Coiled Coil Exposure and Histidine Tags Drive Function of an Intracellular Protein Drug Carrier

Abstract

In recent years, protein engineering efforts have yielded a diverse set of binding proteins that hold promise for various therapeutic applications. Despite this, their inability to reach intracellular targets limits their applications to cell surface or soluble targets. To address this challenge, we previously reported a protein carrier that binds antibodies and delivers them to therapeutic targets inside cancer cells. This carrier, known as the Hex carrier, is comprised of a self-assembling coiled coil hexamer at the core, with each alpha helix fused to a linker, an antibody binding domain, and a six Histidine-tag (His-tag). In this work, we designed different versions of the carrier to determine the role of each building block in cytosolic protein delivery. We found that increasing exposure of the Hex coiled coil on the carriers, through molecular design or removing antibodies, increased internalization, pointing to a role of the coiled coil in promoting endocytosis. We observed a clear increase in endosomal disruption events when His-tags were present on the carrier relative to when they were removed, due to an endosomal buffering effect. Finally, we found that the antibody binding domains of the Hex carrier could be replaced with monomeric ultra-stable GFP for intracellular delivery and endosomal escape. Our results demonstrate that the Hex coiled coil, in conjunction with His-tags, could be a generalizable vehicle for delivering small and large proteins to intracellular targets. This work also highlights new biological applications for oligomeric coiled coils and shows the direct and quantifiable impact of histidine residues on endosomal disruption. These findings could inform the design of future drug delivery vehicles in applications beyond intracellular protein delivery.

Graphical Abstract

Introduction

Despite the progress in drug delivery technologies for small molecules and nucleic acids, intracellular delivery of proteins has remained a challenge. Many therapeutic protein drugs, such as antibodies, are too large and hydrophilic to diffuse across cell membranes or too small to be taken up by endocytosis [1]. While efforts have been made to increase intracellular protein delivery, technologies to load proteins into nanoparticle drug delivery systems face challenges of low drug loading, endosomal entrapment, or loss of functional protein activity [2,3]. The need remains for protein delivery to intracellular targets, which could significantly expand therapeutic options for many currently undruggable targets.

A large number of polymer, metallic, inorganic and lipid carriers have been developed to address this need [4-7]. Although some of these carriers achieve highly efficient intracellular protein delivery, each has inherent limitations that can conflict with distinct biochemical properties of particular proteins or the dosing and delivery strategies needed for specific diseases. Some carriers require chemical modification or genetic engineering of the native proteins [8-10], which may alter the protein activity or toxicity profile. Alternative approaches can achieve cytosolic delivery of a wide range of proteins without the need for chemical or biological modifications. However, these approaches rely on the noncovalent interactions between carriers and proteins, which can reduce control over binding affinity and binding site of the protein [11-15]. For those carriers, the destabilization of proteins is still a concern due to the cooperative binding of carrier’s binding groups to the protein surface [16,17]. While some studies provide insights to serum protein binding and corona formation of these carriers [18-21], the serum stability of carrier-protein complexes remains poorly investigated. The in vivo stability of these carriers has yet to be verified. Though some delivery systems can encapsulate native proteins into the nanoparticle, the loading efficiency is low and homogeneous protein loading is difficult to reach [22,23]. As a result, there is still demand for an effective, convenient, and robust approach for cytosolic delivery of native proteins without chemical or biological modifications. In some cases, the carrier material itself can be difficult to synthesize and may not be biodegradable. Protein and peptide-based carriers are an attractive alternative for cytosolic protein delivery. In general, protein carriers offer the advantages of biocompatibility and biodegradability. The building blocks can be produced recombinantly and the carriers prepared under mild conditions without the use of toxic chemicals or organic solvents. Moreover, some protein carriers can provide precise control over the carrier's sequence, self-assembly, number of protein cargo loaded, bioactivity, stimuli-responsiveness, and polydispersity.

While peptides, including those with alpha helical structure, have been explored as delivery tools for their cell penetration capabilities [24], coiled coil oligomers have yet to be utilized for such applications. Since the discovery of HIV-TAT peptide, many other cell penetrating peptides (CPPs) have been identified. Often their natural function is to promote cell entry of bacteria or viruses and they can be fused to therapeutic proteins to deliver them intracellularly. Some challenges with CPPs that limit their translational benefit are toxicity to cells and low efficiency of cellular uptake and endosomal escape [25]. One strategy that has been used to improve delivery efficiency is using a helix-loop-helix CPP structure or multimers of this structure. Tetramers of the helix-loop-helix amphipathic bundles showed rapid internalization of GFP at nanomolar concentrations[26]. Helix-loop-helix motifs are also often found in coiled coil proteins, which have been identified as key internalization motifs for viral and bacterial proteins. Coiled coil domains in Salmonella effector proteins were found to facilitate membrane association of proteins to promote their internalization [27]. Additionally, the HIV Env protein requires a rearrangement of its subunit gp41 into a six-helix bundle in order to promote membrane fusion and the internalization of HIV virus into host cells [28]. This was found to be the rate limiting step in membrane fusion; without gp41 six helix bundle formation, fusion and pore formation were inhibited. Given the biological role of some coiled coils and the delivery benefits of helix-turn-helix peptide fusions, coiled coil domains may serve as an internalization motif for therapeutic protein delivery.

In addition to uptake efficiency, another significant challenge for protein delivery technologies is escape from endosomes to reach cytosolic targets. Many strategies have been explored for this including the use of histidine residues to create an endosomal buffering effect [29,30]. This effect, also known as the proton sponge effect, occurs when a high concentration of positive charge in an endosome leads to osmotic swelling, rupture, and release of endosome components [31,32]. Histidine residues on proteins are often utilized as purification tags but also are positively charged at endosomal pH and are thought to promote endosomal release [32]. In a biological context, histidine residues in viral fusion proteins are thought to play a role in viral fusion through both structural rearrangements of fusion proteins and protonation of histidines in endosomes [33]. Drug delivery and activity of peptides, virus-like particles, and nucleic acids has previously been enhanced by including histidine residues or imidazole groups in protein, polymer, or lipid nanoparticles [34-42]. In one case, six histidines, the number commonly used in protein purification, were found to deliver GFP to the cytosol of cells better than 3, 9, or 12 residues [43]. Altogether, this suggests that 6x histidine (6x-His) tags may play a role in cytosolic delivery of therapeutic proteins.

Previously, we developed a self-assembled hexameric protein drug carrier to deliver antibodies to intracellular targets [44]. The Hex carrier consists of a hexameric coiled coil at the center, and each coiled coil is fused to a glycine-serine linker, an IgG Fc binding domain (Staphylococcus aureus protein A domain B, SPAB), and a 6x-His tag (Fig. 1A). All six SPAB domains bind a total of three antibodies, forming Hex-IgG complexes with each antibody bound by two SPAB domains (Fig. 1B) [45]. We demonstrated delivery of antibodies to intracellular targets including microtubules, nuclear pore complex, and STAT3 transcription factor, showing in each case that biological activity of the antibodies was maintained after delivery with Hex. [44,46]. In this work, we engineered different variants of the Hex carrier to determine the effect of coiled coils and 6x-His tags on cellular uptake and endosomal escape, to improve intracellular protein delivery, and determine if Hex cargo was restricted to IgG. We hypothesized that the hexameric coiled coils could assist cellular uptake or endosomal escape through non-destructive cell membrane interactions given their amphiphilic nature and similarity to pathogenic delivery motifs. However, the interaction of coiled coils and cell membrane may be hindered due to the fused SPAB domains or loaded antibody cargo, thus not fully benefiting delivery. We also hypothesized that the six 6x-His tags could induce strong osmotic pressure and provide a mechanism for the endosomal disruption we observed previously. To test these hypotheses, we designed Hex variants to modulate the exposure of the coiled coils and remove the 6x-His tags. In doing so, we determined that the Hex coiled coil could be a general platform for delivery of other proteins with a wide range of sizes. To our knowledge, the endosomal disruption capabilities of histidine residues have not been directly and quantitatively measured, as previous examples determined endosomal escape through qualitative methods of colocalization with endolysosomes in confocal microscope images or an indirect readout of successful drug cargo delivery. [40-42,47,48].

Figure 1.

Orientation of different versions of Hex carriers. (A) Representative schematic of Hex carrier with balanced orientation of SPAB domains (PDB ID: 1BDC), three on either side of the Hex coiled coil core (PDB ID: 3R47) connected by glycine-serine linkers. (B) Schematic of balanced Hex carrier bound to three antibodies, drawn to scale. (C) Schematic of HS protein comprised of the Hex coiled coil core and six SPAB domains on one side. (D) Schematic of SHS protein which contains twelve SPAB domains. (A, C, D) are not drawn to scale.

The current studies were designed to broadly investigate the relationship of Hex variants’ structure and bioactivity to establish the hexameric coiled coil fusion protein carrier as a potential platform for the intracellular delivery of proteins. Our use of recombinant fusion proteins to create intracellular protein nanocarriers enables modular design for both understanding and optimizing delivery functions. The ability to precisely change protein sequences and manipulate the spatial organization of Hex nanocarriers and the domains contained within the carriers enabled the testing of delivery hypotheses. This ultimately led to improved protein delivery and to design guidelines that can be applied not only to Hex but other types of protein, or even synthetic, carriers.

Materials and Methods

Materials:

Rabbit immunoglobulin G (IgG) was purchased from Sigma-Aldrich. AlexaFluor™ 488 5-SDP Ester was purchased from ThermoFisher Scientific. Tobacco Etch Virus (TEV) protease, ampicillin, chloramphenicol, and Isopropyl-β-thiogalactoside were purchased from Sigma-Aldrich.

Cell Lines:

HeLa cell line and MDA-MB-231 cell line were purchased from American Type Culture Collection (ATCC). Yellow fluorescent protein-Galectin-8 (YFP-Gal8) MDA-MB-231 cell line was generously provided by Professor Craig Duvall from Vanderbilt University. All cell lines were cultured in Dulbecco’s Modified Eagle Medium (Corning) supplemented with 10% v/v fetal bovine serum (FBS, VWR International) and 5% penicillin/streptomycin (Gibco) at 37°C and 5% CO₂ humidified atmosphere.

Protein Expression and Purification:

The Hex carrier was produced by separate production of two proteins: Hex-SPAB (HS) and SPAB-Hex (SH). The genes for HS and SH were previously cloned into pQE80 vectors and transformed into Top 10 Escherichia coli [44]. The genes for SPAB-Hex-SPAB, His-tag cleavable TEV-HS, and muGFP-Hex were synthesized and cloned into pQE-60 vectors by Genscript and transformed into a BL21-AFIQ strain of E. coli. Proteins were expressed in Lysogeny broth media at 37°C in 1 L cultures supplemented with 0.2 mg/mL ampicillin for HS and SH or 0.2 mg/mL ampicillin and 0.034 mg/mL chloramphenicol for SHS, TEV-HS, and muGFP-Hex. At an optical density of 0.6, protein production was induced by 1 mM isopropyl-β-thiogalactoside. The cells were grown for an additional 4 hours after induction. After collecting cells by centrifugation, all proteins were purified under native conditions by Ni-NTA (Qiagen) affinity chromatography using the manufacturer’s protocol. A concentration of 80 mM imidazole was used in the wash buffer and 250 mM imidazole for elution. TEV-HS was buffer exchanged with Tris buffer (pH 8.0), then treated with TEV enzyme and passed through Ni-NTA column to remove cleaved His-tag, TEV enzyme, and uncleaved TEV-HS according to manufacturer specifications, creating the no-His HS protein (nhHS).The gene for sfGFP-ZE was synthesized and cloned into a pET-28a(+) vector by Genscript and expressed and purified according to a previously published protocol [49]. Purity of all proteins was analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) as shown in Supplementary Fig. S1. Removal of His-tags on nhHS was confirmed by western blot with anti-penta-His AlexaFluor 488 antibody (Qiagen), as shown in Supplementary Fig. S1E. Purified proteins were stored in elution buffer containing 250 mM imidazole at 4 °C for no longer than 2 months.

Sample Preparation:

To produce Hex carriers with a balanced orientation of SPAB domains on either side of the coiled coil core, SH and HS proteins were “disassembled” and “reassembled” as previously reported [44,45]. HS and SH (0.21 mg each) were combined with 8 μL of a 20% w/v solution of sodium dodecyl sulfate (SDS) and phosphate buffered saline (PBS) to a final volume of 2.5 mL. The solution was passed through a PD-10 desalting column (GE Healthcare) to remove the SDS. The resulting Hex solution was concentrated using centrifugal filters (3 kDa MWCO, MilliporeSigma), to a final volume of 0.5 to 1 mg/mL, and sterile filtered with a 0.2 μm membrane filter. The concentration was measured using Protein A280 settings on a NanoDrop 2000 (ThermoFisher). For HS, SHS, nhHS, muGFP-Hex, sfGFP-ZE, purified protein solutions were buffer exchanged into PBS using PD-10 desalting columns and concentrated as described above for Hex. For internalization studies, rabbit serum IgG or HS, Hex, and SHS were labeled with AlexaFluor™ 488 5-SDP Ester using manufacturer’s instructions. Labeling reactions were optimized to yield a degree of labeling of ~1.0 for all proteins.

Protein Assembly Characterization:

HS, Hex, SHS, nhHS, muGFP-Hex, and sfGFP-ZE were analyzed by dynamic light scattering to measure hydrodynamic size of protein assemblies and to ensure no aggregation of stored proteins occurred prior to use in cell culture studies. Measurements were taken with a Malvern Zetasizer Nano ZS equipped with a 4 mW He–Ne laser light source (633 nm). The following solvent settings were used to acquire size (Z-average, Z-ave) and polydispersity index (PDI) for each sample: medium PBS, refractive index 1.33, viscosity 0.8872 cP, measurement temperature 25 °C, cuvette type ZEN 0040, laser wavelength 633 nm, and scattering angle 173 degrees. Three separate readings were obtained for each sample and intensity plots, particle sizes, and polydispersity index were averaged prior to reporting. HS, SHS, and muGFP-Hex were also analyzed by size exclusion chromatography (SEC) to determine the uniformity of assembly formation. Briefly, samples were analyzed by a Malvern OmniSEC integrated system (Malvern Panalytical) with a SRT SEC-300 analytical SEC column (Sepax) for SHS, and a Superdex 200 Increase 10/300GL column (Cytiva Life Sciences) for HS and muGFP-Hex. The UV signal at 254 nm was used to quantify peak area % for each sample.

Internalization Studies of Hex Carriers:

For all internalization studies with fluorescently labeled protein carriers, cells (HeLa or MDA-MB-231) were plated in a 96-well plate at a density of 10,000 cells per well and incubated overnight. Treatment solutions for delivery experiments were prepared by diluting fluorescently labeled Hex carriers or IgG into DMEM + 10% FBS at their respective concentrations and incubating for 5 mins at room temperature. Media was then replaced by treatment solution, and cells were incubated for 24 hours. After 24 hours, cells were removed from plates by trypsin digestion and analyzed by a CytoFLEX flow cytometer (Beckman Coulter). Trypan blue was added to each sample to quench extracellular fluorescence of uninternalized carriers. Flow cytometry data was analyzed using CytExpert software.

MTT and Fluorescence assays:

To evaluate the effect of different Hex carrier proteins and endocytosis inhibitors on viability of HeLa cells, we used an MTT assay (Biotium) for measuring cell metabolic activity as an indicator of cell viability. After incubating cells with protein carriers for 24 hours or endocytosis inhibitors for 1 hour, MTT assay was conducted following manufacturer’s instructions. Absorbance at 570 nm and 630 nm of cell samples was measured by a Synergy 2 plate reader (Biotek). Background absorbance of 630 nm was subtracted from each well and all samples were normalized to the PBS-treated cells to calculate % metabolic activity. To determine the difference in fluorescence of muGFP-Hex and sfGFP-Hex, samples were diluted to various concentrations and fluorescence was analyzed with an excitation wavelength of 488 nm and an emission wavelength of 525 nm, using a Synergy 2 plate reader (Biotek).

Gal8 Recruitment Assay:

The frequency of endosomal disruption by HS, Hex, SHS with and without IgG was investigated by the Gal8 recruitment assay [50]. In brief, YFP-Gal8 MDA-MB-231 cells were seeded into 4-chamber dishes at a density of 10,000 cells/well and incubated overnight. The cells were then treated with HS, Hex, or SHS with and without IgG for 24 h. After incubation, the cells were washed three times with PBS and cell nuclei stained with 16.23 μM Hoechst 33342 for 10 min. Cells were then imaged in complete growth medium using a spinning disc confocal fluorescence microscope (PerkinElmer UltraVIEW VoX). Gain settings were set to be constant for all images acquired. All images were processed using Fiji imageJ using a previously developed processing method [46]. The general workflow for image processing is depicted in Supplementary Fig. S2.

Microscopy Section

For cellular uptake experiments, 1 × 10³ HeLa cells per well were seeded in 35 mm 4 chamber glass bottom dishes (Thermo Fisher Scientific Inc.) at 37 °C for 24 h before treatment. Red fluorescent protein (RFP) tagged Ras-associated binding proteins 5 (Rab5-RFP) and RFP tagged lysosome associated membrane protein-1 (LAMP1-RFP) fusion constructs (CellLight™ Early Endosomes-RFP, BacMam 2.0 and CellLight™ Lysosomes-RFP, BacMam 2.0, Invitrogen) were co-transfected into HeLa cells according to the manufacturer’s instructions. Cells were then washed with PBS and incubated with fresh, serum-supplemented culture media with muGFP-Hex. The sample was incubated at 37 °C for 24 h, washed with cell culture media, and imaged by confocal microscopy (Carl Zeiss LSM 700, or spinning disk confocal microscopy system (PerkinElmer UltraVIEW VoX)) in a live cell imaging chamber (37 °C, 5% CO₂). Nuclei were stained with 16.23 μM Hoechst 33342 for 10 min. For the pulse-chase experiment, Rab5/LAMP1 endosomal/lysosomal labeled HeLa cells were treated with 0.2 μM muGFP-Hex for 1 h at 37 °C. After the pulse period, samples were washed and incubated with fresh media at 37 °C. Endocytosed muGFP-Hex were imaged by confocal microscopy for the indicated times. The fraction of muGFP-Hex escaped to the cytosol was assessed with Mander’s colocalization analysis. Data are mean ± SD from three independent experiments.

Statistics:

All quantitative data was analyzed and plotted using GraphPad Prism. Significance was assessed by a one-way ANOVA and Tukey’s multiple comparisons test at a significance level of p < 0.05. All data shown is representative of at least 3 independent measurements.

Results and Discussion

The Hex carriers used in our experiments had previously been utilized to deliver antibodies inside of cells to intracellular targets [44,46]. The carrier is comprised of the Hex coiled coil at the center, with each alpha helix fused by a 16 residue glycine serine linker to an antibody binding domain, Staphylococcus aureus protein A domain B (SPAB), and a 6x-His tag. The binding affinity of monomeric SPAB to rabbit IgG was 53 nM and the binding affinity of Hex to IgG was found to be stronger than picomolar due to effects of avidity [44]. Upon mixing the carrier with antibodies in a 1 to 3 ratio and incubating at 37 °C, two SPAB domains bind one antibody on either side of the Fc domain, forming a uniform Hex-antibody complex (Fig. 1B). Aside from the original Hex carrier (Fig. 1A), we also produced Hex-SPAB (HS) and SPAB-Hex-SPAB (SHS) to study the effect of binding domain number and orientation on the self-assembly of antibody-carrier complexes (Fig. 1C, 1D) [45]. HS and SHS readily formed complexes with antibodies when mixed and incubated at 37 °C in ratios of 1 to 3 and 1 to 6, respectively. These loading ratios were previously determined by size exclusion chromatography (SEC) and molecular weight estimation of IgG and carrier mixtures [45]. All carrier-antibody complexes were previously found to be stable up to multiple days to two weeks after formation [45]. Additionally, all three carriers were found to be nontoxic at concentrations used for delivery experiments in all cell lines tested (Supplementary Fig. S3) [44]. Characterization of HS and SHS by SEC, and previously published SEC analysis of Hex all showed uniform and highly monodisperse peaks for each carrier (Supplementary Fig. S4) [45]. All three carriers had similar hydrodynamic sizes of ~15 nm as measured by DLS, and all three carrier-IgG complexes had similar sizes of ~25-30 nm (Table 1, Supplementary Fig. S5) [45]. Additionally, the carriers wer e found to be stable in PBS buffer for up to 3 days for Hex and up to 16 days for HS and SHS (Supplementary Fig. S5). The second peaks observed in DLS intensity plots for Hex and HS may be related to transient end-to-end assemblies that the Hex coiled coil peptide is known to form [51]. These assemblies were not detected by SEC and most likely do not form at the lower concentrations and in the presence of serum used in cell experiments. These Hex carrier variants enabled us to test hypotheses relating internalization and endosomal escape properties of carriers with different macromolecular structures, as well as identify the properties of optimal intracellular protein carriers.

Table 1.

Particle diameter and polydispersity data for protein carriers and assemblies.

Protein Assembly	Z-average (nm)	PDI
HS	13	0.283
Hex	15	0.364
SHS	15	0.266
HS-IgG	25	0.161
Hex-IgG	25	0.287
SHS-IgG	28	0.256
nhHS	14	0.315
muGFP-Hex	15	0.199

Effect of coiled coil exposure on internalization

The primary difference between the three carriers, HS, Hex, and SHS, is the relative exposure of the Hex coiled coil in the core of the assembly. When comparing the intracellular delivery of fluorescently labeled antibodies using each of the three carriers, we observed that increasing exposure of the coiled coil in the carriers promoted more delivery of antibodies to HeLa cells, as analyzed by flow cytometry (Fig. 2A, Supplementary Fig. S6). HS-IgG demonstrated the highest delivery, followed by Hex-IgG and then SHS-IgG. Interestingly, the SHS carriers showed no significant difference from the IgG control, suggesting that limited coiled coil exposure, due to the surrounding 12 SPAB domains and 6 bound antibodies, prevented uptake and antibody delivery. In addition, the different orientation of IgG on the carriers could change the curvature of the carrier-IgG complexes, especially between Hex-IgG and HS-IgG, potentially contributing to the differences in uptake. Previously, we had hypothesized that Hex-IgG endocytosis was feasible due to the ~25-30 nm size of the complexes (Table 1), which is well within the range for nanoparticle endocytosis and ~2x larger than the hydrodynamic diameter of soluble IgG [45,52]. However, the limited delivery of SHS-IgG, which has a similar particle size to Hex-IgG and HS-IgG, pointed to the more important role of coiled coil exposure in the endocytosis process.

Figure 2.

Comparison of cellular uptake of different carriers with and without IgG. (A) Flow cytometry analysis of HeLa cell uptake of carriers bound to AlexaFluor 488 labeled IgG over 24 hours. HS and Hex were added at a concentration of 0.2 μM, SHS at 0.1 μM and IgG at 0.6 μM. (B) Flow cytometry analysis of HeLa cell uptake of AlexaFluor 488 labeled SHS, Hex, and HS carriers (0.2 μM for all) with and without unlabeled IgG (1.2 μM IgG for SHS + IgG and 0.6 μM IgG for HS + IgG, Hex + IgG) bound over 24 hours. n.s. means no significance, *** p <0.001, **** p<0.0001

To further increase the exposure of the coiled coil, we tested internalization of the carriers without antibodies. HS, Hex, and SHS were fluorescently labeled and administered to HeLa cells with and without antibodies for 24 hours (Fig. 2B, Supplementary Fig. S6). Across all three carriers, we observed a large increase in internalization of carriers when not loaded with antibodies. We also observed a similar trend amongst carriers, where increasing exposure of the coiled coil in unloaded carriers also increased internalization. SHS had the least internalization of the three, presumably due to double the SPAB domains surrounding the Hex coiled coil. HS demonstrated the best internalization of all tested treatments, as it has six SPAB domains all on one terminus of the protein assembly, leaving the coiled coil exposed on the other side. This data demonstrates the value of exposing the Hex coiled coil in a macromolecular assembly to improve cellular internalization. Importantly, it shows that the Hex coiled coil can deliver smaller proteins that are directly fused, such as SPAB or others, with even higher internalization efficiency than IgG cargo. This supports the use of the Hex coiled coil as a modular platform assembly for intracellular delivery of proteins and motivates the need to understand its delivery properties.

While there is little direct evidence in the literature of coiled coils demonstrating internalization behavior, the Hex coiled coil resembles naturally existing proteins on bacteria and viruses that are involved in membrane association [27,28]. Even though no hexameric coiled coils have been found in nature, the Hex coiled coil in our carriers may be mimicking the membrane association behavior of existing CCs in bacterial effector proteins or the six helix bundles in HIV [28]. We have previously demonstrated that Hex- IgG complexes do not directly disrupt the cell membrane, so the potential membrane association may serve to enhance endocytosis [46]. The role of Hex coiled coils in cellular delivery could motivate the field to explore other uses for promoting drug delivery. Hex could be fused to other protein scaffolds recombinantly, conjugated to lipid or polymer nanoparticles, or adsorbed on solid nanoparticles for enhanced delivery. Beyond Hex, there are a number of coiled coils engineered to have different oligomeric states [53,54]. It will be useful to determine if Hex is unique in its delivery ability or if tetrameric, heptameric, or other coiled coil oligomers have similar abilities, providing a new biological application for the field of de novo coiled coil protein design.

Effect of coiled coil exposure on endosomal disruption

Enhanced endocytosis of any protein carrier is not meaningful without endosomal escape. To investigate the endosomal disruption capacity of Hex variants with different levels of coiled coil exposure, we incubated YFP-Gal8 MDA-MB-231 cells with the carriers both with and without IgG for 24 h and analyzed endosomal disruption events using the microscopy-based YFP-Gal8 recruitment assay. In comparing all three carriers, HS, Hex and SHS, no significant difference in Gal8 recruitment events was found with or without the IgG cargo (Fig. 3, Supplementary Fig. S7). These results show that coiled coil exposure level does not affect the endosomal disruption capacity and the interaction of coiled coil and endosome membrane is not the predominant force that triggers endosomal disruption. Notably, although the uptake of SHS and all IgG loaded carriers was lower, the endosomal disruption was similar to HS and Hex. Given that increasing exposure of the coiled coil in the carriers promoted delivery, this data suggests that higher uptake may not consistently produce a higher incidence of endosomal disruption. Considering that the endosomal disruption rate can be saturated if the uptake concentration of Hex variants is high [55], we also investigated endosomal disruption of Hex variants at lower treatment concentrations. Lower concentrations led to lower Gal8 recruitment events, a dose-dependent effect, which was the same for all three variants (Fig. 4, Supplementary Fig. S8). The same dose-dependent effect can also be observed in the Lipofectamine 3000 positive control, which was tested at several volumes including the manufacturer’s recommendation of 1.5 μL/well (Supplementary Fig. S9). This demonstrated that under the experimental concentration, although the cellular uptake of Hex variants is high, the endosomal disruption did not reach saturation. Results of this study combined with our uptake data indicate that there is no clear relationship between optimal uptake and the endosomal disruption of Hex variants and that components, such as the coiled coils, that enhance uptake may not necessarily have any benefit for endosomal disruption.

Figure 3.

Analysis of endosomal disruption events. Endosomal events per cell were determined from image analysis of the Gal8 recruitment assay (images in Supplementary Fig. S7). YFP-Gal8 MDA-MB-231 cells were treated for 24 hrs with (A) PBS, HS, Hex and SHS carriers; (B) PBS, HS-IgG, Hex-IgG and SHS-IgG complexes. HS, Hex, and SHS were added at a concentration of 0.2 μM. Hex and HS were mixed with 0.6 μM of IgG and SHS with 1.2 μM of IgG. n.s. means no significant difference between carriers.

Figure 4.

Analysis of endosomal disruption events for different concentrations of each carrier. Endosomal events per cell were determined from image analysis of the Gal8 recruitment assay (images in Supplementary Fig. S8). YFP-Gal8 MDA-MB-231 cells were treated for 24 hrs with varying concentrations of (A) HS, (B) Hex and (C) SHS carriers. n.s. means no significance, * p<0.05

Effect of 6x-His tag on endosomal disruption

Understanding the mechanisms of cargo escape from endosomes is important for improving the efficiency of escape through rational carrier design. One route to endosomal escape might be through the proton sponge effect, which refers to the high buffering capacity of some molecules and their ability to swell the endosomes when protonated [31,32]. Protonation causes ions and water to flow into endosomes and can lead to rupture of the endosomal membrane and release of entrapped components. Owing to the six 6x-His tags in each carrier, which protonate at acidic conditions, we hypothesized that these may significantly contribute to endosomal disruption and the cytosolic delivery we previously reported [44,46]. To test this hypothesis, we compared endosomal disruption events of HS and nhHS, with and without IgG. The nhHS variant had a hydrodynamic size of 14 nm, similar to the original HS protein (Table 1, Supplementary Fig. S5). The number of Gal8 recruitment events induced by the HS carrier was similar to the manufacturer’s recommended dose of Lipofectamine 3000 positive control (Supplementary Fig. S9), while for the nhHS carrier, the Gal8 recruitment events were significantly reduced compared to the HS carrier (Fig. 5A, B). However, disruption events in nhHS were still significantly higher than PBS treated control cells. Together, these results indicate that the six 6x-His tags contribute significantly to the endosomal disruption capacity of HS and that the other components of the carrier also contribute to endosomal disruption. Currently, His-tags on protein therapeutics are often removed after purification to reduce immunogenicity [56], as they have been found to affect both antibody and T cell responses to some fusion proteins [57,58]. For drug delivery to cytosolic targets, our results showing endosomal disruption due to His-tags may change this current practice of tag removal. Future work may investigate the balance between using His tags for endosomal escape and reducing His-tag mediated immunogenicity. Moving the tags away from the surface of carriers could reduce immune recognition while still serving as purification tags and endosomal escape domains. The modular design of fusion proteins should allow for such optimization, yielding design parameters for His-tag location and exposure in use of therapeutic proteins in vivo.

Figure 5.

Comparison of endosomal disruption by HS and nhHS and cellular uptake of HS and nHS carriers. (A) Confocal microscopy images showing endosomal disruption events (Bright yellow punctate YFP-Gal8 spots) of YFP-Gal8 MDA-MB-231 cells treated with HS, nhHS, or PBS for 24 hrs. Nuclei were labeled with Hoechst 33342 (blue). Scale bar: 10 μm. (B) Comparison of endosomal events per cell determined from image analysis of the Gal8 recruitment assay. Flow cytometry analysis of HeLa cell uptake of AlexaFluor 488 labeled HS and nhHS carriers without IgG (C) and with IgG bound (D) over 24 hours. HS and nhHS were added at a concentration of 0.2 μM or combined with 0.6 μM of IgG. .n.s. means no significance. * p<0.05

To determine whether reduced endosomal disruption of nhHS carriers was due to lower internalization, we investigated the cellular uptake of HS and nhHS with and without IgG. The flow cytometry results showed that nhHS did not display significantly different uptake than HS (Fig. 5C). Although the cellular uptake of HS-IgG complex was lower than HS carrier alone, no significant difference in uptake was observed between HS-IgG and nhHS-IgG (Fig. 5D). These results indicate that the His-tags have little influence on cellular uptake of both carrier and carrier-cargo complex.

Other studies suggest that local destabilization of the endosomal membrane integrity or nanoscale pore formation in the endosomal membrane leads to a transient local burst release [59-63]. Furthermore, to achieve high endosomal disruption efficacy, it was necessary to reach a critical carrier concentration to achieve a critical membrane tension that strong enough for membrane disruption [64]. Other monovalent His tag containing systems may not reach the critical concentration, resulting in low or no endosomal disruption, while the Hex system contains six 6x-His tags and appears to reach sufficient concentrations for endosomal membrane disruption [65]. Moreover, some polymer-based delivery platforms or cell penetrating peptide conjugated platforms may interact strongly with membranes and subsequently cause membrane damage [66-68]. Although this is beneficial for endosomal disruption, cytotoxicity caused by this mechanism limits these platforms’ in vivo applications [69-71]. It is noteworthy that an imidazole-containing polymeric delivery vehicle showed low cytotoxicity and enhanced endosomal escape capability by optimizing the balance between polymer cationic density with endosomal escape moieties [72]. The imidazole groups can achieve high density, buffer the endosome, and potentially induce its rupture. This successful case suggests that His tags could also serve as the basis for rational design of intracellular protein delivery system with improved endosomal escape capability [37,42,72,73]. Additional strategies can be used to further promote the endosomal escape ability of the Hex carrier. Owing to the simple structure of the Hex carrier, further modifications, such as more His tags and peptides with pH triggered lytic activity, could be easily employed to enhance the endosomal disruptive capacity [74,75]. An alternative approach is to release the cargo from the Hex carrier in the endosome by inducing the endosomal protease cleavage linker [76], which reduces the size of the endosomal pore required for cargo escape.

Hex as a generalizable intracellular protein carrier

Given the importance of both Hex coiled coil exposure and 6x-His tags on internalization and endosomal escape of Hex carriers, we hypothesized that this platform could be utilized more broadly to deliver other proteins into cells besides IgG. We created a new fusion protein, muGFP-Hex, comprised of monomeric ultra-stable green fluorescent protein (muGFP) in place of the SPAB domain as a model protein for delivery (Fig. 6A). There is an N-terminal 6x-His tag on each muGFP-Hex monomer, such that the full assembly also contained 6 tags. As with the prior Hex carriers, muGFP-Hex assemblies were highly uniform when analyzed by SEC (Supplementary Fig. S4) and showed hydrodynamic sizes of 15 nm by DLS analysis that remained stable up to 16 days after sample preparation (Table 1, Supplementary Fig. S5). We compared internalization of muGFP-Hex with another GFP coiled coil: superfolder GFP fused to a leucine zipper (sfGFP-ZE) (Fig. 6B, Supplementary Fig. S6). ZE homodimers have micromolar K_D and will exist in an equilibrium of monomer and dimer forms [77]. This comparison was used to determine if the Hex coiled coil internalization behavior was unique, or generic to coiled coils. For both muGFP-Hex and sfGFP-ZE assemblies, no significant cytotoxicity was observed at the concentrations tested in this experiment (Supplementary Fig. S3). To compare internalization, muGFP-Hex and sfGFP-ZE were incubated with HeLa cells for 24 h at 3 different concentrations and analyzed by flow cytometry (Fig. 6C). muGFP-Hex showed a large and dose-dependent increase in internalization when compared with sfGFP-ZE at all 3 concentrations tested. This relative increase is even greater because sfGFP-ZE has higher fluorescent brightness than muGFP-Hex (Supplementary Fig. S10). These results suggest that internalization properties are unique to the Hex platform, or at least larger coiled coil oligomers, and that other proteins aside from SPAB and muGFP could be delivered by fusing them to the Hex coiled coil. When compared to GFP delivery by fused cell penetrating peptides, the delivery of muGFP by Hex carriers is highly efficient. In one example of GFP delivery with TAT peptides, TAT-GFP required a dose of 150 μM for delivery to 1% of cells [78]. Using a cyclic TAT fusion to GFP, delivery was improved but only ~50% of cells showed uptake at a dose of 50 μM. This direct comparison shows the improved delivery efficiency of muGFP with Hex to 100% of cells at significantly lower doses.

Figure 6.

muGFP-Hex and sfGFP-ZE delivery. Schematic of (A) muGFP-Hex protein (muGFP PDB ID: 5JZL) and (B) sfGFP-ZE protein [27]. Not drawn to scale. (C) Flow cytometry analysis of muGFP-Hex and sfGFP-ZE uptake by HeLa cells over 24 hrs. Inset shows the untreated control and sfGFP-ZE. n.s. means no significance, *** p<0.001

Finally, we investigated if the new muGFP-Hex assembly could escape endosomes and deliver muGFP to the cytosol. Delivery of muGFP-Hex in HeLa cells was analyzed by confocal microscopy with co-staining for endosomal and lysosomal markers (Fig. 7). The YFP-Gal8 MDA-MB-231 assay was not used due to muGFP/YFP spectral overlap. In all samples treated with muGFP-Hex, we observed green punctate spots as well as a diffuse green signal throughout the cell. Colocalization of the bright muGFP-Hex signal with endosomal and lysosomal staining suggests that the punctate spots of muGFP-Hex were proteins in endosomes and lysosomes. However, the Mander’s coefficient was only 0.539 ± 0.074 and the diffuse green signal did not colocalize with red, showing that muGFP-Hex escapes endosomes or lysosomes and reaches the cytosol in HeLa cells. Considering that the extracellular muGFP-Hex proteins were constantly endocytosed, providing co-localized signals, a pulse chase experiment was performed to reveal the trafficking of muGFP-Hex. HeLa cells were exposed to muGFP-Hex for 1 h, followed removal of extracellular muGFP-Hex and confocal imaging. Immediately after the pulse, the Mander’s coefficient was 0.865 ± 0.060, indicating high level of colocalization of the endocytosed muGFP-Hex with endo/lysosomes (Supplementary Fig. S11). After a 30 min chase period, the Mander’s coefficient dropped to 0.550 ± 0.119, suggesting close to half the complexes achieved endosomal escape. After an additional 30 min of chase time, low level of colocalization of muGFP-Hex and endo/lysosomes was reached, with a Mander’s coefficient of 0.370 ± 0.049. These results collectively confirm that the Hex platform has efficient endosomal escape and can be used for intracellular, cytosolic delivery of protein cargo beyond IgG and SPAB.

Figure 7.

Confocal microscopy analysis of muGFP-Hex uptake and endo/lysosomal colocalization. Representative images of Rab5/LAMP1 endosomal/lysosomal labeled HeLa cells treated for 24 h with muGFP-Hex (0.2 μM) or PBS, followed by nuclear staining and imaging. Scale bar: 10 μM

While we have previously tested delivery of antibodies, a number of different binding proteins have been engineered to bind key therapeutic targets in multiple diseases and could be delivered intracellularly by the Hex platform. These range from antibody single chain variable fragments, nanobodies, monobodies, designed ankyrin repeat proteins, affibodies and many more [79]. Evaluating the therapeutic efficacy of these scaffolds is often limited to intracellular expression of the proteins via gene delivery, due to a lack of simple protein delivery techniques [80,81]. The Hex delivery system allows for direct fusion of those domains to the Hex coiled coil for delivery to intracellular targets. Aside from the obvious therapeutic applications and potential for clinical translation, this delivery system could also be a viable laboratory tool for screening and studying new therapeutic proteins and targets in vitro. Monomeric cargos may be directly fused, though cargo proteins with quaternary structure, like IgG, would need to be linked to Hex via affinity or covalent means. As is true with all fusion proteins, the ability for both the Hex and cargo domains to express and fold may depend on the fusion order and the properties of the linker [82]. For any given cargo, these will need to be optimized. Given the role of Hex exposure in delivery, it is also possible that linker length and flexibility affect delivery.

Conclusion

In this work, we used several versions of an intracellular drug delivery vehicle to determine the role of different components on the steps of cytosolic protein delivery. Specifically, use of Hex revealed the significant potential of coiled coil oligomers to induce internalization and the benefit of a spatial arrangement that best exposes these domains to cell membranes. Further, the endosomolytic activity of 6x-His tags was demonstrated in a direct and quantitative approach. The role of the 6x-His tags and Hex coiled coil were combined to effectively deliver muGFP into the cytosol of cells. This work provides a potentially generalizable platform for efficient intracellular delivery of therapeutic proteins and demonstrates the factors that might be critical for the design of future protein or hybrid carriers for cytosolic protein delivery. Altogether, the Hex carriers demonstrated delivery of a wide range of protein cargoes (~20 kDa SPAB; ~26 kDa muGFP; ~150 kDa antibody) with a range of loading ratios (1:12 to 1:3) and retention of protein activity in a nanoassembly of ~15 to 30 nm, ideal for a variety of therapeutic applications. Future work will test delivery of small therapeutic binding proteins to intracellular targets.

Highlights.

• Protein drug carrier can deliver different proteins inside of cells.

• Exposure of the central domain of the carrier increases delivery to cells.

• Presence of histidine tags on the carrier increases escape from endosomes.

• Delivery platform could deliver other therapeutic proteins to new targets.