Aromatic Modification of Low Molecular Weight PEI for Enhanced Gene Delivery

Low molecular weight polyethylenimine (1800 Da, also referred to as oligoethylenimines, OEI) was modified with amino acids, including two aromatic amino acids (tryptophan, phenylalanine) and an aliphatic amino acid (leucine). The substitution degree of amino acids could be controlled by adjusting the feeding mole ratio of the reactants. Fluorescence spectroscopy and circular dichroism experiments demonstrated that the indole ring of tryptophan may intercalate into the DNA base pairs and contribute to efficient DNA condensation. In vitro gene expression results revealed that the modified OEIs (OEI-AAs) may provide higher transfection efficiency even than high molecular weight polyethylenimine (25 kDa, PEI), especially the aromatic tryptophan substituted OEI. Moreover, OEI-AAs exhibited excellent serum tolerance, and up to 137 times higher transfection efficiency than PEI 25 kDa that was obtained in the presence of serum. The cytotoxicity of OEI-AAs is much lower than PEI 25 kDa. This study may afford a new method for the development of low molecular weight oligomeric non-viral gene vectors with both high efficiency and biocompatibility.


Introduction
Nucleic acids are attractive compounds for modulating cell function with high precision [1,2]. Since naked nucleic acids may be easily degraded by nuclease, and the structure of negatively charged phosphates restrict the uptake into cells. Thus, safe and efficient gene delivery vectors are urgently needed [3][4][5][6]. Compared with viral vectors, polycations show lower immunogenicity and a high nucleic acid loading capacity, and they can be easily prepared and their chemical structures may be conveniently modified, exhibiting potential for the construction of ideal non-viral gene delivery systems [7][8][9][10]. Among the "off-the-shelf" polycations, polyethylenimine (PEI) has been regarded as the benchmark for newly designed non-viral polycationic vectors for its high condensation capability toward DNA, strong buffering capacity, and relatively high gene transfection efficiency (TE) [11][12][13]. However, like other polycations for gene delivery, PEI also faces several extracellular and intracellular barriers such as serum stability, cellular uptake, endosomal escape, and nuclear localization [14,15]. The major limitations for its application include the following: firstly, high molecular weight (HMW) PEI (such as 25 kDa) has high TE but considerable toxicity and limited biocompatibility, whereas low molecular weight (LMW) PEI has low toxicity but poor TE [16,17]. It was found that PEI with a molecular weight <2 kDa exhibited almost no in vitro cytotoxicity and low TE even at high concentrations [18,19]. Secondly, PEI has a weak serum tolerance ability, which is necessary for in vivo application [20]. Modification of PEI with some functional groups may overcome specific gene delivery
The ethidium bromide (EB) displacement assay, particle size and zeta-potential measurement (DLS), cellular uptake of plasmid DNA (flow cytometry) and cytotoxicity were carried out according to our previously reported procedures [16,39,40].

Modification of OEI
N-(tert-Butoxycarbonyl)-amino acid (2 mmol), N-(3-dimethylaminopropyl)-N -ethylcarbodiimide hydrochloride (EDCI, 2.4 mmol), 1-hydroxybenzotriazole (HOBT, 2.4 mmol), and DIEA (ethyldiisopropylamine, 2.4 mmol) in dry dichloromethane (50 mL) were stirred for 2 h in an ice bath. Then the desired amount of OEI (20%, 30% and 50%, carboxyl/amine in OEI, mol/mol) was dissolved in dry dichloromethane and added to the reaction mixture and stirred for two days at room temperature. After completion of the reaction, diluted HCl solution (0.5 N in water) was added to remove the protecting t-butyloxycarbonyl (Boc) group, and the reaction mixture was stirred for another day. The solvent was removed under reduced pressure and the residue with a small amount of water was dialyzed (MWCO 1000 Da) against deionized water for three days. The product was obtained as a white or pale yellow solid after lyophilization. Yield: 16.9-57.5%.
OEI-W: 1  The substitution degrees (SD), which represents the mole ratio of amino acid to the amine groups in OEI, were calculated from 1H NMR spectra (OEI-W10 for example, Figure S6A in Supplementary Materials).

Gel Retardation Assay
Polyplexes at different w/w ratios (weight ratios of polymer relative to pDNA) were prepared by adding appropriate volumes of the polymer solution to 5 µL of Puc-19 (0.025 mg/mL). The obtained complex solution was then diluted to the total volume of 10 µL. After incubation at 37 • C for 30 min, 2.5 µL of loading buffer was added. Then the mixture was electrophoresed on a 1% (w/v) agarose gel containing GelRed TM in a Triseacetate (TAE) running buffer at 120 V for 40 min. Then DNA was visualized under an ultraviolet lamp using a BioRad Universal Hood II.

Fluorescence Spectroscopy
Fluorescence spectra were measured at room temperature in air by a HITACHI F-7000 Fluorescence Spectrophotometer (Hitachi High-Tech Science, Schaumburg, IL, USA) and corrected for the system response. Polymers (30 µL, 1 mg/mL) were put into a quartz cuvette containing 2.5 mL Hepes solution. After shaking, the fluorescence intensity of the polymers (F p ) was measured. Then Calf Thymus (CT) DNA with various w/w ratios to polymers was added to the solution and mixed symmetrically, and the measured fluorescence intensity (F) was the result of the interaction between the DNA and the polymers. All the samples were excited at 288 nm and the emission was measured at 378 nm.

Circular Dichroism (CD)
The circular dichroism of DNA was testified by Chirascan (Applied Photophysics, Surrey, UK) with a path length cell of 1 cm at room temperature. CT DNA (1 mg/mL, 100 µL) was added to a Hepes solution (pH 7.4, 2 mL). After the sample was mixed well, the circular dichroism was measured. Then, the solution of polymers (1 mg/mL, 20 or 40 µL) was added each time, and the circular dichroism was measured. The standard scan parameters for all experiments used a wavelength range from 350 to 230 nm. Three scans were made and the average value of them was calculated.

Transmission Electron Microscopy (TEM)
The morphologies of the polyplexes were observed by TEM (Hitachi High-Tech Science, Schaumburg, IL, USA) with an acceleration voltage of 100 kV. A total of 4 µg of pUC-19 was added to the appropriate volume of the polymer solution (weight ratio of polymer relative to pDNA, w/w = 16:1 for OEI-W10&OEI-L10 and w/w = 24:1 for OEI-F10&OEI), and incubated at 37 • C for 0.5 h. Then the polyplex solution was diluted to the total volume of 200 µL. A drop of DNA/polymer polyplex suspension was placed onto the copper grid. After a few minutes, the excess solution was blotted away with filter paper. Then, a drop of 0.5% (w/v) phosphotungstic acid was placed on the above grid. The grid was dried at room temperature at atmospheric pressure for several minutes before observation.

Gene Transfection Efficiency Assay In Vitro
For the expression of the EGFP gene Hela and HEK293 cells were seeded at a density of 1 × 10 5 cells per well in a 24-well plate in DMEM containing 10% FBS and grown to reach 70-80% confluence prior to transfection. Then, the medium of each well was exchanged for fresh, serum-free medium. Subsequently, the cells were treated with polyplexes (containing 0.8 µg of pEGFP) at different (w/w) weight ratios and PEI/DNA at a weight ratio of 1.4 (N/P = 10) for 4 h at 37 • C. The medium was then completely refreshed with the completed culture media. 24 h after transfection, cells were observed with an inverted fluorescence microscope (Nikon Eclipse TS 100, Nikon Corporation, Tokyo, Japan) equipped with a cold Nikon camera. Digital image recording and image analysis were performed with the NIS Elements Advanced Research software (version 4.0).
For the luciferase assays, cells were transfected by polyplexes containing pGL-3. For a typical assay in a 24-well plate, 24 h post transfection as described above, cells were washed with warm PBS and lysed with 100 µL 1 × lysis reporter buffer (Promega, Fitchburg, MA, USA). The luciferase activity was measured by microplate reader (Model 550, BioRad, Hercules, CA, USA). The protein concentration in the cell lysate supernatant was estimated in each case with a Lowry protein assay kit (PIERCE, Rockford, IL, USA). Gene transfection efficiency was expressed as the relative fluorescence intensity per mg of protein (RLU/mg protein). All the experiments were done in triplicate.

Synthesis and Characterization of the Modified OEIs (OEI-AAs)
As shown in Scheme 1, tryptophan (W), phenylalanine (F) and leucine (L), which contain aromatic indole, aromatic phenyl and aliphatic iso-butyl group, respectively, were chosen to modify OEI (1800 Da) with various substitution degrees (SD). The primary amine groups of the amino acids were previously protected by the Boc group. The Boc-amino acids were then coupled with OEI through the formation of amide groups. The SD, which may be calculated from the specific peak integrals of the 1 H NMR spectra, was controlled by adjusting the feeding mole ratio of amino acids toward the amine groups in OEI (M W = 43 g/mol for each ethylenimine unit). Generally, for the steric effect and the reactivity diversity of 1 • /2 • /3 • amines in OEI, the obtained SDs were lower than the relative feeding ratio (Table 1). It is worth noting that the number in the product name refers to the actual (calculated) SD, e.g., OEI-W10 means that the SD of tryptophan is about 10%. Due to the various structures of different amino acids, the same feeding ratio would not lead to OEI-AA with similar SDs. High SD was hard to achieve for the steric issue, while very low SD could not exhibit the role of amino acids. Thus, we chose products with SDs in the range of around 10-20%, which are listed in Table 1. suspension was placed onto the copper grid. After a few minutes, the excess solution was blotted away with filter paper. Then, a drop of 0.5% (w/v) phosphotungstic acid was placed on the above grid. The grid was dried at room temperature at atmospheric pressure for several minutes before observation.

Gene Transfection Efficiency Assay In Vitro
For the expression of the EGFP gene Hela and HEK293 cells were seeded at a density of 1 × 10 5 cells per well in a 24-well plate in DMEM containing 10% FBS and grown to reach 70-80% confluence prior to transfection. Then, the medium of each well was exchanged for fresh, serum-free medium. Subsequently, the cells were treated with polyplexes (containing 0.8 µg of pEGFP) at different (w/w) weight ratios and PEI/DNA at a weight ratio of 1.4 (N/P = 10) for 4 h at 37 °C. The medium was then completely refreshed with the completed culture media. 24 h after transfection, cells were observed with an inverted fluorescence microscope (Nikon Eclipse TS 100, Nikon Corporation, Tokyo, Japan) equipped with a cold Nikon camera. Digital image recording and image analysis were performed with the NIS Elements Advanced Research software (version 4.0).
For the luciferase assays, cells were transfected by polyplexes containing pGL-3. For a typical assay in a 24-well plate, 24 h post transfection as described above, cells were washed with warm PBS and lysed with 100 µL 1 × lysis reporter buffer (Promega, Fitchburg, MA, USA). The luciferase activity was measured by microplate reader (Model 550, BioRad, Hercules, CA, USA). The protein concentration in the cell lysate supernatant was estimated in each case with a Lowry protein assay kit (PIERCE, Rockford, IL, USA). Gene transfection efficiency was expressed as the relative fluorescence intensity per mg of protein (RLU/mg protein). All the experiments were done in triplicate.

Synthesis and Characterization of the Modified OEIs (OEI-AAs)
As shown in Scheme 1, tryptophan (W), phenylalanine (F) and leucine (L), which contain aromatic indole, aromatic phenyl and aliphatic iso-butyl group, respectively, were chosen to modify OEI (1800 Da) with various substitution degrees (SD). The primary amine groups of the amino acids were previously protected by the Boc group. The Boc-amino acids were then coupled with OEI through the formation of amide groups. The SD, which may be calculated from the specific peak integrals of the 1 H NMR spectra, was controlled by adjusting the feeding mole ratio of amino acids toward the amine groups in OEI (MW = 43 g/mol for each ethylenimine unit). Generally, for the steric effect and the reactivity diversity of 1°/2°/3° amines in OEI, the obtained SDs were lower than the relative feeding ratio (Table 1). It is worth noting that the number in the product name refers to the actual (calculated) SD, e.g., OEI-W10 means that the SD of tryptophan is about 10%. Due to the various structures of different amino acids, the same feeding ratio would not lead to OEI-AA with similar SDs. High SD was hard to achieve for the steric issue, while very low SD could not exhibit the role of amino acids. Thus, we chose products with SDs in the range of around 10-20%, which are listed in Table 1.

Interaction with DNA and Formation of Polyplex
The primary requirement for polymeric gene vectors is the ability to bind and condense plasmid DNA. Agarose gel retardation assay was first used to estimate the DNA-binding ability of these polycations. As shown in Figure 1, the studied OEI-AAs could completely retard DNA migration at a small polymer/DNA weight ratio (w/w) of 0.5. On the other hand, this w/w was slightly larger than the cases using OEI or PEI 25 kDa ( Figure S1), suggesting that the surface modification of OEI might have negative steric effect on the interaction with DNA. However, subsequent experiments revealed that the transfection efficiency could be largely improved by the modification, indicating that excessively strong DNA binding might hinder the DNA release and gene transfection. The ethidium bromide (EB) exclusion experiment gave consistent results ( Figure S2). All the tested materials could efficiently displace EB that previously intercalated into DNA base pairs, leading to fluorescence quenching. Unmodified OEI and PEI could quench the fluorescence of EB more quickly than the modified forms, but the difference was small. OEI-W and OEI-F had slightly better binding ability than OEI-L, indicating that the aromatic rings might facilitate their interaction with DNA. To examine the special property of aromatic rings toward DNA binding, we also studied the fluorescence change of OEI-W before and after its interaction with DNA. Results show that the fluorescence intensity of OEI-W10 at 354 nm (the characteristic peak corresponding to the tryptophan residues) gradually lowered with the increase in the amount of DNA (decrease of w/w, Figure S3), suggesting that the aromatic indole rings of tryptophan might intercalate into the DNA base pairs, leading to distinct fluorescence change [37]. On the contrary, a similar phenomenon was not observed for OEI-F10, which has aromatic phenyl rings in the structure. Thus, it was believed that the indole ring may play special role in the interaction with DNA. Further, heparin was used to study the release of DNA from the polyplexes. It was found that DNA could be gradually released with the increasing amount of heparin ( Figure S4). Similar to the above results, since OEI has a higher DNA-binding ability than the modified forms, more heparin was needed to release the DNA from the polyplex formed from OEI.

Interaction with DNA and Formation of Polyplex
The primary requirement for polymeric gene vectors is the ability to bind and condense plasmid DNA. Agarose gel retardation assay was first used to estimate the DNA-binding ability of these polycations. As shown in Figure 1, the studied OEI-AAs could completely retard DNA migration at a small polymer/DNA weight ratio (w/w) of 0.5. On the other hand, this w/w was slightly larger than the cases using OEI or PEI 25 kDa ( Figure S1), suggesting that the surface modification of OEI might have negative steric effect on the interaction with DNA. However, subsequent experiments revealed that the transfection efficiency could be largely improved by the modification, indicating that excessively strong DNA binding might hinder the DNA release and gene transfection. The ethidium bromide (EB) exclusion experiment gave consistent results ( Figure  S2). All the tested materials could efficiently displace EB that previously intercalated into DNA base pairs, leading to fluorescence quenching. Unmodified OEI and PEI could quench the fluorescence of EB more quickly than the modified forms, but the difference was small. OEI-W and OEI-F had slightly better binding ability than OEI-L, indicating that the aromatic rings might facilitate their interaction with DNA. To examine the special property of aromatic rings toward DNA binding, we also studied the fluorescence change of OEI-W before and after its interaction with DNA. Results show that the fluorescence intensity of OEI-W10 at 354 nm (the characteristic peak corresponding to the tryptophan residues) gradually lowered with the increase in the amount of DNA (decrease of w/w, Figure S3), suggesting that the aromatic indole rings of tryptophan might intercalate into the DNA base pairs, leading to distinct fluorescence change [37]. On the contrary, a similar phenomenon was not observed for OEI-F10, which has aromatic phenyl rings in the structure. Thus, it was believed that the indole ring may play special role in the interaction with DNA. Further, heparin was used to study the release of DNA from the polyplexes. It was found that DNA could be gradually released with the increasing amount of heparin ( Figure S4). Similar to the above results, since OEI has a higher DNA-binding ability than the modified forms, more heparin was needed to release the DNA from the polyplex formed from OEI. Free B-type DNA shows a typical circular dichroism (CD) spectrum, and the addition of polycations would make special changes on the CD spectra. As shown in Figure 2, with the addition of OEI-W10, the molar ellipticities of DNA at 245 nm and 279 nm were distinctly suppressed from a w/w of 0.6 together with a certain quantity of red shift. We speculate that the signal change might be ascribed to the intercalation of indole rings into the DNA base pairs [16,41]. On the other hand, other polycations just led to a slight red shift but no obvious suppression of molar ellipticity. These results also indicated the unique role of indole or tryptophan in DNA binding. Free B-type DNA shows a typical circular dichroism (CD) spectrum, and the addition of polycations would make special changes on the CD spectra. As shown in Figure 2, with the addition of OEI-W10, the molar ellipticities of DNA at 245 nm and 279 nm were distinctly suppressed from a w/w of 0.6 together with a certain quantity of red shift. We speculate that the signal change might be ascribed to the intercalation of indole rings into the DNA base pairs [16,41]. On the other hand, other polycations just led to a slight red shift but no obvious suppression of molar ellipticity. These results also indicated the unique role of indole or tryptophan in DNA binding. The physical properties such as particle size and zeta potential of the formed polyplexes were subsequently studied by the dynamic light scattering (DLS) method. The results in Figure 3A show that the particle size became stable from the w/w of six for all polycationic materials, and the values were about 100~500 nm, which is appropriate for cellular uptake [42]. Although the OEI-AAs showed slightly weaker DNA-binding ability than OEI, the particles formed from OEI-AAs were smaller, indicating their better DNA condensation ability. Plasmid DNA is negatively charged in neutral conditions. With the addition of OEI-AAs, the zeta potential of the polyplexes turned positive at a relatively low w/w, <1, and finally reached a plateau of +47~50 mV ( Figure 3B). Such a positive charge might increase their interaction with the negatively charged cell membrane and promote cellular uptake. Moreover, the morphology of the polyplex was directly observed by TEM. The images in Figure 4 show that in deionized water, all materials could condense DNA into uniform regular spherical nanoparticles with a diameter of 80 ± 20 nm. The discrepancy between the sizes measured by DLS and TEM could be ascribed to the different environments of the testing sample-DLS determined the hydrodynamic diameter of the particles in water, while TEM showed the morphology of the particles in their dehydrated state [43].  The physical properties such as particle size and zeta potential of the formed polyplexes were subsequently studied by the dynamic light scattering (DLS) method. The results in Figure 3A show that the particle size became stable from the w/w of six for all polycationic materials, and the values were about 100~500 nm, which is appropriate for cellular uptake [42]. Although the OEI-AAs showed slightly weaker DNA-binding ability than OEI, the particles formed from OEI-AAs were smaller, indicating their better DNA condensation ability. Plasmid DNA is negatively charged in neutral conditions. With the addition of OEI-AAs, the zeta potential of the polyplexes turned positive at a relatively low w/w, <1, and finally reached a plateau of +47~50 mV ( Figure 3B). Such a positive charge might increase their interaction with the negatively charged cell membrane and promote cellular uptake. Moreover, the morphology of the polyplex was directly observed by TEM. The images in Figure 4 show that in deionized water, all materials could condense DNA into uniform regular spherical nanoparticles with a diameter of 80 ± 20 nm. The discrepancy between the sizes measured by DLS and TEM could be ascribed to the different environments of the testing sample-DLS determined the hydrodynamic diameter of the particles in water, while TEM showed the morphology of the particles in their dehydrated state [43]. The physical properties such as particle size and zeta potential of the formed polyplexes were subsequently studied by the dynamic light scattering (DLS) method. The results in Figure 3A show that the particle size became stable from the w/w of six for all polycationic materials, and the values were about 100~500 nm, which is appropriate for cellular uptake [42]. Although the OEI-AAs showed slightly weaker DNA-binding ability than OEI, the particles formed from OEI-AAs were smaller, indicating their better DNA condensation ability. Plasmid DNA is negatively charged in neutral conditions. With the addition of OEI-AAs, the zeta potential of the polyplexes turned positive at a relatively low w/w, <1, and finally reached a plateau of +47~50 mV ( Figure 3B). Such a positive charge might increase their interaction with the negatively charged cell membrane and promote cellular uptake. Moreover, the morphology of the polyplex was directly observed by TEM. The images in Figure 4 show that in deionized water, all materials could condense DNA into uniform regular spherical nanoparticles with a diameter of 80 ± 20 nm. The discrepancy between the sizes measured by DLS and TEM could be ascribed to the different environments of the testing sample-DLS determined the hydrodynamic diameter of the particles in water, while TEM showed the morphology of the particles in their dehydrated state [43].

In Vitro Gene Transfection
To investigate the gene transfection efficiency of OEI-AAs, the luciferase reporter gene pGL-3 plasmid was first used to quantitatively assess their in vitro TE in HEK293 cells. Figure 5 shows the transfection results, which were described as relative TE by normalizing the TE of PEI (w/w = 1.4, N/P = 10) as 1. It is well known that low molecular weight PEI (e.g., OEI) is less toxic than its higher molecular weight counterpart such as PEI 25 kDa, but its TE is lower. This was demonstrated by the experiments, in which OEI indeed showed much lower TE. On the contrary, to our delight, the modification of OEI with amino acids was able to greatly improve the TE, which was even higher than that of PEI 25 kDa. OEI modified by tryptophan exhibited the best transfection results, and the TE reached levels up to ten times higher than PEI (OEI-W10). The modification with phenylalanine or leucine was also able to enhance the TE of OEI, but to a lesser extent. The optimal TE of OEI-F and OEI-L were similar to PEI. On the other hand, results also show that OEI-AAs with an SD of 10% are more suitable for transfection. Further, the transfection experiments were also carried out in serum circumstances to estimate the serum tolerance of the materials ( Figure 5B). With the presence of 10% serum, the OEIs modified by the amino acids containing aromatic rings were able to give much higher relative TE than those without serum, especially for OEI-W10, which gave 96 times higher TE than PEI. We speculate that the excellent serum tolerance might come from their good DNA condensation ability, making the polyplex more stable.

In Vitro Gene Transfection
To investigate the gene transfection efficiency of OEI-AAs, the luciferase reporter gene pGL-3 plasmid was first used to quantitatively assess their in vitro TE in HEK293 cells. Figure 5 shows the transfection results, which were described as relative TE by normalizing the TE of PEI (w/w = 1.4, N/P = 10) as 1. It is well known that low molecular weight PEI (e.g., OEI) is less toxic than its higher molecular weight counterpart such as PEI 25 kDa, but its TE is lower. This was demonstrated by the experiments, in which OEI indeed showed much lower TE. On the contrary, to our delight, the modification of OEI with amino acids was able to greatly improve the TE, which was even higher than that of PEI 25 kDa. OEI modified by tryptophan exhibited the best transfection results, and the TE reached levels up to ten times higher than PEI (OEI-W10). The modification with phenylalanine or leucine was also able to enhance the TE of OEI, but to a lesser extent. The optimal TE of OEI-F and OEI-L were similar to PEI. On the other hand, results also show that OEI-AAs with an SD of 10% are more suitable for transfection. Further, the transfection experiments were also carried out in serum circumstances to estimate the serum tolerance of the materials ( Figure 5B). With the presence of 10% serum, the OEIs modified by the amino acids containing aromatic rings were able to give much higher relative TE than those without serum, especially for OEI-W10, which gave 96 times higher TE than PEI. We speculate that the excellent serum tolerance might come from their good DNA condensation ability, making the polyplex more stable.

In Vitro Gene Transfection
To investigate the gene transfection efficiency of OEI-AAs, the luciferase reporter gene pGL-3 plasmid was first used to quantitatively assess their in vitro TE in HEK293 cells. Figure 5 shows the transfection results, which were described as relative TE by normalizing the TE of PEI (w/w = 1.4, N/P = 10) as 1. It is well known that low molecular weight PEI (e.g., OEI) is less toxic than its higher molecular weight counterpart such as PEI 25 kDa, but its TE is lower. This was demonstrated by the experiments, in which OEI indeed showed much lower TE. On the contrary, to our delight, the modification of OEI with amino acids was able to greatly improve the TE, which was even higher than that of PEI 25 kDa. OEI modified by tryptophan exhibited the best transfection results, and the TE reached levels up to ten times higher than PEI (OEI-W10). The modification with phenylalanine or leucine was also able to enhance the TE of OEI, but to a lesser extent. The optimal TE of OEI-F and OEI-L were similar to PEI. On the other hand, results also show that OEI-AAs with an SD of 10% are more suitable for transfection. Further, the transfection experiments were also carried out in serum circumstances to estimate the serum tolerance of the materials ( Figure 5B). With the presence of 10% serum, the OEIs modified by the amino acids containing aromatic rings were able to give much higher relative TE than those without serum, especially for OEI-W10, which gave 96 times higher TE than PEI. We speculate that the excellent serum tolerance might come from their good DNA condensation ability, making the polyplex more stable.  The OEI-AAs with 10% SD were also applied to the transfection toward tumor cells (HeLa). As shown in Figure 6A, the results were similar to those obtained in HEK293 cells. OEI-W10 showed the highest TE, and all of the OEI-AAs had higher TE than PEI. The green fluorescent protein (GFP) transfection assay ( Figure S5) visually revealed that OEI induced much less green fluorescence than PEI, but the modified form could improve the TE, especially the tryptophan-substituted OEI-W. With the presence of serum, the relative TE of OEI-AAs increased significantly ( Figure 6B), and OEI-W10 showed 25 times higher levels of TE than PEI, further indicating its excellent serum tolerance. OEI-F10 and OEI-L10 could also exhibit much better transfection than PEI in serum circumstances. Moreover, the effect of serum concentrations on the TE was also studied ( Figure 6C). It was confirmed that the TE of OEI-W10 was less negatively affected by serum than that of PEI, and the relative TE (PEI as standard) continuously increased along with the rise of serum concentration. The OEI-AAs with 10% SD were also applied to the transfection toward tumor cells (HeLa). As shown in Figure 6A, the results were similar to those obtained in HEK293 cells. OEI-W10 showed the highest TE, and all of the OEI-AAs had higher TE than PEI. The green fluorescent protein (GFP) transfection assay ( Figure S5) visually revealed that OEI induced much less green fluorescence than PEI, but the modified form could improve the TE, especially the tryptophan-substituted OEI-W.
With the presence of serum, the relative TE of OEI-AAs increased significantly ( Figure 6B), and OEI-W10 showed 25 times higher levels of TE than PEI, further indicating its excellent serum tolerance. OEI-F10 and OEI-L10 could also exhibit much better transfection than PEI in serum circumstances. Moreover, the effect of serum concentrations on the TE was also studied ( Figure 6C). It was confirmed that the TE of OEI-W10 was less negatively affected by serum than that of PEI, and the relative TE (PEI as standard) continuously increased along with the rise of serum concentration. To figure out the positive effect of amino acid modification on both TE and serum tolerance, the cellular uptake of the polyplexes was subsequently studied. Flow cytometry was used to measure the HeLa cell internalization by using Cy5-labelled DNA via the fluorescence activated cell sorting (FACS) technique. After 4 h incubation with the polyplexes, the percentage of cells positive for Cy5-labelled pDNA and mean fluorescence intensity were calculated and shown in Figure 7. To our surprise, in the absence of serum, both the uptake cell percentage and fluorescence intensity induced by OEI and OEI-AAs were lower than those by PEI. Thus the higher TE of OEI-AAs might come from more efficient intracellular delivery [44]. We postulate that polyplexes formed from higher molecular weight PEI would suffer more difficult DNA release, leading to lower TE [31]. With the addition of serum, the uptake cell percentage and fluorescence intensity induced by PEI were To figure out the positive effect of amino acid modification on both TE and serum tolerance, the cellular uptake of the polyplexes was subsequently studied. Flow cytometry was used to measure the HeLa cell internalization by using Cy5-labelled DNA via the fluorescence activated cell sorting (FACS) technique. After 4 h incubation with the polyplexes, the percentage of cells positive for Cy5-labelled pDNA and mean fluorescence intensity were calculated and shown in Figure 7. To our surprise, in the absence of serum, both the uptake cell percentage and fluorescence intensity induced by OEI and OEI-AAs were lower than those by PEI. Thus the higher TE of OEI-AAs might come from more efficient intracellular delivery [44]. We postulate that polyplexes formed from higher molecular weight PEI would suffer more difficult DNA release, leading to lower TE [31]. With the addition of serum, the uptake cell percentage and fluorescence intensity induced by PEI were severely inhibited, and the results were at the same level as those of OEI-AAs. In other words, the decrease of cellular uptake of OEI-AA/DNA polyplexes was less than that of PEI/DNA, suggesting that OEI-AAs have better serum tolerance ability.
Polymers 2017, 9, 362 9 of 12 severely inhibited, and the results were at the same level as those of OEI-AAs. In other words, the decrease of cellular uptake of OEI-AA/DNA polyplexes was less than that of PEI/DNA, suggesting that OEI-AAs have better serum tolerance ability.

In Vitro Cytotoxicity
The cell viability of the polyplexes prepared at various weight ratios was evaluated by comparison with that of PEI/DNA in HEK293 and HeLa cells by using an MTS assay. As shown in Figure 8, the cell viability of the polyplexes formed from OEI-AAs is indeed a bit lower than OEI, indicating that modification with amino acids would lead to slight increase of cytotoxicity. This might be due to the increase of the molecular weight via the modification [45,46]. However, the cytotoxicities under optimal transfection w/w ratios (16~24) were acceptable. On the other hand, the cytotoxicity of OEI-AAs was much lower than PEI, especially at higher weight ratios. Compared to PEI 25 kDa, the much higher TE together with lower cytotoxicity made the OEI-AAs good candidates for further applications.

Conclusions
In summary, PEI with a relatively low molecular weight (1800 Da, also described as OEI) was modified with three amino acids to study the gene delivery capability of the target materials. The substitution degree of the amino acids could be calculated from the specific peak integrals in the 1 H NMR spectra. It was found that the modification would slightly reduce the DNA binding ability, but the OEI-AAs could condense DNA into more compact nanoparticles. CD results indicated that the indole ring of tryptophan might intercalate into the DNA base pairs, resulting in more efficient condensation. In vitro transfection experiments revealed that the modification of OEI, especially

In Vitro Cytotoxicity
The cell viability of the polyplexes prepared at various weight ratios was evaluated by comparison with that of PEI/DNA in HEK293 and HeLa cells by using an MTS assay. As shown in Figure 8, the cell viability of the polyplexes formed from OEI-AAs is indeed a bit lower than OEI, indicating that modification with amino acids would lead to slight increase of cytotoxicity. This might be due to the increase of the molecular weight via the modification [45,46]. However, the cytotoxicities under optimal transfection w/w ratios (16~24) were acceptable. On the other hand, the cytotoxicity of OEI-AAs was much lower than PEI, especially at higher weight ratios. Compared to PEI 25 kDa, the much higher TE together with lower cytotoxicity made the OEI-AAs good candidates for further applications.
Polymers 2017, 9, 362 9 of 12 severely inhibited, and the results were at the same level as those of OEI-AAs. In other words, the decrease of cellular uptake of OEI-AA/DNA polyplexes was less than that of PEI/DNA, suggesting that OEI-AAs have better serum tolerance ability.

In Vitro Cytotoxicity
The cell viability of the polyplexes prepared at various weight ratios was evaluated by comparison with that of PEI/DNA in HEK293 and HeLa cells by using an MTS assay. As shown in Figure 8, the cell viability of the polyplexes formed from OEI-AAs is indeed a bit lower than OEI, indicating that modification with amino acids would lead to slight increase of cytotoxicity. This might be due to the increase of the molecular weight via the modification [45,46]. However, the cytotoxicities under optimal transfection w/w ratios (16~24) were acceptable. On the other hand, the cytotoxicity of OEI-AAs was much lower than PEI, especially at higher weight ratios. Compared to PEI 25 kDa, the much higher TE together with lower cytotoxicity made the OEI-AAs good candidates for further applications.

Conclusions
In summary, PEI with a relatively low molecular weight (1800 Da, also described as OEI) was modified with three amino acids to study the gene delivery capability of the target materials. The substitution degree of the amino acids could be calculated from the specific peak integrals in the 1 H NMR spectra. It was found that the modification would slightly reduce the DNA binding ability, but the OEI-AAs could condense DNA into more compact nanoparticles. CD results indicated that the indole ring of tryptophan might intercalate into the DNA base pairs, resulting in more efficient condensation. In vitro transfection experiments revealed that the modification of OEI, especially

Conclusions
In summary, PEI with a relatively low molecular weight (1800 Da, also described as OEI) was modified with three amino acids to study the gene delivery capability of the target materials. The substitution degree of the amino acids could be calculated from the specific peak integrals in the 1 H NMR spectra. It was found that the modification would slightly reduce the DNA binding ability, but the OEI-AAs could condense DNA into more compact nanoparticles. CD results indicated that the indole ring of tryptophan might intercalate into the DNA base pairs, resulting in more efficient condensation. In vitro transfection experiments revealed that the modification of OEI, especially with aromatic amino acids, could largely improve the TE, which was even much higher than PEI 25 kDa. Moreover, these OEI-AAs exhibited excellent serum tolerance, which PEI lacks. Up to 137 times higher TE than can be found in PEI was obtained by using OEI-W10 as vector. These materials with high TE together with acceptable cytotoxicity may be promising gene vectors for future applications.