Self-assembly and intracellular delivery of DNA by a truncated fragment derived from the Trojan peptide Penetratin

Lucas R. Mello a, Ian W. Hamley b, Valeria Castelletto b, Bianca B. M. Garcia a, Thiago C. Lourenço a, Sandra V. Vassiliades c, Wendel A. Alves c, Sang W. Han a and Emerson R. Silva *a
aDepartamento de Biofísica, Universidade Federal de São Paulo, São Paulo 04023-062, Brazil. E-mail: er.silva@unifesp.br
bDepartment of Chemistry, University of Reading, Reading RGD 6AD, UK
cCentro de Ciências Naturais e Humanas, Universidade Federal do ABC, Santo André 09210-580, Brazil

Received 26th February 2020 , Accepted 17th April 2020

First published on 17th April 2020


Penetratin is a short Trojan peptide that attracts great interest in biomedical research for its capacity to translocate biological membranes. Herein, we study in detail both self-assembly and intracellular delivery of DNA by the heptamer KIWFQNR, a truncated peptide derived from Penetratin. This shortened sequence possesses a unique design with bolaamphiphilic characteristics that preserves the longest noncationic amino acid portion found in Penetratin. These features convey amphipathicity to assist self-assembly and make it a suitable model for exploring the role of hydrophobic residues for peptide interaction and cell uptake. We show that the fragment forms peptiplexes (i.e., peptide–DNA complexes), and aggregates into long nanofibers with clear β-sheet signature. The supramolecular structure of nanofibers is likely composed of DNA cores surrounded by a peptide shell to which the double helix behaves as a template and induces fibrillization. A nucleation and growth mechanism proceeding through liquid–liquid phase separation of coacervates is proposed for describing the self-assembly of peptiplexes. We also demonstrate that peptiplexes deliver double-stranded 200 bp DNA into HeLa cells, indicating its potential for preparing non-viral vectors for oligonucleotides through noncovalent strategies. Since the main structural features of native Penetratin are conserved in this simpler fragment, our findings also highlight the role of uncharged amino acids for structuration, and thus for the ability of Penetratin to cross cell membranes.


Introduction

Cell-penetrating peptides (CPPs), often also called Trojan peptides, are considered one of the most promising candidates for intracellular delivery of bioactive molecules and constitute a major peptide class in future development of therapeutics.1–3 CPPs have been serendipitously discovered about three decades ago during researches on the function of transcriptional regulators in gene activity.4,5 One striking Trojan peptide that emerged from these earlier investigations is Penetratin, a 16-mer sequence present in the Antennapedia homeodomain of Drosophila melanogaster.1,6 This sequence exhibits strong tropism towards biological membranes7 and since its discovery it has been extensively investigated as a transporter for a variety of molecular cargoes.8–10 Recently, Penetratin has been employed in new strategies to overcome the blood brain barrier,11,12 to promote intraocular delivery9,13,14 and for the production of novel cancer therapeutics.10,15 Particularly, the DNA-binding characteristics of the parent Antennapedia homeoprotein has inspired the development of Penetratin-based vectors for delivery of nucleic acid payloads which has potential for use in gene therapy.9,16,17

The Penetratin peptide is composed of the following amino acid sequence: RQIKIWFQNRRMKWKK.6 Similar to vast majority of Trojan sequences,3,18 it is composed of several basic residues making it very competent for noncovalent interactions with anionic species such as DNA. In contrast, it also displays a few hydrophobic amino acids that lead to a “dual mode” of interaction with lipid membranes.19,20 The ability of Penetratin to form ordered aggregates in the presence of anionic surfactants has motivated the study of the synergism between this peptide and lipids.19,21,22 Although understanding on the exact mechanisms by which Penetratin translocates cell membranes is a still evolving process, it has been shown that interaction with negatively-charged phospholipids is mainly driven by electrostatic attraction, followed by limited insertion of peptide into bilayers.23,24 In addition, it has been demonstrated that amphipathicity of Penetratin is not sufficient to allow deep insertion into biomembranes,25 and in this case permeation has been proposed to occur through a process involving the formation of inverted micelles.26

In the present work, we study the heptamer KIWFQNR, a shortened fragment derived from Penetratin. The chemical structure of the peptide is presented in Fig. 1, and it shows several peculiarities that have motivated our choice: first, it contains the longest hydrophobic segment in the Penetratin strand. Since structuration capabilities of Penetratin have been identified as an intermediate step for its translocation capacity,24,26,27 this hydrophobic portion should play a relevant functional role in the permeation properties of the Trojan peptide. Secondly, it carries a tryptophan residue which is well-known to be necessary for efficient membrane translocation by Penetratin.28–30 As shown below, tryptophan also provides a powerful probe to track physicochemical properties of the peptide. Finally, it possesses cationic residues at both termini, enabling electrostatic binding to anionic species (e.g., DNA or biomembranes), and conveying a bolaamphiphile-like design that could assist self-assembly into ordered structures. Putting these features together, it is of great interest to investigate the role of this hydrophobic segment in relation to the structuration and cell penetration behavior of the parent Penetratin peptide.


image file: d0sm00347f-f1.tif
Fig. 1 Chemical structure of the peptide investigated in this work (KIW7). Cationic residues at both termini and the tryptophan indole group in the middle of the sequence appear colored.

Materials and methods

Peptide synthesis, reagents and sample preparation

The heptameric sequence NH2-KIWFQNR-COOH (K: lysine, I: isoleucine, W: tryptophan, F: phenylalanine, Q: glutamine, N: asparagine, R: arginine) was synthetized using routine solid-phase approaches as described elsewhere.31,32 For simplicity, henceforth, the heptameric sequence will be referred to by the shortened form KIW7. The peptide product has been obtained as a TFA salt, being characterized by liquid chromatography coupled to mass spectroscopy assays which revealed purity and molecular mass, respectively, of 97.7% and 991.6 g mol−1 (calculated Mw = 991.15 Da). The corresponding chromatogram and mass spectrum are shown in the ESI (Fig. S1). Calf thymus DNA, containing 10–20 kb, was acquired from Sigma-Aldrich (product code D1501). These polymerized chains were broken into fragments with ∼200 bp through ultrasonication on a Diagenode Bioruptor as previously described,33–35 which were characterized by agarose electrophoresis (ESI, Fig. S2A). These 200 bp DNA fragments possess calculated length at L = 200 × 0.33–67 nm, close to the to the persistence length of double-stranded DNA.36 SAXS measurements on 200 bp DNA solutions indicated the presence of rod-like structures, with diameters around 2.6 nm, consistent with the thickness of DNA duplexes plus a hydration shell (see ESI, Fig. S2B).36 Plasmid vector, pEGFP-N3, was acquired from Clontech (Mountain View, CA, USA) and DH5a bacteria was used for transformation and amplification.37 For large scale production of this plasmid, Endo-free Gigaprep kit from Qiagen (Hilden, Germany) was used following the manufacturer's instructions. The purified plasmid was resuspended in water and maintained in −20 °C freezer until its use. Stock solutions were prepared by weighing lyophilized peptides and 200 bp DNA into test tubes and re-suspending in ultrapure water (or D2O). Peptides formed transparent samples immediately upon dissolution whereas concentrated solutions prepared from lyophilized 200 bp DNA required a few hours of alternating vigorous vortexing cycles to assist homogenization. All preparations were kept in the fridge (4 °C). Complexes were prepared by mixing defined amounts from stock solutions to obtain desired molar ratios between cationic charges on peptide side chains and anionic charges at phosphate groups of DNA (N+[thin space (1/6-em)]:[thin space (1/6-em)]P ratio). Since our preparations were close to neutral pH, a net charge +2 per peptide strand and −2 per DNA base pair was assumed. Spectrofluorimetry assays: fluorescence assays were carried out using a Shimadzu F2500 instrument. Samples with concentrations ranging from 2.2 mg ml−1 down to 3 × 10−4 mg ml−1 were prepared through serial dilutions in water. The presence of a tryptophan residue in the amino acid sequence allowed to use its intrinsic fluorescence to assess changes in the aggregation state of the peptide. Samples were irradiated with λexc = 280 nm and emission spectra were acquired in the range 300–500 nm. Excitation and emission slits were set for the best achievable signal/noise ratio. Circular dichroism (CD). Aqueous solutions containing peptides at 0.1 mg ml−1 were placed into Hellma cells with 1 mm path length. The peptide solutions were titrated with 200 bp DNA aliquots from a concentrated solution in order to scan a range of N+[thin space (1/6-em)]:[thin space (1/6-em)]P ratios below and above the point of molar equivalence between positive and negative charges. CD data were collected in the range 190–260 nm, in steps of 1 nm and 1 s per step, using four accumulations. All data have been background-subtracted prior to further treatment, and conversion into ellipticity units has been performed considering the mean residual weight, peptide concentration and pathlength of cuvette used in the experiments. Only data exhibiting absorbance <2 were considered for analysis. Fourier filtering (7 points window) was used to minimize noise. Small-angle X-ray scattering and fiber X-ray diffraction. SAXS data were collected on SAXS-1 beamline at LNLS (Campinas, Brazil), using X-rays energy of 8 keV and q-range 0.12 nm−1 < q < 3.0 nm−1. Samples were left to equilibrate for 2–3 days at 4 °C prior to SAXS experiments and peptide concentrations of about 5 mg ml−1 were used in the preparations, while the amount of 200 bp DNA was adjusted to match the 1[thin space (1/6-em)]:[thin space (1/6-em)]1 peptide-to-phosphate charge ratio. Aliquots 300 μl of solutions were carefully injected between mica windows of a cell holder with 1 mm pathlength. Ten frames, 30 s each, were acquired and, since no radiation damage was detected, they were averaged, and background subtracted. All experiments were carried out at room temperature. Samples for fiber X-ray diffraction were prepared by suspending droplets from concentrated KIW7/pDNA solutions (5 mg ml−1) between wax-coated capillaries. The capillaries with the suspended droplets were carefully placed inside Petri dishes, sealed with parafilm, and left to dry for about 1 week in the fridge. The resulting stalks were positioned onto a RAXIS IV++ diffractometer and irradiated with 8 keV X-rays. The sample-to-detector distance was kept at 50 mm and data were collected by a Saturn 992 CCD camera. 1D spectra were obtained by using Fit2D software. Atomic Force Microscopy: a Park NX10 atomic force microscope was used for data collection at LNNano (CNPEM, Campinas – Brazil). Samples were prepared by casting about 20 μl from peptides or 1[thin space (1/6-em)]:[thin space (1/6-em)]1 peptide[thin space (1/6-em)]:[thin space (1/6-em)]DNA solutions onto the surface of freshly cleaved mica. Samples were left to rest for about 5 minutes and then rinsed to wash out excess of solution. Substrates were left to dry overnight into desiccators prior to measurements which have been performed in tapping mode, with a cantilever operating at ∼240 kHz. Image visualization and enhancement was performed with software Gwyddion. Cryogenic transmission electron microscopy (TEM). Transmission electron microscopy imaging was carried out on a JEOL 2100 FEG-TEM microscope at LNNano, Campinas, Brazil. Grids for cryo-TEM samples have been treated with a plasma glow discharge to enhance adhesion of particles. Aliquots 3 μl from 5 mg ml−1 peptide solutions were cast onto lacey carbon grids (300 mesh), and quickly plunged into liquid ethane cooled by liquid nitrogen using a Vitribot device. The grids were transferred into the microscope under refrigerated conditions and kept at cryogenic temperatures until the end of analyses. The microscope operated at an acceleration voltage of 200 keV and further data treatment was performed with Image J software. Cell culture and fluorescence imaging using confocal microscopy: HeLa cells were cultured in Dulbecco's modified Eagle medium (DMEM) with the addition of 10% of fetal bovine serum and 2 mM of glutamine (ThermoFisher) and maintained in a humidified chamber with 5% CO2 at 37 °C. In each well of 24-well plates, a sterilized glass coverslip was placed on the bottom and 5 × 104 cells were added and maintained for 24 hours with 5% CO2 at 37 °C. Culture plates were washed three times with PBS to remove excess of serum and cell debris. The complexes formed between KIW7 and 200 bp DNA in serum-free DMEM were added in the plates and incubated for 4 hours. After this, plates were washed with PBS three times to remove unadhered cells and unbonded complexes. The adhered cells were fixated with 4% paraformaldehyde and stained with DAPI fluorescent dye (Thermo Scientific) in PBS for five minutes to stain cell nuclei. After this period, they were washed again three times with PBS and analyzed by confocal microscopy (Leica TCS SP8, Mannheim, Germany).

Results and discussion

Steady-state fluorimetry: association to anionic species

Since the indole group of the tryptophan side chain is highly sensitive to local conditions of the molecule, intrinsic fluorescence of KIW7 could be used to extract valuable information on its chemical environment and provide insights into the structure. KIW7 possesses a bolaamphiphilic design, therefore, this peptide is likely prone to self-aggregate in aqueous solution. In Fig. 2A, we show a series of fluorescence spectra from aqueous solutions at different concentrations whose emission spectra (inset) are characterized by a broad peak at λmax ∼ 350 nm. Fluorescence intensities are modulated by peptide concentration, increasing to such an extent that solutions at ∼0.2 mg ml−1 exhibit a violet glow under UV light (insets in Fig. 2A). To get further understanding on this emission behavior, peak intensities have been plotted as a function of logarithm of peptide concentration. This plot discloses two very distinct regimes, with emission showing moderate increase in the dilute domain and strong growth in the concentrated range. The data suggest stabilization of fluorescence at higher concentrations (i.e., >0.2 mg ml−1), possibly correlated to appearance of peptide aggregates leading to light scattering.38 The overall shape of the curve is well described by a sigmoidal function with inversion point at 0.05 mg ml−1, which is interpreted as a putative critical aggregation concentration (CAC) for KIW7.
image file: d0sm00347f-f2.tif
Fig. 2 (A) Tryptophan fluorescence from KIW7 solutions. Peak intensities at λmax = 350 nm have been plotted as a function of peptide concentration and fitted to a Boltzmann function with inversion point at 0.05 mg ml−1. Insets: Spectra used to construct the sigmoidal plot and photographs from solutions under UV light. (B) Emission spectra from 0.05 mg ml−1KIW7 solutions in the presence or absence of 200 bp DNA and SDS micelles. (C) Stern–Volmer plots from acrylamide quenching assays in solutions containing KIW7 (circles) or peptiplexes (squares), either in water (filled symbols) or in 29 mg ml−1 SDS micelles (open symbols).

Tryptophan fluorescence could be used to probe interactions between KIW7 and other molecules.39 Since this heptamer is a cationic fragment derived from Penetratin, a peptide with recognized cell-penetrating capabilities useful for intracell delivery, we investigated its interaction with 200 bp DNA and with sodium dodecyl sulfate (SDS) micelles. DNA is a potentially bioactive load highly attractive for gene therapy, whereas SDS micelles are extensively used as a first approximation for surfactant interfaces to mimic interaction between proteins and biomembranes due to the anionic nature of their headgroups.40–43Fig. 2B shows fluorescence spectra from mixtures containing 0.05 mg ml−1KIW7 either in the presence of 200 bp DNA or co-solubilized with SDS micelles. The results show a blue shift of about 10 nm in the emission spectrum in the presence of SDS, indicating that indole moieties in these samples are surrounded by a hydrophobic environment, and suggesting that tryptophan residues are partially shielded from the aqueous phase.44 In the absence of SDS micelles, co-solubilization with 200 bp DNA does not induce peak shifts, however, a remarkable decrease of intensity is observed. We verified that absorbance at λ = 280 nm is marginally higher in samples containing DNA, therefore such a strong decline in fluorescence likely results from association between KIW7 and DNA which presumably diminishes the rotational freedom of tryptophan side chains.39 Interestingly, interaction with SDS micelles is also observed in samples containing KIW7/200 bp DNA at equal molar charge ratio, which is an indication that these peptiplexes could be able to interact with anionic surfactant interfaces even in the absence of strong electrostatic attraction.

We also investigated accessibility of tryptophan residues through acrylamide quenching assays. In Fig. 2C, Stern–Volmer plots are shown along with linear fits according to the equation F0/F = 1 + KSV[Q], where F0 is the fluorescence intensity of unquenched solutions, F is the emission at a given acrylamide concentration [Q] and KSV is the Stern–Volmer constant that provides a quantitative index for quenching efficiency.39 Solutions containing only KIW7 in water presented KSV = 24.8 ± 0.3 M−1, whereas those containing the peptide in the presence of SDS micelles exhibited the lowest quenching constant values at KSV = 10.5 ± 0.2 M−1. Not surprisingly, lower quenching constants have been found for SDS-containing samples, reinforcing that peptides are partially buried in hydrophobic core of micelles and this suggests that KIW7 is a good candidate for interacting with cell membranes. In the case of samples containing peptiplexes prepared at equal peptide-to-DNA charge ratio, KSV constants present intermediate values and have been found at 17.3 ± 0.2 and 13.6 ± 0.3 M−1, respectively, for preparations in water and SDS micelles. These results indicate that although complexation with 200 bp DNA decreases the strength of interaction with micelles, association is still possible likely due to the amphipathic nature of the peptide. These findings represent initial physicochemical evidence supporting KIW7 as a potential candidate to prepare peptiplexes for intracell delivery of DNA.

Circular dichroism assays: secondary structure transitions

Circular dichroism (CD) assays were performed to investigate the secondary structure of KIW7 peptide. The CD spectrum from KIW7 in water (Fig. 3A) comprises shallow positive peak at ∼195 nm followed by a sharp negative rotation at ∼205 nm, corresponding to a red shifted signature of disordered conformations.45,46 A positive maximum near 225 nm is likely correlated to electronic transitions of aromatic side chains of tryptophan and phenylalanine residues.45,47 The presence of disordered conformations in KIW7 is consistent with secondary structure displayed by Penetratin, which also exhibits such structures in aqueous solutions.23,24
image file: d0sm00347f-f3.tif
Fig. 3 (A) CD data from KIW7 solutions (0.5 mg ml−1) containing only the peptide or preparations co-solubilized with SDS micelles. (B) CD spectra from KIW7 samples (0.1 mg ml−1) titrated with 200 bp DNA at the indicated peptide-to-phosphate charge molar ratios. 200 bp DNA spectrum (dashed line) is not to scale and is presented just a guide to the eye to allow comparison with peptide samples.

An interesting feature of Penetratin is appearance of random-to-β-sheet transitions upon interaction with anionic species, especially mimetic membrane models.24,48 To investigate if these features are also found in the KIW7 fragment, we measured CD spectra in samples containing SDS micelles. The spectrum shown in Fig. 3A (open symbols) indicates a broad negative rotation centered at ∼220 nm appearing upon interaction with micelles along with a shallow positive peak near to 203 nm, suggesting organization into β-sheets when KIW7 associates with anionic surfactant interfaces.18,49 Similar random-to-β-sheet transitions have been recently identified in Penetratin interacting with nucleic acids.45 In Fig. 3B, we show CD spectra from a peptide solution titrated with concentrated 200 bp DNA aliquots. The amount of DNA in the formulations has been chosen to range from a cationic peptide excess (2.7[thin space (1/6-em)]:[thin space (1/6-em)]1 molar ratio) up to an excess of anionic phosphate charges (1[thin space (1/6-em)]:[thin space (1/6-em)]2.2). Upon titration, spectra become quickly dominated by a strong positive peak at ∼190 nm, which is very characteristic of B-form DNA.46B-form signature is also attested by a maximum at 210 nm and a minimum at 220 nm in CD spectrum from 200 bp DNA solutions (dashed line in Fig. 3B). In addition to the strong rotation at 190 nm, data are characterized by increasing negative bands near to 223 nm, which indicates the emergence of β-sheet conformations upon interaction with anionic species. The presence of an isochoric point at 217 nm suggests two state cooperative transitions.46,50

The findings above demonstrate that KIW7 shows that strong secondary structure similarity with Penetratin, suggesting that this heptameric fragment is crucial for the conformation of its parent sequence, and potentially plays a key role for non-covalent interactions with cell membranes and DNA cargoes.

X-ray, cryo-EM and AFM experiments: nanoscale structure

To extract detailed information on the nanoscopic structure of assemblies based on KIW7 peptides, we have used complementary biophysical techniques comprising in situ small-angle scattering (SAXS) in liquid medium and cryogenic transmission electron microscopy (cryo-TEM) on vitrified specimens, and ex situ atomic force microscopy (AFM) and fiber X-ray diffraction on samples dried from peptide/pDNA solutions.

In SAXS experiments, the q-range investigated corresponds to lengths between 2 to 50 nm; therefore, the data provide insights into the supramolecular arrangement of the assemblies whereas microscopy give information on the morphology of the peptiplexes. In Fig. 4A, we show scattering curves from two samples, one of them containing only KIW7 at 10 mg ml−1 and other with a mixture of peptides and 200 bp DNA at equal molar charge ratio. Comparison between the curves straightforwardly shows that different aggregate structures are present in the solutions depending on the presence of DNA. Data from KIW7 samples are characterized by a linear descent (in log–log representation) at low-q which is followed by a plateau at intermediate range and a new descent at higher q-values, denoting multiple structures across length scales corresponding to the SAXS data. These curves have been successfully fitted using a combination of a simple power law plus a Gaussian coil (GC) form factor. The power law exponent is associated to the slope found at low-q and it accounts for dimensionality of fractal aggregates in solution.31,51 The GC form factor describes smaller structures in solution and it provides the gyration radius of scattering particles and the Flory parameter, which is correlated with interaction between these scatterers and solvent. The low-q scaling exponent has is α = −2.75, corresponding to mass fractals in solution,31,52 and the radius of gyration and Flory parameter are Rg = 0.7 nm and ν = 0.35 respectively, consistent with free peptide chains collapsed in the solvent. These findings correlate well with CD assays which demonstrate that KIW7 secondary structure is dominated by random coil conformations. In the case of KIW7/200 bp DNA samples, the curves have been fitted using a long cylinder shell form factor.35 The fitting parameters are Rc = 5.0 ± 2.8 nm and ΔS = 1.1 nm, respectively, for the core radius and shell thickness, which suggests a biomolecular core surrounded by a counterion shell. It should be noted that sizes found in SAXS assays indicate the presence of rod-like assemblies with total average diameter at 12.2 nm, thus much thicker than DNA duplexes.


image file: d0sm00347f-f4.tif
Fig. 4 (A) SAXS profiles from preparations containing KIW7 at 10 mg ml−1 and mixtures of KIW7/200 bp DNA at 1[thin space (1/6-em)]:[thin space (1/6-em)]1 peptide-to-phosphate charge ratio. (B) Micrograph from vitrified solution containing KIW7 at 10 mg ml−1. (C) cryo-TEM image from KIW7/200 bp DNA solution at 1[thin space (1/6-em)]:[thin space (1/6-em)]1 peptide-to-phosphate charge ratio. (D) and (E) Images from KIW7/200 bp DNA solutions at 2[thin space (1/6-em)]:[thin space (1/6-em)]1 charge ratio.

To further investigate the validity of SAXS fitting parameters, we have performed cryogenic transmission electron microscopy (cryo-TEM). Cryo-TEM is an in situ technique that provides direct-space visualization of vitrified solutions containing peptiplexes at conditions very closeto those used in SAXS assays. Fig. 4B shows a micrograph from KIW7 solution at 10 mg ml−1 revealing only particles with irregular shapes in the samples, in agreement with SAXS data above that suggested the presence of fractal aggregates and unaggregated peptide chains. In Fig. 4C, images from KIW7/DNA at 1[thin space (1/6-em)]:[thin space (1/6-em)]1 peptide-to-phosphate charge ratio display a richer polymorphism with consistent appearance of long unbranched fibers forming an interconnected network. Also, planar tape-like morphologies appear in co-existence with fibers. Images from samples containing KIW7 and DNA at 2[thin space (1/6-em)]:[thin space (1/6-em)]1 charge ratio in Fig. 4D also reveal coexistence between fibrils and elongated tapes. Since fibrillar structures often appear as a continuation along the axis of nanotapes (white arrows), the tape-like assemblies could be due to lateral association between fibrils. At higher magnifications, Fig. 4E, one observes that fibrillar assemblies possess thickness around 7 nm, compatible with sizes of polydisperse rod-like structures revealed by SAXS. Interestingly, lengths of the fibers easily reach a few hundred nanometers, whereas electrophoretic measurements indicate that DNA pieces are much shorter with ∼200 bp (∼67 nm). In this case, 200bp DNA duplexes likely behave as a scaffold directing fibrillization of KIW7 peptide.

The finer structure of peptide/DNA fibers has been investigated through ex situ approaches, namely, AFM and fiber X-rays measurements. To prepare the samples, we used calf thymus DNA with ∼200 bp and plasmid DNA with 4.7 kb (pDNA). The AFM images in Figure clearly reveal the formation of an intricate network of fibers, similar to those revealed by in situ techniques above. The sizes of the fibers are about 20 nm, a bit thicker than those observed in SAXS, probably due to differences in sample preparation between techniques. In Fig. 5B, we observe that pDNA, which is a commonly used vector in gene therapy, also induces structuration of peptides into thin fibers. Overall, the morphology and dimensions of fibers formed by KIW7/pDNA are quite close to those found in samples prepared with 200bp DNA, indicating that a similar structuration pattern is present when long nucleotide sequences are used in the peptiplexes.


image file: d0sm00347f-f5.tif
Fig. 5 (A) AFM topography from KIW7 peptide complexed (A) with 200 bp DNA and (B) with pDNA. The formation of interconnected fiber networks is observed in both samples and the sizes of the structures are found to be around 20 nm, consistent with both cryo-EM and SAXS data. (C) Fiber X-ray diffraction performed on dried stalk from a solution containing KIW7/pDNA complexes. Packing into β-sheets is evidenced by the strong Bragg peak at = 19° [CuKα].

To probe the molecular packing of these fibers, dried stalks were prepared from KIW7/pDNA solutions (1[thin space (1/6-em)]:[thin space (1/6-em)]1 molar ratio) and investigated by X-ray diffraction. The data, shown in Fig. 5C, reveal three Bragg peaks at 2θ = 19°, 20.4 and 21.8°. The peak at 19° corresponds to a repeat distance of d = 4.7 Å, which is ascribed to the separation between hydrogen-bonded peptide strands organized into β-sheets.53 The reflection at 2θ = 20.8° is associated with d = 4.3 Å and it reinforces the presence of β-sheets, since it agrees with the spacing between α-carbons in antiparallel β-sheets.54 The third peak, 2θ = 21.8°, arises from repeats with d = 4.1 Å which is frequent in peptide systems32,55 and usually ascribed to separation between amino acid lateral chains.56 As a whole, X-ray diffraction indicates that fibers exhibit β-sheet structure.

Structural model and self-assembly pathway for KIW7/DNA peptiplexes

From the data presented above, we are now in a position to elaborate a tentative model for describing the structure of KIW7/DNA peptiplexes. The proposed model is represented in Fig. 6A showing a core–shell assembly where DNA double strands serve as a template for condensation of a peptide layer. This templating behavior of DNA agrees with previous observations in virus-like nanostructures,57 and it resembles the spatial arrangement of inverted cylindrical micelles found in DNA/lipid complexes.58 Since KIW7 has cationic residues at both termini, condensation is likely triggered by electrostatic attraction between terminal amino acids and phosphates. In this case, we propose that peptide chains orient perpendicular to the long of the double helix forming an outer surface decorated with basic amino acids at the opposite side of peptide strands that makes fibers soluble. According to this arrangement, thickness of the fibers can be estimated from the sum of two extended heptameric chains (7 × 0.34–2.4 nm) plus the diameter of DNA duplexes (∼2 nm),36 resulting in a size of 6.8 nm, in close agreement with cryo-EM imaging and SAXS data. The model is fully consistent with β-sheet pairing, as shown in Fig. 6B, where hydrophobic effect and H-bonding between adjacent strands hold the assembly together. In addition, this spatial ordering enables π–π interactions between aromatic residues at neighbor peptides. In this case, phenylalanine and tryptophan π-stacks run parallel to the axis of the fiber and convey directionality for self-assembly,59–61 thus complementing the template action exerted by DNA (see Fig. 6C).
image file: d0sm00347f-f6.tif
Fig. 6 Schematic representation of the structural model proposed for complexation between KIW7 and DNA. (A) Fibrillary scaffold with a DNA core (orange ribbons) templating the growth of a peptide shell. (B) β-sheet paring between adjacent KIW7 chains in the fibers. (C) Supramolecular stacks formed by π–π interactions between aromatic residues. (D) Illustration of nucleation and growth of peptiplexes through liquid–liquid phase separation.

The formation of fibrillary peptiplexes can be also viewed in the context of a nucleation-dependent polymerization mechanism.62,63 Recently, nucleation and growth of peptide supramolecular polymers has been described to proceed through liquid–liquid phase separation,62 and such a pathway can be used for explaining complexation between KIW7 and DNA. This process is characterized by the formation of nucleation sites constituted by solute-rich droplets from which the structures growth to further convert to thermodynamically stable nanofibers.62,63 Our proposition for this pathway is schematically illustrated in Fig. 6D, where it appears to be triggered by electrostatic attraction between peptides and DNA. As the next step, association between oppositely charged biomolecules leads to complex coacervation, with liquid–liquid phase separation generating droplets concentrated in DNA and peptides. In the following, these coacervates act as a nucleation center from which elongation of fibers is sustained by coalescence of clusters and incorporation of free biomolecules from bulk solution. Finally, the formation of mature fibers proceeds through usual Ostwald ripening with transient intermediates being added to thermodynamically stable long fibers.62,63

Cell assays: KIW7 promotes cytosolic internalization of DNA fragments

Since KIW7 is a fragment derived from Penetratin, a further step in our study has been to probe the interaction of this peptide with cell membranes and tentatively demonstrate its capacity to assist DNA delivery into HeLa cells. MTT assays have been performed for assessing cell viability in the presence of different concentrations of KIW7. These tests have been carried out using cells incubated during 4 hours with the peptide, the same incubation time used in delivery assays below, in agreement with protocols found elsewhere for Penetratin.64 Also, 4 hours is a period longer than Penetratin plasma half-life observed in vivo.65 Our results show that within this time lapse KIW7 is well tolerated up to concentrations of 4 mg ml−1, indicating low cytotoxicity (ESI, Fig. S3). In the following, we have prepared peptiplexes using KIW7 peptide and 200 bp DNA labelled with YOYO-1, a cell-impermeable dye that becomes highly fluorescent upon intercalation to DNA strands.33,66 Therefore, by monitoring YOYO-1 fluorescence, we have been able to track the presence of exogenous DNA in cell culture. We incubated HeLa cells with KIW7/200 bp DNA peptiplexes at a DNA concentration of 5 μg per well at 1[thin space (1/6-em)]:[thin space (1/6-em)]1 or 2[thin space (1/6-em)]:[thin space (1/6-em)]1 molar ratio between peptide and phosphate charges, respectively. After 4 hours incubation, plates were rinsed to wash out unbonded 200 bp DNA, and cells were fixed with paraformaldehyde for confocal microscopy assays. Two fluorescence channels have been investigated, one corresponding to DAPI blue-fluorescent dye staining nuclei, and the other associated to YOYO-1 green emission.

As shown in Fig. 7, intense green fluorescence is observed throughout cells incubated with KIW7/200 bp DNA complexes prepared at 2[thin space (1/6-em)]:[thin space (1/6-em)]1 molar ratio, indicating cell surface association of 200 bp DNA duplexes and showing that binding to cell membrane is facilitated when peptides are used in the formulation. When only labelled 200 bp DNA is present in culture, the greener glow from YOYO-1 emission is not detected, indicating a complete (or almost complete) removal of non-specific binding.


image file: d0sm00347f-f7.tif
Fig. 7 Confocal fluorescence images from HeLa cells incubated with KIW7/200 bp DNA complexes at 2[thin space (1/6-em)]:[thin space (1/6-em)]1 peptide-to-phosphate charge ratio. DNA strands are stained with green YOYO-1 dye, whereas blue-fluorescent DAPI stains nuclei.

In addition to association to cell surface, some degree of internalization is evidenced by the presence of exogenous DNA in the surroundings of nuclear membranes, as indicated by proximity between DAPI and YOYO-1 fluorescence, which denotes that 200 bp DNA is anchored at eukaryotic barriers (bottom rows in Fig. 7). To further investigate the distribution of DNA in the cells, we recorded images from different focal planes along the z-direction. The image series, shown in ESI (Video 1), reveals that fluorescence from 200 bp DNA is found at different confocal planes, indicating that at least part of the nucleic acid payload enters cells when peptiplexes are formulated with KIW7. Although the data presented above do not allow for discriminating the pathways involved in the uptake, they indicate that binding of DNA to cell membranes, followed by penetration into cytosol, is assisted by peptiplexes based on KIW7. Interaction with cells has also been noticed for peptiplexes prepared at 1[thin space (1/6-em)]:[thin space (1/6-em)]1 charge ratio, however, at much lower intensities in comparison to samples where peptide excess was present (see ESI, Fig. S5) which suggests that excess of cationic charges is relevant for cell binding and penetration.

Putting the findings above together, we observe that KIW7 might be a suitable vector for preparing noncovalent complexes with DNA, where it could assist gene expression and regulate metabolism. Our data demonstrate that this short fragment containing the largest non-cationic portion of Penetratin promotes interaction with cell membranes for internalization, and potentially represents a cheaper and simpler CPP for cytosolic delivery of nucleic acids.

Conclusions

We have performed detailed investigations on both structure and cell membrane interaction of a heptamer derived from Penetratin, an archetypical Trojan peptide of great interest for biomedical research. The peptide has been successfully used to form peptiplexes with DNA, self-assembling into long nanofibers and nanotapes endowed with clear β-sheet features. The supramolecular structure of nanofibers likely comprises DNA cores surrounded by a peptide shell for which the double helix behaves as a template and induces fibrillization. A nucleation and growth mechanism proceeding through liquid–liquid phase separation of coacervates is proposed for describing the self-assembly of peptiplexes. We have demonstrated that KIW7 is able to promote cell surface association and partially deliver DNA with ∼200 bp into HeLa cells, indicating the potential of this truncated version of Penetratin for the transport of nucleotide sequences larger than those successfully transported by CPPs in recent reports.67,68 In fact, 200 bp DNA duplexes internalized with assistance of KIW7 are about 10 times larger than microRNA and siRNA used to regulate cell metabolism and therefore KIW7 is potentially suitable for delivery of these species or even oligonucleotides.

Since KIW7 contains the longest non-cationic segment found in the Penetratin strand, it also represents an interesting model to probe the contribution of uncharged amino acids for overall structure of Penetratin. The capacity of the peptide to interact with anionic micelles was demonstrated even upon charge screening promoted by complexation with DNA, evidencing that hydrophobicity provided by IWF tripeptides in the middle of the strands are a key factor for interaction with biological membranes. In addition, KIW7 conserves main features of Penetratin secondary structure, including random-to-β transitions upon interaction with anionic surfactant interfaces, thus suggesting that structuration of Penetratin in the presence of biomembranes (a key feature for translocation capacity) is closely correlated to the uncharged residues preserved in the truncated peptide. Although the role of non-polar amino acids for bioactivity of Trojan peptides, specially tryptophan residues, has received increasingly attention in recent years,20,35,69 the importance of cationic residues has been the main focus. In this context, the findings presented here undoubtably demonstrate that non-cationic amino acids are directly implicated in cell uptake and, even in the absence of strong cationicity, they make intracell delivery of DNA fragments feasible.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This project has been supported by FAPESP under process no. 2016/24409-3 granted to ERS, who is also a fellow of the National Council for Scientific and Technological Development – CNPq (Proc. no. 310916/2019-4). LRM has been supported by a doctoral fellowship from CAPES-Proex granted to the Graduate Program in Molecular Biology of UNIFESP (Finance Code 001). SAXS data have been collected at the Brazilian National Synchrotron Laboratory-LNLS under proposal 20170958. IWH and VC acknowledge EPSRC for Platform grant EP/L020599/1. This work has been partially supported by the UNIFESP CAPES-PrInt program in the theme “Development of treatment and therapeutics for inflammatory diseases”. WAA is supported by CNPq (grant no. 304389/2019-6), the National Institute of Science and Technology in Bioanalytics (FAPESP grant no. 2014/50867-3 and CNPq grant no. 465389/2014-7), and FAPESP (grant no. 2015/24018-1, 2017/02317-2). AFM and cryo-TEM experiments have been conducted at the National Nanotechnology Laboratory – LNNano under proposals AFM-36338 and TEM-24500. We are very grateful to Dr. Florian Meneau (SAXS1-LNLS), Dr. Carlos Costa (AFM-LNNano) and Dr. Antonio Borges (TEM-LNNano) for kind assistance during data collection.

References

  1. S. Guha, J. Ghimire, E. Wu and W. C. Wimley, Chem. Rev., 2019, 119, 6040–6085 CrossRef CAS PubMed.
  2. I. W. Hamley, Chem. Rev., 2017, 117, 14015–14041 CrossRef CAS PubMed.
  3. K. Kurrikoff and U. Langel, Exp. Opin. Drug Delivery, 2019, 16, 1183–1191 CrossRef CAS PubMed.
  4. A. D. Frankel and C. O. Pabo, Cell, 1988, 55, 1189–1193 CrossRef CAS PubMed.
  5. A. Joliot, C. Pernelle, H. Deagostini-Bazin and A. Prochiantz, Proc. Natl. Acad. Sci. U. S. A., 1991, 88, 1864–1868 CrossRef CAS PubMed.
  6. D. Derossi, A. H. Joliot, G. Chassaing and A. Prochiantz, J. Biol. Chem., 1994, 269, 10444–10450 CAS.
  7. J. S. Bahnsen, H. Franzyk, E. J. Sayers, A. T. Jones and H. M. Nielsen, Pharm. Res., 2015, 32, 1546–1556 CrossRef CAS PubMed.
  8. B. Raudszus, D. Mulac and K. Langer, Int. J. Pharm., 2018, 536, 211–221 CrossRef CAS PubMed.
  9. C. Liu, K. Jiang, L. Tai, Y. Liu, G. Wei, W. Lu and W. Pan, ACS Appl. Mater. Interfaces, 2016, 8, 19256–19267 CrossRef CAS PubMed.
  10. X. Xie, J. E. Kerrigan, T. Minko, O. Garbuzenko, K. C. Lee, A. Scarborough, E. E. Abali, T. Budak-Alpdogan, N. Johnson-Farley, D. Banerjee, K. W. Scotto and J. R. Bertino, Cancer Biol. Ther., 2013, 14, 742–751 CrossRef CAS PubMed.
  11. T. Yin, W. Xie, J. Sun, L. Yang and J. Liu, ACS Appl. Mater. Interfaces, 2016, 8, 19291–19302 CrossRef CAS PubMed.
  12. N. Skrlj, G. Drevensek, S. Hudoklin, R. Romih, V. Curin Serbec and M. Dolinar, Appl. Biochem. Biotechnol., 2013, 169, 159–169 CrossRef CAS PubMed.
  13. K. Jiang, X. Gao, Q. Shen, C. Zhan, Y. Zhang, C. Xie, G. Wei and W. Lu, Acta Biomater., 2017, 63, 123–134 CrossRef CAS PubMed.
  14. L. Tai, C. Liu, K. Jiang, X. Chen, G. Wei, W. Lu and W. Pan, Nanomedicine, 2017, 13, 2091–2100 CrossRef CAS PubMed.
  15. X. Q. Xie, N. Bansal, T. Shaik, J. E. Kerrigan, T. Minko, O. Garbuzenko, E. E. Abali, N. Johnson-Farley, D. Banerjee, K. W. Scotto and J. R. Bertino, Oncotarget, 2014, 5, 901–907 Search PubMed.
  16. I. Nakase, H. Akita, K. Kogure, A. Graslund, U. Langel, H. Harashima and S. Futaki, Acc. Chem. Res., 2012, 45, 1132–1139 CrossRef CAS PubMed.
  17. C. J. Chene and W. M. Saltzman, Biomaterials, 2011, 32, 6194–6203 CrossRef PubMed.
  18. D. Kalafatovic and E. Giralt, Molecules, 2017, 22, 1929 CrossRef PubMed.
  19. E. Dupont, A. Prochiantz and A. Joliot, Cell-Penetrating Peptides: Methods and Protocols, ed. Ü. Langel, Humana Press, Totowa, NJ, 2011, pp. 21–29 Search PubMed.
  20. P. Gehan, V. Vivier, K. Ngo, S. Sagan, A. Walrant, S. Cribier and N. Rodriguez, Biophys. J., 2020, 118, 383a CrossRef.
  21. I. D. Alves, N. Goasdoué, I. Correia, S. Aubry, C. Galanth, S. Sagan, S. Lavielle and G. Chassaing, Biochim. Biophys. Acta, Gen. Subj., 2008, 1780, 948–959 CrossRef CAS PubMed.
  22. C.-Y. Jiao, E. Sachon, I. D. Alves, G. Chassaing, G. Bolbach and S. Sagan, Angew. Chem., Int. Ed., 2017, 56, 8226–8230 CrossRef CAS PubMed.
  23. H. Binder and G. Lindblom, Biophys. J., 2003, 85, 982–995 CrossRef CAS PubMed.
  24. H. Binder and G. Lindblom, Biophys. J., 2004, 87, 332–343 CrossRef CAS PubMed.
  25. G. Drin, H. Déméné, J. Temsamani and R. Brasseur, Biochemistry, 2001, 40, 1824–1834 CrossRef CAS PubMed.
  26. W. B. Kauffman, T. Fuselier, J. He and W. C. Wimley, Trends Biochem. Sci., 2015, 40, 749–764 CrossRef CAS PubMed.
  27. E. Eiríksdóttir, K. Konate, Ü. Langel, G. Divita and S. Deshayes, Biochim. Biophys. Acta, Biomembr., 2010, 1798, 1119–1128 CrossRef PubMed.
  28. C. Bechara and S. Sagan, FEBS Lett., 2013, 587, 1693–1702 CrossRef CAS PubMed.
  29. H. A. Rydberg, M. Matson, H. L. Åmand, E. K. Esbjörner and B. Nordén, Biochemistry, 2012, 51, 5531–5539 CrossRef CAS PubMed.
  30. M.-L. Jobin, M. Blanchet, S. Henry, S. Chaignepain, C. Manigand, S. Castano, S. Lecomte, F. Burlina, S. Sagan and I. D. Alves, Biochim. Biophys. Acta, Biomembr., 2015, 1848, 593–602 CrossRef CAS PubMed.
  31. E. R. Silva, E. Listik, S. W. Han, W. A. Alves, B. M. Soares, M. Reza, J. Ruokolainen and I. W. Hamley, Biophys. Chem., 2018, 233, 1–12 CrossRef CAS PubMed.
  32. C. C. Decandio, E. R. Silva, I. W. Hamley, V. Castelletto, M. S. Liberato, V. X. Oliveira, C. L. P. Oliveira and W. A. Alves, Langmuir, 2015, 31, 4513–4523 CrossRef CAS PubMed.
  33. E. R. T. da Silva, E. A. de Oliveira, A. Fevrier, F. Nallet and L. Navailles, Eur. Phys. J. E: Soft Matter Biol. Phys., 2011, 34, 83 CrossRef PubMed.
  34. E. A. de Oliveira, E. R. T. da Silva, A. Fevrier, E. Grelet, F. Nallet, L. Navailles, E. A. de Oliveira, E. R. T. da Silva, A. Février, G. É. F. Nallet and L. Navailles, EPL, 2010, 91, 28001 CrossRef.
  35. E. R. Silva, G. Cooney, I. W. Hamley, W. A. Alves, S. Lee, B. F. O’Connor, M. Reza, J. Ruokolainen and D. Walls, Soft Matter, 2016, 12, 9158–9169 RSC.
  36. H. H. Strey, R. Podgornik, D. C. Rau and V. A. Parsegian, Curr. Opin. Struct. Biol., 1998, 8, 309–313 CrossRef CAS PubMed.
  37. D. W. Sambrook and J. Russell, Molecular cloning: a laboratory manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 3rd edn, 2001 Search PubMed.
  38. B. M. Soares, A. M. Aguilar, E. R. Silva, M. D. Coutinho-Neto, I. W. Hamley, M. Reza, J. Ruokolainen and W. A. Alves, Phys. Chem. Chem. Phys., 2017, 19, 1181–1189 RSC.
  39. J. R. Lakowicz, Principles of Fluorescence Spectroscopy, Springer, New York, 4th edn, 2006 Search PubMed.
  40. D. V. Tulumello and C. M. Deber, Biochemistry, 2009, 48, 12096–12103 CrossRef CAS PubMed.
  41. M. H. Shabestari, N. J. Meeuwenoord, D. V. Filippov and M. Huber, J. Biol. Phys., 2016, 42, 299–315 CrossRef CAS PubMed.
  42. J.-M. Lin, T.-L. Lin, U.-S. Jeng, Z.-H. Huang and Y.-S. Huang, Soft Matter, 2009, 5, 3913–3919 RSC.
  43. A. Wahlström, L. Hugonin, A. Perálvarez-Marín, J. Jarvet and A. Gräslund, FEBS J., 2008, 275, 5117–5128 CrossRef PubMed.
  44. K. Kilk, M. Magzoub, M. Pooga, L. E. G. Eriksson, Ü. Langel and A. Gräslund, Bioconjugate Chem., 2001, 12, 911–916 CrossRef CAS PubMed.
  45. L. R. De Mello, I. W. Hamley, V. Castelletto, B. B. M. Garcia, S. W. Han, C. L. P. De Oliveira and E. R. Da Silva, J. Phys. Chem. B, 2019, 123, 8861–8871 CrossRef CAS PubMed.
  46. B. Nordén, A. Rodger, T. Dafforn and Royal Society of Chemistry (Great Britain), Linear dichroism and circular dichroism: a textbook on polarized-light spectroscopy, Royal Society of Chemistry, Cambridge, 2010 Search PubMed.
  47. M. J. Krysmann, V. Castelletto, A. Kelarakis, I. W. Hamley, R. A. Hule and D. J. Pochan, Biochemistry, 2008, 47, 4597–4605 CrossRef CAS PubMed.
  48. J. Ye, S. A. Fox, M. Cudic, E. M. Rezler, J. L. Lauer, G. B. Fields and A. C. Terentis, J. Am. Chem. Soc., 2010, 132, 980–988 CrossRef CAS PubMed.
  49. H. L. Åmand, H. A. Rydberg, L. H. Fornander, P. Lincoln, B. Nordén and E. K. Esbjörner, Biochim. Biophys. Acta, Biomembr., 2012, 1818, 2669–2678 CrossRef PubMed.
  50. E. R. da Silva, W. A. Alves, V. Castelletto, M. Reza, J. Ruokolainen, R. Hussain and I. W. Hamley, Chem. Commun., 2015, 51, 11634–11637 RSC.
  51. T. Zemb and P. Lindner, Neutrons, X-rays and light: scattering methods applied to soft condensed matter, Elsevier, Amsterdam; Boston, 1st edn, 2002 Search PubMed.
  52. G. Beaucage, J. Appl. Crystallogr., 1996, 29, 134–146 CrossRef CAS.
  53. K. E. Marshall and L. C. Serpell, Biochem. Soc. Trans., 2009, 37, 671–676 CrossRef CAS PubMed.
  54. A. Baumketner, S. L. Bernstein, T. Wyttenbach, N. D. Lazo, D. B. Teplow, M. T. Bowers and J.-E. Shea, Protein Sci., 2006, 15, 1239–1247 CrossRef CAS PubMed.
  55. O. S. Makin and L. C. Serpell, J. Mol. Biol., 2004, 335, 1279–1288 CrossRef PubMed.
  56. P. Sikorski, E. D. T. Atkins and L. C. Serpell, Structure, 2003, 11, 915–926 CrossRef CAS PubMed.
  57. Y. Ruff, T. Moyer, C. J. Newcomb, B. Demeler and S. I. Stupp, J. Am. Chem. Soc., 2013, 135, 6211–6219 CrossRef CAS PubMed.
  58. I. Koltover, T. Salditt, J. O. Radler and C. R. Safynia, Science, 1998, 281, 78–81 CrossRef CAS PubMed.
  59. Y. Lin, Y. Qiao, P. Tang, Z. Li and J. Huang, Soft Matter, 2011, 7, 2762–2769 RSC.
  60. J. Li, R. Xing, S. Bai and X. Yan, Soft Matter, 2019, 15, 1704–1715 RSC.
  61. C. Yuan, W. Ji, R. Xing, J. Li, E. Gazit and X. Yan, Nat. Rev. Chem., 2019, 3, 567–588 CrossRef CAS.
  62. C. Yuan, A. Levin, W. Chen, R. Xing, Q. Zou, T. W. Herling, P. K. Challa, T. P. J. Knowles and X. Yan, Angew. Chem., Int. Ed., 2019, 58, 18116–18123 CrossRef CAS PubMed.
  63. J. Wang, K. Liu, R. Xing and X. Yan, Chem. Soc. Rev., 2016, 45, 5589–5604 RSC.
  64. B. Christiaens, J. Grooten, M. Reusens, A. Joliot, M. Goethals, J. Vandekerckhove, A. Prochiantz and M. Rosseneu, Eur. J. Biochem., 2004, 271, 1187–1197 CrossRef CAS PubMed.
  65. E. J. B. Nielsen, S. Yoshida, N. Kamei, R. Iwamae, E. S. Khafaby, J. Olsen, U. L. Rahbek, B. L. Pedersen, K. Takayama and M. T. Morishita, J. Controlled Release, 2014, 189, 19–24 CrossRef CAS PubMed.
  66. B. Kundukad, J. Yan and P. S. Doyle, Soft Matter, 2014, 10, 9721–9728 RSC.
  67. W. B. Kauffman, S. Guha and W. C. Wimley, Nat. Commun., 2018, 9, 2568 CrossRef PubMed.
  68. S. Tarvirdipour, C.-A. Schoenenberger, Y. Benenson and C. G. Palivan, Soft Matter, 2020, 16, 1678–1691 RSC.
  69. A. Walrant, A. Bauzá, C. Girardet, I. D. Alves, S. Lecomte, F. Illien, S. Cardon, N. Chaianantakul, M. Pallerla, F. Burlina, A. Frontera and S. Sagan, Biochim. Biophys. Acta, Biomembr., 2020, 1862, 183098 CrossRef PubMed.

Footnote

Electronic supplementary information (ESI) available. See DOI: 10.1039/d0sm00347f

This journal is © The Royal Society of Chemistry 2020