Regular Article
e3MPH16: A D-Glutamic Acid-Substituted Peptide for Efficient and Low-Cytotoxicity Cytosolic Delivery of Macromolecules
2025 Volume 73 Issue 9 Pages 907-913
Details
2025 Volume 73 Issue 9 Pages 907-913
Antibody-based therapeutics have shown remarkable success in targeting extracellular molecules, yet their application to intracellular targets remains largely unexplored due to the absence of efficient delivery systems. The large molecular weight and hydrophilicity of immunoglobulin G (IgG) make cytosolic delivery particularly challenging. Previously, we developed a cytosolic delivery peptide, E3MPH16, based on a modified Mastoparan X sequence, which enabled efficient delivery of macromolecules such as dextran with minimal cytotoxicity. However, effective intracellular delivery of antibodies required high concentrations, limiting its practical utility. In this study, we aimed to enhance delivery efficiency while preserving low toxicity by introducing d-amino acid substitutions into E3MPH16. The resulting peptide, e3MPH16, incorporates d-glutamic acid residues at the N-terminus to improve serum stability and protease resistance. Functional analyses demonstrated that e3MPH16 significantly improves cytosolic delivery of Cre recombinase and antibodies compared with the original E3MPH16, without increasing membrane-lytic activity or cytotoxicity. These results underscore the potential of d-amino acid-substituted peptides such as e3MPH16 as a promising platform for the intracellular delivery of unmodified functional antibodies.
Therapeutic antibodies are widely used to target extracellular molecules due to their high specificity and affinity. Despite this success, their application to intracellular targets remains largely unexplored. This limitation stems from the lack of effective delivery systems that can transport sufficient amounts of functional antibodies into the cytosol. Because antibodies are large (approx. 150 kDa) hydrophilic molecules lacking intrinsic membrane permeability, conventional approaches fail to achieve efficient intracellular access. As a result, intracellular antibody functions have been restricted to fixed-cell applications such as immunostaining. Establishing a versatile, functional intracellular antibody delivery platform applicable to unmodified antibodies could open the door to a new class of antibody therapeutics targeting a broad range of intracellular disease-associated proteins, thereby exerting a significant impact on human health.1–3)
Various strategies have been explored to deliver proteins, including antibodies, into the cytosol. These include cell-penetrating peptides (CPPs),4,5) polymeric or dendrimer-based carriers,6) lipid nanoparticles,7,8) complex coacervates,9–11) and microinjection.12) However, these approaches are often limited by drawbacks such as cytotoxicity, poor delivery efficiency, technical complexity, carrier instability, and endosomal entrapment. Additionally, many require chemical or genetic modification of the antibody, which may compromise antigen-binding affinity. Although intracellularly expressed antibodies (intrabodies) represent one alternative, their application remains limited due to challenges in expression and folding within the cytosol.13) Thus, there is a strong demand for safe, efficient, and facile methods to deliver intact, functional antibodies into cells.
We previously developed cytosolic delivery peptides with reduced cytotoxicity by introducing (via substitution or insertion) glutamic acid residues into amphipathic membrane-lytic peptides derived from natural toxins.14–17) Among these, E3MPH16, derived from the amphiphilic wasp venom peptide Mastoparan X and modified with three N-terminal glutamic acids and a C-terminal stretch of 16 histidines (H16),18) showed high delivery efficiency with low toxicity. E3MPH16 was capable of facilitating endosomal escape and delivering Alexa Fluor 488-labeled 10 kDa dextran (Dex10-Alexa488) into nearly 100% of cells.17) However, when delivering antibodies, high concentrations were required, indicating a need for further improvement in delivery efficiency to advance toward practical therapeutic applications.
In this study, we sought to improve the efficiency of cytosolic delivery by introducing d-amino acid substitutions into E3MPH16 while maintaining its low cytotoxicity. d-Amino acids are known to confer increased stability in serum and resistance to proteolytic degradation.19,20) We generated e3MPH16, an E3MPH16 derivative containing d-glutamic acid substitutions at the three N-terminal positions. This substitution significantly enhanced the intracellular delivery of macromolecules such as Cre recombinase and antibodies, without increasing cytotoxicity or membrane-lytic activity. These findings highlight the utility of d-amino acid-substituted peptides such as e3MPH16 as improved vehicles for intracellular antibody delivery.
HT1080 human fibrosarcoma cells were obtained from the Japanese Collection of Research Bioresources (JCRB) Cell Bank. The cells were cultured in high-glucose Dulbecco’s modified Eagle’s medium supplemented with 10% (v/v) heat-inactivated fetal bovine serum (FBS) (Gibco) [DMEM(+)]. HEK293 Cre reporter cells were established from Flp-InTM-293 cells (Invitrogen, Waltham, MA, U.S.A.) and cultured in DMEM(+). These cells were maintained at 37°C in a humidified incubator with 5% CO2 and passaged every 3–4 d. Expi293 cells (Invitrogen) were maintained at 37°C in shake flasks in a humidified, 8% CO2 incubator, rotating at 120 rpm in Expi293 expression medium (Invitrogen) and subculture was conducted every 3–4 d.
Peptide SynthesisPeptides were synthesized on Link amide AM resin (Nova Biochem, London, U.K.) using standard Fmoc solid-phase peptide synthesis (SPPS) protocols with Syro I peptide synthesizer (Biotage, Uppsala, Sweden).17) After synthesis, peptides were cleaved from the resin and deprotected using a mixture of trifluoroacetic acid (TFA) and ethanedithiol (95 : 5) at room temperature (approx. 23°C) for 3 h. The crude peptides were purified by reverse-phase HPLC (RP-HPLC). Peptide masses were confirmed by matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry (Microflex, Bruker Daltonics, Bremen, Germany). Structures of the peptides, observed m/z: E3MPH16 (EEEINWKGIAAJAKKLLHHHHHHHHHHHHHHHH-amide) 4118.5 (calcd for [M + H]+ 4118.1), e3MPH16 (eeeINWKGIAAJAKKLLHHHHHHHHHHHHHHHH-amide) 4118.8 (calcd for [M + H]+ 4118.1), E3mpH16 (EEEinwkgiaajakkllHHHHHHHHHHHHHHHH-amide) 4118.4 (calcd for [M + H]+ 4118.1), e3mpH16(eeeinwkgiaajakkllHHHHHHHHHHHHHHHH-amide) 4118.4 (calcd for [M + H]+ 4118.1), e3MPh16 (eeeINWKGIAAJAKKLLhhhhhhhhhhhhhhhh-amide) 4118.9 (calcd for [M + H]+ 4118.1), e3mph16(eeeinwkgiaajakkllhhhhhhhhhhhhhhhh-amide) 4118.6 (calcd for [M + H]+ 4118.1), respectively. *J = Norleucine; lowercase = d-amino acids.
Cell Viability Assay (WST-8 Assay)Cell viability was assessed using the Cell Counting Kit-8 (CCK-8) (Dojindo, Kumamoto, Japan) according to the manufacturer’s instructions. HT1080 cells were seeded in 96-well microplates (Iwaki, Tokyo, Japan) and incubated for 24 h until reaching 80–90% confluence. After removing the medium and washing with phosphate-buffered saline (PBS), 100 μL of peptide solution in DMEM(+) was added, followed by incubation at 37°C for 18 h. Cells were then washed twice with PBS containing 0.5 mg/mL heparin sodium, and 100 μL of DMEM(+) containing 10 μL of WST-8 reagent was added, to eliminate any residual peptides and evaluate viability after peptide removal. After 2 h of incubation at 37°C, the color reaction was measured at 450 nm, and the reference was measured at 650 nm with a Multiskan FC microplate photometer (Thermo Fisher Scientific, Waltham, MA, U.S.A.).
Liposome PreparationLipids were purchased from Avanti Polar Lipids. Neutral and acidic large unilamellar vesicles (LUVs) were prepared by mixing 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) and 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphatidylglycerol (POPG) at molar ratios of 1 : 0 and 3 : 1, respectively.14) Lipids dissolved in chloroform were mixed in appropriate ratios, and solvents were removed by rotary evaporation followed by overnight vacuum drying. The lipid film was rehydrated with a dye buffer containing 12.5 mM 8-aminonaphthalene-1,3,6-trisulfonic acid (ANTS), 45 mM p-xylene-bis-pyridinium bromide (DPX), 150 mM NaCl, and 10 mM 2-(N-morpholino)ethanesulfonic acid (MES) (pH 5.0 or 7.0), followed by vortexing and five freeze–thaw cycles. The suspension was extruded through a polycarbonate filter (100 nm) using an extruder (Avanti Polar Lipids). The LUVs were then purified by gel filtration using PD-10 columns (GE Healthcare) with elution buffer containing 150 mM NaCl and 10 mM MES (pH 5.0 or 7.0) to remove unencapsulated dye. Total lipid concentrations were determined enzymatically based on phosphatidylcholine (LabAssay™ Phospholipid, FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan).
ANTS/DPX Leakage AssayMembrane-lytic activity of peptides was evaluated by fluorescence dequenching due to the leakage of ANTS and DPX from LUVs, as previously described.15) Peptide solutions in 10 mM MES buffer (pH 5.0 or 7.0, containing 150 mM NaCl) were added to LUVs containing ANTS/DPX (final lipid concentration: 200 μM). After shaking for 1 h at 37°C, fluorescence changes were measured using a VICTOR Nivo multimode plate reader (PerkinElmer, Inc. Life Sciences) with excitation and emission filters set at 315 nm and 455 nm, respectively. Leakage (%) was calculated as follows: Leakage (%) = 100 × ([Fobs – F0]/[Fmax – F0]), where Fobs represents the observed fluorescence intensity, Fmax is the maximum fluorescence intensity after adding Triton X-100 at a final concentration of 0.1% (v/v), and F0 represents the initial fluorescence intensity.
Microscopic Observation of Dex10-Alexa488 UptakeHT1080 cells were seeded in 35-mm glass-bottomed dishes (Iwaki) and cultured for 24 h until reaching 80–90% confluence. After two PBS washes, cells were incubated with 200 μg/mL Alexa Fluor 488-labeled 10 kDa dextran (Dex10-Alexa488; Invitrogen) and 10 μM E3MPH16 or e3MPH16 in DMEM(+) at 37°C for 18 h. Cells were then washed twice with PBS containing 0.5 mg/mL heparin sodium and stained with 5 μg/mL Hoechst 33342 (Invitrogen). Dextran uptake was visualized using FV3000. NIH ImageJ 1.53k and Python 3.8.8 were used to automatically count cells exhibiting diffuse cytosolic Dextran signals, with more than 200 cells analyzed per condition. Nuclear fluorescence intensities were also quantified using ImageJ.
Cre-loxP Recombination AssayHEK293 Cre reporter cells were seeded in 24-well microplates and cultured in DMEM(+) for 24 h. Cells were treated with 5 μM Cre recombinase, purified according to a previously reported protocol,21) in the presence or absence of peptides for 24 h at 37°C. Recombinant efficiency was assessed by flow cytometric analysis of green fluorescent protein (GFP)-positive cells.
Antibody PreparationAnti-GFP immunoglobulin G (IgG) was produced using the Expi293™ Expression System (Thermo Fisher Scientific). A plasmid encoding anti-GFP IgG (Addgene #114492) (25 μg) was transfected into Expi293 cells (3 × 106 cells/mL, 25 mL) using ExpiFectamine 293 (Thermo Fisher Scientific) according to the manufacturer’s protocol. At 20 h post-transfection, ExpiFectamine293 Transfection Enhancer 1 (150 μL) and Enhancer 2 (1.5 mL) were added, and the culture was incubated for 5 d. After centrifugation, the supernatant containing the antibody was filtered (0.45 μm) and loaded onto a HiTrap Protein A High Performance column (GE Healthcare). The IgG was eluted using a pH gradient with 0.1 M glycine buffer (pH 2.7) and neutralized with 1 M Tris–HCl. Antibody-containing fractions were dialyzed against PBS and stored at −80°C.
Intracellular Delivery of Anti-GFP-IgGIntracellular Antibody Delivery HT1080 cells were transfected with a plasmid encoding GFP-HRas(G12V) using Lipofectamine LTX (Thermo Fisher Scientific) following the manufacturer’s instructions. After washing twice with PBS, the cells were treated with 200 or 400 μg/mL anti-GFP IgG in DMEM(+) with or without 10 μM E3MPH16 or e3MPH16 for 18 h at 37°C. Cells were then washed with PBS containing 0.5 mg/mL heparin sodium, fixed with 4% paraformaldehyde/PBS, and permeabilized with 0.1% Triton X-100/PBS. Samples were incubated with Alexa Fluor 568-labeled anti-mouse IgG secondary antibody, followed by nuclear staining with 5 μg/mL Hoechst 33342. Colocalization of anti-GFP IgG and GFP-HRas(G12V) was observed using a confocal laser scanning microscope FV3000 (Olympus, Tokyo, Japan).
E3MPH16 is a peptide derived from Mastoparan X that promotes macropinocytosis and enables the endosomal escape of macromolecules, although this process requires approximately 18 h. Since E3MPH16 is composed entirely of l-amino acids, its activity may be reduced in serum due to proteolytic degradation. By contrast, d-amino acid substitution has been widely reported to enhance serum stability and confer resistance to proteolysis by reducing recognition by endogenous proteases.20,22) Therefore, five variants of E3MPH16 were designed using d-amino acid substitutions at residues known to modulate cytotoxicity (E3), membrane-disruptive activity (MP), and cellular uptake efficiency (H16). These peptides were synthesized via standard Fmoc solid-phase peptide synthesis (Table 1).
 |
Specifically, e3MPH16 was generated by substituting the three N-terminal glutamic acid residues (E3) with d-enantiomers to suppress degradation by aminopeptidases. E3mpH16 contains d-substituted MP to block endoprotease cleavage. The variant e3mpH16 combines d-substitutions in both E3 and MP, thereby suppressing degradation from both ends. e3MPh16, which also includes a d-substituted H16 segment, is expected to resist both N- and C-terminal proteolysis. Finally, e3mph16, composed entirely of d-amino acids, was anticipated to be the most protease-resistant peptide.
While direct experimental evidence of enhanced proteolytic stability was not obtained in this study, as the stability of the peptides was not systematically evaluated, our design rationale is based on well-established literature demonstrating that d-amino acid substitution improves peptide stability in serum and inhibits enzymatic recognition by endogenous proteases.
d-Amino Acid Substitution Does Not Alter Membrane-Lytic ActivityThe outer leaflet of the plasma membrane primarily consists of neutral phospholipids under physiological pH (approx. 7.0), while the inner leaflet of endosomal membranes contains acidic phospholipids, including bis(monoacylglycero)phosphate (BMP), in an acidic environment (pH 4–6).23,24) We previously demonstrated that E3MPH16 causes minimal damage to neutral lipid large unilamellar vesicles (LUVs) composed of POPC (a plasma membrane model) at pH 7.0, but exerts significant membrane-disruptive activity toward LUVs containing POPG at pH 5.0 (an endosomal membrane model). This pH- and lipid-dependent selectivity allows for efficient endosomal escape while minimizing cytotoxicity.
We evaluated the membrane-lytic activity of E3MPH16 and its d-amino acid variants using a liposome leakage assay. LUVs were prepared using POPC (neutral, pH 7.0) or POPC : POPG = 3 : 1 (acidic, pH 5.0), and encapsulated the fluorophore ANTS and its quencher DPX. Upon membrane disruption, ANTS is released and its fluorescence recovers, allowing quantification of peptide-induced membrane perturbation.
Consistent with prior findings, E3MPH16 showed minimal leakage at pH 7.0 (POPC) and induced strong leakage at pH 5.0 (POPC : POPG = 3 : 1) at a peptide-to-lipid (P/L) ratio of 0.1 (Figs. 1A, 1B). All d-substituted peptides exhibited similar behavior: negligible leakage in the plasma membrane model and pronounced leakage in the endosomal model. Since membrane lytic activity is reported to be independent of peptide chirality,22) our results confirm that d-substituted peptides retain membrane-disruptive function. Moreover, hybrid peptides with mixed l- and d-amino acid regions displayed similar activity, suggesting that stereochemical configurations of E3, MP, and H16 do not significantly affect endosomal membrane disruption. Taken together, d-amino acid substitution improves peptide stability while preserving selective membrane activity beneficial for endosomal escape.
(A) Membrane lytic activity of each peptide against the cell membrane model (POPC LUVs at pH 7.4), and (B) the endosomal membrane model (POPC/POPG (3 : 1) LUVs at pH 5.0). E3MPH16; light green circle, e3MPH16; dark blue, E3mpH16; orange circle, e3mpH16; ligh blue circle, e3MPh16; dark green circle, e3mph16; purple circle. Data are presented as mean ± standard error (S.E.) (n = 3).
We next assessed cytotoxicity of the peptides in serum-containing medium using the WST-8 assay in HT1080 cells. Consistent with previous reports, E3MPH16 exhibited no detectable cytotoxicity up to 100 μM. Likewise, e3MPH16 (d-substituted E3) showed no increase in cytotoxicity. However, e3MPh16, which includes a d-substituted H16 region, caused a slight increase in toxicity. By contrast, peptides containing d-substituted MP (e3mpH16, E3mpH16, e3mph16) exhibited substantial cytotoxicity, with e3mph16 (all-d peptide) showing the highest toxicity (CC50 ≈ 10 μM).
These findings suggest that the E3 region plays a key role in cytotoxicity suppression, and that its d-substitution enhances stability while preserving this function. In e3MPh16, simultaneous N- and C-terminal stabilization may inhibit degradation of the MP segment, resulting in partial toxicity. In the MP-d substituted variants, increased serum and intracellular stability of the MP motif likely led to prolonged membrane-lytic activity, consistent with their higher cytotoxicity. Given that liposome leakage assays showed no differences in membrane-lytic activity (Fig. 2), the observed cytotoxicity is attributed to enhanced peptide persistence rather than intrinsic changes in membrane interaction. Thus, optimal peptide design requires both stable glutamic acids to suppress toxicity and a degradable MP motif to avoid prolonged damage. These criteria are fulfilled by e3MPH16, which retains low cytotoxicity.
Viability of HT1080 cells after 18 h treatment with varying concentrations of E3MPH16 (light green circle), e3MPH16 (dark blue), E3mpH16 (orange circle), e3mpH16 (ligh blue circle), e3MPh16 (dark green circle), and e3mph16 (purple circle). Data are presented as mean ± S.E. (n = 3).
We next evaluated whether e3MPH16 enhances cytosolic delivery of macromolecules. Alexa Fluor 488-labeled 10 kDa dextran (Dex10-Alexa488) was used as a model cargo. HT1080 cells were treated with Dex10-Alexa488 in the presence of 10 μM E3MPH16 or e3MPH16 for 18 h and imaged by confocal laser scanning microscopy (CLSM) (Fig. 3A). In control cells treated with Dex10-Alexa488 alone, punctate endosomal signals were observed. By contrast, co-treatment with either peptide resulted in diffuse cytosolic fluorescence in nearly all cells (Fig. 3B).
(A) Confocal images of HT1080 cells co-treated with 10 μM E3MPH16 or e3MPH16 and 200 μg/mL Dex10-Alexa488 for 1 h. Scale bars, 50 μm. (B) Percentage of cells having diffuse cytosolic Dex10-Alexa488 signal. (C) Quantification of cytosolic fluorescent intensity of Dex10-Alexa488. Data are presented as mean ± S.E. (n = 3). ***p < 0.001 and n.s. signifies not significantly different. The statistical analysis was conducted using a one-way ANOVA, followed by Tukey–Kramer’s honestly significant difference test.
Quantification of fluorescence intensities revealed that a larger proportion of cells exhibited strong cytosolic signals when treated with e3MPH16 (Fig. 3C). These results indicate that d-substitution at the N-terminal E3 region enhances peptide stability in serum and endosomes, leading to greater peptide accumulation and more efficient endosomal escape.
Our previous studies demonstrated that E3MPH16 utilizes macropinocytosis for cellular uptake and facilitates cytosolic delivery via endosomal escape. Given that e3MPH16 retains the same H16 motif, we speculate that a similar macropinocytic pathway underlies its efficient intracellular delivery of macromolecules.
Enhanced Delivery of Cre Recombinase by e3MPH16To test whether the delivery enhancement extends to functional proteins, we used the Cre/loxP recombination assay. We previously established HEK293 reporter cells harboring a loxP-DsRed-loxP-EGFP construct.21) These cells constitutively express DsRed; upon Cre-mediated recombination between loxP sites, EGFP expression is induced, allowing recombination efficiency to be monitored by EGFP expression (Fig. 4A). Cells were treated with 5 μM Cre recombinase in the presence of 10 μM E3MPH16 or e3MPH16 for 24 h, and recombination efficiency was quantified by flow cytometry. E3MPH16 induced GFP expression in approx. 30% of cells, while e3MPH16 increased the percentage to over 40% (Fig. 4B–4D). Notably, in our previous studies, cytosolic delivery peptides such as L17E and HAad achieved 25–40% recombination efficiency at 40 μM in the presence of 5 μM Cre recombinase.25) By contrast, e3MPH16 achieved a comparable recombination efficiency (approx. 40%) at a much lower concentration (10 μM), without detectable cytotoxicity (CC50 > 100 μM). These data demonstrate that e3MPH16 is not only more effective than E3MPH16 but also comparable to other established cytosolic delivery peptides in terms of functional delivery efficiency, even at significantly lower and safer concentrations.
(A) Schematic representation of Cre-loxP recombination assay. (B, C) Recombination efficiency of EGFP-positive cells in loxP-reporter HEK293 cells after treatment with 5 μM Cre or both 5 μM Cre and 10 μM (B) E3MPH16 or (C) e3MPH16 for 24 h. Cre/loxP recombination in HEK293 reporter cells was quantified by flow cytometry. (D) Quantification of recombination efficiency based on EGFP-positive cells. Data are presented as mean ± S.E. (n = 3). ***p < 0.001 and n.s. signifies not significantly different. The statistical analysis was conducted using a one-way ANOVA, followed by Tukey–Kramer’s honestly significant difference test.
We further evaluated whether e3MPH16 can deliver full-length IgG antibodies into cells and enable intracellular antigen recognition. HT1080 cells were transiently transfected with GFP-HRas(G12V), a mutant Ras protein known to activate Raf1 and Rac and serve as a cytosolic antigen.26,27) Anti-GFP IgG was produced using the Expi293 expression system and added to the cells with or without 10 μM peptide for 24 h. Following fixation and permeabilization, cells were stained with Alexa Fluor 568-labeled secondary antibodies (Fig. 5).
HT1080 cells transiently expressing GFP-HR as (G12V) were treated with 200 or 400 μg/mL anti-GFP IgG in the presence of 10 μM E3MPH16 or e3MPH16 for 18 h. Fixed cells were stained with Alexa568-conjugated secondary antibody. Scale bars, 50 μm.
No Alexa568 signal was observed in cells treated with antibody alone, indicating that IgG did not enter cells by passive uptake. Both E3MPH16 and e3MPH16 facilitated the colocalization of IgG and GFP when anti-GFP-IgG was applied at 400 μg/mL. By contrast, at a lower concentration (200 μg/mL), only e3MPH16 resulted in appreciable colocalization, whereas little overlap was observed in E3MPH16-treated cells. These results suggest that e3MPH16 achieves improved delivery efficiency even for large molecules such as IgG.
In this study, we rationally designed d-amino acid-substituted derivatives of the endosomal escape peptide E3MPH16 to improve serum stability and protease resistance. Through a systematic screen of d-substituted variants, we identified e3MPH16, in which the N-terminal glutamic acids were replaced with d-amino acids, as a candidate with low cytotoxicity. e3MPH16 retained membrane-disruptive activity and exhibited enhanced delivery of dextran, Cre recombinase, and full-length IgG into the cytosol. Notably, e3MPH16 outperformed the parent peptide E3MPH16 in delivering functional proteins and antibodies. Furthermore, e3MPH16 efficiently delivers macromolecules into various cell types, such as HT1080 and HEK293 cells, without significant cytotoxicity. These features highlight its potential as a broadly applicable carrier for cytosolic delivery of biologics. By attaching tumor-targeting moieties to enhance cancer cell selectivity, e3MPH16 is expected to expand its potential toward therapeutic applications, including cancer treatment.
Anti-GFP [N86/38.1R] was a gift from James Trimmer (Addgene plasmid # 114492). This work was supported by JSPS KAKENHI [Grant Nos.: JP21H04794 (to S.F.), 23K17412 (to S.F.), 24H00051 (to S.F.), JP22K15250 (to Y.K.), and JP24H00842 (to Y.K.)], and by JST CREST [Grant No.: JPMJCR18H5 (to S.F.)], JST ACT-X (Grant No.: 22715691) to Y.K., by ISHIZUE 2024 of Kyoto University to Y.K. and by The Suzuken Memorial Foundation to Y.K.
Y.K. and S.F. are inventors on a patent application related to the findings reported in this manuscript (Application No. JP2023-174558, by Kyoto University). The other authors declare no conflicts of interest.
