Effects of inorganic nitrogen and litters of Masson Pine on soil organic carbon decomposition

Soil organic matter (SOM) mineralization represents one of the largest fluxes in the global carbon cycle. Numerous studies have shown that soil organic carbon decomposition was largely changed owing to the addition of litter, however very few studies have focused on the role of plant organs in the priming effects (PEs). Here, we studied the effects of different Pinus massoniana organs (fresh leaf, leaf litter, twigs, absorptive fine roots, and transport fine roots) on C4 soil respiration by applying the 13C isotopic natural abundance method. Results showed that the effects of plant organs on PEs were significantly different at the end of 210 days incubation, which can be ascribed to contrasting organs traits especially non-structural carbohydrates and water-soluble compounds. Transport fine roots and fresh leaf induced positive PE, whereas absorptive fine roots induced negative PE. Leaf litter did not change the native SOC decomposition. Plant organ addition can change the microbial community and result in the reduction of bacteria-to-fungi ratio. Our results suggest that with regard to determining the PE of the entire ecosystem, using fresh leaf to represent leaf litter and aboveground to represent underground is implausible.


Introduction
Soil organic matter (SOM) mineralization represents one of the largest fluxes in the global carbon (C) cycle [1]. Forest SOMs are the biggest C pool in the terrestrial ecosystem; hence, a litter change in these SOMs will vastly affect the global C balance [2]. Understanding the factors that regulate SOM turnover is essential to predict the terrestrial feedback on climate change [3]. On one hand, fresh organic matter can form new soil organic carbon (SOC) during litter decomposition [4]. On the other hand, fresh organic substrates can stimulate the decomposition of stabilized SOC through a phenomenon called the priming effect (PE) [5].
Plant litter typically comprises different organs, such as leaf litter, twigs, fine roots, and others (e.g., reproductive organs, bark, and detritus). Fine roots occupy nearly 48% of the annual a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 plant litter input, whereas leaf litter and twigs account for 41% and 11%, respectively [6]. The traditional definition of fine roots are roots with diameters � 2 mm. This definition covered the truth that many traits vary among root orders, and differences in these traits may influence root decomposition rates. Researchers have divided the fine roots into two distinct classes, namely, absorptive (first-to third-order roots) and transport fine roots (higher-order roots) [7], to enable comparisons among functionally similar roots. However, ecological theories generally presume the plant as a whole individual and use leaf traits to represent the whole tree traits [8,9]. This assumption overlooked the role of other organ traits in regulating the ecosystem's C cycle to some degree. Previous studies have explored the correlation between the decomposition rates of leaf litter and other organs, but no general conclusion has been addressed yet. Freschet et al. [10] have demonstrated that for structure-related traits, such as lignin, C controls the decomposability for all plant organs. However, Sun et al. [11] studied the decomposition dynamics of the leaf litter and fine root of 35 tree species in a temperate forest ecosystem and discovered that the decomposition mechanisms of fine roots and leaf litter were different. Phenolic substances are the major controlling factors in the decomposition of fine roots, whereas stoichiometry (e.g. C:N, lignin:N) determines leaf litter decomposition. These contrasting results proved that whole-plant ecological strategies are necessary to account for the impact of different plant organs on decomposition processes.
Priming effect is characterized by the effect of adding of exogenous substrates on SOM mineralization [12]. Numerous studies have shown that soil organic carbon decomposition was largely changed owing to the addition of external organic matter, and PE has become a universal phenomenon. However, the directions of the PE induced by litter vary, showing positive [13], negative [14,15], or no PE [16]. The added high-quality litters, such as fresh leaf and absorptive fine roots (lower C:N ratio), act as an energy source for microorganisms in degrading SOM by producing extracellular enzymes. However, according to nitrogen (N) mining theory, added recalcitrant C substrates, such as twigs and transport fine roots (higher C:N ratio), also stimulate SOM-decomposing microorganisms to chase N for alleviated nutrient limitation [17]. Furthermore, previous studies have demonstrated that the quality and quantity of litter can determine the microbial community [18]. For example, the relative abundance of actinomycetes and fungi increases as recalcitrant substrates are added [19], whereas the relative abundance of bacterial increases owing to the addition of labile substrates [20]. Bacteria that thrive on soluble C compounds can simultaneously facilitate SOM decomposition and stimulate the activity of slow-growing fungi to elevate enzyme production [21]. Yao et al. [22] suggested that the change in microbial community structure might alter the magnitude and direction of SOC mineralization. In this experiment, plant organs have distinct traits; thus, differences in magnitude and direction of the PE induced by plant organs can be observed. Very few studies have focused on the role of plant organs in the PE and microbial community structure [5].
Many experiments that study PE utilize 13 C-or 14 C-labeled fresh leaf. Most model designers use these PE data in representing the entire tree litter, including the above-and belowground litter, to predict the effect of SOM decomposition to environmental change. However, this method may lack precision. Fine roots are another important litter component. Previous literature has demonstrated that root decomposition rates strongly differ among root orders and attribute litter traits [23]. However, the process of PE induction of different functional roots is not fully explored.
For the past two decades, the N deposition in China shows an increase of approximately 8 kg N ha -1 [24]. N deposition can affect the litter decomposition and PE by changing the litter quantity and quality [25,26] and altering the biomass allocation strategy(i.e., aboveground vs. underground) [27]. Moreover N deposition can suppress microbial respiration [28], alleviate lower N availability and hence decrease PE according to "N-mining" [29]. However, different authors reported inconsistent effects of nitrogen addition on PE. Previous studies showed that N can enhance [30], decrease, [31], or not affect [14,32] PE, but these studies failed to indicate the effect if different plant organs are considered.
Masson pine (Pinus massoniana) is a native tree species that accounts for 7.7% of the total forest area in China [33] and occupies an important state in timber production. Masson pine exists mostly in the form of monoculture forests. At present, masson pine pure forests exhibit a decline in soil fertility and biodiversity, which results in the degradation of SOM [34]. Therefore, studying the decomposition process of masson pine litters and the associated PE process is critical to better understand the mechanisms of fertility maintenance and carbon formation balance strategy in southern China.
This study aims to determine the influence of masson pine organs chemistry on the direction and magnitude of PE at the early and the later phase of litter decomposition. To do so, we selected fresh leaf (FL), leaf litter (LL), twigs (T), absorptive fine roots (AR), and transport fine roots (TR) and applied the 13 C isotopic natural abundance method. We hypothesized that the direction and magnitude of the PE will vary with the chemistry of plant organs and differ among decomposition stages. In addition, we speculated that organs with high concentrations of easily degradable C compounds (e.g., water-soluble compounds [WSCs], cellulose, and non-structural carbohydrates [NSCs]) will deteriorate the PE in the early phase, as microorganisms preference, whereas induce increases in PE in the later phase, as production of extracellular enzymes. Moreover, following the N-mining concept, we hypothesized that PE will decrease as external nitrogen is added. Furthermore, we also postulated that microbial community will change differently from the chemistry of plant organs, and these changes will affect PE to some extent.

Ethics statement
The soil used in our experiment was permitted by the experimental station of Heilongjiang Academy of Agricultural Science. There were no other specific permissions required for these locations/activities because the experimental site was located at Huitong Natural Research Station of Forest Ecosystem. We confirmed that the field studies did not involve endangered or protected species and any protected locations.

Plant materials and soil
The study site is located at the Huitong Natural Research Station of Forest Ecosystem (26˚40 0 to 27˚09 0 N and 109˚26 0 to 110˚08 0 E) in Hunan Province, China. The study area has a humid mid-subtropical monsoon climate. The mean annual temperature is 16.5˚C, and the mean annual precipitation is 1,200 mm. The masson pine was planted in 1995 after clear-cutting of Chinese fir (Cunninghamia lanceolata), the stand densities of this forest is approximately 1200 trees ha -1 , the forest average diameter at breast height (DBH) is 14.22 cm, and the average tree height is 9.2 m. The soil is reddish oxisol soil, which developed from shale and slate. We collected Fresh leaf (FL), Twig (T) and Fine roots from three trees with mean DBH in July 2017. The entire root system of the selected trees was excavated from the base of the trunk, with the soil intact for laboratory tests. The rhizosphere soil was washed away with deionized water. Then, the roots were divided into Absorptive fine roots (AR (first-to third-order roots) and Transport fine roots (TF (fourth-and fifth-order roots) [7]. Leaf litter (LL) was collected from litter traps in November 2016. All plant organs were dried at 60˚C immediately after collection. Table 1 shows the chemical properties of the organs.
Experimental soil was collected from the plough layer (0-20 cm) of an agricultural site that has been planted with a C4 maize crop for over 20 years at the experimental station of Heilongjiang Academy of Agricultural Science (45˚69'N, 126˚62'E), Heilongjiang Province, Northeast China. The soil pH is 6.9, and the C and N concentrations are 17.3 g kg -1 and 1.5 g kg -1 , respectively. The soil sample was filtered using a 2-mm mesh sieve and thoroughly mixed. The δ 13 C value of the C4 maize soil is -15.8‰.

Experiment design
The experiment involves 2 N levels (with and without N) and 6 plant organ treatments (nonaddition soil control, FL, LL, AR, TR, and T), with a total of 12 treatments. Plant organs samples were ball-milled into fine powder and sieved through a 0.25-mm-mesh sieve. The amount of each litter organs added to the soil was SOC 10% (1.7 g kg -1 , considering a soil depth of 10 cm and volume weight is 1.1 g cm -3 ) corresponded to a realistic yearly input of litter C in masson pine plantations [35] and mineral N added was 100 ppm N. Each treatment had three replications. One hundred grams of pre-incubated soil (oven-dry equivalent) was placed in a 1-L mason jar, and each litter component was added and thoroughly mixed. Aqueous KNO 3 (1 mL) and deionized water (1 mL) were added to serve as the inorganic and non-N treatment, respectively. The soil moisture was maintained at 60% of the water holding capacity by adjusting the deionized water for the entire incubation period. We punched two holes in the mason jar lids to install the bulkhead connectors (SMC, KQ2E06-00A; Singapore). Polyurethane tubes (TU-0604; SMC, USA) were used to link the bulkhead connectors to a manual valve (VHK2-06F-06F; SMC, Japan). The manual valve was kept open during incubation but was closed between the sampling periods. The jars were then placed in the incubator (SPX-500; Jiangnan, Ningbo, China), which is set at 25˚C. The evolved CO 2 was captured on days 3, 7, 10, 15, 30, 60, 110, 160, and 210. We removed CO 2 inside each jar before gas sampling by circulating CO 2 free air, which was generated using an air compressor (ACO-318; Hailea, Guangdong, China) pumped through a soda lime column for 2 min. The jars were immediately sealed by closing the manual valve. The gas was collected 24 h after sealing using an injection syringe, which was also used to collect the air. The air was tested for CO 2 production. The CO 2 concentration and δ 13 C were analyzed using a high-precision isotopic CO 2 cavity ringdown spectrometer (Picarro G2131-i Analyzer; Picarro, Inc., Santa Clara, CA, USA).

Plant and soil chemical analysis
The oven-dried plant experimental sample was ground and then sieved using a 0.5-mm mesh before the analysis of its chemical properties. C and N are determined using element analyzer (Model CN, vario Macro Elementar, GmbH, Germany). The total P was calorimetrically measured using the phosphomolybdic blue color method, whereas metallic elements (K, Ca, Na, Mg, Cd, and Mn) were gauged using a novAA 350 atomic absorption spectrometer analyzer (Analytik Jena AG; Jena, Germany) and then digested in 1 mol of HCl. WSCs, cellulose, hemicelluloses, and lignin were determined according to the Van Soest extraction protocol [36] using an ANKOM Fiber Analyzer (Ankom Technology; Macedon, New York, USA). NSCs and soluble and total phenolics were calorimetrically determined using a modified phenol-sulfuric acid procedure [37] and Folin-Ciocalteu reagent (gallic acid as the standard), respectively. The soluble and total phenolics were investigated using a 0.5-g sample with 30 mL of water and methanol (50%), respectively, and then shaken for 2 h before filtering. The concentration of ester/ether-bound phenols was determined through sequential extraction according to the method of Tamura and Tharayi [11]. Condensed tannins (CT) were determined using the acid butanol method. Furthermore, the soil's NO 3 concentration was tested by extracting 2 mol of KCl and applying flow injection analysis.

Calculation and statistical analysis
where R t is the CO 2 -C amount (mg CO 2 -C kg −1 dry soil), ΔC/Δt is the emitted CO 2 amount per day (ppm), T is the incubation temperature (25˚C), and V is the chamber volume (m 3 ). The constants 12 and 22.4 are the molecular weight of C and the molar volume of an ideal gas at standard temperature and pressure, respectively. We separated SOM-derived CO 2 -C (R s , mg CO 2 -C kg −1 dry soil) from plant organs-derived CO 2 -C (R p , mg CO 2 -C kg −1 dry soil) using the following mass balance equations: where δ 13 C s is the 13 C abundance of SOM-derived CO 2 -C measured in control soils, δ 13 C f is the 13 C abundance of different organs, and R t and δ 13 C t are the total CO 2 -C emitted on the day of the measurement and its 13 C abundance, respectively. Soil (or plant organs) cumulative production of CO 2 -C (T s (T p ), mg CO 2 -C kg −1 dry soil) at early and later stages of decomposition was calculated by the following equation: where R s(p) i and R s(p) i+1 are soil (or plant organs) CO 2 -C amount at ith and (i + 1)th incubation time (mg CO 2 -C kg −1 dry soil), respectively; t i+1 −t i is the interval between the ith and (i + 1)th incubation time (day); and n is the number of incubation times. The remaining proportion of litter C decomposition (L d , %) was calculated using the following equation: where T d is the cumulative CO 2 (mg CO 2 -C kg −1 dry soil) efflux from litter during the incubation period and M d is the amount of litter C added to soil (mg C kg -1 ). The annual decomposition rate constant (k, year -1 ) was calculated by the following equation: where M d is the amount of litter C added to soil (mg C) and M t is the remaining amount of litter C (mg C) at sampling time t in the year. The PEs was calculated as

PE ¼ ðT s amended soil À T s control soilÞ=T s control soil % ð7Þ
One-way ANOVA, followed by post-hoc LSD test, was employed to test the differences in the initial different organs characteristics, decomposition rate constant (k), the amount of organ-derived CO 2 -C, PEs, and microbial community among the treatments. Two-way ANOVA was employed to analyze the effects of plant organs, nitrogen addition, and interactive effects. Repeated measures ANOVA was employed to evaluate the effects of plant organs, nitrogen addition, and interactive on the mineralization rate of litter organs of P. massoniana; Relationships between plant litter traits and the amount of CO 2 -C from the organs or SOM were determined through linear fitting. Independent t-test was used to test the NO 3 -concentration of soil with and without NO 3 addition. These processes were performed in SPSS 19.0. (SPSS Inc., Chicago, IL, USA).

Litter decomposition
Masson pine organs all displayed a high decomposition rate in the initial stages of litter decomposition (0-15 days) and then remained low during later phase of litter decay (16-210 days) (Fig 1). At the end of the incubation, the amount of LL-derived CO 2 -C was 72.2% higher than that derived from AF. Repeated-measures ANOVA and k showed the CO 2 -C derived from plant organs was significantly affected by litter components (P < 0.001) (Fig 2 and Table 2). The addition of N did not significantly affect the organ-derived CO 2 -C (P > 0.05). In the early phase, N and litter components significantly affected the amount of organ-derived CO 2 -C (P = 0.017 and P < 0.001, respectively). The amount of FL-derived CO 2 -C was 19.2%, 22.8%, 29.3%, and 111.1% higher than that from TR, LL, T, and AR, respectively. The addition of N significantly decreased the amount of organ-derived CO 2 -C (P = 0.017), and no significant interaction between litter components and N was observed during plant organ decomposition (P > 0.05). In the later phase, only the litter components exhibited an effect on the amount of organ-derived CO 2 -C (P = 0.034) (Fig 3). In this case, the amounts of LL-derived CO 2 -C were 31.6%, 47.7%, 72.3%, and 85.1% higher than those derived from FL,T, AR, and TR, respectively.

Fig 2. Effects of organ type and inorganic N addition on decomposition rate constant (k).
Vertical bars are standard errors (n = 3). Different letters over the bars indicate statistically significant differences between treatments (P < 0.05). Two-way ANOVA result was showed above (NS: no significant difference

Priming effects
The addition of N and different litter components to the soil sample influenced the SOC decomposition to varying extents (Fig 4). To evaluate the effects of the observed PEs on the C storage of soil, we calculated the PE for each treatment using (Eq 7). The result shows that PE was significantly affected by the litter components (P < 0.001). In addition, TR and FL induced significant positive PE, whereas AR induced negative PE. We also observed that the addition of LL did not change the native SOC decomposition. Moreover, the effect of the addition of inorganic N was insignificant (P > 0.05).
The extent of PE also depends on the incubation stages (Fig 5). In the early stage of incubation, AR, TR, and LL caused a negative PE, whereas FL and T caused no PE. In the later stage, TR, FL and T caused a significantly positive PE, whereas LL and AR caused no and negative PE, respectively. In these two phases, PE was significantly affected by the litter components (P < 0.001) (Fig 5) but not by N addition (P > 0.05).

Relationship among litter traits, decomposition, and PE
Different litter components have various traits. The amount of organ-derived CO 2 -C during the later phase was positively correlated with Ca concentration (P < 0.05, Fig 6A). Similarly, the amount of CO 2 -C derived from plant organs during the entire incubation time exhibited significant positive correlation with Ca (P < 0.05, Fig 6B). Soluble sugar concentration was positively correlated with SOM-derived CO 2 -C during the early phase and the entire incubation time (P < 0.05, Fig 6C and 6E), WSCs were positively correlated with SOM-derived CO 2 -C derived during the early phase (P < 0.05, Fig 6D).

Soil microbial community
The addition of masson pine organs (with or without N) did not affect the total microbial biomass determined by PLFA (Fig 7A), but it significantly affected the relative abundance of G+, G-, bacteria, fungi, and actinomycetes (Fig 7B-7F). The addition of LL and FL significantly increased the relative abundance of fungi, whereas the addition of TR and T significantly decreased the relative abundance of bacteria. Consequently, the ratio of bacteria to fungi was significantly decreased (Fig 7H). Moreover, the addition of N significantly affected the relative abundance of G-, bacteria, and actinomycetes. Effects of litters on soil organic carbon decomposition

Discussion
This study provides new insights on PE and the effect of dividing the litters into different plant organs on the PE-induction mechanism of the latter. Our findings demonstrate that different organs from the same individual exhibit contrasting impacts on the SOM balance owing to distinct traits. Moreover, we discovered that the PEs among FL, LL, and AR differ (positive, none, and negative PE, respectively). Therefore, the suggested traditional methods in predicting the C balance in the ecosystem are implausible. Effects of litters on soil organic carbon decomposition

Effect of plant organs on litter decomposition
In this study, we found that litter component decomposition exhibits different performances during the two phases. FL decomposed faster than LL during the early phase, whereas LL decomposed faster than the other components during the later phase. This phenomenon can be attributed to the litter quality [39]. The main factors affecting the decomposition rate vary according to the phase of decomposition [40]. In the initial phase, litter mass loss is controlled by easily degradable compounds, such as WSCs and non-protected cellulose [41]. FL contains 4.2 times higher cellulose concentration and 1.7 times higher WSCs than LL (Table 1). In addition, Ca played a significant role in litter decomposition, which is consistent with the study of Berg et al. [42]. This finding was also supported by the measured Ca concentration, in which Effects of litters on soil organic carbon decomposition the highest amount was observed in LL. Moreover, Ca is positively correlated with organderived CO 2 -C (Fig 6A and 6B). We also found that AR has slower decomposition rate than LL. This result agrees with the result of the study conducted by Sun et al. [11], in which they investigated the decomposition of first-order roots and LL from 35 woody plant species in a Chinese temperate forest. They determined that first-order roots decomposed slower than LL after six years of exposure in the field. They attributed this phenomenon to NSCs and polyphenols. In our experiment, AR Effects of litters on soil organic carbon decomposition contained the lowest soluble sugar and soluble phenolics (Table 1), thereby resulting in the slowest decomposition rate.

Effect of plant organs on PEs
A significantly negative PE was observed during the early incubation stage, which is consistent with other previous works [16,30]. This result might be because of the preferential substrate utilization of microorganisms when adding a labile C compared with SOM [43].
In the later phase, all components showed positive PE, except AR. LL and TR exhibited the highest PEs, followed by T and FL. Generally, there are two mechanisms to explain the positive PE. The first possible mechanism is "nitrogen mining" theory: if substrate lack N, microorganisms will accelerated SOM decomposition to access protected N. In our experiment, we can Degrading these low quality (lack N) litter will lead to a higher positive PE. However, our result showed that LL, which has the highest C:N ratio (141.52) and lignin:N ratio (37.21), have no PE during the incubation time. Therefore, the abovementioned mechanism cannot fully explain why LL exhibits no PE. The second possible mechanism is "Co-metabolism" for the degradation of other complex C compounds, microbials must assimilate labile NSCs that can provide the necessary energy for the production of various enzymes, which can co-metabolize the contemporary SOM fraction [16,44], especially some lignin and cellulous extracellular enzymes that can degrade the soil parts that are difficult to decompose [17]. Blagodatskaya and Kuzyakov [43] also found that PEs are related to the energy input of the microorganisms in the soil. In our experiment, LL and AR contain low levels of NSCs and WSCs, thereby preventing microorganisms from achieving sufficient energy. This speculation is supported by our findings that soluble sugar and WSCs are positively correlated with SOM-derived CO 2 -C. Similar to our finding, Wu et al. [45] also declared that the fresh litter of coniferous species with high C:N ratio limits the PE.

Effect of N on decomposition and PE
In the current study, we found that adding inorganic N has a minor effect on plant litter decomposition, which might be caused by the following reason. Zhang et al. [46] declared that the effect of N on low-quality plant litter decomposition is insignificant, and P. massoniana litter can be classified as a low-quality litter. We also found that the effect of the addition of N on PE was insignificant, possibly owing to soil properties. Generally, inorganic N addition can alleviate the limitation of microbial activity, which reduces microbial N-mining and PE [31,47]. However, our materials (C 4 soil) have higher N concentration than normal soils (148mg/ kg NO 3 -) (Fig 8), which can alleviate this phenomenon.

Effects of plant organs on microbial community
Based on PLFA analyses, we determined that organ addition generally increases fungal abundance more than bacterial abundance, causing a decrease in bacteria-to-fungi ratio (  [50] also discovered that fungi communities will increase and use their enzyme production to decompose the SOM when the easily available substrate is exhausted. Moreover, the addition of LL and FL significantly increased the relative abundance of fungi, whereas the addition of TR and T significantly decreased the relative abundance of bacteria. This discrepancy indicates that the response of bacteria and fungi to an external C source was determined by the complexity of the C substrates [51]. Wang et al. [38] suggested that the differences in the substrate's chemical quality can change the relative abundance of bacteria and fungi in soils, which can explain the different responses of the microbial community to the addition of different organs. Furthermore, after adding N, little or no change in microbial composition was noticed within 210 Effects of litters on soil organic carbon decomposition days. This result is consistent with the findings of Hicks et al. [51], where they stated that nutrient treatments are extremely small-relative to the inherent nutrient status of the soil-to change the microbial composition.

Conclusions
This study presents important implications in the assessment of the C balance in soil by considering the relative contribution of the different organs of forest litters. The effects of plant organs on PE were significantly different, which can be ascribed to contrasting organs traits especially NSCs and water-soluble compounds and co-metabolisms (i.e., the promotion in microbial activity as a result of energy supply). TR and FL induced positive PE, whereas AR induced negative PE. LL did not change the native SOC decomposition. With regard to determining the PE of the entire ecosystem, using FL to represent LL and aboveground to represent underground is implausible. Plant organ addition can change the microbial community and result in the reduction of bacteria-to-fungi ratio. SOC decomposition by microorganisms is limited by the energy materials in nutrient-rich soil. To verify the universality of our results, we will investigate different plant varieties (coniferous and broad-leaved or evergreen and deciduous) and build an explicit model that will involve litter decomposition rates, turnover rates, and PE.