RESEARCH ARTICLE | CHEMISTRY

CELL BIOLOGY
OPEN ACCESS

A small-­molecule carrier for the intracellular delivery of a

membrane-­impermeable protein with retained bioactivity
Xiqi Maa,1, Zhixiong Zhanga,1, Andrea Barba-­Bonb , Dongxue Hana, Zichun Qia, Baosheng Gea, Hua Hea , Fang Huanga,2 , Werner M. Naua,b,2 ,
and Xiaojuan Wanga,2

Affiliations are included on p. 8.

Edited by David Weitz, Harvard University, Cambridge, MA; received April 15, 2024; accepted September 24, 2024

Intracellular protein delivery has the potential to revolutionize cell-­biological research

and medicinal therapy, with broad applications in bioimaging, disease treatment, and Significance
genome editing. Herein, we demonstrate successful delivery of a functional protein,
cytochrome c (CYC), by using a boron cluster anion as molecular carrier of the super- The delivery of proteins into cells
chaotropic anion type (B12Br11OPr2–). CYC was delivered into lipid bilayer vesicles as has been achieved by
well as living cells, with a cellular uptake ratio approaching 90%. Mechanistic studies formulations based on
showed that CYC was internalized into cells through a permeation pathway directly into liposomes, polymers, or
the cytoplasm, bypassing endosomal entrapment. Upon carrier-­assisted internalization, nanoparticles. One limitation of
CYC retained its bioactivity, as reflected by an induced cell apoptosis rate of 25% at low these nanocarriers is that uptake
dose (1 µM). This study furbishes a direct protein delivery method by a molecular carrier proceeds through endocytosis
with high efficiency, confirming the potential of inorganic cluster ions as protein trans-
with subsequent endocytic
port vehicles with an extensive range of future cell-­biological or biomedical applications.
entrapment. This work
lipid bilayer permeation | protein delivery | cytochrome C | chaotropic effect | boron clusters introduces a simple molecular
additive, a boron cluster anion,
Intracellular protein delivery is a transformative strategy in biological and pharmaceutical which allows direct permeation
research (1, 2). The delivery of intrinsically impermeable proteins into living cells can of cytochrome C as a target
replace poorly expressed, dysfunctional, or missing proteins and regulate key intracellular protein triggering apoptosis. The
metabolic or signaling pathways (3). Accordingly, the method has broad application poten- cluster forms a supramolecular
Downloaded from https://www.pnas.org by 106.78.52.19 on October 26, 2024 from IP address 106.78.52.19.

tial for bioimaging, cell and stem cell engineering, signaling pathway research, disease complex with the protein, which
therapy, and genome editing (4–8). One clearly defined pharmaceutical potential of pro- permeates directly through the
tein delivery is that even “undruggable” targets can be addressed (9) or that solid tumors membrane into the cytosolic site
can be targeted (10, 11). To affect intracellular protein delivery, any carrier approach needs of action. Because of the
to overcome numerous unfavorable intrinsic properties of most proteins, namely, poor supramolecular interaction, the
membrane permeability, short half-life, large size, and susceptibility to degradation (12).
native protein is set free inside
Therefore, new effective intracellular delivery strategies are highly important in cell biology
and medicine. the cytosol, which ensures that
In response to demand, many techniques have been developed to deliver proteins into cytochrome C remains active.
living cells (13). Physical-mechanical or so-called macro- and micromethods (14), such This method circumvents the
as microinjection (15–17), cell squeezing (18, 19), and electroporation (20–22), are the need for preparing encapsulating
most efficient approaches for directly administering proteins to cells or extracellular ves- formulations because the boron
icles, but their application range is restricted due to the low-throughput and disruptive cluster is highly water soluble
characteristics (23) of these methods, which have very limited in vivo potential. Another and nonencapsulating.
commonly used method is to covalently fuse cargo proteins with cell-penetrating peptides
(CPPs) (24, 25). However, in most cases, CPP-fused proteins enter the cell through
endocytosis, during which they are captured in endosomes and unable to reach their target
subcellular organelles to perform their desired biological functions (26). In addition,
endosomally trapped proteins are routinely discharged by exocytosis or degraded by pro-
teinases (27, 28), either resulting in very low cytoplasmic release of the delivered cargo
(1%) (29) or requiring additional design concepts such as endosomal membrane disruption Competing interest statement: A.B.-­B. and W.M.N. are
(30–32). co-­inventors on a related patent (WO 2021/259668 A1).

In recent years, much progress has also been made toward the use of microheterogeneous This article is a PNAS Direct Submission.

and nanocarrier-based delivery approaches (4, 6, 12, 23, 33–38). The corresponding for- Copyright © 2024 the Author(s). Published by PNAS.
This open access article is distributed under Creative
mulations require the encapsulation of the cargo into liposomes (39) or virus-like particles Commons Attribution License 4.0 (CC BY).
(40) or their conjugation to inorganic (41) or polymeric (42, 43) nanoparticles. While 1
X.M. and Z.Z. contributed equally to this work.
these nanocarrier approaches allow tailoring of the formulations to a specific cargo type, 2
To whom correspondence may be addressed. Email:
endosomal entrapment of the assemblies remains a recurring interference (32, 44, 45) [email protected], [email protected], or
[email protected].
associated with the drawbacks of low delivery efficiency, decreased protein activity, or high
This article contains supporting information online at
toxicity. Finally, there are additional synthetically more elaborate approaches that are based https://www.pnas.org/lookup/suppl/doi:10.1073/pnas.​
on dynamic covalent bond formation to facilitate membrane passage, prominently those 2407515121/-­/DCSupplemental.
based on di- and oligosulfide-modified systems (46). Published October 22, 2024.

PNAS 2024 Vol. 121 No. 44 e2407515121 https://doi.org/10.1073/pnas.2407515121 1 of 10

 Molecular carrier approaches present an appealing alternative
to the use of nanoscale formulations because they can pave the
way for directing membrane permeation pathways into the cyto-
sol; this approach is highly desirable for intracellular activity, as it
allows continuous, endocytosis-independent migration into sub-
cellular organelles or the nucleus (47). However, very few examples
of molecular carriers for peptidic cargos are known, and the
reported ones have traditionally involved amphiphilic molecules
such as pyrene butyrate (48–51) or macrocyclic cyclodextrin
(52, 53) or calixarene (54–58) derivatives, which have limited
cargo scope because they are applicable only to cationic peptides
of the cell-penetrating type rather than native proteins (55, 59).
Recently, inorganic boron cluster (IBC) anions of the dodecabo-
rate (B12X122–
 , X = H, Cl, Br, I) (60) or decaborate type (B10X102–  ,
X = H, Cl, Br, I) (61), and cobalt bis(dicarbollide) (62) as well as
polyoxometalates (63) have been introduced as a new class of
molecular carriers, which draw their carrier potential from their
superchaotropic ionic nature rather than from an amphiphilic
structural design. These boron clusters are chemically highly
inert (64–66), poorly metal coordinating, and weakly basic
inorganic anions (67–69) which have already found numerous
applications in different areas (64, 70–75), including therapy
(65, 76–78). Due to the recently recognized superchaotropic
nature of large cluster ions (69, 79–88), they exhibit unexpect-
edly high affinities for both proteins and membranes, which
can be traced back to the chaotropic effect (60, 89). The chao-
tropic effect presents a generic driving force that accounts for
the intrinsic affinity of chaotropic ions for organic interfaces
and surfaces, including those of polymers, peptides, proteins,
and concave macrocyclic binding sites (82, 83, 89, 90). Most
important with respect to delivery applications is their own
ability to interact with and permeate through lipid bilayer mem-
Downloaded from https://www.pnas.org by 106.78.52.19 on October 26, 2024 from IP address 106.78.52.19.

branes (60, 91, 92). We have now applied the boron cluster
approach to the delivery of a functional membrane-impermeable
protein, cytochrome c (CYC, MW ca. 12 kDa). CYC is a pos-
itively charged protein that has evolved as a key protein target
in delivery studies (93–95), owing to its potential in tumor
therapy by inducing apoptosis (96).
Fig. 1.   Intermolecular interactions between IBC-­Pr and CYC. (A) Chemical-­
Results structural representations (orange represents boron-­bromine bonds or
bromine atoms) of IBC-­Pr. (B) Microcalorimetric titration of CYC with IBC-­Pr.
Intermolecular Interactions between IBC-­Pr and CYC. Among Raw ITC data (Top) for sequential injections of IBC-­Pr into the CYC solution
and apparent reaction heats obtained from the integration of the calorimetric
the transport-­active inorganic clusters (60–63), we focused traces (Bottom). IBC-­Pr/CYC concentrations in mM: 5/0.25. The stoichiometric
on the brominated dodecaborates, because they have the best n value for a one-­set-­of-­sites model was found to be 2.0 ± 0.1. 10% error
balance of high activity and low cellular toxicity (61). Moreover, applies for Ka and ±0.5 kcal mol–1 for ΔH, TΔS, and ΔG (quadruple replicates,
on two instruments); fitting according to a sequential binding model with two
we selected B12Br11OCH2CH2CH32– (IBC-­Pr, as sodium salt, different binding sites was equally possible, but because the thermochemical
Fig. 1A) rather than B12Br122–, to demonstrate the potential of using parameters in this fitting afforded similar values for both binding sites (ΔH1 =
monofunctionalized derivatives (97, 98). IBC-­Pr was synthesized –3.9 ± 0.5 kcal mol–1 versus ΔH2 = –3.4 ± 0.5 kcal mol–1 and TΔS1 = 2.5 ± 0.9 kcal
and purified as reported (97, 99, 100) and characterized via 1H mol–1 versus TΔS2 = 2.2 ± 0.5 kcal mol–1), the fitted binding constants had larger
errors (K1 = (8.5 ± 6.4) × 104 M–1; K2 = (1.5 ± 0.3) × 104 M–1). (C) Size distribution of
NMR and 11B NMR spectroscopy as well as FTIR and ESI-­HRMS CYC (1 µM) in the absence and presence of IBC-­Pr (100 µM), measured by DLS;
measurements (SI Appendix, Fig. S1). The intermolecular interaction percentages derived from DLS intensities. (D) Zeta potential measurements of
between IBC-­Pr and CYC was quantified by isothermal titration CYC (1 µM) in the absence and presence of IBC-­Pr (100 µM). (E) Analysis of two
binding regions (dashed circles); note that the heme ferric active site (shown
calorimetry (ITC, Fig. 1B). The thermodynamic parameters in blue) remains accessible.
showed that the binding is an enthalpically driven process, as
expected for chaotropic interactions (83, 89), while the favorable
entropic contribution may reflect an electrostatic component due or Kd = 29 µM, to ensure significant complex formation at high
to desolvation of ionic residues on the protein (101, 102). The micromolar carrier concentrations but sufficiently low to ensure
interaction stoichiometry (n = 2.0 ± 0.1) suggested the formation of cargo release through a dynamic supramolecular equilibrium. In
ternary 1:2 CYC/IBC-­Pr complexes with two binding sites for boron detail, at typical delivery concentrations of 1 to 10 µM CYC and
clusters on the protein surface. Fitting by assuming a model with 100 µM IBC-­Pr, approximately 85% of CYC is expected to be
two identical binding sites afforded excellent agreement (Fig. 1B), complexed at equilibrium, and ca. 15% remains free.
such that the affinities to the two binding sites in CYC appear to The hydrodynamic size of the complex was measured by
be comparable. The resulting (average) binding constant to both dynamic light scattering (DLS). The theoretical three-dimensional
sites was found to be sufficiently high, Ka = (3.4 ± 0.3) × 104 M–1 parameters calculated from their (native) chemical structure are

2 of 10 https://doi.org/10.1073/pnas.2407515121 pnas.org
 ca. 1.0 × 0.7 × 0.7 nm3 for IBC-Pr and 2.6 × 3.2 × 3.3 nm3 for Structural and Functional Integrity of CYC in the Presence of
CYC (103). DLS revealed an average hydrodynamic diameter of IBC-­Pr. The influence of IBC-­Pr binding on the structure and
4.1 ± 0.6 nm for CYC alone and a distinctly larger one of 5.5 ± activity of CYC was investigated by circular dichroism (CD)
0.7 nm for CYC with IBC-Pr (Fig. 1C), consistent with complex spectroscopy and enzymatic assays. As shown in Fig. 2A, there
formation. Zeta potential measurements (Fig. 1D) showed that was only a subtle change in the secondary structure of CYC in
the binding of IBC-Pr reduced the positive surface potential of the presence of IBC-­Pr (an achiral additive), with a deviation of
CYC from 25.9 ± 1.3 mV to 10.5 ± 1.1 mV. Chemical structure 2%. Similarly, neither a shift in λmax of the CYC fluorescence
analysis showed that CYC is a protein with 8 units of expected emission (SI Appendix, Fig. S3) nor changes in the UV absorption
positive surface charge at physiological conditions (pH 7.4) (104). spectra (108) of CYC were observed in the presence of IBC-­Pr
These are expected to be partially neutralized by the binding of (SI Appendix, Fig. S3). These results confirmed the retention of
two doubly negatively charged IBC-Pr cluster anions, consistent the native conformation of the protein at large.
with the experimentally observed decrease in protein surface Regarding the retention of bioactivity, CYC not only is an
charge. apoptosis-initiating protein (see below) but concomitantly pos-
Superchaotropic anions can be characterized as being strongly sesses peroxidase activity, which is typically exerted in the mito-
hydrophilic and lipophilic at the same time, that is, they can be chondrial respiratory chain. This is exploited in enzyme assays, in
very water-soluble and still have a high affinity for binding to which the ability of CYC to catalyze the oxidation of
hydrophobic interfaces or cavities (69, 80, 89). The hydrophobic- 2,2’-azino-bis(3-ethylbenzthiazoline-6-sulfonate) (ABTS) to its
ity and charge distribution on the surface of CYC were therefore green radical cation (ABTS⋅+) (109) can be monitored by UV
analyzed via simulations to identify the likely binding sites. As spectrophotometry at 418 nm. As depicted in Fig. 2B, the kinetic
shown in Fig. 1E, Region 1 contains a hydrophobic cavity near traces of the CYC-catalyzed reactions were virtually superimpos-
the amino acid residues G-56, I-57, T-58, W-59, K-60, and Y-74. able in the absence and presence of IBC-Pr, even if ca. 85% of
The diameter of the cavity is approximately 1.2 nm (distance CYC was complexed under those conditions. The combined data
between Y-74 and W-59). Region 2 is near the amino acid residues suggest that the interaction between IBC-Pr and CYC affects nei-
K-13, F-82, A-83, G-84, I-85, K-86, and K-87 and is highly ther its structure nor its catalytic activity in a pronounced manner,
positively charged. The width of Region 2 is about 0.8 nm (dis- that is, the CYC/IBC-Pr complex retains these functionally critical
tance between I-85 and K-13). Both concave sites are i) sufficiently properties. It also does not cause any aggregation or precipitation,
large, ii) hydrophobic, and iii) positively charged, such that we as both would invariably reduce activity, and neither were these
assign them as preferred binding sites for IBC-Pr in our 1:2 bind- observed by DLS experiments (see above).
ing model (SI Appendix, Fig. S2).
Experimentally, fluorescence spectra of CYC before and after Delivery of CYC with IBC-­Pr Through Artificial Membranes. The
IBC-Pr binding were also investigated. The intrinsic fluorescence advantage of molecular carriers is that their ability to transport
of CYC is mainly derived from tryptophan, which can be excited cargo across the lipid bilayer can be directly quantified by
Downloaded from https://www.pnas.org by 106.78.52.19 on October 26, 2024 from IP address 106.78.52.19.

at 280 nm and affords an emission with maximum at 350 nm established in vitro transport assays with model membranes. This
(105). Interestingly, the binding of IBC-Pr to CYC resulted in a allows one to obtain evidence for carrier-­assisted lipid-­bilayer
distinct enhancement of the microenvironmentally responsive Trp permeation of the cargo as well as to identify the most suitable
fluorescence intensity and, counterintuitively, a slightly decreased working concentrations, information which needs to be empirically
average lifetime (from 2.05 to 1.99 ns, SI Appendix, Fig. S3). This assessed for microheterogeneous carrier formulations. As a rule,
opposing trend translates to an increased radiative decay rate in molecular carriers that exhibit transport of a particular cargo across
the presence of IBC-Pr (106, 107), which in turn is consistent model membranes are showing cellular activity as well, because
with the relocation of at least one Trp residue, likely W-57 in the pertinent transport mechanism (direct permeation through the
Region 1, into a more polarizable environment, that is, near the lipid bilayer) is omnipresent in cellular membranes, in addition
highly polarizable perbrominated boron cluster (61, 69, 89). The to energy-­dependent pathways (endocytosis).
fluorescence response was consistent with the molecular modeling In experimental detail, the ability of IBC-Pr to deliver CYC
analysis. directly through a phospholipid bilayer membrane was investigated

Fig. 2.   Structural and functional integrity of CYC. (A) CD spectra of CYC (1 µM) in the absence and presence of IBC-­Pr (100 µM). In the inset, the calculated percentage
of secondary structure contributions is shown (Methods), suggesting minimal conformational change upon binding of IBC-­Pr. (B) Kinetics of peroxidase activity
of CYC (10 µM) in the absence and presence of IBC-­Pr (100 µM); the visual formation of the green radical cation (ABTS⋅+) after 20 min is shown in the inset. The
presence of serum (10% (v/v) FBS) had no significant effect on the enzyme kinetics.

PNAS 2024 Vol. 121 No. 44 e2407515121 https://doi.org/10.1073/pnas.2407515121 3 of 10

 in large unilamellar vesicles (ca. 120 diameter, SI Appendix, Fig. S4) In the presence of IBC-Pr, delivery of CYC (5 μM) was signaled
by using the carboxyfluorescein (CF) assay (Fig. 3A) (54). We selected by a time-resolved increase in fluorescence of CF, which was
1,2-dimyristoyl-sn-glycero-3-phosphoethanolamine (DMPE)/1,2- expectedly dependent on the IBC-Pr concentration. To quantita-
dipalmitoyl-sn-glycero-3-phospho-(1’-rac-glycerol) (DPPG)/choles- tively characterize the transport efficiency, the normalized fluo-
terol (CHOL, 1/2/1 molar ratio) liposomes (60, 62) (Fig. 3B) as well rescence response was plotted against the concentration of IBC-Pr,
as egg yolk phosphatidylcholine (EYPC) ones (SI Appendix, Fig. S5); and the resulting dose–response curves were fitted according to
the former have a better resemblance to cellular membranes due to the Hill equation to obtain the characteristic transport parameters,
their negative surface charge and presence of cholesterol. In a typical the maximal transmembrane activity (Ymax), and the concentration
time-resolved fluorescence experiment, CF emission was monitored needed to achieve 50% of Ymax (EC50) (56). Since the IBC-Pr
during the sequential addition of IBC-Pr (t = 60 s) and CYC (t = 120 concentration could not be studied above 1.2 mM due to incipient
s, Fig. 3B). The release of all the encapsulated CF by adding the membrane disruption, the fitting results were approximate (Ymax
surfactant Triton X-100 (TX-100, t = 600 s) allowed normalization ca. 35%; EC50 = 770 ± 80 μM, Fig. 3C) but sufficient to conclude
of the fluorescence intensity data. The addition of only one of the that high micromolar concentrations of clusters induce fast (within
components (IBC-Pr or CYC) did not produce any changes in the 5 min) CYC uptake in model membranes. Empirically, for cellular
fluorescence of the system, suggesting that there was no membrane experiments, the concentration of carrier can be adjusted at least
translocation of either cargo alone nor membrane disruption or other 4 times below the vesicular EC50 because longer incubation times
forms of dye efflux; moreover, the integrity of the vesicles in the are used (on an hourly time scale) (60).
presence of cargo and IBC-Pr (up to 1.2 mM) was confirmed by Strictly speaking, the CF assay does not provide direct evidence
DLS experiments (SI Appendix, Fig. S4). for CYC permeation, but signals it indirectly, by dye efflux. To
Downloaded from https://www.pnas.org by 106.78.52.19 on October 26, 2024 from IP address 106.78.52.19.

Fig. 3.   Delivery of CYC through artificial membranes. (A) Schematic representation of the CF assay. (B) Changes in CF emission (λex = 492 nm, λem = 517 nm) in CF-­
encapsulated DMPE/DPPG/CHOL(1/2/1) liposomes as a function of time during the addition of different concentrations of IBC-­Pr (0 to 1,200 μM) at t = 60 s, CYC (5
μM) at t = 120 s, and TX-­100 at t = 600 s, for calibration, and (C) fractional activity for IBC-­Pr with corresponding Hill curve fit; note that the maximal concentration
of IBC-­Pr was limited by its membrane-­lytic effect above ca. 1.2 mM. (D) Fluorescence microscopy images of EYPC giant vesicles after the addition of CYC-­FITC
(10 μM) and IBC-­Pr (50 μM) recorded after different incubation times (0 to 8 min); note the efficient boron cluster-­induced uptake of CYC-­FITC into the giant vesicle
and the subsequent quenching of extravesicular fluorescence by addition of acid (pH 5.0). See SI Appendix, Fig. S6, for controls with brightfield micrographs.

4 of 10 https://doi.org/10.1073/pnas.2407515121 pnas.org
 directly visualize the lipid bilayer permeation of the protein, we (65, 76–78). Accordingly, micromolar IBC-­Pr concentrations
also used larger, so-called giant vesicles of 50±25 μM diameter, should be sufficiently biocompatible to conduct intracellular
which can be imaged by conventional fluorescence microscopy. protein delivery experiments (112).
For this purpose, CYC was labeled with fluorescein isothiocyanate For fluorescence staining, the labeled protein (CYC-FITC) was
(FITC), by-passing the need for other vesicular additives. Indeed, used. Living HeLa cells were incubated for 3 h with CYC-FITC
the experiments provided evidence that the fluorescently labeled alone or in an in situ mixture of CYC-FITC/IBC-Pr, subsequently
protein was membrane-impermeable in the absence of carrier, but fixed, and scanned by confocal laser-scanning microscopy. We
permeated effectively, within less than 10 min, in the presence of maintained the protein concentration at 1 μM (well below its IC50
IBC-Pr, resulting in an almost equilibrated fluorescence intensity value for inducing long-term apoptosis, see below) and varied the
in bulk solution and inside the vesicles (Fig. 3D, SI Appendix, concentrations of IBC-Pr from 25 to 100 μM, at the onset of
Fig. S6); subsequent addition of acid (HCl, pH 5.0) resulted in a transport activity in the model membranes (Fig. 3C), but well
selective quenching of the extravesicular phase, because the FITC below its membrane-disrupting and cell-toxic concentrations. As
chromophore is pH sensitive (110). shown in Fig. 4A, CYC-FITC alone was unable to enter HeLa
cells, and only very weak green fluorescence was detected in the
Delivery of CYC with IBC-­Pr through Cellular Membranes. To test cells after 3 h of incubation. In contrast, in the presence of IBC-Pr,
the transferability from vesicle models to biological membranes, the fluorescence signal of CYC-FITC became readily observable
intracellular CYC delivery with IBC-­Pr as molecular carrier was and increased with carrier concentration (Fig. 4B). These results
studied in HeLa cells. First, to exclude any significant cytotoxicity are consistent with the in vitro findings with CYC, verifying the
of IBC-­Pr in the concentration range required for transport, successful intracellular delivery of CYC-FITC by IBC-Pr. Enlarged
MTT assays were carried out. As shown in SI Appendix, Fig. S7, images of HeLa cells incubated with a mixture of CYC-FITC and
the physiological state of the cells was not affected when the IBC-Pr further revealed that the internalized CYC-FITC was
IBC-­Pr concentration was below 400 μM. Only at millimolar located evenly within the cells, also in the cytoplasm and nucleus
concentrations was a significant drop in HeLa cell viability (Fig. 4C). A similar cellular distribution was observed when we
observed (to 85%), consistent with the observation of membrane conducted experiments with live instead of fixed HeLa cells
disruption in the vesicle experiments (above 1.2 mM). This is in (SI Appendix, Fig. S8). We also transferred the method to another
line with the generally good biocompatibility of boron clusters cell line, A549, and obtained comparable staining results
(60, 70, 111), which have already found their way into clinical (SI Appendix, Fig. S9). Additionally, the protein was labeled with
applications, e.g., for boron neutron capture therapy (BNCT) carboxytetramethylrhodamine (TAMRA) as an alternative, less
Downloaded from https://www.pnas.org by 106.78.52.19 on October 26, 2024 from IP address 106.78.52.19.

Fig. 4.   Delivery of CYC through cellular membranes. (A) Confocal images of HeLa cells after incubation with CYC-­FITC (1 μM) and a) without or with b) 25 μM, c)
50 μM, or d) 100 μM IBC-­Pr for 3 h; scale bar represents 50 μm. Representative images of three biological replicates. (B) Mean intensity of FITC fluorescence from
the micrographs in panel (A). (C) Enlarged confocal image. (D) Flow cytometry and (E) quantitative analysis of the cytometry results for HeLa cells after incubation
with 1 μM CYC-­FITC and 100 μM IBC-­Pr. Error bars indicate SD (n = 3).

PNAS 2024 Vol. 121 No. 44 e2407515121 https://doi.org/10.1073/pnas.2407515121 5 of 10

 pH-sensitive chromophore (CYC-TAMRA), which afforded also FITC fluorescence stems from inactive proteins that may be dena-
the desirable uptake into the cytosol, albeit with a lower nuclear tured, misfolded, or unfolded when shuttling through the cellular
staining propensity (SI Appendix, Fig. S10). Important to note, membrane. In addition, hypothetically, FITC can be already
while IBC-Pr was able to transport the positively charged CYC cleaved off (hydrolytically or enzymatically) as free dye or peptide
(pI = 9.6) (113), the negatively charged green fluorescent protein fragment from the protein, falsely implying the uptake and pres-
(GFP, pI = 6.2) (114) was not transported into cells (SI Appendix, ence of intact CYC. Consequently, its biological activity needed
Fig. S11). Similarly, the negatively charged bovine serum albumin to be validated (116).
(BSA, pI = 5.5) (115) had been previously found to be inactive
in liposomal transport assays (60). This suggests that the negatively CYC Retains Its Intracellular Bioactivity After Delivery by IBC-­
charged boron clusters, in the absence of other additives, display Pr. CYC is a vital mediator of apoptosis that typically triggers the
a selectivity for transporting positively charged proteins. caspase activation cascade in the cytoplasm only after intracellular
Flow cytometry was also used to quantitatively evaluate the mitochondrial release (95, 117). Determining whether and to
intracellular delivery efficiency. As shown in Fig. 4D, after free what degree any extracellularly internalized CYC is capable of
CYC-FITC was incubated with HeLa cells, the cellular uptake inducing the apoptosis of the target cells is the most critical
ratio was almost negligible (1.3%). When the mixture of measure of successful protein delivery for this particular target.
CYC-FITC with IBC-Pr was applied, the cellular uptake ratio Indeed, CYC (unlabeled, 1 µM) delivered by IBC-­Pr had dose-­
increased to 88%. Quantitative analysis (Fig. 4E and SI Appendix, dependent cytotoxic effects on HeLa cells, with an IC50 value of
Table S1) showed that the average fluorescence intensity of live 7.5 μM for triggering apoptosis after 48 h (Fig. 5A). In contrast,
HeLa cells treated with the CYC-FITC/IBC-Pr mixture was at free CYC displayed no cytotoxicity, even when the concentration
least five times greater than that of cells incubated with was increased from 1 to 16 μM, as expected for a membrane-­
CYC-FITC alone, again signaling highly effective protein deliv- impermeable protein. The results demonstrated that IBC-­Pr is not
ery. CYC transport by the boron cluster was found to be com- only capable of mediating effective cell internalization of CYC but
patible with the presence of serum in the incubation media, also maintains the bioactivity of the protein throughout the course
which showed a similar uptake efficiency (SI Appendix, Table S1). of transmembrane delivery. Therefore, IBC-­Pr acts as a molecular
As a control, we also tested LipofectamineTM 3000, a commer- membrane carrier for an active and fully functional protein. The
cially available nanocarrier frequently used for nucleic acid trans- retained activity is due to the reversibility of the supramolecular
fection, but no cellular uptake of the protein was found equilibria (see ITC results above); accordingly, while the binding
(SI Appendix, Fig. S12). This demonstrates that boron clusters of IBC-­Pr is required to shuttle CYC through the membrane, it is
serve as complementary carriers with a distinct cargo scope. sufficiently dynamic to allow its release inside the cytosol. Upon
Although labeling with FITC is a standard method for follow- release, the apoptotic signaling cascade is initiated through tight,
ing the fate and action of CYC by fluorescence (93, 94, 101), the salt-­bridged binding of CYC by apoptotic protease activating
omnipresent question remains whether all or part of the observed factor 1 (Apaf-­1) (118).
Downloaded from https://www.pnas.org by 106.78.52.19 on October 26, 2024 from IP address 106.78.52.19.

Fig. 5.   Cellular bioactivity of CYC after intracellular delivery by IBC-­Pr. (A) Evaluation of the viability of HeLa cells treated with CYC (0 to 16 μM) with 100 μM or
without IBC-­Pr after 48 h of incubation. Error bars indicate SD (n = 3). (B) Confocal laser scanning microscopy images of HeLa cells after incubation with CYC
(1 μM) in the absence or presence (100 μM) of IBC-­Pr after 24 h of incubation followed by subsequent staining with Annexin V-­FITC/PI; scale bar represents
50 μm. Representative images of three biological replicates are shown. (C) Flow cytometric apoptosis assay of HeLa cells after 24 h of treatment with CYC (1 μM),
IBC-­Pr (100 μM), or both. Q1: necrotic cells, Q2: late apoptotic cells, Q3: early apoptotic cells, and Q4: living cells. The numbers in the Q2 and Q3 areas indicate
the percentage of late and early apoptotic cells, respectively.

6 of 10 https://doi.org/10.1073/pnas.2407515121 pnas.org
 The Annexin V-FITC/PI apoptosis kit was employed to further permeation through the lipid bilayer vesicles applies, and this was
characterize the activity of internalized CYC. As shown in Fig. 5B, visually corroborated by observing the boron-­cluster-­mediated
weak green fluorescence owing to Annexin V-FITC was observed uptake of the fluorescently labeled protein (CYC-­FITC) into giant
in HeLa cells treated with CYC (1 μM) in the absence of carrier. vesicles (Fig. 3D). To determine the mechanism underlying the
In contrast, in the presence of IBC-Pr (100 μM), the apoptotic cells cellular uptake of the CYC/IBC-­Pr system, confocal microscopic
exhibited blebbing of the Annexin V-FITC-stained cellular mem- imaging (Fig. 6 A and B) and flow cytometry (Fig. 6C) were
brane and for the late apoptotic cells staining by PI was observed. performed to evaluate the intracellular delivery efficiency under
The apoptotic activities of the different CYC samples were quanti- different pretreatment conditions. Internalization of CYC-­FITC
fied according to the sum of the fraction of early apoptotic cells was monitored by its relative intracellular fluorescence intensity. At 4
(Q3) to late apoptotic cells after 24 h (Q2, Fig. 5C). In the absence °C, the fluorescence intensity was only slightly reduced, which ruled
of IBC-Pr, no obvious signs of apoptosis induction were observed out a predominant energy-­dependent uptake mechanism (121) as
while in the presence of IBC-Pr, early apoptosis amounted to 19.9% a predominant pathway. When cells were pretreated with NaN3,
and late apoptosis to 5.1%, summing up to 25.0%. This apoptotic which causes ATP depletion, the relative mean fluorescence intensity
effect is comparable to or greater than that observed with micro- decreased only slightly, which also suggested a dominant energy-­
heterogeneous formulations developed for CYC delivery, e.g., those independent internalization mechanism (122). Finally, two clathrin-­
using polyphenol-based (93), calcium carbonate-mineralized (94), dependent endocytic inhibitors, chlorpromazine and sucrose, as well
or mesoporous silica (95) nanoparticles, at comparable loads of as an actin-­dependent endocytic inhibitor, cytochalasin D, were
CYC (12 µg mL–1 versus 5 to 50 µg mL–1) as well as incubation used to pretreat the cells. The experimental results showed that
times (24 h versus 6 to 48 h) (93–95). IBC-Pr, as a highly the uptake of CYC-­FITC delivered by IBC-­Pr was not strongly
water-soluble supramolecular additive, bypasses the need for pre- affected by these conventional inhibitors, and a high delivery
vious loading with CYC and can be homogeneously administered. capacity (70% or greater) remained. These findings, along with
We accordingly classify protein delivery via molecular boron carriers the homogeneous distribution (fluorescence) of CYC-­FITC within
as highly efficient and practically advantageous. the cells, corroborate an “energy-­independent” uptake mechanism,
namely, lipid bilayer permeation, as a dominant pathway. This direct
Membrane Translocation of CYC Occurs by Molecular Carrier-­ permeation pathway is the same that is operative in the vesicle
assisted Direct Permeation. Membrane translocation can experiments, where energy-­dependent pathways are not involved.
occur through energy-­dependent uptake pathways (prominently The major circumvention of endocytic energy-­dependent uptake
endocytosis), while the main energy-­independent internalization routes (which are competitive but functionally unproductive)
pathway is direct permeation (119, 120). In principle, the formation markedly improves the actual protein delivery efficiency into the
of pores caused by a carrier presents another energy-­independent cytosol; this accounts for the excellent bioactivity of CYC when it
pathway, but as no leakage is observed in the CF assays with large is delivered with the boron cluster. The molecular cluster carrier
unilamellar vesicles (up to 1.2 mM, SI Appendix, Figs. S4 and S5), therefore stands out in terms of qualitative (no formulation, no
Downloaded from https://www.pnas.org by 106.78.52.19 on October 26, 2024 from IP address 106.78.52.19.

we do not consider this pathway for IBC-­Pr. Instead, direct loading, direct cytosolic transport) performance indicators.

Fig. 6.   Membrane translocation by molecular carrier-­assisted direct permeation. Cellular uptake efficiency of CYC (1 μM) delivered by IBC-­Pr (100 μM) in the
presence of different endocytosis inhibitors. (A) Confocal microscopy images, (B) relative mean fluorescence intensity per cell calculated from confocal microscopy
images, and (C) quantitative intracellular fluorescence measured by flow cytometry (the threshold line was used to distinguish the FITC intensity of cells from
the autofluorescence of blank cells). Representative images of three biological replicates are shown. Error bars indicate SD (n = 3).

PNAS 2024 Vol. 121 No. 44 e2407515121 https://doi.org/10.1073/pnas.2407515121 7 of 10

 Discussion method can be conveniently used in cell biology for studying
protein uptake and subsequent activity in the cytosol.
Perhalogenated boron cluster anions have recently been shown
to function as broadband molecular carriers (60, 61). The
uptake mechanism has been related to the superchaotropic Methods
nature of these inorganic anions and the chaotropic effect, Na2B12Br11OCH2CH2CH3 (IBC-­Pr) was synthesized as previously reported (97, 99).
which enables dynamically reversible, supramolecular, and Circular dichroism (CD) spectra were recorded on a MOS-­450/AF-­CD instrument
simultaneous interactions of the cluster anions with both the (France) and the ratio of secondary structures was calculated by software K2D2 (123).
cargo and the lipid bilayer membrane (60, 89). By using a The enzyme-­catalytic activity of CYC was measured through the conversion of 2,2
propyloxy derivative as a molecular carrier and cytochrome c ’-­azino-­bis(3-­ethylbenzthiazoline-­6-­sulfonate) (ABTS) to its radical cation, ABTS⋅+
(CYC) as cargo, we have shown that the method can be used (124). Phospholipid concentrations of the prepared vesicles were determined by the
to transport a real protein target, which broadens the applica- Stewart assay (125). Transport experiments in vesicles were performed with the car-
tion scope considerably, in the direction of functional cell boxyfluorescein (CF) assay. Giant vesicles were formed by electroformation (56, 126,
biology and, potentially, therapy. Most importantly, the struc- 127). Labeling of CYC with FITC and TAMRA was performed by standard protocols.
tural integrity as well as biological activity of the protein are The cells for the uptake experiments, flow cytometry, and the cytotoxicity assays were
retained both in vitro and in live cells, affording pronounced cultured in Dulbecco´s modified eagle´s medium (DMEM) supplemented with 10%
biological effects (apoptosis) at low cargo and carrier loads and (v/v) fetal bovine serum (FBS) in a humidified 5% CO2 atmosphere inside an incu-
implying highly efficient uptake. The uptake mechanism pro- bator at 37 °C. The cells used for all the experiments were in the logarithmic phases
ceeds through the preferred route of direct bilayer permeation, of growth. The Annexin V-­FITC/PI kit was used to evaluate the ability of the CYC/
which largely bypasses complications due to endocytic uptake. IBC-­Pr mixture to induce apoptosis. Additional details are provided in SI Appendix.
The use of superchaotropic molecular carriers, namely, boron
Data, Materials, and Software Availability. All study data are included in the
cluster anions, offers numerous advantages because no micro-
article and/or SI Appendix.
heterogeneous formulations are required, cargo and carrier can
be added in situ, sequentially or together, and their relative ACKNOWLEDGMENTS. This study was supported by the National Natural Science
concentrations can be adjusted at will and, beforehand, opti- Foundation of China (Grant no. 42061134020, 22177133, 22278438) and the
mized for vesicles as model membranes. Moreover, the small Deutsche Forschungsgemeinschaft (Grant no. NA 686/14 and NA 686/15). We
size, low molecular weight, and high water solubility of the thank Jinling Zhang for help with the synthesis of the boron cluster and Mathias
clusters are attractive from an administration and biocompat- Winterhalter for assistance with the giant vesicle experiments.
ibility point of view. This method, which is highly complemen-
tary to the use of nanocarriers, should be transferable to other
proteins, prominently positively charged ones, which have been Author affiliations: aCollege of Chemical Engineering, China University of Petroleum (East
notoriously difficult to target. Also conceivable, monofunc- China), Qingdao 266580, China; and bSchool of Science, Constructor University, Bremen
28759, Germany
tionalization of perbrominated boron clusters (97, 98) with
Downloaded from https://www.pnas.org by 106.78.52.19 on October 26, 2024 from IP address 106.78.52.19.

Author contributions: F.H., W.M.N., and X.W. designed research; X.M., Z.Z., A.B.-­B., D.H.,
substituents other than propyloxy offers perspectives for a Z.Q., B.G., and H.H. performed research; X.M., Z.Z., A.B.-­B., D.H., and Z.Q. analyzed data;
direct attachment to the deliverable cargo or to targeting groups A.B.-­B. reviewed and edited the manuscript; B.G. funded the study, reviewed and edited
the manuscript; H.H. supervised the study, reviewed and edited the manuscript; F.H.,
to achieve selective cellular uptake. While therapeutic applica- W.M.N., and X.W. funded and supervised the study, reviewed and edited the manuscript;
tions may appear more far-fetched, the recently introduced and X.M. and Z.Z. wrote the paper.

1. Y. Zhang, J. J. Roise, K. Lee, J. Li, N. Murthy, Recent developments in intracellular protein delivery. 18. G. L. Szeto et al., Microfluidic squeezing for intracellular antigen loading in polyclonal B-­cells as
Curr. Opin. Biotechnol. 52, 25–31 (2018). cellular vaccines. Sci. Rep. 5, 10276 (2015).
2. B. Leader, Q. J. Baca, D. E. Golan, Protein therapeutics: A summary and pharmacological 19. A. Sharei et al., A vector-­free microfluidic platform for intracellular delivery. Proc. Natl. Acad. Sci.
classification. Nat. Rev. Drug Discovery 7, 21–39 (2008). U.S.A. 110, 2082–2087 (2013).
3. S. Du, S. S. Liew, L. Li, S. Q. Yao, Bypassing endocytosis: Direct cytosolic delivery of proteins. J. Am. 20. N. Pathak et al., Cellular delivery of large functional proteins and protein-­nucleic acid constructs via
Chem. Soc. 140, 15986–15996 (2018). localized electroporation. Nano Lett. 23, 3653–3660 (2023).
4. Y. W. Lee et al., Protein delivery into the cell cytosol using non-­viral nanocarriers. Theranostics 9, 21. C. A. Patino et al., Multiplexed high-­throughput localized electroporation workflow with deep
3280–3292 (2019). learning based analysis for cell engineering. Sci. Adv. 8, eabn7637 (2022).
5. W. Zhou, H. Cui, L. Ying, X. F. Yu, Enhanced cytosolic delivery and release of CRISPR/Cas9 by black 22. S.-­O. Choi et al., Intracellular protein delivery and gene transfection by electroporation using a
phosphorus nanosheets for genome editing. Angew. Chem. Int. Ed. 130, 10268–10272 (2018). microneedle electrode array. Small 8, 1081–1091 (2012).
6. R. Mout et al., General strategy for direct cytosolic protein delivery via protein-­nanoparticle co-­ 23. A. Fu, R. Tang, J. Hardie, M. E. Farkas, V. M. Rotello, Promises and pitfalls of intracellular delivery of
engineering. ACS Nano 11, 6416–6421 (2017). proteins. Bioconjugate Chem. 25, 1602–1608 (2014).
7. A. Dinca, W. M. Chien, M. T. Chin, Intracellular delivery of proteins with cell-­penetrating peptides for 24. A. F. L. Schneider, M. Kithil, M. C. Cardoso, M. Lehmann, C. P. R. Hackenberger, Cellular uptake of
therapeutic uses in human disease. Int. J. Mol. Sci. 17, 263 (2016). large biomolecules enabled by cell-­surface-­reactive cell-­penetrating peptide additives. Nat. Chem.
8. J. A. Zuris et al., Cationic lipid-­mediated delivery of proteins enables efficient protein-­based 13, 530–539 (2021).
genome editing in vitro and in vivo. Nat. Biotechnol. 33, 73–80 (2015). 25. P. Säälik et al., Protein cargo delivery properties of cell-­penetrating peptides. A comparative study.
9. T. A. Slastnikova, A. V. Ulasov, A. A. Rosenkranz, A. S. Sobolev, Targeted intracellular delivery of Bioconjugate Chem. 15, 1246–1253 (2004).
antibodies: The state of the art. Front. Pharmacol. 9, 1208 (2018). 26. R. Brock, The uptake of arginine-­rich cell-­penetrating peptides: Putting the puzzle together.
10. M. P. Deonarain, Q. Xue, Tackling solid tumour therapy with small-­format drug conjugates. Antib. Bioconjugate Chem. 25, 863–868 (2014).
Ther. 3, 237–245 (2020). 27. B. Zhitomirsky, Y. G. Assaraf, Lysosomal accumulation of anticancer drugs triggers lysosomal
11. O. Dagher, T. R. King, N. Wellhausen, A. D. Posey Jr., Combination therapy for solid tumors: Taking a exocytosis. Oncotarget 8, 45117–45132 (2017).
classic CAR on new adventures. Cancer Cell 38, 621–623 (2020). 28. M. P. Stewart et al., In vitro and ex vivo strategies for intracellular delivery. Nature 538, 183–192
12. Z. Gu, A. Biswas, M. Zhao, Y. Tang, Tailoring nanocarriers for intracellular protein delivery. Chem. Soc. (2016).
Rev. 40, 3638–3655 (2011). 29. R. Goswami, T. Jeon, H. Nagaraj, S. Zhai, V. M. Rotello, Accessing intracellular targets through
13. V. J. Bruce, B. R. McNaughton, Inside job: Methods for delivering proteins to the interior of nanocarrier-­mediated cytosolic protein delivery. Trends Pharmacol. Sci. 41, 743–754 (2020).
mammalian cells. Cell Chem. Biol. 24, 924–934 (2017). 30. S. Yu et al., Efficient intracellular delivery of proteins by a multifunctional chimaeric peptide in vitro
14. D. Morshedi Rad et al., A comprehensive review on intracellular delivery. Adv. Mater 33, 2005363 (2021). and in vivo. Nat. Commun. 12, 5131 (2021).
15. J. Kreitz et al., Programmable protein delivery with a bacterial contractile injection system. Nature 31. N. Tamemoto et al., Rational design principles of attenuated cationic lytic peptides for intracellular
616, 357–364 (2023). delivery of biomacromolecules. Mol. Pharm. 17, 2175–2185 (2020).
16. F. Jiang et al., N-­terminal signal peptides facilitate the engineering of PVC complex as a potent 32. M. Ray, Y. W. Lee, F. Scaletti, R. Yu, V. M. Rotello, Intracellular delivery of proteins by nanocarriers.
protein delivery system. Sci. Adv. 8, eabm2343 (2022). Nanomedicine (Lond) 12, 941–952 (2017).
17. I. Vlisidou et al., The Photorhabdus asymbiotica virulence cassettes deliver protein effectors directly 33. E. Soprano, E. Polo, B. Pelaz, P. del Pino, Biomimetic cell-­derived nanocarriers in cancer research. J.
into target eukaryotic cells. eLife 8, e46259 (2019). Nanobiotechnol. 20, 538 (2022).

8 of 10 https://doi.org/10.1073/pnas.2407515121 pnas.org
 34. S. Le Saux et al., Interest of extracellular vesicles in regards to lipid nanoparticle based systems for 76. S. O. Oloo, K. M. Smith, M. d. G. H. Vicente, Multi-­functional boron-­delivery agents for boron
intracellular protein delivery. Adv. Drug Deliv. Rev. 176, 113837 (2021). neutron capture therapy of cancers. Cancers 15, 3277 (2023).
35. M. Sakono, R. Hayakawa, Repressor-­like on-­off regulation of protein expression by the DNA-­binding 77. J. Li et al., Designed boron-­rich polymeric nanoparticles based on nano-­ion pairing for boron
transcription activator-­like effector in T7 promoter-­based cell-­free protein synthesis. ChemBioChem delivery. Chem. Eur. J. 26, 14283–14289 (2020).
22, 888–893 (2021). 78. A. H. Soloway et al., The chemistry of neutron capture therapy. Chem. Rev. 98, 2389–2390 (1998).
36. S. Le Saux et al., Nanotechnologies for intracellular protein delivery: Recent progress in inorganic 79. P. Dullinger, D. Horinek, Solvation of nanoions in aqueous solutions. J. Am. Chem. Soc. 145,
and organic nanocarriers. Adv. Ther. 4, 2100009 (2021). 24922–24930 (2023).
37. C. Chau, P. Actis, E. Hewitt, Methods for protein delivery into cells: From current approaches to future 80. K. I. Assaf, W. M. Nau, Large anion binding in water. Org. Biomol. Chem. 21, 6636–6651 (2023).
perspectives. Biochem. Soc. Trans. 48, 357–365 (2020). 81. Y. Hirai et al., Boron clusters alter the membrane permeability of dicationic fluorescent DNA-­
38. A. Hubbell Jeffrey, A. Chilkoti, Nanomaterials for drug delivery. Science 337, 303–305 (2012). staining dyes. ACS Omega 8, 35321–35327 (2023).
39. Y. B. Kim, K. T. Zhao, D. B. Thompson, D. R. Liu, An anionic human protein mediates cationic liposome 82. W. Wei, Hofmeister effects shine in nanoscience. Adv. Sci. 10, e2302057 (2023).
delivery of genome editing proteins into mammalian cells. Nat. Commun. 10, 2905 (2019). 83. S. Yao et al., Hofmeister effect in the Keggin-­type polyoxotungstate series. Inorg. Chem. Front. 8,
40. S. Banskota et al., Engineered virus-­like particles for efficient in vivo delivery of therapeutic proteins. 12–25 (2021).
Cell 185, 250–265 (2022). 84. M. Hohenschutz, I. Grillo, O. Diat, P. Bauduin, How nano-­ions act like ionic surfactants. Angew.
41. F. Scaletti et al., Protein delivery into cells using inorganic nanoparticle-­protein supramolecular Chem. Int. Ed. 59, 8084–8088 (2020).
assemblies. Chem. Soc. Rev. 47, 3421–3432 (2018). 85. K. I. Assaf et al., High-­affinity binding of metallacarborane cobalt bis (dicarbollide) anions to
42. J. Lv, Q. Fan, H. Wang, Y. Cheng, Polymers for cytosolic protein delivery. Biomaterials 218, 119358 cyclodextrins and application to membrane translocation. J. Org. Chem. 84, 11790–11798 (2019).
(2019). 86. T. Buchecker et al., Polyoxometalates in the hofmeister series. Chem. Commun. 54, 1833–1836
43. M. Yan et al., A novel intracellular protein delivery platform based on single-­protein nanocapsules. (2018).
Nat. Nanotechnol. 5, 48–53 (2010). 87. R. Fernandez-­Alvarez, V. Dordovic, M. Uchman, P. Matejicek, Amphiphiles without head-­and-­tail
44. Y. Sun et al., Phase-­separating peptides for direct cytosolic delivery and redox-­activated release of design: Nanostructures based on the self-­assembly of anionic boron cluster compounds. Langmuir
macromolecular therapeutics. Nat. Chem. 14, 274–283 (2022). 34, 3541–3554 (2017).
45. S. A. Smith, L. I. Selby, A. P. R. Johnston, G. K. Such, The endosomal escape of nanoparticles: Toward 88. B. Naskar, O. Diat, V. R. Nardello-­Rataj, P. Bauduin, Nanometer-­size polyoxometalate anions adsorb
more efficient cellular delivery. Bioconjugate Chem. 30, 263–272 (2019). strongly on neutral soft surfaces. J. Phys. Chem. C 119, 20985–20992 (2015).
46. J. López-­Andarias et al., Cell-­penetrating streptavidin: A general tool for bifunctional delivery 89. K. I. Assaf, W. M. Nau, The chaotropic effect as an assembly motif in chemistry. Angew. Chem. Int. Ed.
with spatiotemporal control, mediated by transport systems such as adaptive benzopolysulfane 57, 13968–13981 (2018).
networks. J. Am. Chem. Soc. 142, 4784–4792 (2020). 90. M. Hohenschutz, P. Bauduin, C. G. Lopez, B. Förster, W. Richtering, Superchaotropic nano-­ion
47. C. Liu et al., Natural polyphenols augment cytosolic protein delivery by a functional polymer. Chem. binding as a gelation motif in cellulose ether solutions. Angew. Chem. Int. Ed. 62, e202210208
Mater. 31, 1956–1965 (2019). (2023).
48. D. Argudo, N. P. Bethel, F. V. Marcoline, M. Grabe, Continuum descriptions of membranes and 91. D. Awad et al., Halogenated dodecaborate clusters as agents to trigger release of liposomal
their interaction with proteins: Towards chemically accurate models. Biochim. Biophys. Acta 1858, contents. Chempluschem 80, 656–664 (2015).
1619–1634 (2016). 92. P. Fan, S. Stolte, D. Gabel, Interaction of organic compounds and boron clusters with new
49. S. Katayama et al., Effects of pyrenebutyrate on the translocation of arginine-­rich cell-­penetrating silica matrices containing the phosphatidylcholine headgroup. Anal. Methods 6, 3045–3055
peptides through artificial membranes: Recruiting peptides to the membranes, dissipating liquid-­ (2014).
ordered phases, and inducing curvature. Biochim. Biophys. Acta 1828, 2134–2142 (2013). 93. Y. Han et al., Polyphenol-­based nanoparticles for intracellular protein delivery via competing
50. T. Takeuchi et al., Direct and rapid cytosolic delivery using cell-­penetrating peptides mediated by supramolecular interactions. ACS Nano 14, 12972–12981 (2020).
pyrenebutyrate. ACS Chem. Biol. 1, 299–303 (2006). 94. A. N. Koo et al., Calcium carbonate mineralized nanoparticles as an intracellular transporter of
51. M. Nishihara et al., Arginine magic with new counterions up the sleeve. Org. Biomol. Chem. 3, cytochrome c for cancer therapy. Chem.-­Asian J. 10, 2380–2387 (2015).
1659–1669 (2005). 95. E. Choi, D. K. Lim, S. Kim, Hydrolytic surface erosion of mesoporous silica nanoparticles for efficient
52. I. Pflueger et al., Cyclodextrin-­based facial amphiphiles: Assessing the impact of the hydrophilic– intracellular delivery of cytochrome c. J. Colloid Interface Sci. 560, 416–425 (2020).
lipophilic balance in the self-­assembly, DNA complexation and gene delivery capabilities. Org. 96. S. K. Kim, M. B. Foote, L. Huang, The targeted intracellular delivery of cytochrome C protein to
Biomol. Chem. 14, 10037–10049 (2016). tumors using lipid-­apolipoprotein nanoparticles. Biomaterials 33, 3959–3966 (2012).
53. C. O. Mellet, J. M. G. Fernández, J. M. Benito, Cyclodextrin-­based gene delivery systems. Chem. Soc. 97. J. Zhang, D. Gabel, K. I. Assaf, W. M. Nau, A fluorescein-­substituted perbrominated dodecaborate
Rev. 40, 1586–1608 (2011). cluster as an anchor dye for large macrocyclic hosts and its application in indicator displacement
Downloaded from https://www.pnas.org by 106.78.52.19 on October 26, 2024 from IP address 106.78.52.19.

54. Y. C. Pan et al., An amphiphilic sulfonatocalix[5]arene as an activator for membrane transport of assays. Org. Lett. 24, 9184–9188 (2022).
lysine-­rich peptides and proteins. Angew. Chem. Int. Ed. 60, 1875–1882 (2021). 98. K. Fink, K. Kobak, M. Kasztura, J. Boratyński, T. M. Goszczyński, Synthesis and biological activity of
55. D.-­Y. Zhang et al., Structurally screening calixarenes as peptide transport activators. Chem. Commun. thymosin β4-­anionic boron cluster conjugates. Bioconjugate Chem. 29, 3509–3515 (2018).
57, 12627–12630 (2021). 99. C. Jenne, C. Kirsch, Alkoxy substituted halogenated closo-­dodecaborates as anions for ionic liquids.
56. A. Barba-­Bon et al., Fluorescence monitoring of peptide transport pathways into large and giant Dalton Trans. 44, 13119–13124 (2015).
vesicles by supramolecular host-­dye reporter pairs. J. Am. Chem. Soc. 141, 20137–20145 (2019). 100. V. Geis, K. Guttsche, C. Knapp, H. Scherer, R. Uzun, Synthesis and characterization of
57. S. Peng et al., Phosphorylation-­responsive membrane transport of peptides. Angew. Chem. Int. Ed. synthetically useful salts of the weakly-­coordinating dianion [B12Cl12]2-­. Dalton Trans. 15,
56, 15742–15745 (2017). 2687–2694 (2009).
58. F. Perret et al., Anionic fullerenes, calixarenes, coronenes, and pyrenes as activators of oligo/ 101. M. V. Kuperman et al., Effective binding of perhalogenated closo -­borates to serum albumins
polyarginines in model membranes and live cells. J. Am. Chem. Soc. 127, 1114–1115 (2005). revealed by spectroscopic and ITC studies. J. Mol. Struct. 1141, 75–80 (2017).
59. G. Gasparini, E.-­K. Bang, J. Montenegro, S. Matile, Cellular uptake: Lessons from supramolecular 102. T. M. Goszczyński, K. Fink, K. Kowalski, Z. J. Leśnikowski, J. Boratyński, Interactions of boron clusters
organic chemistry. Chem. Commun. 51, 10389–10402 (2015). and their derivatives with serum albumin. Sci. Rep. 7, 9800 (2017).
60. A. Barba-­Bon et al., Boron clusters as broadband membrane carriers. Nature 603, 637–642 (2022). 103. I. I. Slowing, B. G. Trewyn, V.S.-­Y. Lin, Mesoporous silica nanoparticles for intracellular delivery of
61. G. Salluce et al., Size and polarizability of boron cluster carriers modulate chaotropic membrane membrane-­impermeable proteins. J. Am. Chem. Soc. 129, 8845–8849 (2007).
transport. Angew. Chem. Int. Ed. 63, e202404286 (2024). 104. L. Hannibal et al., Alternative conformations of cytochrome c: Structure, function, and detection.
62. Y. Chen et al., Metallacarborane cluster anions of the cobalt bisdicarbollide-­type as chaotropic Biochemistry 55, 407–428 (2016).
carriers for transmembrane and intracellular delivery of cationic peptides. J. Am. Chem. Soc. 145, 105. A. B. Ghisaidoobe, S. J. Chung, Intrinsic tryptophan fluorescence in the detection and analysis of proteins:
13089–13098 (2023). A focus on Forster resonance energy transfer techniques. Int. J. Mol. Sci. 15, 22518–22538 (2014).
63. A. Barba-­Bon et al., All-­inorganic polyoxometalates act as superchaotropic membrane carriers. Adv. 106. J. Mohanty, W. M. Nau, Refractive index effects on the oscillator strength and radiative decay rate of
Mater 36, 2309219 (2024). 2, 3-­diazabicyclo [2.2. 2] oct-­2-­ene. Photochem. Photobiol. Sci. 3, 1026–1031 (2004).
64. J. L. Barton et al., Perfunctionalized dodecaborate clusters as stable metal-­free active materials for 107. D. Toptygin, R. S. Savtchenko, N. D. Meadow, S. Roseman, L. Brand, Effect of the solvent refractive
charge storage. ACS Appl. Energy Mater. 2, 4907–4913 (2019). index on the excited-­state lifetime of a single tryptophan residue in a protein. J. Phys. Chem. B 106,
65. J. Plesek, Potential applications of the boron cluster compounds. Chem. Rev. 92, 269–278 (1992). 3724–3734 (2002).
66. E. L. Muetterties et al., Salts and acids of B10H10-­2 and B1212–2. Inorg. Chem. 3, 444–451 (1964). 108. I. Aviram, The interaction of chaotropic anions with acid Ferricytochrome C. J. Biol. Chem. 248,
67. S. El Anwar et al., Versatile, one-­pot introduction of nonahalogenated 2-­ammonio-­decaborate 1894–1896 (1973).
ions as boron cluster scaffolds into organic molecules; host-­guest complexation with gamma-­ 109. Y. Chen et al., Insights into the enhanced catalytic activity of cytochrome c when encapsulated in a
cyclodextrin. Chem. Commun. 55, 13669–13672 (2019). metal-­organic framework. J. Am. Chem. Soc. 142, 18576–18582 (2020).
68. W. Wang et al., The chaotropic effect as an orthogonal assembly motif for multi-­responsive 110. F. Le Guern, V. Mussard, A. Gaucher, M. Rottman, D. Prim, Fluorescein derivatives as fluorescent
dodecaborate-­cucurbituril supramolecular networks. Chem. Commun. 54, 2098–2101 (2018). probes for ph monitoring along recent biological applications. Int. J. Mol. Sci. 21, 9217 (2020).
69. K. I. Assaf et al., Water structure recovery in chaotropic anion recognition: high-­affinity binding of 111. J. Cebula, K. Fink, J. Boratyński, T. M. Goszczyński, Supramolecular chemistry of anionic boron
dodecaborate clusters to γ-­cyclodextrin. Angew. Chem. Int. Ed. 54, 6852–6856 (2015). clusters and its applications in biology. Coord. Chem. Rev. 477, 214940 (2023).
70. N. A. Bernier et al., Ex vivo and In vivo evaluation of dodecaborate-­based clusters encapsulated in 112. Y.-­L.P. Ow, D. R. Green, Z. Hao, T. W. Mak, Cytochrome C: Functions beyond respiration. Nat. Rev. Mol.
ferumoxytol nanoparticles. Langmuir 37, 14500–14508 (2021). Cell Biol. 9, 532–542 (2008).
71. E. Hey-­Hawkins, C. V. Teixidor, Boron-­based compounds: Potential and Emerging Applications in 113. S. H. Hristova, A. M. Zhivkov, Isoelectric point of free and adsorbed cytochrome c determined by
Medicine (John Wiley & Sons; Hoboken, NJ, USA, 2018). various methods. Colloids Surf. B Biointerfaces 174, 87–94 (2019).
72. J. C. Axtell, L. M. A. Saleh, E. A. Qian, A. I. Wixtrom, A. M. Spokoyny, Synthesis and applications of 114. S. Gurunathan et al., Enhanced green fluorescent protein-­mediated synthesis of biocompatible
perfunctionalized boron clusters. Inorg. Chem. 57, 2333–2350 (2018). graphene. J. Nanobiotechnology 12, 41 (2014).
73. D. Gabel, Boron clusters in medicinal chemistry: Perspectives and problems. Pure Appl. Chem. 87, 115. K. M. Kanal, G. D. Fullerton, I. L. Cameron, A study of the molecular sources of nonideal osmotic
173–179 (2015). pressure of bovine serum albumin solutions as a function of pH. Biophys. J. 66, 153–160 (1994).
74. N. S. Hosmane, “Boron-­based nanomaterials: Technologies and applications” in Boron Science (CRC 116. P. Ghosh et al., Intracellular delivery of a membrane-­impermeable enzyme in active form using
Press, 2012), vol. 21, pp 489–514. functionalized gold nanoparticles. J. Am. Chem. Soc. 132, 2642–2645 (2010).
75. I. B. Sivaev, V. V. Bregadze, Polyhedral boranes for medical applications: Current status and 117. M. Morales-­Cruz et al., Activation of caspase-­dependent apoptosis by intracellular delivery of
perspectives. Eur. J. Inorg. Chem. 2009, 1433–1450 (2009). Cytochrome C-­based nanoparticles. J. Nanobiotechnology 12, 33 (2014).

PNAS 2024 Vol. 121 No. 44 e2407515121 https://doi.org/10.1073/pnas.2407515121 9 of 10

 118. D. N. Shalaeva, D. V. Dibrova, M. Y. Galperin, A. Y. Mulkidjanian, Modeling of interaction between 123. C. Perez-­Iratxeta, M. A. Andrade-­Navarro, K2D2: Estimation of protein secondary structure from
cytochrome c and the WD domains of Apaf-­1: Bifurcated salt bridges underlying apoptosome circular dichroism spectra. BMC Structural Biology 8, 25 (2008).
assembly. Biol. Direct 10, 29 (2015). 124. C. Guo et al., Gold nanoparticle-­guarded large-­pore mesoporous silica nanocomposites for delivery
119. D. Manzanares, V. Cena, Endocytosis: The nanoparticle and submicron nanocompounds gateway and controlled release of cytochrome c. J. Colloid Interface Sci. 589, 34–44 (2021).
into the cell. Pharmaceutics 12, 371 (2020). 125. J. C. M. Stewart, Colorimetric determination of phospholipids with ammonium ferrothiocyanate.
120. H. Nakamura, S. Watano, Direct permeation of nanoparticles across cell membrane: A review. Kona Anal. Biochem. 104, 10–14 (1980).
Powder Part. J. 35, 2018011 (2018). 126. B. Apellániz, J. L. Nieva, P. Schwille, A. J. García-­Sáez, All-­or-­none versus graded: Single-­vesicle
121. R. Rezgui, K. Blumer, G. Yeoh-­Tan, A. J. Trexler, M. Magzoub, Precise quantification of cellular uptake analysis reveals lipid composition effects on membrane permeabilization. Biophys. J. 99,
of cell-­penetrating peptides using fluorescence-­activated cell sorting and fluorescence correlation 3619–3628 (2010).
spectroscopy. Biochim. Biophys. Acta 1858, 1499–1506 (2016). 127. Z. Boban, I. Mardešić, W. K. Subczynski, M. Raguz, Giant unilamellar vesicle electroformation:
122. Y. Jiang et al., The interplay of size and surface functionality on the cellular uptake of sub-­10 nm What to use, what to avoid, and how to quantify the results. Membranes (Basel) 11, 860
gold nanoparticles. ACS Nano 9, 9986–9993 (2015). (2021).
Downloaded from https://www.pnas.org by 106.78.52.19 on October 26, 2024 from IP address 106.78.52.19.

10 of 10 https://doi.org/10.1073/pnas.2407515121 pnas.org