Soil samples were collected from a warm-temperate forest in a semi-arid region of northern China (39°58′N, 115°25′E; 1,278 m elevation) in July 2017, where a native slow-growing oak (Quercus liaotungensis) was the dominant species. Mean annual precipitation and temperature in the forest were 638.8 mm and 6.5°C, respectively, and 74% of the precipitation occurred in summer (from June to August). The soil type was classified as Lixisols (Wang et al., 2016a). Three plots (30 m × 40 m) were randomly established at the study site. After removing the surface litter, we randomly collected more than 10 soil samples from 0 to 10 cm soil depth and combined these into a composite sample for each plot. The fresh soil samples were immediately sieved through a 2 mm diameter sieve to remove rocks, roots, and visible organic debris. These were then taken and transported to the laboratory and divided into three parts: one part was stored at 4°C before conducting the incubation experiments, one part was freeze-dried to measure soil microbial phospholipid fatty acid (PLFA), while the last part was air-dried and processed to measure the soil properties.

A soil-water suspension (1:2.5 v/v) was used to measure the soil pH with a pH meter (Myron L. Company, California, United States). Soil organic carbon (SOC) and total nitrogen (TN) content were measured using an elemental analyzer (Vario EL III, Elemental Analysis System GmbH, Germany). Soil texture was analyzed using a Mastersizer-2000 laser particle analyzer (Malvern Company, Worcestershire, England), and the soil was classified into sand (250–2,000 μm), silt (50–250 μm), and clay fractions (<50 μm). Soil microbial community composition was determined using PLFA biomarker analysis to obtain fungal, bacterial, and actinomycete content (Frostegård et al., 1993; Bååth and Anderson, 2003). The soil properties and microbial characteristics are shown in Table 1.

Fresh soil (equivalent to 30 g dry soil) was weighed into a 150 mL plastic culture bottles (n = 3), and the headspace volume was approximately 130 mL after placing the soil sample. As the moisture content of fresh soil exceeded 20% WHC, we lightly air-dried the samples, treated with 10 and 20% WHC moisture, and then adjusted the moisture content. The soil samples were then pre-incubated in a 20°C constant temperature incubator for 7 days to activate the microbes and reduce the interference of the “pulse effect” of water on the decomposition of SOM (Li et al., 2022). Additionally, considering that the added LOC contained water would increase the soil moisture, the soil moisture content during pre-culture was less than the predetermined moisture content to ensure that the soil moisture reached 10, 20, 30, 40, 50, 60, 70, 80, and 90% moisture with the addition of LOC. Water loss in the sample bottles was measured daily and corrected by weight.

The automatic temperature control soil flux system (PRI-8800; PRE-ECO, Beijing, China) developed by He et al. (2013) was modified to measure the soil respiration rates. Specifically, this new device with a PICARRO isotope analyzer (G2131-i, PICARRO Inc., Sunnyvale, California, United States) enabled observations of δ¹³C and CO₂ concentrations on a minute-scale. As shown in Supplementary Figure 1, the device consists of five parts: an analysis, sampling, control, temperature control, and calibration system. Specifically, the soil sample bottles were placed in a 16-hole electric water bath controlled by an automatic temperature regulator and held at 20°C. When the instrument began testing, a collection needle controlled by the control system was automatically inserted into the sample bottle and completely covered the mouth of the bottle, creating an enclosed space. During this time, the instrument was connected to the PICARRO G2131-i isotope analyzer through a control system to record the δ¹³C and CO₂ concentrations every second. After measuring one sample, the collection needle was automatically pulled out, and the next sample was measured. During the measurements, each sample bottle was sealed with a preservative film to reduce water loss. Small holes were poked in the preservative film for ventilation, and the sampling needle was inserted through the holes to collect samples from the vials. The dynamics of the respiration rate (R) were measured at 3-min intervals for each sample, and each sample was measured 290 times over a 48-h period. In addition, after every 30 samples, the instrument was automatically calibrated using the calibration system.

We added glucose as a simple analog for LOC input. Specifically, three incubation bottles were removed from the instrument. Then, 1 mL ¹³C-uniformly labeled glucose (δ¹³C = 3,864‰) solution was added slowly and uniformly to the soil by pipetting the glucose onto the soil surface at a controlled and consistent rate. This was done to ensure uniform distribution of the solution in the soil (Li et al., 2022). The amount of LOC added to the soil was 0.4 mg C g soil, which was approximately 1% of the soil organic carbon content (3.9%). Considering the possible pulse effect of the deionized water contained in 1 mL of the glucose solution, a control treatment with 1 mL of deionized water was set up. The R-values were stable (Supplementary Figure 2), as demonstrated in our recent study (Li et al., 2022).

As the soil used in this study is acidic and probably contains low carbonate amounts, the amount of carbonate decomposition in this soil environment should be minimal, because inorganic C is relatively stable. Therefore, we assumed that the released CO₂ is derived directly from microbial respiration and did not consider the C emissions potentially caused by abiotic factors. Additionally, the fractionation during the biodegradation processes was negligible (Mary et al., 1992). The proportions of CO₂ derived from LOC and SOM components were calculated using a mass balance model:

where f_LOC is the fraction of total CO₂ derived from glucose, δ_Total is the measured isotopic value of the sample obtained at each measurement point, δ_SOM is the mean of the isotopic value of the control treatment, and δ_LOC is the isotopic value of the added glucose treatment.

As shown in Figure 1, the corresponding parameters were defined to better determine the microbial response to the LOC input. Tukey’s HSD test was used to identify differences in R_max, time for R_max, absolute change in R, and accumulation of CO₂ released among different soil moisture treatments. Curve fitting was used to explore the relationships between the soil moisture and the above parameters. Statistical analyses were performed using the SPSS software (SPSS for Windows, Version 13.0; Chicago, IL, United States). Graph plotting and curve fitting were performed using Origin software (version 8.5; Northampton, MA, United States). Differences were considered statistically significant at P < 0.05.

Soil moisture content is an important environmental factor that affects microbial activity, the diffusion rate of gasses, and the availability and mobility of soluble SOM (Casals et al., 2000; Liu et al., 2009). Therefore, changes in soil moisture content have a profound impact on soil microbial respiration and SOM mineralization. In general, aerobic microbial respiration is optimal when the soil moisture content is approximately 60% WHC (Das et al., 2019). Here, we observed that extremely high or low moisture content was not conducive to microbial utilization of both LOC and native SOM. However, when soil moisture content was 50–70% of WHC, the responses of soil microbes to added LOC were stronger than those observed under other soil moisture treatments; that is, R_max and CO₂ released were greater. This finding supports our hypothesis, and consistent with findings in related studies (Geisseler et al., 2011; Wang et al., 2016b). We found that soil microbes with lower moisture content (<30%WHC) were less responsive to added LOC (lower R_max and CO_2–LOC). The movement of the glucose solution strongly depends on the soil moisture content (Casals et al., 2000; Liu et al., 2009). During low soil moisture treatment, the downward movement of the glucose solution was much slower than that during high soil moisture treatment. Microbes and their enzymes responsible for macromolecule degradation can be affected by moisture (Šnajdr et al., 2008; Geisseler et al., 2011); low soil moisture content affects the utilization of added LOC by microbes (Xue et al., 2017). That is because the diffusion of substrate may slow down due to lack of water which limits the substrate supply to soil microbes (Schjønning et al., 2003). Additionally, as the matrix potential of soil moisture decreases, soil microbes need to consume additional energy to maintain osmotic balance with their surroundings (Schimel et al., 2007). They also need to produce extracellular substances to buffer soil moisture changes and improve diffusivity (Or et al., 2007). Together, these factors lead to a reduced metabolic activity of soil microbes and a limited rate of reproduction and respiration. The utilization of LOC by microorganisms is also reduced when the moisture content is too high (>70% WHC). High moisture content creates an anaerobic environment (Hackl et al., 2005), which inhibits microbial activity and is unfavorable for the utilization of LOC by microorganisms.

Although we found that microbes preferred the utilization of LOC over native SOM, the simultaneous effect of LOC on microbial activity further promoted the decomposition of native SOM. This decomposition was also influenced by soil moisture. Generally, microbes are limited by the availability of C substrates (Cleveland et al., 2007), and LOC inputs increase substrate availability, relieve microbial C limitations, and increase microbial activity. This in turn promotes the decomposition of native SOM (Weedon et al., 2013). The microbial nitrogen (N)-mining hypothesis can be used to explain this phenomenon (Craine et al., 2007). Although the addition of LOC increases the availability of C substrate to microbes, microbial growth and metabolism will be limited by N, when microbes need to decompose more native SOM to obtain nutrients to meet growth requirements, and thus stimulate the decomposition of native SOM (Chen et al., 2014).

Our results show an interesting phenomenon: when the soil moisture was below 60% WHC, the absolute change in R_SOM was less than 0 but that of R_LOC was greater than 0, after 1 h of LOC input (Figures 3G,H). This implies the microbial preferential utilization of LOC and an inhibition of SOM decomposition (Li et al., 2022). Generally, microbial availability of organic matter depends on its chemical recalcitrance, mineral association, and accessibility to decomposers (Deng et al., 2020). Microbes are limited by the availability of C substrates, and thus soil microbes prefer to utilize labile substrate with a low molecular composition when compared to recalcitrant SOM, as less energy is required for mineralization (Deng et al., 2020). Additionally, it is difficult for microorganisms to access and decompose the dissolved soil C due to the microbial activity is limited by the low moisture content (Curiel Yuste et al., 2007). This allows them to preferentially use LOC, thereby inhibit microbial decomposition of native SOM (Blagodatskaya and Kuzyakov, 2008). In contrast to lower soil moisture content, optimal soil moisture content allows the cleavage and dispersion of soil aggregates, which leads to the dissolution and release of SOM encapsulated within the aggregates, thereby allowing more and better access to and decomposition of native SOM by soil microbes (Marín-Spiotta et al., 2014). In contrast, when soil moisture is saturated (>70% WHC), the oxidation-reduction decreases and anaerobic conditions limit the activity of soil microbial enzymes (Hackl et al., 2005). Furthermore, any soil moisture content above 70% WHC also reduces the diffusion of O₂ and forms a flooded anaerobic environment. This environment inhibits the activity of aerobic microbes and various oxidative enzymes, thereby affecting the mineralization of C by microbes (Schjønning et al., 2003).

The proportions of LOC and native SOM utilized by microbes differed in response to different soil moisture content. This provides quantitative evidence that soil moisture content can affect how microbes utilize LOC and native SOM, through R_LOC/R_SOM, CO_2–LOC/CO_2–Total, CO_2–SOM/CO_2–Total, and CO_2–LOC/CO_2–SOM. Our study found that soils have greater CO₂ fluxes under suitable moisture conditions (approximately 60% WHC). However, this is inconsistent with our alternative hypothesis that microbes would utilize more LOC under higher moisture conditions, as we found that microbes have different strategies for utilizing LOC and native SOM under different moisture conditions. We found that the relative proportions of the metabolism of soil microbes using the two C sources to total respiration were different under different moisture conditions (Figure 5). Under lower moisture conditions (<30%WHC), more native SOM was used for microbial respiration. This may occur because low moisture conditions are not conducive to the diffusion of LOC; thus, microbes predominantly use native SOM (Davidson et al., 2010). Furthermore, the decrease in soil water potentially may stimulate microbial responses to water stress. This stimulation results in an increase in soil microbial C-use efficiency (Herron et al., 2009), which can lead to lower LOC for respiration. However, as soil water content increased, more microbes had access to LOC, resulting in more LOC for respiration. Thus, values of CO_2–LOC/CO_2–Total increased and those of CO_2–SOM/CO_2–Total decreased. Moreover, when the soil moisture content was too high (>60% WHC), the proportion of native SOM utilization by microbes gradually increased again (Figure 5B). This increase occurs because moisture content affects the microbial C use efficiency. Many studies have found that the native SOM mineralization rate is highest under higher moisture content conditions after LOC input, which is consistent with our results (Guenet et al., 2013; Miao et al., 2016). When the water content is too high, it contributes to the disruption of the soil aggregate structure and SOM leaching and increases the contact area between native SOM and decomposers, thus promoting native SOM mineralization (Marín-Spiotta et al., 2014). Moreover, it is possible that LOC input resulted in local energy excesses and nutrient deficits caused by the highly heterogeneous SOM forming many relatively independent microhabitats, Meanwhile, soil flooding resulted in a homogeneous mixing of decomposing substrates and adequate utilization of LOC, which promoted the utilization of native SOM by microbes (Guenet et al., 2013).

Here, we provide quantitative evidence that soil moisture content can influence the way microbes utilize LOC and native SOM and propose potential mechanisms of influence (Figure 6). However, microbial taxa determine the composition and functional properties of microbial communities, which organize and structure their responses to resource availability and changes in environmental conditions and play a key role in the C cycle (Banerjee et al., 2018). Theoretically, soil microbes are divided into r- and K-strategists according to their biological characteristics and utilization of resources. Among them, r-strategists (mainly bacteria) grow faster and generally use LOC, whereas K-strategists grow slowly and can use more resistant organic matter (Fontaine et al., 2003). Therefore, future studies should examine the dynamics of soil microbes. This will help us to fundamentally understand the microbial strategies for the utilization of LOC and native SOM under different moisture conditions.