The total synthesis of proteasome inhibitors TMC-95A and TMC-95B: Discovery of a new method to generate cis-propenyl amides

Article abstract of DOI:10.1002/1521-3773(20020201)41:3<512::AID-ANIE512>3.0.CO;2-R

Silatropic and ene-like bond reorganizations (see scheme, left) were the key steps in the first total synthesis of the title compounds, which only differ in stereochemistry at the remote C36 stereocenter. Other key steps include a Suzuki biaryl construction, a diastereofacial dihydroxylation reaction, and a macrolactamization.

Full text of DOI:10.1002/1521-3773(20020201)41:3<512::AID-ANIE512>3.0.CO;2-R

The Total Synthesis of Proteasome Inhibitors
TMC-95A and TMC-95B: Discovery of a New
Method To Generate cis-Propenyl Amides**
Songnian Lin and Samuel J. Danishefsky*
The ubiquitin proteasome pathway is critical for accom-
plishing proteolysis in both the cytosol and nucleus of all
eukaryotic cells.[1] Understanding of the physiological roles of
a particular proteasome has been aided by studying the effects
of cell-permeable inhibitors. Moreover, specific proteasome
inhibitors are emerging as possible drug candidates.[1c] Ac-
cordingly, the discovery of the cyclic peptides TMC-95A
D
Scheme 1. Structures of TMC-95A D (1a d) and 2. TIPS ˆ triisopro-
pylsilyl, Bn ˆ benzyl.
(1a d; see Scheme 1), first isolated as fermentation products
of Apoispora montagnei Sacc TC 1093 derived from soil
samples,[2] was of great interest.[3, 4] The A and B isomers of 1,
which differ in stereochemistry at the remote, configuratively
labile C36 center, inhibit proteasomal functions of the 20S
proteasome with IC50 values down to low nanomolar levels.[2a]
Accordingly, a total synthesis program, which targets the
more active TMC-95A and B isomers, was launched in our
laboratory. We hoped to gain access to both compounds since
after a successful total synthesis the groundwork would be
more promising for establishing systematic SAR profiles of
these TMC-95 inhibitors. Ideally, simpler structures could be
designed, synthesized, and screened for similar activity in
anticipation of their use as new drug leads.
Recently, we described an approach to accomplish stage I in
our program, that is, the total synthesis of the TMC-95
inhibitors.[4a] The most advanced structure reported in our
previous disclosure was compound 2 (Scheme 1). The attain-
ment of this subgoal, promising as it was for our proposed
program, still left unaddressed several key issues for total
syntheses of 1a and 1b. Thus, provision would be necessary to
allow exposure of a free phenolic hydroxy group at C19 and a
homopyruvoyl group at N33 (see structure 2). Moreover, the
oxidation level at C25 must be upgraded from an alcohol, as
found in 2, to the carboxy state required to reach 1 (see
below). This was no trivial matter, given the free hydroxy
groups at C6 and C7in our model target, 2. We also note that
in the synthesis of 2 the facial sense of cis dihydroxylation at
[4a]
C6 and C7was virtually stereorandom.
Perhaps most
problematic was the cis-propenyl amide linkage that encom-
passes atoms N26 C29 of the multifaceted molecular ensem-
ble of 1. It is with this latter issue that we commence our
report.
Examination of the literature reveals several methods that
might, in principle, be relevant to the enamide problem.[5]
However, the applicability of these methods to the synthesis
of the active TMC compounds is not certain. Compounds 1
have a rich diversity of functionality, particularly in the
extended ™pyruvoyl∫-like (C34 C38) and dihydroxyindoli-
none (positions 22, 23, and 6 8) regions. Anticipation of
potential vulnerability in these sectors, not to speak of the cis-
enamide itself (N26 C29), prompted us to explore a new
modality for reaching such a substructure appropriate for
molecules with multiple sites of potential instability. We
wondered whether a substance of the type 3 might, upon
heating or catalysis, undergo concurrent ene- and silatropic-
like bond reorganizations that would lead to 4 (Scheme 2). If
this hypothesis were to be fruitful, the cis character of the
enamide would be virtually assured at the kinetic level.
Moreover, it seemed at least possible that the intermediate
silyl imidate linkage in 4 could be cleaved, with retrieval of
general substructure 5.
SiEt3
N
O
SiEt3
O
O
SiEt3
O
[*] Prof. S. J. Danishefsky, Dr. S. Lin
Laboratory for Bioorganic Chemistry
Sloan Kettering Institute for Cancer Research
1275 York Avenue, New York, NY 10021 (USA)
Fax : (‡1)212-772-8691
∆
or
cat.
H2O
R
N
¨
R
N
R
N
R
H
H
H
3
5
4
Scheme 2. Formation of cis-propenyl amides 5 by rearrangement hy-
drolysis of a-silylallyl amides 3 (R ˆ aryl, alkenyl, and alkyl groups).
E-mail: s-danishefsky@ski.mskcc.org
and
Department of Chemistry
Havemeyer Hall, Columbia University
New York, NY 10027(USA)
In the event, a range of probe substrates 3 was synthesized
by acylation of the known of amine 6 (Table 1).[6] As seen in
entries 1 3, heating of these compounds at about 1108C for
[**] This work was supported by the National Institutes of Health (grant
CA28824). S.L. gratefully acknowledges the US army breast cancer
research program for a postdoctoral fellowship (DAMD-17-99-1-
9373). We thank Dr. Jun Kohno, Tanabe Seiyaku Co., Japan, for
generously informing us of the conditions required for HPLC
separation of 1a and 1b as well as for authentic samples of these
compounds. We thank Dr. Michael Groll, Max Planck Institut f¸r
Biochemie, of Germany, for sample of 1a. We also thank Dr. George
Sukenick and Ms. Sylvi Rusli of the MSKCC NMR Core Facility for
NMR and mass spectral analyses (NIH Grant CA08748).
the time periods indicated gave rise to silyl imidates 4a
c
(observed by 1H NMR analysis).[7] Aqueous hydrolysis of
these compounds afforded enamides 5a c. The reaction was
also applicable to substrate 3d (entry 4), though a longer
heating time was required for its conversion to 4d. Most
significantly, the rearrangement hydrolysis sequence proved
extendable to the aminoacyl substrate 3e (entry 5) to afford
512
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
1433-7851/02/4103-0512 $ 17.50+.50/0
Angew. Chem. Int. Ed. 2002, 41, No. 3

Table 1. Rearrangement hydrolysis of a-silylallyl amides 3.[a]
a 1:1.3 mixture of a,b-unsaturated lactams (Z)-9 and (E)-9.[4a,e]
Fortunately, the former isomer could be converted to the
latter one through iodine-mediated isomerization.[4a,e] In a
parallel sequence, l-tyrosine (10) was converted to 11 in three
steps (Scheme 3).[9] A high-yielding ortho iodination[4a] of 11
led to 12 and thence, after palladium-mediated borylation,[10]
to 13. Suzuki-type coupling[11] of 13 with (E)-9 afforded
compound 14 (75% yield). We hoped that the biaryl domain
thus presented would be serviceable in the context of our
projected total synthesis (see below).
Hydrolysis of the methyl ester function of 14 led to the
corresponding carboxylic acid, which, after acylation of the
basic nitrogen atom of the asparagine derivative H-Asn-OtBu
(15), afforded 16. The hydroxy groups were introduced at C6
and C7, as shown in Scheme 4. Indeed, the presence of the
Garner N,O-acetonide served to direct, preferentially, the
oxidizing agent to the Re face (C6) of 16 to afford 17 in a 5:1
ratio relative to its 6R,7S stereoisomer (not shown).
RCOCl,
NEt3; or
O
O
TES
TES
a), b)
R
N
R
N
RCO2H,
EDC/HOAT
H2N
H
H
single
82-95%
6
3
5
isomer
Entry R ˆ
Conditions
Yield
[%]
1
2
a) toluene, 1108C, 10 h; b) H2O 81
a
b
a) toluene, 1108C, 20 h; b) H2O
7
3
O
3
4
5
a) toluene, 1108C, 27h; b) H 2O 67
c
d
e
O
a) toluene, 1108C, 3 d; b) H2O
7
2
TIPSO
a) o-xylene, 1358C, 4 d; b) H2O 52
HN
O
Boc
HO
O
N
[a] TES ˆ triethylsilyl, Boc ˆ tert-butoxycarbonyl, EDC ˆ 1-[3-(dimethyla-
mino)propyl]-3-ethylcarbodiimide hydrochloride, HOATˆ 1-hydroxy-7-
azabenzotriazole.
HO
Boc
N
6
O
O
O
Boc
O
a)
BnO
O
OtBu b)
N
OtBu
N
H
14
H
BnO
O
4e and thence 5e. It was of great interest to determine
whether this new method would find application at a very late
stage of our projected total synthesis.
The starting material for our refashioned synthesis was the
previously described iodooxindole 7 (Scheme 3).[4a] Crossed-
aldol condensation of 7 with the Garner aldehyde 8,[8]
followed by b-elimination of the derived mesylate, afforded
NH
N
H
O
NH2
NH2
O
Cbz
Cbz
N
H
HN
16
17
Scheme 4. Synthesis of diol 17. a) 1) LiOH, THF/H2O, 0 8C, 1.5 h; 2) H-
Asn-OtBu (15), EDC/HOAT, THF, RT, 2 h; 85% (two steps); b) OsO4/
NMO, (DHQD)2-PHAL, tBuOH/H2O, RT, 12 h; 88% (drˆ 5:1). Asn ˆ
asparagine, NMO ˆ 4-methylmorpholine N-oxide, (DHQD)2-PHAL ˆ 1,4-
bis(9-O-dihydroquinidine)phthalazine.
O
The timing and manner in which the various heteroatom-
centered functional groups were exposed and protected
proved to be critical to the success of the project. We
proceeded as follows (see Scheme 5): Deprotection of the
N,O-acetonide linkage of 17 afforded N-Boc triol 18. The
primary alcohol at position 25 was protected as its TIPS
derivative (19). Cleavage of the tert-butyl group generated a
carboxylic acid at C10 which set the stage for macrolactam-
ization (see compound 20).[12] In the next step, the benzyl
groups protecting the C19 phenol and N33 were concurrently
cleaved by hydrogenolysis. Acylation of the latter basic
nitrogen atom with racemic 21[13] afforded 22a and 22b as a
1:1 mixture. Both compounds were advanced concurrently.
The TIPS protecting group was cleaved from the primary
alcohol. The four hydroxy groups (positions 6, 7, 19, and 25)
were protected as the TES derivatives (see 23a and 23b). In a
key step of the synthesis, reaction of these compounds with
Jones reagent[14] led to specific oxidation at the primary center
(position 25) to afford acids 24a and 24b. In addition, partial
deprotection occurred at the C19 ether linkage which resulted
in the formation of 25a and 25b. This four-component
mixture led, after condensation with amine 6 as indicated, to
the amide silyl ethers 26a and 26b as well as amide phenols
27a and 27b (26a,b:27a,b ꢀ1:1.5).
N
Boc
H
H
O
O
a)
+
O
N
N
Boc
H
N
O
I
H
I
7
8
(E)-9
(Z)-9
b)
(1.1 equiv)
O
N
Boc
R
BnO
Cbz
HN
f)
c)
L-Tyr
(10)
OMe
N
O
Cbz
H
BnO
O
11 R = H
12 R = I
HN
d)
e)
OMe
O
O
B
13 R =
O
14
Scheme 3. Synthesis of biaryl compound 14. a) LDA (2.0 equiv), THF,
À788C, 1.5 h; NEt3, MsCl, CH2Cl2, À70 ! À 508C, 1.5 h; 81 % (E/Z ˆ
1.3:1); b) I2 (cat.), benzene, 808C, 26 h; DMP/PPTS, toluene, 658C, 5 h;
85% (60% conversion); c) 1) MeOH/SOCl2; 2) CbzCl/K2CO3; 3) BnBr,
Cs2CO3, acetone, reflux; 88% (three steps); d) Ag2SO4/I2, MeOH, RT, 1 h;
99%; e) pinacolatodiborane, [PdCl2(dppf)] ¥ CH2Cl2, KOAc, DMSO, 808C,
10 h, 91%; f) (E)-9, [PdCl2(dppf)] ¥ CH2Cl2, K2CO3, DME, 808C, 2 h; 75%.
Cbz ˆ benzyloxycarbonyl, LDA ˆ lithium diisopropylamide, DMP ˆ 2,2-
dimethoxypropane, PPTS ˆ pyridinium p-toluenesulfonate, dppf ˆ bis(di-
phenylphosphanyl)ferrocene.
Angew. Chem. Int. Ed. 2002, 41, No. 3
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
1433-7851/02/4103-0513 $ 17.50+.50/0
513

by their high-field NMR spectra in comparison with those of
authentic samples.[16]
HO
HO
HO
OR
HO
OTIPS
O
In summary, the total syntheses of TMC-95A and TMC-95B
have been achieved. The program featured a sequential
assembly of oxindole 7, Garner×s aldehyde (8), aryl boronate
13, asparagine derivative 15, 3-methyl-2-oxopentanoic acid
(21), and a-silylallyl amine 6. Highlights of the synthesis
include a Suzuki biaryl construction ((E)-9 ‡ 13 !14), a
diastereofacial dihydroxylation reaction that took advantage
of the Garner method (16 !17), and a macrolactamization
(formation of 20). We also note that new chemistry to
accomplish stereospecific cis-propenyl amide formation (26/
27 !28/29) was inspired by goal system 1. The application to
the delicate case at hand serves to build confidence in its
generality.
HN
O
O
Boc
O
O
c)
HN
a)
N
N
OtBu
17
H
H
BnO
b)
O
BnO
NH
N
H
NH2
O
NH2
O
Cbz
HN
HN
HN
18 R = H
Cbz
20
R2O
19 R = TIPS
R2O
R2O
R2O
O
OR1
O
R1
O
8
O
HN
O
f)
d) 21
R3O
N
N
O
H
O
H
R3O
O
O
NH
NH2
NH
NH2
O
14
O
N
N
Received: November 16, 2001 [Z18229]
H
H
O
O
22a,b R1 = TIPS, R2 = R3 = H
24a,b R1 = OH, R2 = R3 = TES
e)
23a,b R1 = R2 = R3 = TES
25a,b R1 = OH, R2 = TES, R3 = H
TES
g)
[1] a) A. Ciechanover, A. L. Schwartz, Proc. Natl. Acad. Sci. USA 1998,
95, 2727 2730; b) M. Bochtler, L. Ditzel, M. Groll, C. Hartmann, R.
Huber, Annu. Rev. Biophys. Biomol. Struct. 1999, 28, 295 317;
c) A. F. Kisselev, A. L. Goldberg, Chem. Biol. 2001, 8, 739 758.
[2] a) Y. Koguchi, J. Kohno, M. Nishio, K. Takahashi, T. Okuda, T.
Ohnuki, S. Komatsubara, J. Antibiot. 2000, 53, 105 109; b) J. Kohno,
Y. Koguchi, M. Nishio, K. Nakao, M. Kuroda, R. Shimizu, T. Ohnuki,
S. Komatsubara, J. Org. Chem. 2000, 65, 990 995.
, R2 = R3 = TES
26a,b R1
27a,b R1
=
=
N
H
TES
, R2 = TES, R3 = H
N
H
Scheme 5. Synthesis of a-silylallyl amides 26a,b and 27a,b. a) PPTS/
MeOH, reflux, 2 h; b) TIPSCl, imidazole/DMAP, CH2Cl2, RT, 5 h; 88%
(two steps); c) 1) TFA/CH2Cl2 (4:1), RT, 2 h; 2) EDC/HOAT/DIEA,
CH2Cl2/DMF(2 mm), RT, 24 h; 52% (two steps); d) 1) Pd/C, H2, EtOH,
RT, 19 h; 2) (Æ)-3-methyl-2-oxopentanoic acid (21), EDC/HOAT, CH2Cl2/
DMF, RT, 2 h; 85% (2 steps); e) 1) HF/Py; 2) TESOTf, 2,6-lutidine,
CH2Cl2, 08C !RT, 15 h; 3) NaHCO3; 4) citric acid, EtOAc/H2O; 73%
(from 22); f) Jones reagent, acetone, 08C, 2 h; g) 6, EDC/HOAT, CH2Cl2/
DMF, RT, 13 h; 45% (two steps). DMAP ˆ 4-dimethylaminopyridine,
TFA ˆ trifluoroacetic acid, DIEA ˆ N,N-diisopropylethyl amine.
[3] M. Groll, Y. Koguchi, R. Huber, J. Kohno, J. Mol. Biol. 2001, 311,
543 548.
[4] a) S. Lin, S. J. Danishefsky, Angew. Chem. 2001, 113, 2020 2024;
Angew. Chem. Int. Ed. 2001, 40, 1967 1970. For synthetic efforts
directed to the TMC-95 metabolites from other laboratories, see:
b) M. Inoue, H. Furyama, H. Sakazaki, M. Hirama, Org. Lett. 2001, 3,
2863 2865; c) D. Ma, Q. Wu,. Tetrahedron Lett. 2001, 42, 5279 5281;
d) B. K. Albrecht, R. M. Williams, Tetrahedron Lett. 2001, 42, 2755
2757; e) D. Ma, Q. Wu, Tetrahedron Lett. 2000, 41, 9089 9093.
[5] a) J. K. Stille, Y. Becker, J. Org. Chem. 1980, 45, 2139 2145; b) P.
Ribereau, M. Delamare, S. Celanire, G. Queguiner, Terahedron Lett.
2001, 42, 3571 73; c) K. Kuramochi, H. Watanabe, T. Kitahara,
Synlett 2000, 3, 397 399; d) R. Chen, J. A. Porco, Jr., Org. Lett. 2000,
2, 1333 1336.
Construction of the (Z)-1-propenylamide was now achiev-
ed by thermally driven rearrangement of the a-silylallyl
amides corresponding to 3 (Scheme 6). The rearrangement of
[6] Amine 6 (S.-F. Chen, E. Ho, P. S. Mariano, Tetrahedron 1988, 44,
7013 7026) was synthesized by using an improved procedure from
allyl alcohol through a one-pot TES ether formation, retro-Brook
TESO
TESO
R1O
R1O
O
TES
O
N
H
O
N
H
O
O
O
rearrangement, and mesylation, followed by
a displacement of
HN
HN
N
N
a)
mesylate with ammonia. Acylation in entries 1 4 (Table 1) was
accomplished by coupling of 6 with the acid chlorides, while in entry 5
an EDC-mediated coupling with protected amino acid with amine 6
was involved. Details will be forthcoming in a full disclosure.
[7] The formation of silyl imidates 4 was clearly observed by 1H NMR
spectroscopy when these reactions were carried out in deuterated
solvents.
[8] a) P. Garner, J. M. Park, J. Org. Chem. 1987, 52, 2361 2364; b) A.
McKillop, R. K. Taylor, R. J. Watson, N. Lewis, Synthesis 1994, 31 33.
Garner×s aldehyde was employed since it was found that an N,O-
acetonide was crucial for the stereoselective installation of diol
functionality at C6 C7position (see ref. [4a]).
O
O
H
H
O
O
RO
R2O
NH
NH
NH2
NH2
O
O
N
N
H
H
O
O
28a,b R1 = R2 = TES
26a,b R = TES
27a,b R = H
29a,b R1 = TES, R2 = H
1a,b R1 = R2 = H
b)
Scheme 6. Synthesis of TMC-95A (1a) and TMC-95B (1b). a) 1) o-xylene,
1408C, 3 d; 2) H2O; b) HF/Py, THF/Py; then Me3SiOMe; 49% (two steps).
[9] To address the issues raised by problematic deprotection of the
phenolic methyl group at
Scheme 1), a benzyl group, instead of a methyl group, was used for
the protection of the phenol (see ref. [4a]).
a later stage (that is, compound 2,
the complex mixture[15] in anhydrous o-xylene at 1408C
provided (Z)-1-propenylamides 28a,b and 29a,b. The crude
mixture of these compounds was globally deprotected with
pyridine-buffered HF/pyridine to afford a mixture of our total
synthesis goals–TMC-95A and TMC-95B (1a and 1b; 1:1).
This mixture was separated by RP-HPLC[16] to provide the
individual compounds 1a and 1b. These were characterized
[10] a) T. Ishiyama, M. Murata, N. Miyaura, J. Org. Chem. 1995, 60, 7508
7510; b) A. M. Elder, D. H. Rich, Org. Lett. 1999, 1, 1443 1446.
[11] For a review of Suzuki coupling reactions, see: N. Miyaura, A. Suzuki,
Chem. Rev. 1995, 95, 2457 2483.
[12] The atropisomer shown for 20 follows from the C6 stereochemistry
(see ref. [2b]).
514
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
1433-7851/02/4103-0514 $ 17.50+.50/0
Angew. Chem. Int. Ed. 2002, 41, No. 3

[13] 3-Methyl-2-oxopentanoic acid (21) was obtained from its commer-
cially available sodium salt. There would be no purpose in coupling
with enantiomerically defined acid 21 because the C36 stereocenter
epimerizes rapidly throughout the series.
[14] a) K. Bowden, I. M. Heibron, E. R. H. Jones, B. C. L. Weedon, J.
Chem. Soc. 1946, 39; b) R. A. Pilli, M. M. Victor, Tetrahedron Lett.
1998, 39, 4421 4424.
separation was achieved at the stage of the two-component mixture
1a/1b.
[16] HPLC conditions are as follows. Column: YMC-pack ODS-AM,
150 Â 10 mm; eluant: 25% MeCN in water; flow rate: 2.5 mLminÀ1
;
tR(1a) ˆ 37.3 min, tR(1b) ˆ 34.3 min. The synthetic materials were
identical to natural TMC-95A and TMC-95B (TLC, NMR, EIMS, and
HPLC). We thank Dr. Jun Kohno, Tanabe Seiyaku Co., Japan, for
generously providing us with the HPLC conditions and the samples of
TMC-95A and B for comparison.
[15] It was not feasible or necessary to separate at this stage. Rather
the eight-component mixture was advanced as shown, and the
Angew. Chem. Int. Ed. 2002, 41, No. 3
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
1433-7851/02/4103-0515 $ 17.50+.50/0
515