Polyketide Biosynthesis, Modular Polyketide Synthases
Tonia J. Buchholz, Jeffrey D. Kittendorf and David H. Sherman, University of Michigan, Ann Arbor, Michigan
Polyketides constitute a large class of microbial and plant-derived secondary metabolites that displays a vast array of structural diversity. These organic molecules vary in molecular weight, functional group modification, and include linear, polycyclic, and macrocyclic structural forms. Currently, polyketide natural products find clinical use as antibiotics, antiparasitic agents, antifungals, anticancer drugs, and immunosuppressants. Given these impressive and wide-ranging pharmacologic activities, an ever-increasing demand is placed on natural products research to uncover novel polyketide metabolites for the benefit of human and animal health. Modular polyketide synthases are nature's platform for the expansion of chemical diversity. This review provides new perspectives on important biosynthetic mechanisms that contribute to this variety. This includes control of double-bond configuration and regiochemistry, introduction of β-branching during polyketide chain assembly, and other processes that contribute to introduction of unique chemical functionality into these fascinating systems.
Despite the promise of modern synthetic technologies to enhance pharmaceutical discovery significantly, natural products continue to be the greatest source of all new drug leads (1). Currently, many examples of natural product-derived pharmaceuticals are employed to benefit human health (2, 3); polyketides constitute a large class of microbial and plant-derived secondary metabolites that displays a vast array of structural diversity. These organic molecules vary in molecular weight and functional group modification; they include linear, polycyclic, and macrocyclic structural forms. Currently, polyketide natural products find clinical use as antibiotics, antiparasitic agents, antifungals, anticancer drugs, and immunosuppressants (Fig. 1a). Given these impressive and wide-ranging pharmacological activities, an ever-increasing demand is placed on natural products research to uncover novel polyketide metabolites for the benefit of human health.
Although the clinical use of polyketide-inspired pharmaceuticals has been appreciated for decades, polyketide-derived metabolites have been recognized recently for their role in bacterial virulence. For example, the pathogenesis of Mycobacterium ulcerans, the causative agent of the devastating skin disease known as Buruli ulcer, is the result of the secretion of polyketide-derived toxins known as the mycolactones (Fig. 1b) (4). These polyketide toxins are responsible largely for the necrotic lesions that are characteristic of this debilitating condition. As such, the disruption of mycolactone biosynthesis may lead to an effective chemotherapy for Buruli ulcer. Recent findings also suggest that the virulence of another mycobacterial species, Mycobacterium tuberculosis, could be partially dependent on polyketide biosynthesis. The cell surface sulfolipid-1 (SL-1) is among several virulence-associated molecules produced by M. tuberculosis. SL-1 consists of a sulfated disaccharide core (trehalose-2-sulfate) that displays four lipidic substituents; all but one substituent seems to be polyketide-derived (Fig. 1b) (5). Finally, the toxic agent in rice seedling blight, which is a highly destructive fungal disease that inflicts severe agricultural losses worldwide, has been identified recently as the polyketide metabolite, rhizoxin (Fig. 1b) (6). Interestingly, rhizoxin is not produced directly by the fungus (Rhizopus), but rather by the endosymbiotic bacteria Burkholderia that thrives within the fungus. Together, these three examples suggest that inhibition of polyketide biosynthesis may lead to effective chemotherapy for controlling certain human and plant bacterial diseases.
Figure 1. Examples of clinically-relevant polyketide natural product drugs (a) and virulence factors (b).
Prototypical Polyketide Biosynthesis
The biosynthesis of many important polyketide compounds occurs via a stepwise, assembly-line type mechanism that is catalyzed by type I modular polyketide synthases (PKSs). These modular PKSs are composed of several large, multifunctional enzymes that are responsible for catalyzing the initiation, elongation, and processing steps that ultimately give rise to the characteristic macrolactone scaffold (Fig. 2) (7-11). Structural studies have been critical in developing a sophisticated understanding of the overall architecture and mechanism of type I PKSs and their homologs in recent years (12-18). A review from the perspective of the 6-deoxyerythronolide B synthase (a well-studied type I PKS) was published recently by Khosla et al. (19).
It is well established that the sequential arrangement of modules within a PKS system serves effectively as a biosynthetic program, which is responsible for dictating the final size and structure of the polyketide core. Typically, initiation of polyketide biosynthesis begins by the acyltransferase (AT) catalyzed linkage of a coenzyme A (CoA) priming unit (e.g., methylmalonyl-CoA, malonyl-CoA, propionyl-CoA) to the acyl carrier protein (ACP) of the loading module. Once initiated, downstream elongation modules carry out repetitive extensions of the starter unit. In most PKS systems, each elongation module contains at minimum an AT domain, an ACP domain, and a ketosynthase (KS) domain (Fig. 2a). The AT domain is responsible for loading the appropriate CoA extender unit onto the ACP domain (i.e., malonyl-CoA, methylmalonyl-CoA, etc.). The KS domain then, catalyzes a decarboxylative condensation of the extender unit with the growing polyketide chain obtained from the preceding module to generate an ACP-bound β-ketoacyl product. In addition to the three core domains, each elongation module may contain up to three additional domains [ketoreductase (KR), dehydratase (DH), enoyl reductase (ER)] that are responsible for the reductive processing of the β-keto functionality prior to the next extension step (Fig. 2a). These reductive steps contribute greatly to the overall structural diversity that is observed among polyketide natural products. The presence of a KR domain alone generates a β-hydroxyl functionality, the presence of both a KR and a DH domain generates an alkene, whereas the combination of KR, DH, and ER results in complete reduction to the alkane. Finally, termination of polyketide biosynthesis is catalyzed by a thioesterase (TE) domain located at the carboxy terminus of the final elongation module. The activity of this domain results in the cleavage of the acyl chain from the adjacent ACP; typically, intramolecular cyclization results in the formation and release of a macrolactone ring. Tailoring enzymes, such as hydroxylases and glycosyl transferases, often serve to further modify the polyketide to yield the final bioactive compound.
The modular organization of type I PKSs has made them particularly attractive targets for rational bioengineering. Combinatorial biosynthetic efforts centered on prototypical modular PKSs have been the topic of many recent outstanding review articles (20-23). Currently, several strategies are being pursued that attempt to leverage PKS systems for the generation of structurally diverse polyketides. For example, it has been demonstrated that alterations of individual catalytic domains (i.e., inactivation, substitution, addition, deletion) within a PKS module can result in predicted structural alterations of the final PKS product. Likewise, the addition, deletion, or exchange of intact modules can also impart structural variety into polyketide metabolites. Using these and other approaches, hundreds of novel polyketide structures have been generated, which established the tremendous potential of these applications. However, these successes seem to be more the exception rather than the rule, as many efforts result in trace levels, or they fail to provide the desired metabolite. This finding suggests that much remains to be learned regarding the molecular intricacies of these complex biosynthetic machines. This review provides new perspectives on important mechanisms that contribute to structural diversity in modular PKSs. These mechanisms include control of double-bond configuration and regiochemistry, introduction of β-branching during polyketide chain assembly, and other processes that contribute to introduction of unique chemical functionality into these fascinating systems.
Figure 2. Conventional modular type I PKS paradigm. (a) Individual domains in a full type I polyketide synthase extension module. Homodimeric contacts are made in the N-terminal docking, ketosynthase, dehydratase, enoyl reductase, and C-terminal docking domains. (b) PKS system for 10-deoxymethynolide and narbonolide generation.
The presence of unsaturated carbon-carbon bonds within most polyketide compounds exemplifies the overall structural diversity that is a hallmark of this class of important natural products. Typically, the installation of double bonds into nascent polyketide chains relies on the two-step processing at the β-keto group by the successive activity of KR and DH domains that are embedded within a given PKS elongation module. After KS catalyzed chain elongation that extends the growing chain by two carbon atoms, the KR domain, when present, directs the NADPH-dependent reduction of the β-ketone to yield a 3-hydroxyacyl intermediate. Subsequently, an embedded DH domain within the elongation module catalyzes dehydration of the 3-hydroxyacyl intermediate, normally which results in the incorporation of an (E)-trans α,β unsaturated bond into the growing polyketide chain (Fig. 3a).
It should be noted that the KR catalyzed reduction of a β-ketoacyl intermediate has stereochemical consequences because a new chiral center is introduced into the growing oligoketide. Aside from serving to enhance the structural diversity of the final polyketide product even more, the stereochemical outcome of this reaction can have profound effects on any subsequent processing or elongation reactions. As such, an understanding of how KR domains exert stereochemical control of their hydroxylated product is a critical aspect to deciphering the mechanism of DH-mediated double bond formation. Through bioinformatic and biochemical analyses, an appreciation of ketoreductase-influenced stereochemistry has emerged (24, 25). Thus, Caffrey (25) has proposed that KR domains can be divided into two classes, depending on the final configuration of the P-hydroxyl moiety. The so-called “A” class generates an L-3-hydroxy product, whereas the “B” class produces the D-3-hydroxy polyketide intermediate. Although little difference exists between these two putative classes at the amino acid sequence level, the presence of a conserved aspartate residue within an LDD motif correlates well with “B” class. This motif is absent in the defined “A” class of KR domains. An additional diagnostic feature of the “A” class of KR domains is the presence of a conserved tryptophan residue. Recently, Keatinge-Clay (18) has proposed a refinement of the KR class descriptions as originally suggested by Caffrey (25), effectively increasing the number of possible KR types from two to six (18). This new classification takes into consideration whether a given KR domain (either reductively competent or incompetent) is located in an epimerization-competent module. Although this new classification offers a more complete description of PKS KR domains, for simplicity we will continue to use Caffrey’s KR nomenclature throughout our discussion of double-bond formation.
While examples of both D- and L-hydroxyl group configurations can be found within polyketide natural products, recent evidence suggests that DH domains require a stereospecific 3-hydroxyacyl intermediate. Bioinformatic analyses performed on 71 KR domains for which the stereochemical outcome of the reduction is cryptic because of subsequent dehydration revealed that all belong to the “B” class of KR domains (25). As such, it appears that the generally preferred substrate for DH domains is a D-3-hydroxyacyl chain. However, direct experimental evidence has been difficult to obtain because the 3-hydroxyacyl intermediate is transient in modules that contain a DH domain. Recently, biochemical studies of the DH domain found in module 2 of the pikromycin PKS system (Fig. 2b) (26) have supported this hypothesis; inactivation of the DH domain resulted in the exclusive generation of the D-3-triketide acylthioester intermediate from a diketide substrate (27). Aside from this study, no other reports probe the substrate preference or catalytic mechanism of DH domains within PKS systems; therefore, much of what is known has been elucidated from studies of fatty acid biosynthesis (28). Previous studies on the dehydration step that is catalyzed by the yeast fatty acid synthase confirmed the syn elimination of water from a D-(3 R)-hydroxyacylthioester substrate (29). This result is consistent with the stereospecificity of the PKS DH domain and may suggest that trans unsaturated bonds, typically found in polyketides, are likewise formed via syn water elimination.
Figure 3. (a) Traditional view of reductive processing at the β-ketone position in the growing polyketide chain. Presence of a ketoreductase domain leads to formation of an alcohol. An active dehydratase domain can further process the alcohol moiety to an alkene. Complete saturation to the alkane is accomplished by an enoyl reductase domain. (b) Proposed terminal double bond formation for curacin biosynthesis. (c) Proposed terminal double bond formation in tautomycetin biosynthesis.
Cis double bonds
Although rare, several PKS biosynthetic systems can install cis double bonds into the final polyketide product (Fig. 4). Several possible mechanisms could account for the infrequent occurrence of this double bond configuration. One explanation is that an isomerization event occurs that converts a trans double bond into a cis double bond. This isomerization activity could be specified by the PKS elongation module, much like previously identified epimerization activities that are known to exist in some PKS and NRPS modules. Alternatively, the combined activity of KR-DH domains within certain modules could directly establish the cis double bond. Finally, it is possible that after reduction of the β-keto functionality, a trans acting DH could catalyze dehydration to form a (Z)-cis double bond. This trans activity might derive from a discrete enzyme encoded within the biosynthetic gene cluster or from an adjacent module within the PKS pathway. Additionally, examples exist in which some general rules may not hold true. Chivosazol, which is a potential antitumor agent, is one such example (30).
Figure 4. Double bond containing compounds discussed in this review. Cis-double bonds discussed in the text are boxed and modifications produced by the HMGS cassettes are shaded.
The epothilones 6 (and Fig. 1a), which are produced by Sorangium cellulosum, are mixed NRPS-polyketide derived natural products that possess potent antitumor activity. Interestingly, some compounds feature a cis double bond between carbon atoms 12 and 13 that should be generated by PKS elongation module 4; however, sequence analysis of the epothilone biosynthetic gene cluster indicates that module 4 does not contain a DH domain requisite for the formation of the unsaturated bond (31). Thus, Tang et al. (31) hypothesized that the DH activity might occur from the subsequent module or by the action of a post-PKS modifying enzyme. Biochemical experiments later demonstrated that the DH domain of module 5 catalyzed the cis double bond formation (32). To account for this atypical activity, it is proposed that the 3-hydroxythioester intermediate undergoes an ACP4-to-ACP5 transfer (32). After dehydration, the thioester intermediate would then be transferred to the KS5 domain for subsequent elongation. Similarly, the antitumor phoslactomycin compounds 13 feature three cis double bonds, two of which seem to be installed by a KR-DH pair (see below); however, the elongation module (Plm7) that should be responsible for generating the unsaturated bond between carbon atoms 2 and 3 does not appear to encode the required DH activity (33). Thus, it is likely that the source of this catalytic activity comes from either a different module within the PKS system or a separate enzyme that could act either before or after TE mediated termination of polyketide biosynthesis. Additional analysis is required to discriminate between these two possibilities.
Unlike the examples described above, most polyketide cis double bonds are installed through a successive KR-DH pair found embedded within the elongation module. In these cases, the stereochemistry of the 3-hydroxyacyl intermediate appears to be the discriminating factor between (Z)-cis or (E)-trans unsaturation. For example, the antitumor disorazole compounds 5 display up to three cis double bonds per monomer (note: final compound is a condensed dimer). Sequence analyses of KR domains preceding DH domains that would be responsible for cis double bond formation suggest that they all belong to the “A” class (34). Thus, they are predicted to generate a L-3-hydroxyacyl intermediate. It is expected that the subsequent DH domain preferentially recognizes the L-3-hydroxyl group to facilitate the generation of the cis double bond. Interestingly, the module responsible for incorporating the cis unsaturated bond between carbon atoms 11 and 12 is split between two polypeptide chains (34). It cannot be ruled out that this modular dissection may play a role in formation of this particular double bond, as the cleavage point occurs between the DH and KR domain. Furthermore, it is intriguing that the major product, disorazole A1, is composed of two nonidentical monomers that differ in saturation between carbon atoms 5 and 6. It has been suggested that the synthesis of the two different monomers is caused by poor activity of the DH domain (34); however, final proof requires additional experimental verification.
In addition to disorazole 5, several other known examples of polyketide natural products exist that have cis double bonds installed by an embedded KR-DH domain pair. The potent antitumor compound curacin A 3 contains many interesting structural features. Among them is the presence of a cis double bond between carbon atoms 3 and 4. Sequence alignment of the KR domain encoded by curG (encoding the module responsible for generation of the cis double bond) suggests that it belongs to the “A” class of KR domains (35). Thus, this particular KR is predicted to generate a L-3-hydroxyacyl intermediate that is subsequently dehydrated to the cis double bond. Likewise, the KR domains that set up the 3 cis double bonds found in the linear mixed NRPS-polyketide natural product bacillaene 1 are predicted to produce L-3-hydroxyacyl intermediates (36). Interestingly, two of the elongation modules that incorporate cis olefins are split between two polypeptides (36). As in disorazole biosynthesis, it is possible that these modular dissections contribute to the configuration of the unsaturated bond that is introduced by these modules.
Finally, we consider the conjugated cis olefins that span carbon atoms 12-15 of the phoslactomycins 13. Analysis of the KR sequences of elongation modules 1 and 2 could not clearly predict whether these reductive domains belonged to the “A” or “B” class. Therefore, Alhamadsheh et al. (37) employed a comprehensive biochemical study to elucidate the mechanism of cis olefin formation by the first elongation module. Two hypotheses were considered. First, the configuration of the double bond could develop directly from the combined activities of the embedded KR-DH domain pair. Alternatively, the KR-DH domains might establish a trans olefin that is isomerized subsequently to the observed cis configuration. To distinguish between these two possibilities, Alhamadsheh et al. (37) genetically inactivated both the loading module and elongation module of Plm1 and conducted feeding experiments with diketide analogs containing both cis and trans olefins. Results from this work indicated clearly that only the cis olefin containing diketide is accepted as a substrate for elongation module 2, suggesting that the product of module 1 must contain the cis double bond. Furthermore, this work demonstrated nicely that the phoslactomycin biosynthetic pathway cannot process trans diketide intermediates into mature products, which rules out the possibility of an isomerization domain in downstream modules.
Terminal double bonds
Termination of polyketide biosynthesis typically involves the TE mediated cleavage of the ACP-bound thioester, followed by cyclization to generate a macrolactone. Alternatively, the TE catalyzes the simple hydrolysis of the thioester to generate a linear free acid product. Here, we consider two of the relatively few known examples of polyketide natural products that are neither a macrocycle nor a free acid, but instead terminate with a double bond.
Aside from containing a cis double bond noted above, the antitumor polyketide compound curacin A 3 also features a terminal olefin. Previously reported feeding studies suggested that the formation of the terminal double bond develops from successive decarboxylation and dehydration events (35). The biosynthetic gene cluster responsible for curacin A biosynthesis has been identified and initially characterized, which enables the putative assignment of domains within the predicted elongation modules (35). Like most other known polyketide biosynthetic pathways, the final elongation module of the curacin pathway, CurM, contains a terminal thioesterase domain that presumably plays a role in formation of the terminal olefin; however, biochemical evidence for such a role is lacking. Interestingly, domain analysis of CurM also predicts the presence of a sulfotransferase (ST) domain immediately preceding the TE domain. Typically, ST domains are responsible for transferring a sulfuryl group from a donor molecule (such as 3'-phosphoadenosine-5'-phosphosulfate, PAPS) to a variety of acceptor carbohydrates, proteins and other low-molecular weight metabolites (38). Although STs have been characterized Current efforts in our laboratory include elucidating the roles of the ST and TE domains in curacin A biosynthesis, and in particular, their potential functions in terminal olefin formation. One possible mechanism that is currently under consideration is shown in Fig. 3b. On reduction of the β-keto functionality of the ACP-bound thioester intermediate, the ST domain transfers a sulfuryl group from the donor molecule PAPS to the 3-hydroxyl group of the thioester chain. Consistent with this hypothesis, bioinformatic analysis suggests that the putative curacin ST domain contains the signature PAPS binding pocket (unpublished data); however, no experimental evidence suggests that this ST domain can catalyze the sulfuryl transfer or that the ST domain can bind the requisite PAPS donor molecule. Assuming our hypothesis is correct and that the ST domain functions as proposed, transfer of the sulfurylated intermediate to the TE domain would initiate hydrolytic termination of curacin A biosynthesis to produce the linear free acid. At this point, one of several chemical steps can be envisioned. Following hydrolysis, the TE may catalyze decarboxylation, after which the formation of the double bond would occur in a concerted process by displacement of the sulfate leaving group. Alternatively, a separately encoded enzyme might be responsible for decarboxylating the free acid generated by the TE domain. It is also conceivable that on TE catalyzed hydrolysis, the decarboxylation reaction occurs spontaneously due to the presence of the sulfate leaving group at carbon 3.
Similar to curacin, the polyketide metabolite tautomycetin 14 also possesses a terminal olefin. This polyketide metabolite has potential medicinal value because of its novel immunosuppressive activities (39). The tautomycetin biosynthetic gene cluster has been sequenced recently, which enables domain composition analysis of the terminal elongation module (40). Unlike the terminal module involved in curacin biosynthesis, the final tautomycetin elongation module does not contain the unusual ST domain. This finding may suggest that formation of the terminal olefin of tautomycetin 14 occurs by a different chemical mechanism. The final elongation module does contain a TE domain, which presumably terminates tautomycetin biosynthesis through generation of the free acid (Fig. 3c). As described for curacin, it is possible that this TE domain can also catalyze the subsequent decarboxylation event; however, in lieu of an activated leaving group at carbon 3, it is reasonable to expect that dehydration to remove the hydroxyl at carbon 3 would also be a catalyzed event. Alternatively, the terminal olefin could be installed during the post-PKS maturation of the polyketide to the final tautomycetin product. DNA sequence analysis of open reading frames that are downstream of the tautomycetin PKS gene cluster reveals two potential candidates, tmcJ and tmcM, which may be involved in double bond formation (40). Bioinformatic analyses suggest that tmcJ might encode for a putative decarboxylase and that the gene product of tmcM is a potential dehydratase.
Atypical PKS Domains
The wide distribution of PKSs in the microbial world and the extreme chemical diversity of their products do in fact result from a varied use of the well-known catalytic domains described above for the canonical PKS systems. Taking a theoretic view of polyketide diversity, Gonzalez-Lergier et al. (41) have suggested that even if the starter and extender units are fixed, over 100,000 linear heptaketide structures are possible using only the 5 common reductive outcomes at the P-carbon position (ketone, (R- or S-) alcohol, trans-double bond, or alkane). Recently, it has become apparent that even this does not represent the upper limit for polyketide diversification. To create chemical functionalities beyond those mentioned above, nature has recruited some enzymes from sources other than fatty acid synthesis (the mevalonate pathway in primary metabolism is one example) not typically thought of as type I PKS domains. Next, we explore the ways PKS-containing systems have modified these domains for the catalysis of some unique chemistries observed in natural products.
Methyl groups at the α- and β-carbons
As described above for polyketide biosynthesis, the presence or absence of a methyl group on the a-carbon position of the growing polyketide chain is most often governed by the selection of the extender unit (malonyl-CoA versus methylmalonyl-CoA). However, in PKS systems that use trans acyltransferases (AT-less type I PKSs) (8, 42), the module by module control over extender unit selection is sometimes not possible. In most cases, malonyl-CoA is used as the extender unit, and a methyl group can be added to selected positions through the action of an embedded methyl transferase (MT) domain or discrete MT enzyme. For example, the C-6 methyl group of leinamycin 8 is thought to be installed by the MT embedded in LnmJ (43), the C-10 methyl of curacin A 3 is generated via the MT domain in CurJ (35), and the gem dimethyl groups on C-8 and C-18 of bryostatin 2, most likely are the consequence of the MT domains in BryB and BryC (44).
In contrast to the α-carbon methylations, the incorporation of methyl or methylene groups (or functional groups derived from such groups) at the β-position represents the assimilation of a full cassette of enzymes into the typical PKS machinery. Recently, a subset of type I modular PKSs (and hybrid NRPS/PKS megasynthases) have been identified that contain multiple enzymes acting in trans during the traditional linear assembly-line process to accomplish β-branching. Termed HMG-CoA synthase (HMGS) cassettes, these enzyme systems provide a unique method of expanding the repertoire of the traditional reductive domains (KR, DH, ER). These enzymes work in conjunction with the PKS machinery to create unique functionalities observed at the branch points that include the pendant methyl groups of bacillaene 1 (36, 45, 46), mupirocin 9 (47), and virginiamycin M 15 (48), which are the methoxymethyl and ethyl groups of myxovirescin A 10 (49, 50); the exomethylene groups of difficidin 4 (45), onnamide A 11 (51), and pederin 12 (51-53); the cyclopropyl ring of curacin A 3 (35, 54); the vinyl chloride of jamaicamide 7 (55); the unique 1,3-dioxo-1,2-dithiolane moiety of leinamycin 8 (43); and the exocyclic olefins in bryostatin 2 (Fig. 4) (44).
HMG-CoA synthase cassettes
In primary metabolism, HMG-CoA synthase (HMGS) is responsible for the condensation of C-2 of acetyl-CoA onto the β-ketone of acetoacetyl-CoA to form 3-hydroxyl-3-methylglutaryl-CoA and free CoASH (Fig. 5a) (56). Several secondary metabolite pathways have been identified over the past five years that perform an analogous reaction, although they seem to use ACP-tethered acyl groups (Fig. 5b) as opposed to acyl-CoA substrates. After generation of the HMG-ACP analog on the growing polyketide chain, the product is usually dehydrated and decarboxylated to yield the branched intermediate. Found in 11 pathways to date (36, 45-55), included in the cassette are a discrete ACP, a decarboxylative KS (active site cysteine is replaced with a serine), an HMGS, and one or two enoyl CoA hydratase-like (ECH) domains. (Table 1, Figs. 5, 6).
Table 1. HMGS containing biosynthetic pathways and their producing organisms
Myxovirescin A (antibiotic TA)
Bacillus subtilis 168/B. amyloiquefaciens
Candidatus Endobugula sertula
Lyngbya majuscule JHB
Myxococcus xanthus DK1622
Symbiont bacterium of Theonella swinhoei
Symbiont bacterium of Paederus fuscipes
TaB & TaE
TaC & TaF
CurF N-terminal domain
JamJ N-terminal domain
Figure 5. HMGS cassette reaction scheme. (a) HMG-CoA synthase (HMGS) reaction from primary metabolism. (b) An HMGS cassette can convert the β-ketone to an alkene (β,γ or γ,δ double bond) with a pendant methyl (or ethyl) group.
Three HMGS-containing cassettes (those in the curacin A, bacillaene, and myxovirescin pathways) have been validated biochemically in the past two years and will serve as the basis for our analysis of the individual components in this complex (46, 54, 57, 58). The mechanistic and structural details for HMG-CoA synthase in primary metabolism have been elucidated for both bacterial and eukaryotic HMGSs (59-63). Although polyketide HMGSs share only 20-30% sequence identity with their primary metabolism homologs (in both prokaryotes and eukaryotes), multiple sequence alignment reveals that the key catalytic residues (Glu/Cys/His) are conserved. As shown in Fig. 5b, the first step in the formation of the HMG-intermediate is the generation of of acetyl-ACP. This step is accomplished through the loading via an AT of malonyl-CoA [or perhaps methylmalonyl-CoA in at the case of TaE (58)]. Then, the decarboxylative KS converts the malonyl-ACP into acetyl-ACP, after which the tethered acetyl group is condensed onto the β-ketone of the polyketide intermediate. Finally, formation of the HMG-analog is completed on addition of water.
Processing of the HMG-intermediate can vary considerably, but typically proceeds via dehydration and decarboxylation catalyzed by two enoyl-CoA hydratase-like domains (Fig. 5b). Based on sequence similarity, the members of the crotonase fold family observed in these HMGS cassettes can be subdivided into two groups, termed ECH1 and ECH2 (54). The successive dehydration and decarboxylation steps are catalyzed by the ECH1 and ECH2 enzymes/domains, respectively. Evidence for the specific function of the curacin ECH1 and ECH2 enzyme pair from the curacin pathway has been demonstrated using a coupled enzyme assay and ESI-FT-ICR MS (54). Using purified ECH1 (CurE) and ECH2 (the N-terminal domain of CurF) overexpressed in E. coli, (S)-HMG-ACP was converted first to 3-methylglutaconyl-ACP then to 3-methylcrotonyl-ACP, which is the gained intermediate for subsequent formation of the cyclopropyl ring. Insights into the mechanism of the CurF ECH2decarboxylation have been gained based on the recent crystal structure of the curacin ECH2 domain (64). Additional in vitro evidence for the function of these enzymes has been generated using proteins from the PksX pathway of Bacillus subtilis (46) and the myxovirescin pathway from Myxococcus xanthus (58). Prior to the identification of bacillaene as the product of the PksX pathway, Calderone et al. (46) and Dorrestein et al. (57) reported the function of several discrete enzymes. Using radioactive biochemical assays together with mass spectrometry, they assigned functional roles to AcpK, PksC, the tandem ACPs in PksL, PksF, PksG, PksH, and PksI. Using the model acceptor ACP, acetoacetyl-ACP, and malonyl-CoA in combination with the above proteins, a ∆2-isoprenyl-S-carrier protein was generated (46). Most recently, a similar in vitro investigation was conducted using the homologous enzymes from the myxovirescin pathway (58). The HMGS cassette reaction sequence proposed above held fast for the myxovirescin pathway, although the generation of the propionyl- or methylmalonyl-S-ACP could not be demonstrated. The authors have suggested that perhaps additional enzymes are yet to be identified to fill these roles to complete the β-ethylation at C16 in 10. Two variations on the HMGS cassette theme already have been identified. In the biosynthesis of bryostatin 2, and leinamycin 8, one or both of the ECH-mediated steps is likely omitted based on the final natural product structures. The details of these deviations have not yet been established.
Recently, in vivo evidence for the function of these HMGS cassettes has come from the Muller lab (49, 50, 65). To date, all HMGS cassette proteins for myxovirescin A (TaB/TaC, TaE/TaF, TaK, TaX, TaY) has been individually deleted, and the impact on the products of the engineered Myxococcus xanthus strains has been analyzed. In addition, analysis of products from ∆taV (the trans-acting AT), ∆taH (a cytochrome P450 that is thought to hydroxylate the HMGS-installed P-methyl group at C12 of 10), and ∆taQ (an O-methyl transferase necessary for completing the transformation to the final methoxy-methyl functionality) strains have provided insights into this complex pathway. Production of myxovirescin A was abolished (or greatly reduced) in all above deletion strains. Appearance of novel myxovirescin analogs (β-methyl vs β-ethyl at C16) in the ∆taE & ∆taF strains appears to be a result of TaB or TaC complementation, which provides direct evidence for TaE/TaF in the formation of the ethyl branch point. However, independent biochemical verification of TaF function has been difficult to obtain (58).
HMGS cassette architecture
Analysis of the placement of the known HMGS cassettes identified to date into their biosynthetic clusters reveals a variety of possible architectures (Fig. 6). For example, the ECH2 decarboxylase exists as a discrete enzyme downstream of the ECH1 dehydratase (mupirocin and others), as an N-terminal domain of a large PKS (curacin and jamaicamide), and as an embedded domain (pederin and onnamide). Although most clusters published to date are mixed PKS/NRPS systems with in trans ATs and tandem ACPs at the site of HMGS modification, exceptions exist for each example (difficidin is PKS only, curacin and jamaicamide contain embedded ATs, and bryostatin and myxovirescin do not contain tandem ACPs at the site of HMGS modification).
As HMGS enzyme cassettes have been identified and functionally characterized only recently, many mechanistic details as well as the key protein-protein interactions needed to orchestrate communication among the polypeptide components remain unclear. Details on how the individual proteins are brought to the correct place in the pathway to perform their functions are still unknown for most pathways. In the case of the PksX/bacillaene pathway, some intriguing microscopy performed on B. subtilis suggests that the bacillaene proteins are clustered into a huge mega-enzyme factory inside the bacterial cell (66). Whether this organization extends (or is limited) to the other members of HMGS cassette containing pathways remains to be observed. Additionally, in some pathways, key enzymes have yet to be identified. Two lingering questions include, 1) Which AT domain loads the discrete ACP in the embedded AT systems typified by the curacin and jamaicamide pathways? and 2) where are the missing domains located in the incomplete cassettes? Despite these remaining issues, the stage is now set for these unique suites of enzymes to be included and applied in the growing metabolic engineering/combinatorial biosynthesis toolbox.
The goal of this review has been to highlight a series of novel systems for creating chemical diversity in polyketide natural product biosynthesis. This review includes the mechanistic basis for introduction of trans or cis double bonds within linear or macrocyclic compounds, or assembly of the rare terminal alkene in select secondary metabolites such as curacin and tautomycetin. Similarly, introduction of methyl groups to create branch points or gem dimethyl functionality can occur by several processes that have been dissected in several systems over the past few years. Finally, one of the most intriguing new methods for introduction of diverse branching functionality involves the HMGS-containing enzymes that are being identified in a growing number of PKS and mixed NRPS-PKS pathways. The rapidly increasing knowledge and mechanistic understanding of these complex metabolic systems will provide growing opportunities to engineer chemical diversity using rational approaches.
Figure 6. HMGS containing biosynthetic pathways. Portions of the PKS and PKS/NRPS pathways where the HMGS and related enzymes are located. Abbreviations: A - Adenylation, ACP - acyl carrier protein, AT - acyltransferase, Cy - cyclization, DH - dehydratase, ER - enoyl reductase, GNAT - GCN5-related N-acetyltransferase, KS - ketosynthase, KR - ketoreductase, MT - methyltransferase, Ox - Oxidase, Oxy - Oxygenase, PCP - peptide carrier protein, PhyH - phytanoyl-CoA dioxygenase, PS - pyrone synthase, TE - thioesterase, ? - unknown function, * - inactive domain.
Research on modular type I polyketide synthases in the Sherman laboratory is supported generously by grants from the National Institutes of Health (GM076477, CA108874, TW007404), the Hans and Ella McCollum Vahlteich Research Fund at the University of Michigan College of Pharmacy, and the H.W. Vahlteich Professorship in Medicinal Chemistry. J.D.K. is supported by an NRSA postdoctoral fellowship (GM075641) from the NIH.
1. Koehn FE, Carter GT. The evolving role of natural products in drug discovery. Nat. Rev. Drug Discov. 2005; 4:206-220.
2. Newman DJ, Cragg GM. Natural products as sources of new drugs over the last 25 years. J. Nat. Prod. 2007; 70:461-477.
3. Newman DJ, Cragg GM, Snader KM. The influence of natural products upon drug discovery. Nat. Prod. Rep. 2000; 17:215-234.
4. Stinear TP, Mve-Obian A, Small PLC, Frigui W, Pryor MJ, Brosch R, Jenkin GA, Johnson PDR, Davies JK, Lee RE, Adusumilli S, Garnier T, Haydock SF, Leadlay PF, Cole AT. Giant plasmid-encoded polyketide synthases produce the macrolide toxin of Mycobacterium ulcerans. Proc. Natl. Acad. Sci. U.S.A. 2004; 101:1345-1349.
5. Kumar P, Schelle MW, Jain M, Lin FL, Petzold CJ, Leavell MD, Leary JA, Cox JS, Bertozzi CR. PapA1 and PapA2 are acyl-transferases essential for the biosynthesis of the Mycobacterium tuberculosis virulence factor Sulfolipid-1. Proc. Natl. Acad. Sci. U.S.A. 2007; 104:11221-11226.
6. Partida-Martinez LP, Hertweck C. A gene cluster encoding rhizoxin biosynthesis in “Burkholderia rhizoxina”, the bacterial endosymbiont of the fungus Rhizopus microsporus. ChemBioChem. 2007; 8:41-45.
7. Hill AM. The biosynthesis, molecular genetics and enzymology of the polyketide-derived metabolites. J. Nat. Prod. 2006; 23:256-320.
8. Shen B. Polyketide biosynthesis beyond the type I, II and III polyketide synthase paradigms. Curr. Opin. Chem. Biol. 2003; 7:285-295.
9. Staunton J, Weissman KJ. Polyketide biosynthesis: a millennium review. Nat. Prod. Rep. 2001; 18:380-416.
10. Reeves CD. The enzymology of combinatorial biosynthesis. Crit. Rev. Biotechnol. 2003; 23:95-147.
11. Sherman DH, Smith JL. Clearing the skies over modular polyketide synthases. ACS Chem. Biol. 2006; 1:505-509.
12. Tang Y, Kim CY, Mathews II, Cane DE, Khosla C. The 2.7 Angstrom crystal structure of a 194-kDa homodimeric fragment of the 6-deoxyerythronolide B synthase. Proc. Natl. Acad. Sci. 2006; 103:11124-11129.
13. Giraldes JW, Akey DL, Kittendorf JD, Sherman DH, Smith JL, Fecik RA. Structural and mechanistic insights into polyketide macrolactonization from polyketide-based affinity labels. Nat. Chem. Biol. 2006; 2:531-536.
14. Akey DL, Kittendorf JD, Giraldes JW, Fecik RA, Sherman DH, Smith JL. Structural basis for macrolactonization by the pikromycin thioesterase. Nat. Chem. Biol. 2006; 2:537-542.
15. Keatinge-Clay AT, Stroud RM. The structure of a ketoreductase determines the organization of the beta-carbon processing enzymes of modular synthases. Structure 2006; 14:737-748.
16. Tsai SC, Lu H, Cane DE, Khosla C, Stroud RM. Insights into channel architecture and substrate specificity from crystal structures of two macrocycle-forming thioesterases of modular polyketide synthases. Biochemistry 2002; 41:12598-12606.
17. Maier T, Jenni S, Ban N. Architecture of mammalian fatty acid synthase at 4.5 A resolution. Science 2006; 311:1258-1262.
18. Keatinge-Clay AT. A tylosin ketoreductase reveals how chirality is determined in polyketides. Chem. Biol. 2007; 14:898-908.
19. Khosla C, Tang Y, Chen AY, Schnarr NA, Cane DE. Structure and mechanism of the 6-deoxyerythronolide B synthase. Annu. Rev. Biochem. 2007; 76:11.11-11.27.
20. Weissman KJ, Leadlay PF. Combinatorial biosynthesis of reduced polyketides. Nat. Rev. Microbiol. 2005; 3:925-936.
21. Kittendorf JD, Sherman DH. Developing tools for engineering hybrid polyketide synthetic pathways. Curr. Opin. Biotechnol. 2006; 17:597-605.
22. Baltz RH. Molecular engineering approaches to peptide, polyketide and other antibiotics. Nat. Biotechnol. 2006; 24:1533-1540.
23. Rodriquez E, McDaniel R. Combinatorial biosynthesis of antimicrobials and other natural products. Curr. Opin. Microbiol. 2001; 4:526-534.
24. Reid R, Piagentini M, Rodriguez E, Ashley G, Viswanathan N, Carney J, Santi DV, Hutchinson CR, McDaniel R. A model of structure and catalysis for ketoreductase domains in modular polyketide synthases. Biochemistry 2003; 42:72-79.
25. Caffrey P. Conserved amino acid residues correlating with ketore- ductase stereospecificity in modular polyketicle synthases. ChemBioChem. 2003; 4:654-657.
26. Xue Y, Zhao L, Liu HW, Sherman DH. A gene cluster for macrolide antibiotic biosynthesis in Streptomyces venezuelae: architecture of metabolic diversity. Proc. Natl. Acad. Sci. U.S.A 1998; 95:12111-12116.
27. Wu JQ, Zaleski TJ, Valenzano C, Khosla C, Cane DE. Polyketide double bond biosynthesis. Mechanistic analysis of the dehydratase- containing module 2 of the picromycin/methymycin polyketide synthase. J. Am. Chem. Soc. 2005;127:17393-17404.
28. Leesong M, Henderson BS, Gillig JR, Schwab JM, Smith JL. Structure of a dehydratase-isomerase from the bacterial pathway for biosynthesis of unsaturated fatty acids: two catalytic activities in one active site. Structure 1996; 4:253-264.
29. Sedgwick B, Morris C, French SJ. Stereochemical course of dehydration catalyzed by yeast fatty-acid synthetase. J. Chem. Soc. Chem. Commun. 1978; 193-194.
30. Perlova O, Klaus G, Kaiser O, Hans A, Muller R. Identification and analysis of the chivosazol biosynthetic gene cluster from the myxobacterial model strain Sorangium cellulosum So ce56. J. Biotechnol. 2006; 121:174-191.
31. Tang L, Shah S, Chung L, Carney J, Katz L, Khosla C, Julien B. Cloning and heterologous expression of the epothilone gene cluster. Science 2000; 287:640-642.
32. Tang L, Ward S, Chung L, Carney JR, Li Y, Reid R, Katz L. Elucidating the mechanism of cis double bond formation in epothilone biosynthesis. J. Am. Chem. Soc. 2004; 126:46-47.
33. Palaniaappan N, Kim BS, Sekiyama Y, Osada H, Reynolds KA. Enhancement and selective production of phoslactomycin B, a protein phosphatase IIa inhibitor, through identification and engineering of the corresponding biosynthetic gene cluster. J. Biol. Chem. 2003; 278:35552-35557.
34. Carvalho R, Reid R, Viswanathan N, Gramajo H, Julien B. The biosynthetic genes for disorazoles, potent cytotoxic compounds that disrupt microtubule formation. Gene 2005; 359:91-98.
35. Chang Z, Sitachitta N, Rossi JV, Roberts MA, Flatt PM, Jia J, Sherman DH, Gerwick WH. Biosynthetic pathway and gene cluster analysis of curacin A, an antitubulin natural product from the tropical marine cyanobacterium Lyngbya majuscula. J. Nat. Prod. 2004; 67:1356-1367.
36. Butcher RA, Schroeder FC, Fischbach MA, Straight PD, Kolter R, Walsh CT, Clardy J. The identification of bacillaene, the product of the PksX megacomplex in Bacillus subtilis. Proc. Natl. Acad. Sci. 2007; 104:1506-1509.
37. Alhamadsheh MM, Palaniappan N, DasChouduri S, Reynolds KA. Modular polyketide synthases and cis double bond formation: establishment of activated cis-3-cyclohexylpropenoic acid as the diketide intermediate in phoslactomycin biosynthesis. J. Am. Chem. Soc. 2007; 129:1910-1911.
38. Negishi M, Pedersen LG, Petrotchenko E, Shevtsov S, Gorokhov A, Kakuta Y, Pedersen LC. Structure and function of sulfotrans- ferases. Arch. Biochem. Biophys. 2001; 390:149-157.
39. Shim JH, Lee HK, Chang EJ, Chae WJ, Han JH, Han DJ, Morio T, Yang JJ, Bothwell A, Lee SK. Immunosuppressive effects of tautomycetin in vivo and in vitro via T cell-specific apoptosis induction. Proc. Natl. Acad. Sci. U.S.A. 2002; 99:10617-10622.
40. Choi SS, Hur YA, Sherman DH, Kim ES. Isolation of the biosynthetic gene cluster for tautomycetin, a linear polyketide T cell-specific immunomodulator from Streptomyces sp. CK4412. Microbiology 2007; 153:1095-1102.
41. Gonzalez-Lergier J, Broadbelt LJ, Hatzimanikatis V. Theoretical considerations and computational analysis of the complexity in polyketide synthesis pathways. J. Am. Chem. Soc. 2005; 127: 9930-9938.
42. Cheng YQ, Tang GL, Shen B. Type I polyketide synthase requiring a discrete acyltransferase for polyketide biosynthesis. Proc. Natl. Acad. Sci. U.S.A. 2003; 100:3149-3154.
43. Tang GL, Cheng YQ, Shen B. Leinamycin biosynthesis revealing unprecedented architectural complexity for a hybrid polyke- tide synthase and nonribosomal peptide synthetase. Chem. Biol. 2004; 11:33-45.
44. Sudek S, Lopanik NB, Waggoner LE, Hildebrand M, Anderson C, Liu H, Patel A, Sherman DH, Haygood MG. Identification of the putative bryostatin polyketide synthase gene cluster from “Candidatus Endobugula sertula”, the uncultivated microbial symbiont of the marine bryozoan Bugula neritina. J. Nat. Prod. 2007; 70:67-74.
45. Chen XH, Vater J, Piel J, Franke P, Scholz R, Schneider K, Koumoutsi A, Hitzeroth G, Grammel N, Strittmatter AW, Gottschalk G, Sussmuth RD, Borriss R. Structural and functional characterization of three polyketide synthase gene clusters in Bacillus amyloliquefaciens FZB 42. J. Bacteriol. 2006; 188:4024-4036.
46. Calderone CT, Kowtoniuk WE, Kelleher NL, Walsh CT, Dorrestein PC. Convergence of isoprene and polyketide biosynthetic machinery: isoprenyl-S-carrier proteins in the pksX pathway of Bacillus subtilis. Proc. Natl. Acad. Sci. U.S.A. 2006; 103:8977-8982.
47. El-Sayed AK, Hothersall J, Cooper SM, Stephens ER, Simpson TJ, Thomas CM. Characterization of the mupirocin biosynthesis gene cluster from Pseudomonas fluorescens NCIMB 10586. Chem. Biol. 2003; 10:419-430.
48. Pulsawat N, Kitani S, Nihira T. Characterization of biosynthetic gene cluster for the production of virginiamycin M, a streptogramin type A antibiotic, in Streptomyces virginiae. Gene 2007; 393:31-42.
49. Simunovic V, Zapp J, Rachid S, Krug D, Meiser P, Muller R. Myxovirescin A biosynthesis is directed by hybrid polyketide synthases/nonribosomal peptide synthetase, 3-hydroxy-3- methylglutaryl-CoA synthases, and trans-acting acyltransferases. ChemBioChem. 2006; 7:1206-1220.
50. Simunovic V, Muller R. 3-Hydroxy-3-methylglutaryl-CoA-like synthases direct the formation of methyl and ethyl side groups in the biosynthesis of the antibiotic myxovirescin A. ChemBioChem. 2007; 8:497-500.
51. Piel J, Hui D, Wen G, Butzke D, Platzer M, Fusetani N, Matsunaga S. Antitumor polyketide biosynthesis by an uncultivated bacterial symbiont of the marine sponge Theonella swinhoei. Proc. Natl. Acad. Sci. U.S.A. 2004; 101:16222-16227.
52. Piel J, Wen G, Platzer M, Hui D. Unprecedented diversity of catalytic domains in the first four modules of the putative pederin polyketide synthase. ChemBioChem. 2004; 5:93-98.
53. Piel J. A polyketide synthase-peptide synthetase gene cluster from an uncultured bacterial symbiont of Paederus beetles. Proc. Natl. Acad. Sci. U.S.A. 2002; 99:14002-14007.
54. Gu L, Jia J, Liu H, Hakansson K, Gerwick WH, Sherman DH. Metabolic coupling of dehydration and decarboxylation in the curacin A pathway: functional identification of a mechanistically diverse enzyme pair. J. Am. Chem. Soc. 2006; 128:9014-9015.
55. Edwards DJ, Marquez BL, Nogle LM, McPhail K, Goeger DE, Roberts MA, Gerwick WH. Structure and biosynthesis of the jamaicamides, new mixed polyketide-peptide neurotoxins from the marine cyanobacterium Lyngbya majuscula. Chem. Biol. 2004; 11:817-833.
56. Lange BM, Rugan T, Martin W, Croteau R. Isoprenoid biosynthesis: The evolution of two ancient and distinct pathways across genomes. Proc. Natl. Acad. Sci. U.S.A. 2000; 97:13172-13177.
57. Dorrestein PC, Bumpus SB, Calderone CT, Garneau-Tsodikova S, Aron ZD, Straight PD, Kolter R, Walsh CT, Kelleher NL. Facile detection of acyl and peptidyl intermediates on thiotemplate carrier domains via phosphopantetheinyl elimination reactions during tandem mass spectrometry. Biochemistry 2006; 45:12756-12766.
58. Calderone CT, Iwig DF, Dorrestein PC, Kelleher NL, Walsh CT. Incorporation of nonmethyl branches by isoprenoid-like logic: multiple β-alkylation events in the biosynthesis of myxovirescin A1. Chem. Biol. 2007; 14:835-846.
59. Campobasso N, Patel M, Wilding IE, Kallender H, Rosenberg M, Gwynn MN. Staphylococcus aureus 3-hydroxy-3-methylglutaryl- CoA synthase. J. Biol. Chem. 2004; 279:44883-44888.
60. Steussy CN, Robison AD, Tetrick AM, Knight JT, Rodwell VW, Stauffacher CV, Sutherlin AL. A structural limitation on enzyme activity: the case of HMG-CoA synthase. Biochemistry 2006; 45:14407-14414.
61. Theisen ML, Misra I, Saadat D, Campobasso N, Miziorko HM, Harrison DHT. 3-Hydroxy-3-methylglutaryl-CoA synthase intermediate complex observed in “real-time”. Proc. Natl. Acad. Sci. U.S.A. 2004; 101:16442-16447.
62. Steussy CN, Vartia AA, Burgner JW II, Sutherlin A, Rodwell VW, Stauffacher CV. X-ray crystal structures of HMG-CoA synthase from Enterococcus faecalis and a complex with its second substrate/inhibitor acetoacetyl-CoA. Biochemistry 2005; 44:14256-14267.
63. Pojer F, Ferrer JL, Richard SB, Nagegowda DA, Chye ML, Bach TJ, Noel JP. Structural basis for the design of potent and species-specific inhibitors of 3-hydroxy-3-methylglutaryl CoA synthases. Proc. Natl. Acad. Sci. U.S.A. 2006; 103:11491-11496.
64. Geders TW, Gu L, Mowers JC, Liu H, Gerwick WH, Hakansson K, Sherman DH, Smith JL. Crystal structure of the ECH2 catalytic domain of CurF from Lyngbya majuscula: Insights inot a decarboxylase involved in polyketide chain β-branching. J. Biol. Chem. 2007; 282:35954-35963.
65. Simunovic V, Muller R. Mutational analysis of the myxovirescin biosynthetic gene cluster reveals novel insights into the functional elaboration of polyketide backbones. ChemBioChem. 2007; 8:1273-1280.
66. Straight PD, Fischbach MA, Walsh CT, Rudner DZ, Kolter R. A singular enzymatic megacomplex from Bacillus subtilis. Proc. Natl. Acad. Sci. U.S.A. 2007; 104:305-310.
Nature: A Model System for Chemists
Antibiotics, Biosynthesis of
Natural Products in Microbes, Chemical Diversity of
Pharmaceuticals, Natural Products and Natural Product Models for