A simple quick route to fullerene amino acid derivatives

Article abstract of DOI:10.1039/c003019h

Two new fullerene amino acids have been prepared by dipolar addition to C₆₀ of either the Boc- or Fmoc-Nα-protected azido amino acids derived from phenylalanine and lysine. UV-visible and CV studies indicate the as prepared amino acids are a mi

Full text of DOI:10.1039/c003019h

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
A simple quick route to fullerene amino acid derivativesw
T. Amanda Strom and Andrew R. Barron*
Received 12th February 2010, Accepted 14th May 2010
First published as an Advance Article on the web 2nd June 2010
DOI: 10.1039/c003019h
Two new fullerene amino acids have been prepared by dipolar
addition to C60 of either the Boc- or Fmoc-Na-protected azido
amino acids derived from phenylalanine and lysine. UV-visible
and CV studies indicate the as prepared amino acids are a mixture
of 5,6-open (major product) and 6,6-closed (minor product)
derivatives that may be readily separated.
as diazomethanes and organic azides with a yield around
30% for the mono-addition products.7,8 Furthermore, the
reaction conditions can be manipulated to achieve either a
6,6-closed aziridine structure by nitrene addition or a 5,6-open
aza-fulleroid structure by azide addition then N2 extrusion.8
Scheme 1 illustrates our new approach to a fullerene deriva-
tized amino acid which can be completed in 3 days, including
chromatographic purification, with B25% yield of a mixture
of the Boc-protected mono adducts, Boc-aza-C60-Phe (1a) and
Boc-aziridine-C60-Phe (1b).z y The Fmoc-protected derivatives
(2a and 2b) are prepared in an analogous manner (Scheme 1).y
This methodology may also be used to prepare (in 14% yield)
the Boc-protected lysine derivatives, Boc-aza-C60-Lys (3a) and
Boc-aziridine-C60-Lys (3b) (Chart 1).y The formation of the
minor aziridine products (14–17% of isolated product) is due
to the loss of molecular N2 before the addition, resulting in a
[2+1] nitrene addition. The isomers are readily separated with
flash chromatography; however, for many applications this is
not necessary further simplifying the synthetic protocol.
In addition to the mono aza- and aziridino-fullerene
adducts (1–3) multiple addition products are formed with
increased reaction time. However, these may be limited if the
reaction is monitored by reverse phase HPLC.z Changes to
the solvent system, protecting groups, and the temperature
resulted in the most significant yield increases. Initially, a
co-solvent system of toluene (for the C60) and THF (for the
amino acid) was chosen. However, both reactants were readily
soluble in o-dichlorobenzene (ODCB) and the reaction was
complete within 24 hours with less formation of the multiple
addition products. Despite slightly lowering the yield, the Boc
protecting group was chosen due to the ease of the deprotec-
tion. However, it must be noted that, the aza-fullerene linkage
is slightly acid labile so precautions must be taken to protect
the linkage. Piperidine deprotection of Fmoc protected full-
erene compounds has been shown to form adducts with the
fullerene cage making purification very difficult.9 While we did
not observe any formation of adducts by MALDI-ToF, the
Boc approach proved more facile and is consistent with our
current SPPS strategy.
As part of our program to incorporate fullerene into macro-
molecular structures of biological importance we have previously
reported the synthesis of a series of fullerene amino acids in
which the C60 moiety resides on the R side chain of each amino
acid and is connected by a non-hydrolyzable amino linkage.1,2
These characteristics allow the fullerene amino acids to be
readily incorporated into a peptide sequence through solid
phase peptide synthesis (SPPS).2 The hydrophobic nature of
fullerene makes these amino acids interesting pharmacophores
in biologically active molecules, especially where hydrophobic
interactions are important, e.g., with both receptors and
membranes.3 In regard to membranes, we have demonstrated
that attachment of our fullerene substituted phenylalanine
derivative (Baa) to a peptide enables transport into a wide
range of cells where the peptides in the absence of the fullerene
amino acid cannot enter the cell, and shows cancer cell
inhibition without significant toxicity to healthy cells.2,4 The
fullerene amino acid also facilitates penetration of peptides
through porcine skin.5 Recently we have shown that these
fullerene substituted phenylalanine derivatives have high
affinities with HIV-1 PR and carbonic anhydrase, and greater
inhibitory binding than the most active fullerene-based inhibitor
currently available.6 Unfortunately, the current synthesis is
time consuming (up to 2 weeks) and suffers from low yields
(e.g., the yield for the lysine derivative is 10%) even in
experienced hands. Since the fullerene amino acid derivatives
are required in three-fold excess for solid phase ligation to a
peptide sequence, the current synthesis is not ideal, and
progress in research has been limited by the amount of time
required to make the starting material. For these reasons, we
now report a simpler, faster, method for making a range of
fullerene amino acids in order to allow greater flexibility for
future research.
The goal of this synthesis is to create a general method for
the preparation of fullerene amino acids and to improve the
yield of a fullero-lysine derivative. In this regard, the optimal
reaction conditions were substrate dependent, either aryl or alkyl.
The temperature of the reactions were usually maintained
around 80 1C and the survival of the azido functional group
was confirmed by thin film IR (nass = 2100 cmꢀ1).10 Azido-
lysine (alkyl) decomposes around 90 1C, resulting in undesirable
reaction products, but is avoidable by protection of the
C-terminus. However, when the temperature was maintained
around 60 1C, no C-terminus protection was required and no
In our previous approach the fullerene forms a [2+4] Diels–
Alder product with a trimethylsiloxy diene; however, C60 will
also form a [3+2] cycloaddition product with 1,3-dipoles such
Richard E. Smalley Institute for Nanoscale Science and Technology,
Center for Biological and Environmental Nanotechnology,
and Department of Chemistry, Rice University, Houston, Texas
77005, USA. E-mail: arb@rice.edu; Web: www.rice.edu/barron
w Electronic supplementary information (ESI) available: Details of the
synthesis; 1H NMR measurements; and cyclic voltammograms. See
DOI: 10.1039/c003019h
ꢁc
This journal is The Royal Society of Chemistry 2010
4764 | Chem. Commun., 2010, 46, 4764–4766

View Article Online
Scheme 1 Aza-C60-Phe synthesis: (i) 1.2 eq. imidazole-sufonyl-N3, 2 eq. NEt3, ZnCl2 catalyst, MeCN/MeOH, rt, (ii) 1 eq. C60, ODCB, 80–90 1C,
(iii) 25% TFA in DCM, and (iv) Gly-OtBu, HBTU, NEt3, anhydrous DMF.
The absorption spectra of each fullerene amino acid are
consistent with the absorption of C60 in chloroform with maxima
at approximately 260 and 330 nm. Aziridino fullerenes (1b, 2b,
3b) exhibit an additional absorption around 420 nm.12 Fig. 1
shows the UV-visible spectra of the lysine derivatives in com-
parison with pristine C60 confirming the attachment of the amino
acids to the fullerene core. In addition, 3a exhibits no absorptions
with l E 420 nm, while in 3b a small peak at 423 nm is
discernible in the inset, confirming the 6,6-closed aziridino-
fullerene structure determination. The lmax at 327 nm (e 40000)
was used to determine the concentration of fullerene amino
acids in solution. The colors of the aza derivatives are blue/
purple, while the aziridino compounds are either amber
(1b and 2b) or pink (3b).
Chart 1 Boc-aza-C60-Lys 3a and Boc-aziridino-C60-Lys 3b.
Further conformation of the aza versus aziridino structures
was obtained from the electrochemical behavior as determined
undesirable reactions occurred. For Phe derivatives, increasing the
by cyclic voltammetry (CV) in a mixture of DMF/toluene
reaction time beyond 24 h did not improve the yield of the mono-
(3 : 2) at room temperature with [Bu4N][BF4] as the electrolyte
adduct and consumed a great deal more C60 in the form of
(see ESIw, Fig. S1). C60 is known to be less easily reduced with
multiple addition products. In contrast, for 3a the reaction
proceeded much more slowly due to the temperature difference.
These reactions typically went more than 72 hours before forma-
tion of the tris products. As we discovered, the substituent adjacent
to the aza/aziridino N, either aryl or alkyl, has a dramatic effect on
the structure’s stability. The aza-Phe derivatives decomposed after
several days in light and air, whereas the aza-Lys derivative, 3a,
oxidized to a product 32 mass units higher than the Mꢀ ion but
did not decompose. The acute oxygen sensitivity and subsequent
ketolactam formation (+32 mass units) of alkyl aza-fulleroids is
well documented in the literature.11 No such oxidation was
observed for any of our aziridino (6,6-closed) products. All efforts
to convert 3a to the thermodynamic product 3b thermally, as a dry
powder and in refluxing toluene or ODCB, failed.
Fig. 1 UV absorption spectra of 3a, 3b, and C60 in CHCl3.
Chem. Commun., 2010, 46, 4764–4766 | 4765
ꢁc
This journal is The Royal Society of Chemistry 2010

View Article Online
Table 1 Selected cyclic voltammetry data for fullerene amino acidsa
presently investigating the cellular uptake, transport through
porcine skin, and inhibitory binding with HIV-1 PR and
carbonic anhydrase of this new class of fullerene amino acids.
Financial support for this work is provided by the Robert
A. Welch Foundation. We thank Dr Jianhua Yang for the
cellular toxicity studies.
1
2
Compound
Formal potential, E /V
Structure
C60
ꢀ0.247, ꢀ0.741, ꢀ1.327
ꢀ0.280, ꢀ0.892b
Boc-aza-C60-Lys (3a)
Boc-aziridino-C60-Lys (3b)
Boc-aza-C60-Phe (1a)
Boc-Baa
5,6-Open
6,6-Closed
5,6-Open
6,6-Closed
ꢀ0.319, ꢀ0.792b
ꢀ0.270, ꢀ0.744, ꢀ1.318
ꢀ0.377, ꢀ0.847, ꢀ1.489
a
Potentials are given in volts versus a reference electrode, Ag|AgCl.
Third reduction not observed.
b
Notes and references
z An amino derivatized, Na-protected amino acid (1 equiv.), imidazole-
sulfonyl-azideꢂHCl,16 triethylamine (2 equiv.) and 4 mol% ZnCl2 were
dissolved in a solution of MeOH : MeCN (3 : 2) and stirred in the
dark under N2 for 12 h.16 The mixture was concentrated then diluted
with ethyl acetate and washed with AcOH (10%, 3ꢃ) and brine (1ꢃ),
dried over MgSO4, and concentrated under vacuum. Flash chromato-
graphy is necessary at the gram scale. C60 (1 equiv.) was vacuum dried
for 45 min and o-dichlorobenzene (ODCB) was added. To this was
added the Na-protected, azido-amino acid (1 equiv.) with stirring.
The temperature was maintained between 80–100 1C for the Phe
derivatives and 60–70 1C for the Lys derivatives. The reaction was
monitored by RP-HPLC (5–95% IPA in 0.1% TFA over 30 min, then
95% IPA for 20 min). Typically, within 24 h the mono addition
product was isolated by column chromatography.
each successive addition to its core. For C60 derivatives, the
first electron transfer is shifted to more negative potentials
while maintaining the characteristic redox properties of
pristine C60.13 In this respect, Baa (6,6-closed) is consistent
with other similar closed ring structures, with its first reduction
occurring at ꢀ0.377 V. In contrast, in the open ring structure
of the aza-C60 derivatives (1a and 3a), only sp2 hybridized
carbons are present on the cage allowing these compounds to
behave more like pristine C60 in their reductive ability under
the specified conditions.14 Table 1 shows the detailed redox
potential data for the aza and aziridino-fullerene derivatives
relative to C60 and the previously reported Baa. For C60 and
the aza-fulleroid derivatives the first reduction occurs between
ꢀ0.248 and ꢀ0.288 V. However, the closed ring structures
with two sp3 C60 carbons do not undergo their first reduction
until more negative potentials, cathodically shifted approxi-
mately 0.04–0.15 V relative to the aza derivatives.
y MALDI/TOF, DCTB matrix, (Mꢀ): Fmoc-aza-C60-Phe (1a) m/z
1120.23 (calcd 1120.14); Boc-aza-C60-Phe (2a) m/z 998.16 (calcd
998.13); Boc-aza-C60-Lys (3a) m/z 964.06 (calcd 964.14); aza-C60-Phe
(4) m/z 898.02 (calcd 898.07); aza-C60-Phe-Gly-OtBu (matrix CHCA)
(5) m/z (M + 1) 1234.11 (calcd 1234.23).
1 J. Yang and A. R. Barron, Chem. Commun., 2004, 2884.
2 J. Yang, L. B. Alemany, J. Driver, J. D. Hartgerink and
A. R. Barron, Chem.–Eur. J., 2007, 13, 2530.
3 T. Da Ros and M. Prato, Chem. Commun., 1999, 663;
N. Tagmatarchis and H. Shinohara, Mini-Rev. Med. Chem.,
2001, 1, 339.
4 J. G. Rouse, J. Yang, A. R. Barron and N. A. Monteiro-Riviere,
Toxicol. in Vitro, 2006, 20, 1313.
Given that the aza-C60-Phe and aza-C60-Lys are easier to
prepare on a large scale, and therefore of much greater utility
for subsequent studies as compared to Baa, it is important to
ensure that the change in linkage unit does not alter the cellular
toxicity significantly. The toxic response of the deprotected
aza-C60-Phe (4) was determined by MTT assay on three
different neuroblastoma cell lines (JF, SK-N-AS, and LAN-1)
at concentrations from 0.004–0.4 mg mLꢀ1. No appreciable
toxicity was observed for aza-C60-Phe at any of the tested
concentrations after a 24 hour incubation period.
5 J. G. Rouse, J. Yang, J. P. Ryman-Rasmussen, A. R. Barron and
N. A. Monteiro-Riviere, Nano Lett., 2007, 7, 155.
6 S. Durdagi, C. T. Supuran, T. A. Strom, N. Doostdar,
M. K. Kumar, A. R. Barron, T. Mavromoustakos and
M. G. Papadopoulos, J. Chem. Inf. Model., 2009, 49, 1139;
A. Innocenti, S. Durdagi, N. Doostdar, T. A. Strom,
A. R. Barron and C. T. Supuran, Bioorg. Med. Chem., 2010, 18,
2822.
To demonstrate the feasibility of using aza-C60-Phe for SPPS
the coupling reaction of compound 2a to a tert-butyl ester of
glycine (NH2-Gly-OtBu) was carried out in 4 : 1 DCM/DMF
in the presence of HBTU and NEt3 in a sonication bath for 2 h
(5, Scheme 1). The reaction products were characterized by
MALDI MS.y After 2 hours, only starting materials and
products were detected. Given the rapid oxidation of 3a in
air, only Lys derivative 3b will be useful for SPPS.
7 M. Prato, C. Li and F. Wudl, J. Am. Chem. Soc., 1993, 115, 1148.
8 A. Hirsch and M. Brettreich, Fullerenes: Chemistry and Reactions,
Wiley-VCH Verlag GmbH & Co., Weinheim, Germany, 2005,
p. 134.
9 D. Pantarotto, A. Bianco, F. Pellarini, A. Tossi, A. Giangaspero,
I. Zelezetsky, J. Briand and M. Prato, J. Am. Chem. Soc., 2002,
124, 12543.
10 E. F. Scriven, Azides and Nitrenes: Reactivity and Utility,
Academic Press, Inc, Orlando, FL, 1984, p. 2.
11 J. C. Hummelen, M. Prato and F. Wudl, J. Am. Chem. Soc., 1995,
117, 7003; S. Xiao, Y. Li, H. Fang, H. Li, H. Liu, Z. Shi, L. Jiang
and D. Zhu, Org. Lett., 2002, 4, 3063; D. M. Guldi,
In summary, we report a general route to fullerene deriva-
tized amino acids through the dipolar addition of azido amino
acids, both alkyl and aryl. The ease of this procedure, including
purification, is evident over previously reported methods.1,2
The newly reported C60 amino acids differ in their linkage
to C60, resulting in mostly a 5,6-open, aza-fulleroid structure
(1a, 2a, 3a) rather than a 6,6-closed cyclohexyl ring on C60. We
have demonstrated the coupling ability of 2a indicating its
ability for incorporation into a peptide sequence, and the
unprotected phenylalanine derivative has shown inhibitory
binding with HIV-1 PR at nM concentrations.15 We are
H. Hungerbuhler, I. Carmichael, K.-D. Asmus and M. Maggini,
¨
J. Phys. Chem. A, 2000, 104, 8601.
12 C. S. Foote, Top. Curr. Chem., 1994, 169, 347.
13 C. Boudon, J. Gisselbrecht, M. Gross, A. Herrmann,
M. Ruttimann, J. Crassous, F. Cardullo, L. Echegoyen and
¨
F. Diederich, J. Am. Chem. Soc., 1998, 120, 7860.
14 C. Boudon, J. Cisselbrecht and M. Cross, Helv. Chim. Acta, 1995,
78, 1334.
15 S. Durdagi, C. T. Supuran, N. Doostdar, T. A. Strom,
A. R. Barron, T. Mavromoustakos and M. G. Papadopoulos,
2010, submitted.
16 E. D. Goddard-Borger and R. V. Stick, Org. Lett., 2007, 9, 3797.
ꢁc
This journal is The Royal Society of Chemistry 2010
4766 | Chem. Commun., 2010, 46, 4764–4766