Charge modulation of peptide/nucleic acid complexes: an anionic additive enhances gene silencing and CRISPR/Cas9 editing by promoting intracellular nucleic acid release

Small interfering RNA (siRNA) and Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR)/CRISPR-associated Protein 9 (Cas9) complexes are effective approaches to temporarily downregulate protein expression via post-transcription RNA interference or permanently altering protein expression via editing genomic DNA, respectively. However, the efficient delivery of these mediators to targeted cells has been challenging, largely due to their anionic and hydrophilic nature, which hinders their interaction with the cell membrane and cellular internalization. Cell-penetrating peptides (CPPs) exhibit dual characteristics as a carrier for nucleic acid delivery, where positively charged components bind to the negatively charged nucleic acid, and the hydrophobic components enhance interactions with the cell membranes. Arginine/tryptophan-containing peptides provide both characteristics and have been shown to be efficient in delivering nucleic acids. In this project, we evaluated a library of novel linear and hybrid (cyclic/linear) peptides with increasing molecular weight for their physical characteristics, interaction with nucleic acids, cytotoxicity, internalization into different breast cancer cell lines, and their ability to promote and enhance silencing efficiency. The majority of the peptides included in this study demonstrated strong binding affinity to siRNA; however, despite effective cellular internalization, they did not show expected silencing/transfection efficiency, which we hypothesized to be due to the strong bonding between the carrier and cargo and, therefore, poor intracellular release of delivered nucleic acids. The addition of a negatively charged component to the peptide/siRNA and peptide/ribonucleoprotein complexes enhanced the efficiency of the delivered nucleic acid. These data emphasize the importance of the balance between stability of the carrier/cargo complex and timely intracellular release in the efficiency of this approach.

Despite the impressive efficiency of various nucleic acids (including plasmid DNA, messenger RNA or mRNA, small interfering RNA or siRNA, and short guide RNA (sgRNA)/clustered regularly interspaced short palindromic repeat or CRISPR/the associated protein nine or Cas9) in modifying cell function, efficiently delivering these molecules into the targeted cells has been challenging, to say the least. While in vivo delivery of nucleic acids has obviously been a daunting task due to rapid enzymatic degradation, clearance by immune cells, and removal by the liver and kidneys (Gonzalez et al., 2025), even in vitro delivery of the nucleic acids could be a formidable challenge. Internalization into the cell is hindered by limited interaction between nucleic acids and the cell membrane and endosomal entrapment is an ever-present threat (plus the extra step of intranuclear delivery for DNA and CRISPR/Cas9) (Aliabadi et al., 2012). This explains the wide array of carriers that have been studied for nucleic acid delivery in the last 25 years, which range from viral carriers to exosomes, liposomes, micelles, polymers, peptides, gold nanoparticles, and lipid nanoparticles (LNPs; arguably considered the most efficient carrier to date), to name a few (Bian et al., 2025).

Many studies have been conducted to find correlations between the performance of nucleic acid carriers and the structural characteristics of the nanoparticles, including size, zeta potential (indicating the overall electrical charge of the nucleic acid/cargo), and binding affinity. Among the characteristics of an ideal nucleic acid delivery system is a positive charge that would bind to nucleic acids via interionic interaction for a strong cargo/carrier association. Even in LNPs, the presence of an ionizable lipid (that is positively charged at acidic pH) is considered crucial (Schlich et al., 2021). However, a strong binding to nucleic acids, while important for protection of the cargo and stability of the cargo/carrier association before cellular internalization, could negatively affect silencing efficiency due to delayed release of nucleic acid in the cell (Han et al., 2013), which might result in the degradation of nucleic acid molecules before they can exert their effect inside the cell. However, there are contradictory reports indicating that “excessively strong siRNA binding” could, in fact, result in an elevated silencing efficiency (Yang et al., 2017).

We have previously reported different peptide structures for small-molecule and siRNA delivery (Mandal et al., 2023; Mandal et al., 2022; Zoghebi et al., 2022; Mozaffari et al., 2021; Mozaffari et al., 2019; Do et al., 2017) and as part of a multicomponent delivery system for CRISPR/Cas nine delivery (Alhazza et al., 2024). The present study reports an attempt to deliver siRNA and CRISPR/Cas9 to triple-negative breast cancer cell lines and HEK 293 cell line, respectively, using two series of linear and hybrid cyclic/linear peptides. The synthesis and characterization of the linear and hybrid cyclic/linear peptides have been previously reported (Khayyatnejad Shoushtari et al., 2021). The peptides contain a constant five alternative tryptophan (W) and arginine (R) residues (either in linear (RW) 5 or cyclic (RW) 5 form) attached to an increasing number of (RW) n linear section (n = 1–5) via a lysine (K) linker. Positively charged amino acids (e.g., R) have been used in carrier peptides for interionic interaction with nucleic acids (Yu et al., 2020), while hydrophobic amino acids (e.g., W) can enhance the interaction of the carrier with the lipid bilayer of the cell membrane and therefore, the cellular internalization. The hypothesis was that the 10-peptide series (five linear and five cyclic/linear hybrid peptides; L1–L5 and C1–C5, respectively; Figure 1) would offer structure/performance information on the effect of (i) linear vs. cyclic structure of the constant portion of the peptides, and (ii) increasing number of additional RW units in each series.

While some of the peptides included in the study demonstrated promising internalization into the selected cell lines, this achievement did not translate into effective silencing of the selected targets, which we hypothesized was due to the strong binding affinity of some peptides and siRNA observed in the serum stability experiments. Therefore, we added a negatively charged component to the siRNA/peptide complexes to enhance siRNA release in the cytoplasm and therefore boost the silencing efficiency. Dr. Uludag’s laboratory has commercialized a negatively charged transfection aid (additive) called Trans-Booster, which was selected for this purpose. This study demonstrates the need for case-by-case modification of the delivery system based on nanoparticles’ components and characteristics.

Materials

Life Technologies (Grand Island, NY, United States) provided SYBR Green II, Lipofectamine™ 2000, CRISPIRMAX, and Silencer® Negative Control siRNA (Catalog number: AM4635). Western blot material, including 4%–20% Mini-PROTEAN® TGX Stain-Free TM Protein Gels, Mini-PROTEAN® Tetra Electrode Assembly, Clarity™ Western ECL Substrate, Trans-blot® Turbo™ Cassettes, Trans-Blot® Turbo™ Mini PVDF Transfer Packs, Precision Plus Protein Western C Standards, 10X Tris/Tricine/SDS, and Bio-Safe Coomassie G-250 Stain were purchased from Bio-Rad (Hercules, CA, United States). Negative control siRNA labeled with Alexa Fluor 488 (AF488; Catalog number 1027292), Gel Pilot Loading dye (Lot # 166049455), DNeasy Blood and Tissue kit, Buffer AL Lysis buffer, Buffer AW1, Buffer AW2, AE buffer, and proteinase K were supplied by Qiagen (Valencia, CA, United States). The siRNA targeting STAT3 was purchased from (catalog no. SI02662338) was purchased from Thermo Fisher. Trans-Booster was supplied by RJH Biosciences (Edmonton, Alberta, Canada). Cell Signaling Inc (Danvers, MA, United States) provided the STAT3 (124H6) Mouse mAb (Catalog number 9139 S), anti-rabbit IgG, HRP-linked secondary antibody (7074 S), and human beta-actin rabbit monoclonal antibody (Catalog number 4967 S). Green Fluorescence Protein (GFP)-siRNA (sense sequence: 5′-GAA​CUU​CAG​GGU​CAG​CUU​GCC​G-3′), gene Ruler 1 kb Plus DNA Ladder (Lot# 00899446), Alt-R TracrRNA 5′ ATTO 550 (Cat# 1077024), Alt-R negative control crRNA (Cat# 1079138), Alt-R CRISPR-Cas9 HPRT (Cat# 1079132), Alt-R TracrRNA (Cat# 1073189), and Alt-R S. p. HiFi Cas9 Nuclease V3 (Cat# 1081061) were provided by Integrated DNA Technologies (Coralville, IA, United States). R&D Systems (Minneapolis, MN, United States) provided anti-mouse polyclonal secondary antibody (Goat Anti-Mouse IgG HRP-linked; Catalog number HAF007). Biotool LLC (Houston, TX, United States) supplied Cell Counting 8 (CCK8) KIT. Disposable capillary cells and cuvettes used with the Zetasizer were purchased from Malvern Panalytical (Westborough, MA). All other reagents were supplied by VWR (Radnor, PA, United States).

Cell lines

The human triple-negative breast cancer cell lines, MDA-MB-231 (MDA 231; ATCC® HTB-26™), MDA-MB-468 (MDA 468; ATCC® HTB-132™), and AU 565 (HER2-positive; ATCC® CRL-2351™) were purchased from American Type Culture Collection (ATCC; Manassas, VA). MDA 231 and MDA 468 cell lines were cultured and maintained in DMEM low glucose medium, while AU 565 was cultured and maintained in RPMI-1640 medium. Both media were supplemented with 10% (v/v) FBS and 100 U/mL penicillin/100 μg/mL streptomycin. All 3 cell lines were maintained at 37 °C, controlled humidity, and 5% CO 2, and were expanded to a new flask when they reached 80%–90% confluency. Cells were no longer used after 40 passages and were replaced with a new batch of frozen cells that were stored in liquid nitrogen and thawed for the replacement.

Methods

Peptide synthesis and purification

The synthesis and purification of cyclic/linear peptides have been reported previously (Khayyatnejad Shoushtari et al., 2021). Briefly, Fmoc solid-phase peptide synthesis followed by solution-phase cyclization was utilized to synthesize hybrid cyclic/linear peptides. The linear peptides were purchased from Synpeptide Co. (Shanghai, China). A scheme representing the peptide synthesis process is included in Supplementary Figure S1.

Binding affinity

To investigate the spontaneous ionic interaction, peptides and scrambled siRNA were mixed in normal saline at 14 different nitrogen/phosphate (N/P) ratios as follows: 0.05, 01, 0.2, 0.5, 1, 2, 3, 4, 5, 6, 8, 10, 20, 40. The wide range of the N/P ratios is selected to identify the ratio resulting in full binding as well as the ratio required for 50% binding. The mixtures of peptide/siRNA in saline were incubated at room temperature for 30 min to allow complete complex formation. The N/P ratios were calculated using Equation 1:

Samples were prepared in separate tubes. The final concentration of siRNA in all samples was 892.5 nM. Peptides were solubilized in RNase-free water, and the total volume of each sample was 159 μL. Samples were transferred into the wells in BD FluoroBlok MW Insert 96-well plate (in triplicate). Then, 200 μL of diluted SYBR Green Dye II (dilution of 1:10,000 in double-distilled water) was added to all wells. The fluorescent signal was measured using SpectraMax M5 UV-VIS plate reader (λ excitation: 485 nm, λ emission: 527 nm). The percentage of siRNA bound to peptide molecules was calculated based on Equation 2 and plotted against the N/P ratio to calculate the ratio resulting in 50% binding (BR50) based on the linear portion of the graph:

Hydrodynamic diameter and the surface charge

The size and electrical charge of nucleic acid carriers play important roles in their efficiency and their interaction with biological barriers. We used an N/P ratio of 10 to measure the hydrodynamic diameter, and 4 N/P ratios (5, 10, 20, 40) to measure the zeta-potential of siRNA/peptides complexes. The N/P ratio of 10 was selected for hydrodynamic measurements to quantify and compare the sizes of peptide/siRNA complexes at full complex formation; however, four ratios were used for the zeta-potential measurements to illustrate the changes in electrical charge that indicate the effect of an increase in the proportion of the positively charged component. The samples of 183 μL were prepared by mixing scrambled siRNA and peptides in normal saline and were incubated for 30 min in room temperature to ensure complete complex formation. The samples were then transferred to disposable cuvettes. Malvern Nano ZS Zetasizer (Westborough, MA) was used to measure the size and polydispersity. For zeta-potential, we added 540 μL of RNase-free water to a disposable cuvette and transferred the solution into a disposable folded capillary cell, and measured zeta-potential using Malvern Nano ZS Zetasizer.

Serum stability

Protecting nucleic acids against rapid enzymatic degradation by nucleases present in biological environments is one of the main tasks of the delivery system. To determine the efficiency of the peptides in protecting the scrambled siRNA exposed to serum, we selected 5 N/P ratios (4, 6, 10, 20, 40) to determine the ideal peptide/siRNA ratio in maximizing the serum stability of siRNA in the complexes formed with peptides. Additionally, two samples without peptide were used as negative (free siRNA exposed to HBSS) and positive (free siRNA exposed to fetal bovine serum or FBS) controls. Samples were prepared by mixing siRNA and peptides in normal saline to a final volume of 75 μL. After 30 min of incubation at room temperature, 25 μL of HBSS was added to the negative control sample, while the rest of the samples received 25 μL of FBS, and all samples were incubated at 37 °C for 24 h. After the incubation time, 3 μL of 0.5 mM EDTA and 2 μL of 7% heparin solution were added to each sample with 5 min incubation time at room temperature. Then, 5 μL of loading dye was added to each sample, and 35 μL of each sample was loaded into 1% agarose gel (with 1 μg/mL ethidium bromide) at 100 V for 30 min. Ultraviolet illumination (Gel-Doc system, Bio-Rad; Hercules, CA) was used to visualize the gel and the intensity of the bands.

Cytotoxicity

All three human breast cancer cell lines were used in this study to investigate the potential toxicity of the complexes. Cells were seeded in 96-well plates (15,000 cells per well or 150,000 cells per mL) and incubated at 37 °C and 5% CO 2 for 24 h. The peptide/siRNA samples were formed using scrambled siRNA at N/P ratios of 10, 20, 40, and 80. Again, free siRNA and siRNA delivered via Lipofectamine™ 2000 were used as negative control and a point of comparison (Lipofectamine™ is known to be toxic to varying degrees across different cell types), respectively. After 24 h of incubation, cells were exposed to peptide/siRNA complexes and incubated for an additional 48 h. After the incubation period, 10 μL of CCK reagent was added to each well, and the plates were incubated for 60 min at 37 °C and 5% CO 2. A SpectraMax M5 UV-Vis Plate Reader was used to measure absorbance at 450 nm.

Cellular internalization of siRNA

The efficiency of peptides in cellular internalization was evaluated in all three selected human breast cancer cell lines. The cells were seeded in 24-well plates at a confluency of ∼ 40% and were exposed to peptide/siRNA complexes prepared at N/P ratios of 20 and 40 using Alexa Fluor 488-labeled siRNA. Free siRNA and siRNA delivered by Lipofectamine™ 2000 were used as negative and positive controls, respectively. The siRNA complexes were added to the seeded cells after the initial incubation time, and the cells were incubated for 24 h at 37 °C and 5% CO 2. After the incubation period, the media was removed, and the cells were washed with HBSS. The cells were then detached by adding 100 μL of 0.05% trypsin. After confirming complete detachment, 300 μL of 3.7% formaldehyde was added to the suspended cells to neutralize the trypsin and to fix the cells. Suspended cells of each well were then transferred to capped tubes for flow cytometry measurements. FACSVERSE Flow cytometry (BD-FACSVerse, BD Biosciences; San Jose, CA) was used to measure fluorescent signal. Calibration was performed by gating on the cell population treated with saline (i.e., “No Treatment”), so that the auto-fluorescent cells represented ∼1% of the cell population. The percentage of cells with the fluorescence signal and the mean fluorescence in the cell population were determined.

Confocal microscopy was used to trace and visualize the fluorescent-labeled siRNA in the cells. Treatment groups were selected based on cellular uptake results and the most efficient N/P ratio for each peptide. Alexa Fluor 488-labeled siRNA was used to visualize siRNA localization by confocal microscopy. Sterilized cover slips were placed at the bottom of 6-well plates and incubated overnight in DMEM low-glucose media at 37 °C to enhance cell adherence to the cover slips. Media was removed the next day, and MDA-MB-231 cells were seeded in the 6-well plates (on the cover slips) and were incubated overnight at 37 °C and 5% CO 2. After the incubation period, free siRNA or siRNA complexes were added to the wells, and cells were incubated again for 24 h. Lipofectamine/siRNA was used as positive control. Cells were washed with PBS three times after the incubation period, then fixed using 3.7% formaldehyde in 1× PBS for 10 min on cover slips. Cells were rinsed three times with PBS for 5 min. Cell membranes were stained by Texas Red Phalloidin solution (40 μL and 10 mg BSA in 10 mL of 1× PBS) and incubated for 1 h at room temperature. Cells were washed with PBS three times for 5 min. One drop of DAPI was added to the slide to stain the nuclei, after which the coverslips were stored facing down and away from light overnight. Imaging was performed using Nikon A1R high-definition resonant scanning confocal microscope and a NIS-Elements software (AR 4.30.02, 64bit).

Protein silencing by siRNA

For this study, signal transducer and activator of transcription 3 (STAT3) was targeted as a model protein to evaluate the silencing efficiency of peptide/siRNA complexes. MDA-MB-231 cells were seeded in 6-well plates with ∼40% confluency 24 h prior to the silencing experiments. The study groups were no treatment (NT), scrambled siRNA/L1 complexes, and siRNA targeting STAT3 delivered by Lipofectamine™ 2000 or L1 (with or without Trans-Booster). Cells were then incubated for 48 h at 37 °C and 5% CO2. Afterward, cells were washed with 1 mL HBSS and then detached by 0.05% trypsin. Detached cells were suspended in 2 mL of low-glucose DMEM and centrifuged at 600 rpm for 5 min. The pelleted cells were suspended in 1 mL PBS and centrifuged again at 3,400 rpm for 5 min. The pelleted cells were exposed to 1× RIPA buffer containing 10 μL of protease and phosphatase inhibitor for each 1 mL of buffer. All tubes were kept on ice and were intermittently sonicated for 2 min. The tubes were then kept on ice for 1 h and vortexed every 20 min. Then, all samples were centrifuged at 15,000 RPM for 15 min at 4 °C. Supernatant containing the extracted protein was collected in a new 1.5 mL Eppendorf tube, and the pelleted cell debris was discarded. After collecting the protein, the total protein extracted was quantified using a BSA assay. Briefly, 3 μL of the protein sample was diluted in 57 μL of PBS, and a standard curve was obtained using different albumin concentrations (20–2000 μg/mL). 200 μL of the reagent mixture (50:1 Reagent A: Reagent B) was added to each sample. Spectrophotometer was used to analyze the concentration at 562 nm.

For Western blotting, 25 μL dye (2× Laemmli buffer and 2-mercaptoethanol was added to each sample and vortexed. Samples were centrifuged at 13,000 RPM for 1 min and then incubated at 95 °C for 5 min. Samples were then cooled and centrifuged again at 13,000 RPM for 1 minute and the protein fraction was collected at the bottom of the tubes. The samples were loaded onto a Mini-Protean TGX 10% and ran at 100 V and 75 mA for 1 h. The protein was then transferred to a Trans-Blot® Turbo™ Mini PVDF membrane (Catalog No. 1704156). The membrane was washed three times in TBS for 5 min. Then, the membrane was immersed in 5% BSA in TBST for 1 h at room temperature for blocking. Then, the membrane was soaked in primary antibody (1:1000 in 3% BSA in TBST) overnight at 4 °C. The next day, primary antibody was removed and TBST was used to wash the membrane for 5 min four times. Diluted secondary antibody was added to the membrane for 1 h at room temperature. The membrane was then washed by TBST four times for 5 min. After that, the ECL kit was prepared to detect the blots and the ChemiDoc imager (Bio-Rad) was used. Beta-actin was visualized as an endogenous gene using the same approach.

We also studied the silencing efficiency in an MDA-231 cell line that was transfected with Green Fluorescence Protein (GFP) in Dr. Uludag’s laboratories. Naïve and GFP-positive cells were seeded in a 6-well plate (three wells for each study group; approximately 300,000 cells per mL). After 24 h, cells were exposed to normal saline, or scrambled or GFP-targeting siRNA delivered by Lipofectamine™, peptides, or peptide/Trans-Booster. The remaining GFP signal was calculated based on the remaining expression using the following equation:

The equation normalizes the change in GFP signal based on any potential changes in the fluorescence signal that could be due to off-target effect or toxicity.

Cellular internalization of CRISPR

The ribonucleoproteins (RNPs) preparation has been previously reported (Alhazza et al., 2024). Briefly, the annealing of crRNA and tracrRNA was performed in a nuclear-free Duplex Buffer at 95 °C for 5 min. The duplexes were cooled down, and Cas9 enzyme was added to the duplex solution (1:1 M ratio, Cas9:RNA) and incubated for 5 min at room temperature. RNPs were prepared fresh and kept on ice until incorporated into nanoparticles. The efficiency of peptides in cellular internalization of RNPs was evaluated in HEK-293 cell line. The cells were seeded in 24-well plates with confluency of ∼ 40% and were exposed to peptide/RNP complexes prepared at 0.5:1 1:1 and 2:1 ratio using ATTO 550 fluorescent labeled crRNA. Free RNP and RNP delivered by CRISPRMAX® were used as negative and positive control, respectively. The RNP complexes were added to the seeded cells after 24 h, and the cells were incubated for 24 h at 37 °C and 5% CO 2. After the incubation period, the media was removed, and the cells were washed with HBSS. The cells were then detached by adding 100 μL of 0.05% trypsin, and 300 μL of 3.7% formaldehyde was added to the suspended cells to neutralize the trypsin and to fix the cells. Suspended cells of each well were then transferred to capped tubes for flow cytometry measurements. Calibration was performed by gating on the cell population treated with saline (i.e., “No Treatment”), so that the auto-fluorescent cells represented ∼1% of the cell population. The percentage of cells with ATTO-550 fluorescence and the mean fluorescence in the cell population were determined. Confocal microscopy was also utilized to visualize and track the internalized RNPs. Please see the method explained in the “Cellular Internalization of siRNA” section.

Transfection with ribonucleoprotein (RNP)

To assess transfection efficiency, we targeted the housekeeping gene, HPRT. Treatment groups were selected based on cellular uptake results and the most efficient concentration of each peptide. RNP/peptide complexes were added to 6-well plates and then were seeded with HEK-293 cells at high confluency for reverse transfection. For the positive control group, Opti-MEM media (Corning, Cat. No. 40–300-CV) were used to prepare CRISPRMAX/RNP complexes. After 48 h, cells of the treated groups were transferred to flasks to enhance the cell population. After cells reached ∼90% confluency, the DNA of cells included in each study group was extracted using DNeasy kit. Cells were pelleted after centrifuging the cell suspension and were mixed with 200 uL of PBS and 20 uL proteinase K. AL buffer (200 uL) was then added to each sample and incubated at 56 °C for 10 min. After the incubation period, 200 uL of ethanol was added to each sample, and samples were transferred into a DNeasy mini spin column placed into 2 mL collection tubes. Samples were centrifuged at 10,000 RPM for 1 minute. Spin columns were transferred to new collecting tubes, and 500 uL of Buffer AW1 (stringent wash with low concentration of quinidine used to denature the protein) was added to each column and centrifuged at 10,000 RPM for 1 minute. Spin columns were then transferred to new collecting tubes, and 500 uL of Buffer AW2 was added to each column and were centrifuged at 13,000 RPM for 3 min. Spin columns were transferred to new 1.5 mL microcentrifuge tubes. DNA was eluted by adding 50 uL of AE (10 mM Tris-Cl. 0.5 mM EDTA; pH 9.0) buffer and incubating the sample for 1 minute at room temperature and then centrifuge samples at 10,000 RPM for 1 minute. DNA concentration was measured using Nanodrop©. DNA of each sample was diluted to 1:10 ratio and those diluted DNA samples were used to run PCR by mixing each sample with Ready Mix, reverse and forward primers, and nuclease-free water. After PCR, half of the sample, 10 uL, was mixed with two uL of T7E1 reaction buffer and nuclease-free water. After PCR, half of the sample, 10 uL, was mixed with two uL of T7E1 reaction buffer and nuclease-free water. The mixed samples were run into the CRISR ramp cycle for 15 min Two uL of T7 endonuclease was added to each sample, and the samples were incubated for 60 min. For gel electrophoresis, PCR samples were mixed with PBS to make an equal concentration with T7 endonuclease reaction samples. Two uL of DNA loading dye was added to all samples. Samples were loaded into 1.5% agarose gel, and electrophoresis was run at 80 V for 70 min. Ultraviolet illumination (Gel-Doc system, Bio-Rad; Hercules, CA) was used to visualize the gel and to reveal the intensity of the bands.

Binding affinity

A strong affinity to bind to the nucleic acid cargo can be a desirable characteristic for a delivery system due to enhanced encapsulation efficiency and protection against early degradation. However, if bound too strong, it can interfere with the timely release of the cargo at the site of action. To quantify and compare the binding affinity of the 10 peptides included in the study, we plotted the N/P ratio versus the percent of bound siRNA to calculate N/P ratio required for 50% binding (BR50), and the results are summarized in Figures 2A,B. It was immediately clear that as the number of arginine residues increased, the BR50 value decreased. This was expected as arginine would be the siRNA-binding amino acid in the RW combination. While the calculated BR50 was generally lower for the hybrid peptides than for their linear counterparts, the difference was only statistically significant for L2 vs. C2 and L3 vs. C3. This is also expected as reports indicate that cyclic peptides show a higher affinity to bind to the cargo (Hayes et al., 2021; Pal and t Hart, 2022), Overall, the calculated BR50s were between 1.21 and 2.08, which indicates a strong binding affinity compared to other peptides we have studied as siRNA carriers, which mostly represented BR50s > 2.7 (Mandal et al., 2022; Mozaffari et al., 2019), and even in some cases never reached 50% binding (Do et al., 2017).

This strong binding affinity, combined with the relatively low silencing efficiency results will be presented later, prompted the idea of adding an additional component to the peptide/siRNA complexes to weaken binding between the cargo and peptides. We have previously reported that negatively charged additives enhance siRNA silencing efficiency across multiple settings (Parmar et al., 2018; Santadkha et al., 2022; Sundaram et al., 2022). We added Trans-Booster to the peptide and siRNA mixture (1:2 siRNA/Trans-Booster w/w ratio) before the incubation time at room temperature. Figure 2C demonstrates the effect of adding Trans-Booster to the polymer/siRNA complexes on the binding affinity of three of the linear peptides. As expected, adding another negatively charged component reduced binding affinity to siRNA, with the effect being statistically significant for all three linear peptides included in the experiment.

Size and zeta-Potential

The size of the nanoparticles can significantly affect cellular internalization and biodistribution, when they are administered in vivo. While the particle size plays a crucial role in the efficiency of a delivery system (Xu et al., 2023), the ideal size depends largely on the targeted tissue and the purpose of the delivery system. In cancer therapy, and in order to take advantage of Enhanced Permeation and Retention (EPR) effect, a 50–100 nm range seems to be an appropriate size, with successful use of larger sizes up to 200 nm being reported (Xu et al., 2023; Leporatti, 2022; Tang et al., 2014). The size of the peptide/siRNA complexes formed at an N/P ratio of 10 is summarized in Figure 3A. A significant variety in size was not observed among the peptides included in this study, and the average sizes were in the range of 98–115 nm.

The change in the average zeta-potential of the complexes with varying N/P ratio followed the expected pattern, with the zeta-potential increasing gradually as the N/P ratio increased. The average zeta-potential of the complexes with an N/P ratio of 10 or higher was positive for all complexes (Figure 3B). Since peptides contain positively charged arginine, an increase in the proportion of peptide in the complexes is expected to shift the surface charge to positive values., which has been reported in our previous reports as well (Do et al., 2017; Aliabadi et al., 2011). While a distinctive trait was not observed comparing the linear and cyclic peptides, a significantly higher zeta-potential was observed for C2, C4, and C5, compared to their linear counterparts L2, L4, and L5, respectively, at N/P ratios of 10 and higher. While hybrid and linear counterparts contain the same number of arginines and, theoretically, should not present different surface charges after forming complexes with siRNA, hybrid peptides are known to be more rigid in structure and, therefore, in the distribution of the electric charge, which might contribute to the observed differences. Addition of the negatively charged additive lowered the zeta-potential of the complexes formed by three of the cyclic peptides, as expected (Figure 3C), which was statistically significant at lower N/P ratios. At N/P ratios of 20 and 40 the differences disappeared as the contribution of the electric charge of the peptide became overwhelming.

Serum stability

One of the major roles of a nucleic acid carrier is to protect the nucleic acid from degradation before reaching the target cells. RNA is highly vulnerable to RNases, which are ubiquitous in biological systems and readily degrade exposed RNA molecules. To study siRNA stability, we used fetal bovine serum (FBS) as a source of degrading enzymes. Free siRNA was exposed to HBSS or FBS as negative (NC) and positive controls (PC), respectively (Figure 4).

As expected, the free siRNA exposed to HBSS was intact after 24 h incubation. To our surprise, however, many of the siRNA/peptide complexes did not form a similarly distinct band and appeared to not move at all. In fact, for some of the complexes, the band appeared to be in the well where the sample was added, indicating no movement through the gel (see red arrows marking some of those samples; Figure 4). In our method, which we have used over the years with complete release of siRNA, the heparin concentration in the final sample was 1.4 mg/mL, and the heparin/siRNA w/w ratio was 56. While there are reports of incomplete release of siRNA with exposure to heparin (Gujrati et al., 2014; Nguyen et al., 2008; Parnaste et al., 2017), this was unprecedented in our studies. We noticed a study that reported 4 mg/mL as the required concentration of heparin for complete release of siRNA from a cell-penetrating peptide (Parnaste et al., 2017), and therefore, we increased the heparin final concentration to 4 mg/mL and did not see any visible difference in the blots (data not shown). This observation confirmed our concern about the strong binding affinity and the risk of limited siRNA release inside the cell. Only L1 and, to a lesser extent, L2, showed partial release of siRNA, as apparent by the appearance of the free siRNA band. After the addition of the negatively charged additive to siRNA complexes formed with C1, C2, and C3 (which showed no sign of release in the initial experiment), the free siRNA band appeared, which was most prominent at lower N/P ratios. This experiment indicates that the addition of Trans-Booster can facilitate the release of siRNA from the complexes, which we hypothesized could improve silencing efficiency.

Toxicity

The safety profile is among the important characteristics of a delivery system. Given the fact that the cells will be exposed to the complexes in the future in vivo studies and clinical applications, rather than testing the toxicity of the peptide molecules, we studied the toxicity of the peptide/siRNA complexes in three breast cancer cell lines. We also used Lipofectamine™ as a point of reference as a commercially available siRNA delivery agent (Figure 5). The cell viability was reduced the most in AU 565 cells after 48 h of exposure to Lipofectamine™ (42% cell viability at the manufacturer recommended formulation parameters) compared to MDA 468 and MDA 231 cells (50.20% and 59.5% viability, respectively). Among the nanoparticles, the complexes formed with hybrid peptide C3 showed the greatest effect on cell viability of MDA-MB-231 cells (80.15% cell viability at N/P ratio of 80), followed by its linear counterpart, L3 (84.5% cell viability at the highest N/P ratio tested; Figures 5A,B). In MDA-MB-468 cells, linear L4 and hybrid C5 reduced the cell viability to 85.7% and 90.6%, respectively, at N/P ratio of 80 (Figures 5C,D). The lowest cell viability after exposure to peptide/siRNA complexes was observed in AU565 cells, with 72.3% at an N/P ratio of 40 for C1/siRNA complexes (Figure 5E). Overall, the toxicity observed with peptide/siRNA complexes was significantly less than the Lipofectamine™ at all cell lines included and could be described as “mild”.

siRNA internalization

Internalization of nucleic acids into cells is a challenging step, but necessary for the efficiency of protein silencing. We studied the internalization of a scrambled siRNA delivered by the peptides in the three selected breast cancer cell lines with Lipofectamine™ as a point of reference. Both the mean fluorescence signal and the percentage of the cells positive for the fluorescent signal were used to analyze the delivery data.

In MDA-MB-231 cells, the N/P ratio of 20 was more efficient than the N/P ratio of 40 for all included peptides except L1 (Figures 6A,B,H). Only L1 at an N/P ratio of 40 and L3 at an N/P ratio of 20 demonstrated significantly higher internalizations than lipofectamine™. Also, L1 at a ratio of 20 and L3 at ratio of 40 showed internalizations that were comparable to Lipofectamine™. Except L1, internalization decreased with an increase in N/P ratio from 20 to 40. None of the hybrid peptides achieved siRNA internalization levels comparable to those of Lipofectamine™. A similar pattern was observed in the percentage of cells positive for fluorescent signal. Overall, a negative correlation was apparent between the number of arginines and internalization for linear and hybrid peptides in both studied N/P ratios; however, it was only statistically significant for hybrid peptides at N/P ratio of 40 (R = 0.95, P < 0.05; Figure 6G).

Our internalization results in MDA-MB-231 cells were unexpected in certain regards. An increase in the number of tryptophans was expected to enhance the interaction of the peptide/siRNA complexes with the cell membrane and, therefore, increase cellular internalization (Neuberg et al., 2019; Walrant et al., 2020; Konate et al., 2021). This has been observed in our own previous studies as well. For example, in a recent study on cyclic and linear peptides containing alternate WW and RR residues, we noticed a steady increase in siRNA internalization into MDA-MB-231 cells with an increase in the number of tryptophan and arginine residues (Kim et al., 2023). However, the reverse trend has also been reported. In 2019, we reported synthesis and characterization of amphiphilic peptides with a ring composed of arginines (5 or 6) linked to a chain of tryptophans (5 or 6) via lysine ([R 5 K]W 5 and [R 6 K]W 6, respectively (Mozaffari et al., 2019). In that study, we observed an enhanced cellular internalization for [R 5 K]W 5 in MDA-MB-231 and AU 565 cells and a comparable internalization for the two complexes in MDA-MB-468 cells. Overall, the number of these amino acids in the peptide carrier and the extent of siRNA internalization might not be correlated in all cases. The other surprise was the higher efficiency of the N/P ratio of 20 compared to 40. A higher N/P ratio indicates complexes with a higher positive surface charge, which, in theory, can enhance cellular internalization. This has been observed in some of our previous studies using peptides for siRNA delivery (Salehi et al., 2022; Sharma et al., 2017; Rai et al., 2024). However, there have also been exceptions in our observations before. For example, in the same 2019 study, we reported a lower internalization for [R 6 K]W 6 /siRNA complexes formed with an N/P ratio of 40 (compared to an N/P ratio = 20) in MDA-MB-468 cells. Similarly, a drop in internalization was observed for [R 5 K]W 5 /siRNA complexes with an increase in N/P ratio from 20 to 40 in AU 565 cells (Mozaffari et al., 2019). Also, a lower internalization was observed for hybrid cyclic/linear peptides containing arginine and tryptophan when we increased the N/P ratio from 20 to 40 (Mandal et al., 2022). Our data once again provides evidence for the necessity of tailoring the delivery system for each carrier and purpose.

Another explanation could be the overall three-dimensional structure of the peptides and their interactions with siRNA. The conformation adopted by a peptide in solution is influenced by its amino acid composition, sequence, and the balance between hydrophobic and electrostatic forces. As the number of tryptophan and arginine residues increases, the peptide may adopt a more compact or rigid conformation due to increased intramolecular hydrophobic interactions and stacking of the indole rings of tryptophan residues. This structural reorganization could alter the spatial arrangement of the positively charged arginine residues, potentially reducing their accessibility for electrostatic interaction with the negatively charged siRNA backbone. The hypothesis that higher tryptophan/arginine content promotes a more compact or rigid peptide conformation, potentially reducing the accessibility of arginine residues for electrostatic contact with siRNA, can be tested by circular dichroism spectroscopy of the peptides in the absence and presence of siRNA. Such conformational analysis, which was beyond the scope of the present delivery-focused study, is a priority for future mechanistic work.

Conversely, a peptide with fewer hydrophobic residues may adopt a more extended, flexible conformation, allowing more favorable, multivalent contact with siRNA. Furthermore, the three-dimensional structure of the peptide/siRNA complex, including its size, shape, surface charge distribution, and hydrophobicity, can influence how it interacts with the cell membrane, is endocytosed, and escapes the endosome. It is also plausible that at higher N/P ratios, excess free peptide can compete with or disrupt the organized structure of the complex, leading to aggregation or altered surface properties that are less conducive to cellular uptake.

The overall uptake was lower in MDA-MB-468 cells (including Lipofectamine™, 956 Arbitrary units, or A. U., for MDA-MB-468 cells, compared to 1436 A. U. for MDA-MB-231 cells). Interestingly, a similar trend for different N/P ratios was not observed in this cell line. For most of the complexes included in the study, the level of mean fluorescence intensity was higher for complexes prepared with N/P ratio of 40, which was in some cases statistically significant (Figure 6C; Supplementary Figure S2B). While similar trends were observed in data calculated as the percentage of cells positive for the fluorescent signal, the differences were not significant, since most study groups generated close to 90% fluorescent-positive cells (Figure 6D). In contrast to our observations in MDA-MB-231 cell line, here an apparent positive correlation was observed between internalization (based on mean fluorescence) and the number of arginines. Due to small number of data points, however, the correlation factor was not significant for any of the categories of peptides (Supplementary Figure S2A). Our observations in MDA-MB-468 cell line were more aligned with the expectations. Increasing the N/P ratio increased siRNA internalization in most cases and increasing the number of arginines and tryptophans had a slight positive effect on internalization, though this was not statistically significant.

Interestingly, in the AU 565 cell line, we observed a combination of trends for the effect of N/P ratio on internalization (Figure 6E; Supplementary Figure S2D). The internalization (calculated based on mean fluorescence) was higher at an N/P ratio of 20 for L2 (not significant), L3, L5 (not significant), C4, and C5 (not significant). On the other hand, the N/P ratio of 40 showed higher internalization for the remaining peptides included in the study (only significant for C1 and C3). A negative correlation between the number of arginines and mean fluorescence was also observed in this cell line; however, it was not statistically significant (Supplementary Figure S2C).

In order to visualize the internalization of fluorescent-labeled siRNA, we performed confocal microscopy on the delivery of free siRNA and selected peptide/siRNA complexes to MDA-MB-231 cells. DAPI and Texas Red were used to stain the nuclei and cell membrane, respectively (Figure 7). The intensity of the green fluorescence signal in the microscopy images correlates with our internalization data presented in Figure 6. Investigation of the 3-dimensional images confirmed that the location of the signal associated with the siRNA was within the confines of the cells.

Protein silencing via siRNA delivery

We selected STAT3 as a model protein to target in the MDA-MB-231 cell line due to our previous studies using this protein (Mandal et al., 2023; Mandal et al., 2022). We used cells exposed to normal saline as “No Treatment”, to Lipofectamine™ as a point of reference, and to L1/Scrambled siRNA complexes as a negative control. Lipofectamine reduced the STAT3 expression (normalized based on Beta actin expression as the endogenous protein) to 61% on average (Figure 8). This efficiency is comparable to some of the previous reports where Lipofectamine/STAT3 siRNA reduced the expression level of the protein to around 55% in vascular smooth muscle cells (Sun et al., 2012), or showed an approximately 50% reduction in expression in hepatocarcinoma cells (Zhang et al., 2016). However, there are other studies that have reported higher efficiencies for Lipofectamine (Li et al., 2021), including our own studies (Mandal et al., 2023). Using linear peptide L1 to deliver STAT3 -targeting siRNA did not reduce the expression level of STAT3 (92.5% expression vs. 100.5% expression for L1/scrambled siRNA group). Again, we tested the addition of Trans-Booster at three different levels (2, 4, and 8 μg/mL) to the polymer and siRNA mixture before incubation at room temperature. Trans-Booster addition enhanced silencing efficiency and reduced STAT3 expression to ∼ 82, 45, and 41% levels compared to no-treatment cells at 2, 4, and 8 μg/mL Trans-Booster concentration, respectively. This observation once again confirmed the enhanced performance of the complexes upon addition of a negatively charged component.

To confirm our observations using a more quantitative approach, we targeted GFP in the GFP-positive MDA-231 cells using three of the peptides (L1, L3, and C1) and a combination of peptide (L1) and Trans-Booster and quantified the fluorescence signal using flow cytometry. The mean fluorescence of the cell populations is summarized in Figure 9. Delivering scrambled siRNA using Lipofectamine™ reduced the fluorescence of the cells significantly (∼ 2500 A.U. compared to >3500 A.U. for no treatment group of the GFP-positive cells), which could be due to toxicity. However, silencing was efficient with 23.6% of the GFP signal remaining after delivering GFP-targeting siRNA (compared to the cells exposed to scrambled siRNA). While the peptides alone did not reduce the GFP signal significantly (Remaining Expression = 96, 93, and 90% for L1, L3, and C1, respectively), the remaining GFP signal after exposure to L1/Trans-Booster/GFP siRNA was comparable to Lipofectamine™ (%23.7) with no significant reduction in the fluorescence signal as a result of delivering scrambled siRNA with any of the peptide study groups.

Our central observation that strong peptide/siRNA binding can limit intracellular cargo release and thereby suppress silencing despite efficient uptake is mostly consistent with reports from other groups. For example, Han et al. showed that siRNA binding affinity governs in vivo antitumor efficacy of polyplexes, with excessively tight binding impairing activity (Han et al., 2013). Working with amphipathic cell-penetrating peptides, Pärnaste et al. reported that as much as 4 mg/mL heparin was required for complete siRNA release, underscoring how tightly some CPP/siRNA complexes bind, in agreement with our heparin-competition results (Parnaste et al., 2017). Konate et al. found that WRAP-based peptide nanoparticles could rival lipid reagents for siRNA delivery but required careful balancing of hydrophobic and cationic content, consistent with our finding that increasing arginine/tryptophan content does not monotonically improve delivery (Konate et al., 2021). At the same time, opposing observations also exist: Yang et al. reported that ultra-high-affinity dsRNA-binding proteins can enhance cytosolic siRNA delivery, highlighting that the optimal binding strength is system-dependent (Yang et al., 2017). Our data reinforces the emerging consensus that an intermediate, tunable binding affinity. Here achieved by adding a polyanionic component, rather than maximal affinity, is desirable for efficient nucleic acid delivery.

CRISPR/Cas9 internalization

Delivery of CRISPR/Cas9 RNPs was performed to investigate the feasibility of the peptide library for delivering a second form of nucleic acids. The uptake was detected intracellularly using fluorescent-labeled TracrRNA (ATTO 550) in this experiment at three final concentrations for the peptides: 12.5, 25, and 50 μM) delivering 25 nM of RNP. In addition, CRISPRMAX was used as a positive control in this experiment, according to the manufacturer’s guidelines, as a commercially available RNP carrier. In fact, the highest level of internalization was observed with CRISPRMAX, both in mean fluorescence (∼ 255 A.U.) and percentage of fluorescent-positive cells (∼ 47%).

Linear peptide L3 at 25 μM RNP concentration showed the highest ATTO-550 fluorescence signal among all peptides (210 A.U. for mean fluorescence and 35% fluorescent positive cells). Linear peptide L4 at 25 μM concentration was the second largest signal (with 188 mean fluorescence and 33% fluorescent-positive cells).

Hybrid cyclic peptide C4 achieved the highest internalization among the hybrid cyclic peptides at 50 uM in terms of mean fluorescence (165 A.U.), but C1 showed the highest percentage of fluorescent-positive cells (34%). Overall, increasing the final concentration of RNPs from 12.5 to 25 μM increased uptake in most cases; however, increasing the concentration to 50 μM did not improve uptake for most peptides. Also, despite the differences mentioned above, and unlike our observations with siRNA internalization in MDA-MB-231 cell line, a drastic difference between RNP cellular internalization efficiency was not observed among the peptides included in this study (Figure 10).

CRISPR/Cas9 editing

The HPRT was selected as the model gene targeted by CRISPR, and the T7 Endonuclease 1 (T7E1) was used to identify and cleave the Non-homologous end joining (NHEJ) caused by CRISPR/Cas9, as reported previously by us (Alhazza et al., 2024). In transfected cells, in addition to the original band, for the HPRT, we expected to see two extra bands at 471 and 198 base pairs due to the digestion of heteroduplexes by the endonuclease T7E1. While the RNPs delivered by CRISPRMAX™ showed two distinct expected bands, none of the selected peptides (L3, C3, and C4) showed any signs of the T7E1 cleavage (Figure 11). In the study groups that included Trans-Booster, a faint band was observed, further demonstrating the effect of this additional ingredient in enhancing the carrier’s efficiency. This effect was more noticeable for L3 and C3.

Limitations and future directions

Several limitations of the present study should be acknowledged. First, although our functional data (serum-stability release, silencing, and editing, all improved by the anionic additive) support a model in which strong carrier/cargo binding limits intracellular release, we did not obtain direct mechanistic evidence of the release event itself or of the endosomal-escape step; time-resolved intracellular tracking and colocalization studies would address this. Second, the biophysical characterization of the peptide/siRNA interaction was based primarily on a functional dye-exclusion assay (BR50) and DLS. Complementary techniques, such as FTIR and Isothermal Titration Calorimetry (ITC) to characterize molecular interactions and their thermodynamics (including the dissociation constant or kd), Circular Dichroism (CD) spectroscopy to detect siRNA-induced conformational changes, and Transmission Electron Microscopy (TEM)/Cryo-TEM to resolve nanoparticle morphology, would provide a more complete structural picture and are priorities for our follow-up work. Third, stability of the nanoparticles (analyzed by DLS to monitor size and electrical charge) in biologically relevant environments (e.g., serum) and during storage time (for colloidal dispersion as well as lyophilized nanoparticles) will be included in the next stage of this project. Finally, we acknowledge that all experiments were performed in vitro using established cell lines; in vivo validation, including biodistribution, efficacy, and safety, will be required before therapeutic relevance can be established. We view the present study as establishing the design principle that tuning carrier/cargo binding strength with a polyanionic additive improves the efficiency of the delivery, which these future studies can build upon.

In this study, we investigated the two small libraries of linear and hybrid peptides for their efficiency in delivering nucleic acids to triple-negative breast cancer cells. While some of the peptides showed significant cellular internalization of siRNA, this internalization did not translate to the expected silencing efficiency, which we hypothesized to be due to a strong binding of siRNA to the peptides. Adding a commercialized negatively charged component to the complexes significantly enhanced silencing efficiency, as did the transfection efficiency with CRISPR/Cas9 delivery. Our results emphasize the importance of carrier/nucleic acid interaction in the efficiency of the delivery system. We conclude that the addition of a negatively charged component to peptide/nucleic acid complexes can enhance the efficiency of nucleic acid by facilitating the release of the cargo in the cytoplasm.