Introduction

Recent decades have seen an increasing consumer interest in functional foods, which provide health-benefiting effects in supplementary to their general nutritional content [1,2,3]. These health-benefiting properties are attributed to the presence of a wide range of trace level constituents, such as carotenoids (vitamin A), anthocyanins, fatty acids, phytosterols and terpenoids [4, 5]. There is particular interest in phenolic compounds, which may provide antioxidant, anti-inflammatory, anticancer and antimicrobial properties [6,7,8]. A number of surveys have indicated that many consumers will pay a premium price for functional foods demonstrated to contain high levels of such health-benefiting compounds [9,10,11,12]. Several recent reviews have been published outlining the definition, diversity and consumer awareness of functional foods; the interested reader is referred to these publications for more details [13,14,15,16,17]. However, there is a general lack of knowledge around functional food in the Australian horticulture & fresh produce sector.

In Queensland, Australia, the horticultural sector is the second largest primary industry, worth around $2.8 billion AUD p.a. [18]. Major crops include banana, pineapple, tomatoes, mandarins and strawberries. However, there is increasing investment into the establishment of niche crops comprising internationally grown species or of heirloom cultivars of currently grown produce (e.g. pumpkins, tomatoes, melons). In general, such novel or heirloom cultivars may have increased health-benefitting or nutritional properties compared to existing commercial cultivars [19, 20]; hence may be classified as functional foods. For example, the Queen Garnet plum–recently developed in southern Queensland–contains exceptionally high levels of anthocyanins, which may reduce oxidative stress and provide thrombotic properties [21, 22]. Consequently, there may be potential to market other lesser-known horticultural products from Queensland as health-benefitting functional foods.

However, the phytochemical composition of many of these species are poorly studied, particularly when grown under the unique subtropical conditions of Queensland. It is important to have a foundational understanding of the phytochemical contents of any crops/produce marketed as a functional food. In turn, this will provide the seller or processer with the information needed to substantiate any health claims which may be made. Consequently, this study aimed to investigate the phytochemical composition of a number of niche and emerging horticultural crops from central Queensland, with a particular focus on their phenolic profiles. This should help with their ongoing commercialisation and expansion of production.

Material and methods

Reagents

Methanol (AR grade for extractions and gradient-grade for HPLC work) and ascorbic acid (vitamin C) were obtained from Chem Supply (Gillman, SA). All other reagents were of analytical grade or higher and were obtained from Sigma Aldrich Australia (Castle Hill, NSW).

Sample procurement and processing

Red seedless watermelon (Citrullus lanatus), Christmas melon (Cucumis melo cv. 'Sancho'), Golden honeydew (Cucumis melo, Inodorus group), Indian cream cobra rockmelon (Cucumis melo ssp. melo), Queensland Blue pumpkin (Cucurbita maxima), Trombone pumpkin (Cucurbita moschata), Spaghetti squash (Cucurbita pepo subsp. pepo), Apple cucumber (Cucumis sativus cv. ‘Crystal Apple’), Blood orange (Citrus × sinensis cv. ‘Moro’) and Dragonfruit (Selenicereus costaricensis) were obtained from Berry Good Produce in Rockhampton, central Queensland [23]. Citron melon fruit (Citrullus lanatus var. citroides) were collected from a farm in the Wura district, approximately 46 km south south-west of Rockhampton. Mandarin (Citrus reticulata) and Brazilian cherry (Eugenia uniflora) fruit were collected from a farm in The Caves district, approximately 23 km N of Rockhampton.

After subsampling of each fruit or vegetable type for the ascorbic acid extractions, the samples were frozen at − 80 °C and freeze-dried (− 50 °C; 50 mT) for approximately 1 week. The mass difference between the fresh and freeze-dried sample was used to calculate the moisture content. The freeze-dried samples were then ground (Coffee & Spice Grinder; Breville, Botany, NSW) and the subsequent powder used for the methanol extractions.

Extraction of ascorbic acid and phenolic compounds

Ascorbic acid was extracted from the fresh samples following previously reported methods [24]. Briefly, approximately 1 g of sample was homogenised in 13 mL of 3% w/v metaphosphoric acid at 11,000 rpm for 30 secs (Ultra-Turrax T25 blender; Rawang, Malaysia) and sonicated (20 min). After centrifugation (1000 g; 5 min), extracts were syringe filtered (Livingstone 0.45 µm PTFE membrane) prior to HPLC analysis.

For the polar compound extracts, approximately 1 g of freeze-dried powder was combined with 10 mL methanol and 0.5 mL 1 M HCl and sonicated for 30 min. The extracts were then shaken end-over-end at 50 rpm for 3 h (Ratek RM4; Melbourne, Australia) and centrifuged (1000 g; 5 min). The supernatant was collected and a further 10 mL methanol and 0.5 mL 1 M HCl was added to the pellet. After end-over-end shaking for 1 h and centrifuging, the combined supernatants were syringe filtered (0.45 µm PTFE; Livingstone) prior to subsequent use.

The methanol extracts were used for the total phenolics assay, ferric reducing antioxidant potential assay, and measurement of total monomeric anthocyanin content. Extractions were performed in duplicate for each sample and extraction protocol, with results reported as the mean of the duplicate extractions.

Phytochemical profiling

The total phenolic (TP) content, cupric reducing antioxidant capacity (CUPRAC), ferric reducing antioxidant potential (FRAP) and total monomeric anthocyanin (TMA) content were measured in the methanol extracts using previously described methods [25,26,27]. Radical scavenging activity was measured using the ABTS assay, following previous methods [28]. All results were expressed as mg of the respective standard per 100 g of flesh (dried weight). The TP content was expressed in terms of gallic acid equivalents (GAE); the CUPRAC, FRAP and ABTS as Trolox equivalents (TE) and the TMA as equivalents of cyanidin-3-glucoside (cyd-3-glu).

Analysis of ascorbic acid by HPLC–DAD

Ascorbic acid was quantified in the metaphosphoric extracts on an Agilent 1100 high-performance liquid chromatography with diode array detection (HPLC–DAD) system using previously reported methods [24]. Briefly, separation was achieved using an Agilent Eclipse XDB-C18 column (150 × 4.6 mm; 5 µm pore size) and a gradient elution held at 100% 0.01 M phosphoric acid for 3 min, ramping from 0 to 100% methanol between 3 and 5 min, and held at 100% methanol for another 3 min. The injection volume was 5 µL and detection wavelength 250 nm. Ascorbic acid was quantified against an external standard calibration (20–200 mg/L) and expressed in mg/100 g on a fresh weight basis.

LC–MS/MS analysis of phenolic compounds

Targeted liquid chromatograph tandem mass spectrometry (LC–MS/MS) was used to quantify the phenolic compounds in the methanolic extracts, using a Shimadzu LCMS-8040 system. The method used a Raptor biphenyl column (100 mM × 2.1 mM, 2.7 µM) operated at 40 °C, with an injection volume of 5 µL and flow rate of 0.6 mL/min. The mobile phase comprised 5 mM ammonium formate and 0.1% formic acid (phase A) and 5 mM ammonium formate and 0.1% formic acid (phase B). The gradient was held at 5% phase B for 2 min, then ramped to reach 85% phase B at 12 min, then it was held for a further 2 min, with 2 min equilibration time between samples.

The electrospray ionisation (ESI) MS detection was performed in positive and negative ionization modes, with the following conditions: interface temperature of 350 °C, DL temperature 250 °C, heat block temperature 400 °C, interface voltage 4.50 kV. Nitrogen was used as the nebulizing gas and drying gas, at flow rates of 3 and 15 L/min, respectively.

Multiple reaction monitoring (MRM) mode was used to target the selected compounds, with two transitions monitored for most (Table 1). The data were collected and analysed in the LabSolutions software (Shimadzu, Kyoto, Japan). The precursor m/z, product m/z, collision energy, Q1 and Q3 pre-biases were optimised using authentic standards of each compound (1 ppm concentration; data not shown).

Table 1 Analytical parameters of the LC–MS/MS method
Full size table

Data analysis

Statistical analysis was performed in R Studio, running R 4.0.2 [29]. Where applicable, results are presented as mean ± 1 standard deviation. All results are provided on a dry weight basis unless otherwise specified.

Results and discussion

Moisture and ascorbic acid content

The HPLC method for ascorbic acid showed good separation between the target analyte and the other matrix constituents in the samples (see Fig. 1). Table 2 provides the moisture and ascorbic acid content of each of the produce samples. It should be noted that some of the samples (e.g., trombone pumpkin skin) were analysed for ascorbic acid content but were not included in more detailed phytochemical profiling later in the study.

Fig. 1
figure 1

Chromatogram showing the location of the ascorbic acid peak (1) in the HPLC analysis of the apple cucumber sample

Full size image
Table 2 Moisture and ascorbic acid content of the horticultural produce samples (ascorbic acid values given as mean ± SD from 2 replicate extracts for each sample)
Full size table

The highest ascorbic acid content (on a dry weight basis) was found in the blood orange skin (817 mg/100 g), followed by the mandarin skin (610 mg/100 g). Conversely, the lowest ascorbic acid content was found in mature citron melon (15 mg/100 g).

Phytochemical profiling

The results of phytochemical profiling from the CUPRAC, FRAP, ABTS, TP and TMA assays are presented in Table 3. The highest antioxidant capacity was found in the blood orange skin using the CUPRAC assay (6418 mg TE/100 g), although the Brazilian cherry flesh showed the highest antioxidant activity using the FRAP assay (1022 mg TE/100 g). Furthermore, the spaghetti squash sampled contained the highest radical scavenging activity, as measured by the ABTS assay (3795 mg TE/100 g), highlighting the importance of using multiple methods to assess the antioxidant activity of a particular matrix. Other samples demonstrating a high antioxidant capacity included the mandarin skin and flesh, citron melon, dragonfruit skin and apple cucumber.

Table 3 Phytochemical composition of the horticultural produce samples (all values given as mean ± SD from 2 replicate extracts for each sample)
Full size table

The highest amounts of total phenolics were found in the blood orange skin (1988 mg GAE/100 g), followed by mandarin skin, spaghetti squash and blood orange flesh. The flesh of most cucurbit cultivars contained relatively low total phenolic contents (247–519 mg GAE/100 g), although the Indian cream cobra rockmelon sample contained significantly more total phenolics (831 mg GAE/100 g). Most of the cucurbit flesh samples also showed lower antioxidant capacity. Anthocyanins were detected in several of the samples, with the highest concentrations in blood orange and dragonfruit skin.

To investigate the relationship between the phytochemical constituents, correlation analysis was performed [30], with the results presented in Fig. 2. The ABTS, CUPRAC, FRAP, TP and TMA were all positively correlated with one another (P < 0.05), with the strongest correlation observed between the CUPRAC and ABTS values (r17 = 0.74, P < 0.001), similar to results observed in previous studies [31, 32]. Ascorbic acid content was positively correlated with the CUPRAC (r17 = 0.59, P < 0.01) and TP content (r17 = 0.76, P < 0.001).

Fig. 2
figure 2

Correlogram showing the correlations between the phytochemical constituents in the horticultural produce samples (n = 19). The numbers inside each square show the Pearson R correlation values

Full size image

Phenolic profiling by LC–MS/MS

To further explore the phytochemical composition of the produce samples, profiling of individual phenolic acids, flavonoids and other selected polyphenols was conducted via LC–MS/MS. The results are presented in Tables 4, 5, 6, 7, while an example chromatogram is shown in Fig. 3.

Table 4 Hydroxybenzoic acid and ellagic acid and resveratrol contents in the horticultural produce samples
Full size table
Table 5 Hydroxycinnamic acid and ellagic acid contents in the horticultural produce samples
Full size table
Table 6 Flavonoid contents in the horticultural produce samples
Full size table
Table 7 Anthocyanin contents in the horticultural produce samples
Full size table
Fig. 3
figure 3

A typical total ion chromatogram (TIC) of the LC–MS phenolic profiles of one of the horticulture samples (Brazilian cherry). Peak numbers correspond to those provided in Table 1

Full size image

A total of 26 compounds were identified across the samples analysed using LC–MS/MS, comprising 6 hydroxybenzoic acids (Table 4), 6 hydroxycinnamic acids (Table 5), 8 flavonoids (Table 6), 4 anthocyanins (Table 7), as well as ellagic acid and resveratrol.

The most commonly found phenolic acid was salicylic acid, which was present in 13 out of 19 samples. In contrast, a number of compounds including 4-hydroxybenzoic acid, syringic acid, catechin, luteolin, myricetin and ellagic acid were only found in one sample each. No previous literature was found for the phenolic profiles of apple cucumber and several of the Cucumis melo and Cucurbita cultivars. Further discussion on the individual phenolic profiles of the horticultural produce samples is provided in the following sections, grouped by the produce type.

Citrus

Citrus and particularly citrus peel are well-known sources of phenolic acids and flavonoids [33]. Indeed, the second-highest number of compounds (11) were identified in the blood orange flesh and blood orange skin, after Brazilian cherry (12). Previous investigations on the juice from this cultivar have reported the presence of gallic acid, protocatechuic acid, p-hydroxybenzoic acid, sinapic acid, p-coumaric acid, ferulic acid, caffeic acid, chlorogenic acid, cinnamic acid, catechin and rutin [34,35,36,37]. A number of these compounds (sinapic acid, p-coumaric acid, caffeic acid, chlorogenic acid and rutin) were detected in either the flesh or skin in this study, as well as quercetin and quercetin 3-glucoside. Delphinidin was also found in both the blood orange flesh and the skin (Table 7).

Similarly, previous studies on the phenolic acid and flavonoid profiles of mandarin peel, pulp and seeds have reported the presence of polyphenols including gallic acid, protocatechuic acid, p-hydroxybenzoic acid, vanillic acid, syringic acid, chlorogenic acid, p-coumaric acid, caffeic acid, sinapic acid, ferulic acid, catechin and rutin [38,39,40,41]. This study found that gallic acid, protocatechuic acid, p-hydroxybenzoic acid, chlorogenic acid, p-coumaric acid, caffeic acid, ferulic acid and rutin. Additionally, quercetin, quercetin-3-glucoside and neochlorogenic acid were also found. Neochlorogenic acid does not appear to have been previously reported from the mandarin species investigated in this study (Citrus reticulata), although it has been found in other Citrus species [42].

Cucurbits

The cucurbit samples generally contained a lower number of identified compounds (average of 4 compounds/sample) compared to the other fruits, concurring with the lower total phenolic contents found in these samples.

Although there have been several studies on the general phytochemical composition of citron melon and the identification of cucurbitacins from this cultivar [43], there are limited studies on its phenolic acid content. Compounds identified in the present study were gallic acid, 4-hydroxybenzoic acid, p-coumaric acid, protocatechuic acid, chlorogenic acid, caffeic acid, salicylic acid and rutin.

Only one compound (salicylic acid) was identified in the red watermelon flesh; quercetin-3-glucoside was also found in the rind of this sample. This concurred with previous work, which found more a higher abundance and diversity of phenolic acids in watermelon rind compared to the flesh [44,45,46].

The three Cucumis melo samples (Christmas melon, golden honeydew and Indian cream cobra rockmelon) showed broadly differing phenolic profiles, with the Christmas melon and Indian cream cobra rockmelon samples containing four identified polyphenols each. The golden honeydew rind also contained several additional compounds: sinapic acid, myricetin and quercetin-3-glucoside. The results broadly agreed with previous work on another Cucumis melo cultivar (C. melo cv. maazoun), which demonstrated the presence of gallic acid, protocatechuic acid, p-hydroxybenzoic acid, isovanillic acid, chlorogenic acid and m-coumaric acid [47].

Although no previous work was found for the apple cucumber cultivar, a number of the compounds detected were previously identified in regular cucumber (Cucumis sativus), including quercetin compounds [48, 49].

Similarly, no literature was found on the specific Cucurbita cultivars investigated in this study. The trombone pumpkin did not contain any of the targeted polyphenols, while the Queensland blue pumpkin contained only vanillic acid and cyanidin 3-glucoside. However, the spaghetti squash contained 10 compounds, including caffeic acid, salicylic acid, sinapic acid and a number of flavonoids. This largely agreed with Kulczyński, Gramza-Michałowska [50], who reported the most abundant phenolic acids present in a “Spaghetti” cultivar of C. pepo to be protocatechuic acid, p-hydroxybenzoic acid, caffeic acid and sinapic acid.

Other crops

Brazilian cherry contained the highest number of identified compounds (12). Many of these, including gallic acid and quercetin-3-glucoside, have been previously reported from Brazilian cherry [51,52,53]. Although not previously reported from this species, the presence of rutin is quite feasible due to the high concentration of quercetin and other quercetin glycosides observed here and in previous literature [53]. Cyanidin-3-glucoside, which has been reported as the primary anthocyanin from this species [51, 52], was also detected (Table 7).

The majority of phenolic acids and flavonoids detected in the dragonfruit flesh and skin samples were previously reported, including caffeic acid, sinapic acid, rutin and quercetin-3-glucoside [54,55,56,57]. The skin of this species was notable for containing a quercetin-3-glucoside content over three times higher than any other species investigated.

Conclusion

This study investigated the phytochemical composition and phenolic acid contents in a number of horticultural crops from central Queensland, including reporting the phenolic profiles of several species/cultivars for the first time. The highest antioxidant activities were found in Brazilian cherry, blood orange skin, mandarin skin and spaghetti squash. In addition, the levels of over 30 polyphenolic compounds were quantified in the samples using LC–MS/MS. These results may support the prospect of marketing several of the crops/cultivars as functional food crops in domestic or export markets.