associated with carbon metabolism (red), nitrogen metabolism (gray), and phosphorus cycle (purple). b Bugbase predictions of the organism level of microbiome phenotypes. A signifi-cant difference in relative frequencies of functional pathways, functional potential, and phenotypes between two groups were false discovery rate-corrected and is represented by an asterisk (p ≤ 0.05). ASW: Artificial sickle weeding; WDW: woody disc weeding; nWCW: nonwoven cloth weeding relative mRNA levels of genes associated with glucose metabolism were significantly changed. For example, the transcript level of the glucose kinase gene in new needles under nWCW was significantly higher than that under the other two weeding methods (p < 0.05, Fig. 11a). Compared with nWCW, the mRNA levels of sucrose phosphate synthase and sucrose synthase genes in new needles was significantly lower under WDW (p < 0.05, Fig. 11b, c). In contrast, the mRNA level of the hexokinase gene in the nWCW treatment was significantly lower than that under ASW and WDW (p < 0.05, Fig. 11d). These results indicated that the carbohydrate metabolism of new needles of Chinese fir was distinctly affected by the different weeding methods.

Discussion
Effects of weeding method on bacterial community structure In this study, Proteobacteria, Acidobacteria, Chloroflexi, Actinobacteria, and Verrucomicrobia were the predominant bacterial phyla in the soil of the Chinese fir plantation, which was consistent with previous reports (Liu et al. 2020a;Wang et al. 2018). Acidobacteria are among the most abundant and ubiquitous bacterial phyla in global soil environments, and they possess the capability to degrade complex and recalcitrant carbon compounds, including polysaccharides such as cellulose and xylan (Ward et al. 2009;Rawat et al. 2012;Ai et al. 2015). In addition, Acidobacteria harbor genes that encode polyketide synthase and nonribosomal peptide synthase. which are well known for their roles in the synthesis of antibiotics and antifungals (Ward et al. 2009). The significantly raised abundance of Acidobacteria indicated the promotion of carbon absorption and disease suppression on Chinese fir after using nonwoven cloth weeding method. The relative abundance of Verrucomicrobia were also significantly increased using nWCW method, suggesting their possible function in Chinese fir protection. It was reported that the abundance of phyla Acidobacteria and Verrucomicrobia were negatively correlated with the frequency of tobacco bacterial wilt (Niu et al. 2016). High abundance of Actinobacteria after weeding with WDW method indicated the higher carbon availability, because these bacteria exhibit the ability to utilize lignin-derived compounds and consequently are involved in the advanced stage of litter decomposition (Kirby 2006). Actinobacteria and Firmicutesare commonly enriched in disease-suppressive soils (Lee et al. 2021;Zhang et al. 2022). Though they were more abundant in WDW, a severe pathogenic potential was showed by predicted functional phenotypes. These may be explained by the existence of abundant Gram-negative Proteobacteria and Bacteroidetes. Combining the metadata analysis with machine learning, Zhang et al. (2022) revealed that Proteobacteria and Bacteroidetes are enriched in soil containing plant diseases. The increase in abundance of Gemmatimonadetes were significant within ASW treatment, which may associate with the drought tolerance of these bacteria (DeBruyn et al. 2011;Ullah et al. 2019) has revealed that the abundance of Gemmatimonadetes was predominated in the drought-treated rhizosphere of the cotton (G. hirsutum cv. Jin668).
At the genus level, Candidatus_Solibacter plays an important role in reduction of nitrite and nitrate, which contributes to the N cycle (Ward et al. 2009). These bacteria have a versatile genome that is adapted to breakdown recalcitrant organic compounds and carbohydrates, and provide a suitable environment for other microbes involved in the degradation of complex organic materials (Rawat et al. 2012;. Candidatus_Solibacter are abundant in the rhizosphere of black pepper, Panax notoginseng, and rice, and are positively associated with rice cropping (Lopes et al. 2014;Tan et al. 2017;Umadevi et al. 2017). Ktedonobacteria, members of the Chloroflexi phylum, were also highly abundant. The genus Ktedonobacter is characterized as mesophilic, aerobic, heterotrophic, moderately acidophilic bacteria that are also able to grow in a microaerophilic environment (Cavaletti et al. 2006). Ktedonobacter are linked with tobacco disease because these bacteria are nonnitrogen fixers that may compete with the microbial community for a N source (Niu et al. 2016). In the Pinus rhizosphere, Ktedonobacter are postulated to play a critical role in maintaining the microbial community structure . The genus Mucilaginibacter are heterotrophic bacteria capable of degrading pectin, xylan, laminarin, and other polysaccharides (Pankratov et al. 2007;Madhaiyan et al. 2010) isolated two Mucilaginibacter from the rhizosphere soil of cotton, Mucilaginibacter gossypii sp. nov. and Mucilaginibacter gossypiicola sp. nov., which produce abundant amounts of extracellular Fig. 11 Effects of different weeding methods on the transcription of genes associated with energy metabolism of newly developed leaves of Chinese fir seedlings. Relative expression of genes relative to the housekeeping gene GAPDH were estimated by RT-qPCR analysis. a Glucokinase (GK); b sucrose phosphate synthase (SPS); c sucrose synthase (SS); d hexokinase (HXK). Different lowercase letters above bars indicate a significant difference between two treatment groups (p < 0.05). Values are the mean ± standard error of the mean. ASW: Artificial sickle weeding; WDW: woody disc weeding; nWCW: nonwoven cloth weeding polysaccharides and possess plant-growth-promoting traits. The genus Leifsonia comprises Gram-positive, non-spore-forming, and irregular rod-or filamentshaped bacteria that are reported to differ in phenotypic potential. For instance, Leifsonia xyli subsp. xyli, a unique xylem-limited and fastidious bacterial pathogen, is the cause of ratoon stunting disease (Purcell and Hopkins 1996). These bacteria show obvious antagonistic activity, and can produce indole-3-acetic acid and promote root and shoot height of chili seedlings under an in vivo greenhouse environment (Passari et al. 2015). The present results indicated that after weeding using the WDW method, bacteria equipped with the plant-pathogenic potential were more abundant than under the ASW and nWCW weed control methods.

Correlation between soil bacteria and soil properties
Soil properties, such as N and P contents, are critical for plant growth in tropical forest plantations (Chen et al. 2015a, b). Nitrogen is considered to be a limiting factor that influences primary productivity in forest ecosystems, and continued anthropogenic P additions are essential to improve forest primary production (Cleveland et al. 2006;Weand et al. 2010). Although these nutrients may impact on the abundance, diversity, and community composition of soil microbes, which play an important role in regulating soil fertility, soil microorganisms are also important drivers of energy flow and nutrient cycling, such as carbon, N, and P cycling in terrestrial ecosystems (Artursson et al. 2006;Morris and Blackwood 2015). The input of N was previously believed to reduce microbial abundance and biodiversity (Zhou et al. 2017;Wang et al. 2017;Nie et al. 2018). In the present study, the number of 16 S rRNA-based OTUs and alpha diversity were closely associated with NH 4 + -N and TN contents, rather than AP and TP contents, which further confirms the importance of N in altering bacterial diversity. The results of recent studies also support that the inorganic P contention does not affect bacterial diversity (Huang et al. 2016;Liu et al. 2018). The present results suggested that inorganic P may not be a limiting factor that affected bacterial biodiversity in the Chinese fir plantation. Soil microorganisms are sensitive to changes in soil pH (Dai et al. 2018). In our study, soil pH was higher in WDW and nWCW treatment compared with ASW treatment, which may be facilitated by increasing DOC that may decrease soil acidity. Strong negative relationships between pH and DOC concentrations have been found in Scandinavian forest soils (Löfgren and Zetterberg 2011) and agricultural soils (Kemmitt et al. 2006). However, pH was not a decisive factor that affected bacterial diversity in this study, whereas soil moisture content had a significantly positive effect. Li et al. (2021) reported that soil moisture content significantly improves plant productivity, and soil bacterial abundance and diversity in the wheat root zone. In amended soils, bacterial diversity, richness, and community composition at 25% of the waterholding capacity distinctly differs from those at 55% of the water-holding capacity (Chen et al. 2015a, b). In WDW and nWCW treatments, woody disc and nonwoven increased the soil moisture by restricting the amount of evaporation away from the soi and allowing rainwater percolation respectively, which equipped with the similar effects like other inorganic mulch materials (Appleton and Hill 1997). In addition, we observed that the DOC content showed a significant negative correlation with bacterial diversity, which is consistent with the results of previous studies (Tian et al. 2018;Wang et al. 2018).
Different responses of major phyla to soil parameters, such as inorganic N and P contents, might be one reason for the change in bacterial diversity. The present study revealed that Gemmatimonadetes were negatively correlated with NH 4 + -N, NO 3 − -N, and TN contents, and Verrucomicrobia were negatively correlated with TN content. These findings were supported by a long-term fertilization experiment in which the relative abundances of Acidobacteria, Gemmatimonadetes, and Verrucomicrobia declined with N fertilization (Cederlund et al. 2014). However, the AP and TP contents did not affect the relative abundance of the dominant phyla, which may be because the soil available P content was less than 100 mg kg − 1 and soil-borne copiotrophic bacteria are not dominant (Kuramae et al. 2011). On the basis of Spearman's correlation analysis, soil moisture content was significantly correlated with the majority of bacterial phyla, which was supported by the results of a Mantel test and CCA analysis. These results implied that soil moisture content was a more important factor than other soil properties in shaping the bacterial community structure.

3
Vol.: (0123456789) Effect of three weeding strategies on Chinese fir growth Variation in environmental factors such as water, temperature, space and nutrition can lead to a significant decrease in the efficiency of photosynthesis in source tissues such as leaves. Weeding may positively influence the Chinese fir biomass, because the negative interference of grasses on forest species is associated with the competition for essential resources and allelopathic compounds produced by grass (dos Santos et al. 2019). In this study, the contents of sugar, glucose and free fatty acid were higher after weeding with nonwoven cloth, which were strongly associated with soil moisture. Many reports had stressed the importance of mulching in the conservation of soil moisture, which has been found to be an effective method for increasing the total sugar of plants such as plum and strawberry (Melgarejo et al. 2012;Bakshi et al. 2014). However, Guo et al. (2018) reported that the soluble sugar contents of Lycium ruthenicum Murr. seedling were significantly increased under drought treatment. Free fatty acids play a significant role in developmental stages of plant growth and defense against biotic and abiotic stresses. In this study, free fatty acid was closely correlated with bacterial alpha diversity and soil moisture. However, the potential interaction mechanism between soil bacteria and plant free fatty acid were rarely reported.
Glucokinase (GK) is a key enzyme in plant respiration and metabolism and can mediate phosphorylation of glucose to glucose-6-phosphate. Sucrose phosphate synthetase (SPS), synthesis of sucrose-6-phosphate from fructose-6-phophate and UDP-glucose, is crucial in the control of sucrose synthesis in plants. Sucrose synthase (SS) accelerates the reversible conversion of sucrose and UDP into UDP-glucose and fructose (Hirose et al. 2008). Hexokinase (HXK) catalyzes the phosphorylation of glucose and fructose during the glycolysis pathway to provide energy and substrates for plant activities, and acts as a sugar sensor for glucose signaling (Granot et al. 2013). After weeding with nonwoven clothes, the up-regulated expression of genes associated with energy metabolism including GK, SS and HXK were consistent with the contents of sugar and glucose in newly developed leaves of Chinese fir seedlings. Hence, the gene expression pattern on forest trees have become an important indicator to reveal the adaptation mechanisms to environmental condition.

Conclusions
Our results suggested that three non-chemical weed control methods had different effects on soil physicochemical properties, soil bacterial community structure, nutrient status of newly developed needles, expression in the needle of selected genes associated with energy metabolism, and growth of Chinese fir seedlings. Soil properties were closely correlated with bacterial diversity, and nutrient contents of new leaves. Soil moisture was clearly connected with the dominated bacteria at phyla level, and bacterial diversity had a significantly positive effects on the biosynthesis of sugar, glucose and free fatty acids in seedlings. ASW and nWCW methods were recommended to be applied in the weed control practice on the seedling growth of Chinese fir.