The Nuclear Pore Complex Is Involved in Nuclear Transfer of Plasmid DNA Condensed with an Oligolysine–RGD Peptide Containing Nuclear Localisation Properties

Home , Nuclear pore

Gene Therapy (2001) 8, 1643–1653  2001 Nature Publishing Group All rights reserved 0969-7128/01 $15.00 www.nature.com/gt RESEARCH ARTICLE The nuclear pore complex is involved in nuclear transfer of plasmid DNA condensed with an oligolysine–RGD peptide containing nuclear localisation properties

M Colin1, S Moritz1, P Fontanges2, M Kornprobst1, C Delouis3, M Keller4, AD Miller4, J Capeau1, C Coutelle5 and MC Brahimi-Horn1 1INSERM U402, Faculte´ de Me´decine Saint-Antoine, Paris, France; 2Service d’Imagerie Cellulaire IFR 65, Hoˆpital Tenon, Paris, France; 3UMR 955 INRA-ENVA, Paris, France; 4Department of Chemistry, Imperial College, London, UK; and 5Section of Molecular Genetics, Division of Biomedical Sciences, Imperial College, School of Medicine, London, UK

One of the major barriers to efﬁcient gene transfer and energy-dependent process involving the nuclear pore com- expression of nonviral vectors for gene therapy is passage plex, since it is inhibited at 4°C and by treatment with wheat across the nuclear envelope. We have previously shown that germ agglutinin or with an antibody to the nuclear pore com- an oligolysine–RGD peptide that condenses plasmid DNA plex which all block nuclear protein import. In accordance and binds to cell surface integrins can mediate increased with active nuclear transport, we have shown that all these internalisation of plasmid DNA into cells and synergistic treatments inhibit expression of a luciferase reporter plasmid enhancement of gene expression when complexed to a cat- in permeabilised cells. Nuclear transfer of pDNA is ionic lipid. In this report, we show that this enhancement is enhanced in mitotic cells, but cell division is not a prerequi- due to increased nuclear transfer of the plasmid DNA. We site for transfer. We propose that the oligolysine–RGD pep- have applied the digitonin-permeabilised cell system that tide acts as a nuclear localisation signal and that the cationic has been well established for the study of the nuclear trans- lipid is more important for cell entry and endosome destabil- port of proteins to examine the nuclear transfer of plasmid isation than nuclear transfer. Gene Therapy (2001) 8, DNA. Nuclear transfer of plasmid DNA complexed to an olig- 1643–1653. olysine–RGD peptide and lipofectamine appears to be an

Keywords: cationic lipid; gene transfer; integrin; in vitro transfection; nuclear transfer; plasmid/liposomes

Introduction see Refs 19 and 20). Active protein transport requires interaction of specific nuclear localisation signal (NLS) Nonviral vectors for gene therapy enter cells by endo- sequences on proteins with importins, that in turn inter- cytosis and traffic via endosomes before being released act with the protein RanGDP and carry cargo into the 1 into the cytoplasm for nuclear transfer. One of the major nucleus. These amino acid sequences characteristically barriers to efficient gene transfer is the nuclear envelope bare a strong positive charge. An enhancement of nuclear that selectively controls passage of macromolecules from transfer and/or expression of pDNA in the presence of the cytoplasm to the nucleus. Nuclear transport of oligon- a NLS, either covalently5,10,12,21 or noncovalently2,3,6,16,22–26 ucleotides and plasmid DNA (pDNA) has been studied associated with the pDNA was found, which together 1–5 following: (1) transfection of whole cells; (2) microinjec- with the observed interaction of NLS-pDNA constructs 6–13 tion into the cytoplasm; and (3) transfection of plasma with the α-importin NLS nuclear import receptor3,21 sup- 12,14–17 membrane-permeabilised cells, with the aim of ports the hypothesis of an involvement of the NPC determining the mechanism of transfer. Although the machinery in the nuclear transfer of pDNA. However, it nuclear transport of pDNA is less well understood than has also been observed that transfection is optimal in 5,7,13,18 nuclear protein transport a number of these studies actively proliferating cells and thus it has been suggested suggest that transfer occurs via the nuclear pore complex that exogenous pDNA enters the nucleus during mitosis (NPC) similar to protein import and export. The NPC, a when breakdown of the nuclear envelope occurs.27–30 multiprotein complex, allows for active, energy-depen- The majority of studies into nuclear transport of pDNA dent, or passive transport of macromolecules (for review, have been performed with either naked pDNA or a NLS- pDNA, while only a few studies have examined the mechanism of transfer of pDNA in the presence of a cat- 11,31 Correspondence: MC Brahimi-Horn, CNRS UMR 6543, Centre A Lacas- ionic polymer or liposome. Yet the latter methodology sagne, 33 Avenue de Valombrose, 06189 Nice cedex, France is the most widely used for in vitro transfection. Received 22 November 2000; accepted 22 August 2001 We and others have previously demonstrated the prin- Nuclear transfer of plasmid DNA M Colin et al 1644 ciple of integrin-mediated gene transfer using a bifunctional oligolysine–RGD peptide which condenses pDNA and binds to RGD-recognising integrins on the cell surface.32–35 In addition, we have demonstrated that the presence of a liposome enhances substantially the level of gene expression of this construct.1,32,36 It was, however, not clear whether this enhancement is entirely due to the targeting function of the oligo-peptide or if it may also be able to confer a nuclear targeting function by virtue of its strong positive charge. In this study, we set out to investigate the nuclear transport of the pDNA when complexed to an oligolysine–RGD peptide and/or liposome and to examine the role of the NPC in this transfer. We chose to use digitonin-permeabilised cells to examine nuclear transport since this cell system has proven to be invaluable in understanding nuclear protein import and export.37 The detergent digitonin selectively permea- bilises membranes rich in cholesterol, such as the plasma membrane, while leaving intact the membranes of intracellular organelles with low levels of cholesterol, such as the nucleus37 and secretory granules.38 This system thereby allows direct access of various agents to the nucleus for the study of nuclear transfer and unlike microinjection a large population of cells can be examined. In addition, we have examined the role of the cell cycle in Figure 1 Confocal microscopy of digitonin-permeabilised cells showing nuclear transfer with the aim of determining whether dis- intact nuclei. Human tracheal cells were permeabilised with digitonin for mantling of the nuclear envelope during mitosis makes 5 min on ice, rinsed and incubated in the presence of FITC-dextran for 30 min at 37°C. Cells were fixed and nuclei were stained with propidium pDNA transfer more efficient. iodide. A merged image of the two fluorophores (FITC, green and propid- The results presented here indicate that both the oligo- ium iodide, red) is shown. The bar represents 18 ␮m. lysine–RGD and –RGE peptides are able to mediate nuclear transfer of pDNA much more efficiently than oligolysine on its own, indicating that the charge is not the confirmed by serial sectioning at the level of the nucleus main factor involved in this function. We also found that using a confocal microscope.

the liposome lipofectamine promotes nuclear transfer in When FITC-labelled [K]16RGD (Figure 2a) or [K]16RGE digitonin-permeabilised cells, but not in cells microin- (Figure 2b) peptides were incubated with permeabilised jected with the complex. Despite permeabilisation, lipo- cells the label (green) was observed in the nucleus within fectamine may enhance cell entry and thus increase the 1 h by confocal sectioning of several cells at the nuclear amount of pDNA that comes in contact with the nucleus. level. The FITC-[K]16 peptide also transferred to the In digitonin-permeabilised cells oligolysine–RGD and nucleus, but the label appeared to be less intense (Figure –RGE act synergistically with lipofectamine to enhance 2c). To test if the oxidised/reduced state of the peptide nuclear transfer of pDNA. In addition, the mechanism affected nuclear transfer FITC-labelled oxidised (cyclic) of [K]16RGD/RGE nuclear transfer of pDNA appears to or reduced (non-cyclic) peptides were examined for involve the nuclear pore complex machinery. Further- nuclear transfer (Figure 3). However, both peptides more, this study also indicates that although nuclear appeared to enter the nucleus within 30 min. When Cy5- transfer is more efﬁcient in cells undergoing mitosis, labelled pDNA (pseudo-blue) alone was added to digi- substantial nuclear translocation also occurs in non- tonin-permeabilised cells, no label was detected in nuclei dividing cells. counterstained with propidium iodide (red) (Figure 4a). When the Cy5-pDNA was condensed with the [K]16RGD peptide, only a low level of label was seen in nuclei Results (Figure 4b). When lipofectamine was complexed with Cy5-pDNA label was observed in the nuclei (Figure 4c) The oligolysine–RGD peptide enhances transfer of and addition of lipofectamine to the Cy5- pDNA to the nucleus pDNA:[K]16RGD complex clearly enhanced the amount To examine nuclear transport of pDNA we adopted the of pDNA in the nucleus (Figure 4d). Under these con-

well-established digitonin-permeabilised cell system ditions colocalisation of Cy5-pDNA and FITC-[K]16RGD which has been extensively used in the study of the was observed in the nucleus as shown by the white signal nuclear transport of proteins.37,39 To verify that the resulting from the merger of red, blue and green (Figure nuclear membrane of the human tracheal cells used in 5a). Colocalisation of Cy5-pDNA and FITC-[K]16RGE this study remains intact during permeabilisation with (Figure 5b) or FITC-[K]16 (Figure 5c) was also observed digitonin the nuclei of permeabilised cells were tested for in the nucleus after complexing with lipofectamine. How- their capacity to exclude FITC-labelled dextran of 70 kDa ever, the proportion of the two fluorophores varied giv- (Figure 1). No FITC label (green) could be detected by ing either a yellow (green and red) or mauve (blue and confocal microscopy in nuclei counterstained with the red) signal. Examination of the individual fluorophore DNA binding dye propidium iodide (red), thus indicat- images (ie non-merged) also allowed confirmation of col- ing that the nuclear membrane remains intact. This was ocalisation. No Cy5-pDNA and only a limited amount of

Gene Therapy Nuclear transfer of plasmid DNA M Colin et al 1645 peptide was observed in the nucleus when complexed with FITC-[K]16 without lipofectamine (Figure 5d). These results correlate with the results for the luciferase activity levels in non-permeabilised cells, where the presence of both the [K]16RGD or [K]16RGE peptides with lipofectamine led to the highest expression and where the presence of the [K]16 peptide did not signiﬁcantly modify the luciferase expression level mediated by lipofectamine

(Figure 5e). In the absence of lipofectamine, the [K]16RGD peptide enhanced gene transfer by speciﬁc interaction with cell surface integrins since the mutated peptide [K]16RGE which does not bind integrins did not increase the level of gene transfer above that of pDNA alone.32,33 However, in the presence of lipofectamine a loss of speci- ﬁcity is observed since the [K]16RGE peptide gives equiv- 32 alent levels of luciferase expression as does [K]16RGD. The observed synergistic effect on expression of both

[K]16RGD and [K]16RGE in the presence of lipofectamine suggested to us that both peptides may be acting as a NLS and thereby enhancing gene expression by increasing nuclear transfer. Examination of the two peptide sequences on the protein data base PSORT II (prediction of protein sorting signals and localisation sites in amino acid sequences) predicts their nuclear localisation and

thus supports this hypothesis. The basic amino acids [K]16 and R (arginine) in the peptides would together form a bipartite NLS. As lipofectamine, in contrast to the [K]16 RGD and [K]16 RGE peptides does not enter the nucleus in a detectable amount,1 we suggest that the mechanism by which it enhances nuclear transfer on its own is different from that of the two peptides. The overall luciferase expression levels were higher in digitonin-permeabilised compared with non-permeabilised cells (Figure 6). This appears to be primarily due to an increase in the amount of pDNA, that enters cells and thus comes in contact with the nucleus irrespective of its complexing with a peptide. However, importantly the mean expression levels in permeabilised cells for the

[K]16 RGD and [K]16 RGE peptides is higher than for [K]16 alone again suggesting that the RGD and RGE motifs enhance nuclear transfer. When the pDNA:[K]16 RGD complex was microinjected into the cytoplasm of cells, no label was observed in the nucleus at the control zero time (T0) but label was seen in the nucleus after 2 h (T 2h), whether or not the complex was prepared with lipofectamine (Figure 7). The label appeared to remain in a compact form within the nucleus. However, not all nuclei of microinjected cells took up the label (data not shown). No label was observed in the nuclei of cells microinjected with Cy5- labelled pDNA alone. These results suggest that the peptide is the major determining factor in nuclear transfer, but they appear to be in conflict with the data presented in Figure 4, which suggest that lipofectamine favours nuclear transfer. It may be that in digitonin-permeabilised cells lipofectamine enhances cell entry and protection of the pDNA and thus brings increased amounts in contact with the nucleus. It is, however, hazardous to draw quantitative conclusions from confocal images. It Figure 2 Nuclear transfer of peptides [K]16RGD, [K]16RGE and [K]16. Digitonin-permeabilised cells were incubated for 1 h at 37°C in the pres- should also be noted that due to the small amount of ence of (a) FITC-[K]16RGD, (b) FITC-[K]16RGE or (c) FITC-[K]16. Cells sample injected into the cytoplasm of cells (about 1 pl) were fixed and nuclei were stained with propidium iodide. Merged confocal the amount of complex that comes in contact with the images of the two fluorophores (FITC, green and propidium iodide, red) ␮ ␮ ␮ nucleus is potentially lower in microinjected cells than are shown. The bar represents 43 m (a), 30 m (b) and 26 m (c). in permeabilised cells and the complex is presented in a microdrop form rather than as a diffuse solution enrob-

Gene Therapy Nuclear transfer of plasmid DNA M Colin et al 1646

Figure 3 Cyclic and non-cyclic peptides enter the nucleus. (a) Cells pretreated with DTT and incubated in the presence of FITC-dextran; (b) cells

incubated in the presence of cyclic (oxidised) FITC-[K]16RGD peptide; and (c) cells incubated in the presence of non-cyclic (reduced with DTT) FITC- ␮ ␮ [K]16RGD peptide. The bar in (a) represents 40 m and 20 m in (b and c).

Figure 4 Nuclear transfer of pDNA is enhanced in the presence of peptide and lipofectamine. (a–c) Cells were permeabilised with digitonin and incubated for 1 h at 37°C with either (a) Cy5-pDNA alone, (b) Cy5-

pDNA:[K]16RGD, (c) Cy5-pDNA:lipofectamine or (d) Cy5- pDNA:[K]16RGD:lipofectamine. (a) Nuclei were stained with propidium iodide and a merged confocal image of the two ﬂuorophores (Cy5, blue and propidium iodide, red) is shown. (b–d) The phase contrast images are merged with the Cy5 ﬂuorescence images. The bar represents 84 ␮m.

ing the nucleus. In conclusion, we suggest that lipofectamine is more important for cell entry and protection of the pDNA, and possibly destabilisation of endosomes, than for nuclear transfer.

Nuclear transfer of complexed pDNA involves the nuclear pore complex Figure 5 Nuclear translocation of pDNA in the presence of lipofectamine Nuclear transfer of NLS-containing proteins is inhibited is dependent on the [K] RGD/RGE sequence and not the [K] motif of 37 16 16 at 4°C, and is blocked by agents which bind the nuclear the peptide. (a–d) Cells were permeabilised with digitonin and incubated ° pore complex such as the lectin wheat germ agglutinin 1hat37C with (a) Cy5-pDNA:FITC-[K]16RGD:lipofectamine, (b) Cy5- 37,39 (WGA), or specific antibodies to a nuclear pore com- pDNA:FITC-[K]16RGE:lipofectamine, (c) pDNA:FITC-[K]16:lipofectam- plex protein.40 To determine if nuclear transfer of oligoly- ine or (d) with Cy5-pDNA:FITC-[K]16. After fixation, nuclei were stained sine-RGD:lipofectamine complexed Cy5-pDNA was also with propidium iodide. Merged confocal images of the three fluorophores (Cy5, blue; FITC, green; and propidium iodide, red) are shown. The bar inhibited by such treatments permeabilised cells were represents 70 ␮m. (e) Non-permeabilised cells were transfected with the either maintained at 37°C (control) (Figure 8a) or at 4°C indicated components (ratio 1:5:24) and the luciferase activity determined (Figure 8b) or they were pretreated with either WGA 48 h later.

Gene Therapy Nuclear transfer of plasmid DNA M Colin et al 1647 observed in the nucleus when cells were treated with an antibody to the nuclear pore complex (yellow) (Figure 8d). Since quantification of the fluorescence of confocal images is not very reliable, we attempted to confirm and quantify this inhibition by determining the luciferase activity expressed in digitonin-permeabilised cells. We hypothesised that since the protein luciferase localises to peroxisomes,41 it may not leak from permeabilised cells as do small macromolecules like ATP and cytoplasmic proteins such as lactate dehydrogenase (LDH) and thus synthesis of luciferase and detection of its activity could be possible. Permeabilised cells transfected at 4°C showed significantly reduced luciferase expression levels compared with 37°C (Figure 9a). Incubation of non-permeabilised control cells in the presence of WGA or an antibody to the nuclear pore complex resulted in a slight decrease in the level of expression (Figure 9b). However, permeabilisation also allows entry of the inhibitors into the cell and incubation of permeabilised cells in the presence of WGA or an antibody to the nuclear pore complex resulted in an approximate five-fold and three-fold decrease in the level of expression, respectively, com- Figure 6 Luciferase expression levels are higher in digitonin permea- pared with non-permeabilised treated cells. To verify that bilised cells. Cells were either not permeabilised (open bars) or permea- digitonin treatment did indeed result in permeabilisation bilised (full bars) with digitonin and incubated for 4 h at 37°C with either of cells the release of ATP and LDH at different digitonin pDNA (pGL3) alone, pDNA:[K]16, pDNA:[K]16RGE or pDNA:[K]16RGD concentrations was determined. Digitonin at the concen- in the presence of lipofectamine (ratio 1:5:24) and the luciferase activity tration of 20 ␮g/ml resulted in the release of a substantial determined 48 h later. amount of ATP (Figure 9c) and LDH (Figure 9d) into the extracellular medium. Release of LDH at high digitonin (Figure 8c) or a nuclear pore complex antibody (Figure concentrations did not reach 100% probably due to the 8d) before transfection and analysis by confocal presence in cells of LDH in subcellular compartments, microscopy. The 37°C control showed colocalisation of such as mitochondria.42 In addition, de novo protein syn- peptide and pDNA in the nucleus (Figure 8a), but no lab- thesis is already detected 2 h after permeabilisation, eled pDNA or peptide could be seen in the nucleus when although at a reduced level as determined by incorpor- cells were maintained at 4°C (Figure 8b). No peptide or ation of [35S]-methionine into protein (Figure 9e) and 48 pDNA was observed in nuclei when cells were pretreated h later the number of non-viable cells for non-permea- with WGA (Figure 8c). No pDNA but some peptide was bilised and permeabilised cells was 9.0 ± 1.0 and 14.3 ±

Figure 7 Microinjected pDNA: [K]16RGD complexes enter the nucleus in the absence and presence of lipofectamine. Cy5-labelled pDNA was complexed + with FITC-labelled [K]16RGD together without (-LM) or with lipofectamine ( LM) (ratio 1:5:18) and microinjected into the cytoplasm of cells which ° were then incubated for zero (T0) and 2 h (T 2h) at 37 C. Cells were incubated with pGL3: [K]16RGD for 0 min (a) or 2 h (b, c-1 and c-2) or with pGL3:[K]16RGD:lipofectamine for 0 min (d) or 2 h (e-1, e-2 and f) and then fixed and stained with the nuclear dye propidium iodide. a, b, c-2, d, e-2, and f are dark field confocal images showing the merged Cy5, FITC and propidium iodide fluorescence. c-1 and e-1 are phase contrast images superimposed on the Cy5 fluorescence. c-1 and c-2 are images of the same field as are e-1 and e-2. The bars represent 20 ␮m.

Gene Therapy Nuclear transfer of plasmid DNA M Colin et al 1648

Figure 8 Nuclear transport of pDNA is an active process and requires the nuclear pore complex. Permeabilised cells, maintained at 37°C (a) or at 4°C (b) or pretreated at room temperature for 30 min with WGA (50 ␮g/ml) (c) or with an antibody to the nuclear pore complex (d) were incu-

bated for 1 h with Cy5-pDNA:FITC-[K]16RGD:lipofectamine. After ﬁx- ation, nuclei were stained with propidium iodide. Merged confocal images of the three ﬂuorophores (Cy5, blue; FITC, green; and propidium iodide, red) were obtained. The bar represents 70 ␮m.

Figure 9 Binding of WGA or of an antibody to the nuclear pore complex inhibits luciferase expression. (a) Digitonin-permeabilised cells were trans- 1.5, respectively. This result and the measurements of ° luciferase expression suggest that digitonin momentarily fected at either 37 or 4 C with pDNA:[K]16RGD:lipofectamine (ratio 1:5:18) and the luciferase activity determined after a 48-h post-transfection forms pores in the plasma membrane allowing for entry period; (b) Digitonin-permeabilised (P) and non-permeabilised (NP) cells of different agents, but that the cells recommence protein were incubated with or without WGA or an antibody to the nuclear pore

synthesis which reaches a normal level 48 h after the complex, transfected with pDNA:[K]16RGD:lipofectamine and the lucifer- transfection period when the luciferase activity is deter- ase activity determined; (c) Release of ATP from cells permeabilised with mined. To further confirm the presence of the luciferase increasing concentrations of digitonin. (d) Release of lactate dehydrogen- protein in permeabilised cells, fluorescence immunode- ase from cells permeabilised with increasing concentrations of digitonin; (e) Measurement of protein synthesis with [35S]-methionine during the tection of the protein was performed (Figure 9f). Micro- period 2 h after permeabilisation (P) and non-permeabilised (NP) cells; (f) scopic examination showed the presence of the luciferase confocal microscopy showing the immunolocalisation of the protein lucifer- protein in vesicles that were predominantly perinuclear. ase in permeabilised cells. The image shows the superposition of the FITC- In addition, the luciferase protein was detected in the labelled anti-luciferase antibody image and the phase contrast image of the ␮ large majority of cells thus indicating that the detected cells. The bar represents 70 m. luciferase activity was not due to just a few highly expressing cells that had not been permeabilised. fold increase in the luciferase expression level compared Optimal expression in mitotic cells with non-synchronised control cells (Figure 10b, left), To examine the role played by the cell cycle in nuclear while cells in the G0/G1 phase showed a lower but none- transfer, cells were incubated in the presence of either theless considerable level of expression similar to those of demecolcine, which blocks cells in the G2/M phase, or a non-synchronised control (Figure 10b, right). Confocal hydroxyurea, which blocks cells in the G0/G1 phase. microscopy analysis showed the presence of labelled FACS analysis was used to determine the percentage of pDNA in the nuclei of cells treated either with demecol- cells in the respective cycle phases. In the presence of cine (Figure 10c, left) or hydroxyurea (Figure 10c, right) demecolcine about 60% of the cells were in the G2/M when transfected with the Cy5pDNA:[K]16RGD: phase (Figure 10a, left) and in the presence of hydroxyu- lipofectamine complex. These results suggest that the dis- rea about 60% of the cells were in the G0/G1 phase mantling of the nuclear membrane during mitosis fav- (Figure 10a, right). In fact, half the control cells at 80% ours nuclear transfer, but that transfer also occurs when confluence were already in the G0/G1 phase. Treated the nuclear envelope is intact and that mitosis plays a cells were transfected in the presence of the respective relatively limited role in the entry of the [K]16 RGD/E drug that was then removed and the luciferase activity pDNA complexes into the nucleus. was determined after the 48-h post-transfection period. We tested the possibility that an increase in expression Cells in the G2/M phase showed an approximate four- observed in cells in mitosis was due to an increase in the

Gene Therapy Nuclear transfer of plasmid DNA M Colin et al 1649

Figure 11 The rate of uptake of plasmid DNA by cells in mitosis is lower than by cells in G0/G1. Control cells and cells treated with hydroxyurea or demecolcine were exposed to [35S]-labelled plasmid DNA complexed to

[K]16RGD:lipofectamine (ratio1:5:18) for the indicated times and the cell- associated radioactivity determined.

nine. Sebestyeń et al12 covalently attached this NLS to pDNA (up to 11 kb) and observed nuclear transfer of the pDNA in digitonin-permeabilised cells when the pDNA was conjugated with a high ratio of NLS peptide. Trans- fer was inhibited at 4°C and in the presence of wheat Figure 10 Luciferase expression and nuclear transfer in dividing and germ agglutinin (WGA) which is known to inhibit NLS- non-dividing cells. (a) The percentage of cells in the G2/M and G0/G1 protein import through the nuclear pore complex, thus phases was determined by FACS analysis in control cells and cells treated suggesting that nuclear transfer involves an energy- with either demecolcine (0.06 ␮g/ml, left) or hydroxyurea (2.5 mM, right). ␮ dependent active transport through the nuclear pore. The (b) Luciferase activity of cells pretreated with demecolcine (0.06 g/ml, finding that NLS-pDNA constructs can interact with α- left) or with hydroxyurea (2.5 mM, right) transfected with importin, an NLS nuclear import receptor3,21 supports pDNA:[K]16RGD:lipofectamine. (c) Confocal microscopy of permeabilised cells pretreated with demecolcine (0.06 ␮g/ml, left) or hydroxyurea (2.5 this hypothesis. However, no nuclear transfer of this and mM, right) and transfected 1 h with Cy5-pDNA:[K]16RGD:lipofectamine. another NLS–DNA conjugate could be observed after These confocal images are the superposition of the Cy5-labelled pDNA their microinjection into the cytoplasm of cells, possibly (blue) and phase contrast images of cells. The bar represents 64 ␮m (left) 12,21 ␮ as a result of cytoplasmic sequestration. Yet, other and 70 m (right). covalent NLS-pDNA constructs have been shown to be present in nuclear fractions of zebrafish embryos after uptake of the complex by cells and not an increase in microinjection into the cytoplasm,6 and to result in an nuclear transfer. To this aim the uptake of [35S]-labelled increase in expression levels in cells transfected in the 5,21 pDNA complexed to [K]16RGD:lipofectamine was exam- presence of either a cationic lipid or polymer. ined in control cells and cells treated with hydroxyurea Non-covalent association of an NLS with the pDNA or demecolcine. Since the cell-associated radioactivity for has also been reported to enhance expression levels in demecolcine-treated cells was lower than for either con- cells.3,6,16,22,25,43 In addition, using a bifunctional peptide trol or hydroxyurea-treated cells, we conclude that the nucleic acid (PNA)-NLS, where PNA forms duplex increase in expression levels in mitotic cells was due to hybrids with complementary DNA, Brandeń et al2 dem- an increase in the nuclear transfer and not cell uptake onstrated increased association of pDNA with the (Figure 11). It should be pointed out that cell-associated nucleus and expression in cells transfected in the pres- radioactivity includes internalised complex, but also ence of polyethylenimine (PEI). Thus the results concern- complex that may remain bound to the cell surface. ing covalent and non-covalent association suggest that NLS binding to DNA and subsequent enhancement does Discussion not require a cross-linking agent and covalent binding may even be detrimental in abolishing transcription, in Inspired by the knowledge that protein transport into the particular when many NLS molecules are present per nucleus requires specific amino acid sequences referred pDNA.12 to as nuclear localisation signals (NLS), several attempts To determine whether the previously observed to enhance nuclear transfer of plasmid DNA (pDNA) by increase in expression and the increase in the amount of covalent5,10,12,21 or non-covalent association2,3,6,16,22–26 of pDNA in isolated nuclear fractions in the presence of pDNA with an NLS have been reported. The most fre- peptide and/or liposome1,33 is due in part to an increase quently employed NLS is the classic monopartite in the transfer of the pDNA to the nucleus, we examined sequence of the large T antigen of SV40 that is composed the nuclear transfer of the pDNA in digitonin-permea- of a single stretch of basic amino acids; lysine and argi- bilised cells with or without the oligolysine–RGD peptide

Gene Therapy Nuclear transfer of plasmid DNA M Colin et al 1650 and/or liposome. The results of this study demonstrate diffusion channels (approximately 10 nm in diameter).20 that pDNA (approximately 5 kb) alone does not transfer Diffusion of macromolecules appears to be limited to a to the nucleus of digitonin-permeabilised or microin- size of approximately 50 kDa, thus peptides of about 3

jected human tracheal cells, while transfer occurs when kDa such as [K]16RGD can theoretically diffuse into the [K]16RGD or the liposome lipofectamine are complexed nucleus. However, the size of the [K]16RGD:pDNA com- by non-covalent interaction with pDNA. In addition, the plex, estimated to possess a diameter of about 20–100 peptide and the liposome appear to act synergistically for nm,33 would exclude transport by diffusion. The addition nuclear transfer of the pDNA, as suggested previously of lipofectamine, estimated to have a mean diameter of by results of subcellular fractionation experiments.1 The about 90 nm46 would also increase the overall size, observed synergistic increase in luciferase expression though it is possible that this component of the complex mediated by the [K]16RGD or [K]16RGE peptides, but not does not enter the nucleus. Our previous studies into the the [K]16, in the presence of an excess amount of liposome localisation of lipofectamine by electron microscopy in also suggested to us that the peptides may be acting as an the cell did not reveal its presence in the nucleus.1 Thus NLS and thereby enhancing nuclear transfer. Increased the sizes of these components suggest that transfer occurs expression in digitonin-permeabilised cells in the pres- via the central energy-dependent channel unless the ence of the RGD/RGE peptides further supports this pDNA is capable of threading through the NPC in a hypothesis. Examination of the peptide sequences on the worm-like manner as previously suggested.5 protein data base PSORT II predict their nuclear localis- We were also interested in examining the relationship ation and suggest that they are bipartite NLS-like basic between the cell cycle and nuclear transfer since break- peptides. down of the nuclear membrane during mitosis may open The mechanism by which [K]16RGD/E peptides per- the way to nuclear transfer. Several studies have demon- mits nuclear transfer of pDNA appears to involve the strated that cells arrested in the G1 phase show reduced nuclear pore and to be an active process, since transport gene expression after transfection with either a cationic is inhibited at 4°C and under conditions that block trans- lipid or polycation,8,13,29 while cells in mitosis show port through the nuclear pore complex. Thus these increased expression.27,30 Nonetheless significant gene results suggest that the peptide acts as an NLS and that expression was observed in non-dividing cells8 and cell the bipartite nature of the [K]16RGD peptide, ie the [K]16 division was not required for expression of a pDNA com- motif and the arginine residue are necessary since the plex to PEI when injected into the cell cytoplasm.5 Studies 27 47 [K]16 peptide appears to have a lower affinity for the by Brunner et al and Escriou et al showed substantial nucleus and does not enhance expression levels. Chan increases in expression (30–500-fold) for mitotic cells and Jans3 also found that polylysine in contrast to a compared with cells in the G1 phase when transfected polylysine/NLS conjugate does not increase nuclear with a cationic lipid or polymer. However, when an import. Godbey et al31 showed that the polycation PEI adenoviral system was used no substantial enhancement which enhances gene expression undergoes nuclear local- in expression (four-fold) was observed in mitotic cells. isation when added to cells with or without pDNA and This suggests that when an efficient nuclear entry Pollard et al11 showed that this polymer and polylysine machinery is present as in the case of the adenovirus, the promoted gene delivery from the cytoplasm to the breakdown of the nuclear membrane does not influence nucleus in microinjected cells. An alternative explanation the level of transfer.27 Our studies confirm the findings for the peptide-induced transfer observed in this study obtained with nonviral systems and suggest that the dis-

may be that the substantial amount of [K]16RGD/E pep- mantling of the nuclear envelope favours nuclear trans- tides in the nucleus modiﬁes the characteristics of the fer, but is not a prerequiste for transfer in particular if an nucleus which in turn favours nuclear penetration of efﬁcient NLS-like mechanism which employs the nuclear the pDNA. entry machinery is present. The mechanism by which the liposome lipofectamine In conclusion, results presented here indicate that: (1)

enhances nuclear transfer in this study is yet to be the peptide [K]16RGD/RGE enhances nuclear transfer of deﬁned. It is possible that the hydrophobic and/or cat- pDNA, probably as a result of its bipartite NLS-like ionic characteristics of lipofectamine favour interaction sequence and interaction with the NPC; (2) the liposome with the nuclear membrane and/or that the condensation lipofectamine is probably more important for cell entry of the pDNA by a cation reduces its size and thus facili- and possibly endosome escape than nuclear translo- tates passage into the nucleus. It has been shown how- cation; and (3) cells in mitosis possess enhanced nuclear ever that pDNA complexed to the cationic lipids when transfer, but cell division is not a prerequisite for microinjected into the cell cytoplasm did not result in an nuclear transfer. increase in gene expression possibly as a result of cytoplasmic sequestration.11,44 A recent report demonstrated that lactosylated poly-l-lysine resulted in nuclear translo- Materials and methods cation of pDNA by a NPC-dependent mechanism, but that unsubstituted poly-l-lysine or mannosylated poly-l- Cell culture lysine were less efﬁcient for transfer.45 The authors postu- The fetal human tracheal epithelial cell line 56FHTE8o- late that the lactose moiety provides for nuclear localis- was kindly provided by Dr D Gruenert.48 Cells were ation by targeting a potential lectin-like protein of the grown in a 1:1 mixture of Dulbecco’s minimum essential NPC. medium and Ham’s F-12 nutrient medium (DMEM/F- The nuclear pore complex acts as a gateway between 12) and supplemented with foetal calf serum (10% v/v) the cytoplasm and the nucleus and is a multiprotein (Biowest, Nuaille´, France). Cells were maintained at 37°C structure which forms a large central transporter channel in 5% CO2–95% air and the media changed every second (50–60 nm in diameter) surrounded by eight peripheral day. Cells were seeded on either six- or 12-well plates,

Gene Therapy Nuclear transfer of plasmid DNA M Colin et al 1651 25 cm2 or 75 cm2 cultured flasks or Labtek eight-well room temperature before the addition of the complex and plates (Gibco BRL, Cergy Pontoise, France) until approxi- subsequent detection of luciferase activity. This lectin mately 80% confluence. binds N-acetylglucosamine residues present on a class of nuclear pore complex (NPC) proteins. In addition, inhi- Peptides and plasmid DNA bition of nuclear pore transport in permeabilised cells The peptides N-[K]16GGCRGDMFGCA ([K]16RGD) and was performed by pretreatment with the Mab 414 anti- N-[K]16GGCRGEMFGCA ([K]16RGE) were synthesized body which recognizes a related family of NPC proteins 33 40 and cyclised as previously described and [K]16, (Babco, Berkley, USA) for 30 min at room temperature [K]16RGD and [K]16RGE peptides were labelled with FITC at a working dilution of 1/10 000. Cells synchronised in as previously described.32 An American firefly(Photinus the G2/M phase by pretreatment with demecolcine (0.06 pyralis) luciferase pDNA (5256 bp) (pGL3) under the con- ␮g/ml), for 16 h at 37°C or in the G0/G1 phase with trol of an SV40 promoter and enhancer was used as a hydroxyurea (2.5 mm) for 18 h at 37°C were transfected reporter gene. pDNA was purified and labelled with Cy5 in the presence of drug as described above and the 48-h by a modified nick-translation technique giving predomi- post-transfection was performed in the absence of drug. 32,49 nantly an open circle form as previously described. 35 The pDNA (1 ␮g) was labelled in the presence of 10-6 [ S]-methionine incorporation After two rinses in chilled PBS, cells were incubated in units of DNase I and the integrity of the pDNA con- ␮ firmed by electrophoresis on an agarose (0.8%) gel and the presence or absence of digitonin (20 g/ml) staining with ethidium bromide (data not shown). The (Calbiochem, Fontenay-sous-Bois, France) for 5 min on Cy5-label was detected using a Molecular Dynamics flu- ice and washed twice in chilled PBS. Cells were then incubated in the presence of [35S]-methionine (15 orescence scanning system (Storm 860) with excitation at ␮ ° 649 and emission at 670 nm (data not shown). Ci/well, six-well plates) for 2 h at 37 C. In order to determine free label associated with the cells, control cells ° Complex formation were incubated in the same conditions at 4 C. After three rinses in chilled PBS cells were incubated with 5% TCA The [K]16RGD, [K]16RGE peptides (FITC labelled or non- labelled) were mixed with the pDNA at a ratio of five to for 30 min on ice and rinsed twice with 80% ethanol. one (w:w), respectively, and incubated 15 min at room Lysis was performed in the presence of 20% KOH for 15 min on ice. The protein concentration was determined by temperature in HEPES buffered saline. Peptide at this 50 ratio condenses the DNA as shown by electron the Bradford method and the radioactivity measured by microscopy and renders it resistant to Dnase.33 Lipofecta- scintillation counting with Ecolite in an LKB Rackbeta mine (2 mg/ml) at 3:1 (w:w) mixture of a polycationic 1209 counter. lipid (DOSPA) and a neutral lipid (DOPE), was then ATP and lactate dehydrogenase release added (final ratio DNA:peptide:lipofectamine; 1:5:18 or Cell ATP and lactate dehydrogenase (LDH) release was 1:5:24 (w:w:w) and the mixture incubated a further 15 measured after permeabilisation of cells with digitonin min at room temperature. (10–150 ␮g/ml). For ATP release,cells were rinsed three times with DMEM/F-12 medium and incubated with Cell synchronisation ATP degradation inhibitors: P1,P5-di (adenosine-5Ј) pen- Cells were synchronised in the G2/M phase by pretreat- Ј ␮ ° taphosphate (Sigma) and adenosine 5 -diphosphate ment with demecolcine (0.06 g/ml), for 16 h at 37 Cor (Sigma) respectively at 0.5 mm and 1 mm in DMEM/F- in the G0/G1 phase with hydroxyurea (2.5 mm) for 18 h ° ° 12 medium, 3 min at 37 C. Cells were then prechilled and at 37 C. Cells were then harvested in trypsin permeabilised with digitonin (10–150 ␮g/ml) containing (0.05%)/EDTA (0.02%) and fixed in cold 70% ethanol inhibitors, for 5 min on ice (600 ␮l per well). The medium overnight. After washing with PBS, cells were resus- was collected and the ATP measured using an adenosine pended in PBS containing RNase A at 1 mg/ml and pro- Ј ␮ 5 -triphosphate bioluminescence assay kit (Sigma) and pidium iodide at 50 g/ml. After 30 min at room tem- detected as relative light units (RLU) emitted. Results are perature, the red fluorescence of the propidium iodide expressed as extracellular ATP. For detection of the lac- was detected and analysed on a FACScalibur (Becton tate dehydrogenase (LDH) activity cells were prechilled Dickinson, Le Pont de Claix, France) equipped with an and incubated 5 min on ice with the indicated digitonin argon laser emitting at 488 nm. concentration. After two rinses with PBS, cells were scrapped into 1 ml of PBS and centrifuged for 10 min at Transfection and luciferase detection ° ␮ 4 C at 800 r.p.m. The pellet was resuspended in hypo- The complex (pDNA:[K]16RGD:lipofectamine, 0.08 g:0.4 ␮ ␮ tonic lysis buffer (500 mm Tris HCl pH 7.4, 100 mm NaCl, g:1.4 g, per well) was added to permeabilised or non- 3mm MgCl ,1mm EDTA, 1 mm EGTA, 0.5% deoxychol- permeabilised cells in a final volume of 300 ␮l of Opti- 2 ° ate, 0.1% SDS, Nonidet P40) and the cytoplasmic marker MEM (Gibco BRL). Cells were then incubated at 37 C for lactate dehydrogenase was assayed with a Synchron CX4 4 h after which the complex was removed. The cells were CE (Beckman Gagny, France) analyser using pyruvate then washed in PBS, the normal supplemented media and NADH as substrates. Results are expressed as a per- added and incubated for a further 48-h post-transfection centage of the LDH in non-permeabilised cells. Cell period. Forty-eight hours was found to give optimal viability was determined by the trypan blue exclusion luciferase activity (data not shown). Luciferase activity 32 assay and triplicate samples were counted in a haemo- was assayed as previously described and the protein cytometer. concentration of cell lysates was determined using the Bradford method.50 For inhibition of nuclear pore trans- Cell uptake of [35S]-labelled pDNA port, permeabilised cells were pretreated with 50 ␮g/ml Cells were synchronised in the G2/M or G0/G1 phase WGA (Sigma, St-Quentin Fallavier, France) for 30 min at by pretreatment with demecolcine (0.06 ␮g/ml), for 16 h

Gene Therapy Nuclear transfer of plasmid DNA M Colin et al 1652 at 37°C or with hydroxyurea (2.5 mm) for 18 h at 37°C, were permeabilised and the nuclear transfer of the Cy5- respectively. The plasmid pGL3 was labelled with pDNA:[K]16RGD:lipofectamine complex at the optimal [35S]dATP and the complex composed of ratio of 1:5:18 (w/w/w) for 1 h at 37°C examined. To test 35 [ S]pGL3:[K]16RGD:lipofectamine (1:5:18) was prepared the nuclear transfer of cyclised and non-cylised peptide, as previously described.32 Cells were then incubated with cells were permeabilised and rinsed twice with PBS and the complex for the indicated times. After rinsing with incubated with 0.1 mg/ml of cyclised or non- cyclised ° PBS, cells were harverested with trypsin/EDTA, the tryp- [K]16RGD, 30 min at 37 C. Control cells pretreated with sin was blocked with of 1% BSA in PBS and cells reco- DTT were incubated a further 30 min at room tempera- vered by centrifugation (10 min, 1100 g). The cell pellet ture with 500 ␮g/ml FITC-dextran to determine nuclear was rinsed once in PBS by centrifugation before lysis in 1 membrane integrity. Cells were washed twice in PBS, N NaOH. The radioactivity was measured by scintillation fixed for 20 min at room temperature in 4% paraformal- counting in a 1600TR analyser (Packard, Rungis, France) dehyde (PFA), rinsed in 50 mm NH4Cl for 5 min at room and the protein content determined by the method of temperature. Cells were treated with 1 mg/ml of RNase Bradford. Results are expressed as c.p.m/mg of protein. A 10 min at room temperature and nuclei were then stained with propidium iodide (3 ␮g/ml) for 3 min at Disulphide bridge peptide reduction room temperature. Cells were then mounted in To determine the impact on nuclear transport of disul- DABCO/glycergel (Dako, Trappes, France). For detection phide bond formation between the cysteine residues of of the luciferase protein, cells were transfected as the peptide, a sample of the peptide was reduced with described above. After a 48-h post-transfection period, dithiotreitol (DTT). DTT was added in a five-fold excess cells were incubated in the presence or absence of digi- relative to thiol groups (ie 0.3 mm) to a 1 mg/ml solution tonin (20 ␮g/ml) for 5 min on ice, washed twice in PBS, of peptide in 0.1 m ammonium bicarbonate pH 8.0 and fixed for 20 min at room temperature in 4% PFA and

incubated for 6 h at room temperature in a nitrogen rinsed four times in 50 mm NH4Cl for 5 min at room atmosphere. The mixture was then acidified to pH 7.4 temperature. Cells were then incubated with a fluor- with acetic acid. To verify reduction of the disulphide escein-conjugated IgG anti-luciferase (firefly, Photinus bridge in the peptide an Ellman’s assay was performed. pyralis) antibody (Rockland, Gilbertsville, USA) for 1 h at Fifty ␮lof3mm DNTB (5,5Ј-dithio-bis(2-nitrobenzoic- room temperature at a working dilution of 1/7500 in the acid) was added to 1 ml of 0.1 mg/ml peptide solution. presence of 0.% saponin. Cells were then washed four

The absorbance at 412 nm was measured on a spectro- times with 50 mm NH4Cl for 5 min at room temperature photometer after incubation for 15 min. The concen- and treated as described above. Confocal microscopy was tration of sulphydryl groups was calculated from the performed using a TCS SP Leica (Lasertechnichk) micro- molar extinction coefficient of 14150/cm using the follow- scope, equipped with a × 40 objective (plan apo; NA = = ing equation, [SH] [A412(sample) - A412(reference)] /14 150. 1.25). For FITC excitation, an argon-krypton ion laser adjusted at 488 nm was used, 568 nm for propidium iod- Microinjection ide and 647 nm for Cy5 excitation. For each optical sec- Cells (1.5 × 105/ml) were grown on Esco glass slides tion, triple fluorescence images were obtained in the (22 × 22 mm) (Erie Scientific, Portsmouth, USA) in 35 mm sequential mode (ie FITC first, TRITC second and Cy5 Petri dishes for 48 h. Either pDNA (pGL3) labelled with third). The signal was treated by line averaging to inte- Cy5 (20 ␮g/␮l ) or Cy5-pGL3 complexed to FITC- grate the signal collected over four lines in order to

[K]16RGD in the presence or absence of lipofectamine at reduce noise. The confocal pinhole was adjusted to allow a ratio of 1:5:18 (pGL3: [K]16RGD:lipofectamine) were for a minimum ﬁeld depth. A focal series was collected injected into the cytoplasm of cells with an Eppendorf for each specimen. The focal step between each section 5242 microinjector using glass micropipettes (Clark, was 0.9 ␮m. Selected sections were then processed to pro- Reading, UK). About 1 pl was injected per cell. Cells were duce a single composite overlay image (colour merged). then rinsed twice with OpiMEM and incubated at 37°C Images were printed on a colour ink jet printer (Epson for the indicated times. Stylus color 850) using Photoshop 5.0 software.

Confocal laser scanning microscopy Statistical analysis ± Cells were permeabilised for 5 min on ice with digitonin Data are presented as the mean s.e.m of triplicate deter- (20 ␮g/ml). The wells were rinsed twice with PBS and minations and are representative of results obtained in then incubated at 37°Cor4°C with different components: three or four independent experiments that produced similar relative results. Note that the luciferase mean Cy5-labelled pDNA:FITC-labelled [K]16RGD/E:lipofectamine at the above indicated optimal ratio 1:5:18 values varied for different batches of cells and for cells (w/w/w, 0.3 ␮g of pDNA per well) or Cy5-pDNA:FITC- of distant passage number. Confocal images show rep- ␮ resentative results of at least two individual experiments [K]16RGD (1:5, 0.3 g pDNA per well) or Cy5-pDNA:lipofectamine (1:18, 0.3 ␮g pDNA per well), or FITC- using different batches of labelled pDNA. Statistical ␮ analysis was performed using the nonparametric rank [K]16RGD/E (5 g per well) or FITC-polylysine [K]16 (5 Ͻ ␮g per well) or Cy5-pDNA alone (0.3 ␮g pDNA per well) test of Wilcoxon. Probability values 0.05 were con- in OptiMEM for 1 h. Pretreatment with the lectin wheat sidered statistically signiﬁcant. germ agglutinin (WGA, 50 ␮m) or with the Mab 414 antibody (1/10 000) was performed for 30 min at room tem- Acknowledgements perature. To verify the integrity of the nuclear membrane This work was supported by the Association Franc¸aise of permeabilised cells, they were incubated with 500 de Lutte contre la Mucoviscidose and MC was supported ␮g/ml FITC-dextran (MW 70 000, Sigma) for 30 min at by a grant from this association. We are very grateful to 37°C. Cells synchronised in the G0/G1 or G2/M phase Dr Andrew D Miller of the Department of Chemistry

Gene Therapy Nuclear transfer of plasmid DNA M Colin et al 1653 Imperial College, London, UK, for synthesising the 26 Remy JS et al. Targeted gene transfer into hepatoma cells with [K]16RGD and [K]16RGE peptides. We thank Dr Elizabeth lipopolyamine-condensed DNA particles presenting galactose Lasnier for the assay of LDH. ligands: a stage toward artificial viruses. Proc Natl Acad Sci USA 1995; 92: 1744–1748. References 27 Brunner S et al. Cell cycle dependence of gene transfer by lipo- plex, polyplex and recombinant adenovirus. Gene Therapy 2000; 1 Colin M et al. Cell delivery, intracellular trafficking and 7: 401–407. expression of an integrin-mediated gene transfer vector in tra- 28 Jiang C et al.Efficiency of cationic lipid-mediated transfection cheal epithelial cells. Gene Therapy 2000; 7: 139–152. of polarized and differentiated airway epithelial cells in vitro 2 Brandeń LJ, Mohamed AJ, Smith CIE. A peptide nucleic acid- and in vivo. Hum Gene Ther 1998; 9: 1531–1542. nuclear localisation signal fusion that mediates nuclear trans- 29 Mortimer I et al. Cationic lipid-mediated transfection of cells in port of DNA. Nature Biotechnol 1999; 17: 784–787. culture requires mitotic activity. Gene Therapy 1999; 6: 403–411. 3 Chan CK, Jans DA. Enhancement of polylysine-mediated trans- et al fer infection by nuclear localisation sequences: polylysine does 30 Wilke M .Efficacy of a peptide-based gene delivery system not function as a nuclear localisation sequence. Hum Gene Ther depends on mitotic activity. Gene Therapy 1996; 3: 1133–1142. 1999; 10: 1695–1702. 31 Godbey WT, Wu KK, Mikos AG. Tracking the intracellular path 4 James MB, Giorgio TD. Nuclear-associated plasmid, but not cell- of poly(ethylenimine)/DNA complexes for gene delivery. Proc associated plasmid, is correlated with transgene expression in Natl Acad Sci USA 1999; 96: 5177–5181. cultured mammalian cell. Mol Ther 2000; 1: 339–346. 32 Colin M et al. Liposomes enhance delivery and expression of an 5 Zanta MA, Belguise-Valladier P, Behr JP. Gene delivery: a single RGD-oligolysine gene transfer vector in human tracheal cells. nuclear localisation signal peptide is sufficient to carry DNA to Gene Therapy 1998; 5: 1488–1498. the cell nucleus. Proc Natl Acad Sci USA 1999; 96:91–96. 33 Harbottle RP et al. An RGD-oligolysine peptide: a prototype 6 Collas P, Alestro¨m P. Rapid targeting of plasmid DNA to construct for integrin-mediated gene delivery. Hum Gene Ther zebrafish embryo nuclei by the nuclear localisation signal of 1998; 9: 1037–1047. SV40 T antigen. Mol Marine Biol Biotechnol 1997; 6:48–58. 34 Hart SL et al. Gene delivery and expression mediated by an inte- 7 Dean DA. Import of plasmid DNA into the nucleus is sequence grin-binding peptide. Gene Therapy 1995; 2: 552–554. specific. Exp Cell Res 1997; 230: 293–302. 35 Hart SL et al. Integrin-mediated transfection with peptides con- 8 Dean DA, Byrd JN, Dean BS. Nuclear targeting of plasmid DNA taining arginine-glycine-aspartic acid domains. Gene Therapy in human corneal cells. Curr Eye Res 1999; 19:66–75. 1997; 4: 1225–1230. 9 Escriou V et al. Cationic lipid-mediated gene transfer: analysis 36 Hart SL et al. Lipid-mediated enhancement of transfection by a of cellular uptake and nuclear import of plasmid DNA. Cell Biol nonviral integrin-targeting vector. Hum Gene Ther 1998; 9: 575–585. Toxicol 1998; 14:95–104. 37 Adam SA, Marr RS. Nuclear protein import in permeabilised 10 Neves C et al. Intracellular fate and nuclear targeting of plasmid mammalian cells requires soluble cytoplasmic factors. J Cell Biol DNA. Cell Biol Toxicol 1999; 15: 193–202. 1990; 111: 807–816. 11 Pollard H et al. Polyethylenimine but not cationic lipids pro- 38 Dunn LA, Holz RW. Catecholamine secretion from digitonin- motes transgene delivery to the nucleus in mammalian cells. J treated adrenal medullary chromaffin cells. J Biol Chem 1983; Biol Chem 1998; 273: 7507–7511. 258: 4989–4993. 12 SebestyeńMGet al. DNA vector chemistry: the covalent attach- 39 Moore MS, Blobel G. The two steps of nuclear import, targeting ment of signal peptides to plasmid DNA. Nat Biotechnol 1998; to the nuclear envelope and translocation through the nuclear 16:80–85. pore, require different cytosolic factors. Cell 1992; 69: 939–950. 13 Vacik J et al. Cell-specific nuclear import of plasmid DNA. Gene 40 Davis LI, Blobel G. Identification and characterization of a Therapy 1999; 6: 1006–1014. nuclear pore complex protein. Cell 1986; 45: 699–709. 14 Hagstrom JE et al. Nuclear import of DNA in digitonin-permea- 41 Gould SJ, Keller G-A, Subramani S. Identification of a peroxiso- bilised cells. J Cell Sci 1997; 110: 2323–2331. mal targeting signal at the carboxy terminus of firefly luciferase. 15 Hartig R et al. Active nuclear import of single-stranded oligonu- J Cell Biol 1987; 105: 2923–2931. cleotides and their complexes with non-karyophilic macromol- 42 Rojo EE et al. Sorting of D-lactate dehydrogenase to the inner ecules. Biol Cell 1998; 90: 407–426. membrane of mitochondria. Analysis of topogenic signal and 16 Subramanian A, Ranganathan P, Diamond SL. Nuclear targeting energetic requirements. J Biol Chem 1998; 273: 8040–8047. peptide scaffolds for lipofection of nondividing mammalian 43 Collas P, Husebye H, Alestro¨m P. The nuclear localisation cells. Nat Biotechnol 1999; 17: 873–877. sequence of the SV40 T antigen promotes transgene uptake and 17 Wilson GL et al. Nuclear import of plasmid DNA in digitonin- expression in zebrafish embrio nuclei. Transgenic Res 1996; 5: permeabilised cells requires both cytoplasmic factors and spe- 451–458. cific DNA sequences. J Biol Chem 1999; 274: 22025–22032. 44 Zabner J et al. Cellular and molecular barriers to gene transfer 18 Dowty ME et al. Plasmid DNA entry into postmitotic nuclei of by a cationic lipid. J Biol Chem 1995; 270: 18997–19007. primary rat myotubes. Proc Natl Acad Sci USA 1995; 92: 4572–4576. 45 Klink DT et al. Nuclear translocation of lactosylated poly-l- 19 Stoffler D, Fahrenkrog B, Aebi U. The nuclear pore complex: lysine/cDNA complex in cystic fibrosis airway epithelial cells. from molecular architecture to functional dynamics. Curr Opin Mol Ther 2001; 3: 831–841. Cell Biol 1999; 11: 391–401. 20 Talcott B, Moore MS. Getting across the nuclear pore complex. 46 Yang J-P, Huang L. Time-dependent maturation of cationic lipo- Trends Cell Biol 1999; 9: 312–318. some-DNA complex for serum resistance. Gene Therapy 1998; 5: 21 Neves C et al. Coupling of a targeting peptide to plasmid DNA 380–387. by covalent triple helix formation. FEBS Lett 1999; 453:41–45. 47 Escriou V et al. Critical assessment of the nuclear import of plas- 22 Aronsohn AI, Hughes JA. Nuclear localisation signal peptides mid during cationic lipid-mediated gene transfer. J Gene Med enhance cationic liposome-mediated gene therapy. J Drug Target 2001; 3: 179–187. 1997; 5: 163–169. 48 Gruenert DC et al. Characterization of human tracheal epithelial 23 Collas P, Alestro¨m P. Nuclear localisation signals: a driving cells transformed by an origin-defective simian virus 40. Proc force for nuclear transport of plasmid DNA in zebrafish. Biochem Natl Acad Sci USA 1988; 85: 5951–5955. Cell Biol 1997; 75: 633–640. 49 Wheeler VC, Coutelle C. Nondegradative in vitro labeling of 24 Gao X, Huang L. Potentiation of cationic liposome-mediated plasmid DNA. Anal Biochem 1995; 225: 374–376. gene delivery by polycations. Biochemistry 1996; 35: 1027–1036. 50 Bradford M. A rapid and sensitive method for the quantification 25 Morris MC et al. A novel potent strategy for gene delivery using a of microgram quantities of protein utilizing protein-dye bind- single peptide vector as a carrier. Nucleic Acid Res 1999; 27: 3510–3517. ing. Anal Biochem 1976; 72: 248–254.

Gene Therapy