Effect of ammonium: nitrate nutrient ratio on nitrate and glucosinolate contents of hydroponically-grown rocket salad (Eruca sativa Mill.)

Abstract

The effect of the ammonium-nitrogen (NH⁺ ₄-N) to nitrate-nitrogen (NO⁻ ₃-N) ratio on NO⁻ ₃ and glucosinolate (GSL) contents in rocket salad (Eruca sativa Mill.) was investigated. Rocket salad plants were provided with five nutrient solutions with the same total nitrogen (N) level (10 mmol L⁻¹), but with different (0, 25, 50, 75 and 100) percent mole ratios of NH⁺ ₄ to NO⁻ ₃-nitrogen (PMR-N). Rocket growth (height and dry weight [DW] of the leaves and roots) was severely inhibited at a PMR-N of 100. The leaves were withered and showed chlorotic and necrotic phenomena from NH⁺ ₄ toxicity. Leaf NO⁻ ₃ and sulfate (SO²⁻ ₄) contents sharply decreased with increasing PMR-N. Six GSLs including glucoraphanin, 4-(β-D-glucopyranosyldisulfanyl) butyl GSL, glucoerucin, glucobrassicin, dimeric 4-mercaptobutyl GSL and 4-methoxyglucobrassicin were identified from rocket salad by their retention times using high-pressure liquid chromatography and confirmed using electrospray ionization mass spectrometry (ESI-MS). An unknown compound (m/z 414 as desulfo-GSL) only appeared at a PMR-N of 100. This compound together with the other GSLs appears to be involved in detoxification of NH⁺ ₄ toxicity. Total GSL content in the leaves ranged from 9 to 13 µmol kg⁻¹ DW, with the highest content occurring at a PMR-N of 50 and the lowest value at a PMR-N of 100. In contrast, total GSL content in the roots ranged from 31 to 48 µmol kg⁻¹ DW, with the lowest value occurring at PMR-Ns of 50 and 100. The four GSLs glucoraphanin, 4-(β-D-glucopyranosyldisulfanyl)butyl GSL, glucoerucin and dimeric 4-mercaptobutyl GSL were the major compounds in the leaves, whereas glucoerucin was found in great quantities in the roots.

Key words:

• ammonium
• Eruca sativa
• glucosinolate
• nitrate
• rocket salad

INTRODUCTION

Rocket salad (Eruca sativa Mill.), known as “rucola” in Italy, is an annual herbaceous plant belonging to the Brassicaceae (Cruciferae) family. This leafy vegetable is commonly used as a salad or as a steamed vegetable in European countries, and the seeds are used as an important oil source in India and Pakistan (Jirovetz et al. 2002; Yadava et al. 1998). In Japan, rocket salad crop has spread quickly in the past decade and shows increasing economic potential because of its short biological cycle (40–60 days), spicy hot taste and sesame-seed-like aroma, and it is a popular vegetable among the Japanese, similar to the turnip rape (Brassica rapa L.). When crushed the leaves release a strong hot flavor from certain glucosinolates (GSLs), sulfur-containing glucosides, and their associated breakdown products. At present, more than 120 different GSLs have been identified in 11 plant families, but mainly in the Brassicaceae (Fahey et al. 2001).

Plant GSL levels are affected by genetic, botanical, nutritional and environmental factors including varieties, climatic conditions and agronomic practices (Fenwick et al. 1989). Stress factors such as water stress and high plant density, are also known to increase GSL levels (Macleod and Pikk 1979; Shattuck and Wang 1994). Among the agronomic practices, the intensity of nitrogen (N) and sulfur (S) fertilization is of primary importance for GSL synthesis and levels because N and S are the major elemental constituents of GSLs, which are derived from chain-elongated forms of amino acids, notably from methionine and phenylalanine (Mithen et al. 2000). Nitrogen and S nutrition and their effects on total GSL content in Brassica vegetables have been well documented (Schnug 1989; Zhao et al. 1993). Sulfur supply has the greater influence on seed GSL content (Schnug 1989). However, Zhao et al. (1993) insisted that N supply was a further important factor influencing the variation in seed GSL content. They showed in low GSL oilseed rape (Brassica napus L.) that increasing the N supply decreased the seed GSL content when S was deficient. They suggested that the balance between N and S supply played an important role as a regulator in GSL biosynthesis.

However, the effects of different N forms on total GSL content have received little attention. Ammonium (NH⁺ ₄) and nitrate (NO⁻ ₃) are the two major sources of inorganic nitrogen taken up by the roots of higher plants. The effect of these two forms on plant growth is dependent not only on the plant species, but also on NH⁺ ₄ : NO⁻ ₃ ratios and concentrations, although most plants can use either or both forms as a source of N (Marschner 1997). Plant accumulation of NH⁺ ₄ and NO⁻ ₃ is principally determined by the concentration of the ion and pH of the growth media, with NH⁺ ₄ uptake favored by alkalinity and NO⁻ ₃ uptake by mild acidity (Haynes and Go 1978). Lowering the NO⁻ ₃ concentration in leafy vegetables is desirable because of possible negative effects (the presence of NO⁻ ₂ intoxicant metabolites) of high NO⁻ ₃ uptake (Marschner 1997). Thus, NO⁻ ₃ accumulation in leafy vegetables can be of considerable importance for their quality. The NO⁻ ₃ content of rocket salad has been mandated to not to exceed 2.5–4.0 g kg⁻¹ fresh weight for Italian export (Santamaria 2002, 2006). The NO⁻ ₃ content in plants can be reduced by decreasing NO⁻ ₃ in the nutrient solution or replacing it with NH⁺ ₄ a few days before harvest. Santamaria et al. (1998a) reported that NO⁻ ₃ accumulation by rocket salad and chicory (Cichorium intybus L.) was reduced by more than half without influence on relative growth rate when available NO⁻ ₃-N was diminished in 75% of the NO⁻ ₃ in the nutrient solution (either by transfer from 2 mmol L⁻¹ to 0.25 mmol L⁻¹ N or replacement with NH⁺ ₄) within 5–6 days before harvest. In contrast, NH⁺ ₄, in particular NH₃, is toxic at quite low concentrations. The formation of amino acids, amides and related compounds is the main pathway of detoxification of either NH⁺ ₄ ions or NH₃ (Marschner 1997). NH⁺ ₄ leads to serious physiological and morphological disorders resulting in chlorosis, restricted growth and in some cases death (Barker et al. 1966; Goyal 1982).

In general, good plant yields are obtained through a combined supply of both NO⁻ ₃ and NH⁺ ₄, but also strongly depend on the pre-existing concentrations in the field as well as the total concentrations supplied. However, the effects of percent molar ratios of nitrogenous nutrients on the GSL content in rocket salad are poorly documented. In the present experiment, therefore, we have sought to assess whether NO⁻ ₃ and GSL contents in rocket salad were affected by changing NH⁺ ₄ : NO⁻ ₃ percent molar ratios in the nutrient solution.

MATERIALS AND METHODS

Chemicals

High-pressure liquid chromatography (HPLC)-grade acetonitrile was purchased from Wako Pure Chemical Industries (Osaka, Japan). Aryl sulfatase (Type H-1, EC 3.1.6.1) was purchased from Sigma Chemical Company (St Louis, MO, USA). DEAE-Sephadex A-25 was purchased from Amersham Biosciences (Uppsala, Sweden) and sinigrin (allylGSL) was purchased as an internal standard from Tokyo Kasei Kogyo Company (Tokyo, Japan).

Plant materials

Rocket seeds (E. sativa Mill. “Odyssey”) were purchased from Sakata Seed Company (Yokohama, Japan). The seeds were sown in a cell pot containing vermiculite (13 September 2002) and the seedlings were transferred to a plastic container (300 mm × 130 mm × 120 mm, 1.8 L) 2 weeks later (14 days after sowing [DAS]). The plants were grown in a glasshouse (3 m × 3 m) at the National Agricultural Research Center for Hokkaido Region (longitude 141°21′ E; latitude 43°04′ N). The temperature in the glasshouse was maintained at over 20°C by a boiler system and natural light periods were reinforced for 4 hours (16:00–20:00) with two 500 W fluorescent lamps.

The percent molar ratio of NH⁺ ₄- to NO⁻ ₃-nitrogen (PMR-N) was defined as:

(1)

where, [NH⁺ ₄-N] is the concentration of NH⁺ ₄-N in the nutrient solution (mmol L⁻¹), and [NO⁻ ₃-N] is the concentration of NO⁻ ₃-N in the nutrient solution (mmol L⁻¹).

Five nutrient solutions with the same total N content (10 mmol L⁻¹) but with PMR-N values of 0, 25, 50, 75 and 100 were used. Plants were transplanted on 11 October (28 DAS; Table 1). The micronutrients were supplied in all treatments as: MnCl₂·4H₂O (1.81 g L⁻¹), H₃BO₃ (2.86 g L⁻¹), ZnSO₄·7H₂O (0.22 g L⁻¹), (NH₄)₆Mo₇O₂₄·4H₂O (0.09 g L⁻¹), CuSO₄·5H₂O (0.08 g L⁻¹) and Fe-ethylenediamine–N, N, N¹, N¹–tetraacetic acid (EDTA) (7.54 g L⁻¹) according to the Hoagland type solutions (Hoagland and Arnon 1938). The initial pH of the nutrient solution was adjusted to 6.0.

Table 1 Composition of nutrient solutions for different NH4 : NO3 percent molar ratios (PMR-N) in a 10 mmol L−1 total hydroponic nutrition solution (mmol L−1)

Compound	PMR-N
	0	25	50	75	100
KCI	−	3	5	6	6
(NH4)2SO4	−	−	I	1	1
Ca(NO3)2	2	2	2	−	–
NaNO3	2	−	−	−	–
KNO3	4	1	1	−	−
NH4NO3	−	2.5	−	2.5	−
(NH4)3PO4·3H2O	−	−	1	I	1
CaCl2	−	−	−	2	2
MgSO4	1	1	−	−	−
MgCl2	−	−	1	1	1
K2HPO4	1	1	−	−	−
NH4Cl	−	−	−	−	5
Micronutrients^†	+	+	+	+	+
^†Micronutrients were the same for all treatments applied according to Hoagland type solutions (Hoagland and Arnon 1938).

Two plants were grown in each container, 0.1 m apart, and each treatment was replicated five times. Plants were watered in the morning, every 2–3 days, with the nutrient solution. Five plants, one from each ratio treatment, were randomly harvested on 5 November (53 DAS), representing a treatment duration of 25 days (TD 25). After measuring the length of the longest leaf among approximately 15 leaves, using a ruler, the plant samples were separated into leaves and roots and were lyophilized, ground and stored in a plastic bottle until chemical analysis.

NO⁻ ₃ and SO²⁻ ₄ contents

A 50 mg ground sample was placed in a 100 mL glass bottle and 50 mL of deionized water was added and the suspension shaken for 1 h. The extracts were filtered through Whatman No. 6 filter paper. The NO⁻ ₃ and SO²⁻ ₄ concentrations were determined after dilution using an ion chromato-graph (Dionex Qic Analyzer, Dionex Corporation, Sunnyvale, CA, USA) with an AS4A column (4 mm × 250 mm). The solvent was 2.0 mmol L⁻¹ Na₂CO₃ and NaHCO₃. The chromatograph was operated in the constant flow mode at 1.4 mL min⁻¹ (slightly modified from Takebe et al. 1995).

Extraction of glucosinolates

A 100 mg aliquot of the ground sample was placed in a 2.0 mL microcentrifuge tube and crude GSLs were extracted with 1.5 mL of 70% (v/v) boiling methanol maintained in a water-bath at 70°C for 5 min. The mixture was centrifuged at 12,000 g for 10 min and the supernatant collected. The residue was re-extracted twice as described above. The combined supernatant was taken as the crude extract of GSLs (Bjerg and Sørensen 1987).

Preparation of desulfo-glucosinolates

The crude extracts were applied to a mini column (using a 1,000 µL pipet tip) packed with DEAE-Sephadex A-25. Glucosinolates were desulfated by the addition of a solution of aryl sulfatase onto the column. After overnight reaction at ambient temperature desulfo-GSLs (DS-GSLs) were eluted into a 2.0 mL microcentrifuge tube with 4 × 0.5 mL of de-ionized water. HPLC profiles were obtained by injecting samples of a known volume (20 µL).

High-pressure liquid chromatography analysis

The HPLC analysis was carried out with a CLASS-VP Chromatography Data System (Shimadzu, Kyoto, Japan). The eluate containing DS-GSLs was injected into an Inertsil ODS-2 column (4.6 mm × 250 mm, GL Sciences, Tokyo, Japan). The detection wavelength was set at 227 nm using a Diode Array Detector (SPD-M10Avp) that was set to scan between 190 and 370 nm. Peak areas and retention times were recorded using FmV-6750 CL7S (Fujitsu, Tokyo, Japan). The flow rate was 1.0 mL min⁻¹, and the column oven temperature was set at 35°C (Minchinton et al. 1982; Spinks et al. 1984). The solvent systems used were (A) de-ionized water and (B) 20% (v/v) acetonitrile. The solvent program consisted of a linear gradient from 1 to 99% of (B) solution over a period of 18 min and then kept constant at 99% of (B) solution for 11 min (Macfarlane-Smith and Griffiths 1988).

ESI-MS analysis

The mass spectrometry (MS) data were obtained by electrospray under atmospheric pressure using an API-100 instrument (Perkin-Elmer Sciex Instruments, Pomona, CA, USA). The ESI-MS was conducted by direct injection (10 µL min⁻¹) with a 1% AcOH solution. The MS operating conditions were as follows: ionspray voltage, 4.8 kV (positive mode); orifice voltage, 40 V; nebulizer gas, air; curtain gas, nitrogen. The API 100 instrument with TUNE software (version; C-Preliminary Release) was used for data acquisition and evaluation.

Statistical analysis

Data were subjected to statistical analysis by Tukey's multiple range test using Esumi Statistical Software version 5.0 (Esumi Incorporated, Tokyo, Japan). The anova was done at a 5% level of significance.

RESULTS AND DISCUSSION

Plant growth

Plant height was significantly reduced in the NH⁺ ₄-rich 100 PMR-N treatment compared with the 25 and 50

Table 2 Effect of NH4 : NO3 percent molar ratios (PMR-N) on the growth of rocket salad

PMR-N	Height (cm)	Dry weight (g)
		Leaf	Root
0	25.5 ± 3.59ab^†	2.44 ± 0.30ab	0.22 ± 0.06ab
25	28.3 ± 2.04a	2.92 ± 1.04a	0.28 ± 0.08a
50	28.7 ± 0.77a	2.48 ± 0.29ab	0.22 ± 0.04ab
75	27.1 ± 2.63ab	2.58 ± 0.49ab	0.29 ± 0.08a
100	23.8 ± 1.71b	1.66 ± 0.13b	0.15 ± 0.02b
^†Within each column values followed by the same letters are not significantly different at P < 0.05 using Tukey's multiple range test. Values are mean ± standard deviation (n = 5).

PMR-N treatments, but the 0, 25, 50 and 75 treatments did not differ significantly from one another (Table 2). The dry weights of leaves and roots were similarly lower for the 100 PMR-N treatment, and again the other treatments showed no significant differences. For the 100 PMR-N treatment, the plants at 2 days before harvesting (51 DAS) had withered and showed chlorotic and necrotic leaves because of ammonium toxicity (Barker et al. 1966; Puritch and Barker 1967; Santamaria et al. 1998a,b). It appears that in the case of the 100 PMR-N treatment, NH⁺ ₄ uptake exceeded the plant's capacity to detoxify NH⁺ ₄.

NO⁻ ₃ and SO²⁻ ₄ contents

In general, increasing the PMR-N led to a linear decrease in both leaf and root NO⁻ ₃ and SO²⁻ ₄ contents, particularly for PMR-N values above 50 (Table 3). The changes in NO⁻ ₃ content concurred with that observed in leaves, stems and roots of tomato plants (Hartman et al. 1986). However, others (Santamaria and Elia 1997; Santamaria et al. 1998a,b) have insisted that the SO²⁻ ₄ content in chicory did not change, even when the PMR-N value reached 75, for a total nutrient N level of 4 mmol L⁻¹. In the current study, the greatest NO⁻ ₃ and SO²⁻ ₄ contents in both the leaves and roots occurred for the 0 PMR-N treatment, with or without a significant difference compared to the 25 PMR-N treatment, whereas the lowest value occurred for a PMR-N of 100. In both plant tissues, NO⁻ ₃ content was more strongly affected by increasing PMR-N than SO²⁻ ₄ content. In all treatments, the NO⁻ ₃ and SO²⁻ ₄ contents in the leaves were over two-fold greater than those in the roots, except for the NO⁻ ₃ content for the 100 PMR-N treatment. The NO⁻ ₃ was 13.3- and 6.5-fold greater in the leaves and roots, respectively, in the 0 PMR-N than in the 100 PMR-N treatment and the parallel increase in SO²⁻ ₄ content was over 2.1-fold for both tissues. These differences were very small compared to those reported by Santamaria et al. (1998a). However, the difference in

Table 3 Effect of NH4 : NO3 percent molar ratios (PMR-N) on the nitrate and sulfate content (g kg−1 fresh weight) of rocket salad leaves and roots

PMR-N	Leaf		Root
	NO3	SO4	NO3	SO4
0	7.2 ± 0.66a^†	3.6 ± 0.27a	3.1 ± 0.05a	1.6 ± 0.15a
25	6.4 ± 0.47ab	3.5 ± 0.21a	2.2 ± 0.09b	1.4 ± 0.13a
50	6.1 ± 0.55b	3.1 ± 0.19b	1.8 ± 0.14b	1.1 ± 0.09b
75	2.9 ± 0.26c	2.6 ± 0.27c	1.0 ± 0.09c	0.9 ± 0.06b
100	0.5 ± 0.04d	1.7 ± 0.13d	0.5 ± 0.05d	0.7 ± 0.03c
^†Within each columm values followed by the same letters are not significantly different at P < 0.05 using Tukey's multiple range test. Values are mean ± standard deviation (n = 5).

NO⁻ ₃ content between the 0 and 50 PMR-N treatments was similar to the 1.2-fold difference found between chicory and rocket leaves. Studies of rocket salad obtained from Italian markets have shown it to be the vegetable with the highest NO⁻ ₃ content (963–4305 mg kg⁻¹ fresh weight) (Santamaria et al. 1999). In the present study, NO⁻ ₃ content decreased 60% for the 75 PMR-N treatment compared to the 0 PMR-N treatment without reducing plant growth, suggesting that NO⁻ ₃ content could be decreased by using combined NH⁺ ₄ and NO⁻ ₃ applications, particularly at a PMR-N of 75.

Total glucosinolate content and individual glucosinolates

Six GSLs including glucoraphanin, 4-(β-D-glucopyranosyldisulfanyl)butyl GSL, glucoerucin, glucobrassicin, dimeric 4-mercaptobutyl GSL and 4-methoxyglucobrassicin, were identified by their retention times using HPLC and confirmed using ESI-MS analysis based on their MS data (Table 4). Glucoraphanin is of particular interest because its isothiocyanate (ITC), sulforaphane (4-methylsulfinylbutyl ITC), is considered to be one of most potent inducers of phase II proteins (Tawfig et al. 1995). Glucoerucin possesses direct antioxidative activity because of its ability to decompose hydroperoxides and hydrogen peroxide (Barillari et al. 2005). Two minor indolyl GSLs (glucobrassicin and 4-methoxyglucobrassicin) and two unique GSLs (dimeric 4-mercaptobutyl GSL and 4-[β-D-glucopyranosyldisulfanyl]butyl GSL) were identified in rocket leaves. In addition, we detected another unknown GSL (m/z 414 as DS-GSL) in both leaves and roots under the 100 PMR-N treatment, and its identification is currently under way in our laboratory. We suggest that three GSLs, dimeric 4-mercaptobutyl GSL, 4-(β-D-glucopyranosyldisulfanyl)butyl GSL and the unknown GSL (based on its putative formula), may be closely related to 4-mercaptobutyl GSL (m/z 327 as DS-GSL) in the pathway of GSL biosyntheses in rocket tissues (Bennett et al. 2002; Kim et al. 2004).

Table 4 Structures and names of glucosinolates (GSLs) isolated from rocket salad

No.^†	Structure of side chain (R)-groups	Semisystematic name	Trivial name	Response factor^‡
1	Unknown^§	−	−	1.00^¶
2	CH3-SO-CH2-CH2-CH2-CH2-	Dimeric 4-metbylsulfinylbutyl GSL	Glucoraphanin	1.07
3	C6H11O5-S-S-(CH2)4-	4-(β-D-glucopyranosyldisulfanyl)butyl GSL	−	1.00^¶
4	CH3-S-CH2-CH2-CH2-CH2-	4-Methylthiobutyl GSL	Glucoerucin	1.00^¶
5	Indole-3-CH2-	Indolyl-3-methyl GSL	Glucobrassicin	0.29
6	{-(CH2)4-S-S-(CH2)4-}	dimeric 4-mercaptobutyl GSL	−	1.00^¶
7	Indole-4-OCH3-	4-Metboxyindolyl-3-methyl GSL	4-Methoxyglucobrassicin	0.25
^†Elation order of GSL from high-pressure liquid chromatography profiles.
^‡The International Organization for Standardization (ISO 9167-1, 1992).
^§Unknown has not yet been identified, but showed an electrospray ionization mass spectrometry-based molecular weight of m/z 414 as desulfo-GSL.
^¶Not yet decided by the ISO.

The total GSL content in the leaves ranged from 9 to 13 µmol kg⁻¹ dry weight (DW) (Table 5). Total GSL content was significantly greater for the 50 PMR-N treatment than all treatments except the 75 PMR-N treatment. Conversely, the total GSL content for the 100 PMR-N treatment was less than that of any other treatment. This result showed a similar tendency to that of plant growth (i.e. total GSL content of the leaves was much reduced in the 100 PMR-N treatment because of probable plant injury).

Individual GSL levels in the leaves occurred in the following order of prevalence: glucoerucin, glucoraphanin, dimeric 4-mercaptobutyl GSL, 4-(β-D-glucopyranosyl disulfanyl) butyl GSL and two minor indolyl compounds (glucobrassicin and 4-methoxyglucobrassicin) in all treatments, except for the 75 PMR-N treatment in which glucoraphanin and dimeric 4-mercaptobutyl GSL were reversed (Table 5). A high level (30–80 µmol kg⁻¹ DW) of total GSL content in flower buds of Japanese turnip rape (B. rapa L.) is occasionally a problem because of the resulting bitter taste (Kim et al. 2003). In 1991, the European Community imposed a limit of 35 µmol kg⁻¹ DW for GSL levels in rapeseeds because of their potentially detrimental effects to humans and livestock (Milford and Evans 1991). However, although GSLs in rocket leaves were detected at relatively low levels, rocket salad nonetheless had a strongly hot flavor.

In contrast, the total GSL content in the roots ranged from 31 to 48 µmol kg⁻¹ DW (Table 5), and did not differ significantly between the 0, 25 and 75 PMR-N treatments or between the 50 and 100 PMR-N treatments. Unfortunately, we do not currently have an explanation for the result recorded in the 50 PMR-N treatment. Individual GSL levels in the roots occurred in the following order of prevalence: glucoerucin, glucoraphanin, dimeric 4-mercaptobutyl GSL and two minor compounds (4-[β-D-glucopyranosyldisulfanyl]butyl GSL and Glucobrassicin). However, 4-methoxyglucobrassicin was not found, even as a trace, in any of the treatments. Total root GSL levels were two-fold to five-fold greater than the levels in the leaves. In terms of GSL biosynthesis in E. sativa (Bennett et al. 2002), glucoerucin was present at high levels in the roots and served as a precursor in the formation of glucoraphanin (by S-oxidation) and (dimeric) 4-mercaptobutyl GSL (by S-demethylation), whereas in the leaves glucoerucin served in the formation of three main GSLs: glucoraphanin, 4-(β-D-glucopyranosyldisulfanyl)butyl GSL and (dimeric) 4-mercaptobutyl GSL.

Rocket salad, spinach and Swiss chard have a strong tendency to NO⁻ ₃ accumulation, ranging from 0.96 to 4.3 g kg⁻¹, 0.54 to 3.4 g kg⁻¹ and 1.3 to 4.2 g kg⁻¹ of fresh matter, respectively (Santamaria et al. 1999). Many studies have looked at reducing leaf NO⁻ ₃ content in rocket salad (Santamaria et al. 1997, 1998a, 1998b, 1999, 2002). In the present study, the NO⁻ ₃ content in both leaves and roots of rocket salad could be reduced with a partial NO⁻ ₃N substitution with NH⁺ ₄-N. However, the lowest plant growth and total GSL content was recorded in the 100 PMR-N treatment. This may be closely related to the plant's inability to resist or adapt itself to NH⁺ ₄ toxicity. Once NH⁺ ₄ absorbed, it must be immediately biosynthesized to supply plant proteins, amino acids, nucleic acids and chlorophyll as a source of N during plant growth to avoid accumulation to toxic levels (Gill and Reisenauer 1993). In the present study, the unknown GSL, only detected in the 100 PMR-N treatment, and two indolyl GSLs (glucobrassicin and 4-methoxyglucobrassicin) were probably involved in NH⁺ ₄ detoxification (Table 5). However, this result was unpredictable because of our poor understanding of the metabolic pathway involved. Many studies have examined the chemical and biological properties of indolyl GSLs (Chavadej et al. 1994; Ludwig-Müller et al. 1999; McDanell et al. 1988). 4-Methoxyglucobrassicin was affected by Zn in solution and glucobrassicin linearly responded to increasing Zn levels (Coolong et al. 2004). Indolyl GSLs, mainly 4-methoxyglucobrassicin, have been

Table 5 Effect of NH4 : NO3 percent molar ratios (PMR-N) on individual and total glucosinolate (GSL) content (g kg−1 dry weight) in the leaves and roots of rocket salad

Plant part	PMR-N	GSL							Total GSL
		Unknown A	GRA	4GPdB	GER	GBS	d4MCB	4MGB
Leaf	0	ND	2.96 ± 0.18b^†	1.84 ± 0.20ab	3.75 ± 0.18b	0.04 ± 0.00c	2.24 ± 0.14b	0.12 ± 0.02b	11.0 ± 0.28b
	25	ND	3.36 ± 0.11b	1.39 ± 0.20cd	3.99 ± 0.53ab	0.03 ± 0.00d	2.36 ± 0.43b	0.09 ± 0.00c	11.2 ± 1.14b
	50	ND	4.38 ± 0.42a	1.63 ± 0.131bc	4.38 ± 0.23ab	0.04 ± 0.00c	2.23 ± 0.09b	0.11 ± 0.01bc	12.8 ± 0.54a
	75	ND	2.46 ± 0.08c	1.94 ± 0.07a	4.58 ± 0.55a	0.05 ± 0.00b	3.10 ± 0.36a	0.11 ± 0.0lbc	12.2 ± 0.7ab
	100	0.74 ± 0.01	2.18 ± 0.21c	1.24 ± 0.11d	2.21 ± 0.15c	0.12 ± 0.01a	2.00 ± 0.17b	0.19 ± 0.02a	8.7 ± 0.38c
Root	0	ND	6.19 ± 0.58a	0.46 ± 0.04c	36.3 ± 4.88a	0.06 ± 0.01b	5.12 ± 0.56c	ND	48.1 ± 5.76a
	25	ND	5.43 ± 0.70ab	0.50 ± 0.051bc	36.7 ± 4.06a	0.05 ± 0.00bc	5.60 ± 0.43bc	ND	48.3 ± 4.76a
	50	ND	4.62 ± 0.50b	0.41 ± 0.04c	22.6 ± 3.52b	0.03 ± 0.00c	3.42 ± 0.32d	ND	31.1 ± 3.99b
	75	ND	4.96 ± 0.68b	0.67 ± 0.09a	34.9 ± 3.01a	0.08 ± 0.01a	7.75 ± 1.17a	ND	48.4 ± 4.61a
	100	0.73 ± 0.10	2.93 ± 0.42c	0.63 ± 0.13ab	20.4 ± 5.75b	0.09 ± 0.02a	7.15 ± 1.22ab	ND	31.9 ± 7.50b
Within each column values followed by the same letters are not significantly different at P < 0.05 using Tukey's multiple range test. Values are mean ± standard deviation (n = 5). GRA, glucoraphanin; 4GPdB, 4-(B-D-glucopyranosyldisulfanyl) butyl GSL; GER, glucorucin; GBS, Glucobrassicin; d4MCB, dimeric 4-mercaptobutyl, GSL; 4MGB, 4-methoxyglucobrassicin; ND, not detected.

shown to accumulate systemically in plants treated with jasmonic acid, a signal molecule associated with response to wounding, herbivory and infection (Doughty et al. 1995).

When the roots take up NH⁺ ₄ in excess, it causes lower concentrations of K⁺, Ca⁺ and Mg⁺ in order to maintain both cation–anion balance and intracellular pH (Haynes and Goh 1978; Marschner 1997). In the 100 PMR-N treatment, substantial amounts of NH⁺ ₄ may be stored in the vacuoles where GSLs are also located as potassium salt (Halkier and Du 1997). In general, nearly all of the NH⁺ ₄ taken up has to be assimilated in the roots and the assimilated N transported in the xylem as amino acids and amides to the shoot in the form of N-rich compounds to minimize the carbon costs for root-to-shoot transport (Marschner 1997). Consequently, the unknown GSL (based on its putative formula, m/z 414 as DS-GSL) and indolyl GSLs (derived from tryptophan) may be biosynthetically connected with N applications, probably NH₄-N form, based on their molecular structure, which contains two N atoms. However, the GSL biosynthesis from NH₄-N applications remains to be elucidated. There is probably a second pathway for the formation of GSLs, particularly the unknown GSL, from NH⁺ ₄-N from an N source.

ACKNOWLEDGMENT

The authors wish to thank the Japanese Society for the Promotion of Science (JSPS) for financial assistance.