Use of Thiazoles in the Halogen Dance Reaction: Application to the Total Synthesis of WS75624 B

Article abstract of DOI:10.1021/jo0351217

The total synthesis of the pyridine-thiazole-containing natural product WS75624 B (1) is described. This synthesis proceeds via the Stille coupling of appropriately functionalized pyridine and thiazole components, and this paper details our studies on the use of the halogen dance reaction to prepare the desired thiazole. Various halogen dance reactions on thizoles are described, including a novel one-pot multistep reaction in which 2-bromothiazole is treated with LDA in the presence of a silyl chloride at -78 °C and quenched with an electrophile to provide the highly functionalized thiazole derivatives 27.

Full text of DOI:10.1021/jo0351217

Use of Th ia zoles in th e Ha logen Da n ce Rea ction :
Ap p lica tion to th e Tota l Syn th esis of WS75624 B
Eric L. Stangeland and Tarek Sammakia*
Department of Chemistry and Biochemistry, University of Colorado, Boulder, Colorado 80309-0215
sammakia@colorado.edu
Received J uly 30, 2003
The total synthesis of the pyridine-thiazole-containing natural product WS75624 B (1) is described.
This synthesis proceeds via the Stille coupling of appropriately functionalized pyridine and thiazole
components, and this paper details our studies on the use of the halogen dance reaction to prepare
the desired thiazole. Various halogen dance reactions on thizoles are described, including a novel
one-pot multistep reaction in which 2-bromothiazole is treated with LDA in the presence of a silyl
chloride at -78 °C and quenched with an electrophile to provide the highly functionalized thiazole
derivatives 27.
WS75624 A and B (Figure 1) are two related com-
pounds that were isolated from the fermentation broth
of Saccharothrix sp. No. 75624.1 The structures of these
compounds differ at the hydroxyl-bearing carbon of their
aliphatic side chain; WS75624 A has a tertiary alcohol
and one less methylene in the chain, while WS75624 B
has a secondary alcohol of which the stereochemistry is
unknown. These compounds are potent endothelin con-
verting enzyme (ECE) inhibitors and are potential anti-
hypertensive agents.2 We recently described the synthesis
F IGURE 1. The structures of WS75624 A and B and caeru-
of a structurally related molecule, caerulomycin C (Figure
1, 3), in which we made extensive use of the halogen
dance reaction3 to prepare a suitably functionalized
pyridine subunit.4 The similarity of these molecules
suggests that the synthesis of WS75624 B could be
accomplished using a similar disconnection. In this paper,
we describe the synthesis and determination of the
absolute stereochemistry of WS75624 B and, in the
context of this synthesis, the use of thiazoles in the
halogen dance reaction.
lomycin C.
by Huang and Gordon,6 and one synthesis of the struc-
turally related molecule WS75624 A.7 In Patt and Mas-
sa’s synthesis, the pyridine ring is derived from kojic acid
(Scheme 1). Thus, conversion of kojic acid to the substi-
tuted pyridine derivative 4 took five steps and proceeded
in 23% yield. This compound contains three of the four
pyridyl substituents, and incorporation of the fourth
substituent proceeded by a radical acylation procedure
involving acetaldehyde, tert-butyl hydroperoxide, and
ferrous sulfate to provide compound 5 in a modest 30%
yield. Conversion of 5 to bromo ketone 6 required two
steps and proceeded in 37% yield. Treatment of this
compound with racemic thioamide 7 provided the desired
thiazole (8) in 33% yield along with a major ring-
demethylated byproduct, the structure of which was not
fully characterized. Conversion of this material to the
natural product occurred uneventfully by alkaline hy-
drolysis in 71% yield. This synthesis is notable in its use
of a starting material, kojic acid, which contains much
of the required functionality of the pyridine of the
molecule but suffers from a problematic installation of
the thiazole moiety and several low-yielding transforma-
tions.
There are two reported syntheses of WS75624 B in the
literature, the first by Patt and Massa5 and the second
(1) Yoshimura, S.; Tsuruni, T.; Takase, S.; Okuhara, M. J . Antibiot-
ics 1995, 48, 1073.
(2) Tsuruni, Y.; Ueda, H.; Hayashi, K.; Takase, S.; Nishikawa, M.;
Kiyoto, S.; Okuhara, M. J . Antibiotics 1995, 48, 1066.
(3) For reviews of the halogen dance reaction, see: (a) Bunnett, J .
F. Acc. Chem. Res. 1972, 5, 139. (b) Fro¨hlich, J . Bull. Soc. Chim. Belg.
1996, 105, 615. (c) Fro¨hlich, J . In Progress in Heterocyclic Chemistry;
Suschitzky, H., Scriven, E. F. V., Eds.; Oxford: New York, 1994; Vol
6, pp 1-35. For leading references to more recent work, see: (d) Marzi,
E.; Bigi, A.; Schlosser, M. Eur. J . Org. Chem. 2001, 1371. (e) Trecourt,
F.; Gervais, B.; Mallet, M.; Queguiner, G. J . Org. Chem. 1996, 61, 1673.
(f) Comins, D. L.; Saha, J . K. Tetrahedron Lett. 1995, 36, 7995. (g)
Tre´court, F.; Mallet, M.; Mongin, O.; Que´guiner, G. J . Org. Chem. 1994,
59, 6173. (h) Guillier, F.; Nivoliers, F.; Godard, A.; Marsais, F.;
Que´guiner, G. Tetrahedron Lett. 1994, 35, 6489. (i) Rocca, P.; Cochen-
nec, C.; Marsais, F.; Thomas-dit-Dumont, L.; Mallet, M.; Godard, A.;
Que´guiner, G. J . Org. Chem. 1993, 58, 7832. (j) Marsais, F.; Pineau,
P.; Nivolliers, F.; Mallet, M.; Turck, A.; Godard, A.; Que´guiner, G. J .
Org. Chem. 1992, 57, 565.
(5) (a) Patt, W. C.; Massa, M. A. Tetrahedron Lett. 1997, 38, 1297.
(b) Massa, M. A.; Patt, W. C.; Ahn, K.; Sisneros, A. M.; Herman, S. B.;
Doherty, A. Bioorg. Med. Chem. Lett. 1998, 8, 2117.
(4) Sammakia, T.; Stangeland, E. L.; Whitcomb, M. C. Org. Lett.
2002, 4, 2385.
(6) Huang, S.-T.; Gordon, D. M. Tetrahedron Lett. 1998, 39, 9335.
(7) Bach, T.; Heuser, S. Synlett 2002, 2089.
10.1021/jo0351217 CCC: $27.50 © 2004 American Chemical Society
Published on Web 02/26/2004
J . Org. Chem. 2004, 69, 2381-2385
2381

Stangeland and Sammakia
SCHEME 1
F IGUR E 2. Coupling fragments for the synthesis of
WS75624 B.
SCHEME 3
Huang and Gordon’s synthesis (Scheme 2) begins with
3,4-dimethoxypyridine, which is known to undergo meta-
lation/Negishi coupling at the 2-position.3e Thus, coupling
SCHEME 2
quired an efficient synthesis of the thiazole component
(12), and we turned to the halogen dance reaction to
accomplish this task. The halogen dance reaction pro-
ceeds by a metalation to form a lithio aromatic species,
followed by a transposition of the lithio and halogen
moieties to produce a new organolithium (Scheme 3).8
The driving force for this reaction is the formation of a
more stable organolithium, and typically, the added
stability is derived from having the lithium flanked by
the halogen and an additional carbanion stabilizing
group. However, this is not a required component of the
reaction; if a substrate contains sites that differ in acidity,
there exists a thermodynamic driving force for the
rearrangement of a lithio species to the more acidic site.9
This is the case with thiazoles. Thiazole is most acidic
at the 2-position, less acidic at the 5-position, and least
acidic at the 4-position (Scheme 1).10 For our synthesis
of WS75624 B, we required a substituent at the least
acidic position of the thiazole (the 4-position), and we
therefore wished to force a metalation at this position
and migrate a halogen from either the 2-position or the
5-position to this site.
with bromothiazole 9, which was prepared in three steps
and 75% yield from 2,4-dibromothiazole, provided 10 in
50% yield. Compound 10 requires reduction of the mixed
thioketal, and this proceeds in two steps and 48% yield.
Incorporation of the required carboxylic acid proceeded
in two steps by a Reissert-Hanze reaction via the
N-oxide. This provided nitrile 11, which was hydrolyzed
to the racemic natural product in 30% overall yield for
three steps. This synthesis is notable for its brevity,
convergency, and lack of use of protecting groups, but it
suffers from a few steps that proceed in modest yields.
Our retrosynthetic analysis of WS75624 B is similar
to that for our synthesis of caerulomycin C.4 We first
disconnected the thiazole from the pyridine, simplifying
our precursors to the 6-iodopyridine derivative 13 and
the substituted 4-stannyl thiazole 12 (Figure 2). We
previously described the synthesis of 134 and now re-
To test the viability of the halogen dance reaction on
a simple thiazole substrate, we studied the reactivity of
2-bromothiazole (14, Scheme 4). This compound was
treated with LDA at -78 °C in ethereal solvents and then
subjected to an aqueous workup; however, under all
(8) For leading references on the mechanism of this reaction, see
ref 4 and references therein.
(9) For examples of halogen dance reactions that are driven by
differential acidities on thiophenes, see ref 11a and the following:
Sauter, F.; Fro¨hlich, H.; Kalt, W. Synthesis 1989, 771.
(10) For a review of the chemistry of lithio thiazole, see: Iddon, B.
Heterocycles 1995, 41, 533.
2382 J . Org. Chem., Vol. 69, No. 7, 2004

Total Synthesis of WS75624 B
SCHEME 4
enhanced by the presence of a halogen at the 5-position
of the thiazole, and we sought to stabilize this species
by the installation of an electron-donating group at that
position. We therefore conducted a reaction wherein
2-bromothiazole was treated with 2.2 equiv of LDA at
-78 °C in the presence of 1 equiv of TIPSCl. After an
aqueous workup, 5-triisopropylsilyl-4-bromothiazole (23)
was isolated in 85% yield (eq 1). This remarkable reaction
conditions examined a dark intractable product was
produced. We suspect that this is a result of lithiation at
the moderately acidic 5-position to produce 15, followed
by a halogen dance reaction to produce 2-lithio-5-bro-
mothiazole (16). This species can then undergo an
R-elimination and open to the unstable isocyanide (17),
which decomposes to a complex mixture. Although the
parent 2-lithiothiazole is known to be stable at -78 °C,9
we postulate that the electron-withdrawing nature of the
bromine at the 5-position stabilizes the thiolate of 17 and
facilitates the ring opening. We therefore decided to block
the 2-position of the thiazole and examine the halogen
dance reaction between the 4- and 5-positions, and we
prepared 2-silyl-5-bromothiazole derivatives by treating
2-bromothiazole with n-BuLi and trapping with a variety
of chlorotrialkylsilanes at -78 °C. The 2-silyl species were
then metalated with LDA and quenched with bromine
(Scheme 5). Of the silanes we examined (trimethylsilyl,
likely proceeds by the mechanism shown in Scheme 6.
The reaction is conducted at -78 °C, and at this tem-
perature, LDA does not react with TIPSCl. Thus, the first
SCHEME 6
SCHEME 5
equivalent of LDA deprotonates the thiazole at the more
acidic 5-position to produce 5-lithio-2-bromothiazole (15).
This compound can either undergo a halogen dance
reaction or be trapped by the TIPSCl. The latter is a more
facile process, and 5-triisopropylsilyl-2-bromothiazole (24)
is produced.11 Compound 24 can then be deprotonated
by the second equivalent of LDA in solution to produce
the substituted 4-lithothiazole 25. Since all the TIPSCl
is consumed after the first metalation, 25 can only
undergo a 1,3-halogen dance reaction to produce the
substituted 2-lithiothiazole 26. Compound 26 is now
stable at -78 °C and can be trapped with reactive
electrophiles. Table 1 shows the results of trapping with
a variety of common electrophiles. In cases that display
diminished yields, the balance of the material consists
of the protonated product 23.
triethylsilyl, tert-butyldimethylsilyl, and triisopropylsi-
lyl), we found that only the triisopropylsilyl (TIPS)
derivate (18) underwent the bromination reaction with-
out desilylation. When 2-triisopropylsilyl-5-bromothiazole
(19) was subjected to the halogen dance conditions by
treatment with LDA at -78 °C, it smoothly rearranged
to the 4-bromo derivative 22 in 86% yield (Scheme 5).
This reaction likely proceeds by a deprotonation at the
4-position to provide 20, followed by a 1,2-dance to
provide the 4-bromo-5-lithio product (21) in which the
lithium is at the more acidic site. Confident that the
halogen dance is viable on thiazole derivatives, we
returned to examining the use of 2-bromothiazole as a
starting material.
This reaction is remarkably clean considering its
complexity and the lability of many of the intermediates.
A drawback, however, is the low nucleophilicity of the
2-lithiothiazole 26; at -78 °C, 26 does not react with
hindered lactones and typical alkyl halides, though it will
react with more reactive alkyl halides, such as methyl
(11) For other examples where an in situ trapping with a silyl
chloride proceeds faster than a halogen dance, see: (a) Kano, S.; Yuasa,
Y.; Yokomatsu, T.; Shibuya, S. Heterocycles 1983, 20, 2035. (b) Bury,
P.; Hareau, G.; Kocienski, P.; Dhanak, D. Tetrahedron 1994, 50, 8793.
(c) Cochennec, C.; Rocca, P.; Marsais, F.; Godard, A.; Que´guiner, G.
Synthesis 1995, 321. (d) Arzel, E.; Rocca, P.; Marsais, F.; Godard, A.;
Que´guiner, G. Heterocycles 1999, 50, 215.
Our previous attempts at the use of 2-bromothiazole
as a partner in this reaction suffered from the lability of
the 2-lithiothiazole that is produced by the halogen dance.
As previously mentioned, it is likely that this lability is
J . Org. Chem, Vol. 69, No. 7, 2004 2383

Stangeland and Sammakia
TABLE 1. Th ia zole Ha logen Da n ce a n d Su bsequ en t Tr a p p in g w ith Electr op h iles
SCHEME 7
SCHEME 8
converted to enoate 29 in 63% yield in a one-pot proce-
dure in which the lactone was reduced with diisobutyl-
aluminum hydride at -78 °C and then subjected to an
in situ Horner-Emmons olefination.14 The secondary
alcohol was protected as a TIPS ether, and the ester was
reduced with diisobutylaluminum hydride to alcohol 31,
which was then activated as the chloride (32), bromide
(33), or tosylate (34). Activation as the tosylate using
tosyl chloride was problematic, due to competing dis-
placement of the tosylate by chloride ion. This could be
suppressed using p-toluenesulfonic anhydride and depro-
tonating the alcohol with n-BuLi.15
With the activated electrophiles in hand, we attempted
a coupling reaction between them and lithothiazole 26.
As expected, the reaction was sluggish and did not
provide usable quantities of the desired product. How-
ever, in the presence of 10% CuI, a 42% yield of the
coupled product was obtained with allylic tosylate 34
(Scheme 8). The corresponding bromide (33) and chloride
(32) provided diminished yields and significant amounts
of the SN2′ coupling product under a variety of conditions.
Diimide reduction of the alkene of 34 occurred smoothly;
iodide, and with aldehydes.12 At temperatures above
about -70 °C, it is unstable and undergoes decomposition
faster than alkylation.
To complete the synthesis of WS75624 B, we were faced
with the problem of appending the side chain at the
2-position of the thiazole. Since alkylation of 26 with
simple primary alkyl halides proceeded in low yields, we
sought to increase the reactivity of the alkylating agent
and rendered the leaving group allylic. The synthesis of
our coupling partner proceeded as shown in Scheme 7
beginning with ethyl levulinate. Noyori hydrogenation
using the (S)-BINAP-derived catalyst provided (S)-lac-
tone 28 in 91% yield and >98% ee.13 This compound was
(14) For a related transformation, see: Burke, S. D.; Deaton, D. N.;
Olsen, R. J .; Armistead, D. M.; Blough, B. E. Tetrahedron Lett. 1987,
28, 3905.
(15) This reaction requires prior purification of the anhydride by
recrystallization from ethyl acetate in order to minimize the formation
of several byproducts.
(12) For a review of nucleophilic reactions of 2-lithiothiazole, see:
Dondoni, A. Synthesis 1998, 1681.
(13) Ohkuma, T.; Kitamura, M.; Noyori, R. Tetrahedron Lett. 1990,
31, 5509.
2384 J . Org. Chem., Vol. 69, No. 7, 2004

Total Synthesis of WS75624 B
SCHEME 9
of 39 to WS75624 B simply required hydrolysis of the
amide to the acid and deprotection of the silyl group.
However, we were unable to accomplish the hydrolysis
in a single step and resorted to the two-step procedure
of diisobutylaluminum hydride reduction of the amide
to the aldehyde followed by Lindgren oxidation16 to the
carboxylic acid. Finally, deprotection of the TIPS group
with TBAF provided (S)-WS75624 B in 67% yield for
three steps. The optical rotation of our material was
determined to be [R]D ) +3.1 (c ) 8.0 mg/mL, CH3OH),
while that of the natural product was reported as being
[R]D ) +3 (c ) 8.0 mg/mL, CH3OH).1 This allows us to
assign the stereochemistry of the natural product as the
(S)-configuration at the carbon bearing the secondary
alcohol.
Con clu sion s. We have described the asymmetric
synthesis of WS75624 B and the use of thiazoles as
substrates for the halogen dance reaction. With appro-
priately functionalized thiazoles, this reaction can provide
high yields of compounds that are difficult to obtain by
other methods. In the course of our synthesis, we utilized
a novel, one-pot functionalization of 2-bromothiazole in
which the halogen dance reaction plays a key role.
Studies to further extend the scope of these reactions are
currently in progress.
however, the TIPS blocking group on the thiazole proved
difficult to remove without extensive decomposition. We
therefore repeated the sequence starting from 14 using
a triethylsilyl (TES) blocking group instead and found
that it was smoothly removed upon stirring in basic
methanol (Scheme 8).
Completion of the synthesis proceeded as shown in
Scheme 9. Compound 37 was converted to the tributly-
stannyl derivative by treatment with t-BuLi with an in
situ quench with tributyltin chloride to provide 38. Due
to the steric hindrance of the t-BuLi, halogen-metal
exchange is more facile than attack on the tributyltin
chloride, allowing the use of the tin electrophile as an
internal quenching agent. If the tributyltin chloride is
not present during the halogen-metal exchange, dimin-
ished yields of product are observed, presumably due to
deprotonation of the R-methylene of the thiazole. Cou-
pling of 38 with iodopyridine 13 (see Figure 2) was then
accomplished using Pd(PPh3)4 in dimethylacetamide at
60 °C for 36 h and provided 39 in 53% yield. Conversion
Ack n ow led gm en t. We thank the National Insti-
tutes of Health (GM48498), Array Biopharma, and
Pharmacia (graduate fellowship to E.L.S.) for financial
support of this research. NMR instrumentation used in
this work was supported in part by the National Science
Foundation CRIF program (CHE-0131003).
Su p p or tin g In for m a tion Ava ila ble: Spectroscopic and
analytical data as well as experimental procedures for the
synthesis of all compounds shown in Schemes 5, 7, 8, and 9.
This material is available free of charge via the Internet at
http://pubs.acs.org.
J O0351217
(16) Lindgren, B. O.; Nilsson, T. Acta Chem. Scand. 1973, 27, 888;
Dalcanale, E.; Montanari, F. J . Org. Chem. 1986, 51, 567.
J . Org. Chem, Vol. 69, No. 7, 2004 2385