The dramatic short-term changes in soil organic matter (SOM) caused by unstable factors (e.g. the input of the labile carbon) are called priming effects (Kuzyakov et al., 2000). The priming effect is a key factor affecting the soil carbon (C) cycle and relates to the regulation of SOM mineralization by root-derived C inputs (Cheng, 2009) and the process of nitrogen mining performed by soil microorganisms (Craine, 2007). Labile C flow is huge due to continuous inputs and outputs, despite the small net concentration (de Vries and Caruso, 2016). Labile C can be immediately utilized by the soil microbial community, thus its retention time is short and the products (e.g. CO₂ and microbial biomass) are easy to detect (Creamer et al., 2014). Labile C such as glucose is often used in substrate addition experiments to study the decomposition process of different C sources and C utilization dynamics in microbial communities (Li et al., 2018b, Liu et al., 2017; Zhang et al., 2016).

Nitrogen (N) is one of the factors limiting ecosystem productivity, and increasing atmospheric N deposition has a profound impact on global ecosystems (Galloway et al., 2008). The effect of N fertilization on the priming effect has been found to be positive (Chen et al., 2014; Qiao et al., 2014) and negative (Zang et al., 2016; Zhang et al., 2016). As crucial factors in the priming effect (Li et al., 2018a; Liu et al., 2017), the ratio of glucose to microbial biomass carbon (MBC) determines the direction of priming. A higher ratio of glucose to MBC generally leads to a lower priming effect (Blagodatskaya and Kuzyakov, 2008). Therefore, N fertilization can regulate the priming effect via its impact on MBC. Different mechanisms of priming have been defined based on the activities of microbial communities. The priming effect can associate with microbial “N mining” (Craine et al., 2007) and microbial stoichiometry (Hessen et al., 2014) at the community level. Labile carbon is both the energy and C source for soil microorganisms. Favorable nutrient stoichiometry between C and N enhances the growth of soil microorganisms (Hessen et al., 2014) and increases the utilization of labile C. To sufficiently utilize different substrate-derived C in soil, microbial community composition may change during C decomposition. According to Blagodatskaya and Kuzyakov (2008), “apparent priming” is defined as an acceleration of internal microbial metabolism without effects on SOM decomposition; meanwhile “real priming” happens with the increase of K-strategists and extracellular enzymes. Previous studies have found that priming in N fertilized soil is a short-term effect lasting 3 to 7 days (Zang et al., 2016; Li et al., 2017). However, the type of soil priming under N fertilization scenarios is seldom mentioned.

Soil decomposition driven by microorganisms is an important regulator of C storage. Since fungi are often seen as K-strategists which decompose the relatively recalcitrant SOM, the ratio of fungi to bacteria (F:B ratio) is widely used to predict soil quality or the extent of SOM decomposition (Kaur et al. 2005; Rousk and Frey, 2015). Understanding the decomposition dynamics of the soil microbial community helps to improve the prediction of land ecosystem C cycles (Leff et al., 2015). Therefore, changes in microbial community composition during soil priming may reveal interactions between soil C and the N cycles.

The objectives of this study are to explore the priming effect under different N fertilization gradients and the association with microbial communities. A ¹³C labeled glucose addition experiment was conducted on a long-term N fertilization experimental site with three gradients of urea fertilization (0, 4 and 16 g N m^-2 yr^-1). After the day of glucose addition (day 0), substrate-derived CO₂ and phospholipid fatty acid (PLFA) in soil samples were measured at day 3, 7, 21 and 35. We hypothesized: 1) that N fertilization reduces positive priming effects in semi-arid grassland soil; 2) that low gradient N fertilization favors microbial preference for labile C substrate; and 3) that different types of soil priming occur after N fertilization.

High N application reduces microbial activity and the loss of soil CO₂, and is beneficial to SOM accumulation (Reay et al., 2008). N fertilization suppressed soil CO₂ in the study area regardless of N gradient, based on the results of control CO₂. With the addition of glucose, SOM-derived CO₂ at all N levels was higher than the control on day 3, leading to a positive priming effect (Kuzyakov et al., 2000). Glucose-derived CO₂ in the soil at 4 and 16 g N m^-2 yr^-1 on day 3 was 3.05 times and 2.32 times that of soil at 0 g N m^-2 yr^-1, indicating rapid decomposition of labile C with N fertilization. The positive priming effect disappeared after day 7 at 0 and 4 g N m^-2 yr^-1, and after day 3 at 16 g N m^-2 yr^-1. Consistent with Li et al. (2017), cumulative primed C dynamics changed from positive to negative. At the late sampling stage a higher proportion of ¹³CO₂ in total CO₂ in the glucose-adding group may be the result of cross-predation of soil microorganisms (Creamer et al., 2014; Garcia-Pausas and Paterson, 2011).

The soil priming effect is an important factor in the regulation of the soil C balance (Zhang et al., 2016), and this process is regulated by nitrogen addition (Chen et al., 2014). According to the “N mining” theory (Craine, 2007), due to the limited availability of nitrogen, microorganisms use labile C to synthesize extracellular enzymes to accelerate SOM decomposition, leading to an increase in soil CO₂. The addition of ¹³C labeled glucose stimulated soil CO₂ at all N levels in this study, and the cumulative primed C was highest at 0 g N m^-2 yr^-1, suggesting that the strongest priming effect exists when N is most limited (Garcia-Pausas and Paterson, 2011).

Soil microorganisms switched to labile carbon instead of SOM in the N fertilization treatments, revealed by cumulative SOM-derived C and glucose-derived C at the three N levels. Soil at 0 g N m^-2 yr^-1 released the largest amount of SOM-derived C on day 35; the smallest amount was at 16 g N m^-2 yr^-1. The cumulative glucose-derived C on day 35 was highest at 4 g N m^-2 yr^-1, and showed a trend of 4 > 16 > 0 g N m^-2 yr^-1. Lower levels of N fertilization stimulated larger consumption of labile C, likely due to the favorable stoichiometry for microbial growth (Hessen et al., 2014; Li et al., 2018b).

N fertilization exerted an inhibitive impact on microbial PLFA biomass and the F:B ratio, in accordance with other studies (Wei et al., 2013; Liu et al., 2018). Fungi are considered a major player in recalcitrant C decomposition because of massive oxidative enzyme production (Sinsabaugh, 2010). A reduced F:B ratio following N fertilization might lead to a decrease in oxidative enzymes and consequently the inhibition of SOM-derived CO₂. However, no significant variation in F:B ratios among sampling days suggests poor linkage between the F:B ratio and SOM mineralization (Rousk and Frey, 2015). Since glucose is a universal C and energy source for microbial communities, activation of dormant microorganisms (Blagodatskaya and Kuzyakov, 2008), quick response of r-strategists in saprotrophic fungi (Lemanski and Scheu, 2014; Li et al., 2018b) and functional diversity within bacterial taxa (Lonardo et al., 2017) might obscure the transition between fungal and bacterial communities. To be better indicative of SOM mineralization, ¹³C labeled PLFA rather than total PLFA biomass should be considered in each fungal and bacterial community.

Microbial community composition at the early sampling stage was relatively consistent despite the drastic change in the priming effect in soil, suggesting that microbial successional dynamics did not significantly relate to the priming effect (Li et al., 2017; Rousk et al., 2016). “Apparent priming” is defined as a quick increase in the respiratory activity by the acceleration of microbial metabolism without further composition alteration (Blagodatskaya and Kuzyakov, 2008). Such priming could be associated with the activation of dormant microorganisms (Blagodatskaya and Kuzyakov, 2008) and the catabolic response of active microorganisms (Rousk et al., 2016). However, enhanced SOM decomposition at the early sampling stage echoed the definition of “real priming” (Blagodatskaya and Kuzyakov, 2008). Since N addition was found to regulate microbial carbon and nitrogen use efficiency along with the priming effect (Rousk et al., 2016; Liu et al., 2018), multiple mechanisms of priming effect under N fertilization scenarios need to be classified thoroughly, and microbial functions should be emphasized in future studies on the priming effect.

N fertilization substantially reduced soil CO₂ concentration, and N fertilization shifted microbial utilization of C substrate from bulk SOM to labile C. Glucose addition induced a positive priming in soil with N fertilization in semi-arid grassland soil, and the priming effect diminished after 7 days. Cumulative primed C was highest at 0 g N m^-2 yr^-1, and cumulative glucose-derived C was highest at 4 g N m^-2 yr^-1, in accordance with the theory of “N mining” and “microbial stoichiometry”. The F:B ratio was decreased by N fertilization in this study, echoing the inhibition of SOM-derived CO₂ by N fertilization. However, the relatively stable F:B ratio during sampling suggests that the F:B ratio is a poor indicator of the extent of SOM decomposition. The “apparent priming” was likely to occur at the early stage of sampling, since there was no significant alteration in microbial community biomass or composition. Considering the diverse functions in microbial communities, future studies should look into the different mechanisms of priming when mediated by N fertilization.

We thank the Duolun Restoration Ecology Research Station (part of the Institute of Botany, Chinese Academy of Sciences) for providing access to the study site.