Self-activation in de novo designed mimics of cell-penetrating peptides

Article abstract of DOI:10.1002/anie.201101535

DIY transduction: Cell-penetrating peptides (CPPs) efficiently cross the nonpolar biological membrane, but activation by hydrophobic counterions is essential for transport activity. Polymeric mimics of CPPs are coupled with hydrophobic side chains; the "self-activation" in these species is demonstrated by transport of the fluorescent probe CF out of bilayer vesicles. Copyright

Full text of DOI:10.1002/anie.201101535

DOI: 10.1002/anie.201101535
Peptide Mimics
Self-Activation in De Novo Designed Mimics of Cell-Penetrating
Peptides**
Abhigyan Som, A. Ozgul Tezgel, Gregory J. Gabriel, and Gregory N. Tew*
The unique ability of cell-penetrating peptides (CPPs), also
known as protein transduction domains, to navigate across the
nonpolar biological membrane has been under intense
investigation.[1] In vitro studies have shown that multiple
mechanisms are available, with the precise details being
dependent on the peptide and cell line studied. The several
clearly demonstrated pathways include various forms of
endocytosis,[2] macropinocytosis,[2e] lipid-raft-dependent mac-
ropinocytosis,[3] and protein-dependent translocation.[4] In
addition, an energy-independent pathway, or spontaneous
translocation, has also been illustrated.[1i,5]
Perhaps the clearest example of an energy-independent
pathway is the ability of CPPs, and their synthetic mimics, to
cross model phospholipid bilayer vesicle membranes.[1g,h,6]
Scheme 1. Guanidino copolymers G1–G12.
General consensus in the literature suggests that hydrophobic
counterions play an essential role in this transduction by
complexation around the guanidinium-rich backbone, thus
ꢀcoatingꢁ the highly cationic structure with lipophilic moieties.
For example, an octamer of arginine in the presence of
sodium laurate partitioned into octanol versus water with
better than 95% efficiency.[7] Separately, it was shown that the
simple peptide nonaarginine ((Arg)9) does not in fact trans-
verse membranes very effectively on its own.[1h] However, the
presence of hydrophobic counterions “activates” this mole-
cule, thus turning it into a potent transduction peptide. It was
shown that n-alkyl chain surfactants were good “activators”
and thus efficient at promoting the transport of oligo- and
polyarginines across biological membranes.[1g,i,8]
exhibited significantly enhanced activity (by three orders of
magnitude) without the need for any “counterion activator”.
Monomers were prepared by either Mitsunobu coupling
or nucleophilic substitution reactions (see the Supporting
Information).
Random
copolymers
G1–G12
with
50:50 mol% monomer distribution were targeted at two
molecular weights (Mn) using ring-opening metathesis poly-
merization (ROMP; low Mn 2.9–3.9 kDa and high Mn 11.4–
13.6 kDa of the tert-butyloxycarbonyl (Boc)-protected poly-
mers were obtained). Gel-permeation chromatography gave
monomodal signals and narrow molecular-weight indices
(1.05–1.15). The Boc-protected polymers were deprotected
to obtain G1–G12, and their activities were studied in vesicle
assays.
After the initial discovery that CPP-like behavior could be
emulated in simple norbornene-based polymers,[6b,9] we
wondered if the presence of covalently attached hydrophobic
residues would increase their translocation activity. To
evaluate this hypothesis, we designed and synthesized a
series of norbornene-based guanidine-rich polymers, where
the hydrophobic groups were introduced through a side chain
rather than as counterions (Scheme 1). Remarkably, the
guanidine polymers containing certain alkyl side chains
Using the standard biophysical assay well-accepted in the
CPP literature, the transport activities of G1–G12 were
determined.[6b] Specifically, 5(6)-carboxyfluorescein (CF) was
used as a fluorescent probe in egg yolk phosphatidylcholine
large unilamellar vesicles (EYPC-LUVs). The activity of G1–
G12 transporters increased with increasing polymer content
at a constant vesicle concentration as detected by CF emission
intensity, yielding plots of fluorescence intensity versus
polymer concentration for the series G1–G12 (Supporting
Information, Figure S1). Fitting the Hill equation (Y/
(c/EC50)n) to this data for each individual polymer revealed
a nonlinear dependence of the fractional fluorescence inten-
sity Y on the polymer concentration c. This analysis gave Ymax
(maximal CF release relative to complete release by Triton X-
100), EC50 (effective polymer concentration needed to reach
Ymax/2), and the Hill coefficient n (Supporting Information,
Figure S2, Tables S1 and S2). For direct comparison, it is
worth mentioning that the CPPs heptaarginine and polyargi-
nine were inactive under these conditions; it is known that
polyarginine needs counterions for activation.[1g,h,8]
[*] Dr. A. Som, A. O. Tezgel, Dr. G. J. Gabriel, Prof. G. N. Tew
Polymer Science & Engineering Department
University of Massachusetts
120 Governors Drive, Amherst, MA 01003 (USA)
Fax: (+1)413-545-0082
E-mail: tew@mail.pse.umass.edu
Homepage: http://www.pse.umass.edu/gtew/index.html
[**] We thank the NSF (CHE-0910963) for financial support.
Supporting information for this article, including experimental
details, is available on the WWW under http://dx.doi.org/10.1002/
anie.201101535.
Figure 1 collects the EC50 values for this series of
copolymers. Polymers with lower EC50 values are said to be
Angew. Chem. Int. Ed. 2011, 50, 6147 –6150
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6147

Communications
Figure 2. Hill plot for copolymers G1 and G4 in EYPC/EYPGꢁCF
vesicles with fits to the Hill equation (circles). G1 EC50 =0.4 mm, G4
EC50 =0.04 mm. G1 and G4 remained inactive against EYPC/EYPGꢁ
calcein vesicles (triangles).
Figure 1. Effective concentrations (EC50) of low-molecular-weight and
high-molecular-weight copolymers G1–G12 as a function of alkyl side-
chain length.
“more active”, because the concentration needed to reach
50% activity is lower. In Figure 1, EC50 values are plotted
against the alkyl chain length of the copolymers G1–G12 with
low and high molecular weights. Two trends are evident from
this data. First, within both molecular-weight series, a
comparison of the EC50 values shows that the activity of the
copolymers increases as the length of the hydrophobic side
chain is increased up to butyl (G4). For longer side chains, the
activity decreases.
used, no fluorescence increase was observed (see triangles in
Figure 2). These experiments strongly support the hypothesis
that these new polymers exhibit transduction activity and that
they are “self-activated” by the presence of the alkyl
substituent.
Even further support for transduction comes from reports
of similar studies conducted on CPPs. These studies similarly
investigated calcein release for classical CPPs, including
R8,[11] R9,[12] and TAT[11,12] in various lipid systems. At very
low peptide-to-lipid (P/L) ratios of 0.05, R9 exhibited 7%
leakage from EYPCꢁcalcein vesicles and was inactive against
EYPC/EYPGꢁcalcein vesicles.[12] TAT48–60 showed 15 and
2% leakage from EYPCꢁcalcein and EYPC/EYPGꢁcalcein
vesicles, respectively.[12] Various P/L ratios were not reported.
Similarly, the ability of R8 and TAT48–61 to induce leakage of
DMPC/DMPGꢁcalcein (DMPC = dimyristoyl phosphatidyl-
choline; DMPG = dimyristoyl phosphatidylglycerol) vesicles
was examined as a function of the P/L ratio.[11] Consistent with
the previous findings,[12] at low P/L ratios, little to no leakage
was observed; however, at P/L = 1.2, greater than 10%
leakage was observed (R8 ca. 18% and TAT48–61 ca.
10%).[11] As shown in Figure 2, G1 and G4 induced no
change in calcein emission, despite very high P/L (here:
polymer-to-lipid) ratios above 20 (Supporting Information,
Figure S5). These experiments clearly demonstrate that the
novel polymers reported herein are able to induce increases in
CF emission but not in calcein emission (even at very high P/L
ratios), completely consistent with the numerous reports on
CPPꢁCF transduction.
Although it is not entirely clear why the more hydro-
phobic side chains are less active, it is likely that aggregation
of these relatively nonpolar polymers plays some role, as G9
and G12 are significantly less soluble than G1–G5. This
hypothesis is also supported by the Ymax values for G9 and
G12, which are significantly smaller than those for copoly-
mers G1–G5 (Supporting Information, Tables S1 and S2).
The second trend is that higher-molecular-weight samples
are more active across the entire series, in agreement with the
previously observed “polymer effect”.[6b] For example, G1 has
an EC50 = (20.0 ꢀ 0.9) mm and (6.4 ꢀ 0.2) mm for the low- and
high-molecular-weight samples, respectively. Similarly, G4
has EC50 values of (0.20 ꢀ 0.06) mm and (0.0030 ꢀ 0.0005) mm.
In all cases, the Hill coefficient generally ranged between n =
1 and n = 3, implying poor cooperativity, which supports
transduction[6b] and no requirement for multichain structures
being involved in the transport activity. These results strongly
support the proposed hypothesis that the presence of hydro-
phobic side chains can be used for “self-activation”. At the
same time, this strong support assumes the mechanism of
action is transduction and not some type of general pore
formation. To investigate this aspect further, G1 and G4 were
evaluated against EYPC/EYPG (EYPG = egg yolk phospha-
tidylglycerol) vesicles containing either CF or calcein. Cal-
cein-loaded vesicles are routinely used to demonstrate pore
formation induced by antimicrobial peptides and their
synthetic mimics.[10]
To further explore the molecular design of these CPP-like
polymers, we designed and synthesized another series of
polymers (G1’ and G4’, Scheme 2). Unlike the statistically
random copolymers G1–G12, these new homopolymers
contain a precise sequence of guanidinium and alkyl side
chains on every repeat unit. The monomers for G1’ and G4’
were synthesized in three steps and polymerized by ROMP
(Supporting Information, Scheme S6). G4’, with an EC50
value of (0.0010 ꢀ 0.0004) mm, exhibited three orders of
magnitude better activity than G1’ (EC50 = (1.3 ꢀ 0.2) mm),
Figure 2 shows that both G1 and G4 induced nonlinear
increases in the fractional fluorescence from EYPC/
EYPGꢁCF vesicles as a function of concentration in a
manner similar to that discussed previously. However, and in
sharp contrast, when EYPC/EYPGꢁcalcein vesicles were
which is similar to the trend observed for G4 (EC50
=
6148
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 6147 –6150

classical CPPs is further evidence that these novel polymers
can replicate the essential biochemical features of CPPs.
This study reports new cost-effective, highly efficient
molecular transporters. Biophysical studies clearly demon-
strate CPP-transport-like activity as opposed to general pore
formation for these polymeric PTDMs. There is at least a
three-fold activity enhancement upon moving from methyl to
butyl side chains. Furthermore, we observed that the location
of the alkyl substituents along the polymer backbone did not
influence transport activity, as both designs (random copoly-
mer or homopolymer) yielded similar EC50 values. This study
also shows that the complex-
Scheme 2. Guanidino homopolymers G1’ and G4’.
ation between the guanidi-
nium and anionic functional-
ity of the “activators” does
not play a central role, as in
the studies presented herein
we did not “neutralize” the
guanidinium cationic charge
but simply added “lipophilic-
ity” to the structures. We
continue to design new syn-
thetic polyguanidines to
expand the structure–activity
relationships. At the same
time, it will be critical to see
how these synthetic mimics
compare to natural peptide
Figure 3. Hill plot of a) copolymers G1, G4 and b) homopolymers G1’, G4’ in EYPCꢁCF vesicles with fits to
sequences in cellular uptake
studies for the improved
delivery of cargo into cells.
the Hill equation.
Received: March 2, 2011
Published online: May 17, 2011
(0.0030 ꢀ 0.0005) mm) and G1 (EC50 = (6.4 ꢀ 0.2) mm; all
shown in Figure 3). These data confirm the findings of “self-
activation” reported in Figure 1 and Figure 2. Moreover, the
Keywords: cell-penetrating peptides · peptidomimetics ·
polymers · vesicles · transduction
precise sequence of these homopolymers coupled with the
similar behavior of the copolymers implies that the exact
molecular sequence along the backbone is not critical for
transport activity as measured by these assays.
.
[1] a) I. Nakase, T. Takeuchi, G. Tanaka, S. Futaki, Adv. Drug
Delivery Rev. 2008, 60, 598 – 607; b) P. A. Wender, W. C.
Galliher, E. A. Goun, L. R. Jones, T. H. Pillow, Adv. Drug
Delivery Rev. 2008, 60, 452 – 472; c) A. Ziegler, Adv. Drug
Delivery Rev. 2008, 60, 580 – 597; d) A. Pantos, I. Tsogas, C. A.
Paleos, Biochim. Biophys. Acta Biomembr. 2008, 1778, 811 – 823;
e) B. A. Smith, D. S. Daniels, A. E. Coplin, G. E. Jordan, L. M.
McGregor, A. Schepartz, J. Am. Chem. Soc. 2008, 130, 2948 –
2949; f) S. M. Fuchs, R. T. Raines, ACS Chem. Biol. 2007, 2, 167 –
170; g) N. Sakai, S. Futaki, S. Matile, Soft Matter 2006, 2, 636 –
641; h) N. Sakai, S. Matile, J. Am. Chem. Soc. 2003, 125, 14348 –
14356; i) T. Takeuchi, M. Kosuge, A. Tadokoro, Y. Sugiura, M.
Nishi, M. Kawata, N. Sakai, S. Matile, S. Futaki, ACS Chem. Biol.
2006, 1, 299 – 303.
[2] a) S. M. Fuchs, R. T. Raines, Biochemistry 2004, 43, 2438 – 2444;
b) G. Drin, S. Cottin, E. Blanc, A. R. Rees, J. Temsamani, J. Biol.
Chem. 2003, 278, 31192 – 31201; c) J. P. Richard, K. Melikov, E.
Vives, C. Ramos, B. Verbeure, M. J. Gait, L. V. Chernomordik,
B. Lebleu, J. Biol. Chem. 2003, 278, 585 – 590; d) T. B. Potocky,
A. K. Menon, S. H. Gellman, J. Biol. Chem. 2003, 278, 50188 –
50194; e) I. A. Khalil, K. Kogure, S. Futaki, H. Harashima, J.
Biol. Chem. 2006, 281, 3544 – 3551.
We have established that “self-activation” is possible for
these polymeric mimics of CPPs, or protein transduction
domain mimics (PTDMs). In addition, the butyl side chain
appears to optimize transport activity, with the maximum
degree of self-activation being achieved for copolymer G4
and homopolymers G4’. Upon increasing or decreasing the
length of the side chain, transport activity decreased. The
closest arginine-rich CPP analogues of low- and high-molec-
ular-weight PTDMs, R9 and polyarginine (pR), remained
inactive across the concentration range in which G1 and G4
exhibited concentration-dependent activity (Supporting
Information, Figure S6). In addition, the hydrophobic coun-
teranion pyrenebutyrate (PB) efficiently “activated” G1 with
a low EC50 (3 mm) compared to pR–PB activation (EC50
=
40 mm; Supporting Information, Figure S7a). ꢀInactivationꢁ
was also observed at a fixed concentration of G1 (at 15 mm,
Y= 0.8) with the commonly used hydrophilic counteranion
Cascade Blue pyruvate with an IC50 of 12.5 mm (Supporting
Information, Figure S7b). The fact that this PTDM demon-
strated activation and inactivation properties similar to
[3] J. S. Wadia, R. V. Stan, S. F. Dowdy, Nat. Med. 2004, 10, 310 – 315.
Angew. Chem. Int. Ed. 2011, 50, 6147 –6150
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 6149

Communications
[4] M. Silhol, M. Tyagi, M. Giacca, B. Lebleu, E. Vives, Eur. J.
[9] E. M. Kolonko, L. L. Kiessling, J. Am. Chem. Soc. 2008, 130,
5626 – 5627.
Biochem. 2002, 269, 494 – 501.
[5] a) T. Suzuki, S. Futaki, M. Niwa, S. Tanaka, K. Ueda, Y. Sugiura,
J. Biol. Chem. 2002, 277, 2437 – 2443; b) P. E. G. Thoren, D.
Persson, P. Isakson, M. Goksor, A. Onfelt, B. Norden, Biochem.
Biophys. Res. Commun. 2003, 307, 100 – 107; c) J. L. Zaro, W. C.
Shen, Exp. Cell Res. 2005, 307, 164 – 173.
[10] a) A. Som, G. N. Tew, J. Phys. Chem. B 2008, 112, 3495 – 3502;
b) K. Matsuzaki, S. Yoneyama, K. Miyajima, Biophys. J. 1997,
73, 831 – 838; c) K. Matsuzaki, M. Harada, T. Handa, S.
Funakoshi, N. Fujii, H. Yajima, K. Miyajima, Biochim. Biophys.
Acta Biomembr. 1989, 981, 130 – 134.
[6] a) N. Sakai, T. Takeuchi, S. Futaki, S. Matile, ChemBioChem
2005, 6, 114 – 122; b) A. Hennig, G. J. Gabriel, G. N. Tew, S.
Matile, J. Am. Chem. Soc. 2008, 130, 10338 – 10344.
[7] J. B. Rothbard, T. C. Jessop, R. S. Lewis, B. A. Murray, P. A.
Wender, J. Am. Chem. Soc. 2004, 126, 9506 – 9507.
[8] F. Perret, M. Nishihara, T. Takeuchi, S. Futaki, A. N. Lazar,
A. W. Coleman, N. Sakai, S. Matile, J. Am. Chem. Soc. 2005, 127,
1114 – 1115.
[11] P. Ruzza, B. Biondi, A. Marchiani, N. Antolini, A. Calderan,
Pharmaceuticals 2010, 3, 1045 – 1062.
[12] P. Guterstam, F. Madani, H. Hirose, T. Takeuchi, S. Futaki, S.
El Andaloussi, A. Graslund, U. Langel, Biochim. Biophys. Acta
Biomembr. 2009, 1788, 2509 – 2517.
6150
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 6147 –6150