Accepted Manuscript Impact of fruit texture on the release and perception of aroma compounds during in vivo consumption using fresh and processed mango fruits Adeline Bonneau, Renaud Boulanger, Marc Lebrun, Isabelle Maraval, Jérémy Valette, Élisabeth Guichard, Ziya Gunata PII: DOI: Reference:
S0308-8146(17)31162-7 http://dx.doi.org/10.1016/j.foodchem.2017.07.017 FOCH 21401
To appear in:
Food Chemistry
Received Date: Revised Date: Accepted Date:
29 March 2017 12 June 2017 4 July 2017
Please cite this article as: Bonneau, A., Boulanger, R., Lebrun, M., Maraval, I., Valette, J., Guichard, E., Gunata, Z., Impact of fruit texture on the release and perception of aroma compounds during in vivo consumption using fresh and processed mango fruits, Food Chemistry (2017), doi: http://dx.doi.org/10.1016/j.foodchem.2017.07.017
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Impact of fruit texture on the release and perception of aroma compounds during in vivo consumption using fresh and processed mango fruits Adeline Bonneau1,2, Renaud Boulanger2, Marc Lebrun2, Isabelle Maraval2, Jérémy Valette3, Élisabeth Guichard4, Ziya Gunata1* 1
UMR QualiSud, University of Montpellier, Place E. Bataillon, 34095 Montpellier, Cedex 5,
France 2
UMR QualiSud, CIRAD, 73 Rue J.F. Breton, 34398 Montpellier, Cedex 5, France
3
UPR BioWooEB, CIRAD, 73 Rue J.F. Breton, 34398 Montpellier, Cedex 5, France
4
CSGA, CNRS, INRA, University of Bourgogne Franche-Comté, 17 Rue Sully, 21000 Dijon,
France *Corresponding author. Tel.: +33 4 67149321; Fax: +33 4 67144292 E-mail address:
[email protected]
Abstract Two fresh (fresh cubic pieces, fresh puree) and two dried (dried cubic pieces, dried powder) products were prepared from a homogenous mango fruit batch to obtain four samples differing in texture. The aromatic profiles were determined by SAFE extraction technique and GC-MS analysis. VOCs released during consumption were trapped by a retronasal aroma-trapping device (RATD) and analysed by GC-MS. Twenty-one terpenes and one ester were identified from the exhaled nose-space. They were amongst the major mango volatile compounds, 10 of which were already reported as being potential key flavour compounds in mango. The in vivo release of aroma compounds was affected by the matrix texture. The intact samples (fresh and dried cubic pieces) released significantly more aroma 1
compounds than disintegrated samples (fresh puree, dried powder). The sensory descriptive analysis findings were in close agreement with the in vivo aroma release data regarding fresh products, in contrast to the dried products.
Keywords Fresh and dried mangoes, texture, aroma compounds, in vivo aroma release, sensory analysis.
1. Introduction Understanding flavour perception impact factors is an important challenge for the foodprocessing industry to be able to produce innovative and appreciated food products. Indeed, flavour is the main factor affecting consumers’ preferences. Several studies have reported that flavour perception and in vivo flavour release from food was impacted by the nature and amount of aroma compounds, and also by other matrix components, such as lipids (Delahunty, Piggott, Conner, & Paterson, 1996; Frank, Appelqvist, Piyasiri, Wooster, & Delahunty, 2011), polyphenols (Muñoz-González, MartinAlvarez, Moreno-Arribas, & Pozo-Bayon, 2014), sugars, acids or alcohols (Malundo, Shewfelt, Ware, & Baldwin, 2001; Marsh, Friel, Gunson, Lund, & MacRae, 2005; MuñozGonzález, Rodríguez-Bencomo, Moreno-Arribas, & Pozo-Bayón, 2014) and pectins (Boland, Delahunty, & van Ruth, 2006; Lubbers & Guichard, 2003). On the other hand food texture is a noticeable factor in flavour perception and is taken into consideration in flavour formulation (Aprea, Biasioli, Gasperi, Märk, & van Ruth, 2006; Boland, Buhr, Giannouli, & van Ruth, 2004; Frank et al., 2011; Frank, Eyres, Piyasiri, & Delahunty, 2012). The release of aroma compounds during eating is also known to be affected by human oral physiology and oral processing factors (saliva, chewing, breathing, oral cavity volume, in-mouth
2
temperature, tasting time, etc.) (Buettner & Beauchamp, 2010; Buettner & Schieberle, 2000a, 2000b). Many in vivo and in vitro experimental studies have been conducted on model food systems spiked with volatile organic compounds (VOCs) to gain insight into the contribution of food texture, oral physiology and oral processing factors on the release of aroma compounds during consumption (Feron, Ayed, Qannari, Courcoux, Laboure, & Guichard, 2014; Lubbers et al., 2003; Poinot, Arvisenet, Ledauphin, Gaillard, & Prost, 2013; Van Ruth & Roozen, 2000). However, very few studies have used real food matrices, such as fruits, to study the aroma release during oral processing (Frank et al., 2012; Ting et al., 2015; Ting et al., 2016). Two techniques have been described in the literature for studying the in vivo release of aroma compounds during food consumption. First, atmospheric pressure chemical ionisation mass spectrometry (API-MS) and proton transfer reaction mass spectrometry (PTR-MS) are often used. They enable real-time monitoring of VOC release during consumption (Frank et al., 2012). The conventional MS instruments are equipped with a quadrupole mass filter that hinders discrimination of isobaric compounds, i.e. compounds with the same nominal mass that may occur in foods. Combining PTR with a time-of-flight (TOF) mass analyser, as first mentioned in 2004, could readily overcome this problem by boosting the mass resolution (Zardin, Tyapkova, Buettner, & Beauchamp, 2014). Secondly, quite simple techniques, amenable to any laboratory and based on the trapping of VOCs exhaled through the nasal cavity on either SPME fibres (Pionnier, Sémon, Chabanet, & Salles, 2005) or an adsorbant polymer like Tenax (Buettner et al., 2000a; Muñoz-González, Rodríguez-Bencomo, et al., 2014) have been developed. Further analysis of trapped VOCs by conventional GC-MS-EI leads to easier compound identification. However, these techniques do not deliver temporal profiles of VOC release during consumption, in contrast to PTR or API based techniques. 3
The main objective of this study was to understand the impact of fruit texture both in the release of VOCs during oral processing and in sensory perception. Mango fruit was used because of its richness in aroma compounds (Munafo, Didzbalis, Schnell, Schieberle, & Steinhaus, 2014; Pino, Mesa, Munoz, Marti, & Marbot, 2005). To assess the impact of texture, four mango samples were prepared from a homogenous mango fruit batch to obtain different textures: two fresh (fresh puree, fresh cubic pieces) and two dried (dried powder, dried cubic pieces) mango samples. Hence it was possible to carry out pairwise comparisons of fresh puree and fresh cubic pieces for aroma release in vivo since they were similar in their VOC composition. The same comparison was made between dried powder and dried cubic pieces because of their similar VOC compositions. VOCs from the assessors’ exhaled nostril breath during consumption of the mango products were trapped by a retronasal aromatrapping device (RATD) containing Tenax and then analysed by GC-MS. The assessors watched a visual animation in order to calibrate their chewing and breathing cycles and swallowing time so as to limit intra- and inter-individual variability. VOCs from exhaled nostril breath were compared with those identified in the organic extracts of mango samples obtained by a convenient technique, i.e. solvent-assisted flavour evaporation (SAFE). Finally, four mango samples were evaluated by a trained sensory panel to discuss the relationship between VOCs detected in the assessors’ exhaled nostril breath and the sensory attributes of the panel.
2. Materials and methods 2.1. Chemical reagents The chemical solvents pentane, hexane, dichloromethane and methanol were purchased from Sigma-Aldrich (St. Louis, MO). The internal standard α-cedrene used for SAFE extraction and in vivo experiments was from Sigma-Aldrich. The authentic standards α-
4
pinene, β-pinene, β-myrcene, δ-3-carene, camphene, α-phellandrene, β-phellandrene, αterpinene, γ-terpinene, limonene, p-cymene, β-ocimene, α-terpinolene, α-gurjunene, αcaryophyllene, β-caryophyllene, 3-methylbutyl butanoate, heptanal, 3-hydroxy-2-butanone, nonanal, 2-phenylethanol, (E,Z)-2,6-nonadienal and toluene were from Sigma Aldrich; hexanal, 2-methyl-1-butanol, 5-methylfurfural, 1-octanol and β-ionone were from Fluka Analytical (Steinheim, Germany); 2-pentanol, 3-methyl-1-butanol, (Z)-3-hexen-1-ol, furfural, γ-butyrolactone and γ-hexalactone were from Interchim (Montluçon, France); δ-octalactone was from Fabster (Angerville, France); o-xylene was from Alfa Aesar (Karlsruhe, Germany) and 1-hexanol was from SBI Sanofi (Grasse, France). The series of n-alkanes C8 to C20 was from Sigma-Aldrich.
2.2. Fruit samples Fifty mango fruits (cv. Kent, from Côte d’Ivoire and harvested in spring 2015) were purchased at a local market in Montpellier, France. Mature fruits were selected on the basis of aspect, shape, size, skin and firmness according to the Durofel index (Bonneau, Boulanger, Lebrun, Maraval, & Gunata, 2016). Rectangular fruit slices from each mango were prepared (around 0.8 cm width 0.8 cm height 8 cm length) and divided into four equal batches while taking the same precautions and using the same protocol as previously described (Bonneau et al., 2016). One batch of fresh mangos was blended in a laboratory blender, while another batch was re-cut into cubic pieces with a knife (around 0.8 cm width 0.8 cm height 1 cm length). The two resulting fresh mango samples, i.e. puree and cubic pieces, were immediately frozen under liquid nitrogen and kept at ‒20°C until the experiments. To obtain the two other mango samples, the two remaining batches of rectangular mango slices were dried in a pilot unit (dryer UTA, Villeneuve-sur-lot, France) at 60 °C and 40% 5
relative humidity under a constant air flow for 9 h (Bonneau et al., 2016). The resulting dried rectangular fruit slices were cut into cubic pieces with a knife (around 0.6 cm width 0.3 cm height 1 cm length). Half of the dried batch was ground to a powder under liquid nitrogen with a ball crusher (Dangoumill 300, Prolabo, Fontenay-sous-Bois, France). The resulting dried mango powder and dried mango cubic pieces were stored at ‒20 °C until the experiments.
2.3. Physicochemical characterisation Mango samples were characterised for their total soluble solids (expressed in °Bx), pH, titratable acidity (TA, expressed in % mEq citric acid), water activity (a w) and dry matter (DM, expressed in %) using the previously described materials and procedure (Bonneau et al., 2016). All measurements were performed in quadruplicate.
2.4. Extraction of volatile compounds Volatile compounds were extracted from the mango samples using the solvent-assisted flavour evaporation (SAFE) technique (Bonneau et al., 2016). Briefly, 15 g of dried mango or 60 g of fresh mango (equivalent weight in DM vs dried mango), 100 mL of ultrapure water and 250 µL of internal standard solution (α-cedrene at 120 µg/mL in methanol) were placed in a 500-mL flask. A magnetic stir bar and PTFE boiling stones were added to regulate the agitation and distillation. The water bath temperature was at 45 °C. The volatile fraction was isolated under high vacuum (10‒3 mbar) and collected in a liquid-nitrogen-cooled flask. The distillation process lasted 90 min when high vacuum was attained. The distillate was recovered, thawed at room temperature and extracted first with 100 mL of solvent mixture (pentane:dichloromethane, 2:1, v/v) and then with two other 50-mL volumes of the same solvent. The organic extracts were combined, dried over anhydrous sodium sulfate,
6
concentrated using a Kuderna-Danish column, and then reduced to around 200 µL under a gentle nitrogen stream. SAFE extraction was conducted in triplicate for each mango sample.
2.5. In vivo aroma trapping 2.5.1. In vivo retronasal aroma trapping device (RATD) The device used for the in vivo experiments was adapted from that used in previous studies (Buettner et al., 2000a; Muñoz-González, Rodríguez-Bencomo, et al., 2014) and included three parts (Fig. 1): (i) an olfactory glass port, where the assessors placed their nose and exhaled through the nasal cavity during mango consumption. Different ports were used according to the assessors’ nose shape; (ii) the olfactory port was connected to an adsorbent polymer conditioned in stainless steel tube (200 mg Tenax TA, 60-80 mesh, N9307054, Perkin Elmer, Norwalk, CT); (iii) the Tenax tube was connected to a vacuum pump at the outlet of the device. To ensure a steady 500 mL/min flow during in vivo aroma trapping, a fine valve and a diaphragm gas flow meter (ADM 2000 flow meter, Agilent Technologies, Wilmington, DE) were connected to the device. 2.5.2. In vivo experiment assessor panel Eight volunteers (3 women and 5 men, with an average of 47 years old) were recruited from our QualiSud Research Unit for the in vivo experiments. They were non-smokers, experienced in sensory analysis and were instructed not to eat and drink (except neutral mineral water) 2 h before the experiments. 2.5.3. Regulation of breathing and chewing with visual animation A computer visual animation consisting of a precise protocol to regulate chewing and nose breathing cycles during consumption was watched by the assessors in agreement with the protocol used in earlier studies (Frank et al., 2011; Frank et al., 2012). It included two
7
phases. A pre-swallow phase (32 sec) consisted of 8 cycles of 2 sec of inhalation combined with 2 chewing strokes and 8 cycles of 2 sec of exhalation combined with 2 chewing strokes. Thereafter the assessor swallowed the sample at once (4 sec). This was followed by a postswallow phase (32 sec) consisting of 8 cycles of 2 sec of inhalation and 2 sec of exhalation as above but without the chewing strokes. Hence one tasting involved 16 chewing and breathing cycles and lasted 68 sec. Similar inhalation and exhalation cycles and pre- and post-swallow phase durations were used in an earlier study related to the effect of the food matrix on in vivo aroma release (Frank et al., 2012). The assessors were trained by watching the visual animation with food matrices similar to mango samples of the present study, i.e. industrial mango puree, canned mango cubic pieces, dried apricot cubic pieces, and dried banana powder. Several training sessions (at least two) were performed until the assessors felt at ease with the animation and tasting procedure. 2.5.4. In vivo RATD extraction and tasting protocol The in vivo experiments were performed from 9:00 am to 12:00 am. For a given mango product, replicates were performed by the assessors in a 2 h session; 3 g dried mango samples (powder or cubic pieces) or 10 g fresh mango samples (puree or cubic pieces), that represented the same dry matter weight as the fresh samples, were consumed in each tasting. The mango sample temperature ranged from 10 °C to 12 °C. Before each tasting, the assessors rinsed their oral cavity according to the following washing protocol: alternative palate cleaning with 4 replicates of 30 mL of mineral water to spit out and 4 replicates of 30 mL of mineral water to swallow. Then they had a 30-min break to achieve an unstimulated oral physiology status. Prior to the mango sample assessment, a blank of exhaled nostril breath of the assessor was collected using RATD under the above visual animation protocol. VOCs trapped by the Tenax tube for 68 sec were then
8
analysed by GC-MS. The assessor then consumed the mango sample under the same conditions using another Tenax tube, RATD device and the visual animation. Volatile compounds released from the mango samples during oral processing were trapped for 68 sec and then analysed by GC-MS. Throughout the in vivo aroma trapping time, the assessors maintained their lips closed during consumption to favour volatile compound equilibrium within the oral cavity (Buettner et al., 2000a). Each session per mango product and per panellist led to 4 Tenax tubes dealing with blanks of assessors’ nose-space and 4 Tenax tubes corresponding to the VOCs released from mango samples. A total of 128 Tenax tubes (8 assessors 4 products 4 replicates) representing VOCs extracted during oral processing of mango samples together with 128 Tenax blank tubes of the assessors were subjected to GCMS analysis. Before each in vivo extraction, the Tenax tubes were conditioned with helium (100 mL/min) at 300 °C for 30 min. To estimate the amount of VOCs in the Tenax tubes, 2 µL of α-cedrene (13 ng/µL in hexane) were applied with a syringe onto the adsorbant polymer just before GC-MS analysis.
2.6. Volatile compound analysis 2.6.1. GC-MS analysis Mango sample aroma extracts obtained by the SAFE technique were analysed with an Agilent 6890 series GC (Agilent Technologies, Santa Clara, CA) equipped with a DB-WAX column (30 m 0.25 mm, 0.25 µm phase film thickness, Agilent J&W) coupled to an Agilent 5973 mass spectrometer detector. One microliter of aroma extract was injected into an on-column injector at 45 °C. The initial oven temperature was 45 °C (5 min), and then it increased by 2 °C/min to 115 °C and then by 10 °C/min to 250 °C. The final column temperature was maintained for 10 min. The mass spectrometer was operated at 70 eV in
9
electron impact (EI) ionisation mode with an m/z 35 to 350 scan range. The carrier gas used was hydrogen at 1.5 mL/min constant flow. VOCs trapped by Tenax tubes in the in vivo experiments were thermodesorbed with a PerkinElmer Turbo Matrix 650 automatic thermal desorber (ATD; PerkinElmer, Norwalk, CT). The ATD was programmed in two steps. First the Tenax tube was heated with helium (100 mL/min) to 280 °C for 15 min. Desorbed VOCs were refocused on a Tenax TA trap maintained at ‒30 °C. The trap was then quickly heated to 280 °C at 99 °C/s and maintained for 10 min. Desorbed VOCs were introduced, through a heated transfer line (deactivated column at 280 °C), to the GC injector (splitless mode) of an Agilent 6890 series GC equipped with a DB-1701 column (60 m 0.25 mm, 0.25 µm phase film thickness, Agilent J&W) coupled to an Agilent 5975 mass spectrometer detector. The GC-MS analysis conditions were the same as above except for the carrier gas (helium at 1.5 mL/min constant flow). 2.6.2. Identification of volatile compounds Volatile compounds were identified through their mass spectra (NIST version 2.0), retention indices (RIs) and via injection of the standards when available. The compound RIs were determined on DB-WAX and DB-1701 columns by linear interpolation following the injection of n-alkanes solution (C8‒C20). 2.6.3. Quantification of VOCs VOCs extracted by the SAFE technique were quantified as equivalent to α-cedrene. A calibration curve was plotted for the in vivo experiments. Standard solutions of α-pinene, βmyrcene, δ-3-carene, γ-terpinene, 3-methylbutyl butanoate, α-gurjunene, camphene, βpinene, α-phellandrene, α-terpinene, limonene, p-cymene, β-ocimene, terpinolene, βcaryophyllene and α-caryophyllene were prepared in hexane to obtain 1, 2, 3, 4, 5, 8, 10, 15, 20 and 25 ng.µL‒1 concentrations. δ-3-Carene was by far the major volatile in Kent mango
10
fruit. Standard solutions of δ-3-carene in hexane were thus prepared at 35, 50, 100, 200, 250 and 350 ng µL‒1 final concentration. These solutions were spiked with α-cedrene (internal standard) solution to yield a 13 ng µL‒1 final concentration. Note that, for the in vivo experiments with the assessors, Tenax tubes were loaded with the same solution (2 µL) of αcedrene just before the GC-MS analysis. Two microlitres of each standard solution spiked with α-cedrene were injected into Tenax tubes with a syringe. The resulting Tenax tubes were analysed by GC-MS, as above. The analyses were conducted in triplicate. The volatile compound quantities were estimated from the calibration curves (r2 > 0.99). When volatile compounds were not available, they were quantified as equivalent to α-cedrene.
2.7. Sensory analysis Ten trained assessors (6 women and 4 men, 7 of whom participated in the in vivo experiments with RATD, with an average age of 47 years old) were selected using the same criteria as in the in vivo experiments and for their experience in sensory analysis and mango product evaluation. The panel was trained with industrial mango products similar to those of the study (industrial mango puree, canned mango cubic pieces, dried mango slices transformed into cubic pieces and powder) purchased at a local market in Montpellier, France. Mango samples for the study and training were defrosted in the refrigerator overnight. They were placed in a glass jar sealed with a cap and presented to the panel at 10‒12 °C, as in in vivo experiments with RATD. Samples were randomly presented in monadic mode with a 3-digit anonymous code. The room temperature and humidity were at 23 °C and 35%, respectively. Two training sessions were conducted to familiarise the assessors with pre-defined odour, texture, taste and flavour attributes of fresh and dried mangoes (Ledeker, Chambers, Chambers IV, & Adhikari, 2012; Ledeker, Suwonsichon, Chambers, & Adhikari, 2014;
11
Suwonsichon, Chambers IV, Kongpensook, & Oupadissakoon, 2012). A separate list of attributes was established for the fresh and dried products. A structured discrete scale from 0 (none) to 10 (high intensity) was used to rate the sensory attributes (Table 1).
(Malundo, Baldwill, Ware, & Shewfelt, 1996; Malundo, Shewfelt, & Ware, 2001; Valente, Ribeyre, Self, Berthiot, & Assemat, 2011)
Each mango product was evaluated under red light in duplicate. Evaluations were performed in two morning sessions from 10:00 am to 12:00 am as for the in vivo experiments with RATD. The assessors were free in their mastication, breathing and swallowing cycles, but they were instructed to rate the flavour attributes during around 1-min consumption. Between tastings, mineral water and neutral crackers were presented to the assessors.
2.8. Statistical analysis All statistical tests were performed using XLSTAT software (version 2015.6.01; Addinsoft, Paris, France). Analyses of variance (ANOVA) were done using Tukey’s honestly significant difference (HSD) at a 95% confidence level for pairwise comparison of fresh and dried samples on their aromatic and sensory profiles. Mann-Whitney non-parametric tests at a 95% confidence level were used for pairwise comparison of the amount of volatile compounds released from fresh and dried samples in the in vivo experiments.
3. Results and Discussion 3.1. Main physicochemical analysis of mango samples The main physicochemical properties of mango samples, i.e. brix (°Bx), pH, titratable acidity (TA), water activity (a w) and dry matter (DM) of fresh and dried mango were similar to previously published values (Pott, Neidhart, Mühlbauer, & Carle, 2005; Tharanathan, Yashoda, & Prabha, 2006) (Table 2). Aw was less than 0.6, which is the level required for the microbiological stability of dried fruits (Pott et al., 2005).
3.2. Aromatic profile of mango samples 12
The SAFE technique was used to determine, as accurately as possible, the whole aroma composition of mango samples. This technique generates an aroma extract close to the aroma composition of the targeted product (Munafo et al., 2014). Fifty volatile compounds were identified in mango samples (Table 3). In agreement with earlier studies (Bonneau et al., 2016; Pino & Mesa, 2006), the aromatic profile of fresh and dried mango samples from cv. Kent were rich in terpene compounds, representing 92.8 % and 84.1% of VOCs, respectively. δ-3-Carene was the predominant terpene followed by α-terpinolene, β-myrcene, limonene and α-pinene. Other chemical classes were present in all products, including non-terpene hydrocarbons, alcohols, lactones and aldehydes. The difference in aromatic composition between fresh and dried samples was due to the loss and generation of compounds during the drying process (Bonneau et al., 2016). Significant differences between fresh and dried mango were observed with respect to the levels of many volatile compounds. Amongst them, monoterpenes, sesquiterpenes, aliphatic alcohols (3-pentanol, 1-penten-3-ol, etc.), some aldehydes ((E)-2-nonenal and (E,Z)-2,6nonadienal) and lactones (γ-hexalactone, δ- and γ-octalactone) were present at higher levels in fresh mango than in dried mango. In contrast, dried mango had higher 3-methyl-1-butanol, hexanal and γ-butyrolactone levels than fresh mango. In addition, new compounds appeared upon drying and were specific to dried samples (2-phenylethanol, 3-methylbutyl butanoate, furan derivatives). Of the 50 VOCs identified 20 could be considered as potential aroma contributors in mango samples. Their odour activity value (OAV), calculated on the basis of compound odour thresholds determined from the literature, was found to be higher than 1. Amongst them, α-pinene, δ-3-carene, α-phellandrene, β-myrcene, limonene, β-ocimene, p-cymene, αterpinolene, β-caryophyllene, 1-penten-3-ol, hexanal, nonanal, δ-octalactone and γoctalactone were present in both fresh and dried mango samples, while camphene, (E)-213
nonenal and (E,Z)-2,6-nonadienal were only found in fresh mango, and heptanal, 3methylbutyl butanoate and γ-butyrolactone were only in dried mango samples. Moreover, the 20 aforementioned compounds were reported to be key flavour compounds in mango according to GC-O analysis (Bonneau et al., 2016; Munafo et al., 2014; Pino, 2012; Pino et al., 2006) (Table 3).
(Bonneau et al., 201 6; M unafo et al., 2 014; Pino, 201 2; Pino et al., 20 06)
3.3. Impact of fruit texture on in vivo aroma release during consumption During consumption of a food besides food matrix, oral processing conditions, i.e. saliva flow, saliva composition, breathing and chewing cycles and their duration, are known to affect volatile compound release (Boland et al., 2006; Buettner et al., 2010; Frank et al., 2011; Muñoz-González, Martin-Alvarez, et al., 2014). Therefore the assessors watched a visual animation monitoring breathing and chewing cycles to reduce intra- and interindividual variability in aroma release. This approach has already been used in previous studies on aroma release during oral processing of foods (Frank et al., 2011; Frank et al., 2012). Since the in vivo nose-space technique applied here was not able to monitor volatile release in real time, as in the PTR-MS technique, VOCs from the assessors’ exhaled breath during both the pre- (32 sec) and post-swallow phase (32 sec) were trapped on Tenax and further identified by GC-MS analysis. Preliminary studies involving separate analysis of VOCs from pre- and post-swallow phases did not show any differences in aroma release in both phases regardless of the mango samples tested but they were assessor dependent. An unstimulated oral physiology status was preferred before each tasting to minimize variability in saliva flow and composition between replicates. In fact, successive tastings without any break-time induced a stimulated oral physiology status that caused variations in saliva flow and composition (Drago, Panouillé, Saint-Eve, Neyraud, Feron, & Souchon, 2011; Engelen, de Wijk, Prinz, van der Bilt, & Bosman, 2003). In addition, prior to each
14
tasting, a blank of assessors’ exhaled nostril breath was performed under the same protocol as that used for in vivo aroma release. Indeed, several volatile compounds (acetone, ethanol, terpenes, etc.) were found in the exhaled nostril breath (Miekisch, Schubert, & NoeldgeSchomburg, 2004; Sanchez & Sacks, 2006). Many volatile compounds were detected in the assessors’ blank tests (mainly acetone, ethanol, α-pinene, β-pinene, limonene), some of which (α-pinene, β-pinene, limonene) were detected in the in-vivo analysis of mango samples. Therefore their amount was subtracted from that obtained with the mango samples. The data reported in Table 4 are averages per compound released and detected by the eight assessors in four replicates per mango product. A total of 21 volatile compounds was detected in the in vivo experiments. They were amongst the main volatile compounds of the fresh and dried mango samples. Seventeen monoterpenes and three sesquiterpenes were detected in all samples, whereas 3-methylbul butanoate was only found in dried samples. Furthermore, among the 20 compounds reported as being potential key aroma compounds in mango products on the basis of the OAV and GC-O data (Bonneau et al., 2016; Munafo et al., 2014; Pino, 2012; Pino et al., 2006), only 11 were detected in the in vivo experiments, i.e. α-pinene, camphene, δ-3-carene, α-phellandrene, β-myrcene, limonene, β-ocimene, pcymene, α-terpinolene, β-caryophyllene and 3-methylbutyl butanoate. With respect to the fresh samples, 83.7 µg kg‒1 DM VOCs from fresh cubic pieces was detected in the nose-space, while 27.0 µg kg‒1 (DM) from fresh puree was detected. The ANOVA findings revealed that 17 terpenes were significantly released at higher levels from fresh cubic pieces than from fresh puree. Similarly, in the case of the dried samples, higher levels of VOCs were emitted from dried cubic pieces than from dried powder, i.e. 55.6 µg kg‒1 DM compared to 20.5 µg kg‒1
15
DM. Moreover, 12 terpenes were significantly released at higher levels from dried cubic pieces than from dried powder. The above results indicate that the intact fruit samples (fresh and dried cubic pieces) released more aroma compounds than the disintegrated samples (fresh puree, dried powder). This could be attributed to the higher surface area of intact samples in oral processing than that of the disintegrated samples, which tend to aggregate under mastication. To compare our data with those obtained in previous studies where the PTR-MS technique was applied to assess several real foods (Frank et al., 2012), the area under the curve (AUC) of the VOCs in PTR-MS analysis from pre- and post-sallow phases was taken into consideration. Note that these authors applied a controlled protocol during in vivo tasting similar to that used in the present study. The same trends were observed. VOCs released including pre- (30 sec) and post-swallow (30 sec) phases showed that intact food (pieces of orange, chocolate and peanut) generally released more aroma compounds than disintegrated food (orange jelly and chocolate dessert). These data are supported by the findings on a study on pectin and gelatin gels supplemented with strawberry flavour compounds (Boland et al., 2006). VOCs release based on AUC values during PTR-MS monitored consumption showed higher volatile release when the rigidity of the model fruit was increased by the addition of pectin or gelatin. In a recent study, in vivo PTR-MS analysis revealed that firm apple cultivars release more VOCs based on AUC values than soft cultivars (Ting et al., 2015; Ting et al., 2016). However these data could not be compared with the present data since the aroma composition of the apple cultivars differed.
3.4. Impact of fruit texture on sensory perception
16
A total of 15 selected sensory attributes were chosen to describe the sensory profile of the fresh and dried mango samples. The sensory panel (n = 10) showed good repeatability and homogeneous scores between subjects. For the fresh samples, six attributes related with texture, taste and flavour were significantly different between the two fresh products (Fig. 2a). Conversely, there were no significant differences between these two products regarding odour attributes. Fresh cubic pieces were rated as being more firm, with more particles and fibres than fresh puree. Fresh puree was rated more sweet and sour than fresh cubic pieces, which could be attributed to a higher availability of sweet and sour components for the tongue. Mango and spicy flavour attributes were more perceived in fresh cubic pieces than in fresh puree, in agreement with a previous study involving a comparison of fresh mango pieces (cv. Kent) and mango puree by a sensory panel (Ledeker et al., 2012). Interestingly, there seemed a relationship between the flavour perception of the sensory panel and the in vivo aroma release. VOCs released in the in vivo experiments imparting mango-like (α-pinene, δ-3-carene, α-phellandrene, β-myrcene, limonene, β-ocimene, p-cymene and α-terpinolene) and spicy notes (α-phellandrene, pcymene and β-caryophyllene), were detected at higher levels in fresh cubic pieces, i.e. intact fruits, than in fresh puree. Regarding the dried samples, all attributes were rated as significantly higher in dried powder than in dried cubic pieces, except for the mango flavour attribute, which was similarly perceived in both products (Fig. 2b). Dried powder developed more fruity, spicy and cooked notes than dried cubic pieces. This was out of line with the in vivo experimental results where the dried cubic pieces released more VOCs than dried powder. A plausible explanation for this difference could be that the assessors were not controlled with regard to the breathing, chewing cycles and swallowing during the sensory evaluation. Consequently,
17
the chewing of dried mango pieces might have taken less or more time than in the in vivo aroma release experiments. Sweet and sour taste attributes were rated higher in the dried powder than in the dried cubic pieces. The taste perception (sweetness and sourness) was thus clearly linked to the matrix structure since it was higher in the disintegrated mango samples (fresh puree, dried powder) compared to the intact mango samples (fresh and dried cubic pieces). In this regard, it has been shown that an increased thickness in pectin gels was inversely related to sweet perception (Boland et al., 2006). In addition, flavour perception could also be impacted by food components, sugars, organic acids, macromolecules (polysaccharides, lipids, polyphenols, etc.) (Guichard, 2002; Tournier, Sulmont-Rossé, & Guichard, 2007). Taste components are known to be flavour enhancers. For example, the sugar level was found to be correlated with sweet and peachy flavour attributes in fresh mango (Malundo, Shewfelt, Ware, et al., 2001). Consequently, higher flavour perception in dried mango powder than in the dried cubic pieces was presumably due to greater sugar accessibility for the palate in the powdered mango.
4. Conclusion The current study highlighted the impact of fruit texture on the release and perception of aroma compounds during in vivo consumption. Although the RATD-based experimental methodology was unable to monitor VOC release in real-time in vivo experiments, it enabled us to draw general conclusions matching those reported with PTR-MS on real foods. The implemented methodology was quite simple and also enabled us to identify VOCs at low levels because of the several seconds of trapping during oral processing. Further studies could help to gain further insight into the effect of various factors (sugars, pectins, saliva, etc.) on the release of flavour compounds and flavour perception using a 18
model matrix with varying texture and supplemented with potential key flavour compounds from mango.
Acknowledgements We are very grateful to Christophe Bugaud from the QualiSud Research Unit and Xavier Bry from Montpellier University for their support in the statistical analysis and all the assessors from the QualiSud Research Unit involved in the in vivo experiments.
19
References Aprea, E., Biasioli, F., Gasperi, F., Märk, T. D., & van Ruth, S. (2006). In vivo monitoring of strawberry flavour release from model custards: effect of texture and oral processing. Flavour and Fragrance Journal, 21(1), 53-58. Boland, A. B., Buhr, K., Giannouli, P., & van Ruth, S. M. (2004). Influence of gelatin, starch, pectin and artificial saliva on the release of 11 flavour compounds from model gel systems. Food Chemistry, 86(3), 401-411. Boland, A. B., Delahunty, C. M., & van Ruth, S. M. (2006). Influence of the texture of gelatin gels and pectin gels on strawberry flavour release and perception. Food Chemistry, 96(3), 452-460. Bonneau, A., Boulanger, R., Lebrun, M., Maraval, I., & Gunata, Z. (2016). Aroma compounds in fresh and dried mango fruit (Mangifera indica L. cv. Kent): impact of drying on volatile composition. International Journal of Food Science and Technology, 51(3), 789-800. Buettner, A., & Beauchamp, J. (2010). Chemical input – Sensory output: Diverse modes of physiology–flavour interaction. Food Quality and Preference, 21(8), 915-924. Buettner, A., & Schieberle, P. (2000a). Exhaled Odorant Measurement (EXOM) — A New Approach to Quantify the Degree of In-mouth Release of Food Aroma Compounds. LWT - Food Science and Technology, 33(8), 553-559. Buettner, A., & Schieberle, P. (2000b). Influence of mastication on the concentrations of aroma volatiles — some aspects of flavour release and flavour perception. Food Chemistry, 71(3), 347-354. Delahunty, C. M., Piggott, J. R., Conner, J. M., & Paterson, A. (1996). Comparison of dynamic flavour release from hard cheeses and analysis of headspace volatiles from the mouth with flavour perception during consumption. Journal of the Science of Food and Agriculture, 71(3), 273-281. Drago, S. R., Panouillé, M., Saint-Eve, A., Neyraud, E., Feron, G., & Souchon, I. (2011). Relationships between saliva and food bolus properties from model dairy products. Food Hydrocolloids, 25(4), 659-667. Engelen, L., de Wijk, R. A., Prinz, J. F., van der Bilt, A., & Bosman, F. (2003). The relation between saliva flow after different stimulations and the perception of flavor and texture attributes in custard desserts. Physiology & Behavior, 78(1), 165-169. Feron, G., Ayed, C., Qannari, E. M., Courcoux, P., Laboure, H., & Guichard, E. (2014). Understanding aroma release from model cheeses by a statistical multiblock approach on oral processing. PloS one, 9(4), e93113. Frank, D., Appelqvist, I., Piyasiri, U., Wooster, T. J., & Delahunty, C. (2011). Proton transfer reaction mass spectrometry and time intensity perceptual measurement of flavor release from lipid emulsions using trained human subjects. Journal of Agricultural and Food Chemistry, 59(9), 4891-4903. Frank, D., Eyres, G. T., Piyasiri, U., & Delahunty, C. M. (2012). Effect of food matrix structure and composition on aroma release during oral processing usingin vivomonitoring. Flavour and Fragrance Journal, 27(6), 433-444. Guichard, E. (2002). Interactions between flavor compounds and food ingredients and their influence on flavor perception. Food Reviews International, 18(1), 49-70. Ledeker, C. N., Chambers, D. H., Chambers IV, E., & Adhikari, K. (2012). Changes in the sensory characteristics of mango cultivars during the production of mango puree and sorbet. Journal of Food Science, 77(10), S348-355. 20
Ledeker, C. N., Suwonsichon, S., Chambers, D. H., & Adhikari, K. (2014). Comparison of sensory attributes in fresh mangoes and heat-treated mango purées prepared from Thai cultivars. LWT - Food Science and Technology, 56(1), 138-144. Lubbers, S., & Guichard, E. (2003). The effects of sugars and pectin on flavour release from a fruit pastille model system. Food Chemistry, 81(2), 269-273. Malundo, T. M. M., Baldwill, E. A., Ware, G. O., & Shewfelt, R. L. (1996). Volatile composition and interaction influence flavor properties of mango (Mangifera indica L.). In Proceedings-Florida State Horticultural Society, vol. 109 (pp. 264-268): Florida State Horticultural Society. Malundo, T. M. M., Shewfelt, R. L., & Ware, G. O. (2001). An alternative method for relating consumer and descriptive data used to identify critical flavor properties of mango (Mangifera Indica L.). Journal of sensory studies, 16(2), 199-214. Malundo, T. M. M., Shewfelt, R. L., Ware, G. O., & Baldwin, E. A. (2001). Sugars and Acids Influence Flavor Properties of Mango (Mangifera indica). Journal of the American Society for Horticultural Science, 126(1), 115-121. Marsh, K. B., Friel, E. N., Gunson, A., Lund, C., & MacRae, E. (2005). Perception of flavour in standardised fruit pulps with additions of acids or sugars. Food Quality and Preference, 17(5), 376-386. Miekisch, W., Schubert, J. K., & Noeldge-Schomburg, G. F. E. (2004). Diagnostic potential of breath analysis - focus on volatile organic compounds. Clinica Chimica Acta, 347(1–2), 25-39. Munafo, J. P., Jr., Didzbalis, J., Schnell, R. J., Schieberle, P., & Steinhaus, M. (2014). Characterization of the major aroma-active compounds in mango (Mangifera indica L.) cultivars Haden, White Alfonso, Praya Sowoy, Royal Special, and Malindi by application of a comparative aroma extract dilution analysis. Journal of Agricultural and Food Chemistry, 62(20), 4544-4551. Muñoz-González, C., Martin-Alvarez, P. J., Moreno-Arribas, M. V., & Pozo-Bayon, M. A. (2014). Impact of the nonvolatile wine matrix composition on the in vivo aroma release from wines. Journal of Agricultural and Food Chemistry, 62(1), 66-73. Muñoz-González, C., Rodríguez-Bencomo, J. J., Moreno-Arribas, M. V., & Pozo-Bayón, M. Á. (2014). Feasibility and application of a retronasal aroma-trapping device to study in vivo aroma release during the consumption of model wine-derived beverages. Food Science & Nutrition, 2(4), 361-370. Pino, J. A. (2012). Odour-active compounds in mango (Mangifera indica L. cv. Corazón). International Journal of Food Science and Technology, 47(9), 1944-1950. Pino, J. A., & Mesa, J. (2006). Contribution of volatile compounds to mango (Mangifera indica L.) aroma. Flavour and Fragrance Journal, 21(2), 207-213. Pino, J. A., Mesa, J., Munoz, Y., Marti, M. P., & Marbot, R. (2005). Volatile components from mango (Mangifera indica L.) cultivars. Journal of Agricultural and Food Chemistry, 53(6), 2213-2223. Pionnier, E., Sémon, E., Chabanet, C., & Salles, C. (2005). Évaluation de la technique de microextraction sur phase solide (SPME) pour l'analyse de l'air humain exhalé pendant la consommation d'aliments. Sciences des Aliments, 25(3), 193-206. Poinot, P., Arvisenet, G., Ledauphin, J., Gaillard, J.-L., & Prost, C. (2013). How can aroma– related cross–modal interactions be analysed? A review of current methodologies. Food Quality and Preference, 28(1), 304-316.
21
Pott, I., Neidhart, S., Mühlbauer, W., & Carle, R. (2005). Quality improvement of nonsulphited mango slices by drying at high temperatures. Innovative Food Science & Emerging Technologies, 6(4), 412-419. Sanchez, J. M., & Sacks, R. D. (2006). Development of a multibed sorption trap, comprehensive two-dimensional gas chromatography, and time-of-flight mass spectrometry system for the analysis of volatile organic compounds in human breath. Analytical chemistry, 78(9), 3046-3054. Suwonsichon, S., Chambers IV, E., Kongpensook, V., & Oupadissakoon, C. (2012). Sensory Lexicon for Mango as Affected by Cultivars and Stages of Ripeness. Journal of sensory studies, 27(3), 148-160. Tharanathan, R. N., Yashoda, H. M., & Prabha, T. N. (2006). Mango (Mangifera indica L.), “The King of Fruits”—An Overview. Food Reviews International, 22(2), 95-123. Ting, V. J. L., Romano, A., Silcock, P., Bremer, P. J., Corollaro, M. L., Soukoulis, C., Cappellin, L., Gasperi, F., & Biasioli, F. (2015). Apple flavor: Linking sensory perception to volatile release and textural properties. Journal of Sensory Studies, 30(3), 195-210. Ting, V. J. L., Romano, A., Soukoulis, C., Silcock, P., Bremer, P. J., Cappellin, L., & Biasioli, F. (2016). Investigating the in-vitro and in-vivo flavour release from 21 fresh-cut apples. Food Chemistry, 212, 543-551. Tournier, C., Sulmont-Rossé, C., & Guichard, E. (2007). Flavour perception: aroma, taste and texture interactions. Food, 1(2), 246-257. Valente, M., Ribeyre, F., Self, G., Berthiot, L., & Assemat, S. (2011). Instrumental and sensory characterization of mango fruit texture. Journal of Food Quality, 34(6), 413424. Van Ruth, S. M., & Roozen, J. P. (2000). Influence of mastication and saliva on aroma release in a model mouth system. Food Chemistry, 71(3), 339-345. Zardin, E., Tyapkova, O., Buettner, A., & Beauchamp, J. (2014). Performance assessment of proton-transfer-reaction time-of-flight mass spectrometry (PTR-TOF-MS) for analysis of isobaric compounds in food-flavour applications. LWT - Food Science and Technology, 56(1), 153-160.
22
FIGURE CAPTIONS Fig. 1. RATD used for the in vivo aroma trapping experiments. Fig. 2. Sensory analysis of mango samples: (a) fresh puree and fresh cubic pieces; (b) dried powder and dried cubic pieces.
23
TABLE CAPTIONS Table 1. List of sensory attributes for mango products. Table 2. Physicochemical properties of mango products. Table 3. Aroma compounds in fresh and dried mango samples determined by SAFE-GC/MS. Table 4. Aroma compounds trapped from the exhaled nostril of assessors in the in vivo experiments (eight assessors, four tastings) determined by RATD-GC/MS.
24
Table 1. Sensory attributes Odour
Intensity of overall odour Intensity of overall fruity odour
Intensity of overall cooked odour
Texture
Firmness Particle/Fibre amount Viscosity
Sticky Taste
Sweet Sour
Flavour
Intensity of overall aroma Mango identity Fruity Spicy
Cooked Aroma persistence in the mouth
Definitions* Fresh samples Dried samples (puree, cubic pieces) (powder, cubic pieces) Intensity of odour with the contribution of all odour notes Intensity of fruity odour with the contribution of all odour notes associated with fruits Intensity of cooked odour with the contribution of all odour notes associated with caramel notes, cooked, crystallised, dried, caramelised fruit notes The force required to chew and deform the sample Amount of residues (fibres or particles) in the sample The force required to move and disperse the sample in the mouth The force required to remove the sample sticking to the mouth or teeth The fundamental taste sensation produced by compounds such as sucrose The fundamental taste sensation produced by compounds such as citric acid Intensity of aroma with the contribution of all aromatic notes Contribution of sweet, fruity, green, turpentine aromatic notes (typical of mango flavour) A fruity aromatic note associated with all fruits except mango A spicy aromatic note associated with cumin, clove, black pepper, liquorice A cooked aromatic note associated with caramel, cooked, crystallised, dried, caramelised fruits aroma perception in the mouth following swallowing
*Definition of sensory attributes in reference to the literature: Ledeker et al. (2012); Ledeker et al. (2014); Malundo, Baldwill, Ware, and Shewfelt (1996); Malundo, Shewfelt, and Ware (2001); Suwonsichon, Chambers IV, Kongpensook, and Oupadissakoon (2012); Valente, Ribeyre, Self, Berthiot, and Assemat (2011); NF ISO 5492 (AFNOR, Association Française de Normalisation, Analyse sensorielle, Vocabulaire).
25
Table 2. °Bxa
pHb
TAc
awd
DMe
18.6 ± 0.2
4.0 ± 0.1
0.7 ± 0.1
0.98 ± 0.004
20.6 ± 0.2
81.3 ± 1.4
4.1 ± 0.1
2.8 ± 0.2
0.56 ± 0.008
90.1 ± 1.2
Mango products Fresh samples (puree, cubic pieces) Dried samples (powder, cubic pieces) a
Total soluble solids (°Bx) ± 0.1°Bx
b
pH ± 0.1
c
Titratable acidity (TA) was expressed in % mEq of citric acid.
d
Water activity (aw) was expressed unitless ± 0.0001 at 25°C
e
Dry matter (DM) was expressed in %
26
Table 3.
Compounds Monoterpene hydrocarbons α-pinene† α-fenchene camphene† β-pinene δ-3-carene† α-phellandrene† β-myrcene† α-terpinene sylvestrene limonene† β-phellandrene γ-terpinene β-ocimene† p-cymene† unknown monoterpene§ α-terpinolene† p-cymenene Total Sesquiterpene hydrocarbons α-gurjunene β-caryophyllene† α-caryophyllene Total Norisoprenoid β-ionone Non-terpene hydrocarbons toluene p-xylene m-xylene o-xylene Total Alcohols 3-pentanol 2-pentanol 1-penten-3-ol† 3-methyl-1-butanol$ & 2-methyl-1-butanol$ 1-pentanol 2-penten-1-ol (isomer) 1-hexanol (Z)-3-hexen-1-ol 1-octanol 2-phenylethanol Total Aldehydes hexanal† heptanal†
RI DB-WAXa
RI DB-1701a
IDb
Relative quantities (µg kg‒1 DM)c Fresh Dried samples samples
Significanced
Exp
Lit
Exp
Lit
1027 1056 1069 1108 1147 1159 1163 1174 1192 1194 1202 1234 1243 1252 1260 1265 1433
1032 1054 1075 1113 1148 1166 1150 1178 1177 1188 1209 1238 1242 1257 1275 1426
946 965 967 999 1034 1034 1021 1045 1051 1055 1062 1096 1083 1072 1114 1117 1153
943 n.f. 970 999 1024 1036 1016 1036 n.f. 1065 1059 1090 1075 1073 1113 n.f.
A B A A A A A A B A A A A A C A B
429 12.0 3.30 25.7 21150 122 597 56.7 30.4 530 194 12.5 9.84 51.8 74.0 845 5.52 24149
160 3.33 t 12.0 7583 31.1 153 18.4 10.4 163 49.1 4.59 t 12.6 21.4 254 3.15 8479
** *** ** ** ** ** ** ** ** ** ** ** ** ** ** ** **
1507 1579 1639
1529 1570 1640
1448 1480 1522
1471 1514
A A A
17.7 90.7 127 235
17.1 75.5 93.6 186
n.s. n.s. *
1900
1912
1621
1622
A
t
t
n.s.
1037 1120 1129 1172
1042 1127 1138 1182
816 911 947 919
817 919 n.f. 919
A B B A
779 11.2 22.6 13.3 826
564 7.04 15.0 7.94 594
n.s. * ** *
1112 1123 1158
1110 1117 1157
796 799 -
n.f. 806 784
167 438 30.9
125 315 11.9
* * *
1208
1206
858
863
t
86.7
***
1246 1314 1369 1404 1569 1876
1244 1316 1354 1401 1557 1893
901 878 990 971 1050 1654
891 888 985 970 n.f. 1645
B A B A A B B A A A A
26.2 30.9 19.9 118 6.99 n.d. 838
14.6 11.3 8.48 3.37 4.30 4.79 585
** * ** ** n.s. **
1087 1190
1075 1183
872 974
887 982
A A
t n.d.
5.33 15.9
* *
27
nonanal† (E)-2-nonenal† (E,Z)-2,6-nonadienal† Total Ketone 3-hydroxy-2-butanone Ester 3-methylbutyl butanoate† Furans 2-pentylfuran 2-furfural 5-methylfurfural Total Lactones γ-butyrolactone† γ-hexalactone δ-octalactone† γ-octalactone† Total Total amounts a
1409 1517 1573
1392 1519 1597
1194 1283 1284
1189 1282 1269
A B A
22.5 14.2 68.0 105
19.3 t t 40.5
n.s. *** **
1258
1272
845
n.f.
A
t
89.0
**
1258
1261
1119
n.f.
A
n.d.
52.2
***
1226 1450 1550
1231 1456 1560
1034 970 1109
1030 970 1116
B A A
n.d. n.d. n.d. n.d.
t 21.6 t 21.6
n.s. *** n.s.
1591 1653 sup 1868
1619 1694 1964 1883
1128 1272 1661 1486
1128 n.f. n.f. 1475
A A A B
28.8 60.0 17.2 12.5 119 26271
238 5.89 8.14 t 252 10300
*** ** ** **
Retention index of aroma compounds (RI) from databases (Flavornet; Pherobase; The LRI and Odour
Database). b
ID identification (A) mass spectra database from NIST (NIST version 2.0, 2011), RI and injection of standard
(positive identification), (B) mass spectrum, RI (tentative identification), (C) only mass spectrum. c
Quantification of aroma compounds as equivalent to internal standard (α-cedrene) and expressed in µg per kg
of dry matter (µg.kg-1 DM) in fresh and dried mango samples. d
Significance: analysis of variance (ANOVA) with the honestly significant difference (HSD) in a Tukey test at
different significance levels (XLSTAT software): (n.s.) no significant difference between dried and fresh samples (α > 5%); (*) significant difference with α < 5%; (**) significant difference with α < 1%; (***) significant difference with α < 0.1%. Symbols: ($) coelution of aroma compound; () compound was also observed in in vivo experiments RATD; (†) compound was reported as a key aroma compound in mango flavour (Bonneau et al., 2016; Munafo et al., 2014; Pino, 2012; Pino et al., 2006);
(§) unknown monoterpene compound with specific ions (m/z
91/93/105/121/136); (t) aroma compound in trace amount (t < 3.0 µg/kg DM); (n.d.) aroma compound was not detected; (n.f.) data was not available in the literature; (-) data was missing.
28
Table 4.
Compound
Monoterpene hydrocarbons α-pinene† camphene† α-fenchene$ β-pinene β-myrcene† δ-3-carene† & α-phellandrene† α-terpinene limonene† sylvestrene$ β-phellandrene$ p-cymene† β-ocimene† γ-terpinene unknown monoterpene§ α-terpinolene† p-cymenene Sesquiterpene hydrocarbons α-gurjunene β-caryophyllene† α-caryophyllene Ester 3-methylbutyl butanoate† Total amounts a
Amount of aroma compounds released (ng kg‒1 DM)a Fresh mango products Dried mango products Fresh Dried Fresh Dried b cubic Significance cubic Significanceb puree powder pieces pieces
492 15.8 43.2 74.6 842
1980 38.6 148 121 4322
*** n.s. ** n.s. ***
291 10.3 22.6 47 579
1286 38.5 106 96 2765
*** *** *** n.s. *
23116
65731
***
17240
43070
*
42.1 750 105 229 259 93.0 48.7 85.5 548 191
268 3028 590 1149 810 187 168 493 3453 817
** ** *** *** * n.s. ** *** *** ***
37.9 564 75 174 133 39.0 59.1 94.7 539 189
207 1971 398 666 681 143 109 334 2143 657
** * n.s. * ** ** n.s. n.s. n.s. n.s.
t 41.4 11.2
62.8 242 121
*** *** ***
35.4 91.4 24.5
77.9 302 136
n.s. n.s. **
26988
83729
***
293 20539
389 55575
n.s. *
Average amount of aroma compounds released in in vivo experiments during consumption of mango samples
by eight assessors in four replicates. Quantity expressed in ng per kg of dry matter (ng kg‒1 DM) in fresh and dried mango samples. b
Significance: non-parametric tests of Mann-Whitney were used to compare mango products at different
significance levels (XLSTAT software): (n.s.) no significant difference between mango products (α > 5%); (*) significant difference with α < 5%; (**) significant difference with α < 1%; (***) significant difference with α < 0 .1%. Symbols: ($) quantification of some aroma compounds as equivalent to internal standard (α-cedrene) and expressed in ng per kg of dry matter (ng kg‒1 DM) in fresh and dried mango samples; () compound δ-3-carene coeluted with α-phellandrene that represented less than 1% of signal response of δ-3-carene; (†) compound was
29
reported as a key aroma compound in mango flavour (Bonneau et al., 2016; Munafo et al., 2014; Pino, 2012; Pino et al., 2006); (§) unknown monoterpene compound with specific ions (m/z 91/93/105/121/136); (t) aroma compound was detected in trace quantity (t < 10 ng kg‒1 DM).
30
Highlights Impact of fruit texture on the in vivo aroma release was determined by RATD-GC/MS. Fruit texture had a significant effect on the in vivo release and perception of VOCs. The intact mango samples released more VOCs in vivo than the disintegrated samples.
31
32