Harnessing the biosynthetic code: combinations, permutations, and mutations.

Polyketides and non-ribosomal peptides are two large families of complex natural products that are built from simple carboxylic acid or amino acid monomers, respectively, and that have important medicinal or agrochemical properties. Despite the substantial differences between these two classes of natural products, each is synthesized biologically under the control of exceptionally large, multifunctional proteins termed polyketide synthases (PKSs) and non-ribosomal peptide synthetases (NRPSs) that contain repeated, coordinated groups of active sites called modules, in which each module is responsible for catalysis of one complete cycle of polyketide or polypeptide chain elongation and associated functional group modifications. It has recently become possible to use molecular genetic methodology to alter the number, content, and order of such modules and, in so doing, to alter rationally the structure of the resultant products. This review considers the promise and challenges inherent in the combinatorial manipulation of PKS and NRPS structure in order to generate entirely "unnatural" products.

Aminoacyl-CoAs as probes of condensation domain selectivity in nonribosomal peptide synthesis.

In nonribosomal biosynthesis of peptide antibiotics by multimodular synthetases, amino acid monomers are activated by the adenylation domains of the synthetase and loaded onto the adjacent carrier protein domains as thioesters, then the formation of peptide bonds and translocation of the growing chain are effected by the synthetase's condensation domains. Whether the condensation domains have any editing function has been unknown. Synthesis of aminoacyl-coenzyme A (CoA) molecules and direct enzymatic transfer of aminoacyl-phosphopantetheine to the carrier domains allow the adenylation domain editing function to be bypassed. This method was used to demonstrate that the first condensation domain of tyrocidine synthetase shows low selectivity at the donor residue (D-phenylalanine) and higher selectivity at the acceptor residue (L-proline) in the formation of the chain-initiating D-Phe-L-Pro dipeptidyl-enzyme intermediate.

The MerR metalloregulatory protein binds mercuric ion as a tricoordinate, metal-bridged dimer.

Bacterial MerR proteins are dimeric DNA-binding proteins that mediate the Hg(II)-dependent induction of mercury resistance operons. Site-directed mutagenesis of the Bacillus sp. RC607 MerR protein reveals that three of four Cys residues per monomer are required for Hg(II) binding at the single high-affinity binding site. Inactive mutant homodimers can exchange subunits to form heterodimers active for Hg(II) binding. Studies of a heterodimer retaining only three of eight cysteine residues per dimer reveal that Cys79 in one subunit and Cys114 and Cys123 in the second subunit are necessary and sufficient for high-affinity Hg(II) binding in an asymmetric, subunit bridging coordination complex.

From peptide precursors to oxazole and thiazole-containing peptide antibiotics: microcin B17 synthase.

Esherichia coli microcin B17 is a posttranslationally modified peptide that inhibits bacterial DNA gyrase. It contains four oxazole and four thiazole rings and is representative of a broad class of pharmaceutically important natural products with five-membered heterocycles derived from peptide precursors. An in vitro assay was developed to detect heterocycle formation, and an enzyme complex, microcin B17 synthase, was purified and found to contain three proteins, McbB, McbC, and McbD, that convert 14 residues into the eight mono- and bisheterocyclic moieties in vitro that confer antibiotic activity on mature microcin B17. These enzymatic reactions alter the peptide backbone connectivity. The propeptide region of premicrocin is the major recognition determinant for binding and downstream heterocycle formation by microcin B17 synthase. A general pathway for the enzymatic biosynthesis of these heterocycles is formulated.

Vancomycin resistance: structure of D-alanine:D-alanine ligase at 2.3 A resolution.

The molecular structure of the D-alanine:D-alanine ligase of the ddlB gene of Escherichia coli, co-crystallized with an S,R-methylphosphinate and adenosine triphosphate, was determined by x-ray diffraction to a resolution of 2.3 angstroms. A catalytic mechanism for the ligation of two D-alanine substrates is proposed in which a helix dipole and a hydrogen-bonded triad of tyrosine, serine, and glutamic acid assist binding and deprotonation steps. From sequence comparison, it is proposed that a different triad exists in a recently discovered D-alanine:D-lactate ligase (VanA) present in vancomycin-resistant enterococci. A molecular mechanism for the altered specificity of VanA is suggested.

The residual enzymatic phosphorylation activity of hexokinase II mutants is correlated with glucose repression in Saccharomyces cerevisiae.

Saccharomyces cerevisiae mutants containing different point mutations in the HXK2 gene were used to study the relationship between phosphorylation by hexokinase II and glucose repression in yeast cells. Mutants showing different levels of hexokinase activity were examined for the degree of glucose repression as indicated by the levels of invertase activity. The levels of hexokinase activity and invertase activity showed a strong inverse correlation, with a few exceptions attributable to very unstable hexokinase II proteins. The in vivo hexokinase II activity was determined by measuring growth rates, using fructose as a carbon source. This in vivo hexokinase II activity was similarly inversely correlated with invertase activity. Several hxk2 alleles were transferred to multicopy plasmids to study the effects of increasing the amounts of mutant proteins. The cells that contained the multicopy plasmids exhibited less invertase and more hexokinase activity, further strengthening the correlation. These results strongly support the hypothesis that the phosphorylation activity of hexokinase II is correlated with glucose repression.

Isolation and characterization of mutations in the HXK2 gene of Saccharomyces cerevisiae.

Several hundred new mutations in the gene (HXK2) encoding hexokinase II of Saccharomyces cerevisiae were isolated, and a subset of them was mapped, resulting in a fine-structure genetic map. Among the mutations that were sequenced, 35 were independent missense mutations. The mutations were obtained by mutagenesis of cloned HXK2 DNA carried on a low-copy-number plasmid vector and screened for a number of different phenotypes in yeast strains bearing chromosomal hxk1 and hxk2 null mutations. Some of these mutants were characterized both in vivo and in vitro; they displayed a wide spectrum of residual hexokinase activities, as indicated by three assays: in vitro enzyme activity, ability to grow on glucose and fructose, and ability to repress invertase production when growing on glucose. Of those that failed to support growth on fructose, only a small minority made normal-size, stable, and inactive protein. Analysis of the amino acid changes in these mutants in light of the crystallographically determined three-dimensional structure of hexokinase II suggests important roles in structure or catalysis for six amino acid residues, only two of which are near the active site.

Lck-dependent tyrosyl phosphorylation of the phosphotyrosine phosphatase SH-PTP1 in murine T cells.

The phosphorylation and dephosphorylation of proteins on tyrosyl residues are key regulatory mechanisms in T-cell signal transduction and are controlled by the opposing activities of protein tyrosine kinases and phosphotyrosyl phosphatases (PTPs). In T cells, several nontransmembrane protein tyrosine kinases are associated with receptors; for example, Lck is bound to the coreceptors CD4 and CD8 and becomes activated upon their stimulation. In comparison, little is known about the role of nontransmembrane PTPs in early T-cell signaling. SH-PTP1 (PTP1C, HCP, SHP) is a nontransmembrane PTP expressed primarily in hematopoietic cells, including T cells. We have found that SH-PTP1 is basally phosphorylated on serine in resting T cells. Upon stimulation of CD4 or CD8 either in a T-cell hybridoma cell line or in primary thymocytes, SH-PTP1 becomes tyrosyl phosphorylated. Moreover, SH-PTP1 is constitutively phosphorylated on tyrosine in the Lck-overexpressing lymphoma cell line LSTRA. SH-PTP1 is also a good substrate for recombinant Lck in vitro. Comparisons of the tryptic phosphopeptide maps of wild-type SH-PTP1 and deletion and point mutations establish that the two sites (Y-536 and Y-564) which are directly phosphorylated by Lck in vitro are also phosphorylated in vivo in LSTRA cells. One of these sites (Y-564) is phosphorylated in T cells in response to Lck activation. We conclude that SH-PTP1 undergoes Lck-dependent tyrosyl phosphorylation in T cells and likely plays a role in early T-cell signaling.

The structure of the substrate-free form of MurB, an essential enzyme for the synthesis of bacterial cell walls.

BACKGROUND: The repeating disaccharide and pentapeptide units of the bacterial peptidoglycan layer are connected by a lactyl ether bridge biosynthesized from UDP-N-acetylglucosamine and phosphoenolpyruvate in sequential enol ether transfer and reduction steps catalyzed by MurA and MurB respectively. Knowledge of the structure and mechanism of the MurB enzyme will permit analysis of this unusual enol ether reduction reaction and may facilitate the design of inhibitors as candidate next-generation antimicrobial agents. RESULTS: The crystal structure of UDP-N-acetylenolpyruvylglucosamine reductase, MurB, has been solved at 3.0 A and compared with our previously reported structure of MurB complexed with its substrate enolpyruvyl-UDP-N- acetylglucosamine. Comparison of the liganded structure of MurB with this unliganded form reveals that the binding of substrate induces a substantial movement of domain 3 (residues 219-319) of the enzyme and a significant rearrangement of a loop within this domain. These ligand induced changes disrupt a stacking interaction between two tyrosines (Tyr190 and Tyr254) which lie at the side of the channel leading to the active site of the free enzyme. CONCLUSIONS: The conformational change induced by enolpyruvyl-UDP-N- acetylglucosamine binding to MurB results in the closure of the substrate-binding channel over the substrate. Tyr190 swings over the channel opening and establishes a hydrogen bond with an oxygen of the alpha-phosphate of the sugar nucleotide substrate which is critical to substrate binding.

Structure of the UDP-glucosyltransferase GtfB that modifies the heptapeptide aglycone in the biosynthesis of vancomycin group antibiotics.

BACKGROUND: Members of the vancomycin group of glycopeptide antibiotics have an oxidatively crosslinked heptapeptide scaffold decorated at the hydroxyl groups of 4-OH-Phegly4 or beta-OH-Tyr6 with mono- (residue 6) or disaccharides (residue 4). The disaccharide in vancomycin itself is L-vancosamine-1,2-glucose, and in chloroeremomycin it is L-4-epi-vancosamine-1,2-glucose. The sugars and their substituents play an important role in efficacy, particularly against vancomycin-resistant pathogenic enterococci. RESULTS: The glucosyltransferase, GtfB, that transfers the glucose residue from UDP-glucose to the 4-OH-Phegly4 residue of the vancomycin aglycone, initiating the glycosylation pathway in chloroeremomycin maturation, has been crystallized, and its structure has been determined by X-ray analysis at 1.8 A resolution. The enzyme has a two-domain structure, with a deep interdomain cleft identified as the likely site of UDP-glucose binding. A hydrophobic patch on the surface of the N-terminal domain is proposed to be the binding site of the aglycone substrate. Mutagenesis has revealed Asp332 as the best candidate for the general base in the glucosyltransfer reaction. CONCLUSIONS: The structure of GtfB places it in a growing group of glycosyltransferases, including Escherichia coli MurG and a beta-glucosyltransferase from T4 phage, which together form a subclass of the glycosyltransferase superfamily and give insights into the recognition of the NDP-sugar and aglycone cosubstrates. A single major interdomain linker between the N- and C- terminal domains suggests that reprogramming of sugar transfer or aglycone recognition in the antibiotic glycosyltransferases, including the glycopeptide and also the macrolide antibiotics, will be facilitated by this structural information.

Enzymes of vancomycin resistance: the structure of D-alanine-D-lactate ligase of naturally resistant Leuconostoc mesenteroides.

BACKGROUND: The bacterial cell wall and the enzymes that synthesize it are targets of glycopeptide antibiotics (vancomycins and teicoplanins) and beta-lactams (penicillins and cephalosporins). Biosynthesis of cell wall peptidoglycan requires a crosslinking of peptidyl moieties on adjacent glycan strands. The D-alanine-D-alanine transpeptidase, which catalyzes this crosslinking, is the target of beta-lactam antibiotics. Glycopeptides, in contrast, do not inhibit an enzyme, but bind directly to D-alanine-D-alanine and prevent subsequent crosslinking by the transpeptidase. Clinical resistance to vancomycin in enterococcal pathogens has been traced to altered ligases producing D-alanine-D-lactate rather than D-alanine-D-alanine. RESULTS: The structure of a D-alanine-D-lactate ligase has been determined by multiple anomalous dispersion (MAD) phasing to 2.4 A resolution. Co-crystallization of the Leuconostoc mesenteroides LmDdl2 ligase with ATP and a di-D-methylphosphinate produced ADP and a phosphinophosphate analog of the reaction intermediate of cell wall peptidoglycan biosynthesis. Comparison of this D-alanine-D-lactate ligase with the known structure of DdlB D-alanine-D-alanine ligase, a wild-type enzyme that does not provide vancomycin resistance, reveals alterations in the size and hydrophobicity of the site for D-lactate binding (subsite 2). A decrease was noted in the ability of the ligase to hydrogen bond a substrate molecule entering subsite 2. CONCLUSIONS: Structural differences at subsite 2 of the D-alanine-D-lactate ligase help explain a substrate specificity shift (D-alanine to D-lactate) leading to remodeled cell wall peptidoglycan and vancomycin resistance in Gram-positive pathogens.

Crystal structures of cyclophilin A complexed with cyclosporin A and N-methyl-4-[(E)-2-butenyl]-4,4-dimethylthreonine cyclosporin A.

BACKGROUND: Cyclophilin (CyP) is a ubiquitious intracellular protein that binds the immunosuppressive drug cyclosporin A (CsA). CyP-CsA forms a ternary complex with calcineurin and thereby inhibits T-cell activation. CyP also has enzymatic activity, catalyzing the cis-trans isomerization of peptidyl-prolyl amide bonds. RESULTS: We have determined the structure of human cyclophilin A (CyPA) complexed with CsA to 2.1 A resolution. We also report here the structure of CyPA complexed with an analog of CsA, CsA (MeBm2t1-CsA), which binds less well to CyPA, but has increased immunosuppressive activity. Comparison of these structures with previously determined structures of unligated CyPA and CyPA complexed with a candidate substrate for the isomerase activity, the dipeptide AlaPro, reveals that subtle conformational changes occur in both CsA and CyPA on complex formation. CONCLUSIONS: MeBm2t1-CsA binds to CyPA in an essentially similar manner to CsA. The 100-fold weaker affinity of its binding may be attributable to the close contact between MeBmt1 and the active site residue Ala103 of CyPA, which causes small conformational changes in both protein and drug. One change, the slight movement of MeLeu6 in CsA relative to MeBm2t1-CsA, may be at least partially responsible for the higher affinity of the CyPA-MeBm2t1-CsA complex for calcineurin. Our comparison between CyPA-CsA and CyPA-AlaPro suggests that CsA is probably not an analog of the natural substrate, confirming that the catalytic activity of CyPA is not related to its role in immunosuppression either structurally or functionally.

Increased activation of 1-beta-D-arabinofuranosylcytosine by hydroxyurea in L1210 cells.

Hydroxyurea (0.1 to 10 mM), incubated with L1210 cells in the presence of 1 microM 1-beta-D-arabinofuranosylcytosine (ara-C), produced a concentration- and time-dependent increase in levels of 1-beta-D-arabinofuranosylcytosine 5'-triphosphate (ara-CTP). This effect was abolished upon removal of hydroxyurea from cells. The enhancement of ara-CTP by hydroxyurea was observed for concentrations of ara-C ranging from 0.1 to 100 microM. These changes were not associated with alteration in the activity of deoxycytidine kinase, measured in extracts from L1210 cells both with and without deoxycytidine triphosphate. Hydroxyurea did, however, increase the apparent maximum velocity of this enzyme when determined with intact cell preparations, a finding consistent with alteration in endogenous regulation of this enzyme. Four-hr i.v. infusion of hydroxyurea (192 mg/kg/hr) with ara-C (1 mg/kg/hr) into ascites tumor-bearing mice also resulted in increased L1210 cell concentrations of ara-CTP. ara-C levels in plasma and L1210 cells were unaffected by hydroxyurea. Therefore, the enhancement of ara-CTP concentration in vivo, like in vitro, resulted from changes in ara-C conversion in the L1210 cells.

Initiation, elongation, and termination strategies in polyketide and polypeptide antibiotic biosynthesis.

Progress in sequence analysis of biosynthetic gene clusters encoding polyketides and nonribosomal peptides and in the reconstitution of in vitro activities continues to reveal new insights into the growth of these natural products' acyl chains, which have been revealed as a series of elongating, covalent, acyl enzyme intermediates on their multimodular scaffolds. Studies that focus on the three stages of natural product biosynthesis - initiation, elongation, and termination - have yielded crucial information on monomer substrate specificity, domain and module portability, and product release mechanisms, all of which are important not only for an understanding of this exquisite enzymatic machinery, but also for the rational construction of new, functional synthetases and synthases that are a goal of combinatorial biosynthesis.

Post-translational modification of polyketide and nonribosomal peptide synthases.

The past year has witnessed a major advance in the study of polyketide and nonribosomal peptide biosynthesis with the identification of the phosphopantetheinyl transferase enzyme family, enzymes required to produce active, post-translationally modified polyketide and peptide synthases. Phosphopantetheinyl transferases required for fatty acid, peptide and siderophore biosynthesis have been characterized and a consensus sequence noted in order to facilitate future identification of additional proteins catalyzing phosphopantetheinyl transfer.

Tailoring enzymes that modify nonribosomal peptides during and after chain elongation on NRPS assembly lines.

Nonribosomal peptide synthetases are large enzyme complexes that synthesize a variety of peptide natural products through a thiotemplated mechanism. Assembly of the peptides proceeds through amino acid loading, amide-bond formation and chain translocation, and finally thioester lysis to release the product. The final products are often heavily modified, however, through methylation, epimerization, hydroxylation, heterocyclization, oxidative cross-linking and attachment of sugars. These activities are the province of specialized enzymes (either embedded in the multidomain nonribosomal peptide synthetase structure or standalone).

D-alanine-D-alanine ligase (ADP) from Salmonella typhimurium. Overproduction, purification, crystallization and preliminary X-ray analysis.

The ddlA gene from Salmonella typhimurium coding for D-alanine-D-alanine ligase (ADP-forming) has been subcloned behind the tac promotor in the plasmid pKK223-3, with expression in Escherichia coli JM105. The overexpression system yields 58 mg of active enzyme from 12 g of wet cell paste after 40-fold purification to homogeneity. 5,5'-Dithiobis-(2-nitrobenzoic acid) titrations indicate that all four cysteine residues exist as free thiols. Two crystal forms of the 39,300 Mr enzyme have been produced. A tetragonal form grows at 21 degrees C from 10 to 15% (w/v) polyethylene glycol 8000 in space group P4(1)2(1)2, with two molecules in the asymmetric unit; it has cell constants a = b = 83.8 A, c = 220.0 A, and diffracts to 2.9 A. A monoclinic form grows from 30% (w/v) ammonium sulfate in space group P2(1), with two molecules in the asymmetric unit; it has cell constants a = 60.4 A, b = 102.1 A, c = 64.3 A, beta = 115.7 degrees, and diffracts to 2.2 A resolution.

The antibiotic microcin B17 is a DNA gyrase poison: characterisation of the mode of inhibition.

Microcin B17 is a 3.1-kDa bactericidal peptide; the putative target of this antibiotic is DNA gyrase. Microcin B17 has no detectable effect on gyrase-catalysed DNA supercoiling or relaxation activities in vitro and is unable to stabilise DNA cleavage in the absence of nucleotides. However, in the presence of ATP, or the non-hydrolysable analogue 5'-adenylyl beta,gamma-imidodiphosphate, microcin B17 stabilises a gyrase-dependent DNA cleavage complex in a manner reminiscent of quinolones, Ca(2+), or the bacterial toxin CcdB. The pattern of DNA cleavage produced by gyrase in the presence of microcin B17 is different from that produced by quinolones and more closely resembles Ca(2+)-mediated cleavage. Several gyrase mutants, including well-known quinolone-resistant mutants, are cross resistant to microcin-induced DNA cleavage. We suggest that microcin exerts its effects through a mechanism that has similarities to those of both the bacterial toxin CcdB and the quinolone antibacterial agents.

Coumarin formation in novobiocin biosynthesis: beta-hydroxylation of the aminoacyl enzyme tyrosyl-S-NovH by a cytochrome P450 NovI.

BACKGROUND: Coumarin group antibiotics, such as novobiocin, coumermycin A1 and clorobiocin, are potent inhibitors of DNA gyrase. These antibiotics have been isolated from various Streptomyces species and all possess a 3-amino-4-hydroxy-coumarin moiety as their structural core. Prior labeling experiments on novobiocin established that the coumarin moiety was derived from L-tyrosine, probably via a beta-hydroxy-tyrosine (beta-OH-Tyr) intermediate. Recently the novobiocin gene cluster from Streptomyces spheroides was cloned and sequenced and allows analysis of the biosynthesis of the coumarin at the biochemical level using overexpressed and purified proteins. RESULTS: Two open reading frames (ORFs), NovH and NovI, from the novobiocin producer S. spheroides have been overexpressed in Escherichia coli, purified and characterized for tyrosine activation and oxygenation which are the initial steps in coumarin formation. The 65 kDa NovH has two predicted domains, an adenylation (A) and a peptidyl carrier protein (PCP), reminiscent of non-ribosomal peptide synthetases. Purified NovH catalyzes L-tyrosyl-AMP formation by its A domain, can be posttranslationally phosphopantetheinylated on the PCP domain, and accumulates the covalent L-tyrosyl-S-enzyme intermediate on the holo PCP domain. The second enzyme in the pathway, NovI, is a 45 kDa heme protein that functions as a cytochrome P450-type monooxygenase with specificity for the tyrosyl-S-NovH acyl enzyme. The product beta-OH-tyrosyl-S-NovH was detected by alkaline release and high performance liquid chromatography analysis of radioactive [3H]beta-OH-Tyr and by mass spectrometry. Also detected was 4-OH-benzaldehyde, a retro aldol breakdown product of beta-OH-Tyr. The amino acid released was (3R,2S)-3-OH-Tyr by comparison with authentic standards. CONCLUSIONS: This work establishes that NovH and NovI are responsible for the formation of a beta-OH-Tyr intermediate that is covalently tethered to NovH in novobiocin biosynthesis. Comparable A-PCP/P450 pairs for amino acid beta-hydroxylation are found in various biosynthetic gene clusters, such as ORF19/ORF20 in the chloroeremomycin cluster for tyrosine, CumC/CumD in the coumermycin A1 cluster for tyrosine, and NikP1/NikQ in the nikkomycin cluster for histidine. This phenomenon of covalent docking of the amino acid in a kinetically stable thioester linkage prior to chemical modification by downstream tailoring enzymes, could represent a common strategy for controlling the partitioning of the amino acid for incorporation into secondary metabolites.

Mutational analysis of active-site residues of the enterococcal D-ala-D-Ala dipeptidase VanX and comparison with Escherichia coli D-ala-D-Ala ligase and D-ala-D-Ala carboxypeptidase VanY.

BACKGROUND: Vancomycin-resistant enterococci are pathogenic bacteria that attenuate antibiotic sensitivity by producing peptidoglycan precursors that terminate in D-Ala-D-lactate rather than D-Ala-D-Ala. A key enzyme in effecting antibiotic resistance is the metallodipeptidase VanX, which reduces the cellular pool of the D-Ala-D-Ala dipeptide. RESULTS: We constructed eleven mutants, using the recently determined VanX structure as a basis, to investigate residue function. Mutating Asp142 or Ser114 showed a large effect principally on KM, consistent with roles in recognition of the D-Ala-D-Ala termini. The drastic reduction or absence of activity in the Arg71 mutants correlates with a role in the stabilization of an anionic tetrahedral transition state. Three residues of the Escherichia coli D-Ala-D-Ala ligase (Ddl), Glu15, Ser 281 and Arg255, are similarly conserved and have equivalent functions with respect to VanX, consistent with a convergent evolution of active sites to bind D-Ala-D-Ala and lower energy barriers for formation of the tetrahedral intermediate and transition states. In the N-acyl-D-Ala-D-Ala carboxypeptidase VanY, all active-site residues are conserved (except for the two responsible for recognition of the dipeptide amino terminus). CONCLUSIONS: The mutagenesis results support structure-based functional predictions and explain why the VanX dipeptidase and Ddl ligase show narrow specificity for the D,D-dipeptide substrate. The results reveal that VanX and Ddl, two enzymes that use the same substrate but proceed in opposite directions driven by distinct cofactors (zinc versus ATP), evolved similar architectural solutions to substrate recognition and catalysis acceleration. VanY sequence analysis predicts an active site and mechanism of reaction similar to VanX.