Cell-penetrating peptides (CPPs) hold great potential as tools for drug delivery systems (DDSs). Numerous research groups have developed novel CPPs with enhanced functionality and safety. This review highlights recent advancements in CPP research by my research group and our collaborators. We have designed novel CPPs incorporating unnatural amino acids, explored their applications in DDSs, and elucidated their functional mechanisms.

Cell-penetrating peptides (CPPs), which contain basic amino acids such as L-arginine (Arg) and L-lysine (Lys), can transport membrane-impermeable drugs, proteins, nucleic acids, and nanoparticles into intracellular compartments.^1–5) Due to this capability, CPPs have been widely investigated as potential tools for drug delivery systems (DDSs). Numerous CPP sequences have been identified, and several clinical trials utilizing CPPs for DDS applications have been conducted. However, no CPP-based therapy has yet been clinically implemented.^6,7) The primary obstacles to clinical application are their insufficient cell-penetrating efficiency and non-negligible toxicity. To overcome these challenges, the development of novel CPPs with enhanced cell-penetrating ability and reduced toxicity is essential. This review article highlights recent advances in CPP research conducted by our groups (Fig. 1). We have designed novel CPPs using three distinct strategies. The first approach involves Arg-rich peptides, because Arg often plays a crucial role in the cell-penetrating ability of CPPs. The second strategy focuses on amphipathic helical peptides, where controlling the peptide secondary structure into a helical conformation enhances cell penetration. The third approach explores peptides with modified side-chain structures, incorporating amino acids with unnatural side-chains to improve functionality. Furthermore, we have investigated the functional mechanisms of CPPs. To evaluate their cell-penetrating ability and mechanism, fluorophores have been chemically conjugated to peptides as membrane-impermeable drug models. Because basic peptides carry positive charges, they interact with anionic nucleic acids, including plasmid DNA (pDNA), mRNA, and small interfering RNA (siRNA), leading to the formation of nano-sized complexes. These complexes have potential applications as intracellular DDS tools for gene therapy and nucleic acid therapeutics. This review provides a comprehensive summary of these studies.

Fig. 1. Cell-Penetrating Peptides Described in This Review

We have studied α,α-disubstituted α-amino acids (dAAs),^8–10) which feature two alkyl substituents at the α-position, for the conformational properties of peptides containing dAAs^11–15) and their functional applications.^16,17) Peptides incorporating dAAs tend to adopt a helical conformation^18,19) and exhibit increased resistance to enzymatic degradation,^20,21) leading to enhanced and prolonged functional activity in cultured cells. Based on these properties, we developed Arg-rich CPPs containing dAAs (Fig. 2).

Fig. 2. Design of Arg-Rich Peptides

α-Aminoisobutyric acid (Aib), which has two methyl groups at the α-position, is the simplest dAA and is well known as a helical inducer in peptides.^18,19,22) We designed a nonapeptide (Arg-Arg-Aib)₃, composed of three Aib residues and six Arg residues.²³⁾ Despite its short sequence, (Arg-Arg-Aib)₃ adopted a helical structure and exhibited high cell-penetrating ability due to its loose amphipathicity. Furthermore, intracellular pDNA delivery was evaluated using peptide/pDNA complexes.²⁴⁾ Notably, pDNA-encoded luciferase allows quantification of transfection efficiency using luciferase assay. Compared with an Arg nonapeptide (R9), a representative CPP, (Arg-Arg-Aib)₃ exhibited remarkably low transfection efficiency despite having similar cell-penetrating ability. Analyzing the physicochemical properties of the peptide/pDNA complexes, including zeta potential and size, suggested that (Arg-Arg-Aib)₃ failed to form stable complexes. To enhance the transfection efficiency of Aib-containing peptides, the following design strategies (sections 2.2–2.4) were explored (Fig. 2).

We speculated that the inability of (Arg-Arg-Aib)₃ to form stable pDNA complexes was due to its lower Arg content compared with R9 and the insufficient stability of its helical structure. To address this limitation of (Arg-Arg-Aib)₃, we designed (Arg-Arg-Aib)_n (n = 1–6) with varying peptide lengths²⁵⁾ (Fig. 2). Helicity and cell-penetrating ability of (Arg-Arg-Aib)_n increased with peptide length. These peptides were then applied to intracellular pDNA delivery. For n ≥ 4, the size of peptide/pDNA complexes stabilized at approximately 100 nm, which was suitable for intracellular DDS applications, with positive zeta potentials. Moreover, (Arg-Arg-Aib)_n (n ≥ 4) exhibited more than 10-fold higher transfection efficiency than R9 and demonstrated prolonged transfection activity compared with a commercially available transfection reagent. This prolonged effect was likely due to the high resistance of these peptides to intracellular enzymatic degradation. Thus, elongation of the Aib-containing peptide sequence led to enhanced and sustained intracellular pDNA delivery. Furthermore, (Arg-Arg-Aib)₅ was evaluated for intracellular mRNA delivery, using luciferase-encoding mRNA to quantify delivery efficiency.²⁶⁾ Remarkably, it exhibited approximately 1000-fold higher mRNA delivery efficiency than R9, along with prolonged protein expression. Because exogenous mRNA undergoes more rapid intracellular degradation than pDNA, Aib-containing peptides, which exhibit high resistance to enzymatic degradation, may effectively protect delivered mRNA both extracellularly and intracellularly. Thus, longer (Arg-Arg-Aib)_n peptides can efficiently facilitate the intracellular delivery of both pDNA and mRNA.

Aib is a hydrophobic amino acid, which was expected to be a disadvantage for (Arg-Arg-Aib)₃ in forming peptide/pDNA complexes via electrostatic interactions. To overcome this limitation, we designed five-membered cyclic dAAs with a basic functional group, Ac₅c^X, and developed their synthetic routes²⁷⁾ (Fig. 2). Amino and guanidino groups were chosen as the basic functional groups, and Ac₅c^X contained two chiral centers. In this study, four dAAs, (1S,3R)-Ac₅c^NH2, (1R,3R)-Ac₅c^NH2, (1S,3R)-Ac₅c^Gu, and (1R,3R)-Ac₅c^Gu, were incorporated into (Arg-Arg-Ac₅c^X)₃ in place of Aib in (Arg-Arg-Aib)₃ to investigate the effects of basic functional groups and α-position chirality. (Arg-Arg-Ac₅c^X)₃ peptides exhibited high water solubility due to the introduction of hydrophilic dAAs and adopted a helical structure in aqueous solution. All (Arg-Arg-Ac₅c^X)₃ peptides displayed remarkably high and sustained cell-penetrating ability and efficiently formed pDNA complexes of approximately 100 nm in size. Their transfection efficiencies were 100-fold higher than that of R9. However, no significant differences in cell-penetrating ability or transfection efficiency were observed among the (Arg-Arg-Ac₅c^X)₃ variants. Nevertheless, the incorporation of basic cyclic dAAs into Arg-rich peptides represents a promising strategy for designing CPPs for pDNA delivery. Similarly, we designed and introduced basic dAAs with a piperidine ring in the side chain, Api^X, into Arg-rich peptides.²⁸⁾ Api^X demonstrated effects similar to those of Ac₅c^X: (Arg-Arg-Api^X)₃ exhibited a preferred helical conformation, high and sustained cell-penetrating ability, excellent pDNA complexation, and efficient intracellular pDNA delivery.

Sarcosine, an N-methyl glycine, is highly hydrophilic. Oligosarcosine and polysarcosine serve as hydrophilic and biocompatible elements in biomaterials and have attracted significant interest as potential alternatives to poly(ethylene glycol) (PEG), which is well known as biocompatible materials. In DDS research, oligosarcosine has been utilized as a hydrophilic segment to prevent non-specific interactions with biomacromolecules and cell surfaces, as well as a biocompatible segment to minimize toxicity.^29,30) To address the limitations of Aib-containing Arg-rich peptides, we conjugated oligosarcosine into (Aib-Arg-Arg)₃, varying the oligosarcosine chain length (10, 15, and 20 sarcosine units), with the goal of inhibiting pDNA complex aggregation³¹⁾ (Fig. 2). Oligosarcosine 10 exhibited an insufficient shielding effect, whereas oligosarcosine 15 and 20 facilitated the formation of pDNA complexes of approximately 200 nm with moderate shielding. (Aib-Arg-Arg)₃, with or without oligosarcosine 10, showed almost no pDNA transfection, whereas conjugation with oligosarcosine 15 and 20 resulted in reasonable transfection efficiencies. Thus, conjugating oligosarcosine of an appropriate length to Aib-containing Arg-rich peptides improved pDNA complexation and enabled efficient intracellular pDNA delivery.

Several CPPs achieve cell-penetrating ability by adopting secondary structures such as α-helices and β-sheets, forming a basic/hydrophobic amphipathic structure.^1–5) The key feature of these CPPs is the formation of a stable secondary structure. dAAs are promising tools for inducing helical structures in peptides.

We designed 12-mer peptides based on a model amphipathic peptide (MAP: KLALKLALKALKAALKLA),^32,33) incorporating dAAs³⁴⁾ (Fig. 3a). When these peptides adopt an α-helical structure, they acquire amphipathic properties. Five dAAs, Aib, diethylglycine (Deg), hydroxymethylserine (Hms), 1-aminocyclopentane carboxylic acid (Ac₅c), and dipropylglycine (Dpg), were substituted for L-alanine (Ala) in the basic MAP sequence. Hms peptide failed to form a helical structure, whereas Deg, Ac₅c, and Dpg peptides exhibited high helicity and excellent cell-penetrating ability. Furthermore, intracellular siRNA delivery was evaluated using these peptides. Among them, Dpg peptide formed stable complexes with siRNA, achieving an appropriate size for intracellular delivery. Luciferase-expressing human hepatoma and siRNA for luciferase gene were selected to evaluate intracellular siRNA delivery. Notably, Dpg peptide exhibited effective RNA interference efficiency at low peptide and siRNA concentrations. However, siRNA delivery was only successful in serum-free conditions. To overcome this limitation, we designed Dpg-containing peptides with varying lengths, sequences, and basic amino acids³⁵⁾ (Fig. 3b). Extending the peptide length to a 16-mer enhanced the helicity, preferred an α-helical structure, and enabled effective intracellular siRNA delivery even in the presence of serum with negligible cytotoxicity. Additionally, Dpg-containing peptides were applied to intracellular pDNA delivery.³⁶⁾ The peptides that performed well in siRNA delivery also exhibited high pDNA transfection efficiency. However, two peptides that failed to deliver siRNA demonstrated high pDNA transfection efficiencies. These results suggested that the optimal CPPs for nucleic acid delivery vary depending on the physicochemical and biological properties of nucleic acids, such as their molecular weights (pDNA: very high; siRNA intermediate) and their functional site (pDNA: nucleus; siRNA: cytoplasm).

Fig. 3. Amphipathic Helical Peptides Containing Various dAAs (a) and Dpg (b)

Additionally, we developed amphipathic 3₁₀-helical peptides using the cyclic dAA Ac₅c.³⁷⁾ The 3₁₀-helix is a more compact structure than the α-helix. We designed nonapeptides containing three Arg and six Leu and/or Ac₅c. The number of Ac₅c residues in the peptides influenced their preferred conformations, resulting in α-helical, 3₁₀-helical, and random structures. Notably, Ac₅c-containing peptide that acquired amphipathic properties via 3₁₀-helical structure exhibited strong cell-penetrating ability.

Crosslinking between amino acid side chains is a promising strategy for constructing functional helical peptides.^38,39) We designed CPPs containing a side-chain crosslink between cyclic dAAs⁴⁰⁾ (Fig. 4). Two cyclic dAAs with allyl groups in their side chains were incorporated into R9 every 3 Arg residues, forming a non-stapled peptide. Subsequently, stapling was performed using a first-generation Grubbs catalyst, yielding a stapled peptide. The R9 peptide adopted a random structure, whereas the non-stapled peptide formed a helical structure. The introduction of the staple further enhanced helicity. Additionally, resistance to enzymatic degradation increased with the incorporation of cyclic dAAs, and was further improved by crosslinking. Cellular uptake studies revealed that R9 exhibited strong uptake after a short incubation time, whereas the uptake of non-stapled and stapled peptides increased with longer incubation times. Among these, a stapled peptide displayed significantly higher cell-penetrating ability than a non-stapled peptide. Overall, incorporating cyclic dAAs into the R9 peptide and introducing crosslinking stabilized the helical structures, increased protease resistance in serum and within cells, and led to enhanced and sustained cell-penetrating abilities.

Fig. 4. Peptides with Modified Side-Chain Structure

The key determination of the cell-penetrating ability of Arg-rich peptides is the presence of guanidino groups in Arg residues. CPPs containing not only Arg residues but also Arg derivatives with guanidino groups have been extensively developed.^41,42) Polycations with ethylenediamine structures are well-known as efficient gene delivery carriers.^43,44) Their protonation state plays a crucial role in facilitating endosomal escape.^45,46) At neutral pH, ethylenediamine exists predominantly in a mono-protonated form, which has low membrane-destabilizing capacity. However, at acidic pH, it adopts a di-protonated form, which significantly enhances membrane destabilization. Polycations containing ethylenediamine structures selectively destabilize endosomal membranes, enabling efficient endosomal escape with minimal cytotoxicity. By integrating the properties of Arg-rich peptide and a di-protonated ethylenediamine, we designed a CPP with GEt amine structures composed of Lys derivative (Lys(GEt))⁴⁷⁾ (Fig. 4). The pK_a of the protonated guanidino group (pK_a 12.5 for Arg) is notably higher than that of the protonated primary amino group (pK_a 10.2 for Lys). Thus, Lys(GEt) peptide was expected to adopt a di-protonated form even at neutral pH, leading to enhanced membrane destabilization and high cell-penetrating ability. Indeed, a GEt amine model compound exhibited a di-protonated form at pH 7.4. Cellular uptake studies showed that the Lys(GEt) nonapeptide had a higher uptake efficiency than both R9 and K9 (Lys nonapeptide). However, at high concentrations, Lys(GEt) nonapeptide exhibited increased cytotoxicity, likely due to its strong cell membrane destabilizing capacity. Furthermore, intracellular pDNA delivery was evaluated using these peptides. Lys(GEt) nonapeptide achieved over 10-fold higher transfection efficiency than both R9 and K9, and at longer incubation times, it significantly outperformed commercially available transfection agent.

We have not only developed novel CPPs containing unnatural amino acids but also elucidated their functional mechanisms. To investigate the role of unnatural amino acids in CPP function, we incorporated five amino acids, L-leucine (Leu), Arg, (S)-α-methylleucine ((S)-(α-Me)Leu), Ac₅c, and Ac₅c with dimethoxy groups (Ac₅c^dOM) into (Arg-Arg-AA)₃, where AA represents the variable amino acid⁴⁸⁾ (Fig. 5a). The helicity, enzymatic resistance, and cell-penetrating ability of these peptides were systematically assessed. Peptides containing dAAs, (S)-(α-Me)Leu, Ac₅c, and Ac₅c^dOM, exhibited similar levels of enzymatic resistance, all of which were significantly higher than those of peptides composed of only natural amino acids (Leu peptide). As a result, the cellular uptake of (S)-(α-Me)Leu, Ac₅c, and Ac₅c^dOM peptides was sustained over time, likely due to their high enzymatic stability. Helicity varied significantly among the five peptides. Notably, (S)-(α-Me)Leu and Ac₅c peptides, which exhibited high helicity, also showed stronger cell-penetrating ability than the other peptides. These results suggested that helicity plays a crucial role in the cell-penetrating ability of Arg-rich peptides. However, (Arg-Arg-AA)₃ peptides tend to adopt a loosely amphipathic structure when forming helices, which may influence their functionality. To isolate the effect of helicity from amphipathicity, we designed Arg₂-AA-Arg₄-AA-Arg₂ peptides, where AA was a dAA including Aib, Ac₅c, Deg, and Dpg⁴⁹⁾ (Fig. 5b). Circular dichroism (CD) spectral analysis revealed the following helicity ranking: Dpg peptide > Deg peptide ≥ Ac₅c peptide > Aib peptide. Additionally, hydrophobicity, as estimated by retention time in ultra performance liquid chromatography, correlated with helicity. Dpg peptide, which exhibited the highest helicity and hydrophobicity, also showed the strongest cell-penetrating ability. These findings suggest that both helicity and hydrophobicity contribute to the high cell-penetrating ability of Arg-rich peptides.

Fig. 5. Peptides for Elucidating Functional Mechanisms of Cell-Penetrating Ability

This review summarizes recent advances in CPPs and their applications to DDS as investigated by our groups. Research on CPPs has been ongoing for over three decades, yielding significant progress. However, challenges remain in translating these findings into DDS applications, and clinical implementation has yet to be achieved. To meet this demand, further refinement and optimization of the functions described here will be necessary. In the near future, CPPs with enhanced safety and functionality are expected to drive progress toward successful clinical applications in humans.

I would like to express my sincere gratitude to my colleagues and collaborators for their invaluable support of my research.

The author declares no conflict of interest.

This review of the author’s work was written by the author upon receiving the 2025 Pharmaceutical Society of Japan Award for Divisional Scientific Promotion.

• 1) Oba M., Demizu Y., “Cell-Penetrating Peptides: Design, Development and Applications,” John Wiley & Sons, Hoboken, NJ, 2022.
• 2) Copolovici D. M., Langel K., Eriste E., Langel Ü., ACS Nano, 8, 1972–1994 (2014).
• 3) Oba M., ChemBioChem, 20, 2041–2045 (2019).
• 4) Kim G. C., Cheon D. H., Lee Y., Biochim. Biophys. Acta. Proteins Proteom., 1869, 140604 (2021).
• 5) Futaki C., Chem. Pharm. Bull., 69, 601–607 (2021).
• 6) Guidotti G., Brambilla L., Rossi D., Trends Pharmacol. Sci., 38, 406–424 (2017).
• 7) Xie J., Bi Y., Zhang H., Dong S., Teng L., Lee R. J., Yang Z., Front. Pharmacol., 11, 697 (2020).
• 8) Goodman M. M., Yu W., Jarkas N., J. Labelled Comp. Radiopharm., 61, 272–290 (2018).
• 9) Cativiela C., Ordonez M., Viveros-Ceballos J. L., Tetrahedron, 76, 130875 (2020).
• 10) Zhang Y., Vanderghinste J., Wang J., Das S., Nat. Commun., 15, 1474 (2024).
• 11) Furukawa K., Oba M., Opiyo G. O., Doi M., Tanaka M., Eur. J. Org. Chem., 2016, 2988–2998 (2016).
• 12) Eto R., Oba M., Ueda A., Uku T., Doi M., Matsuo Y., Tanaka T., Demizu Y., Kurihara M., Tanaka M., Chem. Eur. J., 23, 18120–18124 (2017).
• 13) Furukawa K., Oba M., Toyama K., Opiyo G. O., Demizu Y., Kurihara M., Doi M., Tanaka M., Org. Biomol. Chem., 15, 6302–6305 (2017).
• 14) Yamaberi Y., Eto R., Umeno T., Kato T., Doi M., Yokoo H., Oba M., Tanaka M., Org. Lett., 23, 4358–4362 (2021).
• 15) Oba M., Nonaka H., Umeno T., Kato T., Doi M., Ueda A., Tanaka M., ChemPlusChem, 90, e202400772 (2025).
• 16) Oba M., Ito C., Tanaka M., Bioorg. Med. Chem. Lett., 28, 875–877 (2018).
• 17) Umeno T., Ueda A., Doi M., Kato T., Oba M., Tanaka M., Tetrahedron Lett., 60, 151301 (2019).
• 18) Tanaka M., Chem. Pharm. Bull., 55, 349–358 (2007).
• 19) Crisma M., Toniolo C., Pept. Sci. (Hoboken), 104, 46–64 (2015).
• 20) Nachman R. J., Isaac R. E., Coast G. M., Holman G. M., Peptides, 18, 53–57 (1997).
• 21) Yamaguchi H., Kodama H., Osada S., Kato F., Jelokhani-Niaraki M., Kondo M., Biosci. Biotechnol. Biochem., 67, 2269–2272 (2003).
• 22) Toniolo C., Crisma M., Formaggio F., Peggion C., Pept. Sci. (Hoboken), 60, 396–419 (2001).
• 23) Yamashita H., Demizu Y., Shoda T., Sato Y., Oba M., Tanaka M., Kurihara M., Bioorg. Med. Chem., 22, 2403–2408 (2014).
• 24) Oba M., Demizu Y., Yamashita H., Kurihara M., Tanaka M., Bioorg. Med. Chem., 23, 4911–4918 (2015).
• 25) Oba M., Ito Y., Umeno T., Kato T., Tanaka M., ACS Biomater. Sci. Eng., 5, 5660–5668 (2019).
• 26) Uchida S., Yamaberi Y., Tanaka M., Oba M., Nanoscale, 13, 18941–18946 (2021).
• 27) Kato T., Oba M., Nishida K., Tanaka M., ACS Biomater. Sci. Eng., 4, 1368–1376 (2018).
• 28) Kato T., Yamashita H., Misawa T., Nishida K., Kurihara M., Tanaka M., Oba M., Bioorg. Med. Chem., 24, 2681–2687 (2016).
• 29) Tanisaka H., Kizaka-Kondoh S., Makino A., Tanaka S., Hiraoka M., Kimura S., Bioconjug. Chem., 19, 109–117 (2008).
• 30) Miyamoto T., Tsuchiya K., Numata K., Biomacromolecules, 21, 2735–2744 (2020).
• 31) Yokoo H., Dirisala A., Uchida S., Oba M., ACS Biomater. Sci. Eng., 10, 890–896 (2024).
• 32) Oehlke J., Krause E., Wiesner B., Beyermann M., Bienert M., Protein Pept. Lett., 3, 393–398 (1996).
• 33) Wolf Y., Pritz S., Abes S., Bienert M., Lebleu B., Oehlke J., Biochemistry, 45, 14944–14954 (2006).
• 34) Furukawa K., Tanaka M., Oba M., Bioorg. Med. Chem., 28, 115402 (2020).
• 35) Oba M., Shibuya M., Yamaberi Y., Yokoo H., Uchida S., Ueda A., Tanaka M., Chem. Pharm. Bull., 71, 250–256 (2023).
• 36) Naka M., Umeno T., Shibuya M., Yamaberi Y., Ueda A., Tanaka M., Takemoto H., Oba M., Chem. Pharm. Bull., 72, 512–517 (2024).
• 37) Kato T., Oba M., Nishida K., Tanaka M., Bioconjug. Chem., 25, 1761–1768 (2014).
• 38) Schafmeister C. E., Po J., Verdine G. L., J. Am. Chem. Soc., 122, 5891–5892 (2000).
• 39) Lau Y. H., de Andrade P., Wu Y., Spring D. R., Chem. Soc. Rev., 44, 91–102 (2015).
• 40) Oba M., Kunitake M., Kato T., Ueda A., Tanaka M., Bioconjug. Chem., 28, 1801–1806 (2017).
• 41) Wender P. A., Galliher W. C., Goun E. A., Jones L. R., Pillow T. H., Adv. Drug Deliv. Rev., 60, 452–472 (2008).
• 42) Nakase I., Takeuchi T., Tanaka G., Futaki S., Adv. Drug Deliv. Rev., 60, 598–607 (2008).
• 43) Neu M., Fischer D., Kissel T., J. Gene Med., 7, 992–1009 (2005).
• 44) Kanayama N., Fukushima S., Nishiyama N., Itaka K., Jang W.-D., Miyata K., Yamasaki Y., Chung U.-I., Kataoka K., ChemMedChem, 1, 439–444 (2006).
• 45) Miyata K., Oba M., Nakanishi M., Fukushima S., Yamasaki Y., Koyama H., Nishiyama N., Kataoka K., J. Am. Chem. Soc., 130, 16287–16294 (2008).
• 46) Uchida H., Miyata K., Oba M., Ishii T., Suma T., Itaka K., Nishiyama N., Kataoka K., J. Am. Chem. Soc., 133, 15524–15532 (2011).
• 47) Oba M., Kato T., Furukawa K., Tanaka M., Sci. Rep., 6, 19913 (2016).
• 48) Oba M., Nagano Y., Kato T., Tanaka M., Sci. Rep., 9, 1349 (2019).
• 49) Oba M., Nakajima S., Misao K., Yokoo H., Tanaka M., Bioorg. Med. Chem., 91, 117409 (2023).