вход по аккаунту


Epothilones Promising Natural Products with Taxol-Like Activity.

код для вставкиСкачать
Epothilones: Promising Natural Products with Taxol-Like Activity
Ludger Wessjohann*
A new group of macrocyclic natural products, the epothilones
1, has enabled research into structure-activity relationships in
the stabilization of microtubules to be conducted faster and
more efficiently. This could lead to an antitumor agent with
antimitotic activity that is more effective than taxol. The
epothilones were first isolated from myxobacteria of the genus
sorangium by two groups at the Gesellschaft fur biotechnologische Forschung (GBF): Hofle et al., who were also responsible
for the structure elucidation,[’. and Reichenbach et al. As
these bacteria can be cultivated in large amounts, several grams
of epothilones can be obtained in a single batch. The relative
configuration of the seven stereogenic centers of epothilone B
(1 b) was determined by X-ray structure analysis, the absolute
configuration by chemical degradation of C13 - C16 to hydroxysuccinic acid.
Specific antimycotic and cytotoxic activity (2 ngmL- ’) was
established early on.[’] The fundamental breakthrough came
with the results of Bollag et al., who were the first to recognize
the microtubule-stabilizing activity of ep~thilones.[~]
This property is shared by only two other natural products, discodermolide and paclitaxel, the latter being commercially available
under the name of “taxol” or “taxotere”. These pharmaceutical
preparations are the most important results in the development
of antitumor agents in the past decade.14] After more than
20 years and tests on over 140000 (!) synthetic substances and
extracts from natural products[3,51 an activity similar to that of
taxol was found-again in a natural product. In some assays
epothilone B is twice as active as epothiloneA.[’a.dl The
epothilones can even displace taxol from the binding site and act
faster. More importantly, they are between 1000 and 5000 times
more active against multiply resistant cell lines.[ld,3 , ’I Their
mode of action is also similar to that of taxol. However, the
binding sites are probably not identical, but instead overlap.
The markedly different spatial structure of the epothilones,
despite allowing a certain flexibility, could not be made congruent with that of taxol. Another important advantage of
epothilones over taxol is their significantly better solubility. A
substantial risk of the taxol therapy derives from problems of
application, since the necessary solubilizers can lead to powerful, sometimes life-endangering side effects.
The constitution of the epothilones suggests that they are
derived from the polyketide metabolism like similarly built
macrolides. The synthesis apparently commences with N-acetylcysteine, which becomes part of the aromatic thiazole. The compounds show three structural deviations from simple polyketide
macrocycles: apart from the thiazole ring, it is mainly the oxirane ring and the C4 dimethyl group that disturb the usual role
of methyl and hydroxyl branches (“triads”) .[‘I Under basic conditions the dimethyl group promotes retro-aldol cleavage at the
keto group.[g1Retrosynthetic analysis leads to three fragments
containing these functional groups (Scheme 1): an aromatic
unit, an aldol fragment in the “southern half’ of the ring, and
an alkyl epoxide in the “northern half ’. This pattern is essentially followed by the published syntheses.[6. 16]
The first total synthesis of epothiloneA (1 a) was achieved by
Danishefsky’s group from the Sloan Kettering Institute for
Cancer Research,[””] who also recently completed the first synthesis of epothilone B.[61 Shortly afterwards, Nicolaou
et al.[”as and Schinzer et al.113a1
presented their approaches
to epothilone A. Their work, together with some partial sequences by the groups of Mul~er,[’~]
Kalesse,” 51 and Wessjohann[l6]are described herein. In order to limit the discussion to
the essential synthesis of the 16-membered macrocycle, it may
Scheme 1. Structure of epothiloneA (1 a) and B (1 b) as well as retrosynthetic analysis; overview of the disconnections in the macrolide synthesis by: A: all research
groups, D: Danishefsky et al., N : Nicolaou et al., S : Schinzer et al. as well as partial
solutions by: H: Hofle et al., K : Kalesse et al., M : Mulzer et al.. W: Wessjohann
et a].; the main fragments are in bold type.
Dr. L A. Wessjohann
lnstitut fur Organische Chemie der Universitat
Karlstrasse 23, D-80333 Munchen (Germany)
Fax: Int. code +(89)5902-483
e-mail: law($
Angew. Chem. Int. Ed. Engl. 1991,36. N o . 7
l a , R=H
1b, R=Me
Allylborane: N. M .. S
D. W’ ..
0 VCH ~rlugsgesellschufimbH. 0-69451 Weinheim,1997
0570-083319713607-07153 17.50t SO!O
combine the building blocks is more effective than metathesis;
however, a drawback-at least of Nicolaou's approach-is the
less effective access to the precursors.
Since the method applied by Schinzer et al. and the partial
sequences published so far resemble the approach adopted by
Nicolaou et al. they are now treated in parallel. The building
block 6 requires stereoselective introduction of the
hydroxy group at C15, which was achieved very
f N': Metathesis . X = Y = CH., (50%) 7
effectively by Nicolaou et al. using allyl isoMe0 ,,
A (
N2:Wttig, X = 0. Y = PPh3 (- 90%)
pinocampheylborane Ipc,B(allyl) .[' la] The selecN
tive conversion of diene 6 (X = CH, - 0 ) was
by Sharpless dihydroxylation, diol cleav3
age, and oxidation. In contrast, Schinzer et al. first
N': 3 steps (78%)
7 steps (36%)
synthesize the aldehyde 6 (X = 0, seven steps,
N2:6 steps (59%)
14%), which is then converted to the correspond(incl. Wittig)
ing alkene (X = CH,) in a Wittig reaction.['3a1In
building block 7 the configuration at C8 is adjusted
by utilizing the reliable Enders and Evans auxil6
iaries. In the N 2 route, Nicolaou et al. alkylate the
SAMP hydrazone 5;[12]Schinzer et al. methylate a
Aldol (> 2 equiv. LDA)
N': > 45% (- 2 1 )
and very effectively.['3b1
The synthesis of 9 and 10 and their aldol coupling B is of particular interest (Schemes 2 and 3).
Two aldol reactions or carbonyl additions are pos0
4 steps (55%)
sible for the synthesis of the C3 stereocenter. The
C3 -+ C4 bond, in particular, should be sensitive to
basic reaction conditions and could, like in
Scheme 2. Key steps of the total syntheses of 1 a by Nicolaou et al. N': B -+ C -+A, > 9 +4 + 3
epothilone itself, be prone to the reverse reaction.
steps, overall yield: $ 6 % l a ; N': A - B
C, > 17 + 6 +4 steps, overall yield: 1 2 % l a ; DCC =
dicyclohexylcarbodiimide, DMAP = 4-(N,N-dimethylamino)pyridine, Ipc = isopinocampheyl.
Similar solutions to this problem are offered by
LDA = lithium diisopropyl amide, PG = protecting group, for example TBS = SiMe,rBu;
Mulzer et al.[14]and Nicolaou et al.:[llalthe addiYamaguchi esterification: 2.4,6-trichlorobenzoyl chloride, NEt,, then DMAP.
tion of Ipc,B(allyl) (4)to a C3 aldehyde, followed
by oxidative decomposition of the allyl group to
syntheses are the sequence and the method of coupling A. The
give the carboxylic acid (four steps). Whereas Nicolaou et al.
key step of N1 is the ring-closing metathesis A' of a bisalkene,
employ 8 as electrophile, Mulzer et al. use a C3-C9 unit
which is obtained by the aldol reaction of the aldehyde 7
obtained in excellent yield, putting first the B-type aldol reac(Y = CH,, B') with the dianion of 9, followed by esterification
tion (Scheme 2). The route taken by Schinzer et al.[13b1goes
with 6 (X = CH,, C'). The metathesis A' is achieved by the
in the opposite direction; they use Ipc2B(3,3-dimethyl allyl)
dilution principle in surprisingly high yields (50 % of the 2- and
to close the C4 -+ C3 bond (Scheme 1). However, the
35% of the E-isomer) by using the Grubbs catalyst
double bond must not only be subsequently oxidized but the
[RuCl,(=CHPh)(PCy,),] (15 mol%).[""] In a similar routine
additional carbon atoms must be supplicd to the compound
by Schinzer et al., an almost identical precursor was obtained in
to yield 10 (Scheme 3). The apparent possibility of direct94 % yield ( E : Zapprox. 1 : 1).[13a1 The enormous improvements
ly producing building blocks of the type 9 and 10 in an
in the field of metathesis, such as the development of defined,
stablc catalysts and the high tolerance of functional groups,
have allowed this to become a standard method in olefin synthesis.[' 71 Despite this success, it appears that it is not always advisable to follow the latest trends. For example, Danishefsky et al.
were unable to achieve ring-closing metathesis of a fully func10
tionalized precursor with the same catalyst. Model compounds
led to varying results.['0b1 Evidently, at least with epothilone
precursors, the success of the ring-closing metathesis is strongly
dependent on the substituents and the protecting groups employed.
Accordingly, in the second, later route (Scheme 2 ; N2) Nicolaou et al. employ a Wittig reaction of 6 (X = 0) with 7
(Y = PPh,, A')
After an aldol reaction with 9 (see B2) the
ring is closed by macrolactonization according to the approved
Scheme 3. Highly stereoselective C6-C7 aldol reaction (c.f. Scheme 2, B) in the
Yamaguchi method (C'). This more classic approach to
total synthesis by Schinzer et al.
be noted ahead that the epoxidation is the last step in all cases
(Scheme 1). Another feature common to all syntheses is the
introduction of the thiazole group by a Wittig reaction at C16.
Nicolaou et al. presented two similar, highly convergent syntheses (N1 and N2, Scheme 2), in which the basic units 6 , 7 , and
9 are combined.[1'a.121The main differences between the two
VCH Verlu~ge.~ellsckaft
mhH. 0-69451 Weinkeim, 1997
0570-0833/97~3607-0716$ 17.50+ .50/0
Angew. Chem. Int. Ed Engl. 1997. 36. No. 7
ingly, the three-membered ring is preferentially opened solvolytically at the inner bond." Obl Dehalogenation and refunctionalizing ring-opening afford the semiprotected dialdehyde 17. The
chain extension from C9 in 17 to C11 in 21 could only be effected through two consecutive Wittig reactions (with
MeOCH=PPh, and CH,PPh,). This creates the necessary prerequisites for a C l 1 C12 Suzuki coupling, which is an alternative to the described (unsuccessful) metathesis: The alkene 21
activated with 9-borabicyclo[3.1 .l]nonane (9-BBN) is coupled
to the Z-vinyl iodide 20a in 71 % yield under palladium catalysis. The stereoselective macroaldolization of the liberated C3aldehyde 22 is remarkable, and affords up to 70% de of the
correct diastereomer (51 % yield). The oxidation of the etherprotected CS-hydroxy group takes place after some protecting
group transformations with the Dess- Martin reagent. The concluding epoxidation with dimethyldioxirane (45 YO) proceeds
surprisingly diastereoselectively for a macrocycle (20: 1 at
- 50°C)['oa1 and was improved by Schinzer et al. (48% at
- 35 cC).[13a1
When m-chloroperbenzoic acid is used. the epoxidation is less selective, however, a higher overall yield (55 % at
0 "C) of the correct isomer is obtained." l a ]
Epothilone B (1 b)[61 is for the most part synthesized according to the above approach. However, it contains a methyl group
at C12 (Scheme 4), which is obtained by reaction of the unusual
Wittig reagent Ph,P=C(I)CH, with a C11 aldehyde (yield
43%). The subsequent Suzuki coupling that leads to 22b is
more effective than that for epothiloneA (77% yield), as is the
final epoxidation with dimethyl dioxirane (70 %), which affords
the desired diastereomer in excellent excess of 14: 1 despite the
E-methyl group. In this paper by Danishefsky et al., in vitro
studies of the activities of derivatives are discussed for the first
time.[61It is not surprising that the natural products. especially
epothilone B, show the highest activity. The nonepoxidized precursors, the deoxyepothilones 2, are only slightly, their trans
isomers significantly less active.
The first, more extraordinary epothilone synthesis by
Danishefsky et al. is lengthier and less convergent than those by
Schinzer et al. and Nicolaou et al. To date none of the syntheses
can compete with the fermentation, but they offer many possi-
aldol reaction was accomplished by Wessjohann et a1.[l6I in
a one-pot process by applying the chromium Reformatsky
method to suitable 2-bromoacetyl derivatives, using Evans'
auxiliary. In contrast to other aldol procedures, this bears
no risk of a retroaldol reaction. Both the C2 C3 and the
C 4 + C 3 aldol reactions were successful (see W' and W2,
Scheme 1). The diastereoselectivities for chromium-complexed
enolates of x-unsubstituted acetyl oxazolidinones are higher
than with other species: de with Cr: 84-99, with B, Li, Ti:
The C6 C7 aldol reaction (Scheme 2, B, and Scheme 3)
should yield only one of the four possible isomers stereoselectively. All authors obtained only the desired syn diastereomers
when Z-lithium enolates were employed; however, in the reaction of 7 with the dianion of 9 the diastereomers form in approximately equal amounts (Scheme 2, B).[lla.
Mulzer et al.
achieve a ratio of 4: 1 by using an achiral enolate and a chiral
C7-C9-0-Ar building block (see 13 with Ar = 4methoxybenzyl), although C8 offers the only possibility of
Kalesse et al. employ the almost identical C7-C9-0-Ar aldehyde 13
(Ar = benzyl),r'sl while Schinzer et al. use the aldehyde 7
(Y = CH,, Scheme 3),[13"1with which they obtained solely the
desired diastereomer. This important improvement is probably
due to a "matched case" influence of the chiral enolates and
is an important strength of the approach adopted by Schinzer
et al.
The syntheses developed by Danishefsky et al. (Scheme 4) are
substantially different from the other approaches.r6. b1
Whilst the synthesis of 20a-the stereocenter CIS stems from
R-glycidol--is effected under standard conditions, the ring closure and the synthesis of the second main precursor 21 are
unusual. The diastereomer problem of the C6 + C7-aldol reaction is avoided by a chelate-controlled cycloaddition of 13 to the
diene 12. There is. however, a price to pay for this advantage:
functionalization of C9 in the intermediate 14 is difficult to
effect (see also ref. [ 1 4 * l51), and a methyl group is missing at C4.
The latter problem is elegantly, but not exactly practically, solved
by using a Simmons - Smith cyclopropanation, followed by oxidative ring-opening with electrophilic iodine (see 15). Interest-
Suzuki (71%)
Me Me
ma, R=H
2 Deprotection and
Dess-Marlin oxidation
1 AJdol (KHMDS. a 51%.
22b, R=Me
Scheme4. First total syntheses of l a and b by Ddnishefsky et al., > 2 2 +13 steps. overall yield: 1 2 % l a , KHMDS = potassium hexamethyldlsllazlde. MOM
methoxyrnethyl, NIS = 3-iodosuccinimide, IPS = triphenylsilyl. Ar = benzyl, Suzuki coupling see text.
Angew Chmi. Inr Ed. Eng/. 1997, 36, No. 7
VErlagsgesellschufr mhH. 0-69451 Wemheim, 1997
0570-0833197!3607-0717 3 17.50+ SO.0
bilities of derivatization and variation. The course of the in vivo
tests will be pivotal for the further development of the
epothilones and their derivatives. If the tests are positive, the
advantages discussed in the beginning could lead to a replacement of the already established taxoids not only from the receptors. Apart from their immediate medicinal applications, the
epothilones offer the possibility to comprehensively study the
effect of tubulin stabilization with respect to structure-activity
relationships for the first time, as derivatives will be available
faster and at a more reasonable price than with taxol. It remains
to be seen for both epothilone and taxol whether the optimization carried out by nature can significantly be surpassed, since
so far no new or even only considerably better substances with
these effects have been obtained. Notwithstanding this challenge to synthetically based drug screening, the discovery of the
epothilones is already an extremely important advance in the
fields of tumor and mitosis research.
German version- Angeu. Chem. 1997, 109. 739-742
Keywords: antitumor agents epothilones
total synthesis
- natural products -
[l] a) G. Hofle, N. Bedorf, H. Steinmetz, D. Schomburg, K. Gerth, H. Reichenbach, Angew. Chem. 196,108, 1671-1673; Angen. Chem. Int. Ed. Engl. 1996,
35, 1567-1569; b) K. Gerth, N. Bedorf, G. Hofle, H. Irschik, H. Reichenbach,
J. Antihiof. 1996, 49, 560-563; c) G. Hofle in Scientific Annual Report 1991
(Ed.: J. H. E. Walsdorff), Gesellschaft fur Biotechnologische Forschung mbH,
Braunschweig 1992; d) G. Hofle in Wissenschaftlicher Ergebnisbericht 1995
(Ed.: J. H. E. Walsdorff), Gesellschaft fur Biotechnologische Forschung mbH,
Braunschweig 1996. pp. 86-89; e) G. Hofle, N. Bedorf, K. Gerth, H. Reichenbach (Gesellschaft fur Biotechnologische Forschung mbH (GBF), Braunschweig), DE 4138042, 1993 [Chem. Abstr. 1994, 120, 52841rl.
121 In contrast, an earlier modeling and NMR study led to incorrect results: S.
Victory, D. G. Vander Velde, R. K. Jalluri, G . L. Grunewald, G. 1. Georg,
Biochem. Med. Chem. Lett. 1996,6, 893-898.
[3] D. M. Bollag, P. A. McQueney, J. Zhu, 0. Hensens, L. Koupal, J. Liesch, M.
Goetz, E. Lazarides, C. M. Woods, Cancer Res. 1995, 55, 2325-2333; R. J.
Kowalski. E. ter Haar, R. E Longley, S . P. Gunasekera, C. M. Lin, B. W. Day,
E. Hamel, contribution to the Meeting of the Am. Assoc. Cancer Res. 19%
[4] K. C. Nicolaou, W.-M. Dai, R. K. Guy, Angew. Chem. 1994, 106, 38-69;
Angeir. Chem. l n t . Ed. Engl. 1994, 33. 45; L. Wessjohann, ibid. 1994, 106,
1011-1013 and 1994,33,959-961.
[5] F. Lavelle. Exp. Opin. Invest. Drugs 1995. 4, 771 - 775.
161 D.-S. Su. D. Meng, P. Bertinato. A. Balog, E. J. Sorensen. S . J. Danishefsky,
Y-H. Zheng, T.-C. Chou, L. He, S. B. Horwitz, Angen. Chem. 1997, 109,
775-777; Angew. Chem. Int. Ed. Engl. 1997,36,751-I59
[7] D. Schinzer. Eur. Chem Chron. 1996, 1. 7-10
[8] The name “epothilone” derives from the words epoxide, thiazole and ketone.
191 G. Hofle, personal communication and lecture at the Ludwig-MaximiliansUniversitat Munchen, 19th November, 1996 and 2nd December, 1996.
1101 a ) A. Balog, D. Meng, T. Kamenecka, P. Bertinato, D.-S. Su, E. J. Sorensen,
S . J. Danishefsky, Angen. Chem. 1996,108,2976-2978; Angen.. Chem. Int. Ed.
Engl. 1996, 35. 2801 -2803; b) P Bertinato, E. J. Sorensen, D. Meng, S . J.
Danishefsky. J. Org. Chem. 1996,61,8000-8001; c) D. Meng, E. J. Sorensen,
P. Bertinato. S. J. Danishefsky, ibid. 1996, 6f, 7998-7999. d) Note from the
editor: Another total synthesis of epothilone A, in which the key step is also a
ring-closing metathesis, has recently been published: D. Meng, D.-S. Su. A.
Balog, P. Bertinato, E. J. Sorensen, S . J. Danishefsky, Y. H. Zheng, T. C. Chou,
L. He, S. B. Horwitz, J. Am. Chem. SOC.1997, 119, no. 12.
[ l l ] a) Z. Yang. Y He, D. Vourloumis. H. Vallberg, K C. Nicolaou, Angew. Chem.
1997,109,170-172, Angeu. Chem. Int. Ed. Engl. 1997,36,166-168; b) K. C.
Nlcolaou, Y. He. D. Vourloumis, H Vallberg, 2. Yang, hid. 1996, 108, 25542556 and 1996.35.239-2401
[12] K C. Nicolaou. F. Sarabia, S. Ninkovic, Z. Yang, Angex. Chem. 1997. 109.
539-540; Angeu. Chem. I n t . Ed. Engl. 1997.36, 525-527.
1131 a) D. Schinzer, A. Limberg, A. Bauer, 0. M. Bohm, M. Cordes, Angeu. Chem.
1997, 109. 543-544, A n g r x Chem. Int Ed. Engl. 1997, 36, 523-524; b) D.
Schinzer, A. Limberg, 0. M. Bohm, Chem. Eur. J . 1996, 2. 1477-1482.
[14] J. Mulzer, A. Mantoulidis, Tetrahedron Lett. 1996, 37, 9197-9181.
[15] E. Claus, A Pahl, P. G . Jones, H. M. Meyer, M. Kalesse, Tetrahedron Lett.
1997,38, 1359-1362.
[16] T Gabriel, L. Wessjohann. Tetrahedron Lett. 1997, 38, 1363-1366.
[17] H. G . Schmalz. Angew Chem. 1995. 107, 1981-1984; Angeu. Chem. Int. Ed.
Engl. 1995,34, 1833-1836.
Gas-Phase Complexes: Possible Prereactive Gateways for Reactions of
Halogens with NH,, H,O, and H,S
Hans Burger*
Two molecules that collide in the gas phase in the absence of
a surface have to overcome an activation barrier before they can
react with each other. This requires that either one (or both) of
the partners or a prereactive precursor is activated photochemically into excited electronic states, thermally into higher vibrational or rotational levels, or by an increase of translational
energy (for example by collisions). The formation of weakly
bound complexes with a potential minimum that precedes
the activation barrier may occur by prereactive interactions between the partners. These mark a certain point along the reaction pathway and may also determine the structures of the final
reaction products. Such interactions are characterized by a
separation of the partners that is shorter than the van der Waals
contact, but longer than an electron pair bond or an ionic bond.
Although these prereactive complexes (PRCs) are related to the
better-known van der Waals complexes (VDWCs) such as ArCO
and ArHF, which are formed from a noble gas atom and a
suitable partner, they are distinctly different. While the VDWCs
fall apart reversibly into their constituents [Eq. (la)], the PRCs
may undergo a vigorous reaction to form different products
tEq. (1b)l.
[*] Prof. Dr. H. Burger
FB 9 - Anorganische Chemie
Universitit-GH Wuppertal
D-42097 Wuppertal (Germany)
Fax: Int code + (202)439-2901
e-mail: buergerl i.wrcsl
Verlugsgeseflschufi mhH, 0-69451 Weinheim, 1997
The “bonds” between the constituents of the VDWCs and
PRCs are very weak, and are easily stretched. This is evident
from the very small stretching force constant k (0.036 Ncm-’ in
S 1 7 . 5 0 f .SO10
Angeu. Chem. htt. Ed EngI. 1997, 36, No. 7
Без категории
Размер файла
453 Кб
like, natural, activity, product, taxol, promising, epothilone
Пожаловаться на содержимое документа