Phosphorous fertilization alleviates drought effects on Alnus cremastogyne by regulating its antioxidant and osmotic potential

Alnus cremastogyne, a broad-leaved tree endemic to south-western China, has both commercial and restoration importance. However, little is known of its morphological, physiological and biochemical responses to drought and phosphorous (P) application. A randomized experimental design was used to investigate how drought affected A. cremastogyne seedlings, and the role that P applications play in these responses. Drought had significant negative effects on A. cremastogyne growth and metabolism, as revealed by reduced biomass (leaf, shoot and root), leaf area, stem diameter, plant height, photosynthetic rate, leaf relative water content, and photosynthetic pigments, and a weakened antioxidative defence mechanism and high lipid peroxidation level. However, the reduced leaf area and enhanced osmolyte (proline and soluble sugars) accumulation suggests drought avoidance and tolerance strategies in this tree. Applying P significantly improved the leaf relative water content and photosynthetic rate of drought-stressed seedlings, which may reflect increased anti-oxidative enzyme (superoxide dismutase, catalase and peroxidase) activities, osmolyte accumulation, soluble proteins, and decreased lipid peroxidation levels. However, P had only a slight or negligible effect on the well-watered plants. A. cremastogyne is sensitive to drought stress, but P facilitates and improves its metabolism primarily via biochemical and physiological rather than morphological adjustments, regardless of water availability.


Results
Morphological traits. Overall, a strong reduction was observed in the morphological traits of droughtstressed plants when compared with their well-watered counterparts ( Table 1). Irrespective of the P application, there was a significant reduction in leaf biomass (38.29%), shoot biomass (65.73%), leaf area (50.32%) and stem diameter (24.88%) of A. cremastogyne under drought stress compared with well-watered conditions. P application slightly increased the leaf biomass, shoot biomass, root biomass and height under drought stress in comparison with the well-watered plants, but these differences were not significant. Similarly, P application did not have any significant effects on the morphological traits of the well-watered plants. The ration of leaf:root biomass was lower in the plants well-watered (c. 0.47) than in those under drought stress (c. 0.65), irrespective of the P application. Therefore, P fertilization did not change dry mass partitioning in this tree species' seedlings. Leaf relative water content, gas exchange, and chlorophyll fluorescence. Leaf relative water content (LRWC) of A. cremastogyne significantly decreased under drought stress by 32.61% when compared with the well-watered treatment, irrespective of the P application ( Table 2). The P application significantly increased LRWC under drought, by 25.09%, when compared with its counterpart (−P); however, it had no effect on the well-watered plants. Drought significantly decreased the net CO 2 assimilation rate (P n ) (78.14%) in comparison with the well-watered plants; however, the P application significantly increased P n under both watering treatments. Moreover, there was a significant reduction in the intercellular CO 2 concentration (C i ), stomatal conductance (G s) , transpiration rate (E) and maximum quantum efficiency of photosystem II (F v /F m ) in the plants drought-stressed when compared with those well-watered, irrespective of the P application. However, P application did slightly increase the above mentioned leaf physiological traits, though no significant differences were detected.
Photosynthetic pigments. Photosynthetic pigments concentrations chlorophyll a (Chl a), chlorophyll b (Chl b), and carotenoids (Car) were significantly lower in the plants drought-stressed than in those well-watered, regardless of the P application (Fig. 1). Phosphorous fertilization did not significantly increase the Chl a and Chl b concentrations under drought stress when compared with the well-watered treatment. However, the P application apparently had no effect on Car concentration in the drought-stressed plants.
Antioxidative defence system. Lipid peroxidation was significantly higher in the drought-stressed plants than in the well-watered non-fertilized plants (Fig. 2). P application significantly decreased malondialdehyde (MDA) contents by 47 Under drought stress, the activity of SOD, POD, and CAT was similar to that in the well-watered non-fertilized plants. However, applying P significantly increased both POD (22.58%) and SOD (103.81%) activities under drought-stressed plants when compared with their non-fertilized counterparts. The P application also increased CAT activity in the drought-stressed plants, by 50%, but this difference was not significant (Fig. 3).
Osmolytes and soluble proteins. Concentrations of both proline and soluble sugars (SS) were higher, whereas the soluble proteins (SP) was lower, in the plants drought-stressed when compared with those well-watered (Fig. 4). P application under drought significantly increased the proline (34.41%), SS (18.4%) and SP (68.3%) concentrations in the plants when compared with their non-fertilized counterparts. However, the P application had no significant effects on the above biochemical traits of the well-watered plants.

Discussion
Plant growth. Drought stress is the primary abiotic threat to plant growth around the world, especially under arid and semiarid conditions. Permanent or temporary drought events severely hamper plant growth and development and also restricts nutrient uptake by plant organs more so than any other environmental factor 24 . Our result showed that biomass and growth of A. cremastogyne was significantly reduced under drought stress when compared with well-watered growing conditions, which may be explained by associated leaf area and stem diameter reductions (Table 1). A reduction in leaf area is generally ascribed to the suppression of leaf expansion, as cells lose the turgidity necessary for this process and the declining photosynthetic rate diminishes the availability of photoassimilates to build new cells 25 . However, some plant species may reduce their leaf area size to reduce water loss through transpiration, or to enhance proportionally their root water absorption capacity. Hu et al. 23 also suggested that A. cremastogyne adopts a drought avoidance strategy by reducing its individual leaf sizes as well as by shedding leaves (defoliation). Hence, the reduction in A. cremastogyne growth and biomass may be part of seedlings enacting this drought avoiding strategy, as well as being driven by declines in photosynthetic rate that we found (Table 2). Other possible reasons for reduced growth and biomass in drought-stressed plants is a water deficiency that interrupts water flow from soil to the xylem and surrounding cells, which decreases the mobility of available ions, nutrient availabilities and soil microbial activities 26,27 .
Our results also showed that P application did not enhance growth and biomass partitioning of the well-watered A. cremastogyne plants. An explanation for this result is that the already available P in the soil was sufficient to meet the functional requirements of plants during their establishment. Under drought stress conditions, applying P did not improve growth and biomass partitioning of A. cremastogyne, perhaps because its diffusion rate and movement towards the root organs slows under drought stress. Our findings contrast with several researches that reported higher biomass and growth rate in P-fertilized plants, but nonetheless support other studies finding non-significant effects of P fertilization on plant growth parameters 19,28,29 . These varied and inconsistent responses of different plant species to P fertilization in terms of their growth are perhaps not so LRWC, gas exchange, and chlorophyll fluorescence. Water relations in plants are important for maintaining their normal growth and metabolism when challenged by drought. The ability to maintain a high water potential is one of the most imperative challenges faced by many terrestrial plant species. Drought stress affects hydraulic conductivity of plants, driving changes in their shoots and the loss of turgor in their leaves 30 . LRWC is therefore considered an effective and reliable indicator for measuring the drought tolerance of plants. Regardless of the P application, LRWC was significantly lower in the plants drought-stressed than those well-watered, suggesting a low tolerance to drought stress in A. cremastogyne. Previous studies also suggested the same physiological response in different tree species 31,32 . Applying P significantly increased the LRWC under drought stress but not so in the well-watered plants (Table 2), which may be associated with the higher accumulation of osmolytes under P fertilization (Fig. 4) improving its osmotic adjustment. Our findings are in line with prior studies suggesting an improved LRWC in drought-stressed plants under P fertilization 33,34 . However, our findings contrast with several studies that suggested non-significant effects of P application on LRWC for different plant species under drought 19,20 . This disagreement may reflect the interspecific differences in physiological, biochemical and molecular mechanisms of the studied plants, such as their gene expression and protein assimilation.
Drought stress negatively affected P n due to reductions in G s and Ci, as well from damage to the photosynthetic apparatus as shown by the decrease in F v /F m . Reduced stomatal conductance may serve as another drought avoiding strategy of A. cremastogyne, enabling it to reduce its transpiration rate to better tolerate the physiological stress imposed by drought. The reason for a low stomatal conductance, or stomatal closure, may be a drought stress-induced synthesis of high amounts of abscisic acid (ABA) in the roots, which is then transported to the shoots. ABA stimulates the K+ ion efflux from the guard cells, resulting in loss of turgor pressure, which decreases G s 35 . This reduction in G s limits C i and leads to a decreased P n , as shown by our results (Table 2), a process that might be driven by declines in Rubisco activity 36 . However, P fertilization significantly increased P n in the drought-stressed plants without affecting G s , C i or F v /F m . These results are consistent with other studies reporting an improved photosynthetic rate in different drought-stressed plant species that were P-fertilized 19,37 . Thus, our findings suggest that P application can augment the ability of A. cremastogyne seedling to tolerate drought stress (as indicated by their increased LRWC and P n ).
Photosynthetic pigments. Photosynthetic pigments are essential for plants to harvest light; however, a decline in their concentrations can limit the photosynthetic rate of leaves and primary production 38 . In our study, the concentrations of photosynthetic pigments were significantly lower in the drought-stressed plants compared with the well-watered plants, which may be another reason for the decreased F v /F m and P n . This result clearly indicates that drought stress impairs the A. cremastogyne photosystem in leaves by degrading their photosynthetic pigments. Similar findings of drought having negative effects on chloroplast pigments have been reported in other studies 39,40 . Phosphorous application slightly improved the photosynthetic pigments' concentration, but no significant differences were observed. These findings provide an interesting insight: P fertilization significantly improves P n in drought-stressed plants but only slightly improves the pigments' concentrations, possibly for efficient regulation of the available amount of light. Our results agree and disagree with several previous studies, which may be due to the differences in duration and severity of drought investigated by them 18,37,38 . ROS production and antioxidative enzymes. Under stressful conditions, MDA accumulation leads to the damage of the cell membrane in plants 41 . MDA is measured as a suitable indicator for membrane lipid peroxidation. Our results showed that MDA contents were significantly higher in drought-stressed plants than in the well-watered ones due to a significant increase in ROS production. In plant cells, ROS (O 2 •− and H 2 O 2 ) act as signalling molecules; however, their excessive production can trigger fragmentation of DNA, proteins degradation, lipid peroxidation, and may even cause cell death 9 . The ROS level was significantly higher in drought-stressed than in well-watered plants when both were not fertilized. However, among the P-fertilized plants, ROS and MDA were similar between those well-watered and under drought-stressed; hence, fertilization with P avoids an ROS increment under drought stress. An increase in the ROS level under drought conditions was attributed to a reduced photosynthetic rate, as shown in our results ( Table 2) and also by previous studies 42 . The induction of antioxidant enzyme activities is a general defence mechanism that plants use against drought stress; it helps them to overcome oxidative stress and associated cell damage from it. Key antioxidative (SOD, POD and CAT) enzyme activities were consistently found to be higher in drought-stressed plants fertilized with P than in those non-fertilized. Therefore, under drought stress, applying P improves the antioxidant defence mechanism of A. cremastogyne seedling, which was efficient enough to remove ROS and decrease the MDA level with respect to the non-fertilized drought stress plants. Interestingly, P fertilization significantly reduced MDA accumulation and O 2 •production ( Fig. 3) due to significantly increased SOD and POD activities (Fig. 4). In sum, our findings indicate that A. cremastogyne's antioxidative defence mechanism has a weak efficiency when responding to drought-induced damage; however, via P fertilization, its efficiency is strengthened sufficiently to mitigate the negative effects of drought.

Soluble sugars, proline and soluble proteins. Plants typically accumulate different osmolytes in the
cell's cytosol for increasing osmotic potential in order to tolerate drought stress, since it improves or maintains turgidity and the continuation of plant growth processes. Moreover, these osmolytes also detoxify ROS, stabilize membrane and protect macromolecules 43 . Our results showed that plants under drought stress accumulated more proline and SS than did the ones well-watered. However, the SP concentration was lower in plants drought-stressed than in those well-watered. Possible reasons for this result include an associated increased function of protease enzymes, proteolysis or decreased protein synthesis, as well as a lower P n (i.e., less carbon to build any metabolite), in drought-stressed plants. Moreover, the P fertilization significantly increased the SS, proline and SP concentrations in the drought-stressed plants when compared with their non-fertilized counterparts. Shubra et al. 31 similarly reported an increased SS concentration in response to P application under drought conditions, which may have been caused by the reduction in of normal SS transport, utilization and distribution during water stress, or the hydrolysis of starch. Al-Karaki et al. 44 also observed higher accumulation of proline in P-fertilized sorghum plants, while Azcon et al. 45 also found a positive role of P application on the reduction and assimilation of nitrogenous compounds. Our findings clearly indicate that A. cremastogyne can maintain the osmotic potential of its cells under drought stress; however, P fertilization further enhances its osmotic potential and LRWC, which further improve its drought tolerance.
Drought stress had significant negative effects on the growth and metabolism of A. cremastogyne seedlings as indicated by their reduced biomass, leaf relative water content, photosynthetic rate, pigment concentrations, higher malondialdehyde level and inefficient antioxidative defence system. However, the reduced leaf area and higher osmolytes accumulation also reduced water losses and maintained the osmotic potential of the cells, suggesting it may serve as drought-avoidance or tolerant strategy. Phosphorous application had negligible effects on the morphological traits of this tree species when its seedlings were either drought-stressed or well-watered. In the former, however, P fertilization significantly improved the antioxidative enzymes activities, osmolytes accumulation, and also decreased the malondialdehyde level, so that their leaf relative water content and photosynthetic rate were improved. These results reveal that phosphorous fertilization facilitates A. cremastogyne under drought stress, mostly through physiological and biochemical adjustments at leaf level rather than changes made at the whole plant level in terms of growth or dry mass partitioning. Nonetheless, it is now imperative to investigate the underlying biochemical, physiological and molecular mechanisms, as well as the possible role of different levels of P fertilization under drought stress, in order to design effective measures for the proper management of this valuable tree species.

Materials and Methods
Plant collection and experimental design. The experiment was carried out at the Centre for Ecological Studies at the Chinese Academy of Sciences, Sichuan, in south-west China. Healthy and uniform, 2-year-old seedlings of A. cremastogyne were collected from Sichuan Agricultural University, Sichuan province, and transferred into 10-L plastic pots containing approximately 4 kg of homogenized topsoil (pH 7.3; total nitrogen, 0.19%; carbon, 2.67%). These pots were organized in a greenhouse (temperature range: 18-32 °C; relative humidity range 50-85%) following a complete randomized block design and watered regularly. Availability of light inside the greenhouse was kept homogeneous, and there was 6-9% reduction in direct sunlight. After growing for 2 months in the greenhouse, the seedlings were assigned to three replicates of four treatments for 90 days: two water regimes (well-watered vs. water-stressed) and two levels of P fertilization (with vs. without P fertilization). First, total and available P in soil was measured before applying the treatments, and found to be 0.89 g kg −1 and 27.6 mg kg −1 , respectively. Extraction of available P was done with 0.5 M of NaHCO 3 (pH 8.2) according to Olsen and Sommers 46 and colorimetric measurement was done by the molybdate-ascorbic acid as described by Murphy and Riley 47 . Weight method was used to calculate soil relative water content (SRWC) of the two water treatments (control: 80-85%; severe drought: 30-35%) 48 . The pots were weighed daily and watered up to their respective target SRWC, by replacing the amount of water transpired and evaporated. SRWC was expressed as follows: where W soil is the current soil weight (soil + pot + water), W pot is the weight of the empty pot, DW soil is the dry soil weight, and W FC is the soil weight at field capacity (soil + pot + water). Phosphorous fertilization was supplied as sodium di-hydrogen phosphate (NaH 2 PO 4 , 25.5% P), with the dose consisting of 129.3 mg P mixed in 200 mL of water per pot, applied every 30 days (i.e., three times for the entire experiment). To avoid systematic error produced by fluctuations in the local environmental conditions, the pots were rotated after every 5 days during the experiment. Plant samples were collected at the end of the experiment. From each plant, the upper fully-expanded leaves were used for all the physiological and biochemical determinations.

Analysis of growth and biomass.
Plant height (cm), stem diameter (mm) and leaf area (cm 2 ) were quantified in a standard way by using a measuring tape, electronic calipers, and leaf area meter (CI 202, USA), respectively. After removing the experimental plants from the soil, their roots, shoots and leaves of were separated and a subset of them oven-dried at 70 °C for 24 h to obtain their dry weights.
Determination of leaf relative water content. From each pot fully expanded leaves were collected as a single sample and their fresh weight (FW) measured. Then, the samples were immediately dipped into distilled water, in the dark at a 4 °C temperature. After 4 h, leaves were weighed to obtain their turgor weight (TW) and then put in an oven for 24 h at 70 °C to determine their dry weight (DW).
The following equation was used to calculate the LRWC of the samples: Gas exchange and chlorophyll fluorescence measurements. The net CO 2 assimilation rate (P n ), stomatal conductance (G s ), intercellular CO 2 concentration (C i ) and transpiration rate (E) were measured from fully expanded leaves at similar development stages with a portable open-flow gas exchange system (LI-6400, LI-COR Inc., USA) during the late morning (9:00-11:00 h). The relative humidity of air, CO 2 concentration and photon flux density were respectively maintained at 60-70%, 380 µmol mol −1 and 800 µmol m −2 s −1 in all cases. The maximum quantum efficiency of photosystem II (F v /F m ) was measured, from the same leaves as above, with a portable pulse amplitude modulated fluorometer (PAM −2 100, Walz, Effeltrich, Germany). The leaves were dark-adapted with clips for 20 mins. Then, a saturation pulse of 8000 µmol m −2 s −1 was applied for 0. extracted three times with 6 mL of 80% ethanol at 80 °C for 30 min. The resulting supernatant was analysed for soluble sugars (SS) following the anthrone method 50 . Proline was extracted with 2 mL of 10% acetic acid and 5 mL of 3% sulfosalicylic acid. The resulting supernatants were analysed according to the method described by Liu et al. 49 Soluble proteins (SP) concentrations were determined using Bradford G-250 reagent.

Determination of ROS and lipid peroxidation.
We homogenized the fresh leaves (0.2 g) using 2 mL of 65 mM phosphate buffer (pH 7.8) for the measurement of superoxide anion (O 2 •− ) production rate and centrifuged at 5000 g for 10 min 51 . The production rate was measured by observing the nitrite formation from hydroxylamine in the presence of O 2 •− . The composition of incubation mixture was, 0.9 mL of 65-mM phosphate buffer (pH 7.8), 0.1 mL of 10 mM hydroxylammonium chloride and 1 mL of supernatant. After 20 min incubation at 25 °C, sulphanilamide (17 mM) and α-naphthylamine (7 mM) were added to the incubation mixture, which was then kept at 25 °C for 20 min. Ethyl ether was added in the same volume and centrifugation was performed at 1500 g for 5 min. The absorbance wavelength used for the aqueous solution was 530 nm.
Hydrogen peroxide (H 2 O 2 ) determination was done by content was determined by observing the absorbance of the titanium-peroxide complex at 410 nm 52 . We homogenized fresh leaves (0.2 g) using 5 mL of acetone and centrifuged at 3000 g for 10 min. The reactive mixture contained 0.1 mL of a titanium reagent (50 μL of 20%-titanium tetrachloride in concentrated HCl), 0.2 mL of ammonia and 1 mL of supernatant; this was then centrifuged at 3000 g for 10 min. The resultant precipitate was washed for five times using acetone and centrifuged at 10 000 g for 5 min. The ensuing precipitate was then solubilized in 3 mL of 1-M H 2 SO 4 and its absorbance read at 410 nm.
Lipid peroxidation was estimated by measuring the malondialdehyde (MDA) content, according to the thiobarbituric acid (TBA) test, at 450, 532 and 600 nm 53 . For the MDA assay, 0.25 g of fresh leaves were ground in 5 ml of 1% trichloroacetic acid (TCA) and centrifuged at 5000 g for 10 min in a refrigerated centrifuge, then 1 ml of the supernatant was added to 4 ml of 20% TCA (containing 0.5%-thiobarbituric acid). The mixture was heated at 95 °C for 30 min, quickly cooled in an ice bath, and then its absorbance read with a spectrophotometer at 450, 532, and 600 nm 53 . The MDA concentration was calculated using the following equation: Statistical analysis. All the measurements were repeated three times, and data organized in Microsoft Excel 2007 and presented as mean values ± SE. We used SPSS v16.0 (SPSS Inc., 2007) to perform the one-way analysis of variance (ANOVA) on the data. Duncan's multiple range test, with an alpha level of 0.05 for significance, was used to make pairwise comparisons of the mean values for a given response variable. Before fitting the ANOVAs, the data were checked for normality and homogeneity of variances. The tool Origin pro v8.5 was used to draw the figure graphics, all of which show bars as the mean ± SE.