Porphyrin and tetrabenzoporphyrin dendrimers: Tunable membrane-impermeable fluorescent pH nanosensors

Article abstract of DOI:10.1021/ja0341687

The pH dependencies of the UV - vis and fluorescent spectra of new water-soluble dendritic porphyrins and tetrabenzoporphyrins were studied. Because of extended π-conjugation and nonplanar distortion, the absorption and the emission bands of tetraaryltetrabenzoporphyrins (Ar₄TBP) are red-shifted and do not overlap with those of regular tetraarylporphyrins (Ar₄P). When encapsulated inside dendrimers with hydrophilic outer layers, Ar₄Ps and Ar₄TBPs become water soluble and can serve as pH indicators, with pK's adjustable by the peripheral charges on the dendrimers. Two new dendritic porphyrins, Gen 4 polyglutamic porphyrin dendrimer H₂P-Glu⁴OH (1) with 64 peripheral carboxylates and Gen 1 poly(ester amide) Newkome-type tetrabenzoporphyrin dendrimer H₂TBP-Nw¹OH (2) with 36 peripheral carboxylates, were synthesized and characterized. The pK's of the encapsulated porphyrins (pK_H2P-Glu⁴_OH = 6.2 and pK_H2TBP-Nw¹_OH = 6.3) were found to be strongly influenced by the dendrimers, revealing significant electrostatic shielding of the cores by the peripheral charges. The titration curves obtained by differential excitation using the mixtures of the dendrimers were shown to be identical to those determined for the dendrimers individually. Due to their peripheral carboxylates and nanometric molecular size, porphyrin dendrimers cannot penetrate through phospholipid membranes. Dendrimer 1 was captured inside phospholipid liposomes, which were suspended in a solution containing dendrimer 2. No response from 1 was detected upon pH changes in the bulk solution, while the response from 2 was predictably strong. When proton channels were created in the liposome walls, both compounds responded equally to the bulk pH changes. These results suggest that porphyrin dendrimers can be used as fluorescent pH indicators for proton gradient measurements.

Full text of DOI:10.1021/ja0341687

Published on Web 03/29/2003
Porphyrin and Tetrabenzoporphyrin Dendrimers: Tunable
Membrane-Impermeable Fluorescent pH Nanosensors
Olga Finikova,† Alexander Galkin,‡ Vladimir Rozhkov,† Maria Cordero,†
Cecilia Ha¨gerha¨ll,‡ and Sergei Vinogradov*,†
Contribution from the Department of Biochemistry and Biophysics, UniVersity of PennsylVania,
Philadelphia, PennsylVania 19104, and Department of Biochemistry, UniVersity of Lund,
Lund, Sweden
Received January 14, 2003; E-mail: vinograd@mail.med.upenn.edu
Abstract: The pH dependencies of the UV-vis and fluorescent spectra of new water-soluble dendritic
porphyrins and tetrabenzoporphyrins were studied. Because of extended π-conjugation and nonplanar
distortion, the absorption and the emission bands of tetraaryltetrabenzoporphyrins (Ar4TBP) are red-shifted
and do not overlap with those of regular tetraarylporphyrins (Ar4P). When encapsulated inside dendrimers
with hydrophilic outer layers, Ar4Ps and Ar4TBPs become water soluble and can serve as pH indicators,
with pK’s adjustable by the peripheral charges on the dendrimers. Two new dendritic porphyrins, Gen 4
polyglutamic porphyrin dendrimer H2P-Glu4OH (1) with 64 peripheral carboxylates and Gen 1 poly(ester
amide) Newkome-type tetrabenzoporphyrin dendrimer H2TBP-Nw1OH (2) with 36 peripheral carboxylates,
4
were synthesized and characterized. The pK’s of the encapsulated porphyrins (pKH P-Glu OH ) 6.2 and
2
1
pKH TBP-Nw OH ) 6.3) were found to be strongly influenced by the dendrimers, revealing significant electrostatic
2
shielding of the cores by the peripheral charges. The titration curves obtained by differential excitation
using the mixtures of the dendrimers were shown to be identical to those determined for the dendrimers
individually. Due to their peripheral carboxylates and nanometric molecular size, porphyrin dendrimers cannot
penetrate through phospholipid membranes. Dendrimer 1 was captured inside phospholipid liposomes,
which were suspended in a solution containing dendrimer 2. No response from 1 was detected upon pH
changes in the bulk solution, while the response from 2 was predictably strong. When proton channels
were created in the liposome walls, both compounds responded equally to the bulk pH changes. These
results suggest that porphyrin dendrimers can be used as fluorescent pH indicators for proton gradient
measurements.
Introduction
One especially intriguing class of core-modified dendrimers
consists of dendritic porphyrins or porphyrin dendrimers.7 The
initial interest in these molecules stemmed from their resem-
blance to heme-containing proteins. Porphyrins caged inside
dendritic shells are structurally reminiscent of enzymes, in which
The past two decades have been marked by rapid progress
in dendrimer synthesis.1 A broad variety of dendrimers has been
generated during this time, and as a result, current research has
focused primarily on new areas of dendrimer application. In
the interest of discovering their practical potential, dendrimers
with functional moieties and particularly dendrimers incorporat-
ing photo- or electroactive cores have been under extensive
scrutiny in recent years.2 A number of elegant models for
energy3 and electron-transfer studies,4 light-harvesting sys-
tems,3,5 and catalysts6 have been created using these compounds,
suggesting new ways to utilize the unique structural features of
dendrimers.
(3) For examples, see: (a) Stewart, G. M.; Fox, M. A. J. Am. Chem. Soc.
1996, 118, 4354. (b) Devadoss, C.; Bharathi, P.; Moore, J. S. J. Am. Chem.
Soc. 1996, 118, 9635. (c) Plevoets, M.; Vo¨gtle, F.; De Cola, L.; Balzani,
V. New J. Chem. 1999, 23, 63. (d) Gilat, S. L.; Adronov, A.; Fre´chet, J.
M. J. Angew. Chem., Int. Ed. 1999, 38, 1422. (f) Schultze, X.; Serin, J.;
Adronov, A.; Fre´chet, J. M. J. Chem. Commun. 2001, 1160. (g) Choi, M.
S.; Aida, T.; Yamazaki, T.; Yamazaki, I. Angew. Chem., Int. Ed. 2001, 40,
3194.
(4) For examples, see: (a) Sadamoto, R.; Nobuyuki, T.; Aida, T. J. Am. Chem.
Soc. 1996, 118, 3978. (b) Gorman, C. B.; Smith, J. C.; Hager, M. W.;
Parkhurst, B. L.; Sierzputowska-Gracz, H.; Haney, C. A. J. Am. Chem.
Soc. 1999, 121, 9958. (c) Camps, X.; Dietel, E.; Hirsch, A.; Pyo, S.;
Echegoyen, L.; Hackbarth, S.; Roder, R. Chem.sEur. J. 1999, 5, 2362.
(d) Capitosti, G. J.; Cramer, S. J.; Rajesh, C. S.; Modarelli, D. A. Org.
Lett. 2001, 3, 1645. (e) Rajesh, C. S.; Capitosti, G. J.; Cramer, S. J.;
Modarelli, D. A. J. Phys. Chem. B 2001, 105, 10175. (f) Toba, R.; Quintela,
J. M.; Peinador, C.; Roman, E.; Kaifer, A. E. Chem. Commun. 2001, 857.
(g) Ghaddar, T. H.; Wishart, J. F.; Thompson, D. W.; Whitesell, J. K.;
Fox, M. A. J. Am. Chem. Soc. 2002, 124, 8285. (g) Lor, M.; Thielemans,
J.; Viaene, L.; Cotlet, M.; Hofkens, J.; Weil, T.; Hampel, C.; Mullen, K.;
Verhoeven, J. W.; Van der Auweraer, M.; De Schryver, F. C. J. Am. Chem.
Soc. 2002, 124, 9918.
† University of Pennsylvania.
‡ University of Lund.
(1) For selected reviews, see: (a) Fre´chet, J. M. J. Science 1994, 263, 1710.
(b) Fischer, M.; Vo¨gtle, F. Angew. Chem., Int. Ed. Engl. 1999, 38, 885. (c)
Grayson, S. K.; Fre´chet, J. M. J. Chem. ReV. 2001, 101, 3819. (d) Tomalia,
D. A.; Fre´chet, J. M. J. J. Polym. Sci., A 2002, 40, 2719.
(2) For selected reviews, see: (a) Smith, D. K.; Diederich, F. Chem.sEur. J.
1998, 4, 1353. (b) Gorman, C. B.; Smith, J. C. Acc. Chem. Res. 2001, 34,
60. (c) Vo¨gtle, F.; Gestermann, S.; Hesse, R.; Schwierz, H.; Windisch, B.
Prog. Polym. Sci. 2000, 25, 987. (d) Hecht, S.; Fre´chet, J. M. J. Angew.
Chem., Int. Ed. 2001, 40, 74. (e) Dykes, G. M. J. Chem. Technol.
Biotechnol. 2001, 76, 903. (f) Stiriba, S. E.; Frey, H.; Haag, R. Angew.
Chem., Int. Ed. 2002, 41, 1329. (g) Cameron, C. S.; Gorman, C. B. AdV.
Funct. Mater. 2002, 12, 17.
(5) For examples, see: (a) Bar-Haim, A.; Klafter, J. J. Lumin. 1998, 76, 197.
(b) Adronov, A.; Gilat, S. L.; Fre´chet, J. M. J.; Ohta, K.; Neuwahl, F. V.
R.; Fleming, G. R. J. Am. Chem. Soc. 2000, 122, 1175. (c) Balzani, V.;
Juris, A. Coord. Chem. ReV. 2001, 211, 97. (d) Adronov, A.; Robello, D.
R.; Fre´chet, J. M. J. J. Polym. Sci., A 2001, 39, 1366.
9
4882
J. AM. CHEM. SOC. 2003, 125, 4882-4893
10.1021/ja0341687 CCC: $25.00 © 2003 American Chemical Society

Porphyrin and Tetrabenzoporphyrin Dendrimers
A R T I C L E S
hydrophobic hemes are buried deep inside the polypeptide
superstructures and are protected from direct interactions with
solvent and solutes. Mimicking the protein matrix, hydrophobic
dendritic wedges in porphyrin dendrimers modulate various
thermodynamic properties of the encapsulated cores, e.g. their
electrochemical potentials,7b-e,i and control the accessibility of
the cores to small molecules.7a,c,f,j Similar control is offered by
dendimers to other than porphyrin-encapsulated centers.2b,g;8 The
isolation, solubilization, and possibility to modulate the proper-
ties of porphyrins via dendritic encapsulation have all been
employed in catalysis6a-c,e,f and more recently in small molecule
sensing.9
Previously, as part of the studies aiming to design phosphors
for in vivo oxygen sensing, we synthesized a series of water-
soluble polyglutamic porphyrin dendrimers.9 In the case of free-
base porphyrin dendrimers,9c the protonation constants of the
cores were shown to change with the size of the attached
dendritic ligands, suggesting that the negative charges on the
outer layer of dendrimers stabilize the porphyrin dications,
possibly via through-space electrostatic shielding. The observed
effect revealed an opportunity to use the peripheral charges on
dendrimers as a pK-tuning tool. At the same time, the encapsu-
lation itself could be used to increase the solubility of porphyrins
and to prevent their aggregation in aqueous solutions. Consider-
ing the rich spectroscopy of tetrapyrrolic macrocycles and recent
advances in the chemistry of near-infrared absorbing extended
porphyrins,10 it became of interest to explore the possibility of
using porphyrin dendrimers as colorimetric and/or luminescent
pH sensors. Even though many such indicators are available,11
porphyrin dendrimers offer a number of advantages over the
existing commercial dyes, especially for pH measurements in
biological systems. One of these advantages is the complete
membrane impermeability of porphyrin dendrimerssan intrinsic
structural property of large dendritic spheres carrying multiple
peripheral charges. Of a special interest is the possibility of
selecting a number of porphyrins with nonoverlapping absorp-
tion and emission bands and using them simultaneously for pH-
gradient measurements in compartmentalized biological systems.
The design of such a pair was, in fact, the driving force behind
this study, since the elucidation of proton translocation mech-
anisms by respiratory chain enzymes has been the long-term
interest of some of the authors.12 In the following paragraphs
we attempt to clarify the ideas behind our design by summariz-
ing some relevant information from the porphyrin spectroscopy.
Optical absorption spectra of porphyrins contain a number
of bands in the UV-vis and near-infrared regions (Soret and
Q-bands). The positions of these bands depend on the exact
structures of the porphyrins; however, one general trend is that
the absorption maxima shift to the red upon macrocycle
distortion13 and/or upon conjugation of one or more pyrrole rings
with external π-systems.10,14
Porphyrins are highly fluorescent molecules. Their emission
quantum yields are in the range of 10-30%, and their
fluorescence can be excited at different locations throughout
the visible spectrum. The fluorescence of porphyrins exhibits
relatively large Stokes shifts (30-50 nm) and responds to
substitution effects in much the same way as does the absorption;
i.e., emission bands migrate to the red upon macrocycle distor-
tion and/or π-conjugation. Importantly, distortion also usually
diminishes the fluorescence quantum yields of porphyrinssa
fact that must be considered when attempting to use these
molecules as fluorescent markers.
(6) For examples, see: (a) Bhyrappa, P.; Young, J. K.; Moore, J. S.; Suslick,
K. S. J. Am. Chem. Soc. 1996, 118, 5708. (b) Bhyrappa, P.; Vaijayanthimala,
G.; Suslick, K. S. J. Am. Chem. Soc. 1999, 121, 262. (c) Kimura, M.; Shiba,
T.; Yamazaki, M.; Hanabusa, K.; Shirai, H.; Kobayashi, N. J. Am. Chem.
Soc. 2001, 123, 5636-5642. (d) Hecht, S.; Fre´chet, J. M. J. J. Am. Chem.
Soc. 2001, 123, 6959. (e) Davis, A. V.; Driffield, M.; Smith, D. K. Org.
Lett. 2001, 3, 3075. (f) Uyemura, M.; Aida, T. J. Am. Chem. Soc. 2002,
124, 11392. (g) Zhang, J. L.; Zhou, H. B.; Huang, J. S.; Che, C. M. Chem.s
Eur. J. 2002, 8, 1554. (i) Twyman, L. J.; King, A. S. H.; Martin, I. K.
Chem. Soc. ReV. 2002, 31, 69 and references therein.
Two imine nitrogens of a free-base porphyrin (H2P) are able
to attach protons forming porphyrin mono- (H3P+) and dications
(H4P2+): Traditionally, constants K3 and K4 are associated with
(7) (a) Jin, R. H.; Aida, T.; Inoue, S. Chem. Commun. 1993, 1260. (b)
Dandliker, P. J.; Diederich, F.; Gisselbrecht, J. P.; Louati, A.; Gross, M.
Angew. Chem., Int. Ed. Engl. 1996, 34, 2725. (c) Tomoyose, Y.; Jiang, D.
L.; Jin, R. H.; Aida, T.; Yamashita, T.; Horie, K.; Yashima, E.; Okamoto,
Y. Macromolecules 1996, 29, 5236. (d) Collman, J. P.; Fu, L.; Zingg, A.;
Diederich, F. Chem. Commun. 1997, 193. (e) Dandliker, P. J.; Diederich,
F.; Zingg, A.; Gisselbrecht, J. P.; Gross, M.; Louati, A.; Sanford, E. HelV.
Chim. Acta 1997, 80, 1773. (f) Pollak, K. W.; Sanford, E. M.; Fre´chet, J.
M. J. J. Mater. Chem. 1998, 8, 519. (g) Tomioka, N.; Takasu, D.; Takahashi,
T.; Aida, T. Angew. Chem., Int. Ed. 1998, 37, 1531. (i) Weyermann, P.;
Diederich, F. J. Chem. Soc., Perkin Trans. 1 2000, 4231. (j) Matos, M. S.;
Hofkens, J.; Verheijen, W.; De Schryver, F. C.; Hecht, S.; Pollak, K. W.;
Fre´chet, J. M. J.; Forier, B.; Dehaen, W. Macromolecules 2000, 33, 2967.
(k) Weyermann, P.; Diederich, F. HelV. Chim. Acta 2002, 85, 599. (l) Van
Doorslaer, S.; Zingg, A.; Schweiger, A.; Diederich, F. ChemPhysChem
2002, 3, 65.
(8) For examples, see: (a) Issberner, J.; Vo¨gtle, F.; De Cola, L.; Balzani, V.
Chem.sEur. J. 1997, 3, 706. (b) Kawa, M.; Fre´chet, J. M. J. Chem. Mater.
1998, 10, 286. (c) Gorman, C. B.; Hager, M. W.; Parkhurst, B. L.; Smith,
J. C. Macromolecules 1998, 31, 815. (d) Cardona, C. M.; Kaifer, A. E. J.
Am. Chem. Soc. 1998, 120, 4023. (e) Vo¨gtle, F.; Plevoets, M.; Nieger, M.;
Azzellini, G. C.; Credi, A.; De Cola, L.; De Marchis, V.; Venturi, M.;
Balzani, V. J. Am. Chem. Soc. 1999, 121, 6290. (f) Gorman, C. B.; Smith,
J. C. J. Am. Chem. Soc. 2000, 122, 9342. (g) Koenig, S.; Muller, L.; Smith,
D. K. Chem.sEur. J. 2001, 7, 979. (i) Stone, D. L.; Smith, D. K.; McGrail,
P. T. J. Am. Chem. Soc. 2002, 124, 856. (j) Goldsmith, J. I.; Takada, K.;
Abruna, H. D. J. Phys. Chem. B 2002, 106, 8504.
imine-N protonations, while the constants K1 and K2 correspond
to the removal of amine protons and the formation of porphyrin
anions.15 Upon protonation, both the absorption and emission
(fluorescence) spectra of porphyrins undergo strong changes,
which makes it possible to determine pK3 and pK4 by optical
methods. For the majority of porphyrins, constants pK3 and pK4,
are close (less than 2 pH units apart). Therefore, their individual
(9) (a) Vinogradov, S. A.; Wilson, D. F. AdV. Exp. Med. Biol. 1997, 428, 657.
(b) Vinogradov, S. A.; Lo, L. W.; Wilson, D. F. Chem.sEur. J. 1999, 5,
1338. (c) Vinogradov, S. A.; Wilson, D. F. Chem.sEur. J. 2000, 6, 2456.
(d) Rozhkov, V.; Wilson, D. F.; Vinogradov, S. Macromolecules 2002,
35, 1991.
(10) For review, see: Lash, T. D. Syntheses of novel porphyrinoid chromophores.
In The Porphyrin Handbook; Kadish, K. M., Smith, K. M., Guilard, R.,
Eds.; Academic Press: New York, 2000.
(11) Handbook of Fluorescent Probes and Research Products, Web edition at:
http://www.probes.com/handbook/Ch. 21; Molecular Probes, Inc.
(12) (a) Ha¨gerha¨ll, C. Biochim. Biophys. Acta 1997, 1320, 107. (b) Galkin, A.
S.; Grivennikova, V. G.; Vinogradov, A. D. FEBS Lett. 1999, 451, 157.
(13) Senge, M. O. Highly substituted porphyrins. In The Porphyrin Handbook;
Kadish, K. M., Smith, K. M., Guilard, R., Eds.; Academic Press: New
York, 2000.
(14) Lash, T. D. J. Porphyrins Phthalocyanines 2001, 5, 267.
(15) Hambright, P. Chemistry of water soluble porphyrins. In The Porphyrin
Handbook, Kadish, K. M., Smith, K. M., Guilard, R., Eds.; Academic
Press: New York, 2000.
9
J. AM. CHEM. SOC. VOL. 125, NO. 16, 2003 4883

A R T I C L E S
Finikova et al.
values, when extracted from the titration curves, are subject to
significant errors. It is more convenient to use instead apparent
constants (pK or pK1/2),16 which represent a somewhat averaged
proton affinity of the macrocycle. The averaged pK’s are more
reproducible and entirely adequate for drawing the comparisons
between different types of porphyrins.
pK’s of many water-soluble porphyrins have been measured15
and shown to be dependent on the electronic effects of sub-
stituents, macrocycle planarity,17 and the peripheral charges
surrounding the protonation site.9c,16,18 Any one of these factors,
or all of them together, can be used to tune pK’s of porphyrins
and to bring them into the desired range.
Compared to regular tetraarylporphyrins (Ar4P), porphyrins
with extended π-conjugation exhibit considerable red-shifts of
absorption and emission bands in both free-base and dicationic
forms. In the case of tetraaryltetrabenzoporphyrins (Ar4TBP),
the simplest representatives of symmetrically π-extended Ar4-
Ps, such red shifts of the Soret bands are in the range of 40-
50 nm.10,14
Figure 1. Structures and optical spectra of meso-tetrakis((methoxycarbo-
nyl)phenyl)porphyrin (1a) and dodecakis(methoxycarbonyl)tetraphenyl-
tetrabenzoporphyrin (2a) in CH2Cl2 solutions. Insets show the corrected
fluorescence spectra. Porphyrin dications were formed upon addition of
TFA.
The positions of the bands can be further adjusted by adding
electron-donating or electron-withdrawing substituents X and/
or Y. Such a large spectral separation makes it feasible to
measure the fluorescence of Ar4TBPs in the presence of Ar4Ps
and vice versa, by means of differential excitation, completely
avoiding the interference between the spectra. Provided that
solubilizing substituents (e.g. dendrons) can be attached to both
Ar4Ps and Ar4TBPs and that the pK’s of both chromophors can
be brought into the desired pH interval, such a pair of porphyrin
dendrimers can be extremely useful for biological pH measure-
ments.
In this paper we present the studied spectroscopic properties
and protonations of a fluorescent dye couple, consisting of a
porphyrin dendrimer and a tetrabenzoporphyrin dendrimer. The
tetrabenzoporphyrin dendrimer reported in this work is the first
example of a dendrimer with a free-base tetrabenzoporphyrin
core19sa new water-soluble, fluorescent, near infrared dye.
Table 1. UV-Vis and Fluorescence Spectroscopic Data for 1a
and 2aa
compd
λSoret (nm)
λQ (nm)
λ
emis (nm) (λexcit
)
φfl (%)
1a fbb
1a dcc
2a fbb
2a dcc
419
441
480
510
514, 548, 590, 646 654, 717 (515) 15.3 ( 1.3
657
695 (462)
21.9 ( 1.6
5.9 ( 0.8
8.3 ( 0.2
509, 652, 703
652, 711
718, 803 (490)
739, 801 (490)
a Abbreviations: fb, free base; dc, dication. b Solution in CH2Cl2.
c Solution in CH2Cl2/TFA.
dendrimers. The structures and the spectra of the selected
representatives are shown in Figure 1, and the spectroscopic
data are summarized in Table 1. The synthetic details are
discussed in the following section.
The Soret bands of 1a and 2a free bases (fb) are 60 nm apart.
Upon protonation both peaks shift by about 20-30 nm to the
red, which ensures their minimal overlap in dicationic forms
as well. The extinction coefficients at the Soret maxima of both
Results and Discussion
1fb
2fb
porphyrins, i.e., log ꢀSoret ) 5.6 and log ꢀSoret ∼ 4.9 (in
CH2Cl2), provide a means for efficient excitation. Upon pro-
tonation, four visible Q-bands of the free-base 1a merge into a
single peak at 657 nm. A similar transformation occurs with
the somewhat less resolved bands of 2a, which give rise to two
distinct peaks at 652 and 711 nm. Maxima of 2a dication (dc)
differ much less from those of the free base than in the case of
1a. This is perhaps a consequence of a much weaker structural
deformation upon protonation in the case of the already severely
distorted Ar4TBP macrocycle.20 In tetraarylporphyrins (Ar4P)
Porphyrin Cores. Porphyrins from two basic classes, i.e.,
meso-tetraarylporphyrins (Ar4P) and meso-tetraaryltetraben-
zoporphyrins (Ar4TBP), were chosen for this study as cores for
(16) Valiotti, A.; Adeyemo, A.; Williams, R. F. X.; Ricks, L.; North, J.;
Hambright, P. J. Inorg. Nucl. Chem. 1981, 43, 2653.
(17) (a) Medforth, C. J.; Smith, K. M. Tetrahedron Lett. 1990, 31, 5583. (b)
Barkigia, K. M.; Berber, M. D.; Fajer, J.; Medforth, C. J.; Renner, M. W.;
Smith, K. M. J. Am. Chem. Soc. 1990, 112, 8851. (c) Takeda, J.; Ohya, T.;
Sato, M. Inorg. Chem. 1992, 31, 2877.
(18) Kohata, K.; Higashio, H.; Yamaguchi, Y.; Koketsu, M.; Odashima, T. Bull.
Chem. Soc. Jpn. 1994, 67, 668.
(19) A series of Pd tetrabenzoporphyrin dendrimers for in vivo oxygen imaging
were recently synthesized: Rietveld, I.; Kim, E.; Vinogradov, S. Tetra-
hedron 2003, in press.
(20) Finikova, O.; Cheprakov, A.; Carroll, P.; Dalosto S., Vinogradov, S. Inorg.
Chem. 2002, 41, 6944.
9
4884 J. AM. CHEM. SOC. VOL. 125, NO. 16, 2003

Porphyrin and Tetrabenzoporphyrin Dendrimers
A R T I C L E S
such deformations are rather large, since Ar4P dications typically
have pronounced saddle-type geometries,21 while the Ar4P free
bases are nearly flat.
Fluorescence spectra of both 1a and 2a (Figure 1, insets) also
undergo hypochromic shifts upon protonation, although overlaps
between the emission bands of the free-base and dicationic forms
are more severe, especially in the case of 2a. Importantly, the
fluorescence quantum yield of 1a (15.3%) is more than two
times higher than that of 2a (5.9%), which implies that larger
quantities of the latter dye and/or stronger excitation should be
used to obtain equal fluorescence responses. In dicationic form,
the differences between the quantum yields of the dyes remain
about the same. The diminished fluorescence of 2a is probably
due to its strong nonplanar distortion, which is typical for Ar4-
TBPs.20,22 The distortion is known to increase the nonradiative
pathways of excited-state deactivation.23 On the other hand, the
fluorescence of 2a appears to be considerably stronger than that
of similarly distorted but not π-fused porphyrins, such as
tetracyclohexanoporphyrins.23a The reason for this increase is
probably the enhanced rigidity of the tetrabenzoporphyrin
macrocycle due to π-conjugation, which leads to a lower
probability of thermal deactivation. One should keep in mind,
however, that the quantum yield values obtained for tetraben-
zoporphyrins are generally less accurate because of the rapidly
fading efficiency of detection systems above 750-800 nm. This
is true even in the case of the specially IR-enhanced PMT
(Hamamatsu R2658P) used in our experiments.
The hydrolysis of the ester groups in both porphyrins produces
water-soluble meso-tetracarboxyphenylporphyrin (1b) and dode-
cacarboxytetraphenyltetrabenzoporphyrin (2b). Despite the pres-
ence of several carboxylic groups, 1b and even 2b are still quite
hydrophobic and readily aggregate in aqueous solutions upon
acidification (pH < 5). The pK3 of 1b has been reported to
have a value of 5.5,15 while the second constant pK4 could not
be measured due to the aggregation. The relatively high value
of pK3 of 1b is likely to be caused by the four negatively charged
peripheral carboxylates. Similar uncharged tetraarylporphyrins
usually exhibit much lower pK’s (2-4).15 The apparent pK of
2b, determined by spectrophotometric titration, has a value of
7.3,20 which represents a combination of two poorly resolved
sequential protonations with pK3 ) 7.9 and pK4 ) 6.4. In this
case, high pK values are due to the combined effect of
electrostatic shielding by ionized carboxylates, as in the case
of 1b, and the enhanced coupling of the carboxylates with the
protonation site via the π-conjugated benzo rings. Blocking all
12 carboxylates with methoxypoly(ethelene glycol) residues
results in a drop of pK by about 5 pH units.20 Such a strong
change indicates that tetrabenzoporphyrins with several electron-
withdrawing groups are intrinsically very acidicsmuch more
so than other nonplanar porphyrins.13,17 On the other hand,
although the apparent pK of 2b itself falls into the desired
interval (pH ) 6-8), this porphyrin could not be used in
Figure 2. Structures of H2P-Glu4OH (MW 8532) (1) and H2TBP-Nw1OH
(MW 5172) (2).
practical measurements due to its strong lipophilic nature. Not
only does 2b aggregate upon acidification, but it also binds to
phospholipid membranes, which affects its fluorescent proper-
ties. Our attempts to prepare liposomes in the presence of 2b
resulted in the extraction of the porphyrin into the lipid phase,
where its fluorescence was almost insensitive to the pH changes
in the medium. It was, therefore, necessary to create a
hydrophilic surrounding shell around the tetrabenzoporphyrin
2b to prevent its aggregation and improve its solubility.
Synthesis of Porphyrin Dendrimers. As mentioned in the
Introduction, a convenient method for the solubilization and
isolation of porphyrins in the aqueous phase is dendritic
encapsulation. For our particular case, employing dendritic
encapsulation had one extra advantage, namely, the possibility
to use peripheral charges on dendrimers to tune the pK’s of the
central cores. With that in mind, porphyrin 1b and tetraben-
zoporphyrin 2b were modified with dendritic ligands. The
structures of two newly synthesized porphyrin dendrimers are
shown in Figure 2.
(21) (a) Stone, A.; Fleisher, E. B. J. Am. Chem. Soc. 1968, 90, 2735. (b) Cheng,
B.; Munro, O. Q.; Marques, H. M.; Scheidt, W. R. J. Am. Chem. Soc.
1997, 119, 10732.
(22) Cheng, R. J.; Chen, Y. R.; Wang, S. L.; Cheng, C. Y. Polyhedron 1993,
12, 1353.
(23) (a) Gentemann, S.; Medforth, C. J.; Forsyth, T. P.; Nurco, D. J.; Smith, K.
M.; Fajer, J.; Holten, D. J. Am. Chem. Soc. 1994, 116, 7363. (b) Gentemann,
S.; Medforth, C. J.; Ema, T.; Nelson, N. Y.; Smith, K. M.; Fajer, J.; Holten,
D. Chem. Phys. Lett. 1995, 245, 441. (c) Sazanovich, I. V.; Galievsky, V.
A.; van Hoek, A.; Schaafsma, T. J.; Malinovskii, V. L.; Holten, D.;
Chirvony, V. S. J. Phys. Chem. B 2001, 105, 7818.
Compound 1 is the generation 4 (Gen 4) polygutamic
dendrimer with 1b as a core, and 2 is the Gen 1 Newkome-
type poly(ester amide) dendrimer with 2b as a core. Compound
2 is, to our knowledge, the first free-base Ar4TBP-core
dendrimer.19 It can be viewed as an intermediate structure
9
J. AM. CHEM. SOC. VOL. 125, NO. 16, 2003 4885

A R T I C L E S
Finikova et al.
Scheme 1 a
a Key: (i) DCC, pyridine, THF, rt, 48 h; (ii) H2, Pd/C, THF/MeOH, rt, 72 h, 65%; (iii) DCC, pyridine, DMF, rt, 72 h, 75%; (iv) LiOH/THF/MeOH, rt,
95%; (v) DCC, pyridine, DMF, rt, 7 days, 53%; (v) DCC, pyridine, 18-crown-6, DMF, 45 °C, 21 days, 37%.
between porphyrin dendrimers and phthalocyanine dendrimers,24
and constitutes yet another example of a photoactive core
dendrimer with near-infrared absorption and luminescence.
Compounds 1 and 2 and the related porphyrin dendrimers are
also abbreviated in the text as H2P-Glu4OH and H2TBP-Nw1-
OH, where H2P and H2TBP refer to the core porphyrin 1b and
core tetrabenzoporphyrin 2b, Glu4 and Nw1 designate poly-
glutamic and Newkome type poly(ester amide) dendrimers of
generations 4 and 1, respectively, and OH represents the terminal
groups.
OEt with 1b gave the Gen 2 compound H2P-Glu2OEt in 75%
yield, and the subsequent hydrolysis nearly quantitatively
afforded polycarboxylic acid H2P-Glu2OH. Further synthesis of
the Gen 3 compound H2P-Glu3OH followed the divergent path
described earlier.9c Attachment of the next layer of glutamates
to 32 terminal carboxyl groups on H2P-Glu3OH, however,
presented a significant obstacle due to the complete insolubility
of this compound in organic solvents. This problem was due in
part to the fact that Gen 3 porphyrin dendrimer could be isolated
only as a poly-Na+ or -K+ salt (H2P-Glu3O-M+) and not as a
free carboxylic acid. Hoping to facilitate the uptake of H2P-
Glu3O-K+ into the liquid organic phase, we attempted to use
18-crown-6 polyether, which proved to be an effective strategy.
Although the coupling of diethyl glutamates to H2P-Glu3OH
still required a lengthy reaction time (21 days), the progress of
the synthesis could already be detected soon after the addition
of the crown ester to the reaction mixture, as the purple
porphyrin became gradually extracted into the liquid phase. As
a result, the Gen 4 ethyl ester terminated dendrimer H2P-Glu4-
OEt could be synthesized in 37% yield. The MALDI-TOF
spectrum of this compound (MW 10326) revealed a distribution
of peaks spaced between 10349 (M+ + Na) and ∼8500 and
separated by 314 MW units. Notably, this mass difference is
most likely due to the fragmentation of the dendrimer rather
than to the presence of underderivatized products. The latter
would appear if some of the carboxyl groups on H2P-Glu3OH
did not undergo coupling with diethyl glutamates. Such “defec-
tive” dendrimers would give rise to the mass peaks decremented
in steps of 185 but not 314 units. Similar patterns have been
observed in the spectra of Pd tetrabenzoporphyrin dendrimers,
produced in the complementary syntheses.19 H2P-Glu4OEt was
smoothly hydrolyzed under basic conditions, and the resulting
poly-Na+ salt of H2P-Glu4OH (1) was isolated as a fluffy green
hygroscopic powder. Compound 1 has in total 60 glutamic units
and 64 peripheral carboxylate groups.
The attachment of dendritic wedges to the porphyrins required
initial modification of the macrocycles with suitable functional
groups. Porphyrin 1b could be easily prepared via Alder-Longo
condensation25 of pyrrole and 4-(methoxycarbonyl)benzaldehyde
and subsequent hydrolysis. The symmetrically positioned car-
boxyl groups on meso-phenyl rings allowed for the coupling of
four amino-terminal dendrons using conventional DCC chem-
istry. Polyglutamic and poly(ester amide) dendritic derivatives
of 1b have been synthesized in the past using this method.9;4d,e
In our own previous work we used the divergent method to
produce polyglutamic dendrimers with 1b as a core, which
succeeded up to Gen 3.9c Since an increase in the number of
charges on the dendrimer periphery has been shown to shift
the protonation pK’s of porphyrins in the desired direction, we
attempted, in this work, the synthesis of even larger polyglutamic
dendrimers, i.e., H2P-Glu4OEt and H2P-Glu4OH (1). Our initial
intention was to switch entirely to the convergent route, which
is known to give dendrimers with the highest monodispersity.1a,c
The convergent scheme could be implemented via the method
of synthesis of polyglutamic dendrons reported earlier.26,9d Using
this approach we were able to produce appropriate amounts of
the Gen 2 dendron; however, the isolation of the Gen 3
heptaglutamic dendron proved in our hands to be impractical
due to low yields and a tedious purification process. Therefore,
1 was prepared by employing a combination of the convergent
and divergent methods, following Scheme 1.
While Ar4Ps with various substituents are readily available,
synthesis of functionalized Ar4TBPs has long presented a
significant challenge.10 Recently, two improved methods were
proposed, charting the routes to Ar4TBPs with various peripheral
functional groups.27,28 In this work we made use of one of these
methods,28 which is advantageous because of its inexpensive
starting materials and shorter reaction sequence. The synthesis
At first, the Gen 2 triglutamic dendron Glu2OEt was
synthesized convergently in 65% yield. Condensation of Glu2-
(24) (a) Kimura, M.; Nakada, K.; Yamaguchi, Y.; Hanabusa, K.; Shirai, H.;
Kobayashi, N. Chem. Commun. 1997, 1215. (b) Brewis, M.; Clarkson, G.
J.; Goddard, V.; Helliwell, M.; Holder, A. M.; McKeown, N. B. Angew.
Chem., Int. Ed. 1998, 37, 1092. (c) Kimura, M.; Sugihara, Y.; Muto, T.;
Hanabusa, K.; Shirai, H.; Kobayashi, N. Chem.sEur. J. 1999, 5, 3495. (d)
Brewis, M.; Helliwell, M.; McKeown, N. B.; Reynolds, S.; Shawcross, A.
Tetrahedron Lett. 2001, 42, 813.
(25) Adler, A. D.; Longo, F. R.; Shergalis, W. D. J. Am. Chem. Soc. 1964, 86,
3145.
(26) Twyman, L.; Beezer, A. E.; Mitchell, J. C. Tetrahedron Lett. 1994, 35,
4423.
(27) (a) Ito, S.; Murashima, T.; Uno, H.; Ono, N. Chem. Commun. 1998, 1661.
(b) Ito, S.; Uno, H.; Murashima, T.; Ono, N. Tetrahedron Lett. 2001, 42,
45.
(28) Finikova, O.; Cheprakov, A.; Beletskaya, I.; Vinogradov, S. Chem.
Commun. 2001, 261.
9
4886 J. AM. CHEM. SOC. VOL. 125, NO. 16, 2003

Porphyrin and Tetrabenzoporphyrin Dendrimers
A R T I C L E S
Scheme 2 a
a Key: (i) DCC, BtOH, sym-collidine, DMF, 0 °C, 4 days, 81%; (ii) NaOH/MeOH, 30 min, 95%. See ref 28 and the Experimental Section for the details
of the syntheses leading to the core tetrabenzoporphyrin 2b.
of the core 2b, shown in Scheme 2, generally followed the
protocol reported earlier. The critical step in this method is the
oxidative aromatization of the preceding tetraaryltetracyclo-
hexanoporphyrin (Ar4TCHP, iv) into the Ar4TBP (v) by DDQ.
The reaction is very sensitive to the solvent, and in the case of
the porphyrin shown, it must be carried out in acetonitrile.
Important is also the choice of Cu as a temporary metal in Ar4-
TCHP (iv) for promoting the oxidation. Aromatization is
impossible when Ar4TCHPs are taken as free bases. Acidifica-
tion of the medium, which instantly follows the addition of
DDQ, leads to the virtually immediate formation of Ar4TCHP
dications, reflecting the extremely high basicity of these
porphyrins.20 The dications are totally inactive in aromatization,
and the entire process stops. Therefore, it is important to block
Ar4TCHPs with such metal ions that, on one hand, the formed
metalloporphyrins are stable enough to sustain aromatization
but, on the other hand, the metal can be pulled back out to
give the Ar4TBP free bases.
Synthesis of 2b was succeeded by the attachment of 12
Newkome fragments NwOMe.29 This was accomplished using
standard peptide chemistry. The Newkome building block
resembles glutamate in its composition but offers the advantage
of a higher branching number (BN ) 3) and therefore faster
dendrimer expansion with increasing generations. The dodeca-
substituted core combined with Newkome fragments allowed
us to build the dendrimer H2TBP-Nw1OMe (2c) with 36
Figure 3. Absorption and corrected fluorescence spectra of porphyrin
peripheral methoxycarbonyl groups (MW 5676) in a single
coupling step. The purification of 2c was accomplished by GPC
on Biorad S-X1 beads, generating material of quite high purity.
The MALDI-TOF spectrum of the sample revealed that the
compound was contaminated with only about 5% of the satellite
carrying one less Newkome dendron.
dendrimers H2P-Glu4OH (1) and H2TBP-Nw1OH (2) in aqueous solutions.
Properties of Porphyrin Dendrimers and pH Titrations.
Both compounds 1 and 2 are extremely water soluble and do
not precipitate through the entire pH range. We also could not
observe any spectroscopic evidence for aggregation of the
porphyrin dendrimers between pH 2 and pH 10.
The absorption and emission spectra of the compounds, as
well as the titration curves resulting from fluorescence rationing,
are shown in Figures 3 and 4. The excitation of fluorescence in
both cases was carried out at the isosbestic points in the Soret
regions, i.e., at 425 and 500 nm, respectively. The spectroscopic
data and the pK’s obtained from the changes in absorption and
emission spectra, as well as the comparative data for some
reference compounds, are summarized in Table 2.
Hydrolysis of 2c gave the water-soluble compound H2TBP-
Nw1OH (2) in a nearly quantitative yield. Similar to compound
1, 2 could not be precipitated from the aqueous solution even
upon strong acidification, and therefore, the solution had to be
purified from the excess salt by extensive dialysis. Subsequent
lyophilization produced the green hygroscopic powder 2.
(29) (a) Newkome, G. R.; Lin, X. Macromolecules 1991, 24, 1443. (b)
Newkome, G. R.; Lin, X.; Young, J. K. Synlett 1992, 53.
9
J. AM. CHEM. SOC. VOL. 125, NO. 16, 2003 4887

A R T I C L E S
Finikova et al.
seem to produce no spectroscopic effect on the encapsulated
cores. It appears that in compound 1 and in the reference Gen
2 dendrimers the attached branches seem to have almost no
influence on the porphyrin absorption bands. (For comparison,
the maxima of the core H2TCPP (1b) in water are λfb
416
Soret
nm and λdc
439 nm.) This indicates that polyglutamic
Soret
dendrons maintain rather open conformations in aqueous solu-
tions, leaving the porphyrins accessible to the solvent. On the
other hand, the electrochemical7b,e and small-molecule diffusion9
studies of porphyrin dendrimers, conducted earlier, reveal that
polyglutamic and poly(ester-amide) cages do impose encap-
sulation constraints on the porphyrin cores. These constraints,
while not very significant, are still perfectly measurable, indi-
cating that electrochemical and kinetic methods are much more
sensitive to the dendritic effects than static optical spectroscopy.
As mentioned above, the spectroscopic transitions observed
during pH titrations of porphyrins can include two protonations,
which may not be well resolvable if the corresponding pK’s
fall close on the pH scale. A sigmoidal curve, obtained by
plotting the ratio of absorbances vs pH, is, in this case, a
superposition of individual protonations, weighted by different
scaling factors. These factors depend on the actual magnitudes
of the changes induced by each protonation in the spectra at
the wavelengths selected. The product of the superposition
usually resembles a sigmoidal curve too but with a wider span
over the pH axis. If such a situation occurs, fitting the
experimental data using the Hasselbach equation with the fixed
parameter n ) 1 (eq 1) would be incorrect. Instead, one should
revert to an equation describing two sequential protonations or,
alternatively, to use the standard Hasselbach equation (eq 1)
but leaving the parameter n nonfixed. During the fitting process,
n would automatically adapt a value lower than 1, resulting in
an average apparent value for pK. The latter approach seems
more reasonable when pK’s are close, as resolving two
sequential protonations would require almost noiseless data.
In the eq 1, B and C are constants, related to the concentration
of solution and to the extinction coefficients (or emission
intensities) of the porphyrin at the wavelength(s) selected; R(pH)
is the intensity of absorption (or emission) for either the free
base or the dication as a function of pH. Alternatively, R can
refer to the ratio of intensities at two selected wavelengths.
Figure 4. Titration curves for porphyrin dendrimers 1 and 2 obtained from
rationing fluorescence intensities at the selected wavelengths. “pH down”
and “pH up” indicate the direction in pH change during the titration, i.e.,
from basic to acidic and vice versa.
Table 2. Photophysical Data and Apparent Protonation Constants
for Porphyrin Dendrimers 1 and 2 and Related Reference
Compounds
λfbSoret (nm),
λdcSoret (nm)a
pK
(n)
pKfluo (λ1max
,
c
b
abs
compd
λ2max
)
φfl (%)
H2P-Glu4OH (1)
H2P-Glu2OH
416, 440 6.4 (0.9) 6.2 (647, 685) 18.4 (fb); 36.2 (dc)
415, 437 5.4 (1.0)
416, 438 2.5 (1.1)
H2P-Glu2PEG350d
H2TBP-Nw1OH (2) 491, 510 5.6 (0.6) 6.3 (688, 742) 5.2 (fb); 13.6 (dc)
2be
478, 496 7.3 (0.6) 7.2 (808, 728)
475, 494 2.2 (1.1)
e
2b-(PEG350)12
a Abbreviations: fb, free base; dc, dication. b Apparent pKabs obtained
via rationing absorbance at the Soret peak of a free base (λfbSoret) to the
absorbance at the isosbestic point throughout the entire protonation range
and fitting the points with Hasselbach equation with nonfixed parameter n.
10n(pH-pK)
1 + 10n(pH-pK)
c pKfluo obtained via rationing intensities of emission at the wavelength λ1
max
R(pH) ) C
+ B
(1)
and λ2max, as shown in Figure 4, and analyzing the data in the same way as
for absorption. d H2P-Glu2PEG350 is a polyglutamic porphyrin dendrimer
with PEG350 terminal residues, where PEG350 designates methoxypoly-
(ethylene monomethyl ether) with average MW 350. See the Experimental
Section for details. e Data are taken from the ref 20. 2b-(PEG350)12 is 2b
esterified with 12 PEG350 residuessa water-soluble uncharged analogue
of the core porphyrin 2b.
In Table 2 we present the apparent absorption-derived (pKabs
)
and fluorescence-derived (pKfluo) constants obtained using eq
1 with nonfixed n and show them together with the correspond-
ing parameters n. Using this type of analysis is justified in our
case because protonations of the peripheral carboxylates in
dendrimers additionally “smooth out” the shape of the proto-
nation curve, making it more difficult to distinguish between
the individual porphyrin protonations. Accordingly, the pK’s
represent the weighted averages of two sequential protonations.
For compound 2 and the parent tetrabenzoporphyrin 2b the
n values reach as low as 0.6, revealing that the corresponding
transitions encompass more than a single proton addition and
that both protonations contribute significantly to the spectral
changes at the selected wavelengths. The fluorescence-derived
pK’s (pKfluo) are sensitive to the same effects as pKabs, and if
The spectra of the porphyrin dendrimers resemble those of
the cores, although, typical of water-soluble porphyrins, the
bands in aqueous solutions are broadened. In the case of com-
pound 2 and the other two tetrabenzoporphyrins (see footnote
d, Table 2), the significant differences among the absorption
maxima are most likely caused by electronic factors, since the
carboxyl groups at the periphery are coupled directly to the
macrocycles through the conjugated π-system. In contrast, the
three H2P dendrimers shown are different from one another only
by the groups which are quite distant from the porphyrin cores,
and therefore, the cores are spectroscopically insensitive to these
differences. The dendritic cages themselves, on the other hand,
9
4888 J. AM. CHEM. SOC. VOL. 125, NO. 16, 2003

Porphyrin and Tetrabenzoporphyrin Dendrimers
A R T I C L E S
the same color scale, and the surfaces were drawn at the same
distances from the atomic centers. The red color designates the
areas of positive potential (+δΨ), induced by the protons
attached to the imine nitrogens in the porphyrin cores. In the
case of the dendrimer with no peripheral charges (A) the red
spot and, therefore, the +δΨ value are noticeably larger than
in the peripherally ionized dendrimer (B). In the latter case, the
negative (blue) charges partially compensate for the central
positive charge, lowering the electrostatic energy of the dication
and decreasing the total protonation free energy of the porphyrin.
Figure 5. Maps of electrostatic potential (positive, red; negative, blue) of
the Gen 3 porphyrin dendrimers dications: H4P2+-Glu3OH (A) with neutral
surface; H4P2+-Glu3O- (B) with 32 peripheral ionized carboxylates.
To experimentally check whether the increase in pK was
indeed due to the electrostatic stabilization of dications, we
blocked the charges on the dendrimer H2P-Glu2OH with neutral
methoxypoly(ethylene glycol) fragments (PEG350), converting
it into a still completely water-soluble but peripherally uncharged
dendrimer H2P-Glu2PEG350. As seen from the table, the result
was a more than 3 pH units drop in pK, thus confirming our
assumption. The power of the stabilization effect, however,
seems to reduce rapidly with the distance separating these
charges from the core. The difference between the pK’s of 1
(pKabs ) 6.4) and H2P-Glu2OH (pKabs ) 5.4) is only 1 pH unit,
while 1 has 48 carboxylates more than the Gen 2 dendrimer
H2P-Glu2OH. At the same time, just 16 charges on H2P-Glu2-
OH (compared to no charges on H2P-Glu2PEG350) produce a
much stronger effect, probably because of their closer proximity
to the core.
the isosbestic points are not well defined, changing wavelengths
for rationing usually results in slight pK shifts. This “impre-
fection” in the protonation curves also explains why the pKfluo
values in Table 2 differ somewhat from the pKabs values. It also
demonstrates that by carefully choosing the emission wave-
lengths for any particular measurement, one can emphasize one
contributing protonation over another and thus shift the apparent
pKfluo value in the desired direction.
Both pKabs and pKfluo of compound 1 are much higher than
would be normally expected for regular planar tetraarylporphy-
rins without donor substituents (pK ) 2-4). The electrostatic
shielding by peripheral carboxylates, proposed earlier,9c is likely
to be responsible for the pK shift in this case, as well as in the
case of the Gen 2 dendrimer H2P-Glu2OH (pKabs ) 5.4).
The stabilization of the porphyrin dication by the surface
charges on the surrounding dendrimer translates effectively into
lowering of the core positive electrostatic potential and, there-
fore, decreasing the total free energy of the protonated porphyrin.
An illustration of this effect could be obtained by calculating
the maps of electrostatic potential on some arbitrary molecular
surfaces and comparing the contour plots generated for the
dications surrounded by protonated vs ionized peripheral
carboxyls. Such plots, produced for Gen 3 porphyrin dendrimer
dications H4P2+-Glu3OH (A) and H4P2+-Glu3O- (B), are shown
in Figure 5.
To obtain these maps, we selected a relatively opened
conformation of the dendrimer, in which one plane of the
porphyrin is well seen. Sampling of the conformations was done
by running molecular dynamics (MM+ force field, Hyperchem
5.0) at a relatively high temperature (1000 °C) in a vacuum,
followed by geometry optimization of a few selected conforma-
tions. In the selected dications of porphyrin dendrimers the
effective charges of +0.5 were assigned to all four inner-
porphyrin protons, to more realistically approximate the charge
distribution within the porphyrin cavity. In the case of fully
ionized structure H4P2+-Glu3O- (B), the charges of -0.5 were
assigned to all oxygen atoms in the terminal glutamic acid
residues. Modeling of the surface potential was performed using
Poisson-Boltzmann solver as implemented in GRASP soft-
ware.30 The solvent (water) was treated as a continuum with
the dielectric constant of 80, and the concentration of salt was
kept 0.1 M in both cases A and B.
In the case of 2 and the other two Ar4TBPs, more factors
than simple through-space electrostatic shielding affect the proto-
nation pK’s. As mentioned above, tetrabenzoporphyrins are
intrinsically less basic than regular porphyrins due to the
extended π-conjugation.20 A much higher than expected basicity
of 2b (pKabs ) 7.3) is likely to be caused by a combination of
the through-bond and through-space effects of adjacent carbox-
ylates. As a result, the pKabs of this porphyrin drops down to
2.2 when the peripheral groups are esterified with PEG350
chains (Table 2, compound 2b-PEG35012). The 36 peripheral
charges in 2, provided by 12 Newkome dendrons, create another
substantial stabilization shield around the central tetrabenzopor-
phyrin dication, bringing its pKabs up to 5.6 and the pKfluo to
6.3.
As seen in Figure 3, the overlap between the Soret bands of
the porphyrin dendrimers 1 and 2 is rather small, making it
possible to use the corresponding isosbestic points (λmax ) 425
nm and λmax ) 500 nm) for differential excitation of fluores-
cence of the porphyrins. To test whether such a spectral
separation is sufficient for independent pH determinations using
the dyes simultaneously, we carried out a titration of a solution
with both compounds 1 and 2. The corresponding changes in
the spectra are depicted in Figure 6.
In this case the emission measurements were performed using
an instrument with a regular detector (Hamamatsu R928 vs IR-
sensitive Hamamatsu R2658P), and so the shapes of the spectral
lines are somewhat different from those shown in Figure 3. The
insets show the pH-titration curves obtained by fluorescence
rationing.
According to Table 2, the emission quantum yield of the
dendrimer 2 is 3-4 times lower than that of 1. To bring the
fluorescence responses from the dyes to similar levels, the
amounts taken into the mixture were adjusted, which is reflected
by the relative intensities of the Soret absorptions in Figure 6A.
The differences between the peripherally charged (A) and
uncharged (B) forms of the dication H4P2+-Glu3OH are well
seen in the figure. The plots in both cases were produced using
(30) Nicholls, A.; Sharp, K. A.; Honig, H. Proteins 1991, 11, 281.
9
J. AM. CHEM. SOC. VOL. 125, NO. 16, 2003 4889

A R T I C L E S
Finikova et al.
Figure 6. Absorption (A) and fluorescence (B(1) and B(2)) spectra upon titration of the mixture of H2P-Glu4OH (1) and H2TBP-Nw1OH (2). The wavelengths
for fluorescence rationing were chosen to minimize the overlaps of responses from the compounds. The titrations were performed in a buffered (50 mM
phosphate) solution.
The mixture contained about 4 times less of the porphyrin
dendrimer 1. The excitation of fluorescence of 1 was carried
out at 430 nm, which is 5 nm off the absorption isosbestic point.
This emphasized the dication emission relative to that of the
free base and provided a cleaner isosbestic point at 647 nm in
the fluorescence spectrum (Figure 6B(1)). As seen from Figure
6B(1), the fluorescence spectrum of 1 is practically undisturbed
throughout the entire pH range, even in the presence of a large
Figure 7. Phospholipid liposome with H2P-Glu4OH (1) captured inside
excess of 2. The ratiometric curve (inset) is smooth and can be
and H2TBP-Nw1OH (2) contained in the outside solution (10 mM
used for pH measurements between pH 6 and 8 (pK ) 6.6).
When excited at 500 nm, the fluorescence of the mixture
consists mainly of the spectrum of dendrimer 2. At basic pH,
there is, however, a small peak around 647 nm originating from
1. This band is due to an overlap of the excitation with the
absorption Q-band of 1 (λmax ) 517 nm). Nevertheless, choosing
wavelengths for fluorescence rationing (λ1 ) 747 nm and λ2 )
839 nm) away from this peak allows one to monitor the titration
curve of 2 exclusively. As in the case of 1, rationing at different
wavelengths (Figure 6B(2), inset) leads to a more basic pK )
6.6 than in the individual titration of 2 (pK ) 6.3). Such a shift
is caused by an admix of transitions due to the first and the
second protonations in the fluorescence spectra and the resulting
indistinct isosbestic points.
phosphate). Fluorescence changes vs time were recorded upon addition of
acid (HCl) or base (KOH) (indicated by arrows). The amounts of acid and/
or base were calibrated to produce 1 pH unit jumps: A (black), dye 1
(inside); B (red), dye 1 (inside) in the presence of 0.5 mg/mL of gramicidin;
C (green), dye 2 (outside) in the presence or absence of gramicidin.
impermeable, and porphyrin dendrimers, like 1 and 2, seem to
be well suited to perform such a task.
Figure 7 shows a model system that we used to examine the
applicability of porphyrin dendrimers for pH-gradient measure-
ments.
Phospholipid liposomes were prepared by sonication of a
phospholipid film, a standard technique used routinely in bio-
chemical preparations.33 The average diameter of liposomes
prepared by this method is 20-30 nm, and the approximate
diameter of 1 is about 20-30 Å.34 This implies that only a few
molecules of 1 can, in principle, fit inside each vesicle. It must
be kept in mind, however, that at pH 7, and when the liposome
concentration level is such that the combined liposome internal
volume is, for example, 0.001 of the volume of the entire
solution, only 1 out of 10 000 liposomes contains a proton. The
concentration of the dye is usually at a submicromolar level
(10-6-10-7M), and so the distribution of the dye molecules is
close to that of the protons, i.e., about 10-3-10-4 dye molecules/
liposome. Even though measuring a “single” proton with just a
“single” molecule of a probe might at first glance appear unreal-
istic, averaging over a large number of vesicles produces the
same effect as a regular ensemble averaging. Indeed, there are
about 1014 liposomes of 30 nm diameter/1 mL of solution at
the assumed level of liposome concentration (0.001 of total
volume), and the overall pH measurement, therefore, becomes
macroscopic in nature.
The titration experiment described above revealed that
optically porphyrin dendrimers 1 and 2 do not interact and their
protonations are spectroscopically independent. It therefore
became of interest to model a system in which these dyes would
be separated spatially but resolved only optically, permitting
the monitoring of pH changes in separate microscopic compart-
ments. As mentioned in the Introduction, proton compartmen-
talization is the key to metabolic energy conservation.31 Precise
quantification of proton transport processes is critical for
understanding the function of biological machinery involved
in bioenergetics. A phospholipid liposome with a reconstituted
enzyme, i.e., a proteoliposome,32 can serve as a mimic of a
bacterial or mitochondrial membrane, in which the enzyme
exists in its natural environment. If a function of the enzyme is
to translocate protons across the membrane, the proton gradient
is built upon the enzyme working cycle. The quantification of
this gradient could be achieved if proton-sensitive optical probes
are placed inside and outside the liposome, where they would
report independently on pH changes. A system of this kind
would require the reporters themselves to be membrane-
The liposomes were prepared in a solution containing por-
phyrin dendrimer 1, so the dye was included in the interior of
(33) Driessen, A. J.M.; De Vrij, W.; Konings, W. N. Eur. J. Biochem. 1986,
154, 617.
(34) Molecular diameter was determined by taking a few measurements across
the calculated dendrimer molecule and averaging the obtained distances.
(31) Mitchell, P.; Moyle J. Nature 1967, 213, 137.
(32) Rigaud, J. L.; Pitard, B.; Levy, D. Biochim. Biophys. Acta 1995, 1231,
223.
9
4890 J. AM. CHEM. SOC. VOL. 125, NO. 16, 2003

Porphyrin and Tetrabenzoporphyrin Dendrimers
A R T I C L E S
Conclusions
the liposomes during the sonication process. To remove the
outside dye, the liposomes were rigorously purified by chro-
matography on a Sephadex G200 column, which contained a
layer of an anion-exchange resin at the bottom. The affinity of
the anion-exchange resin to the carboxylate-terminated den-
drimers is extremely high, allowing for complete removal of
the dye traces from the outside solution. The controls were set
by gentle centrifugation of the liposomes, followed by the
fluorescence measurements of the supernatant solutions, which
were found virtually fluorescence-free.
In this paper we show that dendritic cages are effective in
modulating porphyrin protonation constants (pK). This effect
is due to the peripheral charges on dendrimers, which are
responsible for the stabilization of porphyrin dications. The
resulting pK’s of normally quite acidic porphyrins can be moved
into the physiological pH range (pH 6-8), suggesting the
usefulness of porphyrin dendrimers as optical pH indicators.
By synthesizing a polyglutamic porphyrin dendrimer with a
tetraarylporphyrin (Ar4P) core and the dendritic tetraaryltetra-
benzoporphyrin (Ar4TBP), we succeeded in demonstrating that
in water solutions two porphyrinoid chromophors with dendritic
jackets and nonoverlapping absorption and emission spectra do
not interact optically and do not influence each other’s proto-
nation curves. The latter can be determined individually from
the mixtures of the dyes using differential excitation/emission
techniques. The peripheral charges on the dendrimers ensure
complete membrane impermeability of the synthesized indica-
tors, which was demonstrated by placing porphyrin dendrimers
with two spectroscopically different cores inside and outside
phospholipid vesicles and determining the pH in both compart-
ments independently using bulk spectroscopic measurements.
The synthesized dyes form an effective indicator pair for proton
gradient determination. This is a new potential application for
porphyrin dendrimers, which capitalizes on dendritic encapsula-
tion as a tool for the isolation and solubilization of optically
active cores, as well as on the modulation and tuning of their
physicochemical properties.
The liposomes containing dye 1 were placed in a 10 mM
phosphate buffer where the second dye 2 was dissolved at pH
7.1. Acid and base additions (indicated by arrows in Figure 7)
were precalibrated to produce 1.0 pH unit jumps, but the actual
pH, measured by a conventional electrode, was also recorded
continuously during the experiment. The fluorometer used in
our measurements (SPEX Fluorolog-2) was equipped with just
one PMT, permitting detection at only one wavelength at a time.
Because fluorescence rationing requires measurements of in-
tensities at two wavelengths, the emission monchromator had
to be moved physically, which significantly increased the detec-
tion time. Since it was desirable to monitor the pH changes
with at least 2 s time resolution, measurements with two dyes
were accomplished in two independent experiments. These
measurements, if performed in a single run, would require a
much longer acquisition times, since the excitation monochro-
mator in this case would have to be moved between the Soret
bands of 1 and 2 and two more emission intensities would have
to be measured. The changes in fluorescence ratios are shown
in Figure 7.
Experimental Section
General Methods. H and 13C spectra were recorded on a Varian
1
Unity-500 spectrometer. Optical spectra were recorded on a Perkin-
Elmer Lambda 35 UV/vis spectrophotometer, and fluorescence mea-
surements were performed on a SPF-500C (SLM Instruments, Inc.)
fluorometer, equipped with a Hamamatsu R928 PMT and on a SPEX
Fluorolog-2 spectrofluorometer (Jobin-Yvon Horiba, Inc.), equipped
with a infrared enhanced R2658P PMT. Matrix-assisted laser desorption
mass spectrometry (MALDI-MS) was performed at the MALDI-TOF
facility of the Wistar Institute of the University of Pennsylvania.
Column chromatography was carried out on alumina (Aldrich,
activated, neutral, Brockmann I, standard grade, ∼150 mesh, 58 Å)
and on silica gel (Merck, 230-400 mesh, 60 Å). GPC was carried out
on Biorad Biobeads S-X1 (mobile phase: THF) and Sephadex G-50
(mobile phase: H2O). All reagents were purchased from Sigma-Aldrich,
Inc., Lancaster Synthesis, Inc., and Acros Organics, Inc. Solvents were
purchased from Fisher Scientific, Inc., and purified using standards
methods.
Analysis of spectroscopic data was performed using Origin 6.0,
Microcal, Inc., software. The optimizations of dendrimer structures and
molecular dynamics simulations were done on a 2.5 GHz PC worksta-
tion using Hyperchem 5.0 (Hypercube, Inc.) software. MM+ force field,
as implemented in Hyperchem, was used in both optimizations and
dynamics calculations. The calculation of electrostatic potential maps
was done using GRASP software30 on an SGI workstation. Poisson-
Boltzmann solver, as implemented in GRASP, was used to build the
electrostatic potential maps.
The differences in responses of the outside and inside dyes
(1 and 2, respectively) to the pH jumps in the medium are
obvious. While dye 2 in the bulk instantaneously responded
upon additions of acid or base, the signal originated from 1
kept changing only slightly, indicating that this dye was well
isolated from the outside solution. A gradual decline of the ratio
R1 (liposome-encapsulated porphyrin dendrimer 1) upon addition
of HCl reflects acidification of the liposome’s internal space.
This process is due to a slow diffusion of protons across the
phospholipid membrane (Figure 7A). When proton channels
were made in the membranes by a special pore-forming peptide
gramicidin,35 fluorescence of dye 1 inside the liposomes reacted
immediately to the additions of acid or base (Figure 7B). Similar
results were obtained when a combination of nigericin and
valinomycin was used to make pores in the liposomes (not
shown). Such pores permit free movement of the protons in
and out of the liposomes, resulting in a fast equilibration of
external and internal pH’s. The pores, however, are too small
for the porphyrin dendrimer molecules, so these remain captured
inside the vesicles.36 As expected, the fluorescence response of
the outside dye 2 (ratio R2) appeared sensitive to the additions
of HCl or KOH regardless of whether the pores were made in
the liposomes (Figure 7C). This experiment clearly demonstrates
that the pair of dyes 1 and 2 can be used to quantify proton
concentrations inside and outside the vesicles using optical
measurements in bulk solutions.
UV-Vis and Fluorescence Measurements. To obtain a series of
samples for pH titrations, a small amount of porphyrin was dissolved
in 75-100 mL of 50 mM K2HPO4 solution and pH was adjusted by
additions of KOH or HCl. The absorbance of the solution (1 cm cuvette)
at a maximum of the porphyrin Soret band was kept below 1.5 OD for
spectrophotometric determination (pKabs) or at about 0.1-0.15 OD for
fluorescence measurements (pKfluo). A small aliquot (∼2-3 mL) of
the solution was transferred into a quartz cuvette, and the spectra were
recorded. Normally, two titrations were performed to establish the
(35) Harold, F. M.; Baarda, J. R. J. Bacteriol. 1967, 94, 53.
(36) Urry, D. W.; Goodall, M. C.; Glickson, J. D.; Mayers, D. F. Proc. Natl.
Acad. Sci. U.S.A. 1971, 68, 1907.
9
J. AM. CHEM. SOC. VOL. 125, NO. 16, 2003 4891

A R T I C L E S
Finikova et al.
reversibility, e.g. beginning at basic pH down to acidic and backward.
The ratio of the absorption intensities at two Soret peaks (free base vs
dication) or a ratio of a Soret peak vs the isosbestic point was plotted
against pH and fitted to the Hasselbach equation. The parameter n in
the Hasselbach equation was allowed to change during the search (see
the main text).
with hydrogen. Pearlman catalyst39 was cautiously added, and the
reaction mixture was stirred under hydrogen, until TLC (CH2Cl2-THF)
showed no presence of the original pyrrole-ester i. The catalyst was
filtered off, the solvent was removed under vacuum, and the resulting
product was refluxed in ethylene glycol for 40 min. The mixture was
poured onto crushed ice, brine was added to reduce the emulsion, and
the product was extracted with CH2Cl2 (4 × 50 mL). The organic phase
was thoroughly washed with water (200 mL), then with brine, and then
dried over Na2SO4, and the solvent was removed in a vacuum. The
resulting material was purified on a short (i.d. 2 × 10 cm) silica gel
column (eluent CH2Cl2-THF, 20:1). Yield of pyrrol ii: 0.73 g, 88%.
1H NMR (CDCl3): δ 8.05 (broad s, 1H), 6.5 (d, 2H, J ) 2.5 Hz), 3.65
(s, 6H), 2.88-3.27 (m, 6H).
Simultaneous titrations involving porphyrin and tetrabenzoporphyrin
dendrimers were done in much the same way, except that the concen-
trations of the compounds had to be adjusted to obtain similar fluor-
escent responses from 1 (H2P-Glu4OH) and 2 (H2TBP-Nw1OH). The
latter dendrimer had to be present in the mixture in excess (3-4 times),
because of its significantly lower fluorescence quantum yield and lower
sensitivity of the detection system in the longer wavelength range.
Fluorescence quantum yields were measured by comparing the
integrals of the corrected emission spectra of the samples with those
of free-base meso-tetraphenylporphyrin (H2TPP) in deaerated benzene.
The spectra were recorded using Hamamatsu R2658P detector, and in
the case of tetrabenzoporphyrins scanning was done up to 1000 nm.
The spectra were normalized by the optical density of the samples at
the excitation wavelength, relative photon intensity of the source, and
quantum efficiency of the detector throughout the entire emission range.
The reported quantum yield for H2TPP in benzene is 11%.37 The
measurements involving the core porphyrins 1a and 2a were performed
in deaerated CH2Cl2 solutions.
Liposome Preparation and Dye Inclusion Experiments. Phos-
pholipid vesicles containing dye 1 were prepared as follows. L-R-
Phosphatidylcholine (200 mg) was dissolved in a mixture of 1 mL of
CHCl3 and 0.5 mL of methanol in a glass test tube with a conic bottom.
The solution was evaporated under a flow of nitrogen from the tube,
which was kept under continuous rotation, leaving a phospholipid film
on the tube walls. A 2 mL volume of solution containing 10 mM K2-
HPO4 (pH 7.0), 50 mM KCl, and porphyrin dendrimer 1 was added to
the tube, and the resulting suspension was sonicated on an ice bath
using a microtip in pulsed mode (4 × 60, output 4:40%, sonicator
Vibracell, Sonic & Materials, Inc.). To remove the outside dye, the
freshly prepared liposomes were passed through a column (i.d. 1 × 7
cm) packed with Sephadex G200 (6 cm layer) and an anion-exchange
resin (QAE-Sephadex A50, Pharmacia) (1 cm layer) at the bottom of
the column to trap any remaining dye. After purification, the concentra-
tion of the dye in the outside solution was negligibly low, as evident
by fluorescence measurements of the supernatant after centrifugation
of the liposome solution.
The Fluorolog-2 system with R928 PMT was used to measure
fluorescence in the experiments involving liposomes. The measurements
were conducted using a standard 1 cm quartz cuvette, equipped with a
stirrer and a pH minielectrode. pH changes inside the vesicles after
addition of small aliquots of acid or base were monitored using
fluorescence right angle detection setup. Gramicidin (0.2 µg/mL, final
concentration) was added directly to the cuvette as the measurements
progressed, from a stock solution in DMSO. The amount of DMSO
added (5 µL) did not affect the proton permeability of the liposomes.
Synthesis of the Core Porphyrins. Core porphyrin 1b was
synthesized as described previously.9 The synthesis of core porphyrin
2b (Scheme 2) generally followed the recently published method.28 The
first three steps of the reaction sequence, leading to pyrrole ester i, are
described in detail in refs 28 and 38. In the reported below UV-vis
data for the porphyrins, the intensities (shown in parentheses) are given
relative to the strongest absorption in the spectrum, i.e., the Soret band,
whose intensity was set to 1.0. The positions of the absorption bands
of the porphyrins were found to be solvent dependent and in the case
of the porphyrin dications also dependent on the concentration and type
of the acid added.
Porphyrin iii. Pyrrole ii (0.59 g, 2.5 mmol) was dissolved in 250
mL of CH2Cl2 and 4-(methoxycarbonyl)benzaldehyde (0.44 g, 2.7
mmol) was added to the mixture. The mixture was kept under
continuous stirring in the dark at rt (room temperature) under Ar. After
10 min, BF3‚Et2O (0.071 g, 0.5 mmol) was added in one portion, and
the mixture was left to react for 2 h. DDQ (0.63 g, 2.8 mmol) was
added, and the mixture was left overnight. The resulting solution was
washed with 10% aqueous Na2SO3 (100 mL), then with 10% aqueous
Na2CO3 (100 mL), and finally with 5% aqueous HCl, after which it
was dried over Na2SO4. The solvent was removed in a vacuum, and
the remaining material was purified on a silica gel column (eluent:
CH2Cl2-THF and then CH2Cl2-THF-AcOH). A bright green band
was collected, and after removal of the solvent, the remaining solid
was purified by repetitive precipitations from CH2Cl2-AcOH (10:1)
1
mixture with hexane. Yield of porphyrin iii: 315 mg, 33%. H NMR
(free base, CDCl3): δ 8.52 (m, 8H), 8.19-8.37 (m, 8H), 4.1 (s, 12H),
2.4-4.0 (m, 48H), -2.4 (br s, 2H). MALDI, m/z: calcd, 1527.5; found,
1528.4. UV-vis, λmax nm: (free base; CH2Cl2-Et3N, 10:1) 445 (1.00),
539 (0.09), 611 (0.04), 675 (0.02); (dication; CH2Cl2-TFA, 10:1) 461
(1.00), 611 (0.05), 669 (0.09).
Copper Porphyrin iv. An excess of Cu(OAc)2‚2H2O (110 mg, 0.5
mmol) was added to a solution of porphyrin iii (180 mg, 0.118 mmol)
in CHCl3-MeOH (9:1), and the mixture was stirred at rt for 10-15
min. The conversion was monitored using UV-vis spectroscopy
(solvent CHCl3-AcOH). The reaction was stopped when the Soret
absorption of the porphyrin dication (460-470 nm) disappeared. The
mixture was washed with 10% aqueous AcOH and then with 10%
aqueous NaHCO3 and water and dried over Na2SO4. The solvent was
evaporated in a vacuum, and the remaining material was purified on a
silica gel column, giving the product as a red amorphous solid. Yield
of Cu-porphyrin iv: 182 mg, 97%. MALDI, m/z: calcd, 1589.1; found.
1587.6. UV-vis, λmax nm (CH2Cl2): 428 (1.00), 559 (0.12), 593 (0.09).
Copper Tetrabenzoporphyrin v. Porphyrin iv (180 mg, 0.10 mmol)
was dissolved in 50 mL of dry MeCN. DDQ (0.36 g, 1.6 mmol) was
added, and the mixture was refluxed for 30 min. The color of the
solution changed from reddish to bright green. The mixture was allowed
to cool, diluted with CH2Cl2, washed with 10% aqueous Na2SO3
solution and then with water and brine, and dried over Na2SO4. The
solvent was removed in a vacuum, and the remaining solid was purified
on a silica gel column (eluent CH2Cl2-THF, 10:1). The first dark-
green fraction was collected. The solvent was evaporated, and the
remaining material was recrystallized from CH2Cl2-ether to give the
product as a blue-green crystalline solid. Yield of Cu-tetrabenzopor-
phyrin v: 169 mg, 95%. MALDI, m/z: calcd, 1572.9; found, 1572.8.
UV-vis, λmax nm (CH2Cl2): 471 (1.00), 613 (0.11), 662 (0.35).
Tetrabenzoporphyrin Ester 2a. Compound v (150 mg, 0.095
mmol) was dissolved in 15-30 mL of concentrated H2SO4 and left
under stirring at rt for 24 h. The conversion was monitored using UV-
vis spectroscopy (solvent: HCl concentrated), as the Soret band of 2a
dication emerged at 510-520 nm, while the Soret band of Cu-
tetrabenzoporphyrin v disappear. The solution in H2SO4 was carefully
poured into ice-cold MeOH (100 mL, ice bath) and left in a closed
vessel overnight. CH2Cl2 (100 mL) and water (300 mL) were added to
Pyrrole ii. Pyrrole-ester i (1.3 g, 3.5 mmol) was dissolved in 35
mL of THF, Et3N (0.35 mL) was added, and the vessel was flushed
(37) Seybold, P. G.; Gouterman, M. J. Mol. Spectrosc. 1969, 31, 1.
(38) (a) Haake, G.; Struve, D.; Montforts, F. P. Tetrahedron Lett. 1994, 35,
9703. (b) Abel, Y.; Haake, E.; Haake, G.; Schmidt, W.; Struve, D.; Walter,
A.; Montforts, F. P. HelV. Chim. Acta 1998, 81, 1978. (c) Hopkins, P. B.;
Fush, P. L. J. Org. Chem. 1978, 43, 1208.
(39) Perlman, W. M. Tetrahedron Lett. 1967, 17, 1663.
9
4892 J. AM. CHEM. SOC. VOL. 125, NO. 16, 2003

Porphyrin and Tetrabenzoporphyrin Dendrimers
A R T I C L E S
the resulting mixture. The organic layer was separated, washed with
10% aqueous Na2CO3 solution and then with water, and dried over
Na2SO4. The solvent was removed in a vacuum, and the remaining
green solid was purified on a short silica gel column (i.d. 2 × 10 cm,
eluent CHCl3-THF, 20:1). The major bright-green fraction was
collected and evaporated to dryness, and the residual material was
recrystallized from CH2Cl2-ether mixture. Yield of 2a: 137 mg, 95%,
(8H, m), 3.58 (326HPEG350, m), 3.35 (∼508HPEG350, m), 2.60-2.14 (8H,
m). 13C NMR (CDCl3): δ 172.5, 171.2, 145.1, 141.3, 134.1, 133.4,
131.1, 125.8, 120.9, 71.9, 70.6 (bs), 69.0 (bs), 64.5, 63.8, 58.9, 53.0,
52.0, 30.5, 30.3, 26.95. MALDI-TOF, m/z: calcd (for average MW)
7650; found, 7600 (broad peak).
H2P-Glu4OH (1). H2P-Glu3OH (1.11 g, 0.25 mmol), glutamic acid
diethyl ester hydrochloride (3.83 g, 16 mmol), DCC (3.3 g, 16 mmol),
18-crown-6 polyether (0.3 g, 1.1 mmol), and pyridine (2 mL) were
left to react in 50 mL of DMF under Ar at 45 °C for 21 days. During
the reaction, the solution slowly turned purple, indicating gradual
extraction of porphyrin into the liquid phase. The solvent was
evaporated under vacuum on a rotary evaporator, and the remaining
brown material was redissolved in CH2Cl2. Didicyclohexylurea was
filtered off, and the product was purified by chromatography on silica
gel using 2:1 CH2Cl2-THF as a mobile phase, followed by GPC on
S-X1 beads. Yield of H2P-Glu4OEt: 0.95 g, 37%. MALDI-TOF, m/z:
calcd, 10 326; found, 10 349 [(M + Na+) + satellites] (see the main
text for details). H2P-Glu4OEt was hydrolyzed under basic conditions
as described previously.9 Porphyrin dendrimer 1 was purified by dialysis
and isolated as a fluffy green powder after lyophilization. 1H NMR: δ
9.12 (bs, 8H), 8.51-8.42 (m, 16H), 4.76-4.20 (m, 60H), 2.80-1.8
(m, 240H). 13C NMR (D2O): δ 181.95, 181.9, 179.6, 176.0, 146.6,
142.7, 136.4, 127.9, 122.3, 112.9, 56.4, 55.0, 54.2, 35.3, 34.8, 34.6,
33.2, 32.9, 29.6, 29.0, 28.9, 28.6, 25.8, 25.3. In the 13C NMR spectrum
of 1 the peaks in the glutamate region (60-20 ppm) consist of multiple
resonances and therefore appear as broadened.
H2TBP-Nw1OMe (2c). Tetrabenzoporphyrin 2b (89 mg, 0.066
mmol), 24 equiv of Newkome dendron (0.6 g, 1.58 mmol), and BtOH
(0.21 g, 1.58 mmol) were dissolved in 50 mL of DMF under Ar at
0 °C. DCC (0.32 g,1.58 mmol) and 0.1 mL of sym-collidine were added,
and the mixture was allowed to gradually warm to rt, after which it
was stirred continuously for 3 days. The solvent was removed under
vacuum, and the product was purified twice by GPC on S-X1 beads.
Yield: 303 mg, 81%. 1H NMR (CDCl3): δ 8.39-8.32 (AA′BB′Ar,
16H), 7.30 (bs, 8H), 7.12 (s, 4H), 6.49 (s, 8H), 4.01 (s, 24H), 3.83 (t,
24H), 3.75-3.70 (m, 96H), 3.69 (s, 36H), 3.50 (s, 72H), 2.67 (t, 24H),
2.57 (t, 48H). 13C NMR (CDCl3): δ 172.0, 171.9, 168.5, 166.2, 146.8,
142.8, 139.1, 136.4, 134.6, 128.2, 123.9, 121.9, 116.1, 69.6, 69.0, 66.9,
66.7, 60.5, 60.2, 51.6, 51.5, 51.4, 51.3, 34.8. MALDI-TOF, m/z: calcd,
5679; found, 5702 (M+ + Na), 5680 (M+), 5343 (see the main text for
details).
1
blue-green crystals. H NMR (CDCl3-TFA): δ 11.06 (TFA + NH),
8.68 (br, 16H), 7.77 (s, 8H), 4.24 (s, 12H), 3.83 (s, 24H). 13C NMR
(CDCl3-TFA): δ 166.8, 166.7, 141.3, 141.3, 136.0, 132.9, 132.8,
131.5, 131.2, 125.7, 116.1, 53.2, 53.1. MALDI, m/z: calcd, 1511.4;
found, 1511.5. (UV-vis: see Table 1.)
Tetrabenzoporphyrin Acid 2b. Tetrabenzoporphyrin 2a (100 mg,
0.066 mmol) was stirred overnight at rt with 0.5 g of LiOH in 20 mL
of THF. Water (about 10 mL) was added, and THF was removed on
a rotary evaporator under reduced pressure. The resulting deep green
solution was filtered, and porphyrin 2b was precipitated upon acidifica-
tion with HCl. The precipitate was collected by centrifugation and dried
1
in a vacuum. Yield of 2b: 86 mg, 97%. H NMR (DMSO-d6-TFA):
δ 8.77-8.55 (m, AA′BB′Ar, 16H), 7.52 (s, 8H). 13C NMR (DMSO-
d6-TFA): δ 167.3, 166.7, 142.5, 141.8, 135.7, 132.2, 131.9, 123.8,
116.4. (UV-vis: see Table 2.)
Synthesis of Porphyrin Dendrimers. All carboxyl-terminated
porphyrin dendrimers were obtained by hydrolysis of the corresponding
esters. The procedures for hydrolysis and isolation by either precipitation
or dialysis/liophilization were described in the earlier papers.9 Gen 4
polyglutamic porphyrin dendrimer ethyl ester (H2P-Glu4OEt) was
synthesized divergently (see below) from its precursor Gen 3 dendrimer
H2P-Glu3OH, which was in turn was synthesized from Gen 2 dendrimer
(H2PorphGlu2OH), according to the previously reported method.9c Gen
2 porphyrin dendrimer was synthesized by coupling Gen 2 triglutamic
dendrons to the core porphyrin 1b. Synthesis of Gen 2 dendron followed
the earlier published method.9d Esterification of H2P-Glu2OH with
monomethoxypoly(ethylene glycol) followed the reported protocol.7b,e;9d
Gen 1 Newkome dendron was prepared as described earlier.29 Solutions
of all light-sensitive compounds were shielded from the ambient light
during the syntheses. The UV-vis spectra of all intermediate porphyrin
dendrimers were identical to those of the core porphyrins and/or of
the compounds 1 and 2, whose spectra are tabulated (Table 2) and
described in the main text in great detail.
H2P-Glu2OEt. Porphyrin 1b (200 mg, 0.25 mmol) and 6 equiv of
Gen 2 triglutamic dendron (0.78 g, 1.5 mmol) were dissolved in 50
mL of DMF under Ar. 1,3-Dicyclohexylcarbodiimide (DCC) (0.32 g,
1.55 mmol) and pyridine (2 mL) were added, and the mixture was
stirred at rt for 3 days. The solvent was evaporated under vacuum at
rt, and the remaining brown-purple residue was redissolved in CH2-
Cl2. The white crystalline precipitate (dicyclohexylurea) was filtered
off, and the product was purified by chromatography on silica gel using
4:1 CH2Cl2-THF as a mobile phase, followed by preparative GPC on
H2TBP-Nw1OH (2). Excess of NaOH (0.5 g) was added to a solution
of 2c (200 mg, 0.035 mmol) in 10 mL of THF. The mixture was stirred
at rt for 4 h, and the solvent was evaporated under vacuum at rt. Water
was added to the remaining solid, and the mixture was stirred at rt for
additional 2 h. The solution was neutralized by HCl, filtered, dialyzed
against distilled water for 4 days, and lyophilized. 1H NMR (D2O): δ
9.14-8.44 (m, 16H), 7.44 (bs, 8H), 9.5-7.0 (bm, 12HNH), 3.75 (bm,
144H), 2.71 (bm, 72H). 13C NMR (D2O): 178.8, 178.4, 178.1, 177.9,
171.5, 171.3, 169.3, 143.6, 142.7, 137.9,137.1, 136.0, 129.8, 125.6,
125.5, 118.2, 115.8, 70.7, 70.3, 69.0, 68.8, 62.8, 62.4, 36.8, 36.6.
1
S-X1 beads. Yield: 0.52 g, 75%. H NMR (CDCl3): δ 8.74 (s, 8H),
8.42-7.20 (m AA′BB′Ar and NH, 28H), 4.88-4.74 (m, 12H), 4.40-
4.20 (m, 32H), 2.80-2.25 (m, 48H), 1.75-1.50 (m, 48H), -2.92 (s,
2H). 13C NMR (DMSO-d6): δ 173.6, 173.4, 172.0, 171.6, 166.4, 147.7,
143.9, 141.2, 135.2, 132.0, 127.1, 120.2, 66.1 (bs), 52.1(bs), 31.9, 30.5
(bs), 30.1, 26.4 (bs). MALDI-TOF, m/z: calcd, 2788; found, 2789.
H2P-Glu2OPEG350. H2P-Glu2OH (50 mg, 0.021 mmol), DCC (0.15
g, 0.7 mmol), hydroxybenzotriazole (BtOH) (0.1 g, 0.7 mmol), 3 mL
of poly(ethylene glycol) monomethyl ether (average MW 350), and a
drop of sym-collidine were stirred together at rt for 7 days. Ice-cold
water (20 mL) was added, and the mixture was acidified with a drop
of HCl and left in the dark at 4 °C overnight. White precipitate was
filtered off, and pH of the solution was adjusted to neutral. The solution
was filtered again on a nitrocellulose 0.25 µm filter, and the product
was isolated by GPS on Sephadex G50 using water as the mobile phase.
The solution was liophilized, and the remaining viscous purple oil was
redissolved in 3 mL of THF and repurified by GPC on S-X1 column.
Acknowledgment. The Grant NS-31465 from the NIH (U.S.),
the Contract No. 02-5403-21-2 with Anteon Corp. (U.S.), and
a grant from Carl Tryggers Foundation (Sweden) are gratefully
acknowledged. C.H. and S.V. acknowledge The Royal Physi-
ographic Society (Lund, Sweden) for the support of the inter-
national collaboration. A.G. acknowledges the postdoctoral
scholarship from the Wenner-Gren Foundation (Sweden). We
thank Prof. David F. Wilson for many invaluable discussions,
Prof. Kim Sharp and Dr. Ninad Prahbu for assistance with
GRASP software, and Ms. Natalie Kim for proofreading the
manuscript.
1
Yield: 145 mg, 92%. H NMR (CDCl3): δ 8.76 (bs, 8H), 8.21-8.19
(m, 16H), 7.65-7.50 (m, 12HNH), 4.63 (32H, m), 4.29 (4H, m), 4.19
JA0341687
9
J. AM. CHEM. SOC. VOL. 125, NO. 16, 2003 4893