Characterisation of ‘Ruby Roman’ Table Grapes (Vitis Labruscana Bailey) by Sensory Evaluation and Analysis of Aroma and Taste Compounds

Abstract

‘Ruby Roman’ is a high-end table grape cultivar which has very large and bright red berries rich in light sweet juice. We characterised ‘Ruby Roman’ grapes by sensory evaluation and analysis of volatile aroma and taste compounds. A sensory evaluation by expert panellists using qualitative data comparisons of five grape cultivars revealed that ‘Ruby Roman’ exhibited a fruity character in common with ‘Pione’. Component profiling of aroma and taste compounds revealed that ‘Ruby Roman’ and ‘Pione’ contained 15 esters, which were absent from the other grape cultivars, indicating that these esters were the causes of the fruity character of ‘Ruby Roman’ and ‘Pione’. Analysis by gas chromatography-olfactometry and a comparison between quantitative values and odour thresholds were performed, and these results led to the observation that ethyl 2-methylbutanoate and ethyl butanoate were the characteristic esters in ‘Ruby Roman’ associated with the fruity aroma characteristic of this grape cultivar.

Introduction

‘Ruby Roman’ is a high-end table grape cultivar (Vitis Labruscana Bailey) bred from open-pollinated seed progeny of ‘Fujiminori’ at the Ishikawa Agriculture Research Center (Shima et al., 2006). It has very large and bright red berries, the individual weight of which is twice that of ‘Kyoho’. ‘Ruby Roman’ is cultivated only in Ishikawa Prefecture, Japan. The strict quality standards for ‘Ruby Roman’ result in high-quality products on the market that have been traded at a luxury price since its first release in 2008. At the first auction at a wholesale market in 2017, a bunch of ‘Ruby Roman’ grapes was auctioned for a record ¥1.11 million ($9 745). The high-price trade has been attracting attention from international media (i, ii), making ‘Ruby Roman’ popular as a high-end table grape product. ‘Ruby Roman’ has a rich juice with a light sweet aroma to the berry. However, the characteristic compounds responsible for the attractive aroma in ‘Ruby Roman’ have not been identified. Identification of important aroma compounds would be useful in improving quality and developing food products derived from ‘Ruby Roman’.

The aroma and taste compounds in grapes have been researched to deepen our understanding of cultivar differences and to facilitate breeding to improve the quality of table grapes. Esters and terpenes contribute to fruit and floral aroma characters (Fenoll et al., 2009; Capone et al., 2013). The green leafy aroma is derived from C6-aldehydes and -alcohols (Kalua and Boss, 2009). Major sugars and organic acids in grapes are known to be fructose, glucose, malic acid and tartaric acid (Shiraishi, 2000). The influence of environment and cultivation practices on aroma and taste components has also been studied (Lund and Bohlmann, 2006).

The first study of aroma and taste compounds in ‘Ruby Roman’ reported analysis of sugars, organic acids and volatile compounds (Arite et al., 2015). With regard to sugars and organic acids, ‘Ruby Roman’ contains mainly fructose, glucose, malic acid and tartaric acid, the concentrations of which are similar to the levels found in most table grapes. As for volatile components, several components detected in ‘Ruby Roman’ by gas chromatography-mass spectrometry (GC-MS) were reported, although varietal differences with respect to the relative concentrations were not identified. Scientific sensory evaluation of the volatiles in ‘Ruby Roman’ has not been published, and a combination of sensory evaluation data and quantitative volatile chemical analysis is necessary to fully characterise ‘Ruby Roman’.

A metabolomics approach using GC-MS is useful to provide a comprehensive analysis of volatile components to characterise the taste and aroma qualities of food samples. This profiling approach permits the identification of differences between samples by multivariate analysis, leading to the detection of the important compounds from among the many volatile compounds present in each sample. Japanese sake, apple, tomato juice, roasted tea, etc., have been studied using this approach to characterise the compounds associated with high taste and aroma quality (Miura et al., 2014; Tanaka et al., 2015; Iijima et al., 2016; Sasaki et al., 2018). The analytical technology of aroma components has advanced in recent years, with the gas chromatography-olfactometry (GC-O) method, in which the human nose evaluates the odour of compounds separated by gas chromatography (GC). This approach has been used in aroma research to detect and identify odour-active compounds in foods (Leland et al., 2001; van Ruth et al., 2001) and to evaluate odour-active compounds in wine, beer, green tea, roasted tea and soy sauce, among others (Kumazawa et al., 2002; Frank et al., 2011; Kaneko et al., 2012; Langos et al., 2013; Sasaki et al., 2017). Against a background of advanced GC technology, a combination of a profiling approach and the GC-O method is likely to be highly effective at characterising grape samples.

The main objective of this study is to characterise ‘Ruby Roman’ grapes by means of both sensory evaluation and instrumental analysis of aroma and taste compounds. The sensory impressions of samples of five table grape cultivars, including ‘Ruby Roman’, were evaluated by experienced panellists using the quantitative descriptive analysis (QDA) method. Volatile compounds were analysed by GC-MS and GC-O with solid-phase microextraction (SPME), and taste compounds were analysed using a high-performance liquid chromatography (HPLC) system.

Materials and Methods

Materials

Grapes: Five grape cultivars (Vitis Labruscana Bailey), namely, ‘Ruby Roman’, ‘Pione’, ‘Kyoho’, ‘Shine Muscat’ and ‘Delaware’, were grown on the same farm at Sand Dune Agricultural Research Center, Ishikawa Agriculture and Forestry Research Center, Ishikawa, Japan. The test cultivars included the top three grape cultivars area in Japan in 2012, namely, ‘Kyoho’ (32%), ‘Delaware’ (17%) and ‘Pione’ (14%) (Yamada and Sato, 2016). ‘Shine Muscat’ was selected because it has rapidly increased in popularity in recent years because of its ease of cultivation and its excellent eating quality (Yamada and Sato, 2016). Bunches of grapes were picked in August 2017 and stored in refrigerator, then used for this study within several days of harvest.

Chemicals: For the analysis of volatile compounds, we purchased cyclohexanol from Wako Pure Chemical Industries Ltd. (Tokyo, Japan) and N-alkanes (n = 6–25) from Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan). For the identification of compounds detected by GC-MS, we purchased volatile compounds 1, 2, 4, 6, 12, 14, 22, 24, 30 and 42 (Table 1) from Wako Pure Chemical Industries Ltd., and volatile compounds 3, 5, 7, 9, 11, 13, 18–21, 23, 25, 26, 28, 29, 31, 32, 35 and 37–39 (Table 1) were purchased from Tokyo Chemical Industry Co., Ltd. For the quantitative analysis of sugars and organic acids, we purchased fructose, glucose, malic acid, citric acid and tartaric acid from Wako Pure Chemical Industries Ltd.

Table 1. Characteristics of the 42 identified volatile compounds detected from grapes of five cultivars.

No.	RTa	RIb	Compound	GC–Od	Type	IDe
1	9.6	887	Ethyl acetate	−	Ester	MS(a, d), RI(a)
2	11.7	924	Ethanol	−		MS(a, d), RI(a)
3	12.5	959	Ethyl propanoate	−	Ester	MS(a, d), RI(a)
4	16.1	1 039	Ethyl butanoate	Fruity	Ester	MS(a, d), RI(a)
5	16.8	1 056	Ethyl 2-methylbutanoate	Fruity	Ester	MS(a, d), RI(a)
6	18.0	1 082	Hexanal	green	C6	MS(a, d), RI(a)
7	20.3	1 138	Ethyl pentanoate	Fruity	Ester	MS(a, d), RI(a)
8	20.3	1 138	Unknown			−
9	21.3	1 166	Ethyl (E)-but-2-enoate		Ester	MS(a, d), RI(a)
10	22.8	1 204	Unknown			−
11	23.4	1 222	(E)-hex-2-enal	Green	C6	MS(a, d), RI(a)
12	24.0	1 237	Ethyl hexanoate	Fruity	Ester	MS(a, d), RI(a)
13	24.1	1 241	Ethyl (E)-2-methylbut-2-enoate		Ester	MS(a, d), RI(a)
14	24.5	1 255	Pentan-1-ol	Fusel		MS(a, d), RI(a)
15	25.7	1 291	3-Hydroxybutan-2-one			MS(a, d), RI(a)
16	26.2	1 306	Ethyl (E)-hex-3-enoate		Ester	MS(a, d), RI(a)
17	26.8	1 325	Unknown			−
18	27.0	1 331	(E)-hept-2-enal	Fatty, green		MS(a, d), RI(a)
19	27.2	1 338	Ethyl heptanoate		Ester	MS(a, d), RI(a)
20	27.7	1 357	Hexan-1-ol		C6	MS(a, d), RI(a)
21	28.7	1 389	(Z)-hex-3-en-1-ol		C6	MS(a, d), RI(a)
22	29.1	1 399	Nonanal	Floral		MS(a, d), RI(a)
23	29.3	1 410	(E)-hex-2-en-1-ol		C6	MS(a, d), RI(a)
24	30.2	1 439	Ethyl octanoate		Ester	MS(a, d), RI(a)
25	30.5	1 452	Oct-1-en-3-ol			MS(a, d), RI(a)
26	30.5	1 454	Linalyl oxidec		Terpene	MS(a, d), RI(a)
27	31.5	1 491	Unknown			−
28	31.6	1 493	2-Ethylhexan-1-ol			MS(a, d), RI(a)
29	32.4	1 528	Ethyl 3-hydroxybutanoate		Ester	MS(a, d), RI(a)
30	32.9	1 543	Unknown			−
31	33.0	1 549	Linalool	Floral	Terpene	MS(a, d), RI(a)
32	33.4	1 562	Ethyl (E)-oct-2-enoate		Ester	MS(a, d), RI(a)
33	33.5	1 565	Unknown			−
34	33.8	1 577	Unknown			−
35	34.5	1 609	4-Methoxy-2,5-dimethylfuran-3-one			MS(a, d), RI(a)
36	34.6	1 615	Hotrienol		Terpene	MS(d), RI(l)
37	35.4	1 645	Ethyl decanoate		Ester	MS(a, d), RI(a)
38	39.0	1 802	Ethyl 2-phenylacetate		Ester	MS(a, d), RI(a)
39	39.9	1 852	Hexanoic acid		C6	MS(a, d), RI(a)
40	40.7	1 877	Unknown			−
41	41.2	1 909	Unknown			−
42	41.7	1 933	2-Phenylethanol			MS(a, d), RI(a)

^a  RT: retention time.

^b  RI: retention index on a DB-Wax column (60 m × 0.25 mm i.d., coated with a film thickness of 0.25 µm) observed for gas chromatography–mass spectrometry (GC.MS).

^c  Optical isomer is unclear.

^d  GC–O: gas chromatography-olfactometry

^e  Method of identification: MS(a), identification by comparison with the mass spectra of authentic reference compounds; MS(d), identification by comparison with the mass spectra stored in the National Institute of Standards and Technology (NIST) and Wiley databases; RI(a), identification by comparison with the RIs of authentic reference compounds; RI(l), identification by comparison with published RIs in the literature (Ito et al., 2002).

Sensory evaluation    A panel, trained for QDA, was recruited from the employees of Ogawa & Co., Ltd (Tokyo, Japan). The panel consisted of 12 men and 8 women (n = 20), with an average age of 34.2 years. All panel members had participated previously in several QDA studies on grapes, so they were experienced in QDA profiling. The sensory qualities of the berries of the five grape cultivars were assessed by QDA on the same day. Samples were half-berries of ‘Ruby Roman’, ‘Pione’, ‘Kyoho’ and ‘Shine Muscat’ and whole (uncut) berries of ‘Delaware’ the berry size of which is much smaller than other berries. Samples were presented to each panellist in randomised order, but each sample name was not disclosed. The panellists evaluated the pulp of the grape berries after removing the skin from their mouth. Before the next sample, each panellist rinsed their mouth with water. Sensory evaluation was performed in an environmentally controlled space. The room temperature was 24 °C, and humidity was approximately 40%. Aroma and taste intensities were evaluated using a seven-point categorical scale (1: weak to 7: strong). The panellists used the following terms to describe the taste and aroma sensory attributes of grapes: fruity, floral, green, foxy, sour or sweet.

Analyses of sugars and organic acids    The grape berries were frozen in a -20 °C freezer and peeled by hand. The frozen berries from each cultivar were mixed before being blended in a domestic food processor. The blended samples were rapidly heated in a microwave oven at 600 W for 1 min to deactivate enzymes in grapes. Samples were frozen again and stored in the freezer until analysis.

Fructose and glucose were analysed using an HPLC system (Agilent 1260; Agilent Technologies Inc., Palo Alto, CA, USA) coupled to a mass spectrometry system (Agilent 6530; Agilent Technologies Inc.). Samples (volume of 1 µL) were separated using an amino column (150 × 3.0 mm, 3.0 µm particle size; Unison UK-Amino; Imtakt Corp., Kyoto, Japan) maintained at 55 °C. The elution rate was 0.5 mL/min with a linear gradient of solvent A (water containing 5 mM ammonium formate) and solvent B (acetonitrile) for 8 min, with the following linear gradient elution: 0 min, 90% B; 1 min, 90% B; 5 min, 80% B; 18 min, 60% B; 20 min, 60% B; and 20.01 min, 90% B. The total run time was 20 min. The external standard method was used, with the calibration curve being made from standard solutions of fructose or glucose at concentrations ranging from 0.125 to 1.0 mg/100 mL. The electrospray ionisation probe was operated in the negative mode at three scans per s across an m/z range of 100–1 050: drying gas, 280 °C, 11 L/min; nebuliser, 52 psig; VCap, 3 500 V; fragmentor, 150 V; and nitrogen gas flow, 1.5 L/min.

Malic acid, citric acid and tartaric acid were analysed using an ion chromatography system (ICS-2100, Thermo Fisher Scientific, Waltham, MA, USA). Samples (initial injection volume of 1 µL) were separated using an IonPac AS20 column (250 × 4.0 mm, 7.5-µm particle size; Thermo Fisher Scientific) with a guard column (IonPac AG20; 50 × 4.0 mm) maintained at 30 °C, and eluted at 1.0 mL/min with the following linear gradient elution: 0 min, 0.5 mM KOH and 60 min 36.5 mM KOH. The total run time was 30 min. Conductivity signals were measured using a DS6 heated conductivity detector (Thermo Fisher Scientific). The external standard method was used, with the calibration curve being constructed from standard solutions of malic acid, citric acid and tartaric acid, at concentrations ranging from 0.125–1.0 mg/100 mL.

Analysis of volatile compounds by GC-MS    The grape berries were frozen in a freezer and peeled by hand. The frozen berries were mixed and blended in 20 wt% NaCl in a domestic food processor. NaCl was used to deactivate enzymes that could damage flavour compounds, and the addition of NaCl also improved the sensitivity of GCMS analysis. The blended samples were stored in the freezer prior to analysis.

Volatile compounds in the samples were analysed using a GC system (Agilent 7890A; Agilent Technologies Inc.) equipped with a mass spectrometer (Agilent 5975C; Agilent Technologies Inc.) and an autosampler (MPS2; Gerstel GmbH & Co. KG, Mülheim an der Ruhr, Germany). After the frozen sample was thawed at room temperature, 100 µL of 100 ppm (v/v) cyclohexanol solution was added to 5 mL of the sample in a 20-mL glass vial as an internal standard. The glass vial was heated at 50 °C in an isothermal agitator rotating with a period of 250 rpm. After equilibrating for 10 min, solid-phase microextraction (SPME) fibres (50/30 µm divinylbenzene/carboxen/polydimethylsiloxane; Supelco, Inc., Bellefonte, PA, USA) were injected into the glass vial (injection depth from septum: 22 mm), and volatile compounds in the vial headspace were allowed to extract into the fibres for 30 min. The extracted volatile compounds were desorbed from the fibres at 230 °C in the injector of the GC system and injected into a 60 m × 0.25-mm internal diameter (i.d.) DB-Wax fused-silica capillary column (J&W Scientific Inc., Folsom, CA, USA) with a film thickness of 0.25 µm. The oven temperature was programmed to increase from 40 °C (10-min hold) to 230 °C (12-min hold) at a rate of 5 °C/min. The split ratio was 1:5. Helium was used as the carrier gas at a linear flow rate of 1.12 mL/min. The mass spectra were obtained by electron-impact ionisation under an ionisation voltage of 70 eV and an ion source temperature of 150 °C. Analysis was carried out in the SCAN mode. The retention index (RI) was calculated following the method of Kováts (1958) using N-alkanes (n = 6–25). Each sample was analysed in triplicate.

For screening of odour-active compounds from the volatile compounds detected by GC-MS, samples were analysed by GC-O. The volatile compounds extracted by the SPME method were injected into the GC system as described above, equipped with a sniffing port (ODP2; Gerstel Inc.). The flow from the capillary column was divided between the mass spectrometry system and the sniffing port at a ratio of 1:1. Nitrogen gas was pumped into the sniffing port at 50 mL/min. The compounds released from the sniffing port were evaluated by two panellists, who were trained using the T&T olfactometer (Daiichi Yakuhin Sangyo Ltd., Tokyo, Japan) and Open Essence (Wako Pure Chemical Industries Ltd.), using their nose to check the sensitivity and identification ability of the human olfactory sense to diagnose olfactory differences.

The quantification analysis of ethyl butanoate, ethyl 2-methylbutanoate and ethyl pentanoate in ‘Ruby Roman’ was performed by the standard addition method using GC-MS. 100 µL of authentic standard solution, containing 0–100 mg/L ethyl butanoate, 0–10 mg/L ethyl 2-methylbutanoate, or 0–10 mg/L ethyl pentanoate with 100 ppm (v/v) cyclohexanol as an internal standard, was spiked into 5 mL of the ‘Ruby Roman’ sample in a 20-mL glass vial. Volatile compounds in the glass vial were analysed using a GC system in the same way as described above. Calibration curve was plotted by using relative peak intensity, GC-MS peak intensity standardised by the use of internal standard. The concentration, the X-intercept of a calibration curve, was obtained by dividing the Y-intercept of the curve by the slope.

Compound identification    All volatile compounds were identified by comparison with the reference mass spectra stored in the Wiley and National Institute of Standards and Technology (NIST) databases (Agilent Technologies Inc.) and the mass spectra and/or the RIs of authentic compounds reported in the literature (Ito et al., 2002).

Statistical analysis    One-way analysis of variance (ANOVA) was performed on the sensory evaluation data, with the Tukey test used for post-hoc multiple pairwise comparison tests to compare samples for significant differences, using the SPSS Statistics v.21.0 software (IBM, Armonk, NY, USA). Multivariate analysis was performed on the data regarding volatile compounds, sugars or organic acids by principal component analysis (PCA) using the SIMCA software ver.14.0 (Umetrics, Umeå, Sweden) after mean centring and unit variance scaling of the data.

Results and Discussion

Sensory evaluation    Figure 1 shows the aroma and taste intensity of the sensory evaluation data from the grape samples. In terms of aroma intensity, ‘Ruby Roman’, ‘Pione’ and ‘Kyoho’ generated high scores of 4.1–4.3. ‘Ruby Roman’ is a third-generation offspring of ‘Pione’ (Shima et al., 2006), and ‘Pione’ is an offspring from a cross between ‘Kyoho’ and ‘tetraploid Muscat of Alexandria’ (Yamane et al. 1978). These three grape cultivars were therefore related, explaining why their aroma intensity scores were at a similar level. The aroma intensity score of ‘Shine Muscat’ was significantly lower than those of the three related grape cultivars. In terms of taste intensity, only ‘Delaware’ exhibited a score significantly different from those of the other cultivars tested, showing a lower score than the other grapes. Therefore, ‘Delaware’ was characterised by possessing the lowest intensity of both aroma and taste.

Fig. 1.

The results of aroma and taste intensities from the sensory evaluation.

The sensory evaluation was carried out by 20 panellists, with a high score meaning high intensities of aroma and taste. All values are means ± standard errors. Any two samples with a common lowercase letter are not significantly different (P > 0.05), using Tukey's test.

Figure 2 shows the results of the sensory evaluation, in which the sensory attributes were described as fruity, floral, green, foxy, sour or sweet. Detailed data and multiple comparison results are shown in the Supplemental data (Table S1). ‘Shine Muscat’ was markedly different in terms of sensory characteristics from the other grape cultivars, showing high scores of floral and green characters and low scores of foxy and sour characters.

Fig. 2.

The flavour profile of attribute scores by sensory evaluation.

The sensory evaluation was carried out by the 20 panellists, with a high score meaning high intensities of described sensory attributes: ‘fruity’, ‘floral’, ‘green’, ‘foxy’, ‘sour’ and ‘sweet’. Significant differences are indicated by *P < 0.05, **P < 0.01 and ***P < 0.001 (one-way ANOVA).

Table S1. The flavour profile data from grape samples from five different cultivars by the sensory evaluation.

	Fruity	Floral	Green	Foxy	Sour	Sweet
Ruby Roman	4.85 ± 0.18a	4.05 ± 0.19 a	4.10 ± 0.21 a	3.35 ± 0.41 a	3.70 ± 0.23 a	4.45 ± 0.22 a
Pione	4.75 ± 0.18a	3.95 ± 0.19 a	3.70 ± 0.25 ab	3.90 ± 0.26 a	3.40 ± 0.31 ab	4.35 ± 0.20 a
Kyoho	3.85 ± 0.19b	3.65 ± 0.22 ab	3.15 ± 0.40 b	4.15 ± 0.32 a	3.65 ± 0.34 ab	4.75 ± 0.25 a
Shine Muscat	4.30 ± 0.21ab	5.65 ± 0.15 c	5.15 ± 0.25 c	2.25 ± 0.43 b	2.70 ± 0.36 b	4.60 ± 0.20 a
Delaware	4.15 ± 0.21ab	3.10 ± 0.33 b	3.40 ± 0.29 ab	3.20 ± 0.39 ab	3.65 ± 0.35 ab	4.55 ± 0.19 a

All values are means ± standard errors. Different letters indicate significant difference at P < 0.05 (Tukey's test)

The National Agriculture and Food Research Organization (Ibaraki, Japan) has focused on a long-term strategy to breed new cultivars of grapes with a Muscat flavour. ‘Shine Muscat’, from a cross between ‘Akitsu-21’ and ‘Hakunan’, is a product of this national breeding programme (Yamada et al., 2008), with ‘Shine Muscat’ having a widely accepted Muscat flavour. ‘Delaware’ showed consistently low scores of all sensory attributes. Its aroma and taste intensities were also low (Fig. 1). ‘Delaware’ has been a preferred table grape cultivar in Japan for many years, although its aroma and taste characters are weak in comparison with modern-day cultivars. The sensory attributes of ‘Ruby Roman’, ‘Pione’ and ‘Kyoho’ showed similar patterns because they are related cultivars. The sensory traits of ‘Ruby Roman’ were particularly similar to those of ‘Pione’, with ‘Pione’ being the cultivar most closely related to ‘Ruby Roman’. ‘Ruby Roman’ and ‘Pione’ showed high ‘fruity’ scores in comparison with the other grapes. The ‘fruity’ score of ‘Kyoho’ was the lowest of the five grape cultivars. The three related cultivars, namely, ‘Ruby Roman’, ‘Pione’ and ‘Kyoho’, all had high ‘foxy’ scores, although the ‘foxy’ score of ‘Ruby Roman’ was intermediate between the other three cultivars. ‘Ruby Roman’ had a score for the ‘green’ sensory character higher than those of the other related cultivars. Considering all sensory data, the important sensory character of ‘Ruby Roman’ is the fruity character which was at the same level as ‘Pione’.

Sugars and organic acids in grapes    Table 2 shows the concentrations of sugars and organic acids in berry samples from the different grape cultivars. Sugar and organic acid concentrations were determined initially in relation to their taste thresholds. With regard to sugars, the mean ± standard error (SE) concentrations of fructose and glucose in ‘Ruby Roman’ were 6.8 ± 0.21 and 7.1 ± 0.42 g/100 g grape fresh weight, respectively, similar to those of ‘Pione’ and ‘Kyoho’. ‘Shine Muscat’ and ‘Delaware’ contained 8.8 ± 0.57 and 8.1 ± 0.03 g/100 g of fructose and 8.4 ± 0.80 and 7.0 ± 0.80 g/100 g of glucose, respectively, concentrations that were higher than those in ‘Ruby Roman’. This result was similar to that from previous research, which reported that the sugar concentrations of ‘Ruby Roman’ were at a moderate level for grapes (Arite et al., 2015). The main sugars in grapes are fructose and glucose (Shiraishi, 1993; Shiraishi, 2000), and the concentration of each sugar in grapes of all five cultivars was considerably higher than the corresponding taste threshold values, which are 0.90 and 0.42 g/100 g for fructose and glucose, respectively (Kawai, 2012), indicating that the sugars could influence the taste of the grapes. Additionally, the concentrations of fructose and glucose might influence the sensory scores of the five grape cultivars. However, no significant difference in ‘sweet’ scores was detected between the five cultivars in the sensory evaluation (Fig. 2 and Table S1). ‘Delaware’, which contained sugar concentrations higher than ‘Ruby Roman’, ‘Pione’ and ‘Kyoho’, showed the lowest score for taste sensory intensity (Fig. 1). Considering these results, the differences in fructose and glucose concentrations did not contribute to the differences in sensory evaluation between grapes of the different cultivars. The low taste intensity of ‘Delaware’, despite its high sugar concentrations, might be influenced by differences in concentrations of volatile compounds from the other cultivars.

Table 2. Concentrations of sugars and organic acids in grapes of five cultivars.

Cultivar	Sugars (g/100 g fresh weight)		Organic acids (mg/100 g grape fresh weight)
	Fructose	Glucose	Malic Acid	Citric Acid	Tartaric Acid
Ruby Roman	6.8 ± 0.21	7.1 ± 0.42	200 ± 10	23 ± 0.2	550 ± 4
Pione	6.8 ± 0.08	6.9 ± 0.33	130 ± 1	21 ± 0.5	620 ± 1
Kyoho	6.6 ± 0.10	6.9 ± 0.31	150 ± 11	17 ± 0.5	540 ± 6
Shine Muscat	8.8 ± 0.57	8.4 ± 0.80	100 ± 2	14 ± 1.1	370 ± 6
Delaware	8.1 ± 0.03	7.0 ± 0.80	59 ± 4	16 ± 0.8	750 ± 3

The mean values ± standard errors obtained from three independent samples are shown.

With regard to the organic acids, the concentrations of malic acid, citric acid and tartaric acid in ‘Ruby Roman’ were 200 ± 10, 23 ± 0.2 and 550 ± 4 mg/100 g, respectively. ‘Ruby Roman’ contained higher concentrations of malic acid and citric acid than the other cultivars, whereas tartaric acid concentrations were similar across all samples. The malic acid concentrations in the three related cultivars, namely, ‘Ruby Roman’, ‘Pione’ and ‘Kyoho’, were similar but higher than those in the other two cultivars, with this result being similar to that reported from a previous study (Arite et al., 2015). The taste threshold concentrations of malic acid, citric acid and tartaric acid are 70, 2 and 770 mg/100 g, respectively (Maeda and Nakao, 1963), which means that the concentrations of malic acid and citric acid in ‘Ruby Roman’ grapes were above each of the respective taste thresholds, although the concentration of tartaric acid in ‘Ruby Roman’ was below the taste threshold. These results indicate that malic acid and citric acid could influence the taste of ‘Ruby Roman’.

Volatile compounds in grapes    We analysed the volatile compounds in grape samples from the five cultivars using GC-MS and GC-O, with the SPME method of extraction. The results are summarised in Table 2. GC-MS detected a total of 42 volatiles across the five grape cultivars, 33 of which could be identified. These included 15 esters, three terpenes, six C6-alcohols and -aldehydes (C6 compounds) and 16 other compounds. Odour-active compounds were screened by GC-O. Ten compounds showed odour activity, and these odour qualities are also described in Table 2.

Compound profiling with PCA    The combined data for 42 volatile compounds and eight taste compounds were analysed by PCA, the results of which are shown in Fig. 3. In the score plot (Fig. 3A), the first two principal components, PC1 and PC2, explained 46.1% and 19.4% of the variance, respectively. PC3 and PC4 explained 15.3% and 10.1% of the variance, respectively, but these data are not shown because PC1 and PC2 provided enough information to analyse the character of the different grape samples. Focussing on the PC1 axis, the highest contributing component, ‘Ruby Roman’ and ‘Pione’ were located in the positive PC1 area (Fig. 3A), with similar PC1 scores, indicating that the volatile and taste components in ‘Ruby Roman’ and ‘Pione’ were similar. ‘Kyoho’ was located near the centre of the plot, whereas ‘Shine Muscat’ and ‘Delaware’ were located in the negative PC1 area (Fig. 3A). ‘Kyoho’ is related to ‘Ruby Roman’ and ‘Pione’, which are the two most closely related cultivars. This relationship was supported by the PCA where ‘Kyoho’ was located near ‘Ruby Roman’ and ‘Pione’, with ‘Shine Muscat’ and ‘Delaware’ being distant from ‘Ruby Roman’ and ‘Pione’ in the score plot. Figure 3B shows the loading plot of the PCA, where compounds are classified and represented by individual symbols, and the names of the primary compounds are indicated alongside the symbols. Underlined names indicate odour-active compounds detected by GC-O. In the loading plot, 33 volatile compounds, approximately 80% of all volatile compounds resolved, were located in the positive PC1 area, indicating that ‘Ruby Roman’ and ‘Pione’ contained a large number of aroma compounds. All 15 esters, detected by GC-MS from the various grape samples, were located in the positive PC1 area, with some having highly positive scores (Fig. 3B). This result agreed with the sensory evaluation result that showed high ‘fruity’ scores for ‘Ruby Roman’ and ‘Pione’ as esters are known to confer a fruity odour (Capone et al., 2013). The difference between ‘Ruby Roman’ and ‘Pione’ was reflected in the differences in the PC2 scores of the score plot (Fig. 3A), where ‘Ruby Roman’ and ‘Pione’ were located in the negative and positive PC2 areas, respectively. In the loading plot, seven esters were located in the negative PC2 area (Fig. 3B). These esters were present in higher concentrations in ‘Ruby Roman’ than in ‘Pione’. Of these esters, ethyl 2-methylbutanoate, ethyl butanoate and ethyl pentanoate were detected by GC-O and showed a ‘fruity’ odour. Combining these results, these two esters were deemed to be compounds characteristic of ‘Ruby Roman’. The remaining eight esters were located in the positive PC2 area of the loading plot (Fig. 3B), and they were present at higher concentrations in ‘Pione’ than in grapes of the other cultivars. Of these esters, ethyl hexanoate was detected by GC-O and showed a fruity odour, suggesting that ethyl hexanoate was a characteristic compound in ‘Pione’. A more detailed discussion of the peak intensity comparisons will be described in the next section. Additionally, malic acid and citric acid were located in the positive PC1 and negative PC2 areas, respectively, indicating that ‘Ruby Roman’ contained higher concentrations of these organic acids.

Fig. 3.

Principal component analysis of the volatile compounds found in samples: (A) score plot and (B) loading plot. The mean values obtained from three independent replicates per sample are shown in A. Compounds that are classified as esters, terpenes, C6 compounds, other volatiles, organic acids or sugars are represented by black circles, open squares, grey diamonds, grey circles, grey triangles or open triangles, respectively. The cultivars of grape samples (A) and the names of the primary volatile and taste compounds (B) are indicated alongside the symbols. Underlined names (B) denote odour-active compounds detected by GC–O.

‘Shine Muscat’ was located in the negative PC1 and positive PC2 areas of the score plot (Fig. 3A). In the loading plot, terpenes, C6 compounds and sugars were located in the negative PC1 and positive PC2 areas, and those compounds were present at higher concentrations in ‘Shine Muscat’ (Fig. 3B). All of the terpenes detected were also located in this area of the loading plot. Of these terpenes, linalool was detected by GC-O, presenting a floral odour. This result agreed with the results of the sensory evaluation that showed a high floral aroma of ‘Shine Muscat’. In the sensory evaluation, ‘Shine Muscat’ also obtained a high score for the ‘green’ odour character. The ‘green’ odour character of grapes is known to be derived from C6 compounds (Kalua and Boss, 2009). (E)-Hex-2-en-1-ol and (Z)-hex-3-en-1-ol were located in the negative PC1 and positive PC2 areas of the loading plot (Fig. 3B), indicating that these compounds were present at high concentrations in ‘Shine Muscat’, contributing to the ‘green’ character of ‘Shine Muscat’. On the other hand, (E)-hex-2-enal and hexanal, located in the positive PC1 and positive PC2 areas of the loading plot, were present at higher concentrations in other grape samples with ‘Pione’. (E)-hex-2-enal and hexanal were detected by GC-O, although (E)-hex-2-en-1-ol and (Z)-hex-3-en-1-ol were not detected. Additionally, it is reported that hexanal shown the highest odour activity value among C6 compounds in ‘Shine Muscat’ (Wu et al., 2016). The ‘green’ odour character of ‘Shine Muscat’ in the sensory evaluation could not be explained by only the amount of C6 compounds. Wu et al. (2016) also reported that there were other high OVA compounds of green aroma in ‘Shine Muscat’. The compounds such as 3-methylbut-2-enal, cedrol or geranic acid, may influence the green character of ‘Shine Muscat’. Additionally, glucose and fructose were also located in the negative PC1 and positive PC2 areas, indicating that ‘Shine Muscat’ contained high concentrations of these sugars. However, the influence of these sugars on the taste of ‘Shine Muscat’ was not confirmed.

‘Delaware’ was located in the negative PC1 and negative PC2 areas of the score plot (Fig. 3A). In the loading plot, there were only three compounds in the negative PC1 and negative PC2 areas (Fig. 3B), and only amyl alcohol had high negative scores. This result indicates that the volatile and taste compounds in ‘Delaware’ were few in number. This fact was associated with the lower sensory evaluation score for ‘Delaware’ relative to the other cultivars. In fact, ‘Delaware’ exhibited the lowest score of aroma and taste intensity of all five grape samples and also showed the lowest scores for all sensory attributes.

Sensory evaluation analysed the difference in the ‘foxy’ character associated with the different grape samples (Fig. 2). The three related cultivars, including ‘Ruby Roman’, had high ‘foxy’ scores. The ‘foxy’ character in grapes has been reported to be derived from ethyl 3-sulfanylpropanoate and methyl 2-aminobenzoate (Kolor, 1983; Wang and De Luca, 2005). Considering that these compounds were not detected from any samples, the compounds contributing to the ‘foxy’ character in grapes of the five cultivars tested are unclear.

Peak intensity comparison of characteristic volatile compounds    Characteristic volatile compounds, selected on the basis of component profiling by PCA and GC-O results, were compared in terms of peak intensity. The relative peak intensities, GC-MS peak intensities standardised by the use of internal standards, of several characteristic volatile compounds for the five cultivars are shown in Fig. 4. Component profiling revealed that ethyl 2-methylbutanoate, ethyl butanoate and ethyl pentanoate were characteristic volatile compounds of ‘Ruby Roman’. The same conclusion was also reached from the peak intensity comparisons (Fig. 4). Ethyl 2-methylbutanoate and ethyl butanoate were present at particularly high concentrations in ‘Ruby Roman’, with these compounds showing fruity and sweet odours. Several agricultural producers have reported that ‘Ruby Roman’ grapes smell sweet (Arite et al., 2015) so that the sweet smell from ‘Ruby Roman’ grapes maybe derived from the high concentrations of ethyl 2-methylbutanoate and ethyl butanoate. The quantification analysis of ethyl butanoate, ethyl 2-methylbutanoate and ethyl pentanoate in ‘Ruby Roman’ was performed by the standard addition method using GC-MS. The concentrations of ethyl butanoate, ethyl 2-methylbutanoate and ethyl pentanoate were 220, 14 and 4.5 ppb (v/v), respectively (relative standard deviations were 17%, 18% and 18%, respectively). Frath et al. (1967) reported that the odour threshold concentrations of ethyl butanoate, ethyl 2-methylbutanoate and ethyl pentanoate in water were 10, 1 and 5 ppb (v/v), respectively, which means that the ethyl butyrate and ethyl 2-methylbutanoate concentrations in ‘Ruby Roman’ exceeded their respective odour threshold values. Therefore, these esters were assumed to contribute to the aroma of ‘Ruby Roman’.

Fig. 4.

Relative peak intensity of characteristic volatile components detected with GC–O.

The characteristic volatile components in this figure were selected based on component profiling and the results of GC–O. Peak intensities were standardised relative to the internal standard, cyclohexanol. The mean values ± standard errors obtained from three independent replicates are shown.

Ethyl hexanoate, detected by GC-O and located in the PC1 positive area of the loading plot (Fig. 3B), was present at similar concentrations in ‘Pione’ and ‘Ruby Roman’ (Fig. 4), and therefore, it is the ester common to both cultivars. Some esters, such as ethyl (E)-hex-3-enoate and ethyl decanoate, that had highly positive PC1 scores in the loading plot (Fig. 3B) are considered to be the characteristic compounds of ‘Pione’. Linalool detected by GC-O was one of the terpenes that was revealed by component profiling to be a characteristic volatile compound in ‘Shine Muscat’. In Fig. 4, linalool was present in only ‘Shine Muscat’ and defined the odour characteristics of ‘Shine Muscat’.

Conclusion

In this study, we attempted to characterise grapes from cultivar ‘Ruby Roman’ by sensory evaluation and instrumental analysis of aroma and taste compounds. The sensory evaluation of the five grape cultivars by expert panellists, using the QDA method, led to the finding that ‘Ruby Roman’ exhibited a fruity character in common with the related cultivar, ‘Pione’. Component profiling of 42 volatile compounds and eight taste compounds by PCA revealed that ‘Ruby Roman’ and ‘Pione’ contained 15 esters absent from the other cultivars, indicating that among these esters were the causes of the fruity character of ‘Ruby Roman’ and the closely related ‘Pione’. Of the seven esters that were at higher concentrations in ‘Ruby Roman’, ethyl 2-methylbutanoate, ethyl butanoate and ethyl pentanoate were confirmed to exhibit odour activity, particularly a fruity odour, by GC–O, with ethyl 2-methylbutanoate and ethyl butanoate being present at higher concentrations in ‘Ruby Roman’ than in ‘Pione’. The fruity character of ‘Ruby Roman’ in sensory evaluation was therefore characterised to be caused by these two esters.

References

•  Arite,  T.,  Ishisaki,  K.,  Date,  A.,  Matsuda,  K.,  Sasaki,  T., and  Miwa,  S. (2015). Analysis of sugars, organic acids and volatile compounds in the ‘Ruby Roman’ grape. Bulletin of Ishikawa Agriculture Research Center, 31, 17-27. (in Japanese).
•  Capone,  S.,  Tufariello,  M., and  Siciliano,  P. (2013). Analytical characterisation of Negroamaro red wines by “Aroma Wheels”. Food Chem., 114, 2906-2915.
•  Fenoll,  J.,  Manso,  A.,  Hellin,  P.,  Ruiz,  L., and  Flores,  P. (2009). Changes in the aromatic composition of the Vitis vinifera grape Muscat Hamburg during ripening. Food Chem., 114, 420-428.
•  Frank,  S.,  Wollmann,  N.,  Schieberle,  P., and  Hofmann,  T. (2011). Reconstitution of the flavor signature of Dornfelder red wine on the basis of the natural concentrations of its key aroma and taste compounds. J. Agric. Food Chem., 59, 8866-8874.
•  Frath,  R.A.,  Black,  D.R.,  Guadagni,  D.G.,  McFadden,  W.H., and  Schultz,  T.H. (1967). Identification and organoleptic evaluation of compounds in Delicious apple essence. J. Agric. Food Chem., 15, 29-35.
•  Iijima,  Y.,  Iwasaki,  Y.,  Otagiri,  Y.,  Bamba,  T., and  Fukusaki,  E. (2016). Flavor characteristics of the juices from fresh market tomatoes differentiated from those from processing tomatoes by combined analysis of volatile profiles with sensory evaluation. Biosci. Biotechnol. Biochem., 80, 2401-2411.
•  Ito,  Y.,  Sugimoto,  A.,  Kakuda,  T., and  Kubota,  K. (2002). Identification of potent odorants in Chinese jasmine green tea scented with flowers of Jasminum sambac. J. Agric. Food Chem., 59, 8866-8874.
•  Kalua,  C.M. and  Boss,  P.K. (2009). Evolution of volatile compounds during the development of cabernet sauvignon grapes (Vitis vinifera L.). J. Agric. Food Chem., 57, 3818-3830.
•  Kaneko,  S.,  Kumazawa,  K., and,  Nishimura,  O. (2012). Comparison of key aroma compounds in five different types of Japanese soy sauces by aroma extract dilution analysis (AEDA). J. Agric. Food Chem., 60, 3831-3836.
•  Kawai,  M. (2012). Amami No Kagaku. Chemistry & Education, 60, 414-417 (in Japanese).
•  Kolor,  M.G. (1983). Identification of an important new flavor compound in Concord grape: ethyl-3-mercaptopropionate. J. Agri. Food Chem., 31, 1125-1127.
•  Kováts  E. (1958). Gas-chromatographische charakterisierung organischer verbindungen. teil 1: Retentionsindices aliphatischer halogenide, alkohole, aldehyde und ketone. Helv. Chim. Acta., 41, 1915-1932.
•  Kumazawa,  K. and  Masuda,  H. (2002). Identification of potent odorants in different green tea varieties using flavor dilution technique. J. Agric. Food Chem., 50, 5660-5663.
•  Langos,  D.,  Granvogl,  M., and  Schieberle,  P. (2013). Characterization of the key aroma compounds in two Bavarian wheat beers by means of the sensomics approach. J. Agric. Food Chem., 61, 11303-11311.
•  Leland,  J.V.,  Schieberle,  P.,  Buettner,  A., and  Acree,  E. (2001). “Gas Chromatography-Olfactometry.” American Chemical Society, Washington D. C.
•  Lund,  S.T. and  Bohlmann,  J. (2006). The molecular basis for wine grape quality--a volatile subject. Science, 311, 804-805.
•  Maeda,  S. and  Nakao,  T. (1963) Sour taste of various acidic substances (Part 1). J. Home Econ. Jpn, 14, 149-154 (in Japanese).
•  Miura,  N.,  Isogai,  A.,  Iwashita,  K.,  Bamba,  T., and  Fukusaki,  E. (2014). Gas chromatography/mass spectrometry-based component profiling and quality prediction for Japanese sake. J. Biosci. Bioeng., 118, 406-414.
•  Sasaki,  T.,  Tanase,  Y.,  Yonezawa,  T.,  Michihata,  T., and  Kawamura-Konishi,  Y. (2014). Metabolomics Profiling of Roasted Stem Tea Using Gas Chromatography-Mass Spectrometry and a Sensory Test. Food Sci. Technol. Res., 24, 1059-1067.
•  Sasaki,  T.,  Koshi,  E.,  Take,  H.,  Michihata,  T.,  Maruya,  M., and  Enomoto,  T. (2017). Characterisation of odorants in roasted stem tea using gas chromatography-mass spectrometry and gas chromatography-olfactometry analysis. Food Chem., 220, 177-183.
•  Shima,  M.,  Tamura,  S.,  Inabe,  Y.,  Nobatake,  S., and  Takayama,  N. (2006). New grape cultivar ‘Ruby Roman’. Bulletin of Ishikawa Agriculture Research Center, 27, 33-36. (in Japanese).
•  Shiraishi,  M. (1993). Three descriptors for sugars to evaluate grape germplasm. Euphytica, 71, 99-106.
•  Shiraishi,  M. (2000). Comparison in changes in sugars, Oorganic acids and amino acids during berry ripening of sucrose-and hexose-accumulating grape cultivars. J. Jpn. Soc. Hort. Sci., 69, 141-148.
•  Tanaka,  F.,  Miyazawa,  T.,  Okazaki,  K.,  Tatsuki,  M., and  Ito,  T. (2015). Sensory and metabolic profiles of ‘Fuji’ apples (Malus domestica Borkh.) grown without synthetic agrochemicals: the role of ethylene production. Biosci. Biotechnol. Biochem., 79, 2034-2043.
•  Van Ruth,  S.M. (2001). Methods for gas chromatography-olfactometry: a review. Biomol. Eng., 17, 4-5.
•  Yamane,  H.,  Kurihara,  A., and  Tanaka,  R. (1978) Studies on polyploidy breeding in grapes. 1. Chromosome numbers of large-berried grape varieties grown in Japan. Bull. Fruit Tree Res. Stn. E., 2, 7-8.
•  Yamada,  M.,  Yamane,  H.,  Sato,  A.,  Hirakawa,  N.,  Iwanami,  H.,  Yoshinaga,  K.,  Ozawa,  T.,  Mitani,  N.,  Shiraishi,  M.,  Yoshioka,  M.,  Nakajima,  I.,  Nakano,  M., and  Nakaune,  R. (2008). New grape cultivar ‘Shine Muscat’. Bull. Natl. Fruit Tree Res. Stn., 7, 21-38 (in Japanese).
•  Yamada,  M. and  Sato,  A. (2016) Advances in table grape breeding in Japan. Breed Sci., 66, 34-45.
•  Wu,  Y.,  Duan,  S.,  Zhao,  L.,  Gao,  Z.,  Luo,  M.,  Song,  S.,  Xu,  W.,  Zhang,  C.,  Ma,  C., and  Wang,  S. (2016). Aroma characterization based on aromatic series analysis in table grapes. Sci. Rep, 6:31116.
•  Wang,  J. and  De Luca,  V. (2005). The biosynthesis and regulation of biosynthesis of Concord grape fruit esters, including ‘foxy’ methyl anthranilate. Plant J., 44, 606-619.
• i) https://www.bbc.com/news/world-asia-36734764 (Dec. 25, 2018)
• ii) http://time.com/4396473/expensive-grapes-japan-auction/ (Dec. 25, 2018)