Caffeic acid oligomers such as rosmarinic acid and lithos-

permic acid B (LAB) are distributed in plants of Labiatae and Boraginaceae. Among these oligomers, LAB, the caffeic acid tetramer, has a variety of attractive pharmacological activities such as antioxidant and radical scavenging, improvement of renal disease, hypotensive, improvement of myocardial infarction, amplification of beating of myocardial muscles, anti-hepatitis, and aldose reductase inhibitory activity. To utilize LAB as a medicinal resource, more than 30 plant sources were surveyed and only two medicinal plants, Salvia officinalis and Lithospermum erythrorhizon SIEB. et ZUCC, were found to contain this substance in an appreciable amount. Many attempts to produce this compound in high quantities using cell or hairy root cultures have been unsuccessful. Recently, we found that L. erythrorhizon cell-suspension cultures, which were established for the industrial production of shikonin, produced four caffeic acid oligomers, rosmarinic acid, LAB, a monoglucoside of LAB, and (1)-rabdosiin. In particular, the production of LAB in the cultured cells was highly stimulated in shikonin production medium M-9, reaching almost the same amount as that of shikonin (LAB: ca. 6—10% of dry wt, and total shikonin derivatives: ca. 6—12% of dry wt.). Since both LAB and shikonin are formed through a common phenylpropanoid pathway (Chart 1), the counterbalance between two metabolic routes either to LAB or to shikonin would be regulated in L. erythrorhizon cells, suggesting that the selection of suitable regulatory factors would result in higher production of LAB. However, the metabolic linkage and the regulatory mechanism for the biosyntheses of phenylpropanoid-derived compounds of diverse types in L. erythrorhizon cell cultures remain to be clarified. In this study, we investigated the critical regulatory factors to control these two biosynthetic pathways separately and to produce LAB efficiently in the culture system of L. erythrorhizon. Experimental Cell Cultures The culture strain M18TOM of Lithospermum erythrorhizon, which had been maintained in Linsmaier-Skoog (LS) medium containing IAA 1 mM and kinetin 10 mM, was used in the present study. Cellsuspension cultures in 100-ml flasks each containing 30 ml of medium were agitated on a rotary shaker (80 rpm) at 23 °C in the dark and subcultured at intervals of 2 weeks. The cells cultured in LS medium for 2 weeks were inoculated into M-9 medium, modified M-9 or LS media containing 3% sucrose, and the same growth regulators as above (inoculum size: 1 g of fresh cells/30 ml of medium in a 100-ml flask) and cultured for 3 weeks under the above-mentioned culture conditions. For the investigation of the effect of light color on the production of secondary metabolites, each flask was covered with a commercially available colored cellophane sheet, and irradiated with white light (10000 lx) from fluorescent lamps during the entire culture period. Under such conditions, light intensities in blue, red, and green cellophane-covered flasks were 600, 1000, and 900 lx, respectively. Quantitative Analysis of Secondary Metabolites Sample preparations and quantitative analyses of secondary metabolites were carried out according to the methods described in the previous report, and the sum of contents of b-hydroxyisovalerylshikonin, acetylshikonin, isobutylshikonin, b ,bdimethylacrylshikonin, isovalerylshikonin, and a-methyl-n-butylshikonin was expressed as shikonin. In addition to shikonin and caffeic acid oligomers, L. erythrorhizon cells produced several phenylpropanoid-derived secondary metabolites, including p-hydroxybenzoic acid 1-O-glucoside (PHBOG), furobenzoquinone and furohydroquinone derivatives such as echinofurans B and C, dihydroechinofuran, deoxyshikonofuran, and shikonofuran E, all of which were biosynthesized from p-hydroxybenzoic acid (PHB). In the present study, we determined the contents of all these compounds and found that the total content of furobenzoquinone and furohydroquinone derivatives was within about 10% of that of shikonin. The contents of caffeic acid oligomers, rosmarinic acid, LAB monoglucoside, and (1)-rabdosiin were also low compared with that of LAB, and were not affected by the supplements of modulators used. Therefore the contents of these compounds are not given in the Results section.