Self-sorting of cyclic peptide homodimers into a heterodimeric assembly featuring an efficient photoinduced intramolecular electron-transfer process

DOI: 10.1002/chem.201304200

Full Paper

&

Electron Transfer

Self-Sorting of Cyclic Peptide Homodimers into a Heterodimeric

Assembly Featuring an Efficient Photoinduced Intramolecular

Electron-Transfer Process

Gemma Aragay,^[a]Barbara Ventura,^[b]Arcadio Guerra,^[c]Inmaculada Pintre,^[a]

Claudio Chiorboli,^[b]Rebeca Garcꢀa-FandiÇo,^[c]Lucia Flamigni,*^[b]Juan R. Granja,*^[c]and

Pablo Ballester*^[a]

Abstract: We describe the thermodynamic characterisation

of the self-sorting process experienced by two homodimers

assembled by hydrogen-bonding interactions through their

cyclopeptide scaffolds and decorated with Zn–porphyrin

and fullerene units into a heterodimeric assembly that con-

tains one electron-donor (Zn–porphyrin) and one electron-

acceptor group (fullerene). The fluorescence of the Zn–por-

phyrin unit is strongly quenched upon heterodimer forma-

tion. This phenomenon is demonstrated to be the result of

an efficient photoinduced electron-transfer (PET) process oc-

curring between the Zn–porphyrin and the fullerene units of

the heterodimeric system. The recombination lifetime of the

charge-separated state of the heterodimer complex is in the

order of 180 ns. In solution, both homo- and heterodimers

are present as a mixture of three regioisomers: two stag-

gered and one eclipsed. At the concentration used for this

study, the high stability constant determined for the hetero-

dimer suggests that the eclipsed conformer is the main

component in solution. The application of the bound-state

scenario allowed us to calculate that the heterodimer exists

mainly as the eclipsed regioisomer (75–90%). The attractive

interaction that exists between the donor and acceptor

chromophores in the heterodimeric assembly favours their

arrangement in close contact. This is confirmed by the pres-

ence of charge-transfer bands centred at 720 nm in the ab-

sorption spectrum of the heterodimer. PET occurs in approxi-

mately 75% of the chromophores after excitation of both

Zn–porphyrin and fullerene chromophores. Conversely, anal-

ogous systems, reported previously, decorated with extend-

ed tetrathiafulvalene and fullerene units showed a PET pro-

cess in a significantly reduced extent (33%). We conclude

that the strength (stability constant (K)ꢁeffective molarity

(EM)) of the intramolecular interaction established between

the two chromophores in the Zn–porphyrin/fullerene cyclo-

peptide-based heterodimers controls the regioisomeric dis-

tribution and regulates the high extent to which the PET

process takes place in this system.

Introduction

[a] Dr. G. Aragay, Dr. I. Pintre, Prof. Dr. P. Ballester

The Institute of Chemical Research of Catalonia (ICIQ)

and Catalan Institution of Research and Advanced Studies (ICREA)

Avda. Pa¨ısos Catalans 16, 43007 Tarragona (Spain)

Fax: (+34)977-920-221

In living organisms, molecular assemblies with tubular topolo-

gy carry out remarkable tasks such as molecular recognition^[1]

or ion transport.^[2]For this reason, they have inspired synthetic

and supramolecular chemists.^[3]In this vein and over the past

two decades, the molecular self-assembly of cyclic peptides

(CP) has been used for the construction of remarkable nano-

tubular structures.^[4–7]The assembled nanotubular structures

possess unique properties from which applications for the de-

velopment of bio-inspired functional materials have been de-

rived.^[8–10]Owing to the ease of their synthesis and functional

versatility, CP self-assembled architectures have been used in

diverse areas of chemical research ranging from molecular rec-

ognition studies,^[11,12]transmembrane transport^[13,14]and inves-

tigation of antiamyloidogenic properties,^[15]to charge transport

in vertically aligned peptide nanotubes.^[16]

E-mail: pballester@iciq.es

[b] B. Ventura, C. Chiorboli, L. Flamigni

Dr. B. Ventura, Dr. C. Chiorboli, Dr. L. Flamigni

Istituto ISOF-CNR

Via P. Gobetti 101, 40129 Bologna (Italy)

Fax: (+39)051-639-9844

E-mail: flamigni@isof.cnr.it

[c] A. Guerra, R. Garcꢀa-FandiÇo, J. R. Granja

Mr. A. Guerra, Dr. R. Garcꢀa-FandiÇo, Prof. Dr. J.R. Granja

Departamento de Quꢀmica Orgꢁnica

Centro Singular de Investigaciꢂn en Quꢀmica Biolꢂgica y

Materiales Moleculares, Campus Vida, Universidad de Santiago

15782 Santiago de Compostela (Spain)

Fax: (+34)881-815-704

The intermolecular, b-sheet-like hydrogen-bonding network

that is established between C=O and NꢀH groups pertaining

to the backbone amino acids of CP monomers constitutes the

E-mail: juanr.granja@usc.es

Supporting information for this article is available on the WWW under

http://dx.doi.org/10.1002/chem.201304200.

Chem. Eur. J. 2014, 20, 3427 – 3438

3427

ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

Scheme 1. Dimerisation process of cyclopeptides CP6 (top) and CP7 (bottom) (HOBt=1-hydroxybenzotriazole, Bn=benzyl). The line-drawing structures of

CP6 and CP7 are shown together with their schematic representation upon dimer formation. The three regioisomers of each dimer are represented as car-

toon models.

essential driving force for the assembly of infinite supramolec-

ular architectures with tubular topology. The desire to obtain

an accurate characterisation of the thermodynamic and kinetic

properties of these architectures, with the aim of better under-

standing their function and potential applications, brought

with it the necessity of blocking the formation of extensive

and highly ordered b-sheet-like CP assemblies that have unde-

fined and uncontrollable length. This objective was achieved

by installing N-methyl axial substituents in one of the two

faces of the CP monomer.^[17]In this manner, only dimeric tubu-

lar structures can be assembled. The dimers constitute simple

model systems suitable for further characterisation and investi-

gation of more complex self-assembly processes leading to

larger tubular structures. a,g-CPs are cyclic peptide hybrids

composed of several units of a- and g-amino acids that exhibit

large dimerisation constants.^[18,19]The g-amino acids are usually

constructed using cyclopentane (Acp) or cyclohexane (Ach)

scaffolds, as shown in the structures of CP6 and CP7, respec-

tively (Scheme 1).^[20]In addition to their tendency to homodi-

merise (i.e., CP6·CP6 and CP7·CP7), the blocked CPs show sig-

nificant self-sorting abilities yielding the heterodimeric aggre-

gates in amounts that are significantly larger than expected in

a self-sorting process involving isoenergetic dimers (CP6·CP7

in Scheme 2). Previous reports assigned a 20-fold increase in

the thermodynamic stability of the cyclopeptide heterodimer

with respect to the parent homodimers.^[21]The equilibrium

constant for the self-sorting process of the homodimers into

the heterodimer assembly is estimated to be 400, which is

about one hundred times larger than the statistically predicted

value (K_SS=4; see Scheme 3). This observation was explained

Scheme 2. Binding model used for the calculation of the stability constant

of heterodimer CP6·CP7. The monomers and dimers are shown in schematic

representation. Only the eclipsed regioisomers of the dimers is depicted.

by attributing to the heterodimeric assembly the existence of

energetically more favourable intermolecular interactions be-

tween the backbones of the cyclohexyl–cyclopentyl (Ach:Acp)

components than the corresponding ones in the cyclohexyl–

cyclohexyl (Ach:Ach) and cyclopentyl–cyclopentyl (Acp:Acp)

homodimers.^[21,22]

Supramolecular systems bearing photo- and electroactive

functionalities are particularly interesting for the development

of efficient arrays of chromophores suitable for energy collec-

tion and charge separation, which might have potential appli-

Chem. Eur. J. 2014, 20, 3427 – 3438

www.chemeurj.org

3428

ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

which are decorated at their peripheries with one Zn–porphy-

rin and one fullerene residue, respectively (Scheme 1). In solu-

tion, CP monomers CP6 and CP7 readily self-assemble into ho-

modimeric aggregates CP6·CP6 and CP7·CP7. We disclose our

attempts towards the accurate thermodynamic characterisa-

tion of the homodimers. We also report on the self-sorting pro-

cess of the homodimers, which produces the exclusive forma-

tion of the heterodimeric hybrid CP6·CP7. The CP6·CP7 dimer

bears complementary and photoactive Zn–porphyrin and full-

erene chromophores at the periphery of the two adjacent

monomers. We also examine in detail the effect that the attrac-

tive intramolecular interaction established between the chro-

mophores produces in the distribution of the three possible re-

gioisomers of the CP6·CP7 heterodimer. Finally, we present the

results obtained in the photophysical investigations of the

heterodimer demonstrating the existence of a highly efficient

intramolecular electron-transfer process. The extent of the pro-

cess is clearly related to the distribution of regioisomers of the

heterodimer. The Zn–porphyrin/fullerene CP6·CP7 hybrid

powers the electron transfer in significantly larger magnitude

than the p-extended tetrathiafulvalene/fullerene cyclopeptide-

hybrid analogue previously reported.^[28]Furthermore, whereas

in the former case only the fullerene’s excited state was able

to yield the charge-separated state, in the present case excita-

tion of any of the two chromophores (C₆₀or Zn–porphyrin) in-

duces the electron-transfer process. The complete characterisa-

tion of the photoinduced processes that take place within the

heterodimeric assembly CP6·CP7 in solution is also described.

The presented findings are highly relevant both for further un-

derstanding the characteristics of the electron-transfer process-

es occurring in supramolecular systems, as well as for the

eventual exploitation of this particular Zn–porphyrin/fullerene

CP6·CP7 hybrid in functional devices.

Scheme 3. Self-sorting equilibria showing the schematic representation of

the eclipsed regioisomer for the heterodimer. The relationship that exists be-

tween the equilibrium constant of the self-sorting/disproportionation pro-

cess and the thermodynamic stability constants of the dimeric assemblies is

indicated. Please note that even if homo- and heterodimers were held to-

gether by intermolecular forces of analogous strength (same microscopic

binding constant) K_SSwill take a value of 4, owing to the statistical advant-

age provided by the reduced symmetry of the heterodimer.

cations in artificial photosynthesis.^[23–27]In this context, some

years ago one of us reported the physicochemical properties

of a CP heterodimeric assembly of the type Ach:Acp decorated

with complementary electron-active chromophores at its pe-

riphery.^[28]In brief, a photoexcited fullerene C₆₀unit covalently

attached to one cyclopeptidic component of the heterodimer

underwent a one-electron transfer process from a p-extended

tetrathiafulvalene (exTTF) residue located on the complemen-

tary monomer. Unfortunately, the extent of the charge-separa-

tion process was quite limited. The main reason being that

only one of the three possible regioisomeric heterodimers

(eclipsed) was efficient in promoting the process. The other

two regioisomers (staggered) were less likely to mediate the

electron-transfer process owing to the large distances featured

between donor and acceptor units.

Previously, we described the installation of Zn–porphyrin res-

idues in dimeric assemblies of cyclic octapeptides as a way to

control the relative distribution of the rotatable regioisom-

ers.^[29]The interaction of a bis-Zn–porphyrin homodimer with

a

ditopic diamine, namely, 1,4-diazobicyclo[2.2.2]octane

(DABCO), induces the mixture of regioisomers to adopt exclu-

sively the conformation in which DABCO is sandwiched be- Results and Discussion

tween two porphyrin moieties. We hypothesised that the com-

Synthesis

bination of a fullerene C₆₀unit with a Zn–porphyrin residue in

a dimer of cyclic peptides based on Ach and Acp backbones

should yield a thermodynamically highly stable heterodimer

that may exhibit interesting physicochemical properties. The

interaction between the flat Zn–porphyrin and the curved sur-

face of the fullerenes is known to be attractive and is expected

to induce the preferential assembly of the heterodimer as the

eclipsed regioisomer (Scheme 2).^[30]In addition, the direct

“non-covalent” interactions between Zn–porphyrin and ful-

lerenes have important implications for the generation of effi-

cient photoactive architectures. Supramolecular donor–accept-

or hybrids containing these units are well known to have the

ability to generate charge-separated species by light incidence,

thus mimicking the natural photosynthetic process.^[31]In gener-

al, the construction of supramolecular assemblies decorated

with Zn–porphyrin and fullerene residues provides significant

synthetic advantages with respect to the covalent construction

of closely related dyads.^[32]

The experimental procedures for the synthesis and the charac-

terisation data of the parent cyclic peptides CP1–4 are de-

scribed in the Supporting Information (Scheme S1). The com-

pounds were prepared by using procedures that are quite sim-

ilar to those reported previously.^[33]The cyclic peptide CP5

based on the Acp backbone and decorated with the free-base

porphyrin residue was prepared by an esterification reaction

(Scheme 1) between CP3 alcohol and the porphyrin carboxylic

acid 1 using N,N-diisopropylcarboiimide (DIC) as activating

agent and a catalytic amount of 4-dimethylaminopyridine

(DMAP). The metallation of the porphyrin-substituted cyclo-

peptide CP5 was carried out by treatment with zinc acetate in

a dichloromethane/methanol mixture yielding the Zn–porphy-

rin cyclopeptide CP6 in a 98% yield after column chromatogra-

phy purification. The fullerene derivative CP7 was obtained by

amide coupling between the CP4 amine and phenyl-C61-buty-

ric acid (PCBA), which was activated using N-[(dimethylamino)-

1H-1,2,3-triazolo[4,5-b]pyridin-1-ylmethylene]-N-methylmetha-

naminium hexafluorophosphate N-oxide (N-HATU) and diiso-

In this manuscript, we describe the synthesis of two cyclo-

peptides CP6 and CP7 based on Acp and Ach backbones,

Chem. Eur. J. 2014, 20, 3427 – 3438

www.chemeurj.org

3429

ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

propylethylamine (DIEA) as base in chloroform (Scheme 1).

Both CP derivatives, CP6 and CP7, were fully characterised

through high-resolution techniques (see the Supporting Infor-

mation).

assigned to the protons of the homodimers indicates the ex-

clusive formation of the heterodimer, possibly as a mixture of

regioisomers. Considering that the chemical exchange be-

1

tween regioisomers of the homodimers is slow on the H NMR

It has been shown that fullerene derivatives quite similar to

CP7^[28]and structural analogues of the porphyrin derivative

CP6^[29]display a strong tendency to self-assemble into homo-

dimers when dissolved in non-polar organic solvents. The pro-

duced self-assembled homodimers exist in solution as an equi-

librium mixture of three possible regioisomers (Scheme 1). We

designate the regioisomers as follows: 1) eclipsed isomer (if

the larger substituents on adjacent monomers display a syn ar-

rangement (D(S))) and 2) staggered regioisomer (when the

same larger substituents are located in an anti arrangement).

In addition, two staggered regioisomers are possible. They can

be distinguished by the sense of rotation required for the top

CP units to achieve a syn arrangement with the larger substitu-

ent in the adjacent monomers (clockwise (D(Acw)) or counter-

clockwise (D(Accw))). In solution, the observed molar ratio be-

tween staggered and eclipsed regioisomers depends on sever-

al factors. Among them, the type of substituents and the

nature of the cyclopeptidic backbone are the most important.

Based on statistical grounds and assuming that the three re-

gioisomers are isoenergetic, a 2:1 staggered/eclipsed distribu-

tion would be expected.

spectroscopic timescale, it is sensible to expect that a mixture

of regioisomers of the heterodimer should also show slow ex-

1

change dynamics. Unfortunately, the complexity of the H NMR

spectrum assigned to the heterodimer, as a mixture of re-

gioisomers, precludes any precise calculation of the molar

composition. The exclusive formation of the heterodimer

CP6·CP7 from an equimolar combination of homodimers is in

complete agreement with previous findings for analogous self-

sorting equilibria involving cyclopeptide dimers with Acp and

Ach backbone scaffolds.^[21]Most likely, and as already men-

tioned above, the hydrogen bonds that stabilise the heterodi-

meric assembly Acp:Ach are energetically more favourable

than those in the respective homodimers. However, the exis-

tence of additional attractive interactions between substituents

in the heterodimer cannot be ruled out. The exclusive observa-

tion of the heterodimeric assembly CP6·CP7 at 1 mm concen-

tration indicates that the equilibrium constant for the self-sort-

ing process of homodimers can be estimated as K_SS>10⁴.

Thermodynamic characterisation of the assemblies

1

The complexity of the H NMR spectrum corresponding to

Dimerisation-constant values for hexacyclopeptidic homodim-

ers closely related to CP6 and CP7 have been previously re-

ported in the literature.^[28,21]The magnitudes of the reported

constants are in the range of K_Dhomo=10⁵–10⁶m^ꢀ1. By using UV/

Vis absorption spectroscopy, we performed a dilution experi-

ment of a solution of CP6 (10^ꢀ5–10^ꢀ7m) in dichloromethane.

Upon dilution, we expected to observe shifts in the maxima of

the bands of the absorption spectra in response to the con-

centration changes experienced by the monomer CP6 and the

dimer CP6·CP6. Unfortunately, the maxima of the absorption

bands in the series of UV/Vis absorption spectra registered at

different concentrations, as well as the corresponding extinc-

tion coefficients, did not experience noticeable changes. This

fact complicates the determination of an accurate dimerisa-

tion-constant value for CP6. The application of the same dilu-

tion technique to a concentrated solution of CP7 in dichloro-

methane was again unsuccessful for the accurate determina-

tion of its dimerisation-constant value owing to the complete

lack of spectral changes. In view of these limitations, we decid-

ed to simply estimate the values of the dimerisation constant

for both CP6 and CP7 as K_Dhomo=1ꢁ10⁶m^ꢀ1. The magnitude of

the estimated dimerisation-constant values is in complete

agreement with those previously reported for related cyclo-

peptides.^[28,34]At this point, it is important to highlight that the

use of “non-covalent” reversible interactions in the assembly of

dimeric molecular aggregates or aggregates of any other stoi-

chiometric order suffers from a strong drawback when com-

pared with the connection of molecular units by covalent

bonding. That is, as a result of the use of a molecular self-as-

sembly process, the species involved in the equilibrium will be

present in solution in varying amounts. The magnitude of the

stability constant of the assembly and the concentration used

a solution of the porphyrin derivative CP6 in CDCl₃prevented

the accurate determination of the regioisomeric ratio for the

dimer CP6·CP6 but supported its quantitative formation. Con-

versely, the ¹H NMR spectrum of a solution of CP7 in CDCl₃

clearly showed three different doublets for the NH amide

proton of the lysine (Lys) side chain resonating at d=8.44, 8.42

and 8.29 ppm and also supported quantitative formation of

the homodimer. The observation of three separated signals for

the NH protons of the lysine side chain indicates the presence

of three regioisomeric CP7·CP7 dimers in solution. The three

regioisomers are in an equilibrium that shows slow exchange

dynamics on the ¹H NMR spectroscopic chemical-shift time-

scale. Integration of the NH proton signals assigned to each re-

gioisomeric dimer allowed the calculation of the molar ratio of

regioisomers D(Acw)/D(Accw)/D(S) as 1:1:2. The fact that the

eclipsed regioisomer is the major component in the equilibri-

um mixture of CP7·CP7 regioisomers hints to the existence of

attractive interactions between the two fullerene residues or,

perhaps to the participation of the amide NH of the Lys side

chain in a weak hydrogen-bonding interaction in the eclipsed

form as denoted by the downfield shift of its signals in the

NMR spectrum. This favourable interaction alters the expected

statistical distribution and rules out the idea that the steric

strain present in the eclipsed isomers should always bias the

equilibrium towards the preferential formation of staggered re-

gioisomers.

1

The H NMR spectrum of an equimolar mixture of CP6 and

CP7 (ꢁ1 mm) in CDCl₃showed several sets of proton signals

but none of them coincided with those previously observed

for the regioisomeric mixtures of any of the two homodimers

(CP6·CP6 and CP7·CP7). Thus, the lack of signals that could be

Chem. Eur. J. 2014, 20, 3427 – 3438

www.chemeurj.org

3430

ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

to perform the self-assembly are key factors in determining the

distribution of species in solution. In the case at hand, the

magnitude of the dimerisation constants estimated for CP6

and CP7 dictates that both monomer and dimeric species will

be present in solution, although to a different extent and that

the ratio of the dimer/monomer concentrations will depend

on the total monomer concentration used.

of the titration data to the theoretical binding model was

good and returned the stability-constant value for the hetero-

dimer CP6·CP7 as K(CP6·CP7)=1.8ꢁ10⁸m^ꢀ1(Figure S1 in the

Supporting Information). Remarkably, the magnitude of the

stability constant for the heterodimer CP6·CP7 turned out to

be one order of magnitude larger than those described/esti-

mated for related Acp:Ach-type heterodimers either with no

additional substituents^[21]or bearing one fullerene and one

exTTF unit^[28]covalently attached to the periphery of the CP

monomers.

To assess the relative thermodynamic stability of the homo-

dimers CP6·CP6 and CP7·CP7 with respect to the heterodimer

CP6·CP7, we performed a UV/Vis absorption titration experi-

ment. We used a solution (ꢁ0.7 mm) of the porphyrin deriva-

tive CP6 in dichloromethane, to which we added incremental

amounts of a concentrated solution of CP7 in the same sol-

vent. The concentration of the porphyrin derivative CP6 was

kept constant throughout the titration. The incremental addi-

tion of the fullerene derivative CP7 provoked a significant de-

crease in the absorption intensity of the Soret band of the por-

phyrin unit of CP6 and a concomitant bathochromic shift from

422 to 430 nm (Figure 1a). The titration gave clear isosbestic

The higher thermodynamic stability determined for CP6·CP7

could be attributed to the existence of attractive interactions

between the porphyrin and the fullerene substituents. This hy-

pothesis is supported by the existence of charge-transfer

bands in the UV/Vis absorption spectrum of the CP6·CP7 hete-

rodimer (see below). It is well known that fullerenes form ther-

modynamically stable complexes with metalloporphyrin

dimers^[36]and trimers.^[37]The additional interactions between

C₆₀and the Zn–porphyrin residues will add up to the hydrogen

bonds established between the CP backbones of the mono-

mers in the heterodimer, thus providing a sensible increase in

the thermodynamic stability of the heterodimeric aggregate.

In any supramolecular complex held together by multiple

non-covalent interactions (multivalency), the bound state can

consist of a mixture of complexes displaying partially and fully

bound geometries.^[36,38,39]In the case discussed here, the mea-

sured stability constant (macroscopic) for the heterodimer,

K(CP6·CP7), is considered to be the sum of the stability con-

stants corresponding to the three possible regioisomers (three

bound states). We also consider that the eclipsed isomer must

correspond to the fully-bound geometry, in which both the hy-

drogen-bonding interactions established between the cyclo-

peptidic backbones of the monomers and the fullerene–por-

phyrin interactions are simultaneously present. However, in

any of the two staggered regioisomers, owing to geometrical

constrains, the fullerene–porphyrin interaction is less likely.

Thus, the calculated value for K(CP6·CP7) is shown in Equa-

tion (1):

KðCP6 ꢂ CP7Þ ¼ 2 K₀þK₀K₁EM

ð1Þ

Figure 1. a) UV/Vis absorption and b) fluorescence spectra registered during

the titration of [CP6]=7ꢁ10^ꢀ7m with CP7 (0–1.5 equiv).

in which K₀is a stability constant of a model system for a hete-

rodimer Acp:Ach lacking substituents, K₁is the association con-

stant for the porphyrin–fullerene interaction, EM is the effec-

tive molarity of formation of the intramolecular porphyrin–full-

erene interaction and the 2 is a statistical factor that accounts

for the degeneracy of two isoenergetic staggered regioisom-

ers. By considering the partially bound state scenario discussed

above, it is possible to calculate the relative population of fully

bound (eclipsed) and partially bound (staggered) states simply

by selecting a reference value for K₀. Unfortunately, an undis-

putable value of K₀is not available from previous example in

the literature, but it is sensible to assume that it should be

close to one order of magnitude larger (ꢁ10⁷m^ꢀ1) than the di-

merisation constants of the homodimers. This magnitude ac-

counts for the more than statistically predicted assembly of

spectra. The UV/Vis absorption data were analysed mathemati-

cally by using a theoretical binding model that considers two

species in four coloured states: the monomer of CP6,^[35]the

homodimers CP6·CP6 and CP7·CP7 and the heterodimer

CP6·CP7 (Scheme 2). To reduce the number of variables in the

fit, the absorption spectra (molar absorption coefficient scale)

of the dimers CP6·CP6 and CP7·CP7 were fixed to those regis-

tered using 10^ꢀ5m solutions, respectively. At this concentra-

tion, the dimer is considered to be the almost exclusive species

present in solution for K_Dhomo=1ꢁ10⁶m^ꢀ1. The dimerisation

constants were also fixed to the values described in the litera-

ture for closely related systems (1ꢁ10⁶m^ꢀ1; see above). The fit

Chem. Eur. J. 2014, 20, 3427 – 3438

www.chemeurj.org

3431

ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

the heterodimer Acp:Ach that was experimentally observed in

self-sorting processes of equimolar mixtures of homodimers

[Acp]₂and [Ach]₂with K_Dhomovalues of approximately 1ꢁ

10⁶m^ꢀ1. By using Equation (1), the experimentally determined

value for K(CP6·CP7)=1.8ꢁ10⁸m^ꢀ1and K₀=10⁷m^ꢀ1, the popu-

lation of the staggered isomers (partially bound state) in the

heterodimer can be calculated to be [2K_o/K(CP6·CP7)]ꢁ100=

11%. Whereas the eclipsed isomer (fully bound state) must

represent the rest of the population of the bound-state:

{[K(CP6·CP7)ꢀ2K_o]/K(CP6·CP7)}ꢁ100=89%. The use of the

partially bound state scenario for the calculation of the distri-

bution of bound states (regioisomers) gives an important

weight to the K₀value used as reference. For example, by

simply assuming that K₀=2ꢁ10⁷m^ꢀ1, just a twofold increase

with respect to our previous estimate, the calculated distribu-

tion of regioisomers turns out to be 22% staggered and 78%

eclipsed.^[40]The calculation of the percentage of eclipsed and

staggered isomers that are present in solution for the heterodi-

mer CP6·CP7 will become more relevant at the end of the sec-

tion dedicated to the study of its photophysical properties (see

below).

duced electron-transfer (PET) process mediated by the fuller-

ene residue. This hypothesis will be demonstrated in detail in

the following section through the results obtained from

a series of accurate photophysical studies.

It is worth noting that the fit of the fluorescence titration

data to the same binding model used for the UV/Vis absorp-

tion titration data, but considering only two fluorescent spe-

cies (CP6 and CP6·CP6), returned a stability-constant value for

the heterodimer CP6·CP7 complex in complete agreement

with that determined in the UV/Vis absorption titration experi-

ment.

Photophysical properties of the individual components and

related models

The spectroscopic and photophysical properties of suitable

models of the photoactive substituents (zinc tetrapentylphe-

nylporphyrin (ZnTPPP), phenyl-C61-butyric acid methyl ester

(PCBM); Figure S4 in the Supporting Information) as well as

those of the cyclopeptides CP6 and CP7 were determined in

dichloromethane. Some of the calculated parameters are sum-

marised in Table 1. The molar absorption coefficients of all spe-

cies were determined at a 6ꢁ10^ꢀ6m concentration. In the

cases of CP6 and CP7, increasing the concentration by two

The above calculation of the percentage of regioisomers for

the CP6·CP7 heterodimer indicates that the eclipsed isomer is

highly populated. By having at hand values for the stability

constants of homo (CP6·CP6 and CP7·CP7) and

hetero (CP6·CP7) dimeric assemblies, the calculation

Table 1. Luminescence properties and energy levels of the excited states of models

ZnTPPP, PCBM, CP6 and CP7 in dichloromethane.

of the equilibrium constant value for the self-sorting

process of homodimers into the heterodimer is

straightforward (Scheme 3).

295 K

t [ns]^[c]

77 K

l_max[nm]^[a]

f ^[b]

l_max[nm]^[a]

t [ns]^[c]

E [eV]^[d]

In complete agreement with the large magnitude

determined for the stability constant of the heterodi-

mer, the value of the calculated equilibrium constant

for the self-sorting process, K_SS(CP6·CP7)=1.3ꢁ10⁵, is

three orders of magnitude larger than the one we

could estimate based on reported data for related

equilibria involving Acp and Ach homodimers but

lacking substituents involved in attractive/repulsive

intramolecular interactions.^[21]Clearly, the attractive

interactions established between the peripheral sub-

stituents (Zn–porphyrin and C₆₀) in the heterodimer

CP6·CP7 must be responsible both for the high ther-

modynamic stability of the complex and for the high

fl

CP6

601, 648

0.041

1.6

613, 670,

800^[e]

707, 789

614, 669,

808^[e]

1.7

2.02

19.7ꢁ10^6[e]

1.4

1.55^[e]

1.75

CP7

ZnTPPP

707, 783

599, 648

0.00078

0.040

1.3

1.6

2.02

17.0ꢁ10^6[e]

1.4

1.53^[e]

1.74

PCBM

707, 780

0.00083

1.2

713, 793

[a] From corrected spectra. [b] Fluorescence quantum yields calculated against PCBM

(f=8.3ꢁ10^ꢀ4in toluene) and ZnTPPP (f=4.0ꢁ10^ꢀ2in CH₂Cl₂). [c] Excitation at

560 nm for porphyrins (548 nm for phosphorescence) and at 331 nm for fullerenes.

[d] Energy level of the excited state calculated from the luminescence maximum at

77 K. [e] Phosphorescence from triplet excited state.

population of the eclipsed isomer in the regioisomeric mixture.

We also recorded the emission spectra during the titration

of CP6 with CP7 by exciting at the isosbestic point (427 nm).

The emission spectrum of a solution of CP6 in dichlorome-

thane ([CP6]=0.5 mm) displayed two characteristic and well-

defined bands at 596 and 642 nm (Figure 1b). The intensity of

the fluorescence bands of the CP6 solution was significantly

quenched upon addition of incremental amounts of fullerene

derivative CP7. This finding together with the observed red-

shift of the Soret band in the titration monitored using absorp-

tion spectroscopy support the existence of porphyrin–fullerene

p–p interactions to a significant extent in the structure of the

heterodimer CP6·CP7 complex. Probably, these interactions are

only present in the eclipsed regioisomer and they facilitate the

quenching of the porphyrin chromophore through a photoin-

orders of magnitude did not result in any noticeable change in

the values of the molar absorption coefficients and spectral

distribution. Similar results were obtained in dilution experi-

ments (see above). These results suggest that in the concentra-

tion range of 6–600 mm the registered absorption spectra cor-

respond to the homodimers CP6·CP6 and CP7·CP7, thereby

supporting the estimated values of K_D(CP6·CP6)=K_D-

(CP7·CP7)>10⁶m^ꢀ1. It is worth noting that the solutions of the

functionalised cyclopeptides CP6 and CP7, containing the cor-

responding self-assembled homodimers as major species, dis-

play slightly broader absorption bands and lower molar ab-

sorption coefficients than the reference models ZnTPPP and

PCBM (Figure S4 in the Supporting Information). Conversely, at

room temperature the emission spectral features and lifetimes

of solutions of the functionalised cyclopeptides CP6 and CP7

Chem. Eur. J. 2014, 20, 3427 – 3438

www.chemeurj.org

3432

ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

in dichloromethane, self-assembled in dimeric aggregates, are

almost identical to those of the corresponding model referen-

ces (Figure S5 in the Supporting Information and Table 1). Fluo-

rescence spectra of all these compounds were also determined

at 77 K in a glassy dichloromethane matrix (Figure S6 in the

Supporting Information). For the porphyrin-functionalised

cyclopeptide CP6 and ZnTPPP, the delayed spectra exhibited

a phosphorescence band at 800 and 808 nm, respectively, as-

cribable to the triplet excited state of the Zn–porphyrin

(Table 1 and Figure S7 in the Supporting Information).

Spectroscopic and photophysical properties of CP6·CP7

For practical reasons the photophysical determinations have

been conducted at higher concentrations than those used for

the determination of association constants. However, it was

previously confirmed that the self-sorting process takes place

at any concentration. During these experiments, carried out at

a 6.5ꢁ10^ꢀ5m equimolar concentration of CP6 and CP7,

charge-transfer (CT) bands around 720 nm were detected in

the absorption spectrum (Figure S8 in the Supporting Informa-

tion). These are due to the electronic interactions in the

ground state between the porphyrin and C₆₀chromophore in

close contact in the CP6·CP7 complex, and are often reported

for arrays containing these chromophores.^[41]This confirms the

interaction between the two components, at least in some re-

gioisomeric form. Molar absorption coefficients of the order of

10², or 10³m^ꢀ1cm^ꢀ1as a maximum, are generally reported for

this type of interaction. These figures, with a total absorbance

of 0.05 on the CT band of a 6.5ꢁ10^ꢀ5m concentration (Fig-

ure S8 in the Supporting Information), would allow one to es-

tablish that a large fraction of the partners are interacting, that

is, they must be present in the eclipsed conformation. This

agrees with the former discussion identifying this conforma-

tion as the major one, accounting for almost 80–90% of the

CP6·CP7 heterodimer.

In addition, a series of luminescence experiments in steady-

state and time-resolved regime were performed using an equi-

molar mixture of CP6 and CP7 at 1ꢁ10^ꢀ5m concentration in

dichloromethane, in which, according to the simulated specia-

tion profiles, the CP6·CP7 heterodimer represents approxi-

mately 95% of the total chromophores present (Figure S2 in

the Supporting Information). The luminescence spectra ob-

tained by exciting at 556 nm solutions of CP6 and the

CP6·CP7 complex in dichloromethane (equimolar mixture of

CP6 and CP7) are shown in Figure S9 in the Supporting Infor-

mation. At this wavelength almost exclusive excitation of the

porphyrin chromophore is achieved. The obtained results indi-

cate that there is a residual porphyrin luminescence of the

order of 28% for the solution containing the CP6·CP7 complex

compared with that of CP6. Excitation of the solution contain-

ing the CP6·CP7 complex at 327 nm, in which photons are ab-

sorbed predominantly by the fullerene unit (ca. 70%), leads to

a residual fluorescence at 710 nm of about 25% compared

with a CP7 solution under the same conditions (Figure 2a).

Taken together, these results indicate that in the solution con-

taining the heterodimer CP6·CP7 as the main component,

Figure 2. a) Uncorrected emission spectra of CP7 in CH₂Cl₂(1ꢁ10^ꢀ5m) and

of a solution containing CP6 and CP7 in CH₂Cl₂(1ꢁ10^ꢀ5m) upon excitation

at 327 nm. The full dots represent the calculated contribution owing to the

fullerene component in the mixture. Inset: magnification of the spectra.

b) and c) Integrated streak-camera intensity in the CP6 emission range, 580–

650 nm (b), and in the CP7 emission range, 740–830 nm (c), after a 35 ps ex-

citation at 532 and 355 nm, respectively.

a quarter of the total fluorescence of the two chromophores,

Zn–porphyrin and fullerene, is not quenched.

To better solve the photophysical dynamics of the supra-

molecular self-sorting process involving homodimers and hete-

rodimer CP6·CP7, time-resolved luminescence experiments

with nano- and picosecond resolution were performed.

Single-photon-counting determination at 600 nm, after exci-

tation at 560 nm of a 1ꢁ10^ꢀ5m equimolar solution of CP6 and

CP7 yielded a monoexponential decay with a lifetime of 1.6 ns

corresponding to the singlet excited state of the porphyrin

1

chromophore CP6. Any attempt to also detect by this tech-

nique the lifetime of the unquenched fullerene emission in the

same solution after excitation at 331 nm failed because of its

poor emission properties and overlap with the porphyrin ab-

sorption in this region. Streak-camera measurements upon ex-

citation at 532 nm of the same equimolar mixture of CP6 and

CP7 evidenced again a monoexponential decay of 1.6 ns in the

porphyrin fluorescence region (580–650 nm) without any de-

tectable fast component (Figure 2b). In the spectral region of

the fullerene emission (740–830 nm), we could observe, upon

355 nm excitation, a weak emission with a biexponential decay

(Figure 2c). The derived lifetimes are a short component of

60 ps representing approximately 71% of the decay and

a longer one of 1.3 ns accounting for the residual 29%. The

latter corresponds to the lifetime of the singlet state of the un-

quenched fullerene chromophore and represents the fraction

of chromophore not liable to quenching. On the basis of the

above data we can assume that in the case of the porphyrin

chromophore, the quenching detected by steady-state meth-

ods has to be ascribed to a reaction faster than the apparatus

Chem. Eur. J. 2014, 20, 3427 – 3438

www.chemeurj.org

3433

ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

instrumental resolution, 10 ps, because we are not able to ob-

serve it. On the contrary, in the case of the fullerene derivative,

the short lifetime of 60 ps can be attributed to a quenching re-

rated (CS) state of the heterodimeric complex resulting from

an intramolecular LUMO–LUMO electron transfer from the sin-

glet excited state of the porphyrin to the fullerene. A slower

decay on the microsecond scale (t=20 ms) with maxima at

480–490 nm and around 600–680 nm is ascribed to the por-

phyrin triplet (right inset, Figure 3).

1

action depleting the singlet excited state CP7. For both chro-

mophores, we could detect decay components with lifetimes

identical to that determined for the homodimers of CP6 and

CP7 (1.6 and 1.3 ns for CP6 and CP7 singlet excited states, re-

spectively). This observation confirms the existence of un-

quenched chromophores in the equimolar solution of CP6 and

CP7.

When the triplet spectrum corresponding to the equimolar

solution of CP6 and CP7 is compared, in the wavelength range

free from ground-state absorption (l>670 nm), with that mea-

sured from a solution of CP6, it becomes clear that the bands

of the porphyrin triplet excited state are reduced to approxi-

mately 25% in the former (Figure S11 in the Supporting Infor-

mation). In addition, the porphyrin triplet lifetime of the solu-

tion containing CP6·CP7 is significantly reduced (t=20 ms), rel-

ative to the lifetime of the porphyrin triplet in CP6 (180 ms).

A flash photolysis experiment after excitation at 355 nm, in

which the fullerene absorbs 50% of the light, was performed

by using an equimolar mixture of CP6 and CP7. A fast decay

with t=180 ns and centred on 460 and 670 nm can be ob-

served (Figure S12 in the Supporting Information). Like the

case described above for the selective excitation of porphyrin,

these bands are assigned to the porphyrin cation in the CS

To increase the time resolution, a pump and probe experi-

ment with sub-picosecond resolution was performed. The

measured transient absorption spectra at 1 ps after excitation

at 550 nm (Figure S10 in the Supporting Information), of an

equimolar mixture of CP6 and CP7 at 3ꢁ10^ꢀ5m concentration

(CP6·CP7 complex is the major species in solution, ꢁ95%) was

compared to the one measured from a solution containing

only the porphyrin cyclopeptide derivative CP6 at the same

concentration. The two spectra were markedly different (Fig-

ure S10 in the Supporting Information). In particular, the spec-

trum of the solution containing the CP6·CP7 complex as the

major component displays a higher absorbance around 600–

700 nm, a typical range for the absorption of the porphyrin

cation. From these results, we conclude that the porphyrin

cation is already fully formed in the solution containing the

CP6·CP7 complex at 1 ps delay as a consequence of an ultra-

fast electron-transfer reaction from the porphyrin excited state

to the fullerene. No further spectral evolution was registered

on the picosecond timescale. This finding increases the limit of

the quenching rate of the porphyrin excited state to k>

10¹²s^ꢀ1.

state of CP6⁺·CP7 , formed upon HOMO–HOMO electron

transfer to the fullerene excited state and decaying by recom-

bination to the ground state. At 2 ms after the flash, both the

porphyrin triplet (bands at 470 nm and ꢁ650 nm) and the typ-

ical band of the PCBM triplet (band at 700 nm) were present.

The band of the fullerene chromophore is reduced to approxi-

mately 20% with respect to that of the CP7 solution used as

a model (Figures S12 and S13 in the Supporting Information).

This observation is in fair agreement with the reduction of the

porphyrin triplet signal in the same equimolar mixture and

confirms that the parent singlet excited state was reduced ac-

cordingly, in agreement with fluorescence data. However, the

lifetime for the triplet of the fullerene residue in the equimolar

mixture is identical to that determined for CP7 (13 ms).

C

^ꢀC

Wider timescales were next explored. Transient absorbance

in the nanosecond range of the same equimolar solution (3ꢁ

10^ꢀ5m, CP6·CP7 ꢁ95%) excited at 532 nm (almost selective

excitation of the porphyrin) was registered after a nanosecond

flash photolysis experiment (Figure 3). A fast component with

a lifetime of 170 ns was detected (left inset Figure 3), with ab-

sorption bands around 460–470 nm and 650–680 nm. These

latter bands, as mentioned above, are typical for the porphyrin

Flash photolysis with extended near-infrared (NIR) detection

after predominant excitation of the fullerene moiety at 355 nm

in an equimolar solution of CP6 and CP7 at 3ꢁ10^ꢀ5m was also

performed. The transient absorption spectrum (Figure S14 in

the Supporting Information) shows typical bands for the fuller-

ene triplet and porphyrin cation together with a weak band

around 1040 nm, which is ascribable to the fullerene anion in

cation. Thus, we ascribe them to CP6⁺·CP7 , the charge-sepa-

C

^ꢀC

the CS state of the CP6⁺·CP7 complex. Although the absorb-

ance is rather weak and the signal to noise ratio rather poor,

the measured decay lifetime (160 and 200 ns, measured from

the two radical bands) is in reasonably good agreement with

the previous determinations (see above, 180 ns). This confirms

C

^ꢀC

that the CS state CP6⁺·CP7 recombine to the ground state

with a lifetime of (180ꢃ20) ns and that it is formed upon exci-

tation of both CP6 and CP7.

C

^ꢀC

Molecular modelling studies of CP6·CP7

Simple molecular mechanics studies suggested that both stag-

gered and eclipsed regioisomers of the heterodimeric complex

CP6·CP7 can adopt conformations of similar energy in which

Figure 3. Transient absorption spectra following nanosecond laser flash pho-

tolysis at 532 nm in a solution containing [CP6] and [CP7]=3ꢁ10^ꢀ5in

CH₂Cl₂. Insets show the time profiles at two representative wavelengths.

Chem. Eur. J. 2014, 20, 3427 – 3438

www.chemeurj.org

3434

ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

the porphyrin and fullerene units decorating their periphery

are spatially located at distances allowing direct p–p interac-

tion of chromophores. The shortest separations between

a carbon atom of the C₆₀residue and the Zn–porphyrin centre

are 2.9 ꢃ (staggered) and 2.8 ꢃ (eclipsed) (Figure S3 in the Sup-

porting Information). More accurate quantum calculations

starting from the previous energy-minimised structures were

performed with the Gaussian 09 package at the ab initio level

of theory (for methods, see the Supporting Information). The

shortest separation between a carbon atom of the C₆₀residue

and the Zn–porphyrin centre increased in both the staggered

and eclipsed conformations up to 3.8 ꢃ. The six hydrogen

bonds between CP6 and CP7 in the regioisomeric heterodim-

ers are maintained in both structures (Table S2 in the Support-

ing Information), and whereas the average values are quite

similar for both structures (3.051 and 3.042 ꢃ for the eclipsed

and staggered conformation, respectively), the hydrogen-bond

lengths are more uniform in the eclipsed one, as indicated by

their standard deviation (0.031 and 0.067 ꢃ for the eclipsed

and staggered conformation, respectively), which suggests

a slightly more distorted structure in this latter case. Single-

point calculations on the optimised geometries led to an

eclipsed regioisomer 6.70 kcalmol^ꢀ1more stable than the stag-

gered one (see the Supporting Information).

Besides static calculations, molecular dynamics (MD) simula-

tions were performed to analyse the structural and dynamical

behaviour of both conformers (40 ns, using explicit chloroform

as solvent; for methods, see the Supporting Information).

These calculations constitute a powerful tool to analyse the

fine details of supramolecular structures, and provide informa-

tion that cannot be obtained by another technique. The MD

simulations starting from a structure in which the chromo-

phores are involved in a p–p interaction (Figure S3 in the Sup-

porting Information) produce an energy-equilibrated structure

with the porphyrin and fullerene residues separated by

a centre-to-centre distance of (21.2ꢃ6.3) ꢃ (Figure 4a). This

distance is in agreement with that previously reported from

Fçrster resonance energy transfer (FRET) experiments involving

pyrene and Dapoxyl chromophores attached to cyclic pep-

tides.^[34]Conversely, MD simulations run starting from the

eclipsed regioisomer with the stacked chromophores provided

equilibrated structures in which the fullerene and porphyrin

residues are much closer together ((6.6ꢃ1.1) ꢃ, Figure 4b). In

spite of having high root-mean-square deviation (RMSD)

values for both regioisomers ((4.04ꢃ0.56) and (6.96ꢃ1.95) ꢃ

for the eclipsed and staggered conformation, respectively), the

six hydrogen bonds between CP6 and CP7 that can be ideally

established between the two monomers are maintained for

both conformations. For the eclipsed regioisomer an extra hy-

drogen bond can be found in more than 50% of the structures

of the trajectory, corresponding to the CO and NH groups

present in the side chains of the a-amino acids to which the

fullerene and porphyrin units are attached.

Figure 4. Energy-equilibrated molecular structures produced in the molecu-

lar dynamic simulations: a) Eclipsed and b) staggered regioisomers of the

CP6·CP7 complex. It is worth noting that the initial structures for both re-

gioisomers displayed a p–p interaction between the chromophores. See the

text for details.

ble for the eclipsed structure (interaction energies approxi-

mately 28 kcalmol^ꢀ1more stable than for the staggered regio-

isomer). Cluster analyses for the structures in the trajectory

generated from MD simulations led to two representative

structures of the most populated clusters for the eclipsed and

staggered conformations that are shown in Figure 4. It is

worth mentioning that during the MD run, a chloroform mole-

cule is encapsulated early on in the inner cavity of both re-

gioisomers, similarly to what has been described in related

studies. The chloroform molecule, once included, remains in

the cavity for the duration of the MD simulation.^[42]In conclu-

sion, molecular modelling experiments suggest that the fast

photoreactivity observed leading to the quenching of the sin-

glet excited states of both chromophores is most likely to

occur in the eclipsed regioisomer of CP6·CP7.

Photoinduced processes within CP6·CP7

Figure 5 depicts an energy-level diagram for the deactivation

processes occurring from the two possible singlet states of the

CP6·CP7 complex based on the calculated energies for the sin-

glets and ZnTPPP triplet (Table 1) and on the reported triplet

excited state of a methanofullerene related to PCBM of

1.57 eV.^[43]The energy level of the CS state corresponding to

Analyses of the different contributions to the energy during

the MD simulations suggest that whereas the Coulombic ener-

gies for the eclipsed and staggered conformations are very

similar, the Lennard-Jones contribution is much more favoura-

Chem. Eur. J. 2014, 20, 3427 – 3438

www.chemeurj.org

3435

ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

a charge-recombination process in the order of 10^ꢀ7s, a re-

markable lifetime for reaction partners in close contact, but

not unusual for fullerene-based arrays.^{[28,30,47–49]}

As stressed above, approximately 25% of the species in the

equimolar mixture of CP6 and CP7 survive quenching when

excited to the singlet state. These singlet excited states decay

unperturbed (as the models) and also lead to triplet excited

states ³CP6 and ³CP7. This also corresponds, roughly, to the

observed yield of triplet excited states that survive, which are

of the order of approximately 20%. The triplet excited state of

the porphyrin chromophore in solution of pure CP6 is ex-

tremely long lived, 180 ms, and this could allow either inter-

and/or intramolecular electron-transfer processes yielding

CP6⁺and CP7 in solutions containing the two cyclopeptidic

derivatives. Interestingly, our results show that the lifetime of

the triplet state of the porphyrin is significantly reduced to

20 ms in the equimolar solution of CP6 and CP7. In striking

C

^ꢀC

Figure 5. Energy-level scheme and deactivation processes in the complex

CP6·CP7 occurring upon excitation.

3

contrast, the CP7 state with 13 ms lifetime does not live long

enough to react either diffusively or intramolecularly with the

porphyrinic chromophore in the heterodimer CP6·³CP7. Conse-

3

the transfer of an electron from the porphyrin to the fullerene

unit can be derived from the redox data reported on the litera-

ture. PCBM is reduced at ꢀ1.08 V versus ferrocene/ferrocenium

(Fc/Fc⁺) in polar solvents,^[44]which can be converted to

ꢀ0.63 V versus the saturated calomel electrode (SCE). The oxi-

dation potential for ZnTPPP was measured as +0.87 V versus

Ag/AgCl,^[45]which converts to +0.82 V versus SCE. The energy

level of the CS state in dichloromethane can simply be as-

sumed to coincide with the energy necessary to oxidise the

donor and to reduce the acceptor, that is, approximately

1.4 eV obtained by adding the two potentials.^[46]

quently, the lifetime of CP7 is unaltered in the equimolar mix-

ture of CP6 and CP7 with respect to that determined for solu-

tions of pure CP7.

The assignment of structure to the species/aggregates that

do not experience fluorescence quenching in an equimolar

mixture of CP6 and CP7 is not simple and free of discussion.

First of all, we consider that in such a mixture the free mono-

mers and the homodimers, which are in equilibrium with the

heterodimer, are not involved in PET quenching processes.

However, the sum of free monomers and homodimers can be

reasonably set to a maximum of 5% in the solutions used for

the photophysical studies in which the heterodimer constitutes

95% of the mixture. Thus, we are forced to consider that the

staggered regioisomers of the heterodimeric complex CP6·CP7

are also responsible for the residual observed fluorescence.

From the data discussed above, we estimate that the percent-

age of staggered regioisomers in solution is close to 20%

(0.22ꢁ95%). This figure would nicely complement the percent-

age of monomers and/or homodimers to an overall value of

25%, thereby providing a plausible explanation to the experi-

mental finding of 25% residual fluorescence.

The quenching reaction of the fluorescence of the chromo-

phores can be ascribed to either LUMO–LUMO electron trans-

fer (from excited singlet state of the porphyrin to fullerene,

step (2)), or HOMO–HOMO electron transfer (from porphyrin to

excited singlet fullerene, step (3)). We cannot exclude the oc-

currence of a parallel intramolecular energy-transfer reaction

from the excited singlet state of the porphyrin yielding the ex-

cited singlet state of the fullerene, step (1), since it is a spin-

and thermodynamically allowed process and therefore might

be very fast.

However, the sensitised CP6·¹CP7 state, formed by reaction

step (1), reacts with a lifetime of 60 ps to yield the same CS

The greater distances that exist between chromophores in

the staggered regioisomers are unlikely to power any intramo-

lecular PET process. A similar interpretation was already put

forward by some of us in explaining the reduced percentage

of electron transfer detected in related heterodimers contain-

ing fullerene and exTTF chromophores.^[28]Fluorescence lifetime

measurements of a solution containing the fullerene/exTTF

heterodimer as the main component showed a biexponential

decay with a 1:2 ratio and lifetimes of <0.1 and approximately

1.5 ns, respectively. The two lifetimes were assigned to the full-

erene units experiencing, on one side, electron transfer and on

the other, no quenching since the lifetime of the fullerene sin-

glet excited state was unaltered. The ratio of lifetimes fits

nicely with the distribution of the regioisomers of the hetero-

complex in the ground state, which was determined using

¹H NMR spectroscopy as 1:1:1 for the three possible forms.

1

state, step (3), as the excited CP6·CP7 state. Therefore, there

is very little difference if energy-transfer step (1) is competitive

1

or not. The exceptionally high reactivity of the CP6·CP7 state

that decays with an overall rate of k₁+k₂>10¹²s^ꢀ1compared

with that of CP6·¹CP7 (k₃=1.6ꢁ10¹⁰s^ꢀ1) is ascribable to the

higher driving force for the LUMO–LUMO reaction step (2)

(ꢀDG⁰=0.6 eV) compared with that of HOMO–HOMO reaction

step (3) (ꢀDG⁰=0.35 eV). In addition, the energy-transfer pro-

cess (step (1)) might also contribute to the increase of the por-

phyrin decay. Once formed, the CS state CP6⁺·CP7 , the

energy of which is lower than that of both triplet excited

states of the chromophores, can only deactivate to the ground

state. The DG⁰related to the charge recombination is of the

order of ꢀ1.4 eV, which is sufficiently inverted to display

C

^ꢀC

Chem. Eur. J. 2014, 20, 3427 – 3438

www.chemeurj.org

3436

ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

However, in the study at hand, we observed a significant in-

crease in the electron-transfer process, a 75:25 ratio, occurring

in the equimolar mixture of cyclopeptides, compared to the

33:66 described in the aforementioned example. Our results

are in line with the formation of the eclipsed isomer in a signifi-

cantly larger extent. The stability constant of the CP6·CP7

complex is at least one order of magnitude larger than that es-

timated for the fullerene/exTTF analogue. We explain the in-

crease in thermodynamic stability of CP6·CP7 by making use

of the attractive interaction between the two chromophores.

This additional interaction has also been claimed to be respon-

sible for distribution populations of eclipsed/staggered re-

gioisomers in the case of CP6·CP7 that are in the range of

90:10 to 78:22 depending on the reference value used for K₀

(see above). The latter distribution ratio fits nicely, as discussed

above, with the experimental findings.

tide CP6·CP7 powers the energetic advantage of the eclipsed

regioisomer. Molecular modelling studies, the calculated distri-

bution of the regioisomers’ population using the partially

bound state scenario and the determined stability-constant

value of the heterodimer CP6·CP7 are also in support of the

highly (energetic) preference for the eclipsed versus staggered

conformation in this complex.

Acknowledgements

We are grateful for financial funding of our research by

MINECO Gobierno de EspaÇa (CTQ2011-23014 and CTQ2010-

15725), DURSI Generalitat de Catalunya (2009SGR6868), Xunta

de Galicia (GPC2013-039) and ICIQ Foundation, by Italian CNR

(project PM.P04.010 “MACOL”), and PRIN 2010CX2TLM.

It should also be noted that in the present cyclopeptide

array, excitation of both chromophores is effective in produc-

ing electron transfer, at variance with the aforementioned case

in which excitation of the corresponding electron donor, exTTF,

was unable to produce electron transfer in any conformation,

very likely because of the lifetime (subnanosecond) of the ex-

cited state being too short. This fact results in a much higher

efficiency in photoconversion upon sunlight illumination for

the CP6·CP7 system discussed here.

Keywords: electron transfer

porphyrinoids · self-assembly

·

fullerenes

·

peptides

·

[1] B. Hille, Ionic Channels of Excitable Membranes, Sinauer Associates, Sun-

derland, MA, 2001.

[2] T. M. Fyles, Chem. Soc. Rev. 2007 , 36 , 33 5 –3 47.

[3] F. De Riccardis, I. Izzo, D. Montesarchio, P. Tecilla, Acc. Chem. Res. 2013 ,

46 , 278 1 –2 790.

[4] J. D. Hartgerink, T. D. Clark, M. R. Ghadiri, Chem. Eur. J. 1998 , 4 , 136 7 –

1372.

[5] D. T. Bong, T. D. Clark, J. R. Granja, M. R. Ghadiri, Angew. Chem. 2001 ,

113 , 101 6 –1 041; Angew. Chem. Int. Ed. 2001 , 40 , 98 8 –1 011.

[6] R. Chapman, M. Danial, M. L. Koh, K. A. Jolliffe, S. Perrier, Chem. Soc. Rev.

2012 , 41 , 602 3 –6 041.

[7] R. Garcia-FandiÇo, M. Amorꢀn, J. R. Granja, Synthesis of Supramolecular

Nanotubes. In Supramolecular Chemistry: From Molecules to Nanomateri-

als, Vol. 5 (Eds.: P. A. Gale, J. W. Steed), Wiley, New York, 2012, p. 2149.

[8] M. R. Ghadiri, J. R. Granja, R. A. Milligan, D. E. Mcree, N. Khazanovich,

Nature 1993 , 366 , 32 4 –3 27.

[9] M. R. Ghadiri, J. R. Granja, L. K. Buehler, Nature 1994 , 369 , 30 1 –3 04.

[10] N. Khazanovich, J. R. Granja, D. E. Mcree, R. A. Milligan, M. R. Ghadiri, J.

Am. Chem. Soc. 1994 , 116 , 601 1 –6 012.

[11] J. Sanchez-Quesada, M. R. Ghadiri, H. Bayley, O. Braha, J. Am. Chem. Soc.

2000 , 122 , 1175 7 –1 1766.

[12] R. J. Brea, C. Reiriz, J. R. Granja, Chem. Soc. Rev. 2010 , 39 , 144 8 –1 456.

[13] D. Ranganathan, Acc. Chem. Res. 2001 , 34 , 91 9 –9 30.

[14] J. R. Granja, M. R. Ghadiri, J. Am. Chem. Soc. 1994 , 116 , 1078 5 –1 0786.

[15] M. Richman, S. Wilk, M. Chemerovski, S. K. T. S. Warmlander, A. Wahl-

strom, A. Graslund, S. Rahimipour, J. Am. Chem. Soc. 2013 , 135 , 347 4 –

3484.

[16] M. Mizrahi, A. Zakrassov, J. Lerner-Yardeni, N. Ashkenasy, Nanoscale

2012 , 4 , 51 8 –5 24.

[17] T. D. Clark, J. M. Buriak, K. Kobayashi, M. P. Isler, D. E. McRee, M. R. Gha-

diri, J. Am. Chem. Soc. 1998 , 120 , 894 9 –8 962.

Conclusion

In summary, we have thermodynamically characterised the

self-sorting process leading to the formation of the cyclopep-

tide heterodimeric assembly CP6·CP7. The assembled heterodi-

mer bears one electron-donor unit (zinc–porphyrin) and one

photoactive electron-acceptor unit (C₆₀) in the two adjacent

monomers. A high association constant value (ꢁ10⁸m^ꢀ1) has

been determined for a solution of heterodimer CP6·CP7 in di-

chloromethane, which hints to the existence of strong interac-

tions between the electron-active groups. Such interaction has

been evidenced by UV/Vis absorption and fluorescence spec-

troscopy titration experiments. In addition, accurate photo-

physical studies of the solution containing the heterodimer

have been undertaken. An efficient photoinduced electron-

transfer process, between the porphyrin core and the fullerene

moiety, takes place in approximately 80% of the heterodimers

CP6·CP7 upon excitation of both chromophores. This yields

a radical-ion pair charge-separated state, CP6⁺·CP7 , with an

electron localised on the fullerene unit and a hole localised on

the porphyrin component. This charge-separated state has

been found to have a lifetime of 180 ns and recombine to the

ground state.

C

^ꢀC

[18] C. Reiriz, R. J. Brea, R. Arranz, J. L. Carrascosa, A. Garibotti, B. Manning,

J. M. Valpuesta, R. Eritja, L. Castedo, J. R. Granja, J. Am. Chem. Soc. 2009 ,

131 , 11335.

[19] M. Amorꢀn, L. Castedo, J. R. Granja, J. Am. Chem. Soc. 2003 , 125 , 284 4 –

2845.

[20] N. Rodrꢀguez-Vꢄzquez, S. Salzinger, L. F. Silva, M. Amorin, J. R. Granja,

Eur. J. Org. Chem. 2013 , 347 7 –3 493.

[21] R. J. Brea, M. Amorin, L. Castedo, J. R. Granja, Angew. Chem. 2005 , 117 ,

585 6 –5 859; Angew. Chem. Int. Ed. 2005 , 44 , 571 0 –5 713.

[22] R. Garcꢀa-FandiÇo, L. Castedo, J. R. Granja, S. A. Vazquez, J. Phys. Chem. B

2010, 114, 4973–4983.

[23] M. R. Wasielewski, Acc. Chem. Res. 2009 , 42 , 191 0 –1 921.

[24] A. Magnuson, M. Anderlund, O. Johansson, P. Lindblad, R. Lomoth, T.

Polivka, S. Ott, K. Stensjç, S. Styring, V. Sundstrçm, L. Hammarstrçm,

Acc. Chem. Res. 2009 , 42 , 189 9 –1 909.

The large amount of electron transfer observed in the equi-

molar mixture of CP6 and CP7 (ca. 75%) compared with the

previously described counterparts (ca. 33%) is attributed to

the preferential assembly of heterodimer CP6·CP7 in an

eclipsed conformation for the former, whereas the staggered

conformation was preferred in the latter heterodimer. The exis-

tence of attractive interactions between the peripheral sub-

stituents, C₆₀and Zn–porphyrin, in the heterodimeric cyclopep-

Chem. Eur. J. 2014, 20, 3427 – 3438

www.chemeurj.org

3437

ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper

[25] L. Flamigni, J.-P. Collin, J.-P. Sauvage, Acc. Chem. Res. 2008 , 41 , 85 7 –8 71.

[26] D. Gust, T. A. Moore, A. L. Moore, Acc. Chem. Res. 2009 , 42 , 189 0 –1 898.

[27] N. Martꢀn, L. Sꢄnchez, M. A. Herranz, B. Illescas, D. M. Guldi, Acc. Chem.

Res. 2007, 40, 1015–1024.

[28] R. J. Brea, L. Castedo, J. R. Granja, M. A. Herranz, L. Sanchez, N. Martin,

W. Seitz, D. M. Guldi, Proc. Natl. Acad. Sci. USA 2007 , 104 , 529 1 –5 294.

[29] L. P. Hernꢄndez-Eguꢀa, R. J. Brea, L. Castedo, P. Ballester, J. R. Granja,

Chem. Eur. J. 2011 , 17 , 122 0 –1 229.

mum impact in the stability-constant value determined for the hetero-

dimer (2.1ꢁ10⁸m^ꢀ1). However, this increase is inherently associated

with a rise on K₀(10⁸m^ꢀ1). The use of K₀=10⁸m^ꢀ1in the calculation of

the distribution of bound states provides a complete inversion of the

population mixture that does not explain the experimental results ob-

tained in the photophysical studies.

[41] A. Hosseini, S. Taylor, G. Accorsi, N. Armaroli, C. A. Reed, P. D. W. Boyd, J.

Am. Chem. Soc. 2006 , 128 , 1590 3 –1 5913.

[30] L. P. Hernꢄndez-Eguꢀa, E. C. Escudero-Adan, I. C. Pintre, B. Ventura, L. Fla-

migni, P. Ballester, Chem. Eur. J. 2011 , 17 , 1456 4 –1 4577.

[42] R. Garcꢀa-FandiÇo, J. R. Granja, J. Phys. Chem. C 2013 , 117 , 1014 3 –

10162.

[31] T. Umeyama, H. Imahori, J. Phys. Chem. C 2013 , 117 , 319 5 –3 209.

[32] R. F. Enes, J.-J. Cid, A. Hausmann, O. Trukhina, A. Gouloumis, P. Vꢄzquez,

J. A. S. Cavaleiro, A. C. Tomꢅ, D. M. Guldi, T. Torres, Chem. Eur. J. 2012 ,

18 , 172 7 –1 736.

[43] N. Armaroli, G. Accorsi, F. Y. Song, A. Palkar, L. Echegoyen, D. Bonifazi, F.

Diederich, ChemPhysChem 2005 , 6 , 73 2 –7 43.

[44] F. B. Kooistra, J. Knol, F. Kastenberg, L. M. Popescu, W. J. H. Verhees, J. M.

Kroon, J. C. Hummelen, Org. Lett. 2007 , 9 , 55 1 –5 54.

[33] M. Amorꢀn, L. Castedo, J. R. Granja, Chem. Eur. J. 2005 , 11 , 654 3 –6 551.

[34] R. J. Brea, M. E. Vazquez, M. Mosquera, L. Castedo, J. R. Granja, J. Am.

Chem. Soc. 2007 , 129 , 165 3 –1 657.

[45] A. I. Oliva, B. Ventura, F. Wurthner, A. Camara-Campos, C. A. Hunter, P.

Ballester, L. Flamigni, Dalton Trans. 2009 , 402 3 –4 037.

[46] By introducing the usual corrections for Coulombic interactions and as-

suming a centre-to-centre distance between partners of the order of

10 ꢃ, a value of 1.3 eV would be obtained for the CS state energy level.

At the present level of discussion, this value can be considered equiva-

lent to the used one of 1.4 eV.

[47] A. Ciammaichella, P. O. Dral, T. Clark, P. Tagliatesta, M. Sekita, D. M.

Guldi, Chem. Eur. J. 2012 , 18 , 1400 8 –1 4016.

[48] Y. Terazono, G. Kodis, P. A. Liddell, V. Garg, T. A. Moore, A. L. Moore, D.

Gust, J. Phys. Chem. B 2009 , 113 , 714 7 –7 155.

[35] The monomer of the fullerene derivative CP7 has very low extinction

coefficients, and is present in small concentration during the titration

experiment. For these reasons, and to reduce the number of variables

of the fit, we considered that this species does not contribute signifi-

cantly to the registered absorption spectra (not coloured).

[36] K. Tashiro, T. Aida, Chem. Soc. Rev. 2007 , 36 , 18 9 –1 97.

[37] G. Gil-Ramꢀrez, S. D. Karlen, A. Shundo, K. Porfyrakis, Y. Ito, G. A. D.

Briggs, J. J. L. Morton, H. L. Anderson, Org. Lett. 2010, 12, 3544–3547.

[38] E. Chekmeneva, C. A. Hunter, M. J. Packer, S. M. Turega, J. Am. Chem.

Soc. 2008 , 130 , 1771 8 –1 7725.

[49] F. D’Souza, E. Maligaspe, K. Ohkubo, M. E. Zandler, N. K. Subbaiyan, S.

Fukuzumi, J. Am. Chem. Soc. 2009, 131, 8787–8797.

[39] C. A. Hunter, M. C. Misuraca, S. M. Turega, J. Am. Chem. Soc. 2011 , 133 ,

58 2 –5 94.

[40] We have observed that assuming an increase of one order of magni-

tude in the dimerisation constant of homodimers (10⁷m^ꢀ1) has a mini-

Received: October 28, 2013

Published online on February 13, 2014

Chem. Eur. J. 2014, 20, 3427 – 3438

www.chemeurj.org

3438

ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Article Doi

Self-sorting of cyclic peptide homodimers into a heterodimeric assembly featuring an efficient photoinduced intramolecular electron-transfer process

DOI: 10.1002/chem.201304200

Source and publish data:

Authors:

Article abstract of DOI:10.1002/chem.201304200

Full text of DOI:10.1002/chem.201304200

Products guided by the article

R&D Labs maybe for 1432038-38-4

Relevant to this article

Hot Product