1536
C. Niu et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1535–1538
O
O
port studies on the absolute configuration of the synthetic analogs
HTI-286 and HTI-042.
1. MeNH2
2. BH3.Py
MeI
OH
OH
OH
R
R
N
The stereochemistry of compounds 2 and 3 was expected to
be S,S,S based on the synthetic strategy, HPLC retention time,
NaOH
O
O
and bioassay data.8,11 To confirm the stereochemistry at the
a-
4
5
carbon of the N-Me-b-dimethyl phenylalanine residue in these
hemiasterlin derivatives and to assign the stereochemistry of 2,
we utilized the double-pulsed field gradient spin echo (DPFGSE)
1D NOE experiment. All the NOE data discussed in this Letter
were acquired using THF-d8. This solvent provided a very well
dispersed 1D 1H NMR spectrum and appeared to encourage the
occurrence of one significant solution conformation. This obser-
vation was evidenced by the extreme magnitudes of the homo-
nuclear coupling constants (both small and large) for 2. An
example of the 1D NOE data acquired from selective excitation
of the proR-methyl group of the N-Me-b-dimethyl phenylalanine
residue is shown in Figure 2. As can be seen, the strong NOE cor-
relations between the proR-methyl (CH3-20), the N-3 Me, and
the tert-butyl group (12a–c; Fig. 3) require that all of these
atoms be on the same face of the molecule. When this informa-
tion is coupled with the fact that the proS-methyl (CH3-19)
shows correlations only to CH3-20 and the H-14 proton, it be-
comes clear that this methyl must be on the opposite face of
the molecule from CH3-20 and the stereochemistry at C-14 must
be S.8,12
O
O
H2N
O
O
7
R
HN
1. PyBOP, DIPEA
2. HPLC
6
O
O
O
N
N
R,S,S-isomer
+
R
H
HN
O
8
(S,S,S)
9
O
O
N
N
LiOH
OH
R
H
O
HN
2
3
HTI-286, R = H
HTI-042, R = OCH3
Some of the most important NOE correlations for the analysis of
HTI-286 are shown in Figure 3. According to these studies, the ste-
reochemistry of HTI-286 was in fact the predicted S,S,S configura-
tion at the three chiral centers, and this compound appeared to
fold back on itself along its length. These observations were con-
firmed by single-crystal X-ray diffraction.
Scheme 1. Synthesis of HTI-286 and analogs.
Crystals of HTI-286 (2) and the hydrochloride salt of HTI-042 (3)
were grown from acetonitrile/water solutions as colorless prisms
for a pentahydrate acetonitrile solvate, and colorless needles for
a tris(acetonitrile) solvate, respectively. The Flack absolute struc-
ture parameter for 3 refined to À0.03(11) for (S,S,S) and 1.01(11)
for (R,R,R). These values are consistent with the correctness of
the (S,S,S) assignment for 3 based on anomalous dispersion. Using
3 as a heavy-atom derivative, the absolute configuration of 2 is also
(S,S,S).13
Perspective views of 2 and 3 are given in Figure 4. Also included
is a view of the methyl ester of 1 to facilitate comparisons. Note
that it was drawn as the (R,R,R) instead of the (S,S,S) isomer in an
earlier publication.10 The tripeptides 1–3 differ only in their A-
piece amino acids, so it is perhaps not surprising that the most
noticeable conformational differences in their X-ray structures ap-
pear near the A-piece ends of these molecules. These differences
are largely due to torsional degrees of freedom in the O@C–C–N
units in the A-piece amino acid. Referring to Figure 4, from top
to bottom, the O@C–C–N torsion angles are 155.7(4)°, À45.5(5)°,
and À45.9(3)°, respectively. Note that the A-piece geminal di-
methyl group is a key component for potent activity, yet the tor-
sion angles for 2 and 3 forces their dimethyl groups to be
completely out of synch with the dimethyl groups in 1. Presumably
these inhibitors assume the same torsion angle upon binding to
tubulin,14 since they have similar biological activity.
7
6
5
4
3
2
1
ppm
Figure 2. DPFGSE 1D NOE with selective excitation of tert-butyl group of HTI-286.
predicted as (S,S,S) was then hydrolyzed with lithium hydroxide,
providing the target analog 10. SAR studies revealed that changing
the stereochemistry at each chiral center of the tripeptides resulted
in reduction of the activity.8 To facilitate the design and synthesis
of novel analogs with demonstrated cytotoxic and tubulin binding
abilities, it was necessary to confirm the stereochemistry of HTI-
286 and its analogs. To date there have been no reports on deter-
mination of the stereochemistry of synthesized analogs such as 2
and 3 by X-ray crystal analysis. However, the structure of the
methyl ester of natural product 1 was reported.10 We hereby re-
12a
N-2
12c
19
20
14
O
O
1
O
CH3
N
O
H
16
3
9
2
4
N
17
18
15
10
13
N
H
OH
N-1
N
H
OH
5
H
NH
NH
O
8
O
CH3
N-3
7
6
Figure 3. Key NOE enhancements observed for the assignment of C-14 stereochemistry.