Skip to main content
NCBI home page
UKPMC Funders Author Manuscripts logoLink to UKPMC Funders Author Manuscripts
. Author manuscript; available in PMC: 2023 Mar 15.
Published in final edited form as: Nat Chem Biol. 2022 Sep 15;18(12):1410–1416. doi: 10.1038/s41589-022-01127-y

Diene Incorporation by a Dehydratase Domain Variant in Modular Polyketide Synthases

Christian Hobson 1,‡,#, Matthew Jenner 1,2,#, Xinyun Jian 1,2, Daniel Griffiths 1,‡, Douglas M Roberts 1, Matias Rey-Carrizo 1,‡, Gregory L Challis 1,2,3,4,*
PMCID: PMC7613849  EMSID: EMS151800  PMID: 36109649

Abstract

Modular polyketide synthases (PKSs) are biosynthetic assembly lines that construct structurally diverse natural products with wide-ranging applications in medicine and agriculture. Various mechanisms contribute to structural diversification during PKS-mediated chain assembly, including dehydratase (DH) domain-mediated elimination of water from R and S-configured 3-hydroxy thioesters to introduce E and Z-configured carbon-carbon double bonds, respectively. Here we report the discovery of a novel DH domain variant that catalyses the sequential elimination of two molecules of water from a (3R, 5S)-3, 5-dihydroxy thioester during polyketide chain assembly, introducing a - conjugated E, Z-diene into various modular PKS products. We show that the reaction proceeds via a (2E, 5S)-2-enoyl-5-hydroxy thioester intermediate and involves an additional universally conserved histidine residue that is absent from the active site of most conventional DH domains. These findings expand the diverse range of chemistries mediated by DH-like domains in modular PKSs, highlighting the catalytic versatility of the double hot dog fold.

Introduction

Polyketides are an important family of natural products, with numerous applications in human medicine, animal health and crop protection1. Although they have exceptionally complex and diverse structures, the chemical transformations used to assemble their carbon chains are deceptively straightforward. A thioester starter unit is elongated with a series of (alkyl)malonyl thioester extender units via a decarboxylative condensation reaction, and the β-ketothioesters resulting from each chain elongation undergo one or more α- and/or β-carbon modifications.2,3

In bacteria, structurally complex polyketides are typically assembled by type I modular polyketide synthase (PKS) multienzymes. These remarkable molecular machines employ sets of catalytic domains grouped into modules, each of which is usually responsible for a single round of carbon chain extension and subsequent α/β-carbon modification (Fig. 1a).4 An acyl transferase (AT) loads the extender unit onto the post-translationally installed phosphopantetheine (ppant) prosthetic group of an acyl carrier protein (ACP) domain in each module (Fig. 1a). In cis-AT PKSs these ATs are integrated into the modules and are typically able to utilise a range of (alkyl)malonyl-CoA thioesters, whereas in trans-AT PKSs a single standalone AT usually acylates the ACP domain in each module with malonyl-CoA. Each chain extension module also contains a ketosynthase (KS) domain that receives the acyl chain attached to the ACP domain in the upstream module (via transthioesterification onto a conserved active site Cys residue) and elongates it with the extender unit (Fig. 1a). Optional α/β-carbon modifying domains alter the structure of the resulting β-ketothioester prior to chain elongation by the downstream module. These include C-methyltransferase (MT) domains, which methylate the α-carbon, ketoreductase (KR) domains, which reduce the β-keto group, dehydratase (DH) domains that eliminate water from the resulting β-hydroxy thioester and enoyl reductase (ER) domains, which saturate the alkene formed by the dehydration reaction (Fig. 1a).4,5 O-methyltransferase (OMT) domains catalyse O-methylation of the β-hydroxy thioester intermediate and several trans-acting enzymes are responsible for other types of modification.6–8

Figure 1. Overview of polyketide assembly by cis-AT and trans-AT modular PKSs, stereochemical details of E- and Z-alkene formation, examples of trans-AT PKS products containing E, Z-dienes and organisation of modules 12 and 13 of the glaldiolin PKS.

Figure 1

Open in a new tab

a, Comparison of canonical domain architectures in cis-AT and trans-AT PKSs for assembly of a hypothetical product (boxed). Domain abbreviations are as follows: AT = acyltransferase, ACP = acyl carrier protein, KS = ketosynthase, KR = ketoreductase, DH = dehydratase, ER = enoyl reductase, GNAT = GCN5-related acyltransferase, C-MT = C-methyltransferase. b, E- and Z-alkene formation in modular PKSs proceeds via reduction of the β-keto thioester by type B and type A KR domains to (3R)-3-hydroxy and (3S)-3-hydroxy thioesters, respectively, followed by DH domain-catalysed removal of the pro-S hydrogen atom at C-2 and elimination of the C-3 hydroxyl group. c, Structures of trans-AT PKS products containing E, Z-dienes (highlighted by dashed boxes). d, Organisation of modules 12 and 13 of the trans-AT PKS responsible for the assembly of gladiolin. Modules 12 and 13 contain type A and type B KR domains, respectively, which are predicted to result in formation of a (3R, 5S)-3, 5-dihydroxy thioester attached to the module 13 ACP domain. Only module 13 contains a DH-like domain, which is hypothesised to catalyse the elimination of two molecules of water from the 3, 5-dihydroxy thioester to form the corresponding E, Z-diene.

Many α/β-carbon modifying domains are highly stereoselective and sequence analyses can predict the stereochemistry of the products they form. For example, “type A” and “type B” KR domains generate (3S) and (3R)-3-hydroxythioester products,4 respectively (Fig. 1b), and are distinguishable using sequence alignments. Similarly, DH domains typically convert (3R)-3-hydroxythioesters to E-configured alkenes, although it has recently been established that some DH domains can convert S-configured substrates to Z-configured products (Fig. 1b).5,9

Several trans-AT PKSs assemble polyketides containing a conjugated E, Z-diene. Examples include gladiolin 1, leinamycin 2, kirromycin 3 and elansolid A 4 (Fig. 1c). Analysis of the catalytic domains in the module pairs proposed to be responsible for incorporating these dienes (e.g., modules 12 and 13 of the gladiolin PKS; Fig. 1d and Supplementary Fig. 1 and 2) show that the first contains a type A KR domain, whereas the second contains a type B KR domain. This is consistent with formation of a (3R, 5S)-3, 5-dihydroxy thioester intermediate. However, only the second module contains a DH-like domain, raising intriguing questions about the mechanism for E, Z-diene formation.

Here we elucidate the catalytic function and mechanism of the DH-like domain from module 13 of the gladiolin PKS (Supplementary Fig. 2) using chemically synthesised analogues of intermediates in polyketide chain assembly, intact protein mass spectrometry and site-directed mutagenesis. Our data show that this domain can convert an ACP-bound (3R, 5S)-3, 5-dihydroxyacyl thioester to a (2E, 4Z)-2, 4-dienoyl thioester, via a (2E, 5S)-2-enoyl-5-hydroxy thioester intermediate, further expanding the catalytic repertoire of DH-like domains in type I PKSs.

Results

Dehydration of 3-hydroxy thioesters

The apo-DH-ACP di-domain from module 13 of the gladiolin PKS was overproduced and purified as reported previously (Supplementary Fig. 3).10 Initially we investigated N-acetylcysteamine and pantetheine thioester mimics of the proposed intermediates in chain assembly attached to the ACP domains in modules 12 and 13 of the PKS as substrates for the DH-like domain. Consistent with studies of excised DH domains from other modular PKSs,11 they were found to be poor substrates for the DH-like domain in our construct and no dehydration products could be detected. To circumvent this problem, we exploited the CoaA, CoaD and CoaE enzymes to convert the pantetheine thioesters to the corresponding coenzyme A thioesters,12–14 which were then used, in conjunction with MgCl2, ATP and the substrate tolerant 4’-phosphopantethienyl transferase Sfp,15 to convert the apo-ACP domain of the DH-ACP di-domain to the corresponding acylated holo-forms (Fig. 2a). This allowed dehydration of the ACP-bound thioesters, resulting in a -18 Da mass shift, to be detected directly by intact protein mass spectrometry (Fig. 2a). The pantetheine thioesters were also used to convert the excised apo-ACP domain to the corresponding acylated holo-forms, which were analysed by intact protein mass spectrometry as a negative control (Supplementary Fig. 3 and 4)

Figure 2. Stereospecificity of the DH-like domain from module 13 of the gladiolin PKS towards 3-hydroxyacyl and 2-enoyl thioesters.

Figure 2

Open in a new tab

a, Principle of intact protein mass spectrometry assay used. A mass shift of -18 Da indicates dehydration of the loaded substrate. B, c, The (3R)-3-hydroxyhexanoyl thioester (5) undergoes dehydration, but the (3S)-3-hydroxyhexanoyl thioester (6) does not. d, e, The (2E)-2-hexenoyl thioester (7) is rehydrated, whereas the (2Z)-2-hexenoyl thioester (8) is not. Overall, the data are consistent with conversion of a (3R)-3-hydroxyacyl ACP thioester intermediate in gladiolin biosynthesis to the corresponding (2E)-2-enoyl thioester by the module 13 DH-like domain. Experiments were performed in triplicate and representative data are shown.

As initial probes of dehydratase activity, we synthesised (3R) and (3S)-3-hydroxyhexanoyl pantetheine thioesters, 5 and 6 (Supplementary Note) and incubated them separately with purified recombinant CoaA, CoaD, CoaE, Sfp and the apo-DH-ACP di-domain. Given the structural diversity of the predicted substrates of the putative diene-forming DH-like domains in trans-AT PKSs (Supplementary Fig. 1), we reasoned they are likely to have broad substrate tolerance and that the ACP-bound thioesters derived from 5 and 6 would be accepted by the module 13 DH-like domain. After 3 hours, the mixtures were analysed using UHPLC-ESI-Q-TOF-MS, which showed that the R-configured thioester had undergone dehydration, whereas the S-configured thioester had not (Fig. 2b, 2c and Supplementary Table 1). To elucidate the stereochemical outcome of the transformation with the R-configured thioester, we exploited the intrinsic reversibility of the dehydration reaction, following the precedent set by previous studies of dehydratases involved in bacterial fatty acid biosynthesis and a DH domain from the pikromycin PKS.11,16 The (2E) and (2Z)-2-hexenoyl panetheine thioesters, 7 and 8, were synthesised (Supplementary Note) and loaded via their CoA thioesters onto the apo-ACP domain of the DH-ACP di-domain construct. UHPLC-ESI-Q-TOF-MS analyses of the reaction mixtures showed that only the E-configured thioester underwent rehydration (Fig. 2d, 2e and Supplementary Table 1). Based on the principle of microscopic reversibility,16 these results indicate that the DH-like domain in module 13 of the gladiolin PKS, like canonical modular PKS DH domains, catalyses the dehydration of (3R)-3-hydroxyacyl thioesters to form (2E)-configured products. The DH-like domain cannot, however, catalyse the formation of a Z-configured product from a (3S)-3-hydroxyacyl thioester, as reported recently for the DH domains from module 2 of the fostriecin cis-AT PKS,5 and modules 2 and 5 of the bongkrekic acid and oxazolomycin trans-AT PKSs, respectively.9 Formation of the Z-configured C12-C13 double bond in gladiolin is therefore very unlikely to involve direct dehydration of the (3S)-3-hydroxyacyl thioester produced by the type A KR domain in module 12 of the PKS.

Double dehydration of a (3R, 5S)-3,5-dihydroxy thioester

The organisation of catalytic domains in modules 12 and 13 of the gladiolin PKS suggests a (3R, 5S)-3, 5-dihydroxyacyl thioester attached to the module 13 ACP domain might be an intermediate in polyketide chain assembly. It is therefore conceivable that the module 13 DH-like domain catalyses two successive dehydrations of this diol to form the (10E, 12Z)-diene of gladiolin. To test this possibility, we synthesised (3R, 5S)-3, 5-dihydroxyhexanoyl panthetheine thioester 9 (Supplementary Note), converted it to the corresponding CoA thioester, loaded it onto the apo-DH-ACP di-domain and determined its fate by intact protein MS. In addition to the diol, singly and doubly dehydrated species were observed (Fig. 3a and Supplementary Table 1), indicating that the DH-like domain is indeed capable of diene formation. To gain further insight into this process, we synthesised (2E, 4Z)-2, 4-hexadienoyl pantetheine thioester 10 (Supplementary Note), which corresponds to the presumed product of the double dehydration reaction and loaded it onto the apo-DH-ACP di-domain. Intact protein MS analysis showed that singly and doubly hydrated products are formed (Fig. 3b and Supplementary Table 1). Significant quantities of the holo-DH-ACP di-domain were observed in both experiments, likely resulting from facile offloading of the diol from the ACP domain via δ-lactonisation. Taken together, these data are consistent with formation of the (10E, 12Z)-diene of gladiolin via two sequential dehydrations of a (3R, 5S)-3, 5-dihydroxyacyl thioester by the DH-like domain in module 13 of the PKS. To further substantiate this hypothesis, we used cysteamine to cleave the products of the reaction with the (3R, 5S)-3, 5-dihydroxyhexanoyl thioester from the ACP domain.17 LC-MS comparisons with the cysteamine adducts derived from the synthetic panthetheine thioesters of (2E, 4Z) and (2E, 4E)-2, 4-hexadienoate (10 and 11, respectively; Supplementary Note), showed that the (2E, 4Z)-hexadienoyl thioester is the major diene-containing product of the reaction (Fig. 3c). Significant quantities of the (2E, 4E)-2, 4-hexadienoyl cysteamine adduct were also observed. This is likely due primarily to facile isomerisation of the (2E, 4Z)-2, 4-hexadienoyl thioester to the thermodynamically more stable E, E-stereoisomer, as observed during the synthesis of (2E, 4Z)-2, 4-hexadienoyl pantetheine thioester 10. However, we cannot exclude that the dehydration leading to formation of the 4, 5-double bond results in small amounts of the 4E-configured product, in addition to the major 4Z-configured product of the reaction.

Figure 3. The DH-like domain from module 13 of the gladiolin PKS catalyses double dehydration of a (3R,5S)-3,5-dihydroxyacyl thioester and double rehydration of a (2E,4Z)-2,4-dienoyl thioester.

Figure 3

Open in a new tab

a, Deconvoluted mass spectrum of the DH-ACP di-domain following conversion of pantetheine thioester 9 to the corresponding CoA thioester and loading onto the apo-ACP domain, showing the formation of singly and doubly dehydrated species. b, Deconvoluted mass spectrum of the DH-ACP di-domain following conversion of pantetheine thioester 10 to the corresponding CoA thioester and loading onto the apo-ACP domain, showing the formation of singly and doubly rehydrated species. Overall, the data are consistent with conversion of a (3R,5S)-3,5-dihydroxyacyl thioester intermediate in gladiolin biosynthesis to the corresponding (2E,4Z)-2,4-dienoyl thioester by the DH-like domain in module 13 of the PKS. Experiments were performed in triplicate and representative data are shown. c, Extracted ion chromatograms (EICs) at m/z = 247.09 ± 0.02 from LC-MS comparisons of the cysteamine adducts of the 2, 4-dienoyl thioester resulting from double dehydration of the (3R,5S)-3,5-dihydroxyacyl-ACP thioester (9) by the DH-like domain (top) with synthetic standards of (2E,4Z) and (2E, 4E)-2,4-dienoyl panthetheine thioesters (10 and 11, respectively). Note that (2E,4Z)-2,4-dienoyl panthetheine thioester (10) undergoes facile isomerisation to the thermodynamically more stable 2E, 4E-isomer, resulting in significant quantities of the latter in the authentic standard and products of the enzymatic reaction.

Activity with stereoisomeric 3, 5-dihydroxy thioesters

To further probe the stereochemical preferences of the DH-like domain, we synthesised all three stereoisomers (12, 13 and 14) of (3R, 5S)-3, 5-dihydroxyhexanoyl panthetheine thioester 9 (Supplementary Note) and studied their fate upon conversion to the corresponding CoA thioesters and subsequent loading onto the apo-DH-ACP di-domain (Extended Data Fig. 1 and 2). Negligible dehydration was observed for stereoisomers 12 and 13 with 3S-configured hydroxyl groups, relative to the control reactions employing the excised ACP domain (Extended Data Fig. 1a, 1b, Supplementary Table 1 and Supplementary Fig. 4). Interestingly, the major product of the reaction employing 13 is the holo-DH-ACP di-domain, which likely results from offloading of the substrate via δ-lactonisation (Extended Data Fig. 1b) These data are consistent with that obtained for the (3S)-3-hydroxyhexanoyl pantetheine thioester 6 (Fig. 2) and indicate that diene formation proceeds via elimination of the C-3 hydroxyl group, followed by the C-5 hydroxyl group.

In contrast, the (3R, 5R)-3, 5-dihydroxyacyl thioester 14 underwent almost complete conversion to a mixture of singly and doubly dehydrated products (Extended Data Fig. 2a and Supplementary Table 1), indicating that the DH-like domain is not stereoselective towards the C-5 hydroxyl group. Comparison of the cysteamine cleavage adducts from this reaction with that for the panthetheine thioesters of (2E, 4Z) and (2E, 4E)-2, 4-hexadienoate (10 and 11, respectively) showed that the (2E, 4E)-2, 4-dienoyl thioester is the sole diene-containing product (Extended Data Fig. 2b). Accordingly, conversion of the (2E, 4E)-2, 4-dienoyl pantetheine thioester 11 to the corresponding CoA thioester and subsequent loading onto the apo-DH-ACP di-domain yielded predominantly a mono-rehydrated product (Extended Data Fig. 2c and Supplementary Table 1). The differences in product profile observed for the reactions employing the (3R, 5S)-3, 5-dihydroxyhexanoyl / (2E, 4Z)-2, 4-dienoyl panthetheine thioesters (9 and 10), where a significant proportion of the diol remains / is produced, and the (3S, 5S)-3, 5-dihydroxyhexanoyl / (2E, 4E)-2, 4-dienoyl pantetheine thioesters (14 and 11), where very little of the diol remains / is produced, can be explained by the greater thermodynamic stability of the E, E-diene than the E, Z-diene, and are consistent with the hypothesis that the (2E, 4Z)-2, 4-dienoyl thioester is the major product of the enzyme-catalysed dehydration of (3R, 5S)-3, 5-dihydroxyhexanoyl thioester.

Mechanism of diene formation

Consistent with the observation that 3S-configured 3,5-dihydroxy thioesters are poor substrates for the DH-like domain, mechanistic logic suggests that diene formation likely proceeds via initial elimination of the C-3 hydroxyl group. This is because the protons attached to C-2 are significantly more acidic than those attached to C-4 in the 3, 5-dihydroxy thioester and the acidity of the protons attached to C-4 increases dramatically in the 5-hydroxy-2-enoyl thioester resulting from elimination of the C-3 hydroxyl group. To test this hypothesis, we synthesised (2E, 5S)-2-hexenoyl-5-hydroxy pantetheine thioester 15 (Supplementary Note), which corresponds to the putative mono-dehydrated intermediate in diene formation from the (3R, 5S)-3,5-dihydroxyhexanoyl thioester. Loading of thioester 15, via the corresponding CoA thioester, onto the apo-DH-ACP di-domain produced a mixture of hydrated and dehydrated products (Fig. 4 and Supplementary Table 1), providing clear evidence that it is an intermediate in the interconversion of the diol and the diene. In an analogous experiment with the synthetic (2E, 5R)-2-hexenoyl-5-hydroxy pantetheine thioester 16 (Supplementary Note) the diene was produced (Extended Data Fig. 2d and Supplementary Table 1). However, as for the (2E, 4E)-2,4-hexadienoyl thioester 11, negligible diol formation was observed (Extended Data Fig. 2d and Supplementary Table 1).

Figure 4. Dehydration of a (2E, 5S)-2-enoyl-5-hydroxy thioester by the DH-like domain and proposed mechanism for diene formation.

Figure 4

Open in a new tab

a, Deconvoluted mass spectrum of the DH-ACP didomain following conversion of pantetheine thioester 15 to the corresponding CoA thioester and loading onto the apo-ACP domain, showing conversion to both diene and diol products. The experiment was performed in triplicate and representative data are shown. The stereochemistry of the major diene / diol formed was assigned by analogy with the data shown in Figure 3. (b) Proposed catalytic mechanism of the DH-like domain. Hydrogen bonding of the C-3 and C-5 hydroxyl groups to Asp200 orients the substrate for E, Z-diene formation. The catalytic His23 and His158 residues deprotonate C-2 and C-4 to form the enolate and vinylogous enolate intermediates, respectively.

To illuminate the molecular basis for the unusual catalytic activity of the DH-like domain, we created a structural model using AlphaFold (Fig. 5).18 An overlay of this model with the X-ray crystal structure of a DH domain from module 11 of the difficidin polyketide synthase, which has been proposed to convert a 3, 5-dihydroxy thioester to a diene (although no experimental evidence is available to validate this),19 shows a very high degree of structural similarity (Extended Data Fig. 3).19 DH domains typically contain a conserved His-Asp catalytic dyad in their active site. A one-base mechanism, in which the His residue of the catalytic dyad first deprotonates C-2 of the thioester to form an enolate, then protonates the C-3 hydroxyl group as it is eliminated from this intermediate, has been proposed for PKS DH domains (Extended Data Fig. 4),16 and recently verified using pH-rate profiles.20 The Asp residue of the catalytic dyad appears to position the substrate by accepting a hydrogen bond from the C-3 hydroxyl group. The AlphaFold model of the DH-like domain contains an additional active site His residue on an adjacent β-strand ~3.5 Å away from the His residue of the catalytic dyad (Fig. 5). In canonical DH domains, such as that from the EryAII subunit of the 6-deoxyerythronolide B (6-DEB) PKS, this residue is Tyr (Fig. 5 and Extended Data Fig. 5). A multiple sequence alignment of all the DH-like domains in trans-AT PKSs predicted to catalyse diene formation revealed that this additional His residue is universally conserved (Extended Data Fig. 3). These observations led us to propose that the 2-enoyl-5-hydroxy thioester intermediate in diene-forming DH-like domains likely results from deprotonation of C-2 by the His residue of the conserved His-Asp dyad, whereas the additional His residue (His158 in the DH-like domain from module 13 of the gladiolin PKS) deprotonates C-4 of this intermediate to form a vinylogous enolate (Fig. 4). The C-5 hydroxyl group is then eliminated from the vinylogous enolate with concomitant protonation by the additional His residue (Fig. 4). A similar mechanism has been proposed for formation of a 4E-configured double bond by DH domains in the cis-AT PKS that assembles curacin. However, these domains lack the additional active site His residue in diene-forming DH-like domains.21 The Asp residue in the conserved His-Asp dyad of the trans-AT PKS DH-like domains likely accepts hydrogen bonds from both the C-3 and C-5 hydroxyl groups of the 3, 5-dihydroxy thioester substrates. This would explain why substrates with 3R, 5S and 3R, 5R configurations give rise to predominantly 2E, 4Z and dienes 2E, 4E-configured dienes, respectively (Fig. 4b and Extended Data Fig. 2e).

Figure 5. Structural basis for diene formation by DH-like domains.

Figure 5

Open in a new tab

a, Comparison of the active sites of the AlphaFold model of the DH-like domain from module 13 of the gladiolin PKS (top) and the EryAII DH domain (bottom; from PDB entry 3EL6) rendered using PyMol. Although both domains possess the His-Asp catalytic dyad typically found in DH domains, the DH-like domain from module 13 of the gladiolin PKS has an additional His residue (His158) adjacent to that in the catalytic dyad (His23). His158 in the DH-like domain is hypothesised to deprotonate C-4 of the (5S)-5-hydroxy-(2E)-2-enoyl thioester intermediate, forming a vinylogous enolate which eliminates the (5S)-configured hydroxyl group. In canonical DH domains, such as that from EryAII, the additional His residue is replaced by Tyr. b, Deconvoluted mass spectra of the H158Y mutant of the DH-ACP di-domain following conversion of pantetheine thioester 15 to the corresponding CoA thioester and loading onto the apo-ACP domain. The mutated di-domain catalyses rehydration of the intermediate analogue to the diol but is unable to dehydrate it to the diene, confirming that His158 plays a key role in diene formation. The experiment was performed in triplicate and representative data are shown.

To investigate the role played by the additional His residue in diene formation, a H158Y mutant of the DH-ACP di-domain from module 13 of the gladiolin PKS was constructed (Supplementary Fig. 5). Loading of (2E, 5S)-2-hexenoyl-5-hydroxy pantetheine thioester 15, via the corresponding CoA thioester, onto the mutant di-domain resulted in formation of only the rehydrated (3R, 5S)-3, 5-dihydroxyacyl thioester (Fig. 5 and Supplementary Table 1). None of the diene product resulting from elimination of the C-5 hydroxyl group could be detected (Fig. 5 and Supplementary Table 1). These data are consistent with the proposed roles of the His-Asp dyad and the additional His residue in diene formation.

Discussion

In this study, we solve the conundrum of how conjugated E, Z-dienes are incorporated into the products of trans-AT PKSs. Our results show that a novel DH domain variant in these assembly lines containing an additional active site His residue converts (3R, 5S)-3, 5-dihydroxyacyl ACP thioester intermediates in polyketide chain assembly to the corresponding (2E, 4Z)-2, 4-dienoyl thioesters. This reaction is proposed to proceed via a (2E, 5S)-2-enoyl-5-hydroxy thioester intermediate resulting from deprotonation of C-2 in the 3, 5-dihydroxy thioester substrate by the His residue in the universally conserved His-Asp catalytic dyad of canonical DH domains. Deprotonation of C-4 in the intermediate by the additional active site His residue is hypothesised to trigger a second dehydration reaction, resulting in diene formation. We propose that the Asp residue in the His-Asp catalytic dyad forms hydrogen bonds with the C-3 and C-5 hydroxyl groups, positioning the C-2 and C-4 protons near the conserved and additional His residues, respectively. This orients the substrate in a conformation that favours E, Z-diene formation via two sequential E1CB reactions, each of which is a net syn-elimination.

Interestingly, in a minority of cases (the (2Z, 4E)-dienes in basiliskamide and macrolactin, the (12Z, 14E)-diene in chivosazol, and the (8Z, 10Z)-diene in difficidin) the configuration of the diene in the final product differs from the predicted 2E, 4Z-configuration of the diene in the corresponding biosynthetic intermediate (Supplementary Figure 1). Further experiments will be required to establish the reason for this discrepancy. One possibility, is that the corresponding DH-like domains bind the substrate in a different orientation, leading to the formation of a (2Z, 4E)-2, 4-dienoyl thioester. In the case of difficidin, the 8,10-diene is known to be configurationally labile.22 It is thus possible that the (8Z, 10Z)-configuration could result from isomerisation of an initially formed (8E, 10Z) or (8Z, 10E)-configured isomer.

We identified module pairs containing a type A KR domain in the first module and a type B KR domain alongside a DH-like domain in the second module in the cis-AT PKSs that assemble spirangien and thuggacin (Supplementary Fig. 1). The corresponding regions in the structures of spirangien and thuggacin contain dienes with the same configuration as the (2E, 4Z)-2, 4-dienoyl thioester intermediate predicted to be assembled by these modules, suggesting similar mechanisms may be employed for conjugated E, Z-diene incorporation in both trans and cis-AT PKSs. A related module pair containing an additional ER domain in the second module is present in the PKS that assembles epothilone C, which contains a Z-alkene in the corresponding region of its structure (Supplementary Fig. 1). Katz and co-workers proposed that the DH-like domain catalyses formation of a (2E, 4Z)-2, 4-dienoyl thioester, which is reduced by the ER domain to form the corresponding (4Z)-4-enoyl thioester (Supplementary Fig. 1).23 However, while inactivation of this DH-like domain resulted in production of an epothilone derivative with the expected hydroxyl group at C-13, this compound unexpectedly contained a 10, 11-double bond.23 It was suggested that a DH domain in a subsequent module may catalyse formation of the 10, 11-double bond, but there is no evidence to support this hypothesis. These putative diene-forming DH-like domains are phylogenetically distinct from both their counterparts in trans-AT PKSs and canonical DH domains in cis-AT PKSs (Extended Data Fig. 6). Sequence alignments with the DH domain from the EryAII subunit of the 6-DEB PKS, show that they contain the conserved His-Asp catalytic dyad (Extended Data Fig. 5). However, the additional active site His residue that plays a key catalytic role in diene-forming DH-like domains in trans-AT PKSs (vide supra), is replaced with a conserved Trp residue (Extended Data Fig. 5). It is therefore likely that a different catalytic mechanism is employed, akin to that used by FabA for the conversion of a (3R)-3-hydroxydecanoyl thioester to the corresponding (3Z)-3-enoyl thioester,24 involving sequential deprotonation of C-2 and C-4 by the single active site His residue. Further experiments will be required to confirm that these cis-AT PKS DH domain variants are responsible for E, Z-diene incorporation and to elucidate their catalytic mechanism.

In addition to dehydration of 3-hydroxyacyl thioesters, DH-like domains in trans-AT PKSs have been reported to catalyse the conversion of 3-enoyl-7-hydroxy thioesters and 3-enoyl-6-hydroxy thioesters to tetrahydropyrans and tetrahydrofurans, respectively, and the isomerisation of 2-enoyl thioesters to 3-enoyl thioesters.25–29 Our finding that such domains also catalyse the conversion of 3, 5-dihydroxy thioesters to 2, 4-dienoyl thioesters further expands the catalytic repertoire of DH-like domains. Phylogenetic analysis of all such domains from trans-AT PKSs, which possess a common double hotdog fold, shows that they clade according to their catalytic function (Extended Data Fig. 6). This will aid the predictive analysis of novel trans-AT PKS systems and prompts us to propose a self-consistent nomenclature for the various classes of domain: DHe, 2-enoyl thioester synthases; DHi, 2-enoyl thioester isomerases; DHt, tetrahydropyran synthases; DHd, 2, 4-dienoyl thioester synthases. It is interesting to note that DHd domains and one group of the DHe domains appear to have evolved from a common ancestor with an additional His residue at position 168 (EryAII DH domain numbering) in the active site (Extended Data Fig. 5 and 6). This group of DHe domains is embedded in the same type of “split module” architecture as the DHd domains (Extended Data Fig. 6), but the upstream KS domain is non-elongating, and the downstream KR domain is absent. Thus, the DHe domains introduce an alkene rather than a diene into the corresponding metabolic products. DH-like domains have been proposed to play yet further roles in modular PKSs, such as diene isomerisation and aromatic ring formation,30,31 although experimental evidence for these is currently lacking. Going forward, use of this nomenclature to annotate DH-like domains should facilitate understanding of the relationship between active site architecture, catalytic function, and product structure.

In conclusion, our discovery of a novel DH domain variant that converts 3, 5-dihydroxythioesters to 2, 4-dienes expands the β-keto thioester processing toolkit of modular polyketide synthases. This may open up new opportunities for biosynthetic engineering via introduction of an additional His residue into the active site of canonical DH domains, or domain / module swaps, and further underscores the remarkable catalytic versatility of the double hotdog fold.

Methods

Cloning and Mutagenesis

The cloning of the DH-ACP di-domain and ACP(DH) domain from module 13 of the gladiolin PKS in B. gladioli BCC0238 was reported previously.10 The H158Y mutant of the DH-ACP di-domain was constructed using the Q5 site-directed mutagenesis kit (NEB), with the following primers: DH-ACP(H158Y)_For-(5’-CGGCGTGGTATATGGCCCGTTCATGC-3’), DH-ACP(H158Y)_Rev-(5’-AGCTCGCGCACGCCGCGA-3’). Plasmids were isolated from cultures using a miniprep kit (Thermo), and the inserts were sequenced to verify their integrity.

Protein Overproduction and Purification

Overproduction and purification of module 13 DH-ACP di-domain and the excised module 13 ACP(DH) domain was conducted as described previously.10 The DH-ACP(H158Y) di-domain mutant was overproduced and purified using the same protocol as that employed for the wild type di-domain.

Substrate Loading and Dehydration Assay

Pantetheine thioesters were converted to the corresponding coenzyme A thioesters and attached to wild type and mutant DH-ACP di-domain, or the excised ACP(DH) domain using previously described methodology.12–14 Reactions were typically carried out using the DH-ACP di-domain or excised ACP(DH) domain (200 μM) in 20 mM Tris, 100 mM NaCl, 10 mM MgCl2, 10 mM ATP (total volume: 50 μL). Assays were initiated by addition of 2 μM Sfp, 1 μM CoaA, 1 μM CoaD, 1 μM CoaE and 1 mM pantetheine thioester, and were incubated for 3 h at 25 °C. The mixture was diluted to 20 μM with dH2O prior to UHPLC-ESI-Q-TOF-MS analysis.

UHPLC-ESI-Q-TOF-MS Analysis of Intact Proteins

All assays were analysed on a Bruker MaXis II ESI-Q-TOF-MS connected to a Dionex 3000 RS UHPLC fitted with an ACE C4-300 RP column (100 x 2.1 mm, 5 μm, 30 °C). The column was eluted with a linear gradient of 5–100% MeCN containing 0.1% formic acid over 30 min. The mass spectrometer was operated in positive ion mode with a scan range of 200-3000 m/z. Source conditions were: end plate offset at −500 V; capillary at −4500 V; nebulizer gas (N2) at 1.8 bar; dry gas (N2) at 9.0 L min−1 ; dry temperature at 200 °C. Ion transfer conditions were: ion funnel RF at 400 Vpp; multiple RF at 200 Vpp; quadrupole low mass at 200 m/z; collision energy at 8.0 eV; collision RF at 2000 Vpp; transfer time at 110.0 μs; pre-pulse storage time at 10.0 μs.

Cysteamine-Mediated Cleavage of Products from the DH-ACP Di-domain

Reactions were carried out using the DH-ACP di-domain (200 μM) in 20 mM Tris, 100 mM NaCl, 10 mM MgCl2, 10 mM ATP (total volume: 400 μL). Assays were initiated by addition of 2 μM Sfp, 1 μM CoaA, 1 μM CoaD, 1 μM CoaE and 1 mM pantetheine thioester, and were incubated for 1.5 h at 25 °C before addition of cysteamine to a final concentration of 0.5 mM. Cystamine thioesters of (2E, 4Z)- and (2E, 4E)-hexadienoate were prepared using 1 mM of the (2E, 4Z)- or (2E, 4E)-hexadienoyl panthetheine thioesters (compounds 10 and 11, respectively) in the same reaction conditions, except with the GbnD5 DH-ACP di-domain omitted. After 20 h at 4 °C, reaction solutions were extracted twice with EtOAc (1 mL). The combined extracts were evaporated to dryness, redissolved in 100 μL of MeOH and analysed using UHPLC-ESI-Q-TOF-MS.

UHPLC-ESI-Q-TOF-MS Analysis of Cysteamine Cleavage Products

Cystamine adducts resulting from the cleavage reactions were analysed by UHPLC-ESI-Q-TOF-MS, using a Zorbax Eclipse Plus C18 column (100 × 2.1 mm, 1.8 μm) connected to a Dionex UltiMate 3000 UHPLC coupled to a Bruker MaXis IMPACT mass spectrometer. Mobile phases consisted of water (A) and acetonitrile (B), each supplemented with 0.1% formic acid. A gradient of 5 % B to 100 % B over 42 minutes was employed at a flow rate of 0.2 mL/min. The mass spectrometer was operated in positive ion mode with a scan range of 50 - 3000 m/z. Source conditions were as follows: end plate offset at −500 V; capillary at −4500 V; nebulizer gas (N2) at 1.6 bar; dry gas (N2) at 8 L min−1; dry temperature at 180 °C. Ion transfer conditions were: ion funnel RF at 200 Vpp; multiple RF at 200 Vpp; quadrupole low mass at 55 m/z; collision energy at 5.0 eV; collision RF at 600 Vpp; ion cooler RF at 50 – 350 Vpp; transfer time at 121 μs; pre-pulse storage time at 1 μs. Calibration was performed with 1 mM sodium formate through a loop injection of 20 μL at the start of each run.

Structural Model of DH-like Domain from Module 13 of the Gladiolin PKS

A structural model of the DH-like domain from module 13 of the gladiolin PKS was generated using AlphaFold.18 The full amino acid sequence of the excised domain (minus the His-Tag) was submitted to the AlphaFold Colab notebook, which uses a slightly simplified version of AlphaFold v2.1.0, with the run_relax parameter enabled.

Extended Data

Extended Data Fig. 1. Activity of the DH-like domain from module 13 of the gladiolin PKS towards 3S, 5S and 3S, 5R-configured stereoisomers ofthe natural substrate.

Extended Data Fig. 1

Open in a new tab

a, Deconvoluted mass spectrum of the DH-ACP di-domain following conversion of pantetheine thioester 12 to the corresponding CoA thioester and loading onto the apo-ACP domain. Negligible dehydration is observed. b, Deconvoluted mass spectrum of the DH-ACP di-domain following conversion of pantetheine thioester 13 to the corresponding CoA thioester and loading onto the apo-ACP domain. While a small quantity of mono-dehydrated product is observed, the major product is the holo-di-domain, which likely results from offloading of the substrate by d-lactonization. The stereochemistry of the putative 2-enoyl-5-hydroxy and 2, 4-dienoyl thioesters resulting from single and double dehydration of the ACP thioesters derived from 12 and 13 are tentatively assigned based on established stereochemical relationships for other PKS DH domains (i.e., R-configured hydroxyl groups result in E-configured alkenes and S-configured hydroxyl groups result in Z-configured alkenes). These stereochemical assignments could not be experimentally confirmed due to the very small quantities of singly and doubly dehydrated species produced.

Extended Data Fig. 2. Activity of the DH-like domain from module 13 of the gladiolin PKS towards the 3R, 5R-configured stereoisomer of the natural substrate.

Extended Data Fig. 2

Open in a new tab

a, Deconvoluted mass spectrum of the DH-ACP di-domain following conversion of pantetheine thioester 14 to the corresponding CoA thioester and loading onto the apo-ACP domain, showing the formation of singly and doubly dehydrated products. b, Extracted ion chromatograms (EICs) at m/z = 247.09 ± 0. 02 from LC-MS comparisons of the cysteamine adducts of the 2, 4-dienoyl thioester resulting from double dehydration of the (3R, 5R)-3,5-dihydroxyacyl-ACP thioester by the DH-like domain (top) with synthetic standards of (2E, 4Z) (middle) and (2E, 4E)-2,4-dienoyl (bottom) panthetheine thioesters (10 and 11, respectively). c, Deconvoluted mass spectrum of the DH-ACP di-domain following conversion of pantetheine thioester 11 to the corresponding coenzyme A thioester and loading onto the apo-ACP domain, showing the formation of primarily a singly rehydrated species. d, Deconvoluted mass spectrum of the DH-ACP di-domain following conversion of thioester 16 to the corresponding coenzyme A thioester and subsequent loading onto the apo-ACP domain, showing the formation of primarily the 2, 4-dienoyl thioester. All experiments were performed in triplicate and representative data are shown. e, Proposed mechanism for conversion of the (3R, 5R)-3,5-dihydroxyacyl-ACP thioester derived from 14 to the corresponding (2E, 4E)-2,4-dienoyl thioester by the DH-like domain. Hydrogen bonding of the C-3 and C-5 hydroxyl groups to Asp200 orients the substrate for E, E-diene formation.

Extended Data Fig. 3. Comparison of a model of the DH-like domain from module 13 of the gladiolin PKS with the structure of the DH-like domain from module 11 of the difficidin PKS and conservation of twin active site His residues in putative diene-forming DH-like domains from trans-AT PKSs.

Extended Data Fig. 3

Open in a new tab

a, Overlay of the AlphaFold model of the DH-like domain from the gladiolin PKS with the X-ray crystal structure of the DH-like domain from the difficidin PKS, highlighting the near identical positioning of the Asp and two His active site residues. b, Regions from a sequence alignment encompassing the two His residues observed in the active sites of the DH-like domains from the gladiolin and difficidin PKSs (H23/25 and H158/155, respectively), highlighting that they are universally conserved in other putative diene-forming DH-like domains from trans-AT PKSs. The labelling of DH-like domains corresponds to the subunit names and module numbers illustrated in Supplementary Fig. 1 (e.g., Gbn_M13 is the DH-like domain from module 13 of GbnD5 subunit of the gladiolin PKS).

Extended Data Fig. 4. Mechanism of DH domain-catalyzed dehydration of 3-hydroxyacyl-ACP thioesters.

Extended Data Fig. 4

Open in a new tab

a, Dehydration of (3R)-3-hydroxyacyl thioesters yields the corresponding (2E)-2- enoyl thioester. b, Dehydration of (3S)-3- hydroxyacyl thioesters yields the Z-configured product. These reactions proceed via an E1CB mechanism involving a conserved active site His residue, which acts a general base / acid, and involve a net syn-elimination of the C-3 hydroxyl group and Hb. Hydrogen bonding of the C-3 hydroxyl group to a conserved active site Asp residue orients it in a gauche conformation with respect to Hb.

Extended Data Fig. 5. Sequence alignment of DH-like domains in cis-AT and trans-AT PKSs.

Extended Data Fig. 5

Open in a new tab

R1, R2 and R3 are residues corresponding to H44, Y168 and D206, respectively, of the DH domain from the EryAII subunit of the 6-DEB PKS (labelled EryDH1 in the figure). The functional group(s) introduced by each type of domain in trans-AT PKSs are shown to the left of the alignments. DH domains are labelled according to the order they appear in the PKS (i.e., going from the N to C-terminus the first DH domain is labelled DH1, the second DH2 and so forth).

Extended Data Fig. 6. Phylogenetic analysis of DH-like domains in trans-AT and cis-AT PKSs.

Extended Data Fig. 6

Open in a new tab

Neighbor-joining tree constructed from the sequences of DH-like domains from characterized trans-AT and cis-AT PKSs. Colored clades correspond to different catalytic functions in trans-AT PKSs, as follows: blue - 2-enoyl thioester synthases (DHe domains); red - 2, 4-dienoyl thioester synthases (DHd domains); green - 2-enoyl thioester isomerases (DHi domains); orange - tetrahydropyran / tetrahydrofuran synthases, (DH* domains). R1, R2 and R3 are the three conserved active site residues in the diene forming DH-like domains from trans-AT PKSs and the corresponding residues in other types of DH-like domain (see Extended Data Fig. 5). The structurally characterized DH-like domain from module 11 of the difficidin PKS (labelled DifDH8) and DH-like domain from module 13 of the gladiolin PKS (labelled GbnDH5) studied here are marked with asterisks, and examples of DH-like domains where the configuration of the diene in the final product differs from the predicted 2E, 4Z-configuration of the 2,4-dienoyl thioester intermediate are marked with black squares. An X between the DH and ACP domains indicates the presence of one or more other catalytic domains. DH domains are labelled according to the order they appear in the PKS (i.e., going from the N to C-terminus, the first DH domain is labelled DH1, the second DH2 and so forth).

Supplementary Material

Supplementary Information

Acknowledgements

This research was supported by the BBSRC (BB/L021692/1, BB/K002341/1, and BB/M017982/1 to G.L.C.). M.J. and D.G. were supported by a BBSRC Discovery Fellowship (BB/R012121/1) and the EPSRC MOAC Doctoral Training Centre (EP/F500378/1), respectively. M.R.-C. was supported by a mobility grant from the Government of Andorra (AM025-AND-2015). C.H. was supported by a PhD studentship from the University of Warwick. G.L.C. was the recipient of a Wolfson Research Merit Award from the Royal Society (WM130033).

Footnotes

Author Contributions

M.J., C.H. and G.L.C. designed the study. C.H. D.M.R. and M.R.-C. synthesised the pantetheine thioesters. M.J. overproduced and purified all recombinant proteins. M.J. and C.H. conducted substrate loading and LC-MS analyses of intact proteins. X.J. conducted the cysteamine-mediated product cleavage experiments. M.J. and D.G. conducted bioinformatics analyses. M.J., C.H. and G.L.C. wrote the paper with input from the other authors.

Competing Interests

D.M.R. is an employee and shareholder of Erebagen Ltd. G.L.C. is a Non-Executive Director, shareholder, and paid consultant of Erebagen Ltd. The other authors declare no competing interests.

Data Availability

The minimum dataset necessary to interpret, verify and extend the research is provided in the main manuscript, extended data, and supplementary information. The corresponding raw data (Figures 2, 3, 4 and 5, Extended Data Figures 1 and 2, Supplementary Table 1, and Supplementary Figures 4 and 5), which were processed via standard deconvolution (intact protein mass spectrometry) and extracted ion chromatogram (LC-MS analyses) procedures, are available upon written request to the corresponding author.

References

  • 1.Newman DJ, Cragg GM. Natural products as sources of new drugs over the last 25 years. J Nat Prod. 2007;70:461–477. doi: 10.1021/np068054v. [DOI] [PubMed] [Google Scholar]
  • 2.Staunton J, Weissman KJ. Polyketide biosynthesis: a millennium review. Nat Prod Rep. 2001;18:380–416. doi: 10.1039/a909079g. [DOI] [PubMed] [Google Scholar]
  • 3.Hertweck C. The Biosynthetic Logic of Polyketide Diversity. Angew Chem Int Ed. 2009;48:4688–4716. doi: 10.1002/anie.200806121. [DOI] [PubMed] [Google Scholar]
  • 4.Keatinge-Clay AT. Stereocontrol within polyketide assembly lines. Nat Prod Rep. 2016;33:141–149. doi: 10.1039/c5np00092k. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Shah DD, You Y-O, Cane DE. Stereospecific Formation of E-and Z-Disubstituted Double Bonds by Dehydratase Domains from Modules 1 and 2 of the Fostriecin Polyketide Synthase. J Am Chem Soc. 2017;139:14322–14330. doi: 10.1021/jacs.7b08896. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Skiba MA, et al. Structural Basis of Polyketide Synthase O-Methylation. ACS Chem Biol. 2018;13:3221–3228. doi: 10.1021/acschembio.8b00687. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Walker PD, Weir ANM, Willis CL, Crump MP. Polyketide β-branching: diversity, mechanism and selectivity. Nat Prod Rep. 2021;38:723–756. doi: 10.1039/d0np00045k. [DOI] [PubMed] [Google Scholar]
  • 8.Meoded RA, et al. A Polyketide Synthase Component for Oxygen Insertion into Polyketide Backbones. Angew Chemie Int Ed. 2018;57:11644–11648. doi: 10.1002/anie.201805363. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Xie X, Cane DE. Stereospecific Formation of Z-Trisubstituted Double Bonds by the Successive Action of Ketoreductase and Dehydratase Domains from trans-AT Polyketide Synthases. Biochemistry. 2018;57:3126–3129. doi: 10.1021/acs.biochem.7b01253. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Jenner M, et al. Mechanism of intersubunit ketosynthase-dehydratase interaction in polyketide synthases. Nat Chem Biol. 2018;14:270–275. doi: 10.1038/nchembio.2549. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Li Y, et al. Functional Characterization of a Dehydratase Domain from the Pikromycin Polyketide Synthase. J Am Chem Soc. 2015;137:7003–7006. doi: 10.1021/jacs.5b02325. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Nazi I, Koteva KP, Wright GD. One-pot chemoenzymatic preparation of coenzyme A analogues. Anal Biochem. 2004;324:100–5. doi: 10.1016/j.ab.2003.09.005. [DOI] [PubMed] [Google Scholar]
  • 13.Agarwal V, et al. Chemoenzymatic Synthesis of Acyl Coenzyme A Substrates Enables in Situ Labeling of Small Molecules and Proteins. Org Lett. 2015;17:4452–4455. doi: 10.1021/acs.orglett.5b02113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Worthington AS, et al. One-pot chemo-enzymatic synthesis of reporter-modified proteins. Org Biomol Chem. 2006;4:44–46. doi: 10.1039/b512735a. [DOI] [PubMed] [Google Scholar]
  • 15.Quadri LE, et al. Characterization of Sfp, a Bacillus subtilis Phosphopantetheinyl Transferase for Peptidyl Carrier Protein Domains in Peptide Synthetases. Biochemistry. 1998;37:1585–1595. doi: 10.1021/bi9719861. [DOI] [PubMed] [Google Scholar]
  • 16.Labonte JW, Townsend CA. Active Site Comparisons and Catalytic Mechanisms of the Hot Dog Superfamily. Chem Rev. 2012;113:2182–2204. doi: 10.1021/cr300169a. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Belecki K, Townsend CA. Biochemical Determination of Enzyme-Bound Metabolites: Preferential Accumulation of a Programmed Octaketide on the Enediyne Polyketide Synthase CalE8. J Am Chem Soc. 2013;135:14339–14348. doi: 10.1021/ja406697t. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Jumper J, et al. Highly accurate protein structure prediction with AlphaFold. Nature. 2021;596:583–589. doi: 10.1038/s41586-021-03819-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Wagner DT, et al. Structural and Functional Trends in Dehydrating Bimodules from trans-Acyltransferase Polyketide Synthases. Structure. 2017;25:1045–1055. doi: 10.1016/j.str.2017.05.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Xie X, Cane DE. pH-Rate profiles establish that polyketide synthase dehydratase domains utilize a single-base mechanism. Org Biomol Chem. 2018;16:9165–9170. doi: 10.1039/c8ob02637h. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Fiers WD, Dodge GJ, Sherman DH, Smith JL, Aldrich CC. Vinylogous Dehydration by a Polyketide Dehydratase Domain in Curacin Biosynthesis. J Am Chem Soc. 2016;138:16024–16036. doi: 10.1021/jacs.6b09748. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.KE W, et al. Difficidin and oxydifficidin: novel broad spectrum antibacterial antibiotics produced by Bacillus subtilis. II. Isolation and physico-chemical characterization. J Antibiot. 1987;40:1682–1691. doi: 10.7164/antibiotics.40.1682. [DOI] [PubMed] [Google Scholar]
  • 23.Tang L, et al. Elucidating the mechanism of cis double bond formation in epothilone biosynthesis. J Am Chem Soc. 2004;126:46–47. doi: 10.1021/ja030503f. [DOI] [PubMed] [Google Scholar]
  • 24.Dodge GJ, et al. Structural and dynamical rationale for fatty acid unsaturation in Escherichia coli. Proc Natl Acad Sci. 2019;116:6775–6783. doi: 10.1073/pnas.1818686116. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Pöplau P, Frank S, Morinaka BI, Piel J. An Enzymatic Domain for the Formation of Cyclic Ethers in Complex Polyketides. Angew Chem Int Ed. 2013;52:13215–13218. doi: 10.1002/anie.201307406. [DOI] [PubMed] [Google Scholar]
  • 26.Berkhan G, Hahn F. A dehydratase domain in ambruticin biosynthesis displays additional activity as a pyran-forming cyclase. Angew Chem Int Ed. 2014;53:14240–14244. doi: 10.1002/anie.201407979. [DOI] [PubMed] [Google Scholar]
  • 27.Kusebauch B, Busch B, Scherlach K, Roth M, Hertweck C. Functionally Distinct Modules Operate Two Consecutive α,β→β,γ Double-Bond Shifts in the Rhizoxin Polyketide Assembly Line. Angew Chem Int Ed. 2010;49:1460–1464. doi: 10.1002/anie.200905467. [DOI] [PubMed] [Google Scholar]
  • 28.Gay DC, Spear PJ, Keatinge-Clay AT. A Double-Hotdog with a New Trick: Structure and Mechanism of the trans-Acyltransferase Polyketide Synthase Enoyl-isomerase. ACS Chem Biol. 2014;9:2374–2381. doi: 10.1021/cb500459b. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Berkhan G, Merten C, Holec C, Hahn F. The interplay between a multifunctional dehydratase domain and a C-methyltransferase effects olefin shift in ambruticin biosynthesis. Angew Chem Int Ed. 2016;55:13589–13592. doi: 10.1002/anie.201607827. [DOI] [PubMed] [Google Scholar]
  • 30.Wagner DT, et al. Structural and Functional Studies of a Pyran Synthase Domain from a trans-Acyltransferase Assembly Line. ACS Chem Biol. 2018;13:975–983. doi: 10.1021/acschembio.8b00049. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.N SP, et al. Mining Symbionts of a Spider-Transmitted Fungus Illuminates Uncharted Biosynthetic Pathways to Cytotoxic Benzolactones. Angew Chem Int Ed. 2020;59:7766–7771. doi: 10.1002/anie.201916007. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Information
EMS151800-supplement-Supplementary_Information.pdf (3.5MB, pdf)

Data Availability Statement

The minimum dataset necessary to interpret, verify and extend the research is provided in the main manuscript, extended data, and supplementary information. The corresponding raw data (Figures 2, 3, 4 and 5, Extended Data Figures 1 and 2, Supplementary Table 1, and Supplementary Figures 4 and 5), which were processed via standard deconvolution (intact protein mass spectrometry) and extracted ion chromatogram (LC-MS analyses) procedures, are available upon written request to the corresponding author.

RESOURCES