Head-to-tail interactions in tyrosine/benzophenone dyads in the ground and the excited state: NMR and laser flash photolysis studies

FULL PAPER

DOI: 10.1002/chem.200800315

Head-to-Tail Interactions in Tyrosine/Benzophenone Dyads in the Ground

and the Excited State: NMR and Laser Flash Photolysis Studies

Gerald Hçrner,*^{[a, b]}Gordon L. Hug,^{[a, c]}Dariusz Pogocki,^[d]Piotr Filipiak,^[a]

Walter Bauer,^[e]Andreas Grohmann,^[f]Anica Lämmermann,^[a]Tomasz Pedzinski,^[a]and

Bronislaw Marciniak^[a]

Abstract: The formation of head-to-tail

contacts in de novo synthesized benzo-

phenone/tyrosine dyads, bp[Tyr, was

probed in the ground and excited trip-

let state by NMR techniques and laser

flash photolysis, respectively. The high

affinity of triplet-excited ketones to-

wards phenols was used to trace the

geometric demands for high reactivity

vent the efficiencies of the intramolec-

ular hydrogen-atom-transfer reaction

depend strongly on the properties of

the linker: rate constants for the intra-

molecular quenching of the triplet state

cover the range of 10⁵to 10⁸s^À1. The

observed order of reactivity correlates

to a) the probability of close contacts

(from molecular-dynamics simulations)

and b) the extent of the electronic

overlap between the p systems of the

donor and acceptor moieties (from

NMR). A broad survey of the NMR

spectra in nine different solvents

showed that head-to-tail interactions

between the aromatic moieties of the

bp[Tyr dyads already exist in the

ground state. Favourable aromatic–aro-

matic interactions in the ground state

appear to correspond to high excited-

state reactivity.

in the excited state.

A retardation

effect on the rates with increasing hy-

drogen-bond-acceptor ability of the

solvent is correlated with ground-state

masking of the phenol. In a given sol-

Keywords: ketones

·

peptides

·

photochemistry · pi interactions ·

tyrosine

Introduction

The radical chemistry of the underlying phenol moiety of

Tyr was and is the subject of intense research activity.^[2]

C

Radical centers in proteins have gained considerable interest

in the past decades due to their roles as functional units or

as mediators of pathological pathways. In particular, the ty-

One well-established path giving Tyr(O ) radicals in vitro

is based on photoinduced reactions between tyrosine and ar-

omatic ketones. Triplet-excited ketones are biradicaloid in

nature and, thus, share many reactivity characteristics with

the alkoxy radicals.^[3]Numerous studies identify H-atom

transfer between the phenol and the triplet state as the

C

rosyl radicals, Tyr(O ), as the main oxidation product of ty-

rosine after formal loss of one hydrogen atom, have been

identified to be an integral cofactor of enzyme activity.^[1]

[a] Dr. G. Hçrner, Dr. G. L. Hug, Dr. P. Filipiak,

Dipl.-Chem. A. Lämmermann, Dr. T. Pedzinski, Prof. B. Marciniak

Faculty of Chemistry, Adam MickiewiczUniversity Poznan

Grunwaldzka 6, 60-780 Poznan (Poland)

[d] Prof. D. Pogocki

Faculty of Chemistry, Rzeszow University of Technology

35-959 Rzeszow (Poland)

and

Fax : (+48)61-829-4367

E-mail: hoerner@amu.edu.pl

Institute of Nuclear Chemistry and Technology

03-195 Warszawa (Poland)

[b] Dr. G. Hçrner

[e] Prof. W. Bauer

Interdisciplinary Group Time Resolved Spectroscopy

University of Leipzig, Permoserstrasse 15

04303 Leipzig (Germany)

Institute of Organic Chemistry

University of Erlangen-Nürnberg

91056 Erlangen (Germany)

[c] Dr. G. L. Hug

[f] Prof. A. Grohmann

Radiation Laboratory, University of Notre Dame

Notre Dame, IN 46556 (USA)

Institute of Chemistry, Technical University Berlin

10623 Berlin (Germany)

Supporting information for this article is available on the WWW

under http://dx.doi.org/10.1002/chem.200800315 or from the author.

Chem. Eur. J. 2008, 14, 7913 – 7929

ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

7913

dominate quenching process in mono- and bimolecular sys-

tems.^[4–7]Irrespective of the electronic nature of the lowest

excited triplet state (n!p* vs p!p*), H-atom transfer

from phenols occurs with high efficiency (usually close to

unity) and at much higher rates than observed for aliphatic

alcohols. High efficiency of radical intermediate formation

is usually contrasted by very inefficient product formation.^[4]

Thus, the excited-state energy is dissipated through consecu-

tive H-atom-transfer steps that regenerate the starting mate-

rial in the electronic ground state.^[8]We used the high affini-

ty of phenols towards triplet-excited ketones to trace the

steric demands of the formation of reactive conformations

in peptide-linked benzophenone/Tyr dyads.

The mechanism of the H-atom-transfer quenching of aro-

matic triplet states by phenols is dependent on the medium.

Clear evidence for an electron-transfer-initiated two-step

process is limited to aqueous media.^[9]For organic media,

the mechanism is interpreted in terms of a proton-coupled

electron transfer.^[7a]Importantly, the involvement of charge-

transfer (CT) complexes or exciplexes in H-atom-transfer

reactions has been invoked in several studies concerned

with bi- and monomolecular ketone/phenol systems.^[5b,d,7d]In

a recent example, the formation of encounter complexes be-

tween the phenol and the excited ketone benzoylthiophene

with a p!p* lowest triplet has been concluded from non-

linear Stern–Volmer plots and DFT calculations.^[6b,e]

Scheme 1. a) Folding of the dyads and b) steric constraints of substitu-

tion.

acids and (S)- and (R)-tyrosine methylester, Tyr-OMe. The

flexibility of the alkyl linkage is tuned by the relative posi-

tion and orientation of a rigid amide bond between the two

chromophores (Scheme 1b). The benzamides 1a and 1b

differ in the bp-substitution pattern, and the alkylamides 1c

and 1d differ in the stereochemistry of the Tyr a-carbon

atom. The structures of the dyads and of the primary amides

2a–c that were synthesized as reference compounds are

summarized in Scheme 2. Theoretical data on the ground-

Our interest in this topic began with the question of

whether, and if so, how the accessible conformational space

in ketone/phenol dyads could be modulated by tuning the

flexibility and directionality of the linkage between the in-

teracting groups in the ground state and in the excited state.

The few studies dealing with the effect of steric constraints

on monomolecular reaction rates in the ketone/phenol

system revealed remarkable effects of substitution geometry

on quenching rates.^[5c,d,6d]In addition, significant stereoselec-

tivity^[6a,c]of the quenching process has been reported for

monomolecular quenching of ketone triplets by phenols co-

valently linked by flexible spacers. Such constellations were

shown to allow the approach of the interacting groups on a

time scale much faster than the H-atom transfer itself.^[5a,b]

Both the formation of intramolecular complexes and the

H-atom-transfer step itself require close contacts between

the remote aromatic moieties (Scheme 1a). We show that

the ease of formation of head-to-tail conformations in ben-

zophenone/tyrosine, bp[Tyr, dyads is affected by the molec-

ular structure, that is, by the relative orientation and the dis-

tance between the reacting moieties, and the nature of the

spacer SP. The question arising is whether steric conditions

that are favourable for electronic overlap of the reacting

partners in the excited state can be probed by ground-state

interactions between the remote aromatic moieties also in

the ground state. The effects of molecular geometry on the

ground-state conformations and on the rates and efficiencies

of the intramolecular triplet quenching in ketone/phenol

dyads are compared.

Scheme 2. Structures of the dyads 1a–d and the monochromophores

2a–c.

state conformations of 1a–d from molecular-dynamics simu-

lations (Langevin dynamics) were used to trace the effects

of steric constraints on the probability of head-to-tail inter-

actions between the remote groups in the ground state. Ex-

tended NMR-spectroscopic studies yielded information

about the distribution of Tyr side-chain rotamers, the solva-

tion of the phenol and the amide functions, the possibility of

NOE contacts, and intramolecular aromatic–aromatic inter-

actions.

We used the four dyads 1a–d, bp[Tyr, which were synthe-

sized de novo from three benzophenone (bp) carboxylic

7914

www.chemeurj.org

ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2008, 14, 7913 – 7929

Tyrosine/Benzophenone Dyads

FULL PAPER

The probabilities of head-to-tail contacts are compared

with the effects of the dyad structures on the rate constants

and selectivity of the intramolecular H-atom-transfer reac-

tion in the triplet-excited state. Mechanistic features of the

overall UV-light-induced reactions, that is, transient life-

times, were studied by nanosecond laser flash photolysis

(LFP). The determination of quantum yields of the primary

reactions is based on spectral resolution techniques. Inter-

molecular and intramolecular quenching rate constants were

obtained by LFP and the efficiencies of irreversible reaction

pathways were tested by steady-state UV-light photolysis. In

addition, medium effects were studied in a selection of sol-

vents of similar viscosity, but of different polarity: dichloro-

methane (CH₂Cl₂, nonpolar), acetonitrile (CH₃CN, polar

nonprotic), and methanol (MeOH, protic).

Figure 1. a) Newman projections of the (S)-Tyr side chain along the

a

b

À

C

C bond with R=4-phenol; the three possible staggered rotamers are

Results and Discussion

Ground-state structures and conformations

¹H NMR studies

assigned as the two gauche- (gÀ with c¹=À608 and g+ with c¹=+608)

and one antiperiplanar (t with c¹=À1808) conformations with respect to

the relative orientation of the highest priority substituents -NH- and R,

respectively. b) Assignment of hydrogen atoms in the bp (3/4-substitu-

tion) and Tyr moieties of bp[Tyr dyads.

a) Amide proton resonances: The chemical shifts of the

amide-proton resonances, d (NH), their temperature coeffi-

cients, dd(NH)/dT, and the constants for vicinal coupling

bonding, and are taken as a strong argument against aggre-

gation. The same argument should hold for the much lower

concentrations of the dyads employed in the photochemical

part of the study (<2 and <0.1 mm at 355 and 266 nm, re-

spectively).

The amide resonances of 1a–d are split into well-resolved

doublets in all cases by coupling with the protons on the a-

carbon atom of Tyr. The underlying vicinal coupling con-

with the a protons, ³J(H^aNH), were used to trace the confor-

ACHTREUNG

mations of the peptide backbone and the presence of elec-

1

tronic effects. H NMR spectra of the dyads 1a–d in CDCl₃

reveal significant differences in the amide-resonance pat-

tern, which can be related to the respective chemical nature

of the benzamide (1a,b) and alkylamide (1c,d) linkers of

the dyads. Similar differences are observed for the mono-

chromophoric amides 2a–c. The chemical shifts, d (NH),

range from 5.9 ppm for the alkylamides 1c,d and 2c to

6.9 ppm for the benzamides 1a,b and 2a,b in CDCl₃. The

strong downfield shifts in 1a,b and 2a,b by 1 ppm reflect the

electronic-withdrawing effect of the p system of benzophe-

none on the amide group due to conjugation of the amide

group with the bp-aromatic system. Notably, this withdraw-

ing effect is transferred also through the Tyr side chain of

the benzamides 1a and 1b. With respect to 1c,d, the H^aand

H^bresonances of 1a,b are downfield shifted by +0.4 and

+0.2 ppm, respectively (see Figure 1a for labeling of the hy-

drogens).

The temperature coefficient of the amide-proton reso-

nance, dd(NH)/dT, is a valuable tool for tracing the hydro-

gen-bonding (HB) situation of peptides in solution. Intramo-

lecular HB of the amide group cannot be expected to be sig-

nificant for the structures in this study. However, intermo-

lecular HB due to aggregation might well be a significant

mechanism for the concentrations used in NMR (ꢀ25 mm).

Values of À7.1 and À7.6 ppbK^À1were obtained for 1c and

1d, respectively, in CDCl₃solutions in the temperature

range 223<T<313 K. These strong upfield shifts of the

amide resonances as temperature increases are in accord

with free amide protons that are not involved in hydrogen

3

stants, J

A

a

the N(H) C bond in the peptide backbone. These coupling

constants are widely used to infer the secondary structure of

peptides in solution.^[10]It was reported that most linear pep-

tides give rise to coupling constants higher than 7.0 Hz.

These values are not far from the coupling constants of

8.5 Hzin b sheets, which represent the prototype of “back-

bone-extended structures”. Note that the terms “extended

structure” and “folded structure” will be used below in the

discussion to describe the distance between the aromatic

moieties in the dyads, which do not necessarily refer to the

torsion of the backbone. In our case the dyads consistently

yielded values of 7.8Æ0.1 Hzfor the vicinal coupling con-

stants. No influence of the medium on the coupling con-

stants was observed during experiments in eight nonprotic

deuterated solvents. A dominance of backbone-extended

structures can be concluded for all of the dyads 1a–d in all

media under investigation. In contrast, the chemical shifts of

the amide resonances are strongly dependent on the proper-

ties of the solvent. Downfield shifts of >2.0 ppm were ob-

served in solvents such as DMSO, THF, and DMF, and

these are ascribable to H-bonding of the amide NH group

to the solvent.

b) Side-chain rotamer populations: The molecular structure

of the dyads can be clarified further through a study of their

Chem. Eur. J. 2008, 14, 7913 – 7929

ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

7915

G. Hçrner et al.

¹H NMR coupling constants. In particular, the vicinal cou-

pling constants between the diastereotopic b protons and

the

ACHTREUNG

a ACHTREUNG ) and

proton of the Tyr side chain, ³J(H^a,H^bR

³J(H^a,H^bS), are key in probing the rotamer populations of

the Tyr side chain (Figure 1a). With a view to the anticipat-

ed solvent dependence of the photochemical reactivity, the

spin systems were analyzed in a number of deuterated sol-

vents. Based on the Karplus equation,^[11]which correlates a

coupling constant with a dihedral angle, the vicinal coupling

constants, ³J(H^a,H^bR) and ³J(H^a,H^bS), were used to derive

ACHTREUNG ACHTREUGN

fractional populations p_gÀ, p_t, p_g+of rotamers in the tyrosine

side chain by using Equations (1–3) (see Figure 1b for label-

ing of the Tyr protons).^[12,13]The notation of the rotamers is

based on the dihedral angle c¹between the N-C^a-C^bplane

and the C^a-C^b-C^gplane of the Tyr side chain (the super-

scripts n of the dihedral angles cⁿdefine the position in the

amino acid side chain). Conformations gÀ, t, and g+ corre-

spond to c¹=À60, À180, and +608, respectively. Note that

the 1d configuration at C^ais reversed, that is, the assign-

ment of the rotamers g+ and gÀ is reversed to c¹= À60

and +608, respectively. The geminal protons H^bRand H^bS

(Figure 1a) were assigned as the high-field and the low-field,

respectively, part of the AB subspectrum at 3.0 ppm.^[14]

Figure 2. Solvent dependence of the averaged rotamer populations, p_i, of

*

1c,d (gÀ: , g+: &); filled symbols: data from ref. [18b] for Ac-Tyr-OEt;

dotted lines: linear trends; code: 1, [D₈]dioxane; 2, CDCl₃; 3, [D₈]THF;

4, CD₂Cl₂; 5, [D₆]acetone; 6, [D₃]CH₃CN; 7, [D₄]MeOH; 8, [D₇]DMF; 9,

[D₆]DMSO.

3

p_gÀ¼ ½³JðH^a,H^bRÞÀ J_g=D³J

ð1Þ

ð2Þ

ð3Þ

for all compounds in this study. In CDCl₃the computed dis-

tributions are, within the error limits, not far from a simple

uniform statistical weighting with the factor of about one

third for all compounds, although there appears to be a

slight overrepresentation of the rotamer g+. Similar rota-

mer populations are obtained in CD₂Cl₂solutions (solvent 4

in Figure 2; e_solv=8.93).^[17]In contrast, the same computation

procedure leads to a substantial loss in the statistical weight

for g+ and a concomitantly significant increase in the statis-

tical weight of gÀ if solvents of higher “polarity”, such as

CH₃CN (solvent 6 in Figure 2; e_solv=35.94)^[17]and MeOH

(solvent 7 in Figure 2; e_solv=32.66),^[17]are considered.

p_t¼ ½³JðH^a,H^bSÞÀ J_g=D³J

3

p_gþ¼ 1Àp_gÀÀp_t

In Equations (1–3), D³J=³J_tÀ J_g, in which J_tand J_gare the

nominal values of the coupling constants for the vicinal pro-

tons in the trans and gauche conformations, respectively.

3

Values of J_t=13.56 Hzand ³J_g=2.60 Hzwere used in the

3

calculations.^[15]Experimental coupling constants were de-

rived by simulation of the ABX spin systems with the Mest-

ReC program. With an error in ³J

ACHTREUNG

AHCTREUNG

The gÀ selectivity of aromatic amino acid side chains at

high solvent polarity is a general phenomenon irrespective

of the nature of the aromatic group.^[18–20]It has been associ-

ated with steric effects, however, a satisfying rationalization

for this is lacking. As can be seen in Figure 2, the computed

rotamer populations p_iof the alkylamides 1c,d (given as

mean values with an error Dp_i<0.02), follow a linear (R>

0.92) dependence on the solvent permittivity. A similar de-

pendence (R>0.98) holds for the benzamides 1a,b (some

deviation between the two sets is noted at low permittivity;

data not shown). In particular, the populations are correlat-

ed with a bulk property of the solvent. Our data are qualita-

tively in accord with data reported for tyrosine derivatives

in several solvents and solvent mixtures (filled symbols in

Figure 2).^[18b]However, in the cited work the rotamer popu-

lations of N-Ac-Tyr-OEt depended linearly on the decadic

logarithm of the permittivity, loge_solv. Because this study was

partly based on solvent mixtures, the discrepancy between

the data sets might be due to effects of preferential solva-

tion in solvent mixtures. Irrespective of this, all compounds

(also those from ref. [18b]) give very similar values of p_t

the resulting errors in rotamer populations are estimated to

be Æ0.03 and Æ0.05 for p_gÀ(p_t) and p_g+, respectively. The

measured coupling constants for 1a–d and the computed

fractional distributions for rotamers gÀ, g+, and t are sum-

marized in the Supporting Information, Tables S1 and S2.

Note that the intrinsic uncertainty of the population of g+

is large for small absolute values of p_g+. It has recently been

found that the chosen formalism of population derivation

tends to overestimate the population of g+ systematically

[16]

at low p_g+

.

All four bichromophores 1a–d in CDCl₃gave values for

3

³J(H^a,H^bR) and J(H^a,H^bS) of around 6.2 and 5.6 Hz, respec-

ACHTREUNG ACHTREUNG

tively (see Table S1). This invariance suggests that the chem-

ical environments of the H^bRand H^bSprotons in 1a–d are

closely related. Accordingly, only small differences between

chemical shifts of H^bRand H^bSjDd(H^bRÀH^bS)j ꢀ0.06–

R

0.09 ppm) are observed. This similarity of structures is fur-

ther corroborated by an analysis of the rotamer distribu-

tions. Thus, in CDCl₃(solvent 2 in Figure 2; e_solv=4.81),^[17]

there appears to be no expressed preferential conformation

7916

www.chemeurj.org

ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2008, 14, 7913 – 7929

Tyrosine/Benzophenone Dyads

FULL PAPER

that are not solvent dependent (data not shown). Essentially,

the summed population of the conformers g+ and gÀ,

which are capable of nominally short distances, that is,

which fulfil one requisite of reactive contacts between the

aromatic moieties (t, in contrast, does not, see Figure 1a) is

independent of the solvent.

c) Chemical shifts of aromatic protons: Steric conditions that

are favourable for p–p interactions between the remote aro-

matic moieties in the ground state might be reasonably

taken to be consonant with favourable conditions for elec-

tronic overlap of the reacting partners also in the excited

1

state. Thus, the H NMR pattern of the aromatic protons of

1a–d in CDCl₃was used to trace the effects of the substitu-

tion geometry on the possibility and extent of intramolecu-

lar interactions. Both distinct subsections of the region of ar-

omatic protons at 7.3–8.2 ppm (bp) and 6.5–7.0 ppm (Tyr)

exhibit evidence for such interactions.

1

Figure 3. Region of Tyr aromatic proton resonances of the H NMR spec-

tra (200 MHz, CDCl₃) of 1d (upper trace), 1c (middle trace), and 1a

(lower trace); asterisk: amide-proton resonance.

One singlet signal assigned as the H²resonance of bp (see

Figure 1b for numbering of the protons in the bp rings) is

affected selectively in the presence of Tyr. A subtle high-

field shift from 7.77 for 2c to 7.68 for 1c and 7.64 ppm for

1d is observed. The remaining bp subspectra of 1c and 1d

are convincingly reproduced by the model compound 2c.

Selective shielding of the H²proton is also observed for the

benzamide 1a. However, the introduction of a second car-

bonyl substituent to the meta position of the benzamides 1a

and 2a significantly affects also the resonances for H⁴, H⁵,

and H⁶of bp (see Figure 1b for numbering of the bp rings).

Two contrary effects can be separated: 1) Chemical shifts as-

sociated with the resonances of H², H⁴, and H⁶for 2a are

significantly shifted to lower field by +0.2<Dd<+0.4 ppm

with respect to the alkyl-amide substituted 2c. This shift is

attributed to the additional withdrawing effect of the second

carbonyl group. On the other hand, 2) the presence of the

Tyr residue in 1a causes a high-field shift of the resonance

of H²by À0.15 ppm with respect to the monochromophore

2a. For the para-substituted 1b and 2b no significant differ-

ences between the bp subspectra are observed.

The characteristic “pair of doublets” pattern of the

AA’BB’ spin system of the Tyr moiety in CDCl₃yields valid

additional evidence for the interactions within the bp[Tyr

dyads (Figure 3). The assignment of H^dand H^e(see Fig-

ure 1b for labeling of the Tyr protons) as the low-field and

high-field pseudo-doublets, respectively, was based on NOE

experiments. The observed effects are limited to the keto-

profen-based dyads 1c,d; 1) selective shielding of aromatic

Tyr proton resonances and 2) decreasing chemical-shift dif-

ferences between the H^dand H^eresonances of Tyr. The

dyads 1c,d experience a significant high-field shift of their

H^dand H^eresonances with respect to 1a,b and free Tyr, for

which values of d(H^d)=7.16 ppm and d(H^e)=6.85 ppm have

been reported.^[21]For 1c,d the meta protons H^dare signifi-

cantly shielded by À0.15 and À0.35 ppm, respectively, rela-

tive to the respective chemical shifts in 1b. Weaker shielding

is observed for the ortho protons H^ewith À0.03 and

À0.15 ppm for 1c and 1d, respectively, relative to the re-

spective chemical shifts in 1b.

Similarly affected Tyr subspectra were reported for other

Tyr/ketone dyads in CDCl₃, but no explanation was given.^[6a]

A rationalization of the upfield shifts could invoke intramo-

lecular H-bonding between the bp and the Tyr moieties in

weakly HB-accepting solvents such as CDCl₃. However, the

observed chemical shifts of the resonances at 6.1 and

6.0 ppm for the acidic phenol protons of 1c and 1d in

CDCl₃rule out any significant contribution from intramolec-

ular H-bonding. H-bonding usually gives rise to strong low-

field shifts of the phenol proton resonance to d>10 ppm.^[22]

Strong downfield shifts of the resonances of the phenol hy-

droxyl group were actually observed in HB-accepting sol-

vents. These solvent-dependent shifts are fully in accord

with the linear dependence of the hydroxyl chemical shifts

on the HB-acceptor ability of the solvent.^[23]In addition, an-

alogues of 1c,d in which the phenolic functions are protect-

ed as methoxy groups show NMR spectra in CDCl₃that are

very similar to the spectra of their parent compounds.^[24]

However, it is well known that aromatic amino acids are

susceptible to aromatic–aromatic (A–A) interactions.^[25]Il-

lustrative cases of strong A–A interactions have been re-

ported for neighboring aromatic residues in a cyclic dipep-

tide and across the strands of a b-hairpin mimic. Here, the

proton H^dof Tyr is located inside the anisotropic cone of an-

other aromatic side chain connected to a high-field shift of

the resonance frequency even beyond the one of H^e.^[26]It is

suggested that the observed shielding of H², H^d, and H^ein

1c,d also reflects a “ring-current” effect due to close con-

tacts between the aromatic moieties of 1c,d (see Figure 3).

The interpretation is corroborated by the results of NOE ex-

periments (see below) and by the solvent dependence of the

NMR pattern of the aromatic Tyr protons.

d) Aromatic–aromatic interactions—Medium und steric ef-

fects: Interactions between the aromatic moieties were

probed in a study of the aromatic-proton resonance pattern

in the NMR spectra of 1a–d in several deuterated solvents.

Chem. Eur. J. 2008, 14, 7913 – 7929

ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

7917

G. Hçrner et al.

Two parameters were chosen to analyze the effects on the

chemical shifts of the aromatic protons: 1) The spectral

width of the Tyr AA’BB’ spin system of each dyad,

bonding between the phenol and the solvent is taken from

the medium effect on the chemical shift of the phenolic-

proton resonance. A substantial deshielding of this reso-

nance by >3.0 ppm is observed in highly HB-accepting sol-

vents like DMSO and THF. The H-bonding increases the

electron density on the phenolic oxygen^[2h]and, consequent-

ly, the electron density on the ortho and para carbon atoms

of the phenol. This electronic effect can be read from the

shielding of the ortho protons (H^e) (e.g., Dd(H^e)=

À0.11 ppm for 1b) upon going from CDCl₃to [D₆]DMSO.

The concomitant deshielding of H^dby 0.10 ppm indicates a

decreased electron density at the meta carbon atoms.

jDd

(H^dÀH^e)j=jd(H^d)Àd(H^e)j, and 2) ring-current effects

G

(rce) that focus on single aromatic protons (H^dor H^e) and

are defined as the difference between the chemical shift of

one of these protons in a reference compound (i.e., without

A–A interactions) and the chemical shift of an analogous

proton in a dyad under study: Dd

ACHTREUNG

G

(dyad)Àd-

cant overlap of the aromatic moieties in 1b (see below), this

dyad was chosen as a reference compound.

As mentioned above, the Tyr AA’BB’ spectral widths ob-

tained for the alkylamides 1c,d in CDCl₃are substantially

smaller than the ones of the benzamides 1a,b. Interestingly,

this general order of spectral widths (1aꢀ1b@1c>1d) is

conserved throughout a survey of nine solvents with greatly

varying properties. The spectral widths (Figure 4) correlate

The contributions of the solvents and the dyadsꢁ structures

to the observed spectral widths can be separated. We sug-

gest that the dyad-structure-controlled A–A interactions

manifest themselves in the variation of the intercepts in

Figure 4. This conclusion is further corroborated if ring-cur-

rent effects, rce, on the chemical shifts of H^dand H^e, as de-

fined above, are considered (Figure 5). In fact, the obtained

rce values depend only weakly on the solvent (slope ꢀ70Æ

10 ppb), indicating similar probabilities for head-to-tail in-

teractions in all solvents. This is in accord with the previous-

ly discussed invariance of the summed population of the ro-

tamers g+ and gÀ (see discussion of Figure 2).

H

to the parameter b₂that characterizes the solventꢁs ability

to act as an acceptor for H-bonding. This quantity with

values from 0.00 for non-HB-accepting solvents to 1.00 for

strongly HB-accepting solvents was defined by Abraham

and co-workers in order to construct a Gibbs-energy-related

HB-basicity scale.^[28]As was suggested recently,^[7f]a correct-

ed value of b₂^H=0.15 for CH₂Cl₂is used in our study.

Figure 5. Solvent dependence of the ring-current effects, rce (see text), of

1a (circles), 1c (squares), and 1d (triangles); full and open symbols

denote the aromatic Tyr protons H^dand H^e, respectively; asterisk: data

in [D₄]MeOH; lines correspond to linear fits to the experimental data;

experimental uncertainty of the rce is estimated to be Æ0.02 ppm.

Figure 4. Solvent dependence of ¹H NMR spectral widths of the AA’BB’

~

!

*

spin system of 1a ( ), 1b (&), 1c ( ), and 1d ( ); number code as in

Figure 2; the experimental uncertainty of the spectral widths is estimated

to be Æ0.02 ppm.

Structure-dependent variation in the absolute values of

the rce dominates the plots. The alkylamides 1c,d generally

show higher rce values than the benzamides 1a,b. In addi-

tion, the A–A interactions induce site-specific ring-current

effects in 1c,d. The meta protons H^dof Tyr appear to be

closer to the anisotropic cone(s) of one or both of the bp-

phenyl moieties than the ortho protons H^e. This conclusion

was further addressed by NOE experiments (see below). Im-

portantly, the ring-current effects are significantly different

for the epimeric dyads 1c and 1d. Thus, the probability of

head-to-tail contacts between the aromatic moieties, that is,

The spectral widths of the Tyr resonances of 1a–d obey

linear dependencies on the solventsꢁ HB-acceptor abilities.

Significant deviations from linearity are solely observed in

MeOH solutions (number 7 in Figure 4; b₂^H=0.41) as the

only protic solvent under study. Interestingly, the dyads 1a–

d yield common slopes but dyad-specific intercepts. The

H

structure-independent slopes as b₂

increases (slope

ꢀ260 ppb) can be explained by specific interactions of the

phenol hydroxyls with the solvents. Direct evidence for H-

7918

www.chemeurj.org

ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2008, 14, 7913 – 7929

Tyrosine/Benzophenone Dyads

FULL PAPER

the electronic overlap of the chromophores, is affected by

the regiochemistry (as in 1a and 1b) and the stereochemis-

try of the chromophores.

Instructive NOE signals are obtained upon saturation of

the Tyr H^dresonances (d(H^d)=6.80 and 7.0 ppm in CDCl₃

and CH₃CN, respectively; negative signals in Figure 6). In

CDCl₃solutions, contacts of the Tyr H^dwith H atoms on

both benzophenone rings are evident (Figure 6a’). The Tyr

H^dprotons clearly interact with the H²/H⁴and H⁸/H^8’pro-

tons of the bp moiety. These interactions of the Tyr phenol

ring with the remote bp phenyl rings are suggested to be the

source of the observed high-field shift of the H^dresonance

of 1c,d in CDCl₃(Figure 3). As previously discussed, the

ring-current effects in 1c,d (Figure 5) are expressed more

for the H^dprotons than for the H^eprotons. This site-selectiv-

ity of ring-current effects could be interpreted as a conse-

quence of the H^dprotons being at a closer distance to the

bp-ring system than the H^eprotons.

e) NOE spectroscopy—Intramolecular folding: Nuclear

Overhauser enhancement spectroscopy, NOESY, supple-

ments the information available from usual NMR experi-

ments. In addition to scalar-coupling phenomena, NOESY

yields information on the coupling of nuclear spins through

space. NOESY is performed by saturating the NMR fre-

quency of a selected nuclear spin and measuring the en-

hancements, I_NOE, in the intensities at the resonance fre-

quencies of neighboring nuclear spins. The key feature of

NOESY of interest to this investigation is that the measured

I_NOEare a function (I_NOE/r^À6) of the distances, r, between

the saturated nuclear spin and its respective neighbors.

1D-NOESY experiments in CDCl₃were undertaken for

1a and 1c,d in order to address the structural peculiarities

of the interaction between the aromatic moieties. Similar

NOE patterns were obtained for both alkylamide dyads

Very similar NOE intensities are observed after saturation

of the H^eresonances of 1c,d. This finding points to a similar

distance distribution in both compounds. Thus, ring-current

effects and NOE intensities that describe the geometry of

remote aromatic moieties have to be taken as complementa-

ry parameters.

1

1c,d. Because the above H NMR spectra showed aromatic

interactions and, hence, indicated folding of the dyads, NOE

signals that are indicative of short contacts between the aro-

matic Tyr and other parts of the molecule are of special in-

terest.

This conclusion is in accord with the observed NOEs of

benzamide 1a, that exhibits moderate and strong NOEs

with H²/H⁴on saturation of the H^dand H^eresonances, re-

spectively. The strong NOEs are contrasted by negligible

A strong NOE is observed

for the interaction of Tyr H^d

with the amide proton. This is

in accord with the computed

rotamer populations of 1c and

1d in CDCl₃(see Table S2 and

Figure 2), which exhibited

a

statistical weight of approxi-

mately 70% for the conforma-

tions (p_gÀ+p_g+) that allow for

small distances between NH

and the phenol part of Tyr (Fig-

ure 1a).

The amide itself strongly in-

teracts with the H²of bp. This

finding outlines substantial

folding of the dyads 1c,d. The

position that would normally be

suspect in any folding of the

moieties is the chiral aliphatic

linkage between bp and Tyr.

Apparently, folding of the mol-

ecule produces structures in

which the protons of the keto-

profen-based alkyl group and

the Tyr phenol point in oppo-

site directions. This is suggested

because, apart from trivial con-

tacts, NOE contacts of these

protons are limited to the

amide proton.

Figure 6. Left column: Schematic views of selected NOE contacts obtained for 1c in a) chloroform and b) ace-

tonitrile. Right column: Comparison of 500 MHz ¹H NMR (upper traces) obtained for 1c in aꢁ) chloroform

and bꢁ) acetonitrile and respective NOE spectra (lower traces) obtained after saturation of the H^dresonance

of 1c.

Chem. Eur. J. 2008, 14, 7913 – 7929

ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

7919

G. Hçrner et al.

ring-current effects (Figure 5). Due to the short linker be-

tween the chromophores in 1a, close contacts between the

aromatic systems are inevitable. Thus, the close contacts be-

tween the aromatic moieties as in 1a are not necessarily as-

sociated with effective geometric overlap of the p systems.

Additional NOE experiments on 1c in CD₃CN solution

(Figure 6bꢁ) show that the probability of close contacts be-

tween the remote aromatic moieties is greatly reduced in

this solvent of higher HB-acceptor ability. In particular,

there is no significant NOE between the Tyr H^dand the bp

H². In addition, the NOE between the amide proton and the

H²that was prominent in CDCl₃solution is missing in

CH₃CN. However, the selective shielding due to ring-current

effects was shown to be only weakly solvent dependent

(Figure 5). This necessarily means that the change from

chloroform to CH₃CN induces a change in the average

structure of 1c from a condensed (many NOE contacts in

chloroform) to a more extended one (few NOE contacts in

CH₃CN), which still allows for interactions between the aro-

matic moieties.

Figure 7. Pair-distribution functions of carbonyl–hydroxylic group dis-

tance r_O···O, obtained with umbrella sampling according to Equation (14)

for 1a (c), 1b (g), and 1c (d).

The pair-distribution functions in Figure 7 for 1a–c were ob-

tained after averaging three differential distributions that

were computed for three individual starting conformations

of the Tyr side chain, corresponding to gÀ, t, and g+, re-

spectively. For 1a,b the resulting three differential pair-dis-

tribution functions were very similar to each other. Thus,

the method can be safely assumed to sample efficiently the

accessible conformational space for these two compounds.

The situation appeared to be different for 1c,d (1d exhib-

its a distribution similar to 1c with a tendency to less-fav-

oured close contacts). Here, the individual pair-distribution

functions for each molecule differed significantly among

themselves, and the average distribution for 1c should only

be used for qualitative conclusions. Umbrella sampling ap-

pears not to cover the whole conformational space for these

two ketoprofen-based dyads; that is, the conformational

freedom of the (S)-2-phenylpropionylamide moiety extends

the range of possible ground-state minima that have to be

accounted for and sampled properly. A current study ad-

dresses the dynamics of 1c,d in further detail. Nevertheless,

the main results remain unaffected. The probability of short

contacts between the ketone and the phenol is strongly de-

pendent on the molecular structure. The presence of a rigid

amide bond in the para position of the benzamide 1b con-

fines the accessible space of the remote groups in a way

such that close contacts between the ketone and the phenol

are completely inhibited (g in Figure 7). Moving the

amide substituent to the meta position partly relieves the

steric constraints for close contacts (shown curve in

Figure 7). Furthermore, this effect appears to be enhanced

by increasing the flexibility of the linkage between the chro-

mophores on going from 1a to 1c (d in Figure 7).

In contrast, the NOE contacts observed for 1d remain to-

tally unaffected by this change in the solution medium. It is

concluded that the probability of the condensed structures

of 1d is preserved also in solvents of high HB-acceptor abili-

ty.

Langevin dynamics for distance distributions: The observed

close contacts between the two aromatic moieties point to

the possibility of conformers with close hydroxyl/carbonyl

contacts r_O···Oalso. Intrinsically, NOESY cannot supply

quantitative information on the distance between the phenol

hydroxyl and the bp carbonyl. However, the latter distance

can be investigated with molecular-dynamics simulations

(Langevin dynamics) using umbrella sampling for 1a–d. The

starting structures were obtained by steric energy minimiz-

ing. Sampling of 1c and 1d in a dielectric continuum with

the permittivity set to e=37 as a mimic of acetonitrile solu-

tion similarly gave a very narrow pair-distribution function

with a maximum at a distance r_O···O=2.7 . Such short dis-

tances are usually indicative of strong hydrogen bridges,

which were also observed in the solid state,^[6a]and are seen

as DFT ground-state minima for dyads that are structurally

closely related^[6d]to the ones in the current work.

1

However, H NMR of the hydroxyl protons at d<6 ppm

and <7 ppm rule out that there are any significant contribu-

tions from intramolecular hydrogen bridges at room temper-

ature both in CDCl₃and CD₃CN, respectively. Thus, results

from molecular dynamics are treated to be unrealistic in the

sense that they artificially overestimate the formation of hy-

drogen bridges.

To overcome the limitations of the methodology, the per-

mittivity of the dielectric continuum was set to infinity to

suppress the influence of Coulombic forces. Under these re-

straints, contributions from hydrogen bonding are effectively

inhibited. However, the resulting molecular-dynamics simu-

lation still yields close contacts of hydroxyl and carbonyl oc-

curring with significant probability in 1a and 1c (Figure 7).

Excited-state dynamics

Photophysical and photochemical properties—Steady-state

experiments: The photophysical properties of the dyads

were studied to trace possible effects of the substitution ge-

ometry on the electronic properties of the bp moiety. The

ultraviolet absorption spectra of the compounds 1a–d are

7920

www.chemeurj.org

ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2008, 14, 7913 – 7929

Tyrosine/Benzophenone Dyads

FULL PAPER

dominated by the intense absorption bands of the aromatic

chromophores (Figure 8). The respective band structure de-

pends on the position of ring substitutions and strongly re-

289 kJmol^À1that was found (this work) and reported^[30]for

benzophenone (Table S3). Note that in CH₃CN, the triplet

energies are systematically higher by approximately 3–

4 kJmol^À1. The triplet lifetimes of 1a,b and bp, obtained

from phosphorescence decays, are all about 6 ms (77 K). All

these emission characteristics are in accord with literature

data for bp derivatives.^[30]We conclude that the electronic

structure of the benzophenone unit is not affected by substi-

tution with tyrosine moieties; that is, the lowest-lying triplet

states possess n!p* character.

In contrast to the similar photophysical properties of the

dyads 1a–d, the steady-state photochemical behavior

showed significant differences between the benzamides 1a,b

and the alkylamides 1c,d. Steady-state consumption during

irradiation of solutions of 1a–d at 254 or 313 nm in CH₃CN,

MeOH, and CH₂Cl₂was quantified by following the disap-

pearance of the characteristic bp p!p* absorption at

250 nm by means of UV-visible spectroscopy and HPLC

with UV detection. No attempts were made to identify the

stable products. Quantum yields of disappearance (F_irr) for

313 nm irradiation, based on the oxalate actinometer, are

summarized in Table 1. In inert solvents (CH₃CN, CH₂Cl₂)

Figure 8. Absorption spectra of 1c (c) and 1d (b) and phosphores-

cence spectra of 1c, 1d, 1a, and bp (continuous lines) and 1b (dotted

line); inset: zoom on the n!p* absorption. Absorption spectra were ob-

tained at 258C in CH₃CN; phosphorescence was observed at 77 K in

methanol/ethanol (1:1) with concentrations of 10^À4–10^À3m.

Table 1. Summary of steady-state disappearance quantum yields, F_irr, of

1a–d.

sembles the spectra of the underlying benzophenone parent

compounds. The UV-visible spectra of all the compounds

are typical for meta- and para-substituted benzophenones

with a maximum absorption at 250 nm (in acetonitrile,

[a]

Compound

F_irr

CH₂Cl₂

CH₃CN

MeOH

1a

1b

1c

1d

0.004

0.049

0.006

0.004

0.038

0.036

0.001

0.37

0.23

0.074

0.008

e_bpACHTREUNG

(250 nm)=16000m^À1cm^À1). As usually observed for ben-

zophenones with an S₁state of n!p* character, the S₀!S₁

transition is only weakly allowed with molar absorption co-

efficients in the range of 10²m^À1cm^À1. A shoulder at 280 nm,

observed for the dyads 1a–d, is assigned to the absorption

of the Tyr moiety. Both the p!p* transition at 255 nm and

the weaker n!p* transition at 340 nm (in acetonitrile,

[a] Irradiation at 313 nm, [1]ꢀ10^À3m; experimental error Æ10%.

quantum yields for disappearance (F_irr) are generally much

lower than in MeOH, in which H-atom abstraction from the

solvent is an important process. Note that 1c shows a signifi-

cantly lower affinity towards solvent H-abstraction than do

1a,b, whereas 1d shows low F_irrin all three solvents. The

quantum yields obtained for 1a,b in CH₃CN depend strong-

ly on concentration. Quantum yields obtained upon irradia-

tion of 1a,b at 254 nm decrease by a factor more than ten

upon lowering the dyad concentration from 10^À3to 10^À5m.

Intermolecular processes are concluded to account for the

concentration dependence. The same concentration effect

was observed for 1b consumption in CH₂Cl₂. In contrast, 1a

was virtually inert in CH₂Cl₂solution.

e_bp

A

between the diastereomers 1c and 1d in solvents of differ-

ent polarity, for example, acetonitrile, methanol, chloroform,

benzene, and cyclohexane. A bathochromic shift of the p!

p* resonances and a hypsochromic shift of the n!p* ab-

sorption as solvent polarity increases are in accord with the

literature.^[29]

Phosphorescence emission spectra of 1a–d were recorded

in CH₃CN and methanol/ethanol (1:1) glasses at 77 K with

bp as a spectral reference. All of the compounds exhibit the

typical emission-band structure of benzophenone (Figure 8

and Table S3). The well-defined vibrational progressions

with an energy spacing of approximately 1700cm^À1are indi-

cative of the participation of the >C=O functionality in the

radiative process. The emission spectra of bp and all of the

meta-substituted derivatives 1a and 1c,d coincide, whereas

the spectrum for the para-substituted 1b is significantly red-

shifted in both of the media studied. The triplet energies of

all of the tested compounds, calculated from the wavelength

corresponding to the 0–0 emission in methanol/ethanol (1:1)

Flash photolysis: To study how the structural differences of

the dyads translate into their photochemical dynamics, time-

resolved laser flash photolysis experiments were performed.

Excitation of the strong p!p* transition (l_max=250 nm; e_bp

(250 nm)=16000m^À1cm^À1) or the much weaker n!p* tran-

(340 nm)=100–200m^À1cm^À1) of 1–2

sition (l_max=340 nm; e_bpCAHTREUNG

in CH₃CN with 266 or 355 nm pulses of a Nd:YAG laser

(FWHM 7–9 ns; 5–7 mJ) equally gives rise to strongly ab-

sorbing transients with similar absorption patterns immedi-

glasses, fall into

a narrow range between 284 and

290 kJmol^À1. These values are comparable to an E_Tof

Chem. Eur. J. 2008, 14, 7913 – 7929

ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

7921

G. Hçrner et al.

ately after the flash. Concentrations in the range of 4–8

10^À5m and 2–310^À3m were used at 266 and 355 nm, respec-

tively. The predominance of bimolecular-quenching process-

es observed for 1a and 1b (see below) suggested excitation

at 266 nm, with concomitantly lower dyad concentration, in

order to study intramolecular processes. For the other com-

pounds (and 1a in CH₂Cl₂), excitation at 266 and 355 nm

generally gave similar results with respect to reaction-rate

constants and spectral shape. However, spectra, obtained

with 355-nm photolysis at short delays, experience convolu-

tion with solvent Raman scattering in the spectral range of

380–400 nm. In the case of 266-nm photolysis, excitation of

both chromophores (bp and Tyr) cannot be avoided. Ac-

cordingly, 266-nm excitation produced time profiles at 300–

330 nm that exhibited significant but short-lived emission of

light from Tyr fluorescence. Based on molar-absorption co-

efficients measured for separate bp and Tyr moieties, the

fraction of light intensity exciting the bp chromophore at

266 nm is estimated to be >90%.

bichromophores 1a and 1b in acetonitrile. The transient

decays for these compounds are not connected with any

significant spectral evolutions.

4) Transient decays of 1b and of the monochromophores

2a–c in inert solvents (CH₃CN, CH₂Cl₂) follow mixed ki-

netics. It is suggested that the second-order component is

due to TT annihilation [Eq. (5)]. Clean first-order

decays, with lifetimes >5 ms, are observed upon a sub-

stantial decrease in laser power [Eq. (6)]. Phosphores-

cence emission from the lowest triplet state was observed

for the dyads 1a,b in CH₃CN solutions.

5) The transient absorptions of 1a,b and 2a–c are effective-

ly quenched in the presence of molecular oxygen.

All individual decay processes (Figure 9) convolute to

give a triplet-decay rate constant, k_d. At constant light inten-

sity and dyad concentration, the contributions from intrinsic

decay (k_T), self-quenching (k_SQ[bp[R]), and solvent-ab-

straction reaction (k_S[S]) can be summed to give k₀

[Eq. (11)]. Herein, the small contribution from TT annihila-

tion (k_TT[3bp[R]²) is assumed to be negligible at low laser

power.

a) Identification of transients and relaxation paths: The pri-

mary transients in all cases can safely be assigned as the

triplet states [Eq. (4)] (Figure 9) of the respective benzophe-

none moieties for the following reasons:

k_d¼ k_Tþ k_SQ½bp[R þ k_S½S þ k_H¼ k₀þ k_H

ð11Þ

The general results for the LFP of the dyads are exemplarily

discussed for the cases of 1c,d. For the bichromophores 1c

and 1d in any solvent, the decay of the initial signals is very

rapid, that is, shorter than the triplet decay of monochromo-

phoric 2c by a factor of >100. Triplet lifetimes (1/k_d) were

obtained from mono- or biexponential fits to the transient

decays at 600 and 510 nm, respectively. Transient decay is

biexponential at 510 nm due to overlap of the absorptions of

3

Figure 9. General reaction pattern of the photophysical and photochemi-

cal processes of the monochromophoric compounds 2a–c [Eqs. (4–7)];

R=NH₂) and dyads 1a–d [Eqs. (4–10)]; R=Tyr); S denotes a solvent

with abstractable hydrogen atoms.

C

bp and the ketyl radical, bpH (see below). Triplet lifetimes

are between 10 and 60 ns for 1c and near or below 10 ns for

1d (see Table 2 for kinetic parameters in CH₂Cl₂, CH₃CN,

and MeOH). The latter values are close to the resolution

limit of the laser set-up (because of the laser pulse width)

and are thus of qualitative value only. Shorter lifetimes in

both cases are observed in dichloromethane solution, where-

as the longest triplet lifetimes are observed in MeOH.

1) The small ratios of the observed transient absorbances

DA at 325 and 525 nm (values of 1.6:1) are in accord

with reported triplet–triplet (TT) absorption spectra of

bp derivatives.^[5b]

2) The molecular absorption

Table 2. Solvent dependence of triplet-quenching rate constants k_d, biradical-formation quantum yields F_BR

,

and biradical-decay rate constants k_BRobtained during the laser flash photolysis of 1a–d in deoxygenized sol-

vents at 355 nm.

coefficients of 1–2 at their

respective triplet absorption

maxima compare well with a

value of 6500m^À1cm^À1re-

[a,b]

[c]

[d]

k_d[10⁶s^À1

]

k_BR[10⁶s^À1

]

F_BR

solvent

CH₃CN

MeOH^[e]CH₂Cl₂

CH₃CN MeOH CH₂Cl₂CH₃CN MeOH CH₂Cl₂

3

ported for bp in acetonitrile

compound

solution.^[31]Note that for 1b

and 2b, that is, para-substi-

tuted benzophenones, both

absorption maxima are red-

shifted by 20 nm.

[g]

1a

1b

1c

1d

<0.5^[f](0.30)

5.9

5.5

37

ꢀ100

5.3 (3.2)

–

1.3

–

12.0

25.0

ꢁ0.10

<0.05

1.0

<0.05 >0.9

<0.05 <0.05

0.8₅>0.9

<0.3^[f](0.07)

<1.0^[f](0.80)

[g]

45

ꢀ100

80

>100

5.6

17.2

4.3

11.1

1.0

0.9₅>0.9

[a] From monoexponential fits to transient decays at 470 and 600 nm or biexponential fits at 510 nm with esti-

mated experimental error of Æ510⁶s^À1for 1c,d and Æ0.510⁶s^À1for 1a,b. [b] Values in brackets: intercepts

of Stern–Volmer dilution plots. [c] From monoexponential fits to transient decays at 405 nm with estimated ex-

perimental error of Æ0.510⁶s^À1. [d] Uncertainties are estimated to sum up to Æ0.05. [e] In MeOH bimolecu-

lar quenching by the solvent^[32]has to be taken into account [(Eq. (7)]. [f] Laser flash photolysis at 266 nm

3) Long lifetimes of the transi-

ents in the microsecond

range are observed for all

monochromophores and the (see text). [g] Not determined due to insignificant formation of biradicals.

7922

www.chemeurj.org

ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2008, 14, 7913 – 7929

Tyrosine/Benzophenone Dyads

FULL PAPER

The triplet decays of 1c and 1d were generally accompa-

nied by spectral shifts from 325 to 335 nm (not shown), 525

to 540 nm, and by growths at 405 nm (Figure 10). This spec-

Figure 11. Stern–Volmer plots of triplet-decay rate constants of 1a

(squares) and 1b (circles) in CH₃CN (open symbols) and CH₂Cl₂(filled

symbols); laser photolysis at 355 nm with 7-mJ pulse^À1

.

for k_SQ(ca. 310⁹m^À1s^À1in CH₂Cl₂and ca. 710⁸m^À1s^À1in

CH₃CN) remain well below the diffusion-controlled limit for

both of these solvents. These results are fully in accord with

numerous literature data on bimolecular quenching of

ketone triplets by phenols, both in regard to absolute val-

ues^[4,6b,7e]and solvent dependencies.^[7a,c,f]Interestingly, the in-

tercepts of the Stern–Volmer plots for 1a and 1b differ

greatly in CH₂Cl₂. The triplet-decay rate constants k₀are

not expected to differ markedly for these two dyads because

the triplet lifetimes of the monochromophoric compounds

2a and 2b were both about 1.5 ms at high dilution in

CH₂Cl₂. Thus, in Figure 11, the difference in the intercepts

for 1a and 1b in CH₂Cl₂has to be looked for in the rate

constants for intramolecular quenching, k_H[Eq. (9)]. In

methanol solutions the triplet lifetimes of 1a,b and 2a–c

were already significantly shortened to 160Æ30 ns at con-

centrations in the sub-millimolar range. No attempts were

made to study the self-quenching of 1a,b in MeOH.

The short, intrinsic triplet lifetimes of 1c,d, due to rapid

intramolecular quenching, prevented a direct study of the

self-quenching for these compounds. Intermolecular H-

atom-transfer rates were addressed instead by quenching of

the 2c triplet with either (R)- or (S)-configured Boc-Tyr-

OMe in CH₃CN (Boc=tert-butoxycarbonyl). Stern–Volmer

plots for both experiments yielded bimolecular-quenching

rate constants of 2.310⁸m^À1s^À1, which are in reasonable

agreement with the measured self-quenching rate constants

of 1a,b (see above). These rate constants are also very close

to the results obtained by Miranda et al. for the quenching

of a p!p* triplet by Boc-Tyr-OMe.^[6e]The rate constants

for the quenching of 2c with Boc-Tyr-OMe are somewhat

smaller than the self-quenching rate constants of 1a,b. This

might be rationalized by the presence of the bulky Boc

group, which could potentially hinder a reactive approach.

Nevertheless, an important conclusion can be drawn from

bimolecular-quenching experiments: irrespective of differing

molecular structure and concomitant differences in triplet

energies (Table S3), there are no significant differences be-

tween the reactivities of 1a–d toward intermolecular H-

atom abstraction from phenols.

Figure 10. Transient absorption spectra obtained during laser flash pho-

tolysis at 355 nm of deoxygenated solutions (210^À3m) of a) 1c and

b) 1d in dry CH₃CN; time delays after flash (from top to bottom): 17, 27,

50, 75 ns; initial triplet concentration [T]₀=55Æ5 mm.

tral evolution is in accord with fast triplet quenching due to

C

ketyl radical bpH formation. Concomitant formation of a

C

transient, identified as the tyrosyl radical Tyr(O ) based on

its characteristic double-peak structure at 385 and 405 nm,

clearly suggests there is H-atom transfer between a tyrosine

moiety and the excited triplet for both diastereomers

[Eqs (8,9)]. Furthermore, because the triplet decays of 1c,d

were virtually independent of their dyad concentrations, an

intramolecular H-atom-transfer mechanism is surmised for

their triplet-state quenching reactions in CH₂Cl₂and

CH₃CN. In CH₂Cl₂and CH₃CN transient absorptions decay

to zero within, at most, one microsecond for the dyads 1c,d,

whereas the overall transient decays were much slower in

MeOH solutions. A residual absorbance at 545 nm, with a

lifetime of tens of microseconds, points to the presence of

C

bpH formed through a different path [Eq. (7)]. A quantita-

tive analysis of the quenching process of 1c,d in CH₃CN sol-

utions is presented below.

b) Inter- versus intramolecular H-atom transfer: The dyads

1a,b were generally far less reactive than their ketoprofen-

based analogues. In contrast to 1c,d, their triplet lifetimes in

CH₃CN and CH₂Cl₂were strongly concentration dependent

(Figure 11). Linear Stern–Volmer plots were obtained for

both compounds. The observed spectral evolution during

triplet decay again is readily interpreted in terms of H-atom

C

transfer. Yields of bpH and Tyr(O ) were formed in equal

amounts and increased monotonously as dyad concentration

increased. Thus, the quenching mechanism is safely assigned

as being dominated by self-quenching [Eq. (8)].

The respective self-quenching rate constants k_SQ, obtained

from the slopes of the linear plots for 1a and 1b, do not

differ significantly from each other (Figure 11). The values

Chem. Eur. J. 2008, 14, 7913 – 7929

ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

7923

G. Hçrner et al.

To minimize contributions from self-quenching (see

above), transient spectra for 1a,b were also recorded after

laser excitation at 266 nm with substantially lower dyad con-

centrations. Representative spectra recorded for 1a in differ-

ent solvents are collected in Figure 12a. In CH₃CN the trip-

let state, with its characteristic absorption at 530 nm, is the

dominate species 200 ns after excitation (squares in Fig-

ure 12a). In MeOH solution the triplet decay is accompa-

nied by formation of a long-lived transient with absorptions

at 335 and 545 nm (350 and 575 nm for para benzophenones

1b and 2b), decaying on a timescale of several tens of mi-

croseconds with second-order kinetics. The spectral signa-

ture (triangles in Figure 12a) and the observed rate con-

stants allow us to make the assignment of the transient to

decay processes (Figure 9). In inert solvents such as CH₃CN

and CH₂Cl₂, the contribution from solvent reactions to the

decay constant k₀in Equation (11) can be neglected. How-

ever, in MeOH solutions, the rate constant for triplet-in-

duced H-atom abstraction from the solvent dominates k₀

and must be considered in order to calculate k_Hfrom the ex-

perimentally available k_dvalues. The required k₀values in

MeOH solutions were obtained as the rate constant of the

triplet decay of the monochromophores 2a–c in MeOH.

These k₀values ((5.5Æ0.5)10⁶s^À1) do not differ significant-

ly from one another and are in accord with literature data

for the triplet decay of bp in MeOH.^[32]

High selectivity towards H-atom transfer is a common

feature of triplet-excited ketone/phenol systems.^[4–7]Howev-

er, we are not aware of any reports of quantum yields for

the formation of biradicals in ketone/phenol dyads. To ad-

dress the selectivity of the intramolecular quenching reac-

tion, the transient spectra for 1c,d in CH₃CN in this study

were deconvoluted into spectral components. Over the com-

C

the ketyl radical bpH . Its formation is rationalized as an ab-

straction of a hydrogen atom from the solvent [Eq. (7)], and

second-order decay is consistent with disproportionation

and dimerization. The results for 1b are very similar to 1a

in methanol and CH₃CN, respectively. For 1a in CH₂Cl₂sol-

utions 200 ns after the laser flash (circles in Figure 12a), the

spectral region around 545 nm strongly resembles the one

obtained for 1a in MeOH. This part of the transient spec-

plete timescale of the transient decay only the spectral com-

3

C

ponents for bp, bpH , and Tyr(O ) are needed to simulate

the experimental data quantitatively (Figure 13; upper part).

Concentration/time profiles were constructed from spectral

resolutions for the reactions of 1c,d in CH₃CN (Figure 13;

lower part). The rate of formation of the radical species is,

within experimental error, identical to the triplet-decay rate

of 1c. For 1d the triplet decay is already strongly convoluted

with the excitation laser profile. These triplet decays are

C

trum is presumed to be due to bpH formation. An addition-

al feature at 405 nm is limited to the reaction of 1a in

CH₂Cl₂, and is indicative of an efficient formation of

C

Tyr(O ). No such transient could be observed for 1b in

CH₂Cl₂at comparable concentrations.

C

connected with formation of equimolar amounts of bpH

C

and Tyr(O ). Formation and decay of both radicals appear to

follow the same time law, which is in accord with an intra-

molecular-reaction mechanism. The formation of biradica-

C

loid intermediates bpH [Tyr(O ) [Eq. (9)] is analogous to in-

termediates that were reported for a number of related dy-

ads.^{[4,5a–d,6a,c,d]}The biradicals, BR, decay exponentially on a

longer timescale [Eq. (10)] than do the triplets with k_dꢀ5–

10k_BR. Note that the rate constants for triplet quenching

(k

_dA

G

(1d)/k

(1c)ꢀ2–3)

exhibit some diastereoselectivity, irrespective of the solvent.

Similar observations have been made with related

dyads.^[6a,c,d]

C

Quantum yields for bpH [Tyr(O ) formation, F_BR

(Table 2), are extracted from the concentration profiles in

Figure 13 by extrapolation back to the end of the excitation

pulse. These quantum yields for 1c,d in CH₃CN are not dif-

ferent from 1.00 if account is taken for the experimental

error, which is estimated to be not smaller than Æ0.05. Bi-

radical formation from 1c,d is generally efficient, irrespec-

tive of the solvent (Table 2). The somewhat lower quantum

yields observed in MeOH are readily explained by competi-

tive triplet quenching through H-atom abstraction from the

solvent [Eq. (7)]. This conclusion follows from the slight

Figure 12. a) Transient spectra obtained 200 ns after 266-nm (5 mJ) laser

pulsing of 1a in CH₃CN (squares), MeOH (triangles), and CH₂Cl₂(cir-

cles); b) Transient decay profiles (510 nm) on 266-nm laser pulsing of 1a

in CH₃CN (squares) and CH₂Cl₂(circles), and a profile of 1b (550 nm) in

CH₂Cl₂(filled symbols); the latter profile is scaled by a factor of 0.6 for

the sake of comparison; concentrations [1a,b]=810^À5m; initial triplet

concentration [T]₀=40Æ5 mm.

c) Quantitative analysis of intramolecular quenching: De-

pending on the molecular structure and the solvent, the life-

times of the excited triplet decays of 1a–d range from ap-

proximately 10 ns to >5 ms. The triplet-decay rate constants

k_dobtained from exponential fits to the experimental transi-

ent profiles are the sum of rate constants of the individual

C

non-stoichiometry between bpH and Tyr(O ) and from the

presence of a long-lived spectral component at 545 nm,

C

identified as bpH . Though quantum yields for biradical for-

mation are close to unity, the dyads 1c,d are very inert on

steady-state irradiation, as inferred from the negligible

7924

www.chemeurj.org

ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2008, 14, 7913 – 7929

Tyrosine/Benzophenone Dyads

FULL PAPER

modeling studies (see above),

which are assumed to apply

also in the triplet-excited state

of the dyads. Apparently, the

molecular structure of dyads

1c,d allows reactive contact of

the excited triplet state with the

remote phenol group of the Tyr

residue

with

high

yields

(Table 2) in any of the solvents

studied. In contrast, the triplet

states of the dyads 1a,b are

very inert in CH₃CN or are

C

transformed into bpH radicals

predominantly through H-atom

abstraction from the solvent in

MeOH. Although 1b remains

inactive towards intramolecular

processes also in CH₂Cl₂, 1a ef-

ficiently undergoes intramolec-

ular H-atom transfer to yield

C

biradicals, bpH [Tyr(O ). Clear-

ly, the approach of the remote

reaction sites depends on the

geometric constraints on intra-

molecular motions. The general

pattern of intramolecular reac-

Figure 13. Upper part: Spectral resolutions of transient absorption spectra taken a) 35 ns and b) 400 ns after

355-nm laser pulsing (7 mJ) of a solution of 1c; in (a) and (b) the squares represent the triplet-state bp, circles

3

C

the ketyl radical bpH , triangles the tyrosyl radical Tyr(O ), and * the experimental data; solid curves are the

resulting fits from the regression analysis. Lower part: Concentration profiles for the transients ³bp (squares),

C

bpH (circles), and Tyr(O ) (triangles), obtained from deconvolution of transient absorption spectra of c) 1c

and d) 1d after delays of 15–350 ns with a concentration [1c,d]=1.510^À3m; initial triplet concentration [T]₀=

38Æ2 mm.

tivity, k_H(1d)

> k_HACHTREUNG(1c) @

k

_HA(1a) @ k_HA(1b), as found in

CHTREUNG

CH₂Cl₂solution, can be qualita-

tively rationalized on grounds

quantum yields for dyad consumption in inert solvents

(F_irr<0.01; Table 1). It appears that both triplet quenching

and biradical decay are strictly intramolecular reactions at

sufficiently low concentrations of starting materials. This

type of “reversible” H-atom transfer presents an efficient

path for energy dissipation in the triplet-excited bp moiety.

of molecular-geometry considerations.

The dyads 1a–d are expected to be intrinsically less flexi-

ble than the dyads in ref. [5c]. This difference can be associ-

ated with the chemical nature of the linking group. Oxyalkyl

groups, as they were used in the cited work, are known to

increase chain flexibility and to favour folded structures.^[34b]

In contrast, the presence of a rigid amide spacer in our

dyads decreases the flexibility of the chain, even though the

chain is nominally even longer than in the cited work. In ad-

dition, the directionality of the amide function renders the

distribution of the conformations anisotropic. This situation

was already illustrated in Scheme 1b. In that simplified pic-

ture, the molecular dynamics were broken down into the

space accessible for the Tyr residue by mapping out the an-

gular limits of its amide nitrogen. The predictions from this

intuitive picture are fully confirmed by the results from the

molecular-dynamics simulations (Figure 7), in which it can

be seen that close contacts between the carbonyl and the

remote phenol group are not favoured for 1a and are strict-

ly excluded for 1b. In the ketoprofen-based dyads 1c,d, the

insertion of an alkyl group between benzophenone and the

amide group extends the chain length with respect to 1a,b.

The directionality of the precession cone from 1a (1208) is

preserved, however, the additional hinge in 1c,d greatly ex-

tends the opening angle of the cone (Scheme 1b). This an-

ticipated higher flexibility of the 1c,d dyads is convincingly

Effects of molecular geometry on the intramolecular H-atom-

transfer rates: Photochemical and photophysical parameters

of aromatic carbonyls have been referred to aspects of mo-

lecular geometry before.^[33]Hydrogen-atom transfer, as a

typical reaction of triplet-excited carbonyls, demands close

contacts of the reacting units.^[34a]Remarkable geometrical

effects on H-atom transfer in oxyethyl-linked acetophe-

none–phenol dyads have been reported.^[5c]In the cited work

the substitution pattern of ketone and phenol were system-

atically varied. Semiempirical (PM3) calculations revealed

that the flexible linker allows the formation of possibly reac-

tive exciplexes, irrespective of the substitution pattern. The

striking contrast between completely inert and highly active

regioisomers was accordingly referred to symmetrical re-

strictions of electronic overlap of phenol and ketone p sys-

tems.

It is appealing to interpret also the reactivity order of the

dyads 1a–d in terms of molecular geometry. Discussion is

based on the results of ground-state NMR and molecular-

Chem. Eur. J. 2008, 14, 7913 – 7929

ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

7925

G. Hçrner et al.

À

reflected in the r_O···Odistributions obtained from molecular-

dynamics simulations. Already within this simplified ap-

proach important qualitative conclusions can be drawn: For

1b, the geometrical constraints posed by concerted effects

of 1) unfavourable directionality of substitution and 2) the

lack of intrinsic flexibility of the linker combine to exclude

any close contacts between the carbonyl and the phenol

functions. Accordingly, 1b is not able to dissipate the excita-

tion energy by means of intramolecular H-atom transfer.

For 1a, condition (2) is still valid, whereas the change in di-

rectionality upon change in substitution (from para in 1b to

meta in 1a) allows for the close approach of the reacting

groups, at least to some extent (Figure 7).

In addition to effects of the probability of close contacts

(Figure 7) the orientation of the reacting moieties may

affect the overall reactivity, as was concluded in ref. [5c]. A

similar conclusion can be drawn from a discussion of the

steric constraints that have to be fulfilled for 1a to reach a

reactive conformation. In Scheme 3, the combined structural

the phenyl C(O) bond (not shown). Thus, only a minor

fraction of the molecules occupies conformations that are

reactive for H-atom transfer between hydroxyl and carbonyl

groups (Figure 7 and Scheme 3). However, triplet-excitation

of the dyad 1a yields the biradical BR with unity quantum

yield in CH₂Cl₂(Table 2). It is concluded that the conforma-

tional exchange between the active and unreactive rotamers

occurs on a timescale faster than the intrinsic triplet decay

[Eq. (6)]. This finding is in agreement with the results ob-

tained in similar systems.^[5a,b]

Importantly, neither of the two possibly reactive confor-

mations of 1a allow for significant enough overlap of the p

systems of Tyr and bp in order to form exciplexes. This is in

accord with the negligible ring-current effects on the aro-

matic Tyr protons of 1a (Figures 4 and 5). Vice versa, the

higher reactivity of the alkylamides 1c,d (and their reactivi-

ty order, k_HACHTRE(UNG 1d) > k_HAHCTER(UNG 1c)) corresponds to significant

ground-state interactions between the aromatic moieties.

Because exciplexes have been commonly proposed to be

crucial intermediates of H-atom

transfer in the ketone/phenol

system,^{[5a–d,6a,d,e,7a,d]}we conclude

that the low rate constants of

1a (relative to 1c,d) are a con-

sequence of the missing overlap

between the p systems of Tyr

and bp.

Intramolecular H-atom trans-

fer—Media effects: It can be in-

ferred from an inspection of

Table 2 that the rate constants

for triplet decay, k_d, and, con-

comitantly, the rate constants

for

intramolecular

H-atom

Scheme 3. Interchromophore distances (arrows) of carbonyl (white circle) and hydroxyl (grey circle) moieties

in 1a for the side-chain rotamers gÀ (A), g+ (B), and t (C).

transfer, k_H, of 1a and 1c are

dependent on the reaction

medium. In contrast, the H-

atom-transfer rate constants of

peculiarities of the benzamide group and of the peptide

backbone greatly confine the possible conformations in the

dyad 1a. 1) An extended conformation of the peptide back-

bone is strongly favored, as was concluded from the high

1d vary only slightly upon going from CH₂Cl₂to MeOH sol-

utions. Generally, the highest rates are observed in non-

polar solvents such as CH₂Cl₂. This might be attributed to

favourable conditions towards H-atom transfer being pres-

ent in solvents of lower “polarity”. This is corroborated by

the observation of highly reactive triplet states of 1c,d in

1

NH–H^acoupling constants in H NMR. The amide function,

C(O)–NH, and the C^a–C(O) group are almost co-planar.

2) In addition, the amide group of N-methyl-benzamide has

been reported to be tilted from co-planarity with the phenyl

p system by a well-defined angle of 208.^[35]

benzene and chloroform, or of 1a in n-butylchloride.^[36]

A

general result from related studies appears to be that there

is a correlation between the rate constants and the H-bond-

ing properties of the solvents. In particular, the commonly

observed decrease of H-atom-transfer reactivity as HB-ac-

ceptor ability increases is interpreted as being due to a re-

duction in the concentration of free phenol by hydrogen

bonding to the solvent.^[37]The concept of a solvent-depen-

dent “masking” of the phenol was recently applied to the

qualitative^[6b,7a,38]and quantitative^[7f]analysis of the solvent

dependence of the triplet quenching by phenols. The ob-

served proportionality between the chemical shifts of the

The consequences of rotation around c¹in the Tyr side

chain in 1a for the distance between the phenol (grey circle)

and the carbonyl (white circle) are illustrated in Scheme 3

(for the sake of visual clarity the small deviations from co-

planarity of the N-benzoyl amide and the peptide backbone

are neglected). The given situation affords close contacts be-

tween the phenol and the carbonyl only in the rotamer state

gÀ (structure 1a, A) in Scheme 3). Alternatively, the struc-

ture 1a, B) yields close contacts after rotation by 1808 about

7926

www.chemeurj.org

ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2008, 14, 7913 – 7929

Tyrosine/Benzophenone Dyads

FULL PAPER

phenolic protons and the HB-acceptor ability of the solvents

(see above) supplies direct evidence for an H-bonding equi-

librium also for the dyads 1a–d.

This simultaneous dependence on distance and orientation

is a strong argument for exciplexes being important inter-

mediates on the H-atom-transfer reaction coordinate, with

the stability of the respective exciplex being reflected by the

strength of the ground-state ring-current effects.

Our results are qualitatively in accord with the results of

the cited studies in that we observe a decrease in the rate

constants as HB-acceptor ability of the solvent increases.

However, the extent of the kinetic solvent effect (KSE) ap-

pears to be strongly dependent on the dyad structures. The

KSE that was observed for 1a strongly exceeds the ones for

1c and especially for 1d. This contradicts the model that

predicts uniform KSEs, expressed as the slopes in linear

plots of logk_Hversus b₂^H, for a uniform H-atom donor

(here: phenol). Interestingly, the observed order in KSEs

These results have implications on the interpretation of

the chiral discrimination that was reported for H-atom

transfer in related dyads.^[6a,c,d]In the cited work, the stereo-

selectivity was associated with “nonbonding electronic inter-

actions between the aromatic p systems”,^[6a]which was sug-

gested to be different for different diastereomers. The argu-

mentation was based mainly on X-ray crystallographic data.

In favour of this interpretation given by Miranda et al., the

epimeric dyads 1c and 1d in our study, with (S) and (R)-

configuration of the Tyr moieties, respectively, exhibit ste-

reoselectivity in the H-atom-transfer rates and in the proba-

bility of ground-state head-to-tail contacts. Thus, it is appeal-

ing to interpret the stereoselectivity in excited-state reactivi-

ty as a consequence of sterically controlled electronic over-

lap between the aromatic moieties.

(KSEACHTREUNG(1d) < KSEACHTREUNG(1c) ! KSEACHTER(UGN 1a)) appears to mirror the

order of ground-state interactions between the aromatic Tyr

and bp moieties, which were derived from the ring-current

effects on the Tyr resonances (Figures 4 and 5). Variations

in the solvent dependencies are indicative of structure-spe-

cific deviations from the assumptions of the Ingold model.

A possible correlation between the dyad structure and the

KSE serves as the working hypothesis for ongoing work that

addresses the solvent dependencies of H-atom transfer in

bp[Tyr dyads in further detail.^[36]

Experimental Section

General: Routine ¹H and ¹³C NMR spectra (200/500 MHzand

50.32 MHz, respectively) were recorded in deuterated solvents. Chemical

shifts d are reported in ppm downfield from TMS (experimental error:

DdÆ0.01 ppm). Assignment of resonances was based on ¹H–¹H and

¹H–¹³C COSY spectra. Parameters of the 1D-NOE measurements for 1a

and 1c,d in CDCl₃and CD₃CN (500 MHz): double PFG spin-echo

(DPFGSE)-NOE sequence,^[39]spectral width 4390 Hz, 16 k complex data

points, 2880 scans per irradiation point, acquisition time 3.7 s, relaxation

delay 2.0 s, NOE-buildup delay 500 ms, gradient strengths 14/6/4/À4%

(100%=140 Gcm^À1), gradient duration 1 ms each, selective Gauss pulse

40 ms, exponential line broadening with BF=1.0 Hz.

Conclusion

Hydrogen-atom transfer in the triplet-excited n!p* state of

bp[Tyr dyads is a prominent pathway of intramolecular trip-

let-energy dissipation. As the formation of reactive head-to-

tail contacts between the remote groups requires intramo-

lecular motions, it is expected to be modulated by the

nature of the linker. In fact, the H-atom-transfer efficiencies

of the dyads in this study depend markedly on the dyad

structure and the reaction medium. In this study we have

identified molecular parameters that control the overall

rates and efficiencies of the intramolecular reaction. In a

given solvent the directionality of and the flexibility within

the amide-type linker are selectors of the distance and the

orientation of the interacting moieties. A qualitative order

of reactivity has been established based on geometrical con-

siderations, NMR investigations, and molecular-dynamics

simulations, which address the distance distributions and rel-

ative orientations between the reacting groups as a function

of the molecular structure.

Within a series of three dyads (1a and 1c,d) that differ

solely in the nature of the amide-linking group between the

phenol and the triplet-excited ketone, the efficiency of the

reactive approach can be correlated with the probability of

close contacts and the extent of electronic overlap. Geomet-

rically restricted overlap, as evidenced by insignificant aro-

matic–aromatic interactions in the benzamide 1a, leads to

substantially smaller H-atom-transfer rates than in the alky-

lamides 1c,d, although similar distance distributions are ob-

served in all three cases. The high reactivity of 1c,d is paral-

leled by strong NMR ring-current effects that indicate effec-

tive head-to-tail interactions already in the ground state.

UV/Vis spectra were recorded by using a Varian Cary 300 Bio UV-visible

spectrophotometer in 10^À6to 10^À5molar concentrations. Millimolar con-

centrations were employed for the determination of the n!p* transitions

of benzophenone chromophores around 340 nm. Phosphorescence emis-

sion and excitation spectra and phosphorescence decays were recorded

by using a Spex FluoroMax-P (Horiba Jobin Yvon) in acetonitrile and a

Perkin–Elmer LS 50B in MeOH/EtOH 1:1 glass at an excitation wave-

length of 330 nm at 77 K. Concentrations were set in the millimolar

range, corresponding to absorbances A at 330 nm of 0.3–0.5.

Steady-state experiments were carried out in 1-cm1-cm rectangular UV

cells on standard optical-bench systems. A high-pressure mercury lamp

HBO 200 (Narva) together with water filter, quartzwindows, interfer-

ence filter (313 nm), and cut-off glass filters (<290 nm) were used as the

excitation source for 313 nm irradiations. A low-pressure mercury lamp

(Original Hanau TNN 15/30) was used as the excitation source for the

254 nm irradiation. Reaction progress was monitored by UV/Vis spec-

troscopy with a HP Diode Array 8452A Spectrophotometer or by HPLC

using a Waters 600E Multisolvent Delivery System Pump, as described

previously.^[40]The detection system consisted of a Waters 996 Photodiode

Array UV/Vis Detector. Analytical HPLC analyses were carried out on a

Waters XTerra RP₁₈reverse phase column (4.6250 mm, 5-mm particle

size). Uranyl oxalate actinometry was used to measure the light intensity.

The quantum yields for this actinometry were taken to be 0.602 and

0.561 for the 254 and 313 nm irradiations, respectively.^[30]Light flux was

within the range of 10^À4Einsteindm^À3min^À1at both wavelengths.

The equipment for laser flash photolysis has been described in detail pre-

viously.^[41]It is based on the 266 or 355 nm output of a Nd/YAG laser

with a full width at half maximum of approximately 7 or 9 ns and a dose

Chem. Eur. J. 2008, 14, 7913 – 7929

ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

7927

G. Hçrner et al.

[43]

^ÀC

of 5-mJpulse^À1or 7-mJpulse^À1, respectively. Transient decays were re-

corded at individual wavelengths by the step-scan method with a step dis-

tance of 5 nm in the range of 300 to 800 nm and obtained as the mean

signals of 6 to 16 pulses. Spectral resolution was in the range of Æ3 nm.

Samples for LFP were rigorously deoxygenated by flushing with analyti-

cal-grade nitrogen in a circulating flow system (55-mm Suprasil quartz

cell) for 20 min prior to and kept under nitrogen during measurement.

known spectrum of bp was used in the regression. The tyrosyl radical

C

Tyr(O ) was represented in the regression analysis by a spectrum mea-

sured in water by pulse radiolysis.^[44a]The LFP spectra in our study were

recorded at substantially lower resolution, which leads to dampening of

C

the very narrow absorption feature of Tyr(O ) at 405 nm. Thus, the molar

C

absorption coefficient reported in the cited work for Tyr(O ) (e_Tyr

(405 nm)=3200m^À1cm^À1) appeared to be inadequate. A value of e_Tyr

(405 nm)=1900m^À1cm^À1for acetonitrile solution, which is in accord with

Synthesis: Chemicals and solvents (Sigma–Aldrich and BACHEM) for

synthesis were of highest available analytical grade and were used with-

out further purification. Synthesis of the dyads followed the routines of

carbodiimide-induced amide-coupling (for details of procedures and char-

acterization, see Supporting Information). The identity and purity of the

[44b]

C

literature data for a protein-bound Tyr(O ),

was extracted from bimo-

lecular quenching experiments of bp with N-Boc-Tyr-OMe, based on

C

equimolar formation of Tyr(O ) and bpH .

Computational methods: The statistical distribution of the distances

(pair-distribution function), r_O···O, between the carbonyl oxygen atom of

the bp moiety and the Tyr hydroxylic oxygen was calculated from the po-

1

dyads were studied by H and ¹³C NMR and combustion analysis. The ob-

tained NMR spectra were in accord with the molecular structures of 1a–

d. Combustion analysis gave slight deviations from ideal stoichiometry

for the benzamide 1b due to inclusion of residual solvent (diethyl ether,

ethyl acetate) in the crystalline foams, which could not be removed even

on extended evacuation. Attempts to produce the dyads as crystalline

material were not successful. However, because HPLC analysis proved

the absence of contaminants absorbing light with l>220 nm in all cases,

the dyads were used in this study without further work-up.

tential of mean force, w

(r_O···O), by using Equation (14).^[45]

ACHTREUNG

gðr_OÁÁÁOÞ ¼ expfÀwðr_OÁÁÁOÞ=k_BTg

ð14Þ

These calculations were accomplished with umbrella sampling,^[46]done in

conjunction with Langevin dynamics (LD)^[47–49]and the CHARMM^[50]all-

atom empirical potential. In the LD no explicit solvent molecules were

used, but the usual LD frictional drag of the solvent was factored in by

using a collision frequency for all heavy atoms equal to approximately

20 ps^À1, which reflects the viscosity of acetonitrile. We set the sampling

Spectral analysis of LFP intermediates: Quantum yields for the forma-

tion of the ith transient, F_i, were obtained by using relative actinometry

with benzophenone solutions of matched optical density in acetonitrile,

CH₃CN [Eq. (12)]. The absorbed light intensity I_ais expressed as the re-

sulting triplet-state concentration in the actinometer solution with the

measured DA_Tat 525 nm. The known values of molar absorption coeffi-

windows at intervals of 0.1 and a force constant for the umbrella po-

tential of 20 kcalmol^À1

(83.7 kJmol^{À1 À2}). Each window simulation

À2

was run with a 1.5-fs time step for 50 ps, preceded by 9 ps of heating

from 0 to 300 K and 10 ps of equilibration. For the CHARMM27 force-

field calculations, we used version 7 of the HyperChem PC molecular-

modeling package.^[51]To suppress unrealistic over-representation (see

above) of short distances, r_O···O, with carbonyl and hydroxyl in hydrogen

bonding, the atom charges of all the atoms of the molecules were set to

zero which, in practice, is the equivalent of setting the dielectric permit-

tivity equal to infinity.

cient e_T

ACHTREUNG

(525 nm)=6500m^À1cm^À1and triplet quantum yield F_T(bp)=1.00

were also used in Equation (12).^[31]

DAðl_jÞ

eðl_jÞ

c_i

I_a

F_TðbpÞ Â e_Tð525 nmÞ

DA_Tð525 nmÞ

F_i¼

¼

Â

ð12Þ

Here, DA(l_j) and e(l_j) denote the measured change in optical density

after laser pulsing of the sample solution, and the molar absorption coef-

ficient of the transient at the wavelength l_j, respectively. The procedure

summarized in Equation (12) is limited to cases in which wavelengths

exist that are characteristic for a single transient. In the more common

case of convoluted spectra, a multi-regression analysis has to be done on

the optical transient spectra resulting from pulsed irradiation in order to

extract the individual transient concentrations c_i. Within any time

window, following the excitation pulse, the absorbance of the signal is re-

lated to the concentrations and molar absorption coefficients of the tran-

Acknowledgements

This work was supported by the Research Training Network SULFRAD

(HPRN-CT-2002-00184) and by COST Chemistry CM0603. The authors

thank Prof. O. Brede and Dr. R. Hermann (University Leipzig, Germa-

ny) for access to the LFP equipment and for technical support. The com-

putation was performed employing the computation resources of the In-

terdisciplinary Centre for Mathematical and Computation Modelling of

the Warsaw University (G-24-13) and the Albigowa Biotechnology

Centre of the Rzeszow University of Technology. This paper is Document

No. NDRL-4751 from the Notre Dame Radiation Laboratory. G.L.H.

was a Fulbright Scholar at AMU.

sients through Beers Law DA=logCAHTREU(GN I₀/I).

n

X

DAðl_jÞ ¼

c_iÂ e_iðl_jÞ Â l

ð13Þ

i¼1

In the regression analysis of the experimental spectra by Equation (13)

the concentrations of the individual transients times the optical path-

length, lc_i, are the regression parameters to be fit.^[42]The sets of e_i(l_j)

are the reference spectra of the underlying transients enumerated by the

i-th subscript. The uncertainties for the c_iare computed from the square

roots of the diagonal matrix elements of the covariance matrix for each

linear regression.

[1] a) J. Stubbe, W. A. van der Donk, Chem. Rev. 1998 , 98 , 70 5 –7 62;

b) C. W. Hoganson, C. Tommos, Biochim. Biophys. Acta 2004, 1655,

116–122.

[2] For representative examples, see: a) G. W. Burton, K. U. Ingold,

Acc. Chem. Res. 1986 , 19 , 19 4 –2 01; b) C. Evans, J. C. Scaiano, K. U.

Ingold, J. Am. Chem. Soc. 1992 , 114 , 458 9 –4 593; c) J. M. Mayer,

D. A. Hrovat, J. L. Thomas, W. T. Borden, J. Am. Chem. Soc. 2002,

124, 11142—11147; d) M. Sjçdin, S. Styring, H. Wolpher, Y. Xu, L.

Sun, L. Hammarstrçm, J. Am. Chem. Soc. 2005, 127, 3855–3863;

e) I. J. Rhile, T. F. Markle, H. Nagao, A. G. DiPasquale, O. P. Lam,

M. A. Lockwood, J. M. Mayer, J. Am. Chem. Soc. 2006 , 128 , 607 5 –

6088; f) O. Brede, S. Naumov, J. Phys. Chem. A 2006 , 110 , 1190 6 –

11918; g) M. Sjçdin, T. Irebo, J. E. Utas, J. Lind, G. Merenyi, B.

Akermark, L. Hammarstrçm, J. Am. Chem. Soc. 2006, 128, 13076–

13083; h) F. Thomas, O. Jarjayes, H. Jamet, S. Hamman, E. Saint-

Aman, C. Duboc, J.-L. Pierre, Angew. Chem. 2004 , 116 , 60 4 –6 07;

Angew. Chem. Int. Ed. 2004 , 43 , 59 4 –5 97.

Reference spectra for the excited triplet states of the bp[Tyr dyads 1a–d

were obtained by LFP of diluted solutions of the model compounds 2a–c

in CH₃CN and dichloromethane, CH₂Cl₂. Molar absorption coefficients

were obtained by comparison of the transient absorptions DA

ACHTER(UGN l_max) and

DA_TA(525 nm) of the actinometer solution at time delays for which no sig-

CHTREUNG

nificant decay has taken place, and under the assumption of unity triplet

yield.^[31]Spectra of radical species relevant to this study were obtained by

bimolecular quenching of the respective triplet states of 2a–c in CH₃CN

with methanol, MeOH, or 2-propanol. Spectra of the resulting ketyl radi-

C

cals (bpH , also: hemipinacol) were corrected by scaling with known

^ÀC

molar absorption coefficients. Because ketyl anion radicals (bp ) turned

out to be of minor importance in this study, no attempts were made to es-

tablish a set of reference spectra for these species. If needed, a literature-

7928

www.chemeurj.org

ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2008, 14, 7913 – 7929

Tyrosine/Benzophenone Dyads

FULL PAPER

[3] P. K. Das, M. V. Encinas, S. Steenken, J. C. Scaiano, J. Am. Chem.

Soc. 1981 , 103 , 416 2 –4 166.

[4] P. K. Das, M. V. Encinas, J. C. Scaiano, J. Am. Chem. Soc. 1981 , 103 ,

415 4 –4 162.

[25] a) E. A. Meyer, R. K. Castellano, F. Diederich, Angew. Chem. 2003 ,

115 , 124 4 –1 287; Angew. Chem. Int. Ed. 2003 , 42 , 121 0 –1 250; b) R.

Chelli, F. L. Gervasio, P. Procacci, V. Schettino, J. Am. Chem. Soc.

2002 , 124 , 613 3 –6 143.

[26] a) M. Sheinblatt, J. Chem. Soc. Perkin Trans. 2 1990 , 12 7 –1 32; b) R.

Mahalakshmi, S. Raghothama, P. Balaram, J. Am. Chem. Soc. 2006 ,

128 , 112 5 –1 138.

[27] V. Rizzo, H. Jäckle, J. Am. Chem. Soc. 1983 , 105 , 419 5 –4 205.

[28] M. H. Abraham, P. L. Grellier, D. V. Prior, J. J. Morris, P. J. Taylor, J.

Chem. Soc. Perkin Trans. 2 1990 , 52 1 –5 29.

[29] N. J. Turro, Modern Molecular Photochemistry, University Science

Books, CA, USA, 1991.

[30] S. L. Murov, I. Carmichael, G. L. Hug, Handbook of Photochemistry,

Marcel Dekker, New York, 1993.

[5] a) J. C. Scaiano, W. G. McGimpsey, W. J. Leigh, S. Jakobs, J. Org.

Chem. 1987 , 52 , 454 0 –4 544; b) W. J. Leigh, E. C. Lathioor, M. J. St.

Pierre, J. Am. Chem. Soc. 1996 , 118 , 1233 9 –1 2349; c) E. C. Lathioor,

W. J. Leigh, M. J. St. Pierre, J. Am. Chem. Soc. 1999 , 121 , 1198 4 –

11992; d) E. C. Lathioor, W. J. Leigh, Can. J. Chem. 2001 , 79 , 1851;

e) E. C. Lathioor, W. J. Leigh, Photochem. Photobiol. 2006 , 82 , 29 1 –

300.

[6] a) M. A. Miranda, A. Lahoz, R. Martinez-Manez, F. Bosca, J. V.

Castell, J. Perez-Prieto, J. Am. Chem. Soc. 1999 , 121 , 1156 9 –1 1570;

b) J. Perez-Prieto, F. Bosca, R. E. Galian, A. Lahoz, L. R. Domingo,

M. A. Miranda, J. Org. Chem. 2003 , 68 , 510 4 –5 113; c) J. Perez-

Prieto, A. Lahoz, F. Bosca, R. Martinez-Manez, M. A. Miranda, J.

Org. Chem. 2004 , 69 , 37 4 –3 81; d) J. Perez-Prieto, S.-E. Stiriba, F.

Bosca, A. Lahoz, L. R. Domingo, F. Mourabit, S. Monti, M. A. Mir-

anda, J. Org. Chem. 2004 , 69 , 861 8 –8 625; e) J. Perez-Prieto, R. E.

Galian, M. C. Morant-Mirana, M. A. Miranda, Chem. Commun.

2005, 3180–3182; f) R. E. Galian, L. Pastor-Perez, M. A. Miranda, J.

Perez-Prieto, Chem. Eur. J. 2005 , 11 , 344 3 –3 448.

[7] a) L. Biczok, T. Berces, H. Linschitz, J. Am. Chem. Soc. 1997 , 119 ,

1107 1 –1 1077; b) U. Pischel, W. M. Nau, Photochem. Photobiol. Sci.

2002 , 1 , 14 1 –1 47; c) M. Yamaji, Y. Aoyama, T. Furukawa, T. Ito, S.

Tobita, Chem. Phys. Lett. 2006 , 420 , 18 7 –1 91; d) U. Pischel, D.

Patra, A. L. Koner, W. M. Nau, Photochem. Photobiol. 2006 , 82 ,

31 0 –3 17; e) M. R. Seyedsayamdost, S. Y. Reece, D. G. Nocera, J.

Stubbe, J. Am. Chem. Soc. 2006 , 128 , 156 9 –1 579; f) R. E. Galian, G.

Litwinienko, J. Perez-Prieto, K. U. Ingold, J. Am. Chem. Soc. 2007 ,

129 , 928 0 –9 281.

[8] N. J. Turro, R. Engel, J. Am. Chem. Soc. 1969 , 91 , 711 3 –7 121.

[9] a) S. Canonica, B. Hellrung, J. Wirz, J. Phys. Chem. A 2000 , 104 ,

122 6 –1 233; b) N. B. Sultimova, P. P. Levin, O. N. Chaikovskaya,

Russ. Chem. Bull. 2005, 54, 1439–1444.

[31] R. V. Bensasson, J.-C. Gramain, J. Chem. Soc. Faraday Trans. 1

1980 , 76 , 180 1 –1 810.

[32] T. Kluge, O. Brede, Chem. Phys. Lett. 1998 , 289 , 31 9 –3 23.

[33] P. J. Wagner, Acc. Chem. Res. 1983 , 16 , 46 1 –4 67.

[34] a) P. J. Wagner, R. Pabon, B.-S. Park, A. R. Zand, D. L. Ward, J.

Am. Chem. Soc. 1994 , 116 , 58 9 –5 96; b) P. J. Wagner, P. Klan, J. Am.

Chem. Soc. 1999 , 121 , 962 6 –9 635.

[35] R. Vargas, J. Garza, D. Dixon, B. P. Hay, J. Phys. Chem. A 2001 , 105 ,

77 4 –7 78.

[36] G. Hçrner, G. L. Hug, D. Pogocki, B. Marciniak, unpublished re-

sults.

[37] a) D. W. Snelgrove, J. Lusztyk, J. T. Banks, P. Mulder, K. U. Ingold,

J. Am. Chem. Soc. 2001 , 123 , 46 9 –4 77; b) G. Litwinienko, K. U.

Ingold, J. Org. Chem. 2003 , 68 , 343 3 –3 438; c) G. Litwinienko, K. U.

Ingold, J. Org. Chem. 2004 , 69 , 588 8 –5 896; d) G. Litwinienko, K. U.

Ingold, J. Org. Chem. 2005 , 70 , 898 2 –8 990; e) G. Litwinienko, K. U.

Ingold, Acc. Chem. Res. 2007 , 40 , 22 2 –2 30.

[38] A. L. Koner, U. Pischel, W. M. Nau, Org. Lett. 2007 , 9 , 289 9 –2 902.

[39] a) S. Braun, H.-O. Kalinowski, S. Berger, 150 and More Basic NMR

Experiments, Wiley-VCH, Weinheim, 1998, p. 460; b) K. Stott, J.

Keeler, Q. N. Van, A. J. Shaka, J. Magn. Reson. 1997 , 125 , 30 2 –3 24.

[40] P. Filipiak, G. L. Hug, K. Bobrowski, B. Marciniak, J. Photochem.

Photobiol. A 2005 , 172 , 32 2 –3 30.

[41] a) R. Hermann, G. R. Mahalaxmi, S. Jochum, S. Naumov, O. Brede,

J. Phys. Chem. A 2002 , 106 , 237 9 –2 389; b) O. Brede, T. Leichtner, S.

Kapoor, S. Naumov, R. Hermann, Chem. Phys. Lett. 2002 , 366 , 37 7 –

382.

[42] P. Wisniowski, G. L. Hug, J. Mirkowski, DECOM2006, Warsaw,

2006.

[10] M. P. Williamson, J. P. Waltho, Chem. Soc. Rev. 1992 , 21 , 22 7 –2 36.

[11] M. Karplus, J. Chem. Phys. 1959 , 30 , 1 1 –1 5.

[12] M. Kraszni, Z. Skakacs, B. Noszal, Anal. Bioanal. Chem. 2004 , 378 ,

144 9 –1 463.

[13] R. L. Laws, J. B. A. Ross, H. R. Wyssbrod, J. M. Beechem, L. Brand,

J. C. Sutherland, Biochemistry 1986 , 25 , 59 9 –6 07.

[14] G. W. Kirby, J. Michael, J. Chem. Soc. D 1971 , 18 7 –1 88.

[15] K. G. R. Pachler, Spectrochim. Acta 1964 , 20 , 58 1 –5 87.

[16] J. R. Unruh, M. R. Liyanage, C. K. Johnson, J. Phys. Chem. B 2007 ,

111 , 549 4 –5 502.

[43] E. Hayon, T. Ibata, N. N. Lichtin, M. Simic, J. Phys. Chem. 1972 , 76 ,

207 2 –2 078.

[44] a) K. M. Bansal, R. W. Fessenden, Radiat. Res. 1976 , 67 , 1 –8; b) M.

Fontecave, P. Nordlund, H. Eklund, P. Reichardt, Adv. Enzymol.

Relat. Areas Mol. Biol. 1992 , 65 , 14 7 –1 83.

[17] C. Reichardt, Solvents and Solvent Effects in Organic Chemistry, 2nd

ed., VCH, Weinheim, 1990, p. 408.

[18] a) J. Kobayashi, U. Nagai, Biopolymers 1978 , 17 , 226 5 –2 277; b) J.

Kobayashi, T. Higashijima, S. Sekido, T. Miyazawa, Int. J. Peptide

Protein Res. 1981, 17, 486–494; c) J. Kobayashi, T. Higashijima, T.

Miyazawa, Int. J. Peptide Protein Res. 1984, 24, 40–47.

[19] J. B. A. Ross, W. R. Laws, A. Buku, J. C. Sutherland, H. R. Wyss-

brod, Biochemistry 1986 , 25 , 60 7 –6 12.

[45] J. G. Kirkwood, J. Chem. Phys. 1935 , 3 , 30 0 –3 13.

[46] G. M. Torrie, J. P. Valleau, J. Comp. Physiol. 1977, 23, 187–199.

[47] R. M. Levy, M. Karplus, J. A. McCammon, Chem. Phys. Lett. 1979 ,

65 , 4 –1 1.

[48] G. Widmalm, R. W. Pastor, J. Chem. Soc. Faraday Trans. 1992 , 88 ,

174 7 –1 754.

[20] a) M. Kainosho, K. Ajisaka, J. Am. Chem. Soc. 1975 , 97 , 563 0 –5 631;

b) K. Guzow, R. Ganzynkowicz, A. Rzeska, J. Mrozek, M. Szabelski,

J. Karolczak, A. Liwo, W. Wiczk, J. Phys. Chem. B 2004 , 108 , 387 9 –

3889.

[21] G. Scherer, M. L. Kramer, M. Schutkowski, U. Reimer, G. Fischer, J.

Am. Chem. Soc. 1998 , 120 , 556 8 –5 574.

[22] R. W. Taft, M. J. Kamlet, Org. Magn. Reson. 1980 , 14 , 48 5 –4 93.

[23] a) T. Steiner, Angew. Chem. 2002 , 114 , 5 0 –8 0; Angew. Chem. Int.

Ed. 2002 , 41 , 4 8 –7 6; b) G. Jaccard, J. Lauterwein, Helv. Chim. Acta

1986, 69, 1469–1485.

[49] S. He, H. A. Scheraga, J. Chem. Phys. 1998 , 108 , 28 7 –3 00.

[50] A. D. MacKerell Jr. , D. Bashford, R. L. Bellott, R. L.Dunbrack Jr. ,

J. D. Evanseck, M. J. Field, S. Fischer, J. Gao, H. Guo, S. Ha, D.

Joseph-McCarthy, L. Kuchnir, K. Kuczera, F. T. K. Lau, C. Mattos,

S. Michnick, T. Ngo, D. T. Nguyen, B. Prodhom, W. W. Reiher III, B.

Roux, M. Schlenkrich, J. C. Smith, R. Stote, J. Straub, M. Watanabe,

J. Wiorkiewicz-Kuczera, D. Yin, M. Karplus, J. Phys. Chem. B. 1998 ,

102 , 358 6 –3 616.

[51] HyperChem Computational Chemistry, Hypercube Inc., Gainesville,

USA, 2002.

[24] J. Koput, G. Hçrner, F. Kazmierczak, B. Marciniak, unpublished re-

sults.

Received: February 20, 2008

Published online: July 11, 2008

Chem. Eur. J. 2008, 14, 7913 – 7929

ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

7929

Article Doi

Head-to-tail interactions in tyrosine/benzophenone dyads in the ground and the excited state: NMR and laser flash photolysis studies

DOI: 10.1002/chem.200800315

Source and publish data:

Authors:

Article abstract of DOI:10.1002/chem.200800315

Full text of DOI:10.1002/chem.200800315

Products guided by the article

R&D Labs maybe for 1083386-10-0

Relevant to this article

Hot Product