Synthesis of the C1-C21 (C1'-C21') fragment of the dimeric polyketide natural product SCH 351448.

Article abstract of DOI:10.1021/ol016938u

[structure: see text] A convergent, stereoselective assembly of the C1-C21 (C1'-C21') fragment of SCH 351448, a 28-membered bis-lactone natural product, has been developed. A highly efficient approach to this fragment assembles 75% of the carbon skeleton

Full text of DOI:10.1021/ol016938u

ORGANIC
LETTERS
2002
Vol. 4, No. 4
481-484
Synthesis of the C1−C21 (C1′-C21′)
Fragment of the Dimeric Polyketide
Natural Product SCH 351448
Ashoke Bhattacharjee, Omid Soltani, and Jef K. De Brabander*
Department of Biochemistry, The UniVersity of Texas Southwestern
Medical Center at Dallas, Dallas, Texas 75390-9038
jdebra@biochem.swmed.edu
Received October 19, 2001
ABSTRACT
A convergent, stereoselective assembly of the C1−C21 (C1′−C21′) fragment of SCH 351448, a 28-membered bis-lactone natural product, has
been developed. A highly efficient approach to this fragment assembles 75% of the carbon skeleton and all the stereochemical elements
present in the natural product. In addition, an interesting boron ligand effect on the diastereoselectivity of a key aldol reaction with methyl
ketone-derived enolborinates is reported.
Recently, a team working at the Schering-Plough Research
Institute disclosed the structure of an intriguing microbial
metabolite from Micromonospora sp., named SCH 351448
(1).1 The isolation of 1 was guided by its unique ability to
specifically activate transcription of a reporter construct under
control of the low-density lipoprotein (LDL) receptor
promoter. Clearing serum cholesterol levels through in-
creased uptake of LDL by the LDL receptor is important
for the prevention and treatment of coronary heart disease.2
Besides their potential as therapeutic leads, small molecules
that uncouple LDL receptor transcription from regulatory
feedback mechanisms are potentially valuable tools to study
the proteolytic processing of the sterol regulatory element-
binding proteins (SREBPs), which are the membrane-bound
transcription factors that control activation of the LDL
receptor gene.3
The structure and relative stereochemistry of SCH 351448
(1) was unambiguously established via single-crystal X-ray
analysis.4 SCH 351448 is a unique manifestation of the rich
polyketide-producing biosynthetic machinery expressed in
microorganisms. Composed of two identical diacids, this 28-
membered C2-symmetric macrocyclic metabolite integrates
two salicylic acid moieties and four tetrahydropyranyl rings
into a bis-lactone template featuring 14 stereocenters and
two quaternary centers. Our synthetic approach is shown in
Figure 1 and will exploit the dimeric nature of 1. Dimeri-
zation/cyclization is envisioned to proceed through a se-
quential esterification/lactonization, a tandem intermolecular/
intramolecular olefin metathesis, or a sequential esterification/
ring-closing olefin metathesis. Each of the required monomers
ultimately will derive from a common precursor, C1-C21
(C1′-C21′) fragment 2 in turn accessible from two tetrahy-
(2) Brown, M. S.; Goldstein, J. L. Science 1986, 232, 34-47.
(3) Brown, M. S.; Goldstein, J. L. Cell 1997, 89, 331-340.
(4) The absolute stereochemistry of SCH 351448 has not been assigned.
(1) Hegde, V. R.; Puar, M. S.; Dai, P.; Patel, M.; Gullo, V. P.; Das, P.
R.; Bond, R. W.; McPhail, A. T. Tetrahedron Lett. 2000, 41, 1351-1354.
10.1021/ol016938u CCC: $22.00 © 2002 American Chemical Society
Published on Web 01/19/2002

Scheme 1a
a Reagents and conditions: (a) (+)-Ipc2BOMe, allylMgBr, -78
°C, Et2O (80%, 94% ee); (b) CH2dCHCOCl, i-Pr2NEt, DMAP,
CH2Cl2 (83%); (c) 5 mol % 8, CH2Cl2, 4 h, rt (76%); (d)
NiCl2‚6H2O, NaBH4, THF, 0 °C, 2 h (80% 10 + 7% 11); (e)
DIBAL-H, CH2Cl2, -78 °C, 1 h (96%); (f) 1-triphenylphosphor-
anylidene-2-propanone, MeCN, 90 °C, 24 h (70%); (g) t-BuOK,
THF, -78 °C, 20 min (86%, 13:3 ) 9:1); (h) t-BuOK, THF, 0 °C,
5-10 min (90%, only 3).
Figure 1. Synthetic strategy for the synthesis of SCH 351448.
dropyranyl fragments 3 and 4 via an aldol reaction. Herein,
we report a highly efficient synthesis of fragment 2, which
constitutes 75% of the carbon skeleton and all the stereo-
chemical elements present in the natural product.
conditions (THF, 0 °C) provided the thermodynamic tetra-
hydropyranyl ring 3.10 Equilibration did not occur at -78
°C (otherwise similar conditions), and a separable (flash
chromatography) 9:1 mixture of trans- and cis-tetrahydro-
pyrans 13 and 3 was obtained.11 When subjected to t-BuOK
in THF at 0 °C, purified trans-isomer 13 was converted to
cis-isomer 3 within 10 min (91%).12
The synthesis of the C11-C21 (C11′-C21′) fragment 4
commenced with the addition of known methyl 7-oxo-2-
heptenoate 1413 to a -78 °C solution of B-allylbis-(4-
isocaranyl)borane5 to provide homoallyl alcohol 15 (92% ee
via Mosher ester analysis, Scheme 2).7 Unlike the smooth
ring-closure of compound 12, treatment of 15 with t-BuOK
at 0 °C resulted in a complex mixture of intractable products.
Lowering the temperature did provide for a cleaner trans-
A short, seven-step synthesis of fragment 3 is delineated
in Scheme 1. Asymmetric allylation5 of known aldehyde 56
provided homoallyl alcohol 6 (94% ee via Mosher ester
analysis),7 which was further transformed into the acryloate
derivative 7 (CH2dCHCOCl, i-Pr2EtN, CH2Cl2). Subsequent
intramolecular olefin metathesis was best performed with
Grubbs’ imidazolidinyl ruthenium carbene 8 (5 mol %).8 A
variety of conditions were explored for converting R,â-
unsaturated lactone 9 into the saturated lactol 11. A reducing
mixture composed of CoCl2 and NaBH4 in THF provided
the desired lactol 11 in one operation but was not reproduc-
ible on a larger scale. Instead, a two-step procedure was
followed whereby the conjugated double bond was first
saturated with NiCl2/NaBH4 in THF9 (80%, 7% of lactol 11
was also present) followed by reduction of lactone 10 to
lactol 11 with DIBAL-H (96%). Homologation of 11 with
1-triphenylphosphoranyliden-2-propanone proceeded un-
eventful and set the stage for a base-catalyzed cyclization
of enone 12. As expected, treatment of δ-hydroxy enone 12
with a catalytic amount of t-BuOK under equilibrating
(10) Evans, D. A.; Carreira, E. M. Tetrahedron Lett. 1990, 31, 4703-
4706.
(11) The stereochemistry was assigned on the basis of 1H NMR coupling
3
3
constants and NOE experiments (3, JH3-H4e ) 1.6 Hz, JH3-H4a ) 11.2
Hz, 3JH7-H6e ) 1.6 Hz, 3JH7-H6a ) 10.8 Hz; 13, 3JH3-H4e ) 1.6 Hz, 3JH3-H4a
) 11.2 Hz, JH7-H6e ) 0 Hz, JH7-H6a ) 5.2 Hz).
3
3
(5) Racherla, U. S.; Brown, H. C. J. Org. Chem. 1991, 56, 401-404.
(6) Kim, H.-O.; Carroll, B.; Lee, M. S. Synth. Commun. 1997, 27, 2505-
2515.
(7) Sullivan, G. R.; Dale, J. A.; Mosher, H. S. J. Org. Chem. 1973, 38,
2143-2147.
(8) Chatterjee, A. K.; Morgan, J. P.; Scholl, M.; Grubbs, R. H. J. Am.
Chem. Soc. 2000, 122, 3783-3784.
(9) Satoh, T.; Nanba, K.; Suzuki, S. Chem. Pharm. Bull. 1971, 19, 817-
820.
(12) Evans, D. A.; Ripin, D. H. B.; Halstead, D. P.; Campos, K. R. J.
Am. Chem. Soc. 1999, 121, 6816-6826.
(13) Denmark, S. E.; Middleton, D. S. J. Org. Chem. 1998, 63, 1604-
1618.
482
Org. Lett., Vol. 4, No. 4, 2002

Scheme 2a
Table 1. Aldol Reactions of Enolborinates Derived from 3
entry
RCHO
enolization
yielda
products (ratio)b
1
2
3
4
5
6
7
8
9
4
4
4
Chx2BClc
Bu2BOTfd
Et2BOTfc
Chx2BClc
Bu2BOTfd
Et2BOTfd
Chx2BClc
Bu2BOTfd
Et2BOTfd
68%
45%
86%
87%
91%
91%
66%
90%
90%
23a :24a (1.5:1)
23a :24a (3:1)
23a :24a (5.5:1)
23c:24c (1.8:1)
23c:24c (3:1)
a Reagents and conditions: (a) B-allyl bis(4-isocaranyl)borane,
Et2O, -78 °C, (70%, 92% ee); (b) LiAlH4, Et2O, 0 °C (99%); (c)
PCC, CH2Cl2, 3 h, RT (70%); (d) (2S)-N-propionyl bornanesultam,
TiCl4, i-Pr2NEt, -78 °C; then add 19, -78 °C (71%); (e)
PhOC(S)Cl, py, CH2Cl2, 14 h, RT (81%); (f) Bu3SnH, 10% AIBN,
PhH, 7 h, reflux (87%); (g) DIBAL-H, CH2Cl2, -78 °C, 1 h; then
MeOH, -78 °C, 10 min (70%); or LiAlH4, Et2O (97%); then Dess-
Martin periodinane, CH2Cl2 (87%).
EtCHO
EtCHO
EtCHO
epi-4
epi-4
epi-4
23c:24c (5:1)
23b:24b (1.4:1)
23b:24b (2.4:1)
23b:24b (3.4:1)
a Isolated yield of aldol adducts. b Ratios determined by 1H NMR analysis
(entries 1-3) or HPLC (entries 4-9) of the unpurified product mixture.
c Enolization with Et3N. d Enolization with i-Pr2NEt.
formation, and the cis-isomer 17 was the only product
isolated (80%) after cyclization of 15 at -35 °C (0.6 equiv
of t-BuOK, THF).14,15 A reduction/oxidation sequence gave
aldehyde 19 which was treated with the Z(O)-titanium enolate
derived from (2S)-N-propionyl bornanesultam (-78 °C) to
maximum stereocontrol via 1,5-anti asymmetric induction.19
Disappointingly, the use of dicyclohexyl or dibutyl enol-
borinates derived from 320 in combination with aldehyde 4
provided aldol products with low levels of diastereoselection
(23a:24a ) 1.5:1 and 3:1 respectively, entries 1-2).21 We
noticed, however, that the selectivity was inherently better
with the smaller dibutyl enolborinate, encouraging us to
engage the diethyl enolborinate derived from 3.22 Indeed,
reaction of aldehyde 4 with this enolborinate proceeded with
an increased level of diastereoselectivity (23a:24a ) 5.5:1,
entry 3). These results raise two important questions: (1)
what is the reason for the reduced selectivity as compared
to the examples reported by the Evans and Patterson groups
and (2) what is the origin of the boron ligand effect on
reaction stereoselectivity? To answer these questions, we had
to dissect the influence of aldehyde structure on the stereo-
selectivity of reactions with enolborinates derived from
methyl ketone 3. As indicated by the similar selectivities
obtained for the aldol reactions of 3 with propionaldehyde
1
deliver the single (by H NMR analysis) â-hydroxy amide
diastereomer 20.16 A modified Barton-McCombie deoxy-
genation17 via the corresponding phenyl thionocarbonate 21
was followed by a one-step reductive transformation of
N-acyl bornanesultam derivative 22 to provide target alde-
hyde 4.18
With efficient synthetic access to compounds 3 and 4, we
next examined their joining via aldol bond construction
(Table 1). Analysis of this double stereodifferentiating
process dictated the use of enolborinates to provide for
(14) The trans-isomer 16 predominated when the reaction was conducted
at -78 °C (see also: Schneider, C.; Schuffenhauer, A. Eur. J. Org. Chem.
2000, 73-82).
(15) The stereochemistry was assigned on the basis of 1H NMR coupling
3
3
constants and NOE experiments (17, JH15-H16e ) 1.6 Hz, JH15-H16a
)
3
3
3
10.8 Hz, JH19-H18e ) 1.6 Hz, JH19-H18a ) 10.8 Hz; 16, JH15-H16e ) 3.6
Hz, JH15-H16a ) 6.4 Hz, JH19-H18e ) 3.6 Hz, JH19-H18a ) 8.8 Hz).
3
3
3
(19) (a) Evans, D. A.; Coleman, P. J.; Coˆte´, B. J. Org. Chem. 1997, 62,
788-789. (b) Paterson, I.; Gibson, K. R.; Oballa, R. M. Tetrahedron Lett.
1996, 37, 8585-8588.
(20) For examples of excellent 1,5-anti stereoinduction by a â-tetrahy-
dropyranyl moiety, see: (a) Evans, D. A.; Fitch, D. M.; Smith, T. E.; Cee,
V. J. J. Am. Chem. Soc. 2000, 122, 10033-10046. (b) Kozmin, S. A. Org.
Lett. 2001, 3, 755-758.
(16) Evans, D. A.; Rieger, D. L.; Bilodeau, M. T.; Urp´ı, F. J. Am. Chem.
Soc. 1991, 113, 1047-1049.
(17) Robins, M. J.; Wilson, J. S.; Hansske, F. J. Am. Chem. Soc. 1983,
105, 4059-4065.
(21) The inferred19,20 1,5-anti stereochemical relationship in 23a was
confirmed by assigning the absolute configuration at C11 via the Mosher
ester method (see Supporting Information): Dale, J. A.; Mosher, H. S. J.
Am. Chem. Soc. 1973, 95, 512-519.
(18) Oppolzer, W.; Darcel, C.; Rochet, P.; Rosset, S.; De Brabander, J.
HelV. Chim. Acta 1997, 80, 1319-1337. A two-step reduction/oxidation
sequence is more convenient for a large scale preparation of aldehyde 4
(see Supporting Information).
(22) Changes in the size of boron substituents have been shown to
influence the diastereoselectivity of methyl ketone aldol reactions with
R-Me-â-alkoxy aldehydes: Gustin, D. J.; VanNieuwenhze, M. S.; Roush,
W. R. Tetrahedron Lett. 1995, 36, 3443-3446.
Org. Lett., Vol. 4, No. 4, 2002
483

and 4 (Table 1, entries 4-6 vs entries 1-3), there are no
substantial additional nonbonded interactions imposed by the
anti-Felkin arrangement of the R-Me group of aldehyde 4
in the transition state, demonstrating that the reaction
diastereoselectivity is the result of the intrinsic facial bias
of the enolborinate partner.23 In contrast, nonbonded interac-
tions between the methyl ketone substituent and the R-Me
substituent of aldehyde epi-424 in a Felkin orientation are
most likely responsible for the lower selectivities tabulated
in entries 7-9.23,25
Scheme 3a
Perhaps more importantly, the set of experiments described
in Table 1 features the general dependence of aldol 1,5-
stereoinduction on the size of the enolborinate ligands,
irrespective of aldehyde structure.26 Thus, reactions with the
smaller diethyl enolborinates are always more selective for
the 1,5-anti aldol products 23a-c than the corresponding
reactions with dibutyl and dicyclohexyl enolborinates. With-
out a proper knowledge of the reacting conformation of the
enolborinates and the integrity of a presumed twist-boat
transition state, it is premature to speculate on the origin of
the boron ligand effect in these aldol reactions.27
The final step for the synthesis of the C1-C21 (C1′-
C21′) fragment 2 required an anti-selective reduction of
â-hydroxy ketone 23a (Scheme 3). For practical reasons, the
mixture of aldolates was reduced with Me4NBH(OAc)3 to
give anti diol 2 in 82% isolated yield, together with two
minor diastereomeric diols (12%).28 The 1,3-anti diol ster-
a Reagents and conditions: (a) Me4NBH(OAc)3, HOAc, MeCN,
-10 °C (82%); (b) Me2CO, Me2C(OMe)2, CSA, rt (67%).
eochemistry was confirmed by 13C NMR analysis of the
corresponding acetonide 25 (isopropylidene carbon reso-
nances at 24.6, 25.4 and 100.3 ppm).29
In summary, we have accomplished a highly efficient and
stereoselective synthesis of the C1-C21 (C1′-C21′) frag-
ment of SCH 351448 in 17 total steps (10 steps longest linear
sequence; 14% overall yield) from readily available materials
5 and 14. We also have observed an interesting boron ligand
effect on the stereoselectivity of aldol reactions with methyl
ketone-derived enolborinates. With all the stereochemical
issues solved, we are now in a position to explore the final
steps of the synthesis, which include incorporation of an
ortho-substituted salicylic acid fragment and dimerization to
obtain the natural product.
(23) It has been postulated that twist-boat transition structures are
preferred in aldol reactions with methyl ketone-derived enolborinates, see
refs19, 22, and (a) Li, Y.; Paddon-Row, M. N.; Houk, K. N. J. Org. Chem.
1990, 55, 481-493. (b) Bernardi, A.; Capelli, A. M.; Gennari, C.; Goodman,
J. M.; Paterson, I. J. Org. Chem. 1990, 55, 3576-3581. (c) Bernardi, F.;
Robb, M. A.; Suzzi-Valli, G.; Tagliavini, E.; Trombini, C.; Umani-Ronchi,
A. J. Org. Chem. 1991, 56, 6472-6475.
(24) The C12-epimer of aldehyde 4 (epi-4) was synthesized from
aldehyde 19 by following the protocol outlined in Scheme 2 with the
exception of the use of enantiomeric (2R)-N-propionyl bornanesultam for
the installation of the C12-stereocenter.
Acknowledgment. Financial support provided by the
Robert A. Welch Foundation, the National Institutes of
Health (CA 90349), and junior faculty awards administered
through the Howard Hughes Medical Institute and the
University of Texas Southwestern Medical Center is grate-
fully acknowledged. HRMS analyses were performed at the
NIH regional mass spectrometry facility at the University
of Washington, St. Louis, MO. J. K. De Brabander is a fellow
of the Alfred P. Sloan Foundation.
(25) Roush, W. R.; Bannister, T. D.; Wendt, M. D. Tetrahedron Lett.
1993, 34, 8387-8390.
(26) Roush has formulated a hypothesis to rationalize the steric effect
of metal enolate ligands with the aldehyde component, see ref 22.
(27) A detailed study to rationalize the boron ligand effect and explore
its scope will be reported in due course. At present, we speculate that the
smaller boron ligands have a larger intrinsic (i.e., irrespective of aldehyde
or enolate structure) preference for the twist-boat transition structure. Both
ab initio calculations19a,23a and force field models23b have indicated that a
B-H f B-Me substitution decreases the energy gap between twist-boat and
chair transition states.
Supporting Information Available: Experimental pro-
cedures and characterization data for new compounds. This
material is available free of charge via the Internet at
http://pubs.acs.org.
OL016938U
(28) The separation of 23a from 24a requires preparative HPLC, in
contrast to the corresponding reduced products, which could be conveniently
separated by flash chromatography.
(29) Rychnovsky, S. D.; Rogers, B.; Yang, G. J. Org. Chem. 1993, 58,
3511-3515.
484
Org. Lett., Vol. 4, No. 4, 2002