The total synthesis of the oxopolyene macrolide RK-397

Article abstract of DOI:10.1002/anie.200602601

It works both ways: The convergent total synthesis of the oxopolyene macrolide RK-397 utilizes remote asymmetric induction and a two-directional chain synthesis to prepare the polyol portion of the molecule, as well as a cross-metathesis reaction of a trienal with a terminal alkene to append the polyene to the polyol. (Chemical Equation Presented).

Full text of DOI:10.1002/anie.200602601

Communications
DOI: 10.1002/anie.200602601
Natural Products
The Total Synthesis of the Oxopolyene Macrolide
RK-397**
Mark J. Mitton-Fry, Aaron J. Cullen, and Tarek Sammakia*
Angewandte
Chemie
1066
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 1066 –1070

Angewandte
Chemie
The stereochemically complex natural product RK-397 (1,
Scheme 1) was isolated in 1993 from a strain of soil bacteria
and was shown to possess antifungal, antitumor, and anti-
bacterial activities.[1] RK-397 is an oxopolyene macrolide[2]
whose structure[3] corresponds to that of C14-demethyl
mycoticin A,[4] in which the stereocenters at C19 and C21
have the opposite configuration.[5] Members of this class of
natural products have been popular targets for synthesis,[6]
and Burova and McDonald,[7] as well as Denmark and
Fujimori[8] have recently reported syntheses of RK-397.
Herein, we describe our recently completed route to this
molecule.
approach for the synthesis of 3, and a late-stage cross-
metathesis of an olefin to append the polyene chain.
Scheme 2 depicts our retrosynthetic analysis of methyl
ketone 2. From the outset, we wished to make use of the C30/
Scheme 2. Retrosynthesis of methyl ketone 2.
C31 stereodiad to dictate the stereochemistry at C27 and C25
through remote asymmetric induction. We chose to accom-
plish this by using the method reported by the research groups
of Evans and Paterson, in which a b-alkoxy methyl ketone
undergoes an aldol reaction with relay of stereochemistry
between the b-alkoxy group and the newly formed stereo-
center (1,5 induction).[10] We therefore required an alkoxy
group at C29 to relay asymmetric induction to C25 and then
planned to subject this group to an elimination reaction in
order to install the C28–C29 alkene.
Our synthesis of 2 began with a crossed aldol condensa-
tion between ethyl propionate and isobutyraldehyde
(Scheme 3). This modified literature procedure[11] provided
enoate 4 in good yield, excellent E selectivity (64%, E/Z >
50:1), and was readily conducted on a molar scale. Reduction
Scheme 1. Retrosynthesis of RK-397 (1). PG=protecting group,
TBS=tert-butyldimethylsilyl, PMB=para-methoxybenzyl.
This molecule presents several synthetic challenges,
including the efficient construction of the stereochemically
complex polyol chain, the introduction of the sensitive
polyene moiety, and the formation of the macrocycle. In
analogy to previous oxopolyene macrolide syntheses, we
planned to form the macrocycle through a macrolactoniza-
tion.[9] The polyol was disconnected into two fragments of
roughly equal complexity, namely methyl ketone 2 (C22–C33)
and aldehyde 3 (C10–C21, Scheme 1). We envisioned an
asymmetric catalysis approach with relay of stereochemistry
across the alkene for the synthesis of 2, a two-directional
[12]
of the enoate with LiAlH4 was followed by a Sharpless
asymmetric epoxidation[13] to afford the known epoxy alcohol
5 with good enantiocontrol (96% ee). Rearrangement of 5 to
the b-silyloxyaldehyde 6 proceeded with good diastereose-
[*] Dr. M. J. Mitton-Fry,[+] A. J. Cullen, Prof. Dr. T. Sammakia
Department of Chemistry and Biochemistry
University of Colorado
215 UCB, Boulder, CO 80309-0215 (USA)
Fax: (+1)303-492-0439
E-mail: Sammakia@colorado.edu
[+] Current address:
Pfizer Global Research and Development
Groton, CT 06340 (USA)
Scheme 3. Preparation of hydroxy ketone 10. Reagents and conditions:
a) ethyl propionate, NaH, EtOH (64%, E/Z >50:1); b) LiAlH4, then
(+)-DIPT, Ti(OiPr)4, tBuOOH (96% ee); c) TMSOTf, iPr2NEt
(24:1 d.r.); d) acetone, LDA (40% from 4, >10:1 d.r.); e) MOMCl,
iPr2NEt (73%); f) Bu2BOTf, iPr2NEt, then 9 (85%, anti/syn 4:1).
TMS=trimethylsilyl, Tf=triflate=trifluoromethanesulfonyl, MOM=
methoxymethyl; LDA=lithium diisopropylamide, DIPT=diisopropyl
tartrate.
[**] This work was supported by a grant from the National Institutes of
Health (GM48498). We wish to thank Prof. H. Osada for providing
NMR spectra of RK-397. M.J.M. gratefully acknowledges the
National Science Foundation and also Pharmacia and Upjohn for
fellowships.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2007, 46, 1066 –1070
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1067

Communications
lectivity (24:1) on treatment with TMSOTf and N,N-diiso-
Our synthesis of 3 began with the reduction of glutarate
ester 12 to the corresponding glutaraldehyde derivative 13[19]
(Scheme 5). Bisallylation of this species using the (+)-3-
carene-derived allylboration reagent 14 reported by Brown
propylethylamine by using a procedure described by Jung and
DꢀAmico for related systems.[14] Addition of the lithium
enolate of acetone to 6 at À1008C afforded the aldol adduct 7
with good selectivity and yield over the four-step sequence
(40% from 4, > 10:1 d.r.). Protection of the C29 hydroxy
group as a MOM ether provided ketone 8 in 73% yield. This
compound was then subjected to the methyl ketone aldol
reaction with 1,5-anti induction to set the stereochemistry at
C25. By using the protocol described by Evans et al.
(Bu2BOTf, iPr2NEt)[10b] with aldehyde 9, compound 10 was
produced in high yield with useful levels of diastereoselectiv-
ity (85%, anti/syn 4:1).[15]
With compound 10 in hand, we now required a chemo-
selective elimination of the OMOM group at C29 in
preference to the OH group at C25. We initially attempted
this transformation using two equivalents of LDA at À788C;
however, these conditions provided mixtures of the desired
elimination product 11 and recovered 10. After much
experimentation, we were able to accomplish this trans-
formation using K2CO3 (0.2 equiv) in 95% ethanol at 08C in
good yield and with excellent E selectivity (81%, E/Z > 50:1,
Scheme 4). Conversion of 11 into 2 was then accomplished by
Scheme 5. Preparation of aldehyde 3. Reagents and conditions:
a) DIBAL-H, then 1258C at 2 mmHg; b) 14, then H2O2/NaOH (53%
over two steps, >98% ee, 10:1 d.r.); c) 1. amberlyst-15, MeOH; 2. ace-
tone, CuSO4, PPTS (cat.); d) ozone, then NaBH4; e) p-anisaldehyde
dimethyl acetal, PPTS (cat.) (73% over 4 steps); f) oxalyl chloride,
DMSO, Et3N (94%); g) 14, then H2O2/NaOH (87%, >30:1 d.r.);
h) TBSCl, imidazole (98%); i) DIBAL-H (81%); j) oxalyl chloride,
DMSO, Et3N (85%). PMP=p-methoxyphenyl; PPTS=pyridinium p-
toluene sulfonate; DIBAL-H=diisobutylaluminum hydride.
and co-workers[20] provided diol 15 in good yield and high
selectivity (53% from 12, > 98% ee, 10:1 d.r.). Differentia-
tion of the diastereotopic termini of this substrate began with
removal of the TBS group at C17 by using amberlyst-15.
Subsequent formation of the thermodynamically more
favored syn acetonide between the C15 and C17 hydroxy
groups (as compared to the anti acetonide which would be
formed between the C17 and C19 hydroxy groups) with
acetone and a catalytic amount of PPTS provided compound
16 (syn acetonide/anti acetonide > 20:1). While this sequence
provides internal differentiation, it does not discriminate
between the two terminal alkenes. This was accomplished
through ozonolytic cleavage of the alkene with a reductive
(NaBH4) workup to provide the corresponding triol. Subse-
quent treatment of this triol with catalytic PPTS and p-
anisaldehyde dimethyl acetal provided 17 in 73% yield over
the four-step sequence. The use of mild conditions and a
reactive acetal precursor avoids acetonide migration in this
reaction. Compound 17 was then subjected to Swern oxida-
tion and a second asymmetric allylation, again with the (+)-3-
carene-derived reagent, to provide 18 (> 30:1 d.r.). The
hydroxy group of 18 was next protected as the TBS ether to
provide 19 (80% from 17). Reduction of the p-methoxyben-
zylidene acetal group of 19 using DIBAL-H (08C to RT)
provided the hydroxy group at C21, which was oxidized to
provide 3 in 11 steps and 21% overall yield for the sequence.
Having assembled methyl ketone 2 and aldehyde 3, these
fragments were coupled by using the Paterson variant of the
1,5-anti aldol procedure to provide 20 in good yield and
selectivity (83%, > 10:1 d.r., Scheme 6).[21] The hydroxy
Scheme 4. Preparation of methyl ketone 2. Reagents and conditions:
a) K2CO3 (0.2 equiv), EtOH/H2O 95:5 (81%, E/Z >50:1);
b) 1. Et2BOMe, NaBH4; 2. FeCl3, acetone, Desispheres (68%, 2 steps).
reduction by using a procedure reported by Prasad and co-
workers[16] followed by treatment with FeCl3[17] in dry acetone
containing the alumina-based desiccant desispheres. This
transformation removes the protecting groups at C23 and
C31 and installs an acetonide between C25 and C27. This
sequence provides compound 2 with remote asymmetric
induction across the alkene in ten steps and proceeds in about
10% overall yield from commercially available starting
materials.
We designed the synthesis of compound 3 to take
advantage of the symmetry within the molecule and
employed a two-directional allylation approach.[18] We chose
C17 as the central carbon atom because the hydroxy group at
this site is syn to the hydroxy group at C15, and anti to that at
C19. As such, this stereochemical triad could be synthesized
from a 3-alkoxy glutaraldehyde derivative through a double
asymmetric allylation with reagent control and subsequent
terminus differentiation.
1068 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 1066 –1070

Angewandte
Chemie
Scheme 6. Preparation of protected polyol chain 21. Reagents and conditions: a) Cy2BCl, Et3N, then 3 (83%, anti/syn >10:1); b) 1. Et2BOMe,
NaBH4; 2. 2-methoxypropene, PPTS (cat.) (84% over 2 steps, syn/anti >10:1). Cy=cyclohexyl.
Scheme 7. Synthesis of RK-397. Reagents and conditions: a) CH2Cl2 reflux (72%, E/Z 4:1); b) 24, LDA, then 23 (82%); c) 1. LiOH, THF/H2O/
MeOH (quant.); 2. 2,4,6-trichlorobenzoyl chloride, Et3N; 3. DMAP, toluene; d) HCl (12m, 100 equiv)/MeOH (70% from 25). DMAP=N,N-
dimethylaminopyridine.
group at C31 did not prove problematic in the reaction using
2.4 and 3 equivalents of Cy2BCl and NEt3, respectively.
Prasad reduction of 20 and protection of the resulting diol
as the acetonide provided the fully functionalized and
appropriately protected C10–C33 polyol domain of RK-397
(21, 84% from 20, > 10:1 d.r.). This synthesis proceeds with a
longest linear sequence of 14 steps and has allowed us to
prepare multigram quantities of the protected polyol chain.
The terminal alkene of 21 can serve as a handle for the
installation of the polyene through an olefin cross-metathesis
reaction, and after much experimentation we found that
treatment of 21 and 2,4,6-hexatrienal (22) with Grubbs first
generation catalyst[22] in refluxing dichloromethane provided
compound 23 in 72% yield as a 4:1 mixture of isomers at the
C10–C11 alkene (Scheme 7). Conversion of this material into
RK-397 was accomplished by a Horner–Wadsworth–Emmons
reaction with 24[23] (82% yield), followed by saponification
and Yamaguchi macrolactonization,[24] and subsequent global
deprotection with concentrated HCl in methanol (70% over
four steps).[8,25] While milder acidic conditions such as Dowex
50W-X8 resin were capable of removing all the protecting
groups on model systems, the PMB ether at C19 resisted
cleavage under these conditions on the real system.
reaction as another convergent step in the synthesis. The final
steps, including the macrocyclization and deprotection, pro-
ceeded in high yield and provided RK-397 with a longest
linear sequence of 20 steps.
Received: June 29, 2006
Published online: October 18, 2006
Keywords: asymmetric synthesis · macrocycles · macrolides ·
.
metathesis · total synthesis
[1] K. Kobinata, H. Koshino, T. Kudo, K. Isono, H. Osada, J.
Antibiot. 1993, 46, 1616 – 1618.
[2] For a review, see: S. D. Rychnovsky, Chem. Rev. 1995, 95, 2021 –
2040.
[3] H. Koshino, K. Kobinata, K. Isono, H. Osada, J. Antibiot. 1993,
46, 1619 – 1621.
[4] S. L. Schreiber, M. T. Goulet, J. Am. Chem. Soc. 1987, 109, 8120 –
8122.
[5] T. Suenaga, H. Nakamura, H. Koshino, K. Kobinata, H. Osada,
T. Nakata, Tennen Yuki Kagobutsu Toronkai Koen Yoshishu
1997, 39, 607 – 612.
[6] For partial syntheses of RK-397, see: a) S. A. Burova, F. E.
McDonald, J. Am. Chem. Soc. 2002, 124, 8188 – 8189; b) C.
Schneider, F. Tolksdorf, M. Rehfeuter, Synlett 2002, 2098 – 2100;
c) S. Gerber-Lemaire, A. T. Carmona, K. T. Meilert, P. Vogel,
Eur. J. Org. Chem. 2006, 891 – 900.
In summary, this route takes advantage of a streamlined
strategy for the synthesis of the polyol segment which features
a two-directional approach to the right half and a remote
asymmetric induction approach to the left. The polyene was
appended to the polyol by using a novel cross-metathesis
[7] S. A. Burova, F. E. McDonald, J. Am. Chem. Soc. 2004, 126,
2495 – 2500.
Angew. Chem. Int. Ed. 2007, 46, 1066 –1070
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 1069

Communications
[8] S. E. Denmark, S. Fujimori, J. Am. Chem. Soc. 2005, 127, 8971 –
C22–C33 ketone intermediate 2. In practice, treatment of the
mixture of partially deprotected compounds with FeCl3 in dry
acetone containing desispheres resulted in complete desilyla-
tion, benzodioxepine removal, and acetonide installation to
afford 2 in good yield (68%, 2 steps).
8973.
[9] For related macrolactonization approaches to oxopolyene
macrolides, see: a) Y. Mori, M. Asai, J-i. Kawade, A. Okumura,
H. Furukawa, Tetrahedron Lett. 1994, 35, 6503 – 6506; b) D. A.
Evans, B. T. Connell, J. Am. Chem. Soc. 2003, 125, 10899 –
10905; c) C. S. Poss, S. D. Rychnovsky, S. L. Schreiber, J. Am.
Chem. Soc. 1993, 115, 3360 – 3361; and also see Ref. [8].
[10] a) I. Paterson, K. R. Gibson, R. M. Oballa, Tetrahedron Lett.
1996, 37, 8585 – 8588; b) D. A. Evans, P. J. Coleman, B. CôtØ, J.
Org. Chem. 1997, 62, 788 – 789; for syntheses utilizing these
methods, see: c) Ref. [9b]; d) Ref. [8]; e) Ref. [6b]; f) I. Pater-
son, L. A. Collett, Tetrahedron Lett. 2001, 42, 1187 – 1191.
[11] C. S. Marvel, W. B. King, Organic Syntheses Collective, Vol. 1
(Ed.: H. Gilman), Wiley, New York, 1941, p. 252.
[17] a) S. E. Sen, S. L. Roach, J. K. Boggs, G. J. Ewing, J. Magrath, J.
Org. Chem. 1997, 62, 6684 – 6686; b) P. P. Singh, M. M. Gharia, F.
Dasgupta, H. C. Srivastava, Tetrahedron Lett. 1977, 439 – 440.
[18] a) C. S. Poss, S. L. Schreiber, Acc. Chem. Res. 1994, 27, 9 – 17;
b) S. R. Magnuson, Tetrahedron 1995, 51, 2167 – 2213; c) R. W.
Hoffmann, Angew. Chem. 2003, 115, 1128 – 1142; Angew. Chem.
Int. Ed. 2003, 42, 1096 – 1109.
[19] Compound 13 is prone to hydration and polymerization. We
have found this polymerization to be reversible and that the
bisaldehyde can be obtained by dehydration (see the Supporting
Information for details). For a similar observation and trans-
formation, reported while this work was in progress, see: J. N.
Shepherd, D. C. Myles, Org. Lett. 2003, 5, 1027 – 1030. For a
related reaction, see: Z. Wang, D. DeschÞnes, J. Am. Chem. Soc.
1992, 114, 1090 – 1091.
[12] V. J. Lee, A. R. Branfman, T. R. Herrin, K. L. Rinehart, J. Am.
Chem. Soc. 1978, 100, 4225 – 4236.
[13] a) T. Katsuki, K. B. Sharpless, J. Am. Chem. Soc. 1980, 102,
5974 – 5976; b) Y. Gao, R. M. Hanson, J. M. Klunder, S. Y. Ko,
H. Masamune, K. B. Sharpless, J. Am. Chem. Soc. 1987, 109,
5765 – 5780; c) R. Göttlich, U. Schopfer, M. Stahl, R. W.
Hoffmann, Liebigs Ann. 1997, 1757 – 1764; d) S. H. Kang, H.-S.
Jun, J.-H. Youn, Synlett 1998, 1045 – 1046.
[20] P. K. Jadhav, K. S. Bhat, T. Perumal, H. C. Brown, J. Org. Chem.
1986, 51, 432 – 439. For other reagent-controlled asymmetric
allylations and crotylations, see: C. H. Burgos, E. Canales, K.
Matos, J. A. Soderquist, J. Am. Chem. Soc. 2005, 127, 8044 –
8049, and references therein.
[14] M. E. Jung, D. C. DꢀAmico, J. Am. Chem. Soc. 1993, 115, 12208 –
12209.
[15] Despite extensive experimentation, which included variations of
protecting groups at C29 and C31, we saw no significant increase
in selectivity. Moreover, the use of the conditions reported by
Paterson et al. (Cy2BCl, Et3N, see Ref. [10a]) on similar sub-
strates provided lower selectivity (< 2:1).
[16] K. M. Chen, G. E. Hardtmann, K. Prasad, O. Repic, M. J.
Shapiro, Tetrahedron Lett. 1987, 28, 155 – 158. Reduction of
hydroxyenone 11 using conditions reported in the reference
above (Et2BOMe, NaBH4) gave the 1,3-syn diol as a single
diastereomer on a small scale. On larger scales, this trans-
formation was accompanied by partial cleavage of the C23 and
C31 protecting groups during workup. Fortunately, both of these
groups must be removed in the next step to give our desired
[21] See Ref. [10a]. In this case, the Paterson conditions proved
superior to the Evans conditions because of a greater preference
for enolization at C22.
[22] a) Handbook of Metathesis (Ed.: R. H. Grubbs) Wiley-VCH,
Weinheim, 2003; b) T. W. Funk, J. Efskind, R. H. Grubbs, Org.
Lett. 2005, 7, 187.
[23] M. Kinoshita, H. Takami, M. Taniguchi, T. Tamai, Bull. Chem.
Soc. Jpn. 1987, 60, 2151 – 2162.
[24] J. H. K. Inanaga, H. Saeki, T. Katsuki, M. Yamaguchi Bull.
Chem. Soc. Jpn. 1979, 52, 1989 – 1993.
[25] This material was purified by reverse-phase HPLC to provide a
sample for characterization (see the Supporting Information for
details).
1070 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 1066 –1070