P. S. Shankar et al. / Tetrahedron Letters 54 (2013) 6137–6141
6139
O
OH
O
OH
(a)
72%
(b)
95%
O
O
(b)
BocHN
N
COOR
+
Br
N
H
N
COOEt
PPh3
Br
*
*
Br
RHN
55-94%
O
O
O
OH
*
*
S
S
9
8
17a-d R = Et
18a-d R = H
7a-d
16a-d
R = Boc
R = H.TFA
(a)
(c)
O
(c)
PPh3
96%
Ph
O
OH
O
10
(d)
(e)
60-95%
BocHN
N
N
H
N
COOMe
Ph
*
*
*
*
57-94%
Ph
BocHN
H
Ph
BocHN
S
(d)
O
10
(e)
70-78%
H
OH
BocHN
60-86%
*
*
*
19a-k
20a-k
R = Boc
R = H.TFA
O
O
(a)
11a-b
12a-b
Ph
Ph
Ph
O
OH
O
(f)
( )n
O
H
N
O
O
(f)
48-98%
+
BocHN
N
BocHN
87-90%
*
*
N
R
N
*
N
COOMe
O
*
*
*
*
H
H
O
S
13c-d
13a-b
21a-m
(
g)
90-95%
(
g)
90-95%
R = H, Me
n = 1, 2
O
Ph
Ph
Ph
O
H
O
O
( )
n
-
+
.
-
+
.
Cl H3N
COOX
(X = H)
Cl H3N
COOX
(X = H)
N
N
N
R
N
H
N
H
COOH
14c
15c
*
*
*
14a
(h)
(h)
>98%
O
S
>98%
15a
(X = Me)
(X = Me)
1a-m
R = Ac, Me
n = 1, 2
Ph
Ph
-
+
.
Cl H3N
COOX
(X = H)
-
+
.
Cl H3N
COOX
(X = H)
Scheme 4. Final assembly of fragments. Reagents and conditions: (a) TFA:CH2Cl2
(1:4), 0 °C to rt, 1 h; (b) HOBt, EDC HCl, then Boc-Ile, DIPEA, CH2Cl2, 0 °C to rt, 3 h; (c)
LiOHÁH2O, THF:H2O (4:1), 0 °C to rt, 5 h; (d) HOAt, HATU, then Tup-OMe 15, Et3N,
CH2Cl2, 0 °C to rt, 3 h; (e) HOAt, HATU, then N-methyl pipecolic acid or pipecolic
acid or N-methyl proline, Et3N, 0 °C to rt, 3 h; (f) (i) 1 N LiOH, THF, 0 °C to rt, 2–
3 days; (ii) Ac2O, pyridine, overnight.
14d
15d
14b
15b
(h)
(h)
>98%
>98%
(X = Me)
(X = Me)
Scheme 3. Synthesis of diastereoisomers of Tup. Reagents and conditions: (a) Et3N,
dry THF, 0 °C to rt, 2 h; (b) PPh3, THF, 2 h, 0.38 N NaOH, toluene, 3 h; (c) MeI, CH2Cl2,
0 °C to rt, overnight; (d) Dess–Martin periodinane, CH2Cl2, 6 h; (e) CH2Cl2, 0 °C to rt,
8 h; (f) H2, Pd/C, EtOAc, overnight; (g) 6 N HCl, 130 °C, 1.5 h; (h) 2,2-dimethoxy-
propane, concd HCl, MeOH, 60 °C, overnight.
21a–m in good yields and with no detectable loss of sterochemical
purity. LiOH mediated hydrolysis of the methyl ester followed by
acetylation using acetic anhydride in pyridine afforded the final
tubulysin analogues 1a–m in their stereopure form (Scheme 4
and Fig. 2).
The
a,b-unsaturated aminoester 12 thus obtained was further
hydrogenolyzed using Pd–C (10%) to give N,C-protected Tup pre-
cursors 13a–d as a mixture of their corresponding diastereoiso-
mers (2:1 ratio, evidenced by 1H NMR measurements), separable
by simple flash chromatography. Separation of the isomers fol-
lowed by full deprotection to the free Tup stereoisomers 14a–d
and subsequent carboxyl group methylation using 2,2-dimethoxy
propane afforded the corresponding Tup-methyl esters 15a–d as
hydrochloride salts in good yields. The relative stereochemistry
was assigned based on the single crystal X-ray analysis of one of
the diastereoisomers.8
With all the stereoisomers of both fragments in hand, the final
assembly of tubulysins via standard solution phase peptide synthe-
sis was initiated by Boc deprotection of the stereoisomers 7a–d to
16a–d with 20% TFA:CH2Cl2, followed by a subsequent coupling
with N-Boc isoleucine (Boc-Ile) affording the corresponding dipep-
tides 17a–d. Saponification of the ethyl ester under mild alkaline
conditions to afford 18a–d and subsequent coupling with the re-
quired isomer of Tup-OMe 15a–d gave the tripeptides 19a–k in
satisfactory yields. A second Boc-deprotection of 19 to 20a–k fol-
lowed by further coupling to N-methyl pipecolic acid, or N-acetyl
pipecolic acid or N-methyl proline (the last two were used in case
of tubulysin U precursors bearing the natural stereochemistry on
all carbons) using HOAT-HATU-Et3N yielded the tetrapeptides
Biological tests
Most of the tubulysins 1a–m, with the exception of 1b and 1i
(whose samples were accidentally lost after characterization),
were subjected to cytotoxicity assays on two different cell lines:
HL60 (human promyelocytic leukaemia) and C6 (rat glial tu-
mour). Results are summarized in Table 2. Modification of Tup
stereochemistry has a minor influence on the cytotoxicity of
tubulysins, as witnessed by the fact that 1a, which is the C-2
Tup epimer, is nearly as potent as tubulysin U (1c), whereas 1d,
which is the C-4 Tup epimer, is only ca. 10 times less potent than
1c. These results are in line with those published by Yang et al.,
who observed that 1d is only 3-times less potent than 1c on
HL-60 cells.6 In contrast, modification of either Tuv stereocenter
has a dramatic effect on the cytotoxicity of tubulysins, as demon-
strated by the fact that all of the derivatives 1e–k are at least
1000 times less potent than 1c. Finally, replacement of the Mep
N-methyl group with an acetyl in 1l also results in a dramatic
drop of cytotoxicity, whereas the Mep ring contraction to N-
methyl proline is quite well tolerated as seen with 1m which is
only ca. 10 times less potent than 1c.