Communication
p
In the acyclic peptide control CBLLpC 2, the activity of the
longest stapled peptide CBLLCLL 12 was the highest chemo-
selectivity observed in this study. Corresponding to A/D=63, it
indicates nearly complete suppression of the undesired decar-
boxylation product 27. This record performance presumably
originated from better organization of the stapled turn with
additional hydrogen bonds to and from the exocyclic α helix
elongation. Additional contributions from remote control of
anion-π catalysis by the strengthened α-helix macrodipole
would be consistent with its orientation (Figures 1a, 3).[17]
Movement of the tertiary amine base B in CBLLC 1 (qAD=
0.94) by one position toward the C terminus in CLBLC 14
removed all activity (qAD=0.00, Figure 3, Table 1). Indeed, the
NDI-stapled 14 was less selective than the acyclic control
pCLBLpC 15 (qAD=0.18). Hypersensitivity toward base position[3]
supported that proximity to the π-acidic NDI surface is essential
for CBLLC 1, while in isomer CLBLC 14, the base is too far from
the NDI surface for operational anion-π catalysis. C-terminal
modification to adamantyl amide failed to activate the CLBLC
motif in 16 (qAD=0.25).
1
2
3
4
5
6
7
8
9
NDI-stapled CBLLC 1 was completely lost (pC stands for StBu
protected C, Figure 1b, Table 1). Chemoselectivity was worse
than with the TEA standard (qAD=À 0.12). Increasing the space
between π surface and peptide with one and two homocys-
teines in CBLLhC 3 and hCBLLhC 4 resulted in qAD=0.40 and
0.51, respectively (Figure 3). Expansion of CBLLC 1 (qAD=0.94)
by one or two atoms thus clearly weakened anion-π catalysis.
hCBLLhC 4 was the first example in the series for which two
atropisomers have been isolated in pure form (Figure 1a). While
catalytic activity of these two diastereomers differed clearly
(qAD=0.51, 0.18), assignment of their structures was not
possible without single crystals (Table 1).
Contraction of the peptide turn from CBLLC 1 to CBLC 5
nearly removed all activity (qAD=0.18, Figure 3, Table 1).
Exocyclic addition of the removed L at the N terminus of
LCBLC 6 (qAD=0.35) failed to restore the activity of the
constitutional isomer CBLLC 1 (qAD=0.94). Expansion of the
peptide turn with one more L in the NDI-stapled CBLLLC
hexapeptide 7 afforded similarly poor activity (qAD=0.30).
These losses in activity upon contraction and expansion implied
that topologically correct stapling of an α-helix turn in CBLLC 1
is essential for powerful anion-π catalysis.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Movement of the tertiary amine base B one more position
toward the C terminus in CLLBC 17 (qAD=0.68, 0.60) restored
some but not all activity of the original CBLLC 1 (qAD=0.94).
Moreover, the acyclic control pCLLBpC 18 (qAD=0.30) was more
selective than the acyclic original pCBLLpC 2 (qAD=À 0.12).
These trends were consistent with the positioning of the amine
bases according to ring current effects in 1H NMR
spectroscopy.[12] In the CxxBC series, NDI stapling caused a
downfield shift of the dimethyl amine resonances (CLLBC 17a/
The bulky, solubilizing adamantyl amide in CBLLC 8 did
much not disturb anion-π catalysis (qAD=0.72, Figure 3,
Table 1). The massive drop in activity upon removal of the NDI
was reproduced with CBLLpC 9 (qAD=0.10). Exocyclic peptide
p
elongation at the C terminus of the thus solubilized CBLLC 8
(qAD=0.72) caused a strong increase in chemoselectivity.
Hexapeptide CBLLCL 10 gave qAD=0.97, heptapeptide
CBLLCLL 12 gave qAD=1.20. The acyclic, NDI-free controls of
these longer pCBLLpCL 11 and pCBLLpCLL 13 remained consis-
tently inactive (qAD=À 0.12). The qAD=1.20 obtained for the
p
b: 2.33/2.31 ppm, Figure S129/S132, CLLBpC 18: 2.28 ppm, Fig-
ure S128). In contrast, those of CBxxC series shifted upfield
upon NDI stapling (CBLLC 1: 2.08 ppm, Figure S30, pCBLLpC 2:
2.25 ppm, Figure S18). These opposing ring current effects
demonstrated that the positioning of the amine base right
above the π surface, needed to stabilize the enolate intermedi-
ate by anion-π interactions as soon as it is produced,[3] is
achieved only in CBLLC 1 (TS-1, Figure 2).
Asymmetric anion-π catalysis has been realized for many
reactions.[3] However, anion-π catalysis of the addition of 24 to
enolate acceptors 25 never occurred with significant enantiose-
lectivities, except for anion-π enzymes.[3] Record values reported
so far come from axially chiral NDI catalysts.[19] With peptide-
bridged NDIs, enantioselectivities were better yet still not very
impressive. The highest enantiomeric ratio er 66:34 was
observed with CLBLC 14, an NDI-stapled peptide that does not
operate with anion-π catalysis and thus fails to influence
chemoselectivity (qAD=0.00). For the best performing CBLLC
motif, er >60:40 could be observed, while the intermediate
CLLBC isomers gave the other enantiomer preferentially at er
<36:64. Comparison of er 65:35 of CBLLC 1 with er 41:59 of
acyclic pCBLLpC 2 revealed that NDI stapling occurs with an
inversion of the absolute enantioselectivity in this series, a trend
p
that was confirmed by CBLLC 8 (er 61:39) against CBLLpC 9 (er
43:57). This inversion of configuration upon NDI stapling could
be one of the reasons why the enantioselectivity of CBLLC
anion-π catalysts is not as impressive as their chemoselectivity.
Reaction kinetics were measured first for the cyclic (C)
CBLLC catalyst 1 compared to the acyclic, open (O) pCBLLpC
Figure 3. Full structure of selected catalysts. All catalyst structures are
defined in Table 1.
Chem Asian J. 2020, 15, 1–6
3
© 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
��
These are not the final page numbers!