Tolerance and specificity of recombinant 6-methylsalicyclic acid synthase.

BACKGROUND: 6-Methylsalicylic acid synthase (MSAS), a fungal polyketide synthase from Penicillium patulum, is perhaps the simplest polyketide synthase that embodies several hallmarks of this family of multifunctional enzymes--a large multidomain protein, a high degree of specificity toward acetyl-CoA and malonyl-CoA substrates, chain length control, and regiospecific ketoreduction. MSAS has recently been functionally expressed in Escherichia coli and Saccharomyces cerevisiae, leading to the engineered biosynthesis of 6-methylsalicylic acid in these hosts. These developments have set the stage for detailed mechanistic studies of this model system. RESULTS: A three--step purification procedure was developed to obtain >95% pure MSAS from extracts of E. coli. As reported earlier for the enzyme isolated from P. patulum, the recombinant enzyme produced 6-methylsalicylic acid (a reduced tetraketide) in the presence of acetyl-CoA, malonyl-CoA, and NADPH, but triacetic acid lactone (an unreduced triketide) in the absence of NADPH. Consistent with this observation, point mutations in the highly conserved nucleotide-binding motif of the ketoreductase domain also led to production of triacetic acid lactone in vivo. The enzyme showed some tolerance toward nonnatural primer units including propionyl- and butyryl-CoA, but was incapable of incorporating extender units from (R, S)-methylmalonyl-CoA. Interestingly, MSAS readily accepted the N-acetylcysteamine (NAC) analog of malonyl-CoA as a substrate. CONCLUSIONS: NAC thioesters are simple, cost-effective analogs of CoA thioester substrates, and therefore provide a facile strategy for probing the molecular recognition features of polyketide synthases using unnatural building blocks. The ability to produce 4-hydroxy-6-methyl-2-pyrone in both E. coli and yeast illustrates the feasibility of metabolic engineering of these hosts to produce unnatural polyketides. Finally, the abundant source of recombinant MSAS described here provides an opportunity to study this fascinating model system using a combination of structural, mechanistic, and mutagenesis approaches.

Characterization of the macrolide P-450 hydroxylase from Streptomyces venezuelae which converts narbomycin to picromycin.

The post-polyketide synthase (PKS) biosynthetic tailoring of macrolide antibiotics usually involves one or more oxidation reactions catalyzed by cytochrome P450 monooxygenases. As the specificities of members from this class of enzymes vary significantly among PKS gene clusters, the identification and study of new macrolide P450s are important to the growing field of combinatorial biosynthesis. We have isolated the cytochrome P450 gene picK from Streptomyces venezuelae which is responsible for the C-12 hydroxylation of narbomycin to picromycin. The gene was located by searching regions proximal to modular PKS genes with a probe for macrolide P450 monooxygenases. The overproduction of PicK with a C-terminal six-His affinity tag (PicK/6-His) in Escherichia coli aided the purification of the enzyme for kinetic analysis. PicK/6-His was shown to catalyze the in vitro C-12 hydroxylation of narbomycin with a kcat of 1.4 s-1, which is similar to the value reported for the related C-12 hydroxylation of erythromycin D by the EryK hydroxylase. The unique specificity of this enzyme should be useful for the modification of novel macrolide substrates similar to narbomycin, in particular, ketolides, a promising class of semisynthetic macrolides with activity against erythromycin-resistant pathogens.

Production of a polyketide natural product in nonpolyketide-producing prokaryotic and eukaryotic hosts.

The polyketides are a diverse group of natural products with great significance as human and veterinary pharmaceuticals. A significant barrier to the production of novel genetically engineered polyketides has been the lack of available heterologous expression systems for functional polyketide synthases (PKSs). Herein, we report the expression of an intact functional PKS in Escherichia coli and Saccharomyces cerevisiae. The fungal gene encoding 6-methylsalicylic acid synthase from Penicillium patulum was expressed in E. coli and S. cerevisiae and the polyketide 6-methylsalicylic acid (6-MSA) was produced. In both bacterial and yeast hosts, polyketide production required coexpression of 6-methylsalicylic acid synthase and a heterologous phosphopantetheinyl transferase that was required to convert the expressed apo-PKS to its holo form. Production of 6-MSA in E. coli was both temperature- and glycerol-dependent and levels of production were lower than those of P. patulum, the native host. In yeast, however, 6-MSA levels greater than 2-fold higher than the native host were observed. The heterologous expression systems described will facilitate the manipulation of PKS genes and consequent production of novel engineered polyketides and polyketide libraries.

Macrolide biosynthesis: a single cytochrome P450, PicK, is responsible for the hydroxylations that generate methymycin, neomethymycin, and picromycin in Streptomyces venezuelae.

The final step in the biosynthesis of methymycin, neomethymycin, and picromycin is an hydroxylation, shown to be carried out by the cytochrome P-450 monooxygenase, PicK. Direct comparison of the relative Kcat/K(m) values for the two substrates, YC-17 and narbomycin, showed a threefold rate preference of picK for narbomycin.

The dynamic NMR structure of the T psi C-loop: implications for the specificity of tRNA methylation.

tRNA (m5U54)-methyltransferase (RUMT) catalyzes the S-adenosylmethionine-dependent methylation of uridine-54 in the T psi C-loop of all transfer RNAs in E. coli to form the 54-ribosylthymine residue. However, in all tRNA structures, residue 54 is completely buried and the question arises as to how RUMT gains access to the methylation site. A 17-mer RNA hairpin consisting of nucleotides 49-65 of the T psi-loop is a substrate for RUMT. Homonuclear NMR methods in conjunction with restrained molecular dynamics (MD) methods were used to determine the solution structure of the 17-mer T-arm fragment. The loop of the hairpin exhibits enhanced flexibility which renders the conventional NMR average structure less useful compared to the more commonly found situation where a molecule exists in predominantly one major conformation. However, when resorting to softer refinement methods such as MD with time-averaged restraints, the conflicting restraints in the loop can be satisfied much better. The dynamic structure of the T-arm is represented as an ensemble of 10 time-clusters. In all of these, U54 is completely exposed. The flexibility of the T psi-loop in solution in conjunction with extensive binding studies of RUMT with the T psi C-loop and tRNA suggest that the specificity of the RUMT/ tRNA recognition is associated with tRNA tertiary structure elements. For the methylation, RUMT would simply have to break the tertiary interactions between the D- and T-loops, leading to a melting of the T-arm structure and making U54 available for methylation.

Role of the conserved tryptophan 82 of Lactobacillus casei thymidylate synthase.

BACKGROUND: Thymidylate synthase (TS; EC 2.1.1.45) catalyzes the reductive methylation of 2'-deoxyuridine-5'-monophosphate (dUMP) by 5,10-methylene-5,6,7,8-tetrahydrofolate (CH2H4folate) to produce 2'-deoxythymidine-5'-monophosphate (dTMP) and 7,8-dihydrofolate (H2folate). Major advances in the understanding of the mechanism of TS have been made by studying site-specific mutants of the enzyme. Trp82 is completely conserved in all of the 20 TS sequences known. It forms part of the CH2H4folate binding pocket, is reported to be a component of a catalytically important H-bond network, and is suspected to be the source of an unusual absorbance change at 330 nm when TS forms a ternary complex with 5-fluoro-dTMP and CH2H4folate. We therefore prepared and characterized a set of 12 mutants at position 82 of Lactobacillus casei TS. RESULTS: Eight Trp82 mutants were active enough for us to determine their kinetic constants for dTMP production, while four were inactive. The active mutants had higher Km values for dUMP (2- to 10-fold) and CH2H4folate (2- to 27-fold), and lower kcat values (12- to 250-fold) than wild-type TS. The most active mutants were those containing the aromatic side chains Phe and His at position 82. All of the Trp82 mutants catalyzed the debromination of 5-bromo-dUMP with kinetic parameters similar to those of wild-type TS, and all formed ternary complexes with 5-fluoro-dUMP and CH2H4folate. The absence of Trp82 did not prevent the absorbance change at 330 nm on ternary complex formation. CONCLUSIONS: Trp82, a completely conserved residue that was shown by X-ray crystallography to interact directly with CH2H4folate and indirectly with dUMP, does not appear to be essential for binding or catalysis. We do, however, find a preference for an aromatic side chain at position 82. Trp82 does not contribute to the unique spectral change at 330 nm that accompanies TS ternary complex formation.

Stereochemistry of tRNA(m5U54)-methyltransferase catalysis: 19F NMR spectroscopy of an enzyme-FUraRNA covalent complex.

The catalytic mechanism of tRNA(m5U54)-methyltransferase (RUMT) involves the formation of a covalent adduct between Cys324 of RUMT and C6 of Ura54 in tRNA. The covalent adduct is subsequently methylated at C5 by S-adenosyl-L-methionine (AdoMet). We used an RNA substrate analog containing 5-fluorouracil (FUra) in place of Ura54 to trap the covalent complex and analyzed the adduct by 19F NMR spectroscopy. The 19F NMR spectrum of the adduct consisted of an overlapping doublet of quartets, with an H6-F coupling constant of 4 Hz and a CH3-F coupling constant of 22.4 Hz. On the basis of the magnitude of the H6-F coupling constant, we determined that Cys324 of RUMT and the methyl moiety from AdoMet added across the 5,6-double bond of FUra54 in cis fashion. We deduced that the nucleophilic addition was also cis in the normal enzymatic reaction and that the subsequent beta-elimination of the 5-H and catalytic cysteine was trans. Further, on the basis of chemical considerations, we proposed several conformational adaptations of enzyme-substrate complexes that must occur on the reaction pathway. Together with previous studies, this study enables the proposal of the complete stereochemical pathway for the RUMT-catalyzed methylation of Ura54 in tRNA.

High-level expression and rapid purification of tRNA (m5U54)-methyltransferase.

We report an extremely high-level expression system for tRNA (m5U54)-methyltransferase (RUMT), and a purification strategy which routinely yields 20 to 50 mg of homogeneous RUMT per liter of Escherichia coli cells. The RUMT gene (trmA) was cloned into a pET vector and transformed into E. coli BL21 (DE3) cells. Following induction, this system produces active enzyme at a level approaching 50% of the total soluble protein. A purification scheme consisting of DEAE-cellulose chromatography to remove nucleic acids, followed by phosphocellulose chromatography, provides homogeneous enzyme. The entire procedure, from cell growth to purified enzyme, takes less than 2 days. This represents a significant improvement over the previously published expression/purification protocol for RUMT (Gu, X, and Santi, D.V., Protein Expression Purif. 2, 66-68, 1991), which typically nets 5- to 10-fold less enzyme per liter of cells and is substantially more labor intensive.

Enzymatic mechanism of tRNA (m5U54)methyltransferase.

tRNA (m5U54)methyltransferase (RUMT) catalyzes the methylation of uridine 54 of transfer RNA by S-adenosyl-L-methionine. In this report, we present the enzymatic mechanism of RUMT, including the stereochemical course of the methylation reaction, and discuss RUMT-tRNA recognition. As part of its enzymatic mechanism, we postulate that RUMT catalyzes the disruption of RNA-RNA contacts. We also show that many nucleotide substitutions can be made in the T-loop of tRNA without affecting RUMT binding, indicating that the recognition of the T-loop by RUMT is not stringent.

Purification methods for recombinant Lactobacillus casei thymidylate synthase and mutants: a general, automated procedure.

General procedures for the rapid, large-scale purification of recombinant Lactobacillus casei thymidylate synthase and its mutants have been established. An effective method employs sequential phosphocellulose and hydroxylapatite chromatography. Crude cell extracts are directly applied to phosphocellulose, and the enzyme is obtained in a nearly pure state by stepwise elution with KCl. The eluate is directly applied to hydroxylapatite, and the homogeneous enzyme is eluted with a gradient of potassium phosphate. The method has been successful for the purification of recombinant wild-type enzyme and all mutants thus far examined. The entire purification procedure has been automated using a commonly available FPLC system and can be performed in several hours with minimal operator time.

Stereochemistry of methyl transfer catalyzed by tRNA (m5U54)-methyltransferase--evidence for a single displacement mechanism.

tRNA (m5U54)-methyltransferase (RUMT) catalyzes the transfer of a methyl group from S-adenosyl-L-methionine (AdoMet) to the 5-carbon of uridine 54 of tRNA. We have determined the steric course of methyl transfer, using (methyl-R)- and (methyl-S)-[methyl-2H1,3H]-AdoMet as the chiral methyl donors, and tRNA lacking the 5-methyl group at position 54 as the acceptor. Following methyl transfer, ribothymidine was isolated and degraded to chiral acetic acid for configurational analysis. Transfer of the chiral methyl group to U54 proceeded with inversion of configuration of the chiral methyl group, suggesting that RUMT catalyzed methyl transfer occurs by a single SN2 displacement mechanism.

Identification of the catalytic nucleophile of tRNA (m5U54)methyltransferase.

A covalent complex between tRNA (m5U54)methyltransferase, 5-fluorouridine tRNA(Phe), and S-adenosyl-L-[methyl-3H]methionine was formed in vitro and purified. Previously, it was shown that in this complex the 6-position of fluorouridine-54 is covalently linked to a catalytic nucleophile and the 5-position is bound to the transferred methyl group of AdoMet [Santi, D. V., & Hardy, L. W. (1987) Biochemistry 26, 8599-8606]. Proteolysis of the complex generated a [3H]methyl-FUtRNA-bound peptide, which was purified by 7 M urea-15% polyacrylamide gel electrophoresis. The peptide component of the complex was sequenced by gas-phase Edman degradation and found to contain two cysteines. The tritium was shown to be associated with Cys 324 of the methyltransferase, which unequivocally identifies this residue as the catalytic nucleophile.

Human mast cell carboxypeptidase. Purification and characterization.

A carboxypeptidase activity was recently identified in highly purified human lung mast cells and dispersed mast cells from skin. Using affinity chromatography with potato-tuber carboxypeptidase inhibitor as ligand, mast cell carboxypeptidase was purified to homogeneity from whole skin extracts. The purified enzyme yielded a single staining band of approximately 34,500 D on SDS-PAGE. Carboxypeptidase enzyme content estimated by determination of specific activity, was 0.5, 5, and 16 micrograms/10(6) mast cells from neonatal foreskin, adult facial skin, and adult foreskin, respectively. Human mast cell carboxypeptidase resembled bovine carboxypeptidase A with respect to hydrolysis of synthetic dipeptides and angiotensin I, but was distinguished from carboxypeptidase A in its inability to hydrolyze des-Arg9 bradykinin. The amino acid composition of human mast cell carboxypeptidase was similar to the composition of rat mast cell carboxypeptidase. The amino-terminal amino acid sequence of mast cell carboxypeptidase demonstrated 65% positional identity with human pancreatic carboxypeptidase B, but only 19% with human carboxypeptidase A. Thus, human mast cell carboxypeptidase is a novel member of the protein family of zinc-containing carboxypeptidases, in that it is functionally similar but not identical to bovine carboxypeptidase A, but has structural similarity to bovine and human pancreatic carboxypeptidase B.