Cross-linked self-assembled micelle based nanosensor for intracellular pH measurements

Article abstract of DOI:10.1039/c4tb00446a

A micelle based nanosensor was synthesized and investigated as a ratiometric pH sensor for use in measurements in living cells by fluorescent microscopy. The nanosensor synthesis was based on self-assembly of an amphiphilic triblock copolymer, which was chemically cross-linked after micelle formation. The copolymer, poly(ethylene glycol)-b-poly(2-aminoethyl methacrylate)-b-poly(styrene) (PEG-b-PAEMA-b-PS), was synthesized by isolated macroinitiator atom transfer radical polymerization that forms micelles spontaneously in water. The PAEMA shell of the micelle was hereafter cross-linked by an amidation reaction using 3,6,9-trioxaundecandioic acid cross-linker. The cross-linked micelle was functionalized with two pH sensitive fluorophores and one reference fluorophore, which resulted in a highly uniform ratiometric pH nanosensor with a diameter of 29 nm. The use of two sensor fluorophores provided a sensor with a very broad measurement range that seems to be influenced by the chemical design of the sensor. Cell experiments show that the sensor is capable of monitoring the pH distributions in HeLa cells. This journal is

Full text of DOI:10.1039/c4tb00446a

Journal of
Materials Chemistry B
View Article Online
View Journal | View Issue
PAPER
Cross-linked self-assembled micelle based
nanosensor for intracellular pH measurements†
Cite this: J. Mater. Chem. B, 2014, 2,
652
6
ab
ab
a
E. K. Pramod Kumar, Rikke Vicki Søndergaard, Barbara Windschiegl,
a
ab
Kristoffer Almdal and Thomas L. Andresen*
A micelle based nanosensor was synthesized and investigated as a ratiometric pH sensor for use in
measurements in living cells by fluorescent microscopy. The nanosensor synthesis was based on self-
assembly of an amphiphilic triblock copolymer, which was chemically cross-linked after micelle
formation. The copolymer, poly(ethylene glycol)-b-poly(2-aminoethyl methacrylate)-b-poly(styrene)
(PEG-b-PAEMA-b-PS), was synthesized by isolated macroinitiator atom transfer radical polymerization
that forms micelles spontaneously in water. The PAEMA shell of the micelle was hereafter cross-linked
by an amidation reaction using 3,6,9-trioxaundecandioic acid cross-linker. The cross-linked micelle was
functionalized with two pH sensitive fluorophores and one reference fluorophore, which resulted in a
highly uniform ratiometric pH nanosensor with a diameter of 29 nm. The use of two sensor fluorophores
provided a sensor with a very broad measurement range that seems to be influenced by the chemical
design of the sensor. Cell experiments show that the sensor is capable of monitoring the pH
distributions in HeLa cells.
Received 21st March 2014
Accepted 23rd July 2014
DOI: 10.1039/c4tb00446a
www.rsc.org/MaterialsB
15,16
with high structural control.
However, such micelles can be
Introduction
prone to disintegration under diluted biological conditions,
In cells, intracellular pH plays an important role for the struc- which hampers their use as nanosensors and more advanced
tural stability and function of proteins, in organelles such as strategies are therefore needed.
lysosomes and mitochondria, as well as in cell cycle progression
Controlled radical polymerization, especially redox-active
1–4
and apoptosis. In nanoparticle based drug delivery systems, transition metal complex catalyzed atom transfer radical poly-
17
the pH in endosomes and lysosomes has been a target for merization (ATRP) can provide well dened block co-polymers
controlling drug release specically in diseased cells, and an for use in advanced functional materials with high control of
increased understanding of pH proles in the endosome-lyso- composition and molecular architecture. Amphiphilic block
some pathway is important for future development of pH copolymers having denite hydrophilic to hydrophobic block
5
sensitive nanocarrier systems. A number of uorescent probe size ratios can provide a variety of structural morphologies on
18,19
based nanosensors have been developed in the last decade for nanoscale such as micelles, lamellae, tubes and rods.
In a
intracellular pH measurement by various groups and the efforts biological setting these structures behave differently and it is an
in this area is high as several challenges remains in creating interesting perspective to develop nanosensors based on
highly reproducible sensor particles that interact with cells in a different morphologies using self-assembly principles. In
6
–11
controlled way.
develop nanosensors that can be used for in vivo measure- control of the surface chemistry in comparison to other
A long term objective is furthermore to addition, self-assembly of nanostructures can enhance the
12
20
ments, where polymeric micelles with a PEG corona may be of methods and thereby provide a platform for introducing bio-
particular interest due to their low immunogenicity and ability logical targeting molecules. However, the self-assembly also
to overcome clearance by macrophages.
culties in synthesizing well dened nanosensors have enhanced nanoparticle constructs that are designed for self-assembly is
13,14
In addition, diffi- introduces a risk of disintegration and post-cross-linking of
21,22
the interest in developing self-assembled polymeric micelles therefore highly important to ensure nanosensor integrity.
The focus of this article is to design a nanosensor that is
synthesized by the use of triblock copolymers that self-assemble
into core–shell–corona micelles in water. In the shell region we
installed functional groups that allowed for chemical cross-
a
DTU Nanotech, Department of Micro-and Nanotechnology, Technical University of
Denmark, Building 423, 2800 Lyngby, Denmark. E-mail: thomas.andresen@
nanotech.dtu.dk
b
Center for Nanomedicine and Theranostics, Technical University of Denmark, linking of the micelle to ensure nanosensor integrity under
Building 423, 2800 Lyngby, Denmark
biological conditions. The formed cross-linked micelle could
further be post-functionalized with uorescent dyes that are
†
Electronic supplementary information (ESI) available: See DOI:
0.1039/c4tb00446a
1
6652 | J. Mater. Chem. B, 2014, 2, 6652–6659
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
Journal of Materials Chemistry B
sensitive to pH resulting in a nanosensor with a broad pH duration of 2 minutes each. Zeta potential measurements were
spectrum that allowed for cellular measurements.
carried out using a Brookhaven ZETA PALS analyzer. The
measurements were made in MilliQ water at 25 C and the zeta
ꢂ
potential (x) was calculated using the Smoluchowski equation,
i.e., electrophoretic mobility (m) ¼ x3/h, where h and 3 are the
absolute viscosity and dielectric constant of the medium,
respectively. The mean value of x was calculated from ten data
accumulations. Fluorescence measurements were carried out
using an EDINBURGH F-900 uorometer. A PSIA XE-150
microscope was used for AFM measurements. Argon atmo-
sphere (99.9999%) used in the reactions was provided by AGA
Denmark.
Experimental section
Materials
Styrene (99.5%) was obtained from Fluka and the radical
inhibitors were removed by passing through a column lled
with basic alumina. 2-Aminoethyl methacrylate hydrochloride
(AEMA$HCl) (90%), bis(tert-butyl) dicarbonate (99%), triethyl-
amine (TEA) (99.5%), 2-bromo-isobutyryl-bromide (98%), CuCl
0
(
(
(
99.995%), 2,2 bipyridyl (bpy) (99%), PMDETA (99%), CuCl
2
99.995%), triuoroacetic acid (TFA) (99%), dialysis tubing
0
MWCO
¼
12 kDa), N-(3-dimetylaminopropyl)-N -ethyl-
Synthesis of PEG127-b-PAEMA12-b-PS28
carbodiimide methiodide (EDC$MeI), rhodamine
thiocyanate (RhBITC) and uorescein 5(6)-isothiocyanate
FITC) (90%) were purchased from Sigma Aldrich and used as
B iso-
The macroinitiator PEG127Br and the monomer 2-[N-(tert-
butoxycarbonyl)amino]ethyl methacrylate (AEMA(Boc)) were
(
23,24
synthesized by previously reported procedures.
obtained. 3,6,9-Trioxaundecandioic acid (TUDA) was purchased
from iris biotech GMBH, Oregon Green 488 isothiocyanate
PEG127-b-P(AEMA(Boc))12Cl. PEG127-Br (2 gram, 0.36 mmol),
0
AEMABoc (852 mg, 5 mmol), 2,2 bipyridyl (122 mg, 0.76 mmol)
(OGITC) was purchased from Invitrogen, and CH
3
O-PEG-OH
and 10 mL dry MeOH were added to a 25 mL schlenk ask
equipped with a stir bar. The ask was frozen in liquid nitrogen
and CuCl catalyst (40 mg, 0.40 mmol) was added. The reaction
mixture was degassed with 3 freeze–pump–thaw cycles (each 15
minute long) to remove the oxygen and the polymerization was
(
poly (ethylene glycol) monomethylether (M
n
¼ 5000, was
measured to M
n
¼ 5500 by NMR) was from Fluka. Solvents used
for atom transfer polymerization (ATRP) were puried by
distillation over the drying agents indicated in parentheses
MeOH (Mg(OMe) ), DMF (CaH ). Solvents are stored over
molecular sieves (MeOH 3 A and DMF 4 A) and transferred
under argon. Other solvents and commercially available
chemicals were used as obtained. Water used was collected
from Millipore and aqueous buffer solutions were prepared
from the reported procedures.
2
2
ꢂ
carried out at 40 C for 24 h under argon atmosphere. The dark
˚
˚
brown reaction solution was passed through a silica gel column
to remove the copper catalyst using MeOH as solvent. On
exposure to air, the solution turned to blue, which indicated the
aerial oxidation of the Cu(I) catalyst. Aer the removal of most of
the MeOH by rotary evaporation, the polymer was precipitated
into excess cold diethyl ether. The precipitate was then isolated
Instrumentation details
1
by ltration and dried under vacuum to yield 1.86 g (62%) of the
1
H-NMR spectra were recorded on a Bruker 250 MHz in solvents diblock copolymer. H-NMR (250 MHz, CDCl
3
): d ¼ 5.50 (br s,
CH O), 3.37 (br m,
backbone), 1.43 (s,
–C(CH ) ), 1.11–0.81 (m, –C(CH ) , –CH backbone); FT-IR
as indicated. The chemical shis (d) were given in ppm relative –NH), 4.0 (br s, –OCH
to TMS. The residual solvent signals were used as a reference, –OCH CH NH, –OCH
2
CH
2
NH), 3.63 (s, –CH
2
2
2
2
3
), 1.82 (br, –CH
2
and the chemical shis converted into TMS scale (CDCl : d
7
¼
3
H
3 3
3 2
3
6
ꢀ1
.24 ppm, d-DMSO: dH ¼ 2.50 ppm, D O: d ¼ 4.79 ppm). (cm ): 3387, 2892, 1716, 1520, 1466, 1391, 1361, 1342, 1279,
2
H
Infrared spectra were recorded using a Perkin Elmer FT-IR 1241, 1150, 1112, 1060, 996, 965, 843; Mn (NMR) ¼ 8340; Mn
Spectrometer (KBr pellets method) and the wavenumbers of (GPC) ¼ 5600, M
w
¼ 6460, PD (M
w
/M
n
) ¼ 1.15.
ꢀ1
recorded IR signals quoted in cm . The number-average
molecular weight (M ), weight average molecular weight (M
n 2
and polydispersity (M /M ) of block copolymers were deter- mmol), CuCl (13 mg, 0.096 mmol), PMDETA (0.087 mL, 0.42
Synthesis of PEG127-b-P(AEMA(Boc))12-b-PS28 (1). PEG127-b-
n
w
), P(AEMA(Boc))12-Cl (1.0 g, 0.12 mmol), styrene (0.48 mL, 4.2
w
mined by GPC analysis based on poly styrene calibration stan- mmol), and 3 mL of DMF were added into a 25 mL schlenk ask
dards. Measurements were carried out by using a Mixed-D GPC equipped with a stir bar. The ask was frozen in liquid nitrogen
column from Polymer Laboratories (7.4 ꢁ 300 mm) and a and the CuCl catalyst (12 mg, 0.12 mmol) was added to it. Aer
RID10A-SHIMADZU refractive index detector. DMF with 50 mM being degassed with 3 freeze–pump–thaw cycles (each cycle was
ꢀ
1
ꢂ
LiCl solution was used as eluent (1 mL min ) at 25 C. LiCl was 15 minutes long) to remove the oxygen, the polymerization was
ꢂ
used to minimize polymer aggregation in DMF and hence carried out at 130 C for 27 h under an argon atmosphere. The
misleading micelle formation (hydrodynamic diameter issues) reaction mixture was concentrated under vacuum and the
h
can be avoided. Hydrodynamic diameters (D ) and size distri- polymer was precipitated in cold diethyl ether. The precipitate
butions of the amphiphilic colloidal dispersion in MilliQ water was dried under vacuum to yield 0.74 g (55%) of the triblock
ꢂ
1
6
at 25 C were determined by Brookhaven ZETA PALS instru- copolymer. H-NMR (250 MHz, d-DMSO): d ¼ 7.3–6.3 (m, ArH),
ment. Calculation of the particle size distribution and distri- 3.84 (br, –OCH CH NH), 3.50 (s, –CH CH O), 3.18 (br, –OCH -
2
2
2
2
2
bution averages were performed with the ISDA soware package CH NH), 2.10–1.60 (br, CH and –CH backbone), 1.36 (s,
2
2
ꢀ
1
(
3 3 3
from Brookhaven) through CONTIN particle size distribution –C(CH ) ), 1.20–0.59 (m, –CH backbone); FT-IR (cm ): 3370,
analysis routines. All determinations were made in triplicate for 3025, 2887, 1716, 1602, 1496, 1451, 1342, 1279, 1243, 1146,
This journal is © The Royal Society of Chemistry 2014
J. Mater. Chem. B, 2014, 2, 6652–6659 | 6653

View Article Online
Journal of Materials Chemistry B
Paper
1
105, 961, 841; M
n
(NMR) ¼ 11 250; M
n
(GPC) ¼ 9280, M
w
¼
Green (OG) and Rhodamine (RhB), uorescence intensity (I)
ratios (IOG + IFA)/IRhB were calculated. These uorescence
11 800, PD (M /M ) ¼ 1.27.
w
n
PEG -b-PAEMA -b-PS (2). The triblock copolymer, PEG- intensity ratios were then plotted against corresponding pH to
1
27
12
28
b-PAEMA(Boc)-b-PS (1.0 g, 0.088 mmol) was dissolved in 4 mL of obtain the ratiometric pH calibration curves. The triple labelled
DCM, followed by 4 mL TFA added drop wise, and the mixture nanosensor's calibration curve was tted by the following
11
was stirred at room temperature for 10 h. Aer evaporating equation,
most of the solvent under reduced pressure, the polymer was
precipitated into excess cold diethyl ether and dried under
vacuum. The complete de-protection of the amino group was
R
1
R
pH
2
R ¼ 1
þ 10pKa2-
þ R
0
0pKa1
-pH
þ 1
þ 1
where, R is the ratio of emission intensities of the nanosensor
excited at 488 (OG and FA) and 543 nm (RhB). pKa1 and pKa2
describes the specic pK values of the two pH-sensitive uo-
a
conrmed by the disappearance of the Boc group signal at d
1
6
1
.36 (s, –C(CH ) ) H-NMR (250 MHz in d-DMSO) and quali-
3 3
25
1
6
tatively by conducting a Kaiser test. H-NMR (250 MHz, d-
CH NH), 3.50 (s,
NH), 2.10–1.60 (br, CH and
CH backbone), 1.57–1.25 (br, –NH ), 1.20–0.59 (m, –CH
rophores (OG and FA) conjugated to the micelle. Rmin ¼ R is the
0
DMSO): d ¼ 7.3–6.3 (m, ArH), 3.73 (br, –OCH
2
2
ratio for the fully protonated sensor urophors and Rmax ¼ R +
0
–
–
CH
2
CH
2
O), 3.07 (br, –OCH
2
CH
2
2
R1 + R2 are the ratio for the fully deprotonated sensor uo-
rophores. More details are given in the ESI† (see calibration
curve tting).
2
3
ꢀ
1
backbone); FT-IR (cm ): 3439, 3021, 2910, 1733, 1692, 1602,
494, 1453, 1350, 1277, 1245, 1150, 1109, 953, 843. A GPC
1
chromatogram of the polymer is given in ESI Fig. S8† and
compared to starting materials and synthesis intermediates.
Intracellular pH measurements
HeLa cells were maintained in DMEM supplemented with 10%
ꢀ
1
Shell cross-linked micelle nanosensor (SCMN) 5 synthesis
fetal bovine serum and 100 UI mL penicillin and strepto-
mycin. Cell cultures were incubated in a 5% CO
incubator at 37 C. Cells were seeded on 9 mm cover glasses in
2
humidied
The deprotected amphiphilic triblock copolymer PEG127-b-
ꢂ
PAEMA -b-PS (2) (100 mg, 0.0088 mmol) was dissolved in 10
1
2
28
4
2
4 well plates at a density of 2 ꢁ 10 HeLa cells per well the day
mL DMF by stirring overnight. To the polymeric solution under
stirring, 2 mL of MilliQ water was added dropwise within 30
minutes followed by 20 mL more MilliQ water added dropwise.
The cloudy micelle solution was transferred into a dialysis tube
of molecular weight cut off (MWCO ¼ 12 kDa) and dialyzed
against MilliQ water for 5 days. Number-averaged hydrody-
before treatment. Cells were incubated in the presence of 10 mg
ꢀ1
ꢂ
mL nanoparticle for 20 hours at 37 C. Cells were then washed
three times with ice-cold phosphate buffered saline (PBS) sup-
plemented with heparin (20 units per mL), once with PBS and
kept in growth medium without phenol red and bicarbonate but
supplemented with 30 mM HEPES for control of pH without
h
namic diameters (D ) and zeta potentials (x) were found to be 29
2
CO incubation for observation by confocal microscopy. Cells
ꢃ 2 nm and +29 ꢃ 1 mV, respectively. The nal block copolymer
ꢀ
1
were either imaged immediately or treated with 100 nM Ba-
lomycin A1 for 45 min. before imaging. The images were
captured by a Leica TCS SP5 confocal microscope with a 63ꢁ
water-immersion objective (Leica Microsystems, Germany). A
calibration turve was obtained on the same day of the
measurements by diluting the nanosensor in buffers (20 mM
phosphate/20 mM borate/20 mM acetate/100 mM NaCl) with
different pH to a nal nanosensor concentration of 1.25 mg
concentration aer dialysis was 2 mg mL . To the micelle
solution (25 mL (50 mg, 0.0044 mmol, 0.053 mmol of amine)),
3
(
,6,9-trioxaundecandioic acid (3.6 mg, 0.016 mmol) and 1-[3-
dimethylamino)propyl]-3-ethylcarbodiimide methiodide (9.5
mg, 0.032 mmol) was added and stirred at room temperature for
0 h. The reaction mixture was then dialyzed against MilliQ
water for 3 days (D
(number-averaged) ¼ 24 ꢃ 1 nm and x ¼ +18
ꢃ 2), then against 0.1 M carbonate buffer (pH ¼ 9.6) for another
days. To 3 mL of the basic amphiphilic colloidal dispersion,
1
h
ꢀ1
mL . Images of nanosensor in buffer were acquired at the
3
microscope with the same settings as for the cell measure-
ments. Image analysis was performed as described previously
with a pixel based method resulting in images with pixels
colored according to a pH scale and frequency histograms of pH
for the different treatments.
RhBITC (0.014 mg, 0.0264 mmol), FITC (0.026 mg, 0.066 mmol)
and OGITC (0.028 mg, 0.066 mmol) were added and stirred in
the absence of light at room temperature for 12 h. The reaction
mixture was transferred into a dialysis tube (MWCO ¼ 12 kDa)
and dialysis was conducted against 0.1 M carbonate buffer for 3
11
h
days, then against MilliQ water for another 5 days (D (Number-
Results and discussion
Synthesis and characterization of the shell cross-linked
averaged) ¼ 25 ꢃ 2 nm and x ¼ 16 ꢃ 2 mV). The nal nano-
sensor concentration was 1.9 mg of polymers per mL of water.
micelle nanosensor (SCMN)
pH calibration curve from spectrouorometer
Fig. 1 illustrates the design strategy of the ratiometric
The pH calibration curve was constructed by uorescence nanosensor that has been developed. The sensor design is
measurements. 50 ml of the SCMN were added to 1 mL of buffer based on a hydrophobic styrene core, a functionalizable
solutions having different pH. The nanosensor solutions were shell layer and a PEG corona. The amphiphilic triblock
excited at 488 and 543 nm, successively. Fluorescence emission copolymer, poly(ethylene glycol)127-b-poly(2-[N-(tert-butox-
spectra of the nanosensors at different pH were plotted. From ycarbonyl)-amino]ethyl methacrylate)12-b-poly(styrene)28 (PEG127-b-
the uorescence emission spectra of Fluorescein (FA), Oregon PAEMA(Boc)12-b-PS28
)
(1) was synthesized by isolated
6654 | J. Mater. Chem. B, 2014, 2, 6652–6659
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
Journal of Materials Chemistry B
Table 1 Molecular weights of amphiphilic triblock copolymer, PEG127
-
b-PAEMA12(Boc)-b-PS28 (1)
a
a
w
a
n
b
c
c
c
M
n
M
M
w
/M
M
n
n
m
p
9280
11 800
1.27
11 250
127
12
28
a
b
c
Determined by GPC. Determined by NMR. Number of repeating
units of PEG block (n), PAEMA(Boc) block (m) and PS block (p).
Fig. 1 Schematic representation of shell cross-linked core–shell–
corona micelle pH nanosensor synthesis. AEMA(Boc) ¼ 2-[N-(tert-
butoxycarbonyl)amino]ethyl methacrylate.
macroinitiator atom transfer radical polymerization (Scheme 1)
and the triblock copolymer was characterized by NMR and
1
FT-IR spectroscopy. From the H-NMR data (ESI Fig. S1†), the
number-average degree of polymerization (DP ) was deter-
n
mined. GPC analysis was used to calculate the polydispersity
index (M /M ) of the amphiphilic triblock copolymer. The NMR
w
n
and GPC results are summarized in Table 1. The Boc groups
from PEG-b-PAEMA(Boc)-b-PS were deprotected by TFA in DCM
and the complete deprotection was conrmed by the disap-
3 3
pearance of the Boc group signal at d 1.36 ppm (s, –C(CH ) )
H-NMR (250 MHz in d DMSO) (ESI Fig. S2†).
The stepwise synthesis of the SCMN is shown in Scheme 2.
1
6
Scheme 2 Synthesis of shell cross-linked ratiometric micelle nano-
sensor. OGITC ¼ Oregon Green isothiocyanate, FITC ¼ Fluorescein
isothiocyanate, and RhBITC ¼ Rhodamine B isothiocyanate.
First, the entropy driven self-assembly of PEG127-b-PAEMA -b-
12
PS28 (2) in water was achieved by dissolving the triblock copoly-
mer in DMF that was slowly displaced by addition of excess water.
The resulting polymeric micelle solution was then dialyzed
against MilliQ water to remove the DMF. Stabilization of the
resulting PEG127-b-PAEMA12-b-PS28 micelle (3) was achieved by
covalent cross-linking between the unimers via cross-linking
the shell region using an amidation reaction where water
soluble 3,6,9-trioxaundecandioic acid was used as a cross-
linking agent. The diacid cross-linker was rst converted into
o-acylisourea by treatment with 1-[3-(dimethylamino)propyl]-
was then reacted with the free amino groups present at the
shell layer (PAEMA shell) of the polymer micelle. Stoichio-
metrically, 60% of the amino groups were used for the cross-
linking reactions. The shell cross-linked micelles were rst
dialyzed against MilliQ water to remove the urea by-products
and then against 0.1 M carbonate buffer of pH 9.6. The cross-
linked micelle 4 in basic buffer was converted into ratiometric
pH nanosensor 5 by treatment with suitable ratios of pH
sensitive Fluorescein isothiocyanate (FITC), Oregon Green
isothiocyanate (OGITC), and pH insensitive reference uo-
rophore, Rhodamine B isothiocyanate (RhBITC). Some of the
free amino groups present at the shell layer of the cross-linked
micelle were used for nucleophilic attack of the central elec-
trophilic carbon atoms of the pH sensitive and reference iso-
thiocyanates. Excess un-reacted free dyes were removed by
dialysis against carbonate buffer (0.1 M carbonate buffer (pH
3
-ethylcarbodiimide methiodide (EDC$MeI). This active ester
¼
9.6)) followed by dialysis against MilliQ water, which gave a
pH nanosensor dispersion in MilliQ water.
The morphology and size distribution of the micelle nano-
sensor was investigated by atomic force microscopy (AFM)
under ambient conditions (Fig. 2). AFM images of the air dried
nanosensor shows that the particles are round and relatively
uniform in size. The spherical morphology can be observed in
the phase contrast AFM image (Fig. 2b), which depicts the phase
shi of the cantilever oscillation and visualizes differences in
material properties of the sample. The AFM image also indi-
cates low/no intermicellar cross-linking or aggregation of the
nanosensor.
Scheme 1 Synthesis of PEG127-b-PAEMA12(Boc)-b-PS28 (1) by ATRP
0
and PEG127-b-PAEMA12-b-PS28 (2) by deprotection of 1. bpy ¼ 2,2 -
0
00
0
bipyridyl, PMDETA ¼ N,N,N ,N ,N -pentamethyldiethylenetriamine.
This journal is © The Royal Society of Chemistry 2014
J. Mater. Chem. B, 2014, 2, 6652–6659 | 6655

View Article Online
Journal of Materials Chemistry B
Paper
Fig.
2 Representative AFM images of shell cross-linked micelle
nanosensor. (a) Topographic image and (b) phase contrast image.
Scale bar 1 mm.
Characterizations of the shell cross-linked core–shell–corona
1
micelles by H-NMR was not possible; in D
2
O the hydrophobic
styrene core of the lyophilized micelle was not detectable, and
6
NMR spectrum of the particles in d-DMSO was inconclusive
(ESI Fig. S3†). A FT-IR spectra of the lyophilized micelle aer
cross-linking showed additional overlapping amide I and II
ꢀ1
bands appearing at approximately 1640–1550 cm . This
conrmed the amide bond formation at the shell region of the
micelle (ESI Fig. S4†).
The inner shell stabilization was not expected to consider-
ably affect the outer corona mobility. Since the inner shell does
not entropically impair the stabilizing capability of the hydro-
philic PEG domains, an exclusive intramicellar cross-linking
was anticipated. The intramicellar cross-linking was supported
by dynamic light scattering (DLS) and zeta potential (x)
measurements. DLS and x-potential measurements showed a
h
small decrease in hydrodynamic diameter (D ) and a substan-
tial decrease in x-potential aer shell cross-linking (Table 2).
The decrease in charge density also conrmed the participation
of amino groups of the shell domain in amidation cross-linking.
Due to the low concentration of uorophores in the nano-
sensor, the binding of uorophores to the cross-linked micelle
did not show any signicant change in hydrodynamic diameter
and x-potential.
DLS and size exclusion chromatography (SEC) further
conrms the micelle cross-linking (Fig. 3). DLS measurements
shows that cross-linked micelle 4 has a slightly lower particle
size distribution than non-cross-linked micelle 3 (Fig. 3a),
ꢂ
which indicates the absence of intermicellar crosslinking or Fig. 3 (a) Particle size distribution by DLS in MilliQ water at 25 C, (b)
size exclusion chromatogram using DMF with 50 mM LiCl solution as
aggregation aer cross-linking. Furthermore, lyophilized cross-
ꢀ1
ꢂ
eluent (0.5 mL min ) at 25 C and (c) variation of number-averaged
hydrodynamic diameter with respect to dilution/concentration of the
non-cross-linked micelle 3 and cross-linked micelle 4, respectively.
Table 2 DLS and Zeta potential measurements of nanoparticles at
ꢂ
2
5 C in MilliQ water
linked and non-cross-linked micelles had different solubility in
a
b
Particles
DLS (D
h
) (nm)
Zeta (x) (mV) THF. The non-cross-linked micelle 3 was readily soluble in THF,
whereas the cross-linked micelle 4 was sparingly soluble only by
Micelle before cross-linking (3)
Micelle aer cross-linking (4)
pH nanosensor (5)
50 ꢃ 2
45 ꢃ 1
46 ꢃ 2
29 ꢃ 1
18 ꢃ 2
16 ꢃ 2
ꢂ
ultrasonication for 1 h at 60 C. SEC studies of these solutions
showed different elution time for cross-linked and non-cross-
linked micelles (Fig. 3b), which indicates a higher molecular
weight for cross-linked micelle due to successful cross-linking
between the unimers. Micelle cross-linking was also conrmed
a
Intensity-averaged hydrodynamic diameters of nanoparticles by
b
dynamic light scattering.
measurements, each having 10 cycles.
Zeta potential average from 10
6656 | J. Mater. Chem. B, 2014, 2, 6652–6659
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
Journal of Materials Chemistry B
by DLS measurement by investigating the variations of number- and IRhB represents the maximum emission intensity of the
averaged hydrodynamic diameters of the micelle 3 and 4 as a nanosensor excited at 543 nm. Fluorescence intensity ratios (IOG
function of dilution in MilliQ water (Fig. 3c). The micelle 3 and 4 + IFA)/IRhB were plotted against pH to obtain a pH calibration
ꢀ1
were diluted with water from a concentration of 0.125 mg mL
curve (Fig. 4b). The tted calibration data shows that the sensor
to a concentration of 0.0039 mg mL and number-averaged is sensitive in an interval between pH 3.5 and 7.5 with dual pK
hydrodynamic diameters were measured. The measurement values (pKa1 ¼ 4.3, pKa2 ¼ 6.6, which are obtained as tting
showed that the measurements of cross-linked micelle 4 were parameters). This indicates that the uorophore pK values can
ꢀ1
a
a
independent of dilution whereas the non-cross-linked micelle 3 change due to the local chemical environment in the sensor as
8
was concentration dependent. The disintegration of non-cross- previously described, which can expand the sensitivity range
linked micelle into unimers occurred at a critical micelle compared to using free dyes. In the present case the local
ꢀ
1
concentration (cmc) of approximately 0.010 mg mL
.
chemical environment where the uorophores are conjugated
For sensor synthesis, two pH sensitive dyes and one refer- resulted in a change in the pK values of the uorophores and a
a
ence dye was chosen to expand the measurement range of the broadening of the nanosensor sensitivity range. The nano-
sensor and cover the full pH range inside cells. The choice of the sensor can thereby cover a pH range from 3.5–7.5.
pH sensitive dyes, oregon green and uorescein, was based on
their respective pK values of 4.7 and 6.4. These two dyes are uorescein and rhodamine B conjugated in the shell region of
thereby well suited to cover a pH range from approximately pH the micelle. As expected this sensor covered a much smaller pH
–7.5 when combined. As a pH insensitive reference dye we interval from approximately 5.8 to 7.8 (ESI Fig. S6†).
used rhodamine B. The usability of these dyes were tested The response time and reversibility of the pH nanosensor
For comparison we also synthesized a sensor with only
a
4
before sensor synthesis by a uorometer study where a cali- was also tested by uorescence measurements (Fig. 5a and b).
bration curve was formed by mixing the dyes in buffers of This shows the nanosensors respond quickly (in less than a
different pH. This study showed that the dyes can cover a pH second) towards a change in pH and are reversible between two
range of approximately pH 4.1–7.7 (ESI Fig. S5†) when free in pH's within the pH range.
solution and excited at (lex) 488 nm and 543 nm.
Conjugation of the pH sensitive dyes and reference dye to the
shell region of the cross-linked micelle gave the nanosensor. A
calibration curve was formed by adding the nanosensor to
buffer solutions of different pH. The sensor was excited at (lex)
Intracellular pH measurements
The sensor is spontaneously taken up by the HeLa cells into
small compartments probably through endocytosis due to the
cationic surface chemistry of the sensor. Aer 20 hours, punc-
4
88 nm and 543 nm on a uorometer and emission scans were
tuate structures in the cytosol mainly around the nucleus are
labelled. Fig. 6a show representative pH images of two cells, one
without treatment (ꢀBaf) and one aer treatment with balo-
recorded. From the uorescence emission spectra of the
nanosensor (Fig. 4a) uorescence intensity maxima (I) of pH
sensitive uorophores (IOG + IFA) and reference uorophore
mycin A (+Baf). Cells were imaged by sequential scanning of
(I
RhB) were determined. IOG + IFA represents the maximum
1
the green pH sensitive uorophores and the red reference u-
orophore. The intensity ratio of green to red for each pixel was
converted to pH via the calibration curve (ESI Fig. S7†) and the
single pixels were thus colored according to a linear pH scale.
The low pH of approx. 4.3 in the untreated cells and the punc-
tuate pattern indicates that the sensor resides in the lysosomes.
emission intensity of the nanosensor when excited at 488 nm
Fig. 4 (a) Fluorescence emission curve for the pH nanosensor. (b) pH
calibration curve of the nanosensor made by plotting fluorescence
intensity ratio (IOG + IFA/IRhB) against pH. OG ¼ Oregon Green, FA ¼
Fluorescein, RhB ¼ Rhodamine B. The calibration data was used to find Fig. 5 (a) Response time of the nanosensor towards changes in pH. (b)
the best fit sigmoid curve using the formula, R ¼ R1/(10^(pKa1-pH) + 1) Reversibility of the nanosensor, tested by repeatedly measuring the
+
R2/(10^(pKa2-pH) + 1) + R0.
fluorescence intensity ratio (IOG + IFA/IRhB) between two different pH's.
This journal is © The Royal Society of Chemistry 2014
J. Mater. Chem. B, 2014, 2, 6652–6659 | 6657

View Article Online
Journal of Materials Chemistry B
Paper
hydrophobic interactions between the unimers of the micelle
was provided by covalent cross-linking using an amidation
reaction. The stabilized micelles were converted to ratiometric
pH nanosensors by uorophore conjugation at the shell region.
An In vitro pH calibration curve for the nanosensor was con-
structed by uorescence measurements, which showed a broad
pH sensitivity range of the triple-labelled nanosensors. Fluo-
rescence measurements also conrmed the fast response and
reversibility of these pH nanosensors. Cell uptake experiments
of the nanosensors were performed and pH distributions of
HeLa cells were monitored. This showed that the sensor was
able to pass the cell membrane by cellular internalization and
respond to pH changes within the endosome-lysosome cellular
environment. This design of triple uorophore labelled sensors
enable ratiometric pH measurements within the whole endo-
cytic pathway in cells.
Acknowledgements
The authors would like to thank Kræens Bekæmpelse and the
Danish Research Council for Technology and Production for

nancial support.
References
1
D. Lagadic-Gossmann, L. Huc and V. Lecureur, Cell Death
Differ., 2004, 11, 953–961.
Fig. 6 HeLa cells were treated with nanosensor for 20 h, then washed
and either imaged immediately or treated with 100 nM Bafilomycin A1
(
Baf) for 45 min. before imaging. (A) pH images before and after
2 S. T. Whitten, E. B. Garc ´ı a-Moreno and V. J. Hilser, Proc. Natl.
Acad. Sci. U. S. A., 2005, 102, 4282–4287.
treatment with Bafilomycin A1. The ratio of green to red in every pixel is
converted to pH via the calibration curve and color coded on a linear
scale according to pH. The N indicates the location of the nucleus.
Scale bar, 10 mm. (B) Histogram showing pH distributions by pixel
intensity quantification in approximately 1000 individual endosomes/
lysosomes based on the type of images presented in (A). Mean ꢃ SEM
3
J. R. Casey, S. Grinstein and J. Orlowski, Nat. Rev. Mol. Cell
Biol., 2010, 11, 50–61.
4
A. Asokan and M. J. Cho, J. Pharm. Sci., 2002, 91, 903–913.
5 T. L. Andresen, S. S. Jensen and K. Jørgensen, Prog. Lipid Res.,
2005, 44, 68–97.
(
standard error of the mean)(n ¼ 9 images).
6
H. A. Clark, R. Kopelman, R. Tjalkens and M. A. Philbert,
Anal. Chem., 1999, 71, 4837–4843.
Aer the treatment the color has clearly changed from yellow/
orange to blue/green indicating an increase in pH. As balo-
7 J. Ji, N. Rosenzweig, C. Griffin and Z. Rosenzweig, Anal.
Chem., 2000, 72, 3497–3503.
mycin A
1
inhibits the proton pump that acidies lysosomes and
8 H. Sun, T. L. Andresen, R. V. Benjaminsen and K. Almdal, J.
Biomed. Nanotechnol., 2009, 5, 676–682.
9 H. Sun, K. Almdal and T. L. Andresen, Chem. Commun., 2011,
47, 5268–5270.
thus increases the pH of the lysosomes, which is also supportive
of lysosomal localization. The pH histograms presented in
Fig. 6b gives a mean pH of 4.3 before treatment and 6.3 aer the
treatment conrming what is observed in the images. From 10 J. Peng, X. He, K. Wang, W. Tan, Y. Wang and Y. Liu, Anal.
these pH distributions, it is clear that the sensor is well suited to
Bioanal. Chem., 2007, 388, 645–654.
measure pH in the endosome-lysosome pathway as only 2% of 11 R. V. Benjaminsen, H. Sun, J. R. Henriksen,
measurements of the ratios obtained from the images of the
balomycin treated sample falls outside the sensor's range.
N. M. Christensen, K. Almdal and T. L. Andresen, ACS
Nano, 2011, 5, 5864–5873.
1
2 V. M. Chauhan, G. Orsi, A. Brown, D. I. Pritchard and
J. W. Aylott, ACS Nano, 2013, 7, 5577–5587.
3 S. M. Moghimi, A. C. Hunter and J. C. Murray, Pharm. Rev.,
2001, 53, 283–318.
Conclusions
1
We have demonstrated the rst synthesis, characterization and
utilization of a shell cross-linked core–shell–corona micelle 14 V. P. Torchilin, J. Controlled Release, 2001, 73, 137–172.
based ratiometric pH nanosensor (SCMN) for intracellular 15 E. K. Pramod Kumar, K. Almdal and T. L. Andresen, Chem.
measurements. ATRP and the principle of self-assembly were
used to prepare polymeric micelles having a well-dened core– 16 E. K. P. Kumar, L. N. Feldborg, K. Almdal and T. L. Andresen,
shell–corona morphology. The reinforcement of the weak Chem. Mater., 2013, 25, 1496–1501.
Commun., 2012, 48, 4776–4778.
6658 | J. Mater. Chem. B, 2014, 2, 6652–6659
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
Journal of Materials Chemistry B
17 K. Matyjaszewski and N. V. Tsarevsky, Nat. Chem., 2009, 1, 22 E. S. Read and S. P. Armes, Chem. Commun., 2007, 3021–
276–288.
3035.
1
1
8 L. Zhang and A. Eisenberg, Science, 1995, 268, 1728–1731.
9 L. Zhang, K. Yu and A. Eisenberg, Science, 1996, 272, 1777–
23 S. Liu, J. V. M. Weaver, Y. Tang, N. C. Billingham, S. P. Armes
and K. Tribe, Macromolecules, 2002, 35, 6121–6131.
24 M.-H. l. n. Dufresne and J.-C. Leroux, Pharm. Res., 2004, 21,
160–169.
25 E. Kaiser, R. L. Colescott, C. D. Bossinger and P. I. Cook,
Anal. Biochem., 1970, 34, 595–598.
1779.
20 H. Sun, R. V. Benjaminsen, K. Almdal and T. L. Andresen,
Bioconjugate Chem., 2012, 23, 2247–2255.
2
1 R. K. O'Reilly, C. J. Hawker and K. L. Wooley, Chem. Soc. Rev.,
2006, 35, 1068–1083.
This journal is © The Royal Society of Chemistry 2014
J. Mater. Chem. B, 2014, 2, 6652–6659 | 6659