Chemistry, Occurrence and Biosynthesis of C-Glycosyl Compounds in Plants

Home , Bergenin, Dihydroxybenzoic acid, Protocatechuic acid, Trihydroxybenzoic acid

Journal of Planta 1983, Vol.47, pp.131—140,© Hippokrates Verlag GmbH Medicinal .PlantResearchmediCa

Gerhard Franz, Michael Grun1

Lehrstuhl für Pharmazeutische Biologie, Universität Regensburg, Federal Republic of Germany

Received: August 2, 1982

Introduction

Therapid development of natural product chemistry has led to the isolation of a wide varieLty of secondary metabolites, which, in many cases, was shown to be of glycosidic nature. In particular our knowledge of the chemistry, occurrence and biogenesis of glyco- Flavone Xanthone sides has grown substantially during the last two de- cades [1, 2]. Most of this effort has been directed to- ward the large group of the 0-glycosides as well as the S- and N-glycosides. Far less effort has been given to the C-glycosides due to analytical difficulties. C-Glycosides are a special type of glycoside since the aglycone is directly attached to carbon 1 of a pyra- nose ring of a sugar while 0-glycosides possess a he- miacetal linkage. It is more correct to regard them as C-glycosyls or anhydropolyols as proposed by Chromone Anthrone BANDYUKOVA and YUGIN [3]. Since the original term C-glycoside is well established, it should be maintai- ned. One of the main characteristics of C-glycosides is COOH resistance towards acid hydrolysis. Even after pro- longed acid treatment the sugar residue, which is attached by a carbon-carbon-bond, remains attached HOOH to the aglycone. Cleavage of the C-C-linkage is achie- ved only under drastic conditions. When oxidizing reagents are used partial degradation of the sugar re- Gallic acid sidue is likely [4]. Classification of C-glycosides is usually based on Fig. 1. Basic structures of C-glycoside aglycones. the aglycone [5,6].Five aromatic ring systems are known to occur C-glycosylated (Fig. 1). The largest group, the C-glycosyl flavonoids, has been intensive-also exist. Application of such analytical tools as UV, ly studied by many authors and was the subject of twoJR. NMR and mass spectroscopy as well as optical recent review articles [3, 7]. Xanthone C-glycosidesmeasurements has resulted in the identification of appear to be biogenetically related to C-glycosyl fla-large numbers of such compounds in recent years. vonoid compounds. Another group, C-glycosylated Interestingly, C-glycosides are not restricted to chromone derivatives, are relatively rare. Aromaticplants. Carminic acid is an anthraquinone C-glycosi- three-ring systems, notably the anthrones, are C-lin-de from the insect Dactylopius coccus (references gi- ked to glucose. C-Glycosyl derivatives of gallic acidven in [3]). Kidamycin and toromycin are antibiotics produced by Streptomyces sp. and represent a new Present address: Institute of Biological Chemistry, Washing- type of polycyclic microbial metabolite with unusual ton State University, Pullmann, Washington 99164-6340, USA. C-glycosyl moieties [8, 9]. Recently, C-conjugated 132 Franz, GrOn

compounds with glucuronic acid have been found inits specific substitution pattern. The same aglycone mammals, which are probably analogous to C-glyco-may possess a second C-linked sugar or 0-linked su- sylated compounds in plants. C-Glucuronides ofgar residues attached either to phenolic OH-groups drugs in man [10] and other mammals [11, 12] repre-or alcoholic OH-groups of the C-linked sugar. In the sent a novel type of drug metabolite. latter case, C,O-diglycosides and the C,O-triglycosi- As regards the biosynthetic pathway which leads todes are formed. It is interesting to note that in most the formation of the carbon-carbon bond in C-glyco-cases D-glucose serves as C-linked sugar but in the sides, very little is known. In some cases it is thoughtcase of flavonoid derivatives the only C-glycosyl com- that the attachment of the sugar moiety takes placepounds currently known contain other C-linked su- during an early stage of the aglycone formation butgar residues, i.e. galactose, rhamnose, xylose and ar- definite proof is still missing and cannot be generali-abinose [3]. A wide variety of 0-linked sugars may be zed for all groups of the C-glycosides. present as well. As for the aglycone moiety five types The aim of this article is to present a short review ofof flavonoid structures have been demonstrated, i.e., the principal structural features and occurrence offlavones, flavonols, flavanones, isoflavones and di- the different classes of C-glycosides and further, tohydrochalcones [7]. No anthocyanidines have been discuss the few biosynthetic findings of the C-glycosylidentified as C-glycosides. The most widespread attachment which have been published so far. aglycone structures which have been reported in lite- rature and which show different glycosidic variations are apigenin and luteolin [3]. More than 30 derivati- C-GlycosylFlavonoids ves of these aglycones have been reported so far.

Chemistryof C-glycosyl Flavonoids Occurrence of C-Glycosyl Flavonoids As it has been pointed out, this group of C-glycosi- The review of ALSTON [13] on C-glycosyl flavonoids des has been the subject of a still increasing numbermentions about 20 Angiosperm families where these of reports, which mainly deal with the occurrencecompounds occur. There are no reports of C-glycosyl and structural investigation of these compounds. Sin-flavonoids in Gymnosperms and only a few of mosses ce they have been the subject of several review artic- and algae. Recently BANDYUKOVA and YUGIN [3] listed les [3, 6, 7, 13, 14] only the main features are given54 Angiosperm families with a varying number of here. species that contain C-glycosyl flavones in significant At the present over one hundred C-glycosyl fla-quantity. Gymnosperms and several lower plants al- vonoids have been shown to exist in the plant king-so possess these "unusual" flavonoid derivatives. dom. The model compound and one of the first to beOne can conclude that the C-glycosyl flavonoids are isolated was vitexin (8-C-3-D-glucopyranosylapige-widespread among plants, but with restricted chemo- nm) (Fig. 2), the structure of which was establishedtaxonomic significance as CHOPIN and BOUILLANT by NMR spectroscopy by HOROWITZ and GENTJLINIhave pointed out [7]. The most abi2ndant examples [15]. The 6-C-glycosyl isomer of vitexin, named iso-seem to be vitexin, isovitexin and their derivatives. vitexin (saponaretin), may be formed by an intramo- lecular rearrangement [16]. Biosynthesis of C-Glycosyl Flavonoids Most types of substituted flavonoid aglycones oc- OH HO cur as the O-glycoside with different sugars attached to alcoholic or phenolic groups. Biosynthesis of "nor- CH2OH mal" flavonoids and their glycosylated derivatives in- HO 0 volves three acetate units (ring A) and phenylalanine (ring B and the C3 segment) as shown by in vivo and in vitro tracer studies (see reviews by HAHLBROCK and GRISEBACH [17], HAHLBROCK [18]). In the case of C- HO1OS II glycosyl flavonoid aglycones, participation of phenyl- HO 0 alanine and p-coumaric acid in ring B and C formation has been established [19, 20]. Vitexn The mechanism of flavonoid 0-glycosylation, Fig. 2. Vitexin as a representative C-glycosyl flavonoid. whereby the sugar moiety of a nucleotide sugar is transferred via more or less specific transferases to the aglycone acceptor is well understood and is the A free ortho-position to one or two phenolic hy-subject of recent reviews [2, 21]. The second type of droxyl groups appears to be a common feature that isglycoside formation, i.e., the biosynthesis of the C- prerequisite for the formation of a C-glycosidic linka-glycosidic linkage, has never been explained in de- ge in flavonoids. Therefore, the glycosyl residue istail. A few reports, however, give some indication linked to carbon 6 or 8 of ring A (Fig. 1) according toabout the biosynthetic stage of the aglycone at which C-Glycosyl Compounds in Plants 133

the C-glycosidic linkage might be formed. WALLACE Radioactive phenylalanine is rapidly incorpora- et al. [22]feddifferent radioactive precursors such asted into C-glycosyl flavones of the primary leaf of oat the flavone aglycones, apigenin and luteolin, to Spi-plants (A vena sativa, Poaceae). Two vitexin-derived rodela sp. or Lemna sp. (Lemnaceae) and detected0-rhamnosides and one isovitexin 0-arabinoside are radioactivity on 0-glycosylated and 0-methylatedformed and show a slow turnover during the experi- flavones. Radioactivity was not incorporated in fla-mental period of ten days, thus indicating that the C- vones containing C-glycosyl linkages, indicating thatglycosides are not necessarily metabolic end products 0-glycosylation of the flavones occurs readily in vi-[26, 27]. However, MARGNA and VAJNJARV [28] have vo, but C-glycosylation does not occur under condi-demonstrated that amongst the maj or flavonoids of tions in which both compounds are known to be syn-buckwheat seedlings (Fagopyrum esculentum, Faba- thesized in the plant. WALLACE and GRISEBACH [23], ceae), rutin and the C-glycosyl flavonoids do not on the other hand, showed that C-glycosylation wasundergo appreciable turnover during a seven-day only possible at the flavanone level using in vitro cul-period. tured Spirodela polyrhiza (Lemnaceae) clones. Genetic control of flavonoid biosynthesis and Again Wallace [24] stated that formation of the C-mainly 0-glycosylation has been extensively studied. glycosidic bond is always prior to 0-glycosylation.Vitexin and isovitexin glycosylation in Melandrium Pre-existing 0-glycosyl flavonoids are not C-glycosy-sp. and Silene sp. (Caryophyllaceae) are the subject lated. C-Glycosyl flavonoids, however, might be 0-of a series of publications (see review by HOSEL [21] glycosylated in a later step of biosynthesis (Fig. 3).and [29, 30, 31, 32]). Nothing, however, is shown for INouE and FUJITA [25] and Fujita and Inoue [20] supp- the control of the C-glycosylation process. orted these findings by feeding possible 14C-labeled Besides these in vivo studies concerning the forma- precursors to Pueraria lobata (Fabaceae) and Swertiation of C-glycosyl flavonoids essentially nothing is japonica (Gentianaceae) plants. C-Glycosylationknown about the enzymatic mechanisms which lead again occurred at the chalcone of flavanone level. to the formation of the C-glycosidic bond.

H0pOy-©1

Naringenin Vitexin 7—0—glucoside \c_gtycosytation / \odation ,,//O-glycosylatlon

0—glyco— sylation oxidation

Vitexin

Gic OH

0-glycosylation Fig. 3. Sequence of glycosylation and oxidation HO 0 HO 0 reactions involved in the biosynthesis of flavonoid 0-and C-glycosides according to [22,23]. Apigenin7-0—glucoside Apigenin 134 Franz, Grun

Xanthone C-Glycosides cond known xanthone C-glucoside [43]. Mangiferin derivatives, acetylated or benzoylated in the glucose Chemistryof Xanthone C-Glycosides moiety of the molecule, have been reported by More than ten glycoxanthones from plant sources MARKHAM and WALLACE lfl 1980 [44]. 13C-NMR stu- are now known. Some information about the chemo-dies indicate that the glucose 6-hydroxyl is acetylated taxonomic distribution, isolation and structural de-whereas benzoylation occurred at 2'-0-4'-0- and 6'- termination, synthesis and pharmacology of most0-atoms of mangiferin. However, those mangiferin compounds is given by HOSTE-ITMANN and WAGNER lfl derivatives were not obtained pure but could be sepa- a review on xanthone glycosides [33]. rated by paper chromatography as a mixture from an- The first compound with a C-glucosyl residue lin-other component which represented a mangiferin ked to a polyoxygenated xanthone nucleus was obtai-esterified with two mols of benzoic acid. ned by WIEcHowsKI in 1908 [34] but the final structure SMITE and HARBORNE [45] and GLYZIN et a!. [46] of mangiferin (Fig. 4) as 2-C--D-glucopyranosyl-found hydrolyzable derivatives of mangifenn and 1 ,3,6,7-tetrahydroxyxanthone was not establishedisomangiferin which yielded glucose and the two iso- until 60 years later [35—41]. Mangiferin, being themers on acid treatment. Spectral data suggest that best known C-glucosyl xanthone, represents thethe glucosyl residue is attached to a hydroxy group of most important features of all other hitherto knownthe C-glucosyl moiety. The characteristics of those related substances: the glucosyl residue is either atta-C,0-glucosides are analogous to that of 0-glucosides ched to carbon 2 or 4 of the xanthone nucleus whichof 6- or 8-C-flavone glycosides. Two C-glucosyl xan- is oxygenated in 1, 3, 6 and 5 or 7 positions, respecti-thones with a substitution pattern different from that vely. of mangiferin have been found. 2-C-13-D-Glucopyra- Isomangiferin (Fig. 4), the 4-C-glucosyl isomer ofnosyl-1 ,3 ,5 ,6-tetrahydroxyxanthone (Fig. 4) was isomangiferin, has been isolated and characterized bylated by GHOSAL and CHAUDHURI [47]. The structure ARITOMI and KAWASAKI [42]. Interestingly, there is aof its 5-monomethyl ether (Fig. 4) has been determi- similarity between these compounds and other pairsned by chemical and spectral studies and has been na- of position isomers like the flavonoid 8-C- and 6-C-med irisxanthone [48]. Recently, two highly methyla- glucosides orientin and isoorientin, or vitexin (Fig. 2)ted compounds, 2-C--D-g1ucopyranosyl-1-hydroxy- and isovitexin, respectively. The 3-monomethyl et-3,5 ,6-trimethoxyxanthone and 2-C-13-D-glucopyra- her of mangiferin, homomangiferin (Fig. 4), is the se-nosyl-1-hydroxy-3 ,5 ,6,7-tetramethoxyxanthone were described [491.

R2 Occurrence of Xanthone C-Glycosides 0 56OH Mangiferin is the most wide-spread glycoxanthone in higher plants [50]. First isolated from Mangifera indica (Anacardiaceae) [34], it has been detected la- 8 OH HO 0 ter in many other families. In addition ot that reported by CARPENTER et al. [50] it has been found in the Mangiferin: R1 r5—D—gtucopyrariosyt Orchidaceae [51] and some ferns of the Aspleniaceae R2R3H [45] and the Hymenophyllaceae [44]. A more detai- led view about the distribution of mangiferin among Isomangiferin: R2r5—D—glucopyranosyl R'R3H a number of Gentiana and Swertia sp. (Gentiana- ceae) is given in [33]. Although of little taxonomic si- Homomangiferin:R1 5—D—gIucopyranosy( gnificance in the broad sense, its presence or absence R2H, R3rmethyt at the species level may be significant. Its isomer, isomangiferin, has been identified together with mangiferin in Anemarrhena asphodeloides (Liliaceae) [42], Man gifera indica [52], Hedysar- urn flavescens (Fabaceae) [46], Orchidaceae [51] and Aspleniaceae [45]. The fern genus Asplenium (As- pleniaceae) and Hedysarum flavescens are until now the sole source of 0-glycosylated xanthone C-glyco- :ccoOR sides [45, 46]. Homomangiferin, shown to accompa- ny mangiferin in Mangifera indica [53], was also detected in Hoppea dichotoma (Gentianaceae) toget- 2 -C -5 —D—gtucopyranosyl — her with tn- and tetramethoxy-xanthone C-glycosi- 1,3,5,6—tetrahydroxyxanthone: RrH trisxarithorie: R methyI des [49]. Acetylated and benzoylated mangiferins have Fig. 4. Xanthone C-glycosides. been obtained only from some ferns of the Hymeno- C-Glycosyl Compounds in Plants 135

phyllaceae [44] so far. 2-C-3-D-Glucopyranosyl-by FUJITA and INouE [57,58]. These authors have pro- 1 ,3,5,6-tetrahydroxyxanthone was isolated from thevided strong evidence that the 1 ,3,6,7-tetrahydroxy- roots of Canscora decussata (Gentianaceae) [47] andxanthone moiety of mangiferin is derived from a C6- its methyl ether irisxanthone from Iris florentina (In-C3 compound and two acetate units. When [1-14C]-, daceae) [48]. Often, C-glucosylated xanthones co-[2-'4C]- and [3-14C]phenylalanine are fed to excised occur with related C-glucosyl flavonoids [33, 49], aaerial parts of the plant, the intact C6-C3 unit is incor- fact that will be discussed later. porated into the aglycone moiety of mangiferin. [2- '4C]Malonic acid is also incorporated into mangiferin Biosynthesis of Xanthone C-Glycosides with high efficiency. By feeding [1-'4CJ- and [2- Xanthone derivatives are found in nature as free'4C]phenylalanine and [2-14C]malonic acid, the ra- aglycones and 0- and C-glvcosides [33, 50]. It hasdioactivity of mangiferin is localized in the phloroglu- been shown that biosynthesis of the xanthone nu-cinol ring of the aglycone. Furthermore, [3-'4C]cin- cleus in fungi proceeds via an intermediate with anamic acid and p-[2-14Cjcoumaric acid are also utili- benzophenone structure derived from acetate [54]zed for mangiferin and isomangiferin formation. Ad- and from shikimate-acetate in higher plants [55, 561.ministration of C6-C1 compounds (benzoic acid, p- Gentianaxanthones appear to be formed from a C6-hydroxybenzoic acid or protocatechuic acid, 14C-la- C1 precursor derived from phenylalanine by the lossbeled in the carboxyl carbon) did not produce labeled of two carbon fragments and from three acetate unitsC-glycosides. These results suggest that the postula- yielding a benzophenone as intermediate. Subse-ted benzophenone intermediate of mangiferin bio- quently, the xanthone is synthesized by oxidativesynthesis might be formed by condensation of p-cou- coupling of this intermediate. Earlier findings on themarate with two malonates (Fig. 5). Here, the bio- biosynthesis of xanthones are summarized by CAR-synthetic relationship of mangiferin with the flavono- PENTER et al. [50]. ids becomes evident [59]. Recently, the formation of mangiferin in the Lilia- In further studies of C-glycosylation of mangiferin ceae Anemarrhena asphodeloides has been studiedin the Liliaceae Fujita and Inoue fed radioactive

COOH CoAS-..._O H2C-SCoA

I II .CH2 0 HoOOH HOOC

OHCHOH Macturin SCoA

2Malor-iyl—CoA p-Coumaroyt—CoA /YcosYlation

HO OH OH

Gic OH HO 0

3—C—Gtucosyt —macturin

lisation

HO 0 OH TflITTflC II Gtc OH HO91OOH Fig. 5. Proposed pathway for HO 0 the biosynthesis of mangifenn via maclurin after [19, 25]. Mangiferin 1,3,6,7 — Tetrahydroxyxanthone 136 Franz, Grun

1,3 ,6,7-tetrahydroxyxanthone and the benzopheno-Occurrence of C-Glycosyl Chromones ne maclurin to Anemarrhena asphodeloides and Simple chromones which contain no fused ring sy- found that only maclurin was efficiently incorporatedstem on the chromone nucleus are rare as natural into mangiferin and isomangiferin [60] (Fig. 5). The-compounds. This is especially true for C-glycosylated se results indicate that C-glucosylation of mangiferinbenzopyrones. and isomangiferin occurs at the stage of macturin, Aloesin, the first compound of this type, and its 6"- prior to the formation of the xanthone nucleus, andO-p-coumaroylester were detected in pharmaceuti- that both isomers may be biosynthesized via 3-C-gly-cally-used aloes. Subsequently, aloesin has been cosyl maclunn. When this compound is converted tofound to be present in more than thirty Aloe species C-glucosyl xanthone by ring closure, four isomeric C-(Liliaceae). MCCARTHY and HAYNES [65], MCCARTHY glucosyl xanthones may be expected, three of which[66] and FIOLDSWORTH [67] investigated the distribu- have been reported. tion of aloesin and other C-glycosyl compounds in How the C-glycosidic linkage in glycoxanthones ismore than one hundred Aloe species. The two 2"-O- formed is a problem yet to be solved. acylated derivatives (Fig. 6) were obtained from Aloe arborescens MILL. var. natalensis (Liliaceae) [64]. Thus, the C-glycosyl chromones presently known are C-GlycosylChromones found only in Aloe species, an observation of unique chemotaxonomic significance for this species of the Chemistryof C-Glycosyl Chromones Liliaceae. Chromones are derivatives of benzopyran-4-one. They are usually colourless and produce a strong blueBiosynthesis of C-Glycosyl Chromones fluorescence under UV light. This characteristic fluo- Very little is known about the biosynthesis of chro- rescence led to the detection of the first C-glycosylmones in general (confer [68]). Virtually no experi- chromones [61]. ments have been reported dealing with the formation The isolation and structural elucidation of aloesin,on C-glycosylation of the chromone nucleus. Aloesin identical with aloesin B or aloeresin B (Fig. 6) was re-and related compounds differ in their substitution pat- ported by HAYNES et al. [62] and, independently, bytern from most other chromones which have a 2-C- RATTENBERGER [63]. By means of spectroscopic andmethyl group and a 5-C-phenolic group [68] or are degradative studies these authors established itsunsubstituted in the pyrone ring [69, 70]. Thus, it is structure to be 2-acetonyl-8-C43-D-glucopyranosyl-7-difficult to decide whether the aglycones of these hydroxy-5-methylchromone. All other known chro-compounds are derived from acetate or synthesized mone C-glycosides ar derived from aloesin by este-via a "mixed pathway". It is unlikely that they are de- rification with acid products of phenylpropane met-gradation products of flavonoids [69,70] since their abolism. Aloesin A (Fig. 6) was identified as 6"-O-p-substitution pattern differs from that of flavonoids coumaroylaloesin [63]. Two new 2"-O-acylated C-and no related flavonoids from Aloe have been re- glycosyl chromones, isolated by MAKINO et al. [64],ported. were identified as 2"-O-p-coumaroylaloesin and 2"- 0-feruloylaloesin, respectively (Fig. 6). AnthroneC-Glycosides

Chemistryof Anthrone C-Glycosides Naturally occurring anthracene C-glycosides are OH HO derived from 9(1OH)-anthracenone with a C-glucosyl residue linked to carbon 10. These normally yellow CH OR1 substances have been obtained through chemical R2O synthesis by condensation of tetra-O-acetyl-a-D-glucopyranosyl bromide with the appropriate anthrone HO)OCH2....c_CH3 derivative [71, 72]. Mass spectra of a series of anthro- 6EIi 113 ne glycosides were studied and compared by PROX YYO [73] and EVANS et al. [74]. Aloin (=barbaloin),the H3C 0 first crystalline C-glycosyl compound to be isolated Aloesir, (B) R1R2H as early as in 1851 [5], was characterized structurally in 1952 by MUHLEMANN [71] and shown by chemical Atoesin A R1p-coumaroyI, R2rH 2"—O—p—Coumaroyl —aloesin: R1 H, R2 p —coumaroyt synthesis to be 10-C--D-glucopyranosyl-1 ,8-dihy- droxy-3-(hydroxymethyl)-9(1OH)-anthracenone 2"—O—Feruloyl—aloesin : R1rH, R2 feruIoyl (Fig. 7). The suggestion that aloin might exist in two diastereomeric forms [71, 75, 76] due to opposing Fig. 6. C-glycosyl chromones. configurations of the glucose moiety linked to carbon C-Glycosyl Compounds in Plants 137 R10 0 OH H3CO 0 OH 8 II

CH3 Gic HOGic Aloine A/B R1rR2H Homonataloin Atoinosid B :R1 H, R2 rhamnosyI Cascaroside A/B R gtucosyt, R2 :H

R10 0 OH

cIt:c1I:IItL CH3 HO Gic !1—Deoxyaloin R1H Cassialoin Fig.7.Anthrone C-glycosides. CascarosideC/D: R1gtucosyI

10 of the aglycone was recently confirmed by AUTER-tural determination of those C-10-glycosides by me- HOFF et a!. [77] and GRUN and FRANZ [78]. Genuineans of partial hydrolysis, NMR, mass spectrometry atom extracted from plant material or chemicallyand optical measurements were the subjects of sever- synthesized is separated by high performance liquidal publications by FAIRBAIRN et at. [85, 86, 87] and chromatography on reverse phase columns into aloin WAGNER and DEMUTH [76, 88]. A and B which show differences in optical rotation and circular dichroism. The respective diastereomers are interconvertable which might explain their com-Occurrence of Anthrone C-Glycosides mon anthranol form. Aloin is readily oxidized in The commonest C-glycosylated anthracene deri- aqueous solutions. Aloe emodine and 4-hydroxyalo-vative is aloin. It is a major constituent of many bitter in are two possible degradation products [75, 79].aloes and represents the active principle of these ca- Over one hundred papers deal with the isolation,thartic drugs [89]. It has been detected in the leaf jui- chromatography, quantitative determination and de-ce of more than twenty Aloe species (Liliaceae) [66]. gradation of aloin and some of these studies are dis-Homonataloin also is widely distributed in South Af- cussed in recent dissertations [75, 80, 81]. rican Aloe species [661. Of the more than one hun- 1 1-Deoxyaloin (= chrysaloin, Fig. 7), first obtai-dred species examined, about one-half contain atom ned by catalytic hydrogenation of aloin [821, was laterand homonataloin, roughly in equal proportions. isolated from plant material [83]. The structural elu-Aloesin (see above), appears almost exclusively in cidation and occurrence of a hydroxylated 0-methylthose species which contain aloin or homonataloin. derivative, homonataloin (Fig. 7), is described in [5]. The aloinosides are more restricted, occurring in five Recentlya simple new C-glycosylated anthrone,Aloe species [66]. Outside the Liliaceae, aloin has cassialoin, was isolated by HATA et a!. [84]. It is 10-been identified in only one family, Rhamnaceae hydroxy-10-C-3-D-glucopyranosyl-3-methyl-9(10H)-(Rhamnuspurshiana = Cascara sagrada) where it oc- anthracenone, a lO-hydroxylated li-deoxyaloin (Fig.curs together with cascarosides A, B, C and D and 7). All related compounds are 0-glycosylated aloins11-deoxyalomn ([83], confer [761). Only one instance or li-deoxyaloins which yield glucose or rhamnoseof cassialoin is known. Accompanied by 1 1-deoxya- upon acid hydrolysis. Aloinosid B is the 11-cL-L-rham-loin, it was isolated from the heartwood of Cassia noside of aloin [5] and aloinosid A presumably is thegarrettiana (Caesalpiniaceae) [84]. respective 10-C-glycosyl isomer. The 8-0-3-D-gluco- Thus, li-deoxyaloin, cassialoin and the cascarosi- sides of the diastereomeric aloins have been nameddes are very rare C-glycosy! anthrones which are uni- cascarosides A and B (Fig. 7), whereas the cascarosi-que to two species of the Rhamnaceae or Caesalpi- des C and D (Fig. 7) are analogs derived from the dia-niaceae, respectively, plant families known to con- stereomeric 1 1-deoxyaloins. The isolation and struc-tain anthraquinone O-glycosides. 138 Franz, Grün

Biosynthesis of Anthrone C-Glycosides borescens leaves had a pH optimum between pH 7.0 Biosynthesis of compounds containing the anthra-and 7.3. These experiments were the first to show in quinone and anthrone nucleus were the subject of se-vitro formation of a C-glycosidic bond thus demon- veral publications which were summarized in a recentstrating the similarity of C- and O-glycosylation [2, review on the biosynthesis of plant quinones by LEIST-21]. NER [90]. Anthraquinones may be synthesized via the In naturally occurring C-glycosides the bond be- acetate —polymalonateor the shikimic acid —0-suc-tween the sugar moiety and its aglycone always invol- cinyl benzoic acid —mevalonicacid pathway [90]. De-ves an aglyconic carbon atom of nucleophilic charac- rivatives with substituents in both ring A and ring Cter, such as the C-atom at the ortho- or para-position (Fig. 1) are usually acetate-derived in both higherof relevant phenolic hydroxyl groups [5]. The general and lower plants, while other anthraquinones, with-reaction mechanism of C-glycosylation may be the out substituents in ring A, are generally producedtransfer of an activated sugar form a nucleoside-di- by the shikimic acid pathway in higher plants. It hasphosphate sugar to the nucleophilic C-atom of the never been established which pathway is operativeappropriate aglycone. It follows that formation of 0- during the biosynthesis of C-glycosyl anthrones ofand C-glycosyl compounds in nature may utilize simi- the aloe emodine —orchrysophanol anthrone-type.lar metabolic processess as HAYNES [51 has already From their substitution pattern it appears that thepostulated. aglycones of those compounds are derived from eight acetate units [90]. Recently, GRUN and FRANZ studied the biosynthe-C-GlycosylatedGallic Acids sis of aloin in Aloe arborescens (Liliaceae) [81, 91, 92]. Feeding experiments with 14C02 and quantitati-Chemistryof C-Glycosylated Gallic Acids ve determinations of both diastereomeric aloins in Derivatives of 3,4,5-trihydroxybenzoic acid (gallic leaves of Aloe arborescens at different stages of deve-acid, Fig. 1) may be considered as the simplest natur- lopment demonstrated that only one isomer, aloin B,ally occurring C-glycosides with regard to the aglyco- is preferentially synthesized by the plant. The occur-ne moiety. In these compounds, a f3-D-glucosyl resi- rence of the respective isomer, aloin A (see above),due is C-linked to a hydroxylated phenylcarboxylic is explained by partial conversion of aloin B [92]. Itacid ortho to the carboxyl group. In addition, the car- was shown that the aloin B/aloin A equilibrium is re-boxyl group is esterified with the C-2 hydroxyl group ached non-enzymatically under physiological condi-of the glucosyl moiety to form a b-lactone ring. For- tions. mally these compounds are dihydroisocoumarin de- Biogenesis of aloins in Aloe arborescens seems torivatives [95]. be influenced by environmental conditions [811. An Bergenin, the most prominent representative of earlier study on variations in the content of alointhis class of compounds (Fig. 8), is a colourless sub- throughout a vegetation period in several Aloe spe-stance whose structure is confirmed by synthesis and cies gave similar results. MCCARTHY and VAN RHEEDEdegradation to be 2-C--o-glucopyranosyl-4-O-methylgallic acid b-lactone [96, 97]. It is synthesized VAN OUDTSH0ORN [93,94]found that the content of aloin in Aloe leaf juices varied considerably with afrom tetra-O-acetyl-a-n-glucopyranosyl bromide and 4-0-methylgallic acid [96]. 'H and 13C-NMR maximum production during the summer period sug- gesting that aloin might be metabolized. Feeding ex-spectral data for bergenin are given by RAMAIAH et al. periments with '4C-labeled compounds proved that[98]. These authors also describe three methylated leaf disks of Aloe arborescens preferentially incorpo-bergenin derivatives which are the only other examp- rated [2-14C]acetate into the aglycone moiety of alo-les in this group of C-glycosides (Fig. 8) yet known. ins while [U-'4C]glucose was incorporated into both the aglycone and sugar moiety [92]. Further studies CH2OR' with cell-free extracts demonstrated that crude enzy- 2 OR2 me extracts from Aloe arborescens were able to cata- lyze transglycosylation of [U-14C]glucose from UDP- [U-'4C]glucose to aloe emodine anthrone forming aloin [91]. C-Glycosylation took place only when H3CO9ZOH aloe emodine anthrone was added to the reaction mixture. [U-'4]Glucose containing substrates such as free glucose, glucose-i-phosphate, ADP-glucose, Bergenin R1:R2rR3 :H GDP-glucose and UDP-glucose were tested. Only 8,1O-di—O--methyl ether RR2:H, R3:methyl the latter was shown to be most effective for the iri—O—methyt ethers R1:R3methyt R2 :H transfer of the glucosyl residue to carbon 10 of aloe R2:R3methyI, R1 H emodine anthrone. The crude enzyme found in the 40.000 g supernatant from a homogenate of Aloe ar- Fig. 8. C-Gtycosyf Compounds in P'ants 139

Occurrence of C-Glycosylated Gallic Acids mechanistic schemes in which this compound is bio- Bergenin was isolated first in 1880 by GARREAU andgenetically rearranged to the C-glucosyl structure of MACHELART from rhizomes of three Bergenia sp. (= bergenin. However, attempts to convert the galloyl Saxifraga,Saxifragaceae) [99]. Later, it was found inester to a cmpound of the bergenin type were unsuc- several other families (confer [95], [98]), notably thecessful [106]. Hamamelidaceae and the Saxifragaceae which include An interesting study by MINAMIKAWA et al. [107] many bergenin containing sp. [95]. Of the sixteen Saxi-demonstrates the degradation of bergenin by Erwi- fragaceae sp. examined only those of the Saxifrago-nia herbicola, a strain of soil bacteria isolated from ideae, i.e., Astilbe, Bergenia, Saxifraga, Mitella,the rhizosphere of Bergenia crassifolia by elective Rodgersia, contain bergenin [100]. It is present in allculture with bergenin as a sole carbon source in the plant organs but the highest content usually occurs ingrowth medium. This organism degraded bergenin to the rhizomes. Bergenin methyl esters were isolatedyield 4-O-methylgallic acid among other minor pro- from the heartwood of Macaranga peltata (Euphorbia-ducts and apparently utilized the sugar moiety of the ceae) accompanaied by their parent compound [98].C-glycoside as a source of carbon and energy. Thus, this organism provides a rare example of biological Biosynthesis of C-Glycosylated Gallic Acids cleavage of C-glycosides. Biosynthetic pathways to substituted benzoic acids in plants or microorganisms involve side-chain degra-Acknowledgement dation of cinnamic acids, aromatization of dehydro- shikimic acid, hydroxylation and subsequent oxida- Theauthors wish to thank Prof. Dr. F. A. LoEwus, Pullman, tion of methylphenols as well as condensation of one Washington, USA for the assistance in editing the english manus- acetyl-CoA with three malonyl-CoA units. These cript and Frau K. WOLF for the skillful preparation of the figures. biosynthetic routes are summarized by GROSS m a review on phenolic acids [101]. Aromatization of dehy- droshikimic acid is probably the most important reac-References tion which leads to gallic acid in plants ([101], confer (1) Courtois,E.,F. Percheron in: R. E. Alston and V. C. Runeckles [102]). Pigman, W. and D. Horton (Eds.) (Eds.) Recent Advances in Phyto- With regard to bergenin KINDL lO3] reported in a The Carbohydrates, Vol. 2, pp. chemistry,Vol. 1, p. 305, New 213—240, New York, 1970, Acade- York1968, Appleton-Century- preliminary communication on 4C-tracer experi- micPress. Crofts. ments with leaves of Astilbe chinensis (Saxifraga- (2)Franz, G. in: Loewus, F. A. (14) Chopin, J. in: Wagner, H. and and W. Tanner (Eds.) Plant Carbo- L.Hörhammer (Eds.) Pharmaco- ceae). He introduced [7-'4C]benzoic acid, DL-[3- hydrates I: Intracellular Carbohy- gnosyand Phytochemistry,pp. 14Cjphenylalanine, [3-14C]cinnamic acid, [3-'4C]iso- drates, Encyclopedia of Plant Phy- 111—125,Heidelberg 1971, Sprin- siology, New Series, Vol. 13A, p. gerVerlag. ferulic acid and p-[3-'4CJ-coumaric acid into bergenin 384, Berlin, Heidelberg, New York (15)Horowitz, R. M. and B. Gen- and obtained the following results: hydroxylated cm- 1982,springer-Verlag. tilini:Chem. md. (London), 498 (3) Bandyukova, V. A. and V. A. namic acids and DL-phenylalanine were incorporated (1964). Yugin: Khim. Prir.Soedin, 5 (16) Wallace, J. W. and T. J. Mar- more efficiently than cinnamic acid and the latter in- (1981). bry:Phytochemistry 9,2133(1970). (4) Adinarayana, D. and J. Raja- corporated better than benzoic acid. Utilization of (17) Grisebach, H. and K. HahI- sekhara Rao: Indian J. Chem. brock:Recent Advances in Phyto- [1-'4C]acetate was negligible during short feeding pe- Educ. 4, 2(1974). riods. These findings led to the hypothesis that C-glu- (5) Haynes, L. J.: Adv. Carbo- chemistry8,21(1974). (18) Hahlbrock, K. in: Stumpf, P. cosylation occurs at the stage of a C6-C3 compound hydr. Chem. 18, 227 (1963). K.and E. E. Conn (Eds.) The Bio- (6) Haynes, L. J.:Adv. Carbo- rather than that of a C6-C1 compound. '4C-Labeled chemistry of Plants, Vol. 7, Secon- hydr.Chem. 20,357 (1965). gallic acid was not included in this set of experiments. (7) Chopin. J., M. L. Bouillant in: dary Plant Products, p. 425, New Harborne, 1. B.,T.J. York, London, Toronto, Sydney, On the other hand, TANEYAMA and YOSHIDA [104] Mabry and SanFrancisco 1981, Academic H. Mabry (Eds.) The Flavonoids, Press. incubated leaf disks of Saxifraga stolonifera (Saxifra- p. 866, London 1975, Chapman and gaceae) with D-[U-'4Cjglucose in the presence of un- Hall. (19)Inoue, T. and M. Fujita: Chem. Pharm. Bull. 22,1422 (8) Furukawa, M., I. Hayakawa, labeled gallic acid, 4-0-methylgallic acid and proto- G. Ohta and Y. Iitaka: Tetrahe- (1974). M. and T. Inoue: Ya- catechuic acid (=3,4-dihydroxybenzoicacid). The dron 31, 2989 (1975). (20)Fujita, Zasshi 99, 165 (1979). addition of gallic acid to the incubation medium en- (9) Horii, S., H. Fukase, E. Mizu- kugaku ta, K.Hatano and K. Mizuno: (21) HOsel, W.in:Stumpf,P. K. hanced significantly the incorporation of label fromChem. Pharm. Bull. 28, 3601 andE. F. Conn (Eds.) The Bioche- D-[U-'4C]glucose into bergenin. The other benzoic (1980). mistry of Plants, Vol. 7, Secondary Plant acid derivatives were ineffective. These results indi- (10) Richter,w. J.,K. 0.Alt, W. Products, p. 725, New York, Dieterle, J. W. Faigle, H.-P. London, Toronto, Sydney, San cate that gallic acid may be a likely glucosyl acceptor Kriemler, H. Mory and T. Winkler: Francisco 1981, Academic Press. in bergenin biosynthesis and that methylation may be Helv.Chim. Acta 58, 2512 (1975). (22) Wallace, J. W., T. J. Mabry andR. E. Alston: Phytochemistry a late step in the formation of the molecule. (11) Levy, S.,B. Yagen and R. Mechoulam:Science 200,1391 8, 93 (1969). Considering the fact that bergenin co-occurs with (1978). (23)Wallace, J. W.and H. Grise- the related 2-0-galloyl-arbutin (=gallicacid ester of (12) Abolin, C. R., T. N. Tozer, J. bach: Biochem. Biophys. Acta 304, C. Craig and L.D. Guenke: Scien- 837(1973). the hydroquinone -D-glucosid arbutin) in Bergenia ce 209,703 (1980). (24) Wallace, 1. W.: Phytochemi- sp. (Saxifragaceae) some authors [105, 106] offered (13)Alston, R. E. in: Mabry, T. J., stry 14, 1765 (1975). 140 Franz, Grün

(25) Inoue, T. and M. Fujita: (46) Glyzin, V. I., A. I. Ban'Kows- and K. Venkatamaran: Phytoche- London, Toronto, Sydney, San Chem. Pharm. Bull. 25, 3226 ku, M. G. Pimenov and K. I. Bory- mistry 12, 2033 (1973). Francisco 1981, Academic Press. (1977). dev: Khim. Prir. Soedin 9, 434 (70) Chiji, H., T. Aiba and M. Iza- (91) Grün, M. and 0. Franz: Plan- (26) Effertz, B. and G. Weissen- (1973). Wa: Agric. Biol. Chem. 42, 159 tamed. 152, 562 (1981). bock: Z. Pflanzenphysiol. 92, 319 (47) Ghosal, S. and R. K. Chaud- (1978). (92) Grün, M. and G. Franz: (1979). hun: Phytochemistry 12, 2035 (71) Mühlemann, H.: Pharm. Ac- Arch. Pharm. 315, 231 (1982). (27) Effertz, B. and G. Weissen- (1973). Ia Helv. 27, 17 (1952). (93) McCarthy, T. J. and M. C. B. bock: Phytochemistry 19, 1669 (48) Arisawa, M., N. Morita, Y. (72) Vogt, E. and H. Muhlemann: van Rheede van Oudtshorn: Planta (1980). Kondo and T. Takemoto: Chem. Pharm. Acta Helv. 46,657 (1971). med. 14,62(1966). (28) Margna, U. and T. Vainjärv: Pharm. Bull. 21,2562(1973). (73) Prox, A.: Tetrahedron 24, (94) McCarthy, T. J.: Plants med. Biochem. Physiol. Pflanzen 176,44 (49) Ghosal, S., D. K. Jaiswal and 3697(1968). 16, 348 (1968). (1981). K. Biswas: Phytochemistry 17, (74) Evans, F. J., M. G. Lee and (95) Barry, R. D.: Chem. Rev. 64, (29) Heinsbroek, R., J. van Brede- 2119(1978). D. E. Games: Biomedical Mass 229(1964). rode, G. van Nigtevecht and J. (50) Carpenter, I., H. D. Locksley Spectrometry 6, 374(1979). (96) Hay, J. E. and L. J. Haynes: J. Kamsteeg: Phytochemistry 18, 935 and F. Scheinmann: Phytochemi- (75) Widmaier, W.: Dissertation, Chem. Soc., 2231 (1958). (1979). stry8, 2013 (1969). Tubingen (1975). (97) Posternak, T. and K. Dürr: (30) Brederode van, J., J. Chopin, (51) Williams, C. A.: Phytochemi- (76) Wagner, H. and G. Demuth: Helv. Chim. Acts 41, 1159 (1958). J. Kamsteeg, G. van Nigtevecht stry 18, 803 (1979). Z. Naturforsch. 31b, 267 (1976). (98) Ramaiah, P. A., L. R. Row, and R. Heinsbroek: Phytochemi- (52) Saleh, N. A. M. and M. A. I. (77) Auterhoff, H., E. Graf, G. D. S. Reddy, A. S. R. Anjaneyulu, stry 18, 655 (1979). El-Anseri: Planta med. 28, 124 Eurisch and M. Alexa: Arch. R. S. Ward and A. Pelter: J. Chem. (31) Besson, F., A. Besset, M. L. (1975). Pharm. 313, 113(1980). Soc. Penn 1,2313(1979). Bouillant, J. Chopin, J. van Brede- (53) Antomi, M. and T. Kawasa- (78) Grün, M. and 0. Franz: Phar- (99) Garreau, Machelart: Compt. rode and G. van Nigetvecht: Phyto- ki: Chem. Pharm. Bull. 18, 224 mazie 34, 669 (1979). Rend. 91, 942 (1880) cited by He- chemistry 18, 657(1979). (1970). (79) Graf, E. and M. Alexa: Planta gnauer, R., Chemotaxonomie der (32) Heinsbroek, R., J. van Brede- (54) Birch, A. J., J. Baldas, J. R. med. 38, 121 (1980). Pflanzen, Vol. 6, Basal, Stuttgart rode, G. van Nigtevecht, J. Maas, Hlubucek, T. J. Simpson and P. W. (80) Eurisch, G.: Dissertation, Tu- 1973, Birkhãuser. J. Kamsteeg, E. Besson and J. Cho- Westermann: J. Chem. Soc. Perkin bingen (1979). (100) Taneyama, M. and S. Yoshi- pin: Phytochemistry 19, 1935 I, 898 (1976) and references cited (81) Grun, M.: Dissertation, Re- da: Bot. Mag. (Tokyo) 91, 109 (1980). therein. gensburg (1981). (1978). (33) Hostettmann, K. and H. (55) Floss, H. G. and A. Rettig: Z. (82) Owen, L. N.: Chem. md. (101) Gross, G. G. in: Stumpf, P. Wagner: Phytochemistry 16, 821 Naturforsch. 19b, 1103 (1964). (London), 37 (1956). K. and E. E. Conn (Eds.) The Bio- (1977). (56) Gupta, P. and J. R. Lewis: J. (83) Baumgartner R. and K. Leu- chemistry of Plants, Vol. 7, Secon- (34) Wiechowski, W.: Lotos 56,61 Chem. Soc. (C), 629 (1971). pin: Pharm. Acta Helv. 36, 445 dary Plant Products, p. 301, New (1908); cited in [33]. (57) Fujita, M. and T. Inoue: Te- (1961). York, London, Toronto, Sydney, (35) Ramanathan, J. D. and T. R. trahedron Letters, 4503 (1977). (84) Hata, K., M. Kozawa and K. San Francisco 1981, Academic Seshadri: Curr. Sci. (India) 29, 131 (58) Fujita, M. and T. Inoue: Baba: Chem. Pharm. Bull. 24, 1688 Press. (1960); cited in [6]. Chem. Pharm. Bull. 28, 2476 (1976). (102) Haslam, E. in: Stumpf, P. K. (36) Billet, D., J. Massicot, C. (1980). (85) Fairbairn, J. W., C. A. Fried- and E. F. Conn (Eds.) The Bioche- Mercier, D. Anker, A. Matsehen- (59) Swain, T. in: Harborne,J. B., man and S. Simic: J. Pharm. Phar- mistry of Plants, Vol. 7, Secondary ko, C. Mentzer, M. Chaigneau, G. T. J. Mabry and H. Mabry (Eds.), macol. Suppl. 15,292 T (1963). Plant Products, p. 527, New York, Valdener and H. Pacheco: Bull. The Flavonoids, p. 1097, London (86) Fairbairn, J. W. and S. Simic: London, Toronto, Sydney, San Soc. Chim. France, 3006 (1965). 1975, Chapman and Hall. J. Pharm. Pharmacol. 16, 450 Francisco 1981, Academic Press. (37) Haynes, L. J. and D. R. Tay- (60) Fujita, M. and T. Inoue: (1964). (103) Kind!, H.: Monatsh. Chem. lor: J. Chem. Soc. (C), 1685(1966). Chem. Pharm. Bull. 28, 2482 (87) Fairbairn, J. W., F. J. Evans 95, 1561 (1964). (38) Bhatia, V. K., J. D. Ramana- (1980). and J. D. Phillipson: J. Pharm. Sci. (104) Taneyama, M. and S. Yoshi- than and T. R. Seshadri: Tetrahe- (61) Auterhoff, H. and B. Ball: 66, 1300 (1977). da: Bot. Mag. (Tokyo) 92, 69 dron23, 1363 (1967). Arzneim. Forsch. 4, 725 (1954). (88) Wagner, H. and G. Demuth: (1979). (39) Nott, P. E. andi. C. Roberts: (62) Haynes, L. J., D. K. Holds- Z. Naturforsch. 29c, 444(1974). (105) Wenkert, E.: Chem. md. Phytochemistry 6, 741 (1967). worth and R. Russell: J. Chem. (89) Wade, A. (Ed.) Martindale: (London), 906 (1959). (40) Nott, P. E. andJ. C. Roberts: Soc. (C), 2581 (1970). The Extra Pharmacopoeia, 27the (106) Haslam, E. and M. Uddin: Phytochemistry 6, 1597 (1967). (63) Rattenberger, M.: Disserta- ed., p. 1334, London 1977, The Tetrahedron 24,4015 (1968). (41) Bhatia, V. K. and T. R. Ses- tion, München (1969). Pharmaceutical Press. (107) Minamikawa, T., S. Yoshi- hadri: Tetrahedron Letters, 1741 (64) Makino, K., A. Yags and I. (90) Leistner, F. in: Stumpf, P. K. da, M. Hasegawa, K. Komagata (1968). Nishioka: Chem. Pharm. Bull. 22, and E. E. Conn (Eds.) The Bioche- and K. Kato: Agr. Biol. Chem. 36, (42) Aritomi, M. and T. Kawasa- 1565 (1974). mistry of Plants, Vol. 7, Secondary 773(1972). ki: Chem. Pharm. Bull. 18, 2327 (65) McCarthy, T. J. and L. J. Plant Products, p. 403, New York, (1970). Haynes: Planta med. 15, 342 (43) Aritomi, M. and T. Kawasa- (1967). ki: Chem. Pharm. Bull. 16, 760 (66) McCarthy, T. J.: Plants med. (1968). 17, 1(1969). Address: Prof. Dr. G. Franz, (44) Markham, K. R. and 1. W. (67) Holdsworth, D. K.: Planta Lehrstuhlfür Pharmazeutische Biologie, Wallace: Phytochemistry 19, 415 med. 19, 322 (1971). Fachbereich Chernie und Pharmazie, (1980). (68) Harrison, P. G., B. K. Bailey (45) Smith, D. M. and 1. B. Har- and W. Steck: Can. J. Biochem. 49, Universität Regensburg, borne: Phytochemistry 10, 2117 964 (1971). Universitatsstraf3e 31, D-8400 Regensburg (1971). (69) Pendse, R., A. V. Rama Rao