Exploiting the Biosynthetic Potential of Type III Polyketide Synthases Review Yan Pingexploiting Lim 1,2, Maybelle the K

molecules

Review Exploiting the Biosynthetic Potential of Type III Polyketide Synthases Review Yan PingExploiting Lim 1,2, Maybelle the K. Biosynthetic Go 1,2 and Wen Shan Potential Yew 1,2,* of Type III 1 DepartmentPolyketide of Biochemistry, Synthases Yong Loo Lin School of Medicine, National University of Singapore, 8 Medical Drive, Singapore 117597, Singapore; [email protected] (Y.P.L.); [email protected] (M.K.G.) 2 NUSYan Ping Synthetic Lim 1,2 Biology, Maybelle for Clinical K. Go 1,2 and and Technological Wen Shan Yew Innovation, 1,2,* Centre for Life Sciences, National University of Singapore, 28 Medical Drive, Singapore 117456, Singapore 1 Department of Biochemistry, Yong Loo Lin School of Medicine, National University of Singapore, * Correspondence:8 Medical Drive, [email protected]; 117597, Singapore; [email protected] Tel.: +65-6516-8624 (Y.P.L.); [email protected] (M.K.G.) 2 Academic NUS Editor: Synthetic Kira Biology Weissman for Clinical and Technological Innovation, Centre for Life Sciences, National University of Singapore, 28 Medical Drive, 117456, Singapore Received: 20 May 2016; Accepted: 17 June 2016; Published: 22 June 2016 * Correspondence: [email protected]; Tel.: +65-6516-8624

Abstract:AcademicPolyketides Editor: Kira areWeissman structurally and functionally diverse secondary metabolites that are biosynthesizedReceived: 20 May by polyketide2016; Accepted: synthases 17 June 2016; (PKSs) Published: using June acyl-CoA 2016 precursors. Recent studies in the engineering and structural characterization of PKSs have facilitated the use of target enzymes as biocatalystsAbstract: to Polyketides produce novel are structurally functionally and optimized functiona polyketides.lly diverse secondary These compounds metabolites maythat serveare as biosynthesized by polyketide synthases (PKSs) using acyl-CoA precursors. Recent studies in the potential drug leads. This review summarizes the insights gained from research on type III PKSs, engineering and structural characterization of PKSs have facilitated the use of target enzymes as frombiocatalysts the discovery to produce of chalcone novel functionally synthase in opti plantsmized to polyketides. novel PKSs These in bacteriacompounds and may fungi. serve To as date, at leastpotential 15 families drug leads. of type This III review PKSs havesummarizes been characterized, the insights gained highlighting from research the utilityon type of III PKSs PKSs, in the developmentfrom the discovery of natural of product chalcone libraries synthase for in plants therapeutic to novel development. PKSs in bacteria and fungi. To date, at least 15 families of type III PKSs have been characterized, highlighting the utility of PKSs in the Keywords:developmentchalcone of natural synthase; product stilbene libraries synthase;for therapeutic polyketide; development. precursor-directed combinatorial biosynthesis; synthetic enzymology Keywords: chalcone synthase; stilbene synthase; polyketide; precursor-directed combinatorial biosynthesis; synthetic enzymology

1. Introduction Substantial1. Introduction focus has been awarded to polyketide research since its initial discovery about a century agoSubstantial [1]. Polyketides focus has been are structurally awarded to polyketide and functionally research since diverse its initial secondary discovery metabolites about a century produced in bacteria,ago [1]. fungi,Polyketides and are plants. structurally Many and of thesefunctionally bioactive diverse natural secondary products metabolites have produced signiﬁcant in bacteria, medical or agriculturalfungi, and applications plants. Many [2], of such these as bioactive rapamycin natural (a macrolide products immunosuppressant)have significant medical [ 3or], lovastatinagricultural (used for theapplications treatment of[2], hypercholesterolemia) such as rapamycin (a macrolide [4], actinorhodin immunosuppressant) (an antibiotic) [3], [5 lovastatin], and resveratrol (used for (reported the anti-ageingtreatment properties of hypercholesterolemia) in model organisms) [4], actinorhod [6] (Figurein1 ).(an This antibiotic) highlights [5], theand vital resveratrol role polyketides (reported and theiranti-ageing derivatives properties can play in model the development organisms) [6] of (Figure novel therapeutics 1). This highlights [7]. Although the vital role chemical polyketides synthesis and their derivatives can play in the development of novel therapeutics [7]. Although chemical synthesis can be used to produce polyketide analogues [8], the chemical and structural complexity of polyketides can be used to produce polyketide analogues [8], the chemical and structural complexity of polyketides and their derivatives prevent routine production by this route. and their derivatives prevent routine production by this route.

Molecules 2016, 21, 806; doi:10.3390/molecules21060806Figure 1. Cont. www.mdpi.com/journal/molecules

Molecules 2016, 21, 806; doi:10.3390/molecules21060806 www.mdpi.com/journal/molecules Molecules 2016, 21, 806 2 of 37

Molecules 2016, 21, 806 2 of 36

FigureFigure 1. Examples 1. Examples of polyketidesof polyketides synthesized synthesized byby the different different classes classes of ofPKSs. PKSs.

Insights intoFigure biosynthetic 1. Examples mechanisms of polyketides of synthesized polyketides by only the different achieved classes considerable of PKSs. progress from Insights1953 when into Birch biosynthetic and Donovan mechanisms hypothesized of polyketidesa potential biosynthetic only achieved pathway considerable similar to that progress of fatty from 1953 whenacidsInsights [9]. Birch Further into and advancementsbiosynthetic Donovan mechanisms hypothesized in the polyketide of polyketides a field potential led to only the biosynthetic consensusachieved considerablethat pathway polyketides similarprogress are typicallyto from that of fatty acids1953biosynthesized when [9]. FurtherBirch through and advancements Donovan successive hypothesized decarboxylative in the polyketidea potentia condensationsl biosynthetic ﬁeld led of tocoenzyme pathway the consensus Asimilar (CoA)-derived to that that polyketides of units,fatty are typicallyacidsinto a[9]. complex Further biosynthesized polycyclic advancements multi-carbon through in the polyketide successivecompound field cont decarboxylative ledaining to the keto consensus or hydroxyl condensations that groupspolyketides [10]. of areDownstream coenzyme typically A (CoA)-derivedbiosynthesizedmodifying units, enzymes through into are asuccessive complexthen used decarboxylative polycyclicto obtain flavonoi multi-carbon condensationsds and other compound of bioactive coenzyme containingsubstances. A (CoA)-derived Recent keto or studies units, hydroxyl groupsintoin [ 10the a ].complex engineering Downstream polycyclic and modifying structural multi-carbon characterization enzymes compound are cont then of ainingpolyketide used keto to or obtainsynthases hydroxyl ﬂavonoids (PKSs) groups have [10]. and facilitatedDownstream other bioactive the modifying enzymes are then used to obtain flavonoids and other bioactive substances. Recent studies substances.use of target Recent enzymes studies as in biocatalysts the engineering to produce and novel structural functionally characterization optimized polyketides, of polyketide which synthases can inserve the engineeringas potential anddrug structural leads [11,12]. characterization of polyketide synthases (PKSs) have facilitated the (PKSs)use have of target facilitated enzymes theas biocatalysts use of target to produce enzymes novel as functionally biocatalysts optimized to produce polyketides, novel which functionally can optimizedserve as polyketides, potential drug which leads can [11,12]. serve as potential drug leads [11,12].

Figure 2. Domain organization of the different PKSs. Putative domains are represented by circles. In modular type I PKSs, functional domains are organized into several modules, with each module being Figureresponsible 2. Domain for a singleorganization decarboxylative of the different condensation PKSs. Puta steptive in polyketidedomains are formation. represented For byiterative circles. type In Figure 2. Domain organization of the different PKSs. Putative domains are represented by circles. modularI PKSs, the type functional I PKSs, functional domains domainsare clustered are organized in a single into module, several and modules, each domain with each is used module repeatedly being In modular type I PKSs, functional domains are organized into several modules, with each module being responsibleduring polyketide for a single synthesis. decarboxylative Type II PKSs condensation are dissociable step multi-enzyme in polyketide formation.complexes, For with iterative each protein type responsible for a single decarboxylative condensation step in polyketide formation. For iterative type I bearingPKSs, the a functionalsingle and domains independent are clustered catalytic in adomain single module, that is andused each iteratively domain isduring used repeatedlypolyketide I PKSs,duringformation. the functionalpolyketide Reactions synthesis. domains by type Type are III clusteredPK II SsPKSs are arealso in dissociable aiterative single module,but multi-enzyme do not and require each complexes, an domain ACP forwith is the usedeach attachment protein repeatedly duringbearingof polyketide the growing a single synthesis. polyketide and independent Type chain. II PKSsAT: catalytic Acyltransferas are dissociable domaine; ACP:that multi-enzyme is Acyl used carrier iteratively complexes, protein; during KS: with Ketosynthase; polyketide each protein bearingformation.KR: a singleKetoreductase; Reactions and independent DH: by type Dehydratase; III catalytic PKSs are ER: domain also Enoyl iterative thatreductase; isbut used do TE: not iteratively Thioesterase; require an during ACP CoA: for polyketide Coenzyme the attachment formation.A. Reactionsof the by growing type III polyketide PKSs are alsochain. iterative AT: Acyltransferas but do not requiree; ACP: anAcyl ACP carrier for theprotein; attachment KS: Ketosynthase; of the growing polyketideKR:Generally, Ketoreductase; chain. PKSs AT: Acyltransferase; areDH: grouped Dehydratase; into ACP: ER:three AcylEnoyl classes: carrierreductase; type protein; TE:I, II, Thioesterase; KS:or III Ketosynthase; PKS (FigureCoA: Coenzyme KR:2) [2]. Ketoreductase; Type A. I PKSs DH:are Dehydratase;large multi-domain ER: Enoyl polypeptides reductase; TE:which Thioesterase; can function CoA: in Coenzyme either a modular A. or iterative manner, shuttlingGenerally, substrates PKSs betweenare grouped functional into three domains classes: via type acyl I,carrier II, or IIIproteins PKS (Figure (ACPs). 2) In [2]. modular Type I PKSstype I arePKSs, large functional multi-domain domains polypeptides are organized which into can several function modules, in either with a eachmodular module or iterativebeing responsible manner, Generally,shuttling substrates PKSs are between grouped functional into three domains classes: via typeacyl carrier I, II, or proteins III PKS (ACPs). (Figure In2 modular)[ 2]. Type type I I PKSs are largePKSs, multi-domain functional domains polypeptides are organized which into can several function modules, in eitherwith each a modular module being or iterative responsible manner, shuttling substrates between functional domains via acyl carrier proteins (ACPs). In modular type I PKSs, functional domains are organized into several modules, with each module being responsible for a single decarboxylative condensation step in the formation of the polyketide [2,13]. Conversely for Molecules 2016, 21, 806 3 of 37 iterative type I PKSs, the functional domains are clustered in a single module; hence, each domain is used repeatedly during polyketide synthesis [14]. Type II PKSs are dissociable multi-enzyme complexes, with each protein in the complex bearing a single and independent catalytic domain that is used iteratively during polyketide formation [13]. Type III PKSs are homodimeric enzymes and are structurally simpler and mechanistically different from type I and II PKSs. The relative simplicity, versatility, and unusually broad substrate speciﬁcity of type III PKSs make them ideal candidates for the engineering of biocatalysts, and in doing so, access to bioactive polyketide compound libraries can be readily made available [15]. This review highlights recent progress in the understanding of type III PKSs (in particular with regards to the diversity of naturally occurring reactions and products discovered), and summarizes strategies for exploiting the biosynthetic potential of type III PKSs. Approaches based on random mutagenesis or site-directed mutagenesis of active site cavities to engineer the substrate speciﬁcity and activity of type III PKSs will also be discussed. These studies serve as a primer towards the development of novel and valuable polyketides with therapeutic potential.

2. Insights into Type III PKSs Type III PKSs were previously only found in plants, but in the past 15 years, more bacterial type III PKSs were characterized [16–18]; fungal type III PKSs were only recently discovered [19–21]. Unlike the type I and II PKSs, type III PKSs generally utilize CoA thioesters as substrates and do not entail the involvement of ACP domains. Furthermore, type III PKSs are able to accomplish an entire series of decarboxylative condensations and cyclization reactions in a single active site [2]. By having a simple architecture, it enables type III PKSs to be amenable to in vitro manipulation and examination. Although type III PKSs have been studied since the 1980s [22], structural information, on how a single active site can direct the chemistry of several decarboxylative condensations and subsequent cyclization of the polyketide, was only made known after the determination of the crystal structure of chalcone synthase (CHS) from Medicago sativa (commonly known as alfalfa) [23]. This was the ﬁrst crystal structure of a PKS and it facilitated a framework for engineering type III PKSs to generate novel polyketides.

2.1. Chalcone Synthases and the Basis of Polyketide Synthesis CHSs are the most well studied type III PKSs as they are ubiquitous in higher plants and essential for plant metabolism and defense [15]. Based on the structure of CHS from alfalfa (PDB ID: 1CGK), it was found that the symmetric dimer encapsulates a CoA-binding tunnel, a coumaroyl-binding pocket, and a cyclization pocket in order to facilitate chalcone biosynthesis [23]. Together with the conserved Cys164-His303-Asn336 catalytic triad, and the ‘gatekeeper’ Phe215 which is hypothesized to aid in the orientation of substrates during polyketide chain elongation, these connected cavities constitute the active site architecture of a typical type III PKS [24,25]. Chalcone synthesis by CHS is initiated by the loading of p-coumaroyl-CoA (1) onto the sulfhydryl group at the active site cysteine residue. The reaction then proceeds on with three iterative decarboxylative condensations of the extender substrate, malonyl-CoA (2), with the cysteine-bound starter unit, aided by the active site histidine and asparagine residues. After chain extension, the linear tetraketide, p-coumaroyl triacetyl thioester (4), undergoes an intramolecular C6 to C1 Claisen condensation and cyclizes into chalcone 7, which undergoes a further Michael-type ring closure to give rise to naringenin (8) (Scheme1)[24]. Numerous type III PKSs have been identified till date, and the range of reactions and product classes discovered is due to differences in intramolecular cyclizations of the linear polyketide intermediate, the number of polyketide extension steps involved, and the choice of starter and extender CoAs used [26]. The diverse polyketide scaffolds catalyzed by type III PKSs can be classified into chalcone, stilbene, pyrone, chromone, stilbenecarboxylate, bibenzyl, benzalacetone, curcuminoid, benzophenone, biphenyl, acridone, quinolone, phloroglucinol, resorcinol, and naphthalene families [24,27], respectively. By characterizing various type III PKSs, we can understand reaction mechanisms better and thus, direct the generation of unnatural polyketides using synthetic enzymology [28]. Molecules 2016, 21, 806 4 of 37 Molecules 2016, 21, 806 4 of 36

Scheme 1. Chalcone and naringenin synthesis catalyzed by M. sativa CHS or Pinus sylvestris STS. Scheme 1. Chalcone and naringenin synthesis catalyzed by M. sativa CHS or Pinus sylvestris STS. 2.2. Stilbene Synthases and the Functional Divergence of Type III PKSs 2.2. Stilbene Synthases and the Functional Divergence of Type III PKSs Stilbene synthases (STSs) are found in only a small subset of unrelated plants, unlike CHSs, which areStilbene ubiquitous. synthases Phylogenetic (STSs) studies are found show inthat only no ancestral a small STS subset gene was of unrelated found, and plants, STSs are unlike likely to CHSs, whichhave are evolved ubiquitous. from CHSs Phylogenetic independently studies on showseveral that occasions no ancestral [29]. STSs STS and gene CHSs was are found, structurally and STSs are likelyrelated to and have share evolved at least 65% from amino CHSs acid independently sequence identity on [30]. several These occasions enzymes follow [29]. STSsa similar and overall CHSs are structurallyreaction relatedmechanism and which share involves at least initiation, 65% amino followed acid by sequence the iterative identity decarboxylative [30]. These condensation enzymes follow of a three units of malonyl-CoA to a starter unit, and cyclization of the linear tetraketide intermediate. similar overall reaction mechanism which involves initiation, followed by the iterative decarboxylative Slight variations in active site residues enable STSs to catalyze an intramolecular C2 to C7 aldol condensationcondensation, of three followed units by of decarboxylation malonyl-CoA and to adehydration starter unit, to andgenerate cyclization a stilbene of scaffold the linear instead tetraketide of intermediate.a chalcone Slightbackbone variations (Scheme in 1) active[26]. site residues enable STSs to catalyze an intramolecular C2 to C7 aldol condensation,Although it has followedbeen postulated by decarboxylation that divergence and wasdehydration due to some steric to generate changes a of stilbene the active scaffold insteadsite ofcavity a chalcone affecting backbone the orientation (Scheme of the1 linear)[26]. tetraketide intermediate, the questions of how CHS and STSAlthough can produce it has resveratrol been postulated (3,5,4′-trihydroxystilbene, that divergence was6) and due naringenin to some ( steric8) respectively changes as of minor the active site cavityproducts affecting (Scheme the 1) orientation[30], and how of decarboxylatio the linear tetraketiden occurs in intermediate, STSs, remain theunanswered. questions Previous of how CHS and STShomology can produce modeling resveratrol studies have (3,5,4 shown1-trihydroxystilbene, no substantial structural6) and or naringenin chemical differences (8) respectively in the STS as minor active site, compared to CHS, and no obvious STS or CHS consensus sequence was identified [26,29]. products (Scheme1)[ 30], and how decarboxylation occurs in STSs, remain unanswered. Previous With the report of the first STS crystal structure from Pinus sylvestris (Scots pine) in 2004 (PDB ID: homology modeling studies have shown no substantial structural or chemical differences in the STS 1U0U), the mechanism of stilbene synthesis was finally resolved [26,27]. active site,Structure-based compared to CHS,mutagenic and notransformation obvious STS of or alfalfa CHS consensusCHS into sequenceSTS established was identified that STS’s [26 ,29]. Withalternative the report cyclization of the first specificity STS crystal is due structure to a conformational from Pinus backbone sylvestris change(Scots in pine) a short, in concealed 2004 (PDB ID: 1U0U),loop the between mechanism residues of stilbene132 and 137 synthesis in alfalfa was CHS, finally that resolvedspans the [dimer26,27]. interface in the active site. SeveralStructure-based amino acid mutagenicsubstitutions transformationat buried sites adjace ofnt alfalfa to this loop CHS bring into about STS a establishedslight displacement that STS’s alternative cyclization specificity is due to a conformational backbone change in a short, concealed loop between residues 132 and 137 in alfalfa CHS, that spans the dimer interface in the active site. Several amino acid substitutions at buried sites adjacent to this loop bring about a slight displacement of Molecules 2016, 21, 806 5 of 37

Thr132,Molecules a residue 2016, 21, 806 that is conserved in both CHS and STS. Consequently, the side chain5 hydroxyl of 36 group of Thr132 is brought within the hydrogen bonding distance of a Ser338-stabilized water of Thr132, a residue that is conserved in both CHS and STS. Consequently, the side chain hydroxyl molecule adjacent to the catalytic cysteine, and a different active site hydrogen bond network involving group of Thr132 is brought within the hydrogen bonding distance of a Ser338-stabilized water molecule Glu192-Thr132-Hadjacent to the2O-Ser338 catalytic cysteine, forms in and STS a (Figuredifferent3)[ active24,26 site,31]. hydrogen The nucleophilic bond network water involving activated by this “aldol-switch”Glu192-Thr132-H hydrogen2O-Ser338 bondforms network in STS (Figure subsequently 3) [24,26,31]. participates The nucleophilic in the cleavage water activated of the thioester by bondthis of “aldol-switch” the cysteine-bound hydrogen linear bond tetraketide network subsequently intermediate. participates With thein the absence cleavage of of the the esterthioester bond at C1, thebond intermediate of the cysteine-bound is able to linear undergo tetraketide a C2 interm to C7ediate. aldol With condensation the absence and of the ultimately ester bond generates at C1, a stilbenethe intermediate scaffold. Intriguingly, is able to undergo when a an C2 ester to C7 bond aldol iscondensation present at C1,and aultimately C6 to C1 generates Claisen a condensation stilbene prevailsscaffold. [26,32 Intriguingly,]. Hence inwhen the an case ester of bond CHS is presen and STS,t at C1, the a functionalC6 to C1 Claisen divergence condensation is determined prevails by electronic[26,32]. factors Hence rather in the case than of steric CHS and effects STS, of the the function activeal site divergence cavity. This is determined also explains by electronic the cross-reactivity factors of CHSrather and than STS steric previously effects of identiﬁedthe active site by cavity. Yamaguchi This also and explains co-workers the cross-reactivity [30]. Nevertheless, of CHS and the STS amino previously identified by Yamaguchi and co-workers [30]. Nevertheless, the amino acid substitutions acid substitutions which result in a displacement of Thr132 are varied in different STSs, probably due which result in a displacement of Thr132 are varied in different STSs, probably due to gene duplication to geneand duplication convergent andevolution. convergent Further evolution. biochemical Further and structural biochemical studies and on structural similar proteins studies will on be similar proteinsrequired will beto delineate required the to molecular delineate basis the molecular in greater detail basis [31]. in greater detail [31].

Figure 3. The hydrogen bond network involving Glu192, Thr132, Ser338, and H2O. Differences in this Figure 3. The hydrogen bond network involving Glu192, Thr132, Ser338, and H2O. Differences in this network switches the chemistry of the active site of M. sativa CHS or P. sylvestris STS between chalcone network switches the chemistry of the active site of M. sativa CHS or P. sylvestris STS between chalcone and stilbene synthesis. and stilbene synthesis. 2.3. Truncation Products and Chain Length Specificity 2.3. Truncation Products and Chain Length Speciﬁcity Some CHSs and STSs are promiscuous in terms of the number of elongation steps, and polyketide Someintermediates CHSs can and become STSs derailed are promiscuous without forming in termsthe final of chalcone the number or stilbene of product elongation respectively steps, and polyketide[24]. These intermediates intermediates can usually become unde derailedrgo an intramolecular without forming C5-oxygen the final to C1 chalcone lactonization or stilbene to give rise product respectivelyto heterocyclic [24]. These truncation intermediates products. In usually the case undergo of alfalfa an CHS intramolecular and Scots pine C5-oxygen STS, derailed to triketide C1 lactonization and tetraketide intermediates form bisnoryangonin (3) and coumaroyl triacetic acid lactone (CTAL, 5), to give rise to heterocyclic truncation products. In the case of alfalfa CHS and Scots pine STS, derailed respectively (Scheme 1). Although lactone derailment products can form spontaneously from linear triketide and tetraketide intermediates form bisnoryangonin (3) and coumaroyl triacetic acid lactone polyketide intermediates in acidic solutions [24,30], a 2-pyrone synthase (2-PS) isolated from Gerbera (CTAL,hybrida5), respectively(daisy) was identified (Scheme to1 ).be Althoughable to control lactone the final derailment length of productsthe polyketide can and form thus, spontaneously was the fromfirst linear type polyketide III PKS to intermediateshave a different in pr acidicoduct solutionsprofile from [24 CHS,30], and a 2-pyrone STS [33,34]. synthase (2-PS) isolated from Gerbera hybridaDespite(daisy) 2-PS and was CHS identified sharing at to least be able 70% toamino control acid the sequence final length identity of and the identical polyketide catalytic and thus, was theresidues, first type 2-PS III catalyzes PKS to the have condensation a different productof an acetyl-CoA profilefrom starter CHS molecule and STS with [33 only,34]. two units of Despitemalonyl-CoA 2-PS to and form CHS a triketide, sharing which at leastfurther 70% undergoes amino lactonization acid sequence to generate identity 6-methyl-4- and identical catalytichydroxy-2-pyrone residues, 2-PS (9, Figure catalyzes 4) [34,35]. the condensation A closer inspection of an of acetyl-CoA the crystal structure starter of molecule 2-PS (PDB with ID: only two units1QLV) of revealed malonyl-CoA that the active to form site a of triketide, 2-PS is on whichly approximately further undergoes one third lactonizationof the volume seen to generate in CHS, owing to the greater steric bulk of three active site residues [24]. Intriguingly, the triple mutant 6-methyl-4-hydroxy-2-pyrone (9, Figure4)[ 34,35]. A closer inspection of the crystal structure of 2-PS (PDB ID: 1QLV) revealed that the active site of 2-PS is only approximately one third of the volume seen Molecules 2016, 21, 806 6 of 37 in CHS,Molecules owing 2016 to, 21 the, 806 greater steric bulk of three active site residues [24]. Intriguingly, the triple6 of mutant36 T197L/G256L/S338I of alfalfa CHS was found to be functionally identical to 2-PS. The increased steric T197L/G256L/S338I of alfalfa CHS was found to be functionally identical to 2-PS. The increased steric bulk at positions 197 and 338 occludes the coumaroyl-binding pocket, thus excluding big starter units bulk at positions 197 and 338 occludes the coumaroyl-binding pocket, thus excluding big starter units such as phenylpropanoid CoA thioesters. Substitution of Gly256 in CHS with Leu261 in 2-PS reduces such as phenylpropanoid CoA thioesters. Substitution of Gly256 in CHS with Leu261 in 2-PS reduces the sizethe ofsize the of cyclizationthe cyclization pocket pocket of of the the active active site, site, therefore therefore limitinglimiting the the number number of of chain chain extensions extensions of the polyketideof the polyketide [24,35 [24,35].].

FigureFigure 4. Comparing 4. Comparing residues residues lining lini theng activethe active site ofsiteG. of hybrida G. hybrida2-PS 2-PS and andM. sativa M. sativaCHS. CHS. (A) 2-PS(A) 2-PS catalyzes catalyzes the condensation of acetyl-CoA with two units of malonyl-CoA to form 6-methyl-4-hydroxy- the condensation of acetyl-CoA with two units of malonyl-CoA to form 6-methyl-4-hydroxy-2-pyrone; 2-pyrone; (B) the structures of 2-PS (pink, PDB ID: 1QLV) and CHS (blue, PDB ID: 1CGK) were (B) the structures of 2-PS (pink, PDB ID: 1QLV) and CHS (blue, PDB ID: 1CGK) were aligned to compare aligned to compare the active site cavity between the two structures. The presence of the three bulkier the activeresidues site in cavity 2-PS (L202, between L261, the and two I343) structures. show that The a presencetetraketide of such the threeas naring bulkierenin residueswill not fit in inside 2-PS (L202, L261,the and active I343) site. show Thus, that only a tetraketidea smaller polyketide, such as naringenin such as 6-methyl-4 will not-hydroxy-2-pyrone, fit inside the active may site. be formed. Thus, only a smaller polyketide, such as 6-methyl-4-hydroxy-2-pyrone, may be formed. Conversely, modifications in the active site cavity may also enable the formation of longer polyketides.Conversely, Aloesone modiﬁcations synthase in (ALS) the active from siteRheum cavity palmatum may also(rhubarb) enable catalyzes the formation six successive of longer decarboxylative condensations of malonyl-CoA with an acetyl-CoA starter molecule, followed by an polyketides. Aloesone synthase (ALS) from Rheum palmatum (rhubarb) catalyzes six successive intramolecular C8 to C13 aldol condensation and decarboxylation to generate a heptaketide aloesone decarboxylative condensations of malonyl-CoA with an acetyl-CoA starter molecule, followed by an (2-acetonyl-7-hydroxy-5-methylchromone) (10) (Scheme 2). As with other plant type III PKSs, ALS intramolecularshares more C8 than to C1360% aldolidentity condensation with alfalfa CHS, and decarboxylation and maintains the to Cys-His-Asn generate a heptaketidecatalytic triad aloesone at (2-acetonyl-7-hydroxy-5-methylchromone)positions 164, 303, and 336 respectively (numbering (10) (Scheme in alfalfa2). AsCHS) with [36,37]. other Interestingly, plant type IIIALS PKSs, lacks ALS sharesthe more conserved than CHS 60% residues identity at with positions alfalfa 197, CHS, 256, andand 338, maintains thus affecting the Cys-His-Asn the shape and catalytic volume triadof at positionsthe initiation 164, 303, and and elongation 336 respectively cavity of the (numbering active site. inIn alfalfafact, these CHS) three [36 residues,37]. Interestingly, are also sterically ALS lacks the conservedaltered in the CHS functionally residues divergent at positions G. hybrida 197, 256, 2-PS and (T197L/G256L/S338I) 338, thus affecting [35]. the Further shape research and volume on of the initiationnovel type and III PKSs elongation in the medicinal cavity of plant the activeAloe arborescens site. In fact, (aloe) these also revealed three residues similar arereplacements also sterically alteredin inthe the pentaketide functionally chromone divergent synthaseG. hybrida (PCS)2-PS (T197M/G256L/S338V) (T197L/G256L/S338I) [38], heptaketide [35]. Further synthase research on (PKS3) (T197A/G256L/S338I) [39], and octaketide synthase (OKS) (T197G/G256L/S338V) [40]. These novel type III PKSs in the medicinal plant Aloe arborescens (aloe) also revealed similar replacements enzymes also catalyze successive decarboxylative condensations of malonyl-CoA with an acetyl-CoA in thestarter pentaketide molecule chromoneto form 5,7-dihydroxy-2-methylchromone synthase (PCS) (T197M/G256L/S338V) (11) (pentaketide), [38], aloesone, heptaketide and SEK4 synthase (PKS3)(12 (T197A/G256L/S338I))/SEK4b (13) (octaketide) [respectively39], and octaketide (Scheme synthase2). 2-PS, PCS, (OKS) PKS3, (T197G/G256L/S338V) ALS, and OKS also generate [40]. These enzymestheir alsorespective catalyze products successive by an initial decarboxylative decarboxylation condensations of malonyl-CoA of malonyl-CoA to produce acetyl-CoA, with an acetyl-CoAwhich starteris then molecule utilized to as form the starter 5,7-dihydroxy-2-methylchromone unit. (11) (pentaketide), aloesone, and SEK4 (12)/SEK4b (13) (octaketide) respectively (Scheme2). 2-PS, PCS, PKS3, ALS, and OKS also generate their respective products by an initial decarboxylation of malonyl-CoA to produce acetyl-CoA, which is then utilized as the starter unit.

Molecules 2016, 21, 806 7 of 37 Molecules 2016, 21, 806 7 of 36

Scheme 2. Polyketide synthesis catalyzed by R. palmatum ALS, A. arborescens PKS3, A. arborescens PCS, and Scheme 2. Polyketide synthesis catalyzed by R. palmatum ALS, A. arborescens PKS3, A. arborescens PCS, and A. arborescens OKS. (A) ALS or PKS3 catalyzes the condensation of one unit of acetyl-CoA and six units of A. arborescens OKS. (A) ALS or PKS3 catalyzes the condensation of one unit of acetyl-CoA and six units of malonyl-CoA, generating 2-acetonyl-7-hydroxy-5-methylchromone; (B) PCS catalyzes the condensation of malonyl-CoA,one unit of generating acetyl-CoA 2-acetonyl-7-hydroxy-5-methylchromone; and four units of malonyl-CoA, generating 5,7-dihydroxy-2-methylchromone; (B) PCS catalyzes the condensation of one(C unit) OKS of catalyzes acetyl-CoA the condensation and four units of ofone malonyl-CoA, unit of acetyl-CoA generating and seven 5,7-dihydroxy-2-methylchromone; units of malonyl-CoA to form (C) OKSSEK4 catalyzes and SEK4b. the Acetyl-CoA condensation is biosynthesized of one unit of from acetyl-CoA decarboxylation and seven of malonyl-CoA. units of malonyl-CoA to form SEK4 and SEK4b. Acetyl-CoA is biosynthesized from decarboxylation of malonyl-CoA. Based on homology modeling with CHS, it was predicted that the active site architecture of ALS involves a ‘horizontally restricting’ bulky replacement at residue position 256 from glycine to leucine Based on homology modeling with CHS, it was predicted that the active site architecture of ALS and a ‘downward expanding’ substitution at residue position 197 from threonine to alanine when involvescompared a ‘horizontally to CHS [37]. restricting’ Figure 5 shows bulky a schematic replacement of the at effect residue of the position residues 256at positions from glycine 197, 256, to and leucine and a338 ‘downward for R. palmatum expanding’ ALS, G. hybrida substitution 2-PS, A. at arborescens residue positionPCS, A. arborescens 197 from PKS3, threonine A. arborescens to alanine OKS, when comparedand M. to sativa CHS [CHS.37]. FigureThe residue5 shows at position a schematic 256 is of found the effect to line of thethe residuesinitiation/elongation at positions cavity 197, 256,and and 338 foroccupiesR. palmatum an importantALS, siteG. hybridafor the loading2-PS, A. of arborescensthe starter molecule,PCS, A. hence arborescens facilitatingPKS3, theA. use arborescens of a smallerOKS, and M.acetyl-CoA sativa CHS. rather The than residue the bulky at position p-coumaroyl-CoA. 256 is found The toreduced line the steric initiation/elongation bulk due to the threonine cavity to and occupiesalanine an importantsubstitution site at forresidue the loading position of 197 the in starter ALS molecule,unlocks a henceburied facilitating pocket, thus the expanding use of a smaller a acetyl-CoAputative rather cyclization than cavity. the bulky Site-directedp-coumaroyl-CoA. mutagenesis of The ALS reduced revealed steric that changing bulk due any to amino the threonine acid to alanineat these substitution positions into at the residue CHS conserved position residues 197 in ALS(A197T, unlocks L256G, a and buried T338S) pocket, completely thus prevented expanding a the formation of the heptaketide. In particular, the ALS L256G mutant was able to accept p-coumaroyl putative cyclization cavity. Site-directed mutagenesis of ALS revealed that changing any amino acid at CoA as a starter unit to generate CTAL, thus suggesting that Gly256 controls starter unit selectivity. theseFurther, positions the into ALS the A197T CHS mutant conserved was only residues able to (A197T,form a pentaketide L256G, and 2,7-dihydroxy-5-methylchromone T338S) completely prevented the formationafter a of C4 the to C9 heptaketide. aldol condensation, In particular, which theis a re ALSgioiosmer L256G of mutant the 5,7-dihydroxy-2-methylchromone was able to accept p-coumaroyl (11) CoA as a startersynthesized unit toby generate PCS (C6 CTAL,to C1 Claisen thus suggesting condensation) that (Figure Gly256 6).controls Therefore, starter Thr197 unit is likely selectivity. to control Further, the ALSpolyketide A197T chain mutant length was as only it is ablepositioned to form at the a pentaketide entrance of the 2,7-dihydroxy-5-methylchromone buried pocket. Thr338 is thought to after a C4 toguide C9 aldolthe linear condensation, polyketide intermediate, which is a regioiosmer which is bound of theto the 5,7-dihydroxy-2-methylchromone nearby catalytic Cys164, to extend (11) synthesizedinto the byburied PCS pocket, (C6 to hence C1 Claisen enablin condensation)g the formation (Figureof the heptaketide.6). Therefore, Thr197 is likely to control polyketideInterestingly, chain length further as it mutation is positioned of Ala197 at the in ALS entrance to less bulky of the Gly197 buried in pocket. OKS (numbering Thr338 isin thoughtalfalfa to guideCHS) the linear resulted polyketide in the synthesis intermediate, of SEK4 ( which12) and is SEK4b bound (13 to, Scheme the nearby 2), the catalytic longest polyketides Cys164, to extendknown into to be produced by a wild-type type III PKS [37]. Mutagenesis of Met197 in PCS to Gly197 also led to the buried pocket, hence enabling the formation of the heptaketide. the synthesis of octaketides. Moreover, the cyclization chemistry was altered from a C6 to C1 Claisen Interestingly,condensation of further a pentaketide mutation intermediate, of Ala197 in to ALS a C12 to to less C7 bulkyand C10 Gly197 to C15 in aldol OKS condensation (numbering of in an alfalfa CHS)octaketide resulted inintermediate the synthesis to form of SEK4 SEK4 (12 and) and SEK4b, SEK4b respectively (13, Scheme [38].2 ),Hence, the longest Met197 polyketides (and Thr197 known in to bealfalfa produced CHS) by could a wild-type be important type for III facilitating PKS [37]. Claisen Mutagenesis condensation of Met197 reactions in PCS as well to Gly197 [24]. also led to the synthesisAll in of all, octaketides. these findings Moreover, establish the that cyclization altering the chemistry size and wasshape altered of the fromactive asite C6 cavity to C1 via Claisen condensationsite-directed of amutagenesis pentaketide contributes intermediate, to specificity to a C12 for to starter C7 and molecule C10 to choice, C15 aldol chain condensationelongation, and of an octaketidecyclization intermediate chemistry, to lead forming to SEK4 functional and SEK4b, diversity respectively among type [III38 ].PKSs. Hence, This Met197suggests (andpotentially Thr197 in alfalfadifferent CHS) could strategies be important for the engineered for facilitating biosynthesis Claisen of novel condensation pharmaceutically reactions relevant as well polyketides. [24]. All in all, these ﬁndings establish that altering the size and shape of the active site cavity via site-directed mutagenesis contributes to speciﬁcity for starter molecule choice, chain elongation, and cyclization chemistry, leading to functional diversity among type III PKSs. This suggests potentially different strategies for the engineered biosynthesis of novel pharmaceutically relevant polyketides. Molecules 2016, 21, 806 8 of 37

Molecules 2016, 21, 806 8 of 36

Figure 5. Schematic representation of the active sites of (A) R. Palmatum ALS; (B) G. hybrida 2-PS; Figure 5. Schematic representation of the active sites of (A) R. Palmatum ALS; (B) G. hybrida 2-PS; (C) A. arborescens PCS; (D) A. arborescens PKS3; (E) A. arborescens OKS; and (F) M. sativa CHS. The (C) A.Figureresidue arborescens 5. at Schematic positionPCS; (256D representation) controlsA. arborescens the starterofPKS3; the unitactive (E select) A. sites arborescensivity. of (BulkyA) R.OKS; residuesPalmatum and such( FALS;) M. as ( sativaleucineB) G. CHS.hybrida prevents The 2-PS; the residue at position(entranceC) A. 256arborescens of controls starter PCS;acyl-CoAs the ( starterD) A. such arborescens unit as selectivity.p-coumaroyl PKS3; ( Bulky ECoA.) A. arborescensThe residues residuesuch OKS; at posi as andtion leucine ( F197) M. controls prevents sativa CHS.the the gate The entrance to of starterresiduean additional acyl-CoAs at position buried such 256 pocket controls as p- thatcoumaroyl the extends starter to unit CoA. the select “floor Theivity.” residueof the Bulky active at residues position site. Small such 197 residues as leucine controls such prevents the as alanine gate the to an additionalentranceor glycine buried of allowsstarter pocket chainacyl-CoAs that extension extends such toas toformp-coumaroyl the octaketide “ﬂoor” CoA. ofs. Residue the The active residue at position site. at posi Small 338tion residueslocated 197 controls at such the the proximity as gate alanine to or glycineanof additionalthe allows catalytic chain buried Cys164 extension pocket is at thethat to “ceiling” formextends octaketides. ofto thethe active“floor Residue ”site of cavitythe active at and position site. this Smallguides 338 residues located the growing such at the polyketideas proximity alanine of or glycine allows chain extension to form octaketides. Residue at position 338 located at the proximity the catalyticchain to Cys164extend into is at the the buried “ceiling” pocket of near the activeresidue site position cavity 197. and The this numbering guides the for growing all enzymes polyketide are of the catalytic Cys164 is at the “ceiling” of the active site cavity and this guides the growing polyketide chainbased to extend in M. intosativa the CHS. buried pocket near residue position 197. The numbering for all enzymes are chain to extend into the buried pocket near residue position 197. The numbering for all enzymes are based in M. sativa CHS. based in M. sativa CHS.

Figure 6. Comparison of the active sites of M. sativa CHS and R. palmatum BAS. (A) Synthesis of

benzalacetone; (B) Schematic representations of the active site of BAS and CHS. The presence of Figure 6. Comparison of the active sites of M. sativa CHS and R. palmatum BAS. (A) Synthesis of FigureLeu-208, 6. Comparison Leu-125, and of Cys-190 the active and sitesthe reorientation of M. sativa of CHSSer-331 and in BASR. palmatum changed theBAS. orientation (A) Synthesis of the of benzalacetone;starter acyl-CoA, (B p-) Schematiccoumaroyl-CoA, representations as well as theof theorientation active site of the of growingBAS and polyketide CHS. The intermediate, presence of benzalacetone; (B) Schematic representations of the active site of BAS and CHS. The presence of Leu-208,leading toLeu-125, the formation and Cys-190 of benzalacetone. and the reorientation of Ser-331 in BAS changed the orientation of the Leu-208,starter Leu-125, acyl-CoA, and p-coumaroyl-CoA, Cys-190 and the as reorientation well as the orientation of Ser-331 of the in growing BAS changed polyketide the orientationintermediate, of the starter2.4. Otherleading acyl-CoA, Members to the p-formation coumaroyl-CoA,of Type IIIof benzalacetone.PKS from as wellPlants as the orientation of the growing polyketide intermediate, leading to the formation of benzalacetone. 2.4.2.4.1. Other p-Coumaroyl Members of Triacetic Type III PKSAcid from Synthase Plants and Stilbenecarboxylate Synthase 2.4. Other Members of Type III PKS from Plants 2.4.1. p-p-CoumaroylCoumaroyl triacetic Triacetic acid Acid synthase Synthase (CTAS) and Stilbenecarboxylate from Hydrangea macrophylla Synthase var. thunbergii was found to catalyze three consecutive decarboxylative condensations of malonyl-CoA with a p-coumaroyl- 2.4.1. p-Coumaroyl Triacetic Acid Synthase and Stilbenecarboxylate Synthase CoAp- starterCoumaroyl without triacetic cyclization, acid synthase to form p(CTAS)-coumaroyl from triaceticHydrangea acid macrophylla (14) (Scheme var. 3)thunbergii [41]. A comparison was found to catalyze three consecutive decarboxylative condensations of malonyl-CoA with a p-coumaroyl- p- Coumaroyl triacetic acid synthase (CTAS) from Hydrangea macrophylla var. thunbergii was found CoA starter without cyclization, to form p-coumaroyl triacetic acid (14) (Scheme 3) [41]. A comparison to catalyze three consecutive decarboxylative condensations of malonyl-CoA with a p-coumaroyl-CoA starter without cyclization, to form p-coumaroyl triacetic acid (14) (Scheme3)[ 41]. A comparison Molecules 2016, 21, 806 9 of 37 between the amino acid sequences of H. macrophylla CHS and CTAS showed 76% identity and only Molecules 2016, 21, 806 9 of 36 12 amino acid residues which were signiﬁcantly different in charge and size. Amongst them, replacing any of thebetween nucleophilic the amino residues acid sequences in the of CHS H. macrophylla consensus CHS with and neutralCTAS showed amino 76% acid identity (D/E18G) and only and 12 amides (E264Q,aminoMolecules D321N, acid 2016 and,residues 21, 806 E324Q) which in were CTAS significantly might have different led in to charge the loss and ofsize. cyclization Amongst them, activity replacing9 neededof 36 for any of the nucleophilic residues in the CHS consensus with neutral amino acid (D/E18G) and amides the CHSbetween reaction. the amino Furthermore, acid sequences the of CHS-conserved H. macrophylla CHS Thr197, and CTAS which showed is 76% postulated identity and to playonly 12 a role in Molecules(E264Q, 2016 D321N,, 21, 806 and E324Q) in CTAS might have led to the loss of cyclization activity needed for9 of the 36 Claisenamino condensation, acid residues is which replaced were bysignificantly an asparagine different residuein chargein and CTAS size. Amongst [24]. All them, these replacing residues are CHS reaction. Furthermore, the CHS-conserved Thr197, which is postulated to play a role in Claisen any of the nucleophilic residues in the CHS consensus with neutral amino acid (D/E18G) and amides promisingbetweencondensation, targets the amino for is replaced site-directed acid sequences by an asparagine mutagenesis of H. macrophylla residue and in CHSCTAS such and [24]. studies CTAS All these showed may residues provide 76% areidentity promising valuable and only targets insights 12 into (E264Q, D321N, and E324Q) in CTAS might have led to the loss of cyclization activity needed for the productaminofor speciﬁcity. site-directed acid residues mutagenesis which were and significantly such studies differentmay provide in charge valuable and insights size. Amongst into product them, specificity. replacing anyCHS of reaction. the nucleophilic Furthermore, residues the inCHS-conserved the CHS consensus Thr197, with which neutral is postulated amino acid to (D/E18G) play a role and in amidesClaisen (E264Q,condensation, D321N, is replacedand E324Q) by an in asparagine CTAS might residue have inled CTAS to the [24]. loss All of thesecyclization residues activity are promising needed fortargets the CHSfor site-directed reaction. Furthermore, mutagenesis the and CHS-conserved such studies may Thr197, provide which valuable is postulated insights tointo play product a role specificity.in Claisen condensation, is replaced by an asparagine residue in CTAS [24]. All these residues are promising targets for site-directed mutagenesis and such studies may provide valuable insights into product specificity.

Scheme 3. Synthesis of p-coumaroyl triacetic acid catalyzed by p-coumaroyl triacetic acid synthase Scheme 3. Synthesis of p-coumaroyl triacetic acid catalyzed by p-coumaroyl triacetic acid synthase (CTAS) from Hydrangea macrophylla var. thunbergii. (CTAS) from Hydrangea macrophylla var. thunbergii. StilbenecarboxylateScheme 3. Synthesis synthaseof p-coumaroyl (STCS) triacetic from another acid catalyzed Hydrangea by variety,p-coumaroyl H. macrophylla triacetic acid L., synthasewas found to (CTAS) from Hydrangea macrophylla var. thunbergii. Stilbenecarboxylateutilize p-hydroxyphenylpropionyl-CoA synthase (STCS) fromas a starter another andHydrangea carry out threevariety, consecutiveH. macrophylla decarboxylativeL., was found condensationsScheme 3. ofSynthesis malonyl-CoA, of p-coumaroyl followed triacetic by a C2 acid to C7 catalyzed aldol condensation by p-coumaroyl with triacetic retention acid of synthase the terminal Stilbenecarboxylate synthase (STCS) from another Hydrangea variety, H. macrophylla L., was found to to utilizecarboxylp-hydroxyphenylpropionyl-CoA(CTAS) group from to Hydrangea form 5-hydroxylunularic macrophylla var. asthunbergii a acid starter (.15 ) and(Scheme carry 4) out[42]. threeAlthough consecutive not much decarboxylativeis known utilize p-hydroxyphenylpropionyl-CoA as a starter and carry out three consecutive decarboxylative condensationsabout the of mechanism malonyl-CoA, of action, followed the identification by a C2 of tothis C7 new aldol class condensationof type III PKS adds with possibilities retention of the condensations of malonyl-CoA, followed by a C2 to C7 aldol condensation with retention of the terminal terminalto carboxyldirectStilbenecarboxylate the synthesis group to of synthase form new polyketides. 5-hydroxylunularic (STCS) from another Hydrangea acid (15 variety,) (Scheme H. macrophylla4)[ 42]. L., Although was found not to much carboxyl group to form 5-hydroxylunularic acid (15) (Scheme 4) [42]. Although not much is known is knownutilize about p-hydroxyphenylpropionyl-CoA the mechanism of action, as thea starter identiﬁcation and carry out of three this newconsecutive class ofdecarboxylative type III PKS adds condensationsabout the mechanism of malonyl-CoA, of action, followed the identification by a C2 to C7of thisaldol new condensation class of type with III retention PKS adds of thepossibilities terminal possibilitiescarboxylto direct to directthegroup synthesis to the form synthesis of 5-hydroxylunularic new polyketides. of new polyketides. acid (15) (Scheme 4) [42]. Although not much is known about the mechanism of action, the identification of this new class of type III PKS adds possibilities to direct the synthesis of new polyketides.

Scheme 4. Synthesis of 5-hydroxylunularic acid catalyzed by STCS from H. macrophylla L.

2.4.2. Bibenzyl Synthase Scheme 4. Synthesis of 5-hydroxylunularic acid catalyzed by STCS from H. macrophylla L. SchemeBibenzyl 4. synthaseSynthesis (BBS) of 5-hydroxylunulariccloned from a Phalaenopsis acid catalyzedorchid utilizes by STCSm-hydroxyphenyl from H. macrophyllapropionyl-CoAL. as a starter unit and catalyzes the decarboxylative condensation of three malonyl-CoA units, followed 2.4.2. BibenzylScheme Synthase 4. Synthesis of 5-hydroxylunularic acid catalyzed by STCS from H. macrophylla L. by an aldol condensation and decarboxylation to give 3,3′,5-trihydroxybibenzyl (16) (Scheme 5) [43]. 2.4.2. BibenzylBibenzyl Synthase synthase (BBS) cloned from a Phalaenopsis orchid utilizes m-hydroxyphenylpropionyl-CoA 2.4.2. Bibenzyl Synthase Bibenzylas a starter synthase unit and (BBS) catalyzes cloned the fromdecarboxylative a Phalaenopsis condensationorchid utilizes of three malonyl-CoAm-hydroxyphenylpropionyl-CoA units, followed by an aldol condensation and decarboxylation to give 3,3′,5-trihydroxybibenzyl (16) (Scheme 5) [43]. as a starter unitBibenzyl and synthase catalyzes (BBS) the cloned decarboxylative from a Phalaenopsis condensation orchid utilizes of m three-hydroxyphenyl malonyl-CoApropionyl-CoA units, followed as a starter unit and catalyzes the decarboxylative condensation of three malonyl-CoA units, followed by an aldol condensation and decarboxylation to give 3,31,5-trihydroxybibenzyl (16) (Scheme5)[43]. by an aldol condensation and decarboxylation to give 3,3′,5-trihydroxybibenzyl (16) (Scheme 5) [43].

Scheme 5. Synthesis of 3,3′,5-trihydroxybibenzyl catalyzed by BBS from a Phalaenopsis orchid.

A comparison of amino acid sequences revealed that BBS preferred m-substituted derivatives of dihydrocinnamoyl-CoAScheme 5. Synthesis to oftheir 3,3′ ,5-trihydroxybibenzylcinnamoyl-CoA counterparts catalyzed by because BBS from Val98 a Phalaenopsis is replaced orchid. by alanine in BBS. This substitution is also seen in other divergent type III PKSs, such as Hordeum vulgare (barley) A comparison of amino acid sequences revealed that BBS preferred m-substituted derivatives of homoeriodictyol/eriodictyolScheme 5. Synthesis of 3,3synthase′,5-trihydroxybibenzyl and Ruta graveolens catalyzed acridone by BBS from synthase, a Phalaenopsis which orchid.are known to dihydrocinnamoyl-CoAScheme 5. Synthesis of to 3,3 their1,5-trihydroxybibenzyl cinnamoyl-CoA counterparts catalyzed because by BBS fromVal98 aisPhalaenopsis replaced byorchid. alanine accept starters containing m-substituted phenyl rings [24]. Furthermore, Gly211 is replaced by threonine in BBS. This substitution is also seen in other divergent type III PKSs, such as Hordeum vulgare (barley) in BBS:A comparison this small-to-large of amino mutation acid sequences is thought revealed to alter that the BBSshape preferred of the coumaroyl-binding m-substituted derivatives pocket, ofas homoeriodictyol/eriodictyol synthase and Ruta graveolens acridone synthase, which are known to A comparisondihydrocinnamoyl-CoAthis residue is ofsituated amino at tothe acid their front sequences cinnamoyl-CoA of the active revealed site counterparts cavity, that just BBS beneath because preferred the Val98 ‘gatekeeper’m is-substituted replaced phenylalanine by alanine derivatives of accept starters containing m-substituted phenyl rings [24]. Furthermore, Gly211 is replaced by threonine inpositioned BBS. This at substitution the end of the is CoA-bindingalso seen in other tunnel. divergent Derivatives type of III dihydrocinnamoyl-CoA PKSs, such as Hordeum possessvulgare (barley)reduced dihydrocinnamoyl-CoAin BBS: this small-to-large to their mutation cinnamoyl-CoA is thought to alter counterparts the shape of because the coumaroyl-binding Val98 is replaced pocket, by asalanine homoeriodictyol/eriodictyolpropanoid moieties and have synthase increased and conformational Ruta graveolens flexibility, acridone giving synthase, them morewhich advantage are known over to in BBS.this This residue substitution is situated isat the also front seen of the in active other site divergent cavity, just typebeneath III the PKSs, ‘gatekeeper’ such asphenylalanineHordeum vulgare accepttheir rigid starters counterparts containing in m binding-substituted to a phenylcontorted rings coumaroyl-binding [24]. Furthermore, pocketGly211 in is BBS.replaced The byflexibility threonine of positioned at the end of the CoA-binding tunnel. Derivatives of dihydrocinnamoyl-CoA possess reduced (barley)inthe homoeriodictyol/eriodictyol BBS: tetraketide this small-to-large intermediate mutation due to saturation is synthase thought of to the andalter propanoid Rutathe shape graveolens moiety of the may coumaroyl-bindingacridone also aid it to synthase, attain pocket, a folded which as are propanoid moieties and have increased conformational flexibility, giving them more advantage over knownthisto accept residue starters is situated containing at the front mof-substituted the active site cavity, phenyl just rings beneath [24]. the Furthermore, ‘gatekeeper’ phenylalanine Gly211 is replaced their rigid counterparts in binding to a contorted coumaroyl-binding pocket in BBS. The flexibility of by threoninepositioned in BBS: at the this end small-to-large of the CoA-binding mutation tunnel. Derivatives is thought of to dihydrocinnamoyl-CoA alter the shape of the possess coumaroyl-binding reduced the tetraketide intermediate due to saturation of the propanoid moiety may also aid it to attain a folded pocket,propanoid as this residue moieties is and situated have increased at the frontconformational of the active flexibility, site giving cavity, them just more beneath advantage the ‘gatekeeper’ over their rigid counterparts in binding to a contorted coumaroyl-binding pocket in BBS. The flexibility of phenylalaninethe tetraketide positioned intermediate at the due end to of saturation the CoA-binding of the propanoid tunnel. moiety Derivatives may also aid of dihydrocinnamoyl-CoAit to attain a folded possess reduced propanoid moieties and have increased conformational ﬂexibility, giving them more advantage over their rigid counterparts in binding to a contorted coumaroyl-binding pocket in BBS. Molecules 2016, 21, 806 10 of 37

The ﬂexibility of the tetraketide intermediate due to saturation of the propanoid moiety may also aid it to attain a folded conformation suitable for cyclization more easily within a slightly contorted active site. More structural studies are needed to provide insights into the substrate selectivity and mode of cyclizationMolecules of 2016 the, 21 intermediates, 806 in BBS. 10 of 36

2.4.3.conformation Benzalacetone suitable Synthase for cyclization more easily within a slightly contorted active site. More structural studies are needed to provide insights into the substrate selectivity and mode of cyclization of the Unlike M. sativa CHS, benzalacetone synthase (BAS) from R. palmatum is unable to generate intermediates in BBS. the truncation products of the CHS reaction, bisnoryangonin and CTAL. Instead, BAS catalyzes the one-step2.4.3. decarboxylative Benzalacetone Synthase condensation of malonyl-CoA to p-coumaroyl-CoA, followed by another decarboxylation to form benzalacetone (p-hydroxyphenylbut-3-ene-2-one) (17) (Figure6A) [ 44]. Unlike M. sativa CHS, benzalacetone synthase (BAS) from R. palmatum is unable to generate the The biosynthesis of benzalacetone in R. palmatum was attributed to the replacement of the ‘gatekeeper’ truncation products of the CHS reaction, bisnoryangonin and CTAL. Instead, BAS catalyzes the Phe215one-step with Leu208 decarboxylative in BAS. Incondensation alfalfa CHS, of Phe215 malonyl-CoA is hypothesized to p-coumaroyl-CoA, to aid in orientingsubstrates followed by another during polyketidedecarboxylation chain extension. to form benzalacetone Thus, a replacement (p-hydroxyphenylbut-3-ene-2-one) with leucine in BAS may (17 prevent) (Figure subsequent6A) [44]. The chain extensionbiosynthesis of the diketideof benzalacetone intermediate. in R. palmatum In addition, was the attributed rhubarb to BAS the (PDBreplacement ID: 3A5R) of the uses ‘gatekeeper’ an alternative coumaroyl-bindingPhe215 with Leu208 pocket in BAS. to orient In alfalfa the CHS, starter-CoA, Phe215 is ashypothesized the entrance to aid to thein orientingsubstrates original coumaroyl-binding during pocketpolyketide of the alfalfa chain CHSextension. is blocked Thus, stericallya replacement in BAS with by leucine Leu125, in BAS Leu208 may and prevent Ser331 subsequent (Figure6 B)chain [ 45 ,46]. A singleextension large-to-small of the diketide mutation intermediate. from leucine In addition, to threonine the rhubarb at BAS residue (PDB position ID: 3A5R) 125 uses of an BAS alternative back to the CHScoumaroyl-binding conserved residue pocket was to found orient to the be starter-CoA, sufficient toas the unlock entrance the to buried the original coumaroyl-binding coumaroyl-binding pocket, pocket of the alfalfa CHS is blocked sterically in BAS by Leu125, Leu208 and Ser331 (Figure 6B) [45,46]. and the L125T BAS mutant was able to generate naringenin, CTAL and bisnoryangonin, in addition to A single large-to-small mutation from leucine to threonine at residue position 125 of BAS back to the benzalacetoneCHS conserved [46]. Unlikeresidue STS,was found the crystal to be sufficient structure to of unlock rhubarb the BAS buried also coumaroyl-binding pointed to a hydrogen pocket, bond networkand betweenthe L125T the BAS Cys-His-Asn mutant was able catalytic to generate triad andnaringenin, a putative CTAL nucleophilic and bisnorya waterngonin, molecule in addition thought to to play abenzalacetone role in thioester [46]. bondUnlike cleavage STS, the crystal and the structure final decarboxylation of rhubarb BAS also step pointed [45]. These to a hydrogen findings bond provided a novelnetwork structural between basis the for Cys-His-Asn the cleavage catalytic of the triad thioester and a putative bond of nucleophilic the enzyme-bound water molecule intermediate thought and the subsequentto play a role decarboxylation in thioester bond cleavage to form and benzalacetone. the final decarboxylation step [45]. These findings provided Thea novel crystal structural structure basis of for PcPKS1 the cleavage from Polygonumof the thioester cuspidatum bond of the(Japanese enzyme-bound knotweed) intermediate was also and recently solved,the andsubsequent studies decarboxylation showed PcPKS1 to form to have benzalacetone. both CHS and BAS activities [47]. Unlike rhubarb BAS and the classicalThe crystal BAS structure PcPKS2, of thePcPKS1 ‘gatekeeper’ from Polygonum Phe215 cuspidatum remained (Japanese in PcPKS1, knotweed) thus suggesting was also recently a different solved, and studies showed PcPKS1 to have both CHS and BAS activities [47]. Unlike rhubarb BAS catalytic mechanism in the bifunctional PcPKS1. More engineering studies are needed to shed light on and the classical BAS PcPKS2, the ‘gatekeeper’ Phe215 remained in PcPKS1, thus suggesting a the functionaldifferent catalytic divergence mechanism of this uniquein the bifunctional PKS. PcPKS1. More engineering studies are needed to shed light on the functional divergence of this unique PKS. 2.4.4. Type III Polyketide Synthases from Curcuma longa The2.4.4. biosynthesis Type III Polyke oftide curcuminoids Synthases from in C.Curcuma longa (turmeric)longa has been proposed to involve two type III PKSs. Diketide-CoAThe biosynthesis synthase of curcuminoids (DCS) catalyzesin C. longa the(turmeric) decarboxylative has been proposed condensation to involve of malonyl-CoAtwo type with feruloyl-CoAIII PKSs. Diketide-CoA (18) to give synthase riseto (DCS) feruloyldiketide-CoA catalyzes the decarboxylative (19) (Scheme condensation6A). DCS of is malonyl-CoA unlike most type III PKSswith as feruloyl-CoA it is able to ( release18) to give the rise CoA to feruloyldiketide-CoA product without cleaving (19) (Scheme the thioester 6A). DCS linkage.is unlike most Subsequently, type curcuminIII PKSs synthase as it is able (CURS) to release catalyzes the CoA the decarboxylativeproduct without cleaving condensation the thioes of feruloyldiketideter linkage. Subsequently, acid (20 ), the hydrolyzedcurcuminβ-keto synthase acid (CURS) of feruloyldiketide-CoA, catalyzes the decarboxyl withative another condensation unit of feruloyl-CoAof feruloyldiketide starter acid to (20 produce), the hydrolyzed β-keto acid of feruloyldiketide-CoA, with another unit of feruloyl-CoA starter to curcumin (21) (Scheme6B) via a rarely observed head-to-head condensation of ketide units. In addition, produce curcumin (21) (Scheme 6B) via a rarely observed head-to-head condensation of ketide units. CURS is also able to catalyze the formation of curcumin from feruloyl-CoA and malonyl-CoA, albeit at In addition, CURS is also able to catalyze the formation of curcumin from feruloyl-CoA and malonyl- a muchCoA, slower albeit rateat a much [48]. slower rate [48].

Scheme 6. Bioynthesis of curcuminoids in C. longa. (A) DCS catalyzes the formation of feruloyldiketide- SchemeCoA, 6. oneBioynthesis of the rare of ketides curcuminoids with a CoA in C. still longa attached.(A) DCS to the catalyzes final product; the formation (B) CURS of catalyzes feruloyldiketide- the CoA,formation one of the of rare curcumin ketides from with ferulo a CoAyl-CoA still attached and feruloyldiketide to the final product; via a rare (B head-to-head) CURS catalyzes condensation. the formation of curcumin from feruloyl-CoA and feruloyldiketide via a rare head-to-head condensation.

Molecules 2016, 21, 806 11 of 37

Molecules 2016, 21, 806 11 of 36 Interestingly, curcuminoid synthase (CUS) from Oryza sativa is able to generate bisdemethoxycurcuminInterestingly, curcuminoid (22) from the decarboxylativesynthase (CUS) condensationfrom Oryza ofsativa two unitsis able of p-coumaroyl-CoAto generate and onebisdemethoxycurcumin unit of malonyl-CoA, (22) via from a p the-coumaroyldiketide decarboxylative condensation acid intermediate of two units (Figure of p-coumaroyl-CoA7A). Although the Cys-His-Asnand one catalyticunit of malonyl-CoA, triad and most via a conservedp-coumaroyldiketide residues areacid present intermediate in the (Figure CUS active7A). Although site, the the crystal structureCys-His-Asn of CUS (PDBcatalytic ID: triad 3ALE) and revealedmost conserved a downward residues expandingare present in active the CUS site active that issite, large the crystal enough to accommodatestructure of two CUS starter (PDB CoAsID: 3ALE) and revealed one extender a downward CoA (Figure expanding7B) [active49]. site that is large enough to accommodate two starter CoAs and one extender CoA (Figure 7B) [49].

Figure 7. Comparison of the active sites of M. sativa CHS and O. sativa CUS. (A) CUS catalyzes the Figure 7. Comparison of the active sites of M. sativa CHS and O. sativa CUS. (A) CUS catalyzes condensation between two units of p-coumaroyl-CoA and one unit of malonyl-CoA to form the condensationbisdemethoxycurcumin; between ( twoB) Schematic units of representationsp-coumaroyl-CoA of the active and sites one of unit CUS of and malonyl-CoA CHS. The presence to form bisdemethoxycurcumin;of glycine at position (274B) Schematicand tyrosine representations at position 207 orients of the the active active sites site of to CUS allow and longer CHS. polyketides The presence of glycineto be atgenerated. position 274 and tyrosine at position 207 orients the active site to allow longer polyketides to be generated. Further analysis highlighted that the CHS-conserved residues, Thr132, Thr197, Gly256, and Phe265 were replaced by Asn142, Tyr207, Met265, and Gly274, respectively in CUS. This offers no surprise as Further analysis highlighted that the CHS-conserved residues, Thr132, Thr197, Gly256, and Thr132 plays a role in aldol condensation [26,31], while Thr197 is thought to direct Claisen condensation Phe265 were replaced by Asn142, Tyr207, Met265, and Gly274, respectively in CUS. This offers no [24], and both Thr197 and Gly256 are able to direct polyketide chain lengths in other type III PKSs [35–40]. surpriseHence as Thr132by replacing plays these a role residues, in aldol the condensationproduct specificit [26y ,of31 CUS], while is likely Thr197 to be isvery thought different to from direct other Claisen condensationtype III PKSs. [24], Conformational and both Thr197 differences and Gly256 due to are these able amino to direct acid changes polyketide resulted chain in lengthsa loss of inthe other type IIIcoumaroyl-binding PKSs [35–40]. Hencepocket in by CUS. replacing As for Phe265, these residues, it is located the near product the CoA-binding specificity tunnel of CUS in alfalfa is likely to beCHS, very therefore different a fromless bulky other F265G type mutation III PKSs. in CUS Conformational is able to allow differences larger substrates due to enter these the amino active acid changessite resulted[49]. In addition, in a loss the of smaller the coumaroyl-binding glycine residue opens pocketa gate to ina downward CUS. As forexpanding Phe265, pocket, it is located which is near the CoA-bindingthought to accommodate tunnel in alfalfa the p-coumaroyldiketide CHS, therefore a acid less intermediate bulky F265G before mutation the second in CUS decarboxylative is able to allow largercondensation substrates towith enter another the activeunit of sitep-coumaroyl-CoA. [49]. In addition, the smaller glycine residue opens a gate The crystal structure of CUS revealed a putative nucleophilic water molecule which could form to a downward expanding pocket, which is thought to accommodate the p-coumaroyldiketide acid a hydrogen bond network consisting of Ser351-Asn142-H2O-Tyr207-Glu202 near the catalytic Cys174 intermediate before the second decarboxylative condensation with another unit of p-coumaroyl-CoA. residue in the active site (Figure 8). This is similar to the Glu192-Thr132-H2O-Ser338 “aldol-switch” Thehydrogen crystal bond structure network of CUS in STS, revealed which a participates putative nucleophilic in the cleavage water of moleculethe thioester which bond could of the form a hydrogencysteine-bound bond network linear consistingtetraketide intermediate. of Ser351-Asn142-H The hydrogen2O-Tyr207-Glu202 bond network in near CUS the is catalyticpostulated Cys174 to residuebe inimportant the active for sitethe termination (Figure8). Thisof the is polyketide similar to chain the Glu192-Thr132-Hextension at the diketide2O-Ser338 stage by “aldol-switch” cleaving hydrogenthe thioester bond networkbond of the in cysteine-bou STS, whichnd participates intermediate in so the as to cleavage produce ofp-coumaroyldiketide the thioester bond acid, of the cysteine-boundwhich serves linear as the tetraketide second exten intermediate.der substrate [49]. The These hydrogen findings bond provide network valuable in CUS insights is postulated into the to be importantmechanism for of the action termination of CUS, the of first the type polyketide III PKS capable chain extension of utilizing at two the starters diketide and stage one extender by cleaving the thioesterCoA to generate bond of polyketides. the cysteine-bound intermediate so as to produce p-coumaroyldiketide acid, which serves as the second extender substrate [49]. These findings provide valuable insights into the mechanism of action of CUS, the first type III PKS capable of utilizing two starters and one extender CoA to generate polyketides. Molecules 2016, 21, 806 12 of 37 Molecules 2016, 21, 806 12 of 36

Molecules 2016, 21, 806 12 of 36

FigureFigure 8. Hydrogen-bond 8. Hydrogen-bond network networkof of CUSCUS (PDB(PDB ID: ID: 3A 3ALE)LE) consisting consisting of ofSer351, Ser351, Asn142, Asn142, Tyr207, Tyr207, Glu202,Glu202, and and a water a water molecule molecule near near the the catalytic catalytic Cys174.Cys174.

2.4.5. FigureBenzophenone 8. Hydrogen-bond Synthase network and Biphenyl of CUS Synthase (PDB ID: 3ALE) consisting of Ser351, Asn142, Tyr207, 2.4.5. Benzophenone Synthase and Biphenyl Synthase AmongGlu202, andthe astarter-CoAs water molecule utilized near the by catalytic plant PKSs, Cys174. benzoyl-CoA (23) is a rare starter unit greatly Amongpreferred the only starter-CoAs by two kinds of utilized type III byPKSs, plant benzop PKSs,henone benzoyl-CoA synthase (BPS) (23 and) is biphenyl a rare starter synthase unit (BIS) greatly 2.4.5. Benzophenone Synthase and Biphenyl Synthase preferred[50]. In only cellby cultures two kinds of Hypericum of type androsaemum III PKSs, benzophenone (St. John’s wort synthase family of(BPS) plants), and BPS biphenyl catalyzes synthase the (BIS)decarboxylative [50].Among In cell culturesthe condensation starter-CoAs of Hypericum ofutilized three androsaemum unitsby plant of malonyl-CoA PKSs,(St. benzoyl-CoA John’s with wort one (23 unit family) is ofa rarebenzoyl-CoA of plants),starter unit BPSto formgreatly catalyzes a the decarboxylativetetraketidepreferred only intermediate, by condensationtwo kinds which of type undergoes of III three PKSs,units anbenzop intramol ofhenone malonyl-CoAecular synthase C6 to (BPS)C1 with Claisen and one biphenyl unitcondensation of synthase benzoyl-CoA to (BIS)give to form2,4,6-trihydroxybenzophenone a[50]. tetraketide In cell cultures intermediate, of Hypericum which(25, androsaemumScheme undergoes 7) [51]. (St. an Intriguingly,John’s intramolecular wort familyH. androsaemum C6 of to plants), C1 Claisen CHS BPS showedcatalyzes condensation 60% the to decarboxylative condensation of three units of malonyl-CoA with one unit of benzoyl-CoA to form a give 2,4,6-trihydroxybenzophenoneidentity to BPS, and was able to utilize (25 benzoyl-CoA, Scheme7) as [ 51a starter]. Intriguingly, with low activityH. androsaemum to form benzophenone.CHS showed However,tetraketide BPS intermediate, was unable which to haveundergoes any detectable an intramol activityecular C6with to C1derivatives Claisen condensationof cinnamoyl-CoA. to give 60% identity to BPS, and was able to utilize benzoyl-CoA as a starter with low activity to form Mutagenesis2,4,6-trihydroxybenzophenone of three CHS residues (25, Schemelining the 7) active[51]. Intriguingly, site to their correspondingH. androsaemum counterparts CHS showed in BPS60% benzophenone.(L263M/F265Y/S338G)identity to BPS, However, and was resulted able BPS to utilize wasin a mutant unablebenzoyl-CoA H to. androsaemum have as a starter any withCHS detectable lowwhich activity preferred activity to form withbenzoyl-CoA benzophenone. derivatives to of cinnamoyl-CoA.pHowever,-coumaroyl-CoA. BPS Mutagenesis was This unable suggests ofto a threehave combination any CHS detectable residues of amino activity acid lining changes thewithactive canderivatives affe sitect the toof coumaroyl-binding theircinnamoyl-CoA. corresponding counterpartspocketMutagenesis of in CHS, BPS of resulting (L263M/F265Y/S338G)three CHS in residues a greater lining substrate resultedthe active preference in site a mutant to for their benzoyl-CoA. Hcorresponding. androsaemum Identical counterpartsCHS replacements which in preferred BPS benzoyl-CoA(L263M/F265Y/S338G)(L263M/F265Y/S338G) to p-coumaroyl-CoA. were resulted also founin This a dmutant in suggests the functionallyH. androsaemum a combination similar CHS Garcinia of which amino mangostana preferred acid changes L.benzoyl-CoA (mangosteen) can affect to the coumaroyl-bindingBPSp-coumaroyl-CoA. [52]. In cell pocketcultures This ofsuggests of CHS,the medicinal resultinga combination herb in aCentaurium of greater amino substrateacid erythraea changes preference, 3-hydroxybenzoyl-CoA can affect for the benzoyl-CoA. coumaroyl-binding (24) isIdentical the replacementspreferredpocket of (L263M/F265Y/S338G) starterCHS, resultingCoA for inBPS, a greater giving were substraterise also to 2,3 found pr′,4,6-tetrahydroxybenzophenoneeference in the for functionally benzoyl-CoA. similar Identical (26Garcinia, Scheme replacements mangostana 7), the L. (mangosteen)precursor(L263M/F265Y/S338G) BPS of xanthones [52]. In cell were [53]. cultures alsoMore foun structural ofd the in medicinalthe studiefunctionallys are herb needed similarCentaurium to Garcinia provide erythraea mangostana insights, 3-hydroxybenzoyl-CoA as L. to (mangosteen) why BPS is unableBPS [52]. to Inutilize cell cultures phenylpropanoid-CoA of the medicinal esters herb Centauriumas starters. erythraea, 3-hydroxybenzoyl-CoA (24) is the (24) is the preferred starter CoA for BPS, giving rise to 2,31,4,6-tetrahydroxybenzophenone (26, Scheme7), preferred starter CoA for BPS, giving rise to 2,3′,4,6-tetrahydroxybenzophenone (26, Scheme 7), the the precursorprecursor of of xanthones xanthones [[53].53]. More structural structural studie studiess are are needed needed to provide to provide insights insights as to as why to why BPS is BPS is unableunable to utilize to utilize phenylpropanoid-CoA phenylpropanoid-CoA esters esters as as starters. starters.

Scheme 7. Benzophenone synthesis catalyzed by BPS from H. androsaemum (X = H) and C. erythraea (X = OH).

SimilarScheme to7. BPS,Benzophenone BIS from synthesisyeast extract-treated catalyzed by cellBPS culturesfrom H. androsaemumof Sorbus aucuparia (X = H) andforms C. aerythraea tetraketide Scheme 7. Benzophenone synthesis catalyzed by BPS from H. androsaemum (X = H) and C. erythraea intermediate(X = OH). from one unit of benzoyl-CoA and three units of malonyl-CoA. However, the intermediate (Xundergoes = OH). a C2 to C7 aldol condensation instead of a Claisen condensation, followed by a decarboxylationSimilar to BPS, to give BIS 3,5-dihydroxybiphenylfrom yeast extract-treated (27 cell, Scheme cultures 8) [54].of Sorbus aucuparia forms a tetraketide Similarintermediate to BPS, from BIS one from unit yeastof benzoyl-CoA extract-treated and three cell units cultures of malonyl-CoA. of Sorbus aucupariaHowever, theforms intermediate a tetraketide intermediateundergoes from a oneC2 to unit C7 of aldol benzoyl-CoA condensation and threeinstead units of ofa malonyl-CoA.Claisen condensation, However, followed the intermediate by a undergoesdecarboxylation a C2 to to C7 give aldol 3,5-dihydroxybiphenyl condensation instead (27, Scheme of a 8) Claisen [54]. condensation, followed by a decarboxylation to give 3,5-dihydroxybiphenyl (27, Scheme8)[54].

Molecules 2016, 21, 806 13 of 37 Molecules 2016, 21, 806 13 of 36

Scheme 8. Biosynthesis of 3,5-dihydroxybiphenyl catalyzed by S. aucuparia BIS. Scheme 8. Biosynthesis of 3,5-dihydroxybiphenyl catalyzed by S. aucuparia BIS. BIS was also not reactive towards derivatives of cinnamoyl-CoA, while S. aucuparia CHS was BISable wasto accept also benzoyl-CoA not reactive to towardsgive the derailment derivatives product, of cinnamoyl-CoA, benzoyldiacetic acid while lactone.S. aucuparia Although CHS wasboth able STS to and accept BIS carry benzoyl-CoA out aldol condensation, to give the Thr132 derailment and Ser338 product, are replaced benzoyldiacetic by alanine and acid glycine lactone. Althoughrespectively both STS in BIS. and Hence, BIS carry the “aldol-switch” out aldol condensation, hydrogen bond Thr132 network and consisting Ser338 are of replacedGlu192-Thr132- by alanine and glycineH2O-Ser338 respectively is absent in inBIS. BIS. Furthermore, Hence, the subtle “aldol-switch” point mutations hydrogen used to disrupt bond this network hydrogen consisting bond of network in a mutant CHS with STS activity (18xCHS), revealed that a T132A mutation led to the Glu192-Thr132-H O-Ser338 is absent in BIS. Furthermore, subtle point mutations used to disrupt this production of2 more chalcones instead of stilbenes [26]; this was surprisingly not observed in BIS and hydrogen bond network in a mutant CHS with STS activity (18xCHS), revealed that a T132A mutation suggested a different mode of catalysis in BIS. led to the production of more chalcones instead of stilbenes [26]; this was surprisingly not observed in BIS and2.4.6. suggested Acridone aSynthase different and mode Quinolone of catalysis Synthase in BIS.

2.4.6. AcridoneIn addition Synthase to the and C-C Quinolone bond forming Synthase reactions, two groups of type III PKSs catalyze the C-N bond formation, resulting in the production of anthranilate-derived alkaloids. Acridone synthase In(ACS) addition from cell to the cultures C-C bondof R. forminggraveolens reactions, catalyze the two decarboxylative groups of type co IIIndensation PKSs catalyze of three the units C-N of bond formation,malonyl-CoA resulting with in one the productionunit of N-methylanthraniloyl-CoA of anthranilate-derived (28), alkaloids. followed by Acridone a Claisen synthase condensation (ACS) to from cell culturesgive N-methylaminobenzophenone, of R. graveolens catalyze theand decarboxylativea second ring closure condensation to give 1, 3-dihydroxy- of three unitsN-methylacridone of malonyl-CoA with(30 one, Scheme unit 9) of [55].N-methylanthraniloyl-CoA Although R. graveolens ACS share (28 ),74% followed amino acid by sequence a Claisen identity condensation with R. graveolens to give CHS, ACS was only able to utilize p-coumaroyl-CoA with low activity (10%–20%) to form naringenin. N-methylaminobenzophenone, and a second ring closure to give 1, 3-dihydroxy-N-methylacridone (30, On the other hand, R. graveolens CHS was unable to accept the bulky N-methylanthraniloyl-CoA as a Scheme9)[ 55]. Although R. graveolens ACS share 74% amino acid sequence identity with R. graveolens starter [56]. CHS, ACSQuinolone was only synthase able to utilize (QNS)p -coumaroyl-CoAfrom Aegle marmelos with was low able activity to catalyze (10%–20%) the decarboxylative to form naringenin. On thecondensation other hand, ofR. one graveolens unit of malonyl-CoACHS was unable with toone accept unit of the N bulky-methylanthraniloyl-CoAN-methylanthraniloyl-CoA to form a as a starterdiketide, [56]. which spontaneously cyclizes by amide formation to give 4-hydroxy-N-methylquinolone Quinolone(29, Scheme 9). synthase QNS was (QNS) also able from to formAegle benzalacetone marmelos wasfrom p able-coumaroyl-CoA to catalyze and the malonyl-CoA. decarboxylative condensationHomology of modeling one unit revealed of malonyl-CoA that the CoA-binding with one tunnel, unit the of Cys-His-AsnN-methylanthraniloyl-CoA catalytic triad, and tomost form a diketide,of the which active site spontaneously residues are maintained cyclizes by in amideA. marmelos formation QNS. However, to give 4-hydroxy- the CHS-conservedN-methylquinolone residues (29, SchemeThr132,9 Ser133,). QNS and was Phe265 also able are toreplaced form benzalacetone by Ser132, Ala133, from andp-coumaroyl-CoA Val265, respectively and [57]. malonyl-CoA. These substitutions are also present in R. graveolens ACS, hence, they are likely to play an important role in Homology modeling revealed that the CoA-binding tunnel, the Cys-His-Asn catalytic triad, and starter unit specificity [58]. The decrease in steric bulk of the residues at these three positions results in most of the active site residues are maintained in A. marmelos QNS. However, the CHS-conserved a larger active site cavity that can accommodate the bulkier N-methylanthraniloyl-CoA. Site-directed residuesmutagenesis Thr132, studies Ser133, showed and Phe265 that a are triple replaced ACS mutant by Ser132, (S132T/A133S/V265F) Ala133, and Val265, was respectivelytransformed [57]. Theseinto substitutions a functional are CHS also with present some in ACSR. graveolens side activity.ACS, Intriguingly, hence, they a aresimilar likely triple to play QNS an mutant important role in(S132T/A133S/V265F) starter unit speciﬁcity had no [58 activity,]. The decreasebut a double in stericQNS mutant bulk of (S132T/A133S) the residues was at these converted three into positions a resultsfunctional in a larger CHS activethat utilized site cavity p-coumaroyl-CoA that can accommodate instead of N-methylanthraniloyl-CoA, the bulkier N-methylanthraniloyl-CoA. and the number Site-directedof elongation mutagenesis steps was increased, studies showedleading to that the pr aoduction triple ACS of naringenin mutant (S132T/A133S/V265F) [57]. This suggests slight was transformeddifferences into in the a functional active site architectures CHS with some of ACS ACS and side QNS. activity. Intriguingly, a similar triple QNS mutant (S132T/A133S/V265F)Recently, the crystal structures had no activity, of ACS but (PDB a double ID: 3WD7) QNS mutantand QNS (S132T/A133S) (PDB ID: 3WD8) was convertedfrom Citrus microcarpa were solved and wide active site entrances were found in both PKSs, which enabled into a functional CHS that utilized p-coumaroyl-CoA instead of N-methylanthraniloyl-CoA, and the the access of the bulky N-methylanthraniloyl-CoA [59]. The V98A substitution, hypothesized to be number of elongation steps was increased, leading to the production of naringenin [57]. This suggests important for m-substituted phenyl substrate selectivity [24], was only found in R. graveolens ACS slight[58] differences but not in inthe the other active ACS siteand QNSs architectures identified of so ACS far [59]. and C. QNS.microcarpa ACS was found to be a highly Recently,promiscuous the enzyme crystal with structures a large active of site ACS cavity, (PDB being ID: able 3WD7) to accept and different QNS starter (PDB CoAs ID: to 3WD8) generate from Citruschalcone, microcarpa benzophenone,were solved and and phloroglucinol wide active scaffold site entrancess in addition were to found the production in both PKSs, of 4-hydroxy- which enabledN- the accessmethylquinolone of the bulky (29),N 1,3-dihydroxy--methylanthraniloyl-CoAN-methylacridone [59 (30].), The and V98AN-methyl-anthraniloyltriacetic substitution, hypothesized acid to be importantlactone (31 for, Schemem-substituted 9). phenyl substrate selectivity [24], was only found in R. graveolens ACS [58] but not in the other ACS and QNSs identiﬁed so far [59]. C. microcarpa ACS was found to be a highly promiscuous enzyme with a large active site cavity, being able to accept different starter CoAs to generate chalcone, benzophenone, and phloroglucinol scaffolds in addition to the production of 4-hydroxy-N-methylquinolone (29), 1,3-dihydroxy-N-methylacridone (30), and N-methyl-anthraniloyltriacetic acid lactone (31, Scheme9). Molecules 2016, 21, 806 14 of 37

Molecules 2016, 21, 806 14 of 36

Scheme 9. Synthesis of anthranilate-derived alkaloids catalyzed by R. graveolens ACS (30 only), Scheme 9. Synthesis of anthranilate-derived alkaloids catalyzed by R. graveolens ACS (30 only), A. marmelos QNS (29 and 30), C. microcarpa QNS (29 only), and C. microcarpa ACS (29, 30, and 31). A. marmelos QNS (29 and 30), C. microcarpa QNS (29 only), and C. microcarpa ACS (29, 30, and 31). SchemeC. microcarpa 9. Synthesis QNS can of onlyanthranilate-der produce theived diketide alkaloids quinolone catalyzed as theby R.sole graveolens product, ACS unlike (30 A. only), marmelos C.QNS. microcarpaA. Compared marmelosQNS QNS tocan (R.29 graveolensand only 30 produce), C. microcarpaACS theand diketideQNS A. marmelos (29 only), quinolone QNS,and C. similarmicrocarpa as the solereplacements ACS product, (29, 30, andwere unlike 31 ).found A. marmelos in QNS.C. Compared microcarpa ACS to R. (T132S/S133A/F265V), graveolens ACS and butA. not marmelos in C. microcarpaQNS, similar QNS (T132M/S133A/F265L). replacements were foundThe in C. microcarpavolumeC. microcarpa ofACS the active (T132S/S133A/F265V), QNS site can cavity only ofproduce C. microcarpa the butdiketide QNS not quinolone(Figure in C. microcarpa 9A) as isthe smaller soleQNS product, compared (T132M/S133A/F265L). unlike to R. A. palmatum marmelos QNS.BAS (Figure Compared 9B), dueto R. to graveolens the presence ACS of and methionine A. marmelos residues QNS, at similar positions replacements 132 and 194, were and found tyrosine in The volume of the active site cavity of C. microcarpa QNS (Figure9A) is smaller compared to R. palmatum C.at microcarpaposition 197. ACS These (T132S/S133A/F265V), residues block the but access not in to C. the microcarpa coumaroyl-binding QNS (T132M/S133A/F265L). pocket [59]. Hence, The BAS (Figure9B), due to the presence of methionine residues at positions 132 and 194, and tyrosine volumeC. microcarpa of the QNS active is site unable cavity to ofutilize C. microcarpa p-coumaroyl-CoA QNS (Figure as 9A)a substrate, is smaller and compared chain elongation to R. palmatum with at positionBASN-methylanthraniloyl-CoA (Figure 197. 9B), These due residuesto the cannotpresence block proceed of the methionine accessbeyond totheresidues thediketide coumaroyl-binding at positionsstage. Intriguingly, 132 and pocket194, a large-to-small and [tyrosine59]. Hence, C. microcarpaattyrosine position toQNS alanine197. is These unablemutation residues to at utilize residue blockp -coumaroyl-CoAposition the access 197 into C.the microcarpa ascoumaroyl-binding a substrate, QNS mutant and pocketwas chain able elongation[59]. to gainHence, the with N-methylanthraniloyl-CoAC.production microcarpa of QNSbenzalacetone is unable cannot and to utilizebisnoryangonin. proceed p-coumaroyl-CoA beyond Site thedirected diketide as a mutagenesis substrate, stage. and Intriguingly,of the chain important elongation a large-to-smallactive with site tyrosineNresidues-methylanthraniloyl-CoA to alanine Ser132, mutation Thr194, and at cannot Thr197 residue proceed in position C. microcarpa beyond 197 ACSintheC. diketide to microcarpa Met, Met, stage. andQNS Intriguingly, Tyr mutant respectively wasa large-to-small able resulted to gain in the productiontyrosinea quinolone-producing of tobenzalacetone alanine mutation enzyme and at residuewhich bisnoryangonin. no position longer produces197 Sitein C. directed microcarpaacridone, mutagenesis N QNS-methylanthraniloyltriacetic mutant ofwas the able important to gain acid the active site residuesproductionlactone, Ser132,and of benzalacetonechalcone. Thr194, These and and Thr197 resultsbisnoryangonin. in providedC. microcarpa Site th edirected firstACS structural to mutagenesis Met, Met, basis of and thefor Tyr important the respectively generation active site resultedof in a quinolone-producingresiduesanthranilate-derived Ser132, Thr194, alkaloids enzyme and Thr197 by which type in C. III no microcarpa PKSs. longer producesACS to Met, acridone, Met, and NTyr-methylanthraniloyltriacetic respectively resulted in a quinolone-producing enzyme which no longer produces acridone, N-methylanthraniloyltriacetic acid acid lactone, and chalcone. These results provided the ﬁrst structural basis for the generation of lactone, and chalcone. These results provided the first structural basis for the generation of anthranilate-derived alkaloids by type III PKSs. anthranilate-derived alkaloids by type III PKSs.

Figure 9. Comparison of the active sites (A) C. microcarpa QNS; (B) R. palmatum BAS; and (C) M. sativa CHS. The presence of Tyr197 and Met194 in QNS greatly reduces the active site compared to BAS and CHS; hence, only a diketide may be formed. Figure 9. Comparison of the active sites (A) C. microcarpa QNS; (B) R. palmatum BAS; and (C) M. sativa Figure 9. Comparison of the active sites (A) C. microcarpa QNS; (B) R. palmatum BAS; and (C) M. sativa 2.4.7.CHS. Phloroglucinol The presence Synthase, of Tyr197 andAlkylresor Met194 cinolin QNS Synthase, greatly reduces and Alkylpyrone the active site Synthase compared to BAS and CHS. The presence of Tyr197 and Met194 in QNS greatly reduces the active site compared to BAS and CHS;Most hence,plant type only IIIa diketide PKSs have may evolved be formed. to use CoA-esters derived from phenylpropanoid metabolism CHS; hence, only a diketide may be formed. as starters, but some functionally divergent PKSs are adapted to utilizing long-chain acyl-CoAs. 2.4.7. Phloroglucinol Synthase, Alkylresorcinol Synthase, and Alkylpyrone Synthase

2.4.7. PhloroglucinolMost plant type Synthase, III PKSs have Alkylresorcinol evolved to use Synthase,CoA-esters andderived Alkylpyrone from phenylpropanoid Synthase metabolism Mostas starters, plant but type some III functionally PKSs have divergent evolved PKSs to use are CoA-estersadapted to utilizing derived long-chain from phenylpropanoid acyl-CoAs. metabolism as starters, but some functionally divergent PKSs are adapted to utilizing long-chain Molecules 2016, 21, 806 15 of 37 acyl-CoAs. Previous studies on Petroselinum hortense (parsley) CHS in the 1980s showed that type III PKSsMolecules can 2016 be, promiscuous21, 806 enzymes [22]. In addition to the natural substrate p-coumaroyl-CoA,15 of 36 parsley CHS was able to utilize several aliphatic CoA esters as starters, such as butyryl-CoA and MoleculesPrevious 2016 studies, 21, 806 on Petroselinum hortense (parsley) CHS in the 1980s showed that type III PKSs 15can of be36 hexanoyl-CoA,promiscuous which enzymes were [22]. converted In addition efficiently to the natural to acylphloroglucinols. substrate p-coumaroyl-CoA, Nevertheless, parsley CHS these was short chainPreviousable fatty to acyl-CoAutilize studies several on esters Petroselinum aliphatic are unlikely CoA hortense esters to (parsley) occur as star atters, CHS the such site in the as of butyryl-CoA1980s flavonoid showed biosynthesis, and that hexanoyl-CoA, type III andPKSs the canwhich purpose be of havingpromiscuouswere converted activity enzymes efficiently with these[22]. to In acylph substrates additionloroglucinols. to remainedthe natural Nevertheless, unknown.substrate thesep-coumaroyl-CoA, Recently, short chain phloroglucinol fatty parsley acyl-CoA CHS esters synthasewas (PhlD)ableare from unlikely to utilizeEctocarpus to several occur siliculosusat aliphatic the site CoAof(brown flavonoid esters alga) as biosynthesis, star wasters, found such and as to thebutyryl-CoA be purpose able to synthesizeof and having hexanoyl-CoA, activity phloroglucinol with which these (32) fromweresubstrates three converted units remained of efficiently malonyl-CoA unknown. to acylph Recently, (Schemeloroglucinols. phloroglucinol10A), Nevertheless, and a positivesynthase these correlation(PhlD) short chainfrom wasfattyEctocarpus identifiedacyl-CoA siliculosus esters between reacclimationare(brown unlikely alga) to to seawater was occur found at andthe to site be the ableof phloroglucinol flavonoid to synthesize biosynthesis, contentphloroglucinol and in cells the (purpose32 [60) from]. The of three phloroglucinolhaving units activity of malonyl-CoA with synthase these was also ablesubstrates(Scheme to catalyze 10A),remained and the unknown.a decarboxylative positive Recently,correlation condensation phloroglucinol was identified of synthase three between units (PhlD) reacclimation of malonyl-CoA from Ectocarpus to seawater with siliculosus one and unit of the phloroglucinol content in cells [60]. The phloroglucinol synthase was also able to catalyze the lauroyl-CoA(brown alga) (n = was 9) or found palmitoyl-CoA to be able to ( nsynthesize= 13) to form phloroglucinol tetraketide (32α)-pyrones from three33 unitsand of acylphloroglucinols malonyl-CoA (Schemedecarboxylative 10A), and condensation a positive of correlation three units was of malonyl-CoA identified between with one reacclimation unit of lauroyl-CoA to seawater (n = 9)and or 34 (Scheme 10B). thepalmitoyl-CoA phloroglucinol (n = content 13) to form in cells tetraketide [60]. The α-pyrones phloroglucinol 33 and acylphloroglucinolssynthase was also able 34 (Scheme to catalyze 10B). the decarboxylative condensation of three units of malonyl-CoA with one unit of lauroyl-CoA (n = 9) or palmitoyl-CoA (n = 13) to form tetraketide α-pyrones 33 and acylphloroglucinols 34 (Scheme 10B).

Scheme 10. Synthesis of phloroglucinols and α-pyrones catalyzed by PhlD from E. siliculosus. (A) PhlD Scheme 10. Synthesis of phloroglucinols and α-pyrones catalyzed by PhlD from E. siliculosus.(A) PhlD utilizes three units of malonyl-CoA to form the simplest phloroglucinol; (B) PhlD also accepts longer utilizes three units of malonyl-CoA to form the simplest phloroglucinol; (B) PhlD also accepts longer Schemeacyl-CoAs 10. suchSynthesis as lauryl-CoA of phloroglucinols (n = 9) or and palmitoyl-CoA α-pyrones catalyzed (n = 13) by to PhlDform fromtheir E. respective siliculosus tetraketide. (A) PhlD acyl-CoAs such as lauryl-CoA (n = 9) or palmitoyl-CoA (n = 13) to form their respective tetraketide utilizesα-pyrone three and units tetraketide of malonyl-CoA phloroglucinol. to form the simplest phloroglucinol; (B) PhlD also accepts longer α -pyroneacyl-CoAs and tetraketidesuch as lauryl-CoA phloroglucinol. (n = 9) or palmitoyl-CoA (n = 13) to form their respective tetraketide αPhlD-pyrone was and co-crystallized tetraketide phloroglucinol. with arachidonic acid and the structure of the enzyme was solved at 2.85 Å PhlDresolution was co-crystallizedusing molecularwith replacem arachidonicent (PDB acid ID: and4B0N). the The structure structure of theshowed enzyme that wasan acyl-binding solved at 2.85 Å tunnel is present instead of a coumaroyl-binding pocket, hence enabling the PKS to accommodate long- resolutionPhlD using was molecular co-crystallized replacement with arachidonic (PDB acid ID: and 4B0N). the structure The structure of the enzyme showed was that solved an acyl-bindingat 2.85 Å chain acyl-CoAs (Figure 10) [60]. The residues in the acyl-binding tunnel are not conserved in M. sativa tunnelresolution is present using instead molecular of a replacem coumaroyl-bindingent (PDB ID: 4B0N). pocket, The hence structure enabling showed the that PKS an toacyl-binding accommodate tunnelCHS; whenis present arachidonic instead of acid a coumaroyl-binding is superimposed onpo cket,CHS, hence clashes enabling between the thePKS residues to accommodate lining the long- site long-chain acyl-CoAs (Figure 10) [60]. The residues in the acyl-binding tunnel are not conserved in chainand substrate acyl-CoAs were (Figure observed 10) [60]. (Figure The residues 10, indi catedin the by acyl-binding purple bonds tunnel of the are residues). not conserved in M. sativa M. sativaCHS;CHS; when when arachidonic arachidonic acid is acid superimposed is superimposed on CHS, on clashes CHS, clashesbetween between the residues the residueslining the lining site the site andand substratesubstrate were observed observed (Figure (Figure 10, 10 indi, indicatedcated by purple by purple bonds bonds of the of residues). the residues).

Figure 10. Surface view of arachidonic acid—(A) on the acyl-binding tunnel in E. siliculosus PhlD (PDB ID: 4B0N) and (B) superimposed on M. sativa CHS (PDB ID: 1CHW). Clashes between the corresponding Figureresidues 10. in Surface CHS that view line of the arachidonic “tunnel” andacid—( arachidoA) on nicthe acidacyl-binding were observed tunnel (i inndicated E. siliculosus by purple PhlD bonds(PDB Figure 10. Surface view of arachidonic acid—(A) on the acyl-binding tunnel in E. siliculosus PhlD ID:on 4B0N)the residues). and (B) superimposed on M. sativa CHS (PDB ID: 1CHW). Clashes between the corresponding B M. sativa (PDB residues ID: 4B0N) in CHS and that ( line) superimposed the “tunnel” and on arachidonic acidCHS were (PDB observed ID: 1CHW). (indicated Clashes by purple between bonds the correspondingon the residues). residues in CHS that line the “tunnel” and arachidonic acid were observed (indicated by purple bonds on the residues). Molecules 2016, 21, 806 16 of 37

AnMolecules extra 2016 pocket, 21, 806 is also present at the dimeric interface of the enzyme, when compared16 of 36 to the structures of CHS and STS. This is due to the replacement of a bulky aromatic amino acid in the cavity An extra pocket is also present at the dimeric interface of the enzyme, when compared to the with Gly190Molecules 2016 in brown, 21, 806 alga phloroglucinol synthase, thus increasing the volume of the active16 site of 36 cavity. structures of CHS and STS. This is due to the replacement of a bulky aromatic amino acid in the cavity A higher resolution crystal structure is needed in order to shed more light on the molecular basis of with AnGly190 extra in pocket brown is algaalso phloropresentglucinol at the dimeric synthase, interface thus increasing of the enzyme, the volume when ofcompared the active to sitethe this functionally divergent PKS. structurescavity. A higher of CHS resolution and STS. This crystal is due structure to the replacementis needed in of order a bulky to shed aromatic more amino light onacid the in molecularthe cavity Alkyl-resorcylicwithbasis Gly190of this functionally in brown acid alga divergent synthase phloro glucinolPKS. 2 (ARAS2) synthase, from thus increasingO. sativa thecatalyzes volume of thethe active synthesis site of bis-5-alkylresorcinolscavity.Alkyl-resorcylic A higher resolution using acid synthasecrystal alkanedioic structure 2 (ARAS2) acid is neededN from-acetylcysteamine inO. order sativa to catalyzesshed more (NAC) the light synthesis dithioester on the molecular of bis-5- (35) and malonyl-CoAbasisalkylresorcinols of this (Scheme functionally using 11 alkanedioic)[ divergent61]. One acid PKS. of N the-acetylcysteamine two NAC groups (NAC) isdithioester ﬁrst converted (35) and malonyl-CoA into a resorcinol group,(Scheme generatingAlkyl-resorcylic 11) [61]. an One intermediate. ofacid the synthase two NAC ARAS2 2 groups (ARAS2) subsequently is first from converted O. catalyzes sativa into catalyzes a the resorcinol condensation the group, synthesis generating of malonyl-CoAof bis-5- an with thealkylresorcinolsintermediate. intermediate ARAS2 using and subsequently alkanedioic convert the acid catalyzes remaining N-acetylcysteamine the condensation NAC group (NAC) of into malonyl-CoA dithioester another resorcinol ( 35with) and the malonyl-CoA intermediate group, yielding bis-5-alkylresorcinol(Schemeand convert 11) [61].the remaining One (36 ,of Scheme the NAC two NAC11 group). Thisgroups into isanother is analogous first convertedresorcinol to the intogroup, two-stepa resorcinol yielding reaction group,bis-5-alkylresorcinol generating by curcuminoid an (36, Scheme 11). This is analogous to the two-step reaction by curcuminoid synthase (CUS), although synthaseintermediate. (CUS), although ARAS2 subsequently the mechanism catalyzes of synthesis the condensation is different. of malonyl-CoA ARAS2 uses with two the differentintermediate starters andthe mechanism convert the of remaining synthesis NACis different. group ARAS2into another uses tworesorcinol different group, starters yielding and the bis-5-alkylresorcinol same extenders for and the same extenders for bis-5-alkylresorcinol biosynthesis, while CUS utilizes the same starters and (bis-5-alkylresorcinol36, Scheme 11). This biosynthesis, is analogous whileto the CUS two-step utilizes reaction the same by curcuminoidstarters and two synthase different (CUS), extenders although for two different extenders for the biosynthesis of bisdemethoxycurcumin. More structural studies are the mechanismbiosynthesis of of synthesis bisdemethoxycurcumin. is different. ARAS2 More uses st twoructural different studies starters are neededand the tosame provide extenders insights for neededbis-5-alkylresorcinolinto to the provide unusual insights substrate biosynthesis, into specificities the unusualwhile of CUS ARAS2. substrate utilizes the speciﬁcities same starters of and ARAS2. two different extenders for the biosynthesis of bisdemethoxycurcumin. More structural studies are needed to provide insights into the unusual substrate specificities of ARAS2.

SchemeScheme 11. 11.Synthesis Synthesis of of bis-5-alkylresorcinol bis-5-alkylresorcinol catalyzed catalyzed by by O.O. sativa sativa ARAS2.ARAS2.

Recently, alkylpyroneScheme 11. synthasesSynthesis of from bis-5-alkylresorcinol Arabidopsis thaliana catalyzed (PKSA by and O. sativa PKSB) ARAS2. were found to be able Recently,to produce alkylpyrone triketide and synthases tetraketide from α-pyronesArabidopsis 37 and thaliana 38 from(PKSA the decarboxylative and PKSB) were condensation found to beof able to producemalonyl-CoARecently, triketide withalkylpyrone and medium-chain tetraketide synthases toα from-pyroneslong-chain, Arabidopsis37 anandd thalianahydroxylated38 from (PKSA the fatty and decarboxylative PKSB)acyl-CoAs were (Scheme found condensation to 12) be [62].able of malonyl-CoAtoThese produce tetraketide with triketide medium-chain α -pyronesand tetraketide are to important long-chain, α-pyrones for 37 and sporopolleninand hydroxylated 38 from the biosynthesis decarboxylative fatty acyl-CoAs during condensation (Schemepollen grain 12 of)[ 62]. Thesemalonyl-CoAdevelopment, tetraketide whilewithα-pyrones medium-chainthe triketide are α important-pyrones to long-chain, are forthought an sporopollenind hydroxylated to be derailment biosynthesisfatty products. acyl-CoAs Although during(Scheme these pollen12) PKSs[62]. grain development,Theseare related tetraketide whileto CHS, the αthey-pyrones triketide do not are αdisplay-pyrones important CHS are activifor thought sporopolleninty. A similar to be anther-specific derailment biosynthesis products. chalconeduring pollensynthase-like Although grain these enzyme from Physcomitrella patens (PpASCL) was also discovered; the enzyme catalyzes the PKSsdevelopment, are related while to CHS, the triketide they do α-pyrones not display are thought CHS to activity. be derailment A similar products. anther-speciﬁc Although these PKSs chalcone arecondensation related to CHS,malonyl-CoA they do notwith display saturated CHS acyl-CoAsactivity. A (C6–C20),similar anther-specific unsaturated chalcone acyl-CoAs synthase-like (C16:1 or synthase-like enzyme from Physcomitrella patens (PpASCL) was also discovered; the enzyme catalyzes enzymeC18:1) or fromhydroxyl Physcomitrella fatty acyl-CoAs patens to (PpASCL)generate alkyl was and also hydroxyalkyl discovered; α -pyronesthe enzyme that arecatalyzes important the the condensation malonyl-CoA with saturated acyl-CoAs (C6–C20), unsaturated acyl-CoAs (C16:1 or condensationfor moss spore malonyl-CoA wall formation with [63]. saturated acyl-CoAs (C6–C20), unsaturated acyl-CoAs (C16:1 or C18:1)C18:1) or hydroxyl or hydroxyl fatty fatty acyl-CoAs acyl-CoAs to to generate generate alkylalkyl and and hydroxyalkyl hydroxyalkyl α-pyronesα-pyrones that that are important are important for mossfor moss spore spore wall wall formation formation [63 [63].].

Scheme 12. Synthesis of triketide and tetraketide α-pyrones catalyzed by A. thaliana PKSA and PKSB and P. patens PpASCL. Scheme 12. Synthesis of triketide and tetraketide α-pyrones catalyzed by A. thaliana PKSA and PKSB Scheme 12. Synthesis of triketide and tetraketide α-pyrones catalyzed by A. thaliana PKSA and PKSB and P. patens PpASCL. and P. patens PpASCL. Molecules 2016, 21, 806 17 of 37

BothMolecules PKSA 2016, 21 and, 806 PpASCL are also able to react with p-coumaroyl-CoA to yield bisnoryangonin.17 of 36 Based on ab initio multiple-threading and molecular dynamics simulation methods [63], a bulkier residue likeBoth Thr197 PKSA inand alfalfa PpASCL CHS are is also replaced able to byreact a smallerwith p-coumaroyl-CoA Gly205 in PKSA to yield and Gly225bisnoryangonin. in PpASCL, henceBased potentially on ab initio widening multiple-threading the entrance ofand the molecular acyl-binding dynamics tunnel. simulation In addition, methods the conserved[63], a bulkier Gln212 residue like Thr197 in alfalfa CHS is replaced by a smaller Gly205 in PKSA and Gly225 in PpASCL, in alfalfa CHS is substituted by Ala220 in PKSA and Ala240 in PpASCL. The inability of alanine to hence potentially widening the entrance of the acyl-binding tunnel. In addition, the conserved Gln212 form multiplein alfalfa CHS hydrogen is substituted bonds with by Ala220 other in active PKSA site and residues Ala240 mayin PpASCL. provide The the inability structural of alanine flexibility to for a largerform active multiple site cavity,hydrogen thus bonds enabling with other these active PKSs site to playresidues a role may in provide lipid metabolism. the structural flexibility for a larger active site cavity, thus enabling these PKSs to play a role in lipid metabolism. 2.5. Bacterial Type III PKSs 2.5. Bacterial Type III PKSs 2.5.1. 1,3,6,8-Tetrahydroxynaphthalene Synthase (THNS) Unlike2.5.1. 1,3,6,8-Tetrahydroxynaphthalene plant type III PKSs, bacterial Synthase type III (THNS) PKSs usually share less than 50% identity with each otherUnlike and withplant planttype III type PKSs, III bacter PKSs.ial Bacterial type III PKSs type usua III PKSslly share were less first than characterized 50% identity fromwith the Gram-positiveeach other Streptomycesand with plant griseus type III. RppAPKSs. Bacterial showed type 30% III identity PKSs we tore plantfirst characterized CHSs, and from was the able to utilizeGram-positive malonyl-CoA Streptomyces as a starter griseus and. RppA catalyze showed the 30% decarboxylative identity to plant CHSs, condensation and was able ofanother to utilize four unitsmalonyl-CoA of malonyl-CoA as a starter to form and acatalyze pentaketide, the decarboxylative which undergoes condensation dual of cyclization another four reactions units of and malonyl-CoA to form a pentaketide, which undergoes dual cyclization reactions and decarboxylation decarboxylation to generate 1,3,6,8-tetrahydroxynaphthalene (THN) (39)[16]. THN can be further to generate 1,3,6,8-tetrahydroxynaphthalene (THN) (39) [16]. THN can be further oxidized non- 40 oxidizedenzymatically non-enzymatically to yield a reddish to yield compound, a reddish flaviolin compound, (40, Scheme flaviolin 13). ( Cys138, Scheme serves 13). as Cys138 the catalytic serves as the catalyticresidue residueof RppA, of since RppA, a C138S since RppA a C138S mutant RppA was mutantfound to was have found no enzyme to have activity. no enzyme activity.

Scheme 13. Synthesis of 1,3,6,8-tetrahydroxynaphthalene and flaviolin by S. griseus RppA. Scheme 13. Synthesis of 1,3,6,8-tetrahydroxynaphthalene and flaviolin by S. griseus RppA. RppA also has alkylpyrone synthase and phloroglucinol synthase activities. It accepts aliphatic RppAacyl-CoAs also (C4–C8) has alkylpyrone as starters synthaseand catalyze and decarboxylative phloroglucinol condensations synthase activities. of malonyl-CoA It accepts to aliphaticgive acyl-CoAsα-pyrones (C4–C8) and phloroglucinols as starters and [64]. catalyze The active decarboxylative site cavity of RppA condensations is bigger than of most malonyl-CoA type III PKSs; to give α-pyronesit can andcatalyze phloroglucinols the formation [of64 a]. hexaketide, The active 4-hydroxy-6-(2 site cavity of′,4 RppA′,6′-trioxotridecyl)-2-pyrone, is bigger than most using type IIIone PKSs; it canunit catalyze of octanoyl-CoA the formation and five of a units hexaketide, of malonyl-Co 4-hydroxy-6-(2A. However1,4 unlike1,61-trioxotridecyl)-2-pyrone, most plant PKSs, RppA is unable using one to accept aromatic acyl-CoAs such as benzoyl-CoA and phenylacetyl-CoA, indicating differences in the unit of octanoyl-CoA and five units of malonyl-CoA. However unlike most plant PKSs, RppA is unable active site cavities of RppA and plant PKSs. to accept aromatic acyl-CoAs such as benzoyl-CoA and phenylacetyl-CoA, indicating differences in A similar THNS was also found in Streptomyces coelicolor A3(2), although the malonyl-CoA the activebinding site for cavities THNS was of RppA less efficient and plant compared PKSs. to that of the S. griseus RppA [65]. The crystal structure Aof similar S. coelicolor THNS A3(2) was THNS also (PDB found ID: in 1U0M)Streptomyces reveals coelicolor a cryptic A3(2),tunnel althoughthat extends the into malonyl-CoA the ‘floor’ of bindingthe for THNStraditional was type less III efficient PKS active compared site cavity to [66]. that This of the tunnelS. griseusenables RppAthe select [65ive]. expansion The crystal of available structure of S. coelicoloractive siteA3(2) volume THNS by (PDB allowing ID: 1U0M)the aliphatic reveals tail aof cryptic starter-derived tunnel that polyketide extends intermediates into the ‘floor’ to be of the traditionalsequestered. type III In PKS addition, active the site distance cavity between [66]. This Cys138 tunnel and enables the cryptic the selective tunnel precludes expansion all ofbut available the activelongest site volume reaction by intermediates allowing the from aliphatic accessing tail the of novel starter-derived tunnel. This polyketideprovides a feasible intermediates structural to be sequestered.explanation In addition,for the ability the of distance RppA and between THNS to Cys138 catalyze and an extra the crypticpolyketide tunnel elongation precludes step, instead all but the of fewer polyketide elongation steps, when primed with a longer starter such as octanoyl-CoA in vitro. longest reaction intermediates from accessing the novel tunnel. This provides a feasible structural In addition to the cryptic tunnel, less bulky residues compared to alfalfa CHS were identified at several explanation for the ability of RppA and THNS to catalyze an extra polyketide elongation step, instead positions in the active site, including T132C, T194C, T197C, and S338A [66]. However, Gly256 in alfalfa of fewerCHS polyketide is replaced elongationby Tyr224 in steps, THNS, when resembling primed the with ‘horizontally a longer restricting’ starter such G256L as octanoyl-CoA substitution seenin vitro. In additionin 2-pyrone to the synthase cryptic tunnel,(2-PS), lesspentaketide bulky residues chromone compared synthase to(P alfalfaCS), aloesone CHS were synthase identified (ALS), at and several positionsoctaketide in the synthase active site,(OKS), including which utilize T132C, acetyl-CoA T194C, T197C,instead of and p-coumaroyl-CoA S338A [66]. However, as starters. Gly256 The in alfalfa‘downward CHS is replaced expanding’ by Tyr224 conformation in THNS, of resembling THNS active the site ‘horizontally due to therestricting’ cryptic tunnel G256L and substitution T197C seen insubstitution, 2-pyrone and synthase the ‘horizontally (2-PS), pentaketide restricting’ chromone G256Y substitution synthase provide (PCS), aloesonesome insights synthase into the (ALS), and octaketideability of THNS synthase to utilize (OKS), a small which starter, utilize and yet acetyl-CoA have the required instead volume of p-coumaroyl-CoA for the extra polyketide as starters. The ‘downward expanding’ conformation of THNS active site due to the cryptic tunnel and T197C substitution, and the ‘horizontally restricting’ G256Y substitution provide some insights into the ability of THNS to utilize a small starter, and yet have the required volume for the extra polyketide Molecules 2016, 21, 806 18 of 37

Molecules 2016, 21, 806 18 of 36 elongation steps and dual cyclization reactions needed for THN generation (Figure 11). Furthermore, the replacementelongation steps of Thr132 and dual in cyclization alfalfa CHS reactions with Cys106 needed in for THNS THN generation leads to an (Figure additional 11). Furthermore, cysteine residue in thethe active replacement site, which of Thr132 may in facilitate alfalfa CHS polyketide with Cys106 elongation in THNS beyond leads to thean additional triketide cysteine stage. We residue will need morein liganded the active enzyme site, which structures may facilitate with polyketide substrate, el polyketideongation beyond intermediates, the triketide or stage. product We will before need more informationmore liganded regarding enzyme the structures substrate with speciﬁcity, substrate, polyketide polyketide chainintermediates, length control,or product and before mode more of dual information regarding the substrate specificity, polyketide chain length control, and mode of dual cyclization of THNS can be ascertained. cyclization of THNS can be ascertained.

Figure 11. Schematic representation of the active sites of (A) S. coelicolor THNS and (B) M. sativa CHS. Figure 11. Schematic representation of the active sites of (A) S. coelicolor THNS and (B) M. sativa CHS. 2.5.2. Phloroglucinol Synthase from Pseudomonas fluorescens 2.5.2. Phloroglucinol Synthase from Pseudomonas fluorescens Phloroglucinol synthase from Pseudomonas fluorescens (PfPhlD) catalyzes the synthesis of Phloroglucinolphloroglucinol, a synthaseprecursor of from the anti-fungalPseudomonas 2,4-diacetylphloroglucinol fluorescens (PfPhlD) (Scheme catalyzes 10A). the It synthesis has the of phloroglucinol,same activity a as precursor E. siliculosus of PhlD the anti-fungaland about 40% 2,4-diacetylphloroglucinol identity with RppA [17,67]. PfPhlD (Scheme was 10 alsoA). found It has the sameto activity display as broadE. siliculosus substratePhlD specificity and aboutand formed 40% identity α-pyrones with when RppA incubated [17,67]. with PfPhlD starters was such also as found to displayC4–C12 broad aliphatic substrate acyl-CoAs specificity and phenylacetyl-CoA. and formed α Similar-pyrones to RppA, when when incubated primed with with starters medium such to as C4–C12long-chain aliphatic acyl-CoAs acyl-CoAs like and hexanoyl-CoA, phenylacetyl-CoA. octanoyl-CoA, Similar and to decanoyl-CoA, RppA, when primedPhlD catalyzed with medium extra to polyketide extension steps to generate up to heptaketide products. long-chain acyl-CoAs like hexanoyl-CoA, octanoyl-CoA, and decanoyl-CoA, PhlD catalyzed extra Homology modeling of PfPhlD using the structure of THNS as a reference highlighted the presence polyketideof a similar extension buried steps tunnel to extending generate into up the to heptaketide ‘floor’ of the active products. site cavity, which can potentially aid Homologythe binding modelingof long chain ofPfPhlD acyl-CoAs using [68]. the The structure different of reactivity THNS as towards a reference phenylacetyl-CoA highlighted the might presence of a similarsuggest buried a larger tunnel and more extending flexible into tunnel the in ‘floor’ PfPhlD of that the is active capable site of cavity, incorporating which can the potentiallybulky starter, aid the bindingcompared of long to chain RppA. acyl-CoAs Saturation [ 68mutagenesis]. The different was done reactivity on important towards residues phenylacetyl-CoA lining the buried might tunnel, suggest a largerby replacing and more the flexible codon tunnelencoding in the PfPhlD residue that to isbe capablemutated ofwith incorporating the sequence the‘NNS’. bulky This starter, substitution compared to RppA.with Saturation‘NNS’ allows mutagenesis the replacement was of done the onresidue important with all residues 20 amino lining acids, thewhile buried reducing tunnel, the bynumber replacing the codonof premature encoding stop the codons residue to one to (TAG) be mutated out of three. with After the sequence deconvolution, ‘NNS’. three This mutations substitution M21I, withH24V, ‘NNS’ allowsand the L59M, replacement were found of theto significantl residue withy decrease all 20 aminothe reactivity acids, of while PfPhlD reducing with lauroyl-CoA, the number but of not premature its physiological activity to synthesize phloroglucinol. This showed that even subtle decreases in the stop codons to one (TAG) out of three. After deconvolution, three mutations M21I, H24V, and L59M, tunnel volume could impair the ability of PhlD to accept long chain acyl-CoAs. This highlighted the wereutility found of to a significantlycombination of decrease rational thedesign reactivity and random of PfPhlD mutagenesis with lauroyl-CoA, in changing substrate but not itsspecificities, physiological activitysuggesting to synthesize novel phloroglucinol.strategies for the Thiscombinatorial showed thatbiosynthesis even subtle of unnatural decreases polyketide in the tunnel libraries. volume could impair the ability of PhlD to accept long chain acyl-CoAs. This highlighted the utility of a combination of rational2.5.3. design 3,5-Dihydroxyphenylacetic and random mutagenesis Acid Synthase in changing substrate specificities, suggesting novel strategies for the combinatorial3,5-Dihydroxyphenylacetic biosynthesis ofacid unnatural synthase polyketide from Amycolatopsis libraries. mediterranei (DpgA) catalyzes the decarboxylative condensation of four units of malonyl-CoA, followed by a likely C8 to C3 aldol 2.5.3.condensation 3,5-Dihydroxyphenylacetic to yield 3,5-dihydroxyphenylacetic Acid Synthase acid (41, Scheme 14). The compound is a precursor 3,5-Dihydroxyphenylaceticof (S)-3,5-dihydroxyphenylglycine, acid an synthase amino acid from utilizedAmycolatopsis in the biosynthesis mediterranei of (DpgA)balhimycin, catalyzes a the decarboxylativevancomycin derivative condensation [18]. of four units of malonyl-CoA, followed by a likely C8 to C3 Intriguingly, site-directed mutagenesis of the conserved catalytic Cys160 to alanine or serine in aldol condensation to yield 3,5-dihydroxyphenylacetic acid (41, Scheme 14). The compound is a DpgA did not completely ablate the ability to form products [69]. When malonyl-diketidyl-CoA was precursor of (S)-3,5-dihydroxyphenylglycine, an amino acid utilized in the biosynthesis of balhimycin, used to probe how polyketide elongation and cyclization of the polyketide intermediate occurs, DpgA a vancomycin derivative [18]. Intriguingly, site-directed mutagenesis of the conserved catalytic Cys160 to alanine or serine in DpgA did not completely ablate the ability to form products [69]. When malonyl-diketidyl-CoA was Molecules 2016, 21, 806 19 of 37

usedMolecules to probe 2016 how, 21, 806 polyketide elongation and cyclization of the polyketide intermediate occurs,19 of 36 DpgA was found to accept malonyl-diketidyl-CoA, both as a starter and an extender, thus complicating analyseswas [found70]. Moreto accept structural malonyl-diketidyl-CoA, studies are needed both toas understanda starter and howan extender, polyketide thus extensioncomplicating occurs Moleculesanalyses 2016 [70]., 21 , More806 structural studies are needed to understand how polyketide extension occurs19 of 36 without the catalytic cysteine, and how the C8 to C3 aldol condensation, which is unique to DpgA, without the catalytic cysteine, and how the C8 to C3 aldol condensation, which is unique to DpgA, is catalyzed.wasis catalyzed. found to accept malonyl-diketidyl-CoA, both as a starter and an extender, thus complicating analyses [70]. More structural studies are needed to understand how polyketide extension occurs without the catalytic cysteine, and how the C8 to C3 aldol condensation, which is unique to DpgA, is catalyzed.

Scheme 14. Synthesis of 3,5-dihydroxyphenylacetic acid catalyzed by A. mediterranei DpgA. Scheme 14. Synthesis of 3,5-dihydroxyphenylacetic acid catalyzed by A. mediterranei DpgA. 2.5.4. Biosynthesis of Alkylpyrones by PKS11 and PKS18 2.5.4. BiosynthesisScheme of 14. Alkylpyrones Synthesis of 3,5-dihydroxyphe by PKS11 andnylacetic PKS18 acid catalyzed by A. mediterranei DpgA. Similar to A. thaliana PKSA and P. patens PpASCL (anther-specific chalcone synthase-like enzyme), PKS11 and PKS18 from Mycobacterium tuberculosis were found to be able to accept medium to long- Similar2.5.4. Biosynthesis to A. thaliana of AlkylpyronesPKSA and P. by patens PKS11PpASCL and PKS18 (anther-speciﬁc chalcone synthase-like enzyme), PKS11chain and acyl-CoAs PKS18 from (C6–C20)Mycobacterium to produce tuberculosis tetraketidewere and/or found triketide to be ableα-pyrones to accept (Scheme medium 12) [71]. to long-chain The Similar to A. thaliana PKSA and P. patens PpASCL (anther-specific chalcone synthase-like enzyme), acyl-CoAscrystal (C6–C20)structure of to PKS18 produce (PDB tetraketide ID: 1TED)was and/or independently triketide determinedα-pyrones at (Schemethe same time 12)[ as71 the]. TheTHNS crystal PKS11structure, and and PKS18 the fromorientations Mycobacterium of the catalytic tuberculosis triad were (Cys175, found His313,to be able Asn346) to accept and medium other functional to long- structure of PKS18 (PDB ID: 1TED)was independently determined at the same time as the THNS chainresidues acyl-CoAs were found (C6–C20) to be tosimilar produce to that tetraketide of M. sativa and/or CHS. triketide Subtle αdifferences-pyrones (Scheme between 12) the [71]. primary The structure, and the orientations of the catalytic triad (Cys175, His313, Asn346) and other functional crystalsequences structure of PKS18 of PKS18 and M. (PDB sativa ID :CHS 1TED)was result independentlyin the protrusion determined of Cys205 at in the PKS18 same (corresponding time as the THNS to residuesstructure,Thr194 were in alfalfaand found the CHS) toorientations be into similar the putative of to the that catalytic coumaroy of M. triad satival-binding (Cys175,CHS. pocket SubtleHis313, [72]. differences Asn346) In contrast, andbetween PKS18 other wasfunctional the found primary sequencesresiduesto have of a were PKS18novel found acyl-binding and toM. be sativa similar tunnel,CHS to whichthat result of can inM. thebindsativa protrusion aliphatic CHS. Subtle acyl-CoAs of Cys205differences as instarter PKS18between units (corresponding the (Figure primary 12). to Thr194sequencesComparison in alfalfa of CHS)withPKS18 alfalfa intoand theM.CHS sativa putative crystal CHS structure coumaroyl-bindingresult in showed the protrusion that the pocket ofentrance Cys205 [72]. of in In the PKS18 contrast, binding (corresponding tunnel PKS18 in was CHS to found to haveThr194was a blocked novel in alfalfa acyl-binding by Thr197,CHS) into whereas the tunnel, putative the which corresponding coumaroy can bindl-binding Asn208 aliphatic pocket residue acyl-CoAs [72]. in PKS18 In contrast, as was starter not PKS18 extending units was(Figure found into 12). Comparisontothe have tunnel. a with novel This alfalfa acyl-bindingis analogous CHS crystal totunnel, PKSA structure which and PpASCL can showed bind, whereby aliphatic that the a acyl-CoAsreplacement entrance as of ofstarter the bulky binding units Thr197 (Figure tunnel in alfalfa 12). in CHS was blockedComparisonCHS with by a Thr197,smallerwith alfalfa Gly205 whereas CHS in crystal PKSA the corresponding structureand Gly225 showed in PpASCL Asn208 that the is residue entrancebelieved in ofto PKS18 thewiden binding was the entrancenot tunnel extending in of CHS the into was blocked by Thr197, whereas the corresponding Asn208 residue in PKS18 was not extending into the tunnel.acyl-binding This is tunnel, analogous enabling to PKSAthem to and utilize PpASCL, larger acyl-CoAs whereby aas replacement starters [63]. of bulky Thr197 in alfalfa the tunnel. This is analogous to PKSA and PpASCL, whereby a replacement of bulky Thr197 in alfalfa CHS with a smaller Gly205 in PKSA and Gly225 in PpASCL is believed to widen the entrance of the CHS with a smaller Gly205 in PKSA and Gly225 in PpASCL is believed to widen the entrance of the acyl-bindingacyl-binding tunnel, tunnel, enabling enabling them them to to utilize utilize largerlarger acyl-CoAs as as starters starters [63]. [63 ].

Figure 12. (A) Acyl-binding tunnel in M. tuberculosis PKS18 (PDB ID: 1TED); (B) Myristic acid (from PDB ID: 1TED) superimposed on M. sativa CHS (PDB ID: 1CGK). The residues on the acyl binding tunnel of PKS18 are relatively smaller (A52, S143, C205, A209) and/or more hydrophobic (N208, H221, FigureL266, L348). 12. (A ) Acyl-binding tunnel in M. tuberculosis PKS18 (PDB ID: 1TED); (B) Myristic acid (from Figure 12. (A) Acyl-binding tunnel in M. tuberculosis PKS18 (PDB ID: 1TED); (B) Myristic acid (from PDB ID: 1TED) superimposed on M. sativa CHS (PDB ID: 1CGK). The residues on the acyl binding PDB ID: 1TED) superimposed on M. sativa CHS (PDB ID: 1CGK). The residues on the acyl binding tunnelThe crystal of PKS18 structure are relatively of PKS11 sm wasaller recently(A52, S143, solved C205, (PDB A209) ID: and/ 4JAP),or more and hydrophobic a buried hydrophobic (N208, H221, tunnel tunnelwas alsoL266, of revealed, PKS18L348). are together relatively with smaller a CoA-bi (A52,nding S143, C205,tunnel A209) [73]. Co-crystallizations and/or more hydrophobic and in vitro (N208, enzyme H221, L266,assays L348). revealed that PKS11 can catalyze the decarboxylative condensation of methylmalonyl-CoA with Thepalmitoyl-CoA crystal structure (C16) of orPKS11 stearoyl-CoA was recently (C18), solved followed (PDB ID:by 4JAP),another and decarboxylative a buried hydrophobic condensation tunnel Thewasof malonyl-CoA crystalalso revealed, structure and together lactonization of PKS11 with was a toCoA-bi generate recentlynding methylated solved tunnel (PDB [73]. alkylpyrones. ID: Co-crystallizations 4JAP), and The a ability buried and of hydrophobicin PKS11 vitro enzymeto select tunnel was alsoassaysdifferent revealed, revealed extenders togetherthat at PKS11 different with can a stagescatalyze CoA-binding of thepolyketi decar tunneldeboxylative elongation [73]. condensation Co-crystallizations is intriguing, of methylmalonyl-CoAand andmorein structural vitro enzyme with palmitoyl-CoA (C16) or stearoyl-CoA (C18), followed by another decarboxylative condensation assaysstudies revealed in this that aspect PKS11 will can be catalyzeuseful for the the decarboxylativegeneration of clinically condensation important of unnatural methylmalonyl-CoA polyketides. with of malonyl-CoA and lactonization to generate methylated alkylpyrones. The ability of PKS11 to select palmitoyl-CoAdifferent extenders (C16) or at stearoyl-CoAdifferent stages (C18), of polyketi followedde elongation by another is intriguing, decarboxylative and more condensation structural of malonyl-CoAstudies in this and aspect lactonization will be useful to generate for the generation methylated of clinically alkylpyrones. important The unnatural ability of polyketides. PKS11 to select different extenders at different stages of polyketide elongation is intriguing, and more structural studies in this aspect will be useful for the generation of clinically important unnatural polyketides.

Molecules 2016, 21, 806 20 of 37

Molecules 2016, 21, 806 20 of 36 2.5.5. Pyrone Synthases from S. coelicolor 2.5.5. Pyrone Synthases from S. coelicolor TwoMolecules pyrone 2016, 21, synthases806 from S. coelicolor were recently characterized. Germicidin20 synthaseof 36 (GCS) utilizesTwo pyrone short tosynthases medium-chain from S. coelicolor acyl-CoAs were recently (C4–C8) characterized. and acetoacetyl-CoA Germicidin as synthase starter (GCS) units and utilizes2.5.5. Pyrone short toSynthases medium-chain from S. acyl-CoAs coelicolor (C4–C8) and acetoacetyl-CoA as starter units and malonyl- malonyl-CoA, methylmalonyl-CoA and ethylmalonyl-CoA as extender units to generate various CoA, methylmalonyl-CoA and ethylmalonyl-CoA as extender units to generate various germicidins germicidinsTwo (42 pyrone, Scheme synthases 15)[74 from]. S. coelicolor were recently characterized. Germicidin synthase (GCS) (utilizes42, Scheme short 15) to [74].medium-chain acyl-CoAs (C4–C8) and acetoacetyl-CoA as starter units and malonyl- CoA, methylmalonyl-CoA and ethylmalonyl-CoA as extender units to generate various germicidins (42, Scheme 15) [74].

Scheme 15. Synthesis of various germicidins catalyzed by S. coelicolor GCS. Scheme 15. Synthesis of various germicidins catalyzed by S. coelicolor GCS.

A type III polyketide synthase in S. coelicolor (locus tag: SCO7671) efficiently utilizes benzoyl- Scheme 15. Synthesis of various germicidins catalyzed by S. coelicolor GCS. ACoA type and III the polyketide branched synthaseiso-butyryl-CoA in S. coelicolor as starters(locus to form tag: pyrones. SCO7671) In addition, efficiently SCO7671 utilizes and benzoyl-CoA GCS and thecondensed branchedA type malonyl-CoA III polyketideiso-butyryl-CoA with synthase hexanoyl-ACP as in startersS. coelicolor to toyield form(locus pyrones; pyrones.tag: SCO7671)this is the In first addition,efficiently time type utilizes SCO7671 III PKSs benzoyl- were and GCS condensedfoundCoA and to malonyl-CoA utilizethe branched acyl-ACPs with iso-butyryl-CoA [74,75]. hexanoyl-ACP Both asPKSs starters to are yield capable to pyrones;form of pyrones. accommodating this isIn theaddition, first huge time SCO7671 macromolecular type IIIand PKSs GCS were substrates as acyl carriers, presenting a novel means to manipulate PKSs to expand their repertoire foundcondensed to utilize malonyl-CoA acyl-ACPs [with74,75 hexanoyl-ACP]. Both PKSs to are yield capable pyrones; of accommodatingthis is the first time huge type III macromolecular PKSs were of biosynthesized products. The crystal structure of GCS was recently solved, but more structural substratesfound asto acylutilize carriers, acyl-ACPs presenting [74,75]. Both a novel PKSs means are capable to manipulate of accommodating PKSs to expandhuge macromolecular their repertoire of studies are needed before the molecular basis of the interaction between GCS and ACP can be biosynthesizedsubstrates as products. acyl carriers, The presenting crystal structure a novel of mean GCSs to was manipulate recently solved,PKSs to butexpand more their structural repertoire studies understoodof biosynthesized [76]. products. The crystal structure of GCS was recently solved, but more structural are needed before the molecular basis of the interaction between GCS and ACP can be understood [76]. studies are needed before the molecular basis of the interaction between GCS and ACP can be 2.5.6. Other Bacterial Type III PKSs 2.5.6.understood Other Bacterial [76]. Type III PKSs Recently, a heptaketide pyrone synthase from the soil bacterium Rhizobium etli (RePKS) was Recently,characterized.2.5.6. Other a Bacterial heptaketide Unlike Type the R. III pyrone palmatum PKSs synthase aloesone from synthase the soil(ALS), bacterium RePKS undergoesRhizobium dual etli cyclization(RePKS) was characterized.reactionsRecently, and Unlike generates a heptaketide the aR. heptaketide palmatum pyrone pyronesyaloesonenthase 43 fromfrom synthase onethe soilunit (ALS), bacteriumof acetyl-CoA RePKS Rhizobium undergoesand six unitsetli (RePKS) dual of malonyl- cyclization was reactionsCoAcharacterized. (Scheme and generates 16)Unlike [77]. the aRePKS heptaketideR. palmatum also showed aloesone pyrone broad synthase43 subsfromtrate (ALS), one specificities, unit RePKS of undergoesgenerating acetyl-CoA dualtriketide and cyclization six and/or units of malonyl-CoAtetraketidereactions and (Scheme pyrones generates from16)[ a acyl-CoAs77 heptaketide]. RePKS (C4–C18) pyrone also showed and43 from benzoyl-CoA. broad one unit substrate of Homology acetyl-CoA specificities, modeling and six units generatinghighlighted of malonyl- that triketide the active site cavity of RePKS is slightly larger than the active site cavity of THNS, hence partially and/orCoA tetraketide (Scheme 16) pyrones[77]. RePKS from also acyl-CoAs showed broad (C4–C18) substrate and specificities, benzoyl-CoA. generating Homology triketide and/or modeling explaining why RePKS can carry out more rounds of polyketide extension. More studies will be highlightedtetraketide that pyrones the active from siteacyl-CoAs cavity (C4–C18) of RePKS and is benzoyl-CoA. slightly larger Homology than the modeling active site highlighted cavity of that THNS, needed in order to provide insights into the differences in the mechanisms of action between RePKS, hencethe partially active site explaining cavity of why RePKS RePKS is slightly can carry larger out than more the roundsactive site of polyketidecavity of THNS, extension. hence Morepartially studies THNS,explaining and whyR. palmatum RePKS ALS.can carry out more rounds of polyketide extension. More studies will be will be needed in order to provide insights into the differences in the mechanisms of action between needed in order to provide insights into the differences in the mechanisms of action between RePKS, RePKS, THNS, and R. palmatum ALS. THNS, and R. palmatum ALS.

Scheme 16. Synthesis of heptaketide pyrone from R. etli RePKS.

In Azotobacter vinelandii, two alkylresorcinol synthases are encoded in the ars operon–ArsB and Scheme 16. Synthesis of heptaketide pyrone from R. etli RePKS. ArsC. They share 71%Scheme identity 16. Synthesis with each of other heptaketide but are pyrone enzymatically from R. etlidistinctRePKS. PKSs. Both enzymes can utilizeIn Azotobacter the same vinelandii long-chain, two ac alkylresorcinolyl-CoA as a starter synthases and catalyze are encoded three insuccessive the ars operon–ArsB decarboxylative and In ArsC.Azotobacter They share vinelandii 71% identity, two with alkylresorcinol each other but synthases are enzymatically are encoded distinct in the PKSs.ars operon–ArsBBoth enzymes and ArsC.can They utilize share the 71%same identitylong-chain with acyl-CoA each other as a starter but are and enzymatically catalyze three successive distinct PKSs. decarboxylative Both enzymes can utilize the same long-chain acyl-CoA as a starter and catalyze three successive decarboxylative condensations of malonyl-CoA to yield a tetraketide intermediate. Subtle differences in the active site Molecules 2016, 21, 806 21 of 37 Molecules 2016, 21, 806 21 of 36 residuescondensations of ArsB result of malonyl-CoA in an intramolecular to yield a tetraketide C2 to C7 in aldoltermediate. condensation Subtle differences of the intermediate, in the active followed site by a decarboxylationresiduesMolecules 2016 of ArsB, 21, 806 result to produce in an intramolecular alkylresorcinols C2 to C744 aldol(Scheme condensation 17), which of the are intermediate, vital components followed21 of 36 of the cyst coatby a thatdecarboxylation is generated to during produce bacterial alkylresorcinols dormancy. 44 (Scheme However 17), for which ArsC, are the vital intermediate components undergoesof the a C5-oxygencystcondensations coat tothat C1 is oflactonizationgenerated malonyl-CoA during to to formbacterial yield alkylpyronesa tetraketide dormancy. in However (Schemetermediate. for 17 Subtle)[ArsC,78]. thedifferences ArsB intermediate was in found the undergoes active to be site able to a C5-oxygen to C1 lactonization to form alkylpyrones (Scheme 17) [78]. ArsB was found to be able to acceptresiduesa variety of ArsB of acyl-CoA result in starters an intramolecular (C10–C22) C2 while to C7 ArsC aldolshowed condensation a wider of the substrate intermediate, range followed (C6 to C22). acceptby a decarboxylation a variety of acyl-CoA to produce starters alkylresorcinols (C10–C22) while 44 ArsC (Scheme showed 17), awhich wider are substrate vital components range (C6 to of C22). the cyst coat that is generated during bacterial dormancy. However for ArsC, the intermediate undergoes a C5-oxygen to C1 lactonization to form alkylpyrones (Scheme 17) [78]. ArsB was found to be able to accept a variety of acyl-CoA starters (C10–C22) while ArsC showed a wider substrate range (C6 to C22).

Scheme 17. Synthesis of tetraketide α-pyrones and tetraketide resorcinols catalyzed by A. vinelandii Scheme 17. Synthesis of tetraketide α-pyrones and tetraketide resorcinols catalyzed by A. vinelandii ArsC and ArsB respectively. ArsC and ArsB respectively.

InScheme addition 17. Synthesis to straight of tetraketidechain acyl-CoAs, α-pyrones fatty and acids tetraketide bound resorcinols to the ACP catalyzed domain by ofA. thevinelandii fatty acid Insynthase, additionArsC ArsA,and to ArsB straightwere respectively. found chain to be acyl-CoAs, able to be directly fatty acids transferred bound to to the the catalytic ACP domaincysteines ofof theArsB fatty and acid synthase,ArsC, ArsA, indicating were that found unlike to bemost able type to III be PKSs, directly thes transferrede PKSs can also to theutilize catalytic acyl-ACP cysteines as starters of ArsB [79]. and ArsC,The indicating crystalIn addition structure that to unlike straight of ArsC most chain (PDB type acyl-CoAs, ID: III 3VS8) PKSs, fattyshowed these acids PKSsa hydrophobicbound can to also the acyl-binding utilizeACP domain acyl-ACP tunnel of the as similarfatty starters acid to [79]. The crystalPKS18,synthase, structureexplaining ArsA, were of the ArsC found ability (PDB to ofbe ArsC ID:able 3VS8)to to be utilize directly showed long-chain transferred a hydrophobic acyl-CoAs. to the catalytic acyl-bindingInterestingly, cysteines tunnelmutating of ArsB similar andthe to Trp281ArsC, indicating residue in that ArsB unlike to the most corresponding type III PKSs, Gly thes residuee PKSs in ArsCcan also resulted utilize in acyl-ACP an ArsB W281Gas starters mutant [79]. PKS18, explaining the ability of ArsC to utilize long-chain acyl-CoAs. Interestingly, mutating the thatThe wascrystal unable structure to produce of ArsC any (PDB alkylresorcinol, ID: 3VS8) showed but gained a hydrophobic the synthesis acyl-binding of alkylpyrones. tunnel In similar contrast, to Trp281 residue in ArsB to the corresponding Gly residue in ArsC resulted in an ArsB W281G mutant thePKS18, corresponding explaining G284Wthe ability mutation of ArsC in ArsCto utilize resulted long-chain in a decrease acyl-CoAs. in alkylpyrone Interestingly, production mutating and the a that wasconcomitantTrp281 unable residue toincrease producein ArsB in toalkylresorcinol any the alkylresorcinol,corresponding generation Gly but re [80].sidue gained This in ArsC the highlighted synthesis resulted the in of factan alkylpyrones. ArsB that Trp281W281G inmutant In ArsB contrast, the correspondingandthat Gly284was unable in ArsC G284W to produce play mutationan any important alkylresorcinol, in ArsCrole in resulteddeter butmining gained in a the decrease cyclizationsynthesis in of alkylpyronespecificity alkylpyrones. of the production In PKS. contrast, The and a concomitantcrystalthe corresponding structure increase of G284Wthe in ArsC alkylresorcinol mutation G284W in mutant ArsC generation resultedshowed ina [80 significanta]. decrease This highlighted decreasein alkylpyrone in the the productionactive fact thatsite cavity Trp281and a in ArsBvolumeconcomitant and Gly284 when increase in compared ArsC in play alkylresorcinol to that an importantof the wild-type generation role inArsC determining[80].. Hence, This highlighted the the presence cyclization the of fact an thataromatic specificity Trp281 residue in of ArsB the in PKS. The crystaltheand active Gly284 structure sites in ArsCof ofArsB theplay and ArsC an the important G284WArsC G284W mutantrole inmutant deter showed miningis likely a significantthe to provide cyclization a decrease steric specificity wall in for the of the activethe tetraketide PKS. site The cavity volumeintermediatescrystal when structure compared to beof foldedthe to ArsC that into ofG284W a thesuitable wild-type mutant configur showed ArsC.ation a Hence,for significant aldol thecondensation presencedecrease in ofto the occur. an active aromatic The site presence cavity residue in of the tryptophan residue at position 284 resulted in a modulation of the position of Met293 into the the activevolume sites when of ArsB compared and the to that ArsC of G284Wthe wild-type mutant ArsC is likely. Hence, to the provide presence a steric of an wallaromatic for the residue tetraketide in activethe active site, sites greatly of ArsB decreasing and the the ArsC space G284W for polyketide mutant is likely synthesis to provide (Figure a steric13). wall for the tetraketide intermediates to be folded into a suitable configuration for aldol condensation to occur. The presence intermediates to be folded into a suitable configuration for aldol condensation to occur. The presence of the tryptophan residue at position 284 resulted in a modulation of the position of Met293 into the of the tryptophan residue at position 284 resulted in a modulation of the position of Met293 into the activeactive site, greatlysite, greatly decreasing decreasing the the space space for for polyketide polyketide synthesissynthesis (Figure (Figure 13). 13 ).

Figure 13. Active site cavity of ArsC-WT (blue, PDB ID: 3VS8) and ArsC-G284W mutant (pink, PDB ID: 3VS9). The Cys-His-Asn catalytic triad remained in their positions; the absence of tryptophan in position 284 allowed Met293 to swing out of the cavity, giving the growing polyketide more space.

FigureFigure 13. Active 13. Active site site cavity cavity of ArsC-WTof ArsC-WT (blue, (blue,PDB PDB ID:ID: 3VS8) and and ArsC-G284W ArsC-G284W mutant mutant (pink, (pink, PDB PDB ID: 3VS9). The Cys-His-Asn catalytic triad remained in their positions; the absence of tryptophan in ID: 3VS9). The Cys-His-Asn catalytic triad remained in their positions; the absence of tryptophan in position 284 allowed Met293 to swing out of the cavity, giving the growing polyketide more space. position 284 allowed Met293 to swing out of the cavity, giving the growing polyketide more space.

Molecules 2016, 21, 806 22 of 37

A type III PKS from Rhodospirillum centenum, Rhodospirillum polyketide synthase (RpsA), Molecules 2016, 21, 806 22 of 36 was characterized and found to utilize a variety of acyl-CoAs (C6–C20) as starters, followed by the decarboxylativeA type III PKS condensations from Rhodospirillum of one centenum methylmalonyl-CoA,, Rhodospirillum polyketide one malonyl-CoA, synthase (RpsA), and was another methylmalonyl-CoAMoleculescharacterized 2016, 21 ,and 806 tofound yield to alkyldimethylpyronesutilize a variety of acyl-CoAs45 and/or (C6–C20) alkyldimethylresorcylic as starters, followed by22 acidsofthe 36 46 (Schemedecarboxylative 18)[81]. condensations RpsA is the of one ﬁrst methylmalonyl- bacterial typeCoA, IIIone PKSmalonyl-CoA, capable and of another accepting methylmalonyl- two units of methylmalonyl-CoACoA Ato typeyield IIIalkyldimethylpyrones PKS as from extenders Rhodospirillum efﬁciently. 45 and/or centenum alkyldimethylresorcylic More, Rhodospirillum studies are polyketide neededacids 46 (Scheme tosynthase provide 18) (RpsA), [81]. insights RpsA was into the abilityischaracterized the first of RpsAbacterial and to foundtype selectively III to PKS utilize capable control a variety of the accepting condensationof acyl-CoAs two units (C6–C20) order of methylmalonyl-CoA of as extenders starters, followed at as different extenders by the stages efficiently. More studies are needed to provide insights into the ability of RpsA to selectively control of polyketidedecarboxylative elongation. condensations of one methylmalonyl-CoA, one malonyl-CoA, and another methylmalonyl- theCoA condensation to yield alkyldimethylpyrones order of extenders 45 at and/or different alkyldimethylresorcylic stages of polyketide elongation.acids 46 (Scheme 18) [81]. RpsA is the first bacterial type III PKS capable of accepting two units of methylmalonyl-CoA as extenders efficiently. More studies are needed to provide insights into the ability of RpsA to selectively control the condensation order of extenders at different stages of polyketide elongation.

Scheme 18. Synthesis of alkyl-dimethylpyrones and alkyl-dimethylresorcylic acids by R. centenum RpsA. Scheme 18. Synthesis of alkyl-dimethylpyrones and alkyl-dimethylresorcylic acids by R. centenum RpsA. 2.6. Fungal Type III PKSs 2.6. Fungal Type III PKSs SchemeAlthough 18. Synthesis most fungal of alkyl-dimethylpyrones polyketides such as and lovastat alkyl-dimethylresorcylicin are synthesized acidsby the by iterative R. centenum type RpsA. I PKSs, Althoughfungal type most III PKS fungal genes polyketides were identified such recently as lovastatin in the past are synthesizeddecade. Using by the the alfalfa iterative CHS type amino I PKSs, 2.6. Fungal Type III PKSs fungalacid type sequence III PKS as genesa reference, were four identified putative recently type III PKSs, in the namely past decade. chalcone Using synthase-like the alfalfa A, chalcone CHS amino acid sequencesynthase-likeAlthough as aB, most reference, chalcone fungal syntha fourpolyketides putativese-like such C, type and as lovastat IIIchalcone PKSs,in are namelysynthase-like synthesized chalcone D by (CsyA, the synthase-like iterative CsyB, type CsyC,A, I PKSs, chalconeand fungalCsyD), typewere III found PKS ingenes the werefilamentous identified fungus, recently Aspergillus in the past oryzae decade.. Similar Using CHS-like the alfalfa genes CHS were amino also synthase-like B, chalcone synthase-like C, and chalcone synthase-like D (CsyA, CsyB, CsyC, and acididentified sequence in Fusarium as a reference, graminearum four putative and Neurospora type III crassaPKSs, (onenamely homolog chalcone each), synthase-like Magnaporthe A, grisea chalcone and CsyD), were found in the filamentous fungus, Aspergillus oryzae. Similar CHS-like genes were also synthase-likePodospora anserina B, chalcone (two homologs synthase-like each), PhanerochaeteC, and chalcone chrysosporium synthase-like (three D (CsyA,homologs CsyB, each) CsyC, [19], and identifiedCsyD),several in werespeciesFusarium found of lichen graminearum in the fungi filamentous [82].and Neurospora fungus, Aspergillus crassa (one oryzae homolog. Similar each),CHS-likeMagnaporthe genes were grisea also and PodosporaidentifiedThe anserina first in Fusarium fungal(two homologs typegraminearum III PKS each), andwas PhanerochaeteNeurospora characterized crassa chrysosporiumfrom (one the homolog filamentous(three each), fungus,Magnaporthe homologs N. crassa each) grisea, [and19], and severalshowedPodospora species approximately anserina of lichen (two fungi 26%homologs [identity82]. each), to other Phanerochaete plant and chrysosporium bacterial PKSs (three [83]. 2homologs′-oxoalkylresorcylic each) [19], acidand Thesynthaseseveral first species (ORAS) fungal of fromlichen type N. IIIfungi crassa PKS [82]. was was found characterized to catalyze the from decarboxylative the filamentous condensation fungus, ofN. four crassa units, and showedof malonyl-CoA approximatelyThe first fungal with 26% typea unit identity III of PKS long-chain towas other charac acyl-CoA plantterized and (C18) from bacterial as the the filamentouspreferred PKSs [83 ].starter, 2fungus,1-oxoalkylresorcylic followed N. crassa by, aand C2 acid synthasetoshowed C7 (ORAS) aldol approximately condensation from N.crassa 26% to generateidentitywas found tostearoyl other to catalyzeplant pentaketide andthe bacterial resorcylic decarboxylative PKSs acid [83]. (47 2′)-oxoalkylresorcyliccondensation (Scheme 19) [20]. of four acid units of malonyl-CoAsynthase (ORAS) with from a unit N. ofcrassa long-chain was found acyl-CoA to catalyze (C18) the decarboxylative as the preferred condensation starter, followed of four byunits a C2 to of malonyl-CoA with a unit of long-chain acyl-CoA (C18) as the preferred starter, followed by a C2 C7 aldol condensation to generate stearoyl pentaketide resorcylic acid (47) (Scheme 19)[20]. to C7 aldol condensation to generate stearoyl pentaketide resorcylic acid (47) (Scheme 19) [20].

Scheme 19. Synthesis of stearoyl pentaketide resorcylic acid catalyzed by N. crassa ORAS.

Scheme 19. Synthesis of stearoyl pentaketide resorcylic acid catalyzed by N. crassa ORAS. Scheme 19. Synthesis of stearoyl pentaketide resorcylic acid catalyzed by N. crassa ORAS.

Molecules 2016, 21, 806 23 of 37

ORAS was also capable of accepting a range of acyl-CoAs (C2–C24) as starters to produce triketide Molecules 2016, 21, 806 23 of 36 alkylpyrones (from C2 to C20 acyl-CoA substrates), tetraketide alkylpyrones (from C8–C16 acyl-CoA substrates),ORAS tetraketide was also alkylresorcinols capable of accepting (from a range C10–C24 of acyl-CoAs acyl-CoA (C2–C24) substrates), as starters tetraketide to produce alkylresorcylic triketide acidsalkylpyrones (from C20 acyl-CoA (from C2 substrate),to C20 acyl-CoA pentaketide substrates), alkylresorcinols tetraketide alkylpyrones (from C16–C20 (from C8–C16 acyl-CoA acyl-CoA substrates), and pentaketidesubstrates), tetraketide alkylresorcylic alkylresorcinols acids (from (from C18–C20 C10–C24 acyl-CoA substrates), substrates). tetraketide Similar alkylresorcylic to most bacterial type IIIacids PKSs, (from ORAS C20 acyl-CoA contains substrate), a second pentaketide active site alkylresorcinols thiol (Cys120) (from at aC16–C20 position acyl-CoA corresponding substrates), to the THNSand Cys106, pentaketide which alkylresorcylic may help in steeringacids (from the C18–C20 polyketide acyl-CoA reactivity substrates). [84]. The Similar crystal to structuremost bacterial of ORAS type III PKSs, ORAS contains a second active site thiol (Cys120) at a position corresponding to the THNS (PDB ID: 3EUT) also bears structural resemblance to PKS18 and E. siluculosus phloroglucinol synthase Cys106, which may help in steering the polyketide reactivity [84]. The crystal structure of ORAS (PDB (PhlD),ID: and3EUT) has also a deepbears structur hydrophobical resemblance acyl-binding to PKS18 tunnel and E. forsiluculosus accommodating phloroglucinol long-chain synthase (PhlD), acyl-CoAs, alongand with has the a deep conserved hydrophobic catalytic acyl-binding triad (Cys152, tunnel for His305, accommodating and Asn338) long-chain and acyl-CoAs, the CoA-binding along with tunnel (Figurethe 14 conserved)[83]. However, catalytic triad several (Cys152, critical His305, changes and inAsn338) the active and the site CoA-binding residues result tunnel in (Figure a different 14) [83]. product chemistryHowever, for ORAS,several critical and more changes studies in the are active needed site residues in order result to betterin a different understand product the chemistry mechanism for of actionORAS, of ORAS. and more studies are needed in order to better understand the mechanism of action of ORAS.

Figure 14. Acyl-binding tunnel of the following type III PKSs: M. tuberculosis PKS18 (blue, PDB ID: Figure 14. Acyl-binding tunnel of the following type III PKSs: M. tuberculosis PKS18 (blue, PDB ID: 1TED), E. siliculosus PhlD (pink, PDB ID: 4B0N), and N. crassa ORAS (green, PDB ID: 3EUT). The table 1TED), E. siliculosus PhlD (pink, PDB ID: 4B0N), and N. crassa ORAS (green, PDB ID: 3EUT). The table shows the residues lining the acyl-binding tunnel of the respective PKSs. PKS18 and PhlD are more showssimilar the residues compared lining to ORAS; the acyl-binding it is possible that tunnel these of changes the respective lining the PKSs. acyl-binding PKS18 tunnel and PhlD lead areto a more similardifferent compared product to ORAS;chemistry it isin possibleORAS compared that these to PKS18 changes and liningPhlD. the acyl-binding tunnel lead to a different product chemistry in ORAS compared to PKS18 and PhlD. Both A. oryzae CsyA and CsyB showed 38% identity with ORAS [85]. CsyA is likely to be a pyrone BothsynthaseA. oryzae (SchemeCsyA 20), andcapable CsyB of showed utilizing38% C4–C18 identity acyl-CoAs with ORASwith malonyl-CoA [85]. CsyA isas likely the extender to be a pyroneto synthasegenerate (Scheme triketide 20), pyrones. capable Furthermore, of utilizing C4–C18reactions acyl-CoAswith longer with starters malonyl-CoA (C10–C18) resulted as the extender in the to production of tetraketide pyrones. generate triketide pyrones. Furthermore, reactions with longer starters (C10–C18) resulted in the CsyB is the first type III PKS capable of catalyzing the condensation of two diketide CoA esters, productionfollowed of by tetraketide cyclization pyrones. to yield pyrones. CsyB was found to utilize acetoacetyl-CoA (48) both as a CsyBstarter is and the as first an extender type III to PKS form capable dehydroacetic of catalyzing acid (49 the) (Scheme condensation 21A) [86]. of Similarly, two diketide two units CoA of esters, followedβ-ketohexanoyl-CoA by cyclization to(50) yield can be pyrones. coupled together CsyB was by CsyB found to toform utilize 3-butanoyl-4-hydroxy-6-propyl- acetoacetyl-CoA (48) bothα- as a starterpyrone and as(51 an) (Scheme extender 21B). to Furthermore, form dehydroacetic CsyB is able acid to catalyze (49) (Scheme a single 21 decarboxylativeA) [86]. Similarly, condensation two units of β-ketohexanoyl-CoAof malonyl-CoA with (50 )C4–C10 can be acyl-CoAs coupled to together produce by a β CsyB-keto fatty to form acyl 3-butanoyl-4-hydroxy-6-propyl-intermediate, which can then α-pyroneundergo (51) a (Scheme second condensation 21B). Furthermore, with acetoace CsyBtyl-CoA is able to tocatalyze form 3-acetyl-4-hydroxy-6-alkyl- a single decarboxylativeα-pyrones. condensation of malonyl-CoAIf acetoacetyl-CoA with C4–C10is unavailable, acyl-CoAs CsyB is to also produce able to acatalyzeβ-keto the fatty condensation acyl intermediate, of malonyl-CoA which with can then undergo a second condensation with acetoacetyl-CoA to form 3-acetyl-4-hydroxy- 6-alkyl-α-pyrones. Molecules 2016, 21, 806 24 of 37

If acetoacetyl-CoAMolecules 2016, 21, 806 is unavailable, CsyB is also able to catalyze the condensation of malonyl-CoA24 of 36 with acetyl-CoA to generate acetoacetyl-CoA (Scheme 21A). The broad substrate specificities and unique acetyl-CoA to generate acetoacetyl-CoA (Scheme 21A). The broad substrate specificities and unique couplingMolecules reaction 2016, 21 mechanisms, 806 of CsyB make it an ideal target for future structural studies. 24 of 36 coupling reaction mechanisms of CsyB make it an ideal target for future structural studies. acetyl-CoA to generate acetoacetyl-CoA (Scheme 21A). The broad substrate specificities and unique coupling reaction mechanisms of CsyB make it an ideal target for future structural studies. Molecules 2016, 21, 806 24 of 36

acetyl-CoA to generate acetoacetyl-CoA (Scheme 21A). The broad substrate specificities and unique coupling reaction mechanisms of CsyB make it an ideal target for future structural studies.

Scheme 20. Synthesis of triketide and tetraketide pyrones catalyzed by A. oryzae CsyA. Scheme 20. Synthesis of triketide and tetraketide pyrones catalyzed by A. oryzae CsyA. Scheme 20. Synthesis of triketide and tetraketide pyrones catalyzed by A. oryzae CsyA.

Scheme 20. Synthesis of triketide and tetraketide pyrones catalyzed by A. oryzae CsyA.

Scheme 21. Synthesis of (A) dehydroacetic acid and (B) 3-butanoyl-4-hydroxy-6-propyl-α-pyrone catalyzed by A. oryzae CsyB. Scheme 21. Synthesis of (A) dehydroacetic acid and (B) 3-butanoyl-4-hydroxy-6-propyl-α-pyrone Scheme 21. Synthesis of (A) dehydroacetic acid and (B) 3-butanoyl-4-hydroxy-6-propyl-α-pyrone catalyzedAn unusual by typeA. oryzae III PKS CsyB. with extremely broad substrate specificity was cloned from Aspergillus niger. catalyzed by A. oryzae CsyB. Aspergillus niger polyketide synthase (AnPKS) shares 37% identity with ORAS, and homology modeling usingAn the unusual structure type of III ORAS PKS with as a extremelyreference revealbroad ssubstrate a similar specificity acyl-binding was clonedtunnel from[21]. AspergillusIndeed, AnPKS niger. Scheme 21. Synthesis of (A) dehydroacetic acid and (B) 3-butanoyl-4-hydroxy-6-propyl-α-pyrone AnAspergilluswas found unusual niger to utilize typepolyketide a IIIvariety PKSsynthase of withacyl-CoAs (AnPKS) extremely (C2–C18) shares 37% broad as startersidentity substrate andwith malonyl-CoA ORAS, speciﬁcity and homology as wasthe extender modeling cloned to from catalyzed by A. oryzae CsyB. Aspergillususinggenerate the niger triketidestructure. Aspergillus alkylpyrones. of ORAS niger as a AnPKSreferencepolyketide was reveal also synthases aba similarle to (AnPKS)form acyl-binding tetraketide shares tunnel alkylpyrones, 37% [21]. identity Indeed, tetraketide withAnPKS ORAS, wasalkylresorcinols, found to utilize and a varietypentaketide of acyl-CoAs alkylresorcinols (C2–C18) as53 starters from andC10–C18 malonyl-CoA acyl-CoAs as the(Figure extender 15A). to and homologyAn unusual modeling type III usingPKS with the extremely structure broad of ORAS substrate as specificity a reference was revealscloned from a similar Aspergillus acyl-binding niger. generateFurthermore, triketide acetoacetyl-CoA, alkylpyrones. the AnPKS aromatic was acyl-CoAs also able tosuch form as tetraketidebenzoyl-CoA alkylpyrones, and phenylacetyl-CoA tetraketide tunnelAspergillus [21]. Indeed, niger polyketide AnPKS synthase was found (AnPKS) to utilize shares a37% variety identity of with acyl-CoAs ORAS, and (C2–C18) homology as modeling starters and alkylresorcinols,(54), and branched and acyl-CoAs pentaketide such asalkylresorcinols isobutyryl-CoA 53(55 ),from isovaleryl-CoA C10–C18 acyl-CoAs(56), and methylcrotonyl- (Figure 15A). malonyl-CoAusing the structure as the extenderof ORAS as to a generatereference reveal triketides a similar alkylpyrones. acyl-binding AnPKS tunnel [21]. was Indeed, also able AnPKS to form Furthermore,CoA (57), are readilyacetoacetyl-CoA, accepted bythe AnPKS aromatic (Figure acyl-CoAs 15B). Thesuch promiscuity as benzoyl-CoA of AnPKS and phenylacetyl-CoA makes it an ideal was found to utilize a variety of acyl-CoAs (C2–C18) as starters and malonyl-CoA as the extender53 to tetraketide(target54), and for alkylpyrones, branched the engineered acyl-CoAs tetraketide biosynthesis such as alkylresorcinols,isobutyryl-CoAof unnatural polyketides (55), and isovaleryl-CoA pentaketidewith potential (56 alkylresorcinols), drug and applications.methylcrotonyl- from generate triketide alkylpyrones. AnPKS was also able to form tetraketide alkylpyrones, tetraketide C10–C18CoA ( acyl-CoAs57), are readily (Figure accepted 15A). by Furthermore,AnPKS (Figure acetoacetyl-CoA,15B). The promiscuity the of aromatic AnPKS makes acyl-CoAs it an ideal such as alkylresorcinols, and pentaketide alkylresorcinols 53 from C10–C18 acyl-CoAs (Figure 15A). benzoyl-CoAtarget for the and engineered phenylacetyl-CoA biosynthesis ( 54of ),unnatu andral branched polyketides acyl-CoAs with potential such drug as isobutyryl-CoA applications. (55), Furthermore, acetoacetyl-CoA, the aromatic acyl-CoAs such as benzoyl-CoA and phenylacetyl-CoA isovaleryl-CoA (56), and methylcrotonyl-CoA (57), are readily accepted by AnPKS (Figure 15B). (54), and branched acyl-CoAs such as isobutyryl-CoA (55), isovaleryl-CoA (56), and methylcrotonyl- The promiscuityCoA (57), are readily of AnPKS accepted makes by AnPKS it an ideal (Figure target 15B). for The the promiscuity engineered of AnPKS biosynthesis makes it of anunnatural ideal polyketidestarget for with the engineered potential drug biosynthesis applications. of unnatural polyketides with potential drug applications.

Figure 15. Substrate and product profiles of A. niger AnPKS and B. cinerea BPKS. (A) Products formed, and (B) starter acyl-CoAs accepted by the enzyme. Figure 15. Substrate and product profiles of A. niger AnPKS and B. cinerea BPKS. (A) Products formed, and (B) starter acyl-CoAs accepted by the enzyme.

Figure 15. Substrate and product profiles of A. niger AnPKS and B. cinerea BPKS. (A) Products formed, Figure 15. Substrate and product proﬁles of A. niger AnPKS and B. cinerea BPKS. (A) Products formed, and (B) starter acyl-CoAs accepted by the enzyme. and (B) starter acyl-CoAs accepted by the enzyme.

Molecules 2016, 21, 806 25 of 37

The Botrytis cinerea type III PKS (BPKS) was also recently found to accept short-chain to long-chain acyl-CoAs (C2–C18) as starters with malonyl-CoA as the extender to generate triketide alkylpyrones (Figure 15A) [87]. Tetraketide alkylpyrones were also formed from C10–C18 acyl-CoAs, while pentaketideMolecules 2016 alkylresorcylic, 21, 806 acids 52 and pentaketide alkylresorcinols 53 were also produced25 of 36 from C16–C18 acyl-CoAs. Moreover, BPKS is able to generate a hexaketide alkylresorcylic acid from C18 The Botrytis cinerea type III PKS (BPKS) was also recently found to accept short-chain to long-chain acyl-CoA. BPKS is also able to utilize acetoacetyl-CoA and the aromatic benzoyl-CoA as starters acyl-CoAs (C2–C18) as starters with malonyl-CoA as the extender to generate triketide alkylpyrones to form triketide pyrones (Figure 15B). BPKS shares 65% identity with ORAS and similar to other (Figure 15A) [87]. Tetraketide alkylpyrones were also formed from C10–C18 acyl-CoAs, while pentaketide type IIIalkylresorcylic PKSs which acids utilize 52 and long-chain pentaketide acyl-CoAs alkylresorcinols as substrates, 53 were also homology produced from modeling C16–C18 studies acyl-CoAs. show an acyl-bindingMoreover, tunnel BPKS is in able the to active generate site. a Recently,hexaketide pyrones alkylresorcylic have beenacid from found C18 to acyl-CoA. possess BPKS anti-cancer is also [88] and antibacterialable to utilize [ 89acetoacetyl-CoA] properties, making and the themaromatic potential benzoyl-CoA lead structures as starters for to bioactiveform triketide agents. pyrones However, more(Figure studies 15B). are neededBPKS shares to understand 65% identity how with differential ORAS and chainsimilar extension to other type and III thus PKSs product which utilize profiles of BPKSlong-chain and other acyl-CoAs fungal type as substrates, III PKSs homology can be regulated modeling studies based onshow the an length acyl-binding of the tunnel starter in the units. active site. Recently, pyrones have been found to possess anti-cancer [88] and antibacterial [89] properties, 3. Precursor-Directedmaking them potential Biosynthesis lead structures of Polyketides for bioactive agents. However, more studies are needed to understand how differential chain extension and thus product profiles of BPKS and other fungal type αβαβα AlthoughIII PKSs cantype be regulated III PKSs based share on the a length similar of three-dimensionalthe starter units. fold and a conserved Cys-His-Asn catalytic triad, subtle differences in the active site cavities result in diverse substrate specificities3. Precursor-Directed and different Biosynthesis product profiles. of Polyketides Notably, manipulation of the biosynthetic reaction by the incorporationAlthough type of unnatural III PKSs share substrates a similar has three-dimensional led to the generation αβαβα fold of and novel a conserved polyketide Cys-His- libraries. For instanceAsn catalytic when triad, both subtle the differences starters andin the the active extenders site cavities were result substituted in diverse withsubstrate non-physiological specificities substrates,and different CHS fromproduct the profiles. herb Scutellaria Notably, manipulation baicalensis was of the able biosynthetic to catalyze reaction the formationby the incorporation of unnatural polyketidesof unnatural [90]. substrates Other has than led the to the preferred generation substrates of novel polyketide of p-coumaroyl-CoA libraries. For instance and when malonyl-CoA, both S. baicalensisthe startersCHS and wasthe extenders able to acceptwere substituted benzoyl-CoA with non-physiological as the starter and substrates, methylmalonyl-CoA CHS from the herb as the extenderScutellaria to produce baicalensis methylated was abletriketide to catalyze and the tetraketide formation of pyrones. unnatural Furthermore, polyketides [90]. theCHS Other also than catalyzed the the decarboxylativepreferred substrates condensation of p-coumaroyl-CoA of methylmalonyl-CoA and malonyl-CoA, with S. hexanoyl-CoAbaicalensis CHS was to yield able ato methylated accept benzoyl-CoA as the starter and methylmalonyl-CoA as the extender to produce methylated triketide triketide pyrone. CHSs from the Thai medicinal plant Cassia alata were also found to be promiscuous and tetraketide pyrones. Furthermore, the CHS also catalyzed the decarboxylative condensation of and accepted butyryl-CoA, isovaleryl-CoA, hexanoyl-CoA, benzoyl-CoA, cinnamoyl-CoA, and methylmalonyl-CoA with hexanoyl-CoA to yield a methylated triketide pyrone. CHSs from the Thai p-coumaroyl-CoAmedicinal plant as Cassia starters alata to were form also triketide found to and be promiscuous tetraketidepyrones, and accepted and butyryl-CoA, phloroglucinol isovaleryl- derivatives in theCoA, presence hexanoyl-CoA, of malonyl-CoA benzoyl-CoA, as the cinnamoyl-CoA, extender [91]. and The p-coumaroyl-CoA ability of CHSs as to starters utilize to CoA form esters triketide that are not derivedand tetraketide from the pyrones, phenylpropanoid and phloroglucinol pathway derivatives highlights in the presence the role of CHSs malonyl-CoA can play as the in synthesizingextender diverse[91]. bioactive The ability polyketide of CHSs to compound utilize CoA libraries.esters that are not derived from the phenylpropanoid pathway Similarhighlights to the studies role CHSs on CHSs, can play STS in synthesizing from Arachis diverse hypogaea bioactive(peanut) polyketide was alsocompound found libraries. to accept several unnaturalSimilar substrates to studies to produce on CHSs, unique STS from polyketides Arachis hypogaea [92]. In (peanut) particular, was incubationalso found to of accept peanut several STS with p-fluorocinnamoyl-CoAunnatural substrates to (58 produce), trans unique-3-furanacryloyl-CoA polyketides [92]. In (59 particular,), or trans incubation-3-(3-thienyl)acryloyl-CoA of peanut STS with (60, p-fluorocinnamoyl-CoA (58), trans-3-furanacryloyl-CoA (59), or trans-3-(3-thienyl)acryloyl-CoA (60, Figure 16) in the presence of malonyl-CoA yielded stilbene-like products. Figure 16) in the presence of malonyl-CoA yielded stilbene-like products.

Figure 16. Starter acyl-CoA substrates utilized by A. hypogaea STS. Figure 16. Starter acyl-CoA substrates utilized by A. hypogaea STS. When the peanut STS was incubated with other p-substituted cinnamoyl-CoA analogues (-Cl, When-Br, and the –OCH peanut3), only STS triketide was incubated and tetraketide with pyrones other p were-substituted obtained. cinnamoyl-CoA This indicated potential analogues steric (-Cl, -Br, andeffects –OCH by substituents3), only triketide that are andlarger tetraketide than that of pyronesthe wild-type were substrate obtained. (-OH), This which indicated affected potential the stericstability effects byand/or substituents the conformation that are of larger polyketide than thatintermediates of the wild-type in the cyclization substrate pocket (-OH), of which the active affected the stabilitysite. Interestingly, and/or the when conformation p-coumaroyl-CoA, of polyketide cinnamoyl-CoA, intermediates or phenylpropionyl-CoA in the cyclization were pocket utilized of the as starters, Scots pine STS was able to form pentaketide stilbenes in addition to tetraketide and active site. Interestingly, when p-coumaroyl-CoA, cinnamoyl-CoA, or phenylpropionyl-CoA were triketide products, suggesting a considerable amount of flexibility in chain elongation in the active utilized as starters, Scots pine STS was able to form pentaketide stilbenes in addition to tetraketide and site cavity [93]. triketide products,Octaketide suggestingsynthase (OKS) a considerable from A. arborescens amount is known of ﬂexibility to catalyze in seven chain sequential elongation decarboxylative in the active site cavitycondensations [93]. of malonyl-CoA with an acetyl-CoA starter molecule to form the octaketides SEK4 (12) Octaketideand SEK4b (13 synthase) (Scheme (OKS)2). Further from studiesA. on arborescens substrate toleranceis known revealed to catalyze that aloe OKS seven is able sequential to decarboxylative condensations of malonyl-CoA with an acetyl-CoA starter molecule to form the Molecules 2016, 21, 806 26 of 37 octaketides SEK4 (12) and SEK4b (13) (Scheme2). Further studies on substrate tolerance revealed that aloe OKS is able to accept p-coumaroyl-CoA as a starter to generate a novel hexaketide stilbene (61) andMolecules heptaketide 2016, 21, 806 chalcone (63) in the presence of malonyl-CoA as the extender (Scheme26 of 36 22)[94]. Similarly, OKS condenses malonyl-CoA with hexanoyl-CoA, followed by a C6 to C11 or C12 to accept p-coumaroyl-CoA as a starter to generate a novel hexaketide stilbene (61) and heptaketide C7 aldol-type cyclization to yield a hexaketide resorcinol (62) or heptaketide phloroglucinol (64) chalcone (63) in the presence of malonyl-CoA as the extender (Scheme 22) [94]. Similarly, OKS condenses respectivelymalonyl-CoA (Scheme with 22 hexanoyl-CoA,). The aloe OKS followed also by utilizes a C6 to long-chainC11 or C12 to C8–C20 C7 aldol-type acyl-CoAs cyclization as to starters yield a to form Molecules 2016, 21, 806 26 of 36 triketidehexaketide and tetraketide resorcinolα (62-pyrones) or heptaketide [40]. Thisphloroglucinol indicates (64 that) respectively the ‘downward (Scheme 22). expanding’ The aloe OKS active site architecturealsoaccept utilizes of p aloe-coumaroyl-CoA long-chain OKS is large C8–C20 as enougha starteracyl-CoAs toto accommodategenerate as starters a novelto form bulky hexaketide triketide substrates and stilbene tetraketide and (61 intermediates.) and α-pyrones heptaketide [40]. A similar octaketide-producingThischalcone indicates (63) in that the PKS thepresence (HpPKS2)‘downward of malonyl-CoA fromexpanding’Hypericum as theactive extender site perforatum architecture (Scheme(St. 22) of John’s[94]. aloe Similarly, OKS wort) is large wasOKS enoughcondenses also found to to be accommodate bulky substrates and intermediates. A similar octaketide-producing PKS (HpPKS2) from promiscuousmalonyl-CoA and accepted with hexanoyl-CoA, isobutyryl-CoA, followed hexanoyl-CoA,by a C6 to C11 or C12 and to benzoyl-CoA C7 aldol-type cyclization as starters to yield to generate a a Hypericum perforatum (St. John’s wort) was also found to be promiscuous and accepted isobutyryl-CoA, series ofhexaketide triketide resorcinol to heptaketide (62) or heptaketide products [phloroglucinol95]. (64) respectively (Scheme 22). The aloe OKS hexanoyl-CoA,also utilizes long-chain and benzoyl-CoA C8–C20 acyl-CoAs as starters asto generatestarters to a seriesform triketideof triketide and to tetraketide heptaketide α products-pyrones [95].[40].

This indicates that the ‘downwardOH expanding’ active site architecture of aloe OKS is large enough to accommodate bulky substrates and intermediates. A similar octaketide-producingOH PKSOH (HpPKS2) from CoA-S Hypericum perforatum (St. John’s wort) was also found to be promiscuous and accepted isobutyryl-CoA, O CoA-S OH OKS hexanoyl-CoA, andOR benzoyl-CoA+ 5x as starters to generate a HOseries of triketide to heptaketideHO products [95]. CoA-S O O O OH O OH O HO OH O HO OH O CoA-S CoA-S OH hexaketide stilbene (61) OR hexaketide resorcinol (62) 6x OO O OKS CoA-S OH OKS OR + 5x HO HO CoA-S O O O OH O OH O OH O O HO O HO O

CoA-S OH hexaketide stilbene (61) OR hexaketide resorcinol (62) 6x HO O O OKS OH HO

O O

HOOH O O HOOH O O heptaketide chalcone (63) OR heptaketide phloroglucinol (64) HO Scheme 22. SynthesisOH ofHO various polyketides catalyzed by A. arborescens OKS. Scheme 22. A. arborescens OSynthesis of various polyketidesO catalyzed by OKS.

The rhubarbHO BAS catalyzesO a single decarboxylativeHO O condensation of malonyl-CoA to p-coumaroyl- TheCoA, rhubarb followedheptaketide by BAS another chalcone catalyzes (63)decarboxylationOR heptaketide a single to phloroglucinol form decarboxylative the (64) diketide benzalacetone. condensation As a result of of a malonyl-CoAwide active to site entrance caused by a F215L replacement (numbering in alfalfa CHS) (Figure 6B), rhubarb BAS is able p-coumaroyl-CoA, followedScheme 22. bySynthesis another of various decarboxylation polyketides catalyzed to form by A. the arborescens diketide OKS. benzalacetone. As a result ofto accept a wide a unique active range site of entrancesubstrates, causedincluding bythe abulky F215L N-methylanthraniloyl-CoA replacement (numbering as a starter in alfalfa with malonyl-CoA or methylmalonyl-CoA as the extender to yield quinolone alkaloids [96,97]. CHS) (FigureThe6 B),rhubarb rhubarb BAS catalyzes BAS isa single able decarboxylative to accept a condensation unique range of malonyl-CoA of substrates, to p-coumaroyl- including the Furthermore,CoA, followed theby another promiscuous decarboxylation enzyme tois formable the to diketidecatalyze benzalacetone. the decarboxylative As a result condensation of a wide active of bulky N-methylanthraniloyl-CoA as a starter with malonyl-CoA or methylmalonyl-CoA as the malonyl-CoAsite entrance caused with byphenylalanyl-CoA a F215L replacement (65 (numbering), followed in by alfalfa intramolecular CHS) (Figure lactamization6B), rhubarb BAS to isform able extender2,4-pyrrolidinedioneto toaccept yield a unique quinolone range (66). of alkaloids2,4-pyrrolidinedione substrates, [96 including,97]. and Furthermore, the its bulky anal ogsN-methylanthraniloyl-CoA may the undergo promiscuous further enzymespontaneous as a starter is able to catalyzedimerizationwith the decarboxylativemalonyl-CoA (Scheme or 23)methylmalonyl-CoA condensation [97]. Interestingly, of malonyl-CoA asthe the tetramic extender acid with to dimer, yield phenylalanyl-CoA 2quinolone′,5-1′,2′-dihydro-4-hydroxy- alkaloids (65 ),[96,97]. followed by intramolecular2Furthermore,′,5-bis(phenylmethyl)-[3,3 lactamization the promiscuous to′-Bi-5 form Henzymepyrrole]-2,5 2,4-pyrrolidinedione is able′(1H )-dioneto catalyze (67 (), 66 theexhibited). 2,4-pyrrolidinedionedecarboxylative moderate cytotoxiccondensation and activity its ofanalogs against murine leukemia P388 cells, thus highlighting the utility of precursor-directed biosynthesis. may undergomalonyl-CoA further with spontaneous phenylalanyl-CoA dimerization (65), followed (Scheme by intramolecular23)[97]. Interestingly, lactamization the to tetramic form acid dimer, 22,4-pyrrolidinedione1,5-11,21-dihydro-4-hydroxy-2 (66). 2,4-pyrrolidinedione1,5-bis(phenylmethyl)-[3,3 and its analogs may1-Bi-5Hpyrrole]-2,5 undergo further spontaneous1(1H)-dione (67), dimerization (Scheme 23) [97]. Interestingly, the tetramic acid dimer, 2′,5-1′,2′-dihydro-4-hydroxy- exhibited2′,5-bis(phenylmethyl)-[3,3 moderate cytotoxic activity′-Bi-5Hpyrrole]-2,5 against murine′(1H)-dione leukemia (67), exhibited P388 cells, moderate thus highlightingcytotoxic activity the utility of precursor-directedagainst murine leukemia biosynthesis. P388 cells, thus highlighting the utility of precursor-directed biosynthesis.

Scheme 23. Novel polyketides catalyzed by R. palmatum BAS.

Numerous studies have illustrated the feasibility of performing precursor-directed biosynthesis using type III PKSs to generate novel polyketides. Indeed, by establishing a combinatorial biosynthetic route in Escherichia coli [98] and exploring the substrate promiscuity of a mutant CHS with STS Scheme 23. Novel polyketides catalyzed by R. palmatum BAS. activity (18xCHS)Scheme from alfalfa 23. Novel [26], polyketidesat least 413 catalyzedpotentially by novelR. palmatum polyketideBAS. compounds were biosynthesizedNumerous (manuscript studies have in illustrated preparation). the feasibilityIn this approach, of performing various precursor-combinationsdirected of 69 biosynthesisstarter acyl-

Numeroususing type studiesIII PKSs haveto generate illustrated novel polyketides. the feasibility Indeed, of by performing establishing precursor-directeda combinatorial biosynthetic biosynthesis using typeroute III in PKSs Escherichia to generate coli [98] novel and polyketides.exploring the substrate Indeed, bypromiscuity establishing of a amutant combinatorial CHS with biosynthetic STS activity (18xCHS) from alfalfa [26], at least 413 potentially novel polyketide compounds were route in Escherichia coli [98] and exploring the substrate promiscuity of a mutant CHS with STS activity biosynthesized (manuscript in preparation). In this approach, various combinations of 69 starter acyl- (18xCHS) from alfalfa [26], at least 413 potentially novel polyketide compounds were biosynthesized Molecules 2016, 21, 806 27 of 37

(manuscript in preparation). In this approach, various combinations of 69 starter acyl-CoA and 12 extender acyl-CoA precursors generated by several promiscuous acid-CoA ligases were delivered to the mutant CHS, and the substrate proﬁle was established. Similar work in precursor-directed biosynthesis using a chalcone synthase from O. sativa led to the development of novel polyketides as antimicrobial drug leads [28]. The structural and chemical diversity of products make type III PKSs an excellent platform for the engineered biosynthesis of novel bioactive polyketides. More structural variants of starter and extender substrates can be tested in order to generate comprehensive libraries of compounds which may serve as potential drug leads.

4. Insights into Mutagenesis Studies Crystallographic analyses of the plant type III PKSs (alfalfa CHS [23], Scots pine STS [26], daisy 2-PS [35], rhubarb BAS [45], O. sativa CUS [49], C. microcarpa ACS and QNS [59], and brown alga phloroglucinol synthase [60]), the bacterial PKSs (S. coelicolor THNS [66], M. tuberculosis PKS11 [73] and PKS18 [72], A. vinelandii ArsC [80]), and of the fungal ORAS from N. crassa [83] have enabled the understanding of the basis of starter molecule selectivity, chain elongation, and cyclization of the linear polyketide intermediate. These structural studies have provided insights into the mechanistic details of type III PKSs, illustrating how subtle modiﬁcations in the active site cavity can expand the biosynthetic repertoire and result in the generation of diverse products. Manipulation of active site residues by random mutagenesis [68] and/or site-directed mutagenesis has proved useful in the understanding of and subsequently, engineering of the substrate speciﬁcity and activity of type III PKSs.

4.1. Dissection of Catalytic and Structural Roles of Residues via Mutagenesis The active site of alfalfa CHS (PDB ID: 1CGK) was found to be situated at the intersection of a CoA-binding tunnel and a bi-lobed cavity, which comprises a coumaroyl-binding pocket and a cyclization pocket [27]. Site-directed mutagenesis of the catalytic Cys164 to alanine, serine, or aspartate resulted in no naringenin production, confirming the role of Cys164 as the nucleophile and attachment site for the starter unit [25]. Mutations of Cys164 only modestly affected the malonyl-CoA decarboxylation activity (C164A and C164S), whereas mutations of Phe215, His303, and Asn336 abolished (H303A, H303D, H303N, H303T, N336H, and N336Q) or severely impaired (F215S, F215W, F215Y, N336A, N336D, and N3336K) malonyl-CoA decarboxylation activities, except for the isosteric mutation of His303 to a glutamine. Hence, Cys164 is not important for malonyl-CoA decarboxylation and replacing the nucleophile with a neutral alanine residue only prevents the loading of the starter molecule. The ability of His303 to hydrogen bond with the thioester carbonyl oxygen of the extender substrate and stabilize the negative charge in the acetyl carbanion was found to be essential for malonyl-CoA decarboxylation (Figure 17). In addition, Cys164, Phe215, and His303 mutants did not significantly affect the Kd values of CoA or acetyl-CoA, suggesting that these residues do not contribute to binding of CoA esters in the active site. Although the binding of CoA was not considerably affected, all Asn336 mutants except N336H showed a 15–40 fold increase in Kd values for acetyl-CoA binding compared to wild-type CHS, thus indicating that this residue contributes to the binding of substrates with a thioester carbonyl. The hydrogen bond between Asn336 and the thioester carbonyl oxygen of malonyl-CoA serves as a guide to ensure the proper conformation for interaction with His303, and for decarboxylation and subsequent condensation of the acetyl carbanion to the enzyme-bound starter unit or polyketide intermediate (Figure 17). Similarly, Phe215 may orient the terminal carboxylate of the extender via van der Waals forces, and aid in malonyl-CoA decarboxylation and chain extension of the polyketide intermediate. Mutagenesis studies have also identified various non-catalytic residues lining the active site cavity that are believed to be important in determining the chain length specificity and the substrate and product profiles of PKSs. These residues include Thr132, Ser133, Thr194, Thr197, Gly256, Phe265, and Ser338 (numbering in alfalfa CHS) which are typically replaced in other type III PKSs but are conserved in CHSs [27]. Daisy 2-pyrone synthase (2-PS) shares the same overall structure and catalytic Molecules 2016, 21, 806 28 of 37 triad as alfalfa CHS, but variations in three residues (T197L, G256L, and S338I) lining the active site cavity results in a significant reduction in the active site volume [35]. The 2-PS structure was the first to demonstrateMolecules a2016 simple, 21, 806 means of functional divergence in type III PKSs, whereby steric28 of 36 restriction could altercould the alter selectivity the selectivity of the of starterthe starter molecule, molecule, the number number of chain of chain elongation elongation reactions, reactions, and the and the cyclizationcyclization chemistry chemistry employed employed to terminate to terminate and and releaserelease the the polyketide polyketide product. product.

Figure 17. (A) Decarboxylation of the extender malonyl-CoA is catalyzed by His303 and Asn336 in alfalfa Figure 17.CHS.(A )Phe215 Decarboxylation is thought to orient of the the terminal extender carboxylate malonyl-CoA of malonyl-CoA is catalyzedand aid in the by reaction; His303 (B) The and Asn336 in alfalfa CHS.acetyl carbanion Phe215 subsequently is thought undergoes to orient a condensation the terminal with carboxylate the enzyme-bound of malonyl-CoA triketide intermediate. and aid in the reaction; (B) The acetyl carbanion subsequently undergoes a condensation with the enzyme-bound In H. androsaemum benzophenone synthase (BPS), a T132L mutation (numbering in alfalfa CHS) triketide intermediate. resulted in changes in the active site cavity which limited polyketide chain extension and transformed the enzyme from a biphenyl synthase into a phenylpyrone synthase with a trace of the wild-type In H.activityandrosaemum [99]. In anotherbenzophenone study, structure-based synthase mutagenic (BPS), a transformation T132L mutation of alfalfa (numbering CHS resulted in in alfalfa a CHS) resulted inchange changes in chain in the termination active site chemistry cavity whichfrom a Claisen limited condensation polyketide typical chain of extension CHSs to an and aldol transformed condensation usually catalyzed by STSs, indicating that functional divergence can be determined by the enzymeelectronic from factors a biphenyl rather than synthase steric ef intofects of a the phenylpyrone active site cavity synthase [26]. with a trace of the wild-type activity [99]. InAloe another contains study, two highly structure-based similar type III mutagenic PKSs, pentaketide transformation chromone synthase of alfalfa (PCS) CHS and resulted in a change inoctaketide chain synthase termination (OKS), chemistrywhich share more from than a Claisen 90% identity condensation with each other. typical Although of these CHSs two to an aldol condensationenzymes usually possess catalyzed different bybiochemical STSs, indicating functions, thata single functional mutational divergence change of Met197 can be in determined the by pentaketide-producing PCS to Gly197 in OKS (numbering in alfalfa CHS) resulted in the synthesis of electronicthe factors octaketides, rather SEK4 than (12 steric) and SEK4b effects (13 of) (Scheme the active 2) [38]. site Conversely, cavity [26 a ].reciprocal mutation in aloe AloeOKS contains to the bulky two methionine highly similar residue typeresulted III in PKSs,a loss of pentaketide octaketide synthase chromone activity and synthase a gain in (PCS) and octaketidepentaketide synthase synthase (OKS), activity which [40]. share In addition, more thanvarious 90% small-to-large identity substitutions with each in other. OKS (G197A, Although these two enzymesG197T, possess G197M, G197L, different G197F, biochemical and G197W) resulted functions, in a loss a of single octaketide mutational production changeand a concomitant of Met197 in the generation of shorter chain length polyketides (from heptaketide to triketide), depending on the steric pentaketide-producingbulk of the side chain. PCS Therefore, to Gly197 residue in OKS 197 is (numbering likely to control in polyketide alfalfa CHS)chain length resulted and product in the synthesis of the octaketides,specificity. The SEK4 reduced (12 )steric and bulk SEK4b due to ( the13) M197G (Scheme substitution2)[ 38 in]. the Conversely, PCS mutant unlocks a reciprocal two buried mutation in aloe OKSpockets, to the thus bulky expanding methionine the available residue cyclization resulted cavity in to a accommodate loss of octaketide the octaketide synthase product activity[38]. and a gain in pentaketideInterestingly, synthasethe location activityand orientation [40]. of In one addition, of the pockets various of the PCS small-to-large M197G mutant substitutions superimpose in OKS well with a similar acyl-binding tunnel in OKS [94], in bacterial type III PKSs such as THNS [66] and (G197A, G197T,PKS18 [72], G197M, and in the G197L, fungal G197F, 2′-oxoalkylresorcylic and G197W) acid resultedsynthase (O inRAS) a loss [83]. ofSince octaketide aloe OKS can production utilize and a concomitantlong-chain generation (C6–C20) of shorteracyl-CoAs chain to form length triketide polyketides and tetraketide (from pyrones heptaketide [40], it is topossible triketide), that the depending on the stericplant, bulk bacterial, of the and side fungal chain. type Therefore,III PKSs share residue a similar 197active is site likely architectural to control strategy polyketide to synthesize chain length and productthe long-chain speciﬁcity. polyketides The reduced such as pyrones, steric bulk phloroglucinols, due to the and M197G resorcinols. substitution in the PCS mutant unlocks two4.2. Structure-Based buried pockets, Engineering thus an expandingd the Versatility the of availableType III PKSs cyclization cavity to accommodate the octaketide productManipulation [38]. Interestingly,of the type III PKS the reactions location by andusing orientationrationally engineered of one mutant of the enzymes pockets in of the PCS M197G mutantconjunction superimpose with substrate well analogues with can a similar result in acyl-bindinga further generation tunnel of chemically in OKS and [94 structurally], in bacterial type III PKSs suchdistinct as unnatural THNS [polyketides.66] and PKS18 The ‘gatekeeper’ [72], and Phe215 in the is generally fungal 2conserved1-oxoalkylresorcylic in most type III PKSs acid synthase (ORAS) [83and]. Since is located aloe between OKS can the utilize CoA-binding long-chain tunnel (C6–C20)and the entrance acyl-CoAs of the toactive form site triketide cavity. Despite and tetraketide pyrones [40 ], it is possible that the plant, bacterial, and fungal type III PKSs share a similar active site architectural strategy to synthesize the long-chain polyketides such as pyrones, phloroglucinols, and resorcinols.

4.2. Structure-Based Engineering and the Versatility of Type III PKSs Manipulation of the type III PKS reactions by using rationally engineered mutant enzymes in conjunction with substrate analogues can result in a further generation of chemically and structurally Molecules 2016, 21, 806 29 of 37 distinct unnatural polyketides. The ‘gatekeeper’ Phe215 is generally conserved in most type III PKSs and is located between the CoA-binding tunnel and the entrance of the active site cavity. Despite having a lower malonyl-CoA decarboxylation activity compared to wild-type alfalfa CHS, the F215S mutant is able to utilize p-coumaroyl-CoA, phenylacetyl-CoA, benzoyl-CoA, and C5–C8 acyl-CoAs with less than 3% the catalytic efficiency of the wild-type enzyme to produce tetraketide lactones [100]. Intriguingly, although N-methylanthraniloyl-CoA is not normally utilized by the wild type CHS, the Molecules 2016, 21, 806 29 of 36 F215S mutant was found to accept the bulky starter with a comparable kcat/Km to R. graveolens acridone synthasehaving (ACS), a lower to form malonyl-CoA a tetraketide decarboxylationN-methylanthraniloyltriacetic activity compared to wild-type acid lactone alfalfa CHS, (31). the A modificationF215S into a serinemutant residue is able to at utilize position p-coumaroyl-CoA, 215 results in phenylacetyl-CoA, a reduction in steric benzoyl-CoA, hindrance and andC5–C8 a wideracyl-CoAs active site entrance,with thus less allowing than 3% the catalytic entry of efficiency the bulky of Nthe-methylanthraniloyl-CoA. wild-type enzyme to produce tetraketide lactones [100]. Intriguingly, although N-methylanthraniloyl-CoA is not normally utilized by the wild type Although octaketides are not usually synthesized by the wild type S. baicalensis CHS, the S338V CHS, the F215S mutant was found to accept the bulky starter with a comparable kcat/Km to R. graveolens mutantacridone CHS was synthase found (ACS), to produce to form trace a tetraketide amounts N-methylanthraniloyltriacetic of SEK4 (12) and SEK4b acid (13) lactone using eight(31). A units of malonyl-CoAmodification (Scheme into 2aC) serine [ 101 residue]. This at S338Vposition mutation 215 results wasin a reduction proposed in steric to facilitate hindrance the and extension a wider of the linear polyketideactive site entrance, intermediate thus allowing into a buried the entry pocket of the inbulky the N active-methylanthraniloyl-CoA. site cavity of the enzyme by providing steric guidance.Although It was octaketides demonstrated are not forusually the firstsynthesized time that by the CHS wild could type beS. baicalensis engineered CHS, to the generate S338V longer polyketidesmutant such CHS as was octaketides. found to produce Furthermore, trace amounts the synthesis of SEK4 of (12 SEK4) and (SEK4b12) and (13 SEK4b) using (eight13) by unitsS.baicalensis of malonyl-CoA (Scheme 2C) [101]. This S338V mutation was proposed to facilitate the extension of CHS wasthe dramatically linear polyketide improved intermediate in an into OKS-like a buried triple pocket mutant in the (T197G/G256L/S338V).active site cavity of the enzyme These by results suggestedproviding that rational steric guidance. site-directed It was demonstrated mutagenesis for the efforts first time can that be CHS used could to generate be engineered CHSs to generate with different substratelonger selectivities polyketides and such altered as octaketides. product Furthermore, profiles, thereby the synthesis producing of SEK4 diverse (12) and novel SEK4b polyketides. (13) by Structure-basedS. baicalensis CHS engineering was dramatically efforts improved on the in an aloe OKS- PCSlike and triple OKS mutant further (T197G/G256L/S338V). extended the biosyntheticThese repertoireresults of type suggested III PKSs. that Whenrationaltwo site-directed aromatic mutage residuesnesis atefforts the bottomcan be used of the to buriedgenerate pocketCHSs with of the PCS different substrate selectivities and altered product profiles, thereby producing diverse novel polyketides. M207G mutant (corresponding to Thr197 in alfalfa CHS) were simultaneously replaced by alanine, Structure-based engineering efforts on the aloe PCS and OKS further extended the biosynthetic the activerepertoire site cavity of type was III PKSs. further When expanded two aromatic by approximately residues at the bottom four of times the buried compared pocket toof the PCS wild-type PCS [102M207G]. Interestingly, mutant (corresponding the PCS F80A/Y82A/M207G to Thr197 in alfalfa CHS) mutant were was simultaneously able to catalyze replaced the by decarboxylative alanine, condensationthe active of site nine cavity units was of further malonyl-CoA expanded by with approximately an altered four mechanism times compared of cyclization to the wild-type to yield an unnaturalPCS nonaketide [102]. Interestingly, naphthopyrone the PCS F80A/Y82A/M207G (68) (Scheme 24 muA).tant Similarly was able to in catalyze OKS, when the decarboxylative Asn222 situated at the bottomcondensation of the elongation of nine units cavity of wasmalonyl-CoA replaced with by a an glycine altered residue, mechanism a novel of cyclization decaketide to benzophenoneyield an unnatural nonaketide naphthopyrone (68) (Scheme 24A). Similarly in OKS, when Asn222 situated at SEK15 (69) was produced from ten units of malonyl-CoA (Scheme 24B) [94]. The N222G mutant OKS the bottom of the elongation cavity was replaced by a glycine residue, a novel decaketide benzophenone was alsoSEK15 able ( to69) accept was produced phenylacetyl-CoA from ten units of or malonyl-Co benzoyl-CoAA (Scheme as starters 24B) [94]. to The form N222G novel mutant octaketides OKS and heptaketidewas also benzophenone able to accept phenylacetyl-CoA respectively [103 or]. benz Simultaneousoyl-CoA as starters mutation to form of novel the neighbouringoctaketides and Phe66 residueheptaketide to leucine benzophenone resulted in an respecti OKSvely double [103]. mutant Simultan thateous is mutation capable of of the synthesizing neighbouring an Phe66 unnatural dodecaketideresidue naphthophenone to leucine resulted TW95ain an OKS (70 )double by the muta successivent that is decarboxylative capable of synthesizing condensations an unnatural of 12 units of malonyl-CoA,dodecaketide making naphthophenone it one of TW95a the longest (70) by polyketide the successive synthesized decarboxylative by condensations the structurally of 12 simpleunits type of malonyl-CoA, making it one of the longest polyketide synthesized by the structurally simple type III PKS (SchemeIII PKS (Scheme 24C) [24C)104]. [104].

Scheme 24. Novel polyketides biosynthesized by mutant type III PKSs. (A) PCS F80A/Y82A/M207G Scheme 24. Novel polyketides biosynthesized by mutant type III PKSs. (A) PCS F80A/Y82A/M207G mutant catalyzes the formation of napthopyrone; (B) OKS N222G mutant catalyzes the formation of mutant catalyzesSEK15; and the(C) OKS formation N222G/F66L of napthopyrone; mutant catalyzes (B the) OKS formation N222G of TW95a. mutant catalyzes the formation of SEK15; and (C) OKS N222G/F66L mutant catalyzes the formation of TW95a.

Molecules 2016, 21, 806 30 of 37

AMolecules recently 2016 discovered, 21, 806 CHS from a primitive club moss Huperzia serrata (HsPKS1) was30 found of 36 to display unusually broad substrate selectivity and produced unnatural polyketide-alkaloid hybrid A recently discovered CHS from a primitive club moss Huperzia serrata (HsPKS1) was found to molecules [105]. Using precursor-directed approaches, HsPKS1 synthesized a novel pyridoisoindole display unusually broad substrate selectivity and produced unnatural polyketide-alkaloid hybrid (74) frommolecules the [105]. condensation Using precursor-directed of 2-carbamoylbenzoyl-CoA approaches, HsPKS1 (71 synthesized), a synthetic a novel nitrogen-containing pyridoisoindole non-physiological(74) from the starter,condensation with twoof 2-carbamoylbenzoyl-CoA units of malonyl-CoA (Scheme(71), a synthetic 25A). Conversely, nitrogen-containing the promiscuous non- aloe OKSphysiological was only starter, able to with generate two units a tetraketide of malonyl-CoA pyrone (Scheme as a minor 25A). product Conversely, from the the promiscuous bulky synthetic starter.aloe HsPKS1 OKS was was only also able able to togenerate generate a tetraketide novel alkaloids pyrone fromas a minor the starters, product 3-carbamoylpicolinoyl-CoAfrom the bulky synthetic (72) andstarter. 3-carbamoyl-2-naphthoyl-CoA HsPKS1 was also able to generate novel (73) (Schemealkaloids from25A). the Thestarters, structure-based 3-carbamoylpicolinoyl-CoA S348G mutant (corresponding(72) and 3-carbamoyl-2-naphthoyl-CoA to Ser338 in alfalfa CHS) results (73) in(Scheme an expansion 25A). The of the struct activeure-based site cavity S348G which mutant not only extended(corresponding the polyketide to Ser338 chain in length,alfalfa CHS) but alsoresults altered in an expansion the cyclization of the mechanismactive site cavity to yield which a not biologically only active,extended ring-expanded the polyketide dibenzoazepine chain length, (but75) also from altered the condensationthe cyclization mechanism of 2-carbamoylbenzoyl-CoA to yield a biologically with active, ring-expanded dibenzoazepine (75) from the condensation of 2-carbamoylbenzoyl-CoA with three units of malonyl-CoA (Scheme 25B). three units of malonyl-CoA (Scheme 25B).

Scheme 25. Synthesis of polyketide-alkaloid hybrids catalyzed by (A) H. serrata HsPKS1 and its (B) Scheme 25. Synthesis of polyketide-alkaloid hybrids catalyzed by (A) H. serrata HsPKS1 and its S348G mutant. (B) S348G mutant. Remarkably, the dibenzoazepine alkaloid inhibited biofilm formation by methicillin-susceptible Remarkably,Staphylococcus theaureus dibenzoazepine (MIC = 12.0 μg/mL). alkaloid Hence, inhibited structure-based biofilm engi formationneering of by type methicillin-susceptible III PKSs coupled Staphylococcuswith precursor-directed aureus (MIC biosynthesis = 12.0 µg/mL). using ration Hence,ally structure-baseddesigned substrate engineering analogues can of provide type III an PKSs coupledexcellent with platform precursor-directed for the further biosynthesisdevelopment of using unnatural rationally novel polyketides designed with substrate unique analoguestherapeutic can functions. provide an excellent platform for the further development of unnatural novel polyketides with unique Studies on S. griseus RppA showed that the bulky Tyr224 (corresponding to Gly256 in alfalfa CHS) therapeuticis important functions. for RppA to select malonyl-CoA as a starter unit [106]. Mutations at this site (Y224A, StudiesY224C, Y224G, on S. griseusand Y224L)RppA led showedto a decrease that in the competition bulky Tyr224 for malonyl-CoA (corresponding and the to mutant Gly256 enzyme in alfalfa CHS)was is important able to utilize for phenylacetyl-CoA RppA to select or malonyl-CoA hexanoyl-CoA as instead a starter to synthesize unit [106 tetraketide]. Mutations and triketide at this site (Y224A,pyrones. Y224C, This Y224G, is analogous and Y224L) to the ledG256L to areplacemen decrease int in competition rhubarb ALS for which malonyl-CoA utilizes acetyl-CoA and the as mutant a enzymestarter was instead able of to the utilize bulky phenylacetyl-CoAp-coumaroyl-CoA, suggesting or hexanoyl-CoA that the residue instead at position to synthesize 256 controls tetraketide starter and triketideunit selectivity pyrones. [37]. Similar This is Tyr224 analogous mutations to the in S. G256L coelicolor replacement THNS showed in that rhubarb the mutant ALS enzyme which can utilizes acetyl-CoAproduce as triketide a starter pyrones instead from of theacetyl-CoA, bulky p acetoacetyl-CoA,-coumaroyl-CoA, hexanoyl-CoA, suggesting and that benzoyl-CoA the residue starters at position in significantly higher amounts compared to the wild-type THNS, due to a decrease in loading 256 controls starter unit selectivity [37]. Similar Tyr224 mutations in S. coelicolor THNS showed that efficiency of the malonyl-CoA starter [107]. In addition, Ala305 (corresponding to Ser338 in alfalfa CHS) the mutant enzyme can produce triketide pyrones from acetyl-CoA, acetoacetyl-CoA, hexanoyl-CoA, was found to be important for chain elongation, as a bulky 2-PS-like A305I mutation in S. griseus RppA and benzoyl-CoAresulted in the starters generation in significantlyof triketide pyrones higher only amounts from phenylacetyl-CoA compared to the or wild-type hexanoyl-CoA THNS, starters due to a decrease in loading efficiency of the malonyl-CoA starter [107]. In addition, Ala305 (corresponding to Ser338 in alfalfa CHS) was found to be important for chain elongation, as a bulky 2-PS-like A305I Molecules 2016, 21, 806 31 of 37 mutation in S. griseus RppA resulted in the generation of triketide pyrones only from phenylacetyl-CoA or hexanoyl-CoA starters [106]. These results indicated that the overall backbone architecture of the active site cavity of RppA is similar to that of plant type III PKSs, thus structure-based engineering efforts in plant PKSs are likely to translate well to bacterial type III PKSs with minor differences.

5. Conclusions Recently identiﬁed type III PKSs that possess novel catalytic activities, such as quinolone synthase (QNS), Rhodospirillum polyketide synthase (RpsA), and Botrytis cinerea type III PKS (BPKS), can provide unique polyketide scaffolds for synthetic enzymology. Until recently, biosynthesis of novel polyketide libraries has been dependent upon the broad substrate speciﬁcities of the type III PKSs. By rationally incorporating amino acid substitutions, the active site cavity of type III PKSs can be altered or expanded to facilitate the synthesis of more complex polyketides. This characteristic makes type III PKSs great candidates for structure-based engineering approaches to expand the chemical and structural diversity of natural products derived from these systems. Further studies on the catalytic potential and versatility of the functionally divergent type III PKSs should provide more intimate structural details of the biosynthetic processes. Since the diversity of the polyketides biosynthesized by PKSs is partly dependent on the starter and extender units, variations in the choice of starters and extenders is important. However, there has been a lack of a toolkit that describes the means to introduce novel acyl-CoA precursors to PKSs for polyketide biosynthesis. Instead of chemical synthesis, an alternative way to generate a wide range of novel acyl-CoA esters is to make use of promiscuous acid-CoA ligases with unique carboxylic acid scaffolds [98]. By determining substrate promiscuities of acid-CoA ligases and PKSs beyond conventional substrate pools, we can establish novel enzymatic routes for the biosynthesis of unnatural polyketides which can be subsequently applied to drug screening efforts. These studies serve as a primer towards the development of novel and valuable polyketides with therapeutic potential via synthetic enzymology.

Acknowledgments: Molecular graphics and analyses were performed with the UCSF Chimera package. Chimera is developed by the Resource for Biocomputing, Visualization, and Informatics at the University of California, San Francisco (supported by NIGMS P41-GM103311). This review was supported by grants from the National Medical Research Council, Singapore, and the Academic Research Fund of the Ministry of Education, Singapore. Author Contributions: The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. Conflicts of Interest: The authors declare no conflict of interest. The founding sponsors had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, and in the decision to publish the results.

Abbreviations The following abbreviations are used in this manuscript:

ACP Acyl carrier protein ACS Acridone synthase ALS Aloesone synthase AnPKS Aspergillus niger polyketide synthase ARAS2 Alkyl-resorcylic acid synthase 2 BAS Benzalacetone synthase BBS Bibenzyl synthase BIS Biphenyl synthase BPS Benzophenone synthase CHS Chalcone synthase CoA Coenzyme A CsyA Chalcone synthase-like A CTAL Coumaroyl triacetic acid lactone Molecules 2016, 21, 806 32 of 37

CTAS p-coumaroyl triacetic acid synthase CURS Curcumin synthase CUS Curcuminoid synthase DCS Diketide-CoA synthase DpgA 3,5-dihydroxyphenylacetic acid synthase GCS Germicidin synthase NAC N-acetylcysteamine OKS Octaketide synthase ORAS 21-oxoalkylresorcylic acid synthase PCS Pentaketide chromone synthase PhlD Phloroglucinol synthase PKS Polyketide synthase PpASCL Physcomitrella patens anther-speciﬁc chalcone synthase-like enzyme 2-PS 2-pyrone synthase QNS Quinolone synthase STCS Stilbenecarboxylate synthase STS Stilbene synthase THNS 1,3,6,8-tetrahydroxynaphthalene synthase

References

1. Collie, J.N. Derivatives of the multiple keten group. J. Chem. Soc. Trans. 1907, 91, 1806–1813. [CrossRef] 2. Staunton, J.; Weissman, K.J. Polyketide biosynthesis: A millennium review. Nat. Prod. Rep. 2001, 18, 380–416. [CrossRef][PubMed] 3. Calne, R.Y.; Collier, D.S.; Lim, S.; Pollard, S.G.; Samaan, A.; White, D.J.; Thiru, S. Rapamycin for immunosuppression in organ allografting. Lancet 1989, 2, 227. [CrossRef] 4. Alberts, A.W.; Chen, J.; Kuron, G.; Hunt, J.; Huff, J.; Hoffman, C.; Rothrock, J.; Lopez, M.; Joshua, H.; Harris, E.; et al. Mevinolin—A highly potent competitive inhibitor of hydroxymethylglutaryl-coenzyme a reductase and a cholesterol-lowering agent. Proc. Natl. Acad. Sci. USA 1980, 77, 3957–3961. [CrossRef] [PubMed] 5. Wright, L.F.; Hopwood, D.A. Actinorhodin is a chromosomally-determined antibiotic in Streptomyces coelicolor A3(2). J. Gen. Microbiol. 1976, 96, 289–297. [CrossRef][PubMed] 6. Houtkooper, R.H.; Mouchiroud, L.; Ryu, D.; Moullan, N.; Katsyuba, E.; Knott, G.; Williams, R.W.; Auwerx, J. Mitonuclear protein imbalance as a conserved longevity mechanism. Nature 2013, 497, 451–457. [CrossRef] [PubMed] 7. Newman, D.J.; Cragg, G.M. Natural products as sources of new drugs over the 30 years from 1981 to 2010. J. Nat. Prod. 2012, 75, 311–335. [CrossRef][PubMed] 8. Tatsuta, K.; Hosokawa, S. Total syntheses of polyketide-derived bioactive natural products. Chem. Rec. 2006, 6, 217–233. [CrossRef][PubMed] 9. Birch, A.J.; Donovan, F.W. Studies in relation to biosynthesis. I. Some possible routes to derivatives of orcinol and phloroglucinol. Aust. J. Chem. 1953, 6, 360–368. [CrossRef] 10. Dunn, B.J.; Khosla, C. Engineering the acyltransferase substrate speciﬁcity of assembly line polyketide synthases. J. R. Soc. Interface R. Soc. 2013, 10, 20130297. [CrossRef][PubMed] 11. Abe, I. Novel applications of plant polyketide synthases. Curr. Opin. Chem. Biol. 2012, 16, 179–185. [CrossRef] [PubMed] 12. Till, M.; Race, P.R. Progress challenges and opportunities for the re-engineering of trans-at polyketide synthases. Biotechnol. Lett. 2014, 36, 877–888. [CrossRef][PubMed] 13. Shen, B. Polyketide biosynthesis beyond the type I, II and III polyketide synthase paradigms. Curr. Opin. Chem. Biol. 2003, 7, 285–295. [CrossRef] 14. Cox, R.J. Polyketides, proteins and genes in fungi: Programmed nano-machines begin to reveal their secrets. Org. Biomol. Chem. 2007, 5, 2010–2026. [CrossRef][PubMed] 15. Stewart, C., Jr.; Vickery, C.R.; Burkart, M.D.; Noel, J.P. Conﬂuence of structural and chemical biology: Plant polyketide synthases as biocatalysts for a bio-based future. Curr. Opin. Plant Biol. 2013, 16, 365–372. [CrossRef][PubMed] Molecules 2016, 21, 806 33 of 37

16. Funa, N.; Ohnishi, Y.; Fujii, I.; Shibuya, M.; Ebizuka, Y.; Horinouchi, S. A new pathway for polyketide synthesis in microorganisms. Nature 1999, 400, 897–899. [PubMed] 17. Bangera, M.G.; Thomashow, L.S. Identification and characterization of a gene cluster for synthesis of the polyketide antibiotic 2,4-diacetylphloroglucinol from Pseudomonas fluorescens Q2-87. J. Bacteriol. 1999, 181, 3155–3163. [PubMed] 18. Pfeifer, V.; Nicholson, G.J.; Ries, J.; Recktenwald, J.; Schefer, A.B.; Shawky, R.M.; Schroder, J.; Wohlleben, W.; Pelzer, S. A polyketide synthase in glycopeptide biosynthesis: The biosynthesis of the non-proteinogenic amino acid (S)-3,5-dihydroxyphenylglycine. J. Biol. Chem. 2001, 276, 38370–38377. [CrossRef][PubMed] 19. Seshime, Y.; Juvvadi, P.R.; Fujii, I.; Kitamoto, K. Discovery of a novel superfamily of type III polyketide synthases in Aspergillus oryzae. Biochem. Biophys. Res. Commun. 2005, 331, 253–260. [CrossRef][PubMed] 20. Funa, N.; Awakawa, T.; Horinouchi, S. Pentaketide resorcylic acid synthesis by type III polyketide synthase from Neurospora crassa. J. Biol. Chem. 2007, 282, 14476–14481. [CrossRef][PubMed] 21. Li, J.; Luo, Y.; Lee, J.K.; Zhao, H. Cloning and characterization of a type III polyketide synthase from Aspergillus niger. Bioorg. Med. Chem. Lett. 2011, 21, 6085–6089. [CrossRef][PubMed] 22. Schuz, R.; Heller, W.; Hahlbrock, K. Substrate specificity of chalcone synthase from petroselinum hortense. J. Biol. Chem. 1983, 258, 6730–6734. [PubMed] 23. Ferrer, J.L.; Jez, J.M.; Bowman, M.E.; Dixon, R.A.; Noel, J.P. Structure of chalcone synthase and the molecular basis of plant polyketide biosynthesis. Nat. Struct. Biol. 1999, 6, 775–784. [PubMed] 24. Austin, M.B.; Noel, J.P. The chalcone synthase superfamily of type III polyketide synthases. Nat. Prod. Rep. 2003, 20, 79–110. [CrossRef][PubMed] 25. Jez, J.M.; Ferrer, J.L.; Bowman, M.E.; Dixon, R.A.; Noel, J.P.Dissection of malonyl-coenzyme a decarboxylation from polyketide formation in the reaction mechanism of a plant polyketide synthase. Biochemistry 2000, 39, 890–902. [CrossRef][PubMed] 26. Austin, M.B.; Bowman, M.E.; Ferrer, J.L.; Schroder, J.; Noel, J.P. An aldol switch discovered in stilbene synthases mediates cyclization specificity of type III polyketide synthases. Chem. Biol. 2004, 11, 1179–1194. [CrossRef][PubMed] 27. Abe, I.; Morita, H. Structure and function of the chalcone synthase superfamily of plant type III polyketide synthases. Nat. Prod. Rep. 2010, 27, 809–838. [CrossRef][PubMed] 28. Go, M.K.; Wongsantichon, J.; Cheung, V.W.N.; Chow, J.Y.; Robinson, R.C.; Yew, W.S. Synthetic polyketide enzymology: Platform for biosynthesis of antimicrobial polyketides. ACS Catal. 2015, 5, 4033–4042. [CrossRef] 29. Tropf, S.; Lanz, T.; Rensing, S.A.; Schroder, J.; Schroder, G. Evidence that stilbene synthases have developed from chalcone synthases several times in the course of evolution. J. Mol. Evol. 1994, 38, 610–618. [CrossRef] [PubMed] 30. Yamaguchi, T.; Kurosaki, F.; Suh, D.Y.; Sankawa, U.; Nishioka, M.; Akiyama, T.; Shibuya, M.; Ebizuka, Y. Cross-reaction of chalcone synthase and stilbene synthase overexpressed in Escherichia coli. FEBS Lett. 1999, 460, 457–461. [CrossRef] 31. Shomura, Y.; Torayama, I.; Suh, D.Y.; Xiang, T.; Kita, A.; Sankawa, U.; Miki, K. Crystal structure of stilbene synthase from Arachis hypogaea. Proteins 2005, 60, 803–806. [CrossRef][PubMed] 32. Harris, T.M.; Carney, R.L. Biogenetically modeled synthesis of β-resorcylic acids. J. Am. Chem. Soc. 1966, 88, 2053–2054. [CrossRef] 33. Helariutta, Y.; Elomaa, P.; Kotilainen, M.; Griesbach, R.J.; Schroder, J.; Teeri, T.H. Chalcone synthase-like genes active during corolla development are differentially expressed and encode enzymes with different catalytic properties in Gerbera hybrida (asteraceae). Plant Mol. Biol. 1995, 28, 47–60. [CrossRef][PubMed] 34. Eckermann, S.; Schroder, G.; Schmidt, J.; Strack, D.; Edrada, R.A.; Helariutta, Y.; Elomaa, P.; Kotilainen, M.; Kilpelainen, I.; Proksch, P.; et al. New pathway to polyketides in plants. Nature 1998, 396, 387–390. [CrossRef] 35. Jez, J.M.; Austin, M.B.; Ferrer, J.L.; Bowman, M.E.; Schroder, J.; Noel, J.P. Structural control of polyketide formation in plant-specific polyketide synthases. Chem. Biol. 2000, 7, 919–930. [CrossRef] 36. Abe, I.; Utsumi, Y.; Oguro, S.; Noguchi, H. The first plant type III polyketide synthase that catalyzes formation of aromatic heptaketide. FEBS Lett. 2004, 562, 171–176. [CrossRef] 37. Abe, I.; Watanabe, T.; Lou, W.; Noguchi, H. Active site residues governing substrate selectivity and polyketide chain length in aloesone synthase. FEBS J. 2006, 273, 208–218. [CrossRef][PubMed] Molecules 2016, 21, 806 34 of 37

38. Morita, H.; Kondo, S.; Oguro, S.; Noguchi, H.; Sugio, S.; Abe, I.; Kohno, T. Structural insight into chain-length control and product specificity of pentaketide chromone synthase from Aloe arborescens. Chem. Biol. 2007, 14, 359–369. [CrossRef][PubMed] 39. Mizuuchi, Y.; Shi, S.-P.; Wanibuchi, K.; Kojima, A.; Morita, H.; Noguchi, H.; Abe, I. Novel type III polyketide synthases from Aloe arborescens. FEBS J. 2009, 276, 2391–2401. [CrossRef][PubMed] 40. Abe, I.; Oguro, S.; Utsumi, Y.; Sano, Y.; Noguchi, H. Engineered biosynthesis of plant polyketides: Chain length control in an octaketide-producing plant type III polyketide synthase. J. Am. Chem. Soc. 2005, 127, 12709–12716. [CrossRef][PubMed] 41. Akiyama, T.; Shibuya, M.; Liu, H.M.; Ebizuka, Y. P-coumaroyltriacetic acid synthase, a new homologue of chalcone synthase, from Hydrangea macrophylla var. thunbergii. Eur. J. Biochem. 1999, 263, 834–839. [CrossRef] [PubMed] 42. Eckermann, C.; Schroder, G.; Eckermann, S.; Strack, D.; Schmidt, J.; Schneider, B.; Schroder, J. Stilbenecarboxylate biosynthesis: A new function in the family of chalcone synthase-related proteins. Phytochemistry 2003, 62, 271–286. [CrossRef] 43. Preisig-Muller, R.; Gnau, P.; Kindl, H. The inducible 9,10-dihydrophenanthrene pathway: Characterization and expression of bibenzyl synthase and s-adenosylhomocysteine hydrolase. Arch. Biochem. Biophys. 1995, 317, 201–207. [CrossRef][PubMed] 44. Abe, I.; Takahashi, H.; Morita, H.; Noguchi, H. Benzalacetone synthase. A novel polyketide synthase that plays a crucial role in the biosynthesis of phenylbutanones in Rheum palmatum. Eur. J. Biochem. 2001, 268, 3354–3359. [CrossRef][PubMed] 45. Morita, H.; Shimokawa, Y.; Tanio, M.; Kato, R.; Noguchi, H.; Sugio, S.; Kohno, T.; Abe, I. A structure-based mechanism for benzalacetone synthase from Rheum palmatum. Proc. Natl. Acad. Sci. USA 2010, 107, 669–673. [CrossRef][PubMed] 46. Shimokawa, Y.; Morita, H.; Abe, I. Structure-based engineering of benzalacetone synthase. Bioorg. Med. Chem. Lett. 2010, 20, 5099–5103. [CrossRef][PubMed] 47. Lu, H.; Yang, M.; Liu, C.; Lu, P.; Cang, H.; Ma, L. Protein preparation, crystallization and preliminary X-ray analysis of Polygonum cuspidatum bifunctional chalcone synthase/benzalacetone synthase. Acta Crystallogr. Sec. F Struct. Biol. Cryst. Commun. 2013, 69, 871–875. [CrossRef][PubMed] 48. Katsuyama, Y.; Kita, T.; Funa, N.; Horinouchi, S. Curcuminoid biosynthesis by two type III polyketide synthases in the herb curcuma longa. J. Biol. Chem. 2009, 284, 11160–11170. [CrossRef][PubMed] 49. Morita, H.; Wanibuchi, K.; Nii, H.; Kato, R.; Sugio, S.; Abe, I. Structural basis for the one-pot formation of the diarylheptanoid scaffold by curcuminoid synthase from oryza sativa. Proc. Natl. Acad. Sci. USA 2010, 107, 19778–19783. [CrossRef][PubMed] 50. Beerhues, L.; Liu, B. Biosynthesis of biphenyls and benzophenones—evolution of benzoic acid-specific type III polyketide synthases in plants. Phytochemistry 2009, 70, 1719–1727. [CrossRef][PubMed] 51. Liu, B.; Falkenstein-Paul, H.; Schmidt, W.; Beerhues, L. Benzophenone synthase and chalcone synthase from Hypericum androsaemum cell cultures: Cdna cloning, functional expression, and site-directed mutagenesis of two polyketide synthases. Plant J. 2003, 34, 847–855. [CrossRef][PubMed] 52. Nualkaew, N.; Morita, H.; Shimokawa, Y.; Kinjo, K.; Kushiro, T.; De-Eknamkul, W.; Ebizuka, Y.; Abe, I. Benzophenone synthase from garcinia mangostana l. Pericarps. Phytochemistry 2012, 77, 60–69. [CrossRef] [PubMed] 53. Beerhues, L. Benzophenone synthase from cultured cells of Centaurium erythraea. FEBS Lett. 1996, 383, 264–266. [CrossRef] 54. Liu, B.; Raeth, T.; Beuerle, T.; Beerhues, L. Biphenyl synthase, a novel type III polyketide synthase. Planta 2007, 225, 1495–1503. [CrossRef][PubMed] 55. Baumert, A.; Maier, W.; Groger, D.; Deutzmann, R. Purification and properties of acridone synthase from cell suspension cultures of Ruta graveolens L. Z. Naturforsch. C 1994, 49, 26–32. [PubMed] 56. Springob, K.; Lukacin, R.; Ernwein, C.; Groning, I.; Matern, U. Specifities of functionally expressed chalcone and acridone synthases from Ruta graveolens. Eur. J. Biochem. 2000, 267, 6552–6559. [CrossRef][PubMed] 57. Resmi, M.S.; Verma, P.; Gokhale, R.S.; Soniya, E.V. Identification and characterization of a type III polyketide synthase involved in quinolone alkaloid biosynthesis from Aegle marmelos correa. J. Biol. Chem. 2013, 288, 7271–7281. [CrossRef][PubMed] Molecules 2016, 21, 806 35 of 37

58. Lukacin, R.; Schreiner, S.; Matern, U. Transformation of acridone synthase to chalcone synthase. FEBS Lett. 2001, 508, 413–417. [CrossRef] 59. Mori, T.; Shimokawa, Y.; Matsui, T.; Kinjo, K.; Kato, R.; Noguchi, H.; Sugio, S.; Morita, H.; Abe, I. Cloning and structure-function analyses of quinolone- and acridone-producing novel type III polyketide synthases from Citrus microcarpa. J. Biol. Chem. 2013, 288, 28845–28858. [CrossRef][PubMed] 60. Meslet-Cladiere, L.; Delage, L.; Leroux, C.J.; Goulitquer, S.; Leblanc, C.; Creis, E.; Gall, E.A.; Stiger-Pouvreau, V.; Czjzek, M.; Potin, P. Structure/function analysis of a type III polyketide synthase in the brown alga ectocarpus siliculosus reveals a biochemical pathway in phlorotannin monomer biosynthesis. Plant Cell 2013, 25, 3089–3103. [CrossRef][PubMed] 61. Miyanaga, A.; Horinouchi, S. Enzymatic synthesis of bis-5-alkylresorcinols by resorcinol-producing type III polyketide synthases. J. Antibiot. 2009, 62, 371–376. [CrossRef][PubMed] 62. Kim, S.S.; Grienenberger, E.; Lallemand, B.; Colpitts, C.C.; Kim, S.Y.; Souza Cde, A.; Geoffroy, P.; Heintz, D.; Krahn, D.; Kaiser, M.; et al. LAP6/POLYKETIDE SYNTHASE A and LAP5/POLYKETIDE SYNTHASE B encode hydroxyalkyl α-pyrone synthases required for pollen development and sporopollenin biosynthesis in Arabidopsis thaliana. Plant Cell 2010, 22, 4045–4066. [CrossRef][PubMed] 63. Colpitts, C.C.; Kim, S.S.; Posehn, S.E.; Jepson, C.; Kim, S.Y.; Wiedemann, G.; Reski, R.; Wee, A.G.; Douglas, C.J.; Suh, D.Y. Ppascl, a moss ortholog of anther-specific chalcone synthase-like enzymes, is a hydroxyalkylpyrone synthase involved in an evolutionarily conserved sporopollenin biosynthesis pathway. New Phytol. 2011, 192, 855–868. [CrossRef][PubMed] 64. Funa, N.; Ohnishi, Y.; Ebizuka, Y.; Horinouchi, S. Properties and substrate specificity of rppa, a chalcone synthase-related polyketide synthase in Streptomyces griseus. J. Biol. Chem. 2002, 277, 4628–4635. [CrossRef] [PubMed] 65. Izumikawa, M.; Shipley, P.R.; Hopke, J.N.; O’Hare, T.; Xiang, L.; Noel, J.P.; Moore, B.S. Expression and characterization of the type III polyketide synthase 1,3,6,8-tetrahydroxynaphthalene synthase from Streptomyces coelicolor A3(2). J. Ind. Microbiol. Biotechnol. 2003, 30, 510–515. [CrossRef][PubMed] 66. Austin, M.B.; Izumikawa, M.; Bowman, M.E.; Udwary, D.W.; Ferrer, J.L.; Moore, B.S.; Noel, J.P. Crystal structure of a bacterial type III polyketide synthase and enzymatic control of reactive polyketide intermediates. J. Biol. Chem. 2004, 279, 45162–45174. [CrossRef][PubMed] 67. Achkar, J.; Xian, M.; Zhao, H.; Frost, J.W. Biosynthesis of phloroglucinol. J. Am. Chem. Soc. 2005, 127, 5332–5333. [CrossRef][PubMed] 68. Zha, W.; Rubin-Pitel, S.B.; Zhao, H. Characterization of the substrate specificity of phld, a type III polyketide synthase from Pseudomonas fluorescens. J. Biol. Chem. 2006, 281, 32036–32047. [CrossRef][PubMed] 69. Tseng, C.C.; McLoughlin, S.M.; Kelleher, N.L.; Walsh, C.T. Role of the active site cysteine of dpga, a bacterial type III polyketide synthase. Biochemistry 2004, 43, 970–980. [CrossRef][PubMed] 70. Wu, H.-C.; Li, Y.-S.; Liu, Y.-C.; Lyu, S.-Y.; Wu, C.-J.; Li, T.-L. Chain elongation and cyclization in type III pks dpga. ChemBioChem Eur. J. Chem. Biol. 2012, 13, 862–871. [CrossRef][PubMed] 71. Saxena, P.; Yadav, G.; Mohanty, D.; Gokhale, R.S. A new family of type III polyketide synthases in Mycobacterium tuberculosis. J. Biol. Chem. 2003, 278, 44780–44790. [CrossRef][PubMed] 72. Sankaranarayanan, R.; Saxena, P.; Marathe, U.B.; Gokhale, R.S.; Shanmugam, V.M.; Rukmini, R. A novel tunnel in mycobacterial type III polyketide synthase reveals the structural basis for generating diverse metabolites. Nat. Struct. Mol. Biol. 2004, 11, 894–900. [CrossRef][PubMed] 73. Gokulan, K.; O’Leary, S.E.; Russell, W.K.; Russell, D.H.; Lalgondar, M.; Begley, T.P.; Ioerger, T.R.; Sacchettini, J.C. Crystal structure of Mycobacterium tuberculosis polyketide synthase 11 (pks11) reveals intermediates in the synthesis of methyl-branched alkylpyrones. J. Biol. Chem. 2013, 288, 16484–16494. [CrossRef][PubMed] 74. Song, L.; Barona-Gomez, F.; Corre, C.; Xiang, L.; Udwary, D.W.; Austin, M.B.; Noel, J.P.; Moore, B.S.; Challis, G.L. Type III polyketide synthase beta-ketoacyl-acp starter unit and ethylmalonyl-coa extender unit selectivity discovered by Streptomyces coelicolor genome mining. J. Am. Chem. Soc. 2006, 128, 14754–14755. [CrossRef][PubMed] 75. Gruschow, S.; Buchholz, T.J.; Seufert, W.; Dordick, J.S.; Sherman, D.H. Substrate profile analysis and acp-mediated acyl transfer in Streptomyces coelicolor type III polyketide synthases. ChemBioChem Eur. J. Chem. Biol. 2007, 8, 863–868. [CrossRef][PubMed] Molecules 2016, 21, 806 36 of 37

76. Chemler, J.A.; Buchholz, T.J.; Geders, T.W.; Akey, D.L.; Rath, C.M.; Chlipala, G.E.; Smith, J.L.; Sherman, D.H. Biochemical and structural characterization of germicidin synthase: Analysis of a type III polyketide synthase that employs acyl-acp as a starter unit donor. J. Am. Chem. Soc. 2012, 134, 7359–7366. [CrossRef][PubMed] 77. Jeya, M.; Kim, T.S.; Tiwari, M.K.; Li, J.; Zhao, H.; Lee, J.K. A type III polyketide synthase from Rhizobium etli condenses malonyl coas to a heptaketide pyrone with unusually high catalytic efficiency. Mol. BioSyst. 2012, 8, 3103–3106. [CrossRef][PubMed] 78. Funa, N.; Ozawa, H.; Hirata, A.; Horinouchi, S. Phenolic lipid synthesis by type III polyketide synthases is essential for cyst formation in Azotobacter vinelandii. Proc. Natl. Acad. Sci. USA 2006, 103, 6356–6361. [CrossRef][PubMed] 79. Miyanaga, A.; Funa, N.; Awakawa, T.; Horinouchi, S. Direct transfer of starter substrates from type I fatty acid synthase to type III polyketide synthases in phenolic lipid synthesis. Proc. Natl. Acad. Sci. USA 2008, 105, 871–876. [CrossRef][PubMed] 80. Satou, R.; Miyanaga, A.; Ozawa, H.; Funa, N.; Katsuyama, Y.; Miyazono, K.; Tanokura, M.; Ohnishi, Y.; Horinouchi, S. Structural basis for cyclization specificity of two azotobacter type III polyketide synthases: A single amino acid substitution reverses their cyclization specificity. J. Biol. Chem. 2013, 288, 34146–34157. [CrossRef][PubMed] 81. Awakawa, T.; Sugai, Y.; Otsutomo, K.; Ren, S.; Masuda, S.; Katsuyama, Y.; Horinouchi, S.; Ohnishi, Y. 4-hydroxy-3-methyl-6-(1-methyl-2-oxoalkyl)pyran-2-one synthesis by a type III polyketide synthase from Rhodospirillum centenum. ChemBioChem Eur. J. Chem. Biol. 2013, 14, 1006–1013. [CrossRef][PubMed] 82. Muggia, L.; Grube, M. Type III polyketide synthases in lichen mycobionts. Fungal Biol. 2010, 114, 379–385. [PubMed] 83. Goyal, A.; Saxena, P.; Rahman, A.; Singh, P.K.; Kasbekar, D.P.; Gokhale, R.S.; Sankaranarayanan, R. Structural insights into biosynthesis of resorcinolic lipids by a type III polyketide synthase in Neurospora crassa. J. Struct. Biol. 2008, 162, 411–421. [CrossRef][PubMed] 84. Rubin-Pitel, S.B.; Zhang, H.; Vu, T.; Brunzelle, J.S.; Zhao, H.; Nair, S.K. Distinct structural elements dictate the specificity of the type III pentaketide synthase from Neurospora crassa. Chem. Biol. 2008, 15, 1079–1090. [CrossRef][PubMed] 85. Hashimoto, M.; Nonaka, T.; Fujii, I. Fungal type III polyketide synthases. Nat. Prod. Rep. 2014, 31, 1306–1317. [CrossRef][PubMed] 86. Hashimoto, M.; Koen, T.; Takahashi, H.; Suda, C.; Kitamoto, K.; Fujii, I. Aspergillus oryzae csyb catalyzes the condensation of two beta-ketoacyl-coas to form 3-acetyl-4-hydroxy-6-alkyl-alpha-pyrone. J. Biol. Chem. 2014, 289, 19976–19984. [CrossRef][PubMed] 87. Jeya, M.; Kim, T.S.; Tiwari, M.K.; Li, J.; Zhao, H.; Lee, J.K. The Botrytis cinerea type III polyketide synthase shows unprecedented high catalytic efficiency toward long chain acyl-coas. Mol. BioSyst. 2012, 8, 2864–2867. [CrossRef][PubMed] 88. Yokoigawa, J.; Morimoto, K.; Shiono, Y.; Uesugi, S.; Kimura, K.I.; Kataoka, T. Allantopyrone a, an α-pyrone metabolite from an endophytic fungus, inhibits the tumor necrosis factor α-induced nuclear factor kappab signaling pathway. J. Antibiot. 2015, 68, 71–75. [CrossRef][PubMed] 89. Zhou, X.M.; Zheng, C.J.; Song, X.P.; Han, C.R.; Chen, W.H.; Chen, G.Y. Antibacterial α-pyrone derivatives from a mangrove-derived fungus Stemphylium sp. 33231 from the south China sea. J. Antibiot. 2014, 67, 401–403. [CrossRef][PubMed] 90. Abe, I.; Takahashi, Y.; Lou, W.; Noguchi, H. Enzymatic formation of unnatural novel polyketides from alternate starter and nonphysiological extension substrate by chalcone synthase. Org. Lett. 2003, 5, 1277–1280. [CrossRef][PubMed] 91. Samappito, S.; Page, J.; Schmidt, J.; De-Eknamkul, W.; Kutchan, T.M. Molecular characterization of root-specific chalcone synthases from Cassia alata. Planta 2002, 216, 64–71. [CrossRef][PubMed] 92. Morita, H.; Noguchi, H.; Schroder, J.; Abe, I. Novel polyketides synthesized with a higher plant stilbene synthase. Eur. J. Biochem. 2001, 268, 3759–3766. [CrossRef][PubMed] 93. Li, T.L.; Spiteller, D.; Spencer, J.B. Identification of a pentaketide stilbene produced by a type III polyketide synthase from pinus sylvestris and characterisation of free coenzyme a intermediates. ChemBioChem Eur. J. Chem. Biol. 2009, 10, 896–901. [CrossRef][PubMed] Molecules 2016, 21, 806 37 of 37

94. Shi, S.-P.; Wanibuchi, K.; Morita, H.; Endo, K.; Noguchi, H.; Abe, I. Enzymatic formation of unnatural novel chalcone, stilbene, and benzophenone scaffolds by plant type III polyketide synthase. Org. Lett. 2009, 11, 551–554. [CrossRef][PubMed] 95. Karppinen, K.; Hokkanen, J.; Mattila, S.; Neubauer, P.; Hohtola, A. Octaketide-producing type III polyketide synthase from hypericum perforatum is expressed in dark glands accumulating hypericins. FEBS J. 2008, 275, 4329–4342. [CrossRef][PubMed] 96. Abe, I.; Abe, T.; Wanibuchi, K.; Noguchi, H. Enzymatic formation of quinolone alkaloids by a plant type III polyketide synthase. Org. Lett. 2006, 8, 6063–6065. [CrossRef][PubMed] 97. Wakimoto, T.; Mori, T.; Morita, H.; Abe, I. Cytotoxic tetramic acid derivative produced by a plant type-III polyketide synthase. J. Am. Chem. Soc. 2011, 133, 4746–4749. [CrossRef][PubMed] 98. Go, M.K.; Chow, J.Y.; Cheung, V.W.; Lim, Y.P.; Yew, W.S. Establishing a toolkit for precursor-directed polyketide biosynthesis: Exploring substrate promiscuities of acid-coa ligases. Biochemistry 2012, 51, 4568–4579. [CrossRef][PubMed] 99. Klundt, T.; Bocola, M.; Lutge, M.; Beuerle, T.; Liu, B.; Beerhues, L. A single amino acid substitution converts benzophenone synthase into phenylpyrone synthase. J. Biol. Chem. 2009, 284, 30957–30964. [CrossRef] [PubMed] 100. Jez, J.M.; Bowman, M.E.; Noel, J.P. Expanding the biosynthetic repertoire of plant type III polyketide synthases by altering starter molecule speciﬁcity. Proc. Natl. Acad. Sci. USA 2002, 99, 5319–5324. [CrossRef] [PubMed] 101. Abe, I.; Watanabe, T.; Morita, H.; Kohno, T.; Noguchi, H. Engineered biosynthesis of plant polyketides—Manipulation of chalcone synthase. Org. Lett. 2006, 8, 499–502. [CrossRef][PubMed] 102. Abe, I.; Morita, H.; Oguro, S.; Noma, H.; Wanibuchi, K.; Kawahara, N.; Goda, Y.; Noguchi, H.; Kohno, T. Structure-based engineering of a plant type III polyketide synthase: Formation of an unnatural nonaketide naphthopyrone. J. Am. Chem. Soc. 2007, 129, 5976–5980. [CrossRef][PubMed] 103. Abe, I. Engineered biosynthesis of plant polyketides: Structure-based and precursor-directed approach. Top. Curr. Chem. 2010, 297, 45–66. [PubMed] 104. Wanibuchi, K.; Morita, H.; Noguchi, H.; Abe, I. Enzymatic formation of an aromatic dodecaketide by engineered plant polyketide synthase. Bioorg. Med. Chem. Lett. 2011, 21, 2083–2086. [CrossRef][PubMed] 105. Morita, H.; Yamashita, M.; Shi, S.P.; Wakimoto, T.; Kondo, S.; Kato, R.; Sugio, S.; Kohno, T.; Abe, I. Synthesis of unnatural alkaloid scaffolds by exploiting plant polyketide synthase. Proc. Natl. Acad. Sci. USA 2011, 108, 13504–13509. [CrossRef][PubMed] 106. Funa, N.; Ohnishi, Y.; Ebizuka, Y.; Horinouchi, S. Alteration of reaction and substrate speciﬁcity of a bacterial type III polyketide synthase by site-directed mutagenesis. Biochem. J. 2002, 367, 781–789. [CrossRef] [PubMed] 107. Li, S.; Gruschow, S.; Dordick, J.S.; Sherman, D.H. Molecular analysis of the role of tyrosine 224 in the active site of Streptomyces coelicolor rppa, a bacterial type III polyketide synthase. J. Biol. Chem. 2007, 282, 12765–12772. [CrossRef][PubMed]

© 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).