A genomics-led approach to deciphering the mechanism of thiotetronate antibiotic biosynthesis

Thiolactomycin (TLM) is a thiotetronate antibiotic that selectively targets bacterial fatty acid biosynthesis through inhibition of the β-ketoacyl-acyl carrier protein synthases (KASI/II) that catalyse chain elongation on the type II (dissociated) fatty acid synthase. It has proved effective in in vivo infection models of Mycobacterium tuberculosis and continues to attract interest as a template for drug discovery. We have used a comparative genomics approach to uncover the (hitherto elusive) biosynthetic pathway to TLM and related thiotetronates. Analysis of the whole-genome sequence of Streptomyces olivaceus Tü 3010 producing the more ramified thiotetronate Tü 3010 provided initial evidence that such thiotetronates are assembled by a novel iterative polyketide synthase-nonribosomal peptide synthetase, and revealed the identity of other pathway enzymes, encoded by adjacent genes. Subsequent genome sequencing of three other thiotetronate-producing actinomycetes, including the Lentzea sp. ATCC 31319 that produces TLM, confirmed that near-identical clusters were also present in these genomes. In-frame gene deletion within the cluster for Tü 3010 from Streptomyces thiolactonus NRRL 15439, or within the TLM cluster, led to loss of production of the respective thiotetronate, confirming their identity. Each cluster houses at least one gene encoding a KASI/II enzyme, suggesting plausible mechanisms for self-resistance. A separate genetic locus encodes a cysteine desulfurase and a (thiouridylase-like) sulfur transferase to supply the sulfur atom for thiotetronate ring formation. Transfer of the main Tü 3010 gene cluster (stu gene cluster) into Streptomyces avermitilis led to heterologous production of this thiotetronate, showing that an equivalent sulfur donor can be supplied by this host strain. Mutational analysis of the Tü 3010 and TLM clusters has revealed the unexpected role of a cytochrome P450 enzyme in thiotetronate ring formation. These insights have allowed us to propose a mechanism for sulfur insertion, and have opened the way to engineering of the biosynthesis of TLM and other thiotetronates to produce novel analogues.

Analysis of a C-methyltransferase gene (aviG1) involved in avilamycin biosynthesis in Streptomyces viridochromogenes Tü57 and complementation of a Saccharopolyspora erythraea eryBIII mutant by aviG1

Streptomyces viridochromogenes Tü57 is the principal producer of avilamycin A. aviG1, a putative methyltransferase gene, was detected in the avilamycin biosynthetic gene cluster. To determine the function of aviG1, a targeted gene inactivation experiment was performed. The resulting chromosomal mutant, carrying an in-frame deletion in aviG1, was deficient in avilamycin production. aviG1 was used to complement an eryBIII mutant of the erythromycin A producer Saccharopolyspora erythraea [Gaisser, S., Bohm, G. A., Doumith, M., Raynal, M. C., Dhillon, N., Cortes, J. & Leadlay, P. F. (1998). Mol Gen Genet 258, 78-88]. The presence of erythromycin A in the culture supernatant of the complemented mutant indicated that L-mycarose biosynthesis could be restored and that AviG1 could take over the function of the C-methyltransferase EryBIII.

Engineering of complex polyketide biosynthesis--insights from sequencing of the monensin biosynthetic gene cluster

The biosynthesis of complex reduced polyketides is catalysed in actinomycetes by large multifunctional enzymes, the modular Type I polyketide synthases (PKSs). Most of our current knowledge of such systems stems from the study of a restricted number of macrolide-synthesising enzymes. The sequencing of the genes for the biosynthesis of monensin A, a typical polyether ionophore polyketide, provided the first genetic evidence for the mechanism of oxidative cyclisation through which polyethers such as monensin are formed from the uncyclised products of the PKS. Two intriguing genes associated with the monensin PKS cluster code for proteins, which show strong homology with enzymes that trigger double bond migrations in steroid biosynthesis by generation of an extended enolate of an unsaturated ketone residue. A similar mechanism operating at the stage of an enoyl ester intermediate during chain extension on a PKS could allow isomerisation of an E double bond to the Z isomer. This process, together with epoxidations and cyclisations, form the basis of a revised proposal for monensin formation. The monensin PKS has also provided fresh insight into general features of catalysis by modular PKSs, in particular into the mechanism of chain initiation.

Chain initiation on the soraphen-producing modular polyketide synthase from Sorangium cellulosum

BACKGROUND

Polyketides are structurally diverse natural products with a wide range of useful activities. Bacterial modular polyketide synthases (PKSs) catalyse the production of non-aromatic polyketides using a different set of enzymes for each successive cycle of chain extension. The choice of starter unit is governed by the substrate specificity of a distinct loading module. The unusual loading module of the soraphen modular PKS, from the myxobacterium Sorangium cellulosum, specifies a benzoic acid starter unit. Attempts to design functional hybrid PKSs using this loading module provide a stringent test of our understanding of PKS structure and function, since the order of the domains in the loading and first extension module is non-canonical in the soraphen PKS, and the producing strain is not an actinomycete.

RESULTS

We have constructed bimodular PKSs based on DEBS1-TE, a derivative of the erythromycin PKS that contains only extension modules 1 and 2 and a thioesterase (TE) domain, by substituting one or more domains from the soraphen PKS. A hybrid PKS containing the soraphen acyltransferase domain AT1b instead of extension acyltransferase domain AT1 produced triketide lactones lacking a methyl group at C-4, as expected if AT1b catalyses the addition of malonyl-CoA during the first extension cycle on the soraphen PKS. Substitution of the DEBS1-TE loading module AT domain by the soraphen AT1a domain led to the production of 5-phenyl-substituted triketide lactone, as well as the normal products of DEBS1-TE. This 5-phenyl triketide lactone was also the product of a hybrid PKS containing the entire soraphen PKS loading module as well as part of its first extension module. Phenyl-substituted lactone was only produced when measures were simultaneously taken to increase the intracellular supply of benzoyl-CoA in the host strain of Saccharopolyspora erythraea.

CONCLUSIONS

These results demonstrate that the ability to recruit a benzoate starter unit can be conferred on a modular PKS by the transfer either of a single AT domain, or of multiple domains to produce a chimaeric first extension module, from the soraphen PKS. However, benzoyl-CoA needs to be provided within the cell as a specific precursor. The data also support the respective roles previously assigned to the adjacent AT domains of the soraphen loading/first extension module. Construction of such hybrid actinomycete-myxobacterial enzymes should significantly extend the synthetic repertoire of modular PKSs.

New erythromycin derivatives from Saccharopolyspora erythraea using sugar O-methyltransferases from the spinosyn biosynthetic gene cluster

Using a previously developed expression system based on the erythromycin-producing strain of Saccharopolyspora erythraea, O-methyltransferases from the spinosyn biosynthetic gene cluster of Saccharopolyspora spinosa have been shown to modify a rhamnosyl sugar attached to a 14-membered polyketide macrolactone. The spnI, spnK and spnH methyltransferase genes were expressed individually in the S. erythraea mutant SGT2, which is blocked both in endogenous macrolide biosynthesis and in ery glycosyltransferases eryBV and eryCIII. Exogenous 3-O-rhamnosyl-erythronolide B was efficiently converted into 3-O-(2'-O-methylrhamnosyl)-erythronolide B by the S. erythraea SGT2 (spnI) strain only. When 3-O-(2'-O-methylrhamnosyl)-erythronolide B was, in turn, fed to a culture of S. erythraea SGT2 (spnK), 3-O-(2',3'-bis-O-methylrhamnosyl)-erythronolide B was identified in the culture supernatant, whereas S. erythraea SGT2 (spnH) was without effect. These results confirm the identity of the 2'- and 3'-O-methyltransferases, and the specific sequence in which they act, and they demonstrate that these methyltransferases may be used to methylate rhamnose units in other polyketide natural products with the same specificity as in the spinosyn pathway. In contrast, 3-O-(2',3'-bis-O-methylrhamnosyl)-erythronolide B was found not to be a substrate for the 4'-O-methyltransferase SpnH. Although rhamnosylerythromycins did not serve directly as substrates for the spinosyn methyltransferases, methylrhamnosyl-erythromycins were obtained by subsequent conversion of the corresponding methylrhamnosyl-erythronolide precursors using the S. erythraea strain SGT2 housing EryCIII, the desosaminyltransferase of the erythromycin pathway. 3-O-(2'-O-methylrhamnosyl)-erythromycin D was tested and found to be significantly active against a strain of erythromycin-sensitive Bacillus subtilis.

Engineering a polyketide with a longer chain by insertion of an extra module into the erythromycin-producing polyketide synthase

BACKGROUND

Modular polyketide synthases catalyse the biosynthesis of medically useful natural products by stepwise chain assembly, with each module of enzyme activities catalysing a separate cycle of polyketide chain extension. Domain swapping between polyketide synthases leads to hybrid multienzymes that yield novel polyketides in a more or less predictable way. No experiments have so far been reported which attempt to enlarge a polyketide synthase by interpolating additional modules.

RESULTS

We describe here the construction of tetraketide synthases in which an entire extension module from the rapamycin-producing polyketide synthase is covalently spliced between the first two extension modules of the erythromycin-producing polyketide synthase (DEBS). The extended polyketide synthases thus formed are found to catalyse the synthesis of specific tetraketide products containing an appropriate extra ketide unit. Co-expression in Saccharopolyspora erythraea of the extended DEBS multienzyme with multienzymes DEBS 2 and DEBS 3 leads to the formation, as expected, of novel octaketide macrolactones. In each case the predicted products are accompanied by significant amounts of unextended products, corresponding to those of the unaltered DEBS PKS. We refer to this newly observed phenomenon as 'skipping'.

CONCLUSIONS

The strategy exemplified here shows far-reaching possibilities for combinatorial engineering of polyketide natural products, as well as revealing the ability of modular polyketide synthases to 'skip' extension modules. The results also provide additional insight into the three-dimensional arrangement of modules within these giant synthases.

Molecular basis of Celmer's rules: role of the ketosynthase domain in epimerisation and demonstration that ketoreductase domains can have altered product specificity with unnatural substrates

BACKGROUND

Polyketides are structurally diverse natural products with a range of medically useful activities. Non-aromatic bacterial polyketides are synthesised on modular polyketide synthase multienzymes (PKSs) in which each cycle of chain extension requires a different 'module' of enzymatic activities. Attempts to design and construct modular PKSs that synthesise specified novel polyketides provide a particularly stringent test of our understanding of PKS structure and function.

RESULTS

We show that the ketoreductase (KR) domains of modules 5 and 6 of the erythromycin PKS, housed in the multienzyme subunit DEBS3, exert an unexpectedly low level of stereochemical control in reducing the keto group of a synthetic analogue of the diketide intermediate. This led us to construct a hybrid triketide synthase based on DEBS3 with ketosynthase domain ketosynthase (KS)5 replaced by the loading module and KS1. The construct in vivo produced two major triketide stereoisomers, one expected and one surprising. The latter was of opposite configuration at three out of the four chiral centres: the branching alkyl centre was that produced by KS1 and, surprisingly, both hydroxyl centres produced by the reduction steps carried out by KR5 and KR6 respectively.

CONCLUSIONS

These results demonstrate that the epimerising activity associated with module 1 of the erythromycin PKS can be conferred on module 5 merely by transfer of the KS1 domain. Moreover, the normally precise stereochemical control observed in modular PKSs is lost when KR5 and KR6 are challenged by an unfamiliar substrate, which is much smaller than their natural substrates. This observation demonstrates that the stereochemistry of ketoreduction is not necessarily invariant for a given KR domain and underlines the need for mechanistic understanding in designing genetically engineered PKSs to produce novel products.

Role of type II thioesterases: evidence for removal of short acyl chains produced by aberrant decarboxylation of chain extender units

BACKGROUND

Modular polyketide synthases (PKSs) function as molecular assembly lines in which polyketide chains are assembled by successive addition of chain extension units. At the end of the assembly line, there is usually a covalently linked type I thioesterase domain (TE I), which is responsible for release of the completed acyl chain from its covalent link to the synthase. Additionally, some PKS clusters contain a second thioesterase gene (TE II) for which there is no established role. Disruption of the TE II genes from several PKS clusters has shown that the TE II plays an important role in maintaining normal levels of antibiotic production. It has been suggested that the TE II fulfils this role by removing aberrant intermediates that might otherwise block the PKS complex.

RESULTS

We show that recombinant tylosin TE II behaves in vitro as a TE towards a variety of N-acetylcysteamine and p-nitrophenyl esters. The trends of hydrolytic activity determined by the kinetic parameter k(cat)/K(M) for the analogues tested indicates that simple fatty acyl chains are effective substrates. Analogues that modelled aberrant forms of putative tylosin biosynthetic intermediates were hydrolysed at low rates.

CONCLUSIONS

The behaviour of tylosin TE II in vitro is consistent with its proposed role as an editing enzyme. Aberrant decarboxylation of a malonate-derived moiety attached to an acyl carrier protein (ACP) domain may generate an acetate, propionate or butyrate residue on the ACP thiol. Our results suggest that removal of such groups is a significant role of TE II.

Protection of radical intermediates at the active site of adenosylcobalamin-dependent methylmalonyl-CoA mutase

Adenosylcobalamin-dependent methylmalonyl-CoA mutase catalyzes the interconversion of methylmalonyl-CoA and succinyl-CoA via radical intermediates generated by substrate-induced homolysis of the coenzyme carbon-cobalt bond. From the structure of methylmalonyl-CoA mutase it is evident that the deeply buried active site is completely shielded from solvent with only a few polar contacts made between the protein and the substrate. Site-directed mutants of amino acid His244, a residue close to the inferred site of radical chemistry, were engineered to investigate its role in catalysis. Two mutants, His244Ala and His244Gln, were characterized using kinetic and spectroscopic techniques. These results confirmed that His244 is not an essential residue. However, compared with that of the wild type, k(cat) was lowered by 10(2)- and 10(3)-fold for the His244Gln and His244Ala mutants, respectively, while the K(m) for succinyl-CoA was essentially unchanged in both cases. The primary kinetic tritium isotope effect (k(H)/k(T)) for the His244Gln mutant was 1.5 +/- 0.3, and tritium partitioning was now found to be dependent on the substrate used to initiate the reaction, indicating that the rearrangement of the substrate radical to the product radical was extremely slow. The His244Ala mutant underwent inactivation under aerobic conditions at a rate between 1 and 10% of the initial rate of turnover. The crystal structure of the His244Ala mutant, determined at 2.6 A resolution, indicated that the mutant enzyme is unaltered except for a cavity in the active site which is occupied by an ordered water molecule. Molecular oxygen reaching this cavity may lead directly to inactivation. These results indicate that His244 assists directly in the unusual carbon skeleton rearrangement and that alterations in this residue substantially lower the protection of reactive radical intermediates during catalysis.

A defined system for hybrid macrolide biosynthesis in Saccharopolyspora erythraea

The biological activity of polyketide antibiotics is often strongly dependent on the presence and type of deoxysugar residues attached to the aglycone core. A system is described here, based on the erythromycin-producing strain of Saccharopolyspora erythraea, for detection of hybrid glycoside formation, and this system has been used to demonstrate that an amino sugar characteristic of 14-membered macrolides (D-desosamine) can be efficiently attached to a 16-membered aglycone substrate. First, the S. erythraea mutant strain DM was created by deletion of both eryBV and eryCIII genes encoding the respective ery glycosyltransferase genes. The glycosyltransferase OleG2 from Streptomyces antibioticus, which transfers L-oleandrose, has recently been shown to transfer rhamnose to the oxygen at C-3 of erythronolide B and 6-deoxyerythronolide B. In full accordance with this finding, when oleG2 was expressed in S. erythraea DM, 3-O-rhamnosyl-erythronolide B and 3-O-rhamnosyl-6-deoxyerythronolide B were produced. Having thus validated the expression system, endogenous aglycone production was prevented by deletion of the polyketide synthase (eryA) genes from S. erythraea DM, creating the triple mutant SGT2. To examine the ability of the mycaminosyltransferase TylM2 from Streptomyces fradiae to utilise a different amino sugar, tylM2 was integrated into S. erythraea SGT2, and the resulting strain was fed with the 16-membered aglycone tylactone, the normal TylM2 substrate. A new hybrid glycoside was isolated in good yield and characterized as 5-O-desosaminyl-tylactone, indicating that TylM2 may be a useful glycosyltransferase for combinatorial biosynthesis. 5-O-glucosyl-tylactone was also obtained, showing that endogenous activated sugars and glycosyltransferases compete for aglycone in these cells.

Novel octaketide macrolides related to 6-deoxyerythronolide B provide evidence for iterative operation of the erythromycin polyketide synthase

BACKGROUND

The macrolide antibiotic erythromycin A, like other complex aliphatic polyketides, is synthesised by a bacterial modular polyketide synthase (PKS). Such PKSs, in contrast to other fatty acid and polyketide synthases which work iteratively, contain a separate set or module of enzyme activities for each successive cycle of polyketide chain extension, and the number and type of modules together determine the structure of the polyketide product. Thus, the six extension modules of the erythromycin PKS (DEBS) together catalyse the production of the specific heptaketide 6-deoxyerythronolide B.

RESULTS

A mutant strain of the erythromycin producer Saccharopolyspora erythraea, which accumulates the aglycone intermediate erythronolide B, was found unexpectedly to produce two novel octaketides, both 16-membered macrolides. These compounds were detectable in fermentation broths of wild-type S. erythraea, but not in a strain from which the DEBS genes had been specifically deleted. From their structures, both of these octaketides appear to be aberrant products of DEBS in which module 4 has 'stuttered', that is, has catalysed two successive cycles of chain extension.

CONCLUSIONS

The isolation of novel DEBS-derived octaketides provides the first evidence that an extension module in a modular PKS has the potential to catalyse iterative rounds of chain elongation like other type I FAS and PKS systems. The factors governing the extent of such 'stuttering' remain to be determined.

Efficient purification and kinetic characterization of a bimodular derivative of the erythromycin polyketide synthase

Modular polyketide synthases (PKSs), such as the 6-deoxyerythronolide B synthase (DEBS), are giant multienzymes that biosynthesize a number of clinically important natural products. The modular nature of PKSs suggests the possibility of a combinatorial approach to the synthesis of novel bioactive polyketides, but the efficacy of such a strategy depends critically on gaining fundamental insight into PKS structure and function, most directly through experiments with purified PKS proteins. Several recent investigations into important aspects of the activity of these enzymes have used only partially purified proteins (often 3-4% of total protein), reflecting how difficult it is to purify these multienzymes in amounts adequate for kinetic and structural analysis. We report here the steady-state kinetic analysis of a typical bimodular PKS, 6-deoxyerythronolide B synthase 1-thioesterase (DEBS 1-TE), purified from recombinant Saccharopolyspora erythraea JCB101 by a new, high-yielding procedure consisting of three steps: ammonium sulfate precipitation, hydrophobic interaction chromatography and size-exclusion chromatography. The method provides 13-fold purification with a recovery of 11% of the applied PKS activity. The essentially homogeneous synthase exhibits an intrinsic methylmalonyl-CoA hydrolase activity, which competes with polyketide chain extension. The most reliable value for the kcat for synthesis of (3S,5R)-dihydroxy-(2R,4R)-dimethyl-n-heptanoic acid-delta-lactone is 0.84 min-1, and the apparent Km for (2RS)-methylmalonyl-CoA is 17 microM. This kcat is approximately 10-fold lower than the value reported previously for a differently engineered version of the truncated PKS, DEBS 1+TE. The difference likely reflects the fact that the DEBS 1-TE contains a hybrid acyl carrier protein (ACP) domain in its second module, which lowers its catalytic efficiency.

Knowledge-based design of bimodular and trimodular polyketide synthases based on domain and module swaps: a route to simple statin analogues

BACKGROUND

Polyketides are structurally diverse natural products that have a range of medically useful activities. Nonaromatic bacterial polyketides are synthesised on modular polyketide synthase (PKS) multienzymes, in which each cycle of chain extension requires a different 'module' of enzymatic activities. Attempts to design and construct modular PKSs that synthesise specified novel polyketides provide a particularly stringent test of our understanding of PKS structure and function.

RESULTS

We have constructed bimodular and trimodular PKSs based on DEBS1-TE, a derivative of the erythromycin PKS that contains only modules 1 and 2 and a thioesterase (TE), by substituting multiple domains with appropriate counterparts derived from the rapamycin PKS. Hybrid PKSs were obtained that synthesised the predicted target triketide lactones, which are simple analogues of cholesterol-lowering statins. In constructing intermodular fusions, whether between modules in the same or in different proteins, it was found advantageous to preserve intact the acyl carrier protein-ketosynthase (ACP-KS) didomain that spans the junction between successive modules.

CONCLUSIONS

Relatively simple considerations govern the construction of functional hybrid PKSs. Fusion sites should be chosen either in the surface-accessible linker regions between enzymatic domains, as previously revealed, or just inside the conserved margins of domains. The interaction of an ACP domain with the adjacent KS domain, whether on the same polyketide or not, is of particular importance, both through conservation of appropriate protein-protein interactions, and through optimising molecular recognition of the altered polyketide chain in the key transfer of the acyl chain from the ACP of one module to the KS of the downstream module.

A chain initiation factor common to both modular and aromatic polyketide synthases

Antibiotic-producing polyketide synthases (PKSs) are enzymes responsible for the biosynthesis in Streptomyces and related filamentous bacteria of a remarkably broad range of bioactive metabolites, including antitumour aromatic compounds such as mithramycin and macrolide antibiotics such as erythromycin. The molecular basis for the selection of the starter unit on aromatic PKSs is unknown. Here we show that a component of aromatic PKS, previously named 'chain-length factor', is a factor required for polyketide chain initiation and that this factor has decarboxylase activity towards malonyl-ACP (acyl carrier protein). We have re-examined the mechanism of initiation on modular PKSs and have identified as a specific initiation factor a domain of previously unknown function named KSQ, which operates like chain-length factor. Both KSQ and chain-length factor are similar to the ketosynthase domains that catalyse polyketide chain extension in modular multifunctional PKSs and in aromatic PKSs, respectively, except that the ketosynthase domain active-site cysteine residue is replaced by a highly conserved glutamine in KSQ and in chain-length factor. The glutamine residue is important both for decarboxylase activity and for polyketide synthesis.

Molecular basis of Celmer's rules: the role of two ketoreductase domains in the control of chirality by the erythromycin modular polyketide synthase

BACKGROUND

Polyketides are compounds that possess medically significant activities. The modular nature of the polyketide synthase (PKS) multienzymes has generated interest in bioengineering new PKSs. Rational design of novel PKSs, however, requires a greater understanding of the stereocontrol mechanisms that operate in natural PKS modules.

RESULTS

The N-acetyl cysteamine (NAC) thioester derivative of the natural beta-keto diketide intermediate was incubated with DEBS1-TE, a derivative of the erythromycin PKS that contains only modules 1 and 2. The reduction products of the two ketoreductase (KR) domains of DEBS1-TE were a mixture of the (2S, 3R) and (2R,3S) isomers of the corresponding beta-hydroxy diketide NAC thioesters. Repeating the incubation using a DEBS1-TE mutant that only contains KR1 produced only the (2S,3R) isomer.

CONCLUSIONS

In contrast with earlier results, KR1 selects only the (2S) isomer and reduces it stereospecifically to the (2S, 3R)-3-hydroxy-2-methyl acyl product. The KR domain of module 1 controls the stereochemical outcome at both methyl-and hydroxyl-bearing chiral centres in the hydroxy diketide intermediate. Earlier work showed that the normal enzyme-bound ketoester generated in module 2 is not epimerised, however. The stereochemistry at C-2 is therefore established by a condensation reaction that exclusively gives the (2R)-ketoester, and the stereo-chemistry at C-3 by reduction of the keto group. Two different mechanisms of stereochemical control, therefore, operate in modules 1 and 2 of the erythromycin PKS. These results should provide a more rational basis for designing hybrid PKSs to generate altered stereochemistry in polyketide products.

Structural elucidation studies of erythromycins by electrospray tandem mass spectrometry

Erythromycin A (EryA) was studied by electrospray ionisation tandem mass spectrometry (ESI-MS/MS) with the aim of developing a methodology for the structural elucidation of novel erythromycins developed by biological synthetic methods. Skimmer dissociation along with sequential mass spectrometry studies (up to MS5) have been employed in this study. In the low-resolution MS/MS analysis of the polyketides, there are several fragment ions that are easily assigned to various neutral losses. These have all been confirmed by accurate-mass measurements. There is also a series of peaks due to ring opening and fragmentation that can only be assigned by high-resolution MSn analysis. Further experiments were performed in deuterated media (D2O/CD3OD 50%) which, along with the high-resolution MSn of erythromycin analogues, has enabled us to identify some of the steps in the ring fragmentation, particularly the loss of the polyketide starter acid. This is an essential step for determining structural alterations in the novel polyketides, but further labelling experiments and studies on more erythromycin analogues are required before the complete fragmentation pathway can be confirmed.

Structural elucidation studies of erythromycins by electrospray tandem mass spectrometry II

Erythromycin A (EryA), sec-butyl erythromycin B (SEryB), oleandomycin (Olean) and a synthetic derivative, roxithromycin (Rox), were used to investigate the fragmentation of polyketide macrolide antibiotics by collision induced dissociation (CID) tandem mass spectrometry (MS/MS). Analyses were performed with two commercially available mass spectrometers: a Q-TOF hybrid quadrupole time-of-flight instrument and a BioApex II (4.7 Tesla) Fourier transform ion cyclotron resonance (FTICR) instrument both equipped with electrospray ionisation (ESI) sources. One of the first fragmentation processes is the loss of an H(2)O molecule from the [M+H](+) ion. EryA has three hydroxyl groups on the polyketide ring and loses three H(2)O molecules during CID. This study indicates that these facts are not necessarily related. Deuterium exchange experiments were carried out in order to isotopically label free hydroxyl groups. (18)O-exchange experiments were also carried out in order to label the carbonyl group at the 9-position. In EryA and its analogue the first H(2)O loss shifts in mass from loss of 18 Da to loss of 20 Da in deuterated solvents. For both molecules the loss also shifts in mass from loss of 18 Da to loss of 20 Da during the (18)O-exchange experiments. This suggests that the first loss of H(2)O is from the 9-position carbonyl group, indicating that this, and not the nitrogen of the amino sugar, is the site of protonation of the activated MH(+) ions. For Rox the initial loss of H(2)O is replaced by loss of the 9-position oxime group, the rest of the fragmentation sequence being the same as for EryA. For Olean, there is no H(2)O loss from the parent ion. The results have allowed the proposal of a mechanism for the first loss of H(2)O in the EryA MS/MS fragmentation.

Structural elucidation studies of polyketide tetrasubstituted delta-lactones by gas chromatography/tandem mass spectrometry and electrospray mass spectrometry

A series of tetrasubstituted polyketide delta-lactones were used to evaluate whether gas chromatography/tandem mass spectrometry (GC/MS/MS) and electrospray mass spectrometry (ESI-MS) are useful techniques for probing the structure and stereochemistry of such highly functionalised molecules. Analyses were performed with two commercially available mass spectrometers: a Finnigan/MAT GCQ instrument (CI source) and a Q-TOF Hybrid quadrupole time-of-flight instrument (ESI source). The analyses revealed that a range of variation in the structure and stereochemistry of the lactones did not affect the fragmentation pathway common to these molecules. By accurate mass determination (ESI-MS), the first two fragmentations were assigned to losses of water. Although it was anticipated that the initial dehydration would include the hydroxyl group at the 3-position of the lactones, evidence from deuterium- and (18)O-labelling studies suggests that the losses of water instead involve the oxygen atoms in the ester bond. Attempts to identify further the structures of daughter ions by GC/MS/MS were complicated by extensive rearrangements and non-specific hydrogen/deuterium migrations within the lactones. Together, these results illustrate the limitations of mass spectrometry in the structural elucidation of complex molecules. Copyright 1999 John Wiley & Sons, Ltd.

Evaluating precursor-directed biosynthesis towards novel erythromycins through in vitro studies on a bimodular polyketide synthase

BACKGROUND

Modular polyketide synthases (PKSs) catalyse the biosynthesis of complex polyketides using a different set of enzymes for each successive cycle of chain extension. Directed biosynthesis starting from synthetic diketides is a potentially valuable route to novel polyketides. We have used a purified bimodular derivative of the erythromycin-producing polyketide synthase (DEBS 1-TE) to study chain extension starting from a variety of diketide analogues and, in some cases, from the alternative acyl-CoA thioester substrates.

RESULTS

Chain initiation in vitro by DEBS 1-TE module 2 using a synthetic diketide analogue as a substrate was tolerant of significant structural variation in the starter unit of the synthetic diketide, but other changes completely abolished activity. Interestingly, a racemic beta-keto diketide was found to be reduced in situ on the PKS and utilised in place of its more complex hydroxy analogue as a substrate for chain extension. The presence of a diketide analogue strongly inhibited chain initiation via the loading module. Significantly higher concentrations of diketide N-acetylcysteamine analogues than their corresponding acyl-CoA thioesters are required to achieve comparable yields of triketide lactones.

CONCLUSIONS

Although a broad range of variation in the starter residue is acceptable, the substrate specificity of module 2 of a typical modular PKS in vitro is relatively intolerant of changes at C-2 and C-3. This will restrict the usefulness of approaches to synthesise novel erythromycins using synthetic diketides in vivo. The use of synthetic beta-keto diketides in vivo deserves to be explored.

Novel erythromycins from a recombinant Saccharopolyspora erythraea strain NRRL 2338 pIG1. I. Fermentation, isolation and biological activity

In a previous report, a plasmid, pIG1, which contained the loading domain from the Streptomyces avermitilis polyketide synthase (PKS), promoters from Streptomyces coelicolor and the DEBS1-TE truncated PKS from Saccharopolyspora erythraea, was integrated into the S. erythraea chromosome, effectively replacing the natural erythromycin loading domain with the avermectin loading domain. In this paper, we report the feeding of short-chained fatty acids to this recombinant strain, and its parent, NRRL 2338. Both strains incorporated exogenously supplied fatty acids to produce novel, biologically active, C-13 substituted erythromycins.

Stabilization of radical intermediates by an active-site tyrosine residue in methylmalonyl-CoA mutase

The adenosylcobalamin-dependent methylmalonyl-CoA mutase catalyzes the reversible rearrangement of methylmalonyl-CoA into succinyl-CoA by a free-radical mechanism. The recently solved X-ray crystal structure of methylmalonyl-CoA mutase from Propionibacterium shermanii has shown that tyrosine 89 is an active-site residue involved in substrate binding. The role of tyrosine 89, a conserved residue among methylmalonyl-CoA mutases, has been investigated by using site-directed mutagenesis to replace this residue with phenylalanine. The crystal structure of the Tyr89Phe mutant was determined to 2.2 A resolution and was found to be essentially superimposable on that of wild-type. Mutant and wild-type enzyme have very similar KM values, but kcat for the Tyr89Phe mutant is 580-fold lower than for wild-type. The rate of release of tritium from 5'-[3H]adenosylcobalamin during the enzymatic reaction and its rate of appearance in substrate and product were measured. The tritium released was found to partition unequally between methylmalonyl-CoA and succinyl-CoA, in a ratio of 40:60 when the reaction was initiated by addition of methylmalonyl-CoA and in a ratio of 10:90 when the reaction was initiated by addition of succinyl-CoA. The overall release of tritium was four times faster when succinyl-CoA was used as substrate. The tritium isotope effect on the enzyme catalyzed hydrogen transfer, measured with methylmalonyl-CoA as a substrate, was kH/kT = 30, which is within the expected range for a full primary kinetic tritium isotope effect. The different partitioning of tritium, dependent upon which substrate was used, and the normal value for the kinetic tritium isotope effect contrast markedly with the behavior of wild-type mutase. It appears that the loss of a single interaction involving the hydroxyl group of tyrosine 89 both affects the stability of radical intermediates and decreases the rate of interconversion of the substrate- and product-derived radicals.

Phosphorylation of the periplasmic binding protein in two transport systems for arginine incorporation in Escherichia coli K-12 is unrelated to the function of the transport system

In Escherichia coli K-12, the accumulation of arginine is mediated by two distinct periplasmic binding protein-dependent transport systems, one common to arginine and ornithine (AO system) and one for lysine, arginine, and ornithine (LAO system). Each of these systems includes a specific periplasmic binding protein, the AO-binding protein for the AO system and the LAO-binding protein for the LAO system. The two systems include a common inner membrane transport protein which is able to hydrolyze ATP and also phosphorylate the two periplasmic binding proteins. Previously, a mutant resistant to the toxic effects of canavanine, with low levels of transport activities and reduced levels of phosphorylation of the two periplasmic binding proteins, was isolated and characterized (R. T. F. Celis, J. Biol. Chem. 265:1787-1793, 1990). The gene encoding the transport ATPase enzyme (argK) has been cloned and sequenced. The gene possesses an open reading frame with the capacity to encode 268 amino acids (mass of 29.370 Da). The amino acid sequence of the protein includes two short sequence motifs which constitute a well-defined nucleotide-binding fold (Walker sequences A and B) present in the ATP-binding subunits of many transporters. We report here the isolation of canavanine-sensitive derivatives of the previously characterized mutant. We describe the properties of these suppressor mutations in which the transport of arginine, ornithine, and lysine has been restored. In these mutants, the phosphorylation of the AO- and LAO-binding proteins remains at a low level. This information indicates that whereas hydrolysis of ATP by the transport ATPase is an obligatory requirement for the accumulation of these amino acids in E. coli K-12, the phosphorylation of the periplasmic binding protein is not related to the function of the transport system.

Construction of new vectors for high-level expression in actinomycetes

A new integrative vector (pCJR24) was constructed for use in the erythromycin producer Saccharopolyspora erythraea and in other actinomycetes. It includes the pathway-specific activator gene actII-ORF4 from the actinorhodin biosynthetic gene cluster of Streptomyces coelicolor. The actI promoter and the associated ribosome binding site are located upstream of an NdeI site (5'-CATATG-3') which encompasses the actI start codon allowing protein(s) to be produced at high levels in response to nutritional signals if these signals are faithfully mediated by the ActII-ORF4 activator. Several polyketide synthase genes were cloned in pCJR24 and overexpressed in S. erythraea after integration of the vector into the chromosome by homologous recombination, indicating the possibility that the S. coelicolor promoter/activator functions appropriately in S. erythraea. pCJR24-mediated recombination was also used to place the entire gene set for the erythromycin-producing polyketide synthase under the control of the actI promoter. The resulting strain produced copious quantities of erythromycins and precursor macrolides when compared with wild-type S. erythraea. The use of this system provides the means for rational strain improvement of antibiotic-producing actinomycetes.

Origin of starter units for erythromycin biosynthesis

Modular polyketide synthases (PKSs), such as the 6-deoxyerythronolide B synthase (DEBS), are multifunctional proteins that govern the synthesis of a number of clinically important natural products. The modular arrangement of active sites within these enzymes suggests the possibility of a combinatorial approach to the synthesis of novel bioactive polyketides. The efficacy of combinatorial strategies toward altering the starter unit specificity of polyketide synthases critically depends on controlling the supply of competing endogenous starter acids. Using DEBS 1-TE, a bimodular derivative of DEBS, we aimed to determine whether the beta-ketosynthase (KS) domain responsible for condensation in the first module also has the ability to prime its own biosynthesis by catalyzing the decarboxylation of methylmalonyl-CoA to produce propionyl-CoA. In contrast to earlier reports with a closely similar mini-PKS DEBS 1+TE, we have found that rigorously purified DEBS 1-TE does not catalyze the decarboxylation of methylmalonyl-CoA.

Engineering of a minimal modular polyketide synthase, and targeted alteration of the stereospecificity of polyketide chain extension

BACKGROUND

Polyketides are a large and structurally diverse group of natural products that include antibiotics, antifungal agents and immunosuppressant compounds. Polyketides are biosynthesised in filamentous bacteria on modular polyketide synthases (PKSs) in which each cycle of chain extension requires a different 'module' of enzymatic activities. The recently proposed dimeric model for modular PKSs predicts that even a single-module PKS should be catalytically active in the absence of other PKS components. Researchers are also interested in manipulating the stereochemical outcome of polyketide chain extension using genetic engineering of domains within each module.

RESULTS

We have constructed a minimal modular PKS from the erythromycin-producing PKS (DEBS) of Saccharopolyspora erythraea. The diketide synthase (DKS1-2) consists of a single chimaeric extension module, derived from the DEBS module 1 ketoacyl-ACP synthase (KS), sandwiched between a loading module and a chain-terminating thioesterase. When DKS1-2 was expressed in S. erythraea, the strain preferentially6 accumulated the diketide (2R, 3S)-2-methyl-3-hydroxy pentanoic acid.

CONCLUSIONS

These results demonstrate that, as predicted, even a single-module PKS is catalytically active in the absence of other DEBS proteins. In its normal context, the ketosynthase domain KS1 is thought to generate a (2S)-2methyl-3-hydroxy intermediate by epimerising the initial product of carbon-carbon chain formation, the (2R)-2-methyl-3-ketoester. The observed formation of the alternative (2R)-methyl-3-hydroxy product catalysed by DKS1-2 provides strong support for this proposal, and indicates how targeted alteration of stereospecificity can be achieved on a modular PKS.

Analysis of eryBI, eryBIII and eryBVII from the erythromycin biosynthetic gene cluster in Saccharopolyspora erythraea

The gene cluster (ery) governing the biosynthesis of the macrolide antibiotic erythromycin A by Saccharopolyspora erythraea contains, in addition to the eryA genes encoding the polyketide synthase, two regions containing genes for later steps in the pathway. The region 5' of eryA that lies between the known genes ermE (encoding the erythromycin resistance methyltransferase) and eryBIII (encoding a putative S-adenosylmethionine-dependent methyltransferase), and that contains the gene eryBI (orf2), has now been sequenced. The inferred product of the eryBI gene shows striking sequence similarity to authentic beta-glucosidases. Specific mutants were created in eryBI, and the resulting strains were found to synthesise erythromycin A, showing that this gene, despite its position in the biosynthetic gene cluster, is not essential for erythromycin biosynthesis. A mutant in eryBIII and a double mutant in eryBI and eryBIII were obtained and the analysis of novel erythromycins produced by these strains confirmed the proposed function of EryBIII as a C-methyltransferase. Also, a chromosomal mutant was constructed for the previously sequenced ORF19 and shown to accumulate erythronolide B, as expected for an eryB mutant and consistent with its proposed role as an epimerase in dTDP-mycarose biosynthesis.

Targeted gene inactivation for the elucidation of deoxysugar biosynthesis in the erythromycin producer Saccharopolyspora erythraea

The production of erythromycin A by Saccharopolyspora erythraea requires the synthesis of dTDP-D-desosamine and dTDP-L-mycarose, which serve as substrates for the transfer of the two sugar residues onto the macrolactone ring. The enzymatic activities involved in this process are largely encoded within the ery gene cluster, by two sets of genes flanking the eryA locus that encodes the polyketide synthase. We report here the nucleotide sequence of three such ORFs located immediately downstream of eryA, ORFs 7, 8 and 9. Chromosomal mutants carrying a deletion either in ORF7 or in one of the previously sequenced ORFs 13 and 14 have been constructed and shown to accumulate erythronolide B, as expected for eryB mutants. Similarly, chromosomal mutants carrying a deletion in either ORF8, ORF9, or one of the previously sequenced ORFs 17 and 18 have been constructed and shown to accumulate 3-alpha-mycarosyl erythronolide B, as expected for eryC mutants. The ORF13 (eryBIV), ORF17 (eryCIV) and ORF7 (eryBII) mutants also synthesised small amounts of macrolide shunt metabolites, as shown by mass spectrometry. These results considerably strengthen previous tentative proposals for the pathways for the biosynthesis of dTDP-D-desosamine and dTDP-L-mycarose in Sac. erythraea and reveal that at least some of these enzymes can accommodate alternative substrates.

Mutational biosynthesis of novel rapamycins by a strain of Streptomyces hygroscopicus NRRL 5491 disrupted in rapL, encoding a putative lysine cyclodeaminase

The gene rapL lies within the region of the Streptomyces hygroscopicus chromosome which contains the biosynthetic gene cluster for the immunosuppressant rapamycin. Introduction of a frameshift mutation into rapL by phiC31 phage-mediated gene replacement gave rise to a mutant which did not produce significant amounts of rapamycin. Growth of this rapL mutant on media containing added L-pipecolate restored wild-type levels of rapamycin production, consistent with a proposal that rapL encodes a specific L-lysine cyclodeaminase important for the production of the L-pipecolate precursor. In the presence of added proline derivatives, rapL mutants synthesized novel rapamycin analogs, indicating a relaxed substrate specificity for the enzyme catalyzing pipecolate incorporation into the macrocycle.

Engineering broader specificity into an antibiotic-producing polyketide synthase

The wide-specificity loading module for the avermectin-producing polyketide synthase was grafted onto the first multienzyme component (DEBS1) of the erythromycin-producing polyketide synthase in place of the normal loading module. Expression of this hybrid enzyme in the erythromycin producer Saccharopolyspora erythraea produced several novel antibiotic erythromycins derived from endogenous branched-chain acid starter units typical of natural avermectins. Because the avermectin polyketide synthase is known to accept more than 40 alternative carboxylic acids as starter units, this approach opens the way to facile production of novel analogs of antibiotic macrolides.

The molecular basis of Celmer's rules: the stereochemistry of the condensation step in chain extension on the erythromycin polyketide synthase

Modular polyketide synthases (PKSs), for example, the 6-deoxyerythronolide B synthase (DEBS) responsible for synthesis of the aglycone core of the macrolide antibiotic erythromycin, generate an impressive diversity of asymmetric centers in their polyketide products. However, as noted by Celmer, macrolides have the same absolute configuration at all comparable stereocenters. Understanding how the stereochemistry of chain extension in controlled is therefore crucial to determining the common mechanism of action of these enzymes. We aimed to elucidate the molecular basis of Celmer's rules through in vitro studies with DEBS 1-TE, a bimodular derivative of DEBS from Saccharopolyspora erythraea, which uses (2S)-methylmalonyl-coenzyme. A to produce both D- and L-methyl centers in its triketide lactone product. We show here that condensation of (2S)-methylmalonyl-CoA in module 2 proceeds with decarboxylative inversion without cleavage of the C-H bond adjacent to the methyl group; in contrast, in module 1 the chain extension process involves loss of the hydrogen attached to C-2 of the methylmalonyl-CoA precursor. The production of the D-methyl center in module 2 without loss of hydrogen from the asymmetric center of the (2S)-methylmalonyl-CoA establishes that condensation takes place with inversion of configuration as in fatty acid biosynthesis. The loss of the key hydrogen from the (2S)-methylmalonyl-CoA to produce the L-methyl center generated in module 1 implies that an additional obligatory epimerization step takes place in that module. The nature and timing of the epimerization remain to be established.

Analysis of seven genes from the eryAI-eryK region of the erythromycin biosynthetic gene cluster in Saccharopolyspora erythraea

The gene cluster (ery) governing the biosynthesis of the macrolide antibiotic erythromycin A by Saccharopolyspora erythraea contains, in addition to the eryA genes encoding the polyketide synthase, two regions containing genes for later steps in the pathway. The region 5' of eryA, and lying between eryA and the gene eryK, which is known to encode the C-12 hydroxylase, has been sequenced and shown to contain seven additional open reading frames (ORFs 13-19). On the basis of sequence similarities, roles are proposed for several of these ORFs in the biosynthesis of the deoxysugar mycarose and the deoxyaminosugar desosamine. A chromosomal mutant carrying a deletion in ORF15 has been constructed and shown to accumulate 3-O-mycarosylerythronolide B, as expected for an eryC mutant. Similarly, a chromosomal mutant carrying a deletion in ORF16 has been constructed and shown to accumulate erythronolide B, as expected for an eryB mutant.

Combinatorial approaches to polyketide biosynthesis

Polyketides are a large and structurally diverse family of natural products based on chains of carboxylic acid units. The polyketide synthases that make aromatic polyketides have already been used to generate small combinatorial libraries, by expressing individual genes from different biosynthetic pathways together, so that the enzymes they encode can interact to make novel products. Recent work has shown how to choose these individual components to increase the chances of obtaining such hybrid aromatic compounds. In other polyketide synthases which synthesise complex reduced polyketides, the constituent enzymes are actually linked as domains in a giant multienzyme complex along which the growing polyketide chain is passed. A combinatorial approach here therefore requires the fusing together of individual enzymatic domains from several such synthases in as many productive ways as can be devised, so that the enzyme assembly line produces a library of altered products. A key recent advance has been to demonstrate that such genuinely hybrid enzymes do work as predicted, for example a broad-specificity enzyme that recruits the chain starter unit for an antiparasitic compound has been grafted onto a synthase that makes antimicrobial macrolides.

The pipecolate-incorporating enzyme for the biosynthesis of the immunosuppressant rapamycin--nucleotide sequence analysis, disruption and heterologous expression of rapP from Streptomyces hygroscopicus

An open reading frame (rapP) encoding the putative pipecolate-incorporating enzyme (PIE) has been identified in the gene cluster for the biosynthesis of rapamycin in Streptomyces hygroscopicus. Conserved amino acid sequence motifs for ATP binding, ATP hydrolysis, adenylate formation, and 4'-phosphopantetheine attachment were identified by sequence comparison with authentic peptide synthetases. Disruption of rapP by phage insertion abolished rapamycin production in S. hygroscopicus, and the production of the antibiotic was specifically restored upon loss of the inserted phage by a second recombination event. rapP was expressed in both Escherichia coli and Streptomyces coelicolor, and recombinant PIE was purified to homogeneity from both hosts. Although low-level incorporation of [14C]beta-alanine into recombinant PIE isolated from E. coli was detected, formation of the covalent acylenzyme intermediate could only be shown with the PIE from S. coelicolor, suggesting that while the recombinant PIE from S. coelicolor was phosphopantetheinylated, only a minor proportion of the recombinant enzyme from E. coli was post-translationally modified.

A hybrid modular polyketide synthase obtained by domain swapping

BACKGROUND

Modular polyketide synthases govern the synthesis of a number of medically important antibiotics, and there is therefore great interest in understanding how genetic manipulation may be used to produce hybrid synthases that might synthesize novel polyketides. In particular, we aimed to show whether an individual domain can be replaced by a comparable domain from a different polyketide synthase to form a functional hybrid enzyme. To simplify the analysis, we have used our previously-developed model system DEBS1-TE, consisting of the first two chain-extension modules of the erythromycin-producing polyketide synthase of Saccharopolyspora erythraea.

RESULTS

We show here that replacing the entire acyltransferase (AT) domain from module 1 of DEBS1-TE by the AT domain from module 2 of the rapamycin-producing polyketide synthase leads, as predicted, to the synthesis of two novel triketide lactones in good yield, in place of the two lactones produced by DEBS1-TE. Both of the novel products specifically lack a methyl group at C-4 of the lactone ring.

CONCLUSIONS

Although the AT domain is a core structural domain of a modular polyketide synthase, it has been swapped to generate a truly hybrid multienzyme with a rationally altered specificity of chain extension. Identical manipulations carried out on known polyketide antibiotics might therefore generate families of potentially useful analogues that are inaccessible by chemical synthesis. These results also encourage the belief that other domains may be similarly swapped.

Tritium isotope effects in adenosylcobalamin-dependent methylmalonyl-CoA mutase

Methylmalonyl-CoA mutase from Propionibacterium shermanii is an adenosylcobalamin-dependent enzyme which catalyzes the reversible isomerization of methylmalonyl-CoA and succinyl-CoA. The rate of tritium loss from 5'-[3H]adenosylcobalamin during the enzymic reaction and the relative rates of tritium appearance in substrate and product were examined. Upon the addition of methylmalonyl-CoA to a solution of holoenzyme, tritium was completely released from the cofactor within about 500 ms. No tritium was found either bound to the enzyme or released into the water. The radioactivity was found in methylmalonyl-CoA and succinyl-CoA in a constant ratio of 1 to 3, which did not change during the first 300 ms of the reaction. Upon the addition of succinyl-CoA to a solution of holoenzyme, tritium was released at essentially the same rate, and the radioactivity was found in methylmalonyl-CoA and succinyl-CoA in the identical constant ratio of 1 to 3. The tritium isotope effect on the enzyme-catalyzed hydrogen transfer, measured using 14C-labeled methylmalonyl-CoA as substrate, was kH/kT = 4.9. This low value shows that hydrogen transfer is only partly rate limiting and that at least one subsequent slow step, such as product release, contributes substantially to the overall reaction velocity. The identical partitioning of tritium, regardless of the substrate used, shows that the rearrangement of the substrate radical into the product radical is not rate limiting. The very low tritium isotope effect and the fact that all the tritium is found bound either to the CoA esters or to the cofactor make it very unlikely that a protein radical is an intermediate in the methylmalonyl-CoA mutase-catalyzed rearrangement.

Homology modeling of human methylmalonyl-CoA mutase: a structural basis for point mutations causing methylmalonic aciduria

Point mutations in the human gene encoding coenzyme B12 (adenosylcobalamin)-dependent methylmalonyl-CoA mutase give rise to an inherited disorder of propionic acid metabolism termed mut methylmalonic aciduria. Almost all such mutations alter amino acids in the homodimeric human enzyme that are identical to residues in the catalytic alpha-subunit of the heterodimeric methylmalonyl-CoA mutase from the bacterium Propionibacterium shermanii, to which the mature human enzyme shows an overall 65% sequence identity. To explore how specific mutations might cause the observed clinical phenotype, 12 known mutations were mapped onto a three-dimensional homology model of the subunit of the human enzyme, generated using the program MODELLER on the basis of the recently published 2.0 A X-ray crystal structure of the P. shermanii methylmalonyl-CoA mutase. Eight mutations are found in the C-terminal B12-binding domain, of which 4 (G623R, G626C, G630E, G703R) are in direct contact with the corrin and are clustered around the histidine ligand (H627) provided by the protein to coordinate the cobalt atom of the B12 cofactor. Introduction of a side chain, particularly one that is charged, at any of these positions is expected to disrupt the flavodoxin-like fold and severely impair its binding of B12. Mutation at either of two other highly conserved glycine residues in this domain (G648D, G717V) also disrupts critical elements in the fold as would the introduction of an additional positive charge in the mutation H678R. Mutation of an arginine in a solvent-exposed loop to a hydrophobic residue (R694W) is also pathogenic. The remaining mutations have been mapped to the N-terminal region of the mutase, two of which introduce a buried, uncompensated charge, either near the subunit interface (A377E), or near the narrow channel through which acyl-CoA esters gain access to the active site (W105R). The extreme N-terminus of methylmalonyl-CoA mutase is predicted to make extensive contacts with the other subunit, and a mutant in this region (R93H) may prevent the correct assembly of the dimer.

How coenzyme B12 radicals are generated: the crystal structure of methylmalonyl-coenzyme A mutase at 2 A resolution

BACKGROUND

The enzyme methylmalonyl-coenzyme A (CoA) mutase, an alphabeta heterodimer of 150 kDa, is a member of a class of enzymes that uses coenzyme B12 (adenosylcobalamin) as a cofactor. The enzyme induces the formation of an adenosyl radical from the cofactor. This radical then initiates a free-radical rearrangement of its substrate, succinyl-CoA, to methylmalonyl-CoA.

RESULTS

Reported here is the crystal structure at 2 A resolution of methylmalonyl-CoA mutase from Propionibacterium shermanii in complex with coenzyme B12 and with the partial substrate desulpho-CoA (lacking the succinyl group and the sulphur atom of the substrate). The coenzyme is bound by a domain which shares a similar fold to those of flavodoxin and the B12-binding domain of methylcobalamin-dependent methionine synthase. The cobalt atom is coordinated, via a long bond, to a histidine from the protein. The partial substrate is bound along the axis of a (beta/alpha)8 TIM barrel domain.

CONCLUSIONS

The histidine-cobalt distance is very long (2.5 A compared with 1.95-2.2 A in free cobalamins), suggesting that the enzyme positions the histidine in order to weaken the metal-carbon bond of the cofactor and favour the formation of the initial radical species. The active site is deeply buried, and the only access to it is through a narrow tunnel along the axis of the TIM barrel domain.

Organisation of the biosynthetic gene cluster for rapamycin in Streptomyces hygroscopicus: analysis of genes flanking the polyketide synthase

Analysis of the gene cluster from Streptomyces hygroscopicus that governs the biosynthesis of the polyketide immuno-suppressant rapamycin (Rp) has revealed that it contains three exceptionally large open reading frames (ORFs) encoding the modular polyketide synthase (PKS). Between two of these lies a fourth gene (rapP) encoding a pipecolate-incorporating enzyme that probably also catalyzes closure of the macrolide ring. On either side of these very large genes are ranged a total of 22 further ORFs before the limits of the cluster are reached, as judged by the identification of genes clearly encoding unrelated activities. Several of these ORFs appear to encode enzymes that would be required for Rp biosynthesis. These include two cytochrome P-450 monooxygenases (P450s), designated RapJ and RapN, an associated ferredoxin (Fd) RapO, and three potential SAM-dependent O-methyltransferases (MTases), RapI, RapM and RapQ. All of these are likely to be involved in 'late' modification of the macrocycle. The cluster also contains a novel gene (rapL) whose product is proposed to catalyze the formation of the Rp precursor, L-pipecolate, through the cyclodeamination of L-lysine. Adjacent genes have putative roles in Rp regulation and export. The codon usage of the PKS biosynthetic genes is markedly different from that of the flanking genes of the cluster.

Organization of the biosynthetic gene cluster for rapamycin in Streptomyces hygroscopicus: analysis of the enzymatic domains in the modular polyketide synthase

The three giant multifunctional polypeptides of the rapamycin (Rp)-producing polyketide synthase (RAPS1, RAPS2 and RAPS3) have recently been shown to contain 14 separate sets, or modules, of enzyme activities, each module catalysing a specific round of polyketide chain extension. Detailed sequence comparison between these protein modules has allowed further characterisation of aa that may be important in catalysis or specificity. The acyl-carrier protein (ACP), beta-ketoacyl-ACP synthase (KS) and acyltransferase (AT) domains (the core domains) have an extremely high degree of mutual sequence homology. The KS domains in particular are almost perfect repeats over their entire length. Module 14 shows the least homology and is unique in possessing only core domains. The enoyl reductase (ER), beta-ketoacyl-ACP reductase (KR) and dehydratase (DH) domains are present even in certain modules where they are not apparently required. Four DH domains can be recognised as inactive by characteristic deletions in active site sequences, but for two others, and for KR and ER in module 3, the sequence is not distinguishable from that of active counterparts in other modules. The N terminus of RAPS1 contains a novel coenzyme A ligase (CL) domain that activates and attaches the shikimate-derived starter unit, and an ER activity that may modify the starter unit after attachment. The sequence comparison has revealed the surprisingly high sequence similarity between inter-domain 'linker' regions, and also a potential amphipathic helix at the N terminus of each multienzyme subunit which may promote dimerisation into active species.

Evidence for a double-helical structure for modular polyketide synthases

Modular polyketide synthases are multienzymes responsible for the biosynthesis of a large number of clinically important natural products. They contain multiple sets, or modules, of enzymatic activities, distributed between a few giant multienzymes and there is one module for every successive cycle of polyketide chain extension. We show here that each multienzyme in a typical modular polyketide synthase forms a (possibly helical) parallel dimer, and that each pair of identical modules interacts closely across the dimer interface. Such an arrangement would allow identical modules to share active sites for chain extension, and thus to function independently of flanking modules, which would have important implications both for mechanisms of evolution of polyketide synthases and for their future genetic engineering.

Divergent sequence motifs correlated with the substrate specificity of (methyl)malonyl-CoA:acyl carrier protein transacylase domains in modular polyketide synthases

The amino acid sequences of a large number of polyketide synthase domains that catalyse the transacylation of either methylmalonyl-CoA or malonyl-CoA onto acyl carrier protein (ACP) have been compared. Regions were identified in which the acyltransferase sequences diverged according to whether they were specific for malonyl-CoA or methylmalonyl-CoA. These differences are sufficiently clear to allow unambiguous assignment of newly-sequenced acyltransferase domains in modular polyketide synthases. Comparison with the recently-determined structure of the malonyltransferase from Escherichia coli fatty acid synthase showed that the divergent region thus identified lies near the acyltransferase active site, though not close enough to make direct contact with bound substrate.

Polyketide synthesis in vitro on a modular polyketide synthase

BACKGROUND

The 6-deoxyerythronolide B synthase (DEBS) of Saccharopolyspora erythraea, which synthesizes the aglycone core of the antibiotic erythromycin A, contains some 30 active sites distributed between three multienzyme polypeptides (designated DEBS1-3). This complexity has hitherto frustrated mechanistic analysis of such enzymes. We previously produced a mutant strain of S. erythraea in which the chain-terminating cyclase domain (TE) is fused to the carboxyl-terminus of DEBS1, the multienzyme that catalyzes the first two rounds of polyketide chain extension in S. erythraea. This mutant strain produces triketide lactone in vivo. We set out to purify the chimaeric enzyme and to determine its activity in vitro.

RESULTS

The purified DEBS1-TE multienzyme catalyzes synthesis of triketide lactones in vitro. The synthase specifically uses the (2S)-isomer of methylmalonyl-CoA, as previously proposed, but has a more relaxed specificity for the starter unit than in vivo.

CONCLUSIONS

We have obtained a purified polyketide synthase system, derived from DEBS, which retains catalytic activity. This approach opens the way for mechanistic and structural analyses of active multienzymes derived from any modular polyketide synthase.

Repositioning of a domain in a modular polyketide synthase to promote specific chain cleavage

Macrocyclic polyketides exhibit an impressive range of medically useful activities, and there is great interest in manipulating the genes that govern their synthesis. The 6-deoxyerythronolide B synthase (DEBS) of Saccharopolyspora erythraea, which synthesizes the aglycone core of the antibiotic erythromycin A, has been modified by repositioning of a chain-terminating cyclase domain to the carboxyl-terminus of DEBS1, the multienzyme that catalyzes the first two rounds of polyketide chain extension. The resulting mutant markedly accelerates formation of the predicted triketide lactone, compared to a control in which the repositioned domain is inactive. Repositioning of the cyclase should be generally useful for redirecting polyketide synthesis to obtain polyketides of specified chain lengths.

Limited proteolysis and active-site studies of the first multienzyme component of the erythromycin-producing polyketide synthase

The domain structure of the 6-deoxyerythronolide B synthase 1 component of the erythromycin-producing polyketide synthase from Saccharopolyspora erythraea has been investigated using limited proteolysis and active-site labeling. Trypsin, elastase, endoproteinase Glu-C, and endoproteinase Arg-C were used to cleave the multienzyme, and the sizes of the resulting fragments were assessed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The location of fragments within the primary structure was established by N-terminal sequence analysis. The cleavage pattern followed domain boundaries previously predicted on the basis of sequence alignments, but many predicted interdomain regions were not cleaved, even under the harshest conditions used. Initial proteolysis generated three large fragments: an N-terminal fragment (about 60 kDa) housing an acyltransferase-acyl carrier protein di-domain; a central fragment (about 90 kDa) containing a ketosynthase-acyltransferase di-domain; and a C-terminal fragment (about 220 kDa) containing the remaining six domains of the multienzyme, including the third acyltransferase. The intact multienzyme behaves as a dimer of molecular mass 660 kDa on gel filtration; and the C-terminal fragment remains dimeric. However, the N-terminal and central fragments appear to be monomeric species. After proteolysis of the multienzyme, the N-terminal di-domain was found to be specifically labeled after incubation with [14C]propionyl-CoA, providing the first evidence for its proposed role as a "loading domain" for the propionate starter unit. In contrast, the other two fragments were specifically acylated by [14C]methylmalonyl-CoA, indicating that both the other two acyltransferases remain enzymatically active after proteolysis.

An ABC-transporter from Streptomyces longisporoflavus confers resistance to the polyether-ionophore antibiotic tetronasin

Streptomyces longisporoflavus produces the polyketide-polyether antibiotic, tetronasin, which acts as an ionophore and depolarizes the membrane of bacteria sensitive to the drug. A genomic library of S. longisporoflavus DNA was cloned in Streptomyces lividans and screened to identify tetronasin-resistance determinants. The inclusion of 0.2M NaCl in the growth medium with tetronasin markedly improved the sensitivity of the screen. Two different resistance determinants, designated tnrB (ptetR51) and tnrA (ptetR11) respectively, were identified. The determinant tnrB (ptetR51) but not tnrA (ptetR11), also conferred resistance to tetronasin when cloned into Streptomyces albus. The tnrB determinant was further localized, by subcloning, to a 2.8 kb KpnI fragment. DNA sequence analysis of this insert revealed one incomplete and two complete open reading frames (ORFs 1, 2 and 3). The deduced sequence of the gene product of ORF2 (TnrB2) revealed significant similarity to the ATP-binding domains of the ABC (ATP binding cassette) superfamily of transport-related proteins. The adjacent gene, ORF3, is translationally coupled to ORF2 and would encode a hydrophobic protein (TnrB3) with six transmembrane helices which probably constitutes the integral membrane component of the transporter. The mechanism of tetronasin resistance mediated by tnrB is probably an ATP-dependent efflux system.

Stereospecific acyl transfers on the erythromycin-producing polyketide synthase

During assembly of complex polyketide antibiotics like erythromycin A, molecular recognition by the multienzyme polyketide synthase controls the stereochemical outcome as each successive methylmalonyl-coenzyme A (CoA) extender unit is added. Acylation of the purified erythromycin-producing polyketide synthase has shown that all six acyltransferase domains have identical stereospecificity for their normal substrate, (2S)-methylmalonyl-CoA. In contrast, the configuration of the methyl-branched centers in the product, that are derived from (2S)-methylmalonyl-CoA, is different. Stereoselection during the chain building process must, therefore, involve additional epimerization steps.