Xin-Xian Li 1 *, Kimiaki Fukuhara 1 and Yasuyoshi Hayata 2. Introduction

Transcription

1 J. Japan. Soc. Hort. Sci. 78 (2): Available online at JSHS 2009 Concentrations of Character Impact Odorants in Toyonoka Strawberries Quantified by Standard Addition Method and PQ Column Extraction with GC-MS Analysis Xin-Xian Li 1 *, Kimiaki Fukuhara 1 and Yasuyoshi Hayata 2 1 Fruit Processing Research Center, AOHATA Corporation, Takehara , Japan 2 Department of Agriculture, Meiji University, Kawasaki , Japan In the present study, the standard addition method with Porapak Q column (PQ) extraction and gas chromatography-mass spectrometry analysis was adapted to quantitatively determine the character impact odorants of Toyonoka strawberries. A significant correlation was confirmed between the contents that determine the character impact odorants in the PQ extract and the levels of standards added to the fruit samples; therefore, standard addition analysis might be a useful and reliable method to analyze already-known odorants in fruits. The concentrations of the character impact odorants, especially 2,5-dimethyl-4-hydroxy-2H-furan-3-one, hexanoic acid, γ-dodecalactone, and the esters, were detected at relatively high levels, which could explain why Toyonoka strawberries have abundant aromas. Key Words: Porapak Q column extraction, odorant, standard addition method, strawberry, volatile analysis. Introduction Received; August 11, Accepted; September 16, * Corresponding author ( xinxian_li@aohata.co.jp). The strawberry (Fragaria ananassa Duch.) is cultivated in nearly all countries of the world as an important economic crop. These strawberry fruits are not consumed as fresh fruits but are used in processed foods such as juices, jams, jellies, syrups, liquors, ice creams, and fruit concentrates. F. ananassa Duch. was bred in Europe by crossing F. virginiana Mill. and F. chiloensis (L.) Mill. in the 18th century, and was introduced into Japan in the middle of the 19th century. Since the first strawberry cultivar, Fukuba, was published in 1899 (Oda and Kimura, 2004), a large number of cultivars have been bred, of which Toyonoka, with its well-balanced taste and abundant aroma, has been the most popular cultivar in commercial culture over the last 30 years. Consumers enjoy strawberries immensely for their well-balanced mixture of sweetness and tartness and especially for their unique aroma. Since the first result was reported in 1939 by Coppens and Hoejenbos, many studies on the volatiles of strawberries have been carried out, accompanied by the development of volatile analysis technologies, particularly using gas chromatographymass spectrometry (GC-MS). More than 360 volatile compounds have been identified in strawberries (Latrasse, 1991) by diversified extraction and analysis methods (Gomes da Silva and Chaves das Neves, 1999; Hakala et al., 2002; Park et al., 2000; Pyysalo et al., 1979; Schieberle, 1994; Urruty et al., 2002). Regarding the quantified study of volatiles in strawberries, although high performance liquid chromatography with either ultraviolet radiation (Mura and Zabetakis, 2002; Sanz et al., 1994, 1995) or photo diode array (Zabetakis and Holden, 1996) detection has been used recently, GC analysis is the most frequent method for the separation of volatile compounds (Forney et al., 2000; Zabetakis and Holden, 1997). In analyses with a GC-flame ionization detector or GC-MS, the content of volatile compounds has been generally estimated by the relative concentration of one or several internal standards added during extraction or of external standards added before analysis (Dirinck et al., 1981; Douillard and Guichard, 1990; Fischer and Hammerschmidt, 1992; Hirvi, 1983; Kollmannshberger and Berger, 1994; Park et al., 2000; Perez et al., 1996; Pickenhagen et al., 1981; Pyysalo et al., 1979; Schreier, 1980; Siegmund et al., 2001; Ulrich et al., 1997). It is very difficult to reliably determine the content of volatiles in foods because of differences in the chemical and physical properties of volatiles and due to losses that occur during extraction with some 200

2 J. Japan. Soc. Hort. Sci. 78 (2): treatments, such as heating and/or reducing pressure. Schieberle s group adapted a stable isotope dilution assay (Schieberle, 1994; Sen et al., 1991) to volatile analysis, which is reliable and accurate; however, this method is very limited due to the need for many stable isotopes. In our recent study, we used Porapak Q column (PQ) extraction to extract volatile compounds from Toyonoka strawberries, and evaluated the contribution of volatiles using GC-olfactometry (GC-O) and aroma extract dilution analysis (AEDA) (Fukuhara et al., 2005). Fifty-two odorants, including 4 unknown compounds, were detected from the PQ extract, of which 14 odorants, including one unknown compound, were identified with high flavor dilution (FD) factors of 81 or higher. In the present study, we used the standard addition method and PQ extraction with GC-MS analysis to quantify character impact odorants in Toyonoka strawberries. Materials and Methods Chemicals The authentic aroma standards were obtained as follows: 2,5-dimethyl-4-hydroxy-2H-furan-3-one (DMHF), ethyl 2-methylbutanoate, trans,cis-2,6- nonadienal, trans-3-hexenol, trans-nerolidol, vanillin, γ- decalactone, γ-dodecalactone, δ-decalactone, and δ- dodecalactone (Sigma-Aldrich Inc., St. Louis, MO, USA); butanoic acid, ethyl 3-methylbutanoate, and ethyl hexanoate (Tokyo Kasei Kogyo Co., Ltd., Tokyo, Japan); 2-methylbutanoic acid, hexanoic acid, linalool, and octanol (Wako Pure Chemical Industries Co., Ltd., Osaka, Japan). The solvents diethyl ether and methanol were purchased from Wako. Plant materials Strawberries (F. ananassa Duch. Toyonoka ) were grown in non-heated greenhouses. The fruits were harvested at days after anthesis, the commercial maturity stage (more than 75% red color on the fruit surface), frozen immediately with liquid nitrogen, and stored at 80 C until analysis. Standard solution Seventeen compounds were selected from character impact odorants of Toyonoka strawberries, based on the results of our previous study (Fukuhara et al., 2005). A standard solution of compounds was prepared with diethyl ether as a solvent at 4 C according to the concentrations described in Table 1. Sample preparation and addition of standards Strawberry fruits were thawed and homogenized at 4 C and rpm for 2 min using a NISSEI AM homogenizer (Nihonseiki Ltd., Tokyo). The standard solution was added to 25 g of homogenate. According to the results of a preliminary experiment, four addition levels were determined (Table 1); i.e., the volumes of standard solution added were 0, 10, 30, and 50 μl. The Table 1. Standard solution of character impact odorants of the Toyonoka strawberry. z y Compound Purity of reagent (%) Concentration z (mg L 1 ) Ethyl 2-methylbutanoate Ethyl 3-methylbutanoate Ethyl hexanoate trans-3-hexenol Linalool Octanol trans,cis-2,6-nonadienal Butanoic acid Methylbutanoic acid Hexanoic acid DMHF y trans-nerolidol γ-decalactone δ-decalactone γ-dodecalactone δ-dodecalactone Vanillin Level of standard solution added to a 25 g sample: 0, 10, 30, or 50 μl. 2,5-Dimethyl-4-hydroxy-2H-furan-3-one. homogenate, including the standards, was mixed with 50 ml of deionized water and centrifuged at g for 20 min at 4 C. The residue was again extracted with 25 ml of deionized water. Supernatants were pooled and used for the isolation of volatiles. PQ column preparation Ten milliliters of Porapak Q (polydivinylbenzene, mesh, Waters Co., Ltd., Milford, MA, USA) was packed into a glass column (2 10 cm). For the first use or regeneration, the column was successively washed with 50 ml methanol, 50 ml diethyl ether, 50 ml methanol, and 50 ml deionized water. Isolation of volatiles The process described by Hayata et al. (2003) was followed for the isolation of volatiles. The above supernatant was passed through the PQ column. After the column was rinsed with 50 ml of deionized water, the adsorbed compounds were eluted with 50 ml diethyl ether. The eluate was dried over anhydrous sodium sulfate for at least 3 hr, and the solvent was evaporated to 50 μl under a nitrogen stream at 37 C. Gas chromatography-mass spectrometry (GC-MS) Shimadzu GCMS-QP2010 GC-mass spectrometry (Shimadzu Co., Ltd., Kyoto, Japan) equipped with a DB- Wax column (60 m 0.25 mm i.d., column film thickness of 0.25 μm; J&W Scientific Inc., Folsom, CA, USA) was used to determine the volatile compounds in the extract. The injector, interface, and ion source were maintained at 230, 230, and 200 C, respectively. The

3 202 X-X. Li, K. Fukuhara and Y. Hayata oven temperature program was as follows: maintained at 40 C for 10 min, raised to 220 C at 3 C min 1 and maintained at 220 C for 15 min, and raised to 245 C at 20 C min 1 and maintained at 245 C for 20 min. The flow rate of helium carrier gas was 30 cm s 1. A split mode injector at 1 : 60 with an injection volume of 1 μl was used. Mass spectra were obtained by an electron impact ionization source at 70 ev and scanned from m/z 29 to 350. The data gathering and the identification and quantitative calculation of odorants were performed using GCMS Solution 2.30 (Shimadzu) software. The target ion and reference ion used in the identification and quantitative calculation of each compound are indicated in Table 2. Data analysis Four replications of extraction and analysis for each added level of standards were performed. The mean areas of the target ion of each standard and of methyl hexanoate, trans-2-hexenal, and methyl nicotinate (compounds used as base peak, Table 2) were calculated from 4 replications of each. The ratios of the target ion area of every standard divided by the sum of the target ion area of the three base peak compounds were calculated. The slope, intercept, and correlation coefficient between the concentration of the added level of the standard and the ratio corresponding to the target ion areas were derived by matching a regression line. The statistical significance of linear correlations was tested. The concentration of every character impact odorant was calculated from the slope and the intercept. The coefficient of variation (CV) of the target ion area was indicated by the average CVs of the four levels of standard added, except for the base peak compounds whose CVs were directly calculated from 16 analyzed values. Results and Discussion A significant correlation between the levels of standards added and the results for the determined character impact odorants in the PQ extract from Toyonoka strawberries was shown, because the correlation coefficients of all odorants between the added levels of standards and the ratios of the target ion area of every standard divided by the sum of the target ion areas of the three base peak compounds were >0.998 (P = 0.001), and half were >0.999 (P = ) (Tables 1 and 3). In the present study, as all CVs of the target ion area were <19% (Table 3), higher reproducibility was attained as compared with our previous results using the melon (Hayata et al., 2002); therefore, that the determined values are reliable. Of course, some compounds could not be added; for example, acetic acid could not be selected as an addition standard, although it is an important odorant for Toyonoka with a relatively high FD factor (FD = 27, Fukuhara et al., 2005), for the Table 2. Determination condition of character impact odorants of the Toyonoka strawberry using GCMS with PQ extraction and the standard addition method. z y x Compound Retention time (min) Kovats index on DB-Wax (60 m). 2,5-Dimethyl-4-hydroxy-2H-furan-3-one. Compound used as base peak for determination. Kovats index z Target ion (m/z) Referrence ion 1 st (m/z) Referrence ion 2 nd (m/z) Ethyl 2-methylbutanoate Ethyl 3-methylbutanoate Ethyl hexanoate trans-3-hexenol Linalool Octanol trans,cis-2,6-nonadienal Butanoic acid Methylbutanoic acid Hexanoic acid DMHF y trans-nerolidol γ-decalactone δ-decalactone γ-dodecalactone δ-dodecalactone Vanillin Methyl hexanoate x trans-2-hexenal x Methyl nicotinate x

4 J. Japan. Soc. Hort. Sci. 78 (2): Table 3. Content of character impact odorants of the Toyonoka strawberry using GCMS with PQ extraction and the standard addition method. Compound Concentration (μg kg 1 ) Coefficient of variation of target ion area (%) Correlation coefficient (r) z 2,5-Dimethyl-4-hydroxy-2H-furan-3-one. y Compound used as base peak for determination. x From the Flavor-Base 2004 professional database (Leffingwell & Associates, Canton, GA, USA; Odor threshold in water). w From Moyano et al., Odor threshold (ppb) Ethyl 2-methylbutanoate x Ethyl 3-methylbutanoate x Ethyl hexanoate x trans-3-hexenol w Linalool x Octanol x trans,cis-2,6-nonadienal x Butanoic acid x 2-Methylbutanoic acid x Hexanoic acid x DMHF z x trans-nerolidol x γ-decalactone x δ-decalactone x γ-dodecalactone x δ-dodecalactone x Vanillin x Methyl hexanoate y 5.0 trans-2-hexenal y 5.7 Methyl nicotinate y 3.9 following reason. In a preliminary experiment, we could not acquire a regression line between the added level and the target ion (60 m/z) area of acetic acid (data not shown). It could be presumed that acetic acid possesses high reactivity, although sample preparation, standard addition, and PQ extraction were performed at 4 C. Determining the quantity of standards to be added is also very important. In the present study, the quantity of added standards was 1 8-fold of the contents in the samples due to the addition of 10 μl of standard solution to a 25 g fruit sample. These added levels were confirmed as appropriate concentrations because an extremely close correlation was obtained between the added levels and the target ion areas (Table 3). The concentration of DMHF was found to be 19.7 mg kg 1 FW, the highest in all character impact compounds (Table 3). The contents of DMHF in quantitative experiments are generally lower than 20 mg kg 1 FW, for example, mg L 1 in Red Gaunlet and a wild strawberry in Switzerland (Pickenhagen et al., 1981), 11.7 mg kg 1 FW in Senga Sengana (Larsen and Poll, 1992), and mg kg 1 FW in Dania, Bogota, and Senga (Larsen et al., 1992), cultured in Denmark, 16.2 mg kg 1 FW in Italy cultured in Germany (Schieberle, 1994), 1.1~ 5.4 mg kg 1 FW in Korona, Pandora, Bogota, Pegasus, Senbana, and Elsante cultured in Germany (Sen et al., 1991), <16 mg kg 1 FW in Redcrest, Chandeler, Pajaro, Benton, and Parker cultured in the USA (Sanz et al., 1995), mg kg 1 FW in Cuesta, Seascape, Sunset, Laguna, Carlsbad, Cartcua, Cartuno, and Camarosa cultured in Spain (Perez et al., 1996), about 10 mg kg 1 FW in Elsanta cultured in the UK (Mura and Zabetakis, 2002), and 6.8 mg kg 1 FW in Cigaline cultured in France (Menager et al., 2004). The DMHF concentration was found to be >20 mg kg 1 FW in a hybrid between a wild strawberry and Chandler in Germany (Sen et al., 1991), Douglas and Totem cultured in the USA (Sanz et al., 1995), Elsanta, Pandora Sengana, and Elvira cultured in Denmark (Larsen et al., 1992) and Chandler, Tudla, I-101, and Oso Grande cultured in Spain (Perez et al., 1996); therefore, the level of DMHF in Toyonoka is relatively high for strawberry cultivars. In addition to its highest FD factor (Fukuhara et al., 2005), DMHF is obviously the most important odorant for Toyonoka. On the other hand, DMHF was only detected at <0.01 mg kg 1 FW in Senga Sengana with simultaneous distillation extraction (Fischer and Hammerschmidt, 1992); therefore, the extraction method has a great influence on the quantization of DMHF. PQ extraction should be recommended for determining DMHF. Moreover, obvious differences occur in the DMHF concentration present in the same cultivar when it is cultured in different regions or countries, for example, the Bogota, Elsanta, and Chandeler

5 204 X-X. Li, K. Fukuhara and Y. Hayata cultivars mentioned above. The content of hexanoic acid, detected at a concentration of 12.5 mg kg 1 FW in Toyonoka, was so high that the level might be recognized as the highest reported so far for strawberry fruits (Table 3) except in the cultivar, Elsanta, cultured in Denmark (Larsen et al., 1992). Hexanoic acid plays an important role in the aroma of Toyonoka, especially in strawberries with a sweaty or sour odor, contributing to the mature aroma, accompanied by 2-methylbutanoic acid and butanoic acid, and also including trans-nerolidol at a relatively high concentration. In the retention time of transnerolidol, a sour odor was perceived by GC-O from the PQ extract of Toyonoka fruits (Fukuhara et al., 2005), but the same odor was not perceived in its authentic standard compound. Additionally, many isomers of nerolidol occur in plant organs; therefore, further study is required to draw a conclusion. γ-dodecalactone was found at a concentration of 0.7 mg kg 1 FW (Table 3), and might be the highest level reported for strawberry fruits. In our previous study (Fukuhara et al., 2005), we found that lactones, including γ-dodecalacone, γ-decalactone, δ-dodecalacone, and δ- decalactone, with a sweet, peach-like or coconut-like odor, play an important role in the aroma of Toyonoka, with the FD factor of γ-dodecalactone being the highest of all odorants except DMHF; however, lactones were not so important in Italy strawberries (Schieberle, 1994). The concentration of γ-dodecalactone was about 10 times that of γ-decalactone in Toyonoka, which is consistent with the concentration found in two Korea cultivars (Park et al., 2000); however, an inverse tendency was shown in European strawberry cultivars (Fischer and Hammerschmidt, 1992; Hirvi, 1982; Kollmannsberger, 1994; Menager, 2004; Schreier, 1980). These results might suggest a different formation of these two lactones between Asian and European strawberry cultivars. The concentration of linalool was determined to be relatively high in Toyonoka fruits. The concentration of trans,cis-2,6-nonadienal was not as high, but its threshold was extremely low (Table 3). On the other hand, the concentrations of trans-3-hexenol and octanol were determined to be low. The odor sniffed in the retention time of trans-3-hexenol and octanol in the extract from Toyonoka fruits was similar to that of their standards; however, the FD factor (Fukuhara et al., 2005) in a previous study was incongruous with the concentration determined according to its threshold. Further investigation is therefore necessary. The esters, including ethyl 2-methylbutanoate, ethyl 3-methylbutanoate, and ethyl hexanoate, which are important compounds due to their contribution to the fruity odor of Toyonka fruits, were detected at a relatively high level. In conclusion, the character impact odorants of Toyonoka strawberries were reliably determined using the standard addition method with PQ extract and GC- MS analysis in the present study; therefore, standard addition analysis, combined with PQ extraction, might be a useful method to analyze already-known odorants in fruits. Moreover, the fact that the contents of the character impact odorants, especially DMHF, hexanoic acid, γ-dodecalactone, and the esters, were detected at relatively high levels could explain why Toyonoka strawberries have abundant aromas, in accordance with the meaning of its Japanese name. Literature Cited Coppens, A. and L. Hoejenbos Investigation of the volatile constituents of strawberry juice (Fragaria eliator Ehrh). Rec. Trav. Chim. Pays-Bas. 58: 680. Dirinck, P., L. Schreyen and N. Schamp Aroma quality evaluation of tomatoes, apples, and strawberries. J. Agric. Food Chem. 25: Douillard, C. and E. Guichard The aroma of strawberry (Fragaria ananassa): characterization of some varieties and influence of freezing. J. Sci. Food Agric. 50: Fischer, N. and F. J. Hammerschmidt A contribution to the analysis of fresh strawberry flavour. Chem. Mikrobiol. Technol. Lebensm. 14: Forney, C. F., W. Kalt and M. A. Jordan The composition of strawberry aroma is influenced by cultivar, maturity, and storage. HortScience 35: Fukuhara, K., X. X. Li, M. Okamura, K. Nakahara and Y. Hayata Evaluation of odorants contributing to Toyonoka strawberry aroma in extracts using an adsorptive column and aroma dilution analysis. J. Japan. Soc. Hort. Sci. 74: Gomes da Silva, M. D. R. and H. J. Chaves das Neves Complementary use of hyphenated purge-and-trap gas chromatography techniques and sensory analysis in the aroma profiling of strawberries (Fragaria ananassa). J. Agric. Food Chem. 47: Hakala, M., A. Lapvetelainen and H. Kallio Volatile compounds of selected strawberry varieties analyzed by purge-and-trap headspace GC-MS. J. Agric. Food Chem. 50: Hayata, Y., T. Sakamoto, C. Maneerat, X. X. Li, H. Kozuka and K. Sakamoto Evaluation of aroma compounds contributing to muskmelon flavor in Porapak Q extracts by aroma extract dilution analysis. J. Agric. Food Chem. 51: Hayata, Y., T. Sakamoto, H. Kouzuka, K. Sakamoto and Y. Osajima Analysis of aroma volatile compounds in Miyabi melon (Cucumis melo L.) using the Porapak Q column concentration method. J. Japan. Soc. Hort. Sci. 71: Hirvi, T Mass fragmentographic and sensory analyses in the evaluation of the aroma of some strawberry. Lebensm. Wiss. Technol. 16: Hirvi, T. and E. Honkanen The volatiles of two new strawberry cultivars, Annelie and Alaska Pioneer, obtained by backcrossing of cultivated strawberries with wild strawberries, Fragaria vesca, Rugen and Fragaria virginiana. Z. Lebensm. Unters. Forsch. 175: Kollmannsberger, H. and R. G. Berger Industrial recovery of strawberry favour-fractional distillation versus plate condensation. Z. Lebensm. Unters. Forsch. 198: Larsen, M. and L. Poll Odour thresholds of some important aroma compounds in strawberries. Z. Lebersm. Unters.

6 J. Japan. Soc. Hort. Sci. 78 (2): Forsch. 195: Larsen, M., L. Poll and C. E. Olsen Evaluation of the aroma composition of some strawberry (Fragaria ananassa Duch.) cultivars by use of odour threshold values. Z. Lebensm. Unters. Forsch. 195: Latrasse, A Fruits III. p In: H. Maarse (ed.). Volatile compounds in foods and beverages. Dekker, New York. Menager, I., M. Jost and C. Aubert Changes in physicochemical characteristics and volatile constituents of strawberry (cv. Cigaline) during maturation. J. Agric. Food Chem. 52: Moyano, L., L. Zea, J. Moreno and M. Medina Analytical study of aromatic series in sherry wines subjected to biological aging. J. Agric. Food Chem. 50: Mura, E. and I. Zabetakis The effect of exogenous pectinase on DMHF and derivatives in clarified strawberry juice (Fragaria ananassa, cv. Elsanta). Flavour Fragr. J. 17: Oda, Y. and M. Kimura I. Origin and History of Strawberry. p In: Rural Culture Association (ed). Strawberry, Encyclopedia of Vegetables vol. 3. Rural Culture Association, Tokyo, Japan (in Japanese). Park, E. R., H. J. Lee and K. S. Kim Volatile flavor components in Bogyojosaeng and Suhong cultivars of strawberry (Fragaria ananassa Duch.). J. Food Sci. Nutr. 5: Perez, A. G., R. Olias, C. Sanz and J. M. Olias Furanones in strawberries: evolution during ripening and postharvest shelf life. J. Agric. Food Chem. 44: Pickenhagen, W., A. Velluz, J. P. Passerat and G. Ohloff Estimation of 2,5-dimethyl-4-hydroxy-3(2H)-furanone (Furaneol) in cultivated and wild strawberries, pineapples and mangoes. J. Sci. Food Agric. 32: Pyysalo, T., E. Honkanen and T. Hirvi Volatiles of wild strawberries, Fragaria vesca L., compared to those of cultivated berries, Fragaria ananassa cv. Senga Sengana. J. Agric. Food Chem. 27: Sanz, C., A. G. Perez and D. G. Richardson Simultaneous HPLC determination of 2,5-dimehyl-4-hydroxy-3(2H)- furanone and related flavor compounds in strawberries. J. Food Sci. 59: Sanz, C., D. G. Richardson and A. G. Perez ,5-Dimethyl- 4-hydroxy-3(2H)-furanone and derivatives in strawberries during ripening. p In: R. L. Rouseff and M. M. Leahy (eds.). Fruit Flavors: Biogenesis, Characterization and Authentication. ACS Symp. Series 596. ACS, Washington, D.C. Schieberle, P Heat-induced changes in the most odouractive volatiles of strawberries. p In: H. Maarse and D. G. van der Heij (eds.). Trends in flavour research. Elsevier Science, Amsterdam. Schreier, P Quantitative composition of volatile constituents in cultivated strawberries, Fragaria ananassa cv. Senga Sengana, Senga Litessa and Senga Gourmella. J. Sci. Food Agric. 31: Sen, A., P. Schieberle and W. Grosch Quantitative determination of 2,5-dimethyl-4-hydroxy-3(2H)-furanone and its methyl ether using a stable isotope dilution assay. Lebensm. Wiss. Technol. 24: Siegmund, B., K. Derler and W. Pfannhauser Changes in the aroma of a strawberry drink during storage. J. Agric. Food Chem. 49: Ulrich, D., E. Hoberg, A. Rapp and S. Kecke Analysis of strawberry flavour-discrimination of aroma types by quantification of volatile compounds. Z. Lebensm. Unters. Forsch. 205: Urruty, L., J. L. Giraudel, S. Lek, P. Roudeillac and M. Montury Assessment of strawberry aroma through SPME/GC and ANN methods. Classification and discrimination of varieties. J. Agric. Food Chem. 50: Zabetakis, I. and M. A. Holden The effect of 6-deoxy-Dfructose on favour bioformation from strawberry (Fragaria ananassa cv. Elsanta) callus cultures. Plant Cell Tiss. Org. Cult. 45: Zabetakis, I. and M. A. Holden Strawberry flavour: Analysis and biosynthesis. J. Sci. Food Agric. 74: