Homo-Clavulyne: Evaluation of a One-Atom Ring Expansion Strategy to Stabilize the Nine-Membered Ring Enediyne Biosynthetic Precursor

Abstract

Streptomyces clavuligerus biosynthesizes a nine-membered ring enediyne that spontaneously cycloaromatizes to a 1,4-benzenoid diradical enroute to the clavulynes. Since nature also produces several isolable and exceedingly cytotoxic metabolites featuring 10-membered ring enediynes, we reasoned that a ring-expanded “homo-clavulyne” enediyne might be accessible by the concepts and methods of organic synthesis, isolable, and useful as a new molecular warhead with therapeutic applications. The design of a 10-membered ring homo-clavulyne enediyne and our circumstantial evidence for its existence are described herein.

GRAPHICAL ABSTRACT

Introduction

When the Bergman laboratory presented chemical evidence for the intermediacy of 1,4-benzenoid diradicals in thermal isomerizations of cis-1,5-diyn-3-enes (Fig. 1),¹ the chemical community was unaware that nature had already evolved a family of exceedingly cytotoxic natural products that harness this chemistry to effect cleavages of double-stranded DNA and cell death. The same type of conjugated six-carbon unit at the heart of Bergman’s pioneering studies is also a prominent structural element in calicheamicin γ^I,² esperamicin,³ dynemicin A,⁴ maduropeptin,⁵ and other natural products (not shown)⁶ that bind to DNA and may be triggered to undergo analogous cycloaromatizations to a 1,4-benzene diradical capable of abstracting hydrogen atoms from the phosphate backbone of DNA. Neocarzinostatin is structurally distinct from these aforementioned natural products but can be induced to undergo an analogous Myers–Saito cyclization to a lethal diradical that also abstracts hydrogen atoms from DNA and triggers apoptosis.⁷ These discoveries were captivating and stimulated a great deal of research in the area of organic synthesis, with a particular focus on the context dependence of Bergman cycloaromatization chemistry, for example, correlations between molecular strain and c–d distances with proclivities toward cycloaromatization,⁸ and the therapeutic potential of the enediyne antitumor antibiotics and novel compounds derived from them (e.g., Mylotarg).⁹

Fig. 1.

The Bergman cycloaromatization and the structures of selected members of the enediyne class of antitumor natural products.

Some members of this unique class of natural products, for example, sporolides A and B,¹⁰ the cyanonsporasides,¹¹ fijiolide B,¹² the amycolamycins,¹³ and the clavulynes,¹⁴ originate from enediyne precursors that are too unstable to be isolated.¹⁵ The clavulynes, recently described by Seyedsayamdost and coworkers,¹⁴ are cryptic enediyne-derived metabolites produced by Streptomyces clavuligerus. These metabolites arise from a fleeting nine-membered ring enediyne 1 (Scheme 1a) that cannot be isolated in laboratory settings. The assumption is that S. clavuligerus biosynthesizes the pre-clavulyne enediyne 1 for offensive purposes and that this reactive, unstable compound undergoes a spontaneous Bergman cyclization to a cytotoxic 1,4-benzenoid diradical (2), giving rise to clavulyne E (3) by hydrogen atom transfers to the diradical and the related monohalogenated clavulynes by the Perrin–O’Connor reaction.¹⁶

Scheme 1.

(a) The putative nine-membered ring pre-clavulyne enediyne (1), cytotoxic 1,4-benzenoid diradical (2) and clavulyne E (3). (b) The structures of enantiomeric 10-membered ring homo-clavulyne enediynes 4 and 5, and our synthetic strategy.

To date, there are at least eight structurally distinct natural products with the enediyne array confined to an isolable and relatively stable 10 membered ring. On this basis, we hypothesized that a 10-membered ring ‘homo-clavulyne’ enediyne (e.g., 4 and 5, Scheme 1b) might be accessible by organic synthesis and possibly also isolable and therapeutically useful. A one-atom extension may seem to be a minor structural modification, yet numerous examples have demonstrated that a difference of a single atom can induce substantial changes in chemical and biological properties.¹⁷ The testing of this hypothesis and our initial effort to stabilize the fugitive pre-clavulyne enediyne by a one-atom ring expansion are described herein.

At the outset, our thinking was influenced by the broad applicability of the Sonogashira cross-coupling method¹⁸ to the problem of constructing enediyne arrays, as well as by the successful cyclizations that the groups of Kende,¹⁹ Danishefsky,²⁰ and Nicolaou²¹ had achieved in their syntheses of the challenging molecular architecture of calicheamicin γ^I. In principle, the 10-membered rings in homo-clavulynes 4 and 5 could arise by a strategy comprising a convergent Sonogashira coupling (e.g., between alkyne 6 and vinyl bromide 7) and some type of intramolecular addition of an alkyne-based nucleophile to a carbonyl group. These potential bond constructions seemed reasonable and yet the highly substituted cyclopentene 6, with its three contiguous oxygen-bearing stereocenters, would still be challenging to synthesize in a concise fashion. As matters transpired, we evaluated three distinct pathways for synthesizing compound 6 (Supporting Information) before settling on an approach featuring a 1,4-borylation/oxidation. During this undertaking, ent-homo-clavulyne 5 became our preferred target for synthesis only because we detected a stereochemical correspondence between key intermediate 6 and a known derivative of D-ribose. We also knew that a successful synthesis of 5 could be easily adapted to a synthesis of the enantiomeric homo-clavulyne 4 if desired.

In fact, D-ribose derivative 8 (Scheme 2a) is commercially available, relatively inexpensive, and possesses a cis-fused bicyclic framework that could be leveraged to control the stereochemical outcome of the eventual 1,4-borylation (vide infra). α-Iodination of cyclopentenone 8 afforded iodoenone 9 in 86% yield and was followed by standard Sonogashira coupling with trimethylsilylacetylene to give the alkynyl enone 10 in 82% yield.¹⁸ The two-carbon extension of the carbonyl group in 10 presented the first challenge in our synthesis. Several methods, including Horner–Wadsworth–Emmons, Peterson olefination, and aldol addition with lithio ethyl acetate followed by β-hydroxy elimination, failed to furnish the desired ester 11. To our delight, a Wittig olefination using stabilized phosphorane 12 in refluxing toluene afforded 11 in 69% yield.

Scheme 2.

(a) Synthesis of potential cyclization substrates 20a and 20b featuring a 1,4-borylation/oxidation and a Sonogashira fragment coupling. Reagents and conditions: 1. I₂ (2 equiv), pyridine (1.5 equiv), CH₂Cl₂ (0.2 M), 0 to 23°C, 16 h, 86%; 2. Pd(PPh₃)₄ (1 mol%), CuI (6 mol%), trimethylsilylacetylene (3 equiv), ⁱPr₂NEt (2 equiv), THF (0.1 M), 23°C, 16 h, 82%; 3. 12 (1.8 equiv), PhMe (0.057 M), 110°C, 18 h, 69%; 4. Cu(OAc)₂ (20 mol%), L (24 mol%), B₂pin₂ (1.6 equiv), wet THF (0.1 M), dark, 23°C, 6 h, then NaBO₃·4H₂O (3 equiv), H₂O (0.15 M), 0–23°C, 1 h, 65%; 5. ⁿBu₄NF (1.5 equiv), AcOH (5 equiv), THF (0.1 M), 0–23°C, 1 h, 88%; 6. ^tBuMe₂SiOTf (2 equiv), 2,6-lutidine (3 equiv), CH₂Cl₂ (0.1 M), 0°C, 30 min, 98%; 7. 18 (1 equiv), Pd(PPh₃)₄ (5 mol%), CuI (30 mol%), ⁱPr₂NEt (4 equiv), 2,6-lutidine (2 equiv), DMF (0.1 M), 4Å MS, dark, 23°C, 16 h, 74%; 8a. AgNO₃ (40 mol%), NIS (2 equiv), DMF (0.05 M), dark, 23°C, 3 h, 67%; 8b. AgNO₃ (40 mol%), H₂O (100 equiv), acetone (0.1 M), dark, 23°C, 2 h, 82%. (b) Inappropriate coupling partner 7. B₂pin₂ = Bis(pinacolato)diboron; NIS = N-iodosuccinimide; MS = molecular sieves.

At this stage, we needed to achieve a site- and stereoselective hydration of the newly constructed electron-deficient alkene and were drawn to the effective copper-catalyzed method developed by the Kobayashi laboratory.²² When catalyzed by a chiral Cu^II-bis (hydroxymethyl)bipyridyl complex, Kobayashi’s β-selective hydration of 1,6-unsaturated carbonyls is enantio-selective and performs well on a range of 1,6-unsaturated carbonyl substrates.²² Since the rigid, cis-fused bicyclic structure of 11 was expected to impart substantial facial selectivity, the achiral bis(hydroxymethyl)bipyridyl ligand (L) was employed for this reaction. Following reaction optimization (Supporting Information), we were pleased to find that the Cu^II-catalyzed 1,4-borylation of compound 11 in degassed, wet tetrahydrofuran (THF) under exclusion of light, followed by in situ oxidation of the resulting organoborane, produced alcohol 13 in 65% yield on a gram scale. To set the stage for the key Sonogashira cross-coupling, compound 13 was converted to terminal alkyne 14 via compound 6 by the two-step reaction sequence shown. Notably, trimethylsilyl deprotection of the alkyne in 13 with tetra-n-butyl-ammonium fluoride resulted in decomposition in absence of acetic acid, possibly due to the degradation of the β-hydroxy functionality through retro-aldol or E1cB elimination under basic conditions. To suppress this degradation pathway in subsequent transformations, the alcohol was protected as the corresponding tert-butyldimethylsilyl ether to afford 14 as the alkyne coupling fragment. We had hoped that vinyl bromide 7, bearing a pendant ethenyl group, could serve as a suitable coupling partner for the Sonogashira reaction (Scheme 2b). Unfortunately, however, the outcome we observed suggests that the initially formed organopalladium(II) intermediate 15 isomerizes to 17 prior to transmetalation and this isomerization is thermodynamically favored (Supporting Information). Consequently, vinyl bromide 18b was prepared in one step from the known alcohol 18a as an alternative coupling fragment,²³ equipped with a protected alcohol as a useful functional group handle.

While coupling partners 14 and 18 were in hand, the pivotal Sonogashira reaction required extensive screening to find the most optimal conditions (Supporting Information). Ultimately, we found that conducting this coupling in strictly degassed, anhydrous DMF with exclusion of light and in presence of dual base additives minimizes substrate decomposition and undesired Glaser homocoupling, affording enediyne 19 in 74% yield. In the wake of this fragment coupling, the silyl-protected alkyne was converted to the corresponding iodoalkyne 20a and also the terminal alkyne 20b using AgNO₃ catalysis in preparation for the efforts to construct the desired homo-clavulyne 10-membered ring.^24,25

The two options we favored for achieving ring formation required the seemingly straightforward conversion of the ethyl esters in compounds 20a and 20b to aldehydes. We first attempted the complete reductions of the esters in 20a/20b to primary alcohols, followed by a separate step of Dess–Martin oxidation to acquire the corresponding aldehydes. However, the resulting aldehydes underwent rapid decomposition during column chromatography on silica, deactivated silica, alumina, or Florisil, presumably via facile E1cB eliminations of the β-silyloxy substituents and secondary degradation pathways. Unfortunately, all other efforts, including Piancatelli–Margarita, Stahl, Parikh–Doering, Albright–Goldman, Ley–Griffith, Swern, and Fétizon oxidations, failed to produce the desired aldehydes (Supporting Information). Ultimately, we found that diisobutylaluminum hydride (DIBAL-H) mediated partial reductions of esters 20a and 20b were successful at −78 °C and afforded the sensitive aldehydes 21a and 21b, respectively, as stable, isolable compounds after a Fieser workup of the reactions (Scheme 3).²⁶ In this fashion, the desired aldehydes were produced cleanly with no sign of unreacted starting material and only trace amounts of overreduction products.

Scheme 3.

Ring-closing Nozaki–Hiyama–Kishi coupling to forge the 10-membered enediyne framework and the subsequent premature Bergman cyclization leading to 25. Reagents and conditions: 9. DIBAL-H (1.5 equiv), CH₂Cl₂ (0.05 M), −78 °C, 1 h; 10. NiCl₂ (20 mol %), CrCl₂ (10 equiv), THF (0.005 M), 23°C, 3 h, 50% over two steps. DIBAL-H = diisobutylaluminum hydride; MS = molecular sieves.

With the desired cyclization substrates finally in hand, a one-carbon homologue of the putative pre-clavulyne enediyne 1 was nearly within reach. Could we generate a homo-clavulyne framework in the course of an intramolecular carbonyl addition and would the resulting adduct be isolable and relatively stable? We examined several methods for generating acetylide nucleophiles from 21a or 21b (Supporting Information). Deprotonation of the alkynyl hydrogen on 21b with potassium bis(trimethylsilyl)amide or ^tBuLi resulted in full decomposition of the aldehyde likely due to the competing deprotonation of the α-hydrogens. Carreira’s mild conditions for generating Zn^II-acetylides using stoichiometric Zn(OTf)₂ and NEt₃ also led to decomposition of the aldehyde,²⁷ presumably for the same reason. Eventually, we found that a Nozaki–Hiyama–Kishi reaction²⁸ on substrate 21a using catalytic NiCl₂ and a large excess of CrCl₂ in strictly deoxygenated and anhydrous THF generated a putative Cr^III-acetylide that added intramolecularly to the aldehyde. We hoped and even expected to isolate 10-membered ring enediyne 23 from this reaction but could infer only that this compound formed as a transient intermediate enroute to the cycloaromatized product 25, which we isolated after workup of the reaction. Compound 23 (or its alkoxide precursor 22) underwent a premature Bergman cycloaromatization and the resulting benzenoid diradical 24 abstracted hydrogen atoms, presumably from THF, leading to the polycyclic benzene 25 as a 1:1 mixture of benzylic alcohol epimers. These inseparable epimers were isolated together in an overall yield of 50% from iodoalkyne ester 20a, while the remaining of the mass balance was attributed to 21b via inefficient acetylide addition and uncontrolled radical degradation.

The clavulynes are cryptic gene products that originate from a putative, unstable nine-membered ring enediyne. Since nature also biosynthesizes several 10-membered ring enediynes that may be isolated and even produced in the laboratory by chemical synthesis, we reasoned that a 10-membered ring homo-clavulyne might also be stable, accessible by synthesis, and available for efforts to develop its therapeutic potential. In fact, an appropriate orchestration of Wittig olefination, regio- and stereoselective 1,4-borylation/oxidation, Sonogashira cross-coupling, and ring-closing Nozaki–Hiyama–Kishi reactions are capable of generating a homo-clavulyne enediyne but only as a fugitive intermediate on a path to the product of a premature Bergman cycloaromatization. Like the pre-clavulyne enediyne 1 that inspired it, the homo-clavulyne enediyne is elusive and too close to the margins of chemical stability. Despite the unexpected result, the key strategy of Sonogashira cross-coupling, followed by intramolecular Nozaki–Hiyama–Kishi reaction leading to a fleeting benzenoid diradical intermediate analogous to 2 can be applicable to the biomimetic synthesis of clavulynes themselves. This work is part of a larger, ongoing effort to design and produce clavulyne-inspired compounds with improved stability and potential utility as starting points for the development of enediynes or latent enediynes with genuine therapeutic value.

All reactions were performed in oven-dried or flame-dried glassware under N₂ by standard Schlenk techniques with a dry N₂/vacuum manifold unless otherwise indicated. CuI was purified by washing with refluxing THF in Soxhlet extractor and stored under N₂. All other reagents were purchased and used as received. Molecular sieves (4 Å, beads, 8–12 mesh) were activated in vacuum oven overnight before using. All solvents used were HPLC grade and dried by running through activated alumina columns or from Sure/Seal^™ or AcroSeal^™ bottles. Where indicated as strictly deoxygenated, THF was freshly distilled from sodium/benzophenone before use. Reactions were monitored via analytical thin-layer chromatography (TLC) conducted on Merck Silica gel 60 F254 plates containing a fluorescent indicator, which were visualized with iodine, ceric ammonium molybdate, KMnO₄, p-anisaldehyde stain and/or UV light. Compounds were purified by flash column chromatography on Silicycle SiliaFlash P60 (230–400 mesh) silica gel or preparative TLC on SiliaPlate^™ (20 × 20 cm, 1000 μm) containing a fluorescent indicator with ACS grade solvents. Where used, Celite filtering agent was Celite 545 purchased from Thermo Scientific. Nuclear magnetic resonance (NMR) spectra were obtained on 500 MHz Bruker spectrometers at 22 °C. ¹H and ¹³C spectra were referenced to residual solvent (CDCl₃: δ 7.26 ppm for ¹H NMR and δ 77.16 ppm for ¹³C NMR). The multiplicities are abbreviated as follows: s = singlet, d = doublet, t = triplet, q = quartet, p = pentet, dd = doublet of doublets, ddd = doublet of doublet of doublets, dddd = doublet of doublet of doublet of doublets, dtt = doublet of triplet of triplets, td = triplet of doublets, tdd = triplet of doublet of doublets, qd = quartet of doublets, m = multiplet, and br = broad signal. Coupling constants were measured in Hz and abbreviated as J. Infrared (IR) spectra were recorded on a Nicolet 6700 FT-IR equipped with a 30,000–200 cm⁻¹ diamond. High-resolution mass spectrometry (HRMS) data were obtained on an Agilent 6230 LC/TOF with liquid chromatography electrospray ionization (ESI).

Procedures

Iodide 9

To a 500-mL round-bottom flask charged with a magnetic stir bar was added 8 (5 g, 32.43 mmol, 1 equiv), iodine (16.46 g, 64.87 mmol, 2 equiv), and CH₂Cl₂ (162 mL, 0.2 M). The solution was cooled to 0 °C and added pyridine (3.92 mL, 48.65 mmol, 1.5 equiv) slowly, then stirred for 16 h at room temperature. The reaction was quenched with 10% aq. ascorbic acid solution (150 mL). The organic layer was separated, and the aqueous layer was extracted with CH₂Cl₂ (80 mL × 3). The combined organic layers were dried over anhydrous MgSO₄ and filtered, then concentrated in vacuo. The crude residue was purified via flash chromatography (0–15 % EtOAc/hexanes) to afford iodide 9 (7.80 g, 86%) as a white solid.

TLC R_f = 0.36 (20% EtOAc/hexanes); all characterizations matched with previously reported literature.²⁹

Alkynyl enone 10

To a 1000-mL round-bottom flask charged with a magnetic stir bar was added iodide 9 (7.80 g, 27.85 mmol, 1 equiv), trimethylsilylacetylene (11.6 mL, 83.55 mmol, 3 equiv), ⁱPr₂NEt (9.7 mL, 55.70 mmol, 2 equiv), and anhydrous THF (278 mL, 0.1 M). The solution was degassed by bubbling a steady stream of N₂ for 40 min, then swiftly added Pd(PPh₃)₄ (321.8 mg, 0.2785 mmol, 1 mol%) and CuI (318.3 mg, 1.671 mmol, 6 mol%). The reaction was stirred for 16 h at room temperature before quenching with sat. aq. NH₄Cl (200 mL). The organic layer was separated, and the aqueous layer was extracted with EtOAc (150 mL × 3). The combined organic layers were dried over anhydrous MgSO₄ and filtered, then concentrated in vacuo. The crude residue was purified via flash chromatography (0–12 % EtOAc/hexanes) to afford alkyne 10 (5.72 g, 82%) as a white solid.

TLC R_f = 0.39 (20% EtOAc/hexanes).

¹H NMR (CDCl₃, 500 MHz) δ 7.59 (d, J = 2.5 Hz, 1H), 5.24 (dd, J = 5.6, 2.7 Hz, 1H), 4.52 (d, J = 5.4 Hz, 1H), 1.41 (s, 6H), 0.23 (s, 9H).

¹³C NMR (CDCl₃, 126 MHz) δ 198.6, 159.3, 130.4, 115.7, 105.4, 93.9, 77.4, 76.7, 27.6, 26.3, −0.2.

HRMS (+ESI-TOF) m/z calcd for C₁₃H₁₈NaO₃Si⁺ [M+Na]⁺: 273.0917; found 273.0913.

FT-IR (neat) v 2960, 2095, 1740, 1597, 1375, 1251, 1207, 1153, 1098, 1047, 999, 916, 846, 762 cm⁻¹.

Ester 11

To a 25-ml round-bottom flask charged with a magnetic stir bar was added alkynyl enone 10 (100.0 mg, 0.400 mmol, 1 equiv), ethyl 2-(triphenylphosphoranylidene) acetate 12 (250.5 mg, 0.719 mmol, 1.8 equiv), and toluene (7 mL, 0.057 M). The solution was stirred for 18 h at 110 °C before cooling down to room temperature and concentrating in vacuo. The crude residue was purified via flash chromatography (0–5 % EtOAc/hexanes) to afford ester 11 (87.7 mg, 69%) as a white solid.

TLC R_f = 0.10 (10% EtOAc/hexanes).

¹H NMR (CDCl₃, 500 MHz) δ 6.60 (d, J = 2.3 Hz, 1H), 6.16 (s, 1H), 5.71 (dt, J = 5.5, 1.0 Hz, 1H), 5.17 (dd, J = 5.6, 2.4 Hz, 1H), 4.26 (qd, J = 7.1, 3.8 Hz, 2H), 1.44 (s, 3H), 1.39 (s, 3H), 1.33 (t, J = 7.1 Hz, 3H), 0.24 (s, 9H).

¹³C NMR (CDCl₃, 126 MHz) δ 165.9, 156.6, 145.7, 128.9, 116.4, 112.7, 102.4, 96.2, 82.1, 76.8, 60.6, 27.7, 26.3, 14.4, −0.1.

HRMS (+ESI-TOF) m/z calcd for C₁₇H₂₄NaO₄Si⁺ [M+Na]⁺: 343.1336; found 343.1344.

FT-IR (neat) v 2960, 2043, 1721, 1652, 1581, 1457, 1371, 1331, 1302, 1251, 1209, 1176, 1155, 1081, 1145, 1004, 846, 761, 702 cm⁻¹.

Alcohol 13

To a 100-mL round-bottom flask charged with a magnetic stir bar was added ester 11 (1.13 g, 3.535 mmol, 1 equiv) and wet THF (35 mL, 0.1 M). The solution was degassed by bubbling a steady stream of N₂ for 20 min, then swiftly added Cu(OAc)₂ (128.4 mg, 0.7070 mmol, 20 mol%), Ligand L (183.5 mg, 0.8484 mmol, 24 mol%), and B₂pin₂ (1.44 g, 5.656 mmol, 1.6 equiv). The reaction was covered with aluminum foil and stirred for 6 h at room temperature. After complete consumption of the starting material, the reaction was cooled to 0 °C and slowly added NaBO₃·4H₂O(1.63 g, 10.60 mmol, 3 equiv) and H₂O (24 mL, 0.15 M). The reaction was stirred for another hour at room temperature before quenching with sat. aq. NH₄Cl (20 mL). The organic layer was separated, and the aqueous layer was extracted with EtOAc (30 mL × 3). The combined organic layers were dried over anhydrous MgSO₄ and filtered, then concentrated in vacuo. The crude residue was purified via flash chromatography (0–15 % EtOAc/hexanes) to afford alcohol 13 (778.7 mg, 65%) as a pale yellow oil.

TLC R_f = 0.45 (40% EtOAc/hexanes).

¹H NMR (CDCl₃, 500 MHz) δ 5.36 (dd, J = 5.6, 1.3 Hz, 1H), 4.66 (d, J = 4.6 Hz, 1H), 4.53 (d, J = 5.6 Hz, 1H), 4.17 (qd, J = 7.1, 4.0 Hz, 2H), 3.48 (d, J = 16.4 Hz, 1H), 3.27 (d, J = 16.4 Hz, 1H), 2.10 (d, J = 5.4 Hz, 1H), 1.34 (d, J = 6.4 Hz, 6H), 1.28 (t, J = 7.1 Hz, 4H), 0.21 (s, 9H).

¹³C NMR (CDCl₃, 126 MHz) δ 169.7, 146.5, 126.0, 112.1, 103.2, 98.2, 84.6, 84.2, 81.9, 61.2, 33.7, 27.5, 26.1, 14.3, 0.0.

HRMS (+ESI-TOF) m/z calcd for C₁₇H₂₇O₅Si [M+H]⁺: 339.1622; found 339.1626.

FT-IR (neat) v 3436, 2985, 2959, 2934, 2147, 1738, 1371, 1328, 1251, 1209, 1182, 1158, 1069, 1003, 950, 844, 761, 710 cm⁻¹.

Enyne 6

To a 100-mL round-bottom flask charged with a magnetic stir bar was added alcohol 13 (778.7 mg, 2.30 mmol, 1 equiv) and THF (23 mL, 0.1 M). The reaction was cooled to 0°C and slowly added AcOH (0.66 mL, 11.50 mmol, 5 equiv) and ⁿBu₄NF (3.45 mL, 1 M in THF, 3.45 mmol, 1.5 equiv), then stirred for 1 h at room temperature. After complete consumption of the starting material, the reaction was quenched with sat. aq. NH₄Cl (20 mL). The organic layer was separated, and the aqueous layer was extracted with EtOAc (20 mL × 3). The combined organic layers were dried over anhydrous MgSO₄ and filtered, then concentrated in vacuo. The crude residue was purified via flash chromatography (0–15 % EtOAc/hexanes) to afford enyne 6 (536.4 mg, 88%) as a clear oil.

TLC R_f = 0.26 (40% EtOAc/hexanes).

¹H NMR (CDCl₃, 500 MHz) δ 5.36 (d, J = 5.6 Hz, 1H), 4.69 (dt, J = 5.7, 1.6 Hz, 1H), 4.55 (d, J = 5.6 Hz, 1H), 4.17 (qd, J = 7.2, 3.1 Hz, 2H), 3.50 (d, J = 16.3 Hz, 1H), 3.30 (d, J = 16.3 Hz, 1H), 3.26 (s, 1H), 2.11 (d, J = 5.7 Hz, 1H), 1.35 (d, J = 4.5 Hz, 6H), 1.28 (t, J = 7.2 Hz, 3H).

¹³C NMR (CDCl₃, 126 MHz) δ 169.5, 147.5, 125.0, 112.2, 84.9, 84.6, 84.2, 82.0, 61.3, 33.6, 27.4, 26.1, 14.3.

HRMS (+ESI-TOF) m/z calcd for C₁₄H₁₈NaO₅⁺ [M+Na]⁺: 289.1046; found 289.1050.

FT-IR (neat) v 3435, 3284, 2986, 2935, 1735, 1457, 1372, 1328, 1301, 1254, 1209, 1159, 1069, 1032, 1001, 949, 872, 805 cm⁻¹.

Enyne 14

To a 50-mL round-bottom flask charged with a magnetic stir bar was added enyne 6 (257.0 mg, 0.9651 mmol, 1 equiv), 2,6-lutidine (0.335 mL, 2.895 mmol, 3 equiv) and anhydrous CH₂Cl₂ (9.7 mL, 0.1 M). The reaction was cooled to 0 °C and added ^tBuMe₂SiOTf (0.443 mL, 1.930 mmol, 2 equiv) dropwise, then stirred for 30 min at 0 °C. After complete consumption of the starting material, the reaction was quenched with H₂O (10 mL). The organic layer was separated, and the aqueous layer was extracted with CH₂Cl₂ (10 mL × 3). The combined organic layers were dried over anhydrous MgSO₄ and filtered, then concentrated in vacuo. The crude residue was purified via flash chromatography (0–5 % EtOAc/hexanes) to afford enyne 14 (360.0 mg, 98%) as a clear oil.

TLC R_f = 0.23 (10% EtOAc/hexanes).

¹H NMR (CDCl₃, 500 MHz) δ 5.36 (d, J = 5.6 Hz, 1H), 4.67 (t, J = 1.6 Hz, 1H), 4.41 (d, J = 5.6 Hz, 1H), 4.16 (qd, J = 7.2, 2.6 Hz, 2H), 3.46 (d, J = 16.2 Hz, 1H), 3.30 (d, J = 16.1 Hz, 1H), 3.20 (s, 1H), 1.34 (d, J = 2.7 Hz, 6H), 1.27 (t, J = 7.1 Hz, 3H), 0.92 (s, 9H), 0.16 (d, J = 6.6 Hz, 6H).

¹³C NMR (CDCl₃, 126 MHz) δ 169.6, 146.2, 125.6, 112.1, 85.6, 84.4, 84.4, 82.8, 78.3, 61.1, 33.8, 27.5, 26.3, 26.0, 18.4, 14.3, −4.4, −4.6.

HRMS (+ESI-TOF) m/z calcd for C₂₀H₃₃O₅Si⁺ [M+H]⁺: 381.2092; found 381.2105.

FT-IR (neat) v 3253, 2984, 2955, 2930, 2858, 2032, 1740, 1472, 1371, 1255, 1210, 1179, 1157, 1091, 1029, 1005, 867, 837, 805, 778 cm⁻¹.

Vinyl Bromide 18b

To a 250-mL round-bottom flask charged with a magnetic stir bar was added alcohol 18a (1.773 g, 7.604 mmol, 1 equiv),^{23 i}Pr₂NEt (5.30 mL, 30.42 mmol, 4 equiv) and anhydrous CH₂Cl₂ (76 mL, 0.1 M). The reaction was cooled to −78 °C and added methoxymethyl chloride (1.73 mL, 22.81 mmol, 3 equiv) dropwise, then stirred for 12 h at room temperature. After complete consumption of the starting material, the reaction was cooled to 0 °C and quenched with sat. aq. NaHCO₃ (50 mL). The organic layer was separated, and the aqueous layer was extracted with CH₂Cl₂ (30 mL × 3). The combined organic layers were dried over anhydrous MgSO₄ and filtered, then concentrated in vacuo. The crude residue was purified via flash chromatography (0–2 % EtOAc/hexanes) to afford vinyl bromide 18b (1.594 g, 76%) as a clear oil.

TLC R_f = 0.39 (10% EtOAc/hexanes).

¹H NMR (CDCl₃, 500 MHz) δ 6.30 (s, 1H), 4.65 (s, 2H), 4.26 (s, 2H), 3.38 (s, 3H), 0.22 (s, 9H).

¹³C NMR (CDCl₃, 126 MHz) δ 134.3, 112.0, 102.3, 100.9, 95.6, 71.1, 55.8, −0.1.

HRMS (+ESI-TOF) m/z calcd for C₁₀H₁₈BrO₂Si⁺ [M+H]⁺: 277.0254; found 277.0530.

FT-IR (neat) v 2957, 2897, 2142, 1250, 1152, 1104, 1063, 1035, 924, 842, 760 cm⁻¹.

Enediyne 19

To a 50-mL Schlenk flask charged with a magnetic stir bar was added enyne 14 (288.2 mg, 0.7573 mmol, 1 equiv), vinyl bromide 18b (209.9 mg, 0.7573 mmol, 1 equiv), ⁱPr₂NEt (0.528 mL, 3.029 mmol, 4 equiv), 2,6-lutidine (0.175 mL, 1.515 mmol, 2 equiv), activated 4 Å molecular sieves (379 mg, 50 mg/mL), and anhydrous DMF (7.6 mL, 0.1 M). The system underwent freeze–pump–thaw protocol for three cycles then backfilled with N₂ atmosphere. Pd(PPh₃)₄ (43.8 mg, 37.87 μmol, 5 mol%) and CuI (43.3 mg, 0.2272 mmol, 30 mol%) in a degassed vial was added swiftly under high N₂ outflow. The reaction was covered with aluminum foil and stirred for 16 h at room temperature, then filtered to a separatory funnel and diluted with Et₂O (20 mL) and sat. aq. NH₄Cl (10 mL). The organic layer was separated, and the aqueous layer was extracted with Et₂O (20 mL × 3). The combined organic layers were washed with 5% aq. LiCl (40 mL) and brine (30 mL), then dried over anhydrous MgSO₄, filtered, and concentrated in vacuo. The crude residue was purified via flash chromatography (0–10 % EtOAc/hexanes) to afford enediyne 19 (324.9 mg, 74%) as a pale yellow oil.

TLC R_f = 0.44 (20% EtOAc/hexanes).

¹H NMR (CDCl₃, 500 MHz) δ 6.05 (t, J = 1.7 Hz, 1H), 5.42 (d, J = 5.7 Hz, 1H), 4.69 (t, J = 1.5 Hz, 1H), 4.63 (s, 2H), 4.43 (d, J = 5.6 Hz, 1H), 4.15 (q, J = 7.1 Hz, 2H), 4.12 (d, J = 1.8 Hz, 2H), 3.60 (d, J = 16.4 Hz, 1H), 3.36 (s, 3H), 3.33 (d, J = 16.3 Hz, 1H), 1.34 (s, 6H), 1.26 (t, J = 7.1 Hz, 3H), 0.91 (s, 9H), 0.20 (s, 9H), 0.16 (d, J = 9.3 Hz, 6H).

¹³C NMR (CDCl₃, 126 MHz) δ 169.6, 146.0, 131.4, 126.1, 115.7, 112.0, 103.1, 102.3, 95.8, 92.7, 91.9, 85.7, 84.3, 82.8, 68.3, 61.1, 55.7, 33.9, 27.6, 26.4, 26.0, 18.4, 14.3, 0.1, −4.3, −4.5.

HRMS (+ESI-TOF) m/z calcd for C₃₀H₄₉O₇Si₂⁺ [M+H]⁺: 577.3011; found 577.3023.

FT-IR (neat) v 2955, 2930, 2857, 2035, 1740, 1472, 1371, 1251, 1211, 1152, 1093, 1044, 1007, 921, 842, 778, 761 cm⁻¹.

Iodide 20a

To an 8-mL vial charged with a magnetic stir bar was added enediyne 19 (50.0 mg, 86.67 μmol, 1 equiv), N-iodosuccinimide (39.0 mg, 0.1733 mmol, 2 equiv), AgNO₃ (5.9 mg, 34.67 μmol, 40 mol%), and anhydrous DMF (1.7 mL, 0.05 M). The reaction was covered with aluminum foil and stirred for 3 h at room temperature before quenching with sat. aq. Na₂S₂O₃ (2 mL). The aqueous layer was extracted with Et₂O (5 mL × 3). The combined organic layers were washed with 5% aq. LiCl (10 mL) and brine (10 mL), then dried over anhydrous MgSO₄, filtered, and concentrated in vacuo. The crude residue was purified via flash chromatography (0–10 % EtOAc/hexanes) to afford iodide 20a (36.5 mg, 67%) as a pale yellow oil.

TLC R_f = 0.33 (20% EtOAc/hexanes).

¹H NMR (CDCl₃, 500 MHz) δ 6.15 (t, J = 1.8 Hz, 1H), 5.41 (dd, J = 5.6, 1.2 Hz, 1H), 4.70 (d, J = 1.6 Hz, 1H), 4.64 (s, 2H), 4.43 (d, J = 5.6 Hz, 1H), 4.17 (qd, J = 7.2, 5.7 Hz, 2H), 4.13 (d, J = 1.8 Hz, 2H), 3.58 (d, J = 16.3 Hz, 1H), 3.40 (d, 1H), 3.37 (s, 3H), 1.35 (d, J = 4.5 Hz, 6H), 1.28 (t, J = 7.2 Hz, 3H), 0.92 (s, 9H), 0.17 (d, J = 8.7 Hz, 6H).

¹³C NMR (CDCl₃, 126 MHz) δ 169.7, 146.8, 133.3, 126.0, 115.8, 112.1, 95.9, 92.6, 92.2, 92.0, 85.7, 84.4, 82.6, 68.1, 61.1, 55.7, 34.1, 29.9, 27.6, 26.4, 26.0, 18.4, 14.4, −4.3, −4.5.

HRMS (+ESI-TOF) m/z calcd for C₂₇H₄₀IO₇Si⁺ [M+H]⁺: 631.1583; found 631.1589.

FT-IR (neat) v 2930, 2857, 2007, 1738, 1472, 1372, 1256, 1213, 1154, 1080, 862, 839, 781 cm⁻¹.

Enediyne 20b

To an 8-mL vial charged with a magnetic stir bar was added enediyne 19 (100.0 mg, 0.1733 mmol, 1 equiv), AgNO₃ (5.9 mg, 34.67 μmol, 40 mol%), H₂O (0.31 mL, 17.33 mmol, 100 equiv), and anhydrous acetone (1.7 mL, 0.1 M). The reaction was covered with aluminum foil and stirred for 2 h at room temperature. After complete consumption of the starting material, the reaction was purified directly via preparative TLC (20 % EtOAc/hexanes) to afford Enediyne 20b (71.4 mg, 82%) as a pale yellow oil.

TLC R_f = 0.34 (20% EtOAc/hexanes).

¹H NMR (CDCl₃, 500 MHz) δ 6.01 (q, J = 1.8 Hz, 1H), 5.43 (d, J = 5.6 Hz, 1H), 4.70 (s, 1H), 4.65 (s, 2H), 4.43 (d, J = 5.6 Hz, 1H), 4.21–4.09 (m, 4H), 3.63 (d, J = 16.4 Hz, 1H), 3.38 (s, 3H), 3.36–3.31 (m, 2H), 1.34 (s, 6H), 1.26 (t, J = 7.1 Hz, 3H), 0.92 (s, 9H), 0.16 (d, J = 8.3 Hz, 6H).

¹³C NMR (CDCl₃, 126 MHz) δ 169.6, 146.8, 133.1, 126.1, 114.8, 112.1, 96.0, 92.6, 92.1, 85.7, 85.0, 84.4, 82.5, 81.3, 77.4, 68.2, 61.0, 55.7, 33.8, 27.5, 26.3, 26.0, 18.4, 14.3, −4.3, −4.6.

HRMS (+ESI-TOF) m/z calcd for C₂₇H₄₀NaO₇Si⁺ [M+Na]⁺: 527.2436; found 527.2478.

FT-IR (neat) v 3288, 2930, 2857, 1975, 1738, 1472, 1371, 1331, 1254, 1211, 1183, 1152, 1091, 1060, 1036, 1007, 922, 866, 838, 778 cm⁻¹.

Polycyclic Benzene 25

To a 5-mL round-bottom flask charged with a magnetic stir bar was added iodide 20a (11.7 mg, 18.55 μmol, 1 equiv) and anhydrous CH₂Cl₂ (0.37 mL, 0.05 M). The reaction was cooled to −78 °C and added DIBAL-H (0.028 mL, 1 M in PhMe, 27.83 μmol, 1.5 equiv) dropwise. (Note: dropwise addition along the side of the glass was required to prevent over reduction.) The reaction was stirred at −78 °C for 1 h. After complete consumption of the starting material, the reaction was diluted with Et₂O (5 mL), quenched with one drop of H₂O, one drop of 15% aq. NaOH, and two drops of H₂O, then stirred at room temperature for 15 min. The mixture was dried over anhydrous MgSO₄, filtered, and concentrated in vacuo as aldehyde 21a. The crude residue was used in the next step without purification. (Note: aldehyde 21a decomposes upon chromatographic purification.)

To an 8-mL vial charged with a magnetic stir bar was added NiCl₂ (0.5 mg, 3.71 μmol, 20 mol%), CrCl₂ (22.8 mg, 0.1855 mmol, 10 equiv), activated 4 Å molecular sieves (185 mg, 50 mg/mL), and strictly deoxygenated THF (2.7 mL, freshly distilled from sodium/benzophenone) under N₂ atmosphere. A solution of crude aldehyde 21a (10.9 mg, 18.55 μmol, 1 equiv) in strictly deoxygenated THF (1 mL) was added dropwise over 30 min. The reaction was further stirred at room temperature for 3 h, then filtered to a separatory funnel and diluted with Et₂O (5mL) and sat. aq. NH₄Cl (5 mL). The organic layer was separated, and the aqueous layer was extracted with Et₂O (5 mL × 3). The combined organic layers were washed with brine (15 mL), then dried over anhydrous MgSO₄, filtered, and concentrated in vacuo. The crude residue was purified via preparative TLC (40% EtOAc/hexanes) to afford polycyclic benzene 25 (4.3 mg, 50% over two steps, mixture of inseparable diastereomers, 1:1 d.r.) as a clear oil.

TLC R_f = 0.35 (40% EtOAc/hexanes).

¹H NMR (CDCl₃, 500 MHz) δ 7.45–7.36 (m, 4H), 7.29–7.26 (m, 2H), 5.28 (t, J = 5.0 Hz, 2H), 5.14 (d, J = 11.9 Hz, 2H), 4.94 (q, J = 5.8 Hz, 1H), 4.86 (q, J = 3.9 Hz, 1H), 4.71 (d, J = 3.4 Hz, 4H), 4.60–4.56 (m, 6H), 3.41 (d, J = 3.1 Hz, 6H), 2.80–2.63 (m, 4H), 1.38 (s, 6H), 1.33 (d, J = 11.0 Hz, 6H), 0.90 (d, J = 13.7 Hz, 18H), 0.29–0.22 (m, 12H).

¹³C NMR (CDCl₃, 126 MHz) δ 139.3, 138.9, 138.3, 138.2, 137.0, 136.8, 136.3, 135.7, 130.9, 130.6, 127.7, 127.5, 127.5, 127.4, 124.2, 124.1, 111.9, 111.8, 96.0, 95.9, 87.2, 87.0, 85.4, 84.8, 80.3, 80.2, 77.4, 69.0, 69.0, 68.5, 68.4, 55.5, 55.5, 31.7, 31.5, 29.9, 27.9, 27.7, 26.6, 26.4, 26.0, 26.0, −3.9, −3.9, −4.6, −4.6.

HRMS (+ESI-TOF) m/z calcd for C₂₅H₃₈NaO₆Si⁺ [M+Na]⁺: 485.2330; found 485.2335.

FT-IR (neat) v 3446, 2927, 2856, 1463, 1379, 1254, 1208, 1150, 1075, 1051, 920, 865, 838, 776 cm⁻¹.

Supplementary Material

Supplementary Material is available online at https://doi.org/10.1055/a-2951-5568