- Email: [email protected]
Elemental imaging and classifying rice grains by using laser ablation inductively coupled plasma mass spectrometry and linear discriminant analysis
Journal of Cereal Science 71 (2016) 198e203
Contents lists available at ScienceDirect
Journal of Cereal Science journal homepage: www.elsevier.com/locate/jcs
Elemental imaging and classifying rice grains by using laser ablation inductively coupled plasma mass spectrometry and linear discriminant analysis J. Promchan a, *, D. Günther b, A. Siripinyanond a, J. Shiowatana a a b
Department of Chemistry and Center for Innovation in Chemistry, Faculty of Science, Mahidol University, Rama VI Rd., Bangkok 10400, Thailand ETH Zurich, Department of Chemistry and Applied Biosciences, Laboratory of Inorganic Chemistry, Wolfgang-Pauli-Str. 10, 8093 Zurich, Switzerland
a r t i c l e i n f o
a b s t r a c t
Article history: Received 28 May 2016 Received in revised form 12 August 2016 Accepted 28 August 2016 Available online 30 August 2016
This study aims to investigate elemental imaging in a longitudinal section of single rice grain using laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) and to classify rice according to their origins and types of 16 samples using LA-ICP-MS with linear discriminant analysis (LDA). The distributions of 8 essential elements (Ca, Cu, Fe, K, Mg, Mn, P and Zn) in a single rice grain were visualized as elemental images. Investigation of the elemental imaging of rice grain showed that essential elements were presented in large amounts in embryo and elevated level in endosperm. The elemental distributions of rice grain were not uniform. In addition, the concentration of 20 elements distributed in core endosperm was evaluated and used as chemical indicator to discriminate the origin and type of rice samples. The LDA can successfully differentiate rice samples according to their regions of origin (Northeast or South regions of Thailand) and types. Satisfied classifications are obtained with overall correct classification and cross-validation of 93.8% and 91.1% for origin classification and 100% and 97.9% for type classification. © 2016 Elsevier Ltd. All rights reserved.
Keywords: Elemental imaging Rice grain LA-ICP-MS LDA
1. Introduction Rice (Oryza sativa L.) is the most important food crop in the world, providing over 21% of the calorie intake needs of the world's population and up to 76% for the population of South East Asia (Fitzgerald et al., 2009). It is an important source of energy, vitamins, essential elements and rare proteins for human. Brown rice is composed of four distinct tissues including bran (6e7% w/w), aleurone layer and embryo (2e3% w/w) and endosperm (90% w/w) (Chen et al., 1998). The bran is outermost part which is rich in oil containing oryzanol, tocotrinols, proteins, vitamin, and essential elements (Basnet et al., 2014; Shin et al., 1997). The aleurone layer and embryo are in inner grain which is rich of protein (6e12%w/w) and starch (Glimn-Lacy and Kaufman, 2006). Endosperm which consists of aleurone layer formed outer layer and starchy part, is rich in starch and proteins (Basnet et al., 2014). The nutrients mainly exist in embryo and bran layer which is minor content in rice grain.
* Corresponding author. E-mail address: [email protected] (J. Promchan). http://dx.doi.org/10.1016/j.jcs.2016.08.017 0733-5210/© 2016 Elsevier Ltd. All rights reserved.
Micronutrient deficiencies involving in iron (Fe) and zinc (Zn) are the most prevalent deficiency-related health disorders in the world. Nearly 3.7 billion people worldwide were iron-deficient and the problem was severe enough to cause anemia in 2 billion people (Gregorio et al., 1999). Rice grains which are main food, can be an important source of mineral for human. Therefore, many attentions are to increase amount and bioavailability of essential elements in rice grains. Elemental distribution information is important for improving rice quality. It provides the answer of the information on basic questions for biomedical research as well as enables bioaccumulation and bioavailability studies for ecological and toxicological risk assessment in humans, animals, and plants (Becker et al., 2014). To gain elemental distribution information, several analytical techniques have been used such as X-ray fluorescence spectrometry (Gholap et al., 2010), synchrotron radiation X-ray fluorescence spectrometry (Wang et al., 2010), scanning electromicroscopyzquez et al., 2013), glow energy dispersive X-ray spectrometry (Va discharge optical emission spectrometry (Gamez et al., 2012), auger nchez-Amaya et al., 2012), secondary ion electron spectroscopy (Sa mass spectroscopy (Sui et al., 2015), and LA-ICP-MS (Wu and Becker, 2012). These techniques provide high spatial resolution
J. Promchan et al. / Journal of Cereal Science 71 (2016) 198e203
(nm-mm) with good sensitivities (mg/g or ng/g) and possible quantification. In recent years, bioimaging using LA-ICP-MS has been focused due to its advantages such as minimal sample preparation, high sample throughput, access to isotopic information, the capability to quantify trace elements with high spatial resolution (1e100 mm), and the possibility of analyzing both conductive and non-conductive and opaque and transparent materials (Becker et al., 2014; Günther and Hattendorf, 2005). The LA-ICP-MS with spot ablation analysis was carried out to characterize the localization of arsenic (As) across the rice grain due to uniform carbon (C) signal in longitudinal section of single polished (white) and unpolished (brown) rice grain. The lowest As accumulation was found in core endosperm. However, the quantitative analysis for As accumulation was not performed (Meharg et al., 2008). LA-ICP-MS has been used to observe iron distribution in the endosperm of transgenic rice using single line scan ablation. The iron was accumulated in spots (Wirth et al., 2009). Spatial distributions of trace elements (As, Cd, Pb, Sb and Zn) in single rice grain from contaminated rice fields were imaged by using LA-ICP-MS (Basnet et al., 2014). Nowadays, there is growing interest in research related to identification of the geographic origin of a wide range of agricultural food products. The determination of food authenticity is important issue in quality control and safety of food. The applications of multi-element analysis for discrimination of rice grain have been used to classify the origin or type using several analytical methods and data interpretations. The inductively coupled plasma atomic emission spectrometry (ICP-AES), ICP-MS and flame atomic absorption spectrometry (FAAS) with principal component analysis (PCA) were used to classify various rice samples collected in Australia and Vietnam by 14 elements (Al, As, Ca, Cd, Cu, Fe, K, Mn, Mo, Na, Ni, P, and Zn) (Kokot and Dong Phuong, 1999). The elemental analyzer/isotope ratio mass spectrometry (EA/IRMS) with radar plot was carried out to classify rice from Australia, Japan and USA (Suzuki et al., 2008). The high resolution inductively coupled plasma mass spectrometry (HR-ICP-MS) in coupled with radar plot, PCA and discriminant analysis (DA) was used for the discrimination of the origin of rice samples in Thai jasmine rice samples and foreign rice samples from France, India, Italy, Japan and Pakistan by 21 elements (Al, As, B, Ba, Cd, Co, Cr, Cs, Cu, Fe, Mg, Mn, Mo, Ni, Pb, Rb, Se, Sr, Ti, V and Zn) (Cheajesadagul et al., 2013). The ICP-AES and chemometrics with PCA and partial least-squares discriminant analysis (PLS-DA) was carried out to determine the authenticity of the geographical origin of rice from China, Korea and Philippines by 11 elements (Ag, Ba, Bi, Ca, Cd, Cr, Cu, In, K, Pb and Zn) (Chung et al., 2015). Solution based analytical techniques were carried out to quantify multielement composition in rice grain samples for discriminant analysis. Interestingly, solid based analytical techniques such as LA-ICP-MS has been successfully reported to provide element concentration of various samples due to its advantages over other techniques including direct analysis of solid sample, reduced laboratory preparation and ability to provide information on trace element isotopes (Basnet et al., 2016; Cheajesadagul et al., 2011; Halicz and Günther, 2004; M-M et al., 2011). Thus, LA-ICP-MS is a good option for quantification of multielement composition in rice grain samples for discriminant analysis. In Thailand, rice product is a major agriculture which provides around 30 million tons per year. Major rice agriculture regions are the Northeast and South. The two third of rice production is reserved for nationwide consumption while the rest is exported worldwide. Some rice varieties are more popular and thus more expensive than the others which are grown in specific area in the Northeast while some specific variety is only grown in the South. Therefore, it is essential to investigate the elemental and authenticity information in rice grain from the Northeast and South
199
regions of Thailand. To obtain information of elemental images and authenticity in rice grain, our purposes are (1) to investigate elemental distribution in single rice grain containing embryo and endosperm by elemental imaging using LA-ICP-MS and (2) to classify rice grain samples from the Northeast (NE) and the South (S) regions of Thailand according to their origins and types using data obtained from LA-ICP-MS and LDA. From our knowledge, this is the first time that the information obtained from LA-ICP-MS has been used to classify origin and type of rice grain samples. 2. Materials and methods 2.1. Sample information and preparation For elemental imaging, the white rice grain sample from NE was cut in a longitudinal section by ceramic knife. For elemental composition in core endosperm, 16 rice grain samples from 2 regions of Thailand were collected including 8 samples from NE and 8 samples from S, which were separated by types according to their colors. There are 8 white rice samples from NE and 5 white, 1 black, 1 red and 1 yellow rice samples from S. The rice grain samples were embedded in 2.5 cm diameter of resin. EpoKwick resin (Buehler, Lakebruff, IL, USA) was used. The rice grain samples were placed on sticky tape then covered with resin block. The Epo-Kwick solution was added into resin block for overnight to obtain rigid resin. The block was polished by polishing system, LaboPol-5 (Struers, Ballerup, Denmark). The samples were polished to get flat surface at the middle of grain readily to be analyzed by LA-ICP-MS. 2.2. Quantification Quantification was obtained by using rice flour standard reference material, NIST 1568a, from the National Institute of Standard and Technology (Gaithersburg, MD, USA) as an external standard which is capable of providing acceptable accuracy and precision. Matrix matched standard is used to correct fractionation effects and also compensate for matrix effects produced in the ICP-MS (Hare et al., 2012). The external standard was prepared by mixing with Teflon powder (VHG Labs, Manchester, NH, USA) and pressing as a pellet. For bioimaging, 13C was used as an acceptable internal standard element to compensate the conditions change (Hare et al., 2012). ICP-MS data obtained after the ablation were exported to Microsoft Office Excel 2007 (Microsoft Cooperation) for all data treatment. Data reduction was performed using method reported by Longerich et al. (Longerich et al., 1996). The concentration of the analyte element in sample (CA,SAM) is given by the count rate for the analyte in the sample (RA,SAM) divided by the normalized sensitivity (S), as follows:
CA;SAM ¼
RA;SAM S
The normalized sensitivity (S) is the sensitivity, determined on a calibration standard (STD), corrected for the volume (mass) of sample ablated. When using naturally occurring internal standards, the sensitivity (cps per unit of concentration) normalized to the mass of the sample (SAM) ablated in the determinations is:
S¼
RA;STD RIS;SAM CIS;STD CA;STD RIS;STD CIS;SAM
where RA,STD is the count rate of the analyte in the standard
200
J. Promchan et al. / Journal of Cereal Science 71 (2016) 198e203
material; CA,STD is the concentration of the analyte in the standard material; RIS,SAM is the count rate of the internal standard in the sample; RIS,STD is the count rate of the internal standard in the standard material; CIS,SAM is the concentration of the internal standard in the sample; and CIS,STD is the concentration of the internal standard in the standard material. 2.3. Elemental imaging The longitudinal rice grain section was rastered by linear laser ablation scans over the entire surface. The operating condition is shown in Table 1 (Method I). The Matrix Laboratory (MATLAB) software version 7.8.0.347 (R2009a) was used to build-up the im^a et al., 2016). ages of rice grains (Pesso 2.4. Multi-element composition for classification Multi-element composition (twenty variables: Al, As, Br, Ca, Cd, Cl, Co, Cu, Fe, Hg, K, Mg, Mn, Mo, Na, P, Pb, Rb, S, and Zn) in core endosperm was analyzed by using single line scan mode of laser ablation. Three rice grains of each sample were ablated (4 lines for one grain). Total 12 data of each rice sample were used for classification. The operating condition is shown in Table 1 (Method II). 2.5. Classification The multivariate statistical evaluation of data was performed using SPSS 16.0 software. The LDA was used for region and color classification. The method was based on maximizing the variance within between groups and minimizing the variance within group by creating new variables (discriminant factors) which are linear combinations of the original variables. Classification success was given by the ratio of the correctly classified objects to their total number as percentage of correct classification.
3. Results and discussion 3.1. Elemental imaging To obtain elemental imaging of the white rice sample from NE, macro elements and micro elements in single longitudinal rice grain section containing endosperm and embryo was analyzed by LA-ICP-MS. The concentration of elements was calculated as described above. The high concentration of macro elements including Ca, K, Mg, and P and micro elements including Cu, Fe, Mn, and Zn was mainly found in embryo as in Fig. 1. It is clearly shown that the concentrations of macro elements were high in only embryo, but they were low in endosperm as in Table 2. Among macro elements in embryo, the highest concentration was found for P at 7.15 mg/g while the concentration of K, Mg and Ca were detected at 5.04, 2.95 and 0.31 mg/g, respectively. The concentration ranking in endosperm was K, P, Ca and Mg with average concentration as 0.95, 0.53, 0.07, and 0.04 mg/g, respectively. For micro elements shown in Fig. 1, the concentration ranking in embryo was Zn, Mn, Fe and Cu with average concentration as 83.00, 40.18, 29.91, and 1.58 mg/g, respectively, while those of concentrations in endosperm were obviously decreased. The concentration ranking in endosperm was Zn, Mn, Fe and Cu with average concentration as 30.18, 9.19, 1.39, and 0.50 mg/g, respectively. The elemental distribution patterns in embryo and endosperm were not uniform. Heterogeneous distribution of Fe across rice grain was observed by using LA-ICP-MS and XRF (Wirth et al., 2009). Elements in embryo were found in high concentrations. The highest concentration of Cu, Fe, Mn and Zn in rice grain was found in embryo (Hansen et al., 2012). Quantified elemental bioimages showed that the highest concentration of Zn presented in embryo while elevated levels of Zn were observed in endosperm (Basnet et al., 2014, 2016). The concentration of several elements such as Ca, Cu, Fe, Mg, Mn, P and Zn was decreased by the increase degree of milling and polishing (Wang et al., 2011). Interestingly, our results also showed that the distribution patterns
Table 1 LA-ICP-MS operating conditions for rice grain analysis. Method Laser ablation Model Laser wavelength (nm) Crater diameter (mm) Repetition rate (Hz) Scanning speed (mm/s) Sampling scheme Carrier gas flow rate (L/ min) Length of ablation (mm) ICP-MS Model Rf power (W) Nebulizer gas flow rate (L/ min) Coolant gas flow rate (L/ min) Auxiliary gas flow rate (L/ min) Autolens Detector mode Scanning mode Dwell time (ms) Settling time (ms) Number of sweeps per reading Number of replicate Isotopes monitored (m/z)
I
II
CETAC LSX-500 266 100 10 50 Line scan analysis 1.0 (He)
GeoLas C 193 95 10 20 Line scan analysis 1.0 (He)
6750
1200
Perkin Elmer ELAN 6000 1400 0.90
Perkin Elmer ELAN 6100 DRCþ 1400 0.95
15.0
17.8
0.90
0.80
On Dual Peak hopping 10 3 1
On Dual Peak hopping 10 3 1
1
1
13 66
C, 25Mg, 31P, 39K, 43Ca, 55Mn, 57Fe, 65Cu, Zn
13
C, 23Na, 25Mg, 27Al, 31P, 32S, 35Cl, 39K, 43Ca, 55Mn, 57Fe, 59Co, 65Cu, 66Zn, 75As, 79Br, 85Rb, 98Mo, 111Cd, Hg, 208Pb
202
J. Promchan et al. / Journal of Cereal Science 71 (2016) 198e203
201
Fig. 1. Elemental imaging of the single longitudinal rice grain section of the white rice sample from NE by LA-ICP-MS for macro elements (Ca, Mg, K and P) and micro elements (Cu, Fe, Mn and Zn).
of trace elements in endosperm were not uniform. The high concentration of Cu and Mn was found at the bran layer and in the middle of endosperm, respectively while that of Zn was distributed from the bran layer through the middle of endosperm. These distribution patterns suggest that essential elements can be lost during polishing process. In this work, 100 mm diameter sized laser ablation crater and 50 mm/s scan speed were used for bio-imaging. With this condition, suitable detection limits (DL) of the interested elements were obtained to monitor elemental distribution in embryo and endosperm such as Fe. However, this crater diameter condition was not the best choice in term of high quality and resolution of bio-images but it was suitable for monitoring elemental distribution in rice grain. 3.2. Elemental composition for classification Sixteen rice grain samples were analyzed and quantified by LAICP-MS. Twenty variables including Al, As, Br, Ca, Cd, Cl, Co, Cu, Fe, Hg, K, Mg, Mn, Mo, Na, P, Pb, Rb, S, and Zn in core endosperm were
Table 2 Average element concentrations in embryo and endosperm of the white rice grain sample from NE according to Fig. 1. Element mg/g Ca K Mg P mg/g Cu Fe Mn Zn
Embryo 0.31 5.04 2.95 7.15
± ± ± ±
0.52 2.05 1.09 3.28
1.58 ± 0.96 29.91 ± 12.91 40.18 ± 17.01 83.00 ± 51.49
Endosperm 0.07 0.95 0.04 0.53
± ± ± ±
0.06 0.79 0.03 0.33
0.50 ± 1.67 1.39 ± 1.72 9.19 ± 4.49 30.18 ± 18.75
investigated. The mean values and standard deviations of element concentrations in rice samples according to their types and origins are summarized in Table 3. It was possible to use the concentration of elements for the discrimination of types and origins of rice
202
J. Promchan et al. / Journal of Cereal Science 71 (2016) 198e203
samples. On the basis of this information, all rice samples have similar element concentration ranges. The concentration of Cl in white rice from NE and S was higher than red and yellow rice. Interestingly, the concentration of K in yellow rice was higher than others. The concentration of As, Br and Fe was lower than the detection limit in most rice samples whereas Cd was lower than the detection limit in all rice samples. 3.3. Origin classification The concentration of 20 elements was analyzed with 16 rice grain samples and used as a chemical descriptor in statistical analysis for the discrimination of rice grains according to their geographic origins. As shown in Table 3, the concentration of Br, Cd and Fe in rice samples was below the detection limit of LA-ICP-MS, the concentration values of these elements were replaced with their detection limits to compensate statistical process. In this study, LDA was used to classify the geographical origins of rice samples. After LDA analysis, discriminant function (F1) correlated with the variables was obtained. The variables affecting the classification were Br and Cl as positive coefficients and Mg as negative coefficient, respectively (>0.700 of standardized canonical discriminant function coefficients, CF). The separation between regional groups including NE and S in the discriminant plot was clearly distinguished as shown in Fig. 2. It is suggested that rice samples from different regions were classified by LDA analysis. To characterize the reliability of the developed classification model, cross-validation method was operated to compute the classification and probability of rice samples. The results enable a satisfied classification with an overall correct classification of 93.8% and a cross-validation of 91.1%.
Fig. 2. Scatter plot of individual rice sample scores on the discriminant function for origin classification. NE¼Northeast; S¼South.
obtained. The three dimensional scatter plot of individual rice samples was resulted in the distribution pattern of 8 rice grain samples according to their types defined by discriminant functions (Fig. 3). The variations between groups were explained by the discriminant function 1 (92.1%), 2 (5.6%) and 3 (2.3%), respectively. The variables affecting the classification were considered by large coefficients of each discriminant function. CF1 was mainly related to Br content as strong positive coefficients; CF2 was related to K, Na, Rb and S content as positive coefficients and Cu, Mg and Mo content as negative coefficients; and CF3 was related to Rb and S as positive coefficients and P concentration as the negative coefficient (>0.700 of standardized canonical discriminant function coefficients, CF). The analysis of rice samples provided a satisfied classification with an overall correct classification of 100% and a cross-validation of 97.9%. However, there were limited numbers of rice samples in this work. This study showed potential for rice type discrimination as shown in Fig. 3. For the rigorous statistical analysis, adequate sample number will be needed for further investigation.
3.4. Type classification The classification of 8 rice samples in different types including white, black, red and yellow rice grains from the South region was analyzed using 20 variables. After applying LDA, discriminant functions (F1, F2 and F3) correlated with the variables were
Table 3 Element concentrations in core endosperm of 16 rice grain samples reported by rice types. Element mg/g Cl K P S mg/g Al As Br Ca Cd Cu Fe Mg Mn Mo Na Rb Zn ng/g Co Hg Pb
Whitea (n ¼ 8) 0.23 0.54 0.47 1.04
± ± ± ±
0.08 0.15 0.16 0.43
Whiteb (n ¼ 5) 0.24 0.71 0.63 1.07
± ± ± ±
0.08 0.29 0.12 0.49
Blackb (n ¼ 1) 0.16 0.49 0.50 1.29
± ± ± ±
0.03 0.09 0.20 0.25
Redb (n ¼ 1) 0.13 0.64 0.63 1.29
± ± ± ±
0.02 0.10 0.21 0.25
Yellowb (n ¼ 1) 0.10 0.92 0.67 1.35
± ± ± ±
0.02 0.13 0.09 0.45
DL 0.03 0.0004 0.001 0.01
1.63 ± 2.22 0.28 ± 0.04
0.65 ± 0.25 0.31 ± 0.14
0.78 ± 0.65
0.75 ± 0.38 0.21 ± 0.05
0.66 ± 0.43 0.16 ± 0.01 21.00 ± 2.53 63.38 ± 25.17
0.16 0.06 0.7 27 0.03 0.03 0.3 0.06 0.03 0.007 0.01 0.003 0.08
36.61 ± 19.41 8.14 ± 2.29 16.77 ± 25.95
11.87 ± 3.33 9.50 ± 4.23 8.22 ± 3.30
59.61 ± 43.92 7.88 ± 1.13 21.12 ± 15.39
18.98 ± 10.51 11.88 ± 2.80 10.19 ± 5.03
27.78 ± 22.07 11.25 ± 2.43 19.23 ± 20.11
5 2 2
DL: Detection limit. a Sample from NE region. b Sample from S region.
J. Promchan et al. / Journal of Cereal Science 71 (2016) 198e203
Fig. 3. 3D scatter plot of individual rice sample scores on the first three discriminant functions for type classification of rice grain samples from the South.
4. Conclusions The elemental imaging of a longitudinal section in single rice grain using LA-ICP-MS was investigated. The essential elements were found in large amounts in embryo and in particular area in endosperm such as Zn and Mn in the bottom and the middle parts of endosperm, respectively. The elemental distributions of rice grain were not uniform. Bio-images of essential elements show that brown rice is an important source of essential elements such as Fe and Zn which can be helpful for Fe and Zn deficiency in humans. These findings are in agreement with the previous work (Basnet et al., 2014, 2016). In addition, the classification of rice grains according to their origins and types was successfully performed by LA-ICP-MS and LDA. Satisfied classifications are obtained. Rice grain samples from NE can be separated from S. Moreover, four rice types from S can be clearly separated. As demonstrated in this study, the LA-ICP-MS analytical approach is an alternative approach to investigate the spatial elemental distribution and elemental composition in rice grains, which could be beneficial to further research on seed development, rice cultivation and rice consumption. Acknowledgements The author gratefully acknowledges for the research grants from the Thailand Research Fund (TRF) through the Royal Golden Jubilee Ph.D. Program (RGJ) and the Center of Excellence for Innovation in Chemistry (PERCH-CIC), Commission on Higher Education, Ministry of Education (Thailand). References Basnet, P., Amarasiriwardena, D., Wu, F., Fu, Z., Zhang, T., 2014. Elemental bioimaging of tissue level trace metal distributions in rice seeds (Oryza sativa L.) from a mining area in China. Environ. Pollut. 195, 148e156. Basnet, P., Amarasiriwardena, D., Wu, F., Fu, Z., Zhang, T., 2016. Investigation of tissue level distribution of functional groups and associated trace metals in rice seeds (Oryza sativa L.) using FTIR and LA-ICP-MS. Microchem. J. 127, 152e159. Becker, J.S., Matusch, A., Wu, B., 2014. Bioimaging mass spectrometry of trace elements e recent advance and applications of LA-ICP-MS: a review. Anal. Chim. Acta 835, 1e18. Cheajesadagul, P., Arnaudguilhem, C., Shiowatana, J., Siripinyanond, A., Szpunar, J., 2013. Discrimination of geographical origin of rice based on multi-element fingerprinting by high resolution inductively coupled plasma mass spectrometry. Food Chem. 141, 3504e3509.
203
Cheajesadagul, P., Wananukul, W., Siripinyanond, A., Shiowatana, J., 2011. Metal doped keratin film standard for LA-ICP-MS determination of lead in hair samples. J. Anal. Atomic Spectrom. 26, 493e498. Chen, H., Siebenmorgen, T.J., Griffin, K., 1998. Quality characteristics of long-grain rice milled in two commercial systems. Cereal Chem. 75, 560e565. Chung, I.-M., Kim, J.-K., Lee, J.-K., Kim, S.-H., 2015. Discrimination of geographical origin of rice (Oryza sativa L.) by multielement analysis using inductively coupled plasma atomic emission spectroscopy and multivariate analysis. J. Cereal Sci. 65, 252e259. Fitzgerald, M.A., McCouch, S.R., Hall, R.D., 2009. Not just a grain of rice: the quest for quality. Trends Plant Sci. 14, 133e139. Gamez, G., Voronov, M., Ray, S.J., Hoffmann, V., Hieftje, G.M., Michler, J., 2012. Surface elemental mapping via glow discharge optical emission spectroscopy. Spectrochim. Acta Part B At. Spectrosc. 70, 1e9. Gholap, D.S., Izmer, A., Samber, B.D., Elteren, J.T.v., Selih, V.S., Evens, R., Schamphelaere, K.D., Janssen, C., Balcaen, L., Lindemann, I., Vincze, L., Vanhaecke, F., 2010. Comparison of laser ablation-inductively coupled plasmamass spectrometry and micro-X-ray fluorescence spectrometry for elemental imaging in Daphnia magna. Anal. Chim. Acta 664, 19e26. Glimn-Lacy, J., Kaufman, P.B., 2006. Seed structure and germination. In: Botany Illustrated, second ed. Springer Science, New York. Gregorio, G.B., Senadhira, D., Htut, T., 1999. Improving iron and zinc value of rice for human nutrients. Agric. Dev. 23, 68e87. Günther, D., Hattendorf, B., 2005. Solid sample analysis using laser ablation inductively coupled plasma mass spectrometry. TrAC Trends Anal. Chem. 24, 255e265. Halicz, L., Günther, D., 2004. Quantitative analysis of silicates using LA-ICP-MS with liquid calibration. J. Anal. Atomic Spectrom. 19, 1539e1545. Hansen, T.H., Lombi, E., Fitzgerald, M., Laursen, K.H., Frydenvang, J., Husted, S., Boualaphanh, C., Resurreccion, A., Howard, D.L., de Jonge, M.D., Paterson, D., Schjoerring, J.K., 2012. Losses of essential mineral nutrients by polishing of rice differ among genotypes due to contrasting grain hardness and mineral distribution. J. Cereal Sci. 56, 307e315. Hare, D., Austin, C., Doble, P., 2012. Quantification strategies for elemental imaging of biological samples using laser ablation-inductively coupled plasma-mass spectrometry. Analyst 137, 1527e1537. Kokot, S., Dong Phuong, T., 1999. Elemental content of Vietnamese ricePart 2. Multivariate data analysis. Analyst 124, 561e569. Longerich, H.P., Jackson, S.E., Günther, D., 1996. Inter-laboratory note. Laser ablation inductively coupled plasma mass spectrometric transient signal data acquisition and analyte concentration calculation. J. Anal. Atomic Spectrom. 11, 899e904. M-M, P., Hansawek, R., Shiowatana, J., Siripinyanond, A., 2011. Geographical origin classification of gem corundum using elemental fingerprint analysis by laser ablation inductively coupled plasma mass spectrometry. Int. J. Mass Spectrom. 306, 57e62. Meharg, A.A., Lombi, E., Williams, P.N., Scheckel, K.G., Feldmann, J., Raab, A., Zhu, Y., Islam, R., 2008. Speciation and localization of arsenic in white and brown rice grains. Environ. Sci. Technol. 42, 1051e1057. ^ a, G.d.S., Capelo-Martínez, J.L., Fdez-Riverola, F., Lo pez-Ferna ndez, H., GlezPesso ~ a, D., Reboiro-Jatoef, M., Arruda, M.A.Z., 2016. Laser ablation and inductively Pen coupled plasma mass spectrometry focusing on bioimaging from elemental distribution using MatLab software: a practical guide. J. Anal. Atomic Spectrom. 31, 832e840. S anchez-Amaya, J.M., Blanco, G., Garcia-Garcia, F.J., Bethencourt, M., Botana, F.J., 2012. XPS and AES analyses of cerium conversion coatings generated on AA5083 by thermal activation. Surf. Coatings Technol. 213, 105e116. Shin, T.-S., Godber, J.S., Martin, D.E., Wells, J.H., 1997. Hydrolytic stability and changes in E vitamers and oryzanol of extruded rice bran during storage. J. Food Sci. 62, 704e728. Sui, T., Song, B., Dluhos, J., Lu, L., Korsunsky, A.M., 2015. Nanoscale chemical mapping of Li-ion battery cathode material by FIB-SEM and TOF-SIMS multi-modal microscopy. Nano Energy 17, 254e260. Suzuki, Y., Chikaraishi, Y., Ogawa, N.O., Ohkouchi, N., Korenaga, T., 2008. Geographical origin of polished rice based on multiple element and stable isotope analyses. Food Chem. 109, 470e475. zquez, M.A., Gala n, E., Ortiz, P., Ortiz, R., 2013. Digital image analysis and EDX SEM Va as combined techniques to evaluate salt damp on walls. Constr. Build. Mater. 45, 95e105. Wang, H.-J., Wang, M., Wang, B., Meng, X.-Y., Wang, Y., Li, M., Feng, W.-Y., Zhao, Y.-L., Chai, Z.-F., 2010. Quantitative imaging of element spatial distribution in the brain section of a mouse model of Alzheimer's disease using synchrotron radiation X-ray fluorescence analysis. J. Anal. Atomic Spectrom. 25, 328e333. Wang, K.M., Wu, J.G., Li, G., Zhang, D.P., Yang, Z.W., Shi, C.H., 2011. Distribution of phytic acid and mineral elements in three indica rice (Oryza sativa L.) cultivars. J. Cereal Sci. 54, 116e121. Wirth, J., Poletti, S., Aeschlimann, B., Yakandawala, N., Drosse, B., Osorio, S., Tohge, T., Fernie, A.R., Günther, D., Gruissem, W., Sautter, C., 2009. Rice endosperm iron biofortification by targeted and synergistic action of nicotianamine synthase and ferritin. Plant Biotechnol. J. 7, 631e644. Wu, B., Becker, J.S., 2012. Imaging techniques for elements and element species in plant science. Metallomics 4, 403e416.
Contents lists available at ScienceDirect
Journal of Cereal Science journal homepage: www.elsevier.com/locate/jcs
Elemental imaging and classifying rice grains by using laser ablation inductively coupled plasma mass spectrometry and linear discriminant analysis J. Promchan a, *, D. Günther b, A. Siripinyanond a, J. Shiowatana a a b
Department of Chemistry and Center for Innovation in Chemistry, Faculty of Science, Mahidol University, Rama VI Rd., Bangkok 10400, Thailand ETH Zurich, Department of Chemistry and Applied Biosciences, Laboratory of Inorganic Chemistry, Wolfgang-Pauli-Str. 10, 8093 Zurich, Switzerland
a r t i c l e i n f o
a b s t r a c t
Article history: Received 28 May 2016 Received in revised form 12 August 2016 Accepted 28 August 2016 Available online 30 August 2016
This study aims to investigate elemental imaging in a longitudinal section of single rice grain using laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) and to classify rice according to their origins and types of 16 samples using LA-ICP-MS with linear discriminant analysis (LDA). The distributions of 8 essential elements (Ca, Cu, Fe, K, Mg, Mn, P and Zn) in a single rice grain were visualized as elemental images. Investigation of the elemental imaging of rice grain showed that essential elements were presented in large amounts in embryo and elevated level in endosperm. The elemental distributions of rice grain were not uniform. In addition, the concentration of 20 elements distributed in core endosperm was evaluated and used as chemical indicator to discriminate the origin and type of rice samples. The LDA can successfully differentiate rice samples according to their regions of origin (Northeast or South regions of Thailand) and types. Satisfied classifications are obtained with overall correct classification and cross-validation of 93.8% and 91.1% for origin classification and 100% and 97.9% for type classification. © 2016 Elsevier Ltd. All rights reserved.
Keywords: Elemental imaging Rice grain LA-ICP-MS LDA
1. Introduction Rice (Oryza sativa L.) is the most important food crop in the world, providing over 21% of the calorie intake needs of the world's population and up to 76% for the population of South East Asia (Fitzgerald et al., 2009). It is an important source of energy, vitamins, essential elements and rare proteins for human. Brown rice is composed of four distinct tissues including bran (6e7% w/w), aleurone layer and embryo (2e3% w/w) and endosperm (90% w/w) (Chen et al., 1998). The bran is outermost part which is rich in oil containing oryzanol, tocotrinols, proteins, vitamin, and essential elements (Basnet et al., 2014; Shin et al., 1997). The aleurone layer and embryo are in inner grain which is rich of protein (6e12%w/w) and starch (Glimn-Lacy and Kaufman, 2006). Endosperm which consists of aleurone layer formed outer layer and starchy part, is rich in starch and proteins (Basnet et al., 2014). The nutrients mainly exist in embryo and bran layer which is minor content in rice grain.
* Corresponding author. E-mail address: [email protected] (J. Promchan). http://dx.doi.org/10.1016/j.jcs.2016.08.017 0733-5210/© 2016 Elsevier Ltd. All rights reserved.
Micronutrient deficiencies involving in iron (Fe) and zinc (Zn) are the most prevalent deficiency-related health disorders in the world. Nearly 3.7 billion people worldwide were iron-deficient and the problem was severe enough to cause anemia in 2 billion people (Gregorio et al., 1999). Rice grains which are main food, can be an important source of mineral for human. Therefore, many attentions are to increase amount and bioavailability of essential elements in rice grains. Elemental distribution information is important for improving rice quality. It provides the answer of the information on basic questions for biomedical research as well as enables bioaccumulation and bioavailability studies for ecological and toxicological risk assessment in humans, animals, and plants (Becker et al., 2014). To gain elemental distribution information, several analytical techniques have been used such as X-ray fluorescence spectrometry (Gholap et al., 2010), synchrotron radiation X-ray fluorescence spectrometry (Wang et al., 2010), scanning electromicroscopyzquez et al., 2013), glow energy dispersive X-ray spectrometry (Va discharge optical emission spectrometry (Gamez et al., 2012), auger nchez-Amaya et al., 2012), secondary ion electron spectroscopy (Sa mass spectroscopy (Sui et al., 2015), and LA-ICP-MS (Wu and Becker, 2012). These techniques provide high spatial resolution
J. Promchan et al. / Journal of Cereal Science 71 (2016) 198e203
(nm-mm) with good sensitivities (mg/g or ng/g) and possible quantification. In recent years, bioimaging using LA-ICP-MS has been focused due to its advantages such as minimal sample preparation, high sample throughput, access to isotopic information, the capability to quantify trace elements with high spatial resolution (1e100 mm), and the possibility of analyzing both conductive and non-conductive and opaque and transparent materials (Becker et al., 2014; Günther and Hattendorf, 2005). The LA-ICP-MS with spot ablation analysis was carried out to characterize the localization of arsenic (As) across the rice grain due to uniform carbon (C) signal in longitudinal section of single polished (white) and unpolished (brown) rice grain. The lowest As accumulation was found in core endosperm. However, the quantitative analysis for As accumulation was not performed (Meharg et al., 2008). LA-ICP-MS has been used to observe iron distribution in the endosperm of transgenic rice using single line scan ablation. The iron was accumulated in spots (Wirth et al., 2009). Spatial distributions of trace elements (As, Cd, Pb, Sb and Zn) in single rice grain from contaminated rice fields were imaged by using LA-ICP-MS (Basnet et al., 2014). Nowadays, there is growing interest in research related to identification of the geographic origin of a wide range of agricultural food products. The determination of food authenticity is important issue in quality control and safety of food. The applications of multi-element analysis for discrimination of rice grain have been used to classify the origin or type using several analytical methods and data interpretations. The inductively coupled plasma atomic emission spectrometry (ICP-AES), ICP-MS and flame atomic absorption spectrometry (FAAS) with principal component analysis (PCA) were used to classify various rice samples collected in Australia and Vietnam by 14 elements (Al, As, Ca, Cd, Cu, Fe, K, Mn, Mo, Na, Ni, P, and Zn) (Kokot and Dong Phuong, 1999). The elemental analyzer/isotope ratio mass spectrometry (EA/IRMS) with radar plot was carried out to classify rice from Australia, Japan and USA (Suzuki et al., 2008). The high resolution inductively coupled plasma mass spectrometry (HR-ICP-MS) in coupled with radar plot, PCA and discriminant analysis (DA) was used for the discrimination of the origin of rice samples in Thai jasmine rice samples and foreign rice samples from France, India, Italy, Japan and Pakistan by 21 elements (Al, As, B, Ba, Cd, Co, Cr, Cs, Cu, Fe, Mg, Mn, Mo, Ni, Pb, Rb, Se, Sr, Ti, V and Zn) (Cheajesadagul et al., 2013). The ICP-AES and chemometrics with PCA and partial least-squares discriminant analysis (PLS-DA) was carried out to determine the authenticity of the geographical origin of rice from China, Korea and Philippines by 11 elements (Ag, Ba, Bi, Ca, Cd, Cr, Cu, In, K, Pb and Zn) (Chung et al., 2015). Solution based analytical techniques were carried out to quantify multielement composition in rice grain samples for discriminant analysis. Interestingly, solid based analytical techniques such as LA-ICP-MS has been successfully reported to provide element concentration of various samples due to its advantages over other techniques including direct analysis of solid sample, reduced laboratory preparation and ability to provide information on trace element isotopes (Basnet et al., 2016; Cheajesadagul et al., 2011; Halicz and Günther, 2004; M-M et al., 2011). Thus, LA-ICP-MS is a good option for quantification of multielement composition in rice grain samples for discriminant analysis. In Thailand, rice product is a major agriculture which provides around 30 million tons per year. Major rice agriculture regions are the Northeast and South. The two third of rice production is reserved for nationwide consumption while the rest is exported worldwide. Some rice varieties are more popular and thus more expensive than the others which are grown in specific area in the Northeast while some specific variety is only grown in the South. Therefore, it is essential to investigate the elemental and authenticity information in rice grain from the Northeast and South
199
regions of Thailand. To obtain information of elemental images and authenticity in rice grain, our purposes are (1) to investigate elemental distribution in single rice grain containing embryo and endosperm by elemental imaging using LA-ICP-MS and (2) to classify rice grain samples from the Northeast (NE) and the South (S) regions of Thailand according to their origins and types using data obtained from LA-ICP-MS and LDA. From our knowledge, this is the first time that the information obtained from LA-ICP-MS has been used to classify origin and type of rice grain samples. 2. Materials and methods 2.1. Sample information and preparation For elemental imaging, the white rice grain sample from NE was cut in a longitudinal section by ceramic knife. For elemental composition in core endosperm, 16 rice grain samples from 2 regions of Thailand were collected including 8 samples from NE and 8 samples from S, which were separated by types according to their colors. There are 8 white rice samples from NE and 5 white, 1 black, 1 red and 1 yellow rice samples from S. The rice grain samples were embedded in 2.5 cm diameter of resin. EpoKwick resin (Buehler, Lakebruff, IL, USA) was used. The rice grain samples were placed on sticky tape then covered with resin block. The Epo-Kwick solution was added into resin block for overnight to obtain rigid resin. The block was polished by polishing system, LaboPol-5 (Struers, Ballerup, Denmark). The samples were polished to get flat surface at the middle of grain readily to be analyzed by LA-ICP-MS. 2.2. Quantification Quantification was obtained by using rice flour standard reference material, NIST 1568a, from the National Institute of Standard and Technology (Gaithersburg, MD, USA) as an external standard which is capable of providing acceptable accuracy and precision. Matrix matched standard is used to correct fractionation effects and also compensate for matrix effects produced in the ICP-MS (Hare et al., 2012). The external standard was prepared by mixing with Teflon powder (VHG Labs, Manchester, NH, USA) and pressing as a pellet. For bioimaging, 13C was used as an acceptable internal standard element to compensate the conditions change (Hare et al., 2012). ICP-MS data obtained after the ablation were exported to Microsoft Office Excel 2007 (Microsoft Cooperation) for all data treatment. Data reduction was performed using method reported by Longerich et al. (Longerich et al., 1996). The concentration of the analyte element in sample (CA,SAM) is given by the count rate for the analyte in the sample (RA,SAM) divided by the normalized sensitivity (S), as follows:
CA;SAM ¼
RA;SAM S
The normalized sensitivity (S) is the sensitivity, determined on a calibration standard (STD), corrected for the volume (mass) of sample ablated. When using naturally occurring internal standards, the sensitivity (cps per unit of concentration) normalized to the mass of the sample (SAM) ablated in the determinations is:
S¼
RA;STD RIS;SAM CIS;STD CA;STD RIS;STD CIS;SAM
where RA,STD is the count rate of the analyte in the standard
200
J. Promchan et al. / Journal of Cereal Science 71 (2016) 198e203
material; CA,STD is the concentration of the analyte in the standard material; RIS,SAM is the count rate of the internal standard in the sample; RIS,STD is the count rate of the internal standard in the standard material; CIS,SAM is the concentration of the internal standard in the sample; and CIS,STD is the concentration of the internal standard in the standard material. 2.3. Elemental imaging The longitudinal rice grain section was rastered by linear laser ablation scans over the entire surface. The operating condition is shown in Table 1 (Method I). The Matrix Laboratory (MATLAB) software version 7.8.0.347 (R2009a) was used to build-up the im^a et al., 2016). ages of rice grains (Pesso 2.4. Multi-element composition for classification Multi-element composition (twenty variables: Al, As, Br, Ca, Cd, Cl, Co, Cu, Fe, Hg, K, Mg, Mn, Mo, Na, P, Pb, Rb, S, and Zn) in core endosperm was analyzed by using single line scan mode of laser ablation. Three rice grains of each sample were ablated (4 lines for one grain). Total 12 data of each rice sample were used for classification. The operating condition is shown in Table 1 (Method II). 2.5. Classification The multivariate statistical evaluation of data was performed using SPSS 16.0 software. The LDA was used for region and color classification. The method was based on maximizing the variance within between groups and minimizing the variance within group by creating new variables (discriminant factors) which are linear combinations of the original variables. Classification success was given by the ratio of the correctly classified objects to their total number as percentage of correct classification.
3. Results and discussion 3.1. Elemental imaging To obtain elemental imaging of the white rice sample from NE, macro elements and micro elements in single longitudinal rice grain section containing endosperm and embryo was analyzed by LA-ICP-MS. The concentration of elements was calculated as described above. The high concentration of macro elements including Ca, K, Mg, and P and micro elements including Cu, Fe, Mn, and Zn was mainly found in embryo as in Fig. 1. It is clearly shown that the concentrations of macro elements were high in only embryo, but they were low in endosperm as in Table 2. Among macro elements in embryo, the highest concentration was found for P at 7.15 mg/g while the concentration of K, Mg and Ca were detected at 5.04, 2.95 and 0.31 mg/g, respectively. The concentration ranking in endosperm was K, P, Ca and Mg with average concentration as 0.95, 0.53, 0.07, and 0.04 mg/g, respectively. For micro elements shown in Fig. 1, the concentration ranking in embryo was Zn, Mn, Fe and Cu with average concentration as 83.00, 40.18, 29.91, and 1.58 mg/g, respectively, while those of concentrations in endosperm were obviously decreased. The concentration ranking in endosperm was Zn, Mn, Fe and Cu with average concentration as 30.18, 9.19, 1.39, and 0.50 mg/g, respectively. The elemental distribution patterns in embryo and endosperm were not uniform. Heterogeneous distribution of Fe across rice grain was observed by using LA-ICP-MS and XRF (Wirth et al., 2009). Elements in embryo were found in high concentrations. The highest concentration of Cu, Fe, Mn and Zn in rice grain was found in embryo (Hansen et al., 2012). Quantified elemental bioimages showed that the highest concentration of Zn presented in embryo while elevated levels of Zn were observed in endosperm (Basnet et al., 2014, 2016). The concentration of several elements such as Ca, Cu, Fe, Mg, Mn, P and Zn was decreased by the increase degree of milling and polishing (Wang et al., 2011). Interestingly, our results also showed that the distribution patterns
Table 1 LA-ICP-MS operating conditions for rice grain analysis. Method Laser ablation Model Laser wavelength (nm) Crater diameter (mm) Repetition rate (Hz) Scanning speed (mm/s) Sampling scheme Carrier gas flow rate (L/ min) Length of ablation (mm) ICP-MS Model Rf power (W) Nebulizer gas flow rate (L/ min) Coolant gas flow rate (L/ min) Auxiliary gas flow rate (L/ min) Autolens Detector mode Scanning mode Dwell time (ms) Settling time (ms) Number of sweeps per reading Number of replicate Isotopes monitored (m/z)
I
II
CETAC LSX-500 266 100 10 50 Line scan analysis 1.0 (He)
GeoLas C 193 95 10 20 Line scan analysis 1.0 (He)
6750
1200
Perkin Elmer ELAN 6000 1400 0.90
Perkin Elmer ELAN 6100 DRCþ 1400 0.95
15.0
17.8
0.90
0.80
On Dual Peak hopping 10 3 1
On Dual Peak hopping 10 3 1
1
1
13 66
C, 25Mg, 31P, 39K, 43Ca, 55Mn, 57Fe, 65Cu, Zn
13
C, 23Na, 25Mg, 27Al, 31P, 32S, 35Cl, 39K, 43Ca, 55Mn, 57Fe, 59Co, 65Cu, 66Zn, 75As, 79Br, 85Rb, 98Mo, 111Cd, Hg, 208Pb
202
J. Promchan et al. / Journal of Cereal Science 71 (2016) 198e203
201
Fig. 1. Elemental imaging of the single longitudinal rice grain section of the white rice sample from NE by LA-ICP-MS for macro elements (Ca, Mg, K and P) and micro elements (Cu, Fe, Mn and Zn).
of trace elements in endosperm were not uniform. The high concentration of Cu and Mn was found at the bran layer and in the middle of endosperm, respectively while that of Zn was distributed from the bran layer through the middle of endosperm. These distribution patterns suggest that essential elements can be lost during polishing process. In this work, 100 mm diameter sized laser ablation crater and 50 mm/s scan speed were used for bio-imaging. With this condition, suitable detection limits (DL) of the interested elements were obtained to monitor elemental distribution in embryo and endosperm such as Fe. However, this crater diameter condition was not the best choice in term of high quality and resolution of bio-images but it was suitable for monitoring elemental distribution in rice grain. 3.2. Elemental composition for classification Sixteen rice grain samples were analyzed and quantified by LAICP-MS. Twenty variables including Al, As, Br, Ca, Cd, Cl, Co, Cu, Fe, Hg, K, Mg, Mn, Mo, Na, P, Pb, Rb, S, and Zn in core endosperm were
Table 2 Average element concentrations in embryo and endosperm of the white rice grain sample from NE according to Fig. 1. Element mg/g Ca K Mg P mg/g Cu Fe Mn Zn
Embryo 0.31 5.04 2.95 7.15
± ± ± ±
0.52 2.05 1.09 3.28
1.58 ± 0.96 29.91 ± 12.91 40.18 ± 17.01 83.00 ± 51.49
Endosperm 0.07 0.95 0.04 0.53
± ± ± ±
0.06 0.79 0.03 0.33
0.50 ± 1.67 1.39 ± 1.72 9.19 ± 4.49 30.18 ± 18.75
investigated. The mean values and standard deviations of element concentrations in rice samples according to their types and origins are summarized in Table 3. It was possible to use the concentration of elements for the discrimination of types and origins of rice
202
J. Promchan et al. / Journal of Cereal Science 71 (2016) 198e203
samples. On the basis of this information, all rice samples have similar element concentration ranges. The concentration of Cl in white rice from NE and S was higher than red and yellow rice. Interestingly, the concentration of K in yellow rice was higher than others. The concentration of As, Br and Fe was lower than the detection limit in most rice samples whereas Cd was lower than the detection limit in all rice samples. 3.3. Origin classification The concentration of 20 elements was analyzed with 16 rice grain samples and used as a chemical descriptor in statistical analysis for the discrimination of rice grains according to their geographic origins. As shown in Table 3, the concentration of Br, Cd and Fe in rice samples was below the detection limit of LA-ICP-MS, the concentration values of these elements were replaced with their detection limits to compensate statistical process. In this study, LDA was used to classify the geographical origins of rice samples. After LDA analysis, discriminant function (F1) correlated with the variables was obtained. The variables affecting the classification were Br and Cl as positive coefficients and Mg as negative coefficient, respectively (>0.700 of standardized canonical discriminant function coefficients, CF). The separation between regional groups including NE and S in the discriminant plot was clearly distinguished as shown in Fig. 2. It is suggested that rice samples from different regions were classified by LDA analysis. To characterize the reliability of the developed classification model, cross-validation method was operated to compute the classification and probability of rice samples. The results enable a satisfied classification with an overall correct classification of 93.8% and a cross-validation of 91.1%.
Fig. 2. Scatter plot of individual rice sample scores on the discriminant function for origin classification. NE¼Northeast; S¼South.
obtained. The three dimensional scatter plot of individual rice samples was resulted in the distribution pattern of 8 rice grain samples according to their types defined by discriminant functions (Fig. 3). The variations between groups were explained by the discriminant function 1 (92.1%), 2 (5.6%) and 3 (2.3%), respectively. The variables affecting the classification were considered by large coefficients of each discriminant function. CF1 was mainly related to Br content as strong positive coefficients; CF2 was related to K, Na, Rb and S content as positive coefficients and Cu, Mg and Mo content as negative coefficients; and CF3 was related to Rb and S as positive coefficients and P concentration as the negative coefficient (>0.700 of standardized canonical discriminant function coefficients, CF). The analysis of rice samples provided a satisfied classification with an overall correct classification of 100% and a cross-validation of 97.9%. However, there were limited numbers of rice samples in this work. This study showed potential for rice type discrimination as shown in Fig. 3. For the rigorous statistical analysis, adequate sample number will be needed for further investigation.
3.4. Type classification The classification of 8 rice samples in different types including white, black, red and yellow rice grains from the South region was analyzed using 20 variables. After applying LDA, discriminant functions (F1, F2 and F3) correlated with the variables were
Table 3 Element concentrations in core endosperm of 16 rice grain samples reported by rice types. Element mg/g Cl K P S mg/g Al As Br Ca Cd Cu Fe Mg Mn Mo Na Rb Zn ng/g Co Hg Pb
Whitea (n ¼ 8) 0.23 0.54 0.47 1.04
± ± ± ±
0.08 0.15 0.16 0.43
Whiteb (n ¼ 5) 0.24 0.71 0.63 1.07
± ± ± ±
0.08 0.29 0.12 0.49
Blackb (n ¼ 1) 0.16 0.49 0.50 1.29
± ± ± ±
0.03 0.09 0.20 0.25
Redb (n ¼ 1) 0.13 0.64 0.63 1.29
± ± ± ±
0.02 0.10 0.21 0.25
Yellowb (n ¼ 1) 0.10 0.92 0.67 1.35
± ± ± ±
0.02 0.13 0.09 0.45
DL 0.03 0.0004 0.001 0.01
1.63 ± 2.22 0.28 ± 0.04
0.65 ± 0.25 0.31 ± 0.14
0.78 ± 0.65
0.75 ± 0.38 0.21 ± 0.05
0.66 ± 0.43 0.16 ± 0.01 21.00 ± 2.53 63.38 ± 25.17
0.16 0.06 0.7 27 0.03 0.03 0.3 0.06 0.03 0.007 0.01 0.003 0.08
36.61 ± 19.41 8.14 ± 2.29 16.77 ± 25.95
11.87 ± 3.33 9.50 ± 4.23 8.22 ± 3.30
59.61 ± 43.92 7.88 ± 1.13 21.12 ± 15.39
18.98 ± 10.51 11.88 ± 2.80 10.19 ± 5.03
27.78 ± 22.07 11.25 ± 2.43 19.23 ± 20.11
5 2 2
DL: Detection limit. a Sample from NE region. b Sample from S region.
J. Promchan et al. / Journal of Cereal Science 71 (2016) 198e203
Fig. 3. 3D scatter plot of individual rice sample scores on the first three discriminant functions for type classification of rice grain samples from the South.
4. Conclusions The elemental imaging of a longitudinal section in single rice grain using LA-ICP-MS was investigated. The essential elements were found in large amounts in embryo and in particular area in endosperm such as Zn and Mn in the bottom and the middle parts of endosperm, respectively. The elemental distributions of rice grain were not uniform. Bio-images of essential elements show that brown rice is an important source of essential elements such as Fe and Zn which can be helpful for Fe and Zn deficiency in humans. These findings are in agreement with the previous work (Basnet et al., 2014, 2016). In addition, the classification of rice grains according to their origins and types was successfully performed by LA-ICP-MS and LDA. Satisfied classifications are obtained. Rice grain samples from NE can be separated from S. Moreover, four rice types from S can be clearly separated. As demonstrated in this study, the LA-ICP-MS analytical approach is an alternative approach to investigate the spatial elemental distribution and elemental composition in rice grains, which could be beneficial to further research on seed development, rice cultivation and rice consumption. Acknowledgements The author gratefully acknowledges for the research grants from the Thailand Research Fund (TRF) through the Royal Golden Jubilee Ph.D. Program (RGJ) and the Center of Excellence for Innovation in Chemistry (PERCH-CIC), Commission on Higher Education, Ministry of Education (Thailand). References Basnet, P., Amarasiriwardena, D., Wu, F., Fu, Z., Zhang, T., 2014. Elemental bioimaging of tissue level trace metal distributions in rice seeds (Oryza sativa L.) from a mining area in China. Environ. Pollut. 195, 148e156. Basnet, P., Amarasiriwardena, D., Wu, F., Fu, Z., Zhang, T., 2016. Investigation of tissue level distribution of functional groups and associated trace metals in rice seeds (Oryza sativa L.) using FTIR and LA-ICP-MS. Microchem. J. 127, 152e159. Becker, J.S., Matusch, A., Wu, B., 2014. Bioimaging mass spectrometry of trace elements e recent advance and applications of LA-ICP-MS: a review. Anal. Chim. Acta 835, 1e18. Cheajesadagul, P., Arnaudguilhem, C., Shiowatana, J., Siripinyanond, A., Szpunar, J., 2013. Discrimination of geographical origin of rice based on multi-element fingerprinting by high resolution inductively coupled plasma mass spectrometry. Food Chem. 141, 3504e3509.
203
Cheajesadagul, P., Wananukul, W., Siripinyanond, A., Shiowatana, J., 2011. Metal doped keratin film standard for LA-ICP-MS determination of lead in hair samples. J. Anal. Atomic Spectrom. 26, 493e498. Chen, H., Siebenmorgen, T.J., Griffin, K., 1998. Quality characteristics of long-grain rice milled in two commercial systems. Cereal Chem. 75, 560e565. Chung, I.-M., Kim, J.-K., Lee, J.-K., Kim, S.-H., 2015. Discrimination of geographical origin of rice (Oryza sativa L.) by multielement analysis using inductively coupled plasma atomic emission spectroscopy and multivariate analysis. J. Cereal Sci. 65, 252e259. Fitzgerald, M.A., McCouch, S.R., Hall, R.D., 2009. Not just a grain of rice: the quest for quality. Trends Plant Sci. 14, 133e139. Gamez, G., Voronov, M., Ray, S.J., Hoffmann, V., Hieftje, G.M., Michler, J., 2012. Surface elemental mapping via glow discharge optical emission spectroscopy. Spectrochim. Acta Part B At. Spectrosc. 70, 1e9. Gholap, D.S., Izmer, A., Samber, B.D., Elteren, J.T.v., Selih, V.S., Evens, R., Schamphelaere, K.D., Janssen, C., Balcaen, L., Lindemann, I., Vincze, L., Vanhaecke, F., 2010. Comparison of laser ablation-inductively coupled plasmamass spectrometry and micro-X-ray fluorescence spectrometry for elemental imaging in Daphnia magna. Anal. Chim. Acta 664, 19e26. Glimn-Lacy, J., Kaufman, P.B., 2006. Seed structure and germination. In: Botany Illustrated, second ed. Springer Science, New York. Gregorio, G.B., Senadhira, D., Htut, T., 1999. Improving iron and zinc value of rice for human nutrients. Agric. Dev. 23, 68e87. Günther, D., Hattendorf, B., 2005. Solid sample analysis using laser ablation inductively coupled plasma mass spectrometry. TrAC Trends Anal. Chem. 24, 255e265. Halicz, L., Günther, D., 2004. Quantitative analysis of silicates using LA-ICP-MS with liquid calibration. J. Anal. Atomic Spectrom. 19, 1539e1545. Hansen, T.H., Lombi, E., Fitzgerald, M., Laursen, K.H., Frydenvang, J., Husted, S., Boualaphanh, C., Resurreccion, A., Howard, D.L., de Jonge, M.D., Paterson, D., Schjoerring, J.K., 2012. Losses of essential mineral nutrients by polishing of rice differ among genotypes due to contrasting grain hardness and mineral distribution. J. Cereal Sci. 56, 307e315. Hare, D., Austin, C., Doble, P., 2012. Quantification strategies for elemental imaging of biological samples using laser ablation-inductively coupled plasma-mass spectrometry. Analyst 137, 1527e1537. Kokot, S., Dong Phuong, T., 1999. Elemental content of Vietnamese ricePart 2. Multivariate data analysis. Analyst 124, 561e569. Longerich, H.P., Jackson, S.E., Günther, D., 1996. Inter-laboratory note. Laser ablation inductively coupled plasma mass spectrometric transient signal data acquisition and analyte concentration calculation. J. Anal. Atomic Spectrom. 11, 899e904. M-M, P., Hansawek, R., Shiowatana, J., Siripinyanond, A., 2011. Geographical origin classification of gem corundum using elemental fingerprint analysis by laser ablation inductively coupled plasma mass spectrometry. Int. J. Mass Spectrom. 306, 57e62. Meharg, A.A., Lombi, E., Williams, P.N., Scheckel, K.G., Feldmann, J., Raab, A., Zhu, Y., Islam, R., 2008. Speciation and localization of arsenic in white and brown rice grains. Environ. Sci. Technol. 42, 1051e1057. ^ a, G.d.S., Capelo-Martínez, J.L., Fdez-Riverola, F., Lo pez-Ferna ndez, H., GlezPesso ~ a, D., Reboiro-Jatoef, M., Arruda, M.A.Z., 2016. Laser ablation and inductively Pen coupled plasma mass spectrometry focusing on bioimaging from elemental distribution using MatLab software: a practical guide. J. Anal. Atomic Spectrom. 31, 832e840. S anchez-Amaya, J.M., Blanco, G., Garcia-Garcia, F.J., Bethencourt, M., Botana, F.J., 2012. XPS and AES analyses of cerium conversion coatings generated on AA5083 by thermal activation. Surf. Coatings Technol. 213, 105e116. Shin, T.-S., Godber, J.S., Martin, D.E., Wells, J.H., 1997. Hydrolytic stability and changes in E vitamers and oryzanol of extruded rice bran during storage. J. Food Sci. 62, 704e728. Sui, T., Song, B., Dluhos, J., Lu, L., Korsunsky, A.M., 2015. Nanoscale chemical mapping of Li-ion battery cathode material by FIB-SEM and TOF-SIMS multi-modal microscopy. Nano Energy 17, 254e260. Suzuki, Y., Chikaraishi, Y., Ogawa, N.O., Ohkouchi, N., Korenaga, T., 2008. Geographical origin of polished rice based on multiple element and stable isotope analyses. Food Chem. 109, 470e475. zquez, M.A., Gala n, E., Ortiz, P., Ortiz, R., 2013. Digital image analysis and EDX SEM Va as combined techniques to evaluate salt damp on walls. Constr. Build. Mater. 45, 95e105. Wang, H.-J., Wang, M., Wang, B., Meng, X.-Y., Wang, Y., Li, M., Feng, W.-Y., Zhao, Y.-L., Chai, Z.-F., 2010. Quantitative imaging of element spatial distribution in the brain section of a mouse model of Alzheimer's disease using synchrotron radiation X-ray fluorescence analysis. J. Anal. Atomic Spectrom. 25, 328e333. Wang, K.M., Wu, J.G., Li, G., Zhang, D.P., Yang, Z.W., Shi, C.H., 2011. Distribution of phytic acid and mineral elements in three indica rice (Oryza sativa L.) cultivars. J. Cereal Sci. 54, 116e121. Wirth, J., Poletti, S., Aeschlimann, B., Yakandawala, N., Drosse, B., Osorio, S., Tohge, T., Fernie, A.R., Günther, D., Gruissem, W., Sautter, C., 2009. Rice endosperm iron biofortification by targeted and synergistic action of nicotianamine synthase and ferritin. Plant Biotechnol. J. 7, 631e644. Wu, B., Becker, J.S., 2012. Imaging techniques for elements and element species in plant science. Metallomics 4, 403e416.