Chicken manure application alters microbial community structure and the distribution of antibiotic-resistance genes in rhizosphere soil of Cinnamomum camphora forests

Abstract The distribution of antibiotic-resistance genes (ARGs) in environmental soil is greatly affected by livestock and poultry manure fertilization, the application of manure will lead to antibiotic residues and ARGs pollution, and increase the risk of environmental pollution and human health. Cinnamomum camphora is an economically significant tree species in Fujian Province, China. Here, through high-throughput sequencing analysis, significant differences in the composition of the bacterial community and ARGs were observed between fertilized and unfertilized rhizosphere soil. The application of chicken manure organic fertilizer significantly increased the relative abundance and alpha diversity of the bacterial community and ARGs. The content of organic matter, soluble organic nitrogen, available phosphorus, nitrate reductase, hydroxylamine reductase, urease, acid protease, β-glucosidase, oxytetracycline, and tetracycline in the soil of C. camphora forests have significant effects on bacterial community and ARGs. Significant correlations between environmental factors, bacterial communities, and ARGs were observed in the rhizosphere soil of C. camphora forests according to Mantel tests. Overall, the findings of this study revealed that chicken manure organic fertilizer application has a significant effect on the bacterial community and ARGs in the rhizosphere soil of C. camphora forests, and several environmental factors that affect the bacterial community and ARGs were identified.


Introduction
Soil is the largest reservoir of antibiotic-resistance genes (ARGs) in the environment (Allen et al. 2010 , Nesme andSimonet 2015 ), and ARGs in the soil are a major component of the environmental r esistome, whic h plays a major role in determining the resistance profiles of human pathogens (Forsberg et al. 2012, Nadeem et al. 2020 ).Many ARGs in the environment are derived from livestock man ure (McKinne y et al. 2018 ); se v er al ARGs hav e been identified in c hic ken, cattle, and pig manure (Wang et al. 2019 ).Most animal manure is applied to soil as fertilizer, and this mediates the spread of ARGs in the soil (Knapp et al. 2011, Qiao et al. 2018 ).The antibiotics in liv estoc k manur e will also enter the environment and remain for a long time (Yang and Carlson 2003 ).Although the r esidual concentr ation of antibiotics is v ery low, it will affect the antibiotic r esistance.Antibiotic-r esistant bacteria induced by antibiotic residues in animal feces are likely to enter the rhizosphere and spread in plants, and then enter the human body through the food c hain, whic h will pose a gr eat thr eat to human health (Zhang et al. 2012 ).There is, thus a need to study the accumulation and dissemination of ARGs in manur e-tr eated soil, as such studies can aid the de v elopment of strategies to mitigate the human health risks of ARG dissemination.
Although the application of manure can increase the yield of crops in various soil ecosystems, it can introduce large amounts of ARGs to the soil that e v entuall y r eac h the leav es of plants (Yang et al. 2018, Chen et al. 2018a ).T hus , the plant microbiome is the main route by which humans are exposed to ARGs in the environment.An increasing number of studies of environmental ARGs have been conducted, and the human health threats posed by the ARGs carried by plant microbiomes have received increased r esearc h attention (Abbassi et al. 2022 ).ARGs in the microbiome of plants increase via the absorption of antibiotics from agricultural soil; ARGs carried by microorganisms can also enter the endophytic or epiphytic microbiome of plants via the soil and plant roots (Blau et al. 2018 ).Ho w ever, the ability of ARGs to spread in plants and soil and its driving factors need to be r equir es further examination.
The accumulation of antibiotics in soil can impose strong selection for the evolution of resistance to antibiotics in microorganisms, and this can increase the diversity and abundance of ARGs in the soil (Heuer et al. 2011, Shawver et al. 2021 ).A metaanalysis conducted by Duarte et al. ( 2019 ) revealed a significant relationship between the abundance of ARGs and residual antibiotics, e v en at low antibiotic concentrations.Zhao et al. ( 2019 ) found that the content of oxytetracycline (OTC) in greenhouse soil under long-term cow manure and chick en man ure application was significantly positively correlated with the abundance of tetracycline ARGs ( tetO , tetW , and tetM ).The accumulation of antibiotics in the soil might favor the evolution of antibiotic resistance in soil micr oor ganisms, whic h could accelerate the spread of ARGs; the application of manure can provide sufficient carbon, nitrogen, and phosphorus for microorganisms in the soil; these alterations in the content of soil nutrients can promote the growth and r epr oduction of bacteria carrying ARGs as well as horizontal gene transfer (Peng et al. 2016 ).Changes in the structure of soil bacterial communities induced by long-term fertilization are thought to be the main factor affecting the abundance of ARGs (Forsberg et al. 2014, Chen et al. 2016, Han et al. 2018 ).Changes in the structure of soil bacterial communities induced by manure application explained 39.8% of the variation in ARGs; the second most important variable was soil nutrient conditions (Chen et al. 2016 ).Wang et al. ( 2020 ) found that the application of manur e significantl y c hanged the soil micr obial comm unity, whic h was considered as the main driving factor for the formation of soil ARG profile.Wang et al. ( 2018 ) found that long-term application of se wa ge sludge and c hic k en man ur e c hanged the composition of bacterial community in the phyllospher e, significantl y reduced the bacterial alpha diversity, and significantly increased the abundance of ARG.Variation distribution analysis and network analysis showed that ARG spectrum was closely related to bacterial community composition.These results have improved people's understanding of the diversity of plant-related antibiotic resistance and the factors affecting the distribution of ARG in the phyllosphere; in addition, soil and other environmental characteristics such as soil textur e, or ganic matter, total nitr ogen (TN), and soil enzyme activity affect the suitability of the soil environment for micr oor ganisms and, thus the div ersity and abundance of ARGs (Chen et al. 2016, Zhou et al. 2017, Yan et al. 2020, Li et al. 2021 ).
Cinnamomum camphora is a member of the family Lauraceae that occurs in tropical and subtropical regions of Asia.It is an economicall y important tr ee species in Fujian Pr o vince , China, for its high ornamental, medicinal, and economic value (Singh andJ aw aid 2012 , Chen et al. 2021 ).Most studies of C. camphora for ests hav e anal yzed its physiology, essential oil composition, and related secondary metabolites; ho w ever, few studies have examined the soil micr oor ganisms in C. camphora forests.Recent studies have revealed significant differences in the soil and rootassociated bacterial and fungal communities, as well as the diversity of C. camphora forests on different slopes (Chen et al. 2022a , b ).It is mentioned above that after applying chicken manure fertil-izer, more antibiotics and arginine may accumulate in plant soil, which will pollute the surrounding environment (Wang et al. 2019, Zhao et al. 2019 ).Nonetheless, we still lack a compr ehensiv e understanding of the effects of c hic k en man ure fertilization on the growth of C. camphora forests and the diffusion of ARGs in the soil (Chen et al. 2022a , b ).T hus , it is of great significance to evaluate the effect of a ppl ying c hic k en man ur e or ganic fertilizer on soilrelated ARGs and its growth in C. camphora forests.
Her e, we anal yzed the effects of fertilizer a pplication on the structure of soil microbial communities and the distribution of ARGs using 16S rRNA and metagenomic sequencing.We also anal yzed soil physicoc hemical pr operties , soil enzyme activities , and the content of OTC and tetracycline (Tet) antibiotics in C. camphora forests .T he aim of this study was to c har acterize c hanges in the composition, abundance, and div ersity of the rhizospher e soil bacterial community and ARGs of C. camphora forests following the application of chicken manur e or ganic fertilizer.We also analyzed the effects of various environmental factors on the soil bacterial community and ARGs in C. camphora forests .The results of this study provide new insights that will aid the development of manur e a pplication str ategies as well as the sustainable mana gement of C. camphora forests.Our findings also have implications for the de v elopment of str ategies to limit the abundances of ARGs in plants and soil.

Study site and sample collection
Our study was focused on the C. camphora forests in Nan'an, Fujian Pro vince , China (25 • 16 N, 118 • 31 E, 800 m).Two adjacent experimental plots were established with an interval of about 150 m, and pig manure fertilizer and organic chemical fertilizer were applied to each plot before 2010; the experimental plots were not subsequentl y tr eated with fertilizer for the next 10 years.In one sample plot, c hic k en man ur e or ganic fertilizer was a pplied twice a year for 2 years starting in 2020, with pH 8.8, organic matter (OM) 49%, TN 2.17%, available phosphorus (AP) 5.70%, and available potassium 4.26%, and still no fertilizer was applied to the other sample plot.The region has a subtropical monsoon climate, with hot and r ain y climate in summer and mild and humid climate in winter.T he a v er a ge annual temper atur e was 19.5 • C, and the aver a ge annual rainfall was between 800 and 1900 mm.
For each experimental plot, three subplots (10 m × 10 m) were established, all separated by a minimum distance of 10 m.Measurements of the ground diameter, height, crown width, and total biomass were taken from 20 C. camphora trees in each subplot ( Table S1 , Supporting Information ).
Cinnamomum camphora forests were sampled in mid-December 2022.The five-point and quartering methods were utilized to take soil samples from the 0-20 cm layer, which were then blended into a single sample for each subplot.Soil samples were then sealed and taken to the laboratory for analyses of soil physical and c hemical pr operties, soil enzyme activity, and the content of O TC and Tet antibiotics .Measur ements of se v er al soil physical and chemical properties were taken, including pH, electrical conductivity (EC), soil organic matter (SOM), soil soluble organic nitrogen (SON), AP, total carbon (TC), TN, and soil moisture content (SMC).Measurements of soil enzyme activity were taken for the following enzymes: soil nitrite reductase (S-NiR), soil nitr ate r eductase (S-NR), soil hydr oxylamine r eductase (S-HR), soil urease (S-UE), soil acid phosphatase (S-ACP), soil peroxidase (S-POD), soil acid protease (S-ACPT), soil dehydrogenase (S-DHA), soil-α-glucosidase (S-α-GC), and soil-β-glucosidase (S-β-GC) ( Tables S2 and S3 , Supporting Information ).
Thr ee tr ees wer e r andoml y selected fr om eac h subplot.A sterilized shovel was used to dig the soil of 0-20 cm under gr ound, located about 10 cm a wa y from the root system of C. camphora .After sie ving thr ough a 2-mm sie v e, samples fr om the same subplot were combined together and placed in a 50-ml sterile centrifuge tube, which was labeled as bulk soil.Samples of fine r oots wer e gather ed fr om eac h plant by tr acing the lar ge r oots fr om the base of the trunk in three directions, acquiring 9-15 root segments that measure 5-8 cm from each subplot.
Phosphate buffer was used to elute the rhizosphere soil attached to the roots (Xiao et al. 2017 ).First, 25 ml of phosphate buffer was poured into a 50-ml sterile centrifuge tube; it was then placed in a tube along with the roots and vortexed for 15 s at maximum speed on a vortex shaker (Vortex-Genie ® 2, Mobio Laboratories Inc., USA).The vortexed suspension was filtered through a 100-μm sterile nylon mesh into a new centrifuge tube and centrifuged for 15 min (3200 × g ).The resulting precipitate was labeled as the rhizosphere soil.The washed roots were then surface disinfected by submerging them in 0.5% (v/v) sodium hypochlorite solution and gently shaking for 3 min.Following two rinses with sterile water, root samples were placed on sterilized filter paper for drying.All the bulk soil, rhizosphere soil, and root samples were then stored at −80 • C until the DNA extraction process.

DN A extr action, 16S rRN A gene sequencing, and sequence analysis
The FastDNA ® Spin Kit for Soil (MP Biomedicals, USA) was used to extr act DNA fr om all samples.A NanoDr op 2000 ultr aviolet spectr ophotometer (Invitr ogen, Thermo Fisher Scientific, Waltham, MA, USA) was used to measure the purity of DNA.
The 16S rRNA V3 + V4 v ariable r egions of all DNA samples were amplified using the following primers: 338F (5 -A CTCCTA CGGGA GGCA GCA-3 ) and 806R (5 -GGA CTA CHV GGGTWTCTAAT-3 ) (Xu et al. 2016 ).PCR was conducted in 20 μl reactions containing the following components: 10 ng of genomic DNA, 4 μl of 5 × Fast Pfu buffer 4 μl, 2 μl of dNTP (2.5 mM each), 0.8 μl of upstream and downstream primers, 0.4 μl of Fast Pfu DNA pol ymer ase, 0.2 μl of bovine serum albumin, and the rest with double-distilled water.The thermal cycling conditions were as follows: 95 • C for 3 min, 30 cycles of 95 • C for 30 s, 50 • C for 30 s, and 72 • C for 45 s; and 72 • C for 7 min.An AxyPrep DNA Gel Extraction Kit was used to purify the PCR products, and a miniature fluorometer Quantus™ Fluor ometer (Pr omega, USA) was used to purify the PCR products.Library construction was performed using the NEXTFLEX ® Rapid DNA-seq Kit (BiooScientific, USA), and sequencing was conducted using the Illumina NovaSeq 6000 platform.

Metagenomic sequencing and bioinformatics analysis
Metagenomic sequencing of six rhizosphere soil samples was conducted using the Illumina NovaSeq 6000 sequencing platform.The 3 and 5 adapter sequences of the reads, low-mass sequences with lengths less than 50 bp, sequences with average base mass values less than 20, and sequences with N bases wer e r emov ed using FASTP softwar e (Chen et al. 2018b ).Optimized sequences were assembled using MEGAHIT software (Li et al. 2015b ) ( https:// github.com/voutcn/ megahit , version 1.1.2).Subsequent analysis was conducted on contigs with a length greater than or equal to 300 bp.The open reading frames of contigs were predicted using Prodigal software (Hyatt et al. 2010 ).Genes with lengths greater than or equal to 100 bp were translated into amino acid sequences.CD-HIT (Fu et al. 2012 ) ( http://www.bioinformatics.org/cd-hit/ , version 4.6.1)was used to construct a nonredundant gene set.The original sequences obtained by 16S rRNA and metagenome sequencing have been uploaded to the National Center for Biotechnology Information Sequence Read Arc hiv e database under the BioProject numbers PRJNA979820 and PRJNA980165, r espectiv el y.

Analysis of ARGs, Cluster of Orthologous Groups of proteins, and Kyoto Encyclopedia of Genes and Genomes
The amino acid sequences of the nonredundant gene set wer e compar ed a gainst the Compr ehensiv e Antibiotic Resistance Database (CARD), as well as the Evolutionary Genealogy of Genes: Non-supervised Orthologous Groups (eggNOG) and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases (BLASTP evalue ≤ 1 × 10 −5 , similarity ≥ 90%, comparison length ≥ 25 bp) using DIAMOND softwar e (Buc hfink et al. 2015 , 2021 ) ( http://www.diamondsearc h.or g/index.php, v ersion 2.0.13), and Cluster of Orthologous Groups (COGs) of proteins and KEGG (pathway level) annotation information was obtained for ARGs.Abundances of genes in all rhizosphere soil samples were determined according to the sum of ARGs, as well as the COG and KEGG (pathway le v el) gene abundances.Functional annotation information was used to identify Tet ARGs in drug class le v el.

Sta tistical anal yses
R softwar e v ersion 4.3.0softwar e was used to conduct all statistical analyses.Principal coordinate analysis (PCoA) based on the Bray-Curtis distance was used to evaluate the effects of fertilization on the bacterial community and functional gene composition, and the v egan pac ka ge (Oksanen et al. 2007 ) was used to conduct permutation multivariate analysis of variance (PER-MANOVA) to test the significance of the differences .T he PM-CMR pac ka ge (Pohlert and Pohlert 2018 ) was used to calculate the Shannon index of bacteria in different samples, to study the differences in bacterial diversity among different samples, and Wilcoxon rank-sum tests were conducted to characterize intergr oup differ ences in alpha diversity.The stats R pac ka ge (K eenan et al. 2013 ) was used to analyze the r elativ e abundances of inter gr oup differ ences in species and ARGs between fertilized and unfertilized samples; the threshold for statistical significance was P < .05.The Edger pac ka ge (Robinson et al. 2010 ) was used to analyze the differential expression patterns of highly differentiated OTUs and functional genes.The Linket pac ka ge (Zhu et al. 2022 ) was used to conduct Mantel tests, correlation analysis, and heat ma p anal ysis, combined with r edundancy anal ysis (RDA) based on Bray-Curtis distance to determine the relationship between Figure 1.Beta-diversity and alpha-diversity analysis of bacterial communities and functional genes in fertilized and unfertilized samples.PCoA plot showing variation in (A) bacterial communities based on the Bray-Curtis distance and (B) ARGs, Tet ARGs, COG genes, and KEGG pathway functional genes in fertilized and unfertilized rhizosphere soil samples (the significance of differences was evaluated using PERMANOVA).The x -axis and y -axis r epr esent the two selected principal axes, and the percentage represents the interpretation value of the principal axis for the difference in sample composition; the scale of x -axis and y -axis is a r elativ e distance, whic h has no pr actical significance.(C) Alpha div ersity (based on the Shannon index) of the bacterial community in all fertilized and unfertilized samples, and (D) alpha diversity (based on the Shannon index) of ARGs, Tet ARGs, COG genes, and KEGG pathway functional genes in fertilized and unfertilized rhizosphere soil samples.Each point represents different samples with and without fertilization, and the value of y -axis represents the Shannon index value of each sample.Kruskal-Wallis nonparametric test was used to obtain the P -value of the difference between groups, and Dunn's test was used to test the significance of the difference.* indicates significant differences at P < .05;* * indicates highly significant differences at P < .01. environmental factors and bacterial communities and ARGs.Cooccurrence networks based on the Spearman correlation matrix were constructed using the WGCNA pac ka ge (Langfelder and Horvath 2008 ).The nodes in these networks correspond to bacterial O TUs , and the edges indicate the correlations between them.A false discovery rate (FDR) correction for multiple comparisons was used to e v aluate the significance of the nodes and edges in the network.The igr a ph pac ka ge (Csar di and Nepusz 2006 ) w as used to calculate the network topological features of the bacterial community between fertilized and unfertilized samples (number of nodes , number of edges , degree , clustering coefficient, modularity, av er a ge path length, betweenness centrality, and diameter).The r elativ e importance of differ ent envir onmental factors (soil physicoc hemical pr operties, soil enzyme activity, OTC, and Tet) in determining the network topological features was determined using multiple regression on distance matrices (MRM).Correlations betw een netw ork properties and environmental factors were determined using the "corrplot" and "Hmisc" pac ka ges (Harr ell and Harrell 2019 , Pesenti et al. 2019 ).

Di v ersity of bacterial communities and ARGs
Unconstr ained PCoA r e v ealed significant differ ences in beta div ersity between fertilized and unfertilized samples.Significant differences wer e observ ed in the composition of the bacterial community in fertilized and unfertilized bulk soil, rhizosphere soil, and r oot endospher e (Fig. 1 A).Significant differ ences wer e observ ed in the beta-diversity of ARGs , Tet ARGs , COG, and KEGG pathway functional genes in fertilized and unfertilized rhizosphere soil samples ( P < .05)(Fig. 1 B).A significant difference in the alpha diversity of bacterial communities based on the Shannon index was observed between fertilized and unfertilized samples .T he alpha diversity of bacterial communities was significantly higher in fertilized samples in bulk soil, rhizosphere soil, and r oot endospher e than in unfertilized samples (Fig. 1 C).The alpha diversity of ARGs, Tet ARGs, and COG functional genes was significantly higher in rhizosphere soil in fertilized samples than in unfertilized samples; no significant differences were observed in the alpha diversity of Figure 2. Analysis of the differentially expressed genes of bacterial OTUs and ARGs in fertilized and unfertilized samples.(A) Differences in the number of bacterial OTUs between all fertilized and unfertilized samples, bulk soil, rhizosphere soil, and root endosphere.(B) Enriched and depleted bacterial OTUs in fertilized bulk soil, rhizosphere soil, and root endosphere using unfertilized bulk soil, rhizosphere soil, and root endosphere as contr ols, r espectiv el y.Eac h point r epr esents an individual OTU, and the position along the x -axis r epr esents the abundance fold change compared to unfertilized bulk soil, rhizosphere soil, and root endosphere .T he y -axis is −Log 10 (FDR) obtained by correcting the P -value of the significant difference .T he closer the point is to the top of the graph, the more significant the difference is.(C) Differences in the number of all ARGs, Tet ARGs, COG genes, and KEGG pathway genes.(D) Enriched and depleted ARGs, Tet ARGs, COG genes, and KEGG pathway genes in fertilized samples compared with ARGs, Tet ARGs, COG genes, and KEGG pathway genes in unfertilized samples, r espectiv el y.
KEGG pathway functional genes between fertilized and unfertilized samples (Fig. 1 D).

Significant enrichment and depletion of bacterial OTUs and ARGs
The number of unique OTUs in fertilized samples was as high as 7741, and the number of OTUs was significantly higher in fertilized samples than in unfertilized samples .T he number of unique OTUs was also significantly higher in fertilized bulk soil and rhizosphere soil samples than in unfertilized bulk soil and rhizosphere soil samples.No significant difference was observed in the composition of OTUs between fertilized and unfertilized r oot endospher e samples.Differ ential expr ession anal ysis based on OTU r e v ealed that 288 OTUs were significantly enriched in fertilized samples when the bacterial OTUs of the unfertilized samples were used as a control, and this was significantly higher than the number of depleted O TUs .T he number of enriched OTUs was higher in the fertilized bulk soil and rhizosphere soil than the number of depleted OTUs when the OTUs in bulk soil, rhizosphere soil, and root endospher e wer e used as contr ols; onl y 38 OTUs wer e enric hed in the r oot endospher e, and this was significantl y lo w er than the number of depleted OTUs (Fig. 2 A and B).
A total of 657 ARGs and 142 Tet ARGs were detected at the drug class le v el via the C ARD. T he number of shared ARGs, Tet ARGs , COG genes , and KEGG pathwa y genes the fertilized and unfertilized rhizosphere soil samples was higher than the number of unique genes, and the number of shared ARGs and Tet ARGs was 561 and 136, r espectiv el y.The number of ARGs , Tet ARGs , COG  genes , and KEGG pathwa y enriched genes in fertilized samples was significantly higher than the number of depleted genes when the numbers of ARGs, Tet ARGs, COG genes, and KEGG pathway genes in unfertilized samples were used as controls (Fig. 2 C and D).
The most abundant bacterial phyla were Proteobacteria, Actinobacteriota, and Acidobacteriota.The r elativ e abundance of Pr o-teobacteria was as high as 53.65% and 49.05% in fertilized and unfertilized samples, r espectiv el y.The most abundant bacterial class was Alpha pr oteobacteria, whic h accounted for 34.68% of the fertilized samples and 37.11% of the unfertilized samples .T he abundances of Acidobacteriae, Gamma pr oteobacteria, and Actinobacteria were also relatively high (Fig. 3 A and B; Tables S4 and S5 , Supporting Information ).The analysis of the top 10 ARGs at the antibiotic class le v el sho w ed that the r elativ e abundance of m ultidrug in fertilized and unfertilized rhizosphere soil samples was as high as 45.21% and 45.60%, r espectiv el y, and the abundances of MLS (macr olide-lincosamide-str eptogr amins) and Tet were also r elativ el y high ( Table S6 , Supporting Information ; Fig. 3 C).
Significant differences were observed in bacterial phyla and classes between fertilized and unfertilized samples.Significant differ ences wer e observ ed in the r elativ e abundance of the thr ee dominant bacterial phyla between the fertilized and unfertilized samples.Significant differ ences wer e observ ed in eight dominant bacterial classes between fertilized and unfertilized samples, and the most pronounced difference between fertilized and unfertilized samples was observed for Actinobacteria ( Figures S1 and S2 , Supporting Information ).Significant differ ences wer e observ ed in the r elativ e abundances of all ARGs between fertilized and unfertilized rhizosphere soil samples .T he abundances of the 10 most differ entiall y abundant ARGs wer e significantl y higher in fertilized samples than in unfertilized samples ( Figure S3 , Supporting Information ).

Effects of environmental factors on bacterial communities and ARGs
Following the application of chicken manure organic fertilizer, the gr ound diameter, tr ee height, cr own width, and total biomass of C. camphora for ests significantl y incr eased ( Table S1 , Supporting Information ).To determine the effects of a ppl ying c hic ken manur e or ganic fertilizer on envir onmental factors, we measur ed the pH, EC, SOM, SON, AP, TC, TN, and SMC of fertilized and unfertilized samples from C. camphora forests.We also measured the content of various soil enzymes, including S-NiR, S-NR, S-HR, S-UE, S-ACP, S-POD , S-ACPT , S-DHA, S-α-GC, and S-β-GC.In addition, the content of OTC and Tet was determined.The application of c hic k en man ur e or ganic fertilizer significantl y incr eased the content of SOM, SON, AP, OTC, Tet, S-NR, S-HR, S-UE, S-ACPT, and S-β-GC, and no significant differences were observed in the pH, EC, TC, TN, SMC, S-NiR, S-ACP, S-POD, S-DHA, and S-α-GC between fertilized and unfertilized samples (Fig. 4 ).Mantel tests were conducted on the functional genes, bacterial phyla and classes, and environmental factors .En vironmental factors were significantly correlated with all ARGs, Tet ARGs, COG genes, KEGG pathway genes, and bacterial phyla and classes (Fig. 5 A).
An RDA of ARGs at the antibiotic class le v el was conducted on environmental factors showing significant differences between fertilized and unfertilized samples.In the RDA, these environmental factors explained 98.04% of the variation in the composition of ARGs.MLS ARGs were the dominant ARGs; they were significantl y positiv el y corr elated with SON, AP, OTC, Tet, S-NR, S-HR, S-UE, S-ACPT, and S-β-GC and significantly negatively correlated with SOM.By contr ast, fluor oquinolone, aminocoumarin, and Tet ARGs were significantly positively correlated with SOM (Fig. 5 B).According to the RDA analysis, these environmental factors explained 97.99% and 92.13% of the variation in the composition of bacterial phyla and classes, r espectiv el y.The abundance of Acidobacteria was significantly positively correlated with SON, AP, OTC, Tet, S-NR, S-HR, S-UE, S-ACPT, and S-β-GC, and the abundance of Proteobacteria was significantly positively correlated with SOM ( Figure S4a , Supporting Information ).The abundance of Actinomycetia was significantly negativ el y corr elated with SON, SOM, AP, OTC, Tet, S-NR, S-HR, S-UE, S-ACPT, and Sβ-GC ( Figure S4b , Supporting Information ).

Effects of environmental factors on network-level features
We calculated the network-le v el topological featur es of the bacterial community in fertilized and unfertilized samples.Significant differ ences wer e observ ed in the betweenness centr ality v alues in the fertilized and unfertilized samples, and no significant differences wer e observ ed in the n umber of nodes, n umber of edges, degree, clustering coefficient, modularity, av er a ge path length, and diameter (Fig. 6 A).We determined the correlations between environmental factors and network-level topological features .T he number of nodes was significantly negativ el y corr elated with pH and significantly positively correlated with TC and TN.Between-ness centrality was significantly negatively correlated with SOM, OTC, S-NR, and S-ACP (Fig. 6 B).Ho w e v er, no significant differ ences wer e observ ed in the number of edges, degr ee, clustering coefficient, modularity, av er a ge path length, and diameter.

Composition and di v ersity of the bacterial communities and ARGs in C. camphora forests
Following the application of c hic k en man ur e or ganic fertilizer, all the growth indexes of C. camphora forests significantly increased, which suggested that fertilizer application played a k e y role in pr omoting the gr owth and de v elopment of C. camphora for ests.Fertilization can also lead to changes in the soil microbial comm unity.Micr obial comm unities play k e y roles in maintaining the health of a gr ofor estry ecosystems by mediating the formation of soil and biochemical processes, including residue decomposition and nutrient cycling (Qiu et al. 2014 ).Fertilization is an important mana gement pr actice that has a major effect on the structure of soil microbial communities; consequently, fertilizer application can lead to major changes in the community composition and diversity of soil micr oor ganisms (Allison et al. 2008, Hartman et al. 2015, Su et al. 2015 ).Pr e vious studies have shown that long-term fertilization can lead to substantial increases in the diversity of ARGs in soil under long-term organic fertilizer application (Xie et al. 2018b, Kang et al. 2022 ).This finding is consistent with the significant differences in the composition and structure of the bacterial community and ARGs between fertilized and unfertilized samples observed in our study; the alpha diversity of the soil bacterial community and ARGs significantly increased following c hic k en man ur e or ganic fertilizer.The r etention of some antibiotics in feces might selectiv el y favor the evolution of ARGs in soil (Xie et al. 2018a ).Most of the antibiotics a pplied ar e excr eted in feces and e v entuall y tr ansferr ed to the soil (Binh et al. 2008, Zhu et al. 2013, Jechalke et al. 2014, Tang et al. 2015 ).In our study, the content of OTC and Tet in c hic k en man ur e or ganic fertilizer and unfertilized soil was determined.The application of c hic ken manur e or ganic fertilizer led to significant incr eases in the content of OTC and Tet antibiotics.Ther efor e, the incr ease in the content of Tets antibiotics following fertilization might drive significant increases in the diversity of ARGs.
Analysis of the composition of OTUs in different sample types r e v ealed that the number of unique OTUs was significantly higher in bulk soil and rhizosphere soil than in root endosphere in both fertilized and unfertilized treatments .T he number of significantly enriched OTUs in fertilized root endosphere was the lowest compared with that in unfertilized root endosphere .T his finding suggests that the selection imposed by the roots of plants is unique to that imposed by other plant organs.Most functional genes and ARGs are common in fertilized and unfertilized rhizosphere soil samples; thus, the application of chick en man ure might have little effect on the diversity of functional genes.In addition, the main bacterial phyla detected in this study include Proteobacteria, Actinobacteriota, and Acidobacterota.In the fertilized and unfertilized samples, the r elativ e abundance of Proteobacteria accounted for 53.65% and 49.05% of all bacteria, r espectiv el y.The most abundant bacterial classes were Alphaprotobacteria, Acidobacteria, Gamma pr otebacteria, and Actinobacteria.These patterns are consistent with the results of a pr e vious study (Chen et al. 2022a ) that observed the same dominant bacterial species in C. camphora .
We found that the r elativ e abundance of ARGs was significantly higher in fertilized samples than in unfertilized samples .T his finding is consistent with the results of a pr e vious study (Chen et al. 2016 ) showing that the abundance of ARGs in farmland soil significantl y incr eased under the long-term a pplication of c hic ken man ure fertilizer.Animal man ure mediates the transport of antibiotics and ARGs into the soil envir onment, whic h makes the soil a reservoir of antibiotics and ARGs (Li et al. 2015a ).T hus , the abundances of ARGs in the soil increase following the application of c hic k en man ur e or ganic fertilizer.Ho w e v er, significant increases in the abundances of ARGs also lead to increases in environmental pollution risk, and appropriate management practices need to be implemented to mitigate these risks.For example, hightemper atur e composting can kill bacterial hosts and reduce the number of ARGs in feces, and the tr ansfer ability and r epr oducibility of genes preclude the complete elimination of ARGs (Wang et al. 2015, Xie et al. 2016 ).T hus , the use of antibiotics should be reduced to mitigate increases in the abundances of ARGs in the environment (Vikesland et al. 2017, Xie et al. 2018a ).

Relationships of environmental factors with bacterial communities and ARGs
Soil physical and chemical properties and soil enzyme activities are important properties of the soil environment that play a k e y role in the transformation of nutrients and the decomposition of OM (Qi et al. 2016 ).The application of chicken manure organic fertilizer significantly increased the content of SOM, SON, AP, S-NR, S-HR, S-UE, S-ACPT, and S-β-GC.Pr e vious studies examining the effects of manure fertilizer a pplication hav e shown that animal manure helps increase biological activity and soil fertility, impr ov e soil physical and chemical properties (Adebola et al. 2017, Onunwa et al. 2021 ), and promote the release of soil nutrients and the activity of soil enzymes (Adubasim et al. 2018, Ogumba et al. 2020 ).
Se v er al pr operties of soil affect the enric hment, migr ation, and transformation of ARGs, and the physical and chemical properties and enzyme activities in soil have a significant effect on the dissemination and distribution of ARGs in the environment (You et al. 2018, Qiu et al. 2021 ).According to an RDA of environmental factors and ARGs, these selected environmental factors explained a total of 98.04% of the variation in the composition of ARGs; environmental factors had a significant effect on the abundance of ARGs .T he results of this study sho w ed that MLS ARGs, which were the dominant ARGs, were significantly positively correlated with various physical and chemical factors and soil enzyme activities, and SON, AP, S-NR, S-HR, S-UE, S-ACPT, and S-β-GC might promote increases in the abundances of MLS ARGs.SOM plays a k e y role in mediating the spread of ARGs (Li et al. 2021 ).Increases in the content of SOM might also cause increases in the abundances of fluoroquinolone, aminocoumarin, and Tet.MLS ARGs wer e also significantl y positiv el y corr elated with the content of OTC and Tet, and significant increases in the OTC and Tet content following fertilization might increase the abundances of MLS ARGs to a certain extent.Significant increases in the content of SOM, SON, AP, S-NR, S-HR, S-UE, S-ACPT, and S-β-GC following the long-term application of chick en man ure organic fertilizer partiall y explained incr eases in the abundance of total ARGs.Consistent with the results of Xu et al. ( 2022 ), we found that environmental factors might be the most important for driving changes in the abundance and distribution of ARGs following the application of c hic k en man ur e or ganic fertilizer.Ther efor e, we speculate that environmental factors might have the strongest effects on interactions between microorganisms and ARGs.
An RDA of environmental factors and bacterial communities r e v ealed that the phylum and class le v els explained 97.99% and 92.13% of the variation in bacterial comm unity composition, r espectiv el y.The dominant bacterial phylum Acidobacteria played k e y roles in various biogeochemical cycles; Acidobacteria are metabolicall y div erse and significantl y positiv el y corr elated with most of the measured environmental factors.OM had a significant effect on soil bacterial diversity (Li et al. 2021 ), Proteobacteria play k e y roles in the decomposition and recycling of OM (Dang and Lovell 2016 ), which might promote increases in the SOM content.Actinomycetia was significantly negativ el y corr elated with all measur ed envir onmental factors .T he selected en vironmental factors all had significant effects on the dominant bacterial communities.According to the results of previous studies, the application of organic fertilizers can alter the diversity and composition of bacterial communities by altering the r ele v ant envir onmental factors of the soil (Chen et al. 2016(Chen et al. , 2021 ) ).

Changes in the netw ork-le vel topology of bacterial communities
We analyzed the network-level topological characteristics of bacterial communities in fertilized and unfertilized samples .T he betweenness centrality (i.e. the number of shortest paths passing through a node) was significantly lo w er in fertilized samples than in unfertilized samples, and no significant differences were observed in the eigenvalues for other topological c har acteristics .T he network-le v el topology pr ovides insights into inter actions among microbial species (Deng et al. 2012 ).The degree value (i.e. the number of adjacent edges) indicates the number of direct interactions between connected O TUs .Betweenness centrality indicates the importance of interactions between O TUs , and lo w er betw eenness centr ality v alues indicate that micr obes ar e located further a wa y from the core position in the network.These species might hav e r elativ el y small effects on the interactions between other species in the community (Greenblum et al. 2012, Ma et al. 2016 ).The results of this study indicate that differences in the interactions between micr oor ganisms in fertilized and unfertilized samples were low.We also conducted MRM analysis to identify environmental factors that could explain network-le v el topological changes in bacterial communities between fertilized and unfertilized samples.No significant correlations between environmental factors and topological features were observed, with the exception of the number of nodes and betweenness centrality.Most of the measur ed envir onmental factors had no significant effect on the topological features of the bacterial community.Changes in network-le v el topological featur es of the bacterial comm unity wer e anal yzed in our study; ho w e v er, the inferr ed inter actions are based on associations between two variables, and these corr elations ar e not sufficient for demonstr ating dir ect inter actions between micr oor ganisms and might not r eflect actual corr elations between taxonomic groups.Additional experimental work is needed to verify whether the associations between micr oor ganisms in our study reflect direct interactions.

Conclusions
We used 16S rRNA and metagenomic high-throughput sequencing technology to characterize the effects of c hic k en man ure organic fertilizer on the composition and diversity of bacterial communities and ARGs in the rhizosphere soil of C. camphora forests.We also studied the relationships of environmental factors with bacterial communities and ARGs.Significant differences in the composition of bacterial communities and ARGs were observed between fertilized and unfertilized rhizosphere soil.The abundance and diversity of rhizosphere soil bacterial communities and ARGs significantl y incr eased following the a pplication of c hic ken manur e or ganic fertilizer.The changes of soil physicochemical factors, soil enzyme activity, and Tet antibiotic content have significant effects on bacterial community and ARGs.Tets antibiotics and environmental factors were the main factors affecting changes in the abundance and diversity of rhizosphere soil ARGs following the application of chicken manure organic fertilizer.(grant number 118/KRC16006A), the Forestry Peak Discipline Construction Project of Fujian Agriculture and Forestry University (grant number 72202200205), the Forestry Science and Technology Pr omotion Pr oject of Fujian Pr ovince (gr ant number 2021TG10), the Cinnamomum camphora Genome Sequencing and High-quality Resource Screening Application Project (grant number 2021FKJ19), and the National Nature Science Foundation of China (grant number 32001128).

Figure 3 .
Figure 3. Relative abundances of the top 10 differential bacterial (A) phyla and (B) classes in fertilized and unfertilized samples; (C) relative abundances of the top 10 differential ARGs of different antibiotic classes in fertilized and unfertilized rhizosphere soil samples.

Figur e 4 .
Figur e 4. Soil en vir onmental factors in fertilized and unfertilized sample plots of C. camphora for ests.* indicates significant differ ences ( P < .05).Soil physical and chemical properties include pH, EC, SOM, SON, AP, TC, TN, and SMC.Soil enzymes include S-NiR, S-NR, S-HR, S-UE, S-ACP, S-POD, S-ACPT, S-DHA, S-α-GC, and S-β-GC.Each point represents bulk soil with fertilization and without fertilization, and the y -axis value represents the values of soil physical and chemical properties and enzyme activity of each sample.Kruskal-Wallis nonparametric test was used to obtain the P -value of the difference between groups, and Dunn's test was used to test the significance of the difference.

Figure 5 .
Figure 5. Corr elation anal ysis of functional genes, bacterial comm unities , and en vironmental factors .F ertilized and unfertilized rhizosphere soil samples.(A) Functional gene composition and correlation heatmaps of bacterial phyla and classes and environmental factors.Correlations between differ ent envir onmental factors wer e r epr esented by Spearman corr elation coefficients .T he size and color de pth in the bo xes indicate the correlations.Dunn's test was used to test the significance of the difference.* indicates that the difference is significant ( P < .05),and * * indicates that the difference is highly significant ( P < .01).The thickness of the line indicates the strength of the correlation inferred by the Mantel tests, and the different colors indicate significant differences.(B) RDA based on the Bray-Curtis distance was used to c har acterize the r elationship of ARGs with soil physicochemical factors , O TC, Tet, and soil enzyme activity.The x -axis and y -axis r epr esent the two selected principal axes, and the percentage represents the inter pr etation v alue of the principal axis for the corr elation of samples; the scale of x -axis and y -axis is the r elativ e distance, whic h has no pr actical significance .T he arrows indicate the lengths and angles between explanatory and response variables and reflect their correlations.Different samples are marked with different colors.

Figure 6 .
Figure 6.Corr elation anal ysis between envir onmental factors and network-le v el topological featur es .In both fertilized and unfertilized samples , (A) a Wilcoxon rank-sum test was used to compare the network-level topological features of bacterial communities to evaluate the significance of differences between fertilized and unfertilized samples; (B) the correlations between environmental factors and network-level topological features were calculated based on MRMs, and the size and color depth of bubbles in the squares indicate the magnitude of the correlation (positive R 2 values indicate positive correlations, and negative values indicated negative correlations).Wilcoxon rank-sum test was used to test the significance of the difference.* indicates significant differences ( P < .05),and * * * indicates highly significant differences ( P < .001).The y -axis value represents the value of each network-level topological features.