ARCHIVES

OF BIOCHEMISTRY

Vol. 282, No. 1, October,

AND

BIOPHYSICS

pp. 50-57,

1990

Flavanone-specific 7-0-Glucosyltransferase in Citrus pamdisi Seedlings: Purification and Characterization’ Cecilia

A. McIntosh,*

*Department Department

Received

March

Lilian

Latchinian,?

and Richard

Activity

L. Mansell*v2

of Biology, University of South Florida, Tampa, Florida 33620; and tPlant of Biology, Concordia University, Montreal, Quebec, Canada H3G lM8

16,1990,

and in revised

form

May

The flavonoids are one of the largest and most diverse groups of secondary metabolites found in plants. Their structure and function has been intensively investigated and their taxonomic distribution has been fairly well characterized (1, 2). In addition, the biosynthesis of the Cl5 flavonoid skeleton has been well established with subsequent modifications of the skeleton also being characterized (3). Citrus spp. are noted for their accumulation of flavanone glycosides in fruit and young vegetative tissues (4, in part should

by a Sigma be addressed.

Laboratory,

31,199O

The isolation and characterization of a IIavanonespecific 7-0-glucosyltransferase and its resolution from other glucosyltransferases in Citrus paradisi (grapefruit) seedlings is described. This new enzyme in the subclass 2.4.1 catalyzes the glucosylation of the 7OH group of naringenin (4’,5’,7-trihydroxyflavanone) to prunin and has been purified (943-fold) by fractional precipitation with ammonium sulfate and successive chromatography on Sephadex G-100, hydroxyapatite, UDP-glucuronic acid agarose, Mono Q, and Mono P columns. It has a pH optimum of 7.5-8.0, an apparent pI of 4.3, and an apparent M, of 54,900. This glucosyltransferase has an expressed specificity for the 7-position of the flavanones naringenin (Kmapp 62 pM; Kmapp UDPG 51 pM) and hesperetin (Km,,, 124 pM; KmaPP UDPG 243 pM) and did not accept other flavone or flavonol aglycones. Characteristics of other flavonoid glucosyltransferase activities found in grapefruit seedlings are also described. o ISSO Academic PITSS, IIN.

’ This work was supported Research awarded to C.A.M. * To whom correspondence

Biochemistry

Xi Grant-in-Aid

of

5). The glycoside substitution has an important effect on the organoleptic properties of 7-O-substituted flavanones. Compounds containing 7-0-P-rutinoside substitutions (rhamnosyl (~1-6 glucoside) are tasteless while those with 7-0-P-neohesperidoside substitutions (rhamnosy1 al-2 glucoside) are bitter (4). Naringin (naringenin-7-0-/3-neohesperidoside) is the major flavanone diglycoside found in grapefruit accounting for up to 40-70s of the dry weight of small green fruit (6). Studies of the enzymatic synthesis of naringin in grapefruit have concentrated on those enzymes involved in the earlier stages of flavonoid biosynthesis. Phenylalanine ammonia-lyase, chalcone isomerase, and chalcone cyclase have been isolated and partially characterized (7-9) while the activity of other enzymes, such as chalcone hydroxylase, has been demonstrated by studies of incorporation of radioactively labeled precursors (10,ll). The 7-0-glycosylation of naringenin has been demonstrated in biotransformation studies utilizing citrus liquid suspension cultures (12,13). Recently, both 7-O-glucosylation and 2”-0-rhamnosylation have been reported to occur in vitro in C. maxima and in C. mitis (14). We have recently reported on the isolation and partial characterization of a naringenin-7-0-glucosyltransferase activity in grapefruit (C. paradisi) seedlings (15). In this work, there was glucosylation of naringenin and hesperetin at the 7-position as well as the 4’-glucosylation of 4,2’,4’,6’-tetrahydroxychalcone (Fig. 1). Whereas these data indicated that the glucosylating activities of hesperetin and naringenin differed in thermal stability and pl values, the separation of these two enzyme activities could not be achieved. In addition, the flavonols quercetin and kaempferol were glucosylated although the position of the glucosylation was not determined. This paper describes the partial purification and charac-

50 All

Copyright cc) 1990 rights of reproduction

0003.9861/90 $3.00 by Academic Press, Inc. in any form reserved.

FLAVANONE-SPECIFIC

7-0-GLUCOSYLTRANSFERASE

51

The slurry was filtered through cheesecloth and the residue reextracted with another aliquot of buffer A (1:2, w/v) and filtered. The filtrates were centrifuged at 20,OOOg for 15 min, the supernatant was stirred 20 min with 10% (w/v) Dowex 1X2 which had been previously equilibrated with buffer A, then filtered. The filtrate, designated as the crude extract, was fractionated with solid ammonium sulfate and the protein fraction which precipitated between 40 and 70% salt saturation was collected by centrifugation and dissolved in a minimal volume of buffer B.

W/W),

R= OH, R’ I H. Narlngenln R= OMe, R’

q

OH, Hesperetin

FIG. 1. Structures droxychalcone.

of naringenin,

hesperetin,

and 2’,4’,6’,4-tetrahy-

terization of a flavanone-specific 7-O-glucosyltransferase and its resolution from other glucosyltransferases grapefruit seedlings. MATERIALS

AND

in

METHODS

Plant materials. Seeds of C. paradisi Macf. c.v. tained from the Citrus Budwood Registry (Winter were washed with running tap water for 3-4 days a soil-perlite mixture and grown under greenhouse light green leaves from 2- to 3-month-old seedlings used for enzyme purification.

‘Duncan’ were ohHaven, FL). Seeds then germinated in conditions. Young, were collected and

Chemicals and hochemicals. UDP-[U-‘4C]glucose (301 mCi/ mmol) was purchased from ICN (Irvine, CA). Unlabeled CJDPG,” UDP, BME, naringenin, hesperetin, hesperidin, UDP-GA agarose, and IJDP-agarose were purchased from Sigma (St. Louis, MO). Dyeligand columns (Blue A, Red A, Orange A, Green A, and Green B) were obtained from Amicon Corp. (Lexington, MA.). Standard prunin was obtained from Roth (via Atomergic Chemetals Corp. Farmingdale, NY). Standard Havonoid aglycones and glycosides were from our laboratory collection. Other Havonoid derivatives were synthesized as in Ref. (15). Sephadex G-100, Superose 12 (prep grade), Polybuffer-74, Mono P HR 5/20 and Mono Q HR 5/5 columns as well as the FPLC system (16) were from Pharmacia (Uppsala, Sweden). Prepared 12% acrylamide PhastGel media, buffer strips, silver stain kit, and the PhastSystem were also from Pharmacia. Dowex 1X2 and HA were obtained from Bio-Rad (Richmond, CA). All other chemicals and solvents were of analytical grade. Huffers. The following buffers were used: (A) 0.2 M Tris-HCI, pH T.-i, containing 42 mM BME, 5 mM EDTA, and 10 mM diethylammonium diethyldithiocarbamate; (B) 20 mM Tris+HCl, pH 7.5, containing 14 mM BME; (C) 10 mM potassium-phosphate, pH 6.8, containing 14 mM BME; (D) 25 mM his-Tris-iminodiacetic acid, pH 7.0, containing 14 mM BME; (E) Polybuffer 74.iminodiacetic acid (l:lO, v/v), pH 4.0, containing 14 mM BME; and (F) 50 mM Tris-HCl, pH 7.5, containing 14 mM BME. All buffers were millipore-filtered and degassed before use. Protein extraction. All steps were performed at 4°C unless stated otherwise. Young leaf tissue (ca. 30 g) was homogenized in a Waring blender with buffer A (13, w/v) and polyvinylpolypyrrolidone (lo%,

,’ Abbreviations: IJDPG, UDP-glucose; FPLC, fast protein liquid chromatography; GT, glucosyltransferase; BME, fi-mercaptoethanol; UDP-GA, UDP-glucuronic acid; SDS-PAGE, sodium dodecyl sulfate+polyacrylamide gel electrophoresis; HA, hydroxyapatite; RT, rhamnosyltransferase; Et, ethyl; AC, acetyl; BAW, n-butanol-acetic acid-water.

Enzyme purification. The ammonium sulfate-precipitated protein was applied to a Sephadex G-100 column (2.25 X 28.5 cm) previously equilibrated with buffer B, and eluted with the same buffer. Fractions with the highest specific activity were pooled (designated as the Gpool) and used for subsequent purification procedures. Binding of the GT to different dye ligand supports, affinity media, HA, Mono Q, and Mono P columns was tested utilizing the G-pool. Binding of the enzyme to the latter columns was tested using buffer B except for the HA and Mono P columns, where buffers C and D were used. On the basis of the results obtained with these columns, the following purifcation protocol was used. The G-pool was concentrated under nitrogen (Amicon PM-IO), then applied to a (HA) column (2.25 X 5.5 cm) previously equilibrated with buffer C, and washed with buffer C until Azxo was zero. The bound proteins were eluted using a linear gradient (10 to 200 mM) of potassium phosphate in buffer C (180 ml total volume). Three-milliliter fractions were collected and assayed for enzyme activity. The active fractions were pooled, dialyzed with buffer B, and concentrated. The HA pool was then applied to a 3.5ml UDP-GA agarose column previously equilibrated with buffer B. Glucosyltransferase activity was eluted with a linear salt gradient (0 to 500 mM KCI) in buffer B (80.ml total volume) and 2-ml fractions were collected and assayed for enzyme activity. The active fractions were pooled, dialyzed with buffer B, and concentrated before being applied to a Mono Q HR 5/5 column. The latter column was washed and the bound proteins were eluted with a linear gradient of O-500 mM KC1 in buffer B (30 ml total volume) at a flow rate of 0.5 ml/min; l-ml fractions were collected and tested for enzyme activity. Each peak of GT activity from the Mono Q column was dialyzed with buffer D, concentrated, and applied to a Mono P HR 5/20 column previously equilibrated with buffer D. The bound proteins were eluted with buffer E at a flow rate of 0.3 ml/min, generating a pH gradient between 7 and 4 which was measured using an on-line pH monitor. One-milliliter fractions were collected into tubes containing 0.1 ml of 1.0 M Tris-HCl buffer (pH 7.5, containing 14 mM BME) and assayed for enzyme activity. Glucosyltransferase assay and product identification. The standard assay mixture contained 50 nmol naringenin (or alternate aglycone) in 5 ~1 ethylene glycol monomethyl ether, 10 ~1 10X diluted UDP-[U14C lglucose (55 2000 dpm) and up to 50 wg of protein (in buffer B) in a total volume of 75 ~1. Both naringenin and hesperetin were routinely used as substrates in order to resolve the question concerning the properties of their potentially different glucosylating activities. The assay mixtures were incubated for up to 30 min at 30°C according to the particular experiment, reactions were stopped by the addition of 15 ~1 6 N HCl. The Havonoid glycosides were extracted with 250 ~1 ethyl acetate and an aliquot of the organic phase was counted in a LKB 1217 Rackbeta liquid scintillation counter using a toluene-based scintillation fluid. The flavanone and chalcone reaction products were identified by cochromatography with reference compounds on Polyamide-6 (Machery and Nagel) TLC plates using acetone-chloroform-water (80:20:4.8) (15), as well as on cellulose (Machery and Nagel) TLC plates using water as solvent system, and autoradiographed on X-ray film. Other reaction products were identified by cochromatography with reference compounds on cellulose TLC plates using BAW (3:l:l) as the solvent system and autoradiography. Where more than one reaction product was obtained, the incorporation of label into each

52

MCINTOSH,

IdO Elution

FIG. 2.

Elution profile chromatography on HA. described under Materials

rdo

260

Volume

(ml)

of GT activity (naringenin The elution of the column and Methods.

LATCHINIAN,

2iio

as substrate) was carried

after out as

compound was determined by cutting the spot out of the plate, placing it into scintillation fluid, and counting as previously described (15). Apparent Michaelis constants of the glucosyltransferase for the glucose donor were made by measuring the initial reaction velocities at different concentrations of the glucosyl donor while maintaining the concentration of the flavonoid aglycone acceptor at saturating levels. Apparent K,‘s for the aglycone acceptors were determined in the same manner using saturating levels of UDPG. Molecular weight determination. The concentrated G-pool was dialyzed with buffer F and applied to a Superose 12 HR lo/30 column which was previously calibrated with the following reference proteins: bovine serum albumin (M, 67,000), ovalbumin (M, 45,000), chymotrypsinogen (M, 25,000), and ribonuclease (M, 13,700). The column was eluted with buffer F at a flow rate of 0.5 ml/min and l.O-ml fractions were collected and assayed for glucosyltransferase activity. The apparent molecular weights of the enzymes were estimated from their elution volume from the column (17). Analytical procedures. Protein concentration was determined according to Bradford (18) using bovine serum albumin as standard protein. SDS-polyacrylamide minigel electrophoresis was performed according to Laemmli (19) and the gels were stained for proteins using Coomassie blue stain. SDSPAGE using the PhastSystem was conducted according to manufacturer’s instructions and the gels were stained using the Phast silver stain kit.

RESULTS

AND

AND

MANSELL

jority of the flavonol glucosyltransferase activity. The flavanone 7-0-glucosylating activity did bind to the HA column and eluted at 50 mM phosphate (Fig. 2). This resulted in a 17-fold increase in the specific activity of the enzyme. A flavone 7-0-glucosyltransferase isolated from parsley cell suspension cultures also bound to HA and eluted at 50 mM salt (20). The naringenin 7-0-glucosylating activity in the Gpool bound to Red A, Blue A, and Green A dye-ligand columns under the conditions used. These three ligands are considered to be more general affinity resins as up to 60% of the proteins present in a mixture can bind (21). There was no binding of this GT activity to the relatively more selective Blue B or Orange A columns under these same conditions. In addition, this enzyme activity did not bind to a UDP-agarose column. It did bind to a UDP-GA agarose column as have other UDP-glucosyltransferases (16). The binding of the GT to UDP-GA agarose potentially is a more specific interaction than those obtained with the dye columns, therefore the UDP-GA agarose column was incorporated into the purification protocol. Application of the G-pool to the UDP-GA agarose column and elution with a linear O500 mM KC1 gradient resulted in the elution of the GT’s in a broad peak. When the HA-pool was applied to the UDP-GA agarose column, the flavanone-glucosylating activities eluted in a sharper peak at around 140 mM KC1 (Fig. 3). This resulted in a 46-fold purification of the naringenin 7-0-glucosyltransferase activity as compared to the crude extract. The difference in the elution characteristics of the GT’s when including hydroxyapatite in

1.4

1.2

1.0

1 I

0.5

’ !5I

0.5

DISCUSSION

0.4

Enzyme Purification Binding of the naringenin 7-0-glucosyltransferase to several chromatographic media was tested in order to characterize the enzyme as well as to aid in the determination of a purification protocol. These columns included affinity, size exclusion, dye-ligand, and ion-exchange media. Much of the protein in the Sephadex G-100 pool did not bind to the HA column (Fig. 2). This included a ma-

0.2

60 Butlon

FIG. 3.

100 Volume

Elution profile of GT activity chromatography on UDP-GA agarose. carried out as described under Materials

160 Cm0

(naringenin as substrate) after The elution of the column was and Methods.

FLAVANONE-SPECIFIC

53

7-O-GLUCOSYLTRANSFERASE

Elution

Volume

hII

FIG. 4. Elution profile of GT activity after chromatography on Mono Q. Enzyme sample applied had previously been chromatographed on Sephadex G-100, HA, and UDP-GA agarose. Inset: Elution profile of GT activity after chromatography on Mono Q when HA has been omitted from the purification protocol. Elution was carried out as described under Materials and Methods. *Naringenin was not glucosylated by peak I, therefore hesperetin was used as the substrate for peak I and naringenin as the substrate for peaks II and III.

the purification procedure may be due to the loss of most of the flavonol GT activity which did not bind to the HA column under the applied conditions. Application of the pooled, dialyzed, and concentrated UDP-GA agarose active fractions onto a Mono Q ion exchange column and elution with a linear O-500 mM KC1 gradient resulted in three peaks of naringenin and/or hesperetin glucosylating activity (Fig. 4). The first peak, GT I, eluted at a salt concentration of 140 mM. When the HA column was omitted from the purification procedure, the relative amplitude of this GT I activity peak increased (Fig. 4, inset), indicating that most of the GT I activity does not bind to HA under the conditions used. The second peak, GT II, eluted from the Mono Q column at 230 mM KC1 and the third peak, GT III, eluted at 285 mM KCl. GT I showed highest glucosylating activity with flavonol substrates, GT II with flavanones and flavonols, and GT III with flavanones. There was a 182fold increase in specific activity of the naringenin 7-Oglucosylating activity (289-fold for the hesperetin 7-O GT) in GT III as compared to the crude extract. Chromatofocusing of the pooled, dialyzed, and concentrated UDP-GA agarose active fractions on a Mono P column resulted in three peaks of naringenin and/or hesperetin glucosylating activity (Fig. 5). These activities had apparent PI’S of 5.9,4.8, and 4.3. Application of the GT III peak to the mono P column resulted in a single peak of activity with a PI,,, of 4.3 and an increase

in specific activity of 943-fold for the naringenin 7-O glucosylating activity and 1476-fold for hesperetin 7-0glucosylating activity (Table I). Application of the GT II peak (from a Mono Q column) to the Mono P column resulted in a single peak at a PI,,, of 4.8. Substrate specificity tests of GT I and the 5.9 pl peak from the Mono P column indicated that these two activities corresponded. The purification of the naringenin 7-O- and hesperetin 7-0-glucosylating activities are summarized in Table I. Since earlier purification steps resulted in more than one EtOAc extractable glucoside (Fig. 6), purification

Elution

Volume

(ml)

FIG. 5. Elution profile of GT activity after chromatography on Mono P. Elution was carried out as described under Materials and Methods. *Naringenin was not glucosylated by peak I, therefore hesperetin was used as the substrate for peak I and naringenin as the substrate for peaks II and III.

54

MCINTOSH,

LATCHINIAN,

AND

TABLE

Purification

of Flavanone

MANSELL

I

7-0Glucosyltransferase

A Total protein (md

Procedure Crude homogenate 40-70s Saturated ammonium sulfate Sephadex G-100 Hydroxyapatite UDP-glucuronic acid agarose Mono Q (peak III) Mono P Note.

A, naringenin

Specific activity (pKatal/mg protein)

Total activity (pKata1)

B

Yield (%)

Purification factor

Purification factor

0.03

100

1.0

25.7

0.02

100

1.0

892 170 44

18.3 25.4 22.2

0.02 0.15 0.50

49 68 59

0.6 5.0 16.6

10.3 17.0 18.5

0.01 0.10 0.42

40 66 60

0.5 5.0 21.0

8 0.26 0.01

11.1 1.4 0.3

1.38 5.46 28.30

29 4 1

46.0 182.0 943.3

11.8 1.5 0.3

1.48 5.77 29.53

46 6 1

73.5 289.0 1476.0

as aglycone

substrate;

B, hesperetin

as aglycone

Due to the instability of the GT’s in Polybuffer-74 (even after adjusting the pH of the eluate), characterization of the GT’s was performed using the Mono Q-eluted fractions. GT I was able to glucosylate hesperetin but not naringenin (Table II, A). Chromatography and autoradiography of the hesperetin reaction products showed two products, a minor product that corresponded to hesperetin 7-0-glucoside and a major product that was not hesperetin 7-0-glucoside and did not cochromatograph with hesperidin (Fig. 6). This indicates that there may actually be two activities present in GT I. Using

-‘i

2

Yield (%)

37.7

Properties

1

(pKata1)

Specific activity (pKatal/mg protein)

1258

and yields were calculated on the basis of incorporation only into the 7-0-glucosides as in Ref. (15).

,t

Total activity

PRU

-

H7G

/’ ii

FI 3

-

4

5

6

7

FIG. 6. Autoradiograph of reaction products obtained at different stages of enzyme purification. Lanes: 1, Sephadex G-100 pool with naringenin as substrate; 2, Sephadex G-100 pool with hesperetin as substrate; 3, Mono Q peak I with hesperetin as substrate; 4, Mono Q peak II with naringenin as substrate; 5, Mono Q peak II with hesperetin as substrate; 6, Mono Q peak III with naringenin as substrate; and 7, Mono Q peak III with hesperetin as substrate. PRU, prunin; H7G, hesperetin 7-0-glucoside. Cellulose TLC plates, water as solvent.

substrate.

hesperetin as substrate, CT I gave a pH optima curve with two regions of maximum activity, one at pH 7-7.5 and another peak at pH 9. Analysis of the substrate specificity of GT I showed that the flavonol quercetin was the most active glucose acceptor followed by kaempferol with 81% relative activity. Hesperetin was glucosylated at 27% relative activity and the flavone apigenin at 10%. Naringenin, the flavone luteolin, and the flavonol rhamnetin were not glucosylated. The single product of the quercetin reaction cochromatographed with quercetin 30-glucoside, indicating that the primary activity present in this sample was a flavonol-3-0-GT. Studies on other flavonol GT enzymes have shown that they commonly have pH optima between 8 and 9 (22-24). It is interesting to note that kaempferol was glucosylated while rhamnetin was not glucosylated. This suggests that the presence of a free 7-OH group on the aglycone may be necessary for GT I to catalyze the 3-0-glucosylation of flavonols. GT II was capable of glucosylating naringenin and hesperetin with similar efficiencies (Table II, B). Chromatography and autoradiography of the reaction products showed two products with each substrate. One of the products of the hesperetin reaction corresponded to the 7-0-glucoside and the other to the major product obtained using GT I as the enzyme source (Fig. 6). The products of the naringenin reaction corresponded to the 7-0-glucoside (prunin) and another glucoside (Fig. 6). There was no correlation of the unidentified product from the naringenin reaction with either naringin or narirutin. GT II showed a pH optimum of 6.5 with 60% activity at pH 5.5 and 7.5. Substrate specificity studies showed that kaempferol and quercetin were also glucosylated as were rhamnetin, luteolin, apigenin, and 4,2’,4’,6’-tetrahydroxychalcone. Chromatography and

FLAVANONE-SPECIFIC

55

7-O-GLUCOSYLTRANSFERASE TABLE

II

Substrate Specificity and Properties of Flavanone 7-0-Glucosyltransferase Other Glucosyltransferases Found in Grapefruit Seedlings Percent Glucosyl~ transferase

A

R

C

D

Ih

IIb

IIIb

III”

A&cone

substrate

relative activity

Naringenin Hesperetin Luteolin Apigenin Kaempferol Quercetin Rhamnetin

0 27 0 10 81 100’ 0

Naringenin Hesperetin Luteolin Apigenin Kaempferol Quercetin Rhamnetin 2’,4’,6’,4-Tetrahydroxychalcone

94 89 57 18 100d 75 53

Naringenin Hesperetin Luteolin Apigenin Kaempferol Quercetin Rhamnetin 2’,4’,6’,4-Tetrahydroxychalcone

100’ 98 35 14 67 22 6

Naringenin Hesperetin Luteolin Apigenin Kaempferol Quercetin Rhamnetin

86 100’ 4 6 7 4 0

Aglycone K rnWP (ILM)

UDPG K maw hM)

nd

67

nd

nd

63 43 125

89 45 nd

32 17

61 200

62 124

51 243

and

Apparent Apparent PI

PH optimum

MW (kDa)”

‘i-7.5,9

5.9

49.5

6.5

4.8

49.5

7.5 7.5-8.0

4.3

54.9

18

1 4.2

’ Average of two determinations on a calibrated Superose 12 column. ’ Purification protocol: ammonium sulfate fractionation followed by chromatography ’ 100% = 1.33 pKatal/mg protein. d 100% = 2.12 pKatal/mg protein. ’ 100% = 2.49 pKatal/mg protein. ’ 100% = 29.6 pKatal/mg protein. ’ Purification protocol: ammonium sulfate fractionation followed by chromatography and Mono P.

autoradiography of the reaction products showed that there was one product from the quercetin reaction which cochromatographed with quercetin-3-0-glucoside. Two products were obtained from the luteolin reaction, one of which cochromatographed with luteolin-7-0-glucoside. There were two products from the kaempferol reaction and one product each from the apigenin and rhamnetin reactions and these have not as yet been identified. The chalcone glucoside could not be efficiently extracted

on Sephadex

on Sephadex

G-100,

G-100,

UDP-GA

HA,

agarose,

UDP-GA

and Mono

agarose,

Mono

Q.

Q,

with EtOAc under the conditions used, therefore these reactions were processed as in Ref. (15) and the activity reported here is that of the 4’-0-glucosylation. Labeled prunin was also obtained when using the chalcone as substrate, however, since it is not clear whether this represents glucosylation before or after ring closure (which can occur spontaneously) this data was not included in the calculation of the percentage relative activity of the chalcone.

56

MCINTOSH,

LATCHINIAN,

Apparent Km’s of GT II for the aglycone substrates as well as UDPG are also reported in Table II (B). The K mappof this enzyme for UDPG using quercetin as acceptor is three to five times greater than those obtained when using kaempferol, naringenin, or hesperetin as acceptors. This suggests either that quercetin does not bind the enzyme in a manner which favors binding of UDPG, or UDPG may not be the primary sugar donor when quercetin is the acceptor. This question awaits resolution. Thus, it appears that GT II contains a transferase with broad specificity for both the aglycone and the position of glucose attachment, or it contains multiple transferases with similar chromatographic properties on HA, UDP-GA agarose, Mono Q, and Mono P columns. GT III showed relatively greater specificity for the flavanone aglycones (Table II, C). Naringenin and hesperetin were glucosylated with approximately the same relative activity, and by far the major product was their respective 7-0-glucosides (Fig. 6). Other flavonoids were also glucosylated by GT III, although with lower relative activity than the flavanones. For example, kaempferol was glucosylated with 67% relative activity, yielding one reaction product. Luteolin was glucosylated at 35% relative activity. The product of the reaction using luteolin as the acceptor cochromatographed with luteolin-7-0-glucoside. Subsequent chromatography of GT III on a Mono P column and characterization of the substrate specificity of the single peak of activity which eluted from that column showed only flavanone glucosylation (Table II, D). Glucosylation of naringenin yielded a single product which cochromatographed with prunin, and glucosylation of hesperetin yielded only hesperetin 7-0-glucoside. This indicates that the glucosylation of the other flavonoids by GT III at the Mono Q stage of purification, as well as the activity producing the minor reaction products when naringenin and hesperetin were used as acceptors, was due to the presence of a trace amount of another GT, probably an activity from GT II. GT III had a pH optimum of 7.5 when using naringenin as acceptor, and a pH optimum of 7.5-8.0 when hesperetin was used as acceptor (Table II, D). The broad pH optimum of 6.5-7.5 obtained in earlier studies of the flavanone 7-0-glucosyltransferase in grapefruit seedlings (15) was probably due to the presence of both GT II (optimum 6.5) and GT III (optimum 7.5) in the sample tested. The p1 of GT III was 4.3, which is somewhat lower than most values reported for other flavonoid glucosyltransferases (22, 25, 26). The Kmapp of GT III for naringenin was 63 PM, which is similar to the value of 80 PM obtained with less pure enzyme (15). The Kmapp of GT III for UDPG (using naringenin as acceptor) was 51 PM. This value is lower than that reported in earlier work (15), probably due to the increased level of purification obtained in the present study. The apparent K,,,‘s of GT

AND

MANSELL

31.0

21.5 14.4 1

2

FIG. ‘7. SDS-PAGE PhastGel, pool; 2, Mono Q peak III; 3, Mono standard proteins (kilodaltons).

3

4

silver-stained: 1, UDP-GA Q peak II; 4, low molecular

agarose weight

III using hesperetin as acceptor were 124 PM for hesperetin and 243 PM for UDPG (Table II). The lower Kmapp’s for naringenin and UDPG, as compared to those obtained when using hesperetin as acceptor, indicate that the primary reaction of this GT may be the glucosylation of naringenin. Naringenin glycosides (e.g., naringin) account for the majority of flavonoid glycosides in grapefruit (4, 6). Investigation of the molecular weight of the flavanone 7-0-glucosyltransferase was conducted by applying the UDP-GA agarose active pool onto a calibrated Superose 12 column. The Superose 12 fractions were tested for GT activity using naringenin, hesperetin, and quercetin as acceptors and the ratio of glucosylation of these substrates was compared. Fractions with proportionately greater GT activity with a respective aglycone were applied to a Mono Q column and eluted as previously described. Results indicate that GT I and GT II have an average MW,,, of 49.5 kDa (Table II). GT III has an average MW,,, of 54.9 kDa (Table II). SDS-PAGE of CT III showed the presence of five bands, the middle band having a M, of 55 kDa (Fig. 7). The presence of a flavanone glucosyltransferase activity in citrus has been demonstrated (14,15) and partially characterized (15). We have described the isolation of a flavanone-specific GT, which catalyzes the production of 7-0-monoglucosides of naringenin and hesperetin, and the resolution of this GT from other GT’s present in grapefruit seedlings. The 7-0-glucosylation of naringenin and hesperetin appear to be catalyzed by the same enzyme activity or by activities with identical chromatographic properties on HA, UDP-GA agarose, Mono Q, and Mono P. It appears that differences in the naringenin and hesperetin glucosylating activities noted in an earlier study (15) were due to the presence of other GT activities in the sample which were also capable of glucosylating hesperetin at the 7 position (e.g., GT I and II).

FLAVANONE-SPECIFIC

While the natural occurrence of hesperetin 7-O-glucoside and prunin has not been demonstrated in grapefruit, these compounds have been isolated as one of the products found in the biotransformation of naringenin and hesperetin by citrus suspension cultures (12,13). In addition, in vitro rhamnosylation of prunin and hesperetin 7-0-glucoside has been demonstrated (14, 15) although these activities have not as yet been isolated and purified. This would seem to indicate that naringin biosynthesis in grapefruit may occur via a stepwise addition of the sugar moieties as has been described in other plants (23, 27, 28). Characterization of the rhamnosyltransferase (RT) involved in the biosynthesis of naringin and the investigation of the control of levels of expression of the GT and RT activities through growth and development of Citrus paradisi would be of interest for future studies of this system. ACKNOWLEDGMENT The authors are grateful to Dr. Ragai K. Ihrahim lahoratory where part of this work was conducted.

for the use of his

8. Hasegawa, 2487.

S., and Maier,

9. Raymond,

W. R., and

V. P. (1970) Maier,

Phytochemistty

P. (1977)

V.

9, 248%

of Plants (Corm, New York.

E. E., Ed.),

2. Harborne, .J. B. (Ed.) (1988) The Flavonoids-Advances search Since 1980. Chapman and Hall, New York.

in Re-

10. Fisher,

.J. F. (1968)

Phytochemistry

11. Hasegawa, 1370.

S., and Maier,

12. Lewinsohn, tochemistry

E., Berman, 25,2531&2535.

13. Lewinsohn, Plant Sci.

61,23-28.

14. Lewinsohn, Physiol. 9

1,1323%1328.

15. McIntosh, 15X-1538.

E., Mazur,

E., Berman, E., Britsch,

L., Khouri,

L., Mazur,

Y., and

R. L. (1990)

P. (1965)

18. Bradford,

M. (1976)

Riochem.

19. Laemmli,

U. K. (1970)

Anal.

J. (1986)

Gressel,

Y., and Gressel,

H. E., and Ibrahim,

17. Andrews,

11, 1365-

Phytochemistry

Y., and Gressel,

E., Mazur,

C. A. and Mansell,

16. Latchinian, 388,235-242.

7,769-771.

P. (1972)

V.

29, J. Chrom.

J. 96,595-606. Biochem.

Nature

72,248-254.

(London)

227,680-685.

20. Sutter, A., Ortmann, R., and Grisebach, phys. Acta 248, 71-87.

H. (1972)

Biochim.

Bio-

(Marois,

M.,

22. Jonsson, L. M. V., Aarsman, M. E. G., Bastiaannet, J., DonkerKoopman, W. E., Gerts, A. G. M., and Schram, A. W. (1984) 2. Naturforsch. 39,559-567. 23. Jourdan,

P. S., and Mansell,

R. L. (1982)

Arch.

Biochem.

Biophys.

213,434-443. 24. Kleinehollenhorst, G., Behrens, H., Pegels, G., Srunk, Wiermann, R. (1982) 2. Nuturforsch. 37,587-599.

5. Jourdan, P. S., McIntosh, 1vz.ysi01. 77, 904-908.

26. Bajaj, K. L., Deluca, 72,891-896.

V., and Ibrahim,

27. Barber,

Biochemistry

R. L. (1985)

Plant

Phytochemistry R. K. (1987)

Phy-

J. (1989)

.J. (1989)

3. Heller, W., and Forkmann, G. (1988) in The Flavonoids-Advances in Research Since 1980 (Harhorne, ,J. B., Ed.), pp. 400-425, Chapman and Hall, New York. 4. Horowitz, R. M., and Gentili, B. (1963) 7’etrahedron 19, 773-782. C. A., and Mansell,

16,

Phytochemistry

1535-1539.

21. Fulton, S. (1980) in DyeeLigand Chromatography Ed.), Amicon Corporation, Lexington, MA.

REFERENCES 1. Hosel, W. (1981) inThe Biochemistry Vol. 7, pp. 725-753, Academic Press,

57

7-0-GLUCOSYLTRANSFERASE

Plant

6. Kesterson, J. W., and Hendrickson, R. (1953) E’la. Agric. Exp. Stn. Tech. Rull. 5 11 . 7. Maier, V. P., and Hasegawa, S. (1970) Phytochemistry 9,139-144.

25. Koster, 104.

J., and Bars,

G. A. (1962)

W. (1981)

28. Kamsteeg, J., Van Brederode, Phytochemistry 18,659p660.

Arch.

Biochem.

Biophys.

R. K. (1983)

Plant

N.,

and

212,98I’hysiol.

1,463-468. J., and Van

Nigtevecht,

G. (1979)