A direct route to fluostatin C by a fascinating Diels-Alder reaction

Article abstract of DOI:10.1021/ja7113757

The Diels-Alder reactions between vinylindenes (5 or 6) as the dienes with quinoneketals (7 or 8) or with methacrolein as the dienophiles were investigated. The remarkable regioselectivities of these Diels-Alder adducts suggested that the regiopreferences of these dienes and dienophiles in these cases are not a fixed property of each component of the cycloaddition but are mutually contigent. This paper shows how the Diels-Alder reaction between 6 and 8 was applied to the inaugural total syntheses of fluostatin C and E. Copyright

Full text of DOI:10.1021/ja7113757

Published on Web 02/07/2008
A Direct Route to Fluostatin C by a Fascinating Diels-Alder Reaction
Maolin Yu† and Samuel J. Danishefsky*,†,‡
Department of Chemistry, Columbia UniVersity, HaVemeyer Hall, 3000 Broadway, New York, New York 10027, and
Laboratory for Bioorganic Chemistry, Sloan-Kettering Institute for Cancer Research, 1275 York AVenue,
New York, New York 10065
Received December 23, 2007; E-mail: s-danishefsky@ski.mskcc.org
Scheme 1. Synthetic Strategy toward Fluostatin C
Fluostatin C was isolated from the fermentation broth of
Streptomyces strain Acta 1383.1 It is a member of a larger family
of fluostatins which exhibit varying degrees of antibiotic and
antitumor activity.2 The fluostatins also bear a structural similarity
to the kinamycins, which display related biological profiles.3,4 At
the time we undertook this project, no particular fluostatin had
emerged as the clear-cut leading candidate for a total synthesis/
diverted total synthesis venture. Thus, we defined as our goal the
accomplishment of a concise total synthesis which could be directed
to a selection of fluostatins. For purposes of providing focus, our
initial foray was directed to fluostatin C (see Scheme 1 for structure
1 and numbering system).
Not the least of our considerations in embarking on this program
was the incentive it provided to explore an intriguing possibility,
that is, a Diels-Alder reaction between a substituted vinylindene
(cf. 2) as the diene component with a quinoneketal (cf. 3) as the
dienophile. While each of the component subtypes which we were
envisioning had been shown to function in nonrelated Diels-Alder
contexts, the two component types had not been combined in a
single cycloaddition step. In principle, reaction of 2 and 3 in the
endo mode could, as suggested in Scheme 1, give rise to either 4a
or 4b. The endo outcome pictured in Scheme 1 need not be central
to reaching 1 since the latter lacks sp3-based stereopresentation in
the ring C area.
Scheme 2. Investigations of Diels-Alder Reactions
At this stage of our argument, we do not specify the orientational
preference for the critical Diels-Alder step. At the chemoselectivity
level, we naturally assumed that the active dienophilic double bond
within 3 would be the disubstituted site (see asterisks), correspond-
ing eventually to carbons 4a and 11b of the defining target 1. It
was further supposed that either the keto or the ketal functions of
a Diels-Alder cycloadduct could be managed such as to emerge
as the alcohol function (C1 of structure 1) or the keto group (C4 of
structure 1). Accordingly, our “need to know” the regiochemistry
of the alignment would arise only for the purpose of selecting
among the two possible placement sites of the methyl group in 3
(i.e., next to the ketal or next to the ketone).
Viability of the vinylindene-type dienes (cf. 2) in the projected
cycloaddition was another concern. It had been reported that even
the parent compound (5, Scheme 2) has a pronounced tendency
toward polymerization.5 Conceivably, systems of the type 2 bearing
an electron-donating group in the benzo ring (cf. 6) could be even
more vulnerable.6 The general record of quinoneketals as Diels-
Alder dienophiles suggested that they are not particularly reactive
under strictly thermal conditions.7 Fortunately for our plans, Corey
and associates had described some studies of the Diels-Alder
reaction of quinoneketal 7 under mediation by the Mikami Lewis
acid catalytic system.8 The dienes used by Corey in his demonstra-
tions were well-established in their Diels-Alder “regio” charac-
teristics. Thus, we started by asking whether the Mikami protocols
would be applicable to potentially unstable vinylindene (type 2)
dienes, with either quinoneketals 7 or 8.9 Of course, while we
assumed that either of the regiochemical types of alignments shown
below could be managed, such a proposed route to fluostatin C
would be badly compromised if a mixture of cycloadducts, resulting
from competitive regio-alignments, arose.
Below, we describe how the total synthesis of fluostatin C (1)
was realized. As it turned out, a highly regioselective Diels-Alder
reaction of a type 2 diene with a type 3 dienophile under the
Mikami-Corey protocols did prevail, thereby enabling a straight-
forward total synthesis of fluostatin C (1) and, shortly thereafter,
† Columbia University.
‡ Sloan-Kettering Institute for Cancer Research.
9
10.1021/ja7113757 CCC: $40.75 © 2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 2783-2785
2783

C O M M U N I C A T I O N S
Scheme 3. Synthesis toward Fluostatin Ca
Scheme 4. Completion of Syntheses of Fluostatins C and Ea
a Key: (a) TsOH, benzene, reflux, 2.5 h; (b) SeO2, dioxane, 90 °C, 2 h;
(c) i-Pr2NEt, air, CH2Cl2, rt, 2 h, 47% for less polar series, 38% for more
polar series, over 3 steps; (d) HF, CH3CN, rt, 4 days, 85%; (e) 1 N HCl,
acetone, rt, 10 min, 90%.
methanol (see 16).8c Reduction of the ketone group with superhy-
dride afforded a 92% yield of the equatorial alcohol which was
protected as its triisopropyl silyl derivative (see compound 17). To
complete the setting for nucleophilic epoxidation, compound 17
was treated with PPTS to provide the R,â-unsaturated ketone, 18.
Next, the oxido linkage was introduced in a stereospecific fashion
under the conditions shown12 to afford 19. During the course of
the nucleophilic epoxidation reaction, the trans junction had
undergone partial epimerization. Owing to difficulties in separation,
the two compounds were submitted concurrently to the next step
which involved treatment with osmium tetroxide,13 as shown. This
led to the formation of three products. Apparently, each of the
diastereomers of 19 gave rise to a single dihydroxylation product
(see 20). In addition, there was also obtained an over-oxidation
product shown as ketone 21 (stereochemistry of tertiary alcohol
not formally assigned). Oxidation of the diol mixture with TPAP,
as shown, afforded the previously obtained 21a as well as a new
hydroxylketone shown as 21b. Compounds 20 and 21, whose
relative configurations were not rigorously assigned, were separable
and were advanced individually wherein the two arms converged
at the stage where the stereochemical distinctions were no longer
present in the molecule (see compound 24; Scheme 4).
The transformations, conducted under the same conditions for
both series (i.e., starting with 21a or 21b), consisted of (i)
dehydration of the C6a alcohol concurrently with deprotection of
the C7 phenolic function (see 22a,b), (ii) benzylic (and allylic)
oxidation (see 23a,b),14 and (iii) dehydrogenation-tautomerization
with Hu¨nig’s base in the presence of air (see convergence product
24). Cleavage of the TIPS blocking group served to convert 24 to
fluostatin C (1). It was noted that the NMR spectra of the post-
chromatography, fully synthetic fluostatin C were variable and the
color of the seemingly pure eluates ranged between orange and
purple even following successive attempts at purification. We
wondered whether the two phenolic functions at carbons 6 and 7
manifest variable levels (and perhaps sites) of deprotonation with
attendant intramolecular stabilization of the monophenoxides
through hydrogen bonding. In support of this surmise, it was found
that treatment of the fully synthetic fluostatin C (in acetone-d6)
with a trace of trifluoroacetic acid led to a solution with a
consistently orange color, which provided reproducible NMR
spectra. We received a sample of impure fluostatin C from Professor
H.P. Fiedler which was subjected to chromatographic purification.
When the stable 1H spectra of the natural and total synthesis derived
fluostatins were measured (in the presence of TFA), they were
identical.
a Key: (a) (i-PrO)2TiCl2, (R)-1,1′-bi-2-naphthol, M.S., CH2Cl2, -35 °C,
3 days, 93%, ca. 65% ee; (b) NaOMe, MeOH, 12 h, 92%; (c) superhydride,
-78 °C, THF, 30 min, 92%; (d) TIPSOTf, 2,6-lutidine, CH2Cl2, -20 °C,
3 h, 93%; (e) PPTS, acetone, H2O, reflux, 30 h, 88%; (f) t-BuOOH, Triton
B, THF, 0 °C, 48 h, 72%; (g) OsO4, NMO, t-BuOH, acetone, H2O, 0 °C,
12 h, 20a, ca. 1.5:1, 46%, 20b, 23%; (h) TPAP, NMO, CH2Cl2, rt, 12 h,
21a, 39%, 21b, 22%.
fluostatin E (vide infra). In conjunction with the total synthesis
program, further studies of Diels-Alder reactions of type 2 and
type 3 systems were undertaken. The findings of these studies are
suggestive of an intriguingly subtle balance of forces in controlling
the diene-dienophile alignment of the cycloaddition step.
We anticipated that the prevailing regiochemical alignment in
the reaction of dienophile 7 with vinylindene-type dienes (see 5 or
6) would occur as implied in Scheme 2. In formulating this
expectation, we took note of the Corey precedent,8c wherein 7 and
9, under the Mikami conditions, cleanly afforded 10, thereby
suggesting that, within 7, the prevalent directing function is the
keto group. We also noted a reported Lewis acid catalyzed Diels-
Alder reaction (albeit under guidance by a BINOL/In(III) system)
of 5, with R-bromoacrolein (11).10 This reaction cleanly afforded
12. We took this result to suggest that the governing group within
the diene 5 is the aryl function (see asterisk) attached to carbon 2
of the local 1,3-butadiene moiety. These cases prompted anticipation
that cyclocombination of components 5 and 7 would align as
suggested (see curved double headed arrows). Remarkably, in
practice, this reaction, conducted under the Mikami-Corey pro-
tocols, afforded 13, contrary to the expected alignment. In a similar
vein, the Diels-Alder reaction of 6 and 7 cleanly afforded 14. The
structure assignment of the latter was corroborated by X-ray
crystallography.
We shall revisit the general “Diels-Alder issues” associated with
these cycloadditions. Presently, we focus on how the total synthesis
of fluostatin C was accomplished in light of these findings. In line
with our plan of synthesis adumbrated above, we next examined
the cycloaddition of 6 and 8 under the same Mikami-Corey
protocols. Indeed, this reaction afforded a 93% yield of the required
15 (Scheme 3). The structure of 15, at the time assigned by
extensive NMR analysis, was verified by completion of the total
synthesis of fluostatin C via this intermediate (vide infra).
The steps in the progression of 15 to fluostatin C started with
isomerization of the CD junction from cis to trans.11 This outcome
was accomplished through the agency of sodium methoxide in
Our next goal was the synthesis of fluostatin E (25). Treatment
of fully synthetic fluostatin C with aqueous HCl afforded the
chlorohydrin. In this case, there was no authentic sample of natural
9
2784 J. AM. CHEM. SOC. VOL. 130, NO. 9, 2008

C O M M U N I C A T I O N S
Scheme 5. Diels-Alder Reactions with Methacrolein
Peter Fiedler (Universita¨t Tu¨bingen, Germany) for kindly providing
us with the authentic sample of fluostatin C. We also thank Rebecca
Wilson and Julie Grinstead for assistance with the preparation of
the manuscript, and Daniela Buccella (Columbia University) for
crystal structure analysis.
Supporting Information Available: Experimental procedures,
compound characterization data (PDF), and X-ray structure data for
compound 14 (CIF). This material is available free of charge via the
Internet at http://pubs.acs.org.
product 25 available. However, the very close correspondence in
the high-field 1H and 13C NMR of our synthetic material with data
tabulated in the published report on fluostatin E, as well as our
independent spectral analysis, strongly suggests that the fully
synthetic chlorohydrin does indeed correspond to fluostatin E. This
finding, perhaps surprising at first glance, is actually precedented.15
With the main fluostatin total synthesis goals accomplished, we
revisited the remarkable specificities of the regio-alignments which
were displayed in the course of this program. Thus, the results in
the cycloadditions of 5 with 11 under Lewis acid catalysis (InIII)9
suggested that the “2-aryl” group governs the sense of cycloaddition
with the classical R,â-unsaturated carbonyl type of dienophile.
Moreover, in the Corey example under titaniumIV mediation,8c the
ketonic function of quinoneketal governs the sense of regiochem-
istry. Thus it seems surprising that vinylindenes 5 as well as 6 line
up with 7 such that the 1-methylene group of the local butadiene
rather than the 2-aryl appears to be dominant. We wondered whether
the apparent anomaly might reflect different directing tendencies
of the two different Lewis acid (TiIV vs InIII) systems. However,
the cycloadditions of 5 and 6 with methacrolein (Scheme 5) in the
context of the Mikami system cleanly afford 26 and 27, again
suggesting orientational control by the 2-aryl function. Yet with
each of the dienophilic quinoneketal cases (see 5 + 7, 6 + 7, and
6 + 8), control in the vinylindene seems to be manifested by the
1-methylene group.16 In principle, it could be argued in the
quinoneketals that the governing group on the dienophile is actually
the ketal, following the precedent of Gassman with simple R,â-
unsaturated ketals.17 However, the important Corey precedents (see
for instance trans-piperylene and 7) teach away from such an
interpretation. In summary, dienophiles 5 and 6 behave consistently
with well-understood dienes. Similarly, dienes 7 and 8 behave as
expected with “standard” dienophiles. Yet the cycloadditions of 5
and 6 with 7 and 8 are anomalous. In essence, the regiopreferences
of at least these dienes and dienophiles are not a fixed property of
each one component of the cycloaddition but are contingent on
one another (i.e., “contextual”).
References
(1) (a) Baur, S.; Niehaus, J.; Karagouni, A. D.; Katsifas, E. A.; Chalkou, K.;
Meintanis, C.; Jones, A. L.; Goodfellow, M.; Ward, A. C.; Beil, W.;
Schneider, K.; Sussmuth, R. D.; Fiedler, H.-P. J. Antibiot. 2006, 59, 293.
(b) Schneider, K.; Nicholson, G.; Strobele, M.; Baur, S.; Niehaus, J.;
Fiedler, H.-P.; Sussmuth, R. D. J. Antibiot. 2006, 59, 105.
(2) For studies on fluostatin A and B, see: (a) Akiyama, T.; Nakamura, K.
T.; Takahashi, Y.; Naganawa, H.; Muraoka, Y.; Aoyagi, T.; Takeuchi, T.
J. Antibiot. 1998, 51, 586. (b) Akiyama, T.; Harada, S.; Kojima, F.;
Takahashi, Y.; Imada, C.; Okami, Y.; Muraoka, Y.; Yakeuchi, T. J.
Antibiot. 1998, 51, 553.
(3) (a) Ito, S.; Matsuya, T.; Omura, S.; Otani, M.; Nakagawa, A.; Takeshima,
H.; Iwai, Y.; Ohtani, M.; Hata, T. J. Antibiot. 1970, 23, 315. (b) Hata, T.;
Omura, S.; Iwai, Y.; Nakagawa, A.; Otani, M. J. Antibiot. 1971, 24, 353.
(c) Gould, S. J.; Tamayo, N.; Melville, C. R.; Cone, M. C. J. Am. Chem.
Soc. 1994, 116, 2207. (d) Mithani, S.; Weeratunga, G.; Taylor, N. J.;
Dmitrienko, G. I. J. Am. Chem. Soc. 1994, 116, 2209.
(4) For a proposed biochemical interconversion between kinamycins and
fluostatins, see: Proteau, P. J.; Li, Y.; Chen, J.; Williamson, R. T.; Gould,
S. J.; Laufer, R. S.; Dmitrienko, G. I. J. Am. Chem. Soc. 2000, 122, 8325.
(5) (a) Bergamasco, R.; Porter, Q. N. Aust. J. Chem. 1977, 30, 1051. (b)
Quin, L. D.; Hughes, A. N.; Lawson, H. F.; Good, A. L. Tetrahedron
1983, 39, 401.
(6) For the preparation and full characterization of compound 6, see Supporting
Information.
(7) (a) Carreno, M. C.; Farina, F.; Galan, A.; Ruano, J. L. G. J. Chem. Res.
1979, 296. (b) Carreno, M. C.; Farina, F.; Galan, A.; Ruano, J. L. G. J.
Chem. Res. 1981, 370. (c) Jarvo, E. R.; Boothroyd, S. R.; Kerr, M. A.
Synlett 1996, 9, 897. (d) Gerstenberger, I.; Hansen, M.; Mauvais, A.;
Warichow, R.; Winterfeldt, E. Eur. J. Org. Chem. 1998, 63, 643.
(8) (a) Mikami, K. M.; Terada, M.; Narisawa, S.; Nakai, T. Synlett 1992,
255. (b) Mikami, K. M.; Motoyama, Y.; Terada, M. J. Am. Chem. Soc.
1994, 116, 2812. (c) Breuning, M.; Corey, E. J. Org. Lett. 2001, 3, 1559.
(9) Compounds 7 and 8 were prepared by sequential oxidation of the
corresponding phenols with iodobenzene diacetate in MeOH. Pelter, A.;
Elgendy, S. M. A. J. Chem. Soc., Perkin Trans. 1 1993, 1891. This was
followed by BF3•OEt2-catalyzed transacetalization with ethylene glycol.
Pirrung, M. C.; Nunn, D. S. Tetrahedron Lett. 1992, 33, 6591. The
experimental details are described in Supporting Information.
(10) Teo, Y.-C.; Loh, T.-P. Org. Lett. 2005, 7, 2539.
(11) In the absence of the epimerization, isomerization of the nonenone double
bond to the tetrasubstituted position occurred in the epoxidation stage.
(12) Knight, S. D.; Overman, L. E.; Pairaudean, G. J. Am. Chem. Soc. 1993,
115, 9293. The configuration at C4a in compounds 19a,b is not assigned.
(13) VanRheenen, V.; Kelly, R. C.; Cha, D. Y. Tetrahedron Lett. 1976, 17,
1973. The configuration at C6a in compounds 20a,b and 21a,b is not
assigned.
Obviously, there are various ad hoc theories that can be
summoned to explain these observations. Still we find the high
regiospecificity of each of the seemingly contradicting modalities
remarkable. In the future, we hope to sort out the underlying issues
of context. It may be that context associated specificities reflect
differing steric compatibilities. We hope to explore such issues in
the context of a broader computational level exploration. What is
clear is the continuing power of well-designed Diels-Alder
reactions in enabling well-directed thrusts at complex bioactive
targets (cf. 1 and 25). Apparently, there remain ever more subtle
messages to students of the venerable Diels-Alder reaction.
(14) Bhalerao, U. T.; Rapoport, H. J. Am. Chem. Soc. 1971, 93, 4835.
(15) Edwards, R. L.; Maitland, D. J.; Scowen, I. J.; De Sousa, A. J. T.; Whalley,
A. J. S. J. Chem. Soc., Perkin Trans. 1 2001, 537.
(16) (a) Schmidt, C.; Sabnis, S. D.; Schmidt, E.; Taylor, D. K. Can. J. Chem.
1971, 49, 371. (b) Dickinson, R. A.; Kubela, R.; MacAlpine, G. A.;
Stojnac, Z.; Valenta, Z. Can. J. Chem. 1972, 50, 2377. (c) Danishefsky,
S. J.; Zamboni, R.; Kahn, M.; Etheredge, S. J. J. Am. Chem. Soc. 1980,
102, 2097. (d) Danishefsky, S. J.; Zamboni, R.; Kahn, M.; Etheredge, S.
J. J. Am. Chem. Soc. 1981, 103, 3460. (e) Danishefsky, S. J.; Kahn, M.
Tetrahedron Lett. 1981, 22, 489. (f) Hu, Q.-Y.; Zhou, G.; Corey, E. J. J.
Am. Chem. Soc. 2004, 126, 13708.
(17) (a) Gassman, P. G.; Singleton, D. A.; Wilwerding, J. J.; Chavan, S. P. J.
Am. Chem. Soc. 1987, 109, 2181. (b) Magnuson, S. R.; Sepp-Lorenzino,
L.; Rosen, N.; Danishefsky, S. J. J. Am. Chem. Soc. 1998, 120, 1615.
Acknowledgment. Support for this research was provided by
the National Institutes of Health (HL25848). We thank Prof. Hans-
JA7113757
9
J. AM. CHEM. SOC. VOL. 130, NO. 9, 2008 2785