Angewandte
Research Articles
Chemie
Asn (N) and Met (M) in no particular order, as well as a free
N terminus. To our delight, reactivity and selectivity com-
parable with the ones of other peptide models were observed
(entry 4).
product (entry 21). Longer reaction times did not result in
significant increase in conversion. The isolation of pure
product 25b was also difficult and low yielding. However,
good results were obtained with peptides 26 and 28 resulting
in 77% and 110% relative absorbance, respectively (en-
tries 22 and 23). In this case, the products 27b and 29b could
also be isolated in good yields À44% and 55% respectively.
Stapling was also less efficient with peptide models 30 and 32
containing free N-terminus (entries 24 and 25). Only the
stapling between Cys–Lys was observed by MS/MS analysis of
both models. In contrast, excellent reactivity and full con-
version of the starting peptides 24, 26 and 28 was again
observed with the ortho reagent 9c after 30 minutes (Ta-
ble 2B, entries 26–28). However, all attempts to isolate the
pure products 25c, 27c or 29c were unsuccessful due to the
observed instability of the products in solution.[24] For this
reason, the peptide scope was not further investigated, as the
instability of the ortho-linker made it not suitable for
biological applications. Overall, the trends in reactivity
suggest that the position of the electron withdrawing carboxy
group on the reagent has a strong influence on the efficiency
of stapling.
Next, we turned to the scope of reagents. The cyclohexyl-
based reagent 8d afforded the stapled product with peptides
16, 18 and 20 as efficiently as the iPr-based reagent 8c
(entries 5–7). Introduction of the longer silanol-based linker
using reagent 8e was generally less efficient (entries 8–11),
with the exception of a slightly better result obtained for the
i,i + 7 model (entry 10). This is in accordance with the longer
distance between the two iodine atoms. Due to the observed
low reactivity of 8e, the monothioalkynylation intermediate
was commonly detected as the major product. Longer
reaction times did not increase conversion (Supporting
Information, Table S5). The use of the para- and meta-
substituted phenyl reagents, 8a and 8b was then examined for
both the one-loop model 16 and the two-loop model 20
(entries 12–15). In general, the phenyl- based linkers were less
efficient than the corresponding silicon analogues.
Only the para-linker together with two-loop peptide
model 20 provided a comparable result (entry 13).[23] The
isolation of complex pure peptides is often difficult and
associated with significant loss in yield. We were therefore
pleased to see that selected peptides could be isolated in pure
form after preparative HPLC in 13–48% yield (Entries 1–2, 7,
8–10).
To further investigate the reaction mechanism and the
origin of the observed selectivity, stapling of peptide 26 with
the para reagent 9a was chosen as it displayed quantitative
conversion, and no other side-products were detected by
HPLC. Kinetic MS experiments showed that the thiol attack
onto the EBX core is very fast, complete in under 1 minute,
In order to test the cysteine-lysine (CK) stapling reagents
9, we synthesized the same peptide models as used for Cys– yielding intermediate 34 (Figure 1). The formation of an
Cys stapling (16, 18, 20 and 22), but exchanging the second
cysteine for a lysine— to give peptides 24, 26, 28 and 30
(Table 2B). In addition, the model 32, previously used by
Buchwald and co- workers specifically for Cys–Lys sta-
pling,[12d] albeit with a free N-terminus was also synthesized.
To our delight, the para-reagent 9a stapled the peptide
models 24, 26, and 28 very efficiently in only 30 minutes
(entries 16–18). By HPLC, only the stapled products were
observed. In this case again, higher absorbance is observed for
the stapled products in comparison to the starting materials.
To confirm that the relative absorbance was still reasonably
correlated with the yield, the reaction was performed on
larger scale. The stapled products 25a, 27a and 29a were
isolated in 52, 65 and 87% yields, respectively. Therefore, the
stapling with the Cys–Lys system appeared to be more general
and efficient than with Cys–Cys. This could be due to the
higher reactivity of the hypervalent iodine reagent, or the
higher flexibility of the lysine side chain. The nucleophilic
peptide 30 with a free N-terminus was stapled slightly less
efficiently (entry 19). Interestingly, only one product was
detected by HPLC analysis. MS/MS studies confirmed the
formation of the Cys–Lys stapled product. This selectivity
however, appears to be sequence dependent. In fact, when the
unprotected Buchwald model 32 was examined, both stapled
products— Cys–Lys (33a) and Cys-N-terminus staple (33a’)—
were obtained.
ynamine intermediate instead is highly improbable, as EBX
reagents are known to react rapidly with thiols and not with
amines.[14,15] Overall, the reaction was finished in 10 minutes
even at a 0.1 mM concentration providing the product 27a.
We were pleased to see that intermediates 35 and 36 arising
The meta-reagent 9b showed lower reactivity with peptide
24. When previously using reagent 9a, full conversion was
observed, but meta-reagent 9b provided only 63% conversion
of 24 in 30 minutes with a 34% relative absorbance for the
Figure 1. MS kinetic experiment following reaction between peptide 26
and para reagent 9a. The possible reaction intermediates 34–36 are
shown. See Supporting Information, Figure S3 for details.
9026 www.angewandte.org ꢀ 2021 The Authors. Angewandte Chemie International Edition published by Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 9022 –9031