Angewandte
Chemie
5
The various kinetic constants for [Eq. (3)] range from 10
6
À1 À1
to 10 m
s
(Figure 3). Even taking into account the high
kinetic isotope effect of H/D transfer reactions, this order
of magnitude is completely comparable, for example, to the
[10]
Figure 2. Proportion of the carbon-centered radicals formed by attack
of an HOC radical that are transferred between various nondeuterated
amino acids and deuterated leucine. Gray bars=radical transferred
from the AA to deuterated leucine; white bars=radical transferred
from deuterated leucine to the AA.
Figure 3. a) Ratio of the deuterium transfer from deuterated leucine
measured in the absence of cysteine (T0) to that in the presence of
cysteine (T), as a function of cysteine concentration. b) Kinetic
constant kCC of [Eq. (3)] evaluated by competition.
To check whether radical transfer could be detected for
AAs that are incorporated into peptides, we prepared
peptides with nondeuterated and deuterated leucine residues
at their termini (see section 2 in the Supporting Information).
The leucine residues were separated from the rest of the
peptide by two (L GGL ) or three (L GGGL ) glycine
kinetics of hydrogen abstraction from alanine by methyl
5
À1 À1 [11]
radicals ((1.2 Æ 0.4) ꢀ 10 m s ). As expected from their
[
12]
lower (calculated) stability, the CC radicals that form in
H
D
H
D
residues. For both peptides, hydrogen and deuterium transfer
was identified respectively in the LD and LH residues
proline are more reactive than those that form in leucine and
glutamine. But this higher rate of deuterium transfer is not
associated with more efficient labeling, which suggests that
the transfer efficiency presented in Figure 2 is not directly
related to the kinetics of the transfer reaction, but rather
measures the capacity of the various AAs to sustain radical
reactions without fragmentation or recombination.
(
Figure 2).
For most of these transfer reactions, the fact that the
transfer efficiency is equivalent in [Eq. (2)] and [Eq. (3)] is
a strong indication that the initial formation of a CC radical by
attack of the hydroxyl radical can induce the sequence of
transfers in [Eq. (2)] and [Eq. (3)], and thus can form chain
The same competition method was applied to radical
transfer in peptides. The reaction in [Eq. (3)] is only margin-
ally faster for LGGL relative to the transfer to leucine. In
contrast, no competition was detected for LGGGL. An
increase in the efficiency of deuterium transfer was even
measured upon the introduction of cysteine. We can ration-
alize these observations by assuming that the transfer occurs
intermolecularly for LGGL and intramolecularly for
LGGGL. When intermolecular processes are faster than
intramolecular ones, cysteine acts directly as a competitor of
the labeling step and, thus, favors the formation of unlabeled
products (Scheme 1, top). When intramolecular processes are
faster than intermolecular ones, cysteine reacts with radicals
after their internal transfer. This late reaction of cysteine
increases the labeling by preventing the degradation of
labeled peptides after they have undergone a first hydrogen
abstraction (Scheme 1 bottom). The existence of this protec-
tion reaction is supported by the finding that hydrogen
labeling at deuterated sites increases upon the introduction of
cysteine (Figure S3 in the Supporting Information).
[
6]
reactions. The transfer efficiency is clearly related to the
number of hydrogen atoms that are present in the side chain
of the AAs, with a regular increase when going from glycine
to leucine. Such a dependence on the number of hydrogen
atoms that are available for the reaction is quite common in
[
7]
hydrogen abstraction reactions.
In a few cases we measured the kinetics of [Eq. (3)] by
using a competition strategy (see section 5 in the Supporting
Information). To do this, we measured the decrease in
deuterium transfer upon the introduction of an alternative
hydrogen transfer pathway, through the presence of free
cysteine in solution [Eq. (4)].
The kinetics of the reference reaction in [Eq. (4)] are not
precisely known. The repair of CC radicals by thiols usually
occurs with kinetic constants in the range of 10 –
Our results do not give direct clues about the origin of
these different behaviors of LGGL and LGGGL. However,
the introduction of an additional glycine residue is expected
À7
À8 À1 À1 [8]
1
0 m s , but the equivalent reaction to [Eq. (4)] for
6
À1 À1 [9]
[13]
alanine has a kinetic constant of only 5 ꢀ 10 m s . As we
had no reason to suppose that [Eq. (4)] for other AAs is very
different in efficiency, we used this constant for the reference
reaction to provide a minimal estimate for kCC.
to increase the degrees of freedom of the peptide, and to
[
14]
influence the kinetics of formation of short loops.
To get a more detailed picture of the regioselectivity of
1
hydrogen transfer to leucine, we conducted H NMR experi-
Angew. Chem. Int. Ed. 2012, 51, 2960 –2963
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2961