molecular cloning of a taxa-4(20),11(12)-dien-5α-ol-o-acetyl transferase cdna from taxus and...

E

whT(Gctqdgpc

crmardhmtwzoosf

Archives of Biochemistry and BiophysicsVol. 374, No. 2, February 15, pp. 371–380, 2000doi:10.1006/abbi.1999.1609, available online at http://www.idealibrary.com on

Molecular Cloning of a Taxa-4(20),11(12)-dien-5a-ol-O-Acetyl Transferase cDNA from Taxus and Functional

xpression in Escherichia coli1

Kevin Walker, Anne Schoendorf, and Rodney Croteau2

Institute of Biological Chemistry, Washington State University, Pullman, Washington 99164-6340

Received September 16, 1999

dtt

caTo

The taxa-4(20),11(12)-dien-5a-ol-O-acetyl transferasehich catalyzes the third step of Taxol biosynthesisas been isolated from methyl jasmonate-inducedaxus cells, and partially purified and characterized

K. Walker, R. E. B. Ketchum, M. Hezari, D. Gatfield, M.olenowski, A. Barthol, and R. Croteau, Arch. Bio-

hem. Biophys. 364, 273–279 1999). A revised purifica-ion method allowed internal amino acid microse-uencing of the enzyme, from which primers wereesigned and employed to amplify a transacetylaseene-specific fragment. This radiolabeled, 900-bp am-licon was used as a hybridization probe to screen aDNA library constructed from poly(A)1 RNA isolated

from induced Taxus cells, from which a full-lengthtransacetylase sequence was obtained. Expression ofthis clone from pCWori1 in Escherichia coli JM109ells yielded the functional enzyme, as determined byadiochemical assay and combined capillary gas chro-atographic-mass spectrometric verification of the

cetylated product. The full-length DNA has an open-eading frame of 1317 nucleotides corresponding to aeduced amino acid sequence of 439 residues that ex-ibits high sequence identity to the proteolytic frag-ents of the native enzyme, which the recombinant

ransacetylase resembles in properties. Consistentith the size of the operationally soluble native en-

yme, the DNA appears to encode a monomeric proteinf molecular weight 49,079 that bears no N-terminalrganellar targeting information. Sequence compari-on of the taxadien-5a-ol-O-acetyl transferase with theew other known acyl transferases of plant origin in-

1 This investigation was supported in part by Grant CA-55254from the National Institutes of Health, by Cytoclonal Pharmaceutics,and by McIntire-Stennis Project 0967 from the Washington StateUniversity Agricultural Research Center.

2
To whom correspondence should be addressed. Fax: (509) 335-7643. E-mail: [email protected].
0003-9861/00 $35.00Copyright © 2000 by Academic PressAll rights of reproduction in any form reserved.

icates a significant degree of similarity betweenhese enzymes (64–67%). The efficient conversion ofaxadien-5a-yl acetate to further hydroxylated inter-

mediates of the Taxol pathway confirms the signifi-cance of this acylation step and suggests this taxadi-enol transacetylase to be an important target for ge-netic manipulation to improve Taxol production.© 2000 Academic Press

Key Words: Taxol biosynthesis; paclitaxel; taxadi-enol-O-acetyl transferase; taxadienyl acetate; Taxuscanadensis; Taxus cuspidata.

The novel diterpenoid Taxol3 (paclitaxel) is now well-established as a potent chemotherapeutic drug. Totalsynthesis of this compound is not commercially feasibleand, currently, Taxol is produced primarily by semi-synthesis from advanced taxane metabolites (1) whichare present in the needles (a renewable resource) ofvarious Taxus species. However, because of the in-creasing demand for this drug, both for use earlier inthe course of cancer intervention and for new thera-peutic applications (2), availability and cost remainimportant issues. Thus, the supply of Taxol and itsprecursors for semisynthesis will have to increasinglyrely on biological methods of production, which may bebased upon either intact Taxus plants or Taxus cellultures (3), or, potentially, microbial systems (4). Inll cases, improving the biological production yields ofaxol depends critically upon a detailed understandingf the biosynthetic pathway, the enzymes catalyzing

3 Paclitaxel is the generic name for Taxol, which is now a regis-tered trademark of Bristol-Myers Squibb. Because of the far greater
familiarity of the word Taxol, we use it in this paper in lieu ofpaclitaxel.
371

rbtpmt

sa

raanccvbssst

d

a cetT

372 WALKER, SCHOENDORF, AND CROTEAU

this sequence of reactions, especially the slow steps,and the genes encoding these proteins.

The biosynthesis of Taxol from primary plant metab-olism requires at least 12 distinct steps (5, 6); the earlysteps involve the cyclization of geranylgeranyl diphos-phate (7) to taxa-4(5),11(12)-diene by taxadiene syn-thase (8) followed by cytochrome P450-mediated hy-droxylation, with accompanying double-bond rear-rangement, to yield taxa-4(20),11(12)-dien-5a-ol (9)(Fig. 1). This secondary alcohol is converted to thecorresponding 5a-yl acetate ester, which apparentlyepresents the third specific intermediate in the Taxoliosynthetic pathway (10) since this acylated deriva-ive is further transformed to a series of advancedolyhydroxylated taxoid metabolites in Taxus microso-al preparations optimized for cytochrome P450 reac-

ions (11) (Fig. 1).The taxadien-5a-ol acetyl transferase has been iso-

lated from yew cell cultures (Taxus canadensis andTaxus cuspidata) that were induced with methyl jas-monate for Taxol production, and the operationallysoluble enzyme was partially purified and fully char-acterized (10). Notably, this acetyl-CoA:taxadien-5a-ol-O-acetyl transferase from Taxus (10) appears to beubstantially different in size, substrate selectivity,nd kinetics from an acetyl-CoA:10-hydroxytaxane-O-

acetyl transferase recently isolated from Taxus chinen-sis (12).

In this manuscript, we report the isolation and pu-

FIG. 1. Enzymatic reactions of the Taxol pathway. The cyclizationsynthase (a), followed by hydroxylation/rearrangement by taxadiecetate by taxadien-5a-ol acetyl transferase (c), are illustrated. The aaxol by several subsequent steps (d–g).

rification of taxa-4(20),11(12)-5a-ol acetyl transferase

from Taxus cells, a reverse genetic approach to theisolation of a full-length cDNA encoding this trans-ferase,4 and the functional expression of this gene in E.coli. Acquisition of the gene encoding the acetyl-CoA:taxa-4(20),11(12)-dien-5a-ol-O-acetyl transferase, thatcatalyzes the first acylation step of Taxol biosynthesis,represents an important advance in efforts to increaseTaxol yields by genetic engineering of producing organ-isms.

MATERIALS AND METHODS

Plant cell cultures, reagents, and general methods. Initiation,propagation, and induction of Taxus sp. cell cultures, as well aseagents, procedures for the synthesis of substrates and standards,nd general methods for transacylase isolation, characterization,nd assay have been previously described (3, 7, 10). Since all desig-ated Taxus species are now considered to be closely related subspe-ies (13, 14), the Taxus cell sources were chosen for operationalonsiderations because only minor sequence differences and/or allelicariants between proteins and genes of the various “species” woulde expected. Thus, T. canadensis suspension cells were chosen as theource of the transacetylase because these cells yielded the highesttarting levels of enzyme activity; T. cuspidata suspension cells wereelected for cDNA library construction because these cells affordedhe highest levels of induced Taxol production.

Enzyme isolation, purification, and assay. The purity of the taxa-ienol acetyl transferase after each fractionation step was assessed

4

geranylgeranyl diphosphate to taxa-4(5),11(12)-diene by taxadiene5a-hydroxylase (b) and acetylation to taxa-4(20),11(12)-dien-5a-ylate is further converted to 10-deacetylbaccatin III, baccatin III, and

ofne-

This sequence has been deposited in GenBank under AccessionNumber AF190130.

atG

awviwcnr3cc10

eMc

mwNlwbtwmdo

StVp

mTabstfmlps

mwD(cw

c

t

373cDNA CLONE FOR ACETYL-CoA:TAXADIENOL ACETYL TRANSFERASE

by SDS-PAGE5 according to Laemmli (15); protein quantificationwas carried out by the method of Bradford or by Coommassie bluestaining (16), and was employed in conjunction with the assay todetermine specific activity. The assay is based on the enzymaticconversion of [20-3H]taxa-4(20),11(12)-dien-5a-ol in the presence ofcetyl-CoA (both substrates at saturation) to the corresponding es-er, followed by radiochromatographic separation of the product andC-MS verification (10).The preparation of the T. canadensis suspension cell extracts and

ll subsequent procedures were performed at 0–4°C unless other-ise noted. The induced suspension cells in 40-g batches were har-ested by filtration, frozen in liquid nitrogen, and thoroughly pulver-zed for 1.5 min in a mortar and pestle. The resulting frozen powderas transferred to 225 ml of ice-cold 30 mM Hepes buffer (pH 7.4)

ontaining 3 mM DTT, XAD-4 polystyrene resin (12 g), and polyvi-ylpolypyrrolidone (PVPP, 12 g) to adsorb low molecular weightesinous and phenolic compounds. The slurry was slowly stirred for0 min, and the mixture was filtered through four layers of cheeseloth to remove solid absorbents and particulates. The filtrate wasentrifuged at 7000g for 30 min to remove cellular debris, then at00,000g for 3 h, followed by filtration of the supernatant through.2-mm nylon mesh, to afford the soluble protein fraction (in ;200 ml

buffer) used as the enzyme source.The soluble enzyme fraction was subjected to ultrafiltration (Diaflo

YM 30 membrane, Millipore Inc.) to concentrate this material from200 to 40 ml, as well as to selectively remove proteins of molecularweight lower than the taxadien-5a-ol acetyl transferase (previouslystablished at ;50,000 (10)). Using a peristaltic pump (GilsoniniPuls 2), this concentrate (40 ml) was applied (2 ml/min) to a

olumn of O-diethylaminoethylcellulose (2.8 3 10 cm, WhatmanDE-52) that had been previously equilibrated with 30 mM Hepesbuffer (pH 7.4) containing 3 mM DTT. After washing with 60 ml ofequilibration buffer to remove unbound material, the bound proteinswere eluted with a step gradient of the same buffer containing 50mM (25 ml), 125 mM (50 ml), and 200 mM (50 ml) NaCl. Fractionscontaining taxadien-5a-ol acetyl transferase activity (125 and 200mM fractions) were combined (100 ml, ;160 mM NaCl) and dilutedto 5 mM NaCl by repeated ultrafiltration (Diaflo YM 30 membrane,Millipore Inc.) and dilution with 30 mM Hepes buffer (pH 7.4) con-taining 3 mM DTT.

Further purification was effected by high-resolution anion-ex-change and hydroxyapatite chromatography (Pharmacia FPLC with280 nm effluent monitoring). The preparation from above (155 ml)was first applied to a preparative anion-exchange column (10 3 100

m, Source 15Q, Pharmacia Biotech) that was previously washedith 30 mM Hepes buffer (pH 7.4), containing 3 mM DTT and 1 MaCl, and then equilibrated with wash buffer (without NaCl). After

oading the sample and removing unbound material, bound proteinsere eluted with a linear gradient of 0 to 200 mM NaCl in equili-ration buffer (215 ml total volume; 3 ml/min) (see Fig. 2A). Frac-ions containing transacetylase activity (eluting at ;80 mM NaCl)ere combined and diluted to 5 mM NaCl by ultrafiltration using 30M Hepes buffer (pH 7.4) containing 3 mM DTT as diluent, as

escribed above. This desalted protein sample (70 ml) was loadednto an analytical anion-exchange column (5 3 50 mm, Source 15Q,

Pharmacia Biotech) that was washed and equilibrated as before.

5 Abbreviations used: bp, base pair(s); CoA, coenzyme A; DTT,dithiothreitol; EDTA, ethylenediaminetetraacetic acid; FPLC, fastprotein liquid chromatography; GC, gas chromatography; Hepes,N-(2-hydroxyethyl)piperazine-N9-2-(ethanesulfonic acid); IPTG, iso-propyl-b-D-thiogalactopyranoside; HPLC, high-performance liquidhromatography; Mops, 3-(N-morpholino)propanesulfonic acid; MS,

mass spectrometry; PAGE, polyacrylamide gel electrophoresis; SDS,
sodium dodecyl sulfate; TAT, taxa-4(20),11(12)-dien-5a-ol-O-acetylransferase; Tris, tris-(hydroxymethyl) aminomethane.
This anion-exchange column was then developed using a shallow,linear salt gradient with elution to 200 mM NaCl (275 ml totalvolume, 1.5 ml/min, 3.0-ml fractions). The taxadienol acetyl trans-ferase eluted at 55–60 mM NaCl (see Fig. 2B), and the appropriatefractions were combined (15 ml), diluted to 45 ml in 30 mM Hepesbuffer (pH 6.9), and applied to a ceramic hydroxyapatite column(10 3 100 mm, Bio-Rad Laboratories) that was previously washedwith 200 mM sodium phosphate buffer (pH 6.9) and then equili-brated with 30 mM Hepes buffer (pH 6.9) containing 3 mM DTT(without sodium phosphate). The equilibration buffer was used todesorb weakly associated protein, and the bound proteins wereeluted by a gradient from 0 to 40 mM sodium phosphate in equili-bration buffer (125 ml total volume, 3.0 ml/min, 3.0-ml fractions) (seeFig. 2C). The fractions containing the highest activity, eluting over27 ml at ;10 mM sodium phosphate, were combined and shown bySDS-PAGE to contain the presumptive transacetylase protein (at;50 kDa) at ;95% purity (with a minor protein contaminant presentat ;35 kDa; see Fig. 2D and Table I).

Amino acid microsequencing of taxadienol acetyl transferase. Thepurified protein from multiple preparations as described above(.95% pure, ;100 pmol, ;50 mg) was subjected to preparative

DS-PAGE (15) and the protein band at 50 kDa, corresponding to theaxadienol acetyl transferase, was excised. Whereas treatment with8 protease or treatment with CNBr failed to yield sequencableeptides, in situ proteolysis with endolysC (Caltech Sequence/Struc-

ture Analysis Facility, Pasadena, CA) and trypsin (17) yielded anumber of peptides, as determined by HPLC, and several of thesewere separated, verified by mass spectrometry (18), and subjected toEdman degradative sequencing, from which five distinct and uniqueamino acid sequences (designated Sequence 1 through Sequence 5)were obtained (see Table II).

cDNA library construction and related manipulations. A cDNAlibrary was constructed from mRNA isolated from T. cuspidata sus-pension cultured cells which had been induced to maximal Taxolproduction with methyl jasmonate for 16 h (3, 7). An optimizedprotocol for the isolation of total RNA from T. cuspidata cells wasdeveloped empirically using an extraction buffer containing 100 mMTris–HCl (pH 7.5), 4 M guanidine thiocyanate, 25 mM EDTA, and 14mM 2-mercaptoethanol. Harvested cells (1.5 g in ;20 ml of extrac-tion buffer) were disrupted at 0–4°C using a Polytron ultrasonicator(4 3 15-s bursts at power setting 7), and the resulting homogenatewas adjusted to 2% (v/v) Triton X-100 and allowed to stand 15 min onice. An equal volume of 3 M sodium acetate (pH 6.0) was then added,and the mixed solution was incubated on ice for an additional 15 min,followed by centrifugation at 15,000g for 30 min at 4°C. The resultingsupernatant was next mixed with 0.8 vol of isopropanol and allowedto stand on ice for 5 min, followed by centrifugation at 15,000g for 30

in at 4°C. The resulting pellet was dissolved in 8.0 ml of 20 mMris-HCl (pH 8.0) containing 1 mM EDTA, and the suspension wasdjusted to pH 7.0 by addition of 2 ml of 2 M NaCl in 250 mM Mopsuffer (pH 7.0). Total RNA was then recovered by passing thisuspension over a nucleic acid isolation column (Qiagen) followinghe manufacturer’s instructions. Poly(A)1 mRNA was then purifiedrom total RNA by chromatography on oligo(dT) beads (OligotexRNA Kit, Qiagen), and this material was used to construct a cDNA

ibrary using the lZAPII-cDNA synthesis kit and gigapack III goldackaging kit from Stratagene by following the manufacturer’s in-tructions.Unless otherwise stated, standard procedures were used for DNAanipulations and cloning (19), and for PCR amplification (20). DNAas sequenced using Amplitaq DNA polymerase and ABI Prism 373NA sequencer. The E. coli strains XL1-Blue and XL1-Blue MRF9

Stratagene) were used for routine cloning of PCR products and forDNA library construction, respectively. E. coli XL1-Blue MRF9 cellsere used for in vivo excision of purified positive lZAP clones, and

2
the excised pBluescript SK phagemids were transformed into E. coliSOLR according to the manufacturer’s instructions (Stratagene).

p

ad

Xbc5Fav0wtw2

a

imcyc

tbs

cetsPttateftupcmapa


Degenerate primer design and PCR amplification. Due to codondegeneracy considerations, only one sequence of the five peptidefragments obtained (Sequence 2 of Table II) was suitable for PCRprimer construction. Such a degenerate forward primer, designatedAT-FOR1 (59-ATI (C/T)TI GT(I/C/A) TA(T/C) TA(T/C) CC(I/C/A)CC-39), was designed based on this sequence. Using the NCBIBLAST 2.0 database searching program (21) to search for this se-quence element among the few defined transacylases of plant origin(22–24), and the many deposited sequences of unknown function,allowed the assembly of a set of seemingly related deduced proteins6

in which a consensus sequence element (DFGWGKP) was noted andwas employed for the design of a degenerate reverse primer desig-nated AT-REV1 (39-CT(A/G) AA(A/G) CC(I/C/A) ACC CC(I/C/A)TT(T/C) GG-59). This set of primers incorporated a number of in-osines, and each was 72-fold degenerate. The remaining four peptidefragment sequences (Sequences 1, 3, 4, and 5 of Table II) were notonly less suitable for primer design but also failed to locate veryclosely related sequences by NCBI BLAST searching, thus suggest-ing that these represented more specific sequence elements of theTaxus transacetylase gene.

The forward and reverse primers (150 mM each) were used in PCRerformed with Taq polymerase (3 U/100 ml reaction containing 2

mM MgCl2) that employed the induced T. cuspidata cell librarycDNA (108 PFU) as template under the following conditions: 94°C for5 min, 32 cycles at 94°C for 1 min, 40°C for 1 min and 74°C for 2 minand, finally, 74°C for 5 min. The resulting 900-bp amplicon wasseparated by agarose gel electrophoresis (19) and the product wasextracted from the gel, ligated into pCR TOPOT7 (Invitrogen), andtransformed into E. coli TOP1OF9 cells (Invitrogen). Plasmid DNAwas prepared from several individual transformants and the insertswere fully sequenced to reveal a unique fragment designated Probe1 (see Fig. 3).

Library screening. For library screening, 1 mg of Probe 1 wasmplified by PCR, and the resulting amplicon was gel-purified, ran-omly labeled with [a-32P]CTP using Ready To Go DNA Labelling

Beads (-dCTP) (Amersham Pharmacia Biotech Inc.) and purified bysize-exclusion chromatography (Probe Quant G-50 Micro columns,Amersham Pharmacia Biotech Inc.). The hybridization probe wasused to screen membrane lifts of 5 3 105 plaques grown in E. coli

L1-Blue MRF9. Phage DNA was cross-linked to the nylon mem-ranes by autoclaving on fast cycle for 3–4 min at 120°C. Afterooling, the membranes were washed for 5 min in 23 SSC, then formin in 63 SSC (containing 0.5% SDS, 53 Denhardt’s reagent, 0.5 gicoll (Type 400, Pharmacia), 0.5 g polyvinylpyrrolidone (PVP-10),nd 0.5 g bovine serum albumin (Fraction V, Sigma) in 100 ml totalolume). Hybridization was performed for 20 h at 68°C in 63 SSC,.5% SDS, and 53 Denhardt’s reagent. The nylon membranes wereashed twice for 5 min in 23 SSC with 0.1% SDS at 25°C and then

wice for 30 min with 13 SSC containing 0.1% SDS at 68°C. Afterashing, the membranes were exposed for 17 h to Kodak XAR film at70°C (19).The plaques affording positive signals were purified through two

dditional rounds of hybridization. Purified lZAPII clones were invivo excised as pBluescript II SK2 phagemids and transformed intoE. coli SOLR cells (Stratagene). The size of each cDNA insert wasdetermined by PCR using T3 and T7 promoter primers, and size-selected inserts (.1.5 kb) were partially sequenced from both endsfor the purpose of sorting and acquiring a full-length version.

cDNA expression in E. coli. A full-length insert fragment (ac-quired as above and designated TAT) was PCR-amplified using for-ward primer 59-GG GAA TTC CAT ATG GAG AAG ACA GAT TTAC-39 and reverse primer 59-C TAG TCT AGA TCA TAC TTT AGCCAC AT-39 to introduce 59-NdeI and 39-XbaI restriction sites (italics)

6
The set of sequences compared may be accessed at website www.wsu.edu/;ibc/faculty/;rc.html.
containing start and stop codons (underlined), respectively, at thesequence termini. The amplicon was digested with NdeI and XbaIand ligated in-frame into the expression vector pCWori1 (25) di-gested with the same restriction enzymes. The recombinant pCWori1

plasmid was verified by sequencing to ensure that no errors had beenintroduced by the polymerase reactions and was then transformedinto E. coli JM109 by standard methods (26).

Transformed E. coli JM109 cells were grown to A600 5 0.5 at 37°Cn 50 ml Luria-Bertani medium supplemented with 50 mg ampicillin/

l, and a 1 ml inoculum was then added to a large-scale (100 ml)ulture of Terrific Broth [6 g bacto-tryptone (Difco Laboratories), 12 geast extract (EM Science), and 2 ml glycerol in 500 ml water]ontaining 50 mg ampicillin/ml and thiamine-HCl (320 mM) and

grown at 28°C. Approximately 24 h after induction with 1 mM IPTG,the bacterial cells were harvested by centrifugation and disrupted bysonication in 30 mM potassium phosphate buffer (pH 7.4), followedby centrifugation to give the soluble enzyme preparation that wasassayed for transacylase activity.

RESULTS AND DISCUSSION

Enzyme Purification

Biochemical studies have indicated that the thirdspecific intermediate of the Taxol biosynthesis path-way is taxa-4(20),11(12)-dien-5a-yl acetate, since thismetabolite serves as the precursor of a series of poly-hydroxy taxanes on route to the end product (Fig. 1)(11). The responsible enzyme, taxadienol acetyl trans-ferase, that converts taxadienol to the C5-acetate esteris, thus, an important candidate for cDNA isolation forthe purpose of overexpression in relevant producingorganisms to increase Taxol yields (10). Because thegene has no homologs (i.e., no other terpenoid O-acetylransferases) in the databases to permit similarity-ased cloning approaches, a protein-based cloningtrategy was adopted.The enzyme has been partially purified from Taxus

ells and characterized with respect to reaction param-ters (10); however, the previously described purifica-ion protocol, including an affinity chromatographytep, did not provide a sufficiently simplified SDS-AGE banding pattern to allow assignment of theransacetylase activity to a specific protein (10), nor didhe procedure yield protein in useful amounts formino acid microsequencing. Numerous variations onhe affinity chromatography step, as well as on thearlier hydrophobic interaction chromatography step,ailed to improve the specific activity of the prepara-ions due to the instability of the enzyme upon manip-lation, and a fivefold increase in the scale of thereparation resulted in only marginally improved re-overy (generally ,5% total yield accompanied by re-oval of .99% of total starting protein). Furthermore,

ttempts to improve enzyme stability by the addition ofolyols (sucrose, glycerol), reducing agents (Na2S2O5,scorbate, dithiothreitol, b-mercaptoethanol), and

other proteins (albumin, casein) were also not produc-tive (10); therefore, the published fractionation proto-
col (10) was abandoned.

hl Pant chr


Using the same methyl jasmonate-induced T. cana-densis suspension cells as an enriched enzyme source(3, 7, 10, 27), a fully revised, alternative isolation andpurification protocol was developed. This new protocol,involving a sequence of anion-exchange and hydroxy-

FIG. 2. Purification of the taxadienol acetyl transferase. The anionHR 15Q (10 3 100 mm) (A) and analytical scale Source HR 15Q (5

ydroxyapatite column (C). The solid line is the UV absorbance at 28ine is the elution gradient (sodium chloride or sodium phosphate).axadien-5a-ol acetyl transferase (at ;50 kDa) after hydroxyapatite

apatite chromatography steps (Fig. 2), efficiently

yielded nearly homogeneous protein at 95% puritywith 800-fold purification (Table I), sufficient for mi-crosequencing. Although the protein was N-terminallyblocked and failed to yield sequencable (internal) pep-tides by V8 proteolysis or CNBr cleavage, treatment

change chromatography elution profiles on preparative scale Source50 mm) (B) are illustrated, as is the elution profile on the ceramicm; dashed line is the relative transacetylase activity (dpm); hatchedel D is a silver-stained, 12% SDS-PAGE gel showing the purity ofomatography. A minor protein contaminant is present at ;35 kDa.

-ex30 n

with endolysC and trypsin yielded a mixture of suit-

tvso


able peptides. Five of these could be separated byHPLC and verified by MS and yielded sequence infor-mation useful for a cloning effort (Table II).

cDNA Library Construction

For cDNA library construction, a stable, methyl jas-monate-inducible T. cuspidata suspension cell line waschosen for mRNA isolation because the production ofTaxol was highly inducible in this system (3, 27). Themixing of experimental tools from different Taxus spe-cies is not a significant limitation, since all Taxus spe-cies are known to be very closely related and are con-sidered by several taxonomists to represent geographicvariants of the basic species T. baccata (13, 14). Thus,the genes encoding geranylgeranyl diphosphate syn-thase (7) and taxadiene synthase (8, 28, 29) (early stepsof Taxol biosynthesis) from T. canadensis and T. cus-pidata evidence only very minor sequence differences.A method was therefore developed for the isolation ofRNA from T. cuspidata cells (see Materials and Meth-ods), and this material was employed for the standardseparation of mRNA (Qiagen) as starting material forcDNA library construction using a commercial kit(Stratagene).

Reverse Genetic Cloning

Of the five tryptic and endolysC peptides that weresequenced (Table II), peptide Sequences 2, 3, and 5were found to exhibit some similarity to the sequencesof the only two other acetyl transferases of plant originthat have been documented, deacetylvindolineO-acetyl transferase involved in indole alkaloid biosyn-thesis (24) and benzyl alcohol O-acetyl transferase in-volved in the biosynthesis of aromatic esters of floralscent (23), and lesser resemblance to a putative aro-matic O-benzoyl transferase of plant origin (22). Of the

TABLE I

Purification of Taxadien-5a-ol-O-Acetyl Transferasefrom Induced Taxus Cells

Totalactivity(pkat)

Totalprotein

(mg)

Specificactivity

(pkat/mg)Fold

purification

Crude extract 302 1230 0.25 1YM30 ultrafiltration 136 98 1.4 5.6DE-52 122 69 1.8 7.2YM30 ultrafiltration 54 55 1.0 4Source 15Q

(10 3 100 mm) 47 3 16 63YM30 ultrafiltration 19 2.6 7.3 29Source 15Q

(5 3 50 mm) 13 0.12 108 400Hydroxyapatite 10 0.05 200 800

five peptide sequences (Table II), Sequence 2 was most

suitable for primer design based on codon degeneracyconsiderations and, based on relative placement, sucha degenerate forward primer (AT-FOR1) was synthe-sized. A search of the databases with the Sequence 2element allowed the assembly of a set of seeminglyrelated proteins6 that permitted identification of a dis-al consensus sequence from which a degenerate re-erse primer (AT-REV1) was designed. This databaseearch also revealed the lack of very significant homol-gy between the Taxus peptide sequences and any pre-

viously described genes.PCR amplification using AT-FOR1 (designed to pep-

tide Sequence 2) and AT-REV1 primers, and inducedTaxus cell library cDNA as target, gave rise, by cloningand sequencing, to a 900-bp amplicon that encoded,with near identity, the proteolytic peptides correspond-ing to Sequences 3, 4, and 5 of the purified protein (seeFig. 3), suggesting that this amplicon represented asubstantial fragment of the target gene for taxadienolacetyl transferase. This fragment (designated Probe 1)was, therefore, 32P-labeled and employed as a hybrid-ization probe in a high-stringency screen of the methyljasmonate-induced T. cuspidata suspension cell lZAPcDNA library. Standard hybridization and purificationprocedures directly led to the isolation of the corre-sponding full-length clone designated TAT4 (Fig. 3).

Functional Expression

To confirm the identity of the putative taxadienolacetyl transferase, TAT was subcloned in-frame as de-scribed into the expression vector pCWori1 (25) andexpressed in E. coli JM109. The transformed bacteriawere grown and induced with IPTG, and cell-free ex-tracts were prepared and evaluated for taxadienolacetyl transferase activity using the previously devel-oped assay (10). The TAT clone (corresponding directlyto Probe 1) expressed high levels of taxadienol acetyltransferase activity as determined by radiochemicalassay (20% conversion of substrate to product understandard assay conditions (10)), and the product of thisrecombinant enzyme was confirmed as taxadien-5a-yl

TABLE II

Peptide Sequences Generated from the PurifiedTaxadienol Acetyl Transferase

Fragment Peptide sequence

1 TTLQLSSIDNLPGVRa

2 ILVYYPPFAGRa

3 FTCGGFVVGVSFa

4 KGLAEIARGEVKb

5 NLPNDTNPSSGYYGNa

a
By proteolysis with trypsin.b By proteolysis with endolysC.

S

c


acetate by combined GC-MS (Fig. 4). As with the nativeenzyme, the recombinant acetyl transferase expressedfrom TAT was incapable of acetylating the advancedTaxol precursor 10-deacetyl baccatin III to baccatin III.Thus, based on functional expression of the target ac-tivity, and the resemblance of the recombinant enzyme

FIG. 3. Nucleotide and deduced amino acid sequences of the taxa-dien-5a-ol acetyl transferase. The double-underlined nucleotide se-quence indicates Probe 1 (the 59- and 39-termini of Probe 1 differ fromthe cDNA sequence due to degeneracy in the PCR primers (see text);BOX1 and BOX5 for Probe 1 are ATG and AAA, respectively). Un-derlined amino acid sequences define the matching proteolytic pep-tides from the native protein [BOX2, BOX3, and BOX4 mark regionsof sequence difference between the deduced amino acid sequence andthe tryptic peptide sequences (cf. Table II)].

in substrate specificity and other physical characteris-

tics to the native form, clone TAT was confirmed toencode the Taxus taxadienol acetyl transferase.

equence Analysis

Clone TAT bears an open-reading frame of 1317 nu-leotides4 and encodes a deduced protein of 439 amino

acids (Fig. 3) with a calculated molecular weight of49,079. The size of this encoded protein is consistentwith the molecular weight of the native, monomerictaxadienol transacetylase (MW ;50,000) determinedby gel permeation chromatography (10) and SDS-PAGE. When compared to the amino acid sequenceinformation from the peptide fragments of the nativeT. canadensis enzyme, the deduced TAT sequence fromT. cuspidata exhibits a very close match (91% identity,see Fig. 3); the minor differences are likely attributableto the species (subspecies) differences or to allelic vari-ation. The sequence does not appear to encode anyN-terminal targeting information, consistent with theoperationally soluble nature, and probable cytosoliclocation, of the monomeric, native enzyme.

The deduced amino acid sequence of TAT resemblesthose of other acetyl transferases (50–56% identity;64–67% similarity) involved in different pathways ofsecondary metabolism in plants (23, 24). However,alignment of the TAT sequence with those of the fewdefined acyl transferases of plant origin, and with arange of most closely related deduced protein se-quences (of undefined function) from plants, showsrather little overall homology (Fig. 5). Since the alcoholsubstrates of these encoded enzymes almost certainlyvary widely, the corresponding active site binding de-terminants would also be expected to vary substan-tially. Notably, there are at least three regions of sig-nificant similarity from which obvious consensus se-quences derive. These conserved elements may beinvolved in binding or catalysis involving the commonacyl-CoA cosubstrates. Additionally, TAT possessesthe HXXXDG (H164, D168, G169) motif found in otheracyltransferases (30–33). Site-directed mutagenesisand chemical modification have shown that the histi-dine residue of this element is essential for catalyticactivity of these enzymes (30–33), suggesting that thishistidine may function as a general base in catalyzingthe transfer of the acyl group from acyl-CoA to thecorresponding alcohol (24).

Three Taxus genes involved in Taxol biosynthesishave now been described, including geranylgeranyldiphosphate synthase (7), taxadiene synthase (28), andthe present acetyl transferase, and the Agrobacterium-mediated transformation of Taxus species has beenaccomplished (34). Since none of the early Taxol path-way intermediates, at least up to taxadienyl acetate,accumulate in Taxus (8, 9), it would appear that the
corresponding enzymatic steps are slow relative to

o

a

e


FIG. 4. Coupled gas chromatographic-mass spectrometric (GC-MS) analysis of the biosynthetic taxadien-5a-yl acetate formed by incubationf taxadien-5a-ol and acetyl-CoA with the recombinant enzyme expressed in E. coli. Panels A and B show the respective GC profile (and

retention time) and mass spectrum of authentic taxadien-5a-ol. Panels C and D show the respective GC profile (and retention time) and massspectrum of authentic taxadien-5a-yl acetate. Panel E shows the GC profile of the reaction products generated by incubation of taxadien-5a-olnd acetyl-coenzyme A with the soluble recombinant enzyme. The residual substrate taxadien-5a-ol (11.16 min), the product taxadien-5a-yl

acetate (11.82 min), dehydrated taxadien-5a-ol (“TOH-H2O” peak), and a contaminant from the buffer, bis-(2-ethylhexyl)phthalate (“BEHP”peak, a plasticizer, CAS 117-81-7) are illustrated. Panel F shows the mass spectrum of taxadien-5a-yl acetate biosynthetically formed by therecombinant enzyme (corresponding to the GC peak at 11.82 min in Panel E). Panel G shows the GC profile of the products generated from
taxadien-5a-ol and acetyl-coenzyme A by incubation with the soluble enzyme fraction derived from E. coli JM109 cells transformed withmpty (control) vector, in which no acetylated product was detected.


FIG. 5. Alignment of deduced, partial amino acid sequence of TAT (Accession No. AF190130) with those of defined peptides in the GenBankdatabase, including partial sequences of benzyl alcohol acetyl transferase from Clarkia breweri (Accession No. AF043464), decetylvindoline4-O-acetyl transferase from Catharanthus roseus (Accession No. AF053307), and anthranilate N-hydroxycinnamoyl benzoyl transferase fromDianthus caryophyllus (Accession No. Z84383), and with those of a selection of seemingly related genes of unknown function from thedatabase (Accession Nos. AC002986 and AC004512 from Arabidopsis thaliana, Accession No. X95343 from Nicotiana tabacum, Accession No.Z70521 from Cucumis melo). Residues boxed in black (and gray) indicate the same (or similar) residues for a minimum of four of thesequences compared and illustrate the few regions of conservation. Forward arrow (left to right) shows the conserved region from which the
degenerate forward PCR primer was designed. Reverse arrow (right to left) shows the conserved region from which the degenerate reversePCR primer was designed (cf. Fig. 3).

2

2

3

3


downstream transformations. Therefore, it seemslikely that incorporation of one or more of these genesunder the influence of a strong constitutive promotercould increase the production yields of Taxol and re-lated taxoids by such transformed cells.

ACKNOWLEDGMENTS

We thank M. Wildung and J. Hefner for helpful discussions, andJoyce Tamura for preparation of the manuscript.

REFERENCES

1. Holton, R. A., Biediger, R. J., and Boatman, P. D. (1995) in Taxol:Science and Applications (Suffness, M., Ed.), pp. 97–121, CRCPress, Boca Raton, FL.

2. Goldspiel, B. R. (1997) Pharmacotherapy 17, 110S–125S.3. Ketchum, R. E. B., Gibson, D. M., Croteau, R. B., and Shuler,

M. L. (1999) Biotechnol. Bioeng. 62, 97–105.4. Stierle, A., Strobel, G., and Stierle, D. (1995) J. Nat. Prod. 58,

1315–1324.5. Floss, H. G., and Mocek, U. (1995) in Taxol: Science and Appli-

cations (Suffness, M., Ed.), pp. 191–208, CRC Press, Boca Raton,FL.

6. Croteau, R., Hefner, J., Hezari, M., and Lewis, N. G. (1996) inPhytochemicals and Health (Flores, H. E., and Gustine, D. L.,Eds.), pp. 94–104, Am. Soc. Plant Physiol., Rockville, MD.

7. Hefner, J., Ketchum, R. E. B., and Croteau, R. (1998) Arch.Biochem. Biophys. 360, 62–74.

8. Koepp, A. E., Hezari, M., Zajicek, J., Vogel, B. S., LaFever, R. E.,Lewis, N. G., and Croteau, R. (1995) J. Biol. Chem. 270, 8686–8690.

9. Hefner, J., Rubenstein, S. M., Ketchum, R. E. B., Gibson, D. M.,Williams, R. M., and Croteau, R. (1996) Chem. Biol. 3, 479–489.

10. Walker, K., Ketchum, R. E. B., Hezari, M., Gatfield, D., Gole-nowski, M., Barthol, A., and Croteau, R. (1999) Arch. Biochem.Biophys. 364, 273–279.

11. Hezari, M., and Croteau, R. (1997) Planta Med. 63, 291–295.12. Menhard, B., and Zenk, M. H. (1999) Phytochemistry 50, 763–

774.13. Bolsinger, C. L., and Jaramillo, A. E. (1990) Silvics of Forest

Trees of North America (revised), p. 17, Pacific Northwest Re-search Station, USDA, Portland, OR.

14. Voliotis, D. (1986) Isr. J. Bot. 35, 47–52.

15. Laemmli, U. K. (1970) Nature 227, 680–685.16. Wray, W., Boulikas, T., Wray, V. P., and Hancock, R. (1981)

Anal. Biochem. 118, 197–203.17. Fernandez, J., Andrews, L., and Mische, S. M. (1994) Anal.

Biochem. 218, 112–118.18. Fernandez, J., Gharahdaghi, F., and Mische, S. M. (1998) Elec-

trophoresis 19, 1036–1045.19. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular

Cloning: A Laboratory Manual, 2nd ed., pp. 7.20–7.43, ColdSpring Harbor, NY.

20. Innis, M. A., Gelfand, D. H., Sninsky, J. J., and White, T. J.(1990) PCR Protocols: A Guide to Methods and Applications, pp.3–12, 253–258, Academic Press, San Diego, CA.

21. Genetics Computer Group (1996) Program Manual for the Wis-consin Package, Version 9.0, Madison, WI.

22. Yang, Q., Reinhard, K., Schiltz, E., and Matern, U. (1997) PlantMol. Biol. 35, 777–789.

23. Dudareva, N., D’Auria, J. C., Nam, K. H., Raguso, R. A., andPichersky, E. (1998) Plant J. 14, 297–304.

24. St-Pierre, B., Laflamme, P., Alarco, A.-M., and De Luca, V.(1998) Plant J. 14, 703–713.

25. Barnes, H. J. (1996) Methods Enzymol. 272, 3–14.26. Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seid-

man, J. G., Smith, J. A., and Struhl, K. (1987) Current Protocolsin Molecular Biology, Greene Publ. and Wiley-Interscience, NewYork.

27. Ketchum, R. E. B., and Gibson, D. M. (1996) Plant Cell TissueOrg. Culture 46, 9–16.

8. Wildung, M. R., and Croteau, R. (1996) J. Biol. Chem. 271,9201–9204.

9. Hezari, M., Ketchum, R. E. B., Gibson, D. M., and Croteau, R.(1997) Arch. Biochem. Biophys. 337, 185–190.

30. Brown, N. F., Anderson, R. C., Caplan, S. L., Foster, D. W., andMcGarry, J. D. (1994) J. Biol. Chem. 269, 19157–19162.

31. Carbini, L. A., and Hersh, L. B. (1993) J. Neurochem. 61, 247–253.

2. Hendle, J., Mattevi, A., Westphal, A. H., Spee, J., de Kok, A.,Teplyakov, A., and Hol, W. G. (1995) Biochemistry 34, 4287–4298.

33. Lewendon, A., Murray, I. A., Shaw, W. V., Gibbs, M. R., andLeslie, A. G. (1994) Biochemistry 33, 1944–1950.

4. Han, K.-H., Fleming, P., Walker, K., Loper, M., Chilton, W. S.,
Mocek, U., Gordon, M. P., and Floss, H. G. (1994) Plant Science95, 187–196.

molecular cloning of a taxa-4(20),11(12)-dien-5α-ol-o-acetyl transferase cdna from taxus and...

Documents