Mineral Elements of Subtropical Tree Seedlings in Response to Elevated Carbon Dioxide and Nitrogen Addition

Mineral elements in plants have been strongly affected by increased atmospheric carbon dioxide (CO2) concentrations and nitrogen (N) deposition due to human activities. However, such understanding is largely limited to N and phosphorus in grassland. Using open-top chambers, we examined the concentrations of potassium (K), calcium (Ca), magnesium (Mg), aluminum (Al), copper (Cu) and manganese (Mn) in the leaves and roots of the seedlings of five subtropical tree species in response to elevated CO2 (ca. 700 μmol CO2 mol-1) and N addition (100 kg N ha-1 yr-1) from 2005 to 2009. These mineral elements in the roots responded more strongly to elevated CO2 and N addition than those in the leaves. Elevated CO2 did not consistently decrease the concentrations of plant mineral elements, with increases in K, Al, Cu and Mn in some tree species. N addition decreased K and had no influence on Cu in the five tree species. Given the shifts in plant mineral elements, Schima superba and Castanopsis hystrix were less responsive to elevated CO2 and N addition alone, respectively. Our results indicate that plant stoichiometry would be altered by increasing CO2 and N deposition, and K would likely become a limiting nutrient under increasing N deposition in subtropics.


Introduction
Mineral elements are important for plant growth and ecosystem function [1]. Base cations (potassium, K; calcium, Ca; magnesium, Mg) play a vital role in the capacity of buffering against acidity changes through exchange reactions [2]. They can also help plants against different stresses, such as drought, salinity and high temperature [3,4]. Trace metal cations (aluminium, Al; copper, Cu; manganese, Mn) are important both as micronutrients (10 -5 %~10 -3 %) [5] and as toxins when at high levels [6]. It is crucial to obtain sufficient concentrations of nutrient elements and maintain relatively stable stoichiometry in plant tissues for health [7,8]. Global change induced by human activities, such as increasing atmospheric carbon dioxide (CO 2 ) concentration and nitrogen (N) deposition, has profoundly altered the biogeochemical cycles of several elements [9,10]. However, we know little the influence of increasing CO 2 and N deposition on these mineral elements in plants.
The increasing anthropogenic atmospheric CO 2 concentration stimulates plant growth, which increases C storage on land [11]. The extent to which elevated CO 2 increases plant growth, however, can be controlled or modified by available mineral elements in soil [12]. Elevated CO 2 often increases carbohydrates in plants and might logically be expected to lead to a decrease in the mineral element concentrations in plant tissues [10]. The nutrient dilution may preclude the positive effects of elevated CO 2 on plant growth [13,14]. Consequently, the interaction between CO 2 and nutrient status in plants has significant implications to the responses of forests to global change. While a frequent observation is that plants grown under elevated CO 2 typically have reduced tissue concentrations of N [15], the changes of other mineral element concentrations under elevated CO 2 were much more complex [16]. Using a meta-analysis method, Duval et al. [17] suggested that mineral elements were clearly different in response to elevated CO 2 . Unfortunately, subtropical and tropical forests were not well represented in this meta-analysis. Our previous study has revealed that N and P concentrations in plants were not decreased by elevated CO 2 in subtropics, and instead P concentrations in plants positively responded to elevated CO 2 [18]. The results that challenged the assumption of declines in plant nutrient concentrations under elevated CO 2 [10], raise the question: what would happen to other mineral elements in response to elevated CO 2 in subtropical forests? As subtropical and tropical forests are characterized with multiple-nutrient limitation [19,20], the understanding of plant mineral elements in response to elevated CO 2 is critical for a better modeling plant productivity and biogeochemical cycling in forest ecosystems.
The increasing atmospheric deposition of N-containing compounds could have a pronounced effect on plants in response to elevated CO 2 [21,22]. Enhanced N deposition is associated with the accelerated loss of soil base cations, mobilization of heavy metal elements or lowered concentrations of base cations in forest ecosystems [23,24], and hence leads to nutrient imbalance in plant tissues. Through a meta-analysis, Lucas et al. [2] also suggested that foliar base cations consistently decreased following N addition over periods less than five years. These studies, however, separately treated the effects on nutrient status for N addition and CO 2 increase, and did not combine the effects of N addition with elevated CO 2 except for a few ones [25,26]. As air pollution and climate change are closely linked [25], the information on how N addition affects the dynamics of mineral elements in plant tissues under elevated CO 2 could lead to the development of a better perspective on plant nutrients in the contemporary complex environment.
Many studies on plant mineral elements in response to elevated CO 2 and N deposition focused on leaves [2]. However, different plant organs have different responses to elevated CO 2 and N deposition. For example, as leaves are highly metabolically active, the strength of regulatory control over elements would be stronger in leaves than in roots [27,28]. Therefore, we examined the responses of mineral elements in different plant organs (leaves and roots) to elevated CO 2 and N addition. Previous studies also have reported that the responses of terrestrial plants to elevated CO 2 and N deposition were species-specific, potentially driving a shift of the inter-specific competitive interactions and inducing species composition changes [29,30]. Therefore, it is necessary to examine the responses of multiple species to elevated CO 2 and N addition. However, there are big challenges in conducting the research of elevated CO 2 in mature forests due to their large stature and biological complexity [31]. The adult tree may be less responsive to environmental changes than the seedling. Thus, we used open-top chambers to study the effects of elevated CO 2 and N addition on the mineral elements (K, Ca, Mg, Al, Cu and Mn) in leaves and roots of the seedlings in five subtropical tree species over five years

Ethics statement
The study site was owned by South China Botanical Garden, Chinese Academy of Sciences (CAS). The study was approved by South China Botanical Garden, CAS. All necessary permits were obtained for the described studies. The study did not involve endangered or protected species.

Study site
The study was carried out at South China Botanical Garden, CAS, Guangzhou City, Guangdong Province, China (23°20 0 N and 113°30 0 E). The area is characterized by a monsoon and humid climate. The mean annual temperature is 21.5°C, and the mean relative air humidity is 77%. The annual precipitation rages from 1600 mm to 1900 mm with a distinct seasonal pattern, of which about 80% falls from April to September (wet season) and 20% occurs from October to March (dry season). The N deposition was high at our experimental site, with about 56 kg ha -1 yr -1 for the wet N deposition measured in 2006 [32].

Open-top chamber design
Ten open-top chambers were set up in an open space being exposed to full light and rain. Each chamber had a 3-m diameter, a 0.7-m deep below-ground part and a 3-m high above-ground part (adjusted to 4.5 m later). The below-ground part was delimited by brick walls in order to prevent any lateral or vertical water and/or element fluxes with the outside surrounding soils. Three holes at the bottom of the walls were connected to stainless steel water collection boxes. The above-ground part was wrapped with impermeable and transparent plastic sheets, leaving the top completely open. In the treatments with elevated CO 2 , an additional CO 2 came from a tank, and was distributed by a transparent pipe that entwined the inner wall of the chamber in a snake shape at the height of 0.5-2.5 m. The pipe had pinholes at 1 cm intervals. The pipe was connected to a fan to ensure that CO 2 was equally distributed in the entire chamber. The additional CO 2 was applied daily from 8:00 am to 5:00 pm except for rainy days. The flux of CO 2 from the tank was controlled by a flow meter to reach a target concentration of CO 2 inside the chambers. The CO 2 concentrations on the five planes (0.5, 1.0, 1.5, 2 and 2.5 m in height) in the chambers were monitored once a month using a Licor-6400 (LI-COR Inc., Lincoln, NE, USA).

Experiment design
Soils were collected from a nearby evergreen broad-leaved forest after harvesting in March 2005. Three different soil layers (0-20 cm, 20-40 cm and 40-70 cm) were placed into the belowground part of the chambers correspondingly after being homogenized separately. The bedrock was sandstone and shale. Soils were classified as ultisols following the United States Department of Agriculture (USDA) soil classification system [33]. Eight one-to two-year old seedlings for each tree species were randomly planted with inter-specific mixtures in each chamber at the density of 0.15 m 2 plant -1 . As P. massoniana died in the second year of our experiment, we studied the other five tree species in this experiment.
From April 2005, four treatments with two levels of CO 2 concentrations (elevated CO 2 and ambient CO 2 ) and two levels of N additions (with and without N fertilizer) were randomly applied to the ten chambers. Due to the logistically challenging to maintain the treatments with elevated CO 2 , it is expected that there would be more variations in the treatments with elevated CO 2 than in those with ambient CO 2 . In the face of limited resources, the treatments with elevated CO 2 replicated three times, while those with ambient CO 2 had two replications. That is, three chambers received an elevated CO 2 with N fertilizer (CN), three chambers did an elevated CO 2 without N fertilizer (CC), two chambers did an ambient CO 2 with N fertilizer (NN), and finally two chambers served as controls (ambient CO 2 without N fertilizer (CK). The elevated CO 2 treatments had a concentration of CO 2 at about 700 μmol CO 2 mol -1 . The N fertilized treatments were conducted by spraying once a week with a total amount of NH 4 NO 3 -N at 100 kg N ha -1 yr -1 .

Sample collection and measurement
The initial soil chemical properties were measured before the experiment (See Table 1

Statistical analysis
Normality of the variables was examined with the Kolmogorov-Smirnov test, and the homogeneity of variance was tested with the Levene's test. Data were logarithmically transformed when normality and homogeneity of variances were not conformed. We analyzed data by repeated measures ANOVA using the following mixed linear model for each plant organ: Dependent variables = S + C + N + S×C +S×N + C×N + S×C×N, where S was the effect of different species, C was the effect of the CO 2 treatments, and N was the effect of the N treatments. The effect of the chambers was a random factor in the model. Although there were just two replications for CK and NN, more statistical power could be gained when data were analyzed with repeated measures ANOVA. When there was a significant interaction of the CO 2 treatments and N treatments, the differences between the four treatments (CK, NN, CC and CN) were further analyzed using Tukey multiple comparison test (HSD). The differences were considered to be statistically significant at P < 0.05. Data analyses were performed by the SAS software (SAS Institute Inc., Cary, NC, USA).

Base cations in tree species
Across all the five tree species, the concentrations of K, Ca and Mg were relatively higher in the leaves (7.0 mg g -1 for K, 7.4 mg g -1 for Ca and 1.3 mg g -1 for Mg) than in the roots (2.5 mg g -1 for K, 6.3 mg g -1 for Ca and 0.7 mg g -1 for Mg) (Figs. 1 and 2 and S1 Data). The effects of elevated CO 2 on the base cations did not vary with plant organs (Table 2 and S1 Table). However, N addition significantly reduced K concentrations both in the leaves and roots ( Table 2). The responses of Ca concentrations to N addition were different between roots and leaves, with some decreases in the roots but not in the leaves ( Table 2).
The base cations significantly varied with species and the sampling time ( Table 2). The effects of elevated CO 2 on the concentrations of the base cations largely depended on tree species (Table 2) (Figs. 1 and 2). On the contrary, elevated CO 2 significantly increased the K concentrations in the roots of C. hystrix during the experimental period and those in its leaves in 2008 ( Figs. 1 and 2). N addition consistently decreased K concentrations among the five tree species (Table 2). There were significant influences of N addition on Ca and Mg concentrations of Schima superba Gardn. Champ. (S. superba) and O. pinnata. To be specific, the Ca concentrations in the leaves of S. superba responded positively to the NN treatments in the early period of this experiment before 2007 (Fig. 1), while the lower Ca concentrations were found in the roots of O. pinnata under N addition after 2007 (Fig. 2). The Mg concentrations were significantly decreased by N addition in the roots of S. superba in 2007 and 2008.

Metal cations in tree species
The mean Al concentration across the five tree species was relatively greater in the roots (2.45 mg g -1 ) than in the leaves (0.54 mg g -1 ) (Figs. 3 and 4 and S1 Data). The averaged Cu concentration was 8.92 mg kg -1 for leaves and 6.60 mg kg -1 for roots. The mean Mn concentration was greater in the leaves (178 mg kg -1 ) than in the roots (19 mg kg -1 ). Relative to the leaves, Al concentrations in the roots across the five tree species tended to be lower under elevated CO 2 (P = 0.079) ( Table 2). Elevated CO 2 significantly increased Cu concentrations in the roots by 18% across the five tree species ( Table 2).
The tree species had significantly influences on the metal cations ( Table 2). The metal cations greatly varied with the sampling time ( Table 2). The effects of elevated CO 2 on the metal ). There were some increases in Cu concentrations in the roots of O. pinnata but a decrease in the Mn concentrations in its leaves under elevated CO 2 . Elevated CO 2 increased Al concentrations in the leaves of S. hancei, and the effects were stronger with time (Fig. 3). S. superba exhibited higher foliar Mn concentrations in the CC treatments, but the positive effects of the CC treatment tended to be muted with time (Fig. 3). On the other hand, N addition had no influence on Cu concentrations in the five tree species, but it had significant effects on the Al and Mn concentrations in A. acuminatissima, O. pinnata and S. hancei. Specifically, N addition significantly lowered Al concentrations in the roots of A. acuminatissima in 2005 (Fig. 4).

Effects of elevated CO 2 on plant mineral elements
The concentrations of plant elements were expected to decline if the uptake of elements was not improved at the same rate as dry matter accumulation under elevated CO 2 [10]. Our results showed some declines in the concentrations of the base cations and metal cations under elevated CO 2 . However, the declines did not occur for the whole experimental time. We found no changes or even some increases in the mineral elements in plants under elevated CO 2 . Our results were consistent with other studies, which reported little or even some positive responses of mineral elements in plants to elevated CO 2 [6,26,34]. No decline in plant mineral elements could be explained by the following factors. First, the greater soil moisture content from decreased evapotranspiration under elevated CO 2 detected in our experiment [35] could stimulate soil microbial processes and then facilitate litter decomposition and mineral weathering [36]. Second, plant root growth was increased under elevated CO 2 in our experiment [37], which could improve nutrient uptake. Moreover, elevated CO 2 could indirectly increase the release of cations from the mineral weathering by enhancing carbonic acid [38], which was confirmed by the increased inorganic C leaching and higher cation concentrations in soil water under elevated CO 2 in our experiment [39,40]. Therefore, these mechanisms could be responsible for no changes or some increases in the concentrations of the mineral elements even with biomass stimulation under elevated CO 2 [37]. Across the five tree species, our results showed that leaves were less responsive to elevated CO 2 than roots with regard to the metal cations. This provided the evidence to the suggestion that elements in leaves were relatively constrained to maintain metabolic activity when compared with roots [28]. The lower Al concentrations in the roots under elevated CO 2 could be explained by the growth dilution due to the great allocation of C to root growth [37]. However, it did not appear that the growth dilution was the primary factor that influencing Al concentrations as other metal cations (Cu and Mn) did not decrease under elevated CO 2 . The decreased Al concentrations and increased Cu concentrations in the roots suggest that there would be a biological regulation of metal cations [6]. The down-regulation of Al concentrations in the roots suggests that elevated CO 2 would help plants to alleviate Al toxicity in the contaminated systems.
Compared with the other tree species, Schima superba Gardn. Champ. (S. superba) displayed a competitive advantage at biological regulation of nutrient balance under elevated CO 2 alone, given less changes in the concentrations of the mineral elements.

Effects of N addition on plant mineral elements
Our results showed that N addition led to decreases in the concentrations of base cations, especially K, and increases in Al and Mn in some tree species. Previous studies have reported the elements (e.g. Ca and Mg) in plants were lowered by N addition [24,41], which was partly consistent with our study. The shifts in the mineral elements of the seedlings could be explained by the changes in soil chemistry with increasing N inputs. High N deposition often resulted in a decline in base cations and an increase in soluble metal cations in soil solution [42]. The consequence of the decline in base cations was well reflected by the decrease in K concentrations in the five tree species in our study. Several studies have emphasized the importance of K as a co-limiting nutrient in forest ecosystems as the increased supply of other nutrients [20,43]. Our results also highlight the need to consider K limitation to plant growth under increasing N deposition. On the other hand, in the same experiment, the metal cations (Al and When considering Ca concentrations, roots responded more strongly to N addition than leaves. The results would appear to further support the argument that leaves were less sensitive indicators of soil nutrient availability than roots [28]. On the contrary, K concentrations were decreased by N addition in both the leaves and roots. This probably suggested a restricted mobilization of K from roots towards leaves when K was shortage under N addition. As K dynamics appear to be unique among the base cations (Ca and Mg) [43], further research is necessary to emphasize K cycles under increased N deposition.
When compared with the other tree species, Castanopsis hystrix Hook.f. & Thomson ex A. DC (C. hystrix) was less responsive to high N availability during the experiment. This is corresponding to no significant effects of N addition on the annual NPP of C. hystrix [44]. Further studies are needed to understand the underlying mechanisms of the adaptation of C. hystrix to increasing N deposition.

Interactive effects between elevated CO 2 , N addition and the sampling time
The mineral elements of plants in response to elevated CO 2 and N addition varied with the sampling time, as indicated by their interactions (Table 2 and S1 Table). As mentioned above, elevated CO 2 enhanced base cations in soils, and N addition resulted in the mobilization of metal cations. However, these effects had a lag time to support the faster growth under elevated CO 2 at the beginning of this experiment, thus resulting in lower mineral elements in plants. Moreover, the effects of elevated CO 2 on forests would not be sustained over time [45]. The increased leached amounts of base cations induced by elevated CO 2 in the same experiment were found to be weakening with time [40]. Thus, elevated CO 2 would lead to nutrient limitation to plant growth in the long-time. Finally, the variations between annual precipitations might also influence plant mineral elements in response to elevated CO 2 and N addition. More studies on the relationships between precipitation and the mineral elements in plants are needed.

Conclusions
Elevated CO 2 had more influences on the mineral elements in the roots than in the leaves. Elevated CO 2 did not lead to a consistent decline in plant mineral elements in this experiment. The concentrations of K, Al, Cu and Mn were increased by elevated CO 2 in some tree species. N addition led to a decrease in K across the five tree species. The response of plant mineral elements to elevated CO 2 and N addition varied with tree species, with S. superba and C. hystrix less responsive to elevated CO 2 and N addition alone, respectively. Our results have important implications on the biogeochemical cycles and species composition in subtropical forests under elevated CO 2 and N addition. In the future, the availability of K in the lateritic soils would probably constrain plant growth in response to increasing N deposition in our region.
Supporting Information S1 Data. Concentrations of mineral elements of five subtropical tree species exposed to different CO 2 and N treatments from 2005 to 2009. CK, control; NN, ambient CO 2 with N fertilizer; CC, elevated CO 2 without N fertilizer; CN, elevated CO 2 with N fertilizer. (CSV) S1 Table. Statistical results from repeated measures ANOVA on the effects of different species (S), carbon dioxide (C) and nitrogen (N) treatments and their interactions on the concentrations of mineral elements of five subtropical tree species. (DOCX)