Revisiting the complexation between DNA and polyethylenimine-when and where-S-S-linked PEI is cleaved inside the cell

Article abstract of DOI:10.1039/c4tb00031e

As a non-viral gene vector, long PEI chains are more effective but also more cytotoxic. To solve this problem, people have tried to use disulfide (-S-S-) to link short PEI chains into a long one to generate highly efficient and less cytotoxic gene vectors because -S-S- is degradable inside the cell. In order to investigate when and where -S-S- is cleaved during intracellular trafficking, we designed and synthesized rhodamine B labeled linear PEI chains (M_n ≈ 3 kg mol^-1) with one end modified with a mercapto-group so that they can be coupled together via one disulfide bond in the middle and one rhodamine B molecule on each side of the disulfide bond, where fluorescence is self-quenched because two rhodamine B molecules are closely linked together. The cleavage of the -S-S- bond separates the two rhodamine B molecules and enhances their fluorescent intensity. In addition, plasmid DNA was also modified with bodipy, a FRET donor of rhodamine B. Using this specially prepared PEI, we studied the intracellular trafficking of the PEI/DNA polyplexes by using flow cytometry and confocal laser scanning microscopy. Our results reveal that (1) DNA is gradually dissociated from the polyplex before the disulfide bond's cleavage; (2) some of polyplexes escaped from endosomes before reaching lysosomes; and (3) the disulfide bonds are mainly cleaved inside lysosomes at ～5 h post-transfection. the Partner Organisations 2014.

Full text of DOI:10.1039/c4tb00031e

Journal of
Materials Chemistry B
View Article Online
View Journal | View Issue
PAPER
Revisiting the complexation between DNA and
polyethylenimine – when and where –S–S– linked
PEI is cleaved inside the cell†
a
Cite this: J. Mater. Chem. B, 2014, 2,
3282
Yongzheng Ma and Chi Wuab
*
As a non-viral gene vector, long PEI chains are more effective but also more cytotoxic. To solve this
problem, people have tried to use disulfide (–S–S–) to link short PEI chains into a long one to generate
highly efficient and less cytotoxic gene vectors because –S–S– is degradable inside the cell. In order to
investigate when and where –S–S– is cleaved during intracellular trafficking, we designed and
synthesized rhodamine B labeled linear PEI chains (Mn z 3 kg molꢀ1) with one end modified with a
mercapto-group so that they can be coupled together via one disulfide bond in the middle and one
rhodamine B molecule on each side of the disulfide bond, where fluorescence is self-quenched because
two rhodamine B molecules are closely linked together. The cleavage of the –S–S– bond separates the
two rhodamine B molecules and enhances their fluorescent intensity. In addition, plasmid DNA was also
modified with bodipy, a FRET donor of rhodamine B. Using this specially prepared PEI, we studied the
intracellular trafficking of the PEI/DNA polyplexes by using flow cytometry and confocal laser scanning
microscopy. Our results reveal that (1) DNA is gradually dissociated from the polyplex before the disulfide
bond's cleavage; (2) some of polyplexes escaped from endosomes before reaching lysosomes; and (3)
the disulfide bonds are mainly cleaved inside lysosomes at ꢁ5 h post-transfection.
Received 6th January 2014
Accepted 1st March 2014
DOI: 10.1039/c4tb00031e
www.rsc.org/MaterialsB
vector have to overcome many obstacles.14,15 Some previous
studies have shown that following endocytosis, the PEI/DNA
1. Introduction
Gene therapy has been proposed to cure various diseases.1,2 To
achieve a therapeutic effect, an effective gene vector is needed.3
Due to their large packing size, low immunogenicity, and high
design exibility, different non-viral gene vectors,4–6 especially
cationic polymers, have been developed. In the past three
decades, hundreds, if not thousands, of synthetic cationic
polymers have been prepared and studied. Among them, poly-
ethylenimine (PEI) remains as one of the most efficient non-
viral gene vectors since it was introduced by Behr in 1995.7,8
Many efforts have been devoted to studying the origins of
PEI's high transfection efficiency.9–13 However, PEI chains have
different topologies (linear and branched) for a given molar
mass and different molar masses for a given chain morphology,
making studies complicated and less dened, especially when
in vitro cellular and in vivo animal tests are involved. To deliver a
therapeutic gene into a specic kind of cells, the gene and its
complexes (polyplexes) escaped during the endocytic pathway
via the “proton sponge effect”.16,17 Namely, partially protonated
PEI chains absorb more protons inside the endocytic vesicles
embedded with ATPase proton pumps, accompanied by an
inux of chloride counter ions, so that the higher osmotic
pressure accumulated inside would ultimately rupture the
endocytic vesicles. However, such an assumption could not
explain why longer PEI chains are better than shorter ones at
promoting gene transfection, because short chains, in prin-
ciple, would lead to higher osmotic pressure.
Godbey et al.18 tried to track the intracellular pathway of
uorescence-labeled polyplexes made of branched PEI inside
EA.hy 926 cells. Their results revealed that both free PEI chains
and PEI/DNA polyplexes could be found inside the nuclei. Up to
now, it is still unclear whether large polyplexes could be trans-
located from the cytosol into the nucleus, but their results did
show that PEI/DNA polyplexes can effectively escape during the
endocytic pathway. In addition, the condensation ability of PEI
might also play a key role.19
aDepartment of Chemistry, The Chinese University of Hong Kong, Shatin, N. T., Hong
Kong. E-mail: ericyongzheng@126.com
bThe Hefei National Laboratory of Physical Science at Microscale, Department of
Chemical Physics, University of Science and Technology of China, Hefei, Anhui,
230026, China
Recently, we did a systematic study of the complexation
between DNA and PEI with different chain topologies and
lengths, for the rst time, by using a combination of static and
dynamic laser light scattering.20 We revealed that when the N/P
ratio is higher than 6, most of the polyplexes contain only one
DNA chain and the initial PEI chain topology affects the nal
† Electronic supplementary information (ESI) available: Synthesis details for the
probe, the absorption and emission spectra of B-DNA and Rho-PEI, the
emission intensity of polyplexes and NMR spectra of the probe. See DOI:
10.1039/c4tb00031e
3282 | J. Mater. Chem. B, 2014, 2, 3282–3291
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
Journal of Materials Chemistry B
morphology of the polyplexes. Linear PEI chains lead to an rhodamine B. The FRET occurs once DNA is condensed by the
“onion-like” morphology so that abundant anionic proteins rhodamine B-modied PEI chains, which enable us to monitor
inside the cytosol can peal the cationic PEI chains on/in each the trafficking of the polyplexes and probe when and where the
polyplexes in a layer-by-layer fashion to release the DNA chain disulde bonds are cleaved inside the cell.
inside. Therefore, polyplexes made of linear PEI chains are less
stable than those made of branched PEI chains that act as a
“cross-linking” agent inside, so that releasing DNA from
branched PEI/DNA polyplexes is relatively more difficult.
2. Experimental
2.1 Materials and cell lines
Our results are also proved by the study of uorescence
resonance energy transfer (FRET). Itaka et al.19 monitored the
intracellular trafficking of PEI/DNA polyplexes and found that
DNA condensed with linear PEI chains can be quickly delivered
into the cytoplasm and easily released. Furthermore, we found
that apart from the condensation and dissociation of DNA,
uncomplexed PEI chains with a size larger than ꢁ15 nm which
are free in the solution mixture of DNA and PEI also play a
critical role in promoting gene transfection,21,22 leading to
speculation that it is those free chains that interact with signal
proteins to block inter-vesicular fusion so that the ingested PEI/
DNA polyplexes have less chance of being delivered/developed
into lysosomes and a higher probability of escape from
endosomes.
Previously, in order to reduce the cytotoxicity of long PEI
chains, disulde-connected PEI chains were used because the
cytosol is a redox environment.23 Many degradable PEI chains
have been designed, prepared and tested for gene delivery.24–27
However, most of these studies were focused on polymer
preparation. The exact roles that the disulde bonds play inside
the polyplexes are less studied. Lee et al.28 conrmed the
disulde bonds can be cleaved during intracellular trafficking.
Yang et al.29 evaluated disulde reduction during receptor-
mediated endocytosis by using FRET imaging and found that
the reduction of the disulde bonds occurred along the entire
folate-receptor endocytic pathway. However, the probes they
used were receptor-dependent small molecules, which might
lead to a reduction mechanism different from that of degrad-
able PEI chains inside the polyplexes. In addition, Austin et al.30
reported that recycling endosomes, late-endosomes, and lyso-
somes were not actually reducing but oxidizing, comparable
with the conditions in the endoplasmic reticulum. It is those
controversial and incomplete results that prompted us to
design the current study to nd when and where the –S–S–
bonds of the PEI chains inside the polyplexes are cleaved and
the exact roles of the –S–S– bonds apart from reducing the PEI's
cytotoxicity.
Linear and branched polyethylenimines (lPEI-25k and bPEI25k,
Mw ¼ 25 kg molꢀ1) were purchased from Polysciences and
Sigma-Aldrich, respectively, and used without further purica-
tion. Methyl p-toluenesulfonate from Sigma-Aldrich was
distilled and stored under nitrogen. 2-Ethyl-2-oxazoline was
purchased from Sigma-Aldrich and puried by vacuum distil-
lation over CaH2. Acetonitrile from RCI Labscan was dried over
CaH2 and distilled under dry nitrogen.
N,N-Diisopropylethylamine (DIEA), ethane-1,2-dithiol (EDT),
triisopropylsilane (TIS), piperidine, rhodamine B and Sephadex
G-50 were purchased from Sigma-Aldrich. 1-Ethyl-3-(3-dimethyl
aminopropyl) carbodiimide hydrochloride (EDC), N-hydroxy-
benzotriazole (HOBt), N-Fmoc-S-tritylcysteine (Fmoc-Cys(trt)-
OH), and glycine tert-butyl ester hydrochloride (H-Gly-otBu)
were obtained from GL Biochem Ltd (Shanghai, China). N,N-
Dimethylformamide (DMF) from RCI Labscan was dried over
CaH2 and puried by vacuum distillation. All other chemicals
and solvents were used as received without further purication.
Initial plasmid DNA (pGL3) with a SV40 promoter and an
enhancer sequence encoding rey luciferase was purchased
from Promega (Madison, USA). A large amount of this plasmid
DNA was made with a Qiagen Plasmid Maxi Kit (Qiagen, Ger-
many). Fetal bovine serum (FBS), penicillin–streptomycin and
Dulbecco's modied Eagle's medium (DMEM) were products of
GIBCO (NY, USA). 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenylte-
trazolium bromide (MTT) was purchased from Sigma-Aldrich. A
Label IT Amine Nucleic Acid Modifying Kit from Mirus was used
to modify the plasmid DNA with amino-groups. Lysosensor™
Green, Celllight early endosomes-GFP and bodipy FL NHS ester
were purchased from Invitrogen. Dextran sulfate (DS, Mw ¼ 10
kg molꢀ1) and tris(2-carboxyethyl)phosphine hydrochloride
(TCEP) were products of Sigma-Aldrich. HepG2 cells
were grown at 37 ꢂC, in 5% CO2 in DMEM supplemented
with 10% FBS, penicillin at 100 units per mL and streptomycin
at 100 mg mLꢀ1
.
2.2 Synthesis of linear polyethylenimine (lPEI), Rho-Cys-Gly-
PEI (Rho-PEI) and (Rho-Cys-Gly-PEI)2 ((Rho-PEI)2)
To address these questions, we designed and synthesized a
rhodamine B labeled linear PEI as a probe by coupling two short
linear PEI chains together with uorescent molecules and a
disulde linker. In order to eliminate the chain length effect, we
purposely prepared linear PEI with Mn z 3000 g molꢀ1 as the
initial short chain, which is already long enough to interfere
with the signal proteins involved in inter-vesicular fusion. In
our design, the uorescence of two rhodamine B molecules is
Complete synthesis details for linear polyethylenimine (lPEI),
Rho-Cys-Gly-PEI (Rho-PEI) and (Rho-Cys-Gly-PEI)2 ((Rho-PEI)2)
may be found in the ESI.† Scheme 1 summarizes our synthesis
procedure.
2.3 Characterization
self-quenched because they are linked by a short disulde A Bruker Advanced III 400 NMR spectrometer was used. The
bond.31–33 The cleavage of –S–S– separates the two rhodamine B chemical shis (d) were reported in ppm with the solvent
molecules and enhances the uorescent intensity. In addition, resonance as the internal standard relative to CDCl3 (d 7.26),
the plasmid DNA was modied with bodipy, a FRET donor of CD3OD (d 3.31) or D2O (d 4.79) for 1H. All the 1H NMR
This journal is © The Royal Society of Chemistry 2014
J. Mater. Chem. B, 2014, 2, 3282–3291 | 3283

View Article Online
Journal of Materials Chemistry B
Paper
was used. The MALLS detector used a GaAs laser (685 nm and 30
mW) and had 18 diodes placed at different angles, ranging from
22.5 to 147.0ꢂ. The data were analyzed using the ASTRA for
Windows soware (Ver. 4.90.07, Wyatt). For both SEC-RI and
SEC-MALLS measurements, 1 mL of polymer solution (2–8 mg
mLꢀ1, depending on the molar mass) was pre-ltered through a
0.2 mm PVDF lter before injection.
2.5 Laser light scattering (LLS)
A commercial LLS instrument (ALV5000) with a vertically
polarized 22 mV He–Ne laser (632.8 nm, Uniphase) was used to
detect the molecular weight of (Rho-PEI)2. In static LLS, we can
obtain the weight-averaged molar mass (Mw) and the z-averaged
root-mean square radius of gyration (hRgiz) of scattering objects
in a sufficiently dilute solution/dispersion from the angular and
concentration dependence of the excess absolute scattering
intensity (Rayleigh ratio, Rvv(q)) as KC/Rvv(q) z (1 + q2Rg /3)/Mw
2
where K ¼ 4p2(dn/dc)2/(NAlo4) and q ¼ (4pn/lo)sin(q/2), with dn/
dC, NA, lo, n and q being the specic refractive index increment,
the Avogadro constant number, the incident wavelength in
vacuum, the refractive index of solvent, and the scattering
angle, respectively. Note that we have neglected the concentra-
tion correction here. To measure the weight averaged molar
mass of (Rho-PEI)2, both Rho-PEI and (Rho-PEI)2 solutions were
prepared in a 30 mM NaCl solution and then claried by a 0.45
mm lter. The solutions were measured by dynamic LLS at a low
scattering angle of 20ꢂ. The weight averaged molar mass (Mw) of
(Rho-PEI)2 was calculated from MRho-PEI
ꢃ
(hI0i(Rho-PEI)
/
,
2
C(Rho-PEI) )/(hI0iRho-PEI/CRho-PEI), where CRho-PEI
,
C(Rho-PEI)
2
2
hI0iRho-PEI and hI0i(Rho-PEI) are the concentrations and the time
2
averaged scattering intensities of Rho-PEI and (Rho-PEI)2 at the
zero scattering angle, respectively.
In dynamic LLS, the Laplace inversion of each measured time
correlation function (G(2)(q, t)) can be related to a line-width
distribution G(G). For a diffusion relaxation, G is further related
to the translational diffusion coefficient D by (G/q2)c/0,q/0 ¼ D.
Therefore, G(G) can be converted into a translational diffusion
coefficient distribution G(D) or a hydrodynamic radius distribu-
tion f(Rh) using the Stokes–Einstein equation, Rh ¼ kBT/(6phD),
where kB, T, and h are the Boltzmann constant, the absolute
temperature, and the solvent viscosity, respectively.
Scheme 1 Schematic of the synthesis of probes (1) linear PEI; (2) Rho-
PEI; and (3) (Rho-PEI)2.
measurements were carried out at room temperature. Mass
spectra (ESI-MS) were obtained with a HP 5989B spectrometer
and determined at an ionizing voltage of 70 eV; the relevant data
are given as m/z.
2.6 Fluorescent labeling of plasmid DNA
Plasmid DNA was modied with amino-groups using a Label IT
Amine Nucleic Acid Modifying Kit at a 0.2 : 1 (mass label
reagent/mass nucleic acid) ratio according to the manufactur-
er's instructions. Amine modied DNA was diluted in 200 mM
2.4 Size exclusion chromatography (SEC)
ꢂ
MOPS and stored at ꢀ20 C. 80 mg of puried amine-modied
The relative number- and weight-averaged molar masses (Mn DNA was reacted with 90 mg bodipy FL NHS ester in DMSO (5 mL)
ꢂ
and Mw) were determined at 35 C by SEC (Waters 1515) with and 100 mM NaHCO3 (20 mL, pH ¼ 8.5, freshly prepared) for
three Waters Styragel columns (HR2, HR4, HR6), and two 12 hours at room temperature in the dark. To purify the
detectors: refractive index (RI, Wyatt WREX-02) and multi-angle labeled DNA, the solution was precipitated in ethanol, and
laser light scattering (MALLS, Wyatt DAWN EOS). THF was used washed extensively with 70% ethanol. The nal product bodipy-
as the eluent at a ow rate of 1.0 mL minꢀ1 and the conven- DNA (B-DNA) was dissolved in 1 ꢃ Tris–EDTA buffer and stored
ꢂ
tional universal calibration with linear polystyrene standards at ꢀ20 C.
3284 | J. Mater. Chem. B, 2014, 2, 3282–3291
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
Journal of Materials Chemistry B
F-4500 spectrouorometers, respectively. To detect B-DNA
released from the polyplexes, Rho-PEI/B-DNA, (Rho-PEI)2/
B-DNA and lPEI/B-DNA at N/P ¼ 3 were prepared. The nal
DNA concentration was kept constant (0.04 mg mLꢀ1 B-DNA in
PBS). Aer 5 min of incubation at room temperature, 2 mL DS
(0.1 g mLꢀ1) was added into 200 mL of each polyplex solution.
To study the cleavage of disulde bonds, Rho-PEI/DNA,
(Rho-PEI)2/DNA, Rho-PEI/B-DNA and (Rho-PEI)2/B-DNA
were prepared at N/P ¼ 3. 10 mL of TCEP (0.1 M) was added
into 200 mL of each polyplex solution. The uorescence
intensity of each solution was recorded by a spectrouo-
rometer in real time (excitation/emission: 488/512, 488/580
and 543/580).
2.7 Cytotoxicity of PEI vectors
The MTT assay was used to test the cytotoxicity of each polymer
used. PEI solutions with different concentrations were prepared
in DMEM. HepG2 cells were seeded in 96-well culture plates at
104 cells per well in 100 mL DMEM medium containing 10% FBS
and antibiotics. 24 h later the PEI solution under study was
added to each well. The treated cells were incubated in a
humidied environment with 5% CO2 at 37 ꢂC for 48 h. The
MTT reagent (in 20 mL PBS, 5 mg mLꢀ1) was added to each well
ꢂ
and further incubated at 37 C for 4 h. The medium was then
removed and replaced with 100 mL of DMSO. The plate was
gently agitated for 30 min before the absorbance (A) at 490 nm
was recorded by a microplate reader (Bio-rad, USA). The cell
viability was calculated as viability ¼ (Atreated ꢀ Ablank)/(Acontrol
ꢀ
2.11 Flow cytometry
Ablank) ꢃ 100%, where Atreated, Ablank and Acontrol are the absor-
bance of the cells treated with PEI, fresh culture medium and
without any treatment, respectively. Each experiment condition
was done in triplicate. The data is shown as the mean value plus
standard deviation (ꢄSD).
For the ow cytometry experiments, HepG2 cells were seeded
in a 12-well plate at an initial density of 1.2 ꢃ 105 cells per
well. Aer 24 h, uorescence-labeled PEI/DNA polyplexes (N/P
¼ 3) with an additional 7-fold free unlabeled lPEI chains were
diluted in complete DMEM (500 mL) and then administered
to the cells at a nal concentration of 1.6 mg DNA per well.
Aer 4 h, the transfection medium was removed and fresh
DMEM was supplied. The cells were incubated at 37 ꢂC and
harvested aer different desired incubation times. The har-
vested cells were briey rinsed twice with PBS, and further
detached by 0.05% trypsin–EDTA. Finally, the cells were
washed twice with PBS, then resuspended in 4% para-
formaldehyde and stored at 4 ꢂC. The uorescence intensity
of cells was recorded by using a FC 500 ow cytometry system
(Beckman Coulter, USA). For each sample, 104 gated events
were collected. The uorophore was excited at 488 nm and
detected at 525BP (515 nm–535 nm) and 575BP (567.5 nm–
582.5 nm). The uorescence intensity was displayed using a
logarithmic scale and analyzed with tetraCXP soware
(Beckman Coulter).
2.8 Gel retardation assay of DNA complexation
The PEI/DNA polyplexes with different desired N/P ratios (the
ratio of nitrogen atoms on PEI to phosphates on plasmid DNA)
were prepared by adding an appropriate amount of PEI (5 mL) to
0.4 mg DNA (5 mL) in PBS. The resultant dispersions of poly-
plexes were incubated at room temperature for 5 min and
loaded on a 0.8% (w/v) agarose gel containing ethidium
bromide in TAE (Tris–acetate) buffer. Gel electrophoresis was
carried out at 100 V. Aer 1 h, the DNA bands in the gel were
visualized with a UV (254 nm) illuminator and photographed
with a Vilber Lourmat imaging system.
2.9 In vitro gene transfection
The gene transfection experiments were performed in HepG2
cells using the plasmid pGL3 as an exogenous reporter gene.
Gene transfection was conducted using PEI/DNA polyplexes
with different N/P ratios (10 : 1, 15 : 1, 20 : 1, 25 : 1 and 30 : 1).
The cells were seeded in a 48 well plate at an initial density of 3
ꢃ 104 per well for 24 h before the gene transfection experiments
were carried out. Each PEI/DNA polyplex dispersion with a given
N/P ratio was diluted in complete DMEM medium (10% FBS,
100 units per mL penicillin and 100 mg mLꢀ1 streptomycin) (200
mL), and then administered to the cells at a nal concentration
of 0.4 mg DNA per well. Aer 4 h, the transfection medium was
removed and fresh DMEM medium (500 mL per well) was added.
The cells were further cultured for 48 h before the transgene
expression level was evaluated. A GloMax 96 microplate
luminometer (Promega, USA) and the Bio-Rad protein assay
reagent were used to determine the transfection efficiency of the
PEI/DNA polyplexes, which is expressed as a relative lumines-
cence unit (RLU) per cellular protein (mean ꢄ SD of triplicates).
2.12 Confocal laser scanning microscopy
3 ꢃ 105 HepG2 cells were cultured in a m-Dish 35 mm (ibidi
GmbH, Germany) for 24 h before the transfection experiment
was carried out (optionally, 30 mL Celllight early endosomes-
GFP was administered for 12 h before the transfection
experiment was carried out). Fluorescence-labeled PEI/DNA
polyplexes (N/P ¼ 3) with an additional 7-fold free unlabeled
lPEI chains were diluted in complete DMEM (1 mL) and then
administered to the cells at a nal concentration of 3.2 mg
DNA per dish. Aer 4 h, the transfection medium was
removed. (Optionally, 1 mL Lysosensor™ Green diluted in
1 mL DMEM was added to the cells to stain the lysosome for
15 min.) Aer that, the cells were washed twice with PBS and
supplied with fresh DMEM. Live cell imaging was performed
for 20 h using a Nikon C1si confocal laser scanning micro-
scope equipped with a spectral imaging detector (Nikon,
Japan) and an INU stage-top incubator (Tokai Hit, Japan).
Image sequences were captured at approximately 1 h intervals.
2.10 Fluorometric analysis
The absorption and emission spectra of uorescence-labeled Bodipy, Celllight early endosomes-GFP and Lysosensor™
samples were recorded on Gary 5G UV/Vis/NIR and Hitachi Green were visualized by 488 nm excitation and the
This journal is © The Royal Society of Chemistry 2014
J. Mater. Chem. B, 2014, 2, 3282–3291 | 3285

View Article Online
Journal of Materials Chemistry B
Paper
rhodamine B was excited at 543 nm. The corresponding
emissions were detected at channels 515/30 and 605/75 nm,
respectively. All the images were analyzed using the Nikon EZ-
C1 soware.
3. Results and discussion
The GPC result of poly(2-ethyl-2-oxaline) (PETOZ) shows that
lPEI has a number-averaged molar mass of ꢁ3.2 kg molꢀ1
.
Aer it was labeled with rhodamine B, the integrals of
rhodamine B and lPEI chemical shis in the NMR spectra
reveal that on average only one rhodamine B was conjugated to
each PEI chain, as shown in Fig. 1. Therefore, the number
averaged molar mass of Rho-PEI is ꢁ3.8 kg molꢀ1. The
formation of a disulde bond between the thiol groups on two
Rho-PEI chains results in (Rho-PEI)2 with a number-averaged
molar mass of ꢁ7.6 kg molꢀ1. Consequently, two rhodamine B
molecules were symmetrically located at each end of a disul-
de bond, with reduced emission intensity due to the uo-
rescence self-quenching (Fig. S1†). Fig. 2 shows that when the
dimerized state (Rho-PEI)2 was reduced back to the monomeric
Rho-PEI state in vitro, the emission intensity increases ꢁ1.5
fold.
Fig. 3 shows that the cytotoxicity of both Rho-PEI and (Rho-
PEI)2 increases with the polymer concentration, where both
bPEI-25k and lPEI-25k are used as controls. Note that in a
normal concentration range of 10ꢀ3 to 10ꢀ2 mg mLꢀ1 used for
gene transfection, both Rho-PEI and (Rho-PEI)2 are less cyto-
toxic than bPEI-25k and lPEI-25k. The cytotoxicity of long
cationic chains has been attributed to the disruption of the
anionic membranes of the cell and different organelles so that
the cell undergoes necrosis.8 Also note that linking two Rho-PEI
chains together only slightly increases the cytotoxicity,
compared with single Rho-PEI chains.
Fig. 2 Fluorescence intensities of (Rho-PEI)2 and Rho-PEI in PBS
treated with tris(2-carboxyethyl)phosphine hydrochloride (TCEP, 0.1 M).
form a longer (Rho-PEI)2 chain helps the formation of the
PEI/DNA polyplexes. Note that in comparison with lPEI-25k
and bPEI-25k, a slightly higher concentration of Rho-PEI or
(Rho-PEI)2 is required to condense a given plasmid DNA
chain, because long cationic chains are more effective in
entropy-driven complexation and large rhodamine B moieties
sterically hinder the electrostatic interaction between PEI and
DNA.
Fig. 5 shows that Rho-PEI and (Rho-PEI)2 chains are 10–100
times more efficient for transfection than longer lPEI-25k and
bPEI-25k chains. Due to their high molecular weight, lPEI-25k
and bPEI-25k showed higher cytotoxicity above N/P ratios of
20. Consequently, the transfection efficiencies were decreased
compared to Rho-PEI and (Rho-PEI)2. Our previous studies
showed that most of the cationic polymer chains are actually
not complexed with DNA in a mixture of polymer and DNA,
especially when N/P > 6.20 We have conrmed that it is those
free PEI chains with a size larger than ꢁ15 nm that promote
gene transfection.21 This is why we purposely designed Rho-PEI
with a chain length longer than ꢁ15 nm. In addition, the
hydrophobic xanthene moiety of rhodamine B in Rho-PEI and
(Rho-PEI)2 interacts with cytoplasmic membranes, which may
Fig. 4 shows that both Rho-PEI and (Rho-PEI)2 are able to
effectively condense plasmid DNA when the N/P ratio reaches
ꢁ3, indicated by the disappearance of the free DNA strip.
Indeed, at N/P ꢁ 5, the (Rho-PEI)2/DNA polyplexes start to
migrate in the opposite direction. This indicates that using
the disulde bond to link two Rho-PEI chains together to
Fig. 3 Polymer concentration dependence of relative cell viability of
Rho-PEI, (Rho-PEI)2, lPEI-25k and bPEI-25k in HepG2 cells.
Fig. 1 1H NMR spectra of Rho-PEI in D2O.
3286 | J. Mater. Chem. B, 2014, 2, 3282–3291
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
Journal of Materials Chemistry B
Fig. 4 N/P ratio dependence gel retardation assay of (Rho-PEI)2/DNA
and Rho-PEI/DNA polyplexes, where naked DNA (N/P ¼ 0) was used as
a control.
Fig. 6 Time dependence of the replacement of DNA in polyplexes (N/
P ¼ 3) in PBS by anionic dextran sulfate (0.1 g mLꢀ1).
(acceptor) emission. In addition, we also used a reducing
reagent (TCEP) to cleave the disulde bond to test its effect on
the stability of the polyplexes.
Fig. 7 shows that aer treatment with the reducing reagent
TCEP, the intensity ratio of the uorescence emitted at two
different wavelengths by Rho-PEI/DNA remains constant, but
that of (Rho-PEI)2/B-DNA gradually decreases and approaches
that of Rho-PEI/DNA because the cleavage of the disulde bond
in each (Rho-PEI)2 chain separates two closely linked rhoda-
mine B molecules and removes the self-quenching so that the
uorescence emitted at 580 nm increases. The combination of
the above results demonstrates that (Rho-PEI)2 is an effective
probe for the evaluation of when and where the disulde bond
is cleaved as well as at which point DNA is released inside the
cell.
Fig. 5 N/P ratio dependence of in vitro gene transfection efficiency of
different PEI/DNA polyplexes in HepG2 cells in the presence of serum.
increase the uptake of polyplexes. Consequently, the trans-
fection efficiency is enhanced. Fig. 5 clearly shows that both
Rho-PEI and (Rho-PEI)2 are effective gene vectors, so we can
use them to probe the intracellular trafficking of the
polyplexes.
In order to avoid interference by uorescence generated by
uncomplexed free PEI chains in the study of the intracellular
trafficking, all the uorescence labeled PEI/DNA polyplexes
were prepared at N/P ¼ 3 with an additional 7-fold unlabeled
Since anionic plasmid DNA is labeled with bodipy (denoted
as B-DNA hereaer), a FRET donor of rhodamine B (Fig. S2†), its
complexation with rhodamine labeled cationic PEI results in a
uorescence intensity change (lPEI/B-DNA, Fig. S3 and S4;†
Rho-PEI/B-DNA, Fig. S5;† and (Rho-PEI)2/B-DNA, Fig. S6.†) All
the polyplexes were prepared at N/P ratio of 3. This is because
we want to avoid the interference of free rhodamine B-labeled
PEI chains in the solution when the N/P is higher. Besides, lots
of anionic proteins are present in the cytosol, which can replace
anionic DNA chains inside the polyplexes aer they enter into
the cell, leading to their destabilization. To test this destabili-
zation effect, we used anionic dextran sulfate chains to mimic
proteins.
Fig. 6 shows that aer the addition of dextran sulfate, the
intensity ratio of the uorescence emitted at two different
wavelengths by (Rho-PEI)2/B-DNA increases, but that of Rho-
PEI/B-DNA remains constant in the rst 2 h, and approaches a
plateau value aer ꢁ4 h. This shows that as anionic B-DNA is
replaced by anionic dextran sulfate and released from the
polyplexes, the FRET effect disappears, leading to an increase of
Fig. 7 Time dependence of the fluorescence intensity ratio (I512/I580
)
of (Rho-PEI)2/B-DNA and Rho-PEI/DNA polyplexes (N/P ¼ 3) in PBS
after treatment with tris(2-carboxyethyl)-phosphine hydrochloride
(TCEP, 0.1 M, a reducing reagent to cleave the –S–S– bond in (Rho-
the bodipy (donor) emission and a decrease of the rhodamine B PEI)2).
This journal is © The Royal Society of Chemistry 2014
J. Mater. Chem. B, 2014, 2, 3282–3291 | 3287

View Article Online
Journal of Materials Chemistry B
Paper
free lPEI, so that the nal N/P ratio remains 10.21,22 It should be
noted that the additional 7-fold of unlabeled lPEI chains do not
replace the labeled ones, which combine with DNA in the
polyplexes within our experimental time window because the
two oppositely charged macromolecular chains interact via
multi-point electrostatic interactions driven by the entropy gain
of counter ions.
Fig. 8a shows that aer endocytosis of the lPEI/B-DNA
polyplexes (a control, no FRET effect here), the emission
intensity, as expected, remains nearly constant. Comparatively,
the (Rho-PEI)2/B-DNA polyplexes initially emit less uorescent
light at 512 nm due to the FRET effect between rhodamine B
(the acceptor) and bodipy (the donor). The increase in the
uorescence emitted at 512 nm reects the de-condensation
and a gradual release of B-DNA from the polyplexes. In
contrast, Fig. 8b shows that the uorescence emitted by
rhodamine B only slightly increases ꢁ10 to 15% in the rst ꢁ5
h, and then gradually increases further. Fig. 8b suggests that
the disulde bonds of (Rho-PEI)2 inside the polyplexes can be
cleaved, i.e., the separation of the two initially closely linked
rhodamine B molecules at ꢁ5 h aer their endocytosis.
Furthermore, we carried out additional confocal laser scanning
microscopy (CLSM) studies to support the ow cytometry
results.
Fig. 9 Confocal laser scanning microscopy imagines of the intracel-
lular trafficking of (Rho-PEI)2/B-DNA polyplexes, where left: observed
at channels 515/30 nm (green) and 605/75 nm (red); middle: observed
at channel 515/30 nm (green); and right: observed at channel 605/75
Fig. 9 shows that the cells transfected by (Rho-PEI)2/DNA nm (red) (bar ¼ 50 mm).
polyplexes have yellow-colored spots in the cytoplasm. Two
hours aer the addition of the polyplexes, DNA is still
condensed, reected in the FRET signal, with some weak green
spots. Gradually, the intensity of the green light increases, but
the intensity of the red light remains nearly constant, indicating
the disappearance of FRET and the de-condensation/release of
B-DNA from the polyplexes. 19 hours later, the green light is
visible over a large area of the cytosol although its intensity per
unit area is still not high, but at the same time, the red light
becomes more visible, which is attributed to the cleavage of the
disulde bonds, i.e., the separation of the two closely linked
rhodamine B molecules. For comparison, the intracellular
trafficking of both the Rho-PEI/B-DNA and Rho-PEI/DNA poly-
plexes was also studied by ow cytometry and confocal laser
scanning microscopy.
Fig. 10 shows that during transfection, the emission
intensity of the Rho-PEI/B-DNA polyplexes at 512 nm contin-
uously increases, signaling a diminishing of the FRET effect
and the de-condensation/release of DNA from the polyplexes.
In contrast, the Rho-PEI/B-DNA and Rho-PEI/DNA polyplexes
emit a similar intensity at 580 nm. Since there is no disulde
bond in Rho-PEI, there is no –S–S– cleavage or self-quenching.
The ow cytometry results are consistent with the confocal
microscopy ones, as shown in Fig. 11. Namely, 2 hours aer
the addition of the polyplexes, the green light (bodipy) is weak
due to the FRET effect. As the gene transfection proceeds, the
green light intensity gradually increases, reecting the gradual
de-condensation/release of DNA from the Rho-PEI/B-DNA
polyplexes. It is worth-noting that aer 19 hours, the red light
intensity remains nearly constant and observable over a large
area of the cytosol, indicating that some polyplexes have
Fig. 8 Time dependence of the fluorescence emission intensity of
(Rho-PEI)2/B-DNA, (Rho-PEI)2/DNA and lPEI/B-DNA polyplexes
measured by flow cytometry, where lexcitation ¼ 488 nm.
3288 | J. Mater. Chem. B, 2014, 2, 3282–3291
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
Journal of Materials Chemistry B
already escaped from the endosomes and entered into the
cytosol.
The ow cytometry and CLSM results reveal that (1) the
release of DNA condensed by linear PEI inside the polyplexes
occurs continuously aer intracellular uptake; (2) there is no
visible uorescence inside the nucleus even aer 19 h, at which
the gene transcription has started, indicating that the number
of polyplexes entering the nucleus must be very low; and (3) the
cleavage of the disulde bonds occurs at about 5 h aer the
removal of the transfection reagent, notably, there is no strong
correlation between the cleavage of the disulde linkage and
the release of DNA from the polyplexes. In order to locate where
the disulde bonds are cleaved inside the cell, we labeled early
endosomes with Celllight early endosomes-GFP and lysosomes
with Lysosensor™ Green.
Fig. 12 and S11† show that initially most of polyplexes were
attached to the cellular membranes and endocytosis occurs
gradually. Note that the uorescence signals of rhodamine
and GFP are separated in the cytoplasm; as there are few
observable colocalized yellow spots, indicating the escape of
some polyplexes before they reach the early endosomes.
Fig. 10 Time dependence of the intracellular trafficking of Rho-PEI/B-
DNA, Rho-PEI/DNA and lPEI/B-DNA polyplexes measured by flow
cytometry, where lexcitation ¼ 488 nm.
Fig. 11 Confocal laser scanning microscopy imagines of the intra-
cellular trafficking of Rho-PEI/B-DNA polyplexes, where left: observed
at channels 515/30 nm (green) and 605/75 nm (red); middle: observed
at channel 515/30 nm (green); and right: observed at channel 605/75
nm (red) (bar ¼ 50 mm).
Fig. 12 Colocalization of (a) (Rho-PEI)2/DNA and (b) Rho-PEI/DNA
polyplexes with early endosomes labeled by Celllight early endo-
somes-GFP. In each case, the upper panel: observed at channel 515/
30 nm (green) and channel 605/75 nm (red); and lower panel:
observed at channel 605/75 nm (red) (bar ¼ 50 mm).
This journal is © The Royal Society of Chemistry 2014
J. Mater. Chem. B, 2014, 2, 3282–3291 | 3289

View Article Online
Journal of Materials Chemistry B
Paper
Therefore, it seems that the intracellular trafficking of some
polyplexes does not follow the normal early endosome–late
endosome–endolysosome pathway. In addition, Fig. 13a and
S12a† show that the colocalization of (Rho-PEI)2/DNA poly-
plexes and lysosomes appears aer 3 h, and subsequently
both red and yellow spots are observable inside the cells,
further indicating that some of the polyplexes have escaped
from the endosomes, reecting the red spots in the cytosol;
and the rest are accumulated in the lysosomes, reecting the
increased yellow light intensity.
Fig. 13a also shows the increase in intensity of the red light
in the lysosomes, attributed to the cleavage of the disulde
bond that links the two rhodamine B molecules together, i.e.,
self-dequenching. A similar colocalization between the poly-
plexes and lyososomes is also observed for Rho-PEI/DNA poly-
plexes, as shown in Fig. 13b and S12b.† As expected, the
Scheme 2 Schematic of when and where disulfide bonds inside PEI/
DNA polyplexes are cleaved and polyplexes escape from endosome
intensity of the red light in the lysosome remains the same even during intracellular trafficking.
aer 19 h because there is no disulde bond to be cleaved here.
The above results reveal that the disulde bonds in PEI chains
are cleaved in the lysosomes, although some polyplexes have
escaped from the endosomes before cleavage. Scheme 2
summarizes the intracellular trafficking of degradable PEI/DNA
polyplexes.
4. Conclusions
Coupling two short linear PEI chains together with rhodamine
B and a disulde linker leads to an effective uorescent probe,
which can be applied to trace the intracellular trafficking of
ingested polyplexes. This probe has higher transfection effi-
ciency and lower cytotoxicity compared to commercial PEIs.
Labeling plasmid DNA with bodipy, a FRET donor of rhodamine
B, further enables us to analyze FRET and self-dequenching of
the polyplexes inside the HepG2 cell. We found that soon aer
endocytosis, DNA is continuously de-condensed/released from
the polyplexes prior to the cleavage of the disulde bonds so
that there is little correlation between the cleavage of the
disulde linkage and the release of DNA from the polyplexes.
Some of the PEI/DNA polyplexes escape from the endosomes
rapidly, and the disulde bonds of entrapped polyplexes are
cleaved inside the lysosomes, which occurs at ꢁ5 h aer
endocytosis.
Acknowledgements
The nancial support of the National Natural Scientic Foun-
dation Projects (20934005 and 51173177), the Ministry of
Science and Technology Key Project (2012CB933802), and the
Hong Kong SAR Earmarked Projects (CUHK4036/11P, 2130281;
2060431 and CUHK7/CRF/12G) is gratefully acknowledged.
Notes and references
Fig. 13 Colocalization of (a) (Rho-PEI)2/DNA and (b) Rho-PEI/DNA
polyplexes with lysosomes labeled by Lysosensor™ Green. In each
case, the upper panel: observed at channel 515/30 nm (green) and
channel 605/75 nm (red); and lower panel: observed at channel 605/
75 nm (red) (bar ¼ 50 mm).
1 D. W. Pack, A. S. Hoffman, S. Pun and P. S. Stayton, Nat. Rev.
Drug Discovery, 2005, 4, 581–593.
2 A. Rolland, Adv. Drug Delivery Rev., 2005, 57, 669–673.
3 I. M. Verma and N. Somia, Nature, 1997, 389, 239–242.
3290 | J. Mater. Chem. B, 2014, 2, 3282–3291
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
Journal of Materials Chemistry B
4 P. L. Felgner, T. R. Gadek, M. Holm, R. Roman, H. W. Chan, 18 W. T. Godbey, K. K. Wu and A. G. Mikos, Proc. Natl. Acad. Sci.
M. Wenz, J. P. Northrop, G. M. Ringold and M. Danielsen,
Proc. Natl. Acad. Sci. U. S. A., 1987, 84, 7413–7417.
5 Y. Z. Ma, S. Hou, B. Ji, Y. Yao and X. Z. Feng, Macromol.
Biosci., 2010, 10, 202–210.
U. S. A., 1999, 96, 5177–5181.
19 K. Itaka, A. Harada, Y. Yamasaki, K. Nakamura,
H. Kawaguchi and K. Kataoka, J. Gene Med., 2004, 6, 76–84.
20 Z. J. Dai and C. Wu, Macromolecules, 2012, 45, 4346–4353.
6 Y. Yao, D. F. Feng, Y. P. Wu, Q. J. Ye, L. Liu, X. X. Li, S. Hou, 21 Y. A. Yue, F. Jin, R. Deng, J. G. Cai, Y. C. Chen, M. C. M. Lin,
Y. L. Yang, C. Wang, L. Li and X. Z. Feng, J. Mater. Chem.,
2011, 21, 4538–4545.
7 O. Boussif, F. Lezoualch, M. A. Zanta, M. D. Mergny,
D. Scherman, B. Demeneix and J. P. Behr, Proc. Natl. Acad.
Sci. U. S. A., 1995, 92, 7297–7301.
8 D. Fischer, T. Bieber, Y. X. Li, H. P. Elsasser and T. Kissel,
Pharm. Res., 1999, 16, 1273–1279.
9 A. Akinc and R. Langer, Biotechnol. Bioeng., 2002, 78, 503–
508.
10 A. Akinc, M. Thomas, A. M. Klibanov and R. Langer, J. Gene
Med., 2005, 7, 657–663.
11 M. Breunig, U. Lungwitz, R. Liebl and A. Goepferich, Eur. J.
Pharm. Biopharm., 2006, 63, 156–165.
12 Y. Matsumoto, K. Itaka, T. Yamasoba and K. Kataoka, J. Gene
Med., 2009, 11, 615–623.
13 G. Grandinetti and T. M. Reineke, Mol. Pharm., 2012, 9,
2256–2267.
H. F. Kung and C. Wu, J. Controlled Release, 2011, 155, 67–76.
22 Y. A. Yue, F. Jin, R. Deng, J. G. Cai, Z. J. Dai, M. C. M. Lin,
H. F. Kung, M. A. Mattebjerg, T. L. Andresen and C. Wu, J.
Controlled Release, 2011, 152, 143–151.
23 S. Bauhuber, C. Hozsa, M. Breunig and A. Gopferich, Adv.
Mater., 2009, 21, 3286–3306.
24 M. Breunig, U. Lungwitz, R. Liebl and A. Goepferich, Proc.
Natl. Acad. Sci. U. S. A., 2007, 104, 14454–14459.
25 R. Deng, Y. Yue, F. Jin, Y. C. Chen, H. F. Kung, M. C. M. Lin
and C. Wu, J. Controlled Release, 2009, 140, 40–46.
26 H. Koo, G. W. Jin, H. Kang, Y. Lee, K. Nam, C. Z. Bai and
J. S. Park, Biomaterials, 2010, 31, 988–997.
27 N. Zhao, S. Roesler and T. Kissel, Int. J. Pharm., 2011, 411,
197–205.
28 Y. Lee, H. Mo, H. Koo, J. Y. Park, M. Y. Cho, G. W. Jin and
J. S. Park, Bioconjugate Chem., 2007, 18, 13–18.
29 J. Yang, H. Chen, I. R. Vlahov, J. X. Cheng and P. S. Low, Proc.
Natl. Acad. Sci. U. S. A., 2006, 103, 13872–13877.
30 C. D. Austin, X. H. Wen, L. Gazzard, C. Nelson, R. H. Scheller
and S. J. Scales, Proc. Natl. Acad. Sci. U. S. A., 2005, 102,
17987–17992.
14 C. Pichon, L. Billiet and P. Midoux, Curr. Opin. Biotechnol.,
2010, 21, 640–645.
15 S. N. Xiang, H. J. Tong, Q. Shi, J. C. Fernandes, T. Jin,
K. R. Dai and X. L. Zhang, J. Controlled Release, 2012, 158,
371–378.
31 J. R. Lakowicz, J. Malicka, S. D'Auria and I. Gryczynski, Anal.
Biochem., 2003, 320, 13–20.
16 A. Baker, M. Saltik, H. Lehrmann, I. Killisch, V. Mautner,
G. Lamm, G. Christofori and M. Cotten, Gene Ther., 1997, 32 J. R. Lakowicz, J. Malicka, S. D'Auria and I. Gryczynski,
4, 773–782. Biophys. J., 2003, 84, 291a.
17 W. T. Godbey, K. K. Wu and A. G. Mikos, J. Controlled Release, 33 D. Setiawan, A. Kazaryan, M. A. Martoprawiro and M. Filatov,
1999, 60, 149–160.
Phys. Chem. Chem. Phys., 2010, 12, 11238–11244.
This journal is © The Royal Society of Chemistry 2014
J. Mater. Chem. B, 2014, 2, 3282–3291 | 3291