Biochemical Society Transactions

Transcription

1 702 Plant cell culture in the production of flavour compounds Michael J. C. Rhodes, Andrew Spencer and John D. Hamill* A.F.R.C. Institute of Food Research, Norwich Research Park, Colney Lane, Norwich NR4 7UA Plants are an important source of natural flavourings for food. Plant-derived food flavours may be based on endogenous flavour compounds or be derived from flavour precursors following enzyme action when the tissue is disrupted or during a postharvest treatment. Such flavours are usually based either on a single compound or a small group of compounds with closely related chemical structure. However, many of the plant-derived flavours which are of most significance in the food industry are complex mixtures of compounds which each contribute to the overall flavour; these include fruit flavours and the essential oils. The present paper will describe experiments aimed at developing tissue-culture systems in which to study these more complex flavours. A variety of plant tissue-culture systems are now available for study of flavour compound production. These include callus and cell-suspension cultures which have been used in studies of simple flavours such as capsaicin formation in Capsicum [l] and vanillin formation in VaniZZa [2]. In the case of Capsicum, capsaicin production is enhanced significantly if the cells are immobilized in polyurethane foam particles [ 31. However such systems have generally proved unsuccessful with more complex fruit flavours, where frequently only a small part of the spectrum of flavour compounds is formed in culture [4]. Disorganized plant cell cultures are often genetically heterogeneous and although they are totipotent in that single cells can be induced to regenerate whole plants, it is frequently not possible to induce selective expression of the pathways leading to the formation of the desired flavour compounds. In the intact plant, these pathways are expressed in a co-ordinated fashion within the patterns of differentiation of plant organs, in particular tissues at defined stages in organ development. Although some genes involved in the co-ordination of expression of plant pathways, such as the Zc genes involved in the pathway of formation of anthocyanin pigments, have been identi- Abbreviations used: tmsl, gene coding for tryptophan mono-oxygenase; tms2, gene coding for indoleacetamide hydrolase; ipt, gene coding for dimethylallyl transferase. 'Present address: Department of Genetics and Developmental Biology, Monash University, Clayton 3 168, Melbourne, Australia. tied [S], the regulation of expression of secondary pathways in plants is poorly understood. We have investigated organized plant organ cultures, in which the internal regulatory system is intact, to study the production of complex flavours where both the spectrum of compounds formed and the concentration of the individual components are fundamental to the acceptability of the flavour. In this work, the formation of monoterpene compounds in the essential oil has been studied in two mint species, a Bergamot mint, Mentha piperita var. citrata ('M citrata') and in a peppermint variety (Black mint), Mentha piperita var. vulgaris ('M. piperita'). These were chosen in that they produce a very different spectra of terpene products; M. Citrata makes a very simple oil consisting of over 90% of two acyclic monoterpenes, linalool and linalyl acetate, while M. piperita makes the typical peppermint cyclic monoterpenes, menthol and menthone. These compounds are formed mainly in the leaves of mint plants but attempts to develop shoot cultures of mint from callus by supplying a suitable combination of plant growth substances in the medium have proved unsuccesful [6]. We approached this problem by developing transformed shoot cultures derived using strains of Agrobacterium tumefaciens. This approach developed from earlier successes with the related organism Agrobacterium rhizogenes, which induces root formation at sites at which it infects the plant. These roots can be freed of bacteria and established in axenic culture and have been exploited in the study of root-derived flavour compounds [7, 81. There is no organism similar to A. rhizogenes which has evolved the ability to induce shoot formation. A. tumefaciens is a plant pathogen which causes the crown gall disease of dicotyledonous species. In both species of Agrobacterium, infection involves transfer of bacterial plasmid genes and their integration and expression in the plant genome. In A. tumefaciens the transferred genes code for the factors leading to the formation in the plant of specific bacterial metabolites, opines, and of the factors inducing the tissue to undergo disorganized growth to form a mass of callus-like tissue, a gall. Strains of A. tumefaciens are classified on the basis of the types of opines they produce, the common ones being octopine and nopaline. The processes that lead to oncogenesis are not fully understood but genes coding for

2 Food Biotechnology enzymes of auxin biosynthesis, tmsl and tms2, which code for tryptophan mono-oxygenase and for indoleacetamide hydrolase respectively, and of cytokinin biosynthesis, $t coding for dimethylallyl transferase play an important role [9]. However other genes of the transferred DNA probably also play a role in modifying the action of the tms and $t genes. Much of the initial work on this system was done with tobacco [lo] and in this species it has been shown that expression of the tms and $t genes leads to an increase in the levels of auxin and cytokinin in the transformed plant tissue. Moreover, it was shown that mutagenesis of either of the tms genes by transposon insertion leads to cultures with an enhanced ratio of cytokinins to auxins and a shoot phenotype compared with the galls induced by the wild-type strains. Conversely, mutagenesis of the $t gene leads to cultures with an enhanced auxin to cytokinin ratio and a root phenotype. The shoot inducing strains of A. tumefaciens are known as shooty mutants. Further approaches in developing transformation procedures based on the principle of increasing the relative level of cytokinin in the tissue have proved successful in inducing shoot formation in tobacco. These involved the development of constructs in which the $t gene is expressed behind promoters of various strengths in the absence of the auxin genes [ll]. A further observation is that some nopaline strains of A. tumefaciens will at low frequency induce shoot formation in tobacco [ 121. These various approaches which have proved to be effective in inducing shoot formation in tobacco were tested in the two Mentha species. Table 1 shows the phenotypes resulting from transformation of stem tissue of the two mint species with a range of Agrobacterium strains. Transformation with shooty mutants induced the expected shooty phenotype in tobacco but produced only galls in Mentha. Similarly, the expression of the ipt gene under the control of its own promoter failed to induce shoot formation in Mentha, even though it did so in tobacco. However, expression of the $t gene behind stronger promoters, such as the cauliflower mosaic virus 35s promoter, with or without the duplicated upstream enhancer sequence [ 13, 141, induced shoot formation in both Mentha and tobacco. These results suggest that shoot induction in Mentha as in tobacco may result from high relative levels of expression of the gene of cytokinin biosynthesis but that higher levels of expression are required to induce shooting in Mentha than in Table I Phenotypes resulting from transformation of tobacco and two Mentha species with a range of Agrobacterium tumefaciens strains Abbreviations used: 0, no response; S, shoots arising directly from wound area; G/S, callus and shoots arising area from wound area: G-G, initial callus gall on wound site, maintaining same phenotype on excision and culture in vitro; G-S, initial gall from which shoots develop during subsequent excision and culture in vitro. For experimental details see [ 13, 141. M. M. Agrobacteriurn strain Tobacco citrato piperito LBA 4404 disarmed endogenous S G-G G-G promoter-ipt promoter-ipt S G-S G-S + E355 promoter-ipt S G-S G-S Wild-type octopine strain G G -G G -G (PAW Shooty mutant of pach5 G/S G - G G - G Tn5::tms I Nopaline strain T37 G/S G-S G-S tobacco. Table 1 also shows that the nopaline strain of A. tumefaciens, T37, induces gall formation on tobacco from which shoots develop at low frequency. These nopaline strains also induced shoot formation on both Mentha species. Thus as a result of this initial work two methods proved successful in inducing the formation of shoots in Mentha; (1) expression of the $t gene behind strong promoters and (2) the use of nopaline strains. It is a characteristic of mint that the shooty phenotype was only expressed after the initially formed galls were excised from the explant and decontaminated from free-living bacteria [ 131. Once established in axenic culture the shooty phenotype is stable and shoot cultures have been maintained for over two years with subculturing at two-weekly intervals. They are maintained in a standard plant cell-culture medium with sucrose as the carbon source but without added growth substances. Details of the procedures for establishing and maintaining axenic cultures are given elsewhere [ 1 3, 141. As a result of these initial tissue culture studies, four cultures were selected for further investigation. These are T37-derived cultures of M. citratu (MC-1) and M. piperdu (BM-I) and cultures of M. piperitu derived following transformation with the $t gene behind either the cauliflower mosaic virus 35s (BM-3%) or the enhanced 35s (BM

3 704 E35S) promoters. These cultures were each developed from single shoot tips. They were shown to be transformed by extraction of total DNA and the demonstration of the presence of the $t gene by Southern blotting and hybridization with a 720 bp fragment of the $t gene and by polymerase chain reaction methodology using specific primers for the $t gene. Further, the cultures were shown to be free of bacterial contamination by standard plating techniques and by the absence of Agrobacterium-specific vir sequences in the extracted total DNA [ 13, 141. Both the T37-derived cultures (MC-1 and BM-1) show a typical growth pattern increasing from an inoculum of g dry weight to g dry weight in days at 26 C and a light intensity of 840 Ix. In contrast, the two 35Spromoter derived cultures (BM-35s and BM-E35S) each grew more slowly than the T37-derived ones and showed very different patterns of growth from each other. The culture BM-E35S grew very much more slowly than BM-35s and took over 50 days to reach a biomass of 0.3 g dry weight from an inoculum of 0.1 g while the BM-35s culture achieved nearly twice this biomass in under 40 days. Analysis of the spectrum and content of monoterpenes by g.1.c. showed that the cultures reflected the parent plant from which the culture was derived. The T37 culture of Mentha citrata, MC-1, produced over 90% of its monoterpenes as the acyclic compounds, linalool and linalyl acetate, in close agreement with the composition of the parent plant [13]. All three M. p$m'ta cultures, as with the parent plant, produce almost exclusively cyclic monoterpenes with menthol as the major component (see Table 2). However, some differences were observed in other components such as menthofuran which is a relatively minor component in the plant but which is a significant component in all three M p$erita cultures developed. This increase in menthanofuran is mainly at the expense of menthone. A range of minor components such as 1 -menthy1 acetate, 1,8-cineol, 1 -limonene, sabanene, isomenthone, d-neomenthol, and piperitone are present in extracts of the plant and the shoot cultures. None of these monoterpenes are formed in disorganized cultures of either mint species, such as those formed following transformation with the $t gene under the control of its endogenous promoter. The total monoterpene production of both BM-35s and of BM-E35S rises throughout the 45 days of culture to yield over 9 mg of total terpene per flask from an inoculum of mg dry weight. This yield of terpenes is obtained in the BM-E35S culture in spite of its overall low biomass formation; the concentration of monoterpenes in the BM-E35S culture at 45 days, at about 2.6 mg/g fresh weight, is over twice that of BM-35s. In both cases the formation of total monoterpenes follows Table 2 Growth and monoterpene production of culture BM- I grown under standard culture conditions Culture conditions: - I g fresh weight of shoots were incubated in 50 ml of BS+ medium under the conditions decribed elsewhere [ 13, 141. Values in parentheses are (% of total monoterpene). Menthyl Time of Menthol Menthone Menthofuran acetate lsornenthone Neornenthol culture Fresh Total (days) wt. (g) (pg/cu Itu re) rnonoterpene I079 (70.5) (68.3) (56.2) I97 (57.6) (45.3) 369 (24. I ) 2786 (29.4) I520 (42.2) I28 (37.4) 300 (46.6)

4 Food Biotechnology growth fairly closely and the level remains stable in stationary phase cultures. However this pattern was not found in all cases as is illustrated in Table 2. This shows the growth and monoterpene production of the culture BM-1 on a medium based on Gamborg s B5 medium supplemented with a higher than normal nitrogen content [14]. The total monoterpene level rises, from an initial low value to a peak at 14 days (9.5 mglflask) at about the time the culture enters slow growth phase. Thereafter there is a rapid loss of monoterpene from the system so that by days over 90% of the terpene has been lost. An important feature in the failure to accumulate monoterpenes in dispersed plant cell cultures is the phytotoxicity of these compounds. In the plant, the problem of phytotoxicity is overcome by sequestering the products in gland cells. In the transformed shoot cultures oil glands develop on the leaflets and to a smaller extent on the stems; these structures are absent from non-accumulating dispersed transformed cultures such as those developed with the $t gene under its own promoter. In culture HM-1, loss of terpenes from the culture after day 14 is associated with loss of oil glands from the surface of the leaflets. The fate of the monoterpenes is uncertain; it may be that they are re-absorbed from the glands and metabolized in the tissue or they may be released into the medium where they may be metabolized by extracellular enzymes such as peroxidases [15]. Small amounts of methanol have been detected in the medium late in the culture cycle of BM-1, but not enough to explain the overall loss from the system. The use of extracellular adsorbents to remove this monoterpene from the culture medium and to protect it from degradation is currently being investigated. The loss of monoterpene from culture BM-1 occurs fairly evenly in all components although the relative level of menthofuran tends to rise late in the culture. At its peak, at 14 days, the concentration of total monoterpenes in the culture is similar to that in the parent plant. Similar results have been obtained with the other cultures (BM-35s and MC- 1) with terpene yields at, or just slightly lower than, that of the parent plant tissue. However, the BM- E35S culture produces up to three times more monoterpene than the parent plant. Such comparisons are of necessity approximate, as the yields of terpene in the field and in culture are affected by a number of developmental and environmental factors. Similarly, the spectrum of monoterpenes formed is also influenced by external factors [16]. For instance the level of menthofuran accumulated in the whole plant is known to be affected by environmental factors [ 161 and the differences in composition between the culture and the parent plant tissue may, at least in part, be caused by the influence of these environmental factors. The transformed shoot cultures of Mentha described in this paper represent an important step forward in developing systems in nitro in which to study the biosynthesis of monoterpenes in mint oil. They represent an improvement over previously studied mint cell-culture systems which produce only a very low yield of a partial spectrum of monoterpenoid compounds [6]. The present cultures reflect the parent plant in the range and yield of individual monoterpenes produced except for the increased yield of menthofuran in the M. piperita cultures. It is likely that the approach of establishing transformed shoot cultures described here may provide a model for the development of similar systems to study other essential oil producing species. 1. Yeoman, M. M.. Holden, M. A., Hall, R. D., Holden, P. K. & Holland, S. S. (1989) in Primary and Secondary Metabolism of Plant Cell Cultures 11 (Kurz, W. G. W., ed.), pp , Springer-Verlag, Berlin 2. Knuth, M. E. & Sahai, 0. P. (1989) PCT Patent W089/ Lindsey, K. & Yeoman, M. M. (1984) I lanta 162, Hong, Y. C. & Harlander, S. K. (1989) in Flavour Chemistry of Lipid Foods (Min, D. B. & Smouse, T. H., eds.), pp , AOCS, Champaign, Illinois 5. Ludwig, S. K., Habera, L. F., Dellaporta, S. L. & Wessler, S. R. (1989) Proc. Natl. Acad. Sci. USA. 86, Bricout, J. & Paupardin, C. (1975) C.R. Acad. Sci. Paris 286, Hamill, J. D., Parr, A. J., Khodes, M. J. C., Robins, R. J. & Walton, N. J. (1987) Bio/Technology 5, Hamill, J. D., Robins, R. J. & Khodes, M. J. C. (1989) Planta Medica 55, Morris, K. 0. (1986) Annu. Rev. Plant Physiol. 37, Akiyoshi, D. E., Morris, R. O., Hinz, R., Mischke, B. S., Kosuge, T., Garfinkel, D. J., Gordon, M. P. & Nester, E. W. (1983) Proc. Natl. Acad. Sci. USA. 80, Smigoki, A. C. & Owens, L. D. (1980) Proc. Natl. Acad. Sci. U.S.A. 85,s Gresshof, P. M., Skitnicki, M. L. & Rolfe. B. G. (1979) J. Bacteriol. 137, Spencer, A,, Hamill, J. D. & Rhodes, M. J. C. (1990) Plant Cell Rep. 8,

5 Spencer, A,, Hamill, J. D., Reynolds, J. & Rhodes, 16. Loomis, W. L). (1967) in Terpenes in Plants M. J. C. (1990) in Proc. 7th Int. Congr. Plant Tissue and Organ Culture, Amsterdam (Nijkamp, H. J. J., (Pridham, J. B., ed.), pp , Academic Press, London Van der Plas, I,. H. W. & Van Aartrijk, J. eds.), pp , Kluwer Academic Publishers, Dordrecht 15. Gershenzon, J. & Croteau, R. (1990) Recent Adv. Phytochem. 24, Received 12 April 1991