New Basic Membrane-Destabilizing Peptides for ... - CyberLeninka

6 downloads 0 Views 296KB Size Report
Key words: membrane-destabilizing peptides, -helix, synthetic vector, gene .... amino groups necessary to neutralize the negative charges .... fect human and murine cell lines like SW480, LoVo, 293- ..... KALA, which might need a layer of free peptides, associ- ... plexes, a few animals died after injection of the complexes.
ARTICLE

doi:10.1006/mthe.2002.0523, available online at http://www.idealibrary.com on IDEAL

New Basic Membrane-Destabilizing Peptides for PlasmidBased Gene Delivery in Vitro and in Vivo Karola Rittner,1,* Annie Benavente,1 Albine Bompard-Sorlet,1,† Frédéric Heitz,2 Gilles Divita,2 Robert Brasseur,3 and Eric Jacobs1 1 TRANSGENE S.A., 11 rue de Molsheim, F-67000 Strasbourg, France Centre de Recherche de Biochimie Macromoléculaire UPR 1086–CNRS, 1919, route de Mende, F-34293 Montpellier cedex 5, France 3 Centre de Biophysique Moléculaire Numérique–Faculté Universitaire des Sciences Agronomiques de Gembloux, passage des déportés 2, B-5030 Gembloux, Belgium 2



Present address: Roche Vitamins AG, VFB-Biotechnology-Discovery, CH-4070 Basel, Switzerland

*To whom correspondence and reprint requests should be addressed. Fax + (33) (0)3 88 22 58 07. E-mail: [email protected].

We have designed new basic amphiphilic peptides, ppTG1 and ppTG20 (20 amino acids), and evaluated their efficiencies in vitro and in vivo as single-component gene transfer vectors. ppTG1 and ppTG20 bind to nucleic acids and destabilize liposomes consisting of 1-palmitoyl-2-oleoylphosphatidylcholine (POPC) and cholesterol (3:1 mol/mol) at pH 5 and pH 7. Complexes of plasmid DNA and ppTG1 gave rise to efficient transfection in a variety of human and murine cell lines at low charge ratios ([+/–] between 1 and 2). In cell culture experiments, such vectors were superior to the membrane-destabilizing peptide KALA. In comparison with cationic lipid-, dendrimer-, and polymer-based transfection agents like Superfect, polyethylenimine (PEI), and Lipofectin, ppTG1 vectors showed good transfection efficiencies, especially at low DNA doses. Moreover, we demonstrated for the first time successful gene transfer in living animals with a single-component peptide vector. In the mouse, intravenous injection of a luciferase expression plasmid complexed with ppTG1 or ppTG20 led to significant gene expression in the lung 24 hours after injection. Structure–function studies with ppTG1, ppTG20, and sequence variants suggest that the high gene transfer activity of these peptides is correlated with their propensity to exist in -helical conformation, which seems to be strongly influenced by the nature of the hydrophobic amino acids. Key words: membrane-destabilizing peptides, -helix, synthetic vector, gene transfer

INTRODUCTION Synthetic gene delivery systems have attracted interest for gene therapeutic approaches because of their potential advantages over viral vectors in terms of safety and ease of manufacture [1]. Towards an efficient transfer of plasmid DNA, nonviral delivery systems based on polycations, liposomes, or peptides were designed to mimic advantageous aspects of viral infection with greatly simplified components. These aspects are DNA condensation or compaction, binding to the target cell surface, disruption of cell membrane or endosomes for intracellular release of the genetic information, and DNA delivery into the nucleus [reviewed in 2]. It seems feasible either to use naturally occuring peptides or to adapt and optimize peptide sequences to exert all or part of these tasks. One example of a peptide vector suitable for gene transfer in cell culture was described [3] with two peptides (K8 and JTS1) providing distinct functions: the basic polylysine

104

K8, which binds to plasmid DNA, and the acidic peptide JTS1, which contributes the membrane-destabilizing activity when added to the positively charged plasmid/K8 complexes. JTS1 has a sequence of alternating hydrophilic and hydrophobic amino acid residues, in common with other acidic amphiphilic peptides like INF1 [4,5] and GALA [6]. These peptides assume a random coil structure at pH 7. As the pH decreases, for instance in the late endosomes, the carboxyl groups of aspartate and glutamate side chains are protonated and the disappearance of the anionic repulsion allows the transition into an amphipathic -helical conformation. As a consequence, the peptides can interact with phospholipid membranes to form pores or induce membrane fusion and/or lysis [7]. To combine DNA-binding and membrane-destabilization activities in a single peptide, the basic amphipathic molecule KALA was designed [8]. KALA resembles GALA except for five internal glutamic acid residues, which were replaced by lysines,

MOLECULAR THERAPY Vol. 5, No. 2, February 2002 Copyright © The American Society of Gene Therapy 1525-0016/02 $35.00

doi:10.1006/mthe.2002.0523, available online at http://www.idealibrary.com on IDEAL

ARTICLE

FIG. 1. Gel retardation assay. The uncut plasmid pTG11236 (3 g) was incubated with increasing amounts of the peptides ppTG1, JTS1-K13, or KALA in 30 l of 0.9% NaCl solution. After 20 minutes at room temperature, aliquots were analyzed on a 1% agarose gel. Lane C shows the migration pattern of plasmid DNA in the absence of peptide (arrows). The lanes 1 to 6 correspond to retardation assays in the presence of 1, 2, 3, 4, 5, and 6 g of the respective peptide.

bind to plasmid DNA, to destabilize liposomes, and to promote gene transfer in different cell lines as well as in mice after intravenous injection. The gene transfer efficiencies with ppTG1 and ppTG20 were compared with those obtained with KALA and JTS1-K13, a peptide with a polylysine tail covalently added to JTS-1, as well as to commercially available transfection agents. We carried out transfection assays in the presence of bafilomycin A to study the mechanism of complex internalization. Furthermore, we generated data towards the understanding of sequence and structure requirements for the transfection with ppTG1-related peptides.

RESULTS resulting in a peptide with a net positive charge. KALA was able to bind to plasmid DNA and to mediate efficient transfection of a variety of cell lines at a high charge ratio ([+/–] 10) [8,9]. However, gene transfer with this peptide in living animals was not reported. To improve single-component peptide vectors, we set out to intimately associate the very efficient membranedestabilizing activity of the acidic peptide JTS-1 with a DNAbinding function. Therefore, we replaced all glutamic acid residues of JTS-1 by lysine or arginine residues to generate peptides with two overlapping functions. We studied the capacities of the resulting peptides, ppTG1 and ppTG20, to

We have tested the peptides ppTG1 and JTS1-K13, as well as a series of ppTG1 variants (Table 1A), for their capacity to bind to DNA, to cause leakage of self-quenched calcein from POPC liposomes in the absence or presence of cholesterol, and to mediate gene delivery evaluated by measuring reporter gene expression in vitro and in vivo. These peptides were compared with the “parental” peptide JTS1 [3] (Table 1A) and with the basic membrane-destabilizing peptide KALA [8] (Table 1A). The series of ppTG1 variants comprises peptides with all lysines replaced by arginine (ppTG20) or histidine residues (ppTG21), or all leucines replaced by isoleucine (ppTG30) or valine residues (ppTG32).

TABLE 1A: Sequence of peptides ppTG1 (MW 2297)

Gly-Leu-Phe-Lys-Ala-Leu-Leu-Lys-Leu-Leu-Lys-Ser-Leu-Trp-Lys-Leu-Leu-Leu-Lys-Ala

JTS-1 (MW 2302)

Gly-Leu-Phe-Glu-Ala-Leu-Leu-Glu-Leu-Leu-Glu-Ser-Leu-Trp-Glu-Leu-Leu-Leu-Glu-Ala

JTS-1-K13 (MW 4826) Gly-Leu-Phe-Glu-Ala-Leu-Leu-Glu-Leu-Leu-Glu-Ser-Leu-Trp-Glu-Leu-Leu-Leu-GluAla-Cys_Cys-Tyr-Lys-Ala-Lys-Lys-Lys-Lys-Lys-Lys-Lys-Lys-Trp-Lys-Lys-Lys-Lys-Gln-Ser KALA (MW 3131)

Trp-Glu-Ala-Lys-Leu-Ala-Lys-Ala-Leu-Ala-Lys-Ala-Leu-Ala-Lys-His-Leu-Ala-Lys-Ala-LeuAla-Lys-Ala-Leu-Ala-Lys-Ala-Leu-Lys-Ala-Cys-Glu-Ala

ppTG20 (MW 2437)

Gly-Leu-Phe-Arg-Ala-Leu-Leu-Arg-Leu-Leu-Arg-Ser-Leu-Trp-Arg-Leu-Leu-Leu-Arg-Ala

ppTG21 (MW 2342)

Gly-Leu-Phe-His-Ala-Leu-Leu-His-Leu-Leu-His-Ser-Leu-Trp-His-Leu-Leu-Leu-His-Ala

ppTG30 (MW 2296)

Gly-Leu-Phe-Lys-Ala-Ile-Ile-Lys-Ile-Ile-Lys-Ser-Ile-Trp-Lys-Ile-Ile-Ile-Lys-Ala

ppTG32 (MW 2184)

Gly-Leu-Phe-Lys-Ala-Val-Val-Lys-Val-Val-Lys-Ser-Val-Trp-Lys-Val-Val-Val-Lys-Ala

Peptides used in this study, indicating their molecular weight and amino acid sequence. The peptide JST1-K13 was generated by disulfide bridge formation between JTS1-Cys and NPys-Cys-K13. All peptides were soluble in milliQ water except for ppTG30, which was dissolved in 40% ethanol, and JTS1, which was soluble in 80% dimethylsulfoxid (DMSO).

MOLECULAR THERAPY Vol. 5, No. 2, February 2002 Copyright © The American Society of Gene Therapy

105

ARTICLE

doi:10.1006/mthe.2002.0523, available online at http://www.idealibrary.com on IDEAL

Complex Formation between Plasmid DNA and ppTG1 We carried out gel retardation assays demonstrating the formation of complexes between negatively charged nucleic acids and the positively charged peptides ppTG1, JTS1-K13, and KALA (Fig. 1). More than 90% of 3 g plasmid DNA was retained in the presence of 4 g of the respective peptide. We found that 4 g of peptide provided approximately the number of positively charged amino groups necessary to neutralize the negative charges of the phosphate moieties in the engaged amount of plasmid DNA ([+/–] 0.9). The formation of complexes with these peptides was reversible, as after 1% SDS treatment freely migrating plasmid DNA was detected by agarose gel electrophoresis. The negatively charged peptide JTS-1 was not able to retard DNA migration, as expected. Liposome Leakage Assays We then tested the peptides for their capacity to destabilize membranes in a liposome leakage assay. This assay was carried out [5] with POPC liposomes containing the fluorescent dye calcein in 20 mM sodium citrate buffer at pH 5 or pH 7. The release of calcein led to an increase of the fluorescence signal, which was maximal in the presence of the detergent Triton X-100 (positive control). On POPC liposomes, ppTG1, JTS1, and JTS1-K13 induced a 50% increase of fluorescence at peptide concentrations between 100 pg/l and 320 pg/l at pH 7 and pH 5 (data not shown). This was also the case for KALA at pH 5. At

pH 7, we observed a 50% increase of fluorescence at a slightly lower peptide concentration (64 pg/l; Table 2). This slightly higher POPC liposome-destabilizing activity of KALA at pH 7 has already been shown [8]. Next, we analyzed leakage activity on liposomes consisting of POPC and cholesterol (POPC/chol) at a molar ratio of 3:1 (Fig. 2A). Cholesterol is an important ubiquitous component of natural membranes that determines their fluidity. Assays with such liposomes are thus closer to experiments with living cells than assays with pure POPC liposomes. KALA did not promote calcein leakage from cholesterol-containing liposomes. The peptide JTS-1 mediated leakage at pH 5, at pH 7 this activity was reduced. JTS1-K13 was less efficient than JTS-1 in calcein release at pH 5, at pH 7 such activity was nearly undetectable. In contrast, ppTG1 could efficiently liberate calcein from POPC/chol liposomes at pH 5 and pH 7. The lack of pH-sensitivity was expected, as the protonation status of the lysine residues of ppTG1 does not change under the chosen pH conditions. We tested ppTG1, complexed to the plasmid DNA, for liposome leakage activity on POPC/chol liposomes at pH 7 (Fig. 2B). Complexes with a negative charge ratio [+/–] of 0.8, with all peptides theoretically engaged in DNA binding, still showed significant liposome leakage activity. This suggests that the lytic activity of the peptide is retained even if it is in complex with plasmid DNA. JTS1K13, complexed to plasmid DNA, did not mediate calcein release at charge ratios of [+/–] 0.8 (Table 2).

TABLE 1B: Synthesis and purification of peptides Synthesis

Purification

Peptide

strategy

resin

column

eluant

gradient

ppTG1

Fmoc

Fmoc-Ala-Wang resin

Uptiprep (Interchim) C18 120 Å 15 m

A: milliQ water + 0.1% TFA B: 40% A + 60% CH3CN

30–100% of B in 30 min

JTS-1

Boc

Boc-Ala-PAM resin

no purification

KALA

Fmoc

Fmoc-Ala-Wang resin

Uptiprep (Interchim) C18 120 Å 15 m

A: milliQ water + 0.1% TFA B: 40% A + 60% CH3CN

25–70% of B in 30 min

Uptiprep (Interchim) C18 120 Å 15 m

A: milliQ water + 0.1% TFA B: 40% A + 60% CH3CN

40–100% of B in 30 min

Uptiprep (Interchim) C18 120 Å 15 m

A: milliQ water + 0.1% TFA B: 40% A + 60% CH3CN

40–100% of B in 30 min

Uptiprep (Interchim) C18 120 Å 15 m

A: milliQ water + 0.1% TFA B: 40% A + 60% CH3CN

20–80% of B in 30 min

Uptiprep (Interchim) C18 120 Å 15 m

A: milliQ water + 0.1% TFA B: 40% A + 60% CH3CN

30–90% of B in 30 min

ppTG20 ppTG21 ppTG30 ppTG32 JTS-1-Cys

Fmoc Fmoc Boc Fmoc Boc

NPys-Cys-K13 Fmoc JTS1-K13

Fmoc-Ala-Wang resin Fmoc-Ala-Wang resin Boc-Ala-PAM resin Fmoc-Ala-Wang resin Boc-Cys(4Meb)-PAM resin

no purification

Fmoc-Ser(tBu)-Wang resin

Uptiprep (Interchim) C18 120 Å 15 m

A: milliQ water + 0.1% TFA B: 40% A + 60% CH3CN

0–40% of B in 30 min

Uptiprep (Interchim) C18 120 Å 15 m

A: milliQ water + 0.1% TFA B: 40% A + 60% CH3CN

25–70% of B in 30 min

The conditions for synthesis and purification of each peptide are summarized.

106

MOLECULAR THERAPY Vol. 5, No. 2, February 2002 Copyright © The American Society of Gene Therapy

doi:10.1006/mthe.2002.0523, available online at http://www.idealibrary.com on IDEAL

A

ARTICLE

B

FIG. 2. Liposome leakage assays. (A) The assay was carried out with serial dilutions of JTS1, ppTG1, JTS1-K13, and KALA on POPC/cholesterol (3:1) liposomes at pH 5 and pH 7. Increase of the fluorescence signal emitted at 535 nm as a result of calcein leakage was plotted against peptide concentration. Maximal calcein release was obtained in the presence of Triton X-100 solution (positive control), whereas incubation with H2O served as negative control. (B) Liposome leakage assay on POPC/cholesterol (3:1) liposomes at pH 7 with ppTG1 complexed to the plasmid pTG11236 at the indicated charge ratio.

Transfection Efficiency of ppTG1 Complexes in Vitro We compared the transfection efficiencies of plasmid DNA complexed with ppTG1, KALA or JTS1-K13 in HeLa cells. Complexes were formed with 50 ng of the luciferase expression plasmid pTG11236 and increasing amounts of the respective peptide at final charge ratios [+/–] between 0.2 and 10. We observed maximal luciferase activities for ppTG1/plasmid complexes at charge ratios [+/–] between 1 and 2 (Fig. 3A). For KALA, an optimum of luciferase activity was observed at a charge ratio of [+/–] 10, as described in [8]. Using JTS1-K13, maximal gene transfer was achieved at a charge ratio [+/–] of 4. Luciferase activities obtained with KALA and JTS1-K13/plasmid complexes, however, did not reach those levels obtained with ppTG1/plasmid complexes. Under these conditions, the multi-component peptide vector JTS1/K8 was at least 100fold less efficient than the single-component peptide vector ppTG1 (data not shown). The pH during complex formation does not seem to be of importance for gene transfer efficiency. Complexes formed between 100 ng pTG11236 and ppTG1 at charge ratios of [+/–] 1 and 2 in 0.9% NaCl in the absence (pH 5.8) or presence of 10 mM Tris (pH 8) were equally efficient in transfection studies on HeLa cells (data not shown). We compared the transfection efficiency of plasmid/ppTG1 complexes with those obtained with the

MOLECULAR THERAPY Vol. 5, No. 2, February 2002 Copyright © The American Society of Gene Therapy

commercially available transfection agents PEI [10], Lipofectin, and Superfect. HeLa cells were transfected with 25 ng, 50 ng, or 500 ng of the luciferase expression plasmid pTG11236 complexed with the respective transfection agent. We determined the luciferase activities per mg protein 20 hours after transfection (Fig. 3B). At a low DNA dose (25 ng), ppTG1 was the most efficient among the tested transfection agents. Increasing the DNA amount (50 ng) caused PEI to reach the level of ppTG1, whereas gene transfer with Lipofectin and Superfect remained low. At a DNA dose of 500 ng, transfection with PEI and Superfect was about 10 times more efficient than with ppTG1, whereas luciferase activity obtained with Lipofectin was still inferior. HeLa cells were transfected with 500 ng of a b-galactosidase expression plasmid using the same transfection agents as indicated in Fig. 3B. The relative numbers of blue cells reflected the differences in luciferase activities observed in Fig. 3B (data not shown). ppTG1/plasmid complexes were successfully used to transfect human and murine cell lines like SW480, LoVo, 293EBNA, B16F0, and TC1 (data not shown). These experiments confirmed that ppTG1 is an efficient transfection reagent, especially at low DNA doses. Transfection in the Presence of Bafilomycin A To study the cellular uptake mechanism, we transfected

107

ARTICLE

108 TABLE 2: Features of ppTG1 derivatives JTS1-K13, KALA, and JTS-1 ppTG1

ppTG20 Lys→Arg

ppTG21 Lys→His

ppTG30 Leu→Ile

ppTG32 Leu→Val

JTS-1

JTS1-K13

KALA

Gel retardation assaya

yes

yes

yes at pH 5 no at pH 8

yes

yes

no

yes

yes

Liposome leakage assay with free peptides [pg/l]b

~100

~100

~100

~100

> 2  105

pH 5: 64 pH 7: 1600

pH 5: 1600 inactive on POPC/ pH 7: > 2  105 chol liposomes (on POPC liposomes: 64 pg/l atpH 7 and 320 pg/l at pH 5)

Liposome leakage assay with plasmid/peptide complexes [pg/l]c

~ 320

~ 320

~ 100

> 2  105

not done

not done

pH 5: > 2  105 sharp decline at pH 7: > 2  105 charge ratio lower than [+/–] 2.5 [8]

pH-dependence of liposome leakage activityd

no

no

no

no

not detectable

yes

yes

[on POPC liposomes: yes]

Gene transfer in vitro [RLU/1/5 of cell lysate]e

1  107

1  107

not detectable

1  104

not detectable

not detectable

5  105

1  106

Optimal charge ratiof

1-2

1-2

-

1-2

-

-

4

10

Conformation observed predominant with CD and FTIR studiesg structure: -helix

predominant structure: -helix

random coil

predominant structure: -sheet

predominant structure: -sheet

not done (insolubility under assay conditions)

predominant structure: -helix

random coil - partially folded (estimation: 30% -helicity at pH 7.5 and 5% at pH 5.8)

Gene transfer in vivo [RLU/mg protein]h

2  106

not detectable

not detectable

not detectable

not detectable

9  104

not detectable

5  105

Copyright © The American Society of Gene Therapy

MOLECULAR THERAPY Vol. 5, No. 2, February 2002

Properties of ppTG1 and its derivatives, JTS-1, JTS1-K13, and KALA, in different assay systems are shown. a Gel retardation assay: retardation of plasmid DNA in the presence of the indicated plasmid at a charge ratio of [+/–] 0.9. b Liposome leakage assay: peptide concentration in [pg/l] for 50% increase of fluorescence due to liberation of calcein in leakage assay on POPC/cholesterol (3:1) liposomes. c Liposome leakage assay with peptide/DNA complexes at pH 7 if not indicated differently: peptide concentration in [pg/l] for 50% increase of fluorescence as a consequence of calcein liberation in leakage assay on POPC/cholesterol (3:1) liposomes in the presence of DNA at the charge ratio [+/–] 0.8. d pH-dependence of liposome leakage activity (pH 5 and pH 7). e Gene transfer in vitro: apparent reporter gene expression after transfection of 4  104 HeLa cells with 50 ng pTG11236 complexed with ppTG1-derived peptides at the charge ratio of [+/–] 2, with JTS1-K13 at a charge ratio of [+/–] 4, and with KALA at a charge ratio [+/–] 10. f Optimal charge ratio in transfection studies. g Biophysical studies to study the conformation of peptides. Circular dichroism (CD) studies were carried out for peptides with -helical conformation. In addition, FTIR studies were performed with ppTG30 and ppTG32 to confirm their -sheet conformation. h Gene transfer in vivo: apparent reporter gene expression in the lung observed after intravenous injection of 50 g luciferase expression plasmid (pTG1236) with the indicated peptides.

doi:10.1006/mthe.2002.0523, available online at http://www.idealibrary.com on IDEAL

Assay system

doi:10.1006/mthe.2002.0523, available online at http://www.idealibrary.com on IDEAL

A

ARTICLE

B

FIG. 3. Efficiency of gene transfer with ppTG1 complexes in vitro. (A) Comparison of ppTG1, JTS1-K13, and KALA in HeLa cells: 50 ng pTG11236, complexed with the indicated peptide at charge ratios between [+/–] 0.2 and 10, were used to transfect 4  104 HeLa cells on 24-well plates. Luciferase activity expressed as RLU was measured in 1/5 of total cell extract. (B) Twenty-five ng, 50 ng or 500 ng of plasmid pTG11236 were complexed with ppTG1, PEI, Lipofectin, or Superfect. These complexes were used to transfect 4  104 HeLa cells. Luciferase activities are shown as RLU/mg protein measured day 1 after transfection.

HeLa, A549, and WI-38 cells with ppTG1/plasmid complexes in the absence and presence of 175 nM bafilomycin A, which is a specific inhibitor of the vacuolar proton pump [11]. The results of these studies were expressed as the ratio of luciferase activities in the absence and in the presence of the drug (Table 3). The strong decrease of luciferase activity with PEI complexes in all three cell lines indicated that the acidification of the endosomes, which is essential for gene transfer with PEI [10], was severely disturbed. Under these conditions, gene transfer with KALA was reduced in the presence of bafilomycin A. This suggests that KALA complexes are taken up, at least partially, by endocytosis, with the low pH in the late endosomes being important for ensuring good transfection efficiencies. Gene transfer with ppTG1 complexes, however, was only slightly affected in the presence of bafilomycin A. This could be explained with cellular uptake of ppTG1 complexes across the cytoplasmic membrane or via endocytosis. In the latter case, escape from the endosomes would not depend on their acidification and could occur in early and late endosomes.

protonation status of the histidyl residues (pK around 6) and suggests that mainly electrostatic interactions are responsible for the binding of ppTG1 to plasmid DNA. Liposome leakage activities on POPC/chol (3:1) liposomes with ppTG1, ppTG20, and ppTG21 were comparable, whereas transfection efficiencies gave divergent results (Table 2). ppTG20/plasmid complexes resulted in gene transfer efficiency at least as high as that obtained with ppTG1. ppTG21, however, was not able to transfect HeLa cells, which can be attributed to the lack of binding to and complexing of plasmid DNA in cell culture medium (pH 7.4). To analyze the importance of the hydrophobic residues, all leucines in ppTG1 were replaced by isoleucines (ppTG30) or valines (ppTG32; Table 1A). Gel retardation assays confirmed the capacity of these peptides to bind to plasmid DNA like ppTG1 (Table 2). Liposome leakage assays on POPC/chol liposomes in the absence of DNA revealed a decrease of activity for ppTG32 compared with ppTG1, whereas ppTG30 retained significant activity (Table 2). In the presence of DNA, however, this activity

TABLE 3: Bafilomycin A sensitivity of gene transfer ppTG1 Derivatives and Their ppTG1 KALA PEI Characterization HeLa 2.8-fold 203-fold 4426-fold To identify important sequence requirements we 150 ng pDNA [+/–] 2 500 ng pDNA [+/–] 10 150 ng pDNA tested a series of ppTG1 derivatives in gel retardation, liposome leakage, and transient transfec- A549 2.1-fold 23-fold 280-fold 500 ng pDNA [+/–] 2 500 ng pDNA [+/–] 10 500 ng pDNA tion assays. ppTG20 and ppTG21 are variants of ppTG1 with all lysine residues replaced either by WI-38 4-fold not detectable 125-fold arginines or by histidines (Table 1A). ppTG20 250 ng pDNA [+/–] 2 500 ng pDNA [+/–] 10 250 ng pDNA showed a DNA-binding efficacy in gel retarda- Here is shown the sensitivity of gene transfer with ppTG1, KALA, and PEI against bafilomycin A in HeLa, tion assay comparable to that of ppTG1 (Table 2). A549, and WI-38 cells. These cells were transfected with the indicated amounts of the luciferase expresppTG21 was able to retain plasmid DNA at pH 5, sion plasmid pTG11236. Complexes with KALA were formulated at a charge ratio of [+/–] 10, complexes with ppTG1 at a charge ratio of [+/–] 2. The fold-reduction of luciferase activity in the presence of but not at pH 8 (Table 2). This observation is bafilomycin A is presented. consistent with the pH dependency of the

MOLECULAR THERAPY Vol. 5, No. 2, February 2002 Copyright © The American Society of Gene Therapy

109

ARTICLE

A

doi:10.1006/mthe.2002.0523, available online at http://www.idealibrary.com on IDEAL

B

FIG. 4. Biophysical experiments. FTIR and CD spectra with ppTG1 and ppTG32. ppTG20 was indistinguishable from ppTG1 in these assays, and ppTG30 strongly resembled ppTG32 (data not shown). (A) FTIR spectra of ppTG1 (bottom spectra) and ppTG32 (top spectra) in the absence (full lines) and presence of dioleoyl-phosphatidylcholine (DOPC) (dashed lines). The -helical conformation is indicated by an amide I band centered at 1654 cm–1. Amide I bands centered at 1628 cm–1 are characteristic for -sheet conformation. (B) CD spectra of ppTG1, ppTG32, KALA, and JTS1-K13 in 50 mM phosphate buffer at pH 7.5. The ellipticity is given per residue. Minima at 208 nm indicate -helicity.

was strongly reduced (Table 2). In transfection experiments on HeLa cells, ppTG32 showed no activity at all and the activity of ppTG30 was low (Table 2). Circular dichroism (CD) and Fourier Transform Infrared (FTIR) studies revealed -sheet conformation for ppTG32 in the absence and presence of the lipid DOPC. The FTIR spectra for ppTG32 in the absence and presence of the lipid DOPC revealed amide I bands centered at 1628 cm–1, which are characteristic for -sheet conformation (Fig. 4A), whereas ppTG1 exists predominantly in an -helical conformation as indicated by an amide I band centered at 1654 cm–1 (Fig. 4A). CD spectra taken for ppTG1 and JTS1-K13 showed their propensity to exist predominantly as an helical structure with two characteristic minima at 208 nm and 222 nm, and a maximum at 193 nm. A mainly helical conformation in aqueous solution has already been reported for cationic amphipathic peptides consisting solely of leucines and lysines [12]. In contrast, KALA was only partially folded in an -helix (Fig. 4B and Table 2), which was previously documented [8]. ppTG20 was indistinguishable from ppTG1 in these assays, and ppTG30 strongly resembled ppTG32 (Table 2). Gene Transfer in Mice Using ppTG1 and ppTG20 Furthermore, we have investigated the potential of gene transfer with ppTG1/peptide complexes in living

110

animals. Five B6SJL mice per group were intravenously injected with 250 l 5% glucose solution containing 60 g or 50 g pTG11236 complexed with ppTG1, ppTG20, and ppTG32. Mice were sacrificed 24 hours after injection. The lungs were harvested and tested for luciferase activity standardized for total protein amount. Gene transfer with ppTG1 complexes at charge ratios [+/-] between 1.8 and 2.1 led to significant luciferase activities in the lung (Fig. 5A). Complexes with ppTG20 resulted in slightly better gene transfer. Complexes with the peptide ppTG32, which showed no membranolytic activity, did not give rise to detectable reporter gene expression. This indicates that the combination of DNAbinding and membrane-destabilizing activities is a prerequisite for gene transfer in vivo. We compared complexes with ppTG1 to those formed with JTS1-K13, KALA, K8/JTS-1 [3], and ppTG20. KALA and the multicomponent peptide vector K8/JTS-1 were inefficient (data not shown). Luciferase activities observed in the lung at day 1 after intravenous injection of 50 g pTG11236 complexed with increasing amounts of ppTG1, JTS-1-K13, and ppTG20 are shown in Fig. 5B. Gene transfer with the peptides ppTG1 and ppTG20 led to higher gene expression than with JTS1-K13. These data are the first showing single-component peptide vectors that enable significant gene transfer in vivo.

MOLECULAR THERAPY Vol. 5, No. 2, February 2002 Copyright © The American Society of Gene Therapy

doi:10.1006/mthe.2002.0523, available online at http://www.idealibrary.com on IDEAL

A

ARTICLE

B

FIG. 5. In vivo gene transfer efficiencies with ppTG1 and ppTG20 complexes. Lipoplexes or complexes consisting of peptide and plasmid DNA pTG11236 (p) were formed in 250 l 5% glucose. These solutions were intravenously injected into five B6SJL mice per group. At day 1 after injection the animals were sacrificed, and lungs were recovered and analyzed for luciferase expression. Asterisks indicate the number of mice per group that died before the end of the experiment. All other mice showed no (level 1) or only a few mild clinical signs (level 2). (A) Comparison of complexes consisting of plasmid DNA and the indicated amounts of the peptides ppTG1, ppTG20, and ppTG32, which represent charge ratios between 1.8 and 2.3. (B) Comparison of complexes consisting of 50 g of the plasmid pTG11236 (p) and the indicated amounts of the peptides ppTG1, ppTG20, and JTS1-K13.

DISCUSSION Cellular membranes represent a major barrier to gene delivery with the available synthetic vectors [13]. Towards the design of a new single-component peptide vector, we combined a DNA-binding domain with a very efficient membrane-destabilizing function. By modifying the amphipathic peptide JTS-1 [3], we succeeded in designing the short basic peptide ppTG1, and variants like ppTG20, with overlapping DNA-binding and membrane-destabilizing activities. We have demonstrated the DNA-binding property of ppTG1 and its cationic derivatives in gel retardation assays. This activity seems to rely mainly on electrostatic interactions with the basic amino acid residues lysine or arginine, because at pH 8 the uncharged histidine variant ppTG21 could not bind to DNA. This property may explain the lack of transfection efficiency with this peptide. Plasmid/peptide complex formation with ppTG1 as well as ppTG20, ppTG30, ppTG32, KALA, and JTS1-K13 was reversible in the presence of 1% SDS. Reversibility of complex formation within the cell and accessibility of plasmid DNA to the transcriptional machinery can be postulated as essential for gene expression. We have tested membrane destabilizing activity on liposomes consisting of POPC and cholesterol encapsulating self-quenched calcein. These liposomes appeared to be a useful tool to predict the efficiency of peptides of the ppTG1 family in gene transfer experiments. Indeed, the peptides ppTG30 and ppTG32 showed reduced liposome leakage and transfection efficiencies, in contrast to ppTG1 and ppTG20. In addition, the pH sensitivity of JTS-1 in leakage assays on erythrocytes [3] could be mimicked in the leakage assay with POPC/chol liposomes (24-fold more peptide was needed at pH 7 than at pH 5 for 50% increase

MOLECULAR THERAPY Vol. 5, No. 2, February 2002 Copyright © The American Society of Gene Therapy

of fluorescence), whereas we did not observe significant pH dependence for JTS-1 on pure POPC liposomes [3]. In the case of KALA, the capacity to transfect cells in culture was not reflected in its POPC/chol liposome leakage activity, as no calcein release was observed. On POPC liposomes, however, KALA did induce calcein leakage [8] (Table 2), as well as on the more complex POPC/1-palmitoyl-2oleoylphosphatidylglycerol (POPG) liposomes [8]. In contrast to the negatively charged POPC/POPG liposomes, there is no electrostatic interaction between KALA and the neutral POPC/chol liposomes, which can therefore be suggested to be a prerequisite for KALA to destabilize such liposomes. Gene transfer in cell culture with KALA might be successful because cellular membranes are negatively charged. In in vitro transfection assays, ppTG1 led to the highest luciferase activity at low charge ratios (between 1 and 2). No great excess of peptide is necessary for reaching high transfection levels, in contrast to other single-component peptide vectors [14–19]. This feature may be correlated to the fact that complexes between ppTG1 and plasmid DNA, formulated at neutral or negative charge ratio, still allowed for efficient induction of calcein release. This is in contrast to observations with JTS1-K13, which showed loss of leakage activity when complexed to plasmid DNA at a charge ratio of [+/–] 0.8 (Table 2), or KALA, which showed a sharp decline of leakage activity on POPC liposomes when the charge ratio of peptide/oligonucleotide complexes was smaller than [+/–] 2.5 [8]. We hypothesize that ppTG1 peptides, in complex with plasmid molecules, preserve structures necessary for membrane-destabilizing activity. Molecular modeling of ppTG1 demonstrated that all lysines segregate on the same side when the structure is an -helix, thus the helix is

111

ARTICLE

doi:10.1006/mthe.2002.0523, available online at http://www.idealibrary.com on IDEAL

amphipathic. The lysine residues could then interact with the negative charges of a B type DNA because the mean distance between two NH3+ is 7.11 Å, similar to the mean distance between two phosphates on the DNA strands (6.55 Å), without destroying the conformation necessary for membrane destabilization. This conservation of an “active structure” can not be postulated for JTS1-K13 and KALA, which might need a layer of free peptides, associated to the plasmid/peptide complexes via hydrophobic interactions, to destabilize membranes. Furthermore, there seems to be a correlation between peptide conformation, liposome leakage, and transfection efficiency. Replacement of leucines in ppTG1 by valines (ppTG32) strongly reduced liposome leakage activity, which we observed with the isoleucine variant (ppTG30) in the presence of DNA. This finding was reflected in an abolished/reduced gene transfer activity of ppTG32 and ppTG30 in vitro and in vivo. Using two complementary biophysical techniques (CD and FTIR), we have shown that the replacement of leucine residues by isoleucines or valines impaired the formation of the -helical structure displacing the folding equilibrium towards a -sheet. It remains to be elucidated how ppTG30, which assumes -sheet conformation in the absence of plasmid DNA, and ppTG21, which exists as random coil, interact with liposomes to induce efficient calcein leakage. KALA exists only partially in -helical conformation [8]. From pH 7 to pH 9, 43% of this peptide exists in -helical conformation, at pH 5, only 24%. We have observed less -helicity in our hands, which might be due to the differences in the experimental conditions (Table 2). In the presence of POPC/POPG liposomes, an increasing percentage of -helicity was observed with decreasing pH [8], which correlated with KALA’s liposome leakage activity. We could thus explain KALA’s transfection capacity by the fact that in a cellular context, membrane components might favor the helical conformation of the peptide and its membranedestabilizing activity. Biophysical studies in the presence of natural membranes would be of interest, as well as conformational studies in the presence of plasmid DNA. The latter type of experiments, however, would be difficult to interpret due to the strong absorption of DNA in the far UV light. Gene transfer with KALA is sensitive to bafilomycin A, a drug that inhibits the acidification of late endosomes. This suggests that such complexes are taken up via the endosomal route. We hypothesize that ppTG1 complexes are also, at least partially, taken up by the endosomal route, but their plasmid-delivery activity would not depend on acidification of the endosomes. This correlates with the observation that ppTG1 induced equal destabilization of POPC/chol liposomes at pH 5 and pH 7. We expect that membrane-destabilizing activity in the cellular system has a role not only in late, acidified endosomes, but also in early endosomes, as well as on other cellular membranes. This suggests that plasmid DNA can be

112

liberated with the help of the pH-insensitive ppTG1 from early endosomes, where DNA is less prone to degradation than in late acidified endosomes. In the direct comparison, ppTG1 was more efficient in gene transfer experiments than JTS1-K13 and KALA. Differences in gene transfer efficiencies may be due to the reduced membrane-destabilizing activity of JTS1-K13 at neutral pH, limiting the escape of plasmid DNA from early endosomes. The same hypothesis can be envisaged for KALA. Moreover, differences with respect to the compaction of plasmid DNA with a peptide carrying positive and negative charges might influence gene transfer efficiency. ppTG1 has shown gene transfer efficiencies at low DNA doses that are higher than those observed for PEI and the dendrimer Superfect. At higher DNA doses, ppTG1 efficiency remains within the range of efficiencies observed for commercially available transfection reagents. So far, none of the single-component peptide vectors described in the literature has proven its gene transfer capacity in vivo [14–19]. In contrast, the peptides ppTG1 and ppTG20 were able to confer gene transfer into the lung of living animals after intravenous injection. The comparison with ppTG32, which lacks liposome leakage activity, has clearly demonstrated the need for both DNAcompacting and membrane-destabilizing activity to achieve successful gene transfer in vivo. One explanation for the success of ppTG1 and ppTG20 could be the low optimal charge ratio of the active complexes, which might be due to the fact that DNA-binding and membrane-destabilizing activities within these peptides do not exclude each other. In the case of gene transfer with ppTG1 complexes, a few animals died after injection of the complexes. The remaining mice per group and experiment, however, showed no or only a few mild clinical signs (level 1 to 2). Like observed for other synthetic vectors, gene transfer was high in the lung and low in the liver. It can be assumed that, as with cationic liposome–DNA complexes [20], mainly endothelial cells are transfected by the peptide/plasmid complexes injected intravenously. Compared with gene transfer with the most efficient synthetic vectors for intravenous administration, DOTAP/chol (1:1) [+/–] 5 [21] and PEI [22], ppTG1, and ppTG20 are 10-fold to 40-fold less efficient. Despite this fact, peptide vectors represent a valuable alternative to artificial lipids and polymers, for instance with respect to their metabolization in the organism. Peptides as components of synthetic vectors have the advantage of being low molecular weight compounds that can be chemically synthesized, as well as easily purified and characterized. It can be envisaged to further improve the membrane-destabilizing activity of ppTG1-derived vectors by introducing subtle changes in the peptide sequence. Additional functional domains may help to target specific cells and organs. Given the ease of formulation of plasmid/peptide complexes, as well as their efficiencies in vitro and in vivo, these

MOLECULAR THERAPY Vol. 5, No. 2, February 2002 Copyright © The American Society of Gene Therapy

doi:10.1006/mthe.2002.0523, available online at http://www.idealibrary.com on IDEAL

new vectors represent effective transfection agents and promising components of nonviral vector formulations for therapeutic gene transfer applications.

MATERIALS

AND

METHODS

Cells. In this study we used the human cell lines HeLa (American Type Culture Collection [ATCC] CCL-2), A549 (CCL-185), and WI-38 (CCL-75). These cells were cultivated in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal calf serum (FCS), 1% gentamycin, 1% glutamine, and 3 g/l glucose in an incubator at 37C and 5% CO2. Plasmids. We used the plasmid pTG11236 (5738 bp; [23]) with a luciferase expression cassette comprising the CMV-IE1 promoter, a short SV40 16S/19S intron, and the SV40 poly(A) signal. This plasmid was purified using the CsCl-ethidium bromide gradient centrifugation method, was tested for the presence of endotoxins (20 endotoxin units/mg), and was quantitated by UV absorbance at 260 nm. Polypeptides. All peptides shown in Table 1A were chemically synthesized by Neosystem (Strasbourg, France). The peptides were assembled by stepwise solid phase synthesis as summarized in Table 1B with the C termini being free acids. The crude peptide preparations were purified by HPLC. Fractions with the desired purity were pooled and lyophilized. The final products were controlled for purity and identity by analytical HPLC, electrospray ionization mass spectrometry. In the case of the cationic peptides the ion exchange was carried out on Dowex 1X2 ion exchange resins, which had been regenerated with acetate. The peptides were delivered as lyophilized powder at a purity of 80–97%. The peptide JTS1-K13 was synthesized in two steps (Table 1B). First, JTS1 with a C-terminal cysteine residue was synthesized (JTS1-Cys), as well as a polylysine (K13) with a N-terminal cysteine residue protected by 3nitro-2-pyridinesulfenyl (NPys-Cys-K13). Second, disulfide bridges were formed between JTS1-Cys and NPys-Cys-K13 as follows: 120 mg of JTS1Cys and 118 mg NPys-Cys-K13, dissolved in 13 ml methanol, respectively, were mixed and incubated at room temperature overnight. The reaction was followed by HPLC analysis and mass spectrometry. Next day the reaction mix was centrifuged and the supernatant evaporated before purification by HPLC. JTS1-K13 was obtained at a purity of 92%. We dissolved the peptides in milliQ water to a final concentration of at least 1 mg/l. The final peptide concentrations were verified by HPLC and quantification of amino acid residues. Preparation of peptide/plasmid complexes-gel retardation assay. We have mixed 15 l of plasmid DNA solution, diluted in 0.9% NaCl to a concentration of 200 ng/l, with peptides, diluted in 15 l of a 0.9% NaCl solution, to result in the desired weight or charge ratios [9]. After 20 minutes at room temperature, 6 l 5 loading buffer (glycerol and bromphenol blue in Tris-acetate-EDTA buffer, pH 8) were added and 10 l of these solutions were analyzed on a 1% agarose gel in the presence of ethidium bromide. To evaluate the reversibility of complex formation, we added 6 l of 5 loading buffer containing 5% SDS (final SDS concentration, 1%). After 5 minutes at room temperature, we analyzed the plasmid–peptide complexes on agarose gels. Liposome leakage assay. We have prepared POPC liposomes by the repeated freeze-thaw method followed by extrusion [24]. POPC (10 moles) in chloroform with or without cholesterol was placed in a glass tube and solvent was evaporated under reduced pressure using a Labconco Rapidvap vortex evaporator (Uniequip, Martinsried, Germany). The resulting lipid film was hydrated with 0.5 ml of an aqueous calcein solution. The lipid suspension was sonicated until the solution became clear using a sonic water bath (BRANSONIC 221, Branson Ultrasonics Corp., Danbury, CT). After five freeze-thaw cycles, we extruded the liposomes by four passages through 200-nm pore diameter polycarbonate membranes (Nuclepore, Costar, Cambridge, MA). Free calcein was removed by gel exclusion chromatography using a Sephadex G-50 column and a 200 mM NaCl, 25 mM Hepes, pH 7.3, solution as the elution buffer. The mean diameter of

MOLECULAR THERAPY Vol. 5, No. 2, February 2002 Copyright © The American Society of Gene Therapy

ARTICLE

liposomes determined by dynamic laser light scattering using a Coulter N4 Plus submicron particle sizer (Coultronics France S.A., Margency, France) was 166 nm. We performed this measurement at 25C at a liposome-suspension concentration of 65 M in 200 mM NaCl and 20 mM Hepes at pH 7.5 within a 3- to 10,000-nm size window with a fixed 90 scattered light angle. The liposome leakage assay was carried out as described [5]: the liposome stock solution was diluted to a lipid concentration of 45 mM in 1.8 assay buffer (360 mM NaCl, 36 mM sodium citrate, pH 5 and pH 7). A 1:5 serial dilution of the test peptide was carried out in a 96-well microtiter plate by transferring 20 l of the peptide solution from one well to the next, diluting with 80 l H2O. We added 100 l liposome solution in each well (final lipid concentration, 25 M; final buffer concentration, 200 mM NaCl, 20 mM sodium citrate, pH 5 and pH 7). After 30 minutes at room temperature, fluorescence was analyzed at 515 nm (excitation, 485 nm) on a microtiter plate fluorescence spectrometer (WALLAC, 1420 multilabel counter Victor). The value for 100% leakage was obtained by addition of 1 l of a 10% Triton X-100 solution. We have plotted the increase of fluorescence as a result of calcein leakage activity against the peptide concentration. Transfection assays and detection of reporter gene expression. We have plated 4  104 cells per well on 24-well plates, cultivated in DMEM supplemented with 10% FCS. The next day, we replaced the medium with 200 l serum-free DMEM adding plasmid/peptide complexes which had been prepared in 30 l 0.9% NaCl solution 20 minutes before. After 2 hours, 1 ml serum-containing DMEM was added. In the case of the luciferase expression plasmid pTG11236, cells were lysed 20 hours later with 100 l cell lysis buffer (Promega), 20 l (= 1/5) of the lysates were analyzed for luciferase activity. Total protein per well was measured using the Pierce BCA protein assay. We have shown the results either as relative light units (RLU) per 1/5 of cell lysate or as RLU/mg protein. To convert RLU per 1/5 of cell lysate into RLU/mg protein, numbers have to be multiplied by a factor of 100. Gene transfer in vitro according to [3] or with the transfection reagents Lipofectin (Gibco BRL), PEI [10], and Superfect (Qiagen) was carried out as recommended by the authors or manufacturers. Briefly, Lipofectin was added to plasmid DNA in a fourfold weight excess in 200 ml serum-free medium. PEI was diluted to a 1 mM solution and, for example, 7.5 l were added to 500 ng DNA in 30 l 5% glucose solution. In the case of Superfect (Qiagen), for instance 12 l of the transfection agent were added to 2 g plasmid DNA in 150 l serum- and antibiotic-free medium, the solution was vortexed for 10 seconds, and 10 minutes later 650 l medium was added. We added 200 l, which corresponds to 500 ng plasmid DNA, to washed cells. All these transfection mixtures were incubated at room temperature for 20 to 30 minutes before being added to the cells. Two hours later, 1 ml serum-containing DMEM was added, and reporter gene expression was analyzed the next day. Transfection in the presence of bafilomycin A. We incubated cells, seeded on 24-well plates, in 200 l serum-free medium in the presence of 175 nM bafilomycin A (Fluka) 30 minutes before adding 30 l of the respective transfection mix. One hour later cells were washed and 1 ml of serumcontaining medium was added. The cells were incubated at 37C overnight before reporter gene expression was measured. Circular dichroism (CD) and Fourier transform infrared (FTIR) studies. Circular dichroic spectra were recorded at 25C on a Jasco-810 spectropolarimeter, calibrated with camphorsulfonic acid, at a peptide concentration of 0.1 mg/ml in 1-mm cuvettes. The spectra were taken in water (pH 5.8) or in 50 mM phosphate buffer at pH 7.5 in a wavelength range between 185 nm and 260 nm. For the FTIR spectroscopy, peptide solutions were spread on fluorinated slides. After evaporation of the water under a nitrogen flux, spectra were recorded on a Bruker IFS 28 spectrophotometer. For each peptide, the results correspond to an average of two scans per two independent experiments (in total four scans). In vivo experiments. We administered peptide/plasmid complexes via tail vein injection to 9-week-old female B6SJLF1 mice (Iffa-Credo, l’Arbresle, France). Each animal (five per group) received 50 g or 60 g of plasmid DNA complexed with peptide in 250 l 5% glucose. Lung and liver were

113

ARTICLE

doi:10.1006/mthe.2002.0523, available online at http://www.idealibrary.com on IDEAL

recovered at the indicated time point after injection, homogenized, and tested for luciferase activity with the Promega Luciferase assay kit. We indicate luciferase activities as RLU standardized against the total amount of protein (Pierce BCA protein assay) resulting in RLU/mg protein. To describe the effects of a formulation on the animal, we have defined five levels of severity of clinical signs (low body temperature, weight loss, animal is prostrate or curled up, tense with convulsions, has piloerections, is swollen, is hyperactive and turns in cycles): level 1, no clinical signs were observed; level 2, a few mild clinical signs were observed; level 3, clinical signs are severe and numerous; level 4, clinical signs are very severe and numerous, animal has to be sacrificed; level 5, animal died.

ACKNOWLEDGMENTS We thank Otmane Boussif (Aventis Pharma, France), Gilles Cauet, Olivier Meyer and François Nicol (Transgene, France), and Annick Thomas (FSA Gembloux, Belgium) for discussion; and Denise Stuber, Pascale Cordier, Didier Elmlinger, Michele Klein, and Benoit Heller (Transgene, France) for technical assistance. RECEIVED FOR PUBLICATION JUNE 6; ACCEPTED NOVEMBER 29, 2001.

REFERENCES 1. Ledley, F. D. (1995). Nonviral gene therapy: the promise of genes as pharmaceutical products. Hum. Gene Ther. 6: 1129–1144. 2. Mahato, R. I., Monera, O. D., Smith, L. C., and Rolland, A. (1999). Peptide-base gene therapy. Curr. Opin. Mol. Ther. 1: 226–243. 3. Gottschalk, S., et al. (1996). A novel DNA-peptide complex for efficient gene transfer and expression in mammalian cells. Gene Ther. 3: 448–457. 4. Lear, J. D., and Grado, W. F. (1987). Membrane binding and conformational properties of peptides representing the NH2 terminus of influenza HA-2. J. Biol. Chem. 262: 6500–6505. 5. Plank, C., Oberhauser, B., Mechtler, K., Knoch, C., and Wagner, E. (1994). The influence of endosome-disruptive peptides on gene transfer using synthetic virus-like gene transfer systems. J. Biol. Chem. 269: 12918–12924. 6. Subbarao, N. K., Parente, R. A., Szoka, F. C., Nadasdi, L., and Pongracz, K. (1987). pH dependent bilayer destabilization by an amphipathic peptide. Biochemistry 26: 2964–2972. 7. Parente, R. A., Nir, S., and Szoka, F. C. (1988). pH-dependent fusion of phosphatidylcholine small vesicles. J. Biol. Chem. 263: 4724–4730.

114

8. Wyman, T. B., et al. (1997). Design, synthesis and characterization of a cationic peptide that binds to nucleic acids and permeabilizes bilayers. Biochemistry 36: 3008–3017. 9. Felgner, P. L., et al. (1997). Nomenclature for synthetic gene delivery systems. Hum. Gene Ther. 8: 511–512. 10. Boussif, O., et al. (1995). A versatile vector for gene and oligonucleotide transfer into cells in culture and in vivo: polyethylenimine. Proc. Natl. Acad. Sci. USA 92: 7297–7301. 11. Bowman, E. J., Siebers, A., and Altendorf, K. (1988). Bafilomycins: a class of inhibitors of membrane ATPases from microorganisms, animal cells, and plant cells. Proc. Natl. Acad. Sci. USA 85: 7972–7976. 12. Cornut, I., Büttner, K., Dasseux, J.-L., and Dufourcq, J. (1994). The amphipathic -helix concept. Application to the de novo design of ideally amphipathic Leu, Lys peptides with hemolytic activity higher than that of melittin. FEBS Lett. 349: 29–33. 13. Zabner, J., Fasbender, A. J., Moninger, T., Poellinger, K. A., and Welsh, M. J. (1995). Cellular and molecular barrieres to gene transfer by cationic lipid. J. Biol. Chem. 270: 18997–19007. 14. Zhang, F., et al. (1999). A transfecting peptide derived from adenovirus fiber protein. Gene Ther. 6: 171–181. 15. Morris, M. C., Vidal, P., Chaloin, L., Heitz, F., and Divita, D. (1997). A new peptide vector for efficient delivery of oligonucleotides into mammalian cells. Nucleic Acids Res. 25: 2730–2736. 16. Morris, M. C., Chaloin, L., Méry, J., Heitz, F., and Divita, D. (1999). A novel potent strategy for gene delivery using a single peptide vector as a carrier. Nucleic Acids Res. 27: 3510–3517. 17. Kichler, A., et al. (2000). Efficient DNA transfection mediated by the C-terminal domain of human immunodeficiency virus type 1 viral protein R. J. Virol. 74: 5424–5431. 18. Fominaya, J., et al. (2000). An optimized amphiphilic cationic peptide as an efficient nonviral delivery vector. J. Gene Med. 2: 455–464. 19. Haines, A. M. R., et al. (2001). CL22—a novel cationic peptide for efficient transfection of mammalian cells. Gene Ther. 8: 99–110. 20. McLean, J. W., et al. (1997). Organ-specific endothelial cell uptake of cationic liposomeDNA complexes in mice. Am. J. Physiol. 273: H387–404. 21. Song, Y. K., Liu, F., Chu, S., and Liu, D. (1997). Characterization of cationic liposomemediated gene transfer in vivo by intravenous administration. Hum. Gene Ther. 8: 1585–1594. 22. Goula, D., et al. (1998). Polyethylenimine-based intravenous delivery of transgenes to mouse lung. Gene Ther. 5: 1291–1295. 23. Meyer, O., Schughart, K., Pavirani, A., and Kolbe, H. V. J. (2000). Multiple systemic expression of human interferon- in mice can be achieved upon repeated administration of optimized pcTG90-lipoplex. Gene Ther. 7: 1606–1611. 24. Olson, F., Hunt, C. A., Szoka, F. C., Vail, W. J., and Papahadjopoulos, D. (1979). Preparation of liposomes of defined size distribution by extrusion through polycarbonate membranes. Biochim. Biophys. Acta 557: 19–23.

MOLECULAR THERAPY Vol. 5, No. 2, February 2002 Copyright © The American Society of Gene Therapy