Total synthesis of the potent antitumor agent roseophilin: A concise approach to the macrotricyclic core

Full text of DOI:10.1021/ja963795b

2944
J. Am. Chem. Soc. 1997, 119, 2944-2945
Scheme 1
Total Synthesis of the Potent Antitumor Agent
Roseophilin: A Concise Approach to the
Macrotricyclic Core
Alois Fu¨rstner*,† and Holger Weintritt
Max-Planck-Institut fu¨r Kohlenforschung
Kaiser-Wilhelm-Platz 1, D-45470 Mu¨lheim/Ruhr, Germany
ReceiVed NoVember 1, 1996
Scheme 2
Roseophilin (1), a structurally unique metabolite isolated from
a culture broth of Streptomyces griseoViridis, exhibits cytotox-
icity against several human epidermoid and leukemia cell lines
in the submicromolar range.1,2 This very promising biological
profile renders 1 a new lead compound in the search for
antitumor agents and a rewarding target for total synthesis.
Obviously, the major challenge toward this end is the preparation
of segment A, which when condensed with the known hetero-
cyclic ring system B2a according to literature procedures2a,3 will
afford the desired ansa-bridged 1-azafulvene core of the natural
product (Scheme 1). As part of our endeavors in the synthesis
of physiologically active compounds4 we now disclose a concise
and highly flexible approach to this intricate macrotricyclic
skeleton which may easily be adapted to the synthesis of
analogues as well.
Our plan is guided by the idea to effect the macrocyclization
such that it also sets the stage for a convenient construction of
the ketopyrrolic entity of the target. Based on the subtle
differences in reactivity of various allylic precursors in palladium
catalyzed substitution reactions,5 we perceived a well-orches-
trated manifold for this very purpose (Scheme 2). Driven by
the release of the ring strain, the oxidative addition of Pd(0)
into a difunctional substrate of the general type I will regiose-
lectively occur at the vinyloxirane site,6 provided that the allylic
OR′ group is properly tuned. The alkoxide II thus formed
deprotonates the tethered prenucleophile of the malonate type
which will attack the allylpalladium complex and lead to the
formation of a macrocyclic ring according to literature prece-
dence.7 By taking advantage of the juxtaposition of the OR′
group and the incoming ester, a simple lactonization of these
entities activates the remaining allylic position. A second
palladium-catalyzed reaction with an amine as the nucleophile
may then not only deliver the desired pyrrole ring encoded in
the 1,4-dioxygen functionality of the substrate (III f IV f
V)8 but also will liberate the acid for an ensuing acylation of
its C-2 position (V f VI).
This concept was reduced to practice as shown in Scheme 3.
O-Silylation of the known alcohol 29 with TBDMSCl, followed
by a chloride for iodide exchange and subsequent reaction of
the rather unstable allylic iodide with tetrahydrothiophene in
the presence of AgBF4 in thoroughly dried acetone, afforded
the nicely crystalline sulfonium salt 4 in good overall yield.10
Its deprotonation with t-BuLi in THF at -78 °C followed by
trapping of the sulfur ylide11 formed in situ with 9-bromonona-
nal12 gave the desired vinyloxirane 5 in 84% isolated yield which
can be alkylated with methyl (phenylsulfonyl)acetate in DMF
under standard conditions at the bromide terminus without
affecting the labile functionality at the other end of the chain.7
This simple sequence provided gram amounts of compound 6
which is suitable to test the palladium-manifold outlined above
as the key strategic element of our synthesis plan.
Gratifyingly, substrate 6 cyclized smoothly to the 13-
membered carbocyclic ring 7 in very well reproducible 85%
isolated yield when slowly added to a refluxing solution of
† Tel. Int. 208-306-2372. Fax: Int. 208-306-2980.
(1) Hayakawa, Y.; Kawakami, K.; Seto, H.; Furihata, K. Tetrahedron
Lett. 1992, 33, 2701-2704.
(2) For synthetic studies on roseophilin segments or model compounds,
see: (a) Nakatani, S.; Kirihara, M.; Yamada, K.; Terashima, S. Tetrahedron
Lett. 1995, 36, 8461-8464. (b) Kim, S. H.; Fuchs, P. L. Tetrahedron Lett.
1996, 37, 2545-2548.
(7) See the following for leading references: (a) Trost, B. M.; Warner,
R. W. J. Am. Chem. Soc. 1982, 104, 6112-6114. (b) Trost, B. M.; Warner,
R. W. J. Am. Chem. Soc. 1983, 105, 5940-5942. (c) Kende, A. S.; Kaldor,
I.; Aslanian, R. J. Am. Chem. Soc. 1988, 110, 6265-6266. (d) Trost, B.
M.; Hane, J. T.; Metz, P. Tetrahedron Lett. 1986, 27, 5695-5698. (e)
Review: Trost, B. M. Angew. Chem. 1989, 101, 1199-1219.
(8) (a) For another palladium-catalyzed pyrrole synthesis, see: Trost,
B. M.; Keinan, E. J. Org. Chem. 1980, 45, 2741-2746. (b) For a timely
compilation of pyrrole chemistry, see: Gossauer, A. in Houben-Weyl,
Methoden der Organischen Chemie; Kreher, R. R., Ed.; Thieme: Stuttgart,
1994; Vol. E 6a, Part 1, pp 556-798.
(9) Chalova, O. B.; Christoedova, G. B.; Kiladze, T. K.; Germash, E.
V.; Kantor, E. A.; Rakhmankulov, D. L. Zh. Prikl. Khim. 1988, 61, 934-
937; CA: 110: 38603b.
(10) Rosenberger, M.; Newkom, C., Aig, E. R. J. Am. Chem. Soc. 1983,
105, 3656-3661. Since the excess of tetrahydrothiophene can be removed
in Vacuo, this sulfide is preferred over the non-volatile PhSPh11a for practical
reasons.
(11) (a) t-BuLi was recommended as the deprotonating agent of choice,
c.f.: LaRochelle, R. W.; Trost, B. M.; Krepski, L. J. Org. Chem. 1971, 36,
1126-1136. (b) Review: Trost, B. M.; Melvin, L. S. Sulfur Ylides;
Academic Press: New York, Organic Chemistry Series, 1975; Vol. 31.
(12) Obtained in 78% yield by oxidation of commercially available
9-bromo-1-nonanol with the Dess-Martin periodinane;14 the analytical data
are in accordance with those reported: Muralikrishna, C.; Dasaradhi, L.;
Rao, S. J.; Bhalerao, U. T. Ind. J. Chem. 1989, 28B, 579-580.
(3) The prodigiosin tripyrrole pigments are the closest relatives to
roseophilin. Similar condensation reactions have been used for their
synthesis, c.f.: (a) Wasserman, H. H.; Keith, D. D.; Nadelson, J. J. Am.
Chem. Soc. 1969, 91, 1264-1265. (b) Boger, D. L.; Patel, M. J. Org. Chem.
1988, 53, 1405-1415. (c) Wasserman, H. H.; Lombardo, L. J. Tetrahedron
Lett. 1989, 1725-1728. (d) Rapoport, H.; Holden, K. G. J. Am. Chem.
Soc. 1962, 84, 635-642. (e) D’Alessio, R.; Rossi, A. Synlett 1996, 513-
514 and lit. cited.
(4) (a) Fu¨rstner, A.; Weintritt, H.; Hupperts, A. J. Org. Chem. 1995, 60,
6637-6641. (b) Fu¨rstner, A.; Ernst, A. Tetrahedron 1995, 51, 773-786.
(c) Fu¨rstner, A.; Ernst, A.; Krause, H.; Ptock, A. Tetrahedron 1996, 52,
7329-7344. (d) Fu¨rstner, A.; Langemann, K. J. Org. Chem. 1996, 61,
8746-8749. (e) Fu¨rstner, A.; Kindler, N. Tetrahedron Lett. 1996, 37, 7005-
7008. (f) Fu¨rstner, A.; Langemann, K. J. Org. Chem. 1996, 61, 3942-
3943. (g) Fu¨rstner, A.; Jumbam, D. N.; Seidel, G. Chem. Ber. 1994, 127,
1125-1130. (h) Fu¨rstner, A.; Nikolakis, K. Liebigs Ann. 1996, 2107-2113.
(j) Fu¨rstner, A.; Konetzki, I. Tetrahedron 1996, 52, 15071-15078. (k)
Fu¨rstner, A.; Baumgartner, J. Tetrahedron 1993, 49, 8541-8560.
(5) Review: Tsuji, J. Palladium Reagents and Catalysts; Wiley, New
York, 1995.
(6) (a) Trost, B. M.; Molander, G. A. J. Am. Chem. Soc. 1981, 103,
5969-5972. (b) Tsuji, J.; Kataoka, H.; Kobayashi, Y. Tetrahedron Lett.
1981, 22, 2575-2578.
S0002-7863(96)03795-X CCC: $14.00 © 1997 American Chemical Society

Communications to the Editor
J. Am. Chem. Soc., Vol. 119, No. 12, 1997 2945
Scheme 3a
appropriate nucleophile to the resulting enone may well effect
this transformation. Since the shielding excerted by the fairly
rigid ansa-chain not only provides effective facial guidance in
the Michael addition step but also will force the protonation of
the resulting enolate to occur from the same side, the proper
relative configuration of the newly formed chiral centers at C-22
and C-23 is likely to ensue.
However, our initial attempts along these lines revealed that
the tricyclic enone formed upon elimination of the sulfone group
with t-BuOK in THF at ambient temperature owes a peculiar
reactivity to its highly strained character.16 Since it undergoes
immediate dimerization,17 we faced the problem to effect the
crucial elimination step in the presence of an appropriate
nucleophile in order to intercept the enone immediately subse-
quent to its formation. Cuprates turned out to be unsuitable
for this purpose because of their insufficient thermal stability
and their poor reactivity in the presence of excess t-BuOK. A
viable alternative, however, was found in zincate chemistry.18
Specifically, addition of an excess of t-BuOK to a solution of
11 and i-PrMe2ZnMgCl (formed in situ from ZnCl2‚TMEDA,
2MeLi, and i-PrMgCl) in THF at ambient temperature afforded
the desired product 12 in 47% isolated yield together with some
dimeric byproducts. Although this result may likely be further
improved, it is respectable in view of the complexity of this
specific two-step/one-pot transformation. The structural integ-
rity of 12 was unambiguously assigned by means of extensive
2D NMR investigations. Specifically, the coupling pattern of
a Compounds 5-9 have been obtained as mixtures of all possible
stereoisomers.13 [a] TBDMSCl, DBU, CH2Cl2, room temperature, 90%;
[b] (i) NaI, acetone, 50 °C; (ii) tetrahydrothiophene, AgBF4, acetone,
room temperature, 73%; [c] t-BuLi, then 9-bromononanal, THF, -78
°C f room temperature, 84%; [d] KH, DMF, room temperature; [e]
Pd(PPh3)4 (10 mol %), dppe (20 mol %), THF, reflux; [f] THF, room
temperature; [g] Dess-Martin periodinane, CH2Cl2, room temperature;
[h] Pd(PPh3)4 (15 mol %), THF, 35 °C; [i] (1) CH2Cl2, room
temperature; (2) 1,2-dichloroethane, reflux; [j] THF, room temperature.
1
H-23 in the H NMR spectrum at 600 MHz (d, 2.63 ppm) as
well as the observed crosspeaks in the nJ(C,H) correlated spectra
clearly locate the isopropyl group in the cyclopentanone ring
rather than in the ansa chain. This rules out that strain and/or
antiaromaticity prevent the elimination of the sulfone into the
catalytic amounts of Pd(PPh3)4 and dppe (dppe ) bis(diphenyl-
phosphino)ethane) in THF over a period of 6 h (≈0.0014 M
final concentration).13 The observed, selective activation of the
vinyloxirane group was very much in line with our anticipation.
It should also be mentioned that the use of polymer-bound
Pd(0) precatalysts previously recommended in order to ensure
pseudohigh dilution conditions7a,b led to significantly lower
yields in this crucial macrocyclization step. Desilylation of 7
afforded the somewhat strained lactone 8. Best results in this
pivotal step were obtained in a buffered medium with a mixture
of TBAF/NH4F as the reagents. Oxidation of the OH group
with the Dess-Martin periodinane14 followed by treatment of
ketone 9 thus formed with benzylamine in the presence of
catalytic amounts of Pd(PPh3)4 in THF resulted in the clean
formation of the desired pyrrole carboxylic acid 10 in 70%
isolated yield. The reaction is best carried out at ≈35 °C in
order to avoid problems caused by premature base-induced
elimination of the sulfone group and/or concomittant decar-
boxylation of the acid formed. Compound 10 was then
converted into the acid chloride under neutral conditions using
the highly convenient R-chloroenamine reagent introduced by
Ghosez et al.15 Subsequent treatment with SnCl4 in refluxing
1,2-dichloroethane cleanly afforded the tricyclic ketone 11 in
71% yield by an intramolecular Friedel-Crafts acylation. This
reaction occurs regioselectively at C-2 rather than C-4 of the
pyrrole ring as can be unambiguously deduced from the
observed nJ(C,H) correlated NMR spectra (for details see
Supporting Information).
3
five-membered ring. Moreover, a JH22,H23 of ≈0 Hz matches
that observed in the spectrum of roseophilin itself and pinpoints
the proper relative configuration at these chiral centers (for
details see Supporting Information).
In summary we have achieved the first synthesis of the
intricate macrotricyclic core segment of roseophilin 1 in only
10 efficient steps. Our approach merges the potential of an
established palladium-catalyzed macrocyclization reaction with
a conceptually new entry into substituted ketopyrroles. Based
on this, one may not only gain access to roseophilin itself but
also can prepare various analogues of this promising antibiotic
for the study of its structure/activity profile. Efforts along these
lines as well as the completion of the total synthesis by
condensation of 12 with the known heterocyclic side chain2a
are actively pursued in our laboratory and will be reported in
due course.
Acknowledgment. H.W. thanks the Fonds der Chemischen Indus-
trie, Frankfurt, for a Chemiefonds-Stipendium.
Supporting Information Available: Relevant preparative proce-
1
dures, listing of the spectral and analytical data, and copies of the H,
13C, and some 2D NMR spectra of all new compounds (41 pages). See
any current masthead page for ordering and Internet access instructions.
JA963795B
(16) The use of other bases and/or lower temperatures in this elimination
step were in vain.
(17) The dimeric structure of the product formed ensues from GC/MS
analyses of the reaction mixtures. We did not yet investigate the structure
of this dimer any further. However, there is some precedence in the recent
literature that structurally related heterocyclic ketones exhibit a strong bias
for dimerization, c.f.: (a) Comer, M. C.; Despinoy, X. L. M.; Gould, R.
O.; McNab, H.; Parsons, S. J. Chem. Soc., Chem. Commun. 1996, 1083-
1084. (b) Neidlein, R.; Jeromin, G. Chem. Ber. 1982, 115, 706-713.
(18) (a) Kjonaas, R. A.; Hoffer, R. K. J. Org. Chem. 1988, 53, 4133-
4135. See also: (b) Kjonaas, R. A.; Vawter, E. J. J. Org. Chem. 1986, 51,
3993-3996. (c) Morita, Y.; Suzuki, M.; Noyori, R. J. Org. Chem. 1989,
54, 1785-1787. (d) Isobe, M.; Kondo, S.; Nagasawa, N.; Goto, T. Chem.
Lett. 1977, 679-682. For a study on the interaction of organozinc reagents
with t-BuOK, see: (e) Rathke, M. W.; Yu, H. J. Org. Chem. 1972, 37,
1732-1734.
We envisaged using the sulfone moiety at C-22 (roseophilin
numbering, Scheme 1) in order to introduce the missing
isopropyl substituent at the adjacent position. A base-induced
elimination of PhSO2H followed by a 1,4-addition of an
(13) Products 5-9 shown in Scheme 3 were obtained as mixtures of all
possible diastereoisomers. Since all of them lead to the final product, no
attempts were made to separate the individual compounds.
(14) (a) Dess, D. B.; Martin, J. C. J. Org. Chem. 1983, 48, 4155-4156.
(b) Meyer, S. D.; Schreiber, S. L. J. Org. Chem. 1994, 59, 7549-7552.
(15) Devos, A.; Remion, J.; Frisque-Hesbain, A.-M.; Colens, A.; Ghosez,
L. J. Chem. Soc., Chem. Commun. 1979, 1180-1181.