The Reduced-Charge Melittin Analogue MelP5 Improves the Transfection of Non-Viral DNA Nanoparticles

Abstract

Melittin is a 26 amino acid amphiphilic alpha-helical peptide derived from honeybee venom. Prior studies have incorporated melittin into non-viral delivery systems to effect endosomal escape of DNA nanoparticles and improve transfection efficiency. Recent advances have led to the development of two newer melittin analogues, MelP5 and Macrolittin 70, with improved pore formation in lipid bilayers while possessing fewer positive charges relative to natural melittin. Consequently, MelP5 and Macrolittin 70 were conjugated through a disulfide bond to a DNA binding polyacridine peptide. The resulting peptide conjugates were used to prepare DNA nanoparticles to compare their relative endosomolytic potency by transfection of HepG2 cells. Melittin and MelP5 conjugates were equally potent at mediating in vitro gene transfer, whereas PEGylation of DNA nanoparticles revealed improved transfection with MelP5 relative to melittin. The results demonstrate the ability to substitute a potent, reduced charge analogue of melittin to improve overall DNA nanoparticle biocompatibility needed for in vivo testing.

Introduction

Upon interacting with cell surfaces, gene delivery nanoparticles are taken up by either pinocytosis, receptor-mediated endocytosis, or both.^1,2 Endocytosed cargoes, however, are generally trafficked through the endolysosomal degradation pathway.¹ Thus, gene delivery nanoparticles require a membrane lytic function to facilitate endosomal escape prior to reaching the lysosome.

Several peptides used to increase nanoparticle endosomolysis have been derived from nature.^3–5 Of these, melittin, a 26-amino acid amphipathic peptide isolated from honeybee venom, has been most successful as a peptide gene delivery carrier,⁶ in combination with PEI,^7,8 and as multicomponent oligolysine vectors.^9,10 Natural melittin is a positively charged peptide that binds ionically to plasmid DNA to form nanoparticles. However, with only five total Lys and Arg residues, melittin’s binding affinity to DNA is too weak to maintain particle stability in normal saline and thereby these fail to mediate in vitro transfection.⁶ Addition of terminal Cys residues to melittin facilitates tight DNA binding through templated sulfhydryl cross-linking, and these polymelittin/DNA nanoparticles achieve potent in vitro gene transfer.⁶ Likewise, conjugation of melittin to a polyintercalating peptide via a reducible disulfide bond also significantly increases DNA binding affinity, resulting in DNA nanoparticles that achieve in vitro gene transfer equivalent to PEI.¹¹

While melittin DNA nanoparticles have achieved impressive gene transfer efficiency in vitro, they lack in vivo efficacy when dosed i.v. because of the overall particle positive charge, which attracts protein binding, leading to aggregation and aberrant biodistribution to the lung.^12,13 Thus, reduced charge melittin analogues that mediate potent gene transfer while minimizing charge to improve blood compatibility are desirable.

Prior studies by Wimley and coworkers led to the discovery of MelP5, a melittin mutant with three fewer cationic residues that is 5 to 20-fold more potent than melittin at small molecule-sized pore formation in lipid vesicles across a range of bilayer compositions.¹⁴ Furthermore, MelP5 was approximately 10-fold more potent for macromolecule (10 kDa dextran) release from phosphatidylcholine vesicles.¹⁵ Macrolittin 70 (Mac70) was further developed from MelP5, possesses a net zero charge at neutral pH, and is approximately 10-fold more potent than MelP5 at release of 40 kDa dextrans from phosphatidylcholine vesicles.¹⁶ Besides their lower overall charge and increased potency, a key property of MelP5 and Mac70 is the ability to form stable pores in lipid bilayers large enough for macromolecules to cross. These features make both analogues attractive for gene delivery systems, although neither has been tested in this context to date.

The present study advances our previous work on polyacridine (PAcr)-melittin conjugate in vitro gene delivery¹¹, using a previously optimized polyacridine peptide (PAcr) for in vivo gene delivery ^17,18,22 (Fig. 1A). MelP5, Mac70, and melittin PAcr conjugates (Fig 1B) were synthesized, used to prepare DNA nanoparticles, and analyzed for in vitro gene transfer efficiency in HepG2 cells to directly compare the potencies of the endosomolytic peptides. The results demonstrate the ability to incorporate the more potent MelP5 into DNA delivery systems to generate nanoparticles that possess maximal transfection potency and minimized nanoparticle charge.

Figure 1. Structures of Polyacridine Peptide, Melittin Analogues and Polyacridine Peptide-Conjugates.

A) The polyacridine peptide (PAcr)^17,18,22 is composed of L-Lys, L-Lys(Acr) and possesses an N-terminal Cys that is modified with X (where X = melittin analogues, PEG or an alkyl group), and binds and condenses plasmid DNA into stable nanoparticles. B) Illustrates sequences of melittin, MelP5, and Macrolittin 70 (Mac70) analogues used in this study. Melittin analogues were modified by rearrangement of the internal Trp to the N-terminus, addition of a Cys reactive handle to allow conjugation to R (where is R = PAcr), and incorporation of two N-terminal Lys to improve Cys reactivity.¹¹ Basic residues are shown in blue and acidic residues in red. Melittin analogues were conjugated to PAcr through a disulfide bond for reductive release.

Materials and Methods

Fmoc-protected amino acids and N,N’-diisopropylcarbodiimide were obtained from Advanced ChemTech (Lexington, KY). Unsubstituted Wang resin, N-hydroxybenzotriazole (HOBt), and 1-[Bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxide hexafluorophosphate (HATU) were from AAPPTec, LLC (Louisville, KY). Piperidine, N,N-dimethylformamide (DMF), N,N-diisopropylethylamine (DIPEA), acetic anhydride, trifluoroacetic acid (TFA), 1,2-ethanedithiol (EDT), acetonitrile, Sephadex G-10 resin, iodoacetic acid, ammonium acetate, and 25 kDa branched polyethylenimine (PEI) were purchased from Sigma Chemical Co. (St. Louis, MO). Diethyl ether and dichloromethane were from VWR International (Radnor, PA). Glacial acetic acid, 2-propanol, sodium bicarbonate, and 2,2’-dithiodipyridine (DTDP) were obtained from Thermo Fisher Scientific (Pittsburgh, PA). Tris(hydroxymethyl)aminomethane base (Tris) and 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) were from Research Products International Corp. (Mount Prospect, IL). gWiz-Luc, a 6732-bp luciferase-expressing plasmid under the control of a cytomegalovirus (CMV) promoter was obtained from Aldevron (Fargo, ND) and pGL3 control vector, a 5256-bp luciferase-expressing plasmid under the control of the SV40 promoter and enhancer, as well as the ONE-Glo luciferase assay, were from Promega (Madison, WI). Minimal Essential Media (MEM), fetal bovine serum (FBS), and penicillin-streptomycin were obtained from Gibco (Pittsburgh, PA). HepG2 cells were acquired from the American Type Culture Collection.

Synthesis and Characterization of Polyacridine Peptide-Conjugates

Melittin analogues were prepared by solid-phase peptide synthesis (30 μmol) from pre-loaded Wang resin (0.5 mmol/g) with an APEX 396 solid-phase peptide synthesizer (Advanced ChemTech, Lexington, KY) using HATU/HOBt double couplings. C-(Acr-K₄)₃-Acr-K (PAcr), where “Acr” refers to Nε-(9-acridinyl)lysine, was synthesized as described previously.^11,17,18 Peptides were cleaved from the resins and sidechains deprotected with TFA/EDT/water (95:2.5:2.5, v/v/v) for 2-3 hrs at room temperature (RT), followed by filtration and precipitation in ice-cold diethyl ether. Crude peptides were pelleted by centrifugation, lyophilized, and stored frozen in 0.1% (v/v) TFA in water. Crude Mac70 and MelP5 required addition of DMF at 9 and 33% (v/v), respectively, to completely dissolve. Peptides were purified by semi-preparative reversed-phase high-pressure liquid chromatography (RP-HPLC) on an Aeris Peptide XB-C18 column (250 x 21.2 mm, Phenomenex, Torrance, CA) eluted at 10 mL/min with 0.1% TFA and an acetonitrile gradient of 17-23% (PAcr), 35-41% (melittin), or 44-50% (MelP5/Mac70) over 30 min while detecting A_280nm or A_409nm. The collected peptide fractions were pooled from multiple runs, concentrated by rotary evaporation, and lyophilized. Purified peptides were dissolved in 0.1% TFA and quantified by absorbance (Lys(Acr) ε_{409 nm} = 9,266 M⁻¹cm⁻¹ or Trp ε_{280 nm} = 5,600 M⁻¹cm⁻¹).

For conjugation to PAcr through a disulfide bond, the Cys residue on the purified melittin analogues was activated with thiol-pyridine by reacting 0.25 μmol in 0.35 mL 2 M acetic acid/2-propanol (10:3) with 2.5 μmol DTDP for 16 hrs at RT to form the ortho-pyridyl disulfide (OPSS) variants. Melittin-OPSS and Mac70-OPSS reactions were purified to homogeneity on a Sephadex G-10 column (39.5 x 1.5 cm) eluted with 0.1% (v/v) acetic acid/water while monitoring A_{280 nm}. The peptide fractions were pooled, lyophilized, and quantified by absorbance following reconstitution in 0.1% TFA as described above (Trp ε_{280 nm} = 5,600 M⁻¹cm⁻¹ and OPSS ε_{280 nm} = 5,100 M⁻¹cm⁻¹). Sephadex G-10 purification resulted in low recovered yields, prompting purification of crude MelP5-OPSS by RP-HPLC as described above on a 30-75% acetonitrile gradient over 30 min. Purified melittin-OPSS analogues and PAcr peptide were characterized by LC-MS on an Aeris Peptide XB-C18 column (250 x 4.6 mm) (Phenomenex) by injecting 1 nmol onto an Agilent 1100 HPLC eluted at 0.7 mL/min with 0.1% TFA and the same acetonitrile gradients as above while monitoring by ESI-ion trap mass spectrometer (Agilent Technologies, Inc., Santa Clara, CA).

Melittin-OPSS, MelP5-OPSS, and Mac70-OPSS (0.15 μmol in 0.375 mL 2 M acetic acid/2-propanol [10:3]) were each reacted with excess PAcr for 48 hrs at RT. Oxidative dimerization of PAcr was a competing side reaction, leading to the use of 3 eq PAcr for melittin, 8 eq for MelP5, and 5.5 eq for Mac70 to push the reactions to completion. Gene delivery conjugates were purified by RP-HPLC as described above on 30 min acetonitrile gradients of 20-65% for melittin-PAcr and 30-75% for MelP5-PAcr and Mac70-PAcr. Purified conjugates were dissolved in water, quantified by absorbance at 409 nm, and characterized by LC-MS as described above.

Three PAcr conjugates were prepared that lacked endosomolytic peptides. PAcr containing an alkylated Cys residue (Alk-PAcr) was generated by reacting PAcr (0.3 μmol in 0.3 mL 100 mM Tris pH 8.5) with iodoacetic acid (0.9 μmol) for 30 min at RT in the dark followed by quenching with 0.3 mL 0.1% TFA. Alk-PAcr was purified by RP-HPLC on a 15-25% acetonitrile gradient over 30 min, dissolved in water, and characterized by LC-MS as described above. The Cys residue on PAcr was also modified with a disulfide (PEG-SS-PAcr) or non-reducible maleimide (PEG-Mal-PAcr) linkage to a 5 kDa PEG chain as described previously.^18,19

Formulation and Characterization of Peptide-Conjugate/DNA Nanoparticles

DNA nanoparticles were formed by combining peptide-conjugates (4.5 nmol in 12-20 μL water) with 15 μg pGL3 in 300 μL 5 mM HEPES pH 7.4 under rapid vortexing. After a 10 min incubation at RT, the nanoparticle solutions were transferred to a 0.5-mL fluorescence cuvette and the particle diameters determined by dynamic light scattering (DLS) on a ZetaPlus (Brookhaven, Holtsville, NY). Data are shown as the average ± standard deviation of the intensity-weighted, mean lognormal diameters from 10 measurements.

For in vitro transfection experiments, nanoparticles were generated as described above but with 2 μg gWiz-Luc in 40 μL 5 mM HEPES pH 7.4 and 0.6 or 1 nmol peptide-conjugates in 2 μL water, corresponding to 0.3 or 0.5 nmol/μg, respectively. DNA nanoparticles were also generated as above using 0.6 nmols of a 50:50 mixture of melittin-, MelP5-, or Mac70-PAcr with Alk-PAcr, PEG-SS-PAcr, or PEG-Mal-PAcr to condense 2 μg gWiz-Luc. Alternatively, polyethylenimine (PEI)-DNA nanoparticles were prepared by mixing 7.5 μg 25 kDa branched PEI in 5 μL water with 2 μg of gWiz-Luc in 40 μL 5 mM HEPES pH 7.4 while vortexing, corresponding to an N/P ratio of 9.

In vitro Transfection of HepG2 Cells in 384-Well Format

HepG2 cells were maintained in MEM supplemented with 10% FBS and 1% penicillin/streptomycin in a humidified 5% CO₂ incubator at 37°C. Cells were plated in Nunc 384-well tissue culture plates (Thermo Fisher, Pittsburgh, PA) at 5000 cells/well in 50 μL media, allowed to settle for 15 min at RT, and cultured for 24 hrs at 37°C/5% CO₂ prior to transfection.²⁰ Peptide-conjugate and PEI-DNA nanoparticles were added to plated HepG2 cells at 300 ng DNA/well (6.5 μL) and allowed to transfect for 48 hrs at 37°C/5% CO₂. Transfected plates were cooled to RT and analyzed for luciferase expression using the ONE-Glo assay as described previously.²⁰

Statistical Analysis

The log₁₀(x+1) transformation of the transfection results were analyzed for statistical significance by one- or two-way ANOVA with the Holm-Sidak multiple comparisons test (alpha = 0.05) in GraphPad Prism version 9.0.0 (GraphPad Software, San Diego, CA).

Results

Design, Synthesis, and Characterization of Peptide-Conjugate/DNA Nanoparticles

Melittin and two analogues with increased potency towards lipid vesicle pore formation, including first generation MelP5¹⁴ and second generation Macrolittin 70,¹⁶ were tested for their ability to enhance gene transfer to HepG2 cells when conjugated to a DNA-binding gene delivery peptide (Figure 1). To facilitate bioconjugation and promote synthetic tractability, each sequence was modified with a W19L mutation and addition of a CWKK sequence to the N-terminus as described previously (Fig. 1B).¹¹ According to SOPMA analysis, modified sequences were predicted to form similar α-helical structures as the native analogues.²¹ All three analogues were conjugated to a polyacridine peptide (PAcr) previously optimized for DNA binding, compaction, and stability through in vivo structure-activity analysis.^17,18,22 A key design feature was the linkage of PAcr and melittin analogue through a disulfide bond to allow for reductive release and activation of pore forming activity (Fig. 1A).¹¹ This was accomplished by synthesizing the melittin analogues with an ortho-pyridyl disulfide (OPSS)-activated Cys residue for directed disulfide exchange with PAcr (Table 1). RP-HPLC purification of the N-terminal Cys-peptides resulted in dimerization. Thus, all melittin analogues were reduced with TCEP prior to OPSS activation. Following reaction with PAcr and purification, Melittin-PAcr, MelP5-PAcr and Mac70-PAcr conjugates were recovered in > 95% purity as determined by LC-MS with observed masses consistent with the calculated values (Table 1).

Table 1.

Peptide-Conjugate Structures, Characterization, and Synthetic Yields

#	Structure	Abbreviation	Mass (calc/obs, Da)a	Yield (%)
1	C-(Acr-K4)3-Acr-K	PAcr	3006.8/3007.2	30b
2	Alk-C-(Acr-K4)3-Acr-K	Alk-PAcr	3066.2/3065.4	60c
3	OPSS-CWKKGIGAVLKVLTTGLPALISLIKRKRQQ	Melittin-OPSS	3427.0/3427.6	1.5b
4	OPSS-CWKKGIGAVLKVLATGLPALISLIKAAQQL	MelP5-OPSS	3211.9/3212.0	3.9b
5	OPSS-CWKKGIGEVLKELATLLPELQSLIKAAQQL	Mac70-OPSS	3428.9/3428.4	3.6b
6	CWKKGIGAVLKVLTTGLPALISLIKRKRQQ | C(Acr-K4)3-Acr-K	Melittin-PAcr	6322.9/6323.6	24c
7	CWKKGIGAVLKVLATGLPALISLIKAAQQL | C(Acr-K4)3-Acr-K	MelP5-PAcr	6107.7/6109.2	68c
8	CWKKGIGEVLKELATLLPELQSLIKAAQQL | C(Acr-K4)3-Acr-K	Mac70-PAcr	6324.7/6325.2	54c

^aMass obtained by ESI-Ion Trap

^bYield determined by absorbance based on initial resin substitution

^cYield determined by absorbance for last reaction step

Nanoparticles were formed by combining 0.3 nmol peptide-conjugates per μg plasmid DNA.¹¹ Despite markedly different charge and hydrophobicity for Melittin-PAcr, MelP5-PAcr and Mac70-PAcr, the resulting DNA nanoparticles were similar in diameter to PAcr DNA nanoparticles (Figure 2).

Figure 2. Sizes of Peptide-Conjugate/DNA Nanoparticles.

The diameters of nanoparticles formed between pGL3 and polyacridine peptide-conjugates at 0.3 nmol/μg DNA were obtained by DLS. Data are the average mean lognormal diameter ± standard deviation of ten one-minute measurements. Addition of melittin analogues does not change particle sizes by more than 30 nm.

In vitro Gene Transfer of Peptide-Conjugate/DNA Nanoparticles

The transfection efficiency of the Melittin-PAcr, MelP5-PAcr and Mac70-PAcr/DNA nanoparticles in HepG2 cells was investigated using a miniaturized luciferase expression assay in 384-well plates, which demonstrates a dynamic range spanning four orders of magnitude between negative control lacking DNA and polyethylenimine (PEI)/DNA positive control nanoparticles.²⁰ Despite the lack of an endosomal escape element, PAcr DNA nanoparticles resulted in gene expression at approximately 100-fold over background; however, this was 100-fold less than for PEI DNA nanoparticles (Figure 3). Alternatively, conjugation of melittin and MelP5 to PAcr enhanced the luciferase expression to similar levels as PEI, whereas Mac70 incorporation only marginally improved transfection 3-fold relative to PAcr alone (Figure 3). Increasing the peptide-conjugate to DNA stoichiometry from 0.3 to 0.5 nmol/μg did not improve transfection for melittin- and MelP5-PAcr and decreased the transfection of Mac70-PAcr by 5-fold (Figure 3). Surprisingly the melittin- and MelP5-PAcr conjugates displayed similar transfection efficiencies despite having different potencies towards lipid vesicles. Thus, we were curious how 50% dilution with alkylated PAcr prior to DNA formation would affect transfection efficiency. This strategy would allow formulation of nanoparticles using PAcr containing a 5 kDa PEG, to increase biocompatibility and circulatory stability following IV dosing.¹⁹

Figure 3. Effect of Melittin Analogues on in vitro Gene Transfer of Polyacridine Peptide-Conjugates.

HepG2 cells were transfected in 384-well plates with 300 ng of DNA nanoparticles. Peptide-conjugates were added at 0.3 nmol/μg gWiz-Luc for the open bars and 0.5 nmol/μg gWiz-Luc for the checkered bars. At 48 hrs post-transfection, luciferase expression was quantified as relative light units (RLU) following the addition of ONE-Glo. Data are the average and standard deviation of six replicates. ** p < 0.01, **** p < 0.0001 relative to PAcr.

As shown in Figure 4, although melittin- and MelP5-PAcr were not significantly different following 50% dilution with Alk-PAcr, the transfection efficiencies were reduced by 2.5-fold for melittin and increased by 1.4-fold for MelP5 compared to no dilution. The differences were more pronounced under dilution with PEGylated PAcr conjugates, however. A DNA nanoparticle composed of reducible PEG-SS-PAcr, MelP5-PAcr was 17-fold more potent than melittin-PAcr (Figure 4). Overall transfection efficiencies were markedly reduced when the non-reducible PEG-Mal-PAcr was substituted into the formulation, but still produced a non-significant 5.5-fold enhancement for MelP5 (Figure 4). Thus, in addition to the lower overall charge, MelP5 is less resistant to decreases in transfection efficiencies upon incorporation in disulfide-PEGylated DNA nanoparticles.

Figure 4. In vitro Transfection of Peptide-Conjugate/DNA Nanoparticles Containing 50% Melittin Analogue.

HepG2 cells were transfected with 300 ng of DNA nanoparticles prepared with Melittin-PAcr (green) or MelP5-PAcr (orange), combined with 50 mol% of Alk-PAcr, PEG-SS-PAcr, or PEG-Mal-PAcr. Peptide-conjugates were added at 0.3 total nmol/μg gWiz-Luc. At 48 hrs post-transfection, luciferase expression was determined as relative light units (RLU) following the addition of ONE-Glo. Data are the average and standard deviation of six replicates. *** p < 0.001.

Discussion

Despite melittin’s success in enhancing nanoparticle endosomal escape, the peptide is relatively unoptimized for gene delivery use, and several groups have endeavored to improve its biophysical properties or activity. A major focus has been tuning melittins’ broad, pH-independent activity to improve pore formation at endosomal pH or decrease membrane permeabilization at physiological pH to limit non-specific hemolysis and cytotoxicity. Boeckle et al. employed the former approach by replacing the neutral C-terminal glutamine residues with glutamic acid and found that while hemolytic activity of melittin analogue-PEI conjugates at physiological pH were retained, lytic activity and in vitro transfection efficiencies were improved at pH 5.²³ Alternatively, masking of melittin’s primary amines with a pH-labile blocking group mitigates non-specific toxicity while still allowing for activation within endosomes.^24–26 Despite melittin’s apparent toxicity in some gene delivery systems, our lab has demonstrated markedly reduced hemolysis following incorporation of melittin into stable peptide/DNA nanoparticles, reducing the requirement for analogues with narrow pH activity windows.^6,11

Arguably, a more important consideration for gene delivery is that melittin’s pore forming activity may be suboptimal for endosomal release of nanosized cargos. Melittin is proposed to bind to physiologically-relevant zwitterionic membranes in both inactive parallel and active perpendicular orientations, with limited interconversion between the two orientations.²⁷ Thus, the inactive orientation may actually compete with the active orientation and prevent pore formation.²⁷ Above a threshold peptide/lipid (P:L) ratio, the perpendicularly orientated melittins can form toroidal transmembrane pores with diameters of 2.5 – 4.5 nm.^28–31 This pore size range is too small to allow passage of most gene delivery complexes or nanoparticles. Therefore, elevated P:L ratios are likely required to overcome competition with parallelly-oriented melittin and promote detergent-like permeabilization and lysis of the endosomal membrane.³² While achievable in cell culture transfections due to the high gene carrier-to-cell ratio, in vivo success may require melittin analogues with increased potency and lower P:L ratios to achieve endosomal escape. Furthermore, the high cationic character of melittin hinders its application to in vivo gene delivery systems since near-neutral nanoparticle zeta potentials are preferred to limit scavenging by the liver reticuloendothelial system.^19,33

The Wimley and Hristova groups have conducted several high-throughput synthetic molecular evolution studies around melittin to generate more potent analogues and reveal critical amino acid substitutions that improve pore formation.^{14–16,34,35} Of particular relevance to gene therapy are the first generation peptide MelP5¹⁴ and second generation Macrolittin 70 (Mac70),¹⁶ which was designed based off of MelP5. In contrast to the transient pores of melittin,^36,37 MelP5 and Mac70 form equilibrium pores that allow the release of 10- and 40-kDa dextrans from synthetic phosphatidylcholine vesicles, although the latter is approximately 10-fold more potent for 40-kDa dextran release.¹⁶ The increase in potency of MelP5 relative to melittin has been attributed to formation of a longer and more ideal amphipathic α-helix, increasing the proportion of the membrane-inserted (perpendicular/active) orientation and pore stability.^14,15,37,38 The sequence of Mac70 revealed the importance of acidic residues clustered around K7, which may be involved in interpeptide interactions within membranes, allowing larger pore formation while retaining the amphipathicity of MelP5.¹⁶ In addition to the enhanced potencies, both peptides are less cationic than melittin at physiological pH, with three fewer positive charges for MelP5 and six fewer for Mac70.

Based on these properties, all three analogues were incorporated into synthetic PAcr gene delivery conjugates for comparison. The similar nanoparticle sizes formed between plasmid DNA and the conjugates indicate that the PAcr peptide is the main driver for DNA compaction and thereby discounts a size effect on the in vitro transfection levels. Similarly to Baumhover et al., melittin-PAcr carriers had nearly equivalent potency for HepG2 transfection as PEI at N/P 9, which had 100-fold higher luciferase expression than PAcr alone (Figure 3).¹¹ Surprisingly, MelP5 and Mac70 were similar to or less potent than melittin, respectively, contrasting with the expected effects of the native analogues on synthetic liposomes.^14,16 This may indicate a maximal in vitro HepG2 transfection benefit that is similar with both melittin and MelP5. This is supported by the observation that increased carrier-to-DNA stoichiometries did not improve transfection (Figure 3). The lower transfection with Mac70 is not unexpected given that the macrolittin family of peptides has recently been shown to have very little effect on human cells.³⁹

Due to the similarity in the transfection potentials for melittin and MelP5, we reasoned that both peptides were present at such high P:L ratios in the endosome that the membrane was permeabilized to similar degrees, resulting in plateaued transfection efficiencies. To test this, melittin-PAcr and MelP5-PAcr were combined with 50 mol% PAcr or PEGylated PAcr prior to forming nanoparticles. This revealed that while melittin and MelP5 have similar maximal transfection efficacies, MelP5 could better maintain transfection levels despite dilution with a variety of PAcr derivatives. Additionally, nanoparticles formed with peptide combinations highlighted the importance of a reducible PEG-PAcr linkage in enhancing luciferase expression by two orders of magnitude compared to a non-reducible maleimide linkage (Figure 4).

In conclusion, this report represents the first inclusion and testing of the synthetic peptides MelP5 and Macrolittin 70 in a gene delivery system. MelP5 harbors several features that make it a more attractive candidate for gene therapy as an endosomal escape agent compared to melittin. First, MelP5 is more potent than melittin and better able to overcome the negative impact of PEGylation on transfection. Secondly, the lower cationic charge of MelP5 may be more beneficial for in vivo biodistribution and activity, since these gene delivery systems rely on the careful control of nanoparticle surface groups and charge or zeta potential.^18,19,33 Addition of melittin to a gene delivery system may compromise a nanoparticle’s biophysical properties and result in aberrant biodistribution. Finally, this study identifies reducible PEGylation as a key parameter for effective gene delivery with polyacridine peptides.