Enantioselective Synthesis of Sealutomicin C

The sealutomicins are a family of anthraquinone antibiotics featuring an enediyne (sealutomicin A) or Bergman-cyclized aromatic ring (sealutomicins B–D). Herein we report the development of an enantioselective organocatalytic method for the synthesis of dihydroquinolines and the use of the developed method in the total synthesis of sealutomicin C which features a transannular cyclization of an aryllithium onto a γ-lactone as a second key step.


■ INTRODUCTION
A recent report from Igarashi, Sawa, and co-workers described the isolation of four novel anthraquinone-fused natural products, the sealutomicins A−D 1−4, from fermentation of the marine actinomycete Nonomuraea sp.MM565M-173N2 (Figure 1a). 1 In addition to the anthracene-9,10-dione motif shared by all four natural products, sealutomicin A 1 contains a bicyclo[7.3.1]-tridecadiynenecore, placing it in the anthraquinone-fused enediyne (AFE) subfamily of natural products which includes dynemicin A, tiancimycin A, yangpumicin A, and uncialamycin 5. 2−6 AFEs have attracted considerable attention from the synthetic and medicinal chemistry communities owing to their potent antibiotic and antitumor activities, with notable contributions coming from the Myers, Danishefsky, and Nicolaou groups, 7−15 including the synthesis of numerous analogues to develop structure−activity relationships. 7,8,13,14In contrast, sealutomicins B−D 2−4 all possess a bridging aryl ring in place of the enediyne core, proposed to be formed biosynthetically via Bergman cyclization of an enediyne precursor, such as sealutomicin A 1. Such reactivity has previously been implicated in the biosynthesis of the natural product unciaphenol 6 from its enediyne precursor uncialamycin 5. 16 In both cases, aryl ring formation is thought to be preceded by syn-hydrolysis of an epoxide unit, bringing the two alkyne units close enough together to trigger the cyclization.Igarashi, Sawa, and co-workers note in the isolation paper that sealutomicin C 3 may be identical to a compound isolated by Shen and co-workers, namely, the Bergman cyclization product of tiancimycin B 7. 17 The sealutomicins 1−4 all display in vitro antibacterial activity against Gram-positive bacteria with sealutomicin A 1 being more potent than the cycloaromatized sealutomicins B− D 2−4.Sealutomicin A 1 also shows similar effects against Gram-negative bacteria.The antibacterial activity of sealutomicin A 1 is proposed to arise from the ability of the enediyne warhead to form a benzenoid biradical that triggers bacterial DNA scission via hydrogen atom abstraction from the DNA backbone, in the same manner as other AFEs such as uncialamycin 5 and dynemicin A. Sealutomicins B−D all lack this key enediyne motif capable of inducing DNA damage, and as such, the mechanism of action by which they exert their antimicrobial effects is currently unknown.Sealutomicins B−D 2−4 share strong structural similarities with the natural product unciaphenol 6, which was found to display in vitro anti-HIV activity against both wild-type and antiretroviralresistant strains of HIV. 16This raises the possibility that sealutomicins B−D 2−4 may also display similar antiviral properties.In light of the biological activities and intriguing chemical structures, we sought to develop a synthetic route to the sealutomicins.Herein, we report the development of an organocatalytic enantioselective dihydroquinoline synthesis which, along with a transannular cyclization of an aryl lithium onto a γ-lactone, feature as key steps in our total synthesis of sealutomicin C 3. Additionally, our successful synthesis of 3 demonstrates that the Bergman cyclization product of tiancimycin B, isolated by Shen and co-workers, 17 and sealutomicin C 3 have the same structure.

■ RESULTS AND DISCUSSION
Our strategy for the synthesis of sealutomicin C 3 is shown retrosynthetically in Figure 1b.We reasoned that the anthraquinone moiety of sealutomicin C 3 could be installed via a Hauser−Kraus annulation on an iminoquinone formed from oxidation of an alkoxy aniline such as 8, as previously utilized by Nicolaou et al. in the syntheses of uncialamycin 5, 11−13 leading to a protected triol derivative from which 3 could be formed on global deprotection.
The α,β-unsaturated ester functionality in 8 was to be introduced from terminal alkyne 9, which itself would be derived ultimately from ketone 10.The key disconnection in our proposed route to sealutomicin C was to be a 6-exo-trig cyclization of an arylmetal derived from 11 (X = halide) onto a γ-lactone to yield ketone 10.−20 The key cyclization substrate was to be prepared from the dihydroquinoline 12, which we envisaged would be synthesized by condensation of a cinnamaldehyde 14 and an aniline 13 bearing an α-ketoester under enantioselective organocatalysis.
−40 Two early examples involving 1,2-dihydroquinoline synthesis, from the groups of Wang and Coŕdova et al., utilized diarylprolinol silyl ether catalysts to effect an aza-Michael/ aldol cascade coupling of 2-aminobenzaldehydes with cinnamaldehydes in order to access aryl-substituted 1,2dihydroquinolines in high yields and >90% ee. 37,38We sought to adapt these methods to the synthesis of a variety of dihydroquinolines 18 by replacing the 2-aminobenzaldehyde component with the analogous α-ketoester 15 (Table 1).To investigate the proposed modified organocatalytic cascade, representative α-ketoesters 15 were readily prepared from the analogous isatins in two steps (see the Supporting Information, p S3).Using (E)-cinnamaldehyde 16a and α-ketoester 15a, the conditions of Wang et al., 38 namely, catalyst (S)-17a in DCE with 4 Å molecular sieves and sodium acetate as additive, gave poor conversion of reactants to a mixture of the desired 1,2dihydroquinoline (−)-18a (structure confirmed by singlecrystal X-ray diffraction�see the Supporting Information, p S52 and CIF) and a single diastereomer of aldol product (−)-19a [Table 1, entry 1, the configuration of (±)-19a was assigned from 1 H− 1 H NMR coupling constant and 1 H− 1 H NOESY analysis�see the Supporting Information, pp S16− S17 and p S116].Pleasingly, (−)-18a was isolated in high enantiomeric excess, and dehydration of (−)-19a under the reaction conditions both with and without chiral catalyst (S)-17a provided (−)-18a with similarly high enantiomeric excess.
It was found that the additive played a vital role in the catalyst turnover, with benzoic acid in place of sodium acetate resulting in complete consumption of 16a without accumulation of 19a, while requiring a lower excess of 15a (Table 1, entries 3, 4, and 6).Meanwhile, the absence of molecular sieves considerably reduced the turnover (Table 1, entry 5).While the TMS-protected diphenylprolinol catalyst (S)-17b gave the product (−)-18a with notably lower enantiomeric excess in the presence of a basic additive (Table 1, entry 2), with an acidic additive, the product was formed with similar enantiomeric excess as when using catalyst (S)-17a (Table 1, entries 6 and 7).Catalyst (S)-17c bearing a free alcohol resulted in an almost complete lack of reactivity (Table 1, entry 8).With catalyst (S)-17a, the reaction worked similarly well in other chlorinated solvents as well as toluene, but more polar solvents resulted in lower yields and incomplete dehydration to give (−)-18a (Table 1, entries 9−11).Reduction in the loading of molecular sieves was possible and resulted in the optimal reaction conditions which were readily performed on an up to 3 mmol scale (Table 1, entry 12).
identified 18r as having the appropriate core from which to construct sealutomicin C 3, we now embarked on our synthesis of the natural product.The α-ketoester 15r and bromocinnamaldehyde 16r required for the key catalytic enantioselective dihydroquinoline synthesis were both prepared on a scale in a single step from cheap, commercially available starting materials.Carbamate protection of 5-methoxyisatin 20 followed by alcoholysis in acidic isopropanol gave 15r in 78% yield, while Wittig homologation of 2-bromobenzaldehyde 21 using the ylide derived from 22 followed by subsequent acetal hydrolysis gave 16r in 79% yield (Scheme 1). 41,42Under the optimized organocatalytic conditions, using (R)-17a as the catalyst, dihydroquinoline (+)-18r was obtained in 89% on a 7 g scale with an enantiomeric excess of >98%.Subsequent Luche reduction of (+)-18r gave α,β-unsaturated lactone 23. 43It was found that the addition of CeCl 3 •7H 2 O was crucial for suppressing unwanted over-reduction to the corresponding saturated lactone.With lactone 23 in hand, we turned our attention to the installation of the 1,2-syn diol.Ruthenium-(VIII)-catalyzed dihydroxylation of 23 gave the syn-diol 24 as a single diastereomer in 88% yield, which was assigned as the (R,R) diastereomer in the expectation that dihydroxylation had occurred from the less hindered face of the fully substituted alkene (vide infra). 44Protection of 24 as the bis-paramethoxybenzyl (PMB) ether 25 under basic conditions now set the stage for the crucial cyclization to form the bridged bicyclic core of sealutomicin C. Treatment of bromide 25 in THF with a slight excess of n-butyllithium at −78 °C and subsequent warming to room temperature gave the desired ketone 26 in 94% yield on a 1 g scale.By 1 H NMR, the ketone 26 was formed as a mixture with the corresponding lactol, resulting in a complex spectrum; however, simple oxidation of 26 with the Dess−Martin periodinane gave the aldehyde 27 with a much simplified 1 H NMR spectrum.Aldehyde 27 showed a 1 H− 13 C HMBC correlation between the indicated aromatic proton (δ H 7.88 ppm) and the ketone carbonyl carbon (δ C 190 ppm) demonstrating that successful cyclization

Journal of the American Chemical Society
had occurred which also demonstrated that dihydroxylation had indeed occurred on the less hindered face of the α,βunsaturated-γ-lactone 23.The aldehyde 27 was readily converted into the terminal acetylene 28 under standard conditions in the presence of the aromatic ketone. 45,46duction of the aromatic ketone 28 with lithium borohydride gave the corresponding secondary alcohol 29 as a single diastereomer whose configuration was tentatively assigned by 1 H− 1 H NOESY analysis and later confirmed by 1  the benzylic alcohol as its PMB ether, the terminal alkyne was carboxylated under basic conditions with methyl chloroformate to give alkynyl ester 30.Exposure of propargylic ester 30 to an excess of the organocuprate derived from methyllithium and copper(I) cyanide at 0 °C gave alkenyl ester 31 in 47% yield with the double bond configuration assigned through 1 H− 1 H NOESY and ROESY experiments on later synthetic intermediates.−49 To this end, hydrogenolysis using palladium on carbon enabled selective cleavage of the benzyloxycarbamate protecting group to give the corresponding aniline, which on treatment with cerium(IV) ammonium nitrate (CAN) gave iminoquinone 32 in 75% yield over two steps. 50Treatment of 32 with the anion generated from 3cyanophthalide and LiHMDS resulted in rapid iminoquinone consumption, giving anthraquinone 33 in 86% yield; 1 H− 1 H ROESY analysis of 33 was entirely in keeping with the depicted configuration.Attempts to remove the three PMB ether groups from 33 using Lewis-acidic boron trichloride dimethyl sulfide complex 51 led to degradation of the starting material which also occurred when attempting the deprotection with CAN.Global deprotection was ultimately achieved upon treatment of 33 with excess 2,3-dichloro-5,6dicyano-1,4-benzoquinone (DDQ), yielding sealutomicin C 3 in 70% yield.The 1 H and 13 C NMR data (DMSO-d 6 ) for the synthetic sealutomicin C were in good agreement with those reported for the natural product. 1 Additionally, the circular dichroism spectrum of our synthetic material matched closely with that of the isolated natural product, indicating that the natural enantiomer of sealutomicin C had been synthesized.The 1 H and 13 C NMR spectra (acetone-d 6 ) for synthetic sealutomicin C were in agreement with those reported by Shen and co-workers for the Bergman cyclization product of tiancimycin B 7, indicating that sealutomicin C and cycloaromatized tiancimycin B have the same structure.

■ CONCLUSIONS
In conclusion, we have developed the first enantioselective synthesis of sealutomicin C 3 and demonstrated that it has the same structure as the Bergman cyclization product of tiancimycin B. The synthesis proceeded in 16 steps (longest linear sequence) from 5-methoxy isatin 20 and required the development of a methodology for the organocatalytic enantioselective synthesis of dihydroquinolines from αketoesters and cinnamaldehydes, as well as the use of a key intramolecular cyclization of an aryllithium onto a γ-lactone.Work toward the total synthesis of other members of the sealutomicins is ongoing.
Experimental procedures, characterization data, NMR spectra of reported compounds, and CIF files (PDF)

Chart 1 .
Substrate Scope for Enantioselective Organocatalytic Dihydroquinoline Synthesis a Yield based on cinnamaldehyde starting material.b Incomplete conversion to a dihydroquinoline product.
Scheme 1.Total Synthesis of Sealutomicin C a