Synthesis of (Z)-2-acyl-2-enals via retrocycloadditions of 5-acyl-4-alkyl-4H-1,3-dioxins: Application in the total synthesis of the cytotoxin (±)-euplotin A [7]

Full text of DOI:10.1021/ja011470b

J. Am. Chem. Soc. 2001, 123, 9455-9456
9455
Our initial objective was to examine the stereoselectivity of
the retrocycloaddition as a function of the acyl substituent of the
1,3-dioxin 6 (Scheme 2). To that end, the cyclohexyl imine
derivative of commercially available 4H-1,3-dioxin-5-one was
metalated and alkylated to provide the 6-alkyldioxinone 5.
Regioselective conversion of ketone 5 to the corresponding less-
substituted enol triflate was accomplished by kinetic deprotonation
with NaHMDS in the presence of N-phenyltrifluoromethane-
sulfonimide. This enol triflate was then subjected to a variety of
palladium-catalyzed carbon monoxide insertion reactions (6a,b,c)
as well as a Heck reaction with ethyl vinyl ether (6d) to deliver
the desired 5-acyl-4-alkyl-4H-1,3-dioxins.
Synthesis of (Z)-2-Acyl-2-enals via
Retrocycloadditions of 5-Acyl-4-alkyl-4H-1,3-dioxins:
Application in the Total Synthesis of the Cytotoxin
(()-Euplotin A
Ronald A. Aungst, Jr. and Raymond L. Funk*
Department of Chemistry, PennsylVania State UniVersity
UniVersity Park, PennsylVania 16802
ReceiVed June 15, 2001
The 5-acyl-3,4-dihydro-2H-pyran substructure and its reduced
analogue are common to a variety of natural products which
possess a diverse array of biological properties (Scheme 1). A
well-established strategy for the construction of this key structural
unit is via the heterocycloaddition reaction of an unsaturated
carbonyl compound,1 preferably as an activated 3-acyl-1-oxadiene,
with an alkene. Many conceivable implementations of this viable
approach require the stereocontrolled syntheses of (Z)-2-acyl-2-
enals. For example, a projected synthesis of the cytotoxin euplotin
A (1)2 might proceed through an exo cycloaddition (vide infra)
of the dihydrofuran moiety with the (Z)-2-acyl-2-enal 2. Unfor-
tunately, methodology for the stereoselective generation of
2-acylenals under the requisite mild conditions is unavailable.3
We have recently demonstrated that 5-acyl-4H-1,3-dioxins are
viable precursors to 2-acylpropenals via facile retrocycloaddition
reactions.4a The extension of this protocol to the stereoselective
preparation of the more highly functionalized (Z)-2-acylenals and
its application in the first total synthesis of (()-euplotin A are
reported herein.
Scheme 2a
a Reagents: (a) LiNEt2; ICH2CH2OTES, -78 °C to rt, 2 h, 87%; (b)
NaHMDS, PhNTf2, -78 °C to rt, 2 h, 93%; (c) Pd(OAc)2, dppp, i-Pr2NEt,
K2CO3, CO, THF, MeOH, 86% for 6a; Pd(OAc)2, dppp, i-Pr2NEt, DMF,
HN(OMe)Me/HNMe2, 81 and 90% for 6b and 6c, respectively; Pd(OAc)2,
ethyl vinyl ether, DMSO, 48 h, H3O+, 93% for 6d; (d) CDCl3, 50 °C, 36
h, 99% for 7a; toluene, 90 °C, 2 h, 93% for 7b; toluene, 100 °C, 1 h,
85% for 7c; (e) 5 equiv iso-butyl vinyl ether, CH2Cl2, rt, 24 h, 81% for
8a; 12 kbar, 18 h (71 and 85% for 8b and 8c, respectively; (f) toluene,
110 °C, 8 h, 99% for 9a; toluene, 95 °C, 2 h, 95% for 9b.
Scheme 1
On the basis of previous retrocycloadditions of dioxins
analogous to 6 which lacked the 5-acyl substituent,4b we
anticipated that the retrocycloaddition of 6 would proceed
preferentially through the boat conformer eq-6 rather than the
boat conformer ax-6 which suffers from a flagpole-flagpole
interaction with the O(3) axial lone pair. However, in this case,
the preference for eq-6 might be attenuated by an A1,2 interaction
between the acyl substituent and the C(4) equatorial substituent.
Indeed, we were pleased to discover that the amides 6b and 6c
afforded the desired (Z)-2-acyl-2-enal with high or complete
stereoselectivity, respectively, although the more facile (50 °C)
retrocycloaddition of ester 6a was less stereoselective. Moreover,
each of the acylenals could be converted by subsequent treatment
with iso-butyl vinyl ether to an endo/exo mixture of the targeted
5-acyl-3,4-dihydro-2H-pyrans 8. However, further experimenta-
tion revealed that the Z/E stereoisomeric ratios for enals 7 are
likely the result of thermodynamically controlled isomerizations.
Thus, the amide 7b was photoisomerized (Hanovia 500 W,
toluene, 3 h) to a mixture of isomers (Z:E ) 80:20) which upon
heating (toluene, 1 h) afforded a ratio of isomers (Z:E ) 96:4)
nearly identical to that initially obtained.5 It would appear that
2-acylenals are particularly prone to thermal isomerization6 since
the dioxins 9 all afforded a single stereoisomer of enals 10 and
photoisomerized stereoisomeric mixtures of enals 10 remained
(1) For reviews, see: (a) Desimoni, G.; Tacconi, G. Chem. ReV. 1975, 75,
651. (b) Tietze, L. F.; Beifuss, U. Angew. Chem., Int. Ed. Engl. 1993, 32,
131. (c) Waldmann, H. Synthesis 1994, 535. (d) Tietze, L. F. Chem. ReV.
1996, 96, 115. (e) Tietze, L. F.; Kettschau, G. Top. Curr. Chem. 1997, 189,
1. (f) Tietze, L. F.; Modi, A. Med. Res. ReV. 2000, 20, 304.
(2) In addition to euplotin A, two additional cytotoxic natural products,
euplotin B (∆10,11) and euplotin C (C-8 CO ) CH2), were isolated from the
ciliated protist Euplotes crassus and were shown to inhibit the cell division
of or kill the related marine ciliates E. Vannus and E. minuta via cell-to-cell
encounters, thereby imparting a competitive advantage to the former ciliate,
see: (a) Dini, F.; Guella, G.; Giubbilini, P.; Mancini, I.; Pietra, F. Naturwis-
senschaften 1993, 80, 84. (b) Guella, G.; Dini, F.; Tomei, A.; Pietra, F. J.
Chem. Soc. P. T. 1 1994, 17, 161. (c) Guella, G.; Dini, F.; Pietra, F. HelV.
Chim. Acta 1996, 79, 710.
(3) For the preparation of alkylidenemalonaldehydes, see: (a) Tietze, L.-
F.; Glusenkamp, K.-H.; Holla, W. Angew. Chem., Int. Ed. Engl. 1982, 21,
793. (b) Arnold, Z.; Kryshtal, G. V.; Kral, V.; Dvorak, D.; Yanovskaya, L.
A. Tetrahedron Lett. 1988, 29, 2861. For the in situ generation 2-(trichloro-
acetyl)-2-enals, presumably as a mixture of stereoisomers, see: (c) Tietze, L.
F.; Meier, H.; Nutt, H. Chem. Ber. 1989, 122, 643. (d) Tietze, L. F.; Meier,
H.; Nutt, H. Liebigs Ann. Chem. 1990, 253.
(5) In addition, oxidation (Dess-Martin, rt, 1 h) of the known methyl E-2-
(hydroxymethyl)-2-pentenoate (Charette, A. B.; Coˆte´, B.; Monroc, S.; Prescott,
S. J. Org. Chem. 1995, 60, 6888) gave rise to a mixture of stereoisomers
(Z:E ) 70:30) of methyl 2-formyl-2-pentenoate.
(4) (a) Funk, R. L.; Fearnley, S. P.; Gregg, R. Tetrahedron 2000, 56, 10275.
(b) Funk, R. L.; Bolton, G. L. J. Am. Chem. Soc. 1988, 110, 1290.
10.1021/ja011470b CCC: $20.00 © 2001 American Chemical Society
Published on Web 08/31/2001

9456 J. Am. Chem. Soc., Vol. 123, No. 38, 2001
Communications to the Editor
Scheme 3
Scheme 4
unchanged upon heating in toluene. Finally, it should be noted
that the ketoenal (7, R ) Me) derived from thermolysis of dioxin
6d could not be isolated and led to a complex mixture of
dimerization and enolization products. Moreover, a mixture of
four products derived from endo/exo cycloaddition with both the
enal as well as the enone heterodiene moieties was obtained upon
in situ trapping with iso-butyl vinyl ether.
Having completed our preliminary studies, we turned to the
application of this methodology in the total synthesis of euplotin
A. It was now clear that the keto acylenal 2 would be an unsuitable
cycloaddition substrate, and hence, we focused on the preparation/
generation of the 2-(N-methoxy-N-methylamido)enal 11 (Scheme
3). Although the (Z)-2-acyl-2-enal stereochemistry of 11 was now
assured, the stereoselectivity of the subsequent heterocycloadditon
reaction was not. Thus, an endo transition state would lead to
the significantly less strained all-cis isomer rather than the
stereochemistry found in euplotin A which would be derived from
an exo transition state. Molecular mechanics calculations predict
an energy difference for these two cycloaddition products of 6.4
kcal/mol, and the strain present in euplotin A is reflected, in part,
by calculated angles for C(5-6-7) and C(4-3-15) of 126° and
118°, respectively.7 Nonetheless, it was hoped that an early, highly
asynchronous transition state8 as well as the cis-enol ether moiety1
would combine to favor the exo transition state.
The synthesis of the precursor to the acylenal 11, dioxin 16,
was initiated with the preparation of the known Paterno-Buchi
photocycloadduct 129 (Scheme 4). Stereoselective Lewis acid-
catalyzed ring opening of the acetal moiety of 12 followed by
acylation of the resulting hydroxyl gave the dihyrofuran 13.
Samarium diiodide-mediated reductive removal of the acetoxy
functionality followed by reduction (LAH) of the remaining ester
moiety gave an alcohol which was converted to iodide 14. The
aza-enolate derivative of imine 4 was alkylated with iodide 14 to
afford the dioxinone 15 which was converted to the pivotal
acyldioxin 16 using the previously established protocol. We were
pleased to discover that the acyl dioxin 16 underwent a smooth
retrocycloaddition, cycloaddition10 to deliver a 4.5:1 mixture of
5-acyldihyropyrans, the major isomer of which possesses the
desired stereochemical relationship as indicated by the diagnostic
1H NMR proton coupling constants and NOE experiments. This
assignment was confirmed upon single-crystal X-ray analysis of
the major isomer. Indeed, the angles for C(5-6-7) and C(4-
3-15), 126.4° and 117.6°, respectively, were similar to those
calculated (vide supra).
It was with some trepidation that we examined the remaining
two functional group transformations upon the potentially labile,
strained acetal 17, namely, introduction of the isopentyl ketone
and acetal functionalities. While the amide 17 could be directly
converted to the isopentyl ketone upon treatment with isopentyl-
lithium, the yield (39%) was significantly reduced due to a
competing conjugate addition reaction. Consequently, the amide
was hydrolyzed to the corresponding carboxylic acid which could
be cleanly converted to the desired ketone. Fortunately, the
thioacetal moiety11 of this product could be selectively activated
for exchange with an acetoxy substituent by treatment with
mercuric acetate to afford racemic euplotin A (1) as well as the
â-anomer (5.5:1). The spectral properties of the synthetic material
were identical to those found for authentic euplotin A.
In summary, we have developed a mild and versatile protocol
for the preparation of (Z)-2-acyl-2-enals via retrocycloaddition
reactions of the easily available 5-acyl-4-alkyl-4H-1,3-dioxins.
This methodology facilitated a concise total synthesis of the
strained cytotoxin euplotin A. The exploitation of this strategy
in the total syntheses of other natural products which also embody
5-acyl-3,4-dihydro-2H-pyran substructures or close relatives
thereof is now underway.
(6) Thermolysis of dioxin 6b in the presence of an equivalent amount of
an analogous dioxin possessing a 1,1-dideuteriohexyl substituent instead of
the 2-(triethylsilyloxy)ethyl substituent gave enal 7b without incorporation
of deuterium and an enal derived from the deuterated dioxin which was fully
deuterated in the allylic position, thereby ruling out an enolization pathway
for the Z/E isomerization.
(7) Indeed, it has been suggested that this strain accounts for the biological
properties of this natural product by facilitating the ring opening to aldehyde
containing mono- or bicyclic compounds (see ref 2).
(8) This prediction was tenuous, at best, based upon calculations using
PCMODEL (v 6.0). We employed bond order parameters of 0.1 and 0.9 for
the forming C-O and C-C bonds, respectively, and calculated an energy
difference of 0.9 kcal/mol favoring the exo transition state. However, if the
bond orders used by Tietze (based upon preliminary AM1 calculations: 0.1
C-O, 0.7 C-C; Tietze, L. F.; Geissler, H.; Fennen, J.; Brumby, T.; Brand,
S.; Schulz, G. J. Org. Chem. 1994, 59, 182) for a more synchronous
heterocycloaddition were used, then the exo transition state preference was
reduced to 0.5 kcal/mol. Even more disconcerting was the calculation of the
exo transition structure for the stereoisomeric E-2-acyl-2-enal heterodiene (2.8
kcal/mol lower than exo-Z, C-O 0.1, C-C 0.9). This alternative pathway to
the undesired all-cis cycloadduct was an especially serious concern if the rate
of Z/E isomerization is faster than that of the cycloaddition.
Acknowledgment. We thank Professor Francesco Pietra of the
Univerita` Di Trento, Povo-Trento, Italy for kindly furnishing authentic
spectra of euplotin A. We thank Dr. Douglas R. Powell of the University
of Wisconsin for the X-ray crystallographic analysis of compound 17.
We appreciate the financial support provided by the National Institutes
of Health (GM28553).
Supporting Information Available: Spectroscopic data and experi-
mental details for the preparation of all new compounds (PDF). This
material is available free of charge via the Internet at http://pubs.acs.org.
JA011470B
(9) Zamojski, A.; Kozluk, T. J. Org. Chem. 1977, 42, 1089. Initial attempts
in our laboratory to identify a chiral auxiliary for the photocycloaddition of
furan with the derivatized chiral glyoxylate have been unrewarding (8-
phenylmenthol; 57:43), see also: Inoue, Y. Chem. ReV. 1992, 92, 741.
(10) The intermediate acylenal 11 was not detected during the course of
1
the reaction using either TLC or H NMR spectroscopic analysis.
(11) However, the analogous methyl ether gave rise to ring-opened products
in several attempts to effect a similar exchange.