Protein transduction domain mimics: The role of aromatic functionality

Article abstract of DOI:10.1002/anie.201104624

For better or worse: Protein transduction domain mimics built from synthetic polymers demonstrate that aromatic side chains provide better transduction than aliphatic groups at the same relative hydrophobicity. Similarly, a less hydrophobic aromatic side chain is more active than the corresponding aliphatic one containing the same number of carbon atoms (see picture). Copyright

Full text of DOI:10.1002/anie.201104624

.
Angewandte
Communications
DOI: 10.1002/anie.201104624
Peptidomimetics
Protein Transduction Domain Mimics: The Role of Aromatic
Functionality**
Abhigyan Som, Anika Reuter, and Gregory N. Tew*
Cell-penetrating peptides (CPPs), or protein transduction
domains (PTDs), are a special class of membrane-active
proteins that can cross the cell membrane with unusual
efficiency.[1] They have attracted considerable attention
because of their ability to readily cross biological membranes,
in spite of their highly charged nature.[2] While the exact
mechanism of this transport remains under intense investiga-
tion, energy-independent pathways are known.[2a,3] Perhaps
the clearest example is the ability of CPPs, and their synthetic
mimics, to cross model phospholipid bilayer vesicle mem-
branes.[4] One suggested mechanism implies that, in fact,
CPPs like polyarginine (pR) need assistance to cross the
membrane.[1d,4a–c,5] It suggests that hydrophobic counterions
complex around the guanidinium-rich backbone, thus “coat-
ing” the highly cationic structure with lipophilic moieties. This
process has been termed “activation”, in which the lipophilic
anion acts as an activator. In a series of detailed studies it was
shown that aromatic activators outperform aliphatic
ones.[4a–c,5] For example, sodium 4-(pyren-1-yl)butane-1-sul-
fonate gave an EC50 (effective concentration to obtain 50%
activity) of 6.7 mm whereas the value for sodium dodecane-1-
sulfonate was 16 mm.[5] Among other activators studied, the
larger aromatic counterion, coronene, was not better than
pyrene; however, a fullerene analogue was surprisingly
effective.[5] While this work beautifully demonstrated the
role of various counterions for pR activation, it was not clear
if this better activation was due to general hydrophobicity or
to the aromatic nature of these activators.
rules out a dominant effect of hydrogen bonding.[7] It was
suggested that the flat-rigid shape, p-electronic structure, and
associated quadrupolar moments provide unique and highly
favorable interactions with the bilayer interface.[6b] Specific
interactions that have been proposed include p-cation,
electrostatic, dipole–dipole, and entropic factors related to
bilayer perturbation.[6,7,9] Even HIV-TAT, the original protein
that initiated the field of small PTDs, requires tryptophan
(Trp11) for translocation.[10] Moreover, an oligoarginine con-
sisting of seven arginine residues with a C-terminal trypto-
phan (R7W) and a TAT48–60 peptide with residue 59 substi-
tuted with a tryptophan (TAT48–60P59W) exhibit cellular
internalization through energy-independent pathways.[3b,11]
Another classical CPP, penetratin (Pen), contains two trypto-
phan residues. Substitution of tryptophan by phenylalanine
(Pen2W2F) did not significantly impact cell uptake.[11] Among
the aromatic amino acids, phenylalanine has the unique
ability to partition at the interface and in the membrane
core.[9f] In fact, aromatic residues, especially phenylalanine,
are most effective at anchoring proteins in the membrane due
to their “special ability” to form and stabilize essential
interactions with the polar elements of the bilayer.[12] As a
result, aromatic functionality could be a critical element
facilitating the interactions between CPPs and the bilayer
during transduction.
In the past few years, we and others have reported
polymers designed to mimic the transduction activity of
PTDs.[4d,13] More recently, we demonstrated that these protein
transduction domain mimics (PTDMs) have “self-activation”
properties when hydrophobic alkyl side chains were built into
the copolymers.[14] Here, a new series of PTDMs was designed
to determine if an aromatic functionality provides better
transduction efficiency than aliphatic ones, at the same
relative hydrophobicity. Given the importance of aromatic
amino acids in membrane proteins and their interactions with
the bilayer, it was proposed that aromatic side chains would
make better activators, given equal relative hydrophobicity.
Although aromatic groups have been studied in peptide-
based CPPs,[3b,11,15] demonstration of the importance of
aromatic functionality in these synthetic analogues is critical
to establishing them as appropriate mimics, or PTDMs. By
using reversed-phase HPLC to determine side-chain hydro-
phobicity and EC50 values in a classic transduction experi-
ment, it is demonstrated here that it was possible to differ-
entiate between side-chain hydrophobicity and aromaticity.
As shown in Table 1, a series of new PTDM polymers was
prepared by ring-opening metathesis polymerization (see the
Supporting Information for detailed synthesis and character-
ization of monomers and polymers). Reversed-phase HPLC,
commonly used to evaluate relative hydrophobicity,[16] was
There is good reason to think that aromatic functional
groups may play a special role, beyond their general hydro-
phobicity. It is well recognized that membrane proteins are
enriched in aromatic amino acids at the membrane surface.[6]
Their central hydrophobic core, composed mostly of aliphatic
residues, is flanked on both sides by “aromatic belts”.[6a,7]
Although this belt is predominantly composed of tryptophan
and tyrosine, as opposed to phenylalanine, it was shown that
aromatic residues, including N-methylindole, have favorable
free energies of insertion into the bilayer interface.[6b,8] This
[*] Dr. A. Som, A. Reuter, Prof. G. N. Tew
Polymer Science & Engineering Department
University of Massachusetts
120 Governors Drive, Amherst, MA 01003 (USA)
E-mail: tew@mail.pse.umass.edu
Homepage: http://www.pse.umass.edu/gtew/index.html
[**] Generous support was primarily provided by the NSF (CHE-
0910963). Dr. Semra Colak is acknowledged for her contribution in
octyl monomer synthesis.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201104624.
980
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 980 –983

Angewandte
Chemie
Table 1: Oxanorbornene-derived guanidino copolymers.[a]
phobicities and therefore enables the deconvolution of
hydrophobicity and aromaticity in transduction activity.
Transport activities for these novel PTDMs were deter-
mined using the standard biophysical assay that is well
documented in the CPP literature.[2f,4d,5,14] Specifically, 5(6)-
carboxyfluorescein (CF) was used as a fluorescent probe in
egg yolk phosphatidylcholine large unilamellar vesicles
(EYPC-LUVs). The activity of these transporters increased
with increasing polymer concentration at a constant vesicle
concentration as detected by CF emission intensity, thereby
yielding plots of fluorescence intensity versus polymer con-
centration (Supporting Information, Figures S1 and S2).
Fitting the Hill equation [Y/ (c/EC50)n] to these data for
each individual polymer revealed a nonlinear dependence of
the fractional fluorescence intensity Y on the polymer
concentration c, which is classical behavior demonstrated by
CPPs.[4a–d,5,14] 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 (see Table 2). For direct comparison it is worth
mentioning that the CPP polyarginine hydrochloride was
inactive under these conditions, a known fact since pR needs
counterions for activation.
Polymer
GOc
R
DP
40
Mn
Mw/Mn
14900
1.08
GCy
GPh
34
30
12200
10700
1.04
1.05
GNp
GPy
30
27
11600
11100
1.04
1.07
[a] Degree of polymerization (DP), apparent molecular weight (Mn), and
polydispersity index (Mw/Mn) were calculated from GPC. The repeat units
are Cy cyclohexyl, G guanidino, Np naphthyl, Oc octyl, Ph phenyl, and Py
pyrenyl.
Table 2: EC50, Ymax, and Hill coefficient n of the copolymers’ transduction
activity.
Polymer[a]
EC50 [nm][b]
Ymax
n
[c]
performed on each nonpolar monomer. Using a C8 column in
100% acetonitrile (isocratic), the chromatograms of all five
nonpolar monomers were obtained as shown in Figure 1.
GOc (50:50)
GCy (50:50)
GPh (50:50)
GNp (50:50)
GNp (80:20)
GNp (96:04)
GPy (50:50)
11.4ꢀ2.8
9.7ꢀ0.9
4.3ꢀ0.1
3.8ꢀ0.6
7.8ꢀ1.8
73.0ꢀ0.9
6.1ꢀ0.2
0.80ꢀ0.02
0.80ꢀ0.03
0.96ꢀ0.01
0.84ꢀ0.04
0.88ꢀ0.02
0.91ꢀ0.03
0.81ꢀ0.01
1.7ꢀ0.05
2.6ꢀ1.0
1.1ꢀ0.1
1.4ꢀ0.2
1.2ꢀ0.2
0.9ꢀ0.1
2.8ꢀ0.3
[a] The molar ratios between guanidino (G) repeat units and the
hydrophobic repeat units are given in parentheses. [b] EC50: effective
polymer concentration needed to reach Ymax/2. [c] Ymax: maximal CF
release relative to complete release by Triton X-100. Each data point was
collected in three independent experiments.
Figure 2a is a plot of 1/EC50 versus 1/Rt for GOc, GCy,
GPh, GNp, and GPy. The data were plotted in this way to give
the most efficient transporter the highest value as it relates to
effective concentration. Since lower EC50 values are said to be
more active, 1/EC50 directly provides the largest value for the
best transporter. Similarly, it would be ideal to limit the
hydrophobicity of the transporters while maintaining efficient
transport activity; thus, 1/Rt was plotted since the retention
time is larger for more hydrophobic monomers. Figure 2a
shows that while GOc, GCy, and GPy have similar 1/Rt values,
GPy is a more effective transporter (higher 1/EC50). In fact, it
is approximately 1.5 to 2.0 times more active than GOc or
GCy, despite the similar relative hydrophobicities of their
corresponding nonpolar monomers. This activity difference is
similar to that previously reported for pyrene (EC50, 6.7 and
9.3 mm) versus alkyl activators (EC50, 16 and 19 mm),[5a] which
suggests that aromatic functionality may indeed have a special
role in PTD(M) transduction.
Figure 1. Retention time (Rt) on a reversed-phase C8 HPLC column
(under isocratic conditions, 100% acetonitrile) of the corresponding
hydrophobic monomers that were copolymerized with the guanidine
monomers. In, normalized intensity. Individual Rt [min] of the mono-
mers: Ph 4.15, Np 4.27, Oc 4.50, Cy 4.55, Py 4.57.
Pyrene has been commonly used as an activator of pR,[2a,4a–c,5]
so its retention time (Rt) was of particular interest. In this
study, its Rt was 4.57 min while the aliphatic monomers
containing eight carbon atoms yielded similar Rt values of
4.55 (Cy) and 4.50 min (Oc). As a result, these three
monomers have similar relative hydrophobicities. In contrast,
the other two aromatic monomers, Np and Ph, are less
hydrophobic with Rt values of 4.27 and 4.15 min, respectively.
This series of monomers spans a range of relative hydro-
Angew. Chem. Int. Ed. 2012, 51, 980 –983
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
981

.
Angewandte
Communications
Figure 3. Hill plot of GNp copolymers with different G to Np repeat
unit ratios (50:50, 80:20, and 96:4) in EYPCꢁCF vesicles with fit to the
Hill equation.
decreasing molar content of Np, which suggests that no
threshold was present. These data indicate that when more
Np is present in the polymer it is more effective at trans-
duction, although there is most likely an upper limit, at least
because of the solubility of the polymer.
Table 2 summarizes the Hill parameters for these poly-
mers and shows that they all have similar Ymax values and Hill
coefficients n (around 2), which suggests poor cooperativity
and supporting transduction.[4d] However, no detailed under-
standing of the mechanism is available at this time, and other
factors such as polymer conformation and aggregation cannot
be conclusively ruled out. To compare “activators” of varying
EC50 values and total fractional transport activity, the
activator efficiency E was calculated based on the exponential
relationship between Ymax and EC50.[5a] The same arbitrary
scaling factor that was previously used to calibrate E between
0 and 10[5a] was also used here to determine E values for these
covalently activated PTDMs (see the Supporting Informa-
tion, Table S3). For GPh, E was found to be 25 or 2.5 times
larger than the value for the highly active fullerene analogue
and five times better than for pyrene butyrate. These covalent
PTDMs have both low EC50 and high Ymax values, features
previously suggested for the perfect activator.[5a] This is
markedly different from the supramolecular activators in
which more potent activators (lowest EC50 values) also had
low Ymax values.[5] The fact that these covalently activated
PTDMs are more effective than the supramolecular ana-
logues (pR–activator) is not necessarily surprising, since
covalent attachment eliminates the binding equilibrium
between pR and the activator. A previously reported
homopolymer of the guanidine monomer was activated with
pyrene butyrate with an EC50 of 70 mm,[4d] whereas in this
study a very low EC50 value (6.1 nm) was achieved with GPy.
The best activators most likely also have solubility limitations
since they are significantly hydrophobic. At the same time,
the ability to design PTDMs that are significantly more active
than classical CPPs is extremely encouraging.
Figure 2. a) Plot of 1/EC50 (for the PTDM copolymer) versus 1/Rt (for
the corresponding monomers) for GOc, GCy, GPh, GNp, and GPy.
Data represent mean ꢀ standard deviation (s.d.) from three independ-
ent experiments. b) Concentration (c)-dependent activity of copolymers
*
^
*
^
GOc ( ), GCy ( ), GPh ( ), and pR ( ) in EYPCꢁCF vesicles with fit
to the Hill equation. Transmembrane activity Y defines the fractional
fluorescence intensity at 800 s.
Further support for this hypothesis comes from comparing
the values (EC50 and hydrophobicity) for GPh to the others in
Figure 2a. GPh is the least hydrophobic (larger 1/Rt) yet it is
the most active (higher 1/EC50). This is consistent with
phenylalanineꢀs unique ability to partition at the membrane
interface and in the membrane core.[9f] Although GNp is more
hydrophobic than GPh, they show similar activities within the
1/EC50 error level. Figure 2b shows the Hill plots for GOc,
GCy, and GPh which yields their respective EC50 values of
(11.4 ꢀ 2.8), (9.7 ꢀ 0.9), and (4.3 ꢀ 0.1) nm. This comparison is
particularly interesting since all three nonpolar monomers
contain a total of eight carbon atoms. In addition, both GPh
and GCy contain cyclic rings and, in fact, represent the closest
possible structural analogues. While the aromatic group was
expected to be less hydrophobic,[17] it clearly demonstrates
that transduction activity is not solely dominated by hydro-
phobicity, but rather that aromaticity plays a crucial role. It
also shows that the large pyrene ring is not essential and that
smaller, more protein-like aromatic groups can effectively
promote transduction in these PTDMs.
To further examine the role of aromatic size on trans-
duction activity for this system, copolymers containing Np
were prepared. The 50:50 copolymer provided a similar EC50
value ((3.8 ꢀ 0.6) nm, see Figure 3 and Table 2) to those of the
other aromatic-containing polymers. Given the similarity in
values among all three aromatic-containing polymers, the
molar content of Np was lowered to understand whether or
not a “threshold” of aromatic content was needed for activity.
As Figure 3 shows, the activity of GNp decreased with
Using HPLC to determine the relative hydrophobicity of
various side chains, it was possible to demonstrate the
improved transport activity of aromatic functionality. This
provides guidance for building molecules that more favorably
interact with the membrane while reducing the overall
hydrophobicity. Understanding the broader goals of how
macromolecules (synthetic or natural) interact with the
982
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 980 –983

Angewandte
Chemie
[5] a) M. Nishihara, F. Perret, T. Takeuchi, S. Futaki, A. N. Lazar,
biological membrane is critically important. At the same time,
learning to program synthetic polymers with natural protein-
like activity remains an incredibly important task of modern
macromolecular chemistry. Many fundamental questions
remain but these new synthetic PTDMs appear to be useful
tools for studying macromolecule–membrane interactions.
A. W. Coleman, N. Sakai, S. Matile, Org. Biomol. Chem. 2005, 3,
1659 – 1669; b) 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.
[6] a) J. A. Killian, G. von Heijne, Trends Biochem. Sci. 2000, 25,
429 – 434; b) S. H. White, W. C. Wimley, Annu. Rev. Biophys.
Biomol. Struct. 1999, 28, 319 – 365.
Received: July 5, 2011
Revised: September 6, 2011
Published online: December 13, 2011
[7] W. M. Yau, W. C. Wimley, K. Gawrisch, S. H. White, Biochem-
istry 1998, 37, 14713 – 14718.
[8] W. C. Wimley, S. H. White, Nat. Struct. Biol. 1996, 3, 842 – 848.
[9] a) A. Arbuzova, L. B. Wang, J. Y. Wang, G. Hangyas-Mihalyne,
D. Murray, B. Honig, S. McLaughlin, Biochemistry 2000, 39,
10330 – 10339; b) M. M. Gromiha, Biophys. Chem. 2003, 103,
251 – 258; c) M. M. Gromiha, M. Suwa, Int. J. Biol. Macromol.
2005, 35, 55 – 62; d) W. J. Haynes, X. L. Zhou, Z. W. Su, S. H.
Loukin, Y. Saimi, C. Kung, FEBS Lett. 2008, 582, 1514 – 1518;
e) S. Mall, R. Broadbridge, R. P. Sharma, A. G. Lee, J. M. East,
Biochemistry 2000, 39, 2071 – 2078; f) D. Sengupta, J. C. Smith,
G. M. Ullmann, Biochim. Biophys. Acta Biomembr. 2008, 1778,
2234 – 2243; g) E. Strandberg, S. Morein, D. T. S. Rijkers,
R. M. J. Liskamp, P. C. A. van der Wel, J. A. Killian, Biochem-
istry 2002, 41, 7190 – 7198; h) P. Braun, G. von Heijne, Biochem-
istry 1999, 38, 9778 – 9782.
[10] H. Yezid, K. Konate, S. Debaisieux, A. Bonhoure, B. Beaumelle,
J. Biol. Chem. 2009, 284, 22736 – 22746.
[11] C. E. B. Caesar, E. K. Esbjorner, P. Lincoln, B. Norden, Bio-
chemistry 2006, 45, 7682 – 7692.
[12] C. T. K. Yuen, A. R. Davidson, C. M. Deber, Biochemistry 2000,
39, 16155 – 16162.
[13] a) C. B. Cooley, B. M. Trantow, F. Nederberg, M. K. Kiesewetter,
J. L. Hedrick, R. M. Waymouth, P. A. Wender, J. Am. Chem. Soc.
2009, 131, 16401 – 16403; b) E. M. Kolonko, J. K. Pontrello, S. L.
Mangold, L. L. Kiessling, J. Am. Chem. Soc. 2009, 131, 7327 –
7333; c) E. M. Kolonko, L. L. Kiessling, J. Am. Chem. Soc. 2008,
130, 5626 – 5627.
[14] A. Som, A. O. Tezgel, G. J. Gabriel, G. N. Tew, Angew. Chem.
2011, 123, 6271 – 6274; Angew. Chem. Int. Ed. 2011, 50, 6147 –
6150.
[15] a) K. L. Horton, K. M. Stewart, S. B. Fonseca, Q. Guo, S. O.
Kelley, Chem. Biol. 2008, 15, 375 – 382; b) J. R. Maiolo, M.
Ferrer, E. A. Ottinger, Biochim. Biophys. Acta Biomembr. 2005,
1712, 161 – 172.
[16] a) H. D. Thaker, F. Sgolastra, D. Clements, R. W. Scott, G. N.
Tew, J. Med. Chem. 2011, 54, 2241 – 2254; b) G. J. Gabriel, J. G.
Pool, A. Som, J. M. Dabkowski, E. B. Coughlin, M. Muthukur-
nar, G. N. Tew, Langmuir 2008, 24, 12489 – 12495.
Keywords: aromaticity · cell-penetrating peptides ·
.
hydrophobic effect · peptidomimetics · transduction
[1] a) 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; b) I. Nakase, T. Takeuchi, G.
Tanaka, S. Futaki, Adv. Drug Delivery Rev. 2008, 60, 598 – 607;
c) S. Futaki, Adv. Drug Delivery Rev. 2005, 57, 547 – 558; d) J. B.
Rothbard, T. C. Jessop, R. S. Lewis, B. A. Murray, P. A. Wender,
J. Am. Chem. Soc. 2004, 126, 9506 – 9507; e) P. A. Wender, D. J.
Mitchell, K. Pattabiraman, E. T. Pelkey, L. Steinman, J. B.
Rothbard, Proc. Natl. Acad. Sci. USA 2000, 97, 13003 – 13008;
f) P. S. Kabouridis, Trends Biotechnol. 2003, 21, 498 – 503.
[2] a) 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; b) T. Hitz, R. Iten, J. Gardiner, K. Namoto, P. Walde,
D. Seebach, Biochemistry 2006, 45, 5817 – 5829; c) P. A. Wender,
W. C. Galliher, E. A. Goun, L. R. Jones, T. H. Pillow, Adv. Drug
Delivery Rev. 2008, 60, 452 – 472; d) E. A. Goun, T. H. Pillow,
L. R. Jones, J. B. Rothbard, P. A. Wender, ChemBioChem 2006,
7, 1497 – 1515; e) K. Melikov, L. Chernomordik, Cell. Mol. Life
Sci. 2005, 62, 2739 – 2749; f) S. M. Fuchs, R. T. Raines, Biochem-
istry 2004, 43, 2438 – 2444; g) D. J. Mitchell, D. T. Kim, L.
Steinman, C. G. Fathman, J. B. Rothbard, J. Pept. Res. 2000, 56,
318 – 325.
[3] a) T. Suzuki, S. Futaki, M. Niwa, S. Tanaka, K. Ueda, Y. Sugiura,
J. Biol. Chem. 2002, 277, 2437 – 2443; b) P. E. G. Thorꢁn, 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.
[4] a) N. Sakai, S. Matile, J. Am. Chem. Soc. 2003, 125, 14348 –
14356; b) N. Sakai, S. Futaki, S. Matile, Soft Matter 2006, 2,
636 – 641; c) N. Sakai, T. Takeuchi, S. Futaki, S. Matile, Chem-
BioChem 2005, 6, 114 – 122; d) A. Hennig, G. J. Gabriel, G. N.
Tew, S. Matile, J. Am. Chem. Soc. 2008, 130, 10338 – 10344; e) A.
Pantos, I. Tsogas, C. A. Paleos, Biochim. Biophys. Acta Bio-
membr. 2008, 1778, 811 – 823.
[17] R. Talhout, J. B. F. N. Engberts, Org. Biomol. Chem. 2004, 2,
3071 – 3074.
Angew. Chem. Int. Ed. 2012, 51, 980 –983
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
983