Cytosolic protein delivery via protein-bound microparticles based on anionic boron clusters and cationic polymers

Abstract

Direct protein delivery to the cytosol facilitates immediate functional expression of proteins without the risks associated with gene introduction. However, the technology for delivering various proteins to the cytosol is still in its infancy. Herein, the formation of microparticles comprising anionic boron clusters and the cationic polymer hexadimethrine bromide (HDB) is demonstrated. In particular, the microparticles formed from dodecabromododecaborate clusters and HDB are confirmed to be bound with proteins. The protein-bound boron cluster/polymer-based microparticles (protein·BPMs) are internalized into cells via endocytosis. Upon internalization, the protein·BPMs release the proteins with different isoelectric points and sizes into the cytosol. Furthermore, an enzyme is delivered by protein·BPMs into the cytosol of various cell types while maintaining its functional activity. This method, owing to the simple preparation of protein·BPMs, represents a promising approach for delivering diverse proteins to various cell types. Our findings open new avenues for utilizing boron clusters in cytosolic delivery systems.

---

Introduction

Proteins play diverse and essential roles within cells, including signal transduction, transcription, translation, cell motility, and metabolism. Direct intracellular delivery of proteins offers several advantages, such as avoiding the risks associated with gene introduction^1,2 and bypassing the need for transcription, translation, and folding processes, thereby enabling immediate functional activity. However, most proteins are hydrophilic and have high molecular weights, making it challenging for them to traverse the cell membrane and exert their intended intracellular functions. To address this challenge, various approaches have been proposed, including the use of polymers,^3–5 lipids,^6,7 extracellular vesicles,⁸ and peptides.⁹ While these methods offer unique advantages, they are still limited in their applicability, highlighting the need for direct intracellular delivery techniques capable of accommodating a broader range of proteins and biological environments.

Anionic boron clusters¹⁰ form polymer particles with various polymers. It has been reported that cobalt bis(1,2-dicarbollide) (COSAN), a metallacarborane cluster, forms water-insoluble complexes with polyethylene oxide.¹¹ Polymer particles composed of block copolymers and COSAN, which release boron in response to stimuli such as alkali metal cations¹² and heat,¹³ have been developed. Additionally, regarding dodecaborate clusters, charge-compensated polymer particles formed through electrostatic interactions between dodecahydrododecaborate clusters (B₁₂H₁₂²⁻) and cationic polymers have been introduced into cells and are researched as promising tools for boron delivery, particularly targeting boron neutron capture therapy.^14,15

On the other hand, anionic boron clusters, such as dodecaborate clusters (B₁₂X₁₂²⁻, X = Cl, Br, I)^16–18 and COSAN,¹⁹ and other anionic clusters such as polyoxometalates,²⁰ have been shown to facilitate the intracellular transport of cationic and neutral small molecules that cannot naturally traverse the cell membrane. A notable characteristic of these clusters is their superchaotropic effect. This phenomenon was independently discovered by Assaf et al. in the context of boron clusters²¹ and by Naskar et al. for polyoxometalates adsorbed onto the surface of nonionic micelles.²² The superchaotropic effect is considered a water-mediated phenomenon driven by an enthalpically favorable dehydration process,^23,24 and more recently supported by theoretical studies.²⁵ Owing to this effect, cationic or neutral molecules, impermeable to the cell membrane, have been shown to be transported into the cytosol, thereby facilitating intracellular delivery.^16,17 Additionally, anionic boron clusters with superchaotropicity have a high affinity for hydrophilic–hydrophobic interfaces and have been reported to affect lipid chain structures in experiments using zwitterionic membranes.²⁶ Furthermore, COSAN and B₁₂I₁₂²⁻ have been shown to form reversible assemblies with proteins, with minimal impact on their secondary structures.²⁷ A significant achievement was reported in which the anionic boron cluster B₁₂Br₁₁OCH₂CH₂CH₃²⁻ successfully delivered the positively charged protein cytochrome C into the cytosol by permeating the cell membrane.²⁸ However, no studies have yet utilized anionic boron clusters for the cytosolic delivery of negatively charged proteins.

In this study, we demonstrated that boron cluster/polymer-based microparticles (BPMs) were formed by simply mixing dodecabromododecaborate clusters (B₁₂Br₁₂²⁻) with the cationic polymer hexadimethrine bromide (HDB) in water (Fig. 1a). HDB is a cationic polymer containing quaternary ammonium groups and is known to enhance the transduction efficiency of certain viruses into target cells.²⁹ Owing to its low cytotoxicity and high compatibility with cellular uptake, it has been widely used in various biological applications. Furthermore, we confirmed that the addition of proteins to the BPMs resulted in the formation of protein-bound BPMs. For clarity, we designate the protein-bound BPMs as protein·BPMs (Fig. 1a). Analysis revealed that the protein·BPMs were internalized into cells via endocytosis, rather than membrane permeation, and successfully delivered not only positively charged but also negatively charged and high-molecular-weight proteins into the cytosol. These findings will contribute to the establishment of a powerful tool for cytosolic protein delivery using boron clusters.

Fig. 1 (a) Schematic diagram of boron cluster/polymer-based microparticles (BPMs) formed from dodecaborate clusters (B₁₂X₁₂²⁻) and HDB, and protein-bound BPMs (protein·BPMs). (b) Confocal microscopic images of BSA-FITC·2. (c) Transmission electron microscope (TEM) images of BSA·2.

Results and discussion

When B₁₂Br₁₂²⁻ and HDB were simply mixed in water, the formation of numerous microparticles, BPMs of 2, was observed using phase-contrast microscopy and fluorescence microscopy (Fig. S1a,† Water, arrowheads). This occurred under the condition where the anionic charge of B₁₂Br₁₂²⁻ exceeds the cationic charge of HDB (HDB/B₁₂Br₁₂²⁻ = 0.18). In contrast, when B₁₂H₁₂²⁻ and HDB were mixed at the same concentration ratio to form BPMs of 1, only a small number of microparticles were observed (Fig. S1a,† Water, arrows). Dynamic light scattering (DLS) measurements of this solution revealed multimodal peaks, both containing microparticles of around 1000 nm in size (Fig. S1b†). In contrast to 1, which exhibited a positive zeta potential, 2 showed a negative value. A similar phenomenon has also been observed in liposomes composed of zwitterionic lipids, where the addition of B₁₂Br₁₂²⁻ to liposomes with an initial zeta potential near 0 mV causes the zeta potential to shift to a negative value. This serves as evidence that B₁₂Br₁₂²⁻ strongly adsorbs to the liposomal membrane.²⁶ These observations collectively suggest that B₁₂Br₁₂²⁻ interacts more strongly with HDB than B₁₂H₁₂²⁻ does. Next, when 0.5 μM of fluorescein isothiocyanate-labeled BSA ([FITC]-labeled BSA, BSA-FITC) was added to the BPMs, microparticles were obtained in the same mannner (Fig. S1c†). Fluorescence microscopic imaging confirmed the fluorescence signals of BSA-FITC in both BPMs of 1 and 2, indicating that BSA-FITC bound to the each BPMs. The formation of microparticles in 2 or BSA·2 was confirmed by DLS, even when the ratio of B₁₂Br₁₂²⁻ to HDB was varied (Fig. S2†). When the cationic charge of HDB was less than the anionic charge of B₁₂Br₁₂ (HDB/B₁₂Br₁₂ < 1), both 2 and BSA·2 exhibited negative zeta potentials. As the proportion of cationic charge from HDB increased, the absolute values of the zeta potentials of both 2 and BSA·2 decreased. When the cationic charge of HDB exceeded the anionic charge of B₁₂Br₁₂ at an HDB/B₁₂Br₁₂ ratio of 1.12, both 2 and BSA·2 showed positive zeta potentials (Fig. S2†). Additionally, DLS analysis confirmed that microparticles were formed in BSA·2, regardless of changes in its concentration (Fig. S3†). For particle sizes exceeding 1 μm, particle size distribution analysis was performed using confocal microscopic images. Most BSA-FITC·2 particles were concentrated around 2 μm, while a minority were distributed across a broader range of 1–27 μm (Fig. 1b and S1d†). When BSA·2 was analyzed under electron microscopy, irregularly shaped particles were observed (Fig. 1c). These structures were also observed in various complexes such as 1, 2, and BSA·1 (Fig. S4†). Both 2 and BSA-FITC·2 were precipitated by centrifugation at 5000g for 2 minutes (Fig. S5†). Additionally, ultrasonic treatment of BSA·2 reduced particle size, likely due to the fragmentation of larger particles, resulting in a unimodal intensity distribution with a peak around 101 nm (Fig. S6a†). This result was confirmed by TEM, which revealed numerous smaller particles with an estimated particle size distribution of approximately 122 nm (Fig. S6b and S6c†). These findings indicate that ultrasonic treatment can reduce the size of protein·BPMs from the microscale size to the nanoscale size. By mixing anionic dodecaborate clusters with cationic HDB in the presence of an excess amount of NaCl to suppress electrostatic interactions, BPMs of 1 disappeared, whereas BPMs of 2 retained their microparticle structure (Fig. S1a,† 1.0 M NaCl, arrowhead). This observation suggests that B₁₂Br₁₂²⁻ engages in a stronger non-electrostatic interaction with HDB than B₁₂H₁₂²⁻. Indeed, a comparison of B₁₂Br₁₂²⁻ and B₁₂H₁₂²⁻ indicates that B₁₂Br₁₂²⁻ exhibits stronger chaotropic effects,^18,23 further supporting the hypothesis that the chaotropic effect plays a significant role in BPM formation involving B₁₂Br₁₂²⁻ and HDB. Furthermore, the reversible binding between proteins and dodecaborate clusters with strong chaotropic effects has been demonstrated,²⁷ suggesting that the interaction between B₁₂Br₁₂²⁻ and the protein contributes to the formation of protein·BPMs.

Next, to investigate whether protein·BPMs could serve as carriers for cytosolic protein delivery, BSA-FITC·1 was prepared and applied to human osteosarcoma cells (U-2 OS cells). However, when BSA-FITC·1 was applied to the cells, the diffused cytosolic fluorescence signal of BSA-FITC was detected in only 0.4% of cells, which is almost the same as 0.2% when BSA-FITC alone was applied to the cells (Fig. 2a and b). In contrast, when BSA-FITC·2 was applied to the cells, micro-sized structures were observed inside the cells, and the diffused cytosolic fluorescence signal of BSA-FITC was detected in 62% of cells, indicating that successful cytosolic delivery of BSA-FITC was achieved (Fig. 2a and b). Similar results were observed when sodium was used as the counter ion instead of cesium (Fig. S7†). Furthermore, cytosolic localization of BSA-FITC by BSA-FITC·2 was also confirmed in live cells (Fig. S8†), indicating that the cytosolic localization observed in fixed cells is not an artifact of the fixation process. When BSA-FITC·2 was separated by centrifugation and the supernatant and precipitate fractions were added to the cells, BSA-FITC was not delivered to the cytosol in the supernatant fraction, but BSA-FITC was delivered to the cytosol in the precipitate fraction (Fig. 2c, centrifuged). When sonicated BSA-FITC·2 was added, micro-sized structures were no longer observed, and nano-sized structures were detected in the cells (Fig. 2c, sonicated, enlarged). This suggests a correlation with the pre-generated nanoparticles formed by sonication; however, unlike the untreated microparticles condition (Fig. 2c, untreated, enlarged), the fluorescence intensity was weaker, and a smaller amount of cytosolic delivery of BSA-FITC was observed (Fig. 2c, sonicated). The difference in fluorescence intensity depending on particle size may be attributed to the fact that while nanoparticles remain suspended in solution, microparticles sediment and accumulate near the cells at the bottom of the plate, facilitating cellular uptake. Since protein·BPMs are efficiently taken up by cells while maintaining their micro-size, considering the risk of protein denaturation caused by sonication, although proteins were not degraded by sonication, at least under the experimental conditions used in this study (Fig. S6d†), the approach of directly introducing protein·BPMs into cells without further processing is considered rational. Single reagents used to prepare BSA-FITC·2 did not facilitate BSA-FITC delivery into the cytosol (Fig. S9†). BSA-FITC·2 delivered proteins into the cytosol across various HDB to B₁₂Br₁₂²⁻ ratios (Fig. S10†), with no significant difference in cytotoxicity (Fig. S11†). However, the efficiency of BSA-FITC delivery by BSA-FITC·2 was dependent on the concentrations of B₁₂Br₁₂²⁻ (Fig. S12†).

Fig. 2 Cytosolic protein delivery by protein·BPMs. (a) Cytosolic delivery of BSA-FITC by BSA-FITC·BPMs. U-2 OS cells were treated with BSA-FITC·BPMs for 1 h. After washing, the cells were incubated with Dulbecco's Modified Eagle's medium (DMEM) containing 10% fetal bovine serum (FBS) (DMEM (+)) for 24 h, followed by confocal microscopy. (b) The delivery efficiency of BSA-FITC·BPMs. Data are mean ± standard error of the mean (SEM) (n = 3). In each experiment, at least 295 cells were calculated. Statistical significance was assessed by one-way analysis of variance (ANOVA), followed by Tukey's multiple comparisons test with a 95% confidence interval. ****: P < 0.0001, N.S.: not significant. (c) Effect of sonication and centrifugation on cytosolic protein delivery by BSA-FITC·2. U-2 OS cells were treated with untreated, sonicated or centrifuged BSA-FITC·2 for 1 h. After washing, the cells were incubated with DMEM (+) for 24 h, followed by confocal microscopy.

To elucidate the mechanism by which BSA-FITC·2 pass through the cell membrane, we first tested whether BSA-FITC-derived signals could also be labeled by a lysosomal probe, LysoTracker. Some BSA-FITC-derived signals were also labeled by LysoTracker (Fig. S13a,† an arrow), suggesting that BSA-FITC·2 was taken up by the cells by endocytosis. To further examine by which endocytosis BSA-FITC·2 is taken up by cells, we tested the effects of several endocytosis inhibitors on protein delivery. When cells were treated with latrunculin B (an actin polymerization inhibitor that inhibits macropinocytosis), dynasore (a dynamin-dependent endocytosis inhibitor), chlorpromazine (CPZ; a clathrin-mediated endocytosis inhibitor), sucrose (a clathrin-mediated endocytosis inhibitor), or methyl-β-cyclodextrin (MβCD; a lipid raft-dependent endocytosis inhibitor), the cytosolic signals of BSA-FITC observed when the inhibitors were not treated (Fig. S13b,† arrows) were not observed (Fig. S13b†). On the other hand, even when cells were treated with genistein (a caveolin-mediated endocytosis inhibitor), the cytosolic signals of BSA-FITC were observed (Fig. S13b,† Genistein +, an arrowhead). These results suggest a possible involvement of macropinocytosis, dynamin-dependent endocytosis, clathrin-mediated endocytosis and lipid raft-dependent endocytosis in the endocytotic uptake of BSA-FITC·2.

To verify how proteins leave the protein·BPMs and are released from the endosome into the cytosol, we first measured the particle size distribution and the zeta potential of BSA·2 at pH 7.4, 6.8 and 5.0. As the pH decreased, the particle size of BSA·2 increased slightly, and the absolute value of zeta potential declined (Fig. S14a†). The reduction in the absolute value of the zeta potential is likely due to partial protonation of BSA molecules on the BSA·2 surface under acidic conditions. Next, to examine the possibility that proteins are reversibly released from protein·BPMs within endosomes, we added liposomes to centrifuged BSA-FITC·2 and monitored changes in fluorescence. Since most BSA-FITC·2 were sedimented, their fluorescence intensity was initially much lower than that of free BSA-FITC. However, upon the addition of liposomes, we observed a rapid increase in fluorescence intensity (Fig. S14b†). This suggests that BSA-FITC was released into the supernatant from BSA-FITC·2. We attribute this to the phospholipids in the liposomes extracting B₁₂Br₁₂²⁻ from the BSA-FITC·2, thereby disrupting the chaotropic interactions between BSA-FITC and 2, and facilitating their dissociation. The precise mechanism by which these proteins efficiently escape from endosomes remains unclear; however, based on the previously reported finding that B₁₂Br₁₂²⁻ strongly interacts with lipid chains and reduces the thickness of lipid bilayers,²⁸ together with our present results, we currently propose the following mechanism. In the endosomal environment, where interactions with cellular membranes are enhanced, B₁₂Br₁₂²⁻ is transferred from the protein·BPMs to the endosomal membrane, leading to membrane swelling and disruption. In this process, the slight increase in particle size caused by the decrease in pH may also contribute to some extent. Simultaneously, the loss of B₁₂Br₁₂²⁻ from the protein·BPMs weakens their association with the protein, resulting in the release of the protein into the cytosol.

We also investigated the delivery of other proteins into cells using protein·2. Both FITC-labeled lysozyme and Alexa Fluor 555-labeled immunoglobulin G (IgG) (IgG-Alexa 555) were successfully delivered into the cytosol by these protein·2 (Fig. 3a, b, and S14†). Given that the isoelectric points of BSA and lysozyme are 4.7 and 10.8, respectively, and that IgG has a molecular weight of approximately 150 kDa; protein·2 can deliver both negatively and positively charged proteins under physiological conditions, as well as relatively large proteins. The localization pattern of IgG-Alexa 555 when ProteoCarry, a commercially available protein delivery agent, was used at the same protein concentration as IgG-Alexa 555·2 indicated that cytosolic protein delivery by protein·2 was successfully achieved (Fig. 3b). In contrast, enhanced green fluorescent protein (EGFP) with a nuclear localization signal (NLS) added to the C-terminus (EGFP-NLS) was not delivered into the cytosol by the EGFP-NLS·2. Consequently, it was not translocated into the nucleus via the NLS. This was probably owed to the failure in endosomal escape. However, when polyethyleneimine (PEI) Max, which promotes endosomal escape via the proton sponge effect and/or the membrane destabilization,³⁰ was added alongside B₁₂Br₁₂²⁻ and HDB, EGFP-NLS was successfully delivered to the nucleus (Fig. 3c). This finding suggests that specific proteins may require additional reagents, such as PEI Max, to facilitate endosomal escape for effective cytosolic delivery. Finally, to assess whether the delivered protein retained its functionality, we used beta-galactosidase (β-Gal) as a model enzyme. β-Gal delivered into U-2 OS cells by β-Gal·2 exhibited enzymatic activity, catalyzing the hydrolysis of 5-bromo-4-chloro-3-indolyl-β-D-galactoside into an insoluble blue dye (Fig. 3d). This intracellular enzymatic activity was also observed in HeLa and A549 cells (Fig. 3d), suggesting that β-Gal·2 can deliver functional proteins across various cell types.

Fig. 3 Cytosolic delivery of various types of proteins by protein·2. (a–c) U-2 OS cells were treated with lysozyme-FITC·2 (a), IgG-Alexa 555·2 (b), or EGFP-NLS·2 (c) for 1 h. In the case of EGFP-NLS, PEI Max was also added to the solution containing B₁₂Br₁₂²⁻ and HDB. After washing, the cells were incubated with DMEM containing 10% FBS (DMEM (+)) for 24 h, followed by confocal microscopy. (d) U-2 OS cells, HeLa cells or A549 cells were treated with beta-galactosidase (β-Gal)·2 for 1 h. After washing, the cells were incubated with DMEM (+) for 24 h. β-Gal staining was performed, and the blue staining derived from the product was observed by phase-contrast microscopy.

Conclusion

This study successfully demonstrated cytosolic protein delivery using protein·BPMs formed from anionic dodecaborate clusters, cationic polymers, and proteins, making it a promising strategy for developing new protein delivery methods. Testing with other types of proteins and elucidating the precise mechanism of endosomal escape will further refine the method of cytosolic protein delivery using boron clusters and broaden its field of application.

Data availability

The data supporting this article have been included as part of the ESI.†

Conflicts of interest

The authors declare no conflict of interests.

Acknowledgements

This study was supported in part by JSPS KAKENHI (grant number JP23K06574) and (grant number JP25K13059). We extend our gratitude to Dr Yoshihiro Yoshikawa for providing technical assistance.

References

1. C. J. Braun, M. Witzel, A. Paruzynski, K. Boztug, C. von Kalle, M. Schmidt and C. Klein, Rare Dis., 2014, 2, e947749.
2. S. Stein, M. G. Ott, S. Schultze-Strasser, A. Jauch, B. Burwinkel, A. Kinner, M. Schmidt, A. Krämer, J. Schwäble, H. Glimm, U. Koehl, C. Preiss, C. Ball, H. Martin, G. Göhring, K. Schwarzwaelder, W. K. Hofmann, K. Karakaya, S. Tchatchou, R. Yang, P. Reinecke, K. Kühlcke, B. Schlegelberger, A. J. Thrasher, D. Hoelzer, R. Seger, C. von Kalle and M. Grez, Nat. Med., 2010, 16, 198–204.
3. X. Liu, Z. Zhao, F. Wu, Y. Chen and L. Yin, Adv. Mater., 2022, 34, e2108116.
4. X. Liu, Z. Zhao, W. Li, Y. Li, Q. Yang, N. Liu, Y. Chen and L. Yin, Angew. Chem., Int. Ed., 2023, 62, e202307664.
5. P. Chen, W. Yang, T. Hong, T. Miyazaki, A. Dirisala, K. Kataoka and H. Cabral, Biomaterials, 2022, 288, 121748.
6. Y. Hirai, Y. Kawaguchi, C. Kasahara, H. Hirose and S. Futaki, Mol. Pharm., 2024, 21, 1653–1661.
7. H. H. Wang and A. Tsourkas, Proc. Natl. Acad. Sci. U. S. A., 2019, 116, 22132–22139.
8. I. K. Herrmann, M. J. A. Wood and G. Fuhrmann, Nat. Nanotechnol., 2021, 16, 748–759.
9. Y. Kawaguchi and S. Futaki, Curr. Opin. Chem. Biol., 2024, 81, 102482.
10. J. Cebula, K. Fink, J. Boratyński and T. M. Goszczyński, Coord. Chem. Rev., 2023, 477, 214940.
11. P. Matějíček, J. Zedníky, K. Ušelová, J. Pleštil, J. Fanfrlík, A. Nykänen, J. Ruokolainen, P. Hobza and K. Procházka, Macromolecules, 2009, 42, 4829–4837.
12. V. Ďorďovič, M. Uchman, M. Reza, J. Ruokolainen, A. Zhigunov, O. I. Ivankov and P. Matějíček, RSC Adv., 2016, 6, 9884–9892.
13. V. Ďorďovič, B. Verbraeken, R. Hogenboom, S. Kereïche, P. Matějíček and M. Uchman, Chem. – Asian J., 2018, 13, 838–845.
14. J. Li, O. Janoušková, R. Fernandez-Alvarez, S. Mesíková, Z. Tošner, S. Kereïche, M. Uchman and P. Matějíček, Chem. – Eur. J., 2020, 26, 14283–14289.
15. J. Li, R. Fernandez-Alvarez, Z. Tošner, S. Kereïche, M. Uchman and P. Matějíček, J. Mol. Liq., 2021, 336, 116367.
16. A. Barba-Bon, G. Salluce, I. Lostalé-Seijo, K. I. Assaf, A. Hennig, J. Montenegro and W. M. Nau, Nature, 2022, 603, 637–642.
17. Y. Hirai, Y. Makita, J. Asaoka, Y. Aoyagi, A. Nomoto, H. Okamura and S. Fujiwara, ACS Omega, 2023, 8, 35321–35327.
18. G. Salluce, Y. Folgar-Cameán, A. Barba-Bon, I. Nikšić-Franjić, S. El Anwar, B. Grüner, I. Lostalé-Seijo and J. Montenegro, Angew. Chem., Int. Ed., 2024, 63, e202404286.
19. Y. Chen, A. Barba-Bon, B. Grüner, M. Winterhalter, M. A. Aksoyoglu, S. Pangeni, M. Ashjari, K. Brix, G. Salluce, Y. Folgar-Cameán, J. Montenegro and W. M. Nau, J. Am. Chem. Soc., 2023, 145, 13089–13098.
20. A. Barba-Bon, N. I. Gumerova, E. Tanuhadi, M. Ashjari, Y. Chen, A. Rompel and W. M. Nau, Adv. Mater., 2024, 36, e2309219.
21. K. I. Assaf, M. S. Ural, F. Pan, T. Georgiev, S. Simova, K. Rissanen, D. Gabel and W. M. Nau, Angew. Chem., Int. Ed., 2015, 54, 6852–6856.
22. B. Naskar, O. Diat, V. Nardello-Rataj and P. Bauduin, J. Phys. Chem. C, 2015, 119, 20985–20992.
23. K. I. Assaf and W. M. Nau, Angew. Chem., Int. Ed., 2018, 57, 13968–13981.
24. M. Hohenschutz, J.-F. Dufrêche, O. Diat and P. Bauduin, J. Phys. Chem. Lett., 2023, 14, 3602–3608.
25. P. Dullinger and D. Horinek, J. Am. Chem. Soc., 2023, 145, 24922–24930.
26. A. Barba-Bon, A. E. Haitami, C. Pasquier, I. Nikšić-Franjić, O. Diat, P. Bauduin, S. Cantin and W. M. Nau, Angew. Chem., Int. Ed., 2024, 63, e202412834.
27. I. Chazapi, T. Merhi, C. Pasquier, O. Diat, C. Almunia and P. Bauduin, Angew. Chem., Int. Ed., 2024, 63, e202412588.
28. X. Ma, Z. Zhang, A. Barba-Bon, D. Han, Z. Qi, B. Ge, H. He, F. Huang, W. M. Nau and X. Wang, Proc. Natl. Acad. Sci. U. S. A., 2024, 121, e2407515121.
29. H. E. Davis, J. R. Morgan and M. L. Yarmush, Biophys. Chem., 2002, 97, 159–172.
30. M. J. Mehta, H. J. Kim, S. B. Lim, M. Naito and K. Miyata, Macromol. Biosci., 2024, 24, 2300366.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5bm00346f

---