Increasing antioxidant levels in tomatoes through modi cation of the avonoid biosynthetic pathway

Transcription

1 Journal of Experimental Botany, Vol. 53, No. 377, Fruit Development and Ripening Special Issue, pp. 2099±2106, October 2002 DOI: /jxb/erf026 Increasing antioxidant levels in tomatoes through modi cation of the avonoid biosynthetic pathway M. E. Verhoeyen 1,3, A. Bovy 2, G. Collins 1, S. Muir 1, S. Robinson 1,C.H.R.deVos 2 and S. Colliver 1 1 Unilever R&D Colworth, Sharnbrook, Bedford MK44 1LQ, UK 2 Plant Research International, PO Box 16, 6700 AA Wageningen, The Netherlands Received 12 April 2002; Accepted 29 May 2002 Abstract Flavonoids are a diverse group of phenolic secondary metabolites that occur naturally in plants and therefore form an integral component of the human diet. Many of the compounds belonging to this group are potent antioxidants in vitro and epidemiological studies suggest a direct correlation between high avonoid intake and decreased risk of cardiovascular disease, cancer and other age-related diseases. Enhancing avonoid biosynthesis in chosen crops may provide new raw materials that have the potential to be used in foods designed for speci c bene ts to human health. Using genetic modi cation, it was possible to generate several tomato lines with signi cantly altered avonoid content and to probe the role and importance of several key enzymatic steps in the tomato avonoid biosynthetic pathway. Most notably an up to 78-fold increase in total fruit avonols was achieved through ectopic expression of a single biosynthetic enzyme, chalcone isomerase. In addition, chalcone synthase and avonol synthase transgenes were found to act synergistically to up-regulate avonol biosynthesis signi cantly in tomato esh tissues. Key words: Chalcone isomerase, chalcone synthase, avonol synthase, avonols, tomato, transcription factors. Introduction The avonoids form a large family of low molecular weight polyphenolic compounds, which occur naturally in plant tissues and include the avonols, avones, avanones, catechins, anthocyanins, iso avonoids, dihydro avonols, and stilbenes (Haslam, 1998). To date, more than 4000 avonoids have been described. Notably, most are conjugated to sugar molecules and are commonly located in the upper epidermal layers of leaves (Stewart et al., 2000). However, they also occur naturally in fruits, vegetables, nuts, seeds, and owers and therefore form an integral part of the human diet. In plants, the avonoids are thought to have many functions including protection against UV-B radiation, defence against pathogen attack, attractants to pollinating insects, and as signal compounds for the initiation of symbiotic relationships (Parr and Bolwell, 2000). As a dietary component, the avonoids are thought to have health-promoting properties, probably due to their high antioxidant capacity (Duthie and Crozier, 2000; Pietta, 2000). This function/activity is supported by their ability, in vitro, to induce human protective enzyme systems (Nijveldt et al., 2001) and by a number of epidemiological studies which suggest a protective effect against cardiovascular disease in particular (Hertog et al., 1997), but also against cancer (Knekt et al., 1997) and other age-related diseases such as dementia (Commenges et al., 2000). However, more evidence for their bene cial properties is needed from in vivo supplementation studies, and a few encouraging reports have been published recently. For example, Duarte et al. (2001) reported that quercetin has the ability to lower blood pressure in spontaneously hypertensive rats. Another study in humans concluded that alcohol-free red wine extract and one of its components, quercetin, can inhibit LDL oxidation after in vivo supplementation (Chopra et al., 2000). 3 To whom correspondence should be addressed. Fax: +44 (0) Martine.Verhoeyen@unilever.com Abbreviations: cdna, complementary DNA; CHS, chalcone synthase; CHI, chalcone isomerase; F3H, avanone 3-hydroxylase; FLS, avonol synthase; PAL,-phenylalanine ammonia lyase; FW, fresh weight; DW, dry weight; dpa, days post-anthesis; PCR, polymerase chain reaction. ã Society for Experimental Biology 2002

2 2100 Verhoeyen et al. Fig. 1. Flavonoid biosynthetic pathway. Included are the structures of the major avonol glycosides in tomato, rutin and its precursor isoquercitrin and kaempferol-3-rutinoside. Based on these types of studies, there is a growing interest in the development of food crops with enhanced levels of avonoids. Tomato is an excellent candidate: it is a major crop, and the peel of its fruit already contains modest levels of avonols, a sub-class of avonoids thought to be particularly bene cial to human health. Therefore, the aim was to up-regulate avonoid biosynthesis in tomato in order to augment the fruit's antioxidant capacity and widen its range of health bene t properties. Other groups have reported modi cation of the avonoid pathway in other species, however, this research has been mainly directed towards alteration of anthocyanin pigmentation (Dixon and Steele, 1999; Forkmann and Martens, 2001; Mol et al., 1998). In tomato, avonol biosynthesis was characterized at the molecular, biochemical and tissue-speci c levels using a commercial processing tomato variety, FM6203. This knowledge was used to adopt a transgenic approach generating tomato transformants in which sequences encoding key biosynthetic enzymes, either alone or in combination, or encoding regulatory elements are ectopically expressed. These approaches resulted in a number of modi ed tomato

3 Fig. 2. Flavonol (sum of quercetin and kaempferol) levels in the peel of transgenic tomato fruit containing the P. hybrida chalcone isomerase gene (Muir et al., 2001). Flavonoids were extracted from peel tissue under hydrolysing conditions and analysed by HPLC. The black bar represents plant lines transformed with a control plasmid, and represents the mean value from two independent fruit readings from ve independently transformed plants. The grey shaded bars represent fruit transformed with a sequence encoding P. hybrida CHI, pbbc50. In this case, each bar represents the mean of two independent fruit readings from one transgenic plant. The graph re ects data obtained from analysis of the ten highest avonol accumulators. lines with increased levels of avonols. A review of this work is presented in this paper. Antioxidant levels in tomato 2101 Flavonoid biosynthesis in tomatoes The avonoid biosynthetic pathway (Fig. 1) and its regulation have been well studied in plants (Weisshaar and Jenkins, 1998) and many enzymes required for the production of different avonoid classes have been characterized (Winkel-Shirley, 2001; Holton and Cornish, 1995; Mol et al., 1998). Previous studies have indicated that avonoids accumulate in tomato fruit in a tissue-speci c manner (Krause and Galensa, 1992). The analyses of tomato fruit of the commercial variety FM6203, at different ripening stages, showed that the accumulation of avonoids in tomato fruit is indeed regulated in a tissue-speci c, but also in a development-dependent, manner. The main avonoid found is naringenin chalcone. Naringenin chalcone accumulates almost exclusively in peel tissue and is simultaneously formed with colouring of the fruit, peaking at up to ~1% on a dry weight basis in turning peel tissue (Muir et al., 2001). In addition, the avonols quercetin-rutinoside (rutin) and, to a lesser extent, kaempferol-rutinoside, also accumulate almost exclusively in ripening tomato peel. Furthermore, analysis of avonol accumulation during fruit ripening shows that these avonols are also simultaneously formed with colouring of the fruit, peaking at up to ~0.1% on a dry weight basis in overripe (~45 dpa) peel tissue (Muir et al., 2001). This tissue-speci c accumulation is in agreement with other studies which have shown between 95±98% of the naringenin chalcone and ~98% of the avonols (quercetin- and kaempferol-glycosides) to be located in peel tissue (Krause and Galensa, 1992; Stewart et al., 2000). Northern analysis showed that, in peel tissue from FM6203 fruit, chs, f3h and s transcripts were abundantly expressed whereas chi transcripts were below the levels of detection. Similarly, in tomato esh tissues all three transcripts chs, chi and s were below the level of detection (Muir et al., 2001; S Muir, unpublished observations). These observations appear to correlate with the biochemical phenotype, i.e. accumulation of naringenin chalcone in peel tissue and negligible levels of avonoid accumulation in esh (pericarp and columella) tissues. Given the low levels of avonols detected in FM6203, a number of tomato varieties were analysed in order to investigate the extent of natural variation within existing germplasm. The analyses of fruit from 28 varieties, grown simultaneously and under the same conditions indicated a variation of up to ~10-fold in avonol levels. Quanti cation of quercetin present in hydrolysed peel extracts showed that levels were in a range between 6.3± 64.9 mgg ±1 FW, however, the higher values only represent an up to 2.5-fold increase relative to the commercial variety FM6203 (data not shown). Interestingly, Stewart et al. (2000) also recently reported their analyses of a number of different tomato varieties and showed a similar degree of variation (1.3±22.2 mg g ±1 FW) in total fruit avonols. These ndings suggest that signi cant increases in fruit avonol levels, above those found in FM6203, may not be obtained from utilizing existing tomato cultivars. Enhancing avonol accumulation in tomato peel At a biochemical level it has been shown that peel tissue accumulates naringenin chalcone to a high level throughout fruit ripening, but peaks at up to ~1% DW in turning fruit (31 dpa). It has also been reported that, although chs, f3h and s transcripts are expressed abundantly throughout ripening, the level of chi transcript is below the level of detection (Muir et al., 2001; S Muir, unpublished observations). Notably, the enzyme chalcone isomerase utilizes naringenin chalcone as a substrate to produce dihydro- avonols (Fig. 1) and, therefore, the biochemical and molecular analyses suggest that chi expression might be a rate-limiting step in avonol biosynthesis in tomato peel. Clearly, one approach to overcoming a potential ratelimiting step is ectopically to express a sequence encoding the particular enzymic activity. Therefore, tomato was transformed with a sequence from P. hybrida encoding CHI, under the control of the strong constitutive double

4 2102 Verhoeyen et al. Fig. 3. Control and high avonol tomatoes. FM, FM6203, non-transgenic control and parental variety; C11 hz, homozygous line C11 containing P. hybrida CHI gene; C65 hz, homozygous line C65 containing P. hybrida CHI gene; C65 az, azygous line C65 (lost P. hybrida CHI gene through segregation). CaMV35S promoter and the resulting tomato plants analysed for avonoid accumulation in red-ripe (~40 dpa) fruits (Muir et al., 2001). Analysis of the primary transformants, which harbour the CHI transgene, revealed dramatic increases in fruit peel avonol levels compared with control plants (up to 78-fold increase in individual fruits) (Fig. 2). This rise in total avonol accumulation mainly comprised increases in the accumulation of rutin (quercetin 3-O-rutinoside), isoquercitrin (quercetin-3-oglucoside) and kaempferol-3-o-rutinoside in the peel tissues. Furthermore, there was a direct relationship between quercetin and kaempferol levels in hydrolysed extracts from transformed fruit, with those plants displaying a high quercetin phenotype also possessing high levels of kaempferol (the ratio of quercetin:kaempferol being approximately 10:1). The analyses also showed that, in contrast to control fruit, the level of naringenin chalcone accumulation was severely depleted in the high avonol fruit. Given that naringenin chalcone is a substrate for CHI, this nding suggests that ectopic expression of the P. hybrida CHI results in an increased level of CHI activity, which in turn utilizes the naringenin chalcone pool. In agreement, Northern analysis of transgenic fruit peel revealed low but detectable levels of CHI transcript at all stages of fruit development. By contrast, chi transcripts were below the level of detection in control fruit. It is also noteworthy that avonol accumulation was essentially linear at the post-`mature-green' stage (~7 dpa) of fruit ripening, correlating with the kinetics of naringenin chalcone accumulation detected in non-transformed fruit. These results may suggest that during fruit development the availability of naringenin chalcone is limiting in green fruit from CHI transformants rather than CHI activity as CHI transgene transcripts are present throughout fruit ripening. The exact mechanism by which the increase in CHI messenger RNA (mrna) leads to higher levels of avonols in peel is unclear. However, it has been shown that ectopic expression of CHS, F3H, FLS singly and CHS- FLS, CHS-F3H or F3H-FLS in combination does not signi cantly increase the level of avonols in peel tissue (S Muir, S Colliver, unpublished observations). As such, one might predict that the endogenous CHI activity constitutes the sole rate-limiting step in peel in postmature-green tomato fruit. However, it is interesting to note, from a mass balance perspective, that the magnitude of rutin accumulation is signi cantly greater than the initial level of naringenin chalcone detected. This might suggest that, in fact, ectopic expression of CHI utilizes the naringenin chalcone pool and that depletion of this substrate pool removes a point of negative feedback in avonol biosynthesis resulting in increased ux into the pathway. Alternatively, the increase in the level of CHI mrna could lead to changes in the activity of other enzymes in the pathway, for example, by other feedback/forward regulation mechanisms, or by stabilization of biosynthetic complexes containing CHI (Burbulis and Winkel-Shirley,

5 Fig. 4. Flavonoid levels in transgenic tomato fruit containing sequences encoding P. hybrida chalcone synthase, chalcone isomerase, avanone 3-hydroxylase, and avonol synthase (Colliver et al., 2002). Flavonoids were extracted from peel tissue under non-hydrolysing conditions and analysed by HPLC. (A) Rutin levels in peel tissue. (B) Kaempferol glycoside levels in columella tissue. (C) Naringenin-7- glucoside levels in columella tissue. Black bars represent plant lines transformed with a control plasmid, and represent the mean value following analysis of a single fruit from ten independently transformed plants. Grey shaded bars represent fruit transformed with the four P. hybrida sequences (pbbc800). In this case, each bar represents analysis of a single fruit from each independently transformed plant. The graph re ects data obtained from analysis of the 15 highest avonoid accumulators. 1999). It is interesting to note that ectopic expression of CHI did not exclusively result in the accumulation of rutin, the quercetin glycoside found in wild-type tomato peel. In fact, the signi cant accumulation of isoquercitrin, the Antioxidant levels in tomato 2103 immediate precursor to rutin, perhaps raises the possibility that by increasing ux through the avonol pathway the enzyme that converts isoquercitrin into rutin, rhamnosyl transferase, has become limiting. This accumulation of isoquercitrin is very interesting from a nutritional point of view since this compound is thought to be a more bioavailable glycoside of quercetin (Olthof et al, 2000). Analyses of the vegetative phenotype showed that there were no gross differences in appearance between the high avonol tomato lines and control lines (Fig. 3). In addition, it has been shown that this high avonol phenotype is stable and segregates with the CHI transgene through at least four generations. Fruit avonol levels in the best lines were similar to those found in yellow onions, a crop with naturally high levels of avonols (Ewald et al., 1999). Furthermore, the effect of processing on these high avonol fruit has been tested and it was found that most of the avonols were retained in processed paste, which was indistinguishable from control paste both in taste and avour (data not shown). These studies show that ectopic expression of one gene, P. hybrida CHI, is suf cient to increase avonol accumulation in tomato peel. However, despite the fact that CHI transcripts were detected at relatively high levels in leaf samples, and in green, breaker and turning esh (pericarp) from high avonol plants, no increases in avonol levels were observed in these tissues. This indicates that, at least in tomato, avonoid biosynthesis is subject to tissuespeci c regulation and that to achieve signi cant increases in avonol accumulation in tomato esh a different approach is required. Enhancing avonol accumulation in tomato esh tissue In esh (pericarp and columella) tissue, the level of expression of endogenous avonoid genes (pal, chs, chi, f3h, s) was shown to be below detection limits, even when using TAQMAN analysis (A Bovy, unpublished observations). These ndings are in agreement with the biochemical analyses, which show that tomato esh tissue accumulates only trace amounts of avonols. Further, the analyses indicate that to achieve up-regulation of esh avonols would require a different approach to that used successfully in peel tissue. One such approach might be to up-regulate the endogenous pathway by ectopic expression of appropriate regulatory elements. Several groups have successfully demonstrated this principle by using transcription factors such as Lc and C1, to increase anthocyanin levels in plants (Bradley et al., 1998; de Majnik et al., 2000; Goldsbrough et al., 1996). However, none of these groups has reported increases in avonol levels. By contrast, an up to 60-fold increase in kaempferol glycosides has recently been achieved in tomato esh tissue by simultaneous ectopic expression of the two maize

6 2104 Verhoeyen et al. Fig. 5. Flavonoid accumulation in columella tissue of transgenic tomato fruit resulting from crosses between plants containing sequences encoding P. hybrida chalcone synthase and avonol synthase (Colliver et al., 2002). The presence of both genes was con rmed by PCR analysis. Flavonoids were extracted from columella tissue under non-hydrolysing conditions and analysed by HPLC. Black bars represent the level of kaempferol glycoside accumulation. Grey bars represent the level of naringenin-7-glucoside accumulation. In each case the `control' bar represents transformed lines harbouring a control plasmid, and represents the mean value following analysis of single fruit from ve independently transformed plants. The bars corresponding to CHS/FLS, represent lines harbouring both CHS and FLS transgenes and represent the mean value following analysis of single fruit from three independent progeny. transcription factors Lc and C1. This resulted in an overall increase in total fruit avonols of up to 20-fold (Bovy et al., 2002). In an alternative approach, P. hybrida sequences encoding each of the key biosynthetic enzymes leading to avonols, chalcone synthase (CHS), chalcone isomerase (CHI), avanone-3-hydroxylase (F3H), and avonol synthase (FLS) were ectopically expressed simultaneously. HPLC analyses of primary transformants containing all four transgenes showed that a number (~75%) of these tomato lines accumulate very high levels of quercetin glycosides in the peel and, more modest, but signi cantly increased levels of kaempferol- and naringenin-glycosides in columella tissue (Colliver et al., 2002) (Fig. 4). The high quercetin phenotype in the peel was expected because of the presence of the CHI transgene, and it is noteworthy that the levels detected were similar to those found in CHI-only transformants. However, the role of each of the individual transgenes in determining the accumulation of kaempferoland naringenin-glycosides in columella tissue remained unclear. Interestingly, transcript analysis of Lc+C1 transformed fruit showed that chs and f3h gene expression is increased ~160-fold and ~174-fold, respectively. In addition, s gene expression was also found to be slightly increased (~10- fold) (A Bovy et al., unpublished data). This observation might suggest that chs and f3h are the key genes required for avonol accumulation in tomato esh. To test this hypothesis, tomato was transformed with a construct containing the P. hybrida CHS and F3H transgenes. In addition, the P. hybrida CHI transgene was included for enhanced avonol accumulation in peel tissue. Surprisingly, it was found that ectopic expression of CHS and F3H in conjunction with CHI led to the expected increase in peel avonols, but was not suf cient to upregulate avonol accumulation in esh (pericarp and columella) tissues (Colliver et al., 2002). As described above, it was shown that concomitant ectopic expression of CHS, CHI, F3H, and FLS in tomato fruit does result in increased levels of avonol accumulation in both peel and esh tissues. Therefore, comparing these ndings with those from CHS±CHI±F3H transformants suggested that FLS might play an important role in avonol synthesis in esh tissues after all. These studies emphasize the complex nature of avonoid regulation in tomato fruit, at least at the biosynthetic gene level, and the potential hazards of trying to correlate molecular (transcript) and biochemical phenotypes. In an attempt to elucidate the key biosynthetic enzymes that determine accumulation of avonols throughout the tomato fruit tissue, each of the key enzymes have been ectopically expressed both individually and in selected combinations. In esh (pericarp and columella) tissues, analysis following expression of single gene constructs showed that none of the four biosynthetic genes were able, individually, to up-regulate avonol biosynthesis in tomato esh. However, notably, ectopic expression of CHS did indicate that the transgene was capable of directing ux towards avonoid biosynthesis resulting in an increase in the accumulation of naringenin-glycosides, particularly naringenin-7-glucoside, in columella tissue (Colliver et al., 2002). This nding might suggest that additional enzyme(s) are required to `pull' carbon ux through to avonol end-products. In addition to the `single gene' transformants, a number of `two-gene' combinations were generated by crossing of parent plants harbouring single gene constructs. HPLC analyses of fruit from these transformed lines revealed that the genes that appear to be critical in leading to avonol biosynthesis in tomato esh (pericarp and columella) tissue are CHS and FLS (Colliver et al., 2002). As described previously, ectopic expression of CHS resulted in modi ed tomatoes accumulating increased levels of naringenin-glycosides but with no increase in avonols. By contrast, analysis of tomatoes harbouring the FLS transgene showed that no signi cant difference in biochemical phenotype was detectable when compared to control fruit (results not shown). The analyses have shown that concomitant expression of both CHS and FLS has a synergistic effect resulting in a signi cant accumulation of both naringenin- and kaempferol-glycosides in tomato esh (Fig. 5). This nding was unexpected since transcript analysis of esh from tomatoes transformed with Lc and

7 C1 had shown up-regulation of chs and f3h expression, whilst for s only a very small increase was observed. However, these studies clearly indicate that in esh tissue CHS is a key gene in `pushing' carbon into the pathway (naringenin chalcone) and that FLS, but not CHI or F3H, is required to `PULL' carbon ux through from naringenin chalcone to the avonol end-product. Given the organization of the biosynthetic pathway (Fig. 1), these ndings, perhaps, suggest that CHS and FLS play a key role in stabilizing a metabolic complex comprising CHS, CHI, F3H, and FLS in tomato esh tissue. Enhancing avonol accumulation throughout tomato fruit The investigations to date indicate that chi gene activity appears to be key to avonol accumulation in tomato peel, whilst chs and s activities are required for the production of avonols in esh tissue. Therefore, it was reasoned that, to achieve increased avonol accumulation throughout the tomato fruit, ectopic expression of three genes encoding the biosynthetic enzymes CHS, CHI and FLS would be suf cient. Indeed a cross harbouring these three genes accumulates increased levels of quercetin-glycosides in peel and kaempferol-glycosides in esh (S Colliver, unpublished results). It is also noteworthy that a similar phenotype can be achieved by crossing tomatoes containing Lc and C1 transgenes with tomatoes containing the CHI transgene (S Muir, unpublished results). Conclusions These studies demonstrate different routes to up-regulate avonoid biosynthesis in tomatoes, through the ectopic expression of either a select number of key biosynthetic genes or key regulatory elements, or a combination of both. In peel tissue, chalcone isomerase gene activity appears to be critical and expression of a sequence encoding the P. hybrida chalcone isomerase leads to a large increase in the level of quercetin-glycoside accumulation. It was further demonstrated that concomitant expression of the sequences encoding chalcone synthase and avonol synthase from P. hybrida is suf cient to achieve accumulation of kaempferol-glycosides in tomato esh. In addition, these studies have shown that ectopic expression of three genes encoding the biosynthetic enzymes CHS, CHI and FLS is required to achieve increased avonol accumulation throughout the tomato fruit. Alternatively, ectopic expression of the regulatory genes Lc and C1, together with the biosynthetic gene CHI results in a similar phenotype. Finally, as quercetin is a more potent antioxidant than kaempferol, it may be desirable to enhance the accumulation of quercetinglycosides in tomato esh, rather than kaempferol-glycosides. This could be achieved by overexpression of avanone 3 -hydroxylase, the enzyme that converts dihydrokaempferol to dihydroquercetin. Dihydroquercetin would then be converted to quercetin by the action of avonol synthase. New tomato varieties enriched in avonols potentially offer opportunities for tomato-based products with heart health bene ts. References Antioxidant levels in tomato 2105 Bovy A, de Vos, R, Kemper M, et al High- avonol tomatoes through heterologous expression of the maize transcription factor genes Lc and C1. The Plant Cell 14, (in press). Bradley JM, Davies KM, Deroles SC, Bloor SJ, Lewis DH The maize Lc regulatory gene up-regulates the avonoid biosynthetic pathway of Petunia. The Plant Journal 13, 381±392. Burbulis IE, Winkel-Shirley B Interactions among enzymes of the Arabidopsis avonoid biosynthetic pathway. Proceedings of the National Academy of Sciences, USA 96, 12929± Chopra M, Fitzsimons PEE, Strain JJT, Thurnham DI, Howard AN Non-alcoholic red wine extract and quercetin inhibit LDL oxidation without affecting plasma antioxidant vitamin and carotenoid concentrations. Clinical Chemistry 46, 1162±1170. Colliver S, Bovy A, Collins G, Muir S, Robinson S, de Vos CHR, Verhoeyen ME Improving the nutritional content of tomatoes through reprogramming their avonoid biosynthetic pathway. Phytochemistry Reviews (in press). Commenges D, Scotet V, Renaud S, Jacqmin-Gadda H, Barberger-Gateau P, Dartigues JF Intake of avonoids and risk of dementia. European Journal of Epidemiology 16, 357±363. de Majnik J, Weinman JJ, Djordjevic MA, Rolfe BG, Tanner GJ, Joseph RG, Larkin PJ Anthocyanin regulatory gene expression in transgenic white clover can result in an altered pattern of pigmentation. Australian Journal of Plant Physiology 27, 659±667. Dixon RA, Steele CL Flavonoids and iso avonoidsða gold mine for metabolic engineering. Trends in Plant Science 4, 394± 400. Duarte J, Perez-Palencia R, Vargas F, Ocete MA, Perez- Vizcaino F, Zarzuelo A, Tamargo J Antihypertensive effects of the avonoid quercetin in spontaneously hypertensive rats. British Journal of Pharmacology 133, 117±124. Duthie G, Crozier A Plant-derived phenolic antioxidants. Current Opinion in Lipidology 11, 43±47. Ewald C, Fjelkner-Modig S, Johansson K, Sjoholm I, Akesson B Effect of processing on major avonoids in processed onions, green beans, and peas. Food Chemistry 64, 231±235. Forkmann G, Martens S Metabolic engineering and applications of avonoids. Current Opinion in Biotechnology 12, 155±160. Goldsbrough AP, Tong Y, Yoder JI, Tong YS Lc as a nondestructive visual reporter and transposition excision marker gene for tomato. The Plant Journal 9, 927±933. Haslam E Practical polyphenolics: from structure to molecular recognition and physiological action. Cambridge: Cambridge University Press. Hertog MGL, Feskens EJ, Kromhout D Antioxidant avonols and coronary heart disease risk. The Lancet 349, 699. Holton TA, Cornish EC Genetics and biochemistry of anthocyanin biosynthesis. The Plant Cell 7, 1071±1083.

8 2106 Verhoeyen et al. Knekt P, JaÈrvinen R, SeppaÈnen R, HeliaÈvaara M, Teppo L, Pukkala E, Aromaa A Dietary avonoids and the risk of lung cancer and other malignant neoplasms. American Journal of Epidemiology 146, 223±230. Krause M, Galensa R Determination of naringenin and naringenin-chalcone in tomato skins by reversed phase HPLC after solid-phase extraction. Zeitschrift fuèr Lebensmittel Untersuchung und Forschung 194, 29±32. Mol J, Grotewold E, Koes R How genes paint owers and seeds. Trends in Plant Science 3, 212±217. Muir SR, Collins GJ, Robinson S, Hughes S, Bovy A, De Vos CHR, van Tunen AJ, Verhoeyen ME Overexpression of petunia chalcone isomerase in tomato results in fruit containing increased levels of avonols. Nature Biotechnology 19, 470±474. Nijveldt RJ, van Nood E, van Hoorn DEC, Boelens PG, van Norren K, van Leeuwen PAM Flavonoids: a review of probable mechanisms of action and potential applications. American Journal of Clinical Nutrition 74, 418±425. Olthof MR, Hollman PCH, Vree TB, Katan MB Bioavailabilities of quercetin-3-glucoside and quercetin-4 glucoside do not differ in humans Journal of Nutrition 130, 1200±1203. Parr AJ, Bolwell GP Phenols in the plant and in man. The potential for possible nutritional enhancment of the diet by modifying the phenols content or pro le. Journal of the Science of Food and Agriculture 80, 985±1012. Pietta PG Flavonoids as antioxidants. Journal of Natural Products 63, 1035±1042. Stewart AJ, Bozonnet S, Mullen W, Jenkins GI, Lean MEJ, Crozier A Occurrence of avonols in tomatoes and tomato-based products. Journal of Agricultural and Food Chemistry 48, 2663±2669. Weisshaar B, Jenkins GI Phenylpropanoid biosynthesis and its regulation. Current Opinion in Plant Biology 1, 251±257. Winkel-Shirley B Flavonoid biosynthesis. A colourful model for genetics, biochemistry, cell biology, and biotechnology. Plant Physiology 126, 485±493.