(Received for publication, December 17, 1996, and in revised form, March 12, 1997)
From the Department of Applied Chemistry, Faculty of
Engineering, Nagasaki University, Nagasaki 852, Japan,
§ Central Laboratory, Institute of Tropical Medicine,
Nagasaki University, Nagasaki 852, Japan, and ¶ Department of
Bacteriology, Institute of Tropical Medicine, Nagasaki University,
Nagasaki 852, Japan
Polycationic reagents such as cationic lipids and
poly-L-lysine are widely used for gene transfer into
cells in vitro and show promise as vectors for in
vivo gene therapy applications as nonviral gene transfer
techniques. We have developed a novel transfection method using
cationic amphiphilic -helical oligopeptides with repeated sequences.
Oligopeptide has the advantages of being easily designed and modified
because of its simple structure. In this study, we synthesized five
kinds of peptides of which the total chain length and the width of the
hydrophobic region were changed. The binding of the peptides to plasmid
DNA was evaluated by agarose gel electrophoresis. It was found that the
long and/or hydrophobic peptides can strongly bind to the DNA. The
formation of large aggregates with a 0.5-5-µm diameter, which
consisted of the long peptides and the DNA, was observed by electron
microscopy. The transfection abilities of the peptides were determined
by the expression of luciferase from its cDNA in COS-7 cells. The long peptides showed high transfection abilities. As a result, it could
be said that the transfection ability of these peptides was parallel to
their ability to form aggregates with DNA. Furthermore, the
transfection ability was increased by the addition of chloroquine in
the transfection procedure. This result indicated that the internalization of the peptide-DNA aggregates would be mediated by the
endocytosis pathway.
Gene transfer techniques have caused an important advance in the treatment of both inherited and acquired diseases. Most ongoing gene therapy protocols rely on recombinant retrovirus, adenovirus, adeno-associated virus, etc., which have a number of advantages for gene transfer because of their general efficiency and wide range of cell targets (1-4). However, there were some problems with these protocols, i.e. difficulty in insertion of large size DNA fragments into the gene, the possible production of replication-competent viruses by recombination, generation of several types of immune response, and oncogenic effects by random insertion into the host genome (5, 6). To avoid such problems, the development of nonviral gene transfer techniques has been also encouraged, particularly the use of cationic lipid and polycation such as polylysine. Many gene transfer techniques using cationic lipids bearing essentially a single tertiary or quaternary ammonium head group have been tested in vitro (7-14). Furthermore, Remy et al. (15) succeeded in targeted gene transfer into hepatoma cells using lipopolyamine-condensed DNA particles presenting galactose ligands. Polymeric DNA-binding cations such as polylysine, which were linked to cell-targeting ligands such as asialoorosomucoid, transferrin, insulin, galactose, or lactose were also used in the targeted gene transfer. In these systems, the ligands on the polymeric cations trigger receptor-mediated endocytosis into cells (16-20). In these works, the gene transferred into cells was found to be extensively degraded in the acidic lysosomal compartment. Recently, effective gene transfer has been achieved using a hemisynthetic virus that was prepared by coupling polylysine-asialoorosomucoid with adenovirus (21) and using a complex of DNA, polylysine-transferrin, and fusogenic peptide (22).
As synthetic vectors for nonviral gene transfer techniques, various
complexes have been used. These were usually made by a combination of
several functional groups or compounds that were able to transfer genes
into the targeted cell and to perturb the lysosomal membrane to avoid
degradation of the gene in the lysosome. However, the design of the
synthetic vectors is generally complicated due to the introduction of
several functional compounds such as glycoprotein and the fusogenic
peptide. For the purpose of further study of the gene transfer system
based on polycationic reagent, here we investigated gene transfer into
cells using simple molecules, i.e. cationic oligopeptides
that have the potential to take an -helical structure. Previously,
we studied perturbation and fusion of phospholipid membrane by cationic
amphiphilic
-helical model peptides and elucidated the relationship
between their structure and function (23-25). In the present study, we
have examined the correlation between DNA binding ability of the
peptides or morphology in the peptide-DNA aggregation and their
transfection ability. The results indicated that the cationic
amphiphilic
-helical peptides were highly efficient in transferring
genes into cultured cells and that their transfection ability increased
with their ability to cause aggregation.
Reagents used for the synthesis and analysis were of reagent grade. Amino acid derivatives were purchased from Watanabe Chemical (Hiroshima, Japan). Poly-L-lysine (15-30 kDa) was purchased from Sigma, and plasmid DNA that contained a luciferase gene and SV40 promoter (PicaGene control vector, PGV-C)1 was from Toyo Ink (Tokyo, Japan). Closed circular plasmid DNA was purified by ultracentrifugation in CsCl gradients. Lipofectin was purchased from Life Technologies, Inc.
Peptides SynthesisThe peptides were prepared by solid phase synthesis using oxime resin (26-28). The synthesized peptides were purified by high pressure liquid chromatography with reversed-phase column (YMC-Pack ODS A-323 or YMC-Pack C4 A-823, 10 × 250 mm). Elution was carried out with a linear gradient established between 50 and 100% acetonitrile in 0.05% trifluoroacetic acid for 30 min monitored at 220 nm. The final products were identified by elemental analysis, amino acid analysis, and matrix-assisted laser desorption ionization mass spectrometry using a Shimadzu Kratos Kompact MALDI III apparatus.
DNase I Protection AssayThe tests were performed by mixing 0.5 µg of the plasmid DNA (PGV-C) with the peptides, in which the positive (peptide):negative (DNA) charge ratios were 0, 0.10, 0.25, 0.50, 1.0, 2.0, and 4.0, respectively, in 45 µl of HBS (21 mM Hepes-NaOH buffer containing 135 mM NaCl, 5.0 mM KCl, and 0.76 mM Na2HPO4, pH 7.4). After 30 min at room temperature, 5 µl of a solution of 10 mM MgCl2 and 10 mM CaCl2 was added, followed by 5 µl of 0.5 µg/ml DNase I (Worthington, DPPF grade) in water. After 30 min at 42 °C, 50 µl of a stop solution consisting of 4 M ammonium acetate, 20 mM EDTA, and 2 mg/ml glycogen was added, and the reaction mixture was placed on ice. To dissociate the plasmid DNA from the peptide, 15 µl of 1% SDS was added prior to extraction with TE-saturated phenol/chloroform, followed by ethanol precipitation. The final pellet was resuspended in 25 µl of dye mixture (TBE, 0.02% bromphenol blue, and 5% glycerol). TBE consists of 90 mM Tris, 90 mM boric acid, and 2 mM EDTA, pH 8.0. The aliquot of 5.0 µl was applied to 1% agarose gel electrophoresis.
Circular DichroismCD spectra were recorded on a JASCO J-720W spectropolarimeter using a quartz cell of 1.0-mm path length. The peptides were dissolved at a concentration of 10 µM in HBS. Measurements were performed in the presence of the plasmid DNA at a peptide:DNA charge ratio of 1.0 and in the absence of the plasmid DNA.
Electron MicroscopySamples were prepared by mixing 63 µM of peptide (per cationic charge concentration) and 10 µg/ml plasmid DNA in HBS at final concentrations. These were then left standing for 30 min at room temperature. Peptide-DNA complexes were processed for transmission electron microscopy using a negative stain technique. Fifteen-µl drops of freshly prepared samples were placed on glow-discharged carbon-coated 200-mesh copper grids for 3 min. Solution was wicked off with filter paper and replaced with 1% aqueous uranyl acetate for 30 s. After removal of the solution, grids were rinsed in distilled water and allowed to dry. Grids were imaged in a JEOL JEM-100CX transmission electron microscope.
Peptide-mediated TransfectionCOS-7 cells, a simian kidney cell line transformed with simian virus 40 (SV40), were grown to just before confluence in 16-mm dishes in Dulbecco's modified Eagle's medium with 10% fetal calf serum and 100 µg/ml streptomycin in an atmosphere of 5% CO2 at 37 °C, and washed twice with 1 ml of HBS. Plasmid DNA in 125 µl of HBS and peptide in 125 µl of HBS were mixed and allowed to stand for 15 min at room temperature. The mixture was poured gently onto the cells. After incubation for 3 h at 37 °C, 1 ml of Dulbecco's modified Eagle's medium with 10% fetal calf serum was added. The medium was replaced with 1 ml of a fresh medium after 12 h, and the cells were incubated for 48 h from the first addition of the medium. Treatment of the cells with chloroquine was performed at 100 µM during transfection procedure. Harvesting of cells and luciferase assays were performed 48 h after transfection as described in the protocol of PicaGene luminescence kit (Toyo Ink; Tokyo, Japan). The light units were analyzed by luminometer (Maltibiolumat LB9505, Berthold, Germany). The light unit values shown in the figures represent the specific luciferase activity (cpm/mg of protein) which is standardized for total protein content of the cell lysate. The measurement of gene transfer efficiency was performed in triplicate.
As shown in Fig. 1, peptides
43 and 46, which were previously synthesized,
take an amphiphilic -helical structure. These consist of 12 and 24 residues, respectively, with a repeat of tetrapeptide units (23-25).
Peptides 43S and 46S were designed on the basis
of 43 and 46 sequences so as to reduce the
hydrophobicity of 43 and 46 by introducing 3 and 6 hydrophilic Ser residues, respectively, instead of hydrophobic
Leu and Ala residues. As a consequence, the hydrophilic region of the
helix was increased from 1/4 in 43 and
46 to 1/2 in 43S and 46S.
Peptide 46P contains the Pro8-Pro9
sequence, which can disrupt
-helical structure, instead for Leu8-Leu9 of 46 in the middle of
the peptide chain. The 43 and 46 series peptides have 3 and 6 residues of cationic amino acid (Arg),
respectively, in the molecules.
Formation of the Peptide-DNA Complex
The peptide-DNA complex
formation was examined by the electrophoretic mobility of the complex
on an agarose gel (1%, w/v) stained with ethidium bromide at the
various ratio of peptides to a double-stranded DNA. The tests were
performed by mixing 0.1 µg of the plasmid DNA (PGV-C) with the
peptides, in which the positive (peptide)/negative (DNA) charge ratios
were 0, 0.10, 0.25, 0.50, 1.0, 2.0, 4.0, and 8.0, respectively (Fig.
2A). In the case of 43, no
migration of the plasmid DNA band occurred at charge ratio of 2.0. This
lack of migration was due to neutralization of nucleic acid by cationic
peptide and/or formation of a large complex between the peptide and the
DNA. Peptide 43S gave no effect on migration of the DNA.
The long peptides, 46, 46S, and
46P, suppressed the migration of the DNA at charge ratios
of 0.5, 1.0, and 1.0, respectively. These results indicated that
peptides with long length and large hydrophobic region strongly bound
to the plasmid DNA. Introduction of proline residues, which breaks the -helical structure, at the center of 46 slightly reduced
the binding ability.
DNase I Protection Assay
When peptides bind to the plasmid DNA, it is expected that digestion of the DNA by DNase I will be inhibited. The DNA binding abilities of the peptides were evaluated by the nuclease-inhibitory activities. After adding the peptides to the plasmid DNA at various charge ratios, the mixtures were incubated with DNase I. After undigested DNA was extracted, the DNA was analyzed by 1% agarose gel electrophoresis (Fig. 2B). In the cases of 43, 46, 46S, and 46P, the bands of the DNA were remarkably detected at a charge ratio of 2.0, 0.25, 1.0, and 0.25, respectively. In each case, the band upper to the position of supercoiled plasmid DNA appeared. It was considered that the upper bands corresponded to the position of the nicked or linearized DNA digested by the nuclease at one site. However, 43S offered no resistance to digestion of the DNA by the nuclease. In the case of 46, the upper band was scarcely detected at a charge ratio of 4.0. These findings indicated that 46 had the strongest binding ability to the plasmid DNA and were consistent with the results from the agarose gel shift assay described above.
CD SpectraTo analyze the structural features of these
peptides that bound to the plasmid DNA, CD spectra of the peptides were
measured. In these measurements, it was hard to analyze spectra under
205 nm due to large absorbance. As shown in Fig.
3A, 46 showed a typical -helix
CD pattern with double minima at 208 and 222 nm in the HBS. The
ellipticity at 222 nm indicated that the peptide took an almost
complete
-helical structure (>95%). The helical content of
46S and 46P was lower than that of
46. An increase in the hydrophilic region or the
introduction of proline residues into peptides, which disrupts
-helical structure, actually reduced the
-helicity of the
peptides. Peptides 43 and 43S took mainly a
random coil structure in the HBS.
In the presence of the plasmid DNA at charge ratio of 1.0, 43S took a slightly -helical structure (<20%) (Fig.
3B), while the helical content of 43 was
remarkably increased by the addition of the DNA. This result means that
43 binds to the plasmid DNA with a stable
-helical
structure. Peptides 46, 46S, and
46P showed a large valley at 222 nm, and another valley at
208 nm became shallower. This spectrum would mean aggregation of
peptide with DNA as described by Yoshimura et al. (29).
Peptide 46P also showed a large valley at 222 nm, but its
depth was shallower than that of 46 and 46S. It
is possible that two proline residues in the center of 46P
effectively broke the
-helical structure in binding to the DNA.
To assess the structure of the
peptide-DNA complexes, we used transmission electron microscopy with
negative staining. Fig. 4A shows the free
plasmid DNA without peptide, and Fig. 4, B and C,
shows complexes of 46 and the plasmid DNA at a charge ratio of 2.0. Interestingly, aggregates with a diameter of 0.5-5 µm were
found. Although there were several shapes in aggregates, twisted
fiber-like structures were common. The peptides 46S,
46P, and 43 except for 43S caused
aggregation analogous with that of 46 (Fig. 4,
D-F). Although 43 generated similar
aggregations as 46, 46S, and 46P,
the population of the aggregates was lower. The results indicated that
most of the complex of 43 and the plasmid DNA was dissolved
in HBS without forming large aggregates, because 43
indicated the typical -helical CD pattern in the presence of DNA
despite no strong aggregation with DNA. Furthermore, in the case of
46, the aggregates were found at a peptide:DNA charge ratio
of 0.50, while no aggregate was found at a charge ratio of 0.10. These
results were consistent with those from the agarose gel shift assay and
DNase I protection assay.
We furthermore found that the formation of aggregates was inhibited by
reduction of sodium chloride concentration in buffer solution (data not
shown). This result indicates that not only electrostatic interaction
but also hydrophobic interaction between peptide and DNA is necessary
for the formation of complex and aggregate. The mechanism of formation
of large aggregates by peptide and DNA would be explained as follows.
1) Positive charges of the arginine residues in the peptides stand in
line by taking an -helical structure and electrostatically interact
with the negative charge of phosphate in DNA. 2) Formation of a
hydrophobic face on the opposite side of the DNA-binding face in an
-helix peptide causes hydrophobic interaction between peptide-DNA
complexes, and the interaction results in aggregation.
The efficiency of these peptides in
the transfection of COS-7 cells was tested with the plasmid DNA that
contains a reporter gene encoding firefly luciferase. The efficiencies
of the peptides in the expression of luciferase were determined at
48 h after transfection by measurements of the total enzyme
activity in the cell extracts of the cultured cells using a
luminometer. In Fig. 5, the cross-hatched bar
shows the transfection efficiency of the peptides when the peptides
were mixed with 2.5 µg of the plasmid DNA at a charge ratio of 2.0 without chloroquine treatment. In these cases, 125 µl of 126 µM (cationic charge concentration) peptides and 125 µl
of 20 µg/ml plasmid DNA were mixed, and the mixture was added to the
COS-7 cells in a 16-mm dish. Peptides 43 and
43S were found to exhibit no transfection ability, while the long peptides 46, 46S, and 46P
had remarkable abilities; in particular, 46 had 11- and
6-fold higher abilities compared with those of 46S and
46P, respectively. To examine the efficiency of other DNA
carrier molecule, we used poly-L-lysine as a cationic polypeptide. The poly-L-lysine was found to exhibit the
same transfection ability as that of 46S. However, the
efficiencies of methods using calcium phosphate and cationic liposome
(Lipofectin) were 3- and 5-fold higher, respectively, than that of
46.
Cytotoxic activities of the peptides and the complex of the peptides
and the plasmid DNA were also evaluated by the MTT assay (30). Cell
viabilities in the presence of 43, 46, and
46P at a concentration of 65 µM (cationic
charge) decreased to about 40, 50, and 40%, respectively. On the other
hand, the peptide-plasmid DNA complexes at a charge ratio of 2 had high
cell viability of about 80% at the same peptide concentration (Fig.
6). It was clear that the binding of the plasmid DNA to
the peptides reduced the cytotoxity of the peptides.
As shown in Fig. 5 (open bar), concurrent treatment of the cells with chloroquine, which may inhibit the degradation of the DNA by lysosomal hydrolases, made the efficiencies of all peptides increase remarkably. The gene transfer efficiencies of the long peptides (46, 46S) were 30-200-fold higher than those of the short peptides (43, 43S). Surprisingly, the treatment with chloroquine increased the efficiency of 46P by about 400-fold. In this condition, cell viabilities were reduced by 5-25% compared with in the absence of chloroquine treatment (Fig. 6). Since the increase in transfection efficiencies by the addition of the chloroquine was considerably higher than that in the cytotoxity, it could be said that the increase in the transfection efficiency, which is standardized for total protein content of the cell lysate, is not due to a decrease in living cells. These findings indicated that internalization of the peptide-DNA aggregates was mediated by endocytosis pathway.
That 46 has the highest transfection efficiency in the absence of chloroquine treatment can be explained by the fact that 46 can bind more strongly to the plasmid DNA and more efficiently inhibits degradation of the DNA by DNase I compared with 46S and 46P. As a consequence, a large amount of the DNA could survive in the lysosomal vesicles. On the other hand, 43 had little transfection ability, although this peptide could bind to DNA. It could be considered that most of aggregates were degraded in the lysosomal vesicles because 43 could not protect the DNA from DNase digestion adequately, although the aggregates of 43 and DNA might be incorporated into cells. It is clear that 43S lacked transfection ability because it could not bind to DNA. In the presence of chloroquine, degradation of incorporated DNA was restrained. Therefore, transfection efficiency of any peptide was increased. The efficiencies of 46 and 46S were similar, suggesting that both peptides caused similar aggregations and that the aggregates were incorporated with the same efficiency. On the other hand, 46P, which had a similar aggregation pattern to that of 46 and 46S, showed remarkably stronger transfection efficiency. Since 46P had stronger membrane perturbation activity compared with 46 and 46S,2 it is likely that a large amount of DNA could be transferred from endosomal vesicles to cytosol without remarkable degradation of the plasmid DNA. These results indicated that transfer of the peptide-DNA complex from the endosomal compartment to the cytosol is an important step for efficient expression of the gene in the presence of chloroquine. Therefore, it is possible that transfection efficiency depends on endosomal or lysosomal membrane perturbation activity of the peptide. Further investigation will be necessary for pursuing the localization of incorporated DNA in the cell.
The effect of the amount of 46 and the plasmid DNA on the
transfection efficiency is shown in Fig. 7. The most
effective transfection occurred when 2.5 µg of the plasmid DNA and
2.6 nmol of 46 were mixed. Furthermore, in any amount of
the DNA, the peptide:DNA mixing ratio was optimum at 2.0. When the
peptide:DNA charge ratio was smaller than 2.0, the efficiencies were
low. It is probable that transfection-competent aggregates of peptide
and DNA were hardly formed. Reduction of the efficiency at the higher
amounts of both peptide and DNA than those at optimum conditions
suggests that aggregates too large for endocytosis were formed because of the high concentration of both peptide and DNA. When the peptide:DNA charge ratio was higher than 2.0, the transfection efficiencies were
also reduced. This was due to the cytotoxic activities of the excess
free peptides that remained in the solution.
Conclusions
In this study, we examined some abilities of
amphiphilic -helical oligopeptides containing cationic amino acid as
a gene carrier molecule. These peptides transferred plasmid DNA into cells and demonstrated high expression of protein encoded in the plasmid DNA. In particular, the peptides having a continuous large hydrophobic region, i.e. 46, could form stable
aggregation with the plasmid DNA and also efficiently transfer the DNA
into cells. Furthermore, this study indicated that membrane
perturbation activity of the peptide also was important for the
efficient transfection.
To increase the transfection efficiency of the peptide, it will be important to make the aggregates transfer from endosomal vesicle to cytosol efficiently and to achieve cell-specific gene delivery by the peptide with some ligands. Recent methods of peptide synthesis make it possible to design and exactly synthesize complicated peptides, e.g. peptide with ligands recognizable by specific cells, bundled peptide with multifunctions, peptide with a nonpeptidyl component, and peptide-nucleic acid hybrid. Accordingly, peptide has a high potential as a functional gene carrier because of its diversity on construction.
We are indebted to Prof. M. Nakamura (Institute of Tropical Medicine, Nagasaki University) for the use of the Maltibiolumat LB9505 luminometer. We also thank Dr. P. I. Padilla (Institute of Tropical Medicine, Nagasaki University) for critical reading and revisions of the manuscript.