Evidence for Highly Cooperative Binding between Molecular Umbrella

Oct 15, 1997 - Double- and tetrawalled molecular umbrella-spermine conjugates (I and II) have been synthesized, and their binding to calf thymus DNA ...
0 downloads 0 Views 132KB Size
Download PDF
Bioconjugate Chem. 1997, 8, 891−895

891

Evidence for Highly Cooperative Binding between Molecular Umbrella-Spermine Conjugates and DNA Vaclav Janout, Marion Lanier, Gang Deng, and Steven L. Regen* Department of Chemistry and Zettlemoyer Center for Surface Studies, Lehigh University, Bethlehem, Pennsylvania 18015. Received July 29, 1997X

Double- and tetrawalled molecular umbrella-spermine conjugates (I and II) have been synthesized, and their binding to calf thymus DNA (CT-DNA), poly[d(AT)], and poly[d(GC)] compared with that of a single-walled analogue (III). At moderate salt concentrations (8 mM NaCl), I and II show significantly greater affinity toward each DNA, relative to III; at high salt concentrations (150 mM NaCl), strong binding of I and II (but not III) was maintained toward poly[d(GC)]. Examination of the influence of I-III on the melting behavior of poly[d(AT)] has provided strong evidence that the binding of I and II reflects highly cooperative interactions among DNA-bound conjugates and that the DNA duplex serves as a nucleation site for umbrella aggregation. The implications of these findings for the rational design of novel drug conjugates that operate at the nuclear level, and also novel transfection agents, are briefly discussed.

We have recently introduced a new class of surfactants that can cover an attached agent and shield it from an incompatible environment (1, 2). The design principle upon which these “molecular umbrellas” have been based involves the covalent coupling of two or more rigid “amphiphilic walls” to a central scaffold. Our working hypothesis is that such surfactants will enhance the permeability of pendent polar molecules (e.g., peptides, antisense oligonucleotides, etc.) across lipid bilayers by masking their hydrophilicity. One important issue that we have recently sought to clarify is whether or not an umbrella can adversely affect the binding properties of the attached agent. To probe this question, we synthe-

sized two umbrella-spermine conjugates (I and II) and examined their DNA binding behavior. Spermine was specifically chosen as a “mock drug” on the basis of its affinity toward DNA and also its strong hydrophilicity (3). For purposes of comparison, the DNA binding properties of a single-walled analogue (III) were also investigated. In this paper, we document our discovery that the coupling of a molecular umbrella to spermine results in significantly enhanced DNA binding at physiological salt concentrations. We also present compelling evidence that this binding is highly cooperative in character. MATERIALS AND METHODS

General Methods. Unless stated otherwise, all reagents and chemicals were obtained from commercial sources and used without further purification. N-(OSuccinimidylcholeate) was prepared using procedures X Abstract published in Advance ACS Abstracts, October 15, 1997.

S1043-1802(97)00142-0 CCC: $14.00

similar to those previously described (2). EMS silica Gel 60 was used for column chromatrography; preparative thin-layer chromatography employed EM Science silica gel 60 F-254. Detection on TLC was made by a combination of sulfuric acid 10% in water, I2, and UV (254 and 365 nm). House-deionized water was purified using a Millipore Milli-Q filtering system containing one carbon and two ion-exchange stages. All 1H NMR spectra were recorded on a Bruker 360 MHz instrument; chemical shifts are reported in parts per million and were referenced to residual solvents. High-resolution mass spectra were obtained at the Mass Spectrometry Facility of the University of California, Riverside. Elemental analyses were determined by Midwest Microlab (Indianapolis, IN). Surface tension measurements (23 °C) were made using a tensiometer/micro-balance (NIMA, Model ST9000). Molecular Umbrella-Spermine Conjugate I. A solution composed of N1,N3-spermidinebis[cholic acid amide] (100 mg, 0.11 mmol), ethyl acrylate (0.2 mL, 1.85 mmol), CHCl3 (0.2 mL), and CH3OH (0.3 mL) was stirred at 40 °C for 16 h. Subsequent concentration under reduced pressure and column chromatography [silica, CHCl3/CH3OH/H2O, 103:25:3 (v/v/v)] afforded 85 mg (86%) of conjugated addition product having Rf 0.58 and 1 H NMR (CD3OD/CDCl3, 5:1, v/v) δ 0.67 (s, 6 H), 0.872.45 (m, 71 H), 2.45 (m, 4 H), 2.78 (m, 2 H), 3.15 (m, 4 H), 3.33 (m, 2 H), 3.78 (s, 2 H), 3.92 (s, 2 H), 4.11 (q, 2 H). This product was then hydrolyzed by first dissolving it in 2 mL of methanol, followed by addition of 80 mL of 1 M NaOH in methanol. After stirring for 24 h at room temperature, 80 mL of a 1 M HCl solution was added, followed by concentration under reduced pressure. Purification by column chromatography [silica, CHCl3/CH3OH/H2O, 103:25:3 (v/v/v)] afforded 79.0 mg of the corresponding acid having Rf 0.52 and 1H NMR (CD3OD) δ 0.63 (s, 6 H), 0.83-2.15 (m, 68 H), 3.06-3.28 (m, 8 H), 3.32 (m, 2 H), 3.75 (s, 2 H), 3.88 (s, 2 H). Treatment of this acid with O-(N-succinimidyl)-N,N,N′,N′-tetramethyluronium tetrafluoroborate (TSU) (26.4 mg, 0.086 mmol) and diisopropylethylamine (DIPEA) (11.2 mg, 0.086 mmol) in 0.5 mL of DMF for 4 h at room temperature produced the corresponding activated ester, which was then added, dropwise, to a solution of spermine (44.5 mg, 0.22 mmol) in 0.2 mL of anhydrous DMF. The resulting mixture was stirred for 18 h at room temperature and © 1997 American Chemical Society

892 Bioconjugate Chem., Vol. 8, No. 6, 1997

then concentrated under reduced pressure. Subsequent dissolution in 1 mL of CH3OH and precipitation by addition of 10 mL of a saturated aqueous NaHCO3 solution afforded a colorless solid, which was triturated with water (2 × 10 mL) and purified by column chromatography [silica, CH3OH/NH4OH, 3:1 (v/v)] to give 21.0 mg (16%) of I, having Rf 0.45, mp 151-153 °C, and 1H NMR (CD3OD) δ 0.70 (s, 6 H), 0.91-2.25 (m, 74 H), 2.45 (m, 2 H), 2.60-2.72 (m, 12 H), 3.20 (m, 2 H), 3.33 (m, 2 H), 3.78 (d, 2 H), 3.94 (s, 2 H). Anal. Calcd for C68H123N7O9: C, 69.05; H, 10.48; N, 8.29. Found: C, 68.86; H, 10.51; N, 8.22. HRMS for (C68H124N7O9)+ calcd: 1182.9461. Found: 1182.9413.

Molecular Umbrella-Spermine Conjugate II. To a stirred solution of iminodiacetic acid (0.49 g, 3.68 mmol) in a solution of 9 mL of saturated NaHCO3/dioxane (4:5, v/v) was added 9-fluorenylmethoxycarbonyl chloride (FmocCl) (0.95 g, 3.67 mmol) in a few portions at room temperature. After the mixture was stirred in a closed flask for 20 min at room temperature, the solvents were removed under reduced pressure at 40 °C. The resulting oil was then acidified by adding 20 mL of 1 M HCl to give a colorless solid that was dissolved in 20 mL of CHCl3 and sequentially washed with 1 M HCl (2 × 10 mL) and water (3 × 10 mL). The organic phase was concentrated under reduced pressure, and the colorless solid was recrystallized twice in 30 mL of CH3OH to give 0.75 g (58%) of the Fmoc-carbamate of iminodiacetic acid having 1H NMR (CD3OD/CDCl3, 5:1, v/v) δ 4.10 (d, 4 H), 4.20 (t, 1 H), 4.34 (d, 2 H), 7.27 (t, 2 H), 7.31 (t, 2 H), 7.58 (d, 2 H), 7.74 (d, 2 H). To a stirred solution of FmocN(CH2CO2H)2 (19.2 mg, 0.054 mmol) and DIPEA (15.6 mg, 0.12 mmol) in 0.3 mL of anhydrous DMF was added 36.0 mg (0.19 mmol) of TSU. The reaction mixture was stirred at room temperature for 75 min and then diluted with a solution containing 0.2 mL of DMF, 0.1 mL of DIPEA, and 98.5

Janout et al.

mg (0.106 mmol) of N1,N3-spermidinebis[cholic acid amide]. After 24 h of stirring at room temperature, the solvents were removed under reduced pressure, and the residue was dissolved in 1 mL of CH3OH. Subsequent precipitation by addition into 20 mL of a saturated aqueous solution of NaHCO3, trituration with water (2 × 10 mL), and purification by preparative thin-layer chromatography [EM Sciences silica, 1 mm, CHCl3/CH3OH/H2O, 65:25:3 (v/v/v)] afforded 0.072 g (69%) of the Fmoc-protected tetrawalled umbrella having Rf 0.75 and 1 H NMR (CD3OD) δ 0.60 (s, 12 H), 0.63-2.19 (m, 132 H), 2.95-3.25 (m, 16 H), 3.35 (m, 4 H), 3.73 (s, 4 H), 3.86 (s, 4 H), 4.00 (br s, 4 H), 4.14 (d, 2 H), 4.35 (t, 1 H), 7.227.71 (m, 8 H). Subsequent deprotection was carried out by dissolving 150 mg (0.069 mmol) of the Fmoc-protected tetrawalled umbrella plus 0.10 mL (1.01 mmol) of piperidine in 2 mL of CH3OH and stirring the solution for 5 h at 40 °C. Removal of solvent under reduced pressure, followed by column chromatography [silica, CHCl3/CH3OH/H2O, 103:27:3 (v/v/v/)] afforded 91 mg (67%) of the corresponding amine having Rf 0.38 and 1H NMR (CD3OD) δ 0.70 (s, 12 H), 0.91-2.30 (m, 132 H), 3.17 (m, 8 H), 3.35 (m, 12 H), 3.54 (br d, 4 H), 3.78 (br s, 4 H), 3.94 (br s, 4 H). Conjugate addition of 105 mg (0.045 mmol) of the tetrawalled amine to ethyl acrylate (1 mL, 10 mmol) was carried out in 0.5 mL of CH3OH plus 0.2 mL of CHCl3 by stirring at 50 °C for 3 days. Removal of solvent under reduced pressure and purification by column chromatography [silica, CHCl3/CH3OH/H2O, 103:25:3 (v/v/v)] afforded 81.8 mg (78%) of the conjugate addition product having Rf 0.56 and 1H NMR (CD3OD) δ 0.70 (s, 12 H), 0.91-2.25 (m, 137 H), 2.45 (m, 2 H), 2.95 (m, 2 H), 3.15 (m, 8 H), 3.38 (m, 10 H), 3.48 (br d, 4 H), 3.78 (br s, 4 H), 3.94 (br s, 4 H), 4.15 (t, 2 H). Hydrolysis of the tetrawalled ethyl ester (81.8 mg, 0.0395 mmol) in 2 mL of CH3OH was carried out by adding 0.05 mL of a solution made from 1 M NaOH/CH3OH (4:1, v/v) and stirring at room temperature for 1 h, followed by additional stirring for 3 h at 50 °C. After the mixture had cooled to room temperature, 0.05 mL of a 1 M HCl solution was then added and the solvents were removed under reduced pressure to afford 78.0 mg (97%) of the corresponding acid having an Rf 0.25 [CHCl3/CH3OH/H2O, 65:25:4 (v/v/v)] and 1H NMR (CD3OD) δ 0.70 (s, 12 H), 0.90-2.32 (m, 134 H), 2.52 (t, 2 H), 3.18 (m, 10 H), 3.36 (m, 10 H), 3.70 (m, 4 H), 3.79 (s, 4 H), 3.93 (s, 4 H). To a stirred solution of the tetrawalled carboxylic acid (100 mg, 0.05 mmol) and DIPEA (8.9 mg, 0.065 mmol) in 0.6 mL of anhydrous DMF was added TSU (19.7 mg (0.065 mmol). The resulting mixture was stirred at room temperature in a closed flask for 3 h. The solution was then added, dropwise, to a solution of spermine (80 mg, 0.4 mmol) and DIPEA (25 mL). After the reaction mixture had stirred for 24 h at room temperature, the solvents were removed under reduced pressure. Subsequent dissolution in CH3OH (1 mL), precipitation by addition of a saturated solution of NaHCO3 (10 mL), trituration with water (3 × 5 mL) and purification by column chromatography [silica, CH3OH, NH4OH, 4:1 (v/ v)] afforded 58.3 mg (53%) of II having Rf 0.39, mp 159161 °C, and 1H NMR (CD3OD) δ 0.70 (s, 12 H), 0.912.28 (m, 142 H), 2.30-2.90 (m, 12 H), 3.15 (m, 10 H), 3.35 (m, 12 H), 3.55 (br d, 4 H), 3.79 (br s, 4 H), 3.94 (br s, 4 H). Anal. Calcd for C127H221N11O19‚2 H2O: C, 68.03; H, 9.93; N, 6.87. Found: C, 67.97; H, 9.90; N, 6.84. MS for (C127H222N11O19)+ calcd: 2206. Found: 2207. N1-Spermine Cholic Acid Amide (III). Cholic acid (0.200 g, 0.5 mmol) was activated by reaction with TSU

Bioconjugate Chem., Vol. 8, No. 6, 1997 893

Molecular Umbrella−Spermine Conjugates Scheme 1

(0.170 g, 0.56 mmol) and 80 µL (0.5 mmol) of DIPEA in 2 mL of DMF. After the solution had stirred for 5 h at room temperature, a solution that was made from spermine (100 mg, 0.5 mmol) in 8 mL of CHCl3/THF (1:1, v/v) was added dropwise. The reaction mixture was stirred overnight at room temperature and concentrated under reduced pressure. Subsequent addition of chloroform (10 mL), sequential washing with 3 mL of NaHCO3 and 3 mL of brine, drying (Na2SO4), concentration under reduced pressure, and column chromatography [silica, CH3OH/NH4OH (3:1, v/v)] afforded 34 mg (11%) of a colorless solid having 1H NMR (CD3OD) δ 0.70 (s, 3 H), 0.91 (s, 3 H), 1.00-2.30 (m, 35 H), 2.50-3.00 (m, 12 H), 3.12 (m, 2 H), 3.30-3.40 (m, 1 H), 3.78 (br s, 1 H), 3.95 (br s, 1 H). Anal. Calcd for C34H64N4O4‚3/4 H2O: C, 67.34; H, 10.89; N, 9.24. Found: C, 67.11; H, 10.85; N, 9.83. HRMS for (C34H65N4O4)+ calcd: 593.5006. Found: 593.4997. DNA Binding by Ethidium Displacement. Typically, 3 mL of a 0.01 SHE buffer (8 mM NaCl), 10 mL of an ethidium bromide solution (0.3 mM in 0.01 SHE buffer, 8 mM NaCl), and 10 µL of a DNA solution (0.3 mM in 0.01 SHE buffer, 8 mM NaCl) were added to a cuvette. The fluorescence intensity at 595 nm was then determined after injection of various volumes of the polyamine conjugate (0.3-1.0 mM, methanolic solution). The excitation wavelength was 547 nm. Analogous experiments that were carried out at “high” salt concentrations used a SHE buffer that was 150 mM in NaCl. RESULTS AND DISCUSSION

The synthetic approaches that were used to prepare I-III are outlined in Scheme 1. In brief, activation of cholic acid with N-hydroxysuccinimide (NHS) followed by condensation with spermidine afforded a doublewalled precursor, 1. Subsequent addition of ethyl acrylate, followed by saponification, activation with NHS, and condensation with spermine yielded I. The synthesis of II followed a similar strategy. Thus, protection of the amino group of iminodiacetic acid with the Fmoc moiety, followed by activation with TSU, coupling with 1, deprotection, and conjugation with spermine (in a manner similar to that used for the preparation of I), afforded the desired tetrawalled umbrella, II. Finally, direct coupling of cholic acid to spermine (after activation with NHS) afforded III. To assess the affinity of I-III toward DNA, we have used the ethidium displacement technique (4-7). Here, the concentration of each conjugate that is needed to

Table 1. Binding of Molecular Umbrellas to Calf Thymus DNA, Poly[d(AT)], and Poly[d(GC)] C50a (µM) compound

CT-DNA (1.0)c

spermine 0.90 spermidine 30.0 (27)c I 0.40 II 0.32 III 5.52

poly[d(AT)] poly[d(GC)] poly[d(GC)]b 2.11 (2.7)c 1.47 0.91 10.7

1.15 (1.1)c 0.90 0.90 10.5

>200 4.45 2.20 >200

a Concentration necessary to displace 50% of DNA-bound ethidium under the following conditions: [DNA-bp] ) 1.0 µM, [ethidium] ) 1.0 µM, SHE buffer (8 mM NaCl, 2 mM HEPES, 0.05 mM EDTA, pH 7.0); excitation at 547 nm, emission at 595 nm. Values reported are averages of two independent experiments ( 10%. Methanolic solutions of the polyamine (0-50 µL, 1.0 mM) were injected into 3.0 mL of the DNA-containing buffer at 25 °C. b SHE buffer (150 mM NaCl, 2 mM HEPES, 0.05 mM EDTA, pH 7.0); initial fluorescence intensities were ∼50% of that measured using 8 mM NaCl. c Reference 4.

reduce the fluorescence intensity of ethidium by 50% (only DNA-bound ethidium is fluorescent) is taken as a measure of its affinity toward the nucleic acid. Although such “C50” values cannot be directly converted into binding constants, since the mode and stoichiometry of binding by ethidium and by the sterol polyamines are uncertain, they do provide a useful qualitative means for probing the effects of polyamine structure on DNA binding. Using experimental conditions similar to those previously described (0.01 SHE buffer: 8 mM NaCl, 2 mM HEPES, 0.05 mM EDTA; ethidium concentration ) 1.0 µM; DNA base pair concentration ) 1.0 µM; pH 7, 25 °C), we obtained C50 values for spermine that are in good agreement with those previously reported (4). Since independent studies have estimated a dissociation constant for spermine-DNA complexes in the range of 0.11.0 µM under similar conditions, a C50 value for spermine of ∼1 µM reflects very strong DNA binding (7). A summary of the C50 values for I-III, spermine, and spermidine, with respect to calf thymus (CT) DNA, is presented in Table 1. Introduction of one amphiphilic wall to the spermine molecule (i.e., III) resulted in a significant reduction in DNA binding. The fact that III binds more strongly to DNA than spermidine, however, implies that hydrophobic forces and/or hydrogen bonding between the sterol nucleus and DNA contribute to such complexation. That hydrophobic forces are, in fact, of greater importance is indicated by the fact that replacement of the C-7 hydroxyl group of the sterol by a

894 Bioconjugate Chem., Vol. 8, No. 6, 1997

Janout et al.

Figure 1. Melting behavior of poly[d(AT)] [0.01 SHE buffer: 8 mM NaCl, 2 mM HEPES, 0.05 mM EDTA; DNA base pair concentration ) 6.5 µM; pH 7,] in the presence of I, II, and III: (A) III at (a) 0, (b) 0.065, (c) 0.13, (d) 0.26, and (e) 0.39 µM; (B) I at (a) 0, (b) 0.13, (c) 0.65, and (d) 1.56 µM; (C) II at (a) 0, (b) 0.65, (c) 1.3, (d) 1.95, and (e) 2.6 µM; (D) II at 1.3 µM. In all cases, a heating rate of 2 °C/min was used; absorbance was determined at 260 nm.

hydrogen atom (i.e., an analogous deoxycholic acid derivative) resulted in considerably stronger binding (C50 ) 1.4 µM, not listed). It is noteworthy, in this regard, that a recent study by Burrows and co-workers has also shown that hydrophobic forces can contribute to the binding of certain amine-containing sterols to DNA (4). When the spermine molecule was attached to the doubleand tetrawalled umbrellas, however, significantly greater DNA binding was observed. Similar behavior was found with poly[d(AT)] and poly[d(GC)] (Table 1). In control experiments, inclusion of 2 equiv of cholic acid did not significantly alter the affinity of spermine to poly[d(GC)] or spermidine to CT-DNA (not listed). In addition, the double-walled precursor (1) showed negligible binding toward all three types of DNA (C50 > 200). Previous studies have shown that the binding of spermine and spermidine to DNA is strongly dependent upon the concentration of Na+ that is present: the higher the Na+ concentration, the weaker the polyamine binding (8). This result has been attributed to a competition between the proton-ionized nitrogens of the polyamine and Na+, in serving as counterions in the vicinity of the negatively charged DNA. Consistent with these findings, we have observed a dramatic decrease in the binding of spermine to poly[d(GC)] when the Na+ concentration was increased from 8 to 150 mM (Table 1). Exactly analogous results were observed with III. In sharp contrast, the affinity of each of the umbrella conjugates toward DNA was significantly less dependent on salt concentration;

that is, the C50 values for I and II increased by factors of only ca. 5 and 2, respectively. Thus, the difference in DNA binding between spermine and II (both having four proton-ionizable nitrogens) at high salt concentrations is quite large, i.e., 2 orders of magnitude. To obtain further insight into the interactions I-III with DNA, we examined their influence on the melting behavior (denaturation) of the duplex poly[d(AT)]. As can be seen in Figure 1, striking differences among these three systems are readily apparent. Thus, addition of III leads to a continuous increase in the melting temperature (Tm) in a manner that is very similar to that found with spermine and spermidine (not shown). In the absence of the conjugate, poly[d(AT)] shows a melting temperature of 41 °C; with III, a maximum Tm of 56 °C is observed (Figure 1). In sharp contrast, incremental addition of the tetrawalled umbrella, II, resulted in two distinct melting temperatures,i.e., one that was similar to that of the native DNA and one that corresponded to a stabilized form (Tm ) 56 °C). In this case, incremental addition of the umbrella resulted in a continuous diminution of the former component and an increase in the latter. Qualitatively, the double-walled molecular umbrella, I, exhibited behavior that combined those of II and III. Specifically, the low-temperature transition was gradually shifted to a high-temperature transition, and its overall contribution steadily declined. At high umbrella concentrations, only the high-temperature transition was observed.

Bioconjugate Chem., Vol. 8, No. 6, 1997 895

Molecular Umbrella−Spermine Conjugates

Taken together, the effects of I-III on the melting behavior of poly[d(AT)] and the enhanced binding of the umbrella conjugates in the presence of a high salt concentration provide compelling evidence for highly cooperative binding and umbrella aggregation along the DNA duplex. Thus, the two distinct melting temperatures that are associated with II are a likely consequence of “bare” DNA segments and “stretches” of the duplex that are covered by an aggregated form of the umbrella. As the umbrella concentration is increased, the size and/ or number of such aggregates increases. The fact that submicellar concentrations of II were used in these melting experiments [critical micelle concentration (cmc) in the identical buffer is 31 µM; surface tension] further indicates that DNA serves as a nucleation site for aggregation. The continuous increase in temperature of a single transition for III also indicates that this conjugate is bound to DNA in a more random fashion. The apparent “mixed” behavior of I (cmc ) 40 µM) is a likely consequence of a monomer-aggregate equilibrium that is established on the surface of DNA, where both forms are present in significant amounts. Analogous nucleation phenomena have previously been reported for the binding of cetyltrimethylammonium bromide to DNA (9). Since III contains only a single amphiphilic wall, it is not capable of cooperative hydrophobic interactions with neighboring conjugates. With I and II, however (having two and four amphiphilic walls, respectively), such cooperativity is possible. The high cooperativity that is apparent for the DNA binding of II is a likely consequence of the larger number of walls that are present. The relatively modest differences in DNA binding between I and II, as judged by the ethidium displacement assay, can be readily accounted for if it is assumed that the amphiphilic walls do not play a direct role in the displacement of the dye, e.g., by contacting the DNA. Similar conclusions have recently been drawn for simple bile acid-polyamine conjugates (10). Finally, it should be noted that the cooperativity that we have discovered with these umbrella-based conjugates may also be possible with other small molecules; that is, the presence of amphiphilic walls may not be essential. Our findings do suggest, however, that a prerequisite for such cooperativity is that the ligand contain multiple hydrophobic segments. In principle, cooperative DNA binding of the type described herein may be exploitable from the standpoint of rational drug design. One can envision, for example, that the attachment of molecular umbrellas to those drugs that act on DNA could result in increased potency due to stronger DNA binding. Moreover, the possibility of enhanced permeation across biological membranes could lead to increased efficacy by improving the drug’s accessibility to the cytoplasm and to the nucleus of a cell. In prelimary studies, we have compared the abilities of

I-III to extract CT-DNA from 0.01 SHE buffer into an equal volume of 1-octanol using an umbrella/base pair ratio of 5.0. The percentages of DNA that was extracted were found to be 30, 84, and 3% for I, II, and III, respectively (phosphorus analysis). Although water/ octanol partitioning is distinct from water/bilayer partitioning, these results clearly indicate that molecular umbrellas have the ability to transport DNA into a hydrocarbon environment. This capability, together with their ability to maintain strong DNA binding under physiological salt concentrations, is encouraging in terms of their possible use as gene transfer agents. Efforts aimed at exploring the practical potential of umbrellaspermine conjugates are continuing in our laboratories. ACKNOWLEDGMENT

We are grateful to the National Institutes of Health (PHS Grant GM51814) for support of this research. LITERATURE CITED (1) Janout, V., Lanier, M., and Regen, S. L. (1996) Molecular Umbrellas. J. Am. Chem. Soc. 118, 1573-1574. (2) Janout, V., Lanier, M., and Regen, S. L. (1997) Design and Synthesis of Molecular Umbrellas. J. Am. Chem. Soc. 119, 640-647. (3) Schmid, N., and Behr, J-P. (1991) Location of Spermine and Other Polyamines on DNA As Revealed By Photoaffinity Cleavage With Polyaminobenzenediazonium Salts. Biochemistry 30, 4357-4361. (4) Hsieh, H. P., Muller, J. G., and Burrows, C. J. (1994) Structural Effects In Novel Steroidal Polyamine-DNA Binding. J. Am. Chem. Soc. 116, 12077-12078. (5) Cain, B. F., Baguley, B. C., and Denny, W. A. (1978) Potential Antitumor Agents. 28. Deoxyribonucleic Acid Polyintercalating Agents. J. Med. Chem. 21, 658-668. (6) Stewart, K. D. (1988) The Effect Of Structural Changes In A Polyamine Backbone On Its DNA-Binding Properties. Biochem. Biophys. Res. Commun. 152, 1441-1446. (7) Morgan, J. E., Blankenship, J. W., and Matthews, H. R. (1986) Association Constants For The Interaction Of DoubleStranded And Single-Stranded DNA With Spermine, Spermidine, Putrescine, Diaminopropane, N1- and N8-Acetylspermidine, and Magnesium: Determination From Analysis Of The Broadening Of Thermal Denaturation Curves. Arch. Biochem. Biophys. 246, 225-232. (8) Braunlin, W. H., Strick, T. J., and Record, Jr., M. T. (1982) Equilibrium Dialysis Studies Of Polyamine Binding To DNA. Biopolymers 21, 1301-1314. (9) Mel′nikov, S. M., Sergeyev, V. G., and Yoshikawa, K. (1995) Transition Of Double-Stranded DNA Chains Between Random Coil And Globule States Induced By Cooperative Binding Of Cationic Surfactant. J. Am. Chem. Soc. 117, 9951-9956. (10) Walker, S., Sofia, M. J., Kakarla, R., Kogan, N. A., Wierichs, L., Longley, C. B., Bruker, K., Axelrod, H. R., Midha, S., Babu, S., and Kahne, D. (1996) Cationic Facial Amphiphiles: A Promising Class Of Transfection Agents. Proc. Natl. Acad. Sci. U.S.A. 93, 1585-1590.

BC970142R