Reviews
Current Understanding of Direct Translocation of Arginine-Rich Cell-Penetrating Peptides and Its Internalization Mechanisms
2016 Volume 64 Issue 10 Pages 1431-1437
Details
2016 Volume 64 Issue 10 Pages 1431-1437
Arginine-rich cell-penetrating peptides (CPPs) including Tat, Penetratin and oligoarginine peptides are a series of short peptides that can be efficiently internalized into cells and have been widely used as carriers for intracellular delivery of bioactive molecules. In the early phase of the study, CPPs, as well as their conjugates, were thought to rapidly enter cells by direct penetration through membranes, which was later found to be an experimental artifact that was concluded from observations in fixed cells. Although re-evaluation using living unfixed cells revealed that endocytosis has a major role in internalization of these peptides, there are a number of studies reporting that, even if fixation is avoided, direct translocation across plasma membranes and cytosolic distribution of arginine-rich CPPs are still observed in cells without membrane perturbation. In addition, amphiphilic counteranions such as pyrenebutyrate dramatically accelerate direct translocation of these peptides into cells. These results suggest that there are at least two pathways, i.e., endocytosis and direct translocation, both of which would contribute to cellular internalization of arginine-rich CPPs. In this review, we first introduce the story of fixation artifact, which indeed led to the critical progress in CPP study, and then summarize the current understanding for direct translocation of arginine-rich CPPs. Comprehensive understanding of direct translocation of these peptides and its mechanistic elucidation would provide useful knowledge for developing methodologies that would enable efficient intracellular delivery.
The plasma membrane is a tightly regulated barrier to protect cells against influx of exogenous molecules from the outside. This barrier separates intracellular regions from extracellular environment to maintain cellular homeostasis, but at the same time this makes difficult for extracellular hydrophilic molecules including peptides, proteins and nucleic acids to gain access into cells. The only exception is a defined set of ions and small molecules that are transported across plasma membranes through specific channels and transporters upon stimulation with ligands, leading to signal transduction and neuronal firing.
Considerable efforts have been made to develop delivery carriers that efficiently cross the membranes and transport biologically active cargoes into cells. One of the biggest breakthroughs in this field is the discovery of cell-penetrating peptides (CPPs) or protein transduction domains (PTDs), which are a series of short peptides that can be efficiently internalized into cells and have been widely used as carriers for intracellular delivery of bioactive molecules. Among them, arginine-rich CPPs such as Tat peptide from a human immunodeficiency virus (HIV)-1 trans-activator protein TAT,1–3) Penetratin peptide from a Drosophila homeodomain Antennapedia4,5) and oligoarginines6,7) show high efficiency of internalization, facilitating intracellular delivery of a wide range of hydrophilic molecules including small compounds, peptides, proteins, nucleic acids, nanoparticles, liposomes, and others that would be otherwise difficult to enter cells.8–10) Arginine-rich CPPs have, therefore, attracted much attention as one of the promising carriers for intracellular delivery of therapeutic molecules including nucleic acids and drug candidates.8–10) So far, a number of peptide sequences have been reported as CPPs, which are classified by their physicochemical characteristics, such as basic/amphiphilic CPPs that contain several basic amino acids including arginine and lysine, and hydrophobic CPPs (Table 1): interestingly, polyhistidine peptide such as histidine 16-mer (H16) is recently reported as CPPs.11)
| Name | Sequence | Origin | Reference |
|---|---|---|---|
| Basic/Amphiphilic | |||
| Tat | GRKKRRQRRRPPQ | HIV-1 TAT protein (48–60) | 1–3 |
| Penetratin | RQIKIWFQNRRMKWKK | Drosophila Antennapedia homeodomain (43–58) | 4, 5 |
| Oligoarginine | Rn (n=7–16) | Synthetic peptides | 6, 7 |
| FHV | RRRRNRTRRNRRRVR | FHV coat protein (35–49) | 38 |
| pVEC | LLIILRRRIRKQAHAHSK | VE-cadherin (615–632) | 52 |
| Hydrophobic | |||
| TP10 | GWTLNSAGYLLGKINLKALAALAKKIL | Galanin/Mastoparan chimeric peptide | 52 |
| M918 | MVTVLFRRLRIRRACGPPRVRV | p14ARF tumor suppressor protein | 52 |
Although the molecular mechanisms as to how arginine-rich CPPs enter cells have not been fully understood and still remain debated, it is commonly accepted that internalization of these peptides involves endocytosis. We and others have proposed that macropinocytosis, which is one class of endocytosis pathways, is involved in internalization of arginine-rich CPPs such as octa-arginine (R8) and Tat peptides, as well as their fusion proteins.12–14) Binding of these peptides with membrane-associated proteoglycans is an important step that would initiate downstream signal transductions through activation of Rac1, which eventually leads to actin reorganization and induction of macropinocytosis.15) Furthermore, a membrane-associated chemokine receptor CXCR4 is proposed to serve as a cell-surface receptor responsible for internalization of dodeca-arginine (R12) via macropinocytosis.16) On the other hand, there is also strong evidence that other classes of endocytosis pathways such as clathrin-mediated endocytosis17–19) and caveolae-mediated endocytosis20,21) are involved in internalization of these peptides. It is noted that internalization modes of arginine-rich CPPs can be strongly affected by variable factors in each experiment, such as the physicochemical natures of peptides, cell types, incubation temperatures and cargoes to be delivered, which makes mechanistic studies on their internalization exceptionally difficult.22,23)
Molecules that are taken up by cells via endocytosis are first incorporated into endosomes, which are then either trafficked to lysosomes for degradation, or recycled to cell surfaces, depending on the internalization modes of endocytosis. The cellular trafficking of endosomes indicates that biologically active molecules that are transported via endocytosis would not exert their bioactivity in cytosol/nucleus without crossing the endosomal membranes. Carriers that utilize endocytosis for cell entry, including arginine-rich CPPs, therefore need to escape from endosomes for intracellular delivery of their cargoes. However, microscopic observation shows that the majority of arginine-rich CPPs that are internalized by endocytosis apparently remain trapped in endosomes, and seem not to be efficiently released from endosomes even after prolonged incubation.14) Considering accumulating evidence that peptides, proteins and nucleic acids that are delivered by conjugation with arginine-rich CPPs show their bioactivity in cells, at least a limited but effective fraction of internalized conjugates via endocytosis might be released from endosomes, although detailed mechanism as to how they escape from endosomes is not well understood: one possibility is an involvement of a proton–sponge effect, in which accumulation of positively charged molecules would eventually cause osmotic swelling and rupture of the endosomes.24) The methodology that would facilitate endosomal escape has been proposed for improving the efficiency for CPP-mediated delivery, but still needs to be developed.12,25)
Alternatively, it would be quite ideal that exogenous molecules could enter cells not by endocytosis, but by directly passing through the plasma membranes, so that they would be free from degradation by lysosomes without need for endosomal escape. In the early phase of the study, CPPs, as well as their conjugates, were thought to rapidly enter cells by direct penetration through membranes, which was later found to be an experimental artifact that was concluded from observations in fixed cells.26,27) Reevaluation using living unfixed cells revealed that endocytosis has a major role in internalization of these peptides under physiological conditions. However, there are a number of studies reporting that, even if fixation is avoided, direct translocation across membranes and cytosolic distribution of arginine-rich CPPs are still observed. Accumulating evidence related to this phenomena raise the possibility that direct penetration pathways, in addition to endocytosis, would be involved in internalization of arginine-rich CPPs (Fig. 1). Comprehensive understanding of direct translocation of these peptides and its mechanistic elucidation would provide useful knowledge for developing methodologies that would enable efficient intracellular delivery. In the following sections, we first review the story of fixation artifact, which indeed led to the critical progress in CPP study, and then summarize the current understanding for direct penetration of arginine-rich CPPs.
(A–C) In endocytosis-dependent internalization, arginine-rich CPPs bind to cell-surface proteoglycans, such as heparan and chondroitin sulfate proteoglycans, which would induce intracellular downstream signals, leading to actin rearrangement and eventual internalization through endocytosis including macropinocytosis (A). Alternatively, CPPs directly translocate across plasma membranes and distribute throughout cytosol. This process is highly depending on administration concentration, and possibly involves transient membrane deformation (B). Direct translocation of arginine-rich CPPs is also observed in the presence of amphiphilic counteranions such as pyrenebutyrate: with the help of pyrenebutyrate, Alexa488-labeled R8 peptide (R8-Alexa488) gradually translocates into cell, distributes throughout cytosol and reaches nucleus (C).
One of the difficulties in mechanistic elucidation of internalization of arginine-rich CPPs and their conjugates is originated from the unique feature of these peptides. Arginine-rich CPPs have some positive charges and bind strongly to the negatively charged cell surfaces. The peptides bound to cell surfaces remain associated with membranes even after washings, but readily enter cells and distribute cytosol in a diffusive fashion, once the plasma membranes are disturbed. It is known that fixation with organic solvents including methanol, which is commonly conducted before microscopic observation, significantly disturbs membrane structures. The alteration in cellular distribution of these peptides, induced by membrane perturbation, would lead to incorrect interpretations for their internalization.
In the early phase of the mechanistic studies, it was reported with surprise that, despite their highly hydrophilic nature, arginine-rich CPPs, such as Tat, Penetratin and oligoarginine peptides, readily entered cells and were accumulated in nucleus when added to cell culture media. The internalization and nuclear accumulation of these peptides were not affected by low-temperature incubations, by ATP depletion, or by treatment of endocytosis inhibitors.3,5,28) Specific receptors or transporters seemed not to be involved, because internalization of these peptides occurred independently on primary sequences, secondary structures, chirality, and backbone structures of peptides.29,30) Collectively, it was thought that arginine-rich CPPs were internalized into cells independently on endocytosis. Alternatively, direct penetration through cellular membranes was proposed as a possible mechanism of translocation.
A herpes simplex virus protein VP22, a positively charged DNA-binding protein, was reported to translocate across cell membranes and accumulate in cell nucleus,31) similar to Tat and other arginine-rich CPPs. The ability of VP22 to deliver its fusion proteins into cell nucleus was also suggested. Translocation of VP22 through membranes seemed to be a rapid process, and apparently occurred even at 4°C, raising the possibility that energy-independent modes of internalization, which are different from endocytosis, would be involved. Due to these unique features, VP22 was expected to be one of the promising carriers that would translocate across membranes and facilitate efficient delivery of cargoes including genes and proteins to cell nucleus.
In 2001, however, Lundberg and Johansson reported that VP22 conjugated with green fluorescent protein (GFP) (GFP-VP22) shows no nuclear localization but stays at the cell surface when they perform microscopic analysis in living cells, while this protein localizes almost exclusively to the cell nucleus after fixation with methanol.32) They also reported that a DNA-binding protein histone H1, which is considered to be unrelated to the proteins with cell-penetrating activity, rapidly enters cells and accumulates in the nucleus in the methanol-fixed cells, showing apparent membrane translocation ability similar to VP22.33) From these results, the authors firstly proposed that cellular distribution of VP22 in living cells may be different from that in fixed cells, and that the fixation with methanol may induce artifactual import and nuclear localization of VP22. Indeed, nuclear localization of VP22 was reported only in fixed cells, but not in living cells.
In 2003, Richard et al. tested the validity of experimental procedures that were commonly used for mechanistic studies in internalization of arginine-rich CPPs. They performed microscopic observation in living cells, as well as in fixed cells, and found that cell fixation significantly changes cellular distribution of Tat and oligoarginine (R9) peptides, leading to artificial redistribution of these peptides into the nucleus.26) This is supported by the study from Lundberg et al., demonstrating that the similar distribution changes were observed after fixation of cells that were treated with GFP conjugated with VP22, Tat, R8 or oligolysine (K8).27) The both group also pointed out that these peptides are bound strongly to cell surfaces and are not removed completely by repeated washings. This indicates that quantification of the amount of internalized peptides measured by fluorescent-activated cell sorter (FACS) could provide artifactual data unless additional steps are included to remove the bound peptides, such as trypsin treatment26) and heparin washing.27) Indeed, FACS analysis on trypsin-treated cells radically revised mechanistic understanding of CPP internalization that were proposed at that moment, demonstrating that internalization of Tat and R9 peptides, as well as Tat conjugated with peptide nucleic acids (PNAs), is strongly inhibited by low temperature incubation or by depletion of cellular ATP pool. Furthermore, the kinetics of cellular internalization of these peptides is comparable with that of FM 4-64, an endocytosis marker. These results strongly indicate that internalization of arginine-rich CPPs occurs mostly in an endocytosis-dependent manner, and that direct translocation and nuclear accumulation of these peptides that were proposed in the previous literatures would be possibly an experimental artifact that was concluded by misinterpretation of results obtained from fixed cells. A series of studies related to fixation artifact not only revised experimental data for its correct interpretation, but also revealed the unique features of arginine-rich CPPs, both of which strongly contributed to better understanding of internalization mechanisms of these peptides.
It is noted, however, that not all the fixation methods induce artificial redistribution of the arginine-rich CPPs into the nucleus. We tested whether distribution of the internalized dodeca-arginine (R12), one of oligoarginine CPPs, would be affected by fixation using acetone–methanol (1 : 1), 4% paraformaldehyde (PFA), and 2% glutaraldehyde (GA).34) Among these fixation solutions, treatment of acetone–methanol resulted in significant changes in R12 distribution, as previously reported. In contrast, the punctate signals, which are suggestive of endosomes, of R12 peptide were well retained after treatment of 4% PFA and 2% GA, suggesting that these fixation methods would have no apparent effects on cellular distribution of R12 peptide and other arginine-rich CPPs. Mounting medium containing glycerol, which is often used for microscopic observation, should be avoided, as it may induce a similar artificial effect on CPP distribution.3,26)
There is, however, accumulating evidence that arginine-rich CPPs would be internalized into cells not only by endocytosis, but also by direct translocation through plasma membranes. In the cellular uptake experiments performed without fixation, we still observed internalization of R8 into cells at 4°C, where energy-dependent cellular events including endocytosis are strongly suppressed. The internalized R8 peptide at this temperature showed diffuse distribution throughout cytosol and nucleus, which is clearly distinct from that of endosome-like structures as observed in cells incubated at 37°C.13,23,35) Co-incubation of R8 together with propidium iodide (PI), a nuclear probe that is impermeable to intact membranes, resulted in no apparent staining of PI in cell nucleus, suggesting that cytosolic distribution of R8 is not attributed by significant membrane disruption as occurs in methanol-fixed cells. Likewise, similar cytosolic distribution of Tat and R8 is observed in living unfixed cells, when cells are treated with the pharmacological inhibitors for endocytosis such as 5-(N-ethyl-N-isopropyl)amiloride (EIPA)36) and methyl-β-cyclodextrin.37) These results collectively indicate that an endocytosis-independent pathway that enables direct translocation across membranes would be involved in internalization of these peptides.
Administration concentration of arginine-rich CPPs apparently has an important role in determining internalization modes of these peptides. Fretz et al. examined effects of peptide concentration on subcellular distribution for fluorescently labeled R8 in leukemia cells. They found that, while the endosome-like punctate signals are evident at 2 µM, increasing the peptide concentration to 5 µM or higher results in significant increase in the diffuse signals throughout cytosol, in addition to the punctate labeling.37) We also observed in HeLa cells that increasing administration concentration of R12 leads not only to a sigmoidal increase in the amounts of the internalized peptides, but also to dramatic enhancement of cytosolic labeling.23) These results suggest that there are at least two pathways for CPP internalization, i.e., endocytosis and direct translocation, and the latter mode of internalization occurs highly dependently on administration concentration. This idea is in good agreement with the results reported by Duchardt et al., demonstrating that cytosolic translocation of Tat, Penetratin and R9 are enhanced when HeLa cells are treated at relatively high concentration of these peptides (>10 µM), while the vesicular distribution, suggestive of endosomes, is dominant at low concentration.22) Similar concentration dependency of translocation is observed for other arginine-rich peptides including a peptide derived from Flock house virus (FHV) coat protein,38) tryptophan-labeled R9 (WR9)39) and a peptide conjugate with R8.40) These results indicate that, although endocytosis is a major pathway at least at low peptide concentration, direct translocation through plasma membranes that makes cytosolic distribution is prominent at higher concentration than threshold.
Interestingly, direct translocation of arginine-rich CPPs is accompanied by dynamic alterations in local membrane structures. Duchardt et al. performed time-lapse imaging with confocal microscopy, and investigated the dynamics of direct translocation of fluorescently labeled R9 peptide.22) They found that, in the initial phase of internalization, strong fluorescent signals are observed in the small, but limited areas of cell membranes, which then gradually spread throughout the cytosol and nucleus. This result implies that direct translocation would not occur uniformly in cells, but originate from spatially restricted regions on plasma membranes, which they call “nucleation zones (NZs)” as highly efficient internalization platforms. Cell-surface heparan sulfates seem to be necessary for cytosolic translocation through NZs, because enzymatic digestion by heparinase resulted in strong suppression of direct translocation of R9. These findings are consistent with our results, demonstrating that influx of R12 peptide into cells is initiated from specific locations on the plasma membranes, which eventually diffuses throughout cytosol.23) Importantly, differential interference contrast (DIC) imaging revealed that unique “particle-like” structures with a diameter of ca. 1 µm are transiently formed, coincidently on the specific locations of the plasma membranes where influx of R12 peptide is initiated.34) This membrane structure might be related to “dense aggregates,” which are the ones observed on the cellular membranes by DIC imaging when Tat rapidly enters the cytoplasm and nucleus of living fibroblast cells.41) It is known that ganglioside GM1 and sphingomyelin (SM) are both abundant in a membrane microdomain called lipid raft. Microscopic observation shows that accumulation of cholera toxin subunit B (CTxB) and Lysenin, specific markers for GM1 and SM, respectively, is detected specifically on the membrane particles, and thus at the peptide influx sites, suggesting that the membrane microdomains would have roles in direct translocation of R12. In addition, the membrane particles, together with the peptide influx sites, are stained by Annexin V, a specific marker for phosphatidylserine (PS), indicating the presence of PS in the outer leaflet of cell membranes. Because PS normally distributes in the inner leaflet of plasma membranes, but not in the outer leaflet, local inversion of plasma membranes may be involved in formation of membrane particles and subsequent peptide influx. Furthermore, electron microscopic observation revealed that the membrane particles consist of a number of small vesicles with multilamellar membrane structures. These results suggest that oligoarginines and other arginine-rich CPPs interact with various components of plasma membranes with negative charges including heparan sulfate proteoglycans and glycosylated lipids, which may induce dynamic alterations in membrane structures such as particle formation and local membrane inversion, leading to direct translocation of these peptides into cells, although further mechanistic studies are needed.
Due to its high pKa value (pKa ca. 12), the guanidine group of an arginine residue is protonated, existing as a guanidinium cation under the physiological conditions. The proximity of the guanidinium cations in oligoarginines, however, results in charge repulsion between proximal side chains. To minimize charge repulsion, it is highly possible that guanidinium cations would always accept and exchange their counteranions. Dynamic exchange of counteranions would make guanidinium-rich oligomers either hydrophilic or hydrophobic, depending on the nature of counteranions in the environments. This unique feature of guanidinium-rich oligomers is experimentally demonstrated by Matile and colleagues.42–45) They showed that, despite their highly hydrophilic nature, oligoarginines can be partitioned into the hydrophobic phase, e.g., chloroform, in the presence of amphiphilic counteranions including fatty acids, phospholipids and other hydrophobic anions. The oligoarginines would be transferred again into the aqueous phase by exchanging counteranions, making overall translocation of oligoarginines across hydrophobic environment.
Based on this idea, we tested whether amphiphilic counteranions would make translocation of guanidinium-rich oligomers across membranes in cultured cells. To address this, cells were pretreated with amphiphilic counteranions in phosphate-buffered saline (PBS) for 2 min, and then treated with arginine-rich CPPs in PBS for 5 min. PBS was used to avoid possible interference of ionic molecules and proteins in culture medium, making simple evaluation of counteranion effects on CPP translocation. We found that, in the presence of pyrenebutyrate, which is one of the amphiphilic counteranions used in the above studies, arginine-rich CPPs including oligoarginines, Tat and Penetratin peptides translocate through plasma membranes, and distribute throughout cytosol and nucleus.46) Among the amphiphilic counteranions with large aromatic groups, including carboxylate, sulfate and phosphate derivatives of fullerenes, calixarenes and pyrenes, pyrenebutyrate showed the best activity to promote translocation of R8 peptide (unpublished data). It is proposed that pyrenebutyrate forms relatively stable complex with oligoarginine by cation–π and π–π interactions,42) which would probably contribute to the difference in the observed activity. Treatment of pyrenebutyrate neither induces acute damages on cell membranes, nor decreases cell viability after prolonged incubation. This translocation occurs within a few minutes, and is observed in a wide range of cell lines including HeLa, Chinese hamster ovary (CHO), Cos-7, PC12 and RAW264.7 cells. We also found that pyrenebutyrate promotes translocation of enhanced green fluorescent protein (EGFP) conjugated with R8 into the cytosol of cultured cells such as HeLa cells and rat hippocampus primary neurons, suggesting applicability of this method to cytosolic delivery for exogenous molecules. So far, pyrenebutyrate-assisted translocation has been widely used for promoting the intracellular delivery of CPP-conjugated bioactive molecules including an apoptosis-inducing peptide PAD,46) a tyrosinase inhibitor Hydroquinone,47) a tumor suppressor p53,48) and chemically modified fluorescent tags/proteins for fluorescent imaging49,50) (Table 2).
| Cargo | CPP | Cell | Application | Reference |
|---|---|---|---|---|
| EGFP | R8 | HeLa/Rat primary neuron | — | 46 |
| EGFP | R11 | B16 melanoma cells/Skin of guinea pig | — | 47 |
| PAD (Pro-apoptotic domain) | R8 | HeLa | Induction of cell death | 46 |
| Hydroquinone (Tyrosinase inhibitor) | R11 | B16 melanoma cells/Skin of guinea pig | Suppression of UV-induced pigmentation | 47 |
| p53 | R11/R3 | U251 glioma cells | Transcriptional activation/Suppression of cell growth | 48 |
| Cyclodextrin | Penetratin | MDCK | FRET imaging | 49 |
| Leucine zipper peptide | R8 | HeLa | Fluorescent labeling of cellular proteins | 50 |
| Ubiquitin, GB1, FKBP12 | Tat | HeLa | In-cell NMR | 51 |
In-cell NMR spectroscopy is one of the prominent applications of pyrenebutyrate-assisted translocation.51) Inomata et al. reported that CPP-conjugated proteins including ubiquitin, streptococcal protein GB1 and FKBP12 are delivered into HeLa cells with the help of pyrenebutyrate, and that the 1H–15N correlation spectrum of these proteins inside cells is recorded with high resolution. They revealed that several cross-peaks in the in-cell spectrum of FKBP12 are shifted after treatment of its ligand, i.e., FK506 and rapamycin, suggesting that structural changes in proteins of interest that would be induced by ligand binding are successfully detected in the cells. Thus, in-cell NMR spectroscopy in the mammalian cells, assisted by CPP-mediated delivery using pyrenebutyrate, allows for monitoring protein–protein or protein–ligand interactions in cells, providing a powerful tool for drug screening.
Despite several applications reported so far, the exact mechanism as to how arginine-rich CPPs pass through cellular membranes in the presence of pyrenebutyrate remains elucidated. Guterstam et al. showed that pyrenebutyrate accelerates cytosolic translocation of arginine-rich CPPs such as R9 and Tat, but not of rather hydrophobic CPPs including pVEC and M918, supporting the importance of guanidinium cations in this translocation.52) Pyrenebutyrate-assisted translocation is observed even at 4°C, and is not affected by treatment of the pharmacological inhibitors that would suppress endocytosis-related cellular events such as cytochalasin D, nocodazole, sodium azide and EIPA. These results strongly suggest that the different modes of internalization from endocytosis would be involved.46) One key observation is that this translocation is largely suppressed when membrane potential is diminished, implying that membrane potential might work as a driving force for direct influx of these peptides into cells. Fluorescence microscopic observation demonstrates that pyrenebutyrate readily enters cells and localizes in perinuclear regions, and that cellular distribution of pyrenebutyrate seems not to change even after administration of arginine-rich CPPs. In addition, pyrenebutyrate is easily removed from cells by simple washings with buffers. These results suggest that pyrenebutyrate freely diffuses through cell membranes, possibly acting as a translocation catalyst to promote direct translocation of arginine-rich CPPs.
In model membrane studies, treatment of giant unilameller vesicles (GUVs) with pyrenebutyrate results in significant accumulation of R8 peptide on negatively charged membranes, which leads to internalization of R8 peptide into GUVs without perturbation of membrane structures.53) This implies the possibility that pyrenebutyrate would have some effects on structures or physicochemical properties of lipid membranes during R8 translocation. Indeed, pyrenebutyrate seems to increase the membrane fluidity, because phase separation of liquid-ordered and liquid-disordered phases in the GUV membranes disappears in the presence of pyrenebutyrate. In addition, treatment of pyrenebutyrate leads to formation of tubular structures of membranes inside GUVs, which is a characteristic membrane structure with negative curvature. Pyrenebutyrate thus has various effects on the structures of the negatively charged membranes, which would contribute to direct translocation of arginine-rich CPPs, although further mechanistic studies focusing on the relationships between dynamic membrane structural changes and CPP translocation are needed.
It is noted that pyrenebutyrate-mediated translocation of arginine-rich CPPs is strongly suppressed by the presence of serum in culture media: PBS or other buffers are recommended to use for a short incubation period when CPPs or their conjugates are applied. This dependency of serum is probably due to non-specific interaction of serum proteins with pyrenebutyrate, significantly lowering apparent effective concentration of pyrenebutyrate in media. Indeed, cytosolic translocation of CPPs can occur when higher concentration of pyrenebutyrate is used (unpublished data). In addition, direct translocation mediated by pyrenebutyrate is affected by the characteristics of their cargoes, e.g., molecular sizes and charges, similar to intracellular delivery mediated by CPPs. Guterstam et al. demonstrated that non-covalent complex of oligonucleotides (ONs) and R9 shows vesicular distribution even in the presence of pyrenebutyrate, and that chloroquine treatment is necessary for the splice-switching activity of ONs, indicating that treatment of pyrenebutyrate would not make internalization of ONs through direct translocation, but via endocytosis pathways.52) ONs are highly negatively charged molecules, which would make competition with pyrenebutyrate for interaction with R9, resulting in insufficient activation of pyrenebutyrate-assisted translocation of CPP-associated cargoes. On the other hand, Jablonski et al. tested feasibility of cytosolic translocation of quantum dots conjugated with polyarginines (PA-QDs) using confocal microscopy and electron transmission microscopy. They demonstrated that pyrenebutyrate does not accelerate cellular internalization of PA-QDs, but rather increases their cellular binding.54) This is not surprising because PA-QDs, which have a large hydrodynamic diameter of 32 nm, are not expected to directly pass through the lipid bilayers. The other group also reported that pyrenebutyrate has no acceleration effects on internalization of R9-conjugated QDs, while R9-conjugated GFP is internalized efficiently into cytosol under the same conditions.55)
Direct translocation of arginine-rich CPPs across plasma membranes is a unique mode of internalization that is accompanied with particle formation on the plasma membranes and local membrane inversion. The mechanistic link between dynamic changes in membrane structures and peptide influx is not yet clear, and the molecular mechanism as to how structural changes in local membrane regions would cause peptide translocation through membranes remains to be elucidated. On the other hand, direct translocation of these peptides is promoted by amphiphilic counteranions such as pyrenebutyrate. Although pyrenebutyrate has the ability to increase the membrane fluidity and induce negative curvature, detailed mechanism as to how pyrenebutyrate promotes translocation still remains unknown. Because intracellular delivery using direct translocation across membranes is highly expected, further studies should be needed for better understanding of this unique mode of translocation.
Evidence from these experiments raise the possibility that structural alteration of plasma membranes may be a key event that would initiate direct translocation of arginine-rich CPPs. Because arginine-rich CPPs interact with plasma membranes through binding to various cell-surface molecules with negative charges including glycolipids and proteoglycans, membrane structures or their physicochemical properties would be altered at least at the local sites where these peptides are binding. Indeed, Lamaziere et al. reported that interaction of Penetratin with liquid-disordered membranes leads to induction of negative curvature in model membranes.56) Recently, we found that inducing the positive curvature by a short peptide derived from Epsin-1 led to dramatic promotion of direct translocation of R8 across plasma membranes, as well as across artificial membranes of GUVs,57) indicating that direct translocation of R8 can be regulated by artificial manipulation of membrane curvature. Thus, further elucidation of the relationships between membrane alteration and peptide translocation, together with methodological development that would enable curvature engineering, should provide a platform for highly efficient intracellular delivery.
This work was supported by Grants-in-Aid for Japan Society for the Promotion of Science (JSPS) Fellows (to T.T.) and for Young Scientists (A; to T.T.) from the JSPS, Japan.
The authors declare no conflict of interest.
