Purification, Characterization, and Kinetic Analysis … Characterization, and Kinetic Analysis of a 2-Oxoglutarate-dependent Dioxygenase Involved in Vindoline Biosynthesis from Catharanthus

0 1993 by The American Society for Biochemistry and Molecular Biology, Inc THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 268, No. 8, Issue of March 15, pp. 5504-5511,1993

Printed in U. SA.

Purification, Characterization, and Kinetic Analysis of a 2-Oxoglutarate-dependent Dioxygenase Involved in Vindoline Biosynthesis from Catharanthus roseus*

(Received for publication, May 6, 1992)

Emidio De Carolis$ and Vincenzo De Lucaj From the Institut de Recherche en Biologie Vegetale, Departement de Sciences Biologiques, Universite de Montreal, Montreal, Q d b e c HI X 2B2, Canada

A 2-oxoglutarate-dependent dioxygenase (EC 1.14.1 1.1 1) which catalyzes the hydroxylation at position 4 of the indole alkaloid, desacetoxyvindoline has been purified to near homogeneity from Catharanthus roseus. The purification procedure combined conven- tional chromatographic methods and cosubstrate affinity chromatography on a-ketoglutarate-Sepharose. The specific activity of the 4-hydroxylase was enriched over 2000-fold compared to the crude homogenate with a recovery of 1.6%. The molecular mass of the native and denatured 4-hydroxylase was found to be 45 and 44.7 kDa, respectively, suggesting that the native enzyme is a monomer. Two-dimensional isoelectric focusing under denaturing conditions resolved the purified 4-hydroxylase into three charge isoforms of PI values 4.6,4.7, and 4.8. The enzyme did not require most divalent cations, but inactive enzyme was reac- tivated in a time-dependent manner by incubation with ferrous ions.

The mechanism of action of desacetoxyvindoline 4- hydroxylase was investigated. The results of substrate interaction kinetics and product inhibition studies suggest an Ordered Ter Ter mechanism where 2-oxoglutarate is the first substrate to bind followed by the binding of O2 and desacetoxyvindoline. The first product to be released was deacetylvindoline followed by COz and succinate, respectively.

Alkaloids, in general, may be defined as structurally com- plex organic bases which contain nitrogen and are formed in plants and fungi. In addition, alkaloids comprise a variety of important classes of drugs that have widespread experimental and therapeutic application. The Vinca alkaloids represent a class of natural drugs derived from the periwinkle plant, Catharanthus roseus. Two commercially important bis-indole alkaloids, vinblastine (VBL)’ and vincristine, are known to

* This work was supported by grants from the Natural Sciences

Formations De Chercheurs et l’Aide i La Recherche, Quebec (to V. and Engineering Research Council of Canada and Fonds Pour La

D. L.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ Supported by a Fonds Pour La.formations De Chercheurs et l’Aide i La Recherche doctoral scholarship.

J To whom correspondence should be addressed Institut de Re- cherche en Biologie VBgktale, Universiti de Montreal, 4101 Rue Sherbrooke Est, Montreal, Quebec HIX 2B2, Canada. Tel.: 514-872- 8492; Fax: 514-872-9406. ’ The abbreviations used are: VBL, vinblastine, a-Kg, cu-ketoglutaric acid, FPLC, fast protein liquid chromatography, PAGE, polyacrylamide gel electrophoresis, 2-D, two-dimensional.

accumulate in the aerial parts of C. roseus (1). Vinblastine is typically used to treat Hodgkin’s disease while vincristine is used in cases of acute leukemia (2). Since VBL and vincristine accumulate in low amounts, a considerable amount of research has been devoted to study the production of these alkaloids by cell and tissue culture methods (3). Unfortunately, this approach has not yet succeeded in producing bis-indole alkaloids since cell cultures are unable to synthesize vindoline, one of the monomeric precursors of VBL and vincristine (4).

To gain an understanding of why vindoline biosynthesis is repressed in cell cultures, studies have been performed on intact C. roseus plants with particular interest in the isolation of intermediates and enzymes involved in this pathway. Re- cent studies have indicated that key enzymes involved in vindoline biosynthesis are under strict developmental regulation (3). The biosynthesis of vindoline from tabersonine involves three hydroxylations, one 0-methylation, one N- methylation, and an O-acetylation (Fig. 1). Recently, a num- ber of these enzymes have been purified and characterized from C. roseus. The third to last step in vindoline biosynthesis is catalyzed by N-1-desmethyldesacetoxyvindoline (16-meth- oxy-2,3-dihydr0-3-hydroxytabersonine)~ N-methyltransfer- ase which has been partially purified ( 5 ) and is localized in the chloroplast (6). The last step in vindoline biosynthesis is catalyzed by deacetylvindoline 4-0-acetyltransferase which recently has been purified to homogeneity (7). The enzyme which catalyzes the second to last step in vindoline biosynthesis requires indole alkaloid substrate, 2-oxoglutarate, ascorbate, ferrous ions, and molecular oxygen for activity and thus is classified as a 2-oxoglutarate-dependent dioxygenase (8). The fact that hydroxylation at position 4 is critical for the enzymatic synthesis of vindoline and that desacetoxyvindoline 4-hydroxylase is absent in cell cultures prompted us to develop a protocol for the purification to homogeneity of desacetoxyvindoline 4-hydroxylase. Having developed a purification procedure which yields highly purified enzyme preparation as well as having a direct enzyme assay (8) which is simple, fast, and accurate we attempted to elucidate the kinetic mechanism of desacetoxyvindoline 4-hydroxylase.

This report describes the purification to near homogeneity and a detailed kinetic analysis of a 2-oxoglutarate-dependent dioxygenase involved in vindoline biosynthesis from C. roseus.

EXPERIMENTAL PROCEDURES

Plant Material The C. roseus plants used in this report were grown under green-

house conditions. Terminal buds and the first two pairs of fully expanded leaves were used for enzyme extraction.

The numbering system used is as for aspidospermidine in Chem- ical Abstracts.

5504

Desacetoxyvindoline 4-Hydroxylase from C. roseus

1 oc H

TRYPTOPHAN .OCl"

S T R I C T O S I D I N E

SECOLOGANIN

5505

T A 6 E R S O N l N E 1 6 - H Y O R O X Y T A B L R S O N I N L 1 6 - U f 1 t i O X Y 1 A 8 E A S O N I N ~ a, U l T t i T L A ~ l O ~ H T D R O X Y L A I I O N

ti JCO n,co

2-oxoglutarale

succinate + CO,

_____)

" C02CHs CH, C02CH5

DCSACETOXTVlNDOLlNf I 6 - ~ C 1 ~ ~ O X T - 2 . ~ - D l H i D R O - 3-HYOROXYTA8LRSONINL

I A C l T i C COA C o r

0 4 1

H3CO N on I CH) C O Z W cn, C O P J

D L A C f T T L V I N D O L I N L V l N D O L l N E

FIG. 1. Biosynthetic pathway from tryptophan to vindoline. Desacetoxyvindoline .i-hydroxylase, a 2-oxoglutarate-dependent dioxygenase catalyzes the second to last step in vindoline biosynthesis.

Chemicals S-Adenosyl-~-[methyl-'~C]methionine and [l-'4C]2-oxoglutaric

acid were purchased from Amersham (Mississauga, ON). Sephadex G-25, Sephadex G-100, AH-Sepharose 4B, Mono Q HR 5/5, Super- dex-75 HR 10/30, and the FPLC System were all purchased from Pharmacia (Montrkal, PQ). Green 19-agarose, l-ethyl-3(3-dimethy- laminopropy1)-carbodiimide, ferrous sulfate, and ascorbate were from Sigma. The Mini-Protean I1 dual slab cell, 2-D cell, SDS-PAGE, 2- D SDS-PAGE standards, hydroxylapatite, and the protein dye reagent were purchased from Bio-Rad (Mississauga, ON). Unlabeled 2- oxoglutarate and a-ketoglutaric acid were from Boehringer Mann- heim (Montrbal, PQ). All indole alkaloids used in this study were from our laboratory collection. These compounds were synthesized by previously described methods and their identity confirmed by spectroscopic techniques (9). All other chemicals were of analytical grade.

Buffers The following buffers were used 200 mM Tris-HC1, pH 7.5, con-

taining 10 mM dithiothreitol and 5 mM EDTA (A); 50 mM Tris-HC1, pH 7.5, containing 28 mM 2-mercaptoethanol (B); 10 mM sodium phosphate, pH 6.8, containing 28 mM 2-mercaptoethanol (C).

Desacetoxyuindoline 4-Hydroxylase Assay-The direct enzyme assay developed for the 2-oxoglutarate-dependent dioxygenase in C. roseus (8) contained 0.56 nmol of labeled indole alkaloid substrate (containing 44,600 disintegration/minute), 10 mM 2-oxoglutarate, 7.5 mM ascorbate, and up to 150 pg of protein in a final volume of 200 pl. The reaction was started by addition of the enzyme protein and was incubated at 30 "C for 15 min. The reaction was stopped by the addition of 100 pl of 1 M NaOH, and the aqueous phase was extracted for indole alkaloids with 500 pl of ethyl acetate. The two phases were separated by centrifugation, and the organic layer was recovered. The

organic layer which contained the 4-hydroxylated product as well as the unreacted substrate was evaporated to dryness. The residue was dissolved in 10 p1 of methanol and was separated by thin layer chromatography (silica) using 10% methanol in ethyl acetate as the solvent, in order to separate the product (RF 0.22) and substrate ( R F 0.52). After chromatography, the labeled product was isolated on the silica plate and removed. The amount of product formed was based on the radioactivity recovered from the silica plate. The radioactivity was determined by liquid scintillation counting in Optiphase Hisafe I1 (LKB) scintillation fluid.

The 4-hydroxylase was also assayed according to a previously described method (10). Briefly, the assay contained 9.2 p~ [1-I4C]2- oxoglutarate with a specific activity of 50 mCi/mmol, 5 p~ unlabeled indole alkaloid substrate, 7.5 mM ascorbic acid, 0.5 mg of catalase, and up to 0.3 mg of enzyme protein (50 mM Tris-HCI, pH 7.5, containing 28 mM 2-mercaptoethanol) in a final volume of 1 ml. The enzyme reaction was started by the addition of enzyme protein, and the mixture was incubated for 60 min at 30 "C. The reaction was stopped with the addition of 100 pl of 6 M HCI, and the formation of 2,3-dihydro-3,4-dihydroxy-N(1)-methyltabersonine was based on the decarboxylation of [l-"C]2-oxoglutarate leading to the formation of succinate and the liberation 14C02.

Variation in Oxygen Concentration

The Reacti- Vials containing active 4-hydroxylase were sealed with Teflon/silicone discs (22 mm) and placed on ice. The vials were then flushed with N2 for 30 min prior to the addition of known concentrations of the substrates which were further flushed with N2 for another 15 min. The appropriate 0, concentration introduced into each vial was based on the percent by volume in the incubation atmosphere. Once 0 2 was added to the reaction mixture, the vials were incubated in a water bath for 30 min at 30 "C. The reaction was stopped with the addition of 100 p1 of 1 M NaOH into the vials, and the 4-

5506 Desacetoxyvindoline 4-Hydroxylase from C. roseus

hydroxylated derivative was recovered as mentioned in the direct assay method. The K, for O2 is expressed as micromolar concentration on the basis of the known solubility of 0 2 at 30 "C. The accuracy of O2 concentration introduced into the incubation atmosphere was verified by taking a sample of the incubation atmosphere from con- trols and monitoring for O2 using a Thermal Conductivity Detector on a gas chromatograph (Hewlett-Packard 5890 I1 Chromatograph).

Preparation of a-Ketoglutarate-Sepharose The N,N"disubstituted carhodiimides promote the condensation

between a free amino and a free carboxyl group to form a peptide link. Thus AH-Sepharose 4B, which contains free amino groups was coupled randomly with the carboxyl groups of the ligand, a-ketoglutaric acid (11). Briefly, all operations in the carbodiimide coupling of a-ketoglutaric acid were performed at room temperature. Freeze- dried AH-Sepharose 4B (2 g) was allowed to swell in an excess of 0.5 M NaCl (200 ml/g) for a minimum of 15 min. The gel was washed with 0.5 M NaCl (400 ml) to remove additives originally present with the AH-Sepharose 4B powder. The NaCl was then removed by washing the gel with distilled water (adjusted to pH 4.5). The ligand, a-ketoglutaric acid (a-Kg), (170 mg) was dissolved in 5 ml of distilled water (adjusted pH 4.6) and added to the swelled gel. The suspension was gently stirred, and the pH was maintained at 4.5. Carbodiimide hydrochloride (1.4 g) was dissolved in 5 ml of distilled water (adjusted to pH 4.5) and added to the suspension over a 1-h period with continuous stirring at pH 4.5. The reaction was allowed to proceed for 24 h. The gel was washed with 400 ml of distilled water, pH 4.5, followed by equilibration with the appropriate buffer and packed into a column.

Protein Determination The protein was estimated according to the method of Bradford

(12) using Bio-Rad reagent dye concentrate (per manufacturer's instructions) and bovine serum albumin as standard protein.

Molecular Weight Determination The molecular weight of the different standard proteins was esti-

mated by gel filtration on Superdex-75 using an FPLC system. The column was equilibrated in 50 mM Tris-HC1, pH 7.5, containing 28 mM 2-mercaptoethanol and 200 mM NaCl. The reference proteins were bovine serum albumin ( M , 67,000), ovalbumin (M, 45,000), chymotrypsinogen A (Mr 25,0001, and ribonuclease (Mr 13,500). The void volume was measured by the elution of a sample of blue-dextran 2000. The 4-hydroxylase and sample proteins were applied individ- ually in a volume of 0.5 ml using Motorvalve MV-7 of the FPLC system, and the proteins were eluted at a flow-rate of 0.4 ml/min. The apparent molecular weight was estimated from its elution volume.

SDS-Polyacrylamide Gel Electrophoresis Protein preparations at different stages of purification were visu-

alized by gel electrophoresis under denaturing conditions according to the method of Laemmli (13) using 12% acrylamide gels. The molecular weight of the protein in question was determined from a graph of log molecular weight versus migration distance of standard proteins. The proteins were stained and visualized by Coomassie Blue or silver nitrate.

High Resolution Two-dimensional Polyacrylamide Gel Electrophoresis

A broad range of isoelectric focusing standards (Bio-Rad) which included conalbumin Type I (PI values = 6.0, 6.3, and 6.6), bovine serum albumin (PI values = 4.98, 5.07, and 5.18), bovine muscle albumin (PI values = 5.47, 5.5, and 5.53), rabbit muscle glyceralde- hyde-3-phosphate dehydrogenase (PI values = 8.3 and 8.5), bovine carbonic anhydrase (PI values = 5.9 and 6.0), soybean trypsin inhibitor (PI = 4.5), and equine myoglobin (PI = 7.0) were subjected to high resolution two-dimensional electrophoresis according to the method of O'Farrell(l4). The isoelectric points of the purified protein were determined from a graph of relative mobility ( R F ) versus PI values of standard proteins (Bio-Rad).

Purification of Desacetoxyvindoline 4-Hydroxylase Enzyme Extraction-Unless stated otherwise, all extraction and

purification procedures were carried out at 4 "C. The terminal buds

and the first two pairs of fully expanded leaves (-225 g) were mixed with polyvinylpolypyrrolidone (10% w/w) and homogenized in a Waring blender for 1 min at maximum speed containing ice-cold buffer A (1:3 w/v). The homogenate was filtered through cheese- cloth, and the filtrate was centrifuged at 20,000 X g for 15 min. The filtrate was fractionated with solid ammonium sulfate, and the protein which precipitated between 35 and 75% salt saturation was collected by centrifugation.

Chromatography on Sephadex G-1OO"The protein pellet was re- suspended in 30 ml of buffer B and chromatographed on a Sephadex G-100 column (2.6 X 91 cm), which had been previously equilibrated in the same buffer. The column was developed in buffer B, and 4-ml fractions were collected. The fractions were then assayed for 4- hydroxylase activity using the direct assay method.

Chromatography on Green 19-Agarose-The fractions exhibiting 4- hydroxylase activity from the G-100 were pooled and applied on a green 19-agarose column (5 X 5 cm), which had been previously equilibrated in buffer B. The column was washed with buffer B, and the bound 4-hydroxylase was eluted with a linear salt gradient (500 ml) of 0.0-1.0 M NaCl in buffer B. Four-ml fractions were collected and assayed for 4-hydroxylase activity using the direct assay method. The 4-hydroxylase peak activity eluted between 0.49 and 0.68 M NaCl concentration.

Chromatography on Hydroxylapatite-The green 19-agarose-purified fractions exhibiting 4-hydroxylase activity were chromatographed on a hydroxylapatite column (1.6 X 10.5 cm) which had been previously equilibrated in buffer C. The column was washed with buffer C, and the bound 4-hydroxylase was eluted with a linear gradient (200 ml) of 10-200 mM sodium phosphate in buffer C. Fractions of 2.5 ml were collected and assayed for 4-hydroxylase activity using the direct assay method. The active 4-hydroxylase eluted at 165 mM phosphate.

Chromatography on a-Ketoglutarate-Sepharose-Phosphate was removed from the fractions exhibiting 4-hydroxylase activity by chromatography on Sephadex G-25 (2.6 X 33 cm) which had been previously equilibrated in buffer B. The Sephadex G-25 column was developed in buffer B, and 3-ml fractions were collected and assayed for 4-hydroxylase activity. The eluted protein was applied to a- ketoglutarate-Sepharose (a-Kg-Sepharose) column (0.9 X 5.5 cm) which had been previously equilibrated in buffer B. The column was washed with 5 column volumes of buffer B, and the bound 4-hydroxylase was selectively eluted with a linear a-Kg gradient (50.0 ml) of 0.0-50.0 mM a-Kg in buffer B. Fractions of 1.5 ml were collected and assayed for 4-hydroxylase activity using the direct assay method. The 4-hydroxylase eluted at 29.5 mM a-Kg. Other bound proteins were nonselectively eluted from the a-Kg-Sepharose column by applying a linear salt gradient of 0.0-2.0 M NaCl in buffer B. Fractions of 1.5 ml were collected, but no 4-hydroxylase activity was detected in the salt gradient.

Chromatography on Mono Q-Fractions containing 4-hydroxylase activity which eluted from the a-Kg-Sepharose were pooled and chromatographed on a PD-10 column preequilibrated in buffer B. This preparation still contained minor contaminants, and the 4- hydroxylase was further purified by subjecting it to high performance ion-exchange chromatography on Mono Q HR 5/5 column. The protein was applied to the column which had been previously equilibrated in buffer B and after washing with 10 volumes of buffer B the 4-hydroxylase was eluted with a linear salt gradient (30 ml) of 0.0- 0.2 mM NaCl in buffer B, a t a flow-rate of 0.5 ml/min. One-ml fractions were collected and assayed for 4-hydroxylase activity using the direct assay method. The 4-hydroxylase activity eluted at 105 mM NaCl. At this stage of purification, the 4-hydroxylase was free from other contaminating proteins.

RESULTS

Purification of Desacetoxyvindoline 4-Hydroxylase-The desacetoxyvindoline 4-hydroxylase was purified from C. roseus by ammonium sulfate precipitation and successive chromatography on Sephadex G-100, green 19-agarose, hydroxylapatite, affinity chromatography on a-kg-Sepharose and on a Mono Q column using the FPLC system. A summary of the 4-hydroxylase purification protocol is shown in Table I. An average of 2,000-fold increase in specific activity could be obtained with a recovery of 1.6%. The final preparation had a specific activity of 86.15 picokatals/mg protein. The typical

Desacetoxyvindoline 4-Hydroxylase from C. roseus 5507

TABLE I Purification of desacetoxyvindoline 4-hydroxylase

Step Specific PurificationRecovery Drotein activitv mg Picokata1s"lmgPicokataLs -fold %

Crude 1955 0.043 83.5 100 Sephadex G-100 100 1.46 146.6 34 176 Green 19-agarose 4.3 15.85 67.4 371 81 Hydroxylapatite 1.4 24.57 34.3 575 41 n-KG-Sepharose 0.188 56.83 10.7 1331 12.8 Mono-Q 0.015 86.15 1.3 2018 1.6

One katal of desacetoxyvindoline 4-hydroxylase is defined as the amount of enzyme that catalyzes the conversion of one mole of substrate/s using the direct assay method.

KDa A B C D

9 6(

4!

3

2

1

FIG. 2. SDS-PAGE of fractions exhibiting 4-hydroxylase activity at different stages of purification. Lane A, green 19- agarose (-10 pg); lane B, hydroxyapatite (-10 pg); lane C, n-ketoglutarate-Sepharose (-3 pg); lane D, Mono Q (-2 pg). The molecular mass markers are indicated on the left in kDa.

Coomassie Blue-staining pattern obtained a t different stages of purification is illustrated in Fig. 2. The Mono Q-purified 4-hydroxylase migrated as a single band with an apparent molecular weight of 44,700 as observed on SDS-PAGE. To further demonstrate that the protein band observed on SDS- PAGE corresponded to the 4-hydroxylase, the Mono Q-purified enzyme was submitted to gel filtration on a Superdex-75 column. The fraction containing maximum 4-hydroxylase activity corresponded to a protein with a molecular weight of 45,000 suggesting that the native enzyme exists as a monomer.

High Resolution Two-dimensional Polyacrylamide Gel Elec- trophoresis-The Mono Q-purified 4-hydroxylase preparation was subjected to two-dimensional polyacrylamide electrophoresis according to the method of O'Farrell (14) using isoelectric focusing tube gels with a pH gradient of 3-10. Three protein charge isoforms were visualized after SDS-PAGE (12% acrylamide) using silver nitrate staining, and their cor- responding isoelectric points were determined from a standard curve. Two major protein charge isoforms were observed with PI values of 4.7 and 4.8 followed by a minor one with a PI of 4.6 (Fig. 3).

Requirements for Fe*+-During the purification of desacetoxyvindoline 4-hydroxylase, we found that its activity was rapidly lost after the dye-ligand affinity chromatography step. The addition of glycerol (up to lo%), a reagent known to stabilize enzymatic activity (15), did not maintain 4-hydroxylase activity. However, the enzyme was completely reacti- vated by preincubating the enzyme with all its cofactors (Table 11). The results also suggest that Fe'+ is primarily responsible for the reactivation of the 4-hydroxylase. The

PH - 4.5 5.5 6.5 z5

I I I

4

L

1

5 "

!1

4

FIG. 3. Two-dimensional electrophoresis of desacetoxyvindoline 4-hydroxylase. The 2-D electrophoretic pattern of the purified 4-hydroxylase (-2 pg) was carried out in the first dimension (3-10 pH range ampholites) and SDS-PAGE in the second dimension (15% acrylamide). The molecular mass markers are indicated on the left in kDa and the apparent isoelectric points on the top. The pI values of the respective 4-hydroxylase protein spots are referred to under "Results."

TABLE I1 Cofactor wquirements for 4-h.ydrox.ylase stahi1it.y

Incubation mixture",* Relative activitv

ControFd +ff-Kg +e-Kg and ascorbate +(?-Kg and FeY+ +@-Kg, ascorbate, and Fe" +Ascorbate +Ascorbate and Fez+ +Fe'+

1.0 1.5 1.3 3.6 4.0 1.3 3.2 3.8

" T h e control consists of green 19-agarose-purified 4-hydroxylase in buffer R which had a specific activity of 10.2 pkat/mg (relative activity = 1).

The concentration of n-kg, ascorbate, and Fe'+ were 100,250, and 8.5 p ~ , respectively.

Incubation mixture consisted of adding the reagents and incubating for 3 h a t 4 "C after which the mixture was assayed for 4- hydroxylase activity.

'The direct enzyme assay was used.

reactivation observed by the addition of Fez+ (8.5 K M ) alone (Table 11) is time-dependent, and maximal activity is obtained after incubation for 3 h a t 4 "C (data not shown).

The 4-hydroxylase did not require other divalent cations (Ca*+, MgL+, Mn'+, and Zn*+) for activity nor could they replace Fez+. In addition, the enzyme was not inhibited by the sulfhydryl group reagents iodoacetamide and N-ethylmaleim- ide when present a t concentrations of 0.1-10 mM.

Kinetic Analysis-The kinetic analysis was performed using purified 4-hydroxylase in Tris-HC1 containing 28 mM 2-mercaptoethanol, pH 7.5. The enzyme assays used in the kinetic studies were linear with time and enzyme concentration under all conditions tested. Furthermore, the direct enzyme assay established for desacetoxyvindoline 4-hydroxylase (8) was used to monitor the formation of the 4-hydroxylated derivative. In experiments where there was any doubt about the pattern in the reciprocal plot, it was repeated several times until a consistent pattern was established.

Initial Velocity Patterns-Detailed kinetic studies were carried out by varying the concentration of one substrate in the presence of different fixed concentrations of the second substrate, while the concentration of the third was held constant. Double-reciprocal plots with desacetoxyvindoline as the vari-

5508 Desacetoxyvindoline 4-Hydroxylase from C. roseus

able substrate at several fixed concentrations of a-ketoglutarate saturated with a constant O2 concentration (in air) gave an intersecting pattern (Fig. 44). When a-ketoglutarate was plotted as the variable substrate with 0 2 as the changing fixed substrate the same type of pattern was obtained (Fig. 4B). Similar intersecting lines were also obtained when O2 concentration (11, 40, and 55 p ~ ) was limiting at different fixed concentrations of desacetoxyvindoline (Fig. 4C). However, parallel lines were obtained when a-ketoglutarate was the variable substrate at fixed concentrations of desacetoxyvindoline with saturating O2 (960 pM) concentration (Fig. 40).

As calculated from the replots, the Vmax for the conversion of desacetoxyvindoline to deacetylvindoline was 3.85 picokatal/mg protein. The K,,, values for 2-oxoglutarate, 02, and desacetoxyvindoline were found to be 45.0,45.0, and 0.03 PM, respectively (Table 111). In addition, the K,,, values for ascorbate and Fez+ were found to be 0.2 mM and 8.5 p ~ , respectively.

Product Inhibition Patterns-The order of substrate binding and product release was further determined from the product inhibition studies. Succinate was found to be a competitive inhibitor with respect to 2-oxoglutarate (Fig. 5A) but

noncompetitive with respect to desacetoxyvindoline (data not shown). Deacetylvindoline was a noncompetitive inhibitor with respect to 2-oxoglutarate (Fig. 5B), O2 (Fig. 5C), and desacetoxyvindoline (Fig. 50). The Ki values for deacetylvindoline and succinate were 115 p M and 9 mM, respectively (Table IV) whereas the inhibitor concentration of C02 required for 50% inhibition [C02]~.5 was found to be 7.5 mM. The high levels of C02 required to inhibit activity made it impossible to perform further kinetic studies.

DISCUSSION

The desacetoxyvindoline 4-hydroxylase of C. roseus has been purified to near homogeneity using a five-step purification procedure. The protocol described here yielded an overall purification of 2000-fold with a recovery of 1.6%. The pure enzyme had a specific activity of 86.15 picokatal/mg protein. The key features of this purification protocol involved the utilization of green Wagarose and 2-oxoglutarate cosubstrate affinity chromatography on a-kg-Sepharose. The 4-hydroxylase bound tightly to green 19-agarose since salt concentrations below 0.45 M NaC1, a sufficient concentration to elute most proteins from the dye-ligand, did not elute the enzyme.

I O 20 30 4 0

[Z-oxoglularale]:’ (mMr‘ r

o b 10 20 30

[2-axoglularale];’ (mM)”

FIG. 4. Double-reciprocal plots for desacetoxyvindoline 4-hydroxylase catalyzing the conversion of desacetoxyvindoline to deacetylvindoline. The effect of desacetoxyvindoline concentrations on the rate of the 4-hydroxylase reaction at different fixed concentrations of 2-oxoglutarate ( A ) . The effect of 2-oxoglutarate concentrations on the rate of the 4-hydroxylase reaction at different fixed concentrations of 0, ( B ) . The effect of 0, concentrations on the rate of the 4-hydroxylase reaction at different fixed concentrations of desacetoxyvindoline. The 0, concentrations were limiting (11, 40, and 55 p ~ ) ( C ) . The effect of 2-oxoglutarate concentration on the rate of the 4-hydroxylase at different fixed concentrations of desacetoxyvindoline at saturating 0, concentration (960 pM) (D). Desacetoxyvindoline ( B ) and a-kg ( C ) concentrations were fixed at 0.25 and 100 p ~ , respectively. The concentrations of Fez+ and ascorbate were 10 pM and 0.25 mM, respectively, in all reaction mixtures and 0, concentration in air-saturated reaction mixtures was taken as 240 pM. p b t , picokatals.


TABLE I11 Kinetic parameters of substrate interaction kinetics for C. roseus 4-

hydroxylase The abbreviation used is: DAV, desacetoxyvindoline.

Varied Changing Constant Kinetic ligand fixed ligand ligand

Kinetic pattern parametef

P M

DAV a-kg 0 2 Converging K, 45.0 a-kg 0 2 DAV Converging KB 45.0 a-kg DAV 0 2 Converging KC 0.03 0 2 DAV a-kg Converging 02 (lirn? DAV a-kg Converging a-kg DAV O2 Parallel The replots from the substrate interaction kinetics were used to

The varying concentrations for 02 (limiting) were 11, 40, and 55 p M determine the numerical value of kinetic parameters.

while the O2 (saturating) concentration was 960 pM.

The separation on green 19-agarose resulted in a 371-fold purification, a 10-fold increase over the previous step with a typical recovery of 8045%. This was followed by the immo- bilization of the 4-hydroxylase to an a-kg-Sepharose affinity column and selective elution with a 2-oxoglutarate gradient. With this affinity step, a >1330-foldpurification was achieved and the 4-hydroxylase preparation contained several minor contaminants, as observed by SDS-PAGE. Based on the initial 4-hydroxylase activity of the crude preparation, the enzyme constitutes -0.05% of the soluble protein. Due to the low abundance of the enzyme in C. roseus, the establishment of an affinity chromatography step specific for the 4-hydroxylase was essential. A similar affinity chromatography strat- egy using coenzyme A as the affinity ligand has been used to purify to homogeneity the equally low abundant (-0.03%) 0- acetyltransferase which catalyzes the last step in vindoline biosynthesis from C. roseus (7, 16).

Other 2-oxoglutarate-dependent dioxygenases have been purified. The prolyl 4-hydroxylase from chick embryos has been purified by linking a peptide substrate with a high affinity for the enzyme to agarose and specifically eluting the enzyme with a second peptide substrate (17). The dioxygenases involved in scopolamine (18) and cephalosporin (19) biosynthesis were, in contrast, purified by a series of conven- tional column chromatography steps. The a-kg-Sepharose has been shown to be a good affinity ligand for the 2-oxoglutarate- dependent dioxygenase involved in vindoline biosynthesis from C. roseus. Chromatography of partially purified 4-hydroxylase preparations on a-kg-Sepharose yielded highly purified and active enzyme. The same support might be useful for the purification of other 2-oxoglutarate-dependent dioxygenases but this remains to be explored. However, a general problem encountered with this type of substrate affinity column is the gradual decline in binding affinity (20) which was observed after repeated use, and suggests that the ligand is being modified by the extract. The a-kg-Sepharose support was repeatedly used for the purification until >20% of the 4- hydroxylase activity appeared in the breakthrough volume after which new support was prepared.

We obtained evidence that the 44,700 Da band was the desacetoxyvindoline 4-hydroxylase when consecutive fractions across the peak of activity after a-kg-Sepharose affinity chromatography were applied to SDS-PAGE. The 44,700-Da protein band peaked in Coomassie staining intensity in the fraction containing peak enzyme activity (not shown). At the Mono Q chromatography step, the 44,700-Da band cofraction- ated with peak activity (Fig. 2). In addition, results obtained with gel filtration chromatography and SDS-PAGE suggested that the native enzyme exists as a monomeric protein of

44,700 Da. The molecular mass of the 4-hydroxylase coincides with other plant (21, 22) and bacterial (19) monomeric dioxygenases.

Another important factor responsible for the purification of the C. roseus 4-hydroxylase was the reactivation and sub- sequent stabilization of enzyme activity when incubating it with ferrous ions. Initially, we observed that the 4-hydroxylase activity was lost after several steps of purification. Previous studies in our laboratory have shown that Fez+ had little effect on the activity of the partially purified enzyme (8), whereas highly purified preparations showed an absolute requirement for Fez+. Ascorbate has been shown to reactivate pure prolyl 4-hydroxylase by reducing the enzyme-bound Fe3+ back to the Fez+ state (23). The addition of ascorbate to the purified C. roseus 4-hydroxylase did not reactivate the enzyme (Table 11) whereas addition of Fez+ increased the enzyme activity by 3.8- fold. These results suggest that unlike prolyl 4-hydroxylase, Fez+ has dissociated from desacetoxyvindoline 4-hydroxylase during purification. Furthermore, reactivation of the enzyme by Fez+ was time dependent, and maximal activity was recovered after 3 h of incubation.

The reaction mechanism of chick embryo prolyl4-hydroxylase (24, 25), lysyl hydroxylase (26), and thymine 7-hydroxylase (27) have been investigated, but no reports on plant dioxygenases have yet appeared. We report here the elucida- tion of the kinetic mechanism and kinetic parameters for desacetoxyvindoline 4-hydroxylase from C. roseus. The intersecting initial velocity patterns by all possible pairs involving a-ketoglutarate, 02, and desacetoxyvindoline indicate that the binding of these substrates to the 4-hydroxylase occurs by a sequential mechanism (28). The order of binding of O2 and desacetoxyvindoline was determined by studying the effect on the initial velocity plots when limiting and saturating with O2 concentrations. When O2 concentration was limiting and desacetoxyvindoline was the fixed substrate the initial velocity pattern gave intersecting lines (Fig. 4C). If the second substrate to bind is truly 02, then the initial velocity plot of a- ketoglutarate as the variable substrate a t changing fixed concentrations of desacetoxyvindoline and at saturating O2 should yield a parallel pattern as observed (Fig. 40) . This suggests that the normally reversible reaction sequence from E to (EABC) becomes irreversible when saturating with O2 and thus the a-ketoglutarate (A)-desacetoxyvindoline (C) initial velocity pattern becomes parallel (28). As a result, the order of binding of substrates would have a-ketoglutarate binding first, due to its competitive inhibition with succinate followed by the binding of O2 and desacetoxyvindoline. The first product released was deacetylvindoline due to the noncompetitive inhibition observed with a-ketoglutarate, 02, and desacetoxyvindoline. This suggestion would agree with a com- mon feature of ordered mechanisms whereby the product of the last substrate to combine to the enzyme is the first to be released (29). The third product to be released should give competitive inhibition with respect to a-ketoglutarate and noncompetitive with O2 and desacetoxyvindoline. Succinate was the only product which gave this set of patterns and consequently should be the third product to be released. The present results would suggest that deacetylvindoline, COS, and succinate are released in this order. The data discussed above are consistent with an Ordered Ter Ter mechanism (28) and are in agreement with the kinetic mechanism of prolyl4-hydroxylase (25), lysyl hydroxylase (26), and thymine 7-hydroxylase (27) indicating that this is a general feature of 2-oxoglutarate-dependent dioxygenases. However, more com- plete studies on product inhibition will be necessary to verify the place in the reaction mechanism for Fez+ and ascorbate,

5510

10 0

00

r

. E" 6 0 L

X m a Y

B . .- 4 0

2 0

Desacetoxyvindoline 4-Hydroxylase from C. roseus

10 i. 0 30 4.0

[2-oxoglutarate];' (rnM)"

C [Deacetylvindoline], (pM) 300

100

50

0

F- O! I I 1 I I I 1 1 1 I

0.4 0.8 1.2 1.6 2.0

[ O,] ,-I 70-1

2.5

2.0

r - r . .. 1.5

n m X - . > 7

1.0

0.5'2

B [Dtacetylvindoline], OtM)

+"----

300

75

50

0

0 . I 1 I I I I I I

1 .o 2.0 3.0 4 .O [2-oxoglutarate]:' (rnM)' 1

D [Deacetylvindoline], (pM)

100 300 500 700

[Desacetoxyvindoline],-' (pMr'

FIG. 5. Product inhibition of desacetoxyvindoline 4-hydroxylase catalyzing the conversion of desacetoxyvindoline to deacetylvindoline. Inhibition of the 4-hydroxylase reaction by succinic acid with respect to 2-oxoglutarate ( A ) . The inhibition of the 4-hydroxylase reaction by deacetylvindoline with respect to: 2-oxoglutarate ( B ) , O2 (C), and desacetoxyvindoline (D). The desacetoxyvindoline concentration (A-C) was fixed at 0.25 p ~ . The a-kg concentration ( C and D) was fixed at 100 p ~ . The concentrations of Fez' and ascorbate were 10 FM and 0.25 mM, respectively, while the 02 concentration in air-saturated reaction mixtures was taken as 240 p ~ . pkut, picokatals.

both considered not to be true substrates of 2-oxoglutarate- dependent dioxygenase reactions. Prolyl 4-hydroxylase studies have suggested that Fez+ is the first substrate to bind in thermodynamic equilibrium while ascorbate appears to bind before the other substrates or after the release of at least one product (24). We have observed that desacetoxyvindoline 4- hydroxylase becomes bound t o the a-kg-Sepharose affinity column in the presence or absence of Fez+ and ascorbate suggesting that 2-oxoglutarate is the first real substrate to bind to the enzyme.

The high affinity of the 4-hydroxylase for desacetoxyvin-

doline suggests that the metabolite is present in low concentration within the cell. A similar high affinity for plant me- tabolites has been observed by other Z-oxoglutarate-dependent dioxygenases involved in scopolamine (22), gibberellin (30), and flavanone (31) biosynthesis. Since the Kp for deacetylvindoline is over 3800-fold greater than the K,,, for desacetoxyvindoline, this indicates that the hydroxylation step is not inhibited by the high concentration of product formed. Similar lack of product inhibition has been observed for the enzyme catalyzing the last step in vindoline biosynthesis (7).

The availability of a purified desacetoxyvindoline 4-hydrox-


TABLE IV Kinetic parameters of product inhibition kinetics for C. roseus 4-hydroxyhe

Substrate Product Kinetic pattern parametef Kinetic

a-Ketoglutarate Deacetylvindoline Noncompetitive Kp 115 FM

Desacetoxyvindoline Deacetylvindoline Noncompetitive a-Ketoglutarate Succinic acid Competitive KR 9 mM 0 2 Succinic acid Noncompetitive Desacetoxyvindoline Succinic acid Noncompetitive

0 2 Deacetylvindoline Noncompetitive

The replots from product inhibition kinetics were used to determine the numerical value of kinetic parameter.

ylase has been used to develop the necessary tools for cloning this gene. The purified enzyme was submitted to tryptic digestion in order to obtain internal sequence information. A comparison of the peptide amino acid sequences obtained for the C. roseus 4-hydroxylase with that of hyoscyamine 6p- hydroxylase (32) have shown a relatively low but distinct region of h~mology.~ Oligonucleotide probes have been de- signed based on the peptide sequence information, and a cDNA library made from poly(A)+ RNA of 6-day-old devel- oping C. roseus seedlings (33) will be screened with these probes.

The production of vindoline from C. roseus tissue cultures remains a difficult task. The lack of vindoline biosynthesis in cell suspension cultures has been correlated with the lack of expression of the enzymes which catalyze the last three steps of this pathway. Studies at the molecular level will be useful to study the developmental, environmental, and tissue-specific regulation of 4-hydroxylase gene expression in C. roseus. Our ultimate goal is to understand the factors which regulate vindoline biosynthesis in C. roseus and consequently understand the factors which limit the production of the commercially important alkaloids, vinblastine, and vincristine.

Acknowledgments-We thank Dr. David Morse and Dr. Hargur- deep Saini for helpful discussions.

REFERENCES 1. Endo, T., Goodbody, A., and Misawa, M. (1988) Phytochemistry 27,2147-

2. Johnson, I. S., Wright, H. F., Svoboda, G. H., and Vlantis, J. (1960) Cancer 2149

Res. 20, 1016-1022

E. De Carolis and V. De Luca, unpublished results.

3. De Luca, V., and Kurz, W. G. W. (1988) Cell Cult. Somatic Cell Genet.

4. De Luca. V.. Balsevich. J.. and Kurz. W. G. W. (1985) J. Plant Phvsiol. Plants 4,385-401

121.417L428 . ,

Dethier, M., and De Luca, V. (1992) Phytochemistry, in press De Luca, V., and Cutler, A. J. (1987) Plant Physiol. 86,1099-1102 Power, R., Kurz, W. G. W., and De Luca, V. (1990) Arch. Biochem. Biophys.

De Carolis, E., Chan, F., Balsevich, J., and De Luca, V. (1990) Plant Physiol. 279,370-376

5. 6. 7.

8.

9. Balsevich, J., De Luca, V., and Kurz, W. G. W. (1986) Heterocycles 2 4 , 9 4 , 1323-1329

2415-2421 10. Hashimoto, T., and Yamada, Y. (1986) Plant Physiol. 81,619-625 11. Pharmacia LKB Biotechnology (1988) Affinity Chromatography, Principles

12. Bradford, M. M. (1976) Anal. Biochem. 7 2 , 248-254 13. Laemmli, U. K. (1970) Nature 227,680-685 14. O’Farrell, P. (1975) J. Biol. Chem. 260,4007-4021 15. Deutscher, M. P. (1990) Methods Enzymol. 182,83-89 16. Fahn, W., and Stockigt, J. (1990) Plant Cell Rep. 8,613-616

18. Yamada, Y., Okabe, S., and Hashimoto, T. (1990) Proc. Jpn. Acud. 6 6 , 17. Berg, R. A,, and Prockop, D. J. (1973) J. Biol. Chem. 248 , 1175-1182

20. Tonks, N. K., Diltz, C. D., and Fischer, E. H. (1988) J . BioZ. Chem. 2 6 3 , 19. Dotzlaf J. E., and Yeh, W. K. (1987) J. Bacterial. 169, 1611-1618

21. Lange, T., and Graebe, J. E. (1989) Planta 179 , 211-221 22. Hashimoto, T., and Yamada, Y. (1987) Eur. J. Biochem. 164 , 277-285 23. Myllyla, R., Kuutti-Savolainen, E. R., and Kivirikko, K. I. (1978) Biochern.

24. Myllyla, R., Tuderman, L., and Kivirikko, K. I. (1977) Eur. J. Biochem. Biophys. Res. Commun. 83,441-448

25. Tuderman, L., Myllyla, R., and Kivirikko, K. I. (1977) Eur. J. Biochem. 80,349-357

26. Puistola, U., Turpeenniemi-Hujanen, T. M., Myllyla, R., and Kivirikko, K. 80,341-348

27. Holme, E. (1975) Biochemistry 14,4999-5003 I. (1980) Biochim. Biophys. Acta 611,51-60

28. Cleland, W. W. (1967) in The Enzymes (Boyer, P. D., ed) Vol. 2, pp. 1-65,

29. Morrison, J. F., and Ebner, K. E. (1971) J. Biol. Chern. 246,3977-3984 30. Kwak, S. S., Kamiya, Y., Sakurai, A,, Takahashi, N., and Graebe, J. E.

31. Britsch, L., and Grisebach, H. (1986) Eur. J . Bwchem. 156,569-577 32. Matsuda, J., Okahe, S., Hashimoto, T., and Yamada, Y. (1991) J. Biol.

& Methods, pp. 3-45

Ser. B, 73-76

6722-6730

Academic Press, New York

(1988) Plant Cell Physiol. 29(6), 935-943

Chem. 266.9460-9464 33. De~Luca, V., Marineau, C., and Brisson, N. (1989) Proc. Natl. Acud. Sci.

~~~~

U. S. A. 8 6 , 2582-2586

Documents

Purification, Characterization, and Kinetic Analysis … Characterization, and Kinetic Analysis of a 2-Oxoglutarate-dependent Dioxygenase Involved in Vindoline Biosynthesis from Catharanthus