Cellular uptake of substrate-initiated cell-penetrating poly(disulfide)s

Article abstract of DOI:10.1021/ja501581b

Substrate-initiated, self-inactivating, cell-penetrating poly(disulfide)s (siCPDs) are introduced as general transporters for the covalent delivery of unmodified substrates of free choice. With ring-opening disulfide-exchange polymerization, we show that guanidinium-rich siCPDs grow on fluorescent substrates within minutes under the mildest conditions. The most active siCPD transporters reach the cytosol of HeLa cells within 5 min and depolymerize in less than 1 min to release the native substrate. Depolymerized right after use, the best siCPDs are nontoxic under conditions where cell-penetrating peptides (CPPs) are cytotoxic. Intracellular localization (cytosol, nucleoli, endosomes) is independent of the substrate and can be varied on demand, through choice of polymer composition. Insensitivity to endocytosis inhibitors and classical structural variations (hydrophobicity, aromaticity, branching, boronic acids) suggest that the best siCPDs act differently. Supported by experimental evidence, a unique combination of the counterion-mediated translocation of CPPs with the underexplored, thiol-mediated covalent translocation is considered to account for this decisive difference.

Full text of DOI:10.1021/ja501581b

Article
pubs.acs.org/JACS
Cellular Uptake of Substrate-Initiated Cell-Penetrating
Poly(disulfide)s
Giulio Gasparini,† Eun-Kyoung Bang,†,§ Guillaume Molinard,† David V. Tulumello,†,‡ Sandra Ward,†
Shana O. Kelley,‡ Aurelien Roux,† Naomi Sakai,† and Stefan Matile*,†
†School of Chemistry and Biochemistry, National Centre of Competence in Research (NCCR) Chemical Biology, University of
Geneva, Geneva 1211, Switzerland
‡Department of Pharmaceutical Sciences and Department of Biochemistry, University of Toronto, Toronto, Ontario M5S 3M2,
Canada
S
* Supporting Information
ABSTRACT: Substrate-initiated, self-inactivating, cell-penetrating poly-
(disulfide)s (siCPDs) are introduced as general transporters for the covalent
delivery of unmodified substrates of free choice. With ring-opening disulfide-
exchange polymerization, we show that guanidinium-rich siCPDs grow on
fluorescent substrates within minutes under the mildest conditions. The
most active siCPD transporters reach the cytosol of HeLa cells within 5 min
and depolymerize in less than 1 min to release the native substrate.
Depolymerized right after use, the best siCPDs are nontoxic under
conditions where cell-penetrating peptides (CPPs) are cytotoxic. Intra-
cellular localization (cytosol, nucleoli, endosomes) is independent of the
substrate and can be varied on demand, through choice of polymer composition. Insensitivity to endocytosis inhibitors and
classical structural variations (hydrophobicity, aromaticity, branching, boronic acids) suggest that the best siCPDs act differently.
Supported by experimental evidence, a unique combination of the counterion-mediated translocation of CPPs with the
underexplored, thiol-mediated covalent translocation is considered to account for this decisive difference.
proposed to grow CPD transporters directly on molecular
substrates in solution, in situ, right before delivery (Figure 1).8
The term “substrate” is used here to designate any object in
need of assistance to enter cells (e.g., drugs, probes, peptides,
proteins, DNA, RNA, etc.). Reductive depolymerization of the
obtained substrate-initiated (si) CPDs by endogenous gluta-
thione (GSH) in the cytosol would then eliminate toxicity and
liberate the native substrate. The siCPD concept promises
access to a general, fast, and noninvasive method for the
covalent delivery of unmodified substrates, nontoxic, traceless,
avoiding noncovalent formulations, and applicable to any
substrate of free choice.
The growth of the CPD transporter on the molecular
substrate in solution is initiated by a thiol (or a complementary
functional group converter) that acts as an initiator I of the
disulfide-exchange polymerization (Figure 1). Nucleophilic
disulfide exchange opens the strained disulfide in the otherwise
freely variable monomer M, forms a covalent disulfide bond
between I and M, and regenerates a reactive thiol to attack the
next M. The polymerization is terminated with an iodoaceta-
mide T. The simplest possible terminator T is used in this
study, but the introduction of additional drugs or probes with
the freely variable T is of course inviting for the future. In this
study, siCPDs are grown on fluorescent substrates to monitor,
INTRODUCTION
■
Efficient cellular delivery is one of the key problems that
hampers discovery and development of novel drugs and
probes.1 The problem is most pronounced for but not limited
to larger substrates, including oligonucleotides, proteins, and
nanoparticles. Since the discovery of the TAT peptide more
than 20 years ago,2 a broad variety of arginine-rich cell-
penetrating peptides (CPPs) and CPP mimics have been
introduced for this purpose.3 Attached or complexed to CPPs,
otherwise undeliverable molecules such as drugs, fluorophores,
proteins, siRNA, plasmid DNA, or quantum dots are able to
cross the cellular barrier. However, CPPs can be cytotoxic, are
often trapped in the endosomes, and do not always work
reliably. Counterion-mediated translocation has been intro-
duced to improve direct cytosolic delivery and bypass
endosomal capture.4 To reduce toxicity, the sequence as well
as the peptidic backbone of CPPs have been varied extensively.3
One of the most promising modifications is represented by cell-
penetrating poly(disulfide)s (CPDs).5,6 Despite the growing
interest in CPDs, synthetic methods to produce them are
mainly focused on the postmodification of existing polymers
(such as polyethylenimine, PEI) or the reaction of monomers
already containing a disulfide, and their use has been mainly
limited to noncovalent gene transfection.
Inspired by the robustness of the recently discovered ring-
opening disulfide-exchange polymerization to grow complex
functional architectures directly on solid surfaces,7 we have
Received: February 14, 2014
Published: April 15, 2014
© 2014 American Chemical Society
6069
dx.doi.org/10.1021/ja501581b | J. Am. Chem. Soc. 2014, 136, 6069−6074

Journal of the American Chemical Society
Article
the membrane surface,1,6 i.e., a conceptually innovative,
dynamic-covalent9 way to enter cells.
RESULTS AND DISCUSSION
■
Design and Synthesis. To study the capability of siCPDs
to enter cells, fluorescent initiators I1 derived from
carboxyfluorescein (CF) and I2 from 5-carboxytetramethylr-
hodamine (TAMRA) and monomers M1−M6 were envisioned
first (Figure 1). In M1, racemic lipoic acid offers the strained
disulfide needed for ring-opening disulfide-exchange polymer-
ization initiated by the fluorescent thiols I1 and I2, whereas L-
arginine offers the guanidinium cation needed to obtain
polymers that can cross bilayer membranes like CPPs. In M2,
the methyl ester of M1 is replaced by a benzyl ester to increase
hydrophobicity and add π-basicity to the polymer. In M3, the
lipoic acid of M1 is replaced by a more reactive asparagusic
acid. In M4, the spacer between the guanidinium cation and
lipoic acid is as short as possible. The new M5 and M6 contain
the essence of phenylalanine (Phe) and tryptophan (Trp) with
similarly minimalist spacers for copolymerization with the
cationic M4. The synthesis of all initiators and monomers was
very straightforward (Scheme S1, Supporting Information).10
Substrate-initiated polymerization8 and copolymerization11
with M1−M6 were best in buffer at pH 7.0−7.5 in 5−30 min at
room temperature, depending significantly on the nature of the
monomer. The formation of siCPDs was routinely followed by
the appearance of transport activity in fluorogenic vesicles (see
below).8,10,11 All new siCPDs were characterized by gel
permeation chromatography (GPC; Figure S1, Supporting
Information) to determine molecular weight and polydispersity
and MALDI MS to also confirm the composition of the
copolymers (Figures S2−S4, Supporting Information). Under
these conditions, 200 mM M1−M4 polymerized with 5 mM I1,
a thiolated derivative of CF, for example, afforded fluorescent
siCPDs 1−4 with an average molecular weight (Mw) of 6000−
10000 and a polydispersity index (PDI) of ∼1.5 (Table S2,
Supporting Information). The length of the polymers was
variable on demand; shorter polymers were obtained with
higher initiator concentrations or shorter reaction time. All
siCPDs tested were purified by GPC before use to remove
Figure 1. Concept of substrate-initiated, self-inactivating CPD
transporters. Fluorescent initiators (I1 derived from carboxyfluor-
escein and I2 derived from rhodamine) were used to generate
polymers through ring-opening disulfide-exchange polymerization or
copolymerization of monomers M1−M6. The polymerization process
was controlled using iodoacetamide T as the terminator.
for the first time, their entry into HeLa cells. Experimental
evidence is provided that siCPDs (1) enter cells within 5 min,
(2) can depolymerize in less than 1 min after uptake, (3) are
much less toxic than CPPs, (4) operate independent of the
attached substrate, (5) can target different organelles on
demand, and (6) can enter cells by endocytosis or direct
translocation across the membrane barrier, depending on their
structure. Presumably, their uptake is mediated by a so far
underexploited disulfide exchange with endogenous thiols at
Figure 2. Cellular uptake of fluorescent siCPDs. (A) CLSM images of HeLa cells after 15 min of incubation with CF-labeled polymers 1−4, 500 nM
polymer in Leibovitz medium at 37 °C. (B) Same for polymers 1 and 4 prepared from CF (I1) or TAMRA (I2). (C) Spinning disk microscopy
kinetics of the uptake of CF-labeled polymer 4 into HeLa cells transfected in the presence of DRAQ5 (500 nM 4 in Leibovitz medium, 37 °C).
Images were taken with a time interval of 1 min. The first image corresponds to t = 0.
6070
dx.doi.org/10.1021/ja501581b | J. Am. Chem. Soc. 2014, 136, 6069−6074

Journal of the American Chemical Society
Article
Figure 3. Mechanistic and structural insights into the entry of siCPDs into HeLa cells. (A) Flow cytometry analysis counting fluorescent HeLa cells
after incubation with 500 nM initiator I1 (left, light green, overlap with blank) and CF-labeled (co)polymers 1/2 (0:1, 2:1, 8:1, 1:0, with increasing
fluorescence instensity I). Surface-bound material was removed by washing three times with heparin-containing PBS. (B) Uptake quantification for
CF-labeled polymers 1−4 with CLSM (dark) and flow cytometry (light), normalized to fluorescence intensity I with 1. (C) Flow cytometry data for
the uptake of CF-labeled polymers 1 (dark), 4 (light), and 2 (empty) into HeLa cells at 4 and 37 °C and, for 1 and 4 only, also in the presence of
chlorpromazine (Ch), wortmannin (W), and methyl-β-cyclodextrin (βCD), normalized to I at 37 °C. (D) Flow cytometry data for the uptake of low
(light) and high (dark) molecular weight, CF-labeled polymer 4 and copolymers 4/5 (8:1) and 4/6 (8:1) into HeLa cells, normalized to I with 1
(A). Error bars indicate the mean standard deviation, n ≥ 3. (E) Structure of additional monomers prepared to grow siCPDs (compare Figure 1).
unused reagents and eventual side products with low molecular
weight.
medium at 37 °C (Figure 2C). Already after 2 min of
incubation, intracellular fluorescence could be observed. Within
10 min, siCPD 4 accumulated in the cytosol but was unable to
significantly enter into the nucleus. Intracellular fluorescence
intensity was preserved after removal of the polymer solution
used to incubate the cells. This irreversible accumulation of
siCPD 4 in the cytosol was in excellent agreement with fast
reductive depolymerization of siCPD 4 in the cytosol (see
below).
Flow cytometry was used for rapid access to quantitative data
(Figure 3A). Comparison with CLSM data for polymers 1−4
revealed comparable trends but clear underestimates with
CLSM for cytosolic emission from siCPD 4, which performs
the best according to flow cytometry (Figure 3B).
Significant uptake of siCPDs 1, 2, and 4 was observed at 4 °C
(Figure 3C). This demonstrated that uptake does not occur
exclusively by endocytosis.3,4b Compared to that at 37 °C,
reduced activity found at 4 °C could originate simply from less
favorable direct translocation across the rigidified membranes,12
although losses from missing contributions from endocytosis
cannot be excluded.
Cellular Uptake. The ability of the siCPDs 1−4 to
transport a fluorescent CF substrate into HeLa Kyoto cells
was examined using confocal laser scanning microscopy
(CLSM). After incubation for 15 min at 37 °C with a 500
nM solution of the desired polymer in Leibovitz medium, the
cells were washed with PBS containing heparin (20 U/mL) to
remove siCPDs that are bound reversibly at the cell surface (or
can exit the cells as easily as they entered). Significant
intracellular fluorescent signals were recorded from cells
incubated with siCPDs 1, 2, and 4 (Figure 2A). Only polymer
3 grown with asparagusic instead of lipoic acid did not carry CF
substrates into HeLa cells. The inactivity of polymer 3 in HeLa
cells perfectly reflected the inactivity of polymer 3 as a
transporter in fluorogenic vesicles.8 This finding thus
corroborated the validity of model studies in fluorogenic
vesicles to, at least in part, predict the activity of siCPDs as well
as CPPs.8,11 The CF and TAMRA initiators I1 and I2 alone did
not enter HeLa cells under experimental conditions (Figure S7,
Supporting Information).
The intracellular distribution of CF-labeled siCPDs 1−4
differed significantly. The original siCPD 1 accumulated mainly
inside the nucleus, especially in the nucleoli (Figure 2A, top
right, B, top left). The minimalist siCPD 4 localized mainly in
the cytoplasm (Figure 2A, bottom right, B, bottom left). The
more hydrophobic, π-basic siCPD 2 remained mainly trapped
inside the endosomes (Figure 2A, top left). The localization of
different siCPDs in different organelles was relatively
independent of the concentration and incubation time.
To explore the dependence of siCPD uptake on the attached
substrate, polymers 1′ and 4′ were grown with initiator I2, a
thiolated derivative of TAMRA. HeLa Kyoto cells were
incubated under the conditions used for the CF-labeled siCPDs
1 and 4. The results demonstrated that siCPDs 1′ and 4′ grown
with the neutral TAMRA I2 behave exactly like siCPDs 1 and 4
grown with the anionic CF I1 (Figure 2B). Namely, siCPDs
delivered their substrates reliably to the nucleoli and cytoplasm,
respectively, independent of the nature of the substrates.
The kinetics of cellular uptake and localization of CF-labeled
siCPD 4 were measured with HeLa Kyoto cells that were
treated with DRAQ5 to visualize their nuclei in the far red.
Spinning disk microscopy images were recorded immediately
after the addition of the polymer solution and with a time
interval of 1 min during incubation with 500 nM in Leibovitz
Comparably high activity was obtained with siCPD 2,
although endosomal location at 37 °C demonstrated exclusive
uptake by endocytosis, but CLSM images obtained at 4 °C
showed the hydrophobic siCPD 2 mainly at the surface,
probably too deeply buried in the hydrophobic core of the
membrane to be removed by washing with heparin (Figure
S6B, Supporting Information).13 Preserved localization for
siCPDs 1 and 4 in the nucleoli and cytosol, respectively, at 4 °C
was in agreement with a preserved uptake mechanism, i.e.,
dominant direct translocation across the membrane (Figure
S6A,C). The validity of this interpretation was further
supported by insensitivity of siCPDs 1 and 4 to selective
endocytosis inhibitors, i.e., wortmannin for macropinocytosis,
chlorpromazine for clathrin-mediated endocytosis, and methyl-
β-cyclodextrin (MβCD) for caveolar endocytosis.14 The
distinct increase in activity of siCPDs 4 with MβCD could
originate from facilitated translocation across cholesterol-poor
membranes.15
The activity of siCPD 4 increased with increasing length of
the polymer (Figure 3D). The same trend, even more
pronounced, was found for the more hydrophobic copolymers
4/5 and 4/6. Significant length dependence was as expected
because both transport activity in membranes4a,16 and
6071
dx.doi.org/10.1021/ja501581b | J. Am. Chem. Soc. 2014, 136, 6069−6074

Journal of the American Chemical Society
Article
Figure 4. Toxicity, depolymerization, and thiol-mediated translocation of siCPDs. (A) Cell viability measured with the MTT assay for polymer 4
(squares) and poly-L-arginine (circles) at different concentrations. (B) Relative transport activity Y of polymer 4 (filled symbols) and 1 (empty
symbols) in fluorogenic vesicles after incubation with 2.5 mM (circles) and 5 mM (squares) glutathione (GSH) for time t. (C) Effective
concentration of polymer 4 in fluorogenic vesicles in the presence of increasing amounts of octadecanethiol (circles) and GSH (squares). (D) Flow
cytometry data for the uptake of CF-labeled polymers 1 and 4 into HeLa cells with (dark) or without (light) pretreatment with 1.2 mM DTNB for
30 min before addition of siCPDs. Uptake for each polymer is normalized to 1.0 in the absence of DTNB. (E) Proposed mechanism for thiol-
mediated uptake of siCPDs.
contributions from endocytosis (and toxicity) generally
increase with the length of related polymers, including CPPs.3
However, the introduction of hydrophobic benzyl and
indolyl residues in copolymers 4/5 and 4/6 neither increased
the activity of siCPD 4 nor changed the preferential
accumulation in the cytosol (Figure 3D). Consistently
decreasing activity with increasing hydrophobicity, in sharp
contrast to many observations with CPPs and the general
importance of Phe and Trp for membrane protein function,
already provided first indications that siCPDs might be
fundamentally different and function by a different mechanism
(see below). Similarly decreasing activity with increasing
hydrophobicity was found in the amino acid series with π-
basic copolymers 1/2 (Figure 3A) and 1/7 and, slightly less
pronounced, also with superhydrophobic π-acids in copolymers
1/8 and 1/9 (Figure S8, Supporting Information). The
branched polymer 10 or copolymers 1/10 obtained with the
divalent monomer M10 showed particularly poor activity and
high toxicity (Figures S8 and S9, Supporting Information). Also
boronic acids in copolymers 1/11 did not increase activity
despite the possibility to assist uptake with the formation of
dynamic covalent boronic ester bonds with glycosaminoglycans
at the cell surface, thus enhancing the local concentration and
promoting uptake.5f,17 Quite the contrary, a less desired
increase of copolymers 1/11 permanently located at the cell
surface and in the endosomes was noted in CLSM images
(Figure S6D, Supporting Information). This general failure to
improve performance with fairly standard structural modifica-
tions suggested that the cellular uptake of siCPDs 1 and 4 is
already near maximum levels for this class of transporters and
possibly occurs by a different mechanism compared to that of
other CPP mimics, i.e., thiol-mediated uptake.1,6
solution was removed, and the cells were washed with heparin
sulfate. After addition of culture media, the cells were incubated
for 24 h prior to execution of the MTT assay. Under these
conditions, siCPD 4 showed negligible cytotoxicity up to 10
μM, whereas polyarginine (16.4 kDa) exhibited an EC50 below
2 μM (Figure 4A; hexaarginine was also cytotoxic at 10 μM).
The nontoxicity of siCPD 4 was in excellent agreement with
rapid depolymerization as soon as the cytosol is reached (see
below). All siCPDs tested had good cell viability in the
concentration used for the cellular uptake measurements
(Figure S9, Supporting Information). However, contrary to
the nontoxic siCPD 4 in the cytosol, siCPD 1 was increasingly
cytotoxic at higher concentrations. This finding was consistent
with incomplete depolymerization before leaving the cytosol
(see below) and subsequent interference with cellular function
after binding to the oligonucleotides in the nucleoli.
Cytosolic depolymerization was quantified next. It is difficult
to quantify in cells, so transport studies in fluorogenic vesicles
were selected to secure direct evidence. In this firmly validated
assay (see above, Figure S5, Supporting Information),8,11 large
unilamellar vesicles (LUVs) composed of egg yolk phosphati-
dylcholine (EYPC) are loaded with CF at concentrations high
enough to ensure self-quenching. Local dilution upon CF
export by siCPDs,8,11 CPPs,3,4,15 or other anion transporters is
then observed as fluorescence recovery. Consistent with perfect
cellular uptake, siCPDs 1 and 4 showed maximal transport
activity in the CF assay (without activation by amphiphilic
counterions, polyarginine is inactive in EYPC vesicles).4a
Incubation with GSH at cytosolic concentrations resulted in
rapid loss of activity. The cytosolic siCPDs 4 was completely
inactivated within less than 1 min by 2.5 mM GSH (Figure 4B,
●). Depolymerization of the nucleolar siCPDs 1 by 2.5 mM
GSH was complete within about 5 min (Figure 4B, □). The
lifetime of the transporter in 5 mM GSH was clearly shorter
(Figure 4B, ○).
Toxicity, Internal Depolymerization, and Exofacial
Thiols. Why are siCPDs better? What really makes the
difference? Toxicity was quantified first. MTT assays were
performed with the cytosolic siCPD 4 in comparison with poly-
L-arginine as a comparable CPP. In this assay, the metabolic
activity of the cells was assessed by their ability to enzymatically
convert the tetrazolium dye MTT into formazan.18 The
polymers were incubated with HeLa cells for 15 min at a
concentration ranging from 0.1 to 10 μM. Then the polymer
Sensitivity toward exofacial thiols was explored last. The
presence of octadecanethiol in the EYPC membrane signifi-
cantly changed the transport activity of siCPD 4 (Figure 4C,
●). At a molar ratio octadecanethiol/lipid of 1:1, the EC50 of 4,
i.e., the effective siCPD concentration to reach 50% activity,
was weakened by a factor of 4. A similar increase of the EC50
6072
dx.doi.org/10.1021/ja501581b | J. Am. Chem. Soc. 2014, 136, 6069−6074

Journal of the American Chemical Society
Article
was not observed in the presence of equivalent concentrations
of GSH (Figure 4C, □; these GSH concentrations are far
below the ones needed to depolymerize siCPD 4 as in Figure
4B, ●). Activation by thiols at the membrane surface but
insensitivity to equimolar thiols in the water indicated that
disulfide exchange between siCPD 4 and exofacial thiols occurs.
In fluorogenic vesicles, the resulting shortening of siCPD 4
caused a loss in activity. In cells, this covalent binding to the
surface is expected to increase the local concentration of the
siCPD and thus accelerate direct, counterion-mediated trans-
location across the membrane (Figure 4E). Experimental
evidence for thiol-mediated translocation from fluorogenic
vesicles is unprecedented. To evaluate the validity of this
implication from fluorogenic vesicles, the inhibition of thiol-
mediated translocation with Ellman’s reagent (i.e., 5,5′-
dithiobis-2-nitrobenzoic acid, or DTNB) was explored directly
in HeLa cells. The cells were incubated with the cell-
impermeable DTNB for 30 min to convert all free thiols at
the surface into disulfides. According to flow cytometry
measurements, cellular uptake of siCPDs 1 and 4 was
significantly reduced in the absence of exofacial thiols (Figure
4D). Similarly reduced uptake has been observed previously for
other cell-penetrating poly(disulfide)s.1,6 The inactivation by
Ellman’s reagent was more pronounced for siCPD 4 than for
siCPD 1. Considering that siCPD 4 depolymerizes faster than
siCPD 1 (Figure 4C), this difference implied that the efficiency
of thiol-mediated translocation is determined by the velocity of
disulfide exchange.
cross the membrane like CPPs along transient micellar defects,
and detach into the cytosol by disulfide exchange with
intracellular glutathione (Figure 4E). This fascinating, con-
ceptually innovative thiol/counterion-mediated uptake mecha-
nism drives the concept of covalent delivery of unmodified
substrates to the extreme: The self-inactivating transporters not
only grow covalently on the molecular substrate in solution,
they also bind covalently to the membrane they are crossing.
The inability of both initiators I1 and I2 to enter HeLa cells
demonstrates that, contrary to predictions,1 simple thiolation is
insufficient to turn on thiol-mediated uptake. Permanent
capture at the surface and in endosomes found with boronic
acids in copolymers 1/11 confirmed that thiol-mediated
translocation requires more than dynamic covalent bonds at
the cell surface. For covalent translocation, the simplest siCPDs
are then best because they offer disulfide bonds and
guanidinium cations at the highest effective concentration for
thiol/counterion-mediated translocation.
Once the siCPD has arrived in the cytosol, disulfide-
exchange kinetics seem to determine its final destination.
Whereas the instantaneous depolymerization of transporter 4
liberates the unmodified substrates in the cytosol, a lifetime of
less than 5 min is sufficient for transporter 1 to proceed into the
nucleus. Rapid self-inactivation right after uptake also explains
nicely why transporter 4 is completely nontoxic. Most
importantly, it appears that thiol/counterion-mediated trans-
location is also controlled by disulfide-exchange kinetics.
However, the origin of the different depolymerization kinetics
is unknown. Differences in polymer length are insufficient to
account for their extent. Possibly, the guanidinium cations in 4
are in the best position to activate the thiolate leaving groups by
intramolecular ion pairing, but this is just a hypothesis.
Apparently essential to fully understand and exploit the unique
advantages of siCPDs, the origin, the variability, and the
functional consequences of disulfide-exchange kinetics of
siCPDs are currently explored in the greatest detail, with
particular emphasis on thiol/counterion-mediated transloca-
tion.
CONCLUSIONS
■
This study introduces substrate-initiated cell-penetrating poly-
(disulfide)s as general, nontoxic, self-inactivating transporters
for the covalent delivery of native substrates of free choice. We
provide experimental evidence that the formation of siCPDs
can be initiated by fluorescent probes and occurs within
minutes under the mildest conditions (water, room temper-
ature, pH 7) and that the most active siCPDs reach the cytosol
of HeLa cells in 5 min, where they depolymerize in less than 1
min. The most active siCPDs are nontoxic at all tested
concentrations (up to 10 μM), whereas comparable CPPs are
toxic. Intracellular localization and the uptake mechanism are
independent of the substrate and can be varied on demand by
varying the hydrophobicity and disulfide-exchange kinetics.
Namely, more hydrophobic siCPDs enter mainly by
endocytosis and accumulate in endosomes. More hydrophilic
siCPDs with fast disulfide-exchange kinetics accumulate in the
cytosol because they depolymerize as soon as they arrive. In
clear contrast, the lifetime of siCPDs with slow disulfide-
exchange kinetics is sufficient for them to proceed from the
cytosol to the nucleus and accumulate on the anionic
oligonucleotides in the nucleoli.
ASSOCIATED CONTENT
■
S
* Supporting Information
Experimental details and NMR spectra. This material is
available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
stefan.matile@unige.ch
Present Address
§E.-K.B.: Korea Institute of Science and Technology (KIST),
Seoul 136-791, South Korea.
The simplest siCPDs are the best. The most compact siCPD
4 is derived from lipoic acid and a guanidinium cation, the
original siCPD 1 simply from lipoic acid and L-arginine.
Classical structural modifications (hydrophobicity, aromaticity,
branching, boronic acids) do not improve performance. This
unresponsiveness suggests that the best siCPDs act differently.
Insensitivity toward inhibitors demonstrates that endocytosis is
almost irrelevant. Significant dependence on the presence of
exofacial thiols suggests that the counterion-mediated trans-
location known from CPPs4 is coupled with thiol-mediated
translocation.1,6 Namely, siCPDs bind covalently to the
membrane surface by disulfide exchange with exofacial thiols,
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank the NMR, Sciences Mass Spectrometry (SMS), and
Bioimaging (BiP) platforms for their services and the
University of Geneva, European Research Council (ERC
Advanced Investigator), National Centre of Competence in
Research (NCCR) Chemical Biology, and Swiss National
Science Foundation for financial support.
6073
dx.doi.org/10.1021/ja501581b | J. Am. Chem. Soc. 2014, 136, 6069−6074

Journal of the American Chemical Society
Article
331−340. (g) Kim, T.; Rothmund, T.; Kissel, T.; Kim, S. W. J.
Controlled Release 2011, 152, 110−119. (h) Wua, C.; Li, J.; Zhua, Y.;
REFERENCES
■
(1) Torres, A. G.; Gait, M. J. Trends Biotechnol. 2012, 30, 185−190.
(2) (a) Green, M.; Lowenstein, P. M. Cell 1988, 55, 1179−1188.
(b) Frankel, A. O.; Pabo, C. O. Cell 1988, 55, 1189−1193.
Chena, J.; Oupicky, D. Biomaterials 2013, 34, 8843−8850. (i) Zeng,
́
H.; Little, H. C.; Tiambeng, T. N.; Williams, G. A.; Guan, Z. J. Am.
Chem. Soc. 2013, 135, 4962−4965. (j) Drake, C. A.; Aissaoui, A.;
Argyros, O.; Thanou, M.; Steinke, J. H. G.; Miller, A. D. J. Controlled
Release 2013, 171, 81−90.
(3) (a) Stanzl, E. G.; Trantow, B. M.; Vargas, J. R.; Wender, P. A. Acc.
Chem. Res. 2013, 135, 2944−2954. (b) Sgolastra, F.; Deronde, B. M.;
Sarapas, J. M.; Som, A.; Tew, G. N. Acc. Chem. Res. 2013, 135, 2977−
2987. (c) Futaki, S.; Hirose, H.; Nakase, I. Curr. Pharm. Des. 2013, 19,
2863−2868. (d) Bautista, A. D.; Craig, C. J.; Harker, E. A.; Shepartz,
A. Curr. Opin. Chem. Biol. 2007, 11, 685−692. (e) Bechara, C.; Sagan,
S. FEBS Lett. 2013, 587, 1693−1702. (f) Torchilin, V. P. Adv. Drug
Delivery Rev. 2008, 60, 548−558. (g) Stewart, K. M.; Horton, K. L.;
Kelley, S. O. Org. Biomol. Chem. 2008, 6, 2242−2255. (h) Sgolastra, F.;
Minter, L. M.; Osborne, B. A.; Tew, G. N. Biomacromolecules 2014, 10,
812−820. (i) Weinstain, R.; Savariar, E. N.; Felsen, C. N.; Tsien, R. Y.
J. Am. Chem. Soc. 2014, 136, 874−877. (j) Lei, E. K.; Pereira, M. P.;
Kelley, S. O. Angew. Chem., Int. Ed. 2013, 52, 9660−9663. (k) Lnoue,
M.; Tong, W.; Esko, J. D.; Tor, Y. ACS Chem. Biol. 2013, 8, 1383−
1388. (l) Tang, H.; Yin, L.; Kim, K. H.; Cheng, J. Chem. Sci. 2013, 4,
(6) (a) Kang, H. C.; Kang, H.-J.; Bae, Y. H. Biomaterials 2011, 32,
1193−1203. (b) Li, J.; Manickam, D. S.; Chen, J.; Oupicky, D. Eur. J.
Pharm. Sci. 2012, 46, 173−180. (c) Brulisauer, L.; Kathriner, N.;
̈
Prenrecaj, M.; Gauthier, M. A.; Leroux, J.-C. Angew. Chem., Int. Ed.
2012, 51, 12454−12458.
(7) (a) Sakai, N.; Lista, M.; Kel, O.; Sakurai, S.; Emery, D.; Mareda,
J.; Vauthey, E.; Matile, S. J. Am. Chem. Soc. 2011, 133, 15224−15227.
(b) Lista, M.; Areephong, J.; Sakai, N.; Matile, S. J. Am. Chem. Soc.
2011, 133, 15228−15231. (c) Sakai, N.; Matile, S. J. Am. Chem. Soc.
2011, 133, 18542−18545. (d) Orentas, E.; Lista, M.; Lin, N.-T.; Sakai,
N.; Matile, S. Nat. Chem. 2012, 4, 746−750. (e) Sforazzini, G.;
Orentas, E.; Bolag, A.; Sakai, N.; Matile, S. J. Am. Chem. Soc. 2013, 135,
12082−12090.
́
3839−3844. (m) Ornelas-Megiatto, C.; Wich, P. R.; Frechet, J. M. J.
(8) Bang, E.-K.; Gasparini, G.; Molinard, G.; Roux, A.; Sakai, N.;
Matile, S. J. Am. Chem. Soc. 2013, 135, 2088−2091.
(9) (a) Li, J.; Nowak, P.; Otto, S. J. Am. Chem. Soc. 2013, 135, 9222−
9239. (b) Lehn, J. M. Top. Curr. Chem. 2012, 322, 1−32. (c) Black, S.
P.; Sanders, J. K. M.; Stefankiewicz, A. R. Chem. Soc. Rev. 2014, 43,
1861−1872. (d) Wilson, A.; Gasparini, G.; Matile, S. Chem. Soc. Rev.
2014, 43, 1948−1962.
Am. Chem. Soc. 2012, 134, 1902−1905. (n) Suzuki, Y.; Okuro, K.;
Takeuchi, T.; Aida, T. J. Am. Chem. Soc. 2012, 134, 15273−15276.
(o) Som, A.; Reuter, A.; Tew, G. N. Angew. Chem., Int. Ed. 2012, 51,
980−983. (p) Appelbaum, J. S.; LaRochelle, J. R.; Smith, B. A.; Balkin,
D. M.; Holub, J. M.; Schepartz, A. Chem. Biol. 2012, 19, 819−830.
(q) He, J.; Hristova, K.; Wimley, W. C. Angew. Chem., Int. Ed. 2012,
51, 7150−7153. (r) Treat, N. J.; Smith, D.; Teng, C.; Flores, J. D.;
Abel, B. A.; York, A. W.; Huang, F.; McCormick, C. L. ACS Macro Lett.
2012, 1, 100−104. (s) Kolonko, E. M.; Pontrello, J. K.; Mangold, S. L.;
Kiessling, L. L. J. Am. Chem. Soc. 2009, 131, 7327−7333.
(t) McNaughton, B. R.; Cronican, J. J.; Thompson, D. B.; Liu, D. R.
Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 6111−6116. (u) Trantow, B.
M.; Cooley, C. B.; Nederberg, F.; Kiesewetter, M. K.; Hedrick, J. L.;
Waymouth, R. M.; Wender, P. A. J. Am. Chem. Soc. 2009, 131, 16401−
16403. (v) Herce, H. D.; Garcia, A. E.; Litt, J.; Kane, R. S.; Martin, P.;
Enrique, N.; Rebolledo, A.; Milesi, V. Biophys. J. 2009, 97, 1917−1925.
(w) Shimanouchi, T.; Walde, P.; Gardiner, J.; Mahajan, Y. R.; Seebach,
D.; Thomae, A.; Kraemer, S. D.; Voser, M.; R. Kuboi, R. Biochim.
Biophys. Acta, Biomembr. 2007, 1768, 2726−2736. (x) Fernandez-
Carneado, J.; Van Gool, M.; Martos, V.; Castel, S.; Prados, P.; De
Mendoza, J.; Giralt, E. J. Am. Chem. Soc. 2005, 127, 869−874.
(y) Fillon, Y.; Anderson, J.; Chmielewski, J. J. Am. Chem. Soc. 2005,
127, 11798−11803. (z) Pantarotto, D.; Briand, J.-P.; Prato, M.; Bianco,
A. Chem. Commun. 2004, 16−17. (aa) Jiang, T.; Olson, E. S.; Nguyen,
Q. T.; Roy, M.; Jennings, P. A.; Tsien, R. Y. Proc. Natl. Acad. Sci. U.S.A.
2004, 101, 17867−17872. (bb) Zhou, P.; Wang, M.; Du, L.; Fisher, G.
W.; Waggoner, A.; Ly, D. H. J. Am. Chem. Soc. 2003, 125, 6878−6879.
(cc) Luedtke, N. W.; Carmichael, P.; Tor, Y. J. Am. Chem. Soc. 2003,
125, 12374−12375. (dd) Umezawa, N.; Gelman, M. A.; Haigis, M. C.;
Raines, R. T.; Gellman, S. H. J. Am. Chem. Soc. 2002, 124, 368−369.
(ee) Sakai, N.; Baumeister, B.; Matile, S. ChemBioChem 2000, 1, 123−
125. (ff) Schwarze, S. R.; Ho, A.; Vocero-Akbani, A.; Dowdy, S. F.
Science 1999, 285, 1569−1572.
(10) See the Supporting Information.
(11) Bang, E.-K.; Ward, S.; Gasparini, G.; Sakai, N.; Matile, S. Polym.
Chem. 2014, 5, 2433−2441.
(12) van Meer, G.; Voelker, D. R.; Feigenson, G. W. Nat. Rev. Mol.
Cell Biol. 2008, 9, 112−124.
(13) Sakai, N.; Takeuchi, T.; Futaki, S.; Matile, S. ChemBioChem
2005, 6, 114−122.
(14) Khalil, I. A.; Kogure, K.; Akita, H.; Harashima, H. Pharmacol.
Rev. 2006, 58, 32−45.
(15) Nishihara, M.; Perret, F.; Takeuchi, T.; Futaki, S.; Lazar, A. N.;
Coleman, A. W.; Sakai, N.; Matile, S. Org. Biomol. Chem. 2005, 3,
1659−1669.
(16) Hennig, A.; Gabriel, G. J.; Tew, G. N.; Matile, S. J. Am. Chem.
Soc. 2008, 130, 10338−10344.
(17) Ellis, G. A.; Palte, M. J.; Raines, R. T. J. Am. Chem. Soc. 2012,
134, 3631−3634.
(18) Mosmann, T. J. Immunol. Methods 1983, 65, 55−63.
(4) (a) Sakai, N.; Matile, S. J. Am. Chem. Soc. 2003, 125, 14348−
14356. (b) Takeuchi, T.; Kosuge, M.; Tadokoro, A.; Sugiura, Y.; Nishi,
M.; Kawata, M.; Sakai, N.; Matile, S.; Futaki, S. ACS Chem. Biol. 2006,
1, 299−303. (c) Schmidt, D.; Jiang, Q. X.; MacKinnon, R. Nature
2006, 444, 775−779. (d) Inomata, K.; Ohno, A.; Tochio, H.; Isogai,
S.; Tenno, T.; Nakase, I.; Takeuchi, T.; Futaki, S.; Ito, Y.; Hiroaki, H.;
Shirakawa, M. Nature 2009, 458, 106−109. (e) Katayama, S.; Nakase,
I.; Yano, Y.; Murayama, T.; Nakata, Y.; Matsuzaki, K.; Futaki, S.
Biochim. Biophys. Acta, Biomembr. 2013, 1828, 2134−2142.
(5) (a) Lin, C.; Engbersen, J. F. J. Expert Opin. Drug Delivery 2009, 6,
421−439. (b) Meng, F. H.; Hennink, W. E.; Zhong, Z. Y. Biomaterials
2009, 30, 2180−2198. (c) Kim, T.-I.; Kim, S. W. React. Funct. Polym.
2011, 71, 344−349. (d) Bang, E.-K.; Lista, M.; Sforazzini, G.; Sakai,
N.; Matile, S. Chem. Sci. 2012, 3, 1752−1763. (e) Son, S.; Namgung,
R.; Kim, J.; Singha, K.; Kim, W. J. Acc. Chem. Res. 2012, 45, 1100−
1112. (f) Piest, M.; Engbersen, J. F. J. J. Controlled Release 2011, 155,
6074
dx.doi.org/10.1021/ja501581b | J. Am. Chem. Soc. 2014, 136, 6069−6074