The research site was located in the Douling forest farm in the south-central part of Xinbin Manchu Autonomous County, Liaoning Province, China (geographical coordinates: 124° 52.37′E and 41° 30.41’N). This site is in the Longgang mountain range extension of the Changbai Mountain system. The mean altitude of the territory is approximately 500 m, the mountain slopes are gentle (between 10° and 25°), and the mountains and hills are low. The study area falls within the north temperate seasonal continental climate with four distinct seasons, a pleasant climate, and abundant rainfall (mean annual rainfall of 750–850 mm). The main tree species are Larix gmelinii, Pinus sylvestris var. mongolica Litv., Pinus thunbergii Parl, Pinus koraiensis Sieb.et Zucc., Quercus mongolica Fisch. ex Ledeb., Juglans mandshurica Maxim., and F. mandshurica Rupr. Shrubs in the plantations mainly include Euonymus alatus (Thunb.) Sieb, Lonicera japonica Thunb., Corylus L., Aralia elata (Miq.) Seem., and Acanthopanax senticosus (Rupr. et Maxim.) Harms.

Three plantation types of pure L. gmelinii, pure F. mandshurica, and mixed L. gmelinii and F. mandshurica were selected in August 2019 as the experimental plots for this study. The selected plots had similar slope aspects, stand age, and stand density. The mixed Larix and Fraxinus plantation was cultivated in mixed row mode (5 × 5). Three 20 m × 30 m standard plots were established for each forest type using the typical plot method, with a total of nine completely independent and randomly distributed sample plots (Table 1). Samples were randomly collected from 10 points at each sample plot. Soil samples were acquired at depths of 0–10, 10–20, and 20–40 cm in each sample plot. Additionally, in the mixed Larix and Fraxinus plantation, soil samples were acquired from within the Larix row, the Fraxinus row, and between the Larix and Fraxinus rows. After removal of visible plant roots and understory litter, the soil material was sieved through a 2-mm sieve. Aliquots for DNA extraction were stored at −20°C and at 4°C for the complete determination of enzyme activities within a 1-month period. The remaining samples were dried at 65°C to constant weight for the determination of the soil physicochemical properties.

All samples were analyzed for total nitrogen content, nitrate nitrogen content, ammonium nitrogen content, pH, organic matter content, total phosphorus content, available phosphorus content, and moisture content. Analysis of total N content was carried out using an elemental analyzer based on dry combustion (Vario EL III, Germany). Soil organic matter was measured by the external-heat potassium dichromate oxidation method; available nitrogen was measured by the alkaline hydrolysis diffusion method. Total phosphorus was determined by molybdenum antimony resistance colorimetry. Available phosphorus was measured by the NaHCO₃ extraction method (Yang and Yang, 2020). The moisture content of the soil was determined by drying to a constant weight at 65°C. pH values were determined in deionized water (1:2.5 for soil sample, m/v; Fioretto et al., 2000).

We examined the activities of five extracellular enzymes involved in soil C, N, and P degradation: β-glucosidase, sucrase, urease, acid phosphatase, and alkaline phosphatase. All soil samples were compared without the matrix, and a soil control was not used in the experiments. Soil enzyme activities were defined as mg of products produced over 24 h per gram of dried sample.

The activity of β-glucosidase was determined using the p-nitrophenol matrix method (Kourtev et al., 2002). Soil sucrase activity was measured by the 3,5-dinitrosalicylic acid method, urease activity was measured by the phenol sodium hypochlorite colorimetric method, acid and alkaline phosphatase activity was estimated using the disodium phenyl phosphate colorimetric method (Ge et al., 2017).

In accordance with the manufacturer’s instructions, total microbial community genomic DNA was extracted from 500 mg of soil material using the E.Z.N.A.® soil DNA Kit (Omega Bio-tek, Norcross, GA, United States). There were three duplications for each sample. Duplicate DNA extractions were performed for each sample and pooled. DNA extraction results were tested on 1% agarose gels; DNA concentration and purity were determined with a NanoDrop 2000 UV–Vis spectrophotometer (Thermo Scientific, Wilmington, United States).

The hypervariable V3–V4 region of the bacterial 16S rRNA gene was amplified with the primer pairs 515F (5′-GTGCCAGCMGCCGCGG-3′) and 907R (5’-CCGTCAATT CMTTTRAGTTT-3′) by an ABI GeneAmp® 9,700 polymerase chain reaction (PCR) thermocycler (ABI, CA, United States). The primers ITS3/ITS4 were amplified with the primer pairs ITS3F (5′-GCATCGATGAAGAACGCAGC-3′) and ITS4R (5′-TCCTCCGCTTATTGATATGC-3′). The PCR amplification conditions were as follows: 95°C for 3 min; 27 cycles of 95°C for 30 s, 55°C for 30 s, and 72°C for 45 s; 72°C for 10 min; and indefinite hold at 4°C upon completion. The PCR condition was the same except that the number of fungal PCR cycles was 35. The PCR mixtures contained 2 μl of 10 × TransSTART FastPfu buffer, 2 μl of 2.5 mM dNTPs, 0.8 μl of forward primer (5 μM), 0.8 μl of reverse primer (5 μM), 0.2 μl of TaKaRa rTaq DNA Polymerase, BSA 0.2 μl, 10 ng of template DNA, and sufficient ddH₂O to achieve a final volume of 20 μl. PCRs were carried out in triplicate. The PCR product was extracted from 2% agarose gel and purified using the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, United States), in accordance with the manufacturer’s instructions, and then quantified using a Quantus™ Fluorometer (Promega, United States).

The original data were allocated to each sample according to a specific barcode, and then, tags and primer sequences were removed. The FLASH software (V1.2.11) was used to splice paired-end reads (Chen et al., 2020). The Qiime software (V1.9.1) was performed to quality control and filter the spliced sequences (Whelan and Surette, 2017). Chimeric sequences were detected and removed by UCHIME (Edgar et al., 2011). Paired-end reads were obtained by MiSeq-sequencing and then spliced on the basis of their overlapping relationships. Quality control and filtering were carried out on the resulting sequences. After samples had been differentiated, operational taxonomic unit (OTU) clustering analysis and species taxonomy analysis were performed; the OTUs were subjected to multiple diversity index analysis and sequencing depth detection. On the basis of taxonomic information, OTUs were included in statistical analysis of community structure at each classification level. To obtain species classification information regarding each OTU, an RDP classifier Bayesian algorithm was used to analyze representative OTU sequences with a 97% similarity level. The bacterial species classification and comparison database were SILVA (Release138)¹, while the fungal were Unite (Release 8.0).² FAPROTAX database was used to predict the functional potential of bacteria (Sansupa et al., 2021). Functions of fungal communities were classified and analyzed by FUNGuild,³ with the fungi divided into pathotrophs, symbiotrophs, and saprotrophs (Nguyen et al., 2016). Each sample’s bacterial and fungal community compositions were determined at all taxonomic levels. Accordingly, various in-depth statistical and visual analyses were performed to determine the community composition and phylogenetic information of multiple samples; these included multivariate analysis and significant difference tests. All the raw datasets in this study were publicly available in the NCBI Sequence Read Achieve (SRA) database with an accession number SRP355360.

All statistical analyses in this study were performed by SPSS software (SPSS Inc., Chicago, IL). In all analyses, differences were considered statistically significant when p < 0.05 and extremely significant when p < 0.01. When data did not conform to assumptions of normality, the values were transformed into logarithm and/or square root before analysis. ANOVA of split plot was used to test the effects of microbial diversity (Chao, Shannon, and Simpson index) plantation type, soil depth, and their interaction on soil chemical properties and extracellular enzyme activities (Jones and Nachtsheim, 2009). The plantation types were the main plot, and soil depths were the split plot. Kruskal–Wallis H test was performed to analyze the difference of microbial groups in different soil samples. Additional statistical analyses were conducted using R software. Differences in soil microbial community composition and functions, based on changes in plantation types, were evaluated using principal coordinates analysis (PCoA), and analysis of similarities (ANOSIM) was performed in R using the “vegan” package based on Bray–Curtis dissimilarity distances (Calderón et al., 2017; Otsing et al., 2018). The sequences for all the samples were unified by minimum number of sample sequences before PCoA. BIO-ENV package and distance-based redundancy analysis (db-RDA) were implemented by the vegan package in R software to examine relevance among microbial composition, enzyme activity, and soil nutrient content (Wang et al., 2019; Testa et al., 2021).

ANOVA of split plot showed that the contents of total and available soil phosphorus were significantly affected by plantation type and soil depth (p < 0.05; Figure 1). With increasing soil depth, the contents of total and available soil phosphorus gradually decreased (p < 0.05). Among the three plantation types, the available soil phosphorus contents were significantly higher in pure Larix and Fraxinus plantations than in the mixed Larix–Fraxinus plantation (p < 0.05).

Total nitrogen content, nitrate nitrogen content, ammonium nitrogen content, organic matter content, moisture content, and pH in soil were significantly affected by soil depth (p < 0.05). With increasing soil depth, the contents of organic matter, total nitrogen, nitrate nitrogen, ammonium nitrogen, and moisture gradually decreased; in contrast, the pH value gradually increased.

The activities of β-glucosidase, sucrase, and urease were significantly affected by plantation type (p < 0.01; Figure 2). Compared with other soil samples, the β-glucosidase activities were significantly improved in the mixed Larix–Larix plantation soil. The sucrase, alkaline phosphatase, and acid phosphatase activities were significantly affected by soil depth (p < 0.01). The alkaline phosphatase and acid phosphatase activities significantly decreased in all three plantation types at increased soil depths. There were significant differences in sucrase activities among plantation types and soil depths (p < 0.05).

ANOVA of split plot showed that the indices of Shannon were significantly affected by plantation type and soil depth (p < 0.01; Table 2). The index of Simpson was only significantly affected by soil depth (p < 0.05). The index of Chao was only significantly affected by plantation type (p < 0.05). With the increase in soil depth, bacterial community diversity gradually decreased. Compared with Larix plantation, there was higher bacterial community diversity and richness in other plantations.

Principal coordinates analysis showed significant differences in bacterial community composition at family level among soil depth (ANOSIM, R = 0.32, p = 0.001; Figure 3A). There was no significant difference in bacterial community composition among plantation types (ANOSIM, R² = 0.15, p = 0.059). In total, 12 bacterial families were identified with the relative abundances >3% (Figure 3B). Among them, Gaiellaceae and Acidobacteriales were dominant flora. With increasing soil depths, the relative abundance of Gaiellaceae and Acidobacteriales was gradually increased (Figure 3C). Especially, the relative abundance of Gaiellaceae significantly increased in all soil samples (p < 0.05). Under the same soil depth, there was no significant difference in bacterial community composition among different plantation types.

Bacterial functional categories inferred by FAPROTAX were significantly differed among different soil depths (Figure 4). ANOVA of split plot showed that the relative abundances of functional genes for bacterial ligninolysis, nitrate ammonification, and nitrite ammonification were significantly affected by soil depth (p < 0.01). Kruskal–Wallis H test showed that with increasing soil depth, bacterial ligninolysis, nitrate ammonification, and nitrite ammonification functional genes significantly increased (p < 0.05). On the contrary, bacterial denitrification functional genes gradually decreased in different soil samples. There were no significant differences in bacterial metabolism functions in the same soil layer among the three plantation types (p > 0.05).

ANOVA of split plot showed that the indices of Chao were significantly affected by plantation type and soil depth (p < 0.05; Table 3). The index of Shannon and Simpson was not significantly affected by plantation type and soil depth (p > 0.05). Compared with Larix plantation, there was higher fungal community richness in other plantations. With the increase in soil depth, fungal community richness gradually decreased in all plantation types.

Principal coordinates analysis showed significant differences in fungal community composition at family level among plantation types (ANOSIM, R = 0.13, p = 0.037; Figure 5A). There were significant differences in fungal community composition among plantation types (ANOSIM, R² = 0.29, p = 0.001). In total, 20 fungal families were identified with the relative abundances >3% (Figure 5B). Mortierellaceae, Atheliaceae, Hygrophoraceae, and Inocybaceae were dominant flora in all soil samples. Compared with other plantations, the relative abundance of Inocybaceae was significantly higher in the Larix plantation (p < 0.05; Figure 5C). The relative abundances of Atheliaceae were lower in Fraxinus soils than other types of plantation soils. There were no significant differences in fungal community composition with increasing soil depths in the same type of plantation.

Fungal functions were classified by FUNGuild (Figure 6). The results showed that the mean relative abundance of guild-undefined saprotrophs comprised approximately 48% of all fungi OTUs. The mean relative abundance of Ectomycorrhizae comprised approximately 28% of all fungi OTUs. The relative abundances of saprotrophs were significantly higher in Fraxinus and Fraxinus–Fraxinus plantation soils than in other plantation soils (p < 0.05). In contrast, the relative abundance of Ectomycorrhizae was significantly lower in Fraxinus plantation soil than in other plantation soils (p < 0.05).

db-RDA showed that bacterial community compositions at family level were significantly correlated with pH, ammonium nitrogen, and available phosphorus contents (R² = 0.18, p = 0.017; R² = 0.18, p = 0.019; R² = 0.17, p = 0.02; Figure 7A). Among the bacterial families, Gaiellaceae and Acidobacteriales were positively correlated with available phosphorus and ammonium nitrogen contents, while negatively correlated with pH alone. Fungal community compositions were significantly correlated with pH, ammonium nitrogen, total nitrogen, available phosphorus, organic matter, and total phosphorus contents (R² = 0.44, p = 0.001; R² = 0.33, p = 0.001; R² = 0.20, p = 0.01; R² = 0.40, p = 0.001; R² = 0.20, p = 0.008; R² = 0.17, p = 0.017; Figure 7C). Among the fungal families, Inocybaceae was positively correlated with available phosphorus, ammonium nitrogen, and total phosphorus contents and negatively correlated with pH. Atheliaceae was opposite to Inocybaceae.

Bacterial community compositions were significantly correlated with acid phosphatase, β-glucosidase, sucrase, and urease activities (R² = 0.18, p = 0.013; R² = 0.58, p = 0.001; R² = 0.61, p = 0.001; R² = 0.32, p = 0.001; Figure 7B). Among the bacterial families, Gaiellaceae and Acidobacteriales were positively correlated with acid phosphatase and alkaline phosphatase activities. Fungal community compositions were significantly correlated with β-glucosidase, sucrase, acid phosphatase, alkaline phosphatase, and urease activity (R² = 0.26, p = 0.007; R² = 0.30, p = 0.004; R² = 0.53, p = 0.001; R² = 0.42, p = 0.001; R² = 0.19, p = 0.018; Figure 7D). Among the fungal groups, Inocybaceae was positively correlated with acid phosphatase and alkaline phosphatase activities, while negatively correlated with β-glucosidase, sucrase, and urease activities.

It was found that vegetation type and soil depth had significant effects on microbial community diversity. In this study, there was lower microbial community diversity and richness in Larix plantation than other plantations. The increase of microbial community biodiversity was conducive to soil health and forest growth (Van Bruggen and Semenov, 2000). Higher microbial diversity can increase the number and resilience of plant beneficial functions, which enable plants to produce beneficial characteristics that are difficult to obtain from any isolated species (Saleem et al., 2019). Therefore, cultivating coniferous and broad-leaved mixed plantation has the potential advantage of improving forest productivity. With the increase in soil depth, bacterial community diversity and fungal richness were gradually decreased. These findings consist with the previous study (Too et al., 2018). The possible reason for the results was that surface soil was rich in nutrients under the action of soil leaching and forest litter decomposition, which could provide niche for the growth of more microbial groups.

Bacterial communities and functions were shaped by soil depth but not plantation type. It is widely accepted that mixed plantations exhibit significant effects on soil microbial community structure (Chen et al., 2013; Pereira et al., 2017), although this effect might depend on soil depth. For example, Zhang et al. (2021) showed that mixed plantations did not significantly alter bacterial community structure in the upper soil. In this study, ANOSIM analysis revealed that there were significant differences in bacterial community composition among different soil depths rather plantation type. One possible reason was that the composition of soil bacterial community was mainly driven by soil nutrients (Dong et al., 2021). In this study, except phosphorus, other nutrient content had no significant differences among plantation types. Compared with other soil environmental factors such as pH, organic C, and inorganic N, phosphorus has a relatively low weight in shaping bacterial community structure (Hu et al., 2014; Wang et al., 2018). With the increasing soil depth, the nutrient content and pH significantly changed. This led to significant changes in bacterial community composition. These results were in agreement with the research of Zhang et al. (2016) and Li et al. (2019); the changes of total carbon, total nitrogen, and pH in different soil depths led to significant differences in bacterial community composition. Huang et al. (2013) showed that soil edaphic characteristics and soil profile might shape bacterial community functions; metabolic profiles revealed the carbon source utilization capacity in the surface layer. Using the BIOLOG system, Griffiths et al. (2003) and Braun et al. (2006) showed that the bacterial capability for utilizing diverse substrates decreased with soil depth. Our study showed that except for Larix plantation, ligninolysis, nitrate ammonification, and nitrite ammonification functional gene abundances significantly decreased with increasing soil depth in other plantations. This suggests that soil depth, rather than plantation type, significantly affected bacterial community structure and functional potential of nutrient cycling.

The abundance and diversity of fungal communities are closely related to tree species composition, and microbial communities associated with mixed species are more active (Khlifa et al., 2017). Notably, plant species, phylogenetic relationships, and plant traits affect the compositions of rhizosphere fungal communities (Sweeney et al., 2020). Our results suggested that soil fungal community composition significantly differed among the three plantation types. The main soil fungal families were Atheliaceae, Hygrophoraceae, and Inocybaceae, which belonged to the phylum of Basidiomycota. Most Basidiomycota were soil saprotrophs, and the relative abundances of soil saprotrophs were significantly higher in Fraxinus and Fraxinus–Fraxinus plantations than in other plantations. One possible explanation is that Fraxinus litter contains more nutrients than Larix litter, thus providing more niches for soil Basidiomycota. This was confirmed by Wu D. et al. (2019a), who showed that saprophytic fungi were more abundant in plantations with high soil nutrients. The mixed Larix and F. mandshurica plantation also had a higher relative abundance of ectomycorrhizal fungi than did a pure F. mandshurica plantation. Ectomycorrhizal fungi commonly occur in conifer roots, and the mixture of broadleaf trees with conifers would likely increase ectomycorrhizal fungal diversity (Ishida et al., 2007). Wu W. et al. (2019b) showed that mixed plantations of F. mandshurica and P. koraiensis contained a highly diverse and abundant population of ectomycorrhizal fungi. In that study, most Inocybaceae were ectomycorrhizal fungi (Matheny et al., 2009), and the relative abundance of Inocybaceae was significantly higher in Larix plantation soil. Our study indicated that mixed Larix and F. mandshurica plantations had significant effects on the fungal community. Thus, there is a need to closely monitor the impacts of mixed plantations on fungal communities in future sustainable management efforts.

This study found that mixed plantations did not significantly change general soil fertility or organic C and total N contents, compared with monoculture plantations. These results were consistent with the findings of previous studies, in which no significant differences were observed in soil organic carbon, total nitrogen, and available nitrogen contents among the three plantation types (Zhang et al., 2021). However, our study showed that plantation types and soil depths significantly affected the total soil phosphorus and available phosphorus contents. Among the three plantation types, the available soil phosphorus contents were significantly higher in Larix and Fraxinus plantations than in the mixed Larix–Fraxinus plantation. With increasing soil depth, available phosphorus content significantly decreased in all three plantation types. In contrast, some studies found significant increases in available phosphorus contents in the mixed-species treatments (Rachid et al., 2013; Zhang et al., 2021). One explanation is that faster growth leads to decreases in available phosphorus contents in mixed Larix–Fraxinus plantations (Feng et al., 2021).

Changes in soil microbial community structure serve as important indicators of soil quality and soil changes (Sharma et al., 2010). Bacteria and fungi reportedly contribute to nutrient bioavailability in degraded soils (Rashid et al., 2016). Dang et al. (2018) showed that soil organic carbon, total nitrogen, and available phosphorus contents were significantly correlated with soil bacterial communities; soil organic carbon, total nitrogen, total phosphorus, available phosphorus, and nitrate nitrogen contents were significantly correlated with soil fungal communities. The present study showed that bacterial and fungal community compositions were significantly correlated with the available phosphorus content. Our results showed that available phosphorus contents were significantly higher in Larix plantations than in the mixed Larix–Fraxinus plantation. Furthermore, higher relative abundances of Inocybaceae were observed in Larix plantation soils. Some fungi in Inocybaceae family have higher phosphate-solubilizing activity; thus, they can dissolve nearly all types of phosphorus in soil (Carrino-Kyker et al., 2016). This indicated that the higher abundances of Inocybaceae might increase inorganic phosphorus contents in Larix plantation soils.

Spatial and temporal variations in soil microorganisms are affected by differences in dominant tree species; these can ultimately change the availability and dynamics of soil nutrients, as well as shifts in microbial community composition, during adaptation to new environmental conditions. Physicochemical changes along the soil profile may be the main factor that controls the properties of bacterial communities (Ranjard and Richaume, 2001; Fierer et al., 2003). In the soil profile, total phosphorus content, available phosphorus content, pH value, organic matter content, and moisture content were significantly correlated with soil depth; these findings were consistent with the results of previous studies (Blume et al., 2002; Griffiths et al., 2003; Eilers et al., 2012). In the present study, the abundance of Gaiellaceae was positively correlated with available phosphorus content and phosphatase activity. It is reported that the members of family Gaiellaceae had high alkaline and acid phosphatase producing activity (Albuquerque and Costa, 2014; López-Lozano et al., 2020). With increasing soil depth, the relative abundance of Gaiellaceae significantly increased. It may be the distribution and physiological characteristic of Gaillaceae that makes the content of soil available phosphorus decrease with the increase in soil depth.

Soil enzyme activity plays a vital role in the turnover of carbon, nitrogen, and phosphorus; it is closely related to the distribution of soil microbial communities (Midgley and Phillips, 2016; Xu et al., 2017). Our results indicated that plantation type and soil depth had distinct effects on enzyme activities. In general, broad-leaved forests and mixed coniferous broad-leaved forests have higher soil enzyme activities than do coniferous forests alone (Hu et al., 2006; Xing et al., 2009). Similar to the previous results, our study showed significantly increased sucrase and β-glucosidase activities in Larix–Larix plantation soil. Alkaline phosphatase and acid phosphatase activities were significantly decreased with soil depth in the three plantation types, which is consistent with findings by Samuel et al. (2010). Distance-based redundancy analysis revealed that Gaiellaceae and Inocybaceae abundances were positively correlated with available phosphorus content and phosphatase activity. This indicated that soil available phosphorus variations were closely related to the relative abundance of Gaiellaceae at different soil depths and to the relative abundance of Inocybaceae in different plantation types. In general, our study indicated that mixed Larix–Fraxinus plantations resulted in distinct microbial groups, relative to the respective monocultures; these groups had positive effects on soil phosphorus content, potentially increasing the need for phosphorus anthropogenic fertilization.