Scholarly article on topic 'Volatile aroma components and MS-based electronic nose profiles of dogfruit ( Pithecellobium jiringa ) and stink bean ( Parkia speciosa )'

Volatile aroma components and MS-based electronic nose profiles of dogfruit ( Pithecellobium jiringa ) and stink bean ( Parkia speciosa ) Academic research paper on "Chemical sciences"

CC BY-NC-ND
0
0
Share paper
Academic journal
Journal of Advanced Research
OECD Field of science
Keywords
{"Volatile aroma components" / "MS-based electronic nose" / Dogfruit / "Stink bean" / "Ripening stage"}

Abstract of research paper on Chemical sciences, author of scientific article — Yonathan Asikin, Kusumiyati, Takeshi Shikanai, Koji Wada

Abstract Dogfruit (Pithecellobium jiringa) and stink bean (Parkia speciosa) are two typical smelly legumes from Southeast Asia that are widely used in the cuisines of this region. Headspace/gas chromatography/flame ionization detection analysis and mass spectrometry (MS)-based electronic nose techniques were applied to monitor ripening changes in the volatile flavor profiles of dogfruit and stink bean. Compositional analysis showed that the ripening process greatly influenced the composition and content of the volatile aroma profiles of these two smelly food materials, particularly their alcohol, aldehyde, and sulfur components. The quantity of predominant hexanal in stink bean significantly declined (P < 0.05) during the ripening process, whereas the major volatile components of dogfruit changed from 3-methylbutanal and methanol in the unripe state to acetaldehyde and ethanol in the ripe bean. Moreover, the amount of the typical volatile flavor compound 1,2,4-trithiolane significantly increased (P < 0.05) in both ripened dogfruit and stink bean from 1.70 and 0.93%, to relative amounts of 19.97 and 13.66%, respectively. MS-based nose profiling gave further detailed differentiation of the volatile profiles of dogfruit and stink bean of various ripening stages through multivariate statistical analysis, and provided discriminant ion masses, such as m/z 41, 43, 58, 78, and 124, as valuable “digital fingerprint” dataset that can be used for fast flavor monitoring of smelly food resources.

Academic research paper on topic "Volatile aroma components and MS-based electronic nose profiles of dogfruit ( Pithecellobium jiringa ) and stink bean ( Parkia speciosa )"

Journal of Advanced Research xxx (2017) xxx-xxx

Contents lists available at ScienceDirect

Journal of Advanced Research

journal homepage: www.elsevier.com/locate/jare

a Department of Bioscience and Biotechnology, Faculty of Agriculture, University of the Ryukyus, 1 Senbaru, Nishihara, Okinawa 903-0213, Japan b Faculty of Agriculture, Padjadjaran University, Jalan Raya Bandung-Sumedang KM 21, Jatinangor, West Java 45363, Indonesia c Department of Regional Agricultural Engineering, Faculty of Agriculture, University of the Ryukyus, 1 Senbaru, Nishihara, Okinawa 903-0213, Japan

Original Article

Volatile aroma components and MS-based electronic nose profiles of dogfruit (Pithecellobium jiringa) and stink bean (Parkia speciosa)

Yonathan Asikina'*, Kusumiyatib, Takeshi Shikanaic, Koji Wadaa

GRAPHICAL ABSTRACT

ARTICLE INFO

ABSTRACT

Article history:

Received 12 September 2017 Revised 11 November 2017 Accepted 11 November 2017 Available online xxxx

Keywords:

Volatile aroma components MS-based electronic nose Dogfruit Stink bean Ripening stage

Dogfruit (Pithecellobium jiringa) and stink bean (Parkia speciosa) are two typical smelly legumes from Southeast Asia that are widely used in the cuisines of this region. Headspace/gas chromatography/flame ionization detection analysis and mass spectrometry (MS)-based electronic nose techniques were applied to monitor ripening changes in the volatile flavor profiles of dogfruit and stink bean. Compositional analysis showed that the ripening process greatly influenced the composition and content of the volatile aroma profiles of these two smelly food materials, particularly their alcohol, aldehyde, and sulfur components. The quantity of predominant hexanal in stink bean significantly declined (P < 0.05) during the ripening process, whereas the major volatile components of dogfruit changed from 3-methylbutanal and methanol in the unripe state to acetaldehyde and ethanol in the ripe bean. Moreover, the amount of the typical volatile flavor compound 1,2,4-trithiolane significantly increased (P < 0.05) in both ripened dogfruit and stink bean from 1.70 and 0.93%, to relative amounts of 19.97 and 13.66%, respectively. MS-based nose profiling gave further detailed differentiation of the volatile profiles of dogfruit and stink bean of various ripening stages through multivariate statistical analysis, and provided discriminant ion masses, such as m/z 41,43, 58, 78, and 124, as valuable ''digital fingerprint" dataset that can be used for fast flavor monitoring of smelly food resources.

© 2017 Production and hosting by Elsevier B.V. on behalf of Cairo University. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Peer review under responsibility of Cairo University.

* Corresponding author. E-mail address: yonathan.asikin@gmail.com (Y. Asikin).

https://doi.org/10.1016/jjare.2017.11.003

2090-1232/® 2017 Production and hosting by Elsevier B.V. on behalf of Cairo University.

This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

2 Y. Asikin et al./Journal of Advanced Research xxx (2017) xxx-xxx

Introduction

Dogfruit (Pithecellobium jiringa) and stink bean (Parkia speciosa) are popular smelly legumes from Southeast Asia that possess unpleasant aroma characteristics but that are commonly consumed in various local cooked dishes [1-3]. Dogfruit derives from the Mimosa family (Mimosaceae). It has round-flattened, horse chestnut bean shape and grow in large dark purple pods [1]. On the other hand, stink bean which belongs to pea or bean family (Fabaceae), is formed in dry, longitudinal dehiscent, straight or twisted green pods [4]. Dogfruit and stink bean are commercially available in the markets most of the year and are known under different local names across the region: dogfruit is called as jengkol, jering, krakos, yiniking, niang-yai, and ma-niang, whereas stink bean is also known as smelly bean, petai, sataw, sotor, chou-dou, and u'pang. The unfavorable aspects of these beans are their anti-nutritional components and toxicities if they are excessively consumed or improperly cooked, and in some severe cases, these undesirable properties can cause acute and chronic health effects [1,5,6]. On the other hand, the beans contain various bioactive compounds that possess potent beneficial functionalities, for example, the antifungal and antibacterial activities of dogfruit lec-tins and the antidiabetic and antihypertensive potentials of stink bean sterols and peptides, respectively [7-9]. In spite of the drawbacks, dogfruit and stink bean are regarded as regional delicacies, and these food resources have been used as raw materials in the production of various valuable semi-processed or processed food products, such as flours and cookies [3,10,11].

Agricultural crops, including those of legumes, are distinguishable, not only by their primary appearance or physico-chemical traits but also by other important quality attributes, such as sensory perception [12,13]. Moreover, ripening makes critical biochemical contributions to the metabolite development of volatile constituents and other nutritional components of horticultural products that might differentiate their potential food applications [14,15]. The alteration of volatile aroma components, particularly, has an important direct effect on the appeal of raw or cooked foods, as a whole or indirectly, by influencing other flavor properties and thresholds [13,15,16]. Consequently, maturity could be used as a potent indicator for the progression of volatile aroma composition and flavor characteristics in agricultural crops, which might lead to a distinction in their perceived aroma and consumer acceptance [17,18].

Numerous innovative analytical techniques have been developed to complement the use of conservative methods with common analytical instruments for evaluating food quality traits [12,19,20]. The improved analytical approaches include reliable techniques for both qualitative and quantitative measurements, and they are most often combined with robust chemometric statistical analysis to discriminate samples. Electronic nose measurement technologies, such as gas sensor arrays, fast gas chro-matography (GC), and mass spectrometry (MS), have also been effectively used for distinguishing the volatile flavor profiles of various food resources and products [20-22]. The MS-based electronic nose is a non-targeted volatile-profiling technique for differentiating evaluated samples without a chromatography peak separation requirement. This profiling technique works based on the selection of ion masses needed for statistical analysis by pattern-recognition learning methods, and it can display discriminant ion masses of samples' volatile components as valuable ''digital fingerprints" [19,21,23].

Therefore, the aim of this study was to determine the volatile aroma components of dogfruit and stink bean of different ripening stages and to differentiate their volatile profiles through compositional and MS-based nose datasets (Fig. 1). The volatile

constituents of dogfruit and stink bean were examined by using GC with flame ionization detection (GC-FID), and the volatile characteristics were discriminated by using MS-based electronic nose and chemometric analyses. This is the first report on the volatile and MS-based nose profiles of these two smelly plant resources at different stages of maturity.

Methods

Sample preparation and standards

Fresh samples of two dogfruits (unripe and ripe) and three stink beans (unripe, mid-ripe, and ripe), which originated from the same farming source, were collected from a local market at Bandung, Indonesia, in July 2013. The plant species were authenticated by Dr. Kusumiyati (Laboratory of Horticulture, Faculty of Agriculture, Padjadjaran University), in terms of the perceived visual and physical properties of entire pods and beans. Bean type was morphologically characterized for average weight, size, and color (Table 1 and Fig. 2). The dogfruits and stink beans were peeled from their pods and shells, and the beans were cut into small pieces (about 5 mm2) and stored at -30 °C prior to analysis. Authentic standards (carbon disulfide, dimethyl sulfide, dimethyl disulfide, dimethyl trisulfide, acetaldehyde, propanal, 2-methylpropanal, butanal, 2-methylbutanal, 3-methylbutanal, pentanal, hexanal, heptanal, 2-hexenal, octanal, 2-heptenal, nonanal, 2-octenal, benzaldehyde, 2-nonenal, methanol, ethanol, 3-methylbutanol, pentanol, hexanol, octane, acetone, 2-pentanone, ethyl acetate, hexyl acetate, acetic acid, and hexanoic acid) used for the identification of volatile aroma components were purchased from Sigma-Aldrich (St Louis, MO, USA) and Tokyo Chemical Industry (Tokyo, Japan).

Volatile aroma composition analysis

The composition of the volatile aroma components of dogfruit and stink bean were examined by using an Agilent 7890A GC-FID system equipped with an Agilent G1888 headspace sampler and a fused silica capillary DB-Wax column (60 m x 0.25 mm internal dimensions, 0.25 im film thickness, Agilent J&W, Santa Clara, CA, USA) [24]. The volatile aroma compounds were extracted from a 2 g sample, which was placed in a 20 mL headspace vial, at 80 °C for 20 min, and subsequently pressurized at 11 psi for 0.3 min into the injection port. The sample loop and transfer line were set at 170 and 210 °C, respectively. The injector and FID were both programmed at 250 °C, and the injection split ratio was 1:10. The oven was initially held for 5 min at a temperature of 40 °C, which was then raised to 200 °C at a rate of 5 °C/min and was isothermally maintained for 3 min. Helium was used as the carrier gas, and the flow rate was programmed at 23 cm/s.

The volatile compounds were identified by comparison with the linear retention indices (RIs) of a homologous series of n-alkanes (C5-C20) and by assessment of the MS patterns of the samples and authentic standards with MS data obtained from the National Institute of Standards and Technology (NIST) MS Library, Version 2008. For MS detection, an Agilent 5975C mass spectrometer was used with the same headspace extraction, column, and oven conditions as those described above. The electron-impact ion source and interface were both programmed at 230 °C, the electron ionization at 70 eV, and the mass acquisition range (m/z) at 29-300 amu. The relative amounts (%) of the volatile compounds were determined by measurement of the peak area response. All analyses were carried out in triplicate.

Y. Asikin et al./Journal of Advanced Research xxx (2017) xxx-xxx

Compositional GC analysis:

1. Isolation of volatile aroma components (sample headspace generation) ▼

2. Separation of volatile component peaks by GC-FID ▼

3. Peak Identification: retention index, GC-MS data (samples, authentic standards; NIST MS Library) ▼

4. Chemometric differentiation: principal component analysis (PCA) MS-based electronic nose analysis:

1. Isolation of volatile aroma components (sample headspace generation) ▼

2. Generation of total mass spectrum of volatile components by GC-MS ▼

3. Conversion of total mass spectrum to mass fingerprint dataset ▼

4. Chemometric differentiation: principal component analysis (PCA)

hierarchical cluster analysis (HCA)

Fig. 1. Workflow of volatile aroma composition and MS-based electronic nose analyses of dogfruit and stink bean.

Sample:

- Dogfruit —

- Stink bean

Table 1

Morphological traits of dogfruit and stink bean of different ripening stages.

Traits Dogfruit Stink bean

Unripe Ripe Unripe Mid-ripe Ripe

Bean number per pod 1-2 1-2 8-9 12-13 14-15

Coat thickness (mm) 0.45 ± 0.07 0.45 ± 0.07 0.27 ± 0.06 0.29 ± 0.03 0.40 ± 0.07

Bean weight (g) 5.04 ± 0.89 12.45 ±1.61 1.14 ±0.09 1.16 ±0.10 2.81 ± 0.22

Bean length (mm) 25.33 ± 2.28 34.32 ± 2.30 17.44 ± 1.54 18.28 ±0.24 23.80 ± 0.25

Bean width (mm) 26.86 ±1.79 33.80 ±2.28 15.35 ±0.89 15.29 ±0.52 20.35 ± 1.00

Bean height (mm) 14.01 ±1.10 19.89 ±1.98 7.47 ± 0.33 7.59 ± 0.25 10.74 ±0.35

Bean color Light yellowish cream Deep greenish brown Light whitish green Light green Deep green

Each value is expressed as the mean ± standard deviation (n = 5). Colors were determined by visual observation.

MS-based electronic nose analysis

(a) Unripe without coat Unripe with coat

Ripe without coat

The MS-nose profiles of dogfruit and stink bean were acquired by using a GERSTEL Chemsensor (GERSTEL, Mülheim, Germany) in an Agilent G1888 HSS-7890A GC-5975C MS system (Agilent J&W) [19]. The headspace extraction and MS conditions were set as described above, except for the ion source and interface temperatures, which were both maintained at 250 °C. Volatile compounds from the samples were passed through an HP-5MS fused silica capillary column (30 m x 0.25 mm internal dimensions, 0.25 im film thickness, Agilent J&W). The oven was initially held for 1 min at a temperature of 40 °C, which was then raised to 250 °C at a rate of 20 °C/min and was isothermally maintained for 3 min. The total mass spectrum intensities of detected ion masses (m/z 29-300) of volatile components were converted to a mass fingerprint dataset. All analyses were carried out in triplicate.

with coat without coat

with coat without coat

with coat without coat

Л и ••«» i% M

F.........................пцгпт

111111111 ■111111 ■ ■м111111■■I■I•■••11■I■11<■11■■

В"11 ",'........IP........г1 ijj

ÏmmIiiiiImmIiiiiI

Fig. 2. (a) Dogfruit and (b) stink bean with and without bean coats of different ripening stages.

Statistical analysis

The relative concentrations of the volatile aroma components of dogfruit and stink bean were statistically compared by using Microsoft Office Excel 2007 (Microsoft Corp., Redmond, WA, USA) by analysis of variance, followed by Fisher's least significant difference post hoc test at P < 0.05. The chemometric differentiation of volatile compounds in dogfruit and stink bean and a correlation of their ion masses were evaluated by mean-centered principal component analysis (PCA) by using Pirouette 4.5 software (Infome-trix, Bothell, WA, USA). The connection between dogfruit and stink bean was also statistically determined through a hierarchical

4 Y. Asikin et al./Journal of Advanced Research xxx (2017) xxx-xxx

cluster analysis (HCA) plot by using Pirouette 4.5 software. The MS data were preprocessed in a mean-centering structure, and the HCA plot was taken at Euclidean distance and incremental linking.

Results and discussion

Volatile aroma components of dogfruit and stink bean of different ripening stages

Dogfruit and stink bean possessed distinct volatile aroma components that accounted for 94.36-98.24% of identified compounds

at different maturation stages (Table 2). The peak area relative content of these volatiles was 0.64 and 2.82 E+08 in unripe and ripe dogfruits, respectively. They ranged from 1.85 to 1.94 E+08 in stink bean during ripening. There were 24 volatile components in both unripe and ripe dogfruit, whereas stink bean had more complex profiles with 42, 41, and 32 compounds in unripe, mid-ripe, and ripe beans, respectively. The major volatile component groups of unripe dogfruit were 42.74% alcohols (4 compounds) and 42.15% aldehyde compounds (12), followed by 8.05% sulfur compounds (5). The composition due to the alcohols and sulfurs altered to 41.90 and 25.90%, respectively, during ripening, whereas the

Table 2

Relative concentrations (%) of volatile aroma compounds of dogfruit and stink bean.

No RI Compound Dogfruit Stink bean Identification*

Unripe Ripe Unripe Mid-ripe Ripe

1 525 Hydrogen sulfide 0.16 ± 0.04d 0.19 ± 0.03d 2.32 ± 0.02b 1.92 ± 0.13c 5.59 ± 0.39a RI MS

2 670 Methanethiol 0.24 ± 0.08c 0.21 ±0.01c 3.25 ± 0.46b 4.00 ± 0.39b 9.63 ± 0.93a RI, MS

3 724 Carbon disulfide nd. 0.05 ± 0.01a tr. tr. tr. RI, MS, Std

4 739 Dimethyl sulfide 5.84 ± 0.21a 1.12 ± 0.04b 0.20 ± 0.01c 0.11 ±0.01c tr. RI, MS, Std

5 1023 Thiophene tr. tr. 0.05 ± 0.00a 0.07 ± 0.01a 0.05 ± 0.00a RI, MS

6 1071 Dimethyl disulfide nd. nd. 0.15 ± 0.00b 0.20 ± 0.02a 0.11 ±0.01c RI, MS, Std

7 1112 1-(Methylthio)pentane tr. tr. 0.05 ± 0.03a tr. tr. RI, MS

8 1391 Dimethyl trisulfide tr. nd. 0.02 ± 0.01a 0.03 ± 0.00a tr. RI, MS, Std

9 1406 S-Ethyl hexanethioate nd. 0.03 ± 0.00a 0.04 ± 0.01a 0.03±0.01a tr. RI, MS

10 1560 2-Pentylthiophene tr. tr. 0.03 ± 0.00a 0.02 ± 0.00b tr. RI, MS

11 1675 2,3,5-Trithiahexane nd. tr. 0.04 ± 0.01a 0.02 ± 0.00a tr. RI, MS

12 1716 1-Methyl-3-(methylthio)benzene 0.11 ±0.02b 4.34 ± 0.25a 0.02 ± 0.00b tr. 0.09 ± 0.00b RI, MS

13 1785 1,2,4-Trithiolane 1.70 ± 0.52c 19.97 ±0.40a 0.93 ± 0.20c 1.12 ± 0.23c 13.66 ±1.08b RI, MS

Total sulfurs 8.05 25.90 7.10 7.52 29.13

14 698 Acetaldehyde 7.36 ± 1.22d 29.02 ± 0.24a 15.01 ±1.08c 20.72 ±1.12b 6.96 ± 0.23d RI, MS, Std

15 782 Propanal 0.18 ± 0.01c 0.06 ± 0.00e 0.25 ± 0.01a 0.21 ± 0.00b 0.12 ± 0.01d RI, MS, Std

16 807 2-Methylpropanal 5.53 ± 0.44a tr. 0.09 ± 0.00b 0.07 ± 0.01b tr. RI, MS, Std

17 867 Butanal tr. tr. 0.19 ± 0.01b 0.20 ± 0.01a 0.14 ± 0.01c RI, MS, Std

18 908 2-Methylbutanal 4.07 ± 0.30a tr. 0.04 ± 0.00b 0.04 ± 0.00b tr. RI, MS, Std

19 912 3-Methylbutanal 22.13 ±2.44a tr. 0.05 ± 0.00b 0.04 ± 0.00b tr. RI, MS, Std

20 974 Pentanal 0.31 ±0.01c 0.07 ± 0.00d 3.70 ± 0.18a 3.58 ± 0.12a 3.00 ± 0.15b RI, MS, Std

21 1078 Hexanal 1.39 ± 0.17d 0.12 ± 0.00d 56.03 ±1.52a 50.28 ±1.08b 38.79 ±2.41c RI, MS, Std

22 1151 2-Methylhexanal 0.12 ± 0.02b 0.04 ± 0.02b 1.52 ± 0.14a 1.40 ± 0.05a 0.14 ± 0.02b RI, MS

23 1179 Heptanal 0.10 ± 0.02c tr. 0.20 ± 0.01a 0.17 ± 0.00a 0.14 ± 0.01b RI, MS, Std

24 1216 2-Hexenal 0.50 ± 0.13a nd. 0.05 ± 0.00b 0.04 ± 0.00b 0.03 ± 0.00b RI, MS, Std

25 1319 Octanal tr. tr. 0.05 ± 0.00a 0.04 ± 0.01b 0.03±0.01b RI, MS, Std

26 1326 2-Heptenal nd. nd. 0.21 ±0.01a 0.18 ± 0.00b 0.09 ± 0.00c RI, MS, Std

27 1395 Nonanal 0.23 ± 0.04a 0.03 ± 0.00d 0.15 ± 0.01b 0.11 ±0.01c 0.13±0.01bc RI, MS, Std

28 1449 2-Octenal tr. tr. 0.16±0.01a 0.14 ± 0.00b 0.11 ±0.01c RI, MS, Std

29 1637 Benzaldehyde tr. tr. tr. 0.04 ± 0.00a tr. RI, MS, Std

30 1655 2-Nonenal tr. tr. 0.03 ± 0.00a 0.02 ± 0.00b tr. RI, MS, Std

31 1683 2,4-Nonadienal 0.23 ± 0.01a tr. tr. tr. tr. RI, MS

Total aldehydes 42.15 29.33 77.72 77.31 49.67

32 895 Methanol 34.16 ±0.93a 13.89 ±0.33b 6.23 ± 0.30d 6.21 ±0.18d 10.93 ± ±0.71c RI, MS, Std

33 933 Ethanol 7.26 ± 0.18b 27.78 ± 0.64a 0.85 ± 0.05c 0.81 ±0.01c 0.73 ± 0.26c RI, MS, Std

34 1207 3-Methylbutanol 0.94 ± 0.06a 0.08 ± 0.01b tr. tr. nd. RI, MS, Std

35 1250 Pentanol 0.38 ± 0.01c 0.15 ± 0.00d 1.45 ± 0.03a 1.32 ± 0.03b 1.32 ± 0.08b RI, MS, Std

36 1353 Hexanol tr. tr. 0.38 ± 0.10b 0.43 ± 0.12b 0.99 ± 0.06a RI, MS, Std

Total alcohols 42.74 41.90 8.90 8.77 13.97

37 792 Octane tr. 0.03 ± 0.00b 0.04 ± 0.00b 0.03 ± 0.00b 0.16 ± 0.02a RI, MS, Std

38 1436 (Z)-3-Ethyl-2-methyl-1,3-hexadiene nd. nd. 0.17 ± 0.01a 0.13±0.01b 0.10±0.01c RI, MS

Total aliphatic hydrocarbons - 0.03 0.21 0.17 0.25

39 810 Acetone nd. 0.12 ± 0.00c 0.13±0.01bc 0.15 ± 0.00b 0.38 ± 0.01a RI, MS, Std

40 971 2-Pentanone 0.08 ± 0.00d 0.03 ± 0.00d 0.89 ± 0.11b 1.03 ± 0.07a 0.37 ± 0.02c RI, MS, Std

Total ketones 0.08 0.15 1.03 1.18 0.75

41 880 Ethyl acetate tr. 0.04 ± 0.00b 0.04 ± 0.01bc 0.03 ± 0.00c 0.08 ± 0.00a RI, MS, Std

42 1291 Hexyl acetate tr. 0.09 ± 0.00a 0.07 ± 0.00b 0.09 ± 0.02a 0.07 ± 0.00b RI, MS, Std

Total esters - 0.13 0.11 0.12 0.14

43 947 2-Ethylfuran tr. tr. 0.06 ± 0.00a 0.05 ± 0.01b 0.03 ± 0.00c RI, MS

44 1239 2-Pentylfuran 0.82 ± 0.05a 0.32 ± 0.10bc 0.25 ± 0.03c 0.31 ± 0.04c 0.43 ± 0.05b RI, MS

Total heterocycles 0.82 0.32 0.32 0.36 0.47

45 1456 Acetic acid 0.53 ± 0.14b 0.47 ± 0.03b 0.53 ± 0.11b 0.63±0.11b 0.93 ± 0.12a RI, MS, Std

46 1858 Hexanoic acid tr. tr. 0.16 ± 0.05b 0.21 ± 0.02b 0.31 ± 0.06a RI, MS, Std

Total acids 0.53 0.47 0.70 0.84 1.24

Total identified 94.36 98.24 96.08 96.26 95.62

Total content (peak area 1 x E + 08) 0.64 2.82 1.94 1.82 1.85

Each value is expressed as the mean ± standard deviation (n = 3), obtained by GC-FID analysis; nd.: not detected; tr.: trace amount (<0.01%); values in the same row followed by the same letter are not significantly different (P < 0.05).

* RI: identification based on retention index; MS: identification based on the NIST MS library; Std: identification based on pure standards analyzed by mass spectrometry.

Y. Asikin et al./journal of Advanced Research xxx (2017) xxx-xxx

aldehyde content declined from 42.15 to 29.33%, with only half the number of identified compounds remaining. On the other hand, the stink beans predominately contained aldehydes, although the proportion declined from 77.72 to 49.67% during ripening; this was accompanied by elevations in the alcohol and sulfur components from 8.90 to 13.97% and 7.10 to 29.13%, respectively. These results indicated that various biochemical reactions, including lipid and carbohydrate degradations, as well as amino acid and phenyl-propanoid metabolic changes, occur to a large extent during final ripening of these beans and can alter their volatile flavor profiles [13,14]. Conversely, maturation development from the early to intermediate ripening stage has less impact on the overall volatile flavor profile of stink bean, which indicates that slower volatile component generation occur while the plant is using more nutrients for enlarging its size and weight [13,15].

In detail, the predominant volatile components of unripe dogfruit were methanol and 3-methylbutanal (34.16 and 22.13%, respectively), the amounts of which were significantly higher (P <0.05) than those in ripe dogfruit and other smelly beans (Table 2). The composition also comprised intermediate amounts of acetaldehyde, ethanol, dimethyl sulfide, and 2-methylpropanal ranging from 5.53 to 7.36%. These volatile components may provide green, malty, pungent, and sulfurous smells to unripe dogfruit [25,26]. Moreover, unripe dogfruit contained significantly higher minor amounts of 3-methylbutanol, 2-pentylfuran, 2-hexenal, and nonanal than other materials and was the only sample containing 2,4-nonadienal. Conversely, ripe dogfruit had significantly higher acetaldehyde, ethanol, and 1,2,4-trithiolane levels, at 29.02, 27.78, and 19.97%, respectively. These predominant volatiles contribute pungent and ether odors to the characteristic ripening of this food material [26]. Sulfuric 1,2,4-trithiolane, in particular, is known to be one of the key aroma components in shiitake mushrooms that provide the woody and fresh shiitake-mushroom perceptions [27]. However, both unripe and ripe dogfruits lacked dimethyl disulfide, 2-heptenal, and (Z)-3-ethyl-2-methyl-1,3-hexadiene which might exclude sour-putrid cabbage, soap-fat, and nutty characteristics from their volatile flavor profiles, respectively [25,26,28].

Stink bean had a remarkably higher amount of hexanal, which may specify green and grassy aroma traits [29], than that in dog-fruit. In spite of that, the amount of this volatile aldehyde significantly and gradually declined during ripening, from 56.03 to 50.28% in unripe and mid-ripe beans, respectively, and it then reached 38.79% in the ripe stage. Moreover, stink bean had about 15.01% acetaldehyde in the unripe stage, which significantly increased to a level of 20.72% in the mid-ripe period but then dropped to 6.96% during the final ripening process. On the other hand, unripe and mid-ripe stink beans comprised steady intermediate amounts of methanethiol (3.25-4.00%) and methanol (6.216.23%), which were then significantly enhanced to 9.63 and 10.93%, respectively. The ripening process also remarkably improved concentrations of hydrogen sulfide and 1,2,4-trithiolane from 1.92 and 0.93% to 5.59 and 13.66%, respectively. The large portion of sulfuric compounds in the compositional result of the present study is in agreement with the previously reported volatile profile of Malaysian stink beans [2]; these compounds are also important constituents in other strong-aroma plant materials and products, including leeks, onions, and dried mushrooms [27,30]. Taken together, the sulfurous, putrid, cheesy, woody, and shiitake odor characteristics from sulfuric volatile components are enhanced in stink bean during ripening and may impact on the sensory flavor perception when it is consumed or used as a food ingredient [26,27,29].

The distinctiveness of the volatile flavor profiles of dogfruit and stink bean of different ripening stages was also shown from the useful arrangement for the first two principal component (PC) fac-

ю 0.1

CM 0.0

О -0.1

3-M0thylbutanal о

Methanol

2-Methylbutanal

Pentanal, hydrogen sulfide Q Methanethiol Hexanal \¡ 2-Methylpropanal i -0° Dimethyl sulfide .....................

O 1-Methyl-3- (methylthio)benzene

Other aroma compounds 1,2,4-Trithiolane о Ethanol

Acetaldehyde о

-0.8 -0.6 -0.4 -0.2 0.0 Factor 1 (69.2%)

см О

Unripe DF

Unripe SB Ripe

Mid-ripe SB

—4 -^ж....;..............................L...

Ripe DF ,

-150 -100

-50 0 50 Factor 1 (69.2%)

Fig. 3. (a) Factor loadings and (b) principal component score plots of the relative concentrations of the volatile aroma compounds of dogfruit (DF) and stink bean (SB), obtained by GC-FID analysis.

tors in PCA plots that were derived from the relative concentrations of the volatile aroma components (Fig. 3). The factor loadings plotted several distinct volatile components for the first two PC factors that might explain the volatile composition variations of dogfruit and stink bean (Fig. 3a). They were methanol, 3-methylbutanal, 2-methylbutanal, 2-methylpropanal, and dimethyl sulfide, which were plotted in the positive quadrant of both factors, whereas ethanol, acetaldehyde, 1,2,4-trithiolane, and 1-methyl-3-(methylthio)benzene were only positively related to factor 1. On the other hand, hexanal was clearly separated in the outlying negative quadrant of factor 1, along with methanethiol, hydrogen sulfide, and pentanal, but the latter compounds were close to the plot center where other volatile compounds were loaded. These center-loaded plots indicated compositional likeness of the volatiles in dogfruit and stink bean and, thus, suggest common base aroma formations to the two bean materials regardless of the maturity stage. Moreover, the score plots showed opposite separation of the materials to the first PC factor (69.4%), in which dogfruit was recorded in the positive quadrant and stink bean in the negative (Fig. 3b). Therefore, the second PC factor (25.5%) could separate unripe and ripe dogfruit but failed to distinguish the volatile-profile variations in stink bean during ripening. This PCA outcome thus clearly

Y. Asikin et al./Journal of Advanced Research xxx (2017) xxx-xxx

showed separation of the two beans according to their volatile aroma components as discriminatory loading factors. However, stink beans at different maturity stages might be recorded as a single material when stink bean and dogfruit are evaluated together. On the other hand, the maturity stage allowed differentiation of the volatile profile of dogfruit, as indicated by significantly higher amounts of 3-methylbutanal and methanol in the unripe material, whereas the prominent volatiles were acetaldehyde, ethanol, and 1,2,4-trithiolane in fully ripened beans.

MS-based electronic nose profiles of dogfruit and stink bean of different ripening stages

The volatile aroma profiles of dogfruit and stink bean were also differentiated through a PCA plot from MS-nose analysis that accounted for 97.4% in the first two PC factors (Fig. 4). The score plot outlined a separation of the unripe or mid-ripe dogfruit and stink bean from their fully ripened beans that was clearly bordered by the zero line of PC factor 1 (Fig. 4b). Moreover, unripe dogfruit was solely positively associated with both factors and was clearly separated from ripe dogfruit and any stink bean. However, unlike the result in the volatile compositional PCA plot, ripened stink bean was distinctly plotted from unripe and mid-ripe beans (Fig. 3b versus Fig. 4b). This improved volatile-profile separation was due to the MS intensities of influential discriminatory ions that

Fig. 4. (a) Factor loadings and (b) principal component score plots of the volatile profiles of dogfruit (DF) and stink bean (SB), obtained by MS-nose analysis.

were captured from scanned ion masses with a much larger number of loaded variables than that of the volatile compositional method (272 ions [recorded from m/z 29-300] versus 46 identified compounds) (Fig. 3a versus Fig. 4a) [19,21]. The potential association had been found between discriminant ion masses with MS fragmentation. These may derive from the samples' volatiles through comparison of each discriminant ion with the MS fragmentation patterns (target and qualifier ions) of the identified volatiles listed in Table 2 and corresponding authentic standards, analyzed by compositional GC method.

In detail, the corresponding loading plot showed important scattered ions, such as m/z 39, 41, 42, 43, 58, 62, and 71, that positively associated with both PC factors, whereas m/z 60, 78, 124, and 126 were oppositely positioned (Fig. 4a). These discriminatory ion masses revealed key qualifier ions for associated aroma compounds and might be suitable for distinguishing dogfruit and stink bean during ripening. For instance, m/z 41, 43, and 58 which may derive from predominant 3-methylbutanal might contribute to the separation of unripe dogfruit from other beans (Fig. 4b and Table 2). Conversely, m/z 78 and 124, which are qualifier ions for 1,2,4-trithiolane, clearly indicate the ripened beans, and the significantly greater relative concentration of this sulfuric compound in ripe dogfruit located it at a more distant negative plot within PC factor 1. In addition, other recorded ion masses were only positively associated with factor 1, including m/z 55, 56, 57, 67, 72, 76, 81, and 82. These prominent scattered ions might further indicate the influence of hexanal as the predominant compound in unripe and mid-ripe beans, as the qualifier ions m/z 55, 56, 57, 67, and 82 are linked to this green-grassy aroma emitting aldehyde.

Another multivariate statistical analysis also confirmed volatile aroma profile differentiation of dogfruit and stink bean from their recorded ion masses in an HCA dendrogram clustering tree (Fig. 5). Out of the five beans of different ripening stages, four volatile groups were formed at a component similarity of 0.900, wherein stink beans at the unripe and mid-ripe stages comprised a mixed cluster. Moreover, ripe dogfruit and stink bean were presented as closest group to one another and were split from their immature forms, indicating the comparable progression of their volatile component profiles during the bean maturation process. This clustering outline clearly provides a better general view of the volatile aroma component discrimination in plant resources during ripening of different origins, including dogfruit and stink bean [21,22]. These

Mid-ripe SB 3 ]

Unripe SB 3

Unripe SB 2

Mid-ripe SB 2 1

Mid-ripe SB 1 J

Unripe SB 1 -1

Unripe DF 3

Unripe DF 2 i

Unripe DF 1 I

Ripe SB 3 Ripe SB 2 1

Ripe SB 1 1

■ Ripe DF 3

1 Ripe DF 2

Ripe DF 1

Fig. 5. HCA dendrogram of the volatile profiles of dogfruit (DF) and stink bean (SB), obtained by MS-nose analysis. The volatile component similarity was obtained as 0.900.

Y. Asikin et al./journal of Advanced Research xxx (2017) xxx-xxx

MS-based nose results detailed the volatile-profile differentiations and provided an important chemical markers in a form of discriminative MS dataset as ''digital fingerprints" for dogfruit and stink bean during maturity for further development of rapid measurement technology on volatile alterations evaluation of these legumes or their derivative products [21]. The MS-based electronic nose method and chemometric data analysis might thus be applied for monitoring the flavor quality of smelly plant materials in a faster and thorough manner than compositional GC measurement, which confirms the advantageous use of MS-based e-nose profiling technique on differentiation of food flavor [19-21,31,32].

Conclusions

Dogfruit and stink bean had distinctive compositions and contents of volatile aroma components that varied greatly in the alcohol, aldehyde, and sulfur compounds, but stink bean comprised a greater number of volatiles than that of dogfruit. Stink bean mostly contained hexanal at all maturity stages, whereas unripe dogfruit was primarily predominated by 3-methylbutanal and methanol, which then altered to acetaldehyde and ethanol in ripe dogfruit. There were significant changes in the amount of 1,2,4-trithiolane in both dogfruit and stink bean during maturation. The compositional dataset constructed a multivariate PCA plot that displays separation only for dogfruit during ripening. The non-targeted MS-based electronic nose and chemometric analyses further distinguished the volatile profiles of dogfruit and stink bean on an ion-mass basis, and detailed the differentiation of these smelly materials through PCA and HCA arrangements. The MS-based nose technique also provided a valuable recorded MS dataset and discriminative ion masses which may be derived from samples' volatile components, such as m/z 41, 43, 58, 78, and 124, that could be used as ''digital fingerprints" for monitoring volatile flavor changes in dogfruit and stink bean during ripening.

Conflicts of interest

Authors have declared no conflicts of interest. Compliance with Ethics Requirements

This article does not contain any studies with human or animal subjects.

Acknowledgements

The authors are grateful to the Japan Society for the Promotion of Science for an International Research Fellowship awarded to Y.A. (ID No. P14075). We would like to thank Editage (www.editage.jp) and Split Horizons, LCC (Samuel Bernard) for English language editing supports.

References

[1] Barceloux DG. Djenkol bean [Archidendron jiringa (.lack) I. C. Nielsen]. Dis Mon 2009;55:361-4.

[2] Miyazawa M, Osman F. Headspace constituents of Parkia speciosa seeds. Nat Prod Lett 2001;15:171-6.

[3] Sridaran A, Karim AA, Bhat R. Pithecellobium jiringa legume flour for potential food applications: studies on their physico-chemical and functional properties. Food Chem 2012;130:528-35.

[4] National Parks Board [homepage on the Internet]. Singapore: Parks Board; c2013 [cited 2017 Oct 19]. Parkia speciosa Hassk.; [about 1 screen]. Available from: <https://florafaunaweb.nparks.gov.sg/special-pages/plant-detail.aspx? id=3052>.

[5] Mohamed S, Rahman MSA, Sulaiman S, Abdullah F. Some nutritional and anti-nutritional components in jering (Pithecellobium jeringa), keredas (Pithecellobium microcarpum) and petai (Parkia speciosa). Pertanika 1987;10:61-8.

Shukri R, Mohamed S, Mustapha NM, Hamid AA. Evaluating the toxic and beneficial effects of jering beans (Archidendron jiringa) in normal and diabetic rats. J Sci Food Agric 2011;91:2697-706.

Charungchitrak S, Petsom A, Sangvanich P, Karnchanatat A. Antifungal and antibacterial activities of lectin from the seeds of Archidendron jiringa Nielsen. Food Chem 2011;126:1025-32.

Jamaluddin F, Mohamed S, Lajis MN. Hypoglycaemic effect of Parkia speciosa seeds due to the synergistic action of p-sitosterol and stigmasterol. Food Chem 1994;49:339-45.

Siow HL, Gan CY. Extraction of antioxidative and antihypertensive bioactive peptides from Parkia speciosa seeds. Food Chem 2013;141:3435-42. Cheng YF, Bhat R. Functional, physicochemical and sensory properties of novel cookies produced by utilizing underutilized jering (Pithecellobium jiringa Jack.) legume flour. Food Biosci 2016;14:54-61.

Gan C-Y, Latiff AA. Antioxidant Parkia speciosa pod powder as potential functional flour in food application: physicochemical properties' characterization. Food Hydrocoll 2011;25:1174-80.

Longobardi F, Sacco D, Casiello G, Ventrella A, Sacco A. Chemical profile of the Carpino broad bean by conventional and innovative physicochemical analyses. J Food Qual 2015;38:273-84.

Obenland D, Collin S, Sievert J, Negm F, Arpaia ML. Influence of maturity and ripening on aroma volatiles and flavor in 'Hass' avocado. Postharvest Biol Technol 2012;71:41-50.

Agudelo-Romero P, Erban A, Sousa L, Pais MS, Kopka J, Fortes AM. Search for transcriptional and metabolic markers of grape pre-ripening and ripening and insights into specific aroma development in three Portuguese cultivars. PLoS ONE 2013;8:e60422.

Bron IU, Jacomino AP. Ripening and quality of 'Golden' papaya fruit harvested at different maturity stages. Braz J Plant Physiol 2006;18:389-96. Frérot E, Velluz A, Bagnoud A, Delort E. Analysis of the volatile constituents of cooked petai beans (Parkia speciosa) using high-resolution GC/ToF-MS. Flavour Fragr J 2008;23:434-40.

Mkanda AV, Minnaar A, de Kock HL. Relating consumer preferences to sensory and physicochemical properties of dry beans (Phaseolus vulgaris). J Sci Food Agric 2007;87:2868-79.

Talavera-Bianchi M, Adhikari K, Chambers IV E, Carey EE, Chambers DH. Relation between developmental stage, sensory properties, and volatile content of organically and conventionally grown pac choi (Brassica rapa var. Mei Qing Choi). J Food Sci 2010;75:S173-81.

Asikin Y, Takahara W, Takahashi M, Hirose N, Ito S, Wada K. Compositional and electronic discrimination analyses of taste and aroma profiles of non-centrifugal cane brown sugars. Food Anal Methods 2017;10:1844-56. Sliwinska M, Wisniewska P, Dymerski T, Wardencki W, Namiesnik J. Application of electronic nose based on fast GC for authenticity assessment of Polish homemade liqueurs called nalewka. Food Anal Methods 2016;9:2670-81.

Asikin Y, Maeda G, Tamaki H, Mizu M, Oku H, Wada K. Cultivation line and fruit ripening discriminations of Shiikuwasha (Citrus depressa Hayata) peel oils using aroma compositional, electronic nose, and antioxidant analyses. Food Res Int 2015;67:102-10.

Cui S, Wang J, Yang L, Wu J, Wang X. Qualitative and quantitative analysis on aroma characteristics of ginseng at different ages using E-nose and GC-MS combined with chemometrics. J Pharm Biomed Anal 2015;102:64-77. Caelenberg TV, Leuven IV, Dirinck P. An analytical approach for fast odour evaluation of recycled food-grade paperboard materials using HS-SPME-MS-nose technology. Packag Technol Sci 2013;26:161-72. Weerawatanakorn M, Asikin Y, Takahashi M, Tamaki H, Wada K, Ho CT, et al. Physico-chemical properties, wax composition, aroma profiles, and antioxidant activity of granulated non-centrifugal sugars from sugarcane cultivars of Thailand. J Food Sci Technol 2016;53:4084-92. Liu RS, Li DC, Li HM, Tang YJ. Evaluation of aroma active compounds in tuber fruiting bodies by gas chromatography-olfactometry in combination with aroma reconstitution and omission test. Appl Microbiol Biotechnol 2012;94:353-63.

Zhu J, Chen F, Wang L, Niu Y, Yu D, Shu C, et al. Comparison of aroma-active volatiles in oolong tea infusions using GC-olfactometry, GC-FPD, and GC-MS. J Agric Food Chem 2015;63:7499-510.

Hiraide M, Miyazaki Y, Shibata Y. The smell and odorous components of dried shiitake mushroom, Lentinula edodes I: relationship between sensory evaluations and amounts of odorous components. J Wood Sci 2004;50:358-64. Dong L, Piao Y, Zhang X, Zhao C, Hou Y, Shi Z. Analysis of volatile compounds from a malting process using headspace solid-phase micro-extraction and GC-MS. Food Res Int 2013;51:783-9.

Fuchsmann P, Stern MT, Brugger YA, Breme K. Olfactometry profiles and quantitation of volatile sulfur compounds of Swiss Tilsit cheeses. J Agric Food Chem 2015;63:7511-21.

Kusano M, Kobayashi M, Iizuka Y, Fukushima A, Saito K. Unbiased profiling of volatile organic compounds in the headspace of Allium plants using an in-tube extraction device. BMC Res Notes 2016;9:133.

Fenaille F, Visani P, Fumeaux R, Milo C, Guy PA. Comparison of mass spectrometry-based electronic nose and solid phase microextraction gas chromatography-mass spectrometry technique to assess infant formula oxidation. J Agric Food Chem 2003;51:2790-6.

Liberto E, Ruosi MR, Cordero C, Rubiolo P, Bicchi C, Sgorbini B. Non-separative headspace solid phase microextraction-mass spectrometry profile as a marker to monitor coffee roasting degree. J Agric Food Chem 2013;61:1652-60.