Tea oil camellia shell and TOCC were collected from Ceheng County (with an altitude of 664 m, north latitude of 24°59′55.05′, and east longitude of 105°47′14.05″), Qianxinan Prefecture, Guizhou Province, China. Decomposition agents were purchased from South China State Farming Agricultural Technology Development Co. Ltd. (Chengdu, China). Microbial agents were purchased from Shanghai Jiabosen Bioengineering Co., Ltd. (Shanghai, China). Commercial compound fertilizer [CF; total nitrogen (TN) content 20%, total phosphorus (TP) content 10%, and total potassium (TK) content 10%] was purchased from Guizhou Xiyang Industrial Co., LTD (Guiyang, China).

The TOCC underwent a 10-day pre-fermentation process with 3% starter culture and 60% water at 28°C. Upon reaching full mycelial growth, the preparation of pre-fermented TOCC (PTOCC) was subsequently dried for future use.

In this study, four TOCS-based organic fertilizers (labeled as BOF, BOFvo, OFcsc, and OFpd) were prepared used the reported methods with some modifications (Franke-Whittle et al., 2014; Lu et al., 2022; Chung et al., 2023). Composting aerobic fermentation experiments were conducted from June to August 2021 at the Guizhou Academy of Forestry, Nanming District, Guiyang City, Guizhou Province, China. As shown in Table 1, this study has implemented four treatments, each of which was replicated three times: (1) BOF: 80% of TOCS (partical size ≤5 mm) + 15% of PTOCC (partical size ≤2 mm) + 2.9% of urea (to adjust the initial compost C/N ratio to 32/1) + 0.1% of KH₂PO₄ + 2% of decomposition agent; (2) BOFvo: 80% of TOCS (partical size ≤5 mm) + 15% of PTOCC (partical size ≤ 2 mm) + 2.9% of urea + 0.1% of KH₂PO₄ + 2% of decomposition agent, with an addition of 3% microbial agent after complete fermentation of BOF; (3) OFcsc: 80% of TOCS (partical size ≤5 mm) + 15% of TOCC (partical size ≤2 mm) + 2.9% of urea + 0.1% of KH₂PO₄ + 2% of decomposition agent; (4) OFpd: 80% of TOCS (partical size ≤5 mm) + 15% of poultry dung (instead of pre-fermented TOCC to explore the effects of physicochemical properties and microbial communities on organic fertilizers by adding animal manure) + 2.9% of urea + 0.1% of KH₂PO₄ + 2% of decomposition agent. All the raw materials of each treatment were thoroughly mixed and then 50% water was added for moisture control. Subsequently, a plastic film was applied to the surface of each treatment to prevent moisture from evaporating during compost fermentation. The DTSW-2-A digital electronic thermometer (Taian Detu Automation Instrument Co., Ltd., Taian, China) was utilized to measure the temperature of each treatment. Once the temperature exceeded 55°C, the compost piles were turned weekly and left for 1 month until they were fully decomposed.

After fully decomposition, the physicochemical properties of BOF, BOFvo, OFcsc, and OFpd were analyzed using previously established methods with some modifications (Chu and Grogan, 2010). Briefly, pH values were measured using a PH3110 pH meter (WTW Xylem Inc., Munich, Germany) at a ratio of 1:10 [weight (g)/volume (ml)] for each TOCS-based organic fertilizer compared to distilled water. Total organic matter (TOM) content was determined through the wet oxidation method with some modifications (Nieuwenhuize et al., 1994). The contents of TN, TP, and TK were quantified using the Agricultural Industry Standard of the People’s Republic of China (Organic Fertilizer, NY/T 525-2021). The colony-forming units (CFUs) were quantified using a biological microscope (Shanghai Optical Instrument Factory Co., Ltd., Shanghai, China) (Sieuwerts et al., 2008). The unfermented TOCS was used as a negative control (CK). Each treatment was repeated three times.

The field trials of four TOCS-based organic fertilizers were conducted in a tea oil camellia forest located in Ceheng County (mean annual precipitation of 1,340.7 mm/a, mean temperature of 19.2°C), Qianxinan Prefecture, Guizhou Province, China. As illustrated in Figure 1, in January 2022, a total of 2 kg of TOCC-based organic fertilizers or CF were applied to each tea oil camellia tree by mixing them with the topsoil and backfill at a depth of 10–20 cm. The experimental period lasted from January to July 2022, during which, 3 parallel soil samples (6 points per replicate) were collected in July 2022 at a depth of 0–20 cm surrounding the lateral roots of each tea oil camellia tree at the fertilization position (Figure 1). The soil samples were promptly homogenized, mixed, and preserved in an on-site refrigerator before being expeditiously transported to our laboratory for subsequent analysis. CF was used as a positive control, whereas no application of different TOCS-based organic fertilizers or CF was used as a negative control (CK).

The rhizosphere soil porosities of tea oil camellia treated with different TOCC-based organic fertilizers or CF were measured using the micromeritics ASAP 2460 version 3.01 (Micromeritics Instrument Ltd., GA, USA). To gain a comprehensive understanding of soil porosity size distribution, the soil porosity diameters were categorized into five groups: crytopores (0.007–0.1 μm), ultramicropores (0.1–5 μm), micropores (5–30 μm), mesopores (30–75 μm), and macropores (>75 μm) (Cameron and Buchan, 2006; Lu et al., 2014). pH values were measured using a PH3110 pH meter (WTW Xylem Inc., Munich, Germany) at a ratio of 1:10 [weight (g)/volume (ml)] for each soil samples compared to distilled water. TOM content was determined through the wet oxidation method with some modifications (Nieuwenhuize et al., 1994). The contents of TN, available nitrogen (AN), TP, available phosphorus (AP), TK, and available potassium (AK) were quantified using the Forestry Industry Standard of the People’s Republic of China (LY/T1228-2015, LY/T1232-2015, and LY/T1234-2015). Each treatment was repeated for three times.

In July 2022, 3 parallel tea oil camellia leaf samples (6 points per replicate) were collected pre- and post-fertilization for each treatment (Figure 1). The leaf samples were promptly homogenized, mixed, and stored in an on-site refrigerator before being expeditiously transported to our laboratory for analysis of their physicochemical properties. After that, the TN, TP, and TK contents of tea oil camellia leaves were quantified according to the assessment standard in LY/T 1271-1999. Each treatment was repeated for three times.

In this study, the effects of different TOCS-based organic fertilizers on fruit yields of tea oil camellia were investigated by determining the average yield and fruiting rate of individual plants at the mature stage. Subsequently, 10 tea oil camellia trees treated with different TOCC-based organic fertilizers or CF were selected for each treatment and the average yield and fruiting rate of each individual plant were determined at the mature stage using the following formulas, where m is the weight of each camellia fresh fruits, n is the number of camellia flowers of each individual plant, and N is the number of camellia fresh fruits of each individual plant.

The genomic deoxyribonucleic acid (DNA) was extracted from the rhizosphere soil samples (approximately 2.0 g per sample) treated with BOF, BOFvo, OFcsc, and OFpd using a TGuide S96 Magnetic Soil and Stool DNA Kit (Tiangen Biotech Co., Ltd, Beijing, China) according to the manufacturer’s instructions. The DNA quality was assessed by running it on 1% agarose gels prior to its utilization as a template for PCR amplification targeting specific sequencing regions with the corresponding bacterial (341F: 5′-ACTCCTACGGGAGGCAGCA-3′ and 806R: 5′-GGACTACHVGGGTWTCTAAT-3′) and fungal (ITS2F: 5′-GCATCGATGAAGAACGCAGC-3′ and ITS2R: 5′-TCCTCCGCTTATTGATATGC-3′) primers. The PCR products obtained were subjected to sequencing using Illumina HiSeq™ 2000 (Illumina Inc., San Diego, CA, USA). The resulting raw reads underwent filtration with Trimmomatic v0.33 software, followed by primer sequence identification and removal using Cutadapt 1.9.1 software, resulting in clean reads devoid of primer sequences. Subsequently, Usearch v10 software was employed for overlap stitching of the clean reads, which were then filtered based on length range specific to different regions. Finally, the dada2 method implemented in QIIME2 2020.6 software was employed to eliminate the denoised and chimera sequences, resulting in the acquisition of non-chimeric reads.

The Vegan package in R language was utilized to analyze the relative abundance of operational taxonomic units (OTUs) and characterize soil microorganism diversity by calculating Shannon, ACE, and Chao1 indices. Meanwhile, principal component analysis (PCA) and heatmaps were utilized via the MetaboAnalyst website¹ to assess dissimilarities among different TOCS-based organic fertilizers treatments. Redundancy analysis (RDA) was employed in CANOCO5.² In addition, the experimental data mentioned in this article were analyzed using IBM SPSS Statistics software version 20 (IBM SPSS Inc., Chicago, IL, USA) with ANOVA followed by Duncan’s LSD test (p < 0.05).

The physicochemical property results of different TOCS-based organic fertilizers are summarized in Table 2. Intriguingly, compared to the CK group (unfermented TOCS), BOFvo could significantly increase the contents of TOM, TN, TP, and TK by 115.5, 13,492.6, 3,954.8, and 408.0%, respectively, indicating that incorporating a 3% microbial agent after complete fermentation potential can improve the quality of TOCS-based organic fertilizer.

The impact of different TOCS-based organic fertilizers on the pore characteristics of the rhizosphere soils is presented in Table 3. Results indicate that, compared to the CK and CF groups, the addition of TOCS-based organic fertilizers did not significantly affect the average pore diameter of rhizosphere soils. However, when compared to the CK and CF groups, the addition of TOCS-based organic fertilizers, particularly with BOFvo treatment, significantly increased both total pore volume and BET surface area of the rhizosphere soils by 41.00, 14.29 and 39.14, 33.33%, respectively.

Figure 2A demonstrates that, compared to the CK and CF groups, the addition of TOCS-based organic fertilizers enhances the adsorption capacity of rhizosphere soils under identical relative pressure, with BOFvo exhibiting the highest increase followed by BOF, OFpd, and OFcsc. Meanwhile, Figure 2B reveals that, compared to the CK and CF groups, BOFvo significantly improves pore size distribution in the rhizosphere soils. Additionally, Figure 2C illustrates that the soil pore size distribution was primarily within the range of 0.1–30 μm, with ultramicropores, micropores, mesopores, and macropores accounting for average pore volume proportions of 48.66, 48.24, 2.29, and 0.63%, respectively. Compared to the CK and CF groups, the addition of TOCS-based organic fertilizers, particularly with BOFvo treatment, BOFvo significantly increased the average pore volume proportions of ultramicropores, microporosity, and mesoporosity in the rhizosphere soils. These findings indicate that the application of BOFvo has a more significant impact on the pore characteristics of the rhizosphere soils.

As presented in Table 4, the application of TOCS-based organic fertilizers did not result in a significant alteration of soil pH values. However, compared to the CK group, the application of BOFvo led to a substantial increase in TOM, TN, AN, TP, AP, TK, and AK contents of the rhizosphere soils by 20.7, 187.6, 150.1, 264.7, 67.6, 38.8, and 201.5%, respectively. Notably, compared to CF, there was a remarkable enhancement in TOM, TN, AN, TP, TK, and AK contents of the rhizosphere soils by 35.5, 54.2, 22.1, 158.3, 39.6, and 53.8%, respectively. The results suggest that the utilization of BOFvo has a more pronounced influence on the physicochemical characteristics of the rhizosphere soils.

As depicted in Table 5, the application of TOCS-based organic fertilizers had a significant impact on the TN, TP, and TK contents of tea oil camellia leaves compared to CK and CF groups. Significantly, compared to CF, the use of BOFvo resulted in a 40.8 and 186.9% increase in TN and TP contents, respectively, indicating its potential for promoting tea oil camellia growth.

As shown in Table 6, compared to CK and CF groups, the application of TOCS-based organic fertilizers had a significant impact on the average yield and fruiting rate of individual plant at the mature stage. Significantly, in comparison to CK and CF group, BOFvo treatment resulted in a significant increase of 115.0, 35.4 and 252.8, 128.7% in average yield and fruiting rate, respectively, indicating its potential for enhancing the fruit yields of tea oil camellia.

As shown in Table 7, with respect to the sequencing data of bacteria in the rhizosphere soils, after cleaning and quality checking, 79,788, 79,732, 79,991, 79,772, and 79,794 clean reads were generated from CK, BOF, BOFvo, OFcsc, and OFpd treatment, respectively. Analyses of the OTUs revealed that 64,418, 58,466, 58,661, 59,311, and 63,486 OTUs were shared by CK, BOF, BOFvo, OFcsc, and OFpd treatment, respectively. In the meantime, CK, BOF, BOFvo, OFcsc, and OFpd treatments yielded 79,625, 79,828, 74,395, 60,150, and 79,576 clean reads as well as 75,665, 74,843, 64,802, 48,146, and 73,073 OTUs. It is noteworthy that the coverage exceeded over than 99.0%, indicating satisfactory numbers of analyzed OTUs with most bacteria and fungi present in soil samples being detected.

The alpha-diversity of bacterial and fungal communities in the rhizosphere soils was assessed by computing the ACE, Chao1, and Shannon indices. Table 8 reveals that for bacteria, the Chao1, ACE, and Shannon indices were significantly higher in BOFvo treated soils (exhibiting an increase of 85.1, 78.0, and 12.0%) compared to CK group. For fungi, Table 8 also shows that, compared with CK group, the Chao1 and ACE indices in BOFvo treatment (increased by 13.1, and 11.1%, respectively) were significantly increased. The results indicate that the incorporation of TOCS-based organic fertilizers, particularly BOFvo, significantly enhances the microbial diversity in the rhizosphere soils.

To explore the diversity of microbial communities in the rhizosphere soils treated with different TOCS-based organic fertilizers, we conducted PCA at the genus level to evaluate bacterial and fungal diversity. The PCA results for bacterial diversity (Figure 3A) shows that the five rhizosphere soil samples were distributed across different quadrants based on PC1 accounting for 42.6% and PC2 accounting for 33.1%, indicating large differences in the soil bacterial communities with different TOCS-based organic fertilizers treatments. The bacterial community in the rhizosphere soil treated with BOFvo differed significantly from those treated with CK, BOF, OFpd, and OFcsc. Meanwhile, the results of PCA analysis on fungal diversity (Figure 3B) reveals that the five rhizosphere soil samples were distributed in different quadrants based on PC1 accounting for 42.4% and PC2 accounting for 28.9%, indicating significant differences in the soil fungal community among various TOCS-based organic fertilizer treatments. Notably, compared to CK, BOF, BOFvo, and OFpd treatments, OFcsc treatment significant differences were observed at the genus level of fungal community in rhizosphere soil. In summary, the application of TOCS-based organic fertilizers exerted a significant influence on the microbial communities in the rhizosphere soil.

The impact of TOCS-based organic fertilizers on the relative abundance of bacterial and fungal phyla in the rhizosphere soils was assessed and the results were depicted in Supplementary Tables 1, 2 and Figure 4. According to the sequencing detected, a total of 30 bacteria (Supplementary Table 1) and 15 fungi phyla (Supplementary Table 2) were detected and classification in the all rhizosphere soils samples. As shown in Supplementary Table 1 and Figure 4A, Proteobacteria was identified as the dominant bacterial phyla present in the rhizosphere soils, followed by Acidobacteriota and Actinobacteriota. According to the taxonomic analysis, after BOF treatment, the relative abundances of Verrucomicrobiota, Nitrospirota, Myxococcota, Gemmatimonadota, Cyanobacteria, Chloroflexi, Bacteroidota, and Acidobacteriota in the rhizosphere soils significantly increased. After BOFvo treatment, the relative abundances of Verrucomicrobiota, Planctomycetota, Nitrospirota, Myxococcota, Methylomirabilota, Gemmatimonadota, and Bacteroidota in the rhizosphere soils significantly increased. Meanwhile, OFcsc treatment could significantly increase the relative abundances of Verrucomicrobiota, Proteobacteria, Nitrospirota, Myxococcota, Methylomirabilota, Chloroflexi, Bacteroidota, and Acidobacteriota in the rhizosphere soils. In addition, OFpd treatment significantly increased the relative abundances of Proteobacteria, Patescibacteria, Methylomirabilota, Gemmatimonadota, Cyanobacteria, and Actinobacteriota in the rhizosphere soils. Supplementary Table 2 and Figure 4B illustrates that the dominant fungal phylum was Ascomycota in the rhizosphere soils, followed by Basidiomycota, its relative abundance decreased after the addition of BOF, BOFvo, and OFcsc, while increased after the addition of OFpd. The relative abundances of Basidiomycota, Glomeromycota, and Mortierellomycota significantly increased with the BOF and BOFvo treatments. The OFcsc treatment significantly increased the abundances of Basidiomycota and Zoopagomycota, but decreased the abundances of Basidiomycota, Glomeromycota, and Kickxellomycota in the OFpd treatment.

The correlation between TOCS-based organic fertilizers and the structure of microbial communities was analyzed, with the findings presented in Supplementary Tables 3, 4 and Figure 5. According to the sequencing detected, a total of 466 bacteria (Supplementary Table 3) and 532 fungi genera (Supplementary Table 4) were detected and classification in all rhizosphere soils samples. Supplementary Table 3 also illustrates that the most dominant bacterial genus was Rhodanobacter, followed by Acidothermus, Sphingomonas, Acidibacter, Bryobacter, Burkholderia_Caballeronia_Paraburkholderia, and Candidatus Solibacter. Meanwhile, Supplementary Table 4 illustrates that the dominant fungal genus was Fusarium, followed by Leptobacillium, Trichoderma, Penicillium, Saitozyma, and Arthrocladium. As shown in Supplementary Table 3 and Figure 5A, compared to CK group, the abundance of bacterial genera significantly increased the abundance of Candidatus Solibacter, Candidimonas, Acidothermus, Acidipila Silvibacterium, Gemmatimonas, Ellin 6067, Acidibacter, and Chujaibacter with the BOFvo treatment; Gemmatimonas, Bryobacter, Jatrophihabitans, Sphingomonas, Chujaibacter, and Granulicella with the OFpd treatment; Candidatus Solibacter, Haliangium, RB41, Ellin 6067, Bryobacter, and Burkholderia Caballeronia Paraburkholderia with the BOF treatment; and Devosia, Acidibacter, Phenylobacterium, and FCPS473 with the OFcsc treatment. Supplementary Table 4 and Figure 5B shows that the abundance of fungal genera significantly increased the abundance of Fusarium, Mortierella, Saitozyma, Staphylotrichum, and Rhizophagus with the BOFvo treatment; Clitopilus, Metarhizium, Mortierella, Preussia, Purpureocillium, Cladosporium, and Talaromyces with the BOF treatment; Condenascus, Disculoides, Galerella, and Trichoderma with the OFcsc treatment; and Arthrocladium, Arxotrichum, Penicillium, and Talaromyces in OFpd treatments.

Redundancy analysis was employed to evaluate the associations between the physicochemical properties of different TOCS-based organic fertilizers and microbial communities (at a phylum level) of the rhizosphere soils. Table 9 shows that the pore volume (contribution rate 36.6%) and TP (contribution rate 36.4%) had the most significant effect on the bacterial community structure, followed by AP, pH, ultramicropore, surface, and TN. Among of them, pore volume, TP, AP, pH, ultramicropore, and surface significantly affected the bacterial community of the rhizosphere soils. Meanwhile, pore volume, TP, AP, pH, and surface significantly affected the fungal community, with the greatest effect by the surface (contribution rate 72.2%), followed by TK, ultramicropore, AP, TP, and pH, which shows that soil pore structure was the main factor affecting the fungal community of the rhizosphere soils.

As shown in Figure 6, the RDA PC1 and PC2 accounted for 92.44% (PC1 = 83.59% and PC2 = 8.85%) and 95.19% (PC1 = 79.30%, PC2 = 15.89%) of the bacteria and fungi total variation, respectively. Figures 6A, B demonstrate that the physicochemical properties of the rhizosphere soil, including TP, TN, TA, AN, TOM, AK, pore volume, ultramicropore, micropore diameter, and mesopore, have a positive impact on the abundance of bacterial community, such as Actinobacteriota, Bacteroidota, Verrucomicrobiota, Myxococcota, and Chloroflexi, conversely, macropore had a positive impact on the abundance of Proteobacteria, Acidobacteriota, and Gemmatimonadota. Meanwhile, Figures 6C, D show that Basidiomycota, Rozellomycota, Glomeromycota, Kickxellomycota, and Mortierellomycota were positively influenced by TK, TP, TN, AK, AN, AP, pH, TOM, ultramicropore, micropore, pore volume, and surface of the rhizosphere soil, conversely, Chytridiomycota and Ascomycota were primarily affected by the macropore of the rhizosphere soil. Therefore, the RDA results reveal that the physicochemical characteristics of TOCS-based organic fertilizers had a significant impact on the composition and distribution of microbial communities in the rhizosphere soils.