Enhanced photochemical [6π] electrocyclization within the lipophilic protein binding site

Article abstract of DOI:10.1021/ol3003805

Photocyclization of N-methyldiphenylamine to N-methylcarbazole is achieved within the microenvironment provided by site I of serum albumins. Quantum yield determinations, combined with transient absorption spectroscopic detection of the dihydrocarbazole intermediate, demonstrate that protein encapsulation provides a subtle control of the kinetic parameters, leading to optimized efficiencies.

Full text of DOI:10.1021/ol3003805

ORGANIC
LETTERS
2012
Vol. 14, No. 7
1788–1791
Enhanced Photochemical [6π]
Electrocyclization within the Lipophilic
Protein Binding Site
Mireia Marin, Virginie Lhiaubet-Vallet,* and Miguel A. Miranda*
´
Instituto de Tecnologıa Quımica UPV-CSIC, Universidad Politecnica de Valencia,
Avda de Los Naranjos s/n, 46022 Valencia, Spain
´
lvirgini@itq.upv.es; mmiranda@qim.upv.es
Received February 16, 2012
ABSTRACT
Photocyclization of N-methyldiphenylamine to N-methylcarbazole is achieved within the microenvironment provided by site I of serum albumins.
Quantum yield determinations, combined with transient absorption spectroscopic detection of the dihydrocarbazole intermediate, demonstrate
that protein encapsulation provides a subtle control of the kinetic parameters, leading to optimized efficiencies.
Photochemical [6π] electrocyclizations are well-estab-
lished tools for the synthesis of a large variety of carbo-
cyclic or heterocyclic compounds.1 These reactions have
also found application in fields such as actinometry,2
photochromism,3 or design of molecular devices.4 From
a mechanistic point of view, they typically arise from the
first (π,π*) singlet excited state.5 An interesting exception
relies on the oxidative photocyclization of diarylamines
(see Figure 1 for the case of N-methyldiphenylamine),
which occurs through a multistep pathway involving two
consecutive triplet excited states. Initial formation of
4a,4b-dihydrocarbazoles is followed by oxidation to
carbazoles.2a,6 Oxygen plays a dual role: on the one hand,
it disfavors the reaction by quenching the amine triplet
excited state, whereas on the other hand, it is required as a
reagent for the final oxidative step.7 In this context, the
binding sites of serum albumins (SAs) would represent an
ideal microenvironment to optimize the photocyclization
efficiency by providing a subtle control of the involved
kinetic parameters.
Upon encapsulation, the photobehavior of guest mole-
cules is modulated not only as a result of conformational
restrictions imposed by the rigidity of protein cavities, but
also by the controlled diffusion of oxygen.8 These two
aspects have been previously tackled by comparing the
kinetic parameters of N-methyldiphenylamine photocycli-
zation in micro-heterogeneous media with those obtained
in water.9,10 In the case of cyclodextrins, a lower rate of
(1) (a) Bach, T.; Hehn, J. P. Angew. Chem., Int. Ed. 2011, 50, 1000. (b)
Albini, A.; Fagnoni, M. In Handbook of Synthetic Photochemistry;
Wiley-VCH: Weinheim, 2010.
€
(2) (a) Forster, E. W.; Grellmann, K.-H.; Linschitz, H. J. Am. Chem.
Soc. 1973, 95, 3108. (b) Yoshihara, T.; Yamaji, M.; Itoh, T.; Nishimura,
J.; Shizuka, H.; Tobita, S. J. Photochem. Photobiol., A 2001, 140, 7.
(3) (a) Fukumoto, S.; Nakashima, T.; Kawai, T. Angew. Chem Int.
Ed. 2011, 50, 1565. (b) Irie, M. Photochem. Photobiol. Sci. 2010, 9, 1535.
(4) (a) Irie, M. Chem. Rev. 2000, 100, 1685. (b) Golovkova, T. A.;
Kozlov, D. V.; Neckers, D. C. J. Org. Chem. 2005, 70, 5545.
(5) Turro, N. J.; Ramamurthy, V.; Scaiano, J. C. In Modern Mole-
cular Photochemistry of Organic Molecules; University Science Books:
Sausalito, California, 2010; pp 856.
(7) Fischer, G.; Fischer, E.; Grellman, K. H.; Linschitz, H.; Temizer,
A. J. Am. Chem. Soc. 1974, 96, 6267.
(8) (a) Alonso, R.; Jimenez, M. C.; Miranda, M. A. Org. Lett. 2011,
13, 3860. (b) Lhiaubet-Vallet, V.; Bosca, F.; Miranda, M. A. J. Phys.
Chem. B 2007, 111, 423–431. (c) Lhiaubet-Vallet, V.; Sarabia, Z.; Bosca,
F.; Miranda, M. A. J. Am. Chem. Soc. 2004, 126, 9538. (d) Vaya, I.;
Bueno, C. J.; Jimenez, M. C.; Miranda, M. A. ChemMedChem 2006, 1,
1015.
€
(6) (a) Grellmann, K.-H.; Kuhnle, W.; Weller, H.; Wolff, T. J. Am.
€
Chem. Soc. 1981, 103, 6889. (b) Gorner, H. J. Phys. Chem. A 2008, 112,
1245.
r
10.1021/ol3003805
Published on Web 03/14/2012
2012 American Chemical Society

345 nm, which does not overlap with the reactant
absorption spectrum.
Figure 1. Structures of N-methyldiphenylamine (A), N-methyl-
4a,4b-dihydrocarbazole (B), and N-methylcarbazole (C), to-
gether with an energetic diagram of the reaction mechanism.
photocyclization has been determined and attributed to
restricted rotation inside the rigid oligosaccharide struc-
ture.9 By contrast, no significant micellar effect has been
found within SDS or CTAB vesicles.10
Here, it will be shown that the efficiency of N-methyldi-
phenylamine photocyclization is maximized within the
microreactors provided by the lipophilic protein binding
sites. This is achieved by exploiting the retarded intrapro-
tein oxygen diffusion rate, which is markedly slower than
amine triplet quenching in solution and faster than dihydro-
carbazole dehydrogenation.
The reaction mechanism has been clearly established in
previous works (Figure 1).2a,6 After excitation and inter-
system crossing (ISC), a short-lived triplet excited state
(3A*) is generated. It may return back to the ground state
A in the submicrosecond time scale or undergo adiabatic
ring closure, leading to the N-methyl-4a,4b-dihydrocarba-
zole triplet excited state (3B*), which subsequently crosses
to the corresponding ground state (B). In the presence
of oxygen, dehydrogenation of B leads to the final
product, namely, N-methylcarbazole (C). Conversely,
under anaerobic conditions, B suffers an efficient ring-
opening back to the starting material A (>99% in non
polar solvents).2a,6b
The reaction efficiency was determined through the
photocyclization quantum yields. Experiments were per-
formed both under air and oxygen, in acetonitrile or in
bufferedaqueoussolution, inthepresenceorinthe absence
of SA. The overall study was performed with equimolar
amounts of A and SA, in order to avoid multiple occupa-
tion of the protein binding site. Photoconversion of A to C
was followed by UVꢀvis spectroscopy, as usually done for
actinometry measurements. Formation of C was mon-
itored by its characteristic absorption band peaking at
Figure 2. Top: concentration of carbazole C, determined from
the absorbance changes monitored at 345 nm, plotted as a
function of irradiation time for A (8.25 ꢁ 10ꢀ5 M) in MeCN
(squares), PBS/HSA (triangles), and PBS/BSA (circles) under
air (solid symbols) or oxygen (open symbols). Inset: UV spectral
changes upon irradiation of an aerated PBS solution of A in the
presence of HSA (molar ratio 1/1), during 323 nm monochro-
matic irradiation (from 0 to 30 min). Initial absorption spectrum
of HSA was subtracted. Bottom: HPLC traces obtained during
irradiation (λ = 300 nm) of A (1.65 ꢁ 10ꢀ4 M) in aqueous
buffer, in the presence of equimolar amounts of HSA.
Moreover, an isosbestic point at 323 nm was observed,
as expected for a clean process (Figure 2 top, inset). This
was confirmed by HPLC measurements, where C was
detected as the only significant photoproduct both in
solution and in protein media (Figure 2 bottom). Thus,
the absorbance changes at 345 nm were plotted as a
function of the irradiation time (Figure 2 top), and the
quantum yields (φC) were evaluated by the established
procedure using A in aerated acetonitrile as standard
(Table 1).6b An alternative determination of φC was also
performed by steady state fluorescence, monitoring the
increase of the C emission band (λmax = 367 nm) during
irradiation of nonfluorescent A. As shown in Table 1, the
data obtained with both methods were substantially coin-
cident. It is highly remarkable that the value of φC in-
creased largely in the presence of SAs, from 1.4- up to
4.6-fold. Interestingly, the highest φC, obtained for BSA/O2
(9) (a) Sur, D.; Purkayastha, P; Chattopadhyay, N. J. Photochem.
Photobiol., A 2000, 134, 17. (b) Sur, D.; Purkayastha, P.; Chattopadhyay,
N.; Bera, S. C. J. Mol. Liq. 2000, 89, 175.
(10) Roessler, N.; Wolff, T. Photochem. Photobiol. 1980, 31, 547.
Org. Lett., Vol. 14, No. 7, 2012
1789

system, was of 0.84. This is probably an upper limit,
imposed by the intersystem crossing of the starting amine,
which isthe bottleneck and has been reported tooccur with
a φISC = 0.8ꢀ0.9 in solution.
Table 1. Photocyclization Quantum Yields Determined by
UVꢀVis Absorption and (In Parentheses) Steady State
Fluorescence
PBS/MeCNa
MeCN
PBS/HSA
PBS/BSA
φC(air)
φC(O2)
0.46 (0.44)
0.55 (0.52)
0.45 (0.45)b
0.18 (0.19)
0.63 (0.61)
0.73 (0.72)
0.65 (0.67)
0.83 (0.84)
a A 3/1 ratio was used to ensure complete dissolution of A and C.
b Taken as standard value from ref 6b.
A satisfactory understanding of the influence of SAs on
the photocylization was obtained from the study of the
reaction intermediates. In this context, formation and
decay of dihydrocarbazole B turned out to be a good
reporter on the involved processes. This ground state
species was safely characterized by laser flash photolysis
(LFP), where it displayed a transient absorption band
peaking at 610 nm.
Nanosecond LFP was performed at 308 nm with an
excimer Xe/Cl laser, using PBS solutions of A (4.1 ꢁ 10ꢀ5
M) alone or in the presence of SAs (4.1 ꢁ 10ꢀ5 M). Under
aerobic conditions, the transient absorption spectrum ex-
hibited a weak maximum at 440 nm that decreased con-
comitantly with the increase of the much stronger band
centered at 610 nm (Figure 3, top). By comparison with the
literature, these species were assigned to 3B* and B,
respectively.2a,6b
Interestingly, in bulk solution (aerated PBS), the decay
at 610 nm followed a first order exponential law (eq 1) with
a lifetime of 15 μs, whereas in the presence of SAs, a more
complex fitting was required. The presence of two lifetime
components (τf = 15 μs and τb = 451 μs for HSA, or τf =
15 μs and τb = 384 μs for BSA) was established by
regression analysis of the decay curves according to eq 2.
This observation is clearly associated with the location of
the substrate in two different environments. The value of
15 μs, obtained for the shorter lifetime, corresponds to free
intermediate B in bulk solution, while the longer-lived
component reflects a tight interaction with one of the
two well-established SA binding cavities, namely, site I
or site II.11 In this context, an analogous treatment of the
decays for several [HSA]/[A] ratios was performed to
obtain af and ab preexponential factors (see eq2), necessary
to estimate the percentage of the substrate in each envir-
onment. As shown at the bottom of Figure 3, in the
presence of equimolar amounts of HSA, ca. 90% of the
substrate is complexed within the protein. A similar beha-
vior was observed for BSA.
Figure 3. Top: transient absorption spectra of an aerated solu-
tion of A in aqueous buffer, in the presence of HSA (4.1 ꢁ 10ꢀ5
M, molar ratio 1/1) obtained from 0.4 to 4 μs after the 308 nm
laser pulse. Inset: normalized decays monitored at 610 nm for A
alone in PBS (dotted line) or in the presence of HSA (black line)
or BSA (gray line). Bottom: percentage of protein-complexed
substrate, as a function of the number of protein equivalents
[SA]/[A] for HSA (closed squares) and BSA (crosses). Inset:
percentage of the complexed substrate in the presence of equi-
molar amounts of HSA, as a function of added displacer
concentration (IB = (S)-ibuprofen, OA = oleic acid).
ΔA ¼ ΔA0 þ af ꢁ eꢀt=τ
ð1Þ
ð2Þ
f
ΔA ¼ ΔA0 þ af ꢁ eꢀt=τ þ ab ꢁ eꢀt=τ
f
b
Next, site assignment was achieved using oleic acid (OA)
and (S)-ibuprofen (IB) as specific displacement probes for
site I and site II, respectively.12 Forthispurpose, increasing
amounts of OA were added to aerated solutions of A/HSA
(molar ratio 1/1), and transient decays were monitored at
610 nm. This led toa decrease of the long-livedcomponent,
ab, concomitantly with an increase of the free A fraction,
af (Figure 3 bottom, inset). By contrast, no significant
changes were observed when IB was used as displacer.
Similar results were obtained for BSA (see the Supporting
Information). Overall, the above observations clearly
(11) (a) Sudlow, G.; Birkett, D. J.; Wade, D. N. Mol. Pharmacol.
1975, 11, 824. (b) Sudlow, G.; Birkett, D. J.; Wade, D. N. Mol.
Pharmacol. 1976, 12, 1052.
(12) Montanaro, S.; Lhiaubet-Vallet, V.; Jimenez, M. C.; Blanca, M.;
Miranda, M. A. ChemMedChem 2009, 4, 1196.
1790
Org. Lett., Vol. 14, No. 7, 2012

of the efficiency of the cyclization process. For optically
matched MeCN solutions, a clear difference was ob-
served between oxygen, air, and argon bubbled solutions
(Figure 4 top); formation of B was dramatically disfavored
under aerated conditions because of the quenching of 3A*
by oxygen. By contrast, in SA/PBS media, an oxygen
independent growth was observed (Figure 4 bottom), as
anticipated from the fast decay of 3A* compared to
intraprotein diffusion of oxygen. The rate constant for
quenching of the 610 nm absorbing species by oxygen
within the protein was estimated to be 5.0 ꢁ 105 Mꢀ1 sꢀ1 in
HSA and 8.6 ꢁ 105 Mꢀ1 sꢀ1 in BSA (assuming an intra-
protein oxygen concentration similar to that reported
for water). This value is ca. 2 orders of magnitude lower
than oxygen diffusion within protein binding site I, which
explains why the final dehydrogenation to carbazole C is
not affected by complexation.6b
In summary, serum albumins provide an appropriate
environment for optimization of diarylamine photocycli-
zation. This is achieved by a subtle modulation of the
kinetic parameters, as a result of the geometrical restric-
tions imposed by the protein cavities and the controlled
diffusion of oxygen.
Acknowledgment. Thisworkwas fundedby the Spanish
Government (Project CTQ2009-13699, Ramon y Cajal
contract RyC-2007-00476 to V.L.-V. and JAE-predoc
contract to M.M.) and the Generalitat Valenciana
(Prometeo/2008/090).
Figure 4. Top: transient absorption growths (monitored at 610
nm) of argon (black line), air (gray line), and oxygen (dotted
line) bubbled solutions of A (5.5 ꢁ 10ꢀ5 M) in MeCN. Bottom:
parallel experiment in aqueous buffer, in the presence of HSA
(5.5 ꢁ 10ꢀ5 M).
Supporting Information Available. Experimental proce-
dures, displacement experiments and 610 nm transient
absorption growth for A within BSA. This material
is available free of charge via the Internet at http://
pubs.acs.org.
demonstrate that substrateꢀprotein interaction mainly
occurs at the more hydrophobic binding site I of SAs.
In addition, transient absorption at 610 nm (ΔA) corre-
sponds to B intermediate and can be taken as an indicator
The authors declare no competing financial interest.
Org. Lett., Vol. 14, No. 7, 2012
1791