Identification of a freshness marker metabolite in stored soybean sprouts by comprehensive mass-spectrometric analysis of
4.3 Results and discussion
68 amu with dynamic fill time and a scan rate of 10,000 Da s−1, and the resolution of Q1 device set to unit. Nitrogen was used as a collision gas both in multiplexing MRM and product ion scanning detection mode.
4.2.7 Data processing
Data from peaks of CC derivatives were processed using Marker View™
software 1.2.1. (AB-Sciex, Framingham, MA, USA). CC-DH peaks were extracted using Gaussian smoothing of 1.5 points, noise percentage of 50%, a baseline subtraction window of 8 min, a peak splitting factor of 4 points, a retention time tolerance of 1 min, a minimum intensity of 1500 cps, a minimum peak with 2 points, and a minimal signal/noise ratio of 20. Peak areas were normalized to that of the IS and to sample weights. Principal component analysis with discriminant analysis (PCA-DA) was then performed using Pareto scaling and none weighting in order to find out any differential features between sample groups. Selected marker ion candidates were then identified using The Lipidmaps (www.lipidmaps.org), The Human Metabolome Database (www.hmdb.ca), The METLIN Metabolomic Database (https://metlin.scripps.edu), The ChemicalBook Database (www.chemicalbook.com) and compared with a purchased authentic standard.
69 other hand, CO2 production rate at 20 °C was drastically decreased from 9 to 5 mmol kg–1 h–1 during storage. Kader (2002b) mentioned that the rate of perishability of fresh produce is generally proportional to respiration rate and classified fresh commodities according to their respiration rates. Referring to this classification, soybean sprouts could be can be categorized as a „very high‟
respiration rate, thus soybean sprouts is highly perishable.
70 Fig. 4.1. Changes of CO2 production rates of soybean sprouts stored at 5 °C, 10 °C and 20 °C (Vertical lines represent standard deviation, n =3).
0 2 4 6 8 10
0 2 4 6 8 10 12
CO2production rate (mmol kg−1h−1)
Storage duration (d)
5 °C10 °C 20 °C
71 The rate of respiration is affected by not only temperature but also the amount of the time elapsed after harvest. According to Brash et al. (1995), the respiration rate of asparagus stored at 20 °C decreased by over 50% in the first 24 hours after harvest, and the decrement was less significant with decreasing of temperatures. Deterioration of fresh produce is primarily driven by the product‟s own tissue metabolism and there is a tight linkage between metabolism and perishability. Since CO2 productions provide parallel measures of metabolic activity, the cumulative CO2 production could be used as a reference of degree of freshness. In latter sections, we discuss the relationship between the cumulative CO2 production and change of CCs in soybean sprouts during storage to identify the potential freshness marker.
4.3.2 Profile of CCs in soybean sprouts
Figure 4.2 demonstrates CC metabolites features in fresh soybean sprouts (A; cotyledons and B; hypocotyls). Detected CC-DHs were plotted in circles as a function of retention time (RT) and m/z value. The diameter of each circle represents peak area of detected CC-DH normalized to that of IS-DH at a RT of 11.34 min and to respective sample weight. Even in the fresh condition, about 171 of CC-DHs in cotyledons (4.2A) and 228 of CC-DHs in hypocotyls (4.2B) were detected. Most of these CC-DHs were distributed at the range of retention time from 3 min to 15 min and m/z from 350 to 600. Many of CCs in fresh soybean sprouts have a wide range of polarity and molecular weight that might be considered as secondary metabolites because there are numerous CCs classified as secondary metabolites including vitamins, isoflavones, flavonols, chalcones, and their derivatives which have carbonyl skeletons (Di Carlo, Mascolo, Izzo, &
72 Capasso, 1999; Kim, Kim, Chung, Chi, Kim, & Chung, 2006; Gu et al., 2017).
They likely react with DH forming CC-DH derivatives after extraction in polar and semi-polar solvent mixtures thus leading to the detection of numerous CC-DHs. Moreover, secondary metabolites in soybean sprouts vary in hypocotyls and cotyledons depending on the soybean variety (Plaza, Ancos & Cano, 2003; Youn, Kim, Lee & Kim, 2011). From our data, we can assume that, in soybean sprouts of Glycine max, cv. BS5012, higher number of CCs species are distributed in hypocotyls compared to cotyledons.
73 Fig. 4.2. Carbonyl compound metabolites feature in fresh soybean sprouts;
cotyledon (A) and hypocotyl (B). Detected CC-DHs were plotted in circles as a function of retention time (RT) and m/z value. The diameter of each circle represents peak area of detected CC-DH normalized to that of internal standard-dansyl hydrazine (IS-DH) at a RT of 11.34 min and to respective sample weight.
250 350 450 550 650 750
m/z
(A)
250 350 450 550 650 750
0 5 10 15 20 25
m/z
Retetion time ( min) (B)
74 Figure 4.3 demonstrates the score plots (A) and corresponding loading plots of CC-DH signals (B) from hypocotyls of stored soybean sprouts relating to cumulative CO2 production. Four sample clusters were circled and positioned in different areas of the score plot to discriminate differences between groups as a function of cumulative CO2 production, which was observed in first, second, third, and fourth clusters at 0, 102–115, 176–220, 287–303 mmol kg–1, respectively, and increased with D1 scores (Fig. 4.3A). D1 score contains the information of cumulative CO2 production during storage of soybean sprouts.
However, distributions of CC-DHs in cotyledons were not distinguished relating to the cumulative CO2 production in PCA-DA score plots (data not shown). From corresponding loading plot of CC-DH signals from hypocotyls (Fig. 4.3B), increases in cumulative CO2 production are explained by the positive direction of D1 where three signals were circled at the positive edge of D1 axis of the loading plot that responsible for clustering of the samples and had m/z of Q1_m/z of Q3_RT pairs of 364_236.1_8.82, 512_236.1_9.34, and 330_236.1_17.29.
Moreover, the normalized peak area of these signals showed positive and higher correlation with cumulative CO2 production during storage under various temperatures among detected signals (Fig. 4.4). The correlation coefficient (r) of each signal was also significant (p < 0.05).
Figure 4.5 demonstrates the relationship between normalized peak area of each selected signal and cumulative CO2 production during storage under various temperatures. The regression line in the figure was obtained by the least squares method. The coefficient of determination (R2) of 330_236.1_17.29 (Fig. 4.5A), 364_236.1_8.82 (Fig. 4.5B) and 512_236.1_9.34 (Fig. 4.5C) were 0.45, 0.35 and 0.71, respectively. Previously, to identify the quality markers in food samples
75 such as fish (iced Mediterranean hake and yellowtails) and agricultural commodity (wild rocket, basil, lettuce, cabbage and hazelnuts), the value of R2 ranging from 0.63 to 0.98 was considered as acceptable level (Mabuchi, Zhao, Kondo, & Tanimoto, 2017; Luca, Kjær, & Edelenbos, 2017; Cozzolino, Pace, Cefola, Martignetti, Stocchero, Fratianni, Nazzaro, & De Giulio, 2016; Baixas-Noqueras, Bover-Cid, Veciana-noqués, Mariné-Font, & Vidal-Carou, 2005;
Lonchamp, Barry-Ryan and Devereux, 2009; Fallico, Arena, & Zappalà, 2003).
Moreover, in contrast that the signals of 330_236.1_17.29 and 364_236.1_8.82 were detected when the cumulative CO2 production was 0 in soybean sprouts immediately after harvest, the signal of 512_236.1_9.34 was not observed on that point but subsequently rose with increasing of cumulative CO2 production in time.
This trend can be utilized for simple evaluation of food‟s quality and freshness, because it can easily detect whether the status of a test object is good, or not by only indicating the presence of the identified marker (Fallico et al., 2003; Fallico, Zappalá, Arena, & Verzera, 2004). Therefore, in this study, only the signal of 512_236.1_9.34 was selected for subsequent structure elucidation as a candidate for freshness marker of soybean sprouts.
)LJ 6FRUH VFDWWHU SORW $ DQG ORDGLQJ VFDWWHU SORW % RI WKH 3&$'$ RI
&&'+VLQWKHK\SRFRW\OVLQVWRUHGVR\EHDQVSURXWV9DOXHVEHVLGHGDWDSRLQWVLQ )LJ $ LQGLFDWH FXPXODWLYH &2 SURGXFWLRQ GXULQJ VWRUDJH PPRO NJ± ):
9DOXHVEHVLGHGDWDSRLQWVLQ)LJ%LQGLFDWHP]RI4BP]RI4B57SDLUVIRU HDFK&&'+VLJQDO
%
$
77 Fig. 4.4. Correlation coefficient (r) between each normalized peak area of CC-DH signals and cumulative CO2 productions during storage under various temperatures.
-0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
100 367 416 460 517 575 664
Correlation Coefficient, r
m/zof CC-DH precursor ion
364_236.1_8.82
512_236.1_9.34 330_236.1_17.29
300 400 500 600
78 Fig. 4.5. Relationships between normalized peak area of selected CCs candidates and cumulative CO2 production (A: 330_236.1_17.29, B: 364_236.1_8.82, C:
512_236.1_9.34).
0 3000 6000 9000
Normalized area
(364_236.1_8.8) 0
2000 4000 6000
Normalized area
(330_236.1_17.29)
(A)
0 1000 2000 3000 4000
0 100 200 300 400
Normalized area
Cumulative CO2production (mmol kg–1)
(512_236.1_9.34) (C)
(B) y= 5.42x + 505.09
R2 = 0.45
y= 10.22x + 1824.8 R2 = 0.35
y= 7.70x - 169.89 R2 = 0.71
79 4.3.3 Structure elucidation of freshness marker ion
The enhance product ion (EPI) scan detection in Analyst® system of AB-SCIEX was performed against the selected ion of m/z 512 with RT of 9.34 min to identify the selected freshness marker ion. Figure 4.6 demonstrates the product ions mass spectra of the selected ion of m/z 512 at RT of 9.34 min (upper part) and that of the standard abscisic acid (ABA)-DH (lower part). Initially, the losses of fragment ions with m/z of 18 mass units were observed in Fig. 4.6 (upper part). It likely reflects the loss of water molecules from protonated hydroxyl groups in allylic positions (Britton, 1996) and suggesting the presence of hydroxyl ions. Generally, a hydroxyl ion can be derived from fatty acids, carotenoids, flavonoids and their conjugates. Furthermore, since the selected marker ion is a protonated ion molecule, [M+H]+, it comprises CC, DH and H+ (hydrogen-adduct ion). Therefore, to calculate the molecular weight (MW) of the CC from this CC-DH derivative, m/z values of 1 for hydrogen-adduct ions and 263 for DH moieties were subtracted from the detected m/z value of 512, and an m/z value of 16 was added for the atomic mass of oxygen to form a carbonyl skeleton (Fig. 4.7). Based on these calculations, the MW of 264 mass units for the selected freshness marker metabolite was extracted. Using all the information obtained in our analysis, we have searched the online metabolomics databases for the corresponding names and structures of possible candidates. Five compounds were nominated as candidates, and are listed with their formulas, structures, and hydrophobicity (Log P) values in Table 4.1. These compounds are characterized as derivatives of fatty acids, flavonoids, and their conjugates. In addition, since a reverse phase chromatographic separation system was used for separating the complex CC-DHs in the samples, the resulting RTs reflect the polarities.
80 Specifically, the RT of the selected CC-DH derivative was 9.34 min and was faster than that of p-BOBA-DH, which was detected at 11.34 min as an IS-DH derivative. Therefore, the selected freshness marker ion is more polar than p-BOBA-DH, and from the compounds listed with log P values in Table 4.1, only ABA has a lower log P value than p-BOBA. Thus, to confirm that ABA is the present freshness marker metabolite in soybean sprouts, we purchased an authentic ABA standard and conducted EPI detection against the ABA-DH derivative (Fig. 4.7, lower part), and indicated that fragmentation patterns of product ion mass spectra of ABA-DH exactly matched those of the selected CC-DH. In addition, the RT of the ABA-DH derivative was 9.41 min, similar to that of the selected CC-DH. Finally we conclude that ABA is the identified freshness marker metabolite for soybean sprouts.
81 Fig. 4.6. Schematics flowchart for determining the molecular weight of candidate signal (512_236.1_9.34) as a freshness marker metabolite of soybean sprouts.
N
S O O N N
H C H
N
S O O N N
H C
N
S O O N N
H
O R R'
R R'
C
C R O R'
MW of selected CC-DH (511)
m/z 512 (Protonated molecule of selected marker ion)
(- 263, m/z of attached DH moeity)
MW of CC metabolite as freshness marker (264) (- 1, m/z of hydrogen-adduct ion)
MW of uncomplete CC
(248) H
(+ 16, atomic mass of oxygen) R R'
82 Table 4.1 List of the name for the compound with MW of 264 that contain hydroxyl group (loss of m/z=18) and internal standard (RT of 11.4 min)
H3C OH
O
H3C OH
O
O
O OH
OH O
O
CH3 H3C
OHCH3 CH3
OH O
O
H O
No Name Formula Log P (Hydrophobicity) Structure On line databases
1 Norlinolenic acid C17H28O2 5.27 www.lipidmaps.org
2
10E-heptadecen-8-ynoic acid C17H28O2 4.94 www.lipidmaps.org
3 2'-Hydroxyfurano
[2'',3'':4',3']chalcone C17H12O3 4.03 www.lipidmaps.org
4 all-trans-7-hydroxy hexadeca-2,4,8, 10-tetraenoic acid
C16H24O3 3.91 www.lipidmaps.org
5 Abscisic acid C15H20O4 2.54 www.lipidmaps.org
6 p-BOBA (IS) C14H12O2 3.3 www.chemicalbook.com
OH H3C
83 Fig. 4.7. Product ions mass spectra of protonated molecules ion with m/z of 512 at RT of 9.34 min in positive ion mode; (upper part) selected freshness marker ion and (lower part) authentic ABA standard-DH.
133.2170.2218.0 236.1
277.2325.0 415.1450.2 476.2
494.0 512.2
170.2 218
236.1 277.1
325.0 450.0 476.0
494.0
512.0
-120 -100 -80 -60 -40 -20 0 20 40 60 80 100 120
0 100 200 300 400 500
% Intensity
(–18)
(–18)
(Standard ABA-DH ion)
0 100 200 300 400 500 m/z
(Selected CC-DH marker ion)
N
S O O N N
H
CH3 H3C
OHCH3 CH3
OH O
84 ABA is a plant hormone that was discovered at least 50 years ago and has since been shown to regulate many aspects of plant growth and development (Finkelstein, 2013). ABA is a member of monocyclic monoterpene family and comprises the metabolic precursors ketone and enolate (Duffield & Netting, 2001). Thus, ABA can be conjugated to DH through its ketone group. To date, the best known functions of ABA are related to roles as a major phytohormone that contributes to plant abiotic stress resistance. ABA is mainly induced by moisture loss stress, chilling temperature and salt stress (Swamy & Smith, 1999; Lafuente
& Sala, 2002; Romero, Rodrigo, & Lafuente, 2013), and accumulates through the cleavage of a C40 carotenoid precursor (Xiong & Zhu, 2003). According to Becker and Fricke (1996), fresh fruits and vegetables lose their moisture through the transpiration during storage. Hence, the presence of ABA in hypocotyls of soybean sprouts may have been induced by moisture loss. Transpiration is associated with transport and evaporation of moisture from the skin, and with convective mass transport of moisture to the atmosphere. Moreover, transpiration and respiration have been correlated in previous study whereas CO2 and heat from the associated chemical reaction during respiration may accelerate transpiration in fresh produce.
Previous studies have mentioned that volatile compounds detection is potential to indicate the degree of freshness of fresh produce. Luca et al. (2017) reported pentane and 2-ethylfuran contents related with leaf damage index (degree of tissue integration) during storage of wild rocket. Cozzolino et al. (2016) indicated that 1,8-cineole was considered as a potential marker of chilling injury of basil leave because it was increased associating with ion leakage score. In addition, Lonchamp et al. (2009) showed that dimethylethylphenol and ester
85 pentanoic acid were proposed as freshness indicators of stored young leaves of lettuce and cabbage because they increased associating with odor and browning symptom development. However, volatile compound markers proposed varied depending on the commodity that might be a limiting factor of their application in general postharvest management of agriculture products because each of the volatile markers needs specific detection system. On the other hand, ABA is ubiquitous in plants therefore detection system of only ABA could be applicable in wide range of fresh produces. However, the relationship between ABA accumulation and senescence level should be studied for other fresh produces in further studies. Additionally, a simple and nondestructive detection technique of ABA should also be developed for practical use of ABA as freshness marker in postharvest management.
86 4.4 Conclusion
Herein, the first use of HPLC/ESI-MS/MS based metabolomics approach to identify markers of freshness in stored soybean sprouts was reported. ABA was identified as a metabolite that can indicate the degree of soybean sprouts freshness. Although ABA has been associated previously with responses to abiotic stresses such as moisture loss, no studies suggest the use of ABA as a metabolite biomarker for freshness. Hence, the present data are the first to suggest the utility of ABA as a marker for freshness of soybean sprouts, particularly because ABA was absent in freshly harvested sprouts and accumulated during storage. However, these data advise further validation of using ABA as a marker of freshness in soybean sprouts, as well as in other fruits and vegetables.
87