Polyamine functionalized carbon nanotubes: Synthesis, characterization, cytotoxicity and siRNA binding

Article abstract of DOI:10.1039/c0jm04064a

In this work we have synthesized a new series of cationic carbon nanotubes (CNTs) for siRNA binding. Both single- and multi-walled CNTs have been modified by addition or amidation reaction with short linear polyamine chains including putrescine, spermidin

Full text of DOI:10.1039/c0jm04064a

View Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links <
C
Journal of
Materials Chemistry
Cite this: J. Mater. Chem., 2011, 21, 4850
www.rsc.org/materials
PAPER
Polyamine functionalized carbon nanotubes: synthesis, characterization,
cytotoxicity and siRNA binding†
a
a
b
c
Prabhpreet Singh, Cristian Samorı, Francesca Maria Toma, Cyrill Bussy, Antonio Nunes,
c
`
Khuloud T. Al-Jamal, Cecilia Menard-Moyon, Maurizio Prato, Kostas Kostarelos and Alberto Bianco
c
a
b
c
a
*
*
*
ꢀ
ꢀ
Received 23rd November 2010, Accepted 18th January 2011
DOI: 10.1039/c0jm04064a
In this work we have synthesized a new series of cationic carbon nanotubes (CNTs) for siRNA binding.
Both single- and multi-walled CNTs have been modified by addition or amidation reaction with short
linear polyamine chains including putrescine, spermidine and spermine. All the new conjugates have
been characterized with several spectroscopic and microscopic techniques. Their cytotoxic effects have
been assessed on human lung carcinoma cell line. Finally, we have analyzed their capacity to bind
siRNA towards the development of new carriers for gene silencing applications. The dispersibility
properties and the capacity to complex siRNA of the different conjugates were found to be strongly
dependent on the position of the functional groups on CNTs (i.e. mainly at the tips following the
amidation reaction or on the sidewalls by direct C–C addition).
sonication in a mixture of concentrated nitric and sulfuric acid,
or heating in a mixture of sulfuric acid and hydrogen peroxide,
which results in the formation of short open tubes with
oxygenated functions (mainly carboxylic groups). This approach
represents a popular pathway for further modification of the
nanotubes, since the carboxylic functions can react with alcohols
or amines to give ester or amide derivatives. In the second
approach, more elaborate methods have been developed to
attach organic moieties directly onto the nanotube sidewalls
(e.g. cycloadditions, electrophilic, nucleophilic or radical
additions).6
Introduction
Single-walled carbon nanotubes (SWCNTs) and multi-walled
carbon nanotubes (MWCNTs) are an important class of nano-
materials,1,2 and recent studies in the field of nanotechnology and
nanomedicine have reported that CNTs could find a wide range
of biological applications such as drug and gene delivery, ther-
apeutics and diagnostics.
However, to fully exploit the properties of CNTs in nano-
medicine, we need to overcome their extremely poor aqueous
dispersion and their tendency to aggregate in order to obtain
homogeneous dispersions in aqueous media. In comparison to
other nanomaterials, functionalized CNTs (f-CNTs) have shown
higher biocompatibility and the capacity to cross cell membranes
easily.3–5 The chemical modification of the CNT surface through
covalent or non-covalent functionalization increases their solu-
bility in both organic and aqueous solvents.6–8 Two main
approaches have been developed for the covalent functionaliza-
tion of carbon nanotubes: (i) the amidation and the esterification
of oxidized CNTs; and (ii) the addition chemistry to the surface
of CNTs.6,7 The first approach includes treatment of the pristine
material under strong acidic and oxidative conditions, such as
In our search for new cationic carbon nanotubes towards the
development of advanced systems for the delivery of nucleic
acids, we have recently designed simple ammonium-functional-
ized CNTs and more sophisticated dendron-based nanotubes.9,10
The different conjugates were used for gene silencing experiments
in vitro and in vivo demonstrating the capacity of such systems to
deliver and eventually knock-down different genes efficiently.10
In this paper we have extended the synthesis of cationic nano-
tubes for gene silencing through reaction with a series of linear
polyamines. Polyamines, such as putrescine, spermidine and
spermine, are polycationic straight-chain aliphatic compounds,
ubiquitously present in all prokaryotic and eukaryotic cells,11
and are required in important biological processes such as cell
proliferation and differentiation.12–16 Spermidine and spermine
primarily exist in aqueous solution at pH 7.4 as fully protonated
polycations.17 Therefore, they can bind to the negatively charged
nucleic acid either by electrostatic interactions and/or by
hydrogen bonding. They have been shown to condense DNA to
stabilize its conformation or to prevent its degradation,18 and to
induce both B-to-Z and B-to-A transitions in certain DNA
sequences.19 They also modulate the structure of transfer RNA
aCNRS, Institut de Biologie Moleculaire et Cellulaire, Laboratoire
d’Immunologie et Chimie Theꢀrapeutiques, 67000 Strasbourg, France.
E-mail: a.bianco@ibmc-cnrs.unistra.fr
b
ꢁ
Dipartimento di Scienze Farmaceutiche, Universita di Trieste, 34127
ꢀ
Trieste, Italy. E-mail: prato@units.it
cNanomedicine Laboratory, Centre for Drug Delivery Research, The
School of Pharmacy, University of London, London, WC1N 1AX, UK.
E-mail: kostas.kostarelos@pharmacy.ac.uk
† Electronic supplementary information (ESI) available. See DOI:
10.1039/c0jm04064a
4850 | J. Mater. Chem., 2011, 21, 4850–4860
This journal is ª The Royal Society of Chemistry 2011

View Online
and other RNAs.20 These types of polycationic molecules have
been also proposed for plasmid DNA delivery and siRNA
silencing.21
addition reactions. The diazotization-coupling procedure between
CNTs and N-(4-aminobenzoyl)polyamine in o-dichlorobenzene/
CH3CN results in effective functionalization of the exohedral
surface of CNTs via in situ generation of diazonium salts.23
For this purpose, we have synthesized 4-aminobenzoyl deriv-
atives of putrescine, spermidine and spermine and used them for
conjugation to CNTs. We carried out addition and amidation
reactions for the covalent functionalization of both SWCNTs
and MWCNTs. The cationic conjugates were fully characterized
by thermogravimetric analysis (TGA), transmission electron
microscopy (TEM) and Raman spectroscopy. We also per-
formed preliminary studies to analyze the impact of these poly-
amine-functionalized CNTs on mammalian cell cultures in terms
of cytotoxic effects. Aiming to use them in order to condense and
deliver siRNA, we have also studied the siRNA binding capacity
of some of the new conjugates using gel electrophoresis.
Preparation of modified putrescine functionalized SWCNTs (1b)
General procedure: pristine HiPCO SWCNTs (30 mg, 2.5 mmol)
were homogeneously dispersed in 30 mL of o-dichlorobenzene
under Ar atmosphere through bath sonication for 2 h. To this
suspension was added 15 (0.77 g, 2.5 mmol) previously dissolved
in 10 mL of CH3CN. The suspension was further sonicated for
10 min. After bubbling with Ar for 5 min, isoamyl nitrite
(1.34 mL, 10.0 mmol) was quickly added and suspension was
further stirred for 36 h at 60 ꢀC under Ar atmosphere. The
resulting nanotubes were re-precipitated several times from
DMSO/diethyl ether, DMF/diethyl ether and methanol/diethyl
ether by successive bath sonication and centrifugation (4500
rpm, 5 min, Eppendorf Centrifuge 5804R) procedures until the
filtrate became colorless. Sonication and redispersion was
repeated with a mixture of DMF/diethyl ether and then in diethyl
ether. The product was dried under vacuum at room temperature
for 2 h to give polyamine functionalized SWCNTs 1a as a black
powder. Twenty mg of Boc-protected polyamine–CNTs 1a were
suspended in 8 mL of 4 M HCl solution in dioxane and stirred at
room temperature for 12 h under Ar atmosphere to cleave the
Boc group. Then, solvent was evaporated under vacuum. The
resulting nanotubes were washed with a mixture of CH2Cl2/
CH3OH several times and finally with diethyl ether and dried
under vacuum for 1 h to give cationic functionalized CNTs 1b as
a black powder. The nanotubes were characterized by TEM,
Raman spectroscopy andꢀTGA. Raman D/G ratio: 0.171; TGA
mass loss: 20.8% (at 600 C); TGA analysis afforded one func-
tional group every 73 carbon atoms, amount of functional
Experimental section
Materials
HiPCO SWCNTs were purchased from Carbon Nanotechno-
logies Inc. (Lot no. R0496). MWCNTs, produced by the catalytic
carbon vapour deposition (CCVD) process, were purchased as
purified from Nanocyl (Thin MWCNT 95+% C purity, Nanocyl
3100ꢀ, Batch no. 071005, average diameter and length: 9.5 nm
and 1.5 microns, respectively). Reagents and solvents were
purchased from Fluka, Acros, Aldrich and used without further
purification unless otherwise stated. Moisture-sensitive reactions
were performed under Ar or N2 atmosphere. CH2Cl2 was freshly
distilled from CaH2, THF from Na/benzophenone, and DMF
ꢂ
dried over 4 A molecular sieves. Chromatographic purification
was done with silica gel Merck (Kieselgel 60, 40–60 mm, 230–400
mesh ASTM) in standard columns. TLC was performed on
aluminium sheets coated with silica gel 60 F254 (Merck, Darm-
stadt). 1H and 13C NMR spectra were recorded on Bruker DPX
300 (operating at 300 and 75 MHz, respectively). The peak values
were obtained as ppm (d), and referenced to the solvent. The
protected polyamines 13, 19 and 25 were synthesized through
stepwise elongation of the backbone and fully characterized
(see ESI†). Protected polyamines 13, 19 and 25 were further
modified using already reported methods to give N-(4-amino-
benzoyl)polyamine derivatives 15, 21 and 27,22 respectively as
described in ESI†. Human lung carcinoma (A549) cell line was
purchased from American Type Culture Collection (ATCC)
(USA) (ATCC# CCL-185). Non-coding negative siRNA
(siNEG) was purchased from Eurogentec (UK). Dulbecco’s
modified eagle medium (DMEM), fetal bovine serum (FBS),
penicillin/streptomycin, glutamine, and phosphate buffered
saline (PBS) and electrophoresis grade Agarose were from
Gibco, Invitrogen (UK). Tris/borate/EDTA (TBE) buffer and
ethidium bromide were obtained from Sigma-Aldrich (USA).
The orange loading dye solution was purchased from Fermentas
(UK). LDH release was assessed using Promega Cytotox 96ꢀ
Non-radioactive cytotoxicity assay (Promega UK Ltd).
groups: 0.91 mmol gÀ1 24
.
Preparation of modified spermidine functionalized SWCNTs
(3b). The addition reaction was performed as reported for
putrescine above. The product was dried in vacuum at room
temperature for 1 h to give cationic polyamine–CNTs 3b as
a black powder. Raman D/G ratio: 0.133; TGA mass loss: 15.4%
ꢀ
(at 600 C); TGA analysis afforded one functional group every
147 carbon atoms, amount of functional groups: 0.48 mmol gÀ1
.
Preparation of modified spermine functionalized SWCNTs (5b).
The addition reaction was performed as reported for putrescine
above. The product was dried in vacuum at room temperature
for 1 h to give cationic polyamine–CNTs 5b as a black powder.
Raman D/G ratio: 0.185; TGA mass loss: 14.1% (at 600 ꢀC);
TGA analysis afforded one functional group every 211 carbon
atoms, amount of functional groups: 0.34 mmol gÀ1
.
Preparation of modified putrescine functionalized MWCNTs
(2b). The addition reaction was performed on purified MWCNTs
(30 mg) as reported for the preparation of 1b. The product was
dried in vacuum at room temperature for 1 h to give cationic
polyamine–CNTs 2b as a black powder. TGA mass loss: 7.80%
Preparation of modified polyamine functionalized CNTs through
addition reaction
ꢀ
SWCNTs and MWCNTs were covalently functionalized with
modified polyamines (putrescine, spermidine or spermine) using
(at 600 C); TGA analysis afforded one functional group every
225 carbon atoms, amount of functional groups: 0.34 mmol gÀ1
.
This journal is ª The Royal Society of Chemistry 2011
J. Mater. Chem., 2011, 21, 4850–4860 | 4851

View Online
Preparation of modified spermidine functionalized MWCNTs
(4b). The addition reaction was performed as reported for
putrescine above. The product was dried in vacuum at room
temperature for 1 h to give cationic polyamine–CNTs 4b as
a black powder. TGA mass loss: 8.50% (at 600 ꢀC); TGA analysis
afforded one functional group every 288 carbon atoms, amount
resulting nanotubes were washed with a mixture of dichloro-
methane and methanol several times and finally with diethyl
ether and dried under vacuum for 1 h to give cationic function-
alized CNTs 7b as a black powder. The nanotubes were char-
acterized by TEM, Raman spectroscopy and TGA. Raman D/G
ꢀ
ratio: 0.117; TGA mass loss: 21.5% (at 600 C); TGA analysis
of functional groups: 0.26 mmol gÀ1
.
afforded one functional group every 83 carbon atoms, amount of
functional groups: 0.79 mmol gÀ1
.
Preparation of modified spermine functionalized MWCNTs
(6b). The addition reaction was performed as reported for
putrescine above. The product was dried in vacuum at room
temperature for 1 h to give cationic polyamine–CNTs 6b as
a black powder. TGA mass loss: 9.00% (at 600 ꢀC); TGA analysis
afforded one functional group every 349 carbon atoms, amount
Preparation of modified spermidine functionalized SWCNTs
(9b). The amidation reaction was performed as reported for
putrescine above. The product was dried in vacuum at room
temperature for 1 h to give cationic polyamine–CNTs 9b as
a black powder. Raman D/G ratio: 0.113; TGA mass loss: 18.7%
of functional groups: 0.22 mmol gÀ1
.
(at 600 C); TGA analysis afforded one functional group every
132 carbon atoms, amount of functional groups: 0.51 mmol gÀ1
ꢀ
.
Preparation of modified polyamine functionalized CNTs through
amidation reaction
Preparation of modified spermine functionalized SWCNTs
(11b). The amidation reaction was performed as reported for
putrescine above. The product was dried in vacuum at room
temperature for 1 h to give cationic polyamine–CNTs 11b as
a black ꢀpowder. Raman D/G ratio: 0.219; TGA mass loss: 17.5%
(at 600 C); TGA analysis estimates one functional group every
SWCNTs and MWCNTs were also covalently functionalized
with modified polyamine (putrescine, spermidine and spermine)
mainly at their tips through amide bond formation. The
carboxylic acid functional groups were introduced at the tips and
sidewall of CNTs using modified oxidative methods.25 The
carboxylic groups were subsequently converted into acid chlo-
ride, followed by addition of modified polyamine which results in
covalent attachment of polyamines at the tips of CNTs.
180 carbon atoms, amount of functional groups: 0.38 mmol gÀ1
.
Preparation of modified putrescine functionalized MWCNTs
(8b)
One hundred mg of purified MWCNTs were sonicated in a water
bath (20 W, 40 kHz) for 24 h in 15 mL of sulfuric acid/nitric acid
(3 : 1 v/v, 98% and 65%, respectively) at room temperature.25,27
Deionized water was then carefully added and the MWCNTs
were filtered (Omniporeꢀ membrane filtration, 0.45 mm), re-
suspended in water, filtered again until pH became neutral and
dried. A suspension of 10 mg of ox-MWCNTs in 4 mL of neat
oxalyl chloride was stirred at 62 ꢀC for 24 h under Ar atmo-
sphere. After evaporation in vacuum the resulting nanotubes
were dispersed in a solution of 15 (74 mg, 0.24 mmol) in 15 mL of
distilled THF or DMF followed by addition of DIEA (42 mL).
The resulting suspension was heated under reflux for 48 h. After
cooling to room temperature and removing excess of 15 by
washing several times with DMSO/diethyl ether, DMF/diethyl
ether and methanol/diethyl ether by successive bath sonication
and centrifugation (4500 rpm, 5 min) procedures and finally with
diethyl ether. The resulting Boc-protected MWCNTs 8a were
dried at room temperature under vacuum. The Boc protecting
group was removed using 4 M HCl solution in dioxane. Then,
solvent was evaporated under vacuum. The resulting nanotubes
were washed with a mixture of dichloromethane and methanol
several times and finally with diethyl ether and dried under
vacuum for 1 h to give cationic polyamiꢀne–CNTs 8b as a black
powder. TGA mass loss: 25.5% (at 600 C); TGA analysis esti-
mates one functional group every 66 carbon atoms, amount of
Preparation of modified putrescine functionalized SWCNTs (7b)
General procedure: 100 mg of pristine HiPCO SWCNTs were
suspended in 75 mL of a 3 M HNO3 by bath sonication. The
mixture was refluxed for about 48 h, sonicated for 1 h, and
refluxed again for another 48 h. Then, 25 mL of 3 M HNO3 was
added and after bath sonication for 2 h, the mixture was again
refluxed for 12 h. The obtained suspension was then diluted by
deionized water and filtered through a polycarbonate filter
(Isopore, pore size 100 nm), rinsed thoroughly with deionized
water several times until the pH value was ꢁ7. The resulting
SWCNTs were resuspended in deionized water and sonicated for
5 min. The suspension was then again filtered. The black product
obtained was dried and characterized by TEM, AFM and
TGA.26 A suspension of 10 mg of oxidized SWCNTs in 4 mL of
neat oxalyl chloride was stirred at 62 ꢀC for 24 h under Ar
atmosphere. The excess of oxalyl chloride was evaporated under
vacuum, obtaining SWCNT-C(O)Cl. The resulting nanotubes
were suspended in a solution of 15 (74 mg, 0.24 mmol) in 15 mL
of distilled THF or DMF followed by addition of diisopropyle-
thylamine (DIEA) (42 mL). The resulting suspension was heated
under reflux for 48 h. After cooling to room temperature and
removing excess of 15 by washing several times with DMSO/
diethyl ether, DMF/diethyl ether and methanol/diethyl ether by
successive bath sonication and centrifugation (4500 rpm, 5 min)
procedures and finally with diethyl ether. The resulting Boc-
protected SWCNTs 7a were dried at room temperature under
vacuum. Eight mg of Boc-protected polyamine–CNTs 7a were
suspended in 4 mL of 4 M HCl solution in dioxane and stirred at
room temperature for 12 h under Ar atmosphere to cleave the
Boc group. Then solvent was evaporated under vacuum. The
functional groups: 0.94 mmol gÀ1
.
Preparation of modified spermidine functionalized MWCNTs
(10b). The amidation reaction was performed on MWCNTs as
reported for putrescine above. The product was dried in vacuum
at room temperature for 1 h to give cationic polyamine–CNTs
4852 | J. Mater. Chem., 2011, 21, 4850–4860
This journal is ª The Royal Society of Chemistry 2011

View Online
10b as a black powder. TGA mass loss: 23.2% (at 600 ꢀC); TGA
expressed as cell viability. Hence, the percentage cell survival is
expressed as [LDH released from tested cells À blank (media
alone)/LDH released from control untreated cells] Â 100.
analysis estimates one functional group every 100 carbon atoms,
amount of functional groups: 0.63 mmol gÀ1
.
Preparation of modified spermine functionalized MWCNTs
(12b). The amidation reaction was performed on MWCNTs as
reported for putrescine above. The product was dried in vacuum
at room temperature for 1 h to give cationic polyamine–CNTs
12b as a black powder. TGA mass loss: 19.1% (at 600 ꢀC); TGA
analysis estimates one functional group every 161 carbon atoms,
Electrophoretic motility shift assay
siRNA–nanotube complexes were prepared by mixing 0.5 mg of
siRNA in 30 mL of 5% dextrose with polyamine–CNTs at
different charge ratios (1 : 0 to 1 : 18 À/+). 0.5 mg of free siRNA
was used as a control. The pH was monitored during the binding
to siRNA and no variation was observed before and after the
complexation or among the various CNT constructs. The pH of
the dispersions and the complexes was around 6.0–7.5. At this
pH all amino groups of the different constructs are more likely to
be protonated. Complexes were incubated for 30 min at room
temperature to allow complete complexation to occur. siRNA
complexes or siRNA alone were mixed with orange dye solution
(1 : 10 dilution) and loaded onto 1% (w/v) agarose/TBE gel
containing ethidium bromide (0.5 mg mLÀ1). The gel was run for
45 min at 70 mV (BioRad, UK). The gel was then photographed
under UV light using GeneGenius system (PerkinElmer Life and
Analytical Sciences, USA).
amount of functional groups: 0.42 mmol gÀ1
.
Characterization
The thermogravimetric analyses were performed using a TGA
Q500 TA instrument with a ramp of 10 ꢀC minÀ1 under N2 from
100 ꢀC to 800 ꢀC. Transmission electron microscopy (TEM) was
performed on a Hitachi H600 microscope and a Philips 208
working at different accelerating voltage and at different
magnification. 0.1 microgram of the sample was dispersed in
0.1 mL of methanol/water (1 : 1) by ultrasonication for 5 min and
kept for 10–12 h. The solution was again ultrasonicated for 5 min
before depositing and 10 mL was deposited onto a carbon coated
TEM grid and dried. The images are typical and representative of
the samples under observation. Raman spectra were acquired
with a Renishaw instrument, model Invia reflex equipped with
532 nm, 632.8 nm and 785 nm lasers. The spectra were registered
with the laser at 532 nm. After acquisition the spectra were
normalized with respect to the G band and then the amplitude of
the D peak was calculated.
Results and discussion
Design, synthesis and characterization
To explore the ability of polyamine-modified carbon nanotubes
to complex and eventually deliver siRNA for gene silencing, we
have designed and prepared two families of CNTs functionalized
by amidation reaction of the carboxylic groups and by addition
of in situ generated diazonium derivatives to CNT aromatic
surface. For this purpose, we have used both single- and multi-
walled carbon nanotubes and combined them with polyamines
like putrescine (1,4-diaminobutane), spermidine (1,8-diamino-4-
azaoctane) and spermine (1,12-diamino-4,9-diazadodecane).
Cell viability assay
A549 cells were maintained and passaged in DMEM media,
supplemented with 10% fetal bovine serum (FBS), 100 U mLÀ1
penicillin, 100 mg mLÀ1 streptomycin, 1% L-glutamine and 1%
ꢀ
non-essential amino acids, at 37 C in 5% CO2. A549 cells were
Oxidized
SWCNTs
(ox-SWCNTs)
and
MWCNTs
incubated with polyamine–CNTs in complete media for 24 or
72 h. Toxicity was assessed by the modified LDH assay described
previously using the Promega Cytotox 96ꢀ Non-radioactive
cytotoxicity assay (Promega UK Ltd) according to the manu-
facturer instructions.28 The assay was modified to avoid inter-
ference between f-MWCNTs and the LDH in the survived cells.
Briefly, the LDH was quantified in the cells that survived the
treatment after being artificially lysed. A549 cells were seeded at
15 000 cells (24 h) and 8000 cells (72 h) in a 96 well plate and left
to attach overnight. Cells were then treated with polyamine–
CNTs at 1, 10 and 100 mg mLÀ1 concentrations. In addition, 10%
DMSO or 1 mM H2O2 were used as positive controls to induce
cytotoxicity. After 24 h or 72 h treatment, cells were lysed with
10 mL of lysis buꢀffer per 100 mL serum free media and left for
45–60 min at 37 C. Fifty microlitres of cell lysate after centri-
fugation (13 000 rpm, 5 min) were mixed with 50 mL substrate
mixture in a new microtiter plate and incubated for 15 minutes at
room temperature. Absorbance at 490 nm was read in an ELISA
plate reader. A centrifugation step was included to remove any
nanotubes from the assay mixture which could interfere with the
accuracy of the assay. The amount of LDH released was an
indication of the number of cells that survived the treatment and
(ox-MWCNTs) were obtained after modification of earlier
reported procedures.25,29 The strong acid treatment generates
defects on the sidewalls and formation of open ends that are both
terminated by carboxylic acid groups. The Boc protected poly-
amines spermidine and spermine were synthesized in turn
through stepwise elongation of 1,4-diaminobutane backbone
before reaction with acid chlorides. This allows obtaining an
elongated aliphatic chain of the polyamine backbone thus
introducing an increasing number of amino groups per func-
tional group. The solution phase synthesis of these orthogonally
protected asymmetrical or symmetrical polyamines can be
carried out using the combination of Boc, phthalimide and
benzyl protecting groups (see ESI†).
For carrying out the addition reaction, we prepared 4-nitro-
benzoyl derivatives of polyamines, which were eventually
reduced to obtain the corresponding 4-aminobenzoyl derivatives
(15, 21 and 27) of each polyamine in good yields. Pristine
SWCNTs and purified MWCNTs were reacted with 4-amino-
benzoyl derivatives of polyamines in o-dichlorobenzene using the
diazotization coupling procedure to produce polyamine-func-
tionalized SWCNTs (f-SWCNTs) (1a, 3a and 5a) and MWCNTs
(f-MWCNTs) (2a, 4a and 6a) (Scheme 1). This in situ
This journal is ª The Royal Society of Chemistry 2011
J. Mater. Chem., 2011, 21, 4850–4860 | 4853

View Online
Scheme 1 Addition reactions to SWCNTs and MWCNTs of 4-aminobenzoyl derivatives of polyamines (putrescine, spermidine and spermine). (a)
Isoamyl nitrite, o-dichlorobenzene : CH3CN (3 : 1), 4-aminobenzoyl derivatives (15, 21 and 27, respectively), 60 ꢀC, 36 h; (b) 4 M HCl solution in
dioxane, rt, 12 h.
diazotization coupling to CNTs allows effective functionaliza-
tion without the use of unstable aryl diazonium salts.
a higher functionalization degree than MWCNTs. This is most
likely because the content of carbon on MWCNTs is higher than
on SWCNTs due to the presence of the additional concentric
internal layers. Among the different polyamines used, we
observed higher weight loss for putrescine derivative followed by
spermidine and spermine owing to the presence of the increasing
number of the bulky Boc protecting groups that might reduce the
reactive available area, probably inducing steric hindrance that
prevents further additions. The TGA curves of f-SWCNTs 1b, 3b
and 5b showed a gradual weight loss of about 20.8, 15.4 and
Alternatively, using the amidation approach, f-SWCNTs
(7a, 9a and 11a) and f-MWCNTs (8a, 10a and 12a) were
synthesized following the acyl chloride route,28 allowing the
oxidized CNTs to react with oxalyl chloride under argon at
refluxing conditions. The resulting acid chloride derivatives were
then heated under reflux with 4-aminobenzoyl derivatives of
polyamines 15, 21 and 27, respectively (Scheme 2).
All functionalized CNTs were then treated with a solution of
HCl in dioxane to remove the protecting groups affording the
different cationic conjugates (1b–12b).
ꢀ
14.1% at 600 C, respectively, as compared to 7.8, 8.5 and 9.0%
for f-MWCNTs 2b, 4b and 6b, respectively. On the basis of TGA
weight loss, we estimated that the amount of functional groups
per gram (fw, mmol gÀ1) of f-SWCNTs 1b, 3b and 5b is 0.91, 0.48
and 0.34, respectively, corresponding to a degree of functional-
ization of one polyamine group every 73, 147 and 211 carbon
atoms, respectively. In the case of MWCNTs, the number of
functional groups per gram of f-MWCNTs 2b, 4b and 6b is 0.34,
0.26 and 0.22, respectively, corresponding to a degree of func-
tionalization of one functional group every 225, 288 and 349
carbon atoms, respectively.
The degree of functionalization for the two families of poly-
amine-based CNTs (1b–12b) has been evaluated by TGA under
N2 atmosphere (Fig. 1). Pristine SWCNTs andꢀ purified
MWCNTs lost only 1.1 and 1.2% of mass up to 700 C, while
TGA curves of the functionalized nanotubes 1b–12b display
a considerable decrease of weight in the range of 200–600 ꢀC
corresponding to the decomposition of the attached 4-amino-
benzoyl polyamine moiety. Fig. 1(a) and (b) put in evidence the
TGA differences between SWCNTs and MWCNTs, respectively,
functionalized with polyamines through the addition reaction.
We can derive that the exohedral surface of SWCNTs has
Similarly, Fig. 1(c) and (d) shows the TGA curves of SWCNTs
and MWCNTs functionalized with polyamines through
4854 | J. Mater. Chem., 2011, 21, 4850–4860
This journal is ª The Royal Society of Chemistry 2011

View Online
Scheme 2 Amidation reactions on ox-SWCNTs and ox-MWCNTs with 4-aminobenzoyl derivatives of polyamines (putrescine, spermidine and sper-
mine). (a) Neat (COCl)2; (b) 4-aminobenzoyl derivatives (15, 21 and 27 respectively), 60 ꢀC, 48 h, THF (dry); (c) 4 M HCl solution in dioxane, rt, 12 h.
amidation reaction, respectively. The TGA curves of f-SWCNTs
7b, 9b and 11b showed a gradual weight loss of about 21.5, 18.7
and 17.5% at 600 ꢀC, respectively, as compared to 16.0% for the
ox-SWCNTs whereas, TGA curves of f-MWCNTs 8b, 10b and
12b showed a gradual weight loss of about 25.5, 23.2 and 19.1%
at 600 ꢀC, respectively, as compared to 13.1% for the
ox-MWCNTs. From the TGA weight loss, we calculated 0.79,
0.51 and 0.38 mmol of functional groups per gram of f-SWCNTs
7b, 9b and 11b, that is, one group every 83, 132 and 180 SWCNT
carbon atoms. Similarly, 0.94, 0.63 and 0.42 mmol functional
groups per gram of f-MWCNTs 8b, 10b and 12b was calculated
which corresponds to one group every 66, 100 and 161 MWCNT
carbon atoms. In the case of the amidation reaction MWCNTs
seem to give a better functionalization yield in comparison to
SWCNTs, while the opposite situation was found for the direct
addition to the tubes.
shown in Fig. 2. We have observed that a high degree of
de-bundling occurred in the case of polyamine-functionalized
SWCNTs as compared to pristine SWCNTs, which is consistent
with the degree of functionalization assessed by TGA results.
Fig. 2(a)–(f) displays small bundles of polyamine–CNTs depos-
ited from a methanol/H2O (1 : 1) solution onto a carbon coated
TEM grid. Comparing the TEM images of different polyamine–
CNTs, it can be evaluated that polyamine–CNTs 3b, 4b (Fig. 2(b)
and (e)) and polyamine–CNTs 5b, 6b (Fig. 2(c) and (f)) show
more debundling both in the case of SWCNTs and MWCNTs in
comparison to polyamine–CNTs 1b, 2b (Fig. 2(a) and (d)) under
the same conditions. This behaviour can be attributed to the
presence of the longer chains in the case of spermidine and
spermine and the higher number of positive charges that can
produce higher electrostatic repulsions.
Fig. 3 displays TEM images of polyamine–CNTs 7b–12b
synthesized by amidation reaction deposited again from a meth-
anol/H2O (1 : 1) solution onto a carbon coated TEM grid. In
comparison to oxidized CNTs26 bundles of polyamine–CNTs
The polyamine–CNT hybrids were also characterized by
transmission electron microscopy (TEM). The TEM images of
polyamine–CNTs 1b–6b synthesized by the addition reaction are
This journal is ª The Royal Society of Chemistry 2011
J. Mater. Chem., 2011, 21, 4850–4860 | 4855

View Online
Fig. 1 Thermogravimetric analysis for ox-CNTs and f-CNTs 1b–12b. (a) Thermogravimetric analysis of f-SWCNTs 1b, 3b, and 5b; (b) Thermogra-
vimetric analysis of f-MWCNTs 2b, 4b, and 6b; (c) Thermogravimetric analysis of f-SWCNTs 7b, 9b, and 11b and ox-SWCNTs; (d) Thermogravimetric
analysis of f-MWCNTs 8b, 10b, and 12b and ox-MWCNTs, in N2 atmosphere with a ramp of 10 ꢀC minÀ1
.
7b–12b are more dissociated and less tightly bound to each other,
which further confirmed the functionalization with different
polyamines and its role in modulating the dispersibility proper-
ties of carbon nanotubes. The nanotubes are very short in
accordance with the long oxidation process. However, as can be
seen from the images, nanotubes remained largely intact after
functionalization.
Additional characterization was performed using Raman
spectroscopy. This technique allows the evaluation of the
SWCNT properties after functionalization. Although the infor-
mation obtained from Raman spectra of CNTs should be
considered merely qualitative, we have investigated whether the
structural properties of CNTs are preserved after functionaliza-
tion. The D/G ratio provides information about the quality of
Fig. 2 TEM images of polyamine–CNTs prepared using addition reaction; TEM images of putrescine-functionalized SWCNTs 1b (a) and MWCNTs
2b (d), TEM images of spermidine-functionalized SWCNTs 3b (b) and MWCNTs 4b (e) and TEM images of spermine-functionalized SWCNTs 5b (c)
and MWCNTs 6b (f).
4856 | J. Mater. Chem., 2011, 21, 4850–4860
This journal is ª The Royal Society of Chemistry 2011

View Online
Fig. 3 TEM images of polyamine–CNT hybrids prepared using amidation reaction; TEM images of putrescine-functionalized SWCNTs 7b (a) and
MWCNTs 8b (d), TEM images of spermidine-functionalized SWCNTs 9b (b) and MWCNTs 10b (e) and TEM images of spermine-functionalized
SWCNTs 11b (c) and MWCNTs 12b (f).
these materials. The relative amplitude of D and G peaks was
accurately measured for all the spectra, where the G band was
much higher than the D band in all derivatives (Fig. 4). This
shows that polyamine–CNTs bear a relatively low quantity of
defects mainly preserving their characteristic properties, irre-
spective of the different types of functionalization employed.
Furthermore, from a comparison between the D and G ratio on
the one hand, and the radial breathing mode features on the
other (RBM, 150–300 cmÀ1) we can derive information about the
degree of functionalization and, more precisely, whether
f-SWCNTs have maintained their structures after functionali-
zation or not. On average the D/G ratio ranges between 0.133
and 0.185 for SWCNTs 1b, 3b and 5b while the value in the
pristine material was 0.126, indicating that surface functionali-
zation occurred without significantly affecting the structure of
the nanotubes. In the case of ox-SWCNTs 7b, 9b and 11b we
analyzed more carefully the RBM region as the values of the D/G
ratio (0.113–0.219) were thought to be due to the functional
groups mainly at the tips. The RBM bands after the amidation
reaction had high intensity meaning that the structure of the
ox-SWCNTs was not disrupted. This was also considered as
evidence that the oxidation and consequently the amidation
reactions were more selective to the CNT tips with respect to the
in situ formation of the diazonium salt reaction that was likely
occurring across the CNT backbone surface.
Cytotoxicity and siRNA binding
With the intention of using these new families of polyamine-
functionalized carbon nanotubes for binding and delivering
siRNA, we analyzed the impact of their exposure to cells in
terms of cytotoxic adverse effects. Human lung epithelial cells
(A549) were incubated with polyamine–CNTs 1b–12b at
different concentrations (1–100 mg mLÀ1) for 24 or 72 h
(to assess prolonged cytotoxicity). Cytotoxicity was evaluated
by the modified lactate dehydrogenase (LDH) assay (Fig. 5).
Fig. 4 Raman spectra of f-SWCNTs 1b, 3b, and 5b (left), and of f-SWCNTs 7b, 9b, and 11b (right).
This journal is ª The Royal Society of Chemistry 2011
J. Mater. Chem., 2011, 21, 4850–4860 | 4857

View Online
The LDH assay measures membrane integrity as a function of
the amount of cytoplasmic LDH released into the medium.
LDH assay can also be used to express the cell viability by
measuring the total cytoplasmic LDH remaining in the viable
cells. This assay was modified to avoid any interference from the
CNTs during assessment.28 A549 cells were not affected by the
different concentrations of f-SWCNTs both at 24 and 72 h
(Fig. 5 and SI-1†). However, in the case of the f-MWCNTs, we
observed dose-dependent cytotoxicity for the conjugates 2b and
4b. Interestingly, cytotoxicity decreased as the length of the
polyamine increased with almost no observed cell death at 1 and
10 mg mLÀ1 of the spermine-containing nanotubes 6b after 24 h
of incubation (Fig. 5A). It seems that the length of the poly-
amine chain influences the cell viability particularly in the case
of these f-MWCNTs. Similarly, the impact on cell viability was
evaluated for the family of conjugates modified by the amida-
tion reaction. Results in Fig. 5B indicate that no cytotoxicity
was observed up to 100 mg mLÀ1 for 24 h exposure to the
f-SWCNTs and f-MWCNTs conjugates modified by the ami-
dation reaction (7b–12b). However, a certain degree of cyto-
toxicity was observed after prolonged exposure (72 h) only at
the highest concentration tested (100 mg mLÀ1) which was more
significant for f-SWCNTs than f-MWCNTs (see ESI, Fig. SI-
2†). Overall the cytotoxic effects are significant after 24 h only
in the case of MWCNTs functionalized by addition reaction.
This could be related to the dispersibility properties of the
different cationic CNT conjugates. In particular, the MWCNTs
(2b, 4b and 6b) modified by the addition reaction were difficult
to disperse. The increased number of protonated amines in
spermidine- and spermine-functionalized MWCNTs does not
really facilitate the dispersion of the CNTs in the cell
culture medium by comparison with putrescine-function-
alized MWCNTs. The poor dispersibility likely affects cell
viability.
Fig. 5 Cell viability of A549 cells assessed by the modified LDH assay after 24 h exposure to f-CNTs (1–100 mg mLÀ1). (A) LDH assay for f-SWCNTs
1b, 3b and 5b, and f-MWCNTs 2b, 4b and 6b. Cell viability was not affected by exposure to f-SWCNTs but was decreased by the f-MWCNT exposure at
10–100 mg mLÀ1 concentrations. (B) LDH assay of A549 cells treated with f-SWCNTs 7b, 9b, and 11b, and f-MWCNTs 8b, 10b, and 12b. Cell viability
was not affected by 24 h exposure to none of these f-CNTs. Ten percent DMSO and 1 mM H2O2 in complete cell culture medium were used as
positive controls. The vehicle (5% dextrose) was used as a negative control. Percentage viability was expressed as percentage of control untreated cells
(*, p # 0.05).
4858 | J. Mater. Chem., 2011, 21, 4850–4860
This journal is ª The Royal Society of Chemistry 2011

View Online
Fig. 6 Electrophoretic motility of siRNA complexed with polyamine–CNTs 1b (top left), 2b (top right), 7b (bottom left) and 8b (bottom right)
(reconstituted at 1 mg mLÀ1 in 5% dextrose). Various N/P (+/À) ratios of the four conjugates were complexed to a fixed 0.5 mg of siRNA.
We previously demonstrated that cationic carbon nanotubes
were able to condense nucleic acids efficiently and deliver them
into mammalian cells.9,10,30 Similarly, we have performed
preliminary experiments on the complexation between poly-
amine–CNT conjugates and siRNA using agarose gel electro-
phoresis (Fig. 6). Among the three polyamine-functionalized
CNT families, we tested single- and multi-walled CNTs modified
with putrescine by both the addition reaction (1b and 2b) and
amidation (7b and 8b) mainly because these nanotubes were the
most dispersed in water, so we could achieve highest +/À (N/P)
charge ratios at complexation with the nucleic acids. The amount
of uncomplexed (free) siRNA migrating into the gel was reduced
by increasing the polyamine–CNT : siRNA ratio as evidenced by
a reduction in ethidium bromide fluorescence intensity signals at
3 : 1 (+/À) charge ratio, only in the case of the nanotubes func-
tionalized by the addition reaction (1b and 2b). Unfortunately,
no complexation was achieved with amidated nanotubes up to
18 : 1 (+/À) charge ratio (Fig. 6 and SI-4†). We can conclude
from the data obtained with putrescine-functionalized SWCNTs
and MWCNTs that: (i) there was no difference in siRNA binding
capacity between the f-SWCNTs and f-MWCNTs; and (ii) more
efficient complexation was achieved by the nanotubes modified
by the addition than amidation reaction (1b and 2b vs. 7b and
8b), likely due to a more homogeneous distribution of the
functional groups on the sidewall of CNTs in the case of the
addition reaction. Furthermore, increasing the number of amino
groups in the side chain (i.e. from putrescine to spermidine) in the
case of amidated CNTs did not seem to improve the siRNA
binding (see Fig. SI-3†), probably due to the localization of the
functional groups mainly at the tips. Minor complexation of
siRNA occurred at 18 : 1 (+/À) charge ratio in the case of
f-MWCNTs 10b.
In conclusion, we have synthesized and characterized twelve
different polyamine-modified single- and multi-walled carbon
nanotubes. Most of them displayed reduced cytotoxicity on
A549 cells. Aqueous dispersibility was significantly improved for
the conjugates modified by the amidation reaction of oxidized
CNTs. Putrescine-modified CNTs were able to efficiently
complex siRNA as assessed by gel electrophoresis experiments.
In addition, we have also shown that these conjugates interact
and can be internalized by the A549 cells (see Fig. SI-4†) using
light microscopy. This finding opens the possibility to further
develop some of these cationic conjugates for gene delivery and
silencing. Experiments testing gene silencing using these new
conjugates are in due course.
Acknowledgements
This work was partly supported by the French-Indian CEFI-
PRA/IFCPAR collaborative project (Project no. 3705-2), the
University of Trieste, MIUR (PRIN contract No. 20085M27SS),
and ERC Advanced Grant Carbonanobridge (to M.P.). P.S.
wishes to thank CEFIPRA/IFCPAR for
a post-doctoral
fellowship. TEM images were recorded at the RIO Microscopy
Facility Plate-form of Esplanade Campus (Strasbourg, France).
This journal is ª The Royal Society of Chemistry 2011
J. Mater. Chem., 2011, 21, 4850–4860 | 4859

View Online
12 D. H. Russel and T. A. McVicker, Biochim. Biophys. Acta, 1972, 259,
247.
13 S. A. McCormack and L. R. Johnson, Am. J. Physiol., 1991, 260,
G795.
C.B. is funded by Nanotrans project (INSERM/INERIS/
MEEDM, Programme Post-Grenelle de l’Environnement).
14 M. A. Grillo, Int. J. Biochem., 1985, 17, 943.
15 L. J. Marton and A. E. Pegg, Annu. Rev. Pharmacol. Toxicol., 1995,
33, 55.
16 S. S. Cohen, A Guide to the Polyamines, Oxford University Press,
Oxford, 1998.
References
1 (a) M. Monthioux and V. L. Kuznetsov, Carbon, 2006, 44, 1621; (b)
S. Iijima, Nature, 1991, 354, 56; (c) S. Iijima and T. Ichihashi,
Nature, 1993, 363, 603.
2 (a) R. C. Haddon, ed., Carbon Nanotubes. Special Issue. Acc. Chem.
Res., 2002, 35, 997–1113; (b) A. Jorio, G. Dresselhaus and
M. S. Dresselhaus, Carbon Nanotubes: Advanced Topics in the
Synthesis, Structure, Properties and Applications, Springer, Berlin,
2008.
17 R. A. Casero, Jr. and P. M. Woster, J. Med. Chem., 2009, 52, 4551.
18 B. G. Feuerstein, L. D. Williams, H. S. Basu and L. J. Marton, J. Cell.
Biochem., 1991, 46, 37.
19 (a) H.-H. Chen, M. J. Behe and D. C. Rau, Nucleic Acids Res., 1984,
12, 2381; (b) W. C. Russell, B. Precious, S. R. Martin and
P. M. Bayley, EMBO J., 1983, 2, 1647.
ꢀ
`
3 (a) C. Menard-Moyon, E. Venturelli, C. Fabbro, C. Samorı, T. Da
Ros, K. Kostarelos, M. Prato and A. Bianco, Expert Opin. Drug
Delivery, 2010, 5, 691; (b) H.-C. Wu, X. Chang, L. Liu, F. Zhao
and Y. Zhao, J. Mater. Chem., 2010, 20, 1036.
20 K. Igarashi and K. Kashiwagi, Biochem. Biophys. Res. Commun.,
2000, 271, 559.
21 (a) M. K. Soltan, H. M. Ghonaim, M. El Sadek, M. A. Kull, L. A. El-
aziz and I. S. Blagbrough, Pharm. Res., 2009, 26, 286; (b) D. Jere,
J. E. Kim, R. Arote, H. L. Jiang, Y. K. Kim, Y. J. Choi,
C. H. Yun, M. H. Cho and C. S. Cho, Biomaterials, 2009, 30, 1635;
(c) K. Nagane, J. Jo and Y. Tabata, Tissue Eng., Part A, 2010, 16, 21.
22 Z. Akhter, A. Nigar, M. Y. Razzaq and H. M. Siddiqi, J. Organomet.
Chem., 2007, 692, 3542.
23 C. A. Dyke and J. M. Tour, Chem.–Eur. J., 2004, 10, 812.
24 Y. Yang, S. Qiu, X. Wang and R. K. Y. Li, Appl. Surf. Sci., 2010, 256,
3286.
4 (a) F. Cataldo and T. Da Ros, Medicinal Chemistry and
Pharmacological Potential of Fullerenes and Carbon Nanotubes,
Springer, 2008; (b) N. Martin and D. M. Guldi, Carbon Nanotubes
and Related Structures, Wiley-VCH, Weinheim, 2010.
5 (a) L. Lacerda, A. Bianco, M. Prato and K. Kostarelos, J. Mater.
Chem., 2008, 18, 17; (b) F. Lu, L. Gu, M. J. Meziani, X. Wang,
P. G. Luo, L. M. Veca, L. Cao and Y.-P. Sun, Adv. Mater., 2009,
21, 139.
6 (a) P. Singh, S. Campidelli, S. Giordani, D. Bonifazi, A. Bianco and
M. Prato, Chem. Soc. Rev., 2009, 38, 2214; (b) D. Tasis,
N. Tagmatarchis, A. Bianco and M. Prato, Chem. Rev., 2006, 106,
1105.
€
25 S. Li, W. Wu, S. Campidelli, V. Sarnatskaıa, M. Prato, A. Tridon,
A. Nikolaev, V. Nikolaev, A. Bianco and E. Snezhkova, Carbon,
2008, 46, 1091.
7 (a) A. Hirsch, Angew. Chem., Int. Ed., 2002, 41, 1853; (b) N. Karousis
and N. Tagmatarchis, Chem. Rev., 2010, 110, 5366.
26 P. Singh, J. Kumar, F. M. Toma, J. Raya, M. Prato, B. Fabre,
S. Verma and A. Bianco, J. Am. Chem. Soc., 2009, 131, 13555.
27 J. Liu, A. G. Rinzler, H. Dai, J. H. Hafner, R. K. Bradley and
P. J. Boul, Science, 1998, 280, 1253.
`
28 C. Samorı, H. Ali-Boucetta, R. Sainz, C. Guo, F. M. Toma,
C. Fabbro, T. da Ros, M. Prato, K. Kostarelos and A. Bianco,
8 (a) D. Nepal and K. E. Geckeler, Functional Nanomaterials in
Functionalization of Carbon Nanotubes, ed. K. E. Geckeler and E.
Rosenberg, American Scientific Publishers, California, 2006, pp. 57–
79; (b) A. Krueger and M. Monthioux, Strained Hydrocarbons in
Carbon Nanotubes, ed. H. Dodziuk, Wiley-VCH, Weinheim, 2009,
pp. 335–373.
9 (a) J. E. Podesta, K. T. Al-Jamal, M.-A. Herrero, B. Tian, H. Ali-
Boucetta, V. Hegde, A. Bianco, M. Prato and K. Kostarelos, Small,
2009, 5, 1176; (b) M.-A. Herrero, F. M. Toma, K. T. Al-Jamal,
K. Kostarelos, A. Bianco, T. Da Ros, F. Bano, L. Casalis,
G. Scoles and M. Prato, J. Am. Chem. Soc., 2010, 131, 9843.
10 K. T. Al-Jamal, F. M. Toma, A. Yilmazer, H. Ali-Boucetta,
A. Nunes, M.-A. Herrero, B. Tian, A. Eddaoui, W. T. Al-Jamal,
A. Bianco, M. Prato and K. Kostarelos, FASEB J., 2010, 24, 1.
11 C. W. Tabor and H. Tabor, Annu. Rev. Biochem., 1976, 45, 285.
Chem. Commun., 2010, 46, 1494.
`
ꢀ
29 C. Samorı, R. Sainz, C. Menard-Moyon, F. M. Toma, E. Venturelli,
P. Singh, M. Ballestri, M. Prato and A. Bianco, Carbon, 2010, 48,
2447.
30 (a) D. Pantarotto, R. Singh, D. McCarthy, M. Erhardt, J.-P. Briand,
M. Prato, K. Kostarelos and A. Bianco, Angew. Chem., Int. Ed., 2004,
43, 5242; (b) L. Lacerda, A. Bianco, M. Prato and K. Kostarelos, J.
Mater. Chem., 2008, 18, 17; (c) R. Singh, D. Pantarotto,
D. McCarthy, O. Chaloin, J. Hoebeke, C. D. Partidos, J.-P. Briand,
M. Prato, A. Bianco and K. Kostarelos, J. Am. Chem. Soc., 2005,
127, 4388.
4860 | J. Mater. Chem., 2011, 21, 4850–4860
This journal is ª The Royal Society of Chemistry 2011