In Vitro Lipofection with Novel Asymmetric Series of 1,2

Bioconjugate Chem. 2005, 16, 1411−1422

1411

In Vitro Lipofection with Novel Asymmetric Series of 1,2-Dialkoylamidopropane-Based Cytofectins Containing Single Symmetric Bis-(2-dimethylaminoethane) Polar Headgroups Michalakis Savva,* Pensung Chen, Ahmad Aljaberi, Bilge Selvi, and Michael Spelios Division of Pharmaceutical Sciences, Arnold & Marie Schwartz College of Pharmacy and Health Sciences, Long Island University, 75 Dekalb Avenue, Brooklyn, New York 11201. Received May 11, 2005; Revised Manuscript Received August 31, 2005

Novel N,N′-diacyl-1,2-diaminopropyl-3-carbamoyl[bis-(2-dimethylaminoethane)] bivalent cationic lipids were synthesized and evaluated for in vitro transfection activity against a murine melanoma cell line. In the absence of the helper lipid DOPE (1,2-dioleoyl-sn-glycero-3-phosphoethanolamine), only the dioleoyl derivative 22 (1,2lb5) elicited transfection activity. The transfection activity of this lipid was reduced when formulated with DOPE. Contrary to that, the dimyristoyl derivative 19 (1,2lb2) mediated no activity when used alone but induced the highest levels of marker gene expression in the presence of DOPE. In an effort to correlate the transfection activity with cationic lipid structures, the physicochemical properties of cationic lipids in isolation and of lipoplexes were studied with surface tensiometry, photon correlation spectroscopy, gel electrophoresis mobility shift assay, and fluorescence techniques. In regard to the lipoplex properties, gel electrophoresis mobility shift assay and EtBr exclusion fluorescence assay revealed that the 1,2lb5 was the only lipid to associate and condense plasmid DNA, respectively. Photon correlation spectroscopy analysis found that 1,2lb5/DNA complexes were of relatively small size compared to all other lipoplexes. With respect to the properties of isolated lipids, Langmuir monolayer studies and fluorescence anisotropy on cationic lipid dispersions verified high two-plane elasticity and increased fluidity of the transfection competent dioleoyl derivative 1,2lb5, respectively. The results indicate that high transfection activity is mediated by cationic lipids characterized by an expanded mean molecular area, high molecular elasticity, and increased fluidity.

INTRODUCTION

Gene delivery reagents comprising a variety of recombinant replication deficient viruses (1, 2), positively charged synthetic polymers and peptides and cationic lipids (3-5), are currently being used to treat genetic diseases (6) and a wide variety of cancers (7). These delivery systems have their pluses and minuses. Viral vectors unambiguously induce much higher transgene expression levels than their nonviral rivals. Their in vivo application, however, is severely compromised by the innate immunity and by the more refined adaptive immunity, as provided by the natural killer cells, biochemicals and cellular proteins of the patient, and neutralizing antibodies, respectively. On the other hand, cationic lipids are a particularly attractive alternative to the viral vectors because they do not induce any humoral responses. Moreover, it is possible to design and synthesize a wide variety of cationic lipids that, depending on their structural features, promote gene expression by affecting one or more of the essential requirements for transfection, i.e., efficient compaction of plasmid DNA, lipoplex internalization, fusion of the internalized lipoplex with lysosome membrane, and translocation of the plasmid DNA to the nucleus. Unfortunately, the transfection activity of nonviral vectors is currently very low to permit disease treatment. Thus, for gene therapy to become common medicine therapeutics, current vehicles have to be improved significantly. Nevertheless, despite the present lack of clinical success of gene therapy, the * To whom correspondence should be addressed. Phone: 718488-1471. Fax: 718-780-4586. E-mail: [email protected].

promise that it holds to eradicate the cause of a disease rather than to treat the symptoms of it has kept the drive for discovering efficient cationic transfection lipids active, for more than 17 years. Toward this effort, we have recently launched a systematic investigation of the transfection potency of a series of positively charged lipids containing primary and tertiary amine groups attached to the 2-position of the 1,3-diamino-2-propanol backbone, whereas the acyl chains were linked to the backbone at the 1- and 3-positions via amide bonds (8, 9). Herein, we report on the synthesis and transfection activity of novel cationic lipids of the 1,2-diamino-3-propanol series, in which the polar headgroup bis-(2-dimethylaminoethane) is attached via a carbamate spacer at the 3-position of the 1,2-diaminopropan-3-ol, whereas the hydrophobic alkyl chains are bonded to the nitrogen atoms at the 1- and 2-position of the diaminopropanol backbone through acyl linkers. The physicochemical properties of the cationic lipids in isolation were investigated in an effort to correlate structures with transfection activity. MATERIALS AND METHODS

Materials. Bis-(2-chloroethyl)amine hydrochloride, 98%, 2,3-diaminopropionic acid monohydrochloride 98%, lauroyl chloride 98%, myristoyl chloride 97%, palmitoyl chloride 98%, stearoyl chloride 99%, oleoyl chloride technical grade (∼85%), 4-nitrophenyl chloroformate 97%, hydrogen chloride gas anhydrous 99+%, lithium borohydride 2.0 M solution in THF, tetrahydrofuran (THF) anhydrous 99.9%, triethylamine 99.5% (TEA), ethyl alcohol 99.5%, N,N-dimethylformamide (DMF) 99%,

10.1021/bc050138c CCC: $30.25 © 2005 American Chemical Society Published on Web 11/01/2005

1412 Bioconjugate Chem., Vol. 16, No. 6, 2005

Figure 1. Synthetic scheme for the preparation of 1,2-dialkoylamidopropane-based lipids: R ) C11H23 (18, 1,2lb1), C13H27 (19, 1,2lb2), C15H31 (20, 1,2lb3), C17H35 (21, 1,2lb4), C17H33 (22, 1,2lb5).

and pyridine 99+% were purchased from Aldrich. Dimethylamine 40% was purchased from ACROS ORGANICS. Methanol (anhydrous) and all other reagents used were from VWR. All tissue culture reagents were purchased from Life Technologies Inc. (Rockville, MD), unless otherwise specified. Synthesis. The synthesis of the bivalent cationic lipids is summarized in Figure 1. Mobile phase systems v/v for thin layer chromatography (TLC) (A) CH3OH/CHCl3 (1: 1), (B) CH3OH/CHCl3 (1:4), (C) CH3OH/CHCl3 (1:9), (D) CH3OHCHCl3 (1: 20), (E) CH3OH/CHCl3/NH4OH (1:2: 0.15). Bis-(2-(dimethylamino)ethyl)amine. Solid bis-(2chloroethyl)amine hydrochloride (17.85 g, 0.10 mol) was dissolved in a 500 mL round-bottom flask with addition of 152 mL of dimethylamine (54 g, 1.20 mol). After a 72 h of rigorous stirring at room temperature, the mixture was made alkaline with 6 M NaOH (100 mL) and further saturated with anhydrous potassium carbonate. Upon alkalization an oily liquid separates on the upper layer. The mixture was stirred for one-half hour and then transferred to a 500 mL separation funnel where the crude product was extracted with 4 × 150 mL of ethyl ether (ninhydrin test of the organic phase turns negative). The extracts were collected and dried (MgSO4) overnight. MgSO4 was removed by suction filtration and the organic solvent was eliminated under diminished pressure at 45 °C to give a total of 5.0 g of colorless oil (yield 31%). Further purification of the triamine was

Savva et al.

performed by chromatography with 50% recovery (10). MS (FAB and ES) m/z 160.2 [M + H]+; 1H NMR (400 MHz, CDCl3, 20 °C, TMS) δ 2.60-2.63 (t, 4H, J ) 6.2 Hz, NCH2CH2N(CH3)2), 2.31-2.34 (t, 4H, J ) 6.2 Hz, CH2N(CH3)2), 2.13 (s, 12H, N(CH3)2); 13C NMR (100 MHz, CDCl3, 20 °C, TMS) δ 59.21 (2CH2), 47.48 (2CH2), 45.60 (4CH3). Methyl 2,3-Diaminopropionate Dihydrochloride (2). 2,3-Diaminopropionic acid monohydrochloride (2.0 g, 14.2 mmol) was pulverized and transferred to a 500 mL round-bottom flask containing 50 mL of anhydrous methanol. Bound water was removed by evaporating the solvent several times at low pressure at 75 °C. The dry powder was finally dissolved in fresh anhydrous methanol (100 mL) saturated with anhydrous hydrogen chloride gas. The mixture was refluxed at 85-90 °C for 17 h. Methanol was eliminated under diminished pressure to give 2.85 g as a yellowish crystal of 86-90% purity, depending on the batch, as determined by 1H NMR. Rf ) 0.86 (A). 1H NMR (400 MHz, D2O, 20 °C, TMS) δ 4.264.34 (t, 1H, CH2CH(NH2)CO), 3.66 (s, 3H, COOCH3), 3.22-3.44 (m, 2H, CH2CH). Methyl 2,3-Dilauroylamidopropionate (3). Methyl 2,3-dilauroylamidopropionate was synthesized using a procedure similar to that described by Sunamoto and coworkers (11). Briefly, to a 100 mL anhydrous DMF solution of methyl 2,3-diaminopropionate dihydrochloride (1.36 g, 7.11 mmol) in a 250 mL round-bottom flask was added 10.0 mL of TEA (71.1 mmol) immediately followed by dropwise addition of 7.0 mL of lauroyl chloride (28.4 mmol). The mixture was stirred at 60 °C for 6 h. DMF was evaporated with the aid of a rotary evaporator at 90 °C under reduced pressure, and the crude material was transferred with 100 mL of CHCl3 to a 500 mL separation funnel where it was washed once with HCl and once with Na2CO3. The organic phase was collected in a 250 mL round-bottom flask, and the CHCl3 was removed under reduced pressure with the aid of a rotary evaporator. The dried crude was dissolved in a minimum amount of hot methanol and left overnight for crystallization. The product was collected by suction filtration as a white crystalline powder (2.1 g, 61.3%). Rf ) 0.76 (C). 1H NMR (400 MHz, CDCl3, 20 °C, TMS) δ 6.89 (d, 1H, HNCH), 6.25 (t, 1H, HNCH2), 4.57 (m, 1H, CHCOOCH3), 3.72 (s, 3H, COOCH3), 3.60 (t, 2H, HNCH2CH), 2.11-2.21 (m, 4H, COCH2), 1.58 (m, 4H, COCH2CH2), 1.21 (b, 32H, (CH2)8CH3), 0.82-0.85 (t, 6H, CH3). 2,3-Dilauroylamidopropan-3-ol (8). To a 100 mL solution of methyl 2,3-dilauroylamidopropionate (3.52 g, 7.29 mmol) in absolute ethanol in a 250 mL round-bottom flask was added 10.9 mL (21.87 mmol) of 2.0 M lithium borohydride solution in THF. Hydrogenation was performed in a dry N2 gas environment at 50-60 °C for 3 h and then at 23 °C for an additional 12 h. Ethanol was removed, and the crude material was taken with 100 mL of chloroform, transferred to a 500 mL separation funnel, and washed once with 100 mL of 1.0 N HCl and once with Na2CO3. The organic phase was concentrated to an oil of high viscosity, which was applied to a silica gel column (3.2 cm × 30 cm). Elution was performed with 100 mL of CHCl3 and 1%, 1.5%, 2%, 2.5%, 3%, 10%, 20%, and 50% MeOH/CHCl3. Fractions between 2% and 20% were pooled and concentrated to give a total of 2.94 g of 2,3-dilauroylamidopropanol as a white powder (89% yield). Rf ) 0.27 (D). 1H NMR (400 MHz, CDCl3, 20 °C, TMS) δ 6.25-6.50 (b, 2H, HNCH2HNCH), 3.60-3.90 (m, 2H, HNCH2CH), 3.40-3.55 (m, 2H, CH(NH)CH2OH), 3.20-3.30 (m, 1H, CH), 2.12-2.22 (m, 4H, COCH2), 1.59

Novel Cationic Lipids for Gene Delivery

(b, 4H, COCH2CH2), 1.23 (b, 32H, (CH2)8CH3), 0.83-0.86 (t, 6H, CH3). 2,3-Dilauroylamidopropane-1-(p-nitrophenyl) Carbonate (13). 2,3-Dilauroylamidopropanol (2.94 g, 6.47 mmol) was suspended with continuous stirring in 100 mL of anhydrous THF in a 250 mL capacity round-bottom flask. The mixture was kept at 25 °C with the aid of an oil bath. To the suspension was added 4-nitrophenyl chloroformate (1.30 g, 6.47 mmol) followed by a dropwise addition of pyridine (0.52 mL, 6.47 mmol). The reaction was stopped after 5 h. THF was removed under diminished pressure, and the crude was transferred with 100 mL of chloroform to a 500 mL separation funnel and washed once with 100 mL of 1.0 N HCl and once with 100 mL of Na2CO3. The organic phase was concentrated to a viscous oil with the aid of a rotary evaporator and purified by chromatography on a silica gel column, eluting with 100 mL of 1%, 2%, 3%, 4%, 5%, 20%, 50% MeOH/CHCl3. Fractions between 3% and 20% were pooled and dried to give a total of 3.9 g (6.28 mmol, 97% yield) of 2,3-dilauroylamidopropane-1-(p-nitrophenyl) carbonate as a white crystalline material. Rf ) 0.49 (D). 1 HNMR (400 MHz, CDCl3, 20 °C, TMS) δ 8.27-8.24, 7.37-7.23 (two d, each J ) 9.1 Hz, 4H, C6H4), 6.64 (d, 1H, HNCH), 6.02 (t, 1H, HNCH2), 4.18-4.40 (m, 3H, NHCHCH2OCOO), 3.42-3.57 (m, 2H, HNCH2CH), 2.142.20 (m, 4H, COCH2), 1.57-1.79 (m, 4H, COCH2CH2), 1.0-1.24 (bs, 32H, (CH2)8CH3), 0.82-0.90 (t, 6H, CH3). 1,2-Dilauroylamidopropane-3-[bis-(2-dimethylaminoethane)] Carbamate (18, 1,2lb1). To a solution of 2,3-dilauroylamidopropane-1-(p-nitrophenyl) carbonate (3.9 g, 6.28 mmol) in 100 mL of chloroform was added bis-(2-dimethylaminoethyl)amine (1.7 mL, 9.42 mmol). The mixture was stirred at room temperature for 17 h. The reaction mixture was then transferred to a 500 mL separation funnel where it was washed once with 100 mL of 1.0 N Na2CO3. The organic phase was concentrated to an oil, which was chromatographed on a silica gel column, eluting with 100 mL of 2%, 4%, 200 mL of 5%, 100 mL of 7%, 10%, and 200 mL of 15%, 20%, 25%, 30%, 35%, 40% MeOH/CHCl3. Fractions between 15% and 40% were pooled and dried to obtain a total 2.95 g (73%). Rf ) 0.46 (E). Anal. Calcd for C36H73N5O4 (MW 639): C, 67.56; H, 11.50; N, 10.94. Found: C, 66.97; H, 11.77; N, 10.90. MS (ESI) m/z 662.5 [M + Na]+; 1H NMR (400 MHz, CDCl3, 20 °C, TMS) δ 7.15 (d, 1H, HNCH), 6.70 (t, 1H, HNCH2), 4.13 (m, 3H, NHCHCH2OCO), 3.20-3.50 (m, 6H, NHCH2CH(NH)CH2OCON(CH2)2-), 2.35-2.42 (m, 4H, CH2N(CH3)2), 2.19-2.20 (d, 12H, N(CH3)2), 2.05-2.18 (m, 4H, COCH2), 1.45-1.60 (m, 4H, COCH2CH2), 1.187 (b, 32H, (CH2)8CH3), 0.80-0.90 (t, 6H, CH3); 13C NMR (100 MHz, CDCl3, 20 °C) δ 176.04 (NHCO), 175.34 (NHCO), 157.47 (OCO), 65.30, 58.63, 46.83, 37.75, 32.96, 30.68, 30.41, 27.35, 23.73, 15.17. Methyl 2,3-Dimyristoylamidopropionate (4). Methyl 2,3-dimyristoylamidopropionate was crystallized from methanol analogously to compound 3 to afford a total of 1.3 g as white crystals (68%). Rf ) 0.89 (B), Rf ) 0.82 (C). 1H NMR (400 MHz, CDCl3, 20 °C, TMS) δ 6.77 (d, 1H, HNCH), 6.15 (t, 1H, HNCH2), 4.55 (m, 1H, CHCOO), 3.72 (s, 3H, COOCH3), 3.60 (t, 2H, HNCH2), 2.11-2.21 (m, 4H, COCH2), 1.56 (m, 4H, COCH2CH2), 1.22 (b, 40H, (CH2)10CH3), 0.82-0.85 (t, 6H, (CH2)10CH3); 13C NMR (100 MHz, CDCl3, 20 °C) δ 175.77 (NHCO), 174.95 (NHCO), 171.82 (COO), 54.72, 51.74, 46.64, 37.60, 37.48, 32.93, 30.68, 30.38, 26.55, 26.71, 23.70, 15.11. 2,3-Dimyristoylamidopropanol (9). This compound was prepared similarly to compound 8 with lithium borohydride at 60 °C for 5 h and gave the product as a

Bioconjugate Chem., Vol. 16, No. 6, 2005 1413

white powder (2.23 g, 99%). Rf ) 0.27 (D). 1H NMR (400 MHz, CDCl3, 20 °C, TMS) δ 6.44-6.55 (m, 2H, HNCH2HNCH), 3.40-3.90 (m, 4H, HNCH2CH(NH)CH2OH), 3.20-3.30 (m, 1H, HNCH2CH), 2.13-2.21 (m, 4H, COCH2), 1.56 (m, 4H, COCH2CH2), 1.21 (b, 40H, (CH2)10CH3), 0.82-0.85 (t, 6H, CH3). 2,3-Dimyristoylamidopropane-1-(p-nitrophenyl) Carbonate (14). Compound 14 was prepared analogously to 13 and purified by silica gel column chromatography (70-230, 3.2 cm × 30 cm), eluting with 100 mL of 1%, 1.5%, 2%, 2.5%, 3%, 3.5%, 4%, 5%, 40% MeOH/ CHCl3. The 2-4% fractions were pooled to give 1.16 g (1.7 mmol, 79.5%) of the product as a white solid. Rf ) 0.71 (D). 1H NMR (400 MHz, CDCl3, 20 °C, TMS) δ 8.268.23, 7.36-7.34 (two d, each J ) 9.1 Hz, 4H, C6H4), 6.77 (d, 1H, HNCH), 6.25 (t, 1H, HNCH2), 4.20-4.38 (m, 3H, NHCHCH2OCOO), 3.30-3.60 (m, 2H, HNCH2CH), 2.162.18 (m, 4H, COCH2), 1.56 (m, 4H, COCH2CH2), 1.20 (b, 40H, (CH2)10CH3), 0.81-0.85 (t, 6H, CH3). 1,2-Dimyristoylamidopropane-3-[bis-(2-dimethylaminoethane)] Carbamate (19, 1,2lb2). 19 was prepared analogously to 18 on a 1.5 mmol scale to give upon purification the product as a white solid (0.6 g, 0.87 mmol, 57%). Rf ) 0.76 (E). Anal. Calcd for C40H81N5O4 (MW 695): C, 69.02; H, 11.73; N, 10.06. Found: C, 68.74; H, 11.99; N, 9.98. MS (FAB) m/z 697.1 [M + H]+. 1H NMR (400 MHz, CDCl3, 20 °C, TMS) δ 7.15 (d, 1H, HNCH) 6.70 (t, 1H, HNCH2), 4.13 (m, 3H, CH(NH)CH2OCO), 3.24-3.46 (m, 6H, NHCH2CH(NH)CH2OCON(CH2)2), 2.36-2.47 (m, 4H, CH2N(CH3)2), 2.21-2.23 (d, 12H, N(CH3)2), 2.12 (m, 4H, COCH2), 1.56-1.58 (m, 4H, COCH2CH2), 1.24 (b, 40H, (CH2)10CH3), 0.80-0.90 (t, 6H, CH3); 13C NMR (100 MHz, CDCl3, 20 °C) δ 176.16 (NHCO), 175.34 (NHCO), 157.35 (OCO), 65.76, 58.63, 46.83, 46.37, 46.25, 37.69, 32.96, 30.71, 30.41, 27.05, 23.73, 15.20. Methyl 2,3-Dipalmitoylamidopropionate (5). This compound was prepared similarly to 4 and gave the product as white crystals (2.9 g, 4.9 mmol, 68%). Rf ) 0.76 (C). 1H NMR (400 MHz, CDCl3, 20 °C, TMS) δ 6.80 (d, 1H, HNCH), 6.18 (t, 1H, HNCH2), 4.58 (m, 1H, CHCOOCH3), 3.73 (s, 3H, COOCH3), 3.61 (m, 2H, HNCH2CH), 2.14-2.20 (m, 4H, COCH2), 1.57-1.58 (m, 4H, COCH2CH2), 1.24 (b, 48H, (CH2)12CH3), 0.83-0.86 (t, 6H, (CH2)12CH3). 2,3-Dipalmitoylamidopropan-3-ol (10). Compound 10 was prepared analogously to 9 on a 4.9 mmol scale and gave the product, after chromatography (CHCl3/ MeOH), as a white powder (2.5 g, 89%). Rf ) 0.37 (D). 1 H NMR (400 MHz, CDCl3, 20 °C, TMS) δ 6.24-6.36 (b, 2H, HNCH2HNCH), 3.40-3.90 (m, 4H, HNCH2CH(NH)CH2OH), 3.20-3.30 (m, 1H, HNCH2CH), 2.15-2.19 (m, 4H, COCH2), 1.58-1.59 (m, 4H, COCH2CH2), 1.23 (b, 48H, (CH2)12CH3), 0.84-0.87 (t, 6H, CH3). 2,3-Dipalmitoylamidopropane-1-(p-nitrophenyl) Carbonate (15). This compound was prepared similarly to 14 and was recovered as a white powder (2.3 g, 3.1 mmol, 72%). Rf ) 0.71 (D). 1H NMR (400 MHz, CDCl3, 20 °C, TMS) δ 8.27-8.25, 7.38-7.36 (two d, each J ) 9.1 Hz, 4H, C6H4), 6.69 (m, 1H, -HNCH-), 6.13 (m, 1H, HNCH2), 4.23-4.36 (m, 3H, NHCHCH2OCOO), 3.383.59 (m, 2H, HNCH2CH), 2.15-2.20 (m, 4H, COCH2), 1.57-1.58 (m, 4H, COCH2CH2), 1.24 (b, 48H, (CH2)12CH3), 0.82-0.86 (t, 6H, CH3), 1,2-Dipalmitoylamidopropane-3-[bis-(2-dimethylaminoethane)] Carbamate (20, 1,2lb3). 20 was prepared similarly to 19 as a white sticky powder (1.9 g, 2.5 mmol, 58%). Rf ) 0.52 (E). Anal. Calcd for C44H89N5O4 (MW 751): C, 70.26; H, 11.92; N, 9.31. Found: C, 70.14;


H, 11.97; N, 9.32. MS (ESI) m/z 774.7 [M + Na]+; 1H NMR (400 MHz, CDCl3, 20 °C, TMS) δ 7.15 (d, 1H, OCHNCH) 6.70 (t, 1H, OCHNCH2CH), 4.13 (m, 3H, NHCHCH2OCO), 3.20-3.50 (m, 6H, CH2CHCH2OCON(CH2)2), 2.40-2.50 (m, 4H, CH2N(CH3)2), 2.22-2.25 (d, 12H, N(CH3)2), 2.10-2.20 (m, 4H, COCH2), 1.45-1.60 (m, 4H, COCH2CH2), 1.21 (b, 48H, (CH2)12CH3), 0.82-0.86 (t, 6H, (CH2)12CH3); 13C NMR (100 MHz, CDCl3, 20 °C) δ 176.04 (NHCO), 175.34 (NHCO), 157.44 (OCO), 65.67, 58.61, 46.83, 46.58, 37.72, 32.96, 30.74, 30.41, 26.74, 23.73, 15.20. Methyl 2,3-Distearoylamidopropionate (6). Compound 6 was prepared analogously to 3 on a 6.6 mmol scale to give upon crystallization from methanol the product 6 as a brownish powder (2.4 g, 3.7 mmol, 56%). Rf ) 0.92 (A), Rf ) 0.68 (C). 1H NMR (400 MHz, CDCl3, 20 °C, TMS) δ 6.78 (d, 1H, HNCH), 6.08 (s, 1H, HNCH2), 4.58 (m, 1H, CHCOOCH3), 3.72 (s, 3H, COOCH3), 3.60 (m, 2H, HNCH2), 2.13-2.19 (m, 4H, COCH2), 1.56-1.68 (m, 4H, COCH2CH2), 1.20 (b, 56H, (CH2)14CH3), 0.820.85 (t, 6H, (CH2)14CH3). 2,3-Distearoylamidopropanol (11). Compound 11 was prepared by reducing 6 with lithium borohydride in anhydrous THF at 50-60 °C for 25 h. The crude was worked up similarly to 8 to give the product 11 as white crystals (1.0 g, 1.6 mmol, 47%). Rf ) 0.5 (D). 1H NMR (400 MHz, CDCl3, 20 °C, TMS) δ 6.40 (m, 2H, HNCH2(HN)CH), 3.20-3.90 (m, 5H, HNCH2CH(NH)CH2OH), 2.15-2.19 (m, 4H, COCH2), 1.59 (m, 4H, COCH2CH2), 1.23 (b, 56H, (CH2)14CH3), 0.85 (t, 6H, CH3). 2,3-Distearoylamidopropane-1-(p-nitrophenyl) Carbonate (16). This compound was prepared similarly to 13 with p-nitrophenylchloroformate and gave the product as white solid (1.2 g, 1.6 mmol, 96%). Rf ) 0.63 (D). 1HNMR (400 MHz, CDCl3, 20°C, TMS) δ 8.25-8.23, 7.37-7.35 (two d, each J ) 9.1 Hz, 4H, C6H4), 6.64 (s, 1H, HNCH), 6.02 (s, 1H, HNCH2CH), 4.08-4.40 (m, 3H, NHCHCH2OCOO), 3.42-3.75 (m, 2H, HNCH2CH), 2.20 (m, 4H, COCH2), 1.58-1.59 (m, 4H, COCH2CH2), 1.23 (b, 56H, (CH2)14CH3), 0.84-0.87 (t, 6H, CH3). 1,2-Distearoylamidopropane-3-[bis-(2-dimethylaminoethane)] Carbamate (21, 1,2lb4). 21 was prepared similarly to 20 with bis-(2-dimethylaminoethyl) amine and gave the product as a half-white sticky solid (0.6 g, 0.7 mmol, 48%). Rf ) 0.48 (E). Anal. Calcd for C48H97N5O4 (MW 807): C, 71.32; H, 12.09; N, 8.66. Found: C, 71.38; H, 11.92; N, 8.18. MS (ESI) m/z 830.7 [M + Na]+; 1H NMR (400 MHz, CDCl3, 20°C, TMS) δ 7.00 (d, 1H, HNCH), 6.67 (t, 1H, HNCH2CH), 4.13 (m, 3H, NHCHCH2OCON), 3.33 (m, 6H, CH2CHCH2OCON(CH2)2-), 2.10-2.40 (m, 20H, CH2N(CH3)2, COCH2CH2(CH2)14CH3), 1.55-1.57 (m, 4H, COCH2CH2), 1.22 (s, 56H, (CH2)14CH3), 0.84 (t, 6H, CH3). Methyl 2,3-Dioleoylamidopropionate (7). Compound 7 was prepared as a brown oil (impure mixture) analogously to compound 6 on a 6.8 mmol scale. Rf ) 0.94 (A). 1H NMR (400 MHz, CDCl3, 20 °C, TMS) δ 6.80 (b, 1H, HNCH), 6.18 (b, 1H, HNCH2), 5.3 (m, 4H, CHdCH), 3.5-4.7 (m, 6H, HNCH2HNCHCOOCH3), 1.4-2.4 (b, 16H, COCH2CH2(CH2)4CH2CHdCHCH2(CH2)6CH3), 1.24 (b, 40H, CH2(CH2)4CH2CHdCHCH2(CH2)6CH3), 0.830.86 (t, 6H, (CH2)6CH3). 2,3-Dioleoylamidopropan-3-ol (12). Compound 12 was prepared analogously to 11 on a 1.8 mmol scale to give upon purification (column chromatography, CHCl3/ MeOH) the product as a yellow solid (0.49 g, 0.8 mmol, 44%). Rf ) 0.5 (D). 1H NMR (400 MHz, CDCl3, 20 °C, TMS) δ 7.33 (s, 1H, HNCH), 6.80 (s, 1H, HNCH2CH), 5.27 (m, 4H, CHdCH), 3.20-3.80 (m, 5H, HNCH2(HN)-

Savva et al.

CHCH2OH), 2.14 (m, 4H, COCH2), 1.95 (m, 8H, CH2CHd CHCH2), 1.56 (m, 4H, COCH2CH2) 1.23 (b, 40H, COCH2(CH2)4CH2CHdCHCH2(CH2)6CH3), 0.8-0.83 (t, 6H, (CH2)6CH3). 2,3-Dioleoylamidopropane-1-(p-nitrophenyl) Carbonate (17). Compound 17 was prepared analogously to 16 on a 5.8 mmol scale to give upon purification (column chromatography, CHCl3/MeOH) the product as a yellow-brown oil (2.9 g, 3.7 mmol, 63%). Rf ) 0.73 (D). 1 H NMR (400 MHz, CDCl3, 20°C, TMS) δ 8.25-8.23, 7.37-7.35 (two d, each J ) 9.1 Hz, 4H, C6H4), 7.0 (b, 1H, HNCH), 6.82 (b, 1H, HNCH2CH), 5.31 (m, 4H, CHdCH), 4.18-4.40 (m, 3H, NHCHCH2OCOO), 3.22-3.57 (m, 2H, HNCH2CH), 2.17 (m, 4H, COCH2), 1.97 (m, 8H, CH2CHd CHCH2), 1.55 (m, 4H, COCH2CH2), 1.23 (b, 40H, COCH2(CH2)4CH2CHdCHCH2(CH2)6CH3), 0.82-0.86 (t, 6H, (CH2)6CH3). 2,3-Dioleoylamidopropane-1-[bis-(2-dimethylaminoethane)] Carbamate (22, 1,2lb5). 22 was prepared similarly to 21 with bis-(2-dimethylaminoethyl)amine and gave the product as a yellow oil (0.9 g, 1.1 mmol, 35%). Rf ) 0.5 (E). Anal. Calcd for C48H93N5O4 (MW 803): C, 71.68; H, 11.65; N, 8.71. Found: C, 71.03; H, 11.85; N, 7.73. MS (ESI) m/z 826.7 [M + Na]+; 1H NMR (400 MHz, CDCl3, 20°C, TMS) δ 7.3 (b, 1H, HNCH), 6.89 (b, 1H, HNCH2), 5.28 (m, 4H, CHdCH), 4.16 (m, 3H, HNCHCH2OCO), 3.39 (m, 6H, HNCH2(HN)CHCH2OCON(CH2-)2), 1.97-2.42 (m, 28H, COCH2; CH2CHdCHCH2, CH2N(CH3)2), 1.21-1.55 (m, 44H, COCH2CH2(CH2)4CH2CHdCHCH2(CH2)6), 0.80-0.83 (t, 6H, CH2CH3); 13C NMR (100 MHz, CDCl3, 20 °C) δ 175.40 (NHCO), 174.10 (NHCO), 157.38 (OCO), 130.86-129.50 (CdC), 58.24, 46.852, 37.69, 32.96, 30.80, 30.38, 28.25, 26.80, 23.76, 15.17. Lipoplex Formation. Solutions of cationic lipids in chloroform were prepared on a weight to volume basis. Aliquots of cationic lipids, DOPE, and cholesterol were transferred with the aid of graduated glass syringes to 12 mm × 75 mm borosilicate glass tubes (pyrex, Corning Inc., NY) in appropriate amounts to provide the relative lipid concentrations desired upon reconstitution with 1 mL of 40 mM Tris buffer, pH 7.1. Bulk organic solvent was removed with a stream of nitrogen gas, and residual solvent was removed under high vacuum (0.40 mmHg) for 4-5 h. The dried lipids were hydrated with 40 mM Tris buffer, pH 7.1, for 30-45 min at 50-60 °C with occasional vortexing. Lipid dispersions were further diluted in serum-free RPMI media (0.3 µmol/mL) and mixed with pUC19-CMV-β-gal plasmid DNA (0.2 µg/mL), as described elsewhere (8). Lipoplexes were added to the cells within 1 h after mixing. Transfection Assays. Approximately 50 000 B16F0 cells (CRL-6322, ATCC, Rockville, MD) in 0.5 mL of 10% FBS/90% RPMI, supplemented with 100 units/mL penicillin, 100 µg/mL streptomycin, 4 mM L-glutamine, and 1 mM sodium pyruvate were seeded into each well of 48well plates and cultured overnight in a 5% CO2 incubator at 37 °C. After 12-14 h, the medium was removed, and 250 µL of the lipoplexes in serum-free RPMI was added. Lipoplexes were aspirated after 3 h, and 0.5 mL of 10% FBS/90% RPMI was added to each well. After a total of 48 h, the medium was removed and 150 µL of lysis buffer (0.1% Triton X-100 in 0.1 M Tris and 2 mM EDTA, pH 8) was added to lyse the cells. The cell lysate was centrifuged, and an amount of 120 µL was collected and assayed for β-galactosidase activity using a microplate colorimetric assay that employed the substrate o-nitrophenyl-β-D-galactopyranoside (12).


Cytotoxicity Assay. B16F0 cells were subjected to lipoplex formulations and further incubated for 44 h, as described above in the in vitro transfection studies. MTT was dissolved in PBS at 5 mg/mL and filtered through a 0.22 µm membrane to remove insoluble material. The 50 µL of MTT solution was added into each well, and the cells were incubated for 4 h at 37 °C in 5% CO2. The medium was then flicked out, the cells were washed once with cold PBS, and 250 µL of DMSO was added to each well. The cells were subsequently shaken for 30 min at room temperature using an orbital shaker to ensure complete dissolution of the formazan crystals. The emerged deep-purple color was measured at 630 nm with a multiwell scanning spectrophotometer (Elx 800 UV, BioTek Instruments, Inc.). Agarose Gel Electrophoresis. Agarose powder (400 mg) was added to 50 mL of TAE buffer (0.8% dispersion) in an Erlenmeyer flask. The mixture was boiled for a minute until it became clear. To the solution, 2.5 µL (10 mg/mL) of ethidium bromide (dye) was added. The solution was allowed to cool to 50-60 °C, then poured into the electrophoresis apparatus and allowed to solidify. The lipoplexes were prepared as 0.5, 1.0, 2.0, 4.0, 6.0, and 8.0 +/- charge ratios. Calculations were based on a nucleotide average molecular weight of 330 g/mol. The ratio for cationic lipids mixed with neutral lipids was kept constant at cationic lipid/ DOPE/cholesterol 3:2:1 mol/ mol. A fixed quantity of DNA of 0.2 µg was added to each microfuge. Aliquots of cationic liposomes were then added to provide the desired +/- charge ratio. The total volume was brought to 8 µL by addition of appropriate buffer. The microfuges were briefly centrifuged (10-20 s), and they were allowed to stand at room temperature for at least 15 min. The complexes were then loaded into the wells of the gel sequentially with proper mixing. Naked DNA was loaded on the outermost lanes as a control experiment. After loading the complexes in the wells, the gel tank was filled with TAE buffer, pH 7.4, to a few millimeters above the gel surface and the lipoplexes were subjected to an electrophoretic field of 5 V/cm for 25-30 min. Migration of naked DNA and lipoplexes was noted with the aid of a BioRad minitransluminator. Particle Size Determination. Size determination was carried out using a 90° fixed scattering angle nanoparticle analyzer (Nicomp 370, PCS Inc., CA). All measurements were performed at room temperature with a 5 min run time. Particle size measurements were done using buffers of pH 7.2 and pH 8.0, 40 mM Tris buffer. All buffers were filtered using 0.22 µm filters. Lipid dispersions were prepared as described under the Lipoplex Formation section and subsequently diluted with a dust-free buffer to a final concentration of 1.5 mM in borosilicate tubes, 12 mm × 75 mm (KIMAX 51, VWR). Lipoplexes were prepared at a cationic lipid-to-DNA molar ratio of 1:1 (+/- ratio of 2) from 0.3 mM lipid dispersions. Lipoplexes were always freshly prepared by simply adding DNA to the lipid dispersions with gentle agitation. Preparation and manipulation of the liposomes were performed in a dust-free environment. “Equilibration time” was set to 2 min and “run time” to 10 min. Mean particle size and standard deviation reported are the cumulative average and cumulative standard deviation of five consecutive scans as calculated from intensityweighted Gaussian analysis (C370, version 12.4). Despite the high polydispersity index of some samples, individual scans were highly reproducible. Ethidium Bromide Displacement Assay. The studies were performed on a Cary Eclipse spectrofluorometer


at an excitation wavelength of 515 nm (band-pass ) 5 nm, linear reciprocal dispersion D-1 ) 2 nm/mm, slit width ) 2.5 mm) and an emission wavelength of 595605 nm (band-pass ) 10 nm, D-1 ) 2 nm/mm, slit width ) 5 mm). The amount of DNA added was set to 22.5 µg, and ethidium bromide was added in aliquots in order to identify the concentration that gave the maximum signalto-noise ratio. In a typical titration experiment, ethidium bromide (0.8 µg, 1.01 µM) was added to the quartz cuvette that contained a small magnetic stirrer and diluted to 3 mL with appropriate buffer. Addition of 22.5 µg (67.5 nmol, by nucleotide equivalent) of plasmid DNA with continuous gentle stirring yielded a 27-29 times increase in fluorescence intensity. Cationic lipids (1.5 mM) were then added in aliquots (4.5 µL) under continuous stirring at ambient temperature, and fluorescence was monitored continuously until the reading became stable. To minimize the effects of light source fluctuation and to provide a more accurate measurement, the light intensity emitted by the sample was always ratioed with the reference signal from the excitation light source. Processing of the fluorescence values as described before (13, 14) provided the relative ethidium bromide displacement from the plasmid DNA. The binding profiles of cationic reagents to plasmid DNA were modeled as parabolas on a domain and range [0, +∞) and [0, -∞), respectively, with the aid of PSI-Plot (version 7, Polysoftware International, NY). Vertical shifting of the curves yielded parabolas of the form y ) axb, the slopes of which can easily be compared. Instantaneous rates (dF/dr, where r stands for charge ratio) were calculated either by taking the first derivative of the best fit parabolic equation or from the experimental data for 0.2 charge ratio intervals, using dF/dc ) (Fn Fn-1)/0.2. The two methods correlated well. The cumulative average rates were calculated from the experimental data for 0.2 charge ratio intervals using ∆F/∆c ) (Fin Ffinal)/∆c. Monolayers Studies. The studies were conducted on a KSV Minitrough, Langmuir-type film balance equipped with symmetric barriers (KSV Instruments Ltd., Helsinki, Finland) and thermostated with the aid of a circulating water bath (VWR brand, model 1162, PolyScience, IL). Stock solutions of pure cationic lipids were prepared by dissolving lipids in chloroform at concentrations less than 0.5 mg/mL. The buffer (40 mM Tris, pH 7.2) was prepared with HPLC grade water, the surface tension of which was 72.0-72.8 dynes/cm. All experiments were performed in an open-air vibration-free environment. Prior to each experiment, the platinum plate was rinsed with ethanol and burned with a strong flame for 1 min and the barrier and trough were washed with soap and rinsed with distilled water. The proper volume of buffer was poured into the trough, and the platinum plate was adjusted at one-third of its height immersed in the buffer. The balance and barrier position were zeroed before any addition of the lipid. The 10-25 µL of the lipid dissolved in chloroform were applied through a microsyringe drop by drop on the surface of the buffer. After the chloroform was evaporated (15-20 min), the barriers were closed at constant speed of 9.99 mm/min. The data were collected using the proprietary KSV software (WinLB, version 1.61) and were plotted as surface pressure (mN/m) against mean molecular area (Å2/molecule). The apparatus was validated using DPPC as a standard (15). The data were subsequently imported into an Excel spreadsheet, and the compressibility modulus was cal-


Savva et al.

culated from the first derivative of the monolayer forcearea isotherms using

K ) -A(∂Π/∂A)T where K is the isothermal compressibility modulus of the monolayer and where A is the area per molecule at the corresponding surface pressure Π. Values of the compressibility moduli at monolayer collapse pressure were calculated from the plots of the compressibility modulus as a function of the mean molecular area. Fluorescence Polarization Studies. Fluorescence anisotropies of lipid dispersions were measured using a Cary Eclipse spectrofluorometer (Varian Inc., Victoria, Australia) equipped with motorized polarizers. The slit widths of both excitation and emission were 5 nm. Samples of lipid doped DPH dispersions (0.5 mM lipid, lipid/DPH molar ratio of 100:1) in a standard 1 cm × 1 cm square fluorescence cuvette were placed in a thermoelectric temperature-controlled four-window cuvette holder with a magnetic stirrer (temperature control precision of (0.02 °C) and illuminated with a beam at 351 nm. The fluorescence intensities (λmax ) 430 nm) of the emitted light polarized parallel (Ih) and perpendicular (Iv) to the excited light were recorded at different temperatures. Anisotropy values at each temperature were calculated by the Eclipse software according to the equation

r)

Iv - GIh Iv + 2GIh

where G is the instrumental grating factor that corrects for the anisotropy introduced by the optical components of the detection system. A 5 min equilibration time was left between each temperature. The anisotropy-temperature curves were fitted to

r)A-

B 1 + e-C(T-D)

where A, B, C, and D are constant parameters to be determined. The parameter D represents the gel to liquid crystalline phase transition temperature, Tm. Nonlinear fitting of the experimental data was done in an iterative fashion using the Prostat (Poly Software Intenational, version 3, New York). Initial parameters for A and B were rmax and rmax - rmin, respectively, where rmax is the maximum anisotropy or average of the upper third and rmin represents the minimum anisotropy value or the average of the lower third of the anisotropy values of data points. The initial value of D (or Tm) is the average of the temperature values of the crossover area. Best fit parameters were assessed within a 95% confidence interval. RESULTS

Biological Results. To establish the effect of varying the alkyl chain length on transfection efficiency, the ability of cationic lipids to mediate DNA transfection in the absence and presence of helper lipids DOPE and cholesterol was evaluated in B16F0 cells. Figure 2 illustrates the transfection properties of lipoplexes at cationic lipid-to-DNA charge ratios ranging from 1 to 4. As shown in Figure 2A, only the dioleoyl derivative, 1,2lb5, mediated efficient transfection that increased with increasing charge ratio. At charge ratio 8:1 the transfection activity of cationic lipids decreased dramatically

Figure 2. Transfection activity of cationic lipids in B16F0 cells: (A) 1,2lb lipoplexes; (B) 1,2lb/d lipoplexes; (C) 1,2lb/d/ch lipoplexes. The o-nitrophenol formation was converted to units, using a standard curve obtained with commercial β-galactosidase. The data presented are the average of values obtained from seven wells (n ) 7) of two independent experiments performed at different times.

because of toxicity (not shown). Also, it is noted that the transfection potency of complexes containing 1,2lb5 at 2:1 +/- charge ratio is comparable to the potency of complexes containing the commercially available cationic lipid DC-chol, and it is much greater at higher charge ratios. Surprisingly, the activity of the 1,2lb5 is greatly diminished in the presence of helper lipids (Figure 2B,C). Contrary to that, the dimyristoyl derivative 1,2lb2 induced high levels of β-galactosidase when used with DOPE and DOPE/cholesterol (Figure 2). Lipoplexes made by the dilauroyl derivative 1,2lb1 and the distearoyl analogue 1,2lb4 were devoid of activity at all charge ratios and compositions. The dipalmitoyl derivative 1,2lb3 mediated low levels of gene expression at all +/ratios when it was combined with DOPE and DOPE and cholesterol (Figure 2B,C). The cytotoxicity of the cationic lipids alone or in combination with DOPE and DOPE and cholesterol was tested by the MTT reduction method. All lipoplex formulations were well tolerated with cell survival greater than 75% at all different charge ratios (not shown). Physicochemical Properties. Agarose Gel Electrophoresis. Since DNA condensation is considered a requirement for efficient in vitro internalization of lipoplexes by the cells, techniques that provide information about lipid-DNA interaction are useful tools to provide insight about the mechanism of transfection. One such technique is agarose gel electrophoresis, in which lipoplexes are subjected to an electric field and the electrostatic and hydrophobic interactions of cationic



Figure 3. Gel retardation assay (from left to right): (A) 1,2lb1, 1,2lb1/d, 1,2lb1/d/ch; (B) 1,2lb2, 1,2lb2/d, 1,2lb2/d/ch; (C) 1,2lb3, 1,2lb3/d, 1,2lb3/d/ch; (D) 1,2lb4, 1,2lb4/d, 1,2lb4/d/ch; (E) 1,2lb5, 1,2lb5/d, 1,2lb5/d/ch. All gels were repeated at least once. Details of the assay are as described in Materials and Methods.

lipids with DNA can be evaluated from the migration pattern of the plasmid DNA. Electrophoretic migration of molecules is dependent on both molecular charge and size. Thus, retardation of plasmid DNA on agarose gel can occur because of its encapsulation by cationic lipids or because of charge neutralization and compaction to small particles. Plasmid DNA compaction by cationic lipids is ordinarily accompanied by ethidium bromide exclusion yielding DNA bands of reduced intensity. This is the basis for differentiating compacted DNA from encapsulated DNA that has precipitated in the slots of the gel. Figure 3 illustrates the effect of bivalent cationic lipids on plasmid DNA electrophoretic migration. In the absence of helper lipids only the dioleoyl derivative 1,2lb5 promoted complete DNA retardation at charge ratios relevant to those used in the transfection experiments (Figure 3E, left lane). The results suggest that lack of activity of lipoplexes composed of saturated cationic lipids is due to their extremely weak interaction with the plasmid DNA. Moreover, the complete retardation of plasmid DNA promoted by 1,2lb5 at charge ratios of >2 suggests that the higher levels of gene expression observed at the highest charge ratio (Figure 2A) is due to a larger population of DNA associated with the lipid or due to a greater degree of plasmid condensation induced by the cationic lipid. The presence of DOPE in the formulations greatly facilitated association of plasmid DNA with the cationic lipids. The sole exception was the distearoyl derivative 1,2lb4, where DOPE had no effect (Figure 3D, middle lane). The binding profiles of 1,2lb1/d and 1,2lb2/d were similar with a full retention of the plasmid in the slots of the gel occurring at +/- ratio of g4. The results are in agreement with the high transfection activity of the 1,2lb2/d and 1,2lb2/d/ch, but they contradict the lack of activity of the 1,2lb1 formulated with DOPE and DOPE and cholesterol. The binding efficiency of the dipalmitoyl derivative 1,2lb3 also increases when formulated with helper lipids (Figure 3C), and the transfection activity of these lipoplexes increases accordingly (Figure 2C). In contrast, although plasmid DNA migration is inhibited by the 1,2lb4/d/ch (Figure 3D, right lane), its transfection activity is practically zero. A similar contradiction between the gel electrophoresis and transfection activity

occurred with the 1,2lb5/d and 1,2lb5/d/ch. Although the binding efficiency of the cationic lipids to the DNA is increased compared to the 1,2lb5 alone, their transfection activity is reduced (Figure 2B,C). Ethidium Bromide Displacement Assay. Another approach used to characterize lipid-DNA complexes is the EtBr displacement assay. The distinct fluorescence enhancement of intercalated EtBr allows accurate spectrofluorometric quantification of its exclusion from the double helix of plasmid DNA by cationic lipids. This method is used to determine the exact charge ratio at which cationic lipids fully displace EtBr from plasmid DNA. Figure 4A shows that none of the cationic lipids totally displaces EtBr at charge ratios ranging from 0.5 to 2 at physiological pH and ambient temperature. In fact, the only lipid that was appreciably bound to DNA is the 1,2lb5. At a charge ratio of 2, this lipid had accomplished 66% EtBr exclusion. Fitting the experimental data into a parabolic equation as described in the methods section suggested that complete exclusion of EtBr by the 1,2lb5 analogue takes place at a +/- charge ratio equal to 2.7. This is in agreement with the agarose gel electrophoresis results that indicated that complete retention of plasmid DNA in the slots occurred at charge ratios greater than 2. Binding of cationic lipids bearing a tertiary amine as the polar headgroup was also dependent on the pH of the media. Figure 4B signifies the cooperative nature of DNA binding of the 1,2lb5 at both pH values; i.e., the binding rate increases with charge ratio exponentially. As expected, DNA binding of the lipid at pH 7.2 was more effective than at pH 8.0. Paradoxically, binding cooperativity at pH 8 was higher as signified by the higher curvature of the neutralization curve (Figure 4B) and the higher slope of the average binding rate in the +/- range of 0.2-1.2 (Figure 4C). Specifically, at a +/- charge ratio of 0.2, the lipid quenched 4.2% of the total fluorescence at pH 7.2 as opposed to only 0.87% fluorescence quenching at pH 8.0. The average rate difference at the two pH values was 4.9. At a +/- ratio equal to 2, the lipid quenched 66.4% of the total fluorescence at pH 7.2 as opposed to 43.0% fluorescence quenching at pH 8.0, thus reducing the rate difference between the two pH values to 1.5. To explain the higher curvature of the plasmid DNA binding curve at pH 8, one has to account for the


Savva et al. Table 1. Summary of Particle Size Distribution of Lipid Dispersions and Lipoplexes at a +/- Charge Ratio Equal to 2 pH

mean diameter ( SD (nm)

1,2lb1 1,2lb2 1,2lb3

7.2 7.2 7.2

Particle Size Distribution of Cationic Lipid Dispersionsa 5698 ( 5800 1.018 990 ( 704 0.712 488 ( 352 0.723

1,2lb4 1,2lb5 1,2lb5

7.2 7.2 8.0

326 ( 215 429 ( 330 573 ( 295

7.2 7.2 7.2 7.2 7.2 8.0

Particle Size Distribution of Lipid-DNA Complexesb 3363 ( 3869 1.151 1620 ( 1211 0.748 3561 ( 3349 0.941 980 ( 881 0.899 368 ( 160 0.436 15096 ( 22201 1.471

1,2lb1 1,2lb2 1,2lb3 1,2lb4 1,2lb5 1,2lb5

coefficient of variation

0.663 0.771 0.516

macroscopic appearance

turbid turbid slightly turbid/ aggregates slightly turbid clear solution clear solution

clear solution aggregates clear solution clear solution clear solution clear solution

a All of the lipid concentrations are 1.5 mM. b All of the lipid concentrations are 0.1 mM.

Figure 4. (A, B) Titration of plasmid DNA with cationic lipids in the presence of ethidium bromide. Fluorescence is reported relative to the fluorescence in the absence of cationic lipids, after subtracting the baseline of ethidium bromide alone, for each of the buffer conditions examined. (C) The noncumulative average rates dF/dr were calculated from the data shown in (B), as described in Methods.

forces responsible for EtBr exclusion. The main two forces that play a major role in plasmid DNA condensation by cationic lipids are the electrostatic and hydrophobic ones. Electrostatic forces are long range and are responsible for the initial approach of the cationic lipids to the DNA. Neutralization of the negative charges of the double helix causes its collapse and its compaction to a random coil configuration identified in our case with EtBr displacement. The collapse of the double helix is accompanied by a reduction of the water of hydration; a condition that encourages the cationic lipids to surround the plasmid DNA in greater numbers and pack tighter around it, thus facilitating further dehydration and compaction of the DNA. It appears that the electrostatic interactions are weaker at pH 8 than at pH 7.2, whereas dehydrationmediated DNA condensation due to hydrophobic attractive interactions plays a bigger role at pH 8. Dynamic Light Scattering. Hydration of cationic lipids in Tris buffer, pH 7.2, resulted in liposome formation of quite heterogeneous size, as indicated by the large coefficient of variation (Table 1). The size of liposomes made by 1,2lb1 and 1,2lb2 was above 1 µm, whereas dispersions made of the 1,2lb3 and 1,2lb4 had an average particle size of 0.5 µm. Upon addition of DNA, the diameter of all the complexes composed of saturated cationic lipids increased, with the only exception being the complex made of the dioleoyl derivative 1,2lb5. The results of the particle size distribution suggest that, in agreement with the gel electrophoresis and the EtBr

displacement assay, saturated cationic lipids failed to electrostatically interact and effectively condense the plasmid DNA. In addition, the results verify the dependence of DNA binding on the pH. The very big particle size of the 1,2lb5-DNA complex measured at pH 8, as opposed to the compact particle of the complex at pH 7.2, validates the results obtained from the EtBr displacement assay and suggests that at pH 8, the lipid is loosely associated with plasmid DNA mainly through hydrophobic forces. Monolayer Studies. Studies performed on spread monolayer at the air-water interface enabled investigation of surface properties of the novel cationic lipids. Cationic lipids 12lb1 and 1,2lb5 exist at liquid-expanded phases at both 23 and 37 °C, exhibiting almost identical monolayer collapse areas and pressures at corresponding temperatures (Figure 5 and Table 2). Contrary to the 1,2lb1 and 1,2lb5, 1,2lb2 monolayers were at a liquid condensed state at high compressions at 23 °C while monolayers of this lipid displayed an all-liquid expanded behavior at 37 °C, as indicated by the absence of a phase transition in the force-area isotherm of the 1,2lb monolayers and its larger mean molecular area at 37 °C (Table 2). Monolayers composed of the 1,2lb3 derivative were at liquid expanded states at low surface pressures and at liquid condensed phases at high surface pressures, at both temperatures. The distearoyl derivative 1,2lb4 displayed only liquid-condensed behavior at 23 °C as concluded by the relatively small molecular area at monolayer collapse, the high value of the compressibility modulus K, and the absence of any discontinuities in the Π-A isotherm. Isotherms of this cationic lipid at 37 °C displayed two discontinuities at two different surface pressures and mean molecular areas. The first discontinuity was centered at the mean molecular area and surface pressure of approximately 93 Å2 and 10 mN/m, respectively, while the second one occurred at smaller molecular dimensions (63 Å2) and higher compression forces (18 mN/m). Although the exact nature of these two transitions is not known, it is possible that the lower pressure transition is related to the merging of several lipid islets (2D micelles on the interface) into a more uniform surface, whereas the discontinuity that occurred at higher pressure represents an L1-to-L2 phase transition. Nevertheless, the small average cross-sectional molecular areas (Table 2) of this lipid at monolayer



Figure 5. Representative compression isotherms of 1,2lb cationic lipids spread on 40 mM Tris buffer, pH 7.2, at 23 °C (left) and 37 °C (right). Line represents the Π-A isotherm, while dots represent calculated K-A isotherm.

collapse argue in favor of a compact assembly of the 1,2lb4 at both temperatures. Fluorescence Anisotropy Studies. As shown in Figure 6, the nature and length of the acyl chain affected the dependence of fluorescence anisotropy of cationic lipid/DPH dispersions on the temperature profoundly. All lipids exhibited one or more inflections in their profile with the exception of dioleoyl derivative 1,2lb5, suggesting that this monounsaturated derivative existed at the liquid crystalline state within the temperature range studied in agreement with the monolayer studies (Figures 5 and 6). For the DPPC, a phase transition temperature of 41.3 °C was obtained that agrees well with

literature values (16). Not surprisingly, the phase transition of DPPC was strikingly sharp compared to the much broader phase transitions of the other saturated cationic lipids. Specifically, the temperature span of the DPPC phase transition was only 1 °C, compared to approximately 18, 38, 44, and 36 °C for 1,2lb1, 1,2lb2, 1,2lb3, and 1,2lb4, respectively. The larger temperature span of the transition signifies the uncooperative nature of the phase transition of cationic lipids compared to the highly cooperative gel-to-liquid crystalline phase transition of DPPC. As shown in Table 3, the midpoint of the phase transition temperature for 1,2lb1 was found to be equal


Savva et al.

Table 2. Monolayer Parameters of 1,2lb Seriesa mean molecular area (Å2)

Πc (mN/m)

K (mN/m)

23 °C

37 °C

23 °C

37 °C

69.71 (0.93) 63.59 (1.96)

39.66 (1.71) 59.05 (2.7) 28.25b (0.37) 60.77 (2.77) 12.97b (0.67) 41.35 (2.09)

38.13 (0.28) 40.64 (0.71)

88.44 (4.32) 121.3 (36.9)

1,2lb4 (n ) 4)

63.08 (2.96) 39.5 (1.64) 68.6b (4.08) 43.48 (0.80) 91.9b (1.00) 47.54 (0.90)

1,2lb5 (n ) 3)

64.16 (0.94)

57.14 (1.41) 23.45b (0.10) 46.04 (1.98) 17.13b (0.74) 9.59b (0.27) 37.42 (0.99)

1,2lb1 (n ) 4) 1,2lb2 (n ) 4) 1,2lb3 (n ) 3)

40.62 (2.32) 79.4b (1.9) 42.84 (1.06) 63.11b (2.43) 92.82b (1.15) 69.00 (2.06)

38.72 (0.74)

23 °C

phase statec

dΠ/dA 37 °C

23 °C

37 °C

23 °C

37 °C

86.30 (5.04) HV

1.34 (0.12) 3.7 (1.0)

L1 M

L1 L1

239.49 (41.31)

92.25 (18.45)

3.62 (0.34)

1.22 (0.06) HV 28.25b (0.37) 2.85 (0.50)

M

M

211.00 (8.04)

154.13 (15.39)

4.43 (0.27)

3.59 (0.27)

L2

M

84.62 (7.33)

77.82 (3.10)

1.31 (0.12)

1.08 (0.04)

L1

L1

a

Measured in 40 mM Tris buffer, pH 7.2. b Phase transition denoted by a maximum in the dΠ/dA plot as a function of mean molecular area. c M denotes mixture of liquid-expanded and liquid-condensed states. L1 denotes liquid-expanded state. L2 denotes liquid-condensed state. HV denotes high variability.

Figure 6. Fluorescence anisotropy (r) against temperature plots for cationic lipid dispersions at pH 7.2. Broken lines are the curve fits of the experimental data (symbols). Table 3. Phase Transition Parameters of 1,2lb Lipids As Determined by Fluorescence Anisotropy lipid

Tm (°C)

coefficient of determinationa

Tonset (°C)b

1,2lb1 1,2lb2 1,2lb3 1,2lb4 1,2lb5 DPPC

32.9 46.5 54.3 59.1

In Vitro Lipofection with Novel Asymmetric Series of 1,2

Recommend Documents