Letter Cite This: Org. Lett. XXXX, XXX, XXX−XXX
pubs.acs.org/OrgLett
Biochemical Characterization of a Multifunctional Mononuclear Nonheme Iron Enzyme (PtlD) in Neopentalenoketolactone Biosynthesis Qian Deng,†,# Yang Liu,†,# Linyue Chen,† Meiling Xu,‡ Nathchar Naowarojna,‡ Norman Lee,‡ Li Chen,†,‡ Dongqing Zhu,† Xuechuan Hong,† Zixin Deng,† Pinghua Liu,*,‡ and Changming Zhao*,† Downloaded via CARLETON UNIV on September 6, 2019 at 15:31:46 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
†
Key Laboratory of Combinatory Biosynthesis and Drug Discovery (Wuhan University), Ministry of Education, School of Pharmaceutical Sciences, Wuhan University, Hubei 430072, People’s Republic of China ‡ Department of Chemistry, Boston University, 590 Commonwealth Avenue, Boston, Massachusetts 02215, United States S Supporting Information *
ABSTRACT: Pentalenolactone is a microbial sesquiterpenoid with antibiotic activity. Its biosynthetic pathway was elucidated by a combination of genetic and biochemical characterizations of all genes involved. For the related neopentalenoketolactone biosynthetic gene cluster from Streptomyces avermitilis, an α-ketoglutarate-dependent mononuclear nonheme iron enzyme, PtlD, was proposed to catalyze both desaturation and olefin epoxidation reactions. Yet, these activities remained to be validated by in vitro biochemical evidence. In this report, we demonstrated that PtlD has multiple activities, including hydroxylation, desaturation, and epoxidation, and confirmed the presence of the elusive epoxide intermediate in a neopentalenoketolactone pathway. rearrangement (9 → 10) through a cation intermediate.13,14 For the neopentalenoketolactone pathway, the proposed oxidation of neopentalenolactone D (11) to neopentalenolactone F 13 transformation (Scheme 1) has only been partially characterized. Notably, PtlD was suggested to catalyze the 11 → 16 transformation, while the proposed intermediate 13 has not yet been confirmed. PtlD could also catalyze oxidation of the isomeric substrates of the pentalenolactone pathway, the 7 → 8 → 9 sequential transformations.9 In the S. avermitilis ΔptlD mutant, neopentalenolactone D (11) accumulates, which implies that neopentalenolactone D (11) is most likely the PtlD substrate. Because of the high degree of similarity between the neopentalenoketolactone and pentalenolactone biosynthetic pathways and the demonstrated activities of PenD/PntD (Scheme 1), PtlD in the neopentalenoketolactone biosynthetic pathway was proposed to be the enzyme responsible for the sequential transformation 11 → 12 → 13 (Scheme 1).9 From S. avermitilis culture, compounds 14 and 16 have been isolated and characterized.17
P
entalenolactone (10, Scheme 1) is a sesquiterpenoid antibiotic isolated from many species of Streptomyces.1 Using its electrophilic epoxylactone moiety, pentalenolactone can alkylate an active site cysteine of glyceraldehyde-3phosphate dehydrogenase.2−5 To counter this effect, pentalenolactone producer contains an inducible pentalenolactoneinsensitive glyceraldehyde-3-phosphate dehydrogenase.6−8 Over the two decades, pentalenolactone biosynthetic gene clusters from S. exfoliatus (pen cluster) and S. arenae (pnt cluster) have been identified and all of the steps were biochemically characterized in vitro (1 → 10 transformations, Scheme 1).9−16 The closely related neopentalenoketolactone (14) biosynthetic gene cluster (ptl cluster) from S. avermitilis shares a high degree of similarity to that of pentalenolactone (10).17 These two pathways diverge at the point of the flavin-dependent Baeyer− Villiger monooxygenase-catalyzed reactions (PtlE in neopentalenoketolactone and PenE/PntE in pentalenolactone biosynthesis, Scheme 1).17,18 In the pentalenolactone pathway, after the Baeyer−Villiger reaction (6 → 7 transformation), an α-ketoglutarate (α-KG)dependent mononuclear nonheme iron enzyme (PenD/PntD) catalyzes the subsequent two steps (7 → 8 → 9, Scheme 1).9 A Cytochrome P450 enzyme then catalyzes the oxidative © XXXX American Chemical Society
Received: August 13, 2019
A
DOI: 10.1021/acs.orglett.9b02872 Org. Lett. XXXX, XXX, XXX−XXX
Letter
Organic Letters Scheme 1. Proposed Pentalenolactone (10) and Neopentalenoketolactone (14) Biosynthetic Pathways
Figure 1. HPLC traces of PtlD reactions using neopentalenolactone D (11) or neopentalenolactone E (12) as the substrate: (i) neopentalenolactone D (11); (ii) neopentalenolactone D (11) as the substrate, with 1 equiv of α-KG; (iii) neopentalenolactone D (11) as the substrate, with 2 equiv of α-KG; (iv) neopentalenolactone E (12); and (v) neopentalenolactone E (12) as the substrate, with 1 equiv of αKG.
Figure S4 in the Supporting Information). Interestingly, compound 12 was stable under our assay conditions and no obvious hydrolysis of 12 to 15 by PtlD catalysis or through a nonenzymatic pathway was observed (see Figure S5 in the Supporting Information), which suggests that compound 15 is probably the hydroxylation product of compound 11, instead of being the hydrolysis product of compound 12. To provide further evidence for this hypothesis, we examined the oxygen source of compound 15 by oxygen incorporation experiments. The predicted m/z for compound 15 in the positive-ion mode is 303.1203 ([M + Na]+). If there is [18O] incorporation, the expected m/z of product 15b would be 305.1245 ([M + Na]+). The observed m/z of 15b in the reaction under 18O2 gas atmosphere is 305.1247 and the ratio between 15 and 15b is 48.3:51.7 (see Figure S6A in the Supporting Information). It indicates that molecular oxygen could be incorporated into compound 15. The relatively lower incorporation efficiency might be due to an oxygen exchange between FeIV=18O and H216O. Consistent with this prediction, when the PtlD reaction was performed under a 16O2 atmosphere in a H218O buffer, the ratio between 15 and 15b is 50.5:49.5 (Figure S6B in the Supporting Information). We have also characterized the PtlD reaction with 2 equiv of α-KG, relative to that of substrate 11. In this reaction, three new products (13, 14, and 17 in trace iii, Figure 1) were observed while neopentalenolactone E (12) disappeared. This result suggested that the excess α-KG allows PtlD to catalyze a second round of catalysis, specifically the oxidation of neopentalenolactone E (12), to afford three other products (13, 14, and 17). To test this hypothesis, we performed the PtlD reaction using neopentalenolactone E (12) as the substrate with 1 equiv of αKG. The HPLC profile (trace v, Figure 1) clearly indicated that compounds 13, 14, and 17 are derived from neopentalenolactone E (12). Compounds 14 and 17 were purified and characterized by 1H NMR, 13C NMR, 2D-NMR, and highresolution mass spectrometry (Figures S7 and S8 in the Supporting Information). Compound 14 is neopentalenoketolactone and compound 17 is a methyl ester form of compound 16. In this study, we also noticed that the ratio between compounds 13 and 14 varies from time to time (e.g., trace iii vs trace v in Figure 1). This phenomenon led us to suspect that compound 13 might be an unstable intermediate, and, indeed, we did encounter difficulties in isolating this compound for spectroscopic characterizations. Compound 13 is unstable and decomposes during the purification process to produce
In vitro, when neopentalenolactone D (11) was used as the PtlD substrate, compounds 12 and 15 were detected as the products, while neither the proposed epoxide intermediate neopentalenolactone F (13) nor the rearranged compound neopentalenoketolactone (14) was confirmed.9,17 The differences in product profiles between the in vivo results and the recombinant PtlD reaction warrant further examination of this reaction. In this work, we demonstrated three PtlD activities in vitro, including hydroxylation, desaturation, and epoxidation reactions. To prevent the oxidative inactivation of the labile iron center, PtlD was purified anaerobically and the excess Fe2+ was removed at the end of the affinity column chromatography (see Figure S1 in the Supporting Information). Following the literature protocols,19 we generated a S. avermitilis ΔptlD/ΔSAV_7469 double mutant. After several chromatographic steps, a 2-L culture of this mutant provided us with ∼60 mg of neopentalenolactone D (11; see Figure S2 in the Supporting Information). PtlD has been proposed to catalyze two sequential reactions: desaturation (11 → 12) and epoxidation (12 → 13) reactions, while the epoxidation reactivity has never been successfully demonstrated. We first examined how the amount of α-KG affects the PtlD product profiles. In the reaction with 1 equiv of α-KG relative to the substrate 11, there were two products (trace ii, Figure 1). The product with a retention time of 32.5 min (12, Figure 1) is the major product and the one with a retention time of 14.2 min (15, Figure 1) is the minor product. Using LC/MS, the m/z values of these two compounds are consistent with those of neopentalenolactone E (12) and its lactone ring-opened, keto-acid product (15). Characterization by 1H NMR, 13C NMR, and 2D-NMR confirmed the identities of these compounds as neopentalenolactone E (12; see Figure S3 in the Supporting Information) and compound 15 (see B
DOI: 10.1021/acs.orglett.9b02872 Org. Lett. XXXX, XXX, XXX−XXX
Letter
Organic Letters
disappeared and compounds 14 and 17 were the final products (trace v, Figure 2A). These results imply that compounds 14 and 17 are derived from compound 13. PtlD reaction time-course profiles clearly indicated that compound 13 is only stable for a few hours (Figure 2). Because the PtlD homologues in pentalenolactone biosynthesis (PenD/ PntD) are bifunctional enzymes (desaturation and epoxidation),9 the time course of PtlD reaction in Figure 2A and the high-resolution mass spectrometry data (m/z of 279.1227 in positive ion mode, Figure S11 in the Supporting Information) led us suspect that compound 13 is most likely the proposed unstable epoxide intermediate neopentalenolactone F (13; see Scheme 1). Since epoxide protons have a characteristic chemical shift in 1H NMR, we decided to monitor this reaction directly using 1H NMR (Figure 2B). We ran the PtlD reaction in 50 mM KPi buffer prepared using D2O, which allowed us to directly acquire the 1H NMR spectra at various time points. In the first 2 h, a characteristic AB quartet was observed (δ 3.11 ppm, d, J = 3.5 Hz and δ 3.08 ppm, d, J = 3.5 Hz, highlighted in the box of Figure 2B). These are consistent with the properties of germinal epoxide protons. In the next 16 h, this set of signals in 1H NMR slowly disappeared, and their formation and decay in 1H NMR (Figure 2B) correlates with the HPLC profile time course in Figure 2A. Results from these two sets of experiments (Figures 2A and 2B) and the high-resolution mass spectrometry information (Figure S11) are consistent with the presence of an unstable epoxide intermediate 13 in the neopentalenoketolactone biosynthesis (Scheme 1). For α-KG-dependent mononuclear nonheme iron enzymes, hydroxylation is the most common type of reaction.20−22 In recent years, many other types of transformations have also been reported, including epoxidation,23 chlorination,24 epimerization,25 C−C bond cleavage,26 and endoperoxidation.27 There were also examples of multifunctional α-KG-dependent mononuclear nonheme iron enzymes. For example, the clavulanic acid synthase catalyzes three sequential reactions: hydroxylation, oxidative ring closure, and desaturation.28−30 In the biosynthesis of the viridicatin-type alkaloid, the AsqJ enzyme catalyzes two sequential oxidations: desaturation to form a C C bond and an epoxidation reaction.31−35 It is generally believed that the high valent species FeIV=O is the key intermediate in these enzymes.36−39 However, the FeIV=O species may exchange its oxygen with H 2 O, which results in the incorporation of oxygen into the product from water. After demonstrating the existence of an epoxide intermediate 13 in neopentalenoketolactone biosynthesis, we have also examined the source of the epoxide oxygen using two sets of experiments: PtlD reactions under either 18O2 gas atmosphere or in the H218O buffer. The predicted m/z for compound 13 in the negative ion mode is 277.1081 and the observed signal was m/z of 277.1083 (Figure 3). If there is [18O]-incorporation, the expected m/z of product 13b would be 279.1124. The observed m/z of 13b in H218O buffer reaction is 279.1125 and the ratio between 13 and 13b is 96.3:3.7 (Figure 3A), which indicates that the epoxide oxygen in compound 13 is from molecular oxygen (O2) and after FeIV=O species is produced, there is a very low level of exchange between FeIV=16O and H218O. To provide additional evidence to support this conclusion, we have also conducted PtlD under the 18O2 atmosphere in a H216O buffer. In this case, the dominant signal is 13b with m/z of 279.1125 (Figure 3B). The ratio between 13 and 13b in this reaction is 8.9:91.1, establishing that the epoxide oxygen in compound 13 is indeed derived from molecular oxygen. The low level of
compound 16 as the major product (see Figure S9 in the Supporting Information), which had been isolated from the fermentation broth of S. avermitilis previously.17 After these initial characterizations, we have also kinetically characterized PtlD by monitoring the oxygen consumption rates using a Neofox oxygen electrode. Because of the low Km value (
