Biosynthesis and Catabolism of Caffeine in Low-Caffeine-Containing Speci

8/20/2019 Biosynthesis and Catabolism of Caffeine in Low-Caffeine-Containing Speci

http://slidepdf.com/reader/full/biosynthesis-and-catabolism-of-caffeine-in-low-caffeine-containing-speci 1/7

Biosynthesis and Catabolism of Caffeine in Low-Caffeine-ContainingSpecies of Cof f ea

Hiroshi Ashiha ra † and Alan Crozier* ,§

Department of Biology, Faculty of Science, Ochanomizu University, Otsuka, Bunkyo-ku, Tokyo 112-8610,J apan, and Bower Building, Division of Biochemistry and Molecular Biology, Institute of Biomedical and

Life Sciences, U niversity of Gla sgow, G lasgow G 12 8QQ, Un ited Kingdom

Lea ves of Coffea salvatrix , Coffea eugenioides , a n d C. bengalensis conta in ∼3-7-fold lower levels ofcaffeine than those of Coffea arabica. There w a s more extensive biosynt hesis of caffeine from [8-14C ]-adenine in young leaves of C. a r a b i ca t h a n i n C. salvatrix, C. eugenioides , a n d C. bengalensis.

Degradation of [8-14C]ca ffeine, which is n egligible in lea ves of C. a r a b i ca, wa s also very slow in C.

sa l va tr i x a nd C. bengal ensis. In contra st, [8-14C]caffeine was cat abolized ra pidly by young and m at ureleaves of C. eugeni oides primarily by a caffeine f theophylline f 3-methylxanthine f x an t h in e f

uric acid f allantoin f alla ntoic acid f urea f C O2 + N H 3 pathw ay. These results indicat e tha tthe low caffeine accumulation in C. sa l va tr i x , C. eugenioides , a n d C. bengalensis is a consequenceof a slow rate of caffeine biosynthesis, whereas rapid degradation of caffeine also contributes to thelow endogenous caffeine pool in C . eugenioi des. The genes tha t r egulate caffeine accumulat ion a ppearto be t hose encoding N - methyltra nsferase a nd caffeine (7-N ) demethylase activities. The diversityof ca ffeine ca ta bolism observed in C. arabica , C. sa l va tr i x , C . eugeni oides , an d C. bengalensis , other

species of Coffea , a n d Ca m el l i a si n e n si s is discussed.

Keywords: Caff ein e; biosynthesis; cataboli sm; Coffea ar abica; Coffea salva tr ix; Coffea eugeni oides;

Coffea bengal ensis

INTRODUCTION

Co ffee , p rodu ce d p r imari ly in Ce n t ra l a n d So u t hAmerica a nd Africa , is one of the w orld’s most va lua bleagricultural products because of i ts widespread use asa beverage. Fruits of Coffea arabica a n d Coffea cane-

p h o r a a r e t h e r a w m a t e r ia l s f or t h e p rod u ct i on ofAra bica a nd Robusta coffee. These coffee beans conta inre la t iv e ly h ig h con ce n t ra t ion s of ca f fe in e , wh ich isperceived by some members of the general public ash av in g a dv ers e e f fe ct s on h e al t h (M azzafe ra e t a l . ,1991). As a consequence, there is an increasing demandfor decaffeinated coffee, which is produced by solventextraction or, more recently, supercritical fluid extrac-t io n wi t h CO 2. An alterna t ive source of decaffeinat edcoffee would be the use of Coffea species tha t conta inmuch lower levels of caffeine than C . a r a b i c a or C.

canephora. A number of such species exist , but theye it h e r p rodu ce few f ru it s or t h e bea n s y ield a p oorqua lity bevera ge and ar e, therefore, intrinsically unsuit-able for commercial development . Nor is a breedingprogram to tra nsfer the low-caffeine trait to C. a r a b i ca

a straightforward proposit ion because C. a r a b i ca a n d

C. canephora are polyploid, whereas other Coffea speciesare diploid. Under these circumsta nces, genetic engi-neering to produce transgenic caffeine-deficient C. a r a -

bica plants may be a more pract ical long-term propo-sition than a breeding program (Mazzafera et al., 1991;Crozier, 1997). The key genes in this regard are those

encoding the N -methyltransferases associated with caf-feine biosynthesis (see Figure 1) and the N -demethy-lases that catabolize caffeine. Caffeine-deficient trans-genic coffee could be pr oduced by expression of th e geneencoding the a ppropriat e demethylase or a lternat ivelythrough the antisense expression of an N -methyltrans-fe ras e g e n e . F o r t h is ap p ro ac h t o be t ak e n , de t a i le dinformation is first required on the enzymes and genes

contr olling key st eps in t he biosynthesis a nd/or cat abo-lism of ca ffeine.

There is much evidence tha t caffeine a nd other purin ealk alo ids are s y n t h e s ize d in C . a r a b i c a from purinen u cleot ides , p r imari ly v ia t h e p at h wa y i llu st ra t e d inFigure 1. Xanthosine is metabolized to 7-methylxan-thosine, which is the initial step in the caffeine biosyn-thesis pathway in which theobromine (3,7-dimethyl-xanthine) is the immediate precursor of caffeine (Suzukiet al . , 1992; Ashihara and Crozier, 1999). There is ar e p o r t o f a n a l t e r n a t i v e e n t r y i n t h e p a t h w a y t h a tinvolves d irect conversion of XMP to 7-methy l-XMP(Schulthess et al., 1996). Caffeine biosynthesis occursin both fruits a nd leaves of coffee. How ever, as they a reobta ined m ore easily , especially in non-coffee-producingcou n t r ies , mos t re se arch h as bee n ca rr ie d ou t wi t hleav e s . Ca f fe in e bios y n t h es is is e sp ecia l ly a c t ive inyoung leaves of C. a r a b i ca and declines as the leavesag e (F u jimori an d As h ihara , 1994; As h iha ra e t a l . ,1996a). Ca ffeine a ccumulat es in C . a r a b i ca d u e t oextremely slow cata bolism to t heophylline (1,7-dimeth-ylxan thin e) (Ashiha ra et a l., 1996b). To investigat e themechanisms that regulate the accumulat ion of purinealkaloids, the biosynthesis and catabolism of caffeinehave been compared in leaves of C. a r a b i ca a n d t h re e

* Aut hor t o w hom cor r espond ence should be ad d r essed[teleph one (-44) 141-330-4613; f a x (-44) 141-330-5394; e-ma [email protected]].

† Ochanomizu University.§ Gla sgow Univer sit y .

3425J. Agric. Food Chem. 1999, 47, 3425−3431

10.1021/jf981209n CCC: $18.00 © 1999 American Chemical SocietyPublished on Web 07/15/1999



Figure 1. C affeine biosynth esis and purine cat abolic path wa ys operat ing in leaves of C . arabica. Dotted arrows indicat e alternativepathway proposed by Schulthess et al. (1996). Enzymes: AMP deaminase (AMPD); IMP dehydrogenase; (IMPDH) 5′-nucleotidase(5′-NTD); (5) SAM:xanthosine N -7 methy ltra nsfera se (7-NMT); 7-methy lxan thosine n ucleosida se (MXN); SAM:7-meth ylxa nt hineN -3 methyl transferase (3-NMT); SAM:theobromine N -1 m ethyltra nsferase (1-NMT); xa nthosine n ucleosidase (NSD); xan thinedehydrogena se (XDH). In tea leaves, the 3-NMTa nd 1-NMT reactions are cat alyzed by a single enzyme called caffeine syntha se,which has been purified to homogeneity (Kato et al., 1999).

3426 J. Agric. Food Chem., Vol. 47, No. 8, 1999 Ashihara and Crozier



low-caffeine-containing species of coffee, C . sa l va tr i x , C .eugenioides , a n d C . bengalensis .

MATERIALS AND METHODS

Plant Material. S eed s of C . a r a b i ca , C . s a l v a t r i x , C.

eugenioides , and C. bengalensis were obta ined from th e Insti-tut o Agronom ico, C a mpina s, S a o P au lo, B ra zil. The leav esused in this st udy w ere from 2-year old plan ts growing undera nat ur al phot oper iod in a gr eenhouse at t he Univer sit y of

G l a s g ow . Y ou n g l ea v e s w e r e t h e m os t r ece n t ly e me rg edpr im ar y leaf ; m at ur e leaves com pr ised t he fully expand ed ,second an d th ird leaves below the a pex. Typical leaf size a ndfresh weight of each Coffea leaves were as follows: C. sal vat r i x

(young) 30 mm long, 12 mm wide, fresh weight (fw) ) 80 mg;(m at ur e) 170 m m long, 75 m m w id e, fw ) 1800 mg; C.

eugenioides (young) 32 mm long, 12 mm w ide, fw ) 50 mg;(mature) 90 mm long, 42 mm wide, fw ) 450 mg; C. benga-

lensis (young) 50 mm long, 17 mm w ide, fw ) 70 mg; (mature)180 mm long, 85 mm w ide, fw ) 1300 mg; C. arabi ca (young)20 mm long, 7 mm wide, fw ) 20 mg; (mat ure) 160 mm long,70 mm wide, fw ) 1200 mg.

Radiochemicals. [8-14C]Caffeine (specific activity ) 2.07MB q µmo l-1) a n d [ 8 -14C]adenine (1.96 MBq µmo l-1) w e r epurchas ed from Mora vek Biochemica ls Inc., Br ea, CA. [8-14C ]-Theophylline (2.04 MBq µmo l-1) was obtained from American

Radiolabeled Chemicals, I nc. , St . Louis, MO.Metabolism of Radiolabeled Purine Derivatives. Seg-ments of Coffea leaves (∼5 mm strips of young leaves; ∼5 m m× 5 mm for mat ure leaves) were incubat ed in 2 mL of medium,comprising 30 mM potassium phosphate buffer, pH 5.6, 10 mMsucr ose , and a r ad iolabeled subst r at e (37 kBq), in a 30 m LEr lenm eyer f lask, in a shaking w a t er bat h for 18 h at 27 °C .The Er lenm eyer f lask ha d a cent er w ell cont aining a sm allglass t ube into which wa s inserted a piece of filter paper wettedwit h 0.1 mL of a 20%pota ssium hyd roxide solution. In pulse-chase exper im ent s leaves w er e incubat ed w it h 14C-labeledsubstrate (74 kBq) for 4 h, at which point the radiolabel wasremoved and the incubation continued for a further 20 h.

At the end of the incubat ion period the gla ss tube a nd filterpaper from the center w ell were tran sferred to a 50 mL flaskco nt a i n i n g 1 0 m L of d is t i ll ed w a t e r a n d , a f t e r t h or ou g h

shaking, r ad ioact ivit y in a 1 m L a liquot w a s d et er m ined byliquid scintillat ion counting using ACS II scintillant (Amer-sham I nt er nat ional plc) t o d et er m ine t he am ount of 14C O2

released during t he meta bolism period. The leaf segments wereseparat ed from the incubat ion medium by filtering through at ea st r a iner , w a shed w it h 50 m L d ist i l led w a t er t hen m ixedwith 5 mL of extraction medium, comprising 80%methanolin 20 mM sodium diethyldithiocarbamate, and ground usinga pestle and morta r in ice. The homogena te w as centrifugedat 12000g for 5 min. The pellet wa s resuspended in 5 mL ofextraction medium and recentrifuged. The tw o superna ta nts,containing the methanol-soluble metabolites, were pooled andreduced to dryness in vacuo. The dry material was redissolvedin 0.5 m L of 50% m et hanol, a nd al iquot s w er e a nalyzed byTLC and HPLC.

Thin-Layer Chromatography of Radiolabeled Metabo-

lites. The meth an ol-soluble metabolites w ere an alyzed by TLCusing sheet s of m icr ocr yst al l ine cellulose a s d escr ibed inpr evious paper s (Ashihar a e t a l . , 1997; I t o and Ashihar a,1999). The solvent s yst em used w a s n -but a nol/a cetic acid/w a ter(4:1:2, v/v). The locat ion an d qua ntification of 14C-labeledcompounds on t he TLC sh eets w ere determined using a Bio-Ima ging Ana lyzer, Type FLA-2000 (Fuji P hoto Film C o., Ltd.,Tokyo, J apan).

High-Performance Li quid Chromatography of Endog-enous Purine Alkaloids and Radiolabeled Metabolites.HP LC w as carried out as described by Ashihara et a l. (1996b,1997). In th e present stu dy, a S hima dzu LC-10A HP LC syst emwa s used with a 250 × 4.6 mm i.d. ferruleless column (Ca pita lHP LC Lt d., Broxburn, U.K.) packed in house with a 5- µm ODSHypersil support (Sha ndon plc, Runcorn, Cheshire, U .K.). Them obile phase f low r at e w as 1 m L m in-1, and sam ples w er e

ana lyzed using a 25 m in gr a d ient of 0-40%methanol in 50m M sod ium acet at e , pH 5.0, w hich separ at es 11 d if fer entpurine derivat ives and is a lso able to resolve 14C-labeled uricacid, alla ntoin, an d alla ntoic acid (Ashihar a et a l., 1996a ). Theabsorbance at 270 nm a nd subsequent ra dioactivity detectionwere monitored using a Shima dzu SP D-10A UV-vis monitorand a R ayt est r ad ioanalyzer , M od el R am ona 2000 (R ayt estI sot openm essger at e Gm bh, S t r a ubenhar d t , Ger m any).

RESULTS

Endogenous Purine Alkaloids in Coffea L eaves.The levels of caffeine in C. sal va tr i x , C. eugenioides , a n d

C. bengalensis in y o un g a n d ma t u re le av es w e re 13-

36% of t hose found in C. a r a b i ca . In all four species,young leaves contained higher concentrations of caffeinetha n ma tur e leaves (Ta ble 1). Theobromine, w hich wa snot detected in extracts from C . sa l v a t r i x , C. eugenio-

ides , a n d C. bengalensis , w a s p r e s e n t i n y o u n g a n dmature leaves of C. arabica a t concentra tions of 1.8 an d0 . 1 m g g-1 (fw), r espectively. No other endogenousp ur i ne a l k a loi ds w e r e d e t ect e d i n a n y of t h e l ea fextracts.

Metabolism of [8-14C]Adenine. As plants readilyconvert adenine to AMP, isotopically labeled adeninecan be used to invest igate the AMP-derived caffeinebiosynthetic pathway (Suzuki et al., 1992) (see Figure1). Table 2 shows results of the overall metabolism of

[8-14C]adenine. There w a s litt le difference in the incor-porat ion of label into nucleot ides in the four Coffea

species, a lthough t he level of ra dioact ivity associatedwit h t h e o bro min e an d c a f fe in e was mu c h h ig h e r inyoung leaves of C. a r a b i ca t h a n i t w a s i n C. sa l va tr i x ,C. eugenioid es , an d C. bengal ensis. I n C. arabica 46.7%of the radioact ivity ta ken up by the leaf segments wa sincorpora ted into t he tw o purine alkaloids; the f iguresfor C. sa l va tr i x , C. eugenioides , an d C. bengal ensis were15.0, 0.4, and 0.4%, respectively. E xten sive la beling ofthe purine cat abolites, alla ntoin, allant oic acid, and C O2,wa s observed only in t he low-caffeine-conta ining Coff ea

p lan t s . I n C. eugenioides , More >30% of t otal ra dio-ac t iv i t y was re le as e d as 14C O2, wh e re a s t h e u re ide s ,a l l a n t oi n , a n d a l l a n t oi c a ci d w e r e t h e m os t h ea v i lylabeled compounds in C. bengalensis. Thus, in C . a r a -

bica , [8-14C]adenine is m etabolized preferentially to7-methylxanthine and converted to caffeine via theo-bromine, wh ereas in th e low-caffeine-conta ining Coffea

species, it a ppear s t o be converted prima rily to xa nthinean d e n t ers t h e p u rin e c a t abo lis m p at h w ay .

Catabolism of [8-14C]Caffeine.Da ta on the cat abo-lism of [8-14C]caffeine by young an d ma ture leaves fromC . s a l v a t r i x , C. eugeni oides , a n d C. bengalensis a represented in Table 3. Little or no catabolism occurredin leaves of C . s a l v a t r i x a n d C. bengalensis. S imila rresults were obta ined in a n earlier study with C. ar abica

leaves (Ashiha ra et a l., 1996b). In contr a st, very exten-sive catabolism of [8-14C]caffeine wa s observed in leaf

Table 1. Caffeine Content in Young and Mature CoffeeLeavesa

species young lea ves ma ture lea ves

C . arabica cv. K ent 7.1 ( 2.6 (100) 2.1 ( 0.34 (100)C . s al v a t r i x 0.92 ( 0 .0 8 (13 ) 0 .3 0( 0.01 (14)C . eugenioides 1.3 ( 0 .1 1 (19 ) 0 .3 7( 0.03 (18)C . bengalensis 1.2 ( 0 .0 1 (17 ) 0 .7 6( 0.01 (36)

a Data are expressed as mg g-1 of fresh w eight . Values are t heme a n ( S D (n ) 3). The figures in paren theses represent caffeinelevels expressed as a percentage of the amount detected in C .arabica .

Caffeine Metabolism in Low-Caffeine-Containing Coffea Species J. Agric. Food Chem., Vol. 47, No. 8, 1999 3427



segments from C . eugeni oid es. More than 75%of [8-14C]-caffeine taken up by the segments was catabolized inb o t h y o u n g a n d m a t u r e C. eugeni oides le a v e s , wi t hradioactivity recovered as theophylline, 3-methylxan-thine, 1-methylxanthine, xanthine, ureides, urea, andC O2 This indicat es tha t [8-14C]caffeine u ndergoes dem-ethylat ion, probably via the routes indicat ed in Figure2, re s u l t in g in t h e p ro du c t io n o f x an t h in e , wh ic h isdegraded to CO2 a n d N H 3 v ia the conventional purinecat a bol ic p at h wa y .

The ra te of upta ke of [8-14C]caffeine by leaf segment sof C. eugenioides wa s mu c h h ig h e r t h an t h o s e by t h esegments of C . s a l v a t r i x a n d C. bengalensis (Table 3).A s imilar t re n d w a s a ls o o bs erv ed wh e n [8-14C]theo-phylline w as used a s a precursor (see Table 5). Thereasons for this ar e unclear, a lthough it is possible thatpassive tran sport along a concentra t ion gradient , from

t h e in c u bat io n me diu m in t o t h e c e l ls o f t h e le a f , i se nh a n ce d i n k ee pi ng w i t h t h e e xt e nt t o w h i ch t h eabsorbed purine alkaloid is catabolized.

To obta in further informat ion on the pat hwa y utilizedfor the cat a bolism of caffeine in C. eugeni oides , pulse-chase experiments wit h [8-14C]ca ffeine w ere carried outusing mature leaves (Table 4). Caffeine, theophylline,and 3-methylxanthine were the most heavily labeledcompounds after the 4 h pulse. The radioact ivity as-sociated with caffeine declined after the leaves weretra nsferred to th e nonradioact ive medium. In contra st ,14C-labeled theophylline, 3-methylxanthine, 1-methyl-x an t h in e, x an t h in e, u re ides , u re a , an d CO 2 increasedaft er the 4 h chase, with >40%of the radioactivity takenup during the pulse being incorporated into 3-methyl-x an t h in e. Af t e r a fu r t h er 20 h ch as e , ra dioa c t ivi t yassociated with theophylline, ureides, and urea declined,wh e re as t h e le v e l o f 14C as s o cia t e d wi t h 3-me t h y l-x an t h in e, 1-me t h y lx an t h in e, an d x a n t h in e ch a n g e dl it t le a n d 14C O2 evolution increa sed fr om 8.3 to 24.2%of the recovered radioactivity.

Catabolism of [8-14C]Theophylline. C . a r a b i ca

leaves catabolize theophylline much more rapidly thancaffeine (Ashihara et al . , 1996b). I t wa s, t herefore, ofin t ere st t o in v e st ig a t e t h e fa t e of t h e op h y ll in e wh e nincubated with young an d ma ture leaves of C. sa l va tr i x ,C. eugenioid es , a n d C. bengalensis. The da ta obta inedare p res en t e d in Table 5 . [8-14C]Theophylline w a sconverted t o a ra nge of ca ta bolites, most extensively byC. eugeni oides wit h t he evolution of 1 4C O2 from ma turel ea v es b ei n g 3 t i m es g r ea t e r t h a n t h a t f r om y ou n gleaves, from which more tha n ha lf of ra dioact ivity w asrecovered as 3-methylxanthine. More label was associ-a t e d wi t h 3-me t h y lx an t h in e t h an 1-me t h y lx an t h in e ,indicat ing that the main route for catabolism of theo-

Table 2. Metabolism of [8-14C]Adenine by Young Coffee Leavesa

metabolite C. sal vat r i x C. eu gen i oi d es C. ben gal en si s C. ar abi ca

met ha nol-soluble 55.4 ( 1.3 37.4 ( 0.0 64.3 ( 1.3 63.2 ( 1.0nucleot ides 8.2 ( 1.6 20.4 ( 1.5 6.9 ( 0.4 4.4 ( 0.5 b

a denine 2.7 ( 0.1 1.6 ( 0.1 1.7 ( 0.1t heobr omine 12.3 ( 0.3 0.3 ( 0.1 0.3 ( 0.1 26.2 ( 3.2ca ffeine 2.7 ( 0.2 0.1 ( 0.0 0.1 ( 0.0 20.5 ( 3.5xa nt hine 7.1 ( 1.2 3.3 ( 0.1 3.0 ( 0.3 8.4 ( 1.7a lla nt oin 10.2 ( 1.2 18.5 ( 0.5 33.8 ( 1.4a lla nt oic a cid 10.2 ( 0.2 5.1 ( 0.4 6.6 ( 0.5

unknow n 2.1(

0.1 1.7(

0.4 4.6(

0.3C O2 17.6 ( 1.1 31.9 ( 2.3 5.4 ( 1.0 3.2 ( 0.2m et ha nol-insoluble 27.0 ( 0.2 30.8 ( 2.4 30.3 ( 2.3 35.0 ( 0.7t ot a l upt a ke (kB q) 83.4 ( 6.0 189.9 ( 16.6 155.0 ( 13.4 123.0 ( 3.0

a Segments of young leaves were incubated w ith 9.4 µM [8-14C]adenin e for 18 h. Values for C . arabica were taken from Ashihara et a l.(1996a). Incorporation of ra dioactivity int o meta bolites is expressed as percentage of t otal upt ake ( S D (n ) 3). Total u ptake of r adioactivityis expressed a s kBq g-1 of leaf (fresh weight ). b This value includes incorporation into nucleotides, allantoin, allantoic acid, adenine, andunidentified methanol-soluble compounds.

Table 3. Metabolism of [8-14C]Caffeine by Young and Mature Coffee Leavesa

distribution of recovered radioactivity (%of total ( S D )

species lea ves C f Tp 3-mX 1-mX X ureides urea C O2

total upta ke ofradioact ivity(kBq ( S D )

C. s al v a t r i x y ou ng 100 ( 0.0 ndb nd nd nd nd nd nd 25.3 ( 2.9m a t u r e 99 .7 ( 0.1 nd nd nd nd nd nd 0.3 ( 0.1 45.2 ( 0.1

C. eugenioid es y ou ng 23.6 ( 2.5 37. 6 ( 1.3 21. 6 ( 0.1 9.1 ( 0 .5 1 .8 ( 0. 2 0. 6 ( 0 .5 2 .0 ( 0.1 0.8 ( 0.0 91.5 ( 3.2m a t u r e 18 .5 ( 0.7 14. 3 ( 0.9 29. 7 ( 1.4 11. 6 ( 0 .3 4 .1 ( 0. 0 1. 5 ( 0 .1 4 .5 ( 0.1 13. 5 ( 1.8 80.9 ( 7.2

C. bengalensis y ou ng 100 ( 0.0 nd nd nd nd nd nd nd 49.1 ( 6.4m a t u r e 96 .5 ( 1.0 3.3 ( 1.0 nd nd nd nd nd 0.2 ( 0.1 42.8 ( 3.2

a Segments of young leaves were incubated w ith 8.9 µM [8-14C]caffeine for 18 h. Incorporation of radioactivity into metabolites is expressedas percenta ge of total upta ke ( S D (n ) 3). Total uptake of radioactivity is expressed as kBq g-1 of leaf (fw). Ca ffeine (Cf), theophylline(Tp), 3-methylxanthine (3-mX), 1-methylxanthine (1-mX), xanthine (X). b nd, not detected.

Table 4. Metabolism of [8-14C]Caffeine in a Pulse-ChaseExperiment with Mature Leaves of C . eugen io ides a

m et a bol it e 4 h (pu ls e) 8 h (ch a se) 24 h (ch a se)

r e si du a l ca f f ei n e 52 .0 ( 0.1 4.9 ( 0.1 3.4 ( 0.1t heophylline 21.4 ( 0.9 24.2 ( 0.6 8.1 ( 0.53 -m et h y lx a n t h in e 16 .0 ( 0.1 44.0 ( 0.2 46.9 ( 1.91-m et h y lx a nt h in e 6. 1( 0.6 6.6 ( 0.4 7.4 ( 0.0

xa nt hine ndb 3.9 ( 0.4 5.8 ( 0.2a lla nt oin 0.9 ( 0.1 18 ( 0.0 1.4 ( 1.1a lla nt oic a cid 1.1 ( 0.0 2.5 ( 0.0 ndurea 1.2 ( 0.1 3.5 ( 0.1 2.9 ( 0.1C O2 1.3 ( 0.2 8.3 ( 0.9 24.2 ( 2.0

a Leaf segments (100 mg of fresh w eight) were incubated w ith18 µM [8-14C]caffeine for 4 h (pulse), and then the incubationme d iu m wa s re pla c ed b y f re s h me d iu m w it hou t t ra c e r . Theradioact ivity wa s “chased” for a further 4 and 20 h. Incorporat ionof radioact ivity into each compound is expressed as a percentageof the total radioact ivity recovered. Mean values ( S D (n ) 3)shown. Total ra dioact ivity t aken up by the t issues wa s 31.2 ( 4.0k B q g-1 of leaf. b nd, not detected.




p h y ll in e t o x an t h in e is v ia 3-me t h y lx an t h in e in C.eugenioides. Meta bolism of [8-14C]theophylline w ass lowe r in C. bengalensis wi t h p rop ort ion al ly moreradioact ivity being recovered as unmetabolized theo-phylline and less a s 14C O2 a nd intermediates of purinecatabolism. There was relat ively l i t t le catabolism of[8-14C]theophylline by leaves of C. sa l va tr i x , wi t h >90%of the total radioact ivity taken up in leaves being theunmetabolized substra te a nd only minima l incorpora-tion of label into purine catabolites.

D I S C U S S I O N

The biosynthesis of caffeine in coffee has been inves-t ig a t e d p r imari ly in C. a r a b i ca (Looser et al., 1974;

Ba uma nn et al . , 1978; Roberts a nd Waller, 1979; B au-ma nn and Ga briel, 1984; S uzuki and Waller, 1984a,b;Negishi et al . , 1985a; Fujimori and Ashihara, 1994;Ashiha ra et a l., 1996a,b; Mosli Wa ldha user et al., 1997;Crozier et al., 1998; Ashihara and Crozier, 1999), andt h e ro u t e i l lu s t ra t e d in F ig u re 1 a ls o o pe ra t e s in t e a(Camellia sinensis ) (Suzuki and Takahashi, 1976; Negishiet al., 1985b; Ashihara and Kubota, 1986; Fujimori etal . , 1991; Negishi et al . , 1992; Ashihara et al . , 1995,1997; It o an d Ashiha ra , 1999) Although studied in lessd et a i l , t h e m a i n b ios y nt h et i c p a t h w a y f r om p ur i nenucleotides to t heobromine a nd/or caffeine a ppear s t obe similar, i f not identical, in other species of Coffea

(Mazzafera et al . , 1994) and C a m e l l i a i r r a w a d i e n s i s

Figure 2. Caf fe ine and pur ine cat a bolic pat hw a ys in Coffea p lant s . Bold ar r ow s ind icat e t he m a in pat hw ay via w hich caf feineis degraded in C. eugenioides. The conversion of caffeine to theophylline is effectively blocked in C. arabi ca a s w e l l a s i n C.

s a l v a t r i x a nd C. bengalensis in which the catabolism of theophylline is also slow. In C. arabi ca, theophylline is catabolized toxant hine, tra ces of which are converted to 7-methylxant hine, but most enters the purine cat abolism pathw ay an d is degraded toC O2 a n d N H 3. I n C. dewevrei , C. l i beri ca , a n d C. abeokutae , but not t he other species referred to a bove, caffeine is converted totheacrine, methylliberine, and liberine (Ba uman n et a l. , 1976; P etermann an d B auma nn, 1983). In t ea leaves, the conversion ofcaffeine to theophylline is blocked, whereas t heophylline is also demethylat ed to xan thine a s shown in Coffea p lant s , but t her eis a lso resynthesis of caffeine via a theophylline f 3-methylxant hine f theobromine f caf feine salvage pat hw ay (Ashihar a e tal., 1997). En zymes: N -1 demeth yla se (1-NDM), N -3 demethy las e (3-NDM), an d N -7 demethy las e (7-NDM). To dat e, there a re noreports on the prepara tion of cell-free extra cts conta ining t he N - demethylase a ctivities involved in the cata bolism of purine alkaloidsin coffee and t ea.




(As h ih ara an d Ku bot a , 1987) as we l l a s mat e (I l e x

p a r a g u a r i e n si s ) (Ashiha ra , 1993), cocoa tea (Ca me l l i a

p t i l o p h y l l a ) (As h ih a r a e t a l . , 1998), a n d g u a r a n a(P a u l l i n i a c u p an a ) (H. Ashihara an d T. W. B aum an n,unpublished da ta ).

In th e present study, th e biosynthesis and cat abolismof ca f fe in e we re in v es t ig a t e d in C. eugeni oides , C.

sa l va tr i x , a n d C. bengal ensis, three low-caffeine-con-t a in in g s pe cies of cof fee . Th e dat a f rom fee ds w i t h

[8-

14

C ]a d e n in e d em on s t r a t e t h a t , a s i n C . a r a b i ca

(Suzuki et al., 1992; Ashihara et al., 1996a; Crozier etal . , 1997), caffeine is synthesized from purine nucle-ot i de s i n y ou n g l ea v es of a l l t h r ee Coffea speciesHowever, the rate of caffeine biosynthesis is much lowertha n in C. arabica , especia lly in C. bengalensis , and thisis one reason for their low caffeine content.

The size of endogenous caffeine pools can also bere gu la t e d by t h e ra t e of ca f fein e c a t a bol ism, fu r t h ermethylation, and oxidation, and there is some diversityin the pathways operating in different species of coffee(Mazzafera et al., 1991, 1994). For example, leaves ofC. dewevrei , C . l i b er i c a , a n d C. a b eoku ta e convertcaffeine to thea crine (1,3,7,9-tetra methyluric acid),methylliberine [O (2)1,7,9-tet ra meth yluric a cid], liberine

[O (2)1,9-trim ethylur ic acid] an d oth er m ethyluric a cids(B a u m a n n et a l ., 1976; P e t er m a n n a n d B a u m a n n ,1983), wh ereas t he ma in route in C. arabica is caffeinef theophylline f 3-methylxanthine f x an t h in e f uricacid f allantoin f alla ntoic acid f f C O2 + NH 3 (seeFigur e 2). The rea ction velocity of the init ia l step in th ispathway, the conversion of caffeine to theophylline, isvery low and, as a consequence, caffeine accumulatesin C. arabi ca (Ashiha ra et a l., 1996b). The present st udyrevealed tha t a mong thr ee low-ca ffeine-conta ining Cof-

fea species, only C. eugeni oides possessed a strongcapacity for ca ffeine cat a bolism, with the demethyla tionof theophylline to xanthine proceeding mainly via 3-me-thylxanthine and to a lesser extent via 1-methylxan-thine (see Figure 2). The reduced caffeine content of C.

eugenioides is , thus, the r esult of a low ra te of caffeinebios y n t h e sis combin ed wit h a h ig h ra t e of ca f fein ecata bolism. In C. sa l va tr i x a nd C. bengalensis, cat abo-l is m o f bo t h c a f fe in e an d t h e o p h y l l in e is v e ry s lo w;therefore, the low ca ffeine content of th ese Coffea speciesappears to be due primarily to their reduced biosyn-thet ic a ct ivity .

7-Methylxa nth ine, produced via xa nth ine, is a minorcatabolite of [8-14C]theophylline in C. a r a b i ca leaves(Ashiha ra et a l., 1996b), wher eas in t ea a nd ma te lea vess mall amo u n t s o f t h e o p h y l l in e are s a lv ag e d fo r t h esynthesis of caffeine via a theophylline f 3-methyl-xanthine f theobromine f caffeine pathway (Figure 2)(Ashihara et al., 1997; Ito et al., 1997). In the present

s t u dy wi t h C. eugenioides , C . s a l v a t r i x , a n d C. benga-

lensis , [8-14C]t h e op h y ll in e wa s n e it h e r con v ert e d t o7-me t h y lx an t h in e n or s a lv ag e d for t h e s y n t h e sis ofcaffeine. I t ha s been r eported previously t hat 7-meth-ylxanthine is not involved in caffeine catabolism in C.

dewevrei (Mazza fera, 1993). U nlike C. dewevrei , an d C.

l iberica , a n d C. abeokut ae, there is no evidence that C.

arabica , C. eugenioides , C. sa l va tr i x , an d C. bengal ensis

convert caffeine to theacrine, methylliberine, and liberine.

The results of the present study are useful for thelong-term aims of using biotechnology to produce trans-genic, ca ffeine-deficient C. arabi ca plants. C. eugeni oides

degrades caffeine via theophylline and so, unlike theother species, contains a specific 7N -demethyla se a ctiv-ity. The substr a te specificity of this d emethyla se seemsto be different from tha t of the N -demethylase isolatedfrom bacteria, such as Pseudomonas putida a nd Pseudo-

monas cepacia (Ashihara an d Crozier, 1999), whichcatabolizes caffeine to theobromine rather than theo-phylline (Asa no et a l., 1994). Expression of th e bacter ia lde met h y las e in C . a r a b i ca i s u n l ik ely t o re su l t incaf fein e de f icien cy as ca f fein e w i ll be de grade d t ot h e obromin e, wh ich is t h e immedia t e p recu rs or ofca f fein e in cof fee . I n con t ras t , in s er t ion of t h e 7N -

demethylase encoding gene from C. eugenioides into thegenome of C. a r a b i ca is much more likely to producecaffeine deficiency because t he eugeni oid es gene productwill catalyze the conversion to caffeine to theophyllinean d t h e n a t iv e Arabica e n zy me s h av e t h e cap a c i t y t orapidly degrade theophylline. Work on the isolation andcharacterizat ion of this novel N -demethylase from C.

eugenioides is in progress.

ACKNOWLEDGMENT

This research was supported by the Royal Society,which provided H.A. a nd A.C. with J apa n-U .K. t ra v elgrants. Part of this work is also supported by a Grant-in-Aid for Scientif ic Resear ch from the Ministry of

Educat ion, S cience, Sports a nd Culture of J apa n (No.10640627) to H .A. We tha nk D r. Ma ro Sonda hl (Fit olinkCorp., Mount La urel, NJ ) for his help in obta ining seedsof the four species of coffee used in th is stud y from t heInst itut o Agronom ico, Ca mpina s, S.P ., Bra zil.

LITERATURE CITED

Asano, Y. ; Kom ed a, T. ; Yam ad a, H. Enzym es involved intheobromine production from caffeine by Pseudomonas

p u t i d a No. 352. Biosci., Bi otechnol., B iochem. 1994, 58 ,2303-2304.

As h ih a r a , H . P u r i n e m e t a b ol is m a n d t h e b io sy n t h es i s ofcaffeine in ma te leaves. Phytochemistry 1993, 33 , 1427-1430.

Table 5. Metabolism of [8-14C]Theophylline by Young and Mature Coffee Leavesa

distribution of recovered radioactivity (%total ( S D )

species lea ves Tp 3-m X 1-mX X Alln Alla urea C O2

total upta ke ofradioact ivity(kBq ( S D )

C. s al v a t r i x y ou ng 99.5 ( 0.0 ndb nd nd nd nd nd 0.5 ( 0.0 54.8 ( 12.2m a t u re 95. 2( 0.7 2.2 ( 0.2 nd 0.5 ( 0.1 nd nd 0.5 ( 0.0 1.0 ( 0.4 50.2 ( 5.2

C. eugenioid es y ou ng 26.6 ( 1 .0 5 3. 8 ( 1. 0 0. 8 ( 0. 1 1. 3( 0. 2 4. 8( 0. 7 1. 0 ( 0. 0 3. 3( 0.1 7.9 ( 2.9 86.3 ( 14.2m a t u re 36. 4( 0 .3 2 6. 2 ( 0. 7 0. 8 ( 0. 0 4. 8( 0. 3 0. 8( 0. 1 1. 3 ( 0. 0 3. 5( 0 .3 23 .5 ( 0.7 85.7 ( 1.5

C. bengalensis y ou ng 82.4 ( 1.0 6.9 ( 0. 0 2. 0 ( 0.4 nd 3.4 ( 0.2 nd nd 1.3 ( 0.4 41.7 ( 1.8

m a t u re 78. 2( 0 .3 1 4. 2 ( 0. 2 3. 1 ( 0.2 nd 3.7 ( 0.4 nd nd 0.9 ( 0.1 34.5 ( 3.6a Segments of young leaves were incubated w ith 9.1 µM [8-14C]theophylline for 18 h. Incorporation of radioactivity into metabolites is

expressed as percentage of total uptake ( S D (n ) 3). Total upt ake of ra dioactivity is expressed as kBq g-1 of lea f (fw ). Theophy lline (Tp),3-meth ylxan thin e (3-mX), 1-methy lxant hine (1-mX), xant hine (X), a llant oin (Alln), alla ntoic a cid (Alla). b nd, not detected.




Ashihar a, H. ; Cr ozier , C . Biosynt hesis and m et abolism ofcaffeine and related purine alkaloids in plants. In Ad vances

in Botanical Research ; Ca llow, J . R. , E d.; Academic P ress:London, U .K., 1999; Vol. 30, pp 117-205.

Ashihara , H.; Kubota , H. Pa tterns of adenine metabolism andcaffeine biosynt hesis in d if fer ent par t s of t ea seed lings.P hysi ol . P l ant . 1986, 68 , 275-281.

Ashihara, H.; Kubota, H. Biosynthesis of purine alkaloids inCamel l i a p lant s . P l ant Cel l P hysi ol . 1987, 28 , 535-539.

Ashihar a, H.; S himizu, H.; Takeda , Y.; Suzuki, T.; G illies, F.M.; Crozier, A. Ca ffeine meta bolism in high a nd low caffeine

containing cultivars of Camel l i a si nensi s. Z. Nat urf orsch.1995, 50 C , 602-607.

Ashihar a, H. ; M ont eir o, A. M . ; Gil l ies , F . M . ; Cr ozier , A.B iosynth esis of caffeine in leaves of coffee (Coffea arabica

L.). P l ant P hysi ol . 1996a, 11 1 , 747-753.Ashihar a, H. ; M ont eir o, A. M . ; M or it z , T. ; Gil l ies , F . M . ;

C r oz ie r, A. C a t a b ol is m of ca f f ei n e a n d r el a t e d p ur i nealka loids in lea ves of Coffea arabica L . P l a n t a 1996b, 1 98 ,334-339.

Ashiha ra , H.; G illies, F. M.; Crozier, A. Meta bolism of caffeinea n d r el a t e d p u r in e a l ka l oi d s i n l ea v e s o f t e a (C a m e l l i a

sinensis L.). P l ant Cel l P hysi ol . 1997, 38 , 413-419.Ashihar a, H .; Kat o, M.; Ye, C B iosynthesis and meta bolism of

purine alkaloids in leaves of cocoa tea (Camel l i a pt i l ophyl l a ).J. P l ant Res. 1998, 11 1 , 599-604.

Baum ann, T. W.; Gabr ie l , H. M et abolism and excr et ion of

caffeine during germination of Coffea ar abica L . P l ant Cel l Physiol. 1984, 25 , 1431-1436.

Baum ann, T. W.; Oechslin, M . ; Wanner , H. Caffe ine andmethylated uric acids: Chemical pa tterns during vegeta tivedevelopment of Coffea l iberica. Bi ochem. Physiol. Pfla nzen.

1976, 17 0 , 217-225.Baumann,T.W.;Dupont-Looser,E.;Wanner,H.7-Methyxantho-

sinesan intermediate in caffeine biosynthesis. Ph ytochem-

i s t r y 1978, 17 , 2075-2076.Crozier, A. Decaff. Biol. Sci. Rev. 1997, 10 (2), 40-41.Crozier, A.; B au ma nn, T. W.; Ashiha ra , H.; S uzuki, T.; Waller,

G . R. P at hwa ys involved in the biosynthesis and cata bolismof caffeine in Coffea a nd Camel l i a. P r oc. 16t h Col l . S ci . I nt .

Caf e, N ai r obi ; ASIC : P aris, 1998; pp 106-113.Fujimori, N.; Ashihara, H. Biosynthesis of theobromine and

caffeine in developing leaves of Coffea ar abica. Phyt ochem-

i s t r y 1994, 36 , 1359-1361.Fujimori, N.; Suzuki, T.; Ashiha ra , H. Sea sonal va riat ions in

biosynt het ic capacit y for t he synt hesis of caf feine in t ealeaves. Phytochemistry 1991, 30 , 2245-2248.

Ito, E. ; Ashihara, H. Contribution of purine nucleotide bio-synthesis de novo to the formation of caffeine in young tea(Camell ia sinensis) leaves. J . P l ant P hysi ol . 1999, 1 54 , 145-151.

Ito, E.; Crozier, A.; Ashihara, H. Theophylline metabolism inplant s . B i ochi m . B i ophys. A ct a 1997, 1336 , 323-330.

Ka to, M.; Mizuno, K.; Fujimura, T.; Iwa ma , M.; Irie, I.; Crozier,A.; Ashiha ra, H . Purification an d cha ra cterization of caffeine

synt hase fr om t ea leaves. P l ant P hysi ol . 1999, 12 0 , 579-586.

Looser, E. ; Ba uman n, T. W.; Wan ner, H . The biosynthesis ofcaffeine in the coffee plant . Phytochemi stry 1974, 1 3 , 2515-2518.

M azzafer a , P . 7-M et hylxant hine is not involved in caf fe inecatabolism in Coffea dewevrei. J. Agric. Food Chem. 1993,41 , 1541-1543.

Maz za fera, P .; Crozier, A.; Maga lha es, A. C. Ca ffeine meta bo-lism in Cof f ea arabi ca and other species of coffee. P hy-

tochemistry 1991, 30 , 3913-3916.M a z z a f er a , P . ; C r oz i er , A. ; S a n d be rg , G . S t u d ie s o n t h e

metabolic control of caffeine turnover in developing endo-sperms and leaves of Coffea arabica a nd Coffea dewevrei.

J. A gri c. Food Chem. 1994, 42 , 1423-1427.Mosli Waldha user, S. S. ; G illies, F. M.; Crozier, A.; B aum ann ,

T. W. S epar at ion of t he N -7 m et hylt r ansfer ase , t he keyenzyme in caffeine biosynthesis. Phytochemistry 1997, 45 ,1407-1414.

Negishi, O.; Ozawa, T.; Imagawa, H. The role of xanthosinein t he biosynthesis of caffeine in coffee plant s. A gri c. B i ol .

Chem. 1985a, 49 , 2221-2222.Negishi, O.; Ozaw a, T.; Imaga wa , H. Methyla tion of xanth osine

by tea-leaf extracts and caffeine biosynthesis. A gri c. B i ol .

Chem. 1985b, 49 , 887-890.Negishi, O.; Ozawa , T.; Imaga wa , H. B iosynthesis of caffeine

from purine nucleotides in tea plant. Biosci., Biotechnol.

Biochem. 1992, 56 , 499-503.Pet er m ann, J . B . ; Ba um ann, T. W. M et abolic r e lat ionships

bet w een m et hylxant hines a nd m et hylur ic ac ids in Coffea

L. P l ant P hysi ol . 1983, 73 , 961-964.Roberts, M. F .; Waller, G . R. N - Methyltransferase and 7-meth-

yl -N 9 - nucleoside hydrolase activity in Coffea arabica and t hebiosynthesis of caffeine. Phytochemi stry 1979, 1 8 , 451-455.

S c h u l t h e s s , B . H . ; M o r a t h , P . ; B a u m a n n , T . W . C a f f e i n ebiosynthesis starts with the metabolically channelled for-ma tion of 7-meth yl-XMPsA new h ypothesis. Phytochemistry

1996, 41 , 169-175.Suzuki, T.; Takahashi, E. Caffeine biosynthesis in Camel l i a

sinensis. Ph ytochemistr y 1976, 15 , 1235-1239.Suzuki, T.; Waller, G . R. B iosynthesis and biodegrada tion of

caffeine, t heobromine, an d theophylline in Coffea arabica

L. fruits. J. A gri c. Food Chem. 1984a, 32 , 845-848.Suzuki, T.; Waller, G . R. Biodegrada tion of caffeine: Format ion

of theophylline and caffeine in mature Coffea arabica fruits.J. S ci . Food A gri c. 1984b, 35 , 66-70.

S uzuki, T. ; Ashihar a , H. ; Waller , G. R . Pur ine and pur inealkaloid metabolism in Camel l i a a nd Coffea p lant s . P hy-

tochemistry 1992, 31 , 2575-2584.

Received for review November 5, 1998. Revised manuscriptreceived May 18, 1999. Accepted J une 2, 1999.

J F981209N


Biosynthesis and Catabolism of Caffeine in Low-Caffeine-Containing Speci

Documents

Transcript of Biosynthesis and Catabolism of Caffeine in Low-Caffeine-Containing Speci