1 Introduction

Drying and rewetting (DRW) events are among the most prevalent environmental stresses that can significantly affect soil phosphorus (P) release, and lead to P loss from agricultural soils through leaching and runoff (Sun et al. 2017; Zhu et al. 2019). It is commonly believed that P exists as nano and colloidal particles (1–1000 nm) because the soil contains a richness of colloids with high specific surface area and charge, which have been proven to be P carriers in agricultural soils (Gu et al. 2020; Hens and Merckx 2002). Moreover, recent studies revealed that the released P during DRW events was derived from microbial cell lysis, which was also identified as colloidal P (Pcoll) (Dinh et al. 2017). However, Pcoll has low bioavailability and remains suspended in soil solution, contributing more to migration ability than settled particles (Yang et al. 2021a, 2021b). In general, Pcoll in soil environmental systems is considered as a potential risk for surface water eutrophication and has become a worldwide concern (Powers et al. 2016; Schindler et al. 2016).

Related studies have shown that the P released during DRW events is primarily in the form of organic P (Blackwell et al. 2009; Gu et al. 2018). A recent study reported that particulate organic P was the major specie in open bay waters, and colloidal organic P accounted for 33–65% (Yang et al. 2021a, 2021b). Therefore, Pcoll may be affected by soil phosphatase, such as alkaline phosphatase (ALP) and acid phosphatase (ACP) (Sun et al. 2017; Wang et al. 2021a, 2021b, 2021c). Relevant research demonstrated that the enzymatic hydrolysis of organic P and death of microbial cells contributed at least 51% and 16% to the increase in Olsen-P during DRW events, respectively (Wang et al. 2021a, 2021b, 2021c). Similarly, DRW events increased soil phosphatase activity and increased soil available P by at least 37% (Novair et al. 2021). Moreover, a meta-analysis compiled from over 1800 observations summarized that the drying process decreased soil enzymatic activities by 13–21%; and upon rewetting the dried soil, phosphatase activity increased by 12% (Gao et al. 2020). However, the relationship between Pcoll and phosphatase-producing microorganisms during DRW events is limited by technique.

Since phoD gene (ALP encoded) specific primers have been developed by Sakurai, the contribution of microorganisms to ALP activity can be determined by detecting ALP encoded gene (Nannipieri et al. 2018; Sakurai et al. 2008). Recent studies have shown a positive correlation between ALP activity and phoD-harboring microbial communities (Chen et al. 2019a, 2019b; Tian et al. 2021). Furthermore, the phoC gene determines the ACP activity by microorganisms, which is considered superior to the ALP activity in soils with neutral or low pH (Fraser et al. 2017; Luo et al. 2019). However, soil phosphatase activity and phosphatase-producing microorganisms can also be affected by DRW events, which creates a knowledge gap between P migration and microbial response under environmental stress (Chen et al. 2021; Zhu et al. 2019). It is thus of great environmental significance to reveal the relationship between ALP/ACP and Pcoll during DRW events from the genetic perspective.

Applying biochar has been shown to reduce P loss and regulate microbial composition (Wang et al. 2021a, 2021b, 2021c, 2023; Yang et al. 2021a, 2021b). For example, straw and manure biochar can effectively reduce the loss of soil Pcoll by 55.2−56.7%; and the diversity of feedstocks results in different microbial cooperation and competition (Wang et al. 2023). Moreover, recent research has found that adding biochar can enhance soil phosphatase activity and decrease P loss (Yang and Lu 2022). The potential explanations for this mechanism are as follows: (i) the adsorption/desorption capacity of soil P is altered by enhancing the soil characteristics such as soil aggregates (Islam et al. 2021; Xu et al. 2014); (ii) the large specific surface area of carbon frames provides shelter for microorganisms and enhances their resistance to DRW stress, relieving cell lysis and microbial P release (Yang et al. 2023); (iii) the biochar feedstocks provide the soil with labile carbon, which can effectively associate with soil P (Huang et al. 2021). Nevertheless, there are few reports to comprehensively evaluate the response of Pcoll and phosphatase activity during DRW events under different biochar addition strategies.

In this study, we performed a DRW platform to investigate the potential effects of different biochar strategies on Pcoll content in leachate and soil microbial mechanisms. Specifically, the objectives were to (1) evaluate the effect of biochar addition on mediating the content of Pcoll; (2) explore the microbial mechanism from ALP/ACP activity, and phoD/ phoC-harboring microbial community; and (3) identify key taxa species and their possible links.

2 Materials and methods

2.1 Soil sampling and biochar preparation

Top soil samples (0–20 cm) were collected from Pinghu Environmental Monitoring Station (soil A), Jiaxing City (120° 23′ 47″ E, 29° 44′ 09″ N) and Tiantai Environmental Monitoring Station (soil B), Taizhou City (120° 50′ 33″ E, 29° 7′ 38″ N), Zhejiang Province, China. The soil textures of soil A and B are silty loam and clay loam, respectively (United States Department of Agriculture, USDA). The soil sample was air-dried, and passed through a 2-mm mesh after manually removing all visible non-soil materials. Then, the soil sample was stored at 5 °C for further experiment.

The agricultural residues (rice straw and cattle manure) were collected locally for biochar production. Briefly, about 100 g air-dried feedstock was placed in a muffle furnace (QT-2.5-10, Suzhou Qiantong Instrument Equipment Co. Ltd., China), and heated to 400 °C at a rate of 8 °C min−1 with a residence time of 2 h under oxygen-limited conditions using the slow pyrolysis method (Saffari et al. 2020). After cooling to room temperature, the biochar was removed and sieved through a 5-mm mesh. The basic physicochemical properties of the two soils and used biochar are presented in Additional file 1: Table S1.

2.2 Experimental design

To perform the leaching experiment, soil columns were made of acrylic glass with a length of 40 cm and an inner diameter of 10 cm. Each column was filled with about 2 kg of air-dried soil and treated as follows: (1) control without biochar addition (CK); (2) straw biochar addition rate of 0.5% (SBL); (3) straw biochar addition rate of 1.0% (SBH); (4) manure biochar addition rate of 0.5% (MBL); and (5) manure biochar addition rate of 1.0% (MBH). The soil columns were sealed, each with a rubber plug with a hole pierced in the middle. Deionized water (18.2 MΩ cm, Millipore) was used as the influent solution and injected with a soft hose. The leachate was discharged via a peristaltic pump. The bottom of the column was filled with quartz sand and gauze to prevent soil from clogging the water outlet. Three replicates were established for each treatment.

Each treatment ran through three consecutive DRW cycles (named cycle-1, -2, and -3) (Blackwell et al. 2009). The leaching rate was maintained at 1.5 ml min−1. After each leaching for 10 h, the peristaltic pump was disconnected, and the soil was then gradually dried by flushing N2 (90 L h−1, 25 ± 2 ℃) until it reached a constant weight. The soil was air-dried within 3 days before the next leaching stage, regarded as completing the first DRW cycle (Gu et al. 2018). Soil samples were collected after the third DRW cycle. A portion was used for soil physicochemical property analysis; and the frozen soil samples (−80 °C) for DNA extraction.

2.3 Soil chemical and biological analyses

Soil pH was measured in a 1:5 (w:v) soil/water suspension using a pH meter (PHS-3 C, Shanghai, China). Soil total carbon (TC) and total nitrogen (TN) were determined using an Elemental Analyzer (Vario MAX CNS, Elementar, Germany). Soil total P was measured by H2SO4–HClO4 digestion, followed by the molybdenum blue colorimetry (Walker and Adams 1959). Cation exchange capacity (CEC) was measured as described by Gaskin et al. (2008). Soil texture was measured using the hydrometric method (Li et al. 2020). The Pcoll concentration was determined as described by Pagel et al. (2008). Briefly, the leachate (i.e., Sample I) was centrifuged at 3000×g for 10 min, after which the supernatant was filtered through 1000 nm filter paper (Millipore Amicon, Ultra, USA; Sample II). And then the leachate was centrifuged at 4000×g for 45 min using a membrane with a pore size of 3 kDa (Millipore Amicon, Ultra, USA; Sample III) to represent the particle size < 1 nm. The particulate P (PP, > 1000 nm), Pcoll (1–1000 nm), and truly soluble P (TSP, < 1 nm) were calculated based on the difference between the P concentration of Samples I, II, and III (Eltohamy et al. 2021). The total dissolved P (TDP) was the sum of PP, Pcoll, and TSP.

The detection of ALP and ACP activities was described by Tabatabai and Bremner (1969). In brief, ALP and ACP were cultured in the buffer solution with pH 6.5 and 11, respectively, and then p-nitrophenyl phosphate (PNP) was used as a substrate. Enzyme activities were considered to be the amount of p-nitrophenol produced by 1 g of soil per hour.

2.4 Soil DNA and qPCR quantification

The total soil DNA was extracted from 0.5 g of frozen soil using the DNeasy PowerSoil Kits (Qiagen, Hilden, Germany). The DNA purity was determined by NanoDrop ND-2000 spectrophotometer (Thermo Scientific Wilmington, DE, USA), and then tested with 1% (w/v) agarose gels.

The phoD and phoC gene abundances were determined by quantitative polymerase chain reaction (qPCR) using a QuantStudio 3 Real-time PCR System (Applied Biosystems, Foster City, CA, USA). The primer pairs amplifying phoD and phoC genes were ALPS-F730 (5′-CAGTGGGACGACCACGAGGT-3′)/ALPS-R1101 (5′-GAGGCCGATCGGCATGTCG-3′) (Sakurai et al. 2008); and phoc-A-F1 (5′-CGGCTCCTATCCGTCCGG)/phoc-A-R1 (5′-CAACATCGCTTTGCCAGTG) (Fraser et al. 2017). The PCR was run in triplicate using 20 µL reactions comprising 10 µL 2 × Taq MasterMix, 1 µL of each primer (10 μm), 1 µL template DNA, and 7 µL of ddH2O. The cycling conditions were 95 °C for 3 min, followed by 40 cycles at 95 °C for 15 s, 60 °C for 15 s, 72 °C for 20 s, and 60 °C for 1 min. The triplicate PCR products were then purified using the AxyPrep DNA Gel Extraction Kit (Axygen, Hangzhou, China). The standard curve was constructed using 10-fold dilution sequences of linearized plasmids containing target genes.

2.5 Illumina sequencing and sequence analysis

Sequencing was performed on the Illumina MiSeq 2000 platform by Biomarker Technologies Corporation (Beijing, China). The raw sequencing data were performed by the UCluster method in QIIME2 (https://qiime2.org/), and the RDP FunGene pipeline (Bolyen et al. 2019; Fish et al. 2013). Briefly, the overlapping counterpart reads were assembled and the primer sequence was removed. The phoD gene sequences of 200–350 bp length (about 2% discarded), and phoC gene sequences of 100–175 bp length (about 2% discarded) were retained. And then the non-corresponding sequence of phoD/phoC was removed through the RDP FunGene pipeline. In total, 1,116,735 (phoD) and 1,140,791 (phoC) high-quality sequences were obtained. The phoD and phoC purified sequences in each sample were assigned to the operational taxon with a similarity of 80% and 88%, respectively (Luo et al. 2019). Maximum-likelihood phylogeny was calculated by MEGA 7 and visualized using the iTOL webserver (http://itol.embl.de/). The original DNA sequences have been uploaded to the National Center for Biotechnology Information and registered with the BioProject ID number: PRJNA911733 (https://www.ncbi.nlm.nih.gov/bioproject/911733).

2.6 Statistical analysis

Unless otherwise indicated, one-way analysis of variance (ANOVA), followed by least significant difference (LSD) was performed using SPSS software (version 21) to compare means for each variable (p < 0.05). Principal coordinate analysis (PCoA) was performed using the “vegan” package to explore dissimilarity based on Bray-Curtis distances (Wang et al. 2018).

Network analyses were performed at the operational taxonomic units (OTU) level using the Spearman correlation matrix (Wei et al. 2019); and correlation coefficients were calculated using the “psych” package (Yan et al. 2021). The error discovery rate correction was performed using the Benjamini-Hochberg procedure, keeping the merged edges with p-values < 0.05 (Benjamini and Hochberg 1995). Moreover, the final networks were visualized by Gephi software (version 9.0.2) (Jacomy et al. 2014).

A partial least-squares path model (PLSPM) was established to quantify the contribution of phosphatase activity (ALP, ACP), gene copy numbers (phoC, phoD), microbial alpha diversity (Chao 1, ACE, Simpson, and Shannon index), and Pcoll content using the R package “plspm” (Wang et al. 2023).

3 Results

3.1 Distribution of P in leachate during DRW events

DRW produced the initial peak of P concentration in the leachate of the two tested soils (A and B), which then gradually decreased until the end of the leaching cycle (Table 1). TDP concentrations in the leachate of both soils decreased significantly from cycle-1 to cycle-3 (F = 11.9, p < 0.05 for soil A; F = 13.4, p < 0.05 for soil B). As a result, the TSP and PP of the biochar treatments (i.e., SB and MB) were always greater than those of the control. However, the biochar addition resulted in a significant decrease in Pcoll concentration from the beginning to the end of the leaching cycle in both soil A and B (F = 5.8, p < 0.05), with no significant difference between the biochar treatments. Consequently, biochar addition during the three DRW cycles reduced the Pcoll release from soil A by 17.6–32.1% and from soil B by 15.5–17.6%, compared to the CK. Notably, the Pcoll concentration in the leachate of soil A was significantly lower than that of soil B (F = 63.1, p < 0.001), indicating that Pcoll in the weakly acidic soil leaching may be greater than that in the weakly alkaline paddy soil.

Table 1 Phosphorus distribution in leachate during DRW cycles
Full size table

3.2 Phosphatase activities and phoD/phoC genes abundance

Compared to CK, adding biochar increased the potential ALP activities by 7.7–15.1% in soil A and 17.1–49.1% in soil B after the three DRW cycles (Fig. 1a). By contrast, the ACP activities in soil A decreased by 7.3–32.9% after the biochar addition, and they varied between −26.2% and 18.5% in soil B (Fig. 1b), because biochar inhibited ACP activity in alkaline soil. The variation trend of phoD and phoC genes abundance was similar to that of ALP and ACP activity (Fig. 1c, d). Generally, the biochar addition contributed to the increase in ALP activity and phoD gene abundance (p < 0.05), and the decrease in ACP activity and phoC gene abundance (p < 0.05). There was a significant difference in ACP activity and phoC gene abundance between the tested treatments (p < 0.05), whereas ALP activity and phoD gene abundance remained relatively stable. In addition, ALP activity was positively correlated with phoD gene abundance in the two soils (R2 = 0.514, p < 0.05 for soil A, and R2 = 0.713, p < 0.05 for soil B) (Fig. 2a). And the same trend of correlation was found between the ACP activities and the phoC gene abundance (R2 = 0.579, p < 0.05 for soil A, and R2 = 0.702, p < 0.05 for soil B) (Fig. 2b). In addition, the Pcoll was significantly negatively correlated with the phosphatase activity and the gene abundance (p < 0.05) in both soils (Fig. 2c–f).

Fig. 1
figure 1

Soil alkaline phosphatase (ALP) activities (a), the abundance of the phoD gene estimated by the gene copy numbers (b), acid phosphatase (ACP) activities (c), and abundance of the phoC gene estimated by the gene copy numbers (d). Different letters indicate significant differences between treatments (p < 0.05). Error bars were calculated from triplicate experiments conducted in parallel on the same soil. CK no biochar addition, SBL 0.5% straw biochar addition, SBH 1% straw biochar addition, MBL 0.5% manure biochar addition, MBH 1% manure biochar addition

Full size image
Fig. 2
figure 2

Relationships between alkaline/acid phosphatase activities and the abundances of phoD/phoC genes and colloidal phosphorus (Pcoll)

Full size image

3.3 PhoD- and phoC-harboring microbial abundance and community structure

The phylogenetic relationship between phoD- and phoC-harboring microbial communities was established at the OTU levels (Fig. 3). The OTU diversity may encode different phoD- and/or phoC-harboring microbial communities; at the same time, any one phyla may also encode multiple OTUs. The predominant phylum of phoD gene was Proteobacteria, Actinobacteria, Cyanobacteria, Firmicutes, Deinococcus-Thermus, Gemmatimonadetes, and Planctomycetes; the superior phylum within phoC-harboring was Proteobacteria and Cyanobacteria.

Fig. 3
figure 3

Phylogenetic relationships among the phoD- (a) and phoC-harboring (b) microbial communities based on OTU abundance (mean related abundance > 0.1%). Leaf labels indicate representative sequence IDs (OTU). The colors of the branches and OTU represent different taxonomies of the sequences based on the phyla and genus level taxonomy, respectively. Rings from the inner to the outside circles represent the relative abundance of OTU in different treatments. CK no biochar addition, SBL 0.5% straw biochar addition, SBH 1% straw biochar addition, MBL 0.5% manure biochar addition, MBH 1% manure biochar addition

Full size image

The alpha diversity of phoD- and phoC-harboring microbial communities were characterized by ACE, Chao1, Shannon, and Simpson diversity indices (Table 2). Variation in all beta diversity indices variation was affected by soil type and biochar addition (Additional file 1: Fig. S1). Specifically, two principal coordinates fully explained 52.02% (R2 = 0.247, p = 0.008) and 37.98% (R2 = 0.317, p = 0.001) of the variation in phoD- and phoC-harboring microbial communities, respectively.

Table 2 Alpha diversity indices of the soil microbial community
Full size table

The networks showed that phoD-harboring species (Fig. 4a) were more closely connected than phoC-harboring species (Fig. 4b) in both the control group and biochar addition groups. We found a 1.5-fold increase in node number after biochar addition in phoC-harboring species; in contrast, there was no significant change in phoD-harboring species, suggesting biochar addition could lead to more conflicting interactions among the phoC-harboring species (Additional file 1: Table S2). In addition, seven phyla of these OTUs were observed in the phoD-harboring species, belonging to Proteobacteria, Actinobacteria, Cyanobacteria, Acidobacteria, Planctomycetes, Gemmatimonadetes, and Verrucomicrobia. Similarly, the OTU59, 90, and 206 were identified in the phoC-harboring species, which were considered the keystone with more edges, and further analyzed related to Proteobacteria, Cyanobacteria, and Bacteroidetes.

Fig. 4
figure 4

Co-occurrence networks between the phoD- (a) and phoC-harboring (b) microbial communities based on OTU abundance (mean related abundance > 0.1%), and each connection represents a very strong (Spearman’s R > 0.6) and significant (p < 0.01) correlation. Red lines indicate significant positive correlations, and blue lines represent significant negative correlations. CK no biochar addition, SBL 0.5% straw biochar addition, SBH 1% straw biochar addition, MBL 0.5% manure biochar addition, MBH 1% manure biochar addition

Full size image

3.4 Relationship between phosphatase activities, phoD- and phoC-harboring microbial communities, and Pcoll

The structural equation model (SEM) revealed that the biochar addition decreased Pcoll through multiple pathways on phosphatase and related genes (Fig. 5). Specifically, there was a direct positive correlation between biochar addition and Pcoll in the leaching solution, but finally a negative correlation for regulating ALP and ACP and phoD and phoC-harboring microorganisms. Regarding phosphatase activity, biochar addition displayed a stronger impact on Pcoll in ALP (0.95) than in ACP (0.64). At the same time, there was a positive correlation between the abundance of functional genes, composition, diversity, and the activity of the related enzymes. Overall, biochar addition altered Pcoll in the leachate mainly through regulating phoD- and phoC-harboring microbial communities.

Fig. 5
figure 5

Structural equation model (SEM) showing the relationships between biochar amendments, phoD (a) and phoC (b) gene diversity, abundance, and composition, phosphatase activity (ALP and ACP) on colloidal phosphorus. Green arrows indicate negative effects, red arrows indicate positive effects, and dashed arrows indicate non-significant paths, which were removed in the final model. Arrow widths illustrate the strength of the standardized path coefficient, and R2 values represent the proportion of the variance explained for each endogenous variable. Significance levels are indicated by “*” (p < 0.05), “**” (p < 0.01), and “***” (p < 0.001) levels

Full size image

4 Discussion

4.1 Biochar addition inhibited the release of Pcoll under DRW cycles

The concentration of all P-fractions in the leaching solution was reduced in both tested soils during DRW cycles (Table 1). It is worth noting that the biochar addition consistently resulted in decreased Pcoll concentrations compared to the control. This may be related to the physical behavior of P adsorption and increased organic carbon by biochar (Li et al. 2022). Although in this study we focused on the effect of biochar on Pcoll mineralization from the view of the microbial enzymatic hydrolysis, we could not ignore that direct biochar itself directly adsorbed P, and indirectly affected the adsorption of Pcoll by regulating soil characteristics. On the one hand, the porous structure of biochar surface can adsorb Pcoll and provide suitable habitats for microorganisms to improve their environmental resistance, and avoid releasing colloidal-bond P from dead microbial cell lysis (Palansooriya et al. 2019). On the other hand, biochar addition increases soil pH, thus increasing Pcoll adsorption efficiency. This may be because the protonated hydroxyl (–OH) groups on the biochar surface were negatively charged, and the Pcoll was desirable for the adsorption by electrostatic attraction (Yang et al. 2019). Moreover, several studies suggested the P released during the DRW events could also be attributed to the changes in soil pH, as Fe/Al oxides affect P adsorption capacity, which is pH-sensitive (Hens and Merckx 2001; Turner et al. 2003). In our study, the biochar showed satisfactory adsorption properties for Pcoll under weakly alkaline conditions, and a reasonable explanation for this phenomenon is that chemisorption through the formation of inner-sphere complexation was the dominant adsorption process (Yan et al. 2022).

We also compared the surface functional groups of the two types of biochar by FTIR spectroscopy (Additional file 1: Fig. S2). Specifically, we observed abundant surface functional groups, such as –OH stretching vibration at around 3390 cm−1 belonging to hydroxyl groups; the aromatic skeletal C=C vibrations at 1642 and 1420 cm−1; the M–O and O–M–O (M=Mg, Al) vibration weak appearing in the 456–820 cm−1 region; and C=O stretching vibration at 1090 cm−1 mainly associated with carboxyl, trace of aldehydes, and ketones (Feng et al. 2023; Jiang et al. 2020; Yang et al. 2023). And it should be noted that the only difference between straw biochar and manure biochar was the peak at 2823 cm−1, which is usually considered as –CH2, assigning to nonpolar aliphatic fractions in biopolymer (Han et al. 2017). Therefore, we believe that organic carbon in manure biochar is more easily metabolized by microorganisms, which is one of the main reasons for attracting more phosphatase functional microorganisms. As reported in our previous research, Cmic and Pmic were considered as indicators to link soil properties and microbial communities, and biochar addition increased Cmic and Pmic, suggesting that the overall soil microbial abundance improved, which led part of the Pcoll to be absorbed and utilized by microorganisms (Wang et al. 2023). Overall, the inhibition mechanism of biochar on Pcoll loss potential is a comprehensive result, including its direct adsorption effect and indirect regulation of soil properties and phosphatase function microbial community. Field-scale experiments also confirmed this result, and the mechanism is mainly based on physical sorption (Wang et al. 2021a, 2021b, 2021c).

4.2 Biochar addition regulated phosphatase activities and phoD- and phoC-harboring microbial communities

At the same time, numerous studies have confirmed that phosphatase activities were closely associated with P forms and composition (Luo et al. 2019; Tian et al. 2021; Wei et al. 2019). Our results further suggested that Pcoll may be present as organic P and can be affected by phosphatase. It has also been shown that colloidal organic P can be stable with Fe (III) in soil solution and consequently contribute to plant acquisition; our results provide evidence for the bioavailability of Pcoll from the enzymatic perspective (Bollyn et al. 2017). Interestingly, ACP activity was positively correlated with Pcoll in weakly alkaline soil, which could be because some ACPs were stable in the soil colloids (Gao and DeLuca 2018). This phenomenon reflects that the alkaline soil with biochar may reduce the stability and persistence of ACP (Renella et al. 2006). Nevertheless, the structural equation model showed that Pcoll was ultimately negatively correlated with soil phosphatase activity under biochar addition. This may be because ALP played a key role in the Pcoll content rather than ACP. On the other hand, we found that both ALP and ACP activities were affected by the initial soil pH value, and the ACP activity was sensitive to biochar addition in acidic soil (Tian et al. 2021).

We examined the phoD- and phoC-encoding microbial communities to measure the ALP and ACP activity. And the Proteobacteria and Cyanobacteria were identified with potential ACP and ALP activities in both phoD- and phoC-harboring species (Luo et al. 2019). This is because phosphatase genes could exist in multiple copies, and different species could also produce similar genetic sequences and then encode phosphatase (Gaiero et al. 2018). Therefore, it is not only phoD- and phoC gene abundance but also the community composition encoded by the genes that cooperatively determine phosphatase activity (Tian et al. 2021). The ecological interactions between the different microorganisms drive phosphatase production; and contribute to P mobilization (Bergkemper et al. 2016). Furthermore, keystone species, such as Proteobacteria, play a key role in both phoD- and phoC-harboring microbial communities, which was performed by network analysis (Wei et al. 2019). The phoD-harboring species exhibited a higher relative abundance and had more complex connections (e.g. nodes, edges, and graph density) than the phoC-harboring species. This is probably due to the phoD-coding gene having a richer species database (Fraser et al. 2017). In contrast, biochar amendment increased the network density of phoD- and phoC-harboring communities (Ma et al. 2020; Wang et al. 2023). Bello et al. (2020) reported a similar result that straw biochar resulted in a more complex connection than in soil without biochar. This could be due to the fact that biochar containing labile carbon substrates promotes the growth of microbial communities (Chigineva et al. 2009; Spohn and Kuzyakov 2013). Also, we observed that the density of the co-occurrence network varied under different biochar addition, which may be related to the feedstocks (Palansooriya et al. 2019). Similarly, Wang et al. (2023) reported that manure biochar was more effective than straw biochar in promoting competition and cooperation in the co-occurrence network. This phenomenon could be explained by two possible reasons. On the one hand, compared to straw biochar, manure biochar contains more labile carbon, which can be considered a source of microbial nutrients. In contrast, large amounts of labile carbon require more microbial cooperation to be metabolized (Chen et al. 2019a, 2019b). At the same time, the higher abundance of microbial communities, in turn, caused “P starvation” which reflects the actual potential supply of phosphate for microorganisms (Bergkemper et al. 2016). Due to P limitations, the relative abundance of phosphatase functional microbes increases and enhances both competition and cooperation among their microbial communities to mineralize more Pcoll to obtain available P for survival (Heuck et al. 2015). Many existing studies have observed that the abundance of soil nutrients such as carbon and phosphorus is a key factor causing competition and cooperation among microorganisms, which supports our view (Heuck et al. 2015; Luo et al. 2019; Wei et al. 2019).

4.3 Potential interactions between Pcoll and phoD- and phoC-harboring microbial communities

According to the structural equation model, the Pcoll can be directly affected by biochar amendments, and regulated by both phoD- and phoC-harboring microbial communities (Fig. 5). A positive relationship has been found between biochar and Pcoll in the leachate, as it contains colloidal particles itself (Yang et al. 2020). However, the microbial indicators, particularly for ALP and ACP activities, displayed a dominant role in regulating Pcoll in leachate. Our results indicate the expression of ALP and ACP by key species such as Proteobacteria can contribute to Pcoll content (Luo et al. 2019; Wang et al. 2023; Wei et al. 2019). Although we did not specifically track the change in pH, it is reasonable to assume that biochar addition would increase the value of pH and thus the Pcoll content. Recent studies showed that soil pH is a key environmental factor for microorganisms and determines the functional microbial community composition (Bi et al. 2020; Ragot et al. 2016). In fact, changes in soil carbon were key environmental predictors on the potential influence on both nutrient availability as well as certain microbial taxa, abundance, and functional population composition (Aanderud et al. 2018). Microbial phosphatases are involved in organic P mineralization to promote the availability of P and C from organo-phosphorylated compounds (Heuck et al. 2015). Thus, phosphatase activities can be driven by the microbial need for carbon. The phoD- and phoC-harboring microbial populations directly mediate the relationship between microbial communities and ACP/ALP activities. These interactive effects may explain why the present study identified that phoD- and phoC-harboring populations were increased by the amendment of biochar. In general, our findings imply that biochar addition increased the phoD- and phoC-harboring communities, considerably reducing Pcoll in the leaching solution.

5 Conclusions

In this study, we evaluated the effects of different biochar addition on Pcoll released into leachate from the perspective of ALP/ACP activities and phoD- and phoC-harboring microbial communities during DRW events. Our results demonstrated that the biochar addition significantly decreased Pcoll content during three DRW cycles. We also observed Proteobacteria and Cyanobacteria were the key taxa in phoD- and phoC-harboring microbial communities, in which, manure biochar performed more competition and cooperation than straw biochar. More importantly, we revealed that ALP/ACP activities were the key factors in inhibiting the release of Pcoll under biochar addition through an established structural equation model. Our results provide a microbial mechanism for biochar controlling the migration and loss of Pcoll during DRW events, which has great environmental and bioecological significance.