Biocatalysis of precursors to new-generation SB-T-taxanes effective against paclitaxel-resistant cancer cells.

A 10-O-deacetylbaccatin III 10-O-acetyltransferase biocatalyst from Taxus plants was expressed in bacteria whole-cells that were fed 10-O-deacetylbaccatin III and cyclopropane carboxylic acid. Product analysis by qualitative LC/ESI-MS suggested that the C10-acylated products baccatin III, 10-O-n-propionyl-10-O-deacetylbaccatin III, and 10-O-cyclopropanecarbonyl-10-O-deacetylbaccatin III were made in vivo. The results implied that the cells provided non-natural cyclopropanecarbonyl CoA, from a broad-specificity CoA ligase, and natural products, acetyl CoA and n-propionyl CoA, from reserves in the bacteria for use by acyltransferase to acylate 10-O-deacetylbaccatin III in vivo. The 10-acyl-10-O-deacetylbaccatin III are precursors used to synthesize new-generation paclitaxel analogs SB-T-1214 and SB-T-121303, which are effective against cancer cells resistant to paclitaxel and its drug derivatives. The kcat and KM of the acyltransferase for cyclopropanecarbonyl CoA (0.83 s-1, 0.15 M) and n-propionyl CoA (1.2 s-1, 0.15 M) guided scale-up efforts. The 10-acyl-10-O-deacetylbaccatin III analogs (∼45 mg each) were made in vitro by the acyltransferase when incubated with the commercial taxane 10-O-deacetylbaccatin III and synthesized cyclopropanecarbonyl or n-propionyl CoA. The structures of the 10-acyl products were verified by NMR analyses that confirmed C10 acylation of the taxane substrate. LC/ESI-MS/MS analysis also supported the identities of the biocatalyzed products. This effort provides a biocatalysis framework to produce new-generation taxane precursors.

CoA recycling by a benzoate coenzyme A ligase in cascade reactions with aroyltransferases to biocatalyze paclitaxel analogs.

A Pseudomonas CoA ligase (BadA) biocatalyzed aroyl CoA thioesters used by a downstream N-benzoyltransferase (NDTNBT) in a cascade reaction made aroyl analogs of the anticancer drug paclitaxel. BadA kept the high-cost aroyl CoA substrates at saturation for the downstream NDTNBT by recycling CoA when it was added as the limiting reactant. A deacylated taxane substrate N-debenzoyl-2'-deoxypaclitaxel was converted to its benzoylated product at a higher yield, compared to the converted yield in assays in which the BadA ligase chemistry was omitted, and benzoyl CoA was added as a cosubstrate. The resulting benzoylated product 2'-deoxypaclitaxel was made at 196% over the theoretical yield of product that could be made from the CoA added at 50 µM, and the cosubstrates benzoic acid (100 µM), and N-debenzoyl-2'-deoxypaclitaxel (500 µM) added in excess. In addition, a 2-O-benzoyltransferase (mTBT) was incubated with BadA, aroyl acids, CoA, a 2-O-debenzoylated taxane substrate, and cofactors under the CoA-recycling conditions established for the NDTNBT/BadA cascade. The mTBT/BadA combination also made various 2-O-aroylated products that could potentially function as next-generation baccatin III compounds. These ligase/benzoyltransferase cascade reactions show the feasibility of recycling aroyl CoA thioesters in vitro to make bioactive acyl analogs of paclitaxel precursors.

Biocatalysis of a Paclitaxel Analogue: Conversion of Baccatin III to N-Debenzoyl-N-(2-furoyl)paclitaxel and Characterization of an Amino Phenylpropanoyl CoA Transferase.

In this study, we demonstrate an enzyme cascade reaction using a benzoate CoA ligase (BadA), a modified nonribosomal peptide synthase (PheAT), a phenylpropanoyltransferase (BAPT), and a benzoyltransferase (NDTNBT) to produce an anticancer paclitaxel analogue and its precursor from the commercially available biosynthetic intermediate baccatin III. BAPT and NDTNBT are acyltransferases on the biosynthetic pathway to the antineoplastic drug paclitaxel in Taxus plants. For this study, we addressed the recalcitrant expression of BAPT by expressing it as a soluble maltose binding protein fusion (MBP-BAPT). Further, the preparative-scale in vitro biocatalysis of phenylisoserinyl CoA using PheAT enabled thorough kinetic analysis of MBP-BAPT, for the first time, with the cosubstrate baccatin III. The turnover rate of MBP-BAPT was calculated for the product N-debenzoylpaclitaxel, a key intermediate to various bioactive paclitaxel analogues. MBP-BAPT also converted, albeit more slowly, 10-deacetylbaccatin III to N-deacyldocetaxel, a precursor of the pharmaceutical docetaxel. With PheAT available to make phenylisoserinyl CoA and kinetic characterization of MBP-BAPT, we used Michaelis-Menten parameters of the four enzymes to adjust catalyst and substrate loads in a 200-µL one-pot reaction. This multienzyme network produced a paclitaxel analogue N-debenzoyl-N-(2-furoyl)paclitaxel (230 ng) that is more cytotoxic than paclitaxel against certain macrophage cell types. Also in this pilot reaction, the versatile N-debenzoylpaclitaxel intermediate was made at an amount 20-fold greater than the N-(2-furoyl) product. This reaction network has great potential for optimization to scale-up production and is attractive in its regioselective O- and N-acylation steps that remove protecting group manipulations used in paclitaxel analogue synthesis.

Paclitaxel Biosynthesis: Adenylation and Thiolation Domains of an NRPS TycA PheAT Module Produce Various Arylisoserine CoA Thioesters.

Structure-activity relationship studies show that the phenylisoserinyl moiety of paclitaxel (Taxol) is largely necessary for the effective anticancer activity. Several paclitaxel analogues with a variant isoserinyl side chain have improved pharmaceutical properties versus those of the parent drug. To produce the isoserinyl CoAs as intermediates needed for enzyme catalysis on a semibiosynthetic pathway to paclitaxel analogues, we repurposed the adenylation and thiolation domains (Phe-AT) of a nonribosomal peptide synthetase (TycA) so that they would function as a CoA ligase. Twenty-eight isoserine analogue racemates were synthesized by an established procedure based on the Staudinger [2+2] cycloaddition reaction. Phe-AT converted 16 substituted phenylisoserines, one ß-(heteroaryl)isoserine, and one ß-(cyclohexyl)isoserine to their corresponding isoserinyl CoAs. We imagine that these CoA thioesters can likely serve as linchpin biosynthetic acyl donors transferred by a 13-O-acyltransferase to a paclitaxel precursor baccatin III to make drug analogues with better efficacy. It was also interesting to find that an active site mutant [Phe-AT (W227S)] turned over 2-pyridylisoserine and the sterically demanding p-methoxyphenylisoserine substrates to their CoA thioesters, while Phe-AT did not. This mutant is promising for further development to make 3-fluoro-2-pyridylisoserinyl CoA, a biosynthetic precursor of the oral pharmaceutical tesetaxel used for gastric cancers.

Kinetically and Crystallographically Guided Mutations of a Benzoate CoA Ligase (BadA) Elucidate Mechanism and Expand Substrate Permissivity.

A benzoate CoA ligase (BadA), isolated from the bacterium Rhodopseudomonas palustris, catalyzes the conversion of benzoate to benzoyl CoA on the catabolic pathway of aromatic carboxylic acids. Herein, apparent Michaelis constants K(app)cat and K(app)M were determined for an expanded array of 31 substrates chosen to systematically probe the active site architecture of the enzyme and provide a baseline for expansion of wild-type substrate specificity. Acyl CoA products were observed for 25 of the 31 substrates; in general, BadA converted ortho-substituted substrates better than the corresponding meta and para regioisomers, and the turnover number was more affected by steric rather than electronic effects. The kinetic data are interpreted in relation to six crystal structures of BadA in complex with several substrates and a benzoyl-AMP reaction intermediate. In contrast to other known natural substrate-bound benzoate ligase structures, all substrate-bound BadA structures adopted the thiolation conformation instead of the adenylation conformation. We also observed all the aryl carboxylates to be uniquely oriented within the active site, relative to other structures. Together, the kinetics and structural data suggested a mechanism that involves substrate binding in the thiolation conformation, followed by substrate rotation to an active orientation upon the transition to the adenylation conformation. On the basis of this hypothesis and the structural data, sterically demanding active site residues were mutated, and the substrate specificity was expanded substantially versus that of BadA. Novel activities were seen for substrates with larger substituents, including phenyl acetate. Additionally, the mutant Lys427Ala identified this nonconserved residue as essential for the thiolation step of BadA, but not adenylation. These variously acylated CoAs can serve as novel substrates of acyl CoA-dependent acyltransferases in coupled enzyme assays to produce analogues of bioactive natural products.

Genome sequencing and analysis of the paclitaxel-producing endophytic fungus Penicillium aurantiogriseum NRRL 62431.

BACKGROUND: Paclitaxel (Taxol™) is an important anticancer drug with a unique mode of action. The biosynthesis of paclitaxel had been considered restricted to the Taxus species until it was discovered in Taxomyces andreanae, an endophytic fungus of T. brevifolia. Subsequently, paclitaxel was found in hazel (Corylus avellana L.) and in several other endophytic fungi. The distribution of paclitaxel in plants and endophytic fungi and the reported sequence homology of key genes in paclitaxel biosynthesis between plant and fungi species raises the question about whether the origin of this pathway in these two physically associated groups could have been facilitated by horizontal gene transfer. RESULTS: The ability of the endophytic fungus of hazel Penicillium aurantiogriseum NRRL 62431 to independently synthesize paclitaxel was established by liquid chromatography-mass spectrometry and proton nuclear magnetic resonance. The genome of Penicillium aurantiogriseum NRRL 62431 was sequenced and gene candidates that may be involved in paclitaxel biosynthesis were identified by comparison with the 13 known paclitaxel biosynthetic genes in Taxus. We found that paclitaxel biosynthetic gene candidates in P. aurantiogriseum NRRL 62431 have evolved independently and that horizontal gene transfer between this endophytic fungus and its plant host is unlikely. CONCLUSIONS: Our findings shed new light on how paclitaxel-producing endophytic fungi synthesize paclitaxel, and will facilitate metabolic engineering for the industrial production of paclitaxel from fungi.

Separation of alpha- from beta-arylalanines by nickel nitrilotriacetate chromatography.

A method is described to separate alpha- from beta-arylalanines by ligand exchange chromatography on a nickel nitrilotriacetate agarose column with UV monitoring of the effluent. Separate mixtures containing an alpha- and beta-arylalanine pair (1 mg of each) were individually loaded onto the nickel resin pre-equilibrated with the mobile phase at room temperature, and the amino acids were eluted from the column with a gradient from pH 12.0-8.0. The beta-arylalanines eluted first, followed by the alpha-isomers. The four alpha/beta-amino acid pairs tested were well separated with baseline resolution. An aliquot of each fraction was chemically treated to derivatize the amino acids to their N-acyl methyl ester analogs, and their identities were confirmed by GC/MS analysis. The sample recovery was quantitative (>98%), and the column matrix was very resilient, as demonstrated by consistent separation of the solutes after approximately 100 preparative cycles.

The taxol pathway 10-O-acetyltransferase shows regioselective promiscuity with the oxetane hydroxyl of 4-deacetyltaxanes.

The 10-deacetylbaccatin III:10beta-O-acetyltransferase isolated from Taxus cuspidata regiospecifically transfers short-chain alkanoyl groups from their corresponding CoA thioesters to the C10 hydroxyl of 10-deacetylbaccatin III. This 10-O-acetyltransferase along with five other Taxus acyltransferases on the paclitaxel (Taxol) biosynthetic pathway and one additional Taxus-derived acyltransferases of unknown function were screened for 4-O-acetyltransferase activity against 4-deacetylbaccatin III, 7-acetyl-, 13-acetyl-, and 7,13-diacetyl-4-deacetylbaccatin III. These 4-deacyl derivatives were semisynthesized from the natural product baccatin III via silyl protecting group manipulation, regioselective reductive ester cleavage with sodium bis(2-methoxyethoxy)aluminum hydride, and regioselective acetylation with acetic anhydride. Assays with the 4-deacetylated diterpene substrates and acetyl CoA revealed the taxane 10beta-O-acetyltransferase was able to catalyze the 4-O-acetylation of 4-deacetylbaccatin III to baccatin III and 13-acetyl-4-deacetylbacatin III to 13-acetylbaccatin III, although each was converted at lesser efficiency than with the natural substrate. In contrast, this enzyme was unable to acetylate 7-acetyl-4-deacetylbaccatin III and 7,13-diacetyl-4-deacetylbaccatin III substrates at C4, suggesting that the C7 hydroxyl of baccatin III must remain deacylated for enzyme function. The biocatalytic transfer of an acyl group to the tertiary hydroxyl on the oxetane moiety at C4 of the taxane ring demonstrates that the regiochemistry of the 10beta-acetyltransferase is mutable.

Profiling a Taxol pathway 10beta-acetyltransferase: assessment of the specificity and the production of baccatin III by in vivo acetylation in E. coli.

The 10beta-acetyltransferase on the biosynthetic pathway of the antineoplastic drug Taxol catalyzes the regiospecific transfer of the acetyl group of acetyl-coenzyme A (CoA) to 10-deacetylbaccatin III. We demonstrate that in addition to acetyl group transfer, the overexpressed enzyme also catalyzes the exchange of propionyl and n-butyryl from the corresponding CoA thioester to the hydroxyl group at C10 of the cosubstrate. Also, in vivo studies revealed that E. coli, producing endogenous acetyl-CoA and overexpressing the recombinant acetyltransferase, can convert exogenously supplied 10-deacetylbaccatin III to baccatin III. Potentially, this heterologous in vivo production method in bacteria could be optimized to couple various unnatural acyl-CoA analogs to myriad amino and/or hydroxyl acceptors by acyltransferase catalysis; conceivably, this process could facilitate the preparation of second-generation Taxols.

Cloning, heterologous expression, and characterization of a phenylalanine aminomutase involved in Taxol biosynthesis.

Biosynthesis of the N-benzoyl phenylisoserinoyl side chain of the anticancer drug Taxol starts with the conversion of 2S-alpha-phenylalanine to 3R-beta-phenylalanine by phenylalanine aminomutase (PAM). A gene cloning approach was based on the assumption that PAM would resemble the well known plant enzyme phenylalanine ammonia lyase. A phenylalanine ammonia lyase-like sequence acquired from a Taxus cuspidata cDNA library was expressed functionally in Escherichia coli and confirmed as the target aminomutase that is virtually identical to the recombinant enzyme and clone from Taxus chinensis, acquired recently by a reverse genetics approach (Bristol-Myers Squibb (August 14, 2003) U. S. Patent WO 03/066871 A2). The full-length cDNA has an open reading frame of 2094 base pairs and encodes a protein of 698 residues with a calculated molecular mass of 76,530 Da. The recombinant mutase has a pH optimum of 8.5, a k(cat) value of 0.015 s(-1), and a K(m) of 45 +/- 8 microm for 2S-alpha-phenylalanine. The stereochemical mechanism of PAM involves the removal and interchange of the pro-3S hydrogen and the amino group, which rebonds at C-3 with retention of configuration. The recombinant enzyme appears to catalyze both the forward and reverse reactions with specificity for both 2S-alpha-phenylalanine and 3S- or 3R-beta-phenylalanine substrates, respectively, whereas the related phenylpropanoids 2S-aminocyclohexanepropanoic acid, 2R-alpha-phenylalanine, and 2S-alpha-tyrosine are not converted to their beta-isomers by the mutase.