Optimisation of lactic acid fermentation for improved vinegar flavour during rosy vinegar brewing

Editorial Published online in Wiley Interscience: (www.interscience.wiley.com) DOI 10.1002/jsfa.3996 Food for health and wellbeing: 14th World Congress of Food Science and Technology J Sci Food Agric 2010; 90: 1283–1284 fatty acid profiles of pathogenic bacteria to sensitively detect and identify them in foodstuffs.15 We wish to extend our sincerest thanks to all the authors and referees of these articles and hope that the JSFA readership finds them useful We are very much looking forward to the 15th World Congress in Cape Town later in 2010, and hope to see you there Guest Editors Professor Bo Jiang Jiangnan University, China Dr Li Ni Fuzhou University, China Dr Ken Buckle The University of New South Wales, Australia REFERENCES Kopjar M, Andriot I, Saint-Eve A, Souchon I and Guichard E, Retention of aroma compounds: an interlaboratory study on the effect of the composition of food matrices on thermodynamic parameters in comparison with water J Sci Food Agric 90:1285–1292 (2010) Wongwiwat P, Wattanachant S and Siripongvutikorn S, Effect of phosphate treatments on microbiological and physicochemical changes of spent hen muscle marinated with Tom-Yum paste during chilling storage J Sci Food Agric 90:1293–1299 (2010) Yan W-Q, Zhang M, Huang L-L, Tang J, Mujumdar AS and Sun J-C, Study on optimization of puffing characteristics of potato cubes by spouted bed drying enhanced with microwave J Sci Food Agric 90:1300–1307 (2010) Mirhosseini H and Tan CP, Discrimination of orange beverage emulsions with different formulations using multivariate analysis J Sci Food Agric 90:1308–1316 (2010) Chen S, Zhang M and Wang S, Physiological and quality responses of Chinese ‘Suli’ pear fruit to 1-MCP vacuum infiltration treatment J Sci Food Agric 90:1317–1322 (2010) Hu W, Jiang A, Tian M, Liu C and Wang Y, Effect of ethanol treatment on physiological and quality attributes of fresh-cut eggplant J Sci Food Agric 90:1323–1326 (2010) Cheng L, Mu W and Jiang B, Cloning, overexpression and characterization of a thermostable L-arabinose isomerase from Bacillus stearothermophilus IAM 11 001 for D-tagatose production J Sci Food Agric 90:1327–1333 (2010) Jiang Y, Guo J, Li Y, Lin S, Wang L and Li J, Improvement of lactic acid production during rosy vinegar fermentation J Sci Food Agric 90:1334–1339 (2010) Li Q, Chen Q, Ruan H, Zhu D and He G, Isolation and characterization of a novel oxygen, acid and bile resistant Bifidobacterium animalis subsp lactis Qq08 J Sci Food Agric 90:1340–1346 (2010) www.soci.org 1283 Welcome to this special edition of the Journal of the Science of Food and Agriculture, featuring research papers that were presented at the 14th World Congress of Food Science and Technology held in Shanghai, China An important aim of the conference was to encourage dialogue, knowledge exchange, and collaboration between Eastern and Western institutions and industries The theme enabling this was ‘‘Food for health and wellbeing – where tradition meets the future’’, giving expert food scientists and industry professionals the chance to compare the traditional Chinese interest of health and food against some additional modern concepts and technologies The fascinating and cosmopolitan city of Shanghai was the perfect backdrop to this meeting, drawing in delegates from across the globe and allowing easier access for China, which is home to the largest number of food professionals in the world Opening our special issue we have a group of papers that focus on food technology and physico-chemical properties of varied products Kopjar et al begin by studying the ability of differing food matrices to retain aroma compounds, which closely affects food quality and enjoyment.1 Following this Wongwiwat et al investigate whether applying phosphate pre-treatments to marinated chicken meat may improve important qualities such as marinade absorption, tenderness and shelf-life.2 Next, Yan et al use a specialized method of microwave drying to look at its textural effects on a puffed potato snack as a healthier alternative to potato chips fried in oil.3 Then Dr Mirhosseini and team look at how the ingredients of orange beverage emulsions affect the rheological emulsion properties of such a product.4 Representing the postharvest industry we have Chen et al., who designed a form of vacuum infiltration allowing pears to absorb 1MCP and report here the resulting physiological quality changes.5 Following this Dr Hu et al look for innovative ways to preserve fresh-cut egg plant using ethanol vapour treatments.6 Leading our biotechnology section, Cheng et al investigate a method of producing L-arabinose isomerase in order to make a new kind of low-calorie sweetener, D-Tagatose,7 and Jiang and team optimize fermentation parameters to improve the flavor qualities of vinegar.8 Next, a paper by Dr Li Qingqing et al considers how to screen for a more favorable, aerotolerant strain of Bifidobacterium for use in probiotic foods.9 The subsequent papers, by Dr Li Jing-En et al and Dr Xie et al., respectively, investigate innovative uses and activities of plant products that are well known in China for their potential uses in human health and the food industry.10,11 Then papers by Dr Yin et al and Dr Xua et al give us valuable insights into Chinese consumer behavior towards purchasing organic produce and food safety awareness.12,13 Bringing this special edition to a close, Kuang and team look into the two most highly used pesticides on tea plantations and their residues in Chinese tea products.14 Finally, Dr Li Yanyan et al report their effective method of utilizing the c 2010 Society of Chemical Industry www.soci.org 10 Li J-E, Nie S-P, Qiu Z-H, Che M-J, Li C and Xie M-Y, Antimicrobial and antioxidant activities of essential oil from Herba Moslae J Sci Food Agric 90:1347–1352 (2010) 11 Xie J-H, Xie M-Y, Shen M-Y, Nie S-P, Li C and Wang Y-X, Optimization of microwave-assisted extraction of polysaccharides from Cyclocarya paliurus (Batal.) Iljinskaja using response surface methodology J Sci Food Agric 90:1353–1360 (2010) 12 Yin S, Wu L, Du L and Chen M, Consumers’ purchase intention of organic food in China J Sci Food Agric 90:1361–1367 (2010) Editorial 13 Xu L and Wu L, Food safety and consumer willingness to pay for certified traceable food in China J Sci Food Agric 90:1368–1373 (2010) 14 Kuang H, Miao H, Hou X, Zhao Y, Shen J and Wu Y, Determination of enantiomeric fractions of cypermethrin and cis-bifenthrin in Chinese teas by GC-ECD J Sci Food Agric 90:1374–1379 (2010) 15 Li Y, Wu S, Wang L, Li Y, Shi F and Wang X, Differentiation of bacteria using fatty acid profiles from gas chromatography/tandem mass spectrometry J Sci Food Agric 90:1380–1383 (2010) 1284 www.interscience.wiley.com/jsfa c 2010 Society of Chemical Industry J Sci Food Agric 2010; 90: 1283–1284 Research Article Received: 23 December 2008 Revised: 23 June 2009 Accepted: 25 June 2009 Published online in Wiley Interscience: 18 March 2010 (www.interscience.wiley.com) DOI 10.1002/jsfa.3929 Retention of aroma compounds: an interlaboratory study on the effect of the composition of food matrices on thermodynamic parameters in comparison with water Mirela Kopjar,a – d Isabelle Andriot,a – c Anne Saint-Eve,e Isabelle Souchone and Elisabeth Guicharda – c∗ Abstract BACKGROUND: Partition coefficients give an indication of the retention of aroma compounds by the food matrix Data in the literature are obtained by various methods, under various conditions and expressed in various units, and it is thus difficult to compare the results The aim of the present study was first to obtain gas/water and gas/matrix partition coefficients of selected aroma compounds, at different temperatures, in order to calculate thermodynamic parameters and second to compare the retention of these aroma compounds in different food matrices RESULTS: Yogurts containing lipids and proteins induced a higher retention of aroma compounds than model gel matrices The observed effects strongly depend on hydrophobicity of aroma compounds showing a retention for ethyl hexanoate and a salting out effect for ethyl acetate A small but noticeable decrease in enthalpy of affinity is observed for ethyl butyrate and ethyl hexanoate between water and food matrices, suggesting that the energy needed for the volatilization is lower in matrices than in water CONCLUSION: The composition and complexity of a food matrix influence gas/matrix partition coefficients or aroma compounds in function of their hydrophobicity and to a lower extent enthalpy of vaporization c 2010 Society of Chemical Industry Keywords: aroma compounds; partition coefficient; retention; food matrix INTRODUCTION J Sci Food Agric 2010; 90: 1285–1292 ∗ Correspondence to: Elisabeth Guichard, INRA, UMR 1129 FLAVIC, 17 Rue Sully, F-21000 Dijon, France E-mail: guichard@dijon.inra.fr a INRA, UMR 1129 FLAVIC, F-21000 Dijon, France b ENESAD, UMR 1129 FLAVIC, F-21000 Dijon, France c Universit´e de Bourgogne, UMR 1129 FLAVIC, F-21000 Dijon, France d Faculty of Food Technology, F Kuhaca 18, HR-31000 Osijek, Croatia e UMR 782 GMPA, INRA-AgroParisTech BP 1, F-78850 Thiverval-Grignon, France www.soci.org c 2010 Society of Chemical Industry 1285 Acceptability of foods by consumers is highly governed by their organoleptic properties Aroma perception is mainly related to the release in the gas phase of aroma compounds present in the food However, food products are very complex matrices because of their chemical composition and structure The nature and amount of different components such as proteins, lipids and carbohydrates greatly influence aroma release and perception.1,2 Understanding the behavior of aroma compounds, the strength and the nature of interactions between aroma compounds and ingredients of food matrix is thus of major significance for improving the overall aroma and thus the quality of food products The influence of different types of proteins on aroma compounds release has been widely studied, pointing out the existence of molecular interactions, ionic bonding, hydrogen bonding, and hydrophobic bonding.3 – As most of the aroma compounds are hydrophobic, the presence of lipids influences partitioning of aroma compounds between oil and aqueous phase and consequently between matrix and gas phase.6 The effect of fat on the retention of hydrophobic aroma compounds is higher than that of proteins.7 Moreover, the retention of aroma compounds in emulsions depends not only on the amount of fat but also on the composition of the fat.8 Carbohydrates can also lead to the retention of aroma compounds in the matrix For example, starch and more precisely amylose can form complexes with specific aroma compounds.9 – 11 Other types of carbohydrates such as pectin can play an important role in aroma compound retention due to their thickening effect, which reduces aroma diffusion.12 – 14 www.soci.org Table Main physicochemical characteristics of the studied aroma compounds Aroma compound Ethyl acetate Ethyl butyrate Ethyl hexanoate Molecular formula Molecular mass (g mol−1 ) Hydrophobicity log Pa Saturated vapor pressure psi 25 ◦ C (mmHg) Solubilityb in water at 25 ◦ C (g L−1 ) C4 H8 O2 C6 H12 O2 C8 H16 O2 88 116 144 0.73 1.85 2.83 98.3b 101c 99.6∗ 29.93 14.6b 15.6c 15.1∗ 2.70 1.80b 1.68d 1.74∗ 0.31 ∗ Mean value Hydrophobicity (log P) calculated by the method of Rekker.16 b Estimation program EPI Suite c Calculated with Antoine’s equation parameters d Le Thanh et al.17 M Kopjar et al fat (solid fat level at 10 ◦ C) was modified between yogurts enriched with a standard ratio of whey protein to total protein: yogurt with standard anhydrous milk fat (AMF yogurt: solid fat level of 70%) and yogurt with a high solid fat level (HMF yogurt: solid fat level of 82%) Model gel matrices21 with sucrose (350 g kg−1 ), KCl (1.6 g kg−1 ) and CaSO4 (0.3 g kg−1 ) differed in their composition in texturing agents: without texturing agent (S), with starch (14 g kg−1 ) and pectin (4 g kg−1 ) (P), with starch (14 g kg−1 ) and carrageenan (C) C1, C2 and C3 matrices differed in carrageenan content (0.5, 3.3 and g kg−1 , respectively) Model cheeses22 contained micellar casein (75 g kg−1 ), whey protein (2.5 g kg−1 ), lactose (40 g kg−1 ), sodium chloride (10 g kg−1 ), yeast extract (0.2%), mineral solution (1 mL kg−1 ), starter culture (0.2 mL kg−1 ) and water They only differed in chymosin amount (0, and 10 µg kg−1 ), in order to obtain different gel hardness (CM0, CM3 and CM10) a In order to be perceived during the consumption of food, aroma compounds have first to be transferred from food to the gas phase via the interface and second to reach the olfactory epithelium At thermodynamic equilibrium, aroma compounds are found in the gas phase in a concentration that depends on their gas/matrix partition coefficient Partition coefficients measured in different matrices for the same aroma compound give an indication of the retention of aroma compounds by the food matrix.15 Data in the literature are obtained by various methods, under various conditions and expressed in various units and it is thus difficult to compare the results – even those obtained in water Within a French collaborative project (CANAL-ARLE), gas/water and gas/matrix partition coefficients were calculated for selected aroma compounds by using the same method but at different temperatures The aim of the present study was first to make a comparison of data obtained in the literature and within the CANAL-ARLE project for gas/water partition coefficients of selected aroma compounds, at given temperatures, in order to calculate thermodynamic parameters In a second step, the retention of these aroma compounds in different food matrices will be calculated and discussed Determination of gas/matrix partition coefficient The gas/matrix partition coefficient K describes the distribution of volatile compounds between the matrix and the gas phase: K= CG CM (1) with CG the concentration of the volatile compound in the headspace and CM the concentration of the volatile compound in the matrix Determination of gas/matrix partition coefficients of different aroma compounds in model matrices, which varied in composition and structure, was conducted in various laboratories which were all partners in the framework of CANAL-ARLE project, by the phase ratio variation (PRV) method.23 This method is based on the influence of the volume of the sample on the concentration of volatile in the headspace The following equation was established: 1 × (K + β) = A fi × CM (2) where A is the chromatographic peak area at equilibrium, fi is the proportional factor, CM is the initial sample concentration in the matrix and β is the ratio VG /VM Variables in this equation are the phase ratio (β) and the peak area (A) Thus Eqn (2) corresponds to a linear equation of the following type: EXPERIMENTAL 1/A = a + bβ (3) 1286 Material Physicochemical properties of the selected aroma compounds (ethyl acetate, ethyl butyrate and ethyl hexanoate) are given in Table K (= a/b) is calculated from the values of a and b obtained by plotting 1/A against β in the linear zone Composition of model matrices Three types of flavored model food matrices were studied: yogurts, gel matrices and cheeses Flavored stirred yogurts (dry matter (22.5%), fat (4 g kg−1 ) and total proteins (5.4 g kg−1 )) differed either in the nature of added proteins18,19 or in the nature of fat.20 First, three ratios of whey protein to total protein varied between yogurts: yogurt enriched with a low ratio of whey protein to total protein (CAS yogurt) and yogurt enriched with a high ratio of whey protein to total protein (WP yogurt) Secondly, the fraction of fatty acids of anhydrous milk Standardization of partition coefficients from the literature Literature data were obtained with various methods, under various conditions, and gas/water partition coefficients were expressed in various units Actually, partition coefficient can be expressed using the activity coefficient (γι ), mass fraction (km ), molar fraction (Ki ) or molar concentration (ki ) Therefore, to compare the gas/water partition coefficient from the literature with that of the CANAL-ARLE project, recalculation of literature data was necessary The parameters used for conversion are either standard constants (R, T, P, Ma , Mliq and dliq ) or the www.interscience.wiley.com/jsfa c 2010 Society of Chemical Industry J Sci Food Agric 2010; 90: 1285–1292 Retention of aroma compounds in different food matrices www.soci.org saturation pressure of the pure compound at the corresponding temperature (psi ) Conversion of partition coefficient ki expressed in molar concentration, or in mass concentration to Ki expressed as a molar fraction: RTdliq × 106 (4) Ki = ki × PMliq compound at equilibrium between a liquid phase and a gas phase is the ratio of the activities of the compound in each phase: Conversion of partition coefficient km expressed as a mass fraction to Ki expressed as a molar fraction: Assuming that the liquid phase should be considered as a diluted solution, and the vapor phase obeys the Boyle–Mariotte law, [A]liq p and according to pA V = nA RT a(A)vap = A0 and a(A)liq = C0 P (true if vapor phase volume is constant): Ki = km × Ma Mliq (5) Aliq Keq Keq = Conversion of partition coefficient Ki expressed as a molar fraction to γι activity coefficient: P γι = Ki × s pi (6) Conversion of partition coefficient ki expressed as a molar concentration to γι activity coefficient: γι = ki × RTdliq × 106 psi Mliq (7) Avap a(A)vap = a(A)liq [A]vap · RT = K × RT [A]liq (10) (11) where [A]liq is the concentration of compound (A) in the liquid phase (mol L−1 ); C0 is the reference concentration (C0 = mol L−1 ); pA is the partial pressure of compound (A) in the vapor phase (atm); Po is the reference pressure (1 atm); with R = 0.08211 atm K−1 mol−1 , T in K, pA in atm Knowing the linear relationship regression parameters (slope and intercept) of ln Keq versus 1/T, it is possible to calculate: • Keq values at any temperature (in a range smaller than 100 ◦ C); • H0 value ( H0 = slope × R) where P = pressure (Pa) T = temperature (K) R = gas constant (8.314 J mol−1 K−1 ) dliq = liquid density (kg L−1 ) Mliq = molar mass of liquid (g mol−1 ) Ma = molar mass of air (28.8 g mol−1 ) psi = saturated vapor pressure of compound i (Pa) All the data from the literature were thus converted into molar concentration for comparison with data from the CANAL-ARLE project which were expressed in that unit Retention of aroma compounds in different matrices The gas/water (Kwater ) and gas/matrix (Kmatrix ) partition coefficients of aroma compounds were used for the determination of percentage of retention of the aroma compounds (RAC) by different matrices, at a given temperature: Calculation of equilibrium thermodynamic values For equilibrium, the Vant’Hoff law links the equilibrium constant to temperature according the following equation: For yogurts, partition coefficients were determined at several temperatures, and thus the linearity between ln Keq and 1/T was checked and the regression equation was used for prediction of partition coefficients at temperatures other than those studied d ln(Keq ) H0 = dT RT (8) H0 is the enthalpy of the chemical or physicochemical where reaction (expressed in J mol−1 ), R is the perfect gas constant (R = 8.314 J K−1 mol−1 ) and T is temperature (expressed in kelvin) Assuming that H0 remains constant for the considered temperature range (Ellingham approximation), the previous equation should be integrated as follows: ln(Keq ) = − H0 +C RT (9) J Sci Food Agric 2010; 90: 1285–1292 Kmatrix × 100 Kwater (12) RESULTS AND DISCUSSION The main physicochemical characteristics of the studied aroma compounds are presented in Table Gas/water partition coefficient Data were collected from the literature and from the CANALARLE project for the three chosen aroma compounds – ethyl acetate, ethyl butyrate and ethyl hexanoate – in a wide range of temperatures (0–98 ◦ C, 4–80 ◦ C and 0–50 ◦ C, respectively) Partition coefficients collected from the literature were obtained by various methods, under various conditions, and expressed in various units However, data obtained from the CANAL-ARLE project were obtained by the same method and expressed in the same unit In order to make a comparison of all the data, literature data were converted into ratio of molar concentration using the conversion equations (4–7) As an example, the results obtained for ethyl butyrate are presented in Table The differences noted c 2010 Society of Chemical Industry www.interscience.wiley.com/jsfa 1287 If the equilibrium constant values are known at different temperatures, it is possible to perform a linear regression calculation of ln(K) versus 1/T Use of the thermodynamic equilibrium constant is required for the regression calculation The equilibrium constant of a RAC(%) = − www.soci.org M Kopjar et al Table Conversion of gas/water partition coefficient for ethyl butyrate French collaborative project (measured data) Literature (bold: literature data; italic: conversion data) T (◦ C) Ratio of activity coefficient Ratio of mass fraction Ratio of molar fraction Ratio of molar concentration Ratio of molar concentration 10 12 20 1.41E+02 3.35E+02 1.13E+02 1.15E+03 4.30E+02 6.79E+02 7.53E+02±3.66E+02∗ 1.76E+00 4.16E+00 2.25E+00b 1.43E+01 5.34E+00b 8.44E+00 9.36E+00±4.54E+00∗ 2.81E+00a 6.66E+00a 3.60E+00 2.29E+01 8.56E+00 1.35E+01a 1.50E+01±7.28E+00∗ 2.22E−03 5.15E−03 2.77E−03 1.71E−02c 6.39E−03 1.01E−02 1.12E−02±5.44E−03∗ 25 1.07E+03±1.39E+02d 9.21E+02 1.08E+03±1.87E+02 1.06E+03±1.61E+02∗ 1.33E+01±1.71E+00 1.14E+01 1.34E+01±2.35E+00 1.32E+01±2.00E+00∗ 2.12E+01±2.74E+00 1.83E+01 2.15E+01±3.77E+00f 2.11E+01±3.21E+00∗ 1.56E−02±2.01E−03 1.35E−02e 1.58E−02±2.81E−03 1.55E−02±2.37E−03∗ 1.79E−02k 30 1.49E+01 1.53E+01 1.51E+01±2.92E−01∗ 1.17E+01b 2.43E+01 2.21E+01±4.32E+00 2.05E+01±5.88E+00∗ 2.38E+01 2.45E+01h 2.41E+01±4.67E−01∗ 1.87E+01 3.89E+01h 3.54E+01±6.91E+00 3.28E+01±9.40E+00∗ 1.72E−02 1.77E−02 1.75E−02±3.54E−04∗ 1.32E−02 2.75E−02 2.50E−02±4.87E−03i 2.32E−02±6.63E−03∗ 1.83E−02l 2.40E−02m 2.12E−02±4.03E−03∗ 37 1.20E+03g 1.24E+03 1.22E+03±2.47E+01∗ 9.39E+02 1.96E+03 1.79E+03±3 48E+02 1.65E+03±4.74E+02∗ 60 70 80 6.13E+03 9.78E+03 1.36E+04 7.61E+01 1.21E+02 1.68E+02 1.22E+02 1.94E+02 2.69E+02 8.00E−02j 1.24E−01j 1.67E−01j ∗ Mean value of all found literature data Nongonierma et al.8 Seuvre et al.24 c Stanford and McGorrin25 d Mean value from Carelli et al.,26 Le Thanh et al.,17 Harvey et al.,27 Sancho et al.28 e Landy et al.29 f Mean value from Le Thanh et al.,17 Harvey et al.,27 Landy et al., Philippe et al.30 g Carelli et al.26 h Philippe et al.30 i Mean value from van Ruth,31 Bylaite et al.32 j Jouquand et al.33 k Atlan et al.34 l Gierczynski et al.22 m Savary et al.21 a b 1288 www.interscience.wiley.com/jsfa ethyl acetate y = -4841.5x + 14.365 R2 = 0.9852 ethyl butyrate ethyl hexanoate y = -5912.6x + 18.561 R2 = 0.9692 y = -6631.2x + 21.688 R2 = 0.9501 0.0031 0.0035 2.00 1.00 lnKeq between the literature data can be explained by the differences in the methodologies and the differences in the values of saturated pressure used for the conversion of data into the same unit (Table 1) For the investigated aroma compounds, a linear relation between ln Keq and 1/T was obtained (Fig 1) with high correlations (0.9852, 0.9692 and 0.9453 for ethyl acetate, ethyl butyrate and ethyl hexanoate, respectively; Table 3) The corresponding equations were then used to predict values of gas/water partition coefficient at temperatures at which this parameter was not experimentally determined These equations were then used to calculate ln Keq at 30 ◦ C The studied compounds – ethyl acetate, ethyl butyrate and ethyl hexanoate – differ in their hydrophobic chain length and thus a logical linear regression was found between ln Keq at 30 ◦ C and log P value (R2 = 0.9942) and between ln Keq at 30 ◦ C and MW (R2 = 0.9986), both with a positive slope (respectively 0.67 and 0.025) These results are in agreement with results of other 0.00 0.0027 -1.00 0.0029 0.0033 0.0037 -2.00 -3.00 -4.00 1/T Figure Relation between ln Keq and 1/T for ethyl acetate, ethyl butyrate and ethyl hexanoate, in water authors.4,57,58 As partition coefficients not depend only on hydrophobicity but also on solubility in water and saturated c 2010 Society of Chemical Industry J Sci Food Agric 2010; 90: 1285–1292 Retention of aroma compounds in different food matrices www.soci.org Table Equilibrium standard thermodynamical values of three aroma compounds in water and different food matrices Matrix T (◦ C) R2 F/Fcrit Keq (30 ◦ C) H0 eq (J mol−1 ) Ethyl acetate Watera CAS WP C2d 0, 4, 10, 20, 25, 28, 30, 37, 40, 50, 60, 70, 80 4, 7, 10, 20, 25 4, 7, 10, 20, 25 10, 20, 30 0.9852 0.9956 0.9770 0.9369 1195.98/4.413 684.97/10.13 127.64/10.13 14.84/161.45 1.99E−01 3.03E−01 3.16E−01 4.38E−01 40.3E+03 39.2E+03 39.0E+03 38.6E+03 Ethyl butyrate Waterb CAS WP C2d 4, 10, 12, 20, 25, 30, 37, 60, 70, 80 4, 10, 20, 25 4, 10, 20, 25 10, 20, 30 0.9692 0.9909 0.9510 0.9758 345.58/4.844 217.65/18.513 38.81/18.513 40.28/161.45 3.86E−01 2.24E−01 2.16E−01 4.68E−01 49.2E+03 37.3E+03 34.4E+03 21.3E+03 Ethyl hexanoate Waterc CAS WP C2d 0, 4, 10, 14, 20, 25, 30, 37 4, 10, 20, 25 4, 10, 20, 25 10, 20, 30 0.9453 0.9567 0.9479 0.9411 129.06/5.59 44.19/18.513 36.36/18.513 15.97/161.45 8.21E−01 3.92E−02 3.91E−02 7.94E−01 55.1E+03 30.4E+03 31.4E+03 26.3E+03 Aroma compounds a Covarrubias-Cervantes et al.,35 Nongonierma et al.,8 Pividal et al.,36 Le Thanh et al.,17,37,38,44,50 Carelli et al.,26 Harvey et al.,27 Voilley and Bosset,39 Voilley and Sauvageot,40 Druaux et al.,41 Landy et al.,42 Fares et al.,43 Sancho et al.,28 Marinos and Saravacos,45 Philippe et al.,30 Nelson and Hoff,46 Sorrentino et al.,47 Baudot et al.,48 Kolb et al.49 b Values from Table c Covarrubias-Cervantes et al.,35 Nongonierma et al.,8 Stanford and McGorrin,25 Langourieux and Crouzet,51 Lubbers,52 Voilley and Lubbers,53 Jouenne,54 Le Thanh et al.,17 Landy et al.,4,29,55 Philippe et al.,30 Savary et al.,21 Hirata et al.,56 Seuvre et al.,24 Gierczynski et al.22 d Savary et al.21 J Sci Food Agric 2010; 90: 1285–1292 ethyl acetate ethyl butyrate ethyl hexanoate 100 80 60 40 20 -20 -40 -60 AMF HMF CAS WP C2 Figure Percentage of retention of aroma compounds in different matrices at 10 ◦ C behavior to two other compounds In yogurts and model cheeses these results can be explained by hydrophobic interactions between aroma compounds and proteins Ethyl hexanoate, the aroma compound with the highest hydrophobicity value among investigated aroma compounds, was mostly retained in yogurt and model gel matrices, while for ethyl acetate, which has the lowest hydrophobic character, a salting-out effect was observed Tan and Siebert5 investigated binding of aroma compounds to bovine serum albumin and found that the interactions between aroma compounds and protein became stronger with an increase in the hydrophobic chain length of esters Sostmann et al.59 also found a good linear correlation between the logarithm of the binding constant of aroma compounds to β-lactoglobulin measured by affinity chromatography and hydrophobicity of the molecules, for series of ketones, aldehydes, alcohols, esters and lactones In model cheeses different effects were observed for ethyl butyrate and ethyl hexanoate: a salting-out effect for ethyl butyrate (−44 to −28%) and retention for ethyl hexanoate (48–55%) The hydrophobic nature of the molecule explains the differences in affinity for the proteins present in the cheese and, for example, for βlactoglobulin, which presents a higher binding constant with ethyl c 2010 Society of Chemical Industry www.interscience.wiley.com/jsfa 1289 Retention of aroma compounds in different matrices Retention of aroma compounds was calculated (Eqn (12)) for ethyl acetate, ethyl butyrate and ethyl hexanoate in yogurts and model gel matrices, and for ethyl butyrate and ethyl hexanoate in model cheese matrices Gas/matrix partition coefficients were determined at different temperatures In yogurts with different fractions of fatty acids of anhydrous milk fat (AMF and HMF), partition coefficients were only determined at 10 ◦ C, and in model gel matrices only at 30 ◦ C In yogurts containing different proteins (CAS and WP), determination of partition coefficient was done over a wider range of temperatures (4, 7, 10, 20 and 25 ◦ C) In one model gel matrix (C2) determination was done at 10, 20 and 30 ◦ C When possible the equation between ln Keq versus 1/T was established with high correlation coefficients (Table 3), and used to calculate the partition coefficient at another temperature in order to compare aroma compound retention between the different matrices at the same temperature The retention of aroma compounds at 10 ◦ C (Fig 2) was calculated for yogurts and one model gel matrix (C2) The retention of aroma compounds at 30 ◦ C (Fig 3) was calculated for yogurts with different proteins (CAS and WP) and all the model gel matrices and model cheeses A positive value indicates retention of aroma compound in the matrix, while a negative value indicates a salting out effect Whatever the temperature, for the three aroma compounds a higher retention is observed in yogurts than in model gel and model cheeses Ethyl hexanoate, the more hydrophobic compound, is more retained in all matrices than the two other aroma compounds, except at 10 ◦ C in C2 matrix, in which all aroma compounds are salted out, and in HMF yogurt, in which only a small effect is observed Ethyl butyrate presents an intermediate 10 °C retention (%) vapour pressure,15 the relation between ln Keq at 30 ◦ C and these two physicochemical characteristics was also tested An exponential decrease in ln Keq at 30 ◦ C was found by increasing these two parameters www.soci.org ethyl acetate 30 °C M Kopjar et al ethyl butyrate ethyl hexanoate 100 80 retention (%) 60 40 20 -20 -40 -60 CAS WP S P C1 C2 C3 M0 M3 M10 Figure Percentage of retention of aroma compounds in different matrices at 30 ◦ C hexanoate (543 L mol−1 ) than with ethyl butyrate (55 L mol−1 ).60 The salting-out effect observed at both temperatures for ethyl acetate, except for AMF, was also observed by Nongonierna et al.8 in yogurt without fat Moreover Landy et al.4 found that ethyl acetate was the only aroma compound which was not bound to sodium caseinate In yogurts containing different types of fat, the retention of the three esters is higher in AMF yogurts than in HMF yogurts, which can be explained by the higher content of liquid lipids in AMF than in HMF Solid liquids sorb less than liquid ones61 and thus an increase in solid fat level induces an increase in aroma release, as observed previously in milk-based emulsions62 and in model oil-in-water emulsions.63 In the case of model gel matrices, the presence of polysaccharides induces fewer interactions with aroma compounds, as previously observed by Bylaite et al.32 with λ-carrageenan and by Juteau et al.64 for ι-carrageenan A salting-out effect is observed for the three aroma compounds, whatever their hydrophobicity, in C2 model gel matrix at 10 ◦ C Philippe et al.15 also observed a slight salting-out effect of carbohydrates on 2-pentanone which they explained by a decrease of water availability for the solubilization of 2-pentanone and hence an increase in its volatility 1290 Standard equilibrium thermodynamic values For all matrices in which partition coefficients were calculated at different temperatures (water, two yogurts and one model gel matrix), thermodynamic equilibrium values were estimated (Table 3) In water, an increase of hydrophobicity of the aroma compounds induces an increase of Keq and of enthalpy of vaporization ( H0 eq ) The partition coefficient takes into account the vapor pressure of the compound as well as the environment and the interactions between macroconstituents and aroma compounds In contrast, enthalpy of vaporization is directly linked to the attraction or repulsion forces that retain or release the aroma compounds in their environment.65 As was pointed out by Philippe et al.,30 a linear relationship between ln K and 1/T indicates that in the temperature range studied the enthalpies of vaporization can be considered as constant The values of Keq used in the present study are obtained from a wider range of temperatures (Table 3) than in other studies The values of enthalpy of vaporization obtained in the present study included a large number of data obtained over a wide range of temperature These are intermediate between those obtained by Nongonierma et al.8 and Covarrubias-Cervantes et al.,35 which are always higher, and those obtained by Seuvre et al.,24 which are www.interscience.wiley.com/jsfa Table Enthalpy of affinity (kJ mol−1 ) in water obtained for ethyl acetate, ethyl butyrate and ethyl hexanoate in comparison with literature values Aroma compound Ethyl acetate Ethyl butyrate Ethyl hexanoate Our results Nongonierma et al.8 a Phillippe et al.30 b CovarrubiasCervantes et al.35 c Seuvre et al.24 d 40 52 49 64 55 70 39 52 51 64 63 a ◦ 47 ◦ 46 ◦ 4, 10 and 20 C; b 25, 30 and 37 C; c 0, 4, 10 and 25 C; d 12, 30 and 37 ◦ C always lower (Table 4) However, the same tendency is observed for the investigated esters: an increase in enthalpy of vaporization by increasing the hydrophobic chain length A comparison of the values in water and in the different food matrices shows only a small effect of the matrix on H0 eq for ethyl acetate A small but noticeable decrease in H0 eq is observed for the two other aroma compounds between water and the food matrices, with a lower value in model gel, suggesting that the energy needed for volatilization is lower than in water The same tendency was observed for ethyl butyrate between custard dessert and water by Seuvre et al.,24 for the same three esters between oil-in-water emulsion and water by Philippe et al.,30 and for other compounds between skim milk and water by Meynier et al.,65 whereas these authors found a large increase in H0 eq between water and fat due to strong interactions between aroma compounds and fat CONCLUSIONS By collecting a wide range of gas/water partition coefficients for three aroma compounds at different temperatures, it was possible to establish a linear relation between ln Keq and 1/T in order to predict partition coefficients at temperatures other than experimental A compilation of gas/water partition coefficients obtained for the same aroma compound at different temperatures allowed the calculation of standard thermodynamic equilibrium values which seem to be more accurate than those calculated from c 2010 Society of Chemical Industry J Sci Food Agric 2010; 90: 1285–1292 Retention of aroma compounds in different food matrices data obtained at only three to four temperatures The composition and complexity of food matrix influences to a greater extent gas/matrix partition coefficients and to a lower extent enthalpy of vaporization Yogurts retain the aroma compounds more than model gel matrices, due to the presence of proteins which interact with aroma compounds and lipids which solubilize hydrophobic aroma compounds The observed effects (retention or salting out) strongly depend on hydrophobicity of aroma compounds REFERENCES J Sci Food Agric 2010; 90: 1285–1292 22 Gierczynski I, Labour´e H, S´emon E and Guichard E, Impact of hardness of model fresh cheese on aroma release: in vivo and in vitro study J Agric Food Chem 55:3066–3073 (2007) 23 Ettre LS and Kolb B, Headspace–gas chromatography: the influence of sample volume on analytical results Chromatographia 32:5–12 (1991) 24 Seuvre AM, Turci C and Voilley A, Effect of the temperature on the release of aroma compounds and on the rheological behaviour of model dairy custard Food Chem 108:1176–1182 (2008) 25 Stanford MA and McGorrin J, Flavor Volatilization in Microwave Food Model Systems: Thermally Generated Flavors ACS, Washington, DC (1994) 26 Carelli AA, Crapiste GH and Lozano JE, Activity coefficients of aroma compounds in model solutions simulating apple juice J Agric Food Chem 39:1636–1640 (1991) 27 Harvey BA, Druaux C and Voilley A, Effect of protein on the retention and transfer of aroma compounds at the lipid-water interface, in Food Macromolecules and Colloids, ed by Dickinson E and Lorient D Royal Society of Chemistry, London, pp 154–163 (1995) 28 Sancho MF, Rao MA and Downing DL, Infinite dilution activity coefficients of apple juice aroma compounds J Food Eng 34:145–158 (1997) 29 Landy P, Courthaudon J-L, Dubois C and Voilley A, Effect of interface in model food emulsions on the volatility of aroma compounds J Agric Food Chem 44:526–530 (1996) 30 Philippe E, Seuvre A-M, Schippa C and Voilley A, Flavour compounds behaviour as a function of temperature in a model emulsion, in Flavour Research at the Dawn of the Twenty-First Century: Proceedings of the Weurman Flavour Research Symposium, pp 244–247 (2002) 31 van Ruth SM and Villeneuve E, Influence of beta-lactoglobulin, pH and presence of other aroma compounds on the air/liquid partition coefficients of 20 aroma compounds varying in functional group and chain length Food Chem 79:157–164 (2002) 32 Bylaite E, Ilgunaite Z, Meyer AS and Adler-Nissen J, Influence of lambda-carrageenan on the release of systematic series of volatile flavor compounds from viscous food model systems J Agric Food Chem 52:3542–3549 (2004) 33 Jouquand C, Ducruet V and Giampaoli P, Partition coefficients of aroma compounds in polysaccharide solutions by the phase ratio variation method Food Chem 85:467–474 (2004) 34 Atlan S, Trelea IC, Saint-Eve A, Souchon I and Latrille E, Processing gas chromatographic data and confidence interval calculation for partition coefficients determined by the phase ratio variation method J Chromatogr A 1110:146–155 (2006) 35 Covarrubias-Cervantes M, Champion D, Debeaufort F and Voilley A, Aroma volatility from aqueous sucrose solutions at low and subzero temperatures J Agric Food Chem 52:7064–7069 (2004) 36 Pividal KA, Birtigh A and Sandler SI, Infinite dilution activity coefficients for oxygenate systems determined using a differential static cell J Chem Eng Data 37:484–487 (1992) 37 Le Thanh M, Extraction de substances aromatisantes produites par voie microbiologique: e´tude des interactions entre substances d’arˆome et constituants d’un milieu de culture liquide Th`ese de Doctorat de l’Universit´e de Bourgogne (1992) 38 Le Thanh M, Thibeaudeau P, Thibaut MA and Voilley A, Interactions between volatile and non-volatile compounds in the presence of water Food Chem 43:129–135 (1992) 39 Voilley A and Bosset JO, Nouvelle technique de d´etermination rapide ˆ de la volatilit´e de composants d’arome dans les milieux a` forte viscosit´e Lebensm Wiss Technol 19:47–52 (1986) 40 Voilley A and Sauvageot F, Effet du saccharose sur l’intensit´e de l’odeur ˆ de deux substances d’arome en solution aqueuse Sci Aliment 7:59–64 (1987) 41 Druaux C, Le Than M, Seuvre AM and Voilley A, Application of headspace analysis to the study of aroma compounds-lipids interaction J Am Oil Chem Soc 75:127–130 (1998) 42 Landy P, Fares K, Lorient D and Voilley A, Effect of chemical modification of sodium caseinate on diffusivity of aroma compounds in aqueous solutions J Agric Food Chem 45:2649–2653 (1997) 43 Fares K, Landy P, Guilard R and Voilley A, Physicochemical interactions between aroma compounds and milk proteins: effect of water and protein modification J Dairy Sci 81:82–91 (1998) 44 Le Thanh M, Voilley A and Phan Tan Luu R, Influence de la composition d’un milieu de culture mode` le sur le coefficient de partage vapeurliquide de substances aromatisantes SciAliment 13:699–710 (1993) c 2010 Society of Chemical Industry www.interscience.wiley.com/jsfa 1291 Druaux C and Voilley A, Effect of food composition and microstructure on volatile flavour release Trends FoodSciTechnol 8:364–368 (1997) Guichard E, Interactions between flavor compounds and food ingredients and their influence on flavor perception Food Rev Int 18:49–70 (2002) Guichard E, Flavour retention and release from protein solutions Biotechnol Adv 24:226–229 (2006) Landy P, Druaux C and Voilley A, Retention of aroma compounds by proteins in aqueous solution Food Chem 54:387–392 (1995) Tan Y and Siebert KJ, Modeling bovine serum albumin binding of flavor compounds (alcohols, aldehydes, esters, and ketones) as a function of molecular properties J Food Sci 73:S56–S63 (2008) de Roos KB, How lipids influence food flavour Food Technol 51:60–62 (1997) Seuvre AM, Espinosa Diaz MA and Voilley A, Influence of the food matrix structure on the retention of aroma compounds J Agric Food Chem 48:4296–4300 (2000) Nongonierma AB, Springett M, Le Qu´er´e J-L, Cayot P and Voilley A, Flavour release at gas/matrix interfaces of stirred yoghurt models Int Dairy J 16:102–110 (2006) Goubet I, Le Qu´er´e J-L and Voilley AJ, Retention of aroma compounds by carbohydrates: influence of their physicochemical characteristics and of their physical state A review J Agric Food Chem 46:1981–1990 (1998) 10 Heinemann C, Zinsli M, Renggli A, Escher F and Conde-Petit B, Influence of amylose-flavor complexation on build-up and breakdown of starch structures in aqueous food model systems Lebensm Wiss Technol 38:885–894 (2005) 11 Pozo-Bayon M-A, Biais B, Rampon V, Cayot N and Le Bail P, Influence of complexation between amylose and a flavored model sponge cake on the degree of aroma compound release J Agric Food Chem 56:6640–6647 (2008) 12 Godshall MA, How carbohydrates influence food flavor Food Technol 51:63–67 (1997) 13 Hansson A, Andersson J and Leufven A, The effect of sugars and pectin on flavour release from a soft drink-related model system Food Chem 72:363–368 (2001) 14 Rega B, Guichard E and Voilley A, Flavour release from pectin gels: effects of texture, molecular interactions and aroma compounds diffusion Sci Aliment 22:235–248 (2002) 15 Philippe E, Seuvre AM, Colas B, Langendorff V, Schippa C and Voilley A, Behavior of flavor compounds in model food systems: a thermodynamic study J Agric Food Chem 51:1393–1398 (2003) 16 Rekker RF, The Hydrophobic Fragmental Constant: Its Derivation and Application Elsevier, Amsterdam (1977) 17 Le Thanh M, Lamer T, Voilley A and Jose J, D´etermination des coefficients de partage vapeur-liquide et d’activit´e de compose´ s ˆ d’arome a` partir de leurs caract´eristiques physico-chimiques J Chim Phys 90:545–560 (1993) 18 D´el´eris I, Lauverjat C, Trelea IC and Souchon I, Diffusion of aroma compounds in stirred yogurts with different complex viscosities J Agric Food Chem 55:8681–8687 (2007) 19 Saint-Eve A, Juteau A, Atlan S, Martin N and Souchon I, Complex viscosity induced by protein composition variation influences the aroma release of flavored stirred yogurt J Agric Food Chem 54:3997–4004 (2006) 20 Saint-Eve A, D´el´eris I, Meynier A and Souchon I, Impact of fat composition on diffusion and perception of flavour compounds in yogurts In 12th Weurman Flavour Research Symposium, Interlaken, Switzerland, 1–4 July (2008) 21 Savary G, Guichard E, Doublier JL and Cayot N, Mixture of aroma compounds: determination of partition coefficients in complex semi-solid matrices Food Res Int 39:372–379 (2006) www.soci.org Food safety in China www.soci.org labelled beef Dickinson and Bailey6 and Hobbs et al.7 analysed the willingness of consumers to pay for FTS To date, similar research in developing countries remains limited Studies in China are even more limited China is a very large country comprising various regions with different levels of economic development Chinese consumer behaviour regarding traceability systems is somewhat similar, yet in other ways unique and complicated This paper studies consumer satisfaction with food safety and willingness to pay for CT food The results will assist the government in setting up an FTS, in promoting FTS in developing countries and in improving global food safety standards LITERATURE REVIEW J Sci Food Agric 2010; 90: 1368–1373 DATA SOURCE AND RESEARCH METHODOLOGY The sample data came from Jiangsu Province Research Base of Food Safety’s survey of citizens in 13 cities within Jiangsu, namely Xuzhou, Lian Yungang, Yancheng, Suqian, Huan’an, Taizhou, Yangzhou, Nantong, Zhenjiang, Nanjing, Suzhou, Changzhou and Wuxi The Jiangsu Province Research Base of Food Safety is an important consulting institution for food safety research in Jiangsu Province Jiangsu is an eastern coastal and economically advanced province Its GDP per capita for 2007 was US$4300, which is similar to medium income level countries; thus it is capable of establishing an FTS In 2007, China’s Ministry of Agriculture chose Jiangsu as a trial province for FTS However, the trial was smallscale and only explored a few food categories such as pork and vegetables Based on existing literature, the researchers designed an initial questionnaire on consumer satisfaction with food safety and willingness to pay for CT food A pre-survey of 100 consumers was conducted in Yangzhou According to the pre-survey results, the questionnaire was further revised The final questionnaire consisted of three parts The first part measured consumers’ overall satisfaction with food safety and their perception of FTS The second part assessed their purchasing choices and their willingness to pay for CT food The last part collected their demographic data The survey was conducted from 27 September to 25 October 2008 The survey venues were supermarkets and food stores in the 13 participating cities The main considerations in the survey were that (1) CT foods are mainly available in supermarkets and food stores and (2) supermarkets and food stores are highly prevalent in the 13 cities of Jiangsu The staff of the Jiangsu Province Research Base of Food Safety conducted face-to-face random interviews, based on the contents of the questionnaires, with consumers in three to four main supermarkets or food store entrances in each city The researchers chose respondents of at least 18 years of age For the convenience of a possible follow-up, comparison and analysis, each city had 150 consumers in the survey data The survey retrieved 1950 questionnaires, 1757 of which were valid The collected consumer information features and relevant data were mainly analysed through descriptive statistics Significant differences between different consumer groups were established using the χ test Logistic regression analysis was used to assess the factors that affect consumer purchasing choice and willingness to pay for CT food Table presents the demographic data of the respondents Among the respondents, 51% were males and 49% were females The proportions of respondents aged under 25, 26–40, 41–55 and above 55 were 28, 34, 22 and 16% respectively Sixty-two per cent (62%) of the respondents had family members less than 18 years of age In educational level, 39, 36 and 25% of the respondents were high school or below, college, and university or above respectively c 2010 Society of Chemical Industry www.interscience.wiley.com/jsfa 1369 The recent melamine contamination food safety accident increased consumers’ perception of food safety risk and their worries about food safety Food choice is often influenced more by the psychological interpretation of product properties than by the physical properties of products themselves; this is also true with regard to the perception of food safety risk.8 When all other conditions are the same, risk perception and purchasing behaviour are inversely correlated.8 Different consumers have different perceptions towards food safety/risk.9 – 11 De Jong et al.12 discovered that the perception of food safety/risk among customers varies according to (1) demographic and socio-economic factors such as age, education and income, (2) consumer trust in the administration and food supply chain participant and (3) food safety incidents, consumer knowledge regarding labelling, and media reports on food safety problems, among others Based on the attitude of consumers towards food safety, Knight et al.13 divided consumers into five groups, discovering that these groups of consumers have marked demographic differences such as education, income, allergies in the family, the presence of children below 16 years of age, and so on.13 The more risks consumers can sense, the less confident and less satisfied they are with food safety situations If consumer-perceived risk exceeds a certain threshold or tolerance level, then one or more risk reduction strategies (risk relievers) may be employed to reduce the amount of perceived risk to a tolerable level.14 Consumers may seek information from formal or informal sources such as radio, TV and food labels Where food safety is concerned, certification strategies (traceability or quality labels) have been developed to improve consumer perception of food safety, which ultimately result in price increases Some consumers are willing to pay higher beef prices to be assured of safety, but others trade off the price against safety If adequate risks are perceived, consumers will adopt risk reduction strategies, such as paying higher prices for safety-assured food.15 Risk perception associated with beef is one of the main driving forces for premiums16 and the use of meat labels.17 Angulo et al.5 discovered that food safety incidents, perception and worries about the negative impact of agricultural products on health and the environment are main factors affecting consumer purchasing decisions for different beef products Therefore overall food safety satisfaction of consumers significantly influences their purchasing behaviour Apart from satisfaction with food safety, awareness of products with quality assurance labels, as well as demographic features such as age, gender, family structure, educational level and economic conditions, affect consumer purchasing behaviour and willingness to pay for labelled food Gracia and De Magistris18 discovered that income and organic food knowledge have a positive influence on the final purchasing decision for organic food Similarly, Batte et al.19 studied willingness of consumers to pay for organic food They mainly considered consumer understanding of the National Organic Program seal, age, income, presence of children, education, health, race, gender and other demographic factors They discovered that the presence of children, education, health and understanding of the National Organic Program seal have different levels of influence on consumer willingness to pay for organic food www.soci.org Table Statistical features of respondents Variant Table Features of consumers with different food safety satisfaction Sample population Percentage 904 853 51.5 48.5 Gender Male Female Children in the household? Yes No Age Under 25 26–40 41–55 56 or above Educational background Under high school Diploma degree Bachelor degree Graduate degree Annual incomea Under 30 000 30 000–60 000 60 000–100 000 100 000 or above 1083 674 62 38 486 599 397 275 27.7 34.1 22.6 15.7 682 632 393 50 38.8 36.0 22.4 2.8 843 639 209 66 48.0 36.4 11.9 3.8 a Jiangsu Statistics Bureau reports that Jiangsu urban residents’ per capita income is RMB 17 708 (RMB ≈ US$ 0.133, calculated by the 2007 average exchange rate of the RMB yuan against the US dollar) The majority of the respondents’ annual income fell mainly below RMB 30 000 (48.0%) and between RMB 30 000 and 60 000 (36.4%) Annual incomes between RMB 60 000 and 100 000 and of RMB 100 000 and above accounted for 11.9 and 3.8% of the respondents respectively The 1757 valid questionnaires mainly reflected the supermarket and food store consumer features during the survey period Therefore this survey was conducted with consumers in a specific situation and at a specific time The sample’s gender proportion, education level and age structure were not completely consistent with the overall statistical features of gender, education level and age structure of the citizens in the 13 cities of Jiangsu Province SURVEY RESULTS 1370 Overall satisfaction with food safety and awareness of FTS Some food safety incidents such as the melamine incident have shocked the whole country and increased consumers’ perception of food safety risks Different consumers have different perceptions of food safety risk, and their evaluation of food safety conditions vary When asked whether they were satisfied with the regional food safety conditions, 36% of the respondents said they were very unsatisfied or not satisfied, 50% were fairly satisfied and only about 14% were satisfied or very satisfied This shows that even after the outbreak of the nationally exposed major food safety incidents there are still 64% of the respondents who are not strongly dissatisfied with the food safety conditions Two facts possibly account for this First, even though the Chinese government and the local government have strengthened the checking measures and punishment imposed with food safety issues, the media exposure of food safety incidents is far from the actual situation owing to certain systemic problems www.interscience.wiley.com/jsfa L Xu, L Wu Satifaction with food safety (%) Variable Age∗ Below 25 26–40 41–55 Above 55 Education∗ Below high school College Bachelor or above If read information on food labels∗ Occasionally Often If care about food safety information∗ No or occasionally Sometimes Often or very often If heard of FTS Yes No Very unsatisfied or not satisfied Fairly satisfied Satisfied or very satisfied 12.6 40.2 28.8 18.4 34.6 33.3 19.6 12.5 41.3 21.5 17.4 19.8 30.2 36.7 33.1 41.9 35.4 22.7 49.8 36.0 14.2 37.6 60.0 77.3 19.1 68.8 4.1 14.0 20.2 65.8 21.5 55.1 23.4 44.5 35.5 20.0 40.3 59.7 35.3 64.7 35.2 64.8 ∗ Indicates that significance level of difference between groups is less than 5% that have existed for years Second is the poor educational level of the general public These two facts result in lack of knowledge on food safety and other relevant issues Furthermore, the respondents hardly had any knowledge regarding the labels of CT food Only 37% of the respondents had heard of FTS In order to assess the factors influencing consumer perceptions of food safety, a cross-analysis was conducted among the following factors: consumer food safety perception, consumers’ demographic variables, habit of reading food labels, attention to food safety information over the radio or TV, and awareness of FTS The results are presented in Table The table shows that most of the respondents who were not satisfied with the food safety conditions were between 26 and 55 years old, 73% of them had at least college degrees, 60% possessed common reading habits for food labels and 66% often cared about food quality and safety information aired over the radio or TV Although only 40% of them had heard of FTS, this figure was higher than that for the respondents satisfied or fairly satisfied with the food safety situation Among respondents satisfied with the food safety situation, most were under 25 or their educational level was below high school Only 4% possessed the habit of reading food labels Forty-five per cent (45%) did not care about or only occasionally cared about food quality and safety information aired over the radio or TV Purchasing choice and willingness to pay for CT food The Codex Alimentarius Commission (CAC), International Standards Organisation (ISO), Western scholar Moe20 and Chinese scholars Xia and Song21 have defined traceability and FTS Based c 2010 Society of Chemical Industry J Sci Food Agric 2010; 90: 1368–1373 Food safety in China www.soci.org on their research combined with the actual implementation of Chinese FTS, we have defined FTS as a system where ‘in the process of food production, processing and selling, enterprises or agricultural households record, store and transmit various food quality and safety information which is carried by traceability labels’ Consumers can track or key in the codes of traceability labels through the Internet, supermarket tracking machines, telephones, SMS, etc Information such as the kinds of additives or pesticides and the fertilisers used in the production process can be checked, in addition to other basic producer information Government supervisors can also check and verify the information to control the enterprise behaviour and to realise the overall food quality and safety control from farmlands to dinner tables The notion of FTS was introduced to respondents through the survey, and then the respondents were presented with normal food∗ , green food† , organic food and CT food They were asked which food they were willing to buy If the consumer selected CT food, he/she was asked how much extra charge he/she was willing to pay We take the example of edible vegetables – greens An important reason is that consumers generally eat greens, and China earlier implemented a vegetable traceability system When consumers were asked, it was assumed that the normal market price for greens was yuan kg−1 There were five extra charges that were 0, 1–15, 16–30, 31–50 and >51% higher than the normal price The results showed that 569 of the respondents (32.3% of the total sample) opted to buy CT food Among them, 32% were not willing to pay any extra charge, while 53.1, 11.3, 3.1 and 0.5% of the respondents were willing to pay extra charges of 1–15, 16–30, 31–50 and >51% respectively It is obvious that the extra charges consumers were willing to pay were much lower than the price difference between CT food and normal food In the supermarket, CT vegetables are priced 10–20% higher than organic vegetables Their prices are much higher than those of normal vegetables Taking carrots as an example, at Wal Mart the price of normal carrots is yuan kg−1 , while the price of CT carrots is 11.8 yuan kg−1 , nearly four times as much Meanwhile, rare traceable vegetables such as coloured peppers and broccoli are more expensive than normal coloured peppers and broccoli∗ In China, owing to different standards, there are concepts of normal food, green food and organic food, based on the standard ways of categorising food Normal food has two measured standards of quality and sanitation (safety) These standards focus on product standards and seldom mention production and processing management norms They emphasise result control while ignoring process control Usually it is difficult to ensure the basic product quality of normal food † In China, green food specifically implies planting or breeding under zeropollution conditions, using organic fertiliser and processing according to standard techniques and sanitary standards stipulated by relevant national departments Green foods are also safe, good-quality and nutritious foods accredited by national authority agencies In China, green food production/promotion is in its early stages and is relatively highly popular Comparatively speaking, organic food has developed slowly and its market penetration rate is not as high as that of green food It can be considered that China’s green food is in between normal food and organic food ∗ Data source: CT Food Suffers Low Popularity due to Limiting Quantity and Expensive Price, 21 Food and Beverage online, 2007-12-07, http://www.21food.cn/html/news/26/245075.htm.] J Sci Food Agric 2010; 90: 1368–1373 B f (Bi ) = p1i (1 − p1 )1−Bi ; Bi = 0, PP f (PPi ) = p2 i (1 − p2 )1−PPi ; PPi = 0, This paper used the binary logit model to estimate the regression parameters of the various factors affecting consumer purchasing choice and willingness to pay for CT food The basic forms of the two logit models are as follows: p1i = F(λB + β B xiB + α B ATi + viB ) = 1/{1+ exp[−(λB + β B xiB + α B ATi + viB )]} p2i = F(λPP + β PP xiPP + α PP ATi + viPP ) = 1/{1+ exp[−(λPP + β PP xiPP + α PP ATi + viPP )]} Among them, vectors of explanatory variables xiB and xiPP were used to explain purchasing choice and willingness to pay for CT food respectively, with corresponding constants λB and λPP , parameter vectors β B and β PP and random errors viB and viPP The complete list of variables included in the model is presented in Table The explanatory variables of the two models measure the respondents’ awareness of FTS, the respondents’ satisfaction with food safety conditions and the respondents’ demographic features such as age, gender, education, income, etc The parameter estimation results of the two equations are given in Table Among the explanatory factors affecting consumer purchasing choice, as indicated above, consumer overall satisfaction with food safety has a negative and significant influence on purchasing choice The more dissatisfied consumers are with the overall food safety conditions, the more likely they are to purchase CT food The awareness of FTS has a positive and significant influence on consumer purchasing choice Consumers who have prior knowledge of FTS are more likely to buy CT food Among the demographic variables, gender and medium educational level have negative effects Female consumers are more likely to buy CT food, and consumers with medium educational level are less likely to buy CT food Consumers with kids below 18 are more likely to buy CT food, and consumers aged 26–40 years are more likely to buy CT food Among the explanatory variables affecting consumer willingness to pay for CT food, consumers’ overall satisfaction with food safety negatively and significantly influences consumer willingness to pay for CT food The more dissatisfied consumers are with the overall food safety conditions, the more likely they are to pay extra charges for CT food This is consistent with the research findings of Loureiro and Umberger.16 Awareness of FTS has a positive c 2010 Society of Chemical Industry www.interscience.wiley.com/jsfa 1371 ∗ Analysis of influencing factors for purchasing choice and willingness to pay for CT food The research applied the logit model to estimate the degrees of influence of identified factors on respondents’ purchasing choice and willingness to pay for CT food The two dependent variables were defined as (1) choice to buy CT Food (Bi ) and (2) willingness to pay for CT food (PPi ) If the respondents chose to buy CT food, then Bi = 1, otherwise Bi = If the consumers were willing to pay more than 0% extra charge, then PPi = 1, otherwise PPi = Suppose the probabilities of Bi = and PPi = are p1 and p2 respectively Therefore we utilise the following two equations: www.soci.org Table Definitions of variables and sample statistics Variable Table Regression results for purchasing choice and willingness to pay for CT food Definition and measurement Mean SD Whether respondent chose to buy CT food (yes = 1, no = 0) Whether respondent is willing to pay for CT food (yes = 1, no = 0) Respondent’s overall satisfaction with food safety condition (dummy variable that takes the value if the respondent was satisfied or very satisfied with the overall food safety condition, and otherwise) Whether respondent heard of FTS before (dummy variable, yes = 1, no = 0) Whether there are kids under 18 in respondent family (dummy variable, yes = 1, no = 0) Dummy variable, male = 1, female = Low age level (dummy variable, = 26–40 years old, = otherwise) Medium age level (dummy variable, = 41–55 years old, = otherwise) High age level (dummy variable, = above 55 years old, = otherwise) High education level (dummy variable that takes the value if the respondent has a bachelor degree or above, and otherwise) Medium education level (dummy variable that takes the value if the respondent has a junior college degree, and otherwise) High income level (dummy variable that takes the value if the household’s annual income is higher than RMB 60 001, and otherwise) Medium income level (dummy variable that takes the value if the household’s annual income lies between RMB 30 000 and 60 000, and otherwise) 0.32 0.464 0.68 0.451 0.14 0.348 0.37 0.483 0.62 0.486 Variable Bi PPi Satisfy Aware Kid Gender LowAge MedAge HighAge HighEdu MediumEdu HighInc MediumInc Constant Satisfy Aware Kid Gender LowAge MedAge HighAge HighEdu MediumEdu HighInc MediumInc Purchasing choice Willingness to pay −1.236∗∗ −0.478∗∗ 1.171∗∗ 0.222∗ −0.323∗∗ 0.333∗ 0.067 0.131 −0.097 −0.272∗ 0.270 0.152 −0.862∗∗ −0.856∗∗ 0.888∗∗ 0.442∗ 0.350 −0.908∗∗ −0.522 −0.863∗∗ 0.358 0.671∗∗ 2.640∗∗ 1.910∗∗ Asterisks indicate that estimated coefficients are significant at ∗ 5% or ∗∗ 1% level of confidence 0.49 0.500 0.34 0.474 0.23 0.418 0.16 0.363 0.25 0.434 0.36 0.480 0.16 0.363 0.36 0.481 SD, standard deviation 1372 and significant influence on consumer willingness to pay for CT food Consumers who have heard of FTS are more likely to pay extra charges for CT food Similarly, Batte et al.19 indicated that the more consumers understand the National Organic Program seal, the more likely they are to pay extra for different natural organic food Further, consumers with kids below 18 and those with medium educational level are more likely to pay extra charges for CT food As consumer income increases, consumers are more likely to pay extra charges for CT food Similarly, Angulo and Gil15 and Castro22 discovered that consumers with higher income are more willing to pay higher charges for CT food Consumers aged 26–40 years or above 55 years are less likely to pay extra charges for CT food www.interscience.wiley.com/jsfa L Xu, L Wu CONCLUSION AND IMPLICATIONS The recent food safety incidents and the latest nationwide melamine powdered infant formula contamination accident have adversely increased consumer perception towards food safety risk The more risk perception, the more dissatisfied consumers are with the current food safety situation This study found that 36% of the respondents are strongly dissatisfied with the regional food safety situation Consumers with different characteristics and different lifestyles have different levels of satisfaction with food safety Young people and consumers with a low level of education are relatively satisfied with the food safety situation They generally not often read food labelling information and not concern themselves about food safety information aired over the radio or TV China has limited media coverage and reports on food safety Consequently, consumers’ knowledge of the food safety problems is limited Consumer satisfaction with food safety significantly influences consumers’ purchasing choice and willingness to pay for CT food Therefore it is necessary to increase media coverage of food safety knowledge This can help consumers to correctly understand the food safety situation Only 37% of the respondents have heard of FTS One reason is that the publicity of FTS and CT food is inadequate Regression analysis shows that the awareness of FTS affects consumer purchasing choice and willingness to pay for CT food Therefore the publicity of FTS and its function should be reinforced Recommended measures include the set-up and maintenance of information databases to educate people on how to check and understand traceability labels (information carrier) People should also be informed of the proper channels for food-related complaints Five hundred and sixty-nine respondents (32.3% of the total sample) chose to purchase CT food Among these respondents, 32% were not willing to pay extra charges for CT food Consumer willingness to pay for CT food is very low in China Other than satisfaction with food safety conditions and the awareness of FTS, consumers’ willingness to pay for CT food is also highly affected by income If the CT food price is not acceptable or affordable to consumers, the implementation and promotion of FTS will be difficult Therefore government funding support is very important This study suggests that, in the different stages of FTS construction, the CT food price should be reasonably c 2010 Society of Chemical Industry J Sci Food Agric 2010; 90: 1368–1373 Food safety in China www.soci.org controlled In the initial stage, government subsidies can be employed to enhance consumers’ acceptability of FTS When consumers become used to FTS, they can be required to share in the production cost As the level of income increases, the government can gradually increase the consumers’ share in the price of food What’s worth mentioning is that this research dwelt on CT green food as an example to assess consumer willingness to purchase, without considering the willingness to pay for and accept CT food from different production types and techniques This is the limitation of this paper As stated above, CT food has only been limitedly tried in China Consumers know less about the techniques adopted in CT food processing, so it is difficult for them to make any comparative analysis It is certain that, with the increasing popularity of China’s CT food, consumer awareness will increase To further understand consumer acceptance of different types of CT food with different techniques is the focus of the authors in the near future ACKNOWLEDGEMENTS This research is supported by a grant from the 2009 National Soft Science Project of China (2009GXS5B078) and a grant from the 2009 Doctoral Scientific Research Fund Project of Jiangnan University We would also like to thank six students from Jiangnan University for carrying out the survey REFERENCES Verbeke W, The emerging role of traceability and information in demand-oriented livestock production Outlook Agric 30:249–255 (2001) Hobbs JE, Consumer demand for traceability Paper presented at the IATRC Annual Meeting, December 15–17, 2002, Monterey, California Van Rijswijk W, Frewer LJ, Menozzi D and Faioli G, Consumer perceptions of traceability: a cross-national comparison of the associated benefits Food Qual Prefer 19:452–464 (2008) Opara LU and Mazaud F, Food traceability from field to plate Outlook Agric 30:239–247 (2001) Angulo AM, Gil JM and Tamburo L, Food safety and consumers’ willingness to pay for labelled beef in Spain J Food Prod Market 11:89–105 (2005) Dickinson DL and Bailey D, Meat traceability: are U.S consumers willing to pay for it? J Agric Resource Econ 27:348–364 (2002) Hobbs JE, Bailey D, Dickinson DL and Haghiri M, Traceability in the Canadian red meat sector: consumers care? Can J Agric Econ 53:47–65 (2005) Yeung RMW and Morris J, Food safety risk: consumer perception and purchase behaviour Br Food J 103:170–186 (2001) Nayga RM, Toward an understanding of consumers’ perceptions of food labels IFAMR 2:29–45 (1999) 10 Houghton JR, Rowe G, Frewer LJ, Van Kleef E, Chryssochoidis GM and Kehagia O, The quality of food risk management in Europe: perspectives and priorities Food Policy 33:13–26 (2008) 11 Van Kleef E, Rowe G, Houghton J, Chryssochoidis G, Korzen-Bohr S and Crystallis T, Perceptions of food risk management among key stakeholders: results from a cross European study Appetite 47:46–63 (2006) 12 De Jonge J, Frewer L, van Trijp H, Renes RJ, de Wit W and Timmers J, The development of a monitor for consumer confidence in food safety: results of an exploratory study Br Food J 106:837–849 (2004) 13 Knight AJ, Worosz MR and Todd ECD, Segmentation of US consumers based on food safety attitudes Br Food J 110:691–705 (2008) 14 Mitchell VW, Consumer perceived risk: conceptualisations and models Eur J Market 33:163–195 (1998) 15 Angulo AM and Gil JM, Food safety and consumers’ willingness to pay for labelled beef in Spain Food Qual Prefer 18:1106–1117 (2007) 16 Loureiro ML and Umberger WJ, A choice experiment model for beef: what US consumer preferences tell us about relative preferences for food safety, country of origin labelling and traceability Food Policy 32:496–514 (2007) 17 Becker T, Consumer perception of fresh meat quality: a framework for analysis Br Food J 102:158–176 (2000) 18 Gracia A and De Magistris T, Organic food product purchase behaviour: a pilot study for urban consumers in the South of Italy Span J Agric Res 5:439–451 (2007) 19 Batte MT, Hooker NH, Haab TC and Beaverson J, Putting their money where their mouths are: consumer willingness to pay for multiingredient, processed organic food products Food Policy 32:145–159 (2007) 20 Moe T, Perspectives on traceability in food manufacture Trends Food Sci Technol 9:211–214 (1998) 21 Xia Y and Song B, Food safety assurance from quality standard system to supply chain comprehensive management Agric Econ Issues 11:59–62 (2001) 22 Castro PD, Mechanization and traceability of agricultural products: a challenge for the future Paper presented at the Club of Bologna meeting, July 27, 2002 Chicago, IL., USA 1373 J Sci Food Agric 2010; 90: 1368–1373 c 2010 Society of Chemical Industry www.interscience.wiley.com/jsfa Research Article Received: 12 August 2009 Revised: October 2009 Accepted: October 2009 Published online in Wiley Interscience: 22 March 2010 (www.interscience.wiley.com) DOI 10.1002/jsfa.3934 Determination of enantiomeric fractions of cypermethrin and cis-bifenthrin in Chinese teas by GC/ECD Hua Kuang,a,b Hong Miao,c Xiaolin Hou,c Yunfeng Zhao,c Jianzhong Shenb and Yongning Wuc∗ Abstract BACKGROUND: The pyrethroids cypermethrin and cis-bifenthrin are the major pesticides used in tea plantations in China Potential neurotoxic, genotoxic and immunotoxic effects of chronic exposure to pyrethroids have been reported All synthetic pyrethroids are chiral compounds There is a need to investigate the chiral transformation among isomers after tea processing in order to obtain an accurate risk assessment of these compounds RESULTS: The enantiomeric fraction (EF) of cis-bifenthrin residues was close to 0.5 in all tea samples tested, showing that the levels of (+)-isomer and (−)-isomer were equivalent and there was no preferential degradation However, the patterns of EFs of cypermethrin residues varied depending on the type of tea The EF of isomer 1R-3R-αS increased in black and dark tea samples (EF = 0.200–0.343) compared with the reference cypermethrin commercial mixture (EF = 0.116) In one oolong tea sample it was found that the relative abundance of some isomers was preferentially enhanced: 1R-3R-αS (EF = 0.260), 1S-3S-αR (EF = 0.263) and 1R-3S-αS/1S-3R-αR (the last elution peak, EF = 0.275) The relationship between EF and compound concentration was also analysed, showing that the variation in EFs of cypermethrin was concentration-dependent CONCLUSION: These findings appear to be useful for assessing the species-specific risk of exposure to cypermethrin and cis-bifenthrin c 2010 Society of Chemical Industry Keywords: chiral; enantiomeric fractions; tea; GC/ECD; cypermethrin; bifenthrin INTRODUCTION 1374 Photostable synthetic pyrethroids have been used in agricultural insecticides for over 30 years; however, they have proved to be highly toxic to some fish species, as they are easily transferred from water to fish blood via the gills.1 – Previously, pyrethroid pesticides were regarded as being of low toxicity and safe to humans, but more and more reports have shown that chronic exposure to these compounds can result in neurotoxic,4 genotoxic5 and immunotoxic6 effects.7,8 Pyrethroids are composed of an acid moiety (cyclopropane ring) and an alcohol moiety in their chemical structure There are two chiral carbons in the acid moiety For example, bifenthrin (Fig 1) has two pairs of diastereomers, which are commonly designated as cis and trans based on the orientation of the substituents at C-1 and C-3 relative to the cyclopropane ring Cypermethrin (Fig 1) contains another chiral carbon in the alcohol moiety (α-cyano carbon), so, technically, cypermethrin formulations consist of eight steroisomers Pyrethroid insecticides are synthesised and sold in the form of the single, most active isomer or in mixtures containing two, four or eight different stereoisomers At present, commercial formulations of bifenthrin contain only cis-bifenthrin in the two enantiomeric forms 1S-3S and 1R-3R.9 Enantiomers of different configuration show different insecticidal activity.10 – 13 For example, only the 1R-cis-αS and 1R-trans-αS forms of cypermethrin show insecticidal J Sci Food Agric 2010; 90: 1374–1379 activity.14 Most cypermethrin formulations are used in racemic form, and common determinations are based on the total amount of residues using achiral separation.15 – 19 Little information on residue levels has been reported for individual steroisomers However, chiral compounds are degraded selectively in the environment, and knowledge of the real residue levels of individual isomers is of great value in risk evaluation The use of suitable chiral stationary phases makes it possible to determine traces of synthetic pyrethroids in biota.9,20 – 22 Non-racemic proportions of chiral compounds result from enantioselective bioprocesses in the environment, such as metabolism by different enzymes and degradation There have ∗ Correspondence to: Yongning Wu, National Institute for Nutrition and Food Safety, Chinese Center for Disease Control and Prevention, Beijing 100050, China E-mail: echokh@yahoo.cn a School of Food Science and Technology, Jiangnan University, Wuxi, JiangSu 214122, China b Department of Pharmacology and Toxicology, College of Veterinary Medicine, China Agricultural University, Beijing 100094, China c National Institute for Nutrition and Food Safety, Chinese Center for Disease Control and Prevention, Beijing 100050, China www.soci.org c 2010 Society of Chemical Industry Determination of cypermethrin and cis-bifenthrin in Chinese teas www.soci.org H H Cl H C α COO C O CH Cypermethrin CN Cl H CH3 H F3C C CH Bifenthrin COOCH2 Cl Figure Chemical structures of cypermethrin and bifenthrin been some reports on the enantioselective degradation of chiral compounds Wang et al.23 investigated the enantiomeric composition of organochlorine pesticides in soil Wang et al.24 found that the (−)-enantiomer of θ -cypermethrin was enriched in the plasma, fat, muscle, liver, heart and kidney of rats Tea is a popular drink with a long history, and China has the largest area of tea plantations in the world However, tea plants are susceptible to invasive pests, and pyrethroid pesticides are widely used to control such insects Cypermethrin and cisbifenthrin are the pyrethroid pesticides most commonly applied in tea plantations in China.25 In this study we analysed the contamination status of cypermethrin (CP) and cis-bifenthrin (cisBF) in marketed tea samples using a gas chromatograph (GC)/ion trap mass spectrometer (MS) For those samples verified positive, a GC equipped with an electron capture detector (ECD) was used to investigate the possible preferential degradation in different types of tea based on a chiral capillary column MATERIALS AND METHODS Chemicals Racemic standards of CP (purity 98%; isomers: A cis 26%, B cis 20%, C trans 26%, D trans 19%, according to the certificate provided by the manufacturer; four resolved peaks can be obtained on an achiral column) and racemic cis-BF (purity 98%) were purchased from Dr Ehrenstorfer GmbH (Augsburg, Germany) The stereoconfigurations of CP and cis-BF are shown in Fig All solvents used in this study were of pesticide residue grade All other reagents were of analytical grade or better J Sci Food Agric 2010; 90: 1374–1379 Calculation of contamination levels of tea samples The contamination levels of tea samples were detected and confirmed using a Varian 3800 GC (Walnut Creek, CA, USA) (equipped with a 1079 injector)/Varian Saturn 2000 ion trap MS on an achiral capillary column (VF-5 ms low-bleed column, 30 m × 0.25 mm × 0.25 µm, Varian, Inc., Walnut Creek, CA, USA) The GC conditions were as follows: injector temperature 280 ◦ C; spitless injection mode, injection volume µL; helium carrier gas, constant flow rate mL min−1 The oven temperature was initially held at 60 ◦ C for min, then increased to 150 ◦ C at 25 ◦ C min−1 , increased to 250 ◦ C at ◦ C min−1 and finally increased to 280 ◦ C at ◦ C min−1 and held for 5.23 The total run time was 35 MS detection was performed in electron impact tandem MS (EI-MS/MS) mode with a trap temperature of 220 ◦ C, a transfer line temperature of 260 ◦ C and a manifold temperature of 45 ◦ C Quantification was based on the external standard method The calibration curve was constructed as a plot of the combined area of all elution peaks versus the concentration of each compound Chiral separation and enantiomeric fractions of cis-BF and CP The primary analysis was performed using a Shimadzu 2010 GC (Kyoto, Japan) equipped with an ECD and a BGB-172 chiral column (30 m × 0.25 mm × 0.25 µm, 20% tert-butyldimethylsilyl-βcyclodextrin in 15% diphenyl- and 85% dimethylpolysiloxane (v/v); BGB Analytik, Adliswil, Switzerland) The temperature of the ECD was 270 ◦ C, the carrier gas was helium at 50 cm s−1 and the make-up gas was nitrogen at 30 mL min−1 The oven temperature was initially held at 160 ◦ C for min, then increased to 220 ◦ C at ◦ C min−1 and held for 60 and finally increased to 230 ◦ C at ◦ C min−1 and held for 40 min.26 Enantiomeric fractions (EFs) were calculated using the modified formula given by Harner et al.,27 i.e EF = Ei /( ni=1 Ei ), where Ei is the ith eluting enantiomer and n is the total number of elution peaks Enantiomers were identified by their retention times according to Liu and Gan.26 RESULTS Chiral separation of cis-BF and CP Baseline separation of cis-BF was achieved under the analytical conditions used, with two peaks being resolved on the BGB-172 c 2010 Society of Chemical Industry www.interscience.wiley.com/jsfa 1375 Sample preparation For each sample, g of ground tea powder was soaked with 15 mL of hot water (90–100 ◦ C) in a tube for h before extraction After cooling to room temperature, 30 mL of acetone was added and the tube was placed in a water bath for ultrasonic extraction for 20 The extract (20 mL) was partitioned with 25 mL of hexane The upper solution was collected for evaporation, and mL of petroleum ether was used to reconstitute the residues Clean-up was carried out using a sandwich glass cartridge (40 cm × 12 mm i.d.) packed with g of anhydrous sodium, g of florisil and g of anhydrous sodium from bottom to top The column was conditioned with 20 mL of petroleum ether before loading the extract, and the retained targeted compounds were finally eluted using 100 mL of petroleum ether/diethyl ether (85 : 15 v/v) The cleaned extract was condensed to 200 µL for analysis www.soci.org H Kuang et al Figure Separation of (a) cis-bifenthrin and (b) cypermethrin Chromatographic conditions: BGB-172 chiral column (30 m ì 0.25 mm ì 0.25 àm); temperature programme 160 ◦ C (2 min) → 220 ◦ C (at ◦ C min−1 , 60 min) → 230 ◦ C (at ◦ C min−1 , 40 min); helium carrier gas flow rate 50 cm s−1 ; GC/ECD temperature 270 ◦ C Table Enantiomeric fractions (EFs) of separated peaks of cis-bifenthrin and cypermethrin in reference standards Pesticide Retention time (min) Configuration EF Peak 68.903 69.587 127.091 128.683 131.093 132.117 133.873 134.405 (+)-Enantiomer (−)-Enantiomer 1R-3R-αR 1S-3S-αS 1R-3S-αR/1S-3R-αS 1R-3R-αS 1S-3S-αR 1R-3S-αS/1S-3R-αR 0.494 ± 0.009 0.505 ± 0.009 0.167 ± 0.007 0.170 ± 0.004 0.275 ± 0.015 0.116 ± 0.006 0.130 ± 0.011 0.142 ± 0.006 2 cis-Bifenthrin Cypermethrin column (Fig 2a) Six peaks of eight cypermethrin isomers were resolved on the BGB-172 column (Fig 2b); peaks and were not baseline separated Liu and Gan26 previously reported that no enantiomer conversion was found under these analytical conditions The EFs of separated peaks of CP and cis-BF reference standards were determined with standard solutions of different concentrations (1, 0.2 and 0.05 mg kg−1 ) in three repetitions The data are shown in Table 1376 Patterns of EFs of cis-BF and CP in tea samples The contamination levels of 19 tea samples were determined by GC/ion trap MS (Table 2) Eleven tea samples were found to contain cis-BF residues, with levels of contamination ranging from 14.25 to 3071.29 µg kg−1 The EFs of the first peak of cis-BF ranged from 0.484 to 0.510, approaching a value close to the racemic mixture (EF = 0.5) and consistent with the reference standard (EF = 0.494 ± 0.009 for the first elution peak) Seventeen tea samples were found to contain CP residues Typical chromatograms of CP-contaminated samples of different types of tea are shown in Fig The chromatogram obtained for black tea samples (Fig 3a) often featured an interfering peak between the second and third elution peaks of CP The same www.interscience.wiley.com/jsfa elution pattern was observed in dark tea (Fig 3b) and scented tea (Fig 3d) samples Occasionally, the interfering peak appeared in similar elution order in oolong tea samples (Fig 3e), but it was not found in green tea samples (Fig 3c) in the present study From Tables and the values of EF4 (1R-3R-αS) of two black tea (0.228–0.301) and three dark tea (0.200–0.343) samples were significantly different from that of the reference standard (0.116 ± 0.006) The values of EF5 (1S-3S-αR) and EF6 (1R-3SαS/1S-3R-αR) of two oolong tea samples were 0.183–0.263 and 0.222–0.275 respectively, considerably higher than those of the reference standards (0.130±0.011 and 0.142±0.006 respectively) For scented tea samples a clear deviation of EF4 (0.134–0.190) compared with the value of the reference standard (0.116 ± 0.006) was found There was no noticeable deviation of the EFs of six green tea samples in comparison with the values of the reference standards DISCUSSION Pyrethroid pesticides are a class of insecticides with high boiling points (1.8 × 10−4 mPa at 20 ◦ C for CP) and are always the last eluting compounds on the capillary column Tea is a very complex matrix consisting of various amounts of organic acids and c 2010 Society of Chemical Industry J Sci Food Agric 2010; 90: 1374–1379 Determination of cypermethrin and cis-bifenthrin in Chinese teas www.soci.org Table Enantiomeric fractions (EFs) of separated peaks and contamination levels (CLs) of cis-bifenthrin (cis-BF) and cypermethrin (CP) in tea samples Tea type Black Dark Oolong Scented Green EFs of cis-BF EFs of CP Sample CL of cis-BF (µg kg−1 ) CL of CP (µg kg−1 ) EF1 EF2 EF1 EF2 EF3 EF4 EF5 EF6 B-1 B-2 D-1 D-2 D-3 D-4 O-1 O-2 O-3 S-1 S-2 S-3 S-4 G-1 G-2 G-3 G-4 G-5 G-6 – – 39.82 42.12 – 40.23 – – 75.85 233.30 3071.29 421.32 – – 14.25 – 110.77 18.02 43.68 32.84 49.23 69.41 67.54 42.98 – 39.02 43.45 – 30.13 32.38 29.54 31.65 96.12 137.48 144.18 20.13 187.65 145.02 – – 0.491 0.497 – 0.510 – – 0.494 0.482 0.501 0.492 – – 0.487 – 0.484 0.489 0.488 – – 0.509 0.503 – 0.490 – – 0.505 0.517 0.499 0.508 – – 0.513 – 0.516 0.511 0.512 0.146 0.116 0.145 0.123 0.102 – 0.050 0.121 – 0.151 0.168 0.136 0.165 0.172 0.172 0.172 0.165 0.174 0.178 0.141 0.114 0.152 0.126 0.099 – 0.060 0.120 – 0.141 0.165 0.125 0.144 0.170 0.174 0.171 0.159 0.171 0.172 0.254 0.260 0.251 0.231 0.218 – 0.090 0.224 – 0.242 0.271 0.249 0.270 0.271 0.273 0.269 0.333 0.279 0.272 0.228 0.301 0.200 0.276 0.343 – 0.260 0.130 – 0.180 0.134 0.190 0.137 0.106 0.114 0.114 0.102 0.114 0.116 0.120 0.096 0.124 0.113 0.093 – 0.263 0.183 – 0.135 0.122 0.124 0.132 0.133 0.126 0.129 0.117 0.122 0.122 0.112 0.113 0.128 0.131 0.144 – 0.275 0.222 – 0.152 0.141 0.176 0.158 0.148 0.142 0.146 0.126 0.140 0.140 J Sci Food Agric 2010; 90: 1374–1379 Table lists the contamination levels of CP-positive samples The changes in CP levels of green tea samples (from 20.13 to 187.65 µg kg−1 ) were not reflected by the data on EFs, except that the EF of peak (1R-3S-αR/1S-3R-αS) of sample G-4 (20.13 µg kg−1 CP level) was increased in comparison with other green tea samples Similarly, differences in EFs were also observed in dark tea samples, where the EF value of peak (1S-3S-αR) at high CP detected levels (69.41 µg kg−1 ) was lower than EF of peak3 at low detected level (42.98 µg kg−1 ) The value of EF3 of oolong tea sample O-1 (0.090) was significantly lower than that of the reference standard (0.275 ± 0.015), but this was not found in other oolong tea samples Although the number of tea samples was very limited in these tests, the interesting phenomena observed merit further investigation CONCLUSION The two pyrethroid pesticides cis-BF and CP, which are the most widely used insecticides in tea plantations in China, were analysed using a chiral capillary column with ECD The EF of each separated peak was calculated based on the peak area for cis-BF and the peak height for CP The EFs of chiral compounds were compared with the certified reference materials The data showed that no apparent isomer conversion was detectable for cis-BF in 11 tea samples, while some isomers of CP in 17 tea samples degraded faster depending on the type of tea However, changes in the enantiomeric composition of pyrethroids during tea production are still not well understood and further investigations are necessary ACKNOWLEDGEMENTS Financial support for this work was obtained from the National Natural Science Foundation of China (NSFC) (30700664) and the c 2010 Society of Chemical Industry www.interscience.wiley.com/jsfa 1377 derivatives of catechins and alkaloids.28 Separated-out crystals are often found on the walls of micro-vials after drying the solvent from the extracts The interfering substances co-eluting with CP (Figs 3a, 3b and 3d) are probably some similar constituents with high boiling points in tea extracts However, we could not find any relevant reports on this Bidleman and Falconer29 reported that physical processes (leaching, volatilisation and atmospheric deposition) and abiotic reactions (hydrolysis and photolysis) did not affect the EFs of chiral compounds Therefore the data on EFs reflect biodegradation or biotransformation occurring during tea plant growth, manufacture and storage Pyrethroid pesticides are generally applied to the leaves, so little biological transformation takes place during tea plant growth Teas are usually classified on the basis of ‘processing methods’, with ‘degree of fermentation’ being one of the most important characteristics During tea fermentation the main process occurring is the oxidation of catechins, and the tea undergoes a series of biochemical reactions.30 Black teas and dark teas are fully fermented, oolong teas are semi-fermented and green teas are unfermented Scented teas are processed by blending with flowers (e.g jasmine) under high-moisture/temperature (30–48 ◦ C) conditions to absorb their aroma, and a post-fermentation reaction may occur (autoxidation of tea polyphenols) The small deviations of cis-BF from racemic composition observed in the prepared tea samples showed that little biotransformation of cis-BF took place in tea manufacture The different patterns of EFs of CP enantiomers showed preferential enrichment of some enantiomers depending on the type of tea There are some possible explanations for these results The direction of biotransformation depends on the type of tea sample On the other hand, different degrees of fermentation affect the CP chiral metabolism, because no notable variation in the EFs of CP residues in green tea samples was found www.soci.org (b) uV(x10,000) uV(x10,000) (a) H Kuang et al 7.0 6.0 56 12 10.0 7.5 5.0 4.0 56 5.0 3.0 2.5 2.0 109.5 120.5 131.5 uV(x100,000) (c) 1.75 107.2 116.8 126.4 1.50 7.0 1.25 1.00 135.0 (d) uV(x10,000) 8.0 6.0 56 5.0 0.75 4.0 0.50 3.0 0.25 2.0 126.0 135.0 115.6 125.6 135.6 (e) uV(x1,000) 1.00 0.75 0.50 0.25 0.00 117.5 120.0 122.5 125.0 127.5 130.0 132.5 135.0 Figure Typical chromatographs of cypermethrin residues in (a) black tea, (b) dark tea, (c) green tea, (d) scented tea and (e) oolong tea samples: peak 1, 1R-3R-αR; peak 2, 1S-3S-αS; peak 3, 1R-3S-αR/1S-3R-αS pair; peak 4, 1R-3R-αS; peak 5, 1S-3S-αR; peak 6, 1R-3S-αS/1S-3R-αR pair National High-tech Research and Development Project of China (2007AA06Z404) REFERENCES 1378 Hill IR, Effects on non-target organisms in terrestrial and aquatic environments, in The Pyrethroid Insecticides, ed by Leahey JP Taylor and Francis, London, pp 151–262 (1985) Hao XL, Kuang H, Li YL, Yuan Y, Peng CF, Chen W, et al., Development of an enzyme-linked immunosorbent assay for the α-cyano pyrethroids multi-residue in Tai Lake water J Agric Food Chem 57:3033–3039 (2009) Li ZK, Zhu YY, Yin XG, Peng CF, Chen W, Liu LQ, et al., Development of an indirect enzyme-linked immunosorbent assay for the organophosphorus pesticide paraoxon-methyl Immunol Invest 38:508–523 (2009) Patro N, Shrivastava M, Tripathi S and Patro IK, S100 beta upregulation: a possible mechanism of deltamethrin toxicity and motor coordination deficits Neurotoxicol Teratol 31:169–176 (2009) DOI: 10.1016/j.ntt.2008.12.001 Ila HB, Topaktas M, Rencuzogullari E, Kayraldiz A, Donbak L and Daglioglu YK, Genotoxic potential of cyfluthrin Mutat Res 656:49–54 (2008) Righi DA, Xavier FG and Palermo-Neto J, Effects of type II pyrethroid cyhalothrin on rat innate immunity: a flow cytometric study Int Immunopharmacol 9:148–152 (2009) www.interscience.wiley.com/jsfa Xu CL, Jin ZY and Yang SQ, Polyamines induced by heat treatment before cold-storage reduce mealiness and decay in peach fruit J Hort Sci Biotechnol 80:557–560 (2005) Xu CL, Hao K, Peng CF, Cai WH and Jin ZY, Effects of fermentation temperature and time on the physicochemical and sensory characteristics of douchi Food Technol Biotechnol 43:307–311 (2005) Salem H and Olajos EJ, Review of pesticides: chemistry, uses and toxicology Toxicol Ind Health 4:291–321 (1988) 10 Peng CF, Li ZK, Zhu YY, Chen W, Yuan Y, Liu LQ, et al., Simultaneous and sensitive determination of multiplex chemical residues based on multi-color quantum dot probes Biosens Bioelectron 24:3657–3662 (2009) 11 Kuang H, Wu YN, Hou XL, Miao H, Zhang G, Shen JZ, et al., Synthesis of derivatives and production of antiserum for class specific detection of pyrethroids by indirect ELISA Int J Environ Anal Chem 89:423–437 (2009) 12 Peng CF, Kuang H, Chu XG and Xu CL, Evaluation and interlaboratory validation of a GC–MS method for analysis of pesticide residues in teas Chem Papers 61:1–5 (2007) 13 Wang LB, Li C, Peng CF, Li XQ and Xu CL, A rapid multi-residue determination method of herbicides in grain by GC/MS (SIM) J Chromatogr Sci 46:424–429 (2008) 14 Liu W, Gan J, Lee S and Werner I, Isomer selectivity in aquatic toxicity and biodegradation of bifenthrin and permethrin Environ Toxicol Chem 24:1861–1866 (2005) c 2010 Society of Chemical Industry J Sci Food Agric 2010; 90: 1374–1379 Determination of cypermethrin and cis-bifenthrin in Chinese teas 15 Liu W and Gan J, Determination of enantiomers of synthetic pyrethroids in water by solid phase microextraction–enantioselective gas chromatography J Agric Food Chem 52:736–741 (2004) 16 Tagami T, Kajimura K, Nomura C, Taguchi S and Iwagami S, Determination of pyrethroid pesticides in cinnamomi cortex Yakugaku Zasshi 129:173–176 (2009) 17 Kuang H, Chu XG, Hou YX and Xu CL, Simultaneous determination of 13 phenoxy acid herbicide residues in soybean by GC–ECD Anal Lett 39:2617–2627 (2006) 18 Xu CL, Peng CF, Hao K, Jin ZY and Chu XG, Simultaneous determination of anabolic steroids residues in animal muscle tissues by gas chromatography–mass spectrometry Anal Lett 39:555–568 (2006) 19 Xu CL, Chu XG, Peng CF, Jin ZY and Wang LY, Development of a faster determination of 10 anabolic steroids residues in animal muscle tissues by liquid chromatography tandem mass spectrometry J Pharmaceut Biomed Anal 41:616–621 (2006) 20 Bishnu A, Chakrabarti K, Chakraborty A and Saha T, Pesticide residue level in tea ecosystems of Hill and Dooars regions of West Bengal, India Environ Monit Assess 149:457–464 (2009) 21 Liu W, Qin S and Gan J, Chiral stability of synthetic pyrethroid insecticides J Agric Food Chem 53:3814–3820 (2005) 22 Kuang H, Hou YX, Chu XG and Xu CL, Simultaneous determination of phenoxyacid herbicides in soybean by gas chromatography–mass spectrometry Chin J Anal Chem 34:1733–1736 (2006) www.soci.org 23 Wang W, Li XH, Wang XF, Wang XZ, Lu H, Jiang XN, et al., Levels and chiral signatures of organochlorine pesticides in urban soils of Yinchuan, China Bull Environ Contam Toxicol 82:505–509 (2009) DOI: 10.1007/s00128-009-9632-9 24 Wang Q, Qiu J, Zhu W, Jia G, Li J, Bi C, et al., Stereoselective degradation kinetics of theta-cypermethrin in rats Environ Sci Technol 40:721–726 (2006) 25 Hu B, Song W, Xie L and Shao T, Determination of 33 pesticides in tea by accelerated solvent extraction–gel permeation and solid-phase extraction purification–gas chromatography–mass spectrometry Se Pu 26:22–28 (2008) 26 Liu W and Gan J, Separation and analysis of diastereomers and enantiomers of cypermethrin and cyfluthrin by gas chromatography J Agric Food Chem 52:755–761 (2004) 27 Harner T, Wiberg K and Norstrom R, Enantiomer fractions are preferred to enantiomer ratios for describing chiral signatures in environmental analysis Environ Sci Technol 34:218–220 (2000) 28 Li F and Wang J, Discrimination of teas with different degrees of fermentation by SPME–GC analysis of the characteristic volatile flavour compounds Food Chem 109:196–206 (2008) 29 Bidleman TF and Falconer RL, Using enantiomers to trace pesticide emissions Environ Sci Technol 33:206A–209A (1999) 30 Roberts EA and Wood DJ, The fermentation process in tea manufacture; oxidation of substrates by tea oxidase Biochem J 47:175–186 (1950) 1379 J Sci Food Agric 2010; 90: 1374–1379 c 2010 Society of Chemical Industry www.interscience.wiley.com/jsfa Short Communication Received: 10 July 2009 Revised: 20 September 2009 Accepted: 21 September 2009 Published online in Wiley Interscience: 18 March 2010 (www.interscience.wiley.com) DOI 10.1002/jsfa.3931 Differentiation of bacteria using fatty acid profiles from gas chromatography–tandem mass spectrometry Yanyan Li, Shengfang Wu, Liping Wang, Ye Li, Feng Shi and Xiaoyuan Wang∗ Abstract BACKGROUND: Fatty acids are essential components in cell membranes of bacteria, and they determine the different structures of lipids and lipid A Therefore, fatty acids are good targets for development of rapid and reliable methods for differentiation of bacteria RESULTS: Here we report fatty acid distributions in three different bacteria – Escherichia coli, Francisella novicida and Bacillus subtilis – studied by using gas chromatography–tandem mass spectrometry Different fatty acid profiles were observed in the three bacteria Hydroxyl fatty acids were observed in E coli and F novicida, but not in B subtilis Carbocyclic fatty acids were observed in E coli but not in F novicida or B subtilis Methyl-branched fatty acids were observed in B subtilis but not in E coli or F novicida Although saturated and unsaturated fatty acids were all observed in the three bacteria, their chain lengths and distribution patterns were different CONCLUSION: The results indicate that the fatty acid profile in a given bacterium is highly specific, and gas chromatography–tandem mass spectrometry is a rapid and sensitive method for the identification or detection of bacteria c 2010 Society of Chemical Industry Keywords: bacteria; lipids; lipid A; fatty acid profile; GC/MS/MS 1380 INTRODUCTION EXPERIMENTAL In bacteria, fatty acids are synthesized in the type II fatty acid biosynthesis pathway by a set of enzymes.1 Additional enzymes could further modify the structure of fatty acids in response to environmental changes.2 In bacterial membranes fatty acids consist of the hydrophobic parts of glycerophospholipids or lipid A Glycerophospholipids exist in both Gram-positive and Gram-negative bacteria, while lipid A, the hydrophobic anchor of lipopolysaccharide, only exists in Gram-negative bacteria.3 The lipid A molecule usually contains several 3-OH fatty acid chains which can be used as a marker for the detection of Gram-negative bacteria The difference among fatty acids is focused on their hydrophobic hydrocarbon chains, which could be saturated, unsaturated, branched, hydroxyl or cyclopropanal Here we report the fatty acid profiles in three different bacteria – Escherichia coli W3110, Bacillus subtilis 168 and Francisella tularensis subsp novicida U112 – studied using gas chromatography–tandem mass spectrometry (GC/MS/MS) E coli and F novicida are typical Gram-negative bacteria, while B subtilis is a typical Gram-positive bacterium We found that the fatty acid profiles are different in the three bacteria, and GC/MS/MS can provide a rapid and sensitive method for the identification or detection of bacteria Bacteria strains and growth conditions Francisella novicida U112 was grown at 37 ◦ C in TSB-C broth (30 g kg−1 trypticase soy broth supplemented with g kg−1 cysteine) Escherichia coli W3110 and Bacillus subtilis 168 were grown at 37 ◦ C in LB broth (10 g kg−1 tryptone, g kg−1 yeast extract and 10 g kg−1 NaCl) Lipid extraction from bacteria and their analysis by thin-layer chromatography (TLC) In order to recover lipid A together with phospholipids, the bacteria were first treated with 100 mL of 12.5 mmol L−1 sodium acetate, pH 4.5, at 100 ◦ C for 30 The reaction mixture was centrifuged, and the pellets were resuspended and incubated in 114 mL singlephase Bligh–Dyer mixture,4 consisting of chloroform, methanol and water (1 : 2:0.8, v/v/v) for 60 at room temperature Next, the insoluble debris was removed by centrifugation, and the supernatant containing the phospholipids and lipid A was converted to a two-phase Bligh–Dyer mixture by adding 30 mL J Sci Food Agric 2010; 90: 1380–1383 ∗ Correspondence to: Xiaoyuan Wang, State Key Laboratory of Food Science and Technology, Jiangnan University, 1800 Lihu Avenue, Wuxi 214122, China E-mail: xiaoyuanwang@hotmail.com State Key Laboratory of Food Science and Technology, Jiangnan University, Wuxi 214122, China www.soci.org c 2010 Society of Chemical Industry Differentiation of bacteria using fatty acid profiles www.soci.org chloroform and 30 mL water.4 The two phases were separated by low-speed centrifugation, and the lower phase was collected and dried under a stream of nitrogen The extracted lipids were dissolved in mL chloroform and methanol (2 : 1,v/v) and 10 µL was spotted onto a silica gel 60 TLC plate The plate was developed in a solvent of chloroform, methanol, water and ammonia (65 : 25:3.6 : 0.4, v/v/v/v) and dried The lipids on the plate were visualized by spraying with 100 g kg−1 sulfuric acid in ethanol, followed by charring at 200 ◦ C Extraction of fatty acids from whole bacterial cells Bacteria were resuspended in mL of mol L−1 NaOH, and incubated at 90 ◦ C for 90 After cooling, the pH of the sample was adjusted to with HCl Fatty acids were then extracted by adding mL anhydrous diethyl ether and separated by centrifugation at 5500 × g for 10 The upper phase was removed and dehydrated by adding anhydrous sodium sulfate The dehydrated fatty acids were collected and dried under a stream of nitrogen Next, 50 µL bistrimethylsilyl trifluoroacetamide (BSTFA) was added, and the mixture was incubated at 70 ◦ C for 30 and dried under a stream of nitrogen The sample was dissolved in 100 µL hexane for GC/MS or GC/MS/MS analysis GC/MS and GC/MS/MS analysis Fatty acid composition analysis was performed on a gas chromatograph/tandem mass spectrometer (1200L-GC/MS/MS, Varian, Palo Alto, CA, USA) A 30 m ì 0.25 mm ì 0.25 àm silica capillary column (Varian VF-5 MS) was used The temperature of the column was programmed as follows: it was first held at 120 ◦ C for 0.5 min, then increased to 220 ◦ C at a rate of 10 ◦ C min−1 , to 240 ◦ C at a rate of ◦ C min−1 , to 280 ◦ C at a rate of 12 ◦ C min−1 , and held at 280 ◦ C for Helium was used as the carrier gas at a constant flow rate of 0.8 mL min−1 The mass spectrometer was operated in the electron impact mode at 70 eV with a scan range of 50–500 m/z For GC/MS/MS analysis, collision energy was eV and collision cell pressure was 1.75 torr Data acquisition and analysis were performed using the Varian workstation software RESULTS Composition of lipids differs in E coli, B subtilis and F novicida The lipids extracted from bacteria E coli W3110, B subtilis 168 and F tularensis U112 showed different distribution patterns on TLC (Fig 1) Phospholipids were observed in all three samples, but lipid A was only observed in samples from E coli and F novicida (Fig 1) In addition, F novicida contains more lipid species than the other two bacteria, consistent with our previous oberservation.5 The results indicate that these three bacteria can be differentiated by analyzing profiles of their phospholipids and lipid A Since fatty acids are the major components in phospholipids and lipid A, we further studied the fatty acid profiles in these three bacteria J Sci Food Agric 2010; 90: 1380–1383 Table Comparison of fatty acid profiles in three different bacteria The data were derived from TIC shown in Fig Retention time (min) 5.6 7.4 9.1 10.5 10.9 11.0 11.1 12.1 13.0 13.1 13.3 14.1 14.2 14.9 15.1 16.4 16.7 18.3 18.7 Percentage (%) Cellular fatty acids E coli B subtilis F tularensis 10 : 12 : 14 : 15:0M 16 : 14:0OH 16 : 17:0 18 : 16:0OH 18 : 19:0M 19:0 18:0OH 20 : 22 : 22 : 24 : 24 : 0.18 8.61 12.50 ND 5.94 7.22 34.88 17.16 6.56 ND 5.23 ND 1.71 ND ND ND ND ND ND ND ND 1.96 1.70 ND ND 42.27 ND 30.96 ND 8.26 14.85 ND ND ND ND ND ND ND 7.84 0.28 8.13 ND ND ND 17.99 ND 5.59 3.49 9.80 ND ND 12.68 3.04 1.54 9.25 7.44 12.93 ND, not detectable; 15:0M , methyl-myristic acid; 19:0M , methyloctadecanoic acid; 14:0OH , 3-hydroxy-tetradecanoic acid; 16:0OH , 3-hydroxy-hexadecanoic acid; 18:0OH , 3-hydroxy-octadecanoic acid; 17:0 , cyclohexylundecanoic acid acids C15:0M and C19:0M , respectively In F novicida, 13 fatty acid species were observed (Fig 2(C)) Peaks at 13.0 and 18.3 could be derived from unsaturated fatty acids C18 : and C24 : 1, respectively Peaks at 13.1 and 14.9 could be derived from hydroxyl fatty acids C16:0OH and C18:0OH , respectively Hydroxyl fatty acids were only observed in E coli and F novicida, but not in B subtilis Carbocyclic fatty acids were only observed in E coli and not in F novicida or B subtilis Methyl-branched fatty acids C15:0M and C19:0M comprised 1.70% and 14.85% of the total c 2010 Society of Chemical Industry www.interscience.wiley.com/jsfa 1381 Fatty acid profiles differ in E coli, B subtilis and F novicida The fatty lipids extracted from the three bacteria were analyzed by total ion chromatography (TIC) The overall profiles of fatty acids from the three bacteria are clearly different (Table 1) In E.coli, about nine fatty acid species were observed (Fig 2(A)) Peaks at 11.0, 12.1 and 14.2 could be derived from hydroxyl fatty acids C14:0OH , cyclopropanal fatty acids C17:0 and C19:0 , respectively In B subtilis, six fatty acid species were observed (Fig 2(B)) Peaks at 10.5 and 14.1 could be derived from methyl-branched fatty Figure TLC of lipids extracted from three different bacteria: E coli, B subtilis and F novicida www.soci.org Y Li et al Figure TIC of fatty acids extracted from bacteria: (A) E coli; (B) B subtilis; (C) F novicida fatty acids in B subtilis, respectively, but were not observed in E coli or F tularensis Although saturated and unsaturated fatty acids were observed in all three bacteria, their chain length and distribution patterns were different For example, hexadecanoic acid C16 : comprised 34.88% of total fatty acids in E coli, 42.27% in B subtilis and 17.99% in F tularensis; dodecanoic acid C12 : comprised 8.61% of total fatty acids in E coli, 0.28% in F tularensis, but were not observed at all in B subtilis Long-chain fatty acids (C20 : 0, C22 : 0, C22 : 1, C24 : and C24 : 1) were only observed in F tularensis and made up 34.2% of the total fatty acids These data clearly demonstrated that profiles of fatty acids differ in the three bacteria 1382 Confirmation of 3-hydroxyl fatty acids in Gram-negative bacteria Hydroxyl fatty acids were observed in the TIC spectra of fatty acids extracted from Gram-negative bacteria E coli and F novicida (Fig 2) Molecular weights of the derivatives of hydroxyl fatty acids of C14:0OH in E coli, and of C16:0OH and C18:0OH in F novicida, were 388, 416 and 444, respectively The GC–tandem mass spectra of these molecules are shown in Fig Similar patterns of fragment ions were observed in all three spectra Fragment ions at m/z 373.3 in Fig 3(A), at m/z 401.1 in Fig 3(B) and at m/z 429.1 in Fig 3(C) could be derived from their parent molecules by losing a methyl group The further cleavage of the CO2 (CH3 )3 Si+ group could generate fragment ions at m/z 257.3 in Fig 3(A), at m/z 285.1 in Fig 3(B) and at m/z 313.2 in Fig 3(C) Ions at m/z 73, 147 and 233 observed in all three spectra have ion masses expected for (CH3 )3 Si+ , C3 H3 O2 (CH3 )3 Si+ and C3 H3 O2 (CH3 )6 Si2 + , respectively The presence of the ion at m/z 233 suggests that the hydroxyl www.interscience.wiley.com/jsfa Figure GC/MS/MS analysis of 3-hydroxy fatty acids from bacteria: (A) E coli W3110 (14:0OH ): (B) F tularensis U112 (16:0OH ) and (C) (18:0OH ); (D) assignment for fragment ions of 3-hydroxy trimethysilyl fatty acid group in these hydroxyl fatty acids connects to the third carbon (Fig 3(D)) DISCUSSION AND CONCLUSIONS In bacteria there are dozens of genes encoding enzymes synthesizing fatty acids, phospholipids and lipid A.1 – Some of these genes are conserved but others are not For example, FabH in E coli uses straight-chain fatty acids as substrates, but in Grampositive bacteria it prefers larger branched-chain fatty acids as substrates,6,7 consistent with our data showing that branchedchain fatty acids only exist in B subtilis and not in E coli or F tularensis Therefore, the branched-chain fatty acids can be used as a marker for the differentiation of Gram-negative and Gram-positive bacteria Lipid A, known as endotoxin, can cause severe diseases such as sepsis Detection of lipid A is very important in food safety and human health Limulus amebocyte lysate (LAL) assay has been widely used to detect lipid A, but it is not entirely specific because β-1,3-glucans have been found to react with LAL through a separate glucan-sensitive enzymatic pathway.8 Although the detailed structure of lipid A changes from one bacterium to another, all the lipid A examples known so far contain 3-hydroxyl c 2010 Society of Chemical Industry J Sci Food Agric 2010; 90: 1380–1383 Differentiation of bacteria using fatty acid profiles www.soci.org fatty acid Since 3-hydroxyl fatty acid is not synthesized in organisms other than Gram-negative bacteria, it can be used for the identification of Gram-negative bacteria and the detection of endotoxins in foods and in environmental and pharmaceutical products The detection of pathogenic bacteria is very important to prevent and identify problems related to health and food safety Traditional and standard bacterial detection methods usually take up to a few days.9 These methods depend on the specificity of their target molecule and sometimes cause false positive results Our data have demonstrated that GC/MS/MS is a sensitive method to analyze the fatty acids in bacteria; therefore it could be used for the detection or identification of bacteria ACKNOWLEDGEMENTS Funding was provided by grants from the NSFC (30770114 and 30870074), the SKLF (SKLF-MB-200801) and the 111 project (1112-06) REFERENCES Grogan DW and Cronan JE Jr, Cyclopropane ring formation in membrane lipids of bacteria Microbiol Mol Biol Rev 61:429–441 (1997) Raetz CR, Reynolds CM, Trent MS and Bishop RE, Lipid A modification system in Gram-negative bacteria Annu Rev Biochem 76:295–330 (2007) Bligh EG and Dyer WJ, A rapid method of total lipid extraction and purification Can J Biochem Physiol 37:911–917 (1959) Wang X, Ribeiro AA, Guan Z, McGrath SC, Cotter RJ and Raetz CRH, Structure and biosynthesis of free lipid A molecules that replace lipopolysaccharide in Francisella tularensis subsp novicida Biochemistry 45:14427–14440 (2006) Choi KH, Heath RJ and Rock CO, Beta-ketoacyl-acyl carrier protein synthase III (FabH) is a determining factor in branched-chain fatty acid biosynthesis J Bacteriol 182:365–370 (2000) Qiu X, Choudhry AE, Janson CA, Grooms M, Daines RA, Lonsdale JT, et al., Crystal structure and substrate specificity of the beta-ketoacylacyl carrier protein synthase III (FabH) from Staphylococcus aureus Protein Sci 14:2087–2094 (2005) Takaki Y, Seki N, Kawabata S, Iwanaga S and Muta T, Duplicated binding sites for (1–3)-beta-D-glucan in the horseshoe crab coagulation factor G: implications for a molecular basis of the pattern recognition in innate immunity J Biol Chem 277:14281–14287 (2002) ˜ FX, Pathogen detection: a Lazcka O, Del Campo FJ and Munoz perspective of traditional methods and biosensors Biosens Bioelectron 22:1205–1217 (2007) White SW, Zheng J, Zhang YM and Rock CO, The structural biology of type II fatty acid biosynthesis Annu Rev Biochem 74:791–832 (2005) 1383 J Sci Food Agric 2010; 90: 1380–1383 c 2010 Society of Chemical Industry www.interscience.wiley.com/jsfa ... Improvement of lactic acid production during rosy vinegar fermentation J Sci Food Agric 90:1334–1339 (2010) Li Q, Chen Q, Ruan H, Zhu D and He G, Isolation and characterization of a novel oxygen, acid. .. consisted of a cross-head speed of mm s−1 and a 25 kg load cell.22 The shear force perpendicular to the axis of muscle fibres was measured in 10 replicates for each treatment The peak of the shear force... Cleveland, OH, pp 617–780 (1972) 18 AOAC, Official Methods of Analysis of the Association of Official Analytical Chemists (16th edn) Association of Official Analytical Chemists, Gaithersburg,