Aminofutalosine Synthase: Evidence for Captodative and Aryl Radical

Communication pubs.acs.org/JACS

Aminofutalosine Synthase: Evidence for Captodative and Aryl Radical Intermediates Using β‑Scission and SRN1 Trapping Reactions Sumedh Joshi,† Nilkamal Mahanta,† Dmytro Fedoseyenko,† Howard Williams, and Tadhg P. Begley* Department of Chemistry, Texas A&M University, College Station, Texas 77842, United States S Supporting Information *

ABSTRACT: Aminofutalosine synthase (MqnE) is a radical SAM enzyme involved in the menaquinone biosynthetic pathway. In this communication, we propose a novel mechanism for this reaction involving the addition of the adenosyl radical to the substrate double bond to form a captodative radical followed by rearrangement and decarboxylation to form an aryl radical anion which is then oxidized by the [4Fe−4S]+2 cluster. Consistent with this proposal, we describe the trapping of the captodative radical and the aryl radical anion using radical triggered C−Br fragmentation reactions. We also describe the trapping of the captodative radical by replacing the vinylic carboxylic acid with an amide.

Figure 1. Mechanistic proposal for the aminofutalosine synthasecatalyzed conversion of 3-[(1-carboxyvinyl)oxy]benzoic acid 1 to aminofutalosine 8 (A=Adenine).

HPLC assay, were kcat = 0.22 ± 0.08 min−1, Km(1) = 335 ± 98 μM, and Km(SAM) = 13.3 ± 0.9 μM (Figure S3). The slow kcat is consistent with most other radical SAM enzymes.16 Our first strategy for trapping the captodative radical 3 was based on the rapid homolysis of a C−Br bond β to a radical center (kC−Br = 3 × 108 s−1, Figure 2).17 When the brominated

M

enaquinone (vitamin K2) is a lipid-soluble quinone that participates in the bacterial electron transport chain.1 In mammalian cells, vitamin K is essential for the carboxylation of glutamic acid residues in proteins involved in blood clotting and bone morphogenesis.2,3 Two biosynthetic pathways to menaquinone have been identified. The canonical pathway uses carbanion chemistry to assemble the naphthoquinone moiety,4 while a recently discovered pathway uses radical chemistry.5−7 We have previously reported the successful in vitro reconstitution of aminofutalosine 8 formation from chorismate and demonstrated that this complex transformation is mediated by two enzymes: chorismate dehydratase (MqnA) and a remarkable radical SAM enzyme - aminofutalosine synthase (MqnE).6 Our current mechanistic proposal for aminofutalosine synthase (MqnE) is outlined in Figure 1. In this proposal, the 5′-deoxyadenosyl radical 2 adds to 3-[(1-carboxyvinyl)oxy]benzoic acid 1 to form the captodative radical 3. Rearrangement of 3, via the spiroepoxide 4, generates 5. Analogous Oneophyl rearrangements have been previously reported in the synthetic chemistry literature.8−10 Decarboxylation followed by deprotonation to form the aryl radical anion 7 and an electron transfer to the [4Fe−4S]+2 cluster completes the reaction. Aryl radical anions have been proposed in anaerobic metabolism but not directly characterized.11−15 In this communication, we provide the experimental support for this proposal by describing the successful trapping of the captodative radical 3 and the aryl radical anion 7. Expression of the His-tagged Thermus thermophilus MqnE in Escherichia coli gave a catalytically active (∼25 turnovers) enzyme (Figures S1−S2). The Michaelis−Menten constants for the reaction at 25 °C, determined using a discontinuous © 2017 American Chemical Society

Figure 2. Strategy for trapping the captodative radical using the facile fragmentation of a C−Br bond β to a radical center.

substrate analog 9, synthesized as shown in Figures S4−S7, was incubated with MqnE, SAM, and dithionite, HPLC analysis of the resulting reaction mixture showed the formation of a new compound eluting at 15.5 min (Figures 3, S8). This compound was identified as 12 by NMR and MS analysis (Figures S9− S14). MS−MS fragmentation analysis of the minor reaction products also demonstrated the formation of a small quantity of the isomeric adenosyl radical addition product 13 eluting at 15.75 min (Figures S15−S16). Analogous addition products have been reported for NosL.18 Rearranged products resulting from the spirocyclization reaction were not observed. This Received: April 25, 2017 Published: July 12, 2017 10952

DOI: 10.1021/jacs.7b04209 J. Am. Chem. Soc. 2017, 139, 10952−10955

Communication

Journal of the American Chemical Society

Figure 3. HPLC chromatogram of the MqnE-catalyzed reaction with 9. Control reactions run in the absence of dithionite, 9, or enzyme did not show the formation of 12. Small quantities of the isomeric addition product 13 were also formed.

suggests that kspiro < 1.5 × 107 s−1 (5% detection limit for the rearranged product). To obtain a better estimate of kspiro, we also tested the analogous chloro-substituted substrate (Figures S17−S20), which is less prone to the β-scission reaction (kC−Cl = 4 × 106 s−1).17 HPLC and LCMS analysis of the resulting reaction mixture revealed 12 as the major product. No rearranged products were detected. A small amount of the isomeric radical addition product was also observed with the chloro-substituted substrate analog (Figures S21−S23). From this experiment, we can estimate that kspiro < 2 × 105 s−1 (5% detection limit). This is an upper limit because the C−Cl fragmentation rate has not been corrected for captodative stabilization or for conformational effects at the active site that might create a stereoelectronic barrier to the β-scission reaction. We anticipate that the spirocyclization rate will be relatively slow because it involves dearomatization of the benzene ring and formation of a strained intermediate. It is also likely that formation of the spirocyclic intermediate is energetically unfavorable and that the reaction is driven by the facile decarboxylation of 5. Our mechanistic proposal suggests that the substrate vinylic carboxylic acid may function as the acid catalyst for the ring opening of the spiroepoxide 4. This suggests that replacing the vinylic carboxylic acid of 1 with the corresponding amide might block the conversion of 4 to 5 and result in intermediate trapping. Amide analog 14 was synthesized as shown in Figures S24− S28 and treated with MqnE under standard reaction conditions. HPLC and LCMS analysis of the resulting reaction mixture indicated the formation of a single major product (Figure 4) with m/z: 457.1479, consistent with the trapping of the captodative radical, 15 (Figure 5). This was confirmed by the complete NMR characterization of 19 (Figures S29−S36). Running the reaction in 95% D 2 O containing buffer demonstrated that the hydrogen atom is derived from an exchangeable position on MqnE (Figure 4B and C). A mechanistic proposal for the formation of 19 is outlined in Figure 5. We propose that 15 undergoes reversible spirocyclization to 16. However, because the amide is much less acidic than the carboxylic acid in the native substrate 1, ring opening of the spiroepoxide 16 is retarded sufficiently to allow off path trapping of the captodative radical by hydrogen atom transfer. The next set of experiments were targeted toward detection of the aryl radical anion 7 using the propensity of halogenated aryl radical anions to undergo substitution/hydrogen atom abstraction reactions. An example of such a reaction from the

Figure 4. MqnE-catalyzed reaction of the substrate amide analog 14. (A) HPLC chromatogram of the reaction mixture. (B) MS of the major reaction product eluting at 18.75 min. (C) MS of the product for the reaction run in 95% D2O containing buffer.

Figure 5. Mechanistic proposal for the MqnE catalyzed reaction of the amide analog 14.

organic chemistry literature, for a closely related compound, is shown in Figure 6. Electron transfer from the SOMO of the

Figure 6. Radical substitution mechanism of aryl halide radical anions (SRN1).

aryl ring to the C−Br σ* orbital weakens the C−Br bond resulting in rapid loss of bromide to give the aryl radical 21 which can then react with a nucleophile to give 22 (reverse of the fragmentation reaction) or undergo hydrogen atom transfer from a suitable donor to form 23.19,20 Based on this chemistry, we hypothesized that bromo analog 24 should enable us to probe for the aryl radical anion 7 proposed in Figure 1. Compound 24 was synthesized as shown in Figures S37− S41 and treated with MqnE under standard reaction conditions. HPLC and LCMS analysis of the resulting reaction mixture indicated the formation of two major products (Figures 7, S42) with masses that were consistent with those expected for aminofutalosine 8 and bromo-aminofutalosine 26 (Figures S43−S45). This was confirmed by the complete NMR characterization of 26 (Figures S46−S51) and the coelution of the signal at 20.1 min with authentic aminofutalosine 8 10953

DOI: 10.1021/jacs.7b04209 J. Am. Chem. Soc. 2017, 139, 10952−10955

Communication

Journal of the American Chemical Society

considerable potential for the detection of radical intermediates in enzyme-catalyzed reactions.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.7b04209. Detailed procedures for the overexpression and purification of MqnE; synthetic procedures and spectroscopic characterization of compounds 9, 39 (SI), 14, and 24; NMR and MS characterization of compounds 12, 13, 19, 8, and 26 (PDF)

■

Figure 7. HPLC analysis of the MqnE-catalyzed reaction with 24.

generated in the MqnE reaction with the native substrate 1 (Figure S52). These results are consistent with the formation of an aryl radical anion intermediate 25 (Figure 8). An alternative

AUTHOR INFORMATION

Corresponding Author

*[email protected] ORCID

Sumedh Joshi: 0000-0002-0407-2920 Author Contributions †

S.J., N.M., and D.F. contributed equally.

Notes

The authors declare no competing financial interest.

■ Figure 8. Mechanistic proposal for the MqnE-catalyzed rearrangement of the brominated substrate analog 24.

ACKNOWLEDGMENTS This research was supported by a grant from the National Science Foundation (1507191) and by the Robert A. Welch Foundation (A0034).

mechanism involving protonation of the aryl radical anion (25 to 27 in Figure 8) was excluded by running the reaction in 95% deuterated buffer. Under these conditions, 65% of the aminofutalosine 8 was nondeuterated (Figure S43) consistent with the formation of 8 by hydrogen atom transfer from a nonacidic center rather than proton transfer from an exchangeable acidic center. Our mechanistic proposal indicates that the MqnE reaction is completed by the oxidation of the aryl radical anion 7 by the [4Fe−4S]2+ cluster. This implies that there is no net consumption of reducing equivalents during the reaction. This was verified by the observation of multiple enzyme turnovers in the presence of stoichiometric amounts of sodium dithionite (Figures S53−S54).21 In summary, we have described three key experiments that support the mechanistic proposal for aminofutalosine synthase outlined in Figure 1. Bromo-analog 9 used the fragmentation of a C−Br bond β to a radical center to trap the initially formed captodative radical 10. While radical addition to double bonds has abundant precedent in organic synthesis, the addition of the adenosyl radical to a substrate double bond was unprecedented in biosynthesis. Addition of the 5′-deoxyadenosyl radical to a dehydroalanine-containing pyruvate formate lyase variant22 and the inactivation of glutamate mutase (B12-dependent) by 2methylene glutarate23 are the closest examples of related chemistry in engineered systems. Analog 14 blocked the decarboxylation reaction resulting in the trapping of the captodative radical by a hydrogen atom transfer reaction. Bromo-analog 24 enabled the detection of the aryl radical anion intermediate by an SRN1 type C−Br fragmentation reaction. These studies also demonstrate that the small, synthetically accessible C−Br bond is a highly sensitive radical probe with

(1) Nowicka, B.; Kruk, J. Biochim. Biophys. Acta, Bioenerg. 2010, 1797 (9), 1587−1605. (2) Doolittle, R. F. J. Innate Immun. 2016, 8 (1), 23−29. (3) Neve, A.; Corrado, A.; Cantatore, F. P. J. Cell. Physiol. 2013, 228 (6), 1149−1153. (4) Meganathan, R. Vitam. Horm. 2001, 61, 173−218. (5) Hiratsuka, T.; Furihata, K.; Ishikawa, J.; Yamashita, H.; Itoh, N.; Seto, H.; Dairi, T. Science 2008, 321 (5896), 1670−1673. (6) Mahanta, N.; Fedoseyenko, D.; Dairi, T.; Begley, T. P. J. Am. Chem. Soc. 2013, 135 (41), 15318−15321. (7) Cooper, L. E.; Fedoseyenko, D.; Abdelwahed, S. H.; Kim, S.-H.; Dairi, T.; Begley, T. P. Biochemistry 2013, 52 (27), 4592−4594. (8) Bietti, M.; Calcagni, A.; Cicero, D. O.; Martella, R.; Salamone, M. Tetrahedron Lett. 2010, 51 (31), 4129−4131. (9) Baroudi, A.; Alicea, J.; Flack, P.; Kirincich, J.; Alabugin, I. V. J. Org. Chem. 2011, 76 (6), 1521−1537. (10) Baroudi, A.; Flack, P.; Alabugin, I. V. Chem. - Eur. J. 2010, 16 (41), 12316−12320 and S12316/1−S12316/40. (11) Payne, K. A.; Quezada, C. P.; Fisher, K.; Dunstan, M. S.; Collins, F. A.; Sjuts, H.; Levy, C.; Hay, S.; Rigby, S. E.; Leys, D. Nature 2014, 517 (7535), 513−516. (12) Weinert, T.; Huwiler, S. G.; Kung, J. W.; Weidenweber, S.; Hellwig, P.; Stärk, H.-J.; Biskup, T.; Weber, S.; Cotelesage, J. J.; George, G. N. Nat. Chem. Biol. 2015, 11 (8), 586−591. (13) Boll, M.; Fuchs, G.; Heider, J. Curr. Opin. Chem. Biol. 2002, 6 (5), 604−611. (14) Tiedt, O.; Mergelsberg, M.; Boll, K.; Müller, M.; Adrian, L.; Jehmlich, N.; von Bergen, M.; Boll, M. mBio 2016, 7 (4), e00990-16. (15) Schmid, G.; Auerbach, H.; Pierik, A. J.; Schünemann, V.; Boll, M. Biochemistry 2016, 55 (39), 5578−5586. (16) Broderick, J. B.; Duffus, B. R.; Duschene, K. S.; Shepard, E. M. Chem. Rev. 2014, 114 (8), 4229−4317. (17) Wagner, P. J.; Lindstrom, M. J.; Sedon, J. H.; Ward, D. R. J. Am. Chem. Soc. 1981, 103 (13), 3842−9.

■

10954

REFERENCES

DOI: 10.1021/jacs.7b04209 J. Am. Chem. Soc. 2017, 139, 10952−10955

Communication

Journal of the American Chemical Society (18) Bhandari, D. M.; Fedoseyenko, D.; Begley, T. P. J. Am. Chem. Soc. 2016, 138 (50), 16184−7. (19) Rossi, R. A.; Pierini, A. B.; Peñeń ̃ory, A. B. Chem. Rev. 2003, 103 (1), 71−168. (20) Tanner, D. D.; Chen, J. J.; Chen, L.; Luelo, C. J. Am. Chem. Soc. 1991, 113 (21), 8074−8081. (21) Ruszczycky, M. W.; Choi, S.-h.; Liu, H.-w. J. Am. Chem. Soc. 2010, 132 (7), 2359−2369. (22) Wagner, A. V.; Demand, J.; Schilling, G.; Pils, T.; Knappe, J. Biochem. Biophys. Res. Commun. 1999, 254 (2), 306−310. (23) Huhta, M. S.; Ciceri, D.; Golding, B. T.; Marsh, E. N. G. Biochemistry 2002, 41 (9), 3200−3206.

10955

DOI: 10.1021/jacs.7b04209 J. Am. Chem. Soc. 2017, 139, 10952−10955