The Study of the Aroma Characteristics of Chinese Mango Cultivars by GC/MS with Solid Phase Microextraction
Zhuo-Min Zhang ,
In this study manual headspace Solid Phase Microextraction (SPME) coupled with Gas Chromatography-Mass Spectrometry (GC-MS) was used to study the aroma characteristics of three kinds of Chinese mango cultivars (xiangya, xiaoxiang and jidan mangoes). In total, twenty-four aroma volatiles, including alkene, alkane, alcohol, aldehyde, ketone and aromatic compound were identified. Aroma characteristics of different mango cultivars were specified by Principal Component Analysis (PCA). Three kinds of mango cultivars possessed the typical but different aroma characteristics. It is hoped that the work could provide helpful clues for the mango quality control and specimen discrimination.
Mango (Mangifera indica L.) is a kind of important tropical fruits, originated from the Southeastern Asia (He, 1999). The aggregate yields of mango in the Asia-Pacific region are the largest in the world. The mango industry in China develops faster than any other countries, so China has been one of the major bases of mango industry in the world (Li, 2005). There are many kinds of mango cultivars popular in China. Most of the cultivars possess the strong and attractive aroma (Huang, 2000). Although most of the studies have been undertaken focusing on the aroma characteristics of mango cultivars from India (John et al., 1999), Africa (Sakho et al., 1985; Sakho et al., 1997), Spain (Ibáñez et al., 1998), Australia (Bartley and Schwede, 1987; Lalel et al., 2003a; Lalel et al., 2003b), United States (Malundo et al., 1997; Beaulieu and Jeanne, 2003) and Brazil (Andrade, 2000), there are still few systematical reports focusing on the aroma characteristics of different Chinese mango cultivars. Fruit aroma could be considered as a kind of terminal metabolites (Feng and Zhao, 2001). Different mango cultivars show the obviously various aroma characteristics. The modern chromatographic methodologies such as gas chromatography- mass spectrometry (GC-MS) could interpret the fruit aroma characteristics effectively via corresponding aroma chromatograms.
Headspace method (Malundo et al., 1997), solvent extraction (Li et
al., 1998; Ollé et al., 1998; John et al., 1999) and
simultaneous distillation extraction (Andrade, 2000) are used as the common
sampling methods for mango aroma. The consequent multiple steps such as clean-up
make these classical sampling methods unsuitable for the rapid analysis. Supercritical
fluid extraction (SFE) is an advanced sampling method (Tuan and Ilangantileke,
1997; Morales et al., 1998) for the fruit aroma. However, lacking suitable
solvents for polar analytes and the high expense for the SFE analysis still
limit the application. In recent years, solid phase microextraction (SPME),
developed by Pawliszyn and co-workers (Arthur and Pawliszyn, 1990; Pawliszyn,
1995), has been considered as an excellent pre-sampling method, simple and solvent-saving.
SPME has been widely used in the environmental (Peñalve et al.,
1999), biological (Mills and Walke, 2000), pharmaceutical (Ulrich, 2000) and
the field analyses (Koziel et al., 1999). Especially, headspace solid
phase microextraction (HSSPME) has been considered as the suitable sampling
method for the fruit aroma volatiles (Ibáñez et al., 1998).
Recently, HSSPME has been used to study the aroma volatiles of mangoes (Ibáñez
et al., 1998; Lalel et al., 2003a; Beaulieu and Lea, 2003).
In this study HSSPME was used to sample aroma volatiles from three kinds of Chinese mango cultivars (xiangya, xiaoxiang and jidan mango) followed by GC-MS analysis. An original chromatographic data processing system based on Matlab 6.5 was programmed to manage the chromatographic data to specify the corresponding aroma characteristics based on the principal component analysis (PCA). It is hoped that the study would provide helpful clues for the mango quality control and specimen discrimination.
Materials and Methods
Sample Collection and Preparation
Fresh ripe mangoes were booked from the settled stand of the local largest
wholesale fruit market in Guangzhou. All the samples purchased were selected
for uniformity in size and color. Blemished or diseased fruits were discarded.
The samples were considered fresh as soon as they picked up from the market
and analyzed within 24 h. For each measurement, the fruits of mangoes were randomly
distributed into groups of 3 fruit. Before peeling mangoes were washed with
tap water followed by rinsing with deionized water to get rid of the dirt on
the surface and dried naturally. Then, fifty grams of mango pulp from one group
were homogenized with 30 mL of NaCl solution (0.02 g mL-1) using
a commercial blender. After that, five grams of fruit tissue homogenate were
put in a 15 mL glass vial followed by HSSPME for 1 h. Finally, aroma volatiles
were thermally desorbed by inserting the fiber into the GC injector set at 250EC
in splitless mode for 5 min.
The Hewlett-Packard (HP) 6890 gas chromatography-HP 5973 mass detector system
was used in the study. Chromatographic separation was performed with an HP-VOC
(Agilent Scientific, USA) capillary column (60 m lengthx0.32 mm I.D.x1.8 μm
film thickness) with the following instrumental conditions: Ultra-purified helium
flow 1 mL min-1; injector temperature 250°C; transfer line temperature
280°C; energy of electron 70 eV; oven temperature from 65 to 80°C at
ramp rate of 5EC min-1, 80°C for 1 min, from 80 to 130°C
at ramp rate of 2°C min-1, 130°C for 2 min, from 130 to 160°C
at ramp rate of 3°C min-1 and from 160 to 240°C at ramp rate
of 6°C min-1; The parameters of HP 5973 mass detector were: ion
mass/charge ratio, 20-550 m/z; scan model.
Chromatographic Data Processing System
In this study, an original chromatographic data processing system
based on the Matlab 6.5 was coded to manage the chromatographic data. Wavelet
transform and polynomial smoothing were applied to smooth the chromatograms
in this system. The original data of the aroma chromatograms acquired from the
GC-MS were exported and transformed to an mx2 matrix (m
represented frequencies of MS data-collecting).
||The aroma chromatograms of xiangya (A), xiaoxiang (B) and
jidan (C) mangoes. The marks, and,
to the aroma volatiles with the highest fractions (*, D-Limonene;
The first column in this mx2 matrix represented the time of MS
data-collecting, and the second column was on behalf of corresponding detectors
responses. After normalization, the data of the total chromatograms of all the
investigated samples were merged into an mxn matrix (n
represented the numbers of the aroma chromatograms). Finally, PCA analysis was
based on this mxn matrix. In brief, the data processing system was
an important tool to distill the statistical information from the experimental
data in this study.
Results and Discussion
The Optimization of HSSPME
The experimental conditions of HSSPME potentially influencing the extraction
process included the type of SPME fiber coating, extraction time and ionic strength.
Replicated measurements were performed to improve the sampling efficiency in
the study. The type of SPME fiber coating was crucial to the sampling efficiency.
Some useful and specific factors should be taken into consideration, such as
polarity, matrix, etc. Comparing five common commercial SPME fiber coatings,
100 μm polydimethylsiloxane, 75 μm carboxen-polydimethylsiloxane (CAR-PDMS),
65 μm carbowax-divinylbenzene, 85 μm polyacrylate and 65 μm polydimethylsiloxane-
divinylbenzene (Supelco, Inc., PA, USA), 75 μm CAR-PDMS fiber coatings
could sample more species and amounts of aroma volatiles. Therefore, the recommended
SPME fiber was 75 μm CAR-PDMS in the study. Secondly, extraction time was
also highly influential to the sampling efficiency. In this experiment the different
extraction times (15, 30, 45, 60 and 90 min) were performed to obtain the optimized
||Aroma volatiles of three kinds of Chinese mango cultivars
a Fit value was referred to what degree the
target spectrum matched the standard spectrum in the NIST library (100
relates to a perfect fit).
|The SDs of stable main aroma components (Fraction>0.01%)
were calculated in the table
Due to the complexity of aroma composition, the short sampling time (less than
60 min) resulted in incomplete absorption; however, the longer sampling time
(90 min) aroused competitive absorption and also caused lower sampling efficiency.
Finally, the sampling time of 60 min was preferred in the work. Thirdly, ionic
strength was believed to affect the extraction in HSSPME as analytes tended
to be in the vapor phase. Increasing ionic strength in the solution could reduce
the surface tension and make the analytes volatile to vapor phase easily. However,
the overmuch ionic strength would make the analytes dissolved in the matrix
solution. Compared with other concentrations (0, 0.08, 0.20 and 0.35 g mL-1),
0.02 g mL-1 NaCl solution resulted in the best extraction efficiency.
The Detection of the Aroma Volatiles
The aroma volatiles were identified by matching sample mass spectra with
those of the National Institute of Standards and Technology MS spectral library.
Aroma volatiles were considered identified, when their fit values
of mass spectra were at the default value of 85 or above. The percentage of
the area counts of the identified peaks in the aroma chromatogram was more than
92.97% to the total area counts of the peaks in the TIC.
||PCA for the various aroma characteristics of xiangya (◊),
xiaoxiang (*) and jidan (
)mango cultivars. The marks ■ ▲
represented the corresponding unknown samples of xiangya, xiaoxiang and
jidan mango cultivars purchased from the local market
Typical aroma chromatograms of three kinds of Chinese mango cultivars were
shown in Fig. 1.
Table 1 showed the 24 identified aroma volatiles from three various Chinese mango cultivars (xiangya, xiaoxiang and jidan mangoes) thermally desorbed from the SPME fiber coating, which could be divided into six groups such as alkene, alkane, alcohol, aldehyde, ketone and aromatic compound. Alkenes consisted of the major aroma characteristics of mango. Jidan mango possessed the less identified aroma volatiles desorbed from the SPME fiber than xiangya and xiaoxiang mangoes. Many kinds of alkenes identified in the mango aroma composition, such as α-pinene, β-myrcene, carene, α-phellandrene, caryophyllene, limonene, have been considered as the important terpenes from mango aroma in the previous reports (Andrade et al., 2000; Lalel et al., 2003a). The major aroma volatiles of three kinds of mangoes were different. Xiangya mango possessed the highest fraction of 1-methyl-4-(1-methylethyl)-1,3-cyclohexadiene whereas xiaoxiang and jidan mangoes possessed the highest fractions of D-limonene in the aroma characteristics. Limonene was a natural and functional monoterpene and possessed a lot of physiological functions, especially its strong anticancer activity (Wang, 2005). Also, limonene has been identified as one of the major aroma volatile of Brazilian mango cultivars (Andrade et al., 2000). Besides alkenes, there were several other aroma volatiles in the aroma composition, such as 7-oxabicyclo[4.1.0]heptane and 4-methoxy-2,5-dimethyl-3[2H]-furanone, which have not been identified before.
When the fractions of aroma volatiles were more than 0.01%, they possessed
the stable present frequencies. Therefore, the reproducibility of aroma characteristics
could be evaluated from the standard deviation (SD, n = 7). From the SDs listed
in Table 1, the corresponding relative standard deviations
(RSDs) could be calculated with the satisfying range from 0.7% ((R)-1-methyl-4-(1-methylethyl)-cyclohexene
of xiaoxiang mango) to 12.5% (α-cubebene of jidan mango). The fluctuation
of the retention time of all the identified peaks was within 0.05 min.
PCA for the Various Aroma Characteristics of Three Mango Cultivars
Table 1 suggested that the aroma compositions of three
kinds of mango cultivars were various. To statistically specify the difference
of the aroma characteristics, a PCA model was established to study the chromatographic
data of xiangya, xiaoxiang and jidan mangoes. In the PCA model the different
clustering principles of three mango cultivars emerged, which suggested that
the aroma characteristics of xiangya, xiaoxiang and jidan mangoes were obviously
various. Due to the completeness of chromatographic data inducing to the PCA,
the clustering variety could be considered caused by the variety of entire aroma
characteristics but not individual aroma volatiles. Therefore, the aroma characteristics
of different mango cultivars could be effectively specified by PCA. Figure
2 showed the PCA clustering rules of aroma characteristics of xiangya, xiaoxiang
and jidan mangoes. After the PCA model was established, the fresh ripe mangoes
were purchased from the local market and analyzed as the above procedure. The
chromatographic data of three kinds of Chinese mango cultivars could fall into
the corresponding PCA segregations. The preliminary results assumed that various
aroma characteristics might reflect the diversity of xiangya, xiaoxiang and
jidan mangoes and could be used as the potential markers for the mango quality
control and specimen discrimination.
An HSSPME method was developed to study various aroma characteristics of three kinds of Chinese mango cultivars (xiangya, xiaoxiang and jidan mangoes) followed by GC-MS detection. In total, twenty-four aroma volatiles were identified. The aroma characteristics of xiangya, xiaoxiang and jidan mangoes were specified by PCA. Three kinds of Chinese mango cultivars possessed the different aroma characteristics. It is hoped that the work could provide helpful clues for the mango quality control and specimen discrimination. The next phase work would focus on quantifying the important aroma volatiles and further revealing the potential biomarkers of Chinese mango cultivars for the specimen discrimination.
The authors would like to thank the National Natural Science Foundation of China for financially supporting the research under contact No. 20575081. Special thanks to Mr. Gui-Hua Ruan for his useful advices on the chemometic study.
1: Andrade, E.H.A., J.G.S. Maia, M. Das and G.B. Zoghbi, 2000. Aroma volatile constituents of Brazilian varieties of mango fruit. J. Food Comp. Anal., 13: 27-33.
2: Arthur, C.L. and J. Pawliszyn, 1990. Solid phase microextraction with thermal desorption using fused silica optical fibers. J. Anal. Chem., 62: 2145-2148.
CrossRef | Direct Link |
3: Bartley, J.P. and A. Schwede, 1987. Volatile flavour components in the headspace of the Australian or ABowen' mango. J. Food Sci., 52: 353-355.
4: Beaulieu, J.C. and J.M. Lea, 2003. Volatile and quality changes in fresh-cut mangos prepared from firm-ripe and soft-ripe fruit, stored in clamshell containers and passive MAP. Postharvest Biol. Technol., 30: 15-28.
5: Feng, S.Q. and Y.M. Zhao, 2001. The Change of Postharvest Physiology of Fruits and Vegetables. Chemical Industry Press, Beijing, China, pp: 1-15.
6: He, P., 1999. The advances on the study of mango breeding. Trop. Agric. Sci., 2: 60-63.
7: Huang, G.D., 2000. The actual state and developing trend of mango breeding in China. China Fruits, 3: 47-49.
8: Ibanez, E., S. Lopez-Sebastian, E. Ramos, J. Tabera and G. Reglero, 1998. Analysis of volatile fruit components by headspace solid-phase microextraction. Food Chem., 63: 281-286.
9: John, K.S., L.J.M. Rao, S.G. Bhat and U.J.S.P. Rao, 1999. Characterization of aroma components of sap from different Indian mango varieties. Phytochemistry, 52: 891-894.
10: Koziel, J., M.Y. Jia, A. Khaled, J. Noah and J. Pawliszyn, 1999. Field air analysis with SPME device. Anal. Chim. Acta, 400: 153-162.
11: Lalel, H.J.D., Z. Singh and S.C. Tan, 2003. Aroma volatiles production during fruit ripening of Kensington Pride mango. Postharvest Biol. Technol., 27: 323-336.
CrossRef | Direct Link |
12: Lalel, H.J.D., Z. Singh and S.C,. Tan 2003. Glycosidically-bound aroma volatile compounds in the skin and pulp of Kensington Pride mango fruit at different stages of maturity. Postharvest Biol. Technol., 29: 205-218.
13: Li, Q.C., L. Qiu and J. Xiong, 1998. Analysis of aromatic components in mangos from Nujiang Basin of Yunnan Province. Chinese J. Chromatogr., 16: 268-270.
14: Li, Y., 2005. The prediction for the yield and requirement of the major tropical fruits in the world. World Trop. Agric. Inform., 8: 4-6.
15: Malundo, T.M.M., E.A. Baldwin, M.G. Moshonas, R.A. Baker and R.L. Shewfelt, 1997. Method for the rapid headspace analysis of mango (Mangifera indica L.) homogenate volatile constituents and factors affecting quantitative results. J. Agric. Food Chem., 45: 2187-2194.
16: Mills, G.A. and V. Walke, 2000. Headspace solid-phase microextraction procedures for gas chromatographic analysis of biological fluids and materials. J. Chromatogr., 902: 267-287.
Direct Link |
17: Morales, M.T., A.J. Berry, P.S. McIntyre and R. Aparicio, 1998. Tentative analysis of virgin olive oil aroma by supercritical fluid extraction-high-resolution gas chromatography-mass spectrometry. J. Chromatogr., 819: 267-275.
18: Olle, D., R.L. Baumes, C.L. Bayonove, Y.F. Lozano, C. Sznaper and J.M. Brillouet, 1998. Comparison of free and glycosidically linked volatile components from polyembryonic and monoembryonic mango (Mangifera indica L.) cultivars. J. Agric. Food Chem., 46: 1094-1100.
19: Pawliszyn, J., 1995. New directions in sample preparation for analysis of organic compounds. Trends Anal. Chem., 14: 113-122.
20: Penalve, A., E. Pocurull, F. Borrull and R.M. Marce, 1999. Trends in solid-phase microextraction for determining organic pollutants in environmental. Trends Anal. Chem., 18: 557-568.
21: Sakho, M., J. Crouzet and S. Seck, 1985. Volatile components of African mango. J. Food Sci., 50: 548-550.
22: Sakho, M., D. Chassagne and J. Crouzet, 1997. African mango glycosidically bound volatile compounds. J. Agric. Food Chem., 45: 883-888.
23: Tuan, D.Q. and S.G. Ilangantileke, 1997. Liquid CO2 extraction of essential oil from star anise fruit (Illicium verum H.). J. Food Eng., 31: 47-57.
24: Ulrich, S., 2000. Solid-phase microextraction in biomedical analysis. J. Chromatogr., 902: 167-194.
Direct Link |
25: Wang, W.J., 2005. Recent advances on limonene, a natural and active monoterpene. China Food Additives, 1: 33-37.