Critical Review Iterative Type I Polyketide Synthases Involved in Enediyne Natural Product Biosynthesis

Xiaolei Chen1* Rui Ji2* Xin Jiang3 Rongqiang Yang2 Fuli Liu4 Ying Xin5

1

Department of Chemistry, Dartmouth College, Hanover, NH, USA Department of Biochemistry and Molecular Biology, University of Louisville, School of Medicine, Louisville, KY, USA 3 Department of Radiation Oncology, The First Hospital of Jilin University, Changchun, China 4 Department of Physiology and Neurobiology, Geisel Medical School at Dartmouth, Lebanon, NH, USA 5 Key Laboratory of Pathobiology, Ministry of Education, Jilin University, Changchun, China 2

Abstract Enediyne natural products are potent antibiotics structurally characterized by an enediyne core containing two acetylenic groups conjugated to a double bond in a 9- or 10-membered carbocycle. The biosynthetic gene clusters for enediynes encode a novel iterative type I polyketide synthase (PKSE), which is generally believed to initiate the biosynthetic process

Keywords: enediyne; iterative biosynthesis; thioesterase

type

I;

polyketide

synthase;

Introduction Enediyne natural products discovered in 1980s from soil and marine microorganisms are structurally characterized by enediyne cores (warheads) consisting of two acetylenic groups conjugated to a double bond in a 9- or 10-membered carbocycle ring (1–5). The enediyne natural products are categorized into two subfamilies based on enediyne core structures: 9- and 10-membered enediynes. All known 10-membered enediynes are isolated as stable free-standing chromophores, whereas most 9-membered enediynes are associated with pro-

C 2014 International Union of Biochemistry and Molecular Biology V

Volume 66, Number 9, September 2014, Pages 587–595 Address correspondence to: Ying Xin, Key Laboratory of Pathobiology, Ministry of Education, Jilin University, Changchun 130021, China. E-mail: [email protected] *These authors contributed equally to this manuscript. Received 21 July 2014; Accepted 14 September 2014 DOI 10.1002/iub.1316 Published online 3 October 2014 in Wiley Online Library (wileyonlinelibrary.com)

IUBMB Life

of enediyne cores. This review article will cover research efforts made since its discovery to elucidate the role of the PKSE in enediyne core biosynthesis. Topics covered include the unique domain architecture, identification, and characterization of turnover products, and interaction with partner thioC 2014 IUBMB Life, 66(9):587–595, 2014 esterase protein. V

tective apoproteins. Figure 1 shows several examples of 9membered enediynes C-1027 from Streptomyces globisporus (6) and neocarzinostatin from Streptomyces macromomyceticus (7) and 10-membered enediynes calicheamicin from Micromonospora echinospora (3,8) and esperamicin from Actinomadura verrucosopora (9). Enediyne natural products possess strong DNA cleavage activity endowed by the highly unsaturated enediyne cores. DNA cleavage by enediynes is achieved by a common mechanism of action (10): the enediyne core undergoes an electrocyclic rearrangement to yield a transient benzeniod diradical, which is locked in the minor groove of DNA and abstracts hydrogen atoms from deoxyriboses on both strands. In the presence of O2, the DNA will undergo facile double- or singlestranded cleavage through an oxidative radical mechanism. While the enediyne cores serve as active sites of DNA cleavage activity, peripheral moieties such as sugar moieties and aromatic moieties are responsible for DNA binding specificity and stabilization of enediyne cores. This DNA cleavage activity makes enediynes some of the most potent antitumor natural products and thus excellent candidates for developing anticancer drugs. A neocarzinostatin-polymer conjugate SMANCS (11) and calicheamicin-antibody conjugate Mylotarg ((12);

587

IUBMB LIFE

FIG 1

Structures of several enediyne natural products. All enediyne natural products contain highly unsaturated enediyne cores shown in red. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

withdrawn in 2010 due to concerns about effectiveness and safety) are used clinically in Japan and the United States for treatment of hepatocellular carcinoma and refractory acute myeloid leukemia, respectively. Structural complexity and antitumor activities of enediynes have attracted considerable attention to investigate their biosynthesis. Understanding biosynthesis of enediynes, enables the possibility of engineering new combinations of genes to produce novel products potentially with higher antitumor activity and better cancer cell targeting specificity. Although total synthesis of almost every subfamily of enediyne has been achieved (13–16), their native origins, especially key steps towards enediyne core biosynthesis are still elusive to us. Sequencing and bioinformatics analysis of enediyne gene clusters (17,18) revealed that biosynthesis of enediyne core is initialized by catalytic function of an iterative type I polyketide synthase (PKSE; (19)), which synthesizes a carbon skeleton as the first precursor toward enediyne core synthesis. This review article will cover recent research development on this type I enediyne PKSE with a focus on the domain architecture, identification, and characterization of PKSE products and function of thioesterase partner protein.

Enediyne Gene Cluster and Its Iterative Type I PKSE Before discovery of genes involved in enediyne biosynthesis, classical isotope-labeling was the major tool to explore native origin of enediynes. For example, feeding of 13C labeled ace-

588

tate to corresponding enediyne hosts produced 13C-enriched neocarzinostatin (20), dynemicin A (21), and esperamicin (22). 13 C NMR analysis of the 13C-enriched enediyne products suggests that their enediyne cores are derived from a linear precursor with eight acetate units assembled in a head-to-tail fashion. Although several genes of enediynes producing strain were identified to encode proteins involved in enediyne biosynthesis (23,24), the blueprint was not clear until two groundbreaking articles were published in “Science magazine” in 2002. These two articles reported sequencing and partial annotation of the gene clusters for 9-member enediyne C-1027 and 10membered enediyne calichaemicin biosynthesis respectively (17,18). Each of the two gene clusters spans 80 kb on the genome and contains 70–80 genes predicted to encode enzymes, transporter, transcriptional regulators, and unknown function proteins. Bioinformatics analysis revealed that 5 genes in close vicinity to each other are conserved between the two gene clusters. Since the enediyne core is the only moiety shared between C-1027 and calichaemicin, these 5 genes were predicted to be involved in enediyne core biosynthesis and thus were named as the “warhead gene cassette.” The warhead gene cassette contains an iterative type I PKSE gene, a thioesterase (TEBC) gene and three genes encoding proteins with unknown functions (UNBL, UNBV, and UNBU). In the following years, sequencing of gene clusters for neocarzinostatin (25), maduropeptin (26), kedarcidin (27), and dynemcin (28)

Type I PKSE in Enediyne Natural Product Biosynthesis

FIG 2

A: warhead gene cassettes of 10-membered enediynes dynemicin (DYNE) and calichaemicin (CALI), and 9-membered eneidynes neocarzinostatin (NCS), maduropeptin (MADU), and C-1027; (B) domain organization comparison between PKSE and PUFA synthase. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

confirmed the conservation of the warhead gene cassette among enediyne gene clusters (Fig. 2A). Disruption of any of these PKSE genes abolished production of corresponding enediyne metabolite, and gene complementation restored the biosynthesis in the mutant strains. These results firmly established the polyketide origin of the enediyne core. Sequence analysis of PKSE has identified four domains: ketosynthase (KS), acyltransferase (AT), ketoreductase (KR), and dehydratase (DH). In terms of domain organization and sequence homology, PKSE are most related to PUFA synthase involved in biosynthesis of docosahexaenoic acid in Moritella marina and eicosapentaenoic acid in Shewanella japonica ((29,30); Fig. 2B). Yet, the region between AT and KR domain of PKSE was found homologous to no known protein. This region was proposed to be acyl carrier protein (ACP) domain solely based on domain architecture of PUFA synthase. It suggests that PKSE has unique ACP domain among PKSE. Moreover, the C-terminal region of PKSE was predicted to be phosphopantetheinyl transferase (PPTase) responsible for selfactivation of PKSE by loading a phosphopantetheinyl (p-pant) group from coenzyme A (CoA) to the ACP domain. The p-pant group on ACP serves as a flexible arm that harbors the elongating polyketide chain. Self-activation by integrated PPTase domain, which has been seen in a type I FAS from yeast (31), is very rare among PKS with only one precedent: the PPTase domain (SePptI) associated with erythromycin biosynthesis in S. erythraea (32). The unique domain architecture of PKSE will be discussed in detail in “Domain Architecture of PKSE” section with a focus on interaction between ACP and PPTase domain. Though it has been confirmed that PKSE functions in the first step in the biosynthesis of enediyne core, the turnover product of PKSE needs to be identified for assignment of PKSE’s specific role in enediyne biosynthesis. Since PKSE from 9-membered enediyne and 10-membered enediyne share high

Chen et al.

sequence homology, it is tempting to propose that they produce the same precursor, which goes on separate pathways toward 9- and 10-membered enediyne core biosynthesis. Moreover, two types of thioesterases in general are involved in synthesis of polyketides, type I thioesterases (TE I) as integrated domains on PKS and type II thioesterases (TE II) as discrete proteins. With a few exceptions, TE I are responsible for the release and cyclization of final polyketide products (33,34), whereas TE II are responsible for the trans hydrolytic release of misloaded substrates and aberrant intermediates (35–37). It remains mysterious whether the discrete enediyne thioesterases are for release and cyclization of final product or for removal of misloaded substrates and aberrant intermediates. The next several sections will be devoted to discuss recent research attempting to answer these questions.

Domain Architecture of PKSE Bioinformatics analysis unambiguously identified KS, AT, KR, and DH domain of PKSE and revealed similarity in domain organization with PUFA synthase (18). Although region between AT and KR domain and C-terminal region of PKSE were found homologous to no known protein, the former was assigned as ACP domain based on domain architecture of PUFA synthase (38), and structural modeling of the latter revealed significant similarities with the PPTase Sfp associated with surfactin biosynthesis in Bacillus subtillis (39). Therefore, the C-terminal region was assigned as an integrated PPTase domain (38) for phosphopantetheinylation of the ACP. In a study on 9-membered PKSE SgcE (40) in C-1027 biosynthesis, replacement of conserved residues by Ala, including C211 of KS, S659 of AT, S974 of ACP, and D1827 and E1829 of PPTase, abolished production of C-1027. These results lend support to the domain assignment of PKSE. Yet unambiguous identification of ACP and PPTase required in vitro characterization of the two. Peptide mapping of SgcE by Fouriertransform mass spectrometry clearly identified a 1340 amu mass shift associated with S974 of the ACP domain. The mass shift is consistent with phosphopantetheinylation of the ACP domain, and the mass shift was not observed if S974 of ACP or D1827 of PPTase was substituted by Ala. Moreover, the ACP domain of SgcE was expressed as a stand-alone protein in E. coli and was successfully phosphopantetheinylated after in vitro incubation with CoA and Svp, a promiscuous PPTase from the bleomycin-producer Streptomyces verticillus (41). These results undoubtedly established the region between AT and KR domain of PKSE as a novel ACP domain. Liang and coworkers characterized the ACP and PPTase domain of a 10-membered PKSE CalE8 (42,43). The ACP domain of CalE8 was expressed as a stand-alone protein and showed it could be phosphopantetheinylated by Sfp, which further confirmed the identity of the ACP domain. The crystal structure of ACP domain (43) revealed a helix-bundle structure that is widely adopted by conventional ACPs, except that the helix-bundle only consists of three helices, instead of four in other ACPs. Moreover, the orthodox GX(H/ D)S(L/I) motif is

589

IUBMB LIFE

replaced by a H968MSS971I motif in CalE8 ACP with His968 and Ser971 substituting the canonical Gly and Asp/His residues. S971 (counterpart in 9-membered PKSE SgcE: S974) was confirmed to be the anchor that the p-pant group is attached on, and substitution of S971 by Ala abolished activity of CalE8. The PPTase domain of PKSE was proposed to be Sfp-like type II PPTase based on structural modeling results and identification of conserved residues found in type II PPTase (38). The stand-alone protein of CalE8 PPTase expressed in E. coli was found to adopt a pseudo-trimeric structure, distinct from the pseudo-dimeric structure of type II PPTases. In vitro activity assays demonstrated that the purified PPTase domain as a stand-alone protein is capable of phosphopantetheinylation of ACP from E. coli fatty acid synthesis pathway (44) and PCP domain of nonribosomal peptide synthase from B. brevis (45), yet unable to phosphopantetheinylate its cognate ACP domain, which was also expressed as a stand-alone protein. Further sequence analysis revealed that the PPTase domain of PKSE shares higher sequence homology with PcpS (46,47), a type II PPTase that prefers fatty acid synthase-ACPs, than Sfp, which prefers non-ribosomal peptide synthase-peptidyl carrier proteins (PCPs; 20% identity and 32% similarity with PcpS versus 18% identity and 31% similarity with Sfp). Combining the sequence analysis result and substrate preferences of PKSE PPTase domain, Liang proposed that the PKS probably acquired the PPTase domain from a primary fatty acid metabolic pathway during evolution. Chen and coworkers demonstrated SgcE D1827A mutant lost its PKS activity due to inactivation of the PPTase domain, and the PKS activity was moderately restored after incubation with isolated PPTase domain of SgcE (unpublished data), indicating phosphopantetheinylation in trans of the ACP domain of the SgcE D1827A mutant. This finding suggests that PKSE PPTase is able to modify the ACP integrated on PKSE as a domain. The flanking sequences of the ACP on PKSE are essential for recognition by its cogate PPTase domain. It was also proposed that modification of the ACP by PPTase of PKSE adopts a mechanism shared by fatty acid synthase: two PKSE protein molecules form a dimer and modify each other’s ACP domain (48). After establishment and characterization of the unique ACP and PPTase domains of PKSE, research efforts were devoted to study of the AT domain. Structures of potential turnover products of PKSE (see “Products of PKSE” section) suggest the AT domain of PKSE utilize acetyl CoA as a starter unit and malonyl CoA as extender units. However, in vitro activity assay of CalE8 demonstrated that acetyl CoA is unnecessary for the activity of CalE8 (49,50). Moreover, PKSE is not able to convert malonyl CoA to acetyl CoA. This finding indicates that the AT domain may be able to use malonyl CoA as both starter unit and extender unit, which raises interest among researchers about substrate specificity of the AT domain. Crystal structure of the AT domain of DynE8, the PKSE in 10-membered dynemicin biosynthesis, revealed high similarity with malonyl-CoA-specific AT domains of type I modular PKS and provided the molecular basis for substrate

590

preferences of the PKSE AT domain (51). Interestingly, although PKSE AT does not incorporate acetate unit of acetyl CoA to polyketide product, the isolated AT domian is able to hydrolyze acetyl CoA and forms a noncovalent complex with the released the acetyl group. In contrast, the AT domain forms a malonyl-enzyme covalent complex with the malonyl group from hydrolysis of malonyl CoA. Moreover, acetyl CoA is found to be able to inhibit activity of DynE8 (51), which further supports acetyl CoA competes with the substrate of the AT domain. Taken together, evidence show that PKSE does not use acetyl CoA as a starter unit, yet it loads the acetyl group to its active site and then off-loads it by a hydrolytic process. Existence of a malonyl-enzyme complex suggests the AT domain is not able to decarboxylate the malonate unit to generate acetyl starter unit; instead, the decarboxylation is catalyzed by KS domain, as demonstrated in cases of aromatic PKS involved in actinorhodin, tetracycline, and doxorubicin biosynthesis (52,53).

Products of PKSE PKSE functions in the first step in biosynthesis of the enediyne core. Understanding its function by identification of its turnover product is critical to unveil enediyne core biosynthesis. Based on the domain architecture, the product was predicted to be a polyene (Fig. 3). However, the exact structure of the product cannot be predicted by domain organization. Shen and coworkers isolated the first PKSE product 1,3,5,7,9,11,13-pentadecaheptaene (1) by coexpression of 9membered PKSE SgcE (C-1027) or NcsE (neocarzinostatin) with their cognate thioesterase SgcE10 or NcsE10 (40). The structure of the product is consistent with the domain organization of PKSE. SgcE or NcsE catalyzes seven rounds of Claisen condensation in an iterative manner. Within each round, the KR reduces the newly added keto group to a hydroxyl, and then the DH dehydrates the hydroxyl to yield a double bond. Thioesterase releases a linear polyene that undergoes decarboxylation and dehydration to afford the heptaene (1) (Fig. 4A). In vivo experiments further show that SgcE and NcsE are interchangeable within the 9-membered families of enediynes. Based on these findings, Shen proposed that this heptaene (1) produced by SgcE and NcsE is a common intermediate for all 9-membered enediyne core biosynthesis, and that thioesterase SgcE10 or NcsE10 is indispensible for product release. Coexpression of the 10-membered PKSE CalE8 (calichaemicin) with its partner thioesterase CalE7 and in vitro activity reconstitution of CalE8 and CalE7 produced a carbonylconjugated polyene (3,5,7,9,11,13-pentadecen-2-one) (2) (49). Production of 2 by CalE8 has a similar mechanism to production of 1 except that after the last round of Claisen condensation, the function of the KR is skipped and the keto group is left unreduced (Fig. 4B). Different products produced by 9and 10-membered PKSE led to the hypothesis that 9- and 10membered enediyne cores biosynthetically diverge at the PKSE stage. However, it remains mysterious how the homologous

Type I PKSE in Enediyne Natural Product Biosynthesis

FIG 3

Prediction of PKSE’s product. A core PKS consists of AT for loading of substrate units, KS for elongation of polyketide chain by ligating the substrate unit through Claisen condensation, and acyl carrier protein (ACP) phosphopantetheinylated by PPTase domain for harboring the nascent polyketide chain through a thioester linkage. Keto-modification components consist of KR for reducing the keto to hydroxyl and dehydrotase (DH) for dehydration of hydroxyl to double bond. Iteration of above process yields polyene product with conjugated double bonds. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

CalE8 and SgcE generate different products by controlling the oxidation state in the last round of chain extension. Several following reports found that in vitro reconstitution of PKSE’s activity in the presence of thioesterase produces heptaene (1) as the common product for both 9-membered PKSE SgcE and NcsE and 10-membered PKSE CalE8 (50,54). Besides 1 and 2, PKSE also produces several truncated pyrones 3 (Fig. 4B). Amounts of these products are highly dependent on reaction conditions, such as buffer pH value, temperature, and concentration of NADPH. This highly irregular behavior of PKSE raised suspicion among researchers that none of the identified polyketides is the true turnover product of PKSE. It was proposed that production of the correct polyketide by PKSE might need a trans-acting regulatory enzyme. This is not unprecedented. The study of the iterative PKS LovB in lovastatin biosynthesis has found that assembly of the correct polyketide product by LovB relies on the trans-acting enoylreductase (ER), LovC. Without the ER protein, nonaketide synthase LovB does not synthesize a full-length product but releases conjugated hexaketide and heptaketide pyrone derailment products instead (51). As a further effort to reconcile the apparent catalytic difference between 9- and 10-membered PKSE, Shen and coworker resorted to robust in vivo assays in the presence of the mild and regenerative conditions of the cells (55). Coexpression of 9-membered PKSE SgcE (C-1027), NcsE (neocarzinostatin) or MdpE (maduropeptin), or 10-membered PKSE CalE8 (calichaemicin) or DynE (dynemicin) with their cognate thioesterases in several different hosts all produced heptaene

Chen et al.

(1) as the major product in comparable amounts. Furthermore, all the five thioesterases were freely interchangeable with the five PKSE to yield heptaene 1, clearly indicating 9versus 10-membered enediyne core divergence does not originate from the PKSE-TE interaction. Accumulation of 1 detected in the 5 native enediyne producers (55) seems to lend compelling support to the proposal that heptaene 1 is a common intermediate of enediyne core biosynthesis. However, two critical pieces of evidence undermine the above proposal. First, the high amount of 1 accumulated in enediyne native producers questions the relevance of 1 in enediyne biosynthesis. For example, S. globisporus produces 14 mg=L of 1 and 1 mg=L of C-1027 chromophore, which suggests successful C-1027 production in only about 1 of every 70 enzyme turnovers. Second, only 9-membered pksE genes but not 10membered pksE genes could restore 9-membered enediyne production in pksE null mutants. Hence, Shen envisioned a universal PKSE-bound intermediate that is modified by pathway-specific accessory enzymes en route to formation of the 9- or 10-membered enediyne core divergence and the heptaene 1 might be a shunt product. All above-mentioned reports seem to agree that the thioesterase is indispensible for release of PKSE’s product. In vitro activity reconstitution of 9-membered PKSE SgcE (C-1027) by Chen et al. (56) revealed that SgcE synthesizes a nonaketide 3 efficiently as the only product without assistance of its thioesterase SgcE10 (Fig. 4C). Inclusion of SgcE10 at a concentration up to 2 lM does not change the outcome. Further increase of SgcE10 concentration decreases production of 3 and leads to

591

IUBMB LIFE

FIG 4

Chronicle of PKSE product discoveries. A: Heptaene 1 was isolated from 9-membered PKSE coexpression with thioesterase in heterogeneous host (2008; 39). B: Deficient NADPH causes release of truncated polyketide 3 and methylhexaenone 2 from 10membered PKSE catalyzed reaction with thioesterase (2008–2009; 48, 49, 51, 57). C: Without assistance of thioesterase, 9membered enediyne SgcE produces nonaketide 4 as major product; high concentration of thioesterase SgcE10 causes release of premature intermediate 1 and 5 (2010; 53). D: In vivo coexpression of 10-membered CalE8 in heterogeneous host in dark without thioesterase produces b-hydroxy acid 5 as the major product (2012; 55). Note: 1) M-CoA, malonyl CoA; TE, thioesterase; 2) PKSE in blue: 9-membered PKSE represented by SgcE and NcsE, PKSE in orange: 10-membered eneidyne represented by CalE8 and DynE; 3) CO2 and H2O side products released from decarboxylation and dehydration are not shown for the purpose of clarity. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

production with increasing yields of heptaene 1 and another truncated polyketide product 5 (Fig. 4C). The nonaketide 3 was proposed by the authors to be the precursor of the enediyne core biosynthesis. The thioesterase was proposed to be a TE II for removing aberrant products from the synthase (36,57). Yet, at high concentration, it hydrolyzes the biosynthetic intermediates 1 and 5. To identify the PKSE-bound intermediate first proposed by Shen and shared by 9- and 10-membered enediynes (55), Townsend and coworkers expressed 10-membered PKSE CalE8 (calichaemicin) in a heterologous host in dark without its thioesterase CalE7 and isolated b-hydroxy acid 6 as the major product ((58); Fig. 4D). Townsend reasoned that the bhydroxy acid 6 is the result of host-assisted hydrolysis of the PKSE-bound intermediate b-hydroxyhexaene 7. In a later report (59), Townsend and coworkers devised the mild cysteamine-promoted release of intermediates bound to carrier proteins. This technique enabled preparation of intermediate free CalE8 for in vitro assays. Activity assays of the “cleanslate” CalE8 in dark without CalE7 followed by chemical

592

hydrolysis yielded b-hydroxy acid 6 as the major product. Taken together, these findings support that the PKSE-bound bhydroxy thioester 7 is the common intermediate in 9- and 10memebered enediyne core biosynthesis. It seems to contradict Chen’s report that nonaketide 5 is the only product of 9membered PKSE SgcE in the absence of SgcE10 thioesterase (56). However, since pretreatment of CalE8 by cysteamine to yield intermediate free enzyme is essential for CalE8’s activity, it takes further investigation to demonstrate if SgcE also produces the b-hydroxy acid after treatment by cysteamine. Final confirmation of the PKSE-bound b-hydroxyhexaene 7 as the intermediate of the enediyne core synthesis calls for further investigation, such as identification of the next enzyme that can successfully convert the intermediate. The structure of b-hydroxy thioester 7 inspired Townsend to propose the biosynthetic pathways for both 9- and 10-membered enediyne cores ((58,59); Fig. 5). 9- and 10-membered PKSE produce the common enzyme-bound intermediate 7 with which tailoring enzymes specific to each subclass interact to direct the divergence. One of the downstream tailoring enzymes abstracts

Type I PKSE in Enediyne Natural Product Biosynthesis

FIG 5

A proposed biosynthetic pathway for enediyne cores. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

hydrogen atom from the intermediate 7 to yield the resonance equivalent 8, which cyclizes C4-C15 in path A en route to 9membered enediyne core synthesis or C4-C13 in path B en route to 10-membered enediyne core synthesis. The biosynthetic mechanism proposed by Townsend also provides guidance to further discovery of downstream accessary enzymes for cyclization and maturation of enediyne cores.

Thioesterase in the Enediyne Core Biosynthesis In polyketide and nonribosomal peptide natural product biosynthesis, all substrates and intermediates are covalently bound to enzyme by a thioester linkage to the p-pant group of an ACP domain of PKS or PCP domain of NRPS. Thioesterases are able to break the thioester linkage to release the mature product or remove misloaded substrates and aberrant intermediates. Thioesterases that release mature products are categorized as type I thioeterases and are usually fused at the C-terminal of the most downstream module of the PKS or NRPS (33,34). In some cases, TE I were also found with deacylation activity to release a truncated product, such as thioesterase associated with fungal melanin biosynthesis (60). Discrete TE II are generally responsible for removal of misloaded substrates and aberrant intermediates (35–37,61) and are structurally and evolutionarily related to a well-known a/b hydrolase family (62–67). However, several TE II are also involved in product release, such as NanE in nanchangmycin biosynthesis (68) and MonCII in polyether ionophore monensin biosynthesis (69). Sequence analysis of enediyne thioesterases suggests that they do not belong to either type I, TE II, or a/b hydrolase family. Instead, enediyne thioesterases were found to share moderate sequence homology with a family of hotdog fold thioesterases (70–73) characterized by a long central a-helix packed against a five-stranded anti-parallel b-sheet. Liang and coworkers solved the crystal structures of CalE7 (74) and DynE7

Chen et al.

(75), thioesterases in 10-membered enediyne calichaomicin and dynemicin biosynthesis, respectively. The hotdog fold thioesterase structures of both CalE7 and DynE7 had a ec oligomeric arrangement and a kinked substrate-binding channel, confirming the homology between enediyne thioesterases and the hotdog fold thioesterase family. Since most known hotdog fold thioesterases use acyl-CoAs as substrates, enediyne thioesterases that act on PKSEintermediate complex may represent a novel family of hotdog fold thioesterase. Besides, most known hotdog fold thioesterases catalyze the hydrolysis of thioester bond using a Glu/Asp residue as a nucleophile or general base during catalysis. Enediyne thioesterases CalE7 or DynE7 do not contain an acidic residue in the active site. Based on structural and biochemical data, an alternative catalytic mechanism was proposed (74,75) with R37 (CalE7) or R35 (DynE7) playing a critical catalytic role. The thioesterase first recognizes and binds to the polyene intermediate, which could be the correct product of PKSE or an aberrant intermediate depending on the function of thioesterase. The nucleophilic attack of the thioester carbonyl by a water or hydroxide anion is facilitated with R37 (CalE7) or R35 (DynE7) acting as an oxyanion hole to stabilize the transition state. Subsequent collapse of the transition state breaks the thioester bond and releases the product. As discussed in the Products of PKSE section, in the presence of enediyne thioesterases, the PKSE produces a heptaene and other metabolites that may not be the precursors of enediyne core synthesis, indicating the discrete enediyne thioesterases may serve an editing function for the PKSE. However, after downstream tailoring enzymes finish processing the PKSE-bound intermediate, its eventual release for enediyne core maturation may still need function of a thioesterase. Since there is only one thioesterase gene found in eneidyne gene clusters, the enediyne thioesterase may have both editing and product release functions.

Summary The iterative type I PKSE involved in enediyne natural product biosynthesis are a unique PKS family with an uncommon domain organization. Characterization of this type of PKS together with its partner thioesterase by in vivo expression and in vitro assays has led to the discovery of a number of metabolites and has shed light upon enediyne core biosynthesis. Among identified metabolites, the PKSE-bound b-hydroxyhexaene is proposed to be the genuine precursor of both 9- and 10-membered enediyne core biosynthesis. Identification and characterization of downstream enzymes will be the focus of future study. Although it needs to be confirmed by further investigations, the novel hotdog fold thioesterase involved in enediyne biosynthesis is likely to have not only editorial but also product release function.

References [1] Smith, A. L. and Nicolaou, K. C. (1996) The enediyne antibiotics. J. Med. Chem. 39, 2103–2117.

593

IUBMB LIFE

[2] Liang, Z. X. (2010) Complexity and simplicity in the biosynthesis of enediyne natural products. Nat. Prod. Rep. 27, 499–528. [3] Shen, B., Liu, W., and Nonaka, K. (2003) Enediyne natural products: biosynthesis and prospect towards engineering novel antitumor agents. Curr. Med. Chem. 10, 2317–2325. [4] Van Lanen, S. G. and Shen, B. (2008) Biosynthesis of enediyne antitumor antibiotics. Curr. Top. Med. Chem. 8, 448–459. [5] Ji, R., Liu, F., Meng, L., and Chen, X. (2014) Structures and biosynthesis of enediyne natural products. Int. J. Adv. Innovat. Thoughts Ideas 3, 110. [6] Hu, J., Xue, Y. C., Xie, M. Y., Zhang, R., Otani, T., et al. (1988) A new macromolecular antitumor antibiotic, C-1027.1. Discovery, taxonomy of producing organism, fermentation and biological-activity. J. Antibiot. (Tokyo) 41, 1575– 1579. [7] Edo, K., Mizugaki, M., Koide, Y., Seto, H., Furihata, K., et al. (1985) The structure of neocarzinostatin chromophore possessing a novel bicyclo[7,3,0]dodecadiyne system. Tetrahedron Lett. 26, 331–334. [8] Maiese, W. M., Lechevalier, M. P., Lechevalier, H. A., Korshalla, J., Kuck, N., et al. (1989) Calicheamicins, a novel family of antitumor antibiotics - taxonomy, fermentation and biological properties. J. Antibiot. (Tokyo) 42, 558–563. [9] Konishi, M., Ohkuma, H., Saitoh, K. I., Kawaguchi, H., Golik, J., et al. (1985) Esperamicins, a novel class of potent antitumor antibiotics .1. Physicochemical data and partial structure. J. Antibiot. (Tokyo) 38, 1605–1609. [10] Galm, U., Hager, M. H., Van Lanen, S. G., Ju, J., Thorson, J. S., et al. (2005) Antitumor antibiotics: bleomycin, enediynes, and mitomycin. Chem. Rev. 105, 739–758. [11] Maeda, H. (2001) SMANCS and polymer-conjugated macromolecular drugs: advantages in cancer chemotherapy. Adv. Drug Deliv. Rev. 46, 169–185. [12] Sievers, E. L. and Linenberger, M. (2001) Mylotarg: antibody-targeted chemotherapy comes of age. Curr. Opin. Oncol. 13, 522–527. [13] Nicolaou, K. C., Hummel, C. W., Pitsinos, E. N., Nakada, M., Smith, A. L., et al. (1992) Total synthesis of calicheamicin gamma-1(I). J. Am. Chem. Soc. 114, 10082–10084. [14] Nicolaou, K. C., Chen, J. S., Zhang, H. J., and Montero, A. (2008) Asymmetric synthesis and biological properties of uncialamycin and 26-epi-uncialamycin. Angew. Chem. Int. Ed. 47, 185–189. [15] Ren, F., Hogan, P. C., Anderson, A. J., and Myers, A. G. (2007) Kedarcidin chromophore: synthesis of its proposed structure and evidence for a stereochemical revision. J. Am. Chem. Soc. 129, 5381–5383. [16] Ji, R., Meng, L., and Yang, R. (2014) Neuroprotective effects of pituitary adenylate cyclase-activating polypeptide. Int. J. Adv. Innovat. Thoughts Ideas 3, 112. [17] Liu, W., Christenson, S. D., Standage, S., and Shen, B. (2002) Biosynthesis of the enediyne antitumor antibiotic C-1027. Science 297, 1170–1173. [18] Ahlert, J., Shepard, E., Lomovskaya, N., Zazopoulos, E., Staffa, A., et al. (2002) The calicheamicin gene cluster and its iterative type I enediyne PKS. Science 297, 1173–1176. [19] Horsman, G. R., Van Lanen, S. G., and Shen, B. (2009) Iterative type I polyketide synthases for enediyne core biosynthesis. Methods Enzymol. 459, 97–112. [20] Hensens, O. D., Giner, J. L., and Goldberg, I. H. (1989) Biosynthesis of NCS Chrom A, the chromophore of the antitumor antibiotic neocarzinostatin. J. Am. Chem. Soc. 111, 3295–3299. [21] Tokiwa, Y., Miyoshi-Saitoh, M., Kobayashi, H., Sunaga, R., Konishi, M., et al. (1992) Biosynthesis of dynemicin A, a 3-ene-1,5-diyne antitumor antibiotic. J. Am. Chem. Soc. 114, 4107–4110. [22] Lam, K. S., Veitch, J. A., Golik, J., Krishnan, B., Klohr, S. E., et al. (1993) Biosynthesis of esperamicin A1, an enediyne antitumor antibiotic. J. Am. Chem. Soc. 115, 12340–12345. [23] Liu, W. and Shen, B. (2000) Genes for production of the enediyne antitumor antibiotic C-1027 in Streptomyces globisporus are clustered with the cagA gene that encodes the C-1027 apoprotein. Antimicrob. Agents Chemother. 44, 382–392. [24] Otani, T., Yasuhara, T., Minami, Y., Shimazu, T., Zhang, R., et al. (1991) Purification and primary structure of C-1027-AG, a selective antagonist of antitumor antibiotic C-1027, from Streptomyces globisporus. Agric. Biol. Chem. 55, 407–417.

594

[25] Otsuka, M., Ichinose, K., Fujii, I., and Ebizuka, Y. (2004) Cloning, sequencing, and functional analysis of an iterative type I polyketide synthase gene cluster for biosynthesis of the antitumor chlorinated polyenone neocarzilin in "Streptomyces carzinostaticus". Antimicrob. Agents Chemother. 48, 3468–3476. [26] Van Lanen, S. G., Oh, T. J., Liu, W., Wendt-Pienkowski, E., and Shen, B. (2007) Characterization of the maduropeptin biosynthetic gene cluster from Actinomadura madurae ATCC 39144 supporting a unifying paradigm for enediyne biosynthesis. J. Am. Chem. Soc. 129, 13082–13094. [27] Lohman, J. R., Huang, S. X., Horsman, G. P., Dilfer, P. E., Huang, T. T., et al. (2013) Cloning and sequencing of the kedarcidin biosynthetic gene cluster from Streptoalloteichus sp ATCC 53650 revealing new insights into biosynthesis of the enediyne family of antitumor antibiotics. Mol. BioSyst. 9, 478–491. [28] Gao, Q. and Thorson, J. S. (2008) The biosynthetic genes encoding for the production of the dynemicin enediyne core in Micromonospora chersina ATCC53710. FEMS Microbiol. Lett. 282, 105–114. [29] Metz, J. G., Roessler, P., Facciotti, D., Levering, C., Dittrich, F., et al. (2001) Production of polyunsaturated fatty acids by polyketide synthases in both prokaryotes and eukaryotes. Science 293, 290–293. [30] Jiang, H., Zirkle, R., Metz, J. G., Braun, L., Richter, L., et al. (2008) The role of tandem acyl carrier protein domains in polyunsaturated fatty acid biosynthesis. J. Am. Chem. Soc. 130, 6336–6337. [31] Fichtlscherer, F., Wellein, C., Mittag, M., and Schweizer, E. (2000) A novel function of yeast fatty acid synthase - Subunit alpha is capable of self-pantetheinylation. Eur. J. Biochem. 267, 2666–2671. [32] Weissman, K. J., Hong, H., Oliynyk, M., Siskos, A. P., and Leadlay, P. F. (2004) Identification of a phosphopantetheinyl transferase for erythromycin biosynthesis in Saccharopolyspora erythraea. Chembiochem 5, 116–125. [33] Trauger, J. W., Kohli, R. M., Mootz, H. D., Marahiel, M. A., and Walsh, C. T. (2000) Peptide cyclization catalysed by the thioesterase domain of tyrocidine synthetase. Nature 407, 215–218. [34] Akey, D. L., Kittendorf, J. D., Giraldes, J. W., Fecik, R. A., Sherman, D. H., et al. (2006) Structural basis for macrolactonization by the pikromycin thioesterase. Nat. Chem. Biol. 2, 537–542. [35] Heathcote, M. L., Staunton, J., and Leadlay, P. F. (2001) Role of type II thioesterases: evidence for removal of short acyl chains produced by aberrant decarboxylation of chain extender units. Chem. Biol. 8, 207–220. [36] Kim, B. S., Cropp, T. A., Beck, B. J., Sherman, D. H., and Reynolds, K. A. (2002) Biochemical evidence for an editing role of thioesterase II in the biosynthesis of the polyketide pikromycin. J. Biol. Chem. 277, 48028–48034. [37] Hu, Z., Pfeifer, B. A., Chao, E., Murli, S., Kealey, J., et al. (2003) A specific role of the Saccharopolyspora erythraea thioesterase II gene in the function of modular polyketide synthases. Microbiology 149, 2213–2225. [38] Zazopoulos, E., Huang, K. X., Staffa, A., Liu, W., Bachmann, B. O., et al. (2003) A genomics-guided approach for discovering and expressing cryptic metabolic pathways. Nat. Biotechnol. 21, 187–190. [39] Reuter, K., Mofid, M. R., Marahiel, M. A., and Ficner, R. (1999) Crystal structure of the surfactin synthetase-activating enzyme Sfp: a prototype of the 4’phosphopantetheinyl transferase superfamily. EMBO J. 18, 6823–6831. [40] Zhang, J., Van Lanen, S. G., Ju, J. H., Liu, W., Dorrestein, P. C., et al. (2008) A phosphopantetheinylating polyketide synthase producing a linear polyene to initiate enediyne antitumor antibiotic biosynthesis. Proc. Natl. Acad. Sci. USA 105, 1460–1465. [41] Sanchez, C., Du, L., Edwards, D. J., Toney, M. D., and Shen, B. (2001) Cloning and characterization of a phosphopantetheinyl transferase from Streptomyces verticillus ATCC15003, the producer of the hybrid peptide-polyketide antitumor drug bleomycin. Chem. Biol. 8, 725–738. [42] Murugan, E. and Liang, Z. X. (2008) Evidence for a novel phosphopantetheinyl transferase domain in the polyketide synthase for enediyne biosynthesis. FEBS Lett. 582, 1097–1103. [43] Lim, J., Kong, R., Murugan, E., Ho, C. L., Liang, Z. X., et al. (2011) Solution structures of the acyl carrier protein domain from the highly reducing type I iterative polyketide synthase CalE8. PLoS One 6. [44] Jones, A. L., Kille, P., Dancer, J. E., and Harwood, J. L. (1993) The cloning and overexpression of Escherichia-Coli acyl carrier protein. Biochem. Soc. Trans. 21, 202S.

Type I PKSE in Enediyne Natural Product Biosynthesis

[45] Yin, J., Liu, F., Li, X. H., and Walsh, C. T. (2004) Labeling proteins with small molecules by site-specific posttranslational modification. J. Am. Chem. Soc. 126, 7754–7755. [46] Finking, R., Solsbacher, J., Konz, D., Schobert, M., Schafer, A., et al. (2002) Characterization of a new type of phosphopantetheinyl transferase for fatty acid and siderophore synthesis in Pseudomonas aeruginosa. J. Biol. Chem. 277, 50293–50302. [47] Barekzi, N., Joshi, S., Irwin, S., Ontl, T., and Schweizer, H. P. (2004) Genetic characterization of pcpS, encoding the multifunctional phosphopantetheinyl transferase of Pseudomonas aeruginosa. Microbiology 150, 795–803. [48] Fichtlscherer, F., Wellein, C., Mittag, M., and Schweizer, E. (2000) A novel function of yeast fatty acid synthase. Subunit alpha is capable of self-pantetheinylation. Eur. J. Biochem. 267, 2666–2671. [49] Kong, R., Goh, L. P., Liew, C. W., Ho, Q. S., Murugan, E., et al. (2008) Characterization of a carbonyl-conjugated polyene precursor in 10-membered enediyne biosynthesis. J. Am. Chem. Soc. 130, 8142. [50] Sun, H. H., Kong, R., Zhu, D., Lu, M., Ji, Q., et al. (2009) Products of the iterative polyketide synthases in 9-and 10-membered enediyne biosynthesis. Chem. Commun. (Camb.) 7399–7401. [51] Liew, C. W., Nilsson, M., Chen, M. W., Sun, H. H., Cornvik, T., et al. (2012) Crystal structure of the acyltransferase domain of the iterative polyketide synthase in enediyne biosynthesis. J. Biol. Chem. 287, 23203–23215. [52] Bisang, C., Long, P. F., Cortes, J., Westcott, J., Crosby, J., et al. (1999) A chain initiation factor common to both modular and aromatic polyketide synthases. Nature 401, 502–505. [53] Keatinge-Clay, A., Maltby, D. A., Medzihradszky, K. F., Khosla, C., and Stroud, R. M. (2004) An antibiotic factory caught in action. Nat. Struct. Mol. Biol. 11, 888–893. [54] Belecki, K., Crawford, J. M., and Townsend, C. A. (2009) Production of octaketide polyenes by the calicheamicin polyketide synthase CalE8: implications for the biosynthesis of enediyne core structures. J. Am. Chem. Soc. 131, 12564. [55] Horsman, G. P., Chen, Y., Thorson, J. S., and Shen, B. (2010) Polyketide synthase chemistry does not direct biosynthetic divergence between 9- and 10membered enediynes. Proc. Natl. Acad. Sci. USA 107, 11331–11335. [56] Chen, X. L., Guo, Z. F., Lai, P. M., Sze, K. H., and Guo, Z. H. (2010) Identification of a nonaketide product for the iterative polyketide synthase in biosynthesis of the nine-membered enediyne C-1027. Angew. Chem. Int. Ed. 49, 7926–7928. [57] Guo, Z. F., Sun, Y., Zheng, S., and Guo, Z. (2009) Preferential hydrolysis of aberrant intermediates by the type II thioesterase in Escherichia coli nonribosomal enterobactin synthesis: substrate specificities and mutagenic studies on the active-site residues. Biochemistry 48, 1712–1722. [58] Belecki, K. and Townsend, C. A. (2012) Environmental control of the calicheamicin polyketide synthase leads to detection of a programmed octaketide and a proposal for enediyne biosynthesis. Angew. Chem. Int. Ed. Engl. 51, 11316–11319. [59] Belecki, K. and Townsend, C. A. (2013) Biochemical determination of enzyme-bound metabolites: preferential accumulation of a programmed octaketide on the enediyne polyketide synthase CalE8. J. Am. Chem. Soc. 135, 14339–14348.

Chen et al.

[60] Vagstad, A. L., Hill, E. A., Labonte, J. W., and Townsend, C. A. (2012) Characterization of a fungal thioesterase having claisen cyclase and deacetylase activities in melanin biosynthesis. Chem. Biol. 19, 1525–1534. [61] Meng, L., Jiang, X., and Ji, R. (2014) Role of IL6 and TNFa in hippocampal neurogenesis of TAM triple knockout mice. Int. J. Adv. Innovat. Thoughts Ideas 3, 109. [62] Holmquist, M. (2000) Alpha/beta-hydrolase fold enzymes: structures, functions and mechanisms. Curr. Protein Pept. Sci. 1, 209–235. [63] Dunn, G., Montgomery, M. G., Mohammed, F., Coker, A., Cooper, J. B., et al. (2005) The structure of the C-C bond hydrolase MhpC provides insights into its catalytic mechanism. J. Mol. Biol. 346, 253–265. [64] Li, C., Montgomery, M. G., Mohammed, F., Li, J. J., Wood, S. P., et al. (2005) Catalytic mechanism of C-C hydrolase MhpC from Escherichia coli: kinetic analysis of His263 and Ser110 site-directed mutants. J. Mol. Biol. 346, 241–251. [65] Jiang, M., Chen, X. L., Guo, Z. F., Cao, Y., Chen, M. J., et al. (2008) Identification and characterization of (1R,6R)22-succinyl-6-hydroxy-2,4-cyclohexadiene-1-carboxylate synthase in the menaquinone biosynthesis of Escherichia coli. Biochemistry 47, 3426–3434. [66] Jiang, M., Chen, X. L., Wu, X. H., Chen, M. J., Wu, Y. D., et al. (2009) Catalytic mechanism of SHCHC synthase in the menaquinone biosynthesis of Escherichia coli: identification and mutational analysis of the active site residues. Biochemistry 48, 6921–6931. [67] Yan, X.-B., Wang, S.-S., Hou, H.-L., Ji, R., and Zhou, J.-N. (2007) Lithium improves the behavioral disorder in rats subjected to transient global cerebral ischemia. Behav. Brain Res. 177, 282–289. [68] Liu, T. G., You, D. L., Valenzano, C., Sun, Y. H., Li, J. L., et al. (2006) Identification of NanE as the thioesterase for polyether chain release in nanchangmycin biosynthesis. Chem. Biol. 13, 945–955. [69] Harvey, B. M., Hong, H., Jones, M. A., Hughes-Thomas, Z. A., Goss, R. M., et al. (2006) Evidence that a novel thioesterase is responsible for polyketide chain release during biosynthesis of the polyether ionophore monensin. ChemBioChem 7, 1435–1442. [70] Kunishima, N., Asada, Y., Sugahara, M., Ishijima, J., Nodake, Y., et al. (2005) A novel induced-fit reaction mechanism of asymmetric hot dog thioesterase PAAI. J. Mol. Biol. 352, 212–228. [71] Cheng, Z., Song, F., Shan, X., Wei, Z., Wang, Y., et al. (2006) Crystal structure of human thioesterase superfamily member 2. Biochem. Biophys. Res. Commun. 349, 172–177. [72] Angelini, A., Cendron, L., Goncalves, S., Zanotti, G., and Terradot, L. (2008) Structural and enzymatic characterization of HP0496, a YbgC thioesterase from Helicobacter pylori. Proteins 72, 1212–1221. [73] Chen, M., Ma, X., Chen, X., Jiang, M., Song, H., et al. (2013) Identification of a hotdog fold thioesterase involved in the biosynthesis of menaquinone in Escherichia coli. J. Bacteriol. 195, 2768–2775. [74] Kotaka, M., Kong, R., Qureshi, I., Ho, Q. S., Sun, H. H., et al. (2009) Structure and catalytic mechanism of the thioesterase CalE7 in enediyne biosynthesis. J. Biol. Chem. 284, 15739–15749. [75] Liew, C. W., Sharff, A., Kotaka, M., Kong, R., Sun, H. H., et al. (2010) Induced-fit upon ligand binding revealed by crystal structures of the hotdog fold thioesterase in dynemicin biosynthesis. J. Mol. Biol. 404, 291–306.

595