The presence of soybean, but not soybean cropping frequency has influence on SOM priming in crop rotation systems

Legume crops are advocated for integration into crop rotation systems, and cereal-based rotations with the presence of legumes have a substantial effect on improving soil fertility and health. It is not yet clear whether the frequency of legume inclusion in crop rotation systems influences soil biochemical properties and soil organic matter (SOM) mineralization. An incubation experiment was conducted with 13C-glucose addition to evaluate the influences of soybean (Glycine max L.) cropping frequency on SOM mineralization under long-term wheat (Triticum aestivum L.)- and maize (Zea mays L.)-based rotation systems. Phospholipid fatty acids (PLFAs) and 13C-PLFAs were measured to explore microbial biomass, community structure and microbial utilization of glucose in wheat and maize systems. Glucose addition increased native SOM mineralization, i.e. positive priming effect. Compared with less soybean cropping frequency under long-term wheat- and maize-based rotation systems, wheat-soybean-soybean-soybean rotation and maize-soybean-soybean rotation increased the total biomass (PLFAs), fungal biomass and decreased the ratio of bacteria to fungi. Furthermore, the ratio of bacteria to fungi was negatively correlated with PE intensity, indicating that greater fungal biomass played a key role in stimulating SOM priming. That the proportion of 13C-glucose in G- and fungi had a positive relationship with PE intensity also supported this conclusion. The presence of soybean in wheat- and maize-based rotations increased SOM priming, while the soybean cropping frequency had no significant influence on SOM priming. However, in contrast to a maize-based rotation system, the same frequency of soybean in a wheat-based rotation system had lower soil C/N ratio and higher B/F ratio, and resulted in lower PE intensity. Our findings indicated that the presence of soybean in wheat- and maize-based rotation systems increased PE intensity because of higher soil C/N ratio and lower B/F ratio, while the soybean cropping frequency had no significant influence on SOM priming. Furthermore, the presence of soybean in maize system induced more SOM priming than that in wheat system with glucose addition.


Introduction
Compared with continuous cropping systems, crop rotation has been advocated in agriculture for millennia due to its advantages for decreasing the risk of soil degradation, pests and disease, further maintaining soil quality and crop productivity (Ai et al. 2018;Oliveira et al. 2019;Tiemann et al. 2015). Integrating legumes into crop rotation systems has a positive effect on soil physical, chemical and biological properties which are ascribed to their biological nitrogen (N) fixation and lower ratio of carbon (C) to N litter (Ai et al. 2018;Murugan and Kumar 2013;Oliveira et al. 2019). This is also a sustainable way to improve the soil quality, such as promoting the formation and stabilization of soil aggregates, increasing soil organic C (SOC) content and nutrient uptake, and then to maintain crop productivity (Ai et al. 2018;Khakbazan et al. 2019;Novelli et al. 2011;Tiemann et al. 2015). Moreover, the legume crops with a low ratio of C:N can be more easily utilized by microorganisms, and then induce soil organic matter (SOM) mineralization and the magnitude of priming effects (PE). Therefore, further understanding the relationship between legume cropping frequency and SOM priming is of benefit to evaluate the contribution of legume systems for SOC storage.
As a sensitive indicator of soil biochemical properties, soil microbial communities could be altered by the varied available C and N in different crop rotation systems, and further have a substantial effect on SOM mineralization. The 13 C labelled phospholipid fatty acids (PLFAs) represented the specific microbial groups involved the utilization of 13 C labelled substrates (Bore et al. 2017;Novelli et al. 2011). Therefore, it is meaningful to determine soil microbial communities and 13 C-PLFAs for SOM priming in different crop rotation systems. The effect of soybean (Glycine max L.) cropping frequency in different crop rotations on soil biochemical properties and SOC mineralization remains unclear. An incubation experiment with 13 C-labelled glucose addition was carried out to study the differences in SOM priming and soil biochemical properties in the long-term wheat-and maize-based rotation systems with different soybean cropping frequency. Legumes in crop rotation could increase N availability, and then increase soil microbial biomass and improve soil microbial communities (Ai et al. 2018;Oliveira et al. 2019). Therefore, we hypothesized that SOM mineralization would be higher with higher soybean cropping frequency in wheat-and maize-based rotation systems by influencing the microbial biomass and communities. Furthermore, the utilization preference for 13 C-glucose in different microbial groups also affected the SOM priming under long-term wheat-and maize-based rotation systems.

Study site and soil sampling
Soil samples (0-20 cm) were collected from the Longterm Crop Rotation Experiment in Hailun National Observation and Research Station of Agroecosystems of the Chinese Academy of Sciences (47°26′N, 126°38′E). The site has a typical temperate continental monsoon climate, with mean annual precipitation of 550 mm and mean annual temperature of 2.4 °C.
The field experiment was established in since 1991, consisting of eight treatments with three replicates: wheat (Triticum aestivum L.) monoculture, WS0; wheat-soybean rotation, WS1; wheat-soybean-soybean rotation, WS2; wheat-soybean-soybean-soybean rotation, WS3; maize (Zea mays L.) monoculture, MS0; maize-soybean rotation, MS1; maize-soybeansoybean rotation, MS2; soybean monoculture, S. The fertilizers were applied at the following rates: 112.5 kg N ha −1 as urea, 19.6 kg P ha −1 as superphosphate, and 49.8 kg K ha −1 as K 2 SO 4 for maize and wheat; and 13.5 kg N ha −1 , 15.1 kg P ha −1 and 49.8 kg K ha −1 for soybean. In 2018, ten randomized soil cores per plot were taken and mixed to form a composite sample after harvesting (October). Soil properties and the crop grown in each rotation are presented in Table 1. Soil organic C and total N (TN) were measured with an Elemental analyzer (Vario EL III, Elementar, Hanau, Germany). The δ 13 C values of soils and glucose were measured with an isotope ratio mass spectrometer (IRMS) (Mat 253, Thermo Fisher, Bremen, Germany).

Laboratory incubation
The 13 C-labelled glucose (99.0 atom%, American CIL) was diluted by unlabelled glucose to get the glucose solution for the experiment, and the glucose solution used in the study comprised 13 C-labelled glucose and unlabelled glucose, with a ratio of 1:20. The δ 13 C value of glucose solution used in this study was 4581.28‰. Before incubation, soil samples were preincubated at 40% water holding capacity (WHC) and 22 °C in dark for 7 days to recover microbial activity and function. Eight soil samples with and without glucose were incubated at 60% WHC and 22 °C in dark for 21 days to explore SOM mineralization under different rotation systems. All treatments had three replicates. Glucose solution was added into soil samples (30 g dry weight) at a rate of 0.5 g C kg −1 soil in a 120-ml plastic cup, and then placed in a 750ml mason jar with 10 ml distilled water. Three jars without soil and glucose addition were used to determine the background CO 2 concentration and natural 13 C abundance of air, which was used to remove the effect of 13 C signal from air CO 2 in sample jars . To get CO 2 mineralization and SOM priming of each treatment, soil gas was sampled on 6 h, and then again on days 1, 2, 3, 5, 7, 10, 14, and 21. CO 2 sampling was conducted by using the method of Dai et al. (2022). Briefly, all jars were ventilated with air for 10 min and then closed with air-tight lids to sample soil gas (20 ml) immediately to get the background data of CO 2 at the beginning of incubation. Soil gas was sampled again at 6 h, 1, 2, 3, 5, 7, 10, 14 and 21 days to determine the CO 2 concentrations and δ 13 C abundance after the beginning of incubation. After each sampling, all jars were ventilated with air for 10 min and then closed with airtight lids and returned to the incubation conditions described above until the gas was sampled again. The CO 2 concentration was measured with a gas chromatograph (GC2010, Shimadzu, Kyoto, Japan) and δ 13 C of CO 2 was determined by an IRMS (Delta V Advantage, Thermo Fisher, Bremen, Germany).

Soil PLFAs and 13 C-PLFA analysis
At the end of incubation, PLFAs were determined to express the microbial community structure and the PLFAs extraction was conducted according to Bossio et al. (1998). The phospholipids were separated and derivatized to their fatty acid methyl esters (FAMEs) by a gas chromatography (GC). The 13 C/ 12 C ratios of individual PLFAs were analyzed by GC-C-IRMS using a Trace GC Ultra gas chromatograph with combustion column, attached via a GC combustion III to an isotope ratio mass spectrometer (IRMS) (Mat 253, Thermo Fisher, Bremen, Germany).

Calculations and statistical analyses
The PE and PE intensity were calculated as follows (Li et al. 2018): where CO 2 , control is the average value of CO 2 produced in the soil without 13 C-labelled glucose (n = 3). The calculation of CO 2 production derived from native SOM (CO 2 , SOM ) and glucose (CO 2 , glucose ) was as follows: where CO 2 , total represents the total CO 2 produced from the soil with 13 C-labelled glucose.
The proportion of CO 2 produced by 13 C-labelled glucose (f glucose ) was calculated as follows: where δ 13 CO 2 ‰, glucose is the δ 13 C‰ value of CO 2 produced in the soil with 13 C-labelled glucose. The δ 13 CO 2 ‰, control is the average value of δ 13 C‰ in CO 2 produced in the soil without 13 C-labelled glucose (n = 3). The δ 13 C‰, glucose and δ 13 C‰, soil are the δ 13 C‰ value of 13 C-labelled glucose and the original soil, respectively.
The amount of the microbial utilization of 13 C-glucose and SOC were calculated as follows: where PLFA glucose and PLFA soc are the amount of microbial utilization of 13 C-glucose and SOC, respectively. The PLFA, f glucose is the proportion of PLFA produced by 13 C-labelled glucose. The δ 13 C‰, PLFA-glu and δ 13 C‰, PLFA-control are the δ 13 C‰, PLFA in treatments with and without glucose addition, respectively. PLFA PLFA-glu is the PLFAs concentration in treatments with glucose addition.
The residuals from the analyses of variance were examined for normality and homoscedasticity before analysis of variance (ANOVA), which was carried out by R 3.4.3 (R Development Core Team 2017). Nonparametric tests (Kruskal-Wallis test and Scheirer-Ray-Hare test) were applied when no transformation would satisfy the assumptions. The least significant difference (LSD) test was used to compare the means to estimate the significant difference. To show the effects of crop rotation and glucose addition on the cumulative CO 2 , SOM mineralization and PLFAs, we conducted a two-way analysis of variance (ANOVA). A one-way ANOVA was used to determine the effects of crop rotation on SOM priming, PE intensity, glucose decomposition and the percentage of individual 13 C-PLFAs derived from glucose. Repeated measures ANOVA was conducted to analyze the dynamics of cumulative CO 2 and soil priming in soils with different crop rotations. The principal component analysis (PCA) was conducted to analyze the relationships among soil and microbial properties, glucose decomposition and PE intensity. In addition, the relationships between soil properties, microbial properties and PE intensity, glucose decomposition were also showed by Pearson correlation.

Results and discussion
The effect of soybean cropping frequency on soil organic matter Glucose addition significantly increased CO 2 production by 195-230% and 255-300% in wheat-and maize-based rotation systems, respectively ( Fig. 1a and Table S4, p < 0.05). In wheat-and maizebased rotation systems, glucose addition promoted native SOM mineralization and resulted in positive PE ( Fig. 1a and Table S5, p < 0.05). The glucoseinduced PE in wheat-or maize-based rotation system with soybean was 27-38% and 22-26% higher than that in wheat and maize monoculture, respectively ( Fig. 1a and Table S5, p < 0.05). Priming effect intensity was larger under wheat-and maize-based rotation systems in the presence of soybean than that in wheat and maize monoculture, respectively ( Fig. 1b and Table S5, p < 0.05). However, the frequency of soybean cropping in wheat-and maize-based rotation systems had no significant influence on the PE (Fig. 1a, p > 0.05). In addition, glucose decomposition in continuous wheat cropping was 4-12% higher than all other wheat-and maize-based rotation systems ( Fig. 1a and Table S5, p < 0.05). The SOM mineralization was the highest in wheat monoculture without glucose addition, followed by soybean and the lowest in maize (Fig. 1a and Table S4, p < 0.05). During the incubation, crop rotation, glucose addition, incubation time and their interactions affected CO 2 production ( Fig. S1 and Table S2, p < 0.05), and crop rotation, incubation time and their interactions affected PE (Fig. S2 and Table S3, p < 0.05).
In general, glucose addition provided C source for soil microbial growth and resulted in positive priming effect Fig. 1 The CO 2 derived from SOM, PE and glucose in (a) soils of wheat-and maize-based rotation systems, and priming effect intensity in (b) soils of wheat-and maize-based rotation systems at day 21. Wheat monoculture, WS0; wheat-soybean rotation, WS1; wheat-soybean-soybean rotation, WS2; wheatsoybean-soybean-soybean rotation, WS3; soybean monoculture, S; maize monoculture, MS0; maizesoybean rotation, MS1; maize-soybean-soybean rotation, MS2. Vertical bars denote the standard error of the mean (n = 3). Different lowercase letters showed significant differences in PE among different treatments (p < 0.05), and different capital letters indicated significant differences in glucose decomposition among different treatments (p < 0.05). Different letters in Fig. 1b represent significant differences in priming effect intensity (p < 0.05) (de Graaff et al. 2014;Fontaine et al. 2011;SalomÃ et al. 2010;Sun et al. 2014;Zhang et al. 2018). The increased soil microbial biomass preferred to utilize soil available C and promoted native SOM mineralization ( Fig. 2 and Table S7, p < 0.05). In addition, soil PE intensity was positively correlated with the C content ( Fig. 3 and Table S1, p < 0.05), as soils with higher C content usually provide more available C for soil microorganism. The higher glucose decomposition in soil of wheat monoculture than other rotation systems was likely due to its more labile nutrient-rich soil (Zhang and Wang 2012), confirmed by the lower C/N ratio in continuous wheat system (Tables 1 and S1, p < 0.05). Qiao et al. (2015) found that SOC turnover was the fastest under continuous wheat, followed by soybean and the slowest under maize, which is consistent with the present study (Fig. 1a, p < 0.05). The lower C:N ratio of wheat and soybean root, compared with maize root, probably contributed to their faster SOC turnover (Qiao et al. 2015). However, soil PE intensity was the greatest under continuous maize with glucose addition, followed by soybean and the least under wheat (Fig. 1b, p < 0.05). Compared to continuous wheat and soybean, higher soil C/N ratio under continuous maize indicated higher SOM recalcitrance, which could trigger greater microbial N demand and then induced more PE intensity (Ai et al. 2018;Chen et al. 2019;Waring et al. 2013). Therefore, regardless of the presence of soybean, glucose-induced PE intensity in maize-based rotation systems with higher soil C/N ratio was higher than that in wheat-based rotation systems (Fig. 1b, Tables 1 and S1, p < 0.05). In the present study, higher C/N ratio of soils with the soybean treatments was mainly attributed to the larger increase of SOC content than N content (Table 1), which was supported by the fact that legume plants can promote C storage through protecting SOC from microbial decomposition by increasing the formation and stabilization of soil aggregates (Oliveira et al. 2019). Furthermore, the higher C/N ratio of soils with the soybean treatments also promoted higher PE intensity (Fig. 1b, p < 0.05).

Phospholipid fatty acids affected by soybean cropping frequency
Microbial community structure was studied by determining soil PLFAs and 13 C-glucose incorporated into Fig. 2 Bacteria (a), fungi (b), B/F ratio (c) and total PLFAs (d) in soils with or without glucose addition of wheat-and maize-based rotation systems at day 21. Wheat monoculture, WS0; wheat-soybean rotation, WS1; wheat-soybean-soybean rotation, WS2; wheat-soybean-soybean-soybean rotation, WS3; soybean monoculture, S; maize monoculture, MS0; maize-soybean rotation, MS1; maize-soybean-soybean rotation, MS2. Vertical bars denote the standard error of the mean (n = 3). Different letters represent significant differences in bacteria, fungi, B/F ratio and total PLFAs, respectively (p < 0.05) soils of wheat-and maize-based rotation systems at day 21 (Figs. 2 and 4). The soybean cropping frequency affected microbial utilization of glucose in soils of both wheat-and maize-based rotation systems ( Fig. 4 and Table S6, p < 0.05). In monoculture system, the percentage of 13 C-glucose utilized by G + and actinomycetes in WS0 was larger than that in S and MS0, while the percentage of 13 C-glucose utilized by G-and fungi in WS0 was smaller than that in S and MS0. Furthermore, the percentage of 13 C-glucose utilized by G + in soils decreased with the increasing frequency of legume included in wheat-and maizebased rotation systems, while the tendency of G-and fungi was opposite. Regardless of glucose addition, total PLFAs in WS3 and MS2 were larger than other treatments in wheat-and maize-based rotation systems ( Fig. 2 and Table S7, p < 0.05). In addition, the B/F ratio in soils under maize system was lower than that in soils under wheat system ( Fig. 2 and Table S7, p < 0.05).
Bacteria preferentially used the added glucose in the soils of wheat-and maize-based rotation systems, especially G + . The 13 C-glucose was mostly utilized by G + in soils, accounting for 30-40% of 13 C-PLFA derived from glucose (Fig. 4b). However, fungi utilizing the added glucose was the minority of 13 C-PLFA derived from glucose compared with bacteria (Fig. 4b). This result was consistent with a previous study (Fontaine et al. 2003), which showed that the simple and soluble glucose was more easily utilized by fast growing r-strategist microorganisms (bacteria) than the slow growing K-strategist microorganisms (fungi). In the present study, higher glucose decomposition in wheat monoculture was preferentially utilized by G + (Fig. 3), and resulted in lower PE intensity (Fig. 1b, p < 0.05). It was reported that G + utilizes the labile C of added organic substrates firstly and then co-metabolize the native SOC, corresponding to the 'r-strategist' (Wang et al. 2020). Soil PE intensity in the presence of soybean included in wheat-and maize-based rotation systems was higher with larger percentage of G-and fungi utilizing 13 C-labelled glucose than that in wheat and maize monoculture, respectively ( Fig. 3 and Table S1, p < 0.05). Garcia-Pausas and Paterson (2011) had reported that fungi promoted positive SOM priming, consistent with the present study. The different proportion of G-probably contributed the reduction of the magnitude of PLFAs associated with the increasing frequency of soybean in wheat-based rotation system other than maize-based rotation system, further influence the varied PE in different rotation systems (Wang et al. 2020). Soil microbial communities played a key role in explaining the changes in Fig. 3 The principal component analysis (PCA) among soil (C, N and C/N), microbial properties (microbial biomass (bacterial biomass, fungal biomass, B/F and total PLFAs) and the percentage of individual 13 C-PLFAs (G + , G-, GB, Actinomycetes, and Fungi) derived from glucose), glucose decomposition and PE intensity (PI). Wheat monoculture, WS0; wheatsoybean rotation, WS1; wheat-soybean-soybean rotation, WS2; wheatsoybean-soybean-soybean rotation, WS3; soybean monoculture, S; maize monoculture, MS0; maizesoybean rotation, MS1; maize-soybean-soybean rotation, MS2 SOM priming (Wang et al. 2020). The ratio of bacteria to fungi was negatively correlated to PE intensity (Table S1, p < 0.05), indicating that more fungal biomass induced greater SOM priming (Fontaine et al. 2011;Garcia-Pausas and Paterson 2011;Qiao et al. 2019). It was also reported that higher SOM Fig. 4 The relative abundance of individual 13 C-PLFAs derived from glucose in (a) wheat-and maize-based rotation systems, and the percentage of individual 13 C-PLFAs derived from glucose in (b) wheat-and maize-based rotation systems. Wheat monoculture, WS0; wheat-soybean rotation, WS1; wheat-soybean-soybean rotation, WS2; wheat-soybean-soybean-soybean rotation, WS3; soy-bean monoculture, S; maize monoculture, MS0; maize-soybean rotation, MS1; maize-soybean-soybean rotation, MS2. Vertical bars denote the standard error of the mean (n = 3). Different letters represent significant differences in Gram-positive bacteria (G +), Gram-negative bacteria (G-), fungi, actinomycetes and general bacteria (GB), respectively (p < 0.05) recalcitrance (higher soil C/N ratio) favored fungi and produced greater PE intensity (Ai et al. 2018;Chen et al. 2019). Many researches have reported that legume included in crop rotation increased fungal biomass and fungi/bacteria ratio (Ai et al. 2018;Murugan and Kumar 2013), and then induced greater PE intensity, which is consistent with the result in the present study. In addition, the lower B/F ratio under maize system also promoted greater PE intensity than wheat system.

Conclusion
The SOM mineralization was the highest in continuous wheat cropping, followed by soybean and the lowest in maize. However, the PE intensity in maize system was larger than that in wheat system with the same frequency of soybean integration, which was correlated to higher soil C/N ratio and lower B/F ratio in maize system. Compared with the continuous wheat or maize, the presence of soybean included in wheat-and maize-based rotation systems enhanced PE intensity by increasing the fungal biomass and decreasing the ratio of bacteria to fungi. Furthermore, there were no significant changes in SOM priming with the increasing of soybean cropping frequency both in wheat and maize systems. In conclusion, the presence of soybean in wheat-and maize-based rotations increased soil C/N ratio, fungal biomass and decreased the ratio of bacteria to fungi, and then increased PE intensity, while soybean cropping frequency had no significant influences on SOM priming.