Penetrating peptides are short amino acid sequences with membrane transport capabilities.[1 ]
[2 ]
[3 ]
[4 ]
[5 ]
[6 ]
[7 ] Although not fully necessary for membrane activity, the cationic and the amphiphilic
character are common features of a great number of penetrating peptides.[1,6,8,9 ] The mechanism of internalization of penetrating peptides still represents a source
of major controversy.[10 ] Initial passive diffusion assumptions have been ruled out because of artifacts related
with the cytoplasm distribution after cell fixation.[11 ] Currently, it is widely accepted that penetrating peptides can be transported by
several mechanisms operating at the same time. Additionally, the final internalization
pathway of a given peptide is often promiscuous and depends on different parameters
such as concentration, nature of the cargo and size of the corresponding supramolecular
aggregates.[10 ]
[12 ] Furthermore, the highly dynamic processes involved in membrane transport such as
counterion exchange[13 ]
[14 ] and membrane partitioning, strongly hinder the acquisition of robust data and the
development of predictive models.[6 ]
[15 ] Additionally, the conformational changes of the transported peptides are difficult
to discern and to study in real time.[16 ]
[17 ] However, it has been recently shown that the cationic arginine topology dramatically
influences important steps of the delivery process including endosomal escape.[18 ]
Javier Montenegro received his PhD (2009) from the University of Santiago de Compostela (USC, Prof.
Susana López). In 2005 he was a visiting scientist in the group of Prof. Steven Ley
(Cambridge University) and in 2007 he was a visiting scientist with Prof. Reza M.
Ghadiri (Scripps). In 2009 he moved to the University of Geneva for three years of
postdoctoral studies with Prof. Stefan Matile. In 2012 he obtained a Juan de la Cierva
fellowship (Prof. Juan R. Granja) and in 2014 a Ramon y Cajal contract at the Organic
Chemistry Department and the CIQUS at the USC. In 2015 he was awarded with a Starting
Grant from the European Research Council (ERC). His research interests lay at the
interphase between organic synthetic and supramolecular chemistry applied to biological
systems and the discovery of new materials. His current projects are focused in the
topological control of supramolecular assemblies for broad applications such as differential
sensing, controlled delivery and tubular-templated composites.
The formation of dynamic bonds constitutes a powerful strategy for the preparation
of amphiphilic membrane transporters.[14 ]
[19 ]
[20 ]
[21 ]
[22 ]
[23 ] Dynamic bonds such as hydrazones, oximes and disulfides offer two fundamental advantages
for the synthesis of membrane carriers.[24 ]
[25 ]
[26 ]
[27 ]
[28 ] First, there is a synthetic advantage, because certain dynamic bonds (i.e. hydrazone,
oxime, disulfide) can be considered as ‘clickable chemistry ’ as they lead to the final products in short times, with high yields and without
side products. On the other hand, dynamic bonds are sensitive to external stimuli
and therefore they can readapt to changes in the environment along the internalization
pathway.
We have recently started a research program aimed to decipher the critical structure-activity
relationships of penetrating peptides. This approach is based in the implementation
of dynamic chemical bonds to change (modulate) the structures and thus the activity
of initial model peptide sequences. In this paper we report proof of concept transport
experiments, in vesicles and in cells, of an amphiphilic pseudo-helical short peptide
that is functionalized by oxime bond formation with hydrophilic and hydrophobic electrophilic
counterparts.
This parent peptide named AcP(Ox)2 (Figure [1 ]A), where Ac stands for acetyl N-terminal, P for peptide backbone and Ox for the
two alkoxyamine pendants, was synthesized by solid-phase peptide synthesis using orthogonal
protecting groups (Figure S1).[21 ] Lysine amino acids with the lateral protecting group methyltrytil (Mtt) were employed
in order to perform the selective deprotection and the ‘on resin ’ coupling of the Boc-protected alkoxyamine pendants (Figure S1). The peptide was
terminated by acetylation, fully deprotected, purified by reverse-phase HPLC and characterized
by different spectroscopic techniques (see Supporting Information). The sequence of
AcP(Ox)2 was chosen to give rise to an helical peptide with a similar number of cationic and
hydrophobic functionalities after oxime bond formation with a guanidinium aldehyde
(Figure [1 ]B). In this peptide, the alkoxyamine nucleophilic moieties are located at the interphase
between the hydrophilic and the hydrophobic domains in order to minimize the impact
over the secondary structure. The circular dichroism analysis, in water and in trifluoroethanol
(TFE), showed bands at 208 and 222 nm, respectively, indicating a certain helical
behavior of this parent peptide AcP(Ox)2 . Although helical character decreased in aqueous conditions, it was retained at high
temperatures (Figure [1 ]B).
Figure 1 (A) Structure of the peptide R1 P(Ox)2 . Arginine residues are in blue, leucines and alanines in orange and reactive alkoxyamines
in red. Different electrophilic aldehyde tails employed: TIso (isovaleraldehyde), TAcet (acetone), TGly (glyceraldehyde), TGua (guanidinium aldehyde). (B) Top-view helical diagram (heptad) and circular dichroism
spectra at different temperatures (10–60 °C) of the parent peptide AcP(Ox)2 . (C) Acronym definition and scheme for peptide functionalization with the different
aldehydes to afford oxime-modulated peptides for membrane transport applications.
Color code: black: peptide backbone; blue: arginine; orange: leucine; red: reactive
alkylhydroxylamine; purple: isovaleraldehyde; green: glyceraldehyde; brown: acetone.
AcP(Ox)2 is ready to react with aldehydes (H2 O/DMSO/AcOH 1:1:0.05, 60 °C, 1 h) to afford the corresponding oxime-modulated peptides
that can be readily used in transport experiments under physiological conditions and
without any further purification (Figure [1 ]C, see Supporting Information for details).
We carried out initial transport experiments in model membranes to screen the transport
ability of this peptide with a short double stranded nucleotide cargo (herring DNA,
<50 base pairs). For this purpose we prepared egg yolk phosphatidylcholine (EYPC)
large unilamellar vesicles (LUVs) loaded with the anionic fluorophore 8-hydroxy-1,3,6-pyrenetrisulfonate
(HPTS) and the cationic quencher p -xylene-bispyridinium bromide (DPX) (Figure [2 ]).[23 ]
[24 ] In this assay, the intravesicular concentrations of dye (HPTS) and quencher (DPX)
are balanced to afford minimal emission of the resting state vesicles (see Supporting
Information). In the transport experiment, we measured the emission of the HPTS of
vesicles suspensions upon sequential addition of the different components (peptide,
DNA and detergent; Figure [2 ]). The first addition of the peptide alone (at t = 25 s) should not lead to any increase in the fluorescence unless the cationic peptide
behaves as a mere detergent. However, if the following second addition of the anionic
DNA (at t = 50 s) leads to an increase in the fluorescence emission, the peptide behaves as
an activator of the DNA membrane translocation (Figure [2 ]). At the end of the experiment (t = 225 s) we added a potent detergent (Triton-X-100) to release all the vesicle contents
and normalize the data.
Figure 2 Vesicle transport experiments. (A) Scheme of the vesicle transport experiments. (B)
Representative normalized kinetics of transporter-mediated increase in HPTS fluorescence,
following addition of peptide (5 nM, t = 25 s), DNA (herring DNA, <50 bp, 700 nM, t = 50 s) and Triton-X-100 (7.5 µM, t = 225 s). Kinetic traces demonstrating increasing transporter activity for AcP(TGua )2 (filled blue circles), lack of activity for the AcP(TIso )2 (filled purple squares), control experiment with parent peptide AcP(Ox)2 (empty black circles) and control experiment with only vesicles (empty gray squares).
Two different aldehydes (hydrophilic and hydrophobic) were selected to test oxime-modulated
transport in vesicles, namely TGua (guanidinium aldehyde) and TIso (isovaleraldehyde; Figures S3 and S4). The normalized transport kinetics showed that
the parent peptide AcP(Ox)2 was totally inactive (Figure [2 ]B). However, an important increase in the activity was observed after the oxime linkage
of the peptide with the cationic hydrophilic aldehyde TGua (Figure [2 ]B). Increasing the hydrophobic character of the parent peptide with the leucine mimic
TIso did not increase the transport activity. This result is in clear contrast with previous
observations of dendritic hydrazone and oxime-based amphiphiles,[24 ]
[29 ]
[30 ]
[31 ] which required a strong hydrophobic enrichment in order to activate DNA transport
across the lipid bilayer of LUVs.
To further investigate and visualize oxime-modulated peptides in cells, the peptide
was terminated with a carboxyfluorescein (CF) fluorescent probe (Figure [1 ]A, Figure S5). After fluorophore attachment, the labelled parent peptide was not
completely stable so it was directly reacted with three different electrophile modulators,
namely glyceraldehyde (TGly , neutral hydrophilic, Figure S6), guanidinium aldehyde (TGua , cationic hydrophilic, Figure S7) and acetone (TAcet , hydrophobic, Figure S8).
After reaction with the aldehydes the three stable oxime-bonded peptides were incubated
with HeLa cells and the uptake was quantified after cell washings and trypsinization
to remove the excess of the membrane-bound peptides (Figure [3 ]A, see Supporting Information). Interestingly, the three peptides [CFP(TGly )2 , CFP(TAcet )2 , CFP(TGua )2 ] showed higher levels of internalization than the prototypical octaarginine penetrating
peptide CFR8 (Figure [3 ]A). Intriguingly, the incorporation of the cationic TGua lead to the highest level of internal fluorescence, which was in good agreement with
the DNA transport experiments in vesicles.
Figure 3 Cell transport experiments of the three oxime-modulated peptides CFP(T)2 and the control CFR8 . (A) Quantification of the uptake of the peptides by fluorometry. HeLa cells were
incubated for 30 min with the indicated fluorescein labeled peptides (4 μM) in HKR
buffer. Cells were then washed and trypsinized and the cellular uptake was quantified
by fluorometry. (B) Cell viability (MTT assay). Error bars indicate SD of three replicates.
(C) Fluorescence microscopy images, with differential interference contrast (DIC)
insets, of the three peptides and CFR8 in HeLa cells. Cells were thoroughly washed before imaging. Nuclei were stained blue
with Hoechst. [Peptide] = 4 μM in all cases.
Cell viability in the presence of the three oxime peptides was studied by following
the colorimetric 2,5-diphenyltetrazolium bromide reduction to formazan (MTT assay,
see Supporting Information). Interestingly, this assay confirmed that cationic enhancement
did not only improve the uptake but also the cellular viability (Figure [3 ]B). The peptide modified by oxime bond formation with acetone induced a reduction
of cell viability probably due to membrane disruption of the more hydrophobic peptide
(Figure [3 ]B).
The three peptides were also studied using inverted epifluorescence microscopy after
incubation at 4 μM with HeLa cells (Figure [3 ]C). The pictures confirmed a different behavior of the three peptides in cells (Figure
[3 ]C). The octaarginine control peptide (CFR8 ) showed the typical combination of cytoplasm distribution together with tiny punctate
fluorescence, indicating membrane translocation and energy-dependent endocytosis (Figure
[3 ]C).[10 ] The peptide modified with the neutral glyceraldehyde, CFP(TGly )2 , showed a higher level of bigger internal aggregates suggesting an increase in the
size of the transported particles. The peptide capped with acetone, CFP(TAcet )2 , showed a diffuse cytoplasmic distribution. A similar behavior, but to a greater
extent, was observed for the amphiphilic peptide modified with the cationic aldehyde
[CFP(TGua )2 ]. The clear enhancement of transport efficiency in vesicles and in cells confirmed
that the guanidinium derivative R1 P(TGua )2 was the leader of this small series.
To get further insight into the internalization mechanism and confirm the endosomal
pathway, we performed microscopy experiments in HeLa cells in the presence of LysoTracker
Deep Red (Figure [4 ]). These experiments confirmed that the punctate fluorescence was due to the presence
of endosomes for the octa-arginine (CFR8 ) and the glyceraldehyde oxime peptide [CFP(TGly )2 ; Figure [4 ]]. The membrane transport properties of penetrating peptides are highly dependent
on the concentration of the cellular incubation experiment.[1 ]
[4 ] Therefore, to avoid signal saturation and to check for the potential endocytosis
of the highly active peptides bonded with acetone and guanidinium aldehydes we carried
out a one half dilution of the initial concentrations. We observed that [CFP(TGua )2 ] showed a very similar pattern as the CFR8 and the fluorescence image, in the presence of LysoTracker, showed cytoplasm distribution
together with co-localization in the lysosomes (Figure [4 ]). The peptide capped with acetone did not show the similar co-localization, an observation
that could be related to the membrane lytic properties and a potential endosomal disruption
of the more hydrophobic CFP(TAcet )2 .
Figure 4 Lysosomal staining of cells incubated with the peptides. After three hours of incubation
with the fluorescein-labelled peptides (green), cells were stained with LysoTracker
Deep Red (red). Nuclei were counterstained with Hoechst (blue). Some co-localization
between the lysosomal marker and the four peptides is observed with particular high
intensity for the case of CFP(TGly )2 and CFR8 at 4 μM. For the high penetrating CFP(TAcet )2 and CFP(TGua )2 it was necessary to dilute the concentration to 2 μM to avoid signal saturation and
observe the colocalization.
Increasing cationic and hydrophobic character of penetrating peptides can lead to
toxic undesired effects.[1 ] However, in the particular example reported here, increasing the cationic character
by introducing more guanidinium aldehydes in a parent amphiphilic peptide strongly
increased the cellular uptake (Figure [3 ]B). This increase in uptake observed for the peptide [CFP(TGua )2 ] could suggest the importance of the hydrophobicity of the leucine moieties that
would enhance the membrane partition and the penetrating properties of any given peptide.
These results confirmed the important consequences in the transport activity after
covalent/dynamic modification of an initial parent penetrating peptide scaffold. Although
much more data will be required to validate this trend, the results reported here
suggest that the hydrophilicity/hydrophobicity ratio could be an important factor
to consider with implications in the supramolecular properties and in the transport
capabilities of any given peptide. These implications would not only be limited to
the level of delivery efficiency but also to the final mechanism of internalization.
With this paper we wanted to highlight the potential of dynamic bonds as powerful
synthetic tools to modulate the activity of penetrating peptides. In this article
we have focused on the synthetic advantage of the oxime bonds. However, in the future,
we hope that also the dynamic character and the response to external stimuli of dynamic
covalent chemistry will allow us to understand cell penetration and to develop novel
and superior intracellular delivery vehicles.
