Letter pubs.acs.org/OrgLett
MarH, a Bifunctional Enzyme Involved in the Condensation and Hydroxylation Steps of the Marineosin Biosynthetic Pathway Wanli Lu,† Papireddy Kancharla,† and Kevin A. Reynolds* Department of Chemistry, Portland State University, Portland, Oregon 97201, United States S Supporting Information *
ABSTRACT: A novel bifunctional enzyme, MarH, has been identified, and its key functional role in the marineosin biosynthesis successfully probed. MarH catalyzes (1) a condensation step between 4-methoxy-2,2′-bipyrrole-5-carboxaldehyde (MBC) and 2-undecylpyrrole (UP) to form undecylprodiginine (UPG) and (2) hydroxylation of the alkyl chain of UPG to form the (S)-23-hydroxyundecylprodiginine (HUPG), which is essential for MarG catalyzed bicyclization toward the formation of an unusual spiro-tetrahydropyran-aminal ring of marineosins. The final enigmatic steps in the marineosin biosynthesis have now been deciphered.
P
cluster directing the biosynthesis of undecylprodiginine (UPG) and streptorubin B in S. coelicolor A(3)2, but contains an additional gene marA (Figure 1a). Expression of the entire mar gene cluster in an S. venezuelae host (JND2) led to production of marineosins, whereas gene replacement of marG, which is predicted to encode a Rieske nonheme iron-dependent oxygenase, led to accumulation of (S)-23-hydroxyundecylprodiginine (HUPG), a key pathway intermediate, and a shunt product, 23-ketoundecylprodiginine.20 The functional role and specificity of MarG was further confirmed by feeding of the HUPG and its synthetic analogues to the S. venezuelae MarG overexpression strain, which led to production of premarineosins.20,21 Loss of marA, encoding a putative dehydrogenase, led to accumulation of premarineosin A and a shunt product, 16-ketopremarineosin A.20 These studies unequivocally disproved previously published biosynthetic hypotheses2,22 and suggested the last steps of marineosin biosynthesis occur with MarG catalyzed conversion of HUPG to premarineosins, and a final MarA catalyzed reduction (Figure 1b). The modified PG pathway, therefore, passes through a HUPG intermediate (Figure 1b). The striking similarity between the red and mar gene clusters provides no obvious clue as to the origin of the hydroxyl group of HUPG (Figure 1a).20,21 We initially hypothesized that hydroxylation may arise from expression of either marT or marE. Homologues of these are found in the red cluster, but in neither case is their function known. Deletion mutants of both genes in the JND2 strain were constructed to probe their functional roles. Low-level production of marineosins was observed in both JND2ΔT and JND2ΔE strains, demonstrating that while these proteins are involved in marineosin biosynthesis, they are not essential, nor responsible for the hydroxylation step (Figures S1, S2). We also
rodiginines (PGs) and marineosins (Figure 1) belong to a family of pyrrolylpyrromethene (PPM) alkaloids isolated from soil and marine bacteria.1,2 Marineosins differ from PGs in that they contain an unusual spiro-tetrahydropyran-aminal core fused with a 12-membered macrocyclic pyrrole and lack the aromatic ring conjugation between ring-B and ring-C (Figure 1b). The natural and synthetic PPM products are of worldwide interest because of their wide range of biological activities (antimalarial,3−5 immunosuppressive,6 antimicrobial,7,8 antitumor,7,9 and anticancer2,10) and modes of action.1,11,12 Notably, the synthetic PG analogue, GX15-070, has completed phase II clinical trials for the treatment of small cell lung cancer and is engaged in multiple clinical trials for the treatment of other cancer conditions.13,14 There has been significant interest and progress in understanding PGs and modified PGs biosynthetic pathways. A collaborative effort by the Challis, Reynolds, Bibb, and Smith groups has identified and characterized the functions of most of the enzymes encoded by the Streptomyces coelicolor red gene cluster responsible for undecylprodiginine (UPG) and streptorubin B biosynthesis (Figure 1).15−19 More recently, we have investigated the biosynthetic pathway of marineosin.20,21 Our motivation in part was the unusual spirotetrahydropyran-aminal core, which must arise via an intriguing departure from the standard PG pathway. The potent biological activity of the marineosins also provided a motivation for our work, as it might provide access to larger quantities of marineosins and structural analogues. Indeed, there have been numerous independent efforts to accomplish a total synthesis of marineosin, which for many years led only to synthesis of key fragments.22−27 A total synthesis of marineosin has now been reported.28 Our understanding of marineosin biosynthesis originated with the identification and characterization of the marineosin gene cluster (mar) from marine Streptomyces sp. CNQ-617.20 The mar cluster has a high degree of homology to the red gene © XXXX American Chemical Society
Received: January 10, 2017
A
DOI: 10.1021/acs.orglett.7b00093 Org. Lett. XXXX, XXX, XXX−XXX
Letter
Organic Letters
Figure 1. Biosynthetic gene clusters and pathways. (a) Comparison between PG (red) and marineosin (mar) biosynthetic gene clusters from S. coelicolor and Streptomyces sp. CNQ-617, respectively. (b) Biosynthetic pathway of marineosin, UPG, streptorubin B, and metacyloprodiginine via common precursors, UP and MBC.
leading to either UP or HUP. Feeding of the chemically synthesized UP and HUP to both the JND2ΔL and JND2ΔK strains led to production of marineosins (Figures S4, S5). These data provided strong evidence that the hydroxylation occurs after the MarK catalyzed reaction, and not at an earlier stage in the biosynthesis. All told this series of experiments indicated a role for MarH in hydroxylation. The marH gene product is predicted to be a phosphoenolpyruvate-utilizing enzyme with a molecular weight of 100 kDa and was originally proposed to catalyze the condensation of the HUP (rather than UP) with MBC to form HUPG.20,21 Thus, the enzyme which is proposed to catalyze an intermediary step between those catalyzed by MarK and MarG, in the marineosin biosynthetic pathway, must be a bifunctional enzyme: capable of both subunits condensation and alkyl chain hydroxylation. Attempts to obtain a pure soluble MarH protein in vitro were unsuccessful, and thus we undertook a genetic approach with chemical complementation to probe the role of MarH.30 We cloned and transferred pMarH-34S, a high-copy-number Streptomyces replicative plasmid containing marH under the control of the constitutive ermE* promoter, into S. venezuelae to give S. venezuelae/pMarH. Similarly, a S. venezuelae strain host with a redH overexpression plasmid pRedH-34S was generated to serve as control for feeding studies. Feeding of MBC and HUP to the S. venezuelae/pRedH and S. venezuelae/ pMarH strains produced a visible red pigment, and an LC-MS analysis of organic extract of the mycelia of both strains demonstrated the production of HUPG (Figure S6). The yield of HUPG from the S. venezuelae/pMarH strain is much higher
envisioned that there might be a cytochrome P450 monooxygenase (CYP) in or adjacent to the mar cluster responsible for the alkyl chain hydroxylation. Neither FramePlot nor antiSMASH analyses identified a candidate CYP gene in the mar gene cluster, which led us to clarify the precise boundaries of the mar gene cluster.20 The upstream boundary of the mar gene cluster was defined by replacement of the DNA fragment between the SuperCos1 backbone and marD with the spectinomycin-oriT cassette from pIJ778 using a PCR-targeting strategy.29 Similarly, the genes between marA and the SuperCos1 backbone were replaced to identify the downstream boundary. MS analyses of methanol extracts of both the JND2ΔUP and JND2ΔDOWN strains indicated the same phenotype as the JND2 strain (Figure S3). We also demonstrated that heterologous expression of the mar gene cluster in another PG nonproducing S. albus J1074 also produced the marineosins (Figure S3). Together this work supports the premise that the mar gene cluster as initially reported is complete and contains all genes responsible for encoding marineosin biosynthesis. We also explored the possibility that the hydroxyl group could be introduced at an initial stage in the pathway, potentially with a biosynthetic primer. Specifically, we wanted to test if the MarP initiates the (10′S)-hydroxyundecylpyrrole (HUP) formation from (S)-3hydroxybutyric acid-CoA starter unit instead of acetyl CoA as we had initially proposed.20,21 To test this hypothesis marL and marK genes were deleted separately from the mar gene cluster to generate JND2ΔL and JND2ΔK strains. These two gene products are responsible for the final steps of the pathway B
DOI: 10.1021/acs.orglett.7b00093 Org. Lett. XXXX, XXX, XXX−XXX
Letter
Organic Letters than that from the S. venezuelae/pRedH strain. These data were consistent with previous reports that condensing enzymes have relaxed substrate specificity.31−33 An LC-MS/MS analysis of feeding of the MBC and UP to the S. venezuelae/pMarH strain demonstrated production of both HUPG and UPG (Figure 2e
distinguish the biosynthetic pathways and are responsible for the differing structures. The striking similarity between the putative mar cluster and the red cluster and the commonality of the initial steps spurred us to test if replacement of the redH-G in PG producer S. coelicolor, by marH-G, would lead to production of marineosin products. To investigate this possibility, we adopted a gene swapping strategy. Since redH and redG are neighboring genes in the same operon, we deleted the redH gene, as well as redH-redG together, from the chromosome of S. coelicolor using a PCR-targeting mutagenesis method, resulting in non-UPG producing strains. Similarly, marH and marG are neighboring genes with a short linker sequence in between and an obvious ribosomal binding site for marG, so we cloned the pMarHG-34S plasmid, in which marH and marG are expressed in tandem under the control of the ermE* promoter. LC-MS analysis of fermentation broths of S. coelicolorΔredHG-pRedH, S. coelicolorΔredH-pRedH, S. coelicolorΔredHG-pMarH, and S. coelicolorΔredHG-pMarHG demonstrated the production of UPG, streptorubin B, HUPG, and premarineosin, respectively (Figure S9). This work demonstrates a combination of red and mar genes in S. coelicolor might represent an approach for a higher yield of marineosin and related analogues (versus heterologous expression of the entire mar cluster). More importantly these results unequivocally confirmed the parallel initial steps of the marineosin and PGs biosynthetic pathways (Figure 1b) and that MarH plays a bifunctional role in the marineosin biosynthesis. Sequence analysis of marH provides few clues as to how the conversion of UPG to HUPG is achieved. MarH has several functional homologues such as RedH from S. coelicolor, McpH from S. longispororuber, PigC from Serratia sp. ATCC 39006 and Serratia marcescens ATCC 274, and HapC (HCH_06026) from Hahella chejuensis KCTC 2396.19,33,34 Neighbor-joining phylogenetic tree showed that MarH falls into a different subclade (Figure S10). All these MarH homologues have two domains at the N- and C-terminal, share high similarity with the ATP-binding domain and phosphoryl transfer domains of a pyruvate phosphate dikinase (PPDK), and are proposed to be involved in the condensation. A third domain predicted to contain region binding to MBC and alkylpyrrole has also been proposed.19,31,34 Interestingly, protein BLAST analysis showed that MarH has an additional L-asparaginase-like domain (Figure 1a), which is unique among all MarH homologue proteins (Figures S11, S12). We posited that this domain might play a role in the key hydroxylation step, although an L-asparaginase domain would not be associated with such an activity. The presence of L-asparaginase and a pyruvate phosphate dikinase domain is unusual, and to the best of our knowledge, there are no reports either of multifunction proteins with two of these domains or of interactions between two proteins with this functionality. However, it has been our observation that bioinformatics does not always predict functions.35 To elucidate the function of this unprecedented domain, the amino acid sequence of MarH was aligned with the sequence of its homologues RedH and McpH to predict the putative catalytic sites in the L-asparaginase-like domain (Figures S12, S13). Five amino acids were selected for site-directed mutagenesis studies. A point-mutated marHG gene was generated and overexpressed in the S. coelicolorΔredHG strain. In contrast to the S. coelicolorΔredHG-pMarHG strain (Figure S14a), the predominant product of point-mutated marHG complementation strains was UPG (Figure S14b−f, and S15). Thus, the mutants were able to catalyze condensation but were impaired in the
Figure 2. LC-MS (EIC) profiles of feeding UP, MBC, and UPG to S. venezuelae/pMarH. EIC for m/z 393.5−394.5 and m/z 409.5−410.5, corresponding to [M + H]+ for UPG and HUPG, respectively. (a) Standard UPG; (b) standard HUPG; (c) extract of S. venezuelae fed with UP and MBC; (d) extract of S. venezuelae/pRedH fed with UP and MBC; (e) extract of S. venezuelae/pMarH fed with UP and MBC; (f) extract of S. venezuelae fed with UPG; (g) extract of S. venezuelae/ pRedH fed with UPG; (h) extract of S. venezuelae/pMarH fed with UPG. The bimodal peak is due to geometrical isomerism, resulting from rotation around the bond between ring-B and ring-C.
and Figure S7). Conversely, feeding of MBC and UP to the S. venezuelae/pRedH strain produced only UPG (Figure 2d). These data provided strong evidence that MarH is a bifunctional enzyme that catalyzes both the condensation and hydroxylation steps to form the HUPG in the marineosin biosynthetic pathway. To examine the ordering of these two steps, we separately fed UP and UPG to the S. venezuelae/ pMarH strain. LC-MS analyses of these feeding experiments demonstrated that the HUPG is generated from the UPG, while the HUP was not observed even in detectable levels from UP (Figure 2f−h, Figure S8). These findings further support the role of MarH in the hydroxylation step and that this occurs after the condensation of MBC and UP. Coupled with our previous work,20,21 these new data unambiguously demonstrate that the biosynthetic pathway and precursors of marineosins and PGs (UPG, streptorubin B, and metacycloprodiginine) are identical until the condensation of MBC and UP (Figure 1b). The enzymes, MarH, MarG, and MarA, in the marineosin pathway and RedH/McpH, and RedG/McpG in the PGs pathway (UPG, streptorubin B, and metacycloprodiginine), involved in the latter stages (Figure 1b) C
DOI: 10.1021/acs.orglett.7b00093 Org. Lett. XXXX, XXX, XXX−XXX
Letter
Organic Letters
(9) Williams, R. P.; Hearn, W. R. Antibiotics 1967, 2, 410. (10) Regourd, J.; Ali, A. A.; Thompson, A. J. Med. Chem. 2007, 50, 1528. (11) Marchal, E.; Rastogi, S.; Thompson, A.; Davis, J. T. Org. Biomol. Chem. 2014, 12, 7515. (12) Rastogi, S.; Marchal, E.; Uddin, I.; Groves, B.; Colpitts, J.; McFarland, S. A.; Davis, J. T.; Thompson, A. Org. Biomol. Chem. 2013, 11, 3834. (13) Trudel, S.; Li, Z. H.; Rauw, J.; Tiedemann, R. E.; Wen, X. Y.; Stewart, A. K. Blood 2007, 109, 5430. (14) Nguyen, M.; Marcellus, R. C.; Roulston, A.; Watson, M.; Serfass, L.; Murthy Madiraju, S. R.; Goulet, D.; Viallet, J.; Bélec, L.; Billot, X.; Acoca, S.; Purisima, E.; Wiegmans, A.; Cluse, L.; Johnstone, R. W.; Beauparlant, P.; Shore, G. C. Proc. Natl. Acad. Sci. U. S. A. 2007, 104, 19512. (15) Mo, S.; Kim, B. S.; Reynolds, K. A. Chem. Biol. 2005, 12, 191. (16) Mo, S.; Sydor, P. K.; Corre, C.; Alhamadsheh, M. M.; Stanley, A. E.; Haynes, S. W.; Song, L.; Reynolds, K. A.; Challis, G. L. Chem. Biol. 2008, 15, 137. (17) Whicher, J. R.; Florova, G.; Sydor, P. K.; Singh, R.; Alhamadsheh, M.; Challis, G. L.; Reynolds, K. A.; Smith, J. L. J. Biol. Chem. 2011, 286, 22558. (18) Challis, G. L. J. Ind. Microbiol. Biotechnol. 2014, 41, 219. (19) Hu, D. X.; Withall, D. M.; Challis, G. L.; Thomson, R. J. Chem. Rev. 2016, 116, 7818. (20) Salem, S. M.; Kancharla, P.; Florova, G.; Gupta, S.; Lu, W.; Reynolds, K. A. J. Am. Chem. Soc. 2014, 136, 4565. (21) Kancharla, P.; Lu, W.; Salem, S. M.; Kelly, J. X.; Reynolds, K. A. J. Org. Chem. 2014, 79, 11674. (22) Cai, X. C.; Wu, X.; Snider, B. B. Org. Lett. 2010, 12, 1600. (23) Aldrich, L. N.; Dawson, E. S.; Lindsley, C. W. Org. Lett. 2010, 12, 1048. (24) Cai, X. C.; Snider, B. B. J. Org. Chem. 2013, 78, 12161. (25) Panarese, J. D.; Konkol, L. C.; Berry, C. B.; Bates, B. S.; Aldrich, L. N.; Lindsley, C. W. Tetrahedron Lett. 2013, 54, 2231. (26) Li, G.; Zhang, X.; Li, Q.; Feng, P.; Shi, Y. Org. Biomol. Chem. 2013, 11, 2936. (27) Aldrich, L. N.; Berry, C. B.; Bates, B. S.; Konkol, L. C.; So, M.; Lindsley, C. W. Eur. J. Org. Chem. 2013, 2013, 4215. (28) Xu, B.; Li, G.; Li, J.; Shi, Y. Org. Lett. 2016, 18, 2028. (29) Gust, B.; Challis, G. L.; Fowler, K.; Kieser, T.; Chater, K. F. Proc. Natl. Acad. Sci. U. S. A. 2003, 100, 1541. (30) Sydor, P. K.; Challis, G. L. Methods Enzymol. 2012, 516, 195. (31) Haynes, S. W.; Sydor, P. K.; Stanley, A. E.; Song, L.; Challis, G. L. Chem. Commun. 2008, 1865. (32) Chawrai, S. R.; Williamson, N. R.; Salmond, G. P.; Leeper, F. J. Chem. Commun. 2008, 1862. (33) Chawrai, S. R.; Williamson, N. R.; Mahendiran, T.; Salmond, G. P. C.; Leeper, F. J. Chem. Sci. 2012, 3, 447. (34) Williamson, N. R.; Fineran, P. C.; Leeper, F. J.; Salmond, G. P. Nat. Rev. Microbiol. 2006, 4, 887. (35) Palaniappan, N.; Dhote, V.; Ayers, S.; Starosta, A. L.; Wilson, D. N.; Reynolds, K. A. Chem. Biol. 2009, 16, 1180.
hydroxylation, supporting our hypothesis that the L-asparaginase-like domain is involved in the hydroxylation. To the best of our knowledge the hydroxylation on an alkyl chain of the UPG by this domain of MarH is unprecedented and might suggest a novel enzymatic hydroxylation mechanism. The Lasparaginase-like domain likely plays a role as a protein interaction interface that recruits an unknown oxygenase to catalyze the hydroxylation reaction. Further MarH characterization studies are required to conclusively establish its exact role in the hydroxylation step. In conclusion, through a series of systematic studies, we have unambiguously demonstrated that the MarH is a novel bifunctional enzyme and is responsible for catalyzing both the condensation of MBC and UP, and subsequent hydroxylation of UPG to generate the HUPG, which is essential for the formation of unusual spiro-tetrahydropyran-aminal ring. The final enigmatic steps of the unusual modified PG pathway, which provides for a spiroaminal ring, are now deciphered.
■
ASSOCIATED CONTENT
* Supporting Information S
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.orglett.7b00093. Experimental procedures, Tables giving cosmids, strains, primers, and LC-MS gradient, and figures giving MS, MS/MS, and LC-MS profiles (PDF)
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. ORCID
Wanli Lu: 0000-0001-8449-3129 Papireddy Kancharla: 0000-0002-6381-1236 Kevin A. Reynolds: 0000-0003-0270-9809 Author Contributions †
W.L. and P.K.contributed equally.
Notes
The authors declare no competing financial interest.
■ ■
ACKNOWLEDGMENTS This work was supported by a grant from the National Institutes of Health (GM077147). REFERENCES
(1) Fürstner, A. Angew. Chem., Int. Ed. 2003, 42, 3582. (2) Boonlarppradab, C.; Kauffman, C. A.; Jensen, P. R.; Fenical, W. Org. Lett. 2008, 10, 5505. (3) Castro, A. J. Nature 1967, 213, 903. (4) Papireddy, K.; Smilkstein, M.; Kelly, J. X.; Shweta; Salem, S. M.; Alhamadsheh, M.; Haynes, S. W.; Challis, G. L.; Reynolds, K. A. J. Med. Chem. 2011, 54, 5296. (5) Kancharla, P.; Kelly, J. X.; Reynolds, K. A. J. Med. Chem. 2015, 58, 7286. (6) D’Alessio, R.; Bargiotti, A.; Carlini, O.; Colotta, F.; Ferrari, M.; Gnocchi, P.; Isetta, A.; Mongelli, N.; Motta, P.; Rossi, A.; Rossi, M.; Tibolla, M.; Vanotti, E. J. Med. Chem. 2000, 43, 2557. (7) Boger, D. L.; Patel, M. J. Org. Chem. 1988, 53, 1405. (8) Marchal, E.; Uddin, M. I.; Smithen, D. A.; Hawco, C. L. A.; Lanteigne, M.; Overy, D. P.; Kerr, R. G.; Thompson, A. RSC Adv. 2013, 3, 22967. D
DOI: 10.1021/acs.orglett.7b00093 Org. Lett. XXXX, XXX, XXX−XXX
