Introduction

There has been a renewed interest in the evaluation of the phenolic content of fruits and vegetables mainly due to recent epidemiological studies. These have indicated that a diet with a high fruit and vegetable consumption has health benefits in the prevention of chronic diseases, including cardiovascular disease and certain types of cancer [1, 2]. In the last few years there has been increasing evidence of the possible role of phenolic compounds in these health benefits. Attention has been mainly paid to caffeic acid derivatives (caffeoyl-quinic acid, chlorogenic acid, etc [3, 4]) and flavonols (mainly quercetin and its derivatives [2, 5, 6]). The contribution of dietary flavonols to improved health has been demonstrated to be related to their high antioxidant activity [7]. The occurrence of at least two main flavonol glycosides (quercetin 3- O -sophoroside and kaempferol 3- O -sophoroside) in broccoli inflorescences has been reported [8].

A significant variation in the level of phenolic compounds in plant tissue may have resulted from variations in environmental and agronomic factors such as water availability (irrigation), soil composition (mineral and organic nutrients) as well as from intensity of sulphur fertilisation [9, 10], mainly due to different effects on phenolic enzymes. However, some of the topics covered in this report are sparsely covered in the literature, and when information is available, it is often old and contradictory. Rossiter and Barrow [10] showed that sulphur deficiency increased phenolic compounds concentration in clover and that a severe deficiency almost doubled the concentration of phenolics. On the other hand, light intensity and quality (wavelength) also have an important effect on phenolic metabolism as they affect flavonoid biosynthesis in general [11, 12]. Thus, Taylor [13] reported differences in phenolic compounds content between species and between seasons (daylight and photoperiods) in Xanthium pennsylvanicum.

Therefore, the purpose of the present work was to evaluate genetic and environmental influences on the type and quantity of total flavonoids, caffeoyl-quinic acid derivatives and sinapic acid derivatives as well as vitamin C in the edible portions of eight genotypes of fresh harvested broccoli inflorescences ( Brassica oleracea L. var. italica) including three experimental (SG-4515, I-9904, I-9809) and five commercial cultivars (Vencedor, Furia, Pentathlon, Monterrey and Marathon). All cultivars were grown under different climatic (early vs. late crop) and sulphur fertilisation (poor vs. rich) conditions. This report describes phenolic compounds and vitamin C levels in a threefold interaction among cultivar (C), season (S) and fertilisation (F) conditions, ( i.e. C×S, C×F, S×F, C×S×F)in fresh harvested broccoli inflorescences.

Materials and methods

Plant material and agronomic conditions

Broccoli cultivars ( Brassica oleracea L. var. italica) including five commercial and three experimental (hybrids, open-pollinated varieties, Land races) used for the assays were obtained from the Centro de Investigacion y Desarrollo Agroalimentario (Murcia, Spain). Sowing dates were 20 December 2000 and 3 March 2001 for the early and late seasons respectively. For both seasons, seedlings were transplanted 30 days after sowing in one splitted plot of Finca Torreblanca (La Alberca, Murcia, Spain). Harvest of the early season occurred between 60 and 62 days after planting. Harvest of the late season occurred between 50 and 52 days after planting. Water and pesticides were applied according to standard cultural practices in La Alberca (Murcia). The plot was placed on a clay soil. Before transplanting, the soil was fertilised with 150 kg/ha supplied in the form of ammonium nitrate, 75 kg/ha of P2O5, and 200 kg/ha of K2O and S were applied as calcium sulphate (13% S) at two levels, 15  kg/ha (poor sulphur fertilisation) and 150 kg/ha (rich sulphur fertilisation). Three subplots (15×20 m) for each cultivar were used for the statistical design.

Broccoli processing

At optimum maturity, 24 inflorescences were randomly selected from uniform size plants, free from insect and/or mechanical damage, and immediately transported to the laboratory where the edible portions were cut. For analytical purposes, a total of 12 inflorescences were randomly selected, comprising three replicates of four inflorescences each cultivar, season and sulphur fertilisation. Subsamples of 20 g from each plant per replicate were combined, weighed, frozen at −70 °C and freeze-dried. This tissue was ground into a fine powder and stored at −20 °C for further analysis.

Extraction and determination of phenolic compounds

Extraction procedures have been analysed as previously described for phenolic compounds [14]. Samples (20 μl) were analysed on a Merck-Hitachi liquid chromatograph equipped with a pump (model L-6200) and a UV-VIS detector (model L-7420). Separations were achieved on a LiChroCART column (Merck, Darmstadt, Germany, ODS-18, 25×0.4 cm; 5 μm particle size). The mobile phase was water/formic acid (95:5, v/v) (A) and methanol (B), and a linear gradient, starting with 10% B to reach 15% B at 5 min., 30% B at 20 min., 50% B at 35 min. and 90% B at 40 min, was used. The flow rate was 1 ml/min. Chromatograms were recorded at 320 nm and 360 nm. Caffeoyl-quinic acid derivatives were quantified as chlorogenic acid (5-caffeoylquinic acid, Sigma, St. Louis, USA), sinapic acid derivatives as sinapic acid (Sigma), and total flavonoids, as the addition of all the peak areas, as quercetin 3-rutinoside (Sigma). Results were expressed as milligrams of phenolic compound per kilogram of broccoli (fresh weight).

Extraction and determination of vitamin C

Extraction and determination of vitamin C. Ascorbic acid (AA) and dehydroascorbic acid (DHAA) contents were determined according to Zapata and Dufour [15]. HPLC analysis of vitamin C (AA+DHAA) was achieved after derivatisation of DHAA into the fluorophore 3-(1,2-dihydroxyethyl)furo[3,4-b]quinoxaline-1-one (DFQ), with 1,2-phenylenediamine dihydrochloride. Samples (20 μl) were analysed with a Merck-Hitachi (Tokyo, Japan) liquid chromatograph equipped with an L-4000 UV detector and an L-6000 pump. Separations of DFQ and AA were achieved on a Kromasil 100 C-18 column (25×0.4 cm; 5 µm particle size; Tecnokroma, Barcelona, Spain). The mobile phase was methanol-water (5:95, v/v) containing 5 mM cetrimide (cetyltrimethyl ammonium bromide) and 50 mM potassium dihydrogen phosphate at pH 4.5. The flow rate was 0.9 ml/min. The detector wavelength was initially set at 348 nm, and after elution of DFQ, was manually shifted to 261 nm for AA detection. Standard solutions, column conditioning, and derivatisation procedures have been previously described [14].

HPLC-MS analysis

These analyses were performed according to previous work [16], using an Agilent HPLC system (Agilent, Waldbronn, Germany) equipped with a binary pump G1312A, autosampler G1313 A, photodiode array detector G1315B, controlled by Agilent software v. A.08.03 and degasser G1322A, under the same chromatographic conditions as described above for HPLC analyses.

Statistical analysis

A completely randomised experiment design was performed. The results were submitted to a multifactorial analysis of variance for the three main factors (C, S, F) and a unifactorial analysis for each main factor. Means values were compared using the least significant difference method. Statistical analysis permitted evaluation of several effects and their interaction on the determination, and quantification of these.

Results and discussion

Hydroxycinnamoyl derivatives were identified by their chromatographic behaviour and UV spectra, HPLC-MS (Table 1) and chromatographic comparisons with authentic markers. The pattern found in broccoli cultivars (Fig. 1, Table 2 , Table 3), was similar to that previously described by other authors [8, 17, 18], even if quantitative differences were found for the different cultivars and treatments. Thus, hydroxycinnamic acid compounds, found in lower amounts, identified as 1,2,2'-trisinapoylgentiobiose and 1,2,2'-triferuloylgentiobiose (compounds 7 and 8 respectively, Fig. 1, Table 2, Table 3), were not detected in some cultivars with certain treatments. However, the rest of hydroxycinnamic acid derivatives were common to all cultivars, season and fertilisations. In addition, as far as we are aware, 1-sinapoyl-2,2'-diferuloylgentiobiose, 1,2-diferuloylgentiobiose and 1,2,2'-triferuloylgentiobiose were identified in broccoli inflorescences for the first time.

Table 1. HPLC-DAD-MS analysis of broccoli inflorescences phenolics. sh Spectrum shoulder
Full size table
Fig. 1.
figure 1

HPLC chromatograms of Marathon cv. ( F) Total flavonoids recorded at 360 nm. C1Neochlorogenic acid (3- O -caffeoyl-quinic), C2Chlorogenic acid (5- O -caffeoyl-quinic), 11,2-disinapoylgentiobiose, 21-sinapoyl-2-feruloylgentiobiose, 31,2-diferuloylgentiobiose, 41,2,2'-trisinapoylgentiobiose, 51,2'-disinapoyl-2-feruloylgentiobiose, 61-sinapoyl-2,2'-diferuloylgentiobiose, 71,2,2'-trisinapoylgentiobiose, 81,2,2'-triferuloylgentiobiose, recorded at 320 nm

Full size image

Moreover, a number of individual flavonoids (15–20, depending on the cultivar, season and fertilisation) were detected but not fully identified (currently under study). These were mainly quercetin and kaempferol glycosides, in agreement with previous reports on broccoli florets [8, 14]. Other acylated derivatives were also present.

It is also remarkable that, in some cases, total and individual flavonoids, hydroxycinnamic acid derivatives and vitamin C were found with no significant differences in multifactorial analysis among cultivar, season and fertilisation. However, the general tendency showed that there were more phenolic compounds and vitamin C in the late season than in the early and in rich fertilisation than in poor.

Variation of flavonoids

The total flavonoid content of the different broccoli cultivars is shown in Table 2. Of the cultivars, Vencedor and Monterrey showed the highest contents of these compounds while I-9904 and SG-4515 showed the lowest concentrations. The Vencedor cultivar reached the highest value (970.1 mg/kg f.w.) of total flavonoids in the late season and rich fertilisation. This value was threefold higher than that previously reported by Price et al. [8]. In contrast, SG-4515 contained only 68.1 mg/kg f.w. in the early season in poor fertilisation (Table 2). Total flavonoids levels were significantly affected by cultivar, season, fertilisation and all interactions (Table 3).

Table 2. Total flavonoids and caffeoyl-quinic derivatives levels (milligrams/kilogram f.w.) in the rich and poor fertilisations of the eight broccoli cultivars in early and late seasons. Flavonoids as rutin, and caffeoyl-quinic acid derivatives as chlorogenic acid at 320 nm
Full size table
Table 3. Analysis of variance for the total flavonoids and caffeoyl-quinic acid derivatives in eight broccoli cultivars. LSD values are in brackets
Full size table

These results are in agreement with those previously reported by Gil et al. [11] and Dussi [12] who showed that sun irradiation had some effects on phenolic metabolism, as it affected flavonoid biosynthesis in general. Further, according to Tso et al. [19] and Taylor [13] there was a good correlation between higher levels of flavonoids and longer light treatments. In fact, late season had more sunlight (349 h) than early (264 h), (Table 4) which could explain the significant differences found between early and late season with a higher amount of total flavonoids for the late one.

Table 4. Average maximum and minimum air temperatures, sunlight, and rainfall from January to July for La Alberca area
Full size table

In all cases (except Monterrey) the amount of total flavonoids with rich fertilisation was significantly larger than with the poor (about two- to threefold, Table 2). Previous studies have shown that sulphur deficiency increases clover phenolics [10], in contrast to our results, which confirmed a big increase in total flavonoids for rich fertilisation versus poor fertilisation in all broccoli cultivars. An explanation for the large amount of total flavonoids, found here could be the extreme agronomic and environmental conditions (with a high temperature in late season and a high sulphur fertilisation dosage) to which broccoli was exposed. This led to the inhibition of both phenolic enzymes polyphenol oxidase (PPO) and peroxidase (POD) because of the abiotic stress, as previously reported by Tomás-Barberán and Espín [20]. More research on these topics is needed, specifically the different effects on phenolic compounds and phenolic enzymes (phenyl amonio lyase, POD and PPO) depending on factors such as agronomic practices.

Variation of hydroxycinnamoyl derivatives

The hydroxycinnamoyl derivatives contents of the different cultivars are shown in Table 2 andTable 5. Among the cultivars, Monterrey was the richest in these compounds and reached the highest total sinapic and feruloyl acid derivatives value, containing 201.4 mg/kg f.w., similar to that previously reported by other authors [18, 21], in which broccoli inflorescences contained several mixed feruloylsinapoyl esters of gentiobiose. Regarding caffeoyl-quinic derivatives, these were found in large amounts, reaching values close to those found in total sinapic derivatives (Table 2). Thus, Monterey cv. presented the highest total caffeoyl-quinic derivatives value (150.9 mg/kg f.w., Table 2), sevenfold higher than previously reported by other authors [18, 21]. On the other hand, SG-4515 and I-9809 were the cultivars that contained the lowest values of sinapic and caffeoyl-quinic derivatives respectively (Table 2 and Table 5).

Table 5. Total and individual sinapic and ferulic derivatives levels (milligrams/kilogram f.w.) in the rich and poor fertilisations of the eight broccoli cultivars in early and late seasons. Sinapic and feruloyl derivatives at 320 nm. Values represent the mean of three replicates per cultivar ( n =12)
Full size table

When analysing individual hydroxycinnamoyl derivative levels it could be seen that these were significantly affected by cultivar, season, fertilisation and all interactions (Table 3, Table 6). The main sinapic and feruloyl acid derivatives found in all cultivars were 1,2-diferuloylgentiobiose, 1,2,2'-trisinapoylgentiobiose and 1,2'-disinapoyl-2-feruloylgentiobiose (representing values between 23% and 32% of total sinapic and feruloyl acid derivatives in all cultivars). In the present work, the highest individual sinapic and ferulic acid derivative was found for 1,2-diferuloylgentiobiose (59.8 mg/kg f.w.), in poor fertilisation and late season for Monterrey cv. (Table 5). With regard to caffeoyl-quinic derivatives, the highest value was the compound identified as neochlorogenic acid (117.7 mg/kg f.w.) with poor fertilisation in the late season in Monterrey cv. too (Table 2). All the individual compounds were significantly affected by cultivar, season, fertilisation and all interactions except 1,2,2'-trisinapoylgentiobiose, with no significant differences for S×F interaction (Table 6). Another exception was found for 1,2,2'-triferuloylgentiobiose, with no significant differences for season (Table 6).

Table 6. Analysis of variance for the total and individual sinapic and ferulic acid derivatives in eight broccoli cultivars. LSD values are in brackets. NS Not significant
Full size table

Vitamin C content

Vitamin C levels were significantly affected by cultivar, season, fertilisation and all interactions except for S×F (Table 7, Table 8). Therefore, no significant differences were noted in vitamin C when comparing different fertilisation at both growing seasons. These levels were also irregular, ranging from 64.1 mg/100 g f.w. in Furia to 121.7 mg/100 g f.w. in SG-4515 (Table 7). It is remarkable that this last amount is higher than that previously reported by other authors [22, 23]. Results showed that commercial varieties rendered higher amounts of vitamin C than experimental cultivars (similar to what was observed for phenolic compounds). However, the highest amount was found in SG-4515, which reached a value of 121.7 mg/100 g f.w. (Table 7). As a general rule, cultivars grown under rich fertilisation conditions showed no significant differences to those grown under the poor one. However, in general, regarding the season, those cultivars grown in the late season shower higher vitamin C contents than those grown in the early one.

Table 7. Vitamin C levels (milligrams/100 g f.w.) in the rich and poor fertilisation of the eight broccoli cultivars in early and late seasons. Values represent the mean of three replicates per cultivar ( n =12)
Full size table
Table 8. Analysis of variance for the vitamin C in eight broccoli cultivars. LSD values are in brackets. NS Not significant
Full size table

In conclusion, the results obtained in this work showed the significant influence of the sulphur fertilisation on the amount of flavonoids and hydroxycinnamic acid derivatives. Thus, in general, rich fertilisation and late season showed higher concentrations than poor fertilisation and early season. On the contrary, in the case of vitamin C, cultivars in general presented no significant differences of these compounds between rich and poor sulphur fertilisation. However, in general, regarding the season, those cultivars grown in the late season shower higher vitamin C content than those grown in the early one. Also, the large amount of total flavonoids compounds quantified here, in comparison with what has been previously reported for broccoli inflorescences is remarkable. Moreover, Monterrey and Vencedor were the most interesting commercial cultivars due to their large amounts of phenolics and vitamin C. On the other hand, SG-4515 was the only experimental cultivar with high concentrations of total hydroxycinnamoyl acid derivatives. The wide range of variability in phenolics, among and within the eight cultivars that were evaluated in this study, offer an important base for developing cultivars with enhanced health benefits. ,It should be noted that the proper identification of new compounds in broccoli will increase a little more the interest in this well-studied vegetable.

Finally, more research is needed for the final explanation of the relationship among environment, agronomic practices and phenolic enzymes in broccoli inflorescences.