Total syntheses of colchicine in comparison: A journey through 50 years of synthetic organic chemistry

Reviews

H.-G. Schmalz and T. Graening

Natural Products Synthesis

Total Syntheses of Colchicine in Comparison: A Journey

through 50 Years of Synthetic Organic Chemistry

Timm Graening and Hans-Günther Schmalz*

Keywords:

Dedicated to Professor Helmut Schwarz

on the occasion of his 60th birthday

alkaloids · cycloaddition · natural

products · synthesis design · total

synthesis

Angewandte

Chemie

3230

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

DOI: 10.1002/anie.200300615

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

Angewandte

Total Synthesis of Colchicine

Chemie

C

olchicine, the major alkaloid of the meadow saffron, is one of the

From the Contents

most prominent natural products and, like other tubulin-binding

natural products (e.g. taxol and the epothilones), exhibits great phar-

maceutical potential. The first syntheses in the late 1950s were mile-

stones in natural product synthesis. But even today this structurally

supposedly simple molecule poses a challenge to synthetic chemists.

Only in the last years have syntheses been developed that are efficient

enough to provide novel structurally modified colchicine analogues.

The comparative examination of all known colchicine total syntheses

undertaken in this Review not only reveals the tremendous progress in

synthetic organic methodology over the past decades, but also shows

how the unique synthetic problems posed by this molecule can be

solved in an exceptionally creative manner. Only a few target mole-

cules have been synthesized in such multifaceted ways.

1. Structure Elucidation and

Biosynthesis

3231

3233

2. The Colchicine/Tubulin

Interaction

3. Target Structure and

Classification of the Total

Syntheses

3234

3235

3253

4. Colchicine Total Syntheses

5. Summary

1. Structure Elucidation and Biosynthesis

isolated the important natural products strychnine, brucine,

and quinine. Several aspects of this unusual alkaloid have

challenged chemists in the past. In fact, the elucidation of the

chemical structure and the biosynthesis of colchicine repre-

sent historic achievements in natural product chemistry.

Despite considerable efforts, the structure remained uncer-

Colchicine (1) is the main alkaloid of the poisonous plant

meadow saffron (Colchicum autumnale L.),^[1]a common plant

of European and North African

origin that flowers in autumn on a

leafless stalk (Figure 1). Despite its

tain for a long time. After investigations by Zeisel

alternative name, autumn crocus, it

belongs to the family Liliaceae.

The name Colchicum is derived

from Colchis, a region east of the

(1883–1913), which led to the development of a method for

the determination of methoxy groups,^[4]Windaus (1910–1924)

made some more specific structural proposals, which were

erroneously based on a phenanthrene ring system (2).^[5]

Black Sea that is known in Greek

mythology as the home of the

Golden Fleece and the notorious

poisoner Medea. Besides their use

as a poison, the active ingredients of the Colchicum species

belong to the oldest known drugs and have been used for

more than 2000 years in the treatment of acute gout.^[2]

Colchicine (1) was first isolated in 1820 by Pelletier and

Caventou,^[3]the same natural scientists who had previously

The key to the elucidation of the colchicine structure was

the correct interpretation of the experimental data concern-

ing ring C of the tricyclic skeleton. Significant progress was

made by Dewar in 1945, who assumed that ring C was a

cycloheptatrienolone with aromatic character (3).^[6a]Together

with his structural proposal for stipitatic acid in the same year,

this hypothesis marked the starting point of the chemistry of

tropolones, the term coined by Dewar for this class of

compounds.^[6b]The correct structure of colchicine was con-

firmed in 1952 by means of X-ray crystal-structure analysis,^[7]

[*] Dipl.-Chem. T. Graening, Prof. Dr. H.-G. Schmalz

Institut für Organische Chemie

Universität zu Köln

Greinstrasse 4, 50939 Köln (Germany)

Fax : (+49)221-470-3064

E-mail: schmalz@uni-koeln.de

Figure 1. Meadow saffron (colchicum autumnale L.).^[1f]

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

DOI: 10.1002/anie.200300615

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

3231

Reviews

H.-G. Schmalz and T. Graening

Scheme 1. Colchicine biosynthesis.

and the absolute configuration was proven by chemical

degradation.^[8]

The cyclization product 10 is then methylated and

dehydrogenated enzymatically to the enamine 12, before a

subsequent skeletal rearrangement takes place to form the

tropolone ring. It is assumed that this process (12!13!14) is

triggered by a cytochrome P450-dependent oxidation. The

final steps involve deformylation to give demecolcine (15),

followed by demethylation and acetylation to give colchicine

(1).

The biosynthesis of colchicine was puzzling in the begin-

ning, because the unusual ring system did not show any clear

relationship to other types of plant alkaloids. Many different

hypotheses were put forward before the groups of Leete and,

in particular, Battersby proposed the currently accepted

biosynthetic route based on a large number of labeling and

incorporation experiments.^[9]This pathway involves a para/

para phenol coupling of autumnaline (9, Scheme 1).

Hans-Günther Schmalz, born in 1957, stud-

ied chemistry at Johann Wolfgang Goethe-

University, Frankfurt (Germany), where he

obtained his Ph.D. in 1985 with Prof. G.

Quinkert. After a postdoctoral stay at Prince-

ton University with Prof. M. F. Semmelhack,

he returned to Frankfurt in 1988 and com-

pleted his Habilitation in 1993. In 1994 he

became Professor at the Technical University

Berlin, and in 1999 he moved to the Univer-

sity of Cologne. His research interests lie in

organic synthesis, from organometallic meth-

odology to the stereoselective total synthesis

of bioactive natural products and analogues.

Timm Graening, born in 1971 in Flensburg

(Germany), studied chemistry at the Chris-

tian-Albrechts-Universität Kiel and joined the

group of Prof. Schmalz at the University of

Cologne in 2000. In his doctoral thesis he is

currently working on the total synthesis of

colchicine.

3232

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

www.angewandte.org

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

Angewandte

Total Synthesis of Colchicine

Chemie

2. The Colchicine/Tubulin Interaction

Besides mechanically stabilizing cellular structures, they serve

as “highways” for transport and signal processes. Most

importantly, however, they form the mitotic spindle during

cell division (Figure 4).

Colchicine (1) is an important bioactive compound used in

the treatment of a broad variety of diseases and still remains

the sole drug for the therapy of acute gout and familial

Mediterranean fever.^[10]Moreover, colchicine acts as an

antimitotic agent by binding to tubulin (Figure 2). Tubulin is

Figure 2. Electron crystallographic structure of the ab-tubulin dimer

with two alternative binding modes for colchicine (A and B) as found

by molecular modeling.^[18]

Figure 4. Fluorescence microscopy image of the spindle apparatus of

a cell during the process of mitosis in the metaphase. The microtu-

bules are marked green and the chromosomes are colored blue.

a globular protein and was formerly known as colchicine-

binding protein.^[11]In the presence of guanosine triphosphate

(GTP), tubulin forms heterodimers consisting of an a and a

b subunit. These a,b-dimers assemble in a helical fashion into

polymeric tubes, the so-called microtubules (Figure 3).^[12]

At this point colchicine exerts its dramatic effect: By

binding to the tubulin dimer, it distorts the tubulin/micro-

tubule equilibrium in such a manner that mitosis is arrested in

the metaphase. In fact, several so-called spindle poisons exert

their influence at this very stage, making this a very successful

approach for cancer chemotherapy. Whereas colchicine and

vinblastine induce depolymerization of microtubules, other

drugs such as taxol and the epothilones effect their stabiliza-

tion.^[13]Therefore such compounds can be used to selectively

damage the rapidly proliferating cancer cells. However, the

general toxicity of colchicine has hampered its use in cancer

chemotherapy so far. A large number of colchicine analogues

have been prepared in the past from the natural product itself,

with the aim of developing new antitumor agents, and some

QSAR (quantitative structure–activity relationship) studies

have been performed.^[14]Nevertheless, the general toxicity

could hardly be suppressed. For example, thiocolchicine

(17)^[15]and demecolcine (15),^[16]

which show a comparable activity

and enjoy medical application,

exhibit a slightly lower toxicity.

As a result of the success of

Figure 3. Model for the assembly of ab-tubulin dimers to microtu-

bules.

taxol in the treatment of mamma-

lian and ovarian cancer, molecules

that interact with the colchicine-

These long protein fibers, which consist of 12 to 13

protofilaments with an alternating linear arrangement of a-

and b-tubulin units, exist in dynamic equilibrium with the

tubulin dimer. Microtubules are polar structures with a faster-

growing (+) and a less-dynamic (ꢀ) terminus. They have an

outer diameter of 24 nm and form a hollow cylinder with a

diameter of approximately 15 nm. The microtubules are

responsible for several important functions within the cell:

binding site of tubulin have

regained pharmaceutical interest. According to the National

Cancer Institute, they are considered as important lead

structures for the development of new antitumor agents.^[17]

So far, a major obstacle in the identification of more-active

colchicine derivatives was the difficulty in accessing colchi-

cine and its structural analogues by means of total synthesis.

Also, the rational design of new colchicine-related drugs is

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

www.angewandte.org

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

3233

Reviews

H.-G. Schmalz and T. Graening

complicated by the fact that the binding site and the

mechanism leading to the destabilization of microtubules

are not yet fully understood.^[18]Unquestionably, colchicine

binds in a temperature-dependent and irreversible manner to

a high-affinity binding site at the b-tubulin subunit, inducing a

partial unfolding at the carboxy terminus.^[19]It is assumed that

this change in the secondary structure affects the regions of

the protein dimer that are essential for microtubule forma-

tion. Different opinions have been put forward concerning

the conformational changes of colchicine itself during its

interaction with the protein.^[20]Nevertheless, it has been

clearly demonstrated that the helical twist inherent to the

colchicine ring skeleton, seems to be of prime importance for

effective binding. The molecule contains a stereocenter at C7

access would be required. A deeper understanding of the

complex phenomena could ultimately lead to a rational

design of new drugs derived from colchicine as a lead

structure.

as well as an aR-configured chiral axis defined by the pivot 3. Target Structure and Classification of the Total

bond joining rings A and C (Figure 5).^[21]Berg et al. were able

Syntheses

In the 1950s, colchicine was considered as a milestone in

natural product synthesis, and the first syntheses from the

laboratories of Eschenmoser and van Tamelen in 1959 were

highly recognized masterpieces of synthetic chemistry,^[24–26]

a

field that was rapidly developing at this time. By today's

standards of synthetic chemistry—highly complex natural

products containing novel and often sensitive structural

elements are synthesized in the laboratory only a few years

after their discovery—the small molecule colchicine (1)

appears to be a rather simple target at first glance.^[27]For

that reason, it is remarkable that its synthesis still poses a

problem, despite all the modern methodology available. Since

the early syntheses of Eschenmoser and co-workers and

van Tamelen and co-workers, new approaches towards col-

chicine have been brought forward continuously, in most

cases following fundamentally different strategies.^[28–37]How-

ever, only in recent years have syntheses been developed that

show a satisfying overall efficiency.^[38–40]The reason for the

difficulties inherent to the target structure does not result

from a lack of general methods for the synthesis of tropolones.

Also, stereocontrol does not pose too much of a problem,

because the configuration of the chiral axis is thermodynami-

cally predetermined by the stereogenic center at C7. The

main difficulty actually lies in the regiose-

Figure 5. Stereostructure of colchicine.

to separate the atropisomers of desacetamidocolchicine (aR-

18, aS-18) and were able to demonstrate that only the aR

enantiomer binds to tubulin.^[22]

lective construction of the highly oxidized

ring C within the unusual 6,7,7-membered

ring system 20.

The tropolone ring C is deeply embed-

ded within the molecular architecture by

the neighborhood of the stereocenter, the

direct connection to the benzenoid ring,

and its annulation to the seven-membered

ring B. Therefore, its generation forms an

integral part of any synthetic strategy. During many inves-

tigations directed towards the total synthesis of colchicine, it

became clear that in this context standard methods for

tropolone formation either gave bad yields or were compli-

cated by regioselectivity problems. Also the introduction of

tropolone derivatives as ring C building blocks has hitherto

been disadvantageous.^{[30,32,33,41]}Thus, specific methodology

for the construction of annulated tropolones is a precondition

for any efficient total synthesis.

During the search for new drugs, a series of compounds

were discovered by chance which also interact with the

colchicine-binding site of tubulin. Some of them exhibit a

great pharmacological potential. Combretastatin A4 (19), for

instance, is a highly promising antitumor agent which recently

passed phase I clinical trials.^[23]

To study the interaction of colchicine analogues with

tubulin in more detail, an efficient and flexible total synthetic

3234

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

www.angewandte.org

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

Angewandte

Total Synthesis of Colchicine

Chemie

The chemistry of the parent tropolone and simple tropone

and tropolone derivatives has been studied extensively since

the 1950s.^[42]Tropolones exist in a tautomeric equilibrium and

behave similarly to phenols with respect to aromatic sub-

stitution. Their acidity (pK_a= 7) is inbetween those of phenols

and carboxylic acids. The carbonyl group shows little ketone

character and should rather be considered as part of a

vinylogous carboxylic acid. The aromatic character of tropo-

lone, which was a matter of intense debate in the past, should

not be overemphasized.

The free tropolone derived from 1 is named colchiceine. It

is a prominent colchicine precursor, which exists in a

tautomeric equilibrium (21/22) and can be easily converted

into colchicine by treatment with diazomethane (Scheme 2).

Scheme 3. Important colchicine intermediates and starting materials.

proceed with only low to moderate yields. A second major

group encompasses strategies that exploit five- or six-

membered cyclic precursors for the construction of the crucial

tropolone ring C (ABC’!ABC, AB’C’!ABC, AC’!ABC).

A common feature of most of these syntheses is the

preparation of the seven-membered ring by a ring enlarge-

ment of a norcarane precursor. Following this concept, some

fascinating transformations for the construction of ring C in a

highly oxidized form, or even as the complete tropolone ring,

have been designed. The greatest retrosynthetic simplifica-

tion of the fundamental ortho-condensed ring system corre-

sponds to an intramolecular cycloaddition strategy (A!

ABC). Ideally, both seven-membered rings B and C are

formed simultaneously with the correct placement of the

oxygen functions in ring C.

Scheme 2. Interconversion of colchiceine (22), colchicine (1), and iso-

colchicine (23).

The different strategies developed for the total synthesis

of 1 in the course of the past 50 years are presented in the

following sections, with an emphasis on the overall synthetic

scheme.^[44]The order of the syntheses is not strictly chrono-

logical. Instead, the conceptually different synthetic routes

are organized according to the progress that they contributed

to the solution of key synthetic problems.

Actually, a 1:1 mixture of colchicine (1) and isocolchicine (23)

is formed, which can be separated by chromatography. The

undesired isomer can then be recycled by cleavage of the

methyl ether upon heating with dilute hydrochloric acid.^[43]

Besides colchiceine (21/22), other important intermediate

molecules that have been reached in formal total syntheses

are desacetamidocolchiceine (24) and the advanced colchi- 4. Colchicine Total Syntheses

cine precursor trimethylcolchicinic acid (25), both of which

had been converted into the target compound 1 during the

early syntheses (Scheme 3). As a readily available starting

material, purpurogallin (26) was used in several syntheses,

and the majority of the other syntheses started from a

symmetric C5-substituted pyrogallol trimethyl ether of type

27.

To give an overview of the various synthetic approaches

towards colchicine, including formal total syntheses, it is

advantageous to classify them according to the cyclization

strategies used in the construction of the 6,7,7-membered ring

system (Scheme 4).

4.1. The Scott Synthesis of Desacetamidocolchiceine through

Oxidative Coupling

The synthesis reported by Scott et al. in 1965 started from

a tropolone derivative as the ring C building block and

followed the strategy A + C!AC!ABC (Scheme 5).^[30]This

synthesis is based on an earlier biosynthetic hypothesis

postulating the formation of ring B by oxidative phenol–tro-

polone coupling.^[45]

The tropolone part was derived from purpurogallin (26)

from which the anhydride 29 was synthesized by oxidative

degradation of the electron-rich benzene ring followed by

dehydration. By condensation of 29 with the aldehyde 32,

prepared from the acid chloride 30 in a Rosenmund

reduction, the lactone rac-33 was obtained under elimination

Monotopic cyclization strategies involving the formation

of strategic bonds in ring B or C (AC!ABC or AB!ABC,

respectively) make up the majority of the early syntheses. In

these systems, however, the crucial ring-closing reactions

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

www.angewandte.org

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

3235

Reviews

H.-G. Schmalz and T. Graening

Scheme 4. Retrosynthetic classification of colchicine syntheses.

Scheme 5. Synthesis of desacetamidocolchiceine (24) by Scott et al.

Reagents and conditions: a) H₂O₂, OH^ꢀ, 90–958C; b) H₂SO₄; c) (COCl)₂,

room temperature, 18 h; d) Pd/BaSO₄, quinoline, 118–1208C, 5 h; e) neat

29 and 32, <1008C, 30 min; f) Cu-bronze, 190–2008C, 15 min; g) H₂, Pd/

C, EtOAc; h) HBr, reflux, 30 min; i) FeCl₃·6H₂O, 6n H₂SO₄, EtOH, CHCl₃,

room temperature, 72 h; j) MeI, K₂CO₃, acetone, reflux, 10 h; k) H₂SO₄

(aq.), <1008C, 1 h. The empty brackets indicate a “missing” yield, which

was either not determined at this stage or not published at all.

3236

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

www.angewandte.org

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

Angewandte

Total Synthesis of Colchicine

Chemie

of CO₂. After elimination/decarboxylation and subsequent

Azo coupling of 35 and subsequent hydrogenation

ꢀ

hydrogenation of the C C double bond in 34 (which proceeds

afforded the aminotropolone 43, which after diazotization

directly gave rise to desacetamidocolchiceine (24) in a

Pschorr cyclization. Unfortunately the yield did not exceed

5%.

smoothly in the presence of the tropolone moiety), the

secocolchicine intermediate 35 was obtained in good yield. To

prepare for the projected key coupling reaction, the methyl

ether groups were cleaved to give the free pyrogallol

derivative 36. Various reaction conditions for the crucial

coupling (36!37) were tested, and it was found that the

reaction product was overoxidized under the initially chosen

conditions. Eventually, a two-phase oxidation system consist-

ing of FeCl₃in H₂SO₄(6n) and ethanol in chloroform was

used, and the tricyclic product 37 was isolated in 2% yield

after preparative paper chromatography under an inert

atmosphere. After methylation, the obtained mixture of 18

and 38 was hydrolyzed to desacetamidocolchiceine (24),

which had been previously converted into (ꢁ )-colchicine

(rac-1) by Eschenmoser and co-workers (see Section 4.2.).^[25]

Despite the straightforward strategy, the success of the

Scott synthesis is diminished by the extremely low yield of the

key step. However, the problems associated with this putative

biomimetic transformation are in agreement with the fact that

the colchicine biosynthesis does not involve a phenol/

tropolone coupling, as was recognized later.

4.1.2. The Kato Variant of the Scott Synthesis

An alternative access to the seco intermediate 35 was

reported by Kato et al.^[33]For the synthesis of the tropolone

ring, these authors used the solvolysis of the cyclopentadiene/

dichloroketene adduct rac-49 following the Stevens and

Bartlett protocol for tropolone synthesis (Scheme 7).^[46]

4.1.1. The Kaneko Variant of the Scott Synthesis

Kaneko et al. addressed the improvement of the key step

in the Scott synthesis. Except for alternative preparation of

the aldehyde 32 through decarboxylative Claisen rearrange-

ment and ozonolysis, the seco intermediate 35 was prepared

by following the scheme of Scott et al. (Scheme 6).^[32]

Scheme 7. Synthesis of secodesacetamidocolchiceine (35) by Kato

et al. Reagents and conditions: a) H₂(1 atm), Pd/C, MeOH, room

temperature; b) LiAlH₄, Et₂O, 08C, 4 h, !RT, 14 h; c) TsCl, pyridine,

08C, 2 h; d) NaH, cyclopentadiene, THF, 08C, then 45, 3 h; e) 48,

NEt₃, n-hexane, 5 h; f) KOAc, H₂O, HOAc, reflux, 2 days. Ts=para-

toluenesulfonyl.

The cyclopentadiene precursor 47 was prepared

from the tosylate 45 by nucleophilic substitution with

sodium cyclopentadienyl (46) and concomitant

double-bond isomerization. Attempts by Kato et al.

to increase the yield of the Pschorr reaction described

by Kaneko (43!24) were not successful.

4.2. The Pioneering Eschenmoser Synthesis

The first successful total synthesis of colchicine

(1), a seminal work by Eschenmoser and co-workers,

was reported in 1959.^[25]As in the synthesis by Scott

et al. discussed before, purpurogallin (26) was used as

starting material, but in a completely different

strategy, that is, as an AB building block

(Scheme 8). Therefore, the chosen strategy required

Scheme 6. Synthesis of desacetamidocolchiceine (24) by Kaneko et al. Reagents

and conditions: a) N,N-dimethylaniline, reflux, 10 h; b) Me₂SO₄, aqueous

NaOH, MeOH, reflux, 3 h; c) O₃, EtOAc, 30–40 min; then H₂, Pd/C, 258C;

d) reference [30]; e) NaNO₂, HCl, H₂O, p-toluidine, 08C (!41); then 35,

aqueous NaOH, 2–108C, 50 min; f) H₂, Pd/C, MeOH, room temperature, 1.5 h;

g) isoamylnitrite, conc. H₂SO₄, dioxane, 7–108C; then Cu powder, dioxane,

room temperature, 24 h.

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

www.angewandte.org

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

3237

Reviews

H.-G. Schmalz and T. Graening

Scheme 8. Synthesis of desacetamidocolchiceine (24) by Eschenmoser and co-workers. Reagents and conditions: a) Me₂SO₄, NaOH, H₂O, 08C,

3 h!RT, 7 h; b) H₂, Pd, THF, 458C, 7.5 h; c) LiAlH₄, Et₂O, 08C, 2.5 h; d) H₃PO₄, EtOH, 60–708C, 1 h 15 min; e) 53, NEt₃, benzene, tert-amylalco-

hol, reflux, 45 min; f) MeI, K₂CO₃, acetone, room temperature, 18 h; g) 55, 1758C, 1 h 40 min; h) H₂SO₄, MeOH, reflux, 40 min; then CH₂N₂;

i) KOtBu, tBuOH, benzene, room temperature, 2 h 45 min; j) NaOH, H₂O, MeOH, reflux, 30 min; k) OsO₄, pyridine, Et₂O, room temperature,

10 h; then KClO₃, NaHCO₃, MeOH, 1008C, 2.5 h; l) NaOH, H₂O, reflux, 30 min; m) powdered quartz glass, 260–2708C, 9 min; n) TsCl, pyridine,

room temperature, 14 h; o) NH₃, EtOH, 90–958C, 15 h; p) aqueous KOH, EtOH, 1308C, 22 h.

the development of a suitable method for the annulation of

the tropolone ring C.

The tropolone ring of trimethylpurpurogallin 50 was

reduced, and the product rac-51 was used to prepare the

dimethoxybenzosuberone 52. The free phenol functionality

proved to be of importance as several attempts to alkylate the

corresponding trimethoxybenzosuberone had failed. As

shown in Scheme 9, the free phenol function first undergoes

a Michael addition to the electrophilic propyne ester 53 (52!

68). In a second (intramolecular) Michael addition, the

tricyclic ketoester rac-69 is formed. Subsequent b elimination

of the phenol group and cyclization then leads to the isolated

product 71.

The pyrone 54 obtained after methylation of 71 was then

heated with chloromethylmaleic anhydride (55) to give the

Diels–Alder adduct rac-56 (Scheme 8), which served as

starting point for the following key transformation, that is, a

Scheme 9. Intermediates involved in the preparation of the a-pyrone

54 from 52 according to Eschenmoser.

3238

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

www.angewandte.org

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

Angewandte

Total Synthesis of Colchicine

Chemie

norcaradiene–tropilidene valence tautomerization. Starting

from the diester rac-57, treatment with base led to regiose-

lective formation of the norcaradiene derivative rac-58, which

immediately furnished the tropilidene diester 59 by electro-

cyclic ring opening. The sterically less-hindered ester group

was selectively hydrolyzed, before oxidation with OsO₄

afforded the tropolone 61 under decarboxylation. Further

decarboxylation gave the product 62, a regioisomer of the

targeted tropolone. The required conversion of 62 into

desacetamidocolchiceine (24) oxygenated at C9 and C10,

relied on the findings of Nozoe et al.^[47]Initially, an isomeric

mixture of tosylates 63 and 64 was prepared. Ammonolysis

then resulted in a tele-substitution, which again afforded a

mixture from which the aminotropone 67 was separated and

finally converted into desacetamidocolchiceine (24) by sap-

onification. This compound was utilized as a relay product, as

it could be prepared in four steps from natural colchicine

according to a private communication of R. B. Woodward to

Albert Eschenmoser (even though at the time Woodward was

a competitor in the race for the first synthesis of colchicine).

Methylation of 24 led to a mixture of the regioisomers 18

and 38 (Scheme 10). From these regioisomers, only the

isocolchicide 38 was a suitable substrate for radical bromina-

tion at C7 because of the better stabilization of the radical

intermediate. Ammonolysis of the bromination product rac-

73 afforded the diamino compound rac-74, but only in low

yield. After remethylation of the hydrolysis product rac-25

and chromatographic separation of the resulting regioisom-

ers, acetylation of rac-16 finally gave racemic colchicine (rac-

1), which had previously been separated into the enantiom-

ers.^[48]

The Eschenmoser synthesis opened a highly efficient

access to the colchicine ring scaffold with an unsaturated

ring C (59). In spite of that, the inherent difficulties posed by

the target structure became apparent: both the regioselective

placement of the oxygen functionalities at C9 and C10 of the

tropolone ring and the introduction of the acetamido group at

C7 required many additional steps and proceeded with rather

low yield and selectivity.

4.3. The van Tamelen Synthesis

Also in 1959, van Tamelen et al. reported a total synthesis

of colchicine which in part followed the same pathway as the

strategy Eschenmoser.^[26]However, it differs with respect to

the central step for the construction of ring C (Scheme 11).

The synthesis started with the preparation of the ketone

77 from tetramethylpurpurogallin (76) in analogy to the

synthesis discussed above. The alkylation of the reluctant

substrate trimethoxybenzosuberone 77 could successfully be

realized by using acrylamide (78) as a Michael acceptor. In a

Reformatsky reaction, the Michael adduct rac-79 was trans-

formed into the diastereomeric products rac-81 and rac-82,

which were separated by chromatography and fractional

crystallization. These isomers were separately transformed

Scheme 10. Completion of the synthesis of (ꢁ)-colchicine (rac-1) by Eschenmoser et al. utilizing the relay compound 24. Reagents and conditions:

q) CH₂N₂, Et₂O, MeOH, 08C, 1 h; r) NBS, benzoyl peroxide, CCl₄, hn, reflux, 18 min; s) NH₃, EtOH, H₂O, 95–1008C, 15 h; t) aqueous KOH,

EtOH, 1308C, 23 h; u) CH₂N₂, Et₂O, CH₂Cl₂, MeOH, 08C, 30 min; v) Ac₂O, pyridine, 1008C, 15 min. NBS=N-bromosuccinimide.

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

www.angewandte.org

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

3239

Reviews

H.-G. Schmalz and T. Graening

Scheme 11. Synthesis of (ꢁ)-trimethylcolchicinic acid (rac-25) by van Tamelen et al. utilizing the relay compound 24. Reagents and conditions:

a) LiAlH₄; b) Zn, H₂SO₄; c) H₂, Pd/C; d) 78, KOtBu, MeCN, tBuOH, room temperature, 9 h; e) Zn, I₂, methyl bromoacetate, benzene, reflux, 3 h;

f) KOH, MeOH/H₂O, reflux, 21 h; g) DCC, pyridine, room temperature, 2 days; h) CH₂N₂, MeOH, Et₂O; i) Na, xylene, reflux, 4 h; j) Na, liquid

NH₃, 20 min; k) Cu(OAc)₂, MeOH, reflux, 3 h; l) TsOH·H₂O, benzene, 5.5 h; m) NBS, CHCl₃, reflux, 1 h, 15 min; n) CH₂N₂, MeOH; o) NBS, CCl₄,

hn, reflux, 23 min, !RT, 1 h; p) NaN₃, MeOH, 90–958C, 27 h; q) H₂(1 atm), Pd/C, MeOH, room temperature; r) HCl (aqueous, 1n), reflux, 2 h.

into the corresponding lactone esters rac-83 and rac-84. In this

case, the lactone acts as an internal protecting group. During

the subsequent acyloin condensation, deprotonation of the

hydroxy group had to be suppressed to avoid base catalysis of

the competing Dieckmann condensation. Under classical

conditions for the acyloin condensation, both lactones

afforded exclusively the cyclohexenone rac-85 derived from

the Dieckmann condensation product. With sodium in liquid

ammonia, only the minor lactone isomer rac-84 was con-

verted, to give the condensation products rac-86 and rac-87 in

9% yield.

afforded desacetamidocolchiceine (24). Again, the colchi-

ceine derivative 24 was used as a relay compound and was

brominated at C7 by the method of Eschenmoser to give rac-

73. In contrast to the direct ammonolysis, the introduction of

the amino function via the azide rac-89 proceeded in good

yield. Hydrogenation of the azide function and cleavage of

the tropolone methyl ether were also carried out in excellent

yield to give trimethylcolchicinic acid (rac-25).

Although the synthesis addressed the problem of selective

oxygenation at C9 and C10, the construction of ring C, on the

other hand, could only be accomplished in very low yields

both in the key reaction (acyloin condensation) and the

subsequent oxidation steps.

This mixture of hemiketals was oxidized to give rac-88 and

an acid-catalyzed opening of the oxa bridge followed by

further oxidation of the resulting dihydrotropolone with NBS

3240

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

www.angewandte.org

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

Angewandte

Total Synthesis of Colchicine

Chemie

4.4. The Roussel–Uclaf Synthesis

could be obtained under Brønsted acidic conditions. Both

compounds were converted into the cyanoester 98 in satisfy-

ing yields, either in a six-step sequence from 94 or in nine

steps from 99. The use of the cyanoester rather than the

corresponding diester was of advantage because it guaranteed

the formation of the Dieckmann condensation product in a

regioselective manner. The product of this key reaction was

isolated as the enolbenzoate 105 in only 17% yield. The free

enolate anion derived from 105 was oxygenated at the

a position and subsequently transformed into the dihydro-

tropolone 107, which had already been used as an intermedi-

The Roussel–Uclaf synthesis of desacetamidocolchiceine

(24) (Scheme 12) reported in 1965 by Toromanoff and co-

workers strategically resembles the approach of van Tamelen

and co-workers, but brought about improvements with

respect to the critical points mentioned above.^[31]Ring B

was condensed to the aromatic ring A in a Friedel–Crafts-

type cyclization (A!AB!ABC).

The synthesis started with the preparation of the ketoester

93 from which either of the Friedel–Crafts products 94 or 99

Scheme 12. Roussel–Uclaf synthesis of desacetamidocolchiceine (24). Reagents and conditions: a) 91, Na, benzene, reflux, 6 h, !RT, 14 h;

b) NaOH, MeOH/H₂O; then aqueous HCl; c) CH₂N₂, CH₂Cl₂; d) TsOH, benzene, reflux, 16 h; e) N-methylformanilide, POCl₃, 55–608C, 4 h;

f) ethyl cyanoacetate, piperidine, HOAc, benzene, reflux, 20 h; g) H₂(1 atm), PtO₂, EtOH, room temperature, 1 h; h) KOH, MeOH, room tempera-

ture, 17 h; i) 180–2008C; j) CH₂N₂, CH₂Cl₂; k) 85% H₃PO₄, 508C, 16 h; l) Na, NH₃, THF, ꢀ708C, 1 h 35 min; m) NaOEt, EtOH, isoamylnitrite,

08C, 1 h 20 min; then HOAc; n) pyruvic acid, HOAc/H₂O, <1008C, 2 h; o) C₂H₂, Na, liquid NH₃, ꢀ708C, 2.5 h; then addition of rac-101, THF,

ꢀ408C, 5.5 h; p) H₂1 atm, Pd/CaCO₃, pyridine, DMF, room temperature; q) CH₂N₂; r) PBr₃, CHCl₃/ligroin, ꢀ20!ꢀ108C, 5 h; s) CuCN, DMSO,

508C, 25 h; t) 2,4-dinitrobenzenesulfonic acid, benzene, reflux, 2 h; u) K, toluene, reflux, 2 h; then BzCl, pyridine, 108C, !RT, 2 h; v) KOH,

MeOH/H₂O, room temperature, 16 h; w) Na, benzene, reflux, 15 h; then benzoyl peroxide, room temperature, 14 h; x) aqueous NaHCO₃, EtOH/

H₂O, reflux, 10 min; y) NBS, reference [26]. DMF=N,N-dimethylformamide; DMSO=dimethyl sulfoxide; Bz=Benzoyl.

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

www.angewandte.org

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

3241

Reviews

H.-G. Schmalz and T. Graening

ate by van Tamelen et al. The oxidation of 107 to desaceta-

midocolchiceine (24) with NBS completed the formal total

synthesis of 1.

All in all, the selective construction of the C9, C10

oxygenated tropolone was improved. Nevertheless, selective

functionalization at C-7 had still not been implemented.

As shown in Scheme 13, pyrogallol mono(methyl ether)

108 was first treated with the suberone rac-109 in a Pechmann

condensation to give the coumarin 110. The seven-membered

ring later serves as a precursor for ring C. The coumarin also

contains a carbonyl function, which is later required for the

cyclization of ring B. The free phenol functionality was

transformed into an allyl ether (111) and subsequent Claisen

rearrangement (!112) allowed the selective introduction of a

side chain for the projected construction of ring B. The

propanoic acid side chain was established in five subsequent

steps. Hydrolysis of the coumarin 116 afforded the dicarbox-

ylic acid 117. Methylation provided the diester 118, which

proved to be a better starting material than 117 for the crucial

cyclization step (decarboxylative Dieckmann condensation).

The product of this cyclization (119) already contains the

complete 6,7,7-membered carbocyclic scaffold. To introduce

the nitrogen function at C7, a Leuckart–Wallach reaction was

used to give the formamide rac-120 with a concomitant shift

4.5. The Sankyo Synthesis

The problem of the selective functionalization at C7 was

addressed in the Sankyo synthesis of colchiceine (22) reported

by Nakamura et al. in 1962 (Schemes 13 and 14).^[28]In this

synthesis, the amino group was introduced prior to the

oxidation of ring C to a tropolone. The strategy was actually

based on previous works of Boekelheide et al.^[24c]and focused

on the functionalization of C7 by formation of the strategic

ꢀ

C6 C7 bond in ring B (AC!ABC).

Scheme 13. Sankyo synthesis of the colchicine intermediate rac-122. Reagents and conditions: a) MeSO₃H, room temperature, 20 h; b) allyl bro-

mide, K₂CO₃, NaI, MeOH/H₂O, reflux, 16 h; c) N,N-dimethylaniline, reflux, 6 h; d) KOH, MeOH, 110–1208C, 6 h; e) O₃, CH₂Cl₂, ꢀ608C, 1 h;

f) malonic acid, pyridine, aniline, 508C, 22 h; g) H₂, Pd/C, MeOH, 1 h; h) 1808C, 1 mmHg, 8 h; i) Me₂SO₄, KOH, <1008C, 3 h, !RT, 14 h;

j) CH₂N₂, MeOH/Et₂O, room temperature, 1 h; k) KOH, MeOH/H₂O, reflux, 2 h; l) KOAc, Ac₂O, reflux, 2 h, !1388C, 20 h; m) KOtBu, tBuOH,

xylene, 1398C, 20 h; n) HCO₂H, CO(NH₂)₂, 110!1808C over 6 h; o) NH₂OH·HCl, pyridine, EtOH, reflux, 20 h; p) LiAlH₄, THF, reflux, 6 h;

q) Ac₂O, pyridine, room temperature, 12 h.

3242

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

www.angewandte.org

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

Angewandte

Total Synthesis of Colchicine

Chemie

Scheme 15. Tropone synthesis via a ditropyl ether.

ceineamide rac-126 could be isolated in pure form (again in

very low yield). Saponification of the aminotropone finally

furnished (ꢁ )-colchiceine (rac-22).

4.5.1. Synthesis of an Advanced Colchicine Intermediate by

Wenkert et al.

Wenkert et al. reported an alternative synthesis of the key

intermediates 119 and rac-122, which had already occurred in

the Sankyo synthesis.^[37]The Wenkert approach differs from

the Sankyo synthesis mainly in the way that the Dieckmann

condensation precursor, that is, the diester rac-137, is

prepared (Scheme 16). For the construction of the seven-

membered ring C, a divinylcyclopropane rearrangement was

applied.

The hydrocinnamic acid 131 was used as starting material.

The precursor rac-135 required for the cyclization of ring C

was prepared by a Wittig reaction from rac-134 and the

phosphorane obtained from 133. The subsequent [3,3]

sigmatropic rearrangement proceeded in high yield, and the

Scheme 14. Completion of the Sankyo synthesis of (ꢁ)-colchiceine

(rac-22). Reagents and conditions: r) BF₃·OEt₂, benzene, room temper-

ature, 60 h; s) NBS, CCl₄, reflux, 5 min; then collidine 160–1808C, 2 h;

t) PCl₅, CCl₄, room temperature, 24 h; u) aqueous NaOH, 08C; then

conc. HCl, room temperature, 3 days; v) N₂H₄·H₂O, EtOH, reflux,

7.5 h; w) aqueous NaOH

ꢀ

of the C C double bond. Alternatively, the corresponding

ꢀ

oxime 121 was reduced and the resulting amine subsequently

converted into the acetamide rac-122.

diester rac-137 was obtained after a shift of the remote C C

double bond in rac-136 and subsequent esterification. Further

conversion into the ketone 139 by Dieckmann condensation

and demethoxycarbonylation was achieved in a yield com-

parable to that of the analogous transformation described by

Nakamura (32%). The diastereomeric oximes 140 and 141

obtained from the ketone 139 could be separated, and the

major Z isomer 141 was transformed into the acetamide rac-

122 by catalytic hydrogenation with concomitant N-acetyla-

tion. As a second point of overlap with the Sankyo synthesis,

the key intermediate 119 was reached by catalytic hydro-

genation of 139. Reduction of the corresponding oxime

mixture (142) with lithium aluminum hydride followed by

acetylation also provided rac-122. At this stage the synthetic

study was concluded, constituting a formal total synthesis of

racemic colchicine.

The synthesis was completed by using either synthetic rac-

123, which was available by BF₃·OEt₂-mediated double-bond

isomerization of rac-122 (Scheme 14), or enantiomerically

pure material obtained from natural colchicine as a relay

compound. The cycloheptene rac-123 was converted into the

tropilidene rac-124 in a bromination/dehydrobromination

sequence. As the direct oxidation of rac-124 to the tropone

rac-125 was not viable with the various methods tried, a fairly

wasteful deviation had to be made. The Nozoe method for

tropone synthesis via a ditropyl ether that was applied is

illustrated in Scheme 15.^[49]First, an intermediate tropylium

salt 128 was generated by hydride abstraction from cyclo-

heptatriene (127). Reaction with NaOH then gave a ditropyl

ether 129, which upon treatment with concentrated HCl

disproportionated to afford a mixture of the starting tropili-

dene (127) and the oxidation product, tropone (130).

The corresponding reaction of rac-124 afforded the

desired tropone rac-125, albeit in a disappointing low yield

(Scheme 14). One can suspect that not only the desired

colchicide rac-125 but also regioisomeric products were

formed in this step. Without further purification, crude rac-

125 was treated with hydrazine, presumably again with the

formation of a mixture of regioisomers, from which colchi-

Both the synthesis of Wenkert (Scheme 16) and the

Sankyo synthesis (Schemes 13 and 14) provided an improved

solution for the problem of C7 functionalization. While

focusing on this problem, regioselectivity in the construction

of the tropolone ring C was strategically neglected, and the

overall efficiency of the synthetic schemes was spoiled by

substantial losses of material during the completely unselec-

tive introduction of functionality in ring C following the

strategy of Nakamura et al.

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

www.angewandte.org

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

3243

Reviews

H.-G. Schmalz and T. Graening

Scheme 16. Synthesis of colchicine intermediates 119 and rac-122 by Wenkert et al. Reagents and conditions: a) Me₂SO₄, K₂CO₃, acetone, reflux,

2 h; b) MeOCH₂Cl, CHCl₃/HOAc, 208C 18 h; c) PPh₃, benzene, reflux, 24 h; d) LDA, THF, ꢀ108C, !208C, 3 h, then rac-134, THF, ꢀ108C, !RT,

3 h; e) xylene, reflux, 24 h; f) NaOMe, MeOH, reflux, 12 h; then CH₂N₂, Et₂O; g) NaOMe, xylene, reflux, 12 h; h) 5% HCl, HOAc, H₂O, dioxane,

reflux, 4 h; i) NH₂OH·HCl, pyridine, EtOH, reflux, 2 h; j) H₂, Pd/C, Ac₂O; k) H₂(3 atm), Pd/C, EtOAc, room temperature, 2 h; l) LiAlH₄, THF,

reflux, 12 h; m) Ac₂O, pyridine, room temperature, 16 h. LDA=lithium diisopropylamide.

4.6. Synthesis of Desacetamidocolchiceine through [4+3]

of the ketal function (Scheme 17). A one-pot procedure was

used to increase the yield to 70% (rac-146!148). a Func-

tionalization of 148 with hydrazine according to Nozoe et al.

afforded the regioisomeric aminotropones 149 and 67, which

could be separately converted into the regioisomeric tropo-

lones 62 and 24 (desacetamidocolchiceine), respectively, by

saponification.

The sequence elaborated by Boger (54!24) greatly

abbreviates the Eschenmoser synthesis. Although the route

was very efficient, the unselective a oxygenation of the

tropone intermediate 148 still remained a strategic drawback.

Cycloaddition by Boger and Brotherton

In contrast to the approaches discussed above, Boger and

Brotherton developed a very direct entry to annulated

tropone structures by means of a cycloaddition strategy

ABC’!ABC (Scheme 17).^[36]

The cycloaddition of the Eschenmoser a-pyrone 54 and

the cyclopropenone ketal 143 gave either the Diels–Alder

product rac-144 (which could not be converted into the

desired tropone) or the [4+3] cycloaddition product rac-146,

depending on the reaction conditions chosen. Upon heating,

the cyclopropenone ketal 143 is in thermal equilibrium with a

reactive open chain species, which according to Boger is best

described as a nucleophilic delocalized vinyl carbene (150/

151; Scheme 18).

4.7. Synthesis of Desacetamidoisocolchicine by Skeletal

Rearrangement According to Tobinaga and Co-workers

The [4+3] cycloaddition product 146 could easily be

transformed into the tropone 148 in good yield through a

retro-Diels–Alder decarboxylation and subsequent hydrolysis

A rather striking strategy, completely different to all

previous approaches, was described by Tobinaga and co-

workers (Scheme 19).^[34]Their synthesis of desacetamidoiso-

3244

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

www.angewandte.org

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

Angewandte

Total Synthesis of Colchicine

Chemie

Scheme 18. Thermal ring opening of the

cyclopropene ketal 143.

Scheme 17. Synthesis of desacetamidocolchiceine (24) by Boger and Brotherton. Reagents and conditions: a) 143, 6.2 kbar, 258C, 5 days, b) mesi-

tylene, 2008C, 50 min; c) 143, benzene, 758C, 36 h; d) 2008C, 2 min; e) HOAc, H₂O, THF, 258C, 5 min; f) HOAc, H₂O, THF, 1008C, 3.5 h;

g) N₂H₄·H₂O, EtOH, 0!258C, 4.5 h; h) KOH(aqueous, 2 n), EtOH, 100–1108C, 21 h.

Scheme 19. Synthesis of desacetamidoisocolchicine (38) by Tobinaga and co-workers Reagents and conditions: a) condensation, no conditions

given; b) catalytic hydrogenation, no conditions given; c) anodic oxidation, HBF₄, MeCN, 20 min; d) NaBH₄; e) CH₂I₂, Zn/Cu couple;

f) Jones oxidation; g) Ac₂O, H₂SO₄, room temperature.

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

www.angewandte.org

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

3245

Reviews

H.-G. Schmalz and T. Graening

colchicine (38) along the lines of a rearrangement strategy

rac-168 under milder conditions. The rearrangement pro-

ceeded in analogy to the transformation designed by Tobi-

naga et al., with the difference that the presence of the

additional ester function resulted in the formation of 169 as

well as the expected product rac-170. The mixture of these

isomeric products was treated with DDQ to give a 3:7 mixture

of the heptafulvene 171 and the desired tropolone methyl

ether rac-172. These two oxidation products were separated,

and it was found that they can be readily equilibrated back to

the initial 3:7 mixture. Hydrolysis of the ester and decarbox-

ylation of rac-172 gave rise to the Eschenmoser intermediate

38. For the conclusion of the formal total synthesis, the

isomeric mixture 171/rac-172 was converted in a three-step

sequence into the advanced colchicine precursor rac-25 by

hydrolysis of the ester, Curtius degradation, and demethyla-

tion.

In conclusion, the synthesis of Evans et al. provided an

outstanding solution for the construction of the unusual 6,7,7-

membered ring system and included the selective function-

alization of both rings C and B. The unsatisfactory yield for

the generation of the tropolone remained a shortcoming in

this synthesis. Also, the development of an enantioselective

variant seemed extremely difficult owing to the facile

equilibration of 172 and its isomer 171.

AC’!AB’C’!ABC is based, like the Scott synthesis, on a

false hypothesis concerning the biosynthesis of colchicine.^[50]

The synthesis started from the simple aromatic com-

pounds 152 and 153. The chalcone 154 prepared from these

simple building blocks was hydrogenated and the intermedi-

ate 155 was cyclized in an electrochemical oxidation to the

spirocyclic dienone rac-156. Reduction with sodium borohy-

dride afforded a mixture of the diastereomeric alcohols rac-

157 and rac-158, which was transformed into the cyclo-

propane derivative rac-159 by Simmons–Smith cyclopropa-

nation and subsequent Jones oxidation. Although the stereo-

structure of rac-159 remained unknown, treatment with an

Ac₂O/H₂SO₄mixture directly gave rise to 38. Evidently, the

opening of the cyclopropane ring in rac-159 was accompanied

by a migration of the aryl residue followed by dehydration to

give the tropolone methyl ether 38 (migration of the alkyl

residue would have resulted in the formation of a regioiso-

meric tropolone).

This single operation served to unfold the complete 6,7,7-

membered ring skeleton establishing the tropolone ring C

with the desired oxidation state. Furthermore, the desired

regioisomer of the tropolone methyl ether required for the C7

functionalization according to Eschenmoser was generated

selectively. Nevertheless, both the synthesis of Boger and

Brotherton and the synthesis of Tobinaga and co-workers are

inefficient, as they rely on this low-yielding and non-stereo-

selective introduction of the amido functionality.

4.9. The Woodward Synthesis

One of the most innovative studies in the field of

colchicine synthesis was presented by Woodward in 1963 as

a Harvey Lecture.^[29]A remarkable feature of his synthesis

was the implementation of the nitrogen functionality at C7 as

an integral part of the synthetic strategy. The crucial role in

the very straightforward synthetic concept was played by the

isothiazole ring, a heterocycle first described in the course of

this synthesis. It served as a masked amino function and acted

as platform for the cyclization of rings B and C (Scheme 21).

The disubstituted isothiazole 174 was obtained by treat-

ment of the b-aminocrotonate 173 with thiophosgene. Photo-

bromination and reaction with triphenylphosphane afforded

the phosphonium salt 175, which was used for the Wittig

olefination of trimethoxybenzaldehyde 152. After reduction

4.8. The Evans Synthesis of (ꢁ)-Trimethylcolchicinic Acid

Inspired by the interesting work of Tobinaga and co-

workers, Evans et al. reported a synthesis of the Eschenmoser

intermediate 38 in 1978 in which they further investigated the

rearrangement transformation.^[35a]Three years later, the

Evans group came up with a very convincing total synthesis

of trimethylcolchicinic acid following a AC’!ABC strategy in

which the problems of the construction of the tropolone ring

and the introduction of C7 functionality were equally well-

addressed (Scheme 20).^[35b]

The main innovation was that the precursor rac-167 for

the key transformation, that is, the skeletal rearrangement,

only contains rings A and C’. Intermediate rac-167 was

prepared in a convergent fashion and in high yield from the

cyclohexenone derivative rac-163 and the arylbutanoate 165.

The synthesis of rac-163 started with the electrochemical

oxidation of trimethoxybenzene 160 to give 161, which was

selectively hydrolyzed to the quinone monoketal 162. Sub-

sequent methylenation with dimethylsulphoxonium methyl-

ide afforded rac-163. The arylbutanoate 165 in turn was

obtained by chain elongation of the trimethoxycinnamate 44.

Coupling of the ester enolate derived from 165 with the

building block rac-163 afforded a mixture of diastereomers

(rac-166), which was directly hydrolyzed to give the cycliza-

tion precursor rac-167.

ꢀ

of the C C double bond with diimine (the isothiazole sulphur

atom precluded the use of catalytic hydrogenation) and

conversion of the ester into an aldehyde, the resulting

intermediate 177 was subjected to a second Wittig reaction

with the phosphorane 178. Saponification and isomerization

of the Z/E mixture of isomers afforded the E,E-dienecarbox-

ylic acid 179 as a pure stereoisomer. Cyclization to form

ring B was achieved in a Friedel–Crafts-type alkylation, and

subsequent reduction with diimine afforded rac-181. In the

following step, the heterocycle was lithiated at the unsub-

stituted position and carboxylated with CO₂. The correspond-

ing diester rac-182 then served as a precursor for the

cyclization of ring C in a Dieckmann condensation. Subse-

quent decarboxylation gave the tetracyclic ketone rac-183.

C10 was functionalized by Woodward's method for the

introduction of a 1,3-dithiane. For this purpose, the position

The subsequent key transformation was performed with

very high yield in a one-pot procedure or, alternatively, via

3246

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

www.angewandte.org

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

Angewandte

Total Synthesis of Colchicine

Chemie

Scheme 20. Synthesis of (ꢁ)-trimethylcolchicinic acid (rac-25) by Evans et al. Reagents and conditions: a) KOH, MeOH, anodic oxidation, 08C,

25 h; b) (CO₂H)₂, THF/H₂O, room temperature, 12 min; c) Me₃S(O)I, NaH, DMSO, 08C, 15 min, !RT, 30 min, then 162, DMSO, room tempera-

ture, 2 h; d) CH₂N₂, Pd(OAc)₂, Et₂O/CH₂Cl₂, 08C, 15 min; e) H₂(51 psi), Pd/C, MeOH; f) 165, LDA, THF, ꢀ788C, 35 min, then rac-163, !08C,

30 min; g) aqueous (CO₂H)₂, THF, room temperature, 5 h; h) CF₃CO₂H, room temperature, 20 min; i) CF₃CO₂H, reflux, 35 min; j) DDQ, benzene,

reflux, 18 h; k) CDCl₃, room temperature, 24 h, equilibration to a 3:7 mixture of 171/rac-172; l) NaH, MeOH/H₂O, 758C, 45 min;

m) (PhO)₂P(O)N₃, NEt₃, tBuOH, reflux, 17 h; n) HCl (aqueous, 3n), 1108C, 90 min; o) 1808C, 1–2 min. DDQ=2,3-dichloro-5,6-dicyano-1,4-ben-

zoquinone.

a to the carbonyl group of rac-183 was first activated by

formylation (!rac-184) before reaction with the bisthiotosy-

late 185 gave rise to the dithiane rac-186. Hydrolysis of rac-

186 afforded the a-diketone rac-187, which was converted

into the dienol diacetate rac-188. Remarkably, a spontaneous

aerial oxidation under basic conditions yielded the desired

tropolone 189. To liberate the amino function, desulphuriza-

tion was performed with Raney nickel, and (ꢁ )-colchiceine

(rac-22) was obtained after acetylation.

Although this synthesis is 40 years old, it demonstrates

how selectivity problems can be overcome in a masterly

fashion with conceptual creativity. Unfortunately, Woodward

did not report yields in his publication, so that the overall

efficiency and any possible obstacles cannot be judged.

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

www.angewandte.org

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

3247

Reviews

H.-G. Schmalz and T. Graening

Scheme 21. Synthesis of (ꢁ)-colchiceine (rac-22) by Woodward et al. Reagents and conditions: a) CSCl₂, NEt₃, Et₂O; b) NBS, CCl₄, hn; c) PPh₃;

d) NaOMe; then 152; e) N₂H₄, H₂O₂, Cu^II; f) LiAlH₄; g) MnO₂; h) Wittig olefination; i) NaOHaq.; j) I ₂, hn; k) 70% HClO₄, 608C; l) N₂H₄, H₂O₂,

Cu^II; m) o-lithiobiphenyl, THF; then CO₂; n) esterification; o) NaH, dioxane; p) HOAc, H₂SO₄, heating; q) NaH, HCO₂Et, EtOH, THF; r) 185,

KOAc, iPrOH, reflux; s) Hg(OAc)₂, aqueous HClO₄, HOAc, heating; t) Ac₂O, pyridine; u) NaOH, air; v) Raney Ni, aqueous NaOH; w) Ac₂O, pyri-

dine; then aqueous NaOH.

4.10. The Cha Synthesis of (ꢀ)-Colchicine by Oxyallyl [4+3]

the second building block. Compound 196 was first converted

into the borane adduct to assure regioselective lithiation and

to avoid cleavage of the heterocycle. The product obtained

from the reaction of the intermediate 197 with the aldehyde

193 was directly oxidized under Swern conditions, and the

resulting ketone 198 was enantioselectively reduced to the

alcohol 200 with the (S)-diphenylvalinol-derived oxazabor-

olidine 199 following the Itsuno protocol. The transformation

to 201 was carried out through Mitsunobu reaction with

DPPA as an azide source, Staudinger reduction to the amine,

and N-acetylation. Upon heating of 201, the furan 202 was

formed in an intramolecular Diels–Alder/retro-Diels–Alder

reaction cascade under extrusion of HCN. Whereas the

following key step, the cycloaddition of 202 with an oxyallyl

Cycloaddition

In a much more recent synthesis, Cha et al. also made

intensive use of heterocyclic chemistry. The annulated furan

202 was used as the central intermediate for a cycloaddition,

corresponding to the general ABC’!ABC strategy

(Scheme 22).^[39]

The starting material, alcohol 190, was first protected,

then iodinated, coupled with TMS-acetylene in a Sonogashira

coupling, and finally desilylated (deprotected) to give 192. A

subsequent Swern oxidation afforded the aldehyde 193 in

high overall yield. Oxazole (196), which was prepared by

condensation of TosMIC (195) with formaldehyde, served as

3248

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

www.angewandte.org

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

Angewandte

Total Synthesis of Colchicine

Chemie

Scheme 22. Synthesis of (ꢀ)-colchicine (1) by Cha and co-workers. Reagents and conditions: a) TIPSOTf, 2,6-lutidine, CH₂Cl₂, 08C, !RT, 5 h;

ꢂ

b) I₂, Ag(CF₃CO₂), NaHCO₃, CHCl₃, 3 h; c) HC CTMS, [Pd(PPh₃)₂Cl₂], CuI, Et₂NH, DMSO, 908C, 20 h; d) TBAF, THF, 08C, !RT, 7 h; e) (COCl)₂,

DMSO, CH₂Cl₂, ꢀ788C, 1 h 15 min; then NEt₃, ꢀ78!08C; f) (CH₂O)_n, 195, K₂CO₃, DMSO, room temperature, 5 h; then KOH, ethylene glycol,

room temperature, 5 h; g) BH₃·THF, THF, room temperature, 1 h; then tBuLi, ꢀ788C, 40 min; h) THF, ꢀ788C, 3 h; i) BH₃·THF, (S)-(ꢀ)-diphenyl-

valinol, THF, ꢀ788C, 20 min, !RT; then 198, THF, 6 h; j) (PhO)₂P(O)N₃, PPh₃, DIAD, THF, room temperature, 12 h; k) PPh₃, H₂O, THF, room

temperature, 50 h; l) Ac₂O, NEt₃, DMAP, Et₂O, room temperature, 7 h; m) o-dichlorobenzene, reflux, 66 h; n) Boc₂O, NEt₃, DMAP, CH₂Cl₂, 358C,

10 h; o) LiOH, THF, 12 h; p) 203, TMSOTf, EtNO₂, ꢀ788C, 1 h, !ꢀ608C, 12 h; q) TMSOTf, NEt₃, CH₂Cl₂, 08C, 2 h; r) HCl (1n), Et₂O, room tem-

perature, 1 h; s) Ac₂O, NEt₃, DMAP, Et₂O, 258C, 10 h. TIPS=triisopropylsilyl; Tf=trifluoromethanesulfonyl; TMS=trimethylsilyl; TBAF=tetra-n-

butylammonium fluoride; Boc=tert-butoxycarbonyl; DIAD=diisopropylazodicarboxylate; DMAP=4-dimethylaminopyridine.

generated from 203, afforded the undesired regioisomer 204,

the analogous [4+3] cycloaddition of the Boc-protected

compound 205 proceeded with the reversed regioselectivity.

However, the desired product 206 was obtained in a mere

23% yield.

As an explanation for the observed selectivity, the authors

assumed a participation of a hydrogen bond in the case of 202

as a substrate (structure 208 in Scheme 23). By changing the

N-protecting group to Boc (202!205), such a hydrogen bond

does not seem to play a role and steric interactions dominate

the selectivity (structure 209).

The use of related oxa-bridged [4+3]-cycloadducts was

introduced by Föhlisch et al. as a generally applicable method

for tropone synthesis.^[51]The oxygen bridge of 206 was

removed by treatment with TMSOTf/NEt₃to furnish the

tropolone 207 with the correctly placed methyl ether func-

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

www.angewandte.org

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

3249

Reviews

H.-G. Schmalz and T. Graening

Cha and co-workers also reported another method for the

preparation of the advanced furan intermediate 202 in which

they exploited the chiral pool (Scheme 24).^[39b]The protected

l-serine derivative 210 was converted into the iodide 211, and

its organozinc derivative was treated with the acid chloride

212 in a Pd-catalyzed coupling. The resulting aryl ketone 213

was converted into 214 by benzylic hydrogenation and

subsequent iodination at the arene nucleus. A Sonogashira

coupling of the acetamide 215 followed by desilylation and

reduction with DIBAH afforded the aldehyde 217. The

oxazole 218, prepared from the aldehyde with TosMIC (195),

was heated at reflux in ortho-dichlorobenzene to yield the

same furan 202 previously obtained from the isomeric oxazole

201.

Scheme 23. Influence of the amide function on the regio- and stereose-

lectivity of the cycloadditions to give 204 and 206.

4.11. The Banwell Synthesis of (ꢀ)-Colchicine by Biomimetic Ring

Enlargement

tionality (Scheme 22). After N-deprotection and renewed

acetylation, colchicine (1) was obtained (ca. 90% ee).

Thus nonracemic colchicine was synthesized in a rela-

tively short reaction sequence. The strength of this synthesis

lies in the highly efficient preparation of the key intermediate

202. However, the overall scheme is stained by the low yield

of the key cycloaddition (205!206, 23%, 45% with respect

to 50% recovered starting material).

Another efficient enantioselective synthesis, which is

strategically related to the syntheses of Tobinaga and co-

workers and Evans et al. and follows the AC’!ABC’!ABC

ring enlargement strategy, was reported by Banwell in 1996

(Scheme 25).^[38]This synthesis was the culmination of com-

prehensive efforts by Banwell et al.^[38,41,52]and was also guided

by the biosynthesis of the natural product, which according to

our current knowledge (Scheme 1) does, indeed, proceed

Scheme 24. Alternative route to the colchicine intermediate 202 according to Cha and co-workers. Reagents and conditions: a) MsCl, no further

details given; b) NaI, no further details given; c) 211, Zn/Cu couple, benzene, ultrasound, room temperature, 3 h; then [Pd(PPh₃)₂Cl₂], 212, ultra-

sound, 2 h; d) H₂(60 psi), Pd/C, EtOH, H₂O, 22 h; e) I₂, Ag(CF₃CO₂), NaHCO₃, CHCl₃, 08C, 2 h; f) CF₃CO₂H , EtO, room temperature, 3 h; then

2

ꢂ

Ac₂O, NEt₃, DMAP, Et₂O, 10 h; g) HC CTMS, [Pd(PPh₃)₂Cl₂], CuI, Et₂NH, DMSO, 908C, 15 h; h) K₂CO₃, MeOH, room temperature, 11 h;

i) DIBAH, CH₂Cl₂, ꢀ788C, 1 h; j) TosMIC, K₂CO₃, MeOH, reflux, 1 h; then KOH, reflux, 3 h; k) o-dichlorobenzene, reflux, 40 h. Ms=methanesul-

fonyl; DIBAH=diisobutylaluminum hydride; TosMIC=toluenesulfonylmethylisocyanide.

3250

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

www.angewandte.org

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

Angewandte

Total Synthesis of Colchicine

Chemie

Scheme 25. Banwell synthesis of (ꢀ)-colchicine (1). Reagents and conditions: a) NaOH, MeOH, room temperature, 48 h; b) H₂, Pd/C, EtOAc,

158C, 10 h; c) NaBH₄, THF/MeOH, 158C, 1.5 h; d) Pb(OAc)₄, molecular sieves (3 ), CH₂Cl₂, 158C, 1 h; e) CF₃CO₂H, molecular sieves (3 ),

THF/benzene, 08C, 1 h; f) BnBr, K₂CO₃, MeCN, 828C, 4 h; g) NMO, TPAP, molecular sieves (4 ), CH₂Cl₂, 158C, 43 h; h) 224, THF, 158C, 6 h;

i) H₂, Pd/C, EtOAc, 158C, 9 h; j) Tl(NO₃)₃, MeOH, ꢀ208C, 30 min; k) Me₃S(O)I, NaH, DMSO, 158C, 7 h; l) CF₃CO₂H , CHCl₂, 158C, 3 h;

2

m) Zn(N₃)₂·Py₂, DIAD, PPh₃, THF, 158C, 38 h; n) PPh₃, H₂O, THF, 158C, 63 h; o) Ac₂O, pyridine, 158C, 15 min. NMO=N-methylmorpholine-N-

oxide; TPAP=tetra-n-propylammonium perruthenate.

through ring enlargement of a norcarane intermediate as a

precursor for ring C.

reaction. The product 230 was obtained in only low yield, and

partial racemization occurred during the subsequent Stau-

dinger reduction. From the resulting amine, the natural

product 1 was obtained with only 81% ee after acetylation.

In summary, the synthetic problem was solved by Banwell

in a relatively short sequence of rather simple transforma-

tions. Even the rather low yields of some steps do not distort

the beautiful overall scheme.

Starting from the easily obtainable chalcone 220, the

ꢀ

stepwise reduction of the C C double bond and the keto

group under cleavage of the benzyl ether afforded the racemic

diaryl propanol rac-221. Wessely oxidation gave rise to the

cyclohexadienone rac-222, which was converted into the aryl

ketone 223 by cationic cyclization, benzylation of the result-

ing phenol, and oxidation of the alcohol. Enantioselective

reduction of the ketone 223 with stoichiometric quantities of

the CBS reagent 224 afforded 225 (94% ee). Cleavage of the

benzyl ether was followed by a renewed oxidation of the

phenol, in this case with Tl(NO₃)₃according to the procedure

of Taylor and McKillop. Methylenation of 226 with dime-

thylsulphoxonium methylide gave the cyclopropane 227 in

only moderate yield. The ring enlargement to establish the

6,7,7-membered ring system (!229) was achieved by treat-

ment of 227 with CF₃CO₂H. Along these lines, the tropolone

ring C was delivered with the correct oxidation state and

arrangement of the oxygen functions as found in the target

structure. Unexpectedly, the S_N2-type conversion of the OH

function of 229 into a nitrogen functionality was rather

difficult. At first an azido group was introduced by Mitsunobu

4.12. Our Synthesis of an Advanced Colchicine Intermediate

Exploiting a Rh-Catalyzed Cyclization Cascade

Our conceptually very direct approach towards colchicin

(1) is based on a retrosynthetic analysis in which the oxa-

bridged key intermediate 238 is traced back to the open-chain

cyclization-precursor 237 (A!ABC).^[40,53]This strategy was

designed in analogy to the benzotropolone synthesis of

Friedrichsen and Plüg.^[54]The crucial step, a Rh-catalyzed

cyclization cascade (237!238), not only generated the 6,7,7-

membered ring system, but also installed the oxygen sub-

stituents correctly at C9 and C10 in the newly formed ring C

(Scheme 26).

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

www.angewandte.org

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

3251

Reviews

H.-G. Schmalz and T. Graening

Scheme 26. Synthesis of the advanced colchicine intermediates 241 and 245 according to Graening and Schmalz. Reagents and conditions:

ꢂ

a) (COCl)₂, CH₂Cl₂, 08C!RT, 12 h; then NHMeOMe·HCl, pyridine, CH₂Cl₂, 08C, 1 h; b) Ag(CF₃CO₂), I₂, CHCl₃, room temperature, 4 h; c) LiC C-

TMS, THF, ꢀ788C, !ꢀ108C, 45 min; d) 233 (1 mol%), iPrOH, 16 h; e) TBSCl, imidazole, DMF, room temperature, 18 h; f) iPrMgCl, THF,

ꢀ258C, 4.5 h; then 235, THF, ꢀ408C, 30 min; then K₂CO₃, MeOH, room temperature, 3 h; g) iBuOCOCl, NEt₃, THF/Et₂O, ꢀ20!ꢀ158C; then

CH₂N₂, Et₂O, ꢀ58C, 16 h; h) [Rh₂(OAc)₄], toluene, reflux, 4 h; i) L-selectride, THF, ꢀ788C, 3 h; j) TMSOTf, NEt₃, CH₂Cl₂, ꢀ50!ꢀ108C, 1 h; then

K₂CO₃, MeOH, 08C, 45 min; k) DMSO, (CF₃CO)₂O, CH₂Cl₂, ꢀ608C, 1.5 h; then NEt₃, ꢀ608C, 1.5 h; l) Et₂AlCl, CH₂Cl₂, ꢀ78!08C, 2.5 h;

m) NH₂N₂·H₂O, EtOH, 08C!RT, 4.5 h; n) TBAF, THF, 08C, 45 min; o) 2n KOH, EtOH, 1108C, 16 h. TBS=tert-butyldimethylsilyl.

The bifunctional Weinreb amide 231 was prepared from

hydrocinnamic acid 131 and treated with lithium trimethylsi-

lylacetylide to give the alkynone 232. This compound proved

to be an excellent substrate for enantioselective transfer

hydrogenation according to Noyori. In the presence of the Ru

complex 233 as a catalyst, the corresponding propargylic

alcohol was obtained in essentially enantiomerically pure

form and was subsequently silylated to give the protected

derivative 234 in 93% yield (two steps). Acylation to give the

ketocarboxylic acid 236 was achieved utilizing a protocol of

Knochel for iodine–magnesium exchange and reaction of the

resulting Grignard intermediate with succinic anhydride

(235). An advantage of this method was that deprotonation

at the propargylic position, which was observed with the

stronger base nBuLi, could be inhibited. Treatment of the a-

diazoketone 237 (obtained from 236) with catalytic

[Rh₂(OAc)₄] in refluxing toluene gave the tetracyclic key

intermediate 238 with complete diastereoselectivity.

3252

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

www.angewandte.org

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

Angewandte

Total Synthesis of Colchicine

Chemie

Scheme 27. Rh-Catalyzed cyclization cascade to the key intermediate 238.

The mechanism and scope of related transformations

introduced by Ibata^[55a]were investigated in detail by Padwa

et al.^[55b]As shown in Scheme 27 the cyclization cascade is

initiated by the generation of an electrophilic rhodium

carbene 247, which is then attacked by the adjacent Lewis

basic carbonyl group to form a reactive carbonyl ylide 248.

The ylide is trapped in another intramolecular reaction step, a

1,3-dipolar cycloaddition with the tethered alkyne unit.

The cycloaddition product 238 formed in 64% yield has a

formation of the two regioisomeric aminotropones 243 and

244, which were easily separable. From these, the two

regioisomeric tropolones 245 and 246, respectively, were

accessible by desilylation and alkaline hydrolysis. As 241 and

245 can be converted into the Banwell intermediate 229, their

formation completes a formal total synthesis of colchicine (1).

Furthermore, the alternative route via the tropone 242 opens

an easy entry to the pseudo-colchicine series starting from

246.

ꢀ

rather rigid ring system in which the benzylic C O bond is

With a 12% overall yield for the conversion of the simple

starting material 131 into the advanced colchicine precursor

241, our synthesis represents a highly efficient approach

towards colchicinoid structures (Scheme 26). The strategy

should be flexible enough to allow the synthesis of modified

analogues. However, this has to be demonstrated in the

future.

aligned with the p system of the aromatic ring. Therefore, a

selective cleavage of the oxa bridge should be feasible by

exploiting stereoelectronic effects.

In fact, the alcohol 239, prepared from 238 by reduction

with L-selectride, is readily converted into the dienediol 240

upon treatment with TMSOTf/NEt₃

and immediate hydrolysis of the

primarily formed bis(silyl ether) (to

avoid isomerization by a sigmatropic 5. Summary

[1,5] H-shift). Oxidation of 240 with

DMSO/(CF₃CO₂)₂O then immedi-

ately gave rise to the stable tropolone

241. An alternative route for the

conversion of the key intermediate

Over the past 50 years, colchicine has challenged synthetic

chemists like only few other target molecules. The synthetic

efforts of many research groups have culminated in fascinat-

ing and very diverse synthetic solutions (Table 1). These

studies led to the development of novel tropolone syntheses

and served to scrutinize the methodology for the synthesis of

seven-membered rings in general. Although the early syn-

theses represent cultural achievements of highest merit, their

238 into a tropolone started with its

Lewis acid mediated dehydration to

the tropone 242. Following a similar protocol as applied by

Boger, treatment of 242 with hydrazine resulted in the

Table 1: Summary of colchicine syntheses.^[a]

Authors

Year

Starting compound

Target

Number of steps Yield [%]

Eschenmoser and co-workers 1959

(ꢁ)-colchicine (rac-1)

22

0.00006

0.0005

van Tamelen et al.

Nakamura et al.

Woodward

1959

1962

1963

1965

Trimethylcolchicinic acid (rac-25) 18

(ꢁ)-colchiceine (rac-21)

desacetamidocolchiceine (24)

20

23

11

0.0007^[b]

[c]

–

Scott et al.

0.1

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

www.angewandte.org

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

3253

Reviews

H.-G. Schmalz and T. Graening

Table 1: (Continued)

Authors

Year

Starting compound

Target

Number of steps Yield [%]

Toromanoff and co-workers

1965

desacetamidocolchiceine (24)

15^[d]

11

9

0.3^[d]

0.5^[e]

Kaneko et al.

Kato et al.

1968

1974

desacetamidocolchiceine (24)

~1

Tobinaga and co-workers

1974

7

24

5

Evans et al.

1984

1985

trimethylcolchicinic acid (rac-25) 12

Boger and Brotherton

desacetamidocolchiceine (24)

10^[f]

3.5^[f]

Wenkert et al.

1989

10

6.3

Banwell

1996

1998

(ꢀ)-colchicine (1) ~81% ee

(ꢀ)-colchicine (1) ~90% ee

15

18

0.9

1.9

Cha and co-workers

Graening and Schmalz

2004

11

12

[a] The overall yields and the number of steps, as a matter of interpretation, should only be taken for the purposes of comparison. [b] For the racemic

series. [c] No yields were reported. [d] For the shorter and more efficient route via 94. [e] Based on the yields for 26!35 reported by Scott et al.

[f] Based on the preparation of pyrone 54 reported by Eschenmoser and co-workers.

overall efficiency, of course, cannot meet modern standards.

As the comparative analysis of the many syntheses reveals,

solutions for the problem have only slowly evolved and

colchicine must still be considered as a difficult target

molecule. Notwithstanding the powerful arsenal of modern

synthetic technology, even today synthetic chemists require a

high level of creativity and the ability to integrate subtle

structural and reactivity aspects to tame such a recalcitrant

target structure.

come within reach of the efficiency and flexibility levels

required for such purposes. Consequently, it can be expected

that colchicine will remain an important synthetic target and

challenge further creativity in synthetic chemistry.

Received: June 13, 2003 [A615]

Published Online: April 19, 2004

[1] For reviews of the chemistry of colchicine, see: a) J. W. Cook,

J. D. Loudon in The Alkaloids, Vol. 2 (Eds.: R. H. F. Manske,

H. L. Holmes), Academic Press, New York, 1952, p. 261;

b) W. C. Wildman in The Alkaloids, Vol. 6 (Ed.: R. H. F.

Manske), Academic Press, New York, 1960, p. 247; c) A.

Brossi in The Alkaloids, Vol. 23 (Ed.: A. Brossi), Academic

Press, Orlando, 1984, p. 1; d) W. C. Wildman in The Chemisry of

the Alkaloids (Ed.: S. W. Pelletier), Van Nostrand-Reinhold,

Like other tubulin-binding molecules such as taxol and

the epothilones, colchicine enjoys a great interest from

medicinal research and has experienced a renaissance in the

last few years. To explore its pharmaceutical potential further

as a lead structure, flexible synthetic schemes must be

developed that allow a diversity-oriented synthesis of struc-

tural analogues of colchicine. Only the most recent syntheses

3254

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

www.angewandte.org

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

Angewandte

Total Synthesis of Colchicine

Chemie

New York, 1970, chap. 8; e) O. Boyꢀ, A. Brossi in The Alkaloids,

Vol. 41 (Eds.: A. Brossi, G. A. Cordell), Academic Press, San

Diego, 1992, p. 125; f) The picture of the meadow saffron was

taken from: B. Besler, Hortus Eystettensis, Eichstꢁtt, 1613..

[2] a) C. Bryan, The Papyrus Ebers, Appleton, New York, 1931; b) P.

Dioscorides, De Materia Medica, 78 A.D.

[3] P. J. Pelletier, J. B. Caventou, Ann. Chim. Phys. 1820, 14, 69.

[4] a) S. Zeisel, Monatsh. Chem. 1883, 4, 162; b) S. Zeisel, Monatsh.

Chem. 1886, 7, 557; c) S. Zeisel, K. R. von Stockert, Monatsh.

Chem. 1913, 34, 1339.

[22] U. Berg, J. Deinum, P. Lincoln, J. Kvassman, Bioorg. Chem. 1991,

19, 53.

[23] a) D. J. Chaplin, G. R. Pettit, S. A. Hill, Anticancer Res. 1999, 19,

189; b) M. C. Bibby, Drugs Future 2002, 27, 475.

[24] For earlier work in colchicine synthesis, see: a) H. Rapoport,

A. R. Williams, M. E. Cisney, J. Am. Chem. Soc. 1951, 73, 1414;

b) J. W. Cook, J. Jack, J. D. Loudon, G. L. Buchanan, J.

MacMillan, J. Chem. Soc. 1951, 1397; c) V. Boekelheide, F. C.

Pennington, J. Am. Chem. Soc. 1952, 74, 1558; d) H. J. E.

Loewenthal, J. Chem. Soc. 1953, 3962; e) H. J. E. Loewenthal,

J. Chem. Soc. 1958, 1367; f) C. D. Gutsche, F. A. Fleming,J. Am.

Chem. Soc. 1954, 76, 1771; g) H. J. E. Loewenthal, P. Rana, Proc.

Chem. Soc. London 1958, 114.

[5] A. Windaus, Justus Liebigs Ann. Chem. 1924, 439, 59.

[6] a) M. J. S. Dewar, Nature 1945, 155, 141; b) M. J. S. Dewar,

Nature 1945, 155, 50.

[25] a) J. Schreiber, W. Leimgruber, M. Pesaro, P. Schudel, A.

Eschenmoser, Angew. Chem. 1959, 71, 637; b) J. Schreiber, W.

Leimgruber, M. Pesaro, P. Schudel, A. Eschenmoser, Helv.

Chim. Acta 1961, 44, 540.

[26] a) E. E. van Tamelen, T. A. Spencer, Jr., D. S. Allen, Jr., R. L.

Ovis, J. Am. Chem. Soc. 1959, 81, 6341; b) E. E. van Tamelen,

T. A. Spencer, Jr., D. S. Allen, Jr., R. L. Ovis, Tetrahedron 1961,

8.

[27] K. C. Nicolaou, S. A. Snyder, Classics in Total Synthesis II: More

Targets, Strategies, Methods, Wiley-VCH, 2003.

[28] a) T. Nakamura, Y. Murase, R. Hayashi, Y. Endo, Chem. Pharm.

Bull. 1962, 10, 281; b) G. Sunugawa, T. Nakamura, J. Nakazawa,

Chem. Pharm. Bull. 1962, 10, 291; c) T. Nakamura, Chem.

Pharm. Bull. 1962, 10, 299.

[29] R. B. Woodward, The Harvey Lecture Series 1963, 59, 31.

[30] a) A. I. Scott, F. McCapra, R. L. Buchanan, A. C. Day, D. W.

Young, Tetrahedron 1965, 21, 3605.

[31] J. Martel, E. Toromanoff, C. Huynh, J. Org. Chem. 1965, 30,

1752.

[32] S.-i. Kaneko, M. Matsui, Agric. Biol. Chem. 1968, 32, 995.

[33] M. Kato, F. Kido, M.-D. Wu, A. Yoshikoshi, Bull. Chem. Soc.

Jpn. 1974, 47, 1516.

[34] E. Kotani, F. Miyazaki, S. Tobinaga, J. Chem. Soc. Chem.

Commun. 1974, 300.

[35] a) D. A. Evans, D. J. Hart, P. M. Koelsch, J. Am. Chem. Soc.

1978, 100, 4593; b) D. A. Evans, S. P. Tanis, D. J. Hart, J. Am.

Chem. Soc. 1981, 103, 5813.

[36] a) D. L. Boger, C. E. Brotherton, J. Org. Chem. 1985, 50, 3425;

b) D. L. Boger, C. E. Brotherton, J. Am. Chem. Soc. 1986, 108,

6713.

[37] E. Wenkert, H.-S. Kim in Studies in Natural Products Chemistry

(Part B), Vol. 3 (Ed.: Atta-ur-Rahman), Elsevier, Amsterdam,

1989, p. 287.

[38] M. G. Banwell, Pure Appl. Chem. 1996, 68, 539.

[39] a) J. C. Lee, S.-j. Jin, J. Cha, J. Org. Chem. 1998, 63, 2804; b) J. C.

Lee, S.-j. Jin, J. Cha, Tetrahedron 2000, 56, 10175.

[40] a) T. Graening, W. Friedrichsen, J. Lex, H.-G. Schmalz, Angew.

Chem. 2002, 114, 1594; Angew. Chem. Int. Ed. 2002, 41, 1524;

b) T. Graening, H.-G. Schmalz, unpublished results.

[41] For the synthesis of secocolchicine, see: M. G. Banwell, J. N.

Lambert, G. L. Gravatt, J. Chem. Soc. Perkin Trans. 1 1993, 2817.

[42] For reviews on tropolone chemistry, see: a) P. L. Pauson, Chem.

Rev. 1955, 55, 9; b) T. Nozoe in Non-Benzenoid Aromatic

Compounds (Ed.: D. Ginsburg), Interscience, New York, 1959,

p. 339; c) D. Lloyd, Carbocyclic Non-Benzenoid Aromatic

Compounds, Elsevier, Amsterdam, 1966; d) T. Nozoe, Pure

Appl. Chem. 1971, 28, 239; e) F. Pietra,Chem. Rev. 1973, 73, 293;

f) F. Pietra, Acc. Chem. Res. 1979, 12, 132; see also: g) M. G.

Banwell, Aust. J. Chem. 1991, 44, 1.

[7] M. V. King, J. L. De Vries, R. Pepinsky, Acta Crystallogr. Sect. B

1952, 5, 437.

[8] H. Corrodi, E. Hardegger, Helv. Chim. Acta 1955, 38, 2030.

[9] a) E. Leete, P. E. Nemeth, J. Am. Chem. Soc. 1961, 83, 2192;

b) E. Leete, Tetrahedron Lett. 1965, 333; c) E. McDonald, R.

Ramage, R. N. Woodhouse, E. W. Underhill, L. R. Wetter, A. R.

Battersby, J. Chem. Soc. Perkin Trans. 1 1998, 2979; c) A. C.

Barker, D. R. Julian, R. Ramage, R. N. Woodhouse, G. Hardy, E.

McDonald, A. R. Battersby, J. Chem. Soc. Perkin Trans. 1 1998,

2989; d) R. N. Woodhouse, E. McDonald, R. Ramage, A. R.

Battersby, J. Chem. Soc. Perkin Trans. 1 1998, 2995; e) P. W.

Sheldrake, K. E. Suckling, R. N. Woodhouse, A. J. Murtagh,

R. B. Herbert, A. C. Barker, J. Staunton, A. R. Battersby, J.

Chem. Soc. Perkin Trans. 1 1998, 3003 and earlier works in this

journal.

[10] C. Le Hello in The Alkaloids, Vol. 53 (Ed.: G. A. Cordell),

Academic Press, San Diego, 2000, chap. 5.

[11] For a structure determination of the ab-tubulin dimer by

electron crystallography, see: E. Nogales, S. G. Wolf, K. H.

Downing, Nature 1998, 391, 199.

[12] For an overview of the structure and dynamics of microtubules,

see: a) P. Dustin, Microtubules, 2nd ed., Springer, Berlin, 1984;

b) R. H. Wade, A. A. Hyman, Curr. Opin. Cell Biol. 1997, 9, 12;

c) S. Guha, B. Bhattacharyya, Curr. Sci. 1997, 73, 351.

[13] For an overview of tubulin as a target for anticancer drugs and

agents that interact with the mitotic spindle, see: A. Jordan, J. A.

Hadfield, N. J. Lawrence, A. T. McGown, Med. Res. Rev. 1998,

18, 259.

[14] S.-X. Zhang, J. Feng, S.-C. Kuo, A. Brossi, E. Hamel, A. Tropsha,

K.-H. Lee, J. Med. Chem. 2000, 43, 167, and references therein.

During these investigations, active compounds were designed,

which are, however, structurally unrelated to colchicine.

[15] a) L. Velluz, G. Müller, Bull. Soc. Chim. Fr. 1954, 755; b) L.

Velluz, G. Müller, Bull. Soc. Chim. Fr. 1954, 1072.

[16] a) A. C. Banerjee, B. Bhattacharyya, FEBS Lett. 1979, 99, 333;

b) A. S. Sherpinskaya, V. I. Gelfand, B. P. Kopnin, Biochim.

Biophys. Acta 1981, 673, 86.

[17] See, for example: L. M. Leoni, E. Hamel, D. Genini, H. Shih,

C. J. Carrera, H. B. Cottam, D. A. Carson, J. Natl. Cancer Inst.

2000, 92, 217.

[18] R. Bai, D. G. Covell, X.-F. Pei, J. B. Ewelli, N. Y. Nguyeni, A.

Brossi, E. Hamel, J. Biol. Chem. 2000, 275, 40443, and references

therein.

[19] J. M. Andreu, S. N. Timasheff, Biochemistry 1982, 21, 6465.

[20] a) H. W. Detrich III, R. C. Williams, Jr., T. L. Macdonald, L.

Wilson, D. Puett, Biochemistry 1981, 20, 5999; b) B. Bhattachar-

yya, J. Wolff, J. Biol. Chem. 1984, 259, 11836; c) R. P. Rava, S. B.

Hastie, J. C. Myslik, J. Am. Chem. Soc. 1987, 109, 2202; d) W. A.

Donaldson, Tetrahedron 1988, 44, 7409.

[43] M. Sorkin, Helv. Chim. Acta 1946, 29, 246.

[21] Until the correction in 1999, the stereochemical descriptor of the

chiral axis had been erroneously assigned as aS: a) U. Berg, H.

Bladh, Helv. Chim. Acta 1999, 82, 323; b) K.-H. Lee, A. Brossi,

Helv. Chim. Acta 1999, 82, 1223.

[44] For graphical compilations of colchicine syntheses, see: a) N.

Anand, J. S. Bindra, S. Ranganathan, Art in Organic Synthesis,

Holden-Day, San Francisco, 1970, p. 103; b) S. Ito in Natural

Products Chemistry, Vol. 2 (Eds.: K. Nakanishi, T. Goto, S. Ito, S.

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

www.angewandte.org

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

3255

Reviews

H.-G. Schmalz and T. Graening

Natori, S. Nozoe), Kodansha, Tokyo, Academic Press, New

[52] For a synthesis of the colchicine framework, see: a) M. G.

Banwell, M. Berak, D. C. R. Hockless, J. Chem. Soc. Perkin

Trans. 1 1996, 2217; for a synthesis of desacetamidoisocolchicine,

see: b) M. G. Banwell, J. N. Lambert, M. Corbett, R. J. Green-

wood, J. L. Gulbis, M. F. Mackay, J. Chem. Soc. Perkin Trans. 1

1992, 1415; for a synthesis of (ꢁ )-colchicine, see: c) M. G.

Banwell, J. N. Lambert, M. Mackay, R. J. Greenwood, J. Chem.

Soc. Chem. Commun. 1992, 974.

[53] The intention to apply a similar strategy was published after our

disclosure^[40a]: M. C. McMills, D. L. Wright, R. M. Weekly, Synth.

Commun. 2002, 32, 2417.

[54] C. Plüg, W. Friedrichsen, Tetrahedron Lett. 1992, 33, 7509 – 7510.

[55] a) K. Ueda, T. Ibata, M. Takebayashi, Bull. Chem. Soc. Jpn.

1972, 45, 2779; b) A. Padwa, M. D. Weingarten, Chem. Rev. 1996,

96, 223, and references therein.

York, 1975, p. 349; c) G. Quinkert, Synform 1990, 8, 207; for a

textbook discussion of some early syntheses, see: d) I. Fleming,

Selected Organic Syntheses, Wiley, London, 1973, p. 183.

[45] A. I. Scott, Nature 1960, 186, 556.

[46] For a review, see: W. T. Brady, Synthesis 1971, 3, 415.

[47] T. Nozoe, T. Seto, T. Sato, Proc. Jpn. Acad. 1954, 30, 473.

[48] Separation into the enantiomers via trimethylcolchicinic acid

(rac-25): H. Corrodi, E. Hardegger, Helv. Chim. Acta 1957, 40,

193.

[49] T. Ikemi, T. Nozoe, H. Sugiyama, Chem. Ind. 1960, 932.

[50] R. Robinson, The Structural Relations of Natural Products,

Claredon, Oxford, 1954, p. 46.

[51] B. Föhlisch, S. Sendelbach, H. Bauer, Liebigs Ann. Chem. 1987,

1.

3256

ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

www.angewandte.org

Angew. Chem. Int. Ed. 2004, 43,3230 – 3256

Article Doi

Total syntheses of colchicine in comparison: A journey through 50 years of synthetic organic chemistry

DOI: 10.1002/anie.200300615

Source and publish data:

Authors:

Article abstract of DOI:10.1002/anie.200300615

Full text of DOI:10.1002/anie.200300615

Products guided by the article

R&D Labs maybe for 369405-57-2

Relevant to this article

Hot Product