Effects of Elevated CO2 and Temperature on Yield and Fruit Quality of Strawberry (Fragaria × ananassa Duch.) at Two Levels of Nitrogen Application

We investigated if elevated CO2 could alleviate the negative effect of high temperature on fruit yield of strawberry (Fragaria × ananassa Duch. cv. Toyonoka) at different levels of nitrogen and also tested the combined effects of CO2, temperature and nitrogen on fruit quality of plants cultivated in controlled growth chambers. Results show that elevated CO2 and high temperature caused a further 12% and 35% decrease in fruit yield at low and high nitrogen, respectively. The fewer inflorescences and smaller umbel size during flower induction caused the reduction of fruit yield at elevated CO2 and high temperature. Interestingly, nitrogen application has no beneficial effect on fruit yield, and this may be because of decreased sucrose export to the shoot apical meristem at floral transition. Moreover, elevated CO2 increased the levels of dry matter-content, fructose, glucose, total sugar and sweetness index per dry matter, but decreased fruit nitrogen content, total antioxidant capacity and all antioxidant compounds per dry matter in strawberry fruit. The reduction of fruit nitrogen content and antioxidant activity was mainly caused by the dilution effect of accumulated non-structural carbohydrates sourced from the increased net photosynthetic rate at elevated CO2. Thus, the quality of strawberry fruit would increase because of the increased sweetness and the similar amount of fruit nitrogen content, antioxidant activity per fresh matter at elevated CO2. Overall, we found that elevated CO2 improved the production of strawberry (including yield and quality) at low temperature, but decreased it at high temperature. The dramatic fluctuation in strawberry yield between low and high temperature at elevated CO2 implies that more attention should be paid to the process of flower induction under climate change, especially in fruits that require winter chilling for reproductive growth.


Introduction
The Intergovernmental Panel on Climate Change (IPCC) reported that rising temperatures, drought, floods, desertification and weather extremes will severely affect agricultural production, especially in developing countries [1]. The CO 2 concentration near the ground level has risen from 280 mmol mol 21 in the preindustrial times to the present 390 mmol mol 21 [1]. At the present rate of emission, CO 2 concentration is projected to be in the range of 500-1000 mmol mol 21 by the end of this century, which will potentially increase global temperature by 1.8-5.8uC [1]. Higher temperature individually or along with the ongoing global increase of atmospheric CO 2 could affect various physiological and morphological traits of crops that subsequently influence crop growth and final yield. As estimated by Xiong et al. [2], in China, without the CO 2 fertilization effect, grain yields of rice, wheat and maize would fall consistently if temperature rises by 2.5uC; even taking the CO 2 fertilization effect into account, the yield reductions of these crops would still occur if temperature rises by 3.9uC. Therefore, it's necessary and important to conduct research focusing on the combined effect of elevated CO 2 and increased temperature on crop yield. Strawberry (Fragaria 6 ananassa Duch.) is one of the most important fruit crops that is widely planted in North America, Mediterranean Europe, Southwest Asia, and Australia [3]. Shortened photoperiod and low temperature are known to induce flower formation for June-bearing strawberries [4]. Kumakura and Shishido [5] suggested that maximum strawberry yields are associated with a narrow range of temperatures between 15 and 20uC. The yield is reduced when the day temperature exceeded 25uC, even if the diurnal mean temperature is maintained below 20uC. Therefore, in the event of increased temperatures due to global warming, strawberry production would be severely affected. Currently, there is little knowledge of the combined effects of high temperature and elevated CO 2 on strawberries or other crops, although published data suggest that such interaction is critical. Chen et al. [6] reported that elevated CO 2 levels greatly improved yield and fruit quality of strawberry by increasing the total fruit number per plant, average fruit fresh weight, dry matter content, fruit total sugars and sugar/acid ratio. On the contrary, combined effect of elevated CO 2 and temperature on other C 3 crops such as rice, soybean, dry bean, peanut, cowpea, wheat and cotton cultivated in different growth conditions, including growth chambers, open-top chambers and plastic tunnels, showed no beneficial effect on yield [7][8][9][10][11][12][13]. However, strawberries require much lower temperature than these crops, and it is important to test whether elevated CO 2 will ameliorate the negative effects of the increased temperature on its reproductive development.
Nitrogen is one of the most important resources limiting plant growth and seed production in natural and agricultural ecosystems [14]. An increase in carbon availability due to elevated CO 2 may enhance nitrogen limitation, leading to a reduction in plant nitrogen concentration [15]. Studies on spring wheat and rice suggested that under elevated CO 2 concentration, nitrogen fertilization had important influence on the maintenance and continuing increase of crop yield [16][17][18]. The deeper and larger root system with nitrogen fertilization, which is of benefit to the use of soil moisture and nutrient, is thought to be the reason of continuing increase of crop yield [16]. Deng and Woodward [19] reported high CO 2 increased the strawberry fruit yield by 42% at high nitrogen supply and 17% at low nitrogen supply through an increase in flower and fruit number of individual plants. However, they did not analyze the effect of high temperature, elevated CO 2 and nitrogen supply on strawberry fruit production and quality.
Strawberries are a good source of natural antioxidants [20]. In addition to the usual nutrients, such as vitamins and minerals, strawberries are also rich in anthocyanins, flavonoids, and phenolic acids [20]. Strawberries have shown a remarkably high scavenging activity toward chemically generated radicals, thus making them effective in inhibiting oxidation of human lowdensity lipoproteins [20]. At elevated CO 2 , decrease, no change, and an increase in fruit antioxidant activity have been reported [21][22][23][24]. Levine and Paré [24] showed in scallions that both, total phenol and total antioxidant activity decrease under elevated CO 2 . They suggested that besides species differences, in the absence of stress, plant grew with minimum investment in antioxidant compounds to maintain a basal defense level under elevated CO 2 . Contrastingly, the increase of fruit antioxidant activity may stem from the reduction of fruit nitrogen concentration induced by the elevated CO 2 . As a 'physiological trade-off', the amount of secondary metabolites like phenolics increases at low nitrogen to maintain the growth-differentiation balance (GDB) framework [25]. Further, the antioxidant activity of plant tissues also increases as reactive oxygen species (ROS) that are involved in the signaling and perception of nitrogen deficiency increase [26].
Due to the prediction of climate change, a number of studies have examined the effects of rising CO 2 and/or temperature on yield characteristics, notably quantity and nutrition of food crops [7][8][9][10][11][12][13][16][17][18]. However, almost nothing is known regarding the concurrent interaction of CO 2 , temperature and nutrition (e.g. N) on reproductive biology of fruit crops. This is the first study to undertake an assessment of these potential interactions. Further, unlike the crops that have been studied, the temperature requirement for strawberry cultivation is quite low and the response of strawberry to the increased temperature may therefore be different to other crops. Therefore, we assessed the combined effects of CO 2 concentration, air temperature and nitrogen application on the fruit yield and quality of strawberry. Firstly, we examined the fruit yield under these abiotic factors, and tested whether elevated CO 2 can modify the response of fruit yield to elevated temperature. The effects of nitrogen supply on the response to fruit yield at elevated CO 2 concentration and temperature were also studied. Secondly, we examined the combined effects of CO 2 concentration, temperature and nitrogen supply on fruit quality such as carbohydrate accumulation, nitrogen content and antioxidant levels.

Variation in Fruit Weight and Yield
The abbreviations for the combined treatments of different CO 2 concentrations, temperatures and nitrogen concentrations reported below are explained in Table 1.Elevated CO 2 increased fruit yield (viz. total fruit dry weight per plant) at low temperature, but deceased it at high temperature, when compared to the corresponding treatments in ambient CO 2 (Figure 1a). The greatest fruit yield was in high CO 2 , low temperature and low nitrogen treatment (C), while the least was in high CO 2 , high temperature and high nitrogen treatment (CTN). The plants grown at low nitrogen concentration had greater yield than those grown at high nitrogen concentration, except for plants grown in low CO 2 , low temperature and low nitrogen treatment (ck). Similarly, elevated CO 2 increased fruit number per plant (FN) at low temperature, but decreased it at high temperature, when compared to the corresponding treatments in ambient CO 2 (Figure 1b). High nitrogen decreased FN at high temperature, but had no effect on FN at low temperature. FN in high CO 2 , high temperature and low nitrogen treatment (CT) was 1.86 times greater than in CTN treatment. Meanwhile, FN was 2.32 times greater in low CO 2 , high temperature and low nitrogen treatment (T) than in low CO 2 , high temperature and high nitrogen treatment (TN). The greatest FN was found in T treatment, which was significantly greater than in other treatments.
Fruit dry weight (FDW) varied significantly (P,0.0001) both between treatments and between individual plants, thus the healthy fruits were graded in three size classes (grade 1,0.4 g, 0.4# grade 2#0.7 g, grade 3.0.7 g; Figure 1c, 1d and 1e). Frequency distribution (FD) of the fruits in these three grades varied among all treatments. Elevated CO 2 increased FD in grade 1 at high temperature, but deceased it at low temperature, when compared to the corresponding treatments in ambient CO 2 ( Figure 1c). Contrastingly, elevated CO 2 decreased FD in grade 3 at high temperature, but increased it at low temperature, when compared to the corresponding treatments in ambient CO 2 (Figure 1e). High nitrogen decreased FD in grade 1, but increased it in grade 3. The relatively lower fruit number and the highest FD in grade 1 in CT and CTN treatments eventually decreased fruit yield, when compared with the corresponding treatment in ambient CO 2 . While, the highest fruit yield in C and high CO 2 , low temperature and high nitrogen (CN) treatments resulted from the lowest FD in grade 1, second greatest FD in grade 3 and second greatest FN.
A linear regression was performed to compare slopes of relationships between FDW and total achene number (TAN) on the surface of fruit (Figure 2a, 2b, 2c and 2d). The slopes of FDW versus TAN appeared greater for high CO 2 treated plants than low CO 2 treated plants, while the slopes decreased at high nitrogen when averaged over the other factors (Table 2). Pooling TAN across all treatments revealed that the correlation coefficient between TAN and FDW was significant (r 2 = 0.698, P,0.001) and higher than the correlation coefficient between total fertilized achene number (TFA) and FDW (r 2 = 0.506, P,0.001, Figure 3). However, the relationship between total aborted achene number (TAA) and FDW was very week (r 2 = 0.138; Figure 3), and these different patterns suggested that the aborted achenes were not the major limitation to FDW.

Variation in Taste and Health-related Compounds
Compared to the corresponding treatments in ambient CO 2 , elevated CO 2 decreased the antioxidant compounds and total antioxidant capacity (in simple terms, antioxidant activity) of strawberry fruit in both high-temperature and low-temperature treatments ( Table 3). As expected, the response of antioxidant capacity to the CO 2 and temperature treatments was altered by nitrogen application, which increased at elevated CO 2 but decreased in ambient CO 2 with increasing nitrogen supply ( Table 3). CO 2 and nitrogen both significantly affected the total antioxidant capacity and all antioxidant compounds in strawberry fruit (Table 4). Compared to the corresponding treatments in ambient CO 2 , anthocyanin (AC) content decreased 27% in CT treatment and 48% in C treatment, but decreased only 1% and 4% in CTN and CN treatments, respectively (Table 3). There were significant CO 2 -temperature-nitrogen (C6T6N), CO 2 -temperature (C6T), CO 2 -nitrogen (C6N), and temperature-nitrogen (T6N) interactions affecting AC ( Table 4). The treatment effects on total phenolics (TP) closely matched that of AC. Strawberry fruits showed a 27% and 21% decline in TP levels in CT and C treatments, respectively, but decreased only 8% in CTN treatment and 10% in CN treatment, when compared to the corresponding treatments in ambient CO 2 (Table 3). There were significant CO 2 and nitrogen main effects, and significant C6T, C6N and T6N interactions (Table 4) affecting TP levels. Total flavonoid (TF) decreased 31% and 36% in CT and C treatments, respectively, but decreased only 13% in both CTN and CN treatments, when compared to the corresponding treatments in ambient CO 2 ( Table 3). Besides significant CO 2 , temperature and nitrogen main effects, all interactions also affected the TF levels under various treatments (Table 4). Total antioxidant capacity measured using the free radical 2, 2-diphenyl-1-picrylhydrazyl (DPPH) method decreased 28% in CT treatment and 20% in C treatment, but decreased 12% and 13% in CTN and CN treatments, respectively (Table 3). Comparatively, total antioxidant capacity measured using the 2, 29-azino-bis (3-ethylbenzothiazoline-6-sulphonic acid) (ABTS) method, which closely matched DPPH, decreased approximately 19% and 18% in CT and C treatments, respectively, but decreased 12% in CTN treatment and 8% in CN treatment, when compared with the corresponding treatments in ambient CO 2 (Table 3). There were significant C6T6N, C6T, and C6N interactions affecting DPPH levels (Table 4). Similarly, all interactions had significant effects on the levels of ABTS in different treatments (Table 4). Fruit nitrogen contents (FNC) at elevated CO 2 were similar to the corresponding treatments in ambient CO 2 , except the one in C treatment (Table 3). Contrastingly, high temperature increased the levels of FNC among all treatments with only an exception of T treatment. Not surprisingly, nitrogen application significantly increased the levels of FNC. There were only significant nitrogen and temperature main effects on FNC level ( Table 4).
The concentrations of three main sugars (viz., fructose, glucose and sucrose) were also determined for each treatment; fructose and glucose were quantitatively the most important in this study. The contents of fructose and glucose at elevated CO 2 were almost 1.3 times (1.29 and 1.35 times, respectively) higher than in ambient CO 2 regardless of temperature and nitrogen treatments (Table 5). There was no significant difference in sucrose concentration in different treatments. There was significant effect of CO 2 , temperature, and C6T, C6N and C6T6N interactions on fructose concentration (P,0.05), whilst only CO 2 , C6T and C6T6N interactions significantly affected glucose concentration in different treatments (Table 6). Total sugars per gram fresh weight (TSW) averagely increased 43% under elevated CO 2 regardless of temperature and nitrogen treatments (Table 5), and CO 2 had a significant effect on TSW when compared to other factors (Table 6). Despite differences in the sugar distribution among the treatments, the ranking of sweetness index (SI) was similar to the ranking of total sugars (per fresh weight) from 86.4 to 128.8 relative units. CO 2 effect was significant as it resulted in a 49%, 38%, 45% and 36% increase in SI in CT, CTN, C and CN treatments, respectively (Table 5, 6), when compared to the corresponding treatments in ambient CO 2 .

Variation in Fruit Number, Weight and Yield of Strawberry
Fruit yield of strawberry per plant is composed of fruit dry weight (FDW) and fruit number (FN), while FDW is affected by the total achene number (TAN) and dry matter accumulated per achene (DMA), since achenes (actual seeds) are considered to be involved in regulating strawberry fruit development [27]. Therefore, the treatment effects on either FDW or FN will highlight the effect on fruit yield under those treatments.
Compared to the corresponding treatments in ambient CO 2 (T and TN treatments), elevated CO 2 further reduced the fruit yield at high temperature (CT and CTN treatments). Yield reductions, which were further enhanced by elevated CO 2 at high temperature during flowering and fruit development, also have been documented in other crops such as rice, wheat, grain sorghum, kidney bean, dry bean, soybean, peanut and tomato, though the extreme temperatures were much higher than the one used in this study [7,[9][10][28][29][30][31]. Commonly, the increased seed abortion caused by decreased pollen production [30], lower pollen reception by stigma due to anther indehiscence [32], and lower pollen viability due to degeneration of tapetum layer and decreased carbohydrate metabolism [33][34][35] during flower development and opening, resulted in the reduction of crop yield at high temperature. The exact mechanism of the increased susceptibility of these processes to high temperature at elevated CO 2 is still unclear, but the small increase in tissue temperatures (owing to decreased leaf conductance) which reduces the ceiling temperatures for seed-set by about 2uC is one possible explanation [9]. However, in this study, achene abortion (as seed abortion in other crops) caused by the negative impacts of warmer tissue temperatures on flower development and opening were insufficient to explain the reduction of strawberry yield under high CO 2 concentration. Nitsch [36] reported that fruit production of strawberry was proportional to the extent of achene fertilization, and strawberry fruit size was positively related to the number of   fertilized achenes. Thus, if the yield reduction at high temperature and elevated CO 2 was caused by achene abortion, the correlation between FDW and total number of achenes (TAN) will be lower than the correlation between FDW and total number of fertilized achenes (TFA), because aborted achenes will decrease the accumulation of dry matter. However, the correlation between FDW and TFA was 28% lower than the correlation between FDW and TAN, which suggested that the adverse effects of elevated CO 2 and high temperature on achene fertilization were not the major causes of yield reduction. In other words, TAN, rather than TFA, strongly correlating to FDW implied that a possible regulation mechanism existed in strawberry fruit responding to the occurrence of achene abortion. When achene abortion occurs, the remnant fertile achenes may be stimulated to increase the capacity of dry matter accumulation to offset the reduction of aborted achenes. The observation in Figure 4 indicated that even great achene abortion rates occurred the fruits with similar fruit size generally had the similar fruit weight, which also confirmed achene abortion having rather limited effect on FDW and then yield reduction. In addition, the slopes of linear regression which indicated the dry matter accumulated per achene (DMA) increased at elevated CO 2 and low nitrogen, but have no benefit in maintaining fruit yield at high temperature and elevated CO 2 when yield reduction occurred. Therefore, we propose that achene abortion was not the main cause of yield reduction, and the increased DMA at elevated CO 2 also has no benefit in maintaining fruit yield at high temperature. Indeed, TAN, which was determined by flower induction, affected the change of FDW and contributed greatly to the reduction of fruit yield at elevated CO 2 and high temperature. Besides TAN, the inflorescences of strawberry were handpollinated to keep to a minimum inflorescence abortion, thus FN was also determined by flower induction. Therefore, as an alternative explanation, the reduction of strawberry yield in CT and CTN treatments was mainly caused by the inhibition of flower induction, which could be suppressed by high temperature and other environment factors [37]. In other words, the fewer number of inflorescences and the smaller umbel size of strawberry during flower induction resulted in the reduction of fruit yield at elevated CO 2 and high temperature. Low temperature, as one of many environment factors, can affect flower induction in plant through many physiological pathways, including vernalization and gibberellin (GA) biosynthesis [37]. Exposure to the prolonged cold of winter, through this process called vernalization, is required to permit flowering of June-bearing strawberry plants [38][39]. However, the promotion of flowering by vernalization could be reduced or even completely suppressed by high temperature [37], and devernalization by high day temperature over-riding the effect of low night temperature which induced flower bud initiation has been found in some cultivars of strawberry [40]. In this study, the warmer tissue temperatures in day time at elevated CO 2 may cause devernalization of the plants, and eventually inhibit flower Table 3. Effects of carbon dioxide, temperature and nitrogen treatments on anthocyanin (AC), total phenolic (TP), total flavonoid (TF), DPPH radical scavenging assay (DPPH), ABTS radical scavenging assay (ABTS) and fruit nitrogen content (FNC) of strawberry fruits a .  induction. Besides the vernalization pathway, GA biosynthesis and signaling, including genes such as GA 1 , GAI, RGA, FPF 1 and AtMYB 33 , plays an important role in flower induction in Arabidopsis [37]. It was suggested that temperature effect may be mediated by changes in the level of active endogenous GA s [41]. Su et al. [42]reported that the flowering shoots of Phalaenopsis hybrida grown under high temperature contained lower levels of GA 1 , GA 19 , GA 20 and GA 53 than GA 3 -treated and cold-induced plants. They also found relatively low level of GA 1 and high level of GA 8 in shoot-tips of warm control (non-flowering) plants compared to plants whose flowering was promoted with GA 3 or cooltemperatures. Tayor et al. [43] studied the possible role of endogenous GAs in the control of flowering in strawberry and identified eight 13-hydroxylated GAs from leaf tissues of the shortday cv. Elasnta. Thus, the change of endogenous GAs biosynthesis induced by warmer tissue temperatures may be another reason of inhibition of flower induction at elevated CO 2 . Surprisingly, nitrogen application further decreased fruit yield in CTN treatment than in CT treatment. Though, nitrogen application greatly improved fruit size, the FN level at high temperature was reduced to the extent of nearly a half of low nitrogen treatments. This reduction implied that the induction of inflorescences was greatly reduced at high nitrogen, but umbel size increased. As reported, increase of mineral supply to the roots delayed flowering in several mutants of the photoperiod and autonomous pathways, as well as in wild-type plants in Arabidopsis [44]. An important part of this inhibition was presumably due to nitrogen [45]. Corbesier et al. [46] reported high nitratesupplement reduced the export of sucrose towards the shoot apical meristem at floral transition, and the decrease of FN in high nitrogen treatment in this study is probably caused due to this reason.
Our study suggests that, as a fruit tree, strawberry is idiographic with a highly sensitive requirement of a narrow range of temperature for flower induction. We propose that reduced flower induction plays an important role in the reduction of strawberry yield at high temperature and elevated CO 2 . Further, the adverse effect of high temperature and elevated CO 2 on fruit yield were not ameliorated but rather exacerbated under high nitrogen condition. However, nitrogen supply did improve fruit quality by increasing the fruit weight.

Variation in Taste-and Health -related Compounds
The increased dry matter-content (DMC) of the fruits was probably due to the increased non-structural carbohydrates sourced from the increased net photosynthetic rate of strawberry at elevated CO 2 [47][48]. The non-structural carbohydrates including fructose (the dominant sugar), glucose and sucrose, contribute directly to the perceived sweetness of the fruit, and these sugars account for more than 990 g kg 21 of the total sugars in ripe strawberries [49]. Therefore, elevated CO 2 which increased fructose, glucose and total sugar levels relative to other taste related compounds would improve the perception of fruit sweetness. Table 5. Effects of carbon dioxide, temperature and nitrogen treatments on fructose (Fru), glucose (Glu), sucrose (Suc), total sugars (TSW), sweetness index (SI) and dry matter-content (DMC) of strawberry fruits a .  At elevated CO 2 , decrease in tissue nitrogen content has been widely reported, but there was significant variation in different taxa [50]. In this study, fruit nitrogen content (FNC) decreased nearly 11% at elevated CO 2 and this value was in the range of the reduction of seed nitrogen content (15%) at elevated CO 2 [15]. Dilution hypothesis suggests that the decrease in tissue nitrogen content under elevated CO 2 results from the dilution due to accumulation of non-structural carbohydrates or plant secondary compounds [51]. In this study, the decline in FNC at elevated CO 2 may be caused by dilution effect of accumulated nonstructural carbohydrates, since elevated CO 2 greatly increased leaf photosynthesis and accelerated the accumulation of these compounds. At elevated CO 2 , the total antioxidant capacity and all antioxidant compounds in strawberry fruits decreased nearly 27.5% (from 19% to 37%) at low nitrogen and 9.5% (from 3% to 13%) at high nitrogen, whilst DMC increased 23.7% and 12.5% in the corresponding treatments, respectively. The increase of DMC was proportional to the decrease in total antioxidant capacity and all antioxidant compounds at elevated CO 2 , which implied that the reduction of total antioxidant capacity and all antioxidant compounds in strawberry fruits were mainly caused by the dilution effect of accumulated non-structural carbohydrates, though the dilution effect on individual antioxidant compounds varied.
The extent of decrease in total antioxidant capacity and antioxidant compounds was greater at low nitrogen than at high nitrogen, implying that nitrogen application greatly modified the treatment effect of elevated CO 2 on these compounds. From the results, the greater decrease of antioxidant activity at low nitrogen mainly came from the higher antioxidant levels in ambient CO 2 and lower antioxidant levels at elevated CO 2 , when compared to these compounds at high nitrogen. Commonly, the change of antioxidant activity results from the change of ROS, and these antioxidant compounds were evolved to protect plants from oxidative damage [52]. It is known that environment stresses, including nitrogen starvation [25], may increase the production of ROS. We have summarized four possible causes of the increase of ROS under nitrogen deficiency including: (1) ''physiological tradeoff'' between plant growth and secondary metabolite production in GDB framework [25][26]53]; (2) accelerated senescence of plant tissues or organs [54][55]; (3) limitation of CO 2 uptake efficiency and accumulation of reducing power due to accumulation of H 2 O 2 in nitrogen deficient plants, which is known to decrease stomatal opening [56][57]; (4) surplus electron flow leading to enhanced oxygen photo-reduction in the chloroplast via the Mehler reaction as the ratio of Rubisco activity declined under nitrogen deficiency [58]. In this study, the increased antioxidant activity in strawberry fruit at low CO 2 concentration and low nitrogen treatments could not be explained satisfactorily with the reasons mentioned above except the first one. Obviously, reason 3 and 4 were not suitably explained in fruit, while reason 2 contrasted with recent research that elevated CO 2 accelerated senescence of plant tissues or organs and would increase antioxidant level in them [59][60][61]. Therefore, reason 1 will be a possible explanation that secondary metabolites such as phenolics are accumulated at low nitrogen [25]. Meanwhile, ROS which is involved in the signaling and perception of nitrogen deficiency is also increased [26]. The antioxidant levels decreased in CT and C treatments (though the extent was rather small) suggesting that the effect of nitrogen deficiency on antioxidant level has been modified by the elevated CO 2 . We speculate that the reduced FNC in these treatments may inhibit the activity and amount of relevant enzymes involved in perception of nitrogen deficiency and synthesis of secondary metabolites, and negatively affect the antioxidant levels.

Conclusions
Overall, our study illustrates the combined effects of elevated CO 2 , nitrogen and temperature on strawberry yield and quality. At low temperature, elevated CO 2 greatly improved the fruit yield by increasing fruit number and fruit weight. However, at high temperature, elevated CO 2 decreased fruit yield. This decrease was mainly caused by the fewer induced inflorescences and smaller induced umbel size which eventually reduced fruit number and fruit weight, respectively. Moreover, elevated CO 2 increased the levels of dry matter-content, fructose, glucose, total sugar and sweetness index per dry matter, but decreased fruit nitrogen content, total antioxidant capacity and all antioxidant compounds per dry matter in strawberry fruit. The reduction of fruit nitrogen content and antioxidant activity was mainly caused by the dilution effect of accumulated non-structural carbohydrates sourced from the increased net photosynthetic rate during fruit development. Thus, the quality of strawberry fruit would increase because of the increased sweetness and the similar amount of fruit nitrogen content, DPPH, ABTS and all antioxidant compounds per fresh matter at elevated CO 2 . Interestingly, nitrogen application had no beneficial effect on the fruit yield, but greatly increased fruit weight among all treatments. Fruit quality such as antioxidant activity increased at high nitrogen and elevated CO 2 , but decreased at high nitrogen and low CO 2 . Considering all treatment effects, we conclude that elevated CO 2 improved the production of strawberry (including yield and quality) at low temperature, but decreased it at high temperature. In addition, the dramatic fluctuation in strawberry yield between low and high temperature at elevated CO 2 implies that more attention should be paid to the process of flower induction under climate change especially in fruits that require winter chilling for reproductive growth, as chronic and steady reduction in winter chill is expected [62]. Therefore, efforts should be made to develop cultivars that require less winter chill for future climate.

Plant Material and Experimental Design
Four large growth chambers with an internal chamber height of 2.20 m and a growth area of 1.0 m 2 were used for the experiment. All chambers have air temperature, relative humidity and carbon dioxide control. Photosynthetic active radiation (PAR) was about 600 mmol m 22 s 21 , and relative humidity was controlled at 80% by an air humidifier 24 hours a day. CO 2 was injected automatically into the chambers all day and night, and its concentration was controlled using a CO 2 delivery system and chamber vents. An individual LICOR infrared gas analyzer (LI-800 GasHound CO 2 Analyzer, LI-COR, Nebraska, USA) was used to monitor the CO 2 levels for each chamber independently, and the accuracy of the analyzer was 62%.
The experimental design consisted of a three-way randomized block with four replications. The treatments consisted of two day/ night temperature levels [20/15uC (T A ), 25/20uC (T A +5uC)], two CO 2 concentrations [360 and 720 mmol CO 2 mol 21 air], and two nitrogen application levels [0% (distilled water) and 0.01% NH 4 NO 3 ]. The temperature and CO 2 treatments were randomly allocated in each of the four growth chambers as follows: N Chamber 1-T A +5uC and 360 mmol CO 2 mol 21 , N Chamber 2-T A +5uC and 720 mmol CO 2 mol 21 , N Chamber 3-T A and 360 mmol CO 2 mol 21 , N Chamber 4-T A and 720 mmol CO 2 mol 21 .
Fifty milliliter of 0.01% NH 4 NO 3 solution was applied twice a week per plant at the beginning of 1 December 2010 and lasted for nearly 6 months. A fixed day length of 10 h from 7:00 AM to 17:00 PM, which corresponds to the day length of early spring in Zhejiang, was used.
The strawberry cultivar used in this study was Toyonoka (Fragaria 6 ananassa Duch. cv. Toyonoka) a short-day cultivar which need short-day and low temperature (chilling) treatments to accelerate flower bud initiation [63][64], and now is widely planted in Zhejiang. Strawberry seedlings were planted in 25 cm618 cm pots using field soil (red soil, total nitrogen content 0.96 g/kg dry soil). Prior to the treatments in chambers, plants grew under the ambient autumn temperatures of Jinhua, Zhejiang, in an unheated greenhouse from November to December for one month (chilling and short-day treatments), and the mean daily temperature in November was about 13.2uC. All plants were watered daily and fertilized weekly with 150 ml per plant of Peters fertilizer (20:20:20, N/P/K). Plants with similar height and crown diameter were moved to chambers and 8 pots were placed in each chamber and four pots per treatment. The plants in each chamber were rotated inside chambers per week and between chambers per month to reduce the microclimate effects of different chambers. Blossoms were self-pollinated by hand using a small brush. As daily routine, the ripeness of fruit was determined by color, and firm red-ripe fruits free from defects or decay were harvested from each growth chamber during the fruiting stage. Fruit dry weight, fruit number, total achenes and total aborted achenes were determined. All of berries were graded in three size classes (grade 1,0.4 g; grade 2, 0.4-0.7 g; grade 3.0.7 g) according to FDW. The berries of each plant were cut into small slices, mixed, and frozen at 224uC for analyzing until the end of the harvest season.

Fruit Sample Preparation
To prepare the fruit samples, four 100 g samples of berries from four replicates of each treatment were homogenized for 2 min in a rotating blade homogenizer (Midea, JP351, China). Solution of homogenate extract (2 g) in methanol (25 ml) was used for determination of total flavonoid, total phenolic, DPPH and ABTS. Solution of homogenate extract (2 g) in distilled water (25 ml) was used for determination of anthocyanin content. All compounds mentioned above in each sample from each plant were measured in triplicate and four samples of each treatment were determined.

Determination of Antioxidant Compounds Content
The amount of all the antioxidant compounds was determined according to Zheng et al. and Lu et al. [65][66][67][68]. The total flavonoid content was determined by a colorimetric assay with modifications. Briefly, 0.5 mL extract solution was separately mixed with 1.5 mL of methanol, 0.1 mL of 2% aluminum chloride, 0.1 mL of 1 M potassium acetate, and 2.8 mL of distilled water, and left at room temperature for 30 min. The absorbance of the reaction mixture was measured at 415 nm using a UV-vis spectrophotometer (Jinghua, JH752, China). The total flavonoid content was expressed as rutin equivalents in milligrams per gram dry weight of strawberry.
The total phenolic content was determined colorimetrically using Folin-Ciocalteau reagent, with modifications. The total phenolic assay was conducted by mixing 8.25 mL of deionized water, 0.5 mL of extract, 0.75 mL of 20% Na 2 CO 3 , and 0.5 mL of Folin-Ciocalteu reagent. After 40 min of reaction in a water bath at 40uC, the absorbance at 755 nm was measured using a spectrophotometer. Results were expressed as gallic acid equivalents milligrams per gram of dry weight of strawberry.
The total anthocyanins content was determined with a modified pH differential method, using two buffer systems: potassium chloride 0.025 M at pH 1.0 and sodium acetate 0.4 M at pH 4.5. Briefly, 1 mL of sample was transferred to a 10 mL volumetric flask and made up with each buffer. The absorbance of each equilibrated solution was then measured at 510 and 700 nm, using a UV-vis spectrophotometer. Quartz cuvettes of 1 cm path length were used, and all measurements were carried out at room temperature (25uC). Absorbance readings were made against distilled water as a blank. The total anthocyanins content was calculated on the basis of cyanidin-3-glucoside with a molecular weight of 445.2 g/mol and an extinction coefficient of 29600 L/ mol ? cm, as Where MW is the molecular weight of cyanidin-3-glucoside, DF is the dilution factor, L is the path length in cm, and e is the molar extinction coefficient for cyanidin-3-glucoside. Results were expressed as milligram cyanidin-3-glucoside equivalents per gram of dry weight of strawberry.

Determination of Total Antioxidant Capacity (DPPH and ABTS)
The DPPH free radical scavenging activity was evaluated according to the method of our previous study [65][66]. The extracts (0.1 mL) of strawberry in ethanol were reacted with 10 mL of 0.03 g/L DPPH ethanol solution at room temperature. The extract (0.1 mL) with 10 mL distilled water was used as control. The absorbance was measured at 517 nm after 30 min of reaction in the dark. DPPH radical scavenging capacity was expressed as Trolox equivalent antioxidant capacity (mmol of Trolox/1 g of dry strawberry fruits).
The ABTS assay was based on the method of Re et al. [69] with slight modification. ABTS N+ reagent was produced by reacting 10 mL of 7 mM ABTS solution with 178 mL of 140 mM potassium persulfate aqueous in the dark at room temperature for 13 h before use. The ABTS N+ solution was diluted with ethanol to appropriate absorbance. One-tenth of a milliliter of extract was added to 3.9 mL of diluted ABTS N+ solution to react in the dark at room temperature for 6 min, and the absorbance at 732 nm was recorded. Trolox was used as standard with the final concentration ranging from 0 to 16.5 mM. Results were expressed as Trolox equivalent antioxidant capacity (mmol of Trolox/1 g of dry strawberry fruits).

Determination of Fruit Nitrogen Content
Fruit nitrogen content was determined by mico-Kjeldahl digestion method [70], with modifications. Briefly, 0.5 g dry fine powder of strawberry fruit was accurately weighted into mico-Kjedlahl flasks to which the catalyst mixture (0.3% TiO 2 , 0.3% CuSO 4 , and 10% K 2 SO 4 on a weight basis) and concentrated sulfuric acid (10 mL) were added. The digests were heated for 1.5 h beyond the point when the solutions had cleared. They were then cooled and diluted to 50 mL with distilled water. After addition of 3 ml of 20 g/L H 3 BO 3 solution in the inner chamber of a clean Conway dish, 4 mL diluted digest was added in the outer chamber. The covered Conway dishes were sealed and incubated at 40uC for 24 h. The absorbed ammonia in H 3 BO 3 solution was titrated with 0.02 mol/L HCl solution. The results were expressed as milligram per gram of dry weight of strawberry. Each sample from each plant was measured in triplicate and four samples of each treatment were determined.

Analysis of Sugars Using HPLC
For analysis of sugars, 10 g of snap-frozen strawberry powder (wet) were stirred by a magnetic stirring apparatus in 100 ml of extraction solution containing 90 ml of distilled water, 5 ml of 1 mol l 21 zinc acetate and 5 ml of 0.25 mol l 21 potassium ferrocyanide for 30 min at room temperature. The solution was filtered through a membrane-filtered supernatant (Ø 0.26 mm). Glucose, fructose and sucrose were analyzed by injection of a 50 ml sample volume into a DuoFlow HPLC system (Bio-RAD, USA) using a Sepax Amethyst-Amino column, 250 mm64.6 mm diameter, 5 mm particle size (Sepax, USA; Part no. 322305-4625). The column temperature of 20uC was controlled and an acetonitrile: pure water solution (80:20 v/v) was used as mobile phase (flow rate 0.8 ml min 21 ). Carbohydrates were detected with a refractive index detector (RID-10A, Japan) and their concentrations were calculated by comparing sample peak area to standards using OriginPro 8.5 software. Each sample from each plant was measured in triplicate and four samples of each treatment were determined. The results were recalculated per dry mass.
The sweetness index was calculated by multiplying the sweetness coefficient of each individual sugar (glucose = 1, fructose = 2.3 and sucrose = 1.35), as described by Keutgen and Pawelzik [71].

Statistical Analyses
Data in this study were subjected to analysis of variance, and means were compared by least significant difference (LSD). Multivariate general linear model function (MGLM) was performed to analyze the main effects of CO 2 concentration, air temperature and nitrogen input combined with their interactions on the quality of strawberry growing in chambers. Regression analysis was conducted to examine relationships between fruit dry weight and total achene number. In this study, all statistical analyses were conducted using SAS software (SAS Institute Inc., Cary, NC, USA).