A Biomimetic Synthesis of des-Hydroxy Paecilospirone

Dedicated to purveyors of truth: Tomáš Hudlický, Richard H. Schlessinger, Samuel J. Danishefsky

Abstract

The carbon framework of des-hydroxy paecilospirone was rapidly synthesized using a biomimetic approach whereby an enol ether and an ortho-quinone methide (o-QM), each derived from the same lactone, were combined to arrive at the complete carbon skeleton of paecilospirone.

Key words

ortho-quinone methide - Diels–Alder - paecilospirone - benzylic oxidation - enol ether

Scheme 1 The structure of paecilospirone (1) and speculated biosynthesis

Paecilospirone was first collected from the marine fungus Paecilomyces sp. near Yap Island.[1] It was determined to have a rather unique spiro[chroman-2,1′(3′H)-isobenzofuran] structure that reportedly inhibited microtubule assembly. In 2011, Brimble and coworkers reported a total synthesis of (+)-paecilospirone in 19 steps.[2]

Our synthetic curiosity in this natural product emerged from an appreciation of its putative biosynthesis, which we propose (Scheme [1]). We suppose that salicylic alcohol 2 extrudes water by different mechanisms to either form the o-quinone methide (o-QM) 3 or the enol ether 4. Nature then combines these two entities to arrive at the tetracycle 5, which undergoes further site selective benzylic oxidation to afford paecilospirone A (1). We decided to attempt to replicate this efficient process in the laboratory.

Scheme 2 Synthesis of o-QM precursor 8 and enol ether 9

We suspected that the unprotected salicylic alcohol 2 might prove particularly unstable and highly reactive under chromatographic conditions, and thus it would not be a prudent choice of starting material. We therefore chose to begin our investigation with the phthalide 7, which we imagined to be more easily harnessed for our synthetic purposes (Scheme [2]). Compound 7 was prepared in a regioselective manner as first shown by Buehler.[3] Benzoic acid (6) and formalin were combined at ambient temperature for 24 hours in the presence of concentrated hydrogen chloride. These slightly modified conditions afforded the stable lactone 7 in a respectable 35% yield. The lactone 7 displays three sp² atoms, positioning its five-membered ring into a nearly planar arrangement, which positions RC(=O)–O bond orthogonal to the aromatic π-system. Thus, when the phenol is deprotonated, it is not expelled as carboxylate to form the corresponding o-QM, affording stability to this intermediate. For this reason, phenol 7 underwent smooth benzylation under standard basic conditions to afford the benzyl ether 8 in a 95% yield. The carbonyl of the lactone 8 was subsequently found to form the hereto unknown enol ether 9 as a single geometric isomer in an 87% yield by application of Takai’s method with the dibromide 10.[4] [5] [6] [7] Compound 9 thus served as a surrogate of intermediate 4 postulated in the biosynthesis (Scheme [1]).

Scheme 3 Synthesis of a more manageable o-QM precursor from 8

To construct a more manageable o-QM precursor leading to something resembling compound 3 in the proposed biosynthesis, the lactone ring in 8 would be opened, an appropriate electron-deficient leaving group installed on the benzylic alcohol, and the phenol protecting group then cleaved (Scheme [2]). This proposition was accomplished in three steps by exposure of compound 8 to pyrrolidine and trimethylaluminum to form benzyl alcohol 11, subsequent acetylation, and careful debenzylation under as neutral conditions as possible to provide the phenol 12 in 78% overall yield (Scheme [3]).

Scheme 4 Base-triggered o-QM generation and cycloaddition

As anticipated from our earlier work with base-promoted generation of o-QMs,[8] [9] we found that when an ethereal mixture of phenol acetate 12 and enol ether 9 was treated at –78 °C with tert-butylmagnesium chloride, the spiroketal 14 was formed in 64% yield (Scheme [4]). We presume this smooth reaction is due to the controlled generation of the o-QM intermediate 13 by base-mediated elimination of the acetate, and subsequent inverse demand cycloaddition with the enol ether 9.

Table 1 Oxidation of Amide 14

Entry	Conditions	Main product
1	DDQ, H2O, DCM	15
2	DDQ, AcOH, DCM	15
3	DDQ, MeOH	16
4	NaIO4, RuCl3·H2O, CCl4, H2O, MeCN	17

With this material in hand, we were now poised to explore selective oxidation of the chroman core (Table [1]). One of the reasons that we chose the amide 14 (Scheme [4]), instead of something more closely resembling the ketone 5 (Scheme [1]), was that the knowledge that amides could serve as surrogates for the respective ketone and acid functionalities. Unfortunately, the amide 14 failed to undergo the desired oxidation. Instead, we found that upon exposure of compound 14 to DDQ in water or acetic acid, the chromene 15 formed in low yield. On the other hand, when the DDQ oxidation was performed in methanol, a methoxy group was installed on the isobenzofuran ring resulting in a diastereomeric mixture of the acetals 16. Furthermore, we found that application of ruthenium tetroxide caused oxidation of the pyrrolidine ring to afford the corresponding lactam 17. While the chromene 15 might have proven useful in accessing the desired alcohol, the low yield of the chromene (<20%) dissuade further investigations. In addition, our attempts to remotely deprotonate amide 14 at its γ-position and oxidize the corresponding enolate, or some derivative of it, with DMDO or other suitable oxidants, also failed.

Scheme 5 Synthesis of acid 19

We therefore decided to investigate the conversion of the amide into the corresponding carboxylic acid in the hope that this functional group might facilitate oxidation of the chroman ring. This task was accomplished by a two-step procedure. First, the amide 14 was converted into the aldehyde 18 in 63% yield by its reduction with Schwartz’s reagent. Subsequent Pinnick oxidation afforded the acid 19 in 87% yield (Scheme [5]).[10] Unfortunately, we were unable to oxidize this material in the desired manner using a variety of procedures, such as hypervalent iodide,[11] [12] lead tetraacetate,[13] or Pd(II).[14]

Scheme 6 Synthesis of des-hydroxy paecilospirone (5)

Unable to facilitate the desired benzylic oxidation of the either the amide 14 or then acid 19 in an acceptable yield the aldehyde 18 was elaborated into des-hydroxy paecilospirone (5) over three steps (Scheme [6]). Nucleophilic addition of octyl magnesium bromide to the aldehyde 18 afforded the alcohol 20 in 90% yield. This material was converted into ketone 21 in 90% yield by a Swern oxidation. Hydrogenolysis of the benzylated phenol group furnished des-hydroxy paecilospirone (5) in 89% yield.[15]

In summary, a synthetic strategy resembling what we believe to be the biosynthesis of paecilospirone has been completed. Our approach utilizes an o-QM and an enol ether synthesized from the same starting material and arrives at des-hydroxy paecilospirone (5) in just ten steps. The approach may be amenable to enantioselective control through a chiral amide. Perhaps a method for the selective oxidation of such complex systems, as compounds 14, 19, and 5 with their numerous benzylic sites, will be discovered that can emulate nature’s abilities which routinely surpass our own.

• References and Notes

• 1 Namikoshi M. Kobayashi H. Yoshimoto T. Meguro S. Chem. Lett. 2000; 308
• 2a Yuen T.-Y. Yang S.-H. Brimble MA. Angew. Chem. Int. Ed. 2011; 50: 8350
• 2b Jay-Smith M. Furkert DP. Sperry J. Brimble MA. Synlett 2011; 1395
• 3 Buehler CA. Powers TA. Michels JG. J. Am. Chem. Soc. 1944; 66: 417
• 4 Okazoe T. Takai K. Oshima K. Utimoto K. J. Org. Chem. 1987; 52: 4410
• 5 Takai K. Kakiuchi T. Kataoka Y. Utimoto K. J. Org. Chem. 1994; 59: 2668
• 6 Hoffmann RW. Bovicelli P. Synthesis 1990; 657
• 7 Spaggiari A. Vaccari D. Davoli P. Torre G. Prati F. J. Org. Chem. 2007; 72: 2216
• 8 Green JC. Burnett GL. Pettus TR. R. Pure Appl. Chem. 2012; 84: 1621
• 9 Bai W.-J. David JG. Feng Z.-G. Weaver MG. Wu K.-L. Pettus TR. R. Acc. Chem. Res. 2014; 47: 3655
• 10 Bal BS. Childers WE. Pinnick HW. Tetrahedron 1981; 37: 2091
• 11 Togo H. Muraki T. Hoshina Y. Yamaguchi K. Yokoyama M. J. Chem. Soc., Perkin Trans. 1 1997; 787
• 12 Togo H. Muraki T. Yokoyama M. Tetrahedron Lett. 1995; 36: 7089
• 13 Davies DI. Waring C. J. Chem. Soc. C 1968; 2337
• 14 Lyons TW. Sanford MS. Chem. Rev. 2010; 110: 1147
• 15 Compound 5: ¹H NMR (500 MHz, CDCl₃) δ = 7.30–7.24 (m, 2 H), 7.16 (t, J = 7.9 Hz, 1 H), 6.95–6.90 (m, 2 H), 6.80 (d, J = 7.9 Hz, 1 H), 5.30 (s, 1 H), 5.25 (d, J = 12.8 Hz, 1 H), 5.07 (d, J = 12.8 Hz, 1 H), 3.30 (dd, J = 17.6, 5.5 Hz, 1 H), 3.01–2.84 (m, 3 H), 2.36–2.25 (m, 1 H), 1.77–1.67 (m, 2 H), 1.46–1.07 (m, 22 H), 0.87 (t, J = 7.0 Hz, 3 H), 0.83 (t, J = 7.1 Hz, 3 H). ¹³C NMR (126 MHz, CDCl₃) δ = 204.8, 153.5, 150.0, 141.0, 138.2, 129.8, 126.9, 126.7, 122.4, 121.4, 120.6, 115.7, 114.5, 111.6, 70.6, 41.9, 38.2, 31.8, 31.8, 30.6, 29.6, 29.4, 29.4, 29.2, 29.1, 26.8, 26.1, 24.6, 22.7, 22.6, 14.1, 14.1. IR ν_max (neat): 3393, 2925, 2854, 1670, 1607, 1455, 1287, 1258, 927. HRMS (ESI+) calcd for C₃₂H₄₄O₄Na [M+Na]⁺: 515.3137; found: 515.3126