16.IJFS#1675_bozza Ital. J. Food Sci., vol. 32, 2020 - 482 PAPER CHARACTERIZATION OF VOLATILE COMPOUNDS IN FIVE BLUEBERRY VARIETIES USING PURGE AND TRAP COUPLED TO GAS CHROMATOGRAPHY-MASS SPECTROMETRY Z. XINYU1, W. YI1, H. XUE1, Y. ZIXI1, Y. WEIQIONG1 and L. ZHAOLIN*2, 3 1College of Biological Sciences and Biotechnology, Beijing Forestry University, Beijing 100083, P.R. China 2Analysis and Testing Center, Beijing Forestry University, Beijing 100083, P.R. China 3Department of Beijing Key Laboratory of Forest Food Process and Safety, Beijing Forestry University, Beijing 100083, P.R. China *Corresponding author: zhaolinlv@126.com ABSTRACT The volatile composition of five blueberry varieties from two different regions was analysed by dynamic headspace (purge and trap, P&T) coupled to gas chromatography- mass spectrometry (GC-MS). Under the optimized conditions, the P&T method was successfully validated, showing good linearity, high accuracy, good reproducibility and a low limit of detection. A total of 80 volatiles were identified, including 19 esters, 30 alcohols, 18 aldehydes, 7 ketones and 6 other compounds. Furthermore, a spider web diagram was constructed to compare the flavour profiles of these blueberries, and the obtained results demonstrated that blueberries from different locations have different flavour profiles. Keywords: aroma active compounds, blueberry, GC-MS, P&T, volatile compounds Ital. J. Food Sci., vol. 32, 2020 - 483 1. INTRODUCTION Blueberries have been recognized by the scientific community and consumers for their health-promoting potential (SILVA et al., 2017). The history of blueberry cultivation in China is approximately 20 years old, and Chinese blueberries were mainly introduced from the United States and Japan. There are three main types of blueberries: highbush (Vaccinium corymbosum), lowbush (Vaccinium angustifolium), and rabbiteye (Vaccinium virgatum). Highbush blueberries can be further divided into northern highbush and southern highbush blueberries (DU and ROUSEFF, 2014). Northern highbush and lowbush blueberries are the predominant varieties in the Greater Khingan Range. Southern highbush and rabbiteye blueberries are generally grown to the south of the Yangtze River (HE and WU, 2010). In addition to being rich in vitamins and anthocyanins, blueberries are rich in volatile compounds such as ethyl acetate, butyl acetate, and 1-nonanal. Flavour and aroma are two of the most important fruit quality characteristics and ultimately determine consumer acceptability and purchase decisions (DU and ROUSEFF, 2014). Volatile compounds are important contributors to fruit aroma, which is one of the main characteristics that determine blueberry organoleptic quality and style (SUN et al., 2013). Different proportions of volatile components determine the overall aromatic properties (LV and LIN, 2015). People realized the importance of volatile compounds with regards to aroma approximately 50 years ago. However, due to equipment being less advanced, studies of blueberry aroma are still very limited. The volatile compounds of highbush blueberries were analysed by PARLIAMENT and KOLOR in 1975, and 18 individual components were identified by mass spectrometry, infrared analysis and gas chromatographic retention times (PARLlMENT and KOLOR, 1975). HALL et al. (1970) used gas-liquid chromatography (GLC) to examine the aromatic composition of lowbush blueberries. Acetaldehyde, methyl acetate, ethyl acetate and ethyl alcohol were reported as the major aromatic compounds. Currently, with the emergence of detection techniques with high sensitivity and accuracy, such as gas chromatography-olfactometry (GC-O), gas chromatography-mass spectrometry (GC-MS) and liquid chromatography-mass spectrometry (LC-MS), more volatile compounds at relatively low concentrations and thresholds are expected to be detected. Purge and trap (P&T), also known as dynamic headspace, has been widely used for the preconcentration of volatile compounds (LARRETA et al., 2008). With P&T, an inert gas is purged throughout the sample in the same way as which we breathe, making this technique suitable for correlation with organoleptic studies (AZNAR and ARROYO, 2007). It can be applied to solid or liquid matrices (MURAT et al., 2012). Compared to SPME, the high recovery of very volatile compounds and the low dispersion associated with the use of a totally automated system are the main advantages of P&T-GC-MS-based methods (SORIA et al., 2009). In this study, five blueberry varieties from two major blueberry production areas were identified (i) by purge and trap coupled to gas chromatography-mass spectrometry (P&T- GC-MS). To provide a representative analysis of the blueberry volatiles, we first (ii) optimized this method and evaluate its correctness and then (iii) drew a spider web diagram to compare the flavour profiles of these blueberries. Ital. J. Food Sci., vol. 32, 2020 - 484 2. MATERIALS AND METHODS 2.1. Plant materials All the samples were purchased from the Hulun Buiroroqen Pristine Production Co. Ltd. Blueberries were squeezed into juice, diluted three-fold, and filtered for analysis. In this work, a total of five blueberry taxa were used to study volatiles. These taxa included two wild blueberries and three cultivated blueberries (Table 1). Table 1. Blueberry taxa in this study*. Taxa Characters Origin Population Wild blueberry around the humus Mohe area of Greater Khingan Range WH-M Wild blueberry around the stones Mohe area of Greater Khingan Range WS-M Cultivated blueberry Bluecrop Greater Khingan Range CB-G Cultivated blueberry Powderblue Greater Khingan Range CP-G Cultivated blueberry Britewell Yangzhou CB-Y *CB-G is northern highbush blueberry; CP-G and CB-Y are rabbiteye blueberries. 2.2. Chemicals NaCl and n-alkanes (C6-C22) were purchased from Beijing Chemical Reagents Co. Ltd. (Beijing, China). Analytical grade 2-methylbutyraldehyde, ethyl acetate, 2-nonanone, linalool, and ethyl caprylate were purchased from Sigma-Aldrich Co. Ltd. (Shanghai, China). 2.3. Volatile compounds extracted by purge and trap P&T was performed by an Eclipse 4660 purge and trap sample concentrator with a 4551A autosampler (OI Analytical Company, USA) and a #10 trap. Three millilitres of each juice sample was placed in a 5 mL purge tube. Nitrogen gas was utilized as a purge at 10 psi at 25°C. The other analytical conditions were as follows: Trap temperature: purge, 30°C; desorption, 190°C; transfer line, 110°C; and valve oven, 110°C. Time: purge 11 min; desorption 1 min. 2.4. GC-MS conditions Chromatographic analysis was performed in a GC-MS (QP2010 Ultra, Shimadzu Corporation, Japan) system equipped with a Rtx-5MS capillary column (0.25 mm×30 m×0.25 μm) (Restek, USA). Helium was used as the carrier gas at a linear velocity of 1.0 mL/min. The column temperature was held at 50°C for 5 min, increased to 180°C at a rate of 10°C/min, increased to 210°C for 5 min at a rate of 5°C/min, and finally increased to 280°C at a rate of 20°C/min. The mass selective was operated in the electron ionization mode at 70 eV and a scan range m/z of 45-400. Ital. J. Food Sci., vol. 32, 2020 - 485 2.5. Identification of volatile compounds Volatile compounds were identified by matching their mass spectra with those of the known compounds from the NIST 11/11s edition library. The relative odour activity value (ROAV) was calculated to measure the contribution of each volatile compound towards the whole aroma profile and was calculated using the following equation (ZHUANG et al., 2008; GU et al., 2012). ROAVs were calculated by using Eq. (1): ROAVi = !"% !"#$%% × !"#$% !" ×100 (1) where “stan” is the volatile compound that has the highest relative contents; ROAVi is the odour activity value of the compound in sample i; Ci is its content; and Ti is its odour threshold concentration. Compounds with a ROAV ≥1 significantly contribute to the aroma. (ZHUANG et al.,2016). 2.6. Statistical analysis Significant differences in the volatile compounds of the five blueberry varieties obtained from duplicate analysis were determined by one-way ANOVA with SPSS 17.0 for Windows (SPSS Inc., Chicago, IL). Statistically significant differences were determined at p<0.05. The OriginPro system (v8.5 SR6, OriginLab Corporation, Northampton, MA, USA) was used for statistical analysis. 3. RESULTS AND DISCUSSION 3.1. Optimization of the P&T-GC-MS method This study optimized the following P&T extraction parameters: sample volume, purge temperature and purge time. Ethyl acetate, ethyl caprylate, 2-methylbutyraldehyde, 2-nonanone and linalool were used as standard compounds for optimization of the P&T-GC-MS method. As shown in Fig. 1, varying volumes of blueberry juice (3, 4, and 5 mL) were placed in the trapping apparatus flask and purged for 11 min at 25°C. For ethyl octanoate, ethyl acetate, and linalool, there was a considerable difference between the various sample volumes (p<0.05). For 2- methylbutyraldehyde and 2-nonanone, the relative percentages of these standard compounds in the 3 mL groups increased compared with the high sample quality group, but there was no significant difference (p>0.05). This study also showed that the number of volatile substances obtained from 3, 4, and 5 mL was 52, 50, and 49, respectively. The reason for this result may be that a high liquid level is too close to the top of the purge trap, so when a large amount of N2 purifies the liquid, extra water could be purged into the trap, which can shorten the trap life in the same way as a longer purge time (DENG et al., 2011). Ital. J. Food Sci., vol. 32, 2020 - 486 Figure 1. Effect of sample volume on the extraction efficiency; purge temperature = 25°C, purge time = 11 min. Figure 2. Effect of purge temperature on the extraction efficiency. Sample volume = 3 mL, purge time = 11 min. Ital. J. Food Sci., vol. 32, 2020 - 487 Fig. 2 showed the effect of purge temperature on the extraction efficiency. The purge temperature varied between 25-60°C with 3 mL of sample volume for 11 min. The amount of total volatiles detected in blueberries gradually decreased as the purge temperature increased from 25°C to 60°C, probably due to the amount of water that reached the trap and decreased the sensitivity; therefore, ambient temperature was maintained in all the experiments (CAMPILLO et al., 2004). The effect of purge time on the sensitivity is shown in Fig. 3. The purge time was varied between 8 and 14 min with 3 mL of sample volume at 25°C. Finally, a value of 11 min was chosen as the optimal time, since 8 and 14 min led to a slight decrease in the peak area and total number. Eight minutes is too short mainly because the volatile substances are not fully blown out. Indeed, 14 min decreased the signals because a flow of N2 that was too long could move the volatiles from the trap before desorption and reduce the final signal (CAMPILLO et al., 2004). Figure 3. Effect of purge time on the extraction efficiency. Sample volume = 3 mL, purge temperature = 25°C. Verification and quantitative analysis (of P&T-GC-MS method) Once the final purge conditions were selected, these five aroma standards were detected. Table 2 shows the results from method validation: linearity, recovery, reproducibility, LOD and LOQ. Linearity The linearity of the method was evaluated by analysing a series of aromatic standards. Linearity was found in the concentration range between 5 and 160 µg/L, with high Ital. J. Food Sci., vol. 32, 2020 - 488 reproducibility and accuracy. Regression analysis of the experimental data points showed a linear relationship with excellent regression coefficients (r2>99%) for 2-methylbutyraldehyde, ethyl acetate, 2-nonanone, linalool, and ethyl caprylate. Table 2. Performance parameters of the P&T method for the volatile compounds in blueberry. Compounds Linearity (r2) Recovery (%) C.V. (%) LOD (μg/L） LOQ (μg/L) 2-Methylbutyraldehyde 0.9982 117.01 4.3418 0.78 2.60 Ethyl acetate 0.9971 91.52 1.7276 0.90 3.00 2-Nonanone 0.9957 103.47 2.8291 1.02 3.40 Linalool 0.9933 99.33 5.4841 1.06 3.50 Ethyl caprylate 0.9964 106.18 7.1420 1.29 4.30 Recovery Recoveries ranged between 96% and 120%, indicating that the accuracy of the method meets the experimental requirements and that the results are reliable. Reproducibility Reproducibility was evaluated by using the coefficient of variation (CV%) for replicate analyses. The CV% values obtained are shown in Table 3. CV% values were found to be <8% in the case of relative proportions (HAKALA et al., 2002). The smallest CV% was found for ethyl acetate (1.73%), and the largest was found for ethyl octanoate (7.14%). In the range of esters, as the carbon number increases, the coefficient of variation also increased. As shown above, the P&T-GC-MS technique was reproducible enough to allow for comparative comparison studies of the volatiles of different varieties (HAKALA et al., 2002). Determination of the limit of detection (LOD) and the limit of quantification (LOQ). The LOD was calculated as the concentration required to obtain a signal that was three times higher than that of the baseline signal (PINO and OUERIS, 2010). Detection limits were below 1.29 µg/L for all volatiles. The LOQ can also be estimated as the concentration of analyte producing a signal that is 10 times that of the noise (S/N = 10) (PINO and OUERIS, 2010). From the above results, good linearity, high accuracy, very good repeatability and a low limit of detection were achieved (DENG et al., 2011). There were also good recoveries and reproducibility. This method can be applied for research on the volatiles in blueberries. In conclusion, 3 mL of sample purged at 25°C for 11 min were selected as the best extraction conditions for the P&T methodology developed in this study. Ital. J. Food Sci., vol. 32, 2020 - 489 Table 3. Analysis of volatile compounds from different blueberry varieties. No. RI tR (min) Compounds Relative content (%) WH-M WS-M CB-G CP-G CB-Y Esters 1 487 1.603 Methyl acetate 0.14±0.01 2 584 1.644 Ethyl formate 9.9±0.50a 8.81±0.43a 9.54±0.51 a 3 586 1.891 Ethyl acetate 60.94±3.05a 66.21±3.24a 4.89±0.22c 1.58±0.09c 17.36±0.95b 4 686 2.518 Ethyl propionate 0.48±0.02a 0.11±0.01b 5 686 2.540 Propyl acetate 0.14±0.01a 0.10±0.01a 6 686 2.620 Methyl butyrate 0.03±0.00b 0.84±0.03 a 7 778 2.796 Isopentyl formate 0.24±0.02c 0.60±0.03b 1.00±0.04 a 8 785 3.053 Ethyl butyrate 0.04±0.00a 0.04±0.00a 9 785 3.247 Butyl acetate 9.80±0.50a 9.30±0.42a 5.09±0.27 b 10 820 4.680 Ethyl isovalerate 1.78±0.10b 1.32±0.08b 12.66±0.68 a 11 864 5.193 Amyl acetate 3.97±0.21a 3.80±0.24a 12 869 6.330 Prenylacetate 0.02±0.00b 0.24±0.01 a 13 869 6.414 Ethyl 3,3-dimethylacrylate 0.02±0.00 14 983 7.614 Ethyl 2-hydroxy-3-methylbutanoate 0.04±0.00 15 1029 8.875 Hexenyl acetate 0.16±0.01a 0.06±0.00b 16 1043 8.900 Butyl pentanoate 0.01±0.00b 0.02±0.00b 0.17±0.01 a 17 1047 9.117 Hexyl acetate 0.02±0.00a 0.01±0.00a 18 1277 18.500 L-Bornyl acetate 0.10±0.01 19 1294 19.949 2-Methylpropyl benzoate 0.07±0.00c 0.06±0.00c 0.40±0.02a 0.27±0.01b Alcohols 20 662 2.156 n-Butyl alcohol 0.10±0.00b 16.41±0.76 a 21 788 2.275 Cyclopentanol 0.57±0.03 22 700 2.817 2-Methyl-1-butanol 0.12±0.01b 0.32±0.02a 23 769 2.982 2-Penten-1-ol 0.12±0.01 Ital. J. Food Sci., vol. 32, 2020 - 490 24 995 4.659 6-Methyl-heptanol 1.85±0.08a 0.26±0.01 b 25 858 4.934 2-Hexen-1-ol 0.05±0.00d 0.17±0.01cd 0.51±0.03b 1.19±0.07a 0.26±0.01c 26 860 5.020 Hexyl alcohol 0.51±0.03b 0.67±0.03b 2.08±0.12b 3.77±0.20a 2.07±0.98b 27 960 7.780 n-Heptanol 0.10±0.00 28 969 8.092 1-Octen-3-ol 0.07±0.00c 0.06±0.00c 0.35±0.02b 0.38±0.02b 0.94±0.05a 29 969 8.835 Citronellol 1.49±0.06 30 971 9.401 3-Ethyl-4-methyl-1-pentanol 0.12±0.01 31 1042 9.535 4-Isopropyltoluene 0.08±0.00 b 0.31±0.02a 32 1055 9.664 2-Ethylhexanol 1.84±0.11c 1.16±0.01c 11.2±0.52a 9.17±0.43b 10.61±0.59ab 33 1059 9.795 Eucalyptol 0.07±0.00bc 0.10±0.00b 0.39±0.02a 0.10±0.01d 0.06±0.00c 34 1060 11.095 1-Octanol 0.13±0.01b 0.09±0.00b 0.92±0.05a 0.87±0.03ab 0.33±0.02ab 35 1063 11.157 Dihydromyrcenol 1.88±0.10 36 1082 12.093 Linalool 0.05±0.00c 0.03±0.00c 0.88±0.04b 0.93±0.04b 2.68±0.14a 37 1138 12.753 Fenchyl alcohol 0.10±0.01b 0.53±0.03 a 0.18±0.01b 38 1153 14.249 Menthol 0.22±0.02 39 1158 14.598 Borneol 0.96±0.05 40 1159 14.622 1-Nonanol 0.08±0.00c 0.05±0.00c 0.44±0.02b 0.75±0.03a 41 1164 14.817 DL-Menthol 0.68±0.03b 0.85±0.04a 0.65±0.03b 0.47±0.02c 42 1187 14.914 4-Terpineol 0.15±0.01c 0.12±0.01c 0.80±0.03a 0.27±0.01b 43 1198 15.422 (-)-α-Terpineol 0.07±0.00c 0.05±0.00c 0.51±0.02a 0.38±0.02b 0.06±0.00c 44 1228 17.356 Geraniol 0.99±0.05 b 13.28±0.57a 45 1258 18.133 1-Decanol 0.06±0.00b 1.50±0.07 a 46 1200 18.536 cis-Anethol 0.06±0.00b 0.04±0.00b 0.17±0.02a 47 1262 18.978 Thymol 0.10±0.00 48 1457 24.504 1-Dodecanol 0.37±0.15 49 1543 28.659 Cedrol 0.07±0.00 Aldehydes and ketones 50 508 1.589 Propionaldehyde 2.93±0.14 b 2.78±0.12b 3.66±0.20a 51 543 1.725 Isobutyraldehyde 0.16±0.01 b 0.53±0.02 a 52 555 1.815 2-Butanone 0.92±0.04 Ital. J. Food Sci., vol. 32, 2020 - 491 53 643 2.008 2-Methylbutyraldehyde 1.57±0.04 a 1.22±0.07b 54 644 2.279 1-Penten-3-one 0.04±0.00b 0.65±0.03 a 55 654 2.384 3-Pentanone 0.17±0.01 b 0.84±0.05 a 56 715 2.982 2-Pentenal 0.08±0.00 57 791 3.613 4-Methyl-3-pentene-1-one 0.07±0.00 c 22.98±1.13a 21.9±0.99a 8.49±0.37b 58 806 3.654 Hexanal 0.88±0.04b 16.59±0.96 a 17.3±0.87a 59 831 4.225 Furfural 0.14±0.01b 0.12±0.01b 0.12±0.01b 4.13±0.23a 60 853 5.500 2-Heptanone 0.01±0.00c 0.22±0.01 b 1.68±0.07a 61 841 5.816 4-Methylhexanal 0.02±0.00b 0.17±0.01 b 2.07±0.11a 62 913 7.358 2-Heptenal 0.04±0.00 63 982 7.509 Benzaldehyde 0.07±0.00d 0.04±0.00d 0.65±0.03b 0.92±0.06a 0.21±0.01c 64 1005 8.823 Octanal 0.03±0.00d 0.17±0.01 b 0.21±0.01a 0.13±0.00c 65 1013 10.619 2-Octenal 0.04±0.00 66 1052 11.762 2-Nonanone 0.03±0.00b 0.03±0.00b 0.22±0.01 a 67 1104 12.251 Nonanal 1.38±0.06c 1.39±0.07c 7.01±0.42b 9.22±0.51a 2.18±0.14c 68 1112 14.194 (2E)-Nonenal 0.18±0.02 69 1151 15.300 2-Decanone 0.02±0.00 70 1204 15.807 Decanal 1.84±0.01a 0.50±0.03bc 0.58±0.02b 0.47±0.03c 71 1208 16.063 2,4-Dimethylbenzaldehyde 0.16±0.01c 0.10±0.01c 0.57±0.03a 0.45±0.02b 72 1263 16.358 5-Hydroxymethylfurfural 4.46±0.29 73 1402 22.595 Dodecyl aldehyde 0.01±0.00b 0.14±0.01 a 74 1420 23.826 (Z)-Geranyl acetone 0.36±0.02d 0.58±0.03c 0.74±0.04b 1.19±0.06a Others 75 877 4.514 3,7-Dimethyl-1-octene 0.02±0.00b 0.61±0.03 a 76 883 5.574 Phenylethylene 0.05±0.00c 0.04±0.00c 0.18±0.01b 0.27±0.01a 77 1029 10.851 Acetophenone 0.04±0.00b 0.03±0.00b 0.14±0.01b 3.37±0.20a 78 1231 14.992 Naphthalene 0.48±0.02c 0.46±0.02c 1.63±0.08b 2.35±0.13a 0.45±0.02c 79 1407 22.876 Cedarene 0.84±0.05 80 1668 25.545 Butylated hydroxytoluene 0.77±0.04c 1.17±0.06b 2.53±0.11a Ital. J. Food Sci., vol. 32, 2020 - 492 3.2. Identification of the volatile compounds in five blueberry varieties As shown in Table 3, the volatile compounds in the five blueberry varieties were identified. A total of 80 volatiles were identified, including 19 esters, 30 alcohols, 18 aldehydes, 7 ketones and 6 other compounds. The number of identified volatile compounds in each blueberry variety ranged from 30 to 53. WH-M and CB-G had the highest (53) and the second highest number (47) of volatile compounds, respectively, while CB-Y had the smallest number (30) of volatile compounds. Esters are considered to be contributors to fruity and floral notes (WANG et al., 2009). A total of 21 ester compounds were detected in the five blueberry varieties. Ethyl acetate is a common compound that has a strong fruity aroma. Among the 5 groups, the sum of the esters was higher in WS-M and WH-M blueberries than in the other cultivated groups. Esters were abundant in wild blueberries, contributing 87.66-90.44% of the total volatiles (Table 4). Although 13 esters in total were found in wild blueberries in this study, ethyl acetate, ethyl formate, butyl acetate and amyl acetate accounted for more than 80% of the total esters in WH-M and WS-M. The unique esters of WH-M were methyl butyrate, prenylacetate, ethyl 3-methyl-2-butenoate, ethyl 3,3-dimethylacrylate, and ethyl 2- hydroxy-3-methylbutanoate, with the latter two in agreement with previous results (BEAULIEU et al., 2014). L-Bornyl acetate was only detected in CB-G. 2-Methylpropyl benzoate was detected in all varieties except CB-Y. Esters were not considered to be as important as aldehydes to the aroma in highbush blueberries, while they have been identified as important volatiles in some rabbiteye blueberries, which is consistent with previous results (Du and ROUSEFF, 2014). The total content of alcohols accounted for 4.7-35.98% of the total volatiles (Table 3). The content of alcohols was significantly higher in cultivated blueberries than in wild blueberries. Of the 30 alcohols identified in this study, 8 were identified in all five varieties: 2-hexen-1-ol, hexyl alcohol, 1-octene-3-ol, eucalyptol, 1-octanol, linalool, 2- ethylhexanol, and (-)-α-terpineol. Among them, 2-ethylhexanol was dominant, with relative contents ranging from 1.16% to 11.20% (Table 4). 2-Methyl-1-butanol was detected in WS-M and WH-M. 2-Penten-1-ol, 3-ethyl-4-methyl-1-pentanol, borneol and menthol were only detected in CB-G. The unique alcohols in CP-G and CB-Y were citronellol and cyclopentanol, respectively. A total of 25 different aldehydes and ketones in blueberry juice were identified, accounting for 3.16%-68.67% of the total volatiles (Table 4). The sum of the aldehydes and ketones in CP-G was significantly higher than that in other varieties, and it was also significantly higher in cultivated blueberries than in wild blueberries. Nonanal and benzaldehyde were the predominant aldehydes found in the five blueberry varieties. WH- M had a significantly higher decanal content than that of the other aldehydes. In all cultivated groups, 4-methyl-3-pentene-1-one was the major component, accounting for more than 20% of the total aldehydes in CB-G and CP-G. 2-Pentenal, 2-heptenal, and 2- octenal were only detected in CB-G. Additionally, (2E)-nonenal, 5-hydroxymethylfurfural and dodecyl aldehyde could be used to distinguish CP-G from the other varieties. 2- Butanone was only detected in CB-Y. Ital. J. Food Sci., vol. 32, 2020 - 493 Table 4. The aroma-active compounds (ROAV > 1) in different blueberries*. No. Volatile Threshold （μg/L） Sensory attributes Aroma classificatio n ROAV WH-M WS-M CB-G CP-G CB-Y 1 Ethyl formate 150 Fruity 1 0.54 ±0.03b 0.44± 0.02b 1.83 ±0.10a 2 Ethyl acetate 5 Fruity 1 100.00±5.00a 100.00±4.89a 13.95±0.63b 3.43±0.20b 100.00 ±5.47a 3 Butyl acetate 66 Sweet, banana, 1,3 1.22±0.06b 1.06 ±0.05b 2.22 ±0.12a 4 1-Octene-3-ol 1 Mushroom 4 0.57 ±0.00c 0.45 ±0.00c 4.99±0.29b 4.12 ±0.22b 27.07±1.44a 5 Linalool 6 Sweet lemon 1 0.07 ±0.00c 0.04 ±0.00c 2.09 ±0.10b 1.68 ±0.07b 12.86 ±0.67a 6 Geraniol 40 Rose 2 0.19 ±0.01b 4.74 ±0.20a 7 2-Methylbutyraldehyde 1 Stimulating, coffee, sweet 1,3,6 17.03 ±0.43 b 35.14±2.02a 8 Hexanal 5 Fragrant, grassy 5 1.44 ±0.07c 47.33 ±2.74a 37.53±1.89b 9 4-Methylhexanal 3 Fruity, rose 1,2 0.05 ±0.00b 0.81±0.05b 7.48±0.40a 10 Octanal 0.7 Rose, orange 1,2,3 0.35 ±0.00c 3.46 ±0.20b 3.25 ±0.15b 5.35 ±0.00a 11 Nonanal 1 Floral, citrus, slightly spicy 1,2,6 11.32 ±0.49c 10.50 ±0.53c 100.00±5.99a 100.00±5.53a 62.79 ±4.03b 12 Decanal 3 Fruity 1 5.03 ±0.03b 1.26±0.08d 2.76±0.10a 1.70±0.11c ∗Intensity: 1-fruity, 2-floral, 3-sweet, 4-fatty, 5-fragrant, 6-stimulating Ital. J. Food Sci., vol. 32, 2020 - 494 3.3. Determination of the aroma active compounds in different blueberries Considering that volatile compounds have different thresholds and people have different sensitivities to them, the relative content cannot reflect the true contribution that every volatile compound makes to the whole aroma profile. Therefore, we used ROAVs to detect the contribution of volatile compounds to the whole aroma profile (YI et al., 2016). Fourteen aroma active compounds were selected from five blueberry varieties, which are shown in Table 4. There were four aroma active compounds (ethyl acetate, 1-octene-3-ol, linalool, nonanal) with higher ROAVs in five varieties. Ethyl acetate and nonanal possessed the highest ROAVs in wild blueberries and cultivated blueberries, respectively. CB-Y had the highest ROAV summations, which was significantly higher than the other four varieties. Aldehydes were the most abundant chemical group, with aromatic activity found in five blueberry varieties. 2-Methylbutyraldehyde, hexanal, 4-methylhexanal, 1-octanal, nonanal and decanal contributed to stimulating, fragrant, fruity, rose, floral and fruity aroma notes, respectively. 2-Methylbutyraldehyde was observed only in CP-G. 2-Methylbutyraldehyde has stimulating, coffee, and sweet aroma notes, with a very low threshold (1 !g/L) in CP- G. Aldehydes made a major contribution to blueberry aromas, which is in agreement with previous results (Du and ROUSEFF, 2014; HORVAT and SENTER, 1985). Alcohols were the next most abundant group, including 1-octene-3-ol, linalool, and geraniol, contributing mushroom, lemon and rose aroma notes. 1-Octene-3-ol and linalool were identified in the five blueberry varieties. Three esters, including methyl acetate, ethyl acetate and butyl acetate, were aroma active. Ethyl formate had a high threshold value (150 !g/L) and a high relative content. However, its ROAVs were low (0.32-1.53). Ethyl acetate contributed a fruity aroma to the five varieties and possessed the highest ROAV in wild blueberries. Butyl acetate contributed sweet and banana aroma notes. However, it has not been previously reported as contributing to blueberry aroma. Although wild blueberries had higher contents of volatile compounds, their characteristic aroma notes were less than those of cultivated blueberries. The reason may be that the aroma of fruit is not completely dependent on the concentration of the volatile compound but it is closely related to its threshold. The threshold of volatile compounds found differed greatly among the varieties studied. For example, the relative contents of ethyl formate in the two wild blueberries were higher than those in cultivated blueberries, but the ROAVs were lower because the threshold value of methyl acetate was high (150 !g/L). Six descriptors (fruity, floral, sweet, fatty, fragrant, and stimulating) were used to provide an assessment of the five blueberries. To reflect the difference in aroma among different blueberry varieties, the ROAV of each blueberry aroma component was taken as the logarithm base 10, and the aromatic series of the five blueberry juices on the spider web diagram are shown in Fig. 4. Ital. J. Food Sci., vol. 32, 2020 - 495 Figure 4. Aromatic series in blueberries based on aroma activity values. The analysis showed that WH-M and WS-M could mostly be described as having fruity and floral notes due to the higher ROAVs of ethyl acetate and nonanal in the samples. CB- G and CP-G had higher values for the attributes fragrant and floral due to their large quantities of hexanal and nonanal. The difference between CB-G and CP-G lies in the fact that CB-G exhibited a greater sweet component. The ROAV of 1-octene-3-ol was higher in CB-Y; thus, CB-Y was perceived to have a fatty aroma. Considering the volatile composition of these blueberries, samples had higher values for the attributes fruity and fragrant due to their large quantities of aldehydes and alcohols. 4. CONCLUSIONS AND FUTURE WORK The P&T extraction method coupled to GC-MS analysis was a quick and efficient method for the evaluation of blueberry volatiles, and the results demonstrated that 3 mL of sample volume purged at 25°C for 11 min were the best extraction conditions. A total of 80 volatiles were identified in five blueberry varieties using the P&T-GC-MS technique. The volatiles of blueberries were composed of mainly aldehydes, alcohols, esters, and terpenes. Among the identified compounds, 12 compounds (ROAV>1), including ethyl formate, ethyl acetate, butyl acetate, 1-octene-3-ol, linalool, geraniol, 2-methylbutyraldehyde, hexanal, 4-methylhexanal, octanal, nonanal and decanal, were considered aroma active. The spider web diagram showed that the sensory characterization of the five varieties was distinct due to the different quantities of volatile compounds. ACKNOWLEDGEMENTS This work was supported by Water and soil conservation Development and Management Center, Ministry of Water Resources. (Project No.2020 swxyjskf001). -1,5 -1 -0,5 0 0,5 1 1,5 2 2,5 Fruity Floral Sweet Fatty Fragrant Stimulating WH-M WS-M CB-G CP-G CB-W Ital. J. Food Sci., vol. 32, 2020 - 496 REFERENCES Aznar M. and Arroyo T. 2007. Analysis of wine volatile profile by purge-and-trap–gas chromatography–mass spectrometry. J. Chromatogr. A. 1165:151-157. Beaulieu J.C., Stein-Chisholm R.E. and Boykin D.L. 2014. Qualitative analysis of volatiles in rabbiteye blueberry cultivars at various maturities using rapid solid-phase microextraction. J. Amer. Soc. Hort. Sci. 139:167-177. Campillo N., Viñas P., López-Garcı ́a I., Aguinaga N. and Hernández-Córdoba M. 2004. Purge-and-trap capillary gas chromatography with atomic emission detection for volatile halogenated organic compounds determination in waters and beverages. J. Chromatogr. A. 1035:1-8. Deng X., Liang G., Chen J., Qi M. and Xie P. 2011. Simultaneous determination of eight common odors in natural water body using automatic purge and trap coupled to gas chromatography with mass spectrometry. J. Chromatogr. A. 1218:3791-3798. Du X. and Rouseff R. 2014. Aroma active volatiles in four Southern highbush blueberry varieties determined by gas chromatography–olfactometry (GC-O) and gas chromatography-mass spectrometry (GC-MS). J Agric. Food Chem. 62:4537-4543 Gu S.Q., Tao N.P., Wu U. and Zhang J.J. 2012. A new method based on ROAV value to identify the characteristic key volatile compounds of crab flavor. Sci. Technol. Food Indus. 33:410-16. Hakala M.A., Lapveteläinen A.T. and Kallio H.P. 2002. Volatile compounds of selected strawberry varieties analyzed by purge-and-trap headspace GC-MS. J. Agri. Food Chem. 50:1133-1142 Hall I.V., Forsyth F.R. and Lightfoot H.J. 1970. Volatiles from developing fruit of Vaccinium angustifolium. Can. Inst. Food Sci. Technol. 3:1-3 He Q. and Wu L.R. 2010. Biological functions of nutrients in blueberry fruits. Nor Horticul. 24:222-224. Horvat R.J. and Senter S.D. 1985. Comparison of the volatile constituents from rabbiteye blueberries (Vaccinium ashei) duringripening. J. Food Sci. 50:429-431. Larreta J., Bilbao U., Vallejo A., Usobiaga A., Arana G. and Zuloaga O. 2008. Multisimplex optimisation of the purge- and-trap preconcentration of volatile fatty acids, phenols and Indoles in cow slurries. Chromatogr. A. 67:93-99 Lefngwell, and Associates, www.lefngwell.com/odorthre.htm Lv Z.L. and Lin X. 2012. Antioxidant activity of bamboo-leaf extracts from the species Dendrocalamopsisoldhami. Sci. Res. Essays. 44:3789-3796. Murat, C., Gourrat, K., Jerosch, H., and Cayot, N. 2012. Analytical comparison and sensory representativity of SAFE, SPME, and purge and trap extracts of volatile compounds from pea flour. Food Chem. 135:913-920. Parliament T.H. and Kolor M.G. 1975. Identification of the major volatile components of blueberry. J. Food Sci. 40:762- 763. Pino J.A. and Queris O. 2010. Analysis of volatile compounds of pineapple wine using solid-phase microextraction techniques. Food Chem. 122:1241-1246. Silva S., Costa E.M., Calhau C., Morais R.M. and Pintado M.M.E. 2017. Production of a food grade blueberry extract rich in anthocyanins: selection of solvents, extraction conditions and purification method, J. Food Meas. Charact. 11:1248- 1253. Soria A.C., Martínez-Castro I. and Sanz J. 2009. Study of the precision in the purge-and-trap–gas chromatography-mass spectrometry analysis of volatile compounds in honey. J. Chromatogr. A. 1216:3300-3304. Sun S.Y., Che C.Y., Sun T.F., Lv Z.Z., He S.X., Gu H.N., Shen W.J., Chi D.C. and Gao Y. 2013. Evaluation of sequential inoculation of Saccharomyces cerevisiae and Oenococcus oeni strains on the chemical and aromatic profiles of cherry wines. Food Chem. 138:2233-2241. Wang Y., Yang C., Li S., Yang L., Wang Y., Zhao J. and Jiang Q. 2009. Volatile characteristics of 50 peaches and nectarines evaluated by HP-SPME with GC-MS. Food Chem. 116:356-364. Ital. J. Food Sci., vol. 32, 2020 - 497 Yi, Z., Feng, T., Zhuang, H., Ye, R., Li, M., and Liu, T. 2016. Comparison of different extraction methods in the analysis of volatile compounds in pomegranate juice. Food Anal Method. 9:2364-2373. Zhuang K., Wu N., Wang X., Wu X., Wang S., Long X. and Wei X. 2008. “ROAV” method: a new method for determining key odor compounds of Rugao ham. Food Sci. 29:370-374. Paper Received August 28, 2019 Accepted December 20, 2019