Due to the burning of fossil fuels, the production and use of chemical fertilizer and other interferences caused by human activity, anthropogenic reactive nitrogen (N) is increasing and is deposited by precipitation and through dust accumulation in terrestrial and aquatic ecosystems ( Zhang et al. , 2015 ). Worldwide N deposition has caused serious environmental problems, such as the acidification of soil and water, the imbalance of soil-nutrient C/N ratios, increased leaching of soil nitrate and nutrients, and loss of biodiversity ( Bobbink et al. , 2010 ; Stevens et al. , 2018 ). However, N deposition has the potential to constrain the accumulation of anthropogenic CO ₂ in the atmosphere by increasing ecosystem carbon (C) storage. This could thereby slow the pace of climate warming ( Xia & Wan, 2008 ; Frey et al. , 2014 ), for which forest ecosystems play an important role by sequestering atmospheric CO ₂ . In addition, soils can contribute to C sequestration. Soil C sequestration and storage in forested ecosystems are mediated by soil microbial communities and are in dynamic equilibrium with environmental factors ( Lal, 2005 ; Zhu et al. , 2017 ).

Soil microorganisms are essential regulators of nutrient turnover in forest ecosystems ( Wardle et al. , 2004 ). They are involved in key soil biogeochemical processes including mineralization and humification of organic matter in soils ( Cotrufo et al. , 2013 ; Xia et al. , 2016 ). The turnover rate of soil organic matter may depend on the microbial community structure and the capacity of the community to metabolize different organic compounds ( Garcia-Pausas & Pa­terson, 2011 ). N addition is predicted to increase mi­crobial biomass via increasing C or/and N resour­ce availabilities; thus, increased N deposition could alter the structure and composition of soil microbial communities because of different competition among species of soil microorganisms for various forms of N ( Schimel & Bennett, 2004 ; Van Diepen et al. , 2010 ; Zechmeister-Boltenstern et al. , 2011 ). N deposition can affect microorganisms in different ways in different ecosystems. Previous studies have shown that high N deposition significantly reduced soil microbial bio­mass and decreased rhizodeposition, fueling rhizosphere com­­munities ( Mo et al. , 2010 ; Treseder, 2010 ; Eisen­hauer et al. , 2012 ). However, to date, the results are inconsistent. Some studies found that high N addition had no effect on soil microbial characteristics ( Keeler et al. , 2009 ). These inconsistent results indicated that the response of soil microbial communities to N deposition is highly variable among different environments. The­refore, more detailed studies should be conducted on this topic.

The application of a community level method to determine microbial function provides a more sensitive and ecologically sensitive method for determining he­terotrophic microbial community structure. This me­thod employs the commercial Biolog microtiter plate assay (Biolog Inc., Hayward, CA) for monitoring shifts in the metabolic patterns of microbial communities ( Garland, 1997 ; Campbell et al. , 2003 ). In recent years, the method has been widely used in the studies of soil microbial communities, which mainly included evaluations of the effects of different forest types, N deposition, and various factors (such as drought, fertilizers, contaminants, changes in plant cover) on catabolic abilities of soil microorganisms ( Bérarda et al. , 2011 ; Banning et al. , 2012 ; Brackin et al. , 2013 ; Sradnick et al. , 2013 ).

In this study, we designed a long-term (>4 years) N addition experiment in an old-growth subtropical evergreen broad-leaved forest, which faces high atmospheric N deposition (over 10 kg N ha ⁻¹ yr ⁻¹ ), to test the following two hypotheses: (1) N addition has negative effects on microbial biomass and functional capacity of soil microorganisms; and (2) Seasonal variations are much greater than the effects of N addition on the functional ability of the soil microbial community. These changes may be due to changes in species dominance, or they may be due to the constant physiological adaptation of soil microorganisms. The functional capacity of the soil microbial community was also analyzed in relation to chemical and physical soil variables.

The soil microbial biomass comprises only appro­ximately 2-3% of soil organic carbon, but it is an important factor regulating ecosystem processes. Our results indicate that soil microbial biomass in subtropical forest ecosystems exhibited a significant seasonal pat­tern, increasing in autumn and winter and decreasing in spring and summer. The similar seasonality was observed in other ecosystems such as alpine ( Lipson et al. , 2000 ), warm temperate ( Zhao et al. , 2011 ) and subtropical ( Feng et al. , 2009 ). Seasonal variation in soil microbial biomass is controlled by environmental factors (e.g., soil moisture, temperature, pH, nutrient and C availability), plant physiological factors and organic matter input ( Wardle, 1998 ; Feng et al. , 2009 ; Kaschuk et al. , 2010 ). In the subtropical region, the resource availability in soil is usually high due to the rapid turnover of organic matter. However, during the growing season, strong competition may occur between vegetation and soil microbial communities. This may result in low microbial biomass caused by resource limitation for the microbial communities. A decreasing temperature in winter would weaken plant physiological activities and enhance organic matter input by litter and fine-root turnover. This results in a significant change of soil microbes between winter and summer ( Lipson et al. , 2000 ). Conflicting reports on seasonal changes in soil microbial biomass exist ( Wardle, 1998 ; Nieder et al. , 2010 ). Further research is required to reveal the mechanisms regulating the seasonal patterns of soil microbial biomass in different forest ecosystems.

Due to the low content of N and P in the acidic soil in this region (Table 1), the addition of N can alleviate the N limitation of the forest ecosystem ( Di et al. , 2017 ). The addition of N increased the TN content of the soil; meanwhile, the TP content was reduced, and thus, the soil N/P ratio increased. Therefore, N addi­tion is likely to drive P-limitation in this subtropical evergreen broad-leaved forest. In forest ecosystems, soil microorganisms are the key factors determining biodiversity, productivity and energy turnover ( Heijden et al. , 2010 ). A recent Meta-analysis based on 151 experiments showed that atmospheric N deposition negatively affects soil microbial growth, composition, and function across all terrestrial ecosystems, with more pronounced effects with increasing N deposition rate and duration ( Zhang et al. , 2018 ). Nevertheless, other studies found that the effects of N addition on soil microbial biomass and community structure differ greatly among different forest types ( Keeler et al. , 2009 ; Allison et al. , 2010 ; Mo et al. , 2010 ). For nutrient addition in subtropical forests in particular, there is still great uncertainty ( Balser et al. , 2001 ). In this study, there were no significant differences in soil microbial biomass, community activity and diversity between N-amended and control plots. The mechanism is that food and energy required by soil microorganisms are mainly limited by soil carbon. N addition may lead to changes in soil carbon availability, or the unbalanced distribution of soil carbon from above- and below-ground, which may have indirect adverse effects on the survival and reproduction of soil microorganisms. Generally, the soil nutrient availability, enzyme ac­tivity and the change of community structure in the aboveground part can affect the soil microbial community to a great extent. Jing et al. (2017) reported that soil extracellular enzyme activity had no obvious response to N addition at our site. Cui ( Cui et al. , 2017 ) found that N addition had no significant effect on soil microbial diversity in this area by 16S analysis. These results, which indicate that soil microbial survival does not respond to N addition is indicative of N saturation at our site, which would require a long-term N addition experiment to verify.

The results of the RDA showed that the soil mi­crobial community remained relatively stable under the N addition in this study. It is generally believed that soil moisture and pH are two key factors affecting soil microbial community and function ( Bobbink et al. , 2010 ; Brockett et al. , 2012 ). At our site, differences in soil microbial communities caused by N addition and seasonal factors are largely explained by pH, SWC, NO ₃ ^- -N and TP (Fig. 5, Fig. S11-12 [suppl.], p < 0.05). Among various soil properties that affect the soil microbial community structure, pH is a key factor. This is in agreement with many studies, especially in those conducted in acidic soil of warm and humid areas ( Leff et al. , 2015 ; Zhang et al. , 2017 ; Zhang et al. , 2018 ). When soil pH is lower than 4.5, soil microorganisms will be significantly inhibited ( Högberg et al. , 2007 ; Johannes et al. , 2009 ). After 5 years of N addition, soil pH showed a small decline. The change of soil pH in this region had no significant effect on microbial diversity ( Cui et al. , 2017 ). In addition, root exudates and litterfall production in forest ecosystems are controlled by vegetation development and hydrothermal conditions and are the main carbon sources for microbial communities ( Priha et al. , 2001 ). Di et al. (2017) showed that the understory vegetation (such as saplings, surface ferns and herbaceous plants) were sensitive to N addition. This should contribute to the changes of the soil microbial community and its activity.

Our five-year experiment showed that simulated N deposition had no significant effects on soil microbial biomass, community metabolic activity or functional diversity in a subtropical old-growth forest ecosystem. The main soil factors resulting in the change of soil microbial community were pH, soil moisture, NO ₃ ^- -N and total P. This indicates that N addition may likely drive P limitation in this subtropical forest. In addition, the physiological changes of understory vegetation and aboveground litterfall production affected by N addition cannot be ignored and could be an important contributor to the changes of soil microbial community and its activity. Due to the global and long-term nature of atmospheric N deposition, its impact on forest ecosystems is bound to be long-term and to entail complicated processes.

We wish to thank Di Tian, Huirong Wang and Zefu Wang for their help with field soil sampling and laboratory assistance. We appreciate helpful suggestions on the statistical analysis from Chonghua Xu.