SAM-dependent enzyme-catalysed pericyclic reactions in natural product biosynthesis

Abstract

Pericyclic reactions-which proceed in a concerted fashion through a cyclic transition state-are among the most powerful synthetic transformations used to make multiple regioselective and stereoselective carbon-carbon bonds. They have been widely applied to the synthesis of biologically active complex natural products containing contiguous stereogenic carbon centres. Despite the prominence of pericyclic reactions in total synthesis, only three naturally existing enzymatic examples (the intramolecular Diels-Alder reaction, and the Cope and the Claisen rearrangements) have been characterized. Here we report a versatile S-adenosyl-l-methionine (SAM)-dependent enzyme, LepI, that can catalyse stereoselective dehydration followed by three pericyclic transformations: intramolecular Diels-Alder and hetero-Diels-Alder reactions via a single ambimodal transition state, and a retro-Claisen rearrangement. Together, these transformations lead to the formation of the dihydropyran core of the fungal natural product, leporin. Combined in vitro enzymatic characterization and computational studies provide insight into how LepI regulates these bifurcating biosynthetic reaction pathways by using SAM as the cofactor. These pathways converge to the desired biosynthetic end product via the (SAM-dependent) retro-Claisen rearrangement catalysed by LepI. We expect that more pericyclic biosynthetic enzymatic transformations remain to be discovered in naturally occurring enzyme 'toolboxes'. The new role of the versatile cofactor SAM is likely to be found in other examples of enzyme catalysis.

Extended Data Figure 1

LC–MS analysis of in vitro reaction of 3 with LepF. The extracted ion chromatograms (EICs) under positive ionization are shown. The molecular weight of alcohol 4 and 4′ is m/z 354 under positive ionization. Because the enzymatic activity of LepF is low and the alcohol 4 is very unstable, we were not able to obtain sufficient amount of 4 as the substrate for in vitro reaction of LepI. Thus, we obtained the alcohol 4 by using the chemical reduction of the ketone 3 with NaBH₄. Since NaBH₄ reduces the ketone 3 non-stereoselectively, 4 and diastereomer 4′ were formed. After the isolation of 4 and 4′ by HPLC, the fractions containing 4 and 4′ were not concentrated and were immediately used as the substrate. The stereochemistry of the secondary alcohol in 4 and 4′ is not determined.

Extended Data Figure 2

HPLC analysis of chemical reduction of 3 with NaBH₄. The reaction mixture containing 1 mM 3 and 10 mM NaBH₄ with EtOH (50 μL) was incubated at 0 °C for 1 min. Then, the reaction was quenched with water. After centrifugation, the supernatant was analyzed by HPLC. The reduction of 3 gave the alcohol 4 and diastereomer 4′. The spontaneous dehydration of both alcohols resulted in the formation of HDA and IMDA products via the E/Z mixture of quinone methide 5. The isolated 4 and 4′ also readily dehydrated and converted to the mixture of the desired HDA and the undesired HDA and IMDA products, showing the instability of these compounds. The structures show the relative stereochemistry.

Extended Data Figure 3

Reaction analysis of 6–9 under heating condition. 6 (dissolved in 5% DMSO with H₂O) was boiled at 95 °C for 1 h. 7–9 (dissolved in 5% DMSO with H₂O) were boiled 95 °C for 10 h. 6 was completely converted to 2 via [3,3]-sigmatropic retro-Claisen rearrangement. This reaction is irreversible under the conditions. It should be noted that the conversion of 6 to 2 via cycloreversion can be ruled out, since 6 was completely converted to 2 without any other IMDA/HDA side products. No reactions occurred in the case of 7. 8 and 9 can be interconvertible via Claisen rearrangement. In this case, retro-Claisen rearrangement (8 to 9) is more preferable than forward Claisen rearrangement (9 to 8). The structures show the relative stereochemistry.

Extended Data Figure 4

Analysis of the substrate specificity of LepI. (a) In vitro reactions of other IMDA products 7–9 with 30 μM LepI for 12 h. (i) 8 in buffer, (ii) 8 with LepI, (iii) 7 in buffer, (iv) 7 with LepI, (v) 9 in buffer, (vi) 9 with LepI. The experimental details are described in Methods. (b) Elucidation of inhibitory activity for retro-Claisen rearrangement of 7 against LepI. The experimental details are described in Methods. IC₅₀ value is mean ± standard deviation (s.d.) of three independent experiments. The structures show the relative stereochemistry.

Extended Data Figure 5

Time-course analysis of LepI-catalysed retro-Claisen rearrangement of 6 to 2. The experimental details are described in Methods. The data represent one representative experiment from at least three independent replicates.

Extended Data Figure 6

HPLC analysis showing that purified LepI retains SAM. SAM was detected from the supernatant of denatured LepI by acetonitrile. When LepI was denatured by heating the sample at 95 °C for 10 min, a single peak corresponding to 5′-deoxy-5′-(methylthio)adenosine (MTA), a major degradation product of SAM, was also detected from the supernatant of boiled LepI. Since SAM to MTA conversion is nearly quantitative and an MTA standard curve can be readily constructed, we quantified that ~90% of LepI still retains SAM after purification. HPLC profiles of (i) denatured LepI by acetonitrile, (ii) boiled LepI at 95 °C for 10 min, (iii) the authentic reference of SAM, (iv) boiled SAM at 95 °C for 10 min, and (v) the authentic reference of MTA. The experimental details are described in Methods.

Extended Data Figure 7

HPLC analysis of SAM-dependent LepI-catalysed reactions (a) Analysis of in vitro reaction of 240 μM 4 with 300 nM LepI at 30 °C for 5 min in the presence and absence of cofactors. The concentrations of SAH, SAM, and sinefungin used in this experiment are 250 μM, 100 μM, and 100 μM, respectively. The data represent one representative experiment from at least three independent replicates. (b) Analysis of in vitro reaction of 140 μM 6 with 300 nM LepI at 30 °C for 4 min in the presence and absence of cofactors. The concentrations of SAH, SAM, and sinefungin used in this experiment are 250 μM, 100 μM, and 100 μM, respectively. The data represent one representative experiment from at least three independent replicates.

Extended Data Figure 8

SAH is a competitive inhibitor of LepI retro-Claisen rearrangement. (a) Dose-dependent inhibition of retro-Claisen rearrangement by SAH. (b) Dose-dependent recovery of retro-Claisen rearrangement by SAM in the presence of 250 μM SAH. The experimental details are described in Methods. Error bars represent standard deviation of three independent experiments.

Extended Data Figure 9

Time-course analysis of relative production ratio of 2 over sum of 2 and 6. The substrate used in this study is alcohol 4. (a) LepI-catalysed reaction with or without SAH (250 μM). (b) non-enzymatic reaction. The initial production ratio between LepI-catalysed and non-catalysed reaction are clearly different. This data supported that LepI catalysed the competitive IMDA/HDA reactions by changing the preference of the outcome.

Extended Data Figure 10

Calculated free energies and bond distances of ambimodal TS (TS-1) and the retro-Claisen rearrangement TS (TS-2), uncatalysed and with various catalysts, calculated with B3LYP-D3/6-311+G(d,p)//6-31G(d), CPCM water.

Figure 1

Enzyme-catalysed pericyclic reactions and the proposed inverse electron demand hetero-Diels-Alder (HDA) reactions in Nature. (a) Examples of enzymatic pericyclic reactions. (b) The structures of natural products containing dihydropyran, which would be biosynthesized by HDA reaction. Variecolortide A is naturally racemic; the relative stereochemistry of epipyridone and leporin B are shown. (c) The putative leporin biosynthetic gene cluster in A. flavus and assignment of encoded genes and biosynthetic pathway of leporins. PKS–NRPS, polyketide synthase–nonribosomal peptide synthetase; TF, transcription factor; MCT, monocarboxylate transporter; SDR, short-chain dehydrogenase/reductase; ER, enoylreductase; OMT, O-methyltransferase. The structures show the relative stereochemistry. (d) Analysis of metabolites from the transformants of A. nidulans. The peak at 12 min correspond to the tetramic acid product that is biosynthesized by LepA (PKS-NRPS) and LepG (ER).

Figure 2

HPLC analysis showing the reactions catalysed by LepI. (a) in vitro reaction analysis of 4 and 4′ with 3 μM LepI for 0.5 h. (i) 4 in buffer, (ii) 4 with 3 μM LepI, (iii) 4′ in buffer, (iv) 4′ with 3 μM LepI. ^*6 is overlapped with 4′. (b) Time-course analysis of the conversion of 240 μM 4 to 2 in the presence of 300 nM LepI. (c) in vitro reaction analysis of 6 with 300 nM LepI for 3 min. (i) 6 in buffer, (ii) 6 with LepI. (d) Kinetic analysis of LepI-catalysed retro-Claisen rearrangement. (e) Scheme for putative LepI-catalysed retro-Claisen rearrangement. The structures show the relative stereochemistry.

Figure 3

LepI-catalysed reactions are SAM-dependent. (a) Time-course analysis of the consumption of 240 μM 4 in the presence of 300 nM LepI with or without cofactors; SAM (100 μM), SAH (250 μM), sinefungin (100 μM). (b) Time-course analysis of the conversion of 140 μM 6 to 2 in the presence of 300 nM LepI with or without cofactors; SAM (100 μM), SAH (250 μM), sinefungin (100 μM). (c) Time-course analysis of the production of 2 and 6 from 240 μM 4 in the presence of 300 nM LepI with or without 250 μM SAH. (d) Analysis of the relative production ratio of HDA adduct 2 and IMDA adduct 6 from 240 μM 4. In the case of nonenzymatic reaction, the reaction time is 10 min. In the case of LepI (300 nM)-catalysed reactions with or without SAH (250 μM), the reaction time is 4.0 min. Cont. means LepI without cofactors. (e) Structures of SAM, SAH, and sinefungin.

Figure 4

Energetics and transition states of the ambimodal IMDA/HDA, and retro-Claisen pericyclic reactions leading to formation of 2 from 4. (a) Free energy diagram for the non-enzymatic formation of 2 from (E)-5 and (Z)-5 calculated with B3LYP-D3/6-31G(d), gas phase. Gibbs free energies in kcal/mol. The ambimodal TS-1 gives both 6 and 2. (b) (i) Asymmetrical bifurcating PES for the formation of HDA adduct 2 and IMDA adduct 6 from (E)-5. (ii) Catalysed ambimodal TS (TS-1) structure with coordination of a trimethylsulfonium ion model. The same shift towards preference for HDA occurs with ammonium ion catalysis. (c) Summary of LepI-catalysed reactions cascade leading to 2 from 4 via the formation of (E)-5: dehydration and the subsequent reactions, 1) “direct” path (HDA reaction), 2) “byproduct recycle” path (IMDA reaction/retro-Claisen rearrangement). Gibbs free energies (kcal/mol) of TS-1 and TS-2 are calculated with B3LYP-D3/6-311+G(d,p)//6-31G(d), CPCM water. The structures show the relative stereochemistry.