Responses of Soil Organic Carbon Mineralization and Microbial Communities to Leaf Litter Addition under Different Soil Layers

Abstract

The mechanism of how soil carbon pools change when coniferous forests are converted into conifer-broadleaf mixed forests is poorly understood. In this study, the impact of additional carbon inputs on soil organic carbon mineralization and microbial communities was evaluated. In a microcosm incubation experiment, three types of ¹³C-labeled litter (Pinustabulaeformis (PT), Robiniapseudoacacia (RP), and a mixture of PT and RP (1:1, PR)) were added in to top (0–20 cm) and deep (60–80 cm) soil collected from a Chinese pine plantation. The priming effect (PE) and specific microbial groups involved in PE were studied. PT and RP addition to topsoil induced a negative PE. In deep soil, the decomposition rates of soil organic matter (SOM) after adding PT and mixture increased by 16.6% and 34.6% compared to those without litter. The addition of RP with a lower C:N ratio had a stronger negative PE than adding PT or mixture. Moreover, the PE in deep soil was more intense after all litter additions. In topsoil, the litter-derived carbon was mainly incorporated into 16:0, 18:1ω9c, and 18:1ω7c fatty acids. In conclusion, the addition of broadleaf litter into coniferous plantations might be beneficial for enhancing deep soil carbon stocks.

1. Introduction

Forest ecosystems occupy approximately one-third of the world’s land area and store approximately 383 Pg of C in their soils to a depth of 1 m [1,2]. The changes in forest soil organic carbon (SOC) pool have important impacts on global carbon cycling. In recent years, global warming or the change of land use type increased the input of fresh organic carbon to soil, which might affect soil carbon dynamics. Exogenous organic matter inputs can cause rapid changes in native soil organic matter, which is described as the ‘priming effect’(PE) [3]. A positive PE can accelerate SOC decomposition and decrease the content of SOC, while a negative PE induces the opposite [4]. Plant litter quantities and qualities are crucial factors regulating SOC decomposition. Chao et al. [5] demonstrated that litter chemistry controls the magnitude and direction of the PE. Litter hemicellulose and tannin concentrations play important roles in the early stage of decomposition. Lignin and the ratio of lignin:N showed a negative correlation with the PE in the late phase of decomposition. Increasing fresh litter addition resulted in losses of soil organic carbon in agroecosystems [6]. Hick et al. [7] revealed that the complexity of the carbon substrate had a significant impact on the soil PE by affecting microbial carbon-use efficiency in montane forest and lowland soils. The conversion of land use types, such as farmland to forest, pure forest to mixed forest, and broadleaf forest to coniferous forestalter the quantities and qualities of external plant residues and thus cause changes in SOC pools. In addition, forest litter decomposition was important for nutrient return to the soil. Coniferous litter was broken down more slowly than broad leaf litter because of higher lignin concentration and C/N ratio. In recent years, litter mixing experiments (e.g., litters characterized by low and high decomposability) were carried out in order to evaluate the mixture effects on decomposition rates and soil carbon dynamics. The synergistic [8], antagonistic [9], neutral effects [10] were observed because of complex interactions among litter chemical componentand specific soil characteristics. Especially, in arid and semi-arid region, plant species might change frequently because of water shortage, which changed litter composition and chemical component. Litter decomposition rates were essential for nutrient return to soil. There was little known about how litter species diversity affects soil carbon dynamics.

Deep soils are considered a major organic carbon pool in terrestrial ecosystems, containing over half of the total soil C [11]. Small changes in the organic carbon pool in deep soils might have an important impact on the overall soil carbon dynamics and ultimately affect the global carbon cycle [12]. It was reported that theslow decomposition rate of deep SOC is primarily attributed to alack of fresh C supply [11,13]. Fresh leaf litter addition produced a positive PE on SOC mineralization [12]. Nevertheless, some studies found that excessive labile carbon input might transfer into recalcitrant carbon and neutralize some recalcitrant SOC mineralized by priming and thus be useful for soil carbon storage [14]. Litter qualities affected difference in PE through altering microbial activity in mining recalcitrant soil organic matter held nutrients [15]. Fontaine et al. [11] reported that microorganisms might relieve the impact of global warming (e.g., increase of exogenous organic matter inputs) on deep soil mineralization because of the lack of energy. Many studies were conducted in temperate forest ecosystem [16,17]. However, there was not a uniform conclusion. In particular, in arid and semi-arid area, tree species increased the root length in order to better resist droughts, which alter theinput of fresh C to deep soils. Therefore, the response of deep soil carbon pools to fresh carbon supply after the land use conversion in this region needed to be further studied.

Soil microbial communities were reported to be closely related to soil nutrient availability and regulate the direction and intensity of PEs [18]. Soil ‘microbial N mining’ and ‘preferential substrate utilization’ theory have revealed that during the soil PE, microorganisms alter their community structure and their balance of r- and K-strategists to maximize nutrient acquisition [19]. When the soil nutrient availability is low, microbes usually use new labile carbon as energy to obtain nutrients from soil organic carbon and thus induce a positive PE. Conversely, under high nutrient availability, microbes use more fresh organic matter and induce a negative PE [20]. Current studies have reported that microbial species or taxa carry out different functions based on their characteristics during the decomposition of soil organic matter. Gram-negative bacteria preferentially utilize easily decomposed substrates [21]. Gram-positive bacteria and actinomycetes were reported to be more effective at recalcitrant soil organic matter decomposition [22]. During the decomposition of plant residues, bacteria, especially r-strategists, initially mineralized labile C entering the soil [23], speeding up the turnover of bacterial biomass and inducing a PE. In the later stage of decomposition, fungi and certain groups of bacteria (e.g., K-strategists) were the key players decomposing complex substrates (e.g., lignin and cellulose). In addition, the roles of soil microbial communities are different in different ecosystems. Gram-positive bacteria and fungi are the main regulators of rhizosphere PEs, which belong to a K-strategist classification widely observed in a few temperate grassland studies [24]. Nevertheless, there was no connection between specific microbial groups and the PEs in lowland tropical forest soils [25]. Therefore, the microbial species and their contributions involved in the PE remain to be further studied.

In the present study, we want to examine the changes in soil organic carbon pools and microbial communities in the top (0–20 cm) and deep (60–80 cm) soil layers in response to the addition of different kinds of fresh organic matter by using ¹³C-labeled leaf litter, to understand how the conversion of coniferous forest to conifer-broadleaf forest impact soil organic matter (SOM) decomposition. We hypothesized that (1) in all detected soil layers, the addition of locust (Robinia pseudoacacia, RP) leaf litter to Chinese pine (Pinus tabulaeformis, PT) plantations would induce a negative PE and accelerate soil organic carbon storage because of the high litter quality. Mixed litter addition showed the same trend through the selective decomposition of locust leaf litter. Chinese pine litter addition would result in the opposite result. (2) The magnitude and intensity of the PE would be regulated by the activities of specific microbial groups, which would vary with the fresh organic matter availability. (3) The PE would be more intense in the topsoil layer than in the deep soil layer because of the higher microbial activity.

2. Materials and Methods

2.1. Site Description and Soil Sampling

The experimental site was located in the Huai Ping forest region, Shaanxi Province, China, which belongs to the hilly gully area of the Loess Plateau (107°56′–108°200′ E, 34°29′–34°59′ N) (Figure S1). This area features a warm temperate continental monsoon climate. The average annual rainfall is 602 mm, with over 50% falling between July and September. The annual mean temperature is 10.8 °C. The soil is cinnamon soil (Hapludalfs in the USDA classification). The soil properties are described in

Table 1. Physicochemical properties of soil and litter used for incubation.

	Topsoil (0–20cm)	Deep Soil (60–80 cm)	PT	RP	PR
pH	7.63	8.03	-	-	-
TOC (g·kg−1)	11.60	3.32	648.96	484.59	533.73
TN (g·kg−1)	0.88	0.41	14.84	59.18	38.45
C:N	13.26	8.19	44.68	8.24	13.91
${\mathrm{NO}}_3^{-}$-N (mg·kg−1)	4.43	1.72	-	-	-
${\mathrm{NH}}_4^{+}$-N (mg·kg−1)	7.26	3.67	-	-	-
Celulose (%)	-	-	15.31	18.52	16.83
Lignin (%)	-	-	36.28	26.28	32.03
Celulose: Lignin	-	-	0.42	0.69	0.53
δ13C(‰)	−21.64	−21.18	652.75	1464.79	1254.85

PT: Chinese pine leaf litter; RP: locust leaf litter; PR: litter mixture of Chinese pine and locust; TOC: total organic carbon; TN: total nitrogen; C:N: the ratio of carbon to nitrogen; ${\mathrm{NO}}_3^{-}$-N: nitrate nitrogen; ${\mathrm{NH}}_4^{+}$-N: ammonium nitrogen.

. The forests in this region were pure stands planted in the 1960s, which are mainly composed of Chinese pine (PT) and locust (RP). With the development of Chinese pine plantation, greater soil degradation and more pests were found in the pure forest than in a mixed forest. In recent years, broadleaf tree species (e.g., locust) have been introduced into pure forests.

In September 2018, soil samples were collected from 40-year-old Chinese pine plantations. Three 20 m × 20 m plots were selected randomly, and three profiles (1 m width and 1 m depth) were placed in each plot. In each profile, soil samples were obtained from the topsoil (0–20 cm) and deep soil (60–80 cm) using a soil corer (5 cm in diameter). Samples of the same layers in each profile were mixed into a composite sample in each plot. In all, there were nine composite soil samples in the same layer in this study. Each soil sample was divided into two parts after manually removing the visible roots, stones, litter, and debris. One part was stored at 4 °C before the soil incubation experiment and analysis of the enzyme activities and microbial biomass began (about two weeks). The other part was air dried for analysis of the soil properties.

2.2. Litter Properties

The litter samples were oven-dried at 65 °C until constant weight, grinded and passed through a 0.25 mm sieve before subsequent analysis. The litter organic carbon content was measured via the oil-bathK₂Cr₂O₇ titration method [26]. Litter total nitrogen content was determined by the Kjeldahl method after heat digestion with 98% H₂SO₄ [27]. Litter lignin content was measured through the acid detergent digestion method and then calculated based on the weight differences [28]. The cellulose content was determined by the anthronecolorimetric method after acid detergent digestion (1:8:2 (v/v/v) of nitric acid: acetic acid: water) [29].

2.3. Soil Incubation and CO₂ Analysis

The incubation experiment consisted of eight treatments with three replications (based on the above replicate forest plots): (1) topsoil with Chinese pine leaf litter (PTT); (2) topsoil with labeled locust leaf litter (RPT); (3) topsoil with litter mixtures of Chinese pine and locust (PRT); (4) deep soil with Chinese pine leaf litter (PTD); (5) deep soil with locust leaf litter (RPD); (6) deep soil with litter mixtures of Chinese pine and locust (PRD); (7) topsoil without litter; and (8) deep soil without litter. The leaf litters used in the incubation experiment were ¹³Clabeled via pulse labeling with ¹³C-enriched CO₂ over six months in a growth chamber as described by Lyu et al. [30]. The physicochemical properties of the soil and litter are presented in Table 1.

Fifty grams of fresh soil (oven-dry equivalent, sieved to 2 mm) was added into a 500 mL Mason jar and preincubated for seven days at 20 °C with soil moisture at 50% of the field water-holding capacity to allow microbes to adapt to the environment. After preincubation, 1 g of ¹³C-labeled litter (oven dried at 60 °C for 48 h and grounded to pass a 1 mm sieve) was mixed with soil uniformly.According to the average field capacity, the soil moisture was adjusted to 60% of the maximum water-holding capacity with deionized water and remained at this level during the whole incubation period. The alkali absorption method was used to trap CO₂ evolved from the soil. First, one glass vial containing 20 mL of 0.5 M NaOH solution and another vial containing 10 mL of distilled water were placed in each Mason jar. Then, the jars were sealed with airtight caps. Three Mason jars with only NaOH solution and distilled water added were used as controls to trap CO₂ from the air in each jar. The CO₂ trapped by NaOH solution was measured after 1, 3, 7, 14, 22, 30, 42, and 56 days by titration with 0.25 M HCl after precipitating all carbonates with 5 mL of 0.5 M BaCl₂ [6]. The solution was filtered and heated at 50 °C for 24 h to obtain the BaCO₃ precipitate. The δ¹³C abundance of the trapped CO₂ was detected by isotope ratio mass spectrometry (Finnigan MAT-253, Thermo Electron Inc., Bremen, Gemany).

2.4. PLFA and ¹³C PLFA Analysis

The soil microbial communities were analyzed using phospholipid fatty acids (PLFAs) at the end of incubation. PLFA analysis was performed according to the methods described by White et al. [31] with modification. Eight grams of dry-weight equivalent fresh soil was extracted using a solution of chloroform, methanol, and citrate buffer in a ratio of 1.0:2.0:0.8 (v/v/v). Phospholipids were separated on a silicic acid column (Waters Inc., Milford, MA, USA) and then methylated to fatty acid methyl esters (FAMEs). The FAMEs were quantitatively analyzed by gas chromatography (Agilent 6850N, Agilent Technologies Inc., Santa Clara, CA, USA). The amount of each PLFA was calculated according to the internal standard (19:0). The 15:0i, 15:0a, 16:0i, 16:1ω7c, 16:0, 17:0i, 17:0a, 18:1ω5c, and 18:0 PLFAs were considered bacterial markers [32]. The sums of 14:0i, 15:0i, 15:0a, 16:0i, 17:0i, and 17:0a PLFAs and 14:1ω5c, 15:1ω6c, 16:1ω7c, cy17:0, cy19:0, 15:03OH, 16:12OH, 16:1ω9c, 18:1ω5c, 18:1ω7c and cy18:0 PLFAs represented gram-positive bacteria and gram-negative bacteria, respectively [33]. The 18:2ω6c, 18:3ω6c, 18:1ω9c, and 20:1ω9c PLFAs were used to represent fungi-derived fatty acids [34]. The 16:1ω5c fatty acid is used to indicate arbuscular mycorrhizal fungi (AMF) [35]. The concentrations and ¹³C abundances of the FAMEs were analyzed by isotope ratio mass spectrometry [36]. The proportion of C (Pi) from labeled litter in each PLFA was calculated by a mass balance approach [12]:

P_i = (δ¹³C_t − δ¹³C_c)/(δ¹³C_l − δ¹³C_c)

(1)

where δ¹³C_t and δ¹³C_c represent the δ¹³C enrichment of individual PLFAs in the soil samples with and without litter, respectively. δ¹³C_l is the δ¹³C enrichment of labeled CO₂.

2.5. Analyses of Other Soil Properties

Soil organic carbon and total nitrogen were detected by the potassium dichromate oxidation method and Kjeldahl digestion method, respectively [37,38]. The content of soil labile organic carbon (LOC) was assayed with 333 mM potassium permanganate (KMnO₄) solution followed by spectrophotometry [39]. Soil recalcitrant carbon was determined by acid hydrolysis [40]. The carbon lability was expressed as the ratio of labile carbon content to non-labile carbon content [41], which was useful for evaluating the stability of soil organic carbon. The higher the ratio, the less stable the soil carbon pool. Soil microbial biomass carbon was measured by the chloroform fumigation-extraction method following Wu et al. [42].

The activities of β-glucosidase, cellulase and alkaline phosphatase were evaluated using the methods of Guan [43]. For each sample, 5 g of soil was used to determine the alkaline phosphatase activity. One gram of soil was used to determine the β-glucosidase (GLU) and cellulase activities. The substrates were p-nitrophenyl-β-D-glucoside, disodium phenylphosphate and carboxymethyl cellulose. Controls contained distilled water instead of substrates. All three enzyme activities were detected spectrophotometrically at 400, 540, and 510 nm, respectively and expressed based on the soil dry weight. The activity of N-acetylglucosaminidase was evaluated via fluorometric microplate techniques [44]. Briefly, 1.0 g of fresh soil was added to 125 mL of 50 mM sodium acetate buffer (pH 6.0) and homogenized with a vortex mixer. The 200 μL suspension was transferred to a 96-well black plate, and then 50 μL of substrate solution (4-MUB-N-acetyl-β-D-glucosaminide) was added. The negative control included 200 μL of buffer and 50 μL of substrate solution. The sample control included 200 μL of sample suspension and 50 μL of buffer. The quench control included 200 μL of sample suspension and 50 μL of standard. The microplate was incubated for 4 h at 25 °C in darkness. The fluorescence was detected by a fluorescence microplate reader (Infinite M200 Pro, Tecan Inc., Männedorf, Switzerland) at 460 nm.

2.6. PE Calculation

According to the method of Luo et al. [45], the PE (mg C kg⁻¹ soil) induced by litter addition was calculated as follows:

$$\mathsf{\alpha }=\frac{{\mathrm{A}}_{\mathrm{T}}^{13}{-\mathrm{A}}_C^{13}}{{\mathrm{A}}_{\mathrm{S}}^{13}{-\mathrm{A}}_{\mathrm{C}}^{13}}$$

(2)

where α is the percentage of leaf litter-derived C in respired CO₂ released from the NaOH solution. ${\mathrm{A}}_{\mathrm{T}}^{13}$ is the ¹³C abundance in the respired CO₂ from amended soil C. ${\mathrm{A}}_{\mathrm{C}}^{13}$ is the ¹³C abundance in the respired CO₂ from control soil C. A_13Sis the ¹³C abundance of litter.

Rr(mg·kg⁻¹) = αR_T

(3)

Rs = R_T − Rr

(4)

PE = Rs − R_C

(5)

R_T is the amount of accumulated CO₂ emitted from soil with added leaf litter. R_r is the amount of CO₂ released from leaf litter-derived C. Rs is the amount of CO₂ from soil in the treatment with added leaf litter. Rc is the amount of CO₂ derived from the control soil.

2.7. Statistical Analyses

Repeated measures analysis of variance (ANOVA) was used to analyze the temporal impacts of litter addition on SOM mineralization (CO₂ emissions and partitioning) and PE throughout theincubation. One-way ANOVA was conducted to analyze the impacts of litter addition on the soil microbial communities and microbial biomass carbon in each detected soil layer at the end of incubation. For all ANOVAs, the least significant difference (LSD) test was used to compare the means at a significance level of p < 0.05. The Pearson correlation analysis was used to evaluate the relationship between the ¹³C-labled fungal PLFA and soil PE (SPSS version 19.0, IBM, Armonk, NY, USA). The differences in soil microbial communities among treatments were evaluated by principal component analysis. The data of soil microbial communities was standardized by Hellinger. The permutational multivariate analysis of variance (PERMANOVA) analysis was used to evaluate the dissimilaritysignificances. Redundancy analysis (RDA) was carried out to assess the impacts of soil properties on the soil microbial community structure. Before RDA analysis, the data of soil properties and soil microbial community structure were standardized via scale and Hellinger, respectively. Anova with 999 permutations were used to test the significance. Principal component analysis and redundancy analysis were conducted via ‘prcomp’ and ‘rda’ functionin vegan package in R version 2.9 (R Foundation for Statistical Computing, Vienna, Austria), respectively.

3. Results

3.1. Soil Organic Carbon Pool

As shown in

Table 2. Effects of litter addition on soil organic carbon pools at the end of incubation.

Treatments	TOC (g·kg−1)	LOC (g·kg−1)	RC (g·kg−1)	Carbon Lability
CK-T	10.86 ± 0.20b	5.21 ± 0.28a	5.65 ± 0.43b	0.93 ± 0.12b
PTT	12.01 ± 0.20a	5.52 ± 0.16a	6.49 ± 0.31a	0.85 ± 0.06b
RPT	12.19 ± 0.40a	5.35 ± 0.33a	6.84 ± 0.74a	0.79 ± 0.13b
PRT	12.15 ± 0.33a	5.63 ± 0.18a	6.51 ± 0.49a	0.87 ± 0.09b
CK-D	3.54 ± 0.10e	2.34 ± 0.05d	1.20 ± 0.07d	1.96 ± 0.11b
PTD	4.53 ± 0.17d	3.74 ± 0.26b	0.79 ± 0.43d	6.18 ± 3.90a
RPD	5.35 ± 0.04c	3.15 ± 0.37c	2.20 ± 0.40c	1.48 ± 0.42b
PRD	5.10 ± 0.07c	3.15 ± 0.13c	1.95 ± 0.17c	1.63 ± 0.20b

Dates are the means ± standard deviations. Different lowercase letters within the same column indicate significant differences among treatments at p < 0.05 level. PTT: topsoil with Chinese pine leaf litter; RPT: topsoil with locust leaf litter; PRT: topsoil with litter mixtures of Chinese pine and locust; CK-T: topsoil without litter; PTD: deep soil with Chinese pine leaf litter; RPD: deep soil with locust leaf litter; PRD: deep soil with litter mixtures of Chinese pine and locust; CK-D: deep soil without litter; LOC: labile organic carbon; RC: recalcitrant carbon.

, leaf litter addition had no influence on the soil LOC in the topsoil but significantly increased the labile and recalcitrant carbon in the deep soil. In the topsoil, the amount of recalcitrant carbon in the locust, Chinese pine, and mixed litter treatments was 14.87%, 21.06%, and 15.22% higher, respectively, than that in the control soil. In the deep soil, the highest values of soil labile and recalcitrant carbon were observed in the Chinese pine and locust litter treatments, respectively. In addition, there were no significant differences in the carbon lability among all treatments in the topsoil. However, in the deep soil, the carbon lability after Chinese pine litter addition was significantly higher than that in the other treatments.

3.2. Soil Organic Matter Mineralization

Litter addition significantly increased the cumulative CO₂ flux in the top and deep soil, and the highest values appeared with locust litter addition (Figure 1). In the topsoil, the total cumulative CO₂ flux in the locust litter addition treatment increased by 55.47%, 19.05%, and 352.04% compared with the fluxes in the Chinese pine litter addition, mixture litter addition, and no-addition treatments, respectively (Figure 1a). In deep soil, the total cumulative CO₂ flux in the locust litter addition treatment increased by 132.48%, 31.64%, and 936.75% compared with the fluxes in the Chinese pine litter addition, mixture litter addition, and no-addition treatments, respectively (Figure 1b). At the end of incubation, the total cumulative CO₂ flux emitted by topsoil was significantly higher than that emitted by deep soil in all treatments. However, increase in the total cumulative CO₂ flux in topsoil (190.76%–352.04%) was significantly lower than that in deep soil (345.94%–936.75%), which indicated that the impact of litter addition on soil carbon mineralization was more obvious in deep soil than in topsoil.

3.3. Soil- and Litter-Derived CO₂-C and the PE

During 56 days of incubation, the released CO₂-C was mainly derived from litter (Figure 2). Leaf litter type was an important factor affecting the CO₂-C distribution. In the topsoil, during 14 days of incubation, locust and mixed leaf litter addition significantly increased the release of litter-derived CO₂-C (Figure 2b,c). Then, the amount of mineralized C mainly derived from locust and mixed leaf litter decomposition decreased. These results indicated that the two kinds of leaf litter gradually induced native SOC-derived CO₂ production during days 15–56. However, the amount of CO₂ derived from the addition of Chinese pine leaf litter peaked on days 42–56 (Figure 2a). In the deep soil, the release of litter-derived CO₂-C after locust and mixed litter addition showed a similar trend with those in the topsoil (Figure 2e,f). In the Chinese pine and mixed litter treatment, the total amount of CO₂ was mainly derived from litter during days 1–7 of incubation. Subsequently, CO₂ emissions derived from native SOC gradually increased (Figure 2d).

All litter addition treatments induced a PE in the measured soil layers. However, the direction and extent of the PE depended on the litter characteristics (as shown in Table 1) and soil layer. In the topsoil layer, all treatments showed negative PEs during the initial stage of 14 days. Then, positive effects were observed in the locust and mixed litter addition treatments. The extent was higher in the mixed litter addition treatment and stimulated the decomposition of native soil organic matter more intensively. However, at the end of incubation, the cumulative negative priming in the locust litter addition treatment (−389.0 mg C kg⁻¹ soil) demonstrated that the negative PE during the initial stage of incubation balanced the positive PE at the later stage and promoted soil carbon storage (Figure 3). In the deep soil layer, all litter treatments maintained similar PEs as the top layer during incubation. However, the direction and magnitude of cumulative priming at the end of the incubation were different from those in the top layer. The cumulative PE induced by locust, Chinese pine, and mixed litter addition was −191.7 mg C kg⁻¹ soil, 89.9 mg C kg⁻¹ soil, and 187.8 mg C kg⁻¹ soil, respectively (Figure 3). Compared to that in the top layer, the response of soil organic matter mineralization in the deep layer was more sensitive to litter addition.

3.4. Soil Microbial Communities

In the topsoil layer, leaf litter addition significantly affected the soil microbial communities, especially causing an increase in the total PLFAs, bacterial PLFAs, fungal PLFAs, and G⁻PLFAs compared to those in the control soils (

Table 3. Effects of litter addition on the soil microbial community composition at the end of incubation.

Treatments	Total PLFAs (nmol·g−1)	Bacteria (nmol·g−1)	Fungi (nmol·g−1)	G+ (nmol·g−1)	G− (nmol·g−1)	AMF (nmol·g−1)
CK-T	64.75 ± 16.30b	26.02 ± 6.95b	5.47 ± 1.45b	16.20 ± 4.22a	10.56 ± 3.13b	1.58 ± 0.45a
PTT	92.86 ± 13.00ab	36.84 ± 5.60ab	12.68 ± 1.28a	20.09 ± 3.29a	16.44 ± 2.79ab	1.72 ± 0.27a
RPT	95.94 ± 21.50ab	40.50 ± 8.93a	15.09 ± 3.76a	23.26 ± 5.98a	18.11 ± 3.66a	2.13 ± 0.49a
PRT	105.64 ± 22.17a	43.82 ± 9.52a	13.39 ± 3.41a	24.78 ± 5.58a	19.12 ± 4.49a	2.14 ± 0.51a

Dates are the means ± standard deviations. Different lowercase letters within the same column indicate significant difference among treatments at p < 0.05 level. CK-T: topsoil without litter; PTT: topsoil with Chinese pine leaf litter; RPT: topsoil with locust leaf litter; PRT: topsoil with the litter mixtures of Chinese pine and locust; G⁺: Gram-positive bacteria; G⁻: Gram-negative bacteria; AMF: Arbuscular mycorrhizal fungi.

). The highest total PLFAs, bacterial PLFAs, and G⁻PLFAs were observed after mixed leaf litter addition. Fungal PLFAs in the locust litter treatment were highest. However, there was no significance in G⁺PLFAs or AMF among treatments. In addition, the ratio of fungi to bacteria (F:B) increased markedly in all leaf litter treatments. However, the ratio of G⁺ to G⁻ (G⁺:G⁻) showed a reverse trend (Figure 4). In the deep soil, the total concentration about PLFAs was very low and only four PLFAs might meet the demand of isotope analysis. Therefore, we did not analyze the data about PLFAs in the deep soil.

In topsoil layer, the 16:0, 18:1ω9c, and 18:1ω7c fatty acids assimilated a larger proportion of litter-derived carbon in all litter addition treatments. The percentages of litter-derived C within these three PLFAs were 66.70%–86.63%, 68.04%–91.38%, and 53.39%–64.67%, respectively, which illustrated that the bacterial marker 16:0, fungal marker 18:1ω9c, and G⁻ marker 18:1ω7c involved in litter decomposition (Figure 5).

The principal component analysis (PCA) results showed that the microbial community structures differed significantly among the litter addition treatments (Figure 6). Principal components 1 and 2 explained 70.27%and 17.62% of the total variance, respectively. Component 1 was mainly associated with 18:1ω7c, 17:0i, 16:1ω5c, 18:1ω9c, 16:0 and 18:0 fatty acids. In component 2, 16:1ω5c, 17:0i, 10me 18:0, 16:1ω7c, 18:1ω9c and 17:1ω8c fatty acids showed high factor weights. In addition, redundancy analysis (RDA) indicated that the first two axes explained 51.59% of the total variation in microbial communities. Soiltotal nitrogen, total organic carbon and pH were the most important factors influencing the microbial community composition (Figure 7).

3.5. Microbial Biomass Carbon

Leaf litter addition significantly increased soil microbial biomass carbon and had a greater influence on microbial carbon biomass carbon in the deep soil layer than in the topsoil layer (Figure 8). In the topsoil layer, the biomass carbon was significantly higher in the leaf litter treatments than in the control. The microbial biomass carbon was 21.13%, 10.52%, and 131.15% higher in the mixed litter treatment than in the Chinese pine treatment, locust litter treatment and control soil, respectively. In the deep soil layer, leaf litter addition also led to an increase in microbial biomass carbon. However, the litter type had no significant impact on the microbial biomass carbon.

3.6. Soil Enzyme Activities

Leaf litter addition significantly increased all detected enzyme activities, which had a greater influence on the deep soil layer than on the topsoil layer (

Table 4. Effects of litter addition on soil enzyme activities at the end of incubation.

Treatments	β-Glucosidase (mg·g−1·h−1)	Cellulase (mg·g−1·d−1)	N-Acetylglucosaminidase (μmol·g−1·d−1)	Alkaline Phosphatase (mg·g−1·h−1)
CK-T	1.06 ± 0.14c	0.10 ± 0.01d	243.36 ± 20.50d	14.42 ± 1.26cd
PTT	1.17 ± 0.06bc	0.16 ± 0.02c	270.96 ± 18.25c	17.63 ± 0.88c
RPT	1.57 ± 0.24a	0.23 ± 0.00a	291.64 ± 11.81c	26.25 ± 3.06ab
PRT	1.63 ± 0.03a	0.16 ± 0.01c	524.03 ± 8.58a	22.67 ± 2.09b
CK-D	0.11 ± 0.02d	0.11 ± 0.02d	16.38 ± 1.18g	3.47 ± 0.28e
PTD	0.19 ± 0.02d	0.10 ± 0.01d	117.30 ± 6.44f	3.71 ± 0.19e
RPD	1.39 ± 0.25ab	0.20 ± 0.01b	326.61 ± 6.04b	26.62 ± 4.55a
PRD	1.12 ± 0.18bc	0.17 ± 0.01c	152.30 ± 8.54e	12.75 ± 0.53d

Dates are the means ± standard deviations. Different lowercase letters within the same column indicate significant difference among treatments at p < 0.05 level. CK-T: topsoil without litter; PTT: topsoil with Chinese pine leaf litter; RPT: topsoil with locust leaf litter; PRT: topsoil with the litter mixtures of Chinese pine and locust; CK-D: deep soil without litter; PTD: deep soil with Chinese pine leaf litter; RPD: deep soil with locust leaf litter; PRD: deep soil with the litter mixtures of Chinese pine and locust.

). The enzyme activities after leaf litter addition mainly depended on the litter type and the soil layer. The β-glucosidase activity of the broadleaf litter (e.g., locust) treatment in the topsoil layer was significantly higher than that in the deep soil layer. However, the phosphatase and N-acetylglucosaminidase activities after the same litter addition showed reverse trends. In the coniferous leaf litter (e.g., Chinese pine) treatment, all detected enzyme activities were significantly higher in the topsoil layer than in the deep soil layer. Mixed leaf litter addition caused all detected enzyme activities, except the activities of cellulose-related enzymes, to increase greatly in the deep soil layer.

4. Discussion

Leaf litter addition significantly increased the cumulative CO₂ flux in the top and deep soil, which was caused by litter decomposition and soil mineralization (Figure 1 and Figure 2). In the topsoil layer, the release of litter-derived CO₂-C in the locust litter treatment increased rapidly during 14 days of incubation and then decreased gradually (Figure 2). The slow decomposition indicated that the gradual reduction of labile components and concomitant increase in recalcitrant components in the remaining residues. Our results showed that in the topsoil layer, during all incubation stages (a total of 56 days), the negative PE induced by locust litter amendment was −388.99 mg C kg⁻¹ soil (Figure 3). In particular, during the 14 days of incubation, the amount of mineralized C mainly derived from litter decomposition was 2.38 times higher than that induced by Chinese pine litter (Figure 2). This result agreed with previous studies [5,46] and confirmed the ‘preferential substrate utilization’ hypothesis [47]. However, the Chinese pine leaf litter amendment induced a negative PE. The intensity of the PE became stronger with the extension of the incubation time. This was inconsistent with our first hypothesis. There might be a home field advantage in the process of litter decomposition. In addition, the functional breadth hypothesis assumes that decomposer communities from recalcitrant litter environments have a wider functional ability, which degrades more litter species than functional narrow communities [48]. Therefore, we speculated that the soil microbial communities distributed in Chinese pine plantations had wider functional breadth and decomposed coniferous litter more easily, resulting in a negative PE and promoting soil carbon storage. The mixed litter addition caused the native soil organic carbon mineralization more intensely compared to the single litter treatment. It was reported that the slow decomposing species might have a synergistic effect on fast decomposing ones through carbon priming [49]. We deduced that locust litter (a high quality litter, fast decomposing species) was decomposed easily in Chinese pine plantation soils, which lead to the increase of microbial community size and thus showed higher decomposing ability. When the microbial communities were limited by the carbon sources from the locust litter, the added Chinese pine litter (a low quality litter, slow decomposing species) might induce carbon priming to enhance the decomposing ability of soil microbial community. However, we did not fully confirm the mechanism of the synergistic effect of the litter mixture because the differences about species mass loss was not measured.

In our study, we found that leaf litter addition induced more obvious soil carbon mineralization and PEs in the deep soil layer than in the topsoil layer (Figure 1 and Figure 3). This result suggested that the soil organic carbon in deep soil would turnover faster than that in topsoil in response to fresh organic matter input. The main reasons for higher levels of carbon mineralization and priming in deep soil might be as follows: (1) Based on the ‘microbial nitrogen mining’hypothesis, lower nitrogen availability in soils facilitates soil microorganisms to decompose recalcitrant organic matter to obtain nitrogen [50]. The lower content of available nitrogen in deep soil (topsoil: 11.69 mg·kg⁻¹; deep soil: 5.39 mg·kg⁻¹) might enhance microbial nitrogen mining and induce a higher PE. (2) Topsoil is rich in easily utilized substrates (e.g., litter, rhizodeposition, and roots) [11] and is thus suitable for the growth of microorganisms. Therefore, the soil microorganisms did not mineralize more soil organic carbon in the topsoil.

The higher ratio of fungi to bacteria after litter addition suggested changes in microbial substrate utilization (Figure 4a), which might induce the succession and competition between r- (fast-growing microorganisms) and K-strategists (slow-growing microorganisms) and thus determined the magnitude and intensity of the PE [51]. Through ¹³C-PLFA analysis, we found that the litter-derived PLFA composition was different among the three litter treatments (based on the percentage of litter-derived ¹³C in the PLFAs that exceeded 50%, Table S1). This result suggested that the litter type might be the main factor that caused the transformation of the soil microbial community and substrate utilization. In the three litter treatments, the carbon derived from litter was mainly incorporated into 18:1ω9c fungal PLFAs (PTT: 91.38%, RPT: 78.11%, and PRT: 68.04%), implying that the exogenous organic carbon was mainly utilized by fungi, which was in accordance with previous studies [12,52,53]. At the same time, this result reflected the important role of fungi in litter decomposition. The Pearson correlation analysis showed a negative relation between the ¹³C-labeled fungal PLFAs and the soil PE (r = −0.372). It might be that fungi are better able to grow in nutrient-poor environments than bacteria [54] and thus induced less intense positive soil PEs. In this study, we detected the response of soil microbial communities only after litter addition in the topsoil layer, allowing for their activity to reach detectable levels. The magnitude and intensity of the soil PE were obviously different between the top and deep soil layers. Therefore, how soil microorganisms affect the soil PE in the deep soil layer requires further research.

Leaf litter addition promoted microbial activity and boosted enzyme production [55]. In the present study, we found that the β-glucosidase, phosphatase, and cellulase activities in the locust litter treatment were significantly higher than those in the Chinese pine litter treatment (Table 4). However, the SOM decomposition rate increased by 41.7% in the topsoil. According to previous studies, high-quality residue additions decreased specific enzyme activities because of microbial saturation with substrate [56]. If the soil microorganisms are in a state of starvation or nutrient limitation, they might secrete specific enzymes to obtain energy from native soil organic matter, resulting in higher soil enzyme activities [57]. Therefore, our results were inconsistent with these reports, potentially because of the higher contents of easily dissolved substances and lower contents of recalcitrant compounds, such as ligninsand tannins, in the leaves of broadleaf species [58]. The input of labile substrates promoted microbial growth and activity, in turn alleviating the spatial inaccessibility to SOM through their expansion to new areas [59] and thus speeding up the decomposition of SOM.

5. Conclusions

Different kinds of leaf litter addition significantly affected the magnitude and direction of priming effect in all detected soils. In top soil, Chinese pine (PT, a low quality litter with higher lignin concentration and C:N ratio) and locustlitter (RP, a high quality litter) addition induced a negative PE and reduced the losses of soil organic carbon by 46.1% and 26.9%, respectively. However, PT addition gave the primed CO₂ of 89.9 mg C kg⁻¹ soil evolved from the deep soil, indicating coniferous leaf litter input accelerated the decomposition of SOM. The input of locust (broadleaf) litter decreased the release of CO₂-C derived from SOC. The magnitude of the PE in deep soil was more intensive than that in topsoil. Fungi and G^- bacteria were essential for decomposing litter and incorporating new litter C in to soil microbes (average: 69.0% and 55.7%).Soil microbial communities regulated the mineralization and immobilization of SOM through litter quality control. In summary, we speculated that the introduction of broadleaf species into coniferous plantations might be beneficial for enhancing deep soil carbon stocks. However, the effects of litter transplantation among broadleaf and coniferous forests andlitter addition amount on soil PE need further research.