**5.4 Optimal structure of the conjugate molecules for gene transfer**

The facile synthetic route provides a variety of polycationic compounds that can be exploited to examine the effect of the polycationic and hydrophobic portions on transfection efficiency. We hence examined the effect on transfection activity of different polyamine

Polyamine – Lipid Conjugates as Effective Gene Carriers:

the transfection activity.

et al., 2010).

**5.7.1 Formation of complex with DNA** 

**5.7 Morphological analysis of the complexes with DNA** 

Chemical Structure, Morphology, and Gene Transfer Activity 255

to condense DNA molecules, which requires both electrostatic and hydrophobic interactions (Yamazaki, 2000; Dewa et al., 2004a,b). We tested a number of combinations of dialkyl and polyamine portions for their activity in gene transfection. The longer alkyl chain exhibited higher efficiency (Figure 5A). The ζ-potential of the DOPE-based PCL increased with alkyl chain length, for example, 27.6, 33.4, and 37.1 mV for **DLP-spd**(PCL), **DMP-spd**(PCL), and **DCP-spd**(PCL), respectively, showing that the conjugate with the longer chain length provides the higher positive potential. The hydrophobic interaction results in stable incorporation of the conjugate into the PCL, consistent with transfection activity in the order of C16 ≥ C14 > C12. We found that the micellar aggregate of **DCP-PEI** conjugate exhibited slightly higher transfection activity than the low-molecular weight amine conjugates, **DCPspd** and **DCP-spm**, however, the activity of **DCP-PEI**(PCL) was marginal. The size of **DCP-PEI**(PCL)/DNA lipoplex is significantly greater than that of the other lipoplexes (Table 2, entry 14). The **DCP-PEI** conjugate is composed of DCP/PEI(1800) = ~1/1, judged by 1H-NMR (Dewa et al., 2004a). As previously reported, the cetyl-PEI, whose PCL possesses high transfection activity, consists of 10 cetyl portions in the polymer (Matsuura et al., 2003). The cetyl-PEI can attach to the PCL surface via the anchoring of cetyl portions in the lipid bilayer. However, that is not the case for **DCP-PEI**(PCL); the single hydrophobic portion in the conjugate is not enough to provide adequate covering of PEI over the PCL surface. This may cause "PEI-protrusion" from the surface, which gives rise to the large and heterogeneous aggregation seen upon combination with DNA molecules. This is likely the reason for the lower activity of the **DCP-PEI**(PCL). Taken together, these considerations suggest that a homogeneous positive charge distribution on the PCL surface is important to

To elucidate the characteristics of N/P-dependence, formation of **DCP-spd**(PCL)/DNA and **DCP-spm**(PCL)/DNA lipoplexes was analyzed by agarose gel electrophoreses and DLS analysis. Electrophoretic analysis showed N/P ratio-dependent complexation; in the lower N/P range of 5–16, the open circular DNA band vanished and the supercoiled DNA band gradually faded for both of **DCP-spd**(PCL) and **DCP-spm**(PCL) and in the higher N/P range, 20–30, the latter DNA band totally disappeared. These observations indicate that the DNA molecules are completely entrapped within the lipoplex. Ethidium bromide (EtBr) replacement experiments also reveal condensation of DNA in the N/P range of 5–30 (Dewa

The particle size of lipoplexes estimated by DLS analysis is summarized in Table 2. The diameters of the **DCP-spd**(PCL) and **DCP-spm**(PCL) alone are 158 ± 56, and 159 ± 30 nm, respectively (entries 1 and 7). Lipoplexes were larger and their size increased with increasing N/P up to 5 (entries 2 and 8 at N/P = 2 and entries 3 and 9 at N/P = 5). At N/P = 5, the PCL/DNA lipoplexes became larger with a broad distribution from 650 nm to over 1 μm (entries 3 and 9). In the higher N/P range (N/P = 16–24), sizes were reduced, converging at 261 ± 114 nm (**DCP-spd**(PCL), entry 5) and 256 ± 116 nm (**DCP-spm**(PCL), entry 11). ζ-potential measurements indicated the polarity of surface charge of lipoplexes inverts from negative in the lower N/P (5–11) to positive in the higher N/P (>16) regions. AFM images revealed characteristic morphologies of lipoplexes in the both low and high N/P ranges (Figures 6A-D). When the **DCP-spd**(PCL)/DNA and **DCP-spm**(PCL)/DNA lipoplexes at N/P = 5 were put on PLL-treated mica (positively charged surface), large

Fig. 5. (A) Effect of polyamine and hydrophobic portions on PCL-mediated gene transfer efficiency. PCL were composed of polyamine conjugate/DOPE/Cholesterol (1/1/1 mol/mol/mol). N/P ratio was 24. Transfection was done in the presence of 10% FBS. LF represents LipofectamineTM 2000 as a positive control experiment. (B) Effect of the N/P ratio on the transfection efficacy of **DCP-spm**(PCL) (**DCP-spm**/DOPE/cholesterol = 1/1/1 (mol/mol/mol)). Transfection was in the presence of 10% FBS.

conjugates incorporated into PCLs: C12, C14, or C16 alkyl chain in the lipophilic portion and spermidine, spermine, or PEI(1800) as the polycationic head group of the conjugate. The data on these compounds are shown in Figure 5A. This result reveals clear tendencies of longer length of the alkyl group and the lower molecular weight of the polyamines (spermidine, spermine) to enhance transfection. When compared with a commercial product, LipofectamineTM 2000, the **DCP-spd**(PCL) possessed 3.6-fold higher activity.

#### **5.5 N/P-dependent efficacy and complexation of PCL with DNA**

Figure 5B shows the dependence of **DCP-spm**(PCL) efficacy on the ratio of the number of nitrogen atoms in the conjugate to that of phosphate in the DNA (N/P). The efficacy increases with the N/P ratio essentially linearly up to 24. A similar N/P-dependence has been also observed for **DCP-spd**(PCL) (data not shown for clarity), indicating that excess polyamine relative to DNA is needed for effective gene transfer by PCLs. In contrast, the transfection activity of micellar aggregates **DCP-spd** and **DCP-spm** reaches plateau values in the N/P range of 11–16 (1~2 × 109 of RLU/mg protein). This tendency is consistent with our previous data obtained with the β-galactosidase expression system (Dewa et al., 2004a).

#### **5.6 Chemical structure of the polyamine conjugates**

The polyamine–dialkyl phosphate conjugates can be readily synthesized via a two-step reaction: (i) formation of dimerized dialkyl phosphate anhydride and (ii) its nucleophilic substitution with polyamines. The synthetic strategy gives access to a wide variety of polyamine–dialkyl phosphate derivatives. Conjugation of the polyamine and hydrophobic portions is required for an effective gene carrier (Figure 4). Such amphiphilicity is essential

Fig. 5. (A) Effect of polyamine and hydrophobic portions on PCL-mediated gene transfer efficiency. PCL were composed of polyamine conjugate/DOPE/Cholesterol (1/1/1 mol/mol/mol). N/P ratio was 24. Transfection was done in the presence of 10% FBS. LF represents LipofectamineTM 2000 as a positive control experiment. (B) Effect of the N/P ratio on the transfection efficacy of **DCP-spm**(PCL) (**DCP-spm**/DOPE/cholesterol = 1/1/1

conjugates incorporated into PCLs: C12, C14, or C16 alkyl chain in the lipophilic portion and spermidine, spermine, or PEI(1800) as the polycationic head group of the conjugate. The data on these compounds are shown in Figure 5A. This result reveals clear tendencies of longer length of the alkyl group and the lower molecular weight of the polyamines (spermidine, spermine) to enhance transfection. When compared with a commercial product, LipofectamineTM 2000, the **DCP-spd**(PCL) possessed 3.6-fold higher activity.

Figure 5B shows the dependence of **DCP-spm**(PCL) efficacy on the ratio of the number of nitrogen atoms in the conjugate to that of phosphate in the DNA (N/P). The efficacy increases with the N/P ratio essentially linearly up to 24. A similar N/P-dependence has been also observed for **DCP-spd**(PCL) (data not shown for clarity), indicating that excess polyamine relative to DNA is needed for effective gene transfer by PCLs. In contrast, the transfection activity of micellar aggregates **DCP-spd** and **DCP-spm** reaches plateau values in the N/P range of 11–16 (1~2 × 109 of RLU/mg protein). This tendency is consistent with our previous data obtained with the β-galactosidase expression system (Dewa et al., 2004a).

The polyamine–dialkyl phosphate conjugates can be readily synthesized via a two-step reaction: (i) formation of dimerized dialkyl phosphate anhydride and (ii) its nucleophilic substitution with polyamines. The synthetic strategy gives access to a wide variety of polyamine–dialkyl phosphate derivatives. Conjugation of the polyamine and hydrophobic portions is required for an effective gene carrier (Figure 4). Such amphiphilicity is essential

(mol/mol/mol)). Transfection was in the presence of 10% FBS.

**5.5 N/P-dependent efficacy and complexation of PCL with DNA** 

**5.6 Chemical structure of the polyamine conjugates** 

to condense DNA molecules, which requires both electrostatic and hydrophobic interactions (Yamazaki, 2000; Dewa et al., 2004a,b). We tested a number of combinations of dialkyl and polyamine portions for their activity in gene transfection. The longer alkyl chain exhibited higher efficiency (Figure 5A). The ζ-potential of the DOPE-based PCL increased with alkyl chain length, for example, 27.6, 33.4, and 37.1 mV for **DLP-spd**(PCL), **DMP-spd**(PCL), and **DCP-spd**(PCL), respectively, showing that the conjugate with the longer chain length provides the higher positive potential. The hydrophobic interaction results in stable incorporation of the conjugate into the PCL, consistent with transfection activity in the order of C16 ≥ C14 > C12. We found that the micellar aggregate of **DCP-PEI** conjugate exhibited slightly higher transfection activity than the low-molecular weight amine conjugates, **DCPspd** and **DCP-spm**, however, the activity of **DCP-PEI**(PCL) was marginal. The size of **DCP-PEI**(PCL)/DNA lipoplex is significantly greater than that of the other lipoplexes (Table 2, entry 14). The **DCP-PEI** conjugate is composed of DCP/PEI(1800) = ~1/1, judged by 1H-NMR (Dewa et al., 2004a). As previously reported, the cetyl-PEI, whose PCL possesses high transfection activity, consists of 10 cetyl portions in the polymer (Matsuura et al., 2003). The cetyl-PEI can attach to the PCL surface via the anchoring of cetyl portions in the lipid bilayer. However, that is not the case for **DCP-PEI**(PCL); the single hydrophobic portion in the conjugate is not enough to provide adequate covering of PEI over the PCL surface. This may cause "PEI-protrusion" from the surface, which gives rise to the large and heterogeneous aggregation seen upon combination with DNA molecules. This is likely the reason for the lower activity of the **DCP-PEI**(PCL). Taken together, these considerations suggest that a homogeneous positive charge distribution on the PCL surface is important to the transfection activity.
