Conjugation of a New Two-Photon Fluorophore to Poly(ethylenimine

of the objective is slightly overfilled, creating a diffraction-limited focal spot. ...... Hirons, G. T., Fawcett, J. J., and Crissman, H. A. (199...
0 downloads 0 Views 196KB Size
Bioconjugate Chem. 2007, 18, 844−851

844

Conjugation of a New Two-Photon Fluorophore to Poly(ethylenimine) for Gene Delivery Imaging Ali Hayek,† Sebnem Ercelen,‡ Xin Zhang,‡ Fre´de´ric Bolze,† Jean-Franc¸ ois Nicoud,*,† Emmanuel Schaub,‡ Patrice L. Baldeck,§ and Yves Me´ly*,‡ Groupe des Mate´riaux Organiques, Institut de Physique et Chimie des Mate´riaux de Strasbourg, Universite´ Louis Pasteur (CNRS, UMR 7504), BP 43 23 rue du Loess, F-67034 Strasbourg, France, Photophysique des Interactions Biomole´culaires, (CNRS, UMR 7175), Institut Gilbert Laustriat, Faculte´ de Pharmacie, 74, route du Rhin - BP 60024 F-67401 Illkirch-Graffenstaden, France, and Laboratoire de Spectrome´trie Physique, Universite´ Joseph Fourier, (CNRS, UMR 5588), BP 87 F-38402 Saint Martin d’He`res, France. Received November 21, 2006; Revised Manuscript Received January 30, 2007

We report herein the molecular engineering of an efficient two-photon absorbing (TPA) chromophore based on a donor-donor bis-stilbenyl entity to allow conjugation with biologically relevant molecules. The dye has been functionalized using an isothiocyanate moiety to conjugate it with the amine functions of poly(ethylenimine) (PEI), which is a cationic polymer commonly used for nonviral gene delivery. Upon conjugation, the basic architecture and photophysical properties of the active TPA chromophore remain unchanged. At the usual N/P ratio (ratio of the PEI positive charges to the DNA negative charges) of 10 used for transfection, the transfection efficiency and cytotoxicity of the labeled PEI/DNA complexes were found to be comparable to those of the unlabeled PEI/DNA complexes. Moreover, when used in combination with unlabeled PEI (at a ratio of 1 labeled PEI to 3 unlabeled PEI), the labeled PEI does not affect the size of the complexes with DNA. The labeled PEI was successfully used in two-photon fluorescence correlation spectroscopy measurements, showing that at N/P ) 10 most PEI molecules are free and the diffusion coefficient of the complexes is consistent with the 360 nm size measured by quasielastic light scattering. Finally, two-photon images of the labeled PEI/DNA complexes confirmed that the complexes enter into the cytoplasm of HeLa cells by endocytosis and hardly escape from the endosomes. As a consequence, the functionalized TPA chromophore appears to be an adequate tool to label the numerous polyamines used in nonviral gene delivery and characterize their complexes with DNA in two-photon applications.

INTRODUCTION During the past decade, two-photon fluorescence imaging has evolved as a powerful technique for three-dimensional scanning microscopy in biological research (1, 2). This technique offers many advantages over conventional confocal microscopy. First of all, the two-photon absorption only occurs at the laser focus where the laser light intensity is maximum, so that dye bleaching and fluorescence above and below the focused spot are avoided. Furthermore, the excitation source is generally an IR laser, which takes advantage of the transparency window of biological tissues. In addition, the IR light is more penetrating and less damaging to cells than the more energetic UV-vis light, enabling observation of thicker samples and during more extended periods without degradation (3). Furthermore, twophoton excitation was found to be ideally suited for fluorescence correlation spectroscopy (FCS), allowing the determination of the local concentrations and diffusion coefficients of fluorescently labeled molecules in cells (4, 5). However, the use of two-photon microscopy and FCS is still hampered by the availability of appropriate two-photon dyes. Indeed, classical one-photon dyes, well-known in epifluorescence or confocal microscopy, are still largely used in two-photon applications, though they exhibit only low two-photon cross sections. * Corresponding authors. Pr. J.-F. Nicoud (chemistry): Fax 33 (0)3 88 10 72 46; Tel 33 (0)3 88 10 71 55, E-mail: nicoud@ ipcms.u-strasbg.fr. Pr. Y. Me´ly (transfection): Fax 33(0)3 90 24 43 13; Tel 33 (0)3 90 24 42 63, E-mail: [email protected]. † Universite´ Louis Pasteur. ‡ Institut Gilbert Laustriat. § Universite´ Joseph Fourier.

Consequently, the development of new water-soluble dyes, specifically engineered for their multiphoton absorption and fluorescence, is strongly required. Recently, we reported the synthesis of water-soluble efficient blue TPA fluorophores (6). These compounds present a D-π-rigid core-π-D symmetric architecture, where D ) donor group(s), π ) conjugated spacer, and rigid core ) biaryl system (Figure 1a). By introducing oligo(ethylene glycol) moieties, we obtained good water solubility (up to 100 g.L-1 for the fluorophore shown in Figure 1b), which enables the use of these dyes for cellular imaging (7). To go one step further in the validation and application of these watersoluble dyes, our aim in this work was to prepare functionalized TPA fluorophores in order to label biologically relevant molecules and monitor them in cells by two-photon microscopy and FCS. For this purpose, we selected poly(ethylenimine)s (PEIs) that are among the most efficient nonviral gene delivery agents (8, 9). PEIs are rich in amino groups and positively charged at neutral pH. Due to their high cationic charge density, PEIs are proficient condensing agents for DNA that lead to the formation of small particles ( 5) (10). These DNA/PEI complexes exhibit good transfection efficiency in vitro and in various in vivo applications (11). To conjugate the amine functions of PEI, we synthesized the isothiocyanate-functionalized TPA chromophore 1 (Figure 2). This dye was conjugated to PEI at a ratio of 1 dye per 40 amine functions. The corresponding conjugate PEI derivative was shown to substitute nonconjugated PEIs in their complexes with DNA and can be used to monitor the intracellular pathway of the complexes in cells by twophoton microscopy.

10.1021/bc060362h CCC: $37.00 © 2007 American Chemical Society Published on Web 04/03/2007

Conjugation of TPA Fluorophore to PEI

Figure 1. General architecture of a 1D two-photon absorbing chromophore (a) and chemical structure of a symmetrical water-soluble chromophore (b) (8).

Figure 2. Amine-reactive TPEF chromophore 1.

EXPERIMENTAL PROCEDURES Materials. Branched PEI (25 kDa) was purchased from Aldrich. Lysis reagent was from Promega. All other chemicals and reagents were purchased from Aldrich or Acros Organics and were used as received unless specified otherwise. THF was distilled over sodium under argon atmosphere prior to use. Compounds 3, 4, and 7 were prepared as previously described (6). Synthesis. 1H and 13C NMR spectra were recorded with a 300 MHz Brucker Advance 300 instrument in CDCl3 (Aldrich) with CHCl3 as an internal standard (7.24 ppm for 1H and 77 ppm, middle of the three peaks, for 13C spectra). High-resolution ESI mass spectra were recorded with a Microtof instrument at Service de spectrome´trie de masse de l’Institut de chimie LC3ULP- Strasbourg. TLC was run on Merck precoated aluminum plates (Si 60 F254). Column chromatographies were run on Merck silica gel (60-120 mesh) or neutral alumina (Merck) 70-230 mesh. 4-(t-Butyldimethylsilanyloxy)-3,5-dimethoxybenzaldehyde (5). Under argon, in a 250 mL two-necked flask, 4.6 g of tbutyldimethylsilylchloride (30 mmol) and 2.1 g of imidazole (30 mmol) were dissolved in 120 mL of a dimethylformamide/ dichloromethane mixture (1/1 v/v). When the solution became homogeneous, 5 g of syringaldehyde (27 mmol) was added and the mixture was heated at 80 °C overnight. Then, the solvents were evaporated under vacuum. The yellow solid was dissolved in 200 mL of CH2Cl2, washed with water (150 mL), dried on MgSO4, and evaporated under reduced pressure to give a yellow oil. The oil was solubilized in ethanol (40 mL) and cooled down in the freezer overnight. The white precipitate was filtered and dried to give 4.96 g of the desired protected product 5 in 62% yield. 1H NMR (300 MHz, CDCl3, δ ppm): 9.82 (s, 1H), 7.10 (s, 2H), 3.87 (s, 6H), 1.01 (s, 9H), 0.16 (s, 6H). 13C NMR (75.75 MHz, CDCl3, δ ppm): 191.98, 151.92, 140.98, 129.3, 106.66, 55.75, 25.62, 18.73, 4.58. HRMS calculated m/z for C15H24O4SiLi, 303.1604; found, 303.1599. 4-{2-[4-(t-Butyldimethylsilanyloxy)-3,5-dimethoxy]styryl}-4′(diethoxyphosphinylmethyl)-9,9′-di-n-propylfluorene (6). To a solution of 4,4′-bis(diethoxyphosphinylmethyl)-9,9′-dipropy-

Bioconjugate Chem., Vol. 18, No. 3, 2007 845

lfluorene 4 (3.34 g, 6.08 mmol, 1 equiv) in 50 mL of anhydrous THF was added 0.73 g of NaH (30.4 mmol, 5 equiv, 60% in mineral oil). After the effervescence ceased, the suspension was stirred for 1 h at room temperature under an argon atmosphere. After cooling at 0 °C, a solution of 1.80 g of 5 (6.08 mmol, 1 equiv) in anhydrous THF (50 mL) was added slowly. The mixture was stirred for 24 h and neutralized by ethanol (10 mL). The solvents were evaporated under vacuum giving an oil, which was dissolved in CH2Cl2 and washed with brine (50 mL) and water (50 mL). The crude product was purified by column chromatography on silica gel with dichloromethane/methanol as eluant (97/3 in vol). The desired product 6 was obtained as a yellow oil (m ) 1.47 g, 35% yield). 1H NMR (300 MHz, CDCl3, δ ppm): 7.65-7.60 (m, 2H), 7.48-7.45 (m, 2H), 7.297.26 (m, 2H), 7.07 (s, 2H), 6.76 (s, 2H), 4.09-3.98 (m, 4H), 3.87 (s, 6H), 3.24 (d, 2H, J ) 21.49 Hz), 2.20-1.97 (m, 4H), 1.23 (t, 6H, J ) 7.02 Hz), 1.03 (s, 9H), 0.67 (s, 10H), 0.16 (s, 6H). 13C NMR (75.75 MHz, CDCl3, δ ppm): 151.27, 151.27, 151.22, 151.11, 140.23, 104.20, 139.69, 139.64, 136.41, 136.39, 134.42, 130.29, 130.25, 130.17, 128.55, 128.46, 128.27, 127.50, 125.37, 124.28, 124.19, 120.45, 119.75, 119.61, 119.56, 103.61, 62.13, 62.04, 55.74, 55.10, 42.78, 35.00, 33.18, 25.76, 18.76, 17.12, 16.38, 16.30, 14.37, -4.66. HRMS calculated m/z for C40H57O6PSi, 692.3662; found, 692.3664. 4-{2-[4-(t-Butyldimethylsilanyloxy)-3,5-dimethoxy]styryl}-4′{2-[3,4,5-tris(1,4,7,10,13-pentaoxotetradecyl)]styryl}-9,9′-di-npropylfluorene (2). This product was prepared as described for 6 using 500 mg of 6 (0.72 mmol, 1 equiv), 144 mg of NaH (3.6 mmol, 5 equiv), and 523 mg of aldehyde 7 (0.72 mmol, 1 equiv). Product 2 was obtained as a yellow oil (m ) 909.80 mg, 85% yield). 1H NMR (300 MHz, CDCl3, δ ppm): 7.64 (d, 2H J ) 8.33 Hz), 7.48-7.46 (m, 4H), 7.08 (s, 2H), 7.05 (d, J ) 1.32 Hz), 6.80 (s, 2H), 6.76 (s, 2H), 4.25-4.16 (m, 6H), 3.89 (t, 4H, J ) 5.26 Hz), 3.87 (s, 6H), 3.81 (t, 2H, J ) 5.26 Hz), 3.77-3.61 (m, 24H), 3.57-3.53 (m, 6H), 3.38 (s, 3H), 3.37 (s, 6H), 2.02-1.97 (m, 4H), 1.03 (s, 9H), 0.70 (s, 10H), 0.16 (s, 6H). 13C NMR (75.75 MHz, CDCl3, δ ppm): 152.74, 151.72, 151.45, 140.58, 140.23, 138.26, 136.46, 136.04, 134.42, 133.10, 130.26, 128.72, 128.31, 127.64, 127.50, 125.59, 125.44, 120.60, 120.46, 119.81, 119.77, 106.13, 103.62, 71.88, 70.79, 70.65, 70.62, 70.57, 70.54, 70.47, 69.73, 68.81, 58.97, 55.73, 55.10, 53.39, 42.86, 25.76, 18.71, 17.15, 14.45, -4.65. 4-{2-[4-(2,6-Dimethoxy-4-Vinylphenoxy)butyl]isoindolin-1,3dione-4′-[3,4,5-tris-(1,4,7,10,13-pentaoxotetradecyl)]styryl}9,9′-di-n-propylfluorene (9). (a) Deprotection of the phenol moiety: In a 100 mL two-necked flask, under an argon atmosphere, 550 mg of 2 (0.43 mmol, 1 equiv) was dissolved in 50 mL anhydrous THF. A solution of TBAF (0.19 g, 0.65 mmol, 1.5 equiv) in THF (1 mL) was introduced dropwise into the flask. The mixture was stirred at room temperature for 2 h and put in 100 mL of 1 M phosphoric acid. The aqueous phase was extracted with diethyl ether, washed with water, and dried on magnesium sulfate. The solvent was removed under reduced pressure, and the crude product was purified by column chromatography on silica gel (ethyl acetate/hexane mixture as eluant 50/50 in vol). The phenolic product was obtained as a yellow oil (m ) 1.02 g, 90% yield). 1H NMR (300 MHz, CDCl3, δ ppm): 7.64 (d, 2H J ) 8.33 Hz), 7.48-7.46 (m, 4H), 7.08 (d, 2H, J ) 2.20 Hz), 7.05 (d, 2H, J ) 1.09 Hz), 6.809 (s, 2H), 6.801 (s, 2H), 5.58 (s, 1H), 4.23 (t, 4H, J ) 4.83 Hz), 4.18 (t, 2H, J ) 4.83 Hz), 3.97 (s, 6H), 3.89 (t, 4H, J ) 5.26 Hz), 3.81 (t, 2H, J ) 5.26 Hz), 3.77-3.61 (m, 24H), 3.57-3.53 (m, 6H), 3.38 (s, 3H), 3.37 (s, 6H), 2.02-1.97 (m, 4H), 0.70 (s, 10H). 13C NMR (75.75 MHz, CDCl , δ ppm): 152.72, 151.51, 151.45, 3 147.19, 140.55, 140.31, 138.30, 136.34, 136.09, 134.68, 133.08, 129.14, 128.71, 128.09, 127.67, 127.40, 125.59, 125.43, 120.61, 120.44, 119.83, 106.15, 103.21, 77.16, 72.38, 71.90, 70.81,

846 Bioconjugate Chem., Vol. 18, No. 3, 2007

70.65, 70.57, 70.54, 70.51, 70.49, 69.73, 68.83, 59.10, 58.99, 56.27, 55.11, 42.86, 29.66, 24.20, 19.76, 17.15, 14.45, 13.65. (b) Product 9. In a 100 mL three-necked flask were introduced DMF (50 mL), 340 mg of the previous deprotected phenolic compound (0.29 mmol, 1 equiv), 204 mg of K2CO3 (1.48 mol, 5 equiv), and 100.15 mg of N-(4-bromobutyl)phthalimide (0.35 mmol, 1.20 equiv). The mixture was then heated at 80 °C for 4 days under an argon atmosphere. The solvent was removed under vacuum, and the residue dissolved in dichloromethane, washed with water, and dried over magnesium sulfate. After evaporation of the dichloromethane, the crude product was purified by column chromatography on silica gel, with dichloromethane as eluant to give the desired product 9 as a yellow oil (m ) 333 mg, 85% yield). 1H NMR (300 MHz, CDCl3, δ ppm): 7.86-7.83 (m, 2H), 7.72-7.69 (m, 2H), 7.65 (d, 2H J ) 8.33 Hz), 7.50-7.46 (m, 4H), 7.09 (s, 2H), 7.05 (d, 2H, J ) 0.87 Hz), 6.80 (s, 2H), 6.76 (s, 2H), 4.23 (t, 4H, J ) 4.83 Hz), 4.18 (t, 2H, J ) 4.83 Hz), 4.03 (t, 2H, J ) 6.35 Hz), 3.91 (s, 6H), 3.89 (t, 4H, J ) 5.26 Hz), 3.81 (t, 2H, J ) 5.26 Hz), 3.773.61 (m, 24H), 3.57-3.53 (m, 6H), 3.38 (s, 3H), 3.37 (s, 6H), 2.02-1.76 (m, 8H), 0.70 (s, 10H). 13C NMR (75.75 MHz, CDCl3, δ ppm): 168.40, 153.59, 152.74, 151.51, 140.49, 140.31, 138.29, 136.83, 136.18, 136.14, 133.78, 133.11, 133.08, 132.15, 128.70, 128.57, 127.98, 127.70, 125.57, 123.09, 120.61, 120.44, 119.84, 106.16, 103.49, 77.17, 72.55, 72.38, 71.89, 70.79, 70.65, 70.62, 70.58, 70.57, 70.54, 70.52, 70.48, 69.74, 68.82, 58.98, 56.07, 55.13, 42.86, 37.78, 27.44, 25.22, 17.15, 14.45. HRMS m/z calculated for C76H103NO20Na, 1372.6971; found, 1372.7004. 4-{2-[4-(4-Aminobutyloxy)-3,5-dimethoxy]styryl}-4′-{2-[3,4,5tris(1,4,7,10,13-pentaoxotetradecyl)]styryl}-9,9′-di-n-propylfluorene (10). In a 250 mL two-necked flask, 240 mg of 9 (0.2 mmol, 1 equiv) was dissolved in absolute ethanol (30 mL) under argon atmosphere. Hydrazine hydrate (0.2 mL, 0.802 mmol) was then added. The mixture was refluxed for 5 h, then cooled to room temperature, after which aqueous 1 M HCl (5 mL) was added slowly. Then, the reaction mixture was neutralized with dilute sodium carbonate solution to pH ) 7. The aqueous phase was extracted with dichloromethane, the organic phase was dried on MgSO4, and the solvent was removed under vacuum. The crude product was purified by column chromatography on silica gel with methanol/dichloromethane mixture (5/95 in vol) as eluant to give 163 mg of the free amine 10 as a yellow oil (75% yield). 1H NMR (300 MHz, CDCl3, δ ppm): 7.65 (d, 2H J ) 8.33 Hz), 7.50-7.46 (m, 4H), 7.10 (s, 2H), 7.05 (s, 2H), 6.80 (s, 2H), 6.78 (s, 2H), 4.23 (t, 4H, J ) 4.83 Hz), 4.18 (t, 2H, J ) 4.83 Hz), 4.02 (t, 2H, J ) 6.35 Hz), 3.92 (s. 6H), 3.89 (t, 4H, J ) 5.26 Hz), 3.81 (t, 2H, J ) 5.26 Hz), 3.77-3.61 (m, 22H), 3.57-3.53 (m, 6H), 3.38 (s, 3H), 3.37 (s, 6H), 2.80 (t, 2H, J ) 7.01 Hz), 2.02-1.65 (m, 8H), 0.70 (s, 10H). 13C NMR (75.75 MHz, CDCl3) δ (ppm): 153.54, 153.13, 152.73, 152.35, 151.49, 151.23, 140.52, 138.26, 137.00, 136.15, 133.09, 128.65, 127.92, 127.71, 125.57, 122.80, 120.62, 119.84, 119.55, 111.27, 108.05, 106.13, 105.47, 103.57, 77.17, 73.23, 73.08, 72.38, 71.88, 70.78, 70.73, 70.63, 70.61, 70.56, 70.46, 69.73, 68.81, 68.70, 58.97, 56.11, 55.98, 55.13, 55.02, 42.85, 41.87, 38.19, 29.97, 27.45, 17.16, 14.45. HRMS calculated m/z for C68H102NO18, 1220.7096; found, 1220.7090. 4-{2-[4-(4-Isothiocyanatobutyloxy)-3,5-dimethoxy]styryl}-4′{2-[3,4,5-tris(1,4,7,10,13-pentaoxotetradecyl)]styryl}-9,9′-di-npropylfluorene (1). In a 100 mL two-necked flask, 350 mg of the amine derivative 10 (0.28 mmol, 1 equiv) was dissolved in 5 mL of anhydrous dichloromethane under argon atmosphere. The resulting solution was cooled at -20 °C. Then, 169 mg of 4-N,N-dimethylaminopyridine (1.29 mmol, 4.50 equiv) and 0.13 mL of thiophosgene (0.46 mmol, 1.60 equiv) were added. The reaction progress was monitored by TLC (silica, CH2Cl2/MeOH 85/15 in vol as eluant). After the disappearance of the starting

Hayek et al.

materials, the solvent was removed under vacuum, and the crude product was chromatographed on silica gel, with CH2Cl2/MeOH 90/10 in vol as eluant. The desired amine-reactive chromophore 1 was obtained as a yellow oil (m ) 346 mg, 99% yield). 1H NMR (300 MHz, CDCl3, δ ppm): 7.65 (d, 2H, J ) 8.33 Hz), 7.53-7.44 (m, 4H), 7.11 (s, 2H), 7.05 (s, 2H), 6.80 (s, 2H), 6.78 (s, 2H), 4.23 (t, 4H, J ) 4.83 Hz), 4.18 (t, 2H, J ) 4.83 Hz), 4.02 (t, 2H, J ) 6.35 Hz), 3.92 (s, 6H), 3.89 (t, 4H, J ) 5.26 Hz), 3.81 (t, 2H, J ) 5.26 Hz), 3.77-3.61 (m, 24H), 3.573.53 (m, 6H), 3.38 (s, 3H), 3.37 (s, 6H), 2.03-1.85 (m, 8H), 0.70 (s, 10H). 13C NMR (75.75 MHz, CDCl3, δ ppm): 153.50, 152.74, 151.52, 140.58, 140.47, 136.68, 136.18, 136.10, 133.34, 133.08, 128.79, 128.69, 127.86, 127.73, 125.60, 120.63, 119.85, 106.15, 103.43, 77.17, 72.38, 72.07, 71.89, 70.79,70.74, 70.65, 70.62, 70.57, 70.48, 69.73, 68.81, 58.98, 56.06, 55.93, 55.13, 44.79, 42.85, 26.99, 26.84, 17.15, 14.45. HRMS calculated m/z for C69H99NO18S, 1262.5853; found, 1262.5827. Two-Photon Absorption Properties. The TPA cross section spectra were obtained by up-conversion fluorescence using a Nd:YAG pumped optical parametric oscillator that produces 2.6 ns [full width at half-maximum (fwhm)] pulses in the 450650 nm spectral range and using a Ti:sapphire femtosecond laser in the range 700-900 nm. This setup does not allow TPA measurements between 650 and 700 nm. The excitation beam is collimated over the cell length (10 mm). The fluorescence, collected at 90° of the excitation beam, was focused into an optical fiber connected to a spectrometer. The incident beam intensity was adjusted to ensure an intensity-squared dependence of the fluorescence over the whole spectral range. Calibration of the spectra was performed by comparison with p-bis(omethylstyryl)benzene, for which σ2 ) 70 GM (Go¨ppert-Mayer) at 570 nm (1 GM ) 10-50 cm4.s.photon-1) and with the published 700-900 nm Rhodamine B two-photon absorption spectrum (12). PEI-Dye Conjugation. In all experiments, we used a fresh stock solution of branched PEI (Mw ) 25 kDa) with 100 mmol/L of primary amino groups (determination by the TNBS method, using spermine for the calibration curve (13)). Conjugation was achieved in microcentrifuge tubes by adding 50 µL of a solution of dye 1 in water (5 mg/mL) and 450 µL of borate buffer (pH ) 9.4) to 500 µL of the PEI solution. PEI was incubated with dye 1 for 1 h with continuous stirring at room temperature. The mixture was allowed to stand overnight, and the unreacted dye was removed by gel chromatography (Sephadex G-25, Pharmacia, Freiburg Germany). The dye concentration of the PEI-1 conjugate was then determined by absorption spectroscopy, using an extinction coefficient of 3.8 × 104 M-1 cm-1 at 354 nm. Preparation of PEI/DNA Complexes. Complexes of DNA with PEI or PEI-1 were prepared as described previously (8). Briefly, the indicated amounts of plasmid DNA (expressed in phosphate groups) and PEI-1 (expressed in amine groups) prepared in 250 µL HEPES 20 mM, pH 7.4, were briefly mixed and incubated at room temperature. PEI/DNA complexes were formed at molar ratios of PEI nitrogen to DNA phosphate (N/ P) ranging from 0.5 to 10. Spectroscopic Measurements and Particle Size Measurements. One-photon absorption and fluorescence spectra were recorded on a Cary 400 spectrophotometer and a Fluorolog (Jobin-Yvon) spectrofluorimeter, respectively. The condensation of DNA by PEI-1 was monitored by using the fluorescence of the DNA bis-intercalator YOYO-1. Plasmid DNA-YOYO-1 complexes were obtained by mixing equal volumes of solutions of DNA and YOYO-1 in 20 mM HEPES, 150 mM NaCl, pH 7.4. To ensure a uniform distribution of YOYO-1, the solution was incubated overnight at 4 °C under continuous stirring. The sizes of the PEI-1/DNA complexes were determined using a

Conjugation of TPA Fluorophore to PEI

Zetamaster 3000 (Malvern Instruments, Paris, France) with the following specifications: sampling time, 30 s; medium viscosity, 1.054 cP; refractive index (RI) medium, 1.34; RI particle, 1.45; scattering angle, 90°; temperature, 25 °C. Data were analyzed using the multimodal number distribution software of the instrument. Two-Photon Imaging and FCS Measurements. Twophoton FCS and microscopy were performed on a home-built setup. Two-photon excitation (TPE) is provided by a Tsunami Ti:sapphire laser pumped with a Millennia V solid-state laser (Spectra-Physics, Mountain View, CA). Pulses of ∼100 fs are produced with a 80 MHz frequency at 760 nm. After a beam expander, the infrared light is focused into the sample by a water-immersion Olympus objective (60× , NA ) 1.2) mounted on an Olympus IX70 inverted microscope. The back aperture of the objective is slightly overfilled, creating a diffractionlimited focal spot. Samples were placed in eight wells of a LabTek chambered cover glass (Nalge Nunc International, Rochester, NY) positioned in the X and Y directions by a motorized stage (Ma¨rzha¨user, Germany). The fluorescence from the samples was collected through the same objective and directed by a COWL750 dichroic mirror (Coherent, Orsay, France) toward a 50 µm diameter optical fiber coupled to an avalanche photodiode (SPCM 200 FC, EG&G, Canada). The residual infrared light was rejected by a BG39 Filter (Coherent). For FCS measurements, the normalized autocorrelation function (ACF), G(τ), of the fluorescence intensity fluctuations was calculated online by an ALV5000E digital correlator card (ALV, Langen, Germany). Calibration of the system was performed with a 50 nM tetramethylrhodamine (TMR) solution. Assuming a diffusion constant of 2.8 × 10-10 m2 s-1 (14), the equatorial (r0) and axial (z0) radii of the focal volume were, respectively, 0.29 and 1.3 µm, giving an effective volume of 0.2 fL. Assuming a three-dimensional Gaussian distributed excitation intensity, the ACF curves were fitted with eqs 1 and 2 of Clamme et al. (15). The imaging system on the same setup is based on the use of two galvo mirrors (model 6210, Cambridge Technology), in the so-called descanned detection mode. The two mirrors are used to deflect the beam along the X-axis and Y-axis, respectively. Each axis is driven by a closed-loop power amplifier, and the position of the mirrors is controlled through two ADC electronic boards (PCI6711, National Instruments). The photons were counted using a counter/timer board (PCI6602, National Instruments), synchronized with the scan of the galvo-mirrors. Nominally, one picture is 512 × 512 in size, and the dwelling time for each pixel is 4 µs (about 1 s per scan). PEI-1/DNA complexes for two-photon imaging were prepared in 20 mM HEPES, 150 mM NaCl, pH 7.4, buffer. Equal volumes of 5 mM PEI-1 (expressed in amine groups) and 1.4 mM DNA (expressed in phosphate groups) were mixed to reach a 60 µM final concentration of DNA and a N/P ratio of 10. Solutions were briefly vortexed and left for equilibration for 20 min before cell transfection. Prior to transfection, cells were seeded for 24 h in two wells of a Lab-Tek chambered coverglass at 105 cells per well in order to reach 70% confluence. After 24 h, the cells were rinsed and supplemented with 1.5 mL 10% FBS DMEM and 200 µL of complexes in order to deliver 4 µg of plasmid to the cells. After 2 h incubation, the cells were rinsed three times and covered with Hank’s balanced salt solution (HBSS) for two-photon imaging. Cell Transfection and Cytotoxicity. HeLa cells were incubated in serum-free Opti-MEM with pCMVLuc plasmid (2 µg/well) complexed with PEI-1 at N/P ratios ranging from 0.5 to 10. After 3 h 30 min, the transfection medium was replaced with fresh complete medium, and cells were cultured for an additional 24 h. Then, cells were lysed for luciferase

Bioconjugate Chem., Vol. 18, No. 3, 2007 847 Scheme 1. Synthetic Pathway for the Preparation of the Key Intermediate 2

Scheme 2. Preparation of the Isothiocyanate Amine Reactive Chromophore 1a

a (i) (a) Bu4NF, THF, (b) 8, Na2CO3; (ii) H2N-NH2, EtOH; (iii) CSCl2, CH2Cl2/DMAP.

activity quantization. Luciferase expression was recorded over 10 s using a Lumat LB9501/16 instrument (Berthold, Bad Wildbad, Germany) from an aliquot of the supernatant (5-10 µL) and expressed as RLU/mg of protein (luciferase activity per mg of protein in cell lysates). Cytotoxicity of the PEI-1/DNA complexes was evaluated by quantification of the cellular content in proteins. The cells were transfected with PEI-1/DNA complexes in serum-free Opti-MEM. Twenty-four hours after the addition of the transfection solution, the cells were washed three times in PBS and lysed by the addition of 100 µL of cell culture lysis reagent (Promega). The cellular proteins were measured using the bicinchoninic acid assay (Interchim, Montlucon, France) following an incubation time of 30 min at 60 °C. The absorbance was read at 562 nm (16).

RESULTS AND DISCUSSION Synthesis and Characterization of the IsothiocyanateFunctionalized Two-Photon Fluorophore. The reactive group chosen in this study is an isothiocyanate moiety for its versatility in bioconjugation with amines. The basic architecture of the target isothiocyanate fluorophore is identical to our previously described active two-photon excited fluorescent (TPEF) chromophores (6). On one side of this fluorophore, three identical oligo(ethylene glycol) chains are used to provide good water solubility. On the opposite side, we opted for a syringaldehyde derivative (three ether functions, two as methyl ether, one as linker for further conjugation) in order to keep three O-alkyl groups (Figure 2). To prepare the chromophore 1, the key intermediate is the protected phenol 2. To get this intermediate, we prepared first the TBDMS-protected syringaldehyde derivative 5 and followed then the synthetic pathway shown in Scheme 1. The bisphosphonate 4 was reacted with 1 equiv of TBDMSprotected syringaldehyde 5 in basic conditions at 0 °C, and the pure dissymmetric compound 6 was isolated in 35% yield from the reaction mixture containing the monophosphonate compound 6, the unreacted bis-phosphonate 4, and the bis-coupled symmetric molecule. Compound 6 was then reacted at room

848 Bioconjugate Chem., Vol. 18, No. 3, 2007

Hayek et al.

Figure 3. TPA spectra of 1 in water. Table 1. Spectroscopic Properties of the Isothiocyanate-Functionalized Fluorophore 1

Figure 4. Schematic structure of the conjugate PEI-1.

solvent

λmaxabs (nm)

 (M-1 cm-1)

λmaxex (nm)

λmaxem (nm)

Φ (%)

CH2Cl2 H2O

365 354

6.9 × 104 3.8 × 104

365 354

435 450

83 15

temperature with 1 equiv of aldehyde 7 to give the protected phenol derivative 2 in 44% yield. Chromophore 1 was finally obtained following the synthetic way described in Scheme 2. Fluoride deprotection of the TBDMS group affords the reactive phenolic moiety, which is etherified in situ with the commercially available N-(4-bromobutyl)phtalimide 8 according to Williamson conditions, to give the protected amine 9. The amino group is then deprotected by hydrazine in ethanol to yield the free amine 10. The isothiocyanate reactive group was finally obtained by reaction of the primary amine in 10 with thiophosgene. The targeted functionalized dye 1 was obtained in 95% yield. Structures of all new compounds have been confirmed by 1H and 13C NMR, and high-resolution mass spectrometry. Dye 1 has been designed to exhibit good water solubility to allow efficient conjugation in buffered aqueous media. Its solubility is over 10 g.L-1 in saline physiological solution (turbidimetric method). The linear photophysical properties of this novel functionalized dye are similar to those of the parent symmetrical chromophore previously described (6), since the basic photoactive structure remains unchanged. The absorption and emission characteristics of 1 in dichloromethane and water solutions are described in Table 1. This fluorophore presents an intense absorption band in the 365 nm region in dichloromethane and 354 nm in water. This dye also exhibits an intense blue fluorescence in both methylene chloride (435 nm) and water (450 nm). The nonlinear photophysical properties of 1 have been determined in water and methylene chloride from 480 to 650 nm in the visible, and from 700 to 900 nm in the IR (Figure 3). This chromophore exhibits a large TPA absorption in the redIR region over 250 GM at 700 nm and a good resistance to photobleaching. Both the one- and two-photon properties of this blue fluorescent dye are promising for two-photon applications in biology. PEI Conjugation and Spectroscopic Characterization of the Conjugated PEI. To show the potency of the isothiocyanate-functionalized chromophore to be used for two-photon microscopy applications, we conjugated PEI with derivative 1 (Figure 4). Conjugation was performed in borate buffer, pH 9.4, and unreacted dye was removed by gel chromatography. A conjugation of 2.6% of the primary amines of the PEI molecule, corresponding to an average number of 15 fluorophores per PEI molecule, was achieved. The absorption properties of the free chromophore 1 and PEI-1 are compared in Figure 5. Conjugation to PEI slightly affects the vibronic structure of the absorption spectrum of 1. In fact, the most pronounced change is the slight blue shift and relative increase of the vibronic peak

Figure 5. Absorption and fluorescence emission spectra of 1 and PEI1. Both compounds are in 20 mM HEPES and 150 mM NaCl buffer, pH 7.4. λex ) 354 nm.

Figure 6. Average size of the PEI-1/DNA binary complexes as a function of the N/P ratio. The complexes are in 20 mM HEPES and 150 mM NaCl buffer, pH 7.4. PEI/DNA (9); PEI-1/DNA (b); DNA complexes with a mixture of labeled and unlabeled PEI added at a 1:3 ratio (2).

at 361 nm, which becomes the strongest absorption peak of 1 in the conjugate. In the emission spectrum, no shift of the maximum emission accompanies the coupling of 1 to PEI, but only a slight relative increase of the emission shoulder at 480 nm. The quantum yield of fluorescence is not significantly affected by the conjugation of chromophore 1 with PEI (12% for PEI-1 compared to 15% for free dye). It thus appears that conjugation of 1 to PEI only marginally modifies the electronic properties of 1. Size of the PEI-1/DNA Complexes and Condensation of DNA in These Complexes. The size of the complexes of calf thymus DNA with PEI-1 at various N/P ratios was measured by light scattering and compared with the corresponding complexes obtained with nonlabeled PEI (Figure 6). In the case of nonlabeled PEI/DNA complexes, the size of the complexes increases with N/P, reaches a maximum at N/P ) 2, and then decreases at higher N/P ratio. At N/P ) 2, the complexes are neutral and thus have a strong tendency to aggregate, explaining the large size at this N/P ratio (17). In contrast, the smallest sizes are obtained at N/P g 5 when the complexes are strongly

Conjugation of TPA Fluorophore to PEI

Figure 7. Fluorescence spectra of YOYO-1-labeled DNA complexed with PEI and PEI-1. The concentrations of DNA (phosphate) and YOYO-1 were 60 and 1.2 µM, respectively, to ensure a ratio of 1 YOYO-1 molecule per 50 bases. Spectra were recorded in the absence (solid line) or presence of PEI (dashed line) or PEI-1 (dash-dotted line) at N/P ratios of 2 (top spectra) and 5 (bottom spectra). Samples were prepared as described in Figure 4. Excitation was at 470 nm.

positively charged (8) and have thus no tendency to aggregate, due to electrostatic repulsions. A similar qualitative dependence of the size of the complexes as a function of N/P is obtained when PEI is substituted by PEI-1. However, two major differences are immediately apparent. The first one is the significant shift of the N/P ratio (from 2 to 5) at which the complexes are probably neutral and exhibit a size maximum. The second difference is the larger size of the complexes obtained with PEI-1, not only at the size maximum (650 nm vs 500 nm) but also at higher N/P ratios. Indeed, even at N/P ) 10, the DNA/PEI-1 complexes are about 350 nm and, thus, more than two times bigger than the complexes with unlabeled PEI. Both differences may be partly attributed to the decrease of the amine content of PEI-1 due to labeling. This decreases the charge density of PEI and, thus, its ability to condense DNA, since condensation is strongly dependent on the extent of charge neutralization (18). The second and probably most important reason for the differences between labeled and unlabeled complexes may be related to the large size of dye 1 that may sterically hinder the proper neutralization of DNA charges by the labeled PEI and, consequently, the optimal compaction of the bound DNA. These drawbacks could be partially avoided if the complexes are obtained by adding labeled PEI together with unlabeled PEI at a ratio of 1:3 to the DNA. In these conditions, the complexes at N/P g 8 are the same size as the complexes with unlabeled PEI, and the size maximum is shifted to N/P ) 3. Importantly, both the larger maximum size of the latter complexes and the systematic shift of their sizes at N/P < 8 with respect to the unlabeled complexes suggest that both labeled and unlabeled PEI molecules are bound in the complexes with DNA. It may thus be inferred that, in these complexes, unlabeled PEI largely compensates for the defects of DNA compaction induced by the labeled PEI. To further investigate the condensation level of DNA in its complexes with PEI and PEI-1, we used the fluorescent DNA bis-intercalator YOYO-1 (19). This compound is characterized by a very high affinity for DNA (20) and remains bound to DNA during condensation by PEI (19). Moreover, the fluorescence intensity of free YOYO-1 is 3200 times lower than that of the bound form and could thus be neglected. When the level of DNA-bound YOYO-1 is sufficiently large (1 YOYO-1 for 50 bases), condensation by PEI was shown to induce a dramatic fluorescence decrease due to electronic interaction among YOYO-1 molecules bound on the same DNA molecule (19). Fluorescence spectra of YOYO-1 bound to DNA complexed with PEI and PEI-1 at different N/P ratios are given in Figure 7. In excellent agreement with our conclusions above, the fluorescence of YOYO-1 in complexes with PEI-1 was significantly higher

Bioconjugate Chem., Vol. 18, No. 3, 2007 849

Figure 8. Transfection efficiency of PEI-1/DNA complexes. HeLa cells were incubated with pCMVLuc plasmid (2 µg/well) complexed to PEI-1 (left), unlabeled PEI (center), or a mixture of PEI-1 and unlabeled PEI (at a 1:3 ratio, right) in serum-free Opti-MEM. After 3 h 30 min, the transfection medium was replaced with fresh complete medium, and cells were cultured for an additional 24 h, before lysis. Gene expression was determined by the luciferase assay and expressed as RLU/mg of protein (luciferase activity per mg of protein in cell lysates). JetPEI was used as a control.

than in complexes with unlabeled PEI at N/P ratios of 2 and 5, indicating that PEI-1 less efficiently condenses DNA than unlabeled PEI. Transfection Efficiency and Cytotoxicity of DNA/PEI-1 Complexes. PEI-1/DNA complexes were assessed for their in vitro transfection efficiency by monitoring the transient expression of the luciferase reporter gene in Hela cells. At N/P e 5, the complexes gave luciferase signals (103 RLU/mg protein, Figure 8) that were 3 to 4 orders of magnitude lower than those obtained with unlabeled PEI but comparable to that of naked DNA (data not shown). This indicates that complexes with PEI-1 are unable to transfect cells at these ratios. In sharp contrast, PEI-1/DNA complexes and PEI/DNA complexes show comparable transfection efficiencies at N/P > 5. Moreover, complexes obtained by mixing DNA with PEI-1 and PEI at a 1:3 ratio were found to be as active as PEI/DNA complexes at all N/P ratios. These data show that by analogy to a large variety of compounds (poly(ethylene glycol), sugars, peptides, etc.; for a review, see Godbey et al. (9, 21, 22)), the conjugation of 1 does not compromise the transfection properties of PEI. In fact, the poor transfection observed with PEI-1/DNA complexes at N/P e 5 may be rationalized by the size measurements, which suggest that the complexes may be negatively charged or neutral at these ratios and thus cannot efficiently interact with the cell membrane and be internalized by the cells (8, 23). Finally, cell viability experiments using the bicinchoninic acid assay showed that only 10% of dead cells were observed after transfection with PEI-1/DNA or PEI/DNA complexes at N/P ) 10 (data not shown), indicating that conjugation with 1 does not increase the cytotoxicity of PEI. Two-Photon FCS and Microscopy of PEI-1/DNA Complexes. To demonstrate the validity of PEI-1 in two-photon applications, we first performed two-photon FCS measurements of PEI-1 either in its free form or bound to DNA. Two-photon excitation of the dye was provided by a tunable Ti:sapphire laser. Due to the optical specifications of our setup, it was not possible to excite the sample at wavelengths lower than 760 nm. Unfortunately, the TPA cross section of 1 at 760 nm is far from its maximum (below 700 nm, Figure 3). As a consequence, the brightness per molecule did no exceed 0.5 kHz/molecule and the signal-to-noise ratio was fair in the FCS measurements. Nevertheless, the fit of the FCS curves of free PEI-1 (Figure 9) provided an observed diffusion constant of 1.6 × 10-10 m2.s-1 that is consistent with the previously reported diffusion constant of Rhodamine-labeled PEI (15). Moreover, in further agreement with the Rhodamine-labeled PEI data, a population with a slower diffusion constant ((3.8 ( 0.3) × 10-11 m2.s-1) was detected

850 Bioconjugate Chem., Vol. 18, No. 3, 2007

Hayek et al.

10). This shows that conjugation of PEI by the chromophore does not compromise the efficient internalization of free and bound PEI by the cells and their accumulation in perinuclear compartments, identified as endosomes.

CONCLUSION

Figure 9. Autocorrelation curves of free PEI-1 and PEI-1/DNA complexes. The concentration of PEI-1 was 600 µM for both curves. The complexes were obtained at N/P ) 10. The red and blue lines correspond to the best fits of the autocorrelation curves of PEI-1 and PEI-1/DNA complexes, according to eq 2 of Clamme et al. (2003) with the diffusion coefficients given in the text. In the inset, the fitted curves were normalized and their bottom part was highlighted to show their differences.

We described herein the synthesis and characterization of a functionalized two-photon absorbing chromophore with blue emission. This chromophore was successfully conjugated to PEI and was found to substitute unlabeled PEI, without compromising the transfection properties of PEI. In spite of a nonoptimized excitation wavelength, the labeled PEI was successfully used in two-photon FCS and imaging. The obtained results were close to those previously reported for Rhodamine-labeled PEI. However, due to its rather large size, the dye decreases to some extent the condensing properties of PEI, affecting the size of the PEI/DNA complexes. Thus, it is recommended to use them in a mixture with unlabeled PEI, since complexes obtained with a 3:1 ratio of unlabeled to labeled PEI were found to be highly similar to the unlabeled PEI/DNA complexes.

ACKNOWLEDGMENT Part of this work was supported by the CNRS and University Louis Pasteur of Strasbourg. A.H. is grateful to the French ministry of research for a fellowship. Other part of this work and S.E. have been supported by the Agence Franc¸ aise contre les Myopathies (AFM) and CNRS.

LITERATURE CITED Figure 10. Two-photon microscopy of free PEI-1 (A) and PEI-1/ DNA complexes (B) in HeLa cells. Free PEI-1 (600 µM) and PEI1/DNA complexes (60 µM DNA, N/P ) 10) were incubated for 2 h with HeLa cells, as described in Materials and Methods. Images were taken 30 min, post-transfection. Each bar represents 10 µm.

in some autocorrelation curves. However, since this diffusion constant is significantly different from that associated with the complexes and since its contribution to the autocorrelation curve is limited, this population corresponds to a minor species that cannot interfere with PEI-1/DNA measurements. Next, the FCS curve of the PEI-1/DNA complexes at N/P ) 10 was recorded. In full agreement with our previous data using rhodaminelabeled PEI (15), the autocorrelation curves of PEI-1/DNA complexes were similar to the curves obtained with free PEI-1 and, thus, were largely dominated by the contribution of free PEI. This confirms that a large proportion of PEI remains free when PEI is mixed with DNA. In fact, the only difference with free PEI is the presence of a small shoulder at long times in the autocorrelation curve of the complexes. The autocorrelation curves could be adequately fitted with a two-population model. If the correlation times of the two populations were allowed to float, the short correlation time was typical of that of free PEI. According to its small contribution, the correlation time (which is inversely related to the diffusion constant) of the complexes fluctuates in a larger range than that of free PEI. Nevertheless, if we assume that complexes are spherical, we obtained from their diffusion constant ((1.2 ( 0.6) × 10-12 m2.s-1) a diameter of 360 nm that is in excellent agreement with that determined by light scattering (Figure 6). In a next step, two-photon imaging of HeLa cells incubated with free PEI-1 or transfected with PEI-1 complexes at N/P ) 10 was performed with the same setup as for FCS. In excellent agreement with previous studies (15, 24), both free PEI-1 and PEI-1/DNA complexes were found in discrete fluorescence patches that accumulate around the nucleus (Figure

(1) Abbotto, A., Baldini, G., Beverina, L., Chirico, G., Collini, M., D’Alfonso, L., Diaspro, A., Magrassi, R., Nardo, L., and Pagani, G. A. (2005) Dimethyl-pepep: a DNA probe in two-photon excitation cellular imaging. Biophys. Chem. 114, 35-41. (2) Krishna, T. R., Parent, M., Werts, M. H., Moreaux, L., Gmouh, S., Charpak, S., Caminade, A. M., Majoral, J. P., and BlanchardDesce, M. (2006) Water-soluble dendrimeric two-photon tracers for in vivo imaging. Angew. Chem., Int. Ed. 45, 4645-4648. (3) Squirrell, J. M., Wokosin, D. L., White, J. G., and Bavister, B. D. (1999) Long-term two-photon fluorescence imaging of mammalian embryos without compromising viability. Nat. Biotechnol. 17, 763767. (4) Berland, K. M., So, P. T., and Gratton, E. (1995) Two-photon fluorescence correlation spectroscopy: method and application to the intracellular environment. Biophys. J. 68, 694-701. (5) Medina, M. A., and Schwille, P. (2002) Fluorescence correlation spectroscopy for the detection and study of single molecules in biology. Bio-essays 24, 758-764. (6) Hayek, A., Bolze, F., Nicoud, J.-F., Baldeck, P. L., and Me´ly, Y. (2006) Synthesis and characterization of water-soluble two-photon excited blue fluorescent chromophores for bioimaging. Photochem. Photobiol. Sci. 5, 102-106. (7) Hayek, A., Bolze, F., Nicoud, J.-F., Dupperay, A., Grichine, A., Baldeck, P., and Vial, J.-C. (2006) Two-photon water-soluble dyes and their amine reactive derivatives for two-photon bio-imaging applications. Nonlinear Opt., Quantum Opt. 35, 155-164. (8) Boussif, O., Lezoualc’h, F., Zanta, M. A., Mergny, M. D., Scherman, D., Demeneix, B., and Behr, J.-P. (1995) A versatile vector for gene and oligonucleotide transfer into cells in culture and in vivo: polyethylenimine. Proc. Natl. Acad. Sci. U.S.A. 92, 72977301. (9) Godbey, W. T., Wu, K. K., and Mikos, A. G. (1999) Poly(ethylenimine) and its role in gene delivery. J. Controlled Release 60, 149-160. (10) Baker, A., Saltik, M., Lehrmann, H., Killisch, I., Mautner, V., Lamm, G., Christofori, G., and Cotten, M. (1997) Polyethylenimine (PEI) is a simple, inexpensive and effective reagent for condensing and linking plasmid DNA to adenovirus for gene delivery. Gene Ther. 4, 773-782.

Bioconjugate Chem., Vol. 18, No. 3, 2007 851

Conjugation of TPA Fluorophore to PEI (11) Wightman, L., Kircheis, R., Rossler, V., Carotta, S., Ruzicka, R., Kursa, M., and Wagner, E. (2001) Different behavior of branched and linear polyethylenimine for gene delivery in vitro and in vivo. J. Gene Med. 3, 362-372. (12) Xu, C., and Webb, W. W. (1996) Measurement of two-photon excitation cross sections of molecular fluorophores with data from 690 to 1050 nm. J. Opt. Soc. Am. B 13, 481-491. (13) Snyder, S. L., and Sobocinski, P. Z. (1975) An improved 2,4,6trinitobenzenesulfonic acid method for the determination of amines. Anal. Biochem. 64, 284-288. (14) Egele, C., Schaub, E., Pie´mont, E., de Rocquigny, H., and Me´ly, Y. (2005) Investigation by fluorescence correlation spectroscopy of the chaperoning interactions of HIV-1 nucleocapsid protein with the viral DNA initiation sequences. C. R. Biol. 328, 10411051. (15) Clamme, J. P., Krishnamoorthy, G., and Me´ly, Y. (2003) Intracellular dynamics of the gene delivery vehicle polyethylenimine during transfection: investigation by two-photon fluorescence correlation spectroscopy. Biochim. Biophys. Acta 1617, 52-61. (16) Zaric, V., Weltin, D., Erbacher, P., Re´my, J. S., Behr, J.-P., and Stephan, D. (2004) Effective polyethylenimine-mediated gene transfer into human endothelial cells. J. Gene Med. 6 (2), 176184. (17) Choosakoonkriang, S., Lobo, B. A., Koe, G. S., Koe, J. G., and Middaugh, C. R. (2003) Biophysical characterization of PEI/DNA complexes. J. Pharm. Sci. 92, 1710-1722.

(18) Arscott, P., Li, A., and Bloomfield, V. A. (1990) Condensation of DNA by trivalent cations 1: effect of DNA length and topology on the size and shape of condensed particles. Biopolymers 30, 619630. (19) Krishnamoorthy, G., Duportail, G., and Me´ly, Y. (2002) Structure and dynamics of condensed DNA probed by 1,1′-(4,4,8,8-tetramethyl-4,8-diazaundecamethylene)bis[4-[[3- methylbenz-1,3-oxazol2-yl]methylidine]-1,4-dihydroquinolinium] tetraiodide fluorescence. Biochemistry 41, 15277-15287. (20) Hirons, G. T., Fawcett, J. J., and Crissman, H. A. (1994) TOTO and YOYO: new very bright fluorochromes for DNA content analyses by flow cytometry. Cytometry 15, 129-140. (21) Lungwitz, U., Breunig, M., Blunk, T., and Gopferich, A. (2005) Polyethylenimine-based non-viral gene delivery systems. Eur. J. Pharm. Biopharm. 60, 247-266. (22) Kichler, A. (2004) Gene transfer with modified polyethylenimines. J. Gene Med. 6 Suppl. 1, S3-10. (23) Behr, J.-P. (1994) Gene transfer with synthetic cationic amphiphiles: prospects for gene therapy. Bioconjugate Chem. 5, 382389. (24) Remy-Kristensen, A., Clamme, J.-P., Vuilleumier, C., Kuhry, J. G., and Me´ly, Y. (2001) Role of endocytosis in the transfection of L929 fibroblasts by polyethylenimine/DNA complexes. Biochim. Biophys. Acta 1514, 21-32. BC060362H