The effect of nitrogen and phosphorus deficiency on flavonol accumulation in plant tissues

Transcription

1 Blackwell Science, LtdOxford, UK PCEPlant, Cell and Environment Blackwell Science Ltd Flavonol accumulation in N- and P-deficient plant tissues A. J. Stewart et al. Original ArticleBEES SGML Plant, Cell and Environment (2001) 24, The effect of nitrogen and phosphorus deficiency on flavonol accumulation in plant tissues A. J. STEWART, 1 W. CHAPMAN, 2 G. I. JENKINS, 1 I. GRAHAM, 1 * T. MARTIN 1 & A. CROZIER 1 1 Division of Biochemistry and Molecular Biology, Institute of Biomedical and Life Sciences, University of Glasgow, G12 8QQ, UK and 2 Garrion Fruit Farm, Garrion Bridge, Wishaw, ML2 0RR, UK ABSTRACT The flavonol content of Arabidopsis thaliana and tomato seedlings was assessed in conditions of reduced nitrogen or phosphorus availability. In both systems, a significant inverse relationship was observed between nutrient availability and flavonol accumulation, with nitrogen limitation promoting the greatest increase in flavonols. A trial was established to determine the effects of decreased nitrogen and phosphorus availability on the flavonol content of leaf and fruit tissues of tomato plants (Lycopersicon esculentum cv. Chaser) in a commercial situation. Nutrients were supplied by a hydroponic system with nutrient regimes designed to provide the highest and lowest nitrogen and phosphorus levels with which it is possible to support plant growth and fruit set. Fruiting was abundant and tomato fruits were harvested at mature green, breaker and red stages of ripening; leaves were also harvested from the tops of the plants. All tissues were analysed for flavonol content using reversed-phase high-performance liquid chromatography. Flavonol accumulation in the leaves of mature tomato plants was found to increase significantly in response to nitrogen stress, whereas phosphorus deficiency did not elicit this response. Reduced nitrogen availability had no consistent effect on the flavonol content of tomato fruits. Phosphorus deficiency elicited an increase in flavonol content in early stages of ripening. Effects of nutrient stress on the flavonol content of tomato fruits were lost as ripening progressed. The findings suggest that nutrient status may be employed to manipulate the flavonol content of vegetative tissues but cannot be used to elevate the flavonol content of tomato fruit. Key-words: Arabidopsis thaliana; flavonol; HPLC; hydroponic; Lycopersicon esculentum; nitrogen, nutrient deficiency; phosphorus; tomato. Correspondence: Alan Crozier. Fax: ; a.crozier@bio.gla.ac.uk *Present address: CNAP, Central Science Laboratory, Sand Hutton, York, YO41 1LZ, UK. Present address, Department of Plant Sciences, University of Cambridge, Downing Street, Cambridge, CB2 3EA, UK. INTRODUCTION The flavonoids are a large family of polyphenolic compounds found in plant tissues, the main groups being the flavonols, flavonones, flavones, isoflavones, flavan-3-ols and anthocyanins (Haslam 1998) (Fig. 1). Flavonoids, with the exception of flavan-3-ols, occur mainly as sugar conjugates and are concentrated in the upper epidermis of leaves and skins of fruits. A wide range of functions have been proposed for flavonoids in relation to abiotic stresses. Flavonoid production is known to be induced by a wide range of environmental stimuli, including light (Li et al. 1993), pathogen attack (Dixon & Paiva 1995) and low temperature (Leyva et al. 1995). In addition, a variety of nutrient deficiencies in plants are characterized by an accumulation of flavonoids, notably the red/purple-coloured anthocyanins. Indeed, early studies attempted to assess leaf flavonoid content as an indicator of altered plant metabolism due to nutritional deficiency. In such studies, tomato (Lycopersicon esculentum) was commonly used as a model system. A comparison of tomato leaf extracts from control and N-starved plants showed an increase in anthocyanin content, particularly petunidin, as well as higher levels of the flavonol conjugate quercetin-3-o-glucoside in the deficient plants (Bongue- Bartelsman & Phillips 1995). It was also found that nitrogen deprivation greatly increased the levels of chalcone synthase (CHS) and dihydroflavonol reductase (DFR) mrna. An earlier study with apples demonstrated increased accumulation of phenylalanine ammonia lyase (PAL) following reduced availability of nitrogen and potassium (Tan 1980). One explanation for increased flavonoid synthesis under nitrogen stress is that enhanced PAL activity will release nitrogen for amino-acid metabolism whereas the carbon products are shunted via 4-coumaroyl-CoA into the flavonoid biosynthetic pathway (Margna 1977) (see Fig. 1). Alternatively, nitrogen limitation will affect photosynthesis by decreasing available chlorophyll and disrupting photosynthetic membranes due to starch accumulation. This may lead to increased sensitivity to high light levels. The production of photoprotective pigments such as anthocyanins and flavonols may afford protection against light-induced oxidative damage (Guidi et al. 1998). Flavonols are known to accumulate in the skins of tomato fruits (Stewart et al. 2000) and could therefore filter out damaging wavelengths of radiation Blackwell Science Ltd 1189

2 1190 A. J. Stewart et al. Figure 1. Pathways for the biosynthesis of key flavonoids. Key enzymes: PAL, phenylalanine ammonia-lyase; C4H, cinnamate-4-hydroxylase; 4 CL, 4- coumarate: CoA ligase; ACoAC, acetyl CoA carboxylase; CHS, chalcone synthase; CHI, chalcone isomerase; IFS, 2- hydroxyisoflavanone synthase; FNS, flavone synthase; FN3H, flavanone 3-hydroxylase; DFR, dihydroflavonol reductase; LAR, leucoanthocyanidin-4-reductase; ANS, anthocyanin synthase; FLS, flavonol synthase; FL3 H, flavonol-3 -hydroxylase; FL3 MT, flavonol-3 -methylase. 3, condensation of three acetate units from malonyl CoA with 4-coumaroyl-CoA. Much interest in recent years has focused on flavonol intake from the human diet and possible health benefits due to the antioxidant nature of the aromatic flavonol structures (Robak & Gryglewski 1988, Salah et al. 1995; Rice-Evans, Miller & Paganga 1997). It is therefore of interest to determine the mechanisms of flavonol regulation in plant tissues. In addition, environmental influences may be manipulated in order to produce crops rich in flavonols without the need for genetic modification. Although previous studies identified a link between nutrient deficiency and flavonoid accumulation in plant tissues, information on the exact nature of these flavonoids is lacking. Most studies have described only flavonoid groups such as flavonols, flavones and flavonones rather than identifying the individual components. The aim of this study was to identify and quantify the flavonols present in Arabidopsis and tomato tissues and determine whether nutritional deficiency could be employed to increase flavonol accumulation in plant tissues. The flavonol response of Arabidopsis thaliana seedling tissue was determined in conditions of reduced nitrogen and phosphorus availability. Individual flavonols were identified and quantified using a sensitive and selective high-performance liquid chromatography (HPLC) procedure involving postcolumn derivatization. Based on these results a further study was carried out to determine the flavonol response of tomato seedlings to nitrogen and phosphorus stress. In addition, the effect of nitrogen and phosphorus deprivation on the flavonol content of leaves and fruits of mature tomato plants grown in a commercial setting was investigated. The present study provides clear evidence that the flavonol content of plant tissues is influenced by their nutritional status. Manipulation of nutrient availability did not stimulate increased levels of flavonols within tomato fruits. However, nutrient stress could be used to manipulate the flavonol content of vegetative crops to produce flavonolrich food products with potential benefits to human health.

3 Flavonol accumulation in N- and P-deficient plant tissues 1191 MATERIALS AND METHODS Preparation of Murashige and Skoog media Macronutrients Murashige and Skoog (MS) media were prepared containing the following macronutrients; 1 25 mm potassium phosphate (KH 2 PO 4 ), 2 26 mm calcium chloride (CaCl 2 ), 0 73 mm magnesium sulphate (MgSO 4 ). In addition all media contained 20 mm sodium-iron EDTA, 100 mm sucrose and 8 g dm 3 agar. Media were adjusted to ph 5 7 with 0 1 M potassium hydroxide and immediately autoclaved. Micronutrients In order to provide the required micronutrients to support plant growth, the following compounds were added to MS media; 0 1 mm boric acid (H 3 BO 3 ), 0 1 µm cobalt chloride (CoCl 2 ), 0 1 µm copper sulphate (CuSO 4 ), 0 1 mm manganese sulphate (MnSO 4 ), 1 0 µm molybdic acid (Na 2 MoO 4 ), 5 0 µm potassium iodide (KI), 0 03 mm zinc sulphate (ZnSO 4 ). A 200 concentrated solution was prepared and 5 ml was added to 1 litre of medium. Altering nitrogen content of medium Standard nitrogen concentration in medium was 60 mm. Media containing nitrogen at the following concentrations were prepared 0, 0 1, 0 6, 6 0 and 60 0 mm. Nitrogen was added in the ratio of 1 mol ammonia : 2 mol nitrate supplied in the form of ammonium nitrate and potassium nitrate. All media contained 1 25 mm potassium phosphate and 100 mm sucrose. Altering phosphate concentration of medium Media were prepared containing, 0, 0 3, 0 6, 2 5 and 6 3 mm potassium phosphate (KPO 4 ). 1 mm 2-[N-Morpholino] etnanesulfonic acid (MES) buffer was added to the media to enable adjustment of the ph to 5 7. All media contained 60 mm nitrogen and 100 mm sucrose. Potassium in the form of KCl was added to media containing low concentrations of nitrogen and phosphate to compensate for the reduced concentration of KNO 3 or KPO 4 present. Growth of plants on sterile MS medium Seed sterilization Prior to sterilization tomato seeds were imbibed in a constant flow of fresh water for 2 h. Seeds, contained within filter paper packets, were placed within a magenta vessel. They were then immersed for 2 min in ethanol:water (7 : 3, v/v) followed by 10 min in 10% (v/v) sodium hypochlorite with occasional agitation. Seeds were finally rinsed five times in sterile water and left overnight in a laminar flow hood to dry. Growth conditions Arabidopsis seed were sown onto MS media in Petri dishes under sterile conditions. Plates were covered and placed at 4 C in darkness for 3 d. Plates were then transferred for 24 h to white light (100 µmol photon m 2 s 1 ) provided by warm white fluorescent tubes, at 20 C for a further 10 d. Lycopersicon esculentum seed were sown onto MS media in magenta vessels allowing the height required for growth of tomato seedlings. Magenta vessels were sealed and transferred to 24 h white light, as described above, at 20 C. Plants were grown for either 10 or 21 d before being harvested. Harvesting of tissue from sterile medium Tissue was harvested on ice and frozen directly in liquid nitrogen. Tissue for flavonol analysis was freeze-dried and stored at 20 C prior to analysis. Growth of Lycopersicon esculentum in hydroponic conditions Study design Tomato plants were grown in a commercial environment under glass at Garrion Fruit Farm (Garrion Bridge, Wishaw, Lanarkshire, UK). Hydroponic nutrient regimes were designed to control nitrogen and phosphorus availability while still allowing plant growth and fruit set. This allowed determination of the effect of nitrogen and phosphorus fertilizers on the flavonol content of tomato fruit and mature vegetative tissue. Tissue sampling was carried out on two occasions 1 month apart (May June 1998). Red, green and breaker fruits and leaf tissue were collected on each occasion. Tissue was snap-frozen in liquid nitrogen, freeze-dried and stored at 20 C prior to determination of flavonol content. Design of fertilizer stock solutions Stock solution recipes were based on the use of a 40 gallon (182 dm 3 ) stock tank, this is the standard stock tank size used by commercial growers. All solutions were diluted 1 : 100 for application to plants in the trial. Fertilizers for stock solutions were purchased from Clydeside Trading Society, Wishaw, Lanarkshire, UK. Phosphorus manipulation Control phosphate regime. This was designed to produce a nutrient feed with a phosphate concentration of 0 96 M mono-ammonium phosphate (30 µg m 1 phosphate) following dilution. This regime required two stock tanks A and B. Tank A contained calcium nitrate (10 67 kg per 182 dm 3 ), tank B contained potassium nitrate (14 3 kg per 182 dm 3 ), mono-ammonium phosphate (1 97 kg per 182 dm 3 ), ammonium nitrate (1 14 kg per 182 dm 3 ), magnesium sulphate (9 08 kg per 182 dm 3 ) and Solufeed TEC (0 55 kg per 182 dm 3 ). Trace elements boron, copper, iron, manganese,

4 1192 A. J. Stewart et al. molybdenum and zinc were supplied by the Solufeed TEC (Clydeside Trading Society). Low-phosphate regime. This was designed to produce a phosphorus concentration of 0 16 M mono-ammonium phosphate (5 µg g 1 phosphorus). This regime matched the control situation as closely as possible differing only in the concentration of mono-ammonium phosphate (0 33 kg per 182 dm 3 ) and ammonium nitrate (2 27 kg per 182 dm 3 ). High-phosphate regime. This contained 3 13 M monoammonium phosphate (100 µg g 1 phosphorus). This was achieved by increasing the concentration of mono-ammonium phosphate to 6 56 kg per 182 dm 3 and omitting ammonium nitrate. Nitrogen manipulation Control nitrogen regime. This supplied nitrate in the form of calcium nitrate (2 5 M) and potassium nitrate (7 81 M), ammonia was supplied as mono-ammonium phosphate (1 39 M) (total nitrogen 193 µg ml 1 ). Stock tank A contained calcium nitrate (10 67 kg per 182 dm 3 ), stock tank B contained potassium nitrate (14 3 kg per 182 dm 3 ), monoammonium phosphate (2 95 kg per 182 dm 3 ), magnesium sulphate (9 08 kg per 182 dm 3 ) and Solufeed TEC (0 55 kg per 182 dm 3 ). Low-nitrogen regime. This contained calcium nitrate (1 23 M), potassium nitrate (2 37 M) and mono-ammonium phosphate (0 96 M) (total nitrogen 79 µg g 1 ). To achieve this the concentration of calcium nitrate was decreased to 5 33 kg per 182 dm 3, potassium nitrate was decreased to 4 30 kg per 182 dm 3 and mono-ammonium phosphate was decreased to 1 97 kg per 182 dm 3. In addition, magnesium sulphate levels were decreased to 5 45 kg per 182 dm 3. High-nitrogen regime. This contained calcium nitrate (4 96 M), potassium nitrate (15 13 M), ammonium nitrate (1 50 M) and mono-ammonium phosphate at control levels (1 39 M) (total nitrogen 405 µg g 1 ). This involved increasing levels of calcium nitrate to kg per 182 dm 3, potassium nitrate to kg per 182 dm 3 and adding ammonium nitrate at 2 27 kg per 182 dm 3. Mono-ammonium phosphate and magnesium sulphate were retained at control levels. Hydroponic plant growth conditions. Garrion Fruit Farm, supplied tomato plants of variety Chaser at age 3 4 months. These plants were installed (February March 1998) in a commercial glass house receiving nutrients through a controlled drip system. Light levels and temperature would have varied according to outdoor weather conditions. Sampling occurred from May June 1998; at this time plants would receive h of light per day. Light levels measured on-site at midday during sampling were µmol photon m 2 s 1 at the top of the plants and 150 µmol photon m 2 s 1 at truss level. Daytime temperatures varied from 20 to 25 C. Analysis of the flavonol content of plant tissues Sample preparation Tomato fruit skins were peeled from flesh using a sharp kitchen knife prior to processing. Plant tissues were snap frozen in liquid nitrogen, lyophilized and ground to a fine powder prior to acid hydrolysis. Extraction and hydrolysis conditions Optimization of acidic conditions for the hydrolysis of flavonol conjugates has been described by Hertog, Hollman & Venema (1992) following an earlier study by Harbourne (1965) on the release of free flavonols by acidic and enzymatic hydrolyses. Preliminary screening was carried out to ascertain the most effective acid hydrolysis conditions for the tissues involved in this study. Samples of seedling, leaf and tomato fruit tissues (20 mg lyophilized tissue), were all hydrolysed at 90 C for 2 h in a 3 ml glass V-vial containing 2 ml 1 2 M HCl in 50% aqueous methanol and 20 mm sodium diethyldithiocarbamate as an antioxidant. A Tefloncoated magnetic stirrer was placed in the vial, which was sealed tightly with a PTFE-faced septum prior to heating in a Reacti-Therm Heating/Stirring Module (Pierce, Rockford, IL, USA). Extract aliquots of 100 µl, taken both before and after acid hydrolysis, were made up to 250 µl with distilled water adjusted to ph 2 5 with trifluoroacetic acid and filtered through a 0 2 µm Anopore filter (Whatman, Maidstone, Kent, UK), prior to the analysis of 100 µl volumes (1/50th aliquot of total sample) by gradient elution reversed phase HPLC (Crozier et al. 1997). Samples were hydrolysed and analysed in triplicate. Flavonol analysis by HPLC HPLC and post-column derivatization Samples were analysed using a Shimadzu (Kyoto, Japan) LC-10 A series automated liquid chromatograph comprising a SCL-10 A system controller, two LC-10 A pumps, a SIL-10 A autoinjector with sample cooler, a CTO-10 A column oven, and a SPD-10 A UV-vis detector linked to a Reeve Analytical (Glasgow, UK) 2700 data-handling system. Reversed phase separations were carried out at 40 C using a mm i.d., 4 µm Genesis C 18 cartridge column fitted with a mm i.d., 4 µm C 18 Genesis guard column in an integrated holder (Jones Chromatography, Hengoed, UK). The mobile phase was a 20 min, 20 40% gradient of acetonitrile in distilled water adjusted to ph 2 5 with trifluoroacetic acid, eluted at a flow rate of 0 5 ml min 1. Column eluent was first directed to the SPD-10 A absorbance monitor operating at 365 nm, after which post-column derivatization was achieved by the addition of 0 1 M methanolic aluminium nitrate containing 7 5% (v/v) glacial acetic acid (Hollman & Trijp 1996; Aziz et al. 1998) pumped at a flow rate of 0 5 ml min 1 by a pulse free Model 9802 precision mixer/splitter (Reeve Analytical). The mixture was passed through 1 9 m 30/ i.d. loop of PEEK tubing in

5 Flavonol accumulation in N- and P-deficient plant tissues 1193 the column oven before being directed to a RF-10 A fluorimeter and fluorescent flavonol complexes detected at excitation 425 nm and emission 480 nm. The limit of detection at A 365 was < 5 ng and linear ng calibration curves were obtained for morin, rutin, quercetin, kaempferol and isorhamnetin. The fluorescent intensity of the individual flavonol derivatives varied, however, ng linear calibration curves were obtained for myricetin, morin, quercetin, kaempferol and isorhamnetin. Estimates of free and conjugated flavonol levels Free flavonols were detected in the unhydrolysed sample while the hydrolysed samples contained both free and conjugated flavonols. Thus, conjugated flavonol levels were estimated by subtracting the amount found in the unhydrolysed samples from that detected after acid hydrolysis (Crozier et al. 1997). Reference compounds Morin, myricetin, quercetin, rutin and kaempferol were purchased from Sigma Chemicals (Poole, UK). Isorhamnetin was obtained from Apin Chemicals (Abingdon, Oxford, UK). Statistics Data are presented as mean values ± standard error (SE) (n = 3). Where appropriate data were subject to statistical analysis using analysis of variance (ANOVA) to determine the significance of the observed treatment/response relationships. Statistical analyses were performed using MINITAB software, version 12 (Mininc., Addison-Wesley Publishing Co., Reading, MA, USA). RESULTS Seedlings of Arabidopsis thaliana Sterile seeds were plated onto MS media containing either 0 to 60 mm nitrogen or 0 to 6 3 mm phosphate, and allowed to grow for 11 d in conditions of white light (100 µmol photon m 2 s 1 ) at 20 C. Using reversed phase HPLC with absorbance and fluorescence detection following postcolumn derivatization, three flavonols were detected in Arabidopsis seedling extracts. Although no attempt was made in this study to analyse the effects of nutrient deficiency on individual flavonol conjugates, analysis of samples taken before and after acid hydrolysis gave an indication of the proportion of flavonols present in free and conjugated forms. Quercetin and kaempferol were detected almost exclusively as conjugates ( 99%), whereas isorhamnetin was found only in a conjugated form. The standard nitrogen concentration found in MS medium is 60 mm. Arabidopsis seedlings grown in these conditions were found to contain a total flavonol content of ± 18 0 µg g 1 fresh weight (FW) (Table 1). Seedlings grown in conditions of zero nitrogen had a total flavonol content of ± 0 8 µg g 1 FW. Clearly, limiting nitrogen availability induced higher concentrations of quercetin, kaempferol and isorhamnetin in Arabidopsis seedlings. The inverse relationship between nitrogen availability and flavonol content was noted to be highly significant (P = 0 01). Those Arabidopsis seedlings grown on a zero phosphate medium showed a clear increase in total flavonol content compared to plants grown on 6 3 mm phosphate (105 1 ± 11 2 and 27 6 ± 3 3µg g 1, respectively). Again this relationship was found to be statistically significant (P = 0 01). Limiting phosphate availability induced higher concentrations of quercetin, kaempferol and isorhamnetin in Arabidopsis seedlings (Table 1). Having established that the flavonol content of plant tissues can be influenced by their nutritional status it was of interest to determine whether exposure to low levels of nitrogen or phosphorus could induce increased accumulation of flavonols in a commercial crop plant. Tomato was selected as a model system in which to study the effects of nutrient deprivation on flavonol levels in seedlings, mature vegetative tissue and fruit tissue. Tomato seedlings Sterile tomato seeds were sown onto MS media containing zero, 6 0 or 60 mm nitrogen or zero, 2 5 or 6 3 mm phosphate Table 1. Quercetin, kaempferol and isorhamnetin content of Arabidopsis thaliana seedlings in conditions of reduced nitrogen or phosphorus availability. Results represent FW (µg g 1 ) ±SE, where n = 3 Nutrient status Concentration (mm) Quercetin Kaempferol Isorhamnetin Total flavonols Nitrogen ± ± ± ± ± ± ± ± ± ± ± ± ± ± 10 3 ND 44 6 ± ± ± ± ± 18 0 Phosphorus ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 3 3 ND, none detected.

6 1194 A. J. Stewart et al. Table 2. Quercetin, kaempferol and isorhamnetin content of Lycopersicon esculentum seedlings in conditions of reduced nitrogen and phosphorus availability. Results represent FW (µg g 1 ) ±SE, where n = 3 Nutrient status Concentration (mm) Quercetin Kaempferol Isorhamnetin Total flavonols Nitrogen ± ± ± ± ± ± ± ± 0 2 Phosphorus ± ± ± ± ± ± ± ± ± ± ± ± 0 6 and allowed to grow for 21 d in white light (100 µmol photon m 2 s 1 ). HPLC analysis of tomato seedling tissues identified the flavonols quercetin, kaempferol and isorhamnetin. Flavonols were present almost exclusively in the form of conjugates with only very low levels of free quercetin detected. Seedlings grown in conditions of zero nitrogen were found to contain 15 9 ± 0 7 µg flavonol g 1 (Table 2). Increasing the nitrogen concentration of MS media to 6 0 mm reduced the total flavonol content of tomato seedlings to 3 6 ± 0 2 µg g 1. Flavonol levels showed an approximately two-fold increase in conditions of zero phosphate in comparison with 2 5 and 6 3 mm phosphate (6 2, 3 7 and 3 0 µg flavonol g 1, respectively). A significant inverse relationship was observed between both nitrogen and phosphate nutrition and the flavonol content of tomato seedling tissues (P = 0 05). The results indicate that the response of tomato seedlings in conditions of nutrient stress is very similar to that of Arabidopsis. Both species contain the flavonols quercetin, kaempferol and isorhamnetin, the concentration of which is increased in response to either nitrogen or phosphate stress. Nitrogen and phosphorus deprivation of tomato plants in a commercial setting A trial was established at Garrion Fruit Farm to determine the effects of decreased nitrogen and phosphate availability on leaf and fruit tissues of tomato plants grown in a commercial situation. Experimental plants were grown alongside commercial tomato plants in a glasshouse. Nutrients were supplied by a hydroponic system, with nutrient regimes designed to represent high, control and low nitrogen and phosphate concentrations as described. Nutrient concentrations were selected to provide the highest and lowest nitrogen and phosphate levels with which it is possible to support plant growth and fruit set. Such conditions would not be expected to support long-term cultivation. Full nutrient, conductivity and ph analyses were carried out on all hydroponic solutions to ensure that the desired nitrogen and phosphate levels were achieved. Nitrogen levels were found to adhere closely to values expected from feed recipes, phosphorus levels were generally lower than anticipated but were deemed to be satisfactory for this study. Nitrogen and phosphate concentrations employed for commercial use in the Clyde Valley approximate most closely with the high treatment regimes. Three tomato plants cv. Chaser were grown for each treatment level. Tissue sampling was carried out on two occasions 1 month apart (May and June 1998). On each occasion mature green, breaker and red tomato fruits were removed randomly from the plants in each nutrient regime. The breaker stage of tomato fruit development is characterized by mature sized fruit in which chlorophyll degradation results in whitening of the fruit skin and carotenoids (notably the red carotenoid lycopene) begin to be synthesized. Leaf samples were also removed from the top of the plants. Tomato leaf tissue As with tomato seedlings, mature leaf tissue contains conjugated quercetin, kaempferol and isorhamnetin. The predominant flavonol was quercetin (80 90%) followed by kaempferol (8 9%) with relatively low levels of isorhamnetin present. The total flavonol content of tomato leaves was increased in response to reduced nitrogen availability. In trial 1 (May 1998) the total flavonol content varied from 50 5 µg g 1 in the high-nitrogen regime to µg g 1 in the low-nitrogen regime; in trial 2 (June 1998) the flavonol content increased from to µg g 1 (Table 3). The negative correlation between nitrogen availability and flavonol content in trial 1 was found to be significant (P = 0 05). Although a similar trend in flavonol accumulation was observed in trial 2, this was not found to be statistically significant. The flavonol content of tomato leaves collected from plants grown in conditions of reduced phosphate did not show a clear trend in flavonol accumulation. Samples collected in May 1998 showed a decreasing flavonol content as phosphate availability was reduced ( µg g 1 ). Those samples collected in June 1998 showed an increase in flavonol content from high to low-phosphate regimes ( µg g 1 ). In the conditions of our study high and control phosphate levels (64 8 and 6 13 µg g 1, respectively) appear to be the most inductive phosphate concentrations for vegetative tomato leaf tissue (Table 3). Tomato fruits Tomato fruits at three stages of ripening, mature green, breaker and red were harvested randomly from plants in each nutrient regime in May and June 1998 (trial 1 and 2,

7 Flavonol accumulation in N- and P-deficient plant tissues 1195 Table 3. Quercetin, kaempferol and isorhamnetin content of tomato leaf tissue cv. Chaser grown in a hydroponic system supplied with high, low or control levels of nitrogen or phosphorus. Leaves for trial 1 harvested May 1998; leaves for trial 2 harvested June Results represent FW (µg g 1 ) ±SE, where n = 3 Nutrient regime Trial Quercetin Kaempferol Isorhamnetin Total flavonols High N 44 7 ± ± ± ± ± ± ± ± 22 0 Control N 39 9 ± ± ± ± ± ± ± ± 19 0 Low N ± ± ± ± ± ± ± ± 9 3 High P 94 1 ± ± ± ± ± ± ± ± 18 0 Control P 72 0 ± ± ± ± ± ± ± ± 10 0 Low P 49 4 ± ± ± ± ± ± ± ± 15 0 respectively). Despite the altered nutritional status of the plants in the trial, fruiting was abundant and the tomato fruits appeared normal. Fruit skins were removed for flavonol analysis. Tomato fruit skins were found to contain primarily conjugated quercetin with lower levels of conjugated kaempferol. Very low levels of free quercetin and kaempferol were also detected. Skins from mature green fruits showed a small increase in total flavonol content in the low-nitrogen regime as compared to the control and high regimes (Table 4). Altering nitrogen availability had no obvious effect on the flavonol content of tomato fruits at the breaker and red stage of ripening. Reduced phosphate availability caused an increase in flavonol accumulation in fruit skins early in fruit development (mature green stage). The flavonol content in the high regime compared to the low regime was found to increase from 19 6 to 24 3 µg g 1 in trial 1 and µg g 1 in trial 2 (Table 5). The negative correlation between phosphate nutrition and flavonol content in mature green fruit skin was found to be statistically significant in trial 2 (P = 0 01). Although a similar trend in flavonol accumulation was observed in trial 1 this was not found to be statistically significant. Reduced phosphate availability caused increased flavonol content at the breaker stage of fruit development but to a lesser extent than observed at the green stage, µg g 1 in trial 1 and µg g 1 in trial 2 (Table 5). There was no evidence for altered flavonol content due to phosphate nutrition at the red stage of ripening. DISCUSSION Arabidopsis and tomato seedlings grown on nitrogen- or phosphate-deficient media displayed classical symptoms of mineral deficiency. Those seedlings grown on phosphatedeficient media appeared normal in size but were darker green in colour than those grown in standard conditions of 2 5 mm phosphorus. At the highest phosphate concentrations, seedlings were pale yellow in colour. Seedlings grown on nitrogen-deficient media were greatly reduced in size and purple in colour, indicating an increased anthocyanin content. As the nitrogen concentration of the medium was increased, seedlings became light green and then of normal colour and size from 6 0 mm nitrogen onwards. Limiting nitrogen and phosphorus availability also induced higher concentrations of flavonols in seedling tissue of both Arabidopsis thaliana and Lycopersicon esculentum. The data show that flavonol content is inversely related to nitrogen and phosphate nutrition in Arabidopsis seedlings (Tables 1 and 2). Table 4. Quercetin and kaempferol content of tomato fruit skins c.v Chaser grown in high, low or control nitrogen conditions in the Clyde valley, May June Tomatoes for trial 1 collected May 1998; tomatoes for trial 2 collected June Results represent FW (µg g 1 ) ±SE, where n = 3 Developmental stage Trial Nutrient regime Quercetin Kaempferol Total flavonols Green High 16 6 ± ± ± ± ± ± 1 9 Control 17 6 ± ± ± ± ± ± 0 7 Low 22 4 ± ± ± ± ± ± 4 9 Breaker High 23 8 ± ± ± ± ± ± 1 0 Control 19 4 ± ± ± ± ± ± 1 0 Low 20 0 ± ± ± ± ± ± 1 5 Red High 28 2 ± ± ± ± ± ± 3 0 Control 30 9 ± ± ± ± ± ± 1 1 Low 24 6 ± ± ± ± ± ± 1 1

8 1196 A. J. Stewart et al. Table 5. Free and conjugated quercetin and kaempferol content of tomato fruit skins c.v Chaser grown in conditions of high, low and control phosphate in the Clyde valley May June Tomatoes for trial 1 collected May 1998; tomatoes for trial 2 collected June Results represent FW (µg g 1 ) ±SE, where n = 3 Developmental stage Trial Nutrient regime Quercetin Kaempferol Total flavonols Green High 14 8 ± ± ± ± ± ± 2 1 Control 13 7 ± ± ± ± ± ± 0 6 Low 19 5 ± ± ± ± ± ± 2 5 Breaker High 15 0 ± ± ± ± ± ± 1 2 Control 16 5 ± ± ± ± ± ± 1 5 Low 17 5 ± ± ± ± ± ± 3 9 Red High 26 6 ± ± ± ± ± ± 2 5 Control 23 5 ± ± ± ± ± ± 1 0 Low 21 0 ± ± ± ± ± ± 3 4 Overall, flavonol concentrations were lower in tissues from tomato seedlings compared with Arabidopsis. However, the flavonol response to conditions of nutrient stress was very similar. Both species contain the flavonols quercetin, kaempferol and isorhamnetin the concentration of which is increased in response to either nitrogen or phosphate stress. Nitrogen limitation promoted greater increases in flavonols than a reduced supply of phosphate. However, in conditions of high nitrogen, a reduction of phosphate facilitates flavonol induction. Commercial fruit growing in Scotland and the rest of the UK requires the use of glasshouses in order to maintain the temperatures required for fruit set and also to control pests and disease. Due to the costs involved in building and maintaining these glasshouses, tomato growing is highly intensive, with plants tightly packed together and temperature and feeding regimes optimized to obtain the highest possible yields. Consequently commercial tomato plants receive high concentrations of nutrients via hydroponic feeding systems. If nutrient availability can influence flavonol accumulation in plant tissues, the addition of high concentrations of fertilizers may have a negative effect on the flavonol content of UK produce. The aim of this investigation was to decrease the concentration of nitrogen and phosphorus in the commercial plant feed and determine any effect on the flavonol content of mature vegetative and fruit tissue. It was not the intention of this study to impose a level of nutrient stress that would decrease growth or fruit set. Mature tomato leaf tissue was found to contain conjugated quercetin, kaempferol and isorhamnetin, the concentration of which was found to increase in response to reduced nitrogen availability in both trials. Samples collected in trial 2 (June 1998) generally had a higher flavonol content than samples collected in trial 1 (May 1998). This increase may be due to increasing nutrient stress over the additional month that plants had been exposed to the various nutrient regimes. In addition, light levels may have been increasing throughout May June; those samples collected in June may have shown greater light induction of flavonols. Leaf tissue collected from plants grown in conditions of high or low phosphate showed no clear trend in flavonol accumulation. Indeed, in the conditions of our study high and control phosphate regimes appear to be the most inductive phosphate concentrations for vegetative tomato leaf tissue (Table 3). A study by Bongue-Bartelsman & Phillips (1995) determined the effect of nitrogen deficiency on the anthocyanin and flavonol content of tomato leaves using an HPLCbased approach. They reported that anthocyanin content increased 3 4-fold in response to nitrogen stress. The only flavonol reported was quercetin-3-o-β-glucoside, levels of which doubled in response to nitrogen stress. An earlier study by Carpena, Zornoza & Mataix (1982) determined the effect of decreased phosphate nutrition on mature leaves of tomato plants. There was no evidence in their study for an increase in flavonols in tomato leaves due to phosphate deficiency. Tomato fruits at three stages of ripening, mature green, breaker and red were harvested from plants in each nutrient regime. Tomato fruit skins were analysed for flavonol content. Skins from mature green fruits of the low-nitrogen regime showed a small increase in flavonol content in comparison with the control and high regimes (Table 4). Nitrogen availability had no consistent effect on the flavonol content of tomato fruits at the breaker and red stages of ripening. The flavonol content of green tomato fruit skins was found to be higher in fruits from the second trial (June) than those sampled in May. This may indicate increasing nitrogen stress on the plants over the duration of the experiment and increasing light levels. This effect was no longer observed as fruits reached the breaker and red stage of ripening. Reduced phosphate availability caused an increase in flavonol accumulation at the mature green stage and to a lesser extent at the breaker stage of ripening. Any effect of nutrient stress on the flavonol content of the fruits appeared to be lost as ripening progressed. It is possible that green fruits may have to compete with other plant sinks for available nutrients and may therefore suffer nutritional deficiency. During ripening the sink strength of the

9 Flavonol accumulation in N- and P-deficient plant tissues 1197 fruit is likely to increase such that the nutrient deficiency no longer has any effect on flavonol induction. Alternatively, induction of flavonols in the skins of tomato fruits may be important to protect the fruit tissues and developing seeds from penetration by potentially damaging UV-B radiation. CONCLUSIONS The present study provides clear evidence that the flavonol content of seedling and vegetative tissues can be influenced by their nutritional status. Components of the flavonoid biosynthetic pathway are attributed a wide range of possible functions in plants. Upregulation of this pathway in conditions of nutrient deprivation may afford protection against further sources of stress such as pathogen attack or light-induced damage. Nitrogen and phosphorus deprivation caused an increase in flavonols in seedlings of Arabidopsis thaliana and Lycopersicon esculentum. Nitrogen deprivation was also able to increase flavonol accumulation in mature vegetative tissue of tomato plants. Phosphorus deficiency did not elicit this response. Nutrient deficiency appeared to produce an increase in flavonols in tomato fruit tissues only in the early stages of fruit development; as ripening progressed, no increase was observed. Nutrient status may be employed to manipulate the flavonol content of vegetative plant tissues. It would be of interest to test this theory on a leafy crop plant. In addition, future work in this area should investigate phosphate and nitrogen interactions. Further studies may be required to determine the longevity of the flavonol increase in response to a period of nutrient stress and investigate factors such as reduction in yield. Increased dietary intake of flavonols is believed to be linked to potential health benefits. An understanding of the factors influencing flavonol accumulation in plant tissues may allow the production of fruits and vegetables rich in flavonols. ACKNOWLEDGMENTS A.J.S. was supported by a BBSRC CASE postgraduate studentship with Scotland s Tomatoes plc. The authors would like to thank Chris Wand, SAC Advisory Service, Lanark, UK, for design of hydroponic nutrient regimes. REFERENCES Aziz A.A., Edwards C.A., Lean M.E.J. & Crozier A. (1998) Absorption and excretion of conjugated flavonols, including quercetin-4 -O-β-glucoside and isorhamnetin-4-o-β-glucoside by human volunteers after the consumption of onions. Free Radical Research 29, Bongue-Bartelsman M. & Phillips D.A. (1995) Nitrogen stress regulates gene expression of enzymes in the flavonoid biosynthetic pathway of tomato. Plant Physiology & Biochemistry 33, Carpena O., Zornoza P. & Mataix J. (1982) Incidence of P, Mn and B deficiencies on the levels of the whole and individual flavonoid groups in tomato leaves. Journal of Plant Nutrition 5, Crozier A., Lean M.E.J., McDonald M.S. & Black C. (1997) Quantitative analysis of the flavonoid content of commercial tomatoes, onions, lettuce and celery. Journal of Agricultural and Food Chemistry 45, Dixon R.A. & Paiva N.L. (1995) Stress-induced phenylpropanoid metabolism. Plant Cell 7, Guidi L., Lorefice G., Pardossi A., Malorgio F., Tognoni F. & Soldatini G.F. (1998) Growth and photosynthesis of Lycopersicon esculentum (L.) plants as affected by nitrogen deficiency. Biologia Plantarum 40, Harbourne J.B. (1965) Plant polyphenols-xiv. Characterization of flavonoid glycosides by acidic and enzyme hydrolyses. Phytochemistry 4, Haslam E. (1998) Polyphenols, collagen and leather. In Practical Polyphenols from Structure to Molecular Recognition and Physiological Action (ed. E. Haslam), pp Cambridge University Press, Cambridge. Hertog M.G.L., Hollman P.C.H. & Venema D.P. (1992) Optimisation of a quantitative HPLC determination of potentially anticarcinogenic flavonoids in vegetables and fruits. Journal of Agricultural and Food Chemistry 40, Hollman P.C.H. & van Trijp J.M.P.B.N.C.P. (1996) Fluorescence detection of flavonols in HPLC by postcolumn chelation with aluminium. Analytical Chemistry 68, Leyva A., Jarillo J.A., Salinas J. & Martinez-Zapater J.M. (1995) Low temperature induces the accumulation of phenylalanine ammonia-lyase and chalcone synthase mrnas of Arabidopsis thaliana in a light-dependent manner. Plant Physiology 108, Li J., Ou-Lee T.M., Raba R., Amundson R.G. & Last R.L. (1993) Arabidopsis flavonoid mutants are hypersensitive to UV-B irradiation. Plant Cell 5, Margna U. (1977) Control at the level of substrate supply an alternative in the regulation of phenylpropanoid accumulation in plant cells. Phytochemistry 16, Rice-Evans C.A., Miller N.J. & Paganga G. (1997) Antioxidant properties of phenolic compounds. Trends in Plant Science 2, Robak J. & Gryglewski R.J. (1988) Flavonoids are scavengers of superoxide anions. Biochemical Pharmacology 37, Salah N., Miller N.J., Paganga G., Tijburg L., Bolwell G.P. & Rice-Evans C.A. (1995) Polyphenolic flavanols as scavengers of aqeous phase radicals and as chain-breaking antioxidants. Archives of Biochemisty and Biophysics 322, Stewart A.J., Bozonnet S., Mullen W., Jenkins G.I., Lean M.E.J. & Crozier A. (2000) Occurrence of flavonols in tomatoes and tomato-based products. Journal of Agricultural and Food Chemistry 48, Tan S.C. (1980) Phenylalanine ammonia-lyase and the phenylalanine ammonia-lyase inactivating system: effects of light, temperature and mineral deficiencies. Australian Journal of Plant Physiology 7, Received 6 April 2001; received in revised form 3 July 2001; accepted for publication 3 July 2001