Stereoselectivity of the hydrogen-atom transfer in benzophenone-tyrosine dyads: An intramolecular kinetic solvent effect

Article abstract of DOI:10.1002/chem.200802674

The solvent dependence of the intramolecular hydrogen-atom transfer (HAT) quenching in a pair of diastereomeric benzophenone-tyrosine dyads by nanosecond laser flash photolysis was addressed. Resolution of the transparent spectra after the flash reveal th

Full text of DOI:10.1002/chem.200802674

COMMUNICATION
DOI: 10.1002/chem.200802674
Stereoselectivity of the Hydrogen-Atom Transfer in Benzophenone–Tyrosine
Dyads: An Intramolecular Kinetic Solvent Effect
Gerald Hçrner,*[a] Gordon L. Hug,[a, b] Anna Lewandowska,[a]
Franciszek Kazmierczak,[a] and Bronislaw Marciniak[a]
Intramolecular quenching of excited states by successful
electron transfer (ET) or hydrogen-atom transfer (HAT) is
known to produce covalently connected radical–ion pairs[1]
or biradicals (BR) in high yields with a high selectivity.[2]
bond formation between the dyad and the solvent. A kinetic
model is proposed which accounts for the solvent depend-
ence and the intramolecular dynamics and that exhibits
strong parallels to the Ingold et al. kinetic solvent effect
(KSE)[6] in bimolecular HAT reactions (Scheme 1).
C
For instance, tyrosyl radicals (Tyr(O )) that are instrumental
in biochemical metabolism and pathogenesis,[3] can be ob-
tained by the intramolecular quenching of ketone-triplet
states by remote tyrosine. In a recent study we have provid-
ed evidence that the efficiency of the tyrosyl-radical forma-
tion by HAT in short peptide-bound benzophenone[Tyr
dyads (bp[Tyr) is generally governed by regiochemical and
by stereochemical constraints of the peptide chain.[4] Nota-
bly, the HAT rates and the degrees of stereoselectivity were
significantly solvent dependent. The high selectivity of an in-
tramolecular HAT reaction in a related bp[Tyr dyad has
been recently used to address the nature of the long-range
ET reactions of ribonucleotide reductase with external
[5]
C
Tyr(O ) radicals. This approach underlines the potential of
ketone[Tyr dyads as model systems for biologically relevant
questions of radical-site formation and radical transport.
In the current work we address the solvent dependence of
the intramolecular HAT quenching in a pair of diastereo-
meric bp[Tyr dyads 1 by nanosecond laser flash photolysis.
The stereoselectivity of the quenching of the triplet states 2
(Scheme 1) was studied in fifteen different solvents. As a
result, significant stereoselectivity, as measured by the ratio
Scheme 1. a) Chemical structure of the dyads 1a,b; b) reaction scheme
for the deactivation of the triplet excited state 2, and of the biradical 3.
Excitation of the bp chromophores of the dyads 1 with
355 nm pulses yields the excited triplet states 2 with unit
quantum yields within the duration of the laser flash. Reso-
lution of the transient spectra after the flash revealed that
C
C
of the HAT rate constants SD =k
(1b)/k
N
the biradicals 3 (bpH [Tyr(O )) are formed during the decay
of 2 by HAT from the Tyr residue with quantum yields close
to unity in inert solvents. Biradical decay cleanly returns the
ground-state dyads. Details of this process will be discussed
elsewhere. The dyad concentrations were such that contribu-
tions from bimolecular quenching of 2 by ground-state 1 re-
mained below 5%. In inert solvents the experimentally ob-
served rate constant of the triplet decay kD then simply
equals the rate constant of the intramolecular H-atom trans-
fer kH. In solvents with labile H-atoms the competitive HAT
from the solvents as an additional decay channel was taken
into account (see Supporting Information). The obtained ki-
netic data for the decay of 2a,b in fifteen solvents are sum-
specific solvent-solute interactions, that is, by hydrogen-
[a] Dr. G. Hçrner, Dr. G. L. Hug, Mgr. A. Lewandowska,
Dr. F. Kazmierczak, Prof. B. Marciniak
Faculty of Chemistry, Adam Mickiewicz University Poznan
Grunwaldzka 6, 60780 Poznan (Poland)
Fax : (+48)61-829-4367
E-mail: hoerner@amu.edu.pl
[b] Dr. G. L. Hug
Radiation Laboratory, University of Notre Dame
Notre Dame, IN 46556 (USA)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200802674.
Chem. Eur. J. 2009, 15, 3061 – 3064
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3061

marized in Table S1 (Supporting Information). The influ-
ence of the solvent on the value of kH (sꢀ1) that covers a
range of 106–108 for both compounds can be separated into
bulk effects and specific solvent effects.
One effect of the solvent is referred to the influence of
the bulk viscosity hsolv. In particular, plots of 1/kH are linear
with the viscosity, provided that different solvent classes are
analyzed separately. This is shown in Figure 1 for five alco-
hol solvents and three chlorohydrocarbons. Strikingly, ster-
eoselection between 1a and 1b seems to be insignificant in
chlorohydrocarbon solutions (SD <1.3ꢁ0.3; circles in
Figure 1), whereas alcohol solvents give rise to significant
selectivity (SD ꢂ3.0ꢁ0.3; squares in Figure 1).
Scheme 2. Kinetic schemes of the intramolecular H-atom transfer in a) a
two-step mechanism and b) a three-step mechanism including H-bonding
to the solvent.
should give rise to similar stereoselectivity. However, experi-
ments in other low-permittivity solvents like pyridine and
1,4-dioxane exhibit a significant stereoselection (3.0<SD <
4.0) together with a substantial decline of the overall reac-
tivity. This is in stark contrast to the results obtained for the
chlorohydrocarbons. Although we cannot rule out bulk sol-
vent effects on the molecular conformation, we attribute the
failure of equation of Equation (1) to adequately describe
the quenching data in alcohol solvents to neglected specific
solvent–solute interactions.
The importance of specific solvent–solute interactions as a
source of kinetic solvent effects is borne out by a linear free
energy relationship with a solvation-parameter model intro-
duced by Abraham et al. [Eq. (2)].[9] The solvent descriptors
are the solventꢁs dipolarity/dipolarizability p2H, the solventꢁs
effective hydrogen-bond acidity Sa2H, and the solventꢁs ef-
fective hydrogen-bond basicity Sb2H. The respective data are
compiled in Table S1 (Supporting Information). The rate
constants for HAT of 1a,b in twelve solvents (three highly
viscous alcohols were not considered) were subjected to
multi-linear regression in terms of Equation (2). The fit-pa-
rameters are given in Equations (3) and (4). Error analysis
in terms of reduced c2 gave values of 2.41 and 4.34 for 1a
and 1b, respectively. It is noted, that these values substan-
tially improve (reduced c2 ꢂ1.6) when data for ethanol and
benzonitrile are not considered. A plot of the fitted rate
constants versus the experimental rate constants is convinc-
ingly linear with a slope of 0.98ꢁ0.03 (Figure S1, Support-
ing Information).
Figure 1. Viscosity dependence of the reciprocal H-atom transfer rate
constants; squares: alcoholic solvents; circles: chlorohydrocarbon sol-
vents; lines: linear fits to the data.
Scheme 2a serves to rationalize the viscosity dependence
of the decay of 2 within a two-step kinetic model. Therein
k0H is the rate constant of the elementary HAT step. KDyn
=
kfDyn/kbDyn denotes the molecular-dynamics equilibrium be-
tween folded and extended conformations of the dyads (2fold
and 2ext). The rate constants of folding and unfolding, kfDyn
and kbDyn, carry a viscosity dependence. A similar model has
been found adequate to describe the viscosity dependence
of intramolecular quenching of excited tryptophan by
remote cysteine.[7]
1
kH
1
1
1
hsolv
f0
¼
þ
¼
þ
ð1Þ
KDyn ꢃ k0H kfDyn KDyn ꢃ k0H
k
H
H
H
Dyn
ð2Þ
ð3Þ
ln kH ¼ ln k0þsp2 þaSa2 þbSb2
H
In terms of Equation (1) the striking difference between
the intercepts of the “chlorohydrocarbon” and the “alcohol”
plots might be attributed to a substantial impact of the sol-
vent on the molecular conformation, that is, on the equilibri-
um KDyn. Such an effect on the equilibrium conformation is
well established for the side-chain rotamers of aromatic
amino acids, that are sensitive to the bulk solvent permittivi-
ty,[8] and has been recently found to apply also for a set of
structurally related dyads.[4] Provided that this is an impor-
tant factor, solvents of similar viscosity and permittivity
ln kH ð1 aÞ ¼ 18:50ðꢁ0:13Þꢀ0:15ðꢁ0:17Þp2
H
H
ꢀ0:69ðꢁ0:21ÞSa2 ꢀ4:86ðꢁ0:17ÞSb2
H
ln kH ð1 bÞ ¼ 18:62ðꢁ0:18Þꢀ0:15ðꢁ0:23Þp2
ꢀ0:41ðꢁ0:28ÞSa2 ꢀ3:06ðꢁ0:22ÞSb2
ð4Þ
H
H
The fits essentially predict identical HAT rate constants k0
in the absence of solvent effects for both dyads. The fit pa-
rameters s, a, and b are all of negative sign, that is, the inter-
3062
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 3061 – 3064

Benzophenone–Tyrosine Dyads
COMMUNICATION
action of the solute with a solvent induces a reactivity loss.
However, the effects of p2 and Sa2 are small so that the
ed by the availability and the uncertainty of the H-bond
equilibrium constants (box in Figure 3 highlights the uncer-
tainty in pyridine).[13] Irrespective of these limitations, the
H
H
reactivity loss and, equivalently, the increase of the free
¼
energy of activation DG is mainly attributable to H-bond-
ing of the dyad to the solvent, with the solvent being the ac-
ceptor. Interestingly, the dependence on the solventꢁs H-
bond basicity is significantly smaller for the (R,S)-diastereo-
mer 1b. This difference appears to be the dominant source
of the stereoselectivity SD =kHACTHNUTRGENN(GU 1b)/kHAHCTUNGTRNEN(UGN 1a), which is ob-
served for the decay of 2. In particular, a semilogarithmic
plot of SD is linear with Sb2H over the complete set of fifteen
solvents (Figure 2), irrespective of viscosity effects.
Figure 3. Plot of the H-atom transfer rate constants for 1a (squares) and
1b (circles) in terms of Equation (5); filled symbols: range of “thermo-
chemically” derived values for KHB, compiled in ref. [6]; open symbols:
“kinetically” determined values for KHB, taken from ref. [6]; lines: linear
trends of the “kinetically” derived data set; solvent code: 1=acetonitrile;
2=ethyl acetate; 3=methanol; 4=pyridine; 5=dimethylformamide.
slopes of the plots for 1a (0.40 ns) and 1b (0.09 ns) differ
considerably. Provided that solvent-dependent changes of
the dyad conformations are insignificant (i.e., KDyn ꢂconst.),
this finding points to different solvation of the diastereo-
meric dyads. Based on the current results, it cannot be de-
cided yet, whether the discriminating factor can be attribut-
ed to differential solvation of the triplet excited dyads 2a
and 2b before the HAT step and/or in the diastereomeric
transition states of the transfer reaction.
Figure 2. Stereoselectivity of the H-atom transfer in the dyads 1a,b as a
function of the solventsꢁ H-bond basicity; filled symbols: alcoholic sol-
vents.
The propensity of phenols to act as an H-bond donor is
well known,[10] so that the Tyr-phenol moiety can be as-
sumed to be an important source of the specific interaction
with the solvent. Accordingly, a specific ground-state solva-
tion of Tyr has been observed for related compounds.[4,11]
The loss of reactivity in strongly H-bonding environments is
in agreement with earlier observations in the bimolecular
HAT from phenols.[6] We therefore attribute this retardation
in H-bond accepting solvents to the reversible masking of
the phenol by the solvent S in an HB-equilibrium with
1
1
KHB½Sꢄ
¼
þ
ð5Þ
kfDyn KDyn ꢃ k0H
kH
K
HB =[PhOH···S]/ACHTUNGTRENNUNG([PhOH][S]). What results for the special
We conclude that the rates of the intramolecular HAT re-
action are delicately dependent on the H-bonding between
the H-atom donor and the solvent, which shields the donor
from reactive approach of the H-abstracting moiety. In the
present case of diastereomeric dyads, this intramolecular ki-
netic solvent effect induces solvent-dependent stereoselec-
tion (1.2<SD <4.0).
case of strong H-bonding is a three-step kinetic Scheme for
the intramolecular quenching of 2 (Scheme 2b), with the dif-
ferent triplet species
vated/folded), and
sumption of fast H-bond equilibration of
ext
fold
2
(solvated/extended)
2
(sol-
solv
fold
solv
2
(unsolvated/folded). With the as-
free
fold
2
and
solv
fold [12]
2
,
the observed quenching rate constant in this
free
model is given by Equation (5). A detailed derivation can
be found in the Supporting Information. Equation (5) pre-
dicts a linear relationship between the reciprocal HAT rate
constant and the equilibrium constant for the H-bonding of
the phenol to the solvent (multiplied by the solventꢁs molar-
ity [S]) as is implied also by Ingold et al. KSE model for bi-
molecular HAT reactions.[6]
Experimental Section
The dyads 1a,b were synthesized by standard techniques of peptide cou-
pling. Synthetic procedures and characterization of the compounds are
given in the Supporting Information. The laser flash photolysis equip-
ment has been described in detail previously.[14] The excitation light
source was the 355 nm output of a Nd/YAG laser (7–9 ns fwhm; 3–7 mJ
pulseꢀ1). Spectral analysis and data processing have been described in
detail before.[4]
As shown in Figure 3, this prediction is satisfyingly fulfil-
led for five solvents of similar viscosity. The analysis is limit-
Chem. Eur. J. 2009, 15, 3061 – 3064
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3063

G. Hçrner et al.
[4] G. Hçrner, G. L. Hug, D. Pogocki, P. Filipiak, W. Bauer, A. Groh-
mann, A. Lꢃmmermann, T. Pedzinski, B. Marciniak, Chem. Eur. J.
2008, 14, 7913–7929.
[5] S. Y. Reece, M. R. Seyedsayamdost, J. Stubbe, D. G. Nocera, J. Am.
Chem. Soc. 2007, 129, 8500–8509.
[6] a) J. T. Banks, K. U. Ingold, J. Lusztyk, J. Am. Chem. Soc. 1996, 118,
6790–6791; b) D. W. Snelgrove, J. Lusztyk, J. T. Banks, P. Mulder,
K. U. Ingold, J. Am. Chem. Soc. 2001, 123, 469–477; c) G. Litwinien-
ko, K. U. Ingold, Acc. Chem. Res. 2007, 40, 222–230.
Acknowledgements
This work was supported by COST Chemistry CM0603. G.L.H. was a
Fulbright Scholar at AMU. The authors thank Prof. O. Brede and Dr. R.
Hermann (University Leipzig, Germany) for technical support with the
LFP equipment and Prof. A. Grohmann (TU Berlin, Germany) for the
leave of synthesis facilities. This paper is Document No. NDRL-4791
from the Notre Dame Radiation Laboratory, which is supported by the
Office of Basic Energy Sciences of the US Department of Energy.
[7] a) L. J. Lapidus, P. J. Steinbach, W. A. Eaton, A. Szabo, J. Hofricht-
er, J. Phys. Chem. B 2002, 106, 11628–11640; b) R. R. Hudgins, G.
Gramlich, W. M. Nau, J. Am. Chem. Soc. 2002, 124, 556–564; c) F.
Huang, W. M. Nau, Angew. Chem. 2003, 115, 2371–2374; Angew.
Chem. Int. Ed. 2003, 42, 2269–2272.
Keywords: biradicals
photochemistry · solvent effects
·
hydrogen bonds
·
ketones
·
[8] J. Kobayashi, T. Higashijima, T. Miyazawa, Int. J. Pept. Protein Res.
1984, 24, 40–47.
[9] a) M. H. Abraham, Chem. Soc. Rev. 1993, 22, 73–83; b) M. H. Abra-
ham, C. F. Poole, S. K. Poole, J. Chromatogr. A 1999, 842, 79–114.
[10] E. M. Arnett, L. Joris, E. Mitchell, T. S. S. R. Murphy, T. M. Gorrie,
P. v. R. Schleyer, J. Am. Chem. Soc. 1970, 92, 2365–2377.
[11] M. H. Abraham, R. J. Abraham, J. Byrne, L. Griffiths, J. Org. Chem.
2006, 71, 3389–3394.
[12] J. Zheng, M. D. Fayer, J. Am. Chem. Soc. 2007, 129, 4328–4335.
[13] The terms “thermochemical” and “kinetic” refer to the experimen-
tal source of the equilibrium constants: The former are obtained
from direct IR-spectroscopic or calorimetric evaluation of H-bond
formation of the binding partners, typically as dilute solutions in an
inert solvent. The latter are quantified by the effect of an added H-
bond acceptor on the bimolecular rate constants of a HAT reaction
from phenol in the inert solvent CCl4.
[1] Recent examples: a) H. Imahori, K. Tamaki, D. M. Guldi, C. P. Luo,
M. Fujitsuka, O. Ito, Y. Sakata, S. Fukuzumi, J. Am. Chem. Soc.
2001, 123, 2607–2617; b) J. W. Verhoeven, H. J. van Ramesdonk,
M. M. Groeneveld, A. C. Benniston, A. Harriman, ChemPhysChem
2005, 6, 2251–2260; c) H. J. van Ramesdonk, B. H. Bakker, M. M.
Groeneveld, J. W. Verhoeven, B. D. Allen, J. P. Rostron, A. Harri-
man, J. Phys. Chem. B 2006, 110, 13145–13150; d) A. Gouloumis,
D. Gonzales-Rodriguez, P. Vasquesz, T. Torres, S. Liu, L. Echegoy-
en, J. Ramey, G. L. Hug, D. M. Guldi, J. Am. Chem. Soc. 2006, 128,
12674–12684.
[2] a) M. A. Miranda, A. Martinez-Manez, F. Bosca, J. V. Castell, J.
Perez-Prieto, J. Am. Chem. Soc. 1999, 121, 11569–11570; b) J.
Pꢂrez-Pieto, A. Lahoz, F. Bosca, R. Martinez-Manez, M. A. Miran-
da, J. Org. Chem. 2004, 69, 374–381; c) A. Moretto, M. Crisma, F.
Formaggio, L. A. Huck, D. Mangion, W. J. Leigh, C. Toniolo, Chem.
Eur. J. 2009, 15, 67–70.
[14] R. Hermann, G. R. Mahalaxmi, S. Jochum, S. Naumov, O. Brede, J.
Phys. Chem. A 2002, 106, 2379–2389.
Received: December 18, 2008
[3] J. Stubbe, W. A. van der Donk, Chem. Rev. 1998, 98, 705–762.
Published online: February 16, 2009
3064
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 3061 – 3064