Application of H/D Exchange to Hydrogen Bonding
COMMUNICATION
Table 3. HDX exchange data and estimates of hydrogen-bond equilibri-
um for 9–23. All data were gathered at 10 mm, 228C by using 10%
CD3OD/CDCl3. All errors were less than 10% (see the Supporting Infor-
mation).
quite similar to unsubstituted phenylbenzamide 4, and the
slight changes were consistent with a smaller electronic
effect than that observed for para-methoxy derivatives 15
and 16. A cyano group would be expected to have a signifi-
cant electronic effect, but difficulty directing its hydrogen-
bond acceptor nitrogen towards the central amide. Unfortu-
nately, these strong electron-withdrawing derivatives ex-
changed in less than three minutes in 10% CD3OD/CDCl3,
and meaningful data could only be acquired in 1% CD3OD/
CDCl3. Under these conditions, both cyano groups produced
similar rates of H/D exchange, with the para-cyano 24 half-
life of 9.8 min, and the ortho-cyano 25 showing a slightly
shorter half-life of 3.2 min. The slightly faster rate of 25
shows the importance of choosing the appropriate controls,
and in this case, there may be a slight difference in the po-
larizing effects of substituents in the ortho- and para-posi-
tions of a benzene ring. As such, the data for these benzene
derivatives serve as estimates of hydrogen-bond interactions,
because we cannot be certain of the suitability of the con-
trols.
Substrate Rate constant
[minÀ1
Half-life
[min]
ÀRTln(kex/kfree
)
Control
]
[kcalmolÀ1
]
9
0.0074
0.0059
0.00071
0.0014
0.027
0.00053
0.21
0.029
0.10
0.74
0.48
N.d.
0.068
0.13
0.075
93
2.0
0.9
2.9
3.7
1.7
15
16
17
18
19
20
10
11
12
13
14
15
16
17
18
19
20
21
22
23
120
970
500
26
1300
3.4
24
6.7
0.94
1.4
<1
10
> 4
5.5
9.2
N.d.=not determined.
Similar protection from H/D exchange was observed with
carbonyl derivatives. Comparing esters 12 and 18, which
differ in ortho or para substitution, suggested that the six-
membered-ring hydrogen bond has a protection from H/D
exchange that correlated to a hydrogen-bond strength of ap-
proximately 3.7 kcalmolÀ1. A similar comparison of phenyl
ketones 13 and 19 pointed to a weaker hydrogen bond with
a strength of about 1.7 kcalmolÀ1. This weaker interaction
was expected due to the lower hydrogen-bond acceptor abil-
ity of the less electron-rich ketone oxygen, and substantiated
the ability of H/D exchange to make more quantitative esti-
mates of different functional groups. Direct comparison of
14 and 20 is difficult due to the extremely quick exchange of
the doubly activated 20, in which both substituents acceler-
ate the inherent rate of exchange. A bottom limit of 4 kcal
molÀ1 could be estimated based on the exchange of 20 being
faster than one minute. Although all the exchange data de-
scribed above were recorded in 10% CD3OD/CDCl3, the
extremely slow rate of exchange for derivative 14 allowed
analysis in more polar solvents as well. Exchange rates were
measured in up to 50% CD3OD/CDCl3, indicating that this
approach can be applied to more competitive solvents, if the
stability of the hydrogen bond permits.
These model compounds have shown the utility of H/D
exchange in the comparison of individual hydrogen bonds,
and this approach has proven to be useful in other contexts,
such as amino acids and peptidomimetics.[4] It is not any sin-
gular H/D exchange rate itself that indicates the nature of
the hydrogen bond, but rather that character weighed
against the inherent rate of exchange of the functional
group itself, without the possibility of hydrogen bonding.
This is most appropriate with the application of controls
that closely match the overall steric and electronic environ-
ment of the hydrogen-bonding group. With this in mind, H/
D exchange can be an effective tool for investigating the
presence and relative strength of hydrogen-bonding interac-
tions.
tence of hydrogen bonds. A complete analysis must consider
that the rate of exchange is also dependent on the inherent
rate of exchange of the functional group itself and requires
controls that reflect the particular local environment. Con-
trols 15 and 16 most closely match the electron donation of
the methoxy group without the possibility for intramolecular
hydrogen bonding. These two molecules showed exchange
behavior that was distinct from the unsubstituted 4, and
demonstrated the effect of the electronic environment on H/
D exchange (Figure 1). Electron donation to the benzoyl
ring accelerated the rate of H/D exchange, whereas electron
donation to the aniline ring slowed H/D exchange (Table 1).
Opposite effects were observed for electron-withdrawing
esters in 18 and 21.
This analysis allows us to compare the hydrogen-bonding
interactions in 9 and 10 (Table 1) by applying the approach
from Equations (1) and (2). Although both molecules dis-
played similar exchange kinetics, comparison with the ap-
propriate controls suggested that the six-membered ring hy-
drogen bond in 9 had a strength of 2.0 kcalmolÀ1, whereas
the five-membered ring hydrogen bond in 10 had a strength
of 0.9 kcalmolÀ1. Incorporation of both ortho-methoxy
groups as in 11 formed a bifurcated hydrogen bond, and
more significant protection from H/D exchange.[11] The most
appropriate control 17 contains two para-methoxy groups,
and not surprisingly showed an intermediate rate of ex-
change that is again consistent with one group increasing
the inherent rate of exchange, whereas the other group re-
duces it. Comparison of derivatives 11 and 17 suggested a hy-
drogen-bond strength of 2.9 kcalmolÀ1.
Additional controls were performed to gauge the exis-
tence of steric or polarizing effects of substituents. Although
exchange rates have been shown to be affected by the steric
environment,[4,8] the minimal role of steric factors in this
case was demonstrated by using controls 22 and 23 with
ortho-methyl groups. Both showed exchange rates that were
Chem. Eur. J. 2013, 19, 15101 – 15104
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
15103