Ring Expansion of 1-Indanones to 2-Halo-1-naphthols as an Entry Point to Gilvocarcin Natural Products.

Herein, we describe a two-step ring expansion of 1-indanones to afford 2-chloro/bromo-1-naphthols (32 examples). The developed method shows broad functional group tolerance, benefits from mild reaction conditions, and enables rapid access to the tetracyclic core of gilvocarcin natural products. The orthogonally functionalized products allow for selective postmodifications as exemplified in the total synthesis of defucogilvocarcin M. For the selective oxidation of the chromene, a mild and regioselective oxidation protocol (DDQ and TBHP) was developed.

Polyfunctionalized aromatic structures that are derived from 1-naphthols are present in bioactive natural products, numerous pharmaceuticals, and chiral ligands.[1] According to the substituents present at the ortho, meta, and para positions, two major classes can be identified (Scheme a). Class A comprises 3-carboxy-1-naphthols with variations at the ortho and para position as exemplified by chartartin[2] (1), salimabromide[3] (2), and diphyllin[4] (3). On the contrary, parviflorene E[5] (5), the VANOL ligand[6] (4), and the gilvocarcin natural product ravidomycin[7] (6) represent ortho-substituted 1-naphthols with different degrees of substitution at the meta and para position (class B). The potent biological activities associated with these structures as well as their use in asymmetric catalysis have attracted a great deal of attention for the development of efficient methods for their synthesis.[8] Much effort has been spent to access orthogonally functionalized 1-naphthols. Despite significant progress in this area, the developed methods often involve multistep sequences,[9] harsh reaction conditions,[10] and the use of precious transition-metal catalysts[11] or require a β-ketoester functionality to proceed.[12] Other protocols are based on sensitive and uncommon intermediates or reagents (e.g., cyclobutenones, allenes, or nitrones).[13] Moreover, they suffer from noncommercial starting materials, thus preventing rapid access to structurally diverse analogues. Here, we present a robust two-step protocol for the construction of orthogonally functionalized 2-halo-1-naphthols starting from 1-indanones. A plethora of 1-indanones with a broad substitution pattern are commercially available, and functionalized variations thereof are readily accessible via known literature procedures.[14] The applicability of the developed methodology is shown for the synthesis of the natural product defucogilvocarcin M (45).[15]

Scheme 1

Selected Examples and Synthetic Access to Chemically and Biologically Relevant 1-Naphthols

During the course of our investigations to develop novel ring expansion reactions, we gained access to a variety of class A 1-naphthols (Scheme b).[16] This protocol enabled the synthesis of chartarin (1) and also provided access to an advanced intermediate toward salimabromide (2).[17] While a diverse set of (hetero)arenes were generated via this strategy, the inherent ester functionality restricted synthetic access to class A structures and variation of the ortho position was possible only at the stage of the 1-indanone (7a). In addition, several of the 1-indenone intermediates (7b) required for the cyclopropanation were unstable and prone to polymerization. We wanted to address these issues by investigating the ring expansion of gem-dihalocyclopropane[18]10, readily available from indene 7c. During our early investigations, Wang found access to 2-fluoro-1-naphthols via a related cyclo-propanation–ring expansion (CPRE) using (bromodifluoromethyl)trimethylsilane.[10d] The incorporation of a chlorine or bromine atom at this position was not possible via this strategy and restricted further diversification. Inspired by seminal work by Ciamician and Dennstedt and related reports on ring expansion reactions,[18,19] we envisioned an alternative strategy that employs chloroform and bromoform as inexpensive and easy to handle halogen sources for installing the chloride and bromide, respectively. The obtained ortho-chlorinated and brominated 1-naphthols 11b are more valuable substrates than their fluorinated analogues 11a, especially when considering further postmodifications to access a large number of known bioactive 1-naphthols featuring an ortho substituent.[20]

We began our investigations by studying the CPRE of 1-indanone-derived trimethylsilyl enol ether 12 (Table ). While aqueous sodium hydroxide (entry 1) or sodium methanolate (entry 2) led to only desilylation of the starting material, we were delighted to find that upon treatment of 12 with potassium tert-butoxide and chloroform in pentane at cryogenic temperatures ring expansion followed by partial in situ deprotection to 1-naphthol 14 was observed (entry 3). To ensure complete desilylation, hydrochloric acid (entry 4) or tetrabutylammonium fluoride (TBAF, entry 5) was added after full conversion of the starting material. While the use of acidic conditions provides slightly higher yields for 12, TBAF proved to be superior with regard to functional group tolerance. It is noteworthy that the use of sublimed grade potassium tert-butoxide showed significantly higher yields compared to those of reagent grade batches. In this context, we were likewise interested in gaining access to 2-bromo-1-naphthols to expand the range of possible postmodifications of the obtained 1-naphthols (Table ). A simple exchange of chloroform for bromoform gave the desired naphthol 17 in moderate yield (47%) accompanied by large amounts of recovered 1-indanone (entry 1). The competing desilylation was prevented by employing a more stable tert-butyldimethylsilyl enol ether. This allowed for the preparation of 2-bromo-1-naphthol 16 even at ambient temperature (entry 2). However, larger amounts of the base and bromoform were needed to ensure full conversion (entry 3). Although the combination of this protocol with deprotection conditions (DBU in MeCN/H2O or HF·pyr in THF) in a one-pot fashion afforded unprotected naphthol 17 in good yields, we observed reproducibility issues leading to varying yields between 56% and 76%. We also noticed that application of these conditions to a broader substrate scope led to significantly lower yields, not only at the stage of the ring expansion but also for the subsequent deprotection step. The inconsistencies of the subsequent TBS deprotection required another change of the protecting group. We later found that the use of a triisopropyl (TIPS) group was ideally suited as it provided good yields for the enol ethers and could be easily removed upon treatment with either TBAF or a suspension of KOAc in DMF/water (entry 4).[21] Detailed studies showed that the reproducibility of the CPRE step was strongly dependent on the order and temperature at which the substrate and the base were combined. While addition at 23 or 0 °C immediately afforded a deep purple solution, addition at −78 °C led to the formation of a pale-yellow mixture and provided 17 in reproducible 85% yield (entry 5). Efforts to identify and characterize possible side products resulting from a competing aryne formation were unsuccessful.

Table 1

Selected Screening Conditions for the Preparation of 2-Chloro-1-naphthols

entry	reagents	temp	time	solvent	deprotection	yield of 13 (%)	yield of 14 (%)
1	CHCI3, NaOH, BnEt3NCI	45 °C	3 days	CH2CI2, H2O	–	0	0
2	CCI3COOEt, NaOMe	0 °C	4 h	pentane	–	0	0
3	CHCI3, KOt-Bu	–78 to 23 °C	3 h	pentane (0.5 M)	–	10	55
4	CHCI3, KOt-Bu	–78 to 23 °C	3 h	pentane (0.5 M)	aqueous HCI	0	86
5	CHCI3, KOt-Bu	–78 to 23 °C	2 h	pentane (0.2 M)	TBAF	0	80

Table 2

Selected Screening Conditions for the Preparation of 2-Bromo-1-naphthols

entry	R	KOt-Bu (equiv)	CHBr3 (equiv)	base addition	deprotection	yield of 15 (%)	yield of 16 (%)	yield of 17 (%)
1	TMS	2.0	2.2	at –78 °C	TBAF, THF	0	0	47
2	TBS	2.0	2.2	at 23 °C	-	43	36	0
3	TBS	6.0	5.0	at 23 °C	DBU, MeCN/H2O	0	0	67
4	TIPS	4.5	2.0	at 23 °C	KOAc, DMF/H2O	0	0	70
5	TIPS	4.5	2.0	at –78 °C	KOAc, DMF/H2O	0	0	85

With the optimized conditions in hand, we began investigating the conversion of several substrates to the corresponding 2-chloronaphthols (Scheme , protocol A). We found that halogens (19Cl–24Cl), acetals (25Cl), ethers (26Cl–28Cl), esters (30Cl and 31Cl), alkyls (33Cl), and aryls (32Cl) and silyl ethers (29Cl) were stable under the reaction conditions to afford the corresponding 1-naphthols in yields of ≤83%. Unexpectedly, only the presence of methoxy groups led to significantly lower yields under the standard conditions (16% for 27Cl, 57% for 28Cl).[22] This was attributed to the decreased stability of the transient silyl enol ether. We were able to address this issue by adapting the conditions developed for the preparation of 2-bromonaphthols (compare Table ). Under these conditions, 27Cl and 28Cl were obtained in 83% and 81% yields, respectively.

Scheme 2

Scope of 2-Chloro- and 2-Bromo-1-naphthols Obtained via CPRE of 1-Indanones

Via TIPS-silylenol ether.

Yield for TIPS-protected naphthol (see the Supporting Information for details).

TBAF deprotection.

When the substrates mentioned above were subjected to protocol B, comparable yields were obtained for halogenated naphthols 19Br–24Br, benzyl ether 26Br, silyl ether 28Br, and p-phenyl derivative 32Br (Scheme ). However, the protocol was less compatible with electron-donating groups such as an acetal (25Br) or methoxy unit (27Br and 28Br) and failed in the presence of an ester (30Br and 31Br). The slightly decreased yield for ortho,meta-substituted naphthol 33Br can be rationalized by steric hindrance. In the course of investigating further postmodifications to showcase the applicability of the obtained 2-halo-1-naphthols shown in Scheme , we observed an unusual dearomatization reaction. When 2-bromo-5-iodonaphthol 22 was treated with N-chlorosuccinimide (NCS) in acetonitrile, quantitative conversion to bench-stable enone 47 was observed without formation of expected naphthol 46.[23] We found that this rather rare dienone tautomer[24] undergoes conjugate addition[25] with several nucleophiles, thus representing a formal meta-functionalization (see the Supporting Information).

Having prepared a library of 2-halonaphthols, we turned our attention to the synthesis of defucogilvocarcin M [45 (Scheme )].[15] This natural product belongs to a family of >15 antitumor antibiotics, of which the first member was isolated in 1955.[26] Due to their structural and biological properties, defucogilvocarcin M and its related members have become a popular synthetic target.[15b,26] Starting from known indanone 34,[27] 2-bromonaphthol 35 was obtained in 70% yield over two steps on a gram scale. Two-step oxidation gave dihydroquinone 37, which was regioselectively benzylated with 38(28) in the presence of potassium carbonate to give ethers 39. Subsequent methylation provided the key benzyl ethers 40. Among the different known strategies for forming the gilvocarcins’ biaryl bond (e.g., Meerwein, Suzuki, Stille, Heck, and Meyers coupling), no Ar–X–Ar–X (X = halogen) coupling has been reported so far.[15b,26] Somewhat surprisingly, all attempts to realize a Ni- or Pd-catalyzed intramolecular sp2–sp2 cross-coupling or a classical Ullmann coupling[29] failed in our hands. After a survey of alternative methods, Lipshutz’s Cu(I)-mediated biaryl coupling protocol (t-BuLi, CuCN·2LiX) evolved as the first solution for obtaining the full skeleton of 45 (procedure a).[30] When an excess of t-BuLi (11 equiv) was used, simultaneous removal of the benzyl group took place to form 41b, sparing an additional deprotection step (procedure b).[31] Prolonged treatment with 1,3-dinitrobenzene (>1.5 h) led to overoxidation and thus opening of the lactone ring (not shown). Due to unsatisfactory yields, we screened further coupling conditions and were delighted to see that Stille–Kelly coupling[32] afforded the desired tetracycle 41a in 63% yield, with the results for 40Br being better than those of 40I (see the Supporting Information for a detailed screening table). Severe and unanticipated difficulties awaited us when we attempted the oxidation of chromene 41 to install the missing lactone unit. For this purpose, we initially protected the free hydroxy-chromene 41b. Compound 41c resisted oxidation to the corresponding chromenone 43 by several established procedures, including PCC, PDC, SeO2, KMnO4, MnO2, TBHP/KI, or TBHP/I2.[33] In most cases, ring opening of the intermediate lactol to give the corresponding benzoquine or decomposition was observed. Progress was made when we found that treating a solution of 41c in 1,4-dioxane with DDQ and TBHP resulted in the formation of peroxyacetal 42c (R = i-Pr). When the partially purified peroxyacetal 42c was treated with 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) in dichloromethane, a Kornblum–DeLaMare rearrangement[34] to the desired lactone took place. We were pleased to see that this transformation could also be applied in a one-pot fashion affording isopropylated defucogilvocarcin M (43) in 85% yield. Moreover, these conditions not only were completely selective for the chromene core in the presence of a benzyl group (44) but also tolerated the free hydroxyl group of naphthol 41b to directly give defucogilvocarcin M (45) in 80% yield. Spectroscopic data (1H NMR and 13C NMR) for the synthetic material were in full agreement with reported values.[35]

Scheme 3

Application of the CPRE Protocol to the Synthesis of Defucogilvocarcin M

In summary, we have developed a powerful protocol for converting a broad range of readily available 1-indanones into diversely substituted 2-chloro/2-bromo-1-naphthols. The halogen in the ortho position served as a useful handle for further functionalization as demonstrated in the synthesis of defucogilvocarcin M. In addition, a mild protocol for the selective benzylic oxidation of chromenes was developed.