Insights into 6‐Methylsalicylic Acid Bio‐assembly by Using Chemical Probes

Abstract Chemical probes capable of reacting with KS (ketosynthase)‐bound biosynthetic intermediates were utilized for the investigation of the model type I iterative polyketide synthase 6‐methylsalicylic acid synthase (6‐MSAS) in vivo and in vitro. From the fermentation of fungal and bacterial 6‐MSAS hosts in the presence of chain termination probes, a full range of biosynthetic intermediates was isolated and characterized for the first time. Meanwhile, in vitro studies of recombinant 6‐MSA synthases with both nonhydrolyzable and hydrolyzable substrate mimics have provided additional insights into substrate recognition, providing the basis for further exploration of the enzyme catalytic activities.

Abstract: Chemical probes capable of reacting with KS (ketosynthase)-bound biosynthetic intermediates were utilized for the investigation of the model type Ii terative polyketide synthase 6-methylsalicylic acid synthase  in vivo and in vitro.F romt he fermentation of fungal and bacterial 6-MSAS hosts in the presence of chain termination probes,afull range of biosynthetic intermediates was isolated and characterizedf or the first time.M eanwhile,i nv itro studies of recombinant 6-MSA synthases with both nonhydrolyzable and hydrolyzable substrate mimics have provided additional insights into substrate recognition, providing the basis for further exploration of the enzyme catalytic activities.
Together with peptidic molecules,p olyketide natural products constitute one of the most abundant sources of bioderived and bio-inspired pharmaceuticals:t hese include widely utilized antibiotic and antitumor agents (for example, erythromycin and doxorubicin) and cholesterol-lowering (statin) drugs. [1] Polyketide biosynthesis resembles that of fatty acids in the use of decarboxylative Claisen condensation to generate enzyme-bound b-keto thioester intermediates, which are variably processed and ultimately converted to highly diversified products in structure and function. Polyketide synthases (PKSs) can be distinguished as either modular or iterative:t he former (best known as type I modular) are assembly lines comprising multiple sets of domains (modules), with each catalyzing at least one round of chain extension and downstream translocation for further processing. Forthese PKSs the module order and composition determine the sequence of biosynthetic events, [2] which can be predicted and manipulated to generate novel compounds. [3] Conversely iterative synthases are constituted by single enzymes (type Ia nd III) or clusters of discrete proteins (type II) that harbor al imited set of catalytic domains iteratively utilized for intermediate chain growth and elaboration. Ther apid and unpredictable nature of substrate processing makes the investigation of iterative synthases much more challenging compared to their modular counterparts.T ype Ii terative polyketide synthases (iPKSs) most closely resemble fatty acid synthases in their domain organization and modus operandi. They are typical of fungi, [4] although an increasing number have been reported in bacteria. [5] Ty pe Ii PKSs utilize ad iverse range of acyl building blocks (for example,a cetate,h exanoate,a nd benzoate) as starter units for polyketide chain building, and have been classified as non-reducing (NR), partially reducing (PR), or highly reducing (HR) according to the degree of ketone moiety processing occurring throughout product assembly. [4] Tailoring modifications can occur during and after assembly to yield the final bioactive molecules. [8] Established mechanisms of iPKS product release include cyclization and hydrolysis mediated by thioesterase (TE) and Claisen cyclase thioesterase (CLC-TE) domains,t hioester reduction, [4,5] pyrone formation, [6] and product transfer to an onribosomal peptide synthetase (NRPS) assembly line. [7] Important mechanistic insights on iPKSs have been gathered by the in vitro reconstitution of enzyme activity with putative synthetic substrates, [9] by genetic manipulation, [10] and by analysis of protein site-occupancyu sing advanced mass spectrometry of PKS-bound precursors. [11] However,k ey details of the timing of biosynthetic transformations and the basis for substrate discrimination for iPKSs remain elusive, and new tools are needed to uncover them.
6-methylsalicylic acid (6-MSA, 1,S cheme 1) was the first polyketide to be biosynthetically investigated. [12] It is produced in various fungi including Penicillium patulum,where it is ap recursor to the toxin patulin. [13] 6-MSA is also ak ey structural moiety of promising antibiotic and anticancer agents,s uch as chlorothricin, maduropeptin, and neocarzinostatin. [14] Thei PKS 6-methysalicylic acid synthase (6-MSAS) was first purified from P. patulum and characterized as a1 88 kDa tetrameric protein. [15] Early labeling experiments established that 6-MSAS requires one acetyl-CoA (2) and three malonyl-CoA molecules (3, Figure 1A)togenerate 6-MSA. [16] Analysis of its gene cluster revealed it encodes one polypeptide chain harboring ketosynthase (KS), acyltransferase (AT), dehydratase (DH), ketoreductase (KR), and acyl carrier protein (ACP) domains [17] as in avertebrate fatty acid synthase.M echanistic studies of purified 6-MSAS using substrate/intermediate analogues and enzyme inhibitors, [18] as well as enzyme mutagenesis, [14a,19] have led to two distinct biosynthetic proposals:inthe first, DH-catalyzed dehydration of a3 -hydroxytriketide intermediate is followed by af urther round of chain extension, trans to cis isomerization of adouble bond, aromatization, and finally thioester hydrolysis (Scheme 1a). [18c, 19a] In the second, a3-hydroxytriketide intermediate is directly extended to a5-hydroxytetraketide,which cyclizes,d ehydrates,a nd aromatizes prior to final product release (Scheme 1b). [18c,19b] Ar ecent study of the 6-MSAS-like enzyme ATXf rom Aspergillus terreus has supported this second route and provided evidence of involvement of as o-called thioester hydrolase (THID) domain in product release. [19b] TheT HID domain comprises the previously identified dehydratase (DH) domain together with an adjacent region termed the interdomain (ID) linker,o riginally identified as ac ore domain required for subunit-subunit interaction within ATX. [19a] THID has been shown to catalyze 6-MSA release from am utant form of ATX( H972A, which would inactivate the DH function);italso catalyzes hydrolysis of the N-acetylcysteamine thioester of 6-MSA. [19b] This suggests that enzyme-catalyzed dehydration of a3 -hydroxytriketide intermediate is not necessary for 6-MSA formation. However,i n the absence of direct evidence for the biosynthetic intermediates involved, it remains unclear whether triketide dehydration takes place and whether the sole role of THID is the hydrolytic release of enzyme-bound 6-MSA.
To obtain ac omplete mechanistic picture of 6-MSA assembly,w eh ave used chain-termination probes for the capture and identification of polyketide intermediates. [20,21] By competing with ACP-bound malonate extension units for the growing polyketide chain, the chemical probes react with enzyme-bound intermediates and off-load them for LC-MS characterization (Figure 1). Theuse of these tools has already allowed fresh insights into the timing and the mechanism of modular assembly-line biosynthesis in vitro [20] and in vivo, [21] Scheme 1. Overview of biosynthetic hypotheses leading to 6-methylsalicylic acid (6-MSA, 1): a) enzymatic dehydration of a3-hydroxytriketide followed by further chain extension leads to ad ehydrated enzyme-bound tetraketide, eventually giving 1;b)formation of a5-hydroxytetraketide eventually leading to 1 via TH-mediated hydrolysis. [18c, 19b] 6-MSAS comprises ketosynthase (KS), acyltransferase (AT), thioester hydrolase (THID), [19b] ketoreductase (KR), and acyl carrier protein (ACP) domains. Figure 1. A) Chain terminationp robes 9-12,generated in vivo from the hydrolysis of the corresponding esters 4-8, [21] compete with ACP-bound malonate to off-load 6-MSAS-bound intermediates in P. patulum, E. coli heterologously expressing 6-MSAS, [22] and S. antibioticus DSM40725 (Supporting Information, Tables 1S-3S);B)LC-HR-MS detection of putative hydroxytetraketides captured from P. patulum;C )HR-MS n analyses of aputative hydroxypentaketide resulting from the off-loading of aKS-bound hydroxytetraketide in P. patulum. and, more recently,h as unveiled novel opportunities for the generation of unnatural polyketide derivatives. [21c] We initially used our intermediate-capturing probes for in vivo studies on fungal and bacterial strains harboring 6-MSAS genes,i ncluding the natural 6-MSA producer P. patulum,a n engineered E. coli host strain heterologously expressing P. patulum 6-MSAS (E. coli BAP1 pKOS007-109), [22] and S. antibioticus DSM40725 (producer of chlorothricin). [14a] Each strain was grown in the presence of substrates 4-8,which are hydrolyzed in vivo to the corresponding carboxylates 9-12 ( Figure 1A;Supporting Information, Figure 2S). [20] Theo verall outcome of these in vivo experiments is illustrated in Figure 1( for details,s ee the Supporting Information, Tables 1S-3S and following figures). In most of the ethyl acetate extracts from both fungal and bacterial hosts, as eries of trapped intermediates,i ncluding diketides,t riketides,r educed triketides,a nd ar ange of putative hydroxy, dehydrated, and aromatized tetraketides,w ere identified by HR-LC-MS:t hese would directly reflect the nature of ACPbound substrates in 6-MSA assembly.B esides,p utative hydroxy,d ehydrated, and aromatized pentaketides arising from the off-loading of 6-MSAS-bound tetraketides were also identified ( Figure 1C and the Supporting Information). All the captured intermediates,a bsent in control samples,w ere characterized by MS n analysis,s howing diagnostic peaks resulting from the loss of N-acyl chains and subsequent cyclic imine formation (Figure 1Cand the Supporting Information). Forthe majority of the putative tetraketides and pentaketides, multiple peaks were observed:t hese may arise from isomerization, cyclization and dehydration events which can be spontaneous or enzyme-catalyzed. Ad istinction between hydroxy,d ehydrated, and aromatized advanced species was made on the basis of variable LC retention times as well as of detected accurate masses.O nthe same basis,distinct species with masses corresponding to dehydrated triketides could not be identified. From the lack of direct evidence for dehydrated triketides and the identification of the putative hydroxy tetraketides and pentaketides,i ta ppears that, whether the PKS is of fungal or of bacterial origin, route b) of Scheme 1is followed.
Nonetheless,t os eek additional confirmation of these insights and further dissect 6-MSAS catalytic activities,w e also utilized recombinant P. patulum 6-MSAS from heterologous E. coli BAP1 host strain, [23] as well as an additional mutant form of the enzyme (6-MSAS H958A) bearing an alanine in place of ah istidine in the THID active site for in vitro assays. [22] Thec apture of biosynthetic intermediates in vitro proved much more challenging than in vivo.U sing probes 9a-b (generated from pig liver esterase-assisted hydrolysis of 4a-b), [19b] only intermediates from two rounds of chain extension were consistently identified in the ethyl acetate extracts of 6-MSAS assays (Supporting Information, Figure 43S). When recombinant 6-MSAS was primed with acetoacetyl-CoA instead of acetyl-CoA in the attempt to improve advanced intermediate capture,the accumulation of ap ossibly dehydrated triketide was observed (Supporting Information, Figure 46S).
To investigate whether this species could have been enzymatically formed, aracemic 3-hydroxytriketide substrate mimic 13 was synthesized in five steps from 4a (Supporting Information) and tested as as ubstrate for both 6-MSAS and its H958A mutant. In neither case was 13 enzymatically dehydrated (Figure 2A;S upporting Information, Figure 47S);r ather, we found that purified 13 dehydrated over long-term storage.S urprisingly,w hen 6-MSAS and the H958A mutant were incubated with the N-acetylcysteamine thioester of 6-MSA 14, previously utilized to probe the THID function in ATX, [12] no free 6-MSA was generated (Figure 2B;Supporting Information, Figure 48S). An N-decanoyl thioester analogue 15 was synthesized as an additional substrate,w ith the idea of utilizing al ong acyl chain to mimic the phosphopantetheine cofactor of ACP. However, 15 was also resistant to hydrolysis by either enzyme (Supporting Information, Figure 49S). This unexpected outcome suggests that 6-MSAS differs from ATXinthat covalent attachment of tetraketide intermediates to the ACPo rc oenzyme Am ight be necessary for their processing.A lternatively,t his may indicate that readily aromatized thioesters are not true substrates for THID domains.T he results reported herein strongly point towards no enzymatic dehydration taking place at atriketide stage of 6-MSA assembly,soloss of water must occur at the tetraketide stage.Ifthis is enzyme-catalyzed, the configuration of the resulting alkene (from an R-alcohol, as recently established for the 6-MSA-like mellein synthase) [24] would likely be trans. On the basis of multiple peaks observed for dehydrated tetraketides (for example,S upporting Information, Figures 24S), it is tempting to speculate that, along with final thioester hydrolysis,t he THID might act as at emplate domain to aid trans to cis double bond isomerization, and/or cyclization and aromatization;h owever, this remains undetermined. Amongst iPKSs with significant homology to 6-MSAS and leading to 6-MSA related products, the THID domain is highly conserved (Supporting Information, Figure 50S). Its presence in non-reducing iPKSs such as the orsellinic acid synthase supports its role in product cyclization and release. [14a] Intriguingly,aTHID is not present in MicC,a ni PKS responsible for the formation of 6pentasalicylic acid in micacocidin biosynthesis,whose assembly allegedly proceeds similarly to that of 6-MSA. [7b, 25] Although bioinformatic analysis usefully pinpoints the sim-  Figure 51S), it cannot be used to infer subtle differences in substrate recognition and processing between these multienzymes.F or this the use of chemical probes appears apromising approach towards afuller understanding and exploitation of iPKS catalysis.T hrough the use of the chain termination probes 9-12 we have indeed gathered the first direct and comprehensive evidence for the course of substrate processing on at ype Ii terative PKS.T he probes were successfully used in af ungal host and both Grampositive and Gram-negative bacteria, with the putative intermediates providing preliminary insights into the kinetics of 6-MSA assembly.Ingeneral, the first two rounds of chain extension/processing appear relatively slow,w hereas more distinct differences can be observed in advanced intermediate accumulation in the different 6-MSAS hosts (Supporting Information, Tables 1S-3S and following analyses): these may depend on probe efficiencyand uptake to some extent, as well as on specific kinetic programming of product assembly within ap articular host. Significantly 6-MSAS displayed unexpected substrate flexibility,i nt hat it was able to accept malonate surrogates of different chain lengths and bearing various functionalities (including alkyne and fluorine moieties) at different positions and for every round of chain extension, and also to generate novel pentaketide products. This opens the possibility of utilizing iPKSs for the generation of novel unnatural products employing different extender [21c] as well as starter units. [26]